linux/kernel/locking/rwsem-spinlock.c

314 lines
6.9 KiB
C
Raw Normal View History

/* rwsem-spinlock.c: R/W semaphores: contention handling functions for
* generic spinlock implementation
*
* Copyright (c) 2001 David Howells (dhowells@redhat.com).
* - Derived partially from idea by Andrea Arcangeli <andrea@suse.de>
* - Derived also from comments by Linus
*/
#include <linux/rwsem.h>
#include <linux/sched/signal.h>
#include <linux/sched/debug.h>
#include <linux/export.h>
enum rwsem_waiter_type {
RWSEM_WAITING_FOR_WRITE,
RWSEM_WAITING_FOR_READ
};
struct rwsem_waiter {
struct list_head list;
struct task_struct *task;
enum rwsem_waiter_type type;
};
int rwsem_is_locked(struct rw_semaphore *sem)
{
int ret = 1;
unsigned long flags;
if (raw_spin_trylock_irqsave(&sem->wait_lock, flags)) {
ret = (sem->count != 0);
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
}
return ret;
}
EXPORT_SYMBOL(rwsem_is_locked);
/*
* initialise the semaphore
*/
void __init_rwsem(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
sem->count = 0;
raw_spin_lock_init(&sem->wait_lock);
INIT_LIST_HEAD(&sem->wait_list);
}
EXPORT_SYMBOL(__init_rwsem);
/*
* handle the lock release when processes blocked on it that can now run
* - if we come here, then:
* - the 'active count' _reached_ zero
* - the 'waiting count' is non-zero
* - the spinlock must be held by the caller
* - woken process blocks are discarded from the list after having task zeroed
* - writers are only woken if wakewrite is non-zero
*/
static inline struct rw_semaphore *
__rwsem_do_wake(struct rw_semaphore *sem, int wakewrite)
{
struct rwsem_waiter *waiter;
struct task_struct *tsk;
int woken;
waiter = list_entry(sem->wait_list.next, struct rwsem_waiter, list);
if (waiter->type == RWSEM_WAITING_FOR_WRITE) {
if (wakewrite)
/* Wake up a writer. Note that we do not grant it the
* lock - it will have to acquire it when it runs. */
wake_up_process(waiter->task);
goto out;
}
/* grant an infinite number of read locks to the front of the queue */
woken = 0;
do {
struct list_head *next = waiter->list.next;
list_del(&waiter->list);
tsk = waiter->task;
/*
* Make sure we do not wakeup the next reader before
* setting the nil condition to grant the next reader;
* otherwise we could miss the wakeup on the other
* side and end up sleeping again. See the pairing
* in rwsem_down_read_failed().
*/
smp_mb();
waiter->task = NULL;
wake_up_process(tsk);
put_task_struct(tsk);
woken++;
if (next == &sem->wait_list)
break;
waiter = list_entry(next, struct rwsem_waiter, list);
} while (waiter->type != RWSEM_WAITING_FOR_WRITE);
sem->count += woken;
out:
return sem;
}
/*
* wake a single writer
*/
static inline struct rw_semaphore *
__rwsem_wake_one_writer(struct rw_semaphore *sem)
{
struct rwsem_waiter *waiter;
waiter = list_entry(sem->wait_list.next, struct rwsem_waiter, list);
rwsem-spinlock: Implement writer lock-stealing for better scalability We (Linux Kernel Performance project) found a regression introduced by commit: 5a505085f043 mm/rmap: Convert the struct anon_vma::mutex to an rwsem which converted all anon_vma::mutex locks rwsem write locks. The semantics are the same, but the behavioral difference is quite huge in some cases. After investigating it we found the root cause: mutexes support lock stealing while rwsems don't. Here is the link for the detailed regression report: https://lkml.org/lkml/2013/1/29/84 Ingo suggested adding write lock stealing to rwsems: "I think we should allow lock-steal between rwsem writers - that will not hurt fairness as most rwsem fairness concerns relate to reader vs. writer fairness" And here is the rwsem-spinlock version. With this patch, we got a double performance increase in one test box with following aim7 workfile: FILESIZE: 1M POOLSIZE: 10M 10 fork_test /usr/bin/time output w/o patch /usr/bin/time_output with patch -- Percent of CPU this job got: 369% Percent of CPU this job got: 537% Voluntary context switches: 640595016 Voluntary context switches: 157915561 We got a 45% increase in CPU usage and saved about 3/4 voluntary context switches. Reported-by: LKP project <lkp@linux.intel.com> Suggested-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com> Cc: Alex Shi <alex.shi@intel.com> Cc: David Howells <dhowells@redhat.com> Cc: Michel Lespinasse <walken@google.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Anton Blanchard <anton@samba.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: paul.gortmaker@windriver.com Link: http://lkml.kernel.org/r/1359716356-23865-1-git-send-email-yuanhan.liu@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-02-01 10:59:16 +00:00
wake_up_process(waiter->task);
return sem;
}
/*
* get a read lock on the semaphore
*/
void __sched __down_read(struct rw_semaphore *sem)
{
struct rwsem_waiter waiter;
unsigned long flags;
raw_spin_lock_irqsave(&sem->wait_lock, flags);
if (sem->count >= 0 && list_empty(&sem->wait_list)) {
/* granted */
sem->count++;
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
goto out;
}
sched/core: Remove set_task_state() This is a nasty interface and setting the state of a foreign task must not be done. As of the following commit: be628be0956 ("bcache: Make gc wakeup sane, remove set_task_state()") ... everyone in the kernel calls set_task_state() with current, allowing the helper to be removed. However, as the comment indicates, it is still around for those archs where computing current is more expensive than using a pointer, at least in theory. An important arch that is affected is arm64, however this has been addressed now [1] and performance is up to par making no difference with either calls. Of all the callers, if any, it's the locking bits that would care most about this -- ie: we end up passing a tsk pointer to a lot of the lock slowpath, and setting ->state on that. The following numbers are based on two tests: a custom ad-hoc microbenchmark that just measures latencies (for ~65 million calls) between get_task_state() vs get_current_state(). Secondly for a higher overview, an unlink microbenchmark was used, which pounds on a single file with open, close,unlink combos with increasing thread counts (up to 4x ncpus). While the workload is quite unrealistic, it does contend a lot on the inode mutex or now rwsem. [1] https://lkml.kernel.org/r/1483468021-8237-1-git-send-email-mark.rutland@arm.com == 1. x86-64 == Avg runtime set_task_state(): 601 msecs Avg runtime set_current_state(): 552 msecs vanilla dirty Hmean unlink1-processes-2 36089.26 ( 0.00%) 38977.33 ( 8.00%) Hmean unlink1-processes-5 28555.01 ( 0.00%) 29832.55 ( 4.28%) Hmean unlink1-processes-8 37323.75 ( 0.00%) 44974.57 ( 20.50%) Hmean unlink1-processes-12 43571.88 ( 0.00%) 44283.01 ( 1.63%) Hmean unlink1-processes-21 34431.52 ( 0.00%) 38284.45 ( 11.19%) Hmean unlink1-processes-30 34813.26 ( 0.00%) 37975.17 ( 9.08%) Hmean unlink1-processes-48 37048.90 ( 0.00%) 39862.78 ( 7.59%) Hmean unlink1-processes-79 35630.01 ( 0.00%) 36855.30 ( 3.44%) Hmean unlink1-processes-110 36115.85 ( 0.00%) 39843.91 ( 10.32%) Hmean unlink1-processes-141 32546.96 ( 0.00%) 35418.52 ( 8.82%) Hmean unlink1-processes-172 34674.79 ( 0.00%) 36899.21 ( 6.42%) Hmean unlink1-processes-203 37303.11 ( 0.00%) 36393.04 ( -2.44%) Hmean unlink1-processes-224 35712.13 ( 0.00%) 36685.96 ( 2.73%) == 2. ppc64le == Avg runtime set_task_state(): 938 msecs Avg runtime set_current_state: 940 msecs vanilla dirty Hmean unlink1-processes-2 19269.19 ( 0.00%) 30704.50 ( 59.35%) Hmean unlink1-processes-5 20106.15 ( 0.00%) 21804.15 ( 8.45%) Hmean unlink1-processes-8 17496.97 ( 0.00%) 17243.28 ( -1.45%) Hmean unlink1-processes-12 14224.15 ( 0.00%) 17240.21 ( 21.20%) Hmean unlink1-processes-21 14155.66 ( 0.00%) 15681.23 ( 10.78%) Hmean unlink1-processes-30 14450.70 ( 0.00%) 15995.83 ( 10.69%) Hmean unlink1-processes-48 16945.57 ( 0.00%) 16370.42 ( -3.39%) Hmean unlink1-processes-79 15788.39 ( 0.00%) 14639.27 ( -7.28%) Hmean unlink1-processes-110 14268.48 ( 0.00%) 14377.40 ( 0.76%) Hmean unlink1-processes-141 14023.65 ( 0.00%) 16271.69 ( 16.03%) Hmean unlink1-processes-172 13417.62 ( 0.00%) 16067.55 ( 19.75%) Hmean unlink1-processes-203 15293.08 ( 0.00%) 15440.40 ( 0.96%) Hmean unlink1-processes-234 13719.32 ( 0.00%) 16190.74 ( 18.01%) Hmean unlink1-processes-265 16400.97 ( 0.00%) 16115.22 ( -1.74%) Hmean unlink1-processes-296 14388.60 ( 0.00%) 16216.13 ( 12.70%) Hmean unlink1-processes-320 15771.85 ( 0.00%) 15905.96 ( 0.85%) x86-64 (known to be fast for get_current()/this_cpu_read_stable() caching) and ppc64 (with paca) show similar improvements in the unlink microbenches. The small delta for ppc64 (2ms), does not represent the gains on the unlink runs. In the case of x86, there was a decent amount of variation in the latency runs, but always within a 20 to 50ms increase), ppc was more constant. Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dave@stgolabs.net Cc: mark.rutland@arm.com Link: http://lkml.kernel.org/r/1483479794-14013-5-git-send-email-dave@stgolabs.net Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-01-03 21:43:14 +00:00
set_current_state(TASK_UNINTERRUPTIBLE);
/* set up my own style of waitqueue */
kernel/locking: Compute 'current' directly This patch effectively replaces the tsk pointer dereference (which is obviously == current), to directly use get_current() macro. This is to make the removal of setting foreign task states smoother and painfully obvious. Performance win on some archs such as x86-64 and ppc64. On a microbenchmark that calls set_task_state() vs set_current_state() and an inode rwsem pounding benchmark doing unlink: == 1. x86-64 == Avg runtime set_task_state(): 601 msecs Avg runtime set_current_state(): 552 msecs vanilla dirty Hmean unlink1-processes-2 36089.26 ( 0.00%) 38977.33 ( 8.00%) Hmean unlink1-processes-5 28555.01 ( 0.00%) 29832.55 ( 4.28%) Hmean unlink1-processes-8 37323.75 ( 0.00%) 44974.57 ( 20.50%) Hmean unlink1-processes-12 43571.88 ( 0.00%) 44283.01 ( 1.63%) Hmean unlink1-processes-21 34431.52 ( 0.00%) 38284.45 ( 11.19%) Hmean unlink1-processes-30 34813.26 ( 0.00%) 37975.17 ( 9.08%) Hmean unlink1-processes-48 37048.90 ( 0.00%) 39862.78 ( 7.59%) Hmean unlink1-processes-79 35630.01 ( 0.00%) 36855.30 ( 3.44%) Hmean unlink1-processes-110 36115.85 ( 0.00%) 39843.91 ( 10.32%) Hmean unlink1-processes-141 32546.96 ( 0.00%) 35418.52 ( 8.82%) Hmean unlink1-processes-172 34674.79 ( 0.00%) 36899.21 ( 6.42%) Hmean unlink1-processes-203 37303.11 ( 0.00%) 36393.04 ( -2.44%) Hmean unlink1-processes-224 35712.13 ( 0.00%) 36685.96 ( 2.73%) == 2. ppc64le == Avg runtime set_task_state(): 938 msecs Avg runtime set_current_state: 940 msecs vanilla dirty Hmean unlink1-processes-2 19269.19 ( 0.00%) 30704.50 ( 59.35%) Hmean unlink1-processes-5 20106.15 ( 0.00%) 21804.15 ( 8.45%) Hmean unlink1-processes-8 17496.97 ( 0.00%) 17243.28 ( -1.45%) Hmean unlink1-processes-12 14224.15 ( 0.00%) 17240.21 ( 21.20%) Hmean unlink1-processes-21 14155.66 ( 0.00%) 15681.23 ( 10.78%) Hmean unlink1-processes-30 14450.70 ( 0.00%) 15995.83 ( 10.69%) Hmean unlink1-processes-48 16945.57 ( 0.00%) 16370.42 ( -3.39%) Hmean unlink1-processes-79 15788.39 ( 0.00%) 14639.27 ( -7.28%) Hmean unlink1-processes-110 14268.48 ( 0.00%) 14377.40 ( 0.76%) Hmean unlink1-processes-141 14023.65 ( 0.00%) 16271.69 ( 16.03%) Hmean unlink1-processes-172 13417.62 ( 0.00%) 16067.55 ( 19.75%) Hmean unlink1-processes-203 15293.08 ( 0.00%) 15440.40 ( 0.96%) Hmean unlink1-processes-234 13719.32 ( 0.00%) 16190.74 ( 18.01%) Hmean unlink1-processes-265 16400.97 ( 0.00%) 16115.22 ( -1.74%) Hmean unlink1-processes-296 14388.60 ( 0.00%) 16216.13 ( 12.70%) Hmean unlink1-processes-320 15771.85 ( 0.00%) 15905.96 ( 0.85%) Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dave@stgolabs.net Cc: mark.rutland@arm.com Link: http://lkml.kernel.org/r/1483479794-14013-4-git-send-email-dave@stgolabs.net Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-01-03 21:43:13 +00:00
waiter.task = current;
waiter.type = RWSEM_WAITING_FOR_READ;
kernel/locking: Compute 'current' directly This patch effectively replaces the tsk pointer dereference (which is obviously == current), to directly use get_current() macro. This is to make the removal of setting foreign task states smoother and painfully obvious. Performance win on some archs such as x86-64 and ppc64. On a microbenchmark that calls set_task_state() vs set_current_state() and an inode rwsem pounding benchmark doing unlink: == 1. x86-64 == Avg runtime set_task_state(): 601 msecs Avg runtime set_current_state(): 552 msecs vanilla dirty Hmean unlink1-processes-2 36089.26 ( 0.00%) 38977.33 ( 8.00%) Hmean unlink1-processes-5 28555.01 ( 0.00%) 29832.55 ( 4.28%) Hmean unlink1-processes-8 37323.75 ( 0.00%) 44974.57 ( 20.50%) Hmean unlink1-processes-12 43571.88 ( 0.00%) 44283.01 ( 1.63%) Hmean unlink1-processes-21 34431.52 ( 0.00%) 38284.45 ( 11.19%) Hmean unlink1-processes-30 34813.26 ( 0.00%) 37975.17 ( 9.08%) Hmean unlink1-processes-48 37048.90 ( 0.00%) 39862.78 ( 7.59%) Hmean unlink1-processes-79 35630.01 ( 0.00%) 36855.30 ( 3.44%) Hmean unlink1-processes-110 36115.85 ( 0.00%) 39843.91 ( 10.32%) Hmean unlink1-processes-141 32546.96 ( 0.00%) 35418.52 ( 8.82%) Hmean unlink1-processes-172 34674.79 ( 0.00%) 36899.21 ( 6.42%) Hmean unlink1-processes-203 37303.11 ( 0.00%) 36393.04 ( -2.44%) Hmean unlink1-processes-224 35712.13 ( 0.00%) 36685.96 ( 2.73%) == 2. ppc64le == Avg runtime set_task_state(): 938 msecs Avg runtime set_current_state: 940 msecs vanilla dirty Hmean unlink1-processes-2 19269.19 ( 0.00%) 30704.50 ( 59.35%) Hmean unlink1-processes-5 20106.15 ( 0.00%) 21804.15 ( 8.45%) Hmean unlink1-processes-8 17496.97 ( 0.00%) 17243.28 ( -1.45%) Hmean