[linux-dvb] Multiplex dependent mpeg corruption and AverTV A800 with latest driver and 2.6.19-rc1

Klaus Frahm frahm at irsamc.ups-tlse.fr
Sat Oct 14 23:29:56 CEST 2006


I have recently tried the kernel version 2.6.19-rc1 which contains the 
latest (or at least very recent) drivers from the v4l official tree and 
in particular I was hoping, that the patch to solve cold-warm state 
problem of the AverTV A800 (discussed here some months ago) was finally 
incorporated in the kernel. This is indeed the case and the tuner is 
correctly initialized. However, now there is a new strange problem. When 
I am looking certains DVB-T channels with VLC I get the following kind 
of errors:

------------------------------
libdvbpsi error (PSI decoder): TS discontinuity (received 2, expected 1) 
for PID  100
libdvbpsi error (misc PSI): Bad CRC_32 (0x550f8f94) !!!
libdvbpsi error (PSI decoder): TS discontinuity (received 1, expected 0) 
for PID  0
libdvbpsi error (PSI decoder): TS discontinuity (received 7, expected 6) 
for PID  18
libdvbpsi error (PSI decoder): TS discontinuity (received 8, expected 7) 
for PID  0
libdvbpsi error (PSI decoder): TS discontinuity (received 2, expected 1) 
for PID  18
libdvbpsi error (misc PSI): Bad CRC_32 (0x386ee15b) !!!
libdvbpsi error (misc PSI): Bad CRC_32 (0xdef179a0) !!!
libdvbpsi error (PSI decoder): PSI section too long
libdvbpsi error (PSI decoder): TS discontinuity (received 2, expected 1) 
for PID  0
libdvbpsi error (misc PSI): Bad CRC_32 (0x80664799) !!!
libdvbpsi error (misc PSI): Bad CRC_32 (0x0136247e) !!!
libdvbpsi error (PSI decoder): TS discontinuity (received 7, expected 6) 
for PID  0
libdvbpsi error (PSI decoder): TS discontinuity (received 12, expected 
11) for PID 100
----------------------------

The errors with "TS discontinuity" are not serious and I have always a 
modest number of them, however the Bad CRC_32 errors are new and very 
serious, in particular the sound and video are clearly corrupted, with 
very strong perturbations and the strang thing is that it happens to the 
channels of two particular multiplexes while the other three multiplexes 
are okay.
I insist that with an "older" v4l-driver (for exemple that a August 12) 
or with the kernel 2.6.18 or 2.6.17.x this problem does not at all 
appear (provided I apply the cold-warm-state patch for the file a800.c). 
I have also tested on two different machines (a PIII 500 MHz with USB 
1.1 and a Pentium D with 2.8 MHz and USB 2.0) and the probleme is 
identical on both of them.
I have also a very good, an excellant DVB-T reception with my antenna. 
This should and can not be the problem that one or two multiplexes have 
simply a bad antenna reception.

I have the following DVB modules loaded:

dvb_usb_a800           23940  0
dvb_usb_dibusb_common    26756  1 dvb_usb_a800
mt2060                 22148  1 dvb_usb_dibusb_common
dib3000mc              29060  4294967295 dvb_usb_dibusb_common
dibx000_common         20996  1 dib3000mc
dvb_usb                39944  2 dvb_usb_a800,dvb_usb_dibusb_common
dvb_core              101040  1 dvb_usb
firmware_class         27904  1 dvb_usb
dvb_pll                32644  0 dvb_usb_dibusb_common,dvb_usb

I do not know in which exact module the bug/problem appears, maybe in 
dvb_usb_dibusb_common, dib3000mc or dibx000_common ?? The module 
dvb_usb_a800 has also a little changed (a part from this cold-warm-state 
patch) and may be a candidate.

For the moment being I am using the kernel 2.6.19-rc1 but with the 
"older" v4l-dvb drivers from 12 August 2006 which works very well).

Another observation: the scan utility which creates "channels.conf" 
gives for the "bad" multiplexes double entries like:
TF1:730167000:INVERSION_AUTO:BANDWIDTH_8_MHZ:FEC_2_3:FEC_AUTO:QAM_64:TRANSMISSION_MODE_8K:GUARD_INTERVAL_1_32:HIERARCHY_NONE:120:130:1537
TF1:-10:INVERSION_AUTO:BANDWIDTH_8_MHZ:FEC_AUTO:FEC_1_2:QAM_64:TRANSMISSION_MODE_8K:GUARD_INTERVAL_1_32:HIERARCHY_NONE:120:130:1537
the first one being okay but the second one clearly wrong du to the 
negative frequency of "-10".

I hope that some body else may reproduce this bug/problem and that it 
may be solved before 2.6.19-final arrives.

Greetings, Klaus.



More information about the linux-dvb mailing list