On Tue, Dec 20, 2022 at 08:20:03AM -0800, Darrick J. Wong wrote:
> On Tue, Dec 20, 2022 at 08:10:08AM -0800, Darrick J. Wong wrote:
> > On Tue, Dec 20, 2022 at 04:00:01PM +1100, Dave Chinner wrote:
> > > For future consideration, we don't use maxrecs in this calculation
> > > at all - should we just pass minrecs into the function rather than
> > > an array of limits?
>
> I prefer to replace all those mxr/mnr array constructs with an explicit
> structure:
>
> struct xfs_btblock_geometry {
> uint16_t leaf_maxrecs;
> uint16_t leaf_minrecs;
>
> uint16_t node_maxrecs;
> uint16_t node_minrecs;
> };
>
> and get rid of all the mp->m_rmap_mxr[leaf != 0] stuff that slows me
> down every time I have to read it.
Yup, that sounds like a great idea - I have to work it out from
first principles every time, too.
-Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
