From owner-svn-src-stable@freebsd.org Thu Mar 16 01:32:58 2017 Return-Path: Delivered-To: svn-src-stable@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 9EDFED0DC95; Thu, 16 Mar 2017 01:32:58 +0000 (UTC) (envelope-from mjg@FreeBSD.org) Received: from repo.freebsd.org (repo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 6D8B015C7; Thu, 16 Mar 2017 01:32:58 +0000 (UTC) (envelope-from mjg@FreeBSD.org) Received: from repo.freebsd.org ([127.0.1.37]) by repo.freebsd.org (8.15.2/8.15.2) with ESMTP id v2G1WvRW058013; Thu, 16 Mar 2017 01:32:57 GMT (envelope-from mjg@FreeBSD.org) Received: (from mjg@localhost) by repo.freebsd.org (8.15.2/8.15.2/Submit) id v2G1Wvvi058007; Thu, 16 Mar 2017 01:32:57 GMT (envelope-from mjg@FreeBSD.org) Message-Id: <201703160132.v2G1Wvvi058007@repo.freebsd.org> X-Authentication-Warning: repo.freebsd.org: mjg set sender to mjg@FreeBSD.org using -f From: Mateusz Guzik Date: Thu, 16 Mar 2017 01:32:57 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-11@freebsd.org Subject: svn commit: r315341 - in stable/11/sys: kern sys X-SVN-Group: stable-11 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-stable@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: SVN commit messages for all the -stable branches of the src tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 16 Mar 2017 01:32:58 -0000 Author: mjg Date: Thu Mar 16 01:32:56 2017 New Revision: 315341 URL: https://svnweb.freebsd.org/changeset/base/315341 Log: MFC r311172,r311194,r311226,r312389,r312390: mtx: reduce lock accesses Instead of spuriously re-reading the lock value, read it once. This change also has a side effect of fixing a performance bug: on failed _mtx_obtain_lock, it was possible that re-read would find the lock is unowned, but in this case the primitive would make a trip through turnstile code. This is diff reduction to a variant which uses atomic_fcmpset. == Reduce lock accesses in thread lock similarly to r311172 == mtx: plug open-coded mtx_lock access missed in r311172 == rwlock: reduce lock accesses similarly to r311172 == sx: reduce lock accesses similarly to r311172 Modified: stable/11/sys/kern/kern_mutex.c stable/11/sys/kern/kern_rwlock.c stable/11/sys/kern/kern_sx.c stable/11/sys/sys/mutex.h stable/11/sys/sys/rwlock.h stable/11/sys/sys/sx.h Directory Properties: stable/11/ (props changed) Modified: stable/11/sys/kern/kern_mutex.c ============================================================================== --- stable/11/sys/kern/kern_mutex.c Thu Mar 16 01:04:30 2017 (r315340) +++ stable/11/sys/kern/kern_mutex.c Thu Mar 16 01:32:56 2017 (r315341) @@ -94,8 +94,6 @@ PMC_SOFT_DEFINE( , , lock, failed); #define mtx_destroyed(m) ((m)->mtx_lock == MTX_DESTROYED) -#define mtx_owner(m) ((struct thread *)((m)->mtx_lock & ~MTX_FLAGMASK)) - static void assert_mtx(const struct lock_object *lock, int what); #ifdef DDB static void db_show_mtx(const struct lock_object *lock); @@ -452,8 +450,9 @@ __mtx_lock_sleep(volatile uintptr_t *c, lock_delay_arg_init(&lda, NULL); #endif m = mtxlock2mtx(c); + v = MTX_READ_VALUE(m); - if (mtx_owned(m)) { + if (__predict_false(lv_mtx_owner(v) == (struct thread *)tid)) { KASSERT((m->lock_object.lo_flags & LO_RECURSABLE) != 0 || (opts & MTX_RECURSE) != 0, ("_mtx_lock_sleep: recursed on non-recursive mutex %s @ %s:%d\n", @@ -481,8 +480,12 @@ __mtx_lock_sleep(volatile uintptr_t *c, #endif for (;;) { - if (m->mtx_lock == MTX_UNOWNED && _mtx_obtain_lock(m, tid)) - break; + if (v == MTX_UNOWNED) { + if (_mtx_obtain_lock(m, tid)) + break; + v = MTX_READ_VALUE(m); + continue; + } #ifdef KDTRACE_HOOKS lda.spin_cnt++; #endif @@ -491,31 +494,30 @@ __mtx_lock_sleep(volatile uintptr_t *c, * If the owner is running on another CPU, spin until the * owner stops running or the state of the lock changes. */ - v = m->mtx_lock; - if (v != MTX_UNOWNED) { - owner = (struct thread *)(v & ~MTX_FLAGMASK); - if (TD_IS_RUNNING(owner)) { - if (LOCK_LOG_TEST(&m->lock_object, 0)) - CTR3(KTR_LOCK, - "%s: spinning on %p held by %p", - __func__, m, owner); - KTR_STATE1(KTR_SCHED, "thread", - sched_tdname((struct thread *)tid), - "spinning", "lockname:\"%s\"", - m->lock_object.lo_name); - while (mtx_owner(m) == owner && - TD_IS_RUNNING(owner)) - lock_delay(&lda); - KTR_STATE0(KTR_SCHED, "thread", - sched_tdname((struct thread *)tid), - "running"); - continue; - } + owner = lv_mtx_owner(v); + if (TD_IS_RUNNING(owner)) { + if (LOCK_LOG_TEST(&m->lock_object, 0)) + CTR3(KTR_LOCK, + "%s: spinning on %p held by %p", + __func__, m, owner); + KTR_STATE1(KTR_SCHED, "thread", + sched_tdname((struct thread *)tid), + "spinning", "lockname:\"%s\"", + m->lock_object.lo_name); + do { + lock_delay(&lda); + v = MTX_READ_VALUE(m); + owner = lv_mtx_owner(v); + } while (v != MTX_UNOWNED && TD_IS_RUNNING(owner)); + KTR_STATE0(KTR_SCHED, "thread", + sched_tdname((struct thread *)tid), + "running"); + continue; } #endif ts = turnstile_trywait(&m->lock_object); - v = m->mtx_lock; + v = MTX_READ_VALUE(m); /* * Check if the lock has been released while spinning for @@ -534,7 +536,7 @@ __mtx_lock_sleep(volatile uintptr_t *c, * chain lock. If so, drop the turnstile lock and try * again. */ - owner = (struct thread *)(v & ~MTX_FLAGMASK); + owner = lv_mtx_owner(v); if (TD_IS_RUNNING(owner)) { turnstile_cancel(ts); continue; @@ -549,6 +551,7 @@ __mtx_lock_sleep(volatile uintptr_t *c, if ((v & MTX_CONTESTED) == 0 && !atomic_cmpset_ptr(&m->mtx_lock, v, v | MTX_CONTESTED)) { turnstile_cancel(ts); + v = MTX_READ_VALUE(m); continue; } @@ -579,6 +582,7 @@ __mtx_lock_sleep(volatile uintptr_t *c, sleep_time += lockstat_nsecs(&m->lock_object); sleep_cnt++; #endif + v = MTX_READ_VALUE(m); } #ifdef KDTRACE_HOOKS all_time += lockstat_nsecs(&m->lock_object); @@ -636,6 +640,7 @@ _mtx_lock_spin_cookie(volatile uintptr_t { struct mtx *m; struct lock_delay_arg lda; + uintptr_t v; #ifdef LOCK_PROFILING int contested = 0; uint64_t waittime = 0; @@ -662,24 +667,30 @@ _mtx_lock_spin_cookie(volatile uintptr_t #ifdef KDTRACE_HOOKS spin_time -= lockstat_nsecs(&m->lock_object); #endif + v = MTX_READ_VALUE(m); for (;;) { - if (m->mtx_lock == MTX_UNOWNED && _mtx_obtain_lock(m, tid)) - break; + if (v == MTX_UNOWNED) { + if (_mtx_obtain_lock(m, tid)) + break; + v = MTX_READ_VALUE(m); + continue; + } /* Give interrupts a chance while we spin. */ spinlock_exit(); - while (m->mtx_lock != MTX_UNOWNED) { + do { if (lda.spin_cnt < 10000000) { lock_delay(&lda); - continue; + } else { + lda.spin_cnt++; + if (lda.spin_cnt < 60000000 || kdb_active || + panicstr != NULL) + DELAY(1); + else + _mtx_lock_spin_failed(m); + cpu_spinwait(); } - lda.spin_cnt++; - if (lda.spin_cnt < 60000000 || kdb_active || - panicstr != NULL) - DELAY(1); - else - _mtx_lock_spin_failed(m); - cpu_spinwait(); - } + v = MTX_READ_VALUE(m); + } while (v != MTX_UNOWNED); spinlock_enter(); } #ifdef KDTRACE_HOOKS @@ -704,7 +715,7 @@ void thread_lock_flags_(struct thread *td, int opts, const char *file, int line) { struct mtx *m; - uintptr_t tid; + uintptr_t tid, v; struct lock_delay_arg lda; #ifdef LOCK_PROFILING int contested = 0; @@ -746,10 +757,15 @@ retry: m->lock_object.lo_name, file, line)); WITNESS_CHECKORDER(&m->lock_object, opts | LOP_NEWORDER | LOP_EXCLUSIVE, file, line, NULL); + v = MTX_READ_VALUE(m); for (;;) { - if (m->mtx_lock == MTX_UNOWNED && _mtx_obtain_lock(m, tid)) - break; - if (m->mtx_lock == tid) { + if (v == MTX_UNOWNED) { + if (_mtx_obtain_lock(m, tid)) + break; + v = MTX_READ_VALUE(m); + continue; + } + if (v == tid) { m->mtx_recurse++; break; } @@ -760,7 +776,7 @@ retry: &contested, &waittime); /* Give interrupts a chance while we spin. */ spinlock_exit(); - while (m->mtx_lock != MTX_UNOWNED) { + do { if (lda.spin_cnt < 10000000) { lock_delay(&lda); } else { @@ -774,7 +790,8 @@ retry: } if (m != td->td_lock) goto retry; - } + v = MTX_READ_VALUE(m); + } while (v != MTX_UNOWNED); spinlock_enter(); } if (m == td->td_lock) Modified: stable/11/sys/kern/kern_rwlock.c ============================================================================== --- stable/11/sys/kern/kern_rwlock.c Thu Mar 16 01:04:30 2017 (r315340) +++ stable/11/sys/kern/kern_rwlock.c Thu Mar 16 01:32:56 2017 (r315341) @@ -114,9 +114,12 @@ LOCK_DELAY_SYSINIT_DEFAULT(rw_delay); * Return a pointer to the owning thread if the lock is write-locked or * NULL if the lock is unlocked or read-locked. */ -#define rw_wowner(rw) \ - ((rw)->rw_lock & RW_LOCK_READ ? NULL : \ - (struct thread *)RW_OWNER((rw)->rw_lock)) + +#define lv_rw_wowner(v) \ + ((v) & RW_LOCK_READ ? NULL : \ + (struct thread *)RW_OWNER((v))) + +#define rw_wowner(rw) lv_rw_wowner(RW_READ_VALUE(rw)) /* * Returns if a write owner is recursed. Write ownership is not assured @@ -397,7 +400,10 @@ __rw_rlock(volatile uintptr_t *c, const #ifdef KDTRACE_HOOKS all_time -= lockstat_nsecs(&rw->lock_object); - state = rw->rw_lock; +#endif + v = RW_READ_VALUE(rw); +#ifdef KDTRACE_HOOKS + state = v; #endif for (;;) { /* @@ -410,7 +416,6 @@ __rw_rlock(volatile uintptr_t *c, const * completely unlocked rwlock since such a lock is encoded * as a read lock with no waiters. */ - v = rw->rw_lock; if (RW_CAN_READ(v)) { /* * The RW_LOCK_READ_WAITERS flag should only be set @@ -426,6 +431,7 @@ __rw_rlock(volatile uintptr_t *c, const (void *)(v + RW_ONE_READER)); break; } + v = RW_READ_VALUE(rw); continue; } #ifdef KDTRACE_HOOKS @@ -453,9 +459,11 @@ __rw_rlock(volatile uintptr_t *c, const KTR_STATE1(KTR_SCHED, "thread", sched_tdname(curthread), "spinning", "lockname:\"%s\"", rw->lock_object.lo_name); - while ((struct thread*)RW_OWNER(rw->rw_lock) == - owner && TD_IS_RUNNING(owner)) + do { lock_delay(&lda); + v = RW_READ_VALUE(rw); + owner = lv_rw_wowner(v); + } while (owner != NULL && TD_IS_RUNNING(owner)); KTR_STATE0(KTR_SCHED, "thread", sched_tdname(curthread), "running"); continue; @@ -466,11 +474,12 @@ __rw_rlock(volatile uintptr_t *c, const "spinning", "lockname:\"%s\"", rw->lock_object.lo_name); for (i = 0; i < rowner_loops; i++) { - v = rw->rw_lock; + v = RW_READ_VALUE(rw); if ((v & RW_LOCK_READ) == 0 || RW_CAN_READ(v)) break; cpu_spinwait(); } + v = RW_READ_VALUE(rw); #ifdef KDTRACE_HOOKS lda.spin_cnt += rowner_loops - i; #endif @@ -493,7 +502,7 @@ __rw_rlock(volatile uintptr_t *c, const * The lock might have been released while we spun, so * recheck its state and restart the loop if needed. */ - v = rw->rw_lock; + v = RW_READ_VALUE(rw); if (RW_CAN_READ(v)) { turnstile_cancel(ts); continue; @@ -531,6 +540,7 @@ __rw_rlock(volatile uintptr_t *c, const if (!atomic_cmpset_ptr(&rw->rw_lock, v, v | RW_LOCK_READ_WAITERS)) { turnstile_cancel(ts); + v = RW_READ_VALUE(rw); continue; } if (LOCK_LOG_TEST(&rw->lock_object, 0)) @@ -556,6 +566,7 @@ __rw_rlock(volatile uintptr_t *c, const if (LOCK_LOG_TEST(&rw->lock_object, 0)) CTR2(KTR_LOCK, "%s: %p resuming from turnstile", __func__, rw); + v = RW_READ_VALUE(rw); } #ifdef KDTRACE_HOOKS all_time += lockstat_nsecs(&rw->lock_object); @@ -639,13 +650,12 @@ _rw_runlock_cookie(volatile uintptr_t *c LOCK_LOG_LOCK("RUNLOCK", &rw->lock_object, 0, 0, file, line); /* TODO: drop "owner of record" here. */ - + x = RW_READ_VALUE(rw); for (;;) { /* * See if there is more than one read lock held. If so, * just drop one and return. */ - x = rw->rw_lock; if (RW_READERS(x) > 1) { if (atomic_cmpset_rel_ptr(&rw->rw_lock, x, x - RW_ONE_READER)) { @@ -656,6 +666,7 @@ _rw_runlock_cookie(volatile uintptr_t *c (void *)(x - RW_ONE_READER)); break; } + x = RW_READ_VALUE(rw); continue; } /* @@ -672,6 +683,7 @@ _rw_runlock_cookie(volatile uintptr_t *c __func__, rw); break; } + x = RW_READ_VALUE(rw); continue; } /* @@ -707,6 +719,7 @@ _rw_runlock_cookie(volatile uintptr_t *c if (!atomic_cmpset_rel_ptr(&rw->rw_lock, RW_READERS_LOCK(1) | v, x)) { turnstile_chain_unlock(&rw->lock_object); + x = RW_READ_VALUE(rw); continue; } if (LOCK_LOG_TEST(&rw->lock_object, 0)) @@ -772,8 +785,9 @@ __rw_wlock_hard(volatile uintptr_t *c, u lock_delay_arg_init(&lda, NULL); #endif rw = rwlock2rw(c); + v = RW_READ_VALUE(rw); - if (rw_wlocked(rw)) { + if (__predict_false(lv_rw_wowner(v) == (struct thread *)tid)) { KASSERT(rw->lock_object.lo_flags & LO_RECURSABLE, ("%s: recursing but non-recursive rw %s @ %s:%d\n", __func__, rw->lock_object.lo_name, file, line)); @@ -789,11 +803,15 @@ __rw_wlock_hard(volatile uintptr_t *c, u #ifdef KDTRACE_HOOKS all_time -= lockstat_nsecs(&rw->lock_object); - state = rw->rw_lock; + state = v; #endif for (;;) { - if (rw->rw_lock == RW_UNLOCKED && _rw_write_lock(rw, tid)) - break; + if (v == RW_UNLOCKED) { + if (_rw_write_lock(rw, tid)) + break; + v = RW_READ_VALUE(rw); + continue; + } #ifdef KDTRACE_HOOKS lda.spin_cnt++; #endif @@ -808,8 +826,7 @@ __rw_wlock_hard(volatile uintptr_t *c, u * running on another CPU, spin until the owner stops * running or the state of the lock changes. */ - v = rw->rw_lock; - owner = (struct thread *)RW_OWNER(v); + owner = lv_rw_wowner(v); if (!(v & RW_LOCK_READ) && TD_IS_RUNNING(owner)) { if (LOCK_LOG_TEST(&rw->lock_object, 0)) CTR3(KTR_LOCK, "%s: spinning on %p held by %p", @@ -817,9 +834,11 @@ __rw_wlock_hard(volatile uintptr_t *c, u KTR_STATE1(KTR_SCHED, "thread", sched_tdname(curthread), "spinning", "lockname:\"%s\"", rw->lock_object.lo_name); - while ((struct thread*)RW_OWNER(rw->rw_lock) == owner && - TD_IS_RUNNING(owner)) + do { lock_delay(&lda); + v = RW_READ_VALUE(rw); + owner = lv_rw_wowner(v); + } while (owner != NULL && TD_IS_RUNNING(owner)); KTR_STATE0(KTR_SCHED, "thread", sched_tdname(curthread), "running"); continue; @@ -829,6 +848,7 @@ __rw_wlock_hard(volatile uintptr_t *c, u if (!(v & RW_LOCK_WRITE_SPINNER)) { if (!atomic_cmpset_ptr(&rw->rw_lock, v, v | RW_LOCK_WRITE_SPINNER)) { + v = RW_READ_VALUE(rw); continue; } } @@ -843,6 +863,7 @@ __rw_wlock_hard(volatile uintptr_t *c, u } KTR_STATE0(KTR_SCHED, "thread", sched_tdname(curthread), "running"); + v = RW_READ_VALUE(rw); #ifdef KDTRACE_HOOKS lda.spin_cnt += rowner_loops - i; #endif @@ -851,7 +872,7 @@ __rw_wlock_hard(volatile uintptr_t *c, u } #endif ts = turnstile_trywait(&rw->lock_object); - v = rw->rw_lock; + v = RW_READ_VALUE(rw); #ifdef ADAPTIVE_RWLOCKS /* @@ -887,6 +908,7 @@ __rw_wlock_hard(volatile uintptr_t *c, u break; } turnstile_cancel(ts); + v = RW_READ_VALUE(rw); continue; } /* @@ -898,6 +920,7 @@ __rw_wlock_hard(volatile uintptr_t *c, u if (!atomic_cmpset_ptr(&rw->rw_lock, v, v | RW_LOCK_WRITE_WAITERS)) { turnstile_cancel(ts); + v = RW_READ_VALUE(rw); continue; } if (LOCK_LOG_TEST(&rw->lock_object, 0)) @@ -925,6 +948,7 @@ __rw_wlock_hard(volatile uintptr_t *c, u #ifdef ADAPTIVE_RWLOCKS spintries = 0; #endif + v = RW_READ_VALUE(rw); } #ifdef KDTRACE_HOOKS all_time += lockstat_nsecs(&rw->lock_object); Modified: stable/11/sys/kern/kern_sx.c ============================================================================== --- stable/11/sys/kern/kern_sx.c Thu Mar 16 01:04:30 2017 (r315340) +++ stable/11/sys/kern/kern_sx.c Thu Mar 16 01:32:56 2017 (r315341) @@ -543,8 +543,10 @@ _sx_xlock_hard(struct sx *sx, uintptr_t lock_delay_arg_init(&lda, NULL); #endif + x = SX_READ_VALUE(sx); + /* If we already hold an exclusive lock, then recurse. */ - if (sx_xlocked(sx)) { + if (__predict_false(lv_sx_owner(x) == (struct thread *)tid)) { KASSERT((sx->lock_object.lo_flags & LO_RECURSABLE) != 0, ("_sx_xlock_hard: recursed on non-recursive sx %s @ %s:%d\n", sx->lock_object.lo_name, file, line)); @@ -561,12 +563,15 @@ _sx_xlock_hard(struct sx *sx, uintptr_t #ifdef KDTRACE_HOOKS all_time -= lockstat_nsecs(&sx->lock_object); - state = sx->sx_lock; + state = x; #endif for (;;) { - if (sx->sx_lock == SX_LOCK_UNLOCKED && - atomic_cmpset_acq_ptr(&sx->sx_lock, SX_LOCK_UNLOCKED, tid)) - break; + if (x == SX_LOCK_UNLOCKED) { + if (atomic_cmpset_acq_ptr(&sx->sx_lock, x, tid)) + break; + x = SX_READ_VALUE(sx); + continue; + } #ifdef KDTRACE_HOOKS lda.spin_cnt++; #endif @@ -581,11 +586,9 @@ _sx_xlock_hard(struct sx *sx, uintptr_t * running on another CPU, spin until the owner stops * running or the state of the lock changes. */ - x = sx->sx_lock; if ((sx->lock_object.lo_flags & SX_NOADAPTIVE) == 0) { if ((x & SX_LOCK_SHARED) == 0) { - x = SX_OWNER(x); - owner = (struct thread *)x; + owner = lv_sx_owner(x); if (TD_IS_RUNNING(owner)) { if (LOCK_LOG_TEST(&sx->lock_object, 0)) CTR3(KTR_LOCK, @@ -596,9 +599,12 @@ _sx_xlock_hard(struct sx *sx, uintptr_t "lockname:\"%s\"", sx->lock_object.lo_name); GIANT_SAVE(); - while (SX_OWNER(sx->sx_lock) == x && - TD_IS_RUNNING(owner)) + do { lock_delay(&lda); + x = SX_READ_VALUE(sx); + owner = lv_sx_owner(x); + } while (owner != NULL && + TD_IS_RUNNING(owner)); KTR_STATE0(KTR_SCHED, "thread", sched_tdname(curthread), "running"); continue; @@ -625,6 +631,7 @@ _sx_xlock_hard(struct sx *sx, uintptr_t } KTR_STATE0(KTR_SCHED, "thread", sched_tdname(curthread), "running"); + x = SX_READ_VALUE(sx); if (i != asx_loops) continue; } @@ -632,7 +639,7 @@ _sx_xlock_hard(struct sx *sx, uintptr_t #endif sleepq_lock(&sx->lock_object); - x = sx->sx_lock; + x = SX_READ_VALUE(sx); /* * If the lock was released while spinning on the @@ -681,6 +688,7 @@ _sx_xlock_hard(struct sx *sx, uintptr_t break; } sleepq_release(&sx->lock_object); + x = SX_READ_VALUE(sx); continue; } @@ -692,6 +700,7 @@ _sx_xlock_hard(struct sx *sx, uintptr_t if (!atomic_cmpset_ptr(&sx->sx_lock, x, x | SX_LOCK_EXCLUSIVE_WAITERS)) { sleepq_release(&sx->lock_object); + x = SX_READ_VALUE(sx); continue; } if (LOCK_LOG_TEST(&sx->lock_object, 0)) @@ -733,6 +742,7 @@ _sx_xlock_hard(struct sx *sx, uintptr_t if (LOCK_LOG_TEST(&sx->lock_object, 0)) CTR2(KTR_LOCK, "%s: %p resuming from sleep queue", __func__, sx); + x = SX_READ_VALUE(sx); } #ifdef KDTRACE_HOOKS all_time += lockstat_nsecs(&sx->lock_object); @@ -852,20 +862,18 @@ _sx_slock_hard(struct sx *sx, int opts, lock_delay_arg_init(&lda, NULL); #endif #ifdef KDTRACE_HOOKS - state = sx->sx_lock; all_time -= lockstat_nsecs(&sx->lock_object); #endif + x = SX_READ_VALUE(sx); +#ifdef KDTRACE_HOOKS + state = x; +#endif /* * As with rwlocks, we don't make any attempt to try to block * shared locks once there is an exclusive waiter. */ for (;;) { -#ifdef KDTRACE_HOOKS - lda.spin_cnt++; -#endif - x = sx->sx_lock; - /* * If no other thread has an exclusive lock then try to bump up * the count of sharers. Since we have to preserve the state @@ -883,8 +891,13 @@ _sx_slock_hard(struct sx *sx, int opts, (void *)(x + SX_ONE_SHARER)); break; } + x = SX_READ_VALUE(sx); continue; } +#ifdef KDTRACE_HOOKS + lda.spin_cnt++; +#endif + #ifdef HWPMC_HOOKS PMC_SOFT_CALL( , , lock, failed); #endif @@ -898,8 +911,7 @@ _sx_slock_hard(struct sx *sx, int opts, * changes. */ if ((sx->lock_object.lo_flags & SX_NOADAPTIVE) == 0) { - x = SX_OWNER(x); - owner = (struct thread *)x; + owner = lv_sx_owner(x); if (TD_IS_RUNNING(owner)) { if (LOCK_LOG_TEST(&sx->lock_object, 0)) CTR3(KTR_LOCK, @@ -909,9 +921,11 @@ _sx_slock_hard(struct sx *sx, int opts, sched_tdname(curthread), "spinning", "lockname:\"%s\"", sx->lock_object.lo_name); GIANT_SAVE(); - while (SX_OWNER(sx->sx_lock) == x && - TD_IS_RUNNING(owner)) + do { lock_delay(&lda); + x = SX_READ_VALUE(sx); + owner = lv_sx_owner(x); + } while (owner != NULL && TD_IS_RUNNING(owner)); KTR_STATE0(KTR_SCHED, "thread", sched_tdname(curthread), "running"); continue; @@ -924,7 +938,7 @@ _sx_slock_hard(struct sx *sx, int opts, * start the process of blocking. */ sleepq_lock(&sx->lock_object); - x = sx->sx_lock; + x = SX_READ_VALUE(sx); /* * The lock could have been released while we spun. @@ -946,6 +960,7 @@ _sx_slock_hard(struct sx *sx, int opts, owner = (struct thread *)SX_OWNER(x); if (TD_IS_RUNNING(owner)) { sleepq_release(&sx->lock_object); + x = SX_READ_VALUE(sx); continue; } } @@ -960,6 +975,7 @@ _sx_slock_hard(struct sx *sx, int opts, if (!atomic_cmpset_ptr(&sx->sx_lock, x, x | SX_LOCK_SHARED_WAITERS)) { sleepq_release(&sx->lock_object); + x = SX_READ_VALUE(sx); continue; } if (LOCK_LOG_TEST(&sx->lock_object, 0)) @@ -1000,6 +1016,7 @@ _sx_slock_hard(struct sx *sx, int opts, if (LOCK_LOG_TEST(&sx->lock_object, 0)) CTR2(KTR_LOCK, "%s: %p resuming from sleep queue", __func__, sx); + x = SX_READ_VALUE(sx); } #ifdef KDTRACE_HOOKS all_time += lockstat_nsecs(&sx->lock_object); @@ -1034,9 +1051,8 @@ _sx_sunlock_hard(struct sx *sx, const ch if (SCHEDULER_STOPPED()) return; + x = SX_READ_VALUE(sx); for (;;) { - x = sx->sx_lock; - /* * We should never have sharers while at least one thread * holds a shared lock. @@ -1058,6 +1074,8 @@ _sx_sunlock_hard(struct sx *sx, const ch (void *)(x - SX_ONE_SHARER)); break; } + + x = SX_READ_VALUE(sx); continue; } @@ -1074,6 +1092,7 @@ _sx_sunlock_hard(struct sx *sx, const ch __func__, sx); break; } + x = SX_READ_VALUE(sx); continue; } @@ -1095,6 +1114,7 @@ _sx_sunlock_hard(struct sx *sx, const ch SX_SHARERS_LOCK(1) | SX_LOCK_EXCLUSIVE_WAITERS, SX_LOCK_UNLOCKED)) { sleepq_release(&sx->lock_object); + x = SX_READ_VALUE(sx); continue; } if (LOCK_LOG_TEST(&sx->lock_object, 0)) Modified: stable/11/sys/sys/mutex.h ============================================================================== --- stable/11/sys/sys/mutex.h Thu Mar 16 01:04:30 2017 (r315340) +++ stable/11/sys/sys/mutex.h Thu Mar 16 01:32:56 2017 (r315341) @@ -420,9 +420,15 @@ extern struct mtx_pool *mtxpool_sleep; _sleep((chan), &(mtx)->lock_object, (pri), (wmesg), \ tick_sbt * (timo), 0, C_HARDCLOCK) +#define MTX_READ_VALUE(m) ((m)->mtx_lock) + #define mtx_initialized(m) lock_initialized(&(m)->lock_object) -#define mtx_owned(m) (((m)->mtx_lock & ~MTX_FLAGMASK) == (uintptr_t)curthread) +#define lv_mtx_owner(v) ((struct thread *)((v) & ~MTX_FLAGMASK)) + +#define mtx_owner(m) lv_mtx_owner(MTX_READ_VALUE(m)) + +#define mtx_owned(m) (mtx_owner(m) == curthread) #define mtx_recursed(m) ((m)->mtx_recurse != 0) Modified: stable/11/sys/sys/rwlock.h ============================================================================== --- stable/11/sys/sys/rwlock.h Thu Mar 16 01:04:30 2017 (r315340) +++ stable/11/sys/sys/rwlock.h Thu Mar 16 01:32:56 2017 (r315341) @@ -76,6 +76,8 @@ #define rw_recurse lock_object.lo_data +#define RW_READ_VALUE(x) ((x)->rw_lock) + /* Very simple operations on rw_lock. */ /* Try to obtain a write lock once. */ Modified: stable/11/sys/sys/sx.h ============================================================================== --- stable/11/sys/sys/sx.h Thu Mar 16 01:04:30 2017 (r315340) +++ stable/11/sys/sys/sx.h Thu Mar 16 01:32:56 2017 (r315341) @@ -88,6 +88,11 @@ #define sx_recurse lock_object.lo_data +#define SX_READ_VALUE(sx) ((sx)->sx_lock) + +#define lv_sx_owner(v) \ + ((v & SX_LOCK_SHARED) ? NULL : (struct thread *)SX_OWNER(v)) + /* * Function prototipes. Routines that start with an underscore are not part * of the public interface and are wrappered with a macro.