[linux-dvb] DVB-T driver crashing applications?

[Rainer Zocholl <UseNet-Posting-Nospam-74308-@zocki.toppoint.de>]

Hello

i played a bit with "palantir-2.5.5".

Attaching it to /dev/video0 (which is a FF DVB-S 2.1)
works flawlessly and i can get the video stream over the net.

Then i tried the /dev/video3 which is a AverMdia 771 DVB-T (bt748):
plantir starts and tells someting true about the device.
But when i connect remotely to the palatir server it crashes
with a segment fault.
Ok, null pointer reference...bad application...

But now i have a similar effect with VDR 1.3.17.
When VDR is recording from one DVB-T card and i switch
to another transponder VDR crashes earlier or later (10sec) 
(details is VDR-list)

That occurs too if i start VDR after the experiment with palantir above:
/VDR/runvdr: line 76: 15275 Segmentation fault      su -c "$VDRCMD" $VDRUSR

No further infos..


The second DVB-T card is a TwinHan VP3021.



How can this be debugged?


Rainer


5:15 [main] -- palantir 2.5.5 starting --
5:15 [main] Definitions for 7 devices found
5:15 [main] No serial port specified
5:15 [main] No named pipe specified
5:15 [video] Card:   BT878 video (AVerMedia AVerTV D (/dev/video2)
5:15 [video] Capabilities:   171 (capture tuner overlay clipping scales )
5:15 [video] Size:   (48x32)-(924x576)
5:15 [video] Channel no. 0 ('Composite0')    tuners: 0, flags: 2, type: 2, norm: 0
5:15 [video] Channel no. 1 ('S-Video')       tuners: 0, flags: 2, type: 2, norm: 0
5:15 [video] brightness: 32768  hue: 32768  colour: 32768
5:15 [video] contrast: 32768  whiteness: 0
5:15 [video] depth: 24  palette: 4
5:15 [video] Actual capture box is 320x240 (requested 320x240)
5:15 [video] Video buffer size: 17039360, frames: 8
5:15 [video] mmap of 17039360 bytes failed
5:15 [video] mmap() not available - reverting to read()
5:15 [main] Starting device thread
5:16 [audio] Audio device supports half duplex only
5:16 [audio] Can't set mono channel mode
5:16 [main] Audio initialization failed. Audio channel will be disabled
5:16 [main] Listening on port 3000
5:16 [main] Control queue timeout set to 30 s
..attaching client
6:10 [main] 10.10.10.1:1076 #0: connection accepted, spawning server.
6:10 [main] Capture thread started
6:10 [main] 1 client(s) connected
Segmentation fault

On the console "Unable to handle null pointer reference at virtual address
0049" (sorry have currently no free serial port to give the entire message)




In the syslog in found:

6:10 kernel:  printing eip:
6:10 kernel: e2aecb50
6:10 kernel: Oops: 0000
6:10 kernel: CPU:    0
6:10 kernel: EIP:    0010:[<e2aecb50>]    Not tainted
6:10 kernel: EFLAGS: 00010246
6:10 kernel: eax: e2af8d80   ebx: d9d7f008   ecx: 00000001 edx: 00000039
6:10 kernel: esi: 00000000   edi: d9d7f00c   ebp: d488fe9c esp: d488fe8c
6:10 kernel: ds: 0018   es: 0018   ss: 0018
6:10 kernel: Process palantir (pid: 14675, stackpage=d488f000)
6:10 kernel: Stack: 000003c0 0001c200 cbfa74a0 00000001 d488fee0 e2aedc00 e2af8d80 cbfa7548
6:10 kernel:        00000039 00000000 000003c0 000003c0 00000078 00000000 cbfa74a0 00000140
6:10 kernel:        cbfa74a0 e2902816 c15f2000 cbfa74c0 e2aeed80 d488ff00 e2ae5424 e2af8d80
6:10 kernel: Call Trace:    [<e2aedc00>] [<e2af8d80>] [<e2902816>] [<e2aeed80>] [<e2ae5424>]
6:10 kernel:   [<e2af8d80>] [<e2ae54e2>] [<e2af8d80>] [<e2aeeff0>] [<e2902fcd>] [<e29030dc>]
6:10 kernel:   [<e2ae7b77>] [sys_read+148/240][system_call+51/56]
6:10 kernel:
6:10 kernel: Code: 8b 42 10 29 f0 39 45 18 77 1b 8b 45 18 0d 00 00 00 1c 89 03