[PATCH v2 2/2] ext4: disable COLLAPSE_RANGE for bigalloc

[PATCH v2 2/2] ext4: disable COLLAPSE_RANGE for bigalloc

[Namjae Jeon]

From: Namjae Jeon <namjae.jeon@samsung.com>

Once COLLAPSE RANGE is be disable for ext4 with bigalloc feature till finding
root-cause of problem. It will be enable with fixing that regression of
xfstest(generic 075 and 091) again.

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Ashish Sangwan <a.sangwan@samsung.com>
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
---
 fs/ext4/extents.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 4cf8c5b..3276865 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -5406,6 +5406,9 @@ int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
 	if (!S_ISREG(inode->i_mode))
 		return -EINVAL;
 
+	if (EXT4_SB(inode->i_sb)->s_cluster_ratio > 1)
+		return -EOPNOTSUPP;
+
 	trace_ext4_collapse_range(inode, offset, len);
 
 	punch_start = offset >> EXT4_BLOCK_SIZE_BITS(sb);
-- 
1.7.11-rc0

Re: [PATCH v2 2/2] ext4: disable COLLAPSE_RANGE for bigalloc

[Theodore Ts'o]

On Sat, Apr 19, 2014 at 01:53:50PM +0900, Namjae Jeon wrote:
> From: Namjae Jeon <namjae.jeon@samsung.com>
> 
> Once COLLAPSE RANGE is be disable for ext4 with bigalloc feature till finding
> root-cause of problem. It will be enable with fixing that regression of
> xfstest(generic 075 and 091) again.
> 
> Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
> Signed-off-by: Ashish Sangwan <a.sangwan@samsung.com>
> Reviewed-by: Lukas Czerner <lczerner@redhat.com>

Thanks, applied.

What's the status of the "[2/3] ext4: fix ZERO_RANGE test failure in
data journalling mode" patch.  Is it no longer needed?

     		       	       	     - Ted

RE: [PATCH v2 2/2] ext4: disable COLLAPSE_RANGE for bigalloc

[Namjae Jeon]

> On Sat, Apr 19, 2014 at 01:53:50PM +0900, Namjae Jeon wrote:
> > From: Namjae Jeon <namjae.jeon@samsung.com>
> >
> > Once COLLAPSE RANGE is be disable for ext4 with bigalloc feature till finding
> > root-cause of problem. It will be enable with fixing that regression of
> > xfstest(generic 075 and 091) again.
> >
> > Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
> > Signed-off-by: Ashish Sangwan <a.sangwan@samsung.com>
> > Reviewed-by: Lukas Czerner <lczerner@redhat.com>
> 
> Thanks, applied.
> 
> What's the status of the "[2/3] ext4: fix ZERO_RANGE test failure in
> data journalling mode" patch.  Is it no longer needed?
It is needed. Currently, I am considering your suggestion of introducing
EXT4_I(inode)->i_write_mutex which can also include ext4_aio_mutex.

Thanks!
> 
>      		       	       	     - Ted

Re: [PATCH v2 2/2] ext4: disable COLLAPSE_RANGE for bigalloc

[Theodore Ts'o]

On Mon, Apr 21, 2014 at 01:52:28PM +0900, Namjae Jeon wrote:
> > 
> > What's the status of the "[2/3] ext4: fix ZERO_RANGE test failure in
> > data journalling mode" patch.  Is it no longer needed?
> It is needed. Currently, I am considering your suggestion of introducing
> EXT4_I(inode)->i_write_mutex which can also include ext4_aio_mutex.

OK; I've already pushed a set of patches to Linus because it's getting
fairly late in the development cycle, and we really want to get as
much of the bug fixes into -rc3 (having missed -rc2 by a few hours,
sigh).

By the way, in doing some final testing, it appears that we are still
failing generic/127 with a 1k blocksize.  If I block COLLAPSE_RANGE
using the patch that everyone but me seems to hate :-), the problem
goes away.  So we have at least one other issue that needs to be
looked at.

You can reproduce by grabbing the dev branch from the ext4.git tree,
and then cherry-picking the top commit from the unstable branch.

Then run "kvm-xfstests -c 1k generic/127", with and without the
following in config.custom:

EXTRA_ARG="ext4.fallocate_mode_block=0x08"

Cheers,

						- Ted

RE: [PATCH v2 2/2] ext4: disable COLLAPSE_RANGE for bigalloc

[Namjae Jeon]

> On Mon, Apr 21, 2014 at 01:52:28PM +0900, Namjae Jeon wrote:
> > >
> > > What's the status of the "[2/3] ext4: fix ZERO_RANGE test failure in
> > > data journalling mode" patch.  Is it no longer needed?
> > It is needed. Currently, I am considering your suggestion of introducing
> > EXT4_I(inode)->i_write_mutex which can also include ext4_aio_mutex.
> 
> OK; I've already pushed a set of patches to Linus because it's getting
> fairly late in the development cycle, and we really want to get as
> much of the bug fixes into -rc3 (having missed -rc2 by a few hours,
> sigh).
Okay, I will try to fix remaning issues in -rc3.
> 
> By the way, in doing some final testing, it appears that we are still
> failing generic/127 with a 1k blocksize.  If I block COLLAPSE_RANGE
> using the patch that everyone but me seems to hate :-), the problem
> goes away.  So we have at least one other issue that needs to be
> looked at.
> 
> You can reproduce by grabbing the dev branch from the ext4.git tree,
> and then cherry-picking the top commit from the unstable branch.
> 
> Then run "kvm-xfstests -c 1k generic/127", with and without the
> following in config.custom:
> 
> EXTRA_ARG="ext4.fallocate_mode_block=0x08"
Okay, I will look at this issue now.

Thanks Ted!!
> 
> Cheers,
> 
> 						- Ted