Re: Big Endian fix patch

[Lennart Sorensen]

On Wed, Jul 28, 2010 at 01:42:19PM -0400, Lennart Sorensen wrote:
> On Wed, Jul 28, 2010 at 01:12:16PM -0400, Doug Nazar wrote:
> >  On 2010-07-28 1:01 PM, Lennart Sorensen wrote:
> >> Well I don't think that helped (unless I messed up the compile).
> >>
> >
> > The following works for me. Just finished testing it. Damn these 15  
> > minute boot times.
> >
> > Doug
> >
> 
> > === modified file 'normal/completion.c'
> > --- normal/completion.c     2010-06-11 20:31:16 +0000
> > +++ normal/completion.c     2010-07-28 17:02:42 +0000
> > @@ -247,7 +247,7 @@
> >        goto fail;
> >      }
> >  
> > -  dir = grub_strchr (current_word, '/');
> > +  dir = grub_strchr (current_word + grub_strlen(device) + 2, '/');
> >    last_dir = grub_strrchr (current_word, '/');
> >    if (dir)
> >      {
> > 
> 
> I wish the completion would realize that an md device without partitions
> (so no partition table) should not complete with a ','.  Not sure how
> doable that is.

Another thing that puzzles me is that when I boot linux, I always get
a message that says the state is wrong with the superblock on the sda
partitions, but the sdb partitions have the right state.  I have no idea
if grub messes with this, or if the system isn't shutting down the raid
properly on reboots.  It manages to assemble the raid and fix it
automatically, but the error is strange since I have never seen that
before.  Of course I haven't really used 1.x raid before either, so
maybe that's part of it.

-- 
Len Sorensen