Rockbox mail archiveSubject: Disk error handling
Disk error handling
From: Brian King <brking_at_charter.net>
Date: Sun, 10 Nov 2002 20:08:03 -0600
I recently took my archos on a roadtrip - something I haven't done since
upgrading my hard drive to a 40G Toshiba - and found that the Toshiba
tends to be much less resiliant to shock than the stock drive. I started
the trip using firmware from shortly after 1.4 and started having
problems. It would play for about 30-60 seconds, and then the audio
would cut out. After playing around with it I figured out that it would
do this when it went to spin up the drive and read more data. The player
seemed to be happy, the display looked like the song was playing, but no
sound. Thanks to ROLO, I was able to try several different firmware
loads. I tried going back to 1.3 and the behavior I saw was that instead
of dropping the audio, it skipped to the next song - which is what I
would expect from 1.3 when a disk error was hit. I then tried code that
was built from 11/04 and I had improved behavior. Now when encountering
an error, retries were made and things played a bit more smoothly. I got
several skips, but it played much longer than the ~1.4 build until it
finally ran into the same problem - audio got cut off. So here are my
1. What state is the disk ERP code in? Are there improvements that could
still be made?
2. Would it be possible to have a higher high-watermark when the charger
is plugged in? I was thinking this might decrease the skips.
3. Should my Toshiba drive be doing this? Are there 40G disk drives that
people have used that behave better than this under shock?
BTW, I finally just put the archos in my lap for the remainder of the
trip and it worked flawlessly.
-- Some days it's just not worth chewing through the restraints...Received on 2002-11-11