Mailing List archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[linux-dvb] Re: av7110_send_fw_cmd error ect.



> Andreas Share wrote:
> >
> > just insert some debug printkīs in av7110_before_after_tune() before the
two
> > firmware calls, and one printk in SetPIDs() and you see you got several
> > calls from this 2 funktions, especialy without signal.
> > You could also insert in __av7110_send_fw_cmd() (av7110_hw.c) an
> > "rdebi(av7110, DEBINOSWAP, COMMAND, 0, 2" just before the
printk(KERN_ERR
> > "%s: timeout waiting for COMMAND idle\n", __FUNCTION__); and an
additional
> > printk to print out the rdebi return.
> >
> > You get the last command & funktion transfered to the command buffer. I
> > could bet you returned 0x0100 ;) And because the command stay in the
> > command-buffer, no other firmware command will be transfered or
executed.
> > After some time the arm crashed also...
>
> Things are not so easy, because some kinds of messages are first
> pushed into a message queue and executed later. So the last
> command from the PC might not be the one which caused the ARM crash.
> Specifically COMTYPE_PIDFILTER commands (MultiPID etc.) are queued.
> The 0x0100 then means that the ARM crashed before it could push the
> command into the message queue (or while doing it).

I could understand the problematic, especialy because i donīt know how the
firmware works after an command is uploaded to the command buffer. I only
see it from driver side with an hanging MultiPID, and an executed
Scan/FlushTSQueue before.

BTW: COMTYPE_PIDFILTER, AddPIDfilter/DelPIDfilter makeīs no trouble, running
vdr without any section filter runs into the same firmware command error.

Maybe there is a way to workaround the problem from driver side, posibly
there is a chance to set the pes-filter one by one.....just an idea.

Andreas



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



Home | Main Index | Thread Index