2015-04-07 05:35:07

by Davide Italiano

[permalink] [raw]
Subject: [PATCH] Remove redundant check under lock scope in fallocate()

The check for indirect files is already done in ext4_fallocate()
and outside of the lock scope. I removed it entirely from ext4_zero_range
and ext4_collapse range.

Davide Italiano (1):
ext4: Remove redundant check under lock scope

fs/ext4/extents.c | 16 ----------------
1 file changed, 16 deletions(-)

--
2.3.4



2015-04-07 05:51:07

by Davide Italiano

[permalink] [raw]
Subject: [PATCH] ext4: Remove redundant check under lock scope

ext4_zero_range() and ext4_collapse_range() duplicate
the check in ext4_fallocate(). The checks are made with
inode lock held when there's no need for that. Remove them,
reducing the scope of the lock.

Signed-off-by: Davide Italiano <[email protected]>
---
fs/ext4/extents.c | 16 ----------------
1 file changed, 16 deletions(-)

diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index bed4308..9e6fa09 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -4810,15 +4810,6 @@ static long ext4_zero_range(struct file *file, loff_t offset,
flags |= EXT4_GET_BLOCKS_KEEP_SIZE;

mutex_lock(&inode->i_mutex);
-
- /*
- * Indirect files do not support unwritten extnets
- */
- if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
- ret = -EOPNOTSUPP;
- goto out_mutex;
- }
-
if (!(mode & FALLOC_FL_KEEP_SIZE) &&
offset + len > i_size_read(inode)) {
new_size = offset + len;
@@ -5445,13 +5436,6 @@ int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
ret = -EINVAL;
goto out_mutex;
}
-
- /* Currently just for extent based files */
- if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
- ret = -EOPNOTSUPP;
- goto out_mutex;
- }

2015-04-07 10:15:23

by Dmitry Monakhov

[permalink] [raw]
Subject: Re: [PATCH] ext4: Remove redundant check under lock scope

Davide Italiano <[email protected]> writes:

> ext4_zero_range() and ext4_collapse_range() duplicate
> the check in ext4_fallocate(). The checks are made with
> inode lock held when there's no need for that. Remove them,
> reducing the scope of the lock.
NAK. Other task can convert indirect<=>extent via
ioctl(, EXT4_IOC_SETFLAGS, flags)
->ext4_ext_migrate()
->ext4_int_migrate()
So we _must_ recheck EXT4_INODE_EXTENTS after we grab i_mutex.
>
> Signed-off-by: Davide Italiano <[email protected]>
> ---
> fs/ext4/extents.c | 16 ----------------
> 1 file changed, 16 deletions(-)
>
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index bed4308..9e6fa09 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -4810,15 +4810,6 @@ static long ext4_zero_range(struct file *file, loff_t offset,
> flags |= EXT4_GET_BLOCKS_KEEP_SIZE;
>
> mutex_lock(&inode->i_mutex);
> -
> - /*
> - * Indirect files do not support unwritten extnets
> - */
> - if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
> - ret = -EOPNOTSUPP;
> - goto out_mutex;
> - }
> -
> if (!(mode & FALLOC_FL_KEEP_SIZE) &&
> offset + len > i_size_read(inode)) {
> new_size = offset + len;
> @@ -5445,13 +5436,6 @@ int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
> ret = -EINVAL;
> goto out_mutex;
> }
> -
> - /* Currently just for extent based files */
> - if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
> - ret = -EOPNOTSUPP;
> - goto out_mutex;
> - }
> -
> truncate_pagecache(inode, ioffset);
>
> /* Wait for existing dio to complete */
> --
> 2.3.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html


Attachments:
signature.asc (472.00 B)

2015-04-07 18:15:20

by Davide Italiano

[permalink] [raw]
Subject: Re: [PATCH] ext4: Remove redundant check under lock scope

On Tue, Apr 7, 2015 at 3:13 AM, Dmitry Monakhov <[email protected]> wrote:
>
> Davide Italiano <[email protected]> writes:
>
> > ext4_zero_range() and ext4_collapse_range() duplicate
> > the check in ext4_fallocate(). The checks are made with
> > inode lock held when there's no need for that. Remove them,
> > reducing the scope of the lock.
> NAK. Other task can convert indirect<=>extent via
> ioctl(, EXT4_IOC_SETFLAGS, flags)
> ->ext4_ext_migrate()
> ->ext4_int_migrate()
> So we _must_ recheck EXT4_INODE_EXTENTS after we grab i_mutex.
> >
> > Signed-off-by: Davide Italiano <[email protected]>
> > ---
> > fs/ext4/extents.c | 16 ----------------
> > 1 file changed, 16 deletions(-)
> >
> > diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> > index bed4308..9e6fa09 100644
> > --- a/fs/ext4/extents.c
> > +++ b/fs/ext4/extents.c
> > @@ -4810,15 +4810,6 @@ static long ext4_zero_range(struct file *file, loff_t offset,
> > flags |= EXT4_GET_BLOCKS_KEEP_SIZE;
> >
> > mutex_lock(&inode->i_mutex);
> > -
> > - /*
> > - * Indirect files do not support unwritten extnets
> > - */
> > - if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
> > - ret = -EOPNOTSUPP;
> > - goto out_mutex;
> > - }
> > -
> > if (!(mode & FALLOC_FL_KEEP_SIZE) &&
> > offset + len > i_size_read(inode)) {
> > new_size = offset + len;
> > @@ -5445,13 +5436,6 @@ int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
> > ret = -EINVAL;
> > goto out_mutex;
> > }
> > -
> > - /* Currently just for extent based files */
> > - if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
> > - ret = -EOPNOTSUPP;
> > - goto out_mutex;
> > - }
> > -
> > truncate_pagecache(inode, ioffset);
> >
> > /* Wait for existing dio to complete */
> > --
> > 2.3.4
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> > the body of a message to [email protected]
> > More majordomo info at http://vger.kernel.org/majordomo-info.html


Thanks, this explains it. Do you think it's worth adding a comment
explaining this behaviour?

--
Davide