On Wed, Apr 05, 2023 at 08:46:45AM -0700, Christoph Hellwig wrote:
> On Wed, Apr 05, 2023 at 12:36:42PM +0200, Andrey Albershteyn wrote:
> > Hi Christoph,
> >
> > On Tue, Apr 04, 2023 at 08:30:36AM -0700, Christoph Hellwig wrote:
> > > On Tue, Apr 04, 2023 at 04:53:01PM +0200, Andrey Albershteyn wrote:
> > > > Not the whole folio always need to be verified by fs-verity (e.g.
> > > > with 1k blocks). Use passed folio's offset and size.
> > >
> > > Why can't those callers just call fsverity_verify_blocks directly?
> > >
> >
> > They can. Calling _verify_folio with explicit offset; size appeared
> > more clear to me. But I'm ok with dropping this patch to have full
> > folio verify function.
>
> Well, there is no point in a wrapper if it has the exact same signature
> and functionality as the functionality being wrapped.
>
> That being said, right now fsverity_verify_folio, so it might make sense
> to either rename it, or rename fsverity_verify_blocks to
> fsverity_verify_folio. But that's really a question for Eric.
I thought it would be confusing for fsverity_verify_folio() to not actually
verify a whole folio. So, for now we have:
fsverity_verify_page: verify a whole page
fsverity_verify_folio: verify a whole folio
fsverity_verify_blocks: verify a range of blocks in a folio
IMO that makes sense. Note: fsverity_verify_folio() is currently unused, but
ext4 might use it.
So, just use fsverity_verify_blocks().
- Eric
