OpenCloudOS-Kernel/kernel/locking/rtmutex.c

1924 lines
51 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* RT-Mutexes: simple blocking mutual exclusion locks with PI support
*
* started by Ingo Molnar and Thomas Gleixner.
*
* Copyright (C) 2004-2006 Red Hat, Inc., Ingo Molnar <mingo@redhat.com>
* Copyright (C) 2005-2006 Timesys Corp., Thomas Gleixner <tglx@timesys.com>
* Copyright (C) 2005 Kihon Technologies Inc., Steven Rostedt
* Copyright (C) 2006 Esben Nielsen
*
* See Documentation/locking/rt-mutex-design.rst for details.
*/
#include <linux/spinlock.h>
#include <linux/export.h>
#include <linux/sched/signal.h>
#include <linux/sched/rt.h>
#include <linux/sched/deadline.h>
#include <linux/sched/wake_q.h>
#include <linux/sched/debug.h>
#include <linux/timer.h>
#include "rtmutex_common.h"
/*
* lock->owner state tracking:
*
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
* lock->owner holds the task_struct pointer of the owner. Bit 0
* is used to keep track of the "lock has waiters" state.
*
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
* owner bit0
* NULL 0 lock is free (fast acquire possible)
* NULL 1 lock is free and has waiters and the top waiter
* is going to take the lock*
* taskpointer 0 lock is held (fast release possible)
* taskpointer 1 lock is held and has waiters**
*
* The fast atomic compare exchange based acquire and release is only
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
* possible when bit 0 of lock->owner is 0.
*
* (*) It also can be a transitional state when grabbing the lock
* with ->wait_lock is held. To prevent any fast path cmpxchg to the lock,
* we need to set the bit0 before looking at the lock, and the owner may be
* NULL in this small time, hence this can be a transitional state.
*
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
* (**) There is a small time when bit 0 is set but there are no
* waiters. This can happen when grabbing the lock in the slow path.
* To prevent a cmpxchg of the owner releasing the lock, we need to
* set this bit before looking at the lock.
*/
static void
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
rt_mutex_set_owner(struct rt_mutex *lock, struct task_struct *owner)
{
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
unsigned long val = (unsigned long)owner;
if (rt_mutex_has_waiters(lock))
val |= RT_MUTEX_HAS_WAITERS;
WRITE_ONCE(lock->owner, (struct task_struct *)val);
}
static inline void clear_rt_mutex_waiters(struct rt_mutex *lock)
{
lock->owner = (struct task_struct *)
((unsigned long)lock->owner & ~RT_MUTEX_HAS_WAITERS);
}
static void fixup_rt_mutex_waiters(struct rt_mutex *lock)
{
locking/rtmutex: Prevent dequeue vs. unlock race David reported a futex/rtmutex state corruption. It's caused by the following problem: CPU0 CPU1 CPU2 l->owner=T1 rt_mutex_lock(l) lock(l->wait_lock) l->owner = T1 | HAS_WAITERS; enqueue(T2) boost() unlock(l->wait_lock) schedule() rt_mutex_lock(l) lock(l->wait_lock) l->owner = T1 | HAS_WAITERS; enqueue(T3) boost() unlock(l->wait_lock) schedule() signal(->T2) signal(->T3) lock(l->wait_lock) dequeue(T2) deboost() unlock(l->wait_lock) lock(l->wait_lock) dequeue(T3) ===> wait list is now empty deboost() unlock(l->wait_lock) lock(l->wait_lock) fixup_rt_mutex_waiters() if (wait_list_empty(l)) { owner = l->owner & ~HAS_WAITERS; l->owner = owner ==> l->owner = T1 } lock(l->wait_lock) rt_mutex_unlock(l) fixup_rt_mutex_waiters() if (wait_list_empty(l)) { owner = l->owner & ~HAS_WAITERS; cmpxchg(l->owner, T1, NULL) ===> Success (l->owner = NULL) l->owner = owner ==> l->owner = T1 } That means the problem is caused by fixup_rt_mutex_waiters() which does the RMW to clear the waiters bit unconditionally when there are no waiters in the rtmutexes rbtree. This can be fatal: A concurrent unlock can release the rtmutex in the fastpath because the waiters bit is not set. If the cmpxchg() gets in the middle of the RMW operation then the previous owner, which just unlocked the rtmutex is set as the owner again when the write takes place after the successfull cmpxchg(). The solution is rather trivial: verify that the owner member of the rtmutex has the waiters bit set before clearing it. This does not require a cmpxchg() or other atomic operations because the waiters bit can only be set and cleared with the rtmutex wait_lock held. It's also safe against the fast path unlock attempt. The unlock attempt via cmpxchg() will either see the bit set and take the slowpath or see the bit cleared and release it atomically in the fastpath. It's remarkable that the test program provided by David triggers on ARM64 and MIPS64 really quick, but it refuses to reproduce on x86-64, while the problem exists there as well. That refusal might explain that this got not discovered earlier despite the bug existing from day one of the rtmutex implementation more than 10 years ago. Thanks to David for meticulously instrumenting the code and providing the information which allowed to decode this subtle problem. Reported-by: David Daney <ddaney@caviumnetworks.com> Tested-by: David Daney <david.daney@cavium.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Sebastian Siewior <bigeasy@linutronix.de> Cc: Will Deacon <will.deacon@arm.com> Cc: stable@vger.kernel.org Fixes: 23f78d4a03c5 ("[PATCH] pi-futex: rt mutex core") Link: http://lkml.kernel.org/r/20161130210030.351136722@linutronix.de Signed-off-by: Ingo Molnar <mingo@kernel.org>
2016-12-01 05:04:41 +08:00
unsigned long owner, *p = (unsigned long *) &lock->owner;
if (rt_mutex_has_waiters(lock))
return;
/*
* The rbtree has no waiters enqueued, now make sure that the
* lock->owner still has the waiters bit set, otherwise the
* following can happen:
*
* CPU 0 CPU 1 CPU2
* l->owner=T1
* rt_mutex_lock(l)
* lock(l->lock)
* l->owner = T1 | HAS_WAITERS;
* enqueue(T2)
* boost()
* unlock(l->lock)
* block()
*
* rt_mutex_lock(l)
* lock(l->lock)
* l->owner = T1 | HAS_WAITERS;
* enqueue(T3)
* boost()
* unlock(l->lock)
* block()
* signal(->T2) signal(->T3)
* lock(l->lock)
* dequeue(T2)
* deboost()
* unlock(l->lock)
* lock(l->lock)
* dequeue(T3)
* ==> wait list is empty
* deboost()
* unlock(l->lock)
* lock(l->lock)
* fixup_rt_mutex_waiters()
* if (wait_list_empty(l) {
* l->owner = owner
* owner = l->owner & ~HAS_WAITERS;
* ==> l->owner = T1
* }
* lock(l->lock)
* rt_mutex_unlock(l) fixup_rt_mutex_waiters()
* if (wait_list_empty(l) {
* owner = l->owner & ~HAS_WAITERS;
* cmpxchg(l->owner, T1, NULL)
* ===> Success (l->owner = NULL)
*
* l->owner = owner
* ==> l->owner = T1
* }
*
* With the check for the waiter bit in place T3 on CPU2 will not
* overwrite. All tasks fiddling with the waiters bit are
* serialized by l->lock, so nothing else can modify the waiters
* bit. If the bit is set then nothing can change l->owner either
* so the simple RMW is safe. The cmpxchg() will simply fail if it
* happens in the middle of the RMW because the waiters bit is
* still set.
*/
owner = READ_ONCE(*p);
if (owner & RT_MUTEX_HAS_WAITERS)
WRITE_ONCE(*p, owner & ~RT_MUTEX_HAS_WAITERS);
}
/*
locking/rtmutex: Drop usage of __HAVE_ARCH_CMPXCHG The rtmutex code is the only user of __HAVE_ARCH_CMPXCHG and we have a few other user of cmpxchg() which do not care about __HAVE_ARCH_CMPXCHG. This define was first introduced in 23f78d4a0 ("[PATCH] pi-futex: rt mutex core") which is v2.6.18. The generic cmpxchg was introduced later in 068fbad288 ("Add cmpxchg_local to asm-generic for per cpu atomic operations") which is v2.6.25. Back then something was required to get rtmutex working with the fast path on architectures without cmpxchg and this seems to be the result. It popped up recently on rt-users because ARM (v6+) does not define __HAVE_ARCH_CMPXCHG (even that it implements it) which results in slower locking performance in the fast path. To put some numbers on it: preempt -RT, am335x, 10 loops of 100000 invocations of rt_spin_lock() + rt_spin_unlock() (time "total" is the average of the 10 loops for the 100000 invocations, "loop" is "total / 100000 * 1000"): cmpxchg | slowpath used || cmpxchg used | total | loop || total | loop --------|-----------|-------||------------|------- ARMv6 | 9129.4 us | 91 ns || 3311.9 us | 33 ns generic | 9360.2 us | 94 ns || 10834.6 us | 108 ns ----------------------------||-------------------- Forcing it to generic cmpxchg() made things worse for the slowpath and even worse in cmpxchg() path. It boils down to 14ns more per lock+unlock in a cache hot loop so it might not be that much in real world. The last test was a substitute for pre ARMv6 machine but then I was able to perform the comparison on imx28 which is ARMv5 and therefore is always is using the generic cmpxchg implementation. And the numbers: | total | loop -------- |----------- |-------- slowpath | 263937.2 us | 2639 ns cmpxchg | 16934.2 us | 169 ns -------------------------------- The numbers are larger since the machine is slower in general. However, letting rtmutex use cmpxchg() instead the slowpath seem to improve things. Since from the ARM (tested on am335x + imx28) point of view always using cmpxchg() in rt_mutex_lock() + rt_mutex_unlock() makes sense I would drop the define. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: will.deacon@arm.com Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/20150225175613.GE6823@linutronix.de Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-02-26 01:56:13 +08:00
* We can speed up the acquire/release, if there's no debugging state to be
* set up.
*/
locking/rtmutex: Drop usage of __HAVE_ARCH_CMPXCHG The rtmutex code is the only user of __HAVE_ARCH_CMPXCHG and we have a few other user of cmpxchg() which do not care about __HAVE_ARCH_CMPXCHG. This define was first introduced in 23f78d4a0 ("[PATCH] pi-futex: rt mutex core") which is v2.6.18. The generic cmpxchg was introduced later in 068fbad288 ("Add cmpxchg_local to asm-generic for per cpu atomic operations") which is v2.6.25. Back then something was required to get rtmutex working with the fast path on architectures without cmpxchg and this seems to be the result. It popped up recently on rt-users because ARM (v6+) does not define __HAVE_ARCH_CMPXCHG (even that it implements it) which results in slower locking performance in the fast path. To put some numbers on it: preempt -RT, am335x, 10 loops of 100000 invocations of rt_spin_lock() + rt_spin_unlock() (time "total" is the average of the 10 loops for the 100000 invocations, "loop" is "total / 100000 * 1000"): cmpxchg | slowpath used || cmpxchg used | total | loop || total | loop --------|-----------|-------||------------|------- ARMv6 | 9129.4 us | 91 ns || 3311.9 us | 33 ns generic | 9360.2 us | 94 ns || 10834.6 us | 108 ns ----------------------------||-------------------- Forcing it to generic cmpxchg() made things worse for the slowpath and even worse in cmpxchg() path. It boils down to 14ns more per lock+unlock in a cache hot loop so it might not be that much in real world. The last test was a substitute for pre ARMv6 machine but then I was able to perform the comparison on imx28 which is ARMv5 and therefore is always is using the generic cmpxchg implementation. And the numbers: | total | loop -------- |----------- |-------- slowpath | 263937.2 us | 2639 ns cmpxchg | 16934.2 us | 169 ns -------------------------------- The numbers are larger since the machine is slower in general. However, letting rtmutex use cmpxchg() instead the slowpath seem to improve things. Since from the ARM (tested on am335x + imx28) point of view always using cmpxchg() in rt_mutex_lock() + rt_mutex_unlock() makes sense I would drop the define. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: will.deacon@arm.com Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/20150225175613.GE6823@linutronix.de Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-02-26 01:56:13 +08:00
#ifndef CONFIG_DEBUG_RT_MUTEXES
# define rt_mutex_cmpxchg_relaxed(l,c,n) (cmpxchg_relaxed(&l->owner, c, n) == c)
# define rt_mutex_cmpxchg_acquire(l,c,n) (cmpxchg_acquire(&l->owner, c, n) == c)
# define rt_mutex_cmpxchg_release(l,c,n) (cmpxchg_release(&l->owner, c, n) == c)
/*
* Callers must hold the ->wait_lock -- which is the whole purpose as we force
* all future threads that attempt to [Rmw] the lock to the slowpath. As such
* relaxed semantics suffice.
*/
static inline void mark_rt_mutex_waiters(struct rt_mutex *lock)
{
unsigned long owner, *p = (unsigned long *) &lock->owner;
do {
owner = *p;
} while (cmpxchg_relaxed(p, owner,
owner | RT_MUTEX_HAS_WAITERS) != owner);
}
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
/*
* Safe fastpath aware unlock:
* 1) Clear the waiters bit
* 2) Drop lock->wait_lock
* 3) Try to unlock the lock with cmpxchg
*/
static inline bool unlock_rt_mutex_safe(struct rt_mutex *lock,
unsigned long flags)
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
__releases(lock->wait_lock)
{
struct task_struct *owner = rt_mutex_owner(lock);
clear_rt_mutex_waiters(lock);
raw_spin_unlock_irqrestore(&lock->wait_lock, flags);
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
/*
* If a new waiter comes in between the unlock and the cmpxchg
* we have two situations:
*
* unlock(wait_lock);
* lock(wait_lock);
* cmpxchg(p, owner, 0) == owner
* mark_rt_mutex_waiters(lock);
* acquire(lock);
* or:
*
* unlock(wait_lock);
* lock(wait_lock);
* mark_rt_mutex_waiters(lock);
*
* cmpxchg(p, owner, 0) != owner
* enqueue_waiter();
* unlock(wait_lock);
* lock(wait_lock);
* wake waiter();
* unlock(wait_lock);
* lock(wait_lock);
* acquire(lock);
*/
return rt_mutex_cmpxchg_release(lock, owner, NULL);
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
}
#else
# define rt_mutex_cmpxchg_relaxed(l,c,n) (0)
# define rt_mutex_cmpxchg_acquire(l,c,n) (0)
# define rt_mutex_cmpxchg_release(l,c,n) (0)
static inline void mark_rt_mutex_waiters(struct rt_mutex *lock)
{
lock->owner = (struct task_struct *)
((unsigned long)lock->owner | RT_MUTEX_HAS_WAITERS);
}
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
/*
* Simple slow path only version: lock->owner is protected by lock->wait_lock.
*/
static inline bool unlock_rt_mutex_safe(struct rt_mutex *lock,
unsigned long flags)
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
__releases(lock->wait_lock)
{
lock->owner = NULL;
raw_spin_unlock_irqrestore(&lock->wait_lock, flags);
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
return true;
}
#endif
/*
* Only use with rt_mutex_waiter_{less,equal}()
*/
#define task_to_waiter(p) \
&(struct rt_mutex_waiter){ .prio = (p)->prio, .deadline = (p)->dl.deadline }
static inline int
rt_mutex_waiter_less(struct rt_mutex_waiter *left,
struct rt_mutex_waiter *right)
{
sched/deadline: Add SCHED_DEADLINE inheritance logic Some method to deal with rt-mutexes and make sched_dl interact with the current PI-coded is needed, raising all but trivial issues, that needs (according to us) to be solved with some restructuring of the pi-code (i.e., going toward a proxy execution-ish implementation). This is under development, in the meanwhile, as a temporary solution, what this commits does is: - ensure a pi-lock owner with waiters is never throttled down. Instead, when it runs out of runtime, it immediately gets replenished and it's deadline is postponed; - the scheduling parameters (relative deadline and default runtime) used for that replenishments --during the whole period it holds the pi-lock-- are the ones of the waiting task with earliest deadline. Acting this way, we provide some kind of boosting to the lock-owner, still by using the existing (actually, slightly modified by the previous commit) pi-architecture. We would stress the fact that this is only a surely needed, all but clean solution to the problem. In the end it's only a way to re-start discussion within the community. So, as always, comments, ideas, rants, etc.. are welcome! :-) Signed-off-by: Dario Faggioli <raistlin@linux.it> Signed-off-by: Juri Lelli <juri.lelli@gmail.com> [ Added !RT_MUTEXES build fix. ] Signed-off-by: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1383831828-15501-11-git-send-email-juri.lelli@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-11-07 21:43:44 +08:00
if (left->prio < right->prio)
return 1;
/*
sched/deadline: Add SCHED_DEADLINE inheritance logic Some method to deal with rt-mutexes and make sched_dl interact with the current PI-coded is needed, raising all but trivial issues, that needs (according to us) to be solved with some restructuring of the pi-code (i.e., going toward a proxy execution-ish implementation). This is under development, in the meanwhile, as a temporary solution, what this commits does is: - ensure a pi-lock owner with waiters is never throttled down. Instead, when it runs out of runtime, it immediately gets replenished and it's deadline is postponed; - the scheduling parameters (relative deadline and default runtime) used for that replenishments --during the whole period it holds the pi-lock-- are the ones of the waiting task with earliest deadline. Acting this way, we provide some kind of boosting to the lock-owner, still by using the existing (actually, slightly modified by the previous commit) pi-architecture. We would stress the fact that this is only a surely needed, all but clean solution to the problem. In the end it's only a way to re-start discussion within the community. So, as always, comments, ideas, rants, etc.. are welcome! :-) Signed-off-by: Dario Faggioli <raistlin@linux.it> Signed-off-by: Juri Lelli <juri.lelli@gmail.com> [ Added !RT_MUTEXES build fix. ] Signed-off-by: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1383831828-15501-11-git-send-email-juri.lelli@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-11-07 21:43:44 +08:00
* If both waiters have dl_prio(), we check the deadlines of the
* associated tasks.
* If left waiter has a dl_prio(), and we didn't return 1 above,
* then right waiter has a dl_prio() too.
*/
sched/deadline: Add SCHED_DEADLINE inheritance logic Some method to deal with rt-mutexes and make sched_dl interact with the current PI-coded is needed, raising all but trivial issues, that needs (according to us) to be solved with some restructuring of the pi-code (i.e., going toward a proxy execution-ish implementation). This is under development, in the meanwhile, as a temporary solution, what this commits does is: - ensure a pi-lock owner with waiters is never throttled down. Instead, when it runs out of runtime, it immediately gets replenished and it's deadline is postponed; - the scheduling parameters (relative deadline and default runtime) used for that replenishments --during the whole period it holds the pi-lock-- are the ones of the waiting task with earliest deadline. Acting this way, we provide some kind of boosting to the lock-owner, still by using the existing (actually, slightly modified by the previous commit) pi-architecture. We would stress the fact that this is only a surely needed, all but clean solution to the problem. In the end it's only a way to re-start discussion within the community. So, as always, comments, ideas, rants, etc.. are welcome! :-) Signed-off-by: Dario Faggioli <raistlin@linux.it> Signed-off-by: Juri Lelli <juri.lelli@gmail.com> [ Added !RT_MUTEXES build fix. ] Signed-off-by: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1383831828-15501-11-git-send-email-juri.lelli@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-11-07 21:43:44 +08:00
if (dl_prio(left->prio))
return dl_time_before(left->deadline, right->deadline);
return 0;
}
static inline int
rt_mutex_waiter_equal(struct rt_mutex_waiter *left,
struct rt_mutex_waiter *right)
{
if (left->prio != right->prio)
return 0;
/*
* If both waiters have dl_prio(), we check the deadlines of the
* associated tasks.
* If left waiter has a dl_prio(), and we didn't return 0 above,
* then right waiter has a dl_prio() too.
*/
if (dl_prio(left->prio))
return left->deadline == right->deadline;
return 1;
}
static void
rt_mutex_enqueue(struct rt_mutex *lock, struct rt_mutex_waiter *waiter)
{
struct rb_node **link = &lock->waiters.rb_root.rb_node;
struct rb_node *parent = NULL;
struct rt_mutex_waiter *entry;
bool leftmost = true;
while (*link) {
parent = *link;
entry = rb_entry(parent, struct rt_mutex_waiter, tree_entry);
if (rt_mutex_waiter_less(waiter, entry)) {
link = &parent->rb_left;
} else {
link = &parent->rb_right;
leftmost = false;
}
}
rb_link_node(&waiter->tree_entry, parent, link);
rb_insert_color_cached(&waiter->tree_entry, &lock->waiters, leftmost);
}
static void
rt_mutex_dequeue(struct rt_mutex *lock, struct rt_mutex_waiter *waiter)
{
if (RB_EMPTY_NODE(&waiter->tree_entry))
return;
rb_erase_cached(&waiter->tree_entry, &lock->waiters);
RB_CLEAR_NODE(&waiter->tree_entry);
}
static void
rt_mutex_enqueue_pi(struct task_struct *task, struct rt_mutex_waiter *waiter)
{
struct rb_node **link = &task->pi_waiters.rb_root.rb_node;
struct rb_node *parent = NULL;
struct rt_mutex_waiter *entry;
bool leftmost = true;
while (*link) {
parent = *link;
entry = rb_entry(parent, struct rt_mutex_waiter, pi_tree_entry);
if (rt_mutex_waiter_less(waiter, entry)) {
link = &parent->rb_left;
} else {
link = &parent->rb_right;
leftmost = false;
}
}
rb_link_node(&waiter->pi_tree_entry, parent, link);
rb_insert_color_cached(&waiter->pi_tree_entry, &task->pi_waiters, leftmost);
}
static void
rt_mutex_dequeue_pi(struct task_struct *task, struct rt_mutex_waiter *waiter)
{
if (RB_EMPTY_NODE(&waiter->pi_tree_entry))
return;
rb_erase_cached(&waiter->pi_tree_entry, &task->pi_waiters);
RB_CLEAR_NODE(&waiter->pi_tree_entry);
}
static void rt_mutex_adjust_prio(struct task_struct *p)
{
struct task_struct *pi_task = NULL;
sched/rtmutex/deadline: Fix a PI crash for deadline tasks A crash happened while I was playing with deadline PI rtmutex. BUG: unable to handle kernel NULL pointer dereference at 0000000000000018 IP: [<ffffffff810eeb8f>] rt_mutex_get_top_task+0x1f/0x30 PGD 232a75067 PUD 230947067 PMD 0 Oops: 0000 [#1] SMP CPU: 1 PID: 10994 Comm: a.out Not tainted Call Trace: [<ffffffff810b658c>] enqueue_task+0x2c/0x80 [<ffffffff810ba763>] activate_task+0x23/0x30 [<ffffffff810d0ab5>] pull_dl_task+0x1d5/0x260 [<ffffffff810d0be6>] pre_schedule_dl+0x16/0x20 [<ffffffff8164e783>] __schedule+0xd3/0x900 [<ffffffff8164efd9>] schedule+0x29/0x70 [<ffffffff8165035b>] __rt_mutex_slowlock+0x4b/0xc0 [<ffffffff81650501>] rt_mutex_slowlock+0xd1/0x190 [<ffffffff810eeb33>] rt_mutex_timed_lock+0x53/0x60 [<ffffffff810ecbfc>] futex_lock_pi.isra.18+0x28c/0x390 [<ffffffff810ed8b0>] do_futex+0x190/0x5b0 [<ffffffff810edd50>] SyS_futex+0x80/0x180 This is because rt_mutex_enqueue_pi() and rt_mutex_dequeue_pi() are only protected by pi_lock when operating pi waiters, while rt_mutex_get_top_task(), will access them with rq lock held but not holding pi_lock. In order to tackle it, we introduce new "pi_top_task" pointer cached in task_struct, and add new rt_mutex_update_top_task() to update its value, it can be called by rt_mutex_setprio() which held both owner's pi_lock and rq lock. Thus "pi_top_task" can be safely accessed by enqueue_task_dl() under rq lock. Originally-From: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Xunlei Pang <xlpang@redhat.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: Steven Rostedt <rostedt@goodmis.org> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Cc: juri.lelli@arm.com Cc: bigeasy@linutronix.de Cc: mathieu.desnoyers@efficios.com Cc: jdesfossez@efficios.com Cc: bristot@redhat.com Link: http://lkml.kernel.org/r/20170323150216.157682758@infradead.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-03-23 22:56:08 +08:00
lockdep_assert_held(&p->pi_lock);
if (task_has_pi_waiters(p))
pi_task = task_top_pi_waiter(p)->task;
rt_mutex_setprio(p, pi_task);
}
/*
* Deadlock detection is conditional:
*
* If CONFIG_DEBUG_RT_MUTEXES=n, deadlock detection is only conducted
* if the detect argument is == RT_MUTEX_FULL_CHAINWALK.
*
* If CONFIG_DEBUG_RT_MUTEXES=y, deadlock detection is always
* conducted independent of the detect argument.
*
* If the waiter argument is NULL this indicates the deboost path and
* deadlock detection is disabled independent of the detect argument
* and the config settings.
*/
static bool rt_mutex_cond_detect_deadlock(struct rt_mutex_waiter *waiter,
enum rtmutex_chainwalk chwalk)
{
/*
* This is just a wrapper function for the following call,
* because debug_rt_mutex_detect_deadlock() smells like a magic
* debug feature and I wanted to keep the cond function in the
* main source file along with the comments instead of having
* two of the same in the headers.
*/
return debug_rt_mutex_detect_deadlock(waiter, chwalk);
}
/*
* Max number of times we'll walk the boosting chain:
*/
int max_lock_depth = 1024;
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
static inline struct rt_mutex *task_blocked_on_lock(struct task_struct *p)
{
return p->pi_blocked_on ? p->pi_blocked_on->lock : NULL;
}
/*
* Adjust the priority chain. Also used for deadlock detection.
* Decreases task's usage by one - may thus free the task.
*
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
* @task: the task owning the mutex (owner) for which a chain walk is
* probably needed
* @chwalk: do we have to carry out deadlock detection?
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
* @orig_lock: the mutex (can be NULL if we are walking the chain to recheck
* things for a task that has just got its priority adjusted, and
* is waiting on a mutex)
* @next_lock: the mutex on which the owner of @orig_lock was blocked before
* we dropped its pi_lock. Is never dereferenced, only used for
* comparison to detect lock chain changes.
* @orig_waiter: rt_mutex_waiter struct for the task that has just donated
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
* its priority to the mutex owner (can be NULL in the case
* depicted above or if the top waiter is gone away and we are
* actually deboosting the owner)
* @top_task: the current top waiter
*
* Returns 0 or -EDEADLK.
*
* Chain walk basics and protection scope
*
* [R] refcount on task
* [P] task->pi_lock held
* [L] rtmutex->wait_lock held
*
* Step Description Protected by
* function arguments:
* @task [R]
* @orig_lock if != NULL @top_task is blocked on it
* @next_lock Unprotected. Cannot be
* dereferenced. Only used for
* comparison.
* @orig_waiter if != NULL @top_task is blocked on it
* @top_task current, or in case of proxy
* locking protected by calling
* code
* again:
* loop_sanity_check();
* retry:
* [1] lock(task->pi_lock); [R] acquire [P]
* [2] waiter = task->pi_blocked_on; [P]
* [3] check_exit_conditions_1(); [P]
* [4] lock = waiter->lock; [P]
* [5] if (!try_lock(lock->wait_lock)) { [P] try to acquire [L]
* unlock(task->pi_lock); release [P]
* goto retry;
* }
* [6] check_exit_conditions_2(); [P] + [L]
* [7] requeue_lock_waiter(lock, waiter); [P] + [L]
* [8] unlock(task->pi_lock); release [P]
* put_task_struct(task); release [R]
* [9] check_exit_conditions_3(); [L]
* [10] task = owner(lock); [L]
* get_task_struct(task); [L] acquire [R]
* lock(task->pi_lock); [L] acquire [P]
* [11] requeue_pi_waiter(tsk, waiters(lock));[P] + [L]
* [12] check_exit_conditions_4(); [P] + [L]
* [13] unlock(task->pi_lock); release [P]
* unlock(lock->wait_lock); release [L]
* goto again;
*/
static int rt_mutex_adjust_prio_chain(struct task_struct *task,
enum rtmutex_chainwalk chwalk,
struct rt_mutex *orig_lock,
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
struct rt_mutex *next_lock,
struct rt_mutex_waiter *orig_waiter,
struct task_struct *top_task)
{
struct rt_mutex_waiter *waiter, *top_waiter = orig_waiter;
struct rt_mutex_waiter *prerequeue_top_waiter;
int ret = 0, depth = 0;
struct rt_mutex *lock;
bool detect_deadlock;
bool requeue = true;
detect_deadlock = rt_mutex_cond_detect_deadlock(orig_waiter, chwalk);
/*
* The (de)boosting is a step by step approach with a lot of
* pitfalls. We want this to be preemptible and we want hold a
* maximum of two locks per step. So we have to check
* carefully whether things change under us.
*/
again:
/*
* We limit the lock chain length for each invocation.
*/
if (++depth > max_lock_depth) {
static int prev_max;
/*
* Print this only once. If the admin changes the limit,
* print a new message when reaching the limit again.
*/
if (prev_max != max_lock_depth) {
prev_max = max_lock_depth;
printk(KERN_WARNING "Maximum lock depth %d reached "
"task: %s (%d)\n", max_lock_depth,
top_task->comm, task_pid_nr(top_task));
}
put_task_struct(task);
return -EDEADLK;
}
/*
* We are fully preemptible here and only hold the refcount on
* @task. So everything can have changed under us since the
* caller or our own code below (goto retry/again) dropped all
* locks.
*/
retry:
/*
* [1] Task cannot go away as we did a get_task() before !
*/
raw_spin_lock_irq(&task->pi_lock);
/*
* [2] Get the waiter on which @task is blocked on.
*/
waiter = task->pi_blocked_on;
/*
* [3] check_exit_conditions_1() protected by task->pi_lock.
*/
/*
* Check whether the end of the boosting chain has been
* reached or the state of the chain has changed while we
* dropped the locks.
*/
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
if (!waiter)
goto out_unlock_pi;
/*
* Check the orig_waiter state. After we dropped the locks,
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
* the previous owner of the lock might have released the lock.
*/
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
if (orig_waiter && !rt_mutex_owner(orig_lock))
goto out_unlock_pi;
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
/*
* We dropped all locks after taking a refcount on @task, so
* the task might have moved on in the lock chain or even left
* the chain completely and blocks now on an unrelated lock or
* on @orig_lock.
*
* We stored the lock on which @task was blocked in @next_lock,
* so we can detect the chain change.
*/
if (next_lock != waiter->lock)
goto out_unlock_pi;
/*
* Drop out, when the task has no waiters. Note,
* top_waiter can be NULL, when we are in the deboosting
* mode!
*/
if (top_waiter) {
if (!task_has_pi_waiters(task))
goto out_unlock_pi;
/*
* If deadlock detection is off, we stop here if we
* are not the top pi waiter of the task. If deadlock
* detection is enabled we continue, but stop the
* requeueing in the chain walk.
*/
if (top_waiter != task_top_pi_waiter(task)) {
if (!detect_deadlock)
goto out_unlock_pi;
else
requeue = false;
}
}
/*
* If the waiter priority is the same as the task priority
* then there is no further priority adjustment necessary. If
* deadlock detection is off, we stop the chain walk. If its
* enabled we continue, but stop the requeueing in the chain
* walk.
*/
if (rt_mutex_waiter_equal(waiter, task_to_waiter(task))) {
if (!detect_deadlock)
goto out_unlock_pi;
else
requeue = false;
}
/*
* [4] Get the next lock
*/
lock = waiter->lock;
/*
* [5] We need to trylock here as we are holding task->pi_lock,
* which is the reverse lock order versus the other rtmutex
* operations.
*/
if (!raw_spin_trylock(&lock->wait_lock)) {
raw_spin_unlock_irq(&task->pi_lock);
cpu_relax();
goto retry;
}
/*
* [6] check_exit_conditions_2() protected by task->pi_lock and
* lock->wait_lock.
*
* Deadlock detection. If the lock is the same as the original
* lock which caused us to walk the lock chain or if the
* current lock is owned by the task which initiated the chain
* walk, we detected a deadlock.
*/
if (lock == orig_lock || rt_mutex_owner(lock) == top_task) {
debug_rt_mutex_deadlock(chwalk, orig_waiter, lock);
raw_spin_unlock(&lock->wait_lock);
ret = -EDEADLK;
goto out_unlock_pi;
}
/*
* If we just follow the lock chain for deadlock detection, no
* need to do all the requeue operations. To avoid a truckload
* of conditionals around the various places below, just do the
* minimum chain walk checks.
*/
if (!requeue) {
/*
* No requeue[7] here. Just release @task [8]
*/
raw_spin_unlock(&task->pi_lock);
put_task_struct(task);
/*
* [9] check_exit_conditions_3 protected by lock->wait_lock.
* If there is no owner of the lock, end of chain.
*/
if (!rt_mutex_owner(lock)) {
raw_spin_unlock_irq(&lock->wait_lock);
return 0;
}
/* [10] Grab the next task, i.e. owner of @lock */
task = get_task_struct(rt_mutex_owner(lock));
raw_spin_lock(&task->pi_lock);
/*
* No requeue [11] here. We just do deadlock detection.
*
* [12] Store whether owner is blocked
* itself. Decision is made after dropping the locks
*/
next_lock = task_blocked_on_lock(task);
/*
* Get the top waiter for the next iteration
*/
top_waiter = rt_mutex_top_waiter(lock);
/* [13] Drop locks */
raw_spin_unlock(&task->pi_lock);
raw_spin_unlock_irq(&lock->wait_lock);
/* If owner is not blocked, end of chain. */
if (!next_lock)
goto out_put_task;
goto again;
}
/*
* Store the current top waiter before doing the requeue
* operation on @lock. We need it for the boost/deboost
* decision below.
*/
prerequeue_top_waiter = rt_mutex_top_waiter(lock);
/* [7] Requeue the waiter in the lock waiter tree. */
rt_mutex_dequeue(lock, waiter);
/*
* Update the waiter prio fields now that we're dequeued.
*
* These values can have changed through either:
*
* sys_sched_set_scheduler() / sys_sched_setattr()
*
* or
*
* DL CBS enforcement advancing the effective deadline.
*
* Even though pi_waiters also uses these fields, and that tree is only
* updated in [11], we can do this here, since we hold [L], which
* serializes all pi_waiters access and rb_erase() does not care about
* the values of the node being removed.
*/
sched/deadline: Add SCHED_DEADLINE inheritance logic Some method to deal with rt-mutexes and make sched_dl interact with the current PI-coded is needed, raising all but trivial issues, that needs (according to us) to be solved with some restructuring of the pi-code (i.e., going toward a proxy execution-ish implementation). This is under development, in the meanwhile, as a temporary solution, what this commits does is: - ensure a pi-lock owner with waiters is never throttled down. Instead, when it runs out of runtime, it immediately gets replenished and it's deadline is postponed; - the scheduling parameters (relative deadline and default runtime) used for that replenishments --during the whole period it holds the pi-lock-- are the ones of the waiting task with earliest deadline. Acting this way, we provide some kind of boosting to the lock-owner, still by using the existing (actually, slightly modified by the previous commit) pi-architecture. We would stress the fact that this is only a surely needed, all but clean solution to the problem. In the end it's only a way to re-start discussion within the community. So, as always, comments, ideas, rants, etc.. are welcome! :-) Signed-off-by: Dario Faggioli <raistlin@linux.it> Signed-off-by: Juri Lelli <juri.lelli@gmail.com> [ Added !RT_MUTEXES build fix. ] Signed-off-by: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1383831828-15501-11-git-send-email-juri.lelli@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-11-07 21:43:44 +08:00
waiter->prio = task->prio;
waiter->deadline = task->dl.deadline;
rt_mutex_enqueue(lock, waiter);
/* [8] Release the task */
raw_spin_unlock(&task->pi_lock);
put_task_struct(task);
/*
* [9] check_exit_conditions_3 protected by lock->wait_lock.
*
* We must abort the chain walk if there is no lock owner even
* in the dead lock detection case, as we have nothing to
* follow here. This is the end of the chain we are walking.
*/
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
if (!rt_mutex_owner(lock)) {
/*
* 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.
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
*/
if (prerequeue_top_waiter != rt_mutex_top_waiter(lock))
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
wake_up_process(rt_mutex_top_waiter(lock)->task);
raw_spin_unlock_irq(&lock->wait_lock);
return 0;
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
}
/* [10] Grab the next task, i.e. the owner of @lock */
task = get_task_struct(rt_mutex_owner(lock));
raw_spin_lock(&task->pi_lock);
/* [11] requeue the pi waiters if necessary */
if (waiter == rt_mutex_top_waiter(lock)) {
/*
* The waiter became the new top (highest priority)
* waiter on the lock. Replace the previous top waiter
* in the owner tasks pi waiters tree with this waiter
* and adjust the priority of the owner.
*/
rt_mutex_dequeue_pi(task, prerequeue_top_waiter);
rt_mutex_enqueue_pi(task, waiter);
rt_mutex_adjust_prio(task);
} else if (prerequeue_top_waiter == waiter) {
/*
* The waiter was the top waiter on the lock, but is
* no longer the top prority waiter. Replace waiter in
* the owner tasks pi waiters tree with the new top
* (highest priority) waiter and adjust the priority
* of the owner.
* The new top waiter is stored in @waiter so that
* @waiter == @top_waiter evaluates to true below and
* we continue to deboost the rest of the chain.
*/
rt_mutex_dequeue_pi(task, waiter);
waiter = rt_mutex_top_waiter(lock);
rt_mutex_enqueue_pi(task, waiter);
rt_mutex_adjust_prio(task);
} else {
/*
* Nothing changed. No need to do any priority
* adjustment.
*/
}
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
/*
* [12] check_exit_conditions_4() protected by task->pi_lock
* and lock->wait_lock. The actual decisions are made after we
* dropped the locks.
*
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
* Check whether the task which owns the current lock is pi
* blocked itself. If yes we store a pointer to the lock for
* the lock chain change detection above. After we dropped
* task->pi_lock next_lock cannot be dereferenced anymore.
*/
next_lock = task_blocked_on_lock(task);
/*
* Store the top waiter of @lock for the end of chain walk
* decision below.
*/
top_waiter = rt_mutex_top_waiter(lock);
/* [13] Drop the locks */
raw_spin_unlock(&task->pi_lock);
raw_spin_unlock_irq(&lock->wait_lock);
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
/*
* Make the actual exit decisions [12], based on the stored
* values.
*
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
* We reached the end of the lock chain. Stop right here. No
* point to go back just to figure that out.
*/
if (!next_lock)
goto out_put_task;
/*
* If the current waiter is not the top waiter on the lock,
* then we can stop the chain walk here if we are not in full
* deadlock detection mode.
*/
if (!detect_deadlock && waiter != top_waiter)
goto out_put_task;
goto again;
out_unlock_pi:
raw_spin_unlock_irq(&task->pi_lock);
out_put_task:
put_task_struct(task);
return ret;
}
/*
* Try to take an rt-mutex
*
* Must be called with lock->wait_lock held and interrupts disabled
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
*
* @lock: The lock to be acquired.
* @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
*/
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
static int try_to_take_rt_mutex(struct rt_mutex *lock, struct task_struct *task,
struct rt_mutex_waiter *waiter)
{
lockdep_assert_held(&lock->wait_lock);
/*
* Before testing whether we can acquire @lock, we set the
* RT_MUTEX_HAS_WAITERS bit in @lock->owner. This forces all
* other tasks which try to modify @lock into the slow path
* and they serialize on @lock->wait_lock.
*
* The RT_MUTEX_HAS_WAITERS bit can have a transitional state
* as explained at the top of this file if and only if:
*
* - There is a lock owner. The caller must fixup the
* transient state if it does a trylock or leaves the lock
* function due to a signal or timeout.
*
* - @task acquires the lock and there are no other
* waiters. This is undone in rt_mutex_set_owner(@task) at
* the end of this function.
*/
mark_rt_mutex_waiters(lock);
/*
* If @lock has an owner, give up.
*/
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
if (rt_mutex_owner(lock))
return 0;
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
/*
* If @waiter != NULL, @task has already enqueued the waiter
* into @lock waiter tree. If @waiter == NULL then this is a
* trylock attempt.
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
*/
if (waiter) {
/*
* If waiter is not the highest priority waiter of
* @lock, give up.
*/
if (waiter != rt_mutex_top_waiter(lock))
return 0;
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
/*
* We can acquire the lock. Remove the waiter from the
* lock waiters tree.
*/
rt_mutex_dequeue(lock, waiter);
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
} else {
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
/*
* If the lock has waiters already we check whether @task is
* eligible to take over the lock.
*
* If there are no other waiters, @task can acquire
* the lock. @task->pi_blocked_on is NULL, so it does
* not need to be dequeued.
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
*/
if (rt_mutex_has_waiters(lock)) {
/*
* If @task->prio is greater than or equal to
* the top waiter priority (kernel view),
* @task lost.
*/
if (!rt_mutex_waiter_less(task_to_waiter(task),
rt_mutex_top_waiter(lock)))
return 0;
/*
* The current top waiter stays enqueued. We
* don't have to change anything in the lock
* waiters order.
*/
} else {
/*
* No waiters. Take the lock without the
* pi_lock dance.@task->pi_blocked_on is NULL
* and we have no waiters to enqueue in @task
* pi waiters tree.
*/
goto takeit;
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
}
}
/*
* Clear @task->pi_blocked_on. Requires protection by
* @task->pi_lock. Redundant operation for the @waiter == NULL
* case, but conditionals are more expensive than a redundant
* store.
*/
raw_spin_lock(&task->pi_lock);
task->pi_blocked_on = NULL;
/*
* Finish the lock acquisition. @task is the new owner. If
* other waiters exist we have to insert the highest priority
* waiter into @task->pi_waiters tree.
*/
if (rt_mutex_has_waiters(lock))
rt_mutex_enqueue_pi(task, rt_mutex_top_waiter(lock));
raw_spin_unlock(&task->pi_lock);
takeit:
/* We got the lock. */
debug_rt_mutex_lock(lock);
/*
* This either preserves the RT_MUTEX_HAS_WAITERS bit if there
* are still waiters or clears it.
*/
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
rt_mutex_set_owner(lock, task);
return 1;
}
/*
* Task blocks on lock.
*
* Prepare waiter and propagate pi chain
*
* This must be called with lock->wait_lock held and interrupts disabled
*/
static int task_blocks_on_rt_mutex(struct rt_mutex *lock,
struct rt_mutex_waiter *waiter,
struct task_struct *task,
enum rtmutex_chainwalk chwalk)
{
struct task_struct *owner = rt_mutex_owner(lock);
struct rt_mutex_waiter *top_waiter = waiter;
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
struct rt_mutex *next_lock;
[PATCH] clean up and remove some extra spinlocks from rtmutex Oleg brought up some interesting points about grabbing the pi_lock for some protections. In this discussion, I realized that there are some places that the pi_lock is being grabbed when it really wasn't necessary. Also this patch does a little bit of clean up. This patch basically does three things: 1) renames the "boost" variable to "chain_walk". Since it is used in the debugging case when it isn't going to be boosted. It better describes what the test is going to do if it succeeds. 2) moves get_task_struct to just before the unlocking of the wait_lock. This removes duplicate code, and makes it a little easier to read. The owner wont go away while either the pi_lock or the wait_lock are held. 3) removes the pi_locking and owner blocked checking completely from the debugging case. This is because the grabbing the lock and doing the check, then releasing the lock is just so full of races. It's just as good to go ahead and call the pi_chain_walk function, since after releasing the lock the owner can then block anyway, and we would have missed that. For the debug case, we really do want to do the chain walk to test for deadlocks anyway. [oleg@tv-sign.ru: more of the same] Signed-of-by: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Esben Nielsen <nielsen.esben@googlemail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 16:59:44 +08:00
int chain_walk = 0, res;
lockdep_assert_held(&lock->wait_lock);
/*
* Early deadlock detection. We really don't want the task to
* enqueue on itself just to untangle the mess later. It's not
* only an optimization. We drop the locks, so another waiter
* can come in before the chain walk detects the deadlock. So
* the other will detect the deadlock and return -EDEADLOCK,
* which is wrong, as the other waiter is not in a deadlock
* situation.
*/
if (owner == task)
return -EDEADLK;
raw_spin_lock(&task->pi_lock);
waiter->task = task;
waiter->lock = lock;
sched/deadline: Add SCHED_DEADLINE inheritance logic Some method to deal with rt-mutexes and make sched_dl interact with the current PI-coded is needed, raising all but trivial issues, that needs (according to us) to be solved with some restructuring of the pi-code (i.e., going toward a proxy execution-ish implementation). This is under development, in the meanwhile, as a temporary solution, what this commits does is: - ensure a pi-lock owner with waiters is never throttled down. Instead, when it runs out of runtime, it immediately gets replenished and it's deadline is postponed; - the scheduling parameters (relative deadline and default runtime) used for that replenishments --during the whole period it holds the pi-lock-- are the ones of the waiting task with earliest deadline. Acting this way, we provide some kind of boosting to the lock-owner, still by using the existing (actually, slightly modified by the previous commit) pi-architecture. We would stress the fact that this is only a surely needed, all but clean solution to the problem. In the end it's only a way to re-start discussion within the community. So, as always, comments, ideas, rants, etc.. are welcome! :-) Signed-off-by: Dario Faggioli <raistlin@linux.it> Signed-off-by: Juri Lelli <juri.lelli@gmail.com> [ Added !RT_MUTEXES build fix. ] Signed-off-by: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1383831828-15501-11-git-send-email-juri.lelli@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-11-07 21:43:44 +08:00
waiter->prio = task->prio;
waiter->deadline = task->dl.deadline;
/* Get the top priority waiter on the lock */
if (rt_mutex_has_waiters(lock))
top_waiter = rt_mutex_top_waiter(lock);
rt_mutex_enqueue(lock, waiter);
task->pi_blocked_on = waiter;
raw_spin_unlock(&task->pi_lock);
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
if (!owner)
return 0;
raw_spin_lock(&owner->pi_lock);
if (waiter == rt_mutex_top_waiter(lock)) {
rt_mutex_dequeue_pi(owner, top_waiter);
rt_mutex_enqueue_pi(owner, waiter);
rt_mutex_adjust_prio(owner);
[PATCH] clean up and remove some extra spinlocks from rtmutex Oleg brought up some interesting points about grabbing the pi_lock for some protections. In this discussion, I realized that there are some places that the pi_lock is being grabbed when it really wasn't necessary. Also this patch does a little bit of clean up. This patch basically does three things: 1) renames the "boost" variable to "chain_walk". Since it is used in the debugging case when it isn't going to be boosted. It better describes what the test is going to do if it succeeds. 2) moves get_task_struct to just before the unlocking of the wait_lock. This removes duplicate code, and makes it a little easier to read. The owner wont go away while either the pi_lock or the wait_lock are held. 3) removes the pi_locking and owner blocked checking completely from the debugging case. This is because the grabbing the lock and doing the check, then releasing the lock is just so full of races. It's just as good to go ahead and call the pi_chain_walk function, since after releasing the lock the owner can then block anyway, and we would have missed that. For the debug case, we really do want to do the chain walk to test for deadlocks anyway. [oleg@tv-sign.ru: more of the same] Signed-of-by: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Esben Nielsen <nielsen.esben@googlemail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 16:59:44 +08:00
if (owner->pi_blocked_on)
chain_walk = 1;
} else if (rt_mutex_cond_detect_deadlock(waiter, chwalk)) {
[PATCH] clean up and remove some extra spinlocks from rtmutex Oleg brought up some interesting points about grabbing the pi_lock for some protections. In this discussion, I realized that there are some places that the pi_lock is being grabbed when it really wasn't necessary. Also this patch does a little bit of clean up. This patch basically does three things: 1) renames the "boost" variable to "chain_walk". Since it is used in the debugging case when it isn't going to be boosted. It better describes what the test is going to do if it succeeds. 2) moves get_task_struct to just before the unlocking of the wait_lock. This removes duplicate code, and makes it a little easier to read. The owner wont go away while either the pi_lock or the wait_lock are held. 3) removes the pi_locking and owner blocked checking completely from the debugging case. This is because the grabbing the lock and doing the check, then releasing the lock is just so full of races. It's just as good to go ahead and call the pi_chain_walk function, since after releasing the lock the owner can then block anyway, and we would have missed that. For the debug case, we really do want to do the chain walk to test for deadlocks anyway. [oleg@tv-sign.ru: more of the same] Signed-of-by: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Esben Nielsen <nielsen.esben@googlemail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 16:59:44 +08:00
chain_walk = 1;
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
}
[PATCH] clean up and remove some extra spinlocks from rtmutex Oleg brought up some interesting points about grabbing the pi_lock for some protections. In this discussion, I realized that there are some places that the pi_lock is being grabbed when it really wasn't necessary. Also this patch does a little bit of clean up. This patch basically does three things: 1) renames the "boost" variable to "chain_walk". Since it is used in the debugging case when it isn't going to be boosted. It better describes what the test is going to do if it succeeds. 2) moves get_task_struct to just before the unlocking of the wait_lock. This removes duplicate code, and makes it a little easier to read. The owner wont go away while either the pi_lock or the wait_lock are held. 3) removes the pi_locking and owner blocked checking completely from the debugging case. This is because the grabbing the lock and doing the check, then releasing the lock is just so full of races. It's just as good to go ahead and call the pi_chain_walk function, since after releasing the lock the owner can then block anyway, and we would have missed that. For the debug case, we really do want to do the chain walk to test for deadlocks anyway. [oleg@tv-sign.ru: more of the same] Signed-of-by: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Esben Nielsen <nielsen.esben@googlemail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 16:59:44 +08:00
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
/* Store the lock on which owner is blocked or NULL */
next_lock = task_blocked_on_lock(owner);
raw_spin_unlock(&owner->pi_lock);
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
/*
* Even if full deadlock detection is on, if the owner is not
* blocked itself, we can avoid finding this out in the chain
* walk.
*/
if (!chain_walk || !next_lock)
return 0;
[PATCH] clean up and remove some extra spinlocks from rtmutex Oleg brought up some interesting points about grabbing the pi_lock for some protections. In this discussion, I realized that there are some places that the pi_lock is being grabbed when it really wasn't necessary. Also this patch does a little bit of clean up. This patch basically does three things: 1) renames the "boost" variable to "chain_walk". Since it is used in the debugging case when it isn't going to be boosted. It better describes what the test is going to do if it succeeds. 2) moves get_task_struct to just before the unlocking of the wait_lock. This removes duplicate code, and makes it a little easier to read. The owner wont go away while either the pi_lock or the wait_lock are held. 3) removes the pi_locking and owner blocked checking completely from the debugging case. This is because the grabbing the lock and doing the check, then releasing the lock is just so full of races. It's just as good to go ahead and call the pi_chain_walk function, since after releasing the lock the owner can then block anyway, and we would have missed that. For the debug case, we really do want to do the chain walk to test for deadlocks anyway. [oleg@tv-sign.ru: more of the same] Signed-of-by: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Esben Nielsen <nielsen.esben@googlemail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 16:59:44 +08:00
/*
* The owner can't disappear while holding a lock,
* so the owner struct is protected by wait_lock.
* Gets dropped in rt_mutex_adjust_prio_chain()!
*/
get_task_struct(owner);
raw_spin_unlock_irq(&lock->wait_lock);
res = rt_mutex_adjust_prio_chain(owner, chwalk, lock,
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
next_lock, waiter, task);
raw_spin_lock_irq(&lock->wait_lock);
return res;
}
/*
* Remove the top waiter from the current tasks pi waiter tree and
* queue it up.
*
* Called with lock->wait_lock held and interrupts disabled.
*/
static void mark_wakeup_next_waiter(struct wake_q_head *wake_q,
struct rt_mutex *lock)
{
struct rt_mutex_waiter *waiter;
raw_spin_lock(&current->pi_lock);
waiter = rt_mutex_top_waiter(lock);
/*
* Remove it from current->pi_waiters and deboost.
*
* We must in fact deboost here in order to ensure we call
* rt_mutex_setprio() to update p->pi_top_task before the
* task unblocks.
*/
rt_mutex_dequeue_pi(current, waiter);
rt_mutex_adjust_prio(current);
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
/*
* As we are waking up the top waiter, and the waiter stays
* queued on the lock until it gets the lock, this lock
* obviously has waiters. Just set the bit here and this has
* the added benefit of forcing all new tasks into the
* slow path making sure no task of lower priority than
* the top waiter can steal this lock.
*/
lock->owner = (void *) RT_MUTEX_HAS_WAITERS;
/*
* We deboosted before waking the top waiter task such that we don't
* run two tasks with the 'same' priority (and ensure the
* p->pi_top_task pointer points to a blocked task). This however can
* lead to priority inversion if we would get preempted after the
* deboost but before waking our donor task, hence the preempt_disable()
* before unlock.
*
* Pairs with preempt_enable() in rt_mutex_postunlock();
*/
preempt_disable();
wake_q_add(wake_q, waiter->task);
raw_spin_unlock(&current->pi_lock);
}
/*
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
* Remove a waiter from a lock and give up
*
* Must be called with lock->wait_lock held and interrupts disabled. I must
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
* have just failed to try_to_take_rt_mutex().
*/
static void remove_waiter(struct rt_mutex *lock,
struct rt_mutex_waiter *waiter)
{
bool is_top_waiter = (waiter == rt_mutex_top_waiter(lock));
struct task_struct *owner = rt_mutex_owner(lock);
struct rt_mutex *next_lock;
lockdep_assert_held(&lock->wait_lock);
raw_spin_lock(&current->pi_lock);
rt_mutex_dequeue(lock, waiter);
current->pi_blocked_on = NULL;
raw_spin_unlock(&current->pi_lock);
/*
* Only update priority if the waiter was the highest priority
* waiter of the lock and there is an owner to update.
*/
if (!owner || !is_top_waiter)
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
return;
raw_spin_lock(&owner->pi_lock);
rt_mutex_dequeue_pi(owner, waiter);
if (rt_mutex_has_waiters(lock))
rt_mutex_enqueue_pi(owner, rt_mutex_top_waiter(lock));
rt_mutex_adjust_prio(owner);
/* Store the lock on which owner is blocked or NULL */
next_lock = task_blocked_on_lock(owner);
[PATCH] clean up and remove some extra spinlocks from rtmutex Oleg brought up some interesting points about grabbing the pi_lock for some protections. In this discussion, I realized that there are some places that the pi_lock is being grabbed when it really wasn't necessary. Also this patch does a little bit of clean up. This patch basically does three things: 1) renames the "boost" variable to "chain_walk". Since it is used in the debugging case when it isn't going to be boosted. It better describes what the test is going to do if it succeeds. 2) moves get_task_struct to just before the unlocking of the wait_lock. This removes duplicate code, and makes it a little easier to read. The owner wont go away while either the pi_lock or the wait_lock are held. 3) removes the pi_locking and owner blocked checking completely from the debugging case. This is because the grabbing the lock and doing the check, then releasing the lock is just so full of races. It's just as good to go ahead and call the pi_chain_walk function, since after releasing the lock the owner can then block anyway, and we would have missed that. For the debug case, we really do want to do the chain walk to test for deadlocks anyway. [oleg@tv-sign.ru: more of the same] Signed-of-by: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Esben Nielsen <nielsen.esben@googlemail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 16:59:44 +08:00
raw_spin_unlock(&owner->pi_lock);
/*
* Don't walk the chain, if the owner task is not blocked
* itself.
*/
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
if (!next_lock)
return;
[PATCH] clean up and remove some extra spinlocks from rtmutex Oleg brought up some interesting points about grabbing the pi_lock for some protections. In this discussion, I realized that there are some places that the pi_lock is being grabbed when it really wasn't necessary. Also this patch does a little bit of clean up. This patch basically does three things: 1) renames the "boost" variable to "chain_walk". Since it is used in the debugging case when it isn't going to be boosted. It better describes what the test is going to do if it succeeds. 2) moves get_task_struct to just before the unlocking of the wait_lock. This removes duplicate code, and makes it a little easier to read. The owner wont go away while either the pi_lock or the wait_lock are held. 3) removes the pi_locking and owner blocked checking completely from the debugging case. This is because the grabbing the lock and doing the check, then releasing the lock is just so full of races. It's just as good to go ahead and call the pi_chain_walk function, since after releasing the lock the owner can then block anyway, and we would have missed that. For the debug case, we really do want to do the chain walk to test for deadlocks anyway. [oleg@tv-sign.ru: more of the same] Signed-of-by: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Esben Nielsen <nielsen.esben@googlemail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 16:59:44 +08:00
/* gets dropped in rt_mutex_adjust_prio_chain()! */
get_task_struct(owner);
raw_spin_unlock_irq(&lock->wait_lock);
rt_mutex_adjust_prio_chain(owner, RT_MUTEX_MIN_CHAINWALK, lock,
next_lock, NULL, current);
raw_spin_lock_irq(&lock->wait_lock);
}
/*
* Recheck the pi chain, in case we got a priority setting
*
* Called from sched_setscheduler
*/
void rt_mutex_adjust_pi(struct task_struct *task)
{
struct rt_mutex_waiter *waiter;
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
struct rt_mutex *next_lock;
unsigned long flags;
raw_spin_lock_irqsave(&task->pi_lock, flags);
waiter = task->pi_blocked_on;
if (!waiter || rt_mutex_waiter_equal(waiter, task_to_waiter(task))) {
raw_spin_unlock_irqrestore(&task->pi_lock, flags);
return;
}
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
next_lock = waiter->lock;
raw_spin_unlock_irqrestore(&task->pi_lock, flags);
[PATCH] clean up and remove some extra spinlocks from rtmutex Oleg brought up some interesting points about grabbing the pi_lock for some protections. In this discussion, I realized that there are some places that the pi_lock is being grabbed when it really wasn't necessary. Also this patch does a little bit of clean up. This patch basically does three things: 1) renames the "boost" variable to "chain_walk". Since it is used in the debugging case when it isn't going to be boosted. It better describes what the test is going to do if it succeeds. 2) moves get_task_struct to just before the unlocking of the wait_lock. This removes duplicate code, and makes it a little easier to read. The owner wont go away while either the pi_lock or the wait_lock are held. 3) removes the pi_locking and owner blocked checking completely from the debugging case. This is because the grabbing the lock and doing the check, then releasing the lock is just so full of races. It's just as good to go ahead and call the pi_chain_walk function, since after releasing the lock the owner can then block anyway, and we would have missed that. For the debug case, we really do want to do the chain walk to test for deadlocks anyway. [oleg@tv-sign.ru: more of the same] Signed-of-by: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Esben Nielsen <nielsen.esben@googlemail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 16:59:44 +08:00
/* gets dropped in rt_mutex_adjust_prio_chain()! */
get_task_struct(task);
rtmutex: Detect changes in the pi lock chain When we walk the lock chain, we drop all locks after each step. So the lock chain can change under us before we reacquire the locks. That's harmless in principle as we just follow the wrong lock path. But it can lead to a false positive in the dead lock detection logic: T0 holds L0 T0 blocks on L1 held by T1 T1 blocks on L2 held by T2 T2 blocks on L3 held by T3 T4 blocks on L4 held by T4 Now we walk the chain lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all. Brad tried to work around that in the deadlock detection logic itself, but the more I looked at it the less I liked it, because it's crystal ball magic after the fact. We actually can detect a chain change very simple: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T2 times out and blocks on L0 Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; So if we detect that T2 is now blocked on a different lock we stop the chain walk. That's also correct in the following scenario: lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 -> next_lock = T2->pi_blocked_on->lock; drop locks T3 times out and drops L3 T2 acquires L3 and blocks on L4 now Now we continue: lock T2 -> if (next_lock != T2->pi_blocked_on->lock) return; We don't have to follow up the chain at that point, because T2 propagated our priority up to T4 already. [ Folded a cleanup patch from peterz ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reported-by: Brad Mouring <bmouring@ni.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.de Cc: stable@vger.kernel.org
2014-06-05 17:16:12 +08:00
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)
{
debug_rt_mutex_init_waiter(waiter);
RB_CLEAR_NODE(&waiter->pi_tree_entry);
RB_CLEAR_NODE(&waiter->tree_entry);
waiter->task = NULL;
}
/**
* __rt_mutex_slowlock() - Perform the wait-wake-try-to-take loop
* @lock: the rt_mutex to take
* @state: the state the task should block in (TASK_INTERRUPTIBLE
* or TASK_UNINTERRUPTIBLE)
* @timeout: the pre-initialized and started timer, or NULL for none
* @waiter: the pre-initialized rt_mutex_waiter
*
* Must be called with lock->wait_lock held and interrupts disabled
*/
static int __sched
__rt_mutex_slowlock(struct rt_mutex *lock, int state,
struct hrtimer_sleeper *timeout,
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
struct rt_mutex_waiter *waiter)
{
int ret = 0;
for (;;) {
/* Try to acquire the lock: */
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
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 (ret)
break;
}
raw_spin_unlock_irq(&lock->wait_lock);
debug_rt_mutex_print_deadlock(waiter);
schedule();
raw_spin_lock_irq(&lock->wait_lock);
set_current_state(state);
}
__set_current_state(TASK_RUNNING);
return ret;
}
static void rt_mutex_handle_deadlock(int res, int detect_deadlock,
struct rt_mutex_waiter *w)
{
/*
* If the result is not -EDEADLOCK or the caller requested
* deadlock detection, nothing to do here.
*/
if (res != -EDEADLOCK || detect_deadlock)
return;
/*
* Yell lowdly and stop the task right here.
*/
rt_mutex_print_deadlock(w);
while (1) {
set_current_state(TASK_INTERRUPTIBLE);
schedule();
}
}
/*
* 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 rt_mutex_waiter waiter;
unsigned long flags;
int ret = 0;
rt_mutex_init_waiter(&waiter);
/*
* 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);
/* Try to acquire the lock again: */
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
if (try_to_take_rt_mutex(lock, current, NULL)) {
raw_spin_unlock_irqrestore(&lock->wait_lock, flags);
return 0;
}
set_current_state(state);
/* Setup the timer, when timeout != NULL */
if (unlikely(timeout))
hrtimer_start_expires(&timeout->timer, HRTIMER_MODE_ABS);
ret = task_blocks_on_rt_mutex(lock, &waiter, current, chwalk);
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
if (likely(!ret))
/* sleep on the mutex */
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
ret = __rt_mutex_slowlock(lock, state, timeout, &waiter);
if (unlikely(ret)) {
__set_current_state(TASK_RUNNING);
remove_waiter(lock, &waiter);
rt_mutex_handle_deadlock(ret, chwalk, &waiter);
}
/*
* try_to_take_rt_mutex() sets the waiter bit
* unconditionally. We might have to fix that up.
*/
fixup_rt_mutex_waiters(lock);
raw_spin_unlock_irqrestore(&lock->wait_lock, flags);
/* Remove pending timer: */
if (unlikely(timeout))
hrtimer_cancel(&timeout->timer);
debug_rt_mutex_free_waiter(&waiter);
return ret;
}
futex: Avoid violating the 10th rule of futex Julia reported futex state corruption in the following scenario: waiter waker stealer (prio > waiter) futex(WAIT_REQUEUE_PI, uaddr, uaddr2, timeout=[N ms]) futex_wait_requeue_pi() futex_wait_queue_me() freezable_schedule() <scheduled out> futex(LOCK_PI, uaddr2) futex(CMP_REQUEUE_PI, uaddr, uaddr2, 1, 0) /* requeues waiter to uaddr2 */ futex(UNLOCK_PI, uaddr2) wake_futex_pi() cmp_futex_value_locked(uaddr2, waiter) wake_up_q() <woken by waker> <hrtimer_wakeup() fires, clears sleeper->task> futex(LOCK_PI, uaddr2) __rt_mutex_start_proxy_lock() try_to_take_rt_mutex() /* steals lock */ rt_mutex_set_owner(lock, stealer) <preempted> <scheduled in> rt_mutex_wait_proxy_lock() __rt_mutex_slowlock() try_to_take_rt_mutex() /* fails, lock held by stealer */ if (timeout && !timeout->task) return -ETIMEDOUT; fixup_owner() /* lock wasn't acquired, so, fixup_pi_state_owner skipped */ return -ETIMEDOUT; /* At this point, we've returned -ETIMEDOUT to userspace, but the * futex word shows waiter to be the owner, and the pi_mutex has * stealer as the owner */ futex_lock(LOCK_PI, uaddr2) -> bails with EDEADLK, futex word says we're owner. And suggested that what commit: 73d786bd043e ("futex: Rework inconsistent rt_mutex/futex_q state") removes from fixup_owner() looks to be just what is needed. And indeed it is -- I completely missed that requeue_pi could also result in this case. So we need to restore that, except that subsequent patches, like commit: 16ffa12d7425 ("futex: Pull rt_mutex_futex_unlock() out from under hb->lock") changed all the locking rules. Even without that, the sequence: - if (rt_mutex_futex_trylock(&q->pi_state->pi_mutex)) { - locked = 1; - goto out; - } - raw_spin_lock_irq(&q->pi_state->pi_mutex.wait_lock); - owner = rt_mutex_owner(&q->pi_state->pi_mutex); - if (!owner) - owner = rt_mutex_next_owner(&q->pi_state->pi_mutex); - raw_spin_unlock_irq(&q->pi_state->pi_mutex.wait_lock); - ret = fixup_pi_state_owner(uaddr, q, owner); already suggests there were races; otherwise we'd never have to look at next_owner. So instead of doing 3 consecutive wait_lock sections with who knows what races, we do it all in a single section. Additionally, the usage of pi_state->owner in fixup_owner() was only safe because only the rt_mutex owner would modify it, which this additional case wrecks. Luckily the values can only change away and not to the value we're testing, this means we can do a speculative test and double check once we have the wait_lock. Fixes: 73d786bd043e ("futex: Rework inconsistent rt_mutex/futex_q state") Reported-by: Julia Cartwright <julia@ni.com> Reported-by: Gratian Crisan <gratian.crisan@ni.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Julia Cartwright <julia@ni.com> Tested-by: Gratian Crisan <gratian.crisan@ni.com> Cc: Darren Hart <dvhart@infradead.org> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20171208124939.7livp7no2ov65rrc@hirez.programming.kicks-ass.net
2017-12-08 20:49:39 +08:00
static inline int __rt_mutex_slowtrylock(struct rt_mutex *lock)
{
int ret = try_to_take_rt_mutex(lock, current, NULL);
/*
* try_to_take_rt_mutex() sets the lock waiters bit
* unconditionally. Clean this up.
*/
fixup_rt_mutex_waiters(lock);
return ret;
}
/*
* Slow path try-lock function:
*/
static inline int rt_mutex_slowtrylock(struct rt_mutex *lock)
{
unsigned long flags;
int ret;
/*
* If the lock already has an owner we fail to get the lock.
* This can be done without taking the @lock->wait_lock as
* it is only being read, and this is a trylock anyway.
*/
if (rt_mutex_owner(lock))
return 0;
/*
* The mutex has currently no owner. Lock the wait lock and try to
* acquire the lock. We use irqsave here to support early boot calls.
