blkmtd and name_to_dev_t

[Josh Boyer]

On 4/3/06, Eric W. Biederman <ebiederman at lnxi.com> wrote:
> "H. Peter Anvin" <hpa at zytor.com> writes:
>
> > There seems to be a few options, in order from least work to more work:
> >
> > - Removing the blkmtd driver and tell people to use block2mtd;
> > - Have kinit intercept blkmtd configuration options and emulate them using
> > block2mtd;
> > - Adjust blkmtd to have a configuration interface, and then make kinit configure
> > it.
>
> Block devices are not especially interesting when working with flash chips, as they
> are an emulation layer.  The character devices are much more interesting,
> and the native filesystems don't go through a block or a character device.
> So killing blkmtd sounds like a reasonable thing to do.

Erm, you've got the idea behind blkmtd a bit backwards.  It's not to
make an MTD device show up as a block device.  It's to make a block
device show up as an MTD device.  Quite useful for testing JFFS2, etc
when one doesn't have immediate access to flash chips.  But that is
all beside the point.

> So I suggest what happens is that we deprecate blkmtd in Documentation/FeatureRemoval
> of 2.6.17 and delete it in the -mm tree.
>
> Then you can merge klibc.
>
> Does any one have a problem with that plan?

I believe this sounds good.

josh