Re: [PATCH 3/3] minix: calculate s_firstdatazone to support more inodes

[Jan Kara]

On Tue 25-11-25 22:52:07, Jori Koolstra wrote:
> Hello Jan,
> 
> > 
> > This changelog misses explanation why handling s_firstdatazone == 0 is a
> > good idea in the first place. Is it somehow situation present in real
> > filesystems? If not then then why bother? More flexibility in accepting
> > garbage always means more trouble down the road :), we've learned this
> > lesson several times.
> > 
> 
> I believe that the Minix OS implementation of mfs supports it:
> 
> https://github.com/Stichting-MINIX-Research-Foundation/minix/
> blob/4db99f4012570a577414fe2a43697b2f239b699e/minix/usr.sbin/
> mkfs.mfs/mkfs.c#L659

OK, but so far nobody complained that they have images using this feature
they cannot read => no support needed from us. 

> But I understand your concern. Perhaps we should not make too many
> changes to minix as discussed before. I would really like to work on a
> FUSE driver for it, but am a bit stuck. I've discussed this with Shuah
> and she encourages me to do it. However, I need to have some help to vet
> my idea on how to approach it. She would have helped me but told me she
> is not an fs expert. Can I ask you some questions to see if my approach
> is at least sensible? It would make a huge difference to me. Of course, I
> also understand if you have no time for this.

Sure. I suggest you can send your questions to
linux-fsdevel@xxxxxxxxxxxxxxx list (linux-kernel is really a firehose
barely anybody manages to read) and CC me. I don't promise a prompt reply but 
I'll try to get to it. Generally transitioning these legacy filesystems to
FUSE implementations is desirable to lower kernel-code maintenance burden
so we have also selfish motivation to help you with that :)

								Honza

-- 
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR