On Wed 04-09-24 10:39:35, Barry Song wrote:
> On Mon, Sep 2, 2024 at 7:58 PM Michal Hocko <mho...@suse.com> wrote:
> >
> > On Sat 31-08-24 08:28:23, Barry Song wrote:
> > > From: Barry Song <v-songbao...@oppo.com>
> > >
> > > Three points for this change:
> > >
> > > 1. We should consolidate all warnings in one place. Currently, the
> > > order > 1 warning is in the hotpath, while others are in less
> > > likely scenarios. Moving all warnings to the slowpath will reduce
> > > the overhead for order > 1 and increase the visibility of other
> > > warnings.
> > >
> > > 2. We currently have two warnings for order: one for order > 1 in
> > > the hotpath and another for order > costly_order in the laziest
> > > path. I suggest standardizing on order > 1 since it’s been in
> > > use for a long time.
> > >
> > > 3. We don't need to check for __GFP_NOWARN in this case. __GFP_NOWARN
> > > is meant to suppress allocation failure reports, but here we're
> > > dealing with bug detection, not allocation failures. So replace
> > > WARN_ON_ONCE_GFP by WARN_ON_ONCE.
> > >
> > > Suggested-by: Vlastimil Babka <vba...@suse.cz>
> > > Signed-off-by: Barry Song <v-songbao...@oppo.com>
> >
> > Acked-by: Michal Hocko <mho...@suse.com>
> >
> > Updating the doc about order > 1 sounds like it would still fall into
> > the scope of this patch. I don not think we absolutely have to document
> > each unsupported gfp flags combination for GFP_NOFAIL but the order is a
> > good addition with a note that kvmalloc should be used instead in such a
> > case.
>
> Hi Andrew,
> If there are no objections from Michal and David, could you please
> squash the following:
>
> >From fc7a2a49e8d0811d706d13d2080393274f316806 Mon Sep 17 00:00:00 2001
> From: Barry Song <v-songbao...@oppo.com>
> Date: Wed, 4 Sep 2024 10:26:19 +1200
> Subject: [PATCH] mm: also update the doc for __GFP_NOFAIL with order > 1
>
> Obviously we only support order <= 1 __GFP_NOFAIL allocation and if
> someone wants larger memory, they should consider using kvmalloc()
> instead.
>
> Suggested-by: David Hildenbrand <da...@redhat.com>
> Suggested-by: Michal Hocko <mho...@suse.com>
> Signed-off-by: Barry Song <v-songbao...@oppo.com>
LGTM. Thanks!
> ---
> include/linux/gfp_types.h | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/include/linux/gfp_types.h b/include/linux/gfp_types.h
> index 4a1fa7706b0c..65db9349f905 100644
> --- a/include/linux/gfp_types.h
> +++ b/include/linux/gfp_types.h
> @@ -253,7 +253,8 @@ enum {
> * used only when there is no reasonable failure policy) but it is
> * definitely preferable to use the flag rather than opencode endless
> * loop around allocator.
> - * Using this flag for costly allocations is _highly_ discouraged.
> + * Allocating pages from the buddy with __GFP_NOFAIL and order > 1 is
> + * not supported. Please consider using kvmalloc() instead.
> */
> #define __GFP_IO ((__force gfp_t)___GFP_IO)
> #define __GFP_FS ((__force gfp_t)___GFP_FS)
> --
> 2.34.1
>
>
> >
> > > ---
> > > mm/page_alloc.c | 50 ++++++++++++++++++++++++-------------------------
> > > 1 file changed, 25 insertions(+), 25 deletions(-)
> > >
> > > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > > index c81ee5662cc7..e790b4227322 100644
> > > --- a/mm/page_alloc.c
> > > +++ b/mm/page_alloc.c
> > > @@ -3033,12 +3033,6 @@ struct page *rmqueue(struct zone *preferred_zone,
> > > {
> > > struct page *page;
> > >
> > > - /*
> > > - * We most definitely don't want callers attempting to
> > > - * allocate greater than order-1 page units with __GFP_NOFAIL.
> > > - */
> > > - WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
> > > -
> > > if (likely(pcp_allowed_order(order))) {
> > > page = rmqueue_pcplist(preferred_zone, zone, order,
> > > migratetype, alloc_flags);
> > > @@ -4175,6 +4169,7 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int
> > > order,
> > > {
> > > bool can_direct_reclaim = gfp_mask & __GFP_DIRECT_RECLAIM;
> > > bool can_compact = gfp_compaction_allowed(gfp_mask);
> > > + bool nofail = gfp_mask & __GFP_NOFAIL;
> > > const bool costly_order = order > PAGE_ALLOC_COSTLY_ORDER;
> > > struct page *page = NULL;
> > > unsigned int alloc_flags;
> > > @@ -4187,6 +4182,25 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned
> > > int order,
> > > unsigned int zonelist_iter_cookie;
> > > int reserve_flags;
> > >
> > > + if (unlikely(nofail)) {
> > > + /*
> > > + * We most definitely don't want callers attempting to
> > > + * allocate greater than order-1 page units with
> > > __GFP_NOFAIL.
> > > + */
> > > + WARN_ON_ONCE(order > 1);
> > > + /*
> > > + * Also we don't support __GFP_NOFAIL without
> > > __GFP_DIRECT_RECLAIM,
> > > + * otherwise, we may result in lockup.
> > > + */
> > > + WARN_ON_ONCE(!can_direct_reclaim);
> > > + /*
> > > + * PF_MEMALLOC request from this context is rather bizarre
> > > + * because we cannot reclaim anything and only can loop
> > > waiting
> > > + * for somebody to do a work for us.
> > > + */
> > > + WARN_ON_ONCE(current->flags & PF_MEMALLOC);
> > > + }
> > > +
> > > restart:
> > > compaction_retries = 0;
> > > no_progress_loops = 0;
> > > @@ -4404,29 +4418,15 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned
> > > int order,
> > > * Make sure that __GFP_NOFAIL request doesn't leak out and make
> > > sure
> > > * we always retry
> > > */
> > > - if (gfp_mask & __GFP_NOFAIL) {
> > > + if (unlikely(nofail)) {
> > > /*
> > > - * All existing users of the __GFP_NOFAIL are blockable, so
> > > warn
> > > - * of any new users that actually require GFP_NOWAIT
> > > + * Lacking direct_reclaim we can't do anything to reclaim
> > > memory,
> > > + * we disregard these unreasonable nofail requests and still
> > > + * return NULL
> > > */
> > > - if (WARN_ON_ONCE_GFP(!can_direct_reclaim, gfp_mask))
> > > + if (!can_direct_reclaim)
> > > goto fail;
> > >
> > > - /*
> > > - * PF_MEMALLOC request from this context is rather bizarre
> > > - * because we cannot reclaim anything and only can loop
> > > waiting
> > > - * for somebody to do a work for us
> > > - */
> > > - WARN_ON_ONCE_GFP(current->flags & PF_MEMALLOC, gfp_mask);
> > > -
> > > - /*
> > > - * non failing costly orders are a hard requirement which we
> > > - * are not prepared for much so let's warn about these users
> > > - * so that we can identify them and convert them to
> > > something
> > > - * else.
> > > - */
> > > - WARN_ON_ONCE_GFP(costly_order, gfp_mask);
> > > -
> > > /*
> > > * Help non-failing allocations by giving some access to
> > > memory
> > > * reserves normally used for high priority non-blocking
> > > --
> > > 2.34.1
> >
> > --
> > Michal Hocko
> > SUSE Labs
>
> Thanks
> Barry
--
Michal Hocko
SUSE Labs
