Ocfs2 allocation tuning

index 18fc176..9dc34a6 100644
@@ -75,6 +75,34 @@ static int ocfs2_local_alloc_new_window(struct ocfs2_super *osb,
 static int ocfs2_local_alloc_slide_window(struct ocfs2_super *osb,
 					  struct inode *local_alloc_inode);
 
+void ocfs2_la_set_sizes(struct ocfs2_super *osb, int requested_mb)
+{
+	struct super_block *sb = osb->sb;
+	unsigned int la_default_mb = OCFS2_DEFAULT_LOCAL_ALLOC_SIZE;
+	unsigned int la_max_mb;
+
+	la_max_mb = ocfs2_clusters_to_megabytes(sb,
+						ocfs2_local_alloc_size(sb) * 8);
+
+	mlog(0, "requested: %dM, max: %uM, default: %uM\n",
+	     requested_mb, la_max_mb, la_default_mb);
+
+	if (requested_mb == -1) {
+		/* No user request - use defaults */
+		osb->local_alloc_default_bits =
+			ocfs2_megabytes_to_clusters(sb, la_default_mb);
+	} else if (requested_mb > la_max_mb) {
+		/* Request is too big, we give the maximum available */
+		osb->local_alloc_default_bits =
+			ocfs2_megabytes_to_clusters(sb, la_max_mb);
+	} else {
+		osb->local_alloc_default_bits =
+			ocfs2_megabytes_to_clusters(sb, requested_mb);
+	}
+
+	osb->local_alloc_bits = osb->local_alloc_default_bits;
+}
+
 static inline int ocfs2_la_state_enabled(struct ocfs2_super *osb)
 {
 	return (osb->local_alloc_state == OCFS2_LA_THROTTLED ||

index ac5ea9f..04195c6 100644
@@ -30,6 +30,8 @@ int ocfs2_load_local_alloc(struct ocfs2_super *osb);
 
 void ocfs2_shutdown_local_alloc(struct ocfs2_super *osb);
 
+void ocfs2_la_set_sizes(struct ocfs2_super *osb, int requested_mb);
+
 int ocfs2_begin_local_alloc_recovery(struct ocfs2_super *osb,
 				     int node_num,
 				     struct ocfs2_dinode **alloc_copy);

index 4a6f00e..44b95a9 100644
@@ -765,6 +765,12 @@ static inline unsigned int ocfs2_megabytes_to_clusters(struct super_block *sb,
 	return megs << (20 - OCFS2_SB(sb)->s_clustersize_bits);
 }
 
+static inline unsigned int ocfs2_clusters_to_megabytes(struct super_block *sb,
+						       unsigned int clusters)
+{
+	return clusters >> (20 - OCFS2_SB(sb)->s_clustersize_bits);
+}
+
 static inline void ocfs2_init_inode_steal_slot(struct ocfs2_super *osb)
 {
 	spin_lock(&osb->osb_lock);

index db354d1..6dfcb4e 100644
@@ -93,7 +93,7 @@ struct mount_options
 	unsigned long	mount_opt;
 	unsigned int	atime_quantum;
 	signed short	slot;
-	unsigned int	localalloc_opt;
+	int		localalloc_opt;
 	unsigned int	resv_level;
 	char		cluster_stack[OCFS2_STACK_LABEL_LEN + 1];
 };

@@ -1027,8 +1027,8 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	osb->s_atime_quantum = parsed_options.atime_quantum;
 	osb->preferred_slot = parsed_options.slot;
 	osb->osb_commit_interval = parsed_options.commit_interval;
-	osb->local_alloc_default_bits = ocfs2_megabytes_to_clusters(sb, parsed_options.localalloc_opt);
-	osb->local_alloc_bits = osb->local_alloc_default_bits;
+
+	ocfs2_la_set_sizes(osb, parsed_options.localalloc_opt);
 	osb->osb_resv_level = parsed_options.resv_level;
 
 	status = ocfs2_verify_userspace_stack(osb, &parsed_options);

@@ -1288,7 +1288,7 @@ static int ocfs2_parse_options(struct super_block *sb,
 	mopt->mount_opt = 0;
 	mopt->atime_quantum = OCFS2_DEFAULT_ATIME_QUANTUM;
 	mopt->slot = OCFS2_INVALID_SLOT;
-	mopt->localalloc_opt = OCFS2_DEFAULT_LOCAL_ALLOC_SIZE;
+	mopt->localalloc_opt = -1;
 	mopt->cluster_stack[0] = '\0';
 	mopt->resv_level = OCFS2_DEFAULT_RESV_LEVEL;
 

@@ -1381,7 +1381,7 @@ static int ocfs2_parse_options(struct super_block *sb,
 				status = 0;
 				goto bail;
 			}
-			if (option >= 0 && (option <= ocfs2_local_alloc_size(sb) * 8))
+			if (option >= 0)
 				mopt->localalloc_opt = option;
 			break;
 		case Opt_localflocks:

index 9dc34a6..9b70a02 100644
@@ -75,10 +75,120 @@ static int ocfs2_local_alloc_new_window(struct ocfs2_super *osb,
 static int ocfs2_local_alloc_slide_window(struct ocfs2_super *osb,
 					  struct inode *local_alloc_inode);
 
+/*
+ * ocfs2_la_default_mb() - determine a default size, in megabytes of
+ * the local alloc.
+ *
+ * Generally, we'd like to pick as large a local alloc as
+ * possible. Performance on large workloads tends to scale
+ * proportionally to la size. In addition to that, the reservations
+ * code functions more efficiently as it can reserve more windows for
+ * write.
+ *
+ * Some things work against us when trying to choose a large local alloc:
+ *
+ * - We need to ensure our sizing is picked to leave enough space in
+ *   group descriptors for other allocations (such as block groups,
+ *   etc). Picking default sizes which are a multiple of 4 could help
+ *   - block groups are allocated in 2mb and 4mb chunks.
+ *
+ * - Likewise, we don't want to starve other nodes of bits on small
+ *   file systems. This can easily be taken care of by limiting our
+ *   default to a reasonable size (256M) on larger cluster sizes.
+ *
+ * - Some file systems can't support very large sizes - 4k and 8k in
+ *   particular are limited to less than 128 and 256 megabytes respectively.
+ *
+ * The following reference table shows group descriptor and local
+ * alloc maximums at various cluster sizes (4k blocksize)
+ *
+ * csize: 4K	group: 126M	la: 121M
+ * csize: 8K	group: 252M	la: 243M
+ * csize: 16K	group: 504M	la: 486M
+ * csize: 32K	group: 1008M	la: 972M
+ * csize: 64K	group: 2016M	la: 1944M
+ * csize: 128K	group: 4032M	la: 3888M
+ * csize: 256K	group: 8064M	la: 7776M
+ * csize: 512K	group: 16128M	la: 15552M
+ * csize: 1024K	group: 32256M	la: 31104M
+ */
+#define	OCFS2_LA_MAX_DEFAULT_MB	256
+#define	OCFS2_LA_OLD_DEFAULT	8
+unsigned int ocfs2_la_default_mb(struct ocfs2_super *osb)
+{
+	unsigned int la_mb;
+	unsigned int gd_mb;
+	unsigned int megs_per_slot;
+	struct super_block *sb = osb->sb;
+
+	gd_mb = ocfs2_clusters_to_megabytes(osb->sb,
+					    8 * ocfs2_group_bitmap_size(sb));
+
+	/*
+	 * This takes care of files systems with very small group
+	 * descriptors - 512 byte blocksize at cluster sizes lower
+	 * than 16K and also 1k blocksize with 4k cluster size.
+	 */
+	if ((sb->s_blocksize == 512 && osb->s_clustersize <= 8192)
+	    || (sb->s_blocksize == 1024 && osb->s_clustersize == 4096))
+		return OCFS2_LA_OLD_DEFAULT;
+
+	/*
+	 * Leave enough room for some block groups and make the final
+	 * value we work from a multiple of 4.
+	 */
+	gd_mb -= 16;
+	gd_mb &= 0xFFFFFFFB;
+
+	la_mb = gd_mb;
+
+	/*
+	 * Keep window sizes down to a reasonable default
+	 */
+	if (la_mb > OCFS2_LA_MAX_DEFAULT_MB) {
+		/*
+		 * Some clustersize / blocksize combinations will have
+		 * given us a larger than OCFS2_LA_MAX_DEFAULT_MB
+		 * default size, but get poor distribution when
+		 * limited to exactly 256 megabytes.
+		 *
+		 * As an example, 16K clustersize at 4K blocksize
+		 * gives us a cluster group size of 504M. Paring the
+		 * local alloc size down to 256 however, would give us
+		 * only one window and around 200MB left in the
+		 * cluster group. Instead, find the first size below
+		 * 256 which would give us an even distribution.
+		 *
+		 * Larger cluster group sizes actually work out pretty
+		 * well when pared to 256, so we don't have to do this
+		 * for any group that fits more than two
+		 * OCFS2_LA_MAX_DEFAULT_MB windows.
+		 */
+		if (gd_mb > (2 * OCFS2_LA_MAX_DEFAULT_MB))
+			la_mb = 256;
+		else {
+			unsigned int gd_mult = gd_mb;
+
+			while (gd_mult > 256)
+				gd_mult = gd_mult >> 1;
+
+			la_mb = gd_mult;
+		}
+	}
+
+	megs_per_slot = osb->osb_clusters_at_boot / osb->max_slots;
+	megs_per_slot = ocfs2_clusters_to_megabytes(osb->sb, megs_per_slot);
+	/* Too many nodes, too few disk clusters. */
+	if (megs_per_slot < la_mb)
+		la_mb = megs_per_slot;
+
+	return la_mb;
+}
+
 void ocfs2_la_set_sizes(struct ocfs2_super *osb, int requested_mb)
 {
 	struct super_block *sb = osb->sb;
-	unsigned int la_default_mb = OCFS2_DEFAULT_LOCAL_ALLOC_SIZE;
+	unsigned int la_default_mb = ocfs2_la_default_mb(osb);
 	unsigned int la_max_mb;
 
 	la_max_mb = ocfs2_clusters_to_megabytes(sb,

@@ -185,7 +295,7 @@ int ocfs2_load_local_alloc(struct ocfs2_super *osb)
 		     osb->local_alloc_bits, (osb->bitmap_cpg - 1));
 		osb->local_alloc_bits =
 			ocfs2_megabytes_to_clusters(osb->sb,
-						    OCFS2_DEFAULT_LOCAL_ALLOC_SIZE);
+						    ocfs2_la_default_mb(osb));
 	}
 
 	/* read the alloc off disk */

index 04195c6..1be9b58 100644
@@ -31,6 +31,7 @@ int ocfs2_load_local_alloc(struct ocfs2_super *osb);
 void ocfs2_shutdown_local_alloc(struct ocfs2_super *osb);
 
 void ocfs2_la_set_sizes(struct ocfs2_super *osb, int requested_mb);
+unsigned int ocfs2_la_default_mb(struct ocfs2_super *osb);
 
 int ocfs2_begin_local_alloc_recovery(struct ocfs2_super *osb,
 				     int node_num,

index 44b95a9..72e6eef 100644
@@ -339,6 +339,9 @@ struct ocfs2_super
 	 */
 	unsigned int local_alloc_bits;
 	unsigned int local_alloc_default_bits;
+	/* osb_clusters_at_boot can become stale! Do not trust it to
+	 * be up to date. */
+	unsigned int osb_clusters_at_boot;
 
 	enum ocfs2_local_alloc_state local_alloc_state; /* protected
 							 * by osb_lock */

index 7638a38..d58bc76 100644
@@ -340,14 +340,6 @@ struct reflink_arguments {
 #define OCFS2_MIN_JOURNAL_SIZE		(4 * 1024 * 1024)
 
 /*
- * Default local alloc size (in megabytes)
- *
- * The value chosen should be such that most allocations, including new
- * block groups, use local alloc.
- */
-#define OCFS2_DEFAULT_LOCAL_ALLOC_SIZE	8
-
-/*
  * Inline extended attribute size (in bytes)
  * The value chosen should be aligned to 16 byte boundaries.
  */

index 6dfcb4e..bca1ed3 100644
@@ -1499,7 +1499,7 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
 			   (unsigned) (osb->osb_commit_interval / HZ));
 
 	local_alloc_megs = osb->local_alloc_bits >> (20 - osb->s_clustersize_bits);
-	if (local_alloc_megs != OCFS2_DEFAULT_LOCAL_ALLOC_SIZE)
+	if (local_alloc_megs != ocfs2_la_default_mb(osb))
 		seq_printf(s, ",localalloc=%d", local_alloc_megs);
 
 	if (opts & OCFS2_MOUNT_LOCALFLOCKS)

@@ -2247,6 +2247,7 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	}
 
 	osb->bitmap_blkno = OCFS2_I(inode)->ip_blkno;
+	osb->osb_clusters_at_boot = OCFS2_I(inode)->ip_clusters;
 	iput(inode);
 
 	osb->bitmap_cpg = ocfs2_group_bitmap_size(sb) * 8;

index 7fc6cfe..87fa357 100644
@@ -55,9 +55,10 @@ static unsigned int ocfs2_resv_window_bits(struct ocfs2_reservation_map *resmap,
 	if (!(resv->r_flags & OCFS2_RESV_FLAG_DIR)) {
 		/* 8, 16, 32, 64, 128, 256, 512, 1024 */
 		bits = 4 << osb->osb_resv_level;
-	} else
-		bits = OCFS2_RESV_DIR_WINDOW_BITS;
-
+	} else {
+		/* For now, treat directories the same as files. */
+		bits = 4 << osb->osb_resv_level;
+	}
 	return bits;
 }
 

index 34bb308..022aff6 100644
@@ -22,7 +22,7 @@
 
 #include <linux/rbtree.h>
 
-#define OCFS2_DEFAULT_RESV_LEVEL	4
+#define OCFS2_DEFAULT_RESV_LEVEL	2
 #define OCFS2_MAX_RESV_LEVEL	9
 #define OCFS2_MIN_RESV_LEVEL	0
 

index b71acf3..dce39da 100644
@@ -2993,7 +2993,8 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	 * if we only get one now, that's enough to continue. The rest
 	 * will be claimed after the conversion to extents.
 	 */
-	data_ac->ac_resv = &oi->ip_la_data_resv;
+	if (ocfs2_dir_resv_allowed(osb))
+		data_ac->ac_resv = &oi->ip_la_data_resv;
 	ret = ocfs2_claim_clusters(osb, handle, data_ac, 1, &bit_off, &len);
 	if (ret) {
 		mlog_errno(ret);

@@ -3373,7 +3374,8 @@ static int ocfs2_extend_dir(struct ocfs2_super *osb,
 			goto bail;
 		}
 
-		data_ac->ac_resv = &OCFS2_I(dir)->ip_la_data_resv;
+		if (ocfs2_dir_resv_allowed(osb))
+			data_ac->ac_resv = &OCFS2_I(dir)->ip_la_data_resv;
 
 		credits = ocfs2_calc_extend_credits(sb, el, 1);
 	} else {

index 72e6eef..e946936 100644
@@ -353,6 +353,7 @@ struct ocfs2_super
 	struct ocfs2_reservation_map	osb_la_resmap;
 
 	unsigned int	osb_resv_level;
+	unsigned int	osb_dir_resv_level;
 
 	/* Next three fields are for local node slot recovery during
 	 * mount. */

index 87fa357..6497bcc 100644
@@ -44,7 +44,11 @@ DEFINE_SPINLOCK(resv_lock);
 
 #define	OCFS2_MIN_RESV_WINDOW_BITS	8
 #define	OCFS2_MAX_RESV_WINDOW_BITS	1024
-#define	OCFS2_RESV_DIR_WINDOW_BITS	OCFS2_MIN_RESV_WINDOW_BITS
+
+int ocfs2_dir_resv_allowed(struct ocfs2_super *osb)
+{
+	return (osb->osb_resv_level && osb->osb_dir_resv_level);
+}
 
 static unsigned int ocfs2_resv_window_bits(struct ocfs2_reservation_map *resmap,
 					   struct ocfs2_alloc_reservation *resv)

@@ -56,8 +60,7 @@ static unsigned int ocfs2_resv_window_bits(struct ocfs2_reservation_map *resmap,
 		/* 8, 16, 32, 64, 128, 256, 512, 1024 */
 		bits = 4 << osb->osb_resv_level;
 	} else {
-		/* For now, treat directories the same as files. */
-		bits = 4 << osb->osb_resv_level;
+		bits = 4 << osb->osb_dir_resv_level;
 	}
 	return bits;
 }

index 022aff6..25b0c0e 100644
@@ -67,6 +67,8 @@ void ocfs2_resv_init_once(struct ocfs2_alloc_reservation *resv);
 void ocfs2_resv_set_type(struct ocfs2_alloc_reservation *resv,
 			 unsigned int flags);
 
+int ocfs2_dir_resv_allowed(struct ocfs2_super *osb);
+
 /**
  * ocfs2_resv_discard() - truncate a reservation
  * @resmap:

index bca1ed3..9f0e211 100644
@@ -95,6 +95,7 @@ struct mount_options
 	signed short	slot;
 	int		localalloc_opt;
 	unsigned int	resv_level;
+	int		dir_resv_level;
 	char		cluster_stack[OCFS2_STACK_LABEL_LEN + 1];
 };
 

@@ -177,6 +178,7 @@ enum {
 	Opt_usrquota,
 	Opt_grpquota,
 	Opt_resv_level,
+	Opt_dir_resv_level,
 	Opt_err,
 };
 

@@ -204,6 +206,7 @@ static const match_table_t tokens = {
 	{Opt_usrquota, "usrquota"},
 	{Opt_grpquota, "grpquota"},
 	{Opt_resv_level, "resv_level=%u"},
+	{Opt_dir_resv_level, "dir_resv_level=%u"},
 	{Opt_err, NULL}
 };
 

@@ -1030,6 +1033,11 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 
 	ocfs2_la_set_sizes(osb, parsed_options.localalloc_opt);
 	osb->osb_resv_level = parsed_options.resv_level;
+	osb->osb_dir_resv_level = parsed_options.resv_level;
+	if (parsed_options.dir_resv_level == -1)
+		osb->osb_dir_resv_level = parsed_options.resv_level;
+	else
+		osb->osb_dir_resv_level = parsed_options.dir_resv_level;
 
 	status = ocfs2_verify_userspace_stack(osb, &parsed_options);
 	if (status)

@@ -1291,6 +1299,7 @@ static int ocfs2_parse_options(struct super_block *sb,
 	mopt->localalloc_opt = -1;
 	mopt->cluster_stack[0] = '\0';
 	mopt->resv_level = OCFS2_DEFAULT_RESV_LEVEL;
+	mopt->dir_resv_level = -1;
 
 	if (!options) {
 		status = 1;

@@ -1445,6 +1454,17 @@ static int ocfs2_parse_options(struct super_block *sb,
 			    option < OCFS2_MAX_RESV_LEVEL)
 				mopt->resv_level = option;
 			break;
+		case Opt_dir_resv_level:
+			if (is_remount)
+				break;
+			if (match_int(&args[0], &option)) {
+				status = 0;
+				goto bail;
+			}
+			if (option >= OCFS2_MIN_RESV_LEVEL &&
+			    option < OCFS2_MAX_RESV_LEVEL)
+				mopt->dir_resv_level = option;
+			break;
 		default:
 			mlog(ML_ERROR,
 			     "Unrecognized mount option \"%s\" "

@@ -1529,6 +1549,9 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
 	if (osb->osb_resv_level != OCFS2_DEFAULT_RESV_LEVEL)
 		seq_printf(s, ",resv_level=%d", osb->osb_resv_level);
 
+	if (osb->osb_dir_resv_level != osb->osb_resv_level)
+		seq_printf(s, ",dir_resv_level=%d", osb->osb_resv_level);
+
 	return 0;
 }