[Maverick] Removing external display on Vostro 3500 corrupts the screen

[Hankyone]

Binary package hint: xorg

Steps to reproduce:
Unplug external display from VGA port (assuming both monitors were working fine)
Click "Detect Monitors" on the Monitor Preferences panel

At this point, the screen either goes all black or black with white lines.
Mouse cursor is still visible and can be moved around, cursor also changes when hovering link, etc...

ProblemType: Bug
DistroRelease: Ubuntu 10.10
Package: xorg 1:7.5+6ubuntu3
ProcVersionSignature: Ubuntu 2.6.35-19.28-generic 2.6.35.3
Uname: Linux 2.6.35-19-generic i686
NonfreeKernelModules: wl
Architecture: i386
DRM.card0.DisplayPort.1:
 status: disconnected
 enabled: disabled
 dpms: On
 modes:
 edid-base64:
DRM.card0.HDMI_Type_A.1:
 status: disconnected
 enabled: disabled
 dpms: On
 modes:
 edid-base64:
DRM.card0.LVDS.1:
 status: connected
 enabled: enabled
 dpms: On
 modes: 1366x768
 edid-base64: AP///////wAlzAYAAAAAAAATAQOQIhN4CsiFnldUmyYSUFQAAAABAQEBAQEBAQEBAQEBAQEBxhtWf1AAGTAfFTQAWMIQAAAaxhtWf1AAGTAfFTQAWMIQAAAaAAAA/gBXNDY2UoAxNTZHVzAyAAAAAAAAAAAAAAAAAAEBCiAgAO8=
DRM.card0.VGA.1:
 status: disconnected
 enabled: disabled
 dpms: On
 modes:
 edid-base64:
Date: Fri Sep 3 11:25:09 2010
DkmsStatus: bcmwl, 5.60.48.36+bdcom, 2.6.35-19-generic, i686: installed
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Beta i386 (20100901.1)
Lsusb:
 Bus 002 Device 003: ID 0c45:6481 Microdia
 Bus 002 Device 002: ID 8087:0020 Intel Corp. Integrated Rate Matching Hub
 Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
 Bus 001 Device 002: ID 8087:0020 Intel Corp. Integrated Rate Matching Hub
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: Dell Inc. Vostro 3500
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.35-19-generic root=UUID=3b60b647-7ee6-4a07-8c0e-fcb20f80d049 ro quiet splash initcall_debug
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: xorg
Symptom: display
dmi.bios.date: 02/10/2010
dmi.bios.vendor: Dell Inc.
dmi.bios.version: A01
dmi.board.name: 0G2R51
dmi.board.vendor: Dell Inc.
dmi.board.version: A00
dmi.chassis.type: 8
dmi.chassis.vendor: Dell Inc.
dmi.chassis.version: Not Specified
dmi.modalias: dmi:bvnDellInc.:bvrA01:bd02/10/2010:svnDellInc.:pnVostro3500:pvrNotSpecified:rvnDellInc.:rn0G2R51:rvrA00:cvnDellInc.:ct8:cvrNotSpecified:
dmi.product.name: Vostro 3500
dmi.product.version: Not Specified
dmi.sys.vendor: Dell Inc.
glxinfo: Error: [Errno 2] No such file or directory
system:
 distro: Ubuntu
 codename: maverick
 architecture: i686
 kernel: 2.6.35-19-generic

[Chris-linux (chris-linux)]

Created an attachment (id=37244)
The actual messages

[Chris Wilson (ickle)]

I hereby declare that the submitter is of sound mind and neither his vision nor his judgment was impaired by popping too many pills.

Haven't got a clue as to what is going on, yet. It is not typical of the many Arrandale hangs so far reported -- in that it is *not* a hang.

(Whilst noting, Thomas said that his x61 would occasionally hang on unplugging from the dock - a spin not unlike the suspected PP_CONTROL hang.)

[Chris Wilson (ickle)]

I've managed to hit something very similar with:

xrandr --output LVDS1 --off --output DP1 --preferred

The registers look fine but the external monitor loses sync. Of particular note is that this change causes drm_remove_fb() to decouple the old framebuffer from crtc0.

[Chris-linux (chris-linux)]

I'm not sure if this helps or what it means, but the grey seems to come from the gtk theme - I'm guessing it's the window background colour...

