sched/debug: Check for stack overflow in ___might_sleep()

Sometimes a "BUG: sleeping function called from invalid context"
message is not indicative of locking problems, but is the result
of a stack overflow corrupting the thread info.

Witness http://oss.sgi.com/archives/xfs/2014-02/msg00325.html
for example, which took a few go-rounds to sort out.

If we're printing the warning, things are wonky already, and
it'd be informative to check for the stack end corruption at this
point, too.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/5490B158.4060005@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index c0accc00566eb774a022870635c60e63da6ee198..56c9b79772bd4b030f2908d4b67d9d05a57312e3 100644 (file)
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7325,6 +7325,9 @@ void ___might_sleep(const char *file, int line, int preempt_offset)
                        in_atomic(), irqs_disabled(),
                        current->pid, current->comm);
 
+       if (task_stack_end_corrupted(current))
+               printk(KERN_EMERG "Thread overran stack, or stack corrupted\n");
+
        debug_show_held_locks(current);
        if (irqs_disabled())
                print_irqtrace_events(current);