Mailing List Archive: ivtv: users

Driver crash under load?

Index | Next | Previous | View Threaded

martyb at ix

Nov 8, 2007, 6:56 PM

Post #1 of 6 (1525 views)
Permalink

Driver crash under load?

I am using VIA/AMD64 chipset with kernel 2.6.23.1 (the same system mentioned
in my last posting about system instability). Based on a reply to my
last post about system instability, I have disabled DMA in the ivtv driver.
It is still to early to port a result regarding system hangs, but I am
with DMA disabled still seeing a problem that preceeded the DMA change:
the driver seems to "hang" on one of my two tuners, i.e. no video comes
out of the tuner, yet the user program just displays a black image with
no other indication of an error. The frequency of the failure seems to
be directly related to system (i.e. disk I/O) load.

Attempts to reload ivtv (using modprobe -r or rmmod -f) fail. It seems
that the only way to recover tuner function is a reboot.

_______________________________________________
ivtv-users mailing list
ivtv-users [at] ivtvdriver
http://ivtvdriver.org/mailman/listinfo/ivtv-users

gudlaugu at raunvis

Nov 8, 2007, 7:15 PM

Post #2 of 6 (1445 views)
Permalink

Re: Driver crash under load? [In reply to]

I had similar problems with a VIA/AMD64 chipset. I finally decided it was
easiest to switch mainboards. I went for a NVIDIA chipset which did not
let me down for about a year, when I stopped using the box for recording.

On Thu, 08 Nov 2007 18:56:55 -0800, Marty <martyb [at] ix> wrote:

> I am using VIA/AMD64 chipset with kernel 2.6.23.1 (the same system
> mentioned
> in my last posting about system instability). Based on a reply to my
> last post about system instability, I have disabled DMA in the ivtv
> driver.
> It is still to early to port a result regarding system hangs, but I am
> with DMA disabled still seeing a problem that preceeded the DMA change:
> the driver seems to "hang" on one of my two tuners, i.e. no video comes
> out of the tuner, yet the user program just displays a black image with
> no other indication of an error. The frequency of the failure seems to
> be directly related to system (i.e. disk I/O) load.
>
> Attempts to reload ivtv (using modprobe -r or rmmod -f) fail. It seems
> that the only way to recover tuner function is a reboot.
>
> _______________________________________________
> ivtv-users mailing list
> ivtv-users [at] ivtvdriver
> http://ivtvdriver.org/mailman/listinfo/ivtv-users

--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/

_______________________________________________
ivtv-users mailing list
ivtv-users [at] ivtvdriver
http://ivtvdriver.org/mailman/listinfo/ivtv-users

mcree at orcon

Nov 8, 2007, 7:23 PM

Post #3 of 6 (1441 views)
Permalink

Re: Driver crash under load? [In reply to]

On 9/11/2007, at 3:56 PM, Marty wrote:
> I am using VIA/AMD64 chipset with kernel 2.6.23.1 (the same system
> mentioned
> in my last posting about system instability). [snip]
> I have disabled DMA in the ivtv driver.
> [snip] I am
> with DMA disabled still seeing a problem that preceeded the DMA
> change:
> the driver seems to "hang" on one of my two tuners, i.e. no video
> comes
> out of the tuner,

I'm not using VIA/AMD64 (I'm running on an Alpha system),
nevertheless could not run ivtv from the 2.6.23.1 kernel as it lead
to i2c problems in the tda9887 and tuner modules, and zero output
from video capture. Upgrading to the v4l master repository fixed
these problems. Ivtv has always worked for me with prior kernel and
ivtv versions. It seems to me that the ivtv version in kernel 2.6.23
(likewise 2.6.23.1) has some serious problems.

Michael.

_______________________________________________
ivtv-users mailing list
ivtv-users [at] ivtvdriver
http://ivtvdriver.org/mailman/listinfo/ivtv-users

martyb at ix

Nov 8, 2007, 9:26 PM

Post #4 of 6 (1431 views)
Permalink

Re: Driver crash under load? [In reply to]

Marty wrote:
> I am using VIA/AMD64 chipset with kernel 2.6.23.1 (the same system mentioned
> in my last posting about system instability). Based on a reply to my
> last post about system instability, I have disabled DMA in the ivtv driver.
> It is still to early to port a result regarding system hangs, but I am
> with DMA disabled still seeing a problem that preceeded the DMA change:
> the driver seems to "hang" on one of my two tuners, i.e. no video comes
> out of the tuner, yet the user program just displays a black image with
> no other indication of an error. The frequency of the failure seems to
> be directly related to system (i.e. disk I/O) load.
>
> Attempts to reload ivtv (using modprobe -r or rmmod -f) fail. It seems
> that the only way to recover tuner function is a reboot.
>
> _______________________________________________
> ivtv-users mailing list
> ivtv-users [at] ivtvdriver
> http://ivtvdriver.org/mailman/listinfo/ivtv-users
>

I recently got this in syslog, but no crash, and even Mythtv kept running.

