[linux-dvb] Re: budget-ci CI interface testing

["Bernhard Schrenk" <tv@clemy.org>]

> > I am working around this problem at the moment by disabling the error
> > handling in dvb_ca_en50221_set_configoption.
>
> Oh right, so it works, even though that fails? That is weird. Hmmm,
perhaps
> I'm not resetting it enough... If you have a moment, can you bung a printk
in
> to see what the value of configoption is when it fails?
>

On further tests I found out, that it has nothing todo with the second
driver load. Now it often fails at the first load. I have inserted your
suggested printk. The output is:

Apr 13 00:56:04 videobox kernel: Valid DVB CAM detected MANID:ffff DEVID:1
CONFIGBASE:0x1fe CONFIGOPTION:0xf
Apr 13 00:56:04 videobox kernel: dvb_ca_en50221_set_configoption
Apr 13 00:56:04 videobox kernel: configoption=0xff

It reads back 0xff from the config register.

> BTW: Just committed a fix for a stupid bug when reading packets; I've got
the
> transport layer protocol up and running now. I can't guarantee it'll work
for
> packets which require fragmentation, but I'll sort that once I've
implemented
> the higher layers.

It is now working quite well with normal packets, but fragmented packets
cause an endless loop in dvb_ca_en50221_io_read_condition. The bug is in
dvb_ringbuffer.c. The function dvb_ringbuffer_pkt_next does not return the
next packet when called with idx != 0. It is returning the same packet
again.

The communication with the card is now working well. I can use all options
of the card menu. I am looking forward into getting the transport stream
routed through the card.

Regards,
Bernhard

PS: To Andrew: Please excuse the direct mail to you instead to the list, it
was a mistake.



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