Bug 102830 - [QM67] Screen Flickering Until Error
Summary: [QM67] Screen Flickering Until Error
Status: CLOSED INVALID
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard: ReadyForDev
Keywords:
Depends on:
Blocks:
 
Reported: 2017-09-18 02:47 UTC by Yan-Fa Li
Modified: 2018-04-27 07:40 UTC (History)
2 users (show)

See Also:
i915 platform: SNB
i915 features: display/FBC


Attachments
Xorg log file (28.69 KB, text/x-log)
2017-09-18 02:47 UTC, Yan-Fa Li
no flags Details
dmesg output (57.58 KB, text/plain)
2017-09-18 02:47 UTC, Yan-Fa Li
no flags Details
xrandr verbose (4.66 KB, text/plain)
2017-09-18 02:48 UTC, Yan-Fa Li
no flags Details
drm.debug=0x06 (95.40 KB, text/plain)
2017-09-18 02:52 UTC, Yan-Fa Li
no flags Details
drm.debug post error (97.50 KB, text/plain)
2017-09-18 03:05 UTC, Yan-Fa Li
no flags Details
drm-tip dmesg drm.debug=0x06 (94.64 KB, text/plain)
2017-09-20 15:13 UTC, Yan-Fa Li
no flags Details
attachment-23818-0.html (3.20 KB, text/html)
2017-09-21 15:35 UTC, Yan-Fa Li
no flags Details
attachment-16089-0.html (1.93 KB, text/html)
2018-03-29 18:54 UTC, Yan-Fa Li
no flags Details
4.16.3-1-ARCH with no flags (58.16 KB, text/plain)
2018-04-26 16:51 UTC, Yan-Fa Li
no flags Details
4.16.3-1-ARCH drm.debug=14 (100.49 KB, text/plain)
2018-04-26 16:55 UTC, Yan-Fa Li
no flags Details
4.16.3-1-ARCH drm.debug=14 run 2 (99.86 KB, text/plain)
2018-04-26 17:30 UTC, Yan-Fa Li
no flags Details

Description Yan-Fa Li 2017-09-18 02:47:07 UTC
Created attachment 134302 [details]
Xorg log file

Symptoms, screen flashes non-stop logging in X server until this error happens, then problem appears to go away and updates are stable. Is there some sort of reset that happens after this event?

[  444.294724] [drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe A FIFO underrun
[  444.294813] [drm:intel_set_pch_fifo_underrun_reporting [i915]] *ERROR* uncleared pch fifo underrun on pch transcoder A
[  444.294859] [drm:intel_pch_fifo_underrun_irq_handler [i915]] *ERROR* PCH transcoder A FIFO underrun

After this happens the display stabilizes. Until then it's a flickering resizing mess.

Chipset: Intel Corporation QM67 Express Chipset Family LPC Controller rev 04
Xorg: xorg-server 1.19.3-3
Arch: x86_64
Kernel: 4.12.13-1-ARCH
Distro: Arch Linux
Machine: Lenovo T420i
Screen: LVDS-1 connected primary 1600x900+0+0 (normal left inverted right x axis y axis) 309mm x 174mm

Reproduce: 100% of bootups
Comment 1 Yan-Fa Li 2017-09-18 02:47:53 UTC
Created attachment 134303 [details]
dmesg output
Comment 2 Yan-Fa Li 2017-09-18 02:48:40 UTC
Created attachment 134304 [details]
xrandr verbose
Comment 3 Yan-Fa Li 2017-09-18 02:52:08 UTC
Created attachment 134305 [details]
drm.debug=0x06
Comment 4 Yan-Fa Li 2017-09-18 02:53:29 UTC
This is a lot of stuff to ask an end-user to collect. Don't you guys have a bash script people can just download to run all the commands? I almost didn't collect all this stuff because it's so much work. That's a significant barrier to people reporting bugs.
Comment 5 Yan-Fa Li 2017-09-18 03:05:14 UTC
Created attachment 134306 [details]
drm.debug post error
Comment 6 Yan-Fa Li 2017-09-18 03:10:22 UTC
If I use the option i915.enable_fbc=0 the flickering goes away, but now my kernel is tainted.
Comment 7 Yan-Fa Li 2017-09-18 03:11:28 UTC
(In reply to Yan-Fa Li from comment #6)
> If I use the option i915.enable_fbc=0 the flickering goes away, but now my
> kernel is tainted.

