]> git.hungrycats.org Git - linux/commitdiff
locking/ww_mutex: Report recursive ww_mutex locking early
authorChris Wilson <chris@chris-wilson.co.uk>
Thu, 26 May 2016 20:08:17 +0000 (21:08 +0100)
committerSasha Levin <sasha.levin@oracle.com>
Mon, 20 Jun 2016 03:47:43 +0000 (23:47 -0400)
[ Upstream commit 0422e83d84ae24b933e4b0d4c1e0f0b4ae8a0a3b ]

Recursive locking for ww_mutexes was originally conceived as an
exception. However, it is heavily used by the DRM atomic modesetting
code. Currently, the recursive deadlock is checked after we have queued
up for a busy-spin and as we never release the lock, we spin until
kicked, whereupon the deadlock is discovered and reported.

A simple solution for the now common problem is to move the recursive
deadlock discovery to the first action when taking the ww_mutex.

Suggested-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/1464293297-19777-1-git-send-email-chris@chris-wilson.co.uk
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
kernel/locking/mutex.c

index dadbf88c22c48cea044aac4e34fda9deec78636a..e3fe1dda72ca8251209d5632e86a4c7285fc65ba 100644 (file)
@@ -471,9 +471,6 @@ __mutex_lock_check_stamp(struct mutex *lock, struct ww_acquire_ctx *ctx)
        if (!hold_ctx)
                return 0;
 
-       if (unlikely(ctx == hold_ctx))
-               return -EALREADY;
-
        if (ctx->stamp - hold_ctx->stamp <= LONG_MAX &&
            (ctx->stamp != hold_ctx->stamp || ctx > hold_ctx)) {
 #ifdef CONFIG_DEBUG_MUTEXES
@@ -499,6 +496,12 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass,
        unsigned long flags;
        int ret;
 
+       if (use_ww_ctx) {
+               struct ww_mutex *ww = container_of(lock, struct ww_mutex, base);
+               if (unlikely(ww_ctx == READ_ONCE(ww->ctx)))
+                       return -EALREADY;
+       }
+
        preempt_disable();
        mutex_acquire_nest(&lock->dep_map, subclass, 0, nest_lock, ip);