Bug 93181 - [855gm] Kernel panic on JVC MP-XV941 (ASUS J3N platform) unless booted with nomodeset
Summary: [855gm] Kernel panic on JVC MP-XV941 (ASUS J3N platform) unless booted with n...
Status: RESOLVED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - Intel) (show other bugs)
Hardware: i386 Linux
: P3 normal
Assignee: intel-gfx-bugs@lists.freedesktop.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-02-13 14:26 UTC by John Paul Adrian Glaubitz
Modified: 2015-10-07 10:53 UTC (History)
2 users (show)

See Also:
Kernel Version: 3.18.6
Subsystem:
Regression: No
Bisected commit-id:


Attachments
lspci for JVC MP-XV941 (10.69 KB, text/plain)
2015-02-13 14:26 UTC, John Paul Adrian Glaubitz
Details
dmesg for JVC MP-XV941 (46.90 KB, text/plain)
2015-02-13 14:27 UTC, John Paul Adrian Glaubitz
Details
Xorg.0.log for JVC MP-XV941 running xorg-video-vesafb (71.68 KB, text/plain)
2015-02-13 14:27 UTC, John Paul Adrian Glaubitz
Details
Screenshot of the moment the machine panics. (917.71 KB, image/jpeg)
2015-02-16 11:07 UTC, John Paul Adrian Glaubitz
Details
Kernel OOPS logged with netconsole (18.04 KB, text/plain)
2015-02-16 21:08 UTC, John Paul Adrian Glaubitz
Details
dmesg with kernel 4.0.0; system no longer panics (49.67 KB, text/plain)
2015-05-16 20:36 UTC, John Paul Adrian Glaubitz
Details

Description John Paul Adrian Glaubitz 2015-02-13 14:26:43 UTC
Created attachment 166651 [details]
lspci for JVC MP-XV941

Hello!

I recently installed Debian unstable on a JVC MP-XV941 subnotebook which is apparently based on the J3N platform by ASUS according to the kernel messages (asus_laptop driver). The laptop has already the latest BIOS version (0203).

After rebooting the laptop after the installation, I got a kernel panic which turned out to be related to the modesetting. The machine is based around the Intel 855 graphics chipset which means it's using the i915 kernel module.

I have tried both kernel 3.16.7 from Debian unstable as well as kernel 3.18.6 which is the most current kernel we currently have in Debian, both kernel panic unless I adding "nomodeset" to the kernel command line. Thus, I need to resort to running X.Org using the VESA driver since I can't use the Intel driver.

I have the laptop sitting here on my desk running Debian unstable and can install kernels compiled from git to test any patches, so I am open to helping to debug this issue.

Please let me know if you need to know anything else.

Cheers,
Adrian
Comment 1 John Paul Adrian Glaubitz 2015-02-13 14:27:08 UTC
Created attachment 166661 [details]
dmesg for JVC MP-XV941
Comment 2 John Paul Adrian Glaubitz 2015-02-13 14:27:42 UTC
Created attachment 166671 [details]
Xorg.0.log for JVC MP-XV941 running xorg-video-vesafb
Comment 3 John Paul Adrian Glaubitz 2015-02-13 14:28:59 UTC
Addendum: The installation report for Debian can be found here:

> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=778296

It may contain additional hardware information besides what "lspci -vvv" and "dmesg" provide.
Comment 4 Jani Nikula 2015-02-16 08:06:58 UTC
Did you actually see the oops? Can we see it too? ;)
Comment 5 John Paul Adrian Glaubitz 2015-02-16 11:01:59 UTC
(In reply to Jani Nikula from comment #4)
> Did you actually see the oops? Can we see it too? ;)

Sorry, I should have elaborated on that a bit more.

I can't currently see the oops since apparently the machine panics in such a way that the kernel is unable to print the oops. I assumed it was panicing as you can see from the keyboard LEDs flashing the usual way.

I am attaching a photo which shows how far the kernel boots until it panics. I would love to debug the issue using the serial port but the laptop doesn't have one, even on the external port expander I have.

Do you know whether there are other ways to debug it? Probably over USB, Firewire or a PCCard/PCMCIA card?

Adrian
Comment 6 John Paul Adrian Glaubitz 2015-02-16 11:07:45 UTC
Created attachment 167101 [details]
Screenshot of the moment the machine panics.
Comment 7 Jani Nikula 2015-02-16 17:00:26 UTC
(In reply to John Paul Adrian Glaubitz from comment #5)
> Do you know whether there are other ways to debug it? Probably over USB,
> Firewire or a PCCard/PCMCIA card?

netconsole might work https://www.kernel.org/doc/Documentation/networking/netconsole.txt
Comment 8 John Paul Adrian Glaubitz 2015-02-16 21:08:09 UTC
Amazing tip, that worked at first attempt and I got the OOPS :).

See attached output retrieved through netconsole.
Comment 9 John Paul Adrian Glaubitz 2015-02-16 21:08:59 UTC
Created attachment 167171 [details]
Kernel OOPS logged with netconsole
Comment 10 Jani Nikula 2015-02-17 08:05:54 UTC
(In reply to John Paul Adrian Glaubitz from comment #8)
> Amazing tip, that worked at first attempt and I got the OOPS :).

Awesome!

> See attached output retrieved through netconsole.

Ugh, a GPU hang immediately followed by a NULL pointer dereference. I suppose we should fix the oops first to be able to get the error state... unless Chris has some better ideas.
Comment 11 John Paul Adrian Glaubitz 2015-05-16 20:35:33 UTC
Hi there!

Just a short heads-up:

I updated to kernel 4.0.0 and the panic is gone, the machine now boots fine with modesetting enabled. There are still some more error messages from the DRM driver in the dmesg though which I am attaching. The graphics seem to work fine so far, there were some artifacts in the desktop wallpaper of the loaded MATE desktop which might be related to the error messages though.

Let me know if you need more, but in any case, this issue is at least partially resolved, in the sense that the kernel doesn't panic anymore.

Adrian
Comment 12 John Paul Adrian Glaubitz 2015-05-16 20:36:09 UTC
Created attachment 177071 [details]
dmesg with kernel 4.0.0; system no longer panics
Comment 13 Jani Nikula 2015-10-07 10:53:39 UTC
Panics seem to be fixed, closing.

Please file any new bug reports at the freedesktop.org bugzilla [1]. Thank you.

[1] https://bugs.freedesktop.org/enter_bug.cgi?product=DRI&component=DRM/Intel

Note You need to log in before you can comment on or make changes to this bug.