Subject: + mm-thp-avoid-excessive-compaction-latency-during-fault.patch added to -mm tree To: rientjes@xxxxxxxxxx,gthelen@xxxxxxxxxx,mgorman@xxxxxxx,n-horiguchi@xxxxxxxxxxxxx,vbabka@xxxxxxx From: akpm@xxxxxxxxxxxxxxxxxxxx Date: Wed, 07 May 2014 14:20:51 -0700 The patch titled Subject: mm, thp: avoid excessive compaction latency during fault has been added to the -mm tree. Its filename is mm-thp-avoid-excessive-compaction-latency-during-fault.patch This patch should soon appear at http://ozlabs.org/~akpm/mmots/broken-out/mm-thp-avoid-excessive-compaction-latency-during-fault.patch and later at http://ozlabs.org/~akpm/mmotm/broken-out/mm-thp-avoid-excessive-compaction-latency-during-fault.patch Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/SubmitChecklist when testing your code *** The -mm tree is included into linux-next and is updated there every 3-4 working days ------------------------------------------------------ From: David Rientjes <rientjes@xxxxxxxxxx> Subject: mm, thp: avoid excessive compaction latency during fault Synchronous memory compaction can be very expensive: it can iterate an enormous amount of memory without aborting, constantly rescheduling, waiting on page locks and lru_lock, etc, if a pageblock cannot be defragmented. Unfortunately, it's too expensive for transparent hugepage page faults and it's much better to simply fallback to pages. On 128GB machines, we find that synchronous memory compaction can take O(seconds) for a single thp fault. Now that async compaction remembers where it left off without strictly relying on sync compaction, this makes thp allocations best-effort without causing egregious latency during fault. We still need to retry async compaction after reclaim, but this won't stall for seconds. Signed-off-by: David Rientjes <rientjes@xxxxxxxxxx> Acked-by: Mel Gorman <mgorman@xxxxxxx> Cc: Greg Thelen <gthelen@xxxxxxxxxx> Cc: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx> Cc: Vlastimil Babka <vbabka@xxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- mm/page_alloc.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff -puN mm/page_alloc.c~mm-thp-avoid-excessive-compaction-latency-during-fault mm/page_alloc.c --- a/mm/page_alloc.c~mm-thp-avoid-excessive-compaction-latency-during-fault +++ a/mm/page_alloc.c @@ -2578,7 +2578,17 @@ rebalance: &did_some_progress); if (page) goto got_pg; - migration_mode = MIGRATE_SYNC_LIGHT; + + if (gfp_mask & __GFP_NO_KSWAPD) { + /* + * Khugepaged is allowed to try MIGRATE_SYNC_LIGHT, the latency + * of this allocation isn't critical. Everything else, however, + * should only be allowed to do MIGRATE_ASYNC to avoid excessive + * stalls during fault. + */ + if ((current->flags & (PF_KTHREAD | PF_KSWAPD)) == PF_KTHREAD) + migration_mode = MIGRATE_SYNC_LIGHT; + } /* * If compaction is deferred for high-order allocations, it is because _ Patches currently in -mm which might be from rientjes@xxxxxxxxxx are mm-memcontrol-remove-hierarchy-restrictions-for-swappiness-and-oom_control.patch mm-memcontrol-remove-hierarchy-restrictions-for-swappiness-and-oom_control-fix.patch hugetlb-prep_compound_gigantic_page-drop-__init-marker.patch hugetlb-add-hstate_is_gigantic.patch hugetlb-update_and_free_page-dont-clear-pg_reserved-bit.patch hugetlb-move-helpers-up-in-the-file.patch hugetlb-add-support-for-gigantic-page-allocation-at-runtime.patch mem-hotplug-implement-get-put_online_mems.patch slab-get_online_mems-for-kmem_cache_createdestroyshrink.patch mm-debug-make-bad_range-output-more-usable-and-readable.patch memcg-do-not-hang-on-oom-when-killed-by-userspace-oom-access-to-memory-reserves.patch mm-migration-add-destination-page-freeing-callback.patch mm-compaction-return-failed-migration-target-pages-back-to-freelist.patch mm-compaction-add-per-zone-migration-pfn-cache-for-async-compaction.patch mm-compaction-embed-migration-mode-in-compact_control.patch mm-thp-avoid-excessive-compaction-latency-during-fault.patch mm-compaction-terminate-async-compaction-when-rescheduling.patch mm-utilc-add-kstrimdup.patch linux-next.patch -- To unsubscribe from this list: send the line "unsubscribe mm-commits" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html