On Tue 27-02-24 16:49:50, Michal Hocko wrote:

On Tue 27-02-24 14:32:20, Greg KH wrote:

quoted

On Tue, Feb 27, 2024 at 02:29:12PM +0100, Michal Hocko wrote:

quoted

Why is this applied to 5.4?
$ git describe-ver 1a3e1f40962c
v5.9-rc1~97^2~97

I do not see 1a3e1f40962c in 5.4 stable tree. What am I missing?

It is queued up for this next round of releases in the 5.4.y and 4.19.y
trees.

OK, now I remember the partial backport of 1a3e1f40962c
(http://lkml.kernel.org/r/20240222030237.82486-1-gongruiqi1@huawei.com)
but I need to have a look whether the follow up patch is really needed.

AFAICS f1796544a0ca ("memcg: fix use-after-free in uncharge_batch") is
only needed if the full 1a3e1f40962c is backported. The one staged for
5.4 shouldn't need a follow up as it only touches the pcp cache. I would
feel safer if other maintainers double check my thinking though.

Thanks
-- 
Michal Hocko
SUSE Labs