Skip to contentSkip to conversation

Ocfs2: Make ocfs2_find_cpos_for_left_leaf() public.

PATCH New Tristan Ye 2 participants 4 messages Apr 9, 2010 OCFS2

lore
Share & tools
lore

Tristan Ye

author
#0 lore raw

PATCH 1/2 Ocfs2: Make ocfs2_find_cpos_for_left_leaf() public.

The original idea to pull ocfs2_find_cpos_for_left_leaf() out of alloc.c is to benefit punching-holes optimization patch, it however, can also be referred by other funcs in the future who want to do the same job.

Signed-off-by Tristan Ye <[email protected]>
---
 fs/ocfs2/alloc.c |    4 ++--
 fs/ocfs2/alloc.h |    2 ++
 2 files changed, 4 insertions(+), 2 deletions(-)
fs/ocfs2/alloc.c +2 -2
index 6ea34f7..cbbcff5 100644
@@ -2248,8 +2248,8 @@ out:
  *
  * Will return zero if the path passed in is already the leftmost path.
  */
-static int ocfs2_find_cpos_for_left_leaf(struct super_block *sb,
-					 struct ocfs2_path *path, u32 *cpos)
+int ocfs2_find_cpos_for_left_leaf(struct super_block *sb,
+				  struct ocfs2_path *path, u32 *cpos)
 {
 	int i, j, ret = 0;
 	u64 blkno;
fs/ocfs2/alloc.h +2 -0
index 2963ed7..d03f765 100644
@@ -319,6 +319,8 @@ int ocfs2_journal_access_path(struct ocfs2_caching_info *ci,
 			      struct ocfs2_path *path);
 int ocfs2_find_cpos_for_right_leaf(struct super_block *sb,
 				   struct ocfs2_path *path, u32 *cpos);
+int ocfs2_find_cpos_for_left_leaf(struct super_block *sb,
+				  struct ocfs2_path *path, u32 *cpos);
 int ocfs2_find_subtree_root(struct ocfs2_extent_tree *et,
 			    struct ocfs2_path *left,
 			    struct ocfs2_path *right);
Signature · 1 line
1.5.5
author
#1 lore raw

PATCH 2/2 Ocfs2: Optimize punching-hole codes v5.

V5 patch simplifies the logic of handling existing holes and procedures for skipping extent blocks, and removed most of confusing comments.

V5 patch has survived with fill_verify_holes testcase in ocfs2-test, it also passed my manual sanity check and stress tests with enormous extent records.

Currently punching hole on a file with 3+ extent tree depth was really a performance disaster, it even caused several hours to go, though we may not hit this in real life with such a huge extent number.

One simple way to improve the performance is quite straightforward, by learning the logic of truncating codes, means we'd punch hole from hole_end to hole_start, which reduce the overhead of btree operation in a significant way, such as tree rotation and moving.

Following is the testing result when punching hole from 0 to file end in bytes, on a 1G file, 1G file consists of 256k extent records, each record cover 4k data(just one cluster, clustersize is 4k):

===========================================================================
* Former punching-hole mechanism:
===========================================================================

I waited 1 hour for its completion, unfortunately it's still ongoing.

===========================================================================
* Patched punching-hode mechanism:
===========================================================================

real 0m2.518s
user 0m0.000s
sys 0m2.445s

That means we've gained up to 1000 times improvement on performance in this case, whee! It's fairly cool. and it looks like that performance gain will be raising when extent records grow.

The patch was based on my former 2 patches, which were about truncating codes optimization and fixup to handle CoW on punching hole.

Signed-off-by Tristan Ye <[email protected]>
---
 fs/ocfs2/file.c |  173 +++++++++++++++++++++++++++++++++++++++++++++++--------
 1 files changed, 148 insertions(+), 25 deletions(-)
fs/ocfs2/file.c +148 -25
index db2e0c9..907653e 100644
@@ -1423,18 +1423,97 @@ out:
 	return ret;
 }
 
+static int ocfs2_find_rec(struct ocfs2_extent_list *el,
+			  u32 pos,
+			  int include_pos)
+{
+	int i;
+	struct ocfs2_extent_rec *rec = NULL;
+
+	for (i = le16_to_cpu(el->l_next_free_rec) - 1; i >= 0; i--) {
+
+		rec = &el->l_recs[i];
+
+		if (include_pos) {
+			if (le32_to_cpu(rec->e_cpos) <= pos)
+				break;
+		} else {
+			if (le32_to_cpu(rec->e_cpos) < pos)
+				break;
+		}
+	}
+
+	return i;
+}
+
+/*
+ * Helper to calculate the punching pos and length in one run, we handle the
+ * following three cases in order:
+ *
+ * - remove the entire record
+ * - remove a partial record
+ * - no record needs to be removed (hole-punching completed)
+*/
+static void ocfs2_calc_trunc_pos(struct inode *inode,
+				 struct ocfs2_extent_list *el,
+				 struct ocfs2_extent_rec *rec,
+				 u32 trunc_start, u32 *trunc_cpos,
+				 u32 *trunc_len, u32 *trunc_end,
+				 u64 *blkno, int *done)
+{
+	int ret = 0;
+	u32 coff, range;
+
+	range = le32_to_cpu(rec->e_cpos) + ocfs2_rec_clusters(el, rec);
+
+	if (le32_to_cpu(rec->e_cpos) >= trunc_start) {
+		*trunc_cpos = le32_to_cpu(rec->e_cpos);
+		/*
+		 * Skip holes if any.
+		 */
+		if (range < *trunc_end)
+			*trunc_end = range;
+		*trunc_len = *trunc_end - le32_to_cpu(rec->e_cpos);
+		*blkno = le64_to_cpu(rec->e_blkno);
+		*trunc_end = le32_to_cpu(rec->e_cpos);
+	} else if (range > trunc_start) {
+		*trunc_cpos = trunc_start;
+		*trunc_len = range - trunc_start;
+		coff = trunc_start - le32_to_cpu(rec->e_cpos);
+		*blkno = le64_to_cpu(rec->e_blkno) +
+				ocfs2_clusters_to_blocks(inode->i_sb, coff);
+		*trunc_end = trunc_start;
+	} else {
+		/*
+		 * It may have two following possibilities:
+		 *
+		 * - last record has been removed
+		 * - trunc_start was within a hole
+		 *
+		 * both two cases mean the completion of hole punching.
+		 */
+		ret = 1;
+	}
+
+	*done = ret;
+}
+
 static int ocfs2_remove_inode_range(struct inode *inode,
 				    struct buffer_head *di_bh, u64 byte_start,
 				    u64 byte_len)
 {
-	int ret = 0, flags = 0;
-	u32 trunc_start, trunc_len, cpos, phys_cpos, alloc_size;
+	int ret = 0, flags = 0, done = 0, i;
+	u32 trunc_start, trunc_len, trunc_end, trunc_cpos, phys_cpos;
+	u32 cluster_in_el;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_cached_dealloc_ctxt dealloc;
 	struct address_space *mapping = inode->i_mapping;
 	struct ocfs2_extent_tree et;
+	struct ocfs2_path *path = NULL;
+	struct ocfs2_extent_list *el = NULL;
+	struct ocfs2_extent_rec *rec = NULL;
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
-	u64 refcount_loc = le64_to_cpu(di->i_refcount_loc);
+	u64 blkno, refcount_loc = le64_to_cpu(di->i_refcount_loc);
 
 	ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode), di_bh);
 	ocfs2_init_dealloc_ctxt(&dealloc);

@@ -1482,16 +1561,13 @@ static int ocfs2_remove_inode_range(struct inode *inode,
 	}
 
 	trunc_start = ocfs2_clusters_for_bytes(osb->sb, byte_start);
-	trunc_len = (byte_start + byte_len) >> osb->s_clustersize_bits;
-	if (trunc_len >= trunc_start)
-		trunc_len -= trunc_start;
-	else
-		trunc_len = 0;
+	trunc_end = (byte_start + byte_len) >> osb->s_clustersize_bits;
+	cluster_in_el = trunc_end;
 
-	mlog(0, "Inode: %llu, start: %llu, len: %llu, cstart: %u, clen: %u\n",
+	mlog(0, "Inode: %llu, start: %llu, len: %llu, cstart: %u, cend: %u\n",
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     (unsigned long long)byte_start,
-	     (unsigned long long)byte_len, trunc_start, trunc_len);
+	     (unsigned long long)byte_len, trunc_start, trunc_end);
 
 	ret = ocfs2_zero_partial_clusters(inode, byte_start, byte_len);
 	if (ret) {

@@ -1499,32 +1575,79 @@ static int ocfs2_remove_inode_range(struct inode *inode,
 		goto out;
 	}
 
-	cpos = trunc_start;
-	while (trunc_len) {
-		ret = ocfs2_get_clusters(inode, cpos, &phys_cpos,
-					 &alloc_size, &flags);
+	path = ocfs2_new_path_from_et(&et);
+	if (!path) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	while (trunc_end > trunc_start) {
+
+		ret = ocfs2_find_path(INODE_CACHE(inode), path,
+				      cluster_in_el);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		if (alloc_size > trunc_len)
-			alloc_size = trunc_len;
+		el = path_leaf_el(path);
 
-		/* Only do work for non-holes */
-		if (phys_cpos != 0) {
-			ret = ocfs2_remove_btree_range(inode, &et, cpos,
-						       phys_cpos, alloc_size,
-						       &dealloc, refcount_loc,
-						       flags);
-			if (ret) {
+		i = ocfs2_find_rec(el, trunc_end, 0);
+		/*
+		 * Need to go to previous extent block.
+		 */
+		if (i < 0) {
+			if (path->p_tree_depth == 0)
+				break;
+
+			ret = ocfs2_find_cpos_for_left_leaf(inode->i_sb,
+							    path,
+							    &cluster_in_el);
+			if (ret) {                 
 				mlog_errno(ret);
 				goto out;
 			}
+
+			/*
+			 * We've reached the leftmost extent block,
+			 * it's safe to leave.
+			 */
+			if (cluster_in_el == 0)
+				break;
+
+			/*
+			 * The 'pos' searched for previous extent block is
+			 * always one cluster less than actual trunc_end.
+			 */
+			trunc_end = cluster_in_el + 1;
+
+			ocfs2_reinit_path(path, 1);
+
+			continue;
+		
+		} else
+			rec = &el->l_recs[i];
+
+		ocfs2_calc_trunc_pos(inode, el, rec, trunc_start, &trunc_cpos,
+				     &trunc_len, &trunc_end, &blkno, &done);
+		if (done)
+			break;
+
+		flags = rec->e_flags;
+		phys_cpos = ocfs2_blocks_to_clusters(inode->i_sb, blkno);
+
+		ret = ocfs2_remove_btree_range(inode, &et, trunc_cpos,
+					       phys_cpos, trunc_len, &dealloc,
+					       refcount_loc, flags);
+		if (ret < 0) {
+			mlog_errno(ret);
+			goto out;
 		}
 
-		cpos += alloc_size;
-		trunc_len -= alloc_size;
+		cluster_in_el = trunc_end;
+
+		ocfs2_reinit_path(path, 1);
 	}
 
 	ocfs2_truncate_cluster_pages(inode, byte_start, byte_len);
Signature · 1 line
1.5.5
#2 lore raw

PATCH 2/2 Ocfs2: Optimize punching-hole codes v5.

Hi Tristan,

Quoted 16 lines · Tristan Ye

V5 patch simplifies the logic of handling existing holes and procedures for skipping extent blocks, and removed most of confusing comments.

V5 patch has survived with fill_verify_holes testcase in ocfs2-test, it also passed my manual sanity check and stress tests with enormous extent records.

