xrandr: Configure crtc 0 invalid time

Starting from today, my dual screen setup is broken with the "xrandr: Configure crtc 0 invalid time" output as well..

Same here - Intel chipset on Dell Latitude E6400.

xserver-xorg-video-intel/jaunty uptodate 2:2.7.99.901+git20090701.1e4784bf-0ubuntu0sarvatt~jaunty
x11-xserver-utils/jaunty uptodate 7.4+1

Same problem on jaunty with the same xserver-xorg-video-intel 2.7.99.901+git20090701
Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)

If I try to change the screen setup with xrandr I get an error message "xrandr: Configure crtc 0 invalid time"

I can confirm this bug on ThinkPad X200s with Intel GM45.

I am using the following command to connect my external monitor on docking station:
xrandr --verbose --output LVDS --mode 1440x900 --output VGA --mode 1680x1050 --right-of LVDS

This is reported:
screen 0: 3120x1050 565x190 mm 140.25dpi
crtc 0: 1680x1050 60.0 +1440+0 "VGA"
xrandr: Configure crtc 0 invalid time
crtc 0: disable
crtc 1: disable
screen 0: revert
crtc 0: revert
crtc 1: revert

Same here on MSI PR201 with Intel graphics and GM45 on Xubuntu 9.04:

xrandr --verbose --output VGA --auto --output LVDS --off

crtc 0: 1440x900 59.9 +0+0 "VGA"
xrandr: Configure crtc 0 invalid time
crtc 0: disable
crtc 1: disable
screen 0: revert
crtc 0: revert
crtc 1: revert

I have to add that few simple calls work:
xrandr --verbose --output LVDS --off
crtc 1: disable

xrandr --verbose --output VGA --off
crtc 0: disable

And for those who need to change few things surprisingly granrd (xrandr gui frontend) still can change res and reactivate deactivated outputs. In my opinion this bug is due to some new stuff in edge xorg or edge intel driver, as downgrading to:

xserver-xorg-core 2:1.6.0-0ubuntu14
xserver-xorg-video-intel 2:2.6.3-ubuntu9.3

solved this trouble.

Hope this helps

Fabio:
I tryed your suggestion
xrandr --verbose --output LVDS --off
xrandr --verbose --output VGA --off
(before the line to change resolution.)

but I cannot confirm it to work.
Disabled my displays and because setting resolutioon after disabling displays wasnt accepted. So configuration reverted to displays off. Had to restart X.

@ jayprakash

I knew those calls turned screens off (I need to do that sometimes and, as you saw, that still works), and I can't get them work again as you claim too. I did not meant after that call screen worked again.

Anyway have you tried grandr? Try to install it, it still works (nearly anything) on my box: I can enable and disable screens and change res, that seems what you need to do.

Oops... I didn't notice:

"Changing configuration with ubuntu display config works flawlessly."

Sorry

Fabio:
heh, I didn't understand you post like that :)

updated xserver-xorg-video-intel to version
2:2.7.99.901+git20090702.74227141-0ubuntu0sarvatt~jaunty
no change

Same here on Intel GMA X4500.

Can anybody figure out why grandr still works while both xranrd and arandr do not?

Same here with Intel Mobile 945GME.

Only restarting X allows me to set back e.g. the --panning to the normal setting.

Still same on Mobile Intel GM45, with latest version:
xserver-xorg-video-intel 2:2.7.99.901+git20090708.r1.40e7c950-0ubuntu0sarvatt~jaunty

Still same with 20090709

A plead to package maintainer:
Please confirm what is the root of the problem.
intel drivers? -then why does ubuntu display settings work.
xrandr code?

For it looks like, the developers made some big change in driver, and xrandr didn't count with that.
Also I believe that grandr and ubuntu display settings are using underlying libraries of xrandr, and not xrandr binary (as arandr does).
That would mean only some little bug in some if/else part of the xrandr code and not in libxrandr2.

Please...

Can somebody please try to nail down which version of the -intel driver caused the breakage? The best would be to know the exact git commit, but the package version should be fine too, since there are usually very few commits per new package.

If 20090629 was working but 20090701 was not, I would suspect http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=cec9fc6f6cffce186606f39982d0d78ff7c63bbf or http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=53e248af4365232416d2143a791a07c6751f8319 but there were many commits at that time.

Are you all using KMS?

