FS#31393 - [xf86-video-ati] KMS+X broken for ATI Radeon RV250 [Mobility FireGL 9000] after upgrade

Attached to Project: Arch Linux
Opened by Javier (jevv) - Sunday, 02 September 2012, 20:52 GMT
Last edited by Andreas Radke (AndyRTR) - Saturday, 10 November 2012, 22:50 GMT
Task Type Bug Report
Category Upstream Bugs
Status Closed
Assigned To Andreas Radke (AndyRTR)
Architecture i686
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 2
Private No

Details

Description:

I have a dell inspiron 600m laptop with the following card:

01:00.0 VGA compatible controller: Advanced Micro Devices [AMD] nee ATI Radeon RV250 [Mobility FireGL 9000] (rev 01) (prog-if 00 [VGA controller])
Subsystem: Dell Device 011e
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop+ ParErr- Stepping+ SERR+ FastB2B- DisINTx-
Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Latency: 32 (2000ns min), Cache Line Size: 32 bytes
Interrupt: pin A routed to IRQ 11
Region 0: Memory at e8000000 (32-bit, prefetchable) [size=128M]
Region 1: I/O ports at c000 [size=256]
Region 2: Memory at fcff0000 (32-bit, non-prefetchable) [size=64K]
[virtual] Expansion ROM at fc000000 [disabled] [size=128K]
Capabilities: [58] AGP version 2.0
Status: RQ=48 Iso- ArqSz=0 Cal=0 SBA+ ITACoh- GART64- HTrans- 64bit- FW+ AGP3- Rate=x1,x2,x4
Command: RQ=32 ArqSz=0 Cal=0 SBA+ AGP+ GART64- 64bit- FW- Rate=x4
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-

On X I do not use any desktop, just plain fluxbox (I do use xorg-xdm though). My settings on X for the card are:

% cat /etc/X11/xorg.conf.d/20-card.conf
Section "Device"
Identifier "Card0"
Driver "radeon"
Option "ColorTiling" "on" #enabled by default on RV300 and later radeon cards.
EndSection

I perform system upgrades every weekend. Today after the upgrade and reboot (linux got upgraded as well) I got into the X session without problems, and got no error reported on /var/log/Xorg.0.log, however the X session was unworkable. Pieces of whatever window gets loaded are not visible, and instead some letters shown, some others just show some blurry colors, etc.

I attempted commenting out and setting to off my ColorTiling option recommended by https://wiki.archlinux.org/index.php/ATI, but no luck. I attempted moving from early loading of the radeon module:

% grep radeon /etc/mkinitcpio.conf
MODULES="intel_agp radeon"

To a late loading:

% grep radeon /etc/mkinitcpio.conf
# MODULES="intel_agp radeon"

Of course with re-generation of the initramfs images after any modification of the mkinitcpio.conf file and corresponding reboot...

That didn't make a difference. Then what I did was with early loading of the module, disable KMS, and that did work:

% grep radeon /etc/modprobe.d/modprobe.conf
options radeon modeset=0

However, this work around is far from good. The performance of the driver with KMS disabled is really poor. Moving between desktops is really slow for example, and selecting for copy/pasting is really slow, etc. But the worse part, is that this disables console usage, since without KMS the console resolution and fonts size makes impossible to work there...

Notice I've been working fine with xf86-video-ati radeon driver and KMS since the very begining KMS support on linux started. Today's update was pretty unfortunate, and the first time I've had to disable KMS.

Today's uprade included several updates, between them a linux one, ATI drivers one, and xorg-xserver one...


Additional info:

* package version(s)

Here are the latest packages I think related:

core/linux 3.5.3-1 (base) [installed]
extra/xorg-server 1.12.4-1 (xorg) [installed]
extra/xf86-video-ati 1:6.14.6-1 (xorg-drivers xorg) [installed]
extra/ati-dri 8.0.4-3 [installed]

* config and/or log files etc.


Steps to reproduce:
This task depends upon

Closed by  Andreas Radke (AndyRTR)
Saturday, 10 November 2012, 22:50 GMT
Reason for closing:  Fixed
Comment by Laurent Carlier (lordheavy) - Thursday, 06 September 2012, 11:30 GMT
Please provide output of dmesg, also your Xorg.0.log file
Comment by Andreas Radke (AndyRTR) - Thursday, 06 September 2012, 13:39 GMT
For testing you can pacman -Rdd ati-dri or disable any acceleration to check whether the issue is kernel drm related, dri related or in the ddx driver. Also check you last pacman log what was updated.
Comment by Javier (jevv) - Saturday, 08 September 2012, 16:41 GMT
Attached goes the /var/log/Xorg.0.log...
Comment by Javier (jevv) - Saturday, 08 September 2012, 17:14 GMT
Removing the ati-dri package (mesa) didn't help a bit... Attached goes /var/log/Xorg.0.log without mesa dri...
Comment by Sam Stuewe (HalosGhost) - Saturday, 08 September 2012, 18:39 GMT
As confirmation, this bug appears to only be present on i686 architectures. I would also like to add that it affects multiple graphics cards (OPs and mine—Radeon X1200 series—at least). I'd be happy to provide any relevant logs as well should they be helpful.

All the best,

-HG
Comment by Andreas Radke (AndyRTR) - Sunday, 09 September 2012, 13:31 GMT
https://bugs.freedesktop.org/show_bug.cgi?id=54129

does this fit your issue? then help out there if needed.
Comment by Javier (jevv) - Sunday, 23 September 2012, 00:31 GMT
Today's linux update might have included a patch:

local/linux 3.5.4-1 (base)
The linux kernel and modules

Cause now things look normal with KMS enabled. Even though according to the freedesktop bug mentioned the fix was only merged with linux tree 3.6-rc5+:

https://bugs.freedesktop.org/show_bug.cgi?id=54129#c22

So I was waiting for it 3.6-rc6 to try the fix out (such kernel is now available), however there was no need, since current Arch linux image seems OK...

Thanks.


=====================


Bad luck, I might have been doing something wrong, cause Today after util-linux update, the bug still hits me on 3.5 stock linux image.

I then installed linux-mainline from AUR, which is already at 3.6-rc6, and that solved the problem as stated on:

https://bugs.freedesktop.org/show_bug.cgi?id=54129#c22

So, arch linux image still does not fix the problem, but AUR linu-mainline with version 3.6-rc6 already does...

I wanted to make sure to clarify this since the bug is already closed, but the fix seems to be available only when linux kernel stable version reaches 3.6, and arch core adopts it, :-)

Thanks,
Comment by Javier (jevv) - Friday, 09 November 2012, 16:59 GMT
  • Field changed: Percent Complete (100% → 0%)
Might be I did a bad test yesterday, might be Today's update of util-linux affected...

However now I'm having the same thing and I had to disable KMS again, trying to compile linux-mainline from AUR to see if that helps...
Comment by Andreas Radke (AndyRTR) - Friday, 09 November 2012, 17:01 GMT
Please report with recent 3.0.x LTS kernel and 3.6.x kernel - please also try the new ati 7.0.0 driver in testing. And make sure tu use KMS. UMS isn't supported anymore in 7.0.0 driver.

If the error is still there please get in contact upstream. There's not much we can do downstream here.
Comment by Javier (jevv) - Saturday, 10 November 2012, 22:27 GMT
I thought I added a comment about, but current linux stock kernel:

% uname -a
Linux m0 3.6.6-1-ARCH #1 SMP PREEMPT Mon Nov 5 13:14:42 CET 2012 i686 GNU/Linux

Solves the problems as expected from 3.6 kernels, :-)

I just found an issue that I don't know if associated, but stock midori on gtk2 gives blurry fonts at times (not always), specially when typing or doing something... See attachment...
Comment by Javier (jevv) - Saturday, 10 November 2012, 22:29 GMT
sorry, I pushed the button too fast, :-) Now the attachment is there...

midori-git with gtk3 seems more stable BTW, on gmail some times I see the same behavior. Right now is OK... So again I believe not the same issue then.

Thanks,

Javier.

Loading...