From: Mike Holden <rockbox_at_mikeholden.uklinux.net>
Date: Thu, 27 Feb 2003 13:12:56 -0000 (GMT)

Björn Stenberg said:
> I would like you to test this patch for the ata driver (without your
> mpeg modification). It adds a soft reset on disk error:
>

Built this last night (clean CVS apart from that ata.c patch), and tried
it on my daily 1 hour walk this dinnertime. I still got some lockups,
but it didn't seem like as many as usual. THis is of course hit and
miss, and not repeatable testing!

I initially had the anti-skip buffer at 3 seconds, but set it to the
maximum of 7 seconds after the first skip.

The symptoms are as follows with this build:

red led is off.
Track timer continues to count upwards, but nothing is heard.
Pressing PAUSE changes the status icon to paused. Pressing PLAY changes
it back to play, but again no sound is heard.
Pressing NEXT moves on to the next track, and again it looks like it is
playing normally, but nothing is heard.
After the first lock up, the next track started to play properly (with
sound), after the track time reached 11 seconds in. The disk started
activity at this point.
After the second lock up, I left it playing silently again on the next
track after pressing NEXT, but n osound came back (I probably just
didn't leave it long enough however.) After pressing NEXT/PREV a couple
of times, the disk spun and normal service was resumed again.

First guess is that when we get a failed disk read, the buffer is being
populated anyway, but with NULLs. I have no access to the code from
here, but there's a few things there to think about!

-- 
Mike Holden

Received on 2003-02-27