Bug 11240 - doesn't start X - DMA Queue hang
Summary: doesn't start X - DMA Queue hang
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/nouveau (show other bugs)
Version: 7.2 (2007.02)
Hardware: Other All
: medium normal
Assignee: Nouveau Project
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2007-06-11 15:47 UTC by Bill Nottingham
Modified: 2009-04-02 21:13 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
xorg conf (589 bytes, text/plain)
2007-06-11 15:48 UTC, Bill Nottingham
no flags Details
X log (64.09 KB, text/plain)
2007-06-11 15:49 UTC, Bill Nottingham
no flags Details
dmesg with debug=1 passed to drm (28.81 KB, text/plain)
2007-06-11 15:50 UTC, Bill Nottingham
no flags Details
X log (347.70 KB, text/plain)
2007-09-04 10:02 UTC, Bill Nottingham
no flags Details
dmesg with debug=1 passed to drm (34.35 KB, text/plain)
2007-09-04 10:03 UTC, Bill Nottingham
no flags Details
log with debugging (437.11 KB, application/x-bzip2)
2007-09-25 07:44 UTC, Bill Nottingham
no flags Details

Description Bill Nottingham 2007-06-11 15:47:37 UTC
startx dies with:

(II) Module already built-in
crtc 0 1360 1360
writting vpll 00027109
writting vpll2 00000000
output 0 debug_0 31101111
savep->dither 000F0000
0: crtc 0 output0: 0001: twocrt 1 twomon 1
crtc 1 1360 1360
writting vpll 00027109
writting vpll2 00027109
output 1 debug_0 31101111
savep->dither 000F0000
0: crtc 1 output1: 0101: twocrt 1 twomon 1
(EE) AIGLX error: dlopen of /usr/lib64/dri/nouveau_dri.so failed (/usr/lib64/dri/nouveau_dri.so: cannot open shared object file: No such file or directory)
(EE) AIGLX: reverting to software rendering

Fatal server error:
DMA queue hang: dmaPut=8a, current=0, status=0

writting vpll 00045405
writting vpll2 00045405
writting vpll 00045405
writting vpll2 00045405
XIO:  fatal IO error 104 (Connection reset by peer) on X server ":0.0"
      after 0 requests (0 known processed) with 0 events remaining.

01:00.0 VGA compatible controller: nVidia Corporation NV34GL [Quadro FX 500/600 PCI] (rev a1)
on an Intel E7505 chipset. Will attach X conf, x logs, etc.

xorg server 1.3.0, March 30 noveau snapshot (git head of the randr 1.2 branch beahves similarly).

Originally filed as http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=242773
Comment 1 Bill Nottingham 2007-06-11 15:48:41 UTC
Created attachment 10262 [details]
xorg conf
Comment 2 Bill Nottingham 2007-06-11 15:49:08 UTC
Created attachment 10263 [details]
X log
Comment 3 Bill Nottingham 2007-06-11 15:50:40 UTC
Created attachment 10264 [details]
dmesg with debug=1 passed to drm
Comment 4 Ben Skeggs 2007-06-11 18:27:52 UTC
It would be useful if you could bisect the git tree from the point the last working snapshot was known, to track down exactly what broke it (from the fedora bug: 2.0.1-1.fc7).
Comment 5 Bill Nottingham 2007-06-12 10:35:44 UTC
Going to be hard to directly bisect - the last one that worked was with an older copy of drm as well, and the driver and drm are codependent. Will see what I can turn up.
Comment 6 Bill Nottingham 2007-06-15 08:34:43 UTC
So, as best I can track down (due to the ddx driver, libdrm, and the kernel being all intertwined), it broke at some point between February 14th and March 30.
Comment 7 Stephane Marchesin 2007-08-31 03:47:01 UTC
Could you try with current GIT ? 10DE:032A works fine here, this card should be quite close to yours.
Comment 8 Bill Nottingham 2007-09-04 08:50:12 UTC
I still get a DMA queue hang; the hex is slightly different.
Comment 9 Stephane Marchesin 2007-09-04 09:50:21 UTC
Could you provide kernel & X logs with that latest git ? There have been numerous change since june, so that the errors might be totally different now.
Comment 10 Bill Nottingham 2007-09-04 10:02:33 UTC
Created attachment 11413 [details]
X log
Comment 11 Bill Nottingham 2007-09-04 10:03:02 UTC
Created attachment 11414 [details]
dmesg with debug=1 passed to drm
Comment 12 Ben Skeggs 2007-09-22 05:09:02 UTC
Can you:

1) Turn on the NV_DMA_DEBUG option in src/nv_includes.h
2) Try Option "CBLocation" "VRAM" in xorg.conf
Comment 13 Bill Nottingham 2007-09-25 07:44:13 UTC
Created attachment 11746 [details]
log with debugging

OK, here's the log with debugging.
Comment 14 Stuart Bennett 2009-04-02 17:39:00 UTC
From https://bugzilla.redhat.com/show_bug.cgi?id=492399 can it be assumed this bug is resolved?  Also, feel free to open a bug here about your DVI issue (an indication of whether the display works with the open source "nv" or closed "nvidia" drivers would be a good starting point).
Comment 15 Bill Nottingham 2009-04-02 21:13:16 UTC
Yes, it works now.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.