In isolate_freepages(), low_pfn was mistakenly initialized to pageblock_start_pfn() instead of pageblock_end_pfn(), creating a possible underflow, as described by Hugh:
There's a case when that "block_start_pfn -= pageblock_nr_pages" loop can pass through 0 and end up trying to access a pageblock before the start of the mem_map[]. Fixes: mmotm mm-compaction-wrap-calculating-first-and-last-pfn-of-pageblock.patch Reported-by: Hugh Dickins <hu...@google.com> Signed-off-by: Vlastimil Babka <vba...@suse.cz> --- mm/compaction.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/compaction.c b/mm/compaction.c index 759c3ac73ced..6a49d1b35515 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -999,7 +999,7 @@ static void isolate_freepages(struct compact_control *cc) block_start_pfn = pageblock_start_pfn(cc->free_pfn); block_end_pfn = min(block_start_pfn + pageblock_nr_pages, zone_end_pfn(zone)); - low_pfn = pageblock_start_pfn(cc->migrate_pfn); + low_pfn = pageblock_end_pfn(cc->migrate_pfn); /* * Isolate free pages until enough are available to migrate the -- 2.8.1