Signed-off-by Tristan Ye <[email protected]>
---
 fs/ocfs2/file.c |  173 +++++++++++++++++++++++++++++++++++++++++++++++--------
 1 files changed, 148 insertions(+), 25 deletions(-)
fs/ocfs2/file.c +0 -0
index db2e0c9..907653e 100644

fs/ocfs2/file.c out: · line 1428

Quoted 6 lines
fs/ocfs2/file.c +3 -0
@@ -1423,18 +1423,97 @@ out:
 	return ret;
 }
 
+static int ocfs2_find_rec(struct ocfs2_extent_list *el,
+			  u32 pos,
+			  int include_pos)

Why you need this 'include_pos'? btw, there is only one caller with include_pos = 0, so why not remove it? ;)

Quoted 43 lines
fs/ocfs2/file.c +43 -0
+{
+	int i;
+	struct ocfs2_extent_rec *rec = NULL;
+
+	for (i = le16_to_cpu(el->l_next_free_rec) - 1; i >= 0; i--) {
+
+		rec = &el->l_recs[i];
+
+		if (include_pos) {
+			if (le32_to_cpu(rec->e_cpos) <= pos)
+				break;
+		} else {
+			if (le32_to_cpu(rec->e_cpos) < pos)
+				break;
+		}
+	}
+
+	return i;
+}
+
+/*
+ * Helper to calculate the punching pos and length in one run, we handle the
+ * following three cases in order:
+ *
+ * - remove the entire record
+ * - remove a partial record
+ * - no record needs to be removed (hole-punching completed)
+*/
+static void ocfs2_calc_trunc_pos(struct inode *inode,
+				 struct ocfs2_extent_list *el,
+				 struct ocfs2_extent_rec *rec,
+				 u32 trunc_start, u32 *trunc_cpos,
+				 u32 *trunc_len, u32 *trunc_end,
+				 u64 *blkno, int *done)
+{
+	int ret = 0;
+	u32 coff, range;
+
+	range = le32_to_cpu(rec->e_cpos) + ocfs2_rec_clusters(el, rec);
+
+	if (le32_to_cpu(rec->e_cpos) >= trunc_start) {
+		*trunc_cpos = le32_to_cpu(rec->e_cpos);
+		/*

fs/ocfs2/file.c out: · line 1481

Quoted 10 lines
fs/ocfs2/file.c +10 -0
+		 * Skip holes if any.
+		 */
+		if (range < *trunc_end)
+			*trunc_end = range;
+		*trunc_len = *trunc_end - le32_to_cpu(rec->e_cpos);
+		*blkno = le64_to_cpu(rec->e_blkno);
+		*trunc_end = le32_to_cpu(rec->e_cpos);
+	} else if (range > trunc_start) {
+		*trunc_cpos = trunc_start;
+		*trunc_len = range - trunc_start;

I guess there should be a bug? what if the old trunc_end < range? then *trunc_len should be min(range, *trunc_end) - trunc_start?

Quoted 9 lines
fs/ocfs2/file.c +9 -0
+		coff = trunc_start - le32_to_cpu(rec->e_cpos);
+		*blkno = le64_to_cpu(rec->e_blkno) +
+				ocfs2_clusters_to_blocks(inode->i_sb, coff);
+		*trunc_end = trunc_start;
+	} else {
+		/*
+		 * It may have two following possibilities:
+		 *
+		 * - last record has been removed

fs/ocfs2/file.c out: · line 1500

Quoted 10 lines
fs/ocfs2/file.c +10 -0
+		 * - trunc_start was within a hole
+		 *
+		 * both two cases mean the completion of hole punching.
+		 */
+		ret = 1;
+	}
+
+	*done = ret;
+}
+

Regards,
Tao

author
#3 lore raw

PATCH 2/2 Ocfs2: Optimize punching-hole codes v5.

That's really a quick review;)

I appreciated it.

Quoted 23 lines · Tao Ma

Hi Tristan,

Quoted 20 lines · Tristan Ye

V5 patch simplifies the logic of handling existing holes and procedures for skipping extent blocks, and removed most of confusing comments.

V5 patch has survived with fill_verify_holes testcase in ocfs2-test, it also passed my manual sanity check and stress tests with enormous extent records.

