Re: [PATCH] mm/page_alloc: fix memalloc_nocma_{save/restore} APIs

[Joonsoo Kim]

2020ë 7ì 21ì (í) ìí 9:43, Matthew Wilcox <willy@xxxxxxxxxxxxx>ëì ìì:
>
> On Tue, Jul 21, 2020 at 02:38:56PM +0200, Vlastimil Babka wrote:
> > On 7/21/20 2:05 PM, Matthew Wilcox wrote:
> > > On Tue, Jul 21, 2020 at 12:28:49PM +0900, js1304@xxxxxxxxx wrote:
> > >> @@ -4619,8 +4631,10 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
> > >>            wake_all_kswapds(order, gfp_mask, ac);
> > >>
> > >>    reserve_flags = __gfp_pfmemalloc_flags(gfp_mask);
> > >> -  if (reserve_flags)
> > >> +  if (reserve_flags) {
> > >>            alloc_flags = reserve_flags;
> > >> +          alloc_flags = current_alloc_flags(gfp_mask, alloc_flags);
> > >> +  }
> > >
> > > Is this right?  Shouldn't you be passing reserve_flags to
> > > current_alloc_flags() here?  Also, there's no need to add { } here.
> >
> > Note the "alloc_flags = reserve_flags;" line is not being deleted here. But if
> > it was, your points would be true, and yeah it would be a bit more tidy.
>
> Oh ... I should wake up a little more.
>
> Yeah, I'd recommend just doing this:
>
> -               alloc_flags = reserve_flags;
> +               alloc_flags = current_alloc_flags(gfp_mask, reserve_flags);

Okay. I will change it. Just note that the reason that I added it
separately is that
I think that separation is more readable since we can easily notice
that alloc_flags
is changed to reserve_flags without inspecting currect_alloc_flags() function.

Thanks.