On Tue, Aug 27, 2019 at 05:11:06PM +1000, Dave Chinner wrote:
> On Tue, Aug 20, 2019 at 01:30:57PM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> >
> > Create helper functions to perform unit conversions against a runtime
> > filesystem, then remove the open-coded versions in scrub.
>
> .... and there they are...
>
> > +/* Convert fs block number into bytes */
> > +static inline uint64_t
> > +xfrog_fsb_to_b(
> > + const struct xfs_fd *xfd,
> > + uint64_t fsb)
> > +{
> > + return fsb << xfd->blocklog;
> > +}
>
> FWIW, this is for converting linear offsets in fsb /units/, not the
> sparse fsbno (= agno | agbno) to bytes. I've always found it a bit
> nasty that this distinction is not clearly made in the core FSB
> conversion macros.
>
> perhaps off_fsb_to_b?
Ok to both.
--D
> > +/* Convert bytes into (rounded down) fs block number */
> > +static inline uint64_t
> > +xfrog_b_to_fsbt(
> > + const struct xfs_fd *xfd,
> > + uint64_t bytes)
> > +{
> > + return bytes >> xfd->blocklog;
> > +}
>
> Ditto.
>
> Otherwise looks ok.
>
> Cheers,
>
> Dave.
> --
> Dave Chinner
> david@xxxxxxxxxxxxx
