Index: linux-5.6.17-rt10/Documentation/RCU/checklist.txt =================================================================== --- linux-5.6.17-rt10.orig/Documentation/RCU/checklist.txt +++ linux-5.6.17-rt10/Documentation/RCU/checklist.txt @@ -210,8 +210,8 @@ over a rather long period of time, but i the rest of the system. 7. As of v4.20, a given kernel implements only one RCU flavor, - which is RCU-sched for PREEMPT=n and RCU-preempt for PREEMPT=y. - If the updater uses call_rcu() or synchronize_rcu(), + which is RCU-sched for PREEMPTION=n and RCU-preempt for + PREEMPTION=y. If the updater uses call_rcu() or synchronize_rcu(), then the corresponding readers my use rcu_read_lock() and rcu_read_unlock(), rcu_read_lock_bh() and rcu_read_unlock_bh(), or any pair of primitives that disables and re-enables preemption, Index: linux-5.6.17-rt10/Documentation/RCU/stallwarn.txt =================================================================== --- linux-5.6.17-rt10.orig/Documentation/RCU/stallwarn.txt +++ linux-5.6.17-rt10/Documentation/RCU/stallwarn.txt @@ -20,7 +20,7 @@ o A CPU looping with preemption disabled o A CPU looping with bottom halves disabled. -o For !CONFIG_PREEMPT kernels, a CPU looping anywhere in the kernel +o For !CONFIG_PREEMPTION kernels, a CPU looping anywhere in the kernel without invoking schedule(). If the looping in the kernel is really expected and desirable behavior, you might need to add some calls to cond_resched(). @@ -39,7 +39,7 @@ o Anything that prevents RCU's grace-per result in the "rcu_.*kthread starved for" console-log message, which will include additional debugging information. -o A CPU-bound real-time task in a CONFIG_PREEMPT kernel, which might +o A CPU-bound real-time task in a CONFIG_PREEMPTION kernel, which might happen to preempt a low-priority task in the middle of an RCU read-side critical section. This is especially damaging if that low-priority task is not permitted to run on any other CPU, Index: linux-5.6.17-rt10/Documentation/admin-guide/sysctl/vm.rst =================================================================== --- linux-5.6.17-rt10.orig/Documentation/admin-guide/sysctl/vm.rst +++ linux-5.6.17-rt10/Documentation/admin-guide/sysctl/vm.rst @@ -128,6 +128,9 @@ allowed to examine the unevictable lru ( This should be used on systems where stalls for minor page faults are an acceptable trade for large contiguous free memory. Set to 0 to prevent compaction from moving pages that are unevictable. Default value is 1. +On CONFIG_PREEMPT_RT the default value is 0 in order to avoid a page fault, due +to compaction, which would block the task from becomming active until the fault +is resolved. dirty_background_bytes Index: linux-5.6.17-rt10/Documentation/locking/index.rst =================================================================== --- linux-5.6.17-rt10.orig/Documentation/locking/index.rst +++ linux-5.6.17-rt10/Documentation/locking/index.rst @@ -13,6 +13,7 @@ locking mutex-design rt-mutex-design rt-mutex + seqlock spinlocks ww-mutex-design Index: linux-5.6.17-rt10/Documentation/locking/seqlock.rst =================================================================== --- /dev/null +++ linux-5.6.17-rt10/Documentation/locking/seqlock.rst @@ -0,0 +1,242 @@ +====================================== +Sequence counters and sequential locks +====================================== + +Introduction +============ + +Sequence counters are a reader-writer consistency mechanism with +lockless readers (read-only retry loops), and no writer starvation. They +are used for data that's rarely written to (e.g. system time), where the +reader wants a consistent set of information and is willing to retry if +that information changes. + +A data set is consistent when the sequence count at the beginning of the +read side critical section is even and the same sequence count value is +read again at the end of the critical section. The data in the set must +be copied out inside the read side critical section. If the sequence +count has changed between the start and the end of the critical section, +the reader must retry. + +Writers increment the sequence count at the start and the end of their +critical section. After starting the critical section the sequence count +is odd and indicates to the readers that an update is in progress. At +the end of the write side critical section the sequence count becomes +even again which lets readers make progress. + +A sequence counter write side critical section must never be preempted +or interrupted by read side sections. Otherwise the reader will spin for +the entire scheduler tick due to the odd sequence count value and the +interrupted writer. If that reader belongs to a real-time scheduling +class, it can spin forever and the kernel will livelock. + +This mechanism cannot be used if the protected data contains pointers, +as the writer can invalidate a pointer that the reader is following. + +.. _seqcount_t: + +Sequence counters (:c:type:`seqcount_t`) +======================================== + +This is the the raw counting mechanism, which does not protect against +multiple writers. Write side critical sections must thus be serialized +by an external lock. + +If the write serialization primitive is not implicitly disabling +preemption, preemption must be explicitly disabled before entering the +write side section. If the read section can be invoked from hardirq or +softirq contexts, interrupts or bottom halves must also be respectively +disabled before entering the write section. + +If the write serialization mechanism is one of the common kernel locking +primitives, use :ref:`sequence counters with associated locks +` instead. If it's desired to automatically handle +the sequence counter writer serialization and non-preemptibility +requirements, use a :ref:`sequential lock `. + +Initialization: + +.. code-block:: c + + /* dynamic */ + seqcount_t foo_seqcount; + seqcount_init(&foo_seqcount); + + /* static */ + static seqcount_t foo_seqcount = SEQCNT_ZERO(foo_seqcount); + + /* C99 struct init */ + struct { + .seq = SEQCNT_ZERO(foo.seq), + } foo; + +Write path: + +.. _seqcount_write_ops: +.. code-block:: c + + /* Serialized context with disabled preemption */ + + write_seqcount_begin(&foo_seqcount); + + /* ... [[write-side critical section]] ... */ + + write_seqcount_end(&foo_seqcount); + +Read path: + +.. _seqcount_read_ops: +.. code-block:: c + + do { + seq = read_seqcount_begin(&foo_seqcount); + + /* ... [[read-side critical section]] ... */ + + } while (read_seqcount_retry(&foo_seqcount, seq)); + +.. _seqcount_locktype_t: + +Sequence counters with associated locks (:c:type:`seqcount_LOCKTYPE_t`) +----------------------------------------------------------------------- + +As :ref:`earlier discussed `, seqcount write side critical +sections must be serialized and non-preemptible. This variant of +sequence counters associate the lock used for writer serialization at +the seqcount initialization time. This enables lockdep to validate that +the write side critical section is properly serialized. + +This lock association is a NOOP if lockdep is disabled and has neither +storage nor runtime overhead. If lockdep is enabled, the lock pointer is +stored in struct seqcount and lockdep's "lock is held" assertions are +injected at the beginning of the write side critical section to validate +that it is properly protected. + +For lock types which do not implicitly disable preemption, preemption +protection is enforced in the write side function. + +The following seqcounts with associated locks are defined: + + - :c:type:`seqcount_spinlock_t` + - :c:type:`seqcount_raw_spinlock_t` + - :c:type:`seqcount_rwlock_t` + - :c:type:`seqcount_mutex_t` + - :c:type:`seqcount_ww_mutex_t` + +The plain seqcount read and write APIs branch out to the specific +seqcount_LOCKTYPE_t implementation at compile-time. This avoids kernel +API explosion per each new seqcount LOCKTYPE. + +Initialization (replace "LOCKTYPE" with one of the supported locks): + +.. code-block:: c + + /* dynamic */ + seqcount_LOCKTYPE_t foo_seqcount; + seqcount_LOCKTYPE_init(&foo_seqcount, &lock); + + /* static */ + static seqcount_LOCKTYPE_t foo_seqcount = + SEQCNT_LOCKTYPE_ZERO(foo_seqcount, &lock); + + /* C99 struct init */ + struct { + .seq = SEQCNT_LOCKTYPE_ZERO(foo.seq, &lock), + } foo; + +Write path: same as in :ref:`plain seqcount_t `, +while running from a context with the associated LOCKTYPE lock acquired. + +Read path: same as in :ref:`plain seqcount_t `. + +.. _seqlock_t: + +Sequential locks (:c:type:`seqlock_t`) +====================================== + +This contains the :ref:`sequence counting mechanism ` +earlier discussed, plus an embedded spinlock for writer serialization +and non-preemptibility. + +If the read side section can be invoked from hardirq or softirq context, +use the write side function variants which disable interrupts or bottom +halves respectively. + +Initialization: + +.. code-block:: c + + /* dynamic */ + seqlock_t foo_seqlock; + seqlock_init(&foo_seqlock); + + /* static */ + static DEFINE_SEQLOCK(foo_seqlock); + + /* C99 struct init */ + struct { + .seql = __SEQLOCK_UNLOCKED(foo.seql) + } foo; + +Write path: + +.. code-block:: c + + write_seqlock(&foo_seqlock); + + /* ... [[write-side critical section]] ... */ + + write_sequnlock(&foo_seqlock); + +Read path, three categories: + +1. Normal Sequence readers which never block a writer but they must + retry if a writer is in progress by detecting change in the sequence + number. Writers do not wait for a sequence reader. + + .. code-block:: c + + do { + seq = read_seqbegin(&foo_seqlock); + + /* ... [[read-side critical section]] ... */ + + } while (read_seqretry(&foo_seqlock, seq)); + +2. Locking readers which will wait if a writer or another locking reader + is in progress. A locking reader in progress will also block a writer + from entering its critical section. This read lock is + exclusive. Unlike rwlock_t, only one locking reader can acquire it. + + .. code-block:: c + + read_seqlock_excl(&foo_seqlock); + + /* ... [[read-side critical section]] ... */ + + read_sequnlock_excl(&foo_seqlock); + +3. Conditional lockless reader (as in 1), or locking reader (as in 2), + according to a passed marker. This is used to avoid lockless readers + starvation (too much retry loops) in case of a sharp spike in write + activity. First, a lockless read is tried (even marker passed). If + that trial fails (odd sequence counter is returned, which is used as + the next iteration marker), the lockless read is transformed to a + full locking read and no retry loop is necessary. + + .. code-block:: c + + /* marker; even initialization */ + int seq = 0; + do { + read_seqbegin_or_lock(&foo_seqlock, &seq); + + /* ... [[read-side critical section]] ... */ + + } while (need_seqretry(&foo_seqlock, seq)); + done_seqretry(&foo_seqlock, seq); + +API documentation +================= + +.. kernel-doc:: include/linux/seqlock.h Index: linux-5.6.17-rt10/Documentation/printk-ringbuffer.txt =================================================================== --- /dev/null +++ linux-5.6.17-rt10/Documentation/printk-ringbuffer.txt @@ -0,0 +1,377 @@ +struct printk_ringbuffer +------------------------ +John Ogness + +Overview +~~~~~~~~ +As the name suggests, this ring buffer was implemented specifically to serve +the needs of the printk() infrastructure. The ring buffer itself is not +specific to printk and could be used for other purposes. _However_, the +requirements and semantics of printk are rather unique. If you intend to use +this ring buffer for anything other than printk, you need to be very clear on +its features, behavior, and pitfalls. + +Features +^^^^^^^^ +The printk ring buffer has the following features: + +- single global buffer +- resides in initialized data section (available at early boot) +- lockless readers +- supports multiple writers +- supports multiple non-consuming readers +- safe from any context (including NMI) +- groups bytes into variable length blocks (referenced by entries) +- entries tagged with sequence numbers + +Behavior +^^^^^^^^ +Since the printk ring buffer readers are lockless, there exists no +synchronization between readers and writers. Basically writers are the tasks +in control and may overwrite any and all committed data at any time and from +any context. For this reason readers can miss entries if they are overwritten +before the reader was able to access the data. The reader API implementation +is such that reader access to entries is atomic, so there is no risk of +readers having to deal with partial or corrupt data. Also, entries are +tagged with sequence numbers so readers can recognize if entries were missed. + +Writing to the ring buffer consists of 2 steps. First a writer must reserve +an entry of desired size. After this step the writer has exclusive access +to the memory region. Once the data has been written to memory, it needs to +be committed to the ring buffer. After this step the entry has been inserted +into the ring buffer and assigned an appropriate sequence number. + +Once committed, a writer must no longer access the data directly. This is +because the data may have been overwritten and no longer exists. If a +writer must access the data, it should either keep a private copy before +committing the entry or use the reader API to gain access to the data. + +Because of how the data backend is implemented, entries that have been +reserved but not yet committed act as barriers, preventing future writers +from filling the ring buffer beyond the location of the reserved but not +yet committed entry region. For this reason it is *important* that writers +perform both reserve and commit as quickly as possible. Also, be aware that +preemption and local interrupts are disabled and writing to the ring buffer +is processor-reentrant locked during the reserve/commit window. Writers in +NMI contexts can still preempt any other writers, but as long as these +writers do not write a large amount of data with respect to the ring buffer +size, this should not become an issue. + +API +~~~ + +Declaration +^^^^^^^^^^^ +The printk ring buffer can be instantiated as a static structure: + + /* declare a static struct printk_ringbuffer */ + #define DECLARE_STATIC_PRINTKRB(name, szbits, cpulockptr) + +The value of szbits specifies the size of the ring buffer in bits. The +cpulockptr field is a pointer to a prb_cpulock struct that is used to +perform processor-reentrant spin locking for the writers. It is specified +externally because it may be used for multiple ring buffers (or other +code) to synchronize writers without risk of deadlock. + +Here is an example of a declaration of a printk ring buffer specifying a +32KB (2^15) ring buffer: + +.... +DECLARE_STATIC_PRINTKRB_CPULOCK(rb_cpulock); +DECLARE_STATIC_PRINTKRB(rb, 15, &rb_cpulock); +.... + +If writers will be using multiple ring buffers and the ordering of that usage +is not clear, the same prb_cpulock should be used for both ring buffers. + +Writer API +^^^^^^^^^^ +The writer API consists of 2 functions. The first is to reserve an entry in +the ring buffer, the second is to commit that data to the ring buffer. The +reserved entry information is stored within a provided `struct prb_handle`. + + /* reserve an entry */ + char *prb_reserve(struct prb_handle *h, struct printk_ringbuffer *rb, + unsigned int size); + + /* commit a reserved entry to the ring buffer */ + void prb_commit(struct prb_handle *h); + +Here is an example of a function to write data to a ring buffer: + +.... +int write_data(struct printk_ringbuffer *rb, char *data, int size) +{ + struct prb_handle h; + char *buf; + + buf = prb_reserve(&h, rb, size); + if (!buf) + return -1; + memcpy(buf, data, size); + prb_commit(&h); + + return 0; +} +.... + +Pitfalls +++++++++ +Be aware that prb_reserve() can fail. A retry might be successful, but it +depends entirely on whether or not the next part of the ring buffer to +overwrite belongs to reserved but not yet committed entries of other writers. +Writers can use the prb_inc_lost() function to allow readers to notice that a +message was lost. + +Reader API +^^^^^^^^^^ +The reader API utilizes a `struct prb_iterator` to track the reader's +position in the ring buffer. + + /* declare a pre-initialized static iterator for a ring buffer */ + #define DECLARE_STATIC_PRINTKRB_ITER(name, rbaddr) + + /* initialize iterator for a ring buffer (if static macro NOT used) */ + void prb_iter_init(struct prb_iterator *iter, + struct printk_ringbuffer *rb, u64 *seq); + + /* make a deep copy of an iterator */ + void prb_iter_copy(struct prb_iterator *dest, + struct prb_iterator *src); + + /* non-blocking, advance to next entry (and read the data) */ + int prb_iter_next(struct prb_iterator *iter, char *buf, + int size, u64 *seq); + + /* blocking, advance to next entry (and read the data) */ + int prb_iter_wait_next(struct prb_iterator *iter, char *buf, + int size, u64 *seq); + + /* position iterator at the entry seq */ + int prb_iter_seek(struct prb_iterator *iter, u64 seq); + + /* read data at current position */ + int prb_iter_data(struct prb_iterator *iter, char *buf, + int size, u64 *seq); + +Typically prb_iter_data() is not needed because the data can be retrieved +directly with prb_iter_next(). + +Here is an example of a non-blocking function that will read all the data in +a ring buffer: + +.... +void read_all_data(struct printk_ringbuffer *rb, char *buf, int size) +{ + struct prb_iterator iter; + u64 prev_seq = 0; + u64 seq; + int ret; + + prb_iter_init(&iter, rb, NULL); + + for (;;) { + ret = prb_iter_next(&iter, buf, size, &seq); + if (ret > 0) { + if (seq != ++prev_seq) { + /* "seq - prev_seq" entries missed */ + prev_seq = seq; + } + /* process buf here */ + } else if (ret == 0) { + /* hit the end, done */ + break; + } else if (ret < 0) { + /* + * iterator is invalid, a writer overtook us, reset the + * iterator and keep going, entries were missed + */ + prb_iter_init(&iter, rb, NULL); + } + } +} +.... + +Pitfalls +++++++++ +The reader's iterator can become invalid at any time because the reader was +overtaken by a writer. Typically the reader should reset the iterator back +to the current oldest entry (which will be newer than the entry the reader +was at) and continue, noting the number of entries that were missed. + +Utility API +^^^^^^^^^^^ +Several functions are available as convenience for external code. + + /* query the size of the data buffer */ + int prb_buffer_size(struct printk_ringbuffer *rb); + + /* skip a seq number to signify a lost record */ + void prb_inc_lost(struct printk_ringbuffer *rb); + + /* processor-reentrant spin lock */ + void prb_lock(struct prb_cpulock *cpu_lock, unsigned int *cpu_store); + + /* processor-reentrant spin unlock */ + void prb_lock(struct prb_cpulock *cpu_lock, unsigned int *cpu_store); + +Pitfalls +++++++++ +Although the value returned by prb_buffer_size() does represent an absolute +upper bound, the amount of data that can be stored within the ring buffer +is actually less because of the additional storage space of a header for each +entry. + +The prb_lock() and prb_unlock() functions can be used to synchronize between +ring buffer writers and other external activities. The function of a +processor-reentrant spin lock is to disable preemption and local interrupts +and synchronize against other processors. It does *not* protect against +multiple contexts of a single processor, i.e NMI. + +Implementation +~~~~~~~~~~~~~~ +This section describes several of the implementation concepts and details to +help developers better understand the code. + +Entries +^^^^^^^ +All ring buffer data is stored within a single static byte array. The reason +for this is to ensure that any pointers to the data (past and present) will +always point to valid memory. This is important because the lockless readers +may be preempted for long periods of time and when they resume may be working +with expired pointers. + +Entries are identified by start index and size. (The start index plus size +is the start index of the next entry.) The start index is not simply an +offset into the byte array, but rather a logical position (lpos) that maps +directly to byte array offsets. + +For example, for a byte array of 1000, an entry may have have a start index +of 100. Another entry may have a start index of 1100. And yet another 2100. +All of these entry are pointing to the same memory region, but only the most +recent entry is valid. The other entries are pointing to valid memory, but +represent entries that have been overwritten. + +Note that due to overflowing, the most recent entry is not necessarily the one +with the highest lpos value. Indeed, the printk ring buffer initializes its +data such that an overflow happens relatively quickly in order to validate the +handling of this situation. The implementation assumes that an lpos (unsigned +long) will never completely wrap while a reader is preempted. If this were to +become an issue, the seq number (which never wraps) could be used to increase +the robustness of handling this situation. + +Buffer Wrapping +^^^^^^^^^^^^^^^ +If an entry starts near the end of the byte array but would extend beyond it, +a special terminating entry (size = -1) is inserted into the byte array and +the real entry is placed at the beginning of the byte array. This can waste +space at the end of the byte array, but simplifies the implementation by +allowing writers to always work with contiguous buffers. + +Note that the size field is the first 4 bytes of the entry header. Also note +that calc_next() always ensures that there are at least 4 bytes left at the +end of the byte array to allow room for a terminating entry. + +Ring Buffer Pointers +^^^^^^^^^^^^^^^^^^^^ +Three pointers (lpos values) are used to manage the ring buffer: + + - _tail_: points to the oldest entry + - _head_: points to where the next new committed entry will be + - _reserve_: points to where the next new reserved entry will be + +These pointers always maintain a logical ordering: + + tail <= head <= reserve + +The reserve pointer moves forward when a writer reserves a new entry. The +head pointer moves forward when a writer commits a new entry. + +The reserve pointer cannot overwrite the tail pointer in a wrap situation. In +such a situation, the tail pointer must be "pushed forward", thus +invalidating that oldest entry. Readers identify if they are accessing a +valid entry by ensuring their entry pointer is `>= tail && < head`. + +If the tail pointer is equal to the head pointer, it cannot be pushed and any +reserve operation will fail. The only resolution is for writers to commit +their reserved entries. + +Processor-Reentrant Locking +^^^^^^^^^^^^^^^^^^^^^^^^^^^ +The purpose of the processor-reentrant locking is to limit the interruption +scenarios of writers to 2 contexts. This allows for a simplified +implementation where: + +- The reserve/commit window only exists on 1 processor at a time. A reserve + can never fail due to uncommitted entries of other processors. + +- When committing entries, it is trivial to handle the situation when + subsequent entries have already been committed, i.e. managing the head + pointer. + +Performance +~~~~~~~~~~~ +Some basic tests were performed on a quad Intel(R) Xeon(R) CPU E5-2697 v4 at +2.30GHz (36 cores / 72 threads). All tests involved writing a total of +32,000,000 records at an average of 33 bytes each. Each writer was pinned to +its own CPU and would write as fast as it could until a total of 32,000,000 +records were written. All tests involved 2 readers that were both pinned +together to another CPU. Each reader would read as fast as it could and track +how many of the 32,000,000 records it could read. All tests used a ring buffer +of 16KB in size, which holds around 350 records (header + data for each +entry). + +The only difference between the tests is the number of writers (and thus also +the number of records per writer). As more writers are added, the time to +write a record increases. This is because data pointers, modified via cmpxchg, +and global data access in general become more contended. + +1 writer +^^^^^^^^ + runtime: 0m 18s + reader1: 16219900/32000000 (50%) records + reader2: 16141582/32000000 (50%) records + +2 writers +^^^^^^^^^ + runtime: 0m 32s + reader1: 16327957/32000000 (51%) records + reader2: 16313988/32000000 (50%) records + +4 writers +^^^^^^^^^ + runtime: 0m 42s + reader1: 16421642/32000000 (51%) records + reader2: 16417224/32000000 (51%) records + +8 writers +^^^^^^^^^ + runtime: 0m 43s + reader1: 16418300/32000000 (51%) records + reader2: 16432222/32000000 (51%) records + +16 writers +^^^^^^^^^^ + runtime: 0m 54s + reader1: 16539189/32000000 (51%) records + reader2: 16542711/32000000 (51%) records + +32 writers +^^^^^^^^^^ + runtime: 1m 13s + reader1: 16731808/32000000 (52%) records + reader2: 16735119/32000000 (52%) records + +Comments +^^^^^^^^ +It is particularly interesting to compare/contrast the 1-writer and 32-writer +tests. Despite the writing of the 32,000,000 records taking over 4 times +longer, the readers (which perform no cmpxchg) were still unable to keep up. +This shows that the memory contention between the increasing number of CPUs +also has a dramatic effect on readers. + +It should also be noted that in all cases each reader was able to read >=50% +of the records. This means that a single reader would have been able to keep +up with the writer(s) in all cases, becoming slightly easier as more writers +are added. This was the purpose of pinning 2 readers to 1 CPU: to observe how +maximum reader performance changes. Index: linux-5.6.17-rt10/MAINTAINERS =================================================================== --- linux-5.6.17-rt10.orig/MAINTAINERS +++ linux-5.6.17-rt10/MAINTAINERS @@ -9821,7 +9821,7 @@ F: arch/*/include/asm/spinlock*.h F: include/linux/rwlock*.h F: include/linux/mutex*.h F: include/linux/rwsem*.h -F: include/linux/seqlock.h +F: include/linux/seqlock*.h F: lib/locking*.[ch] F: kernel/locking/ X: kernel/locking/locktorture.c Index: linux-5.6.17-rt10/arch/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/arch/Kconfig +++ linux-5.6.17-rt10/arch/Kconfig @@ -31,6 +31,7 @@ config OPROFILE tristate "OProfile system profiling" depends on PROFILING depends on HAVE_OPROFILE + depends on !PREEMPT_RT select RING_BUFFER select RING_BUFFER_ALLOW_SWAP help Index: linux-5.6.17-rt10/arch/alpha/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/alpha/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/alpha/include/asm/spinlock_types.h @@ -2,10 +2,6 @@ #ifndef _ALPHA_SPINLOCK_TYPES_H #define _ALPHA_SPINLOCK_TYPES_H -#ifndef __LINUX_SPINLOCK_TYPES_H -# error "please don't include this file directly" -#endif - typedef struct { volatile unsigned int lock; } arch_spinlock_t; Index: linux-5.6.17-rt10/arch/arm/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/Kconfig +++ linux-5.6.17-rt10/arch/arm/Kconfig @@ -32,6 +32,7 @@ config ARM select ARCH_OPTIONAL_KERNEL_RWX if ARCH_HAS_STRICT_KERNEL_RWX select ARCH_OPTIONAL_KERNEL_RWX_DEFAULT if CPU_V7 select ARCH_SUPPORTS_ATOMIC_RMW + select ARCH_SUPPORTS_RT select ARCH_USE_BUILTIN_BSWAP select ARCH_USE_CMPXCHG_LOCKREF select ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT if MMU @@ -64,7 +65,7 @@ config ARM select HARDIRQS_SW_RESEND select HAVE_ARCH_AUDITSYSCALL if AEABI && !OABI_COMPAT select HAVE_ARCH_BITREVERSE if (CPU_32v7M || CPU_32v7) && !CPU_32v6 - select HAVE_ARCH_JUMP_LABEL if !XIP_KERNEL && !CPU_ENDIAN_BE32 && MMU + select HAVE_ARCH_JUMP_LABEL if !XIP_KERNEL && !CPU_ENDIAN_BE32 && MMU && !PREEMPT_RT select HAVE_ARCH_KGDB if !CPU_ENDIAN_BE32 && MMU select HAVE_ARCH_MMAP_RND_BITS if MMU select HAVE_ARCH_SECCOMP_FILTER if AEABI && !OABI_COMPAT @@ -103,6 +104,7 @@ config ARM select HAVE_PERF_EVENTS select HAVE_PERF_REGS select HAVE_PERF_USER_STACK_DUMP + select HAVE_PREEMPT_LAZY select MMU_GATHER_RCU_TABLE_FREE if SMP && ARM_LPAE select HAVE_REGS_AND_STACK_ACCESS_API select HAVE_RSEQ Index: linux-5.6.17-rt10/arch/arm/include/asm/irq.h =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/include/asm/irq.h +++ linux-5.6.17-rt10/arch/arm/include/asm/irq.h @@ -23,6 +23,8 @@ #endif #ifndef __ASSEMBLY__ +#include + struct irqaction; struct pt_regs; Index: linux-5.6.17-rt10/arch/arm/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/arm/include/asm/spinlock_types.h @@ -2,10 +2,6 @@ #ifndef __ASM_SPINLOCK_TYPES_H #define __ASM_SPINLOCK_TYPES_H -#ifndef __LINUX_SPINLOCK_TYPES_H -# error "please don't include this file directly" -#endif - #define TICKET_SHIFT 16 typedef struct { Index: linux-5.6.17-rt10/arch/arm/include/asm/switch_to.h =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/include/asm/switch_to.h +++ linux-5.6.17-rt10/arch/arm/include/asm/switch_to.h @@ -4,6 +4,13 @@ #include +#if defined CONFIG_PREEMPT_RT && defined CONFIG_HIGHMEM +void switch_kmaps(struct task_struct *prev_p, struct task_struct *next_p); +#else +static inline void +switch_kmaps(struct task_struct *prev_p, struct task_struct *next_p) { } +#endif + /* * For v7 SMP cores running a preemptible kernel we may be pre-empted * during a TLB maintenance operation, so execute an inner-shareable dsb @@ -26,6 +33,7 @@ extern struct task_struct *__switch_to(s #define switch_to(prev,next,last) \ do { \ __complete_pending_tlbi(); \ + switch_kmaps(prev, next); \ last = __switch_to(prev,task_thread_info(prev), task_thread_info(next)); \ } while (0) Index: linux-5.6.17-rt10/arch/arm/include/asm/thread_info.h =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/include/asm/thread_info.h +++ linux-5.6.17-rt10/arch/arm/include/asm/thread_info.h @@ -46,6 +46,7 @@ struct cpu_context_save { struct thread_info { unsigned long flags; /* low level flags */ int preempt_count; /* 0 => preemptable, <0 => bug */ + int preempt_lazy_count; /* 0 => preemptable, <0 => bug */ mm_segment_t addr_limit; /* address limit */ struct task_struct *task; /* main task structure */ __u32 cpu; /* cpu */ @@ -139,7 +140,8 @@ extern int vfp_restore_user_hwstate(stru #define TIF_SYSCALL_TRACE 4 /* syscall trace active */ #define TIF_SYSCALL_AUDIT 5 /* syscall auditing active */ #define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */ -#define TIF_SECCOMP 7 /* seccomp syscall filtering active */ +#define TIF_SECCOMP 8 /* seccomp syscall filtering active */ +#define TIF_NEED_RESCHED_LAZY 7 #define TIF_NOHZ 12 /* in adaptive nohz mode */ #define TIF_USING_IWMMXT 17 @@ -149,6 +151,7 @@ extern int vfp_restore_user_hwstate(stru #define _TIF_SIGPENDING (1 << TIF_SIGPENDING) #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) +#define _TIF_NEED_RESCHED_LAZY (1 << TIF_NEED_RESCHED_LAZY) #define _TIF_UPROBE (1 << TIF_UPROBE) #define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT) @@ -164,7 +167,8 @@ extern int vfp_restore_user_hwstate(stru * Change these and you break ASM code in entry-common.S */ #define _TIF_WORK_MASK (_TIF_NEED_RESCHED | _TIF_SIGPENDING | \ - _TIF_NOTIFY_RESUME | _TIF_UPROBE) + _TIF_NOTIFY_RESUME | _TIF_UPROBE | \ + _TIF_NEED_RESCHED_LAZY) #endif /* __KERNEL__ */ #endif /* __ASM_ARM_THREAD_INFO_H */ Index: linux-5.6.17-rt10/arch/arm/kernel/asm-offsets.c =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/kernel/asm-offsets.c +++ linux-5.6.17-rt10/arch/arm/kernel/asm-offsets.c @@ -53,6 +53,7 @@ int main(void) BLANK(); DEFINE(TI_FLAGS, offsetof(struct thread_info, flags)); DEFINE(TI_PREEMPT, offsetof(struct thread_info, preempt_count)); + DEFINE(TI_PREEMPT_LAZY, offsetof(struct thread_info, preempt_lazy_count)); DEFINE(TI_ADDR_LIMIT, offsetof(struct thread_info, addr_limit)); DEFINE(TI_TASK, offsetof(struct thread_info, task)); DEFINE(TI_CPU, offsetof(struct thread_info, cpu)); Index: linux-5.6.17-rt10/arch/arm/kernel/entry-armv.S =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/kernel/entry-armv.S +++ linux-5.6.17-rt10/arch/arm/kernel/entry-armv.S @@ -206,11 +206,18 @@ __irq_svc: #ifdef CONFIG_PREEMPTION ldr r8, [tsk, #TI_PREEMPT] @ get preempt count - ldr r0, [tsk, #TI_FLAGS] @ get flags teq r8, #0 @ if preempt count != 0 + bne 1f @ return from exeption + ldr r0, [tsk, #TI_FLAGS] @ get flags + tst r0, #_TIF_NEED_RESCHED @ if NEED_RESCHED is set + blne svc_preempt @ preempt! + + ldr r8, [tsk, #TI_PREEMPT_LAZY] @ get preempt lazy count + teq r8, #0 @ if preempt lazy count != 0 movne r0, #0 @ force flags to 0 - tst r0, #_TIF_NEED_RESCHED + tst r0, #_TIF_NEED_RESCHED_LAZY blne svc_preempt +1: #endif svc_exit r5, irq = 1 @ return from exception @@ -225,8 +232,14 @@ svc_preempt: 1: bl preempt_schedule_irq @ irq en/disable is done inside ldr r0, [tsk, #TI_FLAGS] @ get new tasks TI_FLAGS tst r0, #_TIF_NEED_RESCHED + bne 1b + tst r0, #_TIF_NEED_RESCHED_LAZY reteq r8 @ go again - b 1b + ldr r0, [tsk, #TI_PREEMPT_LAZY] @ get preempt lazy count + teq r0, #0 @ if preempt lazy count != 0 + beq 1b + ret r8 @ go again + #endif __und_fault: Index: linux-5.6.17-rt10/arch/arm/kernel/entry-common.S =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/kernel/entry-common.S +++ linux-5.6.17-rt10/arch/arm/kernel/entry-common.S @@ -53,7 +53,9 @@ __ret_fast_syscall: cmp r2, #TASK_SIZE blne addr_limit_check_failed ldr r1, [tsk, #TI_FLAGS] @ re-check for syscall tracing - tst r1, #_TIF_SYSCALL_WORK | _TIF_WORK_MASK + tst r1, #((_TIF_SYSCALL_WORK | _TIF_WORK_MASK) & ~_TIF_SECCOMP) + bne fast_work_pending + tst r1, #_TIF_SECCOMP bne fast_work_pending @@ -90,8 +92,11 @@ __ret_fast_syscall: cmp r2, #TASK_SIZE blne addr_limit_check_failed ldr r1, [tsk, #TI_FLAGS] @ re-check for syscall tracing - tst r1, #_TIF_SYSCALL_WORK | _TIF_WORK_MASK + tst r1, #((_TIF_SYSCALL_WORK | _TIF_WORK_MASK) & ~_TIF_SECCOMP) + bne do_slower_path + tst r1, #_TIF_SECCOMP beq no_work_pending +do_slower_path: UNWIND(.fnend ) ENDPROC(ret_fast_syscall) Index: linux-5.6.17-rt10/arch/arm/kernel/signal.c =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/kernel/signal.c +++ linux-5.6.17-rt10/arch/arm/kernel/signal.c @@ -649,7 +649,8 @@ do_work_pending(struct pt_regs *regs, un */ trace_hardirqs_off(); do { - if (likely(thread_flags & _TIF_NEED_RESCHED)) { + if (likely(thread_flags & (_TIF_NEED_RESCHED | + _TIF_NEED_RESCHED_LAZY))) { schedule(); } else { if (unlikely(!user_mode(regs))) Index: linux-5.6.17-rt10/arch/arm/kernel/smp.c =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/kernel/smp.c +++ linux-5.6.17-rt10/arch/arm/kernel/smp.c @@ -682,11 +682,9 @@ void handle_IPI(int ipinr, struct pt_reg break; case IPI_CPU_BACKTRACE: - printk_nmi_enter(); irq_enter(); nmi_cpu_backtrace(regs); irq_exit(); - printk_nmi_exit(); break; default: Index: linux-5.6.17-rt10/arch/arm/mm/fault.c =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/mm/fault.c +++ linux-5.6.17-rt10/arch/arm/mm/fault.c @@ -414,6 +414,9 @@ do_translation_fault(unsigned long addr, if (addr < TASK_SIZE) return do_page_fault(addr, fsr, regs); + if (interrupts_enabled(regs)) + local_irq_enable(); + if (user_mode(regs)) goto bad_area; @@ -481,6 +484,9 @@ do_translation_fault(unsigned long addr, static int do_sect_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs) { + if (interrupts_enabled(regs)) + local_irq_enable(); + do_bad_area(addr, fsr, regs); return 0; } Index: linux-5.6.17-rt10/arch/arm/mm/highmem.c =================================================================== --- linux-5.6.17-rt10.orig/arch/arm/mm/highmem.c +++ linux-5.6.17-rt10/arch/arm/mm/highmem.c @@ -31,6 +31,11 @@ static inline pte_t get_fixmap_pte(unsig return *ptep; } +static unsigned int fixmap_idx(int type) +{ + return FIX_KMAP_BEGIN + type + KM_TYPE_NR * smp_processor_id(); +} + void *kmap(struct page *page) { might_sleep(); @@ -51,12 +56,13 @@ EXPORT_SYMBOL(kunmap); void *kmap_atomic(struct page *page) { + pte_t pte = mk_pte(page, kmap_prot); unsigned int idx; unsigned long vaddr; void *kmap; int type; - preempt_disable(); + preempt_disable_nort(); pagefault_disable(); if (!PageHighMem(page)) return page_address(page); @@ -76,7 +82,7 @@ void *kmap_atomic(struct page *page) type = kmap_atomic_idx_push(); - idx = FIX_KMAP_BEGIN + type + KM_TYPE_NR * smp_processor_id(); + idx = fixmap_idx(type); vaddr = __fix_to_virt(idx); #ifdef CONFIG_DEBUG_HIGHMEM /* @@ -90,7 +96,10 @@ void *kmap_atomic(struct page *page) * in place, so the contained TLB flush ensures the TLB is updated * with the new mapping. */ - set_fixmap_pte(idx, mk_pte(page, kmap_prot)); +#ifdef CONFIG_PREEMPT_RT + current->kmap_pte[type] = pte; +#endif + set_fixmap_pte(idx, pte); return (void *)vaddr; } @@ -103,44 +112,75 @@ void __kunmap_atomic(void *kvaddr) if (kvaddr >= (void *)FIXADDR_START) { type = kmap_atomic_idx(); - idx = FIX_KMAP_BEGIN + type + KM_TYPE_NR * smp_processor_id(); + idx = fixmap_idx(type); if (cache_is_vivt()) __cpuc_flush_dcache_area((void *)vaddr, PAGE_SIZE); +#ifdef CONFIG_PREEMPT_RT + current->kmap_pte[type] = __pte(0); +#endif #ifdef CONFIG_DEBUG_HIGHMEM BUG_ON(vaddr != __fix_to_virt(idx)); - set_fixmap_pte(idx, __pte(0)); #else (void) idx; /* to kill a warning */ #endif + set_fixmap_pte(idx, __pte(0)); kmap_atomic_idx_pop(); } else if (vaddr >= PKMAP_ADDR(0) && vaddr < PKMAP_ADDR(LAST_PKMAP)) { /* this address was obtained through kmap_high_get() */ kunmap_high(pte_page(pkmap_page_table[PKMAP_NR(vaddr)])); } pagefault_enable(); - preempt_enable(); + preempt_enable_nort(); } EXPORT_SYMBOL(__kunmap_atomic); void *kmap_atomic_pfn(unsigned long pfn) { + pte_t pte = pfn_pte(pfn, kmap_prot); unsigned long vaddr; int idx, type; struct page *page = pfn_to_page(pfn); - preempt_disable(); + preempt_disable_nort(); pagefault_disable(); if (!PageHighMem(page)) return page_address(page); type = kmap_atomic_idx_push(); - idx = FIX_KMAP_BEGIN + type + KM_TYPE_NR * smp_processor_id(); + idx = fixmap_idx(type); vaddr = __fix_to_virt(idx); #ifdef CONFIG_DEBUG_HIGHMEM BUG_ON(!pte_none(get_fixmap_pte(vaddr))); #endif - set_fixmap_pte(idx, pfn_pte(pfn, kmap_prot)); +#ifdef CONFIG_PREEMPT_RT + current->kmap_pte[type] = pte; +#endif + set_fixmap_pte(idx, pte); return (void *)vaddr; } +#if defined CONFIG_PREEMPT_RT +void switch_kmaps(struct task_struct *prev_p, struct task_struct *next_p) +{ + int i; + + /* + * Clear @prev's kmap_atomic mappings + */ + for (i = 0; i < prev_p->kmap_idx; i++) { + int idx = fixmap_idx(i); + + set_fixmap_pte(idx, __pte(0)); + } + /* + * Restore @next_p's kmap_atomic mappings + */ + for (i = 0; i < next_p->kmap_idx; i++) { + int idx = fixmap_idx(i); + + if (!pte_none(next_p->kmap_pte[i])) + set_fixmap_pte(idx, next_p->kmap_pte[i]); + } +} +#endif Index: linux-5.6.17-rt10/arch/arm64/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/arch/arm64/Kconfig +++ linux-5.6.17-rt10/arch/arm64/Kconfig @@ -69,6 +69,7 @@ config ARM64 select ARCH_SUPPORTS_ATOMIC_RMW select ARCH_SUPPORTS_INT128 if CC_HAS_INT128 && (GCC_VERSION >= 50000 || CC_IS_CLANG) select ARCH_SUPPORTS_NUMA_BALANCING + select ARCH_SUPPORTS_RT select ARCH_WANT_COMPAT_IPC_PARSE_VERSION if COMPAT select ARCH_WANT_DEFAULT_BPF_JIT select ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT @@ -163,6 +164,7 @@ config ARM64 select HAVE_PERF_EVENTS select HAVE_PERF_REGS select HAVE_PERF_USER_STACK_DUMP + select HAVE_PREEMPT_LAZY select HAVE_REGS_AND_STACK_ACCESS_API select HAVE_FUNCTION_ARG_ACCESS_API select HAVE_FUTEX_CMPXCHG if FUTEX Index: linux-5.6.17-rt10/arch/arm64/include/asm/preempt.h =================================================================== --- linux-5.6.17-rt10.orig/arch/arm64/include/asm/preempt.h +++ linux-5.6.17-rt10/arch/arm64/include/asm/preempt.h @@ -70,13 +70,36 @@ static inline bool __preempt_count_dec_a * interrupt occurring between the non-atomic READ_ONCE/WRITE_ONCE * pair. */ - return !pc || !READ_ONCE(ti->preempt_count); + if (!pc || !READ_ONCE(ti->preempt_count)) + return true; +#ifdef CONFIG_PREEMPT_LAZY + if ((pc & ~PREEMPT_NEED_RESCHED)) + return false; + if (current_thread_info()->preempt_lazy_count) + return false; + return test_thread_flag(TIF_NEED_RESCHED_LAZY); +#else + return false; +#endif } static inline bool should_resched(int preempt_offset) { +#ifdef CONFIG_PREEMPT_LAZY + u64 pc = READ_ONCE(current_thread_info()->preempt_count); + if (pc == preempt_offset) + return true; + + if ((pc & ~PREEMPT_NEED_RESCHED) != preempt_offset) + return false; + + if (current_thread_info()->preempt_lazy_count) + return false; + return test_thread_flag(TIF_NEED_RESCHED_LAZY); +#else u64 pc = READ_ONCE(current_thread_info()->preempt_count); return pc == preempt_offset; +#endif } #ifdef CONFIG_PREEMPTION Index: linux-5.6.17-rt10/arch/arm64/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/arm64/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/arm64/include/asm/spinlock_types.h @@ -5,10 +5,6 @@ #ifndef __ASM_SPINLOCK_TYPES_H #define __ASM_SPINLOCK_TYPES_H -#if !defined(__LINUX_SPINLOCK_TYPES_H) && !defined(__ASM_SPINLOCK_H) -# error "please don't include this file directly" -#endif - #include #include Index: linux-5.6.17-rt10/arch/arm64/include/asm/thread_info.h =================================================================== --- linux-5.6.17-rt10.orig/arch/arm64/include/asm/thread_info.h +++ linux-5.6.17-rt10/arch/arm64/include/asm/thread_info.h @@ -29,6 +29,7 @@ struct thread_info { #ifdef CONFIG_ARM64_SW_TTBR0_PAN u64 ttbr0; /* saved TTBR0_EL1 */ #endif + int preempt_lazy_count; /* 0 => preemptable, <0 => bug */ union { u64 preempt_count; /* 0 => preemptible, <0 => bug */ struct { @@ -63,6 +64,7 @@ void arch_release_task_struct(struct tas #define TIF_FOREIGN_FPSTATE 3 /* CPU's FP state is not current's */ #define TIF_UPROBE 4 /* uprobe breakpoint or singlestep */ #define TIF_FSCHECK 5 /* Check FS is USER_DS on return */ +#define TIF_NEED_RESCHED_LAZY 6 #define TIF_NOHZ 7 #define TIF_SYSCALL_TRACE 8 /* syscall trace active */ #define TIF_SYSCALL_AUDIT 9 /* syscall auditing */ @@ -83,6 +85,7 @@ void arch_release_task_struct(struct tas #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) #define _TIF_FOREIGN_FPSTATE (1 << TIF_FOREIGN_FPSTATE) +#define _TIF_NEED_RESCHED_LAZY (1 << TIF_NEED_RESCHED_LAZY) #define _TIF_NOHZ (1 << TIF_NOHZ) #define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT) @@ -96,8 +99,9 @@ void arch_release_task_struct(struct tas #define _TIF_WORK_MASK (_TIF_NEED_RESCHED | _TIF_SIGPENDING | \ _TIF_NOTIFY_RESUME | _TIF_FOREIGN_FPSTATE | \ - _TIF_UPROBE | _TIF_FSCHECK) + _TIF_UPROBE | _TIF_FSCHECK | _TIF_NEED_RESCHED_LAZY) +#define _TIF_NEED_RESCHED_MASK (_TIF_NEED_RESCHED | _TIF_NEED_RESCHED_LAZY) #define _TIF_SYSCALL_WORK (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | \ _TIF_SYSCALL_TRACEPOINT | _TIF_SECCOMP | \ _TIF_NOHZ | _TIF_SYSCALL_EMU) Index: linux-5.6.17-rt10/arch/arm64/kernel/asm-offsets.c =================================================================== --- linux-5.6.17-rt10.orig/arch/arm64/kernel/asm-offsets.c +++ linux-5.6.17-rt10/arch/arm64/kernel/asm-offsets.c @@ -30,6 +30,7 @@ int main(void) BLANK(); DEFINE(TSK_TI_FLAGS, offsetof(struct task_struct, thread_info.flags)); DEFINE(TSK_TI_PREEMPT, offsetof(struct task_struct, thread_info.preempt_count)); + DEFINE(TSK_TI_PREEMPT_LAZY, offsetof(struct task_struct, thread_info.preempt_lazy_count)); DEFINE(TSK_TI_ADDR_LIMIT, offsetof(struct task_struct, thread_info.addr_limit)); #ifdef CONFIG_ARM64_SW_TTBR0_PAN DEFINE(TSK_TI_TTBR0, offsetof(struct task_struct, thread_info.ttbr0)); Index: linux-5.6.17-rt10/arch/arm64/kernel/entry.S =================================================================== --- linux-5.6.17-rt10.orig/arch/arm64/kernel/entry.S +++ linux-5.6.17-rt10/arch/arm64/kernel/entry.S @@ -612,9 +612,18 @@ alternative_if ARM64_HAS_IRQ_PRIO_MASKIN mrs x0, daif orr x24, x24, x0 alternative_else_nop_endif - cbnz x24, 1f // preempt count != 0 || NMI return path - bl arm64_preempt_schedule_irq // irq en/disable is done inside + + cbz x24, 1f // (need_resched + count) == 0 + cbnz w24, 2f // count != 0 + + ldr w24, [tsk, #TSK_TI_PREEMPT_LAZY] // get preempt lazy count + cbnz w24, 2f // preempt lazy count != 0 + + ldr x0, [tsk, #TSK_TI_FLAGS] // get flags + tbz x0, #TIF_NEED_RESCHED_LAZY, 2f // needs rescheduling? 1: + bl arm64_preempt_schedule_irq // irq en/disable is done inside +2: #endif #ifdef CONFIG_ARM64_PSEUDO_NMI Index: linux-5.6.17-rt10/arch/arm64/kernel/fpsimd.c =================================================================== --- linux-5.6.17-rt10.orig/arch/arm64/kernel/fpsimd.c +++ linux-5.6.17-rt10/arch/arm64/kernel/fpsimd.c @@ -213,6 +213,16 @@ static void sve_free(struct task_struct __sve_free(task); } +static void *sve_free_atomic(struct task_struct *task) +{ + void *sve_state = task->thread.sve_state; + + WARN_ON(test_tsk_thread_flag(task, TIF_SVE)); + + task->thread.sve_state = NULL; + return sve_state; +} + /* * TIF_SVE controls whether a task can use SVE without trapping while * in userspace, and also the way a task's FPSIMD/SVE state is stored @@ -1010,6 +1020,7 @@ void fpsimd_thread_switch(struct task_st void fpsimd_flush_thread(void) { int vl, supported_vl; + void *mem = NULL; if (!system_supports_fpsimd()) return; @@ -1022,7 +1033,7 @@ void fpsimd_flush_thread(void) if (system_supports_sve()) { clear_thread_flag(TIF_SVE); - sve_free(current); + mem = sve_free_atomic(current); /* * Reset the task vector length as required. @@ -1056,6 +1067,7 @@ void fpsimd_flush_thread(void) } put_cpu_fpsimd_context(); + kfree(mem); } /* Index: linux-5.6.17-rt10/arch/arm64/kernel/signal.c =================================================================== --- linux-5.6.17-rt10.orig/arch/arm64/kernel/signal.c +++ linux-5.6.17-rt10/arch/arm64/kernel/signal.c @@ -912,7 +912,7 @@ asmlinkage void do_notify_resume(struct /* Check valid user FS if needed */ addr_limit_user_check(); - if (thread_flags & _TIF_NEED_RESCHED) { + if (thread_flags & _TIF_NEED_RESCHED_MASK) { /* Unmask Debug and SError for the next task */ local_daif_restore(DAIF_PROCCTX_NOIRQ); Index: linux-5.6.17-rt10/arch/hexagon/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/hexagon/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/hexagon/include/asm/spinlock_types.h @@ -8,10 +8,6 @@ #ifndef _ASM_SPINLOCK_TYPES_H #define _ASM_SPINLOCK_TYPES_H -#ifndef __LINUX_SPINLOCK_TYPES_H -# error "please don't include this file directly" -#endif - typedef struct { volatile unsigned int lock; } arch_spinlock_t; Index: linux-5.6.17-rt10/arch/ia64/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/ia64/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/ia64/include/asm/spinlock_types.h @@ -2,10 +2,6 @@ #ifndef _ASM_IA64_SPINLOCK_TYPES_H #define _ASM_IA64_SPINLOCK_TYPES_H -#ifndef __LINUX_SPINLOCK_TYPES_H -# error "please don't include this file directly" -#endif - typedef struct { volatile unsigned int lock; } arch_spinlock_t; Index: linux-5.6.17-rt10/arch/mips/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/arch/mips/Kconfig +++ linux-5.6.17-rt10/arch/mips/Kconfig @@ -2637,7 +2637,7 @@ config MIPS_CRC_SUPPORT # config HIGHMEM bool "High Memory Support" - depends on 32BIT && CPU_SUPPORTS_HIGHMEM && SYS_SUPPORTS_HIGHMEM && !CPU_MIPS32_3_5_EVA + depends on 32BIT && CPU_SUPPORTS_HIGHMEM && SYS_SUPPORTS_HIGHMEM && !CPU_MIPS32_3_5_EVA && !PREEMPT_RT config CPU_SUPPORTS_HIGHMEM bool Index: linux-5.6.17-rt10/arch/powerpc/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/arch/powerpc/Kconfig +++ linux-5.6.17-rt10/arch/powerpc/Kconfig @@ -141,6 +141,7 @@ config PPC select ARCH_MIGHT_HAVE_PC_SERIO select ARCH_OPTIONAL_KERNEL_RWX if ARCH_HAS_STRICT_KERNEL_RWX select ARCH_SUPPORTS_ATOMIC_RMW + select ARCH_SUPPORTS_RT select ARCH_USE_BUILTIN_BSWAP select ARCH_USE_CMPXCHG_LOCKREF if PPC64 select ARCH_WANT_IPC_PARSE_VERSION @@ -220,6 +221,7 @@ config PPC select HAVE_HARDLOCKUP_DETECTOR_PERF if PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !HAVE_HARDLOCKUP_DETECTOR_ARCH select HAVE_PERF_REGS select HAVE_PERF_USER_STACK_DUMP + select HAVE_PREEMPT_LAZY select MMU_GATHER_RCU_TABLE_FREE select MMU_GATHER_PAGE_SIZE select HAVE_REGS_AND_STACK_ACCESS_API @@ -397,7 +399,7 @@ menu "Kernel options" config HIGHMEM bool "High memory support" - depends on PPC32 + depends on PPC32 && !PREEMPT_RT source "kernel/Kconfig.hz" Index: linux-5.6.17-rt10/arch/powerpc/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/powerpc/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/powerpc/include/asm/spinlock_types.h @@ -2,10 +2,6 @@ #ifndef _ASM_POWERPC_SPINLOCK_TYPES_H #define _ASM_POWERPC_SPINLOCK_TYPES_H -#ifndef __LINUX_SPINLOCK_TYPES_H -# error "please don't include this file directly" -#endif - typedef struct { volatile unsigned int slock; } arch_spinlock_t; Index: linux-5.6.17-rt10/arch/powerpc/include/asm/stackprotector.h =================================================================== --- linux-5.6.17-rt10.orig/arch/powerpc/include/asm/stackprotector.h +++ linux-5.6.17-rt10/arch/powerpc/include/asm/stackprotector.h @@ -24,7 +24,11 @@ static __always_inline void boot_init_st unsigned long canary; /* Try to get a semi random initial value. */ +#ifdef CONFIG_PREEMPT_RT + canary = (unsigned long)&canary; +#else canary = get_random_canary(); +#endif canary ^= mftb(); canary ^= LINUX_VERSION_CODE; canary &= CANARY_MASK; Index: linux-5.6.17-rt10/arch/powerpc/include/asm/thread_info.h =================================================================== --- linux-5.6.17-rt10.orig/arch/powerpc/include/asm/thread_info.h +++ linux-5.6.17-rt10/arch/powerpc/include/asm/thread_info.h @@ -48,6 +48,8 @@ struct thread_info { int preempt_count; /* 0 => preemptable, <0 => BUG */ + int preempt_lazy_count; /* 0 => preemptable, + <0 => BUG */ unsigned long local_flags; /* private flags for thread */ #ifdef CONFIG_LIVEPATCH unsigned long *livepatch_sp; @@ -98,11 +100,12 @@ void arch_setup_new_exec(void); #define TIF_SINGLESTEP 8 /* singlestepping active */ #define TIF_NOHZ 9 /* in adaptive nohz mode */ #define TIF_SECCOMP 10 /* secure computing */ -#define TIF_RESTOREALL 11 /* Restore all regs (implies NOERROR) */ -#define TIF_NOERROR 12 /* Force successful syscall return */ + +#define TIF_NEED_RESCHED_LAZY 11 /* lazy rescheduling necessary */ +#define TIF_SYSCALL_TRACEPOINT 12 /* syscall tracepoint instrumentation */ + #define TIF_NOTIFY_RESUME 13 /* callback before returning to user */ #define TIF_UPROBE 14 /* breakpointed or single-stepping */ -#define TIF_SYSCALL_TRACEPOINT 15 /* syscall tracepoint instrumentation */ #define TIF_EMULATE_STACK_STORE 16 /* Is an instruction emulation for stack store? */ #define TIF_MEMDIE 17 /* is terminating due to OOM killer */ @@ -111,6 +114,9 @@ void arch_setup_new_exec(void); #endif #define TIF_POLLING_NRFLAG 19 /* true if poll_idle() is polling TIF_NEED_RESCHED */ #define TIF_32BIT 20 /* 32 bit binary */ +#define TIF_RESTOREALL 21 /* Restore all regs (implies NOERROR) */ +#define TIF_NOERROR 22 /* Force successful syscall return */ + /* as above, but as bit values */ #define _TIF_SYSCALL_TRACE (1<done.wait, - dev->done.done || kthread_should_stop()); + res = swait_event_interruptible_exclusive(dev->done.wait, + dev->done.done || kthread_should_stop()); if (kthread_should_stop()) res = -EINTR; if (res) { Index: linux-5.6.17-rt10/arch/powerpc/platforms/pseries/iommu.c =================================================================== --- linux-5.6.17-rt10.orig/arch/powerpc/platforms/pseries/iommu.c +++ linux-5.6.17-rt10/arch/powerpc/platforms/pseries/iommu.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include #include @@ -177,6 +178,7 @@ static int tce_build_pSeriesLP(unsigned } static DEFINE_PER_CPU(__be64 *, tce_page); +static DEFINE_LOCAL_IRQ_LOCK(tcp_page_lock); static int tce_buildmulti_pSeriesLP(struct iommu_table *tbl, long tcenum, long npages, unsigned long uaddr, @@ -198,7 +200,8 @@ static int tce_buildmulti_pSeriesLP(stru direction, attrs); } - local_irq_save(flags); /* to protect tcep and the page behind it */ + /* to protect tcep and the page behind it */ + local_lock_irqsave(tcp_page_lock, flags); tcep = __this_cpu_read(tce_page); @@ -209,7 +212,7 @@ static int tce_buildmulti_pSeriesLP(stru tcep = (__be64 *)__get_free_page(GFP_ATOMIC); /* If allocation fails, fall back to the loop implementation */ if (!tcep) { - local_irq_restore(flags); + local_unlock_irqrestore(tcp_page_lock, flags); return tce_build_pSeriesLP(tbl->it_index, tcenum, tbl->it_page_shift, npages, uaddr, direction, attrs); @@ -244,7 +247,7 @@ static int tce_buildmulti_pSeriesLP(stru tcenum += limit; } while (npages > 0 && !rc); - local_irq_restore(flags); + local_unlock_irqrestore(tcp_page_lock, flags); if (unlikely(rc == H_NOT_ENOUGH_RESOURCES)) { ret = (int)rc; @@ -415,13 +418,14 @@ static int tce_setrange_multi_pSeriesLP( DMA_BIDIRECTIONAL, 0); } - local_irq_disable(); /* to protect tcep and the page behind it */ + /* to protect tcep and the page behind it */ + local_lock_irq(tcp_page_lock); tcep = __this_cpu_read(tce_page); if (!tcep) { tcep = (__be64 *)__get_free_page(GFP_ATOMIC); if (!tcep) { - local_irq_enable(); + local_unlock_irq(tcp_page_lock); return -ENOMEM; } __this_cpu_write(tce_page, tcep); @@ -467,7 +471,7 @@ static int tce_setrange_multi_pSeriesLP( /* error cleanup: caller will clear whole range */ - local_irq_enable(); + local_unlock_irq(tcp_page_lock); return rc; } Index: linux-5.6.17-rt10/arch/s390/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/s390/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/s390/include/asm/spinlock_types.h @@ -2,10 +2,6 @@ #ifndef __ASM_SPINLOCK_TYPES_H #define __ASM_SPINLOCK_TYPES_H -#ifndef __LINUX_SPINLOCK_TYPES_H -# error "please don't include this file directly" -#endif - typedef struct { int lock; } __attribute__ ((aligned (4))) arch_spinlock_t; Index: linux-5.6.17-rt10/arch/sh/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/sh/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/sh/include/asm/spinlock_types.h @@ -2,10 +2,6 @@ #ifndef __ASM_SH_SPINLOCK_TYPES_H #define __ASM_SH_SPINLOCK_TYPES_H -#ifndef __LINUX_SPINLOCK_TYPES_H -# error "please don't include this file directly" -#endif - typedef struct { volatile unsigned int lock; } arch_spinlock_t; Index: linux-5.6.17-rt10/arch/sh/kernel/irq.c =================================================================== --- linux-5.6.17-rt10.orig/arch/sh/kernel/irq.c +++ linux-5.6.17-rt10/arch/sh/kernel/irq.c @@ -148,6 +148,7 @@ void irq_ctx_exit(int cpu) hardirq_ctx[cpu] = NULL; } +#ifndef CONFIG_PREEMPT_RT void do_softirq_own_stack(void) { struct thread_info *curctx; @@ -175,6 +176,7 @@ void do_softirq_own_stack(void) "r5", "r6", "r7", "r8", "r9", "r15", "t", "pr" ); } +#endif #else static inline void handle_one_irq(unsigned int irq) { Index: linux-5.6.17-rt10/arch/sparc/kernel/irq_64.c =================================================================== --- linux-5.6.17-rt10.orig/arch/sparc/kernel/irq_64.c +++ linux-5.6.17-rt10/arch/sparc/kernel/irq_64.c @@ -854,6 +854,7 @@ void __irq_entry handler_irq(int pil, st set_irq_regs(old_regs); } +#ifndef CONFIG_PREEMPT_RT void do_softirq_own_stack(void) { void *orig_sp, *sp = softirq_stack[smp_processor_id()]; @@ -868,6 +869,7 @@ void do_softirq_own_stack(void) __asm__ __volatile__("mov %0, %%sp" : : "r" (orig_sp)); } +#endif #ifdef CONFIG_HOTPLUG_CPU void fixup_irqs(void) Index: linux-5.6.17-rt10/arch/x86/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/Kconfig +++ linux-5.6.17-rt10/arch/x86/Kconfig @@ -90,6 +90,7 @@ config X86 select ARCH_SUPPORTS_ACPI select ARCH_SUPPORTS_ATOMIC_RMW select ARCH_SUPPORTS_NUMA_BALANCING if X86_64 + select ARCH_SUPPORTS_RT select ARCH_USE_BUILTIN_BSWAP select ARCH_USE_QUEUED_RWLOCKS select ARCH_USE_QUEUED_SPINLOCKS @@ -135,8 +136,8 @@ config X86 select HAVE_ALIGNED_STRUCT_PAGE if SLUB select HAVE_ARCH_AUDITSYSCALL select HAVE_ARCH_HUGE_VMAP if X86_64 || X86_PAE - select HAVE_ARCH_JUMP_LABEL - select HAVE_ARCH_JUMP_LABEL_RELATIVE + select HAVE_ARCH_JUMP_LABEL if !PREEMPT_RT + select HAVE_ARCH_JUMP_LABEL_RELATIVE if !PREEMPT_RT select HAVE_ARCH_KASAN if X86_64 select HAVE_ARCH_KASAN_VMALLOC if X86_64 select HAVE_ARCH_KGDB @@ -204,6 +205,7 @@ config X86 select HAVE_PCI select HAVE_PERF_REGS select HAVE_PERF_USER_STACK_DUMP + select HAVE_PREEMPT_LAZY select MMU_GATHER_RCU_TABLE_FREE if PARAVIRT select HAVE_REGS_AND_STACK_ACCESS_API select HAVE_RELIABLE_STACKTRACE if X86_64 && (UNWINDER_FRAME_POINTER || UNWINDER_ORC) && STACK_VALIDATION Index: linux-5.6.17-rt10/arch/x86/crypto/aesni-intel_glue.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/crypto/aesni-intel_glue.c +++ linux-5.6.17-rt10/arch/x86/crypto/aesni-intel_glue.c @@ -382,14 +382,14 @@ static int ecb_encrypt(struct skcipher_r err = skcipher_walk_virt(&walk, req, true); - kernel_fpu_begin(); while ((nbytes = walk.nbytes)) { + kernel_fpu_begin(); aesni_ecb_enc(ctx, walk.dst.virt.addr, walk.src.virt.addr, nbytes & AES_BLOCK_MASK); + kernel_fpu_end(); nbytes &= AES_BLOCK_SIZE - 1; err = skcipher_walk_done(&walk, nbytes); } - kernel_fpu_end(); return err; } @@ -404,14 +404,14 @@ static int ecb_decrypt(struct skcipher_r err = skcipher_walk_virt(&walk, req, true); - kernel_fpu_begin(); while ((nbytes = walk.nbytes)) { + kernel_fpu_begin(); aesni_ecb_dec(ctx, walk.dst.virt.addr, walk.src.virt.addr, nbytes & AES_BLOCK_MASK); + kernel_fpu_end(); nbytes &= AES_BLOCK_SIZE - 1; err = skcipher_walk_done(&walk, nbytes); } - kernel_fpu_end(); return err; } @@ -426,14 +426,14 @@ static int cbc_encrypt(struct skcipher_r err = skcipher_walk_virt(&walk, req, true); - kernel_fpu_begin(); while ((nbytes = walk.nbytes)) { + kernel_fpu_begin(); aesni_cbc_enc(ctx, walk.dst.virt.addr, walk.src.virt.addr, nbytes & AES_BLOCK_MASK, walk.iv); + kernel_fpu_end(); nbytes &= AES_BLOCK_SIZE - 1; err = skcipher_walk_done(&walk, nbytes); } - kernel_fpu_end(); return err; } @@ -448,14 +448,14 @@ static int cbc_decrypt(struct skcipher_r err = skcipher_walk_virt(&walk, req, true); - kernel_fpu_begin(); while ((nbytes = walk.nbytes)) { + kernel_fpu_begin(); aesni_cbc_dec(ctx, walk.dst.virt.addr, walk.src.virt.addr, nbytes & AES_BLOCK_MASK, walk.iv); + kernel_fpu_end(); nbytes &= AES_BLOCK_SIZE - 1; err = skcipher_walk_done(&walk, nbytes); } - kernel_fpu_end(); return err; } @@ -505,18 +505,20 @@ static int ctr_crypt(struct skcipher_req err = skcipher_walk_virt(&walk, req, true); - kernel_fpu_begin(); while ((nbytes = walk.nbytes) >= AES_BLOCK_SIZE) { + kernel_fpu_begin(); aesni_ctr_enc_tfm(ctx, walk.dst.virt.addr, walk.src.virt.addr, nbytes & AES_BLOCK_MASK, walk.iv); + kernel_fpu_end(); nbytes &= AES_BLOCK_SIZE - 1; err = skcipher_walk_done(&walk, nbytes); } if (walk.nbytes) { + kernel_fpu_begin(); ctr_crypt_final(ctx, &walk); + kernel_fpu_end(); err = skcipher_walk_done(&walk, 0); } - kernel_fpu_end(); return err; } Index: linux-5.6.17-rt10/arch/x86/crypto/cast5_avx_glue.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/crypto/cast5_avx_glue.c +++ linux-5.6.17-rt10/arch/x86/crypto/cast5_avx_glue.c @@ -46,7 +46,7 @@ static inline void cast5_fpu_end(bool fp static int ecb_crypt(struct skcipher_request *req, bool enc) { - bool fpu_enabled = false; + bool fpu_enabled; struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req); struct cast5_ctx *ctx = crypto_skcipher_ctx(tfm); struct skcipher_walk walk; @@ -61,7 +61,7 @@ static int ecb_crypt(struct skcipher_req u8 *wsrc = walk.src.virt.addr; u8 *wdst = walk.dst.virt.addr; - fpu_enabled = cast5_fpu_begin(fpu_enabled, &walk, nbytes); + fpu_enabled = cast5_fpu_begin(false, &walk, nbytes); /* Process multi-block batch */ if (nbytes >= bsize * CAST5_PARALLEL_BLOCKS) { @@ -90,10 +90,9 @@ static int ecb_crypt(struct skcipher_req } while (nbytes >= bsize); done: + cast5_fpu_end(fpu_enabled); err = skcipher_walk_done(&walk, nbytes); } - - cast5_fpu_end(fpu_enabled); return err; } @@ -197,7 +196,7 @@ static int cbc_decrypt(struct skcipher_r { struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req); struct cast5_ctx *ctx = crypto_skcipher_ctx(tfm); - bool fpu_enabled = false; + bool fpu_enabled; struct skcipher_walk walk; unsigned int nbytes; int err; @@ -205,12 +204,11 @@ static int cbc_decrypt(struct skcipher_r err = skcipher_walk_virt(&walk, req, false); while ((nbytes = walk.nbytes)) { - fpu_enabled = cast5_fpu_begin(fpu_enabled, &walk, nbytes); + fpu_enabled = cast5_fpu_begin(false, &walk, nbytes); nbytes = __cbc_decrypt(ctx, &walk); + cast5_fpu_end(fpu_enabled); err = skcipher_walk_done(&walk, nbytes); } - - cast5_fpu_end(fpu_enabled); return err; } @@ -277,7 +275,7 @@ static int ctr_crypt(struct skcipher_req { struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req); struct cast5_ctx *ctx = crypto_skcipher_ctx(tfm); - bool fpu_enabled = false; + bool fpu_enabled; struct skcipher_walk walk; unsigned int nbytes; int err; @@ -285,13 +283,12 @@ static int ctr_crypt(struct skcipher_req err = skcipher_walk_virt(&walk, req, false); while ((nbytes = walk.nbytes) >= CAST5_BLOCK_SIZE) { - fpu_enabled = cast5_fpu_begin(fpu_enabled, &walk, nbytes); + fpu_enabled = cast5_fpu_begin(false, &walk, nbytes); nbytes = __ctr_crypt(&walk, ctx); + cast5_fpu_end(fpu_enabled); err = skcipher_walk_done(&walk, nbytes); } - cast5_fpu_end(fpu_enabled); - if (walk.nbytes) { ctr_crypt_final(&walk, ctx); err = skcipher_walk_done(&walk, 0); Index: linux-5.6.17-rt10/arch/x86/crypto/glue_helper.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/crypto/glue_helper.c +++ linux-5.6.17-rt10/arch/x86/crypto/glue_helper.c @@ -24,7 +24,7 @@ int glue_ecb_req_128bit(const struct com void *ctx = crypto_skcipher_ctx(crypto_skcipher_reqtfm(req)); const unsigned int bsize = 128 / 8; struct skcipher_walk walk; - bool fpu_enabled = false; + bool fpu_enabled; unsigned int nbytes; int err; @@ -37,7 +37,7 @@ int glue_ecb_req_128bit(const struct com unsigned int i; fpu_enabled = glue_fpu_begin(bsize, gctx->fpu_blocks_limit, - &walk, fpu_enabled, nbytes); + &walk, false, nbytes); for (i = 0; i < gctx->num_funcs; i++) { func_bytes = bsize * gctx->funcs[i].num_blocks; @@ -55,10 +55,9 @@ int glue_ecb_req_128bit(const struct com if (nbytes < bsize) break; } + glue_fpu_end(fpu_enabled); err = skcipher_walk_done(&walk, nbytes); } - - glue_fpu_end(fpu_enabled); return err; } EXPORT_SYMBOL_GPL(glue_ecb_req_128bit); @@ -101,7 +100,7 @@ int glue_cbc_decrypt_req_128bit(const st void *ctx = crypto_skcipher_ctx(crypto_skcipher_reqtfm(req)); const unsigned int bsize = 128 / 8; struct skcipher_walk walk; - bool fpu_enabled = false; + bool fpu_enabled; unsigned int nbytes; int err; @@ -115,7 +114,7 @@ int glue_cbc_decrypt_req_128bit(const st u128 last_iv; fpu_enabled = glue_fpu_begin(bsize, gctx->fpu_blocks_limit, - &walk, fpu_enabled, nbytes); + &walk, false, nbytes); /* Start of the last block. */ src += nbytes / bsize - 1; dst += nbytes / bsize - 1; @@ -148,10 +147,10 @@ int glue_cbc_decrypt_req_128bit(const st done: u128_xor(dst, dst, (u128 *)walk.iv); *(u128 *)walk.iv = last_iv; + glue_fpu_end(fpu_enabled); err = skcipher_walk_done(&walk, nbytes); } - glue_fpu_end(fpu_enabled); return err; } EXPORT_SYMBOL_GPL(glue_cbc_decrypt_req_128bit); @@ -162,7 +161,7 @@ int glue_ctr_req_128bit(const struct com void *ctx = crypto_skcipher_ctx(crypto_skcipher_reqtfm(req)); const unsigned int bsize = 128 / 8; struct skcipher_walk walk; - bool fpu_enabled = false; + bool fpu_enabled; unsigned int nbytes; int err; @@ -176,7 +175,7 @@ int glue_ctr_req_128bit(const struct com le128 ctrblk; fpu_enabled = glue_fpu_begin(bsize, gctx->fpu_blocks_limit, - &walk, fpu_enabled, nbytes); + &walk, false, nbytes); be128_to_le128(&ctrblk, (be128 *)walk.iv); @@ -202,11 +201,10 @@ int glue_ctr_req_128bit(const struct com } le128_to_be128((be128 *)walk.iv, &ctrblk); + glue_fpu_end(fpu_enabled); err = skcipher_walk_done(&walk, nbytes); } - glue_fpu_end(fpu_enabled); - if (nbytes) { le128 ctrblk; u128 tmp; @@ -306,8 +304,14 @@ int glue_xts_req_128bit(const struct com tweak_fn(tweak_ctx, walk.iv, walk.iv); while (nbytes) { + fpu_enabled = glue_fpu_begin(bsize, gctx->fpu_blocks_limit, + &walk, fpu_enabled, + nbytes < bsize ? bsize : nbytes); nbytes = __glue_xts_req_128bit(gctx, crypt_ctx, &walk); + glue_fpu_end(fpu_enabled); + fpu_enabled = false; + err = skcipher_walk_done(&walk, nbytes); nbytes = walk.nbytes; } Index: linux-5.6.17-rt10/arch/x86/entry/common.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/entry/common.c +++ linux-5.6.17-rt10/arch/x86/entry/common.c @@ -131,7 +131,7 @@ static long syscall_trace_enter(struct p #define EXIT_TO_USERMODE_LOOP_FLAGS \ (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_UPROBE | \ - _TIF_NEED_RESCHED | _TIF_USER_RETURN_NOTIFY | _TIF_PATCH_PENDING) + _TIF_NEED_RESCHED_MASK | _TIF_USER_RETURN_NOTIFY | _TIF_PATCH_PENDING) static void exit_to_usermode_loop(struct pt_regs *regs, u32 cached_flags) { @@ -146,9 +146,16 @@ static void exit_to_usermode_loop(struct /* We have work to do. */ local_irq_enable(); - if (cached_flags & _TIF_NEED_RESCHED) + if (cached_flags & _TIF_NEED_RESCHED_MASK) schedule(); +#ifdef ARCH_RT_DELAYS_SIGNAL_SEND + if (unlikely(current->forced_info.si_signo)) { + struct task_struct *t = current; + force_sig_info(&t->forced_info); + t->forced_info.si_signo = 0; + } +#endif if (cached_flags & _TIF_UPROBE) uprobe_notify_resume(regs); Index: linux-5.6.17-rt10/arch/x86/entry/entry_32.S =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/entry/entry_32.S +++ linux-5.6.17-rt10/arch/x86/entry/entry_32.S @@ -1110,8 +1110,26 @@ restore_all: restore_all_kernel: #ifdef CONFIG_PREEMPTION DISABLE_INTERRUPTS(CLBR_ANY) + # preempt count == 0 + NEED_RS set? cmpl $0, PER_CPU_VAR(__preempt_count) +#ifndef CONFIG_PREEMPT_LAZY jnz .Lno_preempt +#else + jz test_int_off + + # atleast preempt count == 0 ? + cmpl $_PREEMPT_ENABLED,PER_CPU_VAR(__preempt_count) + jne .Lno_preempt + + movl PER_CPU_VAR(current_task), %ebp + cmpl $0,TASK_TI_preempt_lazy_count(%ebp) # non-zero preempt_lazy_count ? + jnz .Lno_preempt + + testl $_TIF_NEED_RESCHED_LAZY, TASK_TI_flags(%ebp) + jz .Lno_preempt + +test_int_off: +#endif testl $X86_EFLAGS_IF, PT_EFLAGS(%esp) # interrupts off (exception path) ? jz .Lno_preempt call preempt_schedule_irq Index: linux-5.6.17-rt10/arch/x86/entry/entry_64.S =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/entry/entry_64.S +++ linux-5.6.17-rt10/arch/x86/entry/entry_64.S @@ -670,7 +670,23 @@ retint_kernel: btl $9, EFLAGS(%rsp) /* were interrupts off? */ jnc 1f cmpl $0, PER_CPU_VAR(__preempt_count) +#ifndef CONFIG_PREEMPT_LAZY jnz 1f +#else + jz do_preempt_schedule_irq + + # atleast preempt count == 0 ? + cmpl $_PREEMPT_ENABLED,PER_CPU_VAR(__preempt_count) + jnz 1f + + movq PER_CPU_VAR(current_task), %rcx + cmpl $0, TASK_TI_preempt_lazy_count(%rcx) + jnz 1f + + btl $TIF_NEED_RESCHED_LAZY,TASK_TI_flags(%rcx) + jnc 1f +do_preempt_schedule_irq: +#endif call preempt_schedule_irq 1: #endif @@ -1074,6 +1090,7 @@ SYM_CODE_START_LOCAL_NOALIGN(.Lbad_gs) SYM_CODE_END(.Lbad_gs) .previous +#ifndef CONFIG_PREEMPT_RT /* Call softirq on interrupt stack. Interrupts are off. */ SYM_FUNC_START(do_softirq_own_stack) pushq %rbp @@ -1084,6 +1101,7 @@ SYM_FUNC_START(do_softirq_own_stack) leaveq ret SYM_FUNC_END(do_softirq_own_stack) +#endif #ifdef CONFIG_XEN_PV idtentry hypervisor_callback xen_do_hypervisor_callback has_error_code=0 Index: linux-5.6.17-rt10/arch/x86/include/asm/fpu/api.h =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/include/asm/fpu/api.h +++ linux-5.6.17-rt10/arch/x86/include/asm/fpu/api.h @@ -23,6 +23,7 @@ extern void kernel_fpu_begin(void); extern void kernel_fpu_end(void); extern bool irq_fpu_usable(void); extern void fpregs_mark_activate(void); +extern void kernel_fpu_resched(void); /* * Use fpregs_lock() while editing CPU's FPU registers or fpu->state. Index: linux-5.6.17-rt10/arch/x86/include/asm/preempt.h =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/include/asm/preempt.h +++ linux-5.6.17-rt10/arch/x86/include/asm/preempt.h @@ -89,17 +89,48 @@ static __always_inline void __preempt_co * a decrement which hits zero means we have no preempt_count and should * reschedule. */ -static __always_inline bool __preempt_count_dec_and_test(void) +static __always_inline bool ____preempt_count_dec_and_test(void) { return GEN_UNARY_RMWcc("decl", __preempt_count, e, __percpu_arg([var])); } +static __always_inline bool __preempt_count_dec_and_test(void) +{ + if (____preempt_count_dec_and_test()) + return true; +#ifdef CONFIG_PREEMPT_LAZY + if (preempt_count()) + return false; + if (current_thread_info()->preempt_lazy_count) + return false; + return test_thread_flag(TIF_NEED_RESCHED_LAZY); +#else + return false; +#endif +} + /* * Returns true when we need to resched and can (barring IRQ state). */ static __always_inline bool should_resched(int preempt_offset) { +#ifdef CONFIG_PREEMPT_LAZY + u32 tmp; + tmp = raw_cpu_read_4(__preempt_count); + if (tmp == preempt_offset) + return true; + + /* preempt count == 0 ? */ + tmp &= ~PREEMPT_NEED_RESCHED; + if (tmp != preempt_offset) + return false; + /* XXX PREEMPT_LOCK_OFFSET */ + if (current_thread_info()->preempt_lazy_count) + return false; + return test_thread_flag(TIF_NEED_RESCHED_LAZY); +#else return unlikely(raw_cpu_read_4(__preempt_count) == preempt_offset); +#endif } #ifdef CONFIG_PREEMPTION Index: linux-5.6.17-rt10/arch/x86/include/asm/signal.h =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/include/asm/signal.h +++ linux-5.6.17-rt10/arch/x86/include/asm/signal.h @@ -28,6 +28,19 @@ typedef struct { #define SA_IA32_ABI 0x02000000u #define SA_X32_ABI 0x01000000u +/* + * Because some traps use the IST stack, we must keep preemption + * disabled while calling do_trap(), but do_trap() may call + * force_sig_info() which will grab the signal spin_locks for the + * task, which in PREEMPT_RT are mutexes. By defining + * ARCH_RT_DELAYS_SIGNAL_SEND the force_sig_info() will set + * TIF_NOTIFY_RESUME and set up the signal to be sent on exit of the + * trap. + */ +#if defined(CONFIG_PREEMPT_RT) +#define ARCH_RT_DELAYS_SIGNAL_SEND +#endif + #ifndef CONFIG_COMPAT typedef sigset_t compat_sigset_t; #endif Index: linux-5.6.17-rt10/arch/x86/include/asm/stackprotector.h =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/include/asm/stackprotector.h +++ linux-5.6.17-rt10/arch/x86/include/asm/stackprotector.h @@ -65,7 +65,7 @@ */ static __always_inline void boot_init_stack_canary(void) { - u64 canary; + u64 uninitialized_var(canary); u64 tsc; #ifdef CONFIG_X86_64 @@ -76,8 +76,14 @@ static __always_inline void boot_init_st * of randomness. The TSC only matters for very early init, * there it already has some randomness on most systems. Later * on during the bootup the random pool has true entropy too. + * For preempt-rt we need to weaken the randomness a bit, as + * we can't call into the random generator from atomic context + * due to locking constraints. We just leave canary + * uninitialized and use the TSC based randomness on top of it. */ +#ifndef CONFIG_PREEMPT_RT get_random_bytes(&canary, sizeof(canary)); +#endif tsc = rdtsc(); canary += tsc + (tsc << 32UL); canary &= CANARY_MASK; Index: linux-5.6.17-rt10/arch/x86/include/asm/thread_info.h =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/include/asm/thread_info.h +++ linux-5.6.17-rt10/arch/x86/include/asm/thread_info.h @@ -56,17 +56,24 @@ struct task_struct; struct thread_info { unsigned long flags; /* low level flags */ u32 status; /* thread synchronous flags */ + int preempt_lazy_count; /* 0 => lazy preemptable + <0 => BUG */ }; #define INIT_THREAD_INFO(tsk) \ { \ .flags = 0, \ + .preempt_lazy_count = 0, \ } #else /* !__ASSEMBLY__ */ #include +#define GET_THREAD_INFO(reg) \ + _ASM_MOV PER_CPU_VAR(cpu_current_top_of_stack),reg ; \ + _ASM_SUB $(THREAD_SIZE),reg ; + #endif /* @@ -92,6 +99,7 @@ struct thread_info { #define TIF_NOCPUID 15 /* CPUID is not accessible in userland */ #define TIF_NOTSC 16 /* TSC is not accessible in userland */ #define TIF_IA32 17 /* IA32 compatibility process */ +#define TIF_NEED_RESCHED_LAZY 18 /* lazy rescheduling necessary */ #define TIF_NOHZ 19 /* in adaptive nohz mode */ #define TIF_MEMDIE 20 /* is terminating due to OOM killer */ #define TIF_POLLING_NRFLAG 21 /* idle is polling for TIF_NEED_RESCHED */ @@ -122,6 +130,7 @@ struct thread_info { #define _TIF_NOCPUID (1 << TIF_NOCPUID) #define _TIF_NOTSC (1 << TIF_NOTSC) #define _TIF_IA32 (1 << TIF_IA32) +#define _TIF_NEED_RESCHED_LAZY (1 << TIF_NEED_RESCHED_LAZY) #define _TIF_NOHZ (1 << TIF_NOHZ) #define _TIF_POLLING_NRFLAG (1 << TIF_POLLING_NRFLAG) #define _TIF_IO_BITMAP (1 << TIF_IO_BITMAP) @@ -165,6 +174,8 @@ struct thread_info { #define _TIF_WORK_CTXSW_NEXT (_TIF_WORK_CTXSW) +#define _TIF_NEED_RESCHED_MASK (_TIF_NEED_RESCHED | _TIF_NEED_RESCHED_LAZY) + #define STACK_WARN (THREAD_SIZE/8) /* Index: linux-5.6.17-rt10/arch/x86/kernel/asm-offsets.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/kernel/asm-offsets.c +++ linux-5.6.17-rt10/arch/x86/kernel/asm-offsets.c @@ -38,6 +38,10 @@ static void __used common(void) #endif BLANK(); +#ifdef CONFIG_PREEMPT_LAZY + OFFSET(TASK_TI_flags, task_struct, thread_info.flags); + OFFSET(TASK_TI_preempt_lazy_count, task_struct, thread_info.preempt_lazy_count); +#endif OFFSET(TASK_addr_limit, task_struct, thread.addr_limit); BLANK(); @@ -92,6 +96,7 @@ static void __used common(void) BLANK(); DEFINE(PTREGS_SIZE, sizeof(struct pt_regs)); + DEFINE(_PREEMPT_ENABLED, PREEMPT_ENABLED); /* TLB state for the entry code */ OFFSET(TLB_STATE_user_pcid_flush_mask, tlb_state, user_pcid_flush_mask); Index: linux-5.6.17-rt10/arch/x86/kernel/cpu/mshyperv.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/kernel/cpu/mshyperv.c +++ linux-5.6.17-rt10/arch/x86/kernel/cpu/mshyperv.c @@ -77,12 +77,13 @@ EXPORT_SYMBOL_GPL(hv_remove_vmbus_irq); __visible void __irq_entry hv_stimer0_vector_handler(struct pt_regs *regs) { struct pt_regs *old_regs = set_irq_regs(regs); + u64 ip = regs ? instruction_pointer(regs) : 0; entering_irq(); inc_irq_stat(hyperv_stimer0_count); if (hv_stimer0_handler) hv_stimer0_handler(); - add_interrupt_randomness(HYPERV_STIMER0_VECTOR, 0); + add_interrupt_randomness(HYPERV_STIMER0_VECTOR, 0, ip); ack_APIC_irq(); exiting_irq(); Index: linux-5.6.17-rt10/arch/x86/kernel/fpu/core.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/kernel/fpu/core.c +++ linux-5.6.17-rt10/arch/x86/kernel/fpu/core.c @@ -113,6 +113,18 @@ void kernel_fpu_end(void) } EXPORT_SYMBOL_GPL(kernel_fpu_end); +void kernel_fpu_resched(void) +{ + WARN_ON_FPU(!this_cpu_read(in_kernel_fpu)); + + if (should_resched(PREEMPT_OFFSET)) { + kernel_fpu_end(); + cond_resched(); + kernel_fpu_begin(); + } +} +EXPORT_SYMBOL_GPL(kernel_fpu_resched); + /* * Save the FPU state (mark it for reload if necessary): * Index: linux-5.6.17-rt10/arch/x86/kernel/irq_32.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/kernel/irq_32.c +++ linux-5.6.17-rt10/arch/x86/kernel/irq_32.c @@ -131,6 +131,7 @@ int irq_init_percpu_irqstack(unsigned in return 0; } +#ifndef CONFIG_PREEMPT_RT void do_softirq_own_stack(void) { struct irq_stack *irqstk; @@ -147,6 +148,7 @@ void do_softirq_own_stack(void) call_on_stack(__do_softirq, isp); } +#endif void handle_irq(struct irq_desc *desc, struct pt_regs *regs) { Index: linux-5.6.17-rt10/arch/x86/kernel/process_32.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/kernel/process_32.c +++ linux-5.6.17-rt10/arch/x86/kernel/process_32.c @@ -38,6 +38,7 @@ #include #include #include +#include #include #include @@ -127,6 +128,35 @@ start_thread(struct pt_regs *regs, unsig } EXPORT_SYMBOL_GPL(start_thread); +#ifdef CONFIG_PREEMPT_RT +static void switch_kmaps(struct task_struct *prev_p, struct task_struct *next_p) +{ + int i; + + /* + * Clear @prev's kmap_atomic mappings + */ + for (i = 0; i < prev_p->kmap_idx; i++) { + int idx = i + KM_TYPE_NR * smp_processor_id(); + pte_t *ptep = kmap_pte - idx; + + kpte_clear_flush(ptep, __fix_to_virt(FIX_KMAP_BEGIN + idx)); + } + /* + * Restore @next_p's kmap_atomic mappings + */ + for (i = 0; i < next_p->kmap_idx; i++) { + int idx = i + KM_TYPE_NR * smp_processor_id(); + + if (!pte_none(next_p->kmap_pte[i])) + set_pte(kmap_pte - idx, next_p->kmap_pte[i]); + } +} +#else +static inline void +switch_kmaps(struct task_struct *prev_p, struct task_struct *next_p) { } +#endif + /* * switch_to(x,y) should switch tasks from x to y. @@ -188,6 +218,8 @@ __switch_to(struct task_struct *prev_p, switch_to_extra(prev_p, next_p); + switch_kmaps(prev_p, next_p); + /* * Leave lazy mode, flushing any hypercalls made here. * This must be done before restoring TLS segments so Index: linux-5.6.17-rt10/arch/x86/kvm/x86.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/kvm/x86.c +++ linux-5.6.17-rt10/arch/x86/kvm/x86.c @@ -7350,6 +7350,14 @@ int kvm_arch_init(void *opaque) goto out; } +#ifdef CONFIG_PREEMPT_RT + if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC)) { + pr_err("RT requires X86_FEATURE_CONSTANT_TSC\n"); + r = -EOPNOTSUPP; + goto out; + } +#endif + r = -ENOMEM; x86_fpu_cache = kmem_cache_create("x86_fpu", sizeof(struct fpu), __alignof__(struct fpu), SLAB_ACCOUNT, Index: linux-5.6.17-rt10/arch/x86/mm/highmem_32.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/mm/highmem_32.c +++ linux-5.6.17-rt10/arch/x86/mm/highmem_32.c @@ -33,10 +33,11 @@ EXPORT_SYMBOL(kunmap); */ void *kmap_atomic_prot(struct page *page, pgprot_t prot) { + pte_t pte = mk_pte(page, prot); unsigned long vaddr; int idx, type; - preempt_disable(); + preempt_disable_nort(); pagefault_disable(); if (!PageHighMem(page)) @@ -46,7 +47,10 @@ void *kmap_atomic_prot(struct page *page idx = type + KM_TYPE_NR*smp_processor_id(); vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx); BUG_ON(!pte_none(*(kmap_pte-idx))); - set_pte(kmap_pte-idx, mk_pte(page, prot)); +#ifdef CONFIG_PREEMPT_RT + current->kmap_pte[type] = pte; +#endif + set_pte(kmap_pte-idx, pte); arch_flush_lazy_mmu_mode(); return (void *)vaddr; @@ -89,6 +93,9 @@ void __kunmap_atomic(void *kvaddr) * is a bad idea also, in case the page changes cacheability * attributes or becomes a protected page in a hypervisor. */ +#ifdef CONFIG_PREEMPT_RT + current->kmap_pte[type] = __pte(0); +#endif kpte_clear_flush(kmap_pte-idx, vaddr); kmap_atomic_idx_pop(); arch_flush_lazy_mmu_mode(); @@ -101,7 +108,7 @@ void __kunmap_atomic(void *kvaddr) #endif pagefault_enable(); - preempt_enable(); + preempt_enable_nort(); } EXPORT_SYMBOL(__kunmap_atomic); Index: linux-5.6.17-rt10/arch/x86/mm/iomap_32.c =================================================================== --- linux-5.6.17-rt10.orig/arch/x86/mm/iomap_32.c +++ linux-5.6.17-rt10/arch/x86/mm/iomap_32.c @@ -46,6 +46,7 @@ EXPORT_SYMBOL_GPL(iomap_free); void *kmap_atomic_prot_pfn(unsigned long pfn, pgprot_t prot) { + pte_t pte = pfn_pte(pfn, prot); unsigned long vaddr; int idx, type; @@ -55,7 +56,12 @@ void *kmap_atomic_prot_pfn(unsigned long type = kmap_atomic_idx_push(); idx = type + KM_TYPE_NR * smp_processor_id(); vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx); - set_pte(kmap_pte - idx, pfn_pte(pfn, prot)); + WARN_ON(!pte_none(*(kmap_pte - idx))); + +#ifdef CONFIG_PREEMPT_RT + current->kmap_pte[type] = pte; +#endif + set_pte(kmap_pte - idx, pte); arch_flush_lazy_mmu_mode(); return (void *)vaddr; @@ -106,6 +112,9 @@ iounmap_atomic(void __iomem *kvaddr) * is a bad idea also, in case the page changes cacheability * attributes or becomes a protected page in a hypervisor. */ +#ifdef CONFIG_PREEMPT_RT + current->kmap_pte[type] = __pte(0); +#endif kpte_clear_flush(kmap_pte-idx, vaddr); kmap_atomic_idx_pop(); } Index: linux-5.6.17-rt10/arch/xtensa/include/asm/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/arch/xtensa/include/asm/spinlock_types.h +++ linux-5.6.17-rt10/arch/xtensa/include/asm/spinlock_types.h @@ -2,10 +2,6 @@ #ifndef __ASM_SPINLOCK_TYPES_H #define __ASM_SPINLOCK_TYPES_H -#if !defined(__LINUX_SPINLOCK_TYPES_H) && !defined(__ASM_SPINLOCK_H) -# error "please don't include this file directly" -#endif - #include #include Index: linux-5.6.17-rt10/block/blk-ioc.c =================================================================== --- linux-5.6.17-rt10.orig/block/blk-ioc.c +++ linux-5.6.17-rt10/block/blk-ioc.c @@ -9,6 +9,7 @@ #include #include #include +#include #include "blk.h" @@ -116,7 +117,7 @@ static void ioc_release_fn(struct work_s spin_unlock(&q->queue_lock); } else { spin_unlock_irqrestore(&ioc->lock, flags); - cpu_relax(); + cpu_chill(); spin_lock_irqsave_nested(&ioc->lock, flags, 1); } } Index: linux-5.6.17-rt10/block/blk-iocost.c =================================================================== --- linux-5.6.17-rt10.orig/block/blk-iocost.c +++ linux-5.6.17-rt10/block/blk-iocost.c @@ -408,7 +408,7 @@ struct ioc { enum ioc_running running; atomic64_t vtime_rate; - seqcount_t period_seqcount; + seqcount_spinlock_t period_seqcount; u32 period_at; /* wallclock starttime */ u64 period_at_vtime; /* vtime starttime */ @@ -875,7 +875,6 @@ static void ioc_now(struct ioc *ioc, str static void ioc_start_period(struct ioc *ioc, struct ioc_now *now) { - lockdep_assert_held(&ioc->lock); WARN_ON_ONCE(ioc->running != IOC_RUNNING); write_seqcount_begin(&ioc->period_seqcount); @@ -1961,7 +1960,7 @@ static int blk_iocost_init(struct reques ioc->running = IOC_IDLE; atomic64_set(&ioc->vtime_rate, VTIME_PER_USEC); - seqcount_init(&ioc->period_seqcount); + seqcount_spinlock_init(&ioc->period_seqcount, &ioc->lock); ioc->period_at = ktime_to_us(ktime_get()); atomic64_set(&ioc->cur_period, 0); atomic_set(&ioc->hweight_gen, 0); Index: linux-5.6.17-rt10/block/blk-mq.c =================================================================== --- linux-5.6.17-rt10.orig/block/blk-mq.c +++ linux-5.6.17-rt10/block/blk-mq.c @@ -589,9 +589,17 @@ static void __blk_mq_complete_request(st return; } - cpu = get_cpu(); + cpu = get_cpu_light(); + /* + * Avoid SMP function calls for completions because they acquire + * sleeping spinlocks on RT. + */ +#ifdef CONFIG_PREEMPT_RT + shared = true; +#else if (!test_bit(QUEUE_FLAG_SAME_FORCE, &q->queue_flags)) shared = cpus_share_cache(cpu, ctx->cpu); +#endif if (cpu != ctx->cpu && !shared && cpu_online(ctx->cpu)) { rq->csd.func = __blk_mq_complete_request_remote; @@ -601,7 +609,7 @@ static void __blk_mq_complete_request(st } else { q->mq_ops->complete(rq); } - put_cpu(); + put_cpu_light(); } static void hctx_unlock(struct blk_mq_hw_ctx *hctx, int srcu_idx) @@ -1456,14 +1464,14 @@ static void __blk_mq_delay_run_hw_queue( return; if (!async && !(hctx->flags & BLK_MQ_F_BLOCKING)) { - int cpu = get_cpu(); + int cpu = get_cpu_light(); if (cpumask_test_cpu(cpu, hctx->cpumask)) { __blk_mq_run_hw_queue(hctx); - put_cpu(); + put_cpu_light(); return; } - put_cpu(); + put_cpu_light(); } kblockd_mod_delayed_work_on(blk_mq_hctx_next_cpu(hctx), &hctx->run_work, Index: linux-5.6.17-rt10/block/blk-softirq.c =================================================================== --- linux-5.6.17-rt10.orig/block/blk-softirq.c +++ linux-5.6.17-rt10/block/blk-softirq.c @@ -87,6 +87,7 @@ static int blk_softirq_cpu_dead(unsigned this_cpu_ptr(&blk_cpu_done)); raise_softirq_irqoff(BLOCK_SOFTIRQ); local_irq_enable(); + preempt_check_resched_rt(); return 0; } @@ -138,6 +139,7 @@ do_local: goto do_local; local_irq_restore(flags); + preempt_check_resched_rt(); } static __init int blk_softirq_init(void) Index: linux-5.6.17-rt10/crypto/cryptd.c =================================================================== --- linux-5.6.17-rt10.orig/crypto/cryptd.c +++ linux-5.6.17-rt10/crypto/cryptd.c @@ -36,6 +36,7 @@ static struct workqueue_struct *cryptd_w struct cryptd_cpu_queue { struct crypto_queue queue; struct work_struct work; + spinlock_t qlock; }; struct cryptd_queue { @@ -105,6 +106,7 @@ static int cryptd_init_queue(struct cryp cpu_queue = per_cpu_ptr(queue->cpu_queue, cpu); crypto_init_queue(&cpu_queue->queue, max_cpu_qlen); INIT_WORK(&cpu_queue->work, cryptd_queue_worker); + spin_lock_init(&cpu_queue->qlock); } pr_info("cryptd: max_cpu_qlen set to %d\n", max_cpu_qlen); return 0; @@ -129,8 +131,10 @@ static int cryptd_enqueue_request(struct struct cryptd_cpu_queue *cpu_queue; refcount_t *refcnt; - cpu = get_cpu(); - cpu_queue = this_cpu_ptr(queue->cpu_queue); + cpu_queue = raw_cpu_ptr(queue->cpu_queue); + spin_lock_bh(&cpu_queue->qlock); + cpu = smp_processor_id(); + err = crypto_enqueue_request(&cpu_queue->queue, request); refcnt = crypto_tfm_ctx(request->tfm); @@ -146,7 +150,7 @@ static int cryptd_enqueue_request(struct refcount_inc(refcnt); out_put_cpu: - put_cpu(); + spin_unlock_bh(&cpu_queue->qlock); return err; } @@ -162,16 +166,11 @@ static void cryptd_queue_worker(struct w cpu_queue = container_of(work, struct cryptd_cpu_queue, work); /* * Only handle one request at a time to avoid hogging crypto workqueue. - * preempt_disable/enable is used to prevent being preempted by - * cryptd_enqueue_request(). local_bh_disable/enable is used to prevent - * cryptd_enqueue_request() being accessed from software interrupts. */ - local_bh_disable(); - preempt_disable(); + spin_lock_bh(&cpu_queue->qlock); backlog = crypto_get_backlog(&cpu_queue->queue); req = crypto_dequeue_request(&cpu_queue->queue); - preempt_enable(); - local_bh_enable(); + spin_unlock_bh(&cpu_queue->qlock); if (!req) return; Index: linux-5.6.17-rt10/drivers/block/zram/zcomp.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/block/zram/zcomp.c +++ linux-5.6.17-rt10/drivers/block/zram/zcomp.c @@ -113,12 +113,20 @@ ssize_t zcomp_available_show(const char struct zcomp_strm *zcomp_stream_get(struct zcomp *comp) { - return *get_cpu_ptr(comp->stream); + struct zcomp_strm *zstrm; + + zstrm = *get_local_ptr(comp->stream); + spin_lock(&zstrm->zcomp_lock); + return zstrm; } void zcomp_stream_put(struct zcomp *comp) { - put_cpu_ptr(comp->stream); + struct zcomp_strm *zstrm; + + zstrm = *this_cpu_ptr(comp->stream); + spin_unlock(&zstrm->zcomp_lock); + put_local_ptr(zstrm); } int zcomp_compress(struct zcomp_strm *zstrm, @@ -168,6 +176,7 @@ int zcomp_cpu_up_prepare(unsigned int cp pr_err("Can't allocate a compression stream\n"); return -ENOMEM; } + spin_lock_init(&zstrm->zcomp_lock); *per_cpu_ptr(comp->stream, cpu) = zstrm; return 0; } Index: linux-5.6.17-rt10/drivers/block/zram/zcomp.h =================================================================== --- linux-5.6.17-rt10.orig/drivers/block/zram/zcomp.h +++ linux-5.6.17-rt10/drivers/block/zram/zcomp.h @@ -10,6 +10,7 @@ struct zcomp_strm { /* compression/decompression buffer */ void *buffer; struct crypto_comp *tfm; + spinlock_t zcomp_lock; }; /* dynamic per-device compression frontend */ Index: linux-5.6.17-rt10/drivers/block/zram/zram_drv.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/block/zram/zram_drv.c +++ linux-5.6.17-rt10/drivers/block/zram/zram_drv.c @@ -55,6 +55,40 @@ static void zram_free_page(struct zram * static int zram_bvec_read(struct zram *zram, struct bio_vec *bvec, u32 index, int offset, struct bio *bio); +#ifdef CONFIG_PREEMPT_RT +static void zram_meta_init_table_locks(struct zram *zram, size_t num_pages) +{ + size_t index; + + for (index = 0; index < num_pages; index++) + spin_lock_init(&zram->table[index].lock); +} + +static int zram_slot_trylock(struct zram *zram, u32 index) +{ + int ret; + + ret = spin_trylock(&zram->table[index].lock); + if (ret) + __set_bit(ZRAM_LOCK, &zram->table[index].flags); + return ret; +} + +static void zram_slot_lock(struct zram *zram, u32 index) +{ + spin_lock(&zram->table[index].lock); + __set_bit(ZRAM_LOCK, &zram->table[index].flags); +} + +static void zram_slot_unlock(struct zram *zram, u32 index) +{ + __clear_bit(ZRAM_LOCK, &zram->table[index].flags); + spin_unlock(&zram->table[index].lock); +} + +#else + +static void zram_meta_init_table_locks(struct zram *zram, size_t num_pages) { } static int zram_slot_trylock(struct zram *zram, u32 index) { @@ -70,6 +104,7 @@ static void zram_slot_unlock(struct zram { bit_spin_unlock(ZRAM_LOCK, &zram->table[index].flags); } +#endif static inline bool init_done(struct zram *zram) { @@ -1157,6 +1192,7 @@ static bool zram_meta_alloc(struct zram if (!huge_class_size) huge_class_size = zs_huge_class_size(zram->mem_pool); + zram_meta_init_table_locks(zram, num_pages); return true; } @@ -1219,6 +1255,7 @@ static int __zram_bvec_read(struct zram unsigned long handle; unsigned int size; void *src, *dst; + struct zcomp_strm *zstrm; zram_slot_lock(zram, index); if (zram_test_flag(zram, index, ZRAM_WB)) { @@ -1249,6 +1286,7 @@ static int __zram_bvec_read(struct zram size = zram_get_obj_size(zram, index); + zstrm = zcomp_stream_get(zram->comp); src = zs_map_object(zram->mem_pool, handle, ZS_MM_RO); if (size == PAGE_SIZE) { dst = kmap_atomic(page); @@ -1256,14 +1294,13 @@ static int __zram_bvec_read(struct zram kunmap_atomic(dst); ret = 0; } else { - struct zcomp_strm *zstrm = zcomp_stream_get(zram->comp); dst = kmap_atomic(page); ret = zcomp_decompress(zstrm, src, size, dst); kunmap_atomic(dst); - zcomp_stream_put(zram->comp); } zs_unmap_object(zram->mem_pool, handle); + zcomp_stream_put(zram->comp); zram_slot_unlock(zram, index); /* Should NEVER happen. Return bio error if it does. */ Index: linux-5.6.17-rt10/drivers/block/zram/zram_drv.h =================================================================== --- linux-5.6.17-rt10.orig/drivers/block/zram/zram_drv.h +++ linux-5.6.17-rt10/drivers/block/zram/zram_drv.h @@ -63,6 +63,7 @@ struct zram_table_entry { unsigned long element; }; unsigned long flags; + spinlock_t lock; #ifdef CONFIG_ZRAM_MEMORY_TRACKING ktime_t ac_time; #endif Index: linux-5.6.17-rt10/drivers/char/random.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/char/random.c +++ linux-5.6.17-rt10/drivers/char/random.c @@ -1224,28 +1224,27 @@ static __u32 get_reg(struct fast_pool *f return *ptr; } -void add_interrupt_randomness(int irq, int irq_flags) +void add_interrupt_randomness(int irq, int irq_flags, __u64 ip) { struct entropy_store *r; struct fast_pool *fast_pool = this_cpu_ptr(&irq_randomness); - struct pt_regs *regs = get_irq_regs(); unsigned long now = jiffies; cycles_t cycles = random_get_entropy(); __u32 c_high, j_high; - __u64 ip; unsigned long seed; int credit = 0; if (cycles == 0) - cycles = get_reg(fast_pool, regs); + cycles = get_reg(fast_pool, NULL); c_high = (sizeof(cycles) > 4) ? cycles >> 32 : 0; j_high = (sizeof(now) > 4) ? now >> 32 : 0; fast_pool->pool[0] ^= cycles ^ j_high ^ irq; fast_pool->pool[1] ^= now ^ c_high; - ip = regs ? instruction_pointer(regs) : _RET_IP_; + if (!ip) + ip = _RET_IP_; fast_pool->pool[2] ^= ip; fast_pool->pool[3] ^= (sizeof(ip) > 4) ? ip >> 32 : - get_reg(fast_pool, regs); + get_reg(fast_pool, NULL); fast_mix(fast_pool); add_interrupt_bench(cycles); Index: linux-5.6.17-rt10/drivers/char/tpm/tpm-dev-common.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/char/tpm/tpm-dev-common.c +++ linux-5.6.17-rt10/drivers/char/tpm/tpm-dev-common.c @@ -20,7 +20,6 @@ #include "tpm-dev.h" static struct workqueue_struct *tpm_dev_wq; -static DEFINE_MUTEX(tpm_dev_wq_lock); static ssize_t tpm_dev_transmit(struct tpm_chip *chip, struct tpm_space *space, u8 *buf, size_t bufsiz) Index: linux-5.6.17-rt10/drivers/char/tpm/tpm_tis.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/char/tpm/tpm_tis.c +++ linux-5.6.17-rt10/drivers/char/tpm/tpm_tis.c @@ -49,6 +49,31 @@ static inline struct tpm_tis_tcg_phy *to return container_of(data, struct tpm_tis_tcg_phy, priv); } +#ifdef CONFIG_PREEMPT_RT +/* + * Flushes previous write operations to chip so that a subsequent + * ioread*()s won't stall a cpu. + */ +static inline void tpm_tis_flush(void __iomem *iobase) +{ + ioread8(iobase + TPM_ACCESS(0)); +} +#else +#define tpm_tis_flush(iobase) do { } while (0) +#endif + +static inline void tpm_tis_iowrite8(u8 b, void __iomem *iobase, u32 addr) +{ + iowrite8(b, iobase + addr); + tpm_tis_flush(iobase); +} + +static inline void tpm_tis_iowrite32(u32 b, void __iomem *iobase, u32 addr) +{ + iowrite32(b, iobase + addr); + tpm_tis_flush(iobase); +} + static bool interrupts = true; module_param(interrupts, bool, 0444); MODULE_PARM_DESC(interrupts, "Enable interrupts"); @@ -146,7 +171,7 @@ static int tpm_tcg_write_bytes(struct tp struct tpm_tis_tcg_phy *phy = to_tpm_tis_tcg_phy(data); while (len--) - iowrite8(*value++, phy->iobase + addr); + tpm_tis_iowrite8(*value++, phy->iobase, addr); return 0; } @@ -173,7 +198,7 @@ static int tpm_tcg_write32(struct tpm_ti { struct tpm_tis_tcg_phy *phy = to_tpm_tis_tcg_phy(data); - iowrite32(value, phy->iobase + addr); + tpm_tis_iowrite32(value, phy->iobase, addr); return 0; } Index: linux-5.6.17-rt10/drivers/clocksource/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/drivers/clocksource/Kconfig +++ linux-5.6.17-rt10/drivers/clocksource/Kconfig @@ -434,6 +434,13 @@ config ATMEL_TCB_CLKSRC help Support for Timer Counter Blocks on Atmel SoCs. +config ATMEL_TCB_CLKSRC_USE_SLOW_CLOCK + bool "TC Block use 32 KiHz clock" + depends on ATMEL_TCB_CLKSRC + default y + help + Select this to use 32 KiHz base clock rate as TC block clock. + config CLKSRC_EXYNOS_MCT bool "Exynos multi core timer driver" if COMPILE_TEST depends on ARM || ARM64 Index: linux-5.6.17-rt10/drivers/clocksource/timer-atmel-tcb.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/clocksource/timer-atmel-tcb.c +++ linux-5.6.17-rt10/drivers/clocksource/timer-atmel-tcb.c @@ -28,8 +28,7 @@ * this 32 bit free-running counter. the second channel is not used. * * - The third channel may be used to provide a 16-bit clockevent - * source, used in either periodic or oneshot mode. This runs - * at 32 KiHZ, and can handle delays of up to two seconds. + * source, used in either periodic or oneshot mode. * * REVISIT behavior during system suspend states... we should disable * all clocks and save the power. Easily done for clockevent devices, @@ -143,6 +142,8 @@ static unsigned long notrace tc_delay_ti struct tc_clkevt_device { struct clock_event_device clkevt; struct clk *clk; + bool clk_enabled; + u32 freq; void __iomem *regs; }; @@ -151,15 +152,26 @@ static struct tc_clkevt_device *to_tc_cl return container_of(clkevt, struct tc_clkevt_device, clkevt); } -/* For now, we always use the 32K clock ... this optimizes for NO_HZ, - * because using one of the divided clocks would usually mean the - * tick rate can never be less than several dozen Hz (vs 0.5 Hz). - * - * A divided clock could be good for high resolution timers, since - * 30.5 usec resolution can seem "low". - */ static u32 timer_clock; +static void tc_clk_disable(struct clock_event_device *d) +{ + struct tc_clkevt_device *tcd = to_tc_clkevt(d); + + clk_disable(tcd->clk); + tcd->clk_enabled = false; +} + +static void tc_clk_enable(struct clock_event_device *d) +{ + struct tc_clkevt_device *tcd = to_tc_clkevt(d); + + if (tcd->clk_enabled) + return; + clk_enable(tcd->clk); + tcd->clk_enabled = true; +} + static int tc_shutdown(struct clock_event_device *d) { struct tc_clkevt_device *tcd = to_tc_clkevt(d); @@ -167,8 +179,14 @@ static int tc_shutdown(struct clock_even writel(0xff, regs + ATMEL_TC_REG(2, IDR)); writel(ATMEL_TC_CLKDIS, regs + ATMEL_TC_REG(2, CCR)); + return 0; +} + +static int tc_shutdown_clk_off(struct clock_event_device *d) +{ + tc_shutdown(d); if (!clockevent_state_detached(d)) - clk_disable(tcd->clk); + tc_clk_disable(d); return 0; } @@ -181,9 +199,9 @@ static int tc_set_oneshot(struct clock_e if (clockevent_state_oneshot(d) || clockevent_state_periodic(d)) tc_shutdown(d); - clk_enable(tcd->clk); + tc_clk_enable(d); - /* slow clock, count up to RC, then irq and stop */ + /* count up to RC, then irq and stop */ writel(timer_clock | ATMEL_TC_CPCSTOP | ATMEL_TC_WAVE | ATMEL_TC_WAVESEL_UP_AUTO, regs + ATMEL_TC_REG(2, CMR)); writel(ATMEL_TC_CPCS, regs + ATMEL_TC_REG(2, IER)); @@ -203,12 +221,12 @@ static int tc_set_periodic(struct clock_ /* By not making the gentime core emulate periodic mode on top * of oneshot, we get lower overhead and improved accuracy. */ - clk_enable(tcd->clk); + tc_clk_enable(d); - /* slow clock, count up to RC, then irq and restart */ + /* count up to RC, then irq and restart */ writel(timer_clock | ATMEL_TC_WAVE | ATMEL_TC_WAVESEL_UP_AUTO, regs + ATMEL_TC_REG(2, CMR)); - writel((32768 + HZ / 2) / HZ, tcaddr + ATMEL_TC_REG(2, RC)); + writel((tcd->freq + HZ / 2) / HZ, tcaddr + ATMEL_TC_REG(2, RC)); /* Enable clock and interrupts on RC compare */ writel(ATMEL_TC_CPCS, regs + ATMEL_TC_REG(2, IER)); @@ -234,9 +252,13 @@ static struct tc_clkevt_device clkevt = .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, /* Should be lower than at91rm9200's system timer */ +#ifdef CONFIG_ATMEL_TCB_CLKSRC_USE_SLOW_CLOCK .rating = 125, +#else + .rating = 200, +#endif .set_next_event = tc_next_event, - .set_state_shutdown = tc_shutdown, + .set_state_shutdown = tc_shutdown_clk_off, .set_state_periodic = tc_set_periodic, .set_state_oneshot = tc_set_oneshot, }, @@ -256,8 +278,11 @@ static irqreturn_t ch2_irq(int irq, void return IRQ_NONE; } -static int __init setup_clkevents(struct atmel_tc *tc, int clk32k_divisor_idx) +static const u8 atmel_tcb_divisors[5] = { 2, 8, 32, 128, 0, }; + +static int __init setup_clkevents(struct atmel_tc *tc, int divisor_idx) { + unsigned divisor = atmel_tcb_divisors[divisor_idx]; int ret; struct clk *t2_clk = tc->clk[2]; int irq = tc->irq[2]; @@ -278,7 +303,11 @@ static int __init setup_clkevents(struct clkevt.regs = tc->regs; clkevt.clk = t2_clk; - timer_clock = clk32k_divisor_idx; + timer_clock = divisor_idx; + if (!divisor) + clkevt.freq = 32768; + else + clkevt.freq = clk_get_rate(t2_clk) / divisor; clkevt.clkevt.cpumask = cpumask_of(0); @@ -289,7 +318,7 @@ static int __init setup_clkevents(struct return ret; } - clockevents_config_and_register(&clkevt.clkevt, 32768, 1, 0xffff); + clockevents_config_and_register(&clkevt.clkevt, clkevt.freq, 1, 0xffff); return ret; } @@ -346,8 +375,6 @@ static void __init tcb_setup_single_chan writel(ATMEL_TC_SYNC, tcaddr + ATMEL_TC_BCR); } -static const u8 atmel_tcb_divisors[5] = { 2, 8, 32, 128, 0, }; - static const struct of_device_id atmel_tcb_of_match[] = { { .compatible = "atmel,at91rm9200-tcb", .data = (void *)16, }, { .compatible = "atmel,at91sam9x5-tcb", .data = (void *)32, }, @@ -467,7 +494,11 @@ static int __init tcb_clksrc_init(struct goto err_disable_t1; /* channel 2: periodic and oneshot timer support */ +#ifdef CONFIG_ATMEL_TCB_CLKSRC_USE_SLOW_CLOCK ret = setup_clkevents(&tc, clk32k_divisor_idx); +#else + ret = setup_clkevents(&tc, best_divisor_idx); +#endif if (ret) goto err_unregister_clksrc; Index: linux-5.6.17-rt10/drivers/connector/cn_proc.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/connector/cn_proc.c +++ linux-5.6.17-rt10/drivers/connector/cn_proc.c @@ -18,6 +18,7 @@ #include #include +#include /* * Size of a cn_msg followed by a proc_event structure. Since the @@ -40,10 +41,11 @@ static struct cb_id cn_proc_event_id = { /* proc_event_counts is used as the sequence number of the netlink message */ static DEFINE_PER_CPU(__u32, proc_event_counts) = { 0 }; +static DEFINE_LOCAL_IRQ_LOCK(send_msg_lock); static inline void send_msg(struct cn_msg *msg) { - preempt_disable(); + local_lock(send_msg_lock); msg->seq = __this_cpu_inc_return(proc_event_counts) - 1; ((struct proc_event *)msg->data)->cpu = smp_processor_id(); @@ -56,7 +58,7 @@ static inline void send_msg(struct cn_ms */ cn_netlink_send(msg, 0, CN_IDX_PROC, GFP_NOWAIT); - preempt_enable(); + local_unlock(send_msg_lock); } void proc_fork_connector(struct task_struct *task) Index: linux-5.6.17-rt10/drivers/dma-buf/dma-resv.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/dma-buf/dma-resv.c +++ linux-5.6.17-rt10/drivers/dma-buf/dma-resv.c @@ -50,12 +50,6 @@ DEFINE_WD_CLASS(reservation_ww_class); EXPORT_SYMBOL(reservation_ww_class); -struct lock_class_key reservation_seqcount_class; -EXPORT_SYMBOL(reservation_seqcount_class); - -const char reservation_seqcount_string[] = "reservation_seqcount"; -EXPORT_SYMBOL(reservation_seqcount_string); - /** * dma_resv_list_alloc - allocate fence list * @shared_max: number of fences we need space for @@ -134,9 +128,8 @@ subsys_initcall(dma_resv_lockdep); void dma_resv_init(struct dma_resv *obj) { ww_mutex_init(&obj->lock, &reservation_ww_class); + seqcount_ww_mutex_init(&obj->seq, &obj->lock); - __seqcount_init(&obj->seq, reservation_seqcount_string, - &reservation_seqcount_class); RCU_INIT_POINTER(obj->fence, NULL); RCU_INIT_POINTER(obj->fence_excl, NULL); } @@ -266,7 +259,6 @@ void dma_resv_add_shared_fence(struct dm fobj = dma_resv_get_list(obj); count = fobj->shared_count; - preempt_disable(); write_seqcount_begin(&obj->seq); for (i = 0; i < count; ++i) { @@ -288,7 +280,6 @@ replace: smp_store_mb(fobj->shared_count, count); write_seqcount_end(&obj->seq); - preempt_enable(); dma_fence_put(old); } EXPORT_SYMBOL(dma_resv_add_shared_fence); @@ -315,14 +306,12 @@ void dma_resv_add_excl_fence(struct dma_ if (fence) dma_fence_get(fence); - preempt_disable(); write_seqcount_begin(&obj->seq); /* write_seqcount_begin provides the necessary memory barrier */ RCU_INIT_POINTER(obj->fence_excl, fence); if (old) old->shared_count = 0; write_seqcount_end(&obj->seq); - preempt_enable(); /* inplace update, no shared fences */ while (i--) @@ -400,13 +389,11 @@ retry: src_list = dma_resv_get_list(dst); old = dma_resv_get_excl(dst); - preempt_disable(); write_seqcount_begin(&dst->seq); /* write_seqcount_begin provides the necessary memory barrier */ RCU_INIT_POINTER(dst->fence_excl, new); RCU_INIT_POINTER(dst->fence, dst_list); write_seqcount_end(&dst->seq); - preempt_enable(); dma_resv_list_free(src_list); dma_fence_put(old); Index: linux-5.6.17-rt10/drivers/firmware/efi/efi.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/firmware/efi/efi.c +++ linux-5.6.17-rt10/drivers/firmware/efi/efi.c @@ -68,7 +68,7 @@ struct mm_struct efi_mm = { struct workqueue_struct *efi_rts_wq; -static bool disable_runtime; +static bool disable_runtime = IS_ENABLED(CONFIG_PREEMPT_RT); static int __init setup_noefi(char *arg) { disable_runtime = true; @@ -99,6 +99,9 @@ static int __init parse_efi_cmdline(char if (parse_option_str(str, "noruntime")) disable_runtime = true; + if (parse_option_str(str, "runtime")) + disable_runtime = false; + if (parse_option_str(str, "nosoftreserve")) set_bit(EFI_MEM_NO_SOFT_RESERVE, &efi.flags); Index: linux-5.6.17-rt10/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ linux-5.6.17-rt10/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -257,11 +257,9 @@ static int amdgpu_amdkfd_remove_eviction new->shared_count = k; /* Install the new fence list, seqcount provides the barriers */ - preempt_disable(); write_seqcount_begin(&resv->seq); RCU_INIT_POINTER(resv->fence, new); write_seqcount_end(&resv->seq); - preempt_enable(); /* Drop the references to the removed fences or move them to ef_list */ for (i = j, k = 0; i < old->shared_count; ++i) { Index: linux-5.6.17-rt10/drivers/gpu/drm/i915/display/intel_sprite.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/i915/display/intel_sprite.c +++ linux-5.6.17-rt10/drivers/gpu/drm/i915/display/intel_sprite.c @@ -38,6 +38,7 @@ #include #include #include +#include #include "i915_drv.h" #include "i915_trace.h" @@ -67,6 +68,8 @@ int intel_usecs_to_scanlines(const struc #define VBLANK_EVASION_TIME_US 100 #endif +static DEFINE_LOCAL_IRQ_LOCK(pipe_update_lock); + /** * intel_pipe_update_start() - start update of a set of display registers * @new_crtc_state: the new crtc state @@ -116,7 +119,7 @@ void intel_pipe_update_start(const struc DRM_ERROR("PSR idle timed out 0x%x, atomic update may fail\n", psr_status); - local_irq_disable(); + local_lock_irq(pipe_update_lock); crtc->debug.min_vbl = min; crtc->debug.max_vbl = max; @@ -140,11 +143,11 @@ void intel_pipe_update_start(const struc break; } - local_irq_enable(); + local_unlock_irq(pipe_update_lock); timeout = schedule_timeout(timeout); - local_irq_disable(); + local_lock_irq(pipe_update_lock); } finish_wait(wq, &wait); @@ -177,7 +180,7 @@ void intel_pipe_update_start(const struc return; irq_disable: - local_irq_disable(); + local_lock_irq(pipe_update_lock); } /** @@ -214,7 +217,7 @@ void intel_pipe_update_end(struct intel_ new_crtc_state->uapi.event = NULL; } - local_irq_enable(); + local_unlock_irq(pipe_update_lock); if (intel_vgpu_active(dev_priv)) return; Index: linux-5.6.17-rt10/drivers/gpu/drm/i915/gt/intel_engine_pm.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/i915/gt/intel_engine_pm.c +++ linux-5.6.17-rt10/drivers/gpu/drm/i915/gt/intel_engine_pm.c @@ -64,12 +64,15 @@ static int __engine_unpark(struct intel_ } #if IS_ENABLED(CONFIG_LOCKDEP) +#include + +static DEFINE_LOCAL_IRQ_LOCK(timeline_lock); static inline unsigned long __timeline_mark_lock(struct intel_context *ce) { unsigned long flags; - local_irq_save(flags); + local_lock_irqsave(timeline_lock, flags); mutex_acquire(&ce->timeline->mutex.dep_map, 2, 0, _THIS_IP_); return flags; @@ -79,7 +82,7 @@ static inline void __timeline_mark_unloc unsigned long flags) { mutex_release(&ce->timeline->mutex.dep_map, _THIS_IP_); - local_irq_restore(flags); + local_unlock_irqrestore(timeline_lock, flags); } #else Index: linux-5.6.17-rt10/drivers/gpu/drm/i915/i915_irq.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/i915/i915_irq.c +++ linux-5.6.17-rt10/drivers/gpu/drm/i915/i915_irq.c @@ -803,6 +803,7 @@ bool i915_get_crtc_scanoutpos(struct drm spin_lock_irqsave(&dev_priv->uncore.lock, irqflags); /* preempt_disable_rt() should go right here in PREEMPT_RT patchset. */ + preempt_disable_rt(); /* Get optional system timestamp before query. */ if (stime) @@ -854,6 +855,7 @@ bool i915_get_crtc_scanoutpos(struct drm *etime = ktime_get(); /* preempt_enable_rt() should go right here in PREEMPT_RT patchset. */ + preempt_enable_rt(); spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags); Index: linux-5.6.17-rt10/drivers/gpu/drm/i915/i915_trace.h =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/i915/i915_trace.h +++ linux-5.6.17-rt10/drivers/gpu/drm/i915/i915_trace.h @@ -2,6 +2,10 @@ #if !defined(_I915_TRACE_H_) || defined(TRACE_HEADER_MULTI_READ) #define _I915_TRACE_H_ +#ifdef CONFIG_PREEMPT_RT +#define NOTRACE +#endif + #include #include #include @@ -716,7 +720,7 @@ DEFINE_EVENT(i915_request, i915_request_ TP_ARGS(rq) ); -#if defined(CONFIG_DRM_I915_LOW_LEVEL_TRACEPOINTS) +#if defined(CONFIG_DRM_I915_LOW_LEVEL_TRACEPOINTS) && !defined(NOTRACE) DEFINE_EVENT(i915_request, i915_request_submit, TP_PROTO(struct i915_request *rq), TP_ARGS(rq) Index: linux-5.6.17-rt10/drivers/gpu/drm/radeon/radeon_display.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/radeon/radeon_display.c +++ linux-5.6.17-rt10/drivers/gpu/drm/radeon/radeon_display.c @@ -1812,6 +1812,7 @@ int radeon_get_crtc_scanoutpos(struct dr struct radeon_device *rdev = dev->dev_private; /* preempt_disable_rt() should go right here in PREEMPT_RT patchset. */ + preempt_disable_rt(); /* Get optional system timestamp before query. */ if (stime) @@ -1904,6 +1905,7 @@ int radeon_get_crtc_scanoutpos(struct dr *etime = ktime_get(); /* preempt_enable_rt() should go right here in PREEMPT_RT patchset. */ + preempt_enable_rt(); /* Decode into vertical and horizontal scanout position. */ *vpos = position & 0x1fff; Index: linux-5.6.17-rt10/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ linux-5.6.17-rt10/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -480,19 +480,6 @@ struct vmw_private { bool has_sm4_1; /* - * VGA registers. - */ - - struct vmw_vga_topology_state vga_save[VMWGFX_MAX_DISPLAYS]; - uint32_t vga_width; - uint32_t vga_height; - uint32_t vga_bpp; - uint32_t vga_bpl; - uint32_t vga_pitchlock; - - uint32_t num_displays; - - /* * Framebuffer info. */ @@ -900,7 +887,6 @@ extern void vmw_fifo_commit(struct vmw_p extern void vmw_fifo_commit_flush(struct vmw_private *dev_priv, uint32_t bytes); extern int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *seqno); -extern void vmw_fifo_ping_host_locked(struct vmw_private *, uint32_t reason); extern void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason); extern bool vmw_fifo_have_3d(struct vmw_private *dev_priv); extern bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv); @@ -947,7 +933,6 @@ extern struct ttm_placement vmw_mob_plac extern struct ttm_placement vmw_mob_ne_placement; extern struct ttm_placement vmw_nonfixed_placement; extern struct ttm_bo_driver vmw_bo_driver; -extern int vmw_dma_quiescent(struct drm_device *dev); extern int vmw_bo_map_dma(struct ttm_buffer_object *bo); extern void vmw_bo_unmap_dma(struct ttm_buffer_object *bo); extern const struct vmw_sg_table * @@ -1085,8 +1070,6 @@ int vmw_fb_on(struct vmw_private *vmw_pr int vmw_kms_init(struct vmw_private *dev_priv); int vmw_kms_close(struct vmw_private *dev_priv); -int vmw_kms_save_vga(struct vmw_private *vmw_priv); -int vmw_kms_restore_vga(struct vmw_private *vmw_priv); int vmw_kms_cursor_bypass_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); void vmw_kms_cursor_post_execbuf(struct vmw_private *dev_priv); @@ -1139,7 +1122,6 @@ int vmw_overlay_init(struct vmw_private int vmw_overlay_close(struct vmw_private *dev_priv); int vmw_overlay_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); -int vmw_overlay_stop_all(struct vmw_private *dev_priv); int vmw_overlay_resume_all(struct vmw_private *dev_priv); int vmw_overlay_pause_all(struct vmw_private *dev_priv); int vmw_overlay_claim(struct vmw_private *dev_priv, uint32_t *out); @@ -1186,10 +1168,6 @@ extern void vmw_otables_takedown(struct extern const struct vmw_user_resource_conv *user_context_converter; -extern int vmw_context_check(struct vmw_private *dev_priv, - struct ttm_object_file *tfile, - int id, - struct vmw_resource **p_res); extern int vmw_context_define_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int vmw_extended_context_define_ioctl(struct drm_device *dev, void *data, @@ -1219,7 +1197,6 @@ vmw_context_get_dx_query_mob(struct vmw_ extern const struct vmw_user_resource_conv *user_surface_converter; -extern void vmw_surface_res_free(struct vmw_resource *res); extern int vmw_surface_destroy_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int vmw_surface_define_ioctl(struct drm_device *dev, void *data, @@ -1230,11 +1207,6 @@ extern int vmw_gb_surface_define_ioctl(s struct drm_file *file_priv); extern int vmw_gb_surface_reference_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); -extern int vmw_surface_check(struct vmw_private *dev_priv, - struct ttm_object_file *tfile, - uint32_t handle, int *id); -extern int vmw_surface_validate(struct vmw_private *dev_priv, - struct vmw_surface *srf); int vmw_surface_gb_priv_define(struct drm_device *dev, uint32_t user_accounting_size, SVGA3dSurfaceAllFlags svga3d_flags, Index: linux-5.6.17-rt10/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c +++ linux-5.6.17-rt10/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c @@ -169,10 +169,8 @@ void vmw_fifo_ping_host(struct vmw_priva { u32 *fifo_mem = dev_priv->mmio_virt; - preempt_disable(); if (cmpxchg(fifo_mem + SVGA_FIFO_BUSY, 0, 1) == 0) vmw_write(dev_priv, SVGA_REG_SYNC, reason); - preempt_enable(); } void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo) Index: linux-5.6.17-rt10/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +++ linux-5.6.17-rt10/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c @@ -1897,87 +1897,6 @@ int vmw_kms_write_svga(struct vmw_privat return 0; } -int vmw_kms_save_vga(struct vmw_private *vmw_priv) -{ - struct vmw_vga_topology_state *save; - uint32_t i; - - vmw_priv->vga_width = vmw_read(vmw_priv, SVGA_REG_WIDTH); - vmw_priv->vga_height = vmw_read(vmw_priv, SVGA_REG_HEIGHT); - vmw_priv->vga_bpp = vmw_read(vmw_priv, SVGA_REG_BITS_PER_PIXEL); - if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK) - vmw_priv->vga_pitchlock = - vmw_read(vmw_priv, SVGA_REG_PITCHLOCK); - else if (vmw_fifo_have_pitchlock(vmw_priv)) - vmw_priv->vga_pitchlock = vmw_mmio_read(vmw_priv->mmio_virt + - SVGA_FIFO_PITCHLOCK); - - if (!(vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY)) - return 0; - - vmw_priv->num_displays = vmw_read(vmw_priv, - SVGA_REG_NUM_GUEST_DISPLAYS); - - if (vmw_priv->num_displays == 0) - vmw_priv->num_displays = 1; - - for (i = 0; i < vmw_priv->num_displays; ++i) { - save = &vmw_priv->vga_save[i]; - vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, i); - save->primary = vmw_read(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY); - save->pos_x = vmw_read(vmw_priv, SVGA_REG_DISPLAY_POSITION_X); - save->pos_y = vmw_read(vmw_priv, SVGA_REG_DISPLAY_POSITION_Y); - save->width = vmw_read(vmw_priv, SVGA_REG_DISPLAY_WIDTH); - save->height = vmw_read(vmw_priv, SVGA_REG_DISPLAY_HEIGHT); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); - if (i == 0 && vmw_priv->num_displays == 1 && - save->width == 0 && save->height == 0) { - - /* - * It should be fairly safe to assume that these - * values are uninitialized. - */ - - save->width = vmw_priv->vga_width - save->pos_x; - save->height = vmw_priv->vga_height - save->pos_y; - } - } - - return 0; -} - -int vmw_kms_restore_vga(struct vmw_private *vmw_priv) -{ - struct vmw_vga_topology_state *save; - uint32_t i; - - vmw_write(vmw_priv, SVGA_REG_WIDTH, vmw_priv->vga_width); - vmw_write(vmw_priv, SVGA_REG_HEIGHT, vmw_priv->vga_height); - vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, vmw_priv->vga_bpp); - if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK) - vmw_write(vmw_priv, SVGA_REG_PITCHLOCK, - vmw_priv->vga_pitchlock); - else if (vmw_fifo_have_pitchlock(vmw_priv)) - vmw_mmio_write(vmw_priv->vga_pitchlock, - vmw_priv->mmio_virt + SVGA_FIFO_PITCHLOCK); - - if (!(vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY)) - return 0; - - for (i = 0; i < vmw_priv->num_displays; ++i) { - save = &vmw_priv->vga_save[i]; - vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, i); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY, save->primary); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_X, save->pos_x); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_Y, save->pos_y); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_WIDTH, save->width); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_HEIGHT, save->height); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); - } - - return 0; -} - bool vmw_kms_validate_mode_vram(struct vmw_private *dev_priv, uint32_t pitch, uint32_t height) Index: linux-5.6.17-rt10/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c +++ linux-5.6.17-rt10/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c @@ -354,37 +354,6 @@ static int vmw_overlay_update_stream(str } /** - * Stop all streams. - * - * Used by the fb code when starting. - * - * Takes the overlay lock. - */ -int vmw_overlay_stop_all(struct vmw_private *dev_priv) -{ - struct vmw_overlay *overlay = dev_priv->overlay_priv; - int i, ret; - - if (!overlay) - return 0; - - mutex_lock(&overlay->mutex); - - for (i = 0; i < VMW_MAX_NUM_STREAMS; i++) { - struct vmw_stream *stream = &overlay->stream[i]; - if (!stream->buf) - continue; - - ret = vmw_overlay_stop(dev_priv, i, false, false); - WARN_ON(ret != 0); - } - - mutex_unlock(&overlay->mutex); - - return 0; -} - -/** * Try to resume all paused streams. * * Used by the kms code after moving a new scanout buffer to vram. Index: linux-5.6.17-rt10/drivers/hv/hyperv_vmbus.h =================================================================== --- linux-5.6.17-rt10.orig/drivers/hv/hyperv_vmbus.h +++ linux-5.6.17-rt10/drivers/hv/hyperv_vmbus.h @@ -18,6 +18,7 @@ #include #include #include +#include #include "hv_trace.h" Index: linux-5.6.17-rt10/drivers/hv/vmbus_drv.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/hv/vmbus_drv.c +++ linux-5.6.17-rt10/drivers/hv/vmbus_drv.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include @@ -1247,6 +1248,8 @@ static void vmbus_isr(void) void *page_addr = hv_cpu->synic_event_page; struct hv_message *msg; union hv_synic_event_flags *event; + struct pt_regs *regs = get_irq_regs(); + u64 ip = regs ? instruction_pointer(regs) : 0; bool handled = false; if (unlikely(page_addr == NULL)) @@ -1291,7 +1294,7 @@ static void vmbus_isr(void) tasklet_schedule(&hv_cpu->msg_dpc); } - add_interrupt_randomness(HYPERVISOR_CALLBACK_VECTOR, 0); + add_interrupt_randomness(HYPERVISOR_CALLBACK_VECTOR, 0, ip); } /* Index: linux-5.6.17-rt10/drivers/leds/trigger/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/drivers/leds/trigger/Kconfig +++ linux-5.6.17-rt10/drivers/leds/trigger/Kconfig @@ -64,6 +64,7 @@ config LEDS_TRIGGER_BACKLIGHT config LEDS_TRIGGER_CPU bool "LED CPU Trigger" + depends on !PREEMPT_RT help This allows LEDs to be controlled by active CPUs. This shows the active CPUs across an array of LEDs so you can see which Index: linux-5.6.17-rt10/drivers/md/raid5.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/md/raid5.c +++ linux-5.6.17-rt10/drivers/md/raid5.c @@ -2058,8 +2058,9 @@ static void raid_run_ops(struct stripe_h struct raid5_percpu *percpu; unsigned long cpu; - cpu = get_cpu(); + cpu = get_cpu_light(); percpu = per_cpu_ptr(conf->percpu, cpu); + spin_lock(&percpu->lock); if (test_bit(STRIPE_OP_BIOFILL, &ops_request)) { ops_run_biofill(sh); overlap_clear++; @@ -2118,7 +2119,8 @@ static void raid_run_ops(struct stripe_h if (test_and_clear_bit(R5_Overlap, &dev->flags)) wake_up(&sh->raid_conf->wait_for_overlap); } - put_cpu(); + spin_unlock(&percpu->lock); + put_cpu_light(); } static void free_stripe(struct kmem_cache *sc, struct stripe_head *sh) @@ -6815,6 +6817,7 @@ static int raid456_cpu_up_prepare(unsign __func__, cpu); return -ENOMEM; } + spin_lock_init(&per_cpu_ptr(conf->percpu, cpu)->lock); return 0; } @@ -6929,7 +6932,7 @@ static struct r5conf *setup_conf(struct } else goto abort; spin_lock_init(&conf->device_lock); - seqcount_init(&conf->gen_lock); + seqcount_spinlock_init(&conf->gen_lock, &conf->device_lock); mutex_init(&conf->cache_size_mutex); init_waitqueue_head(&conf->wait_for_quiescent); init_waitqueue_head(&conf->wait_for_stripe); Index: linux-5.6.17-rt10/drivers/md/raid5.h =================================================================== --- linux-5.6.17-rt10.orig/drivers/md/raid5.h +++ linux-5.6.17-rt10/drivers/md/raid5.h @@ -589,7 +589,7 @@ struct r5conf { int prev_chunk_sectors; int prev_algo; short generation; /* increments with every reshape */ - seqcount_t gen_lock; /* lock against generation changes */ + seqcount_spinlock_t gen_lock; /* lock against generation changes */ unsigned long reshape_checkpoint; /* Time we last updated * metadata */ long long min_offset_diff; /* minimum difference between @@ -634,6 +634,7 @@ struct r5conf { int recovery_disabled; /* per cpu variables */ struct raid5_percpu { + spinlock_t lock; /* Protection for -RT */ struct page *spare_page; /* Used when checking P/Q in raid6 */ void *scribble; /* space for constructing buffer * lists and performing address Index: linux-5.6.17-rt10/drivers/net/phy/fixed_phy.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/net/phy/fixed_phy.c +++ linux-5.6.17-rt10/drivers/net/phy/fixed_phy.c @@ -19,7 +19,6 @@ #include #include #include -#include #include #include #include @@ -34,7 +33,6 @@ struct fixed_mdio_bus { struct fixed_phy { int addr; struct phy_device *phydev; - seqcount_t seqcount; struct fixed_phy_status status; bool no_carrier; int (*link_update)(struct net_device *, struct fixed_phy_status *); @@ -80,19 +78,17 @@ static int fixed_mdio_read(struct mii_bu list_for_each_entry(fp, &fmb->phys, node) { if (fp->addr == phy_addr) { struct fixed_phy_status state; - int s; - do { - s = read_seqcount_begin(&fp->seqcount); - fp->status.link = !fp->no_carrier; - /* Issue callback if user registered it. */ - if (fp->link_update) - fp->link_update(fp->phydev->attached_dev, - &fp->status); - /* Check the GPIO for change in status */ - fixed_phy_update(fp); - state = fp->status; - } while (read_seqcount_retry(&fp->seqcount, s)); + fp->status.link = !fp->no_carrier; + + /* Issue callback if user registered it. */ + if (fp->link_update) + fp->link_update(fp->phydev->attached_dev, + &fp->status); + + /* Check the GPIO for change in status */ + fixed_phy_update(fp); + state = fp->status; return swphy_read_reg(reg_num, &state); } @@ -150,8 +146,6 @@ static int fixed_phy_add_gpiod(unsigned if (!fp) return -ENOMEM; - seqcount_init(&fp->seqcount); - if (irq != PHY_POLL) fmb->mii_bus->irq[phy_addr] = irq; Index: linux-5.6.17-rt10/drivers/net/phy/mdio_bus.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/net/phy/mdio_bus.c +++ linux-5.6.17-rt10/drivers/net/phy/mdio_bus.c @@ -740,6 +740,7 @@ EXPORT_SYMBOL(mdiobus_scan); static void mdiobus_stats_acct(struct mdio_bus_stats *stats, bool op, int ret) { + preempt_disable(); u64_stats_update_begin(&stats->syncp); u64_stats_inc(&stats->transfers); @@ -754,6 +755,7 @@ static void mdiobus_stats_acct(struct md u64_stats_inc(&stats->writes); out: u64_stats_update_end(&stats->syncp); + preempt_enable(); } /** Index: linux-5.6.17-rt10/drivers/net/wireless/intersil/orinoco/orinoco_usb.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/net/wireless/intersil/orinoco/orinoco_usb.c +++ linux-5.6.17-rt10/drivers/net/wireless/intersil/orinoco/orinoco_usb.c @@ -693,8 +693,8 @@ static void ezusb_req_ctx_wait(struct ez while (!ctx->done.done && msecs--) udelay(1000); } else { - wait_event_interruptible(ctx->done.wait, - ctx->done.done); + swait_event_interruptible_exclusive(ctx->done.wait, + ctx->done.done); } break; default: Index: linux-5.6.17-rt10/drivers/pci/switch/switchtec.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/pci/switch/switchtec.c +++ linux-5.6.17-rt10/drivers/pci/switch/switchtec.c @@ -52,10 +52,11 @@ struct switchtec_user { enum mrpc_state state; - struct completion comp; + wait_queue_head_t cmd_comp; struct kref kref; struct list_head list; + bool cmd_done; u32 cmd; u32 status; u32 return_code; @@ -77,7 +78,7 @@ static struct switchtec_user *stuser_cre stuser->stdev = stdev; kref_init(&stuser->kref); INIT_LIST_HEAD(&stuser->list); - init_completion(&stuser->comp); + init_waitqueue_head(&stuser->cmd_comp); stuser->event_cnt = atomic_read(&stdev->event_cnt); dev_dbg(&stdev->dev, "%s: %p\n", __func__, stuser); @@ -175,7 +176,7 @@ static int mrpc_queue_cmd(struct switcht kref_get(&stuser->kref); stuser->read_len = sizeof(stuser->data); stuser_set_state(stuser, MRPC_QUEUED); - reinit_completion(&stuser->comp); + stuser->cmd_done = false; list_add_tail(&stuser->list, &stdev->mrpc_queue); mrpc_cmd_submit(stdev); @@ -222,7 +223,8 @@ static void mrpc_complete_cmd(struct swi memcpy_fromio(stuser->data, &stdev->mmio_mrpc->output_data, stuser->read_len); out: - complete_all(&stuser->comp); + stuser->cmd_done = true; + wake_up_interruptible(&stuser->cmd_comp); list_del_init(&stuser->list); stuser_put(stuser); stdev->mrpc_busy = 0; @@ -529,10 +531,11 @@ static ssize_t switchtec_dev_read(struct mutex_unlock(&stdev->mrpc_mutex); if (filp->f_flags & O_NONBLOCK) { - if (!try_wait_for_completion(&stuser->comp)) + if (!READ_ONCE(stuser->cmd_done)) return -EAGAIN; } else { - rc = wait_for_completion_interruptible(&stuser->comp); + rc = wait_event_interruptible(stuser->cmd_comp, + stuser->cmd_done); if (rc < 0) return rc; } @@ -580,7 +583,7 @@ static __poll_t switchtec_dev_poll(struc struct switchtec_dev *stdev = stuser->stdev; __poll_t ret = 0; - poll_wait(filp, &stuser->comp.wait, wait); + poll_wait(filp, &stuser->cmd_comp, wait); poll_wait(filp, &stdev->event_wq, wait); if (lock_mutex_and_test_alive(stdev)) @@ -588,7 +591,7 @@ static __poll_t switchtec_dev_poll(struc mutex_unlock(&stdev->mrpc_mutex); - if (try_wait_for_completion(&stuser->comp)) + if (READ_ONCE(stuser->cmd_done)) ret |= EPOLLIN | EPOLLRDNORM; if (stuser->event_cnt != atomic_read(&stdev->event_cnt)) @@ -1272,7 +1275,8 @@ static void stdev_kill(struct switchtec_ /* Wake up and kill any users waiting on an MRPC request */ list_for_each_entry_safe(stuser, tmpuser, &stdev->mrpc_queue, list) { - complete_all(&stuser->comp); + stuser->cmd_done = true; + wake_up_interruptible(&stuser->cmd_comp); list_del_init(&stuser->list); stuser_put(stuser); } Index: linux-5.6.17-rt10/drivers/scsi/fcoe/fcoe.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/scsi/fcoe/fcoe.c +++ linux-5.6.17-rt10/drivers/scsi/fcoe/fcoe.c @@ -1452,11 +1452,11 @@ err2: static int fcoe_alloc_paged_crc_eof(struct sk_buff *skb, int tlen) { struct fcoe_percpu_s *fps; - int rc; + int rc, cpu = get_cpu_light(); - fps = &get_cpu_var(fcoe_percpu); + fps = &per_cpu(fcoe_percpu, cpu); rc = fcoe_get_paged_crc_eof(skb, tlen, fps); - put_cpu_var(fcoe_percpu); + put_cpu_light(); return rc; } @@ -1641,11 +1641,11 @@ static inline int fcoe_filter_frames(str return 0; } - stats = per_cpu_ptr(lport->stats, get_cpu()); + stats = per_cpu_ptr(lport->stats, get_cpu_light()); stats->InvalidCRCCount++; if (stats->InvalidCRCCount < 5) printk(KERN_WARNING "fcoe: dropping frame with CRC error\n"); - put_cpu(); + put_cpu_light(); return -EINVAL; } @@ -1686,7 +1686,7 @@ static void fcoe_recv_frame(struct sk_bu */ hp = (struct fcoe_hdr *) skb_network_header(skb); - stats = per_cpu_ptr(lport->stats, get_cpu()); + stats = per_cpu_ptr(lport->stats, get_cpu_light()); if (unlikely(FC_FCOE_DECAPS_VER(hp) != FC_FCOE_VER)) { if (stats->ErrorFrames < 5) printk(KERN_WARNING "fcoe: FCoE version " @@ -1718,13 +1718,13 @@ static void fcoe_recv_frame(struct sk_bu goto drop; if (!fcoe_filter_frames(lport, fp)) { - put_cpu(); + put_cpu_light(); fc_exch_recv(lport, fp); return; } drop: stats->ErrorFrames++; - put_cpu(); + put_cpu_light(); kfree_skb(skb); } Index: linux-5.6.17-rt10/drivers/scsi/fcoe/fcoe_ctlr.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/scsi/fcoe/fcoe_ctlr.c +++ linux-5.6.17-rt10/drivers/scsi/fcoe/fcoe_ctlr.c @@ -826,7 +826,7 @@ static unsigned long fcoe_ctlr_age_fcfs( INIT_LIST_HEAD(&del_list); - stats = per_cpu_ptr(fip->lp->stats, get_cpu()); + stats = per_cpu_ptr(fip->lp->stats, get_cpu_light()); list_for_each_entry_safe(fcf, next, &fip->fcfs, list) { deadline = fcf->time + fcf->fka_period + fcf->fka_period / 2; @@ -862,7 +862,7 @@ static unsigned long fcoe_ctlr_age_fcfs( sel_time = fcf->time; } } - put_cpu(); + put_cpu_light(); list_for_each_entry_safe(fcf, next, &del_list, list) { /* Removes fcf from current list */ Index: linux-5.6.17-rt10/drivers/scsi/libfc/fc_exch.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/scsi/libfc/fc_exch.c +++ linux-5.6.17-rt10/drivers/scsi/libfc/fc_exch.c @@ -821,10 +821,10 @@ static struct fc_exch *fc_exch_em_alloc( } memset(ep, 0, sizeof(*ep)); - cpu = get_cpu(); + cpu = get_cpu_light(); pool = per_cpu_ptr(mp->pool, cpu); spin_lock_bh(&pool->lock); - put_cpu(); + put_cpu_light(); /* peek cache of free slot */ if (pool->left != FC_XID_UNKNOWN) { Index: linux-5.6.17-rt10/drivers/thermal/intel/x86_pkg_temp_thermal.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/thermal/intel/x86_pkg_temp_thermal.c +++ linux-5.6.17-rt10/drivers/thermal/intel/x86_pkg_temp_thermal.c @@ -63,7 +63,7 @@ static int max_id __read_mostly; /* Array of zone pointers */ static struct zone_device **zones; /* Serializes interrupt notification, work and hotplug */ -static DEFINE_SPINLOCK(pkg_temp_lock); +static DEFINE_RAW_SPINLOCK(pkg_temp_lock); /* Protects zone operation in the work function against hotplug removal */ static DEFINE_MUTEX(thermal_zone_mutex); @@ -266,12 +266,12 @@ static void pkg_temp_thermal_threshold_w u64 msr_val, wr_val; mutex_lock(&thermal_zone_mutex); - spin_lock_irq(&pkg_temp_lock); + raw_spin_lock_irq(&pkg_temp_lock); ++pkg_work_cnt; zonedev = pkg_temp_thermal_get_dev(cpu); if (!zonedev) { - spin_unlock_irq(&pkg_temp_lock); + raw_spin_unlock_irq(&pkg_temp_lock); mutex_unlock(&thermal_zone_mutex); return; } @@ -285,7 +285,7 @@ static void pkg_temp_thermal_threshold_w } enable_pkg_thres_interrupt(); - spin_unlock_irq(&pkg_temp_lock); + raw_spin_unlock_irq(&pkg_temp_lock); /* * If tzone is not NULL, then thermal_zone_mutex will prevent the @@ -310,7 +310,7 @@ static int pkg_thermal_notify(u64 msr_va struct zone_device *zonedev; unsigned long flags; - spin_lock_irqsave(&pkg_temp_lock, flags); + raw_spin_lock_irqsave(&pkg_temp_lock, flags); ++pkg_interrupt_cnt; disable_pkg_thres_interrupt(); @@ -322,7 +322,7 @@ static int pkg_thermal_notify(u64 msr_va pkg_thermal_schedule_work(zonedev->cpu, &zonedev->work); } - spin_unlock_irqrestore(&pkg_temp_lock, flags); + raw_spin_unlock_irqrestore(&pkg_temp_lock, flags); return 0; } @@ -368,9 +368,9 @@ static int pkg_temp_thermal_device_add(u zonedev->msr_pkg_therm_high); cpumask_set_cpu(cpu, &zonedev->cpumask); - spin_lock_irq(&pkg_temp_lock); + raw_spin_lock_irq(&pkg_temp_lock); zones[id] = zonedev; - spin_unlock_irq(&pkg_temp_lock); + raw_spin_unlock_irq(&pkg_temp_lock); return 0; } @@ -407,7 +407,7 @@ static int pkg_thermal_cpu_offline(unsig } /* Protect against work and interrupts */ - spin_lock_irq(&pkg_temp_lock); + raw_spin_lock_irq(&pkg_temp_lock); /* * Check whether this cpu was the current target and store the new @@ -439,9 +439,9 @@ static int pkg_thermal_cpu_offline(unsig * To cancel the work we need to drop the lock, otherwise * we might deadlock if the work needs to be flushed. */ - spin_unlock_irq(&pkg_temp_lock); + raw_spin_unlock_irq(&pkg_temp_lock); cancel_delayed_work_sync(&zonedev->work); - spin_lock_irq(&pkg_temp_lock); + raw_spin_lock_irq(&pkg_temp_lock); /* * If this is not the last cpu in the package and the work * did not run after we dropped the lock above, then we @@ -452,7 +452,7 @@ static int pkg_thermal_cpu_offline(unsig pkg_thermal_schedule_work(target, &zonedev->work); } - spin_unlock_irq(&pkg_temp_lock); + raw_spin_unlock_irq(&pkg_temp_lock); /* Final cleanup if this is the last cpu */ if (lastcpu) Index: linux-5.6.17-rt10/drivers/tty/serial/8250/8250.h =================================================================== --- linux-5.6.17-rt10.orig/drivers/tty/serial/8250/8250.h +++ linux-5.6.17-rt10/drivers/tty/serial/8250/8250.h @@ -130,12 +130,55 @@ static inline void serial_dl_write(struc up->dl_write(up, value); } +static inline void serial8250_set_IER(struct uart_8250_port *up, + unsigned char ier) +{ + struct uart_port *port = &up->port; + unsigned int flags; + bool is_console; + + is_console = uart_console(port); + + if (is_console) + console_atomic_lock(&flags); + + serial_out(up, UART_IER, ier); + + if (is_console) + console_atomic_unlock(flags); +} + +static inline unsigned char serial8250_clear_IER(struct uart_8250_port *up) +{ + struct uart_port *port = &up->port; + unsigned int clearval = 0; + unsigned int prior; + unsigned int flags; + bool is_console; + + is_console = uart_console(port); + + if (up->capabilities & UART_CAP_UUE) + clearval = UART_IER_UUE; + + if (is_console) + console_atomic_lock(&flags); + + prior = serial_port_in(port, UART_IER); + serial_port_out(port, UART_IER, clearval); + + if (is_console) + console_atomic_unlock(flags); + + return prior; +} + static inline bool serial8250_set_THRI(struct uart_8250_port *up) { if (up->ier & UART_IER_THRI) return false; up->ier |= UART_IER_THRI; - serial_out(up, UART_IER, up->ier); + serial8250_set_IER(up, up->ier); return true; } @@ -144,7 +187,7 @@ static inline bool serial8250_clear_THRI if (!(up->ier & UART_IER_THRI)) return false; up->ier &= ~UART_IER_THRI; - serial_out(up, UART_IER, up->ier); + serial8250_set_IER(up, up->ier); return true; } Index: linux-5.6.17-rt10/drivers/tty/serial/8250/8250_core.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/tty/serial/8250/8250_core.c +++ linux-5.6.17-rt10/drivers/tty/serial/8250/8250_core.c @@ -274,10 +274,8 @@ static void serial8250_backup_timeout(st * Must disable interrupts or else we risk racing with the interrupt * based handler. */ - if (up->port.irq) { - ier = serial_in(up, UART_IER); - serial_out(up, UART_IER, 0); - } + if (up->port.irq) + ier = serial8250_clear_IER(up); iir = serial_in(up, UART_IIR); @@ -300,7 +298,7 @@ static void serial8250_backup_timeout(st serial8250_tx_chars(up); if (up->port.irq) - serial_out(up, UART_IER, ier); + serial8250_set_IER(up, ier); spin_unlock_irqrestore(&up->port.lock, flags); @@ -578,6 +576,14 @@ serial8250_register_ports(struct uart_dr #ifdef CONFIG_SERIAL_8250_CONSOLE +static void univ8250_console_write_atomic(struct console *co, const char *s, + unsigned int count) +{ + struct uart_8250_port *up = &serial8250_ports[co->index]; + + serial8250_console_write_atomic(up, s, count); +} + static void univ8250_console_write(struct console *co, const char *s, unsigned int count) { @@ -663,6 +669,7 @@ static int univ8250_console_match(struct static struct console univ8250_console = { .name = "ttyS", + .write_atomic = univ8250_console_write_atomic, .write = univ8250_console_write, .device = uart_console_device, .setup = univ8250_console_setup, Index: linux-5.6.17-rt10/drivers/tty/serial/8250/8250_fsl.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/tty/serial/8250/8250_fsl.c +++ linux-5.6.17-rt10/drivers/tty/serial/8250/8250_fsl.c @@ -53,9 +53,18 @@ int fsl8250_handle_irq(struct uart_port /* Stop processing interrupts on input overrun */ if ((orig_lsr & UART_LSR_OE) && (up->overrun_backoff_time_ms > 0)) { + unsigned int ca_flags; unsigned long delay; + bool is_console; + is_console = uart_console(port); + + if (is_console) + console_atomic_lock(&ca_flags); up->ier = port->serial_in(port, UART_IER); + if (is_console) + console_atomic_unlock(ca_flags); + if (up->ier & (UART_IER_RLSI | UART_IER_RDI)) { port->ops->stop_rx(port); } else { Index: linux-5.6.17-rt10/drivers/tty/serial/8250/8250_ingenic.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/tty/serial/8250/8250_ingenic.c +++ linux-5.6.17-rt10/drivers/tty/serial/8250/8250_ingenic.c @@ -146,6 +146,8 @@ OF_EARLYCON_DECLARE(x1000_uart, "ingenic static void ingenic_uart_serial_out(struct uart_port *p, int offset, int value) { + unsigned int flags; + bool is_console; int ier; switch (offset) { @@ -167,7 +169,12 @@ static void ingenic_uart_serial_out(stru * If we have enabled modem status IRQs we should enable * modem mode. */ + is_console = uart_console(p); + if (is_console) + console_atomic_lock(&flags); ier = p->serial_in(p, UART_IER); + if (is_console) + console_atomic_unlock(flags); if (ier & UART_IER_MSI) value |= UART_MCR_MDCE | UART_MCR_FCM; Index: linux-5.6.17-rt10/drivers/tty/serial/8250/8250_mtk.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/tty/serial/8250/8250_mtk.c +++ linux-5.6.17-rt10/drivers/tty/serial/8250/8250_mtk.c @@ -212,12 +212,37 @@ static void mtk8250_shutdown(struct uart static void mtk8250_disable_intrs(struct uart_8250_port *up, int mask) { - serial_out(up, UART_IER, serial_in(up, UART_IER) & (~mask)); + struct uart_port *port = &up->port; + unsigned int flags; + unsigned int ier; + bool is_console; + + is_console = uart_console(port); + + if (is_console) + console_atomic_lock(&flags); + + ier = serial_in(up, UART_IER); + serial_out(up, UART_IER, ier & (~mask)); + + if (is_console) + console_atomic_unlock(flags); } static void mtk8250_enable_intrs(struct uart_8250_port *up, int mask) { - serial_out(up, UART_IER, serial_in(up, UART_IER) | mask); + struct uart_port *port = &up->port; + unsigned int flags; + unsigned int ier; + + if (uart_console(port)) + console_atomic_lock(&flags); + + ier = serial_in(up, UART_IER); + serial_out(up, UART_IER, ier | mask); + + if (uart_console(port)) + console_atomic_unlock(flags); } static void mtk8250_set_flow_ctrl(struct uart_8250_port *up, int mode) Index: linux-5.6.17-rt10/drivers/tty/serial/8250/8250_port.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/tty/serial/8250/8250_port.c +++ linux-5.6.17-rt10/drivers/tty/serial/8250/8250_port.c @@ -717,7 +717,7 @@ static void serial8250_set_sleep(struct serial_out(p, UART_EFR, UART_EFR_ECB); serial_out(p, UART_LCR, 0); } - serial_out(p, UART_IER, sleep ? UART_IERX_SLEEP : 0); + serial8250_set_IER(p, sleep ? UART_IERX_SLEEP : 0); if (p->capabilities & UART_CAP_EFR) { serial_out(p, UART_LCR, UART_LCR_CONF_MODE_B); serial_out(p, UART_EFR, efr); @@ -1389,7 +1389,7 @@ static void serial8250_stop_rx(struct ua up->ier &= ~(UART_IER_RLSI | UART_IER_RDI); up->port.read_status_mask &= ~UART_LSR_DR; - serial_port_out(port, UART_IER, up->ier); + serial8250_set_IER(up, up->ier); serial8250_rpm_put(up); } @@ -1407,7 +1407,7 @@ static void __do_stop_tx_rs485(struct ua serial8250_clear_and_reinit_fifos(p); p->ier |= UART_IER_RLSI | UART_IER_RDI; - serial_port_out(&p->port, UART_IER, p->ier); + serial8250_set_IER(p, p->ier); } } static enum hrtimer_restart serial8250_em485_handle_stop_tx(struct hrtimer *t) @@ -1615,7 +1615,7 @@ static void serial8250_disable_ms(struct mctrl_gpio_disable_ms(up->gpios); up->ier &= ~UART_IER_MSI; - serial_port_out(port, UART_IER, up->ier); + serial8250_set_IER(up, up->ier); } static void serial8250_enable_ms(struct uart_port *port) @@ -1631,7 +1631,7 @@ static void serial8250_enable_ms(struct up->ier |= UART_IER_MSI; serial8250_rpm_get(up); - serial_port_out(port, UART_IER, up->ier); + serial8250_set_IER(up, up->ier); serial8250_rpm_put(up); } @@ -2025,14 +2025,7 @@ static void serial8250_put_poll_char(str struct uart_8250_port *up = up_to_u8250p(port); serial8250_rpm_get(up); - /* - * First save the IER then disable the interrupts - */ - ier = serial_port_in(port, UART_IER); - if (up->capabilities & UART_CAP_UUE) - serial_port_out(port, UART_IER, UART_IER_UUE); - else - serial_port_out(port, UART_IER, 0); + ier = serial8250_clear_IER(up); wait_for_xmitr(up, BOTH_EMPTY); /* @@ -2045,7 +2038,7 @@ static void serial8250_put_poll_char(str * and restore the IER */ wait_for_xmitr(up, BOTH_EMPTY); - serial_port_out(port, UART_IER, ier); + serial8250_set_IER(up, ier); serial8250_rpm_put(up); } @@ -2343,7 +2336,7 @@ void serial8250_do_shutdown(struct uart_ */ spin_lock_irqsave(&port->lock, flags); up->ier = 0; - serial_port_out(port, UART_IER, 0); + serial8250_set_IER(up, 0); spin_unlock_irqrestore(&port->lock, flags); synchronize_irq(port->irq); @@ -2628,7 +2621,7 @@ serial8250_do_set_termios(struct uart_po if (up->capabilities & UART_CAP_RTOIE) up->ier |= UART_IER_RTOIE; - serial_port_out(port, UART_IER, up->ier); + serial8250_set_IER(up, up->ier); if (up->capabilities & UART_CAP_EFR) { unsigned char efr = 0; @@ -3093,7 +3086,7 @@ EXPORT_SYMBOL_GPL(serial8250_set_default #ifdef CONFIG_SERIAL_8250_CONSOLE -static void serial8250_console_putchar(struct uart_port *port, int ch) +static void serial8250_console_putchar_locked(struct uart_port *port, int ch) { struct uart_8250_port *up = up_to_u8250p(port); @@ -3101,6 +3094,18 @@ static void serial8250_console_putchar(s serial_port_out(port, UART_TX, ch); } +static void serial8250_console_putchar(struct uart_port *port, int ch) +{ + struct uart_8250_port *up = up_to_u8250p(port); + unsigned int flags; + + wait_for_xmitr(up, UART_LSR_THRE); + + console_atomic_lock(&flags); + serial8250_console_putchar_locked(port, ch); + console_atomic_unlock(flags); +} + /* * Restore serial console when h/w power-off detected */ @@ -3122,6 +3127,32 @@ static void serial8250_console_restore(s serial8250_out_MCR(up, UART_MCR_DTR | UART_MCR_RTS); } +void serial8250_console_write_atomic(struct uart_8250_port *up, + const char *s, unsigned int count) +{ + struct uart_port *port = &up->port; + unsigned int flags; + unsigned int ier; + + console_atomic_lock(&flags); + + touch_nmi_watchdog(); + + ier = serial8250_clear_IER(up); + + if (atomic_fetch_inc(&up->console_printing)) { + uart_console_write(port, "\n", 1, + serial8250_console_putchar_locked); + } + uart_console_write(port, s, count, serial8250_console_putchar_locked); + atomic_dec(&up->console_printing); + + wait_for_xmitr(up, BOTH_EMPTY); + serial8250_set_IER(up, ier); + + console_atomic_unlock(flags); +} + /* * Print a string to the serial port trying not to disturb * any possible real use of the port... @@ -3134,26 +3165,13 @@ void serial8250_console_write(struct uar struct uart_port *port = &up->port; unsigned long flags; unsigned int ier; - int locked = 1; touch_nmi_watchdog(); serial8250_rpm_get(up); + spin_lock_irqsave(&port->lock, flags); - if (oops_in_progress) - locked = spin_trylock_irqsave(&port->lock, flags); - else - spin_lock_irqsave(&port->lock, flags); - - /* - * First save the IER then disable the interrupts - */ - ier = serial_port_in(port, UART_IER); - - if (up->capabilities & UART_CAP_UUE) - serial_port_out(port, UART_IER, UART_IER_UUE); - else - serial_port_out(port, UART_IER, 0); + ier = serial8250_clear_IER(up); /* check scratch reg to see if port powered off during system sleep */ if (up->canary && (up->canary != serial_port_in(port, UART_SCR))) { @@ -3161,14 +3179,16 @@ void serial8250_console_write(struct uar up->canary = 0; } + atomic_inc(&up->console_printing); uart_console_write(port, s, count, serial8250_console_putchar); + atomic_dec(&up->console_printing); /* * Finally, wait for transmitter to become empty * and restore the IER */ wait_for_xmitr(up, BOTH_EMPTY); - serial_port_out(port, UART_IER, ier); + serial8250_set_IER(up, ier); /* * The receive handling will happen properly because the @@ -3180,8 +3200,7 @@ void serial8250_console_write(struct uar if (up->msr_saved_flags) serial8250_modem_status(up); - if (locked) - spin_unlock_irqrestore(&port->lock, flags); + spin_unlock_irqrestore(&port->lock, flags); serial8250_rpm_put(up); } @@ -3202,6 +3221,7 @@ static unsigned int probe_baud(struct ua int serial8250_console_setup(struct uart_port *port, char *options, bool probe) { + struct uart_8250_port *up = up_to_u8250p(port); int baud = 9600; int bits = 8; int parity = 'n'; @@ -3210,6 +3230,8 @@ int serial8250_console_setup(struct uart if (!port->iobase && !port->membase) return -ENODEV; + atomic_set(&up->console_printing, 0); + if (options) uart_parse_options(options, &baud, &parity, &bits, &flow); else if (probe) Index: linux-5.6.17-rt10/drivers/tty/serial/amba-pl011.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/tty/serial/amba-pl011.c +++ linux-5.6.17-rt10/drivers/tty/serial/amba-pl011.c @@ -2198,18 +2198,24 @@ pl011_console_write(struct console *co, { struct uart_amba_port *uap = amba_ports[co->index]; unsigned int old_cr = 0, new_cr; - unsigned long flags; + unsigned long flags = 0; int locked = 1; clk_enable(uap->clk); - local_irq_save(flags); + /* + * local_irq_save(flags); + * + * This local_irq_save() is nonsense. If we come in via sysrq + * handling then interrupts are already disabled. Aside of + * that the port.sysrq check is racy on SMP regardless. + */ if (uap->port.sysrq) locked = 0; else if (oops_in_progress) - locked = spin_trylock(&uap->port.lock); + locked = spin_trylock_irqsave(&uap->port.lock, flags); else - spin_lock(&uap->port.lock); + spin_lock_irqsave(&uap->port.lock, flags); /* * First save the CR then disable the interrupts @@ -2235,8 +2241,7 @@ pl011_console_write(struct console *co, pl011_write(old_cr, uap, REG_CR); if (locked) - spin_unlock(&uap->port.lock); - local_irq_restore(flags); + spin_unlock_irqrestore(&uap->port.lock, flags); clk_disable(uap->clk); } Index: linux-5.6.17-rt10/drivers/tty/serial/omap-serial.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/tty/serial/omap-serial.c +++ linux-5.6.17-rt10/drivers/tty/serial/omap-serial.c @@ -1306,13 +1306,10 @@ serial_omap_console_write(struct console pm_runtime_get_sync(up->dev); - local_irq_save(flags); - if (up->port.sysrq) - locked = 0; - else if (oops_in_progress) - locked = spin_trylock(&up->port.lock); + if (up->port.sysrq || oops_in_progress) + locked = spin_trylock_irqsave(&up->port.lock, flags); else - spin_lock(&up->port.lock); + spin_lock_irqsave(&up->port.lock, flags); /* * First save the IER then disable the interrupts @@ -1341,8 +1338,7 @@ serial_omap_console_write(struct console pm_runtime_mark_last_busy(up->dev); pm_runtime_put_autosuspend(up->dev); if (locked) - spin_unlock(&up->port.lock); - local_irq_restore(flags); + spin_unlock_irqrestore(&up->port.lock, flags); } static int __init Index: linux-5.6.17-rt10/drivers/usb/gadget/function/f_fs.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/usb/gadget/function/f_fs.c +++ linux-5.6.17-rt10/drivers/usb/gadget/function/f_fs.c @@ -1704,7 +1704,7 @@ static void ffs_data_put(struct ffs_data pr_info("%s(): freeing\n", __func__); ffs_data_clear(ffs); BUG_ON(waitqueue_active(&ffs->ev.waitq) || - waitqueue_active(&ffs->ep0req_completion.wait) || + swait_active(&ffs->ep0req_completion.wait) || waitqueue_active(&ffs->wait)); destroy_workqueue(ffs->io_completion_wq); kfree(ffs->dev_name); Index: linux-5.6.17-rt10/drivers/usb/gadget/legacy/inode.c =================================================================== --- linux-5.6.17-rt10.orig/drivers/usb/gadget/legacy/inode.c +++ linux-5.6.17-rt10/drivers/usb/gadget/legacy/inode.c @@ -344,7 +344,7 @@ ep_io (struct ep_data *epdata, void *buf spin_unlock_irq (&epdata->dev->lock); if (likely (value == 0)) { - value = wait_event_interruptible (done.wait, done.done); + value = swait_event_interruptible_exclusive(done.wait, done.done); if (value != 0) { spin_lock_irq (&epdata->dev->lock); if (likely (epdata->ep != NULL)) { @@ -353,7 +353,7 @@ ep_io (struct ep_data *epdata, void *buf usb_ep_dequeue (epdata->ep, epdata->req); spin_unlock_irq (&epdata->dev->lock); - wait_event (done.wait, done.done); + swait_event_exclusive(done.wait, done.done); if (epdata->status == -ECONNRESET) epdata->status = -EINTR; } else { Index: linux-5.6.17-rt10/fs/afs/dir_silly.c =================================================================== --- linux-5.6.17-rt10.orig/fs/afs/dir_silly.c +++ linux-5.6.17-rt10/fs/afs/dir_silly.c @@ -210,7 +210,7 @@ int afs_silly_iput(struct dentry *dentry struct dentry *alias; int ret; - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); _enter("%p{%pd},%llx", dentry, dentry, vnode->fid.vnode); Index: linux-5.6.17-rt10/fs/buffer.c =================================================================== --- linux-5.6.17-rt10.orig/fs/buffer.c +++ linux-5.6.17-rt10/fs/buffer.c @@ -274,8 +274,7 @@ static void end_buffer_async_read(struct * decide that the page is now completely done. */ first = page_buffers(page); - local_irq_save(flags); - bit_spin_lock(BH_Uptodate_Lock, &first->b_state); + spin_lock_irqsave(&first->b_uptodate_lock, flags); clear_buffer_async_read(bh); unlock_buffer(bh); tmp = bh; @@ -288,8 +287,7 @@ static void end_buffer_async_read(struct } tmp = tmp->b_this_page; } while (tmp != bh); - bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); - local_irq_restore(flags); + spin_unlock_irqrestore(&first->b_uptodate_lock, flags); /* * If none of the buffers had errors and they are all @@ -301,8 +299,7 @@ static void end_buffer_async_read(struct return; still_busy: - bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); - local_irq_restore(flags); + spin_unlock_irqrestore(&first->b_uptodate_lock, flags); return; } @@ -371,8 +368,7 @@ void end_buffer_async_write(struct buffe } first = page_buffers(page); - local_irq_save(flags); - bit_spin_lock(BH_Uptodate_Lock, &first->b_state); + spin_lock_irqsave(&first->b_uptodate_lock, flags); clear_buffer_async_write(bh); unlock_buffer(bh); @@ -384,14 +380,12 @@ void end_buffer_async_write(struct buffe } tmp = tmp->b_this_page; } - bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); - local_irq_restore(flags); + spin_unlock_irqrestore(&first->b_uptodate_lock, flags); end_page_writeback(page); return; still_busy: - bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); - local_irq_restore(flags); + spin_unlock_irqrestore(&first->b_uptodate_lock, flags); return; } EXPORT_SYMBOL(end_buffer_async_write); @@ -3396,6 +3390,7 @@ struct buffer_head *alloc_buffer_head(gf struct buffer_head *ret = kmem_cache_zalloc(bh_cachep, gfp_flags); if (ret) { INIT_LIST_HEAD(&ret->b_assoc_buffers); + spin_lock_init(&ret->b_uptodate_lock); preempt_disable(); __this_cpu_inc(bh_accounting.nr); recalc_bh_state(); Index: linux-5.6.17-rt10/fs/cifs/readdir.c =================================================================== --- linux-5.6.17-rt10.orig/fs/cifs/readdir.c +++ linux-5.6.17-rt10/fs/cifs/readdir.c @@ -80,7 +80,7 @@ cifs_prime_dcache(struct dentry *parent, struct inode *inode; struct super_block *sb = parent->d_sb; struct cifs_sb_info *cifs_sb = CIFS_SB(sb); - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); cifs_dbg(FYI, "%s: for %s\n", __func__, name->name); Index: linux-5.6.17-rt10/fs/dcache.c =================================================================== --- linux-5.6.17-rt10.orig/fs/dcache.c +++ linux-5.6.17-rt10/fs/dcache.c @@ -1727,7 +1727,7 @@ static struct dentry *__d_alloc(struct s dentry->d_lockref.count = 1; dentry->d_flags = 0; spin_lock_init(&dentry->d_lock); - seqcount_init(&dentry->d_seq); + seqcount_spinlock_init(&dentry->d_seq, &dentry->d_lock); dentry->d_inode = NULL; dentry->d_parent = dentry; dentry->d_sb = sb; @@ -2484,9 +2484,10 @@ EXPORT_SYMBOL(d_rehash); static inline unsigned start_dir_add(struct inode *dir) { + preempt_disable_rt(); for (;;) { - unsigned n = dir->i_dir_seq; - if (!(n & 1) && cmpxchg(&dir->i_dir_seq, n, n + 1) == n) + unsigned n = dir->__i_dir_seq; + if (!(n & 1) && cmpxchg(&dir->__i_dir_seq, n, n + 1) == n) return n; cpu_relax(); } @@ -2494,26 +2495,30 @@ static inline unsigned start_dir_add(str static inline void end_dir_add(struct inode *dir, unsigned n) { - smp_store_release(&dir->i_dir_seq, n + 2); + smp_store_release(&dir->__i_dir_seq, n + 2); + preempt_enable_rt(); } static void d_wait_lookup(struct dentry *dentry) { - if (d_in_lookup(dentry)) { - DECLARE_WAITQUEUE(wait, current); - add_wait_queue(dentry->d_wait, &wait); - do { - set_current_state(TASK_UNINTERRUPTIBLE); - spin_unlock(&dentry->d_lock); - schedule(); - spin_lock(&dentry->d_lock); - } while (d_in_lookup(dentry)); - } + struct swait_queue __wait; + + if (!d_in_lookup(dentry)) + return; + + INIT_LIST_HEAD(&__wait.task_list); + do { + prepare_to_swait_exclusive(dentry->d_wait, &__wait, TASK_UNINTERRUPTIBLE); + spin_unlock(&dentry->d_lock); + schedule(); + spin_lock(&dentry->d_lock); + } while (d_in_lookup(dentry)); + finish_swait(dentry->d_wait, &__wait); } struct dentry *d_alloc_parallel(struct dentry *parent, const struct qstr *name, - wait_queue_head_t *wq) + struct swait_queue_head *wq) { unsigned int hash = name->hash; struct hlist_bl_head *b = in_lookup_hash(parent, hash); @@ -2527,7 +2532,7 @@ struct dentry *d_alloc_parallel(struct d retry: rcu_read_lock(); - seq = smp_load_acquire(&parent->d_inode->i_dir_seq); + seq = smp_load_acquire(&parent->d_inode->__i_dir_seq); r_seq = read_seqbegin(&rename_lock); dentry = __d_lookup_rcu(parent, name, &d_seq); if (unlikely(dentry)) { @@ -2555,7 +2560,7 @@ retry: } hlist_bl_lock(b); - if (unlikely(READ_ONCE(parent->d_inode->i_dir_seq) != seq)) { + if (unlikely(READ_ONCE(parent->d_inode->__i_dir_seq) != seq)) { hlist_bl_unlock(b); rcu_read_unlock(); goto retry; @@ -2628,7 +2633,7 @@ void __d_lookup_done(struct dentry *dent hlist_bl_lock(b); dentry->d_flags &= ~DCACHE_PAR_LOOKUP; __hlist_bl_del(&dentry->d_u.d_in_lookup_hash); - wake_up_all(dentry->d_wait); + swake_up_all(dentry->d_wait); dentry->d_wait = NULL; hlist_bl_unlock(b); INIT_HLIST_NODE(&dentry->d_u.d_alias); Index: linux-5.6.17-rt10/fs/eventpoll.c =================================================================== --- linux-5.6.17-rt10.orig/fs/eventpoll.c +++ linux-5.6.17-rt10/fs/eventpoll.c @@ -219,12 +219,18 @@ struct eventpoll { /* used to optimize loop detection check */ int visited; - struct list_head visited_list_link; #ifdef CONFIG_NET_RX_BUSY_POLL /* used to track busy poll napi_id */ unsigned int napi_id; #endif + + struct list_head visited_list_link; + +#ifdef CONFIG_DEBUG_LOCK_ALLOC + /* used to track wakeup nests for lockdep validation */ + u8 nests; +#endif }; /* Wait structure used by the poll hooks */ @@ -545,30 +551,43 @@ out_unlock: */ #ifdef CONFIG_DEBUG_LOCK_ALLOC -static DEFINE_PER_CPU(int, wakeup_nest); - -static void ep_poll_safewake(wait_queue_head_t *wq) +static void ep_poll_safewake(struct eventpoll *ep, struct epitem *epi) { + struct eventpoll *ep_src; unsigned long flags; - int subclass; + u8 nests = 0; - local_irq_save(flags); - preempt_disable(); - subclass = __this_cpu_read(wakeup_nest); - spin_lock_nested(&wq->lock, subclass + 1); - __this_cpu_inc(wakeup_nest); - wake_up_locked_poll(wq, POLLIN); - __this_cpu_dec(wakeup_nest); - spin_unlock(&wq->lock); - local_irq_restore(flags); - preempt_enable(); + /* + * If we are not being call from ep_poll_callback(), epi is + * NULL and we are at the first level of nesting, 0. Otherwise, + * we are being called from ep_poll_callback() and if a previous + * wakeup source is not an epoll file itself, we are at depth + * 1 since the wakeup source is depth 0. If the wakeup source + * is a previous epoll file in the wakeup chain then we use its + * nests value and record ours as nests + 1. The previous epoll + * file nests value is stable since its already holding its + * own poll_wait.lock. + */ + if (epi) { + if ((is_file_epoll(epi->ffd.file))) { + ep_src = epi->ffd.file->private_data; + nests = ep_src->nests; + } else { + nests = 1; + } + } + spin_lock_irqsave_nested(&ep->poll_wait.lock, flags, nests); + ep->nests = nests + 1; + wake_up_locked_poll(&ep->poll_wait, EPOLLIN); + ep->nests = 0; + spin_unlock_irqrestore(&ep->poll_wait.lock, flags); } #else -static void ep_poll_safewake(wait_queue_head_t *wq) +static void ep_poll_safewake(struct eventpoll *ep, struct epitem *epi) { - wake_up_poll(wq, EPOLLIN); + wake_up_poll(&ep->poll_wait, EPOLLIN); } #endif @@ -789,7 +808,7 @@ static void ep_free(struct eventpoll *ep /* We need to release all tasks waiting for these file */ if (waitqueue_active(&ep->poll_wait)) - ep_poll_safewake(&ep->poll_wait); + ep_poll_safewake(ep, NULL); /* * We need to lock this because we could be hit by @@ -1258,7 +1277,7 @@ out_unlock: /* We have to call this outside the lock */ if (pwake) - ep_poll_safewake(&ep->poll_wait); + ep_poll_safewake(ep, epi); if (!(epi->event.events & EPOLLEXCLUSIVE)) ewake = 1; @@ -1562,7 +1581,7 @@ static int ep_insert(struct eventpoll *e /* We have to call this outside the lock */ if (pwake) - ep_poll_safewake(&ep->poll_wait); + ep_poll_safewake(ep, NULL); return 0; @@ -1666,7 +1685,7 @@ static int ep_modify(struct eventpoll *e /* We have to call this outside the lock */ if (pwake) - ep_poll_safewake(&ep->poll_wait); + ep_poll_safewake(ep, NULL); return 0; } Index: linux-5.6.17-rt10/fs/ext4/page-io.c =================================================================== --- linux-5.6.17-rt10.orig/fs/ext4/page-io.c +++ linux-5.6.17-rt10/fs/ext4/page-io.c @@ -125,11 +125,10 @@ static void ext4_finish_bio(struct bio * } bh = head = page_buffers(page); /* - * We check all buffers in the page under BH_Uptodate_Lock + * We check all buffers in the page under b_uptodate_lock * to avoid races with other end io clearing async_write flags */ - local_irq_save(flags); - bit_spin_lock(BH_Uptodate_Lock, &head->b_state); + spin_lock_irqsave(&head->b_uptodate_lock, flags); do { if (bh_offset(bh) < bio_start || bh_offset(bh) + bh->b_size > bio_end) { @@ -141,8 +140,7 @@ static void ext4_finish_bio(struct bio * if (bio->bi_status) buffer_io_error(bh); } while ((bh = bh->b_this_page) != head); - bit_spin_unlock(BH_Uptodate_Lock, &head->b_state); - local_irq_restore(flags); + spin_unlock_irqrestore(&head->b_uptodate_lock, flags); if (!under_io) { fscrypt_free_bounce_page(bounce_page); end_page_writeback(page); Index: linux-5.6.17-rt10/fs/fs_struct.c =================================================================== --- linux-5.6.17-rt10.orig/fs/fs_struct.c +++ linux-5.6.17-rt10/fs/fs_struct.c @@ -117,7 +117,7 @@ struct fs_struct *copy_fs_struct(struct fs->users = 1; fs->in_exec = 0; spin_lock_init(&fs->lock); - seqcount_init(&fs->seq); + seqcount_spinlock_init(&fs->seq, &fs->lock); fs->umask = old->umask; spin_lock(&old->lock); @@ -163,6 +163,6 @@ EXPORT_SYMBOL(current_umask); struct fs_struct init_fs = { .users = 1, .lock = __SPIN_LOCK_UNLOCKED(init_fs.lock), - .seq = SEQCNT_ZERO(init_fs.seq), + .seq = SEQCNT_SPINLOCK_ZERO(init_fs.seq, &init_fs.lock), .umask = 0022, }; Index: linux-5.6.17-rt10/fs/fuse/readdir.c =================================================================== --- linux-5.6.17-rt10.orig/fs/fuse/readdir.c +++ linux-5.6.17-rt10/fs/fuse/readdir.c @@ -158,7 +158,7 @@ static int fuse_direntplus_link(struct f struct inode *dir = d_inode(parent); struct fuse_conn *fc; struct inode *inode; - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); if (!o->nodeid) { /* Index: linux-5.6.17-rt10/fs/inode.c =================================================================== --- linux-5.6.17-rt10.orig/fs/inode.c +++ linux-5.6.17-rt10/fs/inode.c @@ -158,7 +158,7 @@ int inode_init_always(struct super_block inode->i_bdev = NULL; inode->i_cdev = NULL; inode->i_link = NULL; - inode->i_dir_seq = 0; + inode->__i_dir_seq = 0; inode->i_rdev = 0; inode->dirtied_when = 0; Index: linux-5.6.17-rt10/fs/io-wq.c =================================================================== --- linux-5.6.17-rt10.orig/fs/io-wq.c +++ linux-5.6.17-rt10/fs/io-wq.c @@ -85,7 +85,7 @@ enum { */ struct io_wqe { struct { - spinlock_t lock; + raw_spinlock_t lock; struct io_wq_work_list work_list; unsigned long hash_map; unsigned flags; @@ -145,7 +145,7 @@ static bool __io_worker_unuse(struct io_ if (current->files != worker->restore_files) { __acquire(&wqe->lock); - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); dropped_lock = true; task_lock(current); @@ -163,7 +163,7 @@ static bool __io_worker_unuse(struct io_ if (worker->mm) { if (!dropped_lock) { __acquire(&wqe->lock); - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); dropped_lock = true; } __set_current_state(TASK_RUNNING); @@ -218,17 +218,17 @@ static void io_worker_exit(struct io_wor worker->flags = 0; preempt_enable(); - spin_lock_irq(&wqe->lock); + raw_spin_lock_irq(&wqe->lock); hlist_nulls_del_rcu(&worker->nulls_node); list_del_rcu(&worker->all_list); if (__io_worker_unuse(wqe, worker)) { __release(&wqe->lock); - spin_lock_irq(&wqe->lock); + raw_spin_lock_irq(&wqe->lock); } acct->nr_workers--; nr_workers = wqe->acct[IO_WQ_ACCT_BOUND].nr_workers + wqe->acct[IO_WQ_ACCT_UNBOUND].nr_workers; - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); /* all workers gone, wq exit can proceed */ if (!nr_workers && refcount_dec_and_test(&wqe->wq->refs)) @@ -463,7 +463,7 @@ static void io_worker_handle_work(struct else if (!wq_list_empty(&wqe->work_list)) wqe->flags |= IO_WQE_FLAG_STALLED; - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); if (put_work && wq->put_work) wq->put_work(old_work); if (!work) @@ -514,14 +514,14 @@ next: worker->cur_work = NULL; spin_unlock_irq(&worker->lock); - spin_lock_irq(&wqe->lock); + raw_spin_lock_irq(&wqe->lock); if (hash != -1U) { wqe->hash_map &= ~BIT_ULL(hash); wqe->flags &= ~IO_WQE_FLAG_STALLED; } if (work && work != old_work) { - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); if (put_work && wq->put_work) { wq->put_work(put_work); @@ -546,7 +546,7 @@ static int io_wqe_worker(void *data) while (!test_bit(IO_WQ_BIT_EXIT, &wq->state)) { set_current_state(TASK_INTERRUPTIBLE); loop: - spin_lock_irq(&wqe->lock); + raw_spin_lock_irq(&wqe->lock); if (io_wqe_run_queue(wqe)) { __set_current_state(TASK_RUNNING); io_worker_handle_work(worker); @@ -557,7 +557,7 @@ loop: __release(&wqe->lock); goto loop; } - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); if (signal_pending(current)) flush_signals(current); if (schedule_timeout(WORKER_IDLE_TIMEOUT)) @@ -569,11 +569,11 @@ loop: } if (test_bit(IO_WQ_BIT_EXIT, &wq->state)) { - spin_lock_irq(&wqe->lock); + raw_spin_lock_irq(&wqe->lock); if (!wq_list_empty(&wqe->work_list)) io_worker_handle_work(worker); else - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); } io_worker_exit(worker); @@ -613,9 +613,9 @@ void io_wq_worker_sleeping(struct task_s worker->flags &= ~IO_WORKER_F_RUNNING; - spin_lock_irq(&wqe->lock); + raw_spin_lock_irq(&wqe->lock); io_wqe_dec_running(wqe, worker); - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); } static bool create_io_worker(struct io_wq *wq, struct io_wqe *wqe, int index) @@ -639,7 +639,7 @@ static bool create_io_worker(struct io_w return false; } - spin_lock_irq(&wqe->lock); + raw_spin_lock_irq(&wqe->lock); hlist_nulls_add_head_rcu(&worker->nulls_node, &wqe->free_list); list_add_tail_rcu(&worker->all_list, &wqe->all_list); worker->flags |= IO_WORKER_F_FREE; @@ -648,7 +648,7 @@ static bool create_io_worker(struct io_w if (!acct->nr_workers && (worker->flags & IO_WORKER_F_BOUND)) worker->flags |= IO_WORKER_F_FIXED; acct->nr_workers++; - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); if (index == IO_WQ_ACCT_UNBOUND) atomic_inc(&wq->user->processes); @@ -700,12 +700,12 @@ static int io_wq_manager(void *data) if (!node_online(node)) continue; - spin_lock_irq(&wqe->lock); + raw_spin_lock_irq(&wqe->lock); if (io_wqe_need_worker(wqe, IO_WQ_ACCT_BOUND)) fork_worker[IO_WQ_ACCT_BOUND] = true; if (io_wqe_need_worker(wqe, IO_WQ_ACCT_UNBOUND)) fork_worker[IO_WQ_ACCT_UNBOUND] = true; - spin_unlock_irq(&wqe->lock); + raw_spin_unlock_irq(&wqe->lock); if (fork_worker[IO_WQ_ACCT_BOUND]) create_io_worker(wq, wqe, IO_WQ_ACCT_BOUND); if (fork_worker[IO_WQ_ACCT_UNBOUND]) @@ -776,10 +776,10 @@ static void io_wqe_enqueue(struct io_wqe } work_flags = work->flags; - spin_lock_irqsave(&wqe->lock, flags); + raw_spin_lock_irqsave(&wqe->lock, flags); wq_list_add_tail(&work->list, &wqe->work_list); wqe->flags &= ~IO_WQE_FLAG_STALLED; - spin_unlock_irqrestore(&wqe->lock, flags); + raw_spin_unlock_irqrestore(&wqe->lock, flags); if ((work_flags & IO_WQ_WORK_CONCURRENT) || !atomic_read(&acct->nr_running)) @@ -897,7 +897,7 @@ static enum io_wq_cancel io_wqe_cancel_c unsigned long flags; bool found = false; - spin_lock_irqsave(&wqe->lock, flags); + raw_spin_lock_irqsave(&wqe->lock, flags); wq_list_for_each(node, prev, &wqe->work_list) { work = container_of(node, struct io_wq_work, list); @@ -907,7 +907,7 @@ static enum io_wq_cancel io_wqe_cancel_c break; } } - spin_unlock_irqrestore(&wqe->lock, flags); + raw_spin_unlock_irqrestore(&wqe->lock, flags); if (found) { io_run_cancel(work); @@ -972,7 +972,7 @@ static enum io_wq_cancel io_wqe_cancel_w * from there. CANCEL_OK means that the work is returned as-new, * no completion will be posted for it. */ - spin_lock_irqsave(&wqe->lock, flags); + raw_spin_lock_irqsave(&wqe->lock, flags); wq_list_for_each(node, prev, &wqe->work_list) { work = container_of(node, struct io_wq_work, list); @@ -982,7 +982,7 @@ static enum io_wq_cancel io_wqe_cancel_w break; } } - spin_unlock_irqrestore(&wqe->lock, flags); + raw_spin_unlock_irqrestore(&wqe->lock, flags); if (found) { io_run_cancel(work); @@ -1097,7 +1097,7 @@ struct io_wq *io_wq_create(unsigned boun } atomic_set(&wqe->acct[IO_WQ_ACCT_UNBOUND].nr_running, 0); wqe->wq = wq; - spin_lock_init(&wqe->lock); + raw_spin_lock_init(&wqe->lock); INIT_WQ_LIST(&wqe->work_list); INIT_HLIST_NULLS_HEAD(&wqe->free_list, 0); INIT_LIST_HEAD(&wqe->all_list); Index: linux-5.6.17-rt10/fs/namei.c =================================================================== --- linux-5.6.17-rt10.orig/fs/namei.c +++ linux-5.6.17-rt10/fs/namei.c @@ -1733,7 +1733,7 @@ static struct dentry *__lookup_slow(cons { struct dentry *dentry, *old; struct inode *inode = dir->d_inode; - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); /* Don't go there if it's already dead */ if (unlikely(IS_DEADDIR(inode))) @@ -3210,7 +3210,7 @@ static int lookup_open(struct nameidata struct dentry *dentry; int error, create_error = 0; umode_t mode = op->mode; - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); if (unlikely(IS_DEADDIR(dir_inode))) return -ENOENT; Index: linux-5.6.17-rt10/fs/namespace.c =================================================================== --- linux-5.6.17-rt10.orig/fs/namespace.c +++ linux-5.6.17-rt10/fs/namespace.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include @@ -321,8 +322,11 @@ int __mnt_want_write(struct vfsmount *m) * incremented count after it has set MNT_WRITE_HOLD. */ smp_mb(); - while (READ_ONCE(mnt->mnt.mnt_flags) & MNT_WRITE_HOLD) - cpu_relax(); + while (READ_ONCE(mnt->mnt.mnt_flags) & MNT_WRITE_HOLD) { + preempt_enable(); + cpu_chill(); + preempt_disable(); + } /* * After the slowpath clears MNT_WRITE_HOLD, mnt_is_readonly will * be set to match its requirements. So we must not load that until Index: linux-5.6.17-rt10/fs/nfs/dir.c =================================================================== --- linux-5.6.17-rt10.orig/fs/nfs/dir.c +++ linux-5.6.17-rt10/fs/nfs/dir.c @@ -461,7 +461,7 @@ void nfs_prime_dcache(struct dentry *par unsigned long dir_verifier) { struct qstr filename = QSTR_INIT(entry->name, entry->len); - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); struct dentry *dentry; struct dentry *alias; struct inode *inode; @@ -1635,7 +1635,7 @@ int nfs_atomic_open(struct inode *dir, s struct file *file, unsigned open_flags, umode_t mode) { - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); struct nfs_open_context *ctx; struct dentry *res; struct iattr attr = { .ia_valid = ATTR_OPEN }; Index: linux-5.6.17-rt10/fs/nfs/nfs4_fs.h =================================================================== --- linux-5.6.17-rt10.orig/fs/nfs/nfs4_fs.h +++ linux-5.6.17-rt10/fs/nfs/nfs4_fs.h @@ -115,7 +115,7 @@ struct nfs4_state_owner { unsigned long so_flags; struct list_head so_states; struct nfs_seqid_counter so_seqid; - seqcount_t so_reclaim_seqcount; + seqcount_spinlock_t so_reclaim_seqcount; struct mutex so_delegreturn_mutex; }; Index: linux-5.6.17-rt10/fs/nfs/nfs4state.c =================================================================== --- linux-5.6.17-rt10.orig/fs/nfs/nfs4state.c +++ linux-5.6.17-rt10/fs/nfs/nfs4state.c @@ -509,7 +509,7 @@ nfs4_alloc_state_owner(struct nfs_server nfs4_init_seqid_counter(&sp->so_seqid); atomic_set(&sp->so_count, 1); INIT_LIST_HEAD(&sp->so_lru); - seqcount_init(&sp->so_reclaim_seqcount); + seqcount_spinlock_init(&sp->so_reclaim_seqcount, &sp->so_lock); mutex_init(&sp->so_delegreturn_mutex); return sp; } Index: linux-5.6.17-rt10/fs/nfs/unlink.c =================================================================== --- linux-5.6.17-rt10.orig/fs/nfs/unlink.c +++ linux-5.6.17-rt10/fs/nfs/unlink.c @@ -13,7 +13,7 @@ #include #include #include -#include +#include #include #include @@ -180,7 +180,7 @@ nfs_async_unlink(struct dentry *dentry, data->cred = get_current_cred(); data->res.dir_attr = &data->dir_attr; - init_waitqueue_head(&data->wq); + init_swait_queue_head(&data->wq); status = -EBUSY; spin_lock(&dentry->d_lock); Index: linux-5.6.17-rt10/fs/ntfs/aops.c =================================================================== --- linux-5.6.17-rt10.orig/fs/ntfs/aops.c +++ linux-5.6.17-rt10/fs/ntfs/aops.c @@ -92,8 +92,7 @@ static void ntfs_end_buffer_async_read(s "0x%llx.", (unsigned long long)bh->b_blocknr); } first = page_buffers(page); - local_irq_save(flags); - bit_spin_lock(BH_Uptodate_Lock, &first->b_state); + spin_lock_irqsave(&first->b_uptodate_lock, flags); clear_buffer_async_read(bh); unlock_buffer(bh); tmp = bh; @@ -108,8 +107,7 @@ static void ntfs_end_buffer_async_read(s } tmp = tmp->b_this_page; } while (tmp != bh); - bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); - local_irq_restore(flags); + spin_unlock_irqrestore(&first->b_uptodate_lock, flags); /* * If none of the buffers had errors then we can set the page uptodate, * but we first have to perform the post read mst fixups, if the @@ -142,8 +140,7 @@ static void ntfs_end_buffer_async_read(s unlock_page(page); return; still_busy: - bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); - local_irq_restore(flags); + spin_unlock_irqrestore(&first->b_uptodate_lock, flags); return; } Index: linux-5.6.17-rt10/fs/proc/base.c =================================================================== --- linux-5.6.17-rt10.orig/fs/proc/base.c +++ linux-5.6.17-rt10/fs/proc/base.c @@ -96,6 +96,7 @@ #include #include #include +#include #include #include "internal.h" #include "fd.h" @@ -1994,7 +1995,7 @@ bool proc_fill_cache(struct file *file, child = d_hash_and_lookup(dir, &qname); if (!child) { - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); child = d_alloc_parallel(dir, &qname, &wq); if (IS_ERR(child)) goto end_instantiate; Index: linux-5.6.17-rt10/fs/proc/kmsg.c =================================================================== --- linux-5.6.17-rt10.orig/fs/proc/kmsg.c +++ linux-5.6.17-rt10/fs/proc/kmsg.c @@ -18,8 +18,6 @@ #include #include -extern wait_queue_head_t log_wait; - static int kmsg_open(struct inode * inode, struct file * file) { return do_syslog(SYSLOG_ACTION_OPEN, NULL, 0, SYSLOG_FROM_PROC); @@ -42,7 +40,7 @@ static ssize_t kmsg_read(struct file *fi static __poll_t kmsg_poll(struct file *file, poll_table *wait) { - poll_wait(file, &log_wait, wait); + poll_wait(file, printk_wait_queue(), wait); if (do_syslog(SYSLOG_ACTION_SIZE_UNREAD, NULL, 0, SYSLOG_FROM_PROC)) return EPOLLIN | EPOLLRDNORM; return 0; Index: linux-5.6.17-rt10/fs/proc/proc_sysctl.c =================================================================== --- linux-5.6.17-rt10.orig/fs/proc/proc_sysctl.c +++ linux-5.6.17-rt10/fs/proc/proc_sysctl.c @@ -702,7 +702,7 @@ static bool proc_sys_fill_cache(struct f child = d_lookup(dir, &qname); if (!child) { - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + DECLARE_SWAIT_QUEUE_HEAD_ONSTACK(wq); child = d_alloc_parallel(dir, &qname, &wq); if (IS_ERR(child)) return false; Index: linux-5.6.17-rt10/fs/squashfs/decompressor_multi_percpu.c =================================================================== --- linux-5.6.17-rt10.orig/fs/squashfs/decompressor_multi_percpu.c +++ linux-5.6.17-rt10/fs/squashfs/decompressor_multi_percpu.c @@ -8,6 +8,7 @@ #include #include #include +#include #include "squashfs_fs.h" #include "squashfs_fs_sb.h" @@ -23,6 +24,8 @@ struct squashfs_stream { void *stream; }; +static DEFINE_LOCAL_IRQ_LOCK(stream_lock); + void *squashfs_decompressor_create(struct squashfs_sb_info *msblk, void *comp_opts) { @@ -77,10 +80,15 @@ int squashfs_decompress(struct squashfs_ { struct squashfs_stream __percpu *percpu = (struct squashfs_stream __percpu *) msblk->stream; - struct squashfs_stream *stream = get_cpu_ptr(percpu); - int res = msblk->decompressor->decompress(msblk, stream->stream, bh, b, - offset, length, output); - put_cpu_ptr(stream); + struct squashfs_stream *stream; + int res; + + stream = get_locked_ptr(stream_lock, percpu); + + res = msblk->decompressor->decompress(msblk, stream->stream, bh, b, + offset, length, output); + + put_locked_ptr(stream_lock, stream); if (res < 0) ERROR("%s decompression failed, data probably corrupt\n", Index: linux-5.6.17-rt10/fs/userfaultfd.c =================================================================== --- linux-5.6.17-rt10.orig/fs/userfaultfd.c +++ linux-5.6.17-rt10/fs/userfaultfd.c @@ -61,7 +61,7 @@ struct userfaultfd_ctx { /* waitqueue head for events */ wait_queue_head_t event_wqh; /* a refile sequence protected by fault_pending_wqh lock */ - struct seqcount refile_seq; + seqcount_spinlock_t refile_seq; /* pseudo fd refcounting */ refcount_t refcount; /* userfaultfd syscall flags */ @@ -1940,7 +1940,7 @@ static void init_once_userfaultfd_ctx(vo init_waitqueue_head(&ctx->fault_wqh); init_waitqueue_head(&ctx->event_wqh); init_waitqueue_head(&ctx->fd_wqh); - seqcount_init(&ctx->refile_seq); + seqcount_spinlock_init(&ctx->refile_seq, &ctx->fault_pending_wqh.lock); } SYSCALL_DEFINE1(userfaultfd, int, flags) Index: linux-5.6.17-rt10/include/linux/bottom_half.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/bottom_half.h +++ linux-5.6.17-rt10/include/linux/bottom_half.h @@ -4,6 +4,10 @@ #include +#ifdef CONFIG_PREEMPT_RT +extern void __local_bh_disable_ip(unsigned long ip, unsigned int cnt); +#else + #ifdef CONFIG_TRACE_IRQFLAGS extern void __local_bh_disable_ip(unsigned long ip, unsigned int cnt); #else @@ -13,6 +17,7 @@ static __always_inline void __local_bh_d barrier(); } #endif +#endif static inline void local_bh_disable(void) { Index: linux-5.6.17-rt10/include/linux/bpf.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/bpf.h +++ linux-5.6.17-rt10/include/linux/bpf.h @@ -886,7 +886,7 @@ int bpf_prog_array_copy(struct bpf_prog_ struct bpf_prog *_prog; \ struct bpf_prog_array *_array; \ u32 _ret = 1; \ - preempt_disable(); \ + migrate_disable(); \ rcu_read_lock(); \ _array = rcu_dereference(array); \ if (unlikely(check_non_null && !_array))\ @@ -899,7 +899,7 @@ int bpf_prog_array_copy(struct bpf_prog_ } \ _out: \ rcu_read_unlock(); \ - preempt_enable(); \ + migrate_enable(); \ _ret; \ }) @@ -933,7 +933,7 @@ _out: \ u32 ret; \ u32 _ret = 1; \ u32 _cn = 0; \ - preempt_disable(); \ + migrate_disable(); \ rcu_read_lock(); \ _array = rcu_dereference(array); \ _item = &_array->items[0]; \ @@ -945,7 +945,7 @@ _out: \ _item++; \ } \ rcu_read_unlock(); \ - preempt_enable(); \ + migrate_enable(); \ if (_ret) \ _ret = (_cn ? NET_XMIT_CN : NET_XMIT_SUCCESS); \ else \ @@ -962,6 +962,36 @@ _out: \ #ifdef CONFIG_BPF_SYSCALL DECLARE_PER_CPU(int, bpf_prog_active); +/* + * Block execution of BPF programs attached to instrumentation (perf, + * kprobes, tracepoints) to prevent deadlocks on map operations as any of + * these events can happen inside a region which holds a map bucket lock + * and can deadlock on it. + * + * Use the preemption safe inc/dec variants on RT because migrate disable + * is preemptible on RT and preemption in the middle of the RMW operation + * might lead to inconsistent state. Use the raw variants for non RT + * kernels as migrate_disable() maps to preempt_disable() so the slightly + * more expensive save operation can be avoided. + */ +static inline void bpf_disable_instrumentation(void) +{ + migrate_disable(); + if (IS_ENABLED(CONFIG_PREEMPT_RT)) + this_cpu_inc(bpf_prog_active); + else + __this_cpu_inc(bpf_prog_active); +} + +static inline void bpf_enable_instrumentation(void) +{ + if (IS_ENABLED(CONFIG_PREEMPT_RT)) + this_cpu_dec(bpf_prog_active); + else + __this_cpu_dec(bpf_prog_active); + migrate_enable(); +} + extern const struct file_operations bpf_map_fops; extern const struct file_operations bpf_prog_fops; Index: linux-5.6.17-rt10/include/linux/buffer_head.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/buffer_head.h +++ linux-5.6.17-rt10/include/linux/buffer_head.h @@ -22,9 +22,6 @@ enum bh_state_bits { BH_Dirty, /* Is dirty */ BH_Lock, /* Is locked */ BH_Req, /* Has been submitted for I/O */ - BH_Uptodate_Lock,/* Used by the first bh in a page, to serialise - * IO completion of other buffers in the page - */ BH_Mapped, /* Has a disk mapping */ BH_New, /* Disk mapping was newly created by get_block */ @@ -76,6 +73,9 @@ struct buffer_head { struct address_space *b_assoc_map; /* mapping this buffer is associated with */ atomic_t b_count; /* users using this buffer_head */ + spinlock_t b_uptodate_lock; /* Used by the first bh in a page, to + * serialise IO completion of other + * buffers in the page */ }; /* Index: linux-5.6.17-rt10/include/linux/completion.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/completion.h +++ linux-5.6.17-rt10/include/linux/completion.h @@ -9,7 +9,7 @@ * See kernel/sched/completion.c for details. */ -#include +#include /* * struct completion - structure used to maintain state for a "completion" @@ -25,7 +25,7 @@ */ struct completion { unsigned int done; - wait_queue_head_t wait; + struct swait_queue_head wait; }; #define init_completion_map(x, m) __init_completion(x) @@ -34,7 +34,7 @@ static inline void complete_acquire(stru static inline void complete_release(struct completion *x) {} #define COMPLETION_INITIALIZER(work) \ - { 0, __WAIT_QUEUE_HEAD_INITIALIZER((work).wait) } + { 0, __SWAIT_QUEUE_HEAD_INITIALIZER((work).wait) } #define COMPLETION_INITIALIZER_ONSTACK_MAP(work, map) \ (*({ init_completion_map(&(work), &(map)); &(work); })) @@ -85,7 +85,7 @@ static inline void complete_release(stru static inline void __init_completion(struct completion *x) { x->done = 0; - init_waitqueue_head(&x->wait); + init_swait_queue_head(&x->wait); } /** Index: linux-5.6.17-rt10/include/linux/console.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/console.h +++ linux-5.6.17-rt10/include/linux/console.h @@ -144,6 +144,7 @@ static inline int con_debug_leave(void) struct console { char name[16]; void (*write)(struct console *, const char *, unsigned); + void (*write_atomic)(struct console *, const char *, unsigned); int (*read)(struct console *, char *, unsigned); struct tty_driver *(*device)(struct console *, int *); void (*unblank)(void); @@ -152,6 +153,8 @@ struct console { short flags; short index; int cflag; + unsigned long printk_seq; + int wrote_history; void *data; struct console *next; }; @@ -232,4 +235,7 @@ extern void console_init(void); void dummycon_register_output_notifier(struct notifier_block *nb); void dummycon_unregister_output_notifier(struct notifier_block *nb); +extern void console_atomic_lock(unsigned int *flags); +extern void console_atomic_unlock(unsigned int flags); + #endif /* _LINUX_CONSOLE_H */ Index: linux-5.6.17-rt10/include/linux/dcache.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/dcache.h +++ linux-5.6.17-rt10/include/linux/dcache.h @@ -89,7 +89,7 @@ extern struct dentry_stat_t dentry_stat; struct dentry { /* RCU lookup touched fields */ unsigned int d_flags; /* protected by d_lock */ - seqcount_t d_seq; /* per dentry seqlock */ + seqcount_spinlock_t d_seq; /* per dentry seqlock */ struct hlist_bl_node d_hash; /* lookup hash list */ struct dentry *d_parent; /* parent directory */ struct qstr d_name; @@ -106,7 +106,7 @@ struct dentry { union { struct list_head d_lru; /* LRU list */ - wait_queue_head_t *d_wait; /* in-lookup ones only */ + struct swait_queue_head *d_wait; /* in-lookup ones only */ }; struct list_head d_child; /* child of parent list */ struct list_head d_subdirs; /* our children */ @@ -236,7 +236,7 @@ extern void d_set_d_op(struct dentry *de extern struct dentry * d_alloc(struct dentry *, const struct qstr *); extern struct dentry * d_alloc_anon(struct super_block *); extern struct dentry * d_alloc_parallel(struct dentry *, const struct qstr *, - wait_queue_head_t *); + struct swait_queue_head *); extern struct dentry * d_splice_alias(struct inode *, struct dentry *); extern struct dentry * d_add_ci(struct dentry *, struct inode *, struct qstr *); extern struct dentry * d_exact_alias(struct dentry *, struct inode *); Index: linux-5.6.17-rt10/include/linux/delay.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/delay.h +++ linux-5.6.17-rt10/include/linux/delay.h @@ -65,4 +65,10 @@ static inline void ssleep(unsigned int s msleep(seconds * 1000); } +#ifdef CONFIG_PREEMPT_RT +extern void cpu_chill(void); +#else +# define cpu_chill() cpu_relax() +#endif + #endif /* defined(_LINUX_DELAY_H) */ Index: linux-5.6.17-rt10/include/linux/dma-resv.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/dma-resv.h +++ linux-5.6.17-rt10/include/linux/dma-resv.h @@ -46,8 +46,6 @@ #include extern struct ww_class reservation_ww_class; -extern struct lock_class_key reservation_seqcount_class; -extern const char reservation_seqcount_string[]; /** * struct dma_resv_list - a list of shared fences @@ -71,7 +69,7 @@ struct dma_resv_list { */ struct dma_resv { struct ww_mutex lock; - seqcount_t seq; + seqcount_ww_mutex_t seq; struct dma_fence __rcu *fence_excl; struct dma_resv_list __rcu *fence; Index: linux-5.6.17-rt10/include/linux/filter.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/filter.h +++ linux-5.6.17-rt10/include/linux/filter.h @@ -561,7 +561,7 @@ DECLARE_STATIC_KEY_FALSE(bpf_stats_enabl #define __BPF_PROG_RUN(prog, ctx, dfunc) ({ \ u32 ret; \ - cant_sleep(); \ + cant_migrate(); \ if (static_branch_unlikely(&bpf_stats_enabled_key)) { \ struct bpf_prog_stats *stats; \ u64 start = sched_clock(); \ @@ -576,8 +576,30 @@ DECLARE_STATIC_KEY_FALSE(bpf_stats_enabl } \ ret; }) -#define BPF_PROG_RUN(prog, ctx) __BPF_PROG_RUN(prog, ctx, \ - bpf_dispatcher_nopfunc) +#define BPF_PROG_RUN(prog, ctx) \ + __BPF_PROG_RUN(prog, ctx, bpf_dispatcher_nopfunc) + +/* + * Use in preemptible and therefore migratable context to make sure that + * the execution of the BPF program runs on one CPU. + * + * This uses migrate_disable/enable() explicitly to document that the + * invocation of a BPF program does not require reentrancy protection + * against a BPF program which is invoked from a preempting task. + * + * For non RT enabled kernels migrate_disable/enable() maps to + * preempt_disable/enable(), i.e. it disables also preemption. + */ +static inline u32 bpf_prog_run_pin_on_cpu(const struct bpf_prog *prog, + const void *ctx) +{ + u32 ret; + + migrate_disable(); + ret = __BPF_PROG_RUN(prog, ctx, bpf_dispatcher_nopfunc); + migrate_enable(); + return ret; +} #define BPF_SKB_CB_LEN QDISC_CB_PRIV_LEN @@ -655,6 +677,7 @@ static inline u8 *bpf_skb_cb(struct sk_b return qdisc_skb_cb(skb)->data; } +/* Must be invoked with migration disabled */ static inline u32 __bpf_prog_run_save_cb(const struct bpf_prog *prog, struct sk_buff *skb) { @@ -680,9 +703,9 @@ static inline u32 bpf_prog_run_save_cb(c { u32 res; - preempt_disable(); + migrate_disable(); res = __bpf_prog_run_save_cb(prog, skb); - preempt_enable(); + migrate_enable(); return res; } @@ -695,9 +718,7 @@ static inline u32 bpf_prog_run_clear_cb( if (unlikely(prog->cb_access)) memset(cb_data, 0, BPF_SKB_CB_LEN); - preempt_disable(); - res = BPF_PROG_RUN(prog, skb); - preempt_enable(); + res = bpf_prog_run_pin_on_cpu(prog, skb); return res; } Index: linux-5.6.17-rt10/include/linux/fs.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/fs.h +++ linux-5.6.17-rt10/include/linux/fs.h @@ -717,7 +717,7 @@ struct inode { struct block_device *i_bdev; struct cdev *i_cdev; char *i_link; - unsigned i_dir_seq; + unsigned __i_dir_seq; }; __u32 i_generation; Index: linux-5.6.17-rt10/include/linux/fs_struct.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/fs_struct.h +++ linux-5.6.17-rt10/include/linux/fs_struct.h @@ -9,7 +9,7 @@ struct fs_struct { int users; spinlock_t lock; - seqcount_t seq; + seqcount_spinlock_t seq; int umask; int in_exec; struct path root, pwd; Index: linux-5.6.17-rt10/include/linux/genhd.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/genhd.h +++ linux-5.6.17-rt10/include/linux/genhd.h @@ -752,9 +752,11 @@ static inline sector_t part_nr_sects_rea static inline void part_nr_sects_write(struct hd_struct *part, sector_t size) { #if BITS_PER_LONG==32 && defined(CONFIG_SMP) + preempt_disable(); write_seqcount_begin(&part->nr_sects_seq); part->nr_sects = size; write_seqcount_end(&part->nr_sects_seq); + preempt_enable(); #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPTION) preempt_disable(); part->nr_sects = size; Index: linux-5.6.17-rt10/include/linux/hardirq.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/hardirq.h +++ linux-5.6.17-rt10/include/linux/hardirq.h @@ -68,7 +68,6 @@ extern void irq_exit(void); #define nmi_enter() \ do { \ arch_nmi_enter(); \ - printk_nmi_enter(); \ lockdep_off(); \ ftrace_nmi_enter(); \ BUG_ON(in_nmi()); \ @@ -85,7 +84,6 @@ extern void irq_exit(void); preempt_count_sub(NMI_OFFSET + HARDIRQ_OFFSET); \ ftrace_nmi_exit(); \ lockdep_on(); \ - printk_nmi_exit(); \ arch_nmi_exit(); \ } while (0) Index: linux-5.6.17-rt10/include/linux/highmem.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/highmem.h +++ linux-5.6.17-rt10/include/linux/highmem.h @@ -8,6 +8,7 @@ #include #include #include +#include #include @@ -90,7 +91,7 @@ static inline void kunmap(struct page *p static inline void *kmap_atomic(struct page *page) { - preempt_disable(); + preempt_disable_nort(); pagefault_disable(); return page_address(page); } @@ -99,7 +100,7 @@ static inline void *kmap_atomic(struct p static inline void __kunmap_atomic(void *addr) { pagefault_enable(); - preempt_enable(); + preempt_enable_nort(); } #define kmap_atomic_pfn(pfn) kmap_atomic(pfn_to_page(pfn)) @@ -111,32 +112,51 @@ static inline void __kunmap_atomic(void #if defined(CONFIG_HIGHMEM) || defined(CONFIG_X86_32) +#ifndef CONFIG_PREEMPT_RT DECLARE_PER_CPU(int, __kmap_atomic_idx); +#endif static inline int kmap_atomic_idx_push(void) { +#ifndef CONFIG_PREEMPT_RT int idx = __this_cpu_inc_return(__kmap_atomic_idx) - 1; -#ifdef CONFIG_DEBUG_HIGHMEM +# ifdef CONFIG_DEBUG_HIGHMEM WARN_ON_ONCE(in_irq() && !irqs_disabled()); BUG_ON(idx >= KM_TYPE_NR); -#endif +# endif return idx; +#else + current->kmap_idx++; + BUG_ON(current->kmap_idx > KM_TYPE_NR); + return current->kmap_idx - 1; +#endif } static inline int kmap_atomic_idx(void) { +#ifndef CONFIG_PREEMPT_RT return __this_cpu_read(__kmap_atomic_idx) - 1; +#else + return current->kmap_idx - 1; +#endif } static inline void kmap_atomic_idx_pop(void) { -#ifdef CONFIG_DEBUG_HIGHMEM +#ifndef CONFIG_PREEMPT_RT +# ifdef CONFIG_DEBUG_HIGHMEM int idx = __this_cpu_dec_return(__kmap_atomic_idx); BUG_ON(idx < 0); -#else +# else __this_cpu_dec(__kmap_atomic_idx); +# endif +#else + current->kmap_idx--; +# ifdef CONFIG_DEBUG_HIGHMEM + BUG_ON(current->kmap_idx < 0); +# endif #endif } Index: linux-5.6.17-rt10/include/linux/hrtimer.