lkml.org 
[lkml]   [2021]   [Feb]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: BUG: KASAN: use-after-free in snd_complete_urb+0x109e/0x1740 [snd_usb_audio] (5.11-rc6)
On Fri, 05 Feb 2021 22:02:16 +0100,
Mikhail Gavrilov wrote:
>
> Hi folks.
> On 5.11-rc6 (git 3aaf0a27ffc2) I caught a new issue.
> For unknown reason sound disappeared in my headset Hyperx orbit s.
> But after reconnecting to another USB port headset stopped being
> detected as USB device in dmesg
> and in log appears a record about bug KASAN: use-after-free.
>
> [37920.370627] xhci_hcd 0000:06:00.3: AMD-Vi: Event logged
> [IO_PAGE_FAULT domain=0x0013 address=0xfe836480 flags=0x0000]
> [37920.370720] xhci_hcd 0000:06:00.3: WARNING: Host System Error
> [37925.514865] xhci_hcd 0000:06:00.3: xHCI host not responding to stop
> endpoint command.
> [37925.514875] xhci_hcd 0000:06:00.3: USBSTS: HCHalted HSE EINT HCE
> [37925.514883] xhci_hcd 0000:06:00.3: xHCI host controller not
> responding, assume dead
> [37925.515700] xhci_hcd 0000:06:00.3: HC died; cleaning up
> [37925.515760] usb 3-1: USB disconnect, device number 2
> [37925.515765] usb 3-1.1: USB disconnect, device number 5
> [37925.515896] ==================================================================
> [37925.515899] BUG: KASAN: use-after-free in
> snd_complete_urb+0x109e/0x1740 [snd_usb_audio]
> [37925.515919] Read of size 8 at addr ffff88813a264260 by task
> kworker/19:0/649395
>
> [37925.515926] CPU: 19 PID: 649395 Comm: kworker/19:0 Tainted: G
> W --------- ---
> 5.11.0-0.rc6.20210203git3aaf0a27ffc2.144.fc34.x86_64 #1
> [37925.515930] Hardware name: System manufacturer System Product
> Name/ROG STRIX X570-I GAMING, BIOS 3402 01/13/2021
> [37925.515934] Workqueue: usb_hub_wq hub_event
> [37925.515940] Call Trace:
> [37925.515943] <IRQ>
> [37925.515945] dump_stack+0xae/0xe5
> [37925.515955] print_address_description.constprop.0+0x18/0x160
> [37925.515960] ? snd_complete_urb+0x109e/0x1740 [snd_usb_audio]
(snip)

This must be a stray URB processed after the disconnection.
Does the patch below help?


thanks,

Takashi

--- a/sound/usb/card.c
+++ b/sound/usb/card.c
@@ -127,6 +127,8 @@ static void snd_usb_stream_disconnect(struct snd_usb_stream *as)
subs = &as->substream[idx];
if (!subs->num_formats)
continue;
+ snd_usb_endpoint_sync_pending_stop(subs->sync_endpoint);
+ snd_usb_endpoint_sync_pending_stop(subs->data_endpoint);
subs->data_endpoint = NULL;
subs->sync_endpoint = NULL;
}
\
 
 \ /
  Last update: 2021-02-06 05:25    [W:0.105 / U:0.236 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site