[PATCH 1/2] debug_core: add a memory barrier before atomic_inc(&slaves_in_kgdb)

[Dongdong Deng]

The atomic_inc operation of slaves_in_kgdb could be executed on
multi-cpus at the same time when kgdb have more than one slave
cpu.

Due to the atomic_inc() function do not imply memory barrier,
a missing memory barrier here maybe could casue an incorrect
value of slaves_in_kgdb, then let kgdb confused.

Add a memory barrier before atomic_inc(&slaves_in_kgdb) to
avoid that potential problem.

Signed-off-by: Dongdong Deng <libfetion@xxxxxxxxx>
---
 kernel/debug/debug_core.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/kernel/debug/debug_core.c b/kernel/debug/debug_core.c
index cefd4a1..bec8af0 100644
--- a/kernel/debug/debug_core.c
+++ b/kernel/debug/debug_core.c
@@ -468,10 +468,13 @@ static int kgdb_cpu_enter(struct kgdb_state *ks, struct pt_regs *regs,
 	kgdb_info[ks->cpu].enter_kgdb++;
 	kgdb_info[ks->cpu].exception_state |= exception_state;
 
-	if (exception_state == DCPU_WANT_MASTER)
+	if (exception_state == DCPU_WANT_MASTER) {
+		smp_mb__before_atomic_inc();
 		atomic_inc(&masters_in_kgdb);
-	else
+	} else {
+		smp_mb__before_atomic_inc();
 		atomic_inc(&slaves_in_kgdb);
+	}
 
 	if (arch_kgdb_ops.disable_hw_break)
 		arch_kgdb_ops.disable_hw_break(regs);
-- 
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/