Patchwork [2/2] ext4: avoid finding next leaf if newext->ee_block smaller than fex->ee_block

login
register
mail settings
Submitter Theodore Ts'o
Date July 11, 2011, 7:52 p.m.
Message ID <20110711195211.GQ28763@thunk.org>
Download mbox | patch
Permalink /patch/104276/
State Accepted
Headers show

Comments

Theodore Ts'o - July 11, 2011, 7:52 p.m.
On Thu, Jun 30, 2011 at 04:42:31PM +0800, Yongqiang Yang wrote:
> How about:
>            next = EXT_MAX_BLOCKS;
>            if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block)
>                    next = ext4_ext_next_leaf_block(inode, path);
>            if (next != EXT_MAX_BLOCKS) {

Agreed, that's a better way of doing things.  Also, Robin, please
compare and contrast your description with mine.  The commit
description shouldn't just be a textual description of the change.  It
should give the larger context of the change and why it's important.

              	    	       	       - Ted

commit 2a3f7e0e0e55f8817fbe92d111ef2a06e6b8ef18
Author: Robin Dong <sanbai@taobao.com>
Date:   Mon Jul 11 15:43:38 2011 -0400

    ext4: avoid unneeded ext4_ext_next_leaf_block() while inserting extents
    
    Optimize ext4_exT_insert_extent() by avoiding
    ext4_ext_next_leaf_block() when the result is not used/needed.
    
    Signed-off-by: Robin Dong <sanbai@taobao.com>
    Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>

--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Robin Dong - July 12, 2011, 1:33 a.m.
2011/7/12 Ted Ts'o <tytso@mit.edu>:
> On Thu, Jun 30, 2011 at 04:42:31PM +0800, Yongqiang Yang wrote:
>> How about:
>>            next = EXT_MAX_BLOCKS;
>>            if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block)
>>                    next = ext4_ext_next_leaf_block(inode, path);
>>            if (next != EXT_MAX_BLOCKS) {
>
> Agreed, that's a better way of doing things.  Also, Robin, please
> compare and contrast your description with mine.  The commit
> description shouldn't just be a textual description of the change.  It
> should give the larger context of the change and why it's important.
>
>                                       - Ted
>
> commit 2a3f7e0e0e55f8817fbe92d111ef2a06e6b8ef18
> Author: Robin Dong <sanbai@taobao.com>
> Date:   Mon Jul 11 15:43:38 2011 -0400
>
>    ext4: avoid unneeded ext4_ext_next_leaf_block() while inserting extents
>
>    Optimize ext4_exT_insert_extent() by avoiding
>    ext4_ext_next_leaf_block() when the result is not used/needed.
>
>    Signed-off-by: Robin Dong <sanbai@taobao.com>
>    Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
>
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index 9cbdcb2..f1c538e 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -1730,9 +1730,10 @@ int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
>
>        /* probably next leaf has space for us? */
>        fex = EXT_LAST_EXTENT(eh);
> -       next = ext4_ext_next_leaf_block(inode, path);
> -       if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block)
> -           && next != EXT_MAX_BLOCKS) {
> +       next = EXT_MAX_BLOCKS;
> +       if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block))
> +               next = ext4_ext_next_leaf_block(inode, path);
> +       if (next != EXT_MAX_BLOCKS) {
>                ext_debug("next leaf block - %d\n", next);
>                BUG_ON(npath != NULL);
>                npath = ext4_ext_find_extent(inode, next, NULL);
>

It seems much better than my description. Thanks!

Patch

diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 9cbdcb2..f1c538e 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -1730,9 +1730,10 @@  int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
 
 	/* probably next leaf has space for us? */
 	fex = EXT_LAST_EXTENT(eh);
-	next = ext4_ext_next_leaf_block(inode, path);
-	if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block)
-	    && next != EXT_MAX_BLOCKS) {
+	next = EXT_MAX_BLOCKS;
+	if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block))
+		next = ext4_ext_next_leaf_block(inode, path);
+	if (next != EXT_MAX_BLOCKS) {
 		ext_debug("next leaf block - %d\n", next);
 		BUG_ON(npath != NULL);
 		npath = ext4_ext_find_extent(inode, next, NULL);