*/
raw_spin_lock_irqsave(&lock->wait_lock, flags);
futex: Avoid violating the 10th rule of futex Julia reported futex state corruption in the following scenario: waiter waker stealer (prio > waiter) futex(WAIT_REQUEUE_PI, uaddr, uaddr2, timeout=[N ms]) futex_wait_requeue_pi() futex_wait_queue_me() freezable_schedule() <scheduled out> futex(LOCK_PI, uaddr2) futex(CMP_REQUEUE_PI, uaddr, uaddr2, 1, 0) /* requeues waiter to uaddr2 */ futex(UNLOCK_PI, uaddr2) wake_futex_pi() cmp_futex_value_locked(uaddr2, waiter) wake_up_q() <woken by waker> <hrtimer_wakeup() fires, clears sleeper->task> futex(LOCK_PI, uaddr2) __rt_mutex_start_proxy_lock() try_to_take_rt_mutex() /* steals lock */ rt_mutex_set_owner(lock, stealer) <preempted> <scheduled in> rt_mutex_wait_proxy_lock() __rt_mutex_slowlock() try_to_take_rt_mutex() /* fails, lock held by stealer */ if (timeout && !timeout->task) return -ETIMEDOUT; fixup_owner() /* lock wasn't acquired, so, fixup_pi_state_owner skipped */ return -ETIMEDOUT; /* At this point, we've returned -ETIMEDOUT to userspace, but the * futex word shows waiter to be the owner, and the pi_mutex has * stealer as the owner */ futex_lock(LOCK_PI, uaddr2) -> bails with EDEADLK, futex word says we're owner. And suggested that what commit: 73d786bd043e ("futex: Rework inconsistent rt_mutex/futex_q state") removes from fixup_owner() looks to be just what is needed. And indeed it is -- I completely missed that requeue_pi could also result in this case. So we need to restore that, except that subsequent patches, like commit: 16ffa12d7425 ("futex: Pull rt_mutex_futex_unlock() out from under hb->lock") changed all the locking rules. Even without that, the sequence: - if (rt_mutex_futex_trylock(&q->pi_state->pi_mutex)) { - locked = 1; - goto out; - } - raw_spin_lock_irq(&q->pi_state->pi_mutex.wait_lock); - owner = rt_mutex_owner(&q->pi_state->pi_mutex); - if (!owner) - owner = rt_mutex_next_owner(&q->pi_state->pi_mutex); - raw_spin_unlock_irq(&q->pi_state->pi_mutex.wait_lock); - ret = fixup_pi_state_owner(uaddr, q, owner); already suggests there were races; otherwise we'd never have to look at next_owner. So instead of doing 3 consecutive wait_lock sections with who knows what races, we do it all in a single section. Additionally, the usage of pi_state->owner in fixup_owner() was only safe because only the rt_mutex owner would modify it, which this additional case wrecks. Luckily the values can only change away and not to the value we're testing, this means we can do a speculative test and double check once we have the wait_lock. Fixes: 73d786bd043e ("futex: Rework inconsistent rt_mutex/futex_q state") Reported-by: Julia Cartwright <julia@ni.com> Reported-by: Gratian Crisan <gratian.crisan@ni.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Julia Cartwright <julia@ni.com> Tested-by: Gratian Crisan <gratian.crisan@ni.com> Cc: Darren Hart <dvhart@infradead.org> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20171208124939.7livp7no2ov65rrc@hirez.programming.kicks-ass.net
2017-12-08 20:49:39 +08:00
ret = __rt_mutex_slowtrylock(lock);
raw_spin_unlock_irqrestore(&lock->wait_lock, flags);
return ret;
}
/*
* Slow path to release a rt-mutex.
*
* 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)
{
unsigned long flags;
/* irqsave required to support early boot calls */
raw_spin_lock_irqsave(&lock->wait_lock, flags);
debug_rt_mutex_unlock(lock);
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
/*
* We must be careful here if the fast path is enabled. If we
* have no waiters queued we cannot set owner to NULL here
* because of:
*
* foo->lock->owner = NULL;
* rtmutex_lock(foo->lock); <- fast path
* free = atomic_dec_and_test(foo->refcnt);
* rtmutex_unlock(foo->lock); <- fast path
* if (free)
* kfree(foo);
* raw_spin_unlock(foo->lock->wait_lock);
*
* So for the fastpath enabled kernel:
*
* Nothing can set the waiters bit as long as we hold
* lock->wait_lock. So we do the following sequence:
*
* owner = rt_mutex_owner(lock);
* clear_rt_mutex_waiters(lock);
* raw_spin_unlock(&lock->wait_lock);
* if (cmpxchg(&lock->owner, owner, 0) == owner)
* return;
* goto retry;
*
* The fastpath disabled variant is simple as all access to
* lock->owner is serialized by lock->wait_lock:
*
* lock->owner = NULL;
* raw_spin_unlock(&lock->wait_lock);
*/
while (!rt_mutex_has_waiters(lock)) {
/* Drops lock->wait_lock ! */
if (unlock_rt_mutex_safe(lock, flags) == true)
return false;
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
/* Relock the rtmutex and try again */
raw_spin_lock_irqsave(&lock->wait_lock, flags);
}
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
/*
* The wakeup next waiter path does not suffer from the above
* race. See the comments there.
*
* Queue the next waiter for wakeup once we release the wait_lock.
rtmutex: Plug slow unlock race When the rtmutex fast path is enabled the slow unlock function can create the following situation: spin_lock(foo->m->wait_lock); foo->m->owner = NULL; rt_mutex_lock(foo->m); <-- fast path free = atomic_dec_and_test(foo->refcnt); rt_mutex_unlock(foo->m); <-- fast path if (free) kfree(foo); spin_unlock(foo->m->wait_lock); <--- Use after free. Plug the race by changing the slow unlock to the following scheme: while (!rt_mutex_has_waiters(m)) { /* Clear the waiters bit in m->owner */ clear_rt_mutex_waiters(m); owner = rt_mutex_owner(m); spin_unlock(m->wait_lock); if (cmpxchg(m->owner, owner, 0) == owner) return; spin_lock(m->wait_lock); } So in case of a new waiter incoming while the owner tries the slow path unlock we have two situations: unlock(wait_lock); lock(wait_lock); cmpxchg(p, owner, 0) == owner mark_rt_mutex_waiters(lock); acquire(lock); Or: unlock(wait_lock); lock(wait_lock); mark_rt_mutex_waiters(lock); cmpxchg(p, owner, 0) != owner enqueue_waiter(); unlock(wait_lock); lock(wait_lock); wakeup_next waiter(); unlock(wait_lock); lock(wait_lock); acquire(lock); If the fast path is disabled, then the simple m->owner = NULL; unlock(m->wait_lock); is sufficient as all access to m->owner is serialized via m->wait_lock; Also document and clarify the wakeup_next_waiter function as suggested by Oleg Nesterov. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-06-12 02:44:04 +08:00
*/
mark_wakeup_next_waiter(wake_q, lock);
raw_spin_unlock_irqrestore(&lock->wait_lock, flags);
return true; /* call rt_mutex_postunlock() */
}
/*
* debug aware fast / slowpath lock,trylock,unlock
*
* The atomic acquire/release ops are compiled away, when either the
* architecture does not support cmpxchg or when debugging is enabled.
*/
static inline int
rt_mutex_fastlock(struct rt_mutex *lock, int state,
int (*slowfn)(struct rt_mutex *lock, int state,
struct hrtimer_sleeper *timeout,
enum rtmutex_chainwalk chwalk))
{
if (likely(rt_mutex_cmpxchg_acquire(lock, NULL, current)))
return 0;
return slowfn(lock, state, NULL, RT_MUTEX_MIN_CHAINWALK);
}
static inline int
rt_mutex_timed_fastlock(struct rt_mutex *lock, int state,
struct hrtimer_sleeper *timeout,
enum rtmutex_chainwalk chwalk,
int (*slowfn)(struct rt_mutex *lock, int state,
struct hrtimer_sleeper *timeout,
enum rtmutex_chainwalk chwalk))
{
if (chwalk == RT_MUTEX_MIN_CHAINWALK &&
likely(rt_mutex_cmpxchg_acquire(lock, NULL, current)))
return 0;
return slowfn(lock, state, timeout, chwalk);
}
static inline int
rt_mutex_fasttrylock(struct rt_mutex *lock,
int (*slowfn)(struct rt_mutex *lock))
{
if (likely(rt_mutex_cmpxchg_acquire(lock, NULL, current)))
return 1;
return slowfn(lock);
}
2017-03-23 22:56:07 +08:00
/*
* Performs the wakeup of the the top-waiter and re-enables preemption.
2017-03-23 22:56:07 +08:00
*/
void rt_mutex_postunlock(struct wake_q_head *wake_q)
2017-03-23 22:56:07 +08:00
{
wake_up_q(wake_q);
/* Pairs with preempt_disable() in rt_mutex_slowunlock() */
preempt_enable();
2017-03-23 22:56:07 +08:00
}
static inline void
rt_mutex_fastunlock(struct rt_mutex *lock,
bool (*slowfn)(struct rt_mutex *lock,
struct wake_q_head *wqh))
{
DEFINE_WAKE_Q(wake_q);
if (likely(rt_mutex_cmpxchg_release(lock, current, NULL)))
return;
if (slowfn(lock, &wake_q))
rt_mutex_postunlock(&wake_q);
}
static inline void __rt_mutex_lock(struct rt_mutex *lock, unsigned int subclass)
{
might_sleep();
mutex_acquire(&lock->dep_map, subclass, 0, _RET_IP_);
rt_mutex_fastlock(lock, TASK_UNINTERRUPTIBLE, rt_mutex_slowlock);
}
#ifdef CONFIG_DEBUG_LOCK_ALLOC
/**
* rt_mutex_lock_nested - lock a rt_mutex
*
* @lock: the rt_mutex to be locked
* @subclass: the lockdep subclass
*/
void __sched rt_mutex_lock_nested(struct rt_mutex *lock, unsigned int subclass)
{
__rt_mutex_lock(lock, subclass);
}
EXPORT_SYMBOL_GPL(rt_mutex_lock_nested);
#else /* !CONFIG_DEBUG_LOCK_ALLOC */
/**
* rt_mutex_lock - lock a rt_mutex
*
* @lock: the rt_mutex to be locked
*/
void __sched rt_mutex_lock(struct rt_mutex *lock)
{
__rt_mutex_lock(lock, 0);
}
EXPORT_SYMBOL_GPL(rt_mutex_lock);
#endif
/**
* rt_mutex_lock_interruptible - lock a rt_mutex interruptible
*
* @lock: the rt_mutex to be locked
*
* Returns:
* 0 on success
* -EINTR when interrupted by a signal
*/
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)
2019-09-20 00:09:40 +08:00
mutex_release(&lock->dep_map, _RET_IP_);
return ret;
}
EXPORT_SYMBOL_GPL(rt_mutex_lock_interruptible);
/*
* Futex variant, must not use fastpath.
*/
int __sched rt_mutex_futex_trylock(struct rt_mutex *lock)
{
return rt_mutex_slowtrylock(lock);
}
futex: Avoid violating the 10th rule of futex Julia reported futex state corruption in the following scenario: waiter waker stealer (prio > waiter) futex(WAIT_REQUEUE_PI, uaddr, uaddr2, timeout=[N ms]) futex_wait_requeue_pi() futex_wait_queue_me() freezable_schedule() <scheduled out> futex(LOCK_PI, uaddr2) futex(CMP_REQUEUE_PI, uaddr, uaddr2, 1, 0) /* requeues waiter to uaddr2 */ futex(UNLOCK_PI, uaddr2) wake_futex_pi() cmp_futex_value_locked(uaddr2, waiter) wake_up_q() <woken by waker> <hrtimer_wakeup() fires, clears sleeper->task> futex(LOCK_PI, uaddr2) __rt_mutex_start_proxy_lock() try_to_take_rt_mutex() /* steals lock */ rt_mutex_set_owner(lock, stealer) <preempted> <scheduled in> rt_mutex_wait_proxy_lock() __rt_mutex_slowlock() try_to_take_rt_mutex() /* fails, lock held by stealer */ if (timeout && !timeout->task) return -ETIMEDOUT; fixup_owner() /* lock wasn't acquired, so, fixup_pi_state_owner skipped */ return -ETIMEDOUT; /* At this point, we've returned -ETIMEDOUT to userspace, but the * futex word shows waiter to be the owner, and the pi_mutex has * stealer as the owner */ futex_lock(LOCK_PI, uaddr2) -> bails with EDEADLK, futex word says we're owner. And suggested that what commit: 73d786bd043e ("futex: Rework inconsistent rt_mutex/futex_q state") removes from fixup_owner() looks to be just what is needed. And indeed it is -- I completely missed that requeue_pi could also result in this case. So we need to restore that, except that subsequent patches, like commit: 16ffa12d7425 ("futex: Pull rt_mutex_futex_unlock() out from under hb->lock") changed all the locking rules. Even without that, the sequence: - if (rt_mutex_futex_trylock(&q->pi_state->pi_mutex)) { - locked = 1; - goto out; - } - raw_spin_lock_irq(&q->pi_state->pi_mutex.wait_lock); - owner = rt_mutex_owner(&q->pi_state->pi_mutex); - if (!owner) - owner = rt_mutex_next_owner(&q->pi_state->pi_mutex); - raw_spin_unlock_irq(&q->pi_state->pi_mutex.wait_lock); - ret = fixup_pi_state_owner(uaddr, q, owner); already suggests there were races; otherwise we'd never have to look at next_owner. So instead of doing 3 consecutive wait_lock sections with who knows what races, we do it all in a single section. Additionally, the usage of pi_state->owner in fixup_owner() was only safe because only the rt_mutex owner would modify it, which this additional case wrecks. Luckily the values can only change away and not to the value we're testing, this means we can do a speculative test and double check once we have the wait_lock. Fixes: 73d786bd043e ("futex: Rework inconsistent rt_mutex/futex_q state") Reported-by: Julia Cartwright <julia@ni.com> Reported-by: Gratian Crisan <gratian.crisan@ni.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Julia Cartwright <julia@ni.com> Tested-by: Gratian Crisan <gratian.crisan@ni.com> Cc: Darren Hart <dvhart@infradead.org> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20171208124939.7livp7no2ov65rrc@hirez.programming.kicks-ass.net
2017-12-08 20:49:39 +08:00
int __sched __rt_mutex_futex_trylock(struct rt_mutex *lock)
{
return __rt_mutex_slowtrylock(lock);
}
/**
* rt_mutex_timed_lock - lock a rt_mutex interruptible
* the timeout structure is provided
* by the caller
*
* @lock: the rt_mutex to be locked
* @timeout: timeout structure or NULL (no timeout)
*
* Returns:
* 0 on success
* -EINTR when interrupted by a signal
* -ETIMEDOUT when the timeout expired
*/
int
rt_mutex_timed_lock(struct rt_mutex *lock, struct hrtimer_sleeper *timeout)
{
int ret;
might_sleep();
mutex_acquire(&lock->dep_map, 0, 0, _RET_IP_);
ret = rt_mutex_timed_fastlock(lock, TASK_INTERRUPTIBLE, timeout,
RT_MUTEX_MIN_CHAINWALK,
rt_mutex_slowlock);
if (ret)
2019-09-20 00:09:40 +08:00
mutex_release(&lock->dep_map, _RET_IP_);
return ret;
}
EXPORT_SYMBOL_GPL(rt_mutex_timed_lock);
/**
* rt_mutex_trylock - try to lock a rt_mutex
*
* @lock: the rt_mutex to be locked
*
* This function can only be called in thread context. It's safe to
* call it from atomic regions, but not from hard interrupt or soft
* interrupt context.
*
* Returns 1 on success and 0 on contention
*/
int __sched rt_mutex_trylock(struct rt_mutex *lock)
{
int ret;
if (WARN_ON_ONCE(in_irq() || in_nmi() || in_serving_softirq()))
return 0;
ret = rt_mutex_fasttrylock(lock, rt_mutex_slowtrylock);
if (ret)
mutex_acquire(&lock->dep_map, 0, 1, _RET_IP_);
return ret;
}
EXPORT_SYMBOL_GPL(rt_mutex_trylock);
/**
* rt_mutex_unlock - unlock a rt_mutex
*
* @lock: the rt_mutex to be unlocked
*/
void __sched rt_mutex_unlock(struct rt_mutex *lock)
{
2019-09-20 00:09:40 +08:00
mutex_release(&lock->dep_map, _RET_IP_);
rt_mutex_fastunlock(lock, rt_mutex_slowunlock);
}
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)
{
lockdep_assert_held(&lock->wait_lock);
debug_rt_mutex_unlock(lock);
if (!rt_mutex_has_waiters(lock)) {
lock->owner = NULL;
return false; /* done */
}
2017-03-23 22:56:07 +08:00
/*
* We've already deboosted, mark_wakeup_next_waiter() will
* retain preempt_disabled when we drop the wait_lock, to
* avoid inversion prior to the wakeup. preempt_disable()
* therein pairs with rt_mutex_postunlock().
2017-03-23 22:56:07 +08:00
*/
mark_wakeup_next_waiter(wake_q, lock);
2017-03-23 22:56:07 +08:00
return true; /* call postunlock() */
}
void __sched rt_mutex_futex_unlock(struct rt_mutex *lock)
{
DEFINE_WAKE_Q(wake_q);
rtmutex: Make rt_mutex_futex_unlock() safe for irq-off callsites When running rcutorture with TREE03 config, CONFIG_PROVE_LOCKING=y, and kernel cmdline argument "rcutorture.gp_exp=1", lockdep reports a HARDIRQ-safe->HARDIRQ-unsafe deadlock: ================================ WARNING: inconsistent lock state 4.16.0-rc4+ #1 Not tainted -------------------------------- inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. takes: __schedule+0xbe/0xaf0 {IN-HARDIRQ-W} state was registered at: _raw_spin_lock+0x2a/0x40 scheduler_tick+0x47/0xf0 ... other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&rq->lock); <Interrupt> lock(&rq->lock); *** DEADLOCK *** 1 lock held by rcu_torture_rea/724: rcu_torture_read_lock+0x0/0x70 stack backtrace: CPU: 2 PID: 724 Comm: rcu_torture_rea Not tainted 4.16.0-rc4+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-20171110_100015-anatol 04/01/2014 Call Trace: lock_acquire+0x90/0x200 ? __schedule+0xbe/0xaf0 _raw_spin_lock+0x2a/0x40 ? __schedule+0xbe/0xaf0 __schedule+0xbe/0xaf0 preempt_schedule_irq+0x2f/0x60 retint_kernel+0x1b/0x2d RIP: 0010:rcu_read_unlock_special+0x0/0x680 ? rcu_torture_read_unlock+0x60/0x60 __rcu_read_unlock+0x64/0x70 rcu_torture_read_unlock+0x17/0x60 rcu_torture_reader+0x275/0x450 ? rcutorture_booster_init+0x110/0x110 ? rcu_torture_stall+0x230/0x230 ? kthread+0x10e/0x130 kthread+0x10e/0x130 ? kthread_create_worker_on_cpu+0x70/0x70 ? call_usermodehelper_exec_async+0x11a/0x150 ret_from_fork+0x3a/0x50 This happens with the following even sequence: preempt_schedule_irq(); local_irq_enable(); __schedule(): local_irq_disable(); // irq off ... rcu_note_context_switch(): rcu_note_preempt_context_switch(): rcu_read_unlock_special(): local_irq_save(flags); ... raw_spin_unlock_irqrestore(...,flags); // irq remains off rt_mutex_futex_unlock(): raw_spin_lock_irq(); ... raw_spin_unlock_irq(); // accidentally set irq on <return to __schedule()> rq_lock(): raw_spin_lock(); // acquiring rq->lock with irq on which means rq->lock becomes a HARDIRQ-unsafe lock, which can cause deadlocks in scheduler code. This problem was introduced by commit 02a7c234e540 ("rcu: Suppress lockdep false-positive ->boost_mtx complaints"). That brought the user of rt_mutex_futex_unlock() with irq off. To fix this, replace the *lock_irq() in rt_mutex_futex_unlock() with *lock_irq{save,restore}() to make it safe to call rt_mutex_futex_unlock() with irq off. Fixes: 02a7c234e540 ("rcu: Suppress lockdep false-positive ->boost_mtx complaints") Signed-off-by: Boqun Feng <boqun.feng@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Lai Jiangshan <jiangshanlai@gmail.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Josh Triplett <josh@joshtriplett.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com> Link: https://lkml.kernel.org/r/20180309065630.8283-1-boqun.feng@gmail.com
2018-03-09 14:56:28 +08:00
unsigned long flags;
bool postunlock;
rtmutex: Make rt_mutex_futex_unlock() safe for irq-off callsites When running rcutorture with TREE03 config, CONFIG_PROVE_LOCKING=y, and kernel cmdline argument "rcutorture.gp_exp=1", lockdep reports a HARDIRQ-safe->HARDIRQ-unsafe deadlock: ================================ WARNING: inconsistent lock state 4.16.0-rc4+ #1 Not tainted -------------------------------- inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. takes: __schedule+0xbe/0xaf0 {IN-HARDIRQ-W} state was registered at: _raw_spin_lock+0x2a/0x40 scheduler_tick+0x47/0xf0 ... other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&rq->lock); <Interrupt> lock(&rq->lock); *** DEADLOCK *** 1 lock held by rcu_torture_rea/724: rcu_torture_read_lock+0x0/0x70 stack backtrace: CPU: 2 PID: 724 Comm: rcu_torture_rea Not tainted 4.16.0-rc4+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-20171110_100015-anatol 04/01/2014 Call Trace: lock_acquire+0x90/0x200 ? __schedule+0xbe/0xaf0 _raw_spin_lock+0x2a/0x40 ? __schedule+0xbe/0xaf0 __schedule+0xbe/0xaf0 preempt_schedule_irq+0x2f/0x60 retint_kernel+0x1b/0x2d RIP: 0010:rcu_read_unlock_special+0x0/0x680 ? rcu_torture_read_unlock+0x60/0x60 __rcu_read_unlock+0x64/0x70 rcu_torture_read_unlock+0x17/0x60 rcu_torture_reader+0x275/0x450 ? rcutorture_booster_init+0x110/0x110 ? rcu_torture_stall+0x230/0x230 ? kthread+0x10e/0x130 kthread+0x10e/0x130 ? kthread_create_worker_on_cpu+0x70/0x70 ? call_usermodehelper_exec_async+0x11a/0x150 ret_from_fork+0x3a/0x50 This happens with the following even sequence: preempt_schedule_irq(); local_irq_enable(); __schedule(): local_irq_disable(); // irq off ... rcu_note_context_switch(): rcu_note_preempt_context_switch(): rcu_read_unlock_special(): local_irq_save(flags); ... raw_spin_unlock_irqrestore(...,flags); // irq remains off rt_mutex_futex_unlock(): raw_spin_lock_irq(); ... raw_spin_unlock_irq(); // accidentally set irq on <return to __schedule()> rq_lock(): raw_spin_lock(); // acquiring rq->lock with irq on which means rq->lock becomes a HARDIRQ-unsafe lock, which can cause deadlocks in scheduler code. This problem was introduced by commit 02a7c234e540 ("rcu: Suppress lockdep false-positive ->boost_mtx complaints"). That brought the user of rt_mutex_futex_unlock() with irq off. To fix this, replace the *lock_irq() in rt_mutex_futex_unlock() with *lock_irq{save,restore}() to make it safe to call rt_mutex_futex_unlock() with irq off. Fixes: 02a7c234e540 ("rcu: Suppress lockdep false-positive ->boost_mtx complaints") Signed-off-by: Boqun Feng <boqun.feng@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Lai Jiangshan <jiangshanlai@gmail.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Josh Triplett <josh@joshtriplett.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com> Link: https://lkml.kernel.org/r/20180309065630.8283-1-boqun.feng@gmail.com
2018-03-09 14:56:28 +08:00
raw_spin_lock_irqsave(&lock->wait_lock, flags);
postunlock = __rt_mutex_futex_unlock(lock, &wake_q);
rtmutex: Make rt_mutex_futex_unlock() safe for irq-off callsites When running rcutorture with TREE03 config, CONFIG_PROVE_LOCKING=y, and kernel cmdline argument "rcutorture.gp_exp=1", lockdep reports a HARDIRQ-safe->HARDIRQ-unsafe deadlock: ================================ WARNING: inconsistent lock state 4.16.0-rc4+ #1 Not tainted -------------------------------- inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. takes: __schedule+0xbe/0xaf0 {IN-HARDIRQ-W} state was registered at: _raw_spin_lock+0x2a/0x40 scheduler_tick+0x47/0xf0 ... other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&rq->lock); <Interrupt> lock(&rq->lock); *** DEADLOCK *** 1 lock held by rcu_torture_rea/724: rcu_torture_read_lock+0x0/0x70 stack backtrace: CPU: 2 PID: 724 Comm: rcu_torture_rea Not tainted 4.16.0-rc4+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-20171110_100015-anatol 04/01/2014 Call Trace: lock_acquire+0x90/0x200 ? __schedule+0xbe/0xaf0 _raw_spin_lock+0x2a/0x40 ? __schedule+0xbe/0xaf0 __schedule+0xbe/0xaf0 preempt_schedule_irq+0x2f/0x60 retint_kernel+0x1b/0x2d RIP: 0010:rcu_read_unlock_special+0x0/0x680 ? rcu_torture_read_unlock+0x60/0x60 __rcu_read_unlock+0x64/0x70 rcu_torture_read_unlock+0x17/0x60 rcu_torture_reader+0x275/0x450 ? rcutorture_booster_init+0x110/0x110 ? rcu_torture_stall+0x230/0x230 ? kthread+0x10e/0x130 kthread+0x10e/0x130 ? kthread_create_worker_on_cpu+0x70/0x70 ? call_usermodehelper_exec_async+0x11a/0x150 ret_from_fork+0x3a/0x50 This happens with the following even sequence: preempt_schedule_irq(); local_irq_enable(); __schedule(): local_irq_disable(); // irq off ... rcu_note_context_switch(): rcu_note_preempt_context_switch(): rcu_read_unlock_special(): local_irq_save(flags); ... raw_spin_unlock_irqrestore(...,flags); // irq remains off rt_mutex_futex_unlock(): raw_spin_lock_irq(); ... raw_spin_unlock_irq(); // accidentally set irq on <return to __schedule()> rq_lock(): raw_spin_lock(); // acquiring rq->lock with irq on which means rq->lock becomes a HARDIRQ-unsafe lock, which can cause deadlocks in scheduler code. This problem was introduced by commit 02a7c234e540 ("rcu: Suppress lockdep false-positive ->boost_mtx complaints"). That brought the user of rt_mutex_futex_unlock() with irq off. To fix this, replace the *lock_irq() in rt_mutex_futex_unlock() with *lock_irq{save,restore}() to make it safe to call rt_mutex_futex_unlock() with irq off. Fixes: 02a7c234e540 ("rcu: Suppress lockdep false-positive ->boost_mtx complaints") Signed-off-by: Boqun Feng <boqun.feng@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Lai Jiangshan <jiangshanlai@gmail.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Josh Triplett <josh@joshtriplett.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com> Link: https://lkml.kernel.org/r/20180309065630.8283-1-boqun.feng@gmail.com
2018-03-09 14:56:28 +08:00
raw_spin_unlock_irqrestore(&lock->wait_lock, flags);
if (postunlock)
rt_mutex_postunlock(&wake_q);
}
/**
* rt_mutex_destroy - mark a mutex unusable
* @lock: the mutex to be destroyed
*
* This function marks the mutex uninitialized, and any subsequent
* use of the mutex is forbidden. The mutex must not be locked when
* this function is called.
*/
void rt_mutex_destroy(struct rt_mutex *lock)
{
WARN_ON(rt_mutex_is_locked(lock));
#ifdef CONFIG_DEBUG_RT_MUTEXES
lock->magic = NULL;
#endif
}
EXPORT_SYMBOL_GPL(rt_mutex_destroy);
/**
* __rt_mutex_init - initialize the rt lock
*
* @lock: the rt lock to be initialized
*
* Initialize the rt lock to unlocked state.
*
* Initializing of a locked rt lock is not allowed
*/
void __rt_mutex_init(struct rt_mutex *lock, const char *name,
struct lock_class_key *key)
{
lock->owner = NULL;
raw_spin_lock_init(&lock->wait_lock);
lock->waiters = RB_ROOT_CACHED;
if (name && key)
debug_rt_mutex_init(lock, name, key);
}
EXPORT_SYMBOL_GPL(__rt_mutex_init);
/**
* rt_mutex_init_proxy_locked - initialize and lock a rt_mutex on behalf of a
* proxy owner
*
* @lock: the rt_mutex to be locked
* @proxy_owner:the task to set as owner
*
* No locking. Caller has to do serializing itself
*
* Special API call for PI-futex support. This initializes the rtmutex and
* assigns it to @proxy_owner. Concurrent operations on the rtmutex are not
* possible at this point because the pi_state which contains the rtmutex
* is not yet visible to other tasks.
*/
void rt_mutex_init_proxy_locked(struct rt_mutex *lock,
struct task_struct *proxy_owner)
{
__rt_mutex_init(lock, NULL, NULL);
debug_rt_mutex_proxy_lock(lock, proxy_owner);
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
rt_mutex_set_owner(lock, proxy_owner);
}
/**
* rt_mutex_proxy_unlock - release a lock on behalf of owner
*
* @lock: the rt_mutex to be locked
*
* No locking. Caller has to do serializing itself
*
* Special API call for PI-futex support. This merrily cleans up the rtmutex
* (debugging) state. Concurrent operations on this rt_mutex are not
* possible because it belongs to the pi_state which is about to be freed
* and it is not longer visible to other tasks.
*/
void rt_mutex_proxy_unlock(struct rt_mutex *lock,
struct task_struct *proxy_owner)
{
debug_rt_mutex_proxy_unlock(lock);
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
rt_mutex_set_owner(lock, NULL);
}
futex: Handle early deadlock return correctly commit 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") changed the locking rules in the futex code so that the hash bucket lock is not longer held while the waiter is enqueued into the rtmutex wait list. This made the lock and the unlock path symmetric, but unfortunately the possible early exit from __rt_mutex_proxy_start() due to a detected deadlock was not updated accordingly. That allows a concurrent unlocker to observe inconsitent state which triggers the warning in the unlock path. futex_lock_pi() futex_unlock_pi() lock(hb->lock) queue(hb_waiter) lock(hb->lock) lock(rtmutex->wait_lock) unlock(hb->lock) // acquired hb->lock hb_waiter = futex_top_waiter() lock(rtmutex->wait_lock) __rt_mutex_proxy_start() ---> fail remove(rtmutex_waiter); ---> returns -EDEADLOCK unlock(rtmutex->wait_lock) // acquired wait_lock wake_futex_pi() rt_mutex_next_owner() --> returns NULL --> WARN lock(hb->lock) unqueue(hb_waiter) The problem is caused by the remove(rtmutex_waiter) in the failure case of __rt_mutex_proxy_start() as this lets the unlocker observe a waiter in the hash bucket but no waiter on the rtmutex, i.e. inconsistent state. The original commit handles this correctly for the other early return cases (timeout, signal) by delaying the removal of the rtmutex waiter until the returning task reacquired the hash bucket lock. Treat the failure case of __rt_mutex_proxy_start() in the same way and let the existing cleanup code handle the eventual handover of the rtmutex gracefully. The regular rt_mutex_proxy_start() gains the rtmutex waiter removal for the failure case, so that the other callsites are still operating correctly. Add proper comments to the code so all these details are fully documented. Thanks to Peter for helping with the analysis and writing the really valuable code comments. Fixes: 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com> Co-developed-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: linux-s390@vger.kernel.org Cc: Stefan Liebler <stli@linux.ibm.com> Cc: Sebastian Sewior <bigeasy@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1901292311410.1950@nanos.tec.linutronix.de
2019-01-30 06:15:12 +08:00
/**
* __rt_mutex_start_proxy_lock() - Start lock acquisition for another task
* @lock: the rt_mutex to take
* @waiter: the pre-initialized rt_mutex_waiter
* @task: the task to prepare
*
* Starts the rt_mutex acquire; it enqueues the @waiter and does deadlock
* detection. It does not wait, see rt_mutex_wait_proxy_lock() for that.
*
* NOTE: does _NOT_ remove the @waiter on failure; must either call
* rt_mutex_wait_proxy_lock() or rt_mutex_cleanup_proxy_lock() after this.
*
* Returns:
* 0 - task blocked on lock
* 1 - acquired the lock for task, caller should wake it up
* <0 - error
*
* Special API call for PI-futex support.
*/
futex: Drop hb->lock before enqueueing on the rtmutex When PREEMPT_RT_FULL does the spinlock -> rt_mutex substitution the PI chain code will (falsely) report a deadlock and BUG. The problem is that it hold hb->lock (now an rt_mutex) while doing task_blocks_on_rt_mutex on the futex's pi_state::rtmutex. This, when interleaved just right with futex_unlock_pi() leads it to believe to see an AB-BA deadlock. Task1 (holds rt_mutex, Task2 (does FUTEX_LOCK_PI) does FUTEX_UNLOCK_PI) lock hb->lock lock rt_mutex (as per start_proxy) lock hb->lock Which is a trivial AB-BA. It is not an actual deadlock, because it won't be holding hb->lock by the time it actually blocks on the rt_mutex, but the chainwalk code doesn't know that and it would be a nightmare to handle this gracefully. To avoid this problem, do the same as in futex_unlock_pi() and drop hb->lock after acquiring wait_lock. This still fully serializes against futex_unlock_pi(), since adding to the wait_list does the very same lock dance, and removing it holds both locks. Aside of solving the RT problem this makes the lock and unlock mechanism symetric and reduces the hb->lock held time. Reported-and-tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Suggested-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: juri.lelli@arm.com Cc: xlpang@redhat.com Cc: rostedt@goodmis.org Cc: mathieu.desnoyers@efficios.com Cc: jdesfossez@efficios.com Cc: dvhart@infradead.org Cc: bristot@redhat.com Link: http://lkml.kernel.org/r/20170322104152.161341537@infradead.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-03-22 18:36:00 +08:00
int __rt_mutex_start_proxy_lock(struct rt_mutex *lock,
struct rt_mutex_waiter *waiter,
struct task_struct *task)
{
int ret;
futex: Handle early deadlock return correctly commit 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") changed the locking rules in the futex code so that the hash bucket lock is not longer held while the waiter is enqueued into the rtmutex wait list. This made the lock and the unlock path symmetric, but unfortunately the possible early exit from __rt_mutex_proxy_start() due to a detected deadlock was not updated accordingly. That allows a concurrent unlocker to observe inconsitent state which triggers the warning in the unlock path. futex_lock_pi() futex_unlock_pi() lock(hb->lock) queue(hb_waiter) lock(hb->lock) lock(rtmutex->wait_lock) unlock(hb->lock) // acquired hb->lock hb_waiter = futex_top_waiter() lock(rtmutex->wait_lock) __rt_mutex_proxy_start() ---> fail remove(rtmutex_waiter); ---> returns -EDEADLOCK unlock(rtmutex->wait_lock) // acquired wait_lock wake_futex_pi() rt_mutex_next_owner() --> returns NULL --> WARN lock(hb->lock) unqueue(hb_waiter) The problem is caused by the remove(rtmutex_waiter) in the failure case of __rt_mutex_proxy_start() as this lets the unlocker observe a waiter in the hash bucket but no waiter on the rtmutex, i.e. inconsistent state. The original commit handles this correctly for the other early return cases (timeout, signal) by delaying the removal of the rtmutex waiter until the returning task reacquired the hash bucket lock. Treat the failure case of __rt_mutex_proxy_start() in the same way and let the existing cleanup code handle the eventual handover of the rtmutex gracefully. The regular rt_mutex_proxy_start() gains the rtmutex waiter removal for the failure case, so that the other callsites are still operating correctly. Add proper comments to the code so all these details are fully documented. Thanks to Peter for helping with the analysis and writing the really valuable code comments. Fixes: 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com> Co-developed-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: linux-s390@vger.kernel.org Cc: Stefan Liebler <stli@linux.ibm.com> Cc: Sebastian Sewior <bigeasy@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1901292311410.1950@nanos.tec.linutronix.de
2019-01-30 06:15:12 +08:00
lockdep_assert_held(&lock->wait_lock);
futex: Drop hb->lock before enqueueing on the rtmutex When PREEMPT_RT_FULL does the spinlock -> rt_mutex substitution the PI chain code will (falsely) report a deadlock and BUG. The problem is that it hold hb->lock (now an rt_mutex) while doing task_blocks_on_rt_mutex on the futex's pi_state::rtmutex. This, when interleaved just right with futex_unlock_pi() leads it to believe to see an AB-BA deadlock. Task1 (holds rt_mutex, Task2 (does FUTEX_LOCK_PI) does FUTEX_UNLOCK_PI) lock hb->lock lock rt_mutex (as per start_proxy) lock hb->lock Which is a trivial AB-BA. It is not an actual deadlock, because it won't be holding hb->lock by the time it actually blocks on the rt_mutex, but the chainwalk code doesn't know that and it would be a nightmare to handle this gracefully. To avoid this problem, do the same as in futex_unlock_pi() and drop hb->lock after acquiring wait_lock. This still fully serializes against futex_unlock_pi(), since adding to the wait_list does the very same lock dance, and removing it holds both locks. Aside of solving the RT problem this makes the lock and unlock mechanism symetric and reduces the hb->lock held time. Reported-and-tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Suggested-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: juri.lelli@arm.com Cc: xlpang@redhat.com Cc: rostedt@goodmis.org Cc: mathieu.desnoyers@efficios.com Cc: jdesfossez@efficios.com Cc: dvhart@infradead.org Cc: bristot@redhat.com Link: http://lkml.kernel.org/r/20170322104152.161341537@infradead.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-03-22 18:36:00 +08:00
if (try_to_take_rt_mutex(lock, task, NULL))
return 1;
/* We enforce deadlock detection for futexes */
ret = task_blocks_on_rt_mutex(lock, waiter, task,
RT_MUTEX_FULL_CHAINWALK);
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
if (ret && !rt_mutex_owner(lock)) {
/*
* Reset the return value. We might have
* returned with -EDEADLK and the owner
* released the lock while we were walking the
* pi chain. Let the waiter sort it out.
