Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

allow cv_signal() immediately followed by cv_destroy().

Browse Source 
this sequence is used by ZFS in a couple places and by supporting it
natively we can undo our local ZFS changes that avoided it.
note that this is only legal when all of the waiters use cv_wait()
and not any of the other variations, and lockdebug will catch
any violations of this rule.
This commit is contained in:
chs 2017-06-08 01:09:52 +00:00
parent da7de9a430
commit 67c81802f1
2 changed files with 45 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
sys/kern/kern_condvar.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: kern_condvar.c,v 1.35 2015/08/07 06:22:12 uebayasi Exp $ */
/* $NetBSD: kern_condvar.c,v 1.36 2017/06/08 01:09:52 chs Exp $ */
/*-
* Copyright (c) 2006, 2007, 2008 The NetBSD Foundation, Inc.
@ -34,7 +34,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: kern_condvar.c,v 1.35 2015/08/07 06:22:12 uebayasi Exp $");
__KERNEL_RCSID(0, "$NetBSD: kern_condvar.c,v 1.36 2017/06/08 01:09:52 chs Exp $");
#include <sys/param.h>
#include <sys/systm.h>
@ -65,9 +65,9 @@ __KERNEL_RCSID(0, "$NetBSD: kern_condvar.c,v 1.35 2015/08/07 06:22:12 uebayasi E
#define CV_DEBUG_P(cv) (CV_WMESG(cv) != nodebug)
#define CV_RA ((uintptr_t)__builtin_return_address(0))
static void cv_unsleep(lwp_t *, bool);
static void cv_wakeup_one(kcondvar_t *);
static void cv_wakeup_all(kcondvar_t *);
static void cv_unsleep(lwp_t *, bool);
static inline void cv_wakeup_one(kcondvar_t *);
static inline void cv_wakeup_all(kcondvar_t *);
static syncobj_t cv_syncobj = {
SOBJ_SLEEPQ_SORTED,
@ -86,6 +86,31 @@ lockops_t cv_lockops = {
static const char deadcv[] = "deadcv";
#ifdef LOCKDEBUG
static const char nodebug[] = "nodebug";
#define CV_LOCKDEBUG_HANDOFF(l, cv) cv_lockdebug_handoff(l, cv)
#define CV_LOCKDEBUG_PROCESS(l, cv) cv_lockdebug_process(l, cv)
static inline void
cv_lockdebug_handoff(lwp_t *l, kcondvar_t *cv)
{
if (CV_DEBUG_P(cv))
l->l_flag |= LW_CVLOCKDEBUG;
}
static inline void
cv_lockdebug_process(lwp_t *l, kcondvar_t *cv)
{
if ((l->l_flag & LW_CVLOCKDEBUG) == 0)
return;
l->l_flag &= ~LW_CVLOCKDEBUG;
LOCKDEBUG_UNLOCKED(true, cv, CV_RA, 0);
}
#else
#define CV_LOCKDEBUG_HANDOFF(l, cv) __nothing
#define CV_LOCKDEBUG_PROCESS(l, cv) __nothing
#endif
/*
@ -214,8 +239,16 @@ cv_wait(kcondvar_t *cv, kmutex_t *mtx)
KASSERT(mutex_owned(mtx));
cv_enter(cv, mtx, l);
/*
* We can't use cv_exit() here since the cv might be destroyed before
* this thread gets a chance to run. Instead, hand off the lockdebug
* responsibility to the thread that wakes us up.
*/
CV_LOCKDEBUG_HANDOFF(l, cv);
(void)sleepq_block(0, false);
(void)cv_exit(cv, mtx, l, 0);
mutex_enter(mtx);
}
/*
@ -302,7 +335,7 @@ cv_signal(kcondvar_t *cv)
cv_wakeup_one(cv);
}
static void __noinline
static inline void
cv_wakeup_one(kcondvar_t *cv)
{
sleepq_t *sq;
@ -321,6 +354,7 @@ cv_wakeup_one(kcondvar_t *cv)
KASSERT(l->l_sleepq == sq);
KASSERT(l->l_mutex == mp);
KASSERT(l->l_wchan == cv);
CV_LOCKDEBUG_PROCESS(l, cv);
sleepq_remove(sq, l);
mutex_spin_exit(mp);
@ -344,7 +378,7 @@ cv_broadcast(kcondvar_t *cv)
cv_wakeup_all(cv);
}
static void __noinline
static inline void
cv_wakeup_all(kcondvar_t *cv)
{
sleepq_t *sq;
@ -360,6 +394,7 @@ cv_wakeup_all(kcondvar_t *cv)
KASSERT(l->l_mutex == mp);
KASSERT(l->l_wchan == cv);
next = TAILQ_NEXT(l, l_sleepchain);
CV_LOCKDEBUG_PROCESS(l, cv);
sleepq_remove(sq, l);
}
mutex_spin_exit(mp);

3
sys/sys/lwp.h
View File

@ -1,4 +1,4 @@
/* $NetBSD: lwp.h,v 1.174 2017/04/21 15:10:35 christos Exp $ */
/* $NetBSD: lwp.h,v 1.175 2017/06/08 01:09:52 chs Exp $ */
/*-
* Copyright (c) 2001, 2006, 2007, 2008, 2009, 2010
@ -231,6 +231,7 @@ extern int maxlwp __read_mostly; /* max number of lwps */
/* These flags are kept in l_flag. */
#define LW_IDLE 0x00000001 /* Idle lwp. */
#define LW_LWPCTL 0x00000002 /* Adjust lwpctl in userret */
#define LW_CVLOCKDEBUG 0x00000004 /* Waker does lockdebug */
#define LW_SINTR 0x00000080 /* Sleep is interruptible. */
#define LW_SYSTEM 0x00000200 /* Kernel thread */
#define LW_WSUSPEND 0x00020000 /* Suspend before return to user */