CK1 [01/13] slab: Simplify bootstrap

Index: linux/include/linux/slab_def.h
===================================================================
@@ -91,7 +91,7 @@ struct kmem_cache {
 	 * is statically defined, so we reserve the max number of cpus.
 	 */
 	struct kmem_list3 **nodelists;
-	struct array_cache *array[NR_CPUS];
+	struct array_cache *array[NR_CPUS + MAX_NUMNODES];
 	/*
 	 * Do not add fields after array[]
 	 */

Index: linux/mm/slab.c
===================================================================
@@ -578,9 +578,7 @@ static struct arraycache_init initarray_
     { {0, BOOT_CPUCACHE_ENTRIES, 1, 0} };
 
 /* internal cache of cache description objs */
-static struct kmem_list3 *kmem_cache_nodelists[MAX_NUMNODES];
 static struct kmem_cache kmem_cache_boot = {
-	.nodelists = kmem_cache_nodelists,
 	.batchcount = 1,
 	.limit = BOOT_CPUCACHE_ENTRIES,
 	.shared = 1,

@@ -1584,6 +1582,15 @@ static void __init set_up_list3s(struct
 }
 
 /*
+ * The memory after the last cpu cache pointer is used for the
+ * the nodelists pointer.
+ */
+static void setup_nodelists_pointer(struct kmem_cache *s)
+{
+	s->nodelists = (struct kmem_list3 **)&s->array[nr_cpu_ids];
+}
+
+/*
  * Initialisation.  Called after the page allocator have been initialised and
  * before smp_init().
  */

@@ -1597,13 +1604,15 @@ void __init kmem_cache_init(void)
 	int node;
 
 	kmem_cache = &kmem_cache_boot;
+	setup_nodelists_pointer(kmem_cache);
 
 	if (num_possible_nodes() == 1)
 		use_alien_caches = 0;
 
+
 	for (i = 0; i < NUM_INIT_LISTS; i++) {
 		kmem_list3_init(&initkmem_list3[i]);
-		if (i < MAX_NUMNODES)
+		if (i < nr_node_ids)
 			kmem_cache->nodelists[i] = NULL;
 	}
 	set_up_list3s(kmem_cache, CACHE_CACHE);

@@ -1643,7 +1652,6 @@ void __init kmem_cache_init(void)
 	list_add(&kmem_cache->list, &slab_caches);
 	kmem_cache->colour_off = cache_line_size();
 	kmem_cache->array[smp_processor_id()] = &initarray_cache.cache;
-	kmem_cache->nodelists[node] = &initkmem_list3[CACHE_CACHE + node];
 
 	/*
 	 * struct kmem_cache size depends on nr_node_ids & nr_cpu_ids

@@ -2454,7 +2462,7 @@ __kmem_cache_create (struct kmem_cache *
 	else
 		gfp = GFP_NOWAIT;
 
-	cachep->nodelists = (struct kmem_list3 **)&cachep->array[nr_cpu_ids];
+	setup_nodelists_pointer(cachep);
 #if DEBUG
 
 	/*

Index: linux/mm/slab.c
===================================================================
@@ -1686,23 +1686,13 @@ void __init kmem_cache_init(void)
 	 * bug.
 	 */
 
-	sizes[INDEX_AC].cs_cachep = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);
-	sizes[INDEX_AC].cs_cachep->name = names[INDEX_AC].name;
-	sizes[INDEX_AC].cs_cachep->size = sizes[INDEX_AC].cs_size;
-	sizes[INDEX_AC].cs_cachep->object_size = sizes[INDEX_AC].cs_size;
-	sizes[INDEX_AC].cs_cachep->align = ARCH_KMALLOC_MINALIGN;
-	__kmem_cache_create(sizes[INDEX_AC].cs_cachep, ARCH_KMALLOC_FLAGS|SLAB_PANIC);
-	list_add(&sizes[INDEX_AC].cs_cachep->list, &slab_caches);
+	sizes[INDEX_AC].cs_cachep = create_kmalloc_cache(names[INDEX_AC].name,
+					sizes[INDEX_AC].cs_size, ARCH_KMALLOC_FLAGS);
 
-	if (INDEX_AC != INDEX_L3) {
-		sizes[INDEX_L3].cs_cachep = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);
-		sizes[INDEX_L3].cs_cachep->name = names[INDEX_L3].name;
-		sizes[INDEX_L3].cs_cachep->size = sizes[INDEX_L3].cs_size;
-		sizes[INDEX_L3].cs_cachep->object_size = sizes[INDEX_L3].cs_size;
-		sizes[INDEX_L3].cs_cachep->align = ARCH_KMALLOC_MINALIGN;
-		__kmem_cache_create(sizes[INDEX_L3].cs_cachep, ARCH_KMALLOC_FLAGS|SLAB_PANIC);
-		list_add(&sizes[INDEX_L3].cs_cachep->list, &slab_caches);
-	}
+	if (INDEX_AC != INDEX_L3)
+		sizes[INDEX_L3].cs_cachep =
+			create_kmalloc_cache(names[INDEX_L3].name,
+				sizes[INDEX_L3].cs_size, ARCH_KMALLOC_FLAGS);
 
 	slab_early_init = 0;
 

@@ -1714,24 +1704,14 @@ 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) {
-			sizes->cs_cachep = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);
-			sizes->cs_cachep->name = names->name;
-			sizes->cs_cachep->size = sizes->cs_size;
-			sizes->cs_cachep->object_size = sizes->cs_size;
-			sizes->cs_cachep->align = ARCH_KMALLOC_MINALIGN;
-			__kmem_cache_create(sizes->cs_cachep, ARCH_KMALLOC_FLAGS|SLAB_PANIC);
-			list_add(&sizes->cs_cachep->list, &slab_caches);
-		}
+		if (!sizes->cs_cachep)
+			sizes->cs_cachep = create_kmalloc_cache(names->name,
+					sizes->cs_size, ARCH_KMALLOC_FLAGS);
+
 #ifdef CONFIG_ZONE_DMA
-		sizes->cs_dmacachep = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);
-		sizes->cs_dmacachep->name = names->name_dma;
-		sizes->cs_dmacachep->size = sizes->cs_size;
-		sizes->cs_dmacachep->object_size = sizes->cs_size;
-		sizes->cs_dmacachep->align = ARCH_KMALLOC_MINALIGN;
-		__kmem_cache_create(sizes->cs_dmacachep,
-			       ARCH_KMALLOC_FLAGS|SLAB_CACHE_DMA| SLAB_PANIC);
-		list_add(&sizes->cs_dmacachep->list, &slab_caches);
+		sizes->cs_dmacachep = create_kmalloc_cache(
+			names->name_dma, sizes->cs_size,
+			SLAB_CACHE_DMA|ARCH_KMALLOC_FLAGS);
 #endif
 		sizes++;
 		names++;

Index: linux/mm/slab.h
===================================================================
@@ -35,6 +35,11 @@ extern struct kmem_cache *kmem_cache;
 /* Functions provided by the slab allocators */
 extern int __kmem_cache_create(struct kmem_cache *, unsigned long flags);
 
+extern struct kmem_cache *create_kmalloc_cache(const char *name, size_t size,
+			unsigned long flags);
+extern void create_boot_cache(struct kmem_cache *, const char *name,
+			size_t size, unsigned long flags);
+
 #ifdef CONFIG_SLUB
 struct kmem_cache *__kmem_cache_alias(const char *name, size_t size,
 	size_t align, unsigned long flags, void (*ctor)(void *));

Index: linux/mm/slab_common.c
===================================================================
@@ -189,3 +189,35 @@ int slab_is_available(void)
 {
 	return slab_state >= UP;
 }
+
+/* Create a cache during boot when no slab services are available yet */
+void __init create_boot_cache(struct kmem_cache *s, const char *name, size_t size,
+		unsigned long flags)
+{
+	int err;
+
+	s->name = name;
+	s->size = s->object_size = size;
+	s->align = ARCH_KMALLOC_MINALIGN;
+	err = __kmem_cache_create(s, flags);
+
+	if (err)
+		panic("Creation of kmalloc slab %s size=%ld failed. Reason %d\n",
+					name, size, err);
+
+	list_add(&s->list, &slab_caches);
+	s->refcount = -1;	/* Exempt from merging for now */
+}
+
+struct kmem_cache *__init create_kmalloc_cache(const char *name, size_t size,
+				unsigned long flags)
+{
+	struct kmem_cache *s = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);
+
+	if (!s)
+		panic("Out of memory when creating slab %s\n", name);
+
+	create_boot_cache(s, name, size, flags);
+	s->refcount = 1;
+	return s;
+}

Index: linux/mm/slub.c
===================================================================
@@ -3253,32 +3253,6 @@ static int __init setup_slub_nomerge(cha
 
 __setup("slub_nomerge", setup_slub_nomerge);
 
-static struct kmem_cache *__init create_kmalloc_cache(const char *name,
-						int size, unsigned int flags)
-{
-	struct kmem_cache *s;
-
-	s = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);
-
-	s->name = name;
-	s->size = s->object_size = size;
-	s->align = ARCH_KMALLOC_MINALIGN;
-
-	/*
-	 * This function is called with IRQs disabled during early-boot on
-	 * single CPU so there's no need to take slab_mutex here.
-	 */
-	if (kmem_cache_open(s, flags))
-		goto panic;
-
-	list_add(&s->list, &slab_caches);
-	return s;
-
-panic:
-	panic("Creation of kmalloc slab %s size=%d failed.\n", name, size);
-	return NULL;
-}
-
 /*
  * Conversion table for small slabs sizes / 8 to the index in the
  * kmalloc array. This is necessary for slabs < 192 since we have non power

Index: linux/include/linux/slab_def.h
===================================================================
@@ -88,16 +88,13 @@ struct kmem_cache {
 	 * We still use [NR_CPUS] and not [1] or [0] because cache_cache
 	 * is statically defined, so we reserve the max number of cpus.
 	 */
-	struct kmem_cache_node **nodelists;
+	struct kmem_cache_node **node;
 	struct array_cache *array[NR_CPUS + MAX_NUMNODES];
 	/*
 	 * Do not add fields after array[]
 	 */
 };
 
-extern struct kmem_cache *cs_cachep[PAGE_SHIFT + MAX_ORDER];
-extern struct kmem_cache *cs_dmacachep[PAGE_SHIFT + MAX_ORDER];
-
 void *kmem_cache_alloc(struct kmem_cache *, gfp_t);
 void *__kmalloc(size_t size, gfp_t flags);
 

@@ -132,10 +129,10 @@ static __always_inline void *kmalloc(siz
 
 #ifdef CONFIG_ZONE_DMA
 		if (flags & GFP_DMA)
-			cachep = cs_dmacachep[i];
+			cachep = kmalloc_dma_caches[i];
 		else
 #endif
-			cachep = cs_cachep[i];
+			cachep = kmalloc_caches[i];
 
 		ret = kmem_cache_alloc_trace(size, cachep, flags);
 

@@ -178,10 +175,10 @@ static __always_inline void *kmalloc_nod
 
 #ifdef CONFIG_ZONE_DMA
 		if (flags & GFP_DMA)
-			cachep = cs_dmacachep[i];
+			cachep = kmalloc_dma_caches[i];
 		else
 #endif
-			cachep = cs_cachep[i];
+			cachep = kmalloc_caches[i];
 
 		return kmem_cache_alloc_node_trace(size, cachep, flags, node);
 	}

574 lines not on this page · raw message

Index: linux/include/linux/slab_def.h
===================================================================
@@ -88,7 +88,7 @@ struct kmem_cache {
 	 * We still use [NR_CPUS] and not [1] or [0] because cache_cache
 	 * is statically defined, so we reserve the max number of cpus.
 	 */
-	struct kmem_list3 **nodelists;
+	struct kmem_cache_node **nodelists;
 	struct array_cache *array[NR_CPUS + MAX_NUMNODES];
 	/*
 	 * Do not add fields after array[]

303 lines not on this page · raw message

Index: linux/mm/slab.c
===================================================================
@@ -572,8 +572,6 @@ static struct cache_names __initdata cac
 #undef CACHE
 };
 
-static struct arraycache_init initarray_cache __initdata =
-    { {0, BOOT_CPUCACHE_ENTRIES, 1, 0} };
 static struct arraycache_init initarray_generic =
     { {0, BOOT_CPUCACHE_ENTRIES, 1, 0} };
 

@@ -1596,12 +1594,9 @@ static void setup_nodelists_pointer(stru
  */
 void __init kmem_cache_init(void)
 {
-	size_t left_over;
 	struct cache_sizes *sizes;
 	struct cache_names *names;
 	int i;
-	int order;
-	int node;
 
 	kmem_cache = &kmem_cache_boot;
 	setup_nodelists_pointer(kmem_cache);

@@ -1645,36 +1640,17 @@ void __init kmem_cache_init(void)
 	 * 6) Resize the head arrays of the kmalloc caches to their final sizes.
 	 */
 
-	node = numa_mem_id();
-
 	/* 1) create the kmem_cache */
-	INIT_LIST_HEAD(&slab_caches);
-	list_add(&kmem_cache->list, &slab_caches);
-	kmem_cache->colour_off = cache_line_size();
-	kmem_cache->array[smp_processor_id()] = &initarray_cache.cache;
 
 	/*
 	 * struct kmem_cache size depends on nr_node_ids & nr_cpu_ids
 	 */
-	kmem_cache->size = offsetof(struct kmem_cache, array[nr_cpu_ids]) +
-				  nr_node_ids * sizeof(struct kmem_list3 *);
-	kmem_cache->object_size = kmem_cache->size;
-	kmem_cache->size = ALIGN(kmem_cache->object_size,
-					cache_line_size());
-	kmem_cache->reciprocal_buffer_size =
-		reciprocal_value(kmem_cache->size);
-
-	for (order = 0; order < MAX_ORDER; order++) {
-		cache_estimate(order, kmem_cache->size,
-			cache_line_size(), 0, &left_over, &kmem_cache->num);
-		if (kmem_cache->num)
-			break;
-	}
-	BUG_ON(!kmem_cache->num);
-	kmem_cache->gfporder = order;
-	kmem_cache->colour = left_over / kmem_cache->colour_off;
-	kmem_cache->slab_size = ALIGN(kmem_cache->num * sizeof(kmem_bufctl_t) +
-				      sizeof(struct slab), cache_line_size());
+	create_boot_cache(kmem_cache, "kmem_cache",
+		offsetof(struct kmem_cache, array[nr_cpu_ids]) +
+				  nr_node_ids * sizeof(struct kmem_list3 *),
+				  SLAB_HWCACHE_ALIGN);
+
+	slab_state = PARTIAL;
 
 	/* 2+3) create the kmalloc caches */
 	sizes = malloc_sizes;

@@ -1722,7 +1698,6 @@ void __init kmem_cache_init(void)
 
 		ptr = kmalloc(sizeof(struct arraycache_init), GFP_NOWAIT);
 
-		BUG_ON(cpu_cache_get(kmem_cache) != &initarray_cache.cache);
 		memcpy(ptr, cpu_cache_get(kmem_cache),
 		       sizeof(struct arraycache_init));
 		/*

