[10/24] vfs: Add no_nrwrite_index_update writeback control flag
diff mbox

Message ID 20100524223016.208581966@clark.site
State Not Applicable, archived
Headers show

Commit Message

gregkh@suse.de May 24, 2010, 10:28 p.m. UTC
2.6.27-stable review patch.  If anyone has any objections, please let us know.

------------------


From: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>

commit 17bc6c30cf6bfffd816bdc53682dd46fc34a2cf4 upstream.

If no_nrwrite_index_update is set we don't update nr_to_write and
address space writeback_index in write_cache_pages.  This change
enables a file system to skip these updates in write_cache_pages and do
them in the writepages() callback.  This patch will be followed by an
ext4 patch that make use of these new flags.

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
CC: linux-fsdevel@vger.kernel.org
[dev@jaysonking.com: Modified the patch to account for subsequent changes in mainline being cherry-picked earlier for 2.6.27.y.]
Signed-off-by: Jayson R. King <dev@jaysonking.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

---
 include/linux/writeback.h |    9 +++++++++
 mm/page-writeback.c       |   14 +++++++++-----
 2 files changed, 18 insertions(+), 5 deletions(-)



--
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

Comments

Christoph Hellwig May 25, 2010, 11:12 a.m. UTC | #1
On Mon, May 24, 2010 at 03:28:06PM -0700, Greg KH wrote:
> 2.6.27-stable review patch.  If anyone has any objections, please let us know.

NACK.  This has been shown to cause severe writeback regression for
other filesystems and is beeing reverted in mianline.
--
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
gregkh@suse.de May 25, 2010, 4:52 p.m. UTC | #2
On Tue, May 25, 2010 at 07:12:08AM -0400, Christoph Hellwig wrote:
> On Mon, May 24, 2010 at 03:28:06PM -0700, Greg KH wrote:
> > 2.6.27-stable review patch.  If anyone has any objections, please let us know.
> 
> NACK.  This has been shown to cause severe writeback regression for
> other filesystems and is beeing reverted in mianline.

Ok, if I drop this one, I think I need to drop an ext4 patch as well,
I'll go figure that out and respin a -rc2 with this change in it.

thanks,

greg k-h
--
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
Jayson King May 25, 2010, 4:53 p.m. UTC | #3
On 05/25/2010 06:12 AM, Christoph Hellwig wrote:
> On Mon, May 24, 2010 at 03:28:06PM -0700, Greg KH wrote:
>> 2.6.27-stable review patch.  If anyone has any objections, please let us know.
> 
> NACK.  This has been shown to cause severe writeback regression for
> other filesystems and is beeing reverted in mianline.

It can't be reverted directly (you'll get a build failure) since the
next ext4 patch in this -stable series depends on it.

There's a commit from Ted Ts'o in the ext4 git, "ext4: Use our own
write_cache_pages()", which removes the dependency, but that patch is
not yet in mainline so isn't a -stable candidate. When/if that patch
hits mainline, it can be added to -stable and this patch dropped/reverted.

Jayson
--
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
gregkh@suse.de May 25, 2010, 4:58 p.m. UTC | #4
On Tue, May 25, 2010 at 07:12:08AM -0400, Christoph Hellwig wrote:
> On Mon, May 24, 2010 at 03:28:06PM -0700, Greg KH wrote:
> > 2.6.27-stable review patch.  If anyone has any objections, please let us know.
> 
> NACK.  This has been shown to cause severe writeback regression for
> other filesystems and is beeing reverted in mianline.

Ok, I've dropped this one, and the two ext4 patches after this one as
well.  If someone wants them back, please feel free to redo them so that
they apply to the .27 tree now.

thanks,

greg k-h
--
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
Jayson King May 25, 2010, 5 p.m. UTC | #5
On 05/25/2010 11:52 AM, Greg KH wrote:
> On Tue, May 25, 2010 at 07:12:08AM -0400, Christoph Hellwig wrote:
>> On Mon, May 24, 2010 at 03:28:06PM -0700, Greg KH wrote:
>>> 2.6.27-stable review patch.  If anyone has any objections, please let us know.
>>
>> NACK.  This has been shown to cause severe writeback regression for
>> other filesystems and is beeing reverted in mianline.
> 
> Ok, if I drop this one, I think I need to drop an ext4 patch as well,
> I'll go figure that out and respin a -rc2 with this change in it.

