Merge branch 'slab-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/christoph/vm

* 'slab-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/christoph/vm:
  slub: fix typo in Documentation/vm/slub.txt
  slab: NUMA slab allocator migration bugfix
  slub: Do not cross cacheline boundaries for very small objects
  slab - use angle brackets for include of kmalloc_sizes.h
  slab numa fallback logic: Do not pass unfiltered flags to page allocator
  slub statistics: Fix check for DEACTIVATE_REMOTE_FREES

Documentation/vm/slub.txt

@@ -50,14 +50,14 @@ F.e. in order to boot just with sanity checks and red zoning one would specify:
 
 Trying to find an issue in the dentry cache? Try
 
-	slub_debug=,dentry_cache
+	slub_debug=,dentry
 
 to only enable debugging on the dentry cache.
 
 Red zoning and tracking may realign the slab.  We can just apply sanity checks
 to the dentry cache with
 
-	slub_debug=F,dentry_cache
+	slub_debug=F,dentry
 
 In case you forgot to enable debugging on the kernel command line: It is
 possible to enable debugging manually when the kernel is up. Look at the

include/linux/slab_def.h

@@ -41,7 +41,7 @@ static inline void *kmalloc(size_t size, gfp_t flags)
 			goto found; \
 		else \
 			i++;
-#include "kmalloc_sizes.h"
+#include <linux/kmalloc_sizes.h>
 #undef CACHE
 		{
 			extern void __you_cannot_kmalloc_that_much(void);
@@ -75,7 +75,7 @@ static inline void *kmalloc_node(size_t size, gfp_t flags, int node)
 			goto found; \
 		else \
 			i++;
-#include "kmalloc_sizes.h"
+#include <linux/kmalloc_sizes.h>
 #undef CACHE
 		{
 			extern void __you_cannot_kmalloc_that_much(void);

mm/slab.c

@@ -333,7 +333,7 @@ static __always_inline int index_of(const size_t size)
 		return i; \
 	else \
 		i++;
-#include "linux/kmalloc_sizes.h"
+#include <linux/kmalloc_sizes.h>
 #undef CACHE
 		__bad_size();
 	} else
@@ -2964,11 +2964,10 @@ static void *cache_alloc_refill(struct kmem_cache *cachep, gfp_t flags)
 	struct array_cache *ac;
 	int node;
 
-	node = numa_node_id();
-
+retry:
 	check_irq_off();
+	node = numa_node_id();
 	ac = cpu_cache_get(cachep);
-retry:
 	batchcount = ac->batchcount;
 	if (!ac->touched && batchcount > BATCHREFILL_LIMIT) {
 		/*
@@ -3280,7 +3279,7 @@ static void *fallback_alloc(struct kmem_cache *cache, gfp_t flags)
 		if (local_flags & __GFP_WAIT)
 			local_irq_enable();
 		kmem_flagcheck(cache, flags);
-		obj = kmem_getpages(cache, flags, -1);
+		obj = kmem_getpages(cache, local_flags, -1);
 		if (local_flags & __GFP_WAIT)
 			local_irq_disable();
 		if (obj) {

mm/slub.c

@@ -1368,7 +1368,7 @@ static void deactivate_slab(struct kmem_cache *s, struct kmem_cache_cpu *c)
 	struct page *page = c->page;
 	int tail = 1;
 
-	if (c->freelist)
+	if (page->freelist)
 		stat(c, DEACTIVATE_REMOTE_FREES);
 	/*
 	 * Merge cpu freelist into slab freelist. Typically we get here
@@ -1856,12 +1856,15 @@ static unsigned long calculate_alignment(unsigned long flags,
 	 * The hardware cache alignment cannot override the specified
 	 * alignment though. If that is greater then use it.
 	 */
-	if ((flags & SLAB_HWCACHE_ALIGN) &&
-			size > cache_line_size() / 2)
-		return max_t(unsigned long, align, cache_line_size());
+	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)
-		return ARCH_SLAB_MINALIGN;
+		align = ARCH_SLAB_MINALIGN;
 
 	return ALIGN(align, sizeof(void *));
 }