commit 1f8763c59c4ec6254d629fe77c0a52220bd907aa upstream.
John Keeping reported and posted a patch for a potential UAF in
rawmidi sequencer destruction: the snd_rawmidi_dev_seq_free() may be
called after the associated rawmidi object got already freed.
After a deeper look, it turned out that the bug is rather the
incorrect private_free call order for a snd_seq_device. The
snd_seq_device private_free gets called at the release callback of the
sequencer device object, while this was rather expected to be executed
at the snd_device call chains that runs at the beginning of the whole
card-free procedure. It's been broken since the rewrite of
sequencer-device binding (although it hasn't surfaced because the
sequencer device release happens usually right along with the card
device release).
This patch corrects the private_free call to be done in the right
place, at snd_seq_device_dev_free().
Fixes: 7c37ae5c625a ("ALSA: seq: Rewrite sequencer device binding with standard bus")
Reported-and-tested-by: John Keeping <john@metanate.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20210930114114.8645-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ALSA sequencer core has a mechanism to load the enumerated devices
automatically, and it's performed in an off-load work. This seems
causing some race when a sequencer is removed while the pending
autoload work is running. As syzkaller spotted, it may lead to some
use-after-free:
BUG: KASAN: use-after-free in snd_rawmidi_dev_seq_free+0x69/0x70
sound/core/rawmidi.c:1617
Write of size 8 at addr ffff88006c611d90 by task kworker/2:1/567
CPU: 2 PID: 567 Comm: kworker/2:1 Not tainted 4.13.0+ #29
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
Workqueue: events autoload_drivers
Call Trace:
__dump_stack lib/dump_stack.c:16 [inline]
dump_stack+0x192/0x22c lib/dump_stack.c:52
print_address_description+0x78/0x280 mm/kasan/report.c:252
kasan_report_error mm/kasan/report.c:351 [inline]
kasan_report+0x230/0x340 mm/kasan/report.c:409
__asan_report_store8_noabort+0x1c/0x20 mm/kasan/report.c:435
snd_rawmidi_dev_seq_free+0x69/0x70 sound/core/rawmidi.c:1617
snd_seq_dev_release+0x4f/0x70 sound/core/seq_device.c:192
device_release+0x13f/0x210 drivers/base/core.c:814
kobject_cleanup lib/kobject.c:648 [inline]
kobject_release lib/kobject.c:677 [inline]
kref_put include/linux/kref.h:70 [inline]
kobject_put+0x145/0x240 lib/kobject.c:694
put_device+0x25/0x30 drivers/base/core.c:1799
klist_devices_put+0x36/0x40 drivers/base/bus.c:827
klist_next+0x264/0x4a0 lib/klist.c:403
next_device drivers/base/bus.c:270 [inline]
bus_for_each_dev+0x17e/0x210 drivers/base/bus.c:312
autoload_drivers+0x3b/0x50 sound/core/seq_device.c:117
process_one_work+0x9fb/0x1570 kernel/workqueue.c:2097
worker_thread+0x1e4/0x1350 kernel/workqueue.c:2231
kthread+0x324/0x3f0 kernel/kthread.c:231
ret_from_fork+0x25/0x30 arch/x86/entry/entry_64.S:425
The fix is simply to assure canceling the autoload work at removing
the device.
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Tested-by: Andrey Konovalov <andreyknvl@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Many drivers bind the sequencer stuff in off-load by another driver
module, so that it's loaded only on demand. In the current code, this
mechanism doesn't work when the driver is built-in while the sequencer
is module. We check with IS_REACHABLE() and enable only when the
sequencer is in the same level of build.
However, this is basically a overshoot. The binder code
(snd-seq-device) is an individual module from the sequencer core
(snd-seq), and we just have to make the former a built-in while
keeping the latter a module for allowing the scenario like the above.
This patch achieves that by rewriting Kconfig slightly. Now, a driver
that provides the manual sequencer device binding should select
CONFIG_SND_SEQ_DEVICE in a way as
select SND_SEQ_DEVICE if SND_SEQUENCER != n
Note that the "!=n" is needed here to avoid the influence of the
sequencer core is module while the driver is built-in.
Also, since rawmidi.o may be linked with snd_seq_device.o when
built-in, we have to shuffle the code to make the linker happy.
(the kernel linker isn't smart enough yet to handle such a case.)
That is, snd_seq_device.c is moved to sound/core from sound/core/seq,
as well as Makefile.
Last but not least, the patch replaces the code using IS_REACHABLE()
with IS_ENABLED(), since now the condition meets always when enabled.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
