LWN.net Logo

[PATCH] spinlock: __raw_spin_is_locked() should return true for UP

From:  Kumar Gala <galak-AT-kernel.crashing.org>
To:  Linus Torvalds <torvalds-AT-osdl.org>
Subject:  [PATCH] spinlock: __raw_spin_is_locked() should return true for UP
Date:  Tue, 18 Aug 2009 17:42:23 -0500
Cc:  peterz-AT-infradead.org, linux-kernel-AT-vger.kernel.org, rostedt-AT-goodmis.org, linuxppc-dev-AT-ozlabs.org, mingo-AT-elte.hu, tglx-AT-linutronix.de
Archive-link:  Article, Thread

For some reason __raw_spin_is_locked() has been returning false for the
uni-processor, non-CONFIG_DEBUG_SPINLOCK.  The UP + CONFIG_DEBUG_SPINLOCK
handles this correctly.

Found this by enabling CONFIG_DEBUG_VM on PPC and hitting always hitting
a BUG_ON that was testing to make sure the pte_lock was held.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
---

Linus, a fix for 2.6.31

 include/linux/spinlock_up.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/include/linux/spinlock_up.h b/include/linux/spinlock_up.h
index d4841ed..2b372e0 100644
--- a/include/linux/spinlock_up.h
+++ b/include/linux/spinlock_up.h
@@ -57,7 +57,7 @@ static inline void __raw_spin_unlock(raw_spinlock_t *lock)
 #define __raw_write_unlock(lock)	do { (void)(lock); } while (0)
 
 #else /* DEBUG_SPINLOCK */
-#define __raw_spin_is_locked(lock)	((void)(lock), 0)
+#define __raw_spin_is_locked(lock)	((void)(lock), 1)
 /* for sched.c and kernel_lock.c: */
 # define __raw_spin_lock(lock)		do { (void)(lock); } while (0)
 # define __raw_spin_lock_flags(lock, flags)	do { (void)(lock); } while (0)
-- 
1.6.0.6


(Log in to post comments)

Copyright © 2009, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds