2009-10-02 16:22:19

by Curt Wohlgemuth

Subject: [PATCH] ext4: fix ext4_dirty_inode() compile warning

Hi Ted:

Apparently my patch for

ext4: Make sure ext4_dirty_inode() updates the inode in no journal mode

didn't play well if CONFIG_JBD_DEBUG wasn't defined.

Here are two patches that fix it. I don't know if the jbd_debug() call has
any importance or not; you can choose which is better perhaps?

Thanks,
Curt

==============================================

Remove jbd_debug() call from ext4_inode_dirty()

Signed-off-by: Curt Wohlgemuth <[email protected]>

--- fs/ext4/inode.c.orig 2009-10-02 09:04:36.000000000 -0700
+++ fs/ext4/inode.c 2009-10-02 09:05:31.000000000 -0700
@@ -5851,14 +5851,12 @@ int ext4_mark_inode_dirty(handle_t *hand
*/
void ext4_dirty_inode(struct inode *inode)
{
- handle_t *current_handle = ext4_journal_current_handle();
handle_t *handle;

handle = ext4_journal_start(inode, 2);
if (IS_ERR(handle))
goto out;

- jbd_debug(5, "marking dirty. outer handle=%p\n", current_handle);
ext4_mark_inode_dirty(handle, inode);

ext4_journal_stop(handle);

==============================================

Make sure ext4_inode_dirty() compiles if CONFIG_JBD_DEBUG is not defined.

Signed-off-by: Curt Wohlgemuth <[email protected]>

--- fs/ext4/inode.c.orig 2009-10-02 09:04:36.000000000 -0700
+++ fs/ext4/inode.c 2009-10-02 09:07:31.000000000 -0700
@@ -5851,7 +5851,9 @@ int ext4_mark_inode_dirty(handle_t *hand
*/
void ext4_dirty_inode(struct inode *inode)
{
+#ifdef CONFIG_JBD_DEBUG
handle_t *current_handle = ext4_journal_current_handle();
+#endif
handle_t *handle;

handle = ext4_journal_start(inode, 2);

2009-10-03 01:08:58

by Theodore Ts'o

Subject: Re: [PATCH] ext4: fix ext4_dirty_inode() compile warning

On Fri, Oct 02, 2009 at 09:22:18AM -0700, Curt Wohlgemuth wrote:
> Hi Ted:
>
> Apparently my patch for
>
> ext4: Make sure ext4_dirty_inode() updates the inode in no journal mode
>
> didn't play well if CONFIG_JBD_DEBUG wasn't defined.
>
> Here are two patches that fix it. I don't know if the jbd_debug() call has
> any importance or not; you can choose which is better perhaps?

I chose the first one, with a slightly elaborated commit message.

Thanks,

- Ted