Arch Linux

edta

Member

Registered: 2011-10-03

Posts: 2

GPU lockup using xf86-video-ati and gnome 3.2

hi everybody, i'm getting a GPU lockup on a fresh install running gnome 3.2 using the open source ati driver.

/var/log/kernel.log tells me this:

Oct  3 17:27:45 myhost kernel: [  232.136663] eth0: no IPv6 routers present
Oct  3 17:27:55 myhost kernel: [  241.573334] radeon 0000:01:05.0: GPU lockup CP stall for more than 10000msec
Oct  3 17:27:55 myhost kernel: [  241.575438] ------------[ cut here ]------------
Oct  3 17:27:55 myhost kernel: [  241.575451] WARNING: at drivers/gpu/drm/radeon/radeon_fence.c:267 radeon_fence_wait+0x443/0x480 [radeon]()
Oct  3 17:27:55 myhost kernel: [  241.575453] Hardware name: TA790GX A3+
Oct  3 17:27:55 myhost kernel: [  241.575454] GPU lockup (waiting for 0x00000636 last fence id 0x00000634)
Oct  3 17:27:55 myhost kernel: [  241.575455] Modules linked in: nls_utf8 udf ipv6 fuse ext2 snd_hda_codec_hdmi snd_hda_codec_realtek radeon ttm firewire_ohci drm_kms_helper firewire_core snd_hda_intel drm snd_hda_codec edac_core btusb sg bluetooth i2c_algo_bit shpchp r8169 evdev rfkill snd_hwdep snd_pcm snd_timer edac_mce_amd snd pci_hotplug i2c_piix4 soundcore k10temp crc_itu_t i2c_core wmi button thermal mii sp5100_tco snd_page_alloc psmouse floppy pcspkr serio_raw processor ext4 mbcache jbd2 crc16 usbhid hid sr_mod cdrom sd_mod pata_acpi ohci_hcd ahci libahci pata_atiixp libata ehci_hcd scsi_mod usbcore
Oct  3 17:27:55 myhost kernel: [  241.575488] Pid: 1296, comm: gnome-shell Not tainted 3.0-ARCH #1
Oct  3 17:27:55 myhost kernel: [  241.575490] Call Trace:
Oct  3 17:27:55 myhost kernel: [  241.575496]  [<ffffffff8105c7ef>] warn_slowpath_common+0x7f/0xc0
Oct  3 17:27:55 myhost kernel: [  241.575499]  [<ffffffff8105c8e6>] warn_slowpath_fmt+0x46/0x50
Oct  3 17:27:55 myhost kernel: [  241.575506]  [<ffffffffa03a6ff3>] radeon_fence_wait+0x443/0x480 [radeon]
Oct  3 17:27:55 myhost kernel: [  241.575509]  [<ffffffff8107f7f0>] ? abort_exclusive_wait+0xb0/0xb0
Oct  3 17:27:55 myhost kernel: [  241.575517]  [<ffffffffa03a7691>] radeon_sync_obj_wait+0x11/0x20 [radeon]
Oct  3 17:27:55 myhost kernel: [  241.575520]  [<ffffffffa02a2c40>] ttm_bo_wait+0xb0/0x1c0 [ttm]
Oct  3 17:27:55 myhost kernel: [  241.575524]  [<ffffffffa02a38a9>] ? ttm_bo_list_ref_sub+0x29/0x30 [ttm]
Oct  3 17:27:55 myhost kernel: [  241.575532]  [<ffffffffa03c145d>] radeon_gem_wait_idle_ioctl+0x8d/0x110 [radeon]
Oct  3 17:27:55 myhost kernel: [  241.575537]  [<ffffffffa034c444>] drm_ioctl+0x3e4/0x4c0 [drm]
Oct  3 17:27:55 myhost kernel: [  241.575545]  [<ffffffffa03c13d0>] ? radeon_gem_busy_ioctl+0x140/0x140 [radeon]
Oct  3 17:27:55 myhost kernel: [  241.575549]  [<ffffffff8116951f>] do_vfs_ioctl+0x8f/0x500
Oct  3 17:27:55 myhost kernel: [  241.575552]  [<ffffffff811575d1>] ? rw_verify_area+0x61/0xf0
Oct  3 17:27:55 myhost kernel: [  241.575554]  [<ffffffff81159002>] ? fget_light+0x92/0x100
Oct  3 17:27:55 myhost kernel: [  241.575556]  [<ffffffff81169a21>] sys_ioctl+0x91/0xa0
Oct  3 17:27:55 myhost kernel: [  241.575559]  [<ffffffff813f4a42>] system_call_fastpath+0x16/0x1b
Oct  3 17:27:55 myhost kernel: [  241.575561] ---[ end trace ab845bf5c8b6dc17 ]---
Oct  3 17:27:55 myhost kernel: [  241.576629] radeon 0000:01:05.0: GPU softreset 
Oct  3 17:27:55 myhost kernel: [  241.576631] radeon 0000:01:05.0:   R_008010_GRBM_STATUS=0xA40034AC
Oct  3 17:27:55 myhost kernel: [  241.576633] radeon 0000:01:05.0:   R_008014_GRBM_STATUS2=0x00010003
Oct  3 17:27:55 myhost kernel: [  241.576634] radeon 0000:01:05.0:   R_000E50_SRBM_STATUS=0x20001040
Oct  3 17:27:55 myhost kernel: [  241.576655] radeon 0000:01:05.0:   R_008020_GRBM_SOFT_RESET=0x00007FEE
Oct  3 17:27:55 myhost kernel: [  241.591655] radeon 0000:01:05.0: R_008020_GRBM_SOFT_RESET=0x00000001
Oct  3 17:27:55 myhost kernel: [  241.607660] radeon 0000:01:05.0:   R_008010_GRBM_STATUS=0xA0003030
Oct  3 17:27:55 myhost kernel: [  241.607662] radeon 0000:01:05.0:   R_008014_GRBM_STATUS2=0x00000003
Oct  3 17:27:55 myhost kernel: [  241.607663] radeon 0000:01:05.0:   R_000E50_SRBM_STATUS=0x20008040
Oct  3 17:27:55 myhost kernel: [  241.608665] radeon 0000:01:05.0: GPU reset succeed
Oct  3 17:27:55 myhost kernel: [  241.629030] radeon 0000:01:05.0: WB enabled
Oct  3 17:27:55 myhost kernel: [  241.660958] [drm] ring test succeeded in 0 usecs
Oct  3 17:27:55 myhost kernel: [  241.660965] [drm] ib test succeeded in 1 usecs

