RFC: kill msdos and vfat modules?

RFC: kill msdos and vfat modules?

[Christoph Hellwig]

There's almost no code left in the msdos and vfat modules with all
the guts in the shared fat module.  Shouldn't we just do away with
the vfat and msdos modules and just provide module aliases in the
main fat module?

Re: RFC: kill msdos and vfat modules?

[OGAWA Hirofumi]

Christoph Hellwig <hch@lst.de> writes:

> There's almost no code left in the msdos and vfat modules with all
> the guts in the shared fat module.  Shouldn't we just do away with
> the vfat and msdos modules and just provide module aliases in the
> main fat module?

I have no objection to it if it works. Maybe, one concern is the order
of detection (rootfs, /proc/filesystems) when it's not moudle. (however,
probably it's not hard to control).

Thanks.
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

Re: RFC: kill msdos and vfat modules?

[Christoph Hellwig]

On Thu, Jun 04, 2009 at 10:26:22AM +0900, OGAWA Hirofumi wrote:
> > There's almost no code left in the msdos and vfat modules with all
> > the guts in the shared fat module.  Shouldn't we just do away with
> > the vfat and msdos modules and just provide module aliases in the
> > main fat module?
> 
> I have no objection to it if it works. Maybe, one concern is the order
> of detection (rootfs, /proc/filesystems) when it's not moudle. (however,
> probably it's not hard to control).

For the non-modular case detection order is already determined by link
order, so nothing would change, at least if we still keep config options
to allow for 8.3 vs vfat naming.  Then again the former is so simple
that we could always compile it in anyway.

Re: RFC: kill msdos and vfat modules?

[OGAWA Hirofumi]

Christoph Hellwig <hch@lst.de> writes:

> On Thu, Jun 04, 2009 at 10:26:22AM +0900, OGAWA Hirofumi wrote:
>> > There's almost no code left in the msdos and vfat modules with all
>> > the guts in the shared fat module.  Shouldn't we just do away with
>> > the vfat and msdos modules and just provide module aliases in the
>> > main fat module?
>> 
>> I have no objection to it if it works. Maybe, one concern is the order
>> of detection (rootfs, /proc/filesystems) when it's not moudle. (however,
>> probably it's not hard to control).
>
> For the non-modular case detection order is already determined by link
> order, so nothing would change, at least if we still keep config options
> to allow for 8.3 vs vfat naming.  Then again the former is so simple
> that we could always compile it in anyway.

If modules was merged to one, msdos and vfat in fs/Makefile will be
merged too. If so, it will not define the link order anymore.  Anyway,
nothing would be hard, but, I guess there is the change which need to care.
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>