Also I get weird chrome behavior when going full screen, so this isn't really an option.
Comment 8 Elizabeth 2017-09-19 15:58:26 UTC
If it's possible, could you try to reproduce with this branch: https://cgit.freedesktop.org/drm-tip?? Thanks.
Comment 9 Yan-Fa Li 2017-09-20 04:47:21 UTC
Hi, frankly this has been going on for a while. I would estimate 3-6 months. It's a laptop I use infrequently so I just hoped you guys would fix it. It finally got so irritating that I decided to report it. Before that it was flawless.

I'm not sure how to build drm-tip on an Arch system. I'll have to go look around for a package to modify. I'm highly skeptical that this would be fixed in tip, as Arch has pretty recent versions of all your packages, but I will give it a shot this weekend hopefully.

I've been following this bug on and off from the bugzilla, but when I noticed you closed it with no further reports I decided to open a new one with the exact information requested. I understand Sandy Bridge is an old chipset and is probably not that common any more, so any help you can provide is most appreciated.
Comment 10 Yan-Fa Li 2017-09-20 04:48:11 UTC
OK, looks like I can just use this https://aur.archlinux.org/packages/linux-drm-tip-git/
Comment 11 Yan-Fa Li 2017-09-20 06:52:08 UTC
I thought I would mention that this is an incredibly painful thing to ask your userbase to do. I've been waiting a couple of hours for a full kernel to build and it's still not complete. I'm pretty sure asking normal non-developers to do this is possibly the most hostile thing you can do. Hopefully one day you guys can come up with a better system, like a diagnostic program that someone can compile easily and just run on their system. I'll let it finish building, but I just wanted to point this out to you folks.
Comment 12 Yan-Fa Li 2017-09-20 15:13:53 UTC
Created attachment 134370 [details]
drm-tip dmesg drm.debug=0x06

uname -a
Linux ptah 4.14.0-1035f22af3e97 #1 SMP PREEMPT Wed Sep 20 00:00:42 PDT 2017 x86_64 GNU/Linux
Comment 13 Yan-Fa Li 2017-09-20 15:14:47 UTC
Still flickering, terminal pops in and out. Let me know if you want me to capture a video of this, I can upload it to youtube.
Comment 14 Yan-Fa Li 2017-09-20 15:15:34 UTC
Flickering stops when this error occurs

[  156.352019] [drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe A FIFO underrun
[  156.352120] [drm:intel_set_pch_fifo_underrun_reporting [i915]] *ERROR* uncleared pch fifo underrun on pch transcoder A
[  156.352169] [drm:intel_pch_fifo_underrun_irq_handler [i915]] *ERROR* PCH transcoder A FIFO underrun
[  156.352311] [drm:intel_fbc_underrun_work_fn [i915]] Disabling FBC due to FIFO underrun.
Comment 15 Elizabeth 2017-09-20 15:22:23 UTC
Thanks for all the time invested on this, seems to be related with synchronization. The video would be helpful too, could you share it?
Comment 16 Yan-Fa Li 2017-09-20 15:31:36 UTC
You're welcome. Video is here: https://youtu.be/p1PlyM4GpwA
Comment 17 Jani Nikula 2017-09-21 09:22:22 UTC
(In reply to Yan-Fa Li from comment #11)
> I thought I would mention that this is an incredibly painful thing to ask
> your userbase to do. I've been waiting a couple of hours for a full kernel
> to build and it's still not complete. I'm pretty sure asking normal
> non-developers to do this is possibly the most hostile thing you can do.
> Hopefully one day you guys can come up with a better system, like a
> diagnostic program that someone can compile easily and just run on their
> system. I'll let it finish building, but I just wanted to point this out to
> you folks.

