sound: GPF in snd_seq_fifo_clear

44 views
Skip to first unread message

Dmitry Vyukov

unread,
Jan 12, 2016, 5:07:36 AM1/12/16
to Jaroslav Kysela, Takashi Iwai, alsa-...@alsa-project.org, LKML, syzkaller, Kostya Serebryany, Alexander Potapenko, Sasha Levin, Eric Dumazet
Hello,

The following program triggers GPF in snd_seq_fifo_clear:

// autogenerated by syzkaller (http://github.com/google/syzkaller)
#include <unistd.h>
#include <sys/syscall.h>
#include <string.h>
#include <stdlib.h>
#include <stdint.h>
#include <pthread.h>

int fd;

void *thr(void *arg)
{
switch ((long)arg) {
case 0:
*(uint32_t*)0x20001fb0 = (uint32_t)0x1;
*(uint64_t*)0x20001fc8 = (uint64_t)0x0;
*(uint64_t*)0x20001fd0 = (uint64_t)0x0;
*(uint8_t*)0x20001fd8 = (uint8_t)0x3;
*(uint8_t*)0x20001fd9 = (uint8_t)0x32be;
*(uint8_t*)0x20001fda = (uint8_t)0x36;
*(uint8_t*)0x20001fdb = (uint8_t)0x5120;
*(uint32_t*)0x20001fdc = (uint32_t)0x0;
*(uint8_t*)0x20001fe0 = (uint8_t)0x4;
*(uint32_t*)0x20001fe4 = (uint32_t)0x0;
*(uint32_t*)0x20001fe8 = (uint32_t)0x0;
*(uint32_t*)0x20001fec = (uint32_t)0x0;
*(uint32_t*)0x20001ff0 = (uint32_t)0x0;
*(uint32_t*)0x20001ff4 = (uint32_t)0x0;
*(uint32_t*)0x20001ff8 = (uint32_t)0x0;
*(uint32_t*)0x20001ffc = (uint32_t)0x0;
*(uint32_t*)0x20002000 = (uint32_t)0x0;
*(uint32_t*)0x20002004 = (uint32_t)0x0;
*(uint32_t*)0x20002008 = (uint32_t)0x0;
syscall(SYS_ioctl, fd, 0x4040534eul, 0x20001fb0ul, 0, 0, 0);
break;
case 1:
*(uint32_t*)0x20006000 = (uint32_t)0xaff;
*(uint32_t*)0x20006004 = (uint32_t)0x5;
*(uint32_t*)0x20006008 = (uint32_t)0x101;
*(uint64_t*)0x20006010 = (uint64_t)0x0;
*(uint64_t*)0x20006018 = (uint64_t)0x989680;
*(uint32_t*)0x20006020 = (uint32_t)0x3;
*(uint32_t*)0x20006024 = (uint32_t)0x4;
*(uint8_t*)0x20006028 = (uint8_t)0x0;
*(uint8_t*)0x20006029 = (uint8_t)0x0;
*(uint8_t*)0x2000602a = (uint8_t)0x0;
*(uint8_t*)0x2000602b = (uint8_t)0x0;
*(uint8_t*)0x2000602c = (uint8_t)0x0;
*(uint8_t*)0x2000602d = (uint8_t)0x0;
*(uint8_t*)0x2000602e = (uint8_t)0x0;
*(uint8_t*)0x2000602f = (uint8_t)0x0;
*(uint8_t*)0x20006030 = (uint8_t)0x0;
*(uint8_t*)0x20006031 = (uint8_t)0x0;
*(uint8_t*)0x20006032 = (uint8_t)0x0;
*(uint8_t*)0x20006033 = (uint8_t)0x0;
*(uint8_t*)0x20006034 = (uint8_t)0x0;
*(uint8_t*)0x20006035 = (uint8_t)0x0;
*(uint8_t*)0x20006036 = (uint8_t)0x0;
*(uint8_t*)0x20006037 = (uint8_t)0x0;
*(uint8_t*)0x20006038 = (uint8_t)0x0;
