From cb2b95e1c6b56e3d2369d3a5f4bc97f4fa180683 Mon Sep 17 00:00:00 2001 From: Andy Whitcroft Date: Fri, 23 Jun 2006 02:03:01 -0700 Subject: [PATCH] zone handle unaligned zone boundaries The buddy allocator has a requirement that boundaries between contigious zones occur aligned with the the MAX_ORDER ranges. Where they do not we will incorrectly merge pages cross zone boundaries. This can lead to pages from the wrong zone being handed out. Originally the buddy allocator would check that buddies were in the same zone by referencing the zone start and end page frame numbers. This was removed as it became very expensive and the buddy allocator already made the assumption that zones boundaries were aligned. It is clear that not all configurations and architectures are honouring this alignment requirement. Therefore it seems safest to reintroduce support for non-aligned zone boundaries. This patch introduces a new check when considering a page a buddy it compares the zone_table index for the two pages and refuses to merge the pages where they do not match. The zone_table index is unique for each node/zone combination when FLATMEM/DISCONTIGMEM is enabled and for each section/zone combination when SPARSEMEM is enabled (a SPARSEMEM section is at least a MAX_ORDER size). Signed-off-by: Andy Whitcroft Cc: Dave Hansen Cc: Mel Gorman Cc: Yasunori Goto Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 253a450c400..fd631c2536a 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -286,22 +286,27 @@ __find_combined_index(unsigned long page_idx, unsigned int order) * we can do coalesce a page and its buddy if * (a) the buddy is not in a hole && * (b) the buddy is in the buddy system && - * (c) a page and its buddy have the same order. + * (c) a page and its buddy have the same order && + * (d) a page and its buddy are in the same zone. * * For recording whether a page is in the buddy system, we use PG_buddy. * Setting, clearing, and testing PG_buddy is serialized by zone->lock. * * For recording page's order, we use page_private(page). */ -static inline int page_is_buddy(struct page *page, int order) +static inline int page_is_buddy(struct page *page, struct page *buddy, + int order) { #ifdef CONFIG_HOLES_IN_ZONE - if (!pfn_valid(page_to_pfn(page))) + if (!pfn_valid(page_to_pfn(buddy))) return 0; #endif - if (PageBuddy(page) && page_order(page) == order) { - BUG_ON(page_count(page) != 0); + if (page_zone_id(page) != page_zone_id(buddy)) + return 0; + + if (PageBuddy(buddy) && page_order(buddy) == order) { + BUG_ON(page_count(buddy) != 0); return 1; } return 0; @@ -352,7 +357,7 @@ static inline void __free_one_page(struct page *page, struct page *buddy; buddy = __page_find_buddy(page, page_idx, order); - if (!page_is_buddy(buddy, order)) + if (!page_is_buddy(page, buddy, order)) break; /* Move the buddy up one level. */ list_del(&buddy->lru); -- cgit v1.2.3-70-g09d2 From 02b694dea473ad3db1e2d1b14c1fef8fbd92e5e6 Mon Sep 17 00:00:00 2001 From: Yasunori Goto Date: Fri, 23 Jun 2006 02:03:08 -0700 Subject: [PATCH] wait_table and zonelist initializing for memory hotadd: change name of wait_table_size() This is just to rename from wait_table_size() to wait_table_hash_nr_entries(). Signed-off-by: Yasunori Goto Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/mmzone.h | 4 ++-- mm/page_alloc.c | 12 +++++++----- 2 files changed, 9 insertions(+), 7 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 9742e3c1622..652673ea92f 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -197,7 +197,7 @@ struct zone { /* * wait_table -- the array holding the hash table - * wait_table_size -- the size of the hash table array + * wait_table_hash_nr_entries -- the size of the hash table array * wait_table_bits -- wait_table_size == (1 << wait_table_bits) * * The purpose of all these is to keep track of the people @@ -220,7 +220,7 @@ struct zone { * free_area_init_core() performs