Re: [PATCH 05/17] luks: Add support for LUKS2 in (proc)/luks_script

[Glenn Washburn]

On Thu, 30 Jul 2020 17:14:54 +0200
Patrick Steinhardt <ps@pks.im> wrote:

> >        {
> > -   size += sizeof ("luks_mount ");
> > +   size += grub_strlen (i->modname);
> > +   size += sizeof ("_mount");
> >     size += grub_strlen (i->uuid);
> >     size += grub_strlen (i->cipher->cipher->name);
> > -   size += 54;
> > +   /* mode + mode_iv + spaces + offset + sector size + ??? +
> > '\n' + NULL */
> > +   size += 5 + 8 + 5 + 20 + 4 + 16 + 1 + 1;
> 
> Is it expected that the `size` is now bigger than before? This adds up
> to `60` now. It's fine as it is more verbose than it previously has
> been, but a comment in the commit message explaining that the
> different size is intentional would've helped.

Yes it the size is expected to be bigger, I also added a "sector size"
field, which is at most 4 digits and a space.  I added an extra byte
for NULL, but later realized I don't need that and that later commit is
in patch 14.  That should account for the 6 extra bytes.  I'm still not
sure why there's an extra 16 bytes unaccounted for, but I've left it
in.  The requested changes will be in the forth coming patchset.

> >     if (i->essiv_hash)
> >       size += grub_strlen (i->essiv_hash->name);
> >     size += i->keysize * 2;