Re: [PATCH 2/2] mm, oom: do not enfore OOM killer for __GFP_NOFAIL automatically

[Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>]

Tetsuo Handa wrote:
> Michal Hocko wrote:
> > On Thu 08-12-16 21:53:44, Tetsuo Handa wrote:
> > > If we could agree
> > > with calling __alloc_pages_nowmark() before out_of_memory() if __GFP_NOFAIL
> > > is given, we can avoid locking up while minimizing possibility of invoking
> > > the OOM killer...
> >
> > I do not understand. We do __alloc_pages_nowmark even when oom is called
> > for GFP_NOFAIL.
> 
> Where is that? I can find __alloc_pages_nowmark() after out_of_memory()
> if __GFP_NOFAIL is given, but I can't find __alloc_pages_nowmark() before
> out_of_memory() if __GFP_NOFAIL is given.
> 
> What I mean is below patch folded into
> "[PATCH 1/2] mm: consolidate GFP_NOFAIL checks in the allocator slowpath".
> 
Oops, I wrongly implemented "__alloc_pages_nowmark() before out_of_memory() if
__GFP_NOFAIL is given." case. Updated version is shown below.

--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3116,23 +3116,27 @@ void warn_alloc(gfp_t gfp_mask, const char *fmt, ...)
 		/* The OOM killer may not free memory on a specific node */
 		if (gfp_mask & __GFP_THISNODE)
 			goto out;
+	} else {
+		/*
+		 * Help non-failing allocations by giving them access to memory
+		 * reserves
+		 */
+		page = get_page_from_freelist(gfp_mask, order,
+					      ALLOC_NO_WATERMARKS|ALLOC_CPUSET, ac);
+		/*
+		 * fallback to ignore cpuset restriction if our nodes
+		 * are depleted
+		 */
+		if (!page)
+			page = get_page_from_freelist(gfp_mask, order,
+						      ALLOC_NO_WATERMARKS, ac);
+		if (page)
+			goto out;
 	}
+
 	/* Exhausted what can be done so it's blamo time */
-	if (out_of_memory(&oc) || WARN_ON_ONCE(gfp_mask & __GFP_NOFAIL)) {
+	if (out_of_memory(&oc))
 		*did_some_progress = 1;
-
-		if (gfp_mask & __GFP_NOFAIL) {
-			page = get_page_from_freelist(gfp_mask, order,
-					ALLOC_NO_WATERMARKS|ALLOC_CPUSET, ac);
-			/*
-			 * fallback to ignore cpuset restriction if our nodes
-			 * are depleted
-			 */
-			if (!page)
-				page = get_page_from_freelist(gfp_mask, order,
-					ALLOC_NO_WATERMARKS, ac);
-		}
-	}
 out:
 	mutex_unlock(&oom_lock);
 	return page;
@@ -3738,6 +3742,11 @@ bool gfp_pfmemalloc_allowed(gfp_t gfp_mask)
 		 */
 		WARN_ON_ONCE(order > PAGE_ALLOC_COSTLY_ORDER);
 
+		/* Try memory reserves and then start killing things. */
+		page = __alloc_pages_may_oom(gfp_mask, order, ac, &did_some_progress);
+		if (page)
+			goto got_pg;
+
 		cond_resched();
 		goto retry;
 	}

I'm calling __alloc_pages_may_oom() from nopage: because we reach without
calling __alloc_pages_may_oom(), for PATCH 1/2 is not for stop enforcing
the OOM killer for __GFP_NOFAIL.