Re: [PATCH] mm/zsmalloc: Use memcpy_from/to_page whereever possible

[Pintu Agarwal]

On Fri, 11 Oct 2024 at 07:30, Sergey Senozhatsky
<senozhatsky@xxxxxxxxxxxx> wrote:
>
> On (24/10/10 23:21), Pintu Kumar wrote:
> [..]
> > @@ -1511,10 +1501,10 @@ static void zs_object_copy(struct size_class *class, unsigned long dst,
> >               d_size -= size;
> >
> >               /*
> > -              * Calling kunmap_atomic(d_addr) is necessary. kunmap_atomic()
> > -              * calls must occurs in reverse order of calls to kmap_atomic().
> > -              * So, to call kunmap_atomic(s_addr) we should first call
> > -              * kunmap_atomic(d_addr). For more details see
> > +              * Calling kunmap_local(d_addr) is necessary. kunmap_local()
> > +              * calls must occurs in reverse order of calls to kmap_local_page().
> > +              * So, to call kunmap_local(s_addr) we should first call
> > +              * kunmap_local(d_addr). For more details see
> >                * Documentation/mm/highmem.rst.
> >                */
>
> I'd prefer this entire comment to be dropped.
Oh I thought the below code for k[un]map_local[_page] stills exists,
so the comments are still valid.
Ok I will remove it in the next patchset.
Looks like there are a few more code improvements possible.

Thank you,
Pintu