Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
Date:   Sat, 16 Jun 2018 10:49:46 -0700
In-Reply-To: <20180616083740.GA23039@kroah.com>
Message-Id: <20180616174946.175109-1-rajatja@google.com>
References: <20180616083740.GA23039@kroah.com>
Subject: [PATCH v3] sysfs: Fix internal_create_group() for named group updates
From:   Rajat Jain <rajatja@google.com>
To:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        linux-kernel@vger.kernel.org
Cc:     Rajat Jain <rajatja@google.com>, rajatxjain@gmail.com
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

There are a couple of problems with named group updates in the code
today:

* sysfs_update_group() will always fail for a named group, because
  internal_create_group() will try to create a new sysfs directory
  unconditionally, which will ofcourse fail with -EEXIST.

* We can leak the kernfs_node for grp->name if some one tries to:
  - rename a group (change grp->name), or
  