Nov 8 23:36:42 algernon kernel: ------------[ cut here ]------------
Nov 8 23:36:42 algernon kernel: kernel BUG at mm/slab.c:603!
Nov 8 23:36:42 algernon kernel: invalid opcode: 0000 [#2]
Nov 8 23:36:42 algernon kernel: SMP
Nov 8 23:36:42 algernon kernel: Modules linked in: wm8775 cx25840 ivtv cx2341x
lirc_mceusb2 lirc_dev
Nov 8 23:36:42 algernon kernel: CPU: 1
Nov 8 23:36:42 algernon kernel: EIP: 0060:[<c0166696>] Tainted: G D VLI
Nov 8 23:36:42 algernon kernel: EFLAGS: 00010002 (2.6.23.1 #7)
Nov 8 23:36:42 algernon kernel: EIP is at free_block+0x126/0x130
Nov 8 23:36:42 algernon kernel: eax: 8000082c ebx: f7960814 ecx: 00000000
edx: c1800000
Nov 8 23:36:42 algernon kernel: esi: f7960800 edi: 00004000 ebp: f7960814
esp: c2151f30
Nov 8 23:36:42 algernon kernel: ds: 007b es: 007b fs: 00d8 gs: 0000 ss: 0068
Nov 8 23:36:42 algernon kernel: Process events/1 (pid: 8, ti=c2150000
task=c213b030 task.ti=c2150000)
Nov 8 23:36:42 algernon kernel: Stack: c213b030 dd393b58 00004000 c225e0c0
00000000 f7960814 f7960800 00004000
Nov 8 23:36:42 algernon kernel: f79e0b00 c0166728 00000000 c225e0c0
f79e0b00 c225e0c0 c2015040 c2101244
Nov 8 23:36:42 algernon kernel: c0167647 00000000 00000000 c2101240
c2015040 c0167590 c0136192 00000003
Nov 8 23:36:42 algernon kernel: Call Trace:
Nov 8 23:36:42 algernon kernel: [<c0166728>] drain_array+0x88/0xc0
Nov 8 23:36:42 algernon kernel: [<c0167647>] cache_reap+0xb7/0x110
Nov 8 23:36:42 algernon kernel: [<c0167590>] cache_reap+0x0/0x110
Nov 8 23:36:42 algernon kernel: [<c0136192>] run_workqueue+0x72/0x100
Nov 8 23:36:42 algernon kernel: [<c0136b70>] worker_thread+0x0/0xb0
Nov 8 23:36:42 algernon kernel: [<c0136be9>] worker_thread+0x79/0xb0
Nov 8 23:36:42 algernon kernel: [<c0139ab0>] autoremove_wake_function+0x0/0x50
Nov 8 23:36:42 algernon kernel: [<c0136b70>] worker_thread+0x0/0xb0
Nov 8 23:36:42 algernon kernel: [<c01397f2>] kthread+0x42/0x70
Nov 8 23:36:42 algernon kernel: [<c01397b0>] kthread+0x0/0x70
Nov 8 23:36:42 algernon kernel: [<c01039af>] kernel_thread_helper+0x7/0x18
Nov 8 23:36:42 algernon kernel: =======================
Nov 8 23:36:42 algernon kernel: Code: 2b 42 20 89 f2 89 47 18 8b 44 24 0c e8 84
fe ff ff e9 2c ff ff ff 83 c4 14 5b 5e 5f 5d c3 8b 52
0c 8b 02 84 c0 0f 88 5d ff ff ff <0f> 0b eb fe 8d b6 00 00 00 00 83 ec 18 85 c9
89 74 24 0c 89 6c

_______________________________________________
ivtv-users mailing list
ivtv-users [at] ivtvdriver
http://ivtvdriver.org/mailman/listinfo/ivtv-users

martyb at ix

Nov 9, 2007, 3:50 PM

Post #5 of 6 (1410 views)
Permalink

Re: Driver crash under load? [In reply to]

Marty wrote:
> Marty wrote:
>> I am using VIA/AMD64 chipset with kernel 2.6.23.1 (the same system mentioned
>> in my last posting about system instability). Based on a reply to my
>> last post about system instability, I have disabled DMA in the ivtv driver.
>> It is still to early to port a result regarding system hangs, but I am
>> with DMA disabled still seeing a problem that preceeded the DMA change:
>> the driver seems to "hang" on one of my two tuners, i.e. no video comes
>> out of the tuner, yet the user program just displays a black image with
>> no other indication of an error. The frequency of the failure seems to
>> be directly related to system (i.e. disk I/O) load.
>>
>> Attempts to reload ivtv (using modprobe -r or rmmod -f) fail. It seems
>> that the only way to recover tuner function is a reboot.
>>
>> _______________________________________________
>> ivtv-users mailing list
>> ivtv-users [at] ivtvdriver
>> http://ivtvdriver.org/mailman/listinfo/ivtv-users
>>
>
> I recently got this in syslog, but no crash, and even Mythtv kept running.

