lkml.org 
[lkml]   [2017]   [Nov]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: usb/sound/usx2y: WARNING in usb_stream_start
On Mon, Nov 6, 2017 at 10:56 AM, Takashi Iwai <tiwai@suse.de> wrote:
> On Fri, 03 Nov 2017 20:52:49 +0100,
> Takashi Iwai wrote:
>>
>> On Fri, 03 Nov 2017 15:44:59 +0100,
>> Andrey Konovalov wrote:
>> >
>> > Hi!
>> >
>> > I've got the following report while fuzzing the kernel with syzkaller.
>> >
>> > On commit 3a99df9a3d14cd866b5516f8cba515a3bfd554ab (4.14-rc7+).
>> >
>> > Looks like there's no check for the actual endpoint types.
>> >
>> > usb 1-1: BOGUS urb xfer, pipe 0 != type 3
>> > ------------[ cut here ]------------
>> > WARNING: CPU: 0 PID: 24 at drivers/usb/core/urb.c:471
>> > usb_submit_urb+0x113e/0x1400
>> > Modules linked in:
>> > CPU: 0 PID: 24 Comm: kworker/0:1 Not tainted
>> > 4.14.0-rc7-44290-gf28444df2601-dirty #52
>> > Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>> > Workqueue: usb_hub_wq hub_event
>> > task: ffff88006bef5c00 task.stack: ffff88006bf60000
>> > RIP: 0010:usb_submit_urb+0x113e/0x1400 drivers/usb/core/urb.c:470
>> > RSP: 0018:ffff88006bf67440 EFLAGS: 00010286
>> > RAX: 0000000000000029 RBX: ffff880064698c80 RCX: ffffffff812495b5
>> > RDX: 0000000000000000 RSI: ffffffff8124d76a RDI: 0000000000000005
>> > RBP: ffff88006bf674b0 R08: ffff88006bef5c00 R09: 0000000000000006
>> > R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
>> > R13: 0000000000000003 R14: ffff880069a25a20 R15: ffff880064698d04
>> > FS: 0000000000000000(0000) GS:ffff88006ca00000(0000) knlGS:0000000000000000
>> > CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> > CR2: 00007fe6fdba3000 CR3: 0000000068470000 CR4: 00000000000006f0
>> > Call Trace:
>> > usb_stream_start+0x48a/0x9f0 sound/usb/usx2y/usb_stream.c:690
>> > us122l_start+0x116/0x290 sound/usb/usx2y/us122l.c:365
>> (snip)
>>
>> OK, so this looks like another typical issue.
>> A totally untested patch is below. Could you check whether it covers?
>
> FYI, below is the patch with a proper description that I queued to
> for-next branch.

This patch fixes the issue.

Thanks!

Tested-by: Andrey Konovalov <andreyknvl@google.com>

>
>
> Takashi
>
> -- 8< --
> From: Takashi Iwai <tiwai@suse.de>
> Subject: [PATCH] ALSA: usx2y: Fix invalid stream URBs
>
> The us122l driver creates URBs per the fixed endpoints, and this may
> end up with URBs with inconsistent pipes when a fuzzer or a malicious
> program deals with the manipulated endpoints. It ends up with a
> kernel warning like:
>
> usb 1-1: BOGUS urb xfer, pipe 0 != type 3
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 24 at drivers/usb/core/urb.c:471
> usb_submit_urb+0x113e/0x1400
> Call Trace:
> usb_stream_start+0x48a/0x9f0 sound/usb/usx2y/usb_stream.c:690
> us122l_start+0x116/0x290 sound/usb/usx2y/us122l.c:365
> us122l_create_card sound/usb/usx2y/us122l.c:502
> us122l_usb_probe sound/usb/usx2y/us122l.c:588
> ....
>
> For avoiding the bad access, this patch adds a few sanity checks of
> the validity of created URBs like previous similar fixes using the new
> usb_urb_ep_type_check() helper function.
>
> Reported-by: Andrey Konovalov <andreyknvl@google.com>
> Signed-off-by: Takashi Iwai <tiwai@suse.de>
> ---
> sound/usb/usx2y/usb_stream.c | 23 +++++++++++++++++------
> 1 file changed, 17 insertions(+), 6 deletions(-)
>
> diff --git a/sound/usb/usx2y/usb_stream.c b/sound/usb/usx2y/usb_stream.c
> index e229abd21652..b0f8979ff2d2 100644
> --- a/sound/usb/usx2y/usb_stream.c
> +++ b/sound/usb/usx2y/usb_stream.c
> @@ -56,7 +56,7 @@ static void playback_prep_freqn(struct usb_stream_kernel *sk, struct urb *urb)
> lb, s->period_size);
> }
>
> -static void init_pipe_urbs(struct usb_stream_kernel *sk, unsigned use_packsize,
> +static int init_pipe_urbs(struct usb_stream_kernel *sk, unsigned use_packsize,
> struct urb **urbs, char *transfer,
> struct usb_device *dev, int pipe)
> {
> @@ -77,6 +77,8 @@ static void init_pipe_urbs(struct usb_stream_kernel *sk, unsigned use_packsize,
> urb->interval = 1;
> if (usb_pipeout(pipe))
> continue;
> + if (usb_urb_ep_type_check(urb))
> + return -EINVAL;
>
> urb->transfer_buffer_length = transfer_length;
> desc = urb->iso_frame_desc;
> @@ -87,9 +89,11 @@ static void init_pipe_urbs(struct usb_stream_kernel *sk, unsigned use_packsize,
> desc[p].length = maxpacket;
> }
> }
> +
> + return 0;
> }
>
> -static void init_urbs(struct usb_stream_kernel *sk, unsigned use_packsize,
> +static int init_urbs(struct usb_stream_kernel *sk, unsigned use_packsize,
> struct usb_device *dev, int in_pipe, int out_pipe)
> {
> struct usb_stream *s = sk->s;
> @@ -103,9 +107,12 @@ static void init_urbs(struct usb_stream_kernel *sk, unsigned use_packsize,
> sk->outurb[u] = usb_alloc_urb(sk->n_o_ps, GFP_KERNEL);
> }
>
> - init_pipe_urbs(sk, use_packsize, sk->inurb, indata, dev, in_pipe);
> - init_pipe_urbs(sk, use_packsize, sk->outurb, sk->write_page, dev,
> - out_pipe);
> + if (init_pipe_urbs(sk, use_packsize, sk->inurb, indata, dev, in_pipe) ||
> + init_pipe_urbs(sk, use_packsize, sk->outurb, sk->write_page, dev,
> + out_pipe))
> + return -EINVAL;
> +
> + return 0;
> }
>
>
> @@ -226,7 +233,11 @@ struct usb_stream *usb_stream_new(struct usb_stream_kernel *sk,
> else
> sk->freqn = get_usb_high_speed_rate(sample_rate);
>
> - init_urbs(sk, use_packsize, dev, in_pipe, out_pipe);
> + if (init_urbs(sk, use_packsize, dev, in_pipe, out_pipe) < 0) {
> + usb_stream_free(sk);
> + return NULL;
> + }
> +
> sk->s->state = usb_stream_stopped;
> out:
> return sk->s;
> --
> 2.14.3
>

\
 
 \ /
  Last update: 2017-11-06 13:15    [W:0.047 / U:0.020 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site