[vdr] Re: softdevice 0.0.7 cvs crash tests...

[Nicolas Huillard <nhuillard@ghs.fr>]

Stefan Lucke wrote:

> On Freitag, 6. August 2004 01:48, Nicolas Huillard wrote:
>
> > I have completed a brand new compile of vdr 1.3.12, softdevice 0.0.7 
> > from CVS tonight, DirectFB from CVS tonight, etc. It seems to work a 
> > little better on live TV. Recordings still crash at the same points than 
> > before.
> > I have edited the recordings I had that demonstated typical crashes : 
> > you can donwload them at http://nicolas.huillard.net/vdr/ :
> >
> > %%24/2004-06-12.23.27.50.50.rec/marks.vdr
> > %%24/2004-06-12.23.27.50.50.rec/summary.vdr
> > %%24/2004-06-12.23.27.50.50.rec/index.vdr
> > %%24/2004-06-12.23.27.50.50.rec/README.txt
> > %%24/2004-06-12.23.27.50.50.rec/001.vdr
> Had a look at this. Crash is reproduceable so it might be a softdevice
> problem but a second look with xlist ( ftp://ftp.cadsoft.de/vdr/xlist.c ) 
> shows some strange output (mplayer has a glicht too. but doesn't segfault):
This recording is one of my very first ones, recorded while using vdr 
1.2.6 + softmpeg, if that matters. Others are newer, with 1.3.12 and the 
latest softdevice.
It shouldn't crash though...

>  2433976 E0 video  1084 04 3C 80 00 00
>  2435066 C0 audio  2042 07 FA 87 80 05 2B B5 DD 44 4F pts: 1836524071          6480
>  2437114 C0 audio   273 01 11 80 00 00
>  2437393 E6 video  2042 07 FA 00 45 55 0A FD 8A DB 40 9D D8 BE 51 E7 4B B0 12 36 9C 9E D0 64 8A 56 95 7E 8F 37 37 36 3D EB 23 AB 53 4A E9 03 7E 5A 74 BD D9 A5 28 FC C9 24 4F 1E C
> 0 90 44 47 BE B7 CD 20 3D E8 13 E0 2F 9F 31 56 77 46 D6 31 7F 8B 5C A2 E6 4D 89 C2 23 A9 42 7D 58 C2 D9 76 C6 40 41 pts: 2137189792          300653892
>  2439441 E6 video  2042 07 FA 0F BA BB FA AA D7 83 21 BD C3 B4 8E 72 61 E4 23 72 C9 1F C4 7E 8D 47 E8 70 1D D6 6E CC E2 BD 07 74 6D 8B 0C 8B 15 87 7C BB 91 F1 1E 8F FC 92 42 40 0
> 7 01 45 57 A6 83 F0 74 3E 27 9A 59 D1 EC A9 B3 07 DD FE 31 DD A3 F9 4D 3F 99 28 F1 B5 12 FB 75 69 2D 16 A6 04 17 BA 11 AA 8B 14 87 02 4B 69 AC 81 73 52 E9 1B CA A0 2A FB EA 4C E5
>  46 2C 04 FF 97 9B 7F 1A 6D 7C 3F 96 8D C8 9D 31 C8 B1 D5 B7 50 82 DA F6 1B EF AE 81 D2 AC E1 B7 E1 D6 2A 93 A4 32 6B 9F 28 03 C7 3E 80 61 49 69 B3 14 B9 18 C4 A7 7D 7D 8D 6A F4
> 97 4A 3A D6 34 D3 EC F9 0F BE 52 69 79 86 08 25 76 A7 AD EA EF 41 pts: 1790296464          3948073968
All this doesn't mean much to me : I hope someone will jump on this ?

> > I don't know if I can do much more to help us debug all this. I think I 
> > can review the code in search of local logic errors or buffer overflow 
> > and such. Is that of any help ?
> ParseStreamIntern ( .. ) .. Help is always welcome .
I'll have a look at cStreamDecoder::ParseStreamIntern() in 
vdr-plugin-softdevice-0.0.7pre2/mpeg2decoder.c
Have you a few hints in how to efficiently use gdb or any text-mode 
interface to it ? (leave debug info in the code, inspect variables, etc. 
: I never used it beside showing the back-trace)

--
NH