public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Ravikiran G Thirumalai <kiran@scalex86.org>
To: Andrew Morton <akpm@osdl.org>
Cc: linux-kernel@vger.kernel.org,
	Alok Kataria <alok.kataria@calsoftinc.com>,
	"Shai Fultheim (Shai@scalex86.org)" <shai@scalex86.org>,
	Christoph Lameter <clameter@engr.sgi.com>
Subject: [patch] slab: Do not use mempolicy for kmalloc_node
Date: Tue, 12 Sep 2006 07:45:18 -0700	[thread overview]
Message-ID: <20060912144518.GA4653@localhost.localdomain> (raw)

The slab should follow the specified memory policy for kmalloc allocations,
which it does.  However, for kmalloc_node allocations, slab should
serve the object from the requested node irrespective of memory policy.
This seems to be broken in slab code.  Following patch fixes this.

Patch abstacts out a __cache_alloc_mempolicy function to be used when
mempolicy is to be applied.
 
Signed-off-by: Alok N Kataria <alok.kataria@calsotinc.com>
Signed-off-by: Ravikiran Thirumalai <kiran@scale86.org>
Signed-off-by: Shai Fultheim <shai@scale86.org>

Index: linux-2.6.18-rc3/mm/slab.c
===================================================================
--- linux-2.6.18-rc3.orig/mm/slab.c	2006-08-04 10:01:46.000000000 -0700
+++ linux-2.6.18-rc3/mm/slab.c	2006-08-08 12:05:21.000000000 -0700
@@ -2963,19 +2963,12 @@ static void *cache_alloc_debugcheck_afte
 #define cache_alloc_debugcheck_after(a,b,objp,d) (objp)
 #endif
 
-static inline void *____cache_alloc(struct kmem_cache *cachep, gfp_t flags)
+static inline void *
+____cache_alloc(struct kmem_cache *cachep, gfp_t flags)
 {
 	void *objp;
 	struct array_cache *ac;
 
-#ifdef CONFIG_NUMA
-	if (unlikely(current->flags & (PF_SPREAD_SLAB | PF_MEMPOLICY))) {
-		objp = alternate_node_alloc(cachep, flags);
-		if (objp != NULL)
-			return objp;
-	}
-#endif
-
 	check_irq_off();
 	ac = cpu_cache_get(cachep);
 	if (likely(ac->avail)) {
@@ -2989,6 +2982,28 @@ static inline void *____cache_alloc(stru
 	return objp;
 }
 
+#ifdef CONFIG_NUMA
+static inline void *__cache_alloc_mempolicy(struct kmem_cache *cachep, gfp_t flags)
+{
+	void *objp;
+
+	if (unlikely(current->flags & (PF_SPREAD_SLAB | PF_MEMPOLICY))) {
+		objp = alternate_node_alloc(cachep, flags);
+		if (objp != NULL)
+			return objp;
+	}
+
+	return ____cache_alloc(cachep, flags);
+}
+#else
+static inline void *__cache_alloc_mempolicy(struct kmem_cache *cachep, gfp_t flags)
+{
+	return ____cache_alloc(cachep, flags);
+}
+#endif
+
+
+
 static __always_inline void *__cache_alloc(struct kmem_cache *cachep,
 						gfp_t flags, void *caller)
 {
@@ -2998,7 +3013,7 @@ static __always_inline void *__cache_all
 	cache_alloc_debugcheck_before(cachep, flags);
 
 	local_irq_save(save_flags);
-	objp = ____cache_alloc(cachep, flags);
+	objp = __cache_alloc_mempolicy(cachep, flags);
 	local_irq_restore(save_flags);
 	objp = cache_alloc_debugcheck_after(cachep, flags, objp,
 					    caller);
@@ -3303,8 +3318,9 @@ void *kmem_cache_alloc_node(struct kmem_
 	cache_alloc_debugcheck_before(cachep, flags);
 	local_irq_save(save_flags);
 
-	if (nodeid == -1 || nodeid == numa_node_id() ||
-			!cachep->nodelists[nodeid])
+	if (nodeid == -1 || !cachep->nodelists[nodeid])
+		ptr = __cache_alloc_mempolicy(cachep, flags);
+	else if (nodeid == numa_node_id())
 		ptr = ____cache_alloc(cachep, flags);
 	else
 		ptr = __cache_alloc_node(cachep, flags, nodeid);

             reply	other threads:[~2006-09-12 14:43 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-09-12 14:45 Ravikiran G Thirumalai [this message]
2006-09-12 17:36 ` [patch] slab: Do not use mempolicy for kmalloc_node Christoph Lameter
2006-09-12 19:52   ` Ravikiran G Thirumalai
2006-09-12 19:52     ` Christoph Lameter
2006-09-13 22:14       ` Ravikiran G Thirumalai
2006-09-13 22:28         ` Christoph Lameter
2006-09-13 23:37           ` Ravikiran G Thirumalai
2006-09-13 23:48             ` Christoph Lameter
2006-09-13 23:59               ` Ravikiran G Thirumalai

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20060912144518.GA4653@localhost.localdomain \
    --to=kiran@scalex86.org \
    --cc=akpm@osdl.org \
    --cc=alok.kataria@calsoftinc.com \
    --cc=clameter@engr.sgi.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=shai@scalex86.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox