Bug 28307

Summary: xf86-video-radeon black screen (Linux) or corrupt screen (FreeBSD) and hard-lock when acceleration enabled
Product: xorg Reporter: Stuart Longland <stuartl>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED INVALID QA Contact: Xorg Project Team <xorg-team>
Severity: critical    
Priority: medium    
Version: unspecified   
Hardware: x86 (IA32)   
OS: All   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Linux Xorg.0.log with acceleration enabled
none
Linux Xorg.0.log with acceleration disabled (NoAccel = "true")
none
Linux kernel log (dmesg)
none
Linux "top" output showing X hogging the CPU
none
FreeBSD Xorg.0.log with acceleration enabled
none
FreeBSD Xorg.0.log with acceleration disabled (NoAccel = "true")
none
FreeBSD kernel log (dmesg)
none
FreeBSD "top" output showing X hogging the CPU
none
FreeBSD xorg.conf as generated by `X -configure`
none
Linux xorg.conf currently in use (autogenerated then NoAccel added)
none
Linux `lspci -v` output
none
Linux: Xorg.0.log with DRI disabled none

Description Stuart Longland 2010-05-28 17:33:41 UTC
Created attachment 35915 [details]
Linux Xorg.0.log with acceleration enabled

Hi,

I recently dusted off an old computer and started setting it up.  I have the machine successfully booting Windows 2000 Pro with full acceleration and no issues, but am having difficulty getting X running with acceleration on FreeBSD 8.0 and Gentoo Linux.

The video card is a HIS Radeon 9200SE (ATI RV280):
01:00.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200 SE] (rev 01) (prog-if 00 [VGA controller])
        Subsystem: Hightech Information System Ltd. Radeon 9200 SE Excalibur
        Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 11
        Memory at d8000000 (32-bit, prefetchable) [size=128M]
        I/O ports at 9000 [size=256]
        Memory at e8020000 (32-bit, non-prefetchable) [size=64K]
        [virtual] Expansion ROM at e8000000 [disabled] [size=128K]
        Capabilities: [58] AGP version 3.0
        Capabilities: [50] Power Management version 2
        Kernel driver in use: radeonfb

Monitor is a Sun Microsystems GDM-5410 20" CRT connected via the standard VGA port (no DVI in use).  Single-head configuration.

Operating systems/kernels affected:
- Linux 2.6.34 ("v2.6.34" git tag; git://repo.or.cz/linux-2.6.git)
- FreeBSD 8.0 "GENERIC" as provided on the FreeBSD 8.0 DVD

In the case of FreeBSD; I am using the binaries as provided by the FreeBSD team.  Under Linux, they are self compiled from Gentoo's package set.  In the latter case; HAL support is enabled.  In both cases, configuration was generated using `X -configure`.

Symptoms:

When an attempt is made to start X; the server appears to start, but then locks up in some kind of loop where it chews most of the CPU time.  Under Linux a black screen is shown to the user.  Under FreeBSD; a corrupted display is shown, likely the residual text-only display data, now being interpreted as graphic data (i.e. the lockup occurs between mode switch and video data being cleared).  In both cases, the system is still working, and can be accessed via SSH.

In both operating systems; I can get a usable display if I enable the NoAccel option for the video card device in xorg.conf.  This results in a usable, but very slow display.

This video card used to perform well under Linux, so I suspect a regression over the last few years since I last used this computer.  The problem is not experienced under Windows 2000, but is experienced under both FreeBSD and Linux, which therefore makes me suspect a component common to both: X.org.

The bug appears to be similar to bug 9284; however disabling DRI doesn't appear to have any effect -- the lockup occurs regardless.

Attached are the logs and other information I have.  I haven't been able to find a duplicate bug, but I could have been using the wrong keywords too.  If there's any additional information that could be useful, let me know and I'll try to provide it.
Comment 1 Stuart Longland 2010-05-28 17:34:31 UTC
Created attachment 35916 [details]
Linux Xorg.0.log with acceleration disabled (NoAccel = "true")
Comment 2 Stuart Longland 2010-05-28 17:35:11 UTC
Created attachment 35917 [details]
Linux kernel log (dmesg)
Comment 3 Stuart Longland 2010-05-28 17:35:46 UTC
Created attachment 35918 [details]
Linux "top" output showing X hogging the CPU
Comment 4 Stuart Longland 2010-05-28 17:36:21 UTC
Created attachment 35919 [details]
FreeBSD Xorg.0.log with acceleration enabled
Comment 5 Stuart Longland 2010-05-28 17:38:05 UTC
Created attachment 35920 [details]
FreeBSD Xorg.0.log with acceleration disabled (NoAccel = "true")
Comment 6 Stuart Longland 2010-05-28 17:38:45 UTC
Created attachment 35921 [details]
FreeBSD kernel log (dmesg)
Comment 7 Stuart Longland 2010-05-28 17:39:24 UTC
Created attachment 35922 [details]
FreeBSD "top" output showing X hogging the CPU
Comment 8 Stuart Longland 2010-05-28 17:40:03 UTC
Created attachment 35923 [details]
FreeBSD xorg.conf as generated by `X -configure`
Comment 9 Stuart Longland 2010-05-28 17:41:30 UTC
Created attachment 35924 [details]
Linux xorg.conf currently in use (autogenerated then NoAccel added)
Comment 10 Stuart Longland 2010-05-28 17:42:58 UTC
Created attachment 35925 [details]
Linux `lspci -v` output
Comment 11 Stuart Longland 2010-05-29 22:27:57 UTC
Created attachment 35948 [details]
Linux: Xorg.0.log with DRI disabled

Well, additional info, I now have accelerated 2D in Linux... which will keep me happy for the short term...  By setting "DRI" to "false" in the Device section for the video card, I was able to disable the 3D acceleration component, and therefore can at least have a reasonable desktop working in 2D.

FreeBSD hasn't been tested at this point ... I have been re-arranging my OSes, and will need to do a re-install of FreeBSD ... but this is a small task.

If anyone has any ideas as to what could be tripping up DRI however, I'm all ears.
Comment 12 Michel Dänzer 2010-06-04 10:37:08 UTC
Does either of the following options help with DRI enabled?

    Option "AGPMode" "4"
    Option "BusMode" "PCI"

(check in the log file that the option actually has an effect)
Comment 13 Adam Jackson 2018-06-12 19:07:34 UTC
Mass closure: This bug has been untouched for more than six years, and is not
obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.

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.