On 5/7/25 08:50, Dmitry Antipov wrote:
> In 'lookup_or_create_module_kobject()', an internal kobject is created
> using 'module_ktype'. So call to 'kobject_put()' on error handling
> path causes an attempt to use an uninitialized completion pointer in
> 'module_kobject_release()'. In this scenario, we just want to release
> kobject without an extra synchronization required for a regular module
> unloading process, so adding an extra check whether 'complete()' is
> actually required makes 'kobject_put()' safe.
>
> Reported-by: syzbot+7fb8a372e1f6add93...@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=7fb8a372e1f6add936dd
> Fixes: 942e443127e9 ("module: Fix mod->mkobj.kobj potentially freed too
> early")
> Suggested-by: Petr Pavlu <petr.pa...@suse.com>
> Signed-off-by: Dmitry Antipov <dmanti...@yandex.ru>
Added on modules-fixes.
I plan to send the fix to Linus on Friday. While the underlying problem
was already there, commit f95bbfe18512 ("drivers: base: handle
module_kobject creation"), merged in v6.15-rc5, made it more exposed and
I think it's better to have it fixed in this cycle. The patch itself is
also safe.
--
Thanks,
Petr
