Waste less memory in dentries.

We don't bother aligining them on a cacheline boundary, since
that is totally excessive in some configurations (especially
P4's with 128-byte cachelines).

Instead, we make the minimum inline string size a bit longer,
and re-order a few fields that allow for better packing on
64-bit architectures, for better memory utilization.

fs/dcache.c

@@ -1558,14 +1558,11 @@ static void __init dcache_init(unsigned long mempages)
 	 * A constructor could be added for stable state like the lists,
 	 * but it is probably not worth it because of the cache nature
 	 * of the dcache. 
-	 * If fragmentation is too bad then the SLAB_HWCACHE_ALIGN
-	 * flag could be removed here, to hint to the allocator that
-	 * it should not try to get multiple page regions.  
 	 */
 	dentry_cache = kmem_cache_create("dentry_cache",
 					 sizeof(struct dentry),
 					 0,
-					 SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT,
+					 SLAB_RECLAIM_ACCOUNT,
 					 NULL, NULL);
 	if (!dentry_cache)
 		panic("Cannot create dentry cache");

include/linux/dcache.h

@@ -74,14 +74,14 @@ full_name_hash(const unsigned char *name, unsigned int len)
 	return end_name_hash(hash);
 }
 
-#define DNAME_INLINE_LEN_MIN 16
+#define DNAME_INLINE_LEN_MIN 24
 
 struct dcookie_struct;
  
 struct dentry {
 	atomic_t d_count;
-	unsigned long d_vfs_flags;	/* moved here to be on same cacheline */
 	spinlock_t d_lock;		/* per dentry lock */
+	unsigned long d_vfs_flags;	/* moved here to be on same cacheline */
 	struct inode  * d_inode;	/* Where the name belongs to - NULL is negative */
 	struct list_head d_lru;		/* LRU list */
 	struct list_head d_child;	/* child of parent list */
@@ -102,7 +102,7 @@ struct dentry {
 	struct hlist_node d_hash;	/* lookup hash list */	
 	struct hlist_head * d_bucket;	/* lookup hash bucket */
 	unsigned char d_iname[DNAME_INLINE_LEN_MIN]; /* small names */
-} ____cacheline_aligned;
+};
 
 #define DNAME_INLINE_LEN	(sizeof(struct dentry)-offsetof(struct dentry,d_iname))