*(uint8_t*)0x20006039 = (uint8_t)0x0;
*(uint8_t*)0x2000603a = (uint8_t)0x0;
*(uint8_t*)0x2000603b = (uint8_t)0x0;
*(uint8_t*)0x2000603c = (uint8_t)0x0;
*(uint8_t*)0x2000603d = (uint8_t)0x0;
*(uint8_t*)0x2000603e = (uint8_t)0x0;
*(uint8_t*)0x2000603f = (uint8_t)0x0;
*(uint8_t*)0x20006040 = (uint8_t)0x0;
*(uint8_t*)0x20006041 = (uint8_t)0x0;
*(uint8_t*)0x20006042 = (uint8_t)0x0;
*(uint8_t*)0x20006043 = (uint8_t)0x0;
*(uint8_t*)0x20006044 = (uint8_t)0x0;
*(uint8_t*)0x20006045 = (uint8_t)0x0;
*(uint8_t*)0x20006046 = (uint8_t)0x0;
*(uint8_t*)0x20006047 = (uint8_t)0x0;
*(uint8_t*)0x20006048 = (uint8_t)0x0;
*(uint8_t*)0x20006049 = (uint8_t)0x0;
*(uint8_t*)0x2000604a = (uint8_t)0x0;
*(uint8_t*)0x2000604b = (uint8_t)0x0;
*(uint8_t*)0x2000604c = (uint8_t)0x0;
*(uint8_t*)0x2000604d = (uint8_t)0x0;
*(uint8_t*)0x2000604e = (uint8_t)0x0;
*(uint8_t*)0x2000604f = (uint8_t)0x0;
*(uint8_t*)0x20006050 = (uint8_t)0x0;
*(uint8_t*)0x20006051 = (uint8_t)0x0;
*(uint8_t*)0x20006052 = (uint8_t)0x0;
*(uint8_t*)0x20006053 = (uint8_t)0x0;
*(uint8_t*)0x20006054 = (uint8_t)0x0;
*(uint8_t*)0x20006055 = (uint8_t)0x0;
*(uint8_t*)0x20006056 = (uint8_t)0x0;
*(uint8_t*)0x20006057 = (uint8_t)0x0;
*(uint8_t*)0x20006058 = (uint8_t)0x0;
*(uint8_t*)0x20006059 = (uint8_t)0x0;
*(uint8_t*)0x2000605a = (uint8_t)0x0;
*(uint8_t*)0x2000605b = (uint8_t)0x0;
*(uint8_t*)0x2000605c = (uint8_t)0x0;
*(uint8_t*)0x2000605d = (uint8_t)0x0;
*(uint8_t*)0x2000605e = (uint8_t)0x0;
*(uint8_t*)0x2000605f = (uint8_t)0x0;
*(uint8_t*)0x20006060 = (uint8_t)0x0;
*(uint8_t*)0x20006061 = (uint8_t)0x0;
*(uint8_t*)0x20006062 = (uint8_t)0x0;
*(uint8_t*)0x20006063 = (uint8_t)0x0;
*(uint8_t*)0x20006064 = (uint8_t)0x0;
*(uint8_t*)0x20006065 = (uint8_t)0x0;
*(uint8_t*)0x20006066 = (uint8_t)0x0;
*(uint8_t*)0x20006067 = (uint8_t)0x0;
syscall(SYS_ioctl, fd, 0x402c5342ul, 0x20006000ul, 0, 0, 0);
break;
case 2:
*(uint8_t*)0x20007fb0 = (uint8_t)0x1037;
*(uint8_t*)0x20007fb1 = (uint8_t)0x7;
*(uint8_t*)0x20007fb2 = (uint8_t)0x30b;
*(uint8_t*)0x20007fb3 = (uint8_t)0x34b0;
*(uint32_t*)0x20007fb4 = (uint32_t)0x5;
*(uint32_t*)0x20007fb8 = (uint32_t)0x7;
*(uint8_t*)0x20007fbc = (uint8_t)0x75d;
*(uint8_t*)0x20007fbd = (uint8_t)0x0;
*(uint8_t*)0x20007fbe = (uint8_t)0x0;
*(uint8_t*)0x20007fbf = (uint8_t)0x0;
*(uint8_t*)0x20007fc0 = (uint8_t)0x0;
*(uint8_t*)0x20007fc1 = (uint8_t)0x0;
*(uint8_t*)0x20007fc2 = (uint8_t)0x0;
*(uint8_t*)0x20007fc3 = (uint8_t)0x0;
