[Letux-kernel] hso madness

H. Nikolaus Schaller hns at goldelico.com
Sun Apr 16 19:12:07 CEST 2017


Hi,

> Am 16.04.2017 um 18:28 schrieb Andreas Kemnade <andreas at kemnade.info>:
> 
> Hi,
> 
> On Sat, 15 Apr 2017 13:59:19 +0200
> Andreas Kemnade <andreas at kemnade.info> wrote:
> 
>> On Sat, 15 Apr 2017 11:35:09 +0200
>> "H. Nikolaus Schaller" <hns at goldelico.com> wrote:
>> 
>>> Hi,
>>> 
>>>> Am 15.04.2017 um 11:07 schrieb Andreas Kemnade <andreas at kemnade.info>:
>>>> 
>>>> Hi,
>>>> 
>>>> I just did a rfkill unblock wwan
>>>> and a rmmod hso afterwards -> kernel oops
>>>> I think hso kindly asks for a git bisect time.
>>>> git bisecting should be easy because there are no relevant patches
>>>> in that area in our tree. For testing, exporting gpio186 should be fine.
>>>> And even the device (and hopefully also the problem) could possibly be
>>>> transferred via usbip to a big computer.. The weather forecast shows
>>>> cold weather here, so some git bisect + kernel compile heating will be
>>>> helpful ;-)
>>>> 
>>>> There is also the ugly commit:
>>>> 
>>>> d073ca63156bc8badea4b7d5f455fa96bb004ee8
>>>> Author: H. Nikolaus Schaller <hns at goldelico.com>
>>>> Date:   Sat Mar 4 15:56:23 2017 +0100
>>>> 
>>>>   drivers: net: hso: hack to avoid NULL pointer dereferencing - the reason is unknown. Perhaps some race condi
>>>> tion in the USB stack?
>>>> Last known good is 4.10?
>>> 
>>> Yes... this 4.11 contains several really bad regressions or the GTA04.
>>> This is the reason why we still do not yet have a really useable 4.11-rc6...
>>> 
>> No, that is older. 4.10 has also the problem.
>> How to reproduce:
>> rfkill unblock wwan
>> rmmod hso
>> 
>> No accesses to tty are needed.
> 
> known good 4.3
> known bad 4.7

good finding.

> 
> ouch. With all these horrible (with big musb bugs) kernels in between
> them.

Yes. Sadly.

Hm. maybe bisecting isn't the best strategy in this case. Trying to understand
what the NULL pointer is thought for and where it is (not) initialized may also
lead to the issue.

Or just submitting the bug report to LKML and have USB maintainers debug the
problem?

BR,
Nikolaus

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.goldelico.com/pipermail/letux-kernel/attachments/20170416/52ea827e/attachment.asc>


More information about the Letux-kernel mailing list