HID: uhid: Fix worker destroying device without any protection
authorJann Horn <jannh@google.com>
Fri, 14 Jan 2022 13:33:30 +0000 (14:33 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 27 Jan 2022 09:53:40 +0000 (10:53 +0100)
commitc2e39d5df028a3cbb61b2e1f6045c4f99a9cf70b
treefd50d7f40e3a1fedde167b76b87803d857cb26f3
parentaa1346113c752783f585d1d08627cfa38aa14e47
HID: uhid: Fix worker destroying device without any protection

commit 4ea5763fb79ed89b3bdad455ebf3f33416a81624 upstream.

uhid has to run hid_add_device() from workqueue context while allowing
parallel use of the userspace API (which is protected with ->devlock).
But hid_add_device() can fail. Currently, that is handled by immediately
destroying the associated HID device, without using ->devlock - but if
there are concurrent requests from userspace, that's wrong and leads to
NULL dereferences and/or memory corruption (via use-after-free).

Fix it by leaving the HID device as-is in the worker. We can clean it up
later, either in the UHID_DESTROY command handler or in the ->release()
handler.

Cc: stable@vger.kernel.org
Fixes: 67f8ecc550b5 ("HID: uhid: fix timeout when probe races with IO")
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/hid/uhid.c