Re: list corruption in the last few days. (block ? crypto ?)

From: Pekka Enberg
Date: Sun Aug 07 2011 - 16:37:06 EST


On Fri, 5 Aug 2011, Dave Jones wrote:
>>  > Argh another debugging code issue. Anyone who sees this please verify that
>>  > it also occurs with debugging off. I ran this code for 9 months in
>>  > various configs but only sporadically with debugging support since that
>>  > switches the lockless stuff.
>>
>> This started turning up immediately as soon as it got merged to mainline.
>> Which leads me to believe no-one is running -next with debugging on either.
>> Perhaps we should force debugging on for a week once per release cycle.

On Fri, Aug 5, 2011 at 9:14 PM, Christoph Lameter <cl@xxxxxxxxx> wrote:
> I have run some tests over the last hours with kvm and
> CONFIG_SLUB_DEBUG_ON as well as CONFIG_DEBUG_LIST set but (aside from
> other unrelated backtraces in other subsystems) did not see this issue.
>
> Can I get a .config? Maybe that will help.

I haven't been able to reproduce the problem on hardware either.
However, I don't have CONFIG_PREEMPT enabled like Dave and our RCU
options seem to be different:

$ diff -u adamo.config ../linux/.config | grep RCU
# RCU Subsystem
-CONFIG_TREE_PREEMPT_RCU=y
-CONFIG_PREEMPT_RCU=y
+CONFIG_TREE_RCU=y
+# CONFIG_PREEMPT_RCU is not set
# CONFIG_RCU_TRACE is not set
CONFIG_RCU_FANOUT=64
# CONFIG_RCU_FANOUT_EXACT is not set
+# CONFIG_RCU_FAST_NO_HZ is not set
# CONFIG_TREE_RCU_TRACE is not set
-CONFIG_RCU_BOOST=y
-CONFIG_RCU_BOOST_PRIO=1
-CONFIG_RCU_BOOST_DELAY=500
-CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
-CONFIG_PROVE_RCU=y
-# CONFIG_PROVE_RCU_REPEATEDLY is not set
-CONFIG_SPARSE_RCU_POINTER=y
+# CONFIG_SPARSE_RCU_POINTER is not set
# CONFIG_RCU_TORTURE_TEST is not set
CONFIG_RCU_CPU_STALL_TIMEOUT=60
-CONFIG_RCU_CPU_STALL_VERBOSE=y

Pekka
--
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/