[Kernel BUG] After Connect to Any WiFi - Kernel Panic - not syncing: Fatal exception in interrupt.

[Tom Li]

According to a few users' report, since Linux 3.11, there is a major bug in the rtl8187 kernel driver. After connect to any wireless access point, we will got a kernel panic message. Blacklist rtl8187 and plug a external wireless card is a temporary solution.

I just noticed that after I was hit by this bug. I installed Linux 3.12.6 and got the kernel panic after connect to a wireless network. I can confirm the bug existed in the pure upstream kernel code. And I think the Loongson related code doesn't cause the bug.

Panic:

    ieee80211_tx_status_irqsafe+0x40/0x483 [mac80211]

    rtl8087_tx_cb+0x204/0x204 [rtl8187]

    usb_hcd_giveback_urb+0xc0/0x2c0

    ....

    loongson2_cpu_wait+0xa0/0xd8 [loongson2_cpufreq]  -> Note: Still panic if I blacklist the module. So it must work correctly.

    See the photos for full panic information.

    http://img.vim-cn.com/5d/d6fc985a61c7e9e9c435c68fb796f6ec9b8b9f.jpeg

    http://img.vim-cn.com/11/7d0014d2b19615b7e51be4f08bcc14a3491450.jpeg

I hope to collection more information before I report it to the upstream.

[Mark H Weaver]

Tom Li <bierga...@gmail.com> writes:
> According to a few users' report, since Linux 3.11, there is a major
> bug in the rtl8187 kernel driver. After connect to any wireless access
> point, we will got a kernel panic message.

I see the exact same problem with linux-libre 3.10.15.

Mark

[Xiyue Deng]

Same symptom on latest 3.12.5 backport kernel for stable. Is there already a bug entry for this?

> --
> You received this message because you are subscribed to the Google Groups "loongson-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev...@googlegroups.com.
> To post to this group, send email to loongs...@googlegroups.com.
> Visit this group at http://groups.google.com/group/loongson-dev.
> For more options, visit https://groups.google.com/groups/opt_out.

[Tom Li]

2014/1/23 Xiyue Deng <man...@gmail.com>:

>
> On Jan 22, 2014 11:09 AM, "Mark H Weaver" <m...@netris.org> wrote:
>>
>> Tom Li <bierga...@gmail.com> writes:
>> > According to a few users' report, since Linux 3.11, there is a major
>> > bug in the rtl8187 kernel driver. After connect to any wireless access
>> > point, we will got a kernel panic message.
>>
>> I see the exact same problem with linux-libre 3.10.15.
>>
>> Mark
>>
>
> Same symptom on latest 3.12.5 backport kernel for stable. Is there already a
> bug entry for this?
>

It seems there isn't any exist bug entry for it.
I want to collect more information before we report a bug.
Would anyone like to test the wireless on Linux 3.9 and Linux 3.8?

Thanks.

[Xiyue Deng]

IMHO you don't have to wait for more information. I suggest you just
file a bug now and add more info later if necessary.

[Tom Li]

2014-01-23 Xiyue Deng <man...@gmail.com>:

> On Wed, Jan 22, 2014 at 1:13 PM, Tom Li <bierga...@gmail.com> wrote:
>> 2014/1/23 Xiyue Deng <man...@gmail.com>:
>>>
>>> On Jan 22, 2014 11:09 AM, "Mark H Weaver" <m...@netris.org> wrote:
>>>>
>>>> Tom Li <bierga...@gmail.com> writes:
>>>> > According to a few users' report, since Linux 3.11, there is a major
>>>> > bug in the rtl8187 kernel driver. After connect to any wireless access
>>>> > point, we will got a kernel panic message.
>>>>
>>>> I see the exact same problem with linux-libre 3.10.15.
>>>>
>>>> Mark
>>>>
>>>
>>> Same symptom on latest 3.12.5 backport kernel for stable. Is there already a
>>> bug entry for this?
>>>
>>
>> It seems there isn't any exist bug entry for it.
>> I want to collect more information before we report a bug.
>> Would anyone like to test the wireless on Linux 3.9 and Linux 3.8?
>>
>> Thanks.
>>
>
> IMHO you don't have to wait for more information. I suggest you just
> file a bug now and add more info later if necessary.
>

I just found a bug on Kernel's Bugzilla which reported in Feb, 2013
with Linux 3.8.0. It seems the same issue:
https://bugzilla.kernel.org/show_bug.cgi?id=54391

I'll add more below the bug and send a mail to LKML for infomation.

[Mark H Weaver]

Tom Li <bierga...@gmail.com> writes:
> According to a few users' report, since Linux 3.11, there is a major
> bug in the rtl8187 kernel driver. After connect to any wireless access
> point, we will got a kernel panic message.

FYI, for those who haven't been following the bug tracker, we now have a
patch to fix this problem. Hopefully it will be upstream soon.

Thanks to Tom Li for doing the git bisect and making this happen! :)

http://www.linux-mips.org/archives/linux-mips/2014-02/msg00071.html

Mark

[Dennis Lan (dlan)]

H

> --
> You received this message because you are subscribed to the Google Groups "loongson-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev...@googlegroups.com.
> To post to this group, send email to loongs...@googlegroups.com.
> Visit this group at http://groups.google.com/group/loongson-dev.
> For more options, visit https://groups.google.com/groups/opt_out.

I still have problem with kernel 3.13 [1] which already include this fix
kernel oops when activate wifi module

dmesg: http://dev.gentoo.org/~dlan/yeeloong/dmesg-3.13.txt
kernel config: http://dev.gentoo.org/~dlan/yeeloong/config-3.13


[1] git://dev.lemote.com/linux-loongson-community.git, branch
remotes/origin/linux-3.13-stable

[Tom Li]

This is not the fault of rtl8187 driver, but something wrong with in
the Fn Key Handler of Loongson kernel module that I discovered when
just get my laptop.

I'm using my own kernel tree with many features. You can checkout the
source code and binary at here [0]. I check the source code of
Loongson kernel module, but I can't figure out the problem. I have to
create a patch [1] to disable to Fn handler (But Fn keys are still
working) to workaround the issue.

It is based on upstream kernel instead of the linux-loongson-community
because of the outdated code and some unknown bug. I added genpatches,
e-sources' (the kernel tree maintained by me in #gentoo-zh) patches,
and more loongson specific drivers ported linux-loongson-community and
more patches for bugfixes and enhancement.

BTW, can you or do you know any kernel developers can help us to
figure out the issue?

[0]: https://github.com/biergaizi/loongson-sources
[1]: patches/0003-loongson-sources/base/0002-loongson-disable-hotkey.patch,
loongson-source.

Regards,

Tom Li

> You received this message because you are subscribed to a topic in the Google Groups "loongson-dev" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/loongson-dev/Cq3Rv_4dXwA/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to loongson-dev...@googlegroups.com.