To be blunt, this bugzilla for drm/i915 really isn't intended for normal non-developers. Sometimes it works out, but more often than not we can't reproduce the bugs ourselves, and we need to have the bug reporters test kernel patches. That means configuring and building the kernel.

The alternative for normal end-users is your distro bug database.

FWIW, the long build time is likely due to using a distro kernel config which will build literally thousands of kernel modules that your machine won't ever need. Also make sure to use make -j<N> option for parallel builds.
Comment 18 Yan-Fa Li 2017-09-21 15:35:01 UTC
Created attachment 134410 [details]
attachment-23818-0.html

Understood. I still think some simple scripting could help enormously for
collecting the information. I agree the drm-tip package is optimized for
generality, but it does facilitate easy.

I'm appreciative of the work you folks do, I would just like the process of
helping you help me be smoother and less awkward. Getting bug reports is a
useful thing in most projects, especially if the quality is good. I'll bear
in mind in future that this bugzilla is meant for distros and developers in
the main. Thank you for your insight and input.

On Thu, Sep 21, 2017, 02:22 <bugzilla-daemon@freedesktop.org> wrote:

> *Comment # 17 <https://bugs.freedesktop.org/show_bug.cgi?id=102830#c17> on
> bug 102830 <https://bugs.freedesktop.org/show_bug.cgi?id=102830> from Jani
> Nikula <jani.nikula@intel.com> *
>
> (In reply to Yan-Fa Li from comment #11 <https://bugs.freedesktop.org/show_bug.cgi?id=102830#c11>)> I thought I would mention that this is an incredibly painful thing to ask
> > your userbase to do. I've been waiting a couple of hours for a full kernel
> > to build and it's still not complete. I'm pretty sure asking normal
> > non-developers to do this is possibly the most hostile thing you can do.
> > Hopefully one day you guys can come up with a better system, like a
> > diagnostic program that someone can compile easily and just run on their
> > system. I'll let it finish building, but I just wanted to point this out to
> > you folks.
>
> To be blunt, this bugzilla for drm/i915 really isn't intended for normal
> non-developers. Sometimes it works out, but more often than not we can't
> reproduce the bugs ourselves, and we need to have the bug reporters test kernel
> patches. That means configuring and building the kernel.
>
> The alternative for normal end-users is your distro bug database.
>
> FWIW, the long build time is likely due to using a distro kernel config which
> will build literally thousands of kernel modules that your machine won't ever
> need. Also make sure to use make -j<N> option for parallel builds.
>
> ------------------------------
> You are receiving this mail because:
>
>    - You reported the bug.
>
>
Comment 19 Jani Saarinen 2018-03-29 07:10:21 UTC
First of all. Sorry about spam.
This is mass update for our bugs. 

Sorry if you feel this annoying but with this trying to understand if bug still valid or not.
If bug investigation still in progress, please ignore this and I apologize!

If you think this is not anymore valid, please comment to the bug that can be closed.
If you haven't tested with our latest pre-upstream tree(drm-tip), can you do that also to see if issue is valid there still and if you cannot see issue there, please comment to the bug.
Comment 20 Yan-Fa Li 2018-03-29 18:54:53 UTC
Created attachment 138434 [details]
attachment-16089-0.html

Hi it was still there last time I checked. I just disable the feature now.

