From: Dave Jones <dave_at_hfpacoustical.com>
Date: Tue, 07 Oct 2003 15:13:56 -0500

At 09:36 AM 10/7/2003 +0200, Linus wrote:
>I think it's the wrong way to fix things. Let's remove the cause of the
>panic instead.
>
>Linus

Sounds like a good idea to me.

I brought up the issue because I've been having *Panic rec upd: -6 errors
during my long term recording sessions. I don't monitor the units so I
don't find out about the errors until I check on them (sometimes days
later) and find them stopped.

The rec upd error is returned by the Xing header code. Apparently, at the
beginning of MPEG_STOP_DONE, the file is closed and then opened again. This
second open fails for some reason. -6 appears to indicate that the file
can't be found in the directory. So it sounds like for some reason the file
is closed and then can't be found during the immediate reopen. Might this
be due to a slow hard drive spin up or a power fluctuation that delays a
write, or are we writing to a buffer here? Sorry, I'm not much of a
programmer and haven't been able to digest all the guts of Rockbox.

I've removed the Xing header creation section and am running a comparison
test between Rockbox with the Xing code and Rockbox without. It might take
a few days before any results come in, though.

Thanks,



Dave



>Dave Jones wrote:
>>What I'd like is to set up my Rockbox so that as soon as it was turned on
>>it would start recording. Then if any errors were encountered during
>>recording, it would Rolo, forcing a restart of the firmware and thus
>>start recording again. It's pretty straight forward to set Rockbox up to
>>start recording right after startup, but is it a really bad idea to Rolo
>>after a panic message?

_______________________
Dave Jones P.E.
Senior Project Engineer
HFP Acoustical Consultants Inc.
6001 Savoy Dr., Suite 115
Houston, Texas 77036
713.789.9400 (phone)
713.789.5493 (fax)
www.hfpacoustical.com
_______________________
Received on 2003-10-07