dev builds
themes manual
device status forums
mailing lists
IRC bugs
dev guide

Rockbox mail archive

Subject: Re: Recording woes (in detail)

Re: Recording woes (in detail)

From: Christophe Avoinne <>
Date: Thu, 14 Nov 2002 17:12:56 +0100

Huh, I'm not sure having two FATs is so safe indeed. Just imagine that there
is a scratch during the second FAT updating, you'll get a damaged second
FAT... Except with some heuristics or a good examination of the partition,
how could you determine which FAT is good ? so far as there are no CRC
checksums integrated in FATs, it doesn't seem for me to have several FATs is
a real safety. Oh yeah maybe having two FATs is interesting lest you should
have a bad sector in the first FAT, so you would still be able to access a
file or a directory through the second FAT.

And yes, to answer to Björn, having more than two FATs is a nonsense since
it would downgrade filesystem performances without being a real safety plus,
so very few people are used to formating a FAT32 partition with more than 2
FATs, but it is possible at the cost of insanity :). I did format with only
one FAT per partition, but I got in trouble when I wanted to defragment this
partition because the software under Windows wants two FATs to do so, so I
wouldn't recommend people to format a FAT32 partition with just one FAT if
using Windows.

In fact, I'm just wondering whether that safity is valuable comparing with
the fact that your harddisk will spin up twice to update two FATs because
the second one would be to far away from the first one to be entirely cached
in the HD buffer during the first access. Could there be any impact about
power-saving ? I'm not able to answer this question.

Again, my 2 cents worth.

> > Just a suggestion :
> >
> > Does the FAT32 code in rockbox update the FAT entries in each FAT
> > in a FAT32
> > partition ? Windows may be very sensitive about the integrity of FATs.
> > the FAT32 partition has the FAT mirroring flag on, you must update as
> > FAT entries as there are FATs in a FAT32 partition when
> > creating/truncating
> > a file. Or do you force the FAT mirroring flag to be off (surely the
> > solution for embedded application since you would only need to update
> > current FAT instead of all) ?
> on the other side, embedded means "running on batteries", not really safe,
> and so one.
> Right ?
> it would be then i good idea to use fat mirroring in order to by on a safe
> side. what do you think ?
> henri
Received on 2002-11-14

Page template was last modified "Tue Sep 7 00:00:02 2021" The Rockbox Crew -- Privacy Policy