@@ -2277,7 +2252,16 @@ static int __init_refok setup_cpu_cache(
 
 	if (slab_state == DOWN) {
 		/*
-		 * Note: the first kmem_cache_create must create the cache
+		 * Note: Creation of first cache (kmem_cache).
+		 * The setup_list3s is taken care
+		 * of by the caller of __kmem_cache_create
+		 */
+		cachep->array[smp_processor_id()] = &initarray_generic.cache;
+		slab_state = PARTIAL;
+	} else
+	if (slab_state == PARTIAL) {
+		/*
+		 * Note: the second kmem_cache_create must create the cache
 		 * that's used by kmalloc(24), otherwise the creation of
 		 * further caches will BUG().
 		 */

@@ -2285,7 +2269,7 @@ static int __init_refok setup_cpu_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,
+		 * the second cache, then we need to set up all its list3s,
 		 * otherwise the creation of further caches will BUG().
 		 */
 		set_up_list3s(cachep, SIZE_AC);

@@ -2294,6 +2278,7 @@ static int __init_refok setup_cpu_cache(
 		else
 			slab_state = PARTIAL_ARRAYCACHE;
 	} else {
+		/* Remaining boot caches */
 		cachep->array[smp_processor_id()] =
 			kmalloc(sizeof(struct arraycache_init), gfp);
 

@@ -2389,22 +2374,6 @@ __kmem_cache_create (struct kmem_cache *
 		size &= ~(BYTES_PER_WORD - 1);
 	}
 
-	/* 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.
-		 */
-		ralign = cache_line_size();
-		while (size <= ralign / 2)
-			ralign /= 2;
-	} else {
-		ralign = BYTES_PER_WORD;
-	}
-
 	/*
 	 * Redzoning and user store require word alignment or possibly larger.
 	 * Note this will be overridden by architecture or caller mandated

@@ -2421,10 +2390,6 @@ __kmem_cache_create (struct kmem_cache *
 		size &= ~(REDZONE_ALIGN - 1);
 	}
 
-	/* 2) arch mandated alignment */
-	if (ralign < ARCH_SLAB_MINALIGN) {
-		ralign = ARCH_SLAB_MINALIGN;
-	}
 	/* 3) caller mandated alignment */
 	if (ralign < cachep->align) {
 		ralign = cachep->align;

Index: linux/include/linux/slab.h
===================================================================
@@ -143,21 +143,6 @@ unsigned int kmem_cache_size(struct kmem
 		(__flags), NULL)
 
 /*
- * The largest kmalloc size supported by the slab allocators is
- * 32 megabyte (2^25) or the maximum allocatable page order if that is
- * less than 32 MB.
- *
- * WARNING: Its not easy to increase this value since the allocators have
- * to do various tricks to work around compiler limitations in order to
- * ensure proper constant folding.
- */
-#define KMALLOC_SHIFT_HIGH	((MAX_ORDER + PAGE_SHIFT - 1) <= 25 ? \
-				(MAX_ORDER + PAGE_SHIFT - 1) : 25)
-
-#define KMALLOC_MAX_SIZE	(1UL << KMALLOC_SHIFT_HIGH)
-#define KMALLOC_MAX_ORDER	(KMALLOC_SHIFT_HIGH - PAGE_SHIFT)
-
-/*
  * Some archs want to perform DMA into kmalloc caches and need a guaranteed
  * alignment larger than the alignment of a 64-bit integer.
  * Setting ARCH_KMALLOC_MINALIGN in arch headers allows that.

@@ -178,8 +163,24 @@ unsigned int kmem_cache_size(struct kmem
 #endif
 
 /*
+ * The largest kmalloc size supported by the slab allocators is
+ * 32 megabyte (2^25) or the maximum allocatable page order if that is
+ * less than 32 MB.
+ *
+ * WARNING: Its not easy to increase this value since the allocators have
+ * to do various tricks to work around compiler limitations in order to
+ * ensure proper constant folding.
+ */
+#define KMALLOC_SHIFT_HIGH	((MAX_ORDER + PAGE_SHIFT - 1) <= 25 ? \
+				(MAX_ORDER + PAGE_SHIFT - 1) : 25)
+
+#define KMALLOC_MAX_SIZE	(1UL << KMALLOC_SHIFT_HIGH)
+#define KMALLOC_MAX_ORDER	(KMALLOC_SHIFT_HIGH - PAGE_SHIFT)
+
+/*
  * Kmalloc subsystem.
  */
+#ifndef CONFIG_SLOB
 #if defined(ARCH_DMA_MINALIGN) && ARCH_DMA_MINALIGN > 8
 #define KMALLOC_MIN_SIZE ARCH_DMA_MINALIGN
 #else

@@ -260,6 +261,7 @@ static __always_inline int kmalloc_size(
 
 	return 0;
 }
+#endif
 
 /*
  * Common kmalloc functions provided by all allocators

Index: linux/mm/slab_common.c
===================================================================
@@ -218,6 +218,8 @@ int slab_is_available(void)
 	return slab_state >= UP;
 }
 
+#ifndef CONFIG_SLOB
+
 /* Create a cache during boot when no slab services are available yet */
 void __init create_boot_cache(struct kmem_cache *s, const char *name, size_t size,
 		unsigned long flags)

@@ -249,3 +251,5 @@ struct kmem_cache *__init create_kmalloc
 	s->refcount = 1;
 	return s;
 }
+
+#endif /* !CONFIG_SLOB */

Index: linux/mm/slab.c
===================================================================
@@ -291,6 +291,11 @@ static inline void clear_obj_pfmemalloc(
 	*objp = (void *)((unsigned long)*objp & ~SLAB_OBJ_PFMEMALLOC);
 }
 
+struct kmem_cache *cs_cachep[PAGE_SHIFT + MAX_ORDER];
+EXPORT_SYMBOL(cs_cachep);
+struct kmem_cache *cs_dmacachep[PAGE_SHIFT + MAX_ORDER];
+EXPORT_SYMBOL(cs_dmacachep);
+
 /*
  * bootstrap: The caches do not work without cpuarrays anymore, but the
  * cpuarrays are allocated from the generic caches...

@@ -334,34 +339,10 @@ static void free_block(struct kmem_cache
 static int enable_cpucache(struct kmem_cache *cachep, gfp_t gfp);
 static void cache_reap(struct work_struct *unused);
 
-/*
- * 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)
-{
-	extern void __bad_size(void);
-
-	if (__builtin_constant_p(size)) {
-		int i = 0;
-
-#define CACHE(x) \
-	if (size <=x) \
-		return i; \
-	else \
-		i++;
-#include <linux/kmalloc_sizes.h>
-#undef CACHE
-		__bad_size();
-	} else
-		__bad_size();
-	return 0;
-}
-
 static int slab_early_init = 1;
 
-#define INDEX_AC index_of(sizeof(struct arraycache_init))
-#define INDEX_L3 index_of(sizeof(struct kmem_list3))
+#define INDEX_AC kmalloc_index(sizeof(struct arraycache_init))
+#define INDEX_L3 kmalloc_index(sizeof(struct kmem_list3))
 
 static void kmem_list3_init(struct kmem_list3 *parent)
 {

@@ -548,30 +529,6 @@ static inline unsigned int obj_to_index(
 	return reciprocal_divide(offset, cache->reciprocal_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) },
-#include <linux/kmalloc_sizes.h>
-	CACHE(ULONG_MAX)
-#undef CACHE
-};
-EXPORT_SYMBOL(malloc_sizes);
-
-/* Must match cache_sizes above. Out of line to keep cache footprint low. */
-struct cache_names {
-	char *name;
-	char *name_dma;
-};
-
-static struct cache_names __initdata cache_names[] = {
-#define CACHE(x) { .name = "size-" #x, .name_dma = "size-" #x "(DMA)" },
-#include <linux/kmalloc_sizes.h>
-	{NULL,}
-#undef CACHE
-};
-
 static struct arraycache_init initarray_generic =
     { {0, BOOT_CPUCACHE_ENTRIES, 1, 0} };
 

@@ -649,19 +606,23 @@ static void slab_set_debugobj_lock_class
 
 static void init_node_lock_keys(int q)
 {
-	struct cache_sizes *s = malloc_sizes;
+	int i;
 
 	if (slab_state < UP)
 		return;
 
-	for (s = malloc_sizes; s->cs_size != ULONG_MAX; s++) {
+	for (i = 1; i < PAGE_SHIFT + MAX_ORDER; i++) {
 		struct kmem_list3 *l3;
+		struct kmem_cache *cache = cs_cachep[i];
 
-		l3 = s->cs_cachep->nodelists[q];
-		if (!l3 || OFF_SLAB(s->cs_cachep))
+		if (!cache)
 			continue;
 
-		slab_set_lock_classes(s->cs_cachep, &on_slab_l3_key,
+		l3 = cache->nodelists[q];
+		if (!l3 || OFF_SLAB(cache))
+			continue;
+
+		slab_set_lock_classes(cache, &on_slab_l3_key,
 				&on_slab_alc_key, q);
 	}
 }

@@ -701,20 +662,19 @@ static inline struct array_cache *cpu_ca
 static inline struct kmem_cache *__find_general_cachep(size_t size,
 							gfp_t gfpflags)
 {
-	struct cache_sizes *csizep = malloc_sizes;
+	int i;
 
 #if DEBUG
 	/* This happens if someone tries to call
 	 * kmem_cache_create(), or __kmalloc(), before
 	 * the generic caches are initialized.
 	 */
-	BUG_ON(malloc_sizes[INDEX_AC].cs_cachep == NULL);
+	BUG_ON(cs_cachep[INDEX_AC] == NULL);
 #endif
 	if (!size)
 		return ZERO_SIZE_PTR;
 
-	while (size > csizep->cs_size)
-		csizep++;
+	i = kmalloc_index(size);
 
 	/*
 	 * Really subtle: The last entry with cs->cs_size==ULONG_MAX

@@ -723,9 +683,9 @@ static inline struct kmem_cache *__find_
 	 */
 #ifdef CONFIG_ZONE_DMA
 	if (unlikely(gfpflags & GFP_DMA))
-		return csizep->cs_dmacachep;
+		return cs_dmacachep[i];
 #endif
-	return csizep->cs_cachep;
+	return cs_cachep[i];
 }
 
 static struct kmem_cache *kmem_find_general_cachep(size_t size, gfp_t gfpflags)

@@ -1594,8 +1554,6 @@ static void setup_nodelists_pointer(stru
  */
 void __init kmem_cache_init(void)
 {
-	struct cache_sizes *sizes;
-	struct cache_names *names;
 	int i;
 
 	kmem_cache = &kmem_cache_boot;

@@ -1653,8 +1611,6 @@ void __init kmem_cache_init(void)
 	slab_state = PARTIAL;
 
 	/* 2+3) create the kmalloc caches */
-	sizes = malloc_sizes;
-	names = cache_names;
 
 	/*
 	 * Initialize the caches that provide memory for the array cache and the

@@ -1662,35 +1618,39 @@ void __init kmem_cache_init(void)
 	 * bug.
 	 */
 
-	sizes[INDEX_AC].cs_cachep = create_kmalloc_cache(names[INDEX_AC].name,
-					sizes[INDEX_AC].cs_size, ARCH_KMALLOC_FLAGS);
+	cs_cachep[INDEX_AC] = create_kmalloc_cache("kmalloc-ac",
+					kmalloc_size(INDEX_AC), ARCH_KMALLOC_FLAGS);
 
 	if (INDEX_AC != INDEX_L3)
-		sizes[INDEX_L3].cs_cachep =
-			create_kmalloc_cache(names[INDEX_L3].name,
-				sizes[INDEX_L3].cs_size, ARCH_KMALLOC_FLAGS);
+		cs_cachep[INDEX_L3] =
+			create_kmalloc_cache("kmalloc-l3",
+				kmalloc_size(INDEX_L3), ARCH_KMALLOC_FLAGS);
 
 	slab_early_init = 0;
 
-	while (sizes->cs_size != ULONG_MAX) {
-		/*
-		 * For performance, all the general caches are L1 aligned.
-		 * This should be particularly beneficial on SMP boxes, as it
-		 * eliminates "false sharing".
-		 * Note for systems short on memory removing the alignment will
-		 * allow tighter packing of the smaller caches.
-		 */
-		if (!sizes->cs_cachep)
-			sizes->cs_cachep = create_kmalloc_cache(names->name,
-					sizes->cs_size, ARCH_KMALLOC_FLAGS);
+	for (i = 1; i < PAGE_SHIFT + MAX_ORDER; i++) {
+		size_t cs_size = kmalloc_size(i);
+
+		if (cs_size < KMALLOC_MIN_SIZE)
+			continue;
+
+		if (!cs_cachep[i]) {
+			/*
+			 * For performance, all the general caches are L1 aligned.
+			 * This should be particularly beneficial on SMP boxes, as it
+			 * eliminates "false sharing".
+			 * Note for systems short on memory removing the alignment will
+			 * allow tighter packing of the smaller caches.
+			 */
+			cs_cachep[i] = create_kmalloc_cache("kmalloc",
+					cs_size, ARCH_KMALLOC_FLAGS);
+		}
 
 #ifdef CONFIG_ZONE_DMA
-		sizes->cs_dmacachep = create_kmalloc_cache(
-			names->name_dma, sizes->cs_size,
+		cs_dmacachep[i] = create_kmalloc_cache(
+			"kmalloc-dma", cs_size,
 			SLAB_CACHE_DMA|ARCH_KMALLOC_FLAGS);
 #endif
-		sizes++;
-		names++;
 	}
 	/* 4) Replace the bootstrap head arrays */
 	{

@@ -1709,17 +1669,16 @@ void __init kmem_cache_init(void)
 
 		ptr = kmalloc(sizeof(struct arraycache_init), GFP_NOWAIT);
 
-		BUG_ON(cpu_cache_get(malloc_sizes[INDEX_AC].cs_cachep)
+		BUG_ON(cpu_cache_get(cs_cachep[INDEX_AC])
 		       != &initarray_generic.cache);
-		memcpy(ptr, cpu_cache_get(malloc_sizes[INDEX_AC].cs_cachep),
+		memcpy(ptr, cpu_cache_get(cs_cachep[INDEX_AC]),
 		       sizeof(struct arraycache_init));
 		/*
 		 * Do not assume that spinlocks can be initialized via memcpy:
 		 */
 		spin_lock_init(&ptr->lock);
 
-		malloc_sizes[INDEX_AC].cs_cachep->array[smp_processor_id()] =
-		    ptr;
+		cs_cachep[INDEX_AC]->array[smp_processor_id()] = ptr;
 	}
 	/* 5) Replace the bootstrap kmem_list3's */
 	{

@@ -1728,17 +1687,39 @@ void __init kmem_cache_init(void)
 		for_each_online_node(nid) {
 			init_list(kmem_cache, &initkmem_list3[CACHE_CACHE + nid], nid);
 
-			init_list(malloc_sizes[INDEX_AC].cs_cachep,
+			init_list(cs_cachep[INDEX_AC],
 				  &initkmem_list3[SIZE_AC + nid], nid);
 
 			if (INDEX_AC != INDEX_L3) {
-				init_list(malloc_sizes[INDEX_L3].cs_cachep,
+				init_list(cs_cachep[INDEX_L3],
 					  &initkmem_list3[SIZE_L3 + nid], nid);
 			}
 		}
 	}
 
 	slab_state = UP;
+
+	/* Create the proper names */
+	for (i = 1; i < PAGE_SHIFT + MAX_ORDER; i++) {
+		char *s;
+		struct kmem_cache *c = cs_cachep[i];
+
+		if (!c)
+			continue;
+
+		s = kasprintf(GFP_NOWAIT, "kmalloc-%d", kmalloc_size(i));
+
+		BUG_ON(!s);
+		c->name = s;
+		
+#ifdef CONFIG_ZONE_DMA
+		c = cs_dmacachep[i];
+		BUG_ON(!c);
+		s = kasprintf(GFP_NOWAIT, "dma-kmalloc-%d", kmalloc_size(i));
+		BUG_ON(!s);
+		c->name = s;
+#endif
+	}
 }
 
 void __init kmem_cache_init_late(void)

45 lines not on this page · raw message

87 lines not on this page · raw message

Index: linux/include/linux/slab.h
===================================================================
@@ -203,6 +203,11 @@ unsigned int kmem_cache_size(struct kmem
 #define KMALLOC_MIN_SIZE (1 << KMALLOC_SHIFT_LOW)
 #endif
 
+extern struct kmem_cache *kmalloc_caches[KMALLOC_SHIFT_HIGH + 1];
+#ifdef CONFIG_ZONE_DMA
+extern struct kmem_cache *kmalloc_dma_caches[KMALLOC_SHIFT_HIGH + 1];
+#endif
+
 /*
  * Figure out which kmalloc slab an allocation of a certain size
  * belongs to.

Index: linux/mm/slab_common.c
===================================================================
@@ -252,4 +252,12 @@ struct kmem_cache *__init create_kmalloc
 	return s;
 }
 
+struct kmem_cache *kmalloc_caches[KMALLOC_SHIFT_HIGH + 1];
+EXPORT_SYMBOL(kmalloc_caches);
+
+#ifdef CONFIG_ZONE_DMA
+struct kmem_cache *kmalloc_dma_caches[KMALLOC_SHIFT_HIGH + 1];
+EXPORT_SYMBOL(kmalloc_dma_caches);
+#endif
+
 #endif /* !CONFIG_SLOB */

Index: linux/include/linux/slub_def.h
===================================================================
@@ -111,6 +111,9 @@ struct kmem_cache {
 	struct kmem_cache_node *node[MAX_NUMNODES];
 };
 
+#define KMEM_CACHE_SIZE (offsetof(struct kmem_cache, node) + \
+			nr_node_ids * sizeof(struct kmem_cache_node *))
+
 #ifdef CONFIG_ZONE_DMA
 #define SLUB_DMA __GFP_DMA
 #else

@@ -119,12 +122,6 @@ struct kmem_cache {
 #endif
 
 /*
- * We keep the general caches in an array of slab caches that are used for
- * 2^x bytes of allocations.
- */
-extern struct kmem_cache *kmalloc_caches[KMALLOC_SHIFT_HIGH + 1];
-
-/*
  * Find the slab cache for a given combination of allocation flags and size.
  *
  * This ought to end up with a global pointer to the right cache

Index: linux/mm/slub.c
===================================================================
@@ -3174,13 +3174,6 @@ int __kmem_cache_shutdown(struct kmem_ca
  *		Kmalloc subsystem
  *******************************************************************/
 
-struct kmem_cache *kmalloc_caches[KMALLOC_SHIFT_HIGH + 1];
-EXPORT_SYMBOL(kmalloc_caches);
-
-#ifdef CONFIG_ZONE_DMA
-static struct kmem_cache *kmalloc_dma_caches[KMALLOC_SHIFT_HIGH + 1];
-#endif
-
 static int __init setup_slub_min_order(char *str)
 {
 	get_option(&str, &slub_min_order);

Index: linux/mm/slub.c
===================================================================
@@ -3649,9 +3649,6 @@ static void __init kmem_cache_bootstrap_
 {
 	int node;
 
-	list_add(&s->list, &slab_caches);
-	s->refcount = -1;
-
 	for_each_node_state(node, N_NORMAL_MEMORY) {
 		struct kmem_cache_node *n = get_node(s, node);
 		struct page *p;

@@ -3668,14 +3665,13 @@ static void __init kmem_cache_bootstrap_
 	}
 }
 
+static __initdata struct kmem_cache boot_kmem_cache,
+			boot_kmem_cache_node;
+
 void __init kmem_cache_init(void)
 {
 	int i;
-	int caches = 0;
-	struct kmem_cache *temp_kmem_cache;
-	int order;
-	struct kmem_cache *temp_kmem_cache_node;
-	unsigned long kmalloc_size;
+	int caches = 2;
 
 	if (debug_guardpage_minorder())
 		slub_max_order = 0;

@@ -3683,53 +3679,32 @@ void __init kmem_cache_init(void)
 	kmem_size = offsetof(struct kmem_cache, node) +
 			nr_node_ids * sizeof(struct kmem_cache_node *);
 
-	/* Allocate two kmem_caches from the page allocator */
-	kmalloc_size = ALIGN(kmem_size, cache_line_size());
-	order = get_order(2 * kmalloc_size);
-	kmem_cache = (void *)__get_free_pages(GFP_NOWAIT | __GFP_ZERO, order);
-
-	/*
-	 * Must first have the slab cache available for the allocations of the
-	 * struct kmem_cache_node's. There is special bootstrap code in
-	 * kmem_cache_open for slab_state == DOWN.
-	 */
-	kmem_cache_node = (void *)kmem_cache + kmalloc_size;
+	kmem_cache_node = &boot_kmem_cache_node;
 
-	kmem_cache_node->name = "kmem_cache_node";
-	kmem_cache_node->size = kmem_cache_node->object_size =
-		sizeof(struct kmem_cache_node);
-	kmem_cache_open(kmem_cache_node, SLAB_HWCACHE_ALIGN | SLAB_PANIC);
+	create_boot_cache(kmem_cache_node, "kmem_cache_node",
+		sizeof(struct kmem_cache_node), SLAB_HWCACHE_ALIGN);
 
 	hotplug_memory_notifier(slab_memory_callback, SLAB_CALLBACK_PRI);
 
 	/* Able to allocate the per node structures */
 	slab_state = PARTIAL;
 
-	temp_kmem_cache = kmem_cache;
-	kmem_cache->name = "kmem_cache";
-	kmem_cache->size = kmem_cache->object_size = kmem_size;
-	kmem_cache_open(kmem_cache, SLAB_HWCACHE_ALIGN | SLAB_PANIC);
+	create_boot_cache(&boot_kmem_cache, "kmem_cache", kmem_size,
+		       SLAB_HWCACHE_ALIGN);
 
-	kmem_cache = kmem_cache_alloc(kmem_cache, GFP_NOWAIT);
-	memcpy(kmem_cache, temp_kmem_cache, kmem_size);
+	kmem_cache = kmem_cache_alloc(&boot_kmem_cache, GFP_NOWAIT);
+	memcpy(kmem_cache, &boot_kmem_cache, kmem_size);
 
 	/*
 	 * Allocate kmem_cache_node properly from the kmem_cache slab.
 	 * kmem_cache_node is separately allocated so no need to
 	 * update any list pointers.
 	 */
-	temp_kmem_cache_node = kmem_cache_node;
-
 	kmem_cache_node = kmem_cache_alloc(kmem_cache, GFP_NOWAIT);
-	memcpy(kmem_cache_node, temp_kmem_cache_node, kmem_size);
+	memcpy(kmem_cache_node, &boot_kmem_cache_node, kmem_size);
 
 	kmem_cache_bootstrap_fixup(kmem_cache_node);
-
-	caches++;
 	kmem_cache_bootstrap_fixup(kmem_cache);
-	caches++;
-	/* Free temporary boot structure */
-	free_pages((unsigned long)temp_kmem_cache, order);
 
 	/* Now we can use the kmem_cache to allocate kmalloc slabs */
 

@@ -3930,6 +3905,10 @@ int __kmem_cache_create(struct kmem_cach
 	if (err)
 		return err;
 
+	/* Mutex is not taken during early boot */
+	if (slab_state <= UP)
+		return 0;
+
 	mutex_unlock(&slab_mutex);
 	err = sysfs_slab_add(s);
 	mutex_lock(&slab_mutex);

Index: linux/mm/slab.h
===================================================================
@@ -32,6 +32,9 @@ extern struct list_head slab_caches;
 /* The slab cache that manages slab cache information */
 extern struct kmem_cache *kmem_cache;
 
+unsigned long calculate_alignment(unsigned long flags,
+		unsigned long align, unsigned long size);
+
 /* Functions provided by the slab allocators */
 extern int __kmem_cache_create(struct kmem_cache *, unsigned long flags);
 

Index: linux/mm/slab_common.c
===================================================================
@@ -71,6 +71,34 @@ static inline int kmem_cache_sanity_chec
 #endif
 
 /*
+ * Figure out what the alignment of the objects will be given a set of
+ * flags, a user specified alignment and the size of the objects.
+ */
+unsigned long calculate_alignment(unsigned long flags,
+		unsigned long align, unsigned long size)
+{
+	/*
+	 * If the user wants hardware cache aligned objects then follow that
+	 * suggestion if the object is sufficiently large.
+	 *
+	 * The hardware cache alignment cannot override the specified
+	 * alignment though. If that is greater then use it.
+	 */
+	if (flags & SLAB_HWCACHE_ALIGN) {
+		unsigned long ralign = cache_line_size();
+		while (size <= ralign / 2)
+			ralign /= 2;
+		align = max(align, ralign);
+	}
+
+	if (align < ARCH_SLAB_MINALIGN)
+		align = ARCH_SLAB_MINALIGN;
+
+	return ALIGN(align, sizeof(void *));
+}
+
+
+/*
  * kmem_cache_create - Create a cache.
  * @name: A string which is used in /proc/slabinfo to identify this cache.
  * @size: The size of objects to be created in this cache.

@@ -115,7 +143,7 @@ struct kmem_cache *kmem_cache_create(con
 	s = kmem_cache_zalloc(kmem_cache, GFP_KERNEL);
 	if (s) {
 		s->object_size = s->size = size;
-		s->align = align;
+		s->align = calculate_alignment(flags, align, size);
 		s->ctor = ctor;
 		s->name = kstrdup(name, GFP_KERNEL);
 		if (!s->name) {

@@ -198,7 +226,7 @@ void __init create_boot_cache(struct kme
 
 	s->name = name;
 	s->size = s->object_size = size;
-	s->align = ARCH_KMALLOC_MINALIGN;
+	s->align = calculate_alignment(flags, ARCH_KMALLOC_MINALIGN, size);
 	err = __kmem_cache_create(s, flags);
 
 	if (err)

Index: linux/mm/slob.c
===================================================================
@@ -124,7 +124,6 @@ static inline void clear_slob_page_free(
 
 #define SLOB_UNIT sizeof(slob_t)
 #define SLOB_UNITS(size) (((size) + SLOB_UNIT - 1)/SLOB_UNIT)
-#define SLOB_ALIGN L1_CACHE_BYTES
 
 /*
  * struct slob_rcu is inserted at the tail of allocated slob blocks, which

@@ -510,20 +509,11 @@ EXPORT_SYMBOL(ksize);
 
 int __kmem_cache_create(struct kmem_cache *c, unsigned long flags)
 {
-	size_t align = c->size;
-
 	if (flags & SLAB_DESTROY_BY_RCU) {
 		/* leave room for rcu footer at the end of object */
 		c->size += sizeof(struct slob_rcu);
 	}
 	c->flags = flags;
-	/* ignore alignment unless it's forced */
-	c->align = (flags & SLAB_HWCACHE_ALIGN) ? SLOB_ALIGN : 0;
-	if (c->align < ARCH_SLAB_MINALIGN)
-		c->align = ARCH_SLAB_MINALIGN;
-	if (c->align < align)
-		c->align = align;
-
 	return 0;
 }
 

Index: linux/mm/slub.c
===================================================================
@@ -2761,32 +2761,6 @@ static inline int calculate_order(int si
 	return -ENOSYS;
 }
 
-/*
- * Figure out what the alignment of the objects will be.
- */
-static unsigned long calculate_alignment(unsigned long flags,
-		unsigned long align, unsigned long size)
-{
-	/*
-	 * If the user wants hardware cache aligned objects then follow that
-	 * suggestion if the object is sufficiently large.
-	 *
-	 * The hardware cache alignment cannot override the specified
-	 * alignment though. If that is greater then use it.
-	 */
-	if (flags & SLAB_HWCACHE_ALIGN) {
-		unsigned long ralign = cache_line_size();
-		while (size <= ralign / 2)
-			ralign /= 2;
-		align = max(align, ralign);
-	}
-
-	if (align < ARCH_SLAB_MINALIGN)
-		align = ARCH_SLAB_MINALIGN;
-
-	return ALIGN(align, sizeof(void *));
-}
-
 static void
 init_kmem_cache_node(struct kmem_cache_node *n)
 {

@@ -2920,7 +2894,6 @@ static int calculate_sizes(struct kmem_c
 {
 	unsigned long flags = s->flags;
 	unsigned long size = s->object_size;
-	unsigned long align = s->align;
 	int order;
 
 	/*

@@ -2992,19 +2965,11 @@ static int calculate_sizes(struct kmem_c
 #endif
 
 	/*
-	 * Determine the alignment based on various parameters that the
-	 * user specified and the dynamic determination of cache line size
-	 * on bootup.
-	 */
-	align = calculate_alignment(flags, align, s->object_size);
-	s->align = align;
-
-	/*
 	 * SLUB stores one object immediately after another beginning from
 	 * offset 0. In order to align the objects we have to simply size
 	 * each object to conform to the alignment.
 	 */
-	size = ALIGN(size, align);
+	size = ALIGN(size, s->align);
 	s->size = size;
 	if (forced_order >= 0)
 		order = forced_order;

@@ -3033,7 +2998,6 @@ static int calculate_sizes(struct kmem_c
 		s->max = s->oo;
 
 	return !!oo_objects(s->oo);
-
 }
 
 static int kmem_cache_open(struct kmem_cache *s, unsigned long flags)

Index: linux/include/linux/slab.h
===================================================================
@@ -162,8 +162,9 @@ unsigned int kmem_cache_size(struct kmem
 #define ARCH_SLAB_MINALIGN __alignof__(unsigned long long)
 #endif
 
+#ifdef CONFIG_SLAB
 /*
- * The largest kmalloc size supported by the slab allocators is
+ * The largest kmalloc size supported by the SLAB allocators is
  * 32 megabyte (2^25) or the maximum allocatable page order if that is
  * less than 32 MB.
  *

@@ -173,9 +174,24 @@ unsigned int kmem_cache_size(struct kmem
  */
 #define KMALLOC_SHIFT_HIGH	((MAX_ORDER + PAGE_SHIFT - 1) <= 25 ? \
 				(MAX_ORDER + PAGE_SHIFT - 1) : 25)
+#define KMALLOC_SHIFT_MAX	KMALLOC_SHIFT_HIGH
+#define KMALLOC_SHIFT_LOW	5
+#else
+/*
+ * SLUB allocates up to order 2 pages directly and otherwise
+ * passes the request to the page allocator.
+ */
+#define KMALLOC_SHIFT_HIGH	(PAGE_SHIFT + 1)
+#define KMALLOC_SHIFT_MAX	(MAX_ORDER + PAGE_SHIFT)
+#define KMALLOC_SHIFT_LOW	3
+#endif
 
-#define KMALLOC_MAX_SIZE	(1UL << KMALLOC_SHIFT_HIGH)
-#define KMALLOC_MAX_ORDER	(KMALLOC_SHIFT_HIGH - PAGE_SHIFT)
+/* Maximum allocatable size */
+#define KMALLOC_MAX_SIZE	(1UL << KMALLOC_SHIFT_MAX)
+/* Maximum size for which we actually use a slab cache */
+#define KMALLOC_MAX_CACHE_SIZE	(1UL << KMALLOC_SHIFT_HIGH)
+/* Maximum order allocatable via the slab allocagtor */
+#define KMALLOC_MAX_ORDER	(KMALLOC_SHIFT_MAX - PAGE_SHIFT)
 
 /*
  * Kmalloc subsystem.

@@ -184,14 +200,8 @@ unsigned int kmem_cache_size(struct kmem
 #if defined(ARCH_DMA_MINALIGN) && ARCH_DMA_MINALIGN > 8
 #define KMALLOC_MIN_SIZE ARCH_DMA_MINALIGN
 #else
-#ifdef CONFIG_SLAB
-#define KMALLOC_MIN_SIZE 32
-#else
-#define KMALLOC_MIN_SIZE 8
+#define KMALLOC_MIN_SIZE (1 << KMALLOC_SHIFT_LOW)
 #endif
-#endif
-
-#define KMALLOC_SHIFT_LOW ilog2(KMALLOC_MIN_SIZE)
 
 /*
  * Figure out which kmalloc slab an allocation of a certain size

Index: linux/include/linux/slub_def.h
===================================================================
@@ -111,19 +111,6 @@ struct kmem_cache {
 	struct kmem_cache_node *node[MAX_NUMNODES];
 };
 
-/*
- * Maximum kmalloc object size handled by SLUB. Larger object allocations
- * are passed through to the page allocator. The page allocator "fastpath"
- * is relatively slow so we need this value sufficiently high so that
- * performance critical objects are allocated through the SLUB fastpath.
- *
- * This should be dropped to PAGE_SIZE / 2 once the page allocator
- * "fastpath" becomes competitive with the slab allocator fastpaths.
- */
-#define SLUB_MAX_SIZE (2 * PAGE_SIZE)
-
-#define SLUB_PAGE_SHIFT (PAGE_SHIFT + 2)
-
 #ifdef CONFIG_ZONE_DMA
 #define SLUB_DMA __GFP_DMA
 #else

@@ -135,7 +122,7 @@ struct kmem_cache {
  * We keep the general caches in an array of slab caches that are used for
  * 2^x bytes of allocations.
  */
-extern struct kmem_cache *kmalloc_caches[SLUB_PAGE_SHIFT];
+extern struct kmem_cache *kmalloc_caches[KMALLOC_SHIFT_HIGH + 1];
 
 /*
  * Find the slab cache for a given combination of allocation flags and size.

@@ -204,7 +191,7 @@ static __always_inline void *kmalloc_lar
 static __always_inline void *kmalloc(size_t size, gfp_t flags)
 {
 	if (__builtin_constant_p(size)) {
-		if (size > SLUB_MAX_SIZE)
+		if (size > KMALLOC_MAX_CACHE_SIZE)
 			return kmalloc_large(size, flags);
 
 		if (!(flags & SLUB_DMA)) {

@@ -240,7 +227,7 @@ kmem_cache_alloc_node_trace(struct kmem_
 static __always_inline void *kmalloc_node(size_t size, gfp_t flags, int node)
 {
 	if (__builtin_constant_p(size) &&
-		size <= SLUB_MAX_SIZE && !(flags & SLUB_DMA)) {
+		size <= KMALLOC_MAX_CACHE_SIZE && !(flags & SLUB_DMA)) {
 			struct kmem_cache *s = kmalloc_slab(size);
 
 		if (!s)

Index: linux/mm/slub.c
===================================================================
@@ -2777,7 +2777,7 @@ init_kmem_cache_node(struct kmem_cache_n
 static inline int alloc_kmem_cache_cpus(struct kmem_cache *s)
 {
 	BUILD_BUG_ON(PERCPU_DYNAMIC_EARLY_SIZE <
-			SLUB_PAGE_SHIFT * sizeof(struct kmem_cache_cpu));
+			KMALLOC_SHIFT_HIGH * sizeof(struct kmem_cache_cpu));
 
 	/*
 	 * Must align to double word boundary for the double cmpxchg

@@ -3174,11 +3174,11 @@ int __kmem_cache_shutdown(struct kmem_ca
  *		Kmalloc subsystem
  *******************************************************************/
 
-struct kmem_cache *kmalloc_caches[SLUB_PAGE_SHIFT];
+struct kmem_cache *kmalloc_caches[KMALLOC_SHIFT_HIGH + 1];
 EXPORT_SYMBOL(kmalloc_caches);
 
 #ifdef CONFIG_ZONE_DMA
-static struct kmem_cache *kmalloc_dma_caches[SLUB_PAGE_SHIFT];
+static struct kmem_cache *kmalloc_dma_caches[KMALLOC_SHIFT_HIGH + 1];
 #endif
 
 static int __init setup_slub_min_order(char *str)

@@ -3280,7 +3280,7 @@ void *__kmalloc(size_t size, gfp_t flags
 	struct kmem_cache *s;
 	void *ret;
 
-	if (unlikely(size > SLUB_MAX_SIZE))
+	if (unlikely(size > KMALLOC_MAX_CACHE_SIZE))
 		return kmalloc_large(size, flags);
 
 	s = get_slab(size, flags);

@@ -3316,7 +3316,7 @@ void *__kmalloc_node(size_t size, gfp_t
 	struct kmem_cache *s;
 	void *ret;
 
-	if (unlikely(size > SLUB_MAX_SIZE)) {
+	if (unlikely(size > KMALLOC_MAX_CACHE_SIZE)) {
 		ret = kmalloc_large_node(size, flags, node);
 
 		trace_kmalloc_node(_RET_IP_, ret,

@@ -3721,7 +3721,7 @@ void __init kmem_cache_init(void)
 		caches++;
 	}
 
-	for (i = KMALLOC_SHIFT_LOW; i < SLUB_PAGE_SHIFT; i++) {
+	for (i = KMALLOC_SHIFT_LOW; i <= KMALLOC_SHIFT_HIGH; i++) {
 		kmalloc_caches[i] = create_kmalloc_cache("kmalloc", 1 << i, 0);
 		caches++;
 	}

@@ -3739,7 +3739,7 @@ void __init kmem_cache_init(void)
 		BUG_ON(!kmalloc_caches[2]->name);
 	}
 
-	for (i = KMALLOC_SHIFT_LOW; i < SLUB_PAGE_SHIFT; i++) {
+	for (i = KMALLOC_SHIFT_LOW; i <= KMALLOC_SHIFT_HIGH; i++) {
 		char *s = kasprintf(GFP_NOWAIT, "kmalloc-%d", 1 << i);
 
 		BUG_ON(!s);

@@ -3751,7 +3751,7 @@ void __init kmem_cache_init(void)
 #endif
 
 #ifdef CONFIG_ZONE_DMA
-	for (i = 0; i < SLUB_PAGE_SHIFT; i++) {
+	for (i = 0; i <= KMALLOC_SHIFT_HIGH; i++) {
 		struct kmem_cache *s = kmalloc_caches[i];
 
 		if (s && s->size) {

@@ -3925,7 +3925,7 @@ void *__kmalloc_track_caller(size_t size
 	struct kmem_cache *s;
 	void *ret;
 
-	if (unlikely(size > SLUB_MAX_SIZE))
+	if (unlikely(size > KMALLOC_MAX_CACHE_SIZE))
 		return kmalloc_large(size, gfpflags);
 
 	s = get_slab(size, gfpflags);

@@ -3948,7 +3948,7 @@ void *__kmalloc_node_track_caller(size_t
 	struct kmem_cache *s;
 	void *ret;
 
-	if (unlikely(size > SLUB_MAX_SIZE)) {
+	if (unlikely(size > KMALLOC_MAX_CACHE_SIZE)) {
 		ret = kmalloc_large_node(size, gfpflags, node);
 
 		trace_kmalloc_node(caller, ret,

@@ -4307,7 +4307,7 @@ static void resiliency_test(void)
 {
 	u8 *p;
 
-	BUILD_BUG_ON(KMALLOC_MIN_SIZE > 16 || SLUB_PAGE_SHIFT < 10);
+	BUILD_BUG_ON(KMALLOC_MIN_SIZE > 16 || KMALLOC_SHIFT_HIGH < 10);
 
 	printk(KERN_ERR "SLUB resiliency testing\n");
 	printk(KERN_ERR "-----------------------\n");

Index: linux/include/linux/slab.h
===================================================================
@@ -178,6 +178,90 @@ unsigned int kmem_cache_size(struct kmem
 #endif
 
 /*
+ * Kmalloc subsystem.
+ */
+#if defined(ARCH_DMA_MINALIGN) && ARCH_DMA_MINALIGN > 8
+#define KMALLOC_MIN_SIZE ARCH_DMA_MINALIGN
+#else
+#ifdef CONFIG_SLAB
+#define KMALLOC_MIN_SIZE 32
+#else
+#define KMALLOC_MIN_SIZE 8
+#endif
+#endif
+
+#define KMALLOC_SHIFT_LOW ilog2(KMALLOC_MIN_SIZE)
+
+/*
+ * Figure out which kmalloc slab an allocation of a certain size
+ * belongs to.
+ * 0 = zero alloc
+ * 1 =  65 .. 96 bytes
+ * 2 = 120 .. 192 bytes
+ * n = 2^(n-1) .. 2^n -1
+ */
+static __always_inline int kmalloc_index(size_t size)
+{
+	if (!size)
+		return 0;
+
+	if (size <= KMALLOC_MIN_SIZE)
+		return KMALLOC_SHIFT_LOW;
+
+	if (KMALLOC_MIN_SIZE <= 32 && size > 64 && size <= 96)
+		return 1;
+	if (KMALLOC_MIN_SIZE <= 64 && size > 128 && size <= 192)
+		return 2;
+	if (size <=          8) return 3;
+	if (size <=         16) return 4;
+	if (size <=         32) return 5;
+	if (size <=         64) return 6;
+	if (size <=        128) return 7;
+	if (size <=        256) return 8;
+	if (size <=        512) return 9;
+	if (size <=       1024) return 10;
+	if (size <=   2 * 1024) return 11;
+	if (size <=   4 * 1024) return 12;
+	if (size <=   8 * 1024) return 13;
+	if (size <=  16 * 1024) return 14;
+	if (size <=  32 * 1024) return 15;
+	if (size <=  64 * 1024) return 16;
+	if (size <= 128 * 1024) return 17;
+	if (size <= 256 * 1024) return 18;
+	if (size <= 512 * 1024) return 19;
+	if (size <= 1024 * 1024) return 20;
+	if (size <=  2 * 1024 * 1024) return 21;
+	if (size <=  4 * 1024 * 1024) return 22;
+	if (size <=  8 * 1024 * 1024) return 23;
+	if (size <=  16 * 1024 * 1024) return 24;
+	if (size <=  32 * 1024 * 1024) return 26;
+	if (size <=  64 * 1024 * 1024) return 27;
+	BUG();
+
+	/* Will never be reached. Needed because the compiler may complain */
+	return -1;
+}
+
+/*
+ * Determine size used for the nth kmalloc cache.
+ * return size or 0 if a kmalloc cache for that
+ * size does not exist
+ */
+static __always_inline int kmalloc_size(int n)
+{
+	if (n > 2)
+		return 1 << n;
+
+	if (n == 1 && KMALLOC_MIN_SIZE <= 32)
+		return 96;
+
+	if (n == 2 && KMALLOC_MIN_SIZE <= 64)
+		return 192;
+
+	return 0;
+}
+
+/*
  * Common kmalloc functions provided by all allocators
  */
 void * __must_check __krealloc(const void *, size_t, gfp_t);

75 lines not on this page · raw message

Index: linux/mm/slab.c
===================================================================
@@ -1623,30 +1623,6 @@ void __init kmem_cache_init(void)
 
 	slab_early_init = 0;
 
-	for (i = 1; i < PAGE_SHIFT + MAX_ORDER; i++) {
-		size_t cs_size = kmalloc_size(i);
-
-		if (cs_size < KMALLOC_MIN_SIZE)
-			continue;
-
-		if (!kmalloc_caches[i]) {
-			/*
-			 * For performance, all the general caches are L1 aligned.
-			 * This should be particularly beneficial on SMP boxes, as it
-			 * eliminates "false sharing".
-			 * Note for systems short on memory removing the alignment will
-			 * allow tighter packing of the smaller caches.
-			 */
-			kmalloc_caches[i] = create_kmalloc_cache("kmalloc",
-					cs_size, ARCH_KMALLOC_FLAGS);
-		}
-
-#ifdef CONFIG_ZONE_DMA
-		kmalloc_dma_caches[i] = create_kmalloc_cache(
-			"kmalloc-dma", cs_size,
-			SLAB_CACHE_DMA|ARCH_KMALLOC_FLAGS);
-#endif
-	}
 	/* 4) Replace the bootstrap head arrays */
 	{
 		struct array_cache *ptr;

@@ -1692,29 +1668,7 @@ void __init kmem_cache_init(void)
 		}
 	}
 
-	slab_state = UP;
-
-	/* Create the proper names */
-	for (i = 1; i < PAGE_SHIFT + MAX_ORDER; i++) {
-		char *s;
-		struct kmem_cache *c = kmalloc_caches[i];
-
-		if (!c)
-			continue;
-
-		s = kasprintf(GFP_NOWAIT, "kmalloc-%d", kmalloc_size(i));
-
-		BUG_ON(!s);
-		c->name = s;
-		
-#ifdef CONFIG_ZONE_DMA
-		c = kmalloc_dma_caches[i];
-		BUG_ON(!c);
-		s = kasprintf(GFP_NOWAIT, "dma-kmalloc-%d", kmalloc_size(i));
-		BUG_ON(!s);
-		c->name = s;
-#endif
-	}
+	create_kmalloc_caches(ARCH_KMALLOC_FLAGS);
 }
 
 void __init kmem_cache_init_late(void)

Index: linux/mm/slab.h
===================================================================
@@ -35,6 +35,12 @@ extern struct kmem_cache *kmem_cache;
 unsigned long calculate_alignment(unsigned long flags,
 		unsigned long align, unsigned long size);
 
+#ifndef CONFIG_SLOB
+/* Kmalloc array related functions */
+void create_kmalloc_caches(gfp_t);
+#endif
+
+
 /* Functions provided by the slab allocators */
 extern int __kmem_cache_create(struct kmem_cache *, unsigned long flags);
 

58 lines not on this page · raw message

74 lines not on this page · raw message