Skip to contentSkip to conversation

fanotify: Disallow permission events for proc filesystem

PATCH In review Jan Kara 2 participants 4 messages May 16, 2019 – May 28, 2019

Subsystems (4)
lore
Share & tools
lore

Jan Kara

author maintainer
#0 lore raw

PATCH v2 1/1 fanotify: Disallow permission events for proc filesystem

Proc filesystem has special locking rules for various files. Thus fanotify which opens files on event delivery can easily deadlock against another process that waits for fanotify permission event to be handled. Since permission events on /proc have doubtful value anyway, just disallow them.

---
 fs/notify/fanotify/fanotify_user.c | 22 ++++++++++++++++++++++
 fs/proc/root.c                     |  2 +-
 include/linux/fs.h                 |  1 +
 3 files changed, 24 insertions(+), 1 deletion(-)

Changes since v1:
* use type flag to detect filesystems not supporting permission events
* return -EINVAL instead of -EOPNOTSUPP for such filesystems

fs/notify/fanotify/fanotify_user.c +22 -0
index a90bb19dcfa2..91006f47e420 100644
@@ -920,6 +920,22 @@ static int fanotify_test_fid(struct path *path, __kernel_fsid_t *fsid)
 	return 0;
 }
 
+static int fanotify_events_supported(struct path *path, __u64 mask)
+{
+	/*
+	 * Some filesystems such as 'proc' acquire unusual locks when opening
+	 * files. For them fanotify permission events have high chances of
+	 * deadlocking the system - open done when reporting fanotify event
+	 * blocks on this "unusual" lock while another process holding the lock
+	 * waits for fanotify permission event to be answered. Just disallow
+	 * permission events for such filesystems.
+	 */
+	if (mask & FANOTIFY_PERM_EVENTS &&
+	    path->mnt->mnt_sb->s_type->fs_flags & FS_DISALLOW_NOTIFY_PERM)
+		return -EINVAL;
+	return 0;
+}
+
 static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 			    int dfd, const char  __user *pathname)
 {
@@ -1018,6 +1034,12 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	if (ret)
 		goto fput_and_out;
 
+	if (flags & FAN_MARK_ADD) {
+		ret = fanotify_events_supported(&path, mask);
+		if (ret)
+			goto path_put_and_out;
+	}
+
 	if (FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
 		ret = fanotify_test_fid(&path, &__fsid);
 		if (ret)
fs/proc/root.c +1 -1
index 8b145e7b9661..522199e9525e 100644
@@ -211,7 +211,7 @@ static struct file_system_type proc_fs_type = {
 	.init_fs_context	= proc_init_fs_context,
 	.parameters		= &proc_fs_parameters,
 	.kill_sb		= proc_kill_sb,
-	.fs_flags		= FS_USERNS_MOUNT,
+	.fs_flags		= FS_USERNS_MOUNT | FS_DISALLOW_NOTIFY_PERM,
 };
 
 void __init proc_root_init(void)
include/linux/fs.h +1 -0
index f7fdfe93e25d..c7136c98b5ba 100644
@@ -2184,6 +2184,7 @@ struct file_system_type {
 #define FS_BINARY_MOUNTDATA	2
 #define FS_HAS_SUBTYPE		4
 #define FS_USERNS_MOUNT		8	/* Can be mounted by userns root */
+#define FS_DISALLOW_NOTIFY_PERM	16	/* Disable fanotify permission events */
 #define FS_RENAME_DOES_D_MOVE	32768	/* FS will handle d_move() during rename() internally. */
 	int (*init_fs_context)(struct fs_context *);
 	const struct fs_parameter_description *parameters;
Signature · 1 line
2.16.4
author maintainer
#1 lore raw
Quoted 8 lines · Jan Kara

Proc filesystem has special locking rules for various files. Thus fanotify which opens files on event delivery can easily deadlock against another process that waits for fanotify permission event to be handled. Since permission events on /proc have doubtful value anyway, just disallow them.

Amir, ping? What do you think about this version of the patch?

Honza

Quoted 78 lines
---
 fs/notify/fanotify/fanotify_user.c | 22 ++++++++++++++++++++++
 fs/proc/root.c                     |  2 +-
 include/linux/fs.h                 |  1 +
 3 files changed, 24 insertions(+), 1 deletion(-)

Changes since v1:
* use type flag to detect filesystems not supporting permission events
* return -EINVAL instead of -EOPNOTSUPP for such filesystems

fs/notify/fanotify/fanotify_user.c +22 -0
index a90bb19dcfa2..91006f47e420 100644
@@ -920,6 +920,22 @@ static int fanotify_test_fid(struct path *path, __kernel_fsid_t *fsid)
 	return 0;
 }
 
+static int fanotify_events_supported(struct path *path, __u64 mask)
+{
+	/*
+	 * Some filesystems such as 'proc' acquire unusual locks when opening
+	 * files. For them fanotify permission events have high chances of
+	 * deadlocking the system - open done when reporting fanotify event
+	 * blocks on this "unusual" lock while another process holding the lock
+	 * waits for fanotify permission event to be answered. Just disallow
+	 * permission events for such filesystems.
+	 */
+	if (mask & FANOTIFY_PERM_EVENTS &&
+	    path->mnt->mnt_sb->s_type->fs_flags & FS_DISALLOW_NOTIFY_PERM)
+		return -EINVAL;
+	return 0;
+}
+
 static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 			    int dfd, const char  __user *pathname)
 {
@@ -1018,6 +1034,12 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	if (ret)
 		goto fput_and_out;
 
+	if (flags & FAN_MARK_ADD) {
+		ret = fanotify_events_supported(&path, mask);
+		if (ret)
+			goto path_put_and_out;
+	}
+
 	if (FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
 		ret = fanotify_test_fid(&path, &__fsid);
 		if (ret)
fs/proc/root.c +1 -1
index 8b145e7b9661..522199e9525e 100644
@@ -211,7 +211,7 @@ static struct file_system_type proc_fs_type = {
 	.init_fs_context	= proc_init_fs_context,
 	.parameters		= &proc_fs_parameters,
 	.kill_sb		= proc_kill_sb,
-	.fs_flags		= FS_USERNS_MOUNT,
+	.fs_flags		= FS_USERNS_MOUNT | FS_DISALLOW_NOTIFY_PERM,
 };
 
 void __init proc_root_init(void)
include/linux/fs.h +1 -0
index f7fdfe93e25d..c7136c98b5ba 100644
@@ -2184,6 +2184,7 @@ struct file_system_type {
 #define FS_BINARY_MOUNTDATA	2
 #define FS_HAS_SUBTYPE		4
 #define FS_USERNS_MOUNT		8	/* Can be mounted by userns root */
+#define FS_DISALLOW_NOTIFY_PERM	16	/* Disable fanotify permission events */
 #define FS_RENAME_DOES_D_MOVE	32768	/* FS will handle d_move() during rename() internally. */
 	int (*init_fs_context)(struct fs_context *);
 	const struct fs_parameter_description *parameters;
Signature · 1 line
2.16.4
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
reviewer Reviewed-by
#2 lore raw
Quoted 8 lines · Jan Kara
Quoted 6 lines · Jan Kara

Proc filesystem has special locking rules for various files. Thus fanotify which opens files on event delivery can easily deadlock against another process that waits for fanotify permission event to be handled. Since permission events on /proc have doubtful value anyway, just disallow them.

Quoted 4 lines
Quoted 2 lines

Amir, ping? What do you think about this version of the patch?

Sorry, I reviewed but forgot to reply.
You may add:

Thanks,
Amir.

author maintainer
#3 lore raw
Quoted 9 lines · Amir Goldstein
Quoted 8 lines · Jan Kara
Quoted 6 lines · Jan Kara

Proc filesystem has special locking rules for various files. Thus fanotify which opens files on event delivery can easily deadlock against another process that waits for fanotify permission event to be handled. Since permission events on /proc have doubtful value anyway, just disallow them.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

Amir, ping? What do you think about this version of the patch?

Sorry, I reviewed but forgot to reply.
You may add:

Thanks! I'll push it to my tree for the next merge window then.

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR