linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2] slab: Ignore internal flags in cache creation
@ 2012-10-01 10:47 Glauber Costa
  2012-10-01 18:00 ` Christoph Lameter
  2012-10-02  3:16 ` David Rientjes
  0 siblings, 2 replies; 3+ messages in thread
From: Glauber Costa @ 2012-10-01 10:47 UTC (permalink / raw)
  To: Pekka Enberg
  Cc: linux-mm, linux-kernel, Glauber Costa, Christoph Lameter,
	Pekka Enberg, David Rientjes

Some flags are used internally by the allocators for management
purposes. One example of that is the CFLGS_OFF_SLAB flag that slab uses
to mark that the metadata for that cache is stored outside of the slab.

No cache should ever pass those as a creation flags. We can just ignore
this bit if it happens to be passed (such as when duplicating a cache in
the kmem memcg patches).

Because such flags can vary from allocator to allocator, we allow them
to make their own decisions on that. Those who want it, can define
CACHE_CREATE_MASK, with all flags that are valid at creation time.
Common code will mask out all flags not belonging to that set.

[ v2: leave the mask out decision up to the allocators ]

Signed-off-by: Glauber Costa <glommer@parallels.com>
CC: Christoph Lameter <cl@linux.com>
CC: Pekka Enberg <penberg@cs.helsinki.fi>
CC: David Rientjes <rientjes@google.com>
---
 include/linux/slab_def.h | 19 +++++++++++++++++++
 mm/slab.c                | 22 ----------------------
 mm/slab_common.c         |  9 +++++++++
 3 files changed, 28 insertions(+), 22 deletions(-)

diff --git a/include/linux/slab_def.h b/include/linux/slab_def.h
index 36d7031..f7ec03d 100644
--- a/include/linux/slab_def.h
+++ b/include/linux/slab_def.h
@@ -15,6 +15,25 @@
 #include <asm/cache.h>		/* kmalloc_sizes.h needs L1_CACHE_BYTES */
 #include <linux/compiler.h>
 
+/* Legal flag mask for kmem_cache_create(). */
+#ifdef CONFIG_DEBUG_SLAB 
+#define CACHE_CREATE_MASK (SLAB_RED_ZONE | 				\
+			SLAB_POISON | SLAB_HWCACHE_ALIGN | 		\
+			SLAB_CACHE_DMA | 				\
+			SLAB_STORE_USER | 				\
+			SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | 		\
+			SLAB_DESTROY_BY_RCU | SLAB_MEM_SPREAD | 	\
+			SLAB_DEBUG_OBJECTS | SLAB_NOLEAKTRACE |	\
+			SLAB_NOTRACK)
+#else
+#define CACHE_CREATE_MASK (SLAB_HWCACHE_ALIGN | 			\
+			SLAB_CACHE_DMA | 				\
+			SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | 		\
+			SLAB_DESTROY_BY_RCU | SLAB_MEM_SPREAD | 	\
+			SLAB_DEBUG_OBJECTS | SLAB_NOLEAKTRACE | 	\
+			SLAB_NOTRACK)
+#endif
+
 /*
  * struct kmem_cache
  *
diff --git a/mm/slab.c b/mm/slab.c
index 8524923..8c1d447 100644
--- a/mm/slab.c
+++ b/mm/slab.c
@@ -162,23 +162,6 @@
  */
 static bool pfmemalloc_active __read_mostly;
 
-/* Legal flag mask for kmem_cache_create(). */
-#if DEBUG
-# define CREATE_MASK	(SLAB_RED_ZONE | \
-			 SLAB_POISON | SLAB_HWCACHE_ALIGN | \
-			 SLAB_CACHE_DMA | \
-			 SLAB_STORE_USER | \
-			 SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | \
-			 SLAB_DESTROY_BY_RCU | SLAB_MEM_SPREAD | \
-			 SLAB_DEBUG_OBJECTS | SLAB_NOLEAKTRACE | SLAB_NOTRACK)
-#else
-# define CREATE_MASK	(SLAB_HWCACHE_ALIGN | \
-			 SLAB_CACHE_DMA | \
-			 SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | \
-			 SLAB_DESTROY_BY_RCU | SLAB_MEM_SPREAD | \
-			 SLAB_DEBUG_OBJECTS | SLAB_NOLEAKTRACE | SLAB_NOTRACK)
-#endif
-
 /*
  * kmem_bufctl_t:
  *
@@ -2385,11 +2368,6 @@ __kmem_cache_create (struct kmem_cache *cachep, unsigned long flags)
 	if (flags & SLAB_DESTROY_BY_RCU)
 		BUG_ON(flags & SLAB_POISON);
 #endif
-	/*
-	 * Always checks flags, a caller might be expecting debug support which
-	 * isn't available.
-	 */
-	BUG_ON(flags & ~CREATE_MASK);
 
 	/*
 	 * Check that size is in terms of words.  This is needed to avoid
diff --git a/mm/slab_common.c b/mm/slab_common.c
index 9c21725..f2682ee 100644
--- a/mm/slab_common.c
+++ b/mm/slab_common.c
@@ -107,6 +107,15 @@ struct kmem_cache *kmem_cache_create(const char *name, size_t size, size_t align
 	if (!kmem_cache_sanity_check(name, size) == 0)
 		goto out_locked;
 
+	/*
+	 * Some allocators will constraint the set of valid flags to a subset
+	 * of all flags. We expect them to define CACHE_CREATE_MASK in this
+	 * case, and we'll just provide them with a sanitized version of the
+	 * passed flags.
+	 */
+#ifdef CACHE_CREATE_MASK
+	flags &= ~CACHE_CREATE_MASK;
+#endif
 
 	s = __kmem_cache_alias(name, size, align, flags, ctor);
 	if (s)
-- 
1.7.11.4


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH v2] slab: Ignore internal flags in cache creation
  2012-10-01 10:47 [PATCH v2] slab: Ignore internal flags in cache creation Glauber Costa
@ 2012-10-01 18:00 ` Christoph Lameter
  2012-10-02  3:16 ` David Rientjes
  1 sibling, 0 replies; 3+ messages in thread
From: Christoph Lameter @ 2012-10-01 18:00 UTC (permalink / raw)
  To: Glauber Costa
  Cc: Pekka Enberg, linux-mm, linux-kernel, Pekka Enberg, David Rientjes

On Mon, 1 Oct 2012, Glauber Costa wrote:

> [ v2: leave the mask out decision up to the allocators ]

Acked-by: Christoph Lameter <cl@linux.com>

I would prefer that this mask be named appropriately and be defined in all
sl*_def.h files for all allocators.

Name could be SLAB_AVAILABLE_FLAGS or some name that makes more
sense than CACHE_CREATE_MASK.


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH v2] slab: Ignore internal flags in cache creation
  2012-10-01 10:47 [PATCH v2] slab: Ignore internal flags in cache creation Glauber Costa
  2012-10-01 18:00 ` Christoph Lameter
@ 2012-10-02  3:16 ` David Rientjes
  1 sibling, 0 replies; 3+ messages in thread
From: David Rientjes @ 2012-10-02  3:16 UTC (permalink / raw)
  To: Glauber Costa
  Cc: Pekka Enberg, linux-mm, linux-kernel, Christoph Lameter, Pekka Enberg

On Mon, 1 Oct 2012, Glauber Costa wrote:

> diff --git a/mm/slab_common.c b/mm/slab_common.c
> index 9c21725..f2682ee 100644
> --- a/mm/slab_common.c
> +++ b/mm/slab_common.c
> @@ -107,6 +107,15 @@ struct kmem_cache *kmem_cache_create(const char *name, size_t size, size_t align
>  	if (!kmem_cache_sanity_check(name, size) == 0)
>  		goto out_locked;
>  
> +	/*
> +	 * Some allocators will constraint the set of valid flags to a subset
> +	 * of all flags. We expect them to define CACHE_CREATE_MASK in this
> +	 * case, and we'll just provide them with a sanitized version of the
> +	 * passed flags.
> +	 */
> +#ifdef CACHE_CREATE_MASK
> +	flags &= ~CACHE_CREATE_MASK;
> +#endif
>  
>  	s = __kmem_cache_alias(name, size, align, flags, ctor);
>  	if (s)

flags &= CACHE_CREATE_MASK

After that's done:

	Acked-by: David Rientjes <rientjes@google.com>

Thanks for working through this.

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2012-10-02  3:16 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-10-01 10:47 [PATCH v2] slab: Ignore internal flags in cache creation Glauber Costa
2012-10-01 18:00 ` Christoph Lameter
2012-10-02  3:16 ` David Rientjes

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).