When a softirq executes in irq_exit(), it can contribute to random complicated and large stack scenario involving task calls, hw interrupt calls, softirq handler calls and then other irqs, interrupting the softirq, that can dig further with an irq handler.
Softirqs executing on the inline hw interrupt stack may favour stack overflows in such circumstances, as it has been reported in powerpc where task -> irq -> softirq -> irq can end up forming a huge calltrace in the single kernel stack. So if there are softirqs pending on hardirq exit, lets execute them on the softirq stack to minimize this. Reported-by: Benjamin Herrenschmidt <[email protected]> Signed-off-by: Frederic Weisbecker <[email protected]> Tested-by: Paul Mackerras <[email protected]> Cc: Benjamin Herrenschmidt <[email protected]> Cc: Paul Mackerras <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Peter Zijlstra <[email protected]> Cc: H. Peter Anvin <[email protected]> Cc: Linus Torvalds <[email protected]> Cc: Paul Mackerras <[email protected]> Cc: James Hogan <[email protected]> Cc: James E.J. Bottomley <[email protected]> Cc: Helge Deller <[email protected]> Cc: Martin Schwidefsky <[email protected]> Cc: Heiko Carstens <[email protected]> Cc: David S. Miller <[email protected]> Cc: Andrew Morton <[email protected]> --- kernel/softirq.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/softirq.c b/kernel/softirq.c index 36112cf..8c8f08b 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -326,7 +326,7 @@ void irq_enter(void) static inline void invoke_softirq(void) { if (!force_irqthreads) - __do_softirq(); + do_softirq_own_stack(); else wakeup_softirqd(); } -- 1.8.3.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