On Thu, Mar 29, 2018, 00:11 <bugzilla-daemon@freedesktop.org> wrote:

> *Comment # 19 <https://bugs.freedesktop.org/show_bug.cgi?id=102830#c19> on
> bug 102830 <https://bugs.freedesktop.org/show_bug.cgi?id=102830> from Jani
> Saarinen <jani.saarinen@intel.com> *
>
> First of all. Sorry about spam.
> This is mass update for our bugs.
>
> Sorry if you feel this annoying but with this trying to understand if bug still
> valid or not.
> If bug investigation still in progress, please ignore this and I apologize!
>
> If you think this is not anymore valid, please comment to the bug that can be
> closed.
> If you haven't tested with our latest pre-upstream tree(drm-tip), can you do
> that also to see if issue is valid there still and if you cannot see issue
> there, please comment to the bug.
>
> ------------------------------
> You are receiving this mail because:
>
>    - You reported the bug.
>
>
Comment 21 Jani Saarinen 2018-04-25 07:01:56 UTC
Jani, any advice here?
Comment 22 Jani Nikula 2018-04-25 12:01:45 UTC
(In reply to Yan-Fa Li from comment #6)
> If I use the option i915.enable_fbc=0 the flickering goes away, but now my
> kernel is tainted.

Please try without any i915 module params set.
Comment 23 Yan-Fa Li 2018-04-25 20:01:54 UTC
Can I use my distro kernel. This machine is pretty slow and it takes quite a while to build one of your kernels. It's arch so it's a 4.16.2 based.
Comment 24 Jani Nikula 2018-04-26 09:57:03 UTC
(In reply to Yan-Fa Li from comment #23)
> Can I use my distro kernel. This machine is pretty slow and it takes quite a
> while to build one of your kernels. It's arch so it's a 4.16.2 based.

Yeah, please try that. No i915 module params, add drm.debug=14, see what happens. Attach dmesg all the way from boot.
Comment 25 Yan-Fa Li 2018-04-26 16:50:23 UTC
sadly it still does it. Here's the dmesg without the debug. I'll do another run with debug and attach it.
Comment 26 Yan-Fa Li 2018-04-26 16:51:34 UTC
Created attachment 139141 [details]
4.16.3-1-ARCH with no flags
Comment 27 Yan-Fa Li 2018-04-26 16:55:13 UTC
Created attachment 139142 [details]
4.16.3-1-ARCH drm.debug=14
Comment 28 Ville Syrjala 2018-04-26 17:03:27 UTC
(In reply to Yan-Fa Li from comment #26)
> Created attachment 139141 [details]
> 4.16.3-1-ARCH with no flags

You still have all kinds of potentially dangerous i915 modparams set:
[    4.911262] i915: unknown parameter 'enable_rc6' ignored
[    4.911263] Setting dangerous option enable_fbc - tainting kernel
[    4.911265] i915: unknown parameter 'semaphores' ignored

Please remove all those.
Comment 29 Ville Syrjala 2018-04-26 17:04:01 UTC
(In reply to Ville Syrjala from comment #28)
> (In reply to Yan-Fa Li from comment #26)
> > Created attachment 139141 [details]
> > 4.16.3-1-ARCH with no flags
> 
> You still have all kinds of potentially dangerous i915 modparams set:
> [    4.911262] i915: unknown parameter 'enable_rc6' ignored
> [    4.911263] Setting dangerous option enable_fbc - tainting kernel
> [    4.911265] i915: unknown parameter 'semaphores' ignored
> 
> Please remove all those.

And any other i915 modparam you have set.
Comment 30 Yan-Fa Li 2018-04-26 17:30:50 UTC
Created attachment 139143 [details]
4.16.3-1-ARCH drm.debug=14 run 2

Looks like it was self inflicted. Thank you for spotting that. I found it under modprobe.d for the driver itself. Please close this if you think the dmesg looks good now.
Comment 31 Jani Nikula 2018-04-27 07:28:31 UTC
(In reply to Yan-Fa Li from comment #30)
> Created attachment 139143 [details]
> 4.16.3-1-ARCH drm.debug=14 run 2
> 
> Looks like it was self inflicted. Thank you for spotting that. I found it
> under modprobe.d for the driver itself. Please close this if you think the
> dmesg looks good now.

Thanks for the follow-up. Closing.
Comment 32 Jani Saarinen 2018-04-27 07:40:30 UTC
Thanks, closing.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.