Re: kernel Oops: 0003 on usbhid_submit_report

On Fri, Aug 23, 2013 at 12:52 PM, Andreas Lillebø Holm
[off-list ref] wrote:

Hi,

I just wanna followup this issue a bit more.

We have now tested on several hardware platform, and it is only on one platform it fails. The platform it fails on is Atom Intel e6xx CPU with EG20T PCH platform controller. We do not observe it on the Intel i7 3517UE CPU with Intel 6/7 series system controller.

We see this error around 4 times over 100 boots / 2 hours runtime.

After the Oops our process is in uninterruptable sleep state (D state). The process is not killable  when in this state. The process also uses around 70% CPU, maybe because it is waiting for the ioctl system call to return?

We haven't found another way to recover from besides rebooting the system.

How can we proceed to fix this issue?

IIRC, Jiri fixed some paging requests with the two following patches:
http://git.kernel.org/cgit/linux/kernel/git/jikos/hid.git/commit/?h=for-next&id=27ce405039bfe6d3f4143415c638f56a3df77dca
http://git.kernel.org/cgit/linux/kernel/git/jikos/hid.git/commit/?h=for-next&id=bc197eedef1ae082ec662c64c3f4aa302821fb7a

At least, these two patches are related to hid_output_report(). I
think they are scheduled for 3.12, so I can not find them in Linus'
tree.

It would worth trying them :)

Cheers,
Benjamin

On Thursday, August 15, 2013 at 8:37 AM, Greg KH wrote:

quoted

quoted

quoted

quoted

When communicating with AT90USB1287, at random intervals (1/25 boots)
the linux hid_output_field Oopses and kills the communicating thread.
The AT90USB1287 microcontroller uses LUFA library for usb/hid
communication. It is trigged by a ioctl call from userspace and fails
in a kernel paging request. The system is after the oops in a state
where no hid commands is sent anymore and only a boot can fix the
system.

Keywords: usbhid hid

Kernel version: Linux version 3.8.13-03081301-generic (apw@gomeisa) (gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) ) #201305311535 SMP Fri May 31 19:44:30 UTC 2013

Oopses:
http://paste.debian.net/24305
http://paste.debian.net/24306
http://paste.debian.net/24307

Code:
The error is triggered by:
ioctl(fd,HIDIOCSUSAGES, &ref_multi_u);
ioctl(fd,HIDIOCSREPORT, &rep_info_u);

Notes:
It is very hard to reproduce so seems like race condition…

Any tips to resolve/workaround this issue is appreciated and please
let me know if my information is incomplete (This is my first kernel
bug report)

Any chance you can try a supported kernel, like 3.10.6 or 3.11-rc5 to
see if that also causes problems? We can't do anything with
distro-specific kernel releases like your 3.8.13 release from Ubuntu,
sorry.

I've now tried with kernel 3.10.6 (http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.10.6-saucy/) and I can trigger the same issue on this kernel.

Here is a paste from last Oops on this kernel:
http://paste.debian.net/24993/

I am also using usbmon to monitor the usb bus traffic, but cannot see anything that should cause the driver to Oops.

Is there any way to find out what can trigger this issue?

--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html