Re: [PATCH 5/6] xfs: refactor by-size extent allocation mode

[Brian Foster <bfoster@redhat.com>]

On Wed, May 15, 2019 at 01:10:16AM -0700, Christoph Hellwig wrote:
> > WRT to merging the functions, I'm a little concerned about the result
> > being too large. What do you think about folding in _vextent_type() but
> > at the same time factoring out the rmap/counter/resv post alloc bits
> > into an xfs_alloc_ag_vextent_accounting() helper or some such?
> 
> Sounds good to me.  I've looked at the function and another nice thing
> to do would be to not pass the ret bno to xfs_alloc_ag_vextent_agfl,
> but let that function fill out the args structure return value itself.
> 

I believe that's what the existing xfs_alloc_ag_vextent_small() function
does if it happens to allocate from the AGFL. I initially found that
inconsistent, but looking at the additional refactoring with the _type()
function folded away and whatnot I think it's actually better. I'll push
the args update back down into the AGFL helper.

> Also for the trace piints that still say near in them - maybe we should
> change that near to ag?

Yeah, I need to make another pass over the tracepoints...

Brian