[GIT PULL] irq fix

[Ingo Molnar <mingo@kernel.org>]

Linus,

Please pull the latest irq-urgent-for-linus git tree from:

   git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git irq-urgent-for-linus

   HEAD: 0d119fb57614fe947aa9048f61d680bbede64170 Merge branch 'irq/urgent-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/linux-dynticks into irq/urgent

Frederic's minimal fix for hardirq/softirq nesting crashes.

 Thanks,

	Ingo

------------------>
Frederic Weisbecker (1):
      irq: Force hardirq exit's softirq processing on its own stack


 kernel/softirq.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/kernel/softirq.c b/kernel/softirq.c
index 53cc09c..d7d498d 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -328,10 +328,19 @@ void irq_enter(void)
 
 static inline void invoke_softirq(void)
 {
-	if (!force_irqthreads)
-		__do_softirq();
-	else
+	if (!force_irqthreads) {
+		/*
+		 * We can safely execute softirq on the current stack if
+		 * it is the irq stack, because it should be near empty
+		 * at this stage. But we have no way to know if the arch
+		 * calls irq_exit() on the irq stack. So call softirq
+		 * in its own stack to prevent from any overrun on top
+		 * of a potentially deep task stack.
+		 */
+		do_softirq();
+	} else {
 		wakeup_softirqd();
+	}
 }
 
 static inline void tick_irq_exit(void)