Re: [patch] mm, oom: remove gfp helper function

[Andrew Morton]

On Wed, 3 Dec 2014 16:52:22 +0100 Michal Hocko <mhocko@xxxxxxx> wrote:

> On Mon 01-12-14 18:30:40, Johannes Weiner wrote:
> > On Thu, Nov 27, 2014 at 11:25:47AM +0100, Michal Hocko wrote:
> > > On Wed 26-11-14 14:17:32, David Rientjes wrote:
> > > > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > > > --- a/mm/page_alloc.c
> > > > +++ b/mm/page_alloc.c
> > > > @@ -2706,7 +2706,7 @@ rebalance:
> > > >  	 * running out of options and have to consider going OOM
> > > >  	 */
> > > >  	if (!did_some_progress) {
> > > > -		if (oom_gfp_allowed(gfp_mask)) {
> > > 		/*
> > > 		 * Do not attempt to trigger OOM killer for !__GFP_FS
> > > 		 * allocations because it would be premature to kill
> > > 		 * anything just because the reclaim is stuck on
> > > 		 * dirty/writeback pages.
> > > 		 * __GFP_NORETRY allocations might fail and so the OOM
> > > 		 * would be more harmful than useful.
> > > 		 */
> > 
> > I don't think we need to explain the individual flags, but it would
> > indeed be useful to remark here that we shouldn't OOM kill from
> > allocations contexts with (severely) limited reclaim abilities.
> 
> Is __GFP_NORETRY really related to limited reclaim abilities? I thought
> it was merely a way to tell the allocator to fail rather than spend too
> much time reclaiming.

That's my understanding of __GFP_NORETRY.  However it seems that I
really didn't have a plan:

: commit 75908778d91e92ca3c9ed587c4550866f4c903fc
: Author: Andrew Morton <akpm@xxxxxxxxx>
: Date:   Sun Apr 20 00:28:12 2003 -0700
: 
:     [PATCH] implement __GFP_REPEAT, __GFP_NOFAIL, __GFP_NORETRY
:     
:     This is a cleanup patch.
:     
:     There are quite a lot of places in the kernel which will infinitely retry a
:     memory allocation.
:     
:     Generally, they get it wrong.  Some do yield(), the semantics of which have
:     changed over time.  Some do schedule(), which can lock up if the caller is
:     SCHED_FIFO/RR.  Some do schedule_timeout(), etc.
:     
:     And often it is unnecessary, because the page allocator will do the retry
:     internally anyway.  But we cannot rely on that - this behaviour may change
:     (-aa and -rmap kernels do not do this, for instance).
:     
:     So it is good to formalise and to centralise this operation.  If an
:     allocation specifies __GFP_REPEAT then the page allocator must infinitely
:     retry the allocation.
:     
:     The semantics of __GFP_REPEAT are "try harder".  The allocation _may_ fail
:     (the 2.4 -aa and -rmap VM's do not retry infinitely by default).
:     
:     The semantics of __GFP_NOFAIL are "cannot fail".  It is a no-op in this VM,
:     but needs to be honoured (or fix up the callers) if the VM ischanged to not
:     retry infinitely by default.
:     
:     The semantics of __GFP_NOREPEAT are "try once, don't loop".  This isn't used
:     at present (although perhaps it should be, in swapoff).  It is mainly for
:     completeness.

(that's a braino in the changelog: it should be
s/__GFP_NOREPEAT/__GFP_NORETRY/)

> If you are referring to __GFP_FS part then I have
> no objections to be less specific, of course, but __GFP_IO would fall
> into the same category but we are not checking for it. I have no idea
> why we consider the first and not the later one, to be honest...

(__GFP_FS && !__GFP_IO) doesn't make much sense and probably doesn't
happen.  "__GFP_FS implies __GFP_IO" is OK.

Anyway, yes, This particular piece of __alloc_pages_slowpath() sorely
needs documenting please.  Once we manage to work out why we're doing
what we're doing!

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/