On Fri, 22 Oct 2010, Miklos Szeredi wrote:
> > Or, more subtle, tries to
> > update atime on some opened struct file on that sb.
>
> Hmm, load_nls() will apparently do that. Drat.
>
> Plan B: remount all vfsmounts read-only before ->remount_fs() but
> remember which ones were read-write and restore on failure. Will
> result in ugly transient write failures if remount_fs() fails, but I
> don't think anybody would care in practice.
And it's actually a pretty simple change. Here's the incremental
patch to "vfs: protect remounting superblock read-only", with "vfs:
allow mnt_want_write() to sleep" reverted.
The full, updated patch with follow.
Thanks,
Miklos
Index: linux-2.6/fs/namespace.c
===================================================================
--- linux-2.6.orig/fs/namespace.c 2010-10-22 16:24:41.000000000 +0200
+++ linux-2.6/fs/namespace.c 2010-10-22 16:17:59.000000000 +0200
@@ -435,11 +435,17 @@ int sb_prepare_remount_readonly(struct s
}
}
}
+ list_for_each_entry(mnt, &sb->s_mounts, mnt_instance) {
+ if (mnt->mnt_flags & MNT_WRITE_HOLD) {
+ if (!err) {
+ mnt->mnt_flags |= MNT_READONLY | MNT_WAS_WRITE;
+ smp_wmb();
+ }
+ mnt->mnt_flags &= ~MNT_WRITE_HOLD;
+ }
+ }
br_write_unlock(vfsmount_lock);
- if (err)
- sb_cancel_remount_readonly(sb);
-
return err;
}
@@ -449,12 +455,11 @@ void sb_cancel_remount_readonly(struct s
br_write_lock(vfsmount_lock);
list_for_each_entry(mnt, &sb->s_mounts, mnt_instance) {
- if (mnt->mnt_flags & MNT_WRITE_HOLD)
- mnt->mnt_flags &= ~MNT_WRITE_HOLD;
+ if (mnt->mnt_flags & MNT_WAS_WRITE)
+ mnt->mnt_flags &= ~(MNT_WAS_WRITE | MNT_READONLY);
}
br_write_unlock(vfsmount_lock);
- wake_up_all(&sb->s_wait_remount_readonly);
}
void sb_finish_remount_readonly(struct super_block *sb)
@@ -463,15 +468,11 @@ void sb_finish_remount_readonly(struct s
br_write_lock(vfsmount_lock);
list_for_each_entry(mnt, &sb->s_mounts, mnt_instance) {
- if (!(mnt->mnt_flags & MNT_READONLY)) {
- mnt->mnt_flags |= MNT_READONLY;
- smp_wmb();
- mnt->mnt_flags &= ~MNT_WRITE_HOLD;
- }
+ if (mnt->mnt_flags & MNT_WAS_WRITE)
+ mnt->mnt_flags &= ~MNT_WAS_WRITE;
}
sb->s_flags |= MS_RDONLY;
br_write_unlock(vfsmount_lock);
- wake_up_all(&sb->s_wait_remount_readonly);
}
void simple_set_mnt(struct vfsmount *mnt, struct super_block *sb)
Index: linux-2.6/include/linux/mount.h
===================================================================
--- linux-2.6.orig/include/linux/mount.h 2010-10-22 16:02:14.000000000 +0200
+++ linux-2.6/include/linux/mount.h 2010-10-22 16:20:26.000000000 +0200
@@ -30,6 +30,7 @@ struct mnt_namespace;
#define MNT_SHRINKABLE 0x100
#define MNT_WRITE_HOLD 0x200
+#define MNT_WAS_WRITE 0x400 /* used by remount to remember write mounts */
#define MNT_SHARED 0x1000 /* if the vfsmount is a shared mount */
#define MNT_UNBINDABLE 0x2000 /* if the vfsmount is a unbindable mount */
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
