From 99dcc3e5a94ed491fbef402831d8c0bbb267f995 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Tue, 5 Jan 2010 15:34:51 +0900 Subject: this_cpu: Page allocator conversion Use the per cpu allocator functionality to avoid per cpu arrays in struct zone. This drastically reduces the size of struct zone for systems with large amounts of processors and allows placement of critical variables of struct zone in one cacheline even on very large systems. Another effect is that the pagesets of one processor are placed near one another. If multiple pagesets from different zones fit into one cacheline then additional cacheline fetches can be avoided on the hot paths when allocating memory from multiple zones. Bootstrap becomes simpler if we use the same scheme for UP, SMP, NUMA. #ifdefs are reduced and we can drop the zone_pcp macro. Hotplug handling is also simplified since cpu alloc can bring up and shut down cpu areas for a specific cpu as a whole. So there is no need to allocate or free individual pagesets. V7-V8: - Explain chicken egg dilemmna with percpu allocator. V4-V5: - Fix up cases where per_cpu_ptr is called before irq disable - Integrate the bootstrap logic that was separate before. tj: Build failure in pageset_cpuup_callback() due to missing ret variable fixed. Reviewed-by: Mel Gorman Signed-off-by: Christoph Lameter Signed-off-by: Tejun Heo --- mm/page_alloc.c | 202 ++++++++++++++++++++------------------------------------ 1 file changed, 71 insertions(+), 131 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 4e9f5cc5fb59..6849e870de54 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1008,10 +1008,10 @@ static void drain_pages(unsigned int cpu) struct per_cpu_pageset *pset; struct per_cpu_pages *pcp; - pset = zone_pcp(zone, cpu); + local_irq_save(flags); + pset = per_cpu_ptr(zone->pageset, cpu); pcp = &pset->pcp; - local_irq_save(flags); free_pcppages_bulk(zone, pcp->count, pcp); pcp->count = 0; local_irq_restore(flags); @@ -1095,7 +1095,6 @@ static void free_hot_cold_page(struct page *page, int cold) arch_free_page(page, 0); kernel_map_pages(page, 1, 0); - pcp = &zone_pcp(zone, get_cpu())->pcp; migratetype = get_pageblock_migratetype(page); set_page_private(page, migratetype); local_irq_save(flags); @@ -1118,6 +1117,7 @@ static void free_hot_cold_page(struct page *page, int cold) migratetype = MIGRATE_MOVABLE; } + pcp = &this_cpu_ptr(zone->pageset)->pcp; if (cold) list_add_tail(&page->lru, &pcp->lists[migratetype]); else @@ -1130,7 +1130,6 @@ static void free_hot_cold_page(struct page *page, int cold) out: local_irq_restore(flags); - put_cpu(); } void free_hot_page(struct page *page) @@ -1180,17 +1179,15 @@ struct page *buffered_rmqueue(struct zone *preferred_zone, unsigned long flags; struct page *page; int cold = !!(gfp_flags & __GFP_COLD); - int cpu; again: - cpu = get_cpu(); if (likely(order == 0)) { struct per_cpu_pages *pcp; struct list_head *list; - pcp = &zone_pcp(zone, cpu)->pcp; - list = &pcp->lists[migratetype]; local_irq_save(flags); + pcp = &this_cpu_ptr(zone->pageset)->pcp; + list = &pcp->lists[migratetype]; if (list_empty(list)) { pcp->count += rmqueue_bulk(zone, 0, pcp->batch, list, @@ -1231,7 +1228,6 @@ again: __count_zone_vm_events(PGALLOC, zone, 1 << order); zone_statistics(preferred_zone, zone); local_irq_restore(flags); - put_cpu(); VM_BUG_ON(bad_range(zone, page)); if (prep_new_page(page, order, gfp_flags)) @@ -1240,7 +1236,6 @@ again: failed: local_irq_restore(flags); - put_cpu(); return NULL; } @@ -2179,7 +2174,7 @@ void show_free_areas(void) for_each_online_cpu(cpu) { struct per_cpu_pageset *pageset; - pageset = zone_pcp(zone, cpu); + pageset = per_cpu_ptr(zone->pageset, cpu); printk("CPU %4d: hi:%5d, btch:%4d usd:%4d\n", cpu, pageset->pcp.high, @@ -2744,10 +2739,29 @@ static void build_zonelist_cache(pg_data_t *pgdat) #endif /* CONFIG_NUMA */ +/* + * Boot pageset table. One per cpu which is going to be used for all + * zones and all nodes. The parameters will be set in such a way + * that an item put on a list will immediately be handed over to + * the buddy list. This is safe since pageset manipulation is done + * with interrupts disabled. + * + * The boot_pagesets must be kept even after bootup is complete for + * unused processors and/or zones. They do play a role for bootstrapping + * hotplugged processors. + * + * zoneinfo_show() and maybe other functions do + * not check if the processor is online before following the pageset pointer. + * Other parts of the kernel may not check if the zone is available. + */ +static void setup_pageset(struct per_cpu_pageset *p, unsigned long batch); +static DEFINE_PER_CPU(struct per_cpu_pageset, boot_pageset); + /* return values int ....just for stop_machine() */ static int __build_all_zonelists(void *dummy) { int nid; + int cpu; #ifdef CONFIG_NUMA memset(node_load, 0, sizeof(node_load)); @@ -2758,6 +2772,23 @@ static int __build_all_zonelists(void *dummy) build_zonelists(pgdat); build_zonelist_cache(pgdat); } + + /* + * Initialize the boot_pagesets that are going to be used + * for bootstrapping processors. The real pagesets for + * each zone will be allocated later when the per cpu + * allocator is available. + * + * boot_pagesets are used also for bootstrapping offline + * cpus if the system is already booted because the pagesets + * are needed to initialize allocators on a specific cpu too. + * F.e. the percpu allocator needs the page allocator which + * needs the percpu allocator in order to allocate its pagesets + * (a chicken-egg dilemma). + */ + for_each_possible_cpu(cpu) + setup_pageset(&per_cpu(boot_pageset, cpu), 0); + return 0; } @@ -3095,121 +3126,33 @@ static void setup_pagelist_highmark(struct per_cpu_pageset *p, pcp->batch = PAGE_SHIFT * 8; } - -#ifdef CONFIG_NUMA -/* - * Boot pageset table. One per cpu which is going to be used for all - * zones and all nodes. The parameters will be set in such a way - * that an item put on a list will immediately be handed over to - * the buddy list. This is safe since pageset manipulation is done - * with interrupts disabled. - * - * Some NUMA counter updates may also be caught by the boot pagesets. - * - * The boot_pagesets must be kept even after bootup is complete for - * unused processors and/or zones. They do play a role for bootstrapping - * hotplugged processors. - * - * zoneinfo_show() and maybe other functions do - * not check if the processor is online before following the pageset pointer. - * Other parts of the kernel may not check if the zone is available. - */ -static struct per_cpu_pageset boot_pageset[NR_CPUS]; - /* - * Dynamically allocate memory for the - * per cpu pageset array in struct zone. + * Allocate per cpu pagesets and initialize them. + * Before this call only boot pagesets were available. + * Boot pagesets will no longer be used by this processorr + * after setup_per_cpu_pageset(). */ -static int __cpuinit process_zones(int cpu) +void __init setup_per_cpu_pageset(void) { - struct zone *zone, *dzone; - int node = cpu_to_node(cpu); - - node_set_state(node, N_CPU); /* this node has a cpu */ + struct zone *zone; + int cpu; for_each_populated_zone(zone) { - zone_pcp(zone, cpu) = kmalloc_node(sizeof(struct per_cpu_pageset), - GFP_KERNEL, node); - if (!zone_pcp(zone, cpu)) - goto bad; - - setup_pageset(zone_pcp(zone, cpu), zone_batchsize(zone)); - - if (percpu_pagelist_fraction) - setup_pagelist_highmark(zone_pcp(zone, cpu), - (zone->present_pages / percpu_pagelist_fraction)); - } - - return 0; -bad: - for_each_zone(dzone) { - if (!populated_zone(dzone)) - continue; - if (dzone == zone) - break; - kfree(zone_pcp(dzone, cpu)); - zone_pcp(dzone, cpu) = &boot_pageset[cpu]; - } - return -ENOMEM; -} + zone->pageset = alloc_percpu(struct per_cpu_pageset); -static inline void free_zone_pagesets(int cpu) -{ - struct zone *zone; - - for_each_zone(zone) { - struct per_cpu_pageset *pset = zone_pcp(zone, cpu); + for_each_possible_cpu(cpu) { + struct per_cpu_pageset *pcp = per_cpu_ptr(zone->pageset, cpu); - /* Free per_cpu_pageset if it is slab allocated */ - if (pset != &boot_pageset[cpu]) - kfree(pset); - zone_pcp(zone, cpu) = &boot_pageset[cpu]; - } -} + setup_pageset(pcp, zone_batchsize(zone)); -static int __cpuinit pageset_cpuup_callback(struct notifier_block *nfb, - unsigned long action, - void *hcpu) -{ - int cpu = (long)hcpu; - int ret = NOTIFY_OK; - - switch (action) { - case CPU_UP_PREPARE: - case CPU_UP_PREPARE_FROZEN: - if (process_zones(cpu)) - ret = NOTIFY_BAD; - break; - case CPU_UP_CANCELED: - case CPU_UP_CANCELED_FROZEN: - case CPU_DEAD: - case CPU_DEAD_FROZEN: - free_zone_pagesets(cpu); - break; - default: - break; + if (percpu_pagelist_fraction) + setup_pagelist_highmark(pcp, + (zone->present_pages / + percpu_pagelist_fraction)); + } } - return ret; } -static struct notifier_block __cpuinitdata pageset_notifier = - { &pageset_cpuup_callback, NULL, 0 }; - -void __init setup_per_cpu_pageset(void) -{ - int err; - - /* Initialize per_cpu_pageset for cpu 0. - * A cpuup callback will do this for every cpu - * as it comes online - */ - err = process_zones(smp_processor_id()); - BUG_ON(err); - register_cpu_notifier(&pageset_notifier); -} - -#endif - static noinline __init_refok int zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages) { @@ -3263,7 +3206,7 @@ static int __zone_pcp_update(void *data) struct per_cpu_pageset *pset; struct per_cpu_pages *pcp; - pset = zone_pcp(zone, cpu); + pset = per_cpu_ptr(zone->pageset, cpu); pcp = &pset->pcp; local_irq_save(flags); @@ -3281,21 +3224,17 @@ void zone_pcp_update(struct zone *zone) static __meminit void zone_pcp_init(struct zone *zone) { - int cpu; - unsigned long batch = zone_batchsize(zone); + /* + * per cpu subsystem is not up at this point. The following code + * relies on the ability of the linker to provide the + * offset of a (static) per cpu variable into the per cpu area. + */ + zone->pageset = &boot_pageset; - for (cpu = 0; cpu < NR_CPUS; cpu++) { -#ifdef CONFIG_NUMA - /* Early boot. Slab allocator not functional yet */ - zone_pcp(zone, cpu) = &boot_pageset[cpu]; - setup_pageset(&boot_pageset[cpu],0); -#else - setup_pageset(zone_pcp(zone,cpu), batch); -#endif - } if (zone->present_pages) - printk(KERN_DEBUG " %s zone: %lu pages, LIFO batch:%lu\n", - zone->name, zone->present_pages, batch); + printk(KERN_DEBUG " %s zone: %lu pages, LIFO batch:%u\n", + zone->name, zone->present_pages, + zone_batchsize(zone)); } __meminit int init_currently_empty_zone(struct zone *zone, @@ -4809,10 +4748,11 @@ int percpu_pagelist_fraction_sysctl_handler(ctl_table *table, int write, if (!write || (ret == -EINVAL)) return ret; for_each_populated_zone(zone) { - for_each_online_cpu(cpu) { + for_each_possible_cpu(cpu) { unsigned long high; high = zone->present_pages / percpu_pagelist_fraction; - setup_pagelist_highmark(zone_pcp(zone, cpu), high); + setup_pagelist_highmark( + per_cpu_ptr(zone->pageset, cpu), high); } } return 0; -- cgit v1.2.2 From d2dbe08ddceb4ba2b274abb84326d7e69d454e5c Mon Sep 17 00:00:00 2001 From: Kazuhisa Ichikawa Date: Fri, 15 Jan 2010 17:01:20 -0800 Subject: mm/page_alloc: fix the range check for backward merging The current check for 'backward merging' within add_active_range() does not seem correct. start_pfn must be compared against early_node_map[i].start_pfn (and NOT against .end_pfn) to find out whether the new region is backward-mergeable with the existing range. Signed-off-by: Kazuhisa Ichikawa Acked-by: David Rientjes Cc: KOSAKI Motohiro Cc: Mel Gorman Cc: Christoph Lameter Cc: Johannes Weiner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 4e9f5cc5fb59..6ea4966a6334 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3998,7 +3998,7 @@ void __init add_active_range(unsigned int nid, unsigned long start_pfn, } /* Merge backward if suitable */ - if (start_pfn < early_node_map[i].end_pfn && + if (start_pfn < early_node_map[i].start_pfn && end_pfn >= early_node_map[i].start_pfn) { early_node_map[i].start_pfn = start_pfn; return; -- cgit v1.2.2 From 6ccf80eb15ccaca4d3f1ab5162b9ded5eecd9971 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 15 Jan 2010 17:01:18 -0800 Subject: page allocator: update NR_FREE_PAGES only when necessary commit f2260e6b (page allocator: update NR_FREE_PAGES only as necessary) made one minor regression. if __rmqueue() was failed, NR_FREE_PAGES stat go wrong. this patch fixes it. Signed-off-by: KOSAKI Motohiro Cc: Mel Gorman Reviewed-by: Minchan Kim Reported-by: Huang Shijie Reviewed-by: Christoph Lameter Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 6ea4966a6334..d2a8889b4c58 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1222,10 +1222,10 @@ again: } spin_lock_irqsave(&zone->lock, flags); page = __rmqueue(zone, order, migratetype); - __mod_zone_page_state(zone, NR_FREE_PAGES, -(1 << order)); spin_unlock(&zone->lock); if (!page) goto failed; + __mod_zone_page_state(zone, NR_FREE_PAGES, -(1 << order)); } __count_zone_vm_events(PGALLOC, zone, 1 << order); -- cgit v1.2.2 From a7016235a61d520e6806f38129001d935c4b6661 Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Fri, 29 Jan 2010 17:46:34 +0000 Subject: mm: fix migratetype bug which slowed swapping After memory pressure has forced it to dip into the reserves, 2.6.32's 5f8dcc21211a3d4e3a7a5ca366b469fb88117f61 "page-allocator: split per-cpu list into one-list-per-migrate-type" has been returning MIGRATE_RESERVE pages to the MIGRATE_MOVABLE free_list: in some sense depleting reserves. Fix that in the most straightforward way (which, considering the overheads of alternative approaches, is Mel's preference): the right migratetype is already in page_private(page), but free_pcppages_bulk() wasn't using it. How did this bug show up? As a 20% slowdown in my tmpfs loop kbuild swapping tests, on PowerMac G5 with SLUB allocator. Bisecting to that commit was easy, but explaining the magnitude of the slowdown not easy. The same effect appears, but much less markedly, with SLAB, and even less markedly on other machines (the PowerMac divides into fewer zones than x86, I think that may be a factor). We guess that lumpy reclaim of short-lived high-order pages is implicated in some way, and probably this bug has been tickling a poor decision somewhere in page reclaim. But instrumentation hasn't told me much, I've run out of time and imagination to determine exactly what's going on, and shouldn't hold up the fix any longer: it's valid, and might even fix other misbehaviours. Signed-off-by: Hugh Dickins Acked-by: Mel Gorman Cc: stable@kernel.org Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index d2a8889b4c58..8deb9d0fd5b1 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -556,8 +556,9 @@ static void free_pcppages_bulk(struct zone *zone, int count, page = list_entry(list->prev, struct page, lru); /* must delete as __free_one_page list manipulates */ list_del(&page->lru); - __free_one_page(page, zone, 0, migratetype); - trace_mm_page_pcpu_drain(page, 0, migratetype); + /* MIGRATE_MOVABLE list may include MIGRATE_RESERVEs */ + __free_one_page(page, zone, 0, page_private(page)); + trace_mm_page_pcpu_drain(page, 0, page_private(page)); } while (--count && --batch_free && !list_empty(list)); } spin_unlock(&zone->lock); -- cgit v1.2.2 From 08677214e318297f228237be0042aac754f48f1d Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 10 Feb 2010 01:20:20 -0800 Subject: x86: Make 64 bit use early_res instead of bootmem before slab Finally we can use early_res to replace bootmem for x86_64 now. Still can use CONFIG_NO_BOOTMEM to enable it or not. -v2: fix 32bit compiling about MAX_DMA32_PFN -v3: folded bug fix from LKML message below Signed-off-by: Yinghai Lu LKML-Reference: <4B747239.4070907@kernel.org> Signed-off-by: H. Peter Anvin --- mm/page_alloc.c | 59 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 58 insertions(+), 1 deletion(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 8deb9d0fd5b1..78821a28e394 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3435,6 +3435,59 @@ void __init free_bootmem_with_active_regions(int nid, } } +int __init add_from_early_node_map(struct range *range, int az, + int nr_range, int nid) +{ + int i; + u64 start, end; + + /* need to go over early_node_map to find out good range for node */ + for_each_active_range_index_in_nid(i, nid) { + start = early_node_map[i].start_pfn; + end = early_node_map[i].end_pfn; + nr_range = add_range(range, az, nr_range, start, end); + } + return nr_range; +} + +void * __init __alloc_memory_core_early(int nid, u64 size, u64 align, + u64 goal, u64 limit) +{ + int i; + void *ptr; + + /* need to go over early_node_map to find out good range for node */ + for_each_active_range_index_in_nid(i, nid) { + u64 addr; + u64 ei_start, ei_last; + + ei_last = early_node_map[i].end_pfn; + ei_last <<= PAGE_SHIFT; + ei_start = early_node_map[i].start_pfn; + ei_start <<= PAGE_SHIFT; + addr = find_early_area(ei_start, ei_last, + goal, limit, size, align); + + if (addr == -1ULL) + continue; + +#if 0 + printk(KERN_DEBUG "alloc (nid=%d %llx - %llx) (%llx - %llx) %llx %llx => %llx\n", + nid, + ei_start, ei_last, goal, limit, size, + align, addr); +#endif + + ptr = phys_to_virt(addr); + memset(ptr, 0, size); + reserve_early_without_check(addr, addr + size, "BOOTMEM"); + return ptr; + } + + return NULL; +} + + void __init work_with_active_regions(int nid, work_fn_t work_fn, void *data) { int i; @@ -4467,7 +4520,11 @@ void __init set_dma_reserve(unsigned long new_dma_reserve) } #ifndef CONFIG_NEED_MULTIPLE_NODES -struct pglist_data __refdata contig_page_data = { .bdata = &bootmem_node_data[0] }; +struct pglist_data __refdata contig_page_data = { +#ifndef CONFIG_NO_BOOTMEM + .bdata = &bootmem_node_data[0] +#endif + }; EXPORT_SYMBOL(contig_page_data); #endif -- cgit v1.2.2 From 2ee78f7b1d8ada2615ecbcd9fea70580008bd6ce Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 17 Feb 2010 11:29:49 -0800 Subject: x86: Fix non-bootmem compilation on PowerPC These build errors on some non-x86 platforms (PowerPC for example): mm/page_alloc.c: In function '__alloc_memory_core_early': mm/page_alloc.c:3468: error: implicit declaration of function 'find_early_area' mm/page_alloc.c:3483: error: implicit declaration of function 'reserve_early_without_check' The function is only needed on CONFIG_NO_BOOTMEM. Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Johannes Weiner Cc: Mel Gorman LKML-Reference: <4B747239.4070907@kernel.org> Signed-off-by: Ingo Molnar --- mm/page_alloc.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 78821a28e394..1fa93bd2bb9f 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3450,6 +3450,7 @@ int __init add_from_early_node_map(struct range *range, int az, return nr_range; } +#ifdef CONFIG_NO_BOOTMEM void * __init __alloc_memory_core_early(int nid, u64 size, u64 align, u64 goal, u64 limit) { @@ -3486,6 +3487,7 @@ void * __init __alloc_memory_core_early(int nid, u64 size, u64 align, return NULL; } +#endif void __init work_with_active_regions(int nid, work_fn_t work_fn, void *data) -- cgit v1.2.2 From f650316c8b80fe61a31b8b575405b37cbf170459 Mon Sep 17 00:00:00 2001 From: Li Hong Date: Fri, 5 Mar 2010 13:41:52 -0800 Subject: mm/page_alloc.c: remove duplicate call to trace_mm_page_free_direct trace_mm_page_free_direct() is called in function __free_pages(). But it is called again in free_hot_page() if order == 0 and produce duplicate records in trace file for mm_page_free_direct event. As below: K-PID CPU# TIMESTAMP FUNCTION gnome-terminal-1567 [000] 4415.246466: mm_page_free_direct: page=ffffea0003db9f40 pfn=1155800 order=0 gnome-terminal-1567 [000] 4415.246468: mm_page_free_direct: page=ffffea0003db9f40 pfn=1155800 order=0 gnome-terminal-1567 [000] 4415.246506: mm_page_alloc: page=ffffea0003db9f40 pfn=1155800 order=0 migratetype=0 gfp_flags=GFP_KERNEL gnome-terminal-1567 [000] 4415.255557: mm_page_free_direct: page=ffffea0003db9f40 pfn=1155800 order=0 gnome-terminal-1567 [000] 4415.255557: mm_page_free_direct: page=ffffea0003db9f40 pfn=1155800 order=0 This patch removes the first call and adds a call to trace_mm_page_free_direct() in __free_pages_ok(). Signed-off-by: Li Hong Cc: Mel Gorman Cc: Rik van Riel Cc: Ingo Molnar Cc: Larry Woodman Cc: Peter Zijlstra Cc: Li Ming Chun Cc: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index a6b17aa4740b..ee37091b191b 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -583,6 +583,7 @@ static void __free_pages_ok(struct page *page, unsigned int order) int bad = 0; int wasMlocked = __TestClearPageMlocked(page); + trace_mm_page_free_direct(page, order); kmemcheck_free_shadow(page, order); for (i = 0 ; i < (1 << order) ; ++i) @@ -2008,7 +2009,6 @@ void __pagevec_free(struct pagevec *pvec) void __free_pages(struct page *page, unsigned int order) { if (put_page_testzero(page)) { - trace_mm_page_free_direct(page, order); if (order == 0) free_hot_page(page); else -- cgit v1.2.2 From c475dab63ae798d81fb597a6a1859986b296d9d0 Mon Sep 17 00:00:00 2001 From: Li Hong Date: Fri, 5 Mar 2010 13:41:53 -0800 Subject: mm/page_alloc.c: adjust a call site to trace_mm_page_free_direct Move a call of trace_mm_page_free_direct() from free_hot_page() to free_hot_cold_page(). It is clearer and close to kmemcheck_free_shadow(), as it is done in function __free_pages_ok(). Signed-off-by: Li Hong Cc: Mel Gorman Cc: Rik van Riel Cc: Ingo Molnar Cc: Larry Woodman Cc: Peter Zijlstra Cc: Li Ming Chun Cc: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index ee37091b191b..caa7df60a4a1 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1083,6 +1083,7 @@ static void free_hot_cold_page(struct page *page, int cold) int migratetype; int wasMlocked = __TestClearPageMlocked(page); + trace_mm_page_free_direct(page, 0); kmemcheck_free_shadow(page, 0); if (PageAnon(page)) @@ -1136,7 +1137,6 @@ out: void free_hot_page(struct page *page) { - trace_mm_page_free_direct(page, 0); free_hot_cold_page(page, 0); } -- cgit v1.2.2 From fc91668eaf9e7ba61e867fc2218b7e9fb67faa4f Mon Sep 17 00:00:00 2001 From: Li Hong Date: Fri, 5 Mar 2010 13:41:54 -0800 Subject: mm: remove free_hot_page() free_hot_page() is just a wrapper around free_hot_cold_page() with parameter 'cold = 0'. After adding a clear comment for free_hot_cold_page(), it is reasonable to remove a level of call. [akpm@linux-foundation.org: fix build] Signed-off-by: Li Hong Cc: Mel Gorman Cc: Rik van Riel Cc: Ingo Molnar Cc: Larry Woodman Cc: Peter Zijlstra Cc: Li Ming Chun Cc: KOSAKI Motohiro Cc: Americo Wang Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index caa7df60a4a1..80bcee0c5034 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1074,8 +1074,9 @@ void mark_free_pages(struct zone *zone) /* * Free a 0-order page + * cold == 1 ? free a cold page : free a hot page */ -static void free_hot_cold_page(struct page *page, int cold) +void free_hot_cold_page(struct page *page, int cold) { struct zone *zone = page_zone(page); struct per_cpu_pages *pcp; @@ -1135,11 +1136,6 @@ out: local_irq_restore(flags); } -void free_hot_page(struct page *page) -{ - free_hot_cold_page(page, 0); -} - /* * split_page takes a non-compound higher-order page, and splits it into * n (1< Date: Fri, 5 Mar 2010 13:41:55 -0800 Subject: mm: restore zone->all_unreclaimable to independence word commit e815af95 ("change all_unreclaimable zone member to flags") changed all_unreclaimable member to bit flag. But it had an undesireble side effect. free_one_page() is one of most hot path in linux kernel and increasing atomic ops in it can reduce kernel performance a bit. Thus, this patch revert such commit partially. at least all_unreclaimable shouldn't share memory word with other zone flags. [akpm@linux-foundation.org: fix patch interaction] Signed-off-by: KOSAKI Motohiro Cc: David Rientjes Cc: Wu Fengguang Cc: KAMEZAWA Hiroyuki Cc: Minchan Kim Cc: Huang Shijie Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 80bcee0c5034..0734bedabd9c 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -530,7 +530,7 @@ static void free_pcppages_bulk(struct zone *zone, int count, int batch_free = 0; spin_lock(&zone->lock); - zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE); + zone->all_unreclaimable = 0; zone->pages_scanned = 0; __mod_zone_page_state(zone, NR_FREE_PAGES, count); @@ -568,7 +568,7 @@ static void free_one_page(struct zone *zone, struct page *page, int order, int migratetype) { spin_lock(&zone->lock); - zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE); + zone->all_unreclaimable = 0; zone->pages_scanned = 0; __mod_zone_page_state(zone, NR_FREE_PAGES, 1 << order); @@ -2262,7 +2262,7 @@ void show_free_areas(void) K(zone_page_state(zone, NR_BOUNCE)), K(zone_page_state(zone, NR_WRITEBACK_TEMP)), zone->pages_scanned, - (zone_is_all_unreclaimable(zone) ? "yes" : "no") + (zone->all_unreclaimable ? "yes" : "no") ); printk("lowmem_reserve[]:"); for (i = 0; i < MAX_NR_ZONES; i++) -- cgit v1.2.2 From 452aa6999e6703ffbddd7f6ea124d3968915f3e3 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 5 Mar 2010 13:42:13 -0800 Subject: mm/pm: force GFP_NOIO during suspend/hibernation and resume There are quite a few GFP_KERNEL memory allocations made during suspend/hibernation and resume that may cause the system to hang, because the I/O operations they depend on cannot be completed due to the underlying devices being suspended. Avoid this problem by clearing the __GFP_IO and __GFP_FS bits in gfp_allowed_mask before suspend/hibernation and restoring the original values of these bits in gfp_allowed_mask durig the subsequent resume. [akpm@linux-foundation.org: fix CONFIG_PM=n linkage] Signed-off-by: Rafael J. Wysocki Reported-by: Maxim Levitsky Cc: Sebastian Ott Cc: Benjamin Herrenschmidt Cc: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 0734bedabd9c..298f307c63a1 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -76,6 +76,31 @@ unsigned long totalreserve_pages __read_mostly; int percpu_pagelist_fraction; gfp_t gfp_allowed_mask __read_mostly = GFP_BOOT_MASK; +#ifdef CONFIG_PM_SLEEP +/* + * The following functions are used by the suspend/hibernate code to temporarily + * change gfp_allowed_mask in order to avoid using I/O during memory allocations + * while devices are suspended. To avoid races with the suspend/hibernate code, + * they should always be called with pm_mutex held (gfp_allowed_mask also should + * only be modified with pm_mutex held, unless the suspend/hibernate code is + * guaranteed not to run in parallel with that modification). + */ +void set_gfp_allowed_mask(gfp_t mask) +{ + WARN_ON(!mutex_is_locked(&pm_mutex)); + gfp_allowed_mask = mask; +} + +gfp_t clear_gfp_allowed_mask(gfp_t mask) +{ + gfp_t ret = gfp_allowed_mask; + + WARN_ON(!mutex_is_locked(&pm_mutex)); + gfp_allowed_mask &= ~mask; + return ret; +} +#endif /* CONFIG_PM_SLEEP */ + #ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE int pageblock_order __read_mostly; #endif -- cgit v1.2.2 From 72f0ba0252e7177965255ed2c663be126b6b5f91 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Fri, 5 Mar 2010 13:42:14 -0800 Subject: mm: suppress pfn range output for zones without pages free_area_init_nodes() emits pfn ranges for all zones on the system. There may be no pages on a higher zone, however, due to memory limitations or the use of the mem= kernel parameter. For example: Zone PFN ranges: DMA 0x00000001 -> 0x00001000 DMA32 0x00001000 -> 0x00100000 Normal 0x00100000 -> 0x00100000 The implementation copies the previous zone's highest pfn, if any, as the next zone's lowest pfn. If its highest pfn is then greater than the amount of addressable memory, the upper memory limit is used instead. Thus, both the lowest and highest possible pfn for higher zones without memory may be the same. The pfn range for zones without memory is now shown as "empty" instead. Signed-off-by: David Rientjes Cc: Mel Gorman Reviewed-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 298f307c63a1..a8182c89de59 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -4392,8 +4392,12 @@ void __init free_area_init_nodes(unsigned long *max_zone_pfn) for (i = 0; i < MAX_NR_ZONES; i++) { if (i == ZONE_MOVABLE) continue; - printk(" %-8s %0#10lx -> %0#10lx\n", - zone_names[i], + printk(" %-8s ", zone_names[i]); + if (arch_zone_lowest_possible_pfn[i] == + arch_zone_highest_possible_pfn[i]) + printk("empty\n"); + else + printk("%0#10lx -> %0#10lx\n", arch_zone_lowest_possible_pfn[i], arch_zone_highest_possible_pfn[i]); } -- cgit v1.2.2 From 2d30a1f6315b8940537e8e98882c6038fbac9ba5 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 10 Mar 2010 15:20:40 -0800 Subject: mm: do not iterate over NR_CPUS in __zone_pcp_update() __zone_pcp_update() iterates over NR_CPUS instead of limiting the access to the possible cpus. This might result in access to uninitialized areas as the per cpu allocator only populates the per cpu memory for possible cpus. This problem was created as a result of the dynamic allocation of pagesets from percpu memory that went in during the merge window - commit 99dcc3e5a94ed491fbef402831d8c0bbb267f995 ("this_cpu: Page allocator conversion"). Signed-off-by: Thomas Gleixner Acked-by: Pekka Enberg Acked-by: Tejun Heo Acked-by: Christoph Lameter Acked-by: Mel Gorman Reviewed-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index a8182c89de59..78ce90dd671f 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3224,7 +3224,7 @@ static int __zone_pcp_update(void *data) int cpu; unsigned long batch = zone_batchsize(zone), flags; - for (cpu = 0; cpu < NR_CPUS; cpu++) { + for_each_possible_cpu(cpu) { struct per_cpu_pageset *pset; struct per_cpu_pages *pcp; -- cgit v1.2.2 From 718a38211bf4375c0a1efad3afbc5dbaef5d33f9 Mon Sep 17 00:00:00 2001 From: Wu Fengguang Date: Wed, 10 Mar 2010 15:20:43 -0800 Subject: mm: introduce dump_page() and print symbolic flag names - introduce dump_page() to print the page info for debugging some error condition. - convert three mm users: bad_page(), print_bad_pte() and memory offline failure. - print an extra field: the symbolic names of page->flags Example dump_page() output: [ 157.521694] page:ffffea0000a7cba8 count:2 mapcount:1 mapping:ffff88001c901791 index:0x147 [ 157.525570] page flags: 0x100000000100068(uptodate|lru|active|swapbacked) Signed-off-by: Wu Fengguang Cc: Ingo Molnar Cc: Alex Chiang Cc: Rik van Riel Cc: Andi Kleen Cc: Mel Gorman Cc: Christoph Lameter Cc: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 83 ++++++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 79 insertions(+), 4 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 78ce90dd671f..d03c946d5566 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -50,6 +50,7 @@ #include #include #include +#include #include #include @@ -288,10 +289,7 @@ static void bad_page(struct page *page) printk(KERN_ALERT "BUG: Bad page state in process %s pfn:%05lx\n", current->comm, page_to_pfn(page)); - printk(KERN_ALERT - "page:%p flags:%p count:%d mapcount:%d mapping:%p index:%lx\n", - page, (void *)page->flags, page_count(page), - page_mapcount(page), page->mapping, page->index); + dump_page(page); dump_stack(); out: @@ -5183,3 +5181,80 @@ bool is_free_buddy_page(struct page *page) return order < MAX_ORDER; } #endif + +static struct trace_print_flags pageflag_names[] = { + {1UL << PG_locked, "locked" }, + {1UL << PG_error, "error" }, + {1UL << PG_referenced, "referenced" }, + {1UL << PG_uptodate, "uptodate" }, + {1UL << PG_dirty, "dirty" }, + {1UL << PG_lru, "lru" }, + {1UL << PG_active, "active" }, + {1UL << PG_slab, "slab" }, + {1UL << PG_owner_priv_1, "owner_priv_1" }, + {1UL << PG_arch_1, "arch_1" }, + {1UL << PG_reserved, "reserved" }, + {1UL << PG_private, "private" }, + {1UL << PG_private_2, "private_2" }, + {1UL << PG_writeback, "writeback" }, +#ifdef CONFIG_PAGEFLAGS_EXTENDED + {1UL << PG_head, "head" }, + {1UL << PG_tail, "tail" }, +#else + {1UL << PG_compound, "compound" }, +#endif + {1UL << PG_swapcache, "swapcache" }, + {1UL << PG_mappedtodisk, "mappedtodisk" }, + {1UL << PG_reclaim, "reclaim" }, + {1UL << PG_buddy, "buddy" }, + {1UL << PG_swapbacked, "swapbacked" }, + {1UL << PG_unevictable, "unevictable" }, +#ifdef CONFIG_MMU + {1UL << PG_mlocked, "mlocked" }, +#endif +#ifdef CONFIG_ARCH_USES_PG_UNCACHED + {1UL << PG_uncached, "uncached" }, +#endif +#ifdef CONFIG_MEMORY_FAILURE + {1UL << PG_hwpoison, "hwpoison" }, +#endif + {-1UL, NULL }, +}; + +static void dump_page_flags(unsigned long flags) +{ + const char *delim = ""; + unsigned long mask; + int i; + + printk(KERN_ALERT "page flags: %#lx(", flags); + + /* remove zone id */ + flags &= (1UL << NR_PAGEFLAGS) - 1; + + for (i = 0; pageflag_names[i].name && flags; i++) { + + mask = pageflag_names[i].mask; + if ((flags & mask) != mask) + continue; + + flags &= ~mask; + printk("%s%s", delim, pageflag_names[i].name); + delim = "|"; + } + + /* check for left over flags */ + if (flags) + printk("%s%#lx", delim, flags); + + printk(")\n"); +} + +void dump_page(struct page *page) +{ + printk(KERN_ALERT + "page:%p count:%d mapcount:%d mapping:%p index:%#lx\n", + page, page_count(page), page_mapcount(page), + page->mapping, page->index); + dump_page_flags(page->flags); +} -- cgit v1.2.2