On 28/06/07, Matthias Scheler <tron@netbsd.org> wrote:
>
>         Hello,
>
> this e-mail is very long because of all the attachment.

Apologies. Never checked the Xorg log size.

http://www.spidersweb.co.uk/~netbsd/ci4ic4-radeon-dri-logs.tar.bz2


> Would it be possible for you to upload the attachments to web server
> and re-send the e-mail with download URLs instead?

Copying and pasting the original mail.

...

Hi,

I have been having problems getting DRM work on my office NetBSD
system; I recently managed to dig up a:
...
% dmesg | grep -i radeon
vga1 at pci1 dev 0 function 0: ATI Technologies Radeon 9800SE AH (rev. 0x00)
radeondrm0 at vga1
info: [drm] ATI Radeon AH 9800 SE (unit 0)
info: [drm] Initialized radeon 1.25.0 20060524
radeondrm0: interrupting at ioapic0 pin 16 (irq 11)
...

card out of oblivion and repeat the stuff I did on my laptop just for
fun. I got quickly to the above stage - kernel with DRM,
modular-xorg-server-1.3.0.0nb2 and the necessary bits to get Mesa-DRI
653 working. glxinfo reports direct rendering, so that seems OK.
Running glxgears, though, hangs the Xorg server; the mouse cursor is
still movable around, but nothing could be sent the the server itself
- no CtrlAltBackspace, no even CtrlAltF[1234]; at the same time I can
ssh to the box and see the state. Top reports:

% top | egrep "glx|drm"
 493 root      64    0  8280K 6548K CPU/1      9:21 99.02% 99.02% glxgears
 859 root       0    0  6320K   15M drmlk2/0   0:01  0.00%  0.00% Xorg

so it looks like some locking issue in this context; if I run 'gdb
Xorg Pid' and trace, I get:

(gdb) bt
#0  0xbba82a6f in ioctl () from /usr/lib/libc.so.12
#1  0xbb950749 in drmGetLock () from /usr/pkg/lib/libdrm.so.2
#2  0xbb9595ab in DRILock ()
  from /usr/pkg/lib/xorg/modules/extensions//libdri.so
#3  0xbb9595fc in DRIDoWakeupHandler ()
  from /usr/pkg/lib/xorg/modules/extensions//libdri.so
#4  0xbb958720 in DRIWakeupHandler ()
  from /usr/pkg/lib/xorg/modules/extensions//libdri.so
#5  0x08086f59 in WakeupHandler ()
#6  0x081855bd in WaitForSomething ()
#7  0x080834ee in Dispatch ()
#8  0x0806d794 in main ()
(gdb)


Any pointers? How to debug this further; actualy, should the Radeon SE
9800 card be supported by the current r300 driver at all?

I've attached the relevant files as well - if they are of any use -
xorg.conf, the output of glxinfo and the Xorg log file.

BTW even without starting glxgears on this one, when I leave the X
server, the console font is very dim, almost invisible.

This is on -current as of 27/06/2007.




>
>         Kind regards
>
...
>
> --
> Matthias Scheler                                  http://zhadum.org.uk/
>
>

Chavdar Ivanov