*/
ret = 0;
}
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
debug_rt_mutex_print_deadlock(waiter);
return ret;
}
futex: Drop hb->lock before enqueueing on the rtmutex When PREEMPT_RT_FULL does the spinlock -> rt_mutex substitution the PI chain code will (falsely) report a deadlock and BUG. The problem is that it hold hb->lock (now an rt_mutex) while doing task_blocks_on_rt_mutex on the futex's pi_state::rtmutex. This, when interleaved just right with futex_unlock_pi() leads it to believe to see an AB-BA deadlock. Task1 (holds rt_mutex, Task2 (does FUTEX_LOCK_PI) does FUTEX_UNLOCK_PI) lock hb->lock lock rt_mutex (as per start_proxy) lock hb->lock Which is a trivial AB-BA. It is not an actual deadlock, because it won't be holding hb->lock by the time it actually blocks on the rt_mutex, but the chainwalk code doesn't know that and it would be a nightmare to handle this gracefully. To avoid this problem, do the same as in futex_unlock_pi() and drop hb->lock after acquiring wait_lock. This still fully serializes against futex_unlock_pi(), since adding to the wait_list does the very same lock dance, and removing it holds both locks. Aside of solving the RT problem this makes the lock and unlock mechanism symetric and reduces the hb->lock held time. Reported-and-tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Suggested-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: juri.lelli@arm.com Cc: xlpang@redhat.com Cc: rostedt@goodmis.org Cc: mathieu.desnoyers@efficios.com Cc: jdesfossez@efficios.com Cc: dvhart@infradead.org Cc: bristot@redhat.com Link: http://lkml.kernel.org/r/20170322104152.161341537@infradead.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-03-22 18:36:00 +08:00
/**
* rt_mutex_start_proxy_lock() - Start lock acquisition for another task
* @lock: the rt_mutex to take
* @waiter: the pre-initialized rt_mutex_waiter
* @task: the task to prepare
*
futex: Handle early deadlock return correctly commit 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") changed the locking rules in the futex code so that the hash bucket lock is not longer held while the waiter is enqueued into the rtmutex wait list. This made the lock and the unlock path symmetric, but unfortunately the possible early exit from __rt_mutex_proxy_start() due to a detected deadlock was not updated accordingly. That allows a concurrent unlocker to observe inconsitent state which triggers the warning in the unlock path. futex_lock_pi() futex_unlock_pi() lock(hb->lock) queue(hb_waiter) lock(hb->lock) lock(rtmutex->wait_lock) unlock(hb->lock) // acquired hb->lock hb_waiter = futex_top_waiter() lock(rtmutex->wait_lock) __rt_mutex_proxy_start() ---> fail remove(rtmutex_waiter); ---> returns -EDEADLOCK unlock(rtmutex->wait_lock) // acquired wait_lock wake_futex_pi() rt_mutex_next_owner() --> returns NULL --> WARN lock(hb->lock) unqueue(hb_waiter) The problem is caused by the remove(rtmutex_waiter) in the failure case of __rt_mutex_proxy_start() as this lets the unlocker observe a waiter in the hash bucket but no waiter on the rtmutex, i.e. inconsistent state. The original commit handles this correctly for the other early return cases (timeout, signal) by delaying the removal of the rtmutex waiter until the returning task reacquired the hash bucket lock. Treat the failure case of __rt_mutex_proxy_start() in the same way and let the existing cleanup code handle the eventual handover of the rtmutex gracefully. The regular rt_mutex_proxy_start() gains the rtmutex waiter removal for the failure case, so that the other callsites are still operating correctly. Add proper comments to the code so all these details are fully documented. Thanks to Peter for helping with the analysis and writing the really valuable code comments. Fixes: 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com> Co-developed-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: linux-s390@vger.kernel.org Cc: Stefan Liebler <stli@linux.ibm.com> Cc: Sebastian Sewior <bigeasy@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1901292311410.1950@nanos.tec.linutronix.de
2019-01-30 06:15:12 +08:00
* Starts the rt_mutex acquire; it enqueues the @waiter and does deadlock
* detection. It does not wait, see rt_mutex_wait_proxy_lock() for that.
*
* NOTE: unlike __rt_mutex_start_proxy_lock this _DOES_ remove the @waiter
* on failure.
*
futex: Drop hb->lock before enqueueing on the rtmutex When PREEMPT_RT_FULL does the spinlock -> rt_mutex substitution the PI chain code will (falsely) report a deadlock and BUG. The problem is that it hold hb->lock (now an rt_mutex) while doing task_blocks_on_rt_mutex on the futex's pi_state::rtmutex. This, when interleaved just right with futex_unlock_pi() leads it to believe to see an AB-BA deadlock. Task1 (holds rt_mutex, Task2 (does FUTEX_LOCK_PI) does FUTEX_UNLOCK_PI) lock hb->lock lock rt_mutex (as per start_proxy) lock hb->lock Which is a trivial AB-BA. It is not an actual deadlock, because it won't be holding hb->lock by the time it actually blocks on the rt_mutex, but the chainwalk code doesn't know that and it would be a nightmare to handle this gracefully. To avoid this problem, do the same as in futex_unlock_pi() and drop hb->lock after acquiring wait_lock. This still fully serializes against futex_unlock_pi(), since adding to the wait_list does the very same lock dance, and removing it holds both locks. Aside of solving the RT problem this makes the lock and unlock mechanism symetric and reduces the hb->lock held time. Reported-and-tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Suggested-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: juri.lelli@arm.com Cc: xlpang@redhat.com Cc: rostedt@goodmis.org Cc: mathieu.desnoyers@efficios.com Cc: jdesfossez@efficios.com Cc: dvhart@infradead.org Cc: bristot@redhat.com Link: http://lkml.kernel.org/r/20170322104152.161341537@infradead.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-03-22 18:36:00 +08:00
* Returns:
* 0 - task blocked on lock
* 1 - acquired the lock for task, caller should wake it up
* <0 - error
*
futex: Handle early deadlock return correctly commit 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") changed the locking rules in the futex code so that the hash bucket lock is not longer held while the waiter is enqueued into the rtmutex wait list. This made the lock and the unlock path symmetric, but unfortunately the possible early exit from __rt_mutex_proxy_start() due to a detected deadlock was not updated accordingly. That allows a concurrent unlocker to observe inconsitent state which triggers the warning in the unlock path. futex_lock_pi() futex_unlock_pi() lock(hb->lock) queue(hb_waiter) lock(hb->lock) lock(rtmutex->wait_lock) unlock(hb->lock) // acquired hb->lock hb_waiter = futex_top_waiter() lock(rtmutex->wait_lock) __rt_mutex_proxy_start() ---> fail remove(rtmutex_waiter); ---> returns -EDEADLOCK unlock(rtmutex->wait_lock) // acquired wait_lock wake_futex_pi() rt_mutex_next_owner() --> returns NULL --> WARN lock(hb->lock) unqueue(hb_waiter) The problem is caused by the remove(rtmutex_waiter) in the failure case of __rt_mutex_proxy_start() as this lets the unlocker observe a waiter in the hash bucket but no waiter on the rtmutex, i.e. inconsistent state. The original commit handles this correctly for the other early return cases (timeout, signal) by delaying the removal of the rtmutex waiter until the returning task reacquired the hash bucket lock. Treat the failure case of __rt_mutex_proxy_start() in the same way and let the existing cleanup code handle the eventual handover of the rtmutex gracefully. The regular rt_mutex_proxy_start() gains the rtmutex waiter removal for the failure case, so that the other callsites are still operating correctly. Add proper comments to the code so all these details are fully documented. Thanks to Peter for helping with the analysis and writing the really valuable code comments. Fixes: 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com> Co-developed-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: linux-s390@vger.kernel.org Cc: Stefan Liebler <stli@linux.ibm.com> Cc: Sebastian Sewior <bigeasy@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1901292311410.1950@nanos.tec.linutronix.de
2019-01-30 06:15:12 +08:00
* Special API call for PI-futex support.
futex: Drop hb->lock before enqueueing on the rtmutex When PREEMPT_RT_FULL does the spinlock -> rt_mutex substitution the PI chain code will (falsely) report a deadlock and BUG. The problem is that it hold hb->lock (now an rt_mutex) while doing task_blocks_on_rt_mutex on the futex's pi_state::rtmutex. This, when interleaved just right with futex_unlock_pi() leads it to believe to see an AB-BA deadlock. Task1 (holds rt_mutex, Task2 (does FUTEX_LOCK_PI) does FUTEX_UNLOCK_PI) lock hb->lock lock rt_mutex (as per start_proxy) lock hb->lock Which is a trivial AB-BA. It is not an actual deadlock, because it won't be holding hb->lock by the time it actually blocks on the rt_mutex, but the chainwalk code doesn't know that and it would be a nightmare to handle this gracefully. To avoid this problem, do the same as in futex_unlock_pi() and drop hb->lock after acquiring wait_lock. This still fully serializes against futex_unlock_pi(), since adding to the wait_list does the very same lock dance, and removing it holds both locks. Aside of solving the RT problem this makes the lock and unlock mechanism symetric and reduces the hb->lock held time. Reported-and-tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Suggested-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: juri.lelli@arm.com Cc: xlpang@redhat.com Cc: rostedt@goodmis.org Cc: mathieu.desnoyers@efficios.com Cc: jdesfossez@efficios.com Cc: dvhart@infradead.org Cc: bristot@redhat.com Link: http://lkml.kernel.org/r/20170322104152.161341537@infradead.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-03-22 18:36:00 +08:00
*/
int rt_mutex_start_proxy_lock(struct rt_mutex *lock,
struct rt_mutex_waiter *waiter,
struct task_struct *task)
{
int ret;
raw_spin_lock_irq(&lock->wait_lock);
ret = __rt_mutex_start_proxy_lock(lock, waiter, task);
futex: Handle early deadlock return correctly commit 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") changed the locking rules in the futex code so that the hash bucket lock is not longer held while the waiter is enqueued into the rtmutex wait list. This made the lock and the unlock path symmetric, but unfortunately the possible early exit from __rt_mutex_proxy_start() due to a detected deadlock was not updated accordingly. That allows a concurrent unlocker to observe inconsitent state which triggers the warning in the unlock path. futex_lock_pi() futex_unlock_pi() lock(hb->lock) queue(hb_waiter) lock(hb->lock) lock(rtmutex->wait_lock) unlock(hb->lock) // acquired hb->lock hb_waiter = futex_top_waiter() lock(rtmutex->wait_lock) __rt_mutex_proxy_start() ---> fail remove(rtmutex_waiter); ---> returns -EDEADLOCK unlock(rtmutex->wait_lock) // acquired wait_lock wake_futex_pi() rt_mutex_next_owner() --> returns NULL --> WARN lock(hb->lock) unqueue(hb_waiter) The problem is caused by the remove(rtmutex_waiter) in the failure case of __rt_mutex_proxy_start() as this lets the unlocker observe a waiter in the hash bucket but no waiter on the rtmutex, i.e. inconsistent state. The original commit handles this correctly for the other early return cases (timeout, signal) by delaying the removal of the rtmutex waiter until the returning task reacquired the hash bucket lock. Treat the failure case of __rt_mutex_proxy_start() in the same way and let the existing cleanup code handle the eventual handover of the rtmutex gracefully. The regular rt_mutex_proxy_start() gains the rtmutex waiter removal for the failure case, so that the other callsites are still operating correctly. Add proper comments to the code so all these details are fully documented. Thanks to Peter for helping with the analysis and writing the really valuable code comments. Fixes: 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com> Co-developed-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: linux-s390@vger.kernel.org Cc: Stefan Liebler <stli@linux.ibm.com> Cc: Sebastian Sewior <bigeasy@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1901292311410.1950@nanos.tec.linutronix.de
2019-01-30 06:15:12 +08:00
if (unlikely(ret))
remove_waiter(lock, waiter);
futex: Drop hb->lock before enqueueing on the rtmutex When PREEMPT_RT_FULL does the spinlock -> rt_mutex substitution the PI chain code will (falsely) report a deadlock and BUG. The problem is that it hold hb->lock (now an rt_mutex) while doing task_blocks_on_rt_mutex on the futex's pi_state::rtmutex. This, when interleaved just right with futex_unlock_pi() leads it to believe to see an AB-BA deadlock. Task1 (holds rt_mutex, Task2 (does FUTEX_LOCK_PI) does FUTEX_UNLOCK_PI) lock hb->lock lock rt_mutex (as per start_proxy) lock hb->lock Which is a trivial AB-BA. It is not an actual deadlock, because it won't be holding hb->lock by the time it actually blocks on the rt_mutex, but the chainwalk code doesn't know that and it would be a nightmare to handle this gracefully. To avoid this problem, do the same as in futex_unlock_pi() and drop hb->lock after acquiring wait_lock. This still fully serializes against futex_unlock_pi(), since adding to the wait_list does the very same lock dance, and removing it holds both locks. Aside of solving the RT problem this makes the lock and unlock mechanism symetric and reduces the hb->lock held time. Reported-and-tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Suggested-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: juri.lelli@arm.com Cc: xlpang@redhat.com Cc: rostedt@goodmis.org Cc: mathieu.desnoyers@efficios.com Cc: jdesfossez@efficios.com Cc: dvhart@infradead.org Cc: bristot@redhat.com Link: http://lkml.kernel.org/r/20170322104152.161341537@infradead.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-03-22 18:36:00 +08:00
raw_spin_unlock_irq(&lock->wait_lock);
return ret;
}
/**
* rt_mutex_next_owner - return the next owner of the lock
*
* @lock: the rt lock query
*
* Returns the next owner of the lock or NULL
*
* Caller has to serialize against other accessors to the lock
* itself.
*
* Special API call for PI-futex support
*/
struct task_struct *rt_mutex_next_owner(struct rt_mutex *lock)
{
if (!rt_mutex_has_waiters(lock))
return NULL;
return rt_mutex_top_waiter(lock)->task;
}
/**
* rt_mutex_wait_proxy_lock() - Wait for lock acquisition
* @lock: the rt_mutex we were woken on
* @to: the timeout, null if none. hrtimer should already have
* been started.
* @waiter: the pre-initialized rt_mutex_waiter
*
* Wait for the the lock acquisition started on our behalf by
* rt_mutex_start_proxy_lock(). Upon failure, the caller must call
* rt_mutex_cleanup_proxy_lock().
*
* Returns:
* 0 - success
* <0 - error, one of -EINTR, -ETIMEDOUT
*
* Special API call for PI-futex support
*/
int rt_mutex_wait_proxy_lock(struct rt_mutex *lock,
struct hrtimer_sleeper *to,
struct rt_mutex_waiter *waiter)
{
int ret;
raw_spin_lock_irq(&lock->wait_lock);
/* sleep on the mutex */
futex,rt_mutex: Fix rt_mutex_cleanup_proxy_lock() Markus reported that the glibc/nptl/tst-robustpi8 test was failing after commit: cfafcd117da0 ("futex: Rework futex_lock_pi() to use rt_mutex_*_proxy_lock()") The following trace shows the problem: ld-linux-x86-64-2161 [019] .... 410.760971: SyS_futex: 00007ffbeb76b028: 80000875 op=FUTEX_LOCK_PI ld-linux-x86-64-2161 [019] ...1 410.760972: lock_pi_update_atomic: 00007ffbeb76b028: curval=80000875 uval=80000875 newval=80000875 ret=0 ld-linux-x86-64-2165 [011] .... 410.760978: SyS_futex: 00007ffbeb76b028: 80000875 op=FUTEX_UNLOCK_PI ld-linux-x86-64-2165 [011] d..1 410.760979: do_futex: 00007ffbeb76b028: curval=80000875 uval=80000875 newval=80000871 ret=0 ld-linux-x86-64-2165 [011] .... 410.760980: SyS_futex: 00007ffbeb76b028: 80000871 ret=0000 ld-linux-x86-64-2161 [019] .... 410.760980: SyS_futex: 00007ffbeb76b028: 80000871 ret=ETIMEDOUT Task 2165 does an UNLOCK_PI, assigning the lock to the waiter task 2161 which then returns with -ETIMEDOUT. That wrecks the lock state, because now the owner isn't aware it acquired the lock and removes the pending robust list entry. If 2161 is killed, the robust list will not clear out this futex and the subsequent acquire on this futex will then (correctly) result in -ESRCH which is unexpected by glibc, triggers an internal assertion and dies. Task 2161 Task 2165 rt_mutex_wait_proxy_lock() timeout(); /* T2161 is still queued in the waiter list */ return -ETIMEDOUT; futex_unlock_pi() spin_lock(hb->lock); rtmutex_unlock() remove_rtmutex_waiter(T2161); mark_lock_available(); /* Make the next waiter owner of the user space side */ futex_uval = 2161; spin_unlock(hb->lock); spin_lock(hb->lock); rt_mutex_cleanup_proxy_lock() if (rtmutex_owner() !== current) ... return FAIL; .... return -ETIMEOUT; This means that rt_mutex_cleanup_proxy_lock() needs to call try_to_take_rt_mutex() so it can take over the rtmutex correctly which was assigned by the waker. If the rtmutex is owned by some other task then this call is harmless and just confirmes that the waiter is not able to acquire it. While there, fix what looks like a merge error which resulted in rt_mutex_cleanup_proxy_lock() having two calls to fixup_rt_mutex_waiters() and rt_mutex_wait_proxy_lock() not having any. Both should have one, since both potentially touch the waiter list. Fixes: 38d589f2fd08 ("futex,rt_mutex: Restructure rt_mutex_finish_proxy_lock()") Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de> Bug-Spotted-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: Darren Hart <dvhart@infradead.org> Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Markus Trippelsdorf <markus@trippelsdorf.de> Link: http://lkml.kernel.org/r/20170519154850.mlomgdsd26drq5j6@hirez.programming.kicks-ass.net Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-05-19 23:48:50 +08:00
set_current_state(TASK_INTERRUPTIBLE);
rtmutex: Simplify PI algorithm and make highest prio task get lock In current rtmutex, the pending owner may be boosted by the tasks in the rtmutex's waitlist when the pending owner is deboosted or a task in the waitlist is boosted. This boosting is unrelated, because the pending owner does not really take the rtmutex. It is not reasonable. Example. time1: A(high prio) onwers the rtmutex. B(mid prio) and C (low prio) in the waitlist. time2 A release the lock, B becomes the pending owner A(or other high prio task) continues to run. B's prio is lower than A, so B is just queued at the runqueue. time3 A or other high prio task sleeps, but we have passed some time The B and C's prio are changed in the period (time2 ~ time3) due to boosting or deboosting. Now C has the priority higher than B. ***Is it reasonable that C has to boost B and help B to get the rtmutex? NO!! I think, it is unrelated/unneed boosting before B really owns the rtmutex. We should give C a chance to beat B and win the rtmutex. This is the motivation of this patch. This patch *ensures* only the top waiter or higher priority task can take the lock. How? 1) we don't dequeue the top waiter when unlock, if the top waiter is changed, the old top waiter will fail and go to sleep again. 2) when requiring lock, it will get the lock when the lock is not taken and: there is no waiter OR higher priority than waiters OR it is top waiter. 3) In any time, the top waiter is changed, the top waiter will be woken up. The algorithm is much simpler than before, no pending owner, no boosting for pending owner. Other advantage of this patch: 1) The states of a rtmutex are reduced a half, easier to read the code. 2) the codes become shorter. 3) top waiter is not dequeued until it really take the lock: they will retain FIFO when it is stolen. Not advantage nor disadvantage 1) Even we may wakeup multiple waiters(any time when top waiter changed), we hardly cause "thundering herd", the number of wokenup task is likely 1 or very little. 2) two APIs are changed. rt_mutex_owner() will not return pending owner, it will return NULL when the top waiter is going to take the lock. rt_mutex_next_owner() always return the top waiter. will not return NULL if we have waiters because the top waiter is not dequeued. I have fixed the code that use these APIs. need updated after this patch is accepted 1) Document/* 2) the testcase scripts/rt-tester/t4-l2-pi-deboost.tst Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> LKML-Reference: <4D3012D5.4060709@cn.fujitsu.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2011-01-14 17:09:41 +08:00
ret = __rt_mutex_slowlock(lock, TASK_INTERRUPTIBLE, to, waiter);
futex,rt_mutex: Fix rt_mutex_cleanup_proxy_lock() Markus reported that the glibc/nptl/tst-robustpi8 test was failing after commit: cfafcd117da0 ("futex: Rework futex_lock_pi() to use rt_mutex_*_proxy_lock()") The following trace shows the problem: ld-linux-x86-64-2161 [019] .... 410.760971: SyS_futex: 00007ffbeb76b028: 80000875 op=FUTEX_LOCK_PI ld-linux-x86-64-2161 [019] ...1 410.760972: lock_pi_update_atomic: 00007ffbeb76b028: curval=80000875 uval=80000875 newval=80000875 ret=0 ld-linux-x86-64-2165 [011] .... 410.760978: SyS_futex: 00007ffbeb76b028: 80000875 op=FUTEX_UNLOCK_PI ld-linux-x86-64-2165 [011] d..1 410.760979: do_futex: 00007ffbeb76b028: curval=80000875 uval=80000875 newval=80000871 ret=0 ld-linux-x86-64-2165 [011] .... 410.760980: SyS_futex: 00007ffbeb76b028: 80000871 ret=0000 ld-linux-x86-64-2161 [019] .... 410.760980: SyS_futex: 00007ffbeb76b028: 80000871 ret=ETIMEDOUT Task 2165 does an UNLOCK_PI, assigning the lock to the waiter task 2161 which then returns with -ETIMEDOUT. That wrecks the lock state, because now the owner isn't aware it acquired the lock and removes the pending robust list entry. If 2161 is killed, the robust list will not clear out this futex and the subsequent acquire on this futex will then (correctly) result in -ESRCH which is unexpected by glibc, triggers an internal assertion and dies. Task 2161 Task 2165 rt_mutex_wait_proxy_lock() timeout(); /* T2161 is still queued in the waiter list */ return -ETIMEDOUT; futex_unlock_pi() spin_lock(hb->lock); rtmutex_unlock() remove_rtmutex_waiter(T2161); mark_lock_available(); /* Make the next waiter owner of the user space side */ futex_uval = 2161; spin_unlock(hb->lock); spin_lock(hb->lock); rt_mutex_cleanup_proxy_lock() if (rtmutex_owner() !== current) ... return FAIL; .... return -ETIMEOUT; This means that rt_mutex_cleanup_proxy_lock() needs to call try_to_take_rt_mutex() so it can take over the rtmutex correctly which was assigned by the waker. If the rtmutex is owned by some other task then this call is harmless and just confirmes that the waiter is not able to acquire it. While there, fix what looks like a merge error which resulted in rt_mutex_cleanup_proxy_lock() having two calls to fixup_rt_mutex_waiters() and rt_mutex_wait_proxy_lock() not having any. Both should have one, since both potentially touch the waiter list. Fixes: 38d589f2fd08 ("futex,rt_mutex: Restructure rt_mutex_finish_proxy_lock()") Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de> Bug-Spotted-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: Darren Hart <dvhart@infradead.org> Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Markus Trippelsdorf <markus@trippelsdorf.de> Link: http://lkml.kernel.org/r/20170519154850.mlomgdsd26drq5j6@hirez.programming.kicks-ass.net Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-05-19 23:48:50 +08:00
/*
* try_to_take_rt_mutex() sets the waiter bit unconditionally. We might
* have to fix that up.
*/
fixup_rt_mutex_waiters(lock);
raw_spin_unlock_irq(&lock->wait_lock);
return ret;
}
/**
* rt_mutex_cleanup_proxy_lock() - Cleanup failed lock acquisition
* @lock: the rt_mutex we were woken on
* @waiter: the pre-initialized rt_mutex_waiter
*
futex: Handle early deadlock return correctly commit 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") changed the locking rules in the futex code so that the hash bucket lock is not longer held while the waiter is enqueued into the rtmutex wait list. This made the lock and the unlock path symmetric, but unfortunately the possible early exit from __rt_mutex_proxy_start() due to a detected deadlock was not updated accordingly. That allows a concurrent unlocker to observe inconsitent state which triggers the warning in the unlock path. futex_lock_pi() futex_unlock_pi() lock(hb->lock) queue(hb_waiter) lock(hb->lock) lock(rtmutex->wait_lock) unlock(hb->lock) // acquired hb->lock hb_waiter = futex_top_waiter() lock(rtmutex->wait_lock) __rt_mutex_proxy_start() ---> fail remove(rtmutex_waiter); ---> returns -EDEADLOCK unlock(rtmutex->wait_lock) // acquired wait_lock wake_futex_pi() rt_mutex_next_owner() --> returns NULL --> WARN lock(hb->lock) unqueue(hb_waiter) The problem is caused by the remove(rtmutex_waiter) in the failure case of __rt_mutex_proxy_start() as this lets the unlocker observe a waiter in the hash bucket but no waiter on the rtmutex, i.e. inconsistent state. The original commit handles this correctly for the other early return cases (timeout, signal) by delaying the removal of the rtmutex waiter until the returning task reacquired the hash bucket lock. Treat the failure case of __rt_mutex_proxy_start() in the same way and let the existing cleanup code handle the eventual handover of the rtmutex gracefully. The regular rt_mutex_proxy_start() gains the rtmutex waiter removal for the failure case, so that the other callsites are still operating correctly. Add proper comments to the code so all these details are fully documented. Thanks to Peter for helping with the analysis and writing the really valuable code comments. Fixes: 56222b212e8e ("futex: Drop hb->lock before enqueueing on the rtmutex") Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com> Co-developed-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: linux-s390@vger.kernel.org Cc: Stefan Liebler <stli@linux.ibm.com> Cc: Sebastian Sewior <bigeasy@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1901292311410.1950@nanos.tec.linutronix.de
2019-01-30 06:15:12 +08:00
* Attempt to clean up after a failed __rt_mutex_start_proxy_lock() or
* rt_mutex_wait_proxy_lock().
*
* Unless we acquired the lock; we're still enqueued on the wait-list and can
* in fact still be granted ownership until we're removed. Therefore we can
* find we are in fact the owner and must disregard the
* rt_mutex_wait_proxy_lock() failure.
*
* Returns:
* true - did the cleanup, we done.
* false - we acquired the lock after rt_mutex_wait_proxy_lock() returned,
* caller should disregards its return value.
*
* Special API call for PI-futex support
*/
bool rt_mutex_cleanup_proxy_lock(struct rt_mutex *lock,
struct rt_mutex_waiter *waiter)
{
bool cleanup = false;
raw_spin_lock_irq(&lock->wait_lock);
futex,rt_mutex: Fix rt_mutex_cleanup_proxy_lock() Markus reported that the glibc/nptl/tst-robustpi8 test was failing after commit: cfafcd117da0 ("futex: Rework futex_lock_pi() to use rt_mutex_*_proxy_lock()") The following trace shows the problem: ld-linux-x86-64-2161 [019] .... 410.760971: SyS_futex: 00007ffbeb76b028: 80000875 op=FUTEX_LOCK_PI ld-linux-x86-64-2161 [019] ...1 410.760972: lock_pi_update_atomic: 00007ffbeb76b028: curval=80000875 uval=80000875 newval=80000875 ret=0 ld-linux-x86-64-2165 [011] .... 410.760978: SyS_futex: 00007ffbeb76b028: 80000875 op=FUTEX_UNLOCK_PI ld-linux-x86-64-2165 [011] d..1 410.760979: do_futex: 00007ffbeb76b028: curval=80000875 uval=80000875 newval=80000871 ret=0 ld-linux-x86-64-2165 [011] .... 410.760980: SyS_futex: 00007ffbeb76b028: 80000871 ret=0000 ld-linux-x86-64-2161 [019] .... 410.760980: SyS_futex: 00007ffbeb76b028: 80000871 ret=ETIMEDOUT Task 2165 does an UNLOCK_PI, assigning the lock to the waiter task 2161 which then returns with -ETIMEDOUT. That wrecks the lock state, because now the owner isn't aware it acquired the lock and removes the pending robust list entry. If 2161 is killed, the robust list will not clear out this futex and the subsequent acquire on this futex will then (correctly) result in -ESRCH which is unexpected by glibc, triggers an internal assertion and dies. Task 2161 Task 2165 rt_mutex_wait_proxy_lock() timeout(); /* T2161 is still queued in the waiter list */ return -ETIMEDOUT; futex_unlock_pi() spin_lock(hb->lock); rtmutex_unlock() remove_rtmutex_waiter(T2161); mark_lock_available(); /* Make the next waiter owner of the user space side */ futex_uval = 2161; spin_unlock(hb->lock); spin_lock(hb->lock); rt_mutex_cleanup_proxy_lock() if (rtmutex_owner() !== current) ... return FAIL; .... return -ETIMEOUT; This means that rt_mutex_cleanup_proxy_lock() needs to call try_to_take_rt_mutex() so it can take over the rtmutex correctly which was assigned by the waker. If the rtmutex is owned by some other task then this call is harmless and just confirmes that the waiter is not able to acquire it. While there, fix what looks like a merge error which resulted in rt_mutex_cleanup_proxy_lock() having two calls to fixup_rt_mutex_waiters() and rt_mutex_wait_proxy_lock() not having any. Both should have one, since both potentially touch the waiter list. Fixes: 38d589f2fd08 ("futex,rt_mutex: Restructure rt_mutex_finish_proxy_lock()") Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de> Bug-Spotted-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: Darren Hart <dvhart@infradead.org> Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Markus Trippelsdorf <markus@trippelsdorf.de> Link: http://lkml.kernel.org/r/20170519154850.mlomgdsd26drq5j6@hirez.programming.kicks-ass.net Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-05-19 23:48:50 +08:00
/*
* Do an unconditional try-lock, this deals with the lock stealing
* state where __rt_mutex_futex_unlock() -> mark_wakeup_next_waiter()
* sets a NULL owner.
*
* We're not interested in the return value, because the subsequent
* test on rt_mutex_owner() will infer that. If the trylock succeeded,
* we will own the lock and it will have removed the waiter. If we
* failed the trylock, we're still not owner and we need to remove
* ourselves.
*/
try_to_take_rt_mutex(lock, current, waiter);
/*
* Unless we're the owner; we're still enqueued on the wait_list.
* So check if we became owner, if not, take us off the wait_list.
*/
if (rt_mutex_owner(lock) != current) {
remove_waiter(lock, waiter);
cleanup = true;
}
futex: Rework futex_lock_pi() to use rt_mutex_*_proxy_lock() By changing futex_lock_pi() to use rt_mutex_*_proxy_lock() all wait_list modifications are done under both hb->lock and wait_lock. This closes the obvious interleave pattern between futex_lock_pi() and futex_unlock_pi(), but not entirely so. See below: Before: futex_lock_pi() futex_unlock_pi() unlock hb->lock lock hb->lock unlock hb->lock lock rt_mutex->wait_lock unlock rt_mutex_wait_lock -EAGAIN lock rt_mutex->wait_lock list_add unlock rt_mutex->wait_lock schedule() lock rt_mutex->wait_lock list_del unlock rt_mutex->wait_lock <idem> -EAGAIN lock hb->lock After: futex_lock_pi() futex_unlock_pi() lock hb->lock lock rt_mutex->wait_lock list_add unlock rt_mutex->wait_lock unlock hb->lock schedule() lock hb->lock unlock hb->lock lock hb->lock lock rt_mutex->wait_lock list_del unlock rt_mutex->wait_lock lock rt_mutex->wait_lock unlock rt_mutex_wait_lock -EAGAIN unlock hb->lock It does however solve the earlier starvation/live-lock scenario which got introduced with the -EAGAIN since unlike the before scenario; where the -EAGAIN happens while futex_unlock_pi() doesn't hold any locks; in the after scenario it happens while futex_unlock_pi() actually holds a lock, and then it is serialized on that lock. Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: juri.lelli@arm.com Cc: bigeasy@linutronix.de Cc: xlpang@redhat.com Cc: rostedt@goodmis.org Cc: mathieu.desnoyers@efficios.com Cc: jdesfossez@efficios.com Cc: dvhart@infradead.org Cc: bristot@redhat.com Link: http://lkml.kernel.org/r/20170322104152.062785528@infradead.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-03-22 18:35:58 +08:00
/*
* try_to_take_rt_mutex() sets the waiter bit unconditionally. We might
* have to fix that up.
*/
fixup_rt_mutex_waiters(lock);
raw_spin_unlock_irq(&lock->wait_lock);
return cleanup;
}