locking/mutex: Further simplify mutex_spin_on_owner()
authorJason Low <jason.low2@hp.com>
Wed, 8 Apr 2015 19:39:19 +0000 (12:39 -0700)
committerIngo Molnar <mingo@kernel.org>
Thu, 9 Apr 2015 06:10:23 +0000 (08:10 +0200)
Similar to what Linus suggested for rwsem_spin_on_owner(), in
mutex_spin_on_owner() instead of having while (true) and
breaking out of the spin loop on lock->owner != owner, we can
have the loop directly check for while (lock->owner == owner) to
improve the readability of the code.

It also shrinks the code a bit:

   text    data     bss     dec     hex filename
   3721       0       0    3721     e89 mutex.o.before
   3705       0       0    3705     e79 mutex.o.after

Signed-off-by: Jason Low <jason.low2@hp.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aswin Chandramouleeswaran <aswin@hp.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
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: Tim Chen <tim.c.chen@linux.intel.com>
Link: http://lkml.kernel.org/r/1428521960-5268-2-git-send-email-jason.low2@hp.com
[ Added code generation info. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
kernel/locking/mutex.c

index 16b2d3cc88b0ca08c58f9ca2f123e3cc0477b0ae..4cccea6b8934f5697fa3dfa609ae4bd10db78b6f 100644 (file)
@@ -224,20 +224,14 @@ ww_mutex_set_context_slowpath(struct ww_mutex *lock,
 static noinline
 bool mutex_spin_on_owner(struct mutex *lock, struct task_struct *owner)
 {
-       bool ret;
+       bool ret = true;
 
        rcu_read_lock();
-       while (true) {
-               /* Return success when the lock owner changed */
-               if (lock->owner != owner) {
-                       ret = true;
-                       break;
-               }
-
+       while (lock->owner == owner) {
                /*
                 * Ensure we emit the owner->on_cpu, dereference _after_
-                * checking lock->owner still matches owner, if that fails,
-                * owner might point to free()d memory, if it still matches,
+                * checking lock->owner still matches owner. If that fails,
+                * owner might point to freed memory. If it still matches,
                 * the rcu_read_lock() ensures the memory stays valid.
                 */
                barrier();