Re: combinatorial explosion in lockdep

[Hugh Dickins]

On Fri, 1 Aug 2008, David Miller wrote:
> 
> I think a non-trivial number of people hit that printk deadlock bug,
> but just didn't report it because the machine essentially hard hangs
> silently.  At best you'd see the:
> 
> ========================================
> 
> initial line from lockdep, but often even that doesn't make it to the
> console.

Not just with lockdep.  A couple of months back CONFIG_DEBUG_SPINLOCK
was giving me spinlock debug messages under load on x86_64 duo laptop,
but hung up before showing the intended backtrace.  I added the patch
below, but IIRC shifted over to other issues (and switched from cfq to
anticipatory scheduler, which made a difference) before finding out if
this patch really helped.  If it makes sense to you, please add it in.

Hugh

[PATCH] set oops_in_progress for spinlock lockup

Set oops_in_progress while reporting spinlock lockup,
to avoid deadlock inside printk() or the backtraces.

Signed-off-by: Hugh Dickins <hugh@xxxxxxxxxxx>
---

 lib/spinlock_debug.c |    2 ++
 1 file changed, 2 insertions(+)

--- 2.6.27-rc1/lib/spinlock_debug.c	2008-01-24 22:58:37.000000000 +0000
+++ linux/lib/spinlock_debug.c	2008-08-01 12:41:52.000000000 +0100
@@ -113,6 +113,7 @@ static void __spin_lock_debug(spinlock_t
 		/* lockup suspected: */
 		if (print_once) {
 			print_once = 0;
+			oops_in_progress = 1;
 			printk(KERN_EMERG "BUG: spinlock lockup on CPU#%d, "
 					"%s/%d, %p\n",
 				raw_smp_processor_id(), current->comm,
@@ -121,6 +122,7 @@ static void __spin_lock_debug(spinlock_t
 #ifdef CONFIG_SMP
 			trigger_all_cpu_backtrace();
 #endif
+			oops_in_progress = 0;
 		}
 	}
 }
--
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/