You'd have to drop the next two ext4 -stable patches to get a buildable
kernel if you drop this one. The last of those ext4 patches is the one
which fixes the deadlock from kernel bugzilla #12579.

Jayson
--
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
gregkh@suse.de May 25, 2010, 5:12 p.m. UTC | #6
On Tue, May 25, 2010 at 12:00:58PM -0500, Jayson R. King wrote:
> On 05/25/2010 11:52 AM, Greg KH wrote:
> > On Tue, May 25, 2010 at 07:12:08AM -0400, Christoph Hellwig wrote:
> >> On Mon, May 24, 2010 at 03:28:06PM -0700, Greg KH wrote:
> >>> 2.6.27-stable review patch.  If anyone has any objections, please let us know.
> >>
> >> NACK.  This has been shown to cause severe writeback regression for
> >> other filesystems and is beeing reverted in mianline.
> > 
> > Ok, if I drop this one, I think I need to drop an ext4 patch as well,
> > I'll go figure that out and respin a -rc2 with this change in it.
> 
> You'd have to drop the next two ext4 -stable patches to get a buildable
> kernel if you drop this one. The last of those ext4 patches is the one
> which fixes the deadlock from kernel bugzilla #12579.

I've now dropped those next two ext4 patches.  If someone wants to get
this bug fixed for the .27 tree, please feel free to send me the
patches.

Personally, I doubt many people care about ext4 on the .27 kernel
release...

thanks,

greg k-h
--
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
Jayson King May 26, 2010, 12:49 a.m. UTC | #7
On 05/25/2010 12:12 PM, Greg KH wrote:
> I've now dropped those next two ext4 patches.  If someone wants to get
> this bug fixed for the .27 tree, please feel free to send me the
> patches.

I'll nominate Ted's "ext4: Use our own write_cache_pages()" patch for
.27-stable when it hits mainline. If it is included in -stable, then
those last two ext4 patches can be re-applied.

> Personally, I doubt many people care about ext4 on the .27 kernel
> release...

In any case, 2.6.27 is a good kernel, and ext4 is a good FS, IMO.

Jayson

--
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

Patch
diff mbox

--- a/include/linux/writeback.h
+++ b/include/linux/writeback.h
@@ -62,6 +62,15 @@  struct writeback_control {
 	unsigned for_writepages:1;	/* This is a writepages() call */
 	unsigned range_cyclic:1;	/* range_start is cyclic */
 	unsigned more_io:1;		/* more io to be dispatched */
+	/*
+	 * write_cache_pages() won't update wbc->nr_to_write and
+	 * mapping->writeback_index if no_nrwrite_index_update
+	 * is set.  write_cache_pages() may write more than we
+	 * requested and we want to make sure nr_to_write and
+	 * writeback_index are updated in a consistent manner
+	 * so we use a single control to update them
+	 */
+	unsigned no_nrwrite_index_update:1;
 };
 
 /*
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -878,6 +878,7 @@  int write_cache_pages(struct address_spa
 	pgoff_t done_index;
 	int cycled;
 	int range_whole = 0;
+	long nr_to_write = wbc->nr_to_write;
 
 	if (wbc->nonblocking && bdi_write_congested(bdi)) {
 		wbc->encountered_congestion = 1;
@@ -985,9 +986,9 @@  continue_unlock:
 				}
  			}
 
-			if (wbc->nr_to_write > 0) {
-				wbc->nr_to_write--;
-				if (wbc->nr_to_write == 0 &&
+			if (nr_to_write > 0) {
+				nr_to_write--;
+				if (nr_to_write == 0 &&
 				    wbc->sync_mode == WB_SYNC_NONE) {
 					/*
 					 * We stop writing back only if we are
@@ -1024,8 +1025,11 @@  continue_unlock:
 		end = writeback_index - 1;
 		goto retry;
 	}
-	if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
-		mapping->writeback_index = done_index;
+	if (!wbc->no_nrwrite_index_update) {
+		if (wbc->range_cyclic || (range_whole && nr_to_write > 0))
+			mapping->writeback_index = done_index;
+		wbc->nr_to_write = nr_to_write;
+	}
 
 	return ret;
 }