Please also attach Xorg.0.log and the output from xrandr --verbose

I really cannot tell definitely if I'm using KMS :)
https://launchpad.net/~xorg-edgers
kernel 2.6.30-9 (when I reboot it wil be 2.6.30-10)
xorg-server - 2:1.6.2~git20090708
xserver-xorg-video-intel - 2:2.7.99.901+git20090708 (after reboot 20090709) (can not reboot now, I'm in work :) )
Default settings (+ disabling TV, virtual size and dontzap):
Section "Monitor"
 Identifier "TV"
 Option "Ignore" "True"
EndSection

Section "Monitor"
 Identifier "Configured Monitor"
EndSection

Section "Screen"
 Identifier "Default Screen"
 Monitor "Configured Monitor"
 Device "Configured Video Device"
 SubSection "Display"
  Virtual 2732 2048
 EndSubSection
EndSection

Section "Device"
 Identifier "Configured Video Device"
 Option "monitor-TV" "TV"
EndSection

Section "ServerFlags"
 Option "DontZap" "False"
EndSection

I'm using the default Jaunty kernel ( 2.6.28 ), so i'm pretty sure KMS is off here..

2.6.30-10
X didnt start, so back on 2.6.30-9 .
With new 20090709 intel driver
xrandr is the same.

If you are not using Karmic, nor the Jaunty kernel in the xorg-edgers PPA, and did not set any options as explained on https://wiki.ubuntu.com/X/KernelModeSetting , you are probably not using KMS. You would see display detection and "mode setting" messages in dmesg if you are running KMS.

Do you have the old packages, for instance in /var/cache/apt/archives, please verify which version works and not.

The xrandr problem appeared after the update from xserver-xorg-video-intel Version 2:2.7.99.901+git20090630.cec9fc6f-0ubuntu0sarvatt~jaunty to 2:2.7.99.901+git20090701.1e4784bf-0ubuntu0sarvatt~jaunty.

So 20090630 worked fine. Starting with 20090701, and in all following versions xrandr doesn't work.
The only thing that works is switching off the laptop screen with xrandr --output LVDS --off. But I can't switch it on again. Trying this, I get:
xrandr --output LVDS --auto
xrandr: Configure crtc 1 invalid time

I don't use KMS.

@Tomrod Volden
With the link you provided I activated KMS on 2.6.30-10 (xorg-edgers on jaunty).
With full update from xorg-edgers ... the performance is ... breath-taking.... wow.. (with compiz)
(2.6.30-10 doesn't like virtual size setting in xorg.conf, Xserver crashes on startup.)

About this bug :
outputs are renamed to <output>1

$ xrandr
Screen 0: minimum 320 x 200, current 1360 x 1536, maximum 2048 x 2048
VGA1 connected 1360x768+0+0 (normal left inverted right x axis y axis) 0mm x 0mm
   1360x768 59.8*
   1152x864 60.0
   1024x768 60.0
   800x600 60.3
   640x480 59.9
LVDS1 connected 1024x768+0+768 (normal left inverted right x axis y axis) 0mm x 0mm
   1024x768 60.0*+ 85.0 75.0 70.1 60.0
   832x624 74.6
   800x600 85.1 72.2 75.0 60.3 56.2
   640x480 85.0 72.8 75.0 59.9
   720x400 85.0
   640x400 85.1
   640x350 85.1
DVI1 disconnected (normal left inverted right x axis y axis)
TV1 disconnected (normal left inverted right x axis y axis)

But problem is the same:
$ xrandr --verbose --output LVDS1 --auto
crtc 1: 1024x768 60.0 +0+768 "LVDS1"
xrandr: Configure crtc 1 invalid time
crtc 0: disable
crtc 1: disable
screen 0: revert
crtc 0: revert
crtc 1: revert

In Display preferences, both displays are "Unknown", but the resolutions they are capable of, are detected correctly.

Packages I have are:
20090611, 20090701, 20090702, 20090708, 20090709

But if you create any package from git, I'll gladly test it, immediately.

Thanks.

tomwu, thanks, that is a wonderfully precise report. This leaves us four possible commits:
- uxa: Fix segfault on source-only picture usage with FallbackDebug.
- Fix 945GM VT switch in UMS
- Load fbcon too if kernel mode setting is checked on
- Don't try to pin buffers in KMS mode
where the last two seem KMS only and the first is not so relevant. The second seems highly relevant:
http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=7e79fc8aa93df4df37c25cf37ee0ec6c7caca1d9

I have made a comment in the corresponding bug report http://bugs.freedesktop.org/show_bug.cgi?id=19578

Actually, I'm more inclined to think it was the xserver update at the same time than the intel driver.

http://cgit.freedesktop.org/xorg/xserver/commit/?h=server-1.6-branch&id=8d65439d5c950ea01ec8e1e4dd989aff0fb4c3f0

https://bugs.freedesktop.org/show_bug.cgi?id=21987

That is most likely like Robert says!

tomwu, did you try these two -intel versions without changing other packages, or was this observed together with updating the xserver-xorg-core package?

you are incredible :)

reverted to xserver-xorg-video-intel_2%3a2.7.99.901+git20090611.6d062e9e-0ubuntu0sarvatt~jaunty_i386.deb

$ xrandr --verbose --output LVDS1 --auto
crtc 1: 1024x768 60.0 +0+768 "LVDS1"
xrandr: Configure crtc 1 invalid time
crtc 0: disable
crtc 1: disable
screen 0: revert
crtc 0: revert
crtc 1: revert

Could anyone having this problem try a karmic livecd from today (or later if you read this later) to see if the problem exists? Karmic at the moment has the 0702 git snapshot of the intel driver and the older xserver 1.6.1.901.

http://cdimage.ubuntu.com/daily-live/current/

test of xserver-xrg-core:
intel driver back to 20090709

xserver-xorg-core reverted to:
xserver-xorg-core_2%3a1.6.1.902~git20090629+server-1.6-branch.5b49aff4-0ubuntu0sarvatt~jaunty_i386.deb
bug still present

xserver-xorg-core reverted to:
xserver-xorg-core_2%3a1.6.1.901+git20090523+server-1.6-branch.5cd5a012-0ubuntu0sarvatt5_i386.deb

working
$ xrandr --verbose --output LVDS1 --autocrtc 1: 1024x768 60.0 +0+768 "LVDS1"

I don't have any package in-between.
But i would say you found the problem :)

xserver-xorg-core (2:1.6.1.901-2ubuntu2)
http://packages.ubuntu.com/en/karmic/xserver-xorg-core

works
(tested on jaunty)

@Tormod: Sorry, I haven't seen there was another update of the xserver-xorg-core. :-(

I reverted xserver-xorg-core to:
xserver-xorg-core_2%3a1.6.1.901+git20090622+server-1.6-branch.dbac41b6-0ubuntu0sarvatt~jaunty_i386.deb
working and xrandr works with
xserver-xorg-video-intel_2%3a2.7.99.901+git20090709.9155cfca-0ubuntu0sarvatt~jaunty_i386.deb

So it seems, not to be the intel driver.

Reassigning to xorg-server, this will be an issue when karmic updates to xserver 1.6.2.

As mentioned in the bug report
https://bugs.freedesktop.org/show_bug.cgi?id=21987

this could well be a bug in the xrandr client tool. The xrandr client tool always uses CurrentTime, instead of the precise timestamp as returned from XRRGetScreenResources. What still escapes me, is why grandr is still working although it uses CurrentTime as well.

FYI: xrandr 1.3.0 built from source works as expected.

Jaunty contains xrandr 1.2.99.4 (see https://launchpad.net/ubuntu/jaunty/+source/x11-xserver-utils/7.4+1). Between 1.2.99.4 and 1.3.0 are only 4 commits. I bisected the problem and beb228842c77fab4a446e028dd501a01f0fb3fd3 (http://cgit.freedesktop.org/xorg/app/xrandr/commit/?id=beb228842c77fab4a446e028dd501a01f0fb3fd3) resolved the problem.

So the resolution is to update to xrandr 1.3.0 in sync with the xserver changes which revealed the problem in xrandr.

Thanks for tracking that down! I updated the package on edgers to fix it in the meantime since its isolated there but we'd want to upgrade it in karmic when we move to xserver 1.6.2.

Oh silly me.. it's already released in karmic! No wonder I couldn't reproduce it. Anyhow, the package is in xorg-edgers now for jaunty, and the bug is limited to the xserver update jaunty will never get officially.

Confirming repaired.
xorg-edgers on jaunty

Thank you.

Solved for me too.

Thanks everybody