[Chris Wilson (ickle)]

Hmm, I managed to bisect the failure on my x201s to

commit 1eb2cbb6d5efe129cd006691267ce513c0aa59da
Author: Al Viro <email address hidden>
Date: Thu May 27 11:11:06 2010 -0400

    Revert "anon_inode: set S_IFREG on the anon_inode"

    This reverts commit a7cf4145bb86aaf85d4d4d29a69b50b688e2e49d.

and reverting just that commit makes 'xrandr --output LVDS1 --off --output DP1 --preferred' work again.

I am baffled.

[Chris Wilson (ickle)]

For extra fun, it appears that udev is both the saviour and the culprit here.

That commit fixes udev running at 100%, and thus blocking uevents. A killall udevd causes the same failure as running without the reverted commit.

[Chris Wilson (ickle)]

Having learnt about the udevd anomaly, I repeated the bisection and found this as the potential culprit for my failure:

commit 7f8a85698f5c8a981641ec0bdf9926768786db9d
Author: Zhenyu Wang <email address hidden>
Date: Thu Apr 1 13:07:53 2010 +0800

    drm/i915: Add the support of memory self-refresh on Ironlake

    Update the self-refresh watermark for display plane/cursor and enable
    the memory self-refresh on Ironlake. The watermark is also updated for
    the active display plane.

    More than 1W idle power is saved on one Ironlake laptop after enabling
    memory self-refresh.

[Chris Wilson (ickle)]

Created an attachment (id=37621)
Update SR after dpms changes.

[Chris Wilson (ickle)]

To add further mystery, booting up with a single connection enabled, self-refresh works. Add an external, self-refresh is disabled as expected. Remove either, self-refresh is re-enabled and the output fails. Notably, during this sequence we also get warnings that we fail to turn off the pipe...

[Hankyone (hankyone)]

Binary package hint: xorg

Steps to reproduce:
Unplug external display from VGA port (assuming both monitors were working fine)
Click "Detect Monitors" on the Monitor Preferences panel

At this point, the screen either goes all black or black with white lines.
Mouse cursor is still visible and can be moved around, cursor also changes when hovering link, etc...

ProblemType: Bug
DistroRelease: Ubuntu 10.10
Package: xorg 1:7.5+6ubuntu3
ProcVersionSignature: Ubuntu 2.6.35-19.28-generic 2.6.35.3
Uname: Linux 2.6.35-19-generic i686
NonfreeKernelModules: wl
Architecture: i386
DRM.card0.DisplayPort.1:
 status: disconnected
 enabled: disabled
 dpms: On
 modes:
 edid-base64:
DRM.card0.HDMI_Type_A.1:
 status: disconnected
 enabled: disabled
 dpms: On
 modes:
 edid-base64:
DRM.card0.LVDS.1:
 status: connected
 enabled: enabled
 dpms: On
 modes: 1366x768
 edid-base64: AP///////wAlzAYAAAAAAAATAQOQIhN4CsiFnldUmyYSUFQAAAABAQEBAQEBAQEBAQEBAQEBxhtWf1AAGTAfFTQAWMIQAAAaxhtWf1AAGTAfFTQAWMIQAAAaAAAA/gBXNDY2UoAxNTZHVzAyAAAAAAAAAAAAAAAAAAEBCiAgAO8=
DRM.card0.VGA.1:
 status: disconnected
 enabled: disabled
 dpms: On
 modes:
 edid-base64:
Date: Fri Sep 3 11:25:09 2010
DkmsStatus: bcmwl, 5.60.48.36+bdcom, 2.6.35-19-generic, i686: installed
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Beta i386 (20100901.1)
Lsusb:
 Bus 002 Device 003: ID 0c45:6481 Microdia
 Bus 002 Device 002: ID 8087:0020 Intel Corp. Integrated Rate Matching Hub
 Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
 Bus 001 Device 002: ID 8087:0020 Intel Corp. Integrated Rate Matching Hub
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: Dell Inc. Vostro 3500
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.35-19-generic root=UUID=3b60b647-7ee6-4a07-8c0e-fcb20f80d049 ro quiet splash initcall_debug
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: xorg
Symptom: display
dmi.bios.date: 02/10/2010
dmi.bios.vendor: Dell Inc.
dmi.bios.version: A01
dmi.board.name: 0G2R51
dmi.board.vendor: Dell Inc.
dmi.board.version: A00
dmi.chassis.type: 8
dmi.chassis.vendor: Dell Inc.
dmi.chassis.version: Not Specified
dmi.modalias: dmi:bvnDellInc.:bvrA01:bd02/10/2010:svnDellInc.:pnVostro3500:pvrNotSpecified:rvnDellInc.:rn0G2R51:rvrA00:cvnDellInc.:ct8:cvrNotSpecified:
dmi.product.name: Vostro 3500
dmi.product.version: Not Specified
dmi.sys.vendor: Dell Inc.
glxinfo: Error: [Errno 2] No such file or directory
system:
 distro: Ubuntu
 codename: maverick
 architecture: i686
 kernel: 2.6.35-19-generic

[Chris Halse Rogers (raof)]

There have been a couple of recent commits to the Intel kernel module regarding Ironlake (this GPU) and VGA, and they've made it into 2.6.36-rc3

Does this work with a kernel from the Kernel team PPA: http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.36-rc3-maverick/ ? If so, we should look at backporting the relevant commits.

[Hankyone (hankyone)]

Nope, does not work.

[Chris Halse Rogers (raof)]

Ok. Worth a shot!

We'll want to send this upstream, then. Could you get a dmesg & Xorg.0.log log after reproducing this bug with “drm.debug=0x06” set on the kernel command line? Thanks.

[Hankyone (hankyone)]

Note: After I recreated the bug, I waited the countdown for the settings to revert back.

[Chris Halse Rogers (raof)]

Hm.

Ok. So, that log shows you booting with just the laptop's internal panel, plugging in a second display to the VGA port, which comes up in an extended desktop situation, then disabling the VGA output, presumably through the Monitors capplet or equivalent, and then letting the timeout expire to get back to the extended desktop.

Once it's back on the extended desktop the display returns to normal?

Given that the mouse pointer continues working properly I'd guess the problem is somewhere in the framebuffer settings for that particular mode change.

[Hankyone (hankyone)]

Yes, the display returns to normal when the countdown expires.

[Chris Wilson (ickle)]

In git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel.git drm-intel-staging we have a patch to disable self-refresh which is how we plan to sweep this bug under the carpet in the short term. /o\

http://git.kernel.org/?p=linux/kernel/git/ickle/drm-intel.git;a=commit;h=71c831d2cce539e0ee8e3b3fadae49355867efc2

[Chris Halse Rogers (raof)]

Aha! With that extra bit of information this looks to be upstream bug: https://bugs.freedesktop.org/show_bug.cgi?id=29173 which has just got an upstream fix http://git.kernel.org/?p=linux/kernel/git/ickle/drm-intel.git;a=commit;h=71c831d2cce539e0ee8e3b3fadae49355867efc2 .

Moving to the kernel, assigning jfo to ensure this gets looked at before kernel freeze.

[Chris Wilson (ickle)]

We've disabled self-refresh [bug 30125].

[Leann Ogasawara (leannogasawara)]

Looks like the kernel patch just landed upstream in 2.6.36-rc4. It was also Cc'd to upstream stable. I'd hope that it would then be included in the next 2.6.35.5 upstream stable release which would then mean we might get this patch automatically when we rebase to 2.6.35.5. Regardless, I've applied this patch and built a test kernel (see url below). Please test asap and let me know your results so we could possibly apply this as a pre-stable patch if 2.6.35.5 doesn't get released before Maverick kernel freeze on Thurs Sept 16. Thanks.

http://people.canonical.com/~ogasawara/lp629711/i386/

commit dd8849c8f59ec1cee4809a0c5e603e045abe860e
Author: Jesse Barnes <email address hidden>
Date: Thu Sep 9 11:58:02 2010 -0700

    drm/i915: don't enable self-refresh on Ironlake

[Hankyone (hankyone)]

Problem solved, I can confirm that Leann's kernel at #10 works.

[Leann Ogasawara (leannogasawara)]

Hi Anouar,

Thanks for the quick turn around testing. I've gone ahead and sent this patch to the Ubuntu kernel team mailing list for it to be considered for Maverick:

https://lists.ubuntu.com/archives/kernel-team/2010-September/012757.html

[kernel-janitor (kernel-janitor)]

Fix released in 2.6.35-22.32

https://launchpad.net/ubuntu/+source/linux/2.6.35-22.32