[PATCH 2/2] debug_core: replace atomic_set() with atomic_xchg() for atomic kgdb_active

[Dongdong Deng]

The atomic_set() don't imply memory barriers, therefor the single
stepping processor could not observe the changing of kgdb_active
value which was done at other cpu immediately.

Using atomic_xchg() to replace atomic_set() to change kgdb_active,
as it implys the memory barrier semantics around the atomic_xchg().

Signed-off-by: Dongdong Deng <libfetion@xxxxxxxxx>
---
 kernel/debug/debug_core.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/debug/debug_core.c b/kernel/debug/debug_core.c
index bec8af0..722968f 100644
--- a/kernel/debug/debug_core.c
+++ b/kernel/debug/debug_core.c
@@ -548,7 +548,7 @@ return_normal:
 	if (atomic_read(&kgdb_cpu_doing_single_step) != -1 &&
 	    (kgdb_info[cpu].task &&
 	     kgdb_info[cpu].task->pid != kgdb_sstep_pid) && --sstep_tries) {
-		atomic_set(&kgdb_active, -1);
+		atomic_xchg(&kgdb_active, -1);
 		raw_spin_unlock(&dbg_master_lock);
 		dbg_touch_watchdogs();
 		local_irq_restore(flags);
@@ -657,7 +657,7 @@ kgdb_restore:
 	smp_mb__before_atomic_dec();
 	atomic_dec(&masters_in_kgdb);
 	/* Free kgdb_active */
-	atomic_set(&kgdb_active, -1);
+	atomic_xchg(&kgdb_active, -1);
 	raw_spin_unlock(&dbg_master_lock);
 	dbg_touch_watchdogs();
 	local_irq_restore(flags);
-- 
1.6.0.4

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/