]> www.infradead.org Git - users/jedix/linux-maple.git/commit
cgroup: avoid null de-ref in css_rstat_exit()
authorJP Kobryn <inwardvessel@gmail.com>
Thu, 7 Aug 2025 00:33:50 +0000 (17:33 -0700)
committerTejun Heo <tj@kernel.org>
Sat, 9 Aug 2025 18:46:32 +0000 (08:46 -1000)
commiteea51c6e3f6675b795f6439eaa960eb2948d6905
treef1a66524a0c8c10531b722c98acaae7f670d3922
parent87eba5bc5ab1d99e31c9d3b2c386187da94a5ab1
cgroup: avoid null de-ref in css_rstat_exit()

css_rstat_exit() may be called asynchronously in scenarios where preceding
calls to css_rstat_init() have not completed. One such example is this
sequence below:

css_create(...)
{
...
init_and_link_css(css, ...);

err = percpu_ref_init(...);
if (err)
goto err_free_css;
err = cgroup_idr_alloc(...);
if (err)
goto err_free_css;
err = css_rstat_init(css, ...);
if (err)
goto err_free_css;
...
err_free_css:
INIT_RCU_WORK(&css->destroy_rwork, css_free_rwork_fn);
queue_rcu_work(cgroup_destroy_wq, &css->destroy_rwork);
return ERR_PTR(err);
}

If any of the three goto jumps are taken, async cleanup will begin and
css_rstat_exit() will be invoked on an uninitialized css->rstat_cpu.

Avoid accessing the unitialized field by returning early in
css_rstat_exit() if this is the case.

Signed-off-by: JP Kobryn <inwardvessel@gmail.com>
Suggested-by: Michal Koutný <mkoutny@suse.com>
Fixes: 5da3bfa029d68 ("cgroup: use separate rstat trees for each subsystem")
Cc: stable@vger.kernel.org # v6.16
Reported-by: syzbot+8d052e8b99e40bc625ed@syzkaller.appspotmail.com
Acked-by: Shakeel Butt <shakeel.butt@linux.dev>
Signed-off-by: Tejun Heo <tj@kernel.org>
kernel/cgroup/rstat.c