unlink1-processes-12 14224.15 ( 0.00%) 17240.21 ( 21.20%) Hmean unlink1-processes-21 14155.66 ( 0.00%) 15681.23 ( 10.78%) Hmean unlink1-processes-30 14450.70 ( 0.00%) 15995.83 ( 10.69%) Hmean unlink1-processes-48 16945.57 ( 0.00%) 16370.42 ( -3.39%) Hmean unlink1-processes-79 15788.39 ( 0.00%) 14639.27 ( -7.28%) Hmean unlink1-processes-110 14268.48 ( 0.00%) 14377.40 ( 0.76%) Hmean unlink1-processes-141 14023.65 ( 0.00%) 16271.69 ( 16.03%) Hmean unlink1-processes-172 13417.62 ( 0.00%) 16067.55 ( 19.75%) Hmean unlink1-processes-203 15293.08 ( 0.00%) 15440.40 ( 0.96%) Hmean unlink1-processes-234 13719.32 ( 0.00%) 16190.74 ( 18.01%) Hmean unlink1-processes-265 16400.97 ( 0.00%) 16115.22 ( -1.74%) Hmean unlink1-processes-296 14388.60 ( 0.00%) 16216.13 ( 12.70%) Hmean unlink1-processes-320 15771.85 ( 0.00%) 15905.96 ( 0.85%) Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dave@stgolabs.net Cc: mark.rutland@arm.com Link: http://lkml.kernel.org/r/1483479794-14013-4-git-send-email-dave@stgolabs.net Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-01-03 21:43:13 +00:00
get_task_struct(current);
list_add_tail(&waiter.list, &sem->wait_list);
/* we don't need to touch the semaphore struct anymore */
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
/* wait to be given the lock */
for (;;) {
if (!waiter.task)
break;
schedule();
sched/core: Remove set_task_state() This is a nasty interface and setting the state of a foreign task must not be done. As of the following commit: be628be0956 ("bcache: Make gc wakeup sane, remove set_task_state()") ... everyone in the kernel calls set_task_state() with current, allowing the helper to be removed. However, as the comment indicates, it is still around for those archs where computing current is more expensive than using a pointer, at least in theory. An important arch that is affected is arm64, however this has been addressed now [1] and performance is up to par making no difference with either calls. Of all the callers, if any, it's the locking bits that would care most about this -- ie: we end up passing a tsk pointer to a lot of the lock slowpath, and setting ->state on that. The following numbers are based on two tests: a custom ad-hoc microbenchmark that just measures latencies (for ~65 million calls) between get_task_state() vs get_current_state(). Secondly for a higher overview, an unlink microbenchmark was used, which pounds on a single file with open, close,unlink combos with increasing thread counts (up to 4x ncpus). While the workload is quite unrealistic, it does contend a lot on the inode mutex or now rwsem. [1] https://lkml.kernel.org/r/1483468021-8237-1-git-send-email-mark.rutland@arm.com == 1. x86-64 == Avg runtime set_task_state(): 601 msecs Avg runtime set_current_state(): 552 msecs vanilla dirty Hmean unlink1-processes-2 36089.26 ( 0.00%) 38977.33 ( 8.00%) Hmean unlink1-processes-5 28555.01 ( 0.00%) 29832.55 ( 4.28%) Hmean unlink1-processes-8 37323.75 ( 0.00%) 44974.57 ( 20.50%) Hmean unlink1-processes-12 43571.88 ( 0.00%) 44283.01 ( 1.63%) Hmean unlink1-processes-21 34431.52 ( 0.00%) 38284.45 ( 11.19%) Hmean unlink1-processes-30 34813.26 ( 0.00%) 37975.17 ( 9.08%) Hmean unlink1-processes-48 37048.90 ( 0.00%) 39862.78 ( 7.59%) Hmean unlink1-processes-79 35630.01 ( 0.00%) 36855.30 ( 3.44%) Hmean unlink1-processes-110 36115.85 ( 0.00%) 39843.91 ( 10.32%) Hmean unlink1-processes-141 32546.96 ( 0.00%) 35418.52 ( 8.82%) Hmean unlink1-processes-172 34674.79 ( 0.00%) 36899.21 ( 6.42%) Hmean unlink1-processes-203 37303.11 ( 0.00%) 36393.04 ( -2.44%) Hmean unlink1-processes-224 35712.13 ( 0.00%) 36685.96 ( 2.73%) == 2. ppc64le == Avg runtime set_task_state(): 938 msecs Avg runtime set_current_state: 940 msecs vanilla dirty Hmean unlink1-processes-2 19269.19 ( 0.00%) 30704.50 ( 59.35%) Hmean unlink1-processes-5 20106.15 ( 0.00%) 21804.15 ( 8.45%) Hmean unlink1-processes-8 17496.97 ( 0.00%) 17243.28 ( -1.45%) Hmean unlink1-processes-12 14224.15 ( 0.00%) 17240.21 ( 21.20%) Hmean unlink1-processes-21 14155.66 ( 0.00%) 15681.23 ( 10.78%) Hmean unlink1-processes-30 14450.70 ( 0.00%) 15995.83 ( 10.69%) Hmean unlink1-processes-48 16945.57 ( 0.00%) 16370.42 ( -3.39%) Hmean unlink1-processes-79 15788.39 ( 0.00%) 14639.27 ( -7.28%) Hmean unlink1-processes-110 14268.48 ( 0.00%) 14377.40 ( 0.76%) Hmean unlink1-processes-141 14023.65 ( 0.00%) 16271.69 ( 16.03%) Hmean unlink1-processes-172 13417.62 ( 0.00%) 16067.55 ( 19.75%) Hmean unlink1-processes-203 15293.08 ( 0.00%) 15440.40 ( 0.96%) Hmean unlink1-processes-234 13719.32 ( 0.00%) 16190.74 ( 18.01%) Hmean unlink1-processes-265 16400.97 ( 0.00%) 16115.22 ( -1.74%) Hmean unlink1-processes-296 14388.60 ( 0.00%) 16216.13 ( 12.70%) Hmean unlink1-processes-320 15771.85 ( 0.00%) 15905.96 ( 0.85%) x86-64 (known to be fast for get_current()/this_cpu_read_stable() caching) and ppc64 (with paca) show similar improvements in the unlink microbenches. The small delta for ppc64 (2ms), does not represent the gains on the unlink runs. In the case of x86, there was a decent amount of variation in the latency runs, but always within a 20 to 50ms increase), ppc was more constant. Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dave@stgolabs.net Cc: mark.rutland@arm.com Link: http://lkml.kernel.org/r/1483479794-14013-5-git-send-email-dave@stgolabs.net Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-01-03 21:43:14 +00:00
set_current_state(TASK_UNINTERRUPTIBLE);
}
sched/core: Remove set_task_state() This is a nasty interface and setting the state of a foreign task must not be done. As of the following commit: be628be0956 ("bcache: Make gc wakeup sane, remove set_task_state()") ... everyone in the kernel calls set_task_state() with current, allowing the helper to be removed. However, as the comment indicates, it is still around for those archs where computing current is more expensive than using a pointer, at least in theory. An important arch that is affected is arm64, however this has been addressed now [1] and performance is up to par making no difference with either calls. Of all the callers, if any, it's the locking bits that would care most about this -- ie: we end up passing a tsk pointer to a lot of the lock slowpath, and setting ->state on that. The following numbers are based on two tests: a custom ad-hoc microbenchmark that just measures latencies (for ~65 million calls) between get_task_state() vs get_current_state(). Secondly for a higher overview, an unlink microbenchmark was used, which pounds on a single file with open, close,unlink combos with increasing thread counts (up to 4x ncpus). While the workload is quite unrealistic, it does contend a lot on the inode mutex or now rwsem. [1] https://lkml.kernel.org/r/1483468021-8237-1-git-send-email-mark.rutland@arm.com == 1. x86-64 == Avg runtime set_task_state(): 601 msecs Avg runtime set_current_state(): 552 msecs vanilla dirty Hmean unlink1-processes-2 36089.26 ( 0.00%) 38977.33 ( 8.00%) Hmean unlink1-processes-5 28555.01 ( 0.00%) 29832.55 ( 4.28%) Hmean unlink1-processes-8 37323.75 ( 0.00%) 44974.57 ( 20.50%) Hmean unlink1-processes-12 43571.88 ( 0.00%) 44283.01 ( 1.63%) Hmean unlink1-processes-21 34431.52 ( 0.00%) 38284.45 ( 11.19%) Hmean unlink1-processes-30 34813.26 ( 0.00%) 37975.17 ( 9.08%) Hmean unlink1-processes-48 37048.90 ( 0.00%) 39862.78 ( 7.59%) Hmean unlink1-processes-79 35630.01 ( 0.00%) 36855.30 ( 3.44%) Hmean unlink1-processes-110 36115.85 ( 0.00%) 39843.91 ( 10.32%) Hmean unlink1-processes-141 32546.96 ( 0.00%) 35418.52 ( 8.82%) Hmean unlink1-processes-172 34674.79 ( 0.00%) 36899.21 ( 6.42%) Hmean unlink1-processes-203 37303.11 ( 0.00%) 36393.04 ( -2.44%) Hmean unlink1-processes-224 35712.13 ( 0.00%) 36685.96 ( 2.73%) == 2. ppc64le == Avg runtime set_task_state(): 938 msecs Avg runtime set_current_state: 940 msecs vanilla dirty Hmean unlink1-processes-2 19269.19 ( 0.00%) 30704.50 ( 59.35%) Hmean unlink1-processes-5 20106.15 ( 0.00%) 21804.15 ( 8.45%) Hmean unlink1-processes-8 17496.97 ( 0.00%) 17243.28 ( -1.45%) Hmean unlink1-processes-12 14224.15 ( 0.00%) 17240.21 ( 21.20%) Hmean unlink1-processes-21 14155.66 ( 0.00%) 15681.23 ( 10.78%) Hmean unlink1-processes-30 14450.70 ( 0.00%) 15995.83 ( 10.69%) Hmean unlink1-processes-48 16945.57 ( 0.00%) 16370.42 ( -3.39%) Hmean unlink1-processes-79 15788.39 ( 0.00%) 14639.27 ( -7.28%) Hmean unlink1-processes-110 14268.48 ( 0.00%) 14377.40 ( 0.76%) Hmean unlink1-processes-141 14023.65 ( 0.00%) 16271.69 ( 16.03%) Hmean unlink1-processes-172 13417.62 ( 0.00%) 16067.55 ( 19.75%) Hmean unlink1-processes-203 15293.08 ( 0.00%) 15440.40 ( 0.96%) Hmean unlink1-processes-234 13719.32 ( 0.00%) 16190.74 ( 18.01%) Hmean unlink1-processes-265 16400.97 ( 0.00%) 16115.22 ( -1.74%) Hmean unlink1-processes-296 14388.60 ( 0.00%) 16216.13 ( 12.70%) Hmean unlink1-processes-320 15771.85 ( 0.00%) 15905.96 ( 0.85%) x86-64 (known to be fast for get_current()/this_cpu_read_stable() caching) and ppc64 (with paca) show similar improvements in the unlink microbenches. The small delta for ppc64 (2ms), does not represent the gains on the unlink runs. In the case of x86, there was a decent amount of variation in the latency runs, but always within a 20 to 50ms increase), ppc was more constant. Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dave@stgolabs.net Cc: mark.rutland@arm.com Link: http://lkml.kernel.org/r/1483479794-14013-5-git-send-email-dave@stgolabs.net Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-01-03 21:43:14 +00:00
__set_current_state(TASK_RUNNING);
out:
;
}
/*
* trylock for reading -- returns 1 if successful, 0 if contention
*/
int __down_read_trylock(struct rw_semaphore *sem)
{
unsigned long flags;
int ret = 0;
raw_spin_lock_irqsave(&sem->wait_lock, flags);
if (sem->count >= 0 && list_empty(&sem->wait_list)) {
/* granted */
sem->count++;
ret = 1;
}
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
return ret;
}
/*
* get a write lock on the semaphore
*/
locking/rwsem: Introduce basis for down_write_killable() Introduce a generic implementation necessary for down_write_killable(). This is a trivial extension of the already existing down_write() call which can be interrupted by SIGKILL. This patch doesn't provide down_write_killable() yet because arches have to provide the necessary pieces before. rwsem_down_write_failed() which is a generic slow path for the write lock is extended to take a task state and renamed to __rwsem_down_write_failed_common(). The return value is either a valid semaphore pointer or ERR_PTR(-EINTR). rwsem_down_write_failed_killable() is exported as a new way to wait for the lock and be killable. For rwsem-spinlock implementation the current __down_write() it updated in a similar way as __rwsem_down_write_failed_common() except it doesn't need new exports just visible __down_write_killable(). Architectures which are not using the generic rwsem implementation are supposed to provide their __down_write_killable() implementation and use rwsem_down_write_failed_killable() for the slow path. Signed-off-by: Michal Hocko <mhocko@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Chris Zankel <chris@zankel.net> Cc: David S. Miller <davem@davemloft.net> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Max Filippov <jcmvbkbc@gmail.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Cc: Signed-off-by: Jason Low <jason.low2@hp.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Tony Luck <tony.luck@intel.com> Cc: linux-alpha@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-ia64@vger.kernel.org Cc: linux-s390@vger.kernel.org Cc: linux-sh@vger.kernel.org Cc: linux-xtensa@linux-xtensa.org Cc: sparclinux@vger.kernel.org Link: http://lkml.kernel.org/r/1460041951-22347-7-git-send-email-mhocko@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2016-04-07 15:12:26 +00:00
int __sched __down_write_common(struct rw_semaphore *sem, int state)
{
struct rwsem_waiter waiter;
unsigned long flags;
locking/rwsem: Introduce basis for down_write_killable() Introduce a generic implementation necessary for down_write_killable(). This is a trivial extension of the already existing down_write() call which can be interrupted by SIGKILL. This patch doesn't provide down_write_killable() yet because arches have to provide the necessary pieces before. rwsem_down_write_failed() which is a generic slow path for the write lock is extended to take a task state and renamed to __rwsem_down_write_failed_common(). The return value is either a valid semaphore pointer or ERR_PTR(-EINTR). rwsem_down_write_failed_killable() is exported as a new way to wait for the lock and be killable. For rwsem-spinlock implementation the current __down_write() it updated in a similar way as __rwsem_down_write_failed_common() except it doesn't need new exports just visible __down_write_killable(). Architectures which are not using the generic rwsem implementation are supposed to provide their __down_write_killable() implementation and use rwsem_down_write_failed_killable() for the slow path. Signed-off-by: Michal Hocko <mhocko@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Chris Zankel <chris@zankel.net> Cc: David S. Miller <davem@davemloft.net> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Max Filippov <jcmvbkbc@gmail.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Cc: Signed-off-by: Jason Low <jason.low2@hp.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Tony Luck <tony.luck@intel.com> Cc: linux-alpha@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-ia64@vger.kernel.org Cc: linux-s390@vger.kernel.org Cc: linux-sh@vger.kernel.org Cc: linux-xtensa@linux-xtensa.org Cc: sparclinux@vger.kernel.org Link: http://lkml.kernel.org/r/1460041951-22347-7-git-send-email-mhocko@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2016-04-07 15:12:26 +00:00
