On Fri, Dec 22, 2023 at 09:31:18AM -0800, Luis Chamberlain wrote:

On Fri, Dec 22, 2023 at 06:17:47PM +0800, Coly Li wrote:

quoted

quoted

2023年12月22日 03:17，Luis Chamberlain [off-list ref] 写道：

On Thu, Dec 21, 2023 at 02:19:56PM +0800, Yu Kuai wrote:

quoted

I can't find this by code review, and I think
maybe it's better to fix this in sysctl error path.

Indeed, we want to fix anything in the way to remove the empty sentinel,
we continue to do that in queued work on sysctl-next [0]. Although I
won't be able to diagnose this right away, could you try the out of
bounds fix by Joel [1] instead?

We want to identify what caused this and fix it within sysctl code.

[0] https://git.kernel.org/pub/scm/linux/kernel/git/mcgrof/linux.git/log/?h=sysctl-next
[1] https://git.kernel.org/pub/scm/linux/kernel/git/mcgrof/linux.git/commit/?h=sysctl-next&id=fd696ee2395755a292f7d49bf4c701a5bab2f076

Hi Luis,

Thanks of the above information. IMHO your code is good, When I cherry
pick the upstream md code for testing, the sysctl related change
leaked from my eyes. please ignore my noise. 

So this was triggered because the tree was missing the changes that
actually handled the removal of the sentinel?

Get back to me if the oops persists even after you have included the
changes in sysctl-next

Best

Great thanks for the heads up! Happy holidays.

  Luis

-- 

Joel Granados