[pkg-nvidia-devel] Bug#566337: Verified on i386

Wed Feb 3 04:28:38 UTC 2010

I can verify this bug on i386 as well, with the following oops message.
At least the stacktrace looks similar enough that it seems to be the
same bug. Interestingly enough, though, this only happens on one of my
two cards, which I'm using to drive a TV output only. I don't know, yet,
if the difference is in the PCI slot, the card itself or the fact that
I'm only using it for TV output.

[  314.125863] NVRM: loading NVIDIA UNIX x86 Kernel Module  190.53  Tue
Dec  8 18:51:41 PST 2009
[  315.280384] BUG: unable to handle kernel NULL pointer dereference at
(null)
[  315.282984] IP: [<(null)>] (null)
[  315.282984] *pde = 00000000 
[  315.282984] Oops: 0000 [#1] SMP 
[  315.282984] last sysfs file: /sys/module/i2c_core/initstate
[  315.282984] Modules linked in: nvidia(P) agpgart binfmt_misc ppdev lp
sco bridge stp bnep rfcomm l2cap crc16 battery powernow_k8
cpufreq_powersave cpufreq_userspac]
[  315.282984] 
[  315.282984] Pid: 0, comm: swapper Tainted: P
(2.6.32-trunk-686 #1)  
[  315.282984] EIP: 0060:[<00000000>] EFLAGS: 00010292 CPU: 1
[  315.282984] EIP is at 0x0
[  315.282984] EAX: 005f88c0 EBX: 000000bc ECX: 00000017 EDX: f601a000
[  315.282984] ESI: 00110000 EDI: f64f5000 EBP: f3e0dc0c ESP: f6c79e08
[  315.282984]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
[  315.282984] Process swapper (pid: 0, ti=f6c78000 task=f6c52200
task.ti=f6c78000)
[  315.282984] Stack:
[  315.282984]  f96397f4 f601a000 f601b000 f3e18000 f64f5000 f95c2625
f64f5000 f601b000
[  315.282984] <0> 0000001a f64f7690 0000035c f95abd2e f64f5000 f601b000
f9cdf1b8 0000035c
[  315.282984] <0> 10f30044 00000000 f601b000 f64f5000 f974a1ef f64f5000
f601b000 0000035c
[  315.282984] Call Trace:
[  315.282984]  [<f96397f4>] ? _nv017798rm+0x24fa/0x372e [nvidia]
[  315.282984]  [<f95c2625>] ? _nv020473rm+0xe81/0x10b5 [nvidia]
[  315.282984]  [<f95abd2e>] ? _nv020312rm+0x57/0xb5 [nvidia]
[  315.282984]  [<f974a1ef>] ? _nv013614rm+0x4f0/0xad2 [nvidia]
[  315.282984]  [<f98a8056>] ? _nv008585rm+0x28a/0x4ab [nvidia]
[  315.282984]  [<f988f1e9>] ? _nv008605rm+0x4d/0x6b [nvidia]
[  315.282984]  [<f993e50c>] ? _nv003827rm+0x80/0xaf [nvidia]
[  315.282984]  [<f994390a>] ? rm_isr_bh+0x52/0x7d [nvidia]
[  315.282984]  [<f9a0f2cc>] ? nv_kern_isr_bh+0x41/0x46 [nvidia]
[  315.282984]  [<c1034b1e>] ? tasklet_action+0x67/0xad
[  315.282984]  [<c1035c80>] ? __do_softirq+0xaa/0x151
[  315.282984]  [<c1035d58>] ? do_softirq+0x31/0x3c
[  315.282984]  [<c1035e2e>] ? irq_exit+0x26/0x58
[  315.282984]  [<c1004c91>] ? do_IRQ+0x78/0x89
[  315.282984]  [<c10037f0>] ? common_interrupt+0x30/0x38
[  315.282984]  [<c101b0c0>] ? native_safe_halt+0x2/0x3
[  315.282984]  [<c1008f1f>] ? default_idle+0x3c/0x5a
[  315.282984]  [<c10092fe>] ? c1e_idle+0xd2/0xd5
[  315.282984]  [<c1002388>] ? cpu_idle+0x89/0xa5
[  315.282984] Code:  Bad EIP value.
[  315.282984] EIP: [<00000000>] 0x0 SS:ESP 0068:f6c79e08
[  315.282984] CR2: 0000000000000000
[  315.601210] ---[ end trace 665e1ee111194523 ]---
[  315.605892] Kernel panic - not syncing: Fatal exception in interrupt

Something very similar appears to happen with the current beta drivers
(195.30) as well:

[  139.273397] NVRM: loading NVIDIA UNIX x86 Kernel Module  195.30  Fri
Dec 18 13:34:49 PST 2009
[  140.501357] BUG: unable to handle kernel NULL pointer dereference at
(null)
[  140.504511] IP: [<(null)>] (null)
[  140.504511] *pde = 00000000 
[  140.504511] Oops: 0000 [#1] SMP 
[  140.504511] last sysfs file: /sys/module/i2c_core/initstate
[  140.504511] Modules linked in: nvidia(P) agpgart binfmt_misc ppdev lp
sco bridge stp bnep rfcomm l2cap crc16 battery powernow_k8
cpufreq_powersave cpufreq_userspac]
[  140.504511] 
[  140.504511] Pid: 0, comm: swapper Tainted: P
(2.6.32-trunk-686 #1)  
[  140.504511] EIP: 0060:[<00000000>] EFLAGS: 00010292 CPU: 1
[  140.504511] EIP is at 0x0
[  140.504511] EAX: 005f88c0 EBX: 000000bc ECX: 00000013 EDX: ef4cb01c
[  140.504511] ESI: f6067000 EDI: ef4cb000 EBP: ef5ddc14 ESP: f6c79e08
[  140.504511]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
[  140.504511] Process swapper (pid: 0, ti=f6c78000 task=f6c52200
task.ti=f6c78000)
[  140.504511] Stack:
[  140.504511]  f9740162 f6067000 ef691000 ef5e8000 ef4cb000 f96c56d8
ef4cb000 ef691000
[  140.504511] <0> 0000001a ef4cc690 0000035c f96aedf4 ef4cb000 ef691000
f9e20278 0000035c
[  140.504511] <0> 10f30044 00000000 ef691000 ef4cb000 f985de63 ef4cb000
ef691000 0000035c
[  140.504511] Call Trace:
[  140.504511]  [<f9740162>] ? _nv020422rm+0x22dc/0x346f [nvidia]
[  140.504511]  [<f96c56d8>] ? _nv023231rm+0xe81/0x10b5 [nvidia]
[  140.504511]  [<f96aedf4>] ? _nv023054rm+0x57/0xb3 [nvidia]
[  140.504511]  [<f985de63>] ? _nv015549rm+0x4f0/0xad2 [nvidia]
[  140.504511]  [<f998150e>] ? _nv009852rm+0x28a/0x4ab [nvidia]
[  140.504511]  [<f9967846>] ? _nv009874rm+0x92/0xb2 [nvidia]
[  140.504511]  [<f9a1cfc0>] ? _nv004485rm+0x80/0xaf [nvidia]
[  140.504511]  [<f9a223b2>] ? rm_isr_bh+0x52/0x7d [nvidia]
[  140.504511]  [<f9af6641>] ? nv_kern_isr_bh+0x42/0x47 [nvidia]
[  140.504511]  [<c1034b1e>] ? tasklet_action+0x67/0xad
[  140.504511]  [<c1035c80>] ? __do_softirq+0xaa/0x151
[  140.504511]  [<c1035d58>] ? do_softirq+0x31/0x3c
[  140.504511]  [<c1035e2e>] ? irq_exit+0x26/0x58
[  140.504511]  [<c1004c91>] ? do_IRQ+0x78/0x89
[  140.504511]  [<c10037f0>] ? common_interrupt+0x30/0x38
[  140.504511]  [<c101b0c0>] ? native_safe_halt+0x2/0x3
[  140.504511]  [<c1008f1f>] ? default_idle+0x3c/0x5a
[  140.504511]  [<c10092fe>] ? c1e_idle+0xd2/0xd5
[  140.504511]  [<c1002388>] ? cpu_idle+0x89/0xa5
[  140.504511] Code:  Bad EIP value.
[  140.504511] EIP: [<00000000>] 0x0 SS:ESP 0068:f6c79e08
[  140.504511] CR2: 0000000000000000
[  140.825890] ---[ end trace 69027cedbdef896b ]---
[  140.830564] Kernel panic - not syncing: Fatal exception in interrupt

Could the fact that this happens with both versions of the driver by any
chance indicate that the flaw is with the kernel configuration or
something? I upgraded the kernel and nvidia-kernel-source at the same
time, so I can't tell anymore. I cannot get the headers for the old
kernels anymore to try and rebuild the new drivers for it, but I'm going
to try it with a vanilla kernel as soon as it gets built (compiling it
as I'm writing this).

Fredrik Tolf