Skip to contentSkip to conversation

Btrfs: fix reported number of inode blocks after buffered append writes

PATCH In review [email protected] 2 participants 2 messages Oct 31, 2017 – Nov 6, 2017 Btrfs file system

lore
Share & tools
lore

[email protected]

author
#0 lore raw

PATCH 2/2 Btrfs: fix reported number of inode blocks after buffered append writes

From: Filipe Manana <[email protected]>

The patch from commit 7e3b975a0f92 ("Btrfs: fix reported number of inode blocks") introduced a regression where if we do a buffered write starting at position equal to or greater than the file's size and then stat(2) the file before writeback is triggered, the number of used blocks does not change (unless there's a prealloc/unwritten extent). Example:

  $ xfs_io -f -c "pwrite -S 0xab 0 64K" foobar
  $ du -h foobar
  0	foobar
  $ sync
  $ du -h foobar
  64K	foobar

The first version of that patch didn't had this regression and the second version, which was the one committed, was made only to address some performance regression detected by the intel test robots using fs_mark.

This fixes the regression by setting the new delaloc bit in the range, and doing it at btrfs_dirty_pages() while setting the regular dealloc bit as well, so that this way we set both bits at once avoiding navigation of the inode's io tree twice. Doing it at btrfs_dirty_pages() is also the most meaninful place, as we should set the new dellaloc bit when if we set the delalloc bit, which happens only if we copied bytes into the pages at __btrfs_buffered_write().

This was making some of LTP's du tests fail, which can be quickly run using a command line like the following:

  $ ./runltp -q -p -l /ltp.log -f commands -s du -d /mnt
Fixes 7e3b975a0f92 ("Btrfs: fix reported number of inode blocks")
Signed-off-by Filipe Manana <[email protected]>
---
 fs/btrfs/ctree.h                 |   1 +
 fs/btrfs/extent_io.h             |   5 +-
 fs/btrfs/file.c                  | 123 +++++++++++++++++++++------------------
 fs/btrfs/inode.c                 |   9 +--
 fs/btrfs/relocation.c            |   3 +-
 fs/btrfs/tests/extent-io-tests.c |   6 +-
 fs/btrfs/tests/inode-tests.c     |  12 ++--
 7 files changed, 86 insertions(+), 73 deletions(-)
fs/btrfs/ctree.h +1 -0
index 8fc690384c58..51f1f3e31608 100644
@@ -3174,6 +3174,7 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput);
 int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput,
 			       int nr);
 int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
+			      unsigned int extra_bits,
 			      struct extent_state **cached_state, int dedupe);
 int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *new_root,
fs/btrfs/extent_io.h +3 -2
index faffa28ba707..580a6c1c296c 100644
@@ -365,10 +365,11 @@ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		       struct extent_state **cached_state);
 
 static inline int set_extent_delalloc(struct extent_io_tree *tree, u64 start,
-		u64 end, struct extent_state **cached_state)
+				      u64 end, unsigned int extra_bits,
+				      struct extent_state **cached_state)
 {
 	return set_extent_bit(tree, start, end,
-			      EXTENT_DELALLOC | EXTENT_UPTODATE,
+			      EXTENT_DELALLOC | EXTENT_UPTODATE | extra_bits,
 			      NULL, cached_state, GFP_NOFS);
 }
fs/btrfs/file.c +66 -57

162 lines not on this page · raw message

fs/btrfs/inode.c +5 -4
index d94e3f68b9b1..185cdc904856 100644
@@ -2036,11 +2036,12 @@ static noinline int add_pending_csums(struct btrfs_trans_handle *trans,
 }
 
 int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
+			      unsigned int extra_bits,
 			      struct extent_state **cached_state, int dedupe)
 {
 	WARN_ON((end & (PAGE_SIZE - 1)) == 0);
 	return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end,
-				   cached_state);
+				   extra_bits, cached_state);
 }
 
 /* see btrfs_writepage_start_hook for details on why this is required */

@@ -2101,7 +2102,7 @@ static void btrfs_writepage_fixup_worker(struct btrfs_work *work)
 		goto out;
 	 }
 
-	btrfs_set_extent_delalloc(inode, page_start, page_end, &cached_state,
+	btrfs_set_extent_delalloc(inode, page_start, page_end, 0, &cached_state,
 				  0);
 	ClearPageChecked(page);
 	set_page_dirty(page);

@@ -4853,7 +4854,7 @@ int btrfs_truncate_block(struct inode *inode, loff_t from, loff_t len,
 			  EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG,
 			  0, 0, &cached_state, GFP_NOFS);
 
-	ret = btrfs_set_extent_delalloc(inode, block_start, block_end,
+	ret = btrfs_set_extent_delalloc(inode, block_start, block_end, 0,
 					&cached_state, 0);
 	if (ret) {
 		unlock_extent_cached(io_tree, block_start, block_end,

@@ -9252,7 +9253,7 @@ int btrfs_page_mkwrite(struct vm_fault *vmf)
 			  EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG,
 			  0, 0, &cached_state, GFP_NOFS);
 
-	ret = btrfs_set_extent_delalloc(inode, page_start, end,
+	ret = btrfs_set_extent_delalloc(inode, page_start, end, 0,
 					&cached_state, 0);
 	if (ret) {
 		unlock_extent_cached(io_tree, page_start, page_end,
fs/btrfs/relocation.c +2 -1
index 9841faef08ea..6432817906c7 100644
@@ -3266,7 +3266,8 @@ static int relocate_file_extent_cluster(struct inode *inode,
 			nr++;
 		}
 
-		btrfs_set_extent_delalloc(inode, page_start, page_end, NULL, 0);
+		btrfs_set_extent_delalloc(inode, page_start, page_end, 0, NULL,
+					  0);
 		set_page_dirty(page);
 
 		unlock_extent(&BTRFS_I(inode)->io_tree,
fs/btrfs/tests/extent-io-tests.c +3 -3
index d06b1c931d05..2e7f64a3b22b 100644
@@ -114,7 +114,7 @@ static int test_find_delalloc(u32 sectorsize)
 	 * |--- delalloc ---|
 	 * |---  search  ---|
 	 */
-	set_extent_delalloc(&tmp, 0, sectorsize - 1, NULL);
+	set_extent_delalloc(&tmp, 0, sectorsize - 1, 0, NULL);
 	start = 0;
 	end = 0;
 	found = find_lock_delalloc_range(inode, &tmp, locked_page, &start,

@@ -145,7 +145,7 @@ static int test_find_delalloc(u32 sectorsize)
 		test_msg("Couldn't find the locked page\n");
 		goto out_bits;
 	}
-	set_extent_delalloc(&tmp, sectorsize, max_bytes - 1, NULL);
+	set_extent_delalloc(&tmp, sectorsize, max_bytes - 1, 0, NULL);
 	start = test_start;
 	end = 0;
 	found = find_lock_delalloc_range(inode, &tmp, locked_page, &start,

@@ -200,7 +200,7 @@ static int test_find_delalloc(u32 sectorsize)
 	 *
 	 * We are re-using our test_start from above since it works out well.
 	 */
-	set_extent_delalloc(&tmp, max_bytes, total_dirty - 1, NULL);
+	set_extent_delalloc(&tmp, max_bytes, total_dirty - 1, 0, NULL);
 	start = test_start;
 	end = 0;
 	found = find_lock_delalloc_range(inode, &tmp, locked_page, &start,
fs/btrfs/tests/inode-tests.c +6 -6
index 8c91d03cc82d..5c329321c936 100644
@@ -969,7 +969,7 @@ static int test_extent_accounting(u32 sectorsize, u32 nodesize)
 
 	/* [BTRFS_MAX_EXTENT_SIZE] */
 	BTRFS_I(inode)->outstanding_extents++;
-	ret = btrfs_set_extent_delalloc(inode, 0, BTRFS_MAX_EXTENT_SIZE - 1,
+	ret = btrfs_set_extent_delalloc(inode, 0, BTRFS_MAX_EXTENT_SIZE - 1, 0,
 					NULL, 0);
 	if (ret) {
 		test_msg("btrfs_set_extent_delalloc returned %d\n", ret);

@@ -986,7 +986,7 @@ static int test_extent_accounting(u32 sectorsize, u32 nodesize)
 	BTRFS_I(inode)->outstanding_extents++;
 	ret = btrfs_set_extent_delalloc(inode, BTRFS_MAX_EXTENT_SIZE,
 					BTRFS_MAX_EXTENT_SIZE + sectorsize - 1,
-					NULL, 0);
+					0, NULL, 0);
 	if (ret) {
 		test_msg("btrfs_set_extent_delalloc returned %d\n", ret);
 		goto out;

@@ -1021,7 +1021,7 @@ static int test_extent_accounting(u32 sectorsize, u32 nodesize)
 	ret = btrfs_set_extent_delalloc(inode, BTRFS_MAX_EXTENT_SIZE >> 1,
 					(BTRFS_MAX_EXTENT_SIZE >> 1)
 					+ sectorsize - 1,
-					NULL, 0);
+					0, NULL, 0);
 	if (ret) {
 		test_msg("btrfs_set_extent_delalloc returned %d\n", ret);
 		goto out;

@@ -1044,7 +1044,7 @@ static int test_extent_accounting(u32 sectorsize, u32 nodesize)
 	ret = btrfs_set_extent_delalloc(inode,
 			BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize,
 			(BTRFS_MAX_EXTENT_SIZE << 1) + 3 * sectorsize - 1,
-			NULL, 0);
+			0, NULL, 0);
 	if (ret) {
 		test_msg("btrfs_set_extent_delalloc returned %d\n", ret);
 		goto out;

@@ -1062,7 +1062,7 @@ static int test_extent_accounting(u32 sectorsize, u32 nodesize)
 	BTRFS_I(inode)->outstanding_extents++;
 	ret = btrfs_set_extent_delalloc(inode,
 			BTRFS_MAX_EXTENT_SIZE + sectorsize,
-			BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize - 1, NULL, 0);
+			BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize - 1, 0, NULL, 0);
 	if (ret) {
 		test_msg("btrfs_set_extent_delalloc returned %d\n", ret);
 		goto out;

@@ -1099,7 +1099,7 @@ static int test_extent_accounting(u32 sectorsize, u32 nodesize)
 	BTRFS_I(inode)->outstanding_extents++;
 	ret = btrfs_set_extent_delalloc(inode,
 			BTRFS_MAX_EXTENT_SIZE + sectorsize,
-			BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize - 1, NULL, 0);
+			BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize - 1, 0, NULL, 0);
 	if (ret) {
 		test_msg("btrfs_set_extent_delalloc returned %d\n", ret);
 		goto out;
Signature · 1 line
2.11.0
#1 lore raw
Quoted 3 lines · On Tue, Oct 31, 2017 at 06:56:12PM +0000, [email protected] wrote:

From: Filipe Manana <[email protected]>

The patch from commit 7e3b975a0f92 ("Btrfs: fix reported number of inode

The commit id should be a7e3b975a0f9, looks like a typo, missing 'a' at the beginning.

Quoted 84 lines

blocks") introduced a regression where if we do a buffered write starting at position equal to or greater than the file's size and then stat(2) the file before writeback is triggered, the number of used blocks does not change (unless there's a prealloc/unwritten extent). Example:

  $ xfs_io -f -c "pwrite -S 0xab 0 64K" foobar
  $ du -h foobar
  0	foobar
  $ sync
  $ du -h foobar
  64K	foobar

The first version of that patch didn't had this regression and the second version, which was the one committed, was made only to address some performance regression detected by the intel test robots using fs_mark.

This fixes the regression by setting the new delaloc bit in the range, and doing it at btrfs_dirty_pages() while setting the regular dealloc bit as well, so that this way we set both bits at once avoiding navigation of the inode's io tree twice. Doing it at btrfs_dirty_pages() is also the most meaninful place, as we should set the new dellaloc bit when if we set the delalloc bit, which happens only if we copied bytes into the pages at __btrfs_buffered_write().

This was making some of LTP's du tests fail, which can be quickly run using a command line like the following:

  $ ./runltp -q -p -l /ltp.log -f commands -s du -d /mnt
Fixes 7e3b975a0f92 ("Btrfs: fix reported number of inode blocks")
Signed-off-by Filipe Manana <[email protected]>
---
 fs/btrfs/ctree.h                 |   1 +
 fs/btrfs/extent_io.h             |   5 +-
 fs/btrfs/file.c                  | 123 +++++++++++++++++++++------------------
 fs/btrfs/inode.c                 |   9 +--
 fs/btrfs/relocation.c            |   3 +-
 fs/btrfs/tests/extent-io-tests.c |   6 +-
 fs/btrfs/tests/inode-tests.c     |  12 ++--
 7 files changed, 86 insertions(+), 73 deletions(-)
fs/btrfs/ctree.h +1 -0
index 8fc690384c58..51f1f3e31608 100644
@@ -3174,6 +3174,7 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput);
 int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput,
 			       int nr);
 int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
+			      unsigned int extra_bits,
 			      struct extent_state **cached_state, int dedupe);
 int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *new_root,
fs/btrfs/extent_io.h +3 -2
index faffa28ba707..580a6c1c296c 100644
@@ -365,10 +365,11 @@ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		       struct extent_state **cached_state);
 
 static inline int set_extent_delalloc(struct extent_io_tree *tree, u64 start,
-		u64 end, struct extent_state **cached_state)
+				      u64 end, unsigned int extra_bits,
+				      struct extent_state **cached_state)
 {
 	return set_extent_bit(tree, start, end,
-			      EXTENT_DELALLOC | EXTENT_UPTODATE,
+			      EXTENT_DELALLOC | EXTENT_UPTODATE | extra_bits,
 			      NULL, cached_state, GFP_NOFS);
 }
fs/btrfs/file.c +5 -0
index aaab1838cece..bcac827d65f0 100644
@@ -477,6 +477,47 @@ static void btrfs_drop_pages(struct page **pages, size_t num_pages)
 	}
 }
 
+static int btrfs_find_new_delalloc_bytes(struct btrfs_inode *inode,
+					 const u64 start,
+					 const u64 len,
+					 struct extent_state **cached_state)
+{

...

Quoted 1 line

+}

Can you please separate moving of btrfs_find_new_delalloc_bytes to another patch?