int ret = 0;
raw_spin_lock_irqsave(&sem->wait_lock, flags);
/* set up my own style of waitqueue */
kernel/locking: Compute 'current' directly This patch effectively replaces the tsk pointer dereference (which is obviously == current), to directly use get_current() macro. This is to make the removal of setting foreign task states smoother and painfully obvious. Performance win on some archs such as x86-64 and ppc64. On a microbenchmark that calls set_task_state() vs set_current_state() and an inode rwsem pounding benchmark doing unlink: == 1. x86-64 == Avg runtime set_task_state(): 601 msecs Avg runtime set_current_state(): 552 msecs vanilla dirty Hmean unlink1-processes-2 36089.26 ( 0.00%) 38977.33 ( 8.00%) Hmean unlink1-processes-5 28555.01 ( 0.00%) 29832.55 ( 4.28%) Hmean unlink1-processes-8 37323.75 ( 0.00%) 44974.57 ( 20.50%) Hmean unlink1-processes-12 43571.88 ( 0.00%) 44283.01 ( 1.63%) Hmean unlink1-processes-21 34431.52 ( 0.00%) 38284.45 ( 11.19%) Hmean unlink1-processes-30 34813.26 ( 0.00%) 37975.17 ( 9.08%) Hmean unlink1-processes-48 37048.90 ( 0.00%) 39862.78 ( 7.59%) Hmean unlink1-processes-79 35630.01 ( 0.00%) 36855.30 ( 3.44%) Hmean unlink1-processes-110 36115.85 ( 0.00%) 39843.91 ( 10.32%) Hmean unlink1-processes-141 32546.96 ( 0.00%) 35418.52 ( 8.82%) Hmean unlink1-processes-172 34674.79 ( 0.00%) 36899.21 ( 6.42%) Hmean unlink1-processes-203 37303.11 ( 0.00%) 36393.04 ( -2.44%) Hmean unlink1-processes-224 35712.13 ( 0.00%) 36685.96 ( 2.73%) == 2. ppc64le == Avg runtime set_task_state(): 938 msecs Avg runtime set_current_state: 940 msecs vanilla dirty Hmean unlink1-processes-2 19269.19 ( 0.00%) 30704.50 ( 59.35%) Hmean unlink1-processes-5 20106.15 ( 0.00%) 21804.15 ( 8.45%) Hmean unlink1-processes-8 17496.97 ( 0.00%) 17243.28 ( -1.45%) Hmean unlink1-processes-12 14224.15 ( 0.00%) 17240.21 ( 21.20%) Hmean unlink1-processes-21 14155.66 ( 0.00%) 15681.23 ( 10.78%) Hmean unlink1-processes-30 14450.70 ( 0.00%) 15995.83 ( 10.69%) Hmean unlink1-processes-48 16945.57 ( 0.00%) 16370.42 ( -3.39%) Hmean unlink1-processes-79 15788.39 ( 0.00%) 14639.27 ( -7.28%) Hmean unlink1-processes-110 14268.48 ( 0.00%) 14377.40 ( 0.76%) Hmean unlink1-processes-141 14023.65 ( 0.00%) 16271.69 ( 16.03%) Hmean unlink1-processes-172 13417.62 ( 0.00%) 16067.55 ( 19.75%) Hmean unlink1-processes-203 15293.08 ( 0.00%) 15440.40 ( 0.96%) Hmean unlink1-processes-234 13719.32 ( 0.00%) 16190.74 ( 18.01%) Hmean unlink1-processes-265 16400.97 ( 0.00%) 16115.22 ( -1.74%) Hmean unlink1-processes-296 14388.60 ( 0.00%) 16216.13 ( 12.70%) Hmean unlink1-processes-320 15771.85 ( 0.00%) 15905.96 ( 0.85%) Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dave@stgolabs.net Cc: mark.rutland@arm.com Link: http://lkml.kernel.org/r/1483479794-14013-4-git-send-email-dave@stgolabs.net Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-01-03 21:43:13 +00:00
waiter.task = current;
waiter.type = RWSEM_WAITING_FOR_WRITE;
list_add_tail(&waiter.list, &sem->wait_list);
rwsem-spinlock: Implement writer lock-stealing for better scalability We (Linux Kernel Performance project) found a regression introduced by commit: 5a505085f043 mm/rmap: Convert the struct anon_vma::mutex to an rwsem which converted all anon_vma::mutex locks rwsem write locks. The semantics are the same, but the behavioral difference is quite huge in some cases. After investigating it we found the root cause: mutexes support lock stealing while rwsems don't. Here is the link for the detailed regression report: https://lkml.org/lkml/2013/1/29/84 Ingo suggested adding write lock stealing to rwsems: "I think we should allow lock-steal between rwsem writers - that will not hurt fairness as most rwsem fairness concerns relate to reader vs. writer fairness" And here is the rwsem-spinlock version. With this patch, we got a double performance increase in one test box with following aim7 workfile: FILESIZE: 1M POOLSIZE: 10M 10 fork_test /usr/bin/time output w/o patch /usr/bin/time_output with patch -- Percent of CPU this job got: 369% Percent of CPU this job got: 537% Voluntary context switches: 640595016 Voluntary context switches: 157915561 We got a 45% increase in CPU usage and saved about 3/4 voluntary context switches. Reported-by: LKP project <lkp@linux.intel.com> Suggested-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com> Cc: Alex Shi <alex.shi@intel.com> Cc: David Howells <dhowells@redhat.com> Cc: Michel Lespinasse <walken@google.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Anton Blanchard <anton@samba.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: paul.gortmaker@windriver.com Link: http://lkml.kernel.org/r/1359716356-23865-1-git-send-email-yuanhan.liu@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-02-01 10:59:16 +00:00
/* wait for someone to release the lock */
for (;;) {
rwsem-spinlock: Implement writer lock-stealing for better scalability We (Linux Kernel Performance project) found a regression introduced by commit: 5a505085f043 mm/rmap: Convert the struct anon_vma::mutex to an rwsem which converted all anon_vma::mutex locks rwsem write locks. The semantics are the same, but the behavioral difference is quite huge in some cases. After investigating it we found the root cause: mutexes support lock stealing while rwsems don't. Here is the link for the detailed regression report: https://lkml.org/lkml/2013/1/29/84 Ingo suggested adding write lock stealing to rwsems: "I think we should allow lock-steal between rwsem writers - that will not hurt fairness as most rwsem fairness concerns relate to reader vs. writer fairness" And here is the rwsem-spinlock version. With this patch, we got a double performance increase in one test box with following aim7 workfile: FILESIZE: 1M POOLSIZE: 10M 10 fork_test /usr/bin/time output w/o patch /usr/bin/time_output with patch -- Percent of CPU this job got: 369% Percent of CPU this job got: 537% Voluntary context switches: 640595016 Voluntary context switches: 157915561 We got a 45% increase in CPU usage and saved about 3/4 voluntary context switches. Reported-by: LKP project <lkp@linux.intel.com> Suggested-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com> Cc: Alex Shi <alex.shi@intel.com> Cc: David Howells <dhowells@redhat.com> Cc: Michel Lespinasse <walken@google.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Anton Blanchard <anton@samba.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: paul.gortmaker@windriver.com Link: http://lkml.kernel.org/r/1359716356-23865-1-git-send-email-yuanhan.liu@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-02-01 10:59:16 +00:00
/*
* That is the key to support write lock stealing: allows the
