diff options
Diffstat (limited to 'mm/page_alloc.c')
| -rw-r--r-- | mm/page_alloc.c | 263 |
1 files changed, 96 insertions, 167 deletions
diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 2b3bf6767d54..f64e7bcb43b7 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -92,7 +92,7 @@ int _node_numa_mem_[MAX_NUMNODES]; #endif #ifdef CONFIG_GCC_PLUGIN_LATENT_ENTROPY -volatile u64 latent_entropy __latent_entropy; +volatile unsigned long latent_entropy __latent_entropy; EXPORT_SYMBOL(latent_entropy); #endif @@ -2058,8 +2058,12 @@ out_unlock: * potentially hurts the reliability of high-order allocations when under * intense memory pressure but failed atomic allocations should be easier * to recover from than an OOM. + * + * If @force is true, try to unreserve a pageblock even though highatomic + * pageblock is exhausted. */ -static void unreserve_highatomic_pageblock(const struct alloc_context *ac) +static bool unreserve_highatomic_pageblock(const struct alloc_context *ac, + bool force) { struct zonelist *zonelist = ac->zonelist; unsigned long flags; @@ -2067,11 +2071,16 @@ static void unreserve_highatomic_pageblock(const struct alloc_context *ac) struct zone *zone; struct page *page; int order; + bool ret; for_each_zone_zonelist_nodemask(zone, z, zonelist, ac->high_zoneidx, ac->nodemask) { - /* Preserve at least one pageblock */ - if (zone->nr_reserved_highatomic <= pageblock_nr_pages) + /* + * Preserve at least one pageblock unless memory pressure + * is really high. + */ + if (!force && zone->nr_reserved_highatomic <= + pageblock_nr_pages) continue; spin_lock_irqsave(&zone->lock, flags); @@ -2085,13 +2094,25 @@ static void unreserve_highatomic_pageblock(const struct alloc_context *ac) continue; /* - * It should never happen but changes to locking could - * inadvertently allow a per-cpu drain to add pages - * to MIGRATE_HIGHATOMIC while unreserving so be safe - * and watch for underflows. + * In page freeing path, migratetype change is racy so + * we can counter several free pages in a pageblock + * in this loop althoug we changed the pageblock type + * from highatomic to ac->migratetype. So we should + * adjust the count once. */ - zone->nr_reserved_highatomic -= min(pageblock_nr_pages, - zone->nr_reserved_highatomic); + if (get_pageblock_migratetype(page) == + MIGRATE_HIGHATOMIC) { + /* + * It should never happen but changes to + * locking could inadvertently allow a per-cpu + * drain to add pages to MIGRATE_HIGHATOMIC + * while unreserving so be safe and watch for + * underflows. + */ + zone->nr_reserved_highatomic -= min( + pageblock_nr_pages, + zone->nr_reserved_highatomic); + } /* * Convert to ac->migratetype and avoid the normal @@ -2103,12 +2124,16 @@ static void unreserve_highatomic_pageblock(const struct alloc_context *ac) * may increase. */ set_pageblock_migratetype(page, ac->migratetype); - move_freepages_block(zone, page, ac->migratetype); - spin_unlock_irqrestore(&zone->lock, flags); - return; + ret = move_freepages_block(zone, page, ac->migratetype); + if (ret) { + spin_unlock_irqrestore(&zone->lock, flags); + return ret; + } } spin_unlock_irqrestore(&zone->lock, flags); } + + return false; } /* Remove an element from the buddy allocator from the fallback list */ @@ -2133,7 +2158,8 @@ __rmqueue_fallback(struct zone *zone, unsigned int order, int start_migratetype) page = list_first_entry(&area->free_list[fallback_mt], struct page, lru); - if (can_steal) + if (can_steal && + get_pageblock_migratetype(page) != MIGRATE_HIGHATOMIC) steal_suitable_fallback(zone, page, start_migratetype); /* Remove the page from the freelists */ @@ -2192,7 +2218,7 @@ static int rmqueue_bulk(struct zone *zone, unsigned int order, unsigned long count, struct list_head *list, int migratetype, bool cold) { - int i; + int i, alloced = 0; spin_lock(&zone->lock); for (i = 0; i < count; ++i) { @@ -2217,13 +2243,21 @@ static int rmqueue_bulk(struct zone *zone, unsigned int order, else list_add_tail(&page->lru, list); list = &page->lru; + alloced++; if (is_migrate_cma(get_pcppage_migratetype(page))) __mod_zone_page_state(zone, NR_FREE_CMA_PAGES, -(1 << order)); } + + /* + * i pages were removed from the buddy list even if some leak due + * to check_pcp_refill failing so adjust NR_FREE_PAGES based + * on i. Do not confuse with 'alloced' which is the number of + * pages added to the pcp list. + */ __mod_zone_page_state(zone, NR_FREE_PAGES, -(i << order)); spin_unlock(&zone->lock); - return i; + return alloced; } #ifdef CONFIG_NUMA @@ -2534,7 +2568,8 @@ int __isolate_free_page(struct page *page, unsigned int order) struct page *endpage = page + (1 << order) - 1; for (; page < endpage; page += pageblock_nr_pages) { int mt = get_pageblock_migratetype(page); - if (!is_migrate_isolate(mt) && !is_migrate_cma(mt)) + if (!is_migrate_isolate(mt) && !is_migrate_cma(mt) + && mt != MIGRATE_HIGHATOMIC) set_pageblock_migratetype(page, MIGRATE_MOVABLE); } @@ -3305,7 +3340,7 @@ retry: * Shrink them them and try again */ if (!page && !drained) { - unreserve_highatomic_pageblock(ac); + unreserve_highatomic_pageblock(ac, false); drain_all_pages(NULL); drained = true; goto retry; @@ -3422,8 +3457,10 @@ should_reclaim_retry(gfp_t gfp_mask, unsigned order, * Make sure we converge to OOM if we cannot make any progress * several times in the row. */ - if (*no_progress_loops > MAX_RECLAIM_RETRIES) - return false; + if (*no_progress_loops > MAX_RECLAIM_RETRIES) { + /* Before OOM, exhaust highatomic_reserve */ + return unreserve_highatomic_pageblock(ac, true); + } /* * Keep reclaiming pages while there is a chance this will lead @@ -3658,7 +3695,7 @@ retry: /* Make sure we know about allocations which stall for too long */ if (time_after(jiffies, alloc_start + stall_timeout)) { warn_alloc(gfp_mask, - "page alloction stalls for %ums, order:%u\n", + "page allocation stalls for %ums, order:%u", jiffies_to_msecs(jiffies-alloc_start), order); stall_timeout += 10 * HZ; } @@ -4224,7 +4261,7 @@ static void show_migration_types(unsigned char type) } *p = '\0'; - printk("(%s) ", tmp); + printk(KERN_CONT "(%s) ", tmp); } /* @@ -4335,7 +4372,8 @@ void show_free_areas(unsigned int filter) free_pcp += per_cpu_ptr(zone->pageset, cpu)->pcp.count; show_node(zone); - printk("%s" + printk(KERN_CONT + "%s" " free:%lukB" " min:%lukB" " low:%lukB" @@ -4382,8 +4420,8 @@ void show_free_areas(unsigned int filter) K(zone_page_state(zone, NR_FREE_CMA_PAGES))); printk("lowmem_reserve[]:"); for (i = 0; i < MAX_NR_ZONES; i++) - printk(" %ld", zone->lowmem_reserve[i]); - printk("\n"); + printk(KERN_CONT " %ld", zone->lowmem_reserve[i]); + printk(KERN_CONT "\n"); } for_each_populated_zone(zone) { @@ -4394,7 +4432,7 @@ void show_free_areas(unsigned int filter) if (skip_free_areas_node(filter, zone_to_nid(zone))) continue; show_node(zone); - printk("%s: ", zone->name); + printk(KERN_CONT "%s: ", zone->name); spin_lock_irqsave(&zone->lock, flags); for (order = 0; order < MAX_ORDER; order++) { @@ -4412,11 +4450,12 @@ void show_free_areas(unsigned int filter) } spin_unlock_irqrestore(&zone->lock, flags); for (order = 0; order < MAX_ORDER; order++) { - printk("%lu*%lukB ", nr[order], K(1UL) << order); + printk(KERN_CONT "%lu*%lukB ", + nr[order], K(1UL) << order); if (nr[order]) show_migration_types(types[order]); } - printk("= %lukB\n", K(total)); + printk(KERN_CONT "= %lukB\n", K(total)); } hugetlb_show_meminfo(); @@ -4977,72 +5016,6 @@ void __ref build_all_zonelists(pg_data_t *pgdat, struct zone *zone) } /* - * Helper functions to size the waitqueue hash table. - * Essentially these want to choose hash table sizes sufficiently - * large so that collisions trying to wait on pages are rare. - * But in fact, the number of active page waitqueues on typical - * systems is ridiculously low, less than 200. So this is even - * conservative, even though it seems large. - * - * The constant PAGES_PER_WAITQUEUE specifies the ratio of pages to - * waitqueues, i.e. the size of the waitq table given the number of pages. - */ -#define PAGES_PER_WAITQUEUE 256 - -#ifndef CONFIG_MEMORY_HOTPLUG -static inline unsigned long wait_table_hash_nr_entries(unsigned long pages) -{ - unsigned long size = 1; - - pages /= PAGES_PER_WAITQUEUE; - - while (size < pages) - size <<= 1; - - /* - * Once we have dozens or even hundreds of threads sleeping - * on IO we've got bigger problems than wait queue collision. - * Limit the size of the wait table to a reasonable size. - */ - size = min(size, 4096UL); - - return max(size, 4UL); -} -#else -/* - * A zone's size might be changed by hot-add, so it is not possible to determine - * a suitable size for its wait_table. So we use the maximum size now. - * - * The max wait table size = 4096 x sizeof(wait_queue_head_t). ie: - * - * i386 (preemption config) : 4096 x 16 = 64Kbyte. - * ia64, x86-64 (no preemption): 4096 x 20 = 80Kbyte. - * ia64, x86-64 (preemption) : 4096 x 24 = 96Kbyte. - * - * The maximum entries are prepared when a zone's memory is (512K + 256) pages - * or more by the traditional way. (See above). It equals: - * - * i386, x86-64, powerpc(4K page size) : = ( 2G + 1M)byte. - * ia64(16K page size) : = ( 8G + 4M)byte. - * powerpc (64K page size) : = (32G +16M)byte. - */ -static inline unsigned long wait_table_hash_nr_entries(unsigned long pages) -{ - return 4096UL; -} -#endif - -/* - * This is an integer logarithm so that shifts can be used later - * to extract the more random high bits from the multiplicative - * hash function before the remainder is taken. - */ -static inline unsigned long wait_table_bits(unsigned long size) -{ - return ffz(~size); -} - -/* * Initially all pages are reserved - free ones are freed * up by free_all_bootmem() once the early boot process is * done. Non-atomic initialization, single-pass. @@ -5304,49 +5277,6 @@ void __init setup_per_cpu_pageset(void) alloc_percpu(struct per_cpu_nodestat); } -static noinline __ref -int zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages) -{ - int i; - size_t alloc_size; - - /* - * The per-page waitqueue mechanism uses hashed waitqueues - * per zone. - */ - zone->wait_table_hash_nr_entries = - wait_table_hash_nr_entries(zone_size_pages); - zone->wait_table_bits = - wait_table_bits(zone->wait_table_hash_nr_entries); - alloc_size = zone->wait_table_hash_nr_entries - * sizeof(wait_queue_head_t); - - if (!slab_is_available()) { - zone->wait_table = (wait_queue_head_t *) - memblock_virt_alloc_node_nopanic( - alloc_size, zone->zone_pgdat->node_id); - } else { - /* - * This case means that a zone whose size was 0 gets new memory - * via memory hot-add. - * But it may be the case that a new node was hot-added. In - * this case vmalloc() will not be able to use this new node's - * memory - this wait_table must be initialized to use this new - * node itself as well. - * To use this new node's memory, further consideration will be - * necessary. - */ - zone->wait_table = vmalloc(alloc_size); - } - if (!zone->wait_table) - return -ENOMEM; - - for (i = 0; i < zone->wait_table_hash_nr_entries; ++i) - init_waitqueue_head(zone->wait_table + i); - - return 0; -} - static __meminit void zone_pcp_init(struct zone *zone) { /* @@ -5367,10 +5297,7 @@ int __meminit init_currently_empty_zone(struct zone *zone, unsigned long size) { struct pglist_data *pgdat = zone->zone_pgdat; - int ret; - ret = zone_wait_table_init(zone, size); - if (ret) - return ret; + pgdat->nr_zones = zone_idx(zone) + 1; zone->zone_start_pfn = zone_start_pfn; @@ -5382,6 +5309,7 @@ int __meminit init_currently_empty_zone(struct zone *zone, zone_start_pfn, (zone_start_pfn + size)); zone_init_free_lists(zone); + zone->initialized = 1; return 0; } @@ -6508,8 +6436,8 @@ unsigned long free_reserved_area(void *start, void *end, int poison, char *s) } if (pages && s) - pr_info("Freeing %s memory: %ldK (%p - %p)\n", - s, pages << (PAGE_SHIFT - 10), start, end); + pr_info("Freeing %s memory: %ldK\n", + s, pages << (PAGE_SHIFT - 10)); return pages; } @@ -6600,38 +6528,39 @@ void __init free_area_init(unsigned long *zones_size) __pa(PAGE_OFFSET) >> PAGE_SHIFT, NULL); } -static int page_alloc_cpu_notify(struct notifier_block *self, - unsigned long action, void *hcpu) +static int page_alloc_cpu_dead(unsigned int cpu) { - int cpu = (unsigned long)hcpu; - if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) { - lru_add_drain_cpu(cpu); - drain_pages(cpu); + lru_add_drain_cpu(cpu); + drain_pages(cpu); - /* - * Spill the event counters of the dead processor - * into the current processors event counters. - * This artificially elevates the count of the current - * processor. - */ - vm_events_fold_cpu(cpu); + /* + * Spill the event counters of the dead processor + * into the current processors event counters. + * This artificially elevates the count of the current + * processor. + */ + vm_events_fold_cpu(cpu); - /* - * Zero the differential counters of the dead processor - * so that the vm statistics are consistent. - * - * This is only okay since the processor is dead and cannot - * race with what we are doing. - */ - cpu_vm_stats_fold(cpu); - } - return NOTIFY_OK; + /* + * Zero the differential counters of the dead processor + * so that the vm statistics are consistent. + * + * This is only okay since the processor is dead and cannot + * race with what we are doing. + */ + cpu_vm_stats_fold(cpu); + return 0; } void __init page_alloc_init(void) { - hotcpu_notifier(page_alloc_cpu_notify, 0); + int ret; + + ret = cpuhp_setup_state_nocalls(CPUHP_PAGE_ALLOC_DEAD, + "mm/page_alloc:dead", NULL, + page_alloc_cpu_dead); + WARN_ON(ret < 0); } /* |