From 264132bc62fe071d0ff378c1103bae9d33212f10 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Mon, 6 Mar 2006 12:10:07 -0800 Subject: Fix "check_slabp" printout size calculation We want to use the "struct slab" size, not the size of the pointer to same. As it is, we'd not print out the last entry pointers in the slab (where is ~10, depending on whether it's a 32-bit or 64-bit kernel). Gaah, that slab code was written by somebody who likes unreadable crud. Signed-off-by: Linus Torvalds --- mm/slab.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index add05d808a4..2b0b1519bb7 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -2554,7 +2554,7 @@ static void check_slabp(struct kmem_cache *cachep, struct slab *slabp) "slab: Internal list corruption detected in cache '%s'(%d), slabp %p(%d). Hexdump:\n", cachep->name, cachep->num, slabp, slabp->inuse); for (i = 0; - i < sizeof(slabp) + cachep->num * sizeof(kmem_bufctl_t); + i < sizeof(*slabp) + cachep->num * sizeof(kmem_bufctl_t); i++) { if ((i % 16) == 0) printk("\n%03x:", i); -- cgit From 9888e6fa7b68d9c8cc2c162a90979825ab45150a Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Mon, 6 Mar 2006 17:44:43 -0800 Subject: slab: clarify and fix calculate_slab_order() If we triggered the 'offslab_limit' test, we would return with cachep->gfporder incremented once too many times. This clarifies the logic somewhat, and fixes that bug. Signed-off-by: Linus Torvalds --- mm/slab.c | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 2b0b1519bb7..f2e92dc1c9c 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -1628,36 +1628,36 @@ static inline size_t calculate_slab_order(struct kmem_cache *cachep, size_t size, size_t align, unsigned long flags) { size_t left_over = 0; + int gfporder; - for (;; cachep->gfporder++) { + for (gfporder = 0 ; gfporder <= MAX_GFP_ORDER; gfporder++) { unsigned int num; size_t remainder; - if (cachep->gfporder > MAX_GFP_ORDER) { - cachep->num = 0; - break; - } - - cache_estimate(cachep->gfporder, size, align, flags, - &remainder, &num); + cache_estimate(gfporder, size, align, flags, &remainder, &num); if (!num) continue; + /* More than offslab_limit objects will cause problems */ - if (flags & CFLGS_OFF_SLAB && cachep->num > offslab_limit) + if ((flags & CFLGS_OFF_SLAB) && num > offslab_limit) break; + /* Found something acceptable - save it away */ cachep->num = num; + cachep->gfporder = gfporder; left_over = remainder; /* * Large number of objects is good, but very large slabs are * currently bad for the gfp()s. */ - if (cachep->gfporder >= slab_break_gfp_order) + if (gfporder >= slab_break_gfp_order) break; - if ((left_over * 8) <= (PAGE_SIZE << cachep->gfporder)) - /* Acceptable internal fragmentation */ + /* + * Acceptable internal fragmentation? + */ + if ((left_over * 8) <= (PAGE_SIZE << gfporder)) break; } return left_over; -- cgit From f78bb8ad482267b92c122f0e37a7dce69c880247 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 8 Mar 2006 10:33:05 -0800 Subject: slab: fix calculate_slab_order() for SLAB_RECLAIM_ACCOUNT Instead of having a hard-to-read and confusing conditional in the caller, just make the slab order calculation handle this special case, since it's simple and obvious there. Signed-off-by: Linus Torvalds --- mm/slab.c | 20 +++++++++----------- 1 file changed, 9 insertions(+), 11 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index f2e92dc1c9c..6ad6bd5a0b3 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -1647,6 +1647,14 @@ static inline size_t calculate_slab_order(struct kmem_cache *cachep, cachep->gfporder = gfporder; left_over = remainder; + /* + * A VFS-reclaimable slab tends to have most allocations + * as GFP_NOFS and we really don't want to have to be allocating + * higher-order pages when we are unable to shrink dcache. + */ + if (flags & SLAB_RECLAIM_ACCOUNT) + break; + /* * Large number of objects is good, but very large slabs are * currently bad for the gfp()s. @@ -1869,17 +1877,7 @@ kmem_cache_create (const char *name, size_t size, size_t align, size = ALIGN(size, align); - if ((flags & SLAB_RECLAIM_ACCOUNT) && size <= PAGE_SIZE) { - /* - * A VFS-reclaimable slab tends to have most allocations - * as GFP_NOFS and we really don't want to have to be allocating - * higher-order pages when we are unable to shrink dcache. - */ - cachep->gfporder = 0; - cache_estimate(cachep->gfporder, size, align, flags, - &left_over, &cachep->num); - } else - left_over = calculate_slab_order(cachep, size, align, flags); + left_over = calculate_slab_order(cachep, size, align, flags); if (!cachep->num) { printk("kmem_cache_create: couldn't create cache %s.\n", name); -- cgit From 07ed76b2a085a31f427c2a912a562627947dc7de Mon Sep 17 00:00:00 2001 From: Jack Steiner Date: Tue, 7 Mar 2006 21:55:46 -0800 Subject: [PATCH] slab: allocate larger cache_cache if order 0 fails kmem_cache_init() incorrectly assumes that the cache_cache object will fit in an order 0 allocation. On very large systems, this is not true. Change the code to try larger order allocations if order 0 fails. Signed-off-by: Jack Steiner Cc: Manfred Spraul Cc: Pekka Enberg Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 6ad6bd5a0b3..61800b88e24 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -1124,6 +1124,7 @@ void __init kmem_cache_init(void) struct cache_sizes *sizes; struct cache_names *names; int i; + int order; for (i = 0; i < NUM_INIT_LISTS; i++) { kmem_list3_init(&initkmem_list3[i]); @@ -1167,11 +1168,15 @@ void __init kmem_cache_init(void) cache_cache.buffer_size = ALIGN(cache_cache.buffer_size, cache_line_size()); - cache_estimate(0, cache_cache.buffer_size, cache_line_size(), 0, - &left_over, &cache_cache.num); + for (order = 0; order < MAX_ORDER; order++) { + cache_estimate(order, cache_cache.buffer_size, + cache_line_size(), 0, &left_over, &cache_cache.num); + if (cache_cache.num) + break; + } if (!cache_cache.num) BUG(); - + cache_cache.gfporder = order; cache_cache.colour = left_over / cache_cache.colour_off; cache_cache.slab_size = ALIGN(cache_cache.num * sizeof(kmem_bufctl_t) + sizeof(struct slab), cache_line_size()); -- cgit From 8fce4d8e3b9e3cf47cc8afeb6077e22ab795d989 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Thu, 9 Mar 2006 17:33:54 -0800 Subject: [PATCH] slab: Node rotor for freeing alien caches and remote per cpu pages. The cache reaper currently tries to free all alien caches and all remote per cpu pages in each pass of cache_reap. For a machines with large number of nodes (such as Altix) this may lead to sporadic delays of around ~10ms. Interrupts are disabled while reclaiming creating unacceptable delays. This patch changes that behavior by adding a per cpu reap_node variable. Instead of attempting to free all caches, we free only one alien cache and the per cpu pages from one remote node. That reduces the time spend in cache_reap. However, doing so will lengthen the time it takes to completely drain all remote per cpu pagesets and all alien caches. The time needed will grow with the number of nodes in the system. All caches are drained when they overflow their respective capacity. So the drawback here is only that a bit of memory may be wasted for awhile longer. Details: 1. Rename drain_remote_pages to drain_node_pages to allow the specification of the node to drain of pcp pages. 2. Add additional functions init_reap_node, next_reap_node for NUMA that manage a per cpu reap_node counter. 3. Add a reap_alien function that reaps only from the current reap_node. For us this seems to be a critical issue. Holdoffs of an average of ~7ms cause some HPC benchmarks to slow down significantly. F.e. NAS parallel slows down dramatically. NAS parallel has a 12-16 seconds runtime w/o rotor compared to 5.8 secs with the rotor patches. It gets down to 5.05 secs with the additional interrupt holdoff reductions. Signed-off-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 65 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 62 insertions(+), 3 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 61800b88e24..d0bd7f07ab0 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -789,6 +789,47 @@ static void __slab_error(const char *function, struct kmem_cache *cachep, char * dump_stack(); } +#ifdef CONFIG_NUMA +/* + * Special reaping functions for NUMA systems called from cache_reap(). + * These take care of doing round robin flushing of alien caches (containing + * objects freed on different nodes from which they were allocated) and the + * flushing of remote pcps by calling drain_node_pages. + */ +static DEFINE_PER_CPU(unsigned long, reap_node); + +static void init_reap_node(int cpu) +{ + int node; + + node = next_node(cpu_to_node(cpu), node_online_map); + if (node == MAX_NUMNODES) + node = 0; + + __get_cpu_var(reap_node) = node; +} + +static void next_reap_node(void) +{ + int node = __get_cpu_var(reap_node); + + /* + * Also drain per cpu pages on remote zones + */ + if (node != numa_node_id()) + drain_node_pages(node); + + node = next_node(node, node_online_map); + if (unlikely(node >= MAX_NUMNODES)) + node = first_node(node_online_map); + __get_cpu_var(reap_node) = node; +} + +#else +#define init_reap_node(cpu) do { } while (0) +#define next_reap_node(void) do { } while (0) +#endif + /* * Initiate the reap timer running on the target CPU. We run at around 1 to 2Hz * via the workqueue/eventd. @@ -806,6 +847,7 @@ static void __devinit start_cpu_timer(int cpu) * at that time. */ if (keventd_up() && reap_work->func == NULL) { + init_reap_node(cpu); INIT_WORK(reap_work, cache_reap, NULL); schedule_delayed_work_on(cpu, reap_work, HZ + 3 * cpu); } @@ -884,6 +926,23 @@ static void __drain_alien_cache(struct kmem_cache *cachep, } } +/* + * Called from cache_reap() to regularly drain alien caches round robin. + */ +static void reap_alien(struct kmem_cache *cachep, struct kmem_list3 *l3) +{ + int node = __get_cpu_var(reap_node); + + if (l3->alien) { + struct array_cache *ac = l3->alien[node]; + if (ac && ac->avail) { + spin_lock_irq(&ac->lock); + __drain_alien_cache(cachep, ac, node); + spin_unlock_irq(&ac->lock); + } + } +} + static void drain_alien_cache(struct kmem_cache *cachep, struct array_cache **alien) { int i = 0; @@ -902,6 +961,7 @@ static void drain_alien_cache(struct kmem_cache *cachep, struct array_cache **al #else #define drain_alien_cache(cachep, alien) do { } while (0) +#define reap_alien(cachep, l3) do { } while (0) static inline struct array_cache **alloc_alien_cache(int node, int limit) { @@ -3497,8 +3557,7 @@ static void cache_reap(void *unused) check_irq_on(); l3 = searchp->nodelists[numa_node_id()]; - if (l3->alien) - drain_alien_cache(searchp, l3->alien); + reap_alien(searchp, l3); spin_lock_irq(&l3->list_lock); drain_array_locked(searchp, cpu_cache_get(searchp), 0, @@ -3548,7 +3607,7 @@ static void cache_reap(void *unused) } check_irq_on(); mutex_unlock(&cache_chain_mutex); - drain_remote_pages(); + next_reap_node(); /* Setup the next iteration */ schedule_delayed_work(&__get_cpu_var(reap_work), REAPTIMEOUT_CPUC); } -- cgit From f205b2fe62d321403525065a4cb31b6bff1bbe53 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Wed, 22 Mar 2006 00:08:02 -0800 Subject: [PATCH] mm: slab less atomics Atomic operation removal from slab Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index d0bd7f07ab0..5988adf010c 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -1402,7 +1402,7 @@ static void *kmem_getpages(struct kmem_cache *cachep, gfp_t flags, int nodeid) atomic_add(i, &slab_reclaim_pages); add_page_state(nr_slab, i); while (i--) { - SetPageSlab(page); + __SetPageSlab(page); page++; } return addr; @@ -1418,8 +1418,8 @@ static void kmem_freepages(struct kmem_cache *cachep, void *addr) const unsigned long nr_freed = i; while (i--) { - if (!TestClearPageSlab(page)) - BUG(); + BUG_ON(!PageSlab(page)); + __ClearPageSlab(page); page++; } sub_page_state(nr_slab, nr_freed); -- cgit From 8fea4e96a8f29ccc34c244f54574680ce9b43631 Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Wed, 22 Mar 2006 00:08:10 -0800 Subject: [PATCH] slab: object to index mapping cleanup Clean up the object to index mapping that has been spread around mm/slab.c. Signed-off-by: Pekka Enberg Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 34 +++++++++++++++++++++++----------- 1 file changed, 23 insertions(+), 11 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 5988adf010c..3d18b711ab8 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -609,6 +609,18 @@ static inline struct slab *virt_to_slab(const void *obj) return page_get_slab(page); } +static inline void *index_to_obj(struct kmem_cache *cache, struct slab *slab, + unsigned int idx) +{ + return slab->s_mem + cache->buffer_size * idx; +} + +static inline unsigned int obj_to_index(struct kmem_cache *cache, + struct slab *slab, void *obj) +{ + return (unsigned)(obj - slab->s_mem) / cache->buffer_size; +} + /* These are the default caches for kmalloc. Custom caches can have other sizes. */ struct cache_sizes malloc_sizes[] = { #define CACHE(x) { .cs_size = (x) }, @@ -1568,18 +1580,18 @@ static void check_poison_obj(struct kmem_cache *cachep, void *objp) * exist: */ struct slab *slabp = virt_to_slab(objp); - int objnr; + unsigned int objnr; - objnr = (unsigned)(objp - slabp->s_mem) / cachep->buffer_size; + objnr = obj_to_index(cachep, slabp, objp); if (objnr) { - objp = slabp->s_mem + (objnr - 1) * cachep->buffer_size; + objp = index_to_obj(cachep, slabp, objnr - 1); realobj = (char *)objp + obj_offset(cachep); printk(KERN_ERR "Prev obj: start=%p, len=%d\n", realobj, size); print_objinfo(cachep, objp, 2); } if (objnr + 1 < cachep->num) { - objp = slabp->s_mem + (objnr + 1) * cachep->buffer_size; + objp = index_to_obj(cachep, slabp, objnr + 1); realobj = (char *)objp + obj_offset(cachep); printk(KERN_ERR "Next obj: start=%p, len=%d\n", realobj, size); @@ -1598,7 +1610,7 @@ static void slab_destroy_objs(struct kmem_cache *cachep, struct slab *slabp) { int i; for (i = 0; i < cachep->num; i++) { - void *objp = slabp->s_mem + cachep->buffer_size * i; + void *objp = index_to_obj(cachep, slabp, i); if (cachep->flags & SLAB_POISON) { #ifdef CONFIG_DEBUG_PAGEALLOC @@ -1631,7 +1643,7 @@ static void slab_destroy_objs(struct kmem_cache *cachep, struct slab *slabp) if (cachep->dtor) { int i; for (i = 0; i < cachep->num; i++) { - void *objp = slabp->s_mem + cachep->buffer_size * i; + void *objp = index_to_obj(cachep, slabp, i); (cachep->dtor) (objp, cachep, 0); } } @@ -2307,7 +2319,7 @@ static void cache_init_objs(struct kmem_cache *cachep, int i; for (i = 0; i < cachep->num; i++) { - void *objp = slabp->s_mem + cachep->buffer_size * i; + void *objp = index_to_obj(cachep, slabp, i); #if DEBUG /* need to poison the objs? */ if (cachep->flags & SLAB_POISON) @@ -2363,7 +2375,7 @@ static void kmem_flagcheck(struct kmem_cache *cachep, gfp_t flags) static void *slab_get_obj(struct kmem_cache *cachep, struct slab *slabp, int nodeid) { - void *objp = slabp->s_mem + (slabp->free * cachep->buffer_size); + void *objp = index_to_obj(cachep, slabp, slabp->free); kmem_bufctl_t next; slabp->inuse++; @@ -2380,7 +2392,7 @@ static void *slab_get_obj(struct kmem_cache *cachep, struct slab *slabp, int nod static void slab_put_obj(struct kmem_cache *cachep, struct slab *slabp, void *objp, int nodeid) { - unsigned int objnr = (unsigned)(objp-slabp->s_mem) / cachep->buffer_size; + unsigned int objnr = obj_to_index(cachep, slabp, objp); #if DEBUG /* Verify that the slab belongs to the intended node */ @@ -2565,10 +2577,10 @@ static void *cache_free_debugcheck(struct kmem_cache *cachep, void *objp, if (cachep->flags & SLAB_STORE_USER) *dbg_userword(cachep, objp) = caller; - objnr = (unsigned)(objp - slabp->s_mem) / cachep->buffer_size; + objnr = obj_to_index(cachep, slabp, objp); BUG_ON(objnr >= cachep->num); - BUG_ON(objp != slabp->s_mem + objnr * cachep->buffer_size); + BUG_ON(objp != index_to_obj(cachep, slabp, objnr)); if (cachep->flags & SLAB_DEBUG_INITIAL) { /* Need to call the slab's constructor so the -- cgit From f30cf7d13eee420f5249b4d7709b46570098ab92 Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Wed, 22 Mar 2006 00:08:11 -0800 Subject: [PATCH] slab: extract setup_cpu_cache Extract setup_cpu_cache() function from kmem_cache_create() to make the latter a little less complex. Signed-off-by: Pekka Enberg Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 109 +++++++++++++++++++++++++++++++------------------------------- 1 file changed, 55 insertions(+), 54 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 3d18b711ab8..4d5c4b93e0e 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -1748,6 +1748,60 @@ static inline size_t calculate_slab_order(struct kmem_cache *cachep, return left_over; } +static void setup_cpu_cache(struct kmem_cache *cachep) +{ + if (g_cpucache_up == FULL) { + enable_cpucache(cachep); + return; + } + if (g_cpucache_up == NONE) { + /* + * Note: the first kmem_cache_create must create the cache + * that's used by kmalloc(24), otherwise the creation of + * further caches will BUG(). + */ + cachep->array[smp_processor_id()] = &initarray_generic.cache; + + /* + * If the cache that's used by kmalloc(sizeof(kmem_list3)) is + * the first cache, then we need to set up all its list3s, + * otherwise the creation of further caches will BUG(). + */ + set_up_list3s(cachep, SIZE_AC); + if (INDEX_AC == INDEX_L3) + g_cpucache_up = PARTIAL_L3; + else + g_cpucache_up = PARTIAL_AC; + } else { + cachep->array[smp_processor_id()] = + kmalloc(sizeof(struct arraycache_init), GFP_KERNEL); + + if (g_cpucache_up == PARTIAL_AC) { + set_up_list3s(cachep, SIZE_L3); + g_cpucache_up = PARTIAL_L3; + } else { + int node; + for_each_online_node(node) { + cachep->nodelists[node] = + kmalloc_node(sizeof(struct kmem_list3), + GFP_KERNEL, node); + BUG_ON(!cachep->nodelists[node]); + kmem_list3_init(cachep->nodelists[node]); + } + } + } + cachep->nodelists[numa_node_id()]->next_reap = + jiffies + REAPTIMEOUT_LIST3 + + ((unsigned long)cachep) % REAPTIMEOUT_LIST3; + + cpu_cache_get(cachep)->avail = 0; + cpu_cache_get(cachep)->limit = BOOT_CPUCACHE_ENTRIES; + cpu_cache_get(cachep)->batchcount = 1; + cpu_cache_get(cachep)->touched = 0; + cachep->batchcount = 1; + cachep->limit = BOOT_CPUCACHE_ENTRIES; +} + /** * kmem_cache_create - Create a cache. * @name: A string which is used in /proc/slabinfo to identify this cache. @@ -2000,60 +2054,7 @@ kmem_cache_create (const char *name, size_t size, size_t align, cachep->name = name; - if (g_cpucache_up == FULL) { - enable_cpucache(cachep); - } else { - if (g_cpucache_up == NONE) { - /* Note: the first kmem_cache_create must create - * the cache that's used by kmalloc(24), otherwise - * the creation of further caches will BUG(). - */ - cachep->array[smp_processor_id()] = - &initarray_generic.cache; - - /* If the cache that's used by - * kmalloc(sizeof(kmem_list3)) is the first cache, - * then we need to set up all its list3s, otherwise - * the creation of further caches will BUG(). - */ - set_up_list3s(cachep, SIZE_AC); - if (INDEX_AC == INDEX_L3) - g_cpucache_up = PARTIAL_L3; - else - g_cpucache_up = PARTIAL_AC; - } else { - cachep->array[smp_processor_id()] = - kmalloc(sizeof(struct arraycache_init), GFP_KERNEL); - - if (g_cpucache_up == PARTIAL_AC) { - set_up_list3s(cachep, SIZE_L3); - g_cpucache_up = PARTIAL_L3; - } else { - int node; - for_each_online_node(node) { - - cachep->nodelists[node] = - kmalloc_node(sizeof - (struct kmem_list3), - GFP_KERNEL, node); - BUG_ON(!cachep->nodelists[node]); - kmem_list3_init(cachep-> - nodelists[node]); - } - } - } - cachep->nodelists[numa_node_id()]->next_reap = - jiffies + REAPTIMEOUT_LIST3 + - ((unsigned long)cachep) % REAPTIMEOUT_LIST3; - - BUG_ON(!cpu_cache_get(cachep)); - cpu_cache_get(cachep)->avail = 0; - cpu_cache_get(cachep)->limit = BOOT_CPUCACHE_ENTRIES; - cpu_cache_get(cachep)->batchcount = 1; - cpu_cache_get(cachep)->touched = 0; - cachep->batchcount = 1; - cachep->limit = BOOT_CPUCACHE_ENTRIES; - } + setup_cpu_cache(cachep); /* cache setup completed, link it into the list */ list_add(&cachep->next, &cache_chain); -- cgit From a737b3e2fcf96f576fa3e2e382236d9ee94f383f Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Wed, 22 Mar 2006 00:08:11 -0800 Subject: [PATCH] slab cleanup slab.c has become a bit revolting again. Try to repair it. - Coding style fixes - Don't do assignments-in-if-statements. - Don't typecast assignments to/from void* Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 596 ++++++++++++++++++++++++++++++++------------------------------ 1 file changed, 304 insertions(+), 292 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 4d5c4b93e0e..7b6f9f10e75 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -50,7 +50,7 @@ * The head array is strictly LIFO and should improve the cache hit rates. * On SMP, it additionally reduces the spinlock operations. * - * The c_cpuarray may not be read with enabled local interrupts - + * The c_cpuarray may not be read with enabled local interrupts - * it's changed with a smp_call_function(). * * SMP synchronization: @@ -266,16 +266,17 @@ struct array_cache { unsigned int batchcount; unsigned int touched; spinlock_t lock; - void *entry[0]; /* - * Must have this definition in here for the proper - * alignment of array_cache. Also simplifies accessing - * the entries. - * [0] is for gcc 2.95. It should really be []. - */ + void *entry[0]; /* + * Must have this definition in here for the proper + * alignment of array_cache. Also simplifies accessing + * the entries. + * [0] is for gcc 2.95. It should really be []. + */ }; -/* bootstrap: The caches do not work without cpuarrays anymore, - * but the cpuarrays are allocated from the generic caches... +/* + * bootstrap: The caches do not work without cpuarrays anymore, but the + * cpuarrays are allocated from the generic caches... */ #define BOOT_CPUCACHE_ENTRIES 1 struct arraycache_init { @@ -310,10 +311,8 @@ struct kmem_list3 __initdata initkmem_list3[NUM_INIT_LISTS]; #define SIZE_L3 (1 + MAX_NUMNODES) /* - * This function must be completely optimized away if - * a constant is passed to it. Mostly the same as - * what is in linux/slab.h except it returns an - * index. + * This function must be completely optimized away if a constant is passed to + * it. Mostly the same as what is in linux/slab.h except it returns an index. */ static __always_inline int index_of(const size_t size) { @@ -351,14 +350,14 @@ static void kmem_list3_init(struct kmem_list3 *parent) parent->free_touched = 0; } -#define MAKE_LIST(cachep, listp, slab, nodeid) \ - do { \ - INIT_LIST_HEAD(listp); \ - list_splice(&(cachep->nodelists[nodeid]->slab), listp); \ +#define MAKE_LIST(cachep, listp, slab, nodeid) \ + do { \ + INIT_LIST_HEAD(listp); \ + list_splice(&(cachep->nodelists[nodeid]->slab), listp); \ } while (0) -#define MAKE_ALL_LISTS(cachep, ptr, nodeid) \ - do { \ +#define MAKE_ALL_LISTS(cachep, ptr, nodeid) \ + do { \ MAKE_LIST((cachep), (&(ptr)->slabs_full), slabs_full, nodeid); \ MAKE_LIST((cachep), (&(ptr)->slabs_partial), slabs_partial, nodeid); \ MAKE_LIST((cachep), (&(ptr)->slabs_free), slabs_free, nodeid); \ @@ -379,8 +378,8 @@ struct kmem_cache { unsigned int buffer_size; /* 2) touched by every alloc & free from the backend */ struct kmem_list3 *nodelists[MAX_NUMNODES]; - unsigned int flags; /* constant flags */ - unsigned int num; /* # of objs per slab */ + unsigned int flags; /* constant flags */ + unsigned int num; /* # of objs per slab */ spinlock_t spinlock; /* 3) cache_grow/shrink */ @@ -390,11 +389,11 @@ struct kmem_cache { /* force GFP flags, e.g. GFP_DMA */ gfp_t gfpflags; - size_t colour; /* cache colouring range */ + size_t colour; /* cache colouring range */ unsigned int colour_off; /* colour offset */ struct kmem_cache *slabp_cache; unsigned int slab_size; - unsigned int dflags; /* dynamic flags */ + unsigned int dflags; /* dynamic flags */ /* constructor func */ void (*ctor) (void *, struct kmem_cache *, unsigned long); @@ -438,8 +437,9 @@ struct kmem_cache { #define OFF_SLAB(x) ((x)->flags & CFLGS_OFF_SLAB) #define BATCHREFILL_LIMIT 16 -/* Optimization question: fewer reaps means less - * probability for unnessary cpucache drain/refill cycles. +/* + * Optimization question: fewer reaps means less probability for unnessary + * cpucache drain/refill cycles. * * OTOH the cpuarrays can contain lots of objects, * which could lock up otherwise freeable slabs. @@ -453,17 +453,19 @@ struct kmem_cache { #define STATS_INC_ALLOCED(x) ((x)->num_allocations++) #define STATS_INC_GROWN(x) ((x)->grown++) #define STATS_INC_REAPED(x) ((x)->reaped++) -#define STATS_SET_HIGH(x) do { if ((x)->num_active > (x)->high_mark) \ - (x)->high_mark = (x)->num_active; \ - } while (0) +#define STATS_SET_HIGH(x) \ + do { \ + if ((x)->num_active > (x)->high_mark) \ + (x)->high_mark = (x)->num_active; \ + } while (0) #define STATS_INC_ERR(x) ((x)->errors++) #define STATS_INC_NODEALLOCS(x) ((x)->node_allocs++) #define STATS_INC_NODEFREES(x) ((x)->node_frees++) -#define STATS_SET_FREEABLE(x, i) \ - do { if ((x)->max_freeable < i) \ - (x)->max_freeable = i; \ - } while (0) - +#define STATS_SET_FREEABLE(x, i) \ + do { \ + if ((x)->max_freeable < i) \ + (x)->max_freeable = i; \ + } while (0) #define STATS_INC_ALLOCHIT(x) atomic_inc(&(x)->allochit) #define STATS_INC_ALLOCMISS(x) atomic_inc(&(x)->allocmiss) #define STATS_INC_FREEHIT(x) atomic_inc(&(x)->freehit) @@ -478,9 +480,7 @@ struct kmem_cache { #define STATS_INC_ERR(x) do { } while (0) #define STATS_INC_NODEALLOCS(x) do { } while (0) #define STATS_INC_NODEFREES(x) do { } while (0) -#define STATS_SET_FREEABLE(x, i) \ - do { } while (0) - +#define STATS_SET_FREEABLE(x, i) do { } while (0) #define STATS_INC_ALLOCHIT(x) do { } while (0) #define STATS_INC_ALLOCMISS(x) do { } while (0) #define STATS_INC_FREEHIT(x) do { } while (0) @@ -488,7 +488,8 @@ struct kmem_cache { #endif #if DEBUG -/* Magic nums for obj red zoning. +/* + * Magic nums for obj red zoning. * Placed in the first word before and the first word after an obj. */ #define RED_INACTIVE 0x5A2CF071UL /* when obj is inactive */ @@ -499,7 +500,8 @@ struct kmem_cache { #define POISON_FREE 0x6b /* for use-after-free poisoning */ #define POISON_END 0xa5 /* end-byte of poisoning */ -/* memory layout of objects: +/* + * memory layout of objects: * 0 : objp * 0 .. cachep->obj_offset - BYTES_PER_WORD - 1: padding. This ensures that * the end of an object is aligned with the end of the real @@ -508,7 +510,8 @@ struct kmem_cache { * redzone word. * cachep->obj_offset: The real object. * cachep->buffer_size - 2* BYTES_PER_WORD: redzone word [BYTES_PER_WORD long] - * cachep->buffer_size - 1* BYTES_PER_WORD: last caller address [BYTES_PER_WORD long] + * cachep->buffer_size - 1* BYTES_PER_WORD: last caller address + * [BYTES_PER_WORD long] */ static int obj_offset(struct kmem_cache *cachep) { @@ -552,8 +555,8 @@ static void **dbg_userword(struct kmem_cache *cachep, void *objp) #endif /* - * Maximum size of an obj (in 2^order pages) - * and absolute limit for the gfp order. + * Maximum size of an obj (in 2^order pages) and absolute limit for the gfp + * order. */ #if defined(CONFIG_LARGE_ALLOCS) #define MAX_OBJ_ORDER 13 /* up to 32Mb */ @@ -573,9 +576,10 @@ static void **dbg_userword(struct kmem_cache *cachep, void *objp) #define BREAK_GFP_ORDER_LO 0 static int slab_break_gfp_order = BREAK_GFP_ORDER_LO; -/* Functions for storing/retrieving the cachep and or slab from the - * global 'mem_map'. These are used to find the slab an obj belongs to. - * With kfree(), these are used to find the cache which an obj belongs to. +/* + * Functions for storing/retrieving the cachep and or slab from the page + * allocator. These are used to find the slab an obj belongs to. With kfree(), + * these are used to find the cache which an obj belongs to. */ static inline void page_set_cache(struct page *page, struct kmem_cache *cache) { @@ -621,7 +625,9 @@ static inline unsigned int obj_to_index(struct kmem_cache *cache, return (unsigned)(obj - slab->s_mem) / cache->buffer_size; } -/* These are the default caches for kmalloc. Custom caches can have other sizes. */ +/* + * These are the default caches for kmalloc. Custom caches can have other sizes. + */ struct cache_sizes malloc_sizes[] = { #define CACHE(x) { .cs_size = (x) }, #include @@ -667,8 +673,8 @@ static DEFINE_MUTEX(cache_chain_mutex); static struct list_head cache_chain; /* - * vm_enough_memory() looks at this to determine how many - * slab-allocated pages are possibly freeable under pressure + * vm_enough_memory() looks at this to determine how many slab-allocated pages + * are possibly freeable under pressure * * SLAB_RECLAIM_ACCOUNT turns this on per-slab */ @@ -687,7 +693,8 @@ static enum { static DEFINE_PER_CPU(struct work_struct, reap_work); -static void free_block(struct kmem_cache *cachep, void **objpp, int len, int node); +static void free_block(struct kmem_cache *cachep, void **objpp, int len, + int node); static void enable_cpucache(struct kmem_cache *cachep); static void cache_reap(void *unused); static int __node_shrink(struct kmem_cache *cachep, int node); @@ -697,7 +704,8 @@ static inline struct array_cache *cpu_cache_get(struct kmem_cache *cachep) return cachep->array[smp_processor_id()]; } -static inline struct kmem_cache *__find_general_cachep(size_t size, gfp_t gfpflags) +static inline struct kmem_cache *__find_general_cachep(size_t size, + gfp_t gfpflags) { struct cache_sizes *csizep = malloc_sizes; @@ -732,8 +740,9 @@ static size_t slab_mgmt_size(size_t nr_objs, size_t align) return ALIGN(sizeof(struct slab)+nr_objs*sizeof(kmem_bufctl_t), align); } -/* Calculate the number of objects and left-over bytes for a given - buffer size. */ +/* + * Calculate the number of objects and left-over bytes for a given buffer size. + */ static void cache_estimate(unsigned long gfporder, size_t buffer_size, size_t align, int flags, size_t *left_over, unsigned int *num) @@ -794,7 +803,8 @@ static void cache_estimate(unsigned long gfporder, size_t buffer_size, #define slab_error(cachep, msg) __slab_error(__FUNCTION__, cachep, msg) -static void __slab_error(const char *function, struct kmem_cache *cachep, char *msg) +static void __slab_error(const char *function, struct kmem_cache *cachep, + char *msg) { printk(KERN_ERR "slab error in %s(): cache `%s': %s\n", function, cachep->name, msg); @@ -918,10 +928,8 @@ static void free_alien_cache(struct array_cache **ac_ptr) if (!ac_ptr) return; - for_each_node(i) kfree(ac_ptr[i]); - kfree(ac_ptr); } @@ -955,7 +963,8 @@ static void reap_alien(struct kmem_cache *cachep, struct kmem_list3 *l3) } } -static void drain_alien_cache(struct kmem_cache *cachep, struct array_cache **alien) +static void drain_alien_cache(struct kmem_cache *cachep, + struct array_cache **alien) { int i = 0; struct array_cache *ac; @@ -998,20 +1007,22 @@ static int __devinit cpuup_callback(struct notifier_block *nfb, switch (action) { case CPU_UP_PREPARE: mutex_lock(&cache_chain_mutex); - /* we need to do this right in the beginning since + /* + * We need to do this right in the beginning since * alloc_arraycache's are going to use this list. * kmalloc_node allows us to add the slab to the right * kmem_list3 and not this cpu's kmem_list3 */ list_for_each_entry(cachep, &cache_chain, next) { - /* setup the size64 kmemlist for cpu before we can + /* + * Set up the size64 kmemlist for cpu before we can * begin anything. Make sure some other cpu on this * node has not already allocated this */ if (!cachep->nodelists[node]) { - if (!(l3 = kmalloc_node(memsize, - GFP_KERNEL, node))) + l3 = kmalloc_node(memsize, GFP_KERNEL, node); + if (!l3) goto bad; kmem_list3_init(l3); l3->next_reap = jiffies + REAPTIMEOUT_LIST3 + @@ -1027,13 +1038,15 @@ static int __devinit cpuup_callback(struct notifier_block *nfb, spin_lock_irq(&cachep->nodelists[node]->list_lock); cachep->nodelists[node]->free_limit = - (1 + nr_cpus_node(node)) * - cachep->batchcount + cachep->num; + (1 + nr_cpus_node(node)) * + cachep->batchcount + cachep->num; spin_unlock_irq(&cachep->nodelists[node]->list_lock); } - /* Now we can go ahead with allocating the shared array's - & array cache's */ + /* + * Now we can go ahead with allocating the shared arrays and + * array caches + */ list_for_each_entry(cachep, &cache_chain, next) { struct array_cache *nc; struct array_cache *shared; @@ -1053,7 +1066,6 @@ static int __devinit cpuup_callback(struct notifier_block *nfb, if (!alien) goto bad; cachep->array[cpu] = nc; - l3 = cachep->nodelists[node]; BUG_ON(!l3); @@ -1073,7 +1085,6 @@ static int __devinit cpuup_callback(struct notifier_block *nfb, } #endif spin_unlock_irq(&l3->list_lock); - kfree(shared); free_alien_cache(alien); } @@ -1095,7 +1106,6 @@ static int __devinit cpuup_callback(struct notifier_block *nfb, /* fall thru */ case CPU_UP_CANCELED: mutex_lock(&cache_chain_mutex); - list_for_each_entry(cachep, &cache_chain, next) { struct array_cache *nc; struct array_cache *shared; @@ -1162,7 +1172,7 @@ free_array_cache: #endif } return NOTIFY_OK; - bad: +bad: mutex_unlock(&cache_chain_mutex); return NOTIFY_BAD; } @@ -1172,7 +1182,8 @@ static struct notifier_block cpucache_notifier = { &cpuup_callback, NULL, 0 }; /* * swap the static kmem_list3 with kmalloced memory */ -static void init_list(struct kmem_cache *cachep, struct kmem_list3 *list, int nodeid) +static void init_list(struct kmem_cache *cachep, struct kmem_list3 *list, + int nodeid) { struct kmem_list3 *ptr; @@ -1187,8 +1198,9 @@ static void init_list(struct kmem_cache *cachep, struct kmem_list3 *list, int no local_irq_enable(); } -/* Initialisation. - * Called after the gfp() functions have been enabled, and before smp_init(). +/* + * Initialisation. Called after the page allocator have been initialised and + * before smp_init(). */ void __init kmem_cache_init(void) { @@ -1213,9 +1225,9 @@ void __init kmem_cache_init(void) /* Bootstrap is tricky, because several objects are allocated * from caches that do not exist yet: - * 1) initialize the cache_cache cache: it contains the struct kmem_cache - * structures of all caches, except cache_cache itself: cache_cache - * is statically allocated. + * 1) initialize the cache_cache cache: it contains the struct + * kmem_cache structures of all caches, except cache_cache itself: + * cache_cache is statically allocated. * Initially an __init data area is used for the head array and the * kmem_list3 structures, it's replaced with a kmalloc allocated * array at the end of the bootstrap. @@ -1238,7 +1250,8 @@ void __init kmem_cache_init(void) cache_cache.array[smp_processor_id()] = &initarray_cache.cache; cache_cache.nodelists[numa_node_id()] = &initkmem_list3[CACHE_CACHE]; - cache_cache.buffer_size = ALIGN(cache_cache.buffer_size, cache_line_size()); + cache_cache.buffer_size = ALIGN(cache_cache.buffer_size, + cache_line_size()); for (order = 0; order < MAX_ORDER; order++) { cache_estimate(order, cache_cache.buffer_size, @@ -1257,24 +1270,26 @@ void __init kmem_cache_init(void) sizes = malloc_sizes; names = cache_names; - /* Initialize the caches that provide memory for the array cache - * and the kmem_list3 structures first. - * Without this, further allocations will bug + /* + * Initialize the caches that provide memory for the array cache and the + * kmem_list3 structures first. Without this, further allocations will + * bug. */ sizes[INDEX_AC].cs_cachep = kmem_cache_create(names[INDEX_AC].name, - sizes[INDEX_AC].cs_size, - ARCH_KMALLOC_MINALIGN, - (ARCH_KMALLOC_FLAGS | - SLAB_PANIC), NULL, NULL); + sizes[INDEX_AC].cs_size, + ARCH_KMALLOC_MINALIGN, + ARCH_KMALLOC_FLAGS|SLAB_PANIC, + NULL, NULL); - if (INDEX_AC != INDEX_L3) + if (INDEX_AC != INDEX_L3) { sizes[INDEX_L3].cs_cachep = - kmem_cache_create(names[INDEX_L3].name, - sizes[INDEX_L3].cs_size, - ARCH_KMALLOC_MINALIGN, - (ARCH_KMALLOC_FLAGS | SLAB_PANIC), NULL, - NULL); + kmem_cache_create(names[INDEX_L3].name, + sizes[INDEX_L3].cs_size, + ARCH_KMALLOC_MINALIGN, + ARCH_KMALLOC_FLAGS|SLAB_PANIC, + NULL, NULL); + } while (sizes->cs_size != ULONG_MAX) { /* @@ -1284,13 +1299,13 @@ void __init kmem_cache_init(void) * Note for systems short on memory removing the alignment will * allow tighter packing of the smaller caches. */ - if (!sizes->cs_cachep) + if (!sizes->cs_cachep) { sizes->cs_cachep = kmem_cache_create(names->name, - sizes->cs_size, - ARCH_KMALLOC_MINALIGN, - (ARCH_KMALLOC_FLAGS - | SLAB_PANIC), - NULL, NULL); + sizes->cs_size, + ARCH_KMALLOC_MINALIGN, + ARCH_KMALLOC_FLAGS|SLAB_PANIC, + NULL, NULL); + } /* Inc off-slab bufctl limit until the ceiling is hit. */ if (!(OFF_SLAB(sizes->cs_cachep))) { @@ -1299,13 +1314,11 @@ void __init kmem_cache_init(void) } sizes->cs_dmacachep = kmem_cache_create(names->name_dma, - sizes->cs_size, - ARCH_KMALLOC_MINALIGN, - (ARCH_KMALLOC_FLAGS | - SLAB_CACHE_DMA | - SLAB_PANIC), NULL, - NULL); - + sizes->cs_size, + ARCH_KMALLOC_MINALIGN, + ARCH_KMALLOC_FLAGS|SLAB_CACHE_DMA| + SLAB_PANIC, + NULL, NULL); sizes++; names++; } @@ -1357,20 +1370,22 @@ void __init kmem_cache_init(void) struct kmem_cache *cachep; mutex_lock(&cache_chain_mutex); list_for_each_entry(cachep, &cache_chain, next) - enable_cpucache(cachep); + enable_cpucache(cachep); mutex_unlock(&cache_chain_mutex); } /* Done! */ g_cpucache_up = FULL; - /* Register a cpu startup notifier callback - * that initializes cpu_cache_get for all new cpus + /* + * Register a cpu startup notifier callback that initializes + * cpu_cache_get for all new cpus */ register_cpu_notifier(&cpucache_notifier); - /* The reap timers are started later, with a module init call: - * That part of the kernel is not yet operational. + /* + * The reap timers are started later, with a module init call: That part + * of the kernel is not yet operational. */ } @@ -1378,16 +1393,13 @@ static int __init cpucache_init(void) { int cpu; - /* - * Register the timers that return unneeded - * pages to gfp. + /* + * Register the timers that return unneeded pages to the page allocator */ for_each_online_cpu(cpu) - start_cpu_timer(cpu); - + start_cpu_timer(cpu); return 0; } - __initcall(cpucache_init); /* @@ -1501,9 +1513,8 @@ static void dump_line(char *data, int offset, int limit) { int i; printk(KERN_ERR "%03x:", offset); - for (i = 0; i < limit; i++) { + for (i = 0; i < limit; i++) printk(" %02x", (unsigned char)data[offset + i]); - } printk("\n"); } #endif @@ -1517,15 +1528,15 @@ static void print_objinfo(struct kmem_cache *cachep, void *objp, int lines) if (cachep->flags & SLAB_RED_ZONE) { printk(KERN_ERR "Redzone: 0x%lx/0x%lx.\n", - *dbg_redzone1(cachep, objp), - *dbg_redzone2(cachep, objp)); + *dbg_redzone1(cachep, objp), + *dbg_redzone2(cachep, objp)); } if (cachep->flags & SLAB_STORE_USER) { printk(KERN_ERR "Last user: [<%p>]", - *dbg_userword(cachep, objp)); + *dbg_userword(cachep, objp)); print_symbol("(%s)", - (unsigned long)*dbg_userword(cachep, objp)); + (unsigned long)*dbg_userword(cachep, objp)); printk("\n"); } realobj = (char *)objp + obj_offset(cachep); @@ -1558,8 +1569,8 @@ static void check_poison_obj(struct kmem_cache *cachep, void *objp) /* Print header */ if (lines == 0) { printk(KERN_ERR - "Slab corruption: start=%p, len=%d\n", - realobj, size); + "Slab corruption: start=%p, len=%d\n", + realobj, size); print_objinfo(cachep, objp, 0); } /* Hexdump the affected line */ @@ -1614,11 +1625,10 @@ static void slab_destroy_objs(struct kmem_cache *cachep, struct slab *slabp) if (cachep->flags & SLAB_POISON) { #ifdef CONFIG_DEBUG_PAGEALLOC - if ((cachep->buffer_size % PAGE_SIZE) == 0 - && OFF_SLAB(cachep)) + if (cachep->buffer_size % PAGE_SIZE == 0 && + OFF_SLAB(cachep)) kernel_map_pages(virt_to_page(objp), - cachep->buffer_size / PAGE_SIZE, - 1); + cachep->buffer_size / PAGE_SIZE, 1); else check_poison_obj(cachep, objp); #else @@ -1650,10 +1660,10 @@ static void slab_destroy_objs(struct kmem_cache *cachep, struct slab *slabp) } #endif -/** +/* * Destroy all the objs in a slab, and release the mem back to the system. - * Before calling the slab must have been unlinked from the cache. - * The cache-lock is not held/needed. + * Before calling the slab must have been unlinked from the cache. The + * cache-lock is not held/needed. */ static void slab_destroy(struct kmem_cache *cachep, struct slab *slabp) { @@ -1674,8 +1684,10 @@ static void slab_destroy(struct kmem_cache *cachep, struct slab *slabp) } } -/* For setting up all the kmem_list3s for cache whose buffer_size is same - as size of kmem_list3. */ +/* + * For setting up all the kmem_list3s for cache whose buffer_size is same as + * size of kmem_list3. + */ static void set_up_list3s(struct kmem_cache *cachep, int index) { int node; @@ -1701,13 +1713,13 @@ static void set_up_list3s(struct kmem_cache *cachep, int index) * high order pages for slabs. When the gfp() functions are more friendly * towards high-order requests, this should be changed. */ -static inline size_t calculate_slab_order(struct kmem_cache *cachep, +static size_t calculate_slab_order(struct kmem_cache *cachep, size_t size, size_t align, unsigned long flags) { size_t left_over = 0; int gfporder; - for (gfporder = 0 ; gfporder <= MAX_GFP_ORDER; gfporder++) { + for (gfporder = 0; gfporder <= MAX_GFP_ORDER; gfporder++) { unsigned int num; size_t remainder; @@ -1742,7 +1754,7 @@ static inline size_t calculate_slab_order(struct kmem_cache *cachep, /* * Acceptable internal fragmentation? */ - if ((left_over * 8) <= (PAGE_SIZE << gfporder)) + if (left_over * 8 <= (PAGE_SIZE << gfporder)) break; } return left_over; @@ -1817,9 +1829,8 @@ static void setup_cpu_cache(struct kmem_cache *cachep) * and the @dtor is run before the pages are handed back. * * @name must be valid until the cache is destroyed. This implies that - * the module calling this has to destroy the cache before getting - * unloaded. - * + * the module calling this has to destroy the cache before getting unloaded. + * * The flags are * * %SLAB_POISON - Poison the slab with a known test pattern (a5a5a5a5) @@ -1837,7 +1848,8 @@ static void setup_cpu_cache(struct kmem_cache *cachep) */ struct kmem_cache * kmem_cache_create (const char *name, size_t size, size_t align, - unsigned long flags, void (*ctor)(void*, struct kmem_cache *, unsigned long), + unsigned long flags, + void (*ctor)(void*, struct kmem_cache *, unsigned long), void (*dtor)(void*, struct kmem_cache *, unsigned long)) { size_t left_over, slab_size, ralign; @@ -1847,12 +1859,10 @@ kmem_cache_create (const char *name, size_t size, size_t align, /* * Sanity checks... these are all serious usage bugs. */ - if ((!name) || - in_interrupt() || - (size < BYTES_PER_WORD) || + if (!name || in_interrupt() || (size < BYTES_PER_WORD) || (size > (1 << MAX_OBJ_ORDER) * PAGE_SIZE) || (dtor && !ctor)) { - printk(KERN_ERR "%s: Early error in slab %s\n", - __FUNCTION__, name); + printk(KERN_ERR "%s: Early error in slab %s\n", __FUNCTION__, + name); BUG(); } @@ -1906,8 +1916,7 @@ kmem_cache_create (const char *name, size_t size, size_t align, * above the next power of two: caches with object sizes just above a * power of two have a significant amount of internal fragmentation. */ - if ((size < 4096 - || fls(size - 1) == fls(size - 1 + 3 * BYTES_PER_WORD))) + if (size < 4096 || fls(size - 1) == fls(size-1 + 3 * BYTES_PER_WORD)) flags |= SLAB_RED_ZONE | SLAB_STORE_USER; if (!(flags & SLAB_DESTROY_BY_RCU)) flags |= SLAB_POISON; @@ -1919,13 +1928,14 @@ kmem_cache_create (const char *name, size_t size, size_t align, BUG_ON(dtor); /* - * Always checks flags, a caller might be expecting debug - * support which isn't available. + * Always checks flags, a caller might be expecting debug support which + * isn't available. */ if (flags & ~CREATE_MASK) BUG(); - /* Check that size is in terms of words. This is needed to avoid + /* + * Check that size is in terms of words. This is needed to avoid * unaligned accesses for some archs when redzoning is used, and makes * sure any on-slab bufctl's are also correctly aligned. */ @@ -1934,12 +1944,14 @@ kmem_cache_create (const char *name, size_t size, size_t align, size &= ~(BYTES_PER_WORD - 1); } - /* calculate out the final buffer alignment: */ + /* calculate the final buffer alignment: */ + /* 1) arch recommendation: can be overridden for debug */ if (flags & SLAB_HWCACHE_ALIGN) { - /* Default alignment: as specified by the arch code. - * Except if an object is really small, then squeeze multiple - * objects into one cacheline. + /* + * Default alignment: as specified by the arch code. Except if + * an object is really small, then squeeze multiple objects into + * one cacheline. */ ralign = cache_line_size(); while (size <= ralign / 2) @@ -1959,7 +1971,8 @@ kmem_cache_create (const char *name, size_t size, size_t align, if (ralign > BYTES_PER_WORD) flags &= ~(SLAB_RED_ZONE | SLAB_STORE_USER); } - /* 4) Store it. Note that the debug code below can reduce + /* + * 4) Store it. Note that the debug code below can reduce * the alignment to BYTES_PER_WORD. */ align = ralign; @@ -2058,7 +2071,7 @@ kmem_cache_create (const char *name, size_t size, size_t align, /* cache setup completed, link it into the list */ list_add(&cachep->next, &cache_chain); - oops: +oops: if (!cachep && (flags & SLAB_PANIC)) panic("kmem_cache_create(): failed to create slab `%s'\n", name); @@ -2109,7 +2122,6 @@ static void smp_call_function_all_cpus(void (*func)(void *arg), void *arg) { check_irq_on(); preempt_disable(); - local_irq_disable(); func(arg); local_irq_enable(); @@ -2120,12 +2132,12 @@ static void smp_call_function_all_cpus(void (*func)(void *arg), void *arg) preempt_enable(); } -static void drain_array_locked(struct kmem_cache *cachep, struct array_cache *ac, - int force, int node); +static void drain_array_locked(struct kmem_cache *cachep, + struct array_cache *ac, int force, int node); static void do_drain(void *arg) { - struct kmem_cache *cachep = (struct kmem_cache *) arg; + struct kmem_cache *cachep = arg; struct array_cache *ac; int node = numa_node_id(); @@ -2273,16 +2285,15 @@ int kmem_cache_destroy(struct kmem_cache *cachep) /* NUMA: free the list3 structures */ for_each_online_node(i) { - if ((l3 = cachep->nodelists[i])) { + l3 = cachep->nodelists[i]; + if (l3) { kfree(l3->shared); free_alien_cache(l3->alien); kfree(l3); } } kmem_cache_free(&cache_cache, cachep); - unlock_cpu_hotplug(); - return 0; } EXPORT_SYMBOL(kmem_cache_destroy); @@ -2305,7 +2316,6 @@ static struct slab *alloc_slabmgmt(struct kmem_cache *cachep, void *objp, slabp->inuse = 0; slabp->colouroff = colour_off; slabp->s_mem = objp + colour_off; - return slabp; } @@ -2333,9 +2343,9 @@ static void cache_init_objs(struct kmem_cache *cachep, *dbg_redzone2(cachep, objp) = RED_INACTIVE; } /* - * Constructors are not allowed to allocate memory from - * the same cache which they are a constructor for. - * Otherwise, deadlock. They must also be threaded. + * Constructors are not allowed to allocate memory from the same + * cache which they are a constructor for. Otherwise, deadlock. + * They must also be threaded. */ if (cachep->ctor && !(cachep->flags & SLAB_POISON)) cachep->ctor(objp + obj_offset(cachep), cachep, @@ -2349,8 +2359,8 @@ static void cache_init_objs(struct kmem_cache *cachep, slab_error(cachep, "constructor overwrote the" " start of an object"); } - if ((cachep->buffer_size % PAGE_SIZE) == 0 && OFF_SLAB(cachep) - && cachep->flags & SLAB_POISON) + if ((cachep->buffer_size % PAGE_SIZE) == 0 && + OFF_SLAB(cachep) && cachep->flags & SLAB_POISON) kernel_map_pages(virt_to_page(objp), cachep->buffer_size / PAGE_SIZE, 0); #else @@ -2365,16 +2375,14 @@ static void cache_init_objs(struct kmem_cache *cachep, static void kmem_flagcheck(struct kmem_cache *cachep, gfp_t flags) { - if (flags & SLAB_DMA) { - if (!(cachep->gfpflags & GFP_DMA)) - BUG(); - } else { - if (cachep->gfpflags & GFP_DMA) - BUG(); - } + if (flags & SLAB_DMA) + BUG_ON(!(cachep->gfpflags & GFP_DMA)); + else + BUG_ON(cachep->gfpflags & GFP_DMA); } -static void *slab_get_obj(struct kmem_cache *cachep, struct slab *slabp, int nodeid) +static void *slab_get_obj(struct kmem_cache *cachep, struct slab *slabp, + int nodeid) { void *objp = index_to_obj(cachep, slabp, slabp->free); kmem_bufctl_t next; @@ -2390,8 +2398,8 @@ static void *slab_get_obj(struct kmem_cache *cachep, struct slab *slabp, int nod return objp; } -static void slab_put_obj(struct kmem_cache *cachep, struct slab *slabp, void *objp, - int nodeid) +static void slab_put_obj(struct kmem_cache *cachep, struct slab *slabp, + void *objp, int nodeid) { unsigned int objnr = obj_to_index(cachep, slabp, objp); @@ -2401,7 +2409,7 @@ static void slab_put_obj(struct kmem_cache *cachep, struct slab *slabp, void *ob if (slab_bufctl(slabp)[objnr] != BUFCTL_FREE) { printk(KERN_ERR "slab: double free detected in cache " - "'%s', objp %p\n", cachep->name, objp); + "'%s', objp %p\n", cachep->name, objp); BUG(); } #endif @@ -2410,7 +2418,8 @@ static void slab_put_obj(struct kmem_cache *cachep, struct slab *slabp, void *ob slabp->inuse--; } -static void set_slab_attr(struct kmem_cache *cachep, struct slab *slabp, void *objp) +static void set_slab_attr(struct kmem_cache *cachep, struct slab *slabp, + void *objp) { int i; struct page *page; @@ -2438,8 +2447,9 @@ static int cache_grow(struct kmem_cache *cachep, gfp_t flags, int nodeid) unsigned long ctor_flags; struct kmem_list3 *l3; - /* Be lazy and only check for valid flags here, - * keeping it out of the critical path in kmem_cache_alloc(). + /* + * Be lazy and only check for valid flags here, keeping it out of the + * critical path in kmem_cache_alloc(). */ if (flags & ~(SLAB_DMA | SLAB_LEVEL_MASK | SLAB_NO_GROW)) BUG(); @@ -2480,14 +2490,17 @@ static int cache_grow(struct kmem_cache *cachep, gfp_t flags, int nodeid) */ kmem_flagcheck(cachep, flags); - /* Get mem for the objs. - * Attempt to allocate a physical page from 'nodeid', + /* + * Get mem for the objs. Attempt to allocate a physical page from + * 'nodeid'. */ - if (!(objp = kmem_getpages(cachep, flags, nodeid))) + objp = kmem_getpages(cachep, flags, nodeid); + if (!objp) goto failed; /* Get slab management. */ - if (!(slabp = alloc_slabmgmt(cachep, objp, offset, local_flags))) + slabp = alloc_slabmgmt(cachep, objp, offset, local_flags); + if (!slabp) goto opps1; slabp->nodeid = nodeid; @@ -2506,9 +2519,9 @@ static int cache_grow(struct kmem_cache *cachep, gfp_t flags, int nodeid) l3->free_objects += cachep->num; spin_unlock(&l3->list_lock); return 1; - opps1: +opps1: kmem_freepages(cachep, objp); - failed: +failed: if (local_flags & __GFP_WAIT) local_irq_disable(); return 0; @@ -2551,8 +2564,8 @@ static void *cache_free_debugcheck(struct kmem_cache *cachep, void *objp, page = virt_to_page(objp); if (page_get_cache(page) != cachep) { - printk(KERN_ERR - "mismatch in kmem_cache_free: expected cache %p, got %p\n", + printk(KERN_ERR "mismatch in kmem_cache_free: expected " + "cache %p, got %p\n", page_get_cache(page), cachep); printk(KERN_ERR "%p is %s.\n", cachep, cachep->name); printk(KERN_ERR "%p is %s.\n", page_get_cache(page), @@ -2562,13 +2575,12 @@ static void *cache_free_debugcheck(struct kmem_cache *cachep, void *objp, slabp = page_get_slab(page); if (cachep->flags & SLAB_RED_ZONE) { - if (*dbg_redzone1(cachep, objp) != RED_ACTIVE - || *dbg_redzone2(cachep, objp) != RED_ACTIVE) { - slab_error(cachep, - "double free, or memory outside" - " object was overwritten"); - printk(KERN_ERR - "%p: redzone 1: 0x%lx, redzone 2: 0x%lx.\n", + if (*dbg_redzone1(cachep, objp) != RED_ACTIVE || + *dbg_redzone2(cachep, objp) != RED_ACTIVE) { + slab_error(cachep, "double free, or memory outside" + " object was overwritten"); + printk(KERN_ERR "%p: redzone 1:0x%lx, " + "redzone 2:0x%lx.\n", objp, *dbg_redzone1(cachep, objp), *dbg_redzone2(cachep, objp)); } @@ -2584,9 +2596,10 @@ static void *cache_free_debugcheck(struct kmem_cache *cachep, void *objp, BUG_ON(objp != index_to_obj(cachep, slabp, objnr)); if (cachep->flags & SLAB_DEBUG_INITIAL) { - /* Need to call the slab's constructor so the - * caller can perform a verify of its state (debugging). - * Called without the cache-lock held. + /* + * Need to call the slab's constructor so the caller can + * perform a verify of its state (debugging). Called without + * the cache-lock held. */ cachep->ctor(objp + obj_offset(cachep), cachep, SLAB_CTOR_CONSTRUCTOR | SLAB_CTOR_VERIFY); @@ -2599,7 +2612,7 @@ static void *cache_free_debugcheck(struct kmem_cache *cachep, void *objp, } if (cachep->flags & SLAB_POISON) { #ifdef CONFIG_DEBUG_PAGEALLOC - if ((cachep->buffer_size % PAGE_SIZE) == 0 && OFF_SLAB(cachep)) { + if ((cachep->buffer_size % PAGE_SIZE)==0 && OFF_SLAB(cachep)) { store_stackinfo(cachep, objp, (unsigned long)caller); kernel_map_pages(virt_to_page(objp), cachep->buffer_size / PAGE_SIZE, 0); @@ -2625,14 +2638,14 @@ static void check_slabp(struct kmem_cache *cachep, struct slab *slabp) goto bad; } if (entries != cachep->num - slabp->inuse) { - bad: - printk(KERN_ERR - "slab: Internal list corruption detected in cache '%s'(%d), slabp %p(%d). Hexdump:\n", - cachep->name, cachep->num, slabp, slabp->inuse); +bad: + printk(KERN_ERR "slab: Internal list corruption detected in " + "cache '%s'(%d), slabp %p(%d). Hexdump:\n", + cachep->name, cachep->num, slabp, slabp->inuse); for (i = 0; i < sizeof(*slabp) + cachep->num * sizeof(kmem_bufctl_t); i++) { - if ((i % 16) == 0) + if (i % 16 == 0) printk("\n%03x:", i); printk(" %02x", ((unsigned char *)slabp)[i]); } @@ -2654,12 +2667,13 @@ static void *cache_alloc_refill(struct kmem_cache *cachep, gfp_t flags) check_irq_off(); ac = cpu_cache_get(cachep); - retry: +retry: batchcount = ac->batchcount; if (!ac->touched && batchcount > BATCHREFILL_LIMIT) { - /* if there was little recent activity on this - * cache, then perform only a partial refill. - * Otherwise we could generate refill bouncing. + /* + * If there was little recent activity on this cache, then + * perform only a partial refill. Otherwise we could generate + * refill bouncing. */ batchcount = BATCHREFILL_LIMIT; } @@ -2715,29 +2729,29 @@ static void *cache_alloc_refill(struct kmem_cache *cachep, gfp_t flags) list_add(&slabp->list, &l3->slabs_partial); } - must_grow: +must_grow: l3->free_objects -= ac->avail; - alloc_done: +alloc_done: spin_unlock(&l3->list_lock); if (unlikely(!ac->avail)) { int x; x = cache_grow(cachep, flags, numa_node_id()); - // cache_grow can reenable interrupts, then ac could change. + /* cache_grow can reenable interrupts, then ac could change. */ ac = cpu_cache_get(cachep); - if (!x && ac->avail == 0) // no objects in sight? abort + if (!x && ac->avail == 0) /* no objects in sight? abort */ return NULL; - if (!ac->avail) // objects refilled by interrupt? + if (!ac->avail) /* objects refilled by interrupt? */ goto retry; } ac->touched = 1; return ac->entry[--ac->avail]; } -static inline void -cache_alloc_debugcheck_before(struct kmem_cache *cachep, gfp_t flags) +static inline void cache_alloc_debugcheck_before(struct kmem_cache *cachep, + gfp_t flags) { might_sleep_if(flags & __GFP_WAIT); #if DEBUG @@ -2746,8 +2760,8 @@ cache_alloc_debugcheck_before(struct kmem_cache *cachep, gfp_t flags) } #if DEBUG -static void *cache_alloc_debugcheck_after(struct kmem_cache *cachep, gfp_t flags, - void *objp, void *caller) +static void *cache_alloc_debugcheck_after(struct kmem_cache *cachep, + gfp_t flags, void *objp, void *caller) { if (!objp) return objp; @@ -2767,15 +2781,14 @@ static void *cache_alloc_debugcheck_after(struct kmem_cache *cachep, gfp_t flags *dbg_userword(cachep, objp) = caller; if (cachep->flags & SLAB_RED_ZONE) { - if (*dbg_redzone1(cachep, objp) != RED_INACTIVE - || *dbg_redzone2(cachep, objp) != RED_INACTIVE) { - slab_error(cachep, - "double free, or memory outside" - " object was overwritten"); + if (*dbg_redzone1(cachep, objp) != RED_INACTIVE || + *dbg_redzone2(cachep, objp) != RED_INACTIVE) { + slab_error(cachep, "double free, or memory outside" + " object was overwritten"); printk(KERN_ERR - "%p: redzone 1: 0x%lx, redzone 2: 0x%lx.\n", - objp, *dbg_redzone1(cachep, objp), - *dbg_redzone2(cachep, objp)); + "%p: redzone 1:0x%lx, redzone 2:0x%lx\n", + objp, *dbg_redzone1(cachep, objp), + *dbg_redzone2(cachep, objp)); } *dbg_redzone1(cachep, objp) = RED_ACTIVE; *dbg_redzone2(cachep, objp) = RED_ACTIVE; @@ -2822,8 +2835,8 @@ static inline void *____cache_alloc(struct kmem_cache *cachep, gfp_t flags) return objp; } -static __always_inline void * -__cache_alloc(struct kmem_cache *cachep, gfp_t flags, void *caller) +static __always_inline void *__cache_alloc(struct kmem_cache *cachep, + gfp_t flags, void *caller) { unsigned long save_flags; void *objp; @@ -2843,7 +2856,8 @@ __cache_alloc(struct kmem_cache *cachep, gfp_t flags, void *caller) /* * A interface to enable slab creation on nodeid */ -static void *__cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int nodeid) +static void *__cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, + int nodeid) { struct list_head *entry; struct slab *slabp; @@ -2854,7 +2868,7 @@ static void *__cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int node l3 = cachep->nodelists[nodeid]; BUG_ON(!l3); - retry: +retry: check_irq_off(); spin_lock(&l3->list_lock); entry = l3->slabs_partial.next; @@ -2881,16 +2895,15 @@ static void *__cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int node /* move slabp to correct slabp list: */ list_del(&slabp->list); - if (slabp->free == BUFCTL_END) { + if (slabp->free == BUFCTL_END) list_add(&slabp->list, &l3->slabs_full); - } else { + else list_add(&slabp->list, &l3->slabs_partial); - } spin_unlock(&l3->list_lock); goto done; - must_grow: +must_grow: spin_unlock(&l3->list_lock); x = cache_grow(cachep, flags, nodeid); @@ -2898,7 +2911,7 @@ static void *__cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int node return NULL; goto retry; - done: +done: return obj; } #endif @@ -2971,7 +2984,7 @@ static void cache_flusharray(struct kmem_cache *cachep, struct array_cache *ac) } free_block(cachep, ac->entry, batchcount, node); - free_done: +free_done: #if STATS { int i = 0; @@ -2992,16 +3005,12 @@ static void cache_flusharray(struct kmem_cache *cachep, struct array_cache *ac) #endif spin_unlock(&l3->list_lock); ac->avail -= batchcount; - memmove(ac->entry, &(ac->entry[batchcount]), - sizeof(void *) * ac->avail); + memmove(ac->entry, &(ac->entry[batchcount]), sizeof(void *)*ac->avail); } /* - * __cache_free - * Release an obj back to its cache. If the obj has a constructed - * state, it must be in this state _before_ it is released. - * - * Called with disabled ints. + * Release an obj back to its cache. If the obj has a constructed state, it must + * be in this state _before_ it is released. Called with disabled ints. */ static inline void __cache_free(struct kmem_cache *cachep, void *objp) { @@ -3020,9 +3029,9 @@ static inline void __cache_free(struct kmem_cache *cachep, void *objp) if (unlikely(slabp->nodeid != numa_node_id())) { struct array_cache *alien = NULL; int nodeid = slabp->nodeid; - struct kmem_list3 *l3 = - cachep->nodelists[numa_node_id()]; + struct kmem_list3 *l3; + l3 = cachep->nodelists[numa_node_id()]; STATS_INC_NODEFREES(cachep); if (l3->alien && l3->alien[nodeid]) { alien = l3->alien[nodeid]; @@ -3106,7 +3115,7 @@ int fastcall kmem_ptr_validate(struct kmem_cache *cachep, void *ptr) if (unlikely(page_get_cache(page) != cachep)) goto out; return 1; - out: +out: return 0; } @@ -3132,7 +3141,7 @@ void *kmem_cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int nodeid) local_irq_save(save_flags); if (nodeid == -1 || nodeid == numa_node_id() || - !cachep->nodelists[nodeid]) + !cachep->nodelists[nodeid]) ptr = ____cache_alloc(cachep, flags); else ptr = __cache_alloc_node(cachep, flags, nodeid); @@ -3249,7 +3258,7 @@ void *__alloc_percpu(size_t size) /* Catch derefs w/o wrappers */ return (void *)(~(unsigned long)pdata); - unwind_oom: +unwind_oom: while (--i >= 0) { if (!cpu_possible(i)) continue; @@ -3352,18 +3361,20 @@ static int alloc_kmemlist(struct kmem_cache *cachep) struct array_cache *nc = NULL, *new; struct array_cache **new_alien = NULL; #ifdef CONFIG_NUMA - if (!(new_alien = alloc_alien_cache(node, cachep->limit))) + new_alien = alloc_alien_cache(node, cachep->limit); + if (!new_alien) goto fail; #endif - if (!(new = alloc_arraycache(node, (cachep->shared * - cachep->batchcount), - 0xbaadf00d))) + new = alloc_arraycache(node, cachep->shared*cachep->batchcount, + 0xbaadf00d); + if (!new) goto fail; - if ((l3 = cachep->nodelists[node])) { - + l3 = cachep->nodelists[node]; + if (l3) { spin_lock_irq(&l3->list_lock); - if ((nc = cachep->nodelists[node]->shared)) + nc = cachep->nodelists[node]->shared; + if (nc) free_block(cachep, nc->entry, nc->avail, node); l3->shared = new; @@ -3372,27 +3383,27 @@ static int alloc_kmemlist(struct kmem_cache *cachep) new_alien = NULL; } l3->free_limit = (1 + nr_cpus_node(node)) * - cachep->batchcount + cachep->num; + cachep->batchcount + cachep->num; spin_unlock_irq(&l3->list_lock); kfree(nc); free_alien_cache(new_alien); continue; } - if (!(l3 = kmalloc_node(sizeof(struct kmem_list3), - GFP_KERNEL, node))) + l3 = kmalloc_node(sizeof(struct kmem_list3), GFP_KERNEL, node); + if (!l3) goto fail; kmem_list3_init(l3); l3->next_reap = jiffies + REAPTIMEOUT_LIST3 + - ((unsigned long)cachep) % REAPTIMEOUT_LIST3; + ((unsigned long)cachep) % REAPTIMEOUT_LIST3; l3->shared = new; l3->alien = new_alien; l3->free_limit = (1 + nr_cpus_node(node)) * - cachep->batchcount + cachep->num; + cachep->batchcount + cachep->num; cachep->nodelists[node] = l3; } return err; - fail: +fail: err = -ENOMEM; return err; } @@ -3404,7 +3415,7 @@ struct ccupdate_struct { static void do_ccupdate_local(void *info) { - struct ccupdate_struct *new = (struct ccupdate_struct *)info; + struct ccupdate_struct *new = info; struct array_cache *old; check_irq_off(); @@ -3414,16 +3425,16 @@ static void do_ccupdate_local(void *info) new->new[smp_processor_id()] = old; } -static int do_tune_cpucache(struct kmem_cache *cachep, int limit, int batchcount, - int shared) +static int do_tune_cpucache(struct kmem_cache *cachep, int limit, + int batchcount, int shared) { struct ccupdate_struct new; int i, err; memset(&new.new, 0, sizeof(new.new)); for_each_online_cpu(i) { - new.new[i] = - alloc_arraycache(cpu_to_node(i), limit, batchcount); + new.new[i] = alloc_arraycache(cpu_to_node(i), limit, + batchcount); if (!new.new[i]) { for (i--; i >= 0; i--) kfree(new.new[i]); @@ -3465,10 +3476,11 @@ static void enable_cpucache(struct kmem_cache *cachep) int err; int limit, shared; - /* The head array serves three purposes: + /* + * The head array serves three purposes: * - create a LIFO ordering, i.e. return objects that are cache-warm * - reduce the number of spinlock operations. - * - reduce the number of linked list operations on the slab and + * - reduce the number of linked list operations on the slab and * bufctl chains: array operations are cheaper. * The numbers are guessed, we should auto-tune as described by * Bonwick. @@ -3484,7 +3496,8 @@ static void enable_cpucache(struct kmem_cache *cachep) else limit = 120; - /* Cpu bound tasks (e.g. network routing) can exhibit cpu bound + /* + * CPU bound tasks (e.g. network routing) can exhibit cpu bound * allocation behaviour: Most allocs on one cpu, most free operations * on another cpu. For these cases, an efficient object passing between * cpus is necessary. This is provided by a shared array. The array @@ -3499,9 +3512,9 @@ static void enable_cpucache(struct kmem_cache *cachep) #endif #if DEBUG - /* With debugging enabled, large batchcount lead to excessively - * long periods with disabled local interrupts. Limit the - * batchcount + /* + * With debugging enabled, large batchcount lead to excessively long + * periods with disabled local interrupts. Limit the batchcount */ if (limit > 32) limit = 32; @@ -3512,8 +3525,8 @@ static void enable_cpucache(struct kmem_cache *cachep) cachep->name, -err); } -static void drain_array_locked(struct kmem_cache *cachep, struct array_cache *ac, - int force, int node) +static void drain_array_locked(struct kmem_cache *cachep, + struct array_cache *ac, int force, int node) { int tofree; @@ -3522,9 +3535,8 @@ static void drain_array_locked(struct kmem_cache *cachep, struct array_cache *ac ac->touched = 0; } else if (ac->avail) { tofree = force ? ac->avail : (ac->limit + 4) / 5; - if (tofree > ac->avail) { + if (tofree > ac->avail) tofree = (ac->avail + 1) / 2; - } free_block(cachep, ac->entry, tofree, node); ac->avail -= tofree; memmove(ac->entry, &(ac->entry[tofree]), @@ -3541,8 +3553,8 @@ static void drain_array_locked(struct kmem_cache *cachep, struct array_cache *ac * - clear the per-cpu caches for this CPU. * - return freeable pages to the main free memory pool. * - * If we cannot acquire the cache chain mutex then just give up - we'll - * try again on the next iteration. + * If we cannot acquire the cache chain mutex then just give up - we'll try + * again on the next iteration. */ static void cache_reap(void *unused) { @@ -3590,9 +3602,8 @@ static void cache_reap(void *unused) goto next_unlock; } - tofree = - (l3->free_limit + 5 * searchp->num - - 1) / (5 * searchp->num); + tofree = (l3->free_limit + 5 * searchp->num - 1) / + (5 * searchp->num); do { p = l3->slabs_free.next; if (p == &(l3->slabs_free)) @@ -3603,9 +3614,9 @@ static void cache_reap(void *unused) list_del(&slabp->list); STATS_INC_REAPED(searchp); - /* Safe to drop the lock. The slab is no longer - * linked to the cache. - * searchp cannot disappear, we hold + /* + * Safe to drop the lock. The slab is no longer linked + * to the cache. searchp cannot disappear, we hold * cache_chain_lock */ l3->free_objects -= searchp->num; @@ -3613,15 +3624,15 @@ static void cache_reap(void *unused) slab_destroy(searchp, slabp); spin_lock_irq(&l3->list_lock); } while (--tofree > 0); - next_unlock: +next_unlock: spin_unlock_irq(&l3->list_lock); - next: +next: cond_resched(); } check_irq_on(); mutex_unlock(&cache_chain_mutex); next_reap_node(); - /* Setup the next iteration */ + /* Set up the next iteration */ schedule_delayed_work(&__get_cpu_var(reap_work), REAPTIMEOUT_CPUC); } @@ -3671,8 +3682,8 @@ static void *s_next(struct seq_file *m, void *p, loff_t *pos) { struct kmem_cache *cachep = p; ++*pos; - return cachep->next.next == &cache_chain ? NULL - : list_entry(cachep->next.next, struct kmem_cache, next); + return cachep->next.next == &cache_chain ? + NULL : list_entry(cachep->next.next, struct kmem_cache, next); } static void s_stop(struct seq_file *m, void *p) @@ -3761,7 +3772,9 @@ static int s_show(struct seq_file *m, void *p) unsigned long node_frees = cachep->node_frees; seq_printf(m, " : globalstat %7lu %6lu %5lu %4lu \ - %4lu %4lu %4lu %4lu", allocs, high, grown, reaped, errors, max_freeable, node_allocs, node_frees); + %4lu %4lu %4lu %4lu", allocs, high, grown, + reaped, errors, max_freeable, node_allocs, + node_frees); } /* cpu stats */ { @@ -3833,13 +3846,12 @@ ssize_t slabinfo_write(struct file *file, const char __user * buffer, mutex_lock(&cache_chain_mutex); res = -EINVAL; list_for_each(p, &cache_chain) { - struct kmem_cache *cachep = list_entry(p, struct kmem_cache, - next); + struct kmem_cache *cachep; + cachep = list_entry(p, struct kmem_cache, next); if (!strcmp(cachep->name, kbuf)) { - if (limit < 1 || - batchcount < 1 || - batchcount > limit || shared < 0) { + if (limit < 1 || batchcount < 1 || + batchcount > limit || shared < 0) { res = 0; } else { res = do_tune_cpucache(cachep, limit, -- cgit From b5d8ca7c50826c0b456b4a646875dc573adfde2b Mon Sep 17 00:00:00 2001 From: Ravikiran G Thirumalai Date: Wed, 22 Mar 2006 00:08:12 -0800 Subject: [PATCH] slab: remove cachep->spinlock Remove cachep->spinlock. Locking has moved to the kmem_list3 and most of the structures protected earlier by cachep->spinlock is now protected by the l3->list_lock. slab cache tunables like batchcount are accessed always with the cache_chain_mutex held. Patch tested on SMP and NUMA kernels with dbench processes running, constant onlining/offlining, and constant cache tuning, all at the same time. Signed-off-by: Ravikiran Thirumalai Cc: Christoph Lameter Cc: Pekka Enberg Cc: Manfred Spraul Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 20 +++++++++----------- 1 file changed, 9 insertions(+), 11 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 7b6f9f10e75..2cd80203984 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -372,17 +372,19 @@ static void kmem_list3_init(struct kmem_list3 *parent) struct kmem_cache { /* 1) per-cpu data, touched during every alloc/free */ struct array_cache *array[NR_CPUS]; +/* 2) Cache tunables. Protected by cache_chain_mutex */ unsigned int batchcount; unsigned int limit; unsigned int shared; + unsigned int buffer_size; -/* 2) touched by every alloc & free from the backend */ +/* 3) touched by every alloc & free from the backend */ struct kmem_list3 *nodelists[MAX_NUMNODES]; + unsigned int flags; /* constant flags */ unsigned int num; /* # of objs per slab */ - spinlock_t spinlock; -/* 3) cache_grow/shrink */ +/* 4) cache_grow/shrink */ /* order of pgs per slab (2^n) */ unsigned int gfporder; @@ -401,11 +403,11 @@ struct kmem_cache { /* de-constructor func */ void (*dtor) (void *, struct kmem_cache *, unsigned long); -/* 4) cache creation/removal */ +/* 5) cache creation/removal */ const char *name; struct list_head next; -/* 5) statistics */ +/* 6) statistics */ #if STATS unsigned long num_active; unsigned long num_allocations; @@ -661,7 +663,6 @@ static struct kmem_cache cache_cache = { .shared = 1, .buffer_size = sizeof(struct kmem_cache), .flags = SLAB_NO_REAP, - .spinlock = SPIN_LOCK_UNLOCKED, .name = "kmem_cache", #if DEBUG .obj_size = sizeof(struct kmem_cache), @@ -2057,7 +2058,6 @@ kmem_cache_create (const char *name, size_t size, size_t align, cachep->gfpflags = 0; if (flags & SLAB_CACHE_DMA) cachep->gfpflags |= GFP_DMA; - spin_lock_init(&cachep->spinlock); cachep->buffer_size = size; if (flags & CFLGS_OFF_SLAB) @@ -3425,6 +3425,7 @@ static void do_ccupdate_local(void *info) new->new[smp_processor_id()] = old; } +/* Always called with the cache_chain_mutex held */ static int do_tune_cpucache(struct kmem_cache *cachep, int limit, int batchcount, int shared) { @@ -3446,11 +3447,9 @@ static int do_tune_cpucache(struct kmem_cache *cachep, int limit, smp_call_function_all_cpus(do_ccupdate_local, (void *)&new); check_irq_on(); - spin_lock(&cachep->spinlock); cachep->batchcount = batchcount; cachep->limit = limit; cachep->shared = shared; - spin_unlock(&cachep->spinlock); for_each_online_cpu(i) { struct array_cache *ccold = new.new[i]; @@ -3471,6 +3470,7 @@ static int do_tune_cpucache(struct kmem_cache *cachep, int limit, return 0; } +/* Called with cache_chain_mutex held always */ static void enable_cpucache(struct kmem_cache *cachep) { int err; @@ -3705,7 +3705,6 @@ static int s_show(struct seq_file *m, void *p) int node; struct kmem_list3 *l3; - spin_lock(&cachep->spinlock); active_objs = 0; num_slabs = 0; for_each_online_node(node) { @@ -3788,7 +3787,6 @@ static int s_show(struct seq_file *m, void *p) } #endif seq_putc(m, '\n'); - spin_unlock(&cachep->spinlock); return 0; } -- cgit From 911851e6ee6ac4e26f07be342a89632f78494fef Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Wed, 22 Mar 2006 00:08:14 -0800 Subject: [PATCH] slab: fix kernel-doc warnings Fix kernel-doc warnings in mm/slab.c. Signed-off-by: Randy Dunlap Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 2cd80203984..5c257498983 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -1615,8 +1615,12 @@ static void check_poison_obj(struct kmem_cache *cachep, void *objp) #if DEBUG /** - * slab_destroy_objs - call the registered destructor for each object in - * a slab that is to be destroyed. + * slab_destroy_objs - destroy a slab and its objects + * @cachep: cache pointer being destroyed + * @slabp: slab pointer being destroyed + * + * Call the registered destructor for each object in a slab that is being + * destroyed. */ static void slab_destroy_objs(struct kmem_cache *cachep, struct slab *slabp) { @@ -1661,7 +1665,11 @@ static void slab_destroy_objs(struct kmem_cache *cachep, struct slab *slabp) } #endif -/* +/** + * slab_destroy - destroy and release all objects in a slab + * @cachep: cache pointer being destroyed + * @slabp: slab pointer being destroyed + * * Destroy all the objs in a slab, and release the mem back to the system. * Before calling the slab must have been unlinked from the cache. The * cache-lock is not held/needed. @@ -3170,6 +3178,7 @@ EXPORT_SYMBOL(kmalloc_node); * kmalloc - allocate memory * @size: how many bytes of memory are required. * @flags: the type of memory to allocate. + * @caller: function caller for debug tracking of the caller * * kmalloc is the normal method of allocating memory * in the kernel. -- cgit From ac2b898ca6fb06196a26869c23b66afe7944e52e Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Wed, 22 Mar 2006 00:08:15 -0800 Subject: [PATCH] slab: Remove SLAB_NO_REAP option SLAB_NO_REAP is documented as an option that will cause this slab not to be reaped under memory pressure. However, that is not what happens. The only thing that SLAB_NO_REAP controls at the moment is the reclaim of the unused slab elements that were allocated in batch in cache_reap(). Cache_reap() is run every few seconds independently of memory pressure. Could we remove the whole thing? Its only used by three slabs anyways and I cannot find a reason for having this option. There is an additional problem with SLAB_NO_REAP. If set then the recovery of objects from alien caches is switched off. Objects not freed on the same node where they were initially allocated will only be reused if a certain amount of objects accumulates from one alien node (not very likely) or if the cache is explicitly shrunk. (Strangely __cache_shrink does not check for SLAB_NO_REAP) Getting rid of SLAB_NO_REAP fixes the problems with alien cache freeing. Signed-off-by: Christoph Lameter Cc: Pekka Enberg Cc: Manfred Spraul Cc: Mark Fasheh Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 13 ++----------- 1 file changed, 2 insertions(+), 11 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 5c257498983..24235506b2a 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -170,12 +170,12 @@ #if DEBUG # define CREATE_MASK (SLAB_DEBUG_INITIAL | SLAB_RED_ZONE | \ SLAB_POISON | SLAB_HWCACHE_ALIGN | \ - SLAB_NO_REAP | SLAB_CACHE_DMA | \ + SLAB_CACHE_DMA | \ SLAB_MUST_HWCACHE_ALIGN | SLAB_STORE_USER | \ SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | \ SLAB_DESTROY_BY_RCU) #else -# define CREATE_MASK (SLAB_HWCACHE_ALIGN | SLAB_NO_REAP | \ +# define CREATE_MASK (SLAB_HWCACHE_ALIGN | \ SLAB_CACHE_DMA | SLAB_MUST_HWCACHE_ALIGN | \ SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | \ SLAB_DESTROY_BY_RCU) @@ -662,7 +662,6 @@ static struct kmem_cache cache_cache = { .limit = BOOT_CPUCACHE_ENTRIES, .shared = 1, .buffer_size = sizeof(struct kmem_cache), - .flags = SLAB_NO_REAP, .name = "kmem_cache", #if DEBUG .obj_size = sizeof(struct kmem_cache), @@ -1848,9 +1847,6 @@ static void setup_cpu_cache(struct kmem_cache *cachep) * %SLAB_RED_ZONE - Insert `Red' zones around the allocated memory to check * for buffer overruns. * - * %SLAB_NO_REAP - Don't automatically reap this cache when we're under - * memory pressure. - * * %SLAB_HWCACHE_ALIGN - Align the objects in this cache to a hardware * cacheline. This can be beneficial if you're counting cycles as closely * as davem. @@ -3584,10 +3580,6 @@ static void cache_reap(void *unused) struct slab *slabp; searchp = list_entry(walk, struct kmem_cache, next); - - if (searchp->flags & SLAB_NO_REAP) - goto next; - check_irq_on(); l3 = searchp->nodelists[numa_node_id()]; @@ -3635,7 +3627,6 @@ static void cache_reap(void *unused) } while (--tofree > 0); next_unlock: spin_unlock_irq(&l3->list_lock); -next: cond_resched(); } check_irq_on(); -- cgit From a07fa3944bf924881450884224cbb2f1269cb9fa Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Wed, 22 Mar 2006 00:08:17 -0800 Subject: [PATCH] slab: use on_each_cpu() Slab duplicates on_each_cpu(). Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 21 ++------------------- 1 file changed, 2 insertions(+), 19 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 24235506b2a..f477acfb732 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -2119,23 +2119,6 @@ static void check_spinlock_acquired_node(struct kmem_cache *cachep, int node) #define check_spinlock_acquired_node(x, y) do { } while(0) #endif -/* - * Waits for all CPUs to execute func(). - */ -static void smp_call_function_all_cpus(void (*func)(void *arg), void *arg) -{ - check_irq_on(); - preempt_disable(); - local_irq_disable(); - func(arg); - local_irq_enable(); - - if (smp_call_function(func, arg, 1, 1)) - BUG(); - - preempt_enable(); -} - static void drain_array_locked(struct kmem_cache *cachep, struct array_cache *ac, int force, int node); @@ -2158,7 +2141,7 @@ static void drain_cpu_caches(struct kmem_cache *cachep) struct kmem_list3 *l3; int node; - smp_call_function_all_cpus(do_drain, cachep); + on_each_cpu(do_drain, cachep, 1, 1); check_irq_on(); for_each_online_node(node) { l3 = cachep->nodelists[node]; @@ -3449,7 +3432,7 @@ static int do_tune_cpucache(struct kmem_cache *cachep, int limit, } new.cachep = cachep; - smp_call_function_all_cpus(do_ccupdate_local, (void *)&new); + on_each_cpu(do_ccupdate_local, (void *)&new, 1, 1); check_irq_on(); cachep->batchcount = batchcount; -- cgit From 84097518d1ecd2330f9488e4c2d09953a3340e74 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Wed, 22 Mar 2006 00:08:34 -0800 Subject: [PATCH] mm: nommu use compound pages Now that compound page handling is properly fixed in the VM, move nommu over to using compound pages rather than rolling their own refcounting. nommu vm page refcounting is broken anyway, but there is no need to have divergent code in the core VM now, nor when it gets fixed. Signed-off-by: Nick Piggin Cc: David Howells (Needs testing, please). Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index f477acfb732..ff0ab772f49 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -590,6 +590,8 @@ static inline void page_set_cache(struct page *page, struct kmem_cache *cache) static inline struct kmem_cache *page_get_cache(struct page *page) { + if (unlikely(PageCompound(page))) + page = (struct page *)page_private(page); return (struct kmem_cache *)page->lru.next; } @@ -600,6 +602,8 @@ static inline void page_set_slab(struct page *page, struct slab *slab) static inline struct slab *page_get_slab(struct page *page) { + if (unlikely(PageCompound(page))) + page = (struct page *)page_private(page); return (struct slab *)page->lru.prev; } @@ -2412,8 +2416,11 @@ static void set_slab_attr(struct kmem_cache *cachep, struct slab *slabp, struct page *page; /* Nasty!!!!!! I hope this is OK. */ - i = 1 << cachep->gfporder; page = virt_to_page(objp); + + i = 1; + if (likely(!PageCompound(page))) + i <<= cachep->gfporder; do { page_set_cache(page, cachep); page_set_slab(page, slabp); -- cgit From 35386e3b0f876bf194982f48f027af0c216499ce Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Wed, 22 Mar 2006 00:09:05 -0800 Subject: [PATCH] slab: cache_reap(): further reduction in interrupt holdoff cache_reap takes the l3->list_lock (disabling interrupts) unconditionally and then does a few checks and maybe does some cleanup. This patch makes cache_reap() only take the lock if there is work to do and then the lock is taken and released for each cleaning action. The checking of when to do the next reaping is done without any locking and becomes racy. Should not matter since reaping can also be skipped if the slab mutex cannot be acquired. The same is true for the touched processing. If we get this wrong once in awhile then we will mistakenly clean or not clean the shared cache. This will impact performance slightly. Note that the additional drain_array() function introduced here will fall out in a subsequent patch since array cleaning will now be very similar from all callers. Signed-off-by: Christoph Lameter Cc: Pekka Enberg Cc: Manfred Spraul Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 57 +++++++++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 43 insertions(+), 14 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index ff0ab772f49..1845c012739 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -292,13 +292,13 @@ struct kmem_list3 { struct list_head slabs_full; struct list_head slabs_free; unsigned long free_objects; - unsigned long next_reap; - int free_touched; unsigned int free_limit; unsigned int colour_next; /* Per-node cache coloring */ spinlock_t list_lock; struct array_cache *shared; /* shared per node */ struct array_cache **alien; /* on other nodes */ + unsigned long next_reap; /* updated without locking */ + int free_touched; /* updated without locking */ }; /* @@ -3539,6 +3539,22 @@ static void drain_array_locked(struct kmem_cache *cachep, } } + +/* + * Drain an array if it contains any elements taking the l3 lock only if + * necessary. + */ +static void drain_array(struct kmem_cache *searchp, struct kmem_list3 *l3, + struct array_cache *ac) +{ + if (ac && ac->avail) { + spin_lock_irq(&l3->list_lock); + drain_array_locked(searchp, ac, 0, + numa_node_id()); + spin_unlock_irq(&l3->list_lock); + } +} + /** * cache_reap - Reclaim memory from caches. * @unused: unused parameter @@ -3572,33 +3588,48 @@ static void cache_reap(void *unused) searchp = list_entry(walk, struct kmem_cache, next); check_irq_on(); + /* + * We only take the l3 lock if absolutely necessary and we + * have established with reasonable certainty that + * we can do some work if the lock was obtained. + */ l3 = searchp->nodelists[numa_node_id()]; + reap_alien(searchp, l3); - spin_lock_irq(&l3->list_lock); - drain_array_locked(searchp, cpu_cache_get(searchp), 0, - numa_node_id()); + drain_array(searchp, l3, cpu_cache_get(searchp)); + /* + * These are racy checks but it does not matter + * if we skip one check or scan twice. + */ if (time_after(l3->next_reap, jiffies)) - goto next_unlock; + goto next; l3->next_reap = jiffies + REAPTIMEOUT_LIST3; - if (l3->shared) - drain_array_locked(searchp, l3->shared, 0, - numa_node_id()); + drain_array(searchp, l3, l3->shared); if (l3->free_touched) { l3->free_touched = 0; - goto next_unlock; + goto next; } tofree = (l3->free_limit + 5 * searchp->num - 1) / (5 * searchp->num); do { + /* + * Do not lock if there are no free blocks. + */ + if (list_empty(&l3->slabs_free)) + break; + + spin_lock_irq(&l3->list_lock); p = l3->slabs_free.next; - if (p == &(l3->slabs_free)) + if (p == &(l3->slabs_free)) { + spin_unlock_irq(&l3->list_lock); break; + } slabp = list_entry(p, struct slab, list); BUG_ON(slabp->inuse); @@ -3613,10 +3644,8 @@ static void cache_reap(void *unused) l3->free_objects -= searchp->num; spin_unlock_irq(&l3->list_lock); slab_destroy(searchp, slabp); - spin_lock_irq(&l3->list_lock); } while (--tofree > 0); -next_unlock: - spin_unlock_irq(&l3->list_lock); +next: cond_resched(); } check_irq_on(); -- cgit From aab2207cf8d9c343b6b5f0e4d27e1732f8618d14 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Wed, 22 Mar 2006 00:09:06 -0800 Subject: [PATCH] slab: make drain_array more universal by adding more parameters And a parameter to drain_array to control the freeing of all objects and then use drain_array() to replace instances of drain_array_locked with drain_array. Doing so will avoid taking locks in those locations if the arrays are empty. Signed-off-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 1845c012739..d73b38e7d7e 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -2126,6 +2126,10 @@ static void check_spinlock_acquired_node(struct kmem_cache *cachep, int node) static void drain_array_locked(struct kmem_cache *cachep, struct array_cache *ac, int force, int node); +static void drain_array(struct kmem_cache *cachep, struct kmem_list3 *l3, + struct array_cache *ac, + int force, int node); + static void do_drain(void *arg) { struct kmem_cache *cachep = arg; @@ -2150,9 +2154,7 @@ static void drain_cpu_caches(struct kmem_cache *cachep) for_each_online_node(node) { l3 = cachep->nodelists[node]; if (l3) { - spin_lock_irq(&l3->list_lock); - drain_array_locked(cachep, l3->shared, 1, node); - spin_unlock_irq(&l3->list_lock); + drain_array(cachep, l3, l3->shared, 1, node); if (l3->alien) drain_alien_cache(cachep, l3->alien); } @@ -3545,12 +3547,11 @@ static void drain_array_locked(struct kmem_cache *cachep, * necessary. */ static void drain_array(struct kmem_cache *searchp, struct kmem_list3 *l3, - struct array_cache *ac) + struct array_cache *ac, int force, int node) { if (ac && ac->avail) { spin_lock_irq(&l3->list_lock); - drain_array_locked(searchp, ac, 0, - numa_node_id()); + drain_array_locked(searchp, ac, force, node); spin_unlock_irq(&l3->list_lock); } } @@ -3571,6 +3572,7 @@ static void cache_reap(void *unused) { struct list_head *walk; struct kmem_list3 *l3; + int node = numa_node_id(); if (!mutex_trylock(&cache_chain_mutex)) { /* Give up. Setup the next iteration. */ @@ -3593,11 +3595,11 @@ static void cache_reap(void *unused) * have established with reasonable certainty that * we can do some work if the lock was obtained. */ - l3 = searchp->nodelists[numa_node_id()]; + l3 = searchp->nodelists[node]; reap_alien(searchp, l3); - drain_array(searchp, l3, cpu_cache_get(searchp)); + drain_array(searchp, l3, cpu_cache_get(searchp), 0, node); /* * These are racy checks but it does not matter @@ -3608,7 +3610,7 @@ static void cache_reap(void *unused) l3->next_reap = jiffies + REAPTIMEOUT_LIST3; - drain_array(searchp, l3, l3->shared); + drain_array(searchp, l3, l3->shared, 0, node); if (l3->free_touched) { l3->free_touched = 0; -- cgit From 1b55253a7f95adc82eb20937b57b3e3e32ba65df Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Wed, 22 Mar 2006 00:09:07 -0800 Subject: [PATCH] slab: remove drain_array_locked Remove drain_array_locked and use that opportunity to limit the time the l3 lock is taken further. Signed-off-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 31 ++++++++++--------------------- 1 file changed, 10 insertions(+), 21 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index d73b38e7d7e..3274144c0d1 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -2123,9 +2123,6 @@ static void check_spinlock_acquired_node(struct kmem_cache *cachep, int node) #define check_spinlock_acquired_node(x, y) do { } while(0) #endif -static void drain_array_locked(struct kmem_cache *cachep, - struct array_cache *ac, int force, int node); - static void drain_array(struct kmem_cache *cachep, struct kmem_list3 *l3, struct array_cache *ac, int force, int node); @@ -3522,40 +3519,32 @@ static void enable_cpucache(struct kmem_cache *cachep) cachep->name, -err); } -static void drain_array_locked(struct kmem_cache *cachep, - struct array_cache *ac, int force, int node) +/* + * Drain an array if it contains any elements taking the l3 lock only if + * necessary. + */ +void drain_array(struct kmem_cache *cachep, struct kmem_list3 *l3, + struct array_cache *ac, int force, int node) { int tofree; - check_spinlock_acquired_node(cachep, node); + if (!ac || !ac->avail) + return; if (ac->touched && !force) { ac->touched = 0; } else if (ac->avail) { tofree = force ? ac->avail : (ac->limit + 4) / 5; if (tofree > ac->avail) tofree = (ac->avail + 1) / 2; + spin_lock_irq(&l3->list_lock); free_block(cachep, ac->entry, tofree, node); + spin_unlock_irq(&l3->list_lock); ac->avail -= tofree; memmove(ac->entry, &(ac->entry[tofree]), sizeof(void *) * ac->avail); } } - -/* - * Drain an array if it contains any elements taking the l3 lock only if - * necessary. - */ -static void drain_array(struct kmem_cache *searchp, struct kmem_list3 *l3, - struct array_cache *ac, int force, int node) -{ - if (ac && ac->avail) { - spin_lock_irq(&l3->list_lock); - drain_array_locked(searchp, ac, force, node); - spin_unlock_irq(&l3->list_lock); - } -} - /** * cache_reap - Reclaim memory from caches. * @unused: unused parameter -- cgit From b18e7e654d7af741d2bf34a90dc34128d0217fea Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Wed, 22 Mar 2006 00:09:07 -0800 Subject: [PATCH] slab: fix drain_array() so that it works correctly with the shared_array The list_lock also protects the shared array and we call drain_array() with the shared array. Therefore we cannot go as far as I wanted to but have to take the lock in a way so that it also protects the array_cache in drain_pages. (Note: maybe we should make the array_cache locking more consistent? I.e. always take the array cache lock for shared arrays and disable interrupts for the per cpu arrays?) Signed-off-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 3274144c0d1..6b691ecbac4 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -3521,7 +3521,8 @@ static void enable_cpucache(struct kmem_cache *cachep) /* * Drain an array if it contains any elements taking the l3 lock only if - * necessary. + * necessary. Note that the l3 listlock also protects the array_cache + * if drain_array() is used on the shared array. */ void drain_array(struct kmem_cache *cachep, struct kmem_list3 *l3, struct array_cache *ac, int force, int node) @@ -3532,16 +3533,18 @@ void drain_array(struct kmem_cache *cachep, struct kmem_list3 *l3, return; if (ac->touched && !force) { ac->touched = 0; - } else if (ac->avail) { - tofree = force ? ac->avail : (ac->limit + 4) / 5; - if (tofree > ac->avail) - tofree = (ac->avail + 1) / 2; + } else { spin_lock_irq(&l3->list_lock); - free_block(cachep, ac->entry, tofree, node); + if (ac->avail) { + tofree = force ? ac->avail : (ac->limit + 4) / 5; + if (tofree > ac->avail) + tofree = (ac->avail + 1) / 2; + free_block(cachep, ac->entry, tofree, node); + ac->avail -= tofree; + memmove(ac->entry, &(ac->entry[tofree]), + sizeof(void *) * ac->avail); + } spin_unlock_irq(&l3->list_lock); - ac->avail -= tofree; - memmove(ac->entry, &(ac->entry[tofree]), - sizeof(void *) * ac->avail); } } -- cgit From 442295c94bf650221af3ef20fc68fa3e93876818 Mon Sep 17 00:00:00 2001 From: Paul Jackson Date: Wed, 22 Mar 2006 00:09:11 -0800 Subject: [PATCH] mm: slab cache interleave rotor fix The alien cache rotor in mm/slab.c assumes that the first online node is node 0. Eventually for some archs, especially with hotplug, this will no longer be true. Fix the interleave rotor to handle the general case of node numbering. Signed-off-by: Paul Jackson Acked-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slab.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/slab.c') diff --git a/mm/slab.c b/mm/slab.c index 6b691ecbac4..1c8f5ee230d 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -830,7 +830,7 @@ static void init_reap_node(int cpu) node = next_node(cpu_to_node(cpu), node_online_map); if (node == MAX_NUMNODES) - node = 0; + node = first_node(node_online_map); __get_cpu_var(reap_node) = node; } -- cgit