Signed-off-by Tristan Ye <[email protected]>

---
fs/ocfs2/file.c | 173
+++++++++++++++++++++++++++++++++++++++++++++++--------

1 files changed, 148 insertions(+), 25 deletions(-)
fs/ocfs2/file.c +0 -0
index db2e0c9..907653e 100644
@@ -1423,18 +1423,97 @@ out:
 return ret;
 }
Quoted 5 lines

fs/ocfs2/file.c out: · line 1428

Quoted 3 lines
fs/ocfs2/file.c +3 -0
+static int ocfs2_find_rec(struct ocfs2_extent_list *el,
+ u32 pos,
+ int include_pos)

Why you need this 'include_pos'? btw, there is only one caller with include_pos = 0, so why not remove it? ;)

I originally used it in two places, anyway, it can be simplified as you told;) that doesn't matter actually, keeing this can make the func looks more common;)

Quoted 48 lines
Quoted 48 lines
fs/ocfs2/file.c +22 -0
+{
+ int i;
+ struct ocfs2_extent_rec *rec = NULL;
+
+ for (i = le16_to_cpu(el->l_next_free_rec) - 1; i >= 0; i--) {
+
+ rec = &el->l_recs[i];
+
+ if (include_pos) {
+ if (le32_to_cpu(rec->e_cpos) <= pos)
+ break;
+ } else {
+ if (le32_to_cpu(rec->e_cpos) < pos)
+ break;
+ }
+ }
+
+ return i;
+}
+
+/*
+ * Helper to calculate the punching pos and length in one run, we

handle the
+ * following three cases in order:
+ *
+ * - remove the entire record
+ * - remove a partial record
+ * - no record needs to be removed (hole-punching completed)

fs/ocfs2/file.c +20 -0
+*/
+static void ocfs2_calc_trunc_pos(struct inode *inode,
+ struct ocfs2_extent_list *el,
+ struct ocfs2_extent_rec *rec,
+ u32 trunc_start, u32 *trunc_cpos,
+ u32 *trunc_len, u32 *trunc_end,
+ u64 *blkno, int *done)
+{
+ int ret = 0;
+ u32 coff, range;
+
+ range = le32_to_cpu(rec->e_cpos) + ocfs2_rec_clusters(el, rec);
+
+ if (le32_to_cpu(rec->e_cpos) >= trunc_start) {
+ *trunc_cpos = le32_to_cpu(rec->e_cpos);
+ /*
+ * Skip holes if any.
+ */
+ if (range < *trunc_end)
+ *trunc_end = range;
Quoted 8 lines
Quoted 6 lines

+ *trunc_len = *trunc_end - le32_to_cpu(rec->e_cpos);
+ *blkno = le64_to_cpu(rec->e_blkno);
+ *trunc_end = le32_to_cpu(rec->e_cpos);
+ } else if (range > trunc_start) {
+ *trunc_cpos = trunc_start;
+ *trunc_len = range - trunc_start;

I guess there should be a bug? what if the old trunc_end < range? then *trunc_len should be min(range, *trunc_end) - trunc_start?

That's definitely a bug, good catch!!

should be trunc_len = trunc_end - trunc_start;

Unlike partial truncating(a partial truncate always from end of a record), while punching could happen anywhere.

Quoted 21 lines
Quoted 15 lines

+ coff = trunc_start - le32_to_cpu(rec->e_cpos);
+ *blkno = le64_to_cpu(rec->e_blkno) +
+ ocfs2_clusters_to_blocks(inode->i_sb, coff);
+ *trunc_end = trunc_start;
+ } else {
+ /*
+ * It may have two following possibilities:
+ *
+ * - last record has been removed
+ * - trunc_start was within a hole
+ *
+ * both two cases mean the completion of hole punching.
+ */
+ ret = 1;
+ }

fs/ocfs2/file.c out: · line 1500

Quoted 4 lines
fs/ocfs2/file.c +4 -0
+
+ *done = ret;
+}
+

Regards,
Tao