*(uint8_t*)0x20007fc4 = (uint8_t)0x0;
*(uint8_t*)0x20007fc5 = (uint8_t)0x0;
*(uint8_t*)0x20007fc6 = (uint8_t)0x0;
*(uint8_t*)0x20007fc7 = (uint8_t)0x0;
*(uint8_t*)0x20007fc8 = (uint8_t)0x0;
*(uint8_t*)0x20007fc9 = (uint8_t)0x0;
*(uint8_t*)0x20007fca = (uint8_t)0x0;
*(uint8_t*)0x20007fcb = (uint8_t)0x0;
*(uint8_t*)0x20007fcc = (uint8_t)0x0;
*(uint8_t*)0x20007fcd = (uint8_t)0x0;
*(uint8_t*)0x20007fce = (uint8_t)0x0;
*(uint8_t*)0x20007fcf = (uint8_t)0x0;
*(uint8_t*)0x20007fd0 = (uint8_t)0x0;
*(uint8_t*)0x20007fd1 = (uint8_t)0x0;
*(uint8_t*)0x20007fd2 = (uint8_t)0x0;
*(uint8_t*)0x20007fd3 = (uint8_t)0x0;
*(uint8_t*)0x20007fd4 = (uint8_t)0x0;
*(uint8_t*)0x20007fd5 = (uint8_t)0x0;
*(uint8_t*)0x20007fd6 = (uint8_t)0x0;
*(uint8_t*)0x20007fd7 = (uint8_t)0x0;
*(uint8_t*)0x20007fd8 = (uint8_t)0x0;
*(uint8_t*)0x20007fd9 = (uint8_t)0x0;
*(uint8_t*)0x20007fda = (uint8_t)0x0;
*(uint8_t*)0x20007fdb = (uint8_t)0x0;
*(uint8_t*)0x20007fdc = (uint8_t)0x0;
*(uint8_t*)0x20007fdd = (uint8_t)0x0;
*(uint8_t*)0x20007fde = (uint8_t)0x0;
*(uint8_t*)0x20007fdf = (uint8_t)0x0;
*(uint8_t*)0x20007fe0 = (uint8_t)0x0;
*(uint8_t*)0x20007fe1 = (uint8_t)0x0;
*(uint8_t*)0x20007fe2 = (uint8_t)0x0;
*(uint8_t*)0x20007fe3 = (uint8_t)0x0;
*(uint8_t*)0x20007fe4 = (uint8_t)0x0;
*(uint8_t*)0x20007fe5 = (uint8_t)0x0;
*(uint8_t*)0x20007fe6 = (uint8_t)0x0;
*(uint8_t*)0x20007fe7 = (uint8_t)0x0;
*(uint8_t*)0x20007fe8 = (uint8_t)0x0;
*(uint8_t*)0x20007fe9 = (uint8_t)0x0;
*(uint8_t*)0x20007fea = (uint8_t)0x0;
*(uint8_t*)0x20007feb = (uint8_t)0x0;
*(uint8_t*)0x20007fec = (uint8_t)0x0;
*(uint8_t*)0x20007fed = (uint8_t)0x0;
*(uint8_t*)0x20007fee = (uint8_t)0x0;
*(uint8_t*)0x20007fef = (uint8_t)0x0;
*(uint8_t*)0x20007ff0 = (uint8_t)0x0;
*(uint8_t*)0x20007ff1 = (uint8_t)0x0;
*(uint8_t*)0x20007ff2 = (uint8_t)0x0;
*(uint8_t*)0x20007ff3 = (uint8_t)0x0;
*(uint8_t*)0x20007ff4 = (uint8_t)0x0;
*(uint8_t*)0x20007ff5 = (uint8_t)0x0;
*(uint8_t*)0x20007ff6 = (uint8_t)0x0;
*(uint8_t*)0x20007ff7 = (uint8_t)0x0;
*(uint8_t*)0x20007ff8 = (uint8_t)0x0;
*(uint8_t*)0x20007ff9 = (uint8_t)0x0;
*(uint8_t*)0x20007ffa = (uint8_t)0x0;
*(uint8_t*)0x20007ffb = (uint8_t)0x0;
*(uint8_t*)0x20007ffc = (uint8_t)0x0;
*(uint8_t*)0x20007ffd = (uint8_t)0x0;
*(uint8_t*)0x20007ffe = (uint8_t)0x0;
*(uint8_t*)0x20007fff = (uint8_t)0x0;
syscall(SYS_ioctl, fd, 0x40505331ul, 0x20007fb0ul, 0, 0, 0);
break;
}
return 0;
}

int main()
{
long i;
pthread_t th;

srand(getpid());
syscall(SYS_mmap, 0x20000000ul, 0x8000ul, 0x3ul, 0x32ul,
0xfffffffffffffffful, 0x0ul);
memcpy((void*)0x20005000,
"\x2f\x64\x65\x76\x2f\x73\x6e\x64\x2f\x73\x65\x71", 12);
fd = syscall(SYS_open, 0x20005000ul, 0x1ul, 0x0ul, 0, 0, 0);
for (i = 0; i < 6; i++) {
pthread_create(&th, 0, thr, (void*)(i%3));
if (rand()%2==0)
usleep(rand()%1000);
}
usleep(10000);
return 0;
}



kasan: CONFIG_KASAN_INLINE enabled[ 146.589109] kasan:
CONFIG_KASAN_INLINE enabledkasan: GPF could be caused by NULL-ptr
deref or user memory accessgeneral protection fault: 0000 [#1] SMP
DEBUG_PAGEALLOC KASAN
Modules linked in:
CPU: 2 PID: 6540 Comm: a.out Not tainted 4.4.0+ #222
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
task: ffff880064ed0000 ti: ffff880063620000 task.ti: ffff880063620000
RIP: 0010:[<ffffffff84b65301>] [<ffffffff84b65301>]
snd_seq_fifo_clear+0x31/0x1d0
RSP: 0018:ffff880063627c48 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: 0000000000000000 RCX: 0000000000000000
RDX: 0000000000000015 RSI: 0000000020001ff0 RDI: 00000000000000a8
RBP: ffff880063627c98 R08: 0000000000000001 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000001 R12: ffff880064de1ec0
R13: dffffc0000000000 R14: ffff880063627d28 R15: 0000000000000001
FS: 00007fc371e37700(0000) GS:ffff88006d600000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000020002000 CR3: 0000000062a62000 CR4: 00000000000006e0
Stack:
ffffffff867a5ce0 00000000000002b9 000000004040534e 0000000020001fb0
ffff880063627c98 1ffff1000c6c4f95 ffff880064de1ec0 dffffc0000000000
ffff880063627d28 0000000000000001 ffff880063627d50 ffffffff84b56918
Call Trace:
[<ffffffff84b56918>] snd_seq_ioctl_remove_events+0x178/0x1b0
sound/core/seq/seq_clientmgr.c:1966
[<ffffffff84b5954a>] snd_seq_do_ioctl+0x19a/0x1c0
sound/core/seq/seq_clientmgr.c:2209
[<ffffffff84b5973d>] snd_seq_ioctl+0x5d/0x80
sound/core/seq/seq_clientmgr.c:2224
[< inline >] vfs_ioctl fs/ioctl.c:43
[<ffffffff817b3531>] do_vfs_ioctl+0x681/0xe40 fs/ioctl.c:607
[< inline >] SYSC_ioctl fs/ioctl.c:622
[<ffffffff817b3d7f>] SyS_ioctl+0x8f/0xc0 fs/ioctl.c:613
[<ffffffff85e748f6>] entry_SYSCALL_64_fastpath+0x16/0x7a
arch/x86/entry/entry_64.S:185
Code: 41 56 41 55 41 54 53 48 89 fb 48 83 ec 28 e8 47 aa 9f fc 48 8d
bb a8 00 00 00 48 b8 00 00 00 00 00 fc ff df 48 89 fa 48 c1 ea 03 <0f>
b6 14 02 48 89 f8 83 e0 07 83 c0 03 38 d0 7c 08 84 d2 0f 85
RIP [< inline >] __write_once_size include/linux/compiler.h:246
RIP [< inline >] atomic_set ./arch/x86/include/asm/atomic.h:39
RIP [<ffffffff84b65301>] snd_seq_fifo_clear+0x31/0x1d0
sound/core/seq/seq_fifo.c:99
RSP <ffff880063627c48>
---[ end trace 3ee37e6a5304c762 ]---


On commit afd2ff9b7e1b367172f18ba7f693dfb62bdcb2dc (Jan 10).

Takashi Iwai

unread,
Jan 12, 2016, 6:44:29 AM1/12/16
to Dmitry Vyukov, alsa-...@alsa-project.org, Jaroslav Kysela, LKML, Eric Dumazet, Alexander Potapenko, Kostya Serebryany, syzkaller, Sasha Levin
Thanks for reporting.

Fortunately this one looks like an easy problem, a simple missing NULL
check. Could you check the patch below?


Takashi

-- 8< --
From: Takashi Iwai <ti...@suse.de>
Subject: [PATCH] ALSA: seq: Fix missing NULL check at remove_events ioctl

snd_seq_ioctl_remove_events() calls snd_seq_fifo_clear()
unconditionally even if there is no FIFO assigned, and this leads to
an Oops due to NULL dereference. The fix is just to add a proper NULL
check.

Reported-by: Dmitry Vyukov <dvy...@google.com>
Cc: <sta...@vger.kernel.org>
Signed-off-by: Takashi Iwai <ti...@suse.de>
---
sound/core/seq/seq_clientmgr.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sound/core/seq/seq_clientmgr.c b/sound/core/seq/seq_clientmgr.c
index b64f20deba90..13cfa815732d 100644
--- a/sound/core/seq/seq_clientmgr.c
+++ b/sound/core/seq/seq_clientmgr.c
@@ -1962,7 +1962,7 @@ static int snd_seq_ioctl_remove_events(struct snd_seq_client *client,
* No restrictions so for a user client we can clear
* the whole fifo
*/
- if (client->type == USER_CLIENT)
+ if (client->type == USER_CLIENT && client->data.user.fifo)
snd_seq_fifo_clear(client->data.user.fifo);
}

--
2.7.0

Dmitry Vyukov

unread,
Jan 12, 2016, 11:20:09 AM1/12/16
to Takashi Iwai, alsa-...@alsa-project.org, Jaroslav Kysela, LKML, Eric Dumazet, Alexander Potapenko, Kostya Serebryany, syzkaller, Sasha Levin
Yes, it fixes the crashes for me. Thanks for quick fix!

Tested-by: Dmitry Vyukov <dvy...@google.com>

Takashi Iwai

unread,
Jan 12, 2016, 11:23:23 AM1/12/16
to Dmitry Vyukov, alsa-...@alsa-project.org, Jaroslav Kysela, LKML, Eric Dumazet, Alexander Potapenko, Kostya Serebryany, syzkaller, Sasha Levin
Thanks for quick testing. I queued the patch now.


Takashi

miles....@gmail.com

unread,
Oct 16, 2017, 4:36:42 AM10/16/17
to syzkaller
Hi,

I encountered the same issue in Android kernel-3.18.

AOSP kernel-4.4 is patched.

But I'm not sure why AOSP kernel-3.18 is not patched?