* task already on CPU to get the lock soon rather than put
* itself into sleep and waiting for system woke it or someone
* else in the head of the wait list up.
*/
if (sem->count == 0)
break;
locking/rwsem: Introduce basis for down_write_killable() Introduce a generic implementation necessary for down_write_killable(). This is a trivial extension of the already existing down_write() call which can be interrupted by SIGKILL. This patch doesn't provide down_write_killable() yet because arches have to provide the necessary pieces before. rwsem_down_write_failed() which is a generic slow path for the write lock is extended to take a task state and renamed to __rwsem_down_write_failed_common(). The return value is either a valid semaphore pointer or ERR_PTR(-EINTR). rwsem_down_write_failed_killable() is exported as a new way to wait for the lock and be killable. For rwsem-spinlock implementation the current __down_write() it updated in a similar way as __rwsem_down_write_failed_common() except it doesn't need new exports just visible __down_write_killable(). Architectures which are not using the generic rwsem implementation are supposed to provide their __down_write_killable() implementation and use rwsem_down_write_failed_killable() for the slow path. Signed-off-by: Michal Hocko <mhocko@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Chris Zankel <chris@zankel.net> Cc: David S. Miller <davem@davemloft.net> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Max Filippov <jcmvbkbc@gmail.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Cc: Signed-off-by: Jason Low <jason.low2@hp.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Tony Luck <tony.luck@intel.com> Cc: linux-alpha@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-ia64@vger.kernel.org Cc: linux-s390@vger.kernel.org Cc: linux-sh@vger.kernel.org Cc: linux-xtensa@linux-xtensa.org Cc: sparclinux@vger.kernel.org Link: http://lkml.kernel.org/r/1460041951-22347-7-git-send-email-mhocko@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2016-04-07 15:12:26 +00:00
if (signal_pending_state(state, current)) {
ret = -EINTR;
goto out;
}
sched/core: Remove set_task_state() This is a nasty interface and setting the state of a foreign task must not be done. As of the following commit: be628be0956 ("bcache: Make gc wakeup sane, remove set_task_state()") ... everyone in the kernel calls set_task_state() with current, allowing the helper to be removed. However, as the comment indicates, it is still around for those archs where computing current is more expensive than using a pointer, at least in theory. An important arch that is affected is arm64, however this has been addressed now [1] and performance is up to par making no difference with either calls. Of all the callers, if any, it's the locking bits that would care most about this -- ie: we end up passing a tsk pointer to a lot of the lock slowpath, and setting ->state on that. The following numbers are based on two tests: a custom ad-hoc microbenchmark that just measures latencies (for ~65 million calls) between get_task_state() vs get_current_state(). Secondly for a higher overview, an unlink microbenchmark was used, which pounds on a single file with open, close,unlink combos with increasing thread counts (up to 4x ncpus). While the workload is quite unrealistic, it does contend a lot on the inode mutex or now rwsem. [1] https://lkml.kernel.org/r/1483468021-8237-1-git-send-email-mark.rutland@arm.com == 1. x86-64 == Avg runtime set_task_state(): 601 msecs Avg runtime set_current_state(): 552 msecs vanilla dirty Hmean unlink1-processes-2 36089.26 ( 0.00%) 38977.33 ( 8.00%) Hmean unlink1-processes-5 28555.01 ( 0.00%) 29832.55 ( 4.28%) Hmean unlink1-processes-8 37323.75 ( 0.00%) 44974.57 ( 20.50%) Hmean unlink1-processes-12 43571.88 ( 0.00%) 44283.01 ( 1.63%) Hmean unlink1-processes-21 34431.52 ( 0.00%) 38284.45 ( 11.19%) Hmean unlink1-processes-30 34813.26 ( 0.00%) 37975.17 ( 9.08%) Hmean unlink1-processes-48 37048.90 ( 0.00%) 39862.78 ( 7.59%) Hmean unlink1-processes-79 35630.01 ( 0.00%) 36855.30 ( 3.44%) Hmean unlink1-processes-110 36115.85 ( 0.00%) 39843.91 ( 10.32%) Hmean unlink1-processes-141 32546.96 ( 0.00%) 35418.52 ( 8.82%) Hmean unlink1-processes-172 34674.79 ( 0.00%) 36899.21 ( 6.42%) Hmean unlink1-processes-203 37303.11 ( 0.00%) 36393.04 ( -2.44%) Hmean unlink1-processes-224 35712.13 ( 0.00%) 36685.96 ( 2.73%) == 2. ppc64le == Avg runtime set_task_state(): 938 msecs Avg runtime set_current_state: 940 msecs vanilla dirty Hmean unlink1-processes-2 19269.19 ( 0.00%) 30704.50 ( 59.35%) Hmean unlink1-processes-5 20106.15 ( 0.00%) 21804.15 ( 8.45%) Hmean unlink1-processes-8 17496.97 ( 0.00%) 17243.28 ( -1.45%) Hmean unlink1-processes-12 14224.15 ( 0.00%) 17240.21 ( 21.20%) Hmean unlink1-processes-21 14155.66 ( 0.00%) 15681.23 ( 10.78%) Hmean unlink1-processes-30 14450.70 ( 0.00%) 15995.83 ( 10.69%) Hmean unlink1-processes-48 16945.57 ( 0.00%) 16370.42 ( -3.39%) Hmean unlink1-processes-79 15788.39 ( 0.00%) 14639.27 ( -7.28%) Hmean unlink1-processes-110 14268.48 ( 0.00%) 14377.40 ( 0.76%) Hmean unlink1-processes-141 14023.65 ( 0.00%) 16271.69 ( 16.03%) Hmean unlink1-processes-172 13417.62 ( 0.00%) 16067.55 ( 19.75%) Hmean unlink1-processes-203 15293.08 ( 0.00%) 15440.40 ( 0.96%) Hmean unlink1-processes-234 13719.32 ( 0.00%) 16190.74 ( 18.01%) Hmean unlink1-processes-265 16400.97 ( 0.00%) 16115.22 ( -1.74%) Hmean unlink1-processes-296 14388.60 ( 0.00%) 16216.13 ( 12.70%) Hmean unlink1-processes-320 15771.85 ( 0.00%) 15905.96 ( 0.85%) x86-64 (known to be fast for get_current()/this_cpu_read_stable() caching) and ppc64 (with paca) show similar improvements in the unlink microbenches. The small delta for ppc64 (2ms), does not represent the gains on the unlink runs. In the case of x86, there was a decent amount of variation in the latency runs, but always within a 20 to 50ms increase), ppc was more constant. Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dave@stgolabs.net Cc: mark.rutland@arm.com Link: http://lkml.kernel.org/r/1483479794-14013-5-git-send-email-dave@stgolabs.net Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-01-03 21:43:14 +00:00
set_current_state(state);
