ext4: do not ask jbd2 to write data for delalloc buffers

Currently we ask jbd2 to write all dirty allocated buffers before
committing a transaction when doing writeback of delay allocated blocks.
However this is unnecessary since we move all pages to writeback state
before dropping a transaction handle and then submit all the necessary
IO. We still need the transaction commit to wait for all the outstanding
writeback before flushing disk caches during transaction commit to avoid
data exposure issues though. Use the new jbd2 capability and ask it to
only wait for outstanding writeback during transaction commit when
writing back data in ext4_writepages().

Tested-by: "HUANG Weller (CM/ESW12-CN)" <Weller.Huang@cn.bosch.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>

diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index f75e9ebd4ca2c1ba90b0d0f1f449cda6993177ad..cb00b1119ec935ae36dd927d93af6c4d7db2dbcc 100644 (file)
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -581,6 +581,9 @@ enum {
 #define EXT4_GET_BLOCKS_ZERO                   0x0200
 #define EXT4_GET_BLOCKS_CREATE_ZERO            (EXT4_GET_BLOCKS_CREATE |\
                                        EXT4_GET_BLOCKS_ZERO)
+       /* Caller will submit data before dropping transaction handle. This
+        * allows jbd2 to avoid submitting data before commit. */
+#define EXT4_GET_BLOCKS_IO_SUBMIT              0x0400
 
 /*
  * The bit position of these flags must not overlap with any of the

diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h
index f1c940b38b30cce172c4aad4bf288ebaa4d4bb37..09c1ef38cbe6aaff2c03185b31da20255df8990b 100644 (file)
--- a/fs/ext4/ext4_jbd2.h
+++ b/fs/ext4/ext4_jbd2.h
@@ -359,7 +359,8 @@ static inline int ext4_journal_force_commit(journal_t *journal)
        return 0;
 }
 
-static inline int ext4_jbd2_file_inode(handle_t *handle, struct inode *inode)
+static inline int ext4_jbd2_inode_add_write(handle_t *handle,
+                                           struct inode *inode)
 {
        if (ext4_handle_valid(handle))
                return jbd2_journal_inode_add_write(handle,
@@ -367,6 +368,15 @@ static inline int ext4_jbd2_file_inode(handle_t *handle, struct inode *inode)
        return 0;
 }
 
+static inline int ext4_jbd2_inode_add_wait(handle_t *handle,
+                                          struct inode *inode)
+{
+       if (ext4_handle_valid(handle))
+               return jbd2_journal_inode_add_wait(handle,
+                                                  EXT4_I(inode)->jinode);
+       return 0;
+}
+
 static inline void ext4_update_inode_fsync_trans(handle_t *handle,
                                                 struct inode *inode,
                                                 int datasync)

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 8ba46ad06aedb9f8f6c7ee56b2979e541c9ce51c..17bfa42ac97132dff751baace2e73c4c9de3d120 100644 (file)
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -695,7 +695,10 @@ out_sem:
                    !(flags & EXT4_GET_BLOCKS_ZERO) &&
                    !IS_NOQUOTA(inode) &&
                    ext4_should_order_data(inode)) {
-                       ret = ext4_jbd2_file_inode(handle, inode);
+                       if (flags & EXT4_GET_BLOCKS_IO_SUBMIT)
+                               ret = ext4_jbd2_inode_add_wait(handle, inode);
+                       else
+                               ret = ext4_jbd2_inode_add_write(handle, inode);
                        if (ret)
                                return ret;
                }
@@ -2319,7 +2322,8 @@ static int mpage_map_one_extent(handle_t *handle, struct mpage_da_data *mpd)
         * the data was copied into the page cache.
         */
        get_blocks_flags = EXT4_GET_BLOCKS_CREATE |
-                          EXT4_GET_BLOCKS_METADATA_NOFAIL;
+                          EXT4_GET_BLOCKS_METADATA_NOFAIL |
+                          EXT4_GET_BLOCKS_IO_SUBMIT;
        dioread_nolock = ext4_should_dioread_nolock(inode);
        if (dioread_nolock)
                get_blocks_flags |= EXT4_GET_BLOCKS_IO_CREATE_EXT;
@@ -3634,7 +3638,7 @@ static int __ext4_block_zero_page_range(handle_t *handle,
                err = 0;
                mark_buffer_dirty(bh);
                if (ext4_should_order_data(inode))
-                       err = ext4_jbd2_file_inode(handle, inode);
+                       err = ext4_jbd2_inode_add_write(handle, inode);
        }
 
 unlock:

diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index 325cef48b39a8d23788dc17ef056ccf79a7b58f6..a920c5d29fac0a5b5ef83c9cca18feeb3dcc0fe7 100644 (file)
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -400,7 +400,7 @@ data_copy:
 
        /* Even in case of data=writeback it is reasonable to pin
         * inode to transaction, to prevent unexpected data loss */
-       *err = ext4_jbd2_file_inode(handle, orig_inode);
+       *err = ext4_jbd2_inode_add_write(handle, orig_inode);
 
 unlock_pages:
        unlock_page(pagep[0]);