saucy regression - three monitors no longer work

[Roger Binns]

I've just upgraded from raring. There I had 3 monitors attached to my Intel HD Graphics 4000 (i7 3770 cpu) where it all works perfectly. (Two monitors are identical flanking my centre displayport monitor.)

In saucy I can get any two monitors working, and any attempt to get all 3 working results in an error about setting crtc with a different number.

ProblemType: Bug
DistroRelease: Ubuntu 13.10
Package: xserver-xorg-video-intel 2:2.99.904-0ubuntu2
ProcVersionSignature: Ubuntu 3.11.0-12.19-generic 3.11.3
Uname: Linux 3.11.0-12-generic x86_64
ApportVersion: 2.12.5-0ubuntu2
Architecture: amd64
Date: Sat Oct 19 14:28:59 2013
MarkForUpload: True
SourcePackage: xserver-xorg-video-intel
UpgradeStatus: Upgraded to saucy on 2013-10-19 (0 days ago)

[Roger Binns (ubuntu-rogerbinns)]

The failures to set occur using the graphical tool from system settings, as well as the xrandr command line

xrandr --output DP2 --primary --auto --output HDMI3 --auto --left-of DP2 --output HDMI1 --auto --right-of DP2

[Roger Binns (ubuntu-rogerbinns)]

I also ran xorg-edgers-live-test as described in https://wiki.ubuntu.com/XorgOnTheEdge and that didn't help. Restored from my backup and everything works fine.

(xorg-edgers-live-test also needs updating - it tries to restart gdm when it should be restarting lightdm)

[Roger Binns (ubuntu-rogerbinns)]

xrandr output running command line from comment 2. Raring:

crtc 0: 1920x1200 60.0 +1920+0 "DP2"
crtc 1: 1920x1080 60.0 +3840+0 "HDMI1"
crtc 2: 1920x1080 60.0 +0+0 "HDMI3"

Saucy livecd:

xrandr: Configure crtc 2 failed
crtc 0: 1920x1080 60.0 +3840+0 "HDMI1"
crtc 1: 1920x1080 60.0 +0+0 "HDMI3"
crtc 2: 1920x1200 60.0 +1920+0 "DP2"
crtc 0: disable
crtc 1: disable
crtc 2: disable
crtc 3: disable
screen 0: revert
crtc 0: revert
crtc 1: revert
crtc 2: revert
crtc 3: revert

[RichardNeill (ubuntu-richardneill)]

FYI, this is an easy way you can test the latest Xorg nightlies:
 http://www.ubuntuupdates.org/ppa/xorg-edgers

[Roger Binns (ubuntu-rogerbinns)]

@RichardNeill - already did all that as mentioned in comment 4 which also has the advantage of working on the livecd.

[Roger Binns (ubuntu-rogerbinns)]

Booting the raring kernel (3.8.0-33) under saucy results in all 3 monitors working correctly.

[Timo Aaltonen (tjaalton)]

so it's a kernel regression, marking as such

[Timo Aaltonen (tjaalton)]

and keeping the ddx entry open so that it's easier to find

[Timo Aaltonen (tjaalton)]

so, things to try:

- does 3.12 work
- if not, does drm-intel-nightly from the mainline ppa work
- if 3.12 works, maybe do a rough bisect through the 3.12-rc's from the ppa

if none of these work, then maybe doing a rough bisect to see which kernel broke it. Likely some -rc1, where most new code is merged.

[Roger Binns (ubuntu-rogerbinns)]

I installed 3.12 from http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.12-saucy/

It failed to boot. With normal boot there was no further progress after ramdisk loading. In recovery mode the kernel said it failed to execute /init but gave no further details.

I am now trying http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-nightly/2013-11-07-saucy/

[Roger Binns (ubuntu-rogerbinns)]

The drm-nightly from previous comment works

[Timo Aaltonen (tjaalton)]

ok, maybe try an earlier 3.12-rc if the final is somehow broken?

[Roger Binns (ubuntu-rogerbinns)]

I'm confused as to what the goal is. The drm-intel-nightly works.

[Timo Aaltonen (tjaalton)]

because of

:: tjaalton@eldon:~/src/kernel/linux.git (master)> git log1 v3.11..i915/drm-intel-next drivers/gpu/drm/i915|wc -l
754

it would be nice to know where it got fixed.. drm-intel-next is basically what'll end up in 3.13-rc1. If you are willing to bisect the diff then fine, but I'd like to get a smaller set of commits to go through :)

[Roger Binns (ubuntu-rogerbinns)]

This didn't work:

http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.11.7-saucy/

This did:

http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.12-rc1-saucy/

It would be ideal if you gave me a list of urls like above in preferred testing order

[Timo Aaltonen (tjaalton)]

Nice, so we're down to 354 commits, which is a start..

[Roger Binns (ubuntu-rogerbinns)]

Can you point me to other builds you want me to try, ideally as ppa links like the above?

[penalvch (penalvch)]

Roger Binns, this bug was reported a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue? If so, could you please test for this with the latest development release of Ubuntu? ISO images are available from http://cdimage.ubuntu.com/daily-live/current/ .

If it remains an issue, could you please run the following command in the development release from a Terminal (Applications->Accessories->Terminal), as it will automatically gather and attach updated debug information to this report:

apport-collect -p xserver-xorg-video-intel REPLACE-WITH-BUG-NUMBER

Please note, given that the information from the prior release is already available, doing this on a release prior to the development one would not be helpful.

Thank you for your understanding.

Helpful bug reporting tips:
https://wiki.ubuntu.com/ReportingBugs

[Roger Binns (ubuntu-rogerbinns)]

Kernel 3.12 even in the prereleases fixed this.

[penalvch (penalvch)]

Roger Binns, thank you for your comment. Would you need a backport to a release prior to Trusty, or may this be closed as Status Invalid?

[Roger Binns (ubuntu-rogerbinns)]

It isn't invalid. This is a very real issue affecting the existing Ubuntu release.

I have been personally compiling 3.12 kernels from kernel.org as a workaround that works for me, so I don't need anything.

I'd suggest closing as Won't Fix since Ubuntu won't provide a fix.

[penalvch (penalvch)]

Roger Binns, thank you for your comment. Of course you are mistaken in your presumption that a fix wouldn't be provided, as it hasn't even been submitted to the SRU process for evaluation as outlined in https://wiki.ubuntu.com/StableReleaseUpdates . Despite this, as Trusty is re-based on the upstream/mainline 3.12.4 kernel, this would be fixed in it. So, getting back to the question not directly answered https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1242180/comments/23 would you like a backport or not?

[Roger Binns (ubuntu-rogerbinns)]

If Ubuntu would actually backport the 3.12 kernel to Saucy then go for it. The kernel team PPA does not do 3.12 for saucy. At a minimum 3.12.6 is needed due to lots of btrfs fixes.

[penalvch (penalvch)]

Roger Binns, thank you for your comments. The scope of this bug report isn't the entire 3.12.x series kernel. It is only the issue noted in the Bug Title "three monitors no longer work". So, staying focused on this report, and it's scope, would you like a backport of the fix to this bug report?

[Roger Binns (ubuntu-rogerbinns)]

You do know that no one knows what the fix is, only that it is in 3.12? You can see that Timo gave up trying to find it in mid-November. If Ubuntu is actually going to figure out the fix and backport it, then go for it. However I am extremely sceptical and had to devise my own workaround.

[penalvch (penalvch)]

Roger Binns, thank you for taking the time to report this bug and helping to make Ubuntu better. However, I am closing it because as per your comment https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1242180/comments/28 the bug has been fixed in the latest development version of Ubuntu - Trusty Tahr.

This is a significant bug in Ubuntu. If you need a fix for the bug in previous versions of Ubuntu, please perform as much as possible of the SRU Procedure [1] to bring the need to a developer's attention.

[1]: https://wiki.ubuntu.com/StableReleaseUpdates#Procedure

[Matthew Braun (powerplug)]

I usually upgrade my system a few months after a release so bugs like this are identified and have plenty of time to be fixed before I upgrade. I just upgraded from 13.04 to 13.10 and ran into this exact bug (and currently stuck only being able to use two of my three monitors). It is unfortunate that this bug was identified and mostly tracked down then stalled without getting fixed.

As a user of Ubuntu, I would most certainly like to answer the question from Christopher "Would you need a backport to a release prior to Trusty" with a resounding YES! Please backport the fix in the kernel to the current "production" release of Ubuntu, Saucy Salamander (13.10).

It seems like Roger was very willing to try other 3.12 kernels to narrow down the bug fix and it would be great if Timo or whoever would continue working with Roger to narrow down the problem so it is easier to identify the bug fix and backport it to the current Ubuntu release.

I would much prefer to use the currently supported kernel instead of running a pre-Ubuntu-release kernel. But in the mean time, Roger, please share your notes in the comments with the workaround you are using to run a newer kernel with the fix (that is also hopefully stable), thanks!

Christopher, can you please reclassify the status of this bug and put in the request to have the fix backported?

[Roger Binns (ubuntu-rogerbinns)]

All I know is that whatever the cause was, it was fixed in 3.12 even in the rcs. I have no idea what the actual fix was.

Kernel 3.11 end of lifed November last year. I also use btrfs for which there have been maintenance patches in 3.12.

I followed the "Kernel Build and Installation" section at https://wiki.ubuntu.com/KernelTeam/GitKernelBuild