Hey Carlos,
On Thu, Oct 18, 2012 at 03:40:33PM -0300, Carlos Maiolino wrote:
> On Thu, Oct 18, 2012 at 12:15:18PM -0500, Mark Tinguely wrote:
> > On 10/18/12 12:04, Carlos Maiolino wrote:
> > >On Thu, Oct 18, 2012 at 11:05:32AM -0500, Mark Tinguely wrote:
> > >>On 10/18/12 11:00, Dave Howorth wrote:
> > >>>Mark Tinguely wrote:
> > >>>>Would "Indicates that XFS is allowed to create inodes at locations up to
> > >>>>32 bits of significance .."
> > >>>
> > >>>I prefer the original wording. Your suggestion says something about what
> > >>>XFS can do, but nothing about what it is not allowed to do, which is
> > >>>rather more important.
> > >>>
> > >>>_______________________________________________
> > >>>xfs mailing list
> > >>>xfs@xxxxxxxxxxx
> > >>>http://oss.sgi.com/mailman/listinfo/xfs
> > >>
> > >>I see your point. Sounds good to me.
> > >>
> > >>--Mark.
> > >>
> > >This means no change is needed?
> > >
> >
> > I am okay with what is written about creating inodes.
> >
> > On a separate question, should something be mentioned that inode32
> > mode can still read/write/unlink any inode, even those number
> > greater than 32 bit, or will that confuse the inode creation point?
> >
> I thought about this when modifying the documentation, but, to be honest I was
> wondering if this might not cause more confusion than expected.
>
> For example,
> afaik, one of the principal reasons we still keep an inode32 allocator is due
> applications which cannot handle 64bit inodes. So, I suppose that users which
> will use inode32 are those who really needs 32bit inodes for this kind of
> 'problem'.
>
> Saying that inode32 mode can still read 64bit inodes might (IMHO) lead users,
> *think* they won't have problems with larger inodes just by using inode32
> mode,
> when, AFAICT, they'll have the same problems with their applications and
> larger
> inode numbers if they have any inode allocated beyong 32bit limit. even using
> inode32 mode.
>
> So, I believe that, not saying it will "force" users who need 32bit inodes to
> use inode32 since its first inode allocation, instead of think they'll fix
> their
> problems only switching to inode32 mode after they already had lots of 64bit
> inodes allocated.
>
>
>
> Hopefully I didn't create more confusion around it :-)
I pulled in the original version of this on Nov 2. That version looked
fine to me. ;)
Regards,
Ben
|