[Bug 94877] Radeon 7500: GPU lockup after starting Xorg

Mon Oct 10 23:18:28 UTC 2016

https://bugs.freedesktop.org/show_bug.cgi?id=94877

erhard_f at mailbox.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|unspecified                 |7.7 (2012.06)

--- Comment #13 from erhard_f at mailbox.org ---
Now as my problem in Bug 97987 is solved I am facing this bug with the radeon
driver. I use a r300 card, but the error messages are very similar so I decided
not to open a new bug. Unlike the original bug report I am running a current
4.8.1-kernel, xorg-server-1.18.4, xf86-video-ati-7.7.1 and mesa-12.0.3.

Having said that Bug 95017, Bug 91893, Bug 96519 look seem very similar.

Sometimes the GPU can reset itself, most of the time I get a hard lock. If I
use "radeon-agpmode=-1" as kernel parameter, X on my G5 holds out longer, but
sooner or later I get a hard lock too.

dmesg-output:
[  774.360780] radeon 0000:f0:10.0: ring 0 stalled for more than 10443msec
[  774.360790] radeon 0000:f0:10.0: GPU lockup (current fence id
0x0000000000000407 last fence id 0x000000000000040e on ring 0)
[  774.511407] Failed to wait GUI idle while programming pipes. Bad things
might happen.
[  774.512511] radeon 0000:f0:10.0: Saved 251 dwords of commands on ring 0.
[  774.512530] radeon 0000:f0:10.0: (r300_asic_reset:425)
RBBM_STATUS=0x80010140
[  775.012522] radeon 0000:f0:10.0: (r300_asic_reset:444)
RBBM_STATUS=0x80010140
[  775.508519] radeon 0000:f0:10.0: (r300_asic_reset:456)
RBBM_STATUS=0x00000140
[  775.508551] radeon 0000:f0:10.0: GPU reset succeed
[  775.508556] radeon 0000:f0:10.0: GPU reset succeeded, trying to resume
[  775.508593] [drm] radeon: 1 quad pipes, 1 Z pipes initialized.
[  775.508601] radeon 0000:f0:10.0: WB disabled
[  775.508607] radeon 0000:f0:10.0: fence driver on ring 0 use gpu addr
0x0000000000000000 and cpu addr 0xd00000000139d000
[  775.508699] [drm] radeon: ring at 0x0000000000001000
[  775.508769] [drm] ring test succeeded in 0 usecs
[  775.559042] [drm] ib test succeeded in 0 usecs

What I also got is some interesting output in Xorg.0.log via sshd from another
machine. Maybe this is of some help:

(EE) [mi] EQ overflowing.  Additional events will be discarded until existing
events are processed.
(EE) 
(EE) Backtrace:
(EE) 0: /usr/bin/X (xorg_backtrace+0x6c) [0x101db7fc]
(EE) 1: /usr/bin/X (mieqEnqueue+0x2d0) [0x101b8eb0]
(EE) 2: /usr/bin/X (QueuePointerEvents+0x94) [0x10053254]
(EE) 3: /usr/bin/X (xf86PostMotionEventM+0x2dc) [0x10097adc]
(EE) 4: /usr/lib/xorg/modules/input/evdev_drv.so (0xcf6c000+0x4dd0) [0xcf70dd0]
(EE) 5: /usr/lib/xorg/modules/input/evdev_drv.so (0xcf6c000+0x56ac) [0xcf716ac]
(EE) 6: /usr/lib/xorg/modules/input/evdev_drv.so (0xcf6c000+0x5960) [0xcf71960]
(EE) 7: /usr/bin/X (0x10000000+0x848cc) [0x100848cc]
(EE) 8: /usr/bin/X (0x10000000+0xb3328) [0x100b3328]
(EE) 9: linux-vdso32.so.1 (__kernel_sigtramp32+0x0) [0x1003f4]
(EE) 10: /lib/libc.so.6 (ioctl+0xe8) [0xf9478d8]
(EE) 11: ?? [0x5b]
(EE) 12: /usr/lib/libdrm.so.2 (drmIoctl+0x54) [0xfd15484]
(EE) 13: /usr/lib/libdrm.so.2 (drmCommandWrite+0x38) [0xfd19618]
(EE) 14: /usr/lib/libdrm_radeon.so.1 (0xee9d000+0x17d8) [0xee9e7d8]
(EE) 15: /usr/lib/libdrm_radeon.so.1 (0xee9d000+0x1a98) [0xee9ea98]
(EE) 16: /usr/lib/libdrm_radeon.so.1 (radeon_bo_map+0x20) [0xeea06f0]
(EE) 17: /usr/lib/xorg/modules/drivers/radeon_drv.so (0xeece000+0x2e1e4)
[0xeefc1e4]
(EE) 18: /usr/lib/xorg/modules/libexa.so (0xed68000+0x6130) [0xed6e130]
(EE) 19: /usr/lib/xorg/modules/libexa.so (0xed68000+0x8fd8) [0xed70fd8]
(EE) 20: /usr/lib/xorg/modules/libexa.so (0xed68000+0x137d0) [0xed7b7d0]
(EE) 21: /usr/lib/xorg/modules/libexa.so (0xed68000+0x10024) [0xed78024]
(EE) 22: /usr/bin/X (0x10000000+0x14fac8) [0x1014fac8]
(EE) 23: /usr/bin/X (CompositePicture+0x294) [0x1013c3f4]
(EE) 24: /usr/lib/xorg/modules/libexa.so (0xed68000+0x10ef0) [0xed78ef0]
(EE) 25: /usr/bin/X (CompositeTrapezoids+0xb8) [0x1013c6b8]
(EE) 26: /usr/bin/X (0x10000000+0x143330) [0x10143330]
(EE) 27: /usr/bin/X (0x10000000+0x13cb34) [0x1013cb34]
(EE) 28: /usr/bin/X (0x10000000+0x340cc) [0x100340cc]
(EE) 29: /usr/bin/X (0x10000000+0x38cac) [0x10038cac]
(EE) 30: /usr/bin/X (0x10000000+0x1e304) [0x1001e304]
(EE) 31: /lib/libc.so.6 (0xf855000+0x21414) [0xf876414]
(EE) 32: /lib/libc.so.6 (__libc_start_main+0xc0) [0xf8765d0]
(EE) 
(EE) [mi] These backtraces from mieqEnqueue may point to a culprit higher up
the stack.
(EE) [mi] mieq is *NOT* the cause.  It is a victim.
[   775.542] [mi] Increasing EQ size to 1024 to prevent dropped events.
[   775.544] [mi] EQ processing has resumed after 85 dropped events.
[   775.544] [mi] This may be caused by a misbehaving driver monopolizing the
server's resources.

If requested I can add full kernel config, Xorg.0.log and dmesg output. Or open
a new bug if more appropriate.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.x.org/archives/xorg-driver-ati/attachments/20161010/0d8ee1e6/attachment-0001.html>