[linux-dvb] Re: Mantis VP-1027/VP-1033/VP-1034/VP-2033/VP-3033

[Manu Abraham]

Marko Ristola wrote:
> 
> Hi
> 
> I tested the Mantis driver. I used stock Fedora Core 5 kernel (AMD64)
> I have the 2033 card, as described in
> http://www.twinhan.com/product_cable_2033.asp
> 
> 05:04.0 Multimedia controller: Twinhan Technology Co. Ltd Mantis DTV PCI
> Bridge Controller [Ver 1.0] (rev 01)
>        Subsystem: Twinhan Technology Co. Ltd Unknown device 0008
>        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
> ParErr- Stepping- SERR- FastB2B-
>        Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort-
> <TAbort+ <MAbort- >SERR- <PERR-
>        Latency: 32 (2000ns min, 63750ns max)
>        Interrupt: pin A routed to IRQ 217
>        Region 0: Memory at c2100000 (32-bit, prefetchable) [size=4K]
> 
> I'm new with this driver: I bought my first digital dvb card on saturday.
> 
> I modified the Mantis driver when I tried to make it work:
> cu1216_set_parameters(): return 0 on success (returned -1 always).
> cu1216_set_parameters(): do thorough scanning:
> - Test inverse and noninverse cases with all three gains.
> - Test error rate ten times and accept the best one, that got FE_HAS_LOCK.
> - Use the best inverse/noninverse case with the lock with the best gain.
> 
> My implementation was unacceptably slow, but cu1216_set_parameters() was
> successful.
> 
> Here is a description of the gained status acquired with
> cu1216_read_status:
> 
> [cu1216_read_status]: status = 31, sync=47 {FE_HAS_SIGNAL |
> FE_HAS_CARRIER} {FE_HAS_SYNC | FE_H
> AS_VITERBI} {FE_HAS_LOCK}
> 
> sync had varying values with status=31:
> sync=47
> sync=63
> sync=15
> 
> I haven't been able to get a picture yet.
> While writing this email, I was able to get Kaffeine to open the pipe
> for picture reading.
> After testing the driver once, I have to reboot: the dvb card does not
> respond otherwise.
> cu1216_sleep() might be the cause for that: it powers down ADC and does
> standby.
> 
> I suspect that doing standby is the problem and thus:
> 
> xine calls cu1216_sleep() before opening the pipe with FC5 and so it
> can't read data stream.
> Kaffeine calls cu1216_sleep() on exit. So restarting Kaffeine won't
> work, but data stream reading
> might work just after reboot as I have seen once.

Just comment out the sleep callback in the ops struct in cu1216. That
will make it not to sleep anymore, for debugging/testing.

> 
> Here is the blocking problem for me (acquired with Kaffeine with lots of
> added debug messages):
> 
> Oct  8 22:12:43 koivu kernel: [cu1216_read_status]: status = 31, sync=47
> {FE_HAS_SIGNAL | FE_HAS_CARRIER} {FE_HAS_SYNC | FE_HAS_VITERBI}
> {FE_HAS_LOCK}
> Oct  8 22:12:43 koivu kernel: mantis start feed & dma
> Oct  8 22:12:43 koivu kernel: CALL cu1216_read_status
> Oct  8 22:12:43 koivu kernel: [cu1216_read_status]: status = 0, sync=0
> (Note: the signal has been lost here!)
> Oct  8 22:12:43 koivu kernel: CALL cu1216_set_parameters
> Oct  8 22:12:43 koivu kernel: CALL cu1216_read_status
> Oct  8 22:12:43 koivu kernel: [cu1216_read_status]: status = 0, sync=0
> ....
> 
> If somebody wants my version of cu1216.c, I can send it by email.
> The slowness makes it unacceptable as a final implementation.


You can send it over/post it here itself.


Regards,
Manu