Skip to contentSkip to conversation

slab: resize the alien caches too

PATCH In review Siddha, Suresh B 3 participants 4 messages Apr 18, 2007 Memory management

lore
Share & tools
lore

Siddha, Suresh B

author
#0 lore raw

PATCH 1/1 slab: resize the alien caches too

Christoph,

While going through the slab code, I observed that alien caches are not getting resized, when user changes the slab tunables. Appended patch tries to fix this. Please review and let me know if I missed anything.

thanks,
suresh
---

Resize the alien caches too based on the slab tunables.

Signed-off-by Suresh Siddha <[email protected]>

---

mm/slab.c +6 -5
index 4cbac24..e0dd9af 100644
@@ -3823,6 +3823,7 @@ static int alloc_kmemlist(struct kmem_cache *cachep)
 		l3 = cachep->nodelists[node];
 		if (l3) {
 			struct array_cache *shared = l3->shared;
+			struct array_cache **alien = l3->alien;
 
 			spin_lock_irq(&l3->list_lock);
 

@@ -3831,15 +3832,15 @@ static int alloc_kmemlist(struct kmem_cache *cachep)
 						shared->avail, node);
 
 			l3->shared = new_shared;
-			if (!l3->alien) {
-				l3->alien = new_alien;
-				new_alien = NULL;
-			}
+			l3->alien = new_alien;
 			l3->free_limit = (1 + nr_cpus_node(node)) *
 					cachep->batchcount + cachep->num;
 			spin_unlock_irq(&l3->list_lock);
 			kfree(shared);
-			free_alien_cache(new_alien);
+			if (alien) {
+				drain_alien_cache(cachep, alien);
+				free_alien_cache(alien);
+			}
 			continue;
 		}
 		l3 = kmalloc_node(sizeof(struct kmem_list3), GFP_KERNEL, node);
#1 lore raw
Quoted 3 lines · Suresh B

While going through the slab code, I observed that alien caches are not getting resized, when user changes the slab tunables. Appended patch tries to fix this. Please review and let me know if I missed anything.

Let it be. We have not resized the alien caches ever and doing so may cause serious memory usage. Size is hardcoded to 12 anyways.

author
#2 lore raw
Quoted 8 lines · Christoph Lameter
Quoted 3 lines · Suresh B

While going through the slab code, I observed that alien caches are not getting resized, when user changes the slab tunables. Appended patch tries to fix this. Please review and let me know if I missed anything.

Let it be. We have not resized the alien caches ever and doing so may cause serious memory usage. Size is hardcoded to 12 anyways.

hmm.. I see.

Let me do some experiments and getback to you(with what I am looking exactly).

thanks,
suresh

#3 lore raw
In reply to Siddha, Suresh B
Quoted 20 lines · Suresh B

Christoph,

While going through the slab code, I observed that alien caches are not getting resized, when user changes the slab tunables. Appended patch tries to fix this. Please review and let me know if I missed anything.

thanks,
suresh
---

Resize the alien caches too based on the slab tunables.

Signed-off-by Suresh Siddha <[email protected]>

---

mm/slab.c +0 -0
index 4cbac24..e0dd9af 100644

mm/slab.c static int alloc_kmemlist(struct kmem_cache *cachep) · line 3829

Quoted 7 lines
mm/slab.c +1 -0
@@ -3823,6 +3823,7 @@ static int alloc_kmemlist(struct kmem_cache *cachep)
 		l3 = cachep->nodelists[node];
 		if (l3) {
 			struct array_cache *shared = l3->shared;
+			struct array_cache **alien = l3->alien;
 
 			spin_lock_irq(&l3->list_lock);

Christoph already rejected the patch, but I have one further comment : l3->alien should be fetched after spin_lock_irq() of course.

It's true that the :
if (limit > 1) limit = 12;
from alloc_alien_cache() is quite disturbing and could be cleaned up.