Update: shortly after posting this, the system became unresponsive with the
(apparently) inaccessible disk problem I reported earlier, although I was able
to log in via ssh are execute the reboot command. (It didn't help, however,
because the mounted partitions were still not cleanly unmounted and required a
fsck check after the restart.) I think I've done all I can with 2.6.23.1, so
unless anyone wants me to experiment with different driver options, I will
revert to 2.6.18 for the time being.

>
> Nov 8 23:36:42 algernon kernel: ------------[ cut here ]------------
> Nov 8 23:36:42 algernon kernel: kernel BUG at mm/slab.c:603!
> Nov 8 23:36:42 algernon kernel: invalid opcode: 0000 [#2]
> Nov 8 23:36:42 algernon kernel: SMP
> Nov 8 23:36:42 algernon kernel: Modules linked in: wm8775 cx25840 ivtv cx2341x
> lirc_mceusb2 lirc_dev
> Nov 8 23:36:42 algernon kernel: CPU: 1
> Nov 8 23:36:42 algernon kernel: EIP: 0060:[<c0166696>] Tainted: G D VLI
> Nov 8 23:36:42 algernon kernel: EFLAGS: 00010002 (2.6.23.1 #7)
> Nov 8 23:36:42 algernon kernel: EIP is at free_block+0x126/0x130
> Nov 8 23:36:42 algernon kernel: eax: 8000082c ebx: f7960814 ecx: 00000000
> edx: c1800000
> Nov 8 23:36:42 algernon kernel: esi: f7960800 edi: 00004000 ebp: f7960814
> esp: c2151f30
> Nov 8 23:36:42 algernon kernel: ds: 007b es: 007b fs: 00d8 gs: 0000 ss: 0068
> Nov 8 23:36:42 algernon kernel: Process events/1 (pid: 8, ti=c2150000
> task=c213b030 task.ti=c2150000)
> Nov 8 23:36:42 algernon kernel: Stack: c213b030 dd393b58 00004000 c225e0c0
> 00000000 f7960814 f7960800 00004000
> Nov 8 23:36:42 algernon kernel: f79e0b00 c0166728 00000000 c225e0c0
> f79e0b00 c225e0c0 c2015040 c2101244
> Nov 8 23:36:42 algernon kernel: c0167647 00000000 00000000 c2101240
> c2015040 c0167590 c0136192 00000003
> Nov 8 23:36:42 algernon kernel: Call Trace:
> Nov 8 23:36:42 algernon kernel: [<c0166728>] drain_array+0x88/0xc0
> Nov 8 23:36:42 algernon kernel: [<c0167647>] cache_reap+0xb7/0x110
> Nov 8 23:36:42 algernon kernel: [<c0167590>] cache_reap+0x0/0x110
> Nov 8 23:36:42 algernon kernel: [<c0136192>] run_workqueue+0x72/0x100
> Nov 8 23:36:42 algernon kernel: [<c0136b70>] worker_thread+0x0/0xb0
> Nov 8 23:36:42 algernon kernel: [<c0136be9>] worker_thread+0x79/0xb0
> Nov 8 23:36:42 algernon kernel: [<c0139ab0>] autoremove_wake_function+0x0/0x50
> Nov 8 23:36:42 algernon kernel: [<c0136b70>] worker_thread+0x0/0xb0
> Nov 8 23:36:42 algernon kernel: [<c01397f2>] kthread+0x42/0x70
> Nov 8 23:36:42 algernon kernel: [<c01397b0>] kthread+0x0/0x70
> Nov 8 23:36:42 algernon kernel: [<c01039af>] kernel_thread_helper+0x7/0x18
> Nov 8 23:36:42 algernon kernel: =======================
> Nov 8 23:36:42 algernon kernel: Code: 2b 42 20 89 f2 89 47 18 8b 44 24 0c e8 84
> fe ff ff e9 2c ff ff ff 83 c4 14 5b 5e 5f 5d c3 8b 52
> 0c 8b 02 84 c0 0f 88 5d ff ff ff <0f> 0b eb fe 8d b6 00 00 00 00 83 ec 18 85 c9
> 89 74 24 0c 89 6c
>
> _______________________________________________
> ivtv-users mailing list
> ivtv-users [at] ivtvdriver
> http://ivtvdriver.org/mailman/listinfo/ivtv-users
>

_______________________________________________
ivtv-users mailing list
ivtv-users [at] ivtvdriver
http://ivtvdriver.org/mailman/listinfo/ivtv-users

martyb at ix

Nov 10, 2007, 6:29 PM

Post #6 of 6 (1445 views)
Permalink

Re: Driver crash under load? [In reply to]

Marty wrote:
I think I've done all I can with 2.6.23.1, so
> unless anyone wants me to experiment with different driver options, I will
> revert to 2.6.18 for the time being.

Having reverted and paid a visit to the driver website, I just noticed that I
was using an old driver, 0.10.5. Sorry for any inconvenience or confusion this
may have caused to the developers. Indeed, a DMA bug was among the items fixed
in the update. I will run this for a while and report the results in a few days.

_______________________________________________
ivtv-users mailing list
ivtv-users [at] ivtvdriver
http://ivtvdriver.org/mailman/listinfo/ivtv-users

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads