Panic in radeon_get_vblank_counter()

[Robert Noland]

On Fri, 2009-03-13 at 23:33 -0500, Sean C. Farley wrote: 
> On Fri, 13 Mar 2009, Robert Noland wrote:
> 
> > On Fri, 2009-03-13 at 15:49 -0500, Sean C. Farley wrote:
> >> After upgrading my laptop (Dell 600m) yesterday, I have begun to 
> >> experience panics on it probably related to the DRM update.
> >>
> >> Killing the X server by Ctrl-Alt-Backspace is a good way to lock the 
> >> system.  Reboot to get a core dump.
> >>
> >> Piece of Xorg.log concerning the hardware using the
> >> xf86-video-ati-6.11.0 driver:
> >> (--) PCI:*(0 at 1:0:0) ATI Technologies Inc Radeon RV250 [Mobility FireGL 9000] rev 1, Mem @ 0xe8000000/0, 0xfcff0000/0, I/O @ 0x0000c000/0, BIOS @ 0x????????/65536
> >> ...
> >> (--) RADEON(0): Chipset: "ATI Radeon Mobility 9000 (M9) Lf (AGP)" (ChipID = 0x4c66)
> >>
> >>
> >> Special options in xorg.conf:
> >> Section "Device"
> >>      ...
> >>      Option     "EnablePageFlip"             "true"
> >>      Option     "AccelMethod"                "EXA"
> >> EndSection
> >>
> >> Section "Screen"
> >>      ...
> >>      Option "AddARGBGLXVisuals" "True"
> >>      Option "DisableGLXRootClipping" "True"
> >>      Option "XAANoOffscreenPixmaps" "True"
> >>      ...
> >> EndSection
> >>
> >>
> >> pciconf -lv of the device:
> >> vgapci0 at pci0:1:0:0:     class=0x030000 card=0x011e1028 chip=0x4c661002 rev=0x01 hdr=0x00
> >>      vendor     = 'ATI Technologies Inc'
> >>      device     = 'ATI MOBILITY RADEON 9000 (Microsoft Corporation - Radeon Mobility M9'
> >>      class      = display
> >>      subclass   = VGA
> >>
> >>
> >> /var/log/messages about drm:
> >> Mar 13 15:41:51 baba kernel: drm0: <ATI Radeon Lf RV250 Mobility 9000 M9 / FireMV 2400 PCI> on vgapci0
> >> Mar 13 15:41:51 baba kernel: vgapci0: child drm0 requested pci_enable_busmaster
> >> Mar 13 15:41:51 baba kernel: info: [drm] AGP at 0xe0000000 128MB
> >> Mar 13 15:41:51 baba kernel: info: [drm] Initialized radeon 1.29.0 20080613
> >> Mar 13 15:41:52 baba kernel: info: [drm] Setting GART location based on new memory map
> >> Mar 13 15:41:52 baba kernel: info: [drm] Loading R200 Microcode
> >> Mar 13 15:41:52 baba kernel: info: [drm] writeback test succeeded in 1 usecs
> >> Mar 13 15:41:52 baba kernel: drm0: [ITHREAD]
> >
> > Ok, I'll try to look at this soon... It would be useful if you could 
> > enable hw.dri.0.debug=1 and send me the debug output as well, if you 
> > can get anything useful before it locks up.
> 
> I have not had luck reproducing a panic, but I can get the computer to 
> lock up at startup of the X server (rarely) and shutdown (often).
> 
> This may not be accurate, but I think I am able to shutdown safely if I 
> switch to a tty and wait until this debug line is printed:
> [drm:pid13:vblank_disable_fn] vblank_disable_allowed=1

Actually, vblank_disable_allowed should be 0 when you have switched out
to a vt..

> If I start rebooting before it is printed, the system locks up.  Of 
> course, this is only after rebooting several times.
> 
> Here is a successful start and shutdown:
> http://people.freebsd.org/~scf/drm-dmesg.log
> http://people.freebsd.org/~scf/Xorg.0.log

Ok, I'll spend some time staring at the current code... Thanks for the
backtrace too, it's nice to get those...

robert.

> Does that help?
> 
> Sean
-- 
Robert Noland <rnoland at FreeBSD.org>
FreeBSD
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: This is a digitally signed message part
Url : http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20090314/c2568820/attachment.pgp