If we flush inodes with pending delalloc in a transaction, we may join
the same transaction handler more than 2 times.
The reason is:
Task use_count of trans handle
commit_transaction 1
|-> btrfs_start_delalloc_inodes 1
|-> run_delalloc_nocow 1
|-> join_transaction 2
|-> cow_file_range 2
|-> join_transaction 3
In fact, cow_file_range needn't join the transaction again because the caller
have joined the transaction, so we fix this problem by this way.
Reported-by: Liu Bo <bo.li....@oracle.com>
Signed-off-by: Miao Xie <mi...@cn.fujitsu.com>
---
fs/btrfs/inode.c | 77 +++++++++++++++++++++++++++++------------------
fs/btrfs/transaction.c | 1 +
2 files changed, 48 insertions(+), 30 deletions(-)
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 5fc0990..aadcdd6 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -803,14 +803,14 @@ static u64 get_extent_allocation_hint(struct inode
*inode, u64 start,
* required to start IO on it. It may be clean and already done with
* IO when we return.
*/
-static noinline int cow_file_range(struct inode *inode,
- struct page *locked_page,
- u64 start, u64 end, int *page_started,
- unsigned long *nr_written,
- int unlock)
+static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
+ struct inode *inode,
+ struct btrfs_root *root,
+ struct page *locked_page,
+ u64 start, u64 end, int *page_started,
+ unsigned long *nr_written,
+ int unlock)
{
- struct btrfs_root *root = BTRFS_I(inode)->root;
- struct btrfs_trans_handle *trans;
u64 alloc_hint = 0;
u64 num_bytes;
unsigned long ram_size;
@@ -823,25 +823,10 @@ static noinline int cow_file_range(struct inode *inode,
int ret = 0;
BUG_ON(btrfs_is_free_space_inode(inode));
- trans = btrfs_join_transaction(root);
- if (IS_ERR(trans)) {
- extent_clear_unlock_delalloc(inode,
- &BTRFS_I(inode)->io_tree,
- start, end, locked_page,
- EXTENT_CLEAR_UNLOCK_PAGE |
- EXTENT_CLEAR_UNLOCK |
- EXTENT_CLEAR_DELALLOC |
- EXTENT_CLEAR_DIRTY |
- EXTENT_SET_WRITEBACK |
- EXTENT_END_WRITEBACK);
- return PTR_ERR(trans);
- }
- trans->block_rsv = &root->fs_info->delalloc_block_rsv;
num_bytes = (end - start + blocksize) & ~(blocksize - 1);
num_bytes = max(blocksize, num_bytes);
disk_num_bytes = num_bytes;
- ret = 0;
/* if this is a small write inside eof, kick off defrag */
if (num_bytes < 64 * 1024 &&
@@ -952,11 +937,9 @@ static noinline int cow_file_range(struct inode *inode,
alloc_hint = ins.objectid + ins.offset;
start += cur_alloc_size;
}
- ret = 0;
out:
- btrfs_end_transaction(trans, root);
-
return ret;
+
out_unlock:
extent_clear_unlock_delalloc(inode,
&BTRFS_I(inode)->io_tree,
@@ -971,6 +954,39 @@ out_unlock:
goto out;
}
+static noinline int cow_file_range(struct inode *inode,
+ struct page *locked_page,
+ u64 start, u64 end, int *page_started,
+ unsigned long *nr_written,
+ int unlock)
+{
+ struct btrfs_trans_handle *trans;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ int ret;
+
+ trans = btrfs_join_transaction(root);
+ if (IS_ERR(trans)) {
+ extent_clear_unlock_delalloc(inode,
+ &BTRFS_I(inode)->io_tree,
+ start, end, locked_page,
+ EXTENT_CLEAR_UNLOCK_PAGE |
+ EXTENT_CLEAR_UNLOCK |
+ EXTENT_CLEAR_DELALLOC |
+ EXTENT_CLEAR_DIRTY |
+ EXTENT_SET_WRITEBACK |
+ EXTENT_END_WRITEBACK);
+ return PTR_ERR(trans);
+ }
+ trans->block_rsv = &root->fs_info->delalloc_block_rsv;
+
+ ret = __cow_file_range(trans, inode, root, locked_page, start, end,
+ page_started, nr_written, unlock);
+
+ btrfs_end_transaction(trans, root);
+
+ return ret;
+}
+
/*
* work queue call back to started compression on a file and pages
*/
@@ -1281,9 +1297,9 @@ out_check:
btrfs_release_path(path);
if (cow_start != (u64)-1) {
- ret = cow_file_range(inode, locked_page, cow_start,
- found_key.offset - 1, page_started,
- nr_written, 1);
+ ret = __cow_file_range(trans, inode, root, locked_page,
+ cow_start, found_key.offset - 1,
+ page_started, nr_written, 1);
if (ret) {
btrfs_abort_transaction(trans, root, ret);
goto error;
@@ -1352,8 +1368,9 @@ out_check:
}
if (cow_start != (u64)-1) {
- ret = cow_file_range(inode, locked_page, cow_start, end,
- page_started, nr_written, 1);
+ ret = __cow_file_range(trans, inode, root, locked_page,
+ cow_start, end,
+ page_started, nr_written, 1);
if (ret) {
btrfs_abort_transaction(trans, root, ret);
goto error;
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 04bbfb1..6d0d5a0 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -312,6 +312,7 @@ static struct btrfs_trans_handle *start_transaction(struct
btrfs_root *root,
WARN_ON(type != TRANS_JOIN && type != TRANS_JOIN_NOLOCK);
h = current->journal_info;
h->use_count++;
+ WARN_ON(h->use_count > 2);
h->orig_rsv = h->block_rsv;
h->block_rsv = NULL;
goto got_it;
--
1.7.6.5
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
