Bug 63053 - (ATI Rage Mobility M4) Screen corruption at gdm login
Summary: (ATI Rage Mobility M4) Screen corruption at gdm login
Alias: None
Product: Red Hat Public Beta
Classification: Retired
Component: XFree86 (Show other bugs)
(Show other bugs)
Version: skipjack-beta2
Hardware: i386 Linux
Target Milestone: ---
Assignee: Mike A. Harris
QA Contact: David Lawrence
Depends On:
TreeView+ depends on / blocked
Reported: 2002-04-09 17:05 UTC by adam.huffman
Modified: 2007-04-18 16:41 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2002-06-24 09:37:43 UTC
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)
X log file (35.20 KB, text/plain)
2002-04-09 17:06 UTC, adam.huffman
no flags Details
X config file (3.67 KB, text/plain)
2002-04-09 17:07 UTC, adam.huffman
no flags Details
syslog as requested (21.46 KB, text/plain)
2002-04-18 13:52 UTC, adam.huffman
no flags Details

Description adam.huffman 2002-04-09 17:05:59 UTC
Description of Problem:
If I logout of GNOME the resulting GDM screen is badly corrupted.  The mouse and
keyboard are both still active but I can't see the login window properly.

Version-Release number of selected component (if applicable):

How Reproducible:
every time

Steps to Reproduce:
1. login to GNOME
2. logout 

Actual Results:

screen corruption

Expected Results:

gdm screen appears normally

Additional Information:
Dell Latitude C800 with ATI Rage Mobility M4

Comment 1 adam.huffman 2002-04-09 17:06:31 UTC
Created attachment 52885 [details]
X log file

Comment 2 adam.huffman 2002-04-09 17:07:25 UTC
Created attachment 52886 [details]
X config file

Comment 3 adam.huffman 2002-04-09 17:09:16 UTC
There are lots of messages like this in the syslog:

kernel: [drm:r128_cce_indirect] *ERROR* process 1978 using buffer owned by 0

Comment 4 Mike A. Harris 2002-04-15 17:13:27 UTC
Please attach your syslog messages file from the start of kernel boot
up to and including these errors.

Comment 5 Alexei Podtelezhnikov 2002-04-15 22:27:33 UTC
Could this be because your confir specifies 1400X1050 as the resolution of your
"Monitor" while "Screen" has the diferent choices. As a matter of fact I've seen
a lot of strange problems (reported here earlier) on Dell C800 at the
resolutions other than 1400x1050 (that is Dewll-recommended).

What will happen if you specify "1400x1050" as the only one available in
"Screen" section of your config?

Comment 6 adam.huffman 2002-04-18 13:52:28 UTC
Created attachment 54325 [details]
syslog as requested

Comment 7 adam.huffman 2002-04-18 13:54:31 UTC
In the past I've tried to use 1400x1050 but it has never worked.  If I do put
that in the screen section I will get 1600x1200 instead.

Comment 8 bryan hunt 2002-04-23 07:52:57 UTC
 I know whats wrong ... I have the same laptop and the problem is caused by dma   
not being enabled .... here is the fix (from dri.sourceforge.net) troubleshoooting 
guide ... PS: 3d accelleration will not work until you reduce resolution to 1024X768 at 16 
bits .... once you do though your graphichs performance will double. Anyhow heres the fix  
... i have put the relevant part into an init script that starts at runlevel 3 ... gdm 
kicks in at runlevel 5  
9.1 Bus Mastering 
DMA-based DRI drivers (that's most DRI drivers) cannot function unless bus mastering is 
enabled for your graphics card. By default, some systems don't having bus mastering on. 
You should enable it in your BIOS.  
Alternately, you can check the status of bus mastering and change the setting from within 
Linux. There may be similar procedures for other operating systems.  
Run lspci (as root) and find the information describing your graphics adapter. For 
    00:00.0 Host bridge: Intel Corporation 440BX/ZX - 82443BX/ZX Host bridge (rev 03) 
    00:01.0 PCI bridge: Intel Corporation 440BX/ZX - 82443BX/ZX AGP bridge (rev 03) 
    00:07.0 ISA bridge: Intel Corporation 82371AB PIIX4 ISA (rev 02) 
    00:07.1 IDE interface: Intel Corporation 82371AB PIIX4 IDE (rev 01) 
    00:07.2 USB Controller: Intel Corporation 82371AB PIIX4 USB (rev 01) 
    00:07.3 Bridge: Intel Corporation 82371AB PIIX4 ACPI (rev 02) 
    00:11.0 Ethernet controller: Intel Corporation 82557 [Ethernet Pro 100] (rev 08) 
    00:12.0 SCSI storage controller: Symbios Logic Inc. (formerly NCR) 53c895 (rev 02) 
    00:14.0 Multimedia audio controller: Ensoniq ES1371 [AudioPCI-97] (rev 08) 
    01:00.0 VGA compatible controller: 3Dfx Interactive, Inc.: Unknown device 0009 (rev 
The bus, device, and function number comprise the device id, which is conventionally 
written in the form bus:dev.func, or in this case 01:00.0.  
Use the setpci command to examine bit two of register 4 for your graphics card. This will 
indicate whether or not bus mastering is enabled.  
        setpci -s 01:00.0 4.w 
A hexadecimal value will be printed. Convert the least significant digit to binary. For 
example, if you see 3, that's 0011 in binary (bit two is 0). If you see 7, that's 0111 in 
binary (bit two is 1). In the first example, bus mastering is disabled. It's enabled in 
the second example.  
The following shell script will enabled bus mastering for your graphics card and host 
bridge. Run it as root.  
    dev=01:00.0   # change as appropriate 
    echo Enabling bus mastering on device $dev 
    setpci -s $dev 4.w=$(printf %x $((0x$(setpci -s $dev 4.w)|4))) 
    echo Enabling bus mastering on host bridge $dev 
    setpci -s $dev 4.w=$(printf %x $((0x$(setpci -s $dev 4.w)|4))) 
You can check if this worked by running the first setpci command again.  

Comment 9 Gene Czarcinski 2002-04-25 13:32:29 UTC
I am having a similar problem on a desktop with a ATI Rage 128 RF/SG AGP video
card (32MB video RAM).

The system is a PIII SMP (running the UP kernel makes no difference).  I have
tried skipjack-2 with both no errata applied and with all errata applied ...
makes no difference.

It does NOT happen every time but does most of the time.

It does not matter if I log out of gnome or kde.  A ctrl-alt-backspace corrects
the screen.

Comment 10 Gene Czarcinski 2002-04-25 19:11:27 UTC
Well, it is not gdm related ... kdm suffers the same problem.

Comment 11 Gene Czarcinski 2002-04-25 19:16:29 UTC
Yes, this is the same problem -- I found the error message in /var/log/messages:
 kernel: [drm:r128_cce_indirect] *ERROR* process 1204 using buffer owned by 0

Comment 12 Mike A. Harris 2002-05-30 05:28:34 UTC
So do you find forcing bus mastering on does actually solve the problem?
If so, I can investigate having the video driver do this itself.

Comment 13 adam.huffman 2002-06-21 11:34:57 UTC
When I upgraded to the final 7.3 release, the problem got a lot better.  I can
now switch from X to the console and back without hanging the system.

However, I do get occasional display crashes when using Xine and/or mplayer.

Comment 14 Mike A. Harris 2002-06-22 06:08:37 UTC
Does "a lot better" mean that the problem no longer occurs and you
do not see gdm login corruption?

xine/mplayer issues should be filed in separate bug reports as that would
be considered a separate problem.

Comment 15 adam.huffman 2002-06-24 09:37:39 UTC
It means that I haven't seen the problem since installing 7.3

Comment 16 Mike A. Harris 2002-06-24 15:31:07 UTC
Ok, beauty.  Thats what I needed to know.  ;o)  Thanks for the update.

Closing as fixed in RHL 7.3 (CURRENTRELEASE)

Note You need to log in before you can comment on or make changes to this bug.