The screen locks, but the mouse still works for about 5 seconds before becoming inactive too. i managed to break out of it by changing to a virtual terminal and shutting down the x server while the mouse was still responsive.

i'm using xf86-video-ati 6-14-2.1 as found in extra currently.

Xfce runs fine, so i'm using that at the moment.
I think the new gnome doesn't play nicely with xf86-video-ati. i can't find anything about a recent problem on the gnome or the radeon bugtracker.

Any suggestions would be very appreciated

---

edited for more info!

Last edited by edta (2011-10-03 16:24:20)

biltong

Member

From: South Africa

Registered: 2011-09-01

Posts: 118

Re: GPU lockup using xf86-video-ati and gnome 3.2

Perhaps it's a specific issue with your card series? What card do you have? Your hardware may also have issues - hear me out. A friend of mine bought an nvidia GT220 last year for is baby gaming rig. It managed to run everything we used to play, except for call of duty 5, it would always lock up the machine right after filling the screen with artifacts. But, every other game we had at the time, including call of duty 4, worked fine for extended periods of time. I told him to take the card back, so he did, and hey presto we can now play call of duty 5 with him just fine

edta

Member

Registered: 2011-10-03

Posts: 2

Re: GPU lockup using xf86-video-ati and gnome 3.2

Mm, I suppose it could be. I'm using

01:05.0 VGA compatible controller: ATI Technologies Inc Radeon HD 3300 Graphics (prog-if 00 [VGA controller])
	Subsystem: Biostar Microtech Int'l Corp Device 0217
	Flags: bus master, fast devsel, latency 0, IRQ 18
	Memory at d0000000 (32-bit, prefetchable) [size=256M]
	I/O ports at d000 [size=256]
	Memory at fe9f0000 (32-bit, non-prefetchable) [size=64K]
	Memory at fe800000 (32-bit, non-prefetchable) [size=1M]
	Expansion ROM at <unassigned> [disabled]
	Capabilities: <access denied>
	Kernel driver in use: fglrx_pci
	Kernel modules: fglrx, radeon

on this machine (lspci -v from my linux mint installation running gnome 2.*). it didn't have a problem with the open source drivers on gnome 3.0, nor fglrx on gnome 2.* and kde4 works great to. strange ...

i'll try out some stuff from testing.

Last edited by edta (2011-10-04 09:11:50)

Finn10111

Member

Registered: 2011-10-05

Posts: 7

Re: GPU lockup using xf86-video-ati and gnome 3.2

The screen locks, but the mouse still works for about 5 seconds before becoming inactive too. i managed to break out of it by changing to a virtual terminal and shutting down the x server while the mouse was still responsive.

i'm using xf86-video-ati 6-14-2.1 as found in extra currently.

Hello,

I have the same problem you described and also I am using the same driver. The only different is that I had this problem also with Gnome 3.0. I have an AMD Mobility HD 2600 graphic card in my notebook. Unfortunetally I could not find some similar errors in my kernel.log but I will have a closer look the next time this lock happens. I am using the free driver for a few months now and before that I never had such problems with catalyst. The strange thing is how often this happens, sometimes 3 times a day and sometimes I have no problems for weeks.

magillos

Member

Registered: 2010-07-03

Posts: 40

Re: GPU lockup using xf86-video-ati and gnome 3.2

Perhaps try to install xf86-video-ati-git from AUR instead of stock kernel radeon driver. The latest seems to fix that issue for me.

ergh.. Forget. I still happens.

Last edited by magillos (2011-10-30 21:16:04)