the initialization of them. */ wait_queue_head_t * wait_table; - unsigned long wait_table_size; + unsigned long wait_table_hash_nr_entries; unsigned long wait_table_bits; /* diff --git a/mm/page_alloc.c b/mm/page_alloc.c index fd631c2536a..27320a0542d 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1727,7 +1727,7 @@ void __init build_all_zonelists(void) */ #define PAGES_PER_WAITQUEUE 256 -static inline unsigned long wait_table_size(unsigned long pages) +static inline unsigned long wait_table_hash_nr_entries(unsigned long pages) { unsigned long size = 1; @@ -2019,13 +2019,15 @@ void zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages) * The per-page waitqueue mechanism uses hashed waitqueues * per zone. */ - zone->wait_table_size = wait_table_size(zone_size_pages); - zone->wait_table_bits = wait_table_bits(zone->wait_table_size); + 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); zone->wait_table = (wait_queue_head_t *) - alloc_bootmem_node(pgdat, zone->wait_table_size + alloc_bootmem_node(pgdat, zone->wait_table_hash_nr_entries * sizeof(wait_queue_head_t)); - for(i = 0; i < zone->wait_table_size; ++i) + for(i = 0; i < zone->wait_table_hash_nr_entries; ++i) init_waitqueue_head(zone->wait_table + i); } -- cgit v1.2.3-70-g09d2 From 86356ab147669bd3bcb2149fd9561d1280835c24 Mon Sep 17 00:00:00 2001 From: Yasunori Goto Date: Fri, 23 Jun 2006 02:03:09 -0700 Subject: [PATCH] wait_table and zonelist initializing for memory hotadd: change to meminit for build_zonelist Change definitions of some functions and data from __init to __meminit. These functions and data can be used after bootup by this patch to be used for hot-add codes. Signed-off-by: Yasunori Goto Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/bootmem.h | 4 ++-- mm/page_alloc.c | 18 +++++++++--------- 2 files changed, 11 insertions(+), 11 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h index da2d107fe2c..22866fa2d96 100644 --- a/include/linux/bootmem.h +++ b/include/linux/bootmem.h @@ -91,8 +91,8 @@ static inline void *alloc_remap(int nid, unsigned long size) } #endif -extern unsigned long __initdata nr_kernel_pages; -extern unsigned long __initdata nr_all_pages; +extern unsigned long nr_kernel_pages; +extern unsigned long nr_all_pages; extern void *__init alloc_large_system_hash(const char *tablename, unsigned long bucketsize, diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 27320a0542d..5ae75bead4d 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -83,8 +83,8 @@ EXPORT_SYMBOL(zone_table); static char *zone_names[MAX_NR_ZONES] = { "DMA", "DMA32", "Normal", "HighMem" }; int min_free_kbytes = 1024; -unsigned long __initdata nr_kernel_pages; -unsigned long __initdata nr_all_pages; +unsigned long __meminitdata nr_kernel_pages; +unsigned long __meminitdata nr_all_pages; #ifdef CONFIG_DEBUG_VM static int page_outside_zone_boundaries(struct zone *zone, struct page *page) @@ -1517,7 +1517,7 @@ void show_free_areas(void) * * Add all populated zones of a node to the zonelist. */ -static int __init build_zonelists_node(pg_data_t *pgdat, +static int __meminit build_zonelists_node(pg_data_t *pgdat, struct zonelist *zonelist, int nr_zones, int zone_type) { struct zone *zone; @@ -1553,7 +1553,7 @@ static inline int highest_zone(int zone_bits) #ifdef CONFIG_NUMA #define MAX_NODE_LOAD (num_online_nodes()) -static int __initdata node_load[MAX_NUMNODES]; +static int __meminitdata node_load[MAX_NUMNODES]; /** * find_next_best_node - find the next node that should appear in a given node's fallback list * @node: node whose fallback list we're appending @@ -1568,7 +1568,7 @@ static int __initdata node_load[MAX_NUMNODES]; * on them otherwise. * It returns -1 if no node is found. */ -static int __init find_next_best_node(int node, nodemask_t *used_node_mask) +static int __meminit find_next_best_node(int node, nodemask_t *used_node_mask) { int n, val; int min_val = INT_MAX; @@ -1614,7 +1614,7 @@ static int __init find_next_best_node(int node, nodemask_t *used_node_mask) return best_node; } -static void __init build_zonelists(pg_data_t *pgdat) +static void __meminit build_zonelists(pg_data_t *pgdat) { int i, j, k, node, local_node; int prev_node, load; @@ -1666,7 +1666,7 @@ static void __init build_zonelists(pg_data_t *pgdat) #else /* CONFIG_NUMA */ -static void __init build_zonelists(pg_data_t *pgdat) +static void __meminit build_zonelists(pg_data_t *pgdat) { int i, j, k, node, local_node; @@ -2071,7 +2071,7 @@ static __meminit void init_currently_empty_zone(struct zone *zone, * - mark all memory queues empty * - clear the memory bitmaps */ -static void __init free_area_init_core(struct pglist_data *pgdat, +static void __meminit free_area_init_core(struct pglist_data *pgdat, unsigned long *zones_size, unsigned long *zholes_size) { unsigned long j; @@ -2159,7 +2159,7 @@ static void __init alloc_node_mem_map(struct pglist_data *pgdat) #endif /* CONFIG_FLAT_NODE_MEM_MAP */ } -void __init free_area_init_node(int nid, struct pglist_data *pgdat, +void __meminit free_area_init_node(int nid, struct pglist_data *pgdat, unsigned long *zones_size, unsigned long node_start_pfn, unsigned long *zholes_size) { -- cgit v1.2.3-70-g09d2 From 718127cc3170454f4aa274fdd2f1e01574fecd66 Mon Sep 17 00:00:00 2001 From: Yasunori Goto Date: Fri, 23 Jun 2006 02:03:10 -0700 Subject: [PATCH] wait_table and zonelist initializing for memory hotadd: add return code for init_current_empty_zone When add_zone() is called against empty zone (not populated zone), we have to initialize the zone which didn't initialize at boot time. But, init_currently_empty_zone() may fail due to allocation of wait table. So, this patch is to catch its error code. Changes against wait_table is in the next patch. Signed-off-by: KAMEZAWA Hiroyuki Signed-off-by: Yasunori Goto Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/mmzone.h | 3 +++ mm/memory_hotplug.c | 15 +++++++++++++-- mm/page_alloc.c | 11 ++++++++--- 3 files changed, 24 insertions(+), 5 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 652673ea92f..e82fc1a52cd 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -333,6 +333,9 @@ void wakeup_kswapd(struct zone *zone, int order); int zone_watermark_ok(struct zone *z, int order, unsigned long mark, int classzone_idx, int alloc_flags); +extern int init_currently_empty_zone(struct zone *zone, unsigned long start_pfn, + unsigned long size); + #ifdef CONFIG_HAVE_MEMORY_PRESENT void memory_present(int nid, unsigned long start, unsigned long end); #else diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index 70df5c0d957..71da5c98c9c 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c @@ -26,7 +26,7 @@ extern void zonetable_add(struct zone *zone, int nid, int zid, unsigned long pfn, unsigned long size); -static void __add_zone(struct zone *zone, unsigned long phys_start_pfn) +static int __add_zone(struct zone *zone, unsigned long phys_start_pfn) { struct pglist_data *pgdat = zone->zone_pgdat; int nr_pages = PAGES_PER_SECTION; @@ -34,8 +34,15 @@ static void __add_zone(struct zone *zone, unsigned long phys_start_pfn) int zone_type; zone_type = zone - pgdat->node_zones; + if (!populated_zone(zone)) { + int ret = 0; + ret = init_currently_empty_zone(zone, phys_start_pfn, nr_pages); + if (ret < 0) + return ret; + } memmap_init_zone(nr_pages, nid, zone_type, phys_start_pfn); zonetable_add(zone, nid, zone_type, phys_start_pfn, nr_pages); + return 0; } extern int sparse_add_one_section(struct zone *zone, unsigned long start_pfn, @@ -50,7 +57,11 @@ static int __add_section(struct zone *zone, unsigned long phys_start_pfn) if (ret < 0) return ret; - __add_zone(zone, phys_start_pfn); + ret = __add_zone(zone, phys_start_pfn); + + if (ret < 0) + return ret; + return register_new_memory(__pfn_to_section(phys_start_pfn)); } diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 5ae75bead4d..4bc66f6b771 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -2050,8 +2050,9 @@ static __meminit void zone_pcp_init(struct zone *zone) zone->name, zone->present_pages, batch); } -static __meminit void init_currently_empty_zone(struct zone *zone, - unsigned long zone_start_pfn, unsigned long size) +__meminit int init_currently_empty_zone(struct zone *zone, + unsigned long zone_start_pfn, + unsigned long size) { struct pglist_data *pgdat = zone->zone_pgdat; @@ -2063,6 +2064,8 @@ static __meminit void init_currently_empty_zone(struct zone *zone, memmap_init(size, pgdat->node_id, zone_idx(zone), zone_start_pfn); zone_init_free_lists(pgdat, zone, zone->spanned_pages); + + return 0; } /* @@ -2077,6 +2080,7 @@ static void __meminit free_area_init_core(struct pglist_data *pgdat, unsigned long j; int nid = pgdat->node_id; unsigned long zone_start_pfn = pgdat->node_start_pfn; + int ret; pgdat_resize_init(pgdat); pgdat->nr_zones = 0; @@ -2118,7 +2122,8 @@ static void __meminit free_area_init_core(struct pglist_data *pgdat, continue; zonetable_add(zone, nid, j, zone_start_pfn, size); - init_currently_empty_zone(zone, zone_start_pfn, size); + ret = init_currently_empty_zone(zone, zone_start_pfn, size); + BUG_ON(ret); zone_start_pfn += size; } } -- cgit v1.2.3-70-g09d2 From cca448fe92246fb59efe55ba2e048ded0971a9af Mon Sep 17 00:00:00 2001 From: Yasunori Goto Date: Fri, 23 Jun 2006 02:03:10 -0700 Subject: [PATCH] wait_table and zonelist initializing for memory hotadd: wait_table initialization Wait_table is initialized according to zone size at boot time. But, we cannot know the maixmum zone size when memory hotplug is enabled. It can be changed.... And resizing of wait_table is hard. So kernel allocate and initialzie wait_table as its maximum size. Signed-off-by: KAMEZAWA Hiroyuki Signed-off-by: Yasunori Goto Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 59 +++++++++++++++++++++++++++++++++++++++++++++++++++------ 1 file changed, 53 insertions(+), 6 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 4bc66f6b771..62564e27b44 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1727,6 +1727,7 @@ void __init build_all_zonelists(void) */ #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; @@ -1745,6 +1746,29 @@ static inline unsigned long wait_table_hash_nr_entries(unsigned long pages) 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 @@ -2010,10 +2034,11 @@ void __init setup_per_cpu_pageset(void) #endif static __meminit -void zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages) +int zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages) { int i; struct pglist_data *pgdat = zone->zone_pgdat; + size_t alloc_size; /* * The per-page waitqueue mechanism uses hashed waitqueues @@ -2023,12 +2048,32 @@ void zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages) wait_table_hash_nr_entries(zone_size_pages); zone->wait_table_bits = wait_table_bits(zone->wait_table_hash_nr_entries); - zone->wait_table = (wait_queue_head_t *) - alloc_bootmem_node(pgdat, zone->wait_table_hash_nr_entries - * sizeof(wait_queue_head_t)); + alloc_size = zone->wait_table_hash_nr_entries + * sizeof(wait_queue_head_t); + + if (system_state == SYSTEM_BOOTING) { + zone->wait_table = (wait_queue_head_t *) + alloc_bootmem_node(pgdat, alloc_size); + } 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 = (wait_queue_head_t *)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) @@ -2055,8 +2100,10 @@ __meminit int init_currently_empty_zone(struct zone *zone, unsigned long size) { struct pglist_data *pgdat = zone->zone_pgdat; - - zone_wait_table_init(zone, size); + 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; -- cgit v1.2.3-70-g09d2 From 6811378e7d8b9aa4fca2a1ca73d24c9d67c9cb12 Mon Sep 17 00:00:00 2001 From: Yasunori Goto Date: Fri, 23 Jun 2006 02:03:11 -0700 Subject: [PATCH] wait_table and zonelist initializing for memory hotadd: update zonelists In current code, zonelist is considered to be build once, no modification. But MemoryHotplug can add new zone/pgdat. It must be updated. This patch modifies build_all_zonelists(). By this, build_all_zonelist() can reconfig pgdat's zonelists. To update them safety, this patch use stop_machine_run(). Other cpus don't touch among updating them by using it. In old version (V2 of node hotadd), kernel updated them after zone initialization. But present_page of its new zone is still 0, because online_page() is not called yet at this time. Build_zonelists() checks present_pages to find present zone. It was too early. So, I changed it after online_pages(). Signed-off-by: Yasunori Goto Signed-off-by: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/memory_hotplug.c | 12 ++++++++++++ mm/page_alloc.c | 26 +++++++++++++++++++++----- 2 files changed, 33 insertions(+), 5 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index 71da5c98c9c..1b1ac3db218 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c @@ -127,6 +127,7 @@ int online_pages(unsigned long pfn, unsigned long nr_pages) unsigned long flags; unsigned long onlined_pages = 0; struct zone *zone; + int need_zonelists_rebuild = 0; /* * This doesn't need a lock to do pfn_to_page(). @@ -139,6 +140,14 @@ int online_pages(unsigned long pfn, unsigned long nr_pages) grow_pgdat_span(zone->zone_pgdat, pfn, pfn + nr_pages); pgdat_resize_unlock(zone->zone_pgdat, &flags); + /* + * If this zone is not populated, then it is not in zonelist. + * This means the page allocator ignores this zone. + * So, zonelist must be updated after online. + */ + if (!populated_zone(zone)) + need_zonelists_rebuild = 1; + for (i = 0; i < nr_pages; i++) { struct page *page = pfn_to_page(pfn + i); online_page(page); @@ -149,5 +158,8 @@ int online_pages(unsigned long pfn, unsigned long nr_pages) setup_per_zone_pages_min(); + if (need_zonelists_rebuild) + build_all_zonelists(); + return 0; } diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 62564e27b44..9dfbe6b7d1c 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -37,6 +37,7 @@ #include #include #include +#include #include #include @@ -1704,14 +1705,29 @@ static void __meminit build_zonelists(pg_data_t *pgdat) #endif /* CONFIG_NUMA */ -void __init build_all_zonelists(void) +/* return values int ....just for stop_machine_run() */ +static int __meminit __build_all_zonelists(void *dummy) { - int i; + int nid; + for_each_online_node(nid) + build_zonelists(NODE_DATA(nid)); + return 0; +} + +void __meminit build_all_zonelists(void) +{ + if (system_state == SYSTEM_BOOTING) { + __build_all_zonelists(0); + cpuset_init_current_mems_allowed(); + } else { + /* we have to stop all cpus to guaranntee there is no user + of zonelist */ + stop_machine_run(__build_all_zonelists, NULL, NR_CPUS); + /* cpuset refresh routine should be here */ + } - for_each_online_node(i) - build_zonelists(NODE_DATA(i)); printk("Built %i zonelists\n", num_online_nodes()); - cpuset_init_current_mems_allowed(); + } /* -- cgit v1.2.3-70-g09d2 From 67de648211fa041fe08a0c25241a4980bbb90698 Mon Sep 17 00:00:00 2001 From: Andy Whitcroft Date: Fri, 23 Jun 2006 02:03:12 -0700 Subject: [PATCH] squash duplicate page_to_pfn and pfn_to_page We have architectures where the size of page_to_pfn and pfn_to_page are significant enough to overall image size that they wish to push them out of line. However, in the process we have grown a second copy of the implementation of each of these routines for each memory model. Share the implmentation exposing it either inline or out-of-line as required. Signed-off-by: Andy Whitcroft Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/asm-generic/memory_model.h | 27 +++++++++++++++------------ mm/page_alloc.c | 32 ++------------------------------ 2 files changed, 17 insertions(+), 42 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/include/asm-generic/memory_model.h b/include/asm-generic/memory_model.h index 0cfb086dd37..8078cbd2c01 100644 --- a/include/asm-generic/memory_model.h +++ b/include/asm-generic/memory_model.h @@ -23,29 +23,23 @@ #endif /* CONFIG_DISCONTIGMEM */ -#ifdef CONFIG_OUT_OF_LINE_PFN_TO_PAGE -struct page; -/* this is useful when inlined pfn_to_page is too big */ -extern struct page *pfn_to_page(unsigned long pfn); -extern unsigned long page_to_pfn(struct page *page); -#else /* * supports 3 memory models. */ #if defined(CONFIG_FLATMEM) -#define pfn_to_page(pfn) (mem_map + ((pfn) - ARCH_PFN_OFFSET)) -#define page_to_pfn(page) ((unsigned long)((page) - mem_map) + \ +#define __pfn_to_page(pfn) (mem_map + ((pfn) - ARCH_PFN_OFFSET)) +#define __page_to_pfn(page) ((unsigned long)((page) - mem_map) + \ ARCH_PFN_OFFSET) #elif defined(CONFIG_DISCONTIGMEM) -#define pfn_to_page(pfn) \ +#define __pfn_to_page(pfn) \ ({ unsigned long __pfn = (pfn); \ unsigned long __nid = arch_pfn_to_nid(pfn); \ NODE_DATA(__nid)->node_mem_map + arch_local_page_offset(__pfn, __nid);\ }) -#define page_to_pfn(pg) \ +#define __page_to_pfn(pg) \ ({ struct page *__pg = (pg); \ struct pglist_data *__pgdat = NODE_DATA(page_to_nid(__pg)); \ (unsigned long)(__pg - __pgdat->node_mem_map) + \ @@ -57,18 +51,27 @@ extern unsigned long page_to_pfn(struct page *page); * Note: section's mem_map is encorded to reflect its start_pfn. * section[i].section_mem_map == mem_map's address - start_pfn; */ -#define page_to_pfn(pg) \ +#define __page_to_pfn(pg) \ ({ struct page *__pg = (pg); \ int __sec = page_to_section(__pg); \ __pg - __section_mem_map_addr(__nr_to_section(__sec)); \ }) -#define pfn_to_page(pfn) \ +#define __pfn_to_page(pfn) \ ({ unsigned long __pfn = (pfn); \ struct mem_section *__sec = __pfn_to_section(__pfn); \ __section_mem_map_addr(__sec) + __pfn; \ }) #endif /* CONFIG_FLATMEM/DISCONTIGMEM/SPARSEMEM */ + +#ifdef CONFIG_OUT_OF_LINE_PFN_TO_PAGE +struct page; +/* this is useful when inlined pfn_to_page is too big */ +extern struct page *pfn_to_page(unsigned long pfn); +extern unsigned long page_to_pfn(struct page *page); +#else +#define page_to_pfn __page_to_pfn +#define pfn_to_page __pfn_to_page #endif /* CONFIG_OUT_OF_LINE_PFN_TO_PAGE */ #endif /* __ASSEMBLY__ */ diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 9dfbe6b7d1c..5af33186a25 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -2879,42 +2879,14 @@ void *__init alloc_large_system_hash(const char *tablename, } #ifdef CONFIG_OUT_OF_LINE_PFN_TO_PAGE -/* - * pfn <-> page translation. out-of-line version. - * (see asm-generic/memory_model.h) - */ -#if defined(CONFIG_FLATMEM) -struct page *pfn_to_page(unsigned long pfn) -{ - return mem_map + (pfn - ARCH_PFN_OFFSET); -} -unsigned long page_to_pfn(struct page *page) -{ - return (page - mem_map) + ARCH_PFN_OFFSET; -} -#elif defined(CONFIG_DISCONTIGMEM) struct page *pfn_to_page(unsigned long pfn) { - int nid = arch_pfn_to_nid(pfn); - return NODE_DATA(nid)->node_mem_map + arch_local_page_offset(pfn,nid); + return __pfn_to_page(pfn); } unsigned long page_to_pfn(struct page *page) { - struct pglist_data *pgdat = NODE_DATA(page_to_nid(page)); - return (page - pgdat->node_mem_map) + pgdat->node_start_pfn; -} -#elif defined(CONFIG_SPARSEMEM) -struct page *pfn_to_page(unsigned long pfn) -{ - return __section_mem_map_addr(__pfn_to_section(pfn)) + pfn; -} - -unsigned long page_to_pfn(struct page *page) -{ - long section_id = page_to_section(page); - return page - __section_mem_map_addr(__nr_to_section(section_id)); + return __page_to_pfn(page); } -#endif /* CONFIG_FLATMEM/DISCONTIGMME/SPARSEMEM */ EXPORT_SYMBOL(pfn_to_page); EXPORT_SYMBOL(page_to_pfn); #endif /* CONFIG_OUT_OF_LINE_PFN_TO_PAGE */ -- cgit v1.2.3-70-g09d2 From bd1e22b8e0a90f9a91e4c27db14ca15773659bf7 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Fri, 23 Jun 2006 02:03:47 -0700 Subject: [PATCH] initialise total_memory() earlier Initialise total_memory earlier in boot. Because if for some reason we run page reclaim early in boot, we don't want total_memory to be zero when we use it as a divisor. And rename total_memory to vm_total_pages to avoid naming clashes with architectures. Cc: Yasunori Goto Cc: KAMEZAWA Hiroyuki Cc: Martin Bligh Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/swap.h | 1 + mm/page_alloc.c | 6 +++--- mm/vmscan.c | 5 ++--- 3 files changed, 6 insertions(+), 6 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/include/linux/swap.h b/include/linux/swap.h index f1a827a972e..dc3f3aa0c83 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h @@ -185,6 +185,7 @@ extern unsigned long try_to_free_pages(struct zone **, gfp_t); extern unsigned long shrink_all_memory(unsigned long nr_pages); extern int vm_swappiness; extern int remove_mapping(struct address_space *mapping, struct page *page); +extern long vm_total_pages; #ifdef CONFIG_NUMA extern int zone_reclaim_mode; diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 5af33186a25..71a0b2a23f5 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1725,9 +1725,9 @@ void __meminit build_all_zonelists(void) stop_machine_run(__build_all_zonelists, NULL, NR_CPUS); /* cpuset refresh routine should be here */ } - - printk("Built %i zonelists\n", num_online_nodes()); - + vm_total_pages = nr_free_pagecache_pages(); + printk("Built %i zonelists. Total pages: %ld\n", + num_online_nodes(), vm_total_pages); } /* diff --git a/mm/vmscan.c b/mm/vmscan.c index 71a02e29503..72babac71de 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -110,7 +110,7 @@ struct shrinker { * From 0 .. 100. Higher means more swappy. */ int vm_swappiness = 60; -static long total_memory; +long vm_total_pages; /* The total number of pages which the VM controls */ static LIST_HEAD(shrinker_list); static DECLARE_RWSEM(shrinker_rwsem); @@ -743,7 +743,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, * how much memory * is mapped. */ - mapped_ratio = (sc->nr_mapped * 100) / total_memory; + mapped_ratio = (sc->nr_mapped * 100) / vm_total_pages; /* * Now decide how much we really want to unmap some pages. The @@ -1482,7 +1482,6 @@ static int __init kswapd_init(void) pgdat->kswapd = find_task_by_pid(pid); read_unlock(&tasklist_lock); } - total_memory = nr_free_pagecache_pages(); hotcpu_notifier(cpu_callback, 0); return 0; } -- cgit v1.2.3-70-g09d2 From 8f9de51a4a98ba32f839903b7d009788bc2c295d Mon Sep 17 00:00:00 2001 From: Kirill Korotaev Date: Fri, 23 Jun 2006 02:03:50 -0700 Subject: [PATCH] printk() should not be called under zone->lock This patch fixes printk() under zone->lock in show_free_areas(). It can be unsafe to call printk() under this lock, since caller can try to allocate/free some memory and selfdeadlock on this lock. I found allocations/freeing mem both in netconsole and serial console. This issue was faced in reallity when meminfo was periodically printed for debug purposes and netconsole was used. Signed-off-by: Kirill Korotaev Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 71a0b2a23f5..423db0db7c0 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1491,7 +1491,7 @@ void show_free_areas(void) } for_each_zone(zone) { - unsigned long nr, flags, order, total = 0; + unsigned long nr[MAX_ORDER], flags, order, total = 0; show_node(zone); printk("%s: ", zone->name); @@ -1502,11 +1502,12 @@ void show_free_areas(void) spin_lock_irqsave(&zone->lock, flags); for (order = 0; order < MAX_ORDER; order++) { - nr = zone->free_area[order].nr_free; - total += nr << order; - printk("%lu*%lukB ", nr, K(1UL) << order); + nr[order] = zone->free_area[order].nr_free; + total += nr[order] << order; } spin_unlock_irqrestore(&zone->lock, flags); + for (order = 0; order < MAX_ORDER; order++) + printk("%lu*%lukB ", nr[order], K(1UL) << order); printk("= %lukB\n", K(total)); } -- cgit v1.2.3-70-g09d2