[vdr] Re: unknwon picture type / error in data stream

[Thomas Breuer <tb@tb-electronic.de>]

On 01/07/2003 11:18 Christian Jacobsen wrote:

> I tried doing a recording on the system as it was still running and that
> one was broken too. After reloading DVB driver and VDR everything was OK.

Yes, reloading sometimes helps. But some time ago I made a test with
several timer recordings and a script that loaded/unloaded the driver
twice (with some delay in between). There were still many corrupted
recordings.

>>Same here, but I don't think it's a DVB driver problem. Since it does
>>not occur on all systems, I believe that it is a hardware dependant PCI
>>timing problem.
> 
> My system is a P3 933 with Intel Chipset (i810 i think) ...

Ok, it seems to clear now that it's not specifically related to the VIA
chipset.

> ...and latency timer
> is set to 128 (you have to do that after loading the driver as it sets it
> to 64) So it seems to be chipset independant.

Question: Does it make any difference if the latency is changed manually
with setpci *after* the DVB driver loaded or if it's done directly by
the driver at a specific time during the initialization of the cards? At
least there are some PCI related actions (IRQ setting) after the latency
is set by the driver:

...
kernel: saa7146_core.o: adjusting pci latency to 128
kernel: DVB: registering frontend 0:0 (Alps BSRV2)...
kernel: DVB: registering new adapter (Siemens/Technotrend/Hauppauge PCI
        rev1.3).
kernel: PCI: Found IRQ 11 for device 00:10.0
kernel: PCI: Sharing IRQ 11 with 00:07.2
kernel: PCI: Sharing IRQ 11 with 00:07.3
kernel: PCI: Sharing IRQ 11 with 00:0c.0
...

AFAIR I did not always succeed with setting the latency to 128 after the
driver had already set it to 64. That's why I patched the saa7146_core.c

Thomas






-- 
Info:
To unsubscribe send a mail to ecartis@linuxtv.org with "unsubscribe vdr" as subject.