ocfs2: Add dir_resv_level mount option

The default behavior for directory reservations stays the same, but we add a
mount option so people can tweak the size of directory reservations
according to their workloads.

Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

6 files changed, 40 insertions, 5 deletions

diff --git a/Documentation/filesystems/ocfs2.txt b/Documentation/filesystems/ocfs2.txt
index 32339e584a9a..1f7ae144f6d8 100644
--- a/Documentation/filesystems/ocfs2.txt
+++ b/Documentation/filesystems/ocfs2.txt
@@ -83,3 +83,7 @@ noacl		(*)	Disables POSIX Access Control Lists support.
 resv_level=2    (*)     Set how agressive allocation reservations will be.
                         Valid values are between 0 (reservations off) to 8
                         (maximum space for reservations).
+dir_resv_level= (*)     By default, directory reservations will scale with file
+                        reservations - users should rarely need to change this
+                        value. If allocation reservations are turned off, this
+                        option will have no effect.
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index 8563f97c58af..6c9a28a2d3ae 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -2977,7 +2977,8 @@ 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);
@@ -3348,7 +3349,8 @@ 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 {
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index 09d7aee3dabe..a388528f485c 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -356,6 +356,7 @@ 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. */
diff --git a/fs/ocfs2/reservations.c b/fs/ocfs2/reservations.c
index 87fa35791f18..6497bcc00fa5 100644
--- a/fs/ocfs2/reservations.c
+++ b/fs/ocfs2/reservations.c
@@ -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;
 }
diff --git a/fs/ocfs2/reservations.h b/fs/ocfs2/reservations.h
index 022aff601e15..25b0c0e31e91 100644
--- a/fs/ocfs2/reservations.h
+++ b/fs/ocfs2/reservations.h
@@ -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:
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 5745682eb1c0..79d7d4cf45b1 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -96,6 +96,7 @@ 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];
 };
 
@@ -178,6 +179,7 @@ enum {
         Opt_usrquota,
         Opt_grpquota,
         Opt_resv_level,
+        Opt_dir_resv_level,
         Opt_err,
 };
 
@@ -205,6 +207,7 @@ 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}
 };
 
@@ -1034,6 +1037,11 @@ 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)
@@ -1295,6 +1303,7 @@ 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;
@@ -1449,6 +1458,17 @@ 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\" "
@@ -1533,6 +1553,9 @@ 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;
 }