Mailing List Archive

On Sun, 2010-10-03 at 13:11 -0400, Dave Mason wrote:
> I sent this to ivtv-users, but no response yet, so I'm really hoping
> someone here can help.

Hi Dave,

I responded yesterday:

http://www.gossamer-threads.com/lists/ivtv/users/41116

Regards,
Andy

> Thanks!
>
> Begin forwarded message:

> > I have a PVR-350 card that I previously used in a different machine
> > which is now 2000 km away. I am now trying to use it on an Atom 410
> > machine. I had a problem for a while where if couldn't get an IRQ.
> > But that's stopped through no action on my part that I can
> > determine. (Weird :-)
> >
> > So the ivtv 1.3.0 driver loads fine, but if I try to cat /dev/video0
> > it just freezes - similarly if I run mythtvsetup and try to set up
> > the card.




_______________________________________________
ivtv-devel mailing list
ivtv-devel@ivtvdriver.org
http://ivtvdriver.org/mailman/listinfo/ivtv-devel

On 2010-Oct-3, at 14:13 , Andy Walls wrote:

> I responded yesterday:
>
> http://www.gossamer-threads.com/lists/ivtv/users/41116


Thanks, Andy. I was having some mailer problems, and missed it.

This is a hard machine freeze. Only hardware reset will get a
response.

I did the modprobe debug=0x7ff, then did a tail -f /var/log/kern.log &
and then cat /dev/video0 >/dev/null

I see (transcribing because the machine is frozen):

Preparing for firmware halt
Stopping VDM
Stopping AO
pinging (?) APU
Stopping VPU
Resetting Hw Blocks
Stopping SPU
init Encoder SDRAM pre-charge
init Encoder SDRAM refresh to 1us
init Decoder SDRAM pre-charge
init Decoder SDRAM refresh to 1us
Sleeping for 600 ms
....freeze...

I hit reset, and when it came back, when I did a modprobe I kept
getting that it couldn't get an IRQ (after varying amounts of time...
sometimes immediately after the modprobe, sometimes after messages
sent from tuner-simple, and everywhere in between), so I halted it and
power-cycled it, and then it installed properly, so I did cat /proc/
kmsg & cat /dev/video0 >/dev/null and saw:

Preparing for firmware halt
Loaded v4l-cx2341x-enc.fw firmware (376836 bytes)
Loading decode image
firmware: requesting v4l-cx2341x-dec.fw
Loaded v4l-cx2341x-dec.fw firmware (262144 byte)
....freeze....

Repeat, this time I realized that rsyslog might be eating some of the
messages, so I stopped it... I got the firmware: and the Loaded
messages for both firmware (enc and dec), then
...freeze...

Repeat this time in single-user, starting udev getting
firmware: requesting v4l-cx2341x-enc.fw
Loaded v4l-cx2341x-enc.fw firmware (376836 bytes)
firmware: requesting v4l-cx2341x-dec.fw
Loaded v4l-cx2341x-dec.fw firmware (262144 byte)
Encode revision: 0x02060039
Decode revision: 0x02020023
firmware: requesting v4l-cx2341x-init.mpg
Loaded v4l-cx2341x-init.mpg firmware (155648 byte)
...not frozen... I could ^C out of the cat, so I tried the cat again,
and got a couple of DMA timeouts, so I ^c again, and then tried a cat /
dev/video0 >video-capture, but the resulting file is 0 bytes long.

So I tried the modprobe -r ivtv;sleep 1;modprobe ivtv debug=0xffff and
then when I tried the cat I got
firmware: requesting v4l-cx2341x-enc.fw
Loaded v4l-cx2341x-enc.fw firmware (376836 bytes)
firmware: requesting v4l-cx2341x-dec.fw
Loaded v4l-cx2341x-dec.fw firmware (262144 byte)
Encoder firmware dead!
Retry loading firmware
firmware: requesting v4l-cx2341x-enc.fw
Loaded v4l-cx2341x-enc.fw firmware (376836 bytes)
firmware: requesting v4l-cx2341x-dec.fw
Loaded v4l-cx2341x-dec.fw firmware (262144 byte)
Encoder firmware dead!
Failed to initialize on minor 0
which seems to be repeatable with additional cat /dev/video0.

So I tried the reboot in single-user mode again and it couldn't get an
IRQ, but it worked on the next attempt. I got the first 4 lines from
last time, then ...freeze...

After a few more reboots for luck... one that couldn't get an IRQ and
the rest with the same 4 lines before freezing, I'm giving up for the
night.

Is my card fried in some way?

Is there some interaction with the Atom D410 that triggers a bug in
the driver? I had hyperthreads on for these experiments, but
previously when I turned it off, it didn't make any difference.

Is there something wrong with my firmware? The MD5 sums end in:
8883f3ecb68 dec.fw
2ef0c50ef48 enc.fw
4557353eacc init.mpg

Thanks for any guidance you may have!

../Dave


_______________________________________________
ivtv-devel mailing list
ivtv-devel@ivtvdriver.org
http://ivtvdriver.org/mailman/listinfo/ivtv-devel

On Sun, 2010-10-03 at 22:14 -0400, Dave Mason wrote:
> On 2010-Oct-3, at 14:13 , Andy Walls wrote:
>
> > I responded yesterday:
> >
> > http://www.gossamer-threads.com/lists/ivtv/users/41116
>
>
> Thanks, Andy. I was having some mailer problems, and missed it.
>
> This is a hard machine freeze. Only hardware reset will get a
> response.
>
> I did the modprobe debug=0x7ff, then did a tail -f /var/log/kern.log &
> and then cat /dev/video0 >/dev/null
>
> I see (transcribing because the machine is frozen):
>
> Preparing for firmware halt
> Stopping VDM
> Stopping AO
> pinging (?) APU
> Stopping VPU
> Resetting Hw Blocks
> Stopping SPU
> init Encoder SDRAM pre-charge
> init Encoder SDRAM refresh to 1us
> init Decoder SDRAM pre-charge
> init Decoder SDRAM refresh to 1us
> Sleeping for 600 ms
> ....freeze...
>
> I hit reset, and when it came back, when I did a modprobe I kept
> getting that it couldn't get an IRQ (after varying amounts of time...
> sometimes immediately after the modprobe, sometimes after messages
> sent from tuner-simple, and everywhere in between), so I halted it and
> power-cycled it, and then it installed properly, so I did cat /proc/
> kmsg & cat /dev/video0 >/dev/null and saw:
>
> Preparing for firmware halt
> Loaded v4l-cx2341x-enc.fw firmware (376836 bytes)
> Loading decode image
> firmware: requesting v4l-cx2341x-dec.fw
> Loaded v4l-cx2341x-dec.fw firmware (262144 byte)
> ....freeze....
>
> Repeat, this time I realized that rsyslog might be eating some of the
> messages, so I stopped it... I got the firmware: and the Loaded
> messages for both firmware (enc and dec), then
> ...freeze...
>
> Repeat this time in single-user, starting udev getting
> firmware: requesting v4l-cx2341x-enc.fw
> Loaded v4l-cx2341x-enc.fw firmware (376836 bytes)
> firmware: requesting v4l-cx2341x-dec.fw
> Loaded v4l-cx2341x-dec.fw firmware (262144 byte)
> Encode revision: 0x02060039
> Decode revision: 0x02020023
> firmware: requesting v4l-cx2341x-init.mpg
> Loaded v4l-cx2341x-init.mpg firmware (155648 byte)
> ...not frozen... I could ^C out of the cat, so I tried the cat again,
> and got a couple of DMA timeouts, so I ^c again, and then tried a cat /
> dev/video0 >video-capture, but the resulting file is 0 bytes long.
>
> So I tried the modprobe -r ivtv;sleep 1;modprobe ivtv debug=0xffff and
> then when I tried the cat I got
> firmware: requesting v4l-cx2341x-enc.fw
> Loaded v4l-cx2341x-enc.fw firmware (376836 bytes)
> firmware: requesting v4l-cx2341x-dec.fw
> Loaded v4l-cx2341x-dec.fw firmware (262144 byte)
> Encoder firmware dead!
> Retry loading firmware
> firmware: requesting v4l-cx2341x-enc.fw
> Loaded v4l-cx2341x-enc.fw firmware (376836 bytes)
> firmware: requesting v4l-cx2341x-dec.fw
> Loaded v4l-cx2341x-dec.fw firmware (262144 byte)
> Encoder firmware dead!
> Failed to initialize on minor 0
> which seems to be repeatable with additional cat /dev/video0.
>
> So I tried the reboot in single-user mode again and it couldn't get an
> IRQ, but it worked on the next attempt. I got the first 4 lines from
> last time, then ...freeze...
>
> After a few more reboots for luck... one that couldn't get an IRQ and
> the rest with the same 4 lines before freezing, I'm giving up for the
> night.
>
> Is my card fried in some way?
>
> Is there some interaction with the Atom D410 that triggers a bug in
> the driver? I had hyperthreads on for these experiments, but
> previously when I turned it off, it didn't make any difference.


Hi Dave,

My gut tells me that you have a power problem. Everything is (sort of)
fine until you really spin up the CX23416 to do anything useful.

Normally when the CX23416's operation is going to screw up a PCI
Southbridge chipset, its going to be a non-Intel chipset, and the PCI
chipset will normally do an NMI or force an immediate system reset.

Your first trouble shooting steps should be to try the board in Windows
and/or on a different machine. Eliminate that the board is defective.


Regards,
Andy

> Is there something wrong with my firmware? The MD5 sums end in:
> 8883f3ecb68 dec.fw
> 2ef0c50ef48 enc.fw
> 4557353eacc init.mpg
>
> Thanks for any guidance you may have!
>
> ../Dave



_______________________________________________
ivtv-devel mailing list
ivtv-devel@ivtvdriver.org
http://ivtvdriver.org/mailman/listinfo/ivtv-devel