Messages in this thread Patch in this message | | | Date | Wed, 21 Mar 2007 17:42:22 +0200 (EET) | From | Pekka J Enberg <> | Subject | Re: [RFC, PATCH] SLAB : [NUMA] keep nodeid in struct page instead of struct slab |
| |
On Wed, 21 Mar 2007, Christoph Lameter wrote: > None of that please. The page flags already contain a node number that is > accessible via page_to_nid(). Just make sure that we never put a slab onto > the wrong node.
Oh well, here's a patch to find out. (I don't have a NUMA box.)
Pekka
From: Pekka Enberg <penberg@cs.helsinki.fi>
This adds a virt_to_nodeid() for looking up the NUMA node ID for a slab object. The current implementation uses slab->nodeid but adds a WARN_ON to catch cases where page nid differs. Eventually, we can change virt_to_nodeid() to use page_to_nid after which slab->nodeid can be removed.
Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi> --- mm/slab.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-)
Index: 2.6/mm/slab.c =================================================================== --- 2.6.orig/mm/slab.c 2007-03-21 16:58:53.000000000 +0200 +++ 2.6/mm/slab.c 2007-03-21 16:59:26.000000000 +0200 @@ -623,6 +623,14 @@ static inline struct slab *virt_to_slab( return page_get_slab(page); } +static inline unsigned short virt_to_nodeid(const void *obj) +{ + struct page *page = virt_to_page(obj); + struct slab *slab = page_get_slab(page); + WARN_ON(slab->nodeid != page_to_nid(page)); + return slab->nodeid; +} + static inline void *index_to_obj(struct kmem_cache *cache, struct slab *slab, unsigned int idx) { @@ -1134,8 +1142,7 @@ int i = 0; static inline int cache_free_alien(struct kmem_cache *cachep, void *objp) { - struct slab *slabp = virt_to_slab(objp); - int nodeid = slabp->nodeid; + int nodeid = virt_to_nodeid(objp); struct kmem_list3 *l3; struct array_cache *alien = NULL; int node; @@ -1146,7 +1153,7 @@ static inline int cache_free_alien(struc * Make sure we are not freeing a object from another node to the array * cache on this cpu. */ - if (likely(slabp->nodeid == node) || unlikely(!use_alien_caches)) + if (likely(nodeid == node) || unlikely(!use_alien_caches)) return 0; l3 = cachep->nodelists[node]; @@ -2665,7 +2672,7 @@ static void *slab_get_obj(struct kmem_ca next = slab_bufctl(slabp)[slabp->free]; #if DEBUG slab_bufctl(slabp)[slabp->free] = BUFCTL_FREE; - WARN_ON(slabp->nodeid != nodeid); + WARN_ON(virt_to_nodeid(objp) != nodeid); #endif slabp->free = next; @@ -2679,7 +2686,7 @@ static void slab_put_obj(struct kmem_cac #if DEBUG /* Verify that the slab belongs to the intended node */ - WARN_ON(slabp->nodeid != nodeid); + WARN_ON(virt_to_nodeid(objp) != nodeid); if (slab_bufctl(slabp)[objnr] + 1 <= SLAB_LIMIT + 1) { printk(KERN_ERR "slab: double free detected in cache " - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |