@@ -480,10 +480,15 @@ unsigned long reclaim_clean_pages_from_list(struct zone *zone,
#define ALLOC_OOM ALLOC_NO_WATERMARKS
#endif
-#define ALLOC_HARDER 0x10 /* try to alloc harder */
-#define ALLOC_HIGH 0x20 /* __GFP_HIGH set */
-#define ALLOC_CPUSET 0x40 /* check for correct cpuset */
-#define ALLOC_CMA 0x80 /* allow allocations from CMA areas */
+#define ALLOC_HARDER 0x10 /* try to alloc harder */
+#define ALLOC_HIGH 0x20 /* __GFP_HIGH set */
+#define ALLOC_CPUSET 0x40 /* check for correct cpuset */
+#define ALLOC_CMA 0x80 /* allow allocations from CMA areas */
+#ifdef CONFIG_ZONE_DMA32
+#define ALLOC_NOFRAGMENT 0x100 /* avoid mixing pageblock types */
+#else
+#define ALLOC_NOFRAGMENT 0x0
+#endif
enum ttu_flags;
struct tlbflush_unmap_batch;
@@ -2375,20 +2375,30 @@ static bool unreserve_highatomic_pageblock(const struct alloc_context *ac,
* condition simpler.
*/
static __always_inline bool
-__rmqueue_fallback(struct zone *zone, int order, int start_migratetype)
+__rmqueue_fallback(struct zone *zone, int order, int start_migratetype,
+ unsigned int alloc_flags)
{
struct free_area *area;
int current_order;
+ int min_order = order;
struct page *page;
int fallback_mt;
bool can_steal;
+ /*
+ * Do not steal pages from freelists belonging to other pageblocks
+ * i.e. orders < pageblock_order. In the event there is on local
+ * zone free, the allocation will retry later.
+ */
+ if (alloc_flags & ALLOC_NOFRAGMENT)
+ min_order = pageblock_order;
+
/*
* Find the largest available free page in the other list. This roughly
* approximates finding the pageblock with the most free pages, which
* would be too costly to do exactly.
*/
- for (current_order = MAX_ORDER - 1; current_order >= order;
+ for (current_order = MAX_ORDER - 1; current_order >= min_order;
--current_order) {
area = &(zone->free_area[current_order]);
fallback_mt = find_suitable_fallback(area, current_order,
@@ -2447,7 +2457,8 @@ __rmqueue_fallback(struct zone *zone, int order, int start_migratetype)
* Call me with the zone->lock already held.
*/
static __always_inline struct page *
-__rmqueue(struct zone *zone, unsigned int order, int migratetype)
+__rmqueue(struct zone *zone, unsigned int order, int migratetype,
+ unsigned int alloc_flags)
{
struct page *page;
@@ -2457,7 +2468,8 @@ __rmqueue(struct zone *zone, unsigned int order, int migratetype)
if (migratetype == MIGRATE_MOVABLE)
page = __rmqueue_cma_fallback(zone, order);
- if (!page && __rmqueue_fallback(zone, order, migratetype))
+ if (!page && __rmqueue_fallback(zone, order, migratetype,
+ alloc_flags))
goto retry;
}
@@ -2472,13 +2484,14 @@ __rmqueue(struct zone *zone, unsigned int order, int migratetype)
*/
static int rmqueue_bulk(struct zone *zone, unsigned int order,
unsigned long count, struct list_head *list,
- int migratetype)
+ int migratetype, unsigned int alloc_flags)
{
int i, alloced = 0;
spin_lock(&zone->lock);
for (i = 0; i < count; ++i) {
- struct page *page = __rmqueue(zone, order, migratetype);
+ struct page *page = __rmqueue(zone, order, migratetype,
+ alloc_flags);
if (unlikely(page == NULL))
break;
@@ -2934,6 +2947,7 @@ static inline void zone_statistics(struct zone *preferred_zone, struct zone *z)
/* Remove page from the per-cpu list, caller must protect the list */
static struct page *__rmqueue_pcplist(struct zone *zone, int migratetype,
+ unsigned int alloc_flags,
struct per_cpu_pages *pcp,
struct list_head *list)
{
@@ -2943,7 +2957,7 @@ static struct page *__rmqueue_pcplist(struct zone *zone, int migratetype,
if (list_empty(list)) {
pcp->count += rmqueue_bulk(zone, 0,
pcp->batch, list,
- migratetype);
+ migratetype, alloc_flags);
if (unlikely(list_empty(list)))
return NULL;
}
@@ -2959,7 +2973,8 @@ static struct page *__rmqueue_pcplist(struct zone *zone, int migratetype,
/* Lock and remove page from the per-cpu list */
static struct page *rmqueue_pcplist(struct zone *preferred_zone,
struct zone *zone, unsigned int order,
- gfp_t gfp_flags, int migratetype)
+ gfp_t gfp_flags, int migratetype,
+ unsigned int alloc_flags)
{
struct per_cpu_pages *pcp;
struct list_head *list;
@@ -2969,7 +2984,7 @@ static struct page *rmqueue_pcplist(struct zone *preferred_zone,
local_irq_save(flags);
pcp = &this_cpu_ptr(zone->pageset)->pcp;
list = &pcp->lists[migratetype];
- page = __rmqueue_pcplist(zone, migratetype, pcp, list);
+ page = __rmqueue_pcplist(zone, migratetype, alloc_flags, pcp, list);
if (page) {
__count_zid_vm_events(PGALLOC, page_zonenum(page), 1 << order);
zone_statistics(preferred_zone, zone);
@@ -2992,7 +3007,7 @@ struct page *rmqueue(struct zone *preferred_zone,
if (likely(order == 0)) {
page = rmqueue_pcplist(preferred_zone, zone, order,
- gfp_flags, migratetype);
+ gfp_flags, migratetype, alloc_flags);
goto out;
}
@@ -3011,7 +3026,7 @@ struct page *rmqueue(struct zone *preferred_zone,
trace_mm_page_alloc_zone_locked(page, order, migratetype);
}
if (!page)
- page = __rmqueue(zone, order, migratetype);
+ page = __rmqueue(zone, order, migratetype, alloc_flags);
} while (page && check_new_pages(page, order));
spin_unlock(&zone->lock);
if (!page)
@@ -3253,6 +3268,36 @@ static bool zone_allows_reclaim(struct zone *local_zone, struct zone *zone)
}
#endif /* CONFIG_NUMA */
+#ifdef CONFIG_ZONE_DMA32
+/*
+ * The restriction on ZONE_DMA32 as being a suitable zone to use to avoid
+ * fragmentation is subtle. If the preferred zone was HIGHMEM then
+ * premature use of a lower zone may cause lowmem pressure problems that
+ * are wose than fragmentation. If the next zone is ZONE_DMA then it is
+ * probably too small. It only makes sense to spread allocations to avoid
+ * fragmentation between the Normal and DMA32 zones.
+ */
+static inline unsigned int alloc_flags_nofragment(struct zone *zone)
+{
+ if (zone_idx(zone) != ZONE_NORMAL)
+ return 0;
+
+ /*
+ * If ZONE_DMA32 exists, assume it is the one after ZONE_NORMAL and
+ * the pointer is within zone->zone_pgdat->node_zones[].
+ */
+ if (!populated_zone(--zone))
+ return 0;
+
+ return ALLOC_NOFRAGMENT;
+}
+#else
+static inline unsigned int alloc_flags_nofragment(struct zone *zone)
+{
+ return 0;
+}
+#endif
+
/*
* get_page_from_freelist goes through the zonelist trying to allocate
* a page.
@@ -3264,11 +3309,14 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
struct zoneref *z = ac->preferred_zoneref;
struct zone *zone;
struct pglist_data *last_pgdat_dirty_limit = NULL;
+ bool no_fallback;
+retry:
/*
* Scan zonelist, looking for a zone with enough free.
* See also __cpuset_node_allowed() comment in kernel/cpuset.c.
*/
+ no_fallback = alloc_flags & ALLOC_NOFRAGMENT;
for_next_zone_zonelist_nodemask(zone, z, ac->zonelist, ac->high_zoneidx,
ac->nodemask) {
struct page *page;
@@ -3307,6 +3355,21 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
}
}
+ if (no_fallback) {
+ int local_nid;
+
+ /*
+ * If moving to a remote node, retry but allow
+ * fragmenting fallbacks. Locality is more important
+ * than fragmentation avoidance.
+ */
+ local_nid = zone_to_nid(ac->preferred_zoneref->zone);
+ if (zone_to_nid(zone) != local_nid) {
+ alloc_flags &= ~ALLOC_NOFRAGMENT;
+ goto retry;
+ }
+ }
+
mark = zone->watermark[alloc_flags & ALLOC_WMARK_MASK];
if (!zone_watermark_fast(zone, order, mark,
ac_classzone_idx(ac), alloc_flags)) {
@@ -3374,6 +3437,15 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
}
}
+ /*
+ * It's possible on a UMA machine to get through all zones that are
+ * fragmented. If avoiding fragmentation, reset and try again
+ */
+ if (no_fallback) {
+ alloc_flags &= ~ALLOC_NOFRAGMENT;
+ goto retry;
+ }
+
return NULL;
}
@@ -4371,6 +4443,12 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, int preferred_nid,
finalise_ac(gfp_mask, &ac);
+ /*
+ * Forbid the first pass from falling back to types that fragment
+ * memory until all local zones are considered.
+ */
+ alloc_flags |= alloc_flags_nofragment(ac.preferred_zoneref->zone);
+
/* First allocation attempt */
page = get_page_from_freelist(alloc_mask, order, alloc_flags, &ac);
if (likely(page))
The page allocator zone lists are iterated based on the watermarks of each zone which does not take anti-fragmentation into account. On x86, node 0 may have multiple zones while other nodes have one zone. A consequence is that tasks running on node 0 may fragment ZONE_NORMAL even though ZONE_DMA32 has plenty of free memory. This patch special cases the allocator fast path such that it'll try an allocation from a lower local zone before fragmenting a higher zone. In this case, stealing of pageblocks or orders larger than a pageblock are still allowed in the fast path as they are uninteresting from a fragmentation point of view. This was evaluated using a benchmark designed to fragment memory before attempting THPs. It's implemented in mmtests as the following configurations configs/config-global-dhp__workload_thpfioscale configs/config-global-dhp__workload_thpfioscale-defrag configs/config-global-dhp__workload_thpfioscale-madvhugepage e.g. from mmtests ./run-mmtests.sh --run-monitor --config configs/config-global-dhp__workload_thpfioscale test-run-1 The broad details of the workload are as follows; 1. Create an XFS filesystem (not specified in the configuration but done as part of the testing for this patch) 2. Start 4 fio threads that write a number of 64K files inefficiently. Inefficiently means that files are created on first access and not created in advance (fio parameterr create_on_open=1) and fallocate is not used (fallocate=none). With multiple IO issuers this creates a mix of slab and page cache allocations over time. The total size of the files is 150% physical memory so that the slabs and page cache pages get mixed 3. Warm up a number of fio read-only threads accessing the same files created in step 2. This part runs for the same length of time it took to create the files. It'll fault back in old data and further interleave slab and page cache allocations. As it's now low on memory due to step 2, fragmentation occurs as pageblocks get stolen. 4. While step 3 is still running, start a process that tries to allocate 75% of memory as huge pages with a number of threads. The number of threads is based on a (NR_CPUS_SOCKET - NR_FIO_THREADS)/4 to avoid THP threads contending with fio, any other threads or forcing cross-NUMA scheduling. Note that the test has not been used on a machine with less than 8 cores. The benchmark records whether huge pages were allocated and what the fault latency was in microseconds 5. Measure the number of events potentially causing external fragmentation, the fault latency and the huge page allocation success rate. 6. Cleanup Note that due to the use of IO and page cache that this benchmark is not suitable for running on large machines where the time to fragment memory may be excessive. Also note that while this is one mix that generates fragmentation that it's not the only mix that generates fragmentation. Differences in workload that are more slab-intensive or whether SLUB is used with high-order pages may yield different results. When the page allocator fragments memory, it records the event using the mm_page_alloc_extfrag event. If the fallback_order is smaller than a pageblock order (order-9 on 64-bit x86) then it's considered an event that may cause external fragmentation issues in the future. Hence, the primary metric here is the number of external fragmentation events that occur with order < 9. The secondary metric is allocation latency and huge page allocation success rates but note that differences in latencies and what the success rate also can affect the number of external fragmentation event which is why it's a secondary metric. 1-socket Skylake machine config-global-dhp__workload_thpfioscale XFS (no special madvise) 4 fio threads, 1 THP allocating thread -------------------------------------- 4.20-rc1 extfrag events < order 9: 1023463 4.20-rc1+patch: 358574 (65% reduction) thpfioscale Fault Latencies 4.20.0-rc1 4.20.0-rc1 vanilla lowzone-v2r4 Min fault-base-1 588.00 ( 0.00%) 557.00 ( 5.27%) Min fault-huge-1 0.00 ( 0.00%) 0.00 ( 0.00%) Amean fault-base-1 663.58 ( 0.00%) 663.65 ( -0.01%) Amean fault-huge-1 0.00 ( 0.00%) 0.00 ( 0.00%) 4.20.0-rc1 4.20.0-rc1 vanilla lowzone-v2r4 Percentage huge-1 0.00 ( 0.00%) 0.00 ( 0.00%) Fault latencies are reduced while allocation success rates remain at zero asthis configuration does not make any heavy effort to allocate THP and fio is heavily active at the time and filling memory. However, a 65% reduction of serious fragmentation events reduces the changes of external fragmentation being a problem in the future. 1-socket Skylake machine global-dhp__workload_thpfioscale-madvhugepage-xfs (MADV_HUGEPAGE) ----------------------------------------------------------------- 4.20-rc1 extfrag events < order 9: 342549 4.20-rc1+patch: 337890 (1% reduction) thpfioscale Fault Latencies 4.20.0-rc1 4.20.0-rc1 vanilla lowzone-v2r4 Amean fault-base-1 1517.06 ( 0.00%) 1531.37 ( -0.94%) Amean fault-huge-1 1129.50 ( 0.00%) 1160.95 ( -2.78%) thpfioscale Percentage Faults Huge 4.20.0-rc1 4.20.0-rc1 vanilla lowzone-v2r4 Percentage huge-1 78.01 ( 0.00%) 78.97 ( 1.23%) Nothing dramatic. Fragmentation events were only reduced slightly which is very different to what was reported in V1. A big difference with V1 is the relative size of Normal to the DMA32 zone. This machine has double the memory so the impact of using a small zone to avoid fragmentation events is much lower. 2-socket Haswell machine config-global-dhp__workload_thpfioscale XFS (no special madvise) 4 fio threads, 5 THP allocating threads ---------------------------------------------------------------- 4.20-rc1 extfrag events < order 9: 209820 4.20-rc1+patch: 185923 (11% reduction) thpfioscale Fault Latencies 4.20.0-rc1 4.20.0-rc1 vanilla lowzone-v2r4 Amean fault-base-5 1324.93 ( 0.00%) 1334.99 ( -0.76%) Amean fault-huge-5 4681.71 ( 0.00%) 2428.43 ( 48.13%) 4.20.0-rc1 4.20.0-rc1 vanilla lowzone-v2r4 Percentage huge-5 1.05 ( 0.00%) 1.13 ( 7.94%) The reduction of external fragmentation events is expected. A careful reader may spot that the reduction is lower than it was on v1. This is due to the removal of __GFP_THISNODE in commit ac5b2c18911f ("mm: thp: relax __GFP_THISNODE for MADV_HUGEPAGE mappings") as THP allocations can now spill over to remote nodes instead of fragmenting local memory. This reduces the impact of the use of a lower zone to avoid fragmentation. It's also worth noting relative to v1 that the allocation success rate is slightly higher. 2-socket Haswell machine global-dhp__workload_thpfioscale-madvhugepage-xfs (MADV_HUGEPAGE) ----------------------------------------------------------------- 4.20-rc1 extfrag events < order 9: 167464 4.20-rc1+patch: 130081 (22% reduction) thpfioscale Fault Latencies 4.20.0-rc1 4.20.0-rc1 vanilla lowzone-v2r4 Amean fault-base-5 7721.82 ( 0.00%) 6652.67 ( 13.85%) Amean fault-huge-5 3896.10 ( 0.00%) 2486.89 * 36.17%* thpfioscale Percentage Faults Huge 4.20.0-rc1 4.20.0-rc1 vanilla lowzone-v2r4 Percentage huge-5 95.02 ( 0.00%) 94.49 ( -0.56%) In this case, there was both a reduction in the external fragmentation causing events and the huge page allocation success latency with little change in the allocation success rates which were already high. A careful reader will note that V1 had very different outcomes both in terms of the number of fragmentation events and the allocation success rates. In this case, it's due to the baseline including the THP __GFP_THISNODE removaal patch. Overall, the patch significantly reduces the number of external fragmentation causing events so the success of THP over long periods of time would be improved for this adverse workload. While there are large differences compared to how V1 behaved, this is almost entirely accounted for by ac5b2c18911f ("mm: thp: relax __GFP_THISNODE for MADV_HUGEPAGE mappings"). Signed-off-by: Mel Gorman <mgorman@techsingularity.net> --- mm/internal.h | 13 +++++--- mm/page_alloc.c | 100 +++++++++++++++++++++++++++++++++++++++++++++++++------- 2 files changed, 98 insertions(+), 15 deletions(-)