2009-10-30 05:38:41

by Kusanagi Kouichi

Subject: [PATCH] fs: Don't ignore MS_RDONLY in bind and not remount case.

A readonly bind mount needed to be created in 2 steps

# mount --bind
# mount --bind -o remount,ro

instead of single

# mount --bind -o ro

because MS_RDONLY is ignored in bind and not remount case.

Signed-off-by: Kusanagi Kouichi <[email protected]>
---
fs/namespace.c | 14 +++++++++-----
fs/pnode.h | 1 +
2 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/fs/namespace.c b/fs/namespace.c
index bdc3cb4..1d350e2 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -562,6 +562,8 @@ static struct vfsmount *clone_mnt(struct vfsmount *old, struct dentry *root,
}

mnt->mnt_flags = old->mnt_flags;
+ if (flag & CL_READONLY)
+ mnt->mnt_flags |= MNT_READONLY;
atomic_inc(&sb->s_active);
mnt->mnt_sb = sb;
mnt->mnt_root = dget(root);
@@ -1447,7 +1449,7 @@ static int do_change_type(struct path *path, int flag)
* do loopback mount.
*/
static int do_loopback(struct path *path, char *old_name,
- int recurse)
+ int flags)
{
struct path old_path;
struct vfsmount *mnt = NULL;
@@ -1469,10 +1471,12 @@ static int do_loopback(struct path *path, char *old_name,
goto out;

err = -ENOMEM;
- if (recurse)
- mnt = copy_tree(old_path.mnt, old_path.dentry, 0);
+ if (flags & MS_REC)
+ mnt = copy_tree(old_path.mnt, old_path.dentry,
+ flags & MS_RDONLY ? CL_READONLY : 0);
else
- mnt = clone_mnt(old_path.mnt, old_path.dentry, 0);
+ mnt = clone_mnt(old_path.mnt, old_path.dentry,
+ flags & MS_RDONLY ? CL_READONLY : 0);

if (!mnt)
goto out;
@@ -1959,7 +1963,7 @@ long do_mount(char *dev_name, char *dir_name, char *type_page,
retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags,
data_page);
else if (flags & MS_BIND)
- retval = do_loopback(&path, dev_name, flags & MS_REC);
+ retval = do_loopback(&path, dev_name, flags);
else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE | MS_UNBINDABLE))
retval = do_change_type(&path, flags);
else if (flags & MS_MOVE)
diff --git a/fs/pnode.h b/fs/pnode.h
index 958665d..59769bb 100644
--- a/fs/pnode.h
+++ b/fs/pnode.h
@@ -23,6 +23,7 @@
#define CL_MAKE_SHARED 0x08
#define CL_PROPAGATION 0x10
#define CL_PRIVATE 0x20
+#define CL_READONLY 0x40

static inline void set_mnt_shared(struct vfsmount *mnt)
{
--
1.6.5

2009-10-30 05:53:34

by Christoph Hellwig

Subject: Re: [PATCH] fs: Don't ignore MS_RDONLY in bind and not remount case.

On Fri, Oct 30, 2009 at 02:31:46PM +0900, Kusanagi Kouichi wrote:
> A readonly bind mount needed to be created in 2 steps
>
> # mount --bind
> # mount --bind -o remount,ro
>
> instead of single
>
> # mount --bind -o ro
>
> because MS_RDONLY is ignored in bind and not remount case.

I'm pretty sure that was intentional, but I don't remember why anymore.

2009-10-30 08:56:41

by Kusanagi Kouichi

Subject: Re: [PATCH] fs: Don't ignore MS_RDONLY in bind and not remount case.

On 2009-10-30 01:53:37 -0400, Christoph Hellwig wrote:
> On Fri, Oct 30, 2009 at 02:31:46PM +0900, Kusanagi Kouichi wrote:
> > A readonly bind mount needed to be created in 2 steps
> >
> > # mount --bind
> > # mount --bind -o remount,ro
> >
> > instead of single
> >
> > # mount --bind -o ro
> >
> > because MS_RDONLY is ignored in bind and not remount case.
>
> I'm pretty sure that was intentional, but I don't remember why anymore.

I could not find any reason to restrict this.
Could you tell me the reason of this restriction?