reassign 595124 linux-2.6 forwarded 595124 https://bugzilla.kernel.org/show_bug.cgi?id=16140 found 595124 2.6.37-1~experimental.1 found 595124 2.6.32-30 thanks
Hi,
I've also seen this bug, also with pm-suspend.
~$ lspci -nn|grep VG
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon RV250
[Mobility FireGL 9000] [1002:4c66] (rev 02)
And with these two kernel images:
ii linux-image-2.6.32-5-686 2.6.32-30 Linux
2.6.32 for modern PCs
ii linux-image-2.6.37-trunk-686 2.6.37-1~experimental.1
Linux 2.6.37 for modern PC
But there is a workaround which makes resume work for ~90% of the cases:
turning agpmode off. To achieve that, I've put the following into
/etc/modprobe.d/radeon-kms.conf
options radeon modeset=1 agpmode=-1
Now resume "usually works". Sometimes resume doesnt work correctly (the screen
stays black, ssh still works, but the machine has to be rebooted to get X
working again), in this case dmesg output is like this:
[ 434.090204] PM: resume of devices complete after 3393.753 msecs
[ 434.090352] PM: Finishing wakeup.
[ 434.090355] Restarting tasks ... done.
[ 434.119551] video LNXVIDEO:00: Restoring backlight state
[ 434.172064] radeon 0000:01:00.0: GPU lockup CP stall for more than 158388msec
[ 434.172070] ------------[ cut here ]------------
[ 434.172104] WARNING: at
/build/buildd-linux-2.6_2.6.37-1~experimental.1-i386-rePuD0/linux-2.6-2.6.37/debian/
build/source_i386_none/drivers/gpu/drm/radeon/radeon_fence.c:244
radeon_fence_wait+0x1e9/0x26f [radeon]()
[ 434.172110] Hardware name: 18299MG
[ 434.172112] GPU lockup (waiting for 0x000051D7 last fence id 0x000051D4)
[ 434.172115] Modules linked in: acpi_cpufreq mperf cpufreq_userspace
cpufreq_stats cpufreq_conservative cpufr
eq_powersave ppdev lp sco bridge stp bnep rfcomm l2cap crc16 bluetooth
binfmt_misc uinput fuse firewire_sbp2 lo
op radeon snd_intel8x0m ttm thinkpad_acpi snd_intel8x0 snd_ac97_codec ac97_bus
pcmcia drm_kms_helper drm i2c_al
go_bit i2c_i801 ipw2200 snd_pcm_oss snd_mixer_oss snd_pcm yenta_socket
pcmcia_rsrc tpm_tis pcmcia_core joydev l
ibipw snd_seq_midi pcspkr tpm rng_core i2c_core battery ac snd_rawmidi tpm_bios
parport_pc power_supply nsc_irc
c evdev nvram parport snd_seq_midi_event processor snd_seq cfg80211 video
output snd_timer snd_seq_device serio
_raw psmouse snd rfkill lib80211 soundcore snd_page_alloc button irda shpchp
pci_hotplug crc_ccitt ext3 jbd mbc
ache sg sr_mod sd_mod cdrom crc_t10dif ata_generic ata_piix libata uhci_hcd
ehci_hcd scsi_mod usbcore firewire_
ohci thermal e1000 firewire_core floppy thermal_sys crc_itu_t nls_base [last
unloaded: scsi_wait_scan]
[ 434.172205] Pid: 1330, comm: Xorg Not tainted 2.6.37-trunk-686 #1
[ 434.172208] Call Trace:
[ 434.172218] [<c102f0c1>] ? warn_slowpath_common+0x6a/0x7b
[ 434.172236] [<f8efe51d>] ? radeon_fence_wait+0x1e9/0x26f [radeon]
[ 434.172241] [<c102f138>] ? warn_slowpath_fmt+0x28/0x2c
[ 434.172259] [<f8efe51d>] ? radeon_fence_wait+0x1e9/0x26f [radeon]
[ 434.172265] [<c10442e6>] ? autoremove_wake_function+0x0/0x29
[ 434.172274] [<f8dca5a0>] ? ttm_bo_wait+0xaa/0x132 [ttm]
[ 434.172293] [<f8f0eb10>] ? radeon_bo_wait+0x59/0x71 [radeon]
[ 434.172312] [<f8f0eb4f>] ? radeon_gem_wait_idle_ioctl+0x27/0x50 [radeon]
[ 434.172326] [<f8c5a3d5>] ? drm_ioctl+0x224/0x2d7 [drm]
[ 434.172344] [<f8f0eb28>] ? radeon_gem_wait_idle_ioctl+0x0/0x50 [radeon]
[ 434.172350] [<c1001eaf>] ? __switch_to+0xb0/0xe2
[ 434.172360] [<f8c5a1b1>] ? drm_ioctl+0x0/0x2d7 [drm]
[ 434.172365] [<c10c47cd>] ? do_vfs_ioctl+0x49e/0x4e9
[ 434.172372] [<c10a475c>] ? remove_vma+0x43/0x48
[ 434.172376] [<c10c485c>] ? sys_ioctl+0x44/0x64
[ 434.172379] [<c1002f9f>] ? sysenter_do_call+0x12/0x28
[ 434.172383] ---[ end trace dc067470c0ec9148 ]---
[ 434.420737] Failed to wait GUI idle while programming pipes. Bad things
might happen.
[ 434.421746] radeon 0000:01:00.0: (r100_asic_reset:2036)
RBBM_STATUS=0x80010140
[ 434.922771] radeon 0000:01:00.0: (r100_asic_reset:2057)
RBBM_STATUS=0x80010140
[ 435.419809] radeon 0000:01:00.0: (r100_asic_reset:2065)
RBBM_STATUS=0x00000140
[ 435.419830] radeon 0000:01:00.0: restoring config space at offset 0x1 (was
0x2b00383, writing 0x2b00387)
[ 435.419842] radeon 0000:01:00.0: GPU reset succeed
[ 435.419845] radeon 0000:01:00.0: GPU reset succeed
[ 435.452058] radeon 0000:01:00.0: WB enabled
[ 435.452136] [drm] radeon: ring at 0x00000000C0001000
[ 435.452157] [drm] ring test succeeded in 1 usecs
[ 435.452171] [drm] ib test succeeded in 2 usecs
[ 435.658037] ADDRCONF(NETDEV_UP): eth0: link is not ready
[ 435.660398] e1000: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control:
RX/TX
[ 435.660833] ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[ 435.661351] ADDRCONF(NETDEV_UP): wlan0: link is not ready
[ 446.368040] eth0: no IPv6 routers present
Just for completeness: with 2.6.32 (using agp) glxgears showed 2421 frames/s,
with 2.6.37 using agp 2685 FPS and without agp still 1800 FPS, with is fast
enough for youtube in fullscreen :-) Resume also worked with 2.6.32 and without
KMS, but then things were slooow. (830 FPS).
cheers,
Holger
signature.asc
Description: This is a digitally signed message part.
