FX 5200/ 1.0-8178 x86_64 U122205 / NMI Watchdog detected LOCKUP on gdm restart

[tru]

Hello,

I am running CentOS-3 (RHEL3 clone) on x86_64 with the latest nvidia driver
(8178 + U122205) on the latest kernel smp version 2.4.21-37.ELsmp.

lscpi:
05:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX 5200] (rev a1)

On GUI logout the nvidia splash screen is displayed but the workstation oops:
ping ok, but no remote access with ssh, nor console access (ctrl-alt-Fn) does
not yield a text console. The machine is barely alive (ping) and the num-lock key
can still be turned on/off, but that's about it. The only way to recover is reset/power button.

When the user only locks the display (xscreensaver), there is no issue, it's only upon
gdm restart -> maybe unloading/reloading the nvidia drivers causes the oops.

remote syslog output:

Jan 11 16:30:54 bobo kernel: NMI Watchdog detected LOCKUP on CPU1, rip ffffffff801242c5, registers:
Jan 11 16:30:54 bobo kernel: CPU 1
Jan 11 16:30:54 bobo kernel: Pid: 23850, comm: X Tainted: P
Jan 11 16:30:54 bobo kernel: RIP: 0010:[<ffffffff801242c5>]{.text.lock.fork+39}
Jan 11 16:30:54 bobo kernel: RSP: 0018:0000010004653e38 EFLAGS: 00000086
Jan 11 16:30:54 bobo kernel: RAX: 0000000000000002 RBX: 0000010002b738b8 RCX: ffffffff803f7240
Jan 11 16:30:54 bobo kernel: RDX: 0000010004652000 RSI: 0000010004653e88 RDI: 00000100263f4fd0
Jan 11 16:30:54 bobo kernel: RBP: 00000100263f4fd0 R08: 0000000000000000 R09: 000001007c54c1c0
Jan 11 16:30:54 bobo kernel: R10: 000001004000b010 R11: 00000100400018b0 R12: 0000010002abb000
Jan 11 16:30:54 bobo kernel: R13: 000000000000000a R14: 0000010004653e88 R15: 0000007fbfffc0e0
Jan 11 16:30:54 bobo kernel: FS: 0000002a95f7a4c0(0000) GS:ffffffff805e5e80(0000) knlGS:0000000046775bb0
Jan 11 16:30:54 bobo kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
Jan 11 16:30:54 bobo kernel: CR2: 0000002a971622c0 CR3: 00000000045a3000 CR4: 00000000000006a0
Jan 11 16:30:54 bobo kernel:
Jan 11 16:30:54 bobo kernel: Call Trace: <EOE> [<ffffffffa006cd0e>]{:usb-ohci:sohci_unlink_urb+542}
Jan 11 16:30:54 bobo kernel: [<ffffffffa005847a>]{:usbcore:usb_unlink_urb_Rsmp_ba76d609+42}
Jan 11 16:30:54 bobo kernel: [<ffffffffa0072456>]{:input:input_close_device_Rsmp_4ee160f7+22}
Jan 11 16:30:54 bobo kernel: [<ffffffffa007a3cc>]{:mousedev:mousedev_release+188}
Jan 11 16:30:54 bobo kernel: [<ffffffff80160863>]{__fput+83} [<ffffffff8015f116>]{filp_close+262}
Jan 11 16:30:54 bobo kernel: [<ffffffff8015f1b5>]{sys_close+133} [<ffffffff801102a7>]{system_call+119}
Jan 11 16:30:54 bobo kernel:
Jan 11 16:30:54 bobo kernel: Process X (pid: 23850, stackpage=10004653000)
Jan 11 16:30:54 bobo kernel: Stack: 0000010004653e38 0000000000000018
Jan 11 16:30:54 bobo kernel: Call Trace: <EOE> [<ffffffffa006cd0e>]{:usb-ohci:sohci_unlink_urb+542}
Jan 11 16:30:54 bobo kernel: [<ffffffffa005847a>]{:usbcore:usb_unlink_urb_Rsmp_ba76d609+42}
Jan 11 16:30:54 bobo kernel: [<ffffffffa0072456>]{:input:input_close_device_Rsmp_4ee160f7+22}
Jan 11 16:30:54 bobo kernel: [<ffffffffa007a3cc>]{:mousedev:mousedev_release+188}
Jan 11 16:30:54 bobo kernel: [<ffffffff80160863>]{__fput+83} [<ffffffff8015f116>]{filp_close+262}
Jan 11 16:30:54 bobo kernel: [<ffffffff8015f1b5>]{sys_close+133} [<ffffffff801102a7>]{system_call+119}
Jan 11 16:30:54 bobo kernel:
Jan 11 16:30:54 bobo kernel:
Jan 11 16:30:54 bobo kernel: Code: f3 90 7e f9 e9 8b e3 ff ff 80 3f 00 f3 90 7e f9 e9 e1 e3 ff
Jan 11 16:30:54 bobo kernel:
Jan 11 16:30:54 bobo kernel: console shuts up ...

Thanks for any idea.

Tru

[zander]

This looks like a known RHEL kernel bug, please see https://bugzilla.redhat.com/bugzilla....cgi?id=171129 for details.

[tru]

hello and thanks for the head up, I will enable/use the PS/2 mouse port, and report ASAP.

The 7th update is in beta now I will try it too.

Tru