FS_CGROUP_WRITEBACK indicates whether a file_system_type supports
cgroup writeback; however, different super_blocks of the same
file_system_type may or may not support cgroup writeback depending on
filesystem options. This patch replaces FS_CGROUP_WRITEBACK with a
kernel-internal super_block->s_flags MS_CGROUPWB. The concatenated
and abbreviated name is for consistency with other MS_* flags.
ext2_fill_super() is updated to assert MS_CGROUPWB.
Signed-off-by: Tejun Heo <[email protected]>
Cc: Alexander Viro <viro-RmSDqhL/[email protected]>
Cc: [email protected]
Cc: Jan Kara <[email protected]>
Cc: [email protected]
---
fs/ext2/super.c | 4 ++--
fs/namespace.c | 2 +-
include/linux/backing-dev.h | 2 +-
include/linux/fs.h | 1 -
include/uapi/linux/fs.h | 1 +
5 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/fs/ext2/super.c b/fs/ext2/super.c
index 549219d..472ed34 100644
--- a/fs/ext2/super.c
+++ b/fs/ext2/super.c
@@ -879,7 +879,7 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent)
if (!parse_options((char *) data, sb))
goto failed_mount;
- sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
+ sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | MS_CGROUPWB |
((EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ?
MS_POSIXACL : 0);
@@ -1543,7 +1543,7 @@ static struct file_system_type ext2_fs_type = {
.name = "ext2",
.mount = ext2_mount,
.kill_sb = kill_block_super,
- .fs_flags = FS_REQUIRES_DEV | FS_CGROUP_WRITEBACK,
+ .fs_flags = FS_REQUIRES_DEV,
};
MODULE_ALIAS_FS("ext2");
diff --git a/fs/namespace.c b/fs/namespace.c
index 1f4f9da..507b90b 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -2669,7 +2669,7 @@ long do_mount(const char *dev_name, const char __user *dir_name,
flags &= ~(MS_NOSUID | MS_NOEXEC | MS_NODEV | MS_ACTIVE | MS_BORN |
MS_NOATIME | MS_NODIRATIME | MS_RELATIME| MS_KERNMOUNT |
- MS_STRICTATIME);
+ MS_STRICTATIME | MS_CGROUPWB);
if (flags & MS_REMOUNT)
retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags,
diff --git a/include/linux/backing-dev.h b/include/linux/backing-dev.h
index dfce808..1489131 100644
--- a/include/linux/backing-dev.h
+++ b/include/linux/backing-dev.h
@@ -260,7 +260,7 @@ static inline bool inode_cgwb_enabled(struct inode *inode)
return bdi_cap_account_dirty(bdi) &&
(bdi->capabilities & BDI_CAP_CGROUP_WRITEBACK) &&
- (inode->i_sb->s_type->fs_flags & FS_CGROUP_WRITEBACK);
+ (inode->i_sb->s_flags & MS_CGROUPWB);
}
/**
diff --git a/include/linux/fs.h b/include/linux/fs.h
index b5e1dcf..66e35dc 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1912,7 +1912,6 @@ struct file_system_type {
#define FS_HAS_SUBTYPE 4
#define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
#define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
-#define FS_CGROUP_WRITEBACK 32 /* Supports cgroup-aware writeback */
#define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
struct dentry *(*mount) (struct file_system_type *, int,
const char *, void *);
diff --git a/include/uapi/linux/fs.h b/include/uapi/linux/fs.h
index 9b964a5..60316e7 100644
--- a/include/uapi/linux/fs.h
+++ b/include/uapi/linux/fs.h
@@ -93,6 +93,7 @@ struct inodes_stat_t {
#define MS_LAZYTIME (1<<25) /* Update the on-disk [acm]times lazily */
/* These sb flags are internal to the kernel */
+#define MS_CGROUPWB (1<<27) /* cgroup-aware writeback enabled */
#define MS_NOSEC (1<<28)
#define MS_BORN (1<<29)
#define MS_ACTIVE (1<<30)
--
2.4.2
On Fri, Jun 12, 2015 at 04:57:33PM -0500, Tejun Heo wrote:
> FS_CGROUP_WRITEBACK indicates whether a file_system_type supports
> cgroup writeback; however, different super_blocks of the same
> file_system_type may or may not support cgroup writeback depending on
> filesystem options. This patch replaces FS_CGROUP_WRITEBACK with a
> kernel-internal super_block->s_flags MS_CGROUPWB. The concatenated
> and abbreviated name is for consistency with other MS_* flags.
Nak. As the uapi part makes it obvious the MS_ namespace is part
of the userspace ABI. Please add a new in-kernel flags field instead.
Hello, Christoph.
On Sat, Jun 13, 2015 at 09:16:08AM -0700, Christoph Hellwig wrote:
> On Fri, Jun 12, 2015 at 04:57:33PM -0500, Tejun Heo wrote:
> > FS_CGROUP_WRITEBACK indicates whether a file_system_type supports
> > cgroup writeback; however, different super_blocks of the same
> > file_system_type may or may not support cgroup writeback depending on
> > filesystem options. This patch replaces FS_CGROUP_WRITEBACK with a
> > kernel-internal super_block->s_flags MS_CGROUPWB. The concatenated
> > and abbreviated name is for consistency with other MS_* flags.
>
> Nak. As the uapi part makes it obvious the MS_ namespace is part
> of the userspace ABI. Please add a new in-kernel flags field instead.
Are MS_ACTIVE and MS_BORN part of userpace ABI? They seem pretty
internal. I don't mind introducing a new internal flag field but it's
weird to put this single flag there with other internal flags in
->s_flags.
Assuming we add a new field, how do sb->s_iflags and SB_I_XXX sound?
Any better suggestions?
Thanks.
--
tejun
On Sun 14-06-15 00:42:36, Tejun Heo wrote:
> Hello, Christoph.
>
> On Sat, Jun 13, 2015 at 09:16:08AM -0700, Christoph Hellwig wrote:
> > On Fri, Jun 12, 2015 at 04:57:33PM -0500, Tejun Heo wrote:
> > > FS_CGROUP_WRITEBACK indicates whether a file_system_type supports
> > > cgroup writeback; however, different super_blocks of the same
> > > file_system_type may or may not support cgroup writeback depending on
> > > filesystem options. This patch replaces FS_CGROUP_WRITEBACK with a
> > > kernel-internal super_block->s_flags MS_CGROUPWB. The concatenated
> > > and abbreviated name is for consistency with other MS_* flags.
> >
> > Nak. As the uapi part makes it obvious the MS_ namespace is part
> > of the userspace ABI. Please add a new in-kernel flags field instead.
>
> Are MS_ACTIVE and MS_BORN part of userpace ABI? They seem pretty
> internal. I don't mind introducing a new internal flag field but it's
> weird to put this single flag there with other internal flags in
> ->s_flags.
So you are right that there are other internal flags allocated from the
top of the i_flags field, however we are pretty much running out of the
flags available for the ABI so it's better to move internal flags elsewhere
as that's simpler than creating a new ABI for mount...
> Assuming we add a new field, how do sb->s_iflags and SB_I_XXX sound?
> Any better suggestions?
Looks good to me.
Honza
--
Jan Kara <[email protected]>
SUSE Labs, CR