Kernel BUG seemingly related to the ATI fglrx drivers

[Lakin Wecker]

I get the following Kernel BUG, which seems to be related to the FGLRX driver.
It occurs when trying to start X.
In searching around on Google I found the following link which relates to a fix
for a similar problem in the nvidia kernel module:
http://www.nvnews.net/vbulletin/showpost.php?p=768227&postcount=4

The relevant lines from /var/log/messages:

Dec 28 14:10:14 zedd kernel: [ 51.120681] ----------- [cut here ] ---------
[please bite here ] ---------
Dec 28 14:10:14 zedd kernel: [ 51.121234] CPU 0
Dec 28 14:10:14 zedd kernel: [ 51.121452] Modules linked in: rfcomm l2cap
bluetooth powernow_k8 cpufreq_userspace cpufreq_stats freq_table
cpufreq_powersave cpufreq_ondemand cpufreq_conservative fglrx video pcc_acpi
hotkey button battery container ac nls_iso8859_1 nls_cp437 vfat fat dm_mod
sr_mod sbp2 scsi_mod lp pcmcia nsc_ircc irda snd_atiixp snd_pcm_oss
snd_mixer_oss snd_atiixp_modem snd_ac97_codec snd_ac97_bus rtc crc_ccitt
parport_pc parport psmouse snd_pcm snd_timer snd soundcore pcspkr ohci1394
ieee1394 yenta_socket rsrc_nonstatic pcmcia_core serio_raw snd_page_alloc
ehci_hcd ohci_hcd tg3 usbcore evdev md_mod ext3 jbd ide_generic ide_cd cdrom
ide_disk generic atiixp thermal processor fan capability commoncap vga16fb
cfbcopyarea vgastate cfbimgblt cfbfillrect fbcon tileblit font bitblit softcursor
Dec 28 14:10:14 zedd kernel: [ 51.132511] Pid: 3895, comm: Xorg Tainted: P
B 2.6.15-8-amd64-generic #1
Dec 28 14:10:14 zedd kernel: [ 51.132680] RIP: 0010:[put_page+90/128]
<ffffffff8015cc1a>{put_page+90}
Dec 28 14:10:14 zedd kernel: [ 51.132977] RSP: 0018:ffff8100395e9de0 EFLAGS:
00210256
Dec 28 14:10:14 zedd kernel: [ 51.133217] RAX: 0000000000000000 RBX:
0000000000000020 RCX: ffffffff803ad580
Dec 28 14:10:14 zedd kernel: [ 51.133387] RDX: ffff810001c3c170 RSI:
ffffffffffffffff RDI: ffff810001c3c170
Dec 28 14:10:14 zedd kernel: [ 51.133558] RBP: ffff8100380369a8 R08:
8000000037ee2067 R09: 0000000000000000
Dec 28 14:10:14 zedd kernel: [ 51.133729] R10: 00002aaaabadda68 R11:
0000000000203246 R12: 00002aaaad736000
Dec 28 14:10:14 zedd kernel: [ 51.133899] R13: 00002aaaad735000 R14:
00000000000b6984 R15: ffff810001c3c170
Dec 28 14:10:14 zedd kernel: [ 51.134069] FS: 00002aaaabae0ae0(0000)
GS:ffffffff803ef800(0000) knlGS:000000005556db60
Dec 28 14:10:14 zedd kernel: [ 51.134285] CS: 0010 DS: 0000 ES: 0000 CR0:
000000008005003b
Dec 28 14:10:14 zedd kernel: [ 51.140861] CR2: 00002aaaaccd3878 CR3:
0000000039fc1000 CR4: 00000000000006e0
Dec 28 14:10:14 zedd kernel: [ 51.147484] Process Xorg (pid: 3895, threadinfo
ffff8100395e8000, task ffff81003b285160)
Dec 28 14:10:14 zedd kernel: [ 51.161271] Stack: ffffffff80161952
ffff81003e67d478 0000000000000000 ffff8100395e9ed8
Dec 28 14:10:14 zedd kernel: [ 51.168803] ffffffffffffffff
0000000000000000 ffff810038c83348 ffff8100395e9ee0
Dec 28 14:10:14 zedd kernel: [ 51.183554] 0000000100203296
0000000000000000
Dec 28 14:10:14 zedd kernel: [ 51.191013] Call
Trace:<ffffffff80161952>{unmap_vmas+1042} <ffffffff8016489e>{exit_mmap+94}
Dec 28 14:10:14 zedd kernel: [ 51.205581] <ffffffff8012de3b>{mmput+27}
<ffffffff801321a3>{do_exit+563}
Dec 28 14:10:14 zedd kernel: [ 51.220158]
<ffffffff80132ba9>{do_group_exit+169} <ffffffff8010e9aa>{system_call+126}
Dec 28 14:10:14 zedd kernel: [ 51.234646]
Dec 28 14:10:14 zedd kernel: [ 51.248371]
Dec 28 14:10:14 zedd kernel: [ 51.248372] Code: 0f 0b 68 59 ce 2f 80 c2 31 00
83 47 08 ff 0f 98 c0 84 c0 74
Dec 28 14:10:14 zedd kernel: [ 51.280090] <1>Fixing recursive fault but
reboot is needed!

[Lakin Wecker (lakin)]

I upgraded to the latest packages of about 4.5 hours ago, specifically the:
"2.6.15-16-amd64-generic" packages, including the fglrx modules for this kernel. It seems that after a clean install from a daily iso (2006/02/21), then updating to these packages and changing the driver in the xorg.conf to fglrx, 3D accelleration works just great, and no kernel bugs.

So, in my opinion, we can close this bug.

[Lakin Wecker (lakin)]

I'm changing this to a fix-released and milestone ubuntu-6.04, as I no longer have this problem, as of ubuntu dapper circa feb-21-2006 and no-one else has made any comments on this bug.

If someone else is still seeing it, feel free to add your comments, and re-open.