[2/2] xfs: Close race between direct IO and xfs_break_layouts()

Message ID 153367990333.37314.16218849614019392916.stgit@djiang5-desk3.ch.intel.com
State Superseded
Headers show
Series
  • [1/2] ext4: Close race between direct IO and ext4_break_layouts()
Related show

Commit Message

Dave Jiang Aug. 7, 2018, 10:11 p.m.
This patch is the duplicate of ross's fix for ext4 for xfs.

If the refcount of a page is lowered between the time that it is returned
by dax_busy_page() and when the refcount is again checked in
xfs_break_layouts() => ___wait_var_event(), the waiting function
xfs_wait_dax_page() will never be called.  This means that
xfs_break_layouts() will still have 'retry' set to false, so we'll stop
looping and never check the refcount of other pages in this inode.

Instead, always continue looping as long as dax_layout_busy_page() gives us
a page which it found with an elevated refcount.

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
---
 fs/xfs/xfs_file.c |    7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

Comments

Jan Kara Aug. 8, 2018, 8:53 a.m. | #1
On Tue 07-08-18 15:11:43, Dave Jiang wrote:
> This patch is the duplicate of ross's fix for ext4 for xfs.
> 
> If the refcount of a page is lowered between the time that it is returned
> by dax_busy_page() and when the refcount is again checked in
> xfs_break_layouts() => ___wait_var_event(), the waiting function
> xfs_wait_dax_page() will never be called.  This means that
> xfs_break_layouts() will still have 'retry' set to false, so we'll stop
> looping and never check the refcount of other pages in this inode.
> 
> Instead, always continue looping as long as dax_layout_busy_page() gives us
> a page which it found with an elevated refcount.
> 
> Signed-off-by: Dave Jiang <dave.jiang@intel.com>

The patch looks good to me. You can add:

Reviewed-by: Jan Kara <jack@suse.cz>

Just one minor nit below:

> @@ -746,9 +744,10 @@ xfs_break_dax_layouts(
>  	if (!page)
>  		return 0;
>  
> +	*did_unlock = true;

I think it would be more understandable to name the argument of
xfs_break_dax_layouts() as 'retry' instead of 'did_unlock' as it's not
about unlocking anymore.

							Honza
Dave Jiang Aug. 8, 2018, 3:47 p.m. | #2
On 08/08/2018 01:53 AM, Jan Kara wrote:
> On Tue 07-08-18 15:11:43, Dave Jiang wrote:
>> This patch is the duplicate of ross's fix for ext4 for xfs.
>>
>> If the refcount of a page is lowered between the time that it is returned
>> by dax_busy_page() and when the refcount is again checked in
>> xfs_break_layouts() => ___wait_var_event(), the waiting function
>> xfs_wait_dax_page() will never be called.  This means that
>> xfs_break_layouts() will still have 'retry' set to false, so we'll stop
>> looping and never check the refcount of other pages in this inode.
>>
>> Instead, always continue looping as long as dax_layout_busy_page() gives us
>> a page which it found with an elevated refcount.
>>
>> Signed-off-by: Dave Jiang <dave.jiang@intel.com>
> 
> The patch looks good to me. You can add:
> 
> Reviewed-by: Jan Kara <jack@suse.cz>
> 
> Just one minor nit below:
> 
>> @@ -746,9 +744,10 @@ xfs_break_dax_layouts(
>>  	if (!page)
>>  		return 0;
>>  
>> +	*did_unlock = true;
> 
> I think it would be more understandable to name the argument of
> xfs_break_dax_layouts() as 'retry' instead of 'did_unlock' as it's not
> about unlocking anymore.

Thanks for the review Jan! I will change. I was trying to decide between
less code change vs more clear definition. :)

> 
> 							Honza
>

Patch

diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index a3e7767a5715..666c93fe5284 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -721,12 +721,10 @@  xfs_file_write_iter(
 
 static void
 xfs_wait_dax_page(
-	struct inode		*inode,
-	bool			*did_unlock)
+	struct inode		*inode)
 {
 	struct xfs_inode        *ip = XFS_I(inode);
 
-	*did_unlock = true;
 	xfs_iunlock(ip, XFS_MMAPLOCK_EXCL);
 	schedule();
 	xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
@@ -746,9 +744,10 @@  xfs_break_dax_layouts(
 	if (!page)
 		return 0;
 
+	*did_unlock = true;
 	return ___wait_var_event(&page->_refcount,
 			atomic_read(&page->_refcount) == 1, TASK_INTERRUPTIBLE,
-			0, 0, xfs_wait_dax_page(inode, did_unlock));
+			0, 0, xfs_wait_dax_page(inode));
 }
 
 int