rwsem-spinlock: Implement writer lock-stealing for better scalability We (Linux Kernel Performance project) found a regression introduced by commit: 5a505085f043 mm/rmap: Convert the struct anon_vma::mutex to an rwsem which converted all anon_vma::mutex locks rwsem write locks. The semantics are the same, but the behavioral difference is quite huge in some cases. After investigating it we found the root cause: mutexes support lock stealing while rwsems don't. Here is the link for the detailed regression report: https://lkml.org/lkml/2013/1/29/84 Ingo suggested adding write lock stealing to rwsems: "I think we should allow lock-steal between rwsem writers - that will not hurt fairness as most rwsem fairness concerns relate to reader vs. writer fairness" And here is the rwsem-spinlock version. With this patch, we got a double performance increase in one test box with following aim7 workfile: FILESIZE: 1M POOLSIZE: 10M 10 fork_test /usr/bin/time output w/o patch /usr/bin/time_output with patch -- Percent of CPU this job got: 369% Percent of CPU this job got: 537% Voluntary context switches: 640595016 Voluntary context switches: 157915561 We got a 45% increase in CPU usage and saved about 3/4 voluntary context switches. Reported-by: LKP project <lkp@linux.intel.com> Suggested-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com> Cc: Alex Shi <alex.shi@intel.com> Cc: David Howells <dhowells@redhat.com> Cc: Michel Lespinasse <walken@google.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Anton Blanchard <anton@samba.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: paul.gortmaker@windriver.com Link: http://lkml.kernel.org/r/1359716356-23865-1-git-send-email-yuanhan.liu@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-02-01 10:59:16 +00:00
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
schedule();
raw_spin_lock_irqsave(&sem->wait_lock, flags);
}
rwsem-spinlock: Implement writer lock-stealing for better scalability We (Linux Kernel Performance project) found a regression introduced by commit: 5a505085f043 mm/rmap: Convert the struct anon_vma::mutex to an rwsem which converted all anon_vma::mutex locks rwsem write locks. The semantics are the same, but the behavioral difference is quite huge in some cases. After investigating it we found the root cause: mutexes support lock stealing while rwsems don't. Here is the link for the detailed regression report: https://lkml.org/lkml/2013/1/29/84 Ingo suggested adding write lock stealing to rwsems: "I think we should allow lock-steal between rwsem writers - that will not hurt fairness as most rwsem fairness concerns relate to reader vs. writer fairness" And here is the rwsem-spinlock version. With this patch, we got a double performance increase in one test box with following aim7 workfile: FILESIZE: 1M POOLSIZE: 10M 10 fork_test /usr/bin/time output w/o patch /usr/bin/time_output with patch -- Percent of CPU this job got: 369% Percent of CPU this job got: 537% Voluntary context switches: 640595016 Voluntary context switches: 157915561 We got a 45% increase in CPU usage and saved about 3/4 voluntary context switches. Reported-by: LKP project <lkp@linux.intel.com> Suggested-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com> Cc: Alex Shi <alex.shi@intel.com> Cc: David Howells <dhowells@redhat.com> Cc: Michel Lespinasse <walken@google.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Anton Blanchard <anton@samba.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: paul.gortmaker@windriver.com Link: http://lkml.kernel.org/r/1359716356-23865-1-git-send-email-yuanhan.liu@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-02-01 10:59:16 +00:00
/* got the lock */
sem->count = -1;
locking/rwsem: Introduce basis for down_write_killable() Introduce a generic implementation necessary for down_write_killable(). This is a trivial extension of the already existing down_write() call which can be interrupted by SIGKILL. This patch doesn't provide down_write_killable() yet because arches have to provide the necessary pieces before. rwsem_down_write_failed() which is a generic slow path for the write lock is extended to take a task state and renamed to __rwsem_down_write_failed_common(). The return value is either a valid semaphore pointer or ERR_PTR(-EINTR). rwsem_down_write_failed_killable() is exported as a new way to wait for the lock and be killable. For rwsem-spinlock implementation the current __down_write() it updated in a similar way as __rwsem_down_write_failed_common() except it doesn't need new exports just visible __down_write_killable(). Architectures which are not using the generic rwsem implementation are supposed to provide their __down_write_killable() implementation and use rwsem_down_write_failed_killable() for the slow path. Signed-off-by: Michal Hocko <mhocko@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Chris Zankel <chris@zankel.net> Cc: David S. Miller <davem@davemloft.net> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Max Filippov <jcmvbkbc@gmail.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Cc: Signed-off-by: Jason Low <jason.low2@hp.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Tony Luck <tony.luck@intel.com> Cc: linux-alpha@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-ia64@vger.kernel.org Cc: linux-s390@vger.kernel.org Cc: linux-sh@vger.kernel.org Cc: linux-xtensa@linux-xtensa.org Cc: sparclinux@vger.kernel.org Link: http://lkml.kernel.org/r/1460041951-22347-7-git-send-email-mhocko@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2016-04-07 15:12:26 +00:00
out:
rwsem-spinlock: Implement writer lock-stealing for better scalability We (Linux Kernel Performance project) found a regression introduced by commit: 5a505085f043 mm/rmap: Convert the struct anon_vma::mutex to an rwsem which converted all anon_vma::mutex locks rwsem write locks. The semantics are the same, but the behavioral difference is quite huge in some cases. After investigating it we found the root cause: mutexes support lock stealing while rwsems don't. Here is the link for the detailed regression report: https://lkml.org/lkml/2013/1/29/84 Ingo suggested adding write lock stealing to rwsems: "I think we should allow lock-steal between rwsem writers - that will not hurt fairness as most rwsem fairness concerns relate to reader vs. writer fairness" And here is the rwsem-spinlock version. With this patch, we got a double performance increase in one test box with following aim7 workfile: FILESIZE: 1M POOLSIZE: 10M 10 fork_test /usr/bin/time output w/o patch /usr/bin/time_output with patch -- Percent of CPU this job got: 369% Percent of CPU this job got: 537% Voluntary context switches: 640595016 Voluntary context switches: 157915561 We got a 45% increase in CPU usage and saved about 3/4 voluntary context switches. Reported-by: LKP project <lkp@linux.intel.com> Suggested-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com> Cc: Alex Shi <alex.shi@intel.com> Cc: David Howells <dhowells@redhat.com> Cc: Michel Lespinasse <walken@google.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Anton Blanchard <anton@samba.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: paul.gortmaker@windriver.com Link: http://lkml.kernel.org/r/1359716356-23865-1-git-send-email-yuanhan.liu@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-02-01 10:59:16 +00:00
list_del(&waiter.list);
