Patchwork ext4: Fix memory leak in ext4_init_new_dir

login
register
mail settings
Submitter Mel Gorman
Date Jan. 11, 2013, 10:06 a.m.
Message ID <20130111100639.GA3302@suse.de>
Download mbox | patch
Permalink /patch/211271/
State Rejected
Headers show

Comments

Mel Gorman - Jan. 11, 2013, 10:06 a.m.
Commit a774f9c2 "ext4: make ext4_init_dot_dotdot for inline dir usage"
is a preparation cleanup for later ext4 patches but there was a snag
when moving code to the newly created ext4_init_new_dir() function.
ext4_bread() is called twice but brelse() is only called one leading to a
memory leak. The user-visible effect is that after a compile-orientated
benchmark 0% of memory could be allocated as transparent huge pages.
This patch deletes the extra call and after it is applied, 76% of memory
could be allocated after the same test.

Signed-off-by: Mel Gorman <mgorman@suse.de>
---
 fs/ext4/namei.c |    1 -
 1 file changed, 1 deletion(-)

--
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
Tao Ma - Jan. 11, 2013, 10:12 a.m.
Hi Mel,

Thanks for the report.

This bug is already spotted by Guo Chao and the patch has been accepted
by Ted already.

Thanks,
Tao
On 01/11/2013 06:06 PM, Mel Gorman wrote:
> Commit a774f9c2 "ext4: make ext4_init_dot_dotdot for inline dir usage"
> is a preparation cleanup for later ext4 patches but there was a snag
> when moving code to the newly created ext4_init_new_dir() function.
> ext4_bread() is called twice but brelse() is only called one leading to a
> memory leak. The user-visible effect is that after a compile-orientated
> benchmark 0% of memory could be allocated as transparent huge pages.
> This patch deletes the extra call and after it is applied, 76% of memory
> could be allocated after the same test.
> 
> Signed-off-by: Mel Gorman <mgorman@suse.de>
> ---
>  fs/ext4/namei.c |    1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
> index 8990165..d068885 100644
> --- a/fs/ext4/namei.c
> +++ b/fs/ext4/namei.c
> @@ -2368,7 +2368,6 @@ static int ext4_init_new_dir(handle_t *handle, struct inode *dir,
>  	}
>  
>  	inode->i_size = EXT4_I(inode)->i_disksize = blocksize;
> -	dir_block = ext4_bread(handle, inode, 0, 1, &err);
>  	if (!(dir_block = ext4_bread(handle, inode, 0, 1, &err))) {
>  		if (!err) {
>  			err = -EIO;
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

--
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
Mel Gorman - Jan. 11, 2013, 10:30 a.m.
On Fri, Jan 11, 2013 at 06:12:17PM +0800, Tao Ma wrote:
> Hi Mel,
> 
> Thanks for the report.
> 
> This bug is already spotted by Guo Chao and the patch has been accepted
> by Ted already.
> 

Ah grand, I see now that it was merged for 3.8-rc3. Thanks for the heads up.

Patch

diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 8990165..d068885 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -2368,7 +2368,6 @@  static int ext4_init_new_dir(handle_t *handle, struct inode *dir,
 	}
 
 	inode->i_size = EXT4_I(inode)->i_disksize = blocksize;
-	dir_block = ext4_bread(handle, inode, 0, 1, &err);
 	if (!(dir_block = ext4_bread(handle, inode, 0, 1, &err))) {
 		if (!err) {
 			err = -EIO;