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/hrtimer.h +++ linux-5.6.17-rt10/include/linux/hrtimer.h @@ -159,7 +159,7 @@ struct hrtimer_clock_base { struct hrtimer_cpu_base *cpu_base; unsigned int index; clockid_t clockid; - seqcount_t seq; + seqcount_raw_spinlock_t seq; struct hrtimer *running; struct timerqueue_head active; ktime_t (*get_time)(void); Index: linux-5.6.17-rt10/include/linux/idr.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/idr.h +++ linux-5.6.17-rt10/include/linux/idr.h @@ -169,10 +169,7 @@ static inline bool idr_is_empty(const st * Each idr_preload() should be matched with an invocation of this * function. See idr_preload() for details. */ -static inline void idr_preload_end(void) -{ - preempt_enable(); -} +void idr_preload_end(void); /** * idr_for_each_entry() - Iterate over an IDR's elements of a given type. Index: linux-5.6.17-rt10/include/linux/interrupt.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/interrupt.h +++ linux-5.6.17-rt10/include/linux/interrupt.h @@ -558,7 +558,7 @@ struct softirq_action asmlinkage void do_softirq(void); asmlinkage void __do_softirq(void); -#ifdef __ARCH_HAS_DO_SOFTIRQ +#if defined(__ARCH_HAS_DO_SOFTIRQ) && !defined(CONFIG_PREEMPT_RT) void do_softirq_own_stack(void); #else static inline void do_softirq_own_stack(void) @@ -573,6 +573,7 @@ extern void __raise_softirq_irqoff(unsig extern void raise_softirq_irqoff(unsigned int nr); extern void raise_softirq(unsigned int nr); +extern void softirq_check_pending_idle(void); DECLARE_PER_CPU(struct task_struct *, ksoftirqd); @@ -637,7 +638,10 @@ static inline void tasklet_unlock(struct static inline void tasklet_unlock_wait(struct tasklet_struct *t) { - while (test_bit(TASKLET_STATE_RUN, &(t)->state)) { barrier(); } + while (test_bit(TASKLET_STATE_RUN, &(t)->state)) { + local_bh_disable(); + local_bh_enable(); + } } #else #define tasklet_trylock(t) 1 Index: linux-5.6.17-rt10/include/linux/irq_work.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/irq_work.h +++ linux-5.6.17-rt10/include/linux/irq_work.h @@ -18,6 +18,8 @@ /* Doesn't want IPI, wait for tick: */ #define IRQ_WORK_LAZY BIT(2) +/* Run hard IRQ context, even on RT */ +#define IRQ_WORK_HARD_IRQ BIT(3) #define IRQ_WORK_CLAIMED (IRQ_WORK_PENDING | IRQ_WORK_BUSY) @@ -56,4 +58,10 @@ static inline bool irq_work_needs_cpu(vo static inline void irq_work_run(void) { } #endif +#if defined(CONFIG_IRQ_WORK) && defined(CONFIG_PREEMPT_RT) +void irq_work_tick_soft(void); +#else +static inline void irq_work_tick_soft(void) { } +#endif + #endif /* _LINUX_IRQ_WORK_H */ Index: linux-5.6.17-rt10/include/linux/irqdesc.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/irqdesc.h +++ linux-5.6.17-rt10/include/linux/irqdesc.h @@ -72,6 +72,7 @@ struct irq_desc { unsigned int irqs_unhandled; atomic_t threads_handled; int threads_handled_last; + u64 random_ip; raw_spinlock_t lock; struct cpumask *percpu_enabled; const struct cpumask *percpu_affinity; Index: linux-5.6.17-rt10/include/linux/irqflags.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/irqflags.h +++ linux-5.6.17-rt10/include/linux/irqflags.h @@ -43,14 +43,6 @@ do { \ do { \ current->hardirq_context--; \ } while (0) -# define lockdep_softirq_enter() \ -do { \ - current->softirq_context++; \ -} while (0) -# define lockdep_softirq_exit() \ -do { \ - current->softirq_context--; \ -} while (0) #else # define trace_hardirqs_on() do { } while (0) # define trace_hardirqs_off() do { } while (0) @@ -63,6 +55,21 @@ do { \ # define lockdep_softirq_enter() do { } while (0) # define lockdep_softirq_exit() do { } while (0) #endif + +#if defined(CONFIG_TRACE_IRQFLAGS) && !defined(CONFIG_PREEMPT_RT) +# define lockdep_softirq_enter() \ +do { \ + current->softirq_context++; \ +} while (0) +# define lockdep_softirq_exit() \ +do { \ + current->softirq_context--; \ +} while (0) + +#else +# define lockdep_softirq_enter() do { } while (0) +# define lockdep_softirq_exit() do { } while (0) +#endif #if defined(CONFIG_IRQSOFF_TRACER) || \ defined(CONFIG_PREEMPT_TRACER) Index: linux-5.6.17-rt10/include/linux/kernel.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/kernel.h +++ linux-5.6.17-rt10/include/linux/kernel.h @@ -218,6 +218,10 @@ extern void __cant_sleep(const char *fil */ # define might_sleep() \ do { __might_sleep(__FILE__, __LINE__, 0); might_resched(); } while (0) + +# define might_sleep_no_state_check() \ + do { ___might_sleep(__FILE__, __LINE__, 0); might_resched(); } while (0) + /** * cant_sleep - annotation for functions that cannot sleep * @@ -249,6 +253,7 @@ extern void __cant_sleep(const char *fil static inline void __might_sleep(const char *file, int line, int preempt_offset) { } # define might_sleep() do { might_resched(); } while (0) +# define might_sleep_no_state_check() do { might_resched(); } while (0) # define cant_sleep() do { } while (0) # define sched_annotate_sleep() do { } while (0) # define non_block_start() do { } while (0) @@ -257,6 +262,13 @@ extern void __cant_sleep(const char *fil #define might_sleep_if(cond) do { if (cond) might_sleep(); } while (0) +#ifndef CONFIG_PREEMPT_RT +# define cant_migrate() cant_sleep() +#else + /* Placeholder for now */ +# define cant_migrate() do { } while (0) +#endif + /** * abs - return absolute value of an argument * @x: the value. If it is unsigned type, it is converted to signed type first. Index: linux-5.6.17-rt10/include/linux/kmsg_dump.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/kmsg_dump.h +++ linux-5.6.17-rt10/include/linux/kmsg_dump.h @@ -46,10 +46,8 @@ struct kmsg_dumper { bool registered; /* private state of the kmsg iterator */ - u32 cur_idx; - u32 next_idx; - u64 cur_seq; - u64 next_seq; + u64 line_seq; + u64 buffer_end_seq; }; #ifdef CONFIG_PRINTK Index: linux-5.6.17-rt10/include/linux/kvm_irqfd.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/kvm_irqfd.h +++ linux-5.6.17-rt10/include/linux/kvm_irqfd.h @@ -42,7 +42,7 @@ struct kvm_kernel_irqfd { wait_queue_entry_t wait; /* Update side is protected by irqfds.lock */ struct kvm_kernel_irq_routing_entry irq_entry; - seqcount_t irq_entry_sc; + seqcount_spinlock_t irq_entry_sc; /* Used for level IRQ fast-path */ int gsi; struct work_struct inject; Index: linux-5.6.17-rt10/include/linux/locallock.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/locallock.h @@ -0,0 +1,257 @@ +#ifndef _LINUX_LOCALLOCK_H +#define _LINUX_LOCALLOCK_H + +#include +#include +#include + +#ifdef CONFIG_PREEMPT_RT + +#ifdef CONFIG_DEBUG_SPINLOCK +# define LL_WARN(cond) WARN_ON(cond) +#else +# define LL_WARN(cond) do { } while (0) +#endif + +/* + * per cpu lock based substitute for local_irq_*() + */ +struct local_irq_lock { + spinlock_t lock; + struct task_struct *owner; + int nestcnt; + unsigned long flags; +}; + +#define DEFINE_LOCAL_IRQ_LOCK(lvar) \ + DEFINE_PER_CPU(struct local_irq_lock, lvar) = { \ + .lock = __SPIN_LOCK_UNLOCKED((lvar).lock) } + +#define DECLARE_LOCAL_IRQ_LOCK(lvar) \ + DECLARE_PER_CPU(struct local_irq_lock, lvar) + +#define local_irq_lock_init(lvar) \ + do { \ + int __cpu; \ + for_each_possible_cpu(__cpu) \ + spin_lock_init(&per_cpu(lvar, __cpu).lock); \ + } while (0) + +static inline void __local_lock(struct local_irq_lock *lv) +{ + if (lv->owner != current) { + spin_lock(&lv->lock); + LL_WARN(lv->owner); + LL_WARN(lv->nestcnt); + lv->owner = current; + } + lv->nestcnt++; +} + +#define local_lock(lvar) \ + do { __local_lock(&get_local_var(lvar)); } while (0) + +static inline int __local_trylock(struct local_irq_lock *lv) +{ + if (lv->owner != current && spin_trylock(&lv->lock)) { + LL_WARN(lv->owner); + LL_WARN(lv->nestcnt); + lv->owner = current; + lv->nestcnt = 1; + return 1; + } else if (lv->owner == current) { + lv->nestcnt++; + return 1; + } + return 0; +} + +#define local_trylock(lvar) \ + ({ \ + int __locked; \ + __locked = __local_trylock(&get_local_var(lvar)); \ + if (!__locked) \ + put_local_var(lvar); \ + __locked; \ + }) + +static inline void __local_unlock(struct local_irq_lock *lv) +{ + LL_WARN(lv->nestcnt == 0); + LL_WARN(lv->owner != current); + if (--lv->nestcnt) + return; + + lv->owner = NULL; + spin_unlock(&lv->lock); +} + +#define local_unlock(lvar) \ + do { \ + __local_unlock(this_cpu_ptr(&lvar)); \ + put_local_var(lvar); \ + } while (0) + +static inline void __local_lock_irq(struct local_irq_lock *lv) +{ + spin_lock_irqsave(&lv->lock, lv->flags); + LL_WARN(lv->owner); + LL_WARN(lv->nestcnt); + lv->owner = current; + lv->nestcnt = 1; +} + +#define local_lock_irq(lvar) \ + do { __local_lock_irq(&get_local_var(lvar)); } while (0) + +static inline void __local_unlock_irq(struct local_irq_lock *lv) +{ + LL_WARN(!lv->nestcnt); + LL_WARN(lv->owner != current); + lv->owner = NULL; + lv->nestcnt = 0; + spin_unlock_irq(&lv->lock); +} + +#define local_unlock_irq(lvar) \ + do { \ + __local_unlock_irq(this_cpu_ptr(&lvar)); \ + put_local_var(lvar); \ + } while (0) + +static inline int __local_lock_irqsave(struct local_irq_lock *lv) +{ + if (lv->owner != current) { + __local_lock_irq(lv); + return 0; + } else { + lv->nestcnt++; + return 1; + } +} + +#define local_lock_irqsave(lvar, _flags) \ + do { \ + if (__local_lock_irqsave(&get_local_var(lvar))) \ + put_local_var(lvar); \ + _flags = __this_cpu_read(lvar.flags); \ + } while (0) + +static inline int __local_unlock_irqrestore(struct local_irq_lock *lv, + unsigned long flags) +{ + LL_WARN(!lv->nestcnt); + LL_WARN(lv->owner != current); + if (--lv->nestcnt) + return 0; + + lv->owner = NULL; + spin_unlock_irqrestore(&lv->lock, lv->flags); + return 1; +} + +#define local_unlock_irqrestore(lvar, flags) \ + do { \ + if (__local_unlock_irqrestore(this_cpu_ptr(&lvar), flags)) \ + put_local_var(lvar); \ + } while (0) + +#define local_spin_trylock_irq(lvar, lock) \ + ({ \ + int __locked; \ + local_lock_irq(lvar); \ + __locked = spin_trylock(lock); \ + if (!__locked) \ + local_unlock_irq(lvar); \ + __locked; \ + }) + +#define local_spin_lock_irq(lvar, lock) \ + do { \ + local_lock_irq(lvar); \ + spin_lock(lock); \ + } while (0) + +#define local_spin_unlock_irq(lvar, lock) \ + do { \ + spin_unlock(lock); \ + local_unlock_irq(lvar); \ + } while (0) + +#define local_spin_lock_irqsave(lvar, lock, flags) \ + do { \ + local_lock_irqsave(lvar, flags); \ + spin_lock(lock); \ + } while (0) + +#define local_spin_unlock_irqrestore(lvar, lock, flags) \ + do { \ + spin_unlock(lock); \ + local_unlock_irqrestore(lvar, flags); \ + } while (0) + +#define get_locked_var(lvar, var) \ + (*({ \ + local_lock(lvar); \ + this_cpu_ptr(&var); \ + })) + +#define put_locked_var(lvar, var) local_unlock(lvar); + +#define get_locked_ptr(lvar, var) \ + ({ \ + local_lock(lvar); \ + this_cpu_ptr(var); \ + }) + +#define put_locked_ptr(lvar, var) local_unlock(lvar); + +#define local_lock_cpu(lvar) \ + ({ \ + local_lock(lvar); \ + smp_processor_id(); \ + }) + +#define local_unlock_cpu(lvar) local_unlock(lvar) + +#else /* PREEMPT_RT */ + +#define DEFINE_LOCAL_IRQ_LOCK(lvar) __typeof__(const int) lvar +#define DECLARE_LOCAL_IRQ_LOCK(lvar) extern __typeof__(const int) lvar + +static inline void local_irq_lock_init(int lvar) { } + +#define local_trylock(lvar) \ + ({ \ + preempt_disable(); \ + 1; \ + }) + +#define local_lock(lvar) preempt_disable() +#define local_unlock(lvar) preempt_enable() +#define local_lock_irq(lvar) local_irq_disable() +#define local_lock_irq_on(lvar, cpu) local_irq_disable() +#define local_unlock_irq(lvar) local_irq_enable() +#define local_unlock_irq_on(lvar, cpu) local_irq_enable() +#define local_lock_irqsave(lvar, flags) local_irq_save(flags) +#define local_unlock_irqrestore(lvar, flags) local_irq_restore(flags) + +#define local_spin_trylock_irq(lvar, lock) spin_trylock_irq(lock) +#define local_spin_lock_irq(lvar, lock) spin_lock_irq(lock) +#define local_spin_unlock_irq(lvar, lock) spin_unlock_irq(lock) +#define local_spin_lock_irqsave(lvar, lock, flags) \ + spin_lock_irqsave(lock, flags) +#define local_spin_unlock_irqrestore(lvar, lock, flags) \ + spin_unlock_irqrestore(lock, flags) + +#define get_locked_var(lvar, var) get_cpu_var(var) +#define put_locked_var(lvar, var) put_cpu_var(var) +#define get_locked_ptr(lvar, var) get_cpu_ptr(var) +#define put_locked_ptr(lvar, var) put_cpu_ptr(var) + +#define local_lock_cpu(lvar) get_cpu() +#define local_unlock_cpu(lvar) put_cpu() + +#endif + +#endif Index: linux-5.6.17-rt10/include/linux/mm_types.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/mm_types.h +++ linux-5.6.17-rt10/include/linux/mm_types.h @@ -12,6 +12,7 @@ #include #include #include +#include #include #include @@ -525,6 +526,9 @@ struct mm_struct { bool tlb_flush_batched; #endif struct uprobes_state uprobes_state; +#ifdef CONFIG_PREEMPT_RT + struct rcu_head delayed_drop; +#endif #ifdef CONFIG_HUGETLB_PAGE atomic_long_t hugetlb_usage; #endif Index: linux-5.6.17-rt10/include/linux/mutex.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/mutex.h +++ linux-5.6.17-rt10/include/linux/mutex.h @@ -22,6 +22,17 @@ struct ww_acquire_ctx; +#ifdef CONFIG_DEBUG_LOCK_ALLOC +# define __DEP_MAP_MUTEX_INITIALIZER(lockname) \ + , .dep_map = { .name = #lockname } +#else +# define __DEP_MAP_MUTEX_INITIALIZER(lockname) +#endif + +#ifdef CONFIG_PREEMPT_RT +# include +#else + /* * Simple, straightforward mutexes with strict semantics: * @@ -108,13 +119,6 @@ do { \ __mutex_init((mutex), #mutex, &__key); \ } while (0) -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define __DEP_MAP_MUTEX_INITIALIZER(lockname) \ - , .dep_map = { .name = #lockname } -#else -# define __DEP_MAP_MUTEX_INITIALIZER(lockname) -#endif - #define __MUTEX_INITIALIZER(lockname) \ { .owner = ATOMIC_LONG_INIT(0) \ , .wait_lock = __SPIN_LOCK_UNLOCKED(lockname.wait_lock) \ @@ -210,4 +214,6 @@ enum mutex_trylock_recursive_enum { extern /* __deprecated */ __must_check enum mutex_trylock_recursive_enum mutex_trylock_recursive(struct mutex *lock); +#endif /* !PREEMPT_RT */ + #endif /* __LINUX_MUTEX_H */ Index: linux-5.6.17-rt10/include/linux/mutex_rt.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/mutex_rt.h @@ -0,0 +1,130 @@ +#ifndef __LINUX_MUTEX_RT_H +#define __LINUX_MUTEX_RT_H + +#ifndef __LINUX_MUTEX_H +#error "Please include mutex.h" +#endif + +#include + +/* FIXME: Just for __lockfunc */ +#include + +struct mutex { + struct rt_mutex lock; +#ifdef CONFIG_DEBUG_LOCK_ALLOC + struct lockdep_map dep_map; +#endif +}; + +#define __MUTEX_INITIALIZER(mutexname) \ + { \ + .lock = __RT_MUTEX_INITIALIZER(mutexname.lock) \ + __DEP_MAP_MUTEX_INITIALIZER(mutexname) \ + } + +#define DEFINE_MUTEX(mutexname) \ + struct mutex mutexname = __MUTEX_INITIALIZER(mutexname) + +extern void __mutex_do_init(struct mutex *lock, const char *name, struct lock_class_key *key); +extern void __lockfunc _mutex_lock(struct mutex *lock); +extern void __lockfunc _mutex_lock_io(struct mutex *lock); +extern void __lockfunc _mutex_lock_io_nested(struct mutex *lock, int subclass); +extern int __lockfunc _mutex_lock_interruptible(struct mutex *lock); +extern int __lockfunc _mutex_lock_killable(struct mutex *lock); +extern void __lockfunc _mutex_lock_nested(struct mutex *lock, int subclass); +extern void __lockfunc _mutex_lock_nest_lock(struct mutex *lock, struct lockdep_map *nest_lock); +extern int __lockfunc _mutex_lock_interruptible_nested(struct mutex *lock, int subclass); +extern int __lockfunc _mutex_lock_killable_nested(struct mutex *lock, int subclass); +extern int __lockfunc _mutex_trylock(struct mutex *lock); +extern void __lockfunc _mutex_unlock(struct mutex *lock); + +#define mutex_is_locked(l) rt_mutex_is_locked(&(l)->lock) +#define mutex_lock(l) _mutex_lock(l) +#define mutex_lock_interruptible(l) _mutex_lock_interruptible(l) +#define mutex_lock_killable(l) _mutex_lock_killable(l) +#define mutex_trylock(l) _mutex_trylock(l) +#define mutex_unlock(l) _mutex_unlock(l) +#define mutex_lock_io(l) _mutex_lock_io(l); + +#define __mutex_owner(l) ((l)->lock.owner) + +#ifdef CONFIG_DEBUG_MUTEXES +#define mutex_destroy(l) rt_mutex_destroy(&(l)->lock) +#else +static inline void mutex_destroy(struct mutex *lock) {} +#endif + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +# define mutex_lock_nested(l, s) _mutex_lock_nested(l, s) +# define mutex_lock_interruptible_nested(l, s) \ + _mutex_lock_interruptible_nested(l, s) +# define mutex_lock_killable_nested(l, s) \ + _mutex_lock_killable_nested(l, s) +# define mutex_lock_io_nested(l, s) _mutex_lock_io_nested(l, s) + +# define mutex_lock_nest_lock(lock, nest_lock) \ +do { \ + typecheck(struct lockdep_map *, &(nest_lock)->dep_map); \ + _mutex_lock_nest_lock(lock, &(nest_lock)->dep_map); \ +} while (0) + +#else +# define mutex_lock_nested(l, s) _mutex_lock(l) +# define mutex_lock_interruptible_nested(l, s) \ + _mutex_lock_interruptible(l) +# define mutex_lock_killable_nested(l, s) \ + _mutex_lock_killable(l) +# define mutex_lock_nest_lock(lock, nest_lock) mutex_lock(lock) +# define mutex_lock_io_nested(l, s) _mutex_lock_io(l) +#endif + +# define mutex_init(mutex) \ +do { \ + static struct lock_class_key __key; \ + \ + rt_mutex_init(&(mutex)->lock); \ + __mutex_do_init((mutex), #mutex, &__key); \ +} while (0) + +# define __mutex_init(mutex, name, key) \ +do { \ + rt_mutex_init(&(mutex)->lock); \ + __mutex_do_init((mutex), name, key); \ +} while (0) + +/** + * These values are chosen such that FAIL and SUCCESS match the + * values of the regular mutex_trylock(). + */ +enum mutex_trylock_recursive_enum { + MUTEX_TRYLOCK_FAILED = 0, + MUTEX_TRYLOCK_SUCCESS = 1, + MUTEX_TRYLOCK_RECURSIVE, +}; +/** + * mutex_trylock_recursive - trylock variant that allows recursive locking + * @lock: mutex to be locked + * + * This function should not be used, _ever_. It is purely for hysterical GEM + * raisins, and once those are gone this will be removed. + * + * Returns: + * MUTEX_TRYLOCK_FAILED - trylock failed, + * MUTEX_TRYLOCK_SUCCESS - lock acquired, + * MUTEX_TRYLOCK_RECURSIVE - we already owned the lock. + */ +int __rt_mutex_owner_current(struct rt_mutex *lock); + +static inline /* __deprecated */ __must_check enum mutex_trylock_recursive_enum +mutex_trylock_recursive(struct mutex *lock) +{ + if (unlikely(__rt_mutex_owner_current(&lock->lock))) + return MUTEX_TRYLOCK_RECURSIVE; + + return mutex_trylock(lock); +} + +extern int atomic_dec_and_mutex_lock(atomic_t *cnt, struct mutex *lock); + +#endif Index: linux-5.6.17-rt10/include/linux/nfs_xdr.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/nfs_xdr.h +++ linux-5.6.17-rt10/include/linux/nfs_xdr.h @@ -1613,7 +1613,7 @@ struct nfs_unlinkdata { struct nfs_removeargs args; struct nfs_removeres res; struct dentry *dentry; - wait_queue_head_t wq; + struct swait_queue_head wq; const struct cred *cred; struct nfs_fattr dir_attr; long timeout; Index: linux-5.6.17-rt10/include/linux/percpu-rwsem.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/percpu-rwsem.h +++ linux-5.6.17-rt10/include/linux/percpu-rwsem.h @@ -3,41 +3,52 @@ #define _LINUX_PERCPU_RWSEM_H #include -#include #include #include +#include #include #include struct percpu_rw_semaphore { struct rcu_sync rss; unsigned int __percpu *read_count; - struct rw_semaphore rw_sem; /* slowpath */ - struct rcuwait writer; /* blocked writer */ - int readers_block; + struct rcuwait writer; + wait_queue_head_t waiters; + atomic_t block; +#ifdef CONFIG_DEBUG_LOCK_ALLOC + struct lockdep_map dep_map; +#endif }; +#ifdef CONFIG_DEBUG_LOCK_ALLOC +#define __PERCPU_RWSEM_DEP_MAP_INIT(lockname) .dep_map = { .name = #lockname }, +#else +#define __PERCPU_RWSEM_DEP_MAP_INIT(lockname) +#endif + #define __DEFINE_PERCPU_RWSEM(name, is_static) \ static DEFINE_PER_CPU(unsigned int, __percpu_rwsem_rc_##name); \ is_static struct percpu_rw_semaphore name = { \ .rss = __RCU_SYNC_INITIALIZER(name.rss), \ .read_count = &__percpu_rwsem_rc_##name, \ - .rw_sem = __RWSEM_INITIALIZER(name.rw_sem), \ .writer = __RCUWAIT_INITIALIZER(name.writer), \ + .waiters = __WAIT_QUEUE_HEAD_INITIALIZER(name.waiters), \ + .block = ATOMIC_INIT(0), \ + __PERCPU_RWSEM_DEP_MAP_INIT(name) \ } + #define DEFINE_PERCPU_RWSEM(name) \ __DEFINE_PERCPU_RWSEM(name, /* not static */) #define DEFINE_STATIC_PERCPU_RWSEM(name) \ __DEFINE_PERCPU_RWSEM(name, static) -extern int __percpu_down_read(struct percpu_rw_semaphore *, int); -extern void __percpu_up_read(struct percpu_rw_semaphore *); +extern bool __percpu_down_read(struct percpu_rw_semaphore *, bool); static inline void percpu_down_read(struct percpu_rw_semaphore *sem) { might_sleep(); - rwsem_acquire_read(&sem->rw_sem.dep_map, 0, 0, _RET_IP_); + rwsem_acquire_read(&sem->dep_map, 0, 0, _RET_IP_); preempt_disable(); /* @@ -48,8 +59,9 @@ static inline void percpu_down_read(stru * and that once the synchronize_rcu() is done, the writer will see * anything we did within this RCU-sched read-size critical section. */ - __this_cpu_inc(*sem->read_count); - if (unlikely(!rcu_sync_is_idle(&sem->rss))) + if (likely(rcu_sync_is_idle(&sem->rss))) + __this_cpu_inc(*sem->read_count); + else __percpu_down_read(sem, false); /* Unconditional memory barrier */ /* * The preempt_enable() prevents the compiler from @@ -58,16 +70,17 @@ static inline void percpu_down_read(stru preempt_enable(); } -static inline int percpu_down_read_trylock(struct percpu_rw_semaphore *sem) +static inline bool percpu_down_read_trylock(struct percpu_rw_semaphore *sem) { - int ret = 1; + bool ret = true; preempt_disable(); /* * Same as in percpu_down_read(). */ - __this_cpu_inc(*sem->read_count); - if (unlikely(!rcu_sync_is_idle(&sem->rss))) + if (likely(rcu_sync_is_idle(&sem->rss))) + __this_cpu_inc(*sem->read_count); + else ret = __percpu_down_read(sem, true); /* Unconditional memory barrier */ preempt_enable(); /* @@ -76,24 +89,36 @@ static inline int percpu_down_read_trylo */ if (ret) - rwsem_acquire_read(&sem->rw_sem.dep_map, 0, 1, _RET_IP_); + rwsem_acquire_read(&sem->dep_map, 0, 1, _RET_IP_); return ret; } static inline void percpu_up_read(struct percpu_rw_semaphore *sem) { + rwsem_release(&sem->dep_map, _RET_IP_); + preempt_disable(); /* * Same as in percpu_down_read(). */ - if (likely(rcu_sync_is_idle(&sem->rss))) + if (likely(rcu_sync_is_idle(&sem->rss))) { __this_cpu_dec(*sem->read_count); - else - __percpu_up_read(sem); /* Unconditional memory barrier */ + } else { + /* + * slowpath; reader will only ever wake a single blocked + * writer. + */ + smp_mb(); /* B matches C */ + /* + * In other words, if they see our decrement (presumably to + * aggregate zero, as that is the only time it matters) they + * will also see our critical section. + */ + __this_cpu_dec(*sem->read_count); + rcuwait_wake_up(&sem->writer); + } preempt_enable(); - - rwsem_release(&sem->rw_sem.dep_map, _RET_IP_); } extern void percpu_down_write(struct percpu_rw_semaphore *); @@ -110,29 +135,19 @@ extern void percpu_free_rwsem(struct per __percpu_init_rwsem(sem, #sem, &rwsem_key); \ }) -#define percpu_rwsem_is_held(sem) lockdep_is_held(&(sem)->rw_sem) - -#define percpu_rwsem_assert_held(sem) \ - lockdep_assert_held(&(sem)->rw_sem) +#define percpu_rwsem_is_held(sem) lockdep_is_held(sem) +#define percpu_rwsem_assert_held(sem) lockdep_assert_held(sem) static inline void percpu_rwsem_release(struct percpu_rw_semaphore *sem, bool read, unsigned long ip) { - lock_release(&sem->rw_sem.dep_map, ip); -#ifdef CONFIG_RWSEM_SPIN_ON_OWNER - if (!read) - atomic_long_set(&sem->rw_sem.owner, RWSEM_OWNER_UNKNOWN); -#endif + lock_release(&sem->dep_map, ip); } static inline void percpu_rwsem_acquire(struct percpu_rw_semaphore *sem, bool read, unsigned long ip) { - lock_acquire(&sem->rw_sem.dep_map, 0, 1, read, 1, NULL, ip); -#ifdef CONFIG_RWSEM_SPIN_ON_OWNER - if (!read) - atomic_long_set(&sem->rw_sem.owner, (long)current); -#endif + lock_acquire(&sem->dep_map, 0, 1, read, 1, NULL, ip); } #endif Index: linux-5.6.17-rt10/include/linux/percpu.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/percpu.h +++ linux-5.6.17-rt10/include/linux/percpu.h @@ -19,6 +19,35 @@ #define PERCPU_MODULE_RESERVE 0 #endif +#ifdef CONFIG_PREEMPT_RT + +#define get_local_var(var) (*({ \ + migrate_disable(); \ + this_cpu_ptr(&var); })) + +#define put_local_var(var) do { \ + (void)&(var); \ + migrate_enable(); \ +} while (0) + +# define get_local_ptr(var) ({ \ + migrate_disable(); \ + this_cpu_ptr(var); }) + +# define put_local_ptr(var) do { \ + (void)(var); \ + migrate_enable(); \ +} while (0) + +#else + +#define get_local_var(var) get_cpu_var(var) +#define put_local_var(var) put_cpu_var(var) +#define get_local_ptr(var) get_cpu_ptr(var) +#define put_local_ptr(var) put_cpu_ptr(var) + +#endif + /* minimum unit size, also is the maximum supported allocation size */ #define PCPU_MIN_UNIT_SIZE PFN_ALIGN(32 << 10) Index: linux-5.6.17-rt10/include/linux/pid.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/pid.h +++ linux-5.6.17-rt10/include/linux/pid.h @@ -3,6 +3,7 @@ #define _LINUX_PID_H #include +#include #include #include Index: linux-5.6.17-rt10/include/linux/posix-timers.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/posix-timers.h +++ linux-5.6.17-rt10/include/linux/posix-timers.h @@ -72,6 +72,7 @@ struct cpu_timer { struct task_struct *task; struct list_head elist; int firing; + int firing_cpu; }; static inline bool cpu_timer_enqueue(struct timerqueue_head *head, @@ -123,6 +124,9 @@ struct posix_cputimers { struct posix_cputimer_base bases[CPUCLOCK_MAX]; unsigned int timers_active; unsigned int expiry_active; +#ifdef CONFIG_PREEMPT_RT + struct task_struct *posix_timer_list; +#endif }; static inline void posix_cputimers_init(struct posix_cputimers *pct) @@ -152,9 +156,16 @@ static inline void posix_cputimers_rt_wa INIT_CPU_TIMERBASE(b[2]), \ } +#ifdef CONFIG_PREEMPT_RT +# define INIT_TIMER_LIST .posix_timer_list = NULL, +#else +# define INIT_TIMER_LIST +#endif + #define INIT_CPU_TIMERS(s) \ .posix_cputimers = { \ .bases = INIT_CPU_TIMERBASES(s.posix_cputimers.bases), \ + INIT_TIMER_LIST \ }, #else struct posix_cputimers { }; Index: linux-5.6.17-rt10/include/linux/preempt.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/preempt.h +++ linux-5.6.17-rt10/include/linux/preempt.h @@ -78,10 +78,8 @@ #include #define hardirq_count() (preempt_count() & HARDIRQ_MASK) -#define softirq_count() (preempt_count() & SOFTIRQ_MASK) #define irq_count() (preempt_count() & (HARDIRQ_MASK | SOFTIRQ_MASK \ | NMI_MASK)) - /* * Are we doing bottom half or hardware interrupt processing? * @@ -96,12 +94,23 @@ * should not be used in new code. */ #define in_irq() (hardirq_count()) -#define in_softirq() (softirq_count()) #define in_interrupt() (irq_count()) -#define in_serving_softirq() (softirq_count() & SOFTIRQ_OFFSET) #define in_nmi() (preempt_count() & NMI_MASK) #define in_task() (!(preempt_count() & \ (NMI_MASK | HARDIRQ_MASK | SOFTIRQ_OFFSET))) +#ifdef CONFIG_PREEMPT_RT + +#define softirq_count() ((long)current->softirq_count) +#define in_softirq() (softirq_count()) +#define in_serving_softirq() (current->softirq_count & SOFTIRQ_OFFSET) + +#else + +#define softirq_count() (preempt_count() & SOFTIRQ_MASK) +#define in_softirq() (softirq_count()) +#define in_serving_softirq() (softirq_count() & SOFTIRQ_OFFSET) + +#endif /* * The preempt_count offset after preempt_disable(); @@ -115,7 +124,11 @@ /* * The preempt_count offset after spin_lock() */ +#if !defined(CONFIG_PREEMPT_RT) #define PREEMPT_LOCK_OFFSET PREEMPT_DISABLE_OFFSET +#else +#define PREEMPT_LOCK_OFFSET 0 +#endif /* * The preempt_count offset needed for things like: @@ -164,6 +177,20 @@ extern void preempt_count_sub(int val); #define preempt_count_inc() preempt_count_add(1) #define preempt_count_dec() preempt_count_sub(1) +#ifdef CONFIG_PREEMPT_LAZY +#define add_preempt_lazy_count(val) do { preempt_lazy_count() += (val); } while (0) +#define sub_preempt_lazy_count(val) do { preempt_lazy_count() -= (val); } while (0) +#define inc_preempt_lazy_count() add_preempt_lazy_count(1) +#define dec_preempt_lazy_count() sub_preempt_lazy_count(1) +#define preempt_lazy_count() (current_thread_info()->preempt_lazy_count) +#else +#define add_preempt_lazy_count(val) do { } while (0) +#define sub_preempt_lazy_count(val) do { } while (0) +#define inc_preempt_lazy_count() do { } while (0) +#define dec_preempt_lazy_count() do { } while (0) +#define preempt_lazy_count() (0) +#endif + #ifdef CONFIG_PREEMPT_COUNT #define preempt_disable() \ @@ -172,13 +199,25 @@ do { \ barrier(); \ } while (0) +#define preempt_lazy_disable() \ +do { \ + inc_preempt_lazy_count(); \ + barrier(); \ +} while (0) + #define sched_preempt_enable_no_resched() \ do { \ barrier(); \ preempt_count_dec(); \ } while (0) -#define preempt_enable_no_resched() sched_preempt_enable_no_resched() +#ifdef CONFIG_PREEMPT_RT +# define preempt_enable_no_resched() sched_preempt_enable_no_resched() +# define preempt_check_resched_rt() preempt_check_resched() +#else +# define preempt_enable_no_resched() preempt_enable() +# define preempt_check_resched_rt() barrier(); +#endif #define preemptible() (preempt_count() == 0 && !irqs_disabled()) @@ -203,6 +242,13 @@ do { \ __preempt_schedule(); \ } while (0) +#define preempt_lazy_enable() \ +do { \ + dec_preempt_lazy_count(); \ + barrier(); \ + preempt_check_resched(); \ +} while (0) + #else /* !CONFIG_PREEMPTION */ #define preempt_enable() \ do { \ @@ -210,6 +256,12 @@ do { \ preempt_count_dec(); \ } while (0) +#define preempt_lazy_enable() \ +do { \ + dec_preempt_lazy_count(); \ + barrier(); \ +} while (0) + #define preempt_enable_notrace() \ do { \ barrier(); \ @@ -248,6 +300,7 @@ do { \ #define preempt_disable_notrace() barrier() #define preempt_enable_no_resched_notrace() barrier() #define preempt_enable_notrace() barrier() +#define preempt_check_resched_rt() barrier() #define preemptible() 0 #endif /* CONFIG_PREEMPT_COUNT */ @@ -268,10 +321,22 @@ do { \ } while (0) #define preempt_fold_need_resched() \ do { \ - if (tif_need_resched()) \ + if (tif_need_resched_now()) \ set_preempt_need_resched(); \ } while (0) +#ifdef CONFIG_PREEMPT_RT +# define preempt_disable_rt() preempt_disable() +# define preempt_enable_rt() preempt_enable() +# define preempt_disable_nort() barrier() +# define preempt_enable_nort() barrier() +#else +# define preempt_disable_rt() barrier() +# define preempt_enable_rt() barrier() +# define preempt_disable_nort() preempt_disable() +# define preempt_enable_nort() preempt_enable() +#endif + #ifdef CONFIG_PREEMPT_NOTIFIERS struct preempt_notifier; @@ -322,4 +387,56 @@ static inline void preempt_notifier_init #endif +#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) + +extern void migrate_disable(void); +extern void migrate_enable(void); + +int __migrate_disabled(struct task_struct *p); + +#elif !defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) + +extern void migrate_disable(void); +extern void migrate_enable(void); +static inline int __migrate_disabled(struct task_struct *p) +{ + return 0; +} + +#else +/** + * migrate_disable - Prevent migration of the current task + * + * Maps to preempt_disable() which also disables preemption. Use + * migrate_disable() to annotate that the intent is to prevent migration, + * but not necessarily preemption. + * + * Can be invoked nested like preempt_disable() and needs the corresponding + * number of migrate_enable() invocations. + */ +static __always_inline void migrate_disable(void) +{ + preempt_disable(); +} + +/** + * migrate_enable - Allow migration of the current task + * + * Counterpart to migrate_disable(). + * + * As migrate_disable() can be invoked nested, only the outermost invocation + * reenables migration. + * + * Currently mapped to preempt_enable(). + */ +static __always_inline void migrate_enable(void) +{ + preempt_enable(); +} + +static inline int __migrate_disabled(struct task_struct *p) +{ + return 0; +} +#endif #endif /* __LINUX_PREEMPT_H */ Index: linux-5.6.17-rt10/include/linux/printk.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/printk.h +++ linux-5.6.17-rt10/include/linux/printk.h @@ -58,6 +58,7 @@ static inline const char *printk_skip_he */ #define CONSOLE_LOGLEVEL_DEFAULT CONFIG_CONSOLE_LOGLEVEL_DEFAULT #define CONSOLE_LOGLEVEL_QUIET CONFIG_CONSOLE_LOGLEVEL_QUIET +#define CONSOLE_LOGLEVEL_EMERGENCY CONFIG_CONSOLE_LOGLEVEL_EMERGENCY extern int console_printk[]; @@ -65,6 +66,7 @@ extern int console_printk[]; #define default_message_loglevel (console_printk[1]) #define minimum_console_loglevel (console_printk[2]) #define default_console_loglevel (console_printk[3]) +#define emergency_console_loglevel (console_printk[4]) static inline void console_silent(void) { @@ -146,18 +148,6 @@ static inline __printf(1, 2) __cold void early_printk(const char *s, ...) { } #endif -#ifdef CONFIG_PRINTK_NMI -extern void printk_nmi_enter(void); -extern void printk_nmi_exit(void); -extern void printk_nmi_direct_enter(void); -extern void printk_nmi_direct_exit(void); -#else -static inline void printk_nmi_enter(void) { } -static inline void printk_nmi_exit(void) { } -static inline void printk_nmi_direct_enter(void) { } -static inline void printk_nmi_direct_exit(void) { } -#endif /* PRINTK_NMI */ - #ifdef CONFIG_PRINTK asmlinkage __printf(5, 0) int vprintk_emit(int facility, int level, @@ -202,8 +192,7 @@ __printf(1, 2) void dump_stack_set_arch_ void dump_stack_print_info(const char *log_lvl); void show_regs_print_info(const char *log_lvl); extern asmlinkage void dump_stack(void) __cold; -extern void printk_safe_flush(void); -extern void printk_safe_flush_on_panic(void); +struct wait_queue_head *printk_wait_queue(void); #else static inline __printf(1, 0) int vprintk(const char *s, va_list args) @@ -268,13 +257,6 @@ static inline void dump_stack(void) { } -static inline void printk_safe_flush(void) -{ -} - -static inline void printk_safe_flush_on_panic(void) -{ -} #endif extern int kptr_restrict; Index: linux-5.6.17-rt10/include/linux/printk_ringbuffer.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/printk_ringbuffer.h @@ -0,0 +1,114 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _LINUX_PRINTK_RINGBUFFER_H +#define _LINUX_PRINTK_RINGBUFFER_H + +#include +#include +#include +#include + +struct prb_cpulock { + atomic_t owner; + unsigned long __percpu *irqflags; +}; + +struct printk_ringbuffer { + void *buffer; + unsigned int size_bits; + + u64 seq; + atomic_long_t lost; + + atomic_long_t tail; + atomic_long_t head; + atomic_long_t reserve; + + struct prb_cpulock *cpulock; + atomic_t ctx; + + struct wait_queue_head *wq; + atomic_long_t wq_counter; + struct irq_work *wq_work; +}; + +struct prb_entry { + unsigned int size; + u64 seq; + char data[0]; +}; + +struct prb_handle { + struct printk_ringbuffer *rb; + unsigned int cpu; + struct prb_entry *entry; +}; + +#define DECLARE_STATIC_PRINTKRB_CPULOCK(name) \ +static DEFINE_PER_CPU(unsigned long, _##name##_percpu_irqflags); \ +static struct prb_cpulock name = { \ + .owner = ATOMIC_INIT(-1), \ + .irqflags = &_##name##_percpu_irqflags, \ +} + +#define PRB_INIT ((unsigned long)-1) + +#define DECLARE_STATIC_PRINTKRB_ITER(name, rbaddr) \ +static struct prb_iterator name = { \ + .rb = rbaddr, \ + .lpos = PRB_INIT, \ +} + +struct prb_iterator { + struct printk_ringbuffer *rb; + unsigned long lpos; +}; + +#define DECLARE_STATIC_PRINTKRB(name, szbits, cpulockptr) \ +static char _##name##_buffer[1 << (szbits)] \ + __aligned(__alignof__(long)); \ +static DECLARE_WAIT_QUEUE_HEAD(_##name##_wait); \ +static void _##name##_wake_work_func(struct irq_work *irq_work) \ +{ \ + wake_up_interruptible_all(&_##name##_wait); \ +} \ +static struct irq_work _##name##_wake_work = { \ + .func = _##name##_wake_work_func, \ + .flags = ATOMIC_INIT(IRQ_WORK_LAZY), \ +}; \ +static struct printk_ringbuffer name = { \ + .buffer = &_##name##_buffer[0], \ + .size_bits = szbits, \ + .seq = 0, \ + .lost = ATOMIC_LONG_INIT(0), \ + .tail = ATOMIC_LONG_INIT(-111 * sizeof(long)), \ + .head = ATOMIC_LONG_INIT(-111 * sizeof(long)), \ + .reserve = ATOMIC_LONG_INIT(-111 * sizeof(long)), \ + .cpulock = cpulockptr, \ + .ctx = ATOMIC_INIT(0), \ + .wq = &_##name##_wait, \ + .wq_counter = ATOMIC_LONG_INIT(0), \ + .wq_work = &_##name##_wake_work, \ +} + +/* writer interface */ +char *prb_reserve(struct prb_handle *h, struct printk_ringbuffer *rb, + unsigned int size); +void prb_commit(struct prb_handle *h); + +/* reader interface */ +void prb_iter_init(struct prb_iterator *iter, struct printk_ringbuffer *rb, + u64 *seq); +void prb_iter_copy(struct prb_iterator *dest, struct prb_iterator *src); +int prb_iter_next(struct prb_iterator *iter, char *buf, int size, u64 *seq); +int prb_iter_wait_next(struct prb_iterator *iter, char *buf, int size, + u64 *seq); +int prb_iter_seek(struct prb_iterator *iter, u64 seq); +int prb_iter_data(struct prb_iterator *iter, char *buf, int size, u64 *seq); + +/* utility functions */ +int prb_buffer_size(struct printk_ringbuffer *rb); +void prb_inc_lost(struct printk_ringbuffer *rb); +void prb_lock(struct prb_cpulock *cpu_lock, unsigned int *cpu_store); +void prb_unlock(struct prb_cpulock *cpu_lock, unsigned int cpu_store); + +#endif /*_LINUX_PRINTK_RINGBUFFER_H */ Index: linux-5.6.17-rt10/include/linux/radix-tree.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/radix-tree.h +++ linux-5.6.17-rt10/include/linux/radix-tree.h @@ -226,6 +226,7 @@ unsigned int radix_tree_gang_lookup(cons unsigned int max_items); int radix_tree_preload(gfp_t gfp_mask); int radix_tree_maybe_preload(gfp_t gfp_mask); +void radix_tree_preload_end(void); void radix_tree_init(void); void *radix_tree_tag_set(struct radix_tree_root *, unsigned long index, unsigned int tag); @@ -243,11 +244,6 @@ unsigned int radix_tree_gang_lookup_tag_ unsigned int max_items, unsigned int tag); int radix_tree_tagged(const struct radix_tree_root *, unsigned int tag); -static inline void radix_tree_preload_end(void) -{ - preempt_enable(); -} - void __rcu **idr_get_free(struct radix_tree_root *root, struct radix_tree_iter *iter, gfp_t gfp, unsigned long max); Index: linux-5.6.17-rt10/include/linux/random.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/random.h +++ linux-5.6.17-rt10/include/linux/random.h @@ -33,7 +33,7 @@ static inline void add_latent_entropy(vo extern void add_input_randomness(unsigned int type, unsigned int code, unsigned int value) __latent_entropy; -extern void add_interrupt_randomness(int irq, int irq_flags) __latent_entropy; +extern void add_interrupt_randomness(int irq, int irq_flags, __u64 ip) __latent_entropy; extern void get_random_bytes(void *buf, int nbytes); extern int wait_for_random_bytes(void); Index: linux-5.6.17-rt10/include/linux/ratelimit.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/ratelimit.h +++ linux-5.6.17-rt10/include/linux/ratelimit.h @@ -59,7 +59,7 @@ static inline void ratelimit_state_exit( return; if (rs->missed) { - pr_warn("%s: %d output lines suppressed due to ratelimiting\n", + pr_info("%s: %d output lines suppressed due to ratelimiting\n", current->comm, rs->missed); rs->missed = 0; } Index: linux-5.6.17-rt10/include/linux/rbtree.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/rbtree.h +++ linux-5.6.17-rt10/include/linux/rbtree.h @@ -19,7 +19,7 @@ #include #include -#include +#include struct rb_node { unsigned long __rb_parent_color; Index: linux-5.6.17-rt10/include/linux/rcu_assign_pointer.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/rcu_assign_pointer.h @@ -0,0 +1,62 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +#ifndef __LINUX_RCU_ASSIGN_POINTER_H__ +#define __LINUX_RCU_ASSIGN_POINTER_H__ +#include +#include + +#ifdef __CHECKER__ +#define rcu_check_sparse(p, space) \ + ((void)(((typeof(*p) space *)p) == p)) +#else /* #ifdef __CHECKER__ */ +#define rcu_check_sparse(p, space) +#endif /* #else #ifdef __CHECKER__ */ + +/** + * RCU_INITIALIZER() - statically initialize an RCU-protected global variable + * @v: The value to statically initialize with. + */ +#define RCU_INITIALIZER(v) (typeof(*(v)) __force __rcu *)(v) + +/** + * rcu_assign_pointer() - assign to RCU-protected pointer + * @p: pointer to assign to + * @v: value to assign (publish) + * + * Assigns the specified value to the specified RCU-protected + * pointer, ensuring that any concurrent RCU readers will see + * any prior initialization. + * + * Inserts memory barriers on architectures that require them + * (which is most of them), and also prevents the compiler from + * reordering the code that initializes the structure after the pointer + * assignment. More importantly, this call documents which pointers + * will be dereferenced by RCU read-side code. + * + * In some special cases, you may use RCU_INIT_POINTER() instead + * of rcu_assign_pointer(). RCU_INIT_POINTER() is a bit faster due + * to the fact that it does not constrain either the CPU or the compiler. + * That said, using RCU_INIT_POINTER() when you should have used + * rcu_assign_pointer() is a very bad thing that results in + * impossible-to-diagnose memory corruption. So please be careful. + * See the RCU_INIT_POINTER() comment header for details. + * + * Note that rcu_assign_pointer() evaluates each of its arguments only + * once, appearances notwithstanding. One of the "extra" evaluations + * is in typeof() and the other visible only to sparse (__CHECKER__), + * neither of which actually execute the argument. As with most cpp + * macros, this execute-arguments-only-once property is important, so + * please be careful when making changes to rcu_assign_pointer() and the + * other macros that it invokes. + */ +#define rcu_assign_pointer(p, v) \ +do { \ + uintptr_t _r_a_p__v = (uintptr_t)(v); \ + rcu_check_sparse(p, __rcu); \ + \ + if (__builtin_constant_p(v) && (_r_a_p__v) == (uintptr_t)NULL) \ + WRITE_ONCE((p), (typeof(p))(_r_a_p__v)); \ + else \ + smp_store_release(&p, RCU_INITIALIZER((typeof(p))_r_a_p__v)); \ +} while (0) + +#endif Index: linux-5.6.17-rt10/include/linux/rcupdate.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/rcupdate.h +++ linux-5.6.17-rt10/include/linux/rcupdate.h @@ -29,6 +29,7 @@ #include #include #include +#include #define ULONG_CMP_GE(a, b) (ULONG_MAX / 2 >= (a) - (b)) #define ULONG_CMP_LT(a, b) (ULONG_MAX / 2 < (a) - (b)) @@ -51,6 +52,11 @@ void __rcu_read_unlock(void); * types of kernel builds, the rcu_read_lock() nesting depth is unknowable. */ #define rcu_preempt_depth() (current->rcu_read_lock_nesting) +#ifndef CONFIG_PREEMPT_RT +#define sched_rcu_preempt_depth() rcu_preempt_depth() +#else +static inline int sched_rcu_preempt_depth(void) { return 0; } +#endif #else /* #ifdef CONFIG_PREEMPT_RCU */ @@ -69,6 +75,8 @@ static inline int rcu_preempt_depth(void return 0; } +#define sched_rcu_preempt_depth() rcu_preempt_depth() + #endif /* #else #ifdef CONFIG_PREEMPT_RCU */ /* Internal to kernel */ @@ -279,7 +287,8 @@ static inline void rcu_preempt_sleep_che #define rcu_sleep_check() \ do { \ rcu_preempt_sleep_check(); \ - RCU_LOCKDEP_WARN(lock_is_held(&rcu_bh_lock_map), \ + if (!IS_ENABLED(CONFIG_PREEMPT_RT)) \ + RCU_LOCKDEP_WARN(lock_is_held(&rcu_bh_lock_map), \ "Illegal context switch in RCU-bh read-side critical section"); \ RCU_LOCKDEP_WARN(lock_is_held(&rcu_sched_lock_map), \ "Illegal context switch in RCU-sched read-side critical section"); \ @@ -300,13 +309,6 @@ static inline void rcu_preempt_sleep_che * (e.g., __srcu), should this make sense in the future. */ -#ifdef __CHECKER__ -#define rcu_check_sparse(p, space) \ - ((void)(((typeof(*p) space *)p) == p)) -#else /* #ifdef __CHECKER__ */ -#define rcu_check_sparse(p, space) -#endif /* #else #ifdef __CHECKER__ */ - #define __rcu_access_pointer(p, space) \ ({ \ typeof(*p) *_________p1 = (typeof(*p) *__force)READ_ONCE(p); \ @@ -335,54 +337,6 @@ static inline void rcu_preempt_sleep_che }) /** - * RCU_INITIALIZER() - statically initialize an RCU-protected global variable - * @v: The value to statically initialize with. - */ -#define RCU_INITIALIZER(v) (typeof(*(v)) __force __rcu *)(v) - -/** - * rcu_assign_pointer() - assign to RCU-protected pointer - * @p: pointer to assign to - * @v: value to assign (publish) - * - * Assigns the specified value to the specified RCU-protected - * pointer, ensuring that any concurrent RCU readers will see - * any prior initialization. - * - * Inserts memory barriers on architectures that require them - * (which is most of them), and also prevents the compiler from - * reordering the code that initializes the structure after the pointer - * assignment. More importantly, this call documents which pointers - * will be dereferenced by RCU read-side code. - * - * In some special cases, you may use RCU_INIT_POINTER() instead - * of rcu_assign_pointer(). RCU_INIT_POINTER() is a bit faster due - * to the fact that it does not constrain either the CPU or the compiler. - * That said, using RCU_INIT_POINTER() when you should have used - * rcu_assign_pointer() is a very bad thing that results in - * impossible-to-diagnose memory corruption. So please be careful. - * See the RCU_INIT_POINTER() comment header for details. - * - * Note that rcu_assign_pointer() evaluates each of its arguments only - * once, appearances notwithstanding. One of the "extra" evaluations - * is in typeof() and the other visible only to sparse (__CHECKER__), - * neither of which actually execute the argument. As with most cpp - * macros, this execute-arguments-only-once property is important, so - * please be careful when making changes to rcu_assign_pointer() and the - * other macros that it invokes. - */ -#define rcu_assign_pointer(p, v) \ -do { \ - uintptr_t _r_a_p__v = (uintptr_t)(v); \ - rcu_check_sparse(p, __rcu); \ - \ - if (__builtin_constant_p(v) && (_r_a_p__v) == (uintptr_t)NULL) \ - WRITE_ONCE((p), (typeof(p))(_r_a_p__v)); \ - else \ - smp_store_release(&p, RCU_INITIALIZER((typeof(p))_r_a_p__v)); \ -} while (0) - -/** * rcu_replace_pointer() - replace an RCU pointer, returning its old value * @rcu_ptr: RCU pointer, whose old value is returned * @ptr: regular pointer Index: linux-5.6.17-rt10/include/linux/rtmutex.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/rtmutex.h +++ linux-5.6.17-rt10/include/linux/rtmutex.h @@ -14,11 +14,15 @@ #define __LINUX_RT_MUTEX_H #include +#include #include -#include extern int max_lock_depth; /* for sysctl */ +#ifdef CONFIG_DEBUG_MUTEXES +#include +#endif + /** * The rt_mutex structure * @@ -31,8 +35,8 @@ struct rt_mutex { raw_spinlock_t wait_lock; struct rb_root_cached waiters; struct task_struct *owner; -#ifdef CONFIG_DEBUG_RT_MUTEXES int save_state; +#ifdef CONFIG_DEBUG_RT_MUTEXES const char *name, *file; int line; void *magic; @@ -82,16 +86,23 @@ do { \ #define __DEP_MAP_RT_MUTEX_INITIALIZER(mutexname) #endif -#define __RT_MUTEX_INITIALIZER(mutexname) \ - { .wait_lock = __RAW_SPIN_LOCK_UNLOCKED(mutexname.wait_lock) \ +#define __RT_MUTEX_INITIALIZER_PLAIN(mutexname) \ + .wait_lock = __RAW_SPIN_LOCK_UNLOCKED(mutexname.wait_lock) \ , .waiters = RB_ROOT_CACHED \ , .owner = NULL \ __DEBUG_RT_MUTEX_INITIALIZER(mutexname) \ - __DEP_MAP_RT_MUTEX_INITIALIZER(mutexname)} + __DEP_MAP_RT_MUTEX_INITIALIZER(mutexname) + +#define __RT_MUTEX_INITIALIZER(mutexname) \ + { __RT_MUTEX_INITIALIZER_PLAIN(mutexname) } #define DEFINE_RT_MUTEX(mutexname) \ struct rt_mutex mutexname = __RT_MUTEX_INITIALIZER(mutexname) +#define __RT_MUTEX_INITIALIZER_SAVE_STATE(mutexname) \ + { __RT_MUTEX_INITIALIZER_PLAIN(mutexname) \ + , .save_state = 1 } + /** * rt_mutex_is_locked - is the mutex locked * @lock: the mutex to be queried @@ -115,6 +126,7 @@ extern void rt_mutex_lock(struct rt_mute #endif extern int rt_mutex_lock_interruptible(struct rt_mutex *lock); +extern int rt_mutex_lock_killable(struct rt_mutex *lock); extern int rt_mutex_timed_lock(struct rt_mutex *lock, struct hrtimer_sleeper *timeout); Index: linux-5.6.17-rt10/include/linux/rwlock_rt.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/rwlock_rt.h @@ -0,0 +1,119 @@ +#ifndef __LINUX_RWLOCK_RT_H +#define __LINUX_RWLOCK_RT_H + +#ifndef __LINUX_SPINLOCK_H +#error Do not include directly. Use spinlock.h +#endif + +extern void __lockfunc rt_write_lock(rwlock_t *rwlock); +extern void __lockfunc rt_read_lock(rwlock_t *rwlock); +extern int __lockfunc rt_write_trylock(rwlock_t *rwlock); +extern int __lockfunc rt_read_trylock(rwlock_t *rwlock); +extern void __lockfunc rt_write_unlock(rwlock_t *rwlock); +extern void __lockfunc rt_read_unlock(rwlock_t *rwlock); +extern int __lockfunc rt_read_can_lock(rwlock_t *rwlock); +extern int __lockfunc rt_write_can_lock(rwlock_t *rwlock); +extern void __rt_rwlock_init(rwlock_t *rwlock, char *name, struct lock_class_key *key); + +#define read_can_lock(rwlock) rt_read_can_lock(rwlock) +#define write_can_lock(rwlock) rt_write_can_lock(rwlock) + +#define read_trylock(lock) __cond_lock(lock, rt_read_trylock(lock)) +#define write_trylock(lock) __cond_lock(lock, rt_write_trylock(lock)) + +static inline int __write_trylock_rt_irqsave(rwlock_t *lock, unsigned long *flags) +{ + /* XXX ARCH_IRQ_ENABLED */ + *flags = 0; + return rt_write_trylock(lock); +} + +#define write_trylock_irqsave(lock, flags) \ + __cond_lock(lock, __write_trylock_rt_irqsave(lock, &(flags))) + +#define read_lock_irqsave(lock, flags) \ + do { \ + typecheck(unsigned long, flags); \ + rt_read_lock(lock); \ + flags = 0; \ + } while (0) + +#define write_lock_irqsave(lock, flags) \ + do { \ + typecheck(unsigned long, flags); \ + rt_write_lock(lock); \ + flags = 0; \ + } while (0) + +#define read_lock(lock) rt_read_lock(lock) + +#define read_lock_bh(lock) \ + do { \ + local_bh_disable(); \ + rt_read_lock(lock); \ + } while (0) + +#define read_lock_irq(lock) read_lock(lock) + +#define write_lock(lock) rt_write_lock(lock) + +#define write_lock_bh(lock) \ + do { \ + local_bh_disable(); \ + rt_write_lock(lock); \ + } while (0) + +#define write_lock_irq(lock) write_lock(lock) + +#define read_unlock(lock) rt_read_unlock(lock) + +#define read_unlock_bh(lock) \ + do { \ + rt_read_unlock(lock); \ + local_bh_enable(); \ + } while (0) + +#define read_unlock_irq(lock) read_unlock(lock) + +#define write_unlock(lock) rt_write_unlock(lock) + +#define write_unlock_bh(lock) \ + do { \ + rt_write_unlock(lock); \ + local_bh_enable(); \ + } while (0) + +#define write_unlock_irq(lock) write_unlock(lock) + +#define read_unlock_irqrestore(lock, flags) \ + do { \ + typecheck(unsigned long, flags); \ + (void) flags; \ + rt_read_unlock(lock); \ + } while (0) + +#define write_unlock_irqrestore(lock, flags) \ + do { \ + typecheck(unsigned long, flags); \ + (void) flags; \ + rt_write_unlock(lock); \ + } while (0) + +#define rwlock_init(rwl) \ +do { \ + static struct lock_class_key __key; \ + \ + __rt_rwlock_init(rwl, #rwl, &__key); \ +} while (0) + +/* + * Internal functions made global for CPU pinning + */ +void __read_rt_lock(struct rt_rw_lock *lock); +int __read_rt_trylock(struct rt_rw_lock *lock); +void __write_rt_lock(struct rt_rw_lock *lock); +int __write_rt_trylock(struct rt_rw_lock *lock); +void __read_rt_unlock(struct rt_rw_lock *lock); +void __write_rt_unlock(struct rt_rw_lock *lock); + +#endif Index: linux-5.6.17-rt10/include/linux/rwlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/rwlock_types.h +++ linux-5.6.17-rt10/include/linux/rwlock_types.h @@ -1,6 +1,10 @@ #ifndef __LINUX_RWLOCK_TYPES_H #define __LINUX_RWLOCK_TYPES_H +#if !defined(__LINUX_SPINLOCK_TYPES_H) +# error "Do not include directly, include spinlock_types.h" +#endif + /* * include/linux/rwlock_types.h - generic rwlock type definitions * and initializers Index: linux-5.6.17-rt10/include/linux/rwlock_types_rt.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/rwlock_types_rt.h @@ -0,0 +1,55 @@ +#ifndef __LINUX_RWLOCK_TYPES_RT_H +#define __LINUX_RWLOCK_TYPES_RT_H + +#ifndef __LINUX_SPINLOCK_TYPES_H +#error "Do not include directly. Include spinlock_types.h instead" +#endif + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +# define RW_DEP_MAP_INIT(lockname) .dep_map = { .name = #lockname } +#else +# define RW_DEP_MAP_INIT(lockname) +#endif + +typedef struct rt_rw_lock rwlock_t; + +#define __RW_LOCK_UNLOCKED(name) __RWLOCK_RT_INITIALIZER(name) + +#define DEFINE_RWLOCK(name) \ + rwlock_t name = __RW_LOCK_UNLOCKED(name) + +/* + * A reader biased implementation primarily for CPU pinning. + * + * Can be selected as general replacement for the single reader RT rwlock + * variant + */ +struct rt_rw_lock { + struct rt_mutex rtmutex; + atomic_t readers; +#ifdef CONFIG_DEBUG_LOCK_ALLOC + struct lockdep_map dep_map; +#endif +}; + +#define READER_BIAS (1U << 31) +#define WRITER_BIAS (1U << 30) + +#define __RWLOCK_RT_INITIALIZER(name) \ +{ \ + .readers = ATOMIC_INIT(READER_BIAS), \ + .rtmutex = __RT_MUTEX_INITIALIZER_SAVE_STATE(name.rtmutex), \ + RW_DEP_MAP_INIT(name) \ +} + +void __rwlock_biased_rt_init(struct rt_rw_lock *lock, const char *name, + struct lock_class_key *key); + +#define rwlock_biased_rt_init(rwlock) \ + do { \ + static struct lock_class_key __key; \ + \ + __rwlock_biased_rt_init((rwlock), #rwlock, &__key); \ + } while (0) + +#endif Index: linux-5.6.17-rt10/include/linux/rwsem-rt.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/rwsem-rt.h @@ -0,0 +1,68 @@ +#ifndef _LINUX_RWSEM_RT_H +#define _LINUX_RWSEM_RT_H + +#ifndef _LINUX_RWSEM_H +#error "Include rwsem.h" +#endif + +#include +#include + +#define READER_BIAS (1U << 31) +#define WRITER_BIAS (1U << 30) + +struct rw_semaphore { + atomic_t readers; + struct rt_mutex rtmutex; +#ifdef CONFIG_DEBUG_LOCK_ALLOC + struct lockdep_map dep_map; +#endif +}; + +#define __RWSEM_INITIALIZER(name) \ +{ \ + .readers = ATOMIC_INIT(READER_BIAS), \ + .rtmutex = __RT_MUTEX_INITIALIZER(name.rtmutex), \ + RW_DEP_MAP_INIT(name) \ +} + +#define DECLARE_RWSEM(lockname) \ + struct rw_semaphore lockname = __RWSEM_INITIALIZER(lockname) + +extern void __rwsem_init(struct rw_semaphore *rwsem, const char *name, + struct lock_class_key *key); + +#define __init_rwsem(sem, name, key) \ +do { \ + rt_mutex_init(&(sem)->rtmutex); \ + __rwsem_init((sem), (name), (key)); \ +} while (0) + +#define init_rwsem(sem) \ +do { \ + static struct lock_class_key __key; \ + \ + __init_rwsem((sem), #sem, &__key); \ +} while (0) + +static inline int rwsem_is_locked(struct rw_semaphore *sem) +{ + return atomic_read(&sem->readers) != READER_BIAS; +} + +static inline int rwsem_is_contended(struct rw_semaphore *sem) +{ + return atomic_read(&sem->readers) > 0; +} + +extern void __down_read(struct rw_semaphore *sem); +extern int __down_read_killable(struct rw_semaphore *sem); +extern int __down_read_trylock(struct rw_semaphore *sem); +extern void __down_write(struct rw_semaphore *sem); +extern int __must_check __down_write_killable(struct rw_semaphore *sem); +extern int __down_write_trylock(struct rw_semaphore *sem); +extern void __up_read(struct rw_semaphore *sem); +extern void __up_write(struct rw_semaphore *sem); +extern void __downgrade_write(struct rw_semaphore *sem); + +#endif Index: linux-5.6.17-rt10/include/linux/rwsem.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/rwsem.h +++ linux-5.6.17-rt10/include/linux/rwsem.h @@ -16,6 +16,11 @@ #include #include #include + +#ifdef CONFIG_PREEMPT_RT +#include +#else /* PREEMPT_RT */ + #ifdef CONFIG_RWSEM_SPIN_ON_OWNER #include #endif @@ -53,12 +58,6 @@ struct rw_semaphore { #endif }; -/* - * Setting all bits of the owner field except bit 0 will indicate - * that the rwsem is writer-owned with an unknown owner. - */ -#define RWSEM_OWNER_UNKNOWN (-2L) - /* In all implementations count != 0 means locked */ static inline int rwsem_is_locked(struct rw_semaphore *sem) { @@ -121,6 +120,13 @@ static inline int rwsem_is_contended(str return !list_empty(&sem->wait_list); } +#endif /* !PREEMPT_RT */ + +/* + * The functions below are the same for all rwsem implementations including + * the RT specific variant. + */ + /* * lock for reading */ Index: linux-5.6.17-rt10/include/linux/sched.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/sched.h +++ linux-5.6.17-rt10/include/linux/sched.h @@ -31,6 +31,7 @@ #include #include #include +#include /* task_struct member predeclarations (sorted alphabetically): */ struct audit_context; @@ -107,12 +108,8 @@ struct task_group; __TASK_TRACED | EXIT_DEAD | EXIT_ZOMBIE | \ TASK_PARKED) -#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0) - #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0) -#define task_is_stopped_or_traced(task) ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0) - #define task_contributes_to_load(task) ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \ (task->flags & PF_FROZEN) == 0 && \ (task->state & TASK_NOLOAD) == 0) @@ -140,6 +137,9 @@ struct task_group; smp_store_mb(current->state, (state_value)); \ } while (0) +#define __set_current_state_no_track(state_value) \ + current->state = (state_value); + #define set_special_state(state_value) \ do { \ unsigned long flags; /* may shadow */ \ @@ -149,6 +149,7 @@ struct task_group; current->state = (state_value); \ raw_spin_unlock_irqrestore(¤t->pi_lock, flags); \ } while (0) + #else /* * set_current_state() includes a barrier so that the write of current->state @@ -193,6 +194,9 @@ struct task_group; #define set_current_state(state_value) \ smp_store_mb(current->state, (state_value)) +#define __set_current_state_no_track(state_value) \ + __set_current_state(state_value) + /* * set_special_state() should be used for those states when the blocking task * can not use the regular condition based wait-loop. In that case we must @@ -230,6 +234,8 @@ extern void io_schedule_finish(int token extern long io_schedule_timeout(long timeout); extern void io_schedule(void); +int cpu_nr_pinned(int cpu); + /** * struct prev_cputime - snapshot of system and user cputime * @utime: time spent in user mode @@ -636,6 +642,8 @@ struct task_struct { #endif /* -1 unrunnable, 0 runnable, >0 stopped: */ volatile long state; + /* saved state for "spinlock sleepers" */ + volatile long saved_state; /* * This begins the randomizable portion of task_struct. Only @@ -705,6 +713,20 @@ struct task_struct { int nr_cpus_allowed; const cpumask_t *cpus_ptr; cpumask_t cpus_mask; +#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) + int migrate_disable; + bool migrate_disable_scheduled; +# ifdef CONFIG_SCHED_DEBUG + int pinned_on_cpu; +# endif +#elif !defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) +# ifdef CONFIG_SCHED_DEBUG + int migrate_disable; +# endif +#endif +#ifdef CONFIG_PREEMPT_RT + int sleeping_lock; +#endif #ifdef CONFIG_PREEMPT_RCU int rcu_read_lock_nesting; @@ -918,11 +940,16 @@ struct task_struct { /* Signal handlers: */ struct signal_struct *signal; struct sighand_struct __rcu *sighand; + struct sigqueue *sigqueue_cache; sigset_t blocked; sigset_t real_blocked; /* Restored if set_restore_sigmask() was used: */ sigset_t saved_sigmask; struct sigpending pending; +#ifdef CONFIG_PREEMPT_RT + /* TODO: move me into ->restart_block ? */ + struct kernel_siginfo forced_info; +#endif unsigned long sas_ss_sp; size_t sas_ss_size; unsigned int sas_ss_flags; @@ -949,6 +976,7 @@ struct task_struct { raw_spinlock_t pi_lock; struct wake_q_node wake_q; + struct wake_q_node wake_q_sleeper; #ifdef CONFIG_RT_MUTEXES /* PI waiters blocked on a rt_mutex held by this task: */ @@ -983,6 +1011,9 @@ struct task_struct { int softirqs_enabled; int softirq_context; #endif +#ifdef CONFIG_PREEMPT_RT + int softirq_count; +#endif #ifdef CONFIG_LOCKDEP # define MAX_LOCK_DEPTH 48UL @@ -1038,7 +1069,7 @@ struct task_struct { /* Protected by ->alloc_lock: */ nodemask_t mems_allowed; /* Seqence number to catch updates: */ - seqcount_t mems_allowed_seq; + seqcount_spinlock_t mems_allowed_seq; int cpuset_mem_spread_rotor; int cpuset_slab_spread_rotor; #endif @@ -1254,6 +1285,12 @@ struct task_struct { unsigned int sequential_io; unsigned int sequential_io_avg; #endif +#ifdef CONFIG_PREEMPT_RT +# if defined CONFIG_HIGHMEM || defined CONFIG_X86_32 + int kmap_idx; + pte_t kmap_pte[KM_TYPE_NR]; +# endif +#endif #ifdef CONFIG_DEBUG_ATOMIC_SLEEP unsigned long task_state_change; #endif @@ -1686,6 +1723,7 @@ extern struct task_struct *find_get_task extern int wake_up_state(struct task_struct *tsk, unsigned int state); extern int wake_up_process(struct task_struct *tsk); +extern int wake_up_lock_sleeper(struct task_struct *tsk); extern void wake_up_new_task(struct task_struct *tsk); #ifdef CONFIG_SMP @@ -1768,6 +1806,89 @@ static inline int test_tsk_need_resched( return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED)); } +#ifdef CONFIG_PREEMPT_LAZY +static inline void set_tsk_need_resched_lazy(struct task_struct *tsk) +{ + set_tsk_thread_flag(tsk,TIF_NEED_RESCHED_LAZY); +} + +static inline void clear_tsk_need_resched_lazy(struct task_struct *tsk) +{ + clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED_LAZY); +} + +static inline int test_tsk_need_resched_lazy(struct task_struct *tsk) +{ + return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED_LAZY)); +} + +static inline int need_resched_lazy(void) +{ + return test_thread_flag(TIF_NEED_RESCHED_LAZY); +} + +static inline int need_resched_now(void) +{ + return test_thread_flag(TIF_NEED_RESCHED); +} + +#else +static inline void clear_tsk_need_resched_lazy(struct task_struct *tsk) { } +static inline int need_resched_lazy(void) { return 0; } + +static inline int need_resched_now(void) +{ + return test_thread_flag(TIF_NEED_RESCHED); +} + +#endif + + +static inline bool __task_is_stopped_or_traced(struct task_struct *task) +{ + if (task->state & (__TASK_STOPPED | __TASK_TRACED)) + return true; +#ifdef CONFIG_PREEMPT_RT + if (task->saved_state & (__TASK_STOPPED | __TASK_TRACED)) + return true; +#endif + return false; +} + +static inline bool task_is_stopped_or_traced(struct task_struct *task) +{ + bool traced_stopped; + +#ifdef CONFIG_PREEMPT_RT + unsigned long flags; + + raw_spin_lock_irqsave(&task->pi_lock, flags); + traced_stopped = __task_is_stopped_or_traced(task); + raw_spin_unlock_irqrestore(&task->pi_lock, flags); +#else + traced_stopped = __task_is_stopped_or_traced(task); +#endif + return traced_stopped; +} + +static inline bool task_is_traced(struct task_struct *task) +{ + bool traced = false; + + if (task->state & __TASK_TRACED) + return true; +#ifdef CONFIG_PREEMPT_RT + /* in case the task is sleeping on tasklist_lock */ + raw_spin_lock_irq(&task->pi_lock); + if (task->state & __TASK_TRACED) + traced = true; + else if (task->saved_state & __TASK_TRACED) + traced = true; + raw_spin_unlock_irq(&task->pi_lock); +#endif + return traced; +} + /* * cond_resched() and cond_resched_lock(): latency reduction via * explicit rescheduling in places that are safe. The return @@ -1820,6 +1941,23 @@ static __always_inline bool need_resched return unlikely(tif_need_resched()); } +#ifdef CONFIG_PREEMPT_RT +static inline void sleeping_lock_inc(void) +{ + current->sleeping_lock++; +} + +static inline void sleeping_lock_dec(void) +{ + current->sleeping_lock--; +} + +#else + +static inline void sleeping_lock_inc(void) { } +static inline void sleeping_lock_dec(void) { } +#endif + /* * Wrappers for p->thread_info->cpu access. No-op on UP. */ @@ -2011,4 +2149,6 @@ int sched_trace_rq_cpu(struct rq *rq); const struct cpumask *sched_trace_rd_span(struct root_domain *rd); +extern struct task_struct *takedown_cpu_task; + #endif Index: linux-5.6.17-rt10/include/linux/sched/mm.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/sched/mm.h +++ linux-5.6.17-rt10/include/linux/sched/mm.h @@ -49,6 +49,17 @@ static inline void mmdrop(struct mm_stru __mmdrop(mm); } +#ifdef CONFIG_PREEMPT_RT +extern void __mmdrop_delayed(struct rcu_head *rhp); +static inline void mmdrop_delayed(struct mm_struct *mm) +{ + if (atomic_dec_and_test(&mm->mm_count)) + call_rcu(&mm->delayed_drop, __mmdrop_delayed); +} +#else +# define mmdrop_delayed(mm) mmdrop(mm) +#endif + /* * This has to be called after a get_task_mm()/mmget_not_zero() * followed by taking the mmap_sem for writing before modifying the Index: linux-5.6.17-rt10/include/linux/sched/wake_q.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/sched/wake_q.h +++ linux-5.6.17-rt10/include/linux/sched/wake_q.h @@ -58,6 +58,17 @@ static inline bool wake_q_empty(struct w extern void wake_q_add(struct wake_q_head *head, struct task_struct *task); extern void wake_q_add_safe(struct wake_q_head *head, struct task_struct *task); -extern void wake_up_q(struct wake_q_head *head); +extern void wake_q_add_sleeper(struct wake_q_head *head, struct task_struct *task); +extern void __wake_up_q(struct wake_q_head *head, bool sleeper); + +static inline void wake_up_q(struct wake_q_head *head) +{ + __wake_up_q(head, false); +} + +static inline void wake_up_q_sleeper(struct wake_q_head *head) +{ + __wake_up_q(head, true); +} #endif /* _LINUX_SCHED_WAKE_Q_H */ Index: linux-5.6.17-rt10/include/linux/seqlock.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/seqlock.h +++ linux-5.6.17-rt10/include/linux/seqlock.h @@ -1,36 +1,15 @@ /* SPDX-License-Identifier: GPL-2.0 */ #ifndef __LINUX_SEQLOCK_H #define __LINUX_SEQLOCK_H + /* - * Reader/writer consistent mechanism without starving writers. This type of - * lock for data where the reader wants a consistent set of information - * and is willing to retry if the information changes. There are two types - * of readers: - * 1. Sequence readers which never block a writer but they may have to retry - * if a writer is in progress by detecting change in sequence number. - * Writers do not wait for a sequence reader. - * 2. Locking readers which will wait if a writer or another locking reader - * is in progress. A locking reader in progress will also block a writer - * from going forward. Unlike the regular rwlock, the read lock here is - * exclusive so that only one locking reader can get it. - * - * This is not as cache friendly as brlock. Also, this may not work well - * for data that contains pointers, because any writer could - * invalidate a pointer that a reader was following. - * - * Expected non-blocking reader usage: - * do { - * seq = read_seqbegin(&foo); - * ... - * } while (read_seqretry(&foo, seq)); - * - * - * On non-SMP the spin locks disappear but the writer still needs - * to increment the sequence variables because an interrupt routine could - * change the state of the data. + * seqcount_t / seqlock_t - a reader-writer consistency mechanism with + * lockless readers (read-only retry loops), and no writer starvation. + * + * See Documentation/locking/seqlock.rst for full description. * - * Based on x86_64 vsyscall gettimeofday - * by Keith Owens and Andrea Arcangeli + * Copyrights: + * - Based on x86_64 vsyscall gettimeofday: Keith Owens, Andrea Arcangeli */ #include @@ -40,10 +19,28 @@ #include /* - * Version using sequence counter only. - * This can be used when code has its own mutex protecting the - * updating starting before the write_seqcountbeqin() and ending - * after the write_seqcount_end(). + * Sequence counters (seqcount_t) + * + * This is the raw counting mechanism, without any writer protection. + * + * Write side critical sections must be serialized and non-preemptible. + * + * If readers can be invoked from hardirq or softirq contexts, + * interrupts or bottom halves must also be respectively disabled before + * entering the write section. + * + * This mechanism can't be used if the protected data contains pointers, + * as the writer can invalidate a pointer that a reader is following. + * + * If the write serialization mechanism is one of the common kernel + * locking primitives, use a sequence counter with associated lock + * (seqcount_LOCKTYPE_t) instead. + * + * If it's desired to automatically handle the sequence counter writer + * serialization and non-preemptibility requirements, use a sequential + * lock (seqlock_t) instead. + * + * See Documentation/locking/seqlock.rst */ typedef struct seqcount { unsigned sequence; @@ -66,6 +63,10 @@ static inline void __seqcount_init(seqco # define SEQCOUNT_DEP_MAP_INIT(lockname) \ .dep_map = { .name = #lockname } \ +/** + * seqcount_init() - runtime initializer for seqcount_t + * @s: Pointer to the &typedef seqcount_t instance + */ # define seqcount_init(s) \ do { \ static struct lock_class_key __key; \ @@ -89,12 +90,15 @@ static inline void seqcount_lockdep_read # define seqcount_lockdep_reader_access(x) #endif -#define SEQCNT_ZERO(lockname) { .sequence = 0, SEQCOUNT_DEP_MAP_INIT(lockname)} - +/** + * SEQCNT_ZERO() - static initializer for seqcount_t + * @name: Name of the &typedef seqcount_t instance + */ +#define SEQCNT_ZERO(name) { .sequence = 0, SEQCOUNT_DEP_MAP_INIT(name) } /** - * __read_seqcount_begin - begin a seq-read critical section (without barrier) - * @s: pointer to seqcount_t + * __read_seqcount_begin() - begin a seq-read critical section (without barrier) + * @s: Pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * Returns: count to be passed to read_seqcount_retry * * __read_seqcount_begin is like read_seqcount_begin, but has no smp_rmb() @@ -105,7 +109,9 @@ static inline void seqcount_lockdep_read * Use carefully, only in critical code, and comment how the barrier is * provided. */ -static inline unsigned __read_seqcount_begin(const seqcount_t *s) +#define __read_seqcount_begin(s) do___read_seqcount_begin(s) + +static inline unsigned __read_seqcount_t_begin(const seqcount_t *s) { unsigned ret; @@ -119,15 +125,18 @@ repeat: } /** - * raw_read_seqcount - Read the raw seqcount - * @s: pointer to seqcount_t + * raw_read_seqcount() - Read the raw seqcount + * @s: Pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * Returns: count to be passed to read_seqcount_retry * * raw_read_seqcount opens a read critical section of the given - * seqcount without any lockdep checking and without checking or - * masking the LSB. Calling code is responsible for handling that. + * seqcount_t, without any lockdep checks and without checking or + * masking the sequence counter LSB. Calling code is responsible for + * handling that. */ -static inline unsigned raw_read_seqcount(const seqcount_t *s) +#define raw_read_seqcount(s) do_raw_read_seqcount(s) + +static inline unsigned raw_read_seqcount_t(const seqcount_t *s) { unsigned ret = READ_ONCE(s->sequence); smp_rmb(); @@ -135,42 +144,46 @@ static inline unsigned raw_read_seqcount } /** - * raw_read_seqcount_begin - start seq-read critical section w/o lockdep - * @s: pointer to seqcount_t + * raw_read_seqcount_begin() - start seq-read critical section w/o lockdep + * @s: Pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * Returns: count to be passed to read_seqcount_retry * * raw_read_seqcount_begin opens a read critical section of the given - * seqcount, but without any lockdep checking. Validity of the critical - * section is tested by checking read_seqcount_retry function. + * seqcount_t, but without any lockdep checking. Validity of the read + * section must be checked with read_seqcount_retry(). */ -static inline unsigned raw_read_seqcount_begin(const seqcount_t *s) +#define raw_read_seqcount_begin(s) do_raw_read_seqcount_begin(s) + +static inline unsigned raw_read_seqcount_t_begin(const seqcount_t *s) { - unsigned ret = __read_seqcount_begin(s); + unsigned ret = __read_seqcount_t_begin(s); smp_rmb(); return ret; } /** - * read_seqcount_begin - begin a seq-read critical section - * @s: pointer to seqcount_t + * read_seqcount_begin() - begin a seq-read critical section + * @s: pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * Returns: count to be passed to read_seqcount_retry * - * read_seqcount_begin opens a read critical section of the given seqcount. - * Validity of the critical section is tested by checking read_seqcount_retry - * function. + * read_seqcount_begin opens a read critical section of the given + * seqcount_t. Validity of the read section must be checked with + * read_seqcount_retry(). */ -static inline unsigned read_seqcount_begin(const seqcount_t *s) +#define read_seqcount_begin(s) do_read_seqcount_begin(s) + +static inline unsigned read_seqcount_t_begin(const seqcount_t *s) { seqcount_lockdep_reader_access(s); - return raw_read_seqcount_begin(s); + return raw_read_seqcount_t_begin(s); } /** - * raw_seqcount_begin - begin a seq-read critical section - * @s: pointer to seqcount_t + * raw_seqcount_begin() - begin a seq-read critical section + * @s: pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * Returns: count to be passed to read_seqcount_retry * - * raw_seqcount_begin opens a read critical section of the given seqcount. + * raw_seqcount_begin opens a read critical section of the given seqcount_t. * Validity of the critical section is tested by checking read_seqcount_retry * function. * @@ -179,7 +192,9 @@ static inline unsigned read_seqcount_beg * read_seqcount_retry() instead of stabilizing at the beginning of the * critical section. */ -static inline unsigned raw_seqcount_begin(const seqcount_t *s) +#define raw_seqcount_begin(s) do_raw_seqcount_begin(s) + +static inline unsigned raw_seqcount_t_begin(const seqcount_t *s) { unsigned ret = READ_ONCE(s->sequence); smp_rmb(); @@ -187,8 +202,8 @@ static inline unsigned raw_seqcount_begi } /** - * __read_seqcount_retry - end a seq-read critical section (without barrier) - * @s: pointer to seqcount_t + * __read_seqcount_retry() - end a seq-read critical section (without barrier) + * @s: pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * @start: count, from read_seqcount_begin * Returns: 1 if retry is required, else 0 * @@ -200,48 +215,54 @@ static inline unsigned raw_seqcount_begi * Use carefully, only in critical code, and comment how the barrier is * provided. */ -static inline int __read_seqcount_retry(const seqcount_t *s, unsigned start) +#define __read_seqcount_retry(s, start) do___read_seqcount_retry(s, start) + +static inline int __read_seqcount_t_retry(const seqcount_t *s, unsigned start) { return unlikely(s->sequence != start); } /** - * read_seqcount_retry - end a seq-read critical section - * @s: pointer to seqcount_t + * read_seqcount_retry() - end a seq-read critical section + * @s: pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * @start: count, from read_seqcount_begin * Returns: 1 if retry is required, else 0 * - * read_seqcount_retry closes a read critical section of the given seqcount. + * read_seqcount_retry closes a read critical section of given seqcount_t. * If the critical section was invalid, it must be ignored (and typically * retried). */ -static inline int read_seqcount_retry(const seqcount_t *s, unsigned start) +#define read_seqcount_retry(s, start) do_read_seqcount_retry(s, start) + +static inline int read_seqcount_t_retry(const seqcount_t *s, unsigned start) { smp_rmb(); - return __read_seqcount_retry(s, start); + return __read_seqcount_t_retry(s, start); } +#define raw_write_seqcount_begin(s) do_raw_write_seqcount_begin(s) - -static inline void raw_write_seqcount_begin(seqcount_t *s) +static inline void raw_write_seqcount_t_begin(seqcount_t *s) { s->sequence++; smp_wmb(); } -static inline void raw_write_seqcount_end(seqcount_t *s) +#define raw_write_seqcount_end(s) do_raw_write_seqcount_end(s) + +static inline void raw_write_seqcount_t_end(seqcount_t *s) { smp_wmb(); s->sequence++; } /** - * raw_write_seqcount_barrier - do a seq write barrier - * @s: pointer to seqcount_t + * raw_write_seqcount_barrier() - do a seq write barrier + * @s: Pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * * This can be used to provide an ordering guarantee instead of the * usual consistency guarantee. It is one wmb cheaper, because we can - * collapse the two back-to-back wmb()s. + * collapse the two back-to-back wmb()s:: * * seqcount_t seq; * bool X = true, Y = false; @@ -269,14 +290,33 @@ static inline void raw_write_seqcount_en * X = false; * } */ -static inline void raw_write_seqcount_barrier(seqcount_t *s) +#define raw_write_seqcount_barrier(s) do_raw_write_seqcount_barrier(s) + +static inline void raw_write_seqcount_t_barrier(seqcount_t *s) { s->sequence++; smp_wmb(); s->sequence++; } -static inline int raw_read_seqcount_latch(seqcount_t *s) +/** + * raw_read_seqcount_latch() - pick even or odd seqcount latch data copy + * @s: pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants + * + * Use seqcount latching to switch between two storage places with + * sequence protection to allow interruptible, preemptible, writer + * sections. + * + * Check raw_write_seqcount_latch() for more details and a full reader + * and writer usage example. + * + * Return: sequence counter. Use the lowest bit as index for picking + * which data copy to read. Full counter must then be checked with + * read_seqcount_retry(). + */ +#define raw_read_seqcount_latch(s) do_raw_read_seqcount_latch(s) + +static inline int raw_read_seqcount_t_latch(seqcount_t *s) { /* Pairs with the first smp_wmb() in raw_write_seqcount_latch() */ int seq = READ_ONCE(s->sequence); /* ^^^ */ @@ -284,8 +324,8 @@ static inline int raw_read_seqcount_latc } /** - * raw_write_seqcount_latch - redirect readers to even/odd copy - * @s: pointer to seqcount_t + * raw_write_seqcount_latch() - redirect readers to even/odd copy + * @s: pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * * The latch technique is a multiversion concurrency control method that allows * queries during non-atomic modifications. If you can guarantee queries never @@ -301,181 +341,561 @@ static inline int raw_read_seqcount_latc * Very simply put: we first modify one copy and then the other. This ensures * there is always one copy in a stable state, ready to give us an answer. * - * The basic form is a data structure like: + * The basic form is a data structure like:: * - * struct latch_struct { - * seqcount_t seq; - * struct data_struct data[2]; - * }; + * struct latch_struct { + * seqcount_t seq; + * struct data_struct data[2]; + * }; * * Where a modification, which is assumed to be externally serialized, does the - * following: + * following:: * - * void latch_modify(struct latch_struct *latch, ...) - * { - * smp_wmb(); <- Ensure that the last data[1] update is visible - * latch->seq++; - * smp_wmb(); <- Ensure that the seqcount update is visible + * void latch_modify(struct latch_struct *latch, ...) + * { + * smp_wmb(); // Ensure that the last data[1] update is visible + * latch->seq++; + * smp_wmb(); // Ensure that the seqcount update is visible * - * modify(latch->data[0], ...); + * modify(latch->data[0], ...); * - * smp_wmb(); <- Ensure that the data[0] update is visible - * latch->seq++; - * smp_wmb(); <- Ensure that the seqcount update is visible + * smp_wmb(); // Ensure that the data[0] update is visible + * latch->seq++; + * smp_wmb(); // Ensure that the seqcount update is visible * - * modify(latch->data[1], ...); - * } + * modify(latch->data[1], ...); + * } * - * The query will have a form like: + * The query will have a form like:: * - * struct entry *latch_query(struct latch_struct *latch, ...) - * { - * struct entry *entry; - * unsigned seq, idx; + * struct entry *latch_query(struct latch_struct *latch, ...) + * { + * struct entry *entry; + * unsigned seq, idx; * - * do { - * seq = raw_read_seqcount_latch(&latch->seq); + * do { + * seq = raw_read_seqcount_latch(&latch->seq); * - * idx = seq & 0x01; - * entry = data_query(latch->data[idx], ...); + * idx = seq & 0x01; + * entry = data_query(latch->data[idx], ...); * - * smp_rmb(); - * } while (seq != latch->seq); + * // read_seqcount_retry() includes necessary smp_rmb() + * } while (read_seqcount_retry(&latch->seq, seq); * - * return entry; - * } + * return entry; + * } * * So during the modification, queries are first redirected to data[1]. Then we * modify data[0]. When that is complete, we redirect queries back to data[0] * and we can modify data[1]. * - * NOTE: The non-requirement for atomic modifications does _NOT_ include - * the publishing of new entries in the case where data is a dynamic - * data structure. - * - * An iteration might start in data[0] and get suspended long enough - * to miss an entire modification sequence, once it resumes it might - * observe the new entry. + * NOTE: + * + * The non-requirement for atomic modifications does _NOT_ include + * the publishing of new entries in the case where data is a dynamic + * data structure. + * + * An iteration might start in data[0] and get suspended long enough + * to miss an entire modification sequence, once it resumes it might + * observe the new entry. * - * NOTE: When data is a dynamic data structure; one should use regular RCU - * patterns to manage the lifetimes of the objects within. + * NOTE: + * + * When data is a dynamic data structure; one should use regular RCU + * patterns to manage the lifetimes of the objects within. */ -static inline void raw_write_seqcount_latch(seqcount_t *s) +#define raw_write_seqcount_latch(s) do_raw_write_seqcount_latch(s) + +static inline void raw_write_seqcount_t_latch(seqcount_t *s) { smp_wmb(); /* prior stores before incrementing "sequence" */ s->sequence++; smp_wmb(); /* increment "sequence" before following stores */ } -/* - * Sequence counter only version assumes that callers are using their - * own mutexing. - */ -static inline void write_seqcount_begin_nested(seqcount_t *s, int subclass) +#define write_seqcount_begin_nested(s, subclass) \ + do_write_seqcount_begin_nested(s, subclass) + +static inline void write_seqcount_t_begin_nested(seqcount_t *s, int subclass) { - raw_write_seqcount_begin(s); + raw_write_seqcount_t_begin(s); seqcount_acquire(&s->dep_map, subclass, 0, _RET_IP_); } -static inline void write_seqcount_begin(seqcount_t *s) +/** + * write_seqcount_begin() - start a seqcount write-side critical section + * @s: Pointer to &typedef seqcount_t + * + * write_seqcount_begin opens a write-side critical section of the given + * seqcount. Seqcount write-side critical sections must be externally + * serialized and non-preemptible. + */ +#define write_seqcount_begin(s) do_write_seqcount_begin(s) + +static inline void write_seqcount_t_begin(seqcount_t *s) { - write_seqcount_begin_nested(s, 0); + write_seqcount_t_begin_nested(s, 0); } -static inline void write_seqcount_end(seqcount_t *s) +/** + * write_seqcount_end() - end a seqcount write-side critical section + * @s: Pointer to &typedef seqcount_t + * + * The write section must've been opened with write_seqcount_begin(). + */ +#define write_seqcount_end(s) do_write_seqcount_end(s) + +static inline void write_seqcount_t_end(seqcount_t *s) { seqcount_release(&s->dep_map, _RET_IP_); - raw_write_seqcount_end(s); + raw_write_seqcount_t_end(s); } /** - * write_seqcount_invalidate - invalidate in-progress read-side seq operations - * @s: pointer to seqcount_t + * write_seqcount_invalidate() - invalidate in-progress read-side seq operations + * @s: Pointer to &typedef seqcount_t or any of the seqcount_locktype_t variants * * After write_seqcount_invalidate, no read-side seq operations will complete * successfully and see data older than this. */ -static inline void write_seqcount_invalidate(seqcount_t *s) +#define write_seqcount_invalidate(s) do_write_seqcount_invalidate(s) + +static inline void write_seqcount_t_invalidate(seqcount_t *s) { smp_wmb(); s->sequence+=2; } +/* + * Sequence counters with associated locks (seqcount_LOCKTYPE_t) + * + * A sequence counter which associates the lock used for writer + * serialization at initialization time. This enables lockdep to validate + * that the write side critical section is properly serialized. + * + * For associated locks which do not implicitly disable preemption, + * preemption protection is enforced in the write side function. + * + * See Documentation/locking/seqlock.rst + */ + +#if defined(CONFIG_LOCKDEP) || defined(CONFIG_PREEMPT_RT) +#define SEQCOUNT_ASSOC_LOCK +#endif + +/** + * typedef seqcount_spinlock_t - sequence count with spinlock associated + * @seqcount: The real sequence counter + * @lock: Pointer to the associated spinlock + * + * A plain sequence counter with external writer synchronization by a + * spinlock. The spinlock is associated to the sequence count in the + * static initializer or init function. This enables lockdep to validate + * that the write side critical section is properly serialized. + */ +typedef struct seqcount_spinlock { + seqcount_t seqcount; +#ifdef SEQCOUNT_ASSOC_LOCK + spinlock_t *lock; +#endif +} seqcount_spinlock_t; + +#ifdef SEQCOUNT_ASSOC_LOCK + +#define SEQCOUNT_LOCKTYPE_ZERO(seq_name, assoc_lock) { \ + .seqcount = SEQCNT_ZERO(seq_name.seqcount), \ + .lock = (assoc_lock), \ +} + +/* Define as macro due to static lockdep key @ seqcount_init() */ +#define seqcount_locktype_init(s, assoc_lock) \ +do { \ + seqcount_init(&(s)->seqcount); \ + (s)->lock = (assoc_lock); \ +} while (0) + +#else /* !SEQCOUNT_ASSOC_LOCK */ + +#define SEQCOUNT_LOCKTYPE_ZERO(seq_name, assoc_lock) { \ + .seqcount = SEQCNT_ZERO(seq_name.seqcount), \ +} + +#define seqcount_locktype_init(s, assoc_lock) \ +do { \ + seqcount_init(&(s)->seqcount); \ +} while (0) + +#endif /* SEQCOUNT_ASSOC_LOCK */ + +/** + * SEQCNT_SPINLOCK_ZERO - static initializer for seqcount_spinlock_t + * @name: Name of the &typedef seqcount_spinlock_t instance + * @lock: Pointer to the associated spinlock + */ +#define SEQCNT_SPINLOCK_ZERO(name, lock) \ + SEQCOUNT_LOCKTYPE_ZERO(name, lock) + +/** + * seqcount_spinlock_init - runtime initializer for seqcount_spinlock_t + * @s: Pointer to the &typedef seqcount_spinlock_t instance + * @lock: Pointer to the associated spinlock + */ +#define seqcount_spinlock_init(s, lock) \ + seqcount_locktype_init(s, lock) + +/** + * typedef seqcount_raw_spinlock_t - sequence count with raw spinlock associated + * @seqcount: The real sequence counter + * @lock: Pointer to the associated raw spinlock + * + * A plain sequence counter with external writer synchronization by a + * raw spinlock. The raw spinlock is associated to the sequence count in + * the static initializer or init function. This enables lockdep to + * validate that the write side critical section is properly serialized. + */ +typedef struct seqcount_raw_spinlock { + seqcount_t seqcount; +#ifdef SEQCOUNT_ASSOC_LOCK + raw_spinlock_t *lock; +#endif +} seqcount_raw_spinlock_t; + +/** + * SEQCNT_RAW_SPINLOCK_ZERO - static initializer for seqcount_raw_spinlock_t + * @name: Name of the &typedef seqcount_raw_spinlock_t instance + * @lock: Pointer to the associated raw_spinlock + */ +#define SEQCNT_RAW_SPINLOCK_ZERO(name, lock) \ + SEQCOUNT_LOCKTYPE_ZERO(name, lock) + +/** + * seqcount_raw_spinlock_init - runtime initializer for seqcount_raw_spinlock_t + * @s: Pointer to the &typedef seqcount_raw_spinlock_t instance + * @lock: Pointer to the associated raw_spinlock + */ +#define seqcount_raw_spinlock_init(s, lock) \ + seqcount_locktype_init(s, lock) + +/** + * typedef seqcount_rwlock_t - sequence count with rwlock associated + * @seqcount: The real sequence counter + * @lock: Pointer to the associated rwlock + * + * A plain sequence counter with external writer synchronization by a + * rwlock. The rwlock is associated to the sequence count in the static + * initializer or init function. This enables lockdep to validate that + * the write side critical section is properly serialized. + */ +typedef struct seqcount_rwlock { + seqcount_t seqcount; +#ifdef SEQCOUNT_ASSOC_LOCK + rwlock_t *lock; +#endif +} seqcount_rwlock_t; + +/** + * SEQCNT_RWLOCK_ZERO - static initializer for seqcount_rwlock_t + * @name: Name of the &typedef seqcount_rwlock_t instance + * @lock: Pointer to the associated rwlock + */ +#define SEQCNT_RWLOCK_ZERO(name, lock) \ + SEQCOUNT_LOCKTYPE_ZERO(name, lock) + +/** + * seqcount_rwlock_init - runtime initializer for seqcount_rwlock_t + * @s: Pointer to the &typedef seqcount_rwlock_t instance + * @lock: Pointer to the associated rwlock + */ +#define seqcount_rwlock_init(s, lock) \ + seqcount_locktype_init(s, lock) + +/** + * typedef seqcount_mutex_t - sequence count with mutex associated + * @seqcount: The real sequence counter + * @lock: Pointer to the associated mutex + * + * A plain sequence counter with external writer synchronization by a + * mutex. The mutex is associated to the sequence counter in the static + * initializer or init function. This enables lockdep to validate that + * the write side critical section is properly serialized. + * + * The write side API functions write_seqcount_begin()/end() automatically + * disable and enable preemption when used with seqcount_mutex_t. + */ +typedef struct seqcount_mutex { + seqcount_t seqcount; +#ifdef SEQCOUNT_ASSOC_LOCK + struct mutex *lock; +#endif +} seqcount_mutex_t; + +/** + * SEQCNT_MUTEX_ZERO - static initializer for seqcount_mutex_t + * @name: Name of the &typedef seqcount_mutex_t instance + * @lock: Pointer to the associated mutex + */ +#define SEQCNT_MUTEX_ZERO(name, lock) \ + SEQCOUNT_LOCKTYPE_ZERO(name, lock) + +/** + * seqcount_mutex_init - runtime initializer for seqcount_mutex_t + * @s: Pointer to the &typedef seqcount_mutex_t instance + * @lock: Pointer to the associated mutex + */ +#define seqcount_mutex_init(s, lock) \ + seqcount_locktype_init(s, lock) + +/** + * typedef seqcount_ww_mutex_t - sequence count with ww_mutex associated + * @seqcount: The real sequence counter + * @lock: Pointer to the associated ww_mutex + * + * A plain sequence counter with external writer synchronization by a + * ww_mutex. The ww_mutex is associated to the sequence counter in the static + * initializer or init function. This enables lockdep to validate that + * the write side critical section is properly serialized. + * + * The write side API functions write_seqcount_begin()/end() automatically + * disable and enable preemption when used with seqcount_ww_mutex_t. + */ +typedef struct seqcount_ww_mutex { + seqcount_t seqcount; +#ifdef SEQCOUNT_ASSOC_LOCK + struct ww_mutex *lock; +#endif +} seqcount_ww_mutex_t; + +/** + * SEQCNT_WW_MUTEX_ZERO - static initializer for seqcount_ww_mutex_t + * @name: Name of the &typedef seqcount_ww_mutex_t instance + * @lock: Pointer to the associated ww_mutex + */ +#define SEQCNT_WW_MUTEX_ZERO(name, lock) \ + SEQCOUNT_LOCKTYPE_ZERO(name, lock) + +/** + * seqcount_ww_mutex_init - runtime initializer for seqcount_ww_mutex_t + * @s: Pointer to the &typedef seqcount_ww_mutex_t instance + * @lock: Pointer to the associated ww_mutex + */ +#define seqcount_ww_mutex_init(s, lock) \ + seqcount_locktype_init(s, lock) + +#include + +/* + * Sequential locks (seqlock_t) + * + * Sequence counters with an embedded spinlock for writer serialization + * and non-preemptibility. + * + * For more info, see: + * - Comments on top of seqcount_t + * - Documentation/locking/seqlock.rst + */ typedef struct { struct seqcount seqcount; spinlock_t lock; } seqlock_t; -/* - * These macros triggered gcc-3.x compile-time problems. We think these are - * OK now. Be cautious. - */ #define __SEQLOCK_UNLOCKED(lockname) \ { \ .seqcount = SEQCNT_ZERO(lockname), \ .lock = __SPIN_LOCK_UNLOCKED(lockname) \ } -#define seqlock_init(x) \ +/** + * seqlock_init() - dynamic initializer for seqlock_t + * @sl: Pointer to the &typedef seqlock_t instance + */ +#define seqlock_init(sl) \ do { \ - seqcount_init(&(x)->seqcount); \ - spin_lock_init(&(x)->lock); \ + seqcount_init(&(sl)->seqcount); \ + spin_lock_init(&(sl)->lock); \ } while (0) -#define DEFINE_SEQLOCK(x) \ - seqlock_t x = __SEQLOCK_UNLOCKED(x) +/** + * DEFINE_SEQLOCK() - Define a statically-allocated seqlock_t + * @sl: Name of the &typedef seqlock_t instance + */ +#define DEFINE_SEQLOCK(sl) \ + seqlock_t sl = __SEQLOCK_UNLOCKED(sl) + +/** + * read_seqbegin() - start a seqlock_t read-side critical section + * @sl: Pointer to &typedef seqlock_t + * + * read_seqbegin opens a read side critical section of the given + * seqlock_t. Validity of the critical section is tested by checking + * read_seqretry(). + * + * Return: count to be passed to read_seqretry() + */ /* - * Read side functions for starting and finalizing a read side section. + * For PREEMPT_RT, preemption cannot be disabled upon entering the write + * side critical section. With disabled preemption: + * + * - The writer cannot be preempted by a task with higher priority + * + * - The writer cannot acquire a spinlock_t since it's a sleeping + * lock. This would invalidate the existing, and non-PREEMPT_RT + * valid, code pattern of acquiring a spinlock_t inside the seqcount + * write side critical section. + * + * To remain preemptible, while avoiding a livelock caused by the reader + * preempting the writer, use a different technique: + * + * - If the sequence counter is even upon entering a read side + * section, then no writer is in progress, and the reader did not + * preempt any write side sections. It can continue. + * + * - If the counter is odd, a writer is in progress and the reader may + * have preempted a write side section. Let the reader acquire the + * lock used for seqcount writer serialization, which is already + * held by the writer. + * + * The higher-priority reader will block on the lock, and the + * lower-priority preempted writer will make progress until it + * finishes its write serialization lock critical section. + * + * Once the reader has the writer serialization lock acquired, the + * writer is finished and the counter is even. Drop the writer + * serialization lock and re-read the sequence counter. + * + * This technique must be implemented for all PREEMPT_RT sleeping locks. */ +#ifdef CONFIG_PREEMPT_RT + static inline unsigned read_seqbegin(const seqlock_t *sl) { - return read_seqcount_begin(&sl->seqcount); + unsigned seq; + + seqcount_lockdep_reader_access(&sl->seqcount); + + do { + seq = READ_ONCE(sl->seqcount.sequence); + if (unlikely(seq & 1)) { + seqlock_t *msl = (seqlock_t *)sl; + spin_lock(&msl->lock); + spin_unlock(&msl->lock); + } + } while (unlikely(seq & 1)); + + smp_rmb(); + return seq; +} + +#else /* !CONFIG_PREEMPT_RT */ + +static inline unsigned read_seqbegin(const seqlock_t *sl) +{ + return read_seqcount_t_begin(&sl->seqcount); } +#endif + +/** + * read_seqretry() - end and validate a seqlock_t read side section + * @sl: Pointer to &typedef seqlock_t + * @start: count, from read_seqbegin() + * + * read_seqretry closes the given seqlock_t read side critical section, + * and checks its validity. If the read section was invalid, it must be + * ignored and retried. + * + * Return: 1 if a retry is required, 0 otherwise + */ static inline unsigned read_seqretry(const seqlock_t *sl, unsigned start) { - return read_seqcount_retry(&sl->seqcount, start); + return read_seqcount_t_retry(&sl->seqcount, start); } -/* - * Lock out other writers and update the count. - * Acts like a normal spin_lock/unlock. - * Don't need preempt_disable() because that is in the spin_lock already. +/** + * write_seqlock() - start a seqlock_t write side critical section + * @sl: Pointer to &typedef seqlock_t + * + * write_seqlock opens a write side critical section of the given + * seqlock_t. It also acquires the spinlock_t embedded inside the + * sequential lock. All the seqlock_t write side critical sections are + * thus automatically serialized and non-preemptible. + * + * Use the ``_irqsave`` and ``_bh`` variants instead if the read side + * can be invoked from a hardirq or softirq context. + * + * The opened write side section must be closed with write_sequnlock(). */ static inline void write_seqlock(seqlock_t *sl) { spin_lock(&sl->lock); - write_seqcount_begin(&sl->seqcount); + write_seqcount_t_begin(&sl->seqcount); } +/** + * write_sequnlock() - end a seqlock_t write side critical section + * @sl: Pointer to &typedef seqlock_t + * + * write_sequnlock closes the (serialized and non-preemptible) write + * side critical section of given seqlock_t. + */ static inline void write_sequnlock(seqlock_t *sl) { - write_seqcount_end(&sl->seqcount); + write_seqcount_t_end(&sl->seqcount); spin_unlock(&sl->lock); } +/** + * write_seqlock_bh() - start a softirqs-disabled seqlock_t write section + * @sl: Pointer to &typedef seqlock_t + * + * ``_bh`` variant of write_seqlock(). Use only if the read side section + * can be invoked from a softirq context. + * + * The opened write section must be closed with write_sequnlock_bh(). + */ static inline void write_seqlock_bh(seqlock_t *sl) { spin_lock_bh(&sl->lock); - write_seqcount_begin(&sl->seqcount); + write_seqcount_t_begin(&sl->seqcount); } +/** + * write_sequnlock_bh() - end a softirqs-disabled seqlock_t write section + * @sl: Pointer to &typedef seqlock_t + * + * write_sequnlock_bh closes the serialized, non-preemptible, + * softirqs-disabled, seqlock_t write side critical section opened with + * write_seqlock_bh(). + */ static inline void write_sequnlock_bh(seqlock_t *sl) { - write_seqcount_end(&sl->seqcount); + write_seqcount_t_end(&sl->seqcount); spin_unlock_bh(&sl->lock); } +/** + * write_seqlock_irq() - start a non-interruptible seqlock_t write side section + * @sl: Pointer to &typedef seqlock_t + * + * This is the ``_irq`` variant of write_seqlock(). Use only if the read + * section of given seqlock_t can be invoked from a hardirq context. + */ static inline void write_seqlock_irq(seqlock_t *sl) { spin_lock_irq(&sl->lock); - write_seqcount_begin(&sl->seqcount); + write_seqcount_t_begin(&sl->seqcount); } +/** + * write_sequnlock_irq() - end a non-interruptible seqlock_t write side section + * @sl: Pointer to &typedef seqlock_t + * + * ``_irq`` variant of write_sequnlock(). The write side section of + * given seqlock_t must've been opened with write_seqlock_irq(). + */ static inline void write_sequnlock_irq(seqlock_t *sl) { - write_seqcount_end(&sl->seqcount); + write_seqcount_t_end(&sl->seqcount); spin_unlock_irq(&sl->lock); } @@ -484,44 +904,98 @@ static inline unsigned long __write_seql unsigned long flags; spin_lock_irqsave(&sl->lock, flags); - write_seqcount_begin(&sl->seqcount); + write_seqcount_t_begin(&sl->seqcount); + return flags; } +/** + * write_seqlock_irqsave() - start a non-interruptible seqlock_t write section + * @lock: Pointer to &typedef seqlock_t + * @flags: Stack-allocated storage for saving caller's local interrupt + * state, to be passed to write_sequnlock_irqrestore(). + * + * ``_irqsave`` variant of write_seqlock(). Use if the read section of + * given seqlock_t can be invoked from a hardirq context. + * + * The opened write section must be closed with write_sequnlock_irqrestore(). + */ #define write_seqlock_irqsave(lock, flags) \ do { flags = __write_seqlock_irqsave(lock); } while (0) +/** + * write_sequnlock_irqrestore() - end non-interruptible seqlock_t write section + * @sl: Pointer to &typedef seqlock_t + * @flags: Caller's saved interrupt state, from write_seqlock_irqsave() + * + * ``_irqrestore`` variant of write_sequnlock(). The write section of + * given seqlock_t must've been opened with write_seqlock_irqsave(). + */ static inline void write_sequnlock_irqrestore(seqlock_t *sl, unsigned long flags) { - write_seqcount_end(&sl->seqcount); + write_seqcount_t_end(&sl->seqcount); spin_unlock_irqrestore(&sl->lock, flags); } -/* - * A locking reader exclusively locks out other writers and locking readers, - * but doesn't update the sequence number. Acts like a normal spin_lock/unlock. - * Don't need preempt_disable() because that is in the spin_lock already. +/** + * read_seqlock_excl() - begin a seqlock_t locking reader critical section + * @sl: Pointer to &typedef seqlock_t + * + * read_seqlock_excl opens a locking reader critical section for the + * given seqlock_t. A locking reader exclusively locks out other writers + * and other *locking* readers, but doesn't update the sequence number. + * + * Locking readers act like a normal spin_lock()/spin_unlock(). + * + * The opened read side section must be closed with read_sequnlock_excl(). */ static inline void read_seqlock_excl(seqlock_t *sl) { spin_lock(&sl->lock); } +/** + * read_sequnlock_excl() - end a seqlock_t locking reader critical section + * @sl: Pointer to &typedef seqlock_t + * + * read_sequnlock_excl closes the locking reader critical section opened + * with read_seqlock_excl(). + */ static inline void read_sequnlock_excl(seqlock_t *sl) { spin_unlock(&sl->lock); } /** - * read_seqbegin_or_lock - begin a sequence number check or locking block - * @lock: sequence lock - * @seq : sequence number to be checked - * - * First try it once optimistically without taking the lock. If that fails, - * take the lock. The sequence number is also used as a marker for deciding - * whether to be a reader (even) or writer (odd). - * N.B. seq must be initialized to an even number to begin with. + * read_seqbegin_or_lock() - begin a seqlock_t lockless or locking reader + * @lock: Pointer to &typedef seqlock_t + * @seq : Marker and return parameter. If the passed value is even, the + * reader will become a *lockless* seqlock_t sequence counter reader as + * in read_seqbegin(). If the passed value is odd, the reader will + * become a fully locking reader, as in read_seqlock_excl(). In the + * first call to read_seqbegin_or_lock(), the caller **must** initialize + * and pass an even value to @seq so a lockless read is optimistically + * tried first. + * + * read_seqbegin_or_lock is an API designed to optimistically try a + * normal lockless seqlock_t read section first, as in read_seqbegin(). + * If an odd counter is found, the normal lockless read trial has + * failed, and the next reader iteration transforms to a full seqlock_t + * locking reader as in read_seqlock_excl(). + * + * This is typically used to avoid lockless seqlock_t readers starvation + * (too much retry loops) in the case of a sharp spike in write + * activity. + * + * The opened read section must be closed with done_seqretry(). Check + * Documentation/locking/seqlock.rst for template example code. + * + * Return: The encountered sequence counter value, returned through the + * @seq parameter, which is overloaded as a return parameter. The + * returned value must be checked with need_seqretry(). If the read + * section must be retried, the returned value must also be passed to + * the @seq parameter of the next read_seqbegin_or_lock() iteration. */ static inline void read_seqbegin_or_lock(seqlock_t *lock, int *seq) { @@ -531,32 +1005,90 @@ static inline void read_seqbegin_or_lock read_seqlock_excl(lock); } +/** + * need_seqretry() - validate seqlock_t "locking or lockless" reader section + * @lock: Pointer to &typedef seqlock_t + * @seq: count, from read_seqbegin_or_lock() + * + * need_seqretry checks if the seqlock_t read-side critical section + * started with read_seqbegin_or_lock() is valid. If it was not, the + * caller must retry the read-side section. + * + * Return: 1 if a retry is required, 0 otherwise + */ static inline int need_seqretry(seqlock_t *lock, int seq) { return !(seq & 1) && read_seqretry(lock, seq); } +/** + * done_seqretry() - end seqlock_t "locking or lockless" reader section + * @lock: Pointer to &typedef seqlock_t + * @seq: count, from read_seqbegin_or_lock() + * + * done_seqretry finishes the seqlock_t read side critical section + * started by read_seqbegin_or_lock(). The read section must've been + * already validated with need_seqretry(). + */ static inline void done_seqretry(seqlock_t *lock, int seq) { if (seq & 1) read_sequnlock_excl(lock); } +/** + * read_seqlock_excl_bh() - start a locking reader seqlock_t section + * with softirqs disabled + * @sl: Pointer to &typedef seqlock_t + * + * ``_bh`` variant of read_seqlock_excl(). Use this variant if the + * seqlock_t write side section, *or other read sections*, can be + * invoked from a softirq context + * + * The opened section must be closed with read_sequnlock_excl_bh(). + */ static inline void read_seqlock_excl_bh(seqlock_t *sl) { spin_lock_bh(&sl->lock); } +/** + * read_sequnlock_excl_bh() - stop a seqlock_t softirq-disabled locking + * reader section + * @sl: Pointer to &typedef seqlock_t + * + * ``_bh`` variant of read_sequnlock_excl(). The closed section must've + * been opened with read_seqlock_excl_bh(). + */ static inline void read_sequnlock_excl_bh(seqlock_t *sl) { spin_unlock_bh(&sl->lock); } +/** + * read_seqlock_excl_irq() - start a non-interruptible seqlock_t locking + * reader section + * @sl: Pointer to &typedef seqlock_t + * + * ``_irq`` variant of read_seqlock_excl(). Use this only if the + * seqlock_t write side critical section, *or other read side sections*, + * can be invoked from a hardirq context. + * + * The opened read section must be closed with read_sequnlock_excl_irq(). + */ static inline void read_seqlock_excl_irq(seqlock_t *sl) { spin_lock_irq(&sl->lock); } +/** + * read_sequnlock_excl_irq() - end an interrupts-disabled seqlock_t + * locking reader section + * @sl: Pointer to &typedef seqlock_t + * + * ``_irq`` variant of read_sequnlock_excl(). The closed section must've + * been opened with read_seqlock_excl_irq(). + */ static inline void read_sequnlock_excl_irq(seqlock_t *sl) { spin_unlock_irq(&sl->lock); @@ -570,15 +1102,59 @@ static inline unsigned long __read_seqlo return flags; } +/** + * read_seqlock_excl_irqsave() - start a non-interruptible seqlock_t + * locking reader section + * @lock: Pointer to &typedef seqlock_t + * @flags: Stack-allocated storage for saving caller's local interrupt + * state, to be passed to read_sequnlock_excl_irqrestore(). + * + * ``_irqsave`` variant of read_seqlock_excl(). Use this only if the + * seqlock_t write side critical section, *or other read side sections*, + * can be invoked from a hardirq context. + * + * Opened section must be closed with read_sequnlock_excl_irqrestore(). + */ #define read_seqlock_excl_irqsave(lock, flags) \ do { flags = __read_seqlock_excl_irqsave(lock); } while (0) +/** + * read_sequnlock_excl_irqrestore() - end non-interruptible seqlock_t + * locking reader section + * @sl: Pointer to &typedef seqlock_t + * @flags: Caller's saved interrupt state, from + * read_seqlock_excl_irqsave() + * + * ``_irqrestore`` variant of read_sequnlock_excl(). The closed section + * must've been opened with read_seqlock_excl_irqsave(). + */ static inline void read_sequnlock_excl_irqrestore(seqlock_t *sl, unsigned long flags) { spin_unlock_irqrestore(&sl->lock, flags); } +/** + * read_seqbegin_or_lock_irqsave() - begin a seqlock_t lockless reader, or + * a non-interruptible locking reader + * @lock: Pointer to &typedef seqlock_t + * @seq: Marker and return parameter. Check read_seqbegin_or_lock(). + * + * This is the ``_irqsave`` variant of read_seqbegin_or_lock(). Use if + * the seqlock_t write side critical section, *or other read side sections*, + * can be invoked from hardirq context. + * + * The validity of the read section must be checked with need_seqretry(). + * The opened section must be closed with done_seqretry_irqrestore(). + * + * Return: + * + * 1. The saved local interrupts state in case of a locking reader, to be + * passed to done_seqretry_irqrestore(). + * + * 2. The encountered sequence counter value, returned through @seq which + * is overloaded as a return parameter. Check read_seqbegin_or_lock(). + */ static inline unsigned long read_seqbegin_or_lock_irqsave(seqlock_t *lock, int *seq) { @@ -592,6 +1168,18 @@ read_seqbegin_or_lock_irqsave(seqlock_t return flags; } +/** + * done_seqretry_irqrestore() - end a seqlock_t lockless reader, or a + * non-interruptible locking reader section + * @lock: Pointer to &typedef seqlock_t + * @seq: Count, from read_seqbegin_or_lock_irqsave() + * @flags: Caller's saved local interrupt state in case of a locking + * reader, also from read_seqbegin_or_lock_irqsave() + * + * This is the ``_irqrestore`` variant of done_seqretry(). The read + * section must've been opened with read_seqbegin_or_lock_irqsave(), and + * validated with need_seqretry(). + */ static inline void done_seqretry_irqrestore(seqlock_t *lock, int seq, unsigned long flags) { Index: linux-5.6.17-rt10/include/linux/seqlock_types_internal.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/seqlock_types_internal.h @@ -0,0 +1,309 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __LINUX_SEQLOCK_TYPES_INTERNAL_H +#define __LINUX_SEQLOCK_TYPES_INTERNAL_H + +/* + * Sequence counters with associated locks + * + * Copyright (C) 2020 Linutronix GmbH + */ + +#ifndef __LINUX_SEQLOCK_H +#error This is an INTERNAL header; it must only be included by seqlock.h +#endif + +#include +#include +#include + +/* + * @s: pointer to seqcount_t or any of the seqcount_locktype_t variants + */ +#define __to_seqcount_t(s) \ +({ \ + seqcount_t *seq; \ + \ + if (__same_type(*(s), seqcount_t)) \ + seq = (seqcount_t *)(s); \ + else if (__same_type(*(s), seqcount_spinlock_t)) \ + seq = &((seqcount_spinlock_t *)(s))->seqcount; \ + else if (__same_type(*(s), seqcount_raw_spinlock_t)) \ + seq = &((seqcount_raw_spinlock_t *)(s))->seqcount; \ + else if (__same_type(*(s), seqcount_rwlock_t)) \ + seq = &((seqcount_rwlock_t *)(s))->seqcount; \ + else if (__same_type(*(s), seqcount_mutex_t)) \ + seq = &((seqcount_mutex_t *)(s))->seqcount; \ + else if (__same_type(*(s), seqcount_ww_mutex_t)) \ + seq = &((seqcount_ww_mutex_t *)(s))->seqcount; \ + else \ + BUILD_BUG_ON_MSG(1, "Unknown seqcount type"); \ + \ + seq; \ +}) + +/* + * seqcount_LOCKTYPE_t -- write APIs + * + * For associated lock types which do not implicitly disable preemption, + * enforce preemption protection in the write side functions. + * + * Never use lockdep for the raw write variants. + */ + +#ifdef CONFIG_PREEMPT_RT + +/* + * Do not disable preemption for PREEMPT_RT. Check comment on top of + * seqlock.h read_seqbegin() for rationale. + */ +#define __enforce_preemption_protection(s) (false) + +#else + +#define __associated_lock_is_preemptible(s) \ +({ \ + bool ret; \ + \ + if (__same_type(*(s), seqcount_t) || \ + __same_type(*(s), seqcount_spinlock_t) || \ + __same_type(*(s), seqcount_raw_spinlock_t) || \ + __same_type(*(s), seqcount_rwlock_t)) { \ + ret = false; \ + } else if (__same_type(*(s), seqcount_mutex_t) || \ + __same_type(*(s), seqcount_ww_mutex_t)) { \ + ret = true; \ + } else \ + BUILD_BUG_ON_MSG(1, "Unknown seqcount type"); \ + \ + ret; \ +}) + +#define __enforce_preemption_protection(s) \ + __associated_lock_is_preemptible(s) + +#endif /* CONFIG_PREEMPT_RT */ + +#ifdef CONFIG_LOCKDEP + +#define __assert_associated_lock_held(s) \ +do { \ + if (__same_type(*(s), seqcount_t)) \ + break; \ + \ + if (__same_type(*(s), seqcount_spinlock_t)) \ + lockdep_assert_held(((seqcount_spinlock_t *)(s))->lock);\ + else if (__same_type(*(s), seqcount_raw_spinlock_t)) \ + lockdep_assert_held(((seqcount_raw_spinlock_t *)(s))->lock); \ + else if (__same_type(*(s), seqcount_rwlock_t)) \ + lockdep_assert_held_write(((seqcount_rwlock_t *)(s))->lock); \ + else if (__same_type(*(s), seqcount_mutex_t)) \ + lockdep_assert_held(((seqcount_mutex_t *)(s))->lock); \ + else if (__same_type(*(s), seqcount_ww_mutex_t)) \ + lockdep_assert_held(&((seqcount_ww_mutex_t *)(s))->lock->base); \ + else \ + BUILD_BUG_ON_MSG(1, "Unknown seqcount type"); \ +} while (0) + +#else + +#define __assert_associated_lock_held(s) \ +do { \ + (void) __to_seqcount_t(s); \ +} while (0) + +#endif /* CONFIG_LOCKDEP */ + +#define do_raw_write_seqcount_begin(s) \ +do { \ + if (__enforce_preemption_protection(s)) \ + preempt_disable(); \ + \ + raw_write_seqcount_t_begin(__to_seqcount_t(s)); \ +} while (0) + +#define do_raw_write_seqcount_end(s) \ +do { \ + raw_write_seqcount_t_end(__to_seqcount_t(s)); \ + \ + if (__enforce_preemption_protection(s)) \ + preempt_enable(); \ +} while (0) + +#define do_write_seqcount_begin_nested(s, subclass) \ +do { \ + __assert_associated_lock_held(s); \ + \ + if (__enforce_preemption_protection(s)) \ + preempt_disable(); \ + \ + write_seqcount_t_begin_nested(__to_seqcount_t(s), subclass); \ +} while (0) + +#define do_write_seqcount_begin(s) \ +do { \ + __assert_associated_lock_held(s); \ + \ + if (__enforce_preemption_protection(s)) \ + preempt_disable(); \ + \ + write_seqcount_t_begin(__to_seqcount_t(s)); \ +} while (0) + +#define do_write_seqcount_end(s) \ +do { \ + write_seqcount_t_end(__to_seqcount_t(s)); \ + \ + if (__enforce_preemption_protection(s)) \ + preempt_enable(); \ +} while (0) + +#define do_write_seqcount_invalidate(s) \ + write_seqcount_t_invalidate(__to_seqcount_t(s)) + +#define do_raw_write_seqcount_barrier(s) \ + raw_write_seqcount_t_barrier(__to_seqcount_t(s)) + +/* + * Latch sequence counters write side critical sections don't need to + * run with preemption disabled. Check @raw_write_seqcount_latch(). + */ +#define do_raw_write_seqcount_latch(s) \ + raw_write_seqcount_t_latch(__to_seqcount_t(s)) + +/* + * seqcount_LOCKTYPE_t -- read APIs + */ + +#ifdef CONFIG_PREEMPT_RT + +/* + * Check comment on top of read_seqbegin() for rationale. + * + * @s: pointer to seqcount_t or any of the seqcount_locktype_t variants + */ +#define __rt_lock_unlock_associated_sleeping_lock(s) \ +do { \ + if (__same_type(*(s), seqcount_t) || \ + __same_type(*(s), seqcount_raw_spinlock_t)) { \ + break; /* NOP */ \ + } \ + \ + if (__same_type(*(s), seqcount_spinlock_t)) { \ + spin_lock(((seqcount_spinlock_t *) s)->lock); \ + spin_unlock(((seqcount_spinlock_t *) s)->lock); \ + } else if (__same_type(*(s), seqcount_rwlock_t)) { \ + read_lock(((seqcount_rwlock_t *) s)->lock); \ + read_unlock(((seqcount_rwlock_t *) s)->lock); \ + } else if (__same_type(*(s), seqcount_mutex_t)) { \ + mutex_lock(((seqcount_mutex_t *) s)->lock); \ + mutex_unlock(((seqcount_mutex_t *) s)->lock); \ + } else if (__same_type(*(s), seqcount_ww_mutex_t)) { \ + ww_mutex_lock(((seqcount_ww_mutex_t *) s)->lock, NULL); \ + ww_mutex_unlock(((seqcount_ww_mutex_t *) s)->lock); \ + } else \ + BUILD_BUG_ON_MSG(1, "Unknown seqcount type"); \ +} while (0) + +/* + * @s: pointer to seqcount_t or any of the seqcount_locktype_t variants + * + * After the lock-unlock operation, re-read the sequence counter since + * the writer made progress. + * + * Do not lock-unlock the seqcount associated sleeping lock again if the + * second counter read value is odd. If the first counter read was odd + * because the reader preempted the write-side critical section, the + * second odd value read must've been the result of a writer running on + * a parallel core instead. + */ +#define __raw_read_seqcount(s) \ +({ \ + unsigned seq = READ_ONCE(__to_seqcount_t(s)->sequence); \ + \ + if (unlikely(seq & 1)) \ + __rt_lock_unlock_associated_sleeping_lock(s); \ + \ + /* no read barrier, no counter stabilization, no lockdep */ \ + READ_ONCE(__to_seqcount_t(s)->sequence); \ +}) + +#define do___read_seqcount_begin(s) \ +({ \ + unsigned seq; \ + \ + do { \ + seq = __raw_read_seqcount(s); \ + cpu_relax(); \ + } while (unlikely(seq & 1)); \ + \ + /* no read barrier, with stabilized counter, no lockdep */ \ + seq; \ +}) + +#define do_raw_read_seqcount(s) \ +({ \ + unsigned seq = __raw_read_seqcount(s); \ + \ + smp_rmb(); \ + \ + /* with read barrier, no counter stabilization, no lockdep */ \ + seq; \ +}) + +#define do_raw_seqcount_begin(s) \ +({ \ + /* with read barrier, no counter stabilization, no lockdep */ \ + (do_raw_read_seqcount(s) & ~1); \ +}) + +#define do_raw_read_seqcount_begin(s) \ +({ \ + unsigned seq = do___read_seqcount_begin(s); \ + \ + smp_rmb(); \ + \ + /* with read barrier, with stabilized counter, no lockdep */ \ + seq; \ +}) + +#define do_read_seqcount_begin(s) \ +({ \ + seqcount_lockdep_reader_access(__to_seqcount_t(s)); \ + \ + /* with read barrier, stabilized counter, and lockdep */ \ + do_raw_read_seqcount_begin(s); \ +}) + +#else /* !CONFIG_PREEMPT_RT */ + +#define do___read_seqcount_begin(s) \ + __read_seqcount_t_begin(__to_seqcount_t(s)) + +#define do_raw_read_seqcount(s) \ + raw_read_seqcount_t(__to_seqcount_t(s)) + +#define do_raw_seqcount_begin(s) \ + raw_seqcount_t_begin(__to_seqcount_t(s)) + +#define do_raw_read_seqcount_begin(s) \ + raw_read_seqcount_t_begin(__to_seqcount_t(s)) + +#define do_read_seqcount_begin(s) \ + read_seqcount_t_begin(__to_seqcount_t(s)) + +#endif /* CONFIG_PREEMPT_RT */ + +/* + * Latch sequence counters allows interruptible, preemptible, writer + * sections. There is no need for a special PREEMPT_RT implementation. + */ +#define do_raw_read_seqcount_latch(s) \ + raw_read_seqcount_t_latch(__to_seqcount_t(s)) + +#define do___read_seqcount_retry(s, start) \ + __read_seqcount_t_retry(__to_seqcount_t(s), start) + +#define do_read_seqcount_retry(s, start) \ + read_seqcount_t_retry(__to_seqcount_t(s), start) + +#endif /* __LINUX_SEQLOCK_TYPES_INTERNAL_H */ Index: linux-5.6.17-rt10/include/linux/serial_8250.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/serial_8250.h +++ linux-5.6.17-rt10/include/linux/serial_8250.h @@ -7,6 +7,7 @@ #ifndef _LINUX_SERIAL_8250_H #define _LINUX_SERIAL_8250_H +#include #include #include #include @@ -124,6 +125,8 @@ struct uart_8250_port { #define MSR_SAVE_FLAGS UART_MSR_ANY_DELTA unsigned char msr_saved_flags; + atomic_t console_printing; + struct uart_8250_dma *dma; const struct uart_8250_ops *ops; @@ -175,6 +178,8 @@ void serial8250_init_port(struct uart_82 void serial8250_set_defaults(struct uart_8250_port *up); void serial8250_console_write(struct uart_8250_port *up, const char *s, unsigned int count); +void serial8250_console_write_atomic(struct uart_8250_port *up, const char *s, + unsigned int count); int serial8250_console_setup(struct uart_port *port, char *options, bool probe); extern void serial8250_set_isa_configurator(void (*v) Index: linux-5.6.17-rt10/include/linux/signal.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/signal.h +++ linux-5.6.17-rt10/include/linux/signal.h @@ -255,6 +255,7 @@ static inline void init_sigpending(struc } extern void flush_sigqueue(struct sigpending *queue); +extern void flush_task_sigqueue(struct task_struct *tsk); /* Test if 'sig' is valid signal. Use this instead of testing _NSIG directly */ static inline int valid_signal(unsigned long sig) Index: linux-5.6.17-rt10/include/linux/skbuff.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/skbuff.h +++ linux-5.6.17-rt10/include/linux/skbuff.h @@ -293,6 +293,7 @@ struct sk_buff_head { __u32 qlen; spinlock_t lock; + raw_spinlock_t raw_lock; }; struct sk_buff; @@ -1895,6 +1896,12 @@ static inline void skb_queue_head_init(s __skb_queue_head_init(list); } +static inline void skb_queue_head_init_raw(struct sk_buff_head *list) +{ + raw_spin_lock_init(&list->raw_lock); + __skb_queue_head_init(list); +} + static inline void skb_queue_head_init_class(struct sk_buff_head *list, struct lock_class_key *class) { Index: linux-5.6.17-rt10/include/linux/smp.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/smp.h +++ linux-5.6.17-rt10/include/linux/smp.h @@ -221,6 +221,9 @@ static inline int get_boot_cpu_id(void) #define get_cpu() ({ preempt_disable(); __smp_processor_id(); }) #define put_cpu() preempt_enable() +#define get_cpu_light() ({ migrate_disable(); __smp_processor_id(); }) +#define put_cpu_light() migrate_enable() + /* * Callback to arch code if there's nosmp or maxcpus=0 on the * boot command line: Index: linux-5.6.17-rt10/include/linux/spinlock.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/spinlock.h +++ linux-5.6.17-rt10/include/linux/spinlock.h @@ -307,7 +307,11 @@ static inline void do_raw_spin_unlock(ra }) /* Include rwlock functions */ -#include +#ifdef CONFIG_PREEMPT_RT +# include +#else +# include +#endif /* * Pull the _spin_*()/_read_*()/_write_*() functions/declarations: @@ -318,6 +322,10 @@ static inline void do_raw_spin_unlock(ra # include #endif +#ifdef CONFIG_PREEMPT_RT +# include +#else /* PREEMPT_RT */ + /* * Map the spin_lock functions to the raw variants for PREEMPT_RT=n */ @@ -438,6 +446,8 @@ static __always_inline int spin_is_conte #define assert_spin_locked(lock) assert_raw_spin_locked(&(lock)->rlock) +#endif /* !PREEMPT_RT */ + /* * Pull the atomic_t declaration: * (asm-mips/atomic.h needs above definitions) Index: linux-5.6.17-rt10/include/linux/spinlock_api_smp.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/spinlock_api_smp.h +++ linux-5.6.17-rt10/include/linux/spinlock_api_smp.h @@ -187,6 +187,8 @@ static inline int __raw_spin_trylock_bh( return 0; } -#include +#ifndef CONFIG_PREEMPT_RT +# include +#endif #endif /* __LINUX_SPINLOCK_API_SMP_H */ Index: linux-5.6.17-rt10/include/linux/spinlock_rt.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/spinlock_rt.h @@ -0,0 +1,154 @@ +#ifndef __LINUX_SPINLOCK_RT_H +#define __LINUX_SPINLOCK_RT_H + +#ifndef __LINUX_SPINLOCK_H +#error Do not include directly. Use spinlock.h +#endif + +#include + +extern void +__rt_spin_lock_init(spinlock_t *lock, const char *name, struct lock_class_key *key); + +#define spin_lock_init(slock) \ +do { \ + static struct lock_class_key __key; \ + \ + rt_mutex_init(&(slock)->lock); \ + __rt_spin_lock_init(slock, #slock, &__key); \ +} while (0) + +extern void __lockfunc rt_spin_lock(spinlock_t *lock); +extern unsigned long __lockfunc rt_spin_lock_trace_flags(spinlock_t *lock); +extern void __lockfunc rt_spin_lock_nested(spinlock_t *lock, int subclass); +extern void __lockfunc rt_spin_unlock(spinlock_t *lock); +extern void __lockfunc rt_spin_lock_unlock(spinlock_t *lock); +extern int __lockfunc rt_spin_trylock_irqsave(spinlock_t *lock, unsigned long *flags); +extern int __lockfunc rt_spin_trylock_bh(spinlock_t *lock); +extern int __lockfunc rt_spin_trylock(spinlock_t *lock); +extern int atomic_dec_and_spin_lock(atomic_t *atomic, spinlock_t *lock); + +/* + * lockdep-less calls, for derived types like rwlock: + * (for trylock they can use rt_mutex_trylock() directly. + * Migrate disable handling must be done at the call site. + */ +extern void __lockfunc __rt_spin_lock(struct rt_mutex *lock); +extern void __lockfunc __rt_spin_trylock(struct rt_mutex *lock); +extern void __lockfunc __rt_spin_unlock(struct rt_mutex *lock); + +#define spin_lock(lock) rt_spin_lock(lock) + +#define spin_lock_bh(lock) \ + do { \ + local_bh_disable(); \ + rt_spin_lock(lock); \ + } while (0) + +#define spin_lock_irq(lock) spin_lock(lock) + +#define spin_do_trylock(lock) __cond_lock(lock, rt_spin_trylock(lock)) + +#define spin_trylock(lock) \ +({ \ + int __locked; \ + __locked = spin_do_trylock(lock); \ + __locked; \ +}) + +#ifdef CONFIG_LOCKDEP +# define spin_lock_nested(lock, subclass) \ + do { \ + rt_spin_lock_nested(lock, subclass); \ + } while (0) + +#define spin_lock_bh_nested(lock, subclass) \ + do { \ + local_bh_disable(); \ + rt_spin_lock_nested(lock, subclass); \ + } while (0) + +# define spin_lock_irqsave_nested(lock, flags, subclass) \ + do { \ + typecheck(unsigned long, flags); \ + flags = 0; \ + rt_spin_lock_nested(lock, subclass); \ + } while (0) +#else +# define spin_lock_nested(lock, subclass) spin_lock(lock) +# define spin_lock_bh_nested(lock, subclass) spin_lock_bh(lock) + +# define spin_lock_irqsave_nested(lock, flags, subclass) \ + do { \ + typecheck(unsigned long, flags); \ + flags = 0; \ + spin_lock(lock); \ + } while (0) +#endif + +#define spin_lock_irqsave(lock, flags) \ + do { \ + typecheck(unsigned long, flags); \ + flags = 0; \ + spin_lock(lock); \ + } while (0) + +static inline unsigned long spin_lock_trace_flags(spinlock_t *lock) +{ + unsigned long flags = 0; +#ifdef CONFIG_TRACE_IRQFLAGS + flags = rt_spin_lock_trace_flags(lock); +#else + spin_lock(lock); /* lock_local */ +#endif + return flags; +} + +/* FIXME: we need rt_spin_lock_nest_lock */ +#define spin_lock_nest_lock(lock, nest_lock) spin_lock_nested(lock, 0) + +#define spin_unlock(lock) rt_spin_unlock(lock) + +#define spin_unlock_bh(lock) \ + do { \ + rt_spin_unlock(lock); \ + local_bh_enable(); \ + } while (0) + +#define spin_unlock_irq(lock) spin_unlock(lock) + +#define spin_unlock_irqrestore(lock, flags) \ + do { \ + typecheck(unsigned long, flags); \ + (void) flags; \ + spin_unlock(lock); \ + } while (0) + +#define spin_trylock_bh(lock) __cond_lock(lock, rt_spin_trylock_bh(lock)) +#define spin_trylock_irq(lock) spin_trylock(lock) + +#define spin_trylock_irqsave(lock, flags) \ + rt_spin_trylock_irqsave(lock, &(flags)) + +#ifdef CONFIG_GENERIC_LOCKBREAK +# define spin_is_contended(lock) ((lock)->break_lock) +#else +# define spin_is_contended(lock) (((void)(lock), 0)) +#endif + +static inline int spin_can_lock(spinlock_t *lock) +{ + return !rt_mutex_is_locked(&lock->lock); +} + +static inline int spin_is_locked(spinlock_t *lock) +{ + return rt_mutex_is_locked(&lock->lock); +} + +static inline void assert_spin_locked(spinlock_t *lock) +{ + BUG_ON(!spin_is_locked(lock)); +} + +#endif Index: linux-5.6.17-rt10/include/linux/spinlock_types.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/spinlock_types.h +++ linux-5.6.17-rt10/include/linux/spinlock_types.h @@ -9,77 +9,15 @@ * Released under the General Public License (GPL). */ -#if defined(CONFIG_SMP) -# include -#else -# include -#endif - -#include - -typedef struct raw_spinlock { - arch_spinlock_t raw_lock; -#ifdef CONFIG_DEBUG_SPINLOCK - unsigned int magic, owner_cpu; - void *owner; -#endif -#ifdef CONFIG_DEBUG_LOCK_ALLOC - struct lockdep_map dep_map; -#endif -} raw_spinlock_t; - -#define SPINLOCK_MAGIC 0xdead4ead - -#define SPINLOCK_OWNER_INIT ((void *)-1L) - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define SPIN_DEP_MAP_INIT(lockname) .dep_map = { .name = #lockname } -#else -# define SPIN_DEP_MAP_INIT(lockname) -#endif +#include -#ifdef CONFIG_DEBUG_SPINLOCK -# define SPIN_DEBUG_INIT(lockname) \ - .magic = SPINLOCK_MAGIC, \ - .owner_cpu = -1, \ - .owner = SPINLOCK_OWNER_INIT, +#ifndef CONFIG_PREEMPT_RT +# include +# include #else -# define SPIN_DEBUG_INIT(lockname) +# include +# include +# include #endif -#define __RAW_SPIN_LOCK_INITIALIZER(lockname) \ - { \ - .raw_lock = __ARCH_SPIN_LOCK_UNLOCKED, \ - SPIN_DEBUG_INIT(lockname) \ - SPIN_DEP_MAP_INIT(lockname) } - -#define __RAW_SPIN_LOCK_UNLOCKED(lockname) \ - (raw_spinlock_t) __RAW_SPIN_LOCK_INITIALIZER(lockname) - -#define DEFINE_RAW_SPINLOCK(x) raw_spinlock_t x = __RAW_SPIN_LOCK_UNLOCKED(x) - -typedef struct spinlock { - union { - struct raw_spinlock rlock; - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define LOCK_PADSIZE (offsetof(struct raw_spinlock, dep_map)) - struct { - u8 __padding[LOCK_PADSIZE]; - struct lockdep_map dep_map; - }; -#endif - }; -} spinlock_t; - -#define __SPIN_LOCK_INITIALIZER(lockname) \ - { { .rlock = __RAW_SPIN_LOCK_INITIALIZER(lockname) } } - -#define __SPIN_LOCK_UNLOCKED(lockname) \ - (spinlock_t ) __SPIN_LOCK_INITIALIZER(lockname) - -#define DEFINE_SPINLOCK(x) spinlock_t x = __SPIN_LOCK_UNLOCKED(x) - -#include - #endif /* __LINUX_SPINLOCK_TYPES_H */ Index: linux-5.6.17-rt10/include/linux/spinlock_types_nort.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/spinlock_types_nort.h @@ -0,0 +1,33 @@ +#ifndef __LINUX_SPINLOCK_TYPES_NORT_H +#define __LINUX_SPINLOCK_TYPES_NORT_H + +#ifndef __LINUX_SPINLOCK_TYPES_H +#error "Do not include directly. Include spinlock_types.h instead" +#endif + +/* + * The non RT version maps spinlocks to raw_spinlocks + */ +typedef struct spinlock { + union { + struct raw_spinlock rlock; + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +# define LOCK_PADSIZE (offsetof(struct raw_spinlock, dep_map)) + struct { + u8 __padding[LOCK_PADSIZE]; + struct lockdep_map dep_map; + }; +#endif + }; +} spinlock_t; + +#define __SPIN_LOCK_INITIALIZER(lockname) \ + { { .rlock = __RAW_SPIN_LOCK_INITIALIZER(lockname) } } + +#define __SPIN_LOCK_UNLOCKED(lockname) \ + (spinlock_t ) __SPIN_LOCK_INITIALIZER(lockname) + +#define DEFINE_SPINLOCK(x) spinlock_t x = __SPIN_LOCK_UNLOCKED(x) + +#endif Index: linux-5.6.17-rt10/include/linux/spinlock_types_raw.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/spinlock_types_raw.h @@ -0,0 +1,55 @@ +#ifndef __LINUX_SPINLOCK_TYPES_RAW_H +#define __LINUX_SPINLOCK_TYPES_RAW_H + +#include + +#if defined(CONFIG_SMP) +# include +#else +# include +#endif + +#include + +typedef struct raw_spinlock { + arch_spinlock_t raw_lock; +#ifdef CONFIG_DEBUG_SPINLOCK + unsigned int magic, owner_cpu; + void *owner; +#endif +#ifdef CONFIG_DEBUG_LOCK_ALLOC + struct lockdep_map dep_map; +#endif +} raw_spinlock_t; + +#define SPINLOCK_MAGIC 0xdead4ead + +#define SPINLOCK_OWNER_INIT ((void *)-1L) + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +# define SPIN_DEP_MAP_INIT(lockname) .dep_map = { .name = #lockname } +#else +# define SPIN_DEP_MAP_INIT(lockname) +#endif + +#ifdef CONFIG_DEBUG_SPINLOCK +# define SPIN_DEBUG_INIT(lockname) \ + .magic = SPINLOCK_MAGIC, \ + .owner_cpu = -1, \ + .owner = SPINLOCK_OWNER_INIT, +#else +# define SPIN_DEBUG_INIT(lockname) +#endif + +#define __RAW_SPIN_LOCK_INITIALIZER(lockname) \ + { \ + .raw_lock = __ARCH_SPIN_LOCK_UNLOCKED, \ + SPIN_DEBUG_INIT(lockname) \ + SPIN_DEP_MAP_INIT(lockname) } + +#define __RAW_SPIN_LOCK_UNLOCKED(lockname) \ + (raw_spinlock_t) __RAW_SPIN_LOCK_INITIALIZER(lockname) + +#define DEFINE_RAW_SPINLOCK(x) raw_spinlock_t x = __RAW_SPIN_LOCK_UNLOCKED(x) + +#endif Index: linux-5.6.17-rt10/include/linux/spinlock_types_rt.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/linux/spinlock_types_rt.h @@ -0,0 +1,48 @@ +#ifndef __LINUX_SPINLOCK_TYPES_RT_H +#define __LINUX_SPINLOCK_TYPES_RT_H + +#ifndef __LINUX_SPINLOCK_TYPES_H +#error "Do not include directly. Include spinlock_types.h instead" +#endif + +#include + +/* + * PREEMPT_RT: spinlocks - an RT mutex plus lock-break field: + */ +typedef struct spinlock { + struct rt_mutex lock; + unsigned int break_lock; +#ifdef CONFIG_DEBUG_LOCK_ALLOC + struct lockdep_map dep_map; +#endif +} spinlock_t; + +#ifdef CONFIG_DEBUG_RT_MUTEXES +# define __RT_SPIN_INITIALIZER(name) \ + { \ + .wait_lock = __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock), \ + .save_state = 1, \ + .file = __FILE__, \ + .line = __LINE__ , \ + } +#else +# define __RT_SPIN_INITIALIZER(name) \ + { \ + .wait_lock = __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock), \ + .save_state = 1, \ + } +#endif + +/* +.wait_list = PLIST_HEAD_INIT_RAW((name).lock.wait_list, (name).lock.wait_lock) +*/ + +#define __SPIN_LOCK_UNLOCKED(name) \ + { .lock = __RT_SPIN_INITIALIZER(name.lock), \ + SPIN_DEP_MAP_INIT(name) } + +#define DEFINE_SPINLOCK(name) \ + spinlock_t name = __SPIN_LOCK_UNLOCKED(name) + +#endif Index: linux-5.6.17-rt10/include/linux/spinlock_types_up.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/spinlock_types_up.h +++ linux-5.6.17-rt10/include/linux/spinlock_types_up.h @@ -1,10 +1,6 @@ #ifndef __LINUX_SPINLOCK_TYPES_UP_H #define __LINUX_SPINLOCK_TYPES_UP_H -#ifndef __LINUX_SPINLOCK_TYPES_H -# error "please don't include this file directly" -#endif - /* * include/linux/spinlock_types_up.h - spinlock type definitions for UP * Index: linux-5.6.17-rt10/include/linux/stop_machine.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/stop_machine.h +++ linux-5.6.17-rt10/include/linux/stop_machine.h @@ -26,6 +26,8 @@ struct cpu_stop_work { cpu_stop_fn_t fn; void *arg; struct cpu_stop_done *done; + /* Did not run due to disabled stopper; for nowait debug checks */ + bool disabled; }; int stop_one_cpu(unsigned int cpu, cpu_stop_fn_t fn, void *arg); Index: linux-5.6.17-rt10/include/linux/swait.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/swait.h +++ linux-5.6.17-rt10/include/linux/swait.h @@ -160,7 +160,9 @@ static inline bool swq_has_sleeper(struc extern void swake_up_one(struct swait_queue_head *q); extern void swake_up_all(struct swait_queue_head *q); extern void swake_up_locked(struct swait_queue_head *q); +extern void swake_up_all_locked(struct swait_queue_head *q); +extern void __prepare_to_swait(struct swait_queue_head *q, struct swait_queue *wait); extern void prepare_to_swait_exclusive(struct swait_queue_head *q, struct swait_queue *wait, int state); extern long prepare_to_swait_event(struct swait_queue_head *q, struct swait_queue *wait, int state); Index: linux-5.6.17-rt10/include/linux/swap.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/swap.h +++ linux-5.6.17-rt10/include/linux/swap.h @@ -12,6 +12,7 @@ #include #include #include +#include #include struct notifier_block; @@ -328,6 +329,7 @@ extern unsigned long nr_free_pagecache_p /* linux/mm/swap.c */ +DECLARE_LOCAL_IRQ_LOCK(swapvec_lock); extern void lru_cache_add(struct page *); extern void lru_cache_add_anon(struct page *page); extern void lru_cache_add_file(struct page *page); Index: linux-5.6.17-rt10/include/linux/thread_info.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/thread_info.h +++ linux-5.6.17-rt10/include/linux/thread_info.h @@ -97,7 +97,17 @@ static inline int test_ti_thread_flag(st #define test_thread_flag(flag) \ test_ti_thread_flag(current_thread_info(), flag) -#define tif_need_resched() test_thread_flag(TIF_NEED_RESCHED) +#ifdef CONFIG_PREEMPT_LAZY +#define tif_need_resched() (test_thread_flag(TIF_NEED_RESCHED) || \ + test_thread_flag(TIF_NEED_RESCHED_LAZY)) +#define tif_need_resched_now() (test_thread_flag(TIF_NEED_RESCHED)) +#define tif_need_resched_lazy() test_thread_flag(TIF_NEED_RESCHED_LAZY)) + +#else +#define tif_need_resched() test_thread_flag(TIF_NEED_RESCHED) +#define tif_need_resched_now() test_thread_flag(TIF_NEED_RESCHED) +#define tif_need_resched_lazy() 0 +#endif #ifndef CONFIG_HAVE_ARCH_WITHIN_STACK_FRAMES static inline int arch_within_stack_frames(const void * const stack, Index: linux-5.6.17-rt10/include/linux/trace_events.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/trace_events.h +++ linux-5.6.17-rt10/include/linux/trace_events.h @@ -67,6 +67,8 @@ struct trace_entry { unsigned char flags; unsigned char preempt_count; int pid; + unsigned char migrate_disable; + unsigned char preempt_lazy_count; }; #define TRACE_EVENT_TYPE_MAX \ Index: linux-5.6.17-rt10/include/linux/uaccess.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/uaccess.h +++ linux-5.6.17-rt10/include/linux/uaccess.h @@ -182,6 +182,7 @@ static __always_inline void pagefault_di */ static inline void pagefault_disable(void) { + migrate_disable(); pagefault_disabled_inc(); /* * make sure to have issued the store before a pagefault @@ -198,6 +199,7 @@ static inline void pagefault_enable(void */ barrier(); pagefault_disabled_dec(); + migrate_enable(); } /* Index: linux-5.6.17-rt10/include/linux/vmstat.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/vmstat.h +++ linux-5.6.17-rt10/include/linux/vmstat.h @@ -60,7 +60,9 @@ DECLARE_PER_CPU(struct vm_event_state, v */ static inline void __count_vm_event(enum vm_event_item item) { + preempt_disable_rt(); raw_cpu_inc(vm_event_states.event[item]); + preempt_enable_rt(); } static inline void count_vm_event(enum vm_event_item item) @@ -70,7 +72,9 @@ static inline void count_vm_event(enum v static inline void __count_vm_events(enum vm_event_item item, long delta) { + preempt_disable_rt(); raw_cpu_add(vm_event_states.event[item], delta); + preempt_enable_rt(); } static inline void count_vm_events(enum vm_event_item item, long delta) Index: linux-5.6.17-rt10/include/linux/wait.h =================================================================== --- linux-5.6.17-rt10.orig/include/linux/wait.h +++ linux-5.6.17-rt10/include/linux/wait.h @@ -10,6 +10,7 @@ #include #include +#include typedef struct wait_queue_entry wait_queue_entry_t; @@ -20,6 +21,7 @@ int default_wake_function(struct wait_qu #define WQ_FLAG_EXCLUSIVE 0x01 #define WQ_FLAG_WOKEN 0x02 #define WQ_FLAG_BOOKMARK 0x04 +#define WQ_FLAG_CUSTOM 0x08 /* * A single wait-queue entry structure: Index: linux-5.6.17-rt10/include/net/gen_stats.h =================================================================== --- linux-5.6.17-rt10.orig/include/net/gen_stats.h +++ linux-5.6.17-rt10/include/net/gen_stats.h @@ -6,6 +6,7 @@ #include #include #include +#include /* Note: this used to be in include/uapi/linux/gen_stats.h */ struct gnet_stats_basic_packed { @@ -42,15 +43,15 @@ int gnet_stats_start_copy_compat(struct spinlock_t *lock, struct gnet_dump *d, int padattr); -int gnet_stats_copy_basic(const seqcount_t *running, +int gnet_stats_copy_basic(net_seqlock_t *running, struct gnet_dump *d, struct gnet_stats_basic_cpu __percpu *cpu, struct gnet_stats_basic_packed *b); -void __gnet_stats_copy_basic(const seqcount_t *running, +void __gnet_stats_copy_basic(net_seqlock_t *running, struct gnet_stats_basic_packed *bstats, struct gnet_stats_basic_cpu __percpu *cpu, struct gnet_stats_basic_packed *b); -int gnet_stats_copy_basic_hw(const seqcount_t *running, +int gnet_stats_copy_basic_hw(net_seqlock_t *running, struct gnet_dump *d, struct gnet_stats_basic_cpu __percpu *cpu, struct gnet_stats_basic_packed *b); @@ -70,13 +71,13 @@ int gen_new_estimator(struct gnet_stats_ struct gnet_stats_basic_cpu __percpu *cpu_bstats, struct net_rate_estimator __rcu **rate_est, spinlock_t *lock, - seqcount_t *running, struct nlattr *opt); + net_seqlock_t *running, struct nlattr *opt); void gen_kill_estimator(struct net_rate_estimator __rcu **ptr); int gen_replace_estimator(struct gnet_stats_basic_packed *bstats, struct gnet_stats_basic_cpu __percpu *cpu_bstats, struct net_rate_estimator __rcu **ptr, spinlock_t *lock, - seqcount_t *running, struct nlattr *opt); + net_seqlock_t *running, struct nlattr *opt); bool gen_estimator_active(struct net_rate_estimator __rcu **ptr); bool gen_estimator_read(struct net_rate_estimator __rcu **ptr, struct gnet_stats_rate_est64 *sample); Index: linux-5.6.17-rt10/include/net/net_seq_lock.h =================================================================== --- /dev/null +++ linux-5.6.17-rt10/include/net/net_seq_lock.h @@ -0,0 +1,24 @@ +#ifndef __NET_NET_SEQ_LOCK_H__ +#define __NET_NET_SEQ_LOCK_H__ + +#ifdef CONFIG_PREEMPT_RT +# define net_seqlock_t seqlock_t +# define net_seq_begin(__r) read_seqbegin(__r) +# define net_seq_retry(__r, __s) read_seqretry(__r, __s) + +static inline int try_write_seqlock(seqlock_t *sl) +{ + if (spin_trylock(&sl->lock)) { + write_seqcount_begin(&sl->seqcount); + return 1; + } + return 0; +} + +#else +# define net_seqlock_t seqcount_t +# define net_seq_begin(__r) read_seqcount_begin(__r) +# define net_seq_retry(__r, __s) read_seqcount_retry(__r, __s) +#endif + +#endif Index: linux-5.6.17-rt10/include/net/netfilter/nf_conntrack.h =================================================================== --- linux-5.6.17-rt10.orig/include/net/netfilter/nf_conntrack.h +++ linux-5.6.17-rt10/include/net/netfilter/nf_conntrack.h @@ -286,7 +286,7 @@ int nf_conntrack_hash_resize(unsigned in extern struct hlist_nulls_head *nf_conntrack_hash; extern unsigned int nf_conntrack_htable_size; -extern seqcount_t nf_conntrack_generation; +extern seqcount_spinlock_t nf_conntrack_generation; extern unsigned int nf_conntrack_max; /* must be called with rcu read lock held */ Index: linux-5.6.17-rt10/include/net/sch_generic.h =================================================================== --- linux-5.6.17-rt10.orig/include/net/sch_generic.h +++ linux-5.6.17-rt10/include/net/sch_generic.h @@ -10,6 +10,7 @@ #include #include #include +#include #include #include #include @@ -100,7 +101,7 @@ struct Qdisc { struct sk_buff_head gso_skb ____cacheline_aligned_in_smp; struct qdisc_skb_head q; struct gnet_stats_basic_packed bstats; - seqcount_t running; + net_seqlock_t running; struct gnet_stats_queue qstats; unsigned long state; struct Qdisc *next_sched; @@ -138,7 +139,11 @@ static inline bool qdisc_is_running(stru { if (qdisc->flags & TCQ_F_NOLOCK) return spin_is_locked(&qdisc->seqlock); +#ifdef CONFIG_PREEMPT_RT + return spin_is_locked(&qdisc->running.lock) ? true : false; +#else return (raw_read_seqcount(&qdisc->running) & 1) ? true : false; +#endif } static inline bool qdisc_is_percpu_stats(const struct Qdisc *q) @@ -162,17 +167,27 @@ static inline bool qdisc_run_begin(struc } else if (qdisc_is_running(qdisc)) { return false; } +#ifdef CONFIG_PREEMPT_RT + if (try_write_seqlock(&qdisc->running)) + return true; + return false; +#else /* Variant of write_seqcount_begin() telling lockdep a trylock * was attempted. */ raw_write_seqcount_begin(&qdisc->running); seqcount_acquire(&qdisc->running.dep_map, 0, 1, _RET_IP_); return true; +#endif } static inline void qdisc_run_end(struct Qdisc *qdisc) { +#ifdef CONFIG_PREEMPT_RT + write_sequnlock(&qdisc->running); +#else write_seqcount_end(&qdisc->running); +#endif if (qdisc->flags & TCQ_F_NOLOCK) spin_unlock(&qdisc->seqlock); } @@ -542,7 +557,7 @@ static inline spinlock_t *qdisc_root_sle return qdisc_lock(root); } -static inline seqcount_t *qdisc_root_sleeping_running(const struct Qdisc *qdisc) +static inline net_seqlock_t *qdisc_root_sleeping_running(const struct Qdisc *qdisc) { struct Qdisc *root = qdisc_root_sleeping(qdisc); Index: linux-5.6.17-rt10/init/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/init/Kconfig +++ linux-5.6.17-rt10/init/Kconfig @@ -891,6 +891,7 @@ config CFS_BANDWIDTH config RT_GROUP_SCHED bool "Group scheduling for SCHED_RR/FIFO" depends on CGROUP_SCHED + depends on !PREEMPT_RT default n help This feature lets you explicitly allocate real CPU bandwidth @@ -1781,6 +1782,7 @@ choice config SLAB bool "SLAB" + depends on !PREEMPT_RT select HAVE_HARDENED_USERCOPY_ALLOCATOR help The regular slab allocator that is established and known to work @@ -1801,6 +1803,7 @@ config SLUB config SLOB depends on EXPERT bool "SLOB (Simple Allocator)" + depends on !PREEMPT_RT help SLOB replaces the stock allocator with a drastically simpler allocator. SLOB is generally more space efficient but @@ -1866,7 +1869,7 @@ config SHUFFLE_PAGE_ALLOCATOR config SLUB_CPU_PARTIAL default y - depends on SLUB && SMP + depends on SLUB && SMP && !PREEMPT_RT bool "SLUB per cpu partial cache" help Per cpu partial caches accelerate objects allocation and freeing Index: linux-5.6.17-rt10/init/init_task.c =================================================================== --- linux-5.6.17-rt10.orig/init/init_task.c +++ linux-5.6.17-rt10/init/init_task.c @@ -73,6 +73,10 @@ struct task_struct init_task .cpus_ptr = &init_task.cpus_mask, .cpus_mask = CPU_MASK_ALL, .nr_cpus_allowed= NR_CPUS, +#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) && \ + defined(CONFIG_SCHED_DEBUG) + .pinned_on_cpu = -1, +#endif .mm = NULL, .active_mm = &init_mm, .restart_block = { @@ -141,7 +145,8 @@ struct task_struct init_task .rcu_tasks_idle_cpu = -1, #endif #ifdef CONFIG_CPUSETS - .mems_allowed_seq = SEQCNT_ZERO(init_task.mems_allowed_seq), + .mems_allowed_seq = SEQCNT_SPINLOCK_ZERO(init_task.mems_allowed_seq, + &init_task.alloc_lock), #endif #ifdef CONFIG_RT_MUTEXES .pi_waiters = RB_ROOT_CACHED, Index: linux-5.6.17-rt10/kernel/Kconfig.preempt =================================================================== --- linux-5.6.17-rt10.orig/kernel/Kconfig.preempt +++ linux-5.6.17-rt10/kernel/Kconfig.preempt @@ -1,5 +1,11 @@ # SPDX-License-Identifier: GPL-2.0-only +config HAVE_PREEMPT_LAZY + bool + +config PREEMPT_LAZY + def_bool y if HAVE_PREEMPT_LAZY && PREEMPT_RT + choice prompt "Preemption Model" default PREEMPT_NONE Index: linux-5.6.17-rt10/kernel/bpf/hashtab.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/bpf/hashtab.c +++ linux-5.6.17-rt10/kernel/bpf/hashtab.c @@ -27,9 +27,62 @@ .map_delete_batch = \ generic_map_delete_batch +/* + * The bucket lock has two protection scopes: + * + * 1) Serializing concurrent operations from BPF programs on differrent + * CPUs + * + * 2) Serializing concurrent operations from BPF programs and sys_bpf() + * + * BPF programs can execute in any context including perf, kprobes and + * tracing. As there are almost no limits where perf, kprobes and tracing + * can be invoked from the lock operations need to be protected against + * deadlocks. Deadlocks can be caused by recursion and by an invocation in + * the lock held section when functions which acquire this lock are invoked + * from sys_bpf(). BPF recursion is prevented by incrementing the per CPU + * variable bpf_prog_active, which prevents BPF programs attached to perf + * events, kprobes and tracing to be invoked before the prior invocation + * from one of these contexts completed. sys_bpf() uses the same mechanism + * by pinning the task to the current CPU and incrementing the recursion + * protection accross the map operation. + * + * This has subtle implications on PREEMPT_RT. PREEMPT_RT forbids certain + * operations like memory allocations (even with GFP_ATOMIC) from atomic + * contexts. This is required because even with GFP_ATOMIC the memory + * allocator calls into code pathes which acquire locks with long held lock + * sections. To ensure the deterministic behaviour these locks are regular + * spinlocks, which are converted to 'sleepable' spinlocks on RT. The only + * true atomic contexts on an RT kernel are the low level hardware + * handling, scheduling, low level interrupt handling, NMIs etc. None of + * these contexts should ever do memory allocations. + * + * As regular device interrupt handlers and soft interrupts are forced into + * thread context, the existing code which does + * spin_lock*(); alloc(GPF_ATOMIC); spin_unlock*(); + * just works. + * + * In theory the BPF locks could be converted to regular spinlocks as well, + * but the bucket locks and percpu_freelist locks can be taken from + * arbitrary contexts (perf, kprobes, tracepoints) which are required to be + * atomic contexts even on RT. These mechanisms require preallocated maps, + * so there is no need to invoke memory allocations within the lock held + * sections. + * + * BPF maps which need dynamic allocation are only used from (forced) + * thread context on RT and can therefore use regular spinlocks which in + * turn allows to invoke memory allocations from the lock held section. + * + * On a non RT kernel this distinction is neither possible nor required. + * spinlock maps to raw_spinlock and the extra code is optimized out by the + * compiler. + */ struct bucket { struct hlist_nulls_head head; - raw_spinlock_t lock; + union { + raw_spinlock_t raw_lock; + spinlock_t lock; + }; }; struct bpf_htab { @@ -68,6 +121,51 @@ struct htab_elem { char key[0] __aligned(8); }; +static inline bool htab_is_prealloc(const struct bpf_htab *htab) +{ + return !(htab->map.map_flags & BPF_F_NO_PREALLOC); +} + +static inline bool htab_use_raw_lock(const struct bpf_htab *htab) +{ + return (!IS_ENABLED(CONFIG_PREEMPT_RT) || htab_is_prealloc(htab)); +} + +static void htab_init_buckets(struct bpf_htab *htab) +{ + unsigned i; + + for (i = 0; i < htab->n_buckets; i++) { + INIT_HLIST_NULLS_HEAD(&htab->buckets[i].head, i); + if (htab_use_raw_lock(htab)) + raw_spin_lock_init(&htab->buckets[i].raw_lock); + else + spin_lock_init(&htab->buckets[i].lock); + } +} + +static inline unsigned long htab_lock_bucket(const struct bpf_htab *htab, + struct bucket *b) +{ + unsigned long flags; + + if (htab_use_raw_lock(htab)) + raw_spin_lock_irqsave(&b->raw_lock, flags); + else + spin_lock_irqsave(&b->lock, flags); + return flags; +} + +static inline void htab_unlock_bucket(const struct bpf_htab *htab, + struct bucket *b, + unsigned long flags) +{ + if (htab_use_raw_lock(htab)) + raw_spin_unlock_irqrestore(&b->raw_lock, flags); + else + spin_unlock_irqrestore(&b->lock, flags); +} + static bool htab_lru_map_delete_node(void *arg, struct bpf_lru_node *node); static bool htab_is_lru(const struct bpf_htab *htab) @@ -82,11 +180,6 @@ static bool htab_is_percpu(const struct htab->map.map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH; } -static bool htab_is_prealloc(const struct bpf_htab *htab) -{ - return !(htab->map.map_flags & BPF_F_NO_PREALLOC); -} - static inline void htab_elem_set_ptr(struct htab_elem *l, u32 key_size, void __percpu *pptr) { @@ -328,8 +421,8 @@ static struct bpf_map *htab_map_alloc(un bool percpu_lru = (attr->map_flags & BPF_F_NO_COMMON_LRU); bool prealloc = !(attr->map_flags & BPF_F_NO_PREALLOC); struct bpf_htab *htab; - int err, i; u64 cost; + int err; htab = kzalloc(sizeof(*htab), GFP_USER); if (!htab) @@ -391,10 +484,7 @@ static struct bpf_map *htab_map_alloc(un else htab->hashrnd = get_random_int(); - for (i = 0; i < htab->n_buckets; i++) { - INIT_HLIST_NULLS_HEAD(&htab->buckets[i].head, i); - raw_spin_lock_init(&htab->buckets[i].lock); - } + htab_init_buckets(htab); if (prealloc) { err = prealloc_init(htab); @@ -602,7 +692,7 @@ static bool htab_lru_map_delete_node(voi b = __select_bucket(htab, tgt_l->hash); head = &b->head; - raw_spin_lock_irqsave(&b->lock, flags); + flags = htab_lock_bucket(htab, b); hlist_nulls_for_each_entry_rcu(l, n, head, hash_node) if (l == tgt_l) { @@ -610,7 +700,7 @@ static bool htab_lru_map_delete_node(voi break; } - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); return l == tgt_l; } @@ -686,15 +776,7 @@ static void htab_elem_free_rcu(struct rc struct htab_elem *l = container_of(head, struct htab_elem, rcu); struct bpf_htab *htab = l->htab; - /* must increment bpf_prog_active to avoid kprobe+bpf triggering while - * we're calling kfree, otherwise deadlock is possible if kprobes - * are placed somewhere inside of slub - */ - preempt_disable(); - __this_cpu_inc(bpf_prog_active); htab_elem_free(htab, l); - __this_cpu_dec(bpf_prog_active); - preempt_enable(); } static void free_htab_elem(struct bpf_htab *htab, struct htab_elem *l) @@ -884,8 +966,7 @@ static int htab_map_update_elem(struct b */ } - /* bpf_map_update_elem() can be called in_irq() */ - raw_spin_lock_irqsave(&b->lock, flags); + flags = htab_lock_bucket(htab, b); l_old = lookup_elem_raw(head, hash, key, key_size); @@ -926,7 +1007,7 @@ static int htab_map_update_elem(struct b } ret = 0; err: - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); return ret; } @@ -964,8 +1045,7 @@ static int htab_lru_map_update_elem(stru return -ENOMEM; memcpy(l_new->key + round_up(map->key_size, 8), value, map->value_size); - /* bpf_map_update_elem() can be called in_irq() */ - raw_spin_lock_irqsave(&b->lock, flags); + flags = htab_lock_bucket(htab, b); l_old = lookup_elem_raw(head, hash, key, key_size); @@ -984,7 +1064,7 @@ static int htab_lru_map_update_elem(stru ret = 0; err: - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); if (ret) bpf_lru_push_free(&htab->lru, &l_new->lru_node); @@ -1019,8 +1099,7 @@ static int __htab_percpu_map_update_elem b = __select_bucket(htab, hash); head = &b->head; - /* bpf_map_update_elem() can be called in_irq() */ - raw_spin_lock_irqsave(&b->lock, flags); + flags = htab_lock_bucket(htab, b); l_old = lookup_elem_raw(head, hash, key, key_size); @@ -1043,7 +1122,7 @@ static int __htab_percpu_map_update_elem } ret = 0; err: - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); return ret; } @@ -1083,8 +1162,7 @@ static int __htab_lru_percpu_map_update_ return -ENOMEM; } - /* bpf_map_update_elem() can be called in_irq() */ - raw_spin_lock_irqsave(&b->lock, flags); + flags = htab_lock_bucket(htab, b); l_old = lookup_elem_raw(head, hash, key, key_size); @@ -1106,7 +1184,7 @@ static int __htab_lru_percpu_map_update_ } ret = 0; err: - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); if (l_new) bpf_lru_push_free(&htab->lru, &l_new->lru_node); return ret; @@ -1144,7 +1222,7 @@ static int htab_map_delete_elem(struct b b = __select_bucket(htab, hash); head = &b->head; - raw_spin_lock_irqsave(&b->lock, flags); + flags = htab_lock_bucket(htab, b); l = lookup_elem_raw(head, hash, key, key_size); @@ -1154,7 +1232,7 @@ static int htab_map_delete_elem(struct b ret = 0; } - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); return ret; } @@ -1176,7 +1254,7 @@ static int htab_lru_map_delete_elem(stru b = __select_bucket(htab, hash); head = &b->head; - raw_spin_lock_irqsave(&b->lock, flags); + flags = htab_lock_bucket(htab, b); l = lookup_elem_raw(head, hash, key, key_size); @@ -1185,7 +1263,7 @@ static int htab_lru_map_delete_elem(stru ret = 0; } - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); if (l) bpf_lru_push_free(&htab->lru, &l->lru_node); return ret; @@ -1325,8 +1403,7 @@ alloc: } again: - preempt_disable(); - this_cpu_inc(bpf_prog_active); + bpf_disable_instrumentation(); rcu_read_lock(); again_nocopy: dst_key = keys; @@ -1335,7 +1412,7 @@ again_nocopy: head = &b->head; /* do not grab the lock unless need it (bucket_cnt > 0). */ if (locked) - raw_spin_lock_irqsave(&b->lock, flags); + flags = htab_lock_bucket(htab, b); bucket_cnt = 0; hlist_nulls_for_each_entry_rcu(l, n, head, hash_node) @@ -1352,10 +1429,9 @@ again_nocopy: /* Note that since bucket_cnt > 0 here, it is implicit * that the locked was grabbed, so release it. */ - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); rcu_read_unlock(); - this_cpu_dec(bpf_prog_active); - preempt_enable(); + bpf_enable_instrumentation(); goto after_loop; } @@ -1364,10 +1440,9 @@ again_nocopy: /* Note that since bucket_cnt > 0 here, it is implicit * that the locked was grabbed, so release it. */ - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); rcu_read_unlock(); - this_cpu_dec(bpf_prog_active); - preempt_enable(); + bpf_enable_instrumentation(); kvfree(keys); kvfree(values); goto alloc; @@ -1418,7 +1493,7 @@ again_nocopy: dst_val += value_size; } - raw_spin_unlock_irqrestore(&b->lock, flags); + htab_unlock_bucket(htab, b, flags); locked = false; while (node_to_free) { @@ -1437,8 +1512,7 @@ next_batch: } rcu_read_unlock(); - this_cpu_dec(bpf_prog_active); - preempt_enable(); + bpf_enable_instrumentation(); if (bucket_cnt && (copy_to_user(ukeys + total * key_size, keys, key_size * bucket_cnt) || copy_to_user(uvalues + total * value_size, values, Index: linux-5.6.17-rt10/kernel/bpf/lpm_trie.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/bpf/lpm_trie.c +++ linux-5.6.17-rt10/kernel/bpf/lpm_trie.c @@ -34,7 +34,7 @@ struct lpm_trie { size_t n_entries; size_t max_prefixlen; size_t data_size; - raw_spinlock_t lock; + spinlock_t lock; }; /* This trie implements a longest prefix match algorithm that can be used to @@ -315,7 +315,7 @@ static int trie_update_elem(struct bpf_m if (key->prefixlen > trie->max_prefixlen) return -EINVAL; - raw_spin_lock_irqsave(&trie->lock, irq_flags); + spin_lock_irqsave(&trie->lock, irq_flags); /* Allocate and fill a new node */ @@ -422,7 +422,7 @@ out: kfree(im_node); } - raw_spin_unlock_irqrestore(&trie->lock, irq_flags); + spin_unlock_irqrestore(&trie->lock, irq_flags); return ret; } @@ -442,7 +442,7 @@ static int trie_delete_elem(struct bpf_m if (key->prefixlen > trie->max_prefixlen) return -EINVAL; - raw_spin_lock_irqsave(&trie->lock, irq_flags); + spin_lock_irqsave(&trie->lock, irq_flags); /* Walk the tree looking for an exact key/length match and keeping * track of the path we traverse. We will need to know the node @@ -518,7 +518,7 @@ static int trie_delete_elem(struct bpf_m kfree_rcu(node, rcu); out: - raw_spin_unlock_irqrestore(&trie->lock, irq_flags); + spin_unlock_irqrestore(&trie->lock, irq_flags); return ret; } @@ -575,7 +575,7 @@ static struct bpf_map *trie_alloc(union if (ret) goto out_err; - raw_spin_lock_init(&trie->lock); + spin_lock_init(&trie->lock); return &trie->map; out_err: Index: linux-5.6.17-rt10/kernel/bpf/percpu_freelist.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/bpf/percpu_freelist.c +++ linux-5.6.17-rt10/kernel/bpf/percpu_freelist.c @@ -25,12 +25,18 @@ void pcpu_freelist_destroy(struct pcpu_f free_percpu(s->freelist); } +static inline void pcpu_freelist_push_node(struct pcpu_freelist_head *head, + struct pcpu_freelist_node *node) +{ + node->next = head->first; + head->first = node; +} + static inline void ___pcpu_freelist_push(struct pcpu_freelist_head *head, struct pcpu_freelist_node *node) { raw_spin_lock(&head->lock); - node->next = head->first; - head->first = node; + pcpu_freelist_push_node(head, node); raw_spin_unlock(&head->lock); } @@ -56,21 +62,16 @@ void pcpu_freelist_populate(struct pcpu_ u32 nr_elems) { struct pcpu_freelist_head *head; - unsigned long flags; int i, cpu, pcpu_entries; pcpu_entries = nr_elems / num_possible_cpus() + 1; i = 0; - /* disable irq to workaround lockdep false positive - * in bpf usage pcpu_freelist_populate() will never race - * with pcpu_freelist_push() - */ - local_irq_save(flags); for_each_possible_cpu(cpu) { again: head = per_cpu_ptr(s->freelist, cpu); - ___pcpu_freelist_push(head, buf); + /* No locking required as this is not visible yet. */ + pcpu_freelist_push_node(head, buf); i++; buf += elem_size; if (i == nr_elems) @@ -78,7 +79,6 @@ again: if (i % pcpu_entries) goto again; } - local_irq_restore(flags); } struct pcpu_freelist_node *__pcpu_freelist_pop(struct pcpu_freelist *s) Index: linux-5.6.17-rt10/kernel/bpf/stackmap.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/bpf/stackmap.c +++ linux-5.6.17-rt10/kernel/bpf/stackmap.c @@ -40,6 +40,9 @@ static void do_up_read(struct irq_work * { struct stack_map_irq_work *work; + if (WARN_ON_ONCE(IS_ENABLED(CONFIG_PREEMPT_RT))) + return; + work = container_of(entry, struct stack_map_irq_work, irq_work); up_read_non_owner(work->sem); work->sem = NULL; @@ -288,10 +291,19 @@ static void stack_map_get_build_id_offse struct stack_map_irq_work *work = NULL; if (irqs_disabled()) { - work = this_cpu_ptr(&up_read_work); - if (atomic_read(&work->irq_work.flags) & IRQ_WORK_BUSY) - /* cannot queue more up_read, fallback */ + if (!IS_ENABLED(CONFIG_PREEMPT_RT)) { + work = this_cpu_ptr(&up_read_work); + if (atomic_read(&work->irq_work.flags) & IRQ_WORK_BUSY) { + /* cannot queue more up_read, fallback */ + irq_work_busy = true; + } + } else { + /* + * PREEMPT_RT does not allow to trylock mmap sem in + * interrupt disabled context. Force the fallback code. + */ irq_work_busy = true; + } } /* Index: linux-5.6.17-rt10/kernel/bpf/syscall.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/bpf/syscall.c +++ linux-5.6.17-rt10/kernel/bpf/syscall.c @@ -171,11 +171,7 @@ static int bpf_map_update_value(struct b flags); } - /* must increment bpf_prog_active to avoid kprobe+bpf triggering from - * inside bpf map update or delete otherwise deadlocks are possible - */ - preempt_disable(); - __this_cpu_inc(bpf_prog_active); + bpf_disable_instrumentation(); if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH || map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH) { err = bpf_percpu_hash_update(map, key, value, flags); @@ -206,8 +202,7 @@ static int bpf_map_update_value(struct b err = map->ops->map_update_elem(map, key, value, flags); rcu_read_unlock(); } - __this_cpu_dec(bpf_prog_active); - preempt_enable(); + bpf_enable_instrumentation(); maybe_wait_bpf_programs(map); return err; @@ -222,8 +217,7 @@ static int bpf_map_copy_value(struct bpf if (bpf_map_is_dev_bound(map)) return bpf_map_offload_lookup_elem(map, key, value); - preempt_disable(); - this_cpu_inc(bpf_prog_active); + bpf_disable_instrumentation(); if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH || map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH) { err = bpf_percpu_hash_copy(map, key, value); @@ -268,8 +262,7 @@ static int bpf_map_copy_value(struct bpf rcu_read_unlock(); } - this_cpu_dec(bpf_prog_active); - preempt_enable(); + bpf_enable_instrumentation(); maybe_wait_bpf_programs(map); return err; @@ -1147,13 +1140,11 @@ static int map_delete_elem(union bpf_att goto out; } - preempt_disable(); - __this_cpu_inc(bpf_prog_active); + bpf_disable_instrumentation(); rcu_read_lock(); err = map->ops->map_delete_elem(map, key); rcu_read_unlock(); - __this_cpu_dec(bpf_prog_active); - preempt_enable(); + bpf_enable_instrumentation(); maybe_wait_bpf_programs(map); out: kfree(key); @@ -1265,13 +1256,11 @@ int generic_map_delete_batch(struct bpf_ break; } - preempt_disable(); - __this_cpu_inc(bpf_prog_active); + bpf_disable_instrumentation(); rcu_read_lock(); err = map->ops->map_delete_elem(map, key); rcu_read_unlock(); - __this_cpu_dec(bpf_prog_active); - preempt_enable(); + bpf_enable_instrumentation(); maybe_wait_bpf_programs(map); if (err) break; Index: linux-5.6.17-rt10/kernel/bpf/trampoline.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/bpf/trampoline.c +++ linux-5.6.17-rt10/kernel/bpf/trampoline.c @@ -367,8 +367,9 @@ out: mutex_unlock(&trampoline_mutex); } -/* The logic is similar to BPF_PROG_RUN, but with explicit rcu and preempt that - * are needed for trampoline. The macro is split into +/* The logic is similar to BPF_PROG_RUN, but with an explicit + * rcu_read_lock() and migrate_disable() which are required + * for the trampoline. The macro is split into * call _bpf_prog_enter * call prog->bpf_func * call __bpf_prog_exit @@ -378,7 +379,7 @@ u64 notrace __bpf_prog_enter(void) u64 start = 0; rcu_read_lock(); - preempt_disable(); + migrate_disable(); if (static_branch_unlikely(&bpf_stats_enabled_key)) start = sched_clock(); return start; @@ -401,7 +402,7 @@ void notrace __bpf_prog_exit(struct bpf_ stats->nsecs += sched_clock() - start; u64_stats_update_end(&stats->syncp); } - preempt_enable(); + migrate_enable(); rcu_read_unlock(); } Index: linux-5.6.17-rt10/kernel/bpf/verifier.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/bpf/verifier.c +++ linux-5.6.17-rt10/kernel/bpf/verifier.c @@ -8207,26 +8207,48 @@ static bool is_tracing_prog_type(enum bp } } +static bool is_preallocated_map(struct bpf_map *map) +{ + if (!check_map_prealloc(map)) + return false; + if (map->inner_map_meta && !check_map_prealloc(map->inner_map_meta)) + return false; + return true; +} + static int check_map_prog_compatibility(struct bpf_verifier_env *env, struct bpf_map *map, struct bpf_prog *prog) { - /* Make sure that BPF_PROG_TYPE_PERF_EVENT programs only use - * preallocated hash maps, since doing memory allocation - * in overflow_handler can crash depending on where nmi got - * triggered. + /* + * Validate that trace type programs use preallocated hash maps. + * + * For programs attached to PERF events this is mandatory as the + * perf NMI can hit any arbitrary code sequence. + * + * All other trace types using preallocated hash maps are unsafe as + * well because tracepoint or kprobes can be inside locked regions + * of the memory allocator or at a place where a recursion into the + * memory allocator would see inconsistent state. + * + * On RT enabled kernels run-time allocation of all trace type + * programs is strictly prohibited due to lock type constraints. On + * !RT kernels it is allowed for backwards compatibility reasons for + * now, but warnings are emitted so developers are made aware of + * the unsafety and can fix their programs before this is enforced. */ - if (prog->type == BPF_PROG_TYPE_PERF_EVENT) { - if (!check_map_prealloc(map)) { + if (is_tracing_prog_type(prog->type) && !is_preallocated_map(map)) { + if (prog->type == BPF_PROG_TYPE_PERF_EVENT) { verbose(env, "perf_event programs can only use preallocated hash map\n"); return -EINVAL; } - if (map->inner_map_meta && - !check_map_prealloc(map->inner_map_meta)) { - verbose(env, "perf_event programs can only use preallocated inner hash map\n"); + if (IS_ENABLED(CONFIG_PREEMPT_RT)) { + verbose(env, "trace type programs can only use preallocated hash map\n"); return -EINVAL; } + WARN_ONCE(1, "trace type BPF program uses run-time allocation\n"); + verbose(env, "trace type programs with run-time allocated hash maps are unsafe. Switch to preallocated hash maps.\n"); } if ((is_tracing_prog_type(prog->type) || Index: linux-5.6.17-rt10/kernel/cgroup/cpuset.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/cgroup/cpuset.c +++ linux-5.6.17-rt10/kernel/cgroup/cpuset.c @@ -345,7 +345,7 @@ void cpuset_read_unlock(void) percpu_up_read(&cpuset_rwsem); } -static DEFINE_SPINLOCK(callback_lock); +static DEFINE_RAW_SPINLOCK(callback_lock); static struct workqueue_struct *cpuset_migrate_mm_wq; @@ -1253,7 +1253,7 @@ static int update_parent_subparts_cpumas * Newly added CPUs will be removed from effective_cpus and * newly deleted ones will be added back to effective_cpus. */ - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); if (adding) { cpumask_or(parent->subparts_cpus, parent->subparts_cpus, tmp->addmask); @@ -1272,7 +1272,7 @@ static int update_parent_subparts_cpumas } parent->nr_subparts_cpus = cpumask_weight(parent->subparts_cpus); - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); return cmd == partcmd_update; } @@ -1377,7 +1377,7 @@ static void update_cpumasks_hier(struct continue; rcu_read_unlock(); - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); cpumask_copy(cp->effective_cpus, tmp->new_cpus); if (cp->nr_subparts_cpus && @@ -1408,7 +1408,7 @@ static void update_cpumasks_hier(struct = cpumask_weight(cp->subparts_cpus); } } - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); WARN_ON(!is_in_v2_mode() && !cpumask_equal(cp->cpus_allowed, cp->effective_cpus)); @@ -1526,7 +1526,7 @@ static int update_cpumask(struct cpuset return -EINVAL; } - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); cpumask_copy(cs->cpus_allowed, trialcs->cpus_allowed); /* @@ -1537,7 +1537,7 @@ static int update_cpumask(struct cpuset cs->cpus_allowed); cs->nr_subparts_cpus = cpumask_weight(cs->subparts_cpus); } - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); update_cpumasks_hier(cs, &tmp); @@ -1731,9 +1731,9 @@ static void update_nodemasks_hier(struct continue; rcu_read_unlock(); - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); cp->effective_mems = *new_mems; - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); WARN_ON(!is_in_v2_mode() && !nodes_equal(cp->mems_allowed, cp->effective_mems)); @@ -1801,9 +1801,9 @@ static int update_nodemask(struct cpuset if (retval < 0) goto done; - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); cs->mems_allowed = trialcs->mems_allowed; - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); /* use trialcs->mems_allowed as a temp variable */ update_nodemasks_hier(cs, &trialcs->mems_allowed); @@ -1894,9 +1894,9 @@ static int update_flag(cpuset_flagbits_t spread_flag_changed = ((is_spread_slab(cs) != is_spread_slab(trialcs)) || (is_spread_page(cs) != is_spread_page(trialcs))); - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); cs->flags = trialcs->flags; - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); if (!cpumask_empty(trialcs->cpus_allowed) && balance_flag_changed) rebuild_sched_domains_locked(); @@ -2405,7 +2405,7 @@ static int cpuset_common_seq_show(struct cpuset_filetype_t type = seq_cft(sf)->private; int ret = 0; - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); switch (type) { case FILE_CPULIST: @@ -2427,7 +2427,7 @@ static int cpuset_common_seq_show(struct ret = -EINVAL; } - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); return ret; } @@ -2740,14 +2740,14 @@ static int cpuset_css_online(struct cgro cpuset_inc(); - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); if (is_in_v2_mode()) { cpumask_copy(cs->effective_cpus, parent->effective_cpus); cs->effective_mems = parent->effective_mems; cs->use_parent_ecpus = true; parent->child_ecpus_count++; } - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &css->cgroup->flags)) goto out_unlock; @@ -2774,12 +2774,12 @@ static int cpuset_css_online(struct cgro } rcu_read_unlock(); - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); cs->mems_allowed = parent->mems_allowed; cs->effective_mems = parent->mems_allowed; cpumask_copy(cs->cpus_allowed, parent->cpus_allowed); cpumask_copy(cs->effective_cpus, parent->cpus_allowed); - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); out_unlock: percpu_up_write(&cpuset_rwsem); put_online_cpus(); @@ -2835,7 +2835,7 @@ static void cpuset_css_free(struct cgrou static void cpuset_bind(struct cgroup_subsys_state *root_css) { percpu_down_write(&cpuset_rwsem); - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); if (is_in_v2_mode()) { cpumask_copy(top_cpuset.cpus_allowed, cpu_possible_mask); @@ -2846,7 +2846,7 @@ static void cpuset_bind(struct cgroup_su top_cpuset.mems_allowed = top_cpuset.effective_mems; } - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); percpu_up_write(&cpuset_rwsem); } @@ -2943,12 +2943,12 @@ hotplug_update_tasks_legacy(struct cpuse { bool is_empty; - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); cpumask_copy(cs->cpus_allowed, new_cpus); cpumask_copy(cs->effective_cpus, new_cpus); cs->mems_allowed = *new_mems; cs->effective_mems = *new_mems; - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); /* * Don't call update_tasks_cpumask() if the cpuset becomes empty, @@ -2985,10 +2985,10 @@ hotplug_update_tasks(struct cpuset *cs, if (nodes_empty(*new_mems)) *new_mems = parent_cs(cs)->effective_mems; - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); cpumask_copy(cs->effective_cpus, new_cpus); cs->effective_mems = *new_mems; - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); if (cpus_updated) update_tasks_cpumask(cs); @@ -3143,7 +3143,7 @@ static void cpuset_hotplug_workfn(struct /* synchronize cpus_allowed to cpu_active_mask */ if (cpus_updated) { - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); if (!on_dfl) cpumask_copy(top_cpuset.cpus_allowed, &new_cpus); /* @@ -3163,17 +3163,17 @@ static void cpuset_hotplug_workfn(struct } } cpumask_copy(top_cpuset.effective_cpus, &new_cpus); - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); /* we don't mess with cpumasks of tasks in top_cpuset */ } /* synchronize mems_allowed to N_MEMORY */ if (mems_updated) { - spin_lock_irq(&callback_lock); + raw_spin_lock_irq(&callback_lock); if (!on_dfl) top_cpuset.mems_allowed = new_mems; top_cpuset.effective_mems = new_mems; - spin_unlock_irq(&callback_lock); + raw_spin_unlock_irq(&callback_lock); update_tasks_nodemask(&top_cpuset); } @@ -3274,11 +3274,11 @@ void cpuset_cpus_allowed(struct task_str { unsigned long flags; - spin_lock_irqsave(&callback_lock, flags); + raw_spin_lock_irqsave(&callback_lock, flags); rcu_read_lock(); guarantee_online_cpus(task_cs(tsk), pmask); rcu_read_unlock(); - spin_unlock_irqrestore(&callback_lock, flags); + raw_spin_unlock_irqrestore(&callback_lock, flags); } /** @@ -3339,11 +3339,11 @@ nodemask_t cpuset_mems_allowed(struct ta nodemask_t mask; unsigned long flags; - spin_lock_irqsave(&callback_lock, flags); + raw_spin_lock_irqsave(&callback_lock, flags); rcu_read_lock(); guarantee_online_mems(task_cs(tsk), &mask); rcu_read_unlock(); - spin_unlock_irqrestore(&callback_lock, flags); + raw_spin_unlock_irqrestore(&callback_lock, flags); return mask; } @@ -3435,14 +3435,14 @@ bool __cpuset_node_allowed(int node, gfp return true; /* Not hardwall and node outside mems_allowed: scan up cpusets */ - spin_lock_irqsave(&callback_lock, flags); + raw_spin_lock_irqsave(&callback_lock, flags); rcu_read_lock(); cs = nearest_hardwall_ancestor(task_cs(current)); allowed = node_isset(node, cs->mems_allowed); rcu_read_unlock(); - spin_unlock_irqrestore(&callback_lock, flags); + raw_spin_unlock_irqrestore(&callback_lock, flags); return allowed; } Index: linux-5.6.17-rt10/kernel/cgroup/rstat.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/cgroup/rstat.c +++ linux-5.6.17-rt10/kernel/cgroup/rstat.c @@ -150,8 +150,9 @@ static void cgroup_rstat_flush_locked(st raw_spinlock_t *cpu_lock = per_cpu_ptr(&cgroup_rstat_cpu_lock, cpu); struct cgroup *pos = NULL; + unsigned long flags; - raw_spin_lock(cpu_lock); + raw_spin_lock_irqsave(cpu_lock, flags); while ((pos = cgroup_rstat_cpu_pop_updated(pos, cgrp, cpu))) { struct cgroup_subsys_state *css; @@ -163,7 +164,7 @@ static void cgroup_rstat_flush_locked(st css->ss->css_rstat_flush(css, cpu); rcu_read_unlock(); } - raw_spin_unlock(cpu_lock); + raw_spin_unlock_irqrestore(cpu_lock, flags); /* if @may_sleep, play nice and yield if necessary */ if (may_sleep && (need_resched() || Index: linux-5.6.17-rt10/kernel/cpu.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/cpu.c +++ linux-5.6.17-rt10/kernel/cpu.c @@ -331,12 +331,12 @@ void lockdep_assert_cpus_held(void) static void lockdep_acquire_cpus_lock(void) { - rwsem_acquire(&cpu_hotplug_lock.rw_sem.dep_map, 0, 0, _THIS_IP_); + rwsem_acquire(&cpu_hotplug_lock.dep_map, 0, 0, _THIS_IP_); } static void lockdep_release_cpus_lock(void) { - rwsem_release(&cpu_hotplug_lock.rw_sem.dep_map, _THIS_IP_); + rwsem_release(&cpu_hotplug_lock.dep_map, _THIS_IP_); } /* @@ -848,6 +848,15 @@ static int take_cpu_down(void *_param) int err, cpu = smp_processor_id(); int ret; +#ifdef CONFIG_PREEMPT_RT + /* + * If any tasks disabled migration before we got here, + * go back and sleep again. + */ + if (cpu_nr_pinned(cpu)) + return -EAGAIN; +#endif + /* Ensure this CPU doesn't handle any more interrupts. */ err = __cpu_disable(); if (err < 0) @@ -877,6 +886,10 @@ static int take_cpu_down(void *_param) return 0; } +#ifdef CONFIG_PREEMPT_RT +struct task_struct *takedown_cpu_task; +#endif + static int takedown_cpu(unsigned int cpu) { struct cpuhp_cpu_state *st = per_cpu_ptr(&cpuhp_state, cpu); @@ -891,11 +904,39 @@ static int takedown_cpu(unsigned int cpu */ irq_lock_sparse(); +#ifdef CONFIG_PREEMPT_RT + WARN_ON_ONCE(takedown_cpu_task); + takedown_cpu_task = current; + +again: + /* + * If a task pins this CPU after we pass this check, take_cpu_down + * will return -EAGAIN. + */ + for (;;) { + int nr_pinned; + + set_current_state(TASK_UNINTERRUPTIBLE); + nr_pinned = cpu_nr_pinned(cpu); + if (nr_pinned == 0) + break; + schedule(); + } + set_current_state(TASK_RUNNING); +#endif + /* * So now all preempt/rcu users must observe !cpu_active(). */ err = stop_machine_cpuslocked(take_cpu_down, NULL, cpumask_of(cpu)); +#ifdef CONFIG_PREEMPT_RT + if (err == -EAGAIN) + goto again; +#endif if (err) { +#ifdef CONFIG_PREEMPT_RT + takedown_cpu_task = NULL; +#endif /* CPU refused to die */ irq_unlock_sparse(); /* Unpark the hotplug thread so we can rollback there */ @@ -914,6 +955,9 @@ static int takedown_cpu(unsigned int cpu wait_for_ap_thread(st, false); BUG_ON(st->state != CPUHP_AP_IDLE_DEAD); +#ifdef CONFIG_PREEMPT_RT + takedown_cpu_task = NULL; +#endif /* Interrupts are moved away from the dying cpu, reenable alloc/free */ irq_unlock_sparse(); Index: linux-5.6.17-rt10/kernel/events/core.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/events/core.c +++ linux-5.6.17-rt10/kernel/events/core.c @@ -9207,7 +9207,6 @@ static void bpf_overflow_handler(struct int ret = 0; ctx.regs = perf_arch_bpf_user_pt_regs(regs); - preempt_disable(); if (unlikely(__this_cpu_inc_return(bpf_prog_active) != 1)) goto out; rcu_read_lock(); @@ -9215,7 +9214,6 @@ static void bpf_overflow_handler(struct rcu_read_unlock(); out: __this_cpu_dec(bpf_prog_active); - preempt_enable(); if (!ret) return; Index: linux-5.6.17-rt10/kernel/exit.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/exit.c +++ linux-5.6.17-rt10/kernel/exit.c @@ -161,7 +161,7 @@ static void __exit_signal(struct task_st * Do this under ->siglock, we can race with another thread * doing sigqueue_free() if we have SIGQUEUE_PREALLOC signals. */ - flush_sigqueue(&tsk->pending); + flush_task_sigqueue(tsk); tsk->sighand = NULL; spin_unlock(&sighand->siglock); @@ -258,6 +258,7 @@ void rcuwait_wake_up(struct rcuwait *w) wake_up_process(task); rcu_read_unlock(); } +EXPORT_SYMBOL_GPL(rcuwait_wake_up); /* * Determine if a process group is "orphaned", according to the POSIX Index: linux-5.6.17-rt10/kernel/fork.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/fork.c +++ linux-5.6.17-rt10/kernel/fork.c @@ -42,6 +42,7 @@ #include #include #include +#include #include #include #include @@ -292,7 +293,7 @@ static inline void free_thread_stack(str return; } - vfree_atomic(tsk->stack); + vfree(tsk->stack); return; } #endif @@ -699,6 +700,19 @@ void __mmdrop(struct mm_struct *mm) } EXPORT_SYMBOL_GPL(__mmdrop); +#ifdef CONFIG_PREEMPT_RT +/* + * RCU callback for delayed mm drop. Not strictly rcu, but we don't + * want another facility to make this work. + */ +void __mmdrop_delayed(struct rcu_head *rhp) +{ + struct mm_struct *mm = container_of(rhp, struct mm_struct, delayed_drop); + + __mmdrop(mm); +} +#endif + static void mmdrop_async_fn(struct work_struct *work) { struct mm_struct *mm; @@ -740,6 +754,15 @@ void __put_task_struct(struct task_struc WARN_ON(refcount_read(&tsk->usage)); WARN_ON(tsk == current); + /* + * Remove function-return probe instances associated with this + * task and put them back on the free list. + */ + kprobe_flush_task(tsk); + + /* Task is done with its stack. */ + put_task_stack(tsk); + cgroup_free(tsk); task_numa_free(tsk, true); security_task_free(tsk); @@ -930,6 +953,7 @@ static struct task_struct *dup_task_stru tsk->splice_pipe = NULL; tsk->task_frag.page = NULL; tsk->wake_q.next = NULL; + tsk->wake_q_sleeper.next = NULL; account_kernel_stack(tsk, 1); @@ -1973,6 +1997,7 @@ static __latent_entropy struct task_stru spin_lock_init(&p->alloc_lock); init_sigpending(&p->pending); + p->sigqueue_cache = NULL; p->utime = p->stime = p->gtime = 0; #ifdef CONFIG_ARCH_HAS_SCALED_CPUTIME @@ -2015,7 +2040,7 @@ static __latent_entropy struct task_stru #ifdef CONFIG_CPUSETS p->cpuset_mem_spread_rotor = NUMA_NO_NODE; p->cpuset_slab_spread_rotor = NUMA_NO_NODE; - seqcount_init(&p->mems_allowed_seq); + seqcount_spinlock_init(&p->mems_allowed_seq, &p->alloc_lock); #endif #ifdef CONFIG_TRACE_IRQFLAGS p->irq_events = 0; Index: linux-5.6.17-rt10/kernel/futex.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/futex.c +++ linux-5.6.17-rt10/kernel/futex.c @@ -962,7 +962,9 @@ static void exit_pi_state_list(struct ta if (head->next != next) { /* retain curr->pi_lock for the loop invariant */ raw_spin_unlock(&pi_state->pi_mutex.wait_lock); + raw_spin_unlock_irq(&curr->pi_lock); spin_unlock(&hb->lock); + raw_spin_lock_irq(&curr->pi_lock); put_pi_state(pi_state); continue; } @@ -1572,6 +1574,7 @@ static int wake_futex_pi(u32 __user *uad struct task_struct *new_owner; bool postunlock = false; DEFINE_WAKE_Q(wake_q); + DEFINE_WAKE_Q(wake_sleeper_q); int ret = 0; new_owner = rt_mutex_next_owner(&pi_state->pi_mutex); @@ -1631,13 +1634,13 @@ static int wake_futex_pi(u32 __user *uad pi_state->owner = new_owner; raw_spin_unlock(&new_owner->pi_lock); - postunlock = __rt_mutex_futex_unlock(&pi_state->pi_mutex, &wake_q); - + postunlock = __rt_mutex_futex_unlock(&pi_state->pi_mutex, &wake_q, + &wake_sleeper_q); out_unlock: raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); if (postunlock) - rt_mutex_postunlock(&wake_q); + rt_mutex_postunlock(&wake_q, &wake_sleeper_q); return ret; } @@ -2261,6 +2264,16 @@ retry_private: requeue_pi_wake_futex(this, &key2, hb2); drop_count++; continue; + } else if (ret == -EAGAIN) { + /* + * Waiter was woken by timeout or + * signal and has set pi_blocked_on to + * PI_WAKEUP_INPROGRESS before we + * tried to enqueue it on the rtmutex. + */ + this->pi_state = NULL; + put_pi_state(pi_state); + continue; } else if (ret) { /* * rt_mutex_start_proxy_lock() detected a @@ -2969,7 +2982,7 @@ retry_private: goto no_block; } - rt_mutex_init_waiter(&rt_waiter); + rt_mutex_init_waiter(&rt_waiter, false); /* * On PREEMPT_RT_FULL, when hb->lock becomes an rt_mutex, we must not @@ -2985,6 +2998,14 @@ retry_private: * before __rt_mutex_start_proxy_lock() is done. */ raw_spin_lock_irq(&q.pi_state->pi_mutex.wait_lock); + /* + * the migrate_disable() here disables migration in the in_atomic() fast + * path which is enabled again in the following spin_unlock(). We have + * one migrate_disable() pending in the slow-path which is reversed + * after the raw_spin_unlock_irq() where we leave the atomic context. + */ + migrate_disable(); + spin_unlock(q.lock_ptr); /* * __rt_mutex_start_proxy_lock() unconditionally enqueues the @rt_waiter @@ -2993,6 +3014,7 @@ retry_private: */ ret = __rt_mutex_start_proxy_lock(&q.pi_state->pi_mutex, &rt_waiter, current); raw_spin_unlock_irq(&q.pi_state->pi_mutex.wait_lock); + migrate_enable(); if (ret) { if (ret == 1) @@ -3141,10 +3163,19 @@ retry: * rt_waiter. Also see the WARN in wake_futex_pi(). */ raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); + /* + * Magic trickery for now to make the RT migrate disable + * logic happy. The following spin_unlock() happens with + * interrupts disabled so the internal migrate_enable() + * won't undo the migrate_disable() which was issued when + * locking hb->lock. + */ + migrate_disable(); spin_unlock(&hb->lock); /* drops pi_state->pi_mutex.wait_lock */ ret = wake_futex_pi(uaddr, uval, pi_state); + migrate_enable(); put_pi_state(pi_state); @@ -3316,7 +3347,7 @@ static int futex_wait_requeue_pi(u32 __u struct hrtimer_sleeper timeout, *to; struct futex_pi_state *pi_state = NULL; struct rt_mutex_waiter rt_waiter; - struct futex_hash_bucket *hb; + struct futex_hash_bucket *hb, *hb2; union futex_key key2 = FUTEX_KEY_INIT; struct futex_q q = futex_q_init; int res, ret; @@ -3337,7 +3368,7 @@ static int futex_wait_requeue_pi(u32 __u * The waiter is allocated on our stack, manipulated by the requeue * code while we sleep on uaddr. */ - rt_mutex_init_waiter(&rt_waiter); + rt_mutex_init_waiter(&rt_waiter, false); ret = get_futex_key(uaddr2, flags & FLAGS_SHARED, &key2, FUTEX_WRITE); if (unlikely(ret != 0)) @@ -3368,20 +3399,55 @@ static int futex_wait_requeue_pi(u32 __u /* Queue the futex_q, drop the hb lock, wait for wakeup. */ futex_wait_queue_me(hb, &q, to); - spin_lock(&hb->lock); - ret = handle_early_requeue_pi_wakeup(hb, &q, &key2, to); - spin_unlock(&hb->lock); - if (ret) - goto out_put_keys; + /* + * On RT we must avoid races with requeue and trying to block + * on two mutexes (hb->lock and uaddr2's rtmutex) by + * serializing access to pi_blocked_on with pi_lock. + */ + raw_spin_lock_irq(¤t->pi_lock); + if (current->pi_blocked_on) { + /* + * We have been requeued or are in the process of + * being requeued. + */ + raw_spin_unlock_irq(¤t->pi_lock); + } else { + /* + * Setting pi_blocked_on to PI_WAKEUP_INPROGRESS + * prevents a concurrent requeue from moving us to the + * uaddr2 rtmutex. After that we can safely acquire + * (and possibly block on) hb->lock. + */ + current->pi_blocked_on = PI_WAKEUP_INPROGRESS; + raw_spin_unlock_irq(¤t->pi_lock); + + spin_lock(&hb->lock); + + /* + * Clean up pi_blocked_on. We might leak it otherwise + * when we succeeded with the hb->lock in the fast + * path. + */ + raw_spin_lock_irq(¤t->pi_lock); + current->pi_blocked_on = NULL; + raw_spin_unlock_irq(¤t->pi_lock); + + ret = handle_early_requeue_pi_wakeup(hb, &q, &key2, to); + spin_unlock(&hb->lock); + if (ret) + goto out_put_keys; + } /* - * In order for us to be here, we know our q.key == key2, and since - * we took the hb->lock above, we also know that futex_requeue() has - * completed and we no longer have to concern ourselves with a wakeup - * race with the atomic proxy lock acquisition by the requeue code. The - * futex_requeue dropped our key1 reference and incremented our key2 - * reference count. + * In order to be here, we have either been requeued, are in + * the process of being requeued, or requeue successfully + * acquired uaddr2 on our behalf. If pi_blocked_on was + * non-null above, we may be racing with a requeue. Do not + * rely on q->lock_ptr to be hb2->lock until after blocking on + * hb->lock or hb2->lock. The futex_requeue dropped our key1 + * reference and incremented our key2 reference count. */ + hb2 = hash_futex(&key2); /* Check if the requeue code acquired the second futex for us. */ if (!q.rt_waiter) { @@ -3390,7 +3456,8 @@ static int futex_wait_requeue_pi(u32 __u * did a lock-steal - fix up the PI-state in that case. */ if (q.pi_state && (q.pi_state->owner != current)) { - spin_lock(q.lock_ptr); + spin_lock(&hb2->lock); + BUG_ON(&hb2->lock != q.lock_ptr); ret = fixup_pi_state_owner(uaddr2, &q, current); if (ret && rt_mutex_owner(&q.pi_state->pi_mutex) == current) { pi_state = q.pi_state; @@ -3401,7 +3468,7 @@ static int futex_wait_requeue_pi(u32 __u * the requeue_pi() code acquired for us. */ put_pi_state(q.pi_state); - spin_unlock(q.lock_ptr); + spin_unlock(&hb2->lock); } } else { struct rt_mutex *pi_mutex; @@ -3415,7 +3482,8 @@ static int futex_wait_requeue_pi(u32 __u pi_mutex = &q.pi_state->pi_mutex; ret = rt_mutex_wait_proxy_lock(pi_mutex, to, &rt_waiter); - spin_lock(q.lock_ptr); + spin_lock(&hb2->lock); + BUG_ON(&hb2->lock != q.lock_ptr); if (ret && !rt_mutex_cleanup_proxy_lock(pi_mutex, &rt_waiter)) ret = 0; Index: linux-5.6.17-rt10/kernel/irq/handle.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/irq/handle.c +++ linux-5.6.17-rt10/kernel/irq/handle.c @@ -185,10 +185,16 @@ irqreturn_t handle_irq_event_percpu(stru { irqreturn_t retval; unsigned int flags = 0; + struct pt_regs *regs = get_irq_regs(); + u64 ip = regs ? instruction_pointer(regs) : 0; retval = __handle_irq_event_percpu(desc, &flags); - add_interrupt_randomness(desc->irq_data.irq, flags); +#ifdef CONFIG_PREEMPT_RT + desc->random_ip = ip; +#else + add_interrupt_randomness(desc->irq_data.irq, flags, ip); +#endif if (!noirqdebug) note_interrupt(desc, retval); Index: linux-5.6.17-rt10/kernel/irq/manage.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/irq/manage.c +++ linux-5.6.17-rt10/kernel/irq/manage.c @@ -1138,6 +1138,12 @@ static int irq_thread(void *data) if (action_ret == IRQ_WAKE_THREAD) irq_wake_secondary(desc, action); +#ifdef CONFIG_PREEMPT_RT + migrate_disable(); + add_interrupt_randomness(action->irq, 0, + desc->random_ip ^ (unsigned long) action); + migrate_enable(); +#endif wake_threads_waitq(desc); } @@ -2720,7 +2726,7 @@ EXPORT_SYMBOL_GPL(irq_get_irqchip_state) * This call sets the internal irqchip state of an interrupt, * depending on the value of @which. * - * This function should be called with preemption disabled if the + * This function should be called with migration disabled if the * interrupt controller has per-cpu registers. */ int irq_set_irqchip_state(unsigned int irq, enum irqchip_irq_state which, Index: linux-5.6.17-rt10/kernel/irq/spurious.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/irq/spurious.c +++ linux-5.6.17-rt10/kernel/irq/spurious.c @@ -443,6 +443,10 @@ MODULE_PARM_DESC(noirqdebug, "Disable ir static int __init irqfixup_setup(char *str) { +#ifdef CONFIG_PREEMPT_RT + pr_warn("irqfixup boot option not supported w/ CONFIG_PREEMPT_RT\n"); + return 1; +#endif irqfixup = 1; printk(KERN_WARNING "Misrouted IRQ fixup support enabled.\n"); printk(KERN_WARNING "This may impact system performance.\n"); @@ -455,6 +459,10 @@ module_param(irqfixup, int, 0644); static int __init irqpoll_setup(char *str) { +#ifdef CONFIG_PREEMPT_RT + pr_warn("irqpoll boot option not supported w/ CONFIG_PREEMPT_RT\n"); + return 1; +#endif irqfixup = 2; printk(KERN_WARNING "Misrouted IRQ fixup and polling support " "enabled\n"); Index: linux-5.6.17-rt10/kernel/irq_work.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/irq_work.c +++ linux-5.6.17-rt10/kernel/irq_work.c @@ -18,6 +18,7 @@ #include #include #include +#include #include @@ -52,13 +53,19 @@ void __weak arch_irq_work_raise(void) /* Enqueue on current CPU, work must already be claimed and preempt disabled */ static void __irq_work_queue_local(struct irq_work *work) { + struct llist_head *list; + bool lazy_work, realtime = IS_ENABLED(CONFIG_PREEMPT_RT); + + lazy_work = atomic_read(&work->flags) & IRQ_WORK_LAZY; + /* If the work is "lazy", handle it from next tick if any */ - if (atomic_read(&work->flags) & IRQ_WORK_LAZY) { - if (llist_add(&work->llnode, this_cpu_ptr(&lazy_list)) && - tick_nohz_tick_stopped()) - arch_irq_work_raise(); - } else { - if (llist_add(&work->llnode, this_cpu_ptr(&raised_list))) + if (lazy_work || (realtime && !(atomic_read(&work->flags) & IRQ_WORK_HARD_IRQ))) + list = this_cpu_ptr(&lazy_list); + else + list = this_cpu_ptr(&raised_list); + + if (llist_add(&work->llnode, list)) { + if (!lazy_work || tick_nohz_tick_stopped()) arch_irq_work_raise(); } } @@ -100,9 +107,16 @@ bool irq_work_queue_on(struct irq_work * preempt_disable(); if (cpu != smp_processor_id()) { + struct llist_head *list; + /* Arch remote IPI send/receive backend aren't NMI safe */ WARN_ON_ONCE(in_nmi()); - if (llist_add(&work->llnode, &per_cpu(raised_list, cpu))) + if (IS_ENABLED(CONFIG_PREEMPT_RT) && !(atomic_read(&work->flags) & IRQ_WORK_HARD_IRQ)) + list = &per_cpu(lazy_list, cpu); + else + list = &per_cpu(raised_list, cpu); + + if (llist_add(&work->llnode, list)) arch_send_call_function_single_ipi(cpu); } else { __irq_work_queue_local(work); @@ -121,9 +135,8 @@ bool irq_work_needs_cpu(void) raised = this_cpu_ptr(&raised_list); lazy = this_cpu_ptr(&lazy_list); - if (llist_empty(raised) || arch_irq_work_has_interrupt()) - if (llist_empty(lazy)) - return false; + if (llist_empty(raised) && llist_empty(lazy)) + return false; /* All work should have been flushed before going offline */ WARN_ON_ONCE(cpu_is_offline(smp_processor_id())); @@ -136,8 +149,12 @@ static void irq_work_run_list(struct lli struct irq_work *work, *tmp; struct llist_node *llnode; +#ifndef CONFIG_PREEMPT_RT + /* + * nort: On RT IRQ-work may run in SOFTIRQ context. + */ BUG_ON(!irqs_disabled()); - +#endif if (llist_empty(list)) return; @@ -170,7 +187,16 @@ static void irq_work_run_list(struct lli void irq_work_run(void) { irq_work_run_list(this_cpu_ptr(&raised_list)); - irq_work_run_list(this_cpu_ptr(&lazy_list)); + if (IS_ENABLED(CONFIG_PREEMPT_RT)) { + /* + * NOTE: we raise softirq via IPI for safety, + * and execute in irq_work_tick() to move the + * overhead from hard to soft irq context. + */ + if (!llist_empty(this_cpu_ptr(&lazy_list))) + raise_softirq(TIMER_SOFTIRQ); + } else + irq_work_run_list(this_cpu_ptr(&lazy_list)); } EXPORT_SYMBOL_GPL(irq_work_run); @@ -180,8 +206,17 @@ void irq_work_tick(void) if (!llist_empty(raised) && !arch_irq_work_has_interrupt()) irq_work_run_list(raised); + + if (!IS_ENABLED(CONFIG_PREEMPT_RT)) + irq_work_run_list(this_cpu_ptr(&lazy_list)); +} + +#if defined(CONFIG_IRQ_WORK) && defined(CONFIG_PREEMPT_RT) +void irq_work_tick_soft(void) +{ irq_work_run_list(this_cpu_ptr(&lazy_list)); } +#endif /* * Synchronize against the irq_work @entry, ensures the entry is not Index: linux-5.6.17-rt10/kernel/kexec_core.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/kexec_core.c +++ linux-5.6.17-rt10/kernel/kexec_core.c @@ -978,7 +978,6 @@ void crash_kexec(struct pt_regs *regs) old_cpu = atomic_cmpxchg(&panic_cpu, PANIC_CPU_INVALID, this_cpu); if (old_cpu == PANIC_CPU_INVALID) { /* This is the 1st CPU which comes here, so go ahead. */ - printk_safe_flush_on_panic(); __crash_kexec(regs); /* Index: linux-5.6.17-rt10/kernel/ksysfs.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/ksysfs.c +++ linux-5.6.17-rt10/kernel/ksysfs.c @@ -138,6 +138,15 @@ KERNEL_ATTR_RO(vmcoreinfo); #endif /* CONFIG_CRASH_CORE */ +#if defined(CONFIG_PREEMPT_RT) +static ssize_t realtime_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%d\n", 1); +} +KERNEL_ATTR_RO(realtime); +#endif + /* whether file capabilities are enabled */ static ssize_t fscaps_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) @@ -229,6 +238,9 @@ static struct attribute * kernel_attrs[] &rcu_expedited_attr.attr, &rcu_normal_attr.attr, #endif +#ifdef CONFIG_PREEMPT_RT + &realtime_attr.attr, +#endif NULL }; Index: linux-5.6.17-rt10/kernel/locking/Makefile =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/Makefile +++ linux-5.6.17-rt10/kernel/locking/Makefile @@ -3,7 +3,7 @@ # and is generally not a function of system call inputs. KCOV_INSTRUMENT := n -obj-y += mutex.o semaphore.o rwsem.o percpu-rwsem.o +obj-y += semaphore.o rwsem.o percpu-rwsem.o ifdef CONFIG_FUNCTION_TRACER CFLAGS_REMOVE_lockdep.o = $(CC_FLAGS_FTRACE) @@ -12,19 +12,23 @@ CFLAGS_REMOVE_mutex-debug.o = $(CC_FLAGS CFLAGS_REMOVE_rtmutex-debug.o = $(CC_FLAGS_FTRACE) endif -obj-$(CONFIG_DEBUG_MUTEXES) += mutex-debug.o obj-$(CONFIG_LOCKDEP) += lockdep.o ifeq ($(CONFIG_PROC_FS),y) obj-$(CONFIG_LOCKDEP) += lockdep_proc.o endif obj-$(CONFIG_SMP) += spinlock.o -obj-$(CONFIG_LOCK_SPIN_ON_OWNER) += osq_lock.o obj-$(CONFIG_PROVE_LOCKING) += spinlock.o obj-$(CONFIG_QUEUED_SPINLOCKS) += qspinlock.o obj-$(CONFIG_RT_MUTEXES) += rtmutex.o obj-$(CONFIG_DEBUG_RT_MUTEXES) += rtmutex-debug.o obj-$(CONFIG_DEBUG_SPINLOCK) += spinlock.o obj-$(CONFIG_DEBUG_SPINLOCK) += spinlock_debug.o +ifneq ($(CONFIG_PREEMPT_RT),y) +obj-y += mutex.o +obj-$(CONFIG_LOCK_SPIN_ON_OWNER) += osq_lock.o +obj-$(CONFIG_DEBUG_MUTEXES) += mutex-debug.o +endif +obj-$(CONFIG_PREEMPT_RT) += mutex-rt.o rwsem-rt.o rwlock-rt.o obj-$(CONFIG_QUEUED_RWLOCKS) += qrwlock.o obj-$(CONFIG_LOCK_TORTURE_TEST) += locktorture.o obj-$(CONFIG_WW_MUTEX_SELFTEST) += test-ww_mutex.o Index: linux-5.6.17-rt10/kernel/locking/lockdep.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/lockdep.c +++ linux-5.6.17-rt10/kernel/locking/lockdep.c @@ -4411,6 +4411,7 @@ static void check_flags(unsigned long fl } } +#ifndef CONFIG_PREEMPT_RT /* * We dont accurately track softirq state in e.g. * hardirq contexts (such as on 4KSTACKS), so only @@ -4425,6 +4426,7 @@ static void check_flags(unsigned long fl DEBUG_LOCKS_WARN_ON(!current->softirqs_enabled); } } +#endif if (!debug_locks) print_irqtrace_events(current); Index: linux-5.6.17-rt10/kernel/locking/mutex-rt.c =================================================================== --- /dev/null +++ linux-5.6.17-rt10/kernel/locking/mutex-rt.c @@ -0,0 +1,223 @@ +/* + * kernel/rt.c + * + * Real-Time Preemption Support + * + * started by Ingo Molnar: + * + * Copyright (C) 2004-2006 Red Hat, Inc., Ingo Molnar + * Copyright (C) 2006, Timesys Corp., Thomas Gleixner + * + * historic credit for proving that Linux spinlocks can be implemented via + * RT-aware mutexes goes to many people: The Pmutex project (Dirk Grambow + * and others) who prototyped it on 2.4 and did lots of comparative + * research and analysis; TimeSys, for proving that you can implement a + * fully preemptible kernel via the use of IRQ threading and mutexes; + * Bill Huey for persuasively arguing on lkml that the mutex model is the + * right one; and to MontaVista, who ported pmutexes to 2.6. + * + * This code is a from-scratch implementation and is not based on pmutexes, + * but the idea of converting spinlocks to mutexes is used here too. + * + * lock debugging, locking tree, deadlock detection: + * + * Copyright (C) 2004, LynuxWorks, Inc., Igor Manyilov, Bill Huey + * Released under the General Public License (GPL). + * + * Includes portions of the generic R/W semaphore implementation from: + * + * Copyright (c) 2001 David Howells (dhowells@redhat.com). + * - Derived partially from idea by Andrea Arcangeli + * - Derived also from comments by Linus + * + * Pending ownership of locks and ownership stealing: + * + * Copyright (C) 2005, Kihon Technologies Inc., Steven Rostedt + * + * (also by Steven Rostedt) + * - Converted single pi_lock to individual task locks. + * + * By Esben Nielsen: + * Doing priority inheritance with help of the scheduler. + * + * Copyright (C) 2006, Timesys Corp., Thomas Gleixner + * - major rework based on Esben Nielsens initial patch + * - replaced thread_info references by task_struct refs + * - removed task->pending_owner dependency + * - BKL drop/reacquire for semaphore style locks to avoid deadlocks + * in the scheduler return path as discussed with Steven Rostedt + * + * Copyright (C) 2006, Kihon Technologies Inc. + * Steven Rostedt + * - debugged and patched Thomas Gleixner's rework. + * - added back the cmpxchg to the rework. + * - turned atomic require back on for SMP. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "rtmutex_common.h" + +/* + * struct mutex functions + */ +void __mutex_do_init(struct mutex *mutex, const char *name, + struct lock_class_key *key) +{ +#ifdef CONFIG_DEBUG_LOCK_ALLOC + /* + * Make sure we are not reinitializing a held lock: + */ + debug_check_no_locks_freed((void *)mutex, sizeof(*mutex)); + lockdep_init_map(&mutex->dep_map, name, key, 0); +#endif + mutex->lock.save_state = 0; +} +EXPORT_SYMBOL(__mutex_do_init); + +void __lockfunc _mutex_lock(struct mutex *lock) +{ + mutex_acquire(&lock->dep_map, 0, 0, _RET_IP_); + __rt_mutex_lock_state(&lock->lock, TASK_UNINTERRUPTIBLE); +} +EXPORT_SYMBOL(_mutex_lock); + +void __lockfunc _mutex_lock_io(struct mutex *lock) +{ + int token; + + token = io_schedule_prepare(); + _mutex_lock(lock); + io_schedule_finish(token); +} +EXPORT_SYMBOL_GPL(_mutex_lock_io); + +int __lockfunc _mutex_lock_interruptible(struct mutex *lock) +{ + int ret; + + mutex_acquire(&lock->dep_map, 0, 0, _RET_IP_); + ret = __rt_mutex_lock_state(&lock->lock, TASK_INTERRUPTIBLE); + if (ret) + mutex_release(&lock->dep_map, _RET_IP_); + return ret; +} +EXPORT_SYMBOL(_mutex_lock_interruptible); + +int __lockfunc _mutex_lock_killable(struct mutex *lock) +{ + int ret; + + mutex_acquire(&lock->dep_map, 0, 0, _RET_IP_); + ret = __rt_mutex_lock_state(&lock->lock, TASK_KILLABLE); + if (ret) + mutex_release(&lock->dep_map, _RET_IP_); + return ret; +} +EXPORT_SYMBOL(_mutex_lock_killable); + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +void __lockfunc _mutex_lock_nested(struct mutex *lock, int subclass) +{ + mutex_acquire_nest(&lock->dep_map, subclass, 0, NULL, _RET_IP_); + __rt_mutex_lock_state(&lock->lock, TASK_UNINTERRUPTIBLE); +} +EXPORT_SYMBOL(_mutex_lock_nested); + +void __lockfunc _mutex_lock_io_nested(struct mutex *lock, int subclass) +{ + int token; + + token = io_schedule_prepare(); + + mutex_acquire_nest(&lock->dep_map, subclass, 0, NULL, _RET_IP_); + __rt_mutex_lock_state(&lock->lock, TASK_UNINTERRUPTIBLE); + + io_schedule_finish(token); +} +EXPORT_SYMBOL_GPL(_mutex_lock_io_nested); + +void __lockfunc _mutex_lock_nest_lock(struct mutex *lock, struct lockdep_map *nest) +{ + mutex_acquire_nest(&lock->dep_map, 0, 0, nest, _RET_IP_); + __rt_mutex_lock_state(&lock->lock, TASK_UNINTERRUPTIBLE); +} +EXPORT_SYMBOL(_mutex_lock_nest_lock); + +int __lockfunc _mutex_lock_interruptible_nested(struct mutex *lock, int subclass) +{ + int ret; + + mutex_acquire_nest(&lock->dep_map, subclass, 0, NULL, _RET_IP_); + ret = __rt_mutex_lock_state(&lock->lock, TASK_INTERRUPTIBLE); + if (ret) + mutex_release(&lock->dep_map, _RET_IP_); + return ret; +} +EXPORT_SYMBOL(_mutex_lock_interruptible_nested); + +int __lockfunc _mutex_lock_killable_nested(struct mutex *lock, int subclass) +{ + int ret; + + mutex_acquire(&lock->dep_map, subclass, 0, _RET_IP_); + ret = __rt_mutex_lock_state(&lock->lock, TASK_KILLABLE); + if (ret) + mutex_release(&lock->dep_map, _RET_IP_); + return ret; +} +EXPORT_SYMBOL(_mutex_lock_killable_nested); +#endif + +int __lockfunc _mutex_trylock(struct mutex *lock) +{ + int ret = __rt_mutex_trylock(&lock->lock); + + if (ret) + mutex_acquire(&lock->dep_map, 0, 1, _RET_IP_); + + return ret; +} +EXPORT_SYMBOL(_mutex_trylock); + +void __lockfunc _mutex_unlock(struct mutex *lock) +{ + mutex_release(&lock->dep_map, _RET_IP_); + __rt_mutex_unlock(&lock->lock); +} +EXPORT_SYMBOL(_mutex_unlock); + +/** + * atomic_dec_and_mutex_lock - return holding mutex if we dec to 0 + * @cnt: the atomic which we are to dec + * @lock: the mutex to return holding if we dec to 0 + * + * return true and hold lock if we dec to 0, return false otherwise + */ +int atomic_dec_and_mutex_lock(atomic_t *cnt, struct mutex *lock) +{ + /* dec if we can't possibly hit 0 */ + if (atomic_add_unless(cnt, -1, 1)) + return 0; + /* we might hit 0, so take the lock */ + mutex_lock(lock); + if (!atomic_dec_and_test(cnt)) { + /* when we actually did the dec, we didn't hit 0 */ + mutex_unlock(lock); + return 0; + } + /* we hit 0, and we hold the lock */ + return 1; +} +EXPORT_SYMBOL(atomic_dec_and_mutex_lock); Index: linux-5.6.17-rt10/kernel/locking/percpu-rwsem.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/percpu-rwsem.c +++ linux-5.6.17-rt10/kernel/locking/percpu-rwsem.c @@ -1,27 +1,29 @@ // SPDX-License-Identifier: GPL-2.0-only #include -#include #include +#include #include #include #include #include +#include #include -#include "rwsem.h" - int __percpu_init_rwsem(struct percpu_rw_semaphore *sem, - const char *name, struct lock_class_key *rwsem_key) + const char *name, struct lock_class_key *key) { sem->read_count = alloc_percpu(int); if (unlikely(!sem->read_count)) return -ENOMEM; - /* ->rw_sem represents the whole percpu_rw_semaphore for lockdep */ rcu_sync_init(&sem->rss); - __init_rwsem(&sem->rw_sem, name, rwsem_key); rcuwait_init(&sem->writer); - sem->readers_block = 0; + init_waitqueue_head(&sem->waiters); + atomic_set(&sem->block, 0); +#ifdef CONFIG_DEBUG_LOCK_ALLOC + debug_check_no_locks_freed((void *)sem, sizeof(*sem)); + lockdep_init_map(&sem->dep_map, name, key, 0); +#endif return 0; } EXPORT_SYMBOL_GPL(__percpu_init_rwsem); @@ -41,73 +43,139 @@ void percpu_free_rwsem(struct percpu_rw_ } EXPORT_SYMBOL_GPL(percpu_free_rwsem); -int __percpu_down_read(struct percpu_rw_semaphore *sem, int try) +static bool __percpu_down_read_trylock(struct percpu_rw_semaphore *sem) { + __this_cpu_inc(*sem->read_count); + /* * Due to having preemption disabled the decrement happens on * the same CPU as the increment, avoiding the * increment-on-one-CPU-and-decrement-on-another problem. * - * If the reader misses the writer's assignment of readers_block, then - * the writer is guaranteed to see the reader's increment. + * If the reader misses the writer's assignment of sem->block, then the + * writer is guaranteed to see the reader's increment. * * Conversely, any readers that increment their sem->read_count after - * the writer looks are guaranteed to see the readers_block value, - * which in turn means that they are guaranteed to immediately - * decrement their sem->read_count, so that it doesn't matter that the - * writer missed them. + * the writer looks are guaranteed to see the sem->block value, which + * in turn means that they are guaranteed to immediately decrement + * their sem->read_count, so that it doesn't matter that the writer + * missed them. */ smp_mb(); /* A matches D */ /* - * If !readers_block the critical section starts here, matched by the + * If !sem->block the critical section starts here, matched by the * release in percpu_up_write(). */ - if (likely(!smp_load_acquire(&sem->readers_block))) - return 1; + if (likely(!atomic_read_acquire(&sem->block))) + return true; - /* - * Per the above comment; we still have preemption disabled and - * will thus decrement on the same CPU as we incremented. - */ - __percpu_up_read(sem); + __this_cpu_dec(*sem->read_count); - if (try) - return 0; + /* Prod writer to re-evaluate readers_active_check() */ + rcuwait_wake_up(&sem->writer); - /* - * We either call schedule() in the wait, or we'll fall through - * and reschedule on the preempt_enable() in percpu_down_read(). - */ - preempt_enable_no_resched(); + return false; +} - /* - * Avoid lockdep for the down/up_read() we already have them. - */ - __down_read(&sem->rw_sem); - this_cpu_inc(*sem->read_count); - __up_read(&sem->rw_sem); +static inline bool __percpu_down_write_trylock(struct percpu_rw_semaphore *sem) +{ + if (atomic_read(&sem->block)) + return false; - preempt_disable(); - return 1; + return atomic_xchg(&sem->block, 1) == 0; +} + +static bool __percpu_rwsem_trylock(struct percpu_rw_semaphore *sem, bool reader) +{ + if (reader) { + bool ret; + + preempt_disable(); + ret = __percpu_down_read_trylock(sem); + preempt_enable(); + + return ret; + } + return __percpu_down_write_trylock(sem); +} + +/* + * The return value of wait_queue_entry::func means: + * + * <0 - error, wakeup is terminated and the error is returned + * 0 - no wakeup, a next waiter is tried + * >0 - woken, if EXCLUSIVE, counted towards @nr_exclusive. + * + * We use EXCLUSIVE for both readers and writers to preserve FIFO order, + * and play games with the return value to allow waking multiple readers. + * + * Specifically, we wake readers until we've woken a single writer, or until a + * trylock fails. + */ +static int percpu_rwsem_wake_function(struct wait_queue_entry *wq_entry, + unsigned int mode, int wake_flags, + void *key) +{ + struct task_struct *p = get_task_struct(wq_entry->private); + bool reader = wq_entry->flags & WQ_FLAG_CUSTOM; + struct percpu_rw_semaphore *sem = key; + + /* concurrent against percpu_down_write(), can get stolen */ + if (!__percpu_rwsem_trylock(sem, reader)) + return 1; + + list_del_init(&wq_entry->entry); + smp_store_release(&wq_entry->private, NULL); + + wake_up_process(p); + put_task_struct(p); + + return !reader; /* wake (readers until) 1 writer */ } -EXPORT_SYMBOL_GPL(__percpu_down_read); -void __percpu_up_read(struct percpu_rw_semaphore *sem) +static void percpu_rwsem_wait(struct percpu_rw_semaphore *sem, bool reader) { - smp_mb(); /* B matches C */ + DEFINE_WAIT_FUNC(wq_entry, percpu_rwsem_wake_function); + bool wait; + + spin_lock_irq(&sem->waiters.lock); /* - * In other words, if they see our decrement (presumably to aggregate - * zero, as that is the only time it matters) they will also see our - * critical section. + * Serialize against the wakeup in percpu_up_write(), if we fail + * the trylock, the wakeup must see us on the list. */ - __this_cpu_dec(*sem->read_count); + wait = !__percpu_rwsem_trylock(sem, reader); + if (wait) { + wq_entry.flags |= WQ_FLAG_EXCLUSIVE | reader * WQ_FLAG_CUSTOM; + __add_wait_queue_entry_tail(&sem->waiters, &wq_entry); + } + spin_unlock_irq(&sem->waiters.lock); + + while (wait) { + set_current_state(TASK_UNINTERRUPTIBLE); + if (!smp_load_acquire(&wq_entry.private)) + break; + schedule(); + } + __set_current_state(TASK_RUNNING); +} - /* Prod writer to recheck readers_active */ - rcuwait_wake_up(&sem->writer); +bool __percpu_down_read(struct percpu_rw_semaphore *sem, bool try) +{ + if (__percpu_down_read_trylock(sem)) + return true; + + if (try) + return false; + + preempt_enable(); + percpu_rwsem_wait(sem, /* .reader = */ true); + preempt_disable(); + + return true; } -EXPORT_SYMBOL_GPL(__percpu_up_read); +EXPORT_SYMBOL_GPL(__percpu_down_read); #define per_cpu_sum(var) \ ({ \ @@ -124,6 +192,8 @@ EXPORT_SYMBOL_GPL(__percpu_up_read); * zero. If this sum is zero, then it is stable due to the fact that if any * newly arriving readers increment a given counter, they will immediately * decrement that same counter. + * + * Assumes sem->block is set. */ static bool readers_active_check(struct percpu_rw_semaphore *sem) { @@ -142,32 +212,36 @@ static bool readers_active_check(struct void percpu_down_write(struct percpu_rw_semaphore *sem) { + might_sleep(); + rwsem_acquire(&sem->dep_map, 0, 0, _RET_IP_); + /* Notify readers to take the slow path. */ rcu_sync_enter(&sem->rss); - down_write(&sem->rw_sem); - /* - * Notify new readers to block; up until now, and thus throughout the - * longish rcu_sync_enter() above, new readers could still come in. + * Try set sem->block; this provides writer-writer exclusion. + * Having sem->block set makes new readers block. */ - WRITE_ONCE(sem->readers_block, 1); + if (!__percpu_down_write_trylock(sem)) + percpu_rwsem_wait(sem, /* .reader = */ false); - smp_mb(); /* D matches A */ + /* smp_mb() implied by __percpu_down_write_trylock() on success -- D matches A */ /* - * If they don't see our writer of readers_block, then we are - * guaranteed to see their sem->read_count increment, and therefore - * will wait for them. + * If they don't see our store of sem->block, then we are guaranteed to + * see their sem->read_count increment, and therefore will wait for + * them. */ - /* Wait for all now active readers to complete. */ + /* Wait for all active readers to complete. */ rcuwait_wait_event(&sem->writer, readers_active_check(sem)); } EXPORT_SYMBOL_GPL(percpu_down_write); void percpu_up_write(struct percpu_rw_semaphore *sem) { + rwsem_release(&sem->dep_map, _RET_IP_); + /* * Signal the writer is done, no fast path yet. * @@ -178,12 +252,12 @@ void percpu_up_write(struct percpu_rw_se * Therefore we force it through the slow path which guarantees an * acquire and thereby guarantees the critical section's consistency. */ - smp_store_release(&sem->readers_block, 0); + atomic_set_release(&sem->block, 0); /* - * Release the write lock, this will allow readers back in the game. + * Prod any pending reader/writer to make progress. */ - up_write(&sem->rw_sem); + __wake_up(&sem->waiters, TASK_NORMAL, 1, sem); /* * Once this completes (at least one RCU-sched grace period hence) the Index: linux-5.6.17-rt10/kernel/locking/rtmutex.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/rtmutex.c +++ linux-5.6.17-rt10/kernel/locking/rtmutex.c @@ -8,6 +8,11 @@ * Copyright (C) 2005-2006 Timesys Corp., Thomas Gleixner * Copyright (C) 2005 Kihon Technologies Inc., Steven Rostedt * Copyright (C) 2006 Esben Nielsen + * Adaptive Spinlocks: + * Copyright (C) 2008 Novell, Inc., Gregory Haskins, Sven Dietrich, + * and Peter Morreale, + * Adaptive Spinlocks simplification: + * Copyright (C) 2008 Red Hat, Inc., Steven Rostedt * * See Documentation/locking/rt-mutex-design.rst for details. */ @@ -19,6 +24,8 @@ #include #include #include +#include +#include #include "rtmutex_common.h" @@ -136,6 +143,12 @@ static void fixup_rt_mutex_waiters(struc WRITE_ONCE(*p, owner & ~RT_MUTEX_HAS_WAITERS); } +static int rt_mutex_real_waiter(struct rt_mutex_waiter *waiter) +{ + return waiter && waiter != PI_WAKEUP_INPROGRESS && + waiter != PI_REQUEUE_INPROGRESS; +} + /* * We can speed up the acquire/release, if there's no debugging state to be * set up. @@ -229,7 +242,7 @@ static inline bool unlock_rt_mutex_safe( * Only use with rt_mutex_waiter_{less,equal}() */ #define task_to_waiter(p) \ - &(struct rt_mutex_waiter){ .prio = (p)->prio, .deadline = (p)->dl.deadline } + &(struct rt_mutex_waiter){ .prio = (p)->prio, .deadline = (p)->dl.deadline, .task = (p) } static inline int rt_mutex_waiter_less(struct rt_mutex_waiter *left, @@ -269,6 +282,27 @@ rt_mutex_waiter_equal(struct rt_mutex_wa return 1; } +#define STEAL_NORMAL 0 +#define STEAL_LATERAL 1 + +static inline int +rt_mutex_steal(struct rt_mutex *lock, struct rt_mutex_waiter *waiter, int mode) +{ + struct rt_mutex_waiter *top_waiter = rt_mutex_top_waiter(lock); + + if (waiter == top_waiter || rt_mutex_waiter_less(waiter, top_waiter)) + return 1; + + /* + * Note that RT tasks are excluded from lateral-steals + * to prevent the introduction of an unbounded latency. + */ + if (mode == STEAL_NORMAL || rt_task(waiter->task)) + return 0; + + return rt_mutex_waiter_equal(waiter, top_waiter); +} + static void rt_mutex_enqueue(struct rt_mutex *lock, struct rt_mutex_waiter *waiter) { @@ -373,6 +407,14 @@ static bool rt_mutex_cond_detect_deadloc return debug_rt_mutex_detect_deadlock(waiter, chwalk); } +static void rt_mutex_wake_waiter(struct rt_mutex_waiter *waiter) +{ + if (waiter->savestate) + wake_up_lock_sleeper(waiter->task); + else + wake_up_process(waiter->task); +} + /* * Max number of times we'll walk the boosting chain: */ @@ -380,7 +422,8 @@ int max_lock_depth = 1024; static inline struct rt_mutex *task_blocked_on_lock(struct task_struct *p) { - return p->pi_blocked_on ? p->pi_blocked_on->lock : NULL; + return rt_mutex_real_waiter(p->pi_blocked_on) ? + p->pi_blocked_on->lock : NULL; } /* @@ -516,7 +559,7 @@ static int rt_mutex_adjust_prio_chain(st * reached or the state of the chain has changed while we * dropped the locks. */ - if (!waiter) + if (!rt_mutex_real_waiter(waiter)) goto out_unlock_pi; /* @@ -696,13 +739,16 @@ static int rt_mutex_adjust_prio_chain(st * follow here. This is the end of the chain we are walking. */ if (!rt_mutex_owner(lock)) { + struct rt_mutex_waiter *lock_top_waiter; + /* * If the requeue [7] above changed the top waiter, * then we need to wake the new top waiter up to try * to get the lock. */ - if (prerequeue_top_waiter != rt_mutex_top_waiter(lock)) - wake_up_process(rt_mutex_top_waiter(lock)->task); + lock_top_waiter = rt_mutex_top_waiter(lock); + if (prerequeue_top_waiter != lock_top_waiter) + rt_mutex_wake_waiter(lock_top_waiter); raw_spin_unlock_irq(&lock->wait_lock); return 0; } @@ -803,9 +849,11 @@ static int rt_mutex_adjust_prio_chain(st * @task: The task which wants to acquire the lock * @waiter: The waiter that is queued to the lock's wait tree if the * callsite called task_blocked_on_lock(), otherwise NULL + * @mode: Lock steal mode (STEAL_NORMAL, STEAL_LATERAL) */ -static int try_to_take_rt_mutex(struct rt_mutex *lock, struct task_struct *task, - struct rt_mutex_waiter *waiter) +static int __try_to_take_rt_mutex(struct rt_mutex *lock, + struct task_struct *task, + struct rt_mutex_waiter *waiter, int mode) { lockdep_assert_held(&lock->wait_lock); @@ -841,12 +889,11 @@ static int try_to_take_rt_mutex(struct r */ if (waiter) { /* - * If waiter is not the highest priority waiter of - * @lock, give up. + * If waiter is not the highest priority waiter of @lock, + * or its peer when lateral steal is allowed, give up. */ - if (waiter != rt_mutex_top_waiter(lock)) + if (!rt_mutex_steal(lock, waiter, mode)) return 0; - /* * We can acquire the lock. Remove the waiter from the * lock waiters tree. @@ -864,14 +911,12 @@ static int try_to_take_rt_mutex(struct r */ if (rt_mutex_has_waiters(lock)) { /* - * If @task->prio is greater than or equal to - * the top waiter priority (kernel view), - * @task lost. + * If @task->prio is greater than the top waiter + * priority (kernel view), or equal to it when a + * lateral steal is forbidden, @task lost. */ - if (!rt_mutex_waiter_less(task_to_waiter(task), - rt_mutex_top_waiter(lock))) + if (!rt_mutex_steal(lock, task_to_waiter(task), mode)) return 0; - /* * The current top waiter stays enqueued. We * don't have to change anything in the lock @@ -918,6 +963,344 @@ takeit: return 1; } +#ifdef CONFIG_PREEMPT_RT +/* + * preemptible spin_lock functions: + */ +static inline void rt_spin_lock_fastlock(struct rt_mutex *lock, + void (*slowfn)(struct rt_mutex *lock)) +{ + might_sleep_no_state_check(); + + if (likely(rt_mutex_cmpxchg_acquire(lock, NULL, current))) + return; + else + slowfn(lock); +} + +static inline void rt_spin_lock_fastunlock(struct rt_mutex *lock, + void (*slowfn)(struct rt_mutex *lock)) +{ + if (likely(rt_mutex_cmpxchg_release(lock, current, NULL))) + return; + else + slowfn(lock); +} +#ifdef CONFIG_SMP +/* + * Note that owner is a speculative pointer and dereferencing relies + * on rcu_read_lock() and the check against the lock owner. + */ +static int adaptive_wait(struct rt_mutex *lock, + struct task_struct *owner) +{ + int res = 0; + + rcu_read_lock(); + for (;;) { + if (owner != rt_mutex_owner(lock)) + break; + /* + * Ensure that owner->on_cpu is dereferenced _after_ + * checking the above to be valid. + */ + barrier(); + if (!owner->on_cpu) { + res = 1; + break; + } + cpu_relax(); + } + rcu_read_unlock(); + return res; +} +#else +static int adaptive_wait(struct rt_mutex *lock, + struct task_struct *orig_owner) +{ + return 1; +} +#endif + +static int task_blocks_on_rt_mutex(struct rt_mutex *lock, + struct rt_mutex_waiter *waiter, + struct task_struct *task, + enum rtmutex_chainwalk chwalk); +/* + * Slow path lock function spin_lock style: this variant is very + * careful not to miss any non-lock wakeups. + * + * We store the current state under p->pi_lock in p->saved_state and + * the try_to_wake_up() code handles this accordingly. + */ +void __sched rt_spin_lock_slowlock_locked(struct rt_mutex *lock, + struct rt_mutex_waiter *waiter, + unsigned long flags) +{ + struct task_struct *lock_owner, *self = current; + struct rt_mutex_waiter *top_waiter; + int ret; + + if (__try_to_take_rt_mutex(lock, self, NULL, STEAL_LATERAL)) + return; + + BUG_ON(rt_mutex_owner(lock) == self); + + /* + * We save whatever state the task is in and we'll restore it + * after acquiring the lock taking real wakeups into account + * as well. We are serialized via pi_lock against wakeups. See + * try_to_wake_up(). + */ + raw_spin_lock(&self->pi_lock); + self->saved_state = self->state; + __set_current_state_no_track(TASK_UNINTERRUPTIBLE); + raw_spin_unlock(&self->pi_lock); + + ret = task_blocks_on_rt_mutex(lock, waiter, self, RT_MUTEX_MIN_CHAINWALK); + BUG_ON(ret); + + for (;;) { + /* Try to acquire the lock again. */ + if (__try_to_take_rt_mutex(lock, self, waiter, STEAL_LATERAL)) + break; + + top_waiter = rt_mutex_top_waiter(lock); + lock_owner = rt_mutex_owner(lock); + + raw_spin_unlock_irqrestore(&lock->wait_lock, flags); + + debug_rt_mutex_print_deadlock(waiter); + + if (top_waiter != waiter || adaptive_wait(lock, lock_owner)) + schedule(); + + raw_spin_lock_irqsave(&lock->wait_lock, flags); + + raw_spin_lock(&self->pi_lock); + __set_current_state_no_track(TASK_UNINTERRUPTIBLE); + raw_spin_unlock(&self->pi_lock); + } + + /* + * Restore the task state to current->saved_state. We set it + * to the original state above and the try_to_wake_up() code + * has possibly updated it when a real (non-rtmutex) wakeup + * happened while we were blocked. Clear saved_state so + * try_to_wakeup() does not get confused. + */ + raw_spin_lock(&self->pi_lock); + __set_current_state_no_track(self->saved_state); + self->saved_state = TASK_RUNNING; + raw_spin_unlock(&self->pi_lock); + + /* + * try_to_take_rt_mutex() sets the waiter bit + * unconditionally. We might have to fix that up: + */ + fixup_rt_mutex_waiters(lock); + + BUG_ON(rt_mutex_has_waiters(lock) && waiter == rt_mutex_top_waiter(lock)); + BUG_ON(!RB_EMPTY_NODE(&waiter->tree_entry)); +} + +static void noinline __sched rt_spin_lock_slowlock(struct rt_mutex *lock) +{ + struct rt_mutex_waiter waiter; + unsigned long flags; + + rt_mutex_init_waiter(&waiter, true); + + raw_spin_lock_irqsave(&lock->wait_lock, flags); + rt_spin_lock_slowlock_locked(lock, &waiter, flags); + raw_spin_unlock_irqrestore(&lock->wait_lock, flags); + debug_rt_mutex_free_waiter(&waiter); +} + +static bool __sched __rt_mutex_unlock_common(struct rt_mutex *lock, + struct wake_q_head *wake_q, + struct wake_q_head *wq_sleeper); +/* + * Slow path to release a rt_mutex spin_lock style + */ +void __sched rt_spin_lock_slowunlock(struct rt_mutex *lock) +{ + unsigned long flags; + DEFINE_WAKE_Q(wake_q); + DEFINE_WAKE_Q(wake_sleeper_q); + bool postunlock; + + raw_spin_lock_irqsave(&lock->wait_lock, flags); + postunlock = __rt_mutex_unlock_common(lock, &wake_q, &wake_sleeper_q); + raw_spin_unlock_irqrestore(&lock->wait_lock, flags); + + if (postunlock) + rt_mutex_postunlock(&wake_q, &wake_sleeper_q); +} + +void __lockfunc rt_spin_lock(spinlock_t *lock) +{ + sleeping_lock_inc(); + rcu_read_lock(); + migrate_disable(); + spin_acquire(&lock->dep_map, 0, 0, _RET_IP_); + rt_spin_lock_fastlock(&lock->lock, rt_spin_lock_slowlock); +} +EXPORT_SYMBOL(rt_spin_lock); + +void __lockfunc __rt_spin_lock(struct rt_mutex *lock) +{ + rt_spin_lock_fastlock(lock, rt_spin_lock_slowlock); +} + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +void __lockfunc rt_spin_lock_nested(spinlock_t *lock, int subclass) +{ + sleeping_lock_inc(); + rcu_read_lock(); + migrate_disable(); + spin_acquire(&lock->dep_map, subclass, 0, _RET_IP_); + rt_spin_lock_fastlock(&lock->lock, rt_spin_lock_slowlock); +} +EXPORT_SYMBOL(rt_spin_lock_nested); +#endif + +void __lockfunc rt_spin_unlock(spinlock_t *lock) +{ + /* NOTE: we always pass in '1' for nested, for simplicity */ + spin_release(&lock->dep_map, _RET_IP_); + rt_spin_lock_fastunlock(&lock->lock, rt_spin_lock_slowunlock); + migrate_enable(); + rcu_read_unlock(); + sleeping_lock_dec(); +} +EXPORT_SYMBOL(rt_spin_unlock); + +void __lockfunc __rt_spin_unlock(struct rt_mutex *lock) +{ + rt_spin_lock_fastunlock(lock, rt_spin_lock_slowunlock); +} +EXPORT_SYMBOL(__rt_spin_unlock); + +/* + * Wait for the lock to get unlocked: instead of polling for an unlock + * (like raw spinlocks do), we lock and unlock, to force the kernel to + * schedule if there's contention: + */ +void __lockfunc rt_spin_lock_unlock(spinlock_t *lock) +{ + spin_lock(lock); + spin_unlock(lock); +} +EXPORT_SYMBOL(rt_spin_lock_unlock); + +int __lockfunc rt_spin_trylock(spinlock_t *lock) +{ + int ret; + + sleeping_lock_inc(); + migrate_disable(); + ret = __rt_mutex_trylock(&lock->lock); + if (ret) { + spin_acquire(&lock->dep_map, 0, 1, _RET_IP_); + rcu_read_lock(); + } else { + migrate_enable(); + sleeping_lock_dec(); + } + return ret; +} +EXPORT_SYMBOL(rt_spin_trylock); + +int __lockfunc rt_spin_trylock_bh(spinlock_t *lock) +{ + int ret; + + local_bh_disable(); + ret = __rt_mutex_trylock(&lock->lock); + if (ret) { + sleeping_lock_inc(); + rcu_read_lock(); + migrate_disable(); + spin_acquire(&lock->dep_map, 0, 1, _RET_IP_); + } else + local_bh_enable(); + return ret; +} +EXPORT_SYMBOL(rt_spin_trylock_bh); + +int __lockfunc rt_spin_trylock_irqsave(spinlock_t *lock, unsigned long *flags) +{ + int ret; + + *flags = 0; + ret = __rt_mutex_trylock(&lock->lock); + if (ret) { + sleeping_lock_inc(); + rcu_read_lock(); + migrate_disable(); + spin_acquire(&lock->dep_map, 0, 1, _RET_IP_); + } + return ret; +} +EXPORT_SYMBOL(rt_spin_trylock_irqsave); + +void +__rt_spin_lock_init(spinlock_t *lock, const char *name, struct lock_class_key *key) +{ +#ifdef CONFIG_DEBUG_LOCK_ALLOC + /* + * Make sure we are not reinitializing a held lock: + */ + debug_check_no_locks_freed((void *)lock, sizeof(*lock)); + lockdep_init_map(&lock->dep_map, name, key, 0); +#endif +} +EXPORT_SYMBOL(__rt_spin_lock_init); + +#endif /* PREEMPT_RT */ + +#ifdef CONFIG_PREEMPT_RT + static inline int __sched +__mutex_lock_check_stamp(struct rt_mutex *lock, struct ww_acquire_ctx *ctx) +{ + struct ww_mutex *ww = container_of(lock, struct ww_mutex, base.lock); + struct ww_acquire_ctx *hold_ctx = READ_ONCE(ww->ctx); + + if (!hold_ctx) + return 0; + + if (unlikely(ctx == hold_ctx)) + return -EALREADY; + + if (ctx->stamp - hold_ctx->stamp <= LONG_MAX && + (ctx->stamp != hold_ctx->stamp || ctx > hold_ctx)) { +#ifdef CONFIG_DEBUG_MUTEXES + DEBUG_LOCKS_WARN_ON(ctx->contending_lock); + ctx->contending_lock = ww; +#endif + return -EDEADLK; + } + + return 0; +} +#else + static inline int __sched +__mutex_lock_check_stamp(struct rt_mutex *lock, struct ww_acquire_ctx *ctx) +{ + BUG(); + return 0; +} + +#endif + +static inline int +try_to_take_rt_mutex(struct rt_mutex *lock, struct task_struct *task, + struct rt_mutex_waiter *waiter) +{ + return __try_to_take_rt_mutex(lock, task, waiter, STEAL_NORMAL); +} + /* * Task blocks on lock. * @@ -950,6 +1333,22 @@ static int task_blocks_on_rt_mutex(struc return -EDEADLK; raw_spin_lock(&task->pi_lock); + /* + * In the case of futex requeue PI, this will be a proxy + * lock. The task will wake unaware that it is enqueueed on + * this lock. Avoid blocking on two locks and corrupting + * pi_blocked_on via the PI_WAKEUP_INPROGRESS + * flag. futex_wait_requeue_pi() sets this when it wakes up + * before requeue (due to a signal or timeout). Do not enqueue + * the task if PI_WAKEUP_INPROGRESS is set. + */ + if (task != current && task->pi_blocked_on == PI_WAKEUP_INPROGRESS) { + raw_spin_unlock(&task->pi_lock); + return -EAGAIN; + } + + BUG_ON(rt_mutex_real_waiter(task->pi_blocked_on)); + waiter->task = task; waiter->lock = lock; waiter->prio = task->prio; @@ -973,7 +1372,7 @@ static int task_blocks_on_rt_mutex(struc rt_mutex_enqueue_pi(owner, waiter); rt_mutex_adjust_prio(owner); - if (owner->pi_blocked_on) + if (rt_mutex_real_waiter(owner->pi_blocked_on)) chain_walk = 1; } else if (rt_mutex_cond_detect_deadlock(waiter, chwalk)) { chain_walk = 1; @@ -1015,6 +1414,7 @@ static int task_blocks_on_rt_mutex(struc * Called with lock->wait_lock held and interrupts disabled. */ static void mark_wakeup_next_waiter(struct wake_q_head *wake_q, + struct wake_q_head *wake_sleeper_q, struct rt_mutex *lock) { struct rt_mutex_waiter *waiter; @@ -1054,7 +1454,10 @@ static void mark_wakeup_next_waiter(stru * Pairs with preempt_enable() in rt_mutex_postunlock(); */ preempt_disable(); - wake_q_add(wake_q, waiter->task); + if (waiter->savestate) + wake_q_add_sleeper(wake_sleeper_q, waiter->task); + else + wake_q_add(wake_q, waiter->task); raw_spin_unlock(¤t->pi_lock); } @@ -1069,7 +1472,7 @@ static void remove_waiter(struct rt_mute { bool is_top_waiter = (waiter == rt_mutex_top_waiter(lock)); struct task_struct *owner = rt_mutex_owner(lock); - struct rt_mutex *next_lock; + struct rt_mutex *next_lock = NULL; lockdep_assert_held(&lock->wait_lock); @@ -1095,7 +1498,8 @@ static void remove_waiter(struct rt_mute rt_mutex_adjust_prio(owner); /* Store the lock on which owner is blocked or NULL */ - next_lock = task_blocked_on_lock(owner); + if (rt_mutex_real_waiter(owner->pi_blocked_on)) + next_lock = task_blocked_on_lock(owner); raw_spin_unlock(&owner->pi_lock); @@ -1131,26 +1535,28 @@ void rt_mutex_adjust_pi(struct task_stru raw_spin_lock_irqsave(&task->pi_lock, flags); waiter = task->pi_blocked_on; - if (!waiter || rt_mutex_waiter_equal(waiter, task_to_waiter(task))) { + if (!rt_mutex_real_waiter(waiter) || + rt_mutex_waiter_equal(waiter, task_to_waiter(task))) { raw_spin_unlock_irqrestore(&task->pi_lock, flags); return; } next_lock = waiter->lock; - raw_spin_unlock_irqrestore(&task->pi_lock, flags); /* gets dropped in rt_mutex_adjust_prio_chain()! */ get_task_struct(task); + raw_spin_unlock_irqrestore(&task->pi_lock, flags); rt_mutex_adjust_prio_chain(task, RT_MUTEX_MIN_CHAINWALK, NULL, next_lock, NULL, task); } -void rt_mutex_init_waiter(struct rt_mutex_waiter *waiter) +void rt_mutex_init_waiter(struct rt_mutex_waiter *waiter, bool savestate) { debug_rt_mutex_init_waiter(waiter); RB_CLEAR_NODE(&waiter->pi_tree_entry); RB_CLEAR_NODE(&waiter->tree_entry); waiter->task = NULL; + waiter->savestate = savestate; } /** @@ -1166,7 +1572,8 @@ void rt_mutex_init_waiter(struct rt_mute static int __sched __rt_mutex_slowlock(struct rt_mutex *lock, int state, struct hrtimer_sleeper *timeout, - struct rt_mutex_waiter *waiter) + struct rt_mutex_waiter *waiter, + struct ww_acquire_ctx *ww_ctx) { int ret = 0; @@ -1175,16 +1582,17 @@ __rt_mutex_slowlock(struct rt_mutex *loc if (try_to_take_rt_mutex(lock, current, waiter)) break; - /* - * TASK_INTERRUPTIBLE checks for signals and - * timeout. Ignored otherwise. - */ - if (likely(state == TASK_INTERRUPTIBLE)) { - /* Signal pending? */ - if (signal_pending(current)) - ret = -EINTR; - if (timeout && !timeout->task) - ret = -ETIMEDOUT; + if (timeout && !timeout->task) { + ret = -ETIMEDOUT; + break; + } + if (signal_pending_state(state, current)) { + ret = -EINTR; + break; + } + + if (ww_ctx && ww_ctx->acquired > 0) { + ret = __mutex_lock_check_stamp(lock, ww_ctx); if (ret) break; } @@ -1223,33 +1631,104 @@ static void rt_mutex_handle_deadlock(int } } -/* - * Slow path lock function: - */ -static int __sched -rt_mutex_slowlock(struct rt_mutex *lock, int state, - struct hrtimer_sleeper *timeout, - enum rtmutex_chainwalk chwalk) +static __always_inline void ww_mutex_lock_acquired(struct ww_mutex *ww, + struct ww_acquire_ctx *ww_ctx) { - struct rt_mutex_waiter waiter; - unsigned long flags; - int ret = 0; +#ifdef CONFIG_DEBUG_MUTEXES + /* + * If this WARN_ON triggers, you used ww_mutex_lock to acquire, + * but released with a normal mutex_unlock in this call. + * + * This should never happen, always use ww_mutex_unlock. + */ + DEBUG_LOCKS_WARN_ON(ww->ctx); - rt_mutex_init_waiter(&waiter); + /* + * Not quite done after calling ww_acquire_done() ? + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->done_acquire); + + if (ww_ctx->contending_lock) { + /* + * After -EDEADLK you tried to + * acquire a different ww_mutex? Bad! + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->contending_lock != ww); + + /* + * You called ww_mutex_lock after receiving -EDEADLK, + * but 'forgot' to unlock everything else first? + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->acquired > 0); + ww_ctx->contending_lock = NULL; + } /* - * Technically we could use raw_spin_[un]lock_irq() here, but this can - * be called in early boot if the cmpxchg() fast path is disabled - * (debug, no architecture support). In this case we will acquire the - * rtmutex with lock->wait_lock held. But we cannot unconditionally - * enable interrupts in that early boot case. So we need to use the - * irqsave/restore variants. + * Naughty, using a different class will lead to undefined behavior! */ - raw_spin_lock_irqsave(&lock->wait_lock, flags); + DEBUG_LOCKS_WARN_ON(ww_ctx->ww_class != ww->ww_class); +#endif + ww_ctx->acquired++; +} + +#ifdef CONFIG_PREEMPT_RT +static void ww_mutex_account_lock(struct rt_mutex *lock, + struct ww_acquire_ctx *ww_ctx) +{ + struct ww_mutex *ww = container_of(lock, struct ww_mutex, base.lock); + struct rt_mutex_waiter *waiter, *n; + + /* + * This branch gets optimized out for the common case, + * and is only important for ww_mutex_lock. + */ + ww_mutex_lock_acquired(ww, ww_ctx); + ww->ctx = ww_ctx; + + /* + * Give any possible sleeping processes the chance to wake up, + * so they can recheck if they have to back off. + */ + rbtree_postorder_for_each_entry_safe(waiter, n, &lock->waiters.rb_root, + tree_entry) { + /* XXX debug rt mutex waiter wakeup */ + + BUG_ON(waiter->lock != lock); + rt_mutex_wake_waiter(waiter); + } +} + +#else + +static void ww_mutex_account_lock(struct rt_mutex *lock, + struct ww_acquire_ctx *ww_ctx) +{ + BUG(); +} +#endif + +int __sched rt_mutex_slowlock_locked(struct rt_mutex *lock, int state, + struct hrtimer_sleeper *timeout, + enum rtmutex_chainwalk chwalk, + struct ww_acquire_ctx *ww_ctx, + struct rt_mutex_waiter *waiter) +{ + int ret; + +#ifdef CONFIG_PREEMPT_RT + if (ww_ctx) { + struct ww_mutex *ww; + + ww = container_of(lock, struct ww_mutex, base.lock); + if (unlikely(ww_ctx == READ_ONCE(ww->ctx))) + return -EALREADY; + } +#endif /* Try to acquire the lock again: */ if (try_to_take_rt_mutex(lock, current, NULL)) { - raw_spin_unlock_irqrestore(&lock->wait_lock, flags); + if (ww_ctx) + ww_mutex_account_lock(lock, ww_ctx); return 0; } @@ -1259,16 +1738,26 @@ rt_mutex_slowlock(struct rt_mutex *lock, if (unlikely(timeout)) hrtimer_start_expires(&timeout->timer, HRTIMER_MODE_ABS); - ret = task_blocks_on_rt_mutex(lock, &waiter, current, chwalk); + ret = task_blocks_on_rt_mutex(lock, waiter, current, chwalk); - if (likely(!ret)) + if (likely(!ret)) { /* sleep on the mutex */ - ret = __rt_mutex_slowlock(lock, state, timeout, &waiter); + ret = __rt_mutex_slowlock(lock, state, timeout, waiter, + ww_ctx); + } else if (ww_ctx) { + /* ww_mutex received EDEADLK, let it become EALREADY */ + ret = __mutex_lock_check_stamp(lock, ww_ctx); + BUG_ON(!ret); + } if (unlikely(ret)) { __set_current_state(TASK_RUNNING); - remove_waiter(lock, &waiter); - rt_mutex_handle_deadlock(ret, chwalk, &waiter); + remove_waiter(lock, waiter); + /* ww_mutex wants to report EDEADLK/EALREADY, let it */ + if (!ww_ctx) + rt_mutex_handle_deadlock(ret, chwalk, waiter); + } else if (ww_ctx) { + ww_mutex_account_lock(lock, ww_ctx); } /* @@ -1276,6 +1765,36 @@ rt_mutex_slowlock(struct rt_mutex *lock, * unconditionally. We might have to fix that up. */ fixup_rt_mutex_waiters(lock); + return ret; +} + +/* + * Slow path lock function: + */ +static int __sched +rt_mutex_slowlock(struct rt_mutex *lock, int state, + struct hrtimer_sleeper *timeout, + enum rtmutex_chainwalk chwalk, + struct ww_acquire_ctx *ww_ctx) +{ + struct rt_mutex_waiter waiter; + unsigned long flags; + int ret = 0; + + rt_mutex_init_waiter(&waiter, false); + + /* + * Technically we could use raw_spin_[un]lock_irq() here, but this can + * be called in early boot if the cmpxchg() fast path is disabled + * (debug, no architecture support). In this case we will acquire the + * rtmutex with lock->wait_lock held. But we cannot unconditionally + * enable interrupts in that early boot case. So we need to use the + * irqsave/restore variants. + */ + raw_spin_lock_irqsave(&lock->wait_lock, flags); + + ret = rt_mutex_slowlock_locked(lock, state, timeout, chwalk, ww_ctx, + &waiter); raw_spin_unlock_irqrestore(&lock->wait_lock, flags); @@ -1336,7 +1855,8 @@ static inline int rt_mutex_slowtrylock(s * Return whether the current task needs to call rt_mutex_postunlock(). */ static bool __sched rt_mutex_slowunlock(struct rt_mutex *lock, - struct wake_q_head *wake_q) + struct wake_q_head *wake_q, + struct wake_q_head *wake_sleeper_q) { unsigned long flags; @@ -1390,7 +1910,7 @@ static bool __sched rt_mutex_slowunlock( * * Queue the next waiter for wakeup once we release the wait_lock. */ - mark_wakeup_next_waiter(wake_q, lock); + mark_wakeup_next_waiter(wake_q, wake_sleeper_q, lock); raw_spin_unlock_irqrestore(&lock->wait_lock, flags); return true; /* call rt_mutex_postunlock() */ @@ -1404,29 +1924,45 @@ static bool __sched rt_mutex_slowunlock( */ static inline int rt_mutex_fastlock(struct rt_mutex *lock, int state, + struct ww_acquire_ctx *ww_ctx, int (*slowfn)(struct rt_mutex *lock, int state, struct hrtimer_sleeper *timeout, - enum rtmutex_chainwalk chwalk)) + enum rtmutex_chainwalk chwalk, + struct ww_acquire_ctx *ww_ctx)) { if (likely(rt_mutex_cmpxchg_acquire(lock, NULL, current))) return 0; - return slowfn(lock, state, NULL, RT_MUTEX_MIN_CHAINWALK); + /* + * If rt_mutex blocks, the function sched_submit_work will not call + * blk_schedule_flush_plug (because tsk_is_pi_blocked would be true). + * We must call blk_schedule_flush_plug here, if we don't call it, + * a deadlock in I/O may happen. + */ + if (unlikely(blk_needs_flush_plug(current))) + blk_schedule_flush_plug(current); + + return slowfn(lock, state, NULL, RT_MUTEX_MIN_CHAINWALK, ww_ctx); } static inline int rt_mutex_timed_fastlock(struct rt_mutex *lock, int state, struct hrtimer_sleeper *timeout, enum rtmutex_chainwalk chwalk, + struct ww_acquire_ctx *ww_ctx, int (*slowfn)(struct rt_mutex *lock, int state, struct hrtimer_sleeper *timeout, - enum rtmutex_chainwalk chwalk)) + enum rtmutex_chainwalk chwalk, + struct ww_acquire_ctx *ww_ctx)) { if (chwalk == RT_MUTEX_MIN_CHAINWALK && likely(rt_mutex_cmpxchg_acquire(lock, NULL, current))) return 0; - return slowfn(lock, state, timeout, chwalk); + if (unlikely(blk_needs_flush_plug(current))) + blk_schedule_flush_plug(current); + + return slowfn(lock, state, timeout, chwalk, ww_ctx); } static inline int @@ -1442,9 +1978,11 @@ rt_mutex_fasttrylock(struct rt_mutex *lo /* * Performs the wakeup of the the top-waiter and re-enables preemption. */ -void rt_mutex_postunlock(struct wake_q_head *wake_q) +void rt_mutex_postunlock(struct wake_q_head *wake_q, + struct wake_q_head *wake_sleeper_q) { wake_up_q(wake_q); + wake_up_q_sleeper(wake_sleeper_q); /* Pairs with preempt_disable() in rt_mutex_slowunlock() */ preempt_enable(); @@ -1453,23 +1991,46 @@ void rt_mutex_postunlock(struct wake_q_h static inline void rt_mutex_fastunlock(struct rt_mutex *lock, bool (*slowfn)(struct rt_mutex *lock, - struct wake_q_head *wqh)) + struct wake_q_head *wqh, + struct wake_q_head *wq_sleeper)) { DEFINE_WAKE_Q(wake_q); + DEFINE_WAKE_Q(wake_sleeper_q); if (likely(rt_mutex_cmpxchg_release(lock, current, NULL))) return; - if (slowfn(lock, &wake_q)) - rt_mutex_postunlock(&wake_q); + if (slowfn(lock, &wake_q, &wake_sleeper_q)) + rt_mutex_postunlock(&wake_q, &wake_sleeper_q); } -static inline void __rt_mutex_lock(struct rt_mutex *lock, unsigned int subclass) +int __sched __rt_mutex_lock_state(struct rt_mutex *lock, int state) { might_sleep(); + return rt_mutex_fastlock(lock, state, NULL, rt_mutex_slowlock); +} + +/** + * rt_mutex_lock_state - lock a rt_mutex with a given state + * + * @lock: The rt_mutex to be locked + * @state: The state to set when blocking on the rt_mutex + */ +static inline int __sched rt_mutex_lock_state(struct rt_mutex *lock, + unsigned int subclass, int state) +{ + int ret; mutex_acquire(&lock->dep_map, subclass, 0, _RET_IP_); - rt_mutex_fastlock(lock, TASK_UNINTERRUPTIBLE, rt_mutex_slowlock); + ret = __rt_mutex_lock_state(lock, state); + if (ret) + mutex_release(&lock->dep_map, _RET_IP_); + return ret; +} + +static inline void __rt_mutex_lock(struct rt_mutex *lock, unsigned int subclass) +{ + rt_mutex_lock_state(lock, subclass, TASK_UNINTERRUPTIBLE); } #ifdef CONFIG_DEBUG_LOCK_ALLOC @@ -1510,16 +2071,7 @@ EXPORT_SYMBOL_GPL(rt_mutex_lock); */ int __sched rt_mutex_lock_interruptible(struct rt_mutex *lock) { - int ret; - - might_sleep(); - - mutex_acquire(&lock->dep_map, 0, 0, _RET_IP_); - ret = rt_mutex_fastlock(lock, TASK_INTERRUPTIBLE, rt_mutex_slowlock); - if (ret) - mutex_release(&lock->dep_map, _RET_IP_); - - return ret; + return rt_mutex_lock_state(lock, 0, TASK_INTERRUPTIBLE); } EXPORT_SYMBOL_GPL(rt_mutex_lock_interruptible); @@ -1537,6 +2089,22 @@ int __sched __rt_mutex_futex_trylock(str } /** + * rt_mutex_lock_killable - lock a rt_mutex killable + * + * @lock: the rt_mutex to be locked + * @detect_deadlock: deadlock detection on/off + * + * Returns: + * 0 on success + * -EINTR when interrupted by a signal + */ +int __sched rt_mutex_lock_killable(struct rt_mutex *lock) +{ + return rt_mutex_lock_state(lock, 0, TASK_KILLABLE); +} +EXPORT_SYMBOL_GPL(rt_mutex_lock_killable); + +/** * rt_mutex_timed_lock - lock a rt_mutex interruptible * the timeout structure is provided * by the caller @@ -1559,6 +2127,7 @@ rt_mutex_timed_lock(struct rt_mutex *loc mutex_acquire(&lock->dep_map, 0, 0, _RET_IP_); ret = rt_mutex_timed_fastlock(lock, TASK_INTERRUPTIBLE, timeout, RT_MUTEX_MIN_CHAINWALK, + NULL, rt_mutex_slowlock); if (ret) mutex_release(&lock->dep_map, _RET_IP_); @@ -1567,6 +2136,18 @@ rt_mutex_timed_lock(struct rt_mutex *loc } EXPORT_SYMBOL_GPL(rt_mutex_timed_lock); +int __sched __rt_mutex_trylock(struct rt_mutex *lock) +{ +#ifdef CONFIG_PREEMPT_RT + if (WARN_ON_ONCE(in_irq() || in_nmi())) +#else + if (WARN_ON_ONCE(in_irq() || in_nmi() || in_serving_softirq())) +#endif + return 0; + + return rt_mutex_fasttrylock(lock, rt_mutex_slowtrylock); +} + /** * rt_mutex_trylock - try to lock a rt_mutex * @@ -1582,10 +2163,7 @@ int __sched rt_mutex_trylock(struct rt_m { int ret; - if (WARN_ON_ONCE(in_irq() || in_nmi() || in_serving_softirq())) - return 0; - - ret = rt_mutex_fasttrylock(lock, rt_mutex_slowtrylock); + ret = __rt_mutex_trylock(lock); if (ret) mutex_acquire(&lock->dep_map, 0, 1, _RET_IP_); @@ -1593,6 +2171,11 @@ int __sched rt_mutex_trylock(struct rt_m } EXPORT_SYMBOL_GPL(rt_mutex_trylock); +void __sched __rt_mutex_unlock(struct rt_mutex *lock) +{ + rt_mutex_fastunlock(lock, rt_mutex_slowunlock); +} + /** * rt_mutex_unlock - unlock a rt_mutex * @@ -1601,16 +2184,13 @@ EXPORT_SYMBOL_GPL(rt_mutex_trylock); void __sched rt_mutex_unlock(struct rt_mutex *lock) { mutex_release(&lock->dep_map, _RET_IP_); - rt_mutex_fastunlock(lock, rt_mutex_slowunlock); + __rt_mutex_unlock(lock); } EXPORT_SYMBOL_GPL(rt_mutex_unlock); -/** - * Futex variant, that since futex variants do not use the fast-path, can be - * simple and will not need to retry. - */ -bool __sched __rt_mutex_futex_unlock(struct rt_mutex *lock, - struct wake_q_head *wake_q) +static bool __sched __rt_mutex_unlock_common(struct rt_mutex *lock, + struct wake_q_head *wake_q, + struct wake_q_head *wq_sleeper) { lockdep_assert_held(&lock->wait_lock); @@ -1627,23 +2207,35 @@ bool __sched __rt_mutex_futex_unlock(str * avoid inversion prior to the wakeup. preempt_disable() * therein pairs with rt_mutex_postunlock(). */ - mark_wakeup_next_waiter(wake_q, lock); + mark_wakeup_next_waiter(wake_q, wq_sleeper, lock); return true; /* call postunlock() */ } +/** + * Futex variant, that since futex variants do not use the fast-path, can be + * simple and will not need to retry. + */ +bool __sched __rt_mutex_futex_unlock(struct rt_mutex *lock, + struct wake_q_head *wake_q, + struct wake_q_head *wq_sleeper) +{ + return __rt_mutex_unlock_common(lock, wake_q, wq_sleeper); +} + void __sched rt_mutex_futex_unlock(struct rt_mutex *lock) { DEFINE_WAKE_Q(wake_q); + DEFINE_WAKE_Q(wake_sleeper_q); unsigned long flags; bool postunlock; raw_spin_lock_irqsave(&lock->wait_lock, flags); - postunlock = __rt_mutex_futex_unlock(lock, &wake_q); + postunlock = __rt_mutex_futex_unlock(lock, &wake_q, &wake_sleeper_q); raw_spin_unlock_irqrestore(&lock->wait_lock, flags); if (postunlock) - rt_mutex_postunlock(&wake_q); + rt_mutex_postunlock(&wake_q, &wake_sleeper_q); } /** @@ -1682,7 +2274,7 @@ void __rt_mutex_init(struct rt_mutex *lo if (name && key) debug_rt_mutex_init(lock, name, key); } -EXPORT_SYMBOL_GPL(__rt_mutex_init); +EXPORT_SYMBOL(__rt_mutex_init); /** * rt_mutex_init_proxy_locked - initialize and lock a rt_mutex on behalf of a @@ -1702,6 +2294,14 @@ void rt_mutex_init_proxy_locked(struct r struct task_struct *proxy_owner) { __rt_mutex_init(lock, NULL, NULL); +#ifdef CONFIG_DEBUG_SPINLOCK + /* + * get another key class for the wait_lock. LOCK_PI and UNLOCK_PI is + * holding the ->wait_lock of the proxy_lock while unlocking a sleeping + * lock. + */ + raw_spin_lock_init(&lock->wait_lock); +#endif debug_rt_mutex_proxy_lock(lock, proxy_owner); rt_mutex_set_owner(lock, proxy_owner); } @@ -1725,6 +2325,26 @@ void rt_mutex_proxy_unlock(struct rt_mut rt_mutex_set_owner(lock, NULL); } +static void fixup_rt_mutex_blocked(struct rt_mutex *lock) +{ + struct task_struct *tsk = current; + /* + * RT has a problem here when the wait got interrupted by a timeout + * or a signal. task->pi_blocked_on is still set. The task must + * acquire the hash bucket lock when returning from this function. + * + * If the hash bucket lock is contended then the + * BUG_ON(rt_mutex_real_waiter(task->pi_blocked_on)) in + * task_blocks_on_rt_mutex() will trigger. This can be avoided by + * clearing task->pi_blocked_on which removes the task from the + * boosting chain of the rtmutex. That's correct because the task + * is not longer blocked on it. + */ + raw_spin_lock(&tsk->pi_lock); + tsk->pi_blocked_on = NULL; + raw_spin_unlock(&tsk->pi_lock); +} + /** * __rt_mutex_start_proxy_lock() - Start lock acquisition for another task * @lock: the rt_mutex to take @@ -1755,6 +2375,34 @@ int __rt_mutex_start_proxy_lock(struct r if (try_to_take_rt_mutex(lock, task, NULL)) return 1; +#ifdef CONFIG_PREEMPT_RT + /* + * In PREEMPT_RT there's an added race. + * If the task, that we are about to requeue, times out, + * it can set the PI_WAKEUP_INPROGRESS. This tells the requeue + * to skip this task. But right after the task sets + * its pi_blocked_on to PI_WAKEUP_INPROGRESS it can then + * block on the spin_lock(&hb->lock), which in RT is an rtmutex. + * This will replace the PI_WAKEUP_INPROGRESS with the actual + * lock that it blocks on. We *must not* place this task + * on this proxy lock in that case. + * + * To prevent this race, we first take the task's pi_lock + * and check if it has updated its pi_blocked_on. If it has, + * we assume that it woke up and we return -EAGAIN. + * Otherwise, we set the task's pi_blocked_on to + * PI_REQUEUE_INPROGRESS, so that if the task is waking up + * it will know that we are in the process of requeuing it. + */ + raw_spin_lock(&task->pi_lock); + if (task->pi_blocked_on) { + raw_spin_unlock(&task->pi_lock); + return -EAGAIN; + } + task->pi_blocked_on = PI_REQUEUE_INPROGRESS; + raw_spin_unlock(&task->pi_lock); +#endif + /* We enforce deadlock detection for futexes */ ret = task_blocks_on_rt_mutex(lock, waiter, task, RT_MUTEX_FULL_CHAINWALK); @@ -1769,6 +2417,9 @@ int __rt_mutex_start_proxy_lock(struct r ret = 0; } + if (ret) + fixup_rt_mutex_blocked(lock); + debug_rt_mutex_print_deadlock(waiter); return ret; @@ -1854,12 +2505,15 @@ int rt_mutex_wait_proxy_lock(struct rt_m raw_spin_lock_irq(&lock->wait_lock); /* sleep on the mutex */ set_current_state(TASK_INTERRUPTIBLE); - ret = __rt_mutex_slowlock(lock, TASK_INTERRUPTIBLE, to, waiter); + ret = __rt_mutex_slowlock(lock, TASK_INTERRUPTIBLE, to, waiter, NULL); /* * try_to_take_rt_mutex() sets the waiter bit unconditionally. We might * have to fix that up. */ fixup_rt_mutex_waiters(lock); + if (ret) + fixup_rt_mutex_blocked(lock); + raw_spin_unlock_irq(&lock->wait_lock); return ret; @@ -1921,3 +2575,97 @@ bool rt_mutex_cleanup_proxy_lock(struct return cleanup; } + +static inline int +ww_mutex_deadlock_injection(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ +#ifdef CONFIG_DEBUG_WW_MUTEX_SLOWPATH + unsigned tmp; + + if (ctx->deadlock_inject_countdown-- == 0) { + tmp = ctx->deadlock_inject_interval; + if (tmp > UINT_MAX/4) + tmp = UINT_MAX; + else + tmp = tmp*2 + tmp + tmp/2; + + ctx->deadlock_inject_interval = tmp; + ctx->deadlock_inject_countdown = tmp; + ctx->contending_lock = lock; + + ww_mutex_unlock(lock); + + return -EDEADLK; + } +#endif + + return 0; +} + +#ifdef CONFIG_PREEMPT_RT +int __sched +ww_mutex_lock_interruptible(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + int ret; + + might_sleep(); + + mutex_acquire_nest(&lock->base.dep_map, 0, 0, + ctx ? &ctx->dep_map : NULL, _RET_IP_); + ret = rt_mutex_slowlock(&lock->base.lock, TASK_INTERRUPTIBLE, NULL, 0, + ctx); + if (ret) + mutex_release(&lock->base.dep_map, _RET_IP_); + else if (!ret && ctx && ctx->acquired > 1) + return ww_mutex_deadlock_injection(lock, ctx); + + return ret; +} +EXPORT_SYMBOL_GPL(ww_mutex_lock_interruptible); + +int __sched +ww_mutex_lock(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + int ret; + + might_sleep(); + + mutex_acquire_nest(&lock->base.dep_map, 0, 0, + ctx ? &ctx->dep_map : NULL, _RET_IP_); + ret = rt_mutex_slowlock(&lock->base.lock, TASK_UNINTERRUPTIBLE, NULL, 0, + ctx); + if (ret) + mutex_release(&lock->base.dep_map, _RET_IP_); + else if (!ret && ctx && ctx->acquired > 1) + return ww_mutex_deadlock_injection(lock, ctx); + + return ret; +} +EXPORT_SYMBOL_GPL(ww_mutex_lock); + +void __sched ww_mutex_unlock(struct ww_mutex *lock) +{ + /* + * The unlocking fastpath is the 0->1 transition from 'locked' + * into 'unlocked' state: + */ + if (lock->ctx) { +#ifdef CONFIG_DEBUG_MUTEXES + DEBUG_LOCKS_WARN_ON(!lock->ctx->acquired); +#endif + if (lock->ctx->acquired > 0) + lock->ctx->acquired--; + lock->ctx = NULL; + } + + mutex_release(&lock->base.dep_map, _RET_IP_); + __rt_mutex_unlock(&lock->base.lock); +} +EXPORT_SYMBOL(ww_mutex_unlock); + +int __rt_mutex_owner_current(struct rt_mutex *lock) +{ + return rt_mutex_owner(lock) == current; +} +EXPORT_SYMBOL(__rt_mutex_owner_current); +#endif Index: linux-5.6.17-rt10/kernel/locking/rtmutex_common.h =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/rtmutex_common.h +++ linux-5.6.17-rt10/kernel/locking/rtmutex_common.h @@ -15,6 +15,7 @@ #include #include +#include /* * This is the control structure for tasks blocked on a rt_mutex, @@ -29,6 +30,7 @@ struct rt_mutex_waiter { struct rb_node pi_tree_entry; struct task_struct *task; struct rt_mutex *lock; + bool savestate; #ifdef CONFIG_DEBUG_RT_MUTEXES unsigned long ip; struct pid *deadlock_task_pid; @@ -130,12 +132,15 @@ enum rtmutex_chainwalk { /* * PI-futex support (proxy locking functions, etc.): */ +#define PI_WAKEUP_INPROGRESS ((struct rt_mutex_waiter *) 1) +#define PI_REQUEUE_INPROGRESS ((struct rt_mutex_waiter *) 2) + extern struct task_struct *rt_mutex_next_owner(struct rt_mutex *lock); extern void rt_mutex_init_proxy_locked(struct rt_mutex *lock, struct task_struct *proxy_owner); extern void rt_mutex_proxy_unlock(struct rt_mutex *lock, struct task_struct *proxy_owner); -extern void rt_mutex_init_waiter(struct rt_mutex_waiter *waiter); +extern void rt_mutex_init_waiter(struct rt_mutex_waiter *waiter, bool savetate); extern int __rt_mutex_start_proxy_lock(struct rt_mutex *lock, struct rt_mutex_waiter *waiter, struct task_struct *task); @@ -153,9 +158,27 @@ extern int __rt_mutex_futex_trylock(stru extern void rt_mutex_futex_unlock(struct rt_mutex *lock); extern bool __rt_mutex_futex_unlock(struct rt_mutex *lock, - struct wake_q_head *wqh); + struct wake_q_head *wqh, + struct wake_q_head *wq_sleeper); + +extern void rt_mutex_postunlock(struct wake_q_head *wake_q, + struct wake_q_head *wake_sleeper_q); + +/* RW semaphore special interface */ +struct ww_acquire_ctx; -extern void rt_mutex_postunlock(struct wake_q_head *wake_q); +extern int __rt_mutex_lock_state(struct rt_mutex *lock, int state); +extern int __rt_mutex_trylock(struct rt_mutex *lock); +extern void __rt_mutex_unlock(struct rt_mutex *lock); +int __sched rt_mutex_slowlock_locked(struct rt_mutex *lock, int state, + struct hrtimer_sleeper *timeout, + enum rtmutex_chainwalk chwalk, + struct ww_acquire_ctx *ww_ctx, + struct rt_mutex_waiter *waiter); +void __sched rt_spin_lock_slowlock_locked(struct rt_mutex *lock, + struct rt_mutex_waiter *waiter, + unsigned long flags); +void __sched rt_spin_lock_slowunlock(struct rt_mutex *lock); #ifdef CONFIG_DEBUG_RT_MUTEXES # include "rtmutex-debug.h" Index: linux-5.6.17-rt10/kernel/locking/rwlock-rt.c =================================================================== --- /dev/null +++ linux-5.6.17-rt10/kernel/locking/rwlock-rt.c @@ -0,0 +1,384 @@ +/* + */ +#include +#include + +#include "rtmutex_common.h" +#include + +/* + * RT-specific reader/writer locks + * + * write_lock() + * 1) Lock lock->rtmutex + * 2) Remove the reader BIAS to force readers into the slow path + * 3) Wait until all readers have left the critical region + * 4) Mark it write locked + * + * write_unlock() + * 1) Remove the write locked marker + * 2) Set the reader BIAS so readers can use the fast path again + * 3) Unlock lock->rtmutex to release blocked readers + * + * read_lock() + * 1) Try fast path acquisition (reader BIAS is set) + * 2) Take lock->rtmutex.wait_lock which protects the writelocked flag + * 3) If !writelocked, acquire it for read + * 4) If writelocked, block on lock->rtmutex + * 5) unlock lock->rtmutex, goto 1) + * + * read_unlock() + * 1) Try fast path release (reader count != 1) + * 2) Wake the writer waiting in write_lock()#3 + * + * read_lock()#3 has the consequence, that rw locks on RT are not writer + * fair, but writers, which should be avoided in RT tasks (think tasklist + * lock), are subject to the rtmutex priority/DL inheritance mechanism. + * + * It's possible to make the rw locks writer fair by keeping a list of + * active readers. A blocked writer would force all newly incoming readers + * to block on the rtmutex, but the rtmutex would have to be proxy locked + * for one reader after the other. We can't use multi-reader inheritance + * because there is no way to support that with + * SCHED_DEADLINE. Implementing the one by one reader boosting/handover + * mechanism is a major surgery for a very dubious value. + * + * The risk of writer starvation is there, but the pathological use cases + * which trigger it are not necessarily the typical RT workloads. + */ + +void __rwlock_biased_rt_init(struct rt_rw_lock *lock, const char *name, + struct lock_class_key *key) +{ +#ifdef CONFIG_DEBUG_LOCK_ALLOC + /* + * Make sure we are not reinitializing a held semaphore: + */ + debug_check_no_locks_freed((void *)lock, sizeof(*lock)); + lockdep_init_map(&lock->dep_map, name, key, 0); +#endif + atomic_set(&lock->readers, READER_BIAS); + rt_mutex_init(&lock->rtmutex); + lock->rtmutex.save_state = 1; +} + +int __read_rt_trylock(struct rt_rw_lock *lock) +{ + int r, old; + + /* + * Increment reader count, if lock->readers < 0, i.e. READER_BIAS is + * set. + */ + for (r = atomic_read(&lock->readers); r < 0;) { + old = atomic_cmpxchg(&lock->readers, r, r + 1); + if (likely(old == r)) + return 1; + r = old; + } + return 0; +} + +void __sched __read_rt_lock(struct rt_rw_lock *lock) +{ + struct rt_mutex *m = &lock->rtmutex; + struct rt_mutex_waiter waiter; + unsigned long flags; + + if (__read_rt_trylock(lock)) + return; + + raw_spin_lock_irqsave(&m->wait_lock, flags); + /* + * Allow readers as long as the writer has not completely + * acquired the semaphore for write. + */ + if (atomic_read(&lock->readers) != WRITER_BIAS) { + atomic_inc(&lock->readers); + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + return; + } + + /* + * Call into the slow lock path with the rtmutex->wait_lock + * held, so this can't result in the following race: + * + * Reader1 Reader2 Writer + * read_lock() + * write_lock() + * rtmutex_lock(m) + * swait() + * read_lock() + * unlock(m->wait_lock) + * read_unlock() + * swake() + * lock(m->wait_lock) + * lock->writelocked=true + * unlock(m->wait_lock) + * + * write_unlock() + * lock->writelocked=false + * rtmutex_unlock(m) + * read_lock() + * write_lock() + * rtmutex_lock(m) + * swait() + * rtmutex_lock(m) + * + * That would put Reader1 behind the writer waiting on + * Reader2 to call read_unlock() which might be unbound. + */ + rt_mutex_init_waiter(&waiter, true); + rt_spin_lock_slowlock_locked(m, &waiter, flags); + /* + * The slowlock() above is guaranteed to return with the rtmutex is + * now held, so there can't be a writer active. Increment the reader + * count and immediately drop the rtmutex again. + */ + atomic_inc(&lock->readers); + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + rt_spin_lock_slowunlock(m); + + debug_rt_mutex_free_waiter(&waiter); +} + +void __read_rt_unlock(struct rt_rw_lock *lock) +{ + struct rt_mutex *m = &lock->rtmutex; + struct task_struct *tsk; + + /* + * sem->readers can only hit 0 when a writer is waiting for the + * active readers to leave the critical region. + */ + if (!atomic_dec_and_test(&lock->readers)) + return; + + raw_spin_lock_irq(&m->wait_lock); + /* + * Wake the writer, i.e. the rtmutex owner. It might release the + * rtmutex concurrently in the fast path, but to clean up the rw + * lock it needs to acquire m->wait_lock. The worst case which can + * happen is a spurious wakeup. + */ + tsk = rt_mutex_owner(m); + if (tsk) + wake_up_process(tsk); + + raw_spin_unlock_irq(&m->wait_lock); +} + +static void __write_unlock_common(struct rt_rw_lock *lock, int bias, + unsigned long flags) +{ + struct rt_mutex *m = &lock->rtmutex; + + atomic_add(READER_BIAS - bias, &lock->readers); + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + rt_spin_lock_slowunlock(m); +} + +void __sched __write_rt_lock(struct rt_rw_lock *lock) +{ + struct rt_mutex *m = &lock->rtmutex; + struct task_struct *self = current; + unsigned long flags; + + /* Take the rtmutex as a first step */ + __rt_spin_lock(m); + + /* Force readers into slow path */ + atomic_sub(READER_BIAS, &lock->readers); + + raw_spin_lock_irqsave(&m->wait_lock, flags); + + raw_spin_lock(&self->pi_lock); + self->saved_state = self->state; + __set_current_state_no_track(TASK_UNINTERRUPTIBLE); + raw_spin_unlock(&self->pi_lock); + + for (;;) { + /* Have all readers left the critical region? */ + if (!atomic_read(&lock->readers)) { + atomic_set(&lock->readers, WRITER_BIAS); + raw_spin_lock(&self->pi_lock); + __set_current_state_no_track(self->saved_state); + self->saved_state = TASK_RUNNING; + raw_spin_unlock(&self->pi_lock); + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + return; + } + + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + + if (atomic_read(&lock->readers) != 0) + schedule(); + + raw_spin_lock_irqsave(&m->wait_lock, flags); + + raw_spin_lock(&self->pi_lock); + __set_current_state_no_track(TASK_UNINTERRUPTIBLE); + raw_spin_unlock(&self->pi_lock); + } +} + +int __write_rt_trylock(struct rt_rw_lock *lock) +{ + struct rt_mutex *m = &lock->rtmutex; + unsigned long flags; + + if (!__rt_mutex_trylock(m)) + return 0; + + atomic_sub(READER_BIAS, &lock->readers); + + raw_spin_lock_irqsave(&m->wait_lock, flags); + if (!atomic_read(&lock->readers)) { + atomic_set(&lock->readers, WRITER_BIAS); + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + return 1; + } + __write_unlock_common(lock, 0, flags); + return 0; +} + +void __write_rt_unlock(struct rt_rw_lock *lock) +{ + struct rt_mutex *m = &lock->rtmutex; + unsigned long flags; + + raw_spin_lock_irqsave(&m->wait_lock, flags); + __write_unlock_common(lock, WRITER_BIAS, flags); +} + +/* Map the reader biased implementation */ +static inline int do_read_rt_trylock(rwlock_t *rwlock) +{ + return __read_rt_trylock(rwlock); +} + +static inline int do_write_rt_trylock(rwlock_t *rwlock) +{ + return __write_rt_trylock(rwlock); +} + +static inline void do_read_rt_lock(rwlock_t *rwlock) +{ + __read_rt_lock(rwlock); +} + +static inline void do_write_rt_lock(rwlock_t *rwlock) +{ + __write_rt_lock(rwlock); +} + +static inline void do_read_rt_unlock(rwlock_t *rwlock) +{ + __read_rt_unlock(rwlock); +} + +static inline void do_write_rt_unlock(rwlock_t *rwlock) +{ + __write_rt_unlock(rwlock); +} + +static inline void do_rwlock_rt_init(rwlock_t *rwlock, const char *name, + struct lock_class_key *key) +{ + __rwlock_biased_rt_init(rwlock, name, key); +} + +int __lockfunc rt_read_can_lock(rwlock_t *rwlock) +{ + return atomic_read(&rwlock->readers) < 0; +} + +int __lockfunc rt_write_can_lock(rwlock_t *rwlock) +{ + return atomic_read(&rwlock->readers) == READER_BIAS; +} + +/* + * The common functions which get wrapped into the rwlock API. + */ +int __lockfunc rt_read_trylock(rwlock_t *rwlock) +{ + int ret; + + sleeping_lock_inc(); + migrate_disable(); + ret = do_read_rt_trylock(rwlock); + if (ret) { + rwlock_acquire_read(&rwlock->dep_map, 0, 1, _RET_IP_); + rcu_read_lock(); + } else { + migrate_enable(); + sleeping_lock_dec(); + } + return ret; +} +EXPORT_SYMBOL(rt_read_trylock); + +int __lockfunc rt_write_trylock(rwlock_t *rwlock) +{ + int ret; + + sleeping_lock_inc(); + migrate_disable(); + ret = do_write_rt_trylock(rwlock); + if (ret) { + rwlock_acquire(&rwlock->dep_map, 0, 1, _RET_IP_); + rcu_read_lock(); + } else { + migrate_enable(); + sleeping_lock_dec(); + } + return ret; +} +EXPORT_SYMBOL(rt_write_trylock); + +void __lockfunc rt_read_lock(rwlock_t *rwlock) +{ + sleeping_lock_inc(); + rcu_read_lock(); + migrate_disable(); + rwlock_acquire_read(&rwlock->dep_map, 0, 0, _RET_IP_); + do_read_rt_lock(rwlock); +} +EXPORT_SYMBOL(rt_read_lock); + +void __lockfunc rt_write_lock(rwlock_t *rwlock) +{ + sleeping_lock_inc(); + rcu_read_lock(); + migrate_disable(); + rwlock_acquire(&rwlock->dep_map, 0, 0, _RET_IP_); + do_write_rt_lock(rwlock); +} +EXPORT_SYMBOL(rt_write_lock); + +void __lockfunc rt_read_unlock(rwlock_t *rwlock) +{ + rwlock_release(&rwlock->dep_map, _RET_IP_); + do_read_rt_unlock(rwlock); + migrate_enable(); + rcu_read_unlock(); + sleeping_lock_dec(); +} +EXPORT_SYMBOL(rt_read_unlock); + +void __lockfunc rt_write_unlock(rwlock_t *rwlock) +{ + rwlock_release(&rwlock->dep_map, _RET_IP_); + do_write_rt_unlock(rwlock); + migrate_enable(); + rcu_read_unlock(); + sleeping_lock_dec(); +} +EXPORT_SYMBOL(rt_write_unlock); + +void __rt_rwlock_init(rwlock_t *rwlock, char *name, struct lock_class_key *key) +{ + do_rwlock_rt_init(rwlock, name, key); +} +EXPORT_SYMBOL(__rt_rwlock_init); Index: linux-5.6.17-rt10/kernel/locking/rwsem-rt.c =================================================================== --- /dev/null +++ linux-5.6.17-rt10/kernel/locking/rwsem-rt.c @@ -0,0 +1,302 @@ +/* + */ +#include +#include +#include +#include +#include + +#include "rtmutex_common.h" + +/* + * RT-specific reader/writer semaphores + * + * down_write() + * 1) Lock sem->rtmutex + * 2) Remove the reader BIAS to force readers into the slow path + * 3) Wait until all readers have left the critical region + * 4) Mark it write locked + * + * up_write() + * 1) Remove the write locked marker + * 2) Set the reader BIAS so readers can use the fast path again + * 3) Unlock sem->rtmutex to release blocked readers + * + * down_read() + * 1) Try fast path acquisition (reader BIAS is set) + * 2) Take sem->rtmutex.wait_lock which protects the writelocked flag + * 3) If !writelocked, acquire it for read + * 4) If writelocked, block on sem->rtmutex + * 5) unlock sem->rtmutex, goto 1) + * + * up_read() + * 1) Try fast path release (reader count != 1) + * 2) Wake the writer waiting in down_write()#3 + * + * down_read()#3 has the consequence, that rw semaphores on RT are not writer + * fair, but writers, which should be avoided in RT tasks (think mmap_sem), + * are subject to the rtmutex priority/DL inheritance mechanism. + * + * It's possible to make the rw semaphores writer fair by keeping a list of + * active readers. A blocked writer would force all newly incoming readers to + * block on the rtmutex, but the rtmutex would have to be proxy locked for one + * reader after the other. We can't use multi-reader inheritance because there + * is no way to support that with SCHED_DEADLINE. Implementing the one by one + * reader boosting/handover mechanism is a major surgery for a very dubious + * value. + * + * The risk of writer starvation is there, but the pathological use cases + * which trigger it are not necessarily the typical RT workloads. + */ + +void __rwsem_init(struct rw_semaphore *sem, const char *name, + struct lock_class_key *key) +{ +#ifdef CONFIG_DEBUG_LOCK_ALLOC + /* + * Make sure we are not reinitializing a held semaphore: + */ + debug_check_no_locks_freed((void *)sem, sizeof(*sem)); + lockdep_init_map(&sem->dep_map, name, key, 0); +#endif + atomic_set(&sem->readers, READER_BIAS); +} +EXPORT_SYMBOL(__rwsem_init); + +int __down_read_trylock(struct rw_semaphore *sem) +{ + int r, old; + + /* + * Increment reader count, if sem->readers < 0, i.e. READER_BIAS is + * set. + */ + for (r = atomic_read(&sem->readers); r < 0;) { + old = atomic_cmpxchg(&sem->readers, r, r + 1); + if (likely(old == r)) + return 1; + r = old; + } + return 0; +} + +static int __sched __down_read_common(struct rw_semaphore *sem, int state) +{ + struct rt_mutex *m = &sem->rtmutex; + struct rt_mutex_waiter waiter; + int ret; + + if (__down_read_trylock(sem)) + return 0; + /* + * If rt_mutex blocks, the function sched_submit_work will not call + * blk_schedule_flush_plug (because tsk_is_pi_blocked would be true). + * We must call blk_schedule_flush_plug here, if we don't call it, + * a deadlock in I/O may happen. + */ + if (unlikely(blk_needs_flush_plug(current))) + blk_schedule_flush_plug(current); + + might_sleep(); + raw_spin_lock_irq(&m->wait_lock); + /* + * Allow readers as long as the writer has not completely + * acquired the semaphore for write. + */ + if (atomic_read(&sem->readers) != WRITER_BIAS) { + atomic_inc(&sem->readers); + raw_spin_unlock_irq(&m->wait_lock); + return 0; + } + + /* + * Call into the slow lock path with the rtmutex->wait_lock + * held, so this can't result in the following race: + * + * Reader1 Reader2 Writer + * down_read() + * down_write() + * rtmutex_lock(m) + * swait() + * down_read() + * unlock(m->wait_lock) + * up_read() + * swake() + * lock(m->wait_lock) + * sem->writelocked=true + * unlock(m->wait_lock) + * + * up_write() + * sem->writelocked=false + * rtmutex_unlock(m) + * down_read() + * down_write() + * rtmutex_lock(m) + * swait() + * rtmutex_lock(m) + * + * That would put Reader1 behind the writer waiting on + * Reader2 to call up_read() which might be unbound. + */ + rt_mutex_init_waiter(&waiter, false); + ret = rt_mutex_slowlock_locked(m, state, NULL, RT_MUTEX_MIN_CHAINWALK, + NULL, &waiter); + /* + * The slowlock() above is guaranteed to return with the rtmutex (for + * ret = 0) is now held, so there can't be a writer active. Increment + * the reader count and immediately drop the rtmutex again. + * For ret != 0 we don't hold the rtmutex and need unlock the wait_lock. + * We don't own the lock then. + */ + if (!ret) + atomic_inc(&sem->readers); + raw_spin_unlock_irq(&m->wait_lock); + if (!ret) + __rt_mutex_unlock(m); + + debug_rt_mutex_free_waiter(&waiter); + return ret; +} + +void __down_read(struct rw_semaphore *sem) +{ + int ret; + + ret = __down_read_common(sem, TASK_UNINTERRUPTIBLE); + WARN_ON_ONCE(ret); +} + +int __down_read_killable(struct rw_semaphore *sem) +{ + int ret; + + ret = __down_read_common(sem, TASK_KILLABLE); + if (likely(!ret)) + return ret; + WARN_ONCE(ret != -EINTR, "Unexpected state: %d\n", ret); + return -EINTR; +} + +void __up_read(struct rw_semaphore *sem) +{ + struct rt_mutex *m = &sem->rtmutex; + struct task_struct *tsk; + + /* + * sem->readers can only hit 0 when a writer is waiting for the + * active readers to leave the critical region. + */ + if (!atomic_dec_and_test(&sem->readers)) + return; + + might_sleep(); + raw_spin_lock_irq(&m->wait_lock); + /* + * Wake the writer, i.e. the rtmutex owner. It might release the + * rtmutex concurrently in the fast path (due to a signal), but to + * clean up the rwsem it needs to acquire m->wait_lock. The worst + * case which can happen is a spurious wakeup. + */ + tsk = rt_mutex_owner(m); + if (tsk) + wake_up_process(tsk); + + raw_spin_unlock_irq(&m->wait_lock); +} + +static void __up_write_unlock(struct rw_semaphore *sem, int bias, + unsigned long flags) +{ + struct rt_mutex *m = &sem->rtmutex; + + atomic_add(READER_BIAS - bias, &sem->readers); + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + __rt_mutex_unlock(m); +} + +static int __sched __down_write_common(struct rw_semaphore *sem, int state) +{ + struct rt_mutex *m = &sem->rtmutex; + unsigned long flags; + + /* Take the rtmutex as a first step */ + if (__rt_mutex_lock_state(m, state)) + return -EINTR; + + /* Force readers into slow path */ + atomic_sub(READER_BIAS, &sem->readers); + might_sleep(); + + set_current_state(state); + for (;;) { + raw_spin_lock_irqsave(&m->wait_lock, flags); + /* Have all readers left the critical region? */ + if (!atomic_read(&sem->readers)) { + atomic_set(&sem->readers, WRITER_BIAS); + __set_current_state(TASK_RUNNING); + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + return 0; + } + + if (signal_pending_state(state, current)) { + __set_current_state(TASK_RUNNING); + __up_write_unlock(sem, 0, flags); + return -EINTR; + } + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + + if (atomic_read(&sem->readers) != 0) { + schedule(); + set_current_state(state); + } + } +} + +void __sched __down_write(struct rw_semaphore *sem) +{ + __down_write_common(sem, TASK_UNINTERRUPTIBLE); +} + +int __sched __down_write_killable(struct rw_semaphore *sem) +{ + return __down_write_common(sem, TASK_KILLABLE); +} + +int __down_write_trylock(struct rw_semaphore *sem) +{ + struct rt_mutex *m = &sem->rtmutex; + unsigned long flags; + + if (!__rt_mutex_trylock(m)) + return 0; + + atomic_sub(READER_BIAS, &sem->readers); + + raw_spin_lock_irqsave(&m->wait_lock, flags); + if (!atomic_read(&sem->readers)) { + atomic_set(&sem->readers, WRITER_BIAS); + raw_spin_unlock_irqrestore(&m->wait_lock, flags); + return 1; + } + __up_write_unlock(sem, 0, flags); + return 0; +} + +void __up_write(struct rw_semaphore *sem) +{ + struct rt_mutex *m = &sem->rtmutex; + unsigned long flags; + + raw_spin_lock_irqsave(&m->wait_lock, flags); + __up_write_unlock(sem, WRITER_BIAS, flags); +} + +void __downgrade_write(struct rw_semaphore *sem) +{ + struct rt_mutex *m = &sem->rtmutex; + unsigned long flags; + + raw_spin_lock_irqsave(&m->wait_lock, flags); + /* Release it and account current as reader */ + __up_write_unlock(sem, WRITER_BIAS - 1, flags); +} Index: linux-5.6.17-rt10/kernel/locking/rwsem.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/rwsem.c +++ linux-5.6.17-rt10/kernel/locking/rwsem.c @@ -28,7 +28,7 @@ #include #include -#include "rwsem.h" +#ifndef CONFIG_PREEMPT_RT #include "lock_events.h" /* @@ -660,8 +660,6 @@ static inline bool rwsem_can_spin_on_own unsigned long flags; bool ret = true; - BUILD_BUG_ON(!(RWSEM_OWNER_UNKNOWN & RWSEM_NONSPINNABLE)); - if (need_resched()) { lockevent_inc(rwsem_opt_fail); return false; @@ -1335,10 +1333,11 @@ static struct rw_semaphore *rwsem_downgr return sem; } + /* * lock for reading */ -inline void __down_read(struct rw_semaphore *sem) +static inline void __down_read(struct rw_semaphore *sem) { if (!rwsem_read_trylock(sem)) { rwsem_down_read_slowpath(sem, TASK_UNINTERRUPTIBLE); @@ -1426,7 +1425,7 @@ static inline int __down_write_trylock(s /* * unlock after reading */ -inline void __up_read(struct rw_semaphore *sem) +static inline void __up_read(struct rw_semaphore *sem) { long tmp; @@ -1485,6 +1484,7 @@ static inline void __downgrade_write(str if (tmp & RWSEM_FLAG_WAITERS) rwsem_downgrade_wake(sem); } +#endif /* * lock for reading @@ -1620,7 +1620,9 @@ void down_read_non_owner(struct rw_semap { might_sleep(); __down_read(sem); +#ifndef CONFIG_PREEMPT_RT __rwsem_set_reader_owned(sem, NULL); +#endif } EXPORT_SYMBOL(down_read_non_owner); @@ -1649,7 +1651,9 @@ EXPORT_SYMBOL(down_write_killable_nested void up_read_non_owner(struct rw_semaphore *sem) { +#ifndef CONFIG_PREEMPT_RT DEBUG_RWSEMS_WARN_ON(!is_rwsem_reader_owned(sem), sem); +#endif __up_read(sem); } EXPORT_SYMBOL(up_read_non_owner); Index: linux-5.6.17-rt10/kernel/locking/rwsem.h =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/rwsem.h +++ linux-5.6.17-rt10/kernel/locking/rwsem.h @@ -1,10 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 */ - -#ifndef __INTERNAL_RWSEM_H -#define __INTERNAL_RWSEM_H -#include - -extern void __down_read(struct rw_semaphore *sem); -extern void __up_read(struct rw_semaphore *sem); - -#endif /* __INTERNAL_RWSEM_H */ Index: linux-5.6.17-rt10/kernel/locking/spinlock.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/spinlock.c +++ linux-5.6.17-rt10/kernel/locking/spinlock.c @@ -124,8 +124,11 @@ void __lockfunc __raw_##op##_lock_bh(loc * __[spin|read|write]_lock_bh() */ BUILD_LOCK_OPS(spin, raw_spinlock); + +#ifndef CONFIG_PREEMPT_RT BUILD_LOCK_OPS(read, rwlock); BUILD_LOCK_OPS(write, rwlock); +#endif #endif @@ -209,6 +212,8 @@ void __lockfunc _raw_spin_unlock_bh(raw_ EXPORT_SYMBOL(_raw_spin_unlock_bh); #endif +#ifndef CONFIG_PREEMPT_RT + #ifndef CONFIG_INLINE_READ_TRYLOCK int __lockfunc _raw_read_trylock(rwlock_t *lock) { @@ -353,6 +358,8 @@ void __lockfunc _raw_write_unlock_bh(rwl EXPORT_SYMBOL(_raw_write_unlock_bh); #endif +#endif /* !PREEMPT_RT */ + #ifdef CONFIG_DEBUG_LOCK_ALLOC void __lockfunc _raw_spin_lock_nested(raw_spinlock_t *lock, int subclass) Index: linux-5.6.17-rt10/kernel/locking/spinlock_debug.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/locking/spinlock_debug.c +++ linux-5.6.17-rt10/kernel/locking/spinlock_debug.c @@ -31,6 +31,7 @@ void __raw_spin_lock_init(raw_spinlock_t EXPORT_SYMBOL(__raw_spin_lock_init); +#ifndef CONFIG_PREEMPT_RT void __rwlock_init(rwlock_t *lock, const char *name, struct lock_class_key *key) { @@ -48,6 +49,7 @@ void __rwlock_init(rwlock_t *lock, const } EXPORT_SYMBOL(__rwlock_init); +#endif static void spin_dump(raw_spinlock_t *lock, const char *msg) { @@ -139,6 +141,7 @@ void do_raw_spin_unlock(raw_spinlock_t * arch_spin_unlock(&lock->raw_lock); } +#ifndef CONFIG_PREEMPT_RT static void rwlock_bug(rwlock_t *lock, const char *msg) { if (!debug_locks_off()) @@ -228,3 +231,5 @@ void do_raw_write_unlock(rwlock_t *lock) debug_write_unlock(lock); arch_write_unlock(&lock->raw_lock); } + +#endif Index: linux-5.6.17-rt10/kernel/panic.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/panic.c +++ linux-5.6.17-rt10/kernel/panic.c @@ -237,7 +237,6 @@ void panic(const char *fmt, ...) * Bypass the panic_cpu check and call __crash_kexec directly. */ if (!_crash_kexec_post_notifiers) { - printk_safe_flush_on_panic(); __crash_kexec(NULL); /* @@ -261,8 +260,6 @@ void panic(const char *fmt, ...) */ atomic_notifier_call_chain(&panic_notifier_list, 0, buf); - /* Call flush even twice. It tries harder with a single online CPU */ - printk_safe_flush_on_panic(); kmsg_dump(KMSG_DUMP_PANIC); /* @@ -524,9 +521,11 @@ static u64 oops_id; static int init_oops_id(void) { +#ifndef CONFIG_PREEMPT_RT if (!oops_id) get_random_bytes(&oops_id, sizeof(oops_id)); else +#endif oops_id++; return 0; Index: linux-5.6.17-rt10/kernel/printk/Makefile =================================================================== --- linux-5.6.17-rt10.orig/kernel/printk/Makefile +++ linux-5.6.17-rt10/kernel/printk/Makefile @@ -1,4 +1,3 @@ # SPDX-License-Identifier: GPL-2.0-only obj-y = printk.o -obj-$(CONFIG_PRINTK) += printk_safe.o obj-$(CONFIG_A11Y_BRAILLE_CONSOLE) += braille.o Index: linux-5.6.17-rt10/kernel/printk/internal.h =================================================================== --- linux-5.6.17-rt10.orig/kernel/printk/internal.h +++ /dev/null @@ -1,72 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0-or-later */ -/* - * internal.h - printk internal definitions - */ -#include - -#ifdef CONFIG_PRINTK - -#define PRINTK_SAFE_CONTEXT_MASK 0x3fffffff -#define PRINTK_NMI_DIRECT_CONTEXT_MASK 0x40000000 -#define PRINTK_NMI_CONTEXT_MASK 0x80000000 - -extern raw_spinlock_t logbuf_lock; - -__printf(5, 0) -int vprintk_store(int facility, int level, - const char *dict, size_t dictlen, - const char *fmt, va_list args); - -__printf(1, 0) int vprintk_default(const char *fmt, va_list args); -__printf(1, 0) int vprintk_deferred(const char *fmt, va_list args); -__printf(1, 0) int vprintk_func(const char *fmt, va_list args); -void __printk_safe_enter(void); -void __printk_safe_exit(void); - -void printk_safe_init(void); -bool printk_percpu_data_ready(void); - -#define printk_safe_enter_irqsave(flags) \ - do { \ - local_irq_save(flags); \ - __printk_safe_enter(); \ - } while (0) - -#define printk_safe_exit_irqrestore(flags) \ - do { \ - __printk_safe_exit(); \ - local_irq_restore(flags); \ - } while (0) - -#define printk_safe_enter_irq() \ - do { \ - local_irq_disable(); \ - __printk_safe_enter(); \ - } while (0) - -#define printk_safe_exit_irq() \ - do { \ - __printk_safe_exit(); \ - local_irq_enable(); \ - } while (0) - -void defer_console_output(void); - -#else - -__printf(1, 0) int vprintk_func(const char *fmt, va_list args) { return 0; } - -/* - * In !PRINTK builds we still export logbuf_lock spin_lock, console_sem - * semaphore and some of console functions (console_unlock()/etc.), so - * printk-safe must preserve the existing local IRQ guarantees. - */ -#define printk_safe_enter_irqsave(flags) local_irq_save(flags) -#define printk_safe_exit_irqrestore(flags) local_irq_restore(flags) - -#define printk_safe_enter_irq() local_irq_disable() -#define printk_safe_exit_irq() local_irq_enable() - -static inline void printk_safe_init(void) { } -static inline bool printk_percpu_data_ready(void) { return false; } -#endif /* CONFIG_PRINTK */ Index: linux-5.6.17-rt10/kernel/printk/printk.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/printk/printk.c +++ linux-5.6.17-rt10/kernel/printk/printk.c @@ -45,6 +45,9 @@ #include #include #include +#include +#include +#include #include #include #include @@ -58,13 +61,13 @@ #include "console_cmdline.h" #include "braille.h" -#include "internal.h" -int console_printk[4] = { +int console_printk[5] = { CONSOLE_LOGLEVEL_DEFAULT, /* console_loglevel */ MESSAGE_LOGLEVEL_DEFAULT, /* default_message_loglevel */ CONSOLE_LOGLEVEL_MIN, /* minimum_console_loglevel */ CONSOLE_LOGLEVEL_DEFAULT, /* default_console_loglevel */ + CONSOLE_LOGLEVEL_EMERGENCY, /* emergency_console_loglevel */ }; EXPORT_SYMBOL_GPL(console_printk); @@ -225,19 +228,7 @@ static int nr_ext_console_drivers; static int __down_trylock_console_sem(unsigned long ip) { - int lock_failed; - unsigned long flags; - - /* - * Here and in __up_console_sem() we need to be in safe mode, - * because spindump/WARN/etc from under console ->lock will - * deadlock in printk()->down_trylock_console_sem() otherwise. - */ - printk_safe_enter_irqsave(flags); - lock_failed = down_trylock(&console_sem); - printk_safe_exit_irqrestore(flags); - - if (lock_failed) + if (down_trylock(&console_sem)) return 1; mutex_acquire(&console_lock_dep_map, 0, 1, ip); return 0; @@ -246,13 +237,9 @@ static int __down_trylock_console_sem(un static void __up_console_sem(unsigned long ip) { - unsigned long flags; - mutex_release(&console_lock_dep_map, ip); - printk_safe_enter_irqsave(flags); up(&console_sem); - printk_safe_exit_irqrestore(flags); } #define up_console_sem() __up_console_sem(_RET_IP_) @@ -267,11 +254,6 @@ static void __up_console_sem(unsigned lo static int console_locked, console_suspended; /* - * If exclusive_console is non-NULL then only this console is to be printed to. - */ -static struct console *exclusive_console; - -/* * Array of consoles built from command line options (console=) */ @@ -367,6 +349,7 @@ enum log_flags { struct printk_log { u64 ts_nsec; /* timestamp in nanoseconds */ + u16 cpu; /* cpu that generated record */ u16 len; /* length of entire record */ u16 text_len; /* length of text buffer */ u16 dict_len; /* length of dictionary buffer */ @@ -382,65 +365,22 @@ __packed __aligned(4) #endif ; -/* - * The logbuf_lock protects kmsg buffer, indices, counters. This can be taken - * within the scheduler's rq lock. It must be released before calling - * console_unlock() or anything else that might wake up a process. - */ -DEFINE_RAW_SPINLOCK(logbuf_lock); - -/* - * Helper macros to lock/unlock logbuf_lock and switch between - * printk-safe/unsafe modes. - */ -#define logbuf_lock_irq() \ - do { \ - printk_safe_enter_irq(); \ - raw_spin_lock(&logbuf_lock); \ - } while (0) - -#define logbuf_unlock_irq() \ - do { \ - raw_spin_unlock(&logbuf_lock); \ - printk_safe_exit_irq(); \ - } while (0) - -#define logbuf_lock_irqsave(flags) \ - do { \ - printk_safe_enter_irqsave(flags); \ - raw_spin_lock(&logbuf_lock); \ - } while (0) - -#define logbuf_unlock_irqrestore(flags) \ - do { \ - raw_spin_unlock(&logbuf_lock); \ - printk_safe_exit_irqrestore(flags); \ - } while (0) +DECLARE_STATIC_PRINTKRB_CPULOCK(printk_cpulock); #ifdef CONFIG_PRINTK -DECLARE_WAIT_QUEUE_HEAD(log_wait); -/* the next printk record to read by syslog(READ) or /proc/kmsg */ +/* record buffer */ +DECLARE_STATIC_PRINTKRB(printk_rb, CONFIG_LOG_BUF_SHIFT, &printk_cpulock); + +static DEFINE_MUTEX(syslog_lock); +DECLARE_STATIC_PRINTKRB_ITER(syslog_iter, &printk_rb); + +/* the last printk record to read by syslog(READ) or /proc/kmsg */ static u64 syslog_seq; -static u32 syslog_idx; static size_t syslog_partial; static bool syslog_time; -/* index and sequence number of the first record stored in the buffer */ -static u64 log_first_seq; -static u32 log_first_idx; - -/* index and sequence number of the next record to store in the buffer */ -static u64 log_next_seq; -static u32 log_next_idx; - -/* the next printk record to write to the console */ -static u64 console_seq; -static u32 console_idx; -static u64 exclusive_console_stop_seq; - /* the next printk record to read after the last 'clear' command */ static u64 clear_seq; -static u32 clear_idx; #ifdef CONFIG_PRINTK_CALLER #define PREFIX_MAX 48 @@ -452,36 +392,16 @@ static u32 clear_idx; #define LOG_LEVEL(v) ((v) & 0x07) #define LOG_FACILITY(v) ((v) >> 3 & 0xff) -/* record buffer */ -#define LOG_ALIGN __alignof__(struct printk_log) -#define __LOG_BUF_LEN (1 << CONFIG_LOG_BUF_SHIFT) -#define LOG_BUF_LEN_MAX (u32)(1 << 31) -static char __log_buf[__LOG_BUF_LEN] __aligned(LOG_ALIGN); -static char *log_buf = __log_buf; -static u32 log_buf_len = __LOG_BUF_LEN; - -/* - * We cannot access per-CPU data (e.g. per-CPU flush irq_work) before - * per_cpu_areas are initialised. This variable is set to true when - * it's safe to access per-CPU data. - */ -static bool __printk_percpu_data_ready __read_mostly; - -bool printk_percpu_data_ready(void) -{ - return __printk_percpu_data_ready; -} - /* Return log buffer address */ char *log_buf_addr_get(void) { - return log_buf; + return printk_rb.buffer; } /* Return log buffer size */ u32 log_buf_len_get(void) { - return log_buf_len; + return (1 << printk_rb.size_bits); } /* human readable text of the record */ @@ -496,180 +416,50 @@ static char *log_dict(const struct print return (char *)msg + sizeof(struct printk_log) + msg->text_len; } -/* get record by index; idx must point to valid msg */ -static struct printk_log *log_from_idx(u32 idx) -{ - struct printk_log *msg = (struct printk_log *)(log_buf + idx); - - /* - * A length == 0 record is the end of buffer marker. Wrap around and - * read the message at the start of the buffer. - */ - if (!msg->len) - return (struct printk_log *)log_buf; - return msg; -} - -/* get next record; idx must point to valid msg */ -static u32 log_next(u32 idx) -{ - struct printk_log *msg = (struct printk_log *)(log_buf + idx); - - /* length == 0 indicates the end of the buffer; wrap */ - /* - * A length == 0 record is the end of buffer marker. Wrap around and - * read the message at the start of the buffer as *this* one, and - * return the one after that. - */ - if (!msg->len) { - msg = (struct printk_log *)log_buf; - return msg->len; - } - return idx + msg->len; -} - -/* - * Check whether there is enough free space for the given message. - * - * The same values of first_idx and next_idx mean that the buffer - * is either empty or full. - * - * If the buffer is empty, we must respect the position of the indexes. - * They cannot be reset to the beginning of the buffer. - */ -static int logbuf_has_space(u32 msg_size, bool empty) -{ - u32 free; - - if (log_next_idx > log_first_idx || empty) - free = max(log_buf_len - log_next_idx, log_first_idx); - else - free = log_first_idx - log_next_idx; - - /* - * We need space also for an empty header that signalizes wrapping - * of the buffer. - */ - return free >= msg_size + sizeof(struct printk_log); -} - -static int log_make_free_space(u32 msg_size) -{ - while (log_first_seq < log_next_seq && - !logbuf_has_space(msg_size, false)) { - /* drop old messages until we have enough contiguous space */ - log_first_idx = log_next(log_first_idx); - log_first_seq++; - } - - if (clear_seq < log_first_seq) { - clear_seq = log_first_seq; - clear_idx = log_first_idx; - } - - /* sequence numbers are equal, so the log buffer is empty */ - if (logbuf_has_space(msg_size, log_first_seq == log_next_seq)) - return 0; - - return -ENOMEM; -} - -/* compute the message size including the padding bytes */ -static u32 msg_used_size(u16 text_len, u16 dict_len, u32 *pad_len) -{ - u32 size; - - size = sizeof(struct printk_log) + text_len + dict_len; - *pad_len = (-size) & (LOG_ALIGN - 1); - size += *pad_len; - - return size; -} - -/* - * Define how much of the log buffer we could take at maximum. The value - * must be greater than two. Note that only half of the buffer is available - * when the index points to the middle. - */ -#define MAX_LOG_TAKE_PART 4 -static const char trunc_msg[] = ""; - -static u32 truncate_msg(u16 *text_len, u16 *trunc_msg_len, - u16 *dict_len, u32 *pad_len) -{ - /* - * The message should not take the whole buffer. Otherwise, it might - * get removed too soon. - */ - u32 max_text_len = log_buf_len / MAX_LOG_TAKE_PART; - if (*text_len > max_text_len) - *text_len = max_text_len; - /* enable the warning message */ - *trunc_msg_len = strlen(trunc_msg); - /* disable the "dict" completely */ - *dict_len = 0; - /* compute the size again, count also the warning message */ - return msg_used_size(*text_len + *trunc_msg_len, 0, pad_len); -} +static void printk_emergency(char *buffer, int level, u64 ts_nsec, u16 cpu, + char *text, u16 text_len); /* insert record into the buffer, discard old ones, update heads */ static int log_store(u32 caller_id, int facility, int level, - enum log_flags flags, u64 ts_nsec, + enum log_flags flags, u64 ts_nsec, u16 cpu, const char *dict, u16 dict_len, const char *text, u16 text_len) { struct printk_log *msg; - u32 size, pad_len; - u16 trunc_msg_len = 0; - - /* number of '\0' padding bytes to next message */ - size = msg_used_size(text_len, dict_len, &pad_len); + struct prb_handle h; + char *rbuf; + u32 size; - if (log_make_free_space(size)) { - /* truncate the message if it is too long for empty buffer */ - size = truncate_msg(&text_len, &trunc_msg_len, - &dict_len, &pad_len); - /* survive when the log buffer is too small for trunc_msg */ - if (log_make_free_space(size)) - return 0; - } + size = sizeof(*msg) + text_len + dict_len; - if (log_next_idx + size + sizeof(struct printk_log) > log_buf_len) { + rbuf = prb_reserve(&h, &printk_rb, size); + if (!rbuf) { /* - * This message + an additional empty header does not fit - * at the end of the buffer. Add an empty header with len == 0 - * to signify a wrap around. + * An emergency message would have been printed, but + * it cannot be stored in the log. */ - memset(log_buf + log_next_idx, 0, sizeof(struct printk_log)); - log_next_idx = 0; + prb_inc_lost(&printk_rb); + return 0; } /* fill message */ - msg = (struct printk_log *)(log_buf + log_next_idx); + msg = (struct printk_log *)rbuf; memcpy(log_text(msg), text, text_len); msg->text_len = text_len; - if (trunc_msg_len) { - memcpy(log_text(msg) + text_len, trunc_msg, trunc_msg_len); - msg->text_len += trunc_msg_len; - } memcpy(log_dict(msg), dict, dict_len); msg->dict_len = dict_len; msg->facility = facility; msg->level = level & 7; msg->flags = flags & 0x1f; - if (ts_nsec > 0) - msg->ts_nsec = ts_nsec; - else - msg->ts_nsec = local_clock(); + msg->ts_nsec = ts_nsec; #ifdef CONFIG_PRINTK_CALLER msg->caller_id = caller_id; #endif - memset(log_dict(msg) + dict_len, 0, pad_len); + msg->cpu = cpu; msg->len = size; /* insert message */ - log_next_idx += msg->len; - log_next_seq++; + prb_commit(&h); return msg->text_len; } @@ -739,9 +529,9 @@ static ssize_t msg_print_ext_header(char do_div(ts_usec, 1000); - return scnprintf(buf, size, "%u,%llu,%llu,%c%s;", + return scnprintf(buf, size, "%u,%llu,%llu,%c%s,%hu;", (msg->facility << 3) | msg->level, seq, ts_usec, - msg->flags & LOG_CONT ? 'c' : '-', caller); + msg->flags & LOG_CONT ? 'c' : '-', caller, msg->cpu); } static ssize_t msg_print_ext_body(char *buf, size_t size, @@ -792,13 +582,18 @@ static ssize_t msg_print_ext_body(char * return p - buf; } +#define PRINTK_SPRINT_MAX (LOG_LINE_MAX + PREFIX_MAX) +#define PRINTK_RECORD_MAX (sizeof(struct printk_log) + \ + CONSOLE_EXT_LOG_MAX + PRINTK_SPRINT_MAX) + /* /dev/kmsg - userspace message inject/listen interface */ struct devkmsg_user { u64 seq; - u32 idx; + struct prb_iterator iter; struct ratelimit_state rs; struct mutex lock; char buf[CONSOLE_EXT_LOG_MAX]; + char msgbuf[PRINTK_RECORD_MAX]; }; static __printf(3, 4) __cold @@ -881,9 +676,11 @@ static ssize_t devkmsg_read(struct file size_t count, loff_t *ppos) { struct devkmsg_user *user = file->private_data; + struct prb_iterator backup_iter; struct printk_log *msg; - size_t len; ssize_t ret; + size_t len; + u64 seq; if (!user) return -EBADF; @@ -892,52 +689,63 @@ static ssize_t devkmsg_read(struct file if (ret) return ret; - logbuf_lock_irq(); - while (user->seq == log_next_seq) { - if (file->f_flags & O_NONBLOCK) { - ret = -EAGAIN; - logbuf_unlock_irq(); - goto out; - } + /* make a backup copy in case there is a problem */ + prb_iter_copy(&backup_iter, &user->iter); - logbuf_unlock_irq(); - ret = wait_event_interruptible(log_wait, - user->seq != log_next_seq); - if (ret) - goto out; - logbuf_lock_irq(); + if (file->f_flags & O_NONBLOCK) { + ret = prb_iter_next(&user->iter, &user->msgbuf[0], + sizeof(user->msgbuf), &seq); + } else { + ret = prb_iter_wait_next(&user->iter, &user->msgbuf[0], + sizeof(user->msgbuf), &seq); } - - if (user->seq < log_first_seq) { - /* our last seen message is gone, return error and reset */ - user->idx = log_first_idx; - user->seq = log_first_seq; + if (ret == 0) { + /* end of list */ + ret = -EAGAIN; + goto out; + } else if (ret == -EINVAL) { + /* iterator invalid, return error and reset */ ret = -EPIPE; - logbuf_unlock_irq(); + prb_iter_init(&user->iter, &printk_rb, &user->seq); + goto out; + } else if (ret < 0) { + /* interrupted by signal */ goto out; } - msg = log_from_idx(user->idx); + user->seq++; + if (user->seq < seq) { + ret = -EPIPE; + goto restore_out; + } + + msg = (struct printk_log *)&user->msgbuf[0]; len = msg_print_ext_header(user->buf, sizeof(user->buf), msg, user->seq); len += msg_print_ext_body(user->buf + len, sizeof(user->buf) - len, log_dict(msg), msg->dict_len, log_text(msg), msg->text_len); - user->idx = log_next(user->idx); - user->seq++; - logbuf_unlock_irq(); - if (len > count) { ret = -EINVAL; - goto out; + goto restore_out; } if (copy_to_user(buf, user->buf, len)) { ret = -EFAULT; - goto out; + goto restore_out; } + ret = len; + goto out; +restore_out: + /* + * There was an error, but this message should not be + * lost because of it. Restore the backup and setup + * seq so that it will work with the next read. + */ + prb_iter_copy(&user->iter, &backup_iter); + user->seq = seq - 1; out: mutex_unlock(&user->lock); return ret; @@ -946,19 +754,22 @@ out: static loff_t devkmsg_llseek(struct file *file, loff_t offset, int whence) { struct devkmsg_user *user = file->private_data; - loff_t ret = 0; + loff_t ret; + u64 seq; if (!user) return -EBADF; if (offset) return -ESPIPE; - logbuf_lock_irq(); + ret = mutex_lock_interruptible(&user->lock); + if (ret) + return ret; + switch (whence) { case SEEK_SET: /* the first record */ - user->idx = log_first_idx; - user->seq = log_first_seq; + prb_iter_init(&user->iter, &printk_rb, &user->seq); break; case SEEK_DATA: /* @@ -966,40 +777,87 @@ static loff_t devkmsg_llseek(struct file * like issued by 'dmesg -c'. Reading /dev/kmsg itself * changes no global state, and does not clear anything. */ - user->idx = clear_idx; - user->seq = clear_seq; + for (;;) { + prb_iter_init(&user->iter, &printk_rb, &seq); + ret = prb_iter_seek(&user->iter, clear_seq); + if (ret > 0) { + /* seeked to clear seq */ + user->seq = clear_seq; + break; + } else if (ret == 0) { + /* + * The end of the list was hit without + * ever seeing the clear seq. Just + * seek to the beginning of the list. + */ + prb_iter_init(&user->iter, &printk_rb, + &user->seq); + break; + } + /* iterator invalid, start over */ + + /* reset clear_seq if it is no longer available */ + if (seq > clear_seq) + clear_seq = 0; + } + ret = 0; break; case SEEK_END: /* after the last record */ - user->idx = log_next_idx; - user->seq = log_next_seq; + for (;;) { + ret = prb_iter_next(&user->iter, NULL, 0, &user->seq); + if (ret == 0) + break; + else if (ret > 0) + continue; + /* iterator invalid, start over */ + prb_iter_init(&user->iter, &printk_rb, &user->seq); + } + ret = 0; break; default: ret = -EINVAL; } - logbuf_unlock_irq(); + + mutex_unlock(&user->lock); return ret; } +struct wait_queue_head *printk_wait_queue(void) +{ + /* FIXME: using prb internals! */ + return printk_rb.wq; +} + static __poll_t devkmsg_poll(struct file *file, poll_table *wait) { struct devkmsg_user *user = file->private_data; + struct prb_iterator iter; __poll_t ret = 0; + int rbret; + u64 seq; if (!user) return EPOLLERR|EPOLLNVAL; - poll_wait(file, &log_wait, wait); + poll_wait(file, printk_wait_queue(), wait); - logbuf_lock_irq(); - if (user->seq < log_next_seq) { - /* return error when data has vanished underneath us */ - if (user->seq < log_first_seq) - ret = EPOLLIN|EPOLLRDNORM|EPOLLERR|EPOLLPRI; - else - ret = EPOLLIN|EPOLLRDNORM; - } - logbuf_unlock_irq(); + mutex_lock(&user->lock); + + /* use copy so no actual iteration takes place */ + prb_iter_copy(&iter, &user->iter); + + rbret = prb_iter_next(&iter, &user->msgbuf[0], + sizeof(user->msgbuf), &seq); + if (rbret == 0) + goto out; + + ret = EPOLLIN|EPOLLRDNORM; + + if (rbret < 0 || (seq - user->seq) != 1) + ret |= EPOLLERR|EPOLLPRI; +out: + mutex_unlock(&user->lock); return ret; } @@ -1029,10 +887,7 @@ static int devkmsg_open(struct inode *in mutex_init(&user->lock); - logbuf_lock_irq(); - user->idx = log_first_idx; - user->seq = log_first_seq; - logbuf_unlock_irq(); + prb_iter_init(&user->iter, &printk_rb, &user->seq); file->private_data = user; return 0; @@ -1072,11 +927,6 @@ const struct file_operations kmsg_fops = */ void log_buf_vmcoreinfo_setup(void) { - VMCOREINFO_SYMBOL(log_buf); - VMCOREINFO_SYMBOL(log_buf_len); - VMCOREINFO_SYMBOL(log_first_idx); - VMCOREINFO_SYMBOL(clear_idx); - VMCOREINFO_SYMBOL(log_next_idx); /* * Export struct printk_log size and field offsets. User space tools can * parse it and detect any changes to structure down the line. @@ -1092,6 +942,8 @@ void log_buf_vmcoreinfo_setup(void) } #endif +/* FIXME: no support for buffer resizing */ +#if 0 /* requested log_buf_len from kernel cmdline */ static unsigned long __initdata new_log_buf_len; @@ -1157,29 +1009,16 @@ static void __init log_buf_add_cpu(void) #else /* !CONFIG_SMP */ static inline void log_buf_add_cpu(void) {} #endif /* CONFIG_SMP */ - -static void __init set_percpu_data_ready(void) -{ - printk_safe_init(); - /* Make sure we set this flag only after printk_safe() init is done */ - barrier(); - __printk_percpu_data_ready = true; -} +#endif /* 0 */ void __init setup_log_buf(int early) { +/* FIXME: no support for buffer resizing */ +#if 0 unsigned long flags; char *new_log_buf; unsigned int free; - /* - * Some archs call setup_log_buf() multiple times - first is very - * early, e.g. from setup_arch(), and second - when percpu_areas - * are initialised. - */ - if (!early) - set_percpu_data_ready(); - if (log_buf != __log_buf) return; @@ -1207,6 +1046,7 @@ void __init setup_log_buf(int early) pr_info("log_buf_len: %u bytes\n", log_buf_len); pr_info("early log buf free: %u(%u%%)\n", free, (free * 100) / __LOG_BUF_LEN); +#endif } static bool __read_mostly ignore_loglevel; @@ -1287,6 +1127,11 @@ static inline void boot_delay_msec(int l static bool printk_time = IS_ENABLED(CONFIG_PRINTK_TIME); module_param_named(time, printk_time, bool, S_IRUGO | S_IWUSR); +static size_t print_cpu(u16 cpu, char *buf) +{ + return sprintf(buf, "%03hu: ", cpu); +} + static size_t print_syslog(unsigned int level, char *buf) { return sprintf(buf, "<%u>", level); @@ -1330,6 +1175,7 @@ static size_t print_prefix(const struct buf[len++] = ' '; buf[len] = '\0'; } + len += print_cpu(msg->cpu, buf + len); return len; } @@ -1375,30 +1221,42 @@ static size_t msg_print_text(const struc return len; } -static int syslog_print(char __user *buf, int size) +static int syslog_print(char __user *buf, int size, char *text, + char *msgbuf, int *locked) { - char *text; + struct prb_iterator iter; struct printk_log *msg; int len = 0; - - text = kmalloc(LOG_LINE_MAX + PREFIX_MAX, GFP_KERNEL); - if (!text) - return -ENOMEM; + u64 seq; + int ret; while (size > 0) { size_t n; size_t skip; - logbuf_lock_irq(); - if (syslog_seq < log_first_seq) { - /* messages are gone, move to first one */ - syslog_seq = log_first_seq; - syslog_idx = log_first_idx; - syslog_partial = 0; + for (;;) { + prb_iter_copy(&iter, &syslog_iter); + ret = prb_iter_next(&iter, msgbuf, + PRINTK_RECORD_MAX, &seq); + if (ret < 0) { + /* messages are gone, move to first one */ + prb_iter_init(&syslog_iter, &printk_rb, + &syslog_seq); + syslog_partial = 0; + continue; + } + break; } - if (syslog_seq == log_next_seq) { - logbuf_unlock_irq(); + if (ret == 0) break; + + /* + * If messages have been missed, the partial tracker + * is no longer valid and must be reset. + */ + if (syslog_seq > 0 && seq - 1 != syslog_seq) { + syslog_seq = seq - 1; + syslog_partial = 0; } /* @@ -1408,131 +1266,213 @@ static int syslog_print(char __user *buf if (!syslog_partial) syslog_time = printk_time; + msg = (struct printk_log *)msgbuf; + skip = syslog_partial; - msg = log_from_idx(syslog_idx); n = msg_print_text(msg, true, syslog_time, text, - LOG_LINE_MAX + PREFIX_MAX); + PRINTK_SPRINT_MAX); if (n - syslog_partial <= size) { /* message fits into buffer, move forward */ - syslog_idx = log_next(syslog_idx); - syslog_seq++; + prb_iter_next(&syslog_iter, NULL, 0, &syslog_seq); n -= syslog_partial; syslog_partial = 0; - } else if (!len){ + } else if (!len) { /* partial read(), remember position */ n = size; syslog_partial += n; } else n = 0; - logbuf_unlock_irq(); if (!n) break; + mutex_unlock(&syslog_lock); if (copy_to_user(buf, text + skip, n)) { if (!len) len = -EFAULT; + *locked = 0; break; } + ret = mutex_lock_interruptible(&syslog_lock); len += n; size -= n; buf += n; + + if (ret) { + if (!len) + len = ret; + *locked = 0; + break; + } } - kfree(text); return len; } -static int syslog_print_all(char __user *buf, int size, bool clear) +static int count_remaining(struct prb_iterator *iter, u64 until_seq, + char *msgbuf, int size, bool records, bool time) { - char *text; + struct prb_iterator local_iter; + struct printk_log *msg; int len = 0; - u64 next_seq; u64 seq; - u32 idx; + int ret; + + prb_iter_copy(&local_iter, iter); + for (;;) { + ret = prb_iter_next(&local_iter, msgbuf, size, &seq); + if (ret == 0) { + break; + } else if (ret < 0) { + /* the iter is invalid, restart from head */ + prb_iter_init(&local_iter, &printk_rb, NULL); + len = 0; + continue; + } + + if (until_seq && seq >= until_seq) + break; + + if (records) { + len++; + } else { + msg = (struct printk_log *)msgbuf; + len += msg_print_text(msg, true, time, NULL, 0); + } + } + + return len; +} + +static void syslog_clear(void) +{ + struct prb_iterator iter; + int ret; + + prb_iter_init(&iter, &printk_rb, &clear_seq); + for (;;) { + ret = prb_iter_next(&iter, NULL, 0, &clear_seq); + if (ret == 0) + break; + else if (ret < 0) + prb_iter_init(&iter, &printk_rb, &clear_seq); + } +} + +static int syslog_print_all(char __user *buf, int size, bool clear) +{ + struct prb_iterator iter; + struct printk_log *msg; + char *msgbuf = NULL; + char *text = NULL; + int textlen; + u64 seq = 0; + int len = 0; bool time; + int ret; - text = kmalloc(LOG_LINE_MAX + PREFIX_MAX, GFP_KERNEL); + text = kmalloc(PRINTK_SPRINT_MAX, GFP_KERNEL); if (!text) return -ENOMEM; + msgbuf = kmalloc(PRINTK_RECORD_MAX, GFP_KERNEL); + if (!msgbuf) { + kfree(text); + return -ENOMEM; + } time = printk_time; - logbuf_lock_irq(); + /* - * Find first record that fits, including all following records, - * into the user-provided buffer for this dump. + * Setup iter to last event before clear. Clear may + * be lost, but keep going with a best effort. */ - seq = clear_seq; - idx = clear_idx; - while (seq < log_next_seq) { - struct printk_log *msg = log_from_idx(idx); - - len += msg_print_text(msg, true, time, NULL, 0); - idx = log_next(idx); - seq++; - } - - /* move first record forward until length fits into the buffer */ - seq = clear_seq; - idx = clear_idx; - while (len > size && seq < log_next_seq) { - struct printk_log *msg = log_from_idx(idx); + prb_iter_init(&iter, &printk_rb, NULL); + prb_iter_seek(&iter, clear_seq); + /* count the total bytes after clear */ + len = count_remaining(&iter, 0, msgbuf, PRINTK_RECORD_MAX, + false, time); + + /* move iter forward until length fits into the buffer */ + while (len > size) { + ret = prb_iter_next(&iter, msgbuf, + PRINTK_RECORD_MAX, &seq); + if (ret == 0) { + break; + } else if (ret < 0) { + /* + * The iter is now invalid so clear will + * also be invalid. Restart from the head. + */ + prb_iter_init(&iter, &printk_rb, NULL); + len = count_remaining(&iter, 0, msgbuf, + PRINTK_RECORD_MAX, false, time); + continue; + } + + msg = (struct printk_log *)msgbuf; len -= msg_print_text(msg, true, time, NULL, 0); - idx = log_next(idx); - seq++; - } - /* last message fitting into this dump */ - next_seq = log_next_seq; + if (clear) + clear_seq = seq; + } + /* copy messages to buffer */ len = 0; - while (len >= 0 && seq < next_seq) { - struct printk_log *msg = log_from_idx(idx); - int textlen = msg_print_text(msg, true, time, text, - LOG_LINE_MAX + PREFIX_MAX); + while (len >= 0 && len < size) { + if (clear) + clear_seq = seq; + + ret = prb_iter_next(&iter, msgbuf, + PRINTK_RECORD_MAX, &seq); + if (ret == 0) { + break; + } else if (ret < 0) { + /* + * The iter is now invalid. Make a best + * effort to grab the rest of the log + * from the new head. + */ + prb_iter_init(&iter, &printk_rb, NULL); + continue; + } - idx = log_next(idx); - seq++; + msg = (struct printk_log *)msgbuf; + textlen = msg_print_text(msg, true, time, text, + PRINTK_SPRINT_MAX); + if (textlen < 0) { + len = textlen; + break; + } + + if (len + textlen > size) + break; - logbuf_unlock_irq(); if (copy_to_user(buf + len, text, textlen)) len = -EFAULT; else len += textlen; - logbuf_lock_irq(); - - if (seq < log_first_seq) { - /* messages are gone, move to next one */ - seq = log_first_seq; - idx = log_first_idx; - } } - if (clear) { - clear_seq = log_next_seq; - clear_idx = log_next_idx; - } - logbuf_unlock_irq(); + if (clear && !seq) + syslog_clear(); kfree(text); + kfree(msgbuf); return len; } -static void syslog_clear(void) -{ - logbuf_lock_irq(); - clear_seq = log_next_seq; - clear_idx = log_next_idx; - logbuf_unlock_irq(); -} - int do_syslog(int type, char __user *buf, int len, int source) { bool clear = false; static int saved_console_loglevel = LOGLEVEL_DEFAULT; + struct prb_iterator iter; + char *msgbuf = NULL; + char *text = NULL; + int locked; int error; + int ret; error = check_syslog_permissions(type, source); if (error) @@ -1550,11 +1490,49 @@ int do_syslog(int type, char __user *buf return 0; if (!access_ok(buf, len)) return -EFAULT; - error = wait_event_interruptible(log_wait, - syslog_seq != log_next_seq); + + text = kmalloc(PRINTK_SPRINT_MAX, GFP_KERNEL); + msgbuf = kmalloc(PRINTK_RECORD_MAX, GFP_KERNEL); + if (!text || !msgbuf) { + error = -ENOMEM; + goto out; + } + + error = mutex_lock_interruptible(&syslog_lock); if (error) - return error; - error = syslog_print(buf, len); + goto out; + + /* + * Wait until a first message is available. Use a copy + * because no iteration should occur for syslog now. + */ + for (;;) { + prb_iter_copy(&iter, &syslog_iter); + + mutex_unlock(&syslog_lock); + ret = prb_iter_wait_next(&iter, NULL, 0, NULL); + if (ret == -ERESTARTSYS) { + error = ret; + goto out; + } + error = mutex_lock_interruptible(&syslog_lock); + if (error) + goto out; + + if (ret == -EINVAL) { + prb_iter_init(&syslog_iter, &printk_rb, + &syslog_seq); + syslog_partial = 0; + continue; + } + break; + } + + /* print as much as will fit in the user buffer */ + locked = 1; + error = syslog_print(buf, len, text, msgbuf, &locked); + if (locked) + mutex_unlock(&syslog_lock); break; /* Read/clear last kernel messages */ case SYSLOG_ACTION_READ_CLEAR: @@ -1599,47 +1577,43 @@ int do_syslog(int type, char __user *buf break; /* Number of chars in the log buffer */ case SYSLOG_ACTION_SIZE_UNREAD: - logbuf_lock_irq(); - if (syslog_seq < log_first_seq) { - /* messages are gone, move to first one */ - syslog_seq = log_first_seq; - syslog_idx = log_first_idx; - syslog_partial = 0; - } + msgbuf = kmalloc(PRINTK_RECORD_MAX, GFP_KERNEL); + if (!msgbuf) + return -ENOMEM; + + error = mutex_lock_interruptible(&syslog_lock); + if (error) + goto out; + if (source == SYSLOG_FROM_PROC) { /* * Short-cut for poll(/"proc/kmsg") which simply checks * for pending data, not the size; return the count of * records, not the length. */ - error = log_next_seq - syslog_seq; + error = count_remaining(&syslog_iter, 0, msgbuf, + PRINTK_RECORD_MAX, true, + printk_time); } else { - u64 seq = syslog_seq; - u32 idx = syslog_idx; - bool time = syslog_partial ? syslog_time : printk_time; - - while (seq < log_next_seq) { - struct printk_log *msg = log_from_idx(idx); - - error += msg_print_text(msg, true, time, NULL, - 0); - time = printk_time; - idx = log_next(idx); - seq++; - } + error = count_remaining(&syslog_iter, 0, msgbuf, + PRINTK_RECORD_MAX, false, + printk_time); error -= syslog_partial; } - logbuf_unlock_irq(); + + mutex_unlock(&syslog_lock); break; /* Size of the log buffer */ case SYSLOG_ACTION_SIZE_BUFFER: - error = log_buf_len; + error = prb_buffer_size(&printk_rb); break; default: error = -EINVAL; break; } - +out: + kfree(msgbuf); + kfree(text); return error; } @@ -1648,144 +1622,128 @@ SYSCALL_DEFINE3(syslog, int, type, char return do_syslog(type, buf, len, SYSLOG_FROM_READER); } -/* - * Special console_lock variants that help to reduce the risk of soft-lockups. - * They allow to pass console_lock to another printk() call using a busy wait. - */ +int printk_delay_msec __read_mostly; -#ifdef CONFIG_LOCKDEP -static struct lockdep_map console_owner_dep_map = { - .name = "console_owner" -}; -#endif +static inline void printk_delay(int level) +{ + boot_delay_msec(level); + if (unlikely(printk_delay_msec)) { + int m = printk_delay_msec; -static DEFINE_RAW_SPINLOCK(console_owner_lock); -static struct task_struct *console_owner; -static bool console_waiter; + while (m--) { + mdelay(1); + touch_nmi_watchdog(); + } + } +} -/** - * console_lock_spinning_enable - mark beginning of code where another - * thread might safely busy wait - * - * This basically converts console_lock into a spinlock. This marks - * the section where the console_lock owner can not sleep, because - * there may be a waiter spinning (like a spinlock). Also it must be - * ready to hand over the lock at the end of the section. - */ -static void console_lock_spinning_enable(void) -{ - raw_spin_lock(&console_owner_lock); - console_owner = current; - raw_spin_unlock(&console_owner_lock); +static void print_console_dropped(struct console *con, u64 count) +{ + char text[64]; + int len; - /* The waiter may spin on us after setting console_owner */ - spin_acquire(&console_owner_dep_map, 0, 0, _THIS_IP_); + len = sprintf(text, "** %llu printk message%s dropped **\n", + count, count > 1 ? "s" : ""); + con->write(con, text, len); } -/** - * console_lock_spinning_disable_and_check - mark end of code where another - * thread was able to busy wait and check if there is a waiter - * - * This is called at the end of the section where spinning is allowed. - * It has two functions. First, it is a signal that it is no longer - * safe to start busy waiting for the lock. Second, it checks if - * there is a busy waiter and passes the lock rights to her. - * - * Important: Callers lose the lock if there was a busy waiter. - * They must not touch items synchronized by console_lock - * in this case. - * - * Return: 1 if the lock rights were passed, 0 otherwise. - */ -static int console_lock_spinning_disable_and_check(void) +static void format_text(struct printk_log *msg, u64 seq, + char *ext_text, size_t *ext_len, + char *text, size_t *len, bool time) { - int waiter; - - raw_spin_lock(&console_owner_lock); - waiter = READ_ONCE(console_waiter); - console_owner = NULL; - raw_spin_unlock(&console_owner_lock); + if (suppress_message_printing(msg->level)) { + /* + * Skip record that has level above the console + * loglevel and update each console's local seq. + */ + *len = 0; + *ext_len = 0; + return; + } - if (!waiter) { - spin_release(&console_owner_dep_map, _THIS_IP_); - return 0; + *len = msg_print_text(msg, console_msg_format & MSG_FORMAT_SYSLOG, + time, text, PRINTK_SPRINT_MAX); + if (nr_ext_console_drivers) { + *ext_len = msg_print_ext_header(ext_text, CONSOLE_EXT_LOG_MAX, + msg, seq); + *ext_len += msg_print_ext_body(ext_text + *ext_len, + CONSOLE_EXT_LOG_MAX - *ext_len, + log_dict(msg), msg->dict_len, + log_text(msg), msg->text_len); + } else { + *ext_len = 0; } +} - /* The waiter is now free to continue */ - WRITE_ONCE(console_waiter, false); +static void printk_write_history(struct console *con, u64 master_seq) +{ + struct prb_iterator iter; + bool time = printk_time; + static char *ext_text; + static char *text; + static char *buf; + u64 seq; - spin_release(&console_owner_dep_map, _THIS_IP_); + ext_text = kmalloc(CONSOLE_EXT_LOG_MAX, GFP_KERNEL); + text = kmalloc(PRINTK_SPRINT_MAX, GFP_KERNEL); + buf = kmalloc(PRINTK_RECORD_MAX, GFP_KERNEL); + if (!ext_text || !text || !buf) + return; - /* - * Hand off console_lock to waiter. The waiter will perform - * the up(). After this, the waiter is the console_lock owner. - */ - mutex_release(&console_lock_dep_map, _THIS_IP_); - return 1; -} + if (!(con->flags & CON_ENABLED)) + goto out; -/** - * console_trylock_spinning - try to get console_lock by busy waiting - * - * This allows to busy wait for the console_lock when the current - * owner is running in specially marked sections. It means that - * the current owner is running and cannot reschedule until it - * is ready to lose the lock. - * - * Return: 1 if we got the lock, 0 othrewise - */ -static int console_trylock_spinning(void) -{ - struct task_struct *owner = NULL; - bool waiter; - bool spin = false; - unsigned long flags; + if (!con->write) + goto out; - if (console_trylock()) - return 1; + if (!cpu_online(raw_smp_processor_id()) && + !(con->flags & CON_ANYTIME)) + goto out; - printk_safe_enter_irqsave(flags); + prb_iter_init(&iter, &printk_rb, NULL); - raw_spin_lock(&console_owner_lock); - owner = READ_ONCE(console_owner); - waiter = READ_ONCE(console_waiter); - if (!waiter && owner && owner != current) { - WRITE_ONCE(console_waiter, true); - spin = true; - } - raw_spin_unlock(&console_owner_lock); - - /* - * If there is an active printk() writing to the - * consoles, instead of having it write our data too, - * see if we can offload that load from the active - * printer, and do some printing ourselves. - * Go into a spin only if there isn't already a waiter - * spinning, and there is an active printer, and - * that active printer isn't us (recursive printk?). - */ - if (!spin) { - printk_safe_exit_irqrestore(flags); - return 0; - } + for (;;) { + struct printk_log *msg; + size_t ext_len; + size_t len; + int ret; - /* We spin waiting for the owner to release us */ - spin_acquire(&console_owner_dep_map, 0, 0, _THIS_IP_); - /* Owner will clear console_waiter on hand off */ - while (READ_ONCE(console_waiter)) - cpu_relax(); - spin_release(&console_owner_dep_map, _THIS_IP_); + ret = prb_iter_next(&iter, buf, PRINTK_RECORD_MAX, &seq); + if (ret == 0) { + break; + } else if (ret < 0) { + prb_iter_init(&iter, &printk_rb, NULL); + continue; + } - printk_safe_exit_irqrestore(flags); - /* - * The owner passed the console lock to us. - * Since we did not spin on console lock, annotate - * this as a trylock. Otherwise lockdep will - * complain. - */ - mutex_acquire(&console_lock_dep_map, 0, 1, _THIS_IP_); + if (seq > master_seq) + break; - return 1; + con->printk_seq++; + if (con->printk_seq < seq) { + print_console_dropped(con, seq - con->printk_seq); + con->printk_seq = seq; + } + + msg = (struct printk_log *)buf; + format_text(msg, master_seq, ext_text, &ext_len, text, + &len, time); + + if (len == 0 && ext_len == 0) + continue; + + if (con->flags & CON_EXTENDED) + con->write(con, ext_text, ext_len); + else + con->write(con, text, len); + + printk_delay(msg->level); + } +out: + con->wrote_history = 1; + kfree(ext_text); + kfree(text); + kfree(buf); } /* @@ -1793,8 +1751,9 @@ static int console_trylock_spinning(void * log_buf[start] to log_buf[end - 1]. * The console_lock must be held. */ -static void call_console_drivers(const char *ext_text, size_t ext_len, - const char *text, size_t len) +static void call_console_drivers(u64 seq, const char *ext_text, size_t ext_len, + const char *text, size_t len, int level, + int facility) { struct console *con; @@ -1804,33 +1763,44 @@ static void call_console_drivers(const c return; for_each_console(con) { - if (exclusive_console && con != exclusive_console) - continue; if (!(con->flags & CON_ENABLED)) continue; + if (!con->wrote_history) { + if (con->flags & CON_PRINTBUFFER) { + printk_write_history(con, seq); + continue; + } + con->wrote_history = 1; + con->printk_seq = seq - 1; + } + if (con->flags & CON_BOOT && facility == 0) { + /* skip boot messages, already printed */ + if (con->printk_seq < seq) + con->printk_seq = seq; + continue; + } if (!con->write) continue; - if (!cpu_online(smp_processor_id()) && + if (!cpu_online(raw_smp_processor_id()) && !(con->flags & CON_ANYTIME)) continue; - if (con->flags & CON_EXTENDED) - con->write(con, ext_text, ext_len); - else - con->write(con, text, len); - } -} + if (con->printk_seq >= seq) + continue; -int printk_delay_msec __read_mostly; + con->printk_seq++; + if (con->printk_seq < seq) { + print_console_dropped(con, seq - con->printk_seq); + con->printk_seq = seq; + } -static inline void printk_delay(void) -{ - if (unlikely(printk_delay_msec)) { - int m = printk_delay_msec; + /* for supressed messages, only seq is updated */ + if (len == 0 && ext_len == 0) + continue; - while (m--) { - mdelay(1); - touch_nmi_watchdog(); - } + if (con->flags & CON_EXTENDED) + con->write(con, ext_text, ext_len); + else + con->write(con, text, len); } } @@ -1850,101 +1820,95 @@ static struct cont { char buf[LOG_LINE_MAX]; size_t len; /* length == 0 means unused buffer */ u32 caller_id; /* printk_caller_id() of first print */ + int cpu_owner; /* cpu of first print */ u64 ts_nsec; /* time of first print */ u8 level; /* log level of first message */ u8 facility; /* log facility of first message */ enum log_flags flags; /* prefix, newline flags */ -} cont; +} cont[2]; -static void cont_flush(void) +static void cont_flush(int ctx) { - if (cont.len == 0) + struct cont *c = &cont[ctx]; + + if (c->len == 0) return; - log_store(cont.caller_id, cont.facility, cont.level, cont.flags, - cont.ts_nsec, NULL, 0, cont.buf, cont.len); - cont.len = 0; + log_store(c->caller_id, c->facility, c->level, c->flags, + c->ts_nsec, c->cpu_owner, NULL, 0, c->buf, c->len); + c->len = 0; } -static bool cont_add(u32 caller_id, int facility, int level, +static void cont_add(int ctx, int cpu, u32 caller_id, int facility, int level, enum log_flags flags, const char *text, size_t len) { + struct cont *c = &cont[ctx]; + + if (cpu != c->cpu_owner || !(flags & LOG_CONT)) + cont_flush(ctx); + /* If the line gets too long, split it up in separate records. */ - if (cont.len + len > sizeof(cont.buf)) { - cont_flush(); - return false; - } + while (c->len + len > sizeof(c->buf)) + cont_flush(ctx); - if (!cont.len) { - cont.facility = facility; - cont.level = level; - cont.caller_id = caller_id; - cont.ts_nsec = local_clock(); - cont.flags = flags; + if (!c->len) { + c->facility = facility; + c->level = level; + c->caller_id = caller_id; + c->ts_nsec = local_clock(); + c->flags = flags; + c->cpu_owner = cpu; } - memcpy(cont.buf + cont.len, text, len); - cont.len += len; + memcpy(c->buf + c->len, text, len); + c->len += len; // The original flags come from the first line, // but later continuations can add a newline. if (flags & LOG_NEWLINE) { - cont.flags |= LOG_NEWLINE; - cont_flush(); + c->flags |= LOG_NEWLINE; + cont_flush(ctx); } - - return true; } -static size_t log_output(int facility, int level, enum log_flags lflags, const char *dict, size_t dictlen, char *text, size_t text_len) +/* ring buffer used as memory allocator for temporary sprint buffers */ +DECLARE_STATIC_PRINTKRB(sprint_rb, + ilog2(PRINTK_RECORD_MAX + sizeof(struct prb_entry) + + sizeof(long)) + 2, &printk_cpulock); + +asmlinkage int vprintk_emit(int facility, int level, + const char *dict, size_t dictlen, + const char *fmt, va_list args) { const u32 caller_id = printk_caller_id(); + int ctx = !!in_nmi(); + enum log_flags lflags = 0; + int printed_len = 0; + struct prb_handle h; + size_t text_len; + u64 ts_nsec; + char *text; + char *rbuf; + int cpu; - /* - * If an earlier line was buffered, and we're a continuation - * write from the same context, try to add it to the buffer. - */ - if (cont.len) { - if (cont.caller_id == caller_id && (lflags & LOG_CONT)) { - if (cont_add(caller_id, facility, level, lflags, text, text_len)) - return text_len; - } - /* Otherwise, make sure it's flushed */ - cont_flush(); - } - - /* Skip empty continuation lines that couldn't be added - they just flush */ - if (!text_len && (lflags & LOG_CONT)) - return 0; + ts_nsec = local_clock(); - /* If it doesn't end in a newline, try to buffer the current line */ - if (!(lflags & LOG_NEWLINE)) { - if (cont_add(caller_id, facility, level, lflags, text, text_len)) - return text_len; + rbuf = prb_reserve(&h, &sprint_rb, PRINTK_SPRINT_MAX); + if (!rbuf) { + prb_inc_lost(&printk_rb); + return printed_len; } - /* Store it in the record log */ - return log_store(caller_id, facility, level, lflags, 0, - dict, dictlen, text, text_len); -} - -/* Must be called under logbuf_lock. */ -int vprintk_store(int facility, int level, - const char *dict, size_t dictlen, - const char *fmt, va_list args) -{ - static char textbuf[LOG_LINE_MAX]; - char *text = textbuf; - size_t text_len; - enum log_flags lflags = 0; + cpu = raw_smp_processor_id(); /* - * The printf needs to come first; we need the syslog - * prefix which might be passed-in as a parameter. + * If this turns out to be an emergency message, there + * may need to be a prefix added. Leave room for it. */ - text_len = vscnprintf(text, sizeof(textbuf), fmt, args); + text = rbuf + PREFIX_MAX; + text_len = vscnprintf(text, PRINTK_SPRINT_MAX - PREFIX_MAX, fmt, args); - /* mark and strip a trailing newline */ + /* strip and flag a trailing newline */ if (text_len && text[text_len-1] == '\n') { text_len--; lflags |= LOG_NEWLINE; @@ -1975,62 +1939,37 @@ int vprintk_store(int facility, int leve if (dict) lflags |= LOG_NEWLINE; - return log_output(facility, level, lflags, - dict, dictlen, text, text_len); -} - -asmlinkage int vprintk_emit(int facility, int level, - const char *dict, size_t dictlen, - const char *fmt, va_list args) -{ - int printed_len; - bool in_sched = false, pending_output; - unsigned long flags; - u64 curr_log_seq; - - /* Suppress unimportant messages after panic happens */ - if (unlikely(suppress_printk)) - return 0; - - if (level == LOGLEVEL_SCHED) { - level = LOGLEVEL_DEFAULT; - in_sched = true; + /* + * NOTE: + * - rbuf points to beginning of allocated buffer + * - text points to beginning of text + * - there is room before text for prefix + */ + if (facility == 0) { + /* only the kernel can create emergency messages */ + printk_emergency(rbuf, level & 7, ts_nsec, cpu, text, text_len); } - boot_delay_msec(level); - printk_delay(); - - /* This stops the holder of console_sem just where we want him */ - logbuf_lock_irqsave(flags); - curr_log_seq = log_next_seq; - printed_len = vprintk_store(facility, level, dict, dictlen, fmt, args); - pending_output = (curr_log_seq != log_next_seq); - logbuf_unlock_irqrestore(flags); - - /* If called from the scheduler, we can not call up(). */ - if (!in_sched && pending_output) { - /* - * Disable preemption to avoid being preempted while holding - * console_sem which would prevent anyone from printing to - * console - */ - preempt_disable(); - /* - * Try to acquire and then immediately release the console - * semaphore. The release will print out buffers and wake up - * /dev/kmsg and syslog() users. - */ - if (console_trylock_spinning()) - console_unlock(); - preempt_enable(); + if ((lflags & LOG_CONT) || !(lflags & LOG_NEWLINE)) { + cont_add(ctx, cpu, caller_id, facility, level, lflags, text, text_len); + printed_len = text_len; + } else { + if (cpu == cont[ctx].cpu_owner) + cont_flush(ctx); + printed_len = log_store(caller_id, facility, level, lflags, ts_nsec, cpu, + dict, dictlen, text, text_len); } - if (pending_output) - wake_up_klogd(); + prb_commit(&h); return printed_len; } EXPORT_SYMBOL(vprintk_emit); +static __printf(1, 0) int vprintk_func(const char *fmt, va_list args) +{ + return vprintk_emit(0, LOGLEVEL_DEFAULT, NULL, 0, fmt, args); +} + asmlinkage int vprintk(const char *fmt, va_list args) { return vprintk_func(fmt, args); @@ -2087,39 +2026,6 @@ asmlinkage __visible int printk(const ch return r; } EXPORT_SYMBOL(printk); - -#else /* CONFIG_PRINTK */ - -#define LOG_LINE_MAX 0 -#define PREFIX_MAX 0 -#define printk_time false - -static u64 syslog_seq; -static u32 syslog_idx; -static u64 console_seq; -static u32 console_idx; -static u64 exclusive_console_stop_seq; -static u64 log_first_seq; -static u32 log_first_idx; -static u64 log_next_seq; -static char *log_text(const struct printk_log *msg) { return NULL; } -static char *log_dict(const struct printk_log *msg) { return NULL; } -static struct printk_log *log_from_idx(u32 idx) { return NULL; } -static u32 log_next(u32 idx) { return 0; } -static ssize_t msg_print_ext_header(char *buf, size_t size, - struct printk_log *msg, - u64 seq) { return 0; } -static ssize_t msg_print_ext_body(char *buf, size_t size, - char *dict, size_t dict_len, - char *text, size_t text_len) { return 0; } -static void console_lock_spinning_enable(void) { } -static int console_lock_spinning_disable_and_check(void) { return 0; } -static void call_console_drivers(const char *ext_text, size_t ext_len, - const char *text, size_t len) {} -static size_t msg_print_text(const struct printk_log *msg, bool syslog, - bool time, char *buf, size_t size) { return 0; } -static bool suppress_message_printing(int level) { return false; } - #endif /* CONFIG_PRINTK */ #ifdef CONFIG_EARLY_PRINTK @@ -2350,187 +2256,23 @@ int is_console_locked(void) } EXPORT_SYMBOL(is_console_locked); -/* - * Check if we have any console that is capable of printing while cpu is - * booting or shutting down. Requires console_sem. - */ -static int have_callable_console(void) -{ - struct console *con; - - for_each_console(con) - if ((con->flags & CON_ENABLED) && - (con->flags & CON_ANYTIME)) - return 1; - - return 0; -} - -/* - * Can we actually use the console at this time on this cpu? - * - * Console drivers may assume that per-cpu resources have been allocated. So - * unless they're explicitly marked as being able to cope (CON_ANYTIME) don't - * call them until this CPU is officially up. - */ -static inline int can_use_console(void) -{ - return cpu_online(raw_smp_processor_id()) || have_callable_console(); -} - /** * console_unlock - unlock the console system * * Releases the console_lock which the caller holds on the console system * and the console driver list. * - * While the console_lock was held, console output may have been buffered - * by printk(). If this is the case, console_unlock(); emits - * the output prior to releasing the lock. - * - * If there is output waiting, we wake /dev/kmsg and syslog() users. - * * console_unlock(); may be called from any context. */ void console_unlock(void) { - static char ext_text[CONSOLE_EXT_LOG_MAX]; - static char text[LOG_LINE_MAX + PREFIX_MAX]; - unsigned long flags; - bool do_cond_resched, retry; - if (console_suspended) { up_console_sem(); return; } - /* - * Console drivers are called with interrupts disabled, so - * @console_may_schedule should be cleared before; however, we may - * end up dumping a lot of lines, for example, if called from - * console registration path, and should invoke cond_resched() - * between lines if allowable. Not doing so can cause a very long - * scheduling stall on a slow console leading to RCU stall and - * softlockup warnings which exacerbate the issue with more - * messages practically incapacitating the system. - * - * console_trylock() is not able to detect the preemptive - * context reliably. Therefore the value must be stored before - * and cleared after the the "again" goto label. - */ - do_cond_resched = console_may_schedule; -again: - console_may_schedule = 0; - - /* - * We released the console_sem lock, so we need to recheck if - * cpu is online and (if not) is there at least one CON_ANYTIME - * console. - */ - if (!can_use_console()) { - console_locked = 0; - up_console_sem(); - return; - } - - for (;;) { - struct printk_log *msg; - size_t ext_len = 0; - size_t len; - - printk_safe_enter_irqsave(flags); - raw_spin_lock(&logbuf_lock); - if (console_seq < log_first_seq) { - len = sprintf(text, - "** %llu printk messages dropped **\n", - log_first_seq - console_seq); - - /* messages are gone, move to first one */ - console_seq = log_first_seq; - console_idx = log_first_idx; - } else { - len = 0; - } -skip: - if (console_seq == log_next_seq) - break; - - msg = log_from_idx(console_idx); - if (suppress_message_printing(msg->level)) { - /* - * Skip record we have buffered and already printed - * directly to the console when we received it, and - * record that has level above the console loglevel. - */ - console_idx = log_next(console_idx); - console_seq++; - goto skip; - } - - /* Output to all consoles once old messages replayed. */ - if (unlikely(exclusive_console && - console_seq >= exclusive_console_stop_seq)) { - exclusive_console = NULL; - } - - len += msg_print_text(msg, - console_msg_format & MSG_FORMAT_SYSLOG, - printk_time, text + len, sizeof(text) - len); - if (nr_ext_console_drivers) { - ext_len = msg_print_ext_header(ext_text, - sizeof(ext_text), - msg, console_seq); - ext_len += msg_print_ext_body(ext_text + ext_len, - sizeof(ext_text) - ext_len, - log_dict(msg), msg->dict_len, - log_text(msg), msg->text_len); - } - console_idx = log_next(console_idx); - console_seq++; - raw_spin_unlock(&logbuf_lock); - - /* - * While actively printing out messages, if another printk() - * were to occur on another CPU, it may wait for this one to - * finish. This task can not be preempted if there is a - * waiter waiting to take over. - */ - console_lock_spinning_enable(); - - stop_critical_timings(); /* don't trace print latency */ - call_console_drivers(ext_text, ext_len, text, len); - start_critical_timings(); - - if (console_lock_spinning_disable_and_check()) { - printk_safe_exit_irqrestore(flags); - return; - } - - printk_safe_exit_irqrestore(flags); - - if (do_cond_resched) - cond_resched(); - } - console_locked = 0; - - raw_spin_unlock(&logbuf_lock); - up_console_sem(); - - /* - * Someone could have filled up the buffer again, so re-check if there's - * something to flush. In case we cannot trylock the console_sem again, - * there's a new owner and the console_unlock() from them will do the - * flush, no worries. - */ - raw_spin_lock(&logbuf_lock); - retry = console_seq != log_next_seq; - raw_spin_unlock(&logbuf_lock); - printk_safe_exit_irqrestore(flags); - - if (retry && console_trylock()) - goto again; } EXPORT_SYMBOL(console_unlock); @@ -2581,24 +2323,10 @@ void console_unblank(void) void console_flush_on_panic(enum con_flush_mode mode) { /* - * If someone else is holding the console lock, trylock will fail - * and may_schedule may be set. Ignore and proceed to unlock so - * that messages are flushed out. As this can be called from any - * context and we don't want to get preempted while flushing, - * ensure may_schedule is cleared. + * FIXME: This is currently a NOP. Emergency messages will have been + * printed, but what about if write_atomic is not available on the + * console? What if the printk kthread is still alive? */ - console_trylock(); - console_may_schedule = 0; - - if (mode == CONSOLE_REPLAY_ALL) { - unsigned long flags; - - logbuf_lock_irqsave(flags); - console_seq = log_first_seq; - console_idx = log_first_idx; - logbuf_unlock_irqrestore(flags); - } - console_unlock(); } /* @@ -2676,7 +2404,6 @@ early_param("keep_bootcon", keep_bootcon void register_console(struct console *newcon) { int i; - unsigned long flags; struct console *bcon = NULL; struct console_cmdline *c; static bool has_preferred; @@ -2792,27 +2519,6 @@ void register_console(struct console *ne if (newcon->flags & CON_EXTENDED) nr_ext_console_drivers++; - if (newcon->flags & CON_PRINTBUFFER) { - /* - * console_unlock(); will print out the buffered messages - * for us. - */ - logbuf_lock_irqsave(flags); - /* - * We're about to replay the log buffer. Only do this to the - * just-registered console to avoid excessive message spam to - * the already-registered consoles. - * - * Set exclusive_console with disabled interrupts to reduce - * race window with eventual console_flush_on_panic() that - * ignores console_lock. - */ - exclusive_console = newcon; - exclusive_console_stop_seq = console_seq; - console_seq = syslog_seq; - console_idx = syslog_idx; - logbuf_unlock_irqrestore(flags); - } console_unlock(); console_sysfs_notify(); @@ -2822,6 +2528,10 @@ void register_console(struct console *ne * boot consoles, real consoles, etc - this is to ensure that end * users know there might be something in the kernel's log buffer that * went to the bootconsole (that they do not see on the real console) + * + * This message is also important because it will trigger the + * printk kthread to begin dumping the log buffer to the newly + * registered console. */ pr_info("%sconsole [%s%d] enabled\n", (newcon->flags & CON_BOOT) ? "boot" : "" , @@ -2965,65 +2675,75 @@ static int __init printk_late_init(void) late_initcall(printk_late_init); #if defined CONFIG_PRINTK -/* - * Delayed printk version, for scheduler-internal messages: - */ -#define PRINTK_PENDING_WAKEUP 0x01 -#define PRINTK_PENDING_OUTPUT 0x02 - -static DEFINE_PER_CPU(int, printk_pending); - -static void wake_up_klogd_work_func(struct irq_work *irq_work) +static int printk_kthread_func(void *data) { - int pending = __this_cpu_xchg(printk_pending, 0); + struct prb_iterator iter; + struct printk_log *msg; + size_t ext_len; + char *ext_text; + u64 master_seq; + size_t len; + char *text; + char *buf; + int ret; - if (pending & PRINTK_PENDING_OUTPUT) { - /* If trylock fails, someone else is doing the printing */ - if (console_trylock()) - console_unlock(); - } + ext_text = kmalloc(CONSOLE_EXT_LOG_MAX, GFP_KERNEL); + text = kmalloc(PRINTK_SPRINT_MAX, GFP_KERNEL); + buf = kmalloc(PRINTK_RECORD_MAX, GFP_KERNEL); + if (!ext_text || !text || !buf) + return -1; - if (pending & PRINTK_PENDING_WAKEUP) - wake_up_interruptible(&log_wait); -} + prb_iter_init(&iter, &printk_rb, NULL); -static DEFINE_PER_CPU(struct irq_work, wake_up_klogd_work) = { - .func = wake_up_klogd_work_func, - .flags = ATOMIC_INIT(IRQ_WORK_LAZY), -}; + /* the printk kthread never exits */ + for (;;) { + ret = prb_iter_wait_next(&iter, buf, + PRINTK_RECORD_MAX, &master_seq); + if (ret == -ERESTARTSYS) { + continue; + } else if (ret < 0) { + /* iterator invalid, start over */ + prb_iter_init(&iter, &printk_rb, NULL); + continue; + } -void wake_up_klogd(void) -{ - if (!printk_percpu_data_ready()) - return; + msg = (struct printk_log *)buf; + format_text(msg, master_seq, ext_text, &ext_len, text, + &len, printk_time); - preempt_disable(); - if (waitqueue_active(&log_wait)) { - this_cpu_or(printk_pending, PRINTK_PENDING_WAKEUP); - irq_work_queue(this_cpu_ptr(&wake_up_klogd_work)); + console_lock(); + console_may_schedule = 0; + call_console_drivers(master_seq, ext_text, ext_len, text, len, + msg->level, msg->facility); + if (len > 0 || ext_len > 0) + printk_delay(msg->level); + console_unlock(); } - preempt_enable(); -} -void defer_console_output(void) -{ - if (!printk_percpu_data_ready()) - return; + kfree(ext_text); + kfree(text); + kfree(buf); - preempt_disable(); - __this_cpu_or(printk_pending, PRINTK_PENDING_OUTPUT); - irq_work_queue(this_cpu_ptr(&wake_up_klogd_work)); - preempt_enable(); + return 0; } -int vprintk_deferred(const char *fmt, va_list args) +static int __init init_printk_kthread(void) { - int r; + struct task_struct *thread; - r = vprintk_emit(0, LOGLEVEL_SCHED, NULL, 0, fmt, args); - defer_console_output(); + thread = kthread_run(printk_kthread_func, NULL, "printk"); + if (IS_ERR(thread)) { + pr_err("printk: unable to create printing thread\n"); + return PTR_ERR(thread); + } - return r; + return 0; +} +late_initcall(init_printk_kthread); + +static int vprintk_deferred(const char *fmt, va_list args) +{ + return vprintk_emit(0, LOGLEVEL_DEFAULT, NULL, 0, fmt, args); } int printk_deferred(const char *fmt, ...) @@ -3145,8 +2865,8 @@ module_param_named(always_kmsg_dump, alw */ void kmsg_dump(enum kmsg_dump_reason reason) { + struct kmsg_dumper dumper_local; struct kmsg_dumper *dumper; - unsigned long flags; if ((reason > KMSG_DUMP_OOPS) && !always_kmsg_dump) return; @@ -3156,21 +2876,18 @@ void kmsg_dump(enum kmsg_dump_reason rea if (dumper->max_reason && reason > dumper->max_reason) continue; - /* initialize iterator with data about the stored records */ - dumper->active = true; + /* + * use a local copy to avoid modifying the + * iterator used by any other cpus/contexts + */ + memcpy(&dumper_local, dumper, sizeof(dumper_local)); - logbuf_lock_irqsave(flags); - dumper->cur_seq = clear_seq; - dumper->cur_idx = clear_idx; - dumper->next_seq = log_next_seq; - dumper->next_idx = log_next_idx; - logbuf_unlock_irqrestore(flags); + /* initialize iterator with data about the stored records */ + dumper_local.active = true; + kmsg_dump_rewind(&dumper_local); /* invoke dumper which will iterate over records */ - dumper->dump(dumper, reason); - - /* reset iterator */ - dumper->active = false; + dumper_local.dump(&dumper_local, reason); } rcu_read_unlock(); } @@ -3197,33 +2914,67 @@ void kmsg_dump(enum kmsg_dump_reason rea bool kmsg_dump_get_line_nolock(struct kmsg_dumper *dumper, bool syslog, char *line, size_t size, size_t *len) { + struct prb_iterator iter; struct printk_log *msg; - size_t l = 0; - bool ret = false; + struct prb_handle h; + bool cont = false; + char *msgbuf; + char *rbuf; + size_t l; + u64 seq; + int ret; if (!dumper->active) - goto out; + return cont; + + rbuf = prb_reserve(&h, &sprint_rb, PRINTK_RECORD_MAX); + if (!rbuf) + return cont; + msgbuf = rbuf; +retry: + for (;;) { + prb_iter_init(&iter, &printk_rb, &seq); + + if (dumper->line_seq == seq) { + /* already where we want to be */ + break; + } else if (dumper->line_seq < seq) { + /* messages are gone, move to first available one */ + dumper->line_seq = seq; + break; + } - if (dumper->cur_seq < log_first_seq) { - /* messages are gone, move to first available one */ - dumper->cur_seq = log_first_seq; - dumper->cur_idx = log_first_idx; + ret = prb_iter_seek(&iter, dumper->line_seq); + if (ret > 0) { + /* seeked to line_seq */ + break; + } else if (ret == 0) { + /* + * The end of the list was hit without ever seeing + * line_seq. Reset it to the beginning of the list. + */ + prb_iter_init(&iter, &printk_rb, &dumper->line_seq); + break; + } + /* iterator invalid, start over */ } - /* last entry */ - if (dumper->cur_seq >= log_next_seq) + ret = prb_iter_next(&iter, msgbuf, PRINTK_RECORD_MAX, + &dumper->line_seq); + if (ret == 0) goto out; + else if (ret < 0) + goto retry; - msg = log_from_idx(dumper->cur_idx); + msg = (struct printk_log *)msgbuf; l = msg_print_text(msg, syslog, printk_time, line, size); - dumper->cur_idx = log_next(dumper->cur_idx); - dumper->cur_seq++; - ret = true; -out: if (len) *len = l; - return ret; + cont = true; +out: + prb_commit(&h); + return cont; } /** @@ -3246,12 +2997,9 @@ out: bool kmsg_dump_get_line(struct kmsg_dumper *dumper, bool syslog, char *line, size_t size, size_t *len) { - unsigned long flags; bool ret; - logbuf_lock_irqsave(flags); ret = kmsg_dump_get_line_nolock(dumper, syslog, line, size, len); - logbuf_unlock_irqrestore(flags); return ret; } @@ -3279,74 +3027,101 @@ EXPORT_SYMBOL_GPL(kmsg_dump_get_line); bool kmsg_dump_get_buffer(struct kmsg_dumper *dumper, bool syslog, char *buf, size_t size, size_t *len) { - unsigned long flags; - u64 seq; - u32 idx; - u64 next_seq; - u32 next_idx; - size_t l = 0; - bool ret = false; + struct prb_iterator iter; bool time = printk_time; + struct printk_log *msg; + u64 new_end_seq = 0; + struct prb_handle h; + bool cont = false; + char *msgbuf; + u64 end_seq; + int textlen; + u64 seq = 0; + char *rbuf; + int l = 0; + int ret; if (!dumper->active) - goto out; + return cont; - logbuf_lock_irqsave(flags); - if (dumper->cur_seq < log_first_seq) { - /* messages are gone, move to first available one */ - dumper->cur_seq = log_first_seq; - dumper->cur_idx = log_first_idx; - } + rbuf = prb_reserve(&h, &sprint_rb, PRINTK_RECORD_MAX); + if (!rbuf) + return cont; + msgbuf = rbuf; - /* last entry */ - if (dumper->cur_seq >= dumper->next_seq) { - logbuf_unlock_irqrestore(flags); - goto out; - } + prb_iter_init(&iter, &printk_rb, NULL); - /* calculate length of entire buffer */ - seq = dumper->cur_seq; - idx = dumper->cur_idx; - while (seq < dumper->next_seq) { - struct printk_log *msg = log_from_idx(idx); + /* + * seek to the start record, which is set/modified + * by kmsg_dump_get_line_nolock() + */ + ret = prb_iter_seek(&iter, dumper->line_seq); + if (ret <= 0) + prb_iter_init(&iter, &printk_rb, &seq); - l += msg_print_text(msg, true, time, NULL, 0); - idx = log_next(idx); - seq++; + /* work with a local end seq to have a constant value */ + end_seq = dumper->buffer_end_seq; + if (!end_seq) { + /* initialize end seq to "infinity" */ + end_seq = -1; + dumper->buffer_end_seq = end_seq; } +retry: + if (seq >= end_seq) + goto out; - /* move first record forward until length fits into the buffer */ - seq = dumper->cur_seq; - idx = dumper->cur_idx; - while (l >= size && seq < dumper->next_seq) { - struct printk_log *msg = log_from_idx(idx); + /* count the total bytes after seq */ + textlen = count_remaining(&iter, end_seq, msgbuf, + PRINTK_RECORD_MAX, 0, time); + + /* move iter forward until length fits into the buffer */ + while (textlen > size) { + ret = prb_iter_next(&iter, msgbuf, PRINTK_RECORD_MAX, &seq); + if (ret == 0) { + break; + } else if (ret < 0 || seq >= end_seq) { + prb_iter_init(&iter, &printk_rb, &seq); + goto retry; + } - l -= msg_print_text(msg, true, time, NULL, 0); - idx = log_next(idx); - seq++; + msg = (struct printk_log *)msgbuf; + textlen -= msg_print_text(msg, true, time, NULL, 0); } - /* last message in next interation */ - next_seq = seq; - next_idx = idx; + /* save end seq for the next interation */ + new_end_seq = seq + 1; + + /* copy messages to buffer */ + while (l < size) { + ret = prb_iter_next(&iter, msgbuf, PRINTK_RECORD_MAX, &seq); + if (ret == 0) { + break; + } else if (ret < 0) { + /* + * iterator (and thus also the start position) + * invalid, start over from beginning of list + */ + prb_iter_init(&iter, &printk_rb, NULL); + continue; + } - l = 0; - while (seq < dumper->next_seq) { - struct printk_log *msg = log_from_idx(idx); + if (seq >= end_seq) + break; - l += msg_print_text(msg, syslog, time, buf + l, size - l); - idx = log_next(idx); - seq++; + msg = (struct printk_log *)msgbuf; + textlen = msg_print_text(msg, syslog, time, buf + l, size - l); + if (textlen > 0) + l += textlen; + cont = true; } - dumper->next_seq = next_seq; - dumper->next_idx = next_idx; - ret = true; - logbuf_unlock_irqrestore(flags); -out: - if (len) + if (cont && len) *len = l; - return ret; +out: + prb_commit(&h); + if (new_end_seq) + dumper->buffer_end_seq = new_end_seq; + return cont; } EXPORT_SYMBOL_GPL(kmsg_dump_get_buffer); @@ -3362,10 +3137,8 @@ EXPORT_SYMBOL_GPL(kmsg_dump_get_buffer); */ void kmsg_dump_rewind_nolock(struct kmsg_dumper *dumper) { - dumper->cur_seq = clear_seq; - dumper->cur_idx = clear_idx; - dumper->next_seq = log_next_seq; - dumper->next_idx = log_next_idx; + dumper->line_seq = 0; + dumper->buffer_end_seq = 0; } /** @@ -3378,12 +3151,89 @@ void kmsg_dump_rewind_nolock(struct kmsg */ void kmsg_dump_rewind(struct kmsg_dumper *dumper) { - unsigned long flags; - - logbuf_lock_irqsave(flags); kmsg_dump_rewind_nolock(dumper); - logbuf_unlock_irqrestore(flags); } EXPORT_SYMBOL_GPL(kmsg_dump_rewind); +static bool console_can_emergency(int level) +{ + struct console *con; + + for_each_console(con) { + if (!(con->flags & CON_ENABLED)) + continue; + if (con->write_atomic && oops_in_progress) + return true; + if (con->write && (con->flags & CON_BOOT)) + return true; + } + return false; +} + +static void call_emergency_console_drivers(int level, const char *text, + size_t text_len) +{ + struct console *con; + + for_each_console(con) { + if (!(con->flags & CON_ENABLED)) + continue; + if (con->write_atomic && oops_in_progress) { + con->write_atomic(con, text, text_len); + continue; + } + if (con->write && (con->flags & CON_BOOT)) { + con->write(con, text, text_len); + continue; + } + } +} + +static void printk_emergency(char *buffer, int level, u64 ts_nsec, u16 cpu, + char *text, u16 text_len) +{ + struct printk_log msg; + size_t prefix_len; + + if (!console_can_emergency(level)) + return; + + msg.level = level; + msg.ts_nsec = ts_nsec; + msg.cpu = cpu; + msg.facility = 0; + + /* "text" must have PREFIX_MAX preceding bytes available */ + + prefix_len = print_prefix(&msg, + console_msg_format & MSG_FORMAT_SYSLOG, + printk_time, buffer); + /* move the prefix forward to the beginning of the message text */ + text -= prefix_len; + memmove(text, buffer, prefix_len); + text_len += prefix_len; + + text[text_len++] = '\n'; + + call_emergency_console_drivers(level, text, text_len); + + touch_softlockup_watchdog_sync(); + clocksource_touch_watchdog(); + rcu_cpu_stall_reset(); + touch_nmi_watchdog(); + + printk_delay(level); +} #endif + +void console_atomic_lock(unsigned int *flags) +{ + prb_lock(&printk_cpulock, flags); +} +EXPORT_SYMBOL(console_atomic_lock); + +void console_atomic_unlock(unsigned int flags) +{ + prb_unlock(&printk_cpulock, flags); +} +EXPORT_SYMBOL(console_atomic_unlock); Index: linux-5.6.17-rt10/kernel/printk/printk_safe.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/printk/printk_safe.c +++ /dev/null @@ -1,406 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0-or-later -/* - * printk_safe.c - Safe printk for printk-deadlock-prone contexts - */ - -#include -#include -#include -#include -#include -#include -#include - -#include "internal.h" - -/* - * printk() could not take logbuf_lock in NMI context. Instead, - * it uses an alternative implementation that temporary stores - * the strings into a per-CPU buffer. The content of the buffer - * is later flushed into the main ring buffer via IRQ work. - * - * The alternative implementation is chosen transparently - * by examinig current printk() context mask stored in @printk_context - * per-CPU variable. - * - * The implementation allows to flush the strings also from another CPU. - * There are situations when we want to make sure that all buffers - * were handled or when IRQs are blocked. - */ - -#define SAFE_LOG_BUF_LEN ((1 << CONFIG_PRINTK_SAFE_LOG_BUF_SHIFT) - \ - sizeof(atomic_t) - \ - sizeof(atomic_t) - \ - sizeof(struct irq_work)) - -struct printk_safe_seq_buf { - atomic_t len; /* length of written data */ - atomic_t message_lost; - struct irq_work work; /* IRQ work that flushes the buffer */ - unsigned char buffer[SAFE_LOG_BUF_LEN]; -}; - -static DEFINE_PER_CPU(struct printk_safe_seq_buf, safe_print_seq); -static DEFINE_PER_CPU(int, printk_context); - -#ifdef CONFIG_PRINTK_NMI -static DEFINE_PER_CPU(struct printk_safe_seq_buf, nmi_print_seq); -#endif - -/* Get flushed in a more safe context. */ -static void queue_flush_work(struct printk_safe_seq_buf *s) -{ - if (printk_percpu_data_ready()) - irq_work_queue(&s->work); -} - -/* - * Add a message to per-CPU context-dependent buffer. NMI and printk-safe - * have dedicated buffers, because otherwise printk-safe preempted by - * NMI-printk would have overwritten the NMI messages. - * - * The messages are flushed from irq work (or from panic()), possibly, - * from other CPU, concurrently with printk_safe_log_store(). Should this - * happen, printk_safe_log_store() will notice the buffer->len mismatch - * and repeat the write. - */ -static __printf(2, 0) int printk_safe_log_store(struct printk_safe_seq_buf *s, - const char *fmt, va_list args) -{ - int add; - size_t len; - va_list ap; - -again: - len = atomic_read(&s->len); - - /* The trailing '\0' is not counted into len. */ - if (len >= sizeof(s->buffer) - 1) { - atomic_inc(&s->message_lost); - queue_flush_work(s); - return 0; - } - - /* - * Make sure that all old data have been read before the buffer - * was reset. This is not needed when we just append data. - */ - if (!len) - smp_rmb(); - - va_copy(ap, args); - add = vscnprintf(s->buffer + len, sizeof(s->buffer) - len, fmt, ap); - va_end(ap); - if (!add) - return 0; - - /* - * Do it once again if the buffer has been flushed in the meantime. - * Note that atomic_cmpxchg() is an implicit memory barrier that - * makes sure that the data were written before updating s->len. - */ - if (atomic_cmpxchg(&s->len, len, len + add) != len) - goto again; - - queue_flush_work(s); - return add; -} - -static inline void printk_safe_flush_line(const char *text, int len) -{ - /* - * Avoid any console drivers calls from here, because we may be - * in NMI or printk_safe context (when in panic). The messages - * must go only into the ring buffer at this stage. Consoles will - * get explicitly called later when a crashdump is not generated. - */ - printk_deferred("%.*s", len, text); -} - -/* printk part of the temporary buffer line by line */ -static int printk_safe_flush_buffer(const char *start, size_t len) -{ - const char *c, *end; - bool header; - - c = start; - end = start + len; - header = true; - - /* Print line by line. */ - while (c < end) { - if (*c == '\n') { - printk_safe_flush_line(start, c - start + 1); - start = ++c; - header = true; - continue; - } - - /* Handle continuous lines or missing new line. */ - if ((c + 1 < end) && printk_get_level(c)) { - if (header) { - c = printk_skip_level(c); - continue; - } - - printk_safe_flush_line(start, c - start); - start = c++; - header = true; - continue; - } - - header = false; - c++; - } - - /* Check if there was a partial line. Ignore pure header. */ - if (start < end && !header) { - static const char newline[] = KERN_CONT "\n"; - - printk_safe_flush_line(start, end - start); - printk_safe_flush_line(newline, strlen(newline)); - } - - return len; -} - -static void report_message_lost(struct printk_safe_seq_buf *s) -{ - int lost = atomic_xchg(&s->message_lost, 0); - - if (lost) - printk_deferred("Lost %d message(s)!\n", lost); -} - -/* - * Flush data from the associated per-CPU buffer. The function - * can be called either via IRQ work or independently. - */ -static void __printk_safe_flush(struct irq_work *work) -{ - static raw_spinlock_t read_lock = - __RAW_SPIN_LOCK_INITIALIZER(read_lock); - struct printk_safe_seq_buf *s = - container_of(work, struct printk_safe_seq_buf, work); - unsigned long flags; - size_t len; - int i; - - /* - * The lock has two functions. First, one reader has to flush all - * available message to make the lockless synchronization with - * writers easier. Second, we do not want to mix messages from - * different CPUs. This is especially important when printing - * a backtrace. - */ - raw_spin_lock_irqsave(&read_lock, flags); - - i = 0; -more: - len = atomic_read(&s->len); - - /* - * This is just a paranoid check that nobody has manipulated - * the buffer an unexpected way. If we printed something then - * @len must only increase. Also it should never overflow the - * buffer size. - */ - if ((i && i >= len) || len > sizeof(s->buffer)) { - const char *msg = "printk_safe_flush: internal error\n"; - - printk_safe_flush_line(msg, strlen(msg)); - len = 0; - } - - if (!len) - goto out; /* Someone else has already flushed the buffer. */ - - /* Make sure that data has been written up to the @len */ - smp_rmb(); - i += printk_safe_flush_buffer(s->buffer + i, len - i); - - /* - * Check that nothing has got added in the meantime and truncate - * the buffer. Note that atomic_cmpxchg() is an implicit memory - * barrier that makes sure that the data were copied before - * updating s->len. - */ - if (atomic_cmpxchg(&s->len, len, 0) != len) - goto more; - -out: - report_message_lost(s); - raw_spin_unlock_irqrestore(&read_lock, flags); -} - -/** - * printk_safe_flush - flush all per-cpu nmi buffers. - * - * The buffers are flushed automatically via IRQ work. This function - * is useful only when someone wants to be sure that all buffers have - * been flushed at some point. - */ -void printk_safe_flush(void) -{ - int cpu; - - for_each_possible_cpu(cpu) { -#ifdef CONFIG_PRINTK_NMI - __printk_safe_flush(&per_cpu(nmi_print_seq, cpu).work); -#endif - __printk_safe_flush(&per_cpu(safe_print_seq, cpu).work); - } -} - -/** - * printk_safe_flush_on_panic - flush all per-cpu nmi buffers when the system - * goes down. - * - * Similar to printk_safe_flush() but it can be called even in NMI context when - * the system goes down. It does the best effort to get NMI messages into - * the main ring buffer. - * - * Note that it could try harder when there is only one CPU online. - */ -void printk_safe_flush_on_panic(void) -{ - /* - * Make sure that we could access the main ring buffer. - * Do not risk a double release when more CPUs are up. - */ - if (raw_spin_is_locked(&logbuf_lock)) { - if (num_online_cpus() > 1) - return; - - debug_locks_off(); - raw_spin_lock_init(&logbuf_lock); - } - - printk_safe_flush(); -} - -#ifdef CONFIG_PRINTK_NMI -/* - * Safe printk() for NMI context. It uses a per-CPU buffer to - * store the message. NMIs are not nested, so there is always only - * one writer running. But the buffer might get flushed from another - * CPU, so we need to be careful. - */ -static __printf(1, 0) int vprintk_nmi(const char *fmt, va_list args) -{ - struct printk_safe_seq_buf *s = this_cpu_ptr(&nmi_print_seq); - - return printk_safe_log_store(s, fmt, args); -} - -void notrace printk_nmi_enter(void) -{ - this_cpu_or(printk_context, PRINTK_NMI_CONTEXT_MASK); -} - -void notrace printk_nmi_exit(void) -{ - this_cpu_and(printk_context, ~PRINTK_NMI_CONTEXT_MASK); -} - -/* - * Marks a code that might produce many messages in NMI context - * and the risk of losing them is more critical than eventual - * reordering. - * - * It has effect only when called in NMI context. Then printk() - * will try to store the messages into the main logbuf directly - * and use the per-CPU buffers only as a fallback when the lock - * is not available. - */ -void printk_nmi_direct_enter(void) -{ - if (this_cpu_read(printk_context) & PRINTK_NMI_CONTEXT_MASK) - this_cpu_or(printk_context, PRINTK_NMI_DIRECT_CONTEXT_MASK); -} - -void printk_nmi_direct_exit(void) -{ - this_cpu_and(printk_context, ~PRINTK_NMI_DIRECT_CONTEXT_MASK); -} - -#else - -static __printf(1, 0) int vprintk_nmi(const char *fmt, va_list args) -{ - return 0; -} - -#endif /* CONFIG_PRINTK_NMI */ - -/* - * Lock-less printk(), to avoid deadlocks should the printk() recurse - * into itself. It uses a per-CPU buffer to store the message, just like - * NMI. - */ -static __printf(1, 0) int vprintk_safe(const char *fmt, va_list args) -{ - struct printk_safe_seq_buf *s = this_cpu_ptr(&safe_print_seq); - - return printk_safe_log_store(s, fmt, args); -} - -/* Can be preempted by NMI. */ -void __printk_safe_enter(void) -{ - this_cpu_inc(printk_context); -} - -/* Can be preempted by NMI. */ -void __printk_safe_exit(void) -{ - this_cpu_dec(printk_context); -} - -__printf(1, 0) int vprintk_func(const char *fmt, va_list args) -{ - /* - * Try to use the main logbuf even in NMI. But avoid calling console - * drivers that might have their own locks. - */ - if ((this_cpu_read(printk_context) & PRINTK_NMI_DIRECT_CONTEXT_MASK) && - raw_spin_trylock(&logbuf_lock)) { - int len; - - len = vprintk_store(0, LOGLEVEL_DEFAULT, NULL, 0, fmt, args); - raw_spin_unlock(&logbuf_lock); - defer_console_output(); - return len; - } - - /* Use extra buffer in NMI when logbuf_lock is taken or in safe mode. */ - if (this_cpu_read(printk_context) & PRINTK_NMI_CONTEXT_MASK) - return vprintk_nmi(fmt, args); - - /* Use extra buffer to prevent a recursion deadlock in safe mode. */ - if (this_cpu_read(printk_context) & PRINTK_SAFE_CONTEXT_MASK) - return vprintk_safe(fmt, args); - - /* No obstacles. */ - return vprintk_default(fmt, args); -} - -void __init printk_safe_init(void) -{ - int cpu; - - for_each_possible_cpu(cpu) { - struct printk_safe_seq_buf *s; - - s = &per_cpu(safe_print_seq, cpu); - init_irq_work(&s->work, __printk_safe_flush); - -#ifdef CONFIG_PRINTK_NMI - s = &per_cpu(nmi_print_seq, cpu); - init_irq_work(&s->work, __printk_safe_flush); -#endif - } - - /* Flush pending messages that did not have scheduled IRQ works. */ - printk_safe_flush(); -} Index: linux-5.6.17-rt10/kernel/ptrace.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/ptrace.c +++ linux-5.6.17-rt10/kernel/ptrace.c @@ -180,7 +180,14 @@ static bool ptrace_freeze_traced(struct spin_lock_irq(&task->sighand->siglock); if (task_is_traced(task) && !__fatal_signal_pending(task)) { - task->state = __TASK_TRACED; + unsigned long flags; + + raw_spin_lock_irqsave(&task->pi_lock, flags); + if (task->state & __TASK_TRACED) + task->state = __TASK_TRACED; + else + task->saved_state = __TASK_TRACED; + raw_spin_unlock_irqrestore(&task->pi_lock, flags); ret = true; } spin_unlock_irq(&task->sighand->siglock); Index: linux-5.6.17-rt10/kernel/rcu/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/kernel/rcu/Kconfig +++ linux-5.6.17-rt10/kernel/rcu/Kconfig @@ -162,8 +162,8 @@ config RCU_FAST_NO_HZ config RCU_BOOST bool "Enable RCU priority boosting" - depends on RT_MUTEXES && PREEMPT_RCU && RCU_EXPERT - default n + depends on (RT_MUTEXES && PREEMPT_RCU && RCU_EXPERT) || PREEMPT_RT + default y if PREEMPT_RT help This option boosts the priority of preempted RCU readers that block the current preemptible RCU grace period for too long. Index: linux-5.6.17-rt10/kernel/rcu/rcutorture.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/rcu/rcutorture.c +++ linux-5.6.17-rt10/kernel/rcu/rcutorture.c @@ -61,10 +61,13 @@ MODULE_AUTHOR("Paul E. McKenney > RCUTORTURE_RDR_SHIFT) > 1); rtrsp->rt_readstate = newstate; - /* First, put new protection in place to avoid critical-section gap. */ + /* + * First, put new protection in place to avoid critical-section gap. + * Disable preemption around the ATOM disables to ensure that + * in_atomic() is true. + */ if (statesnew & RCUTORTURE_RDR_BH) local_bh_disable(); + if (statesnew & RCUTORTURE_RDR_RBH) + rcu_read_lock_bh(); if (statesnew & RCUTORTURE_RDR_IRQ) local_irq_disable(); if (statesnew & RCUTORTURE_RDR_PREEMPT) preempt_disable(); - if (statesnew & RCUTORTURE_RDR_RBH) - rcu_read_lock_bh(); if (statesnew & RCUTORTURE_RDR_SCHED) rcu_read_lock_sched(); + preempt_disable(); + if (statesnew & RCUTORTURE_RDR_ATOM_BH) + local_bh_disable(); + if (statesnew & RCUTORTURE_RDR_ATOM_RBH) + rcu_read_lock_bh(); + preempt_enable(); if (statesnew & RCUTORTURE_RDR_RCU) idxnew = cur_ops->readlock() << RCUTORTURE_RDR_SHIFT; - /* Next, remove old protection, irq first due to bh conflict. */ + /* + * Next, remove old protection, in decreasing order of strength + * to avoid unlock paths that aren't safe in the stronger + * context. Disable preemption around the ATOM enables in + * case the context was only atomic due to IRQ disabling. + */ + preempt_disable(); if (statesold & RCUTORTURE_RDR_IRQ) local_irq_enable(); - if (statesold & RCUTORTURE_RDR_BH) + if (statesold & RCUTORTURE_RDR_ATOM_BH) local_bh_enable(); + if (statesold & RCUTORTURE_RDR_ATOM_RBH) + rcu_read_unlock_bh(); + preempt_enable(); if (statesold & RCUTORTURE_RDR_PREEMPT) preempt_enable(); - if (statesold & RCUTORTURE_RDR_RBH) - rcu_read_unlock_bh(); if (statesold & RCUTORTURE_RDR_SCHED) rcu_read_unlock_sched(); + if (statesold & RCUTORTURE_RDR_BH) + local_bh_enable(); + if (statesold & RCUTORTURE_RDR_RBH) + rcu_read_unlock_bh(); if (statesold & RCUTORTURE_RDR_RCU) cur_ops->readunlock(idxold >> RCUTORTURE_RDR_SHIFT); @@ -1213,6 +1237,12 @@ rcutorture_extend_mask(int oldmask, stru int mask = rcutorture_extend_mask_max(); unsigned long randmask1 = torture_random(trsp) >> 8; unsigned long randmask2 = randmask1 >> 3; + unsigned long preempts = RCUTORTURE_RDR_PREEMPT | RCUTORTURE_RDR_SCHED; + unsigned long preempts_irq = preempts | RCUTORTURE_RDR_IRQ; + unsigned long nonatomic_bhs = RCUTORTURE_RDR_BH | RCUTORTURE_RDR_RBH; + unsigned long atomic_bhs = RCUTORTURE_RDR_ATOM_BH | + RCUTORTURE_RDR_ATOM_RBH; + unsigned long tmp; WARN_ON_ONCE(mask >> RCUTORTURE_RDR_SHIFT); /* Mostly only one bit (need preemption!), sometimes lots of bits. */ @@ -1220,11 +1250,49 @@ rcutorture_extend_mask(int oldmask, stru mask = mask & randmask2; else mask = mask & (1 << (randmask2 % RCUTORTURE_RDR_NBITS)); - /* Can't enable bh w/irq disabled. */ - if ((mask & RCUTORTURE_RDR_IRQ) && - ((!(mask & RCUTORTURE_RDR_BH) && (oldmask & RCUTORTURE_RDR_BH)) || - (!(mask & RCUTORTURE_RDR_RBH) && (oldmask & RCUTORTURE_RDR_RBH)))) - mask |= RCUTORTURE_RDR_BH | RCUTORTURE_RDR_RBH; + + /* + * Can't enable bh w/irq disabled. + */ + tmp = atomic_bhs | nonatomic_bhs; + if (mask & RCUTORTURE_RDR_IRQ) + mask |= oldmask & tmp; + + /* + * Ideally these sequences would be detected in debug builds + * (regardless of RT), but until then don't stop testing + * them on non-RT. + */ + if (IS_ENABLED(CONFIG_PREEMPT_RT)) { + /* + * Can't release the outermost rcu lock in an irq disabled + * section without preemption also being disabled, if irqs + * had ever been enabled during this RCU critical section + * (could leak a special flag and delay reporting the qs). + */ + if ((oldmask & RCUTORTURE_RDR_RCU) && + (mask & RCUTORTURE_RDR_IRQ) && + !(mask & preempts)) + mask |= RCUTORTURE_RDR_RCU; + + /* Can't modify atomic bh in non-atomic context */ + if ((oldmask & atomic_bhs) && (mask & atomic_bhs) && + !(mask & preempts_irq)) { + mask |= oldmask & preempts_irq; + if (mask & RCUTORTURE_RDR_IRQ) + mask |= oldmask & tmp; + } + if ((mask & atomic_bhs) && !(mask & preempts_irq)) + mask |= RCUTORTURE_RDR_PREEMPT; + + /* Can't modify non-atomic bh in atomic context */ + tmp = nonatomic_bhs; + if (oldmask & preempts_irq) + mask &= ~tmp; + if ((oldmask | mask) & preempts_irq) + mask |= oldmask & tmp; + } + return mask ?: RCUTORTURE_RDR_RCU; } Index: linux-5.6.17-rt10/kernel/rcu/srcutree.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/rcu/srcutree.c +++ linux-5.6.17-rt10/kernel/rcu/srcutree.c @@ -25,6 +25,7 @@ #include #include #include +#include #include "rcu.h" #include "rcu_segcblist.h" @@ -735,6 +736,7 @@ static void srcu_flip(struct srcu_struct smp_mb(); /* D */ /* Pairs with C. */ } +static DEFINE_LOCAL_IRQ_LOCK(sp_llock); /* * If SRCU is likely idle, return true, otherwise return false. * @@ -765,13 +767,13 @@ static bool srcu_might_be_idle(struct sr unsigned long tlast; /* If the local srcu_data structure has callbacks, not idle. */ - local_irq_save(flags); + local_lock_irqsave(sp_llock, flags); sdp = this_cpu_ptr(ssp->sda); if (rcu_segcblist_pend_cbs(&sdp->srcu_cblist)) { - local_irq_restore(flags); + local_unlock_irqrestore(sp_llock, flags); return false; /* Callbacks already present, so not idle. */ } - local_irq_restore(flags); + local_unlock_irqrestore(sp_llock, flags); /* * No local callbacks, so probabalistically probe global state. @@ -851,7 +853,7 @@ static void __call_srcu(struct srcu_stru } rhp->func = func; idx = srcu_read_lock(ssp); - local_irq_save(flags); + local_lock_irqsave(sp_llock, flags); sdp = this_cpu_ptr(ssp->sda); spin_lock_rcu_node(sdp); rcu_segcblist_enqueue(&sdp->srcu_cblist, rhp); @@ -867,7 +869,8 @@ static void __call_srcu(struct srcu_stru sdp->srcu_gp_seq_needed_exp = s; needexp = true; } - spin_unlock_irqrestore_rcu_node(sdp, flags); + spin_unlock_rcu_node(sdp); + local_unlock_irqrestore(sp_llock, flags); if (needgp) srcu_funnel_gp_start(ssp, sdp, s, do_norm); else if (needexp) Index: linux-5.6.17-rt10/kernel/rcu/tree.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/rcu/tree.c +++ linux-5.6.17-rt10/kernel/rcu/tree.c @@ -100,8 +100,10 @@ static struct rcu_state rcu_state = { static bool dump_tree; module_param(dump_tree, bool, 0444); /* By default, use RCU_SOFTIRQ instead of rcuc kthreads. */ -static bool use_softirq = 1; +static bool use_softirq = !IS_ENABLED(CONFIG_PREEMPT_RT); +#ifndef CONFIG_PREEMPT_RT module_param(use_softirq, bool, 0444); +#endif /* Control rcu_node-tree auto-balancing at boot time. */ static bool rcu_fanout_exact; module_param(rcu_fanout_exact, bool, 0444); @@ -1113,6 +1115,7 @@ static int rcu_implicit_dynticks_qs(stru !rdp->rcu_iw_pending && rdp->rcu_iw_gp_seq != rnp->gp_seq && (rnp->ffmask & rdp->grpmask)) { init_irq_work(&rdp->rcu_iw, rcu_iw_handler); + atomic_or(IRQ_WORK_HARD_IRQ, &rdp->rcu_iw.flags); rdp->rcu_iw_pending = true; rdp->rcu_iw_gp_seq = rnp->gp_seq; irq_work_queue_on(&rdp->rcu_iw, rdp->cpu); @@ -2719,7 +2722,7 @@ struct kfree_rcu_cpu_work { struct kfree_rcu_cpu { struct rcu_head *head; struct kfree_rcu_cpu_work krw_arr[KFREE_N_BATCHES]; - spinlock_t lock; + raw_spinlock_t lock; struct delayed_work monitor_work; bool monitor_todo; bool initialized; @@ -2741,10 +2744,10 @@ static void kfree_rcu_work(struct work_s krwp = container_of(to_rcu_work(work), struct kfree_rcu_cpu_work, rcu_work); krcp = krwp->krcp; - spin_lock_irqsave(&krcp->lock, flags); + raw_spin_lock_irqsave(&krcp->lock, flags); head = krwp->head_free; krwp->head_free = NULL; - spin_unlock_irqrestore(&krcp->lock, flags); + raw_spin_unlock_irqrestore(&krcp->lock, flags); // List "head" is now private, so traverse locklessly. for (; head; head = next) { @@ -2803,14 +2806,14 @@ static inline void kfree_rcu_drain_unloc krcp->monitor_todo = false; if (queue_kfree_rcu_work(krcp)) { // Success! Our job is done here. - spin_unlock_irqrestore(&krcp->lock, flags); + raw_spin_unlock_irqrestore(&krcp->lock, flags); return; } // Previous RCU batch still in progress, try again later. krcp->monitor_todo = true; schedule_delayed_work(&krcp->monitor_work, KFREE_DRAIN_JIFFIES); - spin_unlock_irqrestore(&krcp->lock, flags); + raw_spin_unlock_irqrestore(&krcp->lock, flags); } /* @@ -2823,11 +2826,11 @@ static void kfree_rcu_monitor(struct wor struct kfree_rcu_cpu *krcp = container_of(work, struct kfree_rcu_cpu, monitor_work.work); - spin_lock_irqsave(&krcp->lock, flags); + raw_spin_lock_irqsave(&krcp->lock, flags); if (krcp->monitor_todo) kfree_rcu_drain_unlock(krcp, flags); else - spin_unlock_irqrestore(&krcp->lock, flags); + raw_spin_unlock_irqrestore(&krcp->lock, flags); } /* @@ -2852,7 +2855,7 @@ void kfree_call_rcu(struct rcu_head *hea local_irq_save(flags); // For safely calling this_cpu_ptr(). krcp = this_cpu_ptr(&krc); if (krcp->initialized) - spin_lock(&krcp->lock); + raw_spin_lock(&krcp->lock); // Queue the object but don't yet schedule the batch. if (debug_rcu_head_queue(head)) { @@ -2874,7 +2877,7 @@ void kfree_call_rcu(struct rcu_head *hea unlock_return: if (krcp->initialized) - spin_unlock(&krcp->lock); + raw_spin_unlock(&krcp->lock); local_irq_restore(flags); } EXPORT_SYMBOL_GPL(kfree_call_rcu); @@ -2887,15 +2890,15 @@ void __init kfree_rcu_scheduler_running( for_each_online_cpu(cpu) { struct kfree_rcu_cpu *krcp = per_cpu_ptr(&krc, cpu); - spin_lock_irqsave(&krcp->lock, flags); + raw_spin_lock_irqsave(&krcp->lock, flags); if (!krcp->head || krcp->monitor_todo) { - spin_unlock_irqrestore(&krcp->lock, flags); + raw_spin_unlock_irqrestore(&krcp->lock, flags); continue; } krcp->monitor_todo = true; schedule_delayed_work_on(cpu, &krcp->monitor_work, KFREE_DRAIN_JIFFIES); - spin_unlock_irqrestore(&krcp->lock, flags); + raw_spin_unlock_irqrestore(&krcp->lock, flags); } } @@ -3780,7 +3783,7 @@ static void __init kfree_rcu_batch_init( for_each_possible_cpu(cpu) { struct kfree_rcu_cpu *krcp = per_cpu_ptr(&krc, cpu); - spin_lock_init(&krcp->lock); + raw_spin_lock_init(&krcp->lock); for (i = 0; i < KFREE_N_BATCHES; i++) krcp->krw_arr[i].krcp = krcp; INIT_DELAYED_WORK(&krcp->monitor_work, kfree_rcu_monitor); Index: linux-5.6.17-rt10/kernel/rcu/tree_plugin.h =================================================================== --- linux-5.6.17-rt10.orig/kernel/rcu/tree_plugin.h +++ linux-5.6.17-rt10/kernel/rcu/tree_plugin.h @@ -287,10 +287,14 @@ void rcu_note_context_switch(bool preemp struct task_struct *t = current; struct rcu_data *rdp = this_cpu_ptr(&rcu_data); struct rcu_node *rnp; + int sleeping_l = 0; trace_rcu_utilization(TPS("Start context switch")); lockdep_assert_irqs_disabled(); - WARN_ON_ONCE(!preempt && rcu_preempt_depth() > 0); +#if defined(CONFIG_PREEMPT_RT) + sleeping_l = t->sleeping_lock; +#endif + WARN_ON_ONCE(!preempt && rcu_preempt_depth() > 0 && !sleeping_l); if (rcu_preempt_depth() > 0 && !t->rcu_read_unlock_special.b.blocked) { Index: linux-5.6.17-rt10/kernel/rcu/update.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/rcu/update.c +++ linux-5.6.17-rt10/kernel/rcu/update.c @@ -54,8 +54,10 @@ #ifndef CONFIG_TINY_RCU module_param(rcu_expedited, int, 0); module_param(rcu_normal, int, 0); -static int rcu_normal_after_boot; +static int rcu_normal_after_boot = IS_ENABLED(CONFIG_PREEMPT_RT); +#ifndef CONFIG_PREEMPT_RT module_param(rcu_normal_after_boot, int, 0); +#endif #endif /* #ifndef CONFIG_TINY_RCU */ #ifdef CONFIG_DEBUG_LOCK_ALLOC Index: linux-5.6.17-rt10/kernel/sched/completion.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/sched/completion.c +++ linux-5.6.17-rt10/kernel/sched/completion.c @@ -29,12 +29,12 @@ void complete(struct completion *x) { unsigned long flags; - spin_lock_irqsave(&x->wait.lock, flags); + raw_spin_lock_irqsave(&x->wait.lock, flags); if (x->done != UINT_MAX) x->done++; - __wake_up_locked(&x->wait, TASK_NORMAL, 1); - spin_unlock_irqrestore(&x->wait.lock, flags); + swake_up_locked(&x->wait); + raw_spin_unlock_irqrestore(&x->wait.lock, flags); } EXPORT_SYMBOL(complete); @@ -58,10 +58,10 @@ void complete_all(struct completion *x) { unsigned long flags; - spin_lock_irqsave(&x->wait.lock, flags); + raw_spin_lock_irqsave(&x->wait.lock, flags); x->done = UINT_MAX; - __wake_up_locked(&x->wait, TASK_NORMAL, 0); - spin_unlock_irqrestore(&x->wait.lock, flags); + swake_up_all_locked(&x->wait); + raw_spin_unlock_irqrestore(&x->wait.lock, flags); } EXPORT_SYMBOL(complete_all); @@ -70,20 +70,20 @@ do_wait_for_common(struct completion *x, long (*action)(long), long timeout, int state) { if (!x->done) { - DECLARE_WAITQUEUE(wait, current); + DECLARE_SWAITQUEUE(wait); - __add_wait_queue_entry_tail_exclusive(&x->wait, &wait); do { if (signal_pending_state(state, current)) { timeout = -ERESTARTSYS; break; } + __prepare_to_swait(&x->wait, &wait); __set_current_state(state); - spin_unlock_irq(&x->wait.lock); + raw_spin_unlock_irq(&x->wait.lock); timeout = action(timeout); - spin_lock_irq(&x->wait.lock); + raw_spin_lock_irq(&x->wait.lock); } while (!x->done && timeout); - __remove_wait_queue(&x->wait, &wait); + __finish_swait(&x->wait, &wait); if (!x->done) return timeout; } @@ -100,9 +100,9 @@ __wait_for_common(struct completion *x, complete_acquire(x); - spin_lock_irq(&x->wait.lock); + raw_spin_lock_irq(&x->wait.lock); timeout = do_wait_for_common(x, action, timeout, state); - spin_unlock_irq(&x->wait.lock); + raw_spin_unlock_irq(&x->wait.lock); complete_release(x); @@ -291,12 +291,12 @@ bool try_wait_for_completion(struct comp if (!READ_ONCE(x->done)) return false; - spin_lock_irqsave(&x->wait.lock, flags); + raw_spin_lock_irqsave(&x->wait.lock, flags); if (!x->done) ret = false; else if (x->done != UINT_MAX) x->done--; - spin_unlock_irqrestore(&x->wait.lock, flags); + raw_spin_unlock_irqrestore(&x->wait.lock, flags); return ret; } EXPORT_SYMBOL(try_wait_for_completion); @@ -322,8 +322,8 @@ bool completion_done(struct completion * * otherwise we can end up freeing the completion before complete() * is done referencing it. */ - spin_lock_irqsave(&x->wait.lock, flags); - spin_unlock_irqrestore(&x->wait.lock, flags); + raw_spin_lock_irqsave(&x->wait.lock, flags); + raw_spin_unlock_irqrestore(&x->wait.lock, flags); return true; } EXPORT_SYMBOL(completion_done); Index: linux-5.6.17-rt10/kernel/sched/core.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/sched/core.c +++ linux-5.6.17-rt10/kernel/sched/core.c @@ -57,7 +57,11 @@ const_debug unsigned int sysctl_sched_fe * Number of tasks to iterate in a single balance run. * Limited because this is done with IRQs disabled. */ +#ifdef CONFIG_PREEMPT_RT +const_debug unsigned int sysctl_sched_nr_migrate = 8; +#else const_debug unsigned int sysctl_sched_nr_migrate = 32; +#endif /* * period over which we measure -rt task CPU usage in us. @@ -411,9 +415,15 @@ static bool set_nr_if_polling(struct tas #endif #endif -static bool __wake_q_add(struct wake_q_head *head, struct task_struct *task) +static bool __wake_q_add(struct wake_q_head *head, struct task_struct *task, + bool sleeper) { - struct wake_q_node *node = &task->wake_q; + struct wake_q_node *node; + + if (sleeper) + node = &task->wake_q_sleeper; + else + node = &task->wake_q; /* * Atomically grab the task, if ->wake_q is !nil already it means @@ -449,7 +459,13 @@ static bool __wake_q_add(struct wake_q_h */ void wake_q_add(struct wake_q_head *head, struct task_struct *task) { - if (__wake_q_add(head, task)) + if (__wake_q_add(head, task, false)) + get_task_struct(task); +} + +void wake_q_add_sleeper(struct wake_q_head *head, struct task_struct *task) +{ + if (__wake_q_add(head, task, true)) get_task_struct(task); } @@ -472,28 +488,39 @@ void wake_q_add(struct wake_q_head *head */ void wake_q_add_safe(struct wake_q_head *head, struct task_struct *task) { - if (!__wake_q_add(head, task)) + if (!__wake_q_add(head, task, false)) put_task_struct(task); } -void wake_up_q(struct wake_q_head *head) +void __wake_up_q(struct wake_q_head *head, bool sleeper) { struct wake_q_node *node = head->first; while (node != WAKE_Q_TAIL) { struct task_struct *task; - task = container_of(node, struct task_struct, wake_q); + if (sleeper) + task = container_of(node, struct task_struct, wake_q_sleeper); + else + task = container_of(node, struct task_struct, wake_q); + BUG_ON(!task); /* Task can safely be re-inserted now: */ node = node->next; - task->wake_q.next = NULL; + if (sleeper) + task->wake_q_sleeper.next = NULL; + else + task->wake_q.next = NULL; /* * wake_up_process() executes a full barrier, which pairs with * the queueing in wake_q_add() so as not to miss wakeups. */ - wake_up_process(task); + if (sleeper) + wake_up_lock_sleeper(task); + else + wake_up_process(task); + put_task_struct(task); } } @@ -529,6 +556,48 @@ void resched_curr(struct rq *rq) trace_sched_wake_idle_without_ipi(cpu); } +#ifdef CONFIG_PREEMPT_LAZY + +static int tsk_is_polling(struct task_struct *p) +{ +#ifdef TIF_POLLING_NRFLAG + return test_tsk_thread_flag(p, TIF_POLLING_NRFLAG); +#else + return 0; +#endif +} + +void resched_curr_lazy(struct rq *rq) +{ + struct task_struct *curr = rq->curr; + int cpu; + + if (!sched_feat(PREEMPT_LAZY)) { + resched_curr(rq); + return; + } + + lockdep_assert_held(&rq->lock); + + if (test_tsk_need_resched(curr)) + return; + + if (test_tsk_need_resched_lazy(curr)) + return; + + set_tsk_need_resched_lazy(curr); + + cpu = cpu_of(rq); + if (cpu == smp_processor_id()) + return; + + /* NEED_RESCHED_LAZY must be visible before we test polling */ + smp_mb(); + if (!tsk_is_polling(curr)) + smp_send_reschedule(cpu); +} +#endif + void resched_cpu(int cpu) { struct rq *rq = cpu_rq(cpu); @@ -1451,7 +1520,7 @@ static inline bool is_cpu_allowed(struct if (!cpumask_test_cpu(cpu, p->cpus_ptr)) return false; - if (is_per_cpu_kthread(p)) + if (is_per_cpu_kthread(p) || __migrate_disabled(p)) return cpu_online(cpu); return cpu_active(cpu); @@ -1500,6 +1569,7 @@ static struct rq *move_queued_task(struc struct migration_arg { struct task_struct *task; int dest_cpu; + bool done; }; /* @@ -1535,6 +1605,11 @@ static int migration_cpu_stop(void *data struct task_struct *p = arg->task; struct rq *rq = this_rq(); struct rq_flags rf; + int dest_cpu = arg->dest_cpu; + + /* We don't look at arg after this point. */ + smp_mb(); + arg->done = true; /* * The original target CPU might have gone down and we might @@ -1557,9 +1632,9 @@ static int migration_cpu_stop(void *data */ if (task_rq(p) == rq) { if (task_on_rq_queued(p)) - rq = __migrate_task(rq, &rf, p, arg->dest_cpu); + rq = __migrate_task(rq, &rf, p, dest_cpu); else - p->wake_cpu = arg->dest_cpu; + p->wake_cpu = dest_cpu; } rq_unlock(rq, &rf); raw_spin_unlock(&p->pi_lock); @@ -1575,9 +1650,18 @@ static int migration_cpu_stop(void *data void set_cpus_allowed_common(struct task_struct *p, const struct cpumask *new_mask) { cpumask_copy(&p->cpus_mask, new_mask); - p->nr_cpus_allowed = cpumask_weight(new_mask); + if (p->cpus_ptr == &p->cpus_mask) + p->nr_cpus_allowed = cpumask_weight(new_mask); } +#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) +int __migrate_disabled(struct task_struct *p) +{ + return p->migrate_disable; +} +EXPORT_SYMBOL_GPL(__migrate_disabled); +#endif + void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask) { struct rq *rq = task_rq(p); @@ -1644,7 +1728,7 @@ static int __set_cpus_allowed_ptr(struct goto out; } - if (cpumask_equal(p->cpus_ptr, new_mask)) + if (cpumask_equal(&p->cpus_mask, new_mask)) goto out; dest_cpu = cpumask_any_and(cpu_valid_mask, new_mask); @@ -1666,7 +1750,8 @@ static int __set_cpus_allowed_ptr(struct } /* Can the task run on the task's current CPU? If so, we're done */ - if (cpumask_test_cpu(task_cpu(p), new_mask)) + if (cpumask_test_cpu(task_cpu(p), new_mask) || + p->cpus_ptr != &p->cpus_mask) goto out; if (task_running(rq, p) || p->state == TASK_WAKING) { @@ -1863,6 +1948,18 @@ out: } #endif /* CONFIG_NUMA_BALANCING */ +static bool check_task_state(struct task_struct *p, long match_state) +{ + bool match = false; + + raw_spin_lock_irq(&p->pi_lock); + if (p->state == match_state || p->saved_state == match_state) + match = true; + raw_spin_unlock_irq(&p->pi_lock); + + return match; +} + /* * wait_task_inactive - wait for a thread to unschedule. * @@ -1907,7 +2004,7 @@ unsigned long wait_task_inactive(struct * is actually now running somewhere else! */ while (task_running(rq, p)) { - if (match_state && unlikely(p->state != match_state)) + if (match_state && !check_task_state(p, match_state)) return 0; cpu_relax(); } @@ -1922,7 +2019,8 @@ unsigned long wait_task_inactive(struct running = task_running(rq, p); queued = task_on_rq_queued(p); ncsw = 0; - if (!match_state || p->state == match_state) + if (!match_state || p->state == match_state || + p->saved_state == match_state) ncsw = p->nvcsw | LONG_MIN; /* sets MSB */ task_rq_unlock(rq, p, &rf); @@ -2510,6 +2608,8 @@ try_to_wake_up(struct task_struct *p, un int cpu, success = 0; preempt_disable(); + +#ifndef CONFIG_PREEMPT_RT if (p == current) { /* * We're waking current, this means 'p->on_rq' and 'task_cpu(p) @@ -2532,7 +2632,7 @@ try_to_wake_up(struct task_struct *p, un trace_sched_wakeup(p); goto out; } - +#endif /* * If we are going to wake up a thread waiting for CONDITION we * need to ensure that CONDITION=1 done by the caller can not be @@ -2541,8 +2641,27 @@ try_to_wake_up(struct task_struct *p, un */ raw_spin_lock_irqsave(&p->pi_lock, flags); smp_mb__after_spinlock(); - if (!(p->state & state)) - goto unlock; + if (!(p->state & state)) { + /* + * The task might be running due to a spinlock sleeper + * wakeup. Check the saved state and set it to running + * if the wakeup condition is true. + */ + if (!(wake_flags & WF_LOCK_SLEEPER)) { + if (p->saved_state & state) { + p->saved_state = TASK_RUNNING; + success = 1; + } + } + raw_spin_unlock_irqrestore(&p->pi_lock, flags); + goto out_nostat; + } + /* + * If this is a regular wakeup, then we can unconditionally + * clear the saved state of a "lock sleeper". + */ + if (!(wake_flags & WF_LOCK_SLEEPER)) + p->saved_state = TASK_RUNNING; trace_sched_waking(p); @@ -2634,9 +2753,12 @@ try_to_wake_up(struct task_struct *p, un ttwu_queue(p, cpu, wake_flags); unlock: raw_spin_unlock_irqrestore(&p->pi_lock, flags); +#ifndef CONFIG_PREEMPT_RT out: +#endif if (success) ttwu_stat(p, cpu, wake_flags); +out_nostat: preempt_enable(); return success; @@ -2659,6 +2781,18 @@ int wake_up_process(struct task_struct * } EXPORT_SYMBOL(wake_up_process); +/** + * wake_up_lock_sleeper - Wake up a specific process blocked on a "sleeping lock" + * @p: The process to be woken up. + * + * Same as wake_up_process() above, but wake_flags=WF_LOCK_SLEEPER to indicate + * the nature of the wakeup. + */ +int wake_up_lock_sleeper(struct task_struct *p) +{ + return try_to_wake_up(p, TASK_UNINTERRUPTIBLE, WF_LOCK_SLEEPER); +} + int wake_up_state(struct task_struct *p, unsigned int state) { return try_to_wake_up(p, state, 0); @@ -2901,6 +3035,9 @@ int sched_fork(unsigned long clone_flags p->on_cpu = 0; #endif init_task_preempt_count(p); +#ifdef CONFIG_HAVE_PREEMPT_LAZY + task_thread_info(p)->preempt_lazy_count = 0; +#endif #ifdef CONFIG_SMP plist_node_init(&p->pushable_tasks, MAX_PRIO); RB_CLEAR_NODE(&p->pushable_dl_tasks); @@ -3228,23 +3365,18 @@ static struct rq *finish_task_switch(str * provided by mmdrop(), * - a sync_core for SYNC_CORE. */ + /* + * We use mmdrop_delayed() here so we don't have to do the + * full __mmdrop() when we are the last user. + */ if (mm) { membarrier_mm_sync_core_before_usermode(mm); - mmdrop(mm); + mmdrop_delayed(mm); } if (unlikely(prev_state == TASK_DEAD)) { if (prev->sched_class->task_dead) prev->sched_class->task_dead(prev); - /* - * Remove function-return probe instances associated with this - * task and put them back on the free list. - */ - kprobe_flush_task(prev); - - /* Task is done with its stack. */ - put_task_stack(prev); - put_task_struct_rcu_user(prev); } @@ -3952,6 +4084,8 @@ restart: BUG(); } +static void migrate_disabled_sched(struct task_struct *p); + /* * __schedule() is the main scheduler function. * @@ -4022,6 +4156,9 @@ static void __sched notrace __schedule(b rq_lock(rq, &rf); smp_mb__after_spinlock(); + if (__migrate_disabled(prev)) + migrate_disabled_sched(prev); + /* Promote REQ to ACT */ rq->clock_update_flags <<= 1; update_rq_clock(rq); @@ -4043,6 +4180,7 @@ static void __sched notrace __schedule(b next = pick_next_task(rq, prev, &rf); clear_tsk_need_resched(prev); + clear_tsk_need_resched_lazy(prev); clear_preempt_need_resched(); if (likely(prev != next)) { @@ -4237,6 +4375,30 @@ static void __sched notrace preempt_sche } while (need_resched()); } +#ifdef CONFIG_PREEMPT_LAZY +/* + * If TIF_NEED_RESCHED is then we allow to be scheduled away since this is + * set by a RT task. Oterwise we try to avoid beeing scheduled out as long as + * preempt_lazy_count counter >0. + */ +static __always_inline int preemptible_lazy(void) +{ + if (test_thread_flag(TIF_NEED_RESCHED)) + return 1; + if (current_thread_info()->preempt_lazy_count) + return 0; + return 1; +} + +#else + +static inline int preemptible_lazy(void) +{ + return 1; +} + +#endif + #ifdef CONFIG_PREEMPTION /* * This is the entry point to schedule() from in-kernel preemption @@ -4250,7 +4412,8 @@ asmlinkage __visible void __sched notrac */ if (likely(!preemptible())) return; - + if (!preemptible_lazy()) + return; preempt_schedule_common(); } NOKPROBE_SYMBOL(preempt_schedule); @@ -4277,6 +4440,9 @@ asmlinkage __visible void __sched notrac if (likely(!preemptible())) return; + if (!preemptible_lazy()) + return; + do { /* * Because the function tracer can trace preempt_count_sub() @@ -6066,7 +6232,9 @@ void init_idle(struct task_struct *idle, /* Set the preempt count _outside_ the spinlocks! */ init_idle_preempt_count(idle, cpu); - +#ifdef CONFIG_HAVE_PREEMPT_LAZY + task_thread_info(idle)->preempt_lazy_count = 0; +#endif /* * The idle tasks have their own, simple scheduling class: */ @@ -6171,6 +6339,8 @@ void sched_setnuma(struct task_struct *p #endif /* CONFIG_NUMA_BALANCING */ #ifdef CONFIG_HOTPLUG_CPU +static DEFINE_PER_CPU(struct mm_struct *, idle_last_mm); + /* * Ensure that the idle task is using init_mm right before its CPU goes * offline. @@ -6186,7 +6356,11 @@ void idle_task_exit(void) current->active_mm = &init_mm; finish_arch_post_lock_switch(); } - mmdrop(mm); + /* + * Defer the cleanup to an alive cpu. On RT we can neither + * call mmdrop() nor mmdrop_delayed() from here. + */ + per_cpu(idle_last_mm, smp_processor_id()) = mm; } /* @@ -6264,6 +6438,7 @@ static void migrate_tasks(struct rq *dea break; next = __pick_migrate_task(rq); + WARN_ON_ONCE(__migrate_disabled(next)); /* * Rules for changing task_struct::cpus_mask are holding @@ -6492,6 +6667,10 @@ int sched_cpu_dying(unsigned int cpu) update_max_interval(); nohz_balance_exit_idle(rq); hrtick_clear(rq); + if (per_cpu(idle_last_mm, cpu)) { + mmdrop_delayed(per_cpu(idle_last_mm, cpu)); + per_cpu(idle_last_mm, cpu) = NULL; + } return 0; } #endif @@ -6723,7 +6902,7 @@ void __init sched_init(void) #ifdef CONFIG_DEBUG_ATOMIC_SLEEP static inline int preempt_count_equals(int preempt_offset) { - int nested = preempt_count() + rcu_preempt_depth(); + int nested = preempt_count() + sched_rcu_preempt_depth(); return (nested == preempt_offset); } @@ -7962,3 +8141,172 @@ const u32 sched_prio_to_wmult[40] = { }; #undef CREATE_TRACE_POINTS + +#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) + +static inline void +update_nr_migratory(struct task_struct *p, long delta) +{ + if (unlikely((p->sched_class == &rt_sched_class || + p->sched_class == &dl_sched_class) && + p->nr_cpus_allowed > 1)) { + if (p->sched_class == &rt_sched_class) + task_rq(p)->rt.rt_nr_migratory += delta; + else + task_rq(p)->dl.dl_nr_migratory += delta; + } +} + +static inline void +migrate_disable_update_cpus_allowed(struct task_struct *p) +{ + p->cpus_ptr = cpumask_of(smp_processor_id()); + update_nr_migratory(p, -1); + p->nr_cpus_allowed = 1; +} + +static inline void +migrate_enable_update_cpus_allowed(struct task_struct *p) +{ + struct rq *rq; + struct rq_flags rf; + + rq = task_rq_lock(p, &rf); + p->cpus_ptr = &p->cpus_mask; + p->nr_cpus_allowed = cpumask_weight(&p->cpus_mask); + update_nr_migratory(p, 1); + task_rq_unlock(rq, p, &rf); +} + +void migrate_disable(void) +{ + preempt_disable(); + + if (++current->migrate_disable == 1) { + this_rq()->nr_pinned++; + preempt_lazy_disable(); +#ifdef CONFIG_SCHED_DEBUG + WARN_ON_ONCE(current->pinned_on_cpu >= 0); + current->pinned_on_cpu = smp_processor_id(); +#endif + } + + preempt_enable(); +} +EXPORT_SYMBOL(migrate_disable); + +static void migrate_disabled_sched(struct task_struct *p) +{ + if (p->migrate_disable_scheduled) + return; + + migrate_disable_update_cpus_allowed(p); + p->migrate_disable_scheduled = 1; +} + +static DEFINE_PER_CPU(struct cpu_stop_work, migrate_work); +static DEFINE_PER_CPU(struct migration_arg, migrate_arg); + +void migrate_enable(void) +{ + struct task_struct *p = current; + struct rq *rq = this_rq(); + int cpu = task_cpu(p); + + WARN_ON_ONCE(p->migrate_disable <= 0); + if (p->migrate_disable > 1) { + p->migrate_disable--; + return; + } + + preempt_disable(); + +#ifdef CONFIG_SCHED_DEBUG + WARN_ON_ONCE(current->pinned_on_cpu != cpu); + current->pinned_on_cpu = -1; +#endif + + WARN_ON_ONCE(rq->nr_pinned < 1); + + p->migrate_disable = 0; + rq->nr_pinned--; +#ifdef CONFIG_HOTPLUG_CPU + if (rq->nr_pinned == 0 && unlikely(!cpu_active(cpu)) && + takedown_cpu_task) + wake_up_process(takedown_cpu_task); +#endif + + if (!p->migrate_disable_scheduled) + goto out; + + p->migrate_disable_scheduled = 0; + + migrate_enable_update_cpus_allowed(p); + + WARN_ON(smp_processor_id() != cpu); + if (!is_cpu_allowed(p, cpu)) { + struct migration_arg __percpu *arg; + struct cpu_stop_work __percpu *work; + struct rq_flags rf; + + work = this_cpu_ptr(&migrate_work); + arg = this_cpu_ptr(&migrate_arg); + WARN_ON_ONCE(!arg->done && !work->disabled && work->arg); + + arg->task = p; + arg->done = false; + + rq = task_rq_lock(p, &rf); + update_rq_clock(rq); + arg->dest_cpu = select_fallback_rq(cpu, p); + task_rq_unlock(rq, p, &rf); + + stop_one_cpu_nowait(task_cpu(p), migration_cpu_stop, + arg, work); + } + +out: + preempt_lazy_enable(); + preempt_enable(); +} +EXPORT_SYMBOL(migrate_enable); + +int cpu_nr_pinned(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + + return rq->nr_pinned; +} + +#elif !defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) +static void migrate_disabled_sched(struct task_struct *p) +{ +} + +void migrate_disable(void) +{ +#ifdef CONFIG_SCHED_DEBUG + current->migrate_disable++; +#endif + barrier(); +} +EXPORT_SYMBOL(migrate_disable); + +void migrate_enable(void) +{ +#ifdef CONFIG_SCHED_DEBUG + struct task_struct *p = current; + + WARN_ON_ONCE(p->migrate_disable <= 0); + p->migrate_disable--; +#endif + barrier(); +} +EXPORT_SYMBOL(migrate_enable); + +#else +static void migrate_disabled_sched(struct task_struct *p) +{ +} + +#endif Index: linux-5.6.17-rt10/kernel/sched/debug.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/sched/debug.c +++ linux-5.6.17-rt10/kernel/sched/debug.c @@ -965,6 +965,10 @@ void proc_sched_show_task(struct task_st P(dl.runtime); P(dl.deadline); } +#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT) + P(migrate_disable); +#endif + P(nr_cpus_allowed); #undef PN_SCHEDSTAT #undef PN #undef __PN Index: linux-5.6.17-rt10/kernel/sched/fair.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/sched/fair.c +++ linux-5.6.17-rt10/kernel/sched/fair.c @@ -4160,7 +4160,7 @@ check_preempt_tick(struct cfs_rq *cfs_rq ideal_runtime = sched_slice(cfs_rq, curr); delta_exec = curr->sum_exec_runtime - curr->prev_sum_exec_runtime; if (delta_exec > ideal_runtime) { - resched_curr(rq_of(cfs_rq)); + resched_curr_lazy(rq_of(cfs_rq)); /* * The current task ran long enough, ensure it doesn't get * re-elected due to buddy favours. @@ -4184,7 +4184,7 @@ check_preempt_tick(struct cfs_rq *cfs_rq return; if (delta > ideal_runtime) - resched_curr(rq_of(cfs_rq)); + resched_curr_lazy(rq_of(cfs_rq)); } static void @@ -4327,7 +4327,7 @@ entity_tick(struct cfs_rq *cfs_rq, struc * validating it and just reschedule. */ if (queued) { - resched_curr(rq_of(cfs_rq)); + resched_curr_lazy(rq_of(cfs_rq)); return; } /* @@ -4452,7 +4452,7 @@ static void __account_cfs_rq_runtime(str * hierarchy can be throttled */ if (!assign_cfs_rq_runtime(cfs_rq) && likely(cfs_rq->curr)) - resched_curr(rq_of(cfs_rq)); + resched_curr_lazy(rq_of(cfs_rq)); } static __always_inline @@ -5176,7 +5176,7 @@ static void hrtick_start_fair(struct rq if (delta < 0) { if (rq->curr == p) - resched_curr(rq); + resched_curr_lazy(rq); return; } hrtick_start(rq, delta); @@ -6686,7 +6686,7 @@ static void check_preempt_wakeup(struct return; preempt: - resched_curr(rq); + resched_curr_lazy(rq); /* * Only set the backward buddy when the current task is still * on the rq. This can happen when a wakeup gets interleaved @@ -10385,7 +10385,7 @@ static void task_fork_fair(struct task_s * 'current' within the tree based on its new key value. */ swap(curr->vruntime, se->vruntime); - resched_curr(rq); + resched_curr_lazy(rq); } se->vruntime -= cfs_rq->min_vruntime; @@ -10412,7 +10412,7 @@ prio_changed_fair(struct rq *rq, struct */ if (rq->curr == p) { if (p->prio > oldprio) - resched_curr(rq); + resched_curr_lazy(rq); } else check_preempt_curr(rq, p, 0); } Index: linux-5.6.17-rt10/kernel/sched/features.h =================================================================== --- linux-5.6.17-rt10.orig/kernel/sched/features.h +++ linux-5.6.17-rt10/kernel/sched/features.h @@ -45,11 +45,19 @@ SCHED_FEAT(DOUBLE_TICK, false) */ SCHED_FEAT(NONTASK_CAPACITY, true) +#ifdef CONFIG_PREEMPT_RT +SCHED_FEAT(TTWU_QUEUE, false) +# ifdef CONFIG_PREEMPT_LAZY +SCHED_FEAT(PREEMPT_LAZY, true) +# endif +#else + /* * Queue remote wakeups on the target CPU and process them * using the scheduler IPI. Reduces rq->lock contention/bounces. */ SCHED_FEAT(TTWU_QUEUE, true) +#endif /* * When doing wakeups, attempt to limit superfluous scans of the LLC domain. Index: linux-5.6.17-rt10/kernel/sched/sched.h =================================================================== --- linux-5.6.17-rt10.orig/kernel/sched/sched.h +++ linux-5.6.17-rt10/kernel/sched/sched.h @@ -1005,6 +1005,10 @@ struct rq { /* Must be inspected within a rcu lock section */ struct cpuidle_state *idle_state; #endif + +#if defined(CONFIG_PREEMPT_RT) && defined(CONFIG_SMP) + int nr_pinned; +#endif }; #ifdef CONFIG_FAIR_GROUP_SCHED @@ -1650,6 +1654,7 @@ static inline int task_on_rq_migrating(s #define WF_SYNC 0x01 /* Waker goes to sleep after wakeup */ #define WF_FORK 0x02 /* Child wakeup after fork */ #define WF_MIGRATED 0x4 /* Internal use, task got migrated */ +#define WF_LOCK_SLEEPER 0x08 /* wakeup spinlock "sleeper" */ /* * To aid in avoiding the subversion of "niceness" due to uneven distribution @@ -1868,6 +1873,15 @@ extern void reweight_task(struct task_st extern void resched_curr(struct rq *rq); extern void resched_cpu(int cpu); +#ifdef CONFIG_PREEMPT_LAZY +extern void resched_curr_lazy(struct rq *rq); +#else +static inline void resched_curr_lazy(struct rq *rq) +{ + resched_curr(rq); +} +#endif + extern struct rt_bandwidth def_rt_bandwidth; extern void init_rt_bandwidth(struct rt_bandwidth *rt_b, u64 period, u64 runtime); Index: linux-5.6.17-rt10/kernel/sched/swait.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/sched/swait.c +++ linux-5.6.17-rt10/kernel/sched/swait.c @@ -32,6 +32,20 @@ void swake_up_locked(struct swait_queue_ } EXPORT_SYMBOL(swake_up_locked); +void swake_up_all_locked(struct swait_queue_head *q) +{ + struct swait_queue *curr; + + while (!list_empty(&q->task_list)) { + + curr = list_first_entry(&q->task_list, typeof(*curr), + task_list); + wake_up_process(curr->task); + list_del_init(&curr->task_list); + } +} +EXPORT_SYMBOL(swake_up_all_locked); + void swake_up_one(struct swait_queue_head *q) { unsigned long flags; @@ -51,6 +65,7 @@ void swake_up_all(struct swait_queue_hea struct swait_queue *curr; LIST_HEAD(tmp); + WARN_ON(irqs_disabled()); raw_spin_lock_irq(&q->lock); list_splice_init(&q->task_list, &tmp); while (!list_empty(&tmp)) { @@ -69,7 +84,7 @@ void swake_up_all(struct swait_queue_hea } EXPORT_SYMBOL(swake_up_all); -static void __prepare_to_swait(struct swait_queue_head *q, struct swait_queue *wait) +void __prepare_to_swait(struct swait_queue_head *q, struct swait_queue *wait) { wait->task = current; if (list_empty(&wait->task_list)) Index: linux-5.6.17-rt10/kernel/sched/topology.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/sched/topology.c +++ linux-5.6.17-rt10/kernel/sched/topology.c @@ -502,6 +502,7 @@ static int init_rootdomain(struct root_d rd->rto_cpu = -1; raw_spin_lock_init(&rd->rto_lock); init_irq_work(&rd->rto_push_work, rto_push_irq_work_func); + atomic_or(IRQ_WORK_HARD_IRQ, &rd->rto_push_work.flags); #endif init_dl_bw(&rd->dl_bw); Index: linux-5.6.17-rt10/kernel/seccomp.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/seccomp.c +++ linux-5.6.17-rt10/kernel/seccomp.c @@ -268,16 +268,14 @@ static u32 seccomp_run_filters(const str * All filters in the list are evaluated and the lowest BPF return * value always takes priority (ignoring the DATA). */ - preempt_disable(); for (; f; f = f->prev) { - u32 cur_ret = BPF_PROG_RUN(f->prog, sd); + u32 cur_ret = bpf_prog_run_pin_on_cpu(f->prog, sd); if (ACTION_ONLY(cur_ret) < ACTION_ONLY(ret)) { ret = cur_ret; *match = f; } } - preempt_enable(); return ret; } #endif /* CONFIG_SECCOMP_FILTER */ Index: linux-5.6.17-rt10/kernel/signal.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/signal.c +++ linux-5.6.17-rt10/kernel/signal.c @@ -20,6 +20,7 @@ #include #include #include +#include #include #include #include @@ -403,13 +404,30 @@ void task_join_group_stop(struct task_st } } +static inline struct sigqueue *get_task_cache(struct task_struct *t) +{ + struct sigqueue *q = t->sigqueue_cache; + + if (cmpxchg(&t->sigqueue_cache, q, NULL) != q) + return NULL; + return q; +} + +static inline int put_task_cache(struct task_struct *t, struct sigqueue *q) +{ + if (cmpxchg(&t->sigqueue_cache, NULL, q) == NULL) + return 0; + return 1; +} + /* * allocate a new signal queue record * - this may be called without locks if and only if t == current, otherwise an * appropriate lock must be held to stop the target task from exiting */ static struct sigqueue * -__sigqueue_alloc(int sig, struct task_struct *t, gfp_t flags, int override_rlimit) +__sigqueue_do_alloc(int sig, struct task_struct *t, gfp_t flags, + int override_rlimit, int fromslab) { struct sigqueue *q = NULL; struct user_struct *user; @@ -431,7 +449,10 @@ __sigqueue_alloc(int sig, struct task_st rcu_read_unlock(); if (override_rlimit || likely(sigpending <= task_rlimit(t, RLIMIT_SIGPENDING))) { - q = kmem_cache_alloc(sigqueue_cachep, flags); + if (!fromslab) + q = get_task_cache(t); + if (!q) + q = kmem_cache_alloc(sigqueue_cachep, flags); } else { print_dropped_signal(sig); } @@ -448,6 +469,13 @@ __sigqueue_alloc(int sig, struct task_st return q; } +static struct sigqueue * +__sigqueue_alloc(int sig, struct task_struct *t, gfp_t flags, + int override_rlimit) +{ + return __sigqueue_do_alloc(sig, t, flags, override_rlimit, 0); +} + static void __sigqueue_free(struct sigqueue *q) { if (q->flags & SIGQUEUE_PREALLOC) @@ -457,6 +485,21 @@ static void __sigqueue_free(struct sigqu kmem_cache_free(sigqueue_cachep, q); } +static void sigqueue_free_current(struct sigqueue *q) +{ + struct user_struct *up; + + if (q->flags & SIGQUEUE_PREALLOC) + return; + + up = q->user; + if (rt_prio(current->normal_prio) && !put_task_cache(current, q)) { + if (atomic_dec_and_test(&up->sigpending)) + free_uid(up); + } else + __sigqueue_free(q); +} + void flush_sigqueue(struct sigpending *queue) { struct sigqueue *q; @@ -470,6 +513,21 @@ void flush_sigqueue(struct sigpending *q } /* + * Called from __exit_signal. Flush tsk->pending and + * tsk->sigqueue_cache + */ +void flush_task_sigqueue(struct task_struct *tsk) +{ + struct sigqueue *q; + + flush_sigqueue(&tsk->pending); + + q = get_task_cache(tsk); + if (q) + kmem_cache_free(sigqueue_cachep, q); +} + +/* * Flush all pending signals for this kthread. */ void flush_signals(struct task_struct *t) @@ -593,7 +651,7 @@ still_pending: (info->si_code == SI_TIMER) && (info->si_sys_private); - __sigqueue_free(first); + sigqueue_free_current(first); } else { /* * Ok, it wasn't in the queue. This must be @@ -630,6 +688,8 @@ int dequeue_signal(struct task_struct *t bool resched_timer = false; int signr; + WARN_ON_ONCE(tsk != current); + /* We only dequeue private signals from ourselves, we don't let * signalfd steal them */ @@ -1313,6 +1373,34 @@ force_sig_info_to_task(struct kernel_sig struct k_sigaction *action; int sig = info->si_signo; + /* + * On some archs, PREEMPT_RT has to delay sending a signal from a trap + * since it can not enable preemption, and the signal code's spin_locks + * turn into mutexes. Instead, it must set TIF_NOTIFY_RESUME which will + * send the signal on exit of the trap. + */ +#ifdef ARCH_RT_DELAYS_SIGNAL_SEND + if (in_atomic()) { + struct task_struct *t = current; + + if (WARN_ON_ONCE(t->forced_info.si_signo)) + return 0; + + if (is_si_special(info)) { + WARN_ON_ONCE(info != SEND_SIG_PRIV); + t->forced_info.si_signo = info->si_signo; + t->forced_info.si_errno = 0; + t->forced_info.si_code = SI_KERNEL; + t->forced_info.si_pid = 0; + t->forced_info.si_uid = 0; + } else { + t->forced_info = *info; + } + + set_tsk_thread_flag(t, TIF_NOTIFY_RESUME); + return 0; + } +#endif spin_lock_irqsave(&t->sighand->siglock, flags); action = &t->sighand->action[sig-1]; ignored = action->sa.sa_handler == SIG_IGN; @@ -1810,7 +1898,8 @@ EXPORT_SYMBOL(kill_pid); */ struct sigqueue *sigqueue_alloc(void) { - struct sigqueue *q = __sigqueue_alloc(-1, current, GFP_KERNEL, 0); + /* Preallocated sigqueue objects always from the slabcache ! */ + struct sigqueue *q = __sigqueue_do_alloc(-1, current, GFP_KERNEL, 0, 1); if (q) q->flags |= SIGQUEUE_PREALLOC; @@ -2206,16 +2295,8 @@ static void ptrace_stop(int exit_code, i if (gstop_done && ptrace_reparented(current)) do_notify_parent_cldstop(current, false, why); - /* - * Don't want to allow preemption here, because - * sys_ptrace() needs this task to be inactive. - * - * XXX: implement read_unlock_no_resched(). - */ - preempt_disable(); read_unlock(&tasklist_lock); cgroup_enter_frozen(); - preempt_enable_no_resched(); freezable_schedule(); cgroup_leave_frozen(true); } else { Index: linux-5.6.17-rt10/kernel/softirq.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/softirq.c +++ linux-5.6.17-rt10/kernel/softirq.c @@ -25,6 +25,9 @@ #include #include #include +#ifdef CONFIG_PREEMPT_RT +#include +#endif #define CREATE_TRACE_POINTS #include @@ -102,6 +105,104 @@ static bool ksoftirqd_running(unsigned l * softirq and whether we just have bh disabled. */ +#ifdef CONFIG_PREEMPT_RT +static DEFINE_LOCAL_IRQ_LOCK(bh_lock); +static DEFINE_PER_CPU(long, softirq_counter); + +void __local_bh_disable_ip(unsigned long ip, unsigned int cnt) +{ + unsigned long __maybe_unused flags; + long soft_cnt; + + WARN_ON_ONCE(in_irq()); + if (!in_atomic()) { + local_lock(bh_lock); + rcu_read_lock(); + } + soft_cnt = this_cpu_inc_return(softirq_counter); + WARN_ON_ONCE(soft_cnt == 0); + current->softirq_count += SOFTIRQ_DISABLE_OFFSET; + +#ifdef CONFIG_TRACE_IRQFLAGS + local_irq_save(flags); + if (soft_cnt == 1) + trace_softirqs_off(ip); + local_irq_restore(flags); +#endif +} +EXPORT_SYMBOL(__local_bh_disable_ip); + +static void local_bh_disable_rt(void) +{ + local_bh_disable(); +} + +void _local_bh_enable(void) +{ + unsigned long __maybe_unused flags; + long soft_cnt; + + soft_cnt = this_cpu_dec_return(softirq_counter); + WARN_ON_ONCE(soft_cnt < 0); + +#ifdef CONFIG_TRACE_IRQFLAGS + local_irq_save(flags); + if (soft_cnt == 0) + trace_softirqs_on(_RET_IP_); + local_irq_restore(flags); +#endif + + current->softirq_count -= SOFTIRQ_DISABLE_OFFSET; + if (!in_atomic()) { + rcu_read_unlock(); + local_unlock(bh_lock); + } +} + +void _local_bh_enable_rt(void) +{ + _local_bh_enable(); +} + +void __local_bh_enable_ip(unsigned long ip, unsigned int cnt) +{ + u32 pending; + long count; + + WARN_ON_ONCE(in_irq()); + lockdep_assert_irqs_enabled(); + + local_irq_disable(); + count = this_cpu_read(softirq_counter); + + if (unlikely(count == 1)) { + pending = local_softirq_pending(); + if (pending && !ksoftirqd_running(pending)) { + if (!in_atomic()) + __do_softirq(); + else + wakeup_softirqd(); + } + trace_softirqs_on(ip); + } + count = this_cpu_dec_return(softirq_counter); + WARN_ON_ONCE(count < 0); + local_irq_enable(); + + if (!in_atomic()) { + rcu_read_unlock(); + local_unlock(bh_lock); + } + + current->softirq_count -= SOFTIRQ_DISABLE_OFFSET; + preempt_check_resched(); +} +EXPORT_SYMBOL(__local_bh_enable_ip); + +#else +static void local_bh_disable_rt(void) { } +static void _local_bh_enable_rt(void) { } + /* * This one is for softirq.c-internal use, * where hardirqs are disabled legitimately: @@ -196,6 +297,7 @@ void __local_bh_enable_ip(unsigned long preempt_check_resched(); } EXPORT_SYMBOL(__local_bh_enable_ip); +#endif /* * We restart softirq processing for at most MAX_SOFTIRQ_RESTART times, @@ -266,7 +368,11 @@ asmlinkage __visible void __softirq_entr pending = local_softirq_pending(); account_irq_enter_time(current); +#ifdef CONFIG_PREEMPT_RT + current->softirq_count |= SOFTIRQ_OFFSET; +#else __local_bh_disable_ip(_RET_IP_, SOFTIRQ_OFFSET); +#endif in_hardirq = lockdep_softirq_start(); restart: @@ -300,9 +406,10 @@ restart: h++; pending >>= softirq_bit; } - +#ifndef CONFIG_PREEMPT_RT if (__this_cpu_read(ksoftirqd) == current) rcu_softirq_qs(); +#endif local_irq_disable(); pending = local_softirq_pending(); @@ -316,11 +423,16 @@ restart: lockdep_softirq_end(in_hardirq); account_irq_exit_time(current); +#ifdef CONFIG_PREEMPT_RT + current->softirq_count &= ~SOFTIRQ_OFFSET; +#else __local_bh_enable(SOFTIRQ_OFFSET); +#endif WARN_ON_ONCE(in_interrupt()); current_restore_flags(old_flags, PF_MEMALLOC); } +#ifndef CONFIG_PREEMPT_RT asmlinkage __visible void do_softirq(void) { __u32 pending; @@ -338,6 +450,7 @@ asmlinkage __visible void do_softirq(voi local_irq_restore(flags); } +#endif /* * Enter an interrupt context. @@ -358,6 +471,16 @@ void irq_enter(void) __irq_enter(); } +#ifdef CONFIG_PREEMPT_RT + +static inline void invoke_softirq(void) +{ + if (this_cpu_read(softirq_counter) == 0) + wakeup_softirqd(); +} + +#else + static inline void invoke_softirq(void) { if (ksoftirqd_running(local_softirq_pending())) @@ -383,6 +506,7 @@ static inline void invoke_softirq(void) wakeup_softirqd(); } } +#endif static inline void tick_irq_exit(void) { @@ -420,6 +544,27 @@ void irq_exit(void) /* * This function must run with irqs disabled! */ +#ifdef CONFIG_PREEMPT_RT +void raise_softirq_irqoff(unsigned int nr) +{ + __raise_softirq_irqoff(nr); + + /* + * If we're in an hard interrupt we let irq return code deal + * with the wakeup of ksoftirqd. + */ + if (in_irq()) + return; + /* + * If were are not in BH-disabled section then we have to wake + * ksoftirqd. + */ + if (this_cpu_read(softirq_counter) == 0) + wakeup_softirqd(); +} + +#else + inline void raise_softirq_irqoff(unsigned int nr) { __raise_softirq_irqoff(nr); @@ -437,6 +582,8 @@ inline void raise_softirq_irqoff(unsigne wakeup_softirqd(); } +#endif + void raise_softirq(unsigned int nr) { unsigned long flags; @@ -564,7 +711,8 @@ void tasklet_kill(struct tasklet_struct while (test_and_set_bit(TASKLET_STATE_SCHED, &t->state)) { do { - yield(); + local_bh_disable(); + local_bh_enable(); } while (test_bit(TASKLET_STATE_SCHED, &t->state)); } tasklet_unlock_wait(t); @@ -594,6 +742,7 @@ static int ksoftirqd_should_run(unsigned static void run_ksoftirqd(unsigned int cpu) { + local_bh_disable_rt(); local_irq_disable(); if (local_softirq_pending()) { /* @@ -602,10 +751,12 @@ static void run_ksoftirqd(unsigned int c */ __do_softirq(); local_irq_enable(); + _local_bh_enable_rt(); cond_resched(); return; } local_irq_enable(); + _local_bh_enable_rt(); } #ifdef CONFIG_HOTPLUG_CPU @@ -679,6 +830,13 @@ static struct smp_hotplug_thread softirq static __init int spawn_ksoftirqd(void) { +#ifdef CONFIG_PREEMPT_RT + int cpu; + + for_each_possible_cpu(cpu) + lockdep_set_novalidate_class(per_cpu_ptr(&bh_lock.lock, cpu)); +#endif + cpuhp_setup_state_nocalls(CPUHP_SOFTIRQ_DEAD, "softirq:dead", NULL, takeover_tasklets); BUG_ON(smpboot_register_percpu_thread(&softirq_threads)); @@ -687,6 +845,75 @@ static __init int spawn_ksoftirqd(void) } early_initcall(spawn_ksoftirqd); +#ifdef CONFIG_PREEMPT_RT + +/* + * On preempt-rt a softirq running context might be blocked on a + * lock. There might be no other runnable task on this CPU because the + * lock owner runs on some other CPU. So we have to go into idle with + * the pending bit set. Therefor we need to check this otherwise we + * warn about false positives which confuses users and defeats the + * whole purpose of this test. + * + * This code is called with interrupts disabled. + */ +void softirq_check_pending_idle(void) +{ + struct task_struct *tsk = __this_cpu_read(ksoftirqd); + static int rate_limit; + bool okay = false; + u32 warnpending; + + if (rate_limit >= 10) + return; + + warnpending = local_softirq_pending() & SOFTIRQ_STOP_IDLE_MASK; + if (!warnpending) + return; + + if (!tsk) + return; + /* + * If ksoftirqd is blocked on a lock then we may go idle with pending + * softirq. + */ + raw_spin_lock(&tsk->pi_lock); + if (tsk->pi_blocked_on || tsk->state == TASK_RUNNING || + (tsk->state == TASK_UNINTERRUPTIBLE && tsk->sleeping_lock)) { + okay = true; + } + raw_spin_unlock(&tsk->pi_lock); + if (okay) + return; + /* + * The softirq lock is held in non-atomic context and the owner is + * blocking on a lock. It will schedule softirqs once the counter goes + * back to zero. + */ + if (this_cpu_read(softirq_counter) > 0) + return; + + printk(KERN_ERR "NOHZ: local_softirq_pending %02x\n", + warnpending); + rate_limit++; +} + +#else + +void softirq_check_pending_idle(void) +{ + static int ratelimit; + + if (ratelimit < 10 && + (local_softirq_pending() & SOFTIRQ_STOP_IDLE_MASK)) { + pr_warn("NOHZ: local_softirq_pending %02x\n", + (unsigned int) local_softirq_pending()); + ratelimit++; + } +} + +#endif + /* * [ These __weak aliases are kept in a separate compilation unit, so that * GCC does not inline them incorrectly. ] Index: linux-5.6.17-rt10/kernel/stop_machine.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/stop_machine.c +++ linux-5.6.17-rt10/kernel/stop_machine.c @@ -86,8 +86,11 @@ static bool cpu_stop_queue_work(unsigned enabled = stopper->enabled; if (enabled) __cpu_stop_queue_work(stopper, work, &wakeq); - else if (work->done) - cpu_stop_signal_done(work->done); + else { + work->disabled = true; + if (work->done) + cpu_stop_signal_done(work->done); + } raw_spin_unlock_irqrestore(&stopper->lock, flags); wake_up_q(&wakeq); Index: linux-5.6.17-rt10/kernel/sysctl.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/sysctl.c +++ linux-5.6.17-rt10/kernel/sysctl.c @@ -212,6 +212,11 @@ static int proc_do_cad_pid(struct ctl_ta void __user *buffer, size_t *lenp, loff_t *ppos); static int proc_taint(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos); +#ifdef CONFIG_COMPACTION +static int proc_dointvec_minmax_warn_RT_change(struct ctl_table *table, + int write, void __user *buffer, + size_t *lenp, loff_t *ppos); +#endif #endif #ifdef CONFIG_PRINTK @@ -1484,7 +1489,7 @@ static struct ctl_table vm_table[] = { .data = &sysctl_compact_unevictable_allowed, .maxlen = sizeof(int), .mode = 0644, - .proc_handler = proc_dointvec, + .proc_handler = proc_dointvec_minmax_warn_RT_change, .extra1 = SYSCTL_ZERO, .extra2 = SYSCTL_ONE, }, @@ -2572,6 +2577,28 @@ int proc_dointvec(struct ctl_table *tabl return do_proc_dointvec(table, write, buffer, lenp, ppos, NULL, NULL); } +#ifdef CONFIG_COMPACTION +static int proc_dointvec_minmax_warn_RT_change(struct ctl_table *table, + int write, void __user *buffer, + size_t *lenp, loff_t *ppos) +{ + int ret, old; + + if (!IS_ENABLED(CONFIG_PREEMPT_RT) || !write) + return proc_dointvec_minmax(table, write, buffer, lenp, ppos); + + old = *(int *)table->data; + ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); + if (ret) + return ret; + if (old != *(int *)table->data) + pr_warn_once("sysctl attribute %s changed by %s[%d]\n", + table->procname, current->comm, + task_pid_nr(current)); + return ret; +} +#endif + /** * proc_douintvec - read a vector of unsigned integers * @table: the sysctl table Index: linux-5.6.17-rt10/kernel/time/hrtimer.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/time/hrtimer.c +++ linux-5.6.17-rt10/kernel/time/hrtimer.c @@ -135,7 +135,11 @@ static const int hrtimer_clock_to_base_t * timer->base->cpu_base */ static struct hrtimer_cpu_base migration_cpu_base = { - .clock_base = { { .cpu_base = &migration_cpu_base, }, }, + .clock_base = { { + .cpu_base = &migration_cpu_base, + .seq = SEQCNT_RAW_SPINLOCK_ZERO(migration_cpu_base.seq, + &migration_cpu_base.lock), + }, }, }; #define migration_base migration_cpu_base.clock_base[0] @@ -1819,7 +1823,7 @@ static void __hrtimer_init_sleeper(struc * expiry. */ if (IS_ENABLED(CONFIG_PREEMPT_RT)) { - if (task_is_realtime(current) && !(mode & HRTIMER_MODE_SOFT)) + if ((task_is_realtime(current) && !(mode & HRTIMER_MODE_SOFT)) || system_state != SYSTEM_RUNNING) mode |= HRTIMER_MODE_HARD; } @@ -1984,6 +1988,38 @@ SYSCALL_DEFINE2(nanosleep_time32, struct } #endif +#ifdef CONFIG_PREEMPT_RT +/* + * Sleep for 1 ms in hope whoever holds what we want will let it go. + */ +void cpu_chill(void) +{ + unsigned int freeze_flag = current->flags & PF_NOFREEZE; + struct task_struct *self = current; + ktime_t chill_time; + + raw_spin_lock_irq(&self->pi_lock); + self->saved_state = self->state; + __set_current_state_no_track(TASK_UNINTERRUPTIBLE); + raw_spin_unlock_irq(&self->pi_lock); + + chill_time = ktime_set(0, NSEC_PER_MSEC); + + current->flags |= PF_NOFREEZE; + sleeping_lock_inc(); + schedule_hrtimeout(&chill_time, HRTIMER_MODE_REL_HARD); + sleeping_lock_dec(); + if (!freeze_flag) + current->flags &= ~PF_NOFREEZE; + + raw_spin_lock_irq(&self->pi_lock); + __set_current_state_no_track(self->saved_state); + self->saved_state = TASK_RUNNING; + raw_spin_unlock_irq(&self->pi_lock); +} +EXPORT_SYMBOL(cpu_chill); +#endif + /* * Functions related to boot-time initialization: */ @@ -1993,8 +2029,11 @@ int hrtimers_prepare_cpu(unsigned int cp int i; for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++) { - cpu_base->clock_base[i].cpu_base = cpu_base; - timerqueue_init_head(&cpu_base->clock_base[i].active); + struct hrtimer_clock_base *clock_b = &cpu_base->clock_base[i]; + + clock_b->cpu_base = cpu_base; + seqcount_raw_spinlock_init(&clock_b->seq, &cpu_base->lock); + timerqueue_init_head(&clock_b->active); } cpu_base->cpu = cpu; Index: linux-5.6.17-rt10/kernel/time/jiffies.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/time/jiffies.c +++ linux-5.6.17-rt10/kernel/time/jiffies.c @@ -58,7 +58,8 @@ static struct clocksource clocksource_ji .max_cycles = 10, }; -__cacheline_aligned_in_smp DEFINE_SEQLOCK(jiffies_lock); +__cacheline_aligned_in_smp DEFINE_RAW_SPINLOCK(jiffies_lock); +__cacheline_aligned_in_smp seqcount_t jiffies_seq; #if (BITS_PER_LONG < 64) u64 get_jiffies_64(void) @@ -67,9 +68,9 @@ u64 get_jiffies_64(void) u64 ret; do { - seq = read_seqbegin(&jiffies_lock); + seq = read_seqcount_begin(&jiffies_seq); ret = jiffies_64; - } while (read_seqretry(&jiffies_lock, seq)); + } while (read_seqcount_retry(&jiffies_seq, seq)); return ret; } EXPORT_SYMBOL(get_jiffies_64); Index: linux-5.6.17-rt10/kernel/time/posix-cpu-timers.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/time/posix-cpu-timers.c +++ linux-5.6.17-rt10/kernel/time/posix-cpu-timers.c @@ -3,8 +3,10 @@ * Implement CPU time clocks for the POSIX clock interface. */ +#include #include #include +#include #include #include #include @@ -15,6 +17,7 @@ #include #include #include +#include #include "posix-timers.h" @@ -27,6 +30,9 @@ void posix_cputimers_group_init(struct p pct->bases[CPUCLOCK_PROF].nextevt = cpu_limit * NSEC_PER_SEC; pct->timers_active = true; } +#ifdef CONFIG_PREEMPT_RT + pct->posix_timer_list = NULL; +#endif } /* @@ -440,6 +446,20 @@ static int posix_cpu_timer_del(struct k_ return ret; } +static DEFINE_PER_CPU(spinlock_t, cpu_timer_expiry_lock) = __SPIN_LOCK_UNLOCKED(cpu_timer_expiry_lock); + +static void posix_cpu_wait_running(struct k_itimer *timer) +{ + int cpu = timer->it.cpu.firing_cpu; + + if (cpu >= 0) { + spinlock_t *expiry_lock = per_cpu_ptr(&cpu_timer_expiry_lock, cpu); + + spin_lock_irq(expiry_lock); + spin_unlock_irq(expiry_lock); + } +} + static void cleanup_timerqueue(struct timerqueue_head *head) { struct timerqueue_node *node; @@ -777,6 +797,7 @@ static u64 collect_timerqueue(struct tim return expires; ctmr->firing = 1; + ctmr->firing_cpu = smp_processor_id(); cpu_timer_dequeue(ctmr); list_add_tail(&ctmr->elist, firing); } @@ -804,7 +825,8 @@ static inline void check_dl_overrun(stru } } -static bool check_rlimit(u64 time, u64 limit, int signo, bool rt, bool hard) +static bool check_rlimit(struct task_struct *tsk, u64 time, u64 limit, + int signo, bool rt, bool hard) { if (time < limit) return false; @@ -812,9 +834,9 @@ static bool check_rlimit(u64 time, u64 l if (print_fatal_signals) { pr_info("%s Watchdog Timeout (%s): %s[%d]\n", rt ? "RT" : "CPU", hard ? "hard" : "soft", - current->comm, task_pid_nr(current)); + tsk->comm, task_pid_nr(tsk)); } - __group_send_sig_info(signo, SEND_SIG_PRIV, current); + __group_send_sig_info(signo, SEND_SIG_PRIV, tsk); return true; } @@ -850,11 +872,11 @@ static void check_thread_timers(struct t /* At the hard limit, send SIGKILL. No further action. */ if (hard != RLIM_INFINITY && - check_rlimit(rttime, hard, SIGKILL, true, true)) + check_rlimit(tsk, rttime, hard, SIGKILL, true, true)) return; /* At the soft limit, send a SIGXCPU every second */ - if (check_rlimit(rttime, soft, SIGXCPU, true, false)) { + if (check_rlimit(tsk, rttime, soft, SIGXCPU, true, false)) { soft += USEC_PER_SEC; tsk->signal->rlim[RLIMIT_RTTIME].rlim_cur = soft; } @@ -949,11 +971,11 @@ static void check_process_timers(struct /* At the hard limit, send SIGKILL. No further action. */ if (hard != RLIM_INFINITY && - check_rlimit(ptime, hardns, SIGKILL, false, true)) + check_rlimit(tsk, ptime, hardns, SIGKILL, false, true)) return; /* At the soft limit, send a SIGXCPU every second */ - if (check_rlimit(ptime, softns, SIGXCPU, false, false)) { + if (check_rlimit(tsk, ptime, softns, SIGXCPU, false, false)) { sig->rlim[RLIMIT_CPU].rlim_cur = soft + 1; softns += NSEC_PER_SEC; } @@ -1110,15 +1132,13 @@ static inline bool fastpath_timer_check( * already updated our counts. We need to check if any timers fire now. * Interrupts are disabled. */ -void run_posix_cpu_timers(void) +static void __run_posix_cpu_timers(struct task_struct *tsk) { - struct task_struct *tsk = current; struct k_itimer *timer, *next; unsigned long flags; + spinlock_t *expiry_lock; LIST_HEAD(firing); - lockdep_assert_irqs_disabled(); - /* * The fast path checks that there are no expired thread or thread * group timers. If that's so, just return. @@ -1126,8 +1146,13 @@ void run_posix_cpu_timers(void) if (!fastpath_timer_check(tsk)) return; - if (!lock_task_sighand(tsk, &flags)) + expiry_lock = this_cpu_ptr(&cpu_timer_expiry_lock); + spin_lock(expiry_lock); + + if (!lock_task_sighand(tsk, &flags)) { + spin_unlock(expiry_lock); return; + } /* * Here we take off tsk->signal->cpu_timers[N] and * tsk->cpu_timers[N] all the timers that are firing, and @@ -1160,6 +1185,7 @@ void run_posix_cpu_timers(void) list_del_init(&timer->it.cpu.elist); cpu_firing = timer->it.cpu.firing; timer->it.cpu.firing = 0; + timer->it.cpu.firing_cpu = -1; /* * The firing flag is -1 if we collided with a reset * of the timer, which already reported this @@ -1169,8 +1195,158 @@ void run_posix_cpu_timers(void) cpu_timer_fire(timer); spin_unlock(&timer->it_lock); } + spin_unlock(expiry_lock); } +#ifdef CONFIG_PREEMPT_RT +#include +#include +DEFINE_PER_CPU(struct task_struct *, posix_timer_task); +DEFINE_PER_CPU(struct task_struct *, posix_timer_tasklist); +DEFINE_PER_CPU(bool, posix_timer_th_active); + +static void posix_cpu_kthread_fn(unsigned int cpu) +{ + struct task_struct *tsk = NULL; + struct task_struct *next = NULL; + + BUG_ON(per_cpu(posix_timer_task, cpu) != current); + + /* grab task list */ + raw_local_irq_disable(); + tsk = per_cpu(posix_timer_tasklist, cpu); + per_cpu(posix_timer_tasklist, cpu) = NULL; + raw_local_irq_enable(); + + /* its possible the list is empty, just return */ + if (!tsk) + return; + + /* Process task list */ + while (1) { + /* save next */ + next = tsk->posix_cputimers.posix_timer_list; + + /* run the task timers, clear its ptr and + * unreference it + */ + __run_posix_cpu_timers(tsk); + tsk->posix_cputimers.posix_timer_list = NULL; + put_task_struct(tsk); + + /* check if this is the last on the list */ + if (next == tsk) + break; + tsk = next; + } +} + +static inline int __fastpath_timer_check(struct task_struct *tsk) +{ + /* tsk == current, ensure it is safe to use ->signal/sighand */ + if (unlikely(tsk->exit_state)) + return 0; + + if (!expiry_cache_is_inactive(&tsk->posix_cputimers)) + return 1; + + if (!expiry_cache_is_inactive(&tsk->signal->posix_cputimers)) + return 1; + + return 0; +} + +void run_posix_cpu_timers(void) +{ + unsigned int cpu = smp_processor_id(); + struct task_struct *tsk = current; + struct task_struct *tasklist; + + BUG_ON(!irqs_disabled()); + + if (per_cpu(posix_timer_th_active, cpu) != true) + return; + + /* get per-cpu references */ + tasklist = per_cpu(posix_timer_tasklist, cpu); + + /* check to see if we're already queued */ + if (!tsk->posix_cputimers.posix_timer_list && __fastpath_timer_check(tsk)) { + get_task_struct(tsk); + if (tasklist) { + tsk->posix_cputimers.posix_timer_list = tasklist; + } else { + /* + * The list is terminated by a self-pointing + * task_struct + */ + tsk->posix_cputimers.posix_timer_list = tsk; + } + per_cpu(posix_timer_tasklist, cpu) = tsk; + + wake_up_process(per_cpu(posix_timer_task, cpu)); + } +} + +static int posix_cpu_kthread_should_run(unsigned int cpu) +{ + return __this_cpu_read(posix_timer_tasklist) != NULL; +} + +static void posix_cpu_kthread_park(unsigned int cpu) +{ + this_cpu_write(posix_timer_th_active, false); +} + +static void posix_cpu_kthread_unpark(unsigned int cpu) +{ + this_cpu_write(posix_timer_th_active, true); +} + +static void posix_cpu_kthread_setup(unsigned int cpu) +{ + struct sched_param sp; + + sp.sched_priority = MAX_RT_PRIO - 1; + sched_setscheduler_nocheck(current, SCHED_FIFO, &sp); + posix_cpu_kthread_unpark(cpu); +} + +static struct smp_hotplug_thread posix_cpu_thread = { + .store = &posix_timer_task, + .thread_should_run = posix_cpu_kthread_should_run, + .thread_fn = posix_cpu_kthread_fn, + .thread_comm = "posixcputmr/%u", + .setup = posix_cpu_kthread_setup, + .park = posix_cpu_kthread_park, + .unpark = posix_cpu_kthread_unpark, +}; + +static int __init posix_cpu_thread_init(void) +{ + /* Start one for boot CPU. */ + unsigned long cpu; + int ret; + + /* init the per-cpu posix_timer_tasklets */ + for_each_possible_cpu(cpu) + per_cpu(posix_timer_tasklist, cpu) = NULL; + + ret = smpboot_register_percpu_thread(&posix_cpu_thread); + WARN_ON(ret); + + return 0; +} +early_initcall(posix_cpu_thread_init); + +#else /* CONFIG_PREEMPT_RT */ +void run_posix_cpu_timers(void) +{ + lockdep_assert_irqs_disabled(); + __run_posix_cpu_timers(current); +} +#endif /* CONFIG_PREEMPT_RT */ + /* * Set one of the process-wide special case CPU timers or RLIMIT_CPU. * The tsk->sighand->siglock must be held by the caller. @@ -1282,6 +1458,8 @@ static int do_cpu_nanosleep(const clocki spin_unlock_irq(&timer.it_lock); while (error == TIMER_RETRY) { + + posix_cpu_wait_running(&timer); /* * We need to handle case when timer was or is in the * middle of firing. In other cases we already freed @@ -1400,6 +1578,7 @@ const struct k_clock clock_posix_cpu = { .timer_del = posix_cpu_timer_del, .timer_get = posix_cpu_timer_get, .timer_rearm = posix_cpu_timer_rearm, + .timer_wait_running = posix_cpu_wait_running, }; const struct k_clock clock_process = { Index: linux-5.6.17-rt10/kernel/time/tick-common.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/time/tick-common.c +++ linux-5.6.17-rt10/kernel/time/tick-common.c @@ -84,13 +84,15 @@ int tick_is_oneshot_available(void) static void tick_periodic(int cpu) { if (tick_do_timer_cpu == cpu) { - write_seqlock(&jiffies_lock); + raw_spin_lock(&jiffies_lock); + write_seqcount_begin(&jiffies_seq); /* Keep track of the next tick event */ tick_next_period = ktime_add(tick_next_period, tick_period); do_timer(1); - write_sequnlock(&jiffies_lock); + write_seqcount_end(&jiffies_seq); + raw_spin_unlock(&jiffies_lock); update_wall_time(); } @@ -162,9 +164,9 @@ void tick_setup_periodic(struct clock_ev ktime_t next; do { - seq = read_seqbegin(&jiffies_lock); + seq = read_seqcount_begin(&jiffies_seq); next = tick_next_period; - } while (read_seqretry(&jiffies_lock, seq)); + } while (read_seqcount_retry(&jiffies_seq, seq)); clockevents_switch_state(dev, CLOCK_EVT_STATE_ONESHOT); Index: linux-5.6.17-rt10/kernel/time/tick-sched.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/time/tick-sched.c +++ linux-5.6.17-rt10/kernel/time/tick-sched.c @@ -65,7 +65,8 @@ static void tick_do_update_jiffies64(kti return; /* Reevaluate with jiffies_lock held */ - write_seqlock(&jiffies_lock); + raw_spin_lock(&jiffies_lock); + write_seqcount_begin(&jiffies_seq); delta = ktime_sub(now, last_jiffies_update); if (delta >= tick_period) { @@ -91,10 +92,12 @@ static void tick_do_update_jiffies64(kti /* Keep the tick_next_period variable up to date */ tick_next_period = ktime_add(last_jiffies_update, tick_period); } else { - write_sequnlock(&jiffies_lock); + write_seqcount_end(&jiffies_seq); + raw_spin_unlock(&jiffies_lock); return; } - write_sequnlock(&jiffies_lock); + write_seqcount_end(&jiffies_seq); + raw_spin_unlock(&jiffies_lock); update_wall_time(); } @@ -105,12 +108,14 @@ static ktime_t tick_init_jiffy_update(vo { ktime_t period; - write_seqlock(&jiffies_lock); + raw_spin_lock(&jiffies_lock); + write_seqcount_begin(&jiffies_seq); /* Did we start the jiffies update yet ? */ if (last_jiffies_update == 0) last_jiffies_update = tick_next_period; period = last_jiffies_update; - write_sequnlock(&jiffies_lock); + write_seqcount_end(&jiffies_seq); + raw_spin_unlock(&jiffies_lock); return period; } @@ -240,6 +245,7 @@ static void nohz_full_kick_func(struct i static DEFINE_PER_CPU(struct irq_work, nohz_full_kick_work) = { .func = nohz_full_kick_func, + .flags = ATOMIC_INIT(IRQ_WORK_HARD_IRQ), }; /* @@ -676,10 +682,10 @@ static ktime_t tick_nohz_next_event(stru /* Read jiffies and the time when jiffies were updated last */ do { - seq = read_seqbegin(&jiffies_lock); + seq = read_seqcount_begin(&jiffies_seq); basemono = last_jiffies_update; basejiff = jiffies; - } while (read_seqretry(&jiffies_lock, seq)); + } while (read_seqcount_retry(&jiffies_seq, seq)); ts->last_jiffies = basejiff; ts->timer_expires_base = basemono; @@ -909,14 +915,7 @@ static bool can_stop_idle_tick(int cpu, return false; if (unlikely(local_softirq_pending())) { - static int ratelimit; - - if (ratelimit < 10 && - (local_softirq_pending() & SOFTIRQ_STOP_IDLE_MASK)) { - pr_warn("NOHZ: local_softirq_pending %02x\n", - (unsigned int) local_softirq_pending()); - ratelimit++; - } + softirq_check_pending_idle(); return false; } Index: linux-5.6.17-rt10/kernel/time/timekeeping.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/time/timekeeping.c +++ linux-5.6.17-rt10/kernel/time/timekeeping.c @@ -39,18 +39,19 @@ enum timekeeping_adv_mode { TK_ADV_FREQ }; +static DEFINE_RAW_SPINLOCK(timekeeper_lock); + /* * The most important data for readout fits into a single 64 byte * cache line. */ static struct { - seqcount_t seq; + seqcount_raw_spinlock_t seq; struct timekeeper timekeeper; } tk_core ____cacheline_aligned = { - .seq = SEQCNT_ZERO(tk_core.seq), + .seq = SEQCNT_RAW_SPINLOCK_ZERO(tk_core.seq, &timekeeper_lock), }; -static DEFINE_RAW_SPINLOCK(timekeeper_lock); static struct timekeeper shadow_timekeeper; /** @@ -63,7 +64,7 @@ static struct timekeeper shadow_timekeep * See @update_fast_timekeeper() below. */ struct tk_fast { - seqcount_t seq; + seqcount_raw_spinlock_t seq; struct tk_read_base base[2]; }; @@ -80,11 +81,13 @@ static struct clocksource dummy_clock = }; static struct tk_fast tk_fast_mono ____cacheline_aligned = { + .seq = SEQCNT_RAW_SPINLOCK_ZERO(tk_fast_mono.seq, &timekeeper_lock), .base[0] = { .clock = &dummy_clock, }, .base[1] = { .clock = &dummy_clock, }, }; static struct tk_fast tk_fast_raw ____cacheline_aligned = { + .seq = SEQCNT_RAW_SPINLOCK_ZERO(tk_fast_raw.seq, &timekeeper_lock), .base[0] = { .clock = &dummy_clock, }, .base[1] = { .clock = &dummy_clock, }, }; @@ -157,7 +160,7 @@ static inline void tk_update_sleep_time( * tk_clock_read - atomic clocksource read() helper * * This helper is necessary to use in the read paths because, while the - * seqlock ensures we don't return a bad value while structures are updated, + * seqcount ensures we don't return a bad value while structures are updated, * it doesn't protect from potential crashes. There is the possibility that * the tkr's clocksource may change between the read reference, and the * clock reference passed to the read function. This can cause crashes if @@ -222,10 +225,10 @@ static inline u64 timekeeping_get_delta( unsigned int seq; /* - * Since we're called holding a seqlock, the data may shift + * Since we're called holding a seqcount, the data may shift * under us while we're doing the calculation. This can cause * false positives, since we'd note a problem but throw the - * results away. So nest another seqlock here to atomically + * results away. So nest another seqcount here to atomically * grab the points we are checking with. */ do { @@ -486,7 +489,7 @@ EXPORT_SYMBOL_GPL(ktime_get_raw_fast_ns) * * To keep it NMI safe since we're accessing from tracing, we're not using a * separate timekeeper with updates to monotonic clock and boot offset - * protected with seqlocks. This has the following minor side effects: + * protected with seqcounts. This has the following minor side effects: * * (1) Its possible that a timestamp be taken after the boot offset is updated * but before the timekeeper is updated. If this happens, the new boot offset @@ -2397,8 +2400,10 @@ EXPORT_SYMBOL(hardpps); */ void xtime_update(unsigned long ticks) { - write_seqlock(&jiffies_lock); + raw_spin_lock(&jiffies_lock); + write_seqcount_begin(&jiffies_seq); do_timer(ticks); - write_sequnlock(&jiffies_lock); + write_seqcount_end(&jiffies_seq); + raw_spin_unlock(&jiffies_lock); update_wall_time(); } Index: linux-5.6.17-rt10/kernel/time/timekeeping.h =================================================================== --- linux-5.6.17-rt10.orig/kernel/time/timekeeping.h +++ linux-5.6.17-rt10/kernel/time/timekeeping.h @@ -25,7 +25,8 @@ static inline void sched_clock_resume(vo extern void do_timer(unsigned long ticks); extern void update_wall_time(void); -extern seqlock_t jiffies_lock; +extern raw_spinlock_t jiffies_lock; +extern seqcount_t jiffies_seq; #define CS_NAME_LEN 32 Index: linux-5.6.17-rt10/kernel/time/timer.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/time/timer.c +++ linux-5.6.17-rt10/kernel/time/timer.c @@ -1783,6 +1783,8 @@ static __latent_entropy void run_timer_s { struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_STD]); + irq_work_tick_soft(); + __run_timers(base); if (IS_ENABLED(CONFIG_NO_HZ_COMMON)) __run_timers(this_cpu_ptr(&timer_bases[BASE_DEF])); Index: linux-5.6.17-rt10/kernel/trace/bpf_trace.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/trace/bpf_trace.c +++ linux-5.6.17-rt10/kernel/trace/bpf_trace.c @@ -83,7 +83,7 @@ unsigned int trace_call_bpf(struct trace if (in_nmi()) /* not supported yet */ return 1; - preempt_disable(); + cant_sleep(); if (unlikely(__this_cpu_inc_return(bpf_prog_active) != 1)) { /* @@ -115,11 +115,9 @@ unsigned int trace_call_bpf(struct trace out: __this_cpu_dec(bpf_prog_active); - preempt_enable(); return ret; } -EXPORT_SYMBOL_GPL(trace_call_bpf); #ifdef CONFIG_BPF_KPROBE_OVERRIDE BPF_CALL_2(bpf_override_return, struct pt_regs *, regs, unsigned long, rc) @@ -1507,10 +1505,9 @@ void bpf_put_raw_tracepoint(struct bpf_r static __always_inline void __bpf_trace_run(struct bpf_prog *prog, u64 *args) { + cant_sleep(); rcu_read_lock(); - preempt_disable(); (void) BPF_PROG_RUN(prog, args); - preempt_enable(); rcu_read_unlock(); } Index: linux-5.6.17-rt10/kernel/trace/trace.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/trace/trace.c +++ linux-5.6.17-rt10/kernel/trace/trace.c @@ -2437,6 +2437,7 @@ tracing_generic_entry_update(struct trac struct task_struct *tsk = current; entry->preempt_count = pc & 0xff; + entry->preempt_lazy_count = preempt_lazy_count(); entry->pid = (tsk) ? tsk->pid : 0; entry->type = type; entry->flags = @@ -2448,8 +2449,11 @@ tracing_generic_entry_update(struct trac ((pc & NMI_MASK ) ? TRACE_FLAG_NMI : 0) | ((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) | ((pc & SOFTIRQ_OFFSET) ? TRACE_FLAG_SOFTIRQ : 0) | - (tif_need_resched() ? TRACE_FLAG_NEED_RESCHED : 0) | + (tif_need_resched_now() ? TRACE_FLAG_NEED_RESCHED : 0) | + (need_resched_lazy() ? TRACE_FLAG_NEED_RESCHED_LAZY : 0) | (test_preempt_need_resched() ? TRACE_FLAG_PREEMPT_RESCHED : 0); + + entry->migrate_disable = (tsk) ? __migrate_disabled(tsk) & 0xFF : 0; } EXPORT_SYMBOL_GPL(tracing_generic_entry_update); @@ -3693,11 +3697,13 @@ static void print_lat_help_header(struct seq_puts(m, "# _------=> CPU# \n" "# / _-----=> irqs-off \n" "# | / _----=> need-resched \n" - "# || / _---=> hardirq/softirq \n" - "# ||| / _--=> preempt-depth \n" - "# |||| / delay \n" - "# cmd pid ||||| time | caller \n" - "# \\ / ||||| \\ | / \n"); + "# || / _----=> need-resched_lazy\n" + "# ||| / _---=> hardirq/softirq \n" + "# |||| / _--=> preempt-depth \n" + "# ||||| / _-=> migrate-disable \n" + "# |||||| / delay \n" + "# cmd pid ||||||| time | caller \n" + "# \\ / |||||| \\ | / \n"); } static void print_event_info(struct array_buffer *buf, struct seq_file *m) @@ -3733,11 +3739,12 @@ static void print_func_help_header_irq(s seq_printf(m, "# %.*s _-----=> irqs-off\n", prec, space); seq_printf(m, "# %.*s / _----=> need-resched\n", prec, space); - seq_printf(m, "# %.*s| / _---=> hardirq/softirq\n", prec, space); - seq_printf(m, "# %.*s|| / _--=> preempt-depth\n", prec, space); - seq_printf(m, "# %.*s||| / delay\n", prec, space); - seq_printf(m, "# TASK-PID %.*sCPU# |||| TIMESTAMP FUNCTION\n", prec, " TGID "); - seq_printf(m, "# | | %.*s | |||| | |\n", prec, " | "); + seq_printf(m, "# %.*s| / _----=> need-resched\n", prec, space); + seq_printf(m, "# %.*s|| / _---=> hardirq/softirq\n", prec, space); + seq_printf(m, "# %.*s||| / _--=> preempt-depth\n", prec, space); + seq_printf(m, "# %.*s||||/ delay\n", prec, space); + seq_printf(m, "# TASK-PID %.*sCPU# ||||| TIMESTAMP FUNCTION\n", prec, " TGID "); + seq_printf(m, "# | | %.*s | ||||| | |\n", prec, " | "); } void @@ -9155,7 +9162,6 @@ void ftrace_dump(enum ftrace_dump_mode o tracing_off(); local_irq_save(flags); - printk_nmi_direct_enter(); /* Simulate the iterator */ trace_init_global_iter(&iter); @@ -9232,7 +9238,6 @@ void ftrace_dump(enum ftrace_dump_mode o atomic_dec(&per_cpu_ptr(iter.array_buffer->data, cpu)->disabled); } atomic_dec(&dump_running); - printk_nmi_direct_exit(); local_irq_restore(flags); } EXPORT_SYMBOL_GPL(ftrace_dump); Index: linux-5.6.17-rt10/kernel/trace/trace.h =================================================================== --- linux-5.6.17-rt10.orig/kernel/trace/trace.h +++ linux-5.6.17-rt10/kernel/trace/trace.h @@ -140,6 +140,7 @@ struct kretprobe_trace_entry_head { * NEED_RESCHED - reschedule is requested * HARDIRQ - inside an interrupt handler * SOFTIRQ - inside a softirq handler + * NEED_RESCHED_LAZY - lazy reschedule is requested */ enum trace_flag_type { TRACE_FLAG_IRQS_OFF = 0x01, @@ -149,6 +150,7 @@ enum trace_flag_type { TRACE_FLAG_SOFTIRQ = 0x10, TRACE_FLAG_PREEMPT_RESCHED = 0x20, TRACE_FLAG_NMI = 0x40, + TRACE_FLAG_NEED_RESCHED_LAZY = 0x80, }; #define TRACE_BUF_SIZE 1024 Index: linux-5.6.17-rt10/kernel/trace/trace_events.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/trace/trace_events.c +++ linux-5.6.17-rt10/kernel/trace/trace_events.c @@ -182,6 +182,8 @@ static int trace_define_common_fields(vo __common_field(unsigned char, flags); __common_field(unsigned char, preempt_count); __common_field(int, pid); + __common_field(unsigned char, migrate_disable); + __common_field(unsigned char, preempt_lazy_count); return ret; } Index: linux-5.6.17-rt10/kernel/trace/trace_output.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/trace/trace_output.c +++ linux-5.6.17-rt10/kernel/trace/trace_output.c @@ -441,6 +441,7 @@ int trace_print_lat_fmt(struct trace_seq { char hardsoft_irq; char need_resched; + char need_resched_lazy; char irqs_off; int hardirq; int softirq; @@ -471,6 +472,9 @@ int trace_print_lat_fmt(struct trace_seq break; } + need_resched_lazy = + (entry->flags & TRACE_FLAG_NEED_RESCHED_LAZY) ? 'L' : '.'; + hardsoft_irq = (nmi && hardirq) ? 'Z' : nmi ? 'z' : @@ -479,14 +483,25 @@ int trace_print_lat_fmt(struct trace_seq softirq ? 's' : '.' ; - trace_seq_printf(s, "%c%c%c", - irqs_off, need_resched, hardsoft_irq); + trace_seq_printf(s, "%c%c%c%c", + irqs_off, need_resched, need_resched_lazy, + hardsoft_irq); if (entry->preempt_count) trace_seq_printf(s, "%x", entry->preempt_count); else trace_seq_putc(s, '.'); + if (entry->preempt_lazy_count) + trace_seq_printf(s, "%x", entry->preempt_lazy_count); + else + trace_seq_putc(s, '.'); + + if (entry->migrate_disable) + trace_seq_printf(s, "%x", entry->migrate_disable); + else + trace_seq_putc(s, '.'); + return !trace_seq_has_overflowed(s); } Index: linux-5.6.17-rt10/kernel/trace/trace_uprobe.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/trace/trace_uprobe.c +++ linux-5.6.17-rt10/kernel/trace/trace_uprobe.c @@ -1333,8 +1333,15 @@ static void __uprobe_perf_func(struct tr int size, esize; int rctx; - if (bpf_prog_array_valid(call) && !trace_call_bpf(call, regs)) - return; + if (bpf_prog_array_valid(call)) { + u32 ret; + + preempt_disable(); + ret = trace_call_bpf(call, regs); + preempt_enable(); + if (!ret) + return; + } esize = SIZEOF_TRACE_ENTRY(is_ret_probe(tu)); Index: linux-5.6.17-rt10/kernel/workqueue.c =================================================================== --- linux-5.6.17-rt10.orig/kernel/workqueue.c +++ linux-5.6.17-rt10/kernel/workqueue.c @@ -145,7 +145,7 @@ enum { /* struct worker is defined in workqueue_internal.h */ struct worker_pool { - spinlock_t lock; /* the pool lock */ + raw_spinlock_t lock; /* the pool lock */ int cpu; /* I: the associated cpu */ int node; /* I: the associated node ID */ int id; /* I: pool ID */ @@ -300,8 +300,9 @@ static struct workqueue_attrs *wq_update static DEFINE_MUTEX(wq_pool_mutex); /* protects pools and workqueues list */ static DEFINE_MUTEX(wq_pool_attach_mutex); /* protects worker attach/detach */ -static DEFINE_SPINLOCK(wq_mayday_lock); /* protects wq->maydays list */ -static DECLARE_WAIT_QUEUE_HEAD(wq_manager_wait); /* wait for manager to go away */ +static DEFINE_RAW_SPINLOCK(wq_mayday_lock); /* protects wq->maydays list */ +/* wait for manager to go away */ +static struct rcuwait manager_wait = __RCUWAIT_INITIALIZER(manager_wait); static LIST_HEAD(workqueues); /* PR: list of all workqueues */ static bool workqueue_freezing; /* PL: have wqs started freezing? */ @@ -826,7 +827,7 @@ static struct worker *first_idle_worker( * Wake up the first idle worker of @pool. * * CONTEXT: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). */ static void wake_up_worker(struct worker_pool *pool) { @@ -879,7 +880,7 @@ void wq_worker_sleeping(struct task_stru return; worker->sleeping = 1; - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); /* * The counterpart of the following dec_and_test, implied mb, @@ -898,7 +899,7 @@ void wq_worker_sleeping(struct task_stru if (next) wake_up_process(next->task); } - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); } /** @@ -909,7 +910,7 @@ void wq_worker_sleeping(struct task_stru * the scheduler to get a worker's last known identity. * * CONTEXT: - * spin_lock_irq(rq->lock) + * raw_spin_lock_irq(rq->lock) * * This function is called during schedule() when a kworker is going * to sleep. It's used by psi to identify aggregation workers during @@ -940,7 +941,7 @@ work_func_t wq_worker_last_func(struct t * Set @flags in @worker->flags and adjust nr_running accordingly. * * CONTEXT: - * spin_lock_irq(pool->lock) + * raw_spin_lock_irq(pool->lock) */ static inline void worker_set_flags(struct worker *worker, unsigned int flags) { @@ -965,7 +966,7 @@ static inline void worker_set_flags(stru * Clear @flags in @worker->flags and adjust nr_running accordingly. * * CONTEXT: - * spin_lock_irq(pool->lock) + * raw_spin_lock_irq(pool->lock) */ static inline void worker_clr_flags(struct worker *worker, unsigned int flags) { @@ -1013,7 +1014,7 @@ static inline void worker_clr_flags(stru * actually occurs, it should be easy to locate the culprit work function. * * CONTEXT: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). * * Return: * Pointer to worker which is executing @work if found, %NULL @@ -1048,7 +1049,7 @@ static struct worker *find_worker_execut * nested inside outer list_for_each_entry_safe(). * * CONTEXT: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). */ static void move_linked_works(struct work_struct *work, struct list_head *head, struct work_struct **nextp) @@ -1126,9 +1127,9 @@ static void put_pwq_unlocked(struct pool * As both pwqs and pools are RCU protected, the * following lock operations are safe. */ - spin_lock_irq(&pwq->pool->lock); + raw_spin_lock_irq(&pwq->pool->lock); put_pwq(pwq); - spin_unlock_irq(&pwq->pool->lock); + raw_spin_unlock_irq(&pwq->pool->lock); } } @@ -1161,7 +1162,7 @@ static void pwq_activate_first_delayed(s * decrement nr_in_flight of its pwq and handle workqueue flushing. * * CONTEXT: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). */ static void pwq_dec_nr_in_flight(struct pool_workqueue *pwq, int color) { @@ -1260,7 +1261,7 @@ static int try_to_grab_pending(struct wo if (!pool) goto fail; - spin_lock(&pool->lock); + raw_spin_lock(&pool->lock); /* * work->data is guaranteed to point to pwq only while the work * item is queued on pwq->wq, and both updating work->data to point @@ -1289,11 +1290,11 @@ static int try_to_grab_pending(struct wo /* work->data points to pwq iff queued, point to pool */ set_work_pool_and_keep_pending(work, pool->id); - spin_unlock(&pool->lock); + raw_spin_unlock(&pool->lock); rcu_read_unlock(); return 1; } - spin_unlock(&pool->lock); + raw_spin_unlock(&pool->lock); fail: rcu_read_unlock(); local_irq_restore(*flags); @@ -1314,7 +1315,7 @@ fail: * work_struct flags. * * CONTEXT: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). */ static void insert_work(struct pool_workqueue *pwq, struct work_struct *work, struct list_head *head, unsigned int extra_flags) @@ -1431,7 +1432,7 @@ retry: if (last_pool && last_pool != pwq->pool) { struct worker *worker; - spin_lock(&last_pool->lock); + raw_spin_lock(&last_pool->lock); worker = find_worker_executing_work(last_pool, work); @@ -1439,11 +1440,11 @@ retry: pwq = worker->current_pwq; } else { /* meh... not running there, queue here */ - spin_unlock(&last_pool->lock); - spin_lock(&pwq->pool->lock); + raw_spin_unlock(&last_pool->lock); + raw_spin_lock(&pwq->pool->lock); } } else { - spin_lock(&pwq->pool->lock); + raw_spin_lock(&pwq->pool->lock); } /* @@ -1456,7 +1457,7 @@ retry: */ if (unlikely(!pwq->refcnt)) { if (wq->flags & WQ_UNBOUND) { - spin_unlock(&pwq->pool->lock); + raw_spin_unlock(&pwq->pool->lock); cpu_relax(); goto retry; } @@ -1488,7 +1489,7 @@ retry: insert_work(pwq, work, worklist, work_flags); out: - spin_unlock(&pwq->pool->lock); + raw_spin_unlock(&pwq->pool->lock); rcu_read_unlock(); } @@ -1757,7 +1758,7 @@ EXPORT_SYMBOL(queue_rcu_work); * necessary. * * LOCKING: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). */ static void worker_enter_idle(struct worker *worker) { @@ -1797,7 +1798,7 @@ static void worker_enter_idle(struct wor * @worker is leaving idle state. Update stats. * * LOCKING: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). */ static void worker_leave_idle(struct worker *worker) { @@ -1935,11 +1936,11 @@ static struct worker *create_worker(stru worker_attach_to_pool(worker, pool); /* start the newly created worker */ - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); worker->pool->nr_workers++; worker_enter_idle(worker); wake_up_process(worker->task); - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); return worker; @@ -1958,7 +1959,7 @@ fail: * be idle. * * CONTEXT: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). */ static void destroy_worker(struct worker *worker) { @@ -1984,7 +1985,7 @@ static void idle_worker_timeout(struct t { struct worker_pool *pool = from_timer(pool, t, idle_timer); - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); while (too_many_workers(pool)) { struct worker *worker; @@ -2002,7 +2003,7 @@ static void idle_worker_timeout(struct t destroy_worker(worker); } - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); } static void send_mayday(struct work_struct *work) @@ -2033,8 +2034,8 @@ static void pool_mayday_timeout(struct t struct worker_pool *pool = from_timer(pool, t, mayday_timer); struct work_struct *work; - spin_lock_irq(&pool->lock); - spin_lock(&wq_mayday_lock); /* for wq->maydays */ + raw_spin_lock_irq(&pool->lock); + raw_spin_lock(&wq_mayday_lock); /* for wq->maydays */ if (need_to_create_worker(pool)) { /* @@ -2047,8 +2048,8 @@ static void pool_mayday_timeout(struct t send_mayday(work); } - spin_unlock(&wq_mayday_lock); - spin_unlock_irq(&pool->lock); + raw_spin_unlock(&wq_mayday_lock); + raw_spin_unlock_irq(&pool->lock); mod_timer(&pool->mayday_timer, jiffies + MAYDAY_INTERVAL); } @@ -2067,7 +2068,7 @@ static void pool_mayday_timeout(struct t * may_start_working() %true. * * LOCKING: - * spin_lock_irq(pool->lock) which may be released and regrabbed + * raw_spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. Does GFP_KERNEL allocations. Called only from * manager. */ @@ -2076,7 +2077,7 @@ __releases(&pool->lock) __acquires(&pool->lock) { restart: - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); /* if we don't make progress in MAYDAY_INITIAL_TIMEOUT, call for help */ mod_timer(&pool->mayday_timer, jiffies + MAYDAY_INITIAL_TIMEOUT); @@ -2092,7 +2093,7 @@ restart: } del_timer_sync(&pool->mayday_timer); - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); /* * This is necessary even after a new worker was just successfully * created as @pool->lock was dropped and the new worker might have @@ -2115,7 +2116,7 @@ restart: * and may_start_working() is true. * * CONTEXT: - * spin_lock_irq(pool->lock) which may be released and regrabbed + * raw_spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. Does GFP_KERNEL allocations. * * Return: @@ -2138,7 +2139,7 @@ static bool manage_workers(struct worker pool->manager = NULL; pool->flags &= ~POOL_MANAGER_ACTIVE; - wake_up(&wq_manager_wait); + rcuwait_wake_up(&manager_wait); return true; } @@ -2154,7 +2155,7 @@ static bool manage_workers(struct worker * call this function to process a work. * * CONTEXT: - * spin_lock_irq(pool->lock) which is released and regrabbed. + * raw_spin_lock_irq(pool->lock) which is released and regrabbed. */ static void process_one_work(struct worker *worker, struct work_struct *work) __releases(&pool->lock) @@ -2236,7 +2237,7 @@ __acquires(&pool->lock) */ set_work_pool_and_clear_pending(work, pool->id); - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); lock_map_acquire(&pwq->wq->lockdep_map); lock_map_acquire(&lockdep_map); @@ -2291,7 +2292,7 @@ __acquires(&pool->lock) */ cond_resched(); - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); /* clear cpu intensive status */ if (unlikely(cpu_intensive)) @@ -2317,7 +2318,7 @@ __acquires(&pool->lock) * fetches a work from the top and executes it. * * CONTEXT: - * spin_lock_irq(pool->lock) which may be released and regrabbed + * raw_spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. */ static void process_scheduled_works(struct worker *worker) @@ -2359,11 +2360,11 @@ static int worker_thread(void *__worker) /* tell the scheduler that this is a workqueue worker */ set_pf_worker(true); woke_up: - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); /* am I supposed to die? */ if (unlikely(worker->flags & WORKER_DIE)) { - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); WARN_ON_ONCE(!list_empty(&worker->entry)); set_pf_worker(false); @@ -2429,7 +2430,7 @@ sleep: */ worker_enter_idle(worker); __set_current_state(TASK_IDLE); - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); schedule(); goto woke_up; } @@ -2483,7 +2484,7 @@ repeat: should_stop = kthread_should_stop(); /* see whether any pwq is asking for help */ - spin_lock_irq(&wq_mayday_lock); + raw_spin_lock_irq(&wq_mayday_lock); while (!list_empty(&wq->maydays)) { struct pool_workqueue *pwq = list_first_entry(&wq->maydays, @@ -2495,11 +2496,11 @@ repeat: __set_current_state(TASK_RUNNING); list_del_init(&pwq->mayday_node); - spin_unlock_irq(&wq_mayday_lock); + raw_spin_unlock_irq(&wq_mayday_lock); worker_attach_to_pool(rescuer, pool); - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); /* * Slurp in all works issued via this workqueue and @@ -2528,7 +2529,7 @@ repeat: * incur MAYDAY_INTERVAL delay inbetween. */ if (need_to_create_worker(pool)) { - spin_lock(&wq_mayday_lock); + raw_spin_lock(&wq_mayday_lock); /* * Queue iff we aren't racing destruction * and somebody else hasn't queued it already. @@ -2537,7 +2538,7 @@ repeat: get_pwq(pwq); list_add_tail(&pwq->mayday_node, &wq->maydays); } - spin_unlock(&wq_mayday_lock); + raw_spin_unlock(&wq_mayday_lock); } } @@ -2555,14 +2556,14 @@ repeat: if (need_more_worker(pool)) wake_up_worker(pool); - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); worker_detach_from_pool(rescuer); - spin_lock_irq(&wq_mayday_lock); + raw_spin_lock_irq(&wq_mayday_lock); } - spin_unlock_irq(&wq_mayday_lock); + raw_spin_unlock_irq(&wq_mayday_lock); if (should_stop) { __set_current_state(TASK_RUNNING); @@ -2642,7 +2643,7 @@ static void wq_barrier_func(struct work_ * underneath us, so we can't reliably determine pwq from @target. * * CONTEXT: - * spin_lock_irq(pool->lock). + * raw_spin_lock_irq(pool->lock). */ static void insert_wq_barrier(struct pool_workqueue *pwq, struct wq_barrier *barr, @@ -2729,7 +2730,7 @@ static bool flush_workqueue_prep_pwqs(st for_each_pwq(pwq, wq) { struct worker_pool *pool = pwq->pool; - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); if (flush_color >= 0) { WARN_ON_ONCE(pwq->flush_color != -1); @@ -2746,7 +2747,7 @@ static bool flush_workqueue_prep_pwqs(st pwq->work_color = work_color; } - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); } if (flush_color >= 0 && atomic_dec_and_test(&wq->nr_pwqs_to_flush)) @@ -2946,9 +2947,9 @@ reflush: for_each_pwq(pwq, wq) { bool drained; - spin_lock_irq(&pwq->pool->lock); + raw_spin_lock_irq(&pwq->pool->lock); drained = !pwq->nr_active && list_empty(&pwq->delayed_works); - spin_unlock_irq(&pwq->pool->lock); + raw_spin_unlock_irq(&pwq->pool->lock); if (drained) continue; @@ -2984,7 +2985,7 @@ static bool start_flush_work(struct work return false; } - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); /* see the comment in try_to_grab_pending() with the same code */ pwq = get_work_pwq(work); if (pwq) { @@ -3000,7 +3001,7 @@ static bool start_flush_work(struct work check_flush_dependency(pwq->wq, work); insert_wq_barrier(pwq, barr, work, worker); - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); /* * Force a lock recursion deadlock when using flush_work() inside a @@ -3019,7 +3020,7 @@ static bool start_flush_work(struct work rcu_read_unlock(); return true; already_gone: - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); rcu_read_unlock(); return false; } @@ -3412,7 +3413,7 @@ static bool wqattrs_equal(const struct w */ static int init_worker_pool(struct worker_pool *pool) { - spin_lock_init(&pool->lock); + raw_spin_lock_init(&pool->lock); pool->id = -1; pool->cpu = -1; pool->node = NUMA_NO_NODE; @@ -3502,6 +3503,18 @@ static void rcu_free_pool(struct rcu_hea kfree(pool); } +/* This returns with the lock held on success (pool manager is inactive). */ +static bool wq_manager_inactive(struct worker_pool *pool) +{ + raw_spin_lock_irq(&pool->lock); + + if (pool->flags & POOL_MANAGER_ACTIVE) { + raw_spin_unlock_irq(&pool->lock); + return false; + } + return true; +} + /** * put_unbound_pool - put a worker_pool * @pool: worker_pool to put @@ -3537,16 +3550,16 @@ static void put_unbound_pool(struct work * Become the manager and destroy all workers. This prevents * @pool's workers from blocking on attach_mutex. We're the last * manager and @pool gets freed with the flag set. + * Because of how wq_manager_inactive() works, we will hold the + * spinlock after a successful wait. */ - spin_lock_irq(&pool->lock); - wait_event_lock_irq(wq_manager_wait, - !(pool->flags & POOL_MANAGER_ACTIVE), pool->lock); + rcuwait_wait_event(&manager_wait, wq_manager_inactive(pool)); pool->flags |= POOL_MANAGER_ACTIVE; while ((worker = first_idle_worker(pool))) destroy_worker(worker); WARN_ON(pool->nr_workers || pool->nr_idle); - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); mutex_lock(&wq_pool_attach_mutex); if (!list_empty(&pool->workers)) @@ -3702,7 +3715,7 @@ static void pwq_adjust_max_active(struct return; /* this function can be called during early boot w/ irq disabled */ - spin_lock_irqsave(&pwq->pool->lock, flags); + raw_spin_lock_irqsave(&pwq->pool->lock, flags); /* * During [un]freezing, the caller is responsible for ensuring that @@ -3725,7 +3738,7 @@ static void pwq_adjust_max_active(struct pwq->max_active = 0; } - spin_unlock_irqrestore(&pwq->pool->lock, flags); + raw_spin_unlock_irqrestore(&pwq->pool->lock, flags); } /* initialize newly alloced @pwq which is associated with @wq and @pool */ @@ -4127,9 +4140,9 @@ static void wq_update_unbound_numa(struc use_dfl_pwq: mutex_lock(&wq->mutex); - spin_lock_irq(&wq->dfl_pwq->pool->lock); + raw_spin_lock_irq(&wq->dfl_pwq->pool->lock); get_pwq(wq->dfl_pwq); - spin_unlock_irq(&wq->dfl_pwq->pool->lock); + raw_spin_unlock_irq(&wq->dfl_pwq->pool->lock); old_pwq = numa_pwq_tbl_install(wq, node, wq->dfl_pwq); out_unlock: mutex_unlock(&wq->mutex); @@ -4358,9 +4371,9 @@ void destroy_workqueue(struct workqueue_ struct worker *rescuer = wq->rescuer; /* this prevents new queueing */ - spin_lock_irq(&wq_mayday_lock); + raw_spin_lock_irq(&wq_mayday_lock); wq->rescuer = NULL; - spin_unlock_irq(&wq_mayday_lock); + raw_spin_unlock_irq(&wq_mayday_lock); /* rescuer will empty maydays list before exiting */ kthread_stop(rescuer->task); @@ -4374,18 +4387,18 @@ void destroy_workqueue(struct workqueue_ mutex_lock(&wq_pool_mutex); mutex_lock(&wq->mutex); for_each_pwq(pwq, wq) { - spin_lock_irq(&pwq->pool->lock); + raw_spin_lock_irq(&pwq->pool->lock); if (WARN_ON(pwq_busy(pwq))) { pr_warn("%s: %s has the following busy pwq\n", __func__, wq->name); show_pwq(pwq); - spin_unlock_irq(&pwq->pool->lock); + raw_spin_unlock_irq(&pwq->pool->lock); mutex_unlock(&wq->mutex); mutex_unlock(&wq_pool_mutex); show_workqueue_state(); return; } - spin_unlock_irq(&pwq->pool->lock); + raw_spin_unlock_irq(&pwq->pool->lock); } mutex_unlock(&wq->mutex); mutex_unlock(&wq_pool_mutex); @@ -4556,10 +4569,10 @@ unsigned int work_busy(struct work_struc rcu_read_lock(); pool = get_work_pool(work); if (pool) { - spin_lock_irqsave(&pool->lock, flags); + raw_spin_lock_irqsave(&pool->lock, flags); if (find_worker_executing_work(pool, work)) ret |= WORK_BUSY_RUNNING; - spin_unlock_irqrestore(&pool->lock, flags); + raw_spin_unlock_irqrestore(&pool->lock, flags); } rcu_read_unlock(); @@ -4766,10 +4779,10 @@ void show_workqueue_state(void) pr_info("workqueue %s: flags=0x%x\n", wq->name, wq->flags); for_each_pwq(pwq, wq) { - spin_lock_irqsave(&pwq->pool->lock, flags); + raw_spin_lock_irqsave(&pwq->pool->lock, flags); if (pwq->nr_active || !list_empty(&pwq->delayed_works)) show_pwq(pwq); - spin_unlock_irqrestore(&pwq->pool->lock, flags); + raw_spin_unlock_irqrestore(&pwq->pool->lock, flags); /* * We could be printing a lot from atomic context, e.g. * sysrq-t -> show_workqueue_state(). Avoid triggering @@ -4783,7 +4796,7 @@ void show_workqueue_state(void) struct worker *worker; bool first = true; - spin_lock_irqsave(&pool->lock, flags); + raw_spin_lock_irqsave(&pool->lock, flags); if (pool->nr_workers == pool->nr_idle) goto next_pool; @@ -4802,7 +4815,7 @@ void show_workqueue_state(void) } pr_cont("\n"); next_pool: - spin_unlock_irqrestore(&pool->lock, flags); + raw_spin_unlock_irqrestore(&pool->lock, flags); /* * We could be printing a lot from atomic context, e.g. * sysrq-t -> show_workqueue_state(). Avoid triggering @@ -4832,7 +4845,7 @@ void wq_worker_comm(char *buf, size_t si struct worker_pool *pool = worker->pool; if (pool) { - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); /* * ->desc tracks information (wq name or * set_worker_desc()) for the latest execution. If @@ -4846,7 +4859,7 @@ void wq_worker_comm(char *buf, size_t si scnprintf(buf + off, size - off, "-%s", worker->desc); } - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); } } @@ -4877,7 +4890,7 @@ static void unbind_workers(int cpu) for_each_cpu_worker_pool(pool, cpu) { mutex_lock(&wq_pool_attach_mutex); - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); /* * We've blocked all attach/detach operations. Make all workers @@ -4891,7 +4904,7 @@ static void unbind_workers(int cpu) pool->flags |= POOL_DISASSOCIATED; - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); mutex_unlock(&wq_pool_attach_mutex); /* @@ -4917,9 +4930,9 @@ static void unbind_workers(int cpu) * worker blocking could lead to lengthy stalls. Kick off * unbound chain execution of currently pending work items. */ - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); wake_up_worker(pool); - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); } } @@ -4946,7 +4959,7 @@ static void rebind_workers(struct worker WARN_ON_ONCE(set_cpus_allowed_ptr(worker->task, pool->attrs->cpumask) < 0); - spin_lock_irq(&pool->lock); + raw_spin_lock_irq(&pool->lock); pool->flags &= ~POOL_DISASSOCIATED; @@ -4985,7 +4998,7 @@ static void rebind_workers(struct worker WRITE_ONCE(worker->flags, worker_flags); } - spin_unlock_irq(&pool->lock); + raw_spin_unlock_irq(&pool->lock); } /** Index: linux-5.6.17-rt10/lib/Kconfig.debug =================================================================== --- linux-5.6.17-rt10.orig/lib/Kconfig.debug +++ linux-5.6.17-rt10/lib/Kconfig.debug @@ -61,6 +61,23 @@ config CONSOLE_LOGLEVEL_QUIET will be used as the loglevel. IOW passing "quiet" will be the equivalent of passing "loglevel=" +config CONSOLE_LOGLEVEL_EMERGENCY + int "Emergency console loglevel (1-15)" + range 1 15 + default "5" + help + The loglevel to determine if a console message is an emergency + message. + + If supported by the console driver, emergency messages will be + flushed to the console immediately. This can cause significant system + latencies so the value should be set such that only significant + messages are classified as emergency messages. + + Setting a default here is equivalent to passing in + emergency_loglevel= in the kernel bootargs. emergency_loglevel= + continues to override whatever value is specified here as well. + config MESSAGE_LOGLEVEL_DEFAULT int "Default message log level (1-7)" range 1 7 @@ -1205,7 +1222,7 @@ config DEBUG_ATOMIC_SLEEP config DEBUG_LOCKING_API_SELFTESTS bool "Locking API boot-time self-tests" - depends on DEBUG_KERNEL + depends on DEBUG_KERNEL && !PREEMPT_RT help Say Y here if you want the kernel to run a short self-test during bootup. The self-test checks whether common types of locking bugs Index: linux-5.6.17-rt10/lib/Makefile =================================================================== --- linux-5.6.17-rt10.orig/lib/Makefile +++ linux-5.6.17-rt10/lib/Makefile @@ -27,7 +27,7 @@ endif lib-y := ctype.o string.o vsprintf.o cmdline.o \ rbtree.o radix-tree.o timerqueue.o xarray.o \ - idr.o extable.o sha1.o irq_regs.o argv_split.o \ + idr.o extable.o sha1.o irq_regs.o argv_split.o printk_ringbuffer.o \ flex_proportions.o ratelimit.o show_mem.o \ is_single_threaded.o plist.o decompress.o kobject_uevent.o \ earlycpio.o seq_buf.o siphash.o dec_and_lock.o \ Index: linux-5.6.17-rt10/lib/bust_spinlocks.c =================================================================== --- linux-5.6.17-rt10.orig/lib/bust_spinlocks.c +++ linux-5.6.17-rt10/lib/bust_spinlocks.c @@ -26,7 +26,6 @@ void bust_spinlocks(int yes) unblank_screen(); #endif console_unblank(); - if (--oops_in_progress == 0) - wake_up_klogd(); + --oops_in_progress; } } Index: linux-5.6.17-rt10/lib/debugobjects.c =================================================================== --- linux-5.6.17-rt10.orig/lib/debugobjects.c +++ linux-5.6.17-rt10/lib/debugobjects.c @@ -537,7 +537,10 @@ __debug_object_init(void *addr, struct d struct debug_obj *obj; unsigned long flags; - fill_pool(); +#ifdef CONFIG_PREEMPT_RT + if (preempt_count() == 0 && !irqs_disabled()) +#endif + fill_pool(); db = get_bucket((unsigned long) addr); Index: linux-5.6.17-rt10/lib/irq_poll.c =================================================================== --- linux-5.6.17-rt10.orig/lib/irq_poll.c +++ linux-5.6.17-rt10/lib/irq_poll.c @@ -37,6 +37,7 @@ void irq_poll_sched(struct irq_poll *iop list_add_tail(&iop->list, this_cpu_ptr(&blk_cpu_iopoll)); raise_softirq_irqoff(IRQ_POLL_SOFTIRQ); local_irq_restore(flags); + preempt_check_resched_rt(); } EXPORT_SYMBOL(irq_poll_sched); @@ -72,6 +73,7 @@ void irq_poll_complete(struct irq_poll * local_irq_save(flags); __irq_poll_complete(iop); local_irq_restore(flags); + preempt_check_resched_rt(); } EXPORT_SYMBOL(irq_poll_complete); @@ -96,6 +98,7 @@ static void __latent_entropy irq_poll_so } local_irq_enable(); + preempt_check_resched_rt(); /* Even though interrupts have been re-enabled, this * access is safe because interrupts can only add new @@ -133,6 +136,7 @@ static void __latent_entropy irq_poll_so __raise_softirq_irqoff(IRQ_POLL_SOFTIRQ); local_irq_enable(); + preempt_check_resched_rt(); } /** @@ -196,6 +200,7 @@ static int irq_poll_cpu_dead(unsigned in this_cpu_ptr(&blk_cpu_iopoll)); __raise_softirq_irqoff(IRQ_POLL_SOFTIRQ); local_irq_enable(); + preempt_check_resched_rt(); return 0; } Index: linux-5.6.17-rt10/lib/locking-selftest.c =================================================================== --- linux-5.6.17-rt10.orig/lib/locking-selftest.c +++ linux-5.6.17-rt10/lib/locking-selftest.c @@ -742,6 +742,8 @@ GENERATE_TESTCASE(init_held_rtmutex); #include "locking-selftest-spin-hardirq.h" GENERATE_PERMUTATIONS_2_EVENTS(irqsafe1_hard_spin) +#ifndef CONFIG_PREEMPT_RT + #include "locking-selftest-rlock-hardirq.h" GENERATE_PERMUTATIONS_2_EVENTS(irqsafe1_hard_rlock) @@ -757,9 +759,12 @@ GENERATE_PERMUTATIONS_2_EVENTS(irqsafe1_ #include "locking-selftest-wlock-softirq.h" GENERATE_PERMUTATIONS_2_EVENTS(irqsafe1_soft_wlock) +#endif + #undef E1 #undef E2 +#ifndef CONFIG_PREEMPT_RT /* * Enabling hardirqs with a softirq-safe lock held: */ @@ -792,6 +797,8 @@ GENERATE_PERMUTATIONS_2_EVENTS(irqsafe2A #undef E1 #undef E2 +#endif + /* * Enabling irqs with an irq-safe lock held: */ @@ -815,6 +822,8 @@ GENERATE_PERMUTATIONS_2_EVENTS(irqsafe2A #include "locking-selftest-spin-hardirq.h" GENERATE_PERMUTATIONS_2_EVENTS(irqsafe2B_hard_spin) +#ifndef CONFIG_PREEMPT_RT + #include "locking-selftest-rlock-hardirq.h" GENERATE_PERMUTATIONS_2_EVENTS(irqsafe2B_hard_rlock) @@ -830,6 +839,8 @@ GENERATE_PERMUTATIONS_2_EVENTS(irqsafe2B #include "locking-selftest-wlock-softirq.h" GENERATE_PERMUTATIONS_2_EVENTS(irqsafe2B_soft_wlock) +#endif + #undef E1 #undef E2 @@ -861,6 +872,8 @@ GENERATE_PERMUTATIONS_2_EVENTS(irqsafe2B #include "locking-selftest-spin-hardirq.h" GENERATE_PERMUTATIONS_3_EVENTS(irqsafe3_hard_spin) +#ifndef CONFIG_PREEMPT_RT + #include "locking-selftest-rlock-hardirq.h" GENERATE_PERMUTATIONS_3_EVENTS(irqsafe3_hard_rlock) @@ -876,6 +889,8 @@ GENERATE_PERMUTATIONS_3_EVENTS(irqsafe3_ #include "locking-selftest-wlock-softirq.h" GENERATE_PERMUTATIONS_3_EVENTS(irqsafe3_soft_wlock) +#endif + #undef E1 #undef E2 #undef E3 @@ -909,6 +924,8 @@ GENERATE_PERMUTATIONS_3_EVENTS(irqsafe3_ #include "locking-selftest-spin-hardirq.h" GENERATE_PERMUTATIONS_3_EVENTS(irqsafe4_hard_spin) +#ifndef CONFIG_PREEMPT_RT + #include "locking-selftest-rlock-hardirq.h" GENERATE_PERMUTATIONS_3_EVENTS(irqsafe4_hard_rlock) @@ -924,10 +941,14 @@ GENERATE_PERMUTATIONS_3_EVENTS(irqsafe4_ #include "locking-selftest-wlock-softirq.h" GENERATE_PERMUTATIONS_3_EVENTS(irqsafe4_soft_wlock) +#endif + #undef E1 #undef E2 #undef E3 +#ifndef CONFIG_PREEMPT_RT + /* * read-lock / write-lock irq inversion. * @@ -990,6 +1011,10 @@ GENERATE_PERMUTATIONS_3_EVENTS(irq_inver #undef E2 #undef E3 +#endif + +#ifndef CONFIG_PREEMPT_RT + /* * read-lock / write-lock recursion that is actually safe. */ @@ -1028,6 +1053,8 @@ GENERATE_PERMUTATIONS_3_EVENTS(irq_read_ #undef E2 #undef E3 +#endif + /* * read-lock / write-lock recursion that is unsafe. */ @@ -2058,6 +2085,7 @@ void locking_selftest(void) printk(" --------------------------------------------------------------------------\n"); +#ifndef CONFIG_PREEMPT_RT /* * irq-context testcases: */ @@ -2070,6 +2098,28 @@ void locking_selftest(void) DO_TESTCASE_6x2("irq read-recursion", irq_read_recursion); // DO_TESTCASE_6x2B("irq read-recursion #2", irq_read_recursion2); +#else + /* On -rt, we only do hardirq context test for raw spinlock */ + DO_TESTCASE_1B("hard-irqs-on + irq-safe-A", irqsafe1_hard_spin, 12); + DO_TESTCASE_1B("hard-irqs-on + irq-safe-A", irqsafe1_hard_spin, 21); + + DO_TESTCASE_1B("hard-safe-A + irqs-on", irqsafe2B_hard_spin, 12); + DO_TESTCASE_1B("hard-safe-A + irqs-on", irqsafe2B_hard_spin, 21); + + DO_TESTCASE_1B("hard-safe-A + unsafe-B #1", irqsafe3_hard_spin, 123); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #1", irqsafe3_hard_spin, 132); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #1", irqsafe3_hard_spin, 213); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #1", irqsafe3_hard_spin, 231); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #1", irqsafe3_hard_spin, 312); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #1", irqsafe3_hard_spin, 321); + + DO_TESTCASE_1B("hard-safe-A + unsafe-B #2", irqsafe4_hard_spin, 123); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #2", irqsafe4_hard_spin, 132); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #2", irqsafe4_hard_spin, 213); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #2", irqsafe4_hard_spin, 231); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #2", irqsafe4_hard_spin, 312); + DO_TESTCASE_1B("hard-safe-A + unsafe-B #2", irqsafe4_hard_spin, 321); +#endif ww_tests(); Index: linux-5.6.17-rt10/lib/nmi_backtrace.c =================================================================== --- linux-5.6.17-rt10.orig/lib/nmi_backtrace.c +++ linux-5.6.17-rt10/lib/nmi_backtrace.c @@ -75,12 +75,6 @@ void nmi_trigger_cpumask_backtrace(const touch_softlockup_watchdog(); } - /* - * Force flush any remote buffers that might be stuck in IRQ context - * and therefore could not run their irq_work. - */ - printk_safe_flush(); - clear_bit_unlock(0, &backtrace_flag); put_cpu(); } Index: linux-5.6.17-rt10/lib/printk_ringbuffer.c =================================================================== --- /dev/null +++ linux-5.6.17-rt10/lib/printk_ringbuffer.c @@ -0,0 +1,589 @@ +// SPDX-License-Identifier: GPL-2.0 +#include +#include +#include +#include +#include + +#define PRB_SIZE(rb) (1 << rb->size_bits) +#define PRB_SIZE_BITMASK(rb) (PRB_SIZE(rb) - 1) +#define PRB_INDEX(rb, lpos) (lpos & PRB_SIZE_BITMASK(rb)) +#define PRB_WRAPS(rb, lpos) (lpos >> rb->size_bits) +#define PRB_WRAP_LPOS(rb, lpos, xtra) \ + ((PRB_WRAPS(rb, lpos) + xtra) << rb->size_bits) +#define PRB_DATA_SIZE(e) (e->size - sizeof(struct prb_entry)) +#define PRB_DATA_ALIGN sizeof(long) + +static bool __prb_trylock(struct prb_cpulock *cpu_lock, + unsigned int *cpu_store) +{ + unsigned long *flags; + unsigned int cpu; + + cpu = get_cpu(); + + *cpu_store = atomic_read(&cpu_lock->owner); + /* memory barrier to ensure the current lock owner is visible */ + smp_rmb(); + if (*cpu_store == -1) { + flags = per_cpu_ptr(cpu_lock->irqflags, cpu); + local_irq_save(*flags); + if (atomic_try_cmpxchg_acquire(&cpu_lock->owner, + cpu_store, cpu)) { + return true; + } + local_irq_restore(*flags); + } else if (*cpu_store == cpu) { + return true; + } + + put_cpu(); + return false; +} + +/* + * prb_lock: Perform a processor-reentrant spin lock. + * @cpu_lock: A pointer to the lock object. + * @cpu_store: A "flags" pointer to store lock status information. + * + * If no processor has the lock, the calling processor takes the lock and + * becomes the owner. If the calling processor is already the owner of the + * lock, this function succeeds immediately. If lock is locked by another + * processor, this function spins until the calling processor becomes the + * owner. + * + * It is safe to call this function from any context and state. + */ +void prb_lock(struct prb_cpulock *cpu_lock, unsigned int *cpu_store) +{ + for (;;) { + if (__prb_trylock(cpu_lock, cpu_store)) + break; + cpu_relax(); + } +} + +/* + * prb_unlock: Perform a processor-reentrant spin unlock. + * @cpu_lock: A pointer to the lock object. + * @cpu_store: A "flags" object storing lock status information. + * + * Release the lock. The calling processor must be the owner of the lock. + * + * It is safe to call this function from any context and state. + */ +void prb_unlock(struct prb_cpulock *cpu_lock, unsigned int cpu_store) +{ + unsigned long *flags; + unsigned int cpu; + + cpu = atomic_read(&cpu_lock->owner); + atomic_set_release(&cpu_lock->owner, cpu_store); + + if (cpu_store == -1) { + flags = per_cpu_ptr(cpu_lock->irqflags, cpu); + local_irq_restore(*flags); + } + + put_cpu(); +} + +static struct prb_entry *to_entry(struct printk_ringbuffer *rb, + unsigned long lpos) +{ + char *buffer = rb->buffer; + buffer += PRB_INDEX(rb, lpos); + return (struct prb_entry *)buffer; +} + +static int calc_next(struct printk_ringbuffer *rb, unsigned long tail, + unsigned long lpos, int size, unsigned long *calced_next) +{ + unsigned long next_lpos; + int ret = 0; +again: + next_lpos = lpos + size; + if (next_lpos - tail > PRB_SIZE(rb)) + return -1; + + if (PRB_WRAPS(rb, lpos) != PRB_WRAPS(rb, next_lpos)) { + lpos = PRB_WRAP_LPOS(rb, next_lpos, 0); + ret |= 1; + goto again; + } + + *calced_next = next_lpos; + return ret; +} + +static bool push_tail(struct printk_ringbuffer *rb, unsigned long tail) +{ + unsigned long new_tail; + struct prb_entry *e; + unsigned long head; + + if (tail != atomic_long_read(&rb->tail)) + return true; + + e = to_entry(rb, tail); + if (e->size != -1) + new_tail = tail + e->size; + else + new_tail = PRB_WRAP_LPOS(rb, tail, 1); + + /* make sure the new tail does not overtake the head */ + head = atomic_long_read(&rb->head); + if (head - new_tail > PRB_SIZE(rb)) + return false; + + atomic_long_cmpxchg(&rb->tail, tail, new_tail); + return true; +} + +/* + * prb_commit: Commit a reserved entry to the ring buffer. + * @h: An entry handle referencing the data entry to commit. + * + * Commit data that has been reserved using prb_reserve(). Once the data + * block has been committed, it can be invalidated at any time. If a writer + * is interested in using the data after committing, the writer should make + * its own copy first or use the prb_iter_ reader functions to access the + * data in the ring buffer. + * + * It is safe to call this function from any context and state. + */ +void prb_commit(struct prb_handle *h) +{ + struct printk_ringbuffer *rb = h->rb; + bool changed = false; + struct prb_entry *e; + unsigned long head; + unsigned long res; + + for (;;) { + if (atomic_read(&rb->ctx) != 1) { + /* the interrupted context will fixup head */ + atomic_dec(&rb->ctx); + break; + } + /* assign sequence numbers before moving head */ + head = atomic_long_read(&rb->head); + res = atomic_long_read(&rb->reserve); + while (head != res) { + e = to_entry(rb, head); + if (e->size == -1) { + head = PRB_WRAP_LPOS(rb, head, 1); + continue; + } + while (atomic_long_read(&rb->lost)) { + atomic_long_dec(&rb->lost); + rb->seq++; + } + e->seq = ++rb->seq; + head += e->size; + changed = true; + } + atomic_long_set_release(&rb->head, res); + + atomic_dec(&rb->ctx); + + if (atomic_long_read(&rb->reserve) == res) + break; + atomic_inc(&rb->ctx); + } + + prb_unlock(rb->cpulock, h->cpu); + + if (changed) { + atomic_long_inc(&rb->wq_counter); + if (wq_has_sleeper(rb->wq)) { +#ifdef CONFIG_IRQ_WORK + irq_work_queue(rb->wq_work); +#else + if (!in_nmi()) + wake_up_interruptible_all(rb->wq); +#endif + } + } +} + +/* + * prb_reserve: Reserve an entry within a ring buffer. + * @h: An entry handle to be setup and reference an entry. + * @rb: A ring buffer to reserve data within. + * @size: The number of bytes to reserve. + * + * Reserve an entry of at least @size bytes to be used by the caller. If + * successful, the data region of the entry belongs to the caller and cannot + * be invalidated by any other task/context. For this reason, the caller + * should call prb_commit() as quickly as possible in order to avoid preventing + * other tasks/contexts from reserving data in the case that the ring buffer + * has wrapped. + * + * It is safe to call this function from any context and state. + * + * Returns a pointer to the reserved entry (and @h is setup to reference that + * entry) or NULL if it was not possible to reserve data. + */ +char *prb_reserve(struct prb_handle *h, struct printk_ringbuffer *rb, + unsigned int size) +{ + unsigned long tail, res1, res2; + int ret; + + if (size == 0) + return NULL; + size += sizeof(struct prb_entry); + size += PRB_DATA_ALIGN - 1; + size &= ~(PRB_DATA_ALIGN - 1); + if (size >= PRB_SIZE(rb)) + return NULL; + + h->rb = rb; + prb_lock(rb->cpulock, &h->cpu); + + atomic_inc(&rb->ctx); + + do { + for (;;) { + tail = atomic_long_read(&rb->tail); + res1 = atomic_long_read(&rb->reserve); + ret = calc_next(rb, tail, res1, size, &res2); + if (ret >= 0) + break; + if (!push_tail(rb, tail)) { + prb_commit(h); + return NULL; + } + } + } while (!atomic_long_try_cmpxchg_acquire(&rb->reserve, &res1, res2)); + + h->entry = to_entry(rb, res1); + + if (ret) { + /* handle wrap */ + h->entry->size = -1; + h->entry = to_entry(rb, PRB_WRAP_LPOS(rb, res2, 0)); + } + + h->entry->size = size; + + return &h->entry->data[0]; +} + +/* + * prb_iter_copy: Copy an iterator. + * @dest: The iterator to copy to. + * @src: The iterator to copy from. + * + * Make a deep copy of an iterator. This is particularly useful for making + * backup copies of an iterator in case a form of rewinding it needed. + * + * It is safe to call this function from any context and state. But + * note that this function is not atomic. Callers should not make copies + * to/from iterators that can be accessed by other tasks/contexts. + */ +void prb_iter_copy(struct prb_iterator *dest, struct prb_iterator *src) +{ + memcpy(dest, src, sizeof(*dest)); +} + +/* + * prb_iter_init: Initialize an iterator for a ring buffer. + * @iter: The iterator to initialize. + * @rb: A ring buffer to that @iter should iterate. + * @seq: The sequence number of the position preceding the first record. + * May be NULL. + * + * Initialize an iterator to be used with a specified ring buffer. If @seq + * is non-NULL, it will be set such that prb_iter_next() will provide a + * sequence value of "@seq + 1" if no records were missed. + * + * It is safe to call this function from any context and state. + */ +void prb_iter_init(struct prb_iterator *iter, struct printk_ringbuffer *rb, + u64 *seq) +{ + memset(iter, 0, sizeof(*iter)); + iter->rb = rb; + iter->lpos = PRB_INIT; + + if (!seq) + return; + + for (;;) { + struct prb_iterator tmp_iter; + int ret; + + prb_iter_copy(&tmp_iter, iter); + + ret = prb_iter_next(&tmp_iter, NULL, 0, seq); + if (ret < 0) + continue; + + if (ret == 0) + *seq = 0; + else + (*seq)--; + break; + } +} + +static bool is_valid(struct printk_ringbuffer *rb, unsigned long lpos) +{ + unsigned long head, tail; + + tail = atomic_long_read(&rb->tail); + head = atomic_long_read(&rb->head); + head -= tail; + lpos -= tail; + + if (lpos >= head) + return false; + return true; +} + +/* + * prb_iter_data: Retrieve the record data at the current position. + * @iter: Iterator tracking the current position. + * @buf: A buffer to store the data of the record. May be NULL. + * @size: The size of @buf. (Ignored if @buf is NULL.) + * @seq: The sequence number of the record. May be NULL. + * + * If @iter is at a record, provide the data and/or sequence number of that + * record (if specified by the caller). + * + * It is safe to call this function from any context and state. + * + * Returns >=0 if the current record contains valid data (returns 0 if @buf + * is NULL or returns the size of the data block if @buf is non-NULL) or + * -EINVAL if @iter is now invalid. + */ +int prb_iter_data(struct prb_iterator *iter, char *buf, int size, u64 *seq) +{ + struct printk_ringbuffer *rb = iter->rb; + unsigned long lpos = iter->lpos; + unsigned int datsize = 0; + struct prb_entry *e; + + if (buf || seq) { + e = to_entry(rb, lpos); + if (!is_valid(rb, lpos)) + return -EINVAL; + /* memory barrier to ensure valid lpos */ + smp_rmb(); + if (buf) { + datsize = PRB_DATA_SIZE(e); + /* memory barrier to ensure load of datsize */ + smp_rmb(); + if (!is_valid(rb, lpos)) + return -EINVAL; + if (PRB_INDEX(rb, lpos) + datsize > + PRB_SIZE(rb) - PRB_DATA_ALIGN) { + return -EINVAL; + } + if (size > datsize) + size = datsize; + memcpy(buf, &e->data[0], size); + } + if (seq) + *seq = e->seq; + /* memory barrier to ensure loads of entry data */ + smp_rmb(); + } + + if (!is_valid(rb, lpos)) + return -EINVAL; + + return datsize; +} + +/* + * prb_iter_next: Advance to the next record. + * @iter: Iterator tracking the current position. + * @buf: A buffer to store the data of the next record. May be NULL. + * @size: The size of @buf. (Ignored if @buf is NULL.) + * @seq: The sequence number of the next record. May be NULL. + * + * If a next record is available, @iter is advanced and (if specified) + * the data and/or sequence number of that record are provided. + * + * It is safe to call this function from any context and state. + * + * Returns 1 if @iter was advanced, 0 if @iter is at the end of the list, or + * -EINVAL if @iter is now invalid. + */ +int prb_iter_next(struct prb_iterator *iter, char *buf, int size, u64 *seq) +{ + struct printk_ringbuffer *rb = iter->rb; + unsigned long next_lpos; + struct prb_entry *e; + unsigned int esize; + + if (iter->lpos == PRB_INIT) { + next_lpos = atomic_long_read(&rb->tail); + } else { + if (!is_valid(rb, iter->lpos)) + return -EINVAL; + /* memory barrier to ensure valid lpos */ + smp_rmb(); + e = to_entry(rb, iter->lpos); + esize = e->size; + /* memory barrier to ensure load of size */ + smp_rmb(); + if (!is_valid(rb, iter->lpos)) + return -EINVAL; + next_lpos = iter->lpos + esize; + } + if (next_lpos == atomic_long_read(&rb->head)) + return 0; + if (!is_valid(rb, next_lpos)) + return -EINVAL; + /* memory barrier to ensure valid lpos */ + smp_rmb(); + + iter->lpos = next_lpos; + e = to_entry(rb, iter->lpos); + esize = e->size; + /* memory barrier to ensure load of size */ + smp_rmb(); + if (!is_valid(rb, iter->lpos)) + return -EINVAL; + if (esize == -1) + iter->lpos = PRB_WRAP_LPOS(rb, iter->lpos, 1); + + if (prb_iter_data(iter, buf, size, seq) < 0) + return -EINVAL; + + return 1; +} + +/* + * prb_iter_wait_next: Advance to the next record, blocking if none available. + * @iter: Iterator tracking the current position. + * @buf: A buffer to store the data of the next record. May be NULL. + * @size: The size of @buf. (Ignored if @buf is NULL.) + * @seq: The sequence number of the next record. May be NULL. + * + * If a next record is already available, this function works like + * prb_iter_next(). Otherwise block interruptible until a next record is + * available. + * + * When a next record is available, @iter is advanced and (if specified) + * the data and/or sequence number of that record are provided. + * + * This function might sleep. + * + * Returns 1 if @iter was advanced, -EINVAL if @iter is now invalid, or + * -ERESTARTSYS if interrupted by a signal. + */ +int prb_iter_wait_next(struct prb_iterator *iter, char *buf, int size, u64 *seq) +{ + unsigned long last_seen; + int ret; + + for (;;) { + last_seen = atomic_long_read(&iter->rb->wq_counter); + + ret = prb_iter_next(iter, buf, size, seq); + if (ret != 0) + break; + + ret = wait_event_interruptible(*iter->rb->wq, + last_seen != atomic_long_read(&iter->rb->wq_counter)); + if (ret < 0) + break; + } + + return ret; +} + +/* + * prb_iter_seek: Seek forward to a specific record. + * @iter: Iterator to advance. + * @seq: Record number to advance to. + * + * Advance @iter such that a following call to prb_iter_data() will provide + * the contents of the specified record. If a record is specified that does + * not yet exist, advance @iter to the end of the record list. + * + * Note that iterators cannot be rewound. So if a record is requested that + * exists but is previous to @iter in position, @iter is considered invalid. + * + * It is safe to call this function from any context and state. + * + * Returns 1 on succces, 0 if specified record does not yet exist (@iter is + * now at the end of the list), or -EINVAL if @iter is now invalid. + */ +int prb_iter_seek(struct prb_iterator *iter, u64 seq) +{ + u64 cur_seq; + int ret; + + /* first check if the iterator is already at the wanted seq */ + if (seq == 0) { + if (iter->lpos == PRB_INIT) + return 1; + else + return -EINVAL; + } + if (iter->lpos != PRB_INIT) { + if (prb_iter_data(iter, NULL, 0, &cur_seq) >= 0) { + if (cur_seq == seq) + return 1; + if (cur_seq > seq) + return -EINVAL; + } + } + + /* iterate to find the wanted seq */ + for (;;) { + ret = prb_iter_next(iter, NULL, 0, &cur_seq); + if (ret <= 0) + break; + + if (cur_seq == seq) + break; + + if (cur_seq > seq) { + ret = -EINVAL; + break; + } + } + + return ret; +} + +/* + * prb_buffer_size: Get the size of the ring buffer. + * @rb: The ring buffer to get the size of. + * + * Return the number of bytes used for the ring buffer entry storage area. + * Note that this area stores both entry header and entry data. Therefore + * this represents an upper bound to the amount of data that can be stored + * in the ring buffer. + * + * It is safe to call this function from any context and state. + * + * Returns the size in bytes of the entry storage area. + */ +int prb_buffer_size(struct printk_ringbuffer *rb) +{ + return PRB_SIZE(rb); +} + +/* + * prb_inc_lost: Increment the seq counter to signal a lost record. + * @rb: The ring buffer to increment the seq of. + * + * Increment the seq counter so that a seq number is intentially missing + * for the readers. This allows readers to identify that a record is + * missing. A writer will typically use this function if prb_reserve() + * fails. + * + * It is safe to call this function from any context and state. + */ +void prb_inc_lost(struct printk_ringbuffer *rb) +{ + atomic_long_inc(&rb->lost); +} Index: linux-5.6.17-rt10/lib/radix-tree.c =================================================================== --- linux-5.6.17-rt10.orig/lib/radix-tree.c +++ linux-5.6.17-rt10/lib/radix-tree.c @@ -26,7 +26,7 @@ #include #include #include - +#include /* * Radix tree node cache. @@ -72,6 +72,7 @@ struct radix_tree_preload { struct radix_tree_node *nodes; }; static DEFINE_PER_CPU(struct radix_tree_preload, radix_tree_preloads) = { 0, }; +static DEFINE_LOCAL_IRQ_LOCK(radix_tree_preloads_lock); static inline struct radix_tree_node *entry_to_node(void *ptr) { @@ -269,12 +270,13 @@ radix_tree_node_alloc(gfp_t gfp_mask, st * succeed in getting a node here (and never reach * kmem_cache_alloc) */ - rtp = this_cpu_ptr(&radix_tree_preloads); + rtp = &get_locked_var(radix_tree_preloads_lock, radix_tree_preloads); if (rtp->nr) { ret = rtp->nodes; rtp->nodes = ret->parent; rtp->nr--; } + put_locked_var(radix_tree_preloads_lock, radix_tree_preloads); /* * Update the allocation stack trace as this is more useful * for debugging. @@ -340,14 +342,14 @@ static __must_check int __radix_tree_pre */ gfp_mask &= ~__GFP_ACCOUNT; - preempt_disable(); + local_lock(radix_tree_preloads_lock); rtp = this_cpu_ptr(&radix_tree_preloads); while (rtp->nr < nr) { - preempt_enable(); + local_unlock(radix_tree_preloads_lock); node = kmem_cache_alloc(radix_tree_node_cachep, gfp_mask); if (node == NULL) goto out; - preempt_disable(); + local_lock(radix_tree_preloads_lock); rtp = this_cpu_ptr(&radix_tree_preloads); if (rtp->nr < nr) { node->parent = rtp->nodes; @@ -389,11 +391,17 @@ int radix_tree_maybe_preload(gfp_t gfp_m if (gfpflags_allow_blocking(gfp_mask)) return __radix_tree_preload(gfp_mask, RADIX_TREE_PRELOAD_SIZE); /* Preloading doesn't help anything with this gfp mask, skip it */ - preempt_disable(); + local_lock(radix_tree_preloads_lock); return 0; } EXPORT_SYMBOL(radix_tree_maybe_preload); +void radix_tree_preload_end(void) +{ + local_unlock(radix_tree_preloads_lock); +} +EXPORT_SYMBOL(radix_tree_preload_end); + static unsigned radix_tree_load_root(const struct radix_tree_root *root, struct radix_tree_node **nodep, unsigned long *maxindex) { @@ -1478,10 +1486,16 @@ EXPORT_SYMBOL(radix_tree_tagged); void idr_preload(gfp_t gfp_mask) { if (__radix_tree_preload(gfp_mask, IDR_PRELOAD_SIZE)) - preempt_disable(); + local_lock(radix_tree_preloads_lock); } EXPORT_SYMBOL(idr_preload); +void idr_preload_end(void) +{ + local_unlock(radix_tree_preloads_lock); +} +EXPORT_SYMBOL(idr_preload_end); + void __rcu **idr_get_free(struct radix_tree_root *root, struct radix_tree_iter *iter, gfp_t gfp, unsigned long max) Index: linux-5.6.17-rt10/lib/scatterlist.c =================================================================== --- linux-5.6.17-rt10.orig/lib/scatterlist.c +++ linux-5.6.17-rt10/lib/scatterlist.c @@ -811,7 +811,7 @@ void sg_miter_stop(struct sg_mapping_ite flush_kernel_dcache_page(miter->page); if (miter->__flags & SG_MITER_ATOMIC) { - WARN_ON_ONCE(preemptible()); + WARN_ON_ONCE(!pagefault_disabled()); kunmap_atomic(miter->addr); } else kunmap(miter->page); Index: linux-5.6.17-rt10/lib/smp_processor_id.c =================================================================== --- linux-5.6.17-rt10.orig/lib/smp_processor_id.c +++ linux-5.6.17-rt10/lib/smp_processor_id.c @@ -23,6 +23,11 @@ unsigned int check_preemption_disabled(c * Kernel threads bound to a single CPU can safely use * smp_processor_id(): */ +#if defined(CONFIG_PREEMPT_RT) && (defined(CONFIG_SMP) || defined(CONFIG_SCHED_DEBUG)) + if (current->migrate_disable) + goto out; +#endif + if (current->nr_cpus_allowed == 1) goto out; Index: linux-5.6.17-rt10/lib/test_bpf.c =================================================================== --- linux-5.6.17-rt10.orig/lib/test_bpf.c +++ linux-5.6.17-rt10/lib/test_bpf.c @@ -6660,14 +6660,14 @@ static int __run_one(const struct bpf_pr u64 start, finish; int ret = 0, i; - preempt_disable(); + migrate_disable(); start = ktime_get_ns(); for (i = 0; i < runs; i++) ret = BPF_PROG_RUN(fp, data); finish = ktime_get_ns(); - preempt_enable(); + migrate_enable(); *duration = finish - start; do_div(*duration, runs); Index: linux-5.6.17-rt10/localversion-rt =================================================================== --- /dev/null +++ linux-5.6.17-rt10/localversion-rt @@ -0,0 +1 @@ +-rt12 Index: linux-5.6.17-rt10/mm/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/mm/Kconfig +++ linux-5.6.17-rt10/mm/Kconfig @@ -370,7 +370,7 @@ config NOMMU_INITIAL_TRIM_EXCESS config TRANSPARENT_HUGEPAGE bool "Transparent Hugepage Support" - depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE + depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE && !PREEMPT_RT select COMPACTION select XARRAY_MULTI help Index: linux-5.6.17-rt10/mm/compaction.c =================================================================== --- linux-5.6.17-rt10.orig/mm/compaction.c +++ linux-5.6.17-rt10/mm/compaction.c @@ -1590,7 +1590,11 @@ typedef enum { * Allow userspace to control policy on scanning the unevictable LRU for * compactable pages. */ +#ifdef CONFIG_PREEMPT_RT +int sysctl_compact_unevictable_allowed __read_mostly = 0; +#else int sysctl_compact_unevictable_allowed __read_mostly = 1; +#endif static inline void update_fast_start_pfn(struct compact_control *cc, unsigned long pfn) @@ -2240,10 +2244,12 @@ check_drain: block_start_pfn(cc->migrate_pfn, cc->order); if (last_migrated_pfn < current_block_start) { - cpu = get_cpu(); + cpu = get_cpu_light(); + local_lock_irq(swapvec_lock); lru_add_drain_cpu(cpu); + local_unlock_irq(swapvec_lock); drain_local_pages(cc->zone); - put_cpu(); + put_cpu_light(); /* No more flushing until we migrate again */ last_migrated_pfn = 0; } Index: linux-5.6.17-rt10/mm/highmem.c =================================================================== --- linux-5.6.17-rt10.orig/mm/highmem.c +++ linux-5.6.17-rt10/mm/highmem.c @@ -31,8 +31,11 @@ #include #include +#ifndef CONFIG_PREEMPT_RT #if defined(CONFIG_HIGHMEM) || defined(CONFIG_X86_32) DEFINE_PER_CPU(int, __kmap_atomic_idx); +EXPORT_PER_CPU_SYMBOL(__kmap_atomic_idx); +#endif #endif /* @@ -108,8 +111,6 @@ static inline wait_queue_head_t *get_pkm atomic_long_t _totalhigh_pages __read_mostly; EXPORT_SYMBOL(_totalhigh_pages); -EXPORT_PER_CPU_SYMBOL(__kmap_atomic_idx); - unsigned int nr_free_highpages (void) { struct zone *zone; Index: linux-5.6.17-rt10/mm/memcontrol.c =================================================================== --- linux-5.6.17-rt10.orig/mm/memcontrol.c +++ linux-5.6.17-rt10/mm/memcontrol.c @@ -63,6 +63,7 @@ #include #include #include "slab.h" +#include #include @@ -92,6 +93,8 @@ int do_swap_account __read_mostly; static DECLARE_WAIT_QUEUE_HEAD(memcg_cgwb_frn_waitq); #endif +static DEFINE_LOCAL_IRQ_LOCK(event_lock); + /* Whether legacy memory+swap accounting is active */ static bool do_memsw_account(void) { @@ -2161,7 +2164,7 @@ static void drain_all_stock(struct mem_c * as well as workers from this path always operate on the local * per-cpu data. CPU up doesn't touch memcg_stock at all. */ - curcpu = get_cpu(); + curcpu = get_cpu_light(); for_each_online_cpu(cpu) { struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu); struct mem_cgroup *memcg; @@ -2182,7 +2185,7 @@ static void drain_all_stock(struct mem_c schedule_work_on(cpu, &stock->work); } } - put_cpu(); + put_cpu_light(); mutex_unlock(&percpu_charge_mutex); } @@ -5428,12 +5431,12 @@ static int mem_cgroup_move_account(struc ret = 0; - local_irq_disable(); + local_lock_irq(event_lock); mem_cgroup_charge_statistics(to, page, compound, nr_pages); memcg_check_events(to, page); mem_cgroup_charge_statistics(from, page, compound, -nr_pages); memcg_check_events(from, page); - local_irq_enable(); + local_unlock_irq(event_lock); out_unlock: unlock_page(page); out: @@ -6497,10 +6500,10 @@ void mem_cgroup_commit_charge(struct pag commit_charge(page, memcg, lrucare); - local_irq_disable(); + local_lock_irq(event_lock); mem_cgroup_charge_statistics(memcg, page, compound, nr_pages); memcg_check_events(memcg, page); - local_irq_enable(); + local_unlock_irq(event_lock); if (do_memsw_account() && PageSwapCache(page)) { swp_entry_t entry = { .val = page_private(page) }; @@ -6569,7 +6572,7 @@ static void uncharge_batch(const struct memcg_oom_recover(ug->memcg); } - local_irq_save(flags); + local_lock_irqsave(event_lock, flags); __mod_memcg_state(ug->memcg, MEMCG_RSS, -ug->nr_anon); __mod_memcg_state(ug->memcg, MEMCG_CACHE, -ug->nr_file); __mod_memcg_state(ug->memcg, MEMCG_RSS_HUGE, -ug->nr_huge); @@ -6577,7 +6580,7 @@ static void uncharge_batch(const struct __count_memcg_events(ug->memcg, PGPGOUT, ug->pgpgout); __this_cpu_add(ug->memcg->vmstats_percpu->nr_page_events, nr_pages); memcg_check_events(ug->memcg, ug->dummy_page); - local_irq_restore(flags); + local_unlock_irqrestore(event_lock, flags); if (!mem_cgroup_is_root(ug->memcg)) css_put_many(&ug->memcg->css, nr_pages); @@ -6738,11 +6741,11 @@ void mem_cgroup_migrate(struct page *old commit_charge(newpage, memcg, false); - local_irq_save(flags); + local_lock_irqsave(event_lock, flags); mem_cgroup_charge_statistics(memcg, newpage, PageTransHuge(newpage), nr_pages); memcg_check_events(memcg, newpage); - local_irq_restore(flags); + local_unlock_irqrestore(event_lock, flags); } DEFINE_STATIC_KEY_FALSE(memcg_sockets_enabled_key); @@ -6924,6 +6927,7 @@ void mem_cgroup_swapout(struct page *pag struct mem_cgroup *memcg, *swap_memcg; unsigned int nr_entries; unsigned short oldid; + unsigned long flags; VM_BUG_ON_PAGE(PageLRU(page), page); VM_BUG_ON_PAGE(page_count(page), page); @@ -6969,10 +6973,14 @@ void mem_cgroup_swapout(struct page *pag * important here to have the interrupts disabled because it is the * only synchronisation we have for updating the per-CPU variables. */ + local_lock_irqsave(event_lock, flags); +#ifndef CONFIG_PREEMPT_RT VM_BUG_ON(!irqs_disabled()); +#endif mem_cgroup_charge_statistics(memcg, page, PageTransHuge(page), -nr_entries); memcg_check_events(memcg, page); + local_unlock_irqrestore(event_lock, flags); if (!mem_cgroup_is_root(memcg)) css_put_many(&memcg->css, nr_entries); Index: linux-5.6.17-rt10/mm/page_alloc.c =================================================================== --- linux-5.6.17-rt10.orig/mm/page_alloc.c +++ linux-5.6.17-rt10/mm/page_alloc.c @@ -61,6 +61,7 @@ #include #include #include +#include #include #include #include @@ -357,6 +358,8 @@ EXPORT_SYMBOL(nr_node_ids); EXPORT_SYMBOL(nr_online_nodes); #endif +static DEFINE_LOCAL_IRQ_LOCK(pa_lock); + int page_group_by_mobility_disabled __read_mostly; #ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT @@ -1236,7 +1239,7 @@ static inline void prefetch_buddy(struct } /* - * Frees a number of pages from the PCP lists + * Frees a number of pages which have been collected from the pcp lists. * Assumes all pages on list are in same zone, and of same order. * count is the number of pages to free. * @@ -1246,15 +1249,57 @@ static inline void prefetch_buddy(struct * And clear the zone's pages_scanned counter, to hold off the "all pages are * pinned" detection logic. */ -static void free_pcppages_bulk(struct zone *zone, int count, - struct per_cpu_pages *pcp) +static void free_pcppages_bulk(struct zone *zone, struct list_head *head, + bool zone_retry) +{ + bool isolated_pageblocks; + struct page *page, *tmp; + unsigned long flags; + + spin_lock_irqsave(&zone->lock, flags); + isolated_pageblocks = has_isolate_pageblock(zone); + + /* + * Use safe version since after __free_one_page(), + * page->lru.next will not point to original list. + */ + list_for_each_entry_safe(page, tmp, head, lru) { + int mt = get_pcppage_migratetype(page); + + if (page_zone(page) != zone) { + /* + * free_unref_page_list() sorts pages by zone. If we end + * up with pages from a different NUMA nodes belonging + * to the same ZONE index then we need to redo with the + * correct ZONE pointer. Skip the page for now, redo it + * on the next iteration. + */ + WARN_ON_ONCE(zone_retry == false); + if (zone_retry) + continue; + } + + /* MIGRATE_ISOLATE page should not go to pcplists */ + VM_BUG_ON_PAGE(is_migrate_isolate(mt), page); + /* Pageblock could have been isolated meanwhile */ + if (unlikely(isolated_pageblocks)) + mt = get_pageblock_migratetype(page); + + list_del(&page->lru); + __free_one_page(page, page_to_pfn(page), zone, 0, mt); + trace_mm_page_pcpu_drain(page, 0, mt); + } + spin_unlock_irqrestore(&zone->lock, flags); +} + +static void isolate_pcp_pages(int count, struct per_cpu_pages *pcp, + struct list_head *dst) + { int migratetype = 0; int batch_free = 0; int prefetch_nr = 0; - bool isolated_pageblocks; - struct page *page, *tmp; - LIST_HEAD(head); + struct page *page; while (count) { struct list_head *list; @@ -1286,7 +1331,7 @@ static void free_pcppages_bulk(struct zo if (bulkfree_pcp_prepare(page)) continue; - list_add_tail(&page->lru, &head); + list_add_tail(&page->lru, dst); /* * We are going to put the page back to the global @@ -1301,26 +1346,6 @@ static void free_pcppages_bulk(struct zo prefetch_buddy(page); } while (--count && --batch_free && !list_empty(list)); } - - spin_lock(&zone->lock); - isolated_pageblocks = has_isolate_pageblock(zone); - - /* - * Use safe version since after __free_one_page(), - * page->lru.next will not point to original list. - */ - list_for_each_entry_safe(page, tmp, &head, lru) { - int mt = get_pcppage_migratetype(page); - /* MIGRATE_ISOLATE page should not go to pcplists */ - VM_BUG_ON_PAGE(is_migrate_isolate(mt), page); - /* Pageblock could have been isolated meanwhile */ - if (unlikely(isolated_pageblocks)) - mt = get_pageblock_migratetype(page); - - __free_one_page(page, page_to_pfn(page), zone, 0, mt); - trace_mm_page_pcpu_drain(page, 0, mt); - } - spin_unlock(&zone->lock); } static void free_one_page(struct zone *zone, @@ -1421,10 +1446,10 @@ static void __free_pages_ok(struct page return; migratetype = get_pfnblock_migratetype(page, pfn); - local_irq_save(flags); + local_lock_irqsave(pa_lock, flags); __count_vm_events(PGFREE, 1 << order); free_one_page(page_zone(page), page, pfn, order, migratetype); - local_irq_restore(flags); + local_unlock_irqrestore(pa_lock, flags); } void __free_pages_core(struct page *page, unsigned int order) @@ -2797,13 +2822,18 @@ void drain_zone_pages(struct zone *zone, { unsigned long flags; int to_drain, batch; + LIST_HEAD(dst); - local_irq_save(flags); + local_lock_irqsave(pa_lock, flags); batch = READ_ONCE(pcp->batch); to_drain = min(pcp->count, batch); if (to_drain > 0) - free_pcppages_bulk(zone, to_drain, pcp); - local_irq_restore(flags); + isolate_pcp_pages(to_drain, pcp, &dst); + + local_unlock_irqrestore(pa_lock, flags); + + if (to_drain > 0) + free_pcppages_bulk(zone, &dst, false); } #endif @@ -2819,14 +2849,21 @@ static void drain_pages_zone(unsigned in unsigned long flags; struct per_cpu_pageset *pset; struct per_cpu_pages *pcp; + LIST_HEAD(dst); + int count; - local_irq_save(flags); + local_lock_irqsave(pa_lock, flags); pset = per_cpu_ptr(zone->pageset, cpu); pcp = &pset->pcp; - if (pcp->count) - free_pcppages_bulk(zone, pcp->count, pcp); - local_irq_restore(flags); + count = pcp->count; + if (count) + isolate_pcp_pages(count, pcp, &dst); + + local_unlock_irqrestore(pa_lock, flags); + + if (count) + free_pcppages_bulk(zone, &dst, false); } /* @@ -2874,9 +2911,9 @@ static void drain_local_pages_wq(struct * cpu which is allright but we also have to make sure to not move to * a different one. */ - preempt_disable(); + migrate_disable(); drain_local_pages(drain->zone); - preempt_enable(); + migrate_enable(); } /* @@ -3025,7 +3062,8 @@ static bool free_unref_page_prepare(stru return true; } -static void free_unref_page_commit(struct page *page, unsigned long pfn) +static void free_unref_page_commit(struct page *page, unsigned long pfn, + struct list_head *dst) { struct zone *zone = page_zone(page); struct per_cpu_pages *pcp; @@ -3054,7 +3092,8 @@ static void free_unref_page_commit(struc pcp->count++; if (pcp->count >= pcp->high) { unsigned long batch = READ_ONCE(pcp->batch); - free_pcppages_bulk(zone, batch, pcp); + + isolate_pcp_pages(batch, pcp, dst); } } @@ -3065,13 +3104,17 @@ void free_unref_page(struct page *page) { unsigned long flags; unsigned long pfn = page_to_pfn(page); + struct zone *zone = page_zone(page); + LIST_HEAD(dst); if (!free_unref_page_prepare(page, pfn)) return; - local_irq_save(flags); - free_unref_page_commit(page, pfn); - local_irq_restore(flags); + local_lock_irqsave(pa_lock, flags); + free_unref_page_commit(page, pfn, &dst); + local_unlock_irqrestore(pa_lock, flags); + if (!list_empty(&dst)) + free_pcppages_bulk(zone, &dst, false); } /* @@ -3082,6 +3125,11 @@ void free_unref_page_list(struct list_he struct page *page, *next; unsigned long flags, pfn; int batch_count = 0; + struct list_head dsts[__MAX_NR_ZONES]; + int i; + + for (i = 0; i < __MAX_NR_ZONES; i++) + INIT_LIST_HEAD(&dsts[i]); /* Prepare pages for freeing */ list_for_each_entry_safe(page, next, list, lru) { @@ -3091,25 +3139,42 @@ void free_unref_page_list(struct list_he set_page_private(page, pfn); } - local_irq_save(flags); + local_lock_irqsave(pa_lock, flags); list_for_each_entry_safe(page, next, list, lru) { unsigned long pfn = page_private(page); + enum zone_type type; set_page_private(page, 0); trace_mm_page_free_batched(page); - free_unref_page_commit(page, pfn); + type = page_zonenum(page); + free_unref_page_commit(page, pfn, &dsts[type]); /* * Guard against excessive IRQ disabled times when we get * a large list of pages to free. */ if (++batch_count == SWAP_CLUSTER_MAX) { - local_irq_restore(flags); + local_unlock_irqrestore(pa_lock, flags); batch_count = 0; - local_irq_save(flags); + local_lock_irqsave(pa_lock, flags); } } - local_irq_restore(flags); + local_unlock_irqrestore(pa_lock, flags); + + for (i = 0; i < __MAX_NR_ZONES; ) { + struct page *page; + struct zone *zone; + + if (list_empty(&dsts[i])) { + i++; + continue; + } + + page = list_first_entry(&dsts[i], struct page, lru); + zone = page_zone(page); + + free_pcppages_bulk(zone, &dsts[i], true); + } } /* @@ -3244,7 +3309,7 @@ static struct page *rmqueue_pcplist(stru struct page *page; unsigned long flags; - local_irq_save(flags); + local_lock_irqsave(pa_lock, flags); pcp = &this_cpu_ptr(zone->pageset)->pcp; list = &pcp->lists[migratetype]; page = __rmqueue_pcplist(zone, migratetype, alloc_flags, pcp, list); @@ -3252,7 +3317,7 @@ static struct page *rmqueue_pcplist(stru __count_zid_vm_events(PGALLOC, page_zonenum(page), 1); zone_statistics(preferred_zone, zone); } - local_irq_restore(flags); + local_unlock_irqrestore(pa_lock, flags); return page; } @@ -3279,7 +3344,7 @@ struct page *rmqueue(struct zone *prefer * allocate greater than order-1 page units with __GFP_NOFAIL. */ WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1)); - spin_lock_irqsave(&zone->lock, flags); + local_spin_lock_irqsave(pa_lock, &zone->lock, flags); do { page = NULL; @@ -3299,7 +3364,7 @@ struct page *rmqueue(struct zone *prefer __count_zid_vm_events(PGALLOC, page_zonenum(page), 1 << order); zone_statistics(preferred_zone, zone); - local_irq_restore(flags); + local_unlock_irqrestore(pa_lock, flags); out: /* Separate test+clear to avoid unnecessary atomics */ @@ -3312,7 +3377,7 @@ out: return page; failed: - local_irq_restore(flags); + local_unlock_irqrestore(pa_lock, flags); return NULL; } @@ -8660,7 +8725,7 @@ void zone_pcp_reset(struct zone *zone) struct per_cpu_pageset *pset; /* avoid races with drain_pages() */ - local_irq_save(flags); + local_lock_irqsave(pa_lock, flags); if (zone->pageset != &boot_pageset) { for_each_online_cpu(cpu) { pset = per_cpu_ptr(zone->pageset, cpu); @@ -8669,7 +8734,7 @@ void zone_pcp_reset(struct zone *zone) free_percpu(zone->pageset); zone->pageset = &boot_pageset; } - local_irq_restore(flags); + local_unlock_irqrestore(pa_lock, flags); } #ifdef CONFIG_MEMORY_HOTREMOVE Index: linux-5.6.17-rt10/mm/slab.c =================================================================== --- linux-5.6.17-rt10.orig/mm/slab.c +++ linux-5.6.17-rt10/mm/slab.c @@ -233,7 +233,7 @@ static void kmem_cache_node_init(struct parent->shared = NULL; parent->alien = NULL; parent->colour_next = 0; - spin_lock_init(&parent->list_lock); + raw_spin_lock_init(&parent->list_lock); parent->free_objects = 0; parent->free_touched = 0; } @@ -558,9 +558,9 @@ static noinline void cache_free_pfmemall page_node = page_to_nid(page); n = get_node(cachep, page_node); - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); free_block(cachep, &objp, 1, page_node, &list); - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); slabs_destroy(cachep, &list); } @@ -688,7 +688,7 @@ static void __drain_alien_cache(struct k struct kmem_cache_node *n = get_node(cachep, node); if (ac->avail) { - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); /* * Stuff objects into the remote nodes shared array first. * That way we could avoid the overhead of putting the objects @@ -699,7 +699,7 @@ static void __drain_alien_cache(struct k free_block(cachep, ac->entry, ac->avail, node, list); ac->avail = 0; - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); } } @@ -772,9 +772,9 @@ static int __cache_free_alien(struct kme slabs_destroy(cachep, &list); } else { n = get_node(cachep, page_node); - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); free_block(cachep, &objp, 1, page_node, &list); - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); slabs_destroy(cachep, &list); } return 1; @@ -815,10 +815,10 @@ static int init_cache_node(struct kmem_c */ n = get_node(cachep, node); if (n) { - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); n->free_limit = (1 + nr_cpus_node(node)) * cachep->batchcount + cachep->num; - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); return 0; } @@ -897,7 +897,7 @@ static int setup_kmem_cache_node(struct goto fail; n = get_node(cachep, node); - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); if (n->shared && force_change) { free_block(cachep, n->shared->entry, n->shared->avail, node, &list); @@ -915,7 +915,7 @@ static int setup_kmem_cache_node(struct new_alien = NULL; } - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); slabs_destroy(cachep, &list); /* @@ -954,7 +954,7 @@ static void cpuup_canceled(long cpu) if (!n) continue; - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); /* Free limit for this kmem_cache_node */ n->free_limit -= cachep->batchcount; @@ -965,7 +965,7 @@ static void cpuup_canceled(long cpu) nc->avail = 0; if (!cpumask_empty(mask)) { - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); goto free_slab; } @@ -979,7 +979,7 @@ static void cpuup_canceled(long cpu) alien = n->alien; n->alien = NULL; - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); kfree(shared); if (alien) { @@ -1163,7 +1163,7 @@ static void __init init_list(struct kmem /* * Do not assume that spinlocks can be initialized via memcpy: */ - spin_lock_init(&ptr->list_lock); + raw_spin_lock_init(&ptr->list_lock); MAKE_ALL_LISTS(cachep, ptr, nodeid); cachep->node[nodeid] = ptr; @@ -1335,11 +1335,11 @@ slab_out_of_memory(struct kmem_cache *ca for_each_kmem_cache_node(cachep, node, n) { unsigned long total_slabs, free_slabs, free_objs; - spin_lock_irqsave(&n->list_lock, flags); + raw_spin_lock_irqsave(&n->list_lock, flags); total_slabs = n->total_slabs; free_slabs = n->free_slabs; free_objs = n->free_objects; - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); pr_warn(" node %d: slabs: %ld/%ld, objs: %ld/%ld\n", node, total_slabs - free_slabs, total_slabs, @@ -2097,7 +2097,7 @@ static void check_spinlock_acquired(stru { #ifdef CONFIG_SMP check_irq_off(); - assert_spin_locked(&get_node(cachep, numa_mem_id())->list_lock); + assert_raw_spin_locked(&get_node(cachep, numa_mem_id())->list_lock); #endif } @@ -2105,7 +2105,7 @@ static void check_spinlock_acquired_node { #ifdef CONFIG_SMP check_irq_off(); - assert_spin_locked(&get_node(cachep, node)->list_lock); + assert_raw_spin_locked(&get_node(cachep, node)->list_lock); #endif } @@ -2145,9 +2145,9 @@ static void do_drain(void *arg) check_irq_off(); ac = cpu_cache_get(cachep); n = get_node(cachep, node); - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); free_block(cachep, ac->entry, ac->avail, node, &list); - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); slabs_destroy(cachep, &list); ac->avail = 0; } @@ -2165,9 +2165,9 @@ static void drain_cpu_caches(struct kmem drain_alien_cache(cachep, n->alien); for_each_kmem_cache_node(cachep, node, n) { - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); drain_array_locked(cachep, n->shared, node, true, &list); - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); slabs_destroy(cachep, &list); } @@ -2189,10 +2189,10 @@ static int drain_freelist(struct kmem_ca nr_freed = 0; while (nr_freed < tofree && !list_empty(&n->slabs_free)) { - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); p = n->slabs_free.prev; if (p == &n->slabs_free) { - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); goto out; } @@ -2205,7 +2205,7 @@ static int drain_freelist(struct kmem_ca * to the cache. */ n->free_objects -= cache->num; - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); slab_destroy(cache, page); nr_freed++; } @@ -2658,7 +2658,7 @@ static void cache_grow_end(struct kmem_c INIT_LIST_HEAD(&page->slab_list); n = get_node(cachep, page_to_nid(page)); - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); n->total_slabs++; if (!page->active) { list_add_tail(&page->slab_list, &n->slabs_free); @@ -2668,7 +2668,7 @@ static void cache_grow_end(struct kmem_c STATS_INC_GROWN(cachep); n->free_objects += cachep->num - page->active; - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); fixup_objfreelist_debug(cachep, &list); } @@ -2834,7 +2834,7 @@ static struct page *get_first_slab(struc { struct page *page; - assert_spin_locked(&n->list_lock); + assert_raw_spin_locked(&n->list_lock); page = list_first_entry_or_null(&n->slabs_partial, struct page, slab_list); if (!page) { @@ -2861,10 +2861,10 @@ static noinline void *cache_alloc_pfmema if (!gfp_pfmemalloc_allowed(flags)) return NULL; - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); page = get_first_slab(n, true); if (!page) { - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); return NULL; } @@ -2873,7 +2873,7 @@ static noinline void *cache_alloc_pfmema fixup_slab_list(cachep, n, page, &list); - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); fixup_objfreelist_debug(cachep, &list); return obj; @@ -2932,7 +2932,7 @@ static void *cache_alloc_refill(struct k if (!n->free_objects && (!shared || !shared->avail)) goto direct_grow; - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); shared = READ_ONCE(n->shared); /* See if we can refill from the shared array */ @@ -2956,7 +2956,7 @@ static void *cache_alloc_refill(struct k must_grow: n->free_objects -= ac->avail; alloc_done: - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); fixup_objfreelist_debug(cachep, &list); direct_grow: @@ -3181,7 +3181,7 @@ static void *____cache_alloc_node(struct BUG_ON(!n); check_irq_off(); - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); page = get_first_slab(n, false); if (!page) goto must_grow; @@ -3199,12 +3199,12 @@ static void *____cache_alloc_node(struct fixup_slab_list(cachep, n, page, &list); - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); fixup_objfreelist_debug(cachep, &list); return obj; must_grow: - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); page = cache_grow_begin(cachep, gfp_exact_node(flags), nodeid); if (page) { /* This slab isn't counted yet so don't update free_objects */ @@ -3380,7 +3380,7 @@ static void cache_flusharray(struct kmem check_irq_off(); n = get_node(cachep, node); - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); if (n->shared) { struct array_cache *shared_array = n->shared; int max = shared_array->limit - shared_array->avail; @@ -3409,7 +3409,7 @@ free_done: STATS_SET_FREEABLE(cachep, i); } #endif - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); slabs_destroy(cachep, &list); ac->avail -= batchcount; memmove(ac->entry, &(ac->entry[batchcount]), sizeof(void *)*ac->avail); @@ -3831,9 +3831,9 @@ static int __do_tune_cpucache(struct kme node = cpu_to_mem(cpu); n = get_node(cachep, node); - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); free_block(cachep, ac->entry, ac->avail, node, &list); - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); slabs_destroy(cachep, &list); } free_percpu(prev); @@ -3958,9 +3958,9 @@ static void drain_array(struct kmem_cach return; } - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); drain_array_locked(cachep, ac, node, false, &list); - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); slabs_destroy(cachep, &list); } @@ -4044,7 +4044,7 @@ void get_slabinfo(struct kmem_cache *cac for_each_kmem_cache_node(cachep, node, n) { check_irq_on(); - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); total_slabs += n->total_slabs; free_slabs += n->free_slabs; @@ -4053,7 +4053,7 @@ void get_slabinfo(struct kmem_cache *cac if (n->shared) shared_avail += n->shared->avail; - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); } num_objs = total_slabs * cachep->num; active_slabs = total_slabs - free_slabs; Index: linux-5.6.17-rt10/mm/slab.h =================================================================== --- linux-5.6.17-rt10.orig/mm/slab.h +++ linux-5.6.17-rt10/mm/slab.h @@ -596,7 +596,7 @@ static inline void slab_post_alloc_hook( * The slab lists for all objects. */ struct kmem_cache_node { - spinlock_t list_lock; + raw_spinlock_t list_lock; #ifdef CONFIG_SLAB struct list_head slabs_partial; /* partial list first, better asm code */ Index: linux-5.6.17-rt10/mm/slub.c =================================================================== --- linux-5.6.17-rt10.orig/mm/slub.c +++ linux-5.6.17-rt10/mm/slub.c @@ -1196,7 +1196,7 @@ static noinline int free_debug_processin unsigned long uninitialized_var(flags); int ret = 0; - spin_lock_irqsave(&n->list_lock, flags); + raw_spin_lock_irqsave(&n->list_lock, flags); slab_lock(page); if (s->flags & SLAB_CONSISTENCY_CHECKS) { @@ -1231,7 +1231,7 @@ out: bulk_cnt, cnt); slab_unlock(page); - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); if (!ret) slab_fix(s, "Object at 0x%p not freed", object); return ret; @@ -1401,6 +1401,12 @@ static inline void dec_slabs_node(struct #endif /* CONFIG_SLUB_DEBUG */ +struct slub_free_list { + raw_spinlock_t lock; + struct list_head list; +}; +static DEFINE_PER_CPU(struct slub_free_list, slub_free_list); + /* * Hooks for other subsystems that check memory allocations. In a typical * production configuration these hooks all should produce no code at all. @@ -1641,10 +1647,18 @@ static struct page *allocate_slab(struct void *start, *p, *next; int idx; bool shuffle; + bool enableirqs = false; flags &= gfp_allowed_mask; if (gfpflags_allow_blocking(flags)) + enableirqs = true; + +#ifdef CONFIG_PREEMPT_RT + if (system_state > SYSTEM_BOOTING && system_state < SYSTEM_SUSPEND) + enableirqs = true; +#endif + if (enableirqs) local_irq_enable(); flags |= s->allocflags; @@ -1703,7 +1717,7 @@ static struct page *allocate_slab(struct page->frozen = 1; out: - if (gfpflags_allow_blocking(flags)) + if (enableirqs) local_irq_disable(); if (!page) return NULL; @@ -1751,6 +1765,16 @@ static void __free_slab(struct kmem_cach __free_pages(page, order); } +static void free_delayed(struct list_head *h) +{ + while (!list_empty(h)) { + struct page *page = list_first_entry(h, struct page, lru); + + list_del(&page->lru); + __free_slab(page->slab_cache, page); + } +} + static void rcu_free_slab(struct rcu_head *h) { struct page *page = container_of(h, struct page, rcu_head); @@ -1762,6 +1786,12 @@ static void free_slab(struct kmem_cache { if (unlikely(s->flags & SLAB_TYPESAFE_BY_RCU)) { call_rcu(&page->rcu_head, rcu_free_slab); + } else if (irqs_disabled()) { + struct slub_free_list *f = this_cpu_ptr(&slub_free_list); + + raw_spin_lock(&f->lock); + list_add(&page->lru, &f->list); + raw_spin_unlock(&f->lock); } else __free_slab(s, page); } @@ -1869,7 +1899,7 @@ static void *get_partial_node(struct kme if (!n || !n->nr_partial) return NULL; - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); list_for_each_entry_safe(page, page2, &n->partial, slab_list) { void *t; @@ -1894,7 +1924,7 @@ static void *get_partial_node(struct kme break; } - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); return object; } @@ -2140,7 +2170,7 @@ redo: * that acquire_slab() will see a slab page that * is frozen */ - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); } } else { m = M_FULL; @@ -2151,7 +2181,7 @@ redo: * slabs from diagnostic functions will not see * any frozen slabs. */ - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); } } @@ -2175,7 +2205,7 @@ redo: goto redo; if (lock) - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); if (m == M_PARTIAL) stat(s, tail); @@ -2214,10 +2244,10 @@ static void unfreeze_partials(struct kme n2 = get_node(s, page_to_nid(page)); if (n != n2) { if (n) - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); n = n2; - spin_lock(&n->list_lock); + raw_spin_lock(&n->list_lock); } do { @@ -2246,7 +2276,7 @@ static void unfreeze_partials(struct kme } if (n) - spin_unlock(&n->list_lock); + raw_spin_unlock(&n->list_lock); while (discard_page) { page = discard_page; @@ -2283,14 +2313,21 @@ static void put_cpu_partial(struct kmem_ pobjects = oldpage->pobjects; pages = oldpage->pages; if (drain && pobjects > s->cpu_partial) { + struct slub_free_list *f; unsigned long flags; + LIST_HEAD(tofree); /* * partial array is full. Move the existing * set to the per node partial list. */ local_irq_save(flags); unfreeze_partials(s, this_cpu_ptr(s->cpu_slab)); + f = this_cpu_ptr(&slub_free_list); + raw_spin_lock(&f->lock); + list_splice_init(&f->list, &tofree); + raw_spin_unlock(&f->lock); local_irq_restore(flags); + free_delayed(&tofree); oldpage = NULL; pobjects = 0; pages = 0; @@ -2358,7 +2395,19 @@ static bool has_cpu_slab(int cpu, void * static void flush_all(struct kmem_cache *s) { + LIST_HEAD(tofree); + int cpu; + on_each_cpu_cond(has_cpu_slab, flush_cpu_slab, s, 1); + for_each_online_cpu(cpu) { + struct slub_free_list *f; + + f = &per_cpu(slub_free_list, cpu); + raw_spin_lock_irq(&f->lock); + list_splice_init(&f->list, &tofree); + raw_spin_unlock_irq(&f->lock); + free_delayed(&tofree); + } } /* @@ -2413,10 +2462,10 @@ static unsigned long count_partial(struc unsigned long x = 0; struct page *page; - spin_lock_irqsave(&n->list_lock, flags); + raw_spin_lock_irqsave(&n->list_lock, flags); list_for_each_entry(page, &n->partial, slab_list) x += get_count(page); - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); return x; } #endif /* CONFIG_SLUB_DEBUG || CONFIG_SYSFS */ @@ -2555,8 +2604,10 @@ static inline void *get_freelist(struct * already disabled (which is the case for bulk allocation). */ static void *___slab_alloc(struct kmem_cache *s, gfp_t gfpflags, int node, - unsigned long addr, struct kmem_cache_cpu *c) + unsigned long addr, struct kmem_cache_cpu *c, + struct list_head *to_free) { + struct slub_free_list *f; void *freelist; struct page *page; @@ -2622,6 +2673,13 @@ load_freelist: VM_BUG_ON(!c->page->frozen); c->freelist = get_freepointer(s, freelist); c->tid = next_tid(c->tid); + +out: + f = this_cpu_ptr(&slub_free_list); + raw_spin_lock(&f->lock); + list_splice_init(&f->list, to_free); + raw_spin_unlock(&f->lock); + return freelist; new_slab: @@ -2637,7 +2695,7 @@ new_slab: if (unlikely(!freelist)) { slab_out_of_memory(s, gfpflags, node); - return NULL; + goto out; } page = c->page; @@ -2650,7 +2708,7 @@ new_slab: goto new_slab; /* Slab failed checks. Next slab needed */ deactivate_slab(s, page, get_freepointer(s, freelist), c); - return freelist; + goto out; } /* @@ -2662,6 +2720,7 @@ static void *__slab_alloc(struct kmem_ca { void *p; unsigned long flags; + LIST_HEAD(tofree); local_irq_save(flags); #ifdef CONFIG_PREEMPTION @@ -2673,8 +2732,9 @@ static void *__slab_alloc(struct kmem_ca c = this_cpu_ptr(s->cpu_slab); #endif - p = ___slab_alloc(s, gfpflags, node, addr, c); + p = ___slab_alloc(s, gfpflags, node, addr, c, &tofree); local_irq_restore(flags); + free_delayed(&tofree); return p; } @@ -2707,6 +2767,10 @@ static __always_inline void *slab_alloc_ struct page *page; unsigned long tid; + if (IS_ENABLED(CONFIG_PREEMPT_RT) && IS_ENABLED(CONFIG_DEBUG_ATOMIC_SLEEP)) + WARN_ON_ONCE(!preemptible() && + (system_state > SYSTEM_BOOTING && system_state < SYSTEM_SUSPEND)); + s = slab_pre_alloc_hook(s, gfpflags); if (!s) return NULL; @@ -2873,7 +2937,7 @@ static void __slab_free(struct kmem_cach do { if (unlikely(n)) { - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); n = NULL; } prior = page->freelist; @@ -2905,7 +2969,7 @@ static void __slab_free(struct kmem_cach * Otherwise the list_lock will synchronize with * other processors updating the list of slabs. */ - spin_lock_irqsave(&n->list_lock, flags); + raw_spin_lock_irqsave(&n->list_lock, flags); } } @@ -2946,7 +3010,7 @@ static void __slab_free(struct kmem_cach add_partial(n, page, DEACTIVATE_TO_TAIL); stat(s, FREE_ADD_PARTIAL); } - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); return; slab_empty: @@ -2961,7 +3025,7 @@ slab_empty: remove_full(s, n, page); } - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); stat(s, FREE_SLAB); discard_slab(s, page); } @@ -3166,8 +3230,13 @@ int kmem_cache_alloc_bulk(struct kmem_ca void **p) { struct kmem_cache_cpu *c; + LIST_HEAD(to_free); int i; + if (IS_ENABLED(CONFIG_PREEMPT_RT) && IS_ENABLED(CONFIG_DEBUG_ATOMIC_SLEEP)) + WARN_ON_ONCE(!preemptible() && + (system_state > SYSTEM_BOOTING && system_state < SYSTEM_SUSPEND)); + /* memcg and kmem_cache debug support */ s = slab_pre_alloc_hook(s, flags); if (unlikely(!s)) @@ -3198,7 +3267,7 @@ int kmem_cache_alloc_bulk(struct kmem_ca * of re-populating per CPU c->freelist */ p[i] = ___slab_alloc(s, flags, NUMA_NO_NODE, - _RET_IP_, c); + _RET_IP_, c, &to_free); if (unlikely(!p[i])) goto error; @@ -3213,6 +3282,7 @@ int kmem_cache_alloc_bulk(struct kmem_ca } c->tid = next_tid(c->tid); local_irq_enable(); + free_delayed(&to_free); /* Clear memory outside IRQ disabled fastpath loop */ if (unlikely(slab_want_init_on_alloc(flags, s))) { @@ -3227,6 +3297,7 @@ int kmem_cache_alloc_bulk(struct kmem_ca return i; error: local_irq_enable(); + free_delayed(&to_free); slab_post_alloc_hook(s, flags, i, p); __kmem_cache_free_bulk(s, i, p); return 0; @@ -3362,7 +3433,7 @@ static void init_kmem_cache_node(struct kmem_cache_node *n) { n->nr_partial = 0; - spin_lock_init(&n->list_lock); + raw_spin_lock_init(&n->list_lock); INIT_LIST_HEAD(&n->partial); #ifdef CONFIG_SLUB_DEBUG atomic_long_set(&n->nr_slabs, 0); @@ -3711,6 +3782,11 @@ static void list_slab_objects(struct kme const char *text) { #ifdef CONFIG_SLUB_DEBUG +#ifdef CONFIG_PREEMPT_RT + /* XXX move out of irq-off section */ + slab_err(s, page, text, s->name); +#else + void *addr = page_address(page); void *p; unsigned long *map; @@ -3730,6 +3806,7 @@ static void list_slab_objects(struct kme slab_unlock(page); #endif +#endif } /* @@ -3743,7 +3820,7 @@ static void free_partial(struct kmem_cac struct page *page, *h; BUG_ON(irqs_disabled()); - spin_lock_irq(&n->list_lock); + raw_spin_lock_irq(&n->list_lock); list_for_each_entry_safe(page, h, &n->partial, slab_list) { if (!page->inuse) { remove_partial(n, page); @@ -3753,7 +3830,7 @@ static void free_partial(struct kmem_cac "Objects remaining in %s on __kmem_cache_shutdown()"); } } - spin_unlock_irq(&n->list_lock); + raw_spin_unlock_irq(&n->list_lock); list_for_each_entry_safe(page, h, &discard, slab_list) discard_slab(s, page); @@ -4025,7 +4102,7 @@ int __kmem_cache_shrink(struct kmem_cach for (i = 0; i < SHRINK_PROMOTE_MAX; i++) INIT_LIST_HEAD(promote + i); - spin_lock_irqsave(&n->list_lock, flags); + raw_spin_lock_irqsave(&n->list_lock, flags); /* * Build lists of slabs to discard or promote. @@ -4056,7 +4133,7 @@ int __kmem_cache_shrink(struct kmem_cach for (i = SHRINK_PROMOTE_MAX - 1; i >= 0; i--) list_splice(promote + i, &n->partial); - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); /* Release empty slabs */ list_for_each_entry_safe(page, t, &discard, slab_list) @@ -4263,6 +4340,12 @@ void __init kmem_cache_init(void) { static __initdata struct kmem_cache boot_kmem_cache, boot_kmem_cache_node; + int cpu; + + for_each_possible_cpu(cpu) { + raw_spin_lock_init(&per_cpu(slub_free_list, cpu).lock); + INIT_LIST_HEAD(&per_cpu(slub_free_list, cpu).list); + } if (debug_guardpage_minorder()) slub_max_order = 0; @@ -4454,7 +4537,7 @@ static int validate_slab_node(struct kme struct page *page; unsigned long flags; - spin_lock_irqsave(&n->list_lock, flags); + raw_spin_lock_irqsave(&n->list_lock, flags); list_for_each_entry(page, &n->partial, slab_list) { validate_slab(s, page); @@ -4476,7 +4559,7 @@ static int validate_slab_node(struct kme s->name, count, atomic_long_read(&n->nr_slabs)); out: - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); return count; } @@ -4655,12 +4738,12 @@ static int list_locations(struct kmem_ca if (!atomic_long_read(&n->nr_slabs)) continue; - spin_lock_irqsave(&n->list_lock, flags); + raw_spin_lock_irqsave(&n->list_lock, flags); list_for_each_entry(page, &n->partial, slab_list) process_slab(&t, s, page, alloc); list_for_each_entry(page, &n->full, slab_list) process_slab(&t, s, page, alloc); - spin_unlock_irqrestore(&n->list_lock, flags); + raw_spin_unlock_irqrestore(&n->list_lock, flags); } for (i = 0; i < t.count; i++) { Index: linux-5.6.17-rt10/mm/swap.c =================================================================== --- linux-5.6.17-rt10.orig/mm/swap.c +++ linux-5.6.17-rt10/mm/swap.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #include @@ -52,6 +53,8 @@ static DEFINE_PER_CPU(struct pagevec, lr #ifdef CONFIG_SMP static DEFINE_PER_CPU(struct pagevec, activate_page_pvecs); #endif +static DEFINE_LOCAL_IRQ_LOCK(rotate_lock); +DEFINE_LOCAL_IRQ_LOCK(swapvec_lock); /* * This path almost never happens for VM activity - pages are normally @@ -254,11 +257,11 @@ void rotate_reclaimable_page(struct page unsigned long flags; get_page(page); - local_irq_save(flags); + local_lock_irqsave(rotate_lock, flags); pvec = this_cpu_ptr(&lru_rotate_pvecs); if (!pagevec_add(pvec, page) || PageCompound(page)) pagevec_move_tail(pvec); - local_irq_restore(flags); + local_unlock_irqrestore(rotate_lock, flags); } } @@ -308,12 +311,13 @@ void activate_page(struct page *page) { page = compound_head(page); if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) { - struct pagevec *pvec = &get_cpu_var(activate_page_pvecs); + struct pagevec *pvec = &get_locked_var(swapvec_lock, + activate_page_pvecs); get_page(page); if (!pagevec_add(pvec, page) || PageCompound(page)) pagevec_lru_move_fn(pvec, __activate_page, NULL); - put_cpu_var(activate_page_pvecs); + put_locked_var(swapvec_lock, activate_page_pvecs); } } @@ -335,7 +339,7 @@ void activate_page(struct page *page) static void __lru_cache_activate_page(struct page *page) { - struct pagevec *pvec = &get_cpu_var(lru_add_pvec); + struct pagevec *pvec = &get_locked_var(swapvec_lock, lru_add_pvec); int i; /* @@ -357,7 +361,7 @@ static void __lru_cache_activate_page(st } } - put_cpu_var(lru_add_pvec); + put_locked_var(swapvec_lock, lru_add_pvec); } /* @@ -404,12 +408,12 @@ EXPORT_SYMBOL(mark_page_accessed); static void __lru_cache_add(struct page *page) { - struct pagevec *pvec = &get_cpu_var(lru_add_pvec); + struct pagevec *pvec = &get_locked_var(swapvec_lock, lru_add_pvec); get_page(page); if (!pagevec_add(pvec, page) || PageCompound(page)) __pagevec_lru_add(pvec); - put_cpu_var(lru_add_pvec); + put_locked_var(swapvec_lock, lru_add_pvec); } /** @@ -603,9 +607,9 @@ void lru_add_drain_cpu(int cpu) unsigned long flags; /* No harm done if a racing interrupt already did this */ - local_irq_save(flags); + local_lock_irqsave(rotate_lock, flags); pagevec_move_tail(pvec); - local_irq_restore(flags); + local_unlock_irqrestore(rotate_lock, flags); } pvec = &per_cpu(lru_deactivate_file_pvecs, cpu); @@ -641,11 +645,12 @@ void deactivate_file_page(struct page *p return; if (likely(get_page_unless_zero(page))) { - struct pagevec *pvec = &get_cpu_var(lru_deactivate_file_pvecs); + struct pagevec *pvec = &get_locked_var(swapvec_lock, + lru_deactivate_file_pvecs); if (!pagevec_add(pvec, page) || PageCompound(page)) pagevec_lru_move_fn(pvec, lru_deactivate_file_fn, NULL); - put_cpu_var(lru_deactivate_file_pvecs); + put_locked_var(swapvec_lock, lru_deactivate_file_pvecs); } } @@ -680,19 +685,20 @@ void mark_page_lazyfree(struct page *pag { if (PageLRU(page) && PageAnon(page) && PageSwapBacked(page) && !PageSwapCache(page) && !PageUnevictable(page)) { - struct pagevec *pvec = &get_cpu_var(lru_lazyfree_pvecs); + struct pagevec *pvec = &get_locked_var(swapvec_lock, + lru_lazyfree_pvecs); get_page(page); if (!pagevec_add(pvec, page) || PageCompound(page)) pagevec_lru_move_fn(pvec, lru_lazyfree_fn, NULL); - put_cpu_var(lru_lazyfree_pvecs); + put_locked_var(swapvec_lock, lru_lazyfree_pvecs); } } void lru_add_drain(void) { - lru_add_drain_cpu(get_cpu()); - put_cpu(); + lru_add_drain_cpu(local_lock_cpu(swapvec_lock)); + local_unlock_cpu(swapvec_lock); } #ifdef CONFIG_SMP Index: linux-5.6.17-rt10/mm/vmalloc.c =================================================================== --- linux-5.6.17-rt10.orig/mm/vmalloc.c +++ linux-5.6.17-rt10/mm/vmalloc.c @@ -1502,7 +1502,7 @@ static void *new_vmap_block(unsigned int struct vmap_block *vb; struct vmap_area *va; unsigned long vb_idx; - int node, err; + int node, err, cpu; void *vaddr; node = numa_node_id(); @@ -1545,11 +1545,12 @@ static void *new_vmap_block(unsigned int BUG_ON(err); radix_tree_preload_end(); - vbq = &get_cpu_var(vmap_block_queue); + cpu = get_cpu_light(); + vbq = this_cpu_ptr(&vmap_block_queue); spin_lock(&vbq->lock); list_add_tail_rcu(&vb->free_list, &vbq->free); spin_unlock(&vbq->lock); - put_cpu_var(vmap_block_queue); + put_cpu_light(); return vaddr; } @@ -1618,6 +1619,7 @@ static void *vb_alloc(unsigned long size struct vmap_block *vb; void *vaddr = NULL; unsigned int order; + int cpu; BUG_ON(offset_in_page(size)); BUG_ON(size > PAGE_SIZE*VMAP_MAX_ALLOC); @@ -1632,7 +1634,8 @@ static void *vb_alloc(unsigned long size order = get_order(size); rcu_read_lock(); - vbq = &get_cpu_var(vmap_block_queue); + cpu = get_cpu_light(); + vbq = this_cpu_ptr(&vmap_block_queue); list_for_each_entry_rcu(vb, &vbq->free, free_list) { unsigned long pages_off; @@ -1655,7 +1658,7 @@ static void *vb_alloc(unsigned long size break; } - put_cpu_var(vmap_block_queue); + put_cpu_light(); rcu_read_unlock(); /* Allocate new block if nothing was found */ Index: linux-5.6.17-rt10/mm/vmstat.c =================================================================== --- linux-5.6.17-rt10.orig/mm/vmstat.c +++ linux-5.6.17-rt10/mm/vmstat.c @@ -321,6 +321,7 @@ void __mod_zone_page_state(struct zone * long x; long t; + preempt_disable_rt(); x = delta + __this_cpu_read(*p); t = __this_cpu_read(pcp->stat_threshold); @@ -330,6 +331,7 @@ void __mod_zone_page_state(struct zone * x = 0; } __this_cpu_write(*p, x); + preempt_enable_rt(); } EXPORT_SYMBOL(__mod_zone_page_state); @@ -341,6 +343,7 @@ void __mod_node_page_state(struct pglist long x; long t; + preempt_disable_rt(); x = delta + __this_cpu_read(*p); t = __this_cpu_read(pcp->stat_threshold); @@ -350,6 +353,7 @@ void __mod_node_page_state(struct pglist x = 0; } __this_cpu_write(*p, x); + preempt_enable_rt(); } EXPORT_SYMBOL(__mod_node_page_state); @@ -382,6 +386,7 @@ void __inc_zone_state(struct zone *zone, s8 __percpu *p = pcp->vm_stat_diff + item; s8 v, t; + preempt_disable_rt(); v = __this_cpu_inc_return(*p); t = __this_cpu_read(pcp->stat_threshold); if (unlikely(v > t)) { @@ -390,6 +395,7 @@ void __inc_zone_state(struct zone *zone, zone_page_state_add(v + overstep, zone, item); __this_cpu_write(*p, -overstep); } + preempt_enable_rt(); } void __inc_node_state(struct pglist_data *pgdat, enum node_stat_item item) @@ -398,6 +404,7 @@ void __inc_node_state(struct pglist_data s8 __percpu *p = pcp->vm_node_stat_diff + item; s8 v, t; + preempt_disable_rt(); v = __this_cpu_inc_return(*p); t = __this_cpu_read(pcp->stat_threshold); if (unlikely(v > t)) { @@ -406,6 +413,7 @@ void __inc_node_state(struct pglist_data node_page_state_add(v + overstep, pgdat, item); __this_cpu_write(*p, -overstep); } + preempt_enable_rt(); } void __inc_zone_page_state(struct page *page, enum zone_stat_item item) @@ -426,6 +434,7 @@ void __dec_zone_state(struct zone *zone, s8 __percpu *p = pcp->vm_stat_diff + item; s8 v, t; + preempt_disable_rt(); v = __this_cpu_dec_return(*p); t = __this_cpu_read(pcp->stat_threshold); if (unlikely(v < - t)) { @@ -434,6 +443,7 @@ void __dec_zone_state(struct zone *zone, zone_page_state_add(v - overstep, zone, item); __this_cpu_write(*p, overstep); } + preempt_enable_rt(); } void __dec_node_state(struct pglist_data *pgdat, enum node_stat_item item) @@ -442,6 +452,7 @@ void __dec_node_state(struct pglist_data s8 __percpu *p = pcp->vm_node_stat_diff + item; s8 v, t; + preempt_disable_rt(); v = __this_cpu_dec_return(*p); t = __this_cpu_read(pcp->stat_threshold); if (unlikely(v < - t)) { @@ -450,6 +461,7 @@ void __dec_node_state(struct pglist_data node_page_state_add(v - overstep, pgdat, item); __this_cpu_write(*p, overstep); } + preempt_enable_rt(); } void __dec_zone_page_state(struct page *page, enum zone_stat_item item) Index: linux-5.6.17-rt10/mm/workingset.c =================================================================== --- linux-5.6.17-rt10.orig/mm/workingset.c +++ linux-5.6.17-rt10/mm/workingset.c @@ -404,6 +404,8 @@ static struct list_lru shadow_nodes; void workingset_update_node(struct xa_node *node) { + struct address_space *mapping; + /* * Track non-empty nodes that contain only shadow entries; * unlink those that contain pages or are being freed. @@ -412,7 +414,8 @@ void workingset_update_node(struct xa_no * already where they should be. The list_empty() test is safe * as node->private_list is protected by the i_pages lock. */ - VM_WARN_ON_ONCE(!irqs_disabled()); /* For __inc_lruvec_page_state */ + mapping = container_of(node->array, struct address_space, i_pages); + lockdep_assert_held(&mapping->i_pages.xa_lock); if (node->count && node->count == node->nr_values) { if (list_empty(&node->private_list)) { Index: linux-5.6.17-rt10/mm/zsmalloc.c =================================================================== --- linux-5.6.17-rt10.orig/mm/zsmalloc.c +++ linux-5.6.17-rt10/mm/zsmalloc.c @@ -57,6 +57,7 @@ #include #include #include +#include #define ZSPAGE_MAGIC 0x58 @@ -74,9 +75,22 @@ */ #define ZS_MAX_ZSPAGE_ORDER 2 #define ZS_MAX_PAGES_PER_ZSPAGE (_AC(1, UL) << ZS_MAX_ZSPAGE_ORDER) - #define ZS_HANDLE_SIZE (sizeof(unsigned long)) +#ifdef CONFIG_PREEMPT_RT + +struct zsmalloc_handle { + unsigned long addr; + struct mutex lock; +}; + +#define ZS_HANDLE_ALLOC_SIZE (sizeof(struct zsmalloc_handle)) + +#else + +#define ZS_HANDLE_ALLOC_SIZE (sizeof(unsigned long)) +#endif + /* * Object location (, ) is encoded as * as single (unsigned long) handle value. @@ -326,7 +340,7 @@ static void SetZsPageMovable(struct zs_p static int create_cache(struct zs_pool *pool) { - pool->handle_cachep = kmem_cache_create("zs_handle", ZS_HANDLE_SIZE, + pool->handle_cachep = kmem_cache_create("zs_handle", ZS_HANDLE_ALLOC_SIZE, 0, 0, NULL); if (!pool->handle_cachep) return 1; @@ -350,10 +364,27 @@ static void destroy_cache(struct zs_pool static unsigned long cache_alloc_handle(struct zs_pool *pool, gfp_t gfp) { - return (unsigned long)kmem_cache_alloc(pool->handle_cachep, - gfp & ~(__GFP_HIGHMEM|__GFP_MOVABLE)); + void *p; + + p = kmem_cache_alloc(pool->handle_cachep, + gfp & ~(__GFP_HIGHMEM|__GFP_MOVABLE)); +#ifdef CONFIG_PREEMPT_RT + if (p) { + struct zsmalloc_handle *zh = p; + + mutex_init(&zh->lock); + } +#endif + return (unsigned long)p; } +#ifdef CONFIG_PREEMPT_RT +static struct zsmalloc_handle *zs_get_pure_handle(unsigned long handle) +{ + return (void *)(handle &~((1 << OBJ_TAG_BITS) - 1)); +} +#endif + static void cache_free_handle(struct zs_pool *pool, unsigned long handle) { kmem_cache_free(pool->handle_cachep, (void *)handle); @@ -372,12 +403,18 @@ static void cache_free_zspage(struct zs_ static void record_obj(unsigned long handle, unsigned long obj) { +#ifdef CONFIG_PREEMPT_RT + struct zsmalloc_handle *zh = zs_get_pure_handle(handle); + + WRITE_ONCE(zh->addr, obj); +#else /* * lsb of @obj represents handle lock while other bits * represent object value the handle is pointing so * updating shouldn't do store tearing. */ WRITE_ONCE(*(unsigned long *)handle, obj); +#endif } /* zpool driver */ @@ -460,6 +497,7 @@ MODULE_ALIAS("zpool-zsmalloc"); /* per-cpu VM mapping areas for zspage accesses that cross page boundaries */ static DEFINE_PER_CPU(struct mapping_area, zs_map_area); +static DEFINE_LOCAL_IRQ_LOCK(zs_map_area_lock); static bool is_zspage_isolated(struct zspage *zspage) { @@ -869,7 +907,13 @@ static unsigned long location_to_obj(str static unsigned long handle_to_obj(unsigned long handle) { +#ifdef CONFIG_PREEMPT_RT + struct zsmalloc_handle *zh = zs_get_pure_handle(handle); + + return zh->addr; +#else return *(unsigned long *)handle; +#endif } static unsigned long obj_to_head(struct page *page, void *obj) @@ -883,22 +927,46 @@ static unsigned long obj_to_head(struct static inline int testpin_tag(unsigned long handle) { +#ifdef CONFIG_PREEMPT_RT + struct zsmalloc_handle *zh = zs_get_pure_handle(handle); + + return mutex_is_locked(&zh->lock); +#else return bit_spin_is_locked(HANDLE_PIN_BIT, (unsigned long *)handle); +#endif } static inline int trypin_tag(unsigned long handle) { +#ifdef CONFIG_PREEMPT_RT + struct zsmalloc_handle *zh = zs_get_pure_handle(handle); + + return mutex_trylock(&zh->lock); +#else return bit_spin_trylock(HANDLE_PIN_BIT, (unsigned long *)handle); +#endif } static void pin_tag(unsigned long handle) { +#ifdef CONFIG_PREEMPT_RT + struct zsmalloc_handle *zh = zs_get_pure_handle(handle); + + return mutex_lock(&zh->lock); +#else bit_spin_lock(HANDLE_PIN_BIT, (unsigned long *)handle); +#endif } static void unpin_tag(unsigned long handle) { +#ifdef CONFIG_PREEMPT_RT + struct zsmalloc_handle *zh = zs_get_pure_handle(handle); + + return mutex_unlock(&zh->lock); +#else bit_spin_unlock(HANDLE_PIN_BIT, (unsigned long *)handle); +#endif } static void reset_page(struct page *page) @@ -1324,7 +1392,7 @@ void *zs_map_object(struct zs_pool *pool class = pool->size_class[class_idx]; off = (class->size * obj_idx) & ~PAGE_MASK; - area = &get_cpu_var(zs_map_area); + area = &get_locked_var(zs_map_area_lock, zs_map_area); area->vm_mm = mm; if (off + class->size <= PAGE_SIZE) { /* this object is contained entirely within a page */ @@ -1378,7 +1446,7 @@ void zs_unmap_object(struct zs_pool *poo __zs_unmap_object(area, pages, off, class->size); } - put_cpu_var(zs_map_area); + put_locked_var(zs_map_area_lock, zs_map_area); migrate_read_unlock(zspage); unpin_tag(handle); Index: linux-5.6.17-rt10/mm/zswap.c =================================================================== --- linux-5.6.17-rt10.orig/mm/zswap.c +++ linux-5.6.17-rt10/mm/zswap.c @@ -18,6 +18,7 @@ #include #include #include +#include #include #include #include @@ -390,6 +391,8 @@ static struct zswap_entry *zswap_entry_f * per-cpu code **********************************/ static DEFINE_PER_CPU(u8 *, zswap_dstmem); +/* Used for zswap_dstmem and tfm */ +static DEFINE_LOCAL_IRQ_LOCK(zswap_cpu_lock); static int zswap_dstmem_prepare(unsigned int cpu) { @@ -919,10 +922,11 @@ static int zswap_writeback_entry(struct dlen = PAGE_SIZE; src = (u8 *)zhdr + sizeof(struct zswap_header); dst = kmap_atomic(page); - tfm = *get_cpu_ptr(entry->pool->tfm); + local_lock(zswap_cpu_lock); + tfm = *this_cpu_ptr(entry->pool->tfm); ret = crypto_comp_decompress(tfm, src, entry->length, dst, &dlen); - put_cpu_ptr(entry->pool->tfm); + local_unlock(zswap_cpu_lock); kunmap_atomic(dst); BUG_ON(ret); BUG_ON(dlen != PAGE_SIZE); @@ -1074,12 +1078,12 @@ static int zswap_frontswap_store(unsigne } /* compress */ - dst = get_cpu_var(zswap_dstmem); - tfm = *get_cpu_ptr(entry->pool->tfm); + local_lock(zswap_cpu_lock); + dst = *this_cpu_ptr(&zswap_dstmem); + tfm = *this_cpu_ptr(entry->pool->tfm); src = kmap_atomic(page); ret = crypto_comp_compress(tfm, src, PAGE_SIZE, dst, &dlen); kunmap_atomic(src); - put_cpu_ptr(entry->pool->tfm); if (ret) { ret = -EINVAL; goto put_dstmem; @@ -1103,7 +1107,7 @@ static int zswap_frontswap_store(unsigne memcpy(buf, &zhdr, hlen); memcpy(buf + hlen, dst, dlen); zpool_unmap_handle(entry->pool->zpool, handle); - put_cpu_var(zswap_dstmem); + local_unlock(zswap_cpu_lock); /* populate entry */ entry->offset = offset; @@ -1131,7 +1135,7 @@ insert_entry: return 0; put_dstmem: - put_cpu_var(zswap_dstmem); + local_unlock(zswap_cpu_lock); zswap_pool_put(entry->pool); freepage: zswap_entry_cache_free(entry); @@ -1176,9 +1180,10 @@ static int zswap_frontswap_load(unsigned if (zpool_evictable(entry->pool->zpool)) src += sizeof(struct zswap_header); dst = kmap_atomic(page); - tfm = *get_cpu_ptr(entry->pool->tfm); + local_lock(zswap_cpu_lock); + tfm = *this_cpu_ptr(entry->pool->tfm); ret = crypto_comp_decompress(tfm, src, entry->length, dst, &dlen); - put_cpu_ptr(entry->pool->tfm); + local_unlock(zswap_cpu_lock); kunmap_atomic(dst); zpool_unmap_handle(entry->pool->zpool, entry->handle); BUG_ON(ret); Index: linux-5.6.17-rt10/net/Kconfig =================================================================== --- linux-5.6.17-rt10.orig/net/Kconfig +++ linux-5.6.17-rt10/net/Kconfig @@ -282,7 +282,7 @@ config CGROUP_NET_CLASSID config NET_RX_BUSY_POLL bool - default y + default y if !PREEMPT_RT config BQL bool Index: linux-5.6.17-rt10/net/bluetooth/rfcomm/sock.c =================================================================== --- linux-5.6.17-rt10.orig/net/bluetooth/rfcomm/sock.c +++ linux-5.6.17-rt10/net/bluetooth/rfcomm/sock.c @@ -64,15 +64,13 @@ static void rfcomm_sk_data_ready(struct static void rfcomm_sk_state_change(struct rfcomm_dlc *d, int err) { struct sock *sk = d->owner, *parent; - unsigned long flags; if (!sk) return; BT_DBG("dlc %p state %ld err %d", d, d->state, err); - local_irq_save(flags); - bh_lock_sock(sk); + spin_lock_bh(&sk->sk_lock.slock); if (err) sk->sk_err = err; @@ -93,8 +91,7 @@ static void rfcomm_sk_state_change(struc sk->sk_state_change(sk); } - bh_unlock_sock(sk); - local_irq_restore(flags); + spin_unlock_bh(&sk->sk_lock.slock); if (parent && sock_flag(sk, SOCK_ZAPPED)) { /* We have to drop DLC lock here, otherwise Index: linux-5.6.17-rt10/net/bpf/test_run.c =================================================================== --- linux-5.6.17-rt10.orig/net/bpf/test_run.c +++ linux-5.6.17-rt10/net/bpf/test_run.c @@ -37,7 +37,7 @@ static int bpf_test_run(struct bpf_prog repeat = 1; rcu_read_lock(); - preempt_disable(); + migrate_disable(); time_start = ktime_get_ns(); for (i = 0; i < repeat; i++) { bpf_cgroup_storage_set(storage); @@ -54,18 +54,18 @@ static int bpf_test_run(struct bpf_prog if (need_resched()) { time_spent += ktime_get_ns() - time_start; - preempt_enable(); + migrate_enable(); rcu_read_unlock(); cond_resched(); rcu_read_lock(); - preempt_disable(); + migrate_disable(); time_start = ktime_get_ns(); } } time_spent += ktime_get_ns() - time_start; - preempt_enable(); + migrate_enable(); rcu_read_unlock(); do_div(time_spent, repeat); Index: linux-5.6.17-rt10/net/core/dev.c =================================================================== --- linux-5.6.17-rt10.orig/net/core/dev.c +++ linux-5.6.17-rt10/net/core/dev.c @@ -79,6 +79,7 @@ #include #include #include +#include #include #include #include @@ -194,7 +195,7 @@ static DEFINE_SPINLOCK(napi_hash_lock); static unsigned int napi_gen_id = NR_CPUS; static DEFINE_READ_MOSTLY_HASHTABLE(napi_hash, 8); -static seqcount_t devnet_rename_seq; +static DECLARE_RWSEM(devnet_rename_sem); static inline void dev_base_seq_inc(struct net *net) { @@ -217,14 +218,14 @@ static inline struct hlist_head *dev_ind static inline void rps_lock(struct softnet_data *sd) { #ifdef CONFIG_RPS - spin_lock(&sd->input_pkt_queue.lock); + raw_spin_lock(&sd->input_pkt_queue.raw_lock); #endif } static inline void rps_unlock(struct softnet_data *sd) { #ifdef CONFIG_RPS - spin_unlock(&sd->input_pkt_queue.lock); + raw_spin_unlock(&sd->input_pkt_queue.raw_lock); #endif } @@ -930,33 +931,28 @@ EXPORT_SYMBOL(dev_get_by_napi_id); * @net: network namespace * @name: a pointer to the buffer where the name will be stored. * @ifindex: the ifindex of the interface to get the name from. - * - * The use of raw_seqcount_begin() and cond_resched() before - * retrying is required as we want to give the writers a chance - * to complete when CONFIG_PREEMPTION is not set. */ int netdev_get_name(struct net *net, char *name, int ifindex) { struct net_device *dev; - unsigned int seq; + int ret; -retry: - seq = raw_seqcount_begin(&devnet_rename_seq); + down_read(&devnet_rename_sem); rcu_read_lock(); + dev = dev_get_by_index_rcu(net, ifindex); if (!dev) { - rcu_read_unlock(); - return -ENODEV; + ret = -ENODEV; + goto out; } strcpy(name, dev->name); - rcu_read_unlock(); - if (read_seqcount_retry(&devnet_rename_seq, seq)) { - cond_resched(); - goto retry; - } - return 0; + ret = 0; +out: + rcu_read_unlock(); + up_read(&devnet_rename_sem); + return ret; } /** @@ -1228,10 +1224,10 @@ int dev_change_name(struct net_device *d likely(!(dev->priv_flags & IFF_LIVE_RENAME_OK))) return -EBUSY; - write_seqcount_begin(&devnet_rename_seq); + down_write(&devnet_rename_sem); if (strncmp(newname, dev->name, IFNAMSIZ) == 0) { - write_seqcount_end(&devnet_rename_seq); + up_write(&devnet_rename_sem); return 0; } @@ -1239,7 +1235,7 @@ int dev_change_name(struct net_device *d err = dev_get_valid_name(net, dev, newname); if (err < 0) { - write_seqcount_end(&devnet_rename_seq); + up_write(&devnet_rename_sem); return err; } @@ -1254,11 +1250,11 @@ rollback: if (ret) { memcpy(dev->name, oldname, IFNAMSIZ); dev->name_assign_type = old_assign_type; - write_seqcount_end(&devnet_rename_seq); + up_write(&devnet_rename_sem); return ret; } - write_seqcount_end(&devnet_rename_seq); + up_write(&devnet_rename_sem); netdev_adjacent_rename_links(dev, oldname); @@ -1279,7 +1275,7 @@ rollback: /* err >= 0 after dev_alloc_name() or stores the first errno */ if (err >= 0) { err = ret; - write_seqcount_begin(&devnet_rename_seq); + down_write(&devnet_rename_sem); memcpy(dev->name, oldname, IFNAMSIZ); memcpy(oldname, newname, IFNAMSIZ); dev->name_assign_type = old_assign_type; @@ -2949,6 +2945,7 @@ static void __netif_reschedule(struct Qd sd->output_queue_tailp = &q->next_sched; raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_restore(flags); + preempt_check_resched_rt(); } void __netif_schedule(struct Qdisc *q) @@ -3011,6 +3008,7 @@ void __dev_kfree_skb_irq(struct sk_buff __this_cpu_write(softnet_data.completion_queue, skb); raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_restore(flags); + preempt_check_resched_rt(); } EXPORT_SYMBOL(__dev_kfree_skb_irq); @@ -3678,7 +3676,11 @@ static inline int __dev_xmit_skb(struct * This permits qdisc->running owner to get the lock more * often and dequeue packets faster. */ +#ifdef CONFIG_PREEMPT_RT + contended = true; +#else contended = qdisc_is_running(q); +#endif if (unlikely(contended)) spin_lock(&q->busylock); @@ -4472,6 +4474,7 @@ drop: rps_unlock(sd); local_irq_restore(flags); + preempt_check_resched_rt(); atomic_long_inc(&skb->dev->rx_dropped); kfree_skb(skb); @@ -4686,7 +4689,7 @@ static int netif_rx_internal(struct sk_b struct rps_dev_flow voidflow, *rflow = &voidflow; int cpu; - preempt_disable(); + migrate_disable(); rcu_read_lock(); cpu = get_rps_cpu(skb->dev, skb, &rflow); @@ -4696,14 +4699,14 @@ static int netif_rx_internal(struct sk_b ret = enqueue_to_backlog(skb, cpu, &rflow->last_qtail); rcu_read_unlock(); - preempt_enable(); + migrate_enable(); } else #endif { unsigned int qtail; - ret = enqueue_to_backlog(skb, get_cpu(), &qtail); - put_cpu(); + ret = enqueue_to_backlog(skb, get_cpu_light(), &qtail); + put_cpu_light(); } return ret; } @@ -4742,11 +4745,9 @@ int netif_rx_ni(struct sk_buff *skb) trace_netif_rx_ni_entry(skb); - preempt_disable(); + local_bh_disable(); err = netif_rx_internal(skb); - if (local_softirq_pending()) - do_softirq(); - preempt_enable(); + local_bh_enable(); trace_netif_rx_ni_exit(err); return err; @@ -5506,7 +5507,7 @@ static void flush_backlog(struct work_st skb_queue_walk_safe(&sd->input_pkt_queue, skb, tmp) { if (skb->dev->reg_state == NETREG_UNREGISTERING) { __skb_unlink(skb, &sd->input_pkt_queue); - kfree_skb(skb); + dev_kfree_skb_irq(skb); input_queue_head_incr(sd); } } @@ -6104,12 +6105,14 @@ static void net_rps_action_and_irq_enabl sd->rps_ipi_list = NULL; local_irq_enable(); + preempt_check_resched_rt(); /* Send pending IPI's to kick RPS processing on remote cpus. */ net_rps_send_ipi(remsd); } else #endif local_irq_enable(); + preempt_check_resched_rt(); } static bool sd_has_rps_ipi_waiting(struct softnet_data *sd) @@ -6187,6 +6190,7 @@ void __napi_schedule(struct napi_struct local_irq_save(flags); ____napi_schedule(this_cpu_ptr(&softnet_data), n); local_irq_restore(flags); + preempt_check_resched_rt(); } EXPORT_SYMBOL(__napi_schedule); @@ -10174,6 +10178,7 @@ static int dev_cpu_dead(unsigned int old raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_enable(); + preempt_check_resched_rt(); #ifdef CONFIG_RPS remsd = oldsd->rps_ipi_list; @@ -10503,7 +10508,7 @@ static int __init net_dev_init(void) INIT_WORK(flush, flush_backlog); - skb_queue_head_init(&sd->input_pkt_queue); + skb_queue_head_init_raw(&sd->input_pkt_queue); skb_queue_head_init(&sd->process_queue); #ifdef CONFIG_XFRM_OFFLOAD skb_queue_head_init(&sd->xfrm_backlog); Index: linux-5.6.17-rt10/net/core/flow_dissector.c =================================================================== --- linux-5.6.17-rt10.orig/net/core/flow_dissector.c +++ linux-5.6.17-rt10/net/core/flow_dissector.c @@ -936,9 +936,7 @@ bool bpf_flow_dissect(struct bpf_prog *p (int)FLOW_DISSECTOR_F_STOP_AT_ENCAP); flow_keys->flags = flags; - preempt_disable(); - result = BPF_PROG_RUN(prog, ctx); - preempt_enable(); + result = bpf_prog_run_pin_on_cpu(prog, ctx); flow_keys->nhoff = clamp_t(u16, flow_keys->nhoff, nhoff, hlen); flow_keys->thoff = clamp_t(u16, flow_keys->thoff, Index: linux-5.6.17-rt10/net/core/gen_estimator.c =================================================================== --- linux-5.6.17-rt10.orig/net/core/gen_estimator.c +++ linux-5.6.17-rt10/net/core/gen_estimator.c @@ -42,7 +42,7 @@ struct net_rate_estimator { struct gnet_stats_basic_packed *bstats; spinlock_t *stats_lock; - seqcount_t *running; + net_seqlock_t *running; struct gnet_stats_basic_cpu __percpu *cpu_bstats; u8 ewma_log; u8 intvl_log; /* period : (250ms << intvl_log) */ @@ -125,7 +125,7 @@ int gen_new_estimator(struct gnet_stats_ struct gnet_stats_basic_cpu __percpu *cpu_bstats, struct net_rate_estimator __rcu **rate_est, spinlock_t *lock, - seqcount_t *running, + net_seqlock_t *running, struct nlattr *opt) { struct gnet_estimator *parm = nla_data(opt); @@ -223,7 +223,7 @@ int gen_replace_estimator(struct gnet_st struct gnet_stats_basic_cpu __percpu *cpu_bstats, struct net_rate_estimator __rcu **rate_est, spinlock_t *lock, - seqcount_t *running, struct nlattr *opt) + net_seqlock_t *running, struct nlattr *opt) { return gen_new_estimator(bstats, cpu_bstats, rate_est, lock, running, opt); Index: linux-5.6.17-rt10/net/core/gen_stats.c =================================================================== --- linux-5.6.17-rt10.orig/net/core/gen_stats.c +++ linux-5.6.17-rt10/net/core/gen_stats.c @@ -137,7 +137,7 @@ __gnet_stats_copy_basic_cpu(struct gnet_ } void -__gnet_stats_copy_basic(const seqcount_t *running, +__gnet_stats_copy_basic(net_seqlock_t *running, struct gnet_stats_basic_packed *bstats, struct gnet_stats_basic_cpu __percpu *cpu, struct gnet_stats_basic_packed *b) @@ -150,15 +150,15 @@ __gnet_stats_copy_basic(const seqcount_t } do { if (running) - seq = read_seqcount_begin(running); + seq = net_seq_begin(running); bstats->bytes = b->bytes; bstats->packets = b->packets; - } while (running && read_seqcount_retry(running, seq)); + } while (running && net_seq_retry(running, seq)); } EXPORT_SYMBOL(__gnet_stats_copy_basic); static int -___gnet_stats_copy_basic(const seqcount_t *running, +___gnet_stats_copy_basic(net_seqlock_t *running, struct gnet_dump *d, struct gnet_stats_basic_cpu __percpu *cpu, struct gnet_stats_basic_packed *b, @@ -204,7 +204,7 @@ ___gnet_stats_copy_basic(const seqcount_ * if the room in the socket buffer was not sufficient. */ int -gnet_stats_copy_basic(const seqcount_t *running, +gnet_stats_copy_basic(net_seqlock_t *running, struct gnet_dump *d, struct gnet_stats_basic_cpu __percpu *cpu, struct gnet_stats_basic_packed *b) @@ -228,7 +228,7 @@ EXPORT_SYMBOL(gnet_stats_copy_basic); * if the room in the socket buffer was not sufficient. */ int -gnet_stats_copy_basic_hw(const seqcount_t *running, +gnet_stats_copy_basic_hw(net_seqlock_t *running, struct gnet_dump *d, struct gnet_stats_basic_cpu __percpu *cpu, struct gnet_stats_basic_packed *b) Index: linux-5.6.17-rt10/net/core/skmsg.c =================================================================== --- linux-5.6.17-rt10.orig/net/core/skmsg.c +++ linux-5.6.17-rt10/net/core/skmsg.c @@ -628,7 +628,6 @@ int sk_psock_msg_verdict(struct sock *sk struct bpf_prog *prog; int ret; - preempt_disable(); rcu_read_lock(); prog = READ_ONCE(psock->progs.msg_parser); if (unlikely(!prog)) { @@ -638,7 +637,7 @@ int sk_psock_msg_verdict(struct sock *sk sk_msg_compute_data_pointers(msg); msg->sk = sk; - ret = BPF_PROG_RUN(prog, msg); + ret = bpf_prog_run_pin_on_cpu(prog, msg); ret = sk_psock_map_verd(ret, msg->sk_redir); psock->apply_bytes = msg->apply_bytes; if (ret == __SK_REDIRECT) { @@ -653,7 +652,6 @@ int sk_psock_msg_verdict(struct sock *sk } out: rcu_read_unlock(); - preempt_enable(); return ret; } EXPORT_SYMBOL_GPL(sk_psock_msg_verdict); @@ -665,9 +663,7 @@ static int sk_psock_bpf_run(struct sk_ps skb->sk = psock->sk; bpf_compute_data_end_sk_skb(skb); - preempt_disable(); - ret = BPF_PROG_RUN(prog, skb); - preempt_enable(); + ret = bpf_prog_run_pin_on_cpu(prog, skb); /* strparser clones the skb before handing it to a upper layer, * meaning skb_orphan has been called. We NULL sk on the way out * to ensure we don't trigger a BUG_ON() in skb/sk operations Index: linux-5.6.17-rt10/net/kcm/kcmsock.c =================================================================== --- linux-5.6.17-rt10.orig/net/kcm/kcmsock.c +++ linux-5.6.17-rt10/net/kcm/kcmsock.c @@ -380,9 +380,7 @@ static int kcm_parse_func_strparser(stru struct bpf_prog *prog = psock->bpf_prog; int res; - preempt_disable(); - res = BPF_PROG_RUN(prog, skb); - preempt_enable(); + res = bpf_prog_run_pin_on_cpu(prog, skb); return res; } Index: linux-5.6.17-rt10/net/netfilter/nf_conntrack_core.c =================================================================== --- linux-5.6.17-rt10.orig/net/netfilter/nf_conntrack_core.c +++ linux-5.6.17-rt10/net/netfilter/nf_conntrack_core.c @@ -178,7 +178,7 @@ EXPORT_SYMBOL_GPL(nf_conntrack_htable_si unsigned int nf_conntrack_max __read_mostly; EXPORT_SYMBOL_GPL(nf_conntrack_max); -seqcount_t nf_conntrack_generation __read_mostly; +seqcount_spinlock_t nf_conntrack_generation __read_mostly; static unsigned int nf_conntrack_hash_rnd __read_mostly; static u32 hash_conntrack_raw(const struct nf_conntrack_tuple *tuple, @@ -2587,7 +2587,8 @@ int nf_conntrack_init_start(void) /* struct nf_ct_ext uses u8 to store offsets/size */ BUILD_BUG_ON(total_extension_size() > 255u); - seqcount_init(&nf_conntrack_generation); + seqcount_spinlock_init(&nf_conntrack_generation, + &nf_conntrack_locks_all_lock); for (i = 0; i < CONNTRACK_LOCKS; i++) spin_lock_init(&nf_conntrack_locks[i]); Index: linux-5.6.17-rt10/net/netfilter/nft_set_rbtree.c =================================================================== --- linux-5.6.17-rt10.orig/net/netfilter/nft_set_rbtree.c +++ linux-5.6.17-rt10/net/netfilter/nft_set_rbtree.c @@ -18,7 +18,7 @@ struct nft_rbtree { struct rb_root root; rwlock_t lock; - seqcount_t count; + seqcount_rwlock_t count; struct delayed_work gc_work; }; @@ -516,7 +516,7 @@ static int nft_rbtree_init(const struct struct nft_rbtree *priv = nft_set_priv(set); rwlock_init(&priv->lock); - seqcount_init(&priv->count); + seqcount_rwlock_init(&priv->count, &priv->lock); priv->root = RB_ROOT; INIT_DEFERRABLE_WORK(&priv->gc_work, nft_rbtree_gc); Index: linux-5.6.17-rt10/net/packet/af_packet.c =================================================================== --- linux-5.6.17-rt10.orig/net/packet/af_packet.c +++ linux-5.6.17-rt10/net/packet/af_packet.c @@ -57,6 +57,7 @@ #include #include #include +#include #include #include #include @@ -661,7 +662,7 @@ static void prb_retire_rx_blk_timer_expi if (BLOCK_NUM_PKTS(pbd)) { while (atomic_read(&pkc->blk_fill_in_prog)) { /* Waiting for skb_copy_bits to finish... */ - cpu_relax(); + cpu_chill(); } } @@ -923,7 +924,7 @@ static void prb_retire_current_block(str if (!(status & TP_STATUS_BLK_TMO)) { while (atomic_read(&pkc->blk_fill_in_prog)) { /* Waiting for skb_copy_bits to finish... */ - cpu_relax(); + cpu_chill(); } } prb_close_block(pkc, pbd, po, status); Index: linux-5.6.17-rt10/net/sched/sch_api.c =================================================================== --- linux-5.6.17-rt10.orig/net/sched/sch_api.c +++ linux-5.6.17-rt10/net/sched/sch_api.c @@ -1248,7 +1248,7 @@ static struct Qdisc *qdisc_create(struct rcu_assign_pointer(sch->stab, stab); } if (tca[TCA_RATE]) { - seqcount_t *running; + net_seqlock_t *running; err = -EOPNOTSUPP; if (sch->flags & TCQ_F_MQROOT) { Index: linux-5.6.17-rt10/net/sched/sch_generic.c =================================================================== --- linux-5.6.17-rt10.orig/net/sched/sch_generic.c +++ linux-5.6.17-rt10/net/sched/sch_generic.c @@ -552,7 +552,11 @@ struct Qdisc noop_qdisc = { .ops = &noop_qdisc_ops, .q.lock = __SPIN_LOCK_UNLOCKED(noop_qdisc.q.lock), .dev_queue = &noop_netdev_queue, +#ifdef CONFIG_PREEMPT_RT + .running = __SEQLOCK_UNLOCKED(noop_qdisc.running), +#else .running = SEQCNT_ZERO(noop_qdisc.running), +#endif .busylock = __SPIN_LOCK_UNLOCKED(noop_qdisc.busylock), .gso_skb = { .next = (struct sk_buff *)&noop_qdisc.gso_skb, @@ -848,7 +852,11 @@ struct Qdisc *qdisc_alloc(struct netdev_ spin_lock_init(&sch->busylock); /* seqlock has the same scope of busylock, for NOLOCK qdisc */ spin_lock_init(&sch->seqlock); +#ifdef CONFIG_PREEMPT_RT + seqlock_init(&sch->running); +#else seqcount_init(&sch->running); +#endif sch->ops = ops; sch->flags = ops->static_flags; @@ -862,7 +870,11 @@ struct Qdisc *qdisc_alloc(struct netdev_ if (sch != &noop_qdisc) { lockdep_set_class(&sch->busylock, &dev->qdisc_tx_busylock_key); lockdep_set_class(&sch->seqlock, &dev->qdisc_tx_busylock_key); +#ifdef CONFIG_PREEMPT_RT + lockdep_set_class(&sch->running.lock, &dev->qdisc_running_key); +#else lockdep_set_class(&sch->running, &dev->qdisc_running_key); +#endif } return sch; Index: linux-5.6.17-rt10/net/sunrpc/svc_xprt.c =================================================================== --- linux-5.6.17-rt10.orig/net/sunrpc/svc_xprt.c +++ linux-5.6.17-rt10/net/sunrpc/svc_xprt.c @@ -411,7 +411,7 @@ void svc_xprt_do_enqueue(struct svc_xprt if (test_and_set_bit(XPT_BUSY, &xprt->xpt_flags)) return; - cpu = get_cpu(); + cpu = get_cpu_light(); pool = svc_pool_for_cpu(xprt->xpt_server, cpu); atomic_long_inc(&pool->sp_stats.packets); @@ -435,7 +435,7 @@ void svc_xprt_do_enqueue(struct svc_xprt rqstp = NULL; out_unlock: rcu_read_unlock(); - put_cpu(); + put_cpu_light(); trace_svc_xprt_do_enqueue(xprt, rqstp); } EXPORT_SYMBOL_GPL(svc_xprt_do_enqueue); Index: linux-5.6.17-rt10/net/xfrm/xfrm_policy.c =================================================================== --- linux-5.6.17-rt10.orig/net/xfrm/xfrm_policy.c +++ linux-5.6.17-rt10/net/xfrm/xfrm_policy.c @@ -122,7 +122,7 @@ struct xfrm_pol_inexact_bin { /* list containing '*:*' policies */ struct hlist_head hhead; - seqcount_t count; + seqcount_spinlock_t count; /* tree sorted by daddr/prefix */ struct rb_root root_d; @@ -155,7 +155,7 @@ static struct xfrm_policy_afinfo const _ __read_mostly; static struct kmem_cache *xfrm_dst_cache __ro_after_init; -static __read_mostly seqcount_t xfrm_policy_hash_generation; +static __read_mostly seqcount_mutex_t xfrm_policy_hash_generation; static struct rhashtable xfrm_policy_inexact_table; static const struct rhashtable_params xfrm_pol_inexact_params; @@ -719,7 +719,7 @@ xfrm_policy_inexact_alloc_bin(const stru INIT_HLIST_HEAD(&bin->hhead); bin->root_d = RB_ROOT; bin->root_s = RB_ROOT; - seqcount_init(&bin->count); + seqcount_spinlock_init(&bin->count, &net->xfrm.xfrm_policy_lock); prev = rhashtable_lookup_get_insert_key(&xfrm_policy_inexact_table, &bin->k, &bin->head, @@ -1906,7 +1906,7 @@ static int xfrm_policy_match(const struc static struct xfrm_pol_inexact_node * xfrm_policy_lookup_inexact_addr(const struct rb_root *r, - seqcount_t *count, + seqcount_spinlock_t *count, const xfrm_address_t *addr, u16 family) { const struct rb_node *parent; @@ -4154,7 +4154,7 @@ void __init xfrm_init(void) { register_pernet_subsys(&xfrm_net_ops); xfrm_dev_init(); - seqcount_init(&xfrm_policy_hash_generation); + seqcount_mutex_init(&xfrm_policy_hash_generation, &hash_resize_mutex); xfrm_input_init(); #ifdef CONFIG_INET_ESPINTCP Index: linux-5.6.17-rt10/net/xfrm/xfrm_state.c =================================================================== --- linux-5.6.17-rt10.orig/net/xfrm/xfrm_state.c +++ linux-5.6.17-rt10/net/xfrm/xfrm_state.c @@ -44,7 +44,7 @@ static void xfrm_state_gc_task(struct wo */ static unsigned int xfrm_state_hashmax __read_mostly = 1 * 1024 * 1024; -static __read_mostly seqcount_t xfrm_state_hash_generation = SEQCNT_ZERO(xfrm_state_hash_generation); +static __read_mostly seqcount_spinlock_t xfrm_state_hash_generation; static struct kmem_cache *xfrm_state_cache __ro_after_init; static DECLARE_WORK(xfrm_state_gc_work, xfrm_state_gc_task); @@ -139,6 +139,11 @@ static void xfrm_hash_resize(struct work return; } + /* XXX - the locking which protects the sequence counter appears + * to be broken here. The sequence counter is global, but the + * spinlock used for the sequence counter write serialization is + * per network namespace... + */ spin_lock_bh(&net->xfrm.xfrm_state_lock); write_seqcount_begin(&xfrm_state_hash_generation); @@ -2557,6 +2562,8 @@ int __net_init xfrm_state_init(struct ne net->xfrm.state_num = 0; INIT_WORK(&net->xfrm.state_hash_work, xfrm_hash_resize); spin_lock_init(&net->xfrm.xfrm_state_lock); + seqcount_spinlock_init(&xfrm_state_hash_generation, + &net->xfrm.xfrm_state_lock); return 0; out_byspi: Index: linux-5.6.17-rt10/virt/kvm/arm/arm.c =================================================================== --- linux-5.6.17-rt10.orig/virt/kvm/arm/arm.c +++ linux-5.6.17-rt10/virt/kvm/arm/arm.c @@ -700,7 +700,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_v * involves poking the GIC, which must be done in a * non-preemptible context. */ - preempt_disable(); + migrate_disable(); kvm_pmu_flush_hwstate(vcpu); @@ -749,7 +749,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_v kvm_timer_sync_hwstate(vcpu); kvm_vgic_sync_hwstate(vcpu); local_irq_enable(); - preempt_enable(); + migrate_enable(); continue; } @@ -825,7 +825,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_v /* Exit types that need handling before we can be preempted */ handle_exit_early(vcpu, run, ret); - preempt_enable(); + migrate_enable(); ret = handle_exit(vcpu, run, ret); } Index: linux-5.6.17-rt10/virt/kvm/eventfd.c =================================================================== --- linux-5.6.17-rt10.orig/virt/kvm/eventfd.c +++ linux-5.6.17-rt10/virt/kvm/eventfd.c @@ -303,7 +303,7 @@ kvm_irqfd_assign(struct kvm *kvm, struct INIT_LIST_HEAD(&irqfd->list); INIT_WORK(&irqfd->inject, irqfd_inject); INIT_WORK(&irqfd->shutdown, irqfd_shutdown); - seqcount_init(&irqfd->irq_entry_sc); + seqcount_spinlock_init(&irqfd->irq_entry_sc, &kvm->irqfds.lock); f = fdget(args->fd); if (!f.file) {