* slab allocators: Remove obsolete SLAB_MUST_HWCACHE_ALIGN
@ 2007-04-17 5:35 Christoph Lameter
2007-04-17 7:31 ` Pekka Enberg
2007-04-17 23:09 ` Matt Mackall
0 siblings, 2 replies; 3+ messages in thread
From: Christoph Lameter @ 2007-04-17 5:35 UTC (permalink / raw)
To: akpm; +Cc: linux-kernel
The flag SLAB_MUST_HWCACHE_ALIGN is
1. Never checked by SLAB at all.
2. A duplicate of SLAB_HWCACHE_ALIGN for SLUB
3. Fulfills the role of SLAB_HWCACHE_ALIGN for SLOB.
The only remaining use is in sparc64 and ppc64 and their use there
reflects some earlier role that the slab flag once may have had. If
its specified then SLAB_HWCACHE_ALIGN is also specified.
The flag is confusing, inconsistent and has no purpose.
Remove it.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Index: linux-2.6.21-rc6/include/linux/slab.h
===================================================================
--- linux-2.6.21-rc6.orig/include/linux/slab.h 2007-04-16 21:55:03.000000000 -0700
+++ linux-2.6.21-rc6/include/linux/slab.h 2007-04-16 21:55:10.000000000 -0700
@@ -26,7 +26,6 @@ typedef struct kmem_cache kmem_cache_t _
#define SLAB_POISON 0x00000800UL /* DEBUG: Poison objects */
#define SLAB_HWCACHE_ALIGN 0x00002000UL /* Align objs on cache lines */
#define SLAB_CACHE_DMA 0x00004000UL /* Use GFP_DMA memory */
-#define SLAB_MUST_HWCACHE_ALIGN 0x00008000UL /* Force alignment even if debuggin is active */
#define SLAB_STORE_USER 0x00010000UL /* DEBUG: Store the last owner for bug hunting */
#define SLAB_RECLAIM_ACCOUNT 0x00020000UL /* Objects are reclaimable */
#define SLAB_PANIC 0x00040000UL /* Panic if kmem_cache_create() fails */
Index: linux-2.6.21-rc6/mm/slab.c
===================================================================
--- linux-2.6.21-rc6.orig/mm/slab.c 2007-04-16 21:55:16.000000000 -0700
+++ linux-2.6.21-rc6/mm/slab.c 2007-04-16 21:55:33.000000000 -0700
@@ -175,12 +175,12 @@
# define CREATE_MASK (SLAB_DEBUG_INITIAL | SLAB_RED_ZONE | \
SLAB_POISON | SLAB_HWCACHE_ALIGN | \
SLAB_CACHE_DMA | \
- SLAB_MUST_HWCACHE_ALIGN | SLAB_STORE_USER | \
+ SLAB_STORE_USER | \
SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | \
SLAB_DESTROY_BY_RCU | SLAB_MEM_SPREAD)
#else
# define CREATE_MASK (SLAB_HWCACHE_ALIGN | \
- SLAB_CACHE_DMA | SLAB_MUST_HWCACHE_ALIGN | \
+ SLAB_CACHE_DMA | \
SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | \
SLAB_DESTROY_BY_RCU | SLAB_MEM_SPREAD)
#endif
Index: linux-2.6.21-rc6/mm/slub.c
===================================================================
--- linux-2.6.21-rc6.orig/mm/slub.c 2007-04-16 21:55:38.000000000 -0700
+++ linux-2.6.21-rc6/mm/slub.c 2007-04-16 21:56:07.000000000 -0700
@@ -1500,7 +1500,7 @@ static int calculate_order(int size)
static unsigned long calculate_alignment(unsigned long flags,
unsigned long align)
{
- if (flags & (SLAB_MUST_HWCACHE_ALIGN | SLAB_HWCACHE_ALIGN))
+ if (flags & SLAB_HWCACHE_ALIGN)
return max_t(unsigned long, align, L1_CACHE_BYTES);
if (align < ARCH_SLAB_MINALIGN)
@@ -3083,8 +3083,7 @@ SLAB_ATTR(reclaim_account);
static ssize_t hwcache_align_show(struct kmem_cache *s, char *buf)
{
- return sprintf(buf, "%d\n", !!(s->flags &
- (SLAB_HWCACHE_ALIGN|SLAB_MUST_HWCACHE_ALIGN)));
+ return sprintf(buf, "%d\n", !!(s->flags & SLAB_HWCACHE_ALIGN));
}
SLAB_ATTR_RO(hwcache_align);
Index: linux-2.6.21-rc6/arch/powerpc/mm/hugetlbpage.c
===================================================================
--- linux-2.6.21-rc6.orig/arch/powerpc/mm/hugetlbpage.c 2007-04-16 21:58:53.000000000 -0700
+++ linux-2.6.21-rc6/arch/powerpc/mm/hugetlbpage.c 2007-04-16 21:59:02.000000000 -0700
@@ -1063,8 +1063,7 @@ static int __init hugetlbpage_init(void)
huge_pgtable_cache = kmem_cache_create("hugepte_cache",
HUGEPTE_TABLE_SIZE,
HUGEPTE_TABLE_SIZE,
- SLAB_HWCACHE_ALIGN |
- SLAB_MUST_HWCACHE_ALIGN,
+ SLAB_HWCACHE_ALIGN,
zero_ctor, NULL);
if (! huge_pgtable_cache)
panic("hugetlbpage_init(): could not create hugepte cache\n");
Index: linux-2.6.21-rc6/arch/powerpc/mm/init_64.c
===================================================================
--- linux-2.6.21-rc6.orig/arch/powerpc/mm/init_64.c 2007-04-16 21:59:08.000000000 -0700
+++ linux-2.6.21-rc6/arch/powerpc/mm/init_64.c 2007-04-16 21:59:19.000000000 -0700
@@ -183,8 +183,7 @@ void pgtable_cache_init(void)
"for size: %08x...\n", name, i, size);
pgtable_cache[i] = kmem_cache_create(name,
size, size,
- SLAB_HWCACHE_ALIGN |
- SLAB_MUST_HWCACHE_ALIGN,
+ SLAB_HWCACHE_ALIGN,
zero_ctor,
NULL);
if (! pgtable_cache[i])
Index: linux-2.6.21-rc6/arch/sparc64/mm/init.c
===================================================================
--- linux-2.6.21-rc6.orig/arch/sparc64/mm/init.c 2007-04-16 21:59:32.000000000 -0700
+++ linux-2.6.21-rc6/arch/sparc64/mm/init.c 2007-04-16 21:59:40.000000000 -0700
@@ -191,8 +191,7 @@ void pgtable_cache_init(void)
{
pgtable_cache = kmem_cache_create("pgtable_cache",
PAGE_SIZE, PAGE_SIZE,
- SLAB_HWCACHE_ALIGN |
- SLAB_MUST_HWCACHE_ALIGN,
+ SLAB_HWCACHE_ALIGN,
zero_ctor,
NULL);
if (!pgtable_cache) {
Index: linux-2.6.21-rc6/arch/sparc64/mm/tsb.c
===================================================================
--- linux-2.6.21-rc6.orig/arch/sparc64/mm/tsb.c 2007-04-16 22:00:02.000000000 -0700
+++ linux-2.6.21-rc6/arch/sparc64/mm/tsb.c 2007-04-16 22:00:12.000000000 -0700
@@ -262,8 +262,7 @@ void __init tsb_cache_init(void)
tsb_caches[i] = kmem_cache_create(name,
size, size,
- SLAB_HWCACHE_ALIGN |
- SLAB_MUST_HWCACHE_ALIGN,
+ SLAB_HWCACHE_ALIGN,
NULL, NULL);
if (!tsb_caches[i]) {
prom_printf("Could not create %s cache\n", name);
Index: linux-2.6.21-rc6/mm/slob.c
===================================================================
--- linux-2.6.21-rc6.orig/mm/slob.c 2007-04-16 21:57:43.000000000 -0700
+++ linux-2.6.21-rc6/mm/slob.c 2007-04-16 21:58:28.000000000 -0700
@@ -21,7 +21,7 @@
*
* SLAB is emulated on top of SLOB by simply calling constructors and
* destructors for every SLAB allocation. Objects are returned with
- * the 8-byte alignment unless the SLAB_MUST_HWCACHE_ALIGN flag is
+ * the 8-byte alignment unless the SLAB_HWCACHE_ALIGN flag is
* set, in which case the low-level allocator will fragment blocks to
* create the proper alignment. Again, objects of page-size or greater
* are allocated by calling __get_free_pages. As SLAB objects know
@@ -295,7 +295,7 @@ struct kmem_cache *kmem_cache_create(con
c->ctor = ctor;
c->dtor = dtor;
/* ignore alignment unless it's forced */
- c->align = (flags & SLAB_MUST_HWCACHE_ALIGN) ? SLOB_ALIGN : 0;
+ c->align = (flags & SLAB_HWCACHE_ALIGN) ? SLOB_ALIGN : 0;
if (c->align < align)
c->align = align;
}
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: slab allocators: Remove obsolete SLAB_MUST_HWCACHE_ALIGN
2007-04-17 5:35 slab allocators: Remove obsolete SLAB_MUST_HWCACHE_ALIGN Christoph Lameter
@ 2007-04-17 7:31 ` Pekka Enberg
2007-04-17 23:09 ` Matt Mackall
1 sibling, 0 replies; 3+ messages in thread
From: Pekka Enberg @ 2007-04-17 7:31 UTC (permalink / raw)
To: Christoph Lameter; +Cc: akpm, linux-kernel
On 4/17/07, Christoph Lameter <clameter@sgi.com> wrote:
> The flag SLAB_MUST_HWCACHE_ALIGN is
>
> 1. Never checked by SLAB at all.
>
> 2. A duplicate of SLAB_HWCACHE_ALIGN for SLUB
>
> 3. Fulfills the role of SLAB_HWCACHE_ALIGN for SLOB.
>
> The only remaining use is in sparc64 and ppc64 and their use there
> reflects some earlier role that the slab flag once may have had. If
> its specified then SLAB_HWCACHE_ALIGN is also specified.
>
> The flag is confusing, inconsistent and has no purpose.
>
> Remove it.
Acked-by: Pekka Enberg <penberg@cs.helsinki.fi>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: slab allocators: Remove obsolete SLAB_MUST_HWCACHE_ALIGN
2007-04-17 5:35 slab allocators: Remove obsolete SLAB_MUST_HWCACHE_ALIGN Christoph Lameter
2007-04-17 7:31 ` Pekka Enberg
@ 2007-04-17 23:09 ` Matt Mackall
1 sibling, 0 replies; 3+ messages in thread
From: Matt Mackall @ 2007-04-17 23:09 UTC (permalink / raw)
To: Christoph Lameter; +Cc: akpm, linux-kernel
On Mon, Apr 16, 2007 at 10:35:21PM -0700, Christoph Lameter wrote:
> The flag SLAB_MUST_HWCACHE_ALIGN is
>
> 1. Never checked by SLAB at all.
>
> 2. A duplicate of SLAB_HWCACHE_ALIGN for SLUB
>
> 3. Fulfills the role of SLAB_HWCACHE_ALIGN for SLOB.
>
> The only remaining use is in sparc64 and ppc64 and their use there
> reflects some earlier role that the slab flag once may have had. If
> its specified then SLAB_HWCACHE_ALIGN is also specified.
>
> The flag is confusing, inconsistent and has no purpose.
>
> Remove it.
Hmm, at one point this was needed by SLOB to boot on x86, I wonder
when that changed? But if SLAB is using the other define now, I'm
happy to switch.
> Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Matt Mackall <mpm@selenic.com>
--
Mathematics is the supreme nostalgia of our time.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2007-04-17 23:25 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-04-17 5:35 slab allocators: Remove obsolete SLAB_MUST_HWCACHE_ALIGN Christoph Lameter
2007-04-17 7:31 ` Pekka Enberg
2007-04-17 23:09 ` Matt Mackall
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®