rwsem-spinlock: Implement writer lock-stealing for better scalability We (Linux Kernel Performance project) found a regression introduced by commit: 5a505085f043 mm/rmap: Convert the struct anon_vma::mutex to an rwsem which converted all anon_vma::mutex locks rwsem write locks. The semantics are the same, but the behavioral difference is quite huge in some cases. After investigating it we found the root cause: mutexes support lock stealing while rwsems don't. Here is the link for the detailed regression report: https://lkml.org/lkml/2013/1/29/84 Ingo suggested adding write lock stealing to rwsems: "I think we should allow lock-steal between rwsem writers - that will not hurt fairness as most rwsem fairness concerns relate to reader vs. writer fairness" And here is the rwsem-spinlock version. With this patch, we got a double performance increase in one test box with following aim7 workfile: FILESIZE: 1M POOLSIZE: 10M 10 fork_test /usr/bin/time output w/o patch /usr/bin/time_output with patch -- Percent of CPU this job got: 369% Percent of CPU this job got: 537% Voluntary context switches: 640595016 Voluntary context switches: 157915561 We got a 45% increase in CPU usage and saved about 3/4 voluntary context switches. Reported-by: LKP project <lkp@linux.intel.com> Suggested-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com> Cc: Alex Shi <alex.shi@intel.com> Cc: David Howells <dhowells@redhat.com> Cc: Michel Lespinasse <walken@google.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Anton Blanchard <anton@samba.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: paul.gortmaker@windriver.com Link: http://lkml.kernel.org/r/1359716356-23865-1-git-send-email-yuanhan.liu@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-02-01 10:59:16 +00:00
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
locking/rwsem: Introduce basis for down_write_killable() Introduce a generic implementation necessary for down_write_killable(). This is a trivial extension of the already existing down_write() call which can be interrupted by SIGKILL. This patch doesn't provide down_write_killable() yet because arches have to provide the necessary pieces before. rwsem_down_write_failed() which is a generic slow path for the write lock is extended to take a task state and renamed to __rwsem_down_write_failed_common(). The return value is either a valid semaphore pointer or ERR_PTR(-EINTR). rwsem_down_write_failed_killable() is exported as a new way to wait for the lock and be killable. For rwsem-spinlock implementation the current __down_write() it updated in a similar way as __rwsem_down_write_failed_common() except it doesn't need new exports just visible __down_write_killable(). Architectures which are not using the generic rwsem implementation are supposed to provide their __down_write_killable() implementation and use rwsem_down_write_failed_killable() for the slow path. Signed-off-by: Michal Hocko <mhocko@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Chris Zankel <chris@zankel.net> Cc: David S. Miller <davem@davemloft.net> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Max Filippov <jcmvbkbc@gmail.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Cc: Signed-off-by: Jason Low <jason.low2@hp.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Tony Luck <tony.luck@intel.com> Cc: linux-alpha@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-ia64@vger.kernel.org Cc: linux-s390@vger.kernel.org Cc: linux-sh@vger.kernel.org Cc: linux-xtensa@linux-xtensa.org Cc: sparclinux@vger.kernel.org Link: http://lkml.kernel.org/r/1460041951-22347-7-git-send-email-mhocko@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2016-04-07 15:12:26 +00:00
return ret;
}
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);
}
/*
* trylock for writing -- returns 1 if successful, 0 if contention
*/
int __down_write_trylock(struct rw_semaphore *sem)
{
unsigned long flags;
int ret = 0;
raw_spin_lock_irqsave(&sem->wait_lock, flags);
if (sem->count == 0) {
rwsem-spinlock: Implement writer lock-stealing for better scalability We (Linux Kernel Performance project) found a regression introduced by commit: 5a505085f043 mm/rmap: Convert the struct anon_vma::mutex to an rwsem which converted all anon_vma::mutex locks rwsem write locks. The semantics are the same, but the behavioral difference is quite huge in some cases. After investigating it we found the root cause: mutexes support lock stealing while rwsems don't. Here is the link for the detailed regression report: https://lkml.org/lkml/2013/1/29/84 Ingo suggested adding write lock stealing to rwsems: "I think we should allow lock-steal between rwsem writers - that will not hurt fairness as most rwsem fairness concerns relate to reader vs. writer fairness" And here is the rwsem-spinlock version. With this patch, we got a double performance increase in one test box with following aim7 workfile: FILESIZE: 1M POOLSIZE: 10M 10 fork_test /usr/bin/time output w/o patch /usr/bin/time_output with patch -- Percent of CPU this job got: 369% Percent of CPU this job got: 537% Voluntary context switches: 640595016 Voluntary context switches: 157915561 We got a 45% increase in CPU usage and saved about 3/4 voluntary context switches. Reported-by: LKP project <lkp@linux.intel.com> Suggested-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com> Cc: Alex Shi <alex.shi@intel.com> Cc: David Howells <dhowells@redhat.com> Cc: Michel Lespinasse <walken@google.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Anton Blanchard <anton@samba.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: paul.gortmaker@windriver.com Link: http://lkml.kernel.org/r/1359716356-23865-1-git-send-email-yuanhan.liu@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-02-01 10:59:16 +00:00
/* got the lock */
sem->count = -1;
ret = 1;
}
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
return ret;
}
/*
* release a read lock on the semaphore
*/
void __up_read(struct rw_semaphore *sem)
{
unsigned long flags;
raw_spin_lock_irqsave(&sem->wait_lock, flags);
if (--sem->count == 0 && !list_empty(&sem->wait_list))
sem = __rwsem_wake_one_writer(sem);
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
}
/*
* release a write lock on the semaphore
*/
void __up_write(struct rw_semaphore *sem)
{
unsigned long flags;
raw_spin_lock_irqsave(&sem->wait_lock, flags);
sem->count = 0;
if (!list_empty(&sem->wait_list))
sem = __rwsem_do_wake(sem, 1);
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
}
/*
* downgrade a write lock into a read lock
* - just wake up any readers at the front of the queue
*/
void __downgrade_write(struct rw_semaphore *sem)
{
unsigned long flags;
raw_spin_lock_irqsave(&sem->wait_lock, flags);
sem->count = 1;
if (!list_empty(&sem->wait_list))
sem = __rwsem_do_wake(sem, 0);
raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
}