On 16 Jul 2001 23:20:37 +0200, Seth Mos wrote:
> On 16 Jul 2001, Eric Sandeen wrote:
> > Hm, I didn't know those errors were still out there... I thought we
> > fixed all the raid controllers? but yes, it should.
> They were not, I just questioned if this is solved in some way?
> Does this mean we always use one size or still switch different sizes but
> in another place of the code.
The block size is always set correctly at mount time, so everything is
fine from then on. The only trouble was in userspace, on a device which
had some other blocksize previously, and had not yet been mounted as xfs
- i.e. during mkfs.xfs time.
So, I guess it's the latter - we switch to the correct size in another
place in the code - at mount time, in the kernel - and it's no longer
necessary to do it in userspace before this.