kms and nonkms use different display names. HDMI-1 vs DVI1
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| xf86-video-intel |
Fix Released
|
Medium
|
|||
| linux (Ubuntu) |
Fix Released
|
Low
|
Tim Gardner | ||
| Karmic |
Fix Released
|
Low
|
Tim Gardner | ||
Bug Description
Binary package hint: xserver-
After KMS was enabled by default my display setup was lost. With xrandr i found what instead of HDMI-1 and HDMI-2 i have DVI1 and DVI2.
Each time i switch between kms and nonkms boot it switch the names.
so to setup my display on nonKMS boot i use:
xrandr --output HDMI-1 --mode 1920x1080 --output HDMI-2 --mode 1280x1024 --left-of HDMI-1
and with KMS:
xrandr --output DVI1 --mode 1920x1080 --output DVI2 --mode 1280x1024 --left-of DVI1
I think it should be fixed for compatibility reasons.
lsb_release -rd
Description: Ubuntu karmic (development branch)
Release: 9.10
apt-cache policy xserver-
xserver-
Installiert: 2:2.7.99.
Kandidat: 2:2.7.99.
Versions-Tabelle:
*** 2:2.7.99.
500 http://
100 /var/lib/
2:
500 http://
I guess this is not a bug, and it may even be deliberate. Jesse should be able to give a stronger reason than me.
Probably true, but it's make annoying to test KMS. specially it's not really stable on my PC.
Binary package hint: xserver-
After KMS was enabled by default my display setup was lost. With xrandr i found what instead of HDMI-1 and HDMI-2 i have DVI1 and DVI2.
Each time i switch between kms and nonkms boot it switch the names.
so to setup my display on nonKMS boot i use:
xrandr --output HDMI-1 --mode 1920x1080 --output HDMI-2 --mode 1280x1024 --left-of HDMI-1
and with KMS:
xrandr --output DVI1 --mode 1920x1080 --output DVI2 --mode 1280x1024 --left-of DVI1
I think it should be fixed for compatibility reasons.
lsb_release -rd
Description: Ubuntu karmic (development branch)
Release: 9.10
apt-cache policy xserver-
xserver-
Installiert: 2:2.7.99.
Kandidat: 2:2.7.99.
Versions-Tabelle:
*** 2:2.7.99.
500 http://
100 /var/lib/
2:
500 http://
+ kms and nonkms use different display names. HDMI-1 vs DVI1
(In reply to comment #2)
> Probably true, but it's make annoying to test KMS. specially it's not really
> stable on my PC.
>
We are more interested in those true KMS issues which make you feel unstable.
Could you file them according to http://
For the name, I think "1" and "-1" won't be fixed, but "HDMI" -> "DVI" may be a problem. Could you elaborate your board model, and which output interfaces it provide and used by you?
I have Intel board DG45ID with with G45 graphic. There is one DVI-I and one HDMI connectors on this board. To the HDMI interface attached iiyama ProLite B2409HDS monitor and to DVI iijama ProLite E1700S. For the BIOS DVI is the primary interface - i case there is two monitors attached it will display on DVI all BIOS configurations. For the intel-linux driver HDMI is the primary interface.
Created an attachment (id=27145)
xorg.log
Created an attachment (id=27146)
dmesg
Created an attachment (id=27147)
xrandr
Hi bug-track,
Please attach the output of `lspci -vvnn`, and attach your /var/log/Xorg.0.log (and maybe Xorg.0.log.old) file from after reproducing this issue. If you've made any customizations to your /etc/X11/xorg.conf please attach that as well.
[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]
removed: needs-lspci-vvnn needs-xorglog
Jesse, this problem does exist in the latest kernel. Can we fix it in Q3?
Created an attachment (id=29941)
Xorg.0.log, lspci and xrandr output on X200s
I have the same/similar problem on my Lenovo Thinkpad X200s running up-to-date Kubuntu Karmic. My system sometimes boots with KMS (fast) and sometimes with non-KMS (slow) - I do not understand what is causing this, but in general I would say it boots 70% in KMS, the rest in non-KMS.
My laptop has only VGA port, but docking station has DisplayPort. When booted in KMS mode, DisplayPort interface is reported as "DVI2", while in non-KMS it is reported as HDMI-2. Also note that laptop's panel is reported differently: LVDS1 (KMS) and LVDS (KMS).
- Xorg.0log, lspci and xrandr logs from X200s Edit (15.3 KiB, application/x-tar)
I have the same problem on X200s: it is not clear to me why some boots are KMS and some nonKMS...
This happens on up-to-date Kubuntu Karmic. I have also tried xorg-edgers packages, but have the same behavior (have downgraded now as I was not sure if PPA packages were causing it).
Should I submit another bug for booting randomly with KMS and nonKMS on X200s?
The DVI vs HDMI issue should be fixed by this commit (queued up in drm-intel-next for 2.6.32):
commit 8d91104aac6e21e
Author: Adam Jackson <email address hidden>
Date: Wed Sep 23 15:08:29 2009 -0400
drm/i915: Initialize HDMI outputs as HDMI connectors, not DVI.
I think there's a separate RFE open for the compatibility names (short story is that moving to the kernel changed the names for all drivers, but at least they all agree now, unlike with the various UMS drivers).
OK, found out that non-KMS boots were caused by kernel bug (module load order) - see bug 430694
According to the upstream bug report, this issue cropped up when stuff moved into the kernel, but should now be fixed (or at least made consistent across all drivers). Not sure if the upstream patch is included in Karmic though; refiling against the kernel for further review.
Upstream has provided proper fix for this - see discussion here:
http://
Patch was recently accepted into 2.6.32-rc4:
Adam Jackson (1):
drm/i915: Initialize HDMI outputs as HDMI connectors, not DVI.
Link to git commit in Linus's tree:
http://
Would it be possible to cherry-pick this fix for Karmic? Current KMS/non-KMS boot changes ID's (breaking user scripts for dock un-dock) and I would not like that upgrade from Karmic -> Lucid will break this again (due to DVI -> HDMI change).
This bug was fixed in the package linux - 2.6.31-13.45
---------------
linux (2.6.31-13.45) karmic; urgency=low
[ Luke Yelavich ]
* [Config] Add sd_mod to scsi-modules udeb for powerpc
[ Mario Limonciello ]
* SAUCE: Update to LIRC 0.8.6
- LP: #432678
[ Tim Gardner ]
* [Config] CONFIG_X86_PAT=y
https:/
https:/
* [Config] Add armel arch to linux-libc-dev arches.
- LP: #449637
[ Upstream Kernel Changes ]
* e1000e: swap max hw supported frame size between 82574 and 82583
* drm/i915: Initialize HDMI outputs as HDMI connectors, not DVI.
- LP: #392017
* ALSA: hda - Add quirks for some HP laptops
- LP: #449742
* ALSA: hda - Add support for HP dv6
- LP: #449742
-- Tim Gardner <email address hidden> 2009年10月12日 11:27:10 -0600