Btrfs: fix a bug in checking whether a inode is already in log

This is based on Josef's "Btrfs: turbo charge fsync".

The current btrfs checks if an inode is in log by comparing
root's last_log_commit to inode's last_sub_trans[2].

But the problem is that this root->last_log_commit is shared among
inodes.

Say we have N inodes to be logged, after the first inode,
root's last_log_commit is updated and the N-1 remained files will
be skipped.

This fixes the bug by keeping a local copy of root's last_log_commit
inside each inode and this local copy will be maintained itself.

[1]: we regard each log transaction as a subset of btrfs's transaction,
i.e. sub_trans

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>

diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 7c7bf818f3c1637f5c407728ea4ffdc4a954c698..ed8ca7ca5eff2d2ce0529f42d32fb394b0331e89 100644 (file)
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -144,6 +144,9 @@ struct btrfs_inode {
        /* flags field from the on disk inode */
        u32 flags;
 
+       /* a local copy of root's last_log_commit */
+       unsigned long last_log_commit;
+
        /*
         * Counters to keep track of the number of extent item's we may use due
         * to delalloc and such.  outstanding_extents is the number of extent
@@ -203,15 +206,10 @@ static inline bool btrfs_is_free_space_inode(struct inode *inode)
 
 static inline int btrfs_inode_in_log(struct inode *inode, u64 generation)
 {
-       struct btrfs_root *root = BTRFS_I(inode)->root;
-       int ret = 0;
-
-       mutex_lock(&root->log_mutex);
        if (BTRFS_I(inode)->logged_trans == generation &&
-           BTRFS_I(inode)->last_sub_trans <= root->last_log_commit)
-               ret = 1;
-       mutex_unlock(&root->log_mutex);
-       return ret;
+           BTRFS_I(inode)->last_sub_trans <= BTRFS_I(inode)->last_log_commit)
+               return 1;
+       return 0;
 }
 
 #endif

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index a6824bd044938bcb3fd3f6af575ad33d641a8561..24745b8f2745a1a797cca4151b67785533815117 100644 (file)
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -6774,6 +6774,7 @@ again:
 
        BTRFS_I(inode)->last_trans = root->fs_info->generation;
        BTRFS_I(inode)->last_sub_trans = BTRFS_I(inode)->root->log_transid;
+       BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->root->last_log_commit;
 
        unlock_extent_cached(io_tree, page_start, page_end, &cached_state, GFP_NOFS);
 
@@ -7018,6 +7019,7 @@ struct inode *btrfs_alloc_inode(struct super_block *sb)
        ei->csum_bytes = 0;
        ei->index_cnt = (u64)-1;
        ei->last_unlink_trans = 0;
+       ei->last_log_commit = 0;
 
        spin_lock_init(&ei->lock);
        ei->outstanding_extents = 0;

diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
index e8b8416c688b2a7b5d6c4f79c844bb07953cb3b5..1a138bfb8f13e8cc5a79accf5834ff4f105fea0a 100644 (file)
--- a/fs/btrfs/transaction.h
+++ b/fs/btrfs/transaction.h
@@ -88,6 +88,7 @@ static inline void btrfs_set_inode_last_trans(struct btrfs_trans_handle *trans,
 {
        BTRFS_I(inode)->last_trans = trans->transaction->transid;
        BTRFS_I(inode)->last_sub_trans = BTRFS_I(inode)->root->log_transid;
+       BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->root->last_log_commit;
 }
 
 int btrfs_end_transaction(struct btrfs_trans_handle *trans,

diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 71e71539ffb753a498406ac7d916c1f61e84fe55..fc0df95c286208df18b35ed5a6389a0370c24908 100644 (file)
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -3185,6 +3185,7 @@ next_slot:
                }
        }
        BTRFS_I(inode)->logged_trans = trans->transid;
+       BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->last_sub_trans;
 out_unlock:
        mutex_unlock(&BTRFS_I(inode)->log_mutex);