Skip to contentSkip to conversation

fs/namespace: notify pollers of legacy propagation changes

PATCH Superseded Guopeng Zhang 2 participants 3 messages May 29 – Jun 1 VFS

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 fs/namespace: notify pollers of legacy propagation changes

From: Guopeng Zhang <[email protected]>

Changing mount propagation through the legacy mount API changes user-visible mountinfo contents, including the shared: and master:
optional fields.

The mount_setattr() path already touches the mount namespace after change_mnt_propagation(), so pollers of /proc/<pid>/mountinfo are woken when the namespace event changes.

The legacy mount --make-* path also changes propagation through change_mnt_propagation(), and MOVE_MOUNT_SET_GROUP updates the propagation relationship of the target mount. Both paths currently return without touching the affected mount namespace.

As a result, userspace polling /proc/<pid>/mountinfo can miss these propagation-only changes even though mountinfo has changed.

A simple reproducer that polls /proc/self/mountinfo while changing propagation shows the inconsistency.

Before this change:

legacy MS_SHARED: poll ret=0 revents=0x0
mount_setattr MS_SHARED: poll ret=1 revents=0xa

After this change:

legacy MS_SHARED: poll ret=1 revents=0xa
mount_setattr MS_SHARED: poll ret=1 revents=0xa

Touch the affected mount namespace after successfully changing propagation state in do_change_type() and do_set_group(). Take the vfsmount lock for write around touch_mnt_namespace(), as required by its locking rules.

---
 fs/namespace.c | 9 +++++++++
 1 file changed, 9 insertions(+)
fs/namespace.c +9 -0
index 9a66a806a9b8..f871c7bf3bc8 100644
@@ -2908,6 +2908,10 @@ static int do_change_type(const struct path *path, int ms_flags)
 	for (m = mnt; m; m = (recurse ? next_mnt(m, mnt) : NULL))
 		change_mnt_propagation(m, type);
 
+	lock_mount_hash();
+	touch_mnt_namespace(mnt->mnt_ns);
+	unlock_mount_hash();
+
 	return 0;
 }
 
@@ -3479,6 +3483,11 @@ static int do_set_group(const struct path *from_path, const struct path *to_path
 		list_add(&to->mnt_share, &from->mnt_share);
 		set_mnt_shared(to);
 	}
+
+	lock_mount_hash();
+	touch_mnt_namespace(to->mnt_ns);
+	unlock_mount_hash();
+
 	return 0;
 }
 
Signature · 1 line
2.43.0
#1 lore raw
Quoted 57 lines · Guopeng Zhang

From: Guopeng Zhang <[email protected]>

Changing mount propagation through the legacy mount API changes user-visible mountinfo contents, including the shared: and master:
optional fields.

The mount_setattr() path already touches the mount namespace after change_mnt_propagation(), so pollers of /proc/<pid>/mountinfo are woken when the namespace event changes.

The legacy mount --make-* path also changes propagation through change_mnt_propagation(), and MOVE_MOUNT_SET_GROUP updates the propagation relationship of the target mount. Both paths currently return without touching the affected mount namespace.

As a result, userspace polling /proc/<pid>/mountinfo can miss these propagation-only changes even though mountinfo has changed.

A simple reproducer that polls /proc/self/mountinfo while changing propagation shows the inconsistency.

Before this change:

legacy MS_SHARED: poll ret=0 revents=0x0
mount_setattr MS_SHARED: poll ret=1 revents=0xa

After this change:

legacy MS_SHARED: poll ret=1 revents=0xa
mount_setattr MS_SHARED: poll ret=1 revents=0xa

Touch the affected mount namespace after successfully changing propagation state in do_change_type() and do_set_group(). Take the vfsmount lock for write around touch_mnt_namespace(), as required by its locking rules.

---
 fs/namespace.c | 9 +++++++++
 1 file changed, 9 insertions(+)
fs/namespace.c +4 -0
index 9a66a806a9b8..f871c7bf3bc8 100644
@@ -2908,6 +2908,10 @@ static int do_change_type(const struct path *path, int ms_flags)
 	for (m = mnt; m; m = (recurse ? next_mnt(m, mnt) : NULL))
 		change_mnt_propagation(m, type);
 
+	lock_mount_hash();
+	touch_mnt_namespace(mnt->mnt_ns);
+	unlock_mount_hash();
+
 	return 0;
 }
 

fs/namespace.c static int do_set_group(const struct path *from_path, const struct path *to_path · line 3489

Quoted 7 lines
fs/namespace.c +4 -0
@@ -3479,6 +3483,11 @@ static int do_set_group(const struct path *from_path, const struct path *to_path
 		list_add(&to->mnt_share, &from->mnt_share);
 		set_mnt_shared(to);
 	}
+
+	lock_mount_hash();
+	touch_mnt_namespace(to->mnt_ns);
+	unlock_mount_hash();

Doing this would cause seqcount readers to retry on mount propagation changes when all of them really only care about mount topology changes. So this can likely use:

guard(mount_locked_reader)();
touch_mnt_namespace(mnt_ns);

Even today, observing an unchanged seqcount across mnt->mnt_flags reads doesn't guarantee that it really wasn't changed.

#2 lore raw
Quoted 70 lines · Christian Brauner
Quoted 57 lines · Guopeng Zhang

From: Guopeng Zhang <[email protected]>

Changing mount propagation through the legacy mount API changes user-visible mountinfo contents, including the shared: and master:
optional fields.

The mount_setattr() path already touches the mount namespace after change_mnt_propagation(), so pollers of /proc/<pid>/mountinfo are woken when the namespace event changes.

The legacy mount --make-* path also changes propagation through change_mnt_propagation(), and MOVE_MOUNT_SET_GROUP updates the propagation relationship of the target mount. Both paths currently return without touching the affected mount namespace.

As a result, userspace polling /proc/<pid>/mountinfo can miss these propagation-only changes even though mountinfo has changed.

A simple reproducer that polls /proc/self/mountinfo while changing propagation shows the inconsistency.

Before this change:

legacy MS_SHARED: poll ret=0 revents=0x0
mount_setattr MS_SHARED: poll ret=1 revents=0xa

After this change:

legacy MS_SHARED: poll ret=1 revents=0xa
mount_setattr MS_SHARED: poll ret=1 revents=0xa

Touch the affected mount namespace after successfully changing propagation state in do_change_type() and do_set_group(). Take the vfsmount lock for write around touch_mnt_namespace(), as required by its locking rules.

---
 fs/namespace.c | 9 +++++++++
 1 file changed, 9 insertions(+)
fs/namespace.c +4 -0
index 9a66a806a9b8..f871c7bf3bc8 100644
@@ -2908,6 +2908,10 @@ static int do_change_type(const struct path *path, int ms_flags)
 	for (m = mnt; m; m = (recurse ? next_mnt(m, mnt) : NULL))
 		change_mnt_propagation(m, type);
 
+	lock_mount_hash();
+	touch_mnt_namespace(mnt->mnt_ns);
+	unlock_mount_hash();
+
 	return 0;
 }
 

fs/namespace.c static int do_set_group(const struct path *from_path, const struct path *to_path · line 3489

Quoted 7 lines
fs/namespace.c +4 -0
@@ -3479,6 +3483,11 @@ static int do_set_group(const struct path *from_path, const struct path *to_path
 		list_add(&to->mnt_share, &from->mnt_share);
 		set_mnt_shared(to);
 	}
+
+	lock_mount_hash();
+	touch_mnt_namespace(to->mnt_ns);
+	unlock_mount_hash();

Doing this would cause seqcount readers to retry on mount propagation changes when all of them really only care about mount topology changes. So this can likely use:

Quoted 5 lines

guard(mount_locked_reader)();
touch_mnt_namespace(mnt_ns);

Even today, observing an unchanged seqcount across mnt->mnt_flags reads doesn't guarantee that it really wasn't changed.

Hi Christian,

Thanks for the review and explanation.

I will send a v2 with the suggested changes.

Thanks,
Guopeng