Merge git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-nmw

Pull GFS2 updates from Steven Whitehouse:
 "The major feature this time is the "rbm" conversion in the resource
  group code.  The new struct gfs2_rbm specifies the location of an
  allocatable block in (resource group, bitmap, offset) form.  There are
  a number of added helper functions, and later patches then rewrite
  some of the resource group code in terms of this new structure.  Not
  only does this give us a nice code clean up, but it also removes some
  of the previous restrictions where extents could not cross bitmap
  boundaries, for example.

  In addition to that, there are a few bug fixes and clean ups, but the
  rbm work is by far the majority of this patch set in terms of number
  of changed lines."

* git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-nmw: (27 commits)
  GFS2: Write out dirty inode metadata in delayed deletes
  GFS2: fix s_writers.counter imbalance in gfs2_ail_empty_gl
  GFS2: Fix infinite loop in rbm_find
  GFS2: Consolidate free block searching functions
  GFS2: Get rid of I_MUTEX_QUOTA usage
  GFS2: Stop block extents at the end of bitmaps
  GFS2: Fix unclaimed_blocks() wrapping bug and clean up
  GFS2: Improve block reservation tracing
  GFS2: Fall back to ignoring reservations, if there are no other blocks left
  GFS2: Fix ->show_options() for statfs slow
  GFS2: Use rbm for gfs2_setbit()
  GFS2: Use rbm for gfs2_testbit()
  GFS2: Eliminate unnecessary check for state > 3 in bitfit
  GFS2: Eliminate redundant calls to may_grant
  GFS2: Combine functions gfs2_glock_dq_wait and wait_on_demote
  GFS2: Combine functions gfs2_glock_wait and wait_on_holder
  GFS2: inline __gfs2_glock_schedule_for_reclaim
  GFS2: change function gfs2_direct_IO to use a normal gfs2_glock_dq
  GFS2: rbm code cleanup
  GFS2: Fix case where reservation finished at end of rgrp
  ...

15 files changed, 710 insertions, 800 deletions

diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index d6526347d386..01c4975da4bc 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -612,6 +612,7 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 	struct gfs2_sbd *sdp = GFS2_SB(mapping->host);
 	struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
 	unsigned int data_blocks = 0, ind_blocks = 0, rblocks;
+	unsigned requested = 0;
 	int alloc_required;
 	int error = 0;
 	pgoff_t index = pos >> PAGE_CACHE_SHIFT;
@@ -641,7 +642,8 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 		if (error)
 			goto out_unlock;
 
-		error = gfs2_inplace_reserve(ip, data_blocks + ind_blocks);
+		requested = data_blocks + ind_blocks;
+		error = gfs2_inplace_reserve(ip, requested);
 		if (error)
 			goto out_qunlock;
 	}
@@ -654,7 +656,7 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 	if (&ip->i_inode == sdp->sd_rindex)
 		rblocks += 2 * RES_STATFS;
 	if (alloc_required)
-		rblocks += gfs2_rg_blocks(ip);
+		rblocks += gfs2_rg_blocks(ip, requested);
 
 	error = gfs2_trans_begin(sdp, rblocks,
 				 PAGE_CACHE_SIZE/sdp->sd_sb.sb_bsize);
@@ -868,8 +870,7 @@ static int gfs2_write_end(struct file *file, struct address_space *mapping,
 	brelse(dibh);
 failed:
 	gfs2_trans_end(sdp);
-	if (gfs2_mb_reserved(ip))
-		gfs2_inplace_release(ip);
+	gfs2_inplace_release(ip);
 	if (ip->i_res->rs_qa_qd_num)
 		gfs2_quota_unlock(ip);
 	if (inode == sdp->sd_rindex) {
@@ -1023,7 +1024,7 @@ static ssize_t gfs2_direct_IO(int rw, struct kiocb *iocb,
 				  offset, nr_segs, gfs2_get_block_direct,
 				  NULL, NULL, 0);
 out:
-	gfs2_glock_dq_m(1, &gh);
+	gfs2_glock_dq(&gh);
 	gfs2_holder_uninit(&gh);
 	return rv;
 }
diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index 49cd7dd4a9fa..1fd3ae237bdd 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -786,7 +786,7 @@ static int do_strip(struct gfs2_inode *ip, struct buffer_head *dibh,
 		goto out_rlist;
 
 	if (gfs2_rs_active(ip->i_res)) /* needs to be done with the rgrp glock held */
-		gfs2_rs_deltree(ip->i_res);
+		gfs2_rs_deltree(ip, ip->i_res);
 
 	error = gfs2_trans_begin(sdp, rg_blocks + RES_DINODE +
 				 RES_INDIRECT + RES_STATFS + RES_QUOTA,
diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index 382000ffac1f..30e21997a1a1 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -441,7 +441,7 @@ static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 		rblocks += data_blocks ? data_blocks : 1;
 	if (ind_blocks || data_blocks) {
 		rblocks += RES_STATFS + RES_QUOTA;
-		rblocks += gfs2_rg_blocks(ip);
+		rblocks += gfs2_rg_blocks(ip, data_blocks + ind_blocks);
 	}
 	ret = gfs2_trans_begin(sdp, rblocks, 0);
 	if (ret)
@@ -845,7 +845,7 @@ retry:
 				&max_bytes, &data_blocks, &ind_blocks);
 
 		rblocks = RES_DINODE + ind_blocks + RES_STATFS + RES_QUOTA +
-			  RES_RG_HDR + gfs2_rg_blocks(ip);
+			  RES_RG_HDR + gfs2_rg_blocks(ip, data_blocks + ind_blocks);
 		if (gfs2_is_jdata(ip))
 			rblocks += data_blocks ? data_blocks : 1;
 
diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 1ed81f40da0d..e6c2fd53cab2 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -186,20 +186,6 @@ static void gfs2_glock_remove_from_lru(struct gfs2_glock *gl)
 }
 
 /**
- * __gfs2_glock_schedule_for_reclaim - Add a glock to the reclaim list
- * @gl: the glock
- *
- * If the glock is demotable, then we add it (or move it) to the end
- * of the glock LRU list.
- */
-
-static void __gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
-{
-	if (demote_ok(gl))
-		gfs2_glock_add_to_lru(gl);
-}
-
-/**
  * gfs2_glock_put_nolock() - Decrement reference count on glock
  * @gl: The glock to put
  *
@@ -883,7 +869,14 @@ static int gfs2_glock_demote_wait(void *word)
 	return 0;
 }
 
-static void wait_on_holder(struct gfs2_holder *gh)
+/**
+ * gfs2_glock_wait - wait on a glock acquisition
+ * @gh: the glock holder
+ *
+ * Returns: 0 on success
+ */
+
+int gfs2_glock_wait(struct gfs2_holder *gh)
 {
 	unsigned long time1 = jiffies;
 
@@ -894,12 +887,7 @@ static void wait_on_holder(struct gfs2_holder *gh)
 		gh->gh_gl->gl_hold_time = min(gh->gh_gl->gl_hold_time +
 					      GL_GLOCK_HOLD_INCR,
 					      GL_GLOCK_MAX_HOLD);
-}
-
-static void wait_on_demote(struct gfs2_glock *gl)
-{
-	might_sleep();
-	wait_on_bit(&gl->gl_flags, GLF_DEMOTE, gfs2_glock_demote_wait, TASK_UNINTERRUPTIBLE);
+	return gh->gh_error;
 }
 
 /**
@@ -929,19 +917,6 @@ static void handle_callback(struct gfs2_glock *gl, unsigned int state,
 	trace_gfs2_demote_rq(gl);
 }
 
-/**
- * gfs2_glock_wait - wait on a glock acquisition
- * @gh: the glock holder
- *
- * Returns: 0 on success
- */
-
-int gfs2_glock_wait(struct gfs2_holder *gh)
-{
-	wait_on_holder(gh);
-	return gh->gh_error;
-}
-
 void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...)
 {
 	struct va_format vaf;
@@ -979,7 +954,7 @@ __acquires(&gl->gl_spin)
 	struct gfs2_sbd *sdp = gl->gl_sbd;
 	struct list_head *insert_pt = NULL;
 	struct gfs2_holder *gh2;
-	int try_lock = 0;
+	int try_futile = 0;
 
 	BUG_ON(gh->gh_owner_pid == NULL);
 	if (test_and_set_bit(HIF_WAIT, &gh->gh_iflags))
@@ -987,7 +962,7 @@ __acquires(&gl->gl_spin)
 
 	if (gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)) {
 		if (test_bit(GLF_LOCK, &gl->gl_flags))
-			try_lock = 1;
+			try_futile = !may_grant(gl, gh);
 		if (test_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags))
 			goto fail;
 	}
@@ -996,9 +971,8 @@ __acquires(&gl->gl_spin)
 		if (unlikely(gh2->gh_owner_pid == gh->gh_owner_pid &&
 		    (gh->gh_gl->gl_ops->go_type != LM_TYPE_FLOCK)))
 			goto trap_recursive;
-		if (try_lock &&
-		    !(gh2->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)) &&
-		    !may_grant(gl, gh)) {
+		if (try_futile &&
+		    !(gh2->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB))) {
 fail:
 			gh->gh_error = GLR_TRYFAILED;
 			gfs2_holder_wake(gh);
@@ -1121,8 +1095,9 @@ void gfs2_glock_dq(struct gfs2_holder *gh)
 		    !test_bit(GLF_DEMOTE, &gl->gl_flags))
 			fast_path = 1;
 	}
-	if (!test_bit(GLF_LFLUSH, &gl->gl_flags))
-		__gfs2_glock_schedule_for_reclaim(gl);
+	if (!test_bit(GLF_LFLUSH, &gl->gl_flags) && demote_ok(gl))
+		gfs2_glock_add_to_lru(gl);
+
 	trace_gfs2_glock_queue(gh, 0);
 	spin_unlock(&gl->gl_spin);
 	if (likely(fast_path))
@@ -1141,7 +1116,8 @@ void gfs2_glock_dq_wait(struct gfs2_holder *gh)
 {
 	struct gfs2_glock *gl = gh->gh_gl;
 	gfs2_glock_dq(gh);
-	wait_on_demote(gl);
+	might_sleep();
+	wait_on_bit(&gl->gl_flags, GLF_DEMOTE, gfs2_glock_demote_wait, TASK_UNINTERRUPTIBLE);
 }
 
 /**
diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index 4bdcf3784187..32cc4fde975c 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -94,6 +94,7 @@ static void gfs2_ail_empty_gl(struct gfs2_glock *gl)
 	/* A shortened, inline version of gfs2_trans_begin() */
 	tr.tr_reserved = 1 + gfs2_struct2blk(sdp, tr.tr_revokes, sizeof(u64));
 	tr.tr_ip = (unsigned long)__builtin_return_address(0);
+	sb_start_intwrite(sdp->sd_vfs);
 	gfs2_log_reserve(sdp, tr.tr_reserved);
 	BUG_ON(current->journal_info);
 	current->journal_info = &tr;
diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index aaecc8085fc5..3d469d37345e 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -99,9 +99,26 @@ struct gfs2_rgrpd {
 #define GFS2_RDF_MASK		0xf0000000 /* mask for internal flags */
 	spinlock_t rd_rsspin;           /* protects reservation related vars */
 	struct rb_root rd_rstree;       /* multi-block reservation tree */
-	u32 rd_rs_cnt;                  /* count of current reservations */
 };
 
+struct gfs2_rbm {
+	struct gfs2_rgrpd *rgd;
+	struct gfs2_bitmap *bi;	/* Bitmap must belong to the rgd */
+	u32 offset;		/* The offset is bitmap relative */
+};
+
+static inline u64 gfs2_rbm_to_block(const struct gfs2_rbm *rbm)
+{
+	return rbm->rgd->rd_data0 + (rbm->bi->bi_start * GFS2_NBBY) + rbm->offset;
+}
+
+static inline bool gfs2_rbm_eq(const struct gfs2_rbm *rbm1,
+			       const struct gfs2_rbm *rbm2)
+{
+	return (rbm1->rgd == rbm2->rgd) && (rbm1->bi == rbm2->bi) && 
+	       (rbm1->offset == rbm2->offset);
+}
+
 enum gfs2_state_bits {
 	BH_Pinned = BH_PrivateStart,
 	BH_Escaped = BH_PrivateStart + 1,
@@ -250,18 +267,11 @@ struct gfs2_blkreserv {
 	/* components used during write (step 1): */
 	atomic_t rs_sizehint;         /* hint of the write size */
 
-	/* components used during inplace_reserve (step 2): */
-	u32 rs_requested; /* Filled in by caller of gfs2_inplace_reserve() */
-
-	/* components used during get_local_rgrp (step 3): */
-	struct gfs2_rgrpd *rs_rgd;    /* pointer to the gfs2_rgrpd */
 	struct gfs2_holder rs_rgd_gh; /* Filled in by get_local_rgrp */
 	struct rb_node rs_node;       /* link to other block reservations */
-
-	/* components used during block searches and assignments (step 4): */
-	struct gfs2_bitmap *rs_bi;    /* bitmap for the current allocation */
-	u32 rs_biblk;                 /* start block relative to the bi */
+	struct gfs2_rbm rs_rbm;       /* Start of reservation */
 	u32 rs_free;                  /* how many blocks are still free */
+	u64 rs_inum;                  /* Inode number for reservation */
 
 	/* ancillary quota stuff */
 	struct gfs2_quota_data *rs_qa_qd[2 * MAXQUOTAS];
diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 753af3d86bbc..381893ceefa4 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -712,14 +712,9 @@ static int gfs2_create_inode(struct inode *dir, struct dentry *dentry,
 	if (error)
 		goto fail_gunlock2;
 
-	/* The newly created inode needs a reservation so it can allocate
-	   xattrs. At the same time, we want new blocks allocated to the new
-	   dinode to be as contiguous as possible. Since we allocated the
-	   dinode block under the directory's reservation, we transfer
-	   ownership of that reservation to the new inode. The directory
-	   doesn't need a reservation unless it needs a new allocation. */
-	ip->i_res = dip->i_res;
-	dip->i_res = NULL;
+	error = gfs2_rs_alloc(ip);
+	if (error)
+		goto fail_gunlock2;
 
 	error = gfs2_acl_create(dip, inode);
 	if (error)
@@ -737,10 +732,7 @@ static int gfs2_create_inode(struct inode *dir, struct dentry *dentry,
 		brelse(bh);
 
 	gfs2_trans_end(sdp);
-	/* Check if we reserved space in the rgrp. Function link_dinode may
-	   not, depending on whether alloc is required. */
-	if (gfs2_mb_reserved(dip))
-		gfs2_inplace_release(dip);
+	gfs2_inplace_release(dip);
 	gfs2_quota_unlock(dip);
 	mark_inode_dirty(inode);
 	gfs2_glock_dq_uninit_m(2, ghs);
@@ -897,7 +889,7 @@ static int gfs2_link(struct dentry *old_dentry, struct inode *dir,
 			goto out_gunlock_q;
 
 		error = gfs2_trans_begin(sdp, sdp->sd_max_dirres +
-					 gfs2_rg_blocks(dip) +
+					 gfs2_rg_blocks(dip, sdp->sd_max_dirres) +
 					 2 * RES_DINODE + RES_STATFS +
 					 RES_QUOTA, 0);
 		if (error)
@@ -1378,7 +1370,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry,
 			goto out_gunlock_q;
 
 		error = gfs2_trans_begin(sdp, sdp->sd_max_dirres +
-					 gfs2_rg_blocks(ndip) +
+					 gfs2_rg_blocks(ndip, sdp->sd_max_dirres) +
 					 4 * RES_DINODE + 4 * RES_LEAF +
 					 RES_STATFS + RES_QUOTA + 4, 0);
 		if (error)
diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index e5af9dc420ef..e443966c8106 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -19,6 +19,7 @@
 #include <linux/mount.h>
 #include <linux/gfs2_ondisk.h>
 #include <linux/quotaops.h>
+#include <linux/lockdep.h>
 
 #include "gfs2.h"
 #include "incore.h"
@@ -766,6 +767,7 @@ fail:
 	return error;
 }
 
+static struct lock_class_key gfs2_quota_imutex_key;
 
 static int init_inodes(struct gfs2_sbd *sdp, int undo)
 {
@@ -803,6 +805,12 @@ static int init_inodes(struct gfs2_sbd *sdp, int undo)
 		fs_err(sdp, "can't get quota file inode: %d\n", error);
 		goto fail_rindex;
 	}
+	/*
+	 * i_mutex on quota files is special. Since this inode is hidden system
+	 * file, we are safe to define locking ourselves.
+	 */
+	lockdep_set_class(&sdp->sd_quota_inode->i_mutex,
+			  &gfs2_quota_imutex_key);
 
 	error = gfs2_rindex_update(sdp);
 	if (error)
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index a3bde91645c2..4021deca61ef 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -765,6 +765,7 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda)
 	struct gfs2_holder *ghs, i_gh;
 	unsigned int qx, x;
 	struct gfs2_quota_data *qd;
+	unsigned reserved;
 	loff_t offset;
 	unsigned int nalloc = 0, blocks;
 	int error;
@@ -781,7 +782,7 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda)
 		return -ENOMEM;
 
 	sort(qda, num_qd, sizeof(struct gfs2_quota_data *), sort_qd, NULL);
-	mutex_lock_nested(&ip->i_inode.i_mutex, I_MUTEX_QUOTA);
+	mutex_lock(&ip->i_inode.i_mutex);
 	for (qx = 0; qx < num_qd; qx++) {
 		error = gfs2_glock_nq_init(qda[qx]->qd_gl, LM_ST_EXCLUSIVE,
 					   GL_NOCACHE, &ghs[qx]);
@@ -811,13 +812,13 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda)
 	 * two blocks need to be updated instead of 1 */
 	blocks = num_qd * data_blocks + RES_DINODE + num_qd + 3;
 
-	error = gfs2_inplace_reserve(ip, 1 +
-				     (nalloc * (data_blocks + ind_blocks)));
+	reserved = 1 + (nalloc * (data_blocks + ind_blocks));
+	error = gfs2_inplace_reserve(ip, reserved);
 	if (error)
 		goto out_alloc;
 
 	if (nalloc)
-		blocks += gfs2_rg_blocks(ip) + nalloc * ind_blocks + RES_STATFS;
+		blocks += gfs2_rg_blocks(ip, reserved) + nalloc * ind_blocks + RES_STATFS;
 
 	error = gfs2_trans_begin(sdp, blocks, 0);
 	if (error)
@@ -1598,7 +1599,7 @@ static int gfs2_set_dqblk(struct super_block *sb, int type, qid_t id,
 		error = gfs2_inplace_reserve(ip, blocks);
 		if (error)
 			goto out_i;
-		blocks += gfs2_rg_blocks(ip);
+		blocks += gfs2_rg_blocks(ip, blocks);
 	}
 
 	/* Some quotas span block boundaries and can update two blocks,
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index c9ed814eeb6f..3cc402ce6fea 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -35,9 +35,6 @@
 #define BFITNOENT ((u32)~0)
 #define NO_BLOCK ((u64)~0)
 
-#define RSRV_CONTENTION_FACTOR 4
-#define RGRP_RSRV_MAX_CONTENDERS 2
-
 #if BITS_PER_LONG == 32
 #define LBITMASK   (0x55555555UL)
 #define LBITSKIP55 (0x55555555UL)
@@ -67,53 +64,48 @@ static const char valid_change[16] = {
 	        1, 0, 0, 0
 };
 
-static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal,
-			unsigned char old_state,
-			struct gfs2_bitmap **rbi);
+static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 minext,
+                         const struct gfs2_inode *ip, bool nowrap);
+
 
 /**
  * gfs2_setbit - Set a bit in the bitmaps
- * @rgd: the resource group descriptor
- * @buf2: the clone buffer that holds the bitmaps
- * @bi: the bitmap structure
- * @block: the block to set
+ * @rbm: The position of the bit to set
+ * @do_clone: Also set the clone bitmap, if it exists
  * @new_state: the new state of the block
  *
  */
 
-static inline void gfs2_setbit(struct gfs2_rgrpd *rgd, unsigned char *buf2,
-			       struct gfs2_bitmap *bi, u32 block,
+static inline void gfs2_setbit(const struct gfs2_rbm *rbm, bool do_clone,
 			       unsigned char new_state)
 {
 	unsigned char *byte1, *byte2, *end, cur_state;
-	unsigned int buflen = bi->bi_len;
-	const unsigned int bit = (block % GFS2_NBBY) * GFS2_BIT_SIZE;
+	unsigned int buflen = rbm->bi->bi_len;
+	const unsigned int bit = (rbm->offset % GFS2_NBBY) * GFS2_BIT_SIZE;
 
-	byte1 = bi->bi_bh->b_data + bi->bi_offset + (block / GFS2_NBBY);
-	end = bi->bi_bh->b_data + bi->bi_offset + buflen;
+	byte1 = rbm->bi->bi_bh->b_data + rbm->bi->bi_offset + (rbm->offset / GFS2_NBBY);
+	end = rbm->bi->bi_bh->b_data + rbm->bi->bi_offset + buflen;
 
 	BUG_ON(byte1 >= end);
 
 	cur_state = (*byte1 >> bit) & GFS2_BIT_MASK;
 
 	if (unlikely(!valid_change[new_state * 4 + cur_state])) {
-		printk(KERN_WARNING "GFS2: buf_blk = 0x%llx old_state=%d, "
-		       "new_state=%d\n",
-		       (unsigned long long)block, cur_state, new_state);
-		printk(KERN_WARNING "GFS2: rgrp=0x%llx bi_start=0x%lx\n",
-		       (unsigned long long)rgd->rd_addr,
-		       (unsigned long)bi->bi_start);
-		printk(KERN_WARNING "GFS2: bi_offset=0x%lx bi_len=0x%lx\n",
-		       (unsigned long)bi->bi_offset,
-		       (unsigned long)bi->bi_len);
+		printk(KERN_WARNING "GFS2: buf_blk = 0x%x old_state=%d, "
+		       "new_state=%d\n", rbm->offset, cur_state, new_state);
+		printk(KERN_WARNING "GFS2: rgrp=0x%llx bi_start=0x%x\n",
+		       (unsigned long long)rbm->rgd->rd_addr,
+		       rbm->bi->bi_start);
+		printk(KERN_WARNING "GFS2: bi_offset=0x%x bi_len=0x%x\n",
+		       rbm->bi->bi_offset, rbm->bi->bi_len);
 		dump_stack();
-		gfs2_consist_rgrpd(rgd);
+		gfs2_consist_rgrpd(rbm->rgd);
 		return;
 	}
 	*byte1 ^= (cur_state ^ new_state) << bit;
 
-	if (buf2) {
-		byte2 = buf2 + bi->bi_offset + (block / GFS2_NBBY);
+	if (do_clone && rbm->bi->bi_clone) {
+		byte2 = rbm->bi->bi_clone + rbm->bi->bi_offset + (rbm->offset / GFS2_NBBY);
 		cur_state = (*byte2 >> bit) & GFS2_BIT_MASK;
 		*byte2 ^= (cur_state ^ new_state) << bit;
 	}
@@ -121,30 +113,21 @@ static inline void gfs2_setbit(struct gfs2_rgrpd *rgd, unsigned char *buf2,
 
 /**
  * gfs2_testbit - test a bit in the bitmaps
- * @rgd: the resource group descriptor
- * @buffer: the buffer that holds the bitmaps
- * @buflen: the length (in bytes) of the buffer
- * @block: the block to read
+ * @rbm: The bit to test
  *
+ * Returns: The two bit block state of the requested bit
  */
 
-static inline unsigned char gfs2_testbit(struct gfs2_rgrpd *rgd,
-					 const unsigned char *buffer,
-					 unsigned int buflen, u32 block)
+static inline u8 gfs2_testbit(const struct gfs2_rbm *rbm)
 {
-	const unsigned char *byte, *end;
-	unsigned char cur_state;
+	const u8 *buffer = rbm->bi->bi_bh->b_data + rbm->bi->bi_offset;
+	const u8 *byte;
 	unsigned int bit;
 
-	byte = buffer + (block / GFS2_NBBY);
-	bit = (block % GFS2_NBBY) * GFS2_BIT_SIZE;
-	end = buffer + buflen;
-
-	gfs2_assert(rgd->rd_sbd, byte < end);
+	byte = buffer + (rbm->offset / GFS2_NBBY);
+	bit = (rbm->offset % GFS2_NBBY) * GFS2_BIT_SIZE;
 
-	cur_state = (*byte >> bit) & GFS2_BIT_MASK;
-
-	return cur_state;
+	return (*byte >> bit) & GFS2_BIT_MASK;
 }
 
 /**
@@ -192,7 +175,7 @@ static inline u64 gfs2_bit_search(const __le64 *ptr, u64 mask, u8 state)
  */
 static inline int rs_cmp(u64 blk, u32 len, struct gfs2_blkreserv *rs)
 {
-	u64 startblk = gfs2_rs_startblk(rs);
+	u64 startblk = gfs2_rbm_to_block(&rs->rs_rbm);
 
 	if (blk >= startblk + rs->rs_free)
 		return 1;
@@ -202,36 +185,6 @@ static inline int rs_cmp(u64 blk, u32 len, struct gfs2_blkreserv *rs)
 }
 
 /**
- * rs_find - Find a rgrp multi-block reservation that contains a given block
- * @rgd: The rgrp
- * @rgblk: The block we're looking for, relative to the rgrp
- */
-static struct gfs2_blkreserv *rs_find(struct gfs2_rgrpd *rgd, u32 rgblk)
-{
-	struct rb_node **newn;
-	int rc;
-	u64 fsblk = rgblk + rgd->rd_data0;
-
-	spin_lock(&rgd->rd_rsspin);
-	newn = &rgd->rd_rstree.rb_node;
-	while (*newn) {
-		struct gfs2_blkreserv *cur =
-			rb_entry(*newn, struct gfs2_blkreserv, rs_node);
-		rc = rs_cmp(fsblk, 1, cur);
-		if (rc < 0)
-			newn = &((*newn)->rb_left);
-		else if (rc > 0)
-			newn = &((*newn)->rb_right);
-		else {
-			spin_unlock(&rgd->rd_rsspin);
-			return cur;
-		}
-	}
-	spin_unlock(&rgd->rd_rsspin);
-	return NULL;
-}
-
-/**
  * gfs2_bitfit - Search an rgrp's bitmap buffer to find a bit-pair representing
  *       a block in a given allocation state.
  * @buf: the buffer that holds the bitmaps
@@ -262,8 +215,6 @@ static u32 gfs2_bitfit(const u8 *buf, const unsigned int len,
 	u64 mask = 0x5555555555555555ULL;
 	u32 bit;
 
-	BUG_ON(state > 3);
-
 	/* Mask off bits we don't care about at the start of the search */
 	mask <<= spoint;
 	tmp = gfs2_bit_search(ptr, mask, state);
@@ -285,6 +236,131 @@ static u32 gfs2_bitfit(const u8 *buf, const unsigned int len,
 }
 
 /**
+ * gfs2_rbm_from_block - Set the rbm based upon rgd and block number
+ * @rbm: The rbm with rgd already set correctly
+ * @block: The block number (filesystem relative)
+ *
+ * This sets the bi and offset members of an rbm based on a
+ * resource group and a filesystem relative block number. The
+ * resource group must be set in the rbm on entry, the bi and
+ * offset members will be set by this function.
+ *
+ * Returns: 0 on success, or an error code
+ */
+
+static int gfs2_rbm_from_block(struct gfs2_rbm *rbm, u64 block)
+{
+	u64 rblock = block - rbm->rgd->rd_data0;
+	u32 goal = (u32)rblock;
+	int x;
+
+	if (WARN_ON_ONCE(rblock > UINT_MAX))
+		return -EINVAL;
+	if (block >= rbm->rgd->rd_data0 + rbm->rgd->rd_data)
+		return -E2BIG;
+
+	for (x = 0; x < rbm->rgd->rd_length; x++) {
+		rbm->bi = rbm->rgd->rd_bits + x;
+		if (goal < (rbm->bi->bi_start + rbm->bi->bi_len) * GFS2_NBBY) {
+			rbm->offset = goal - (rbm->bi->bi_start * GFS2_NBBY);
+			break;
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * gfs2_unaligned_extlen - Look for free blocks which are not byte aligned
+ * @rbm: Position to search (value/result)
+ * @n_unaligned: Number of unaligned blocks to check
+ * @len: Decremented for each block found (terminate on zero)
+ *
+ * Returns: true if a non-free block is encountered
+ */
+
+static bool gfs2_unaligned_extlen(struct gfs2_rbm *rbm, u32 n_unaligned, u32 *len)
+{
+	u64 block;
+	u32 n;
+	u8 res;
+
+	for (n = 0; n < n_unaligned; n++) {
+		res = gfs2_testbit(rbm);
+		if (res != GFS2_BLKST_FREE)
+			return true;
+		(*len)--;
+		if (*len == 0)
+			return true;
+		block = gfs2_rbm_to_block(rbm);
+		if (gfs2_rbm_from_block(rbm, block + 1))
+			return true;
+	}
+
+	return false;
+}
+
+/**
+ * gfs2_free_extlen - Return extent length of free blocks
+ * @rbm: Starting position
+ * @len: Max length to check
+ *
+ * Starting at the block specified by the rbm, see how many free blocks
+ * there are, not reading more than len blocks ahead. This can be done
+ * using memchr_inv when the blocks are byte aligned, but has to be done
+ * on a block by block basis in case of unaligned blocks. Also this
+ * function can cope with bitmap boundaries (although it must stop on
+ * a resource group boundary)
+ *
+ * Returns: Number of free blocks in the extent
+ */
+
+static u32 gfs2_free_extlen(const struct gfs2_rbm *rrbm, u32 len)
+{
+	struct gfs2_rbm rbm = *rrbm;
+	u32 n_unaligned = rbm.offset & 3;
+	u32 size = len;
+	u32 bytes;
+	u32 chunk_size;
+	u8 *ptr, *start, *end;
+	u64 block;
+
+	if (n_unaligned &&
+	    gfs2_unaligned_extlen(&rbm, 4 - n_unaligned, &len))
+		goto out;
+
+	n_unaligned = len & 3;
+	/* Start is now byte aligned */
+	while (len > 3) {
+		start = rbm.bi->bi_bh->b_data;
+		if (rbm.bi->bi_clone)
+			start = rbm.bi->bi_clone;
+		end = start + rbm.bi->bi_bh->b_size;
+		start += rbm.bi->bi_offset;
+		BUG_ON(rbm.offset & 3);
+		start += (rbm.offset / GFS2_NBBY);
+		bytes = min_t(u32, len / GFS2_NBBY, (end - start));
+		ptr = memchr_inv(start, 0, bytes);
+		chunk_size = ((ptr == NULL) ? bytes : (ptr - start));
+		chunk_size *= GFS2_NBBY;
+		BUG_ON(len < chunk_size);
+		len -= chunk_size;
+		block = gfs2_rbm_to_block(&rbm);
+		gfs2_rbm_from_block(&rbm, block + chunk_size);
+		n_unaligned = 3;
+		if (ptr)
+			break;
+		n_unaligned = len & 3;
+	}
+
+	/* Deal with any bits left over at the end */
+	if (n_unaligned)
+		gfs2_unaligned_extlen(&rbm, n_unaligned, &len);
+out:
+	return size - len;
+}
+
+/**
  * gfs2_bitcount - count the number of bits in a certain state
  * @rgd: the resource group descriptor
  * @buffer: the buffer that holds the bitmaps
@@ -487,6 +563,8 @@ int gfs2_rs_alloc(struct gfs2_inode *ip)
 	if (!res)
 		error = -ENOMEM;
 
+	RB_CLEAR_NODE(&res->rs_node);
+
 	down_write(&ip->i_rw_mutex);
 	if (ip->i_res)
 		kmem_cache_free(gfs2_rsrv_cachep, res);
@@ -496,11 +574,12 @@ int gfs2_rs_alloc(struct gfs2_inode *ip)
 	return error;
 }
 
-static void dump_rs(struct seq_file *seq, struct gfs2_blkreserv *rs)
+static void dump_rs(struct seq_file *seq, const struct gfs2_blkreserv *rs)
 {
-	gfs2_print_dbg(seq, "  r: %llu s:%llu b:%u f:%u\n",
-		       rs->rs_rgd->rd_addr, gfs2_rs_startblk(rs), rs->rs_biblk,
-		       rs->rs_free);
+	gfs2_print_dbg(seq, "  B: n:%llu s:%llu b:%u f:%u\n",
+		       (unsigned long long)rs->rs_inum,
+		       (unsigned long long)gfs2_rbm_to_block(&rs->rs_rbm),
+		       rs->rs_rbm.offset, rs->rs_free);
 }
 
 /**
@@ -508,41 +587,26 @@ static void dump_rs(struct seq_file *seq, struct gfs2_blkreserv *rs)
  * @rs: The reservation to remove
  *
  */
-static void __rs_deltree(struct gfs2_blkreserv *rs)
+static void __rs_deltree(struct gfs2_inode *ip, struct gfs2_blkreserv *rs)
 {
 	struct gfs2_rgrpd *rgd;
 
 	if (!gfs2_rs_active(rs))
 		return;
 
-	rgd = rs->rs_rgd;
-	/* We can't do this: The reason is that when the rgrp is invalidated,
-	   it's in the "middle" of acquiring the glock, but the HOLDER bit
-	   isn't set yet:
-	   BUG_ON(!gfs2_glock_is_locked_by_me(rs->rs_rgd->rd_gl));*/
-	trace_gfs2_rs(NULL, rs, TRACE_RS_TREEDEL);
-
-	if (!RB_EMPTY_ROOT(&rgd->rd_rstree))
-		rb_erase(&rs->rs_node, &rgd->rd_rstree);
-	BUG_ON(!rgd->rd_rs_cnt);
-	rgd->rd_rs_cnt--;
+	rgd = rs->rs_rbm.rgd;
+	trace_gfs2_rs(rs, TRACE_RS_TREEDEL);
+	rb_erase(&rs->rs_node, &rgd->rd_rstree);
+	RB_CLEAR_NODE(&rs->rs_node);
 
 	if (rs->rs_free) {
 		/* return reserved blocks to the rgrp and the ip */
-		BUG_ON(rs->rs_rgd->rd_reserved < rs->rs_free);
-		rs->rs_rgd->rd_reserved -= rs->rs_free;
+		BUG_ON(rs->rs_rbm.rgd->rd_reserved < rs->rs_free);
+		rs->rs_rbm.rgd->rd_reserved -= rs->rs_free;
 		rs->rs_free = 0;
-		clear_bit(GBF_FULL, &rs->rs_bi->bi_flags);
+		clear_bit(GBF_FULL, &rs->rs_rbm.bi->bi_flags);
 		smp_mb__after_clear_bit();
 	}
-	/* We can't change any of the step 1 or step 2 components of the rs.
-	   E.g. We can't set rs_rgd to NULL because the rgd glock is held and
-	   dequeued through this pointer.
-	   Can't: atomic_set(&rs->rs_sizehint, 0);
-	   Can't: rs->rs_requested = 0;
-	   Can't: rs->rs_rgd = NULL;*/
-	rs->rs_bi = NULL;
-	rs->rs_biblk = 0;
 }
 
 /**
@@ -550,17 +614,16 @@ static void __rs_deltree(struct gfs2_blkreserv *rs)
  * @rs: The reservation to remove
  *
  */
-void gfs2_rs_deltree(struct gfs2_blkreserv *rs)
+void gfs2_rs_deltree(struct gfs2_inode *ip, struct gfs2_blkreserv *rs)
 {
 	struct gfs2_rgrpd *rgd;
 
-	if (!gfs2_rs_active(rs))
-		return;
-
-	rgd = rs->rs_rgd;
-	spin_lock(&rgd->rd_rsspin);
-	__rs_deltree(rs);
-	spin_unlock(&rgd->rd_rsspin);
+	rgd = rs->rs_rbm.rgd;
+	if (rgd) {
+		spin_lock(&rgd->rd_rsspin);
+		__rs_deltree(ip, rs);
+		spin_unlock(&rgd->rd_rsspin);
+	}
 }
 
 /**
@@ -572,8 +635,7 @@ void gfs2_rs_delete(struct gfs2_inode *ip)
 {
 	down_write(&ip->i_rw_mutex);
 	if (ip->i_res) {
-		gfs2_rs_deltree(ip->i_res);
-		trace_gfs2_rs(ip, ip->i_res, TRACE_RS_DELETE);
+		gfs2_rs_deltree(ip, ip->i_res);
 		BUG_ON(ip->i_res->rs_free);
 		kmem_cache_free(gfs2_rsrv_cachep, ip->i_res);
 		ip->i_res = NULL;
@@ -597,7 +659,7 @@ static void return_all_reservations(struct gfs2_rgrpd *rgd)
 	spin_lock(&rgd->rd_rsspin);
 	while ((n = rb_first(&rgd->rd_rstree))) {
 		rs = rb_entry(n, struct gfs2_blkreserv, rs_node);
-		__rs_deltree(rs);
+		__rs_deltree(NULL, rs);
 	}
 	spin_unlock(&rgd->rd_rsspin);
 }
@@ -1270,211 +1332,276 @@ out:
 
 /**
  * rs_insert - insert a new multi-block reservation into the rgrp's rb_tree
- * @bi: the bitmap with the blocks
  * @ip: the inode structure
- * @biblk: the 32-bit block number relative to the start of the bitmap
- * @amount: the number of blocks to reserve
  *
- * Returns: NULL - reservation was already taken, so not inserted
- *          pointer to the inserted reservation
  */
-static struct gfs2_blkreserv *rs_insert(struct gfs2_bitmap *bi,
-				       struct gfs2_inode *ip, u32 biblk,
-				       int amount)
+static void rs_insert(struct gfs2_inode *ip)
 {
 	struct rb_node **newn, *parent = NULL;
 	int rc;
 	struct gfs2_blkreserv *rs = ip->i_res;
-	struct gfs2_rgrpd *rgd = rs->rs_rgd;
-	u64 fsblock = gfs2_bi2rgd_blk(bi, biblk) + rgd->rd_data0;
+	struct gfs2_rgrpd *rgd = rs->rs_rbm.rgd;
+	u64 fsblock = gfs2_rbm_to_block(&rs->rs_rbm);
+
+	BUG_ON(gfs2_rs_active(rs));
 
 	spin_lock(&rgd->rd_rsspin);
 	newn = &rgd->rd_rstree.rb_node;
-	BUG_ON(!ip->i_res);
-	BUG_ON(gfs2_rs_active(rs));
-	/* Figure out where to put new node */
-	/*BUG_ON(!gfs2_glock_is_locked_by_me(rgd->rd_gl));*/
 	while (*newn) {
 		struct gfs2_blkreserv *cur =
 			rb_entry(*newn, struct gfs2_blkreserv, rs_node);
 
 		parent = *newn;
-		rc = rs_cmp(fsblock, amount, cur);
+		rc = rs_cmp(fsblock, rs->rs_free, cur);
 		if (rc > 0)
 			newn = &((*newn)->rb_right);
 		else if (rc < 0)
 			newn = &((*newn)->rb_left);
 		else {
 			spin_unlock(&rgd->rd_rsspin);
-			return NULL; /* reservation already in use */
+			WARN_ON(1);
+			return;
 		}
 	}
 
-	/* Do our reservation work */
-	rs = ip->i_res;
-	rs->rs_free = amount;
-	rs->rs_biblk = biblk;
-	rs->rs_bi = bi;
 	rb_link_node(&rs->rs_node, parent, newn);
 	rb_insert_color(&rs->rs_node, &rgd->rd_rstree);
 
-	/* Do our inode accounting for the reservation */
-	/*BUG_ON(!gfs2_glock_is_locked_by_me(ip->i_gl));*/
-
 	/* Do our rgrp accounting for the reservation */
-	rgd->rd_reserved += amount; /* blocks reserved */
-	rgd->rd_rs_cnt++; /* number of in-tree reservations */
+	rgd->rd_reserved += rs->rs_free; /* blocks reserved */
 	spin_unlock(&rgd->rd_rsspin);
-	trace_gfs2_rs(ip, rs, TRACE_RS_INSERT);
-	return rs;
-}
-