Bug 21967 - xpress200m GPU hang, kernel/x loop when attempting DRI/DRM
xpress200m GPU hang, kernel/x loop when attempting DRI/DRM
Status: RESOLVED FIXED
Product: DRI
Classification: Unclassified
Component: DRM/Radeon
XOrg git
x86-64 (AMD64) Linux (All)
: medium normal
Assigned To: Default DRI bug account
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2009-05-27 08:18 UTC by William L. Thomson Jr.
Modified: 2011-08-01 12:59 UTC (History)
1 user (show)

See Also:


Attachments
Xorg log of loop at end (111.86 KB, text/plain)
2009-05-27 08:18 UTC, William L. Thomson Jr.
no flags Details
xorg config (1.34 KB, text/plain)
2009-05-27 08:19 UTC, William L. Thomson Jr.
no flags Details
another X log file this time with no loop, also xorg-server 1.6 with xf86-driver-ati from git (47.18 KB, text/plain)
2009-05-27 08:22 UTC, William L. Thomson Jr.
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description William L. Thomson Jr. 2009-05-27 08:18:21 UTC
Created attachment 26249 [details]
Xorg log of loop at end

No go on DRI. I can load kernel modules, radeon and drm. X starts, but graphics are corrupted, top half screen still black. Stops half way it seems. Mouse works. I can log into gdm though I can't see it. Immediately x crashes and/or enters kernel loop. Same if I try to restart or just stop X. Anything beyond moving mouse, causes issue.

So no dri at all. Without everything works fine, just horrible slow...

x11-base/xorg-x11-7.4
x11-base/x11-drm-20090320
x11-base/xorg-server-1.5.3-r6
x11-drivers/xf86-video-ati-6.12.2
x11-libs/libdrm-2.4.11
media-libs/mesa-7.4.2

kernel /2.6.29-gentoo-r4 ro root=/dev/hda6 vga=804 nomodeset

Also tried x11-drm and xf86-drivers-ati from git. xf86-drivers-ati does seem to be a bit better and worse than version mentioned above. With from source, screen is totally black. Better than half corrupted. Mouse works, but without any other action. I get the kernel/gpu loop. Message at the bottom repeating over and over.

Currently running
x11-base/xorg-server-1.6.1.901-r3

No diff between 1.5 or 1.6 wrt to my problem.

dmesg stuffx11-base/xorg-server-1.6.1.901-r3

[drm] Initialized drm 1.1.0 20060810
pci 0000:01:05.0: BAR 0: can't reserve mem region [0xc8000000-0xcfffffff]
[drm] Initialized radeon 1.29.0 20080613 on minor 0
[drm] Used old pci detect: framebuffer loaded
[drm] Setting GART location based on new memory map
[drm] Loading R300 Microcode
[drm] Num pipes: 1
[drm] writeback test succeeded in 1 usecs
[drm] wait idle failed status : 0x80010140 0x00000000
[drm] wait for fifo failed status : 0x8001C100 0x00000000

last line repeats till hard reboot :(
Comment 1 William L. Thomson Jr. 2009-05-27 08:19:20 UTC
Created attachment 26250 [details]
xorg config
Comment 2 William L. Thomson Jr. 2009-05-27 08:22:06 UTC
Created attachment 26251 [details]
another X log file this time with no loop, also xorg-server 1.6 with xf86-driver-ati from git
Comment 3 William L. Thomson Jr. 2009-05-29 15:20:26 UTC
Updated to a vanilla 2.6.30-rc7 kernel. With drm and radeon driver from kernel. I get a similar issue with dri enabled. Black screen, only a mouse, X is running, but 100% cpu.

However this time kernel reports

drm] writeback test failed

Where before that succeeded. Just some additional info, not sure what else I can do to provide useful info, testing feedback :)
Comment 4 Jason Wilson 2009-06-16 18:17:31 UTC
Have a very similar problem with a:
  ATI Technologies Inc RC410 [Radeon Xpress 200M]
This is on Fedora 11.

With me I dont even get a mouse, just a blank screen and 100% in the X server.

By disabling DRI (added Option "DRI" "off" to xorg.conf) I can at least now login.

Using:
  xorg-x11-drv-ati-6.12.2-16.fc11.i586.rpm
found at:
  http://koji.fedoraproject.org/koji/packageinfo?packageID=95

Happy to provide additional details if of assistance.
Comment 5 William L. Thomson Jr. 2009-06-16 19:47:11 UTC
Blank screen, mouse, and 100% CPU is what I get now, or with any 2.6.30 kernel. Using in kernel drm and radeon drivers/modules. Also at one point I recall seeing the kernel say the number of pipes was 2, not 1.
Comment 6 Alex Deucher 2009-12-15 08:25:52 UTC
This might be related to bug 22140. Does the patch there help?
Comment 7 William L. Thomson Jr. 2009-12-15 13:15:16 UTC
Not sure, presently using in kernel stuff 2.6.32 vanilla, and appears that patch is for newer sources. Might have to pull from git and load as a module or something. Thanks for keeping this bug in mind. I will for sure try out the patch ASAP and report back. Very much appreciated :)
Comment 8 Alex Deucher 2009-12-15 13:42:28 UTC
(In reply to comment #7)
> Not sure, presently using in kernel stuff 2.6.32 vanilla, and appears that
> patch is for newer sources. Might have to pull from git and load as a module or
> something. Thanks for keeping this bug in mind. I will for sure try out the
> patch ASAP and report back. Very much appreciated :)
> 

Patch is against the ddx (xf86-video-ati) not the kernel.
Comment 9 William L. Thomson Jr. 2009-12-15 13:48:43 UTC
Ah ok thanks, my bad, will look into that then and comment in a few :)
Comment 10 William L. Thomson Jr. 2009-12-15 15:10:41 UTC
No change, tried that patch and one other for same file. Really no change :( Still failed with 99% cpu usage, blank screen, etc.

[drm] Setting GART location based on new memory map
[drm] Loading R300 Microcode
platform radeon_cp.0: firmware: using built-in firmware radeon/R300_cp.bin
[drm] Num pipes: 1
[drm] writeback test failed

xorg-x11-7.4-r1
xorg-server-1.7.3
xf86-video-ati-8.12.4 (patched)
vanilla 2.6.32 kernel