Re: [PATCH 14/20] mm: zswap: function ordering: public lru api

[Johannes Weiner <hannes@cmpxchg.org>]

On Tue, Jan 30, 2024 at 03:47:25PM -0800, Nhat Pham wrote:
> On Mon, Jan 29, 2024 at 5:42 PM Johannes Weiner <hannes@cmpxchg.org> wrote:
> >
> > The zswap entry section sits awkwardly in the middle of LRU-related
> > functions. Group the external LRU API functions first.
> >
> > Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
> > ---
> >  mm/zswap.c | 37 +++++++++++++++++++------------------
> >  1 file changed, 19 insertions(+), 18 deletions(-)
> >
> > diff --git a/mm/zswap.c b/mm/zswap.c
> > index e650fc587116..511bfafc1456 100644
> > --- a/mm/zswap.c
> > +++ b/mm/zswap.c
> > @@ -746,6 +746,10 @@ static int zswap_enabled_param_set(const char *val,
> >         return ret;
> >  }
> >
> > +/*********************************
> > +* lru functions
> > +**********************************/
> > +
> 
> nit: looks like there are 2 "lru functions" headers after this patch?
> You remove the "lruvec functions" header, then add another "lru
> functions" header it seems. The next patch removes one of them, so end
> result is fine I guess - just seems a bit odd.

Yeah that's an artifact of trying to make git produce readable
diffs. Since the lru functions are right next to the entry functions,
I went through several attempts where it wouldn't generate clean moves
but instead would interleave entry and lru functions line by line to
overwrite one with the other in place. I think the above helped in
making it not do that, although I'm not positive it would still be
required in the final form of this patch. It was kind of brute force.

> That asides:
> Reviewed-by: Nhat Pham <nphamcs@gmail.com>

Thanks!