Thanks,

Miles

Dmitry Vyukov

unread,
Oct 16, 2017, 4:38:20 AM10/16/17
to Miles Fuzzer, syzkaller, Rom Lemarchand
On Mon, Oct 16, 2017 at 10:36 AM, <miles....@gmail.com> wrote:
> Hi,
>
> I encountered the same issue in Android kernel-3.18.
>
> AOSP kernel-4.4 is patched.
> https://android.googlesource.com/kernel/common/+blame/android-4.4/sound/core/seq/seq_clientmgr.c#1964
>
> But I'm not sure why AOSP kernel-3.18 is not patched?
> https://android.googlesource.com/kernel/common/+blame/android-3.18/sound/core/seq/seq_clientmgr.c#1969


+Rom for AOSP 3.18.
> --
> You received this message because you are subscribed to the Google Groups
> "syzkaller" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to syzkaller+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

Dmitry Vyukov

unread,
Oct 16, 2017, 4:54:30 AM10/16/17
to Miles Fuzzer, syzkaller, Rom Lemarchand, Dmitry Shmidt, Mark Salyzyn
On Mon, Oct 16, 2017 at 10:37 AM, Dmitry Vyukov <dvy...@google.com> wrote:
> On Mon, Oct 16, 2017 at 10:36 AM, <miles....@gmail.com> wrote:
>> Hi,
>>
>> I encountered the same issue in Android kernel-3.18.
>>
>> AOSP kernel-4.4 is patched.
>> https://android.googlesource.com/kernel/common/+blame/android-4.4/sound/core/seq/seq_clientmgr.c#1964
>>
>> But I'm not sure why AOSP kernel-3.18 is not patched?
>> https://android.googlesource.com/kernel/common/+blame/android-3.18/sound/core/seq/seq_clientmgr.c#1969
>
>
> +Rom for AOSP 3.18.

Also +Mark and +Dmitry

Mark Salyzyn

unread,
Oct 16, 2017, 10:33:53 AM10/16/17
to Dmitry Vyukov, Miles Fuzzer, syzkaller, Rom Lemarchand, Dmitry Shmidt, Greg Kroah-Hartman
This fix came through -stable.

 3.18.y does not have this fix. 3.10.y, by example, does. GregKH can you get Takashi Iwai's patch 030e2c78d3a91dd0d27fef37e91950dde333eba1 into 3.18 stable?

-- Mark

Greg KH

unread,
Oct 16, 2017, 11:46:24 AM10/16/17
to Mark Salyzyn, Dmitry Vyukov, Miles Fuzzer, syzkaller, Rom Lemarchand, Dmitry Shmidt
On Mon, Oct 16, 2017 at 07:33:51AM -0700, Mark Salyzyn wrote:
> This fix came through -stable.
>
> 3.18.y does not have this fix. 3.10.y, by example, does. GregKH can you
> get Takashi Iwai's patch 030e2c78d3a91dd0d27fef37e91950dde333eba1 into 3.18
> stable?

Now applied. In the future, please send any stable kernel patch
requests to sta...@vger.kernel.org so that all stable tree maintainers
can see this, in case their trees do not contain the missing patch.

thanks,

greg k-h

Mark Salyzyn

unread,
Oct 16, 2017, 12:39:34 PM10/16/17
to Greg KH, Dmitry Vyukov, Miles Fuzzer, syzkaller, Rom Lemarchand, Dmitry Shmidt
Thanks!

Was not sure about history of who (Takashi Iwai?) asked for it to be back-ported to the stable trees. And yes, assumed the process was followed and 3.18.y was skipped, because GregKH is the (silent? labour of love?) 3.18 stable maintainer. I admittedly did not think that there could be others (beyond Android focused set) that could have missed it, so next time a similar situation is hits, sta...@vger.kernel.org will be contacted.

Begs the question, I do not have 3.2.y or 3.16.y pulled to check, they may have been missed too ...

-- Mark
Reply all
Reply to author
Forward
0 new messages