]> git.proxmox.com Git - mirror_ubuntu-eoan-kernel.git/commitdiff
mm, page_alloc: fix a division by zero error when boosting watermarks v2
authorMel Gorman <mgorman@techsingularity.net>
Thu, 21 Feb 2019 06:19:49 +0000 (22:19 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 21 Feb 2019 17:01:00 +0000 (09:01 -0800)
Yury Norov reported that an arm64 KVM instance could not boot since
after v5.0-rc1 and could addressed by reverting the patches

  1c30844d2dfe272d58c ("mm: reclaim small amounts of memory when an external
  73444bc4d8f92e46a20 ("mm, page_alloc: do not wake kswapd with zone lock held")

The problem is that a division by zero error is possible if boosting
occurs very early in boot if the system has very little memory.  This
patch avoids the division by zero error.

Link: http://lkml.kernel.org/r/20190213143012.GT9565@techsingularity.net
Fixes: 1c30844d2dfe ("mm: reclaim small amounts of memory when an external fragmentation event occurs")
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Reported-by: Yury Norov <yury.norov@gmail.com>
Tested-by: Yury Norov <yury.norov@gmail.com>
Tested-by: Will Deacon <will.deacon@arm.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/page_alloc.c

index 7f79b78bc8294557480a6df4fdb328f757895a7d..0b9f577b1a2aee9f2b959f2d717e936802cbfc75 100644 (file)
@@ -2170,6 +2170,18 @@ static inline void boost_watermark(struct zone *zone)
 
        max_boost = mult_frac(zone->_watermark[WMARK_HIGH],
                        watermark_boost_factor, 10000);
+
+       /*
+        * high watermark may be uninitialised if fragmentation occurs
+        * very early in boot so do not boost. We do not fall
+        * through and boost by pageblock_nr_pages as failing
+        * allocations that early means that reclaim is not going
+        * to help and it may even be impossible to reclaim the
+        * boosted watermark resulting in a hang.
+        */
+       if (!max_boost)
+               return;
+
        max_boost = max(pageblock_nr_pages, max_boost);
 
        zone->watermark_boost = min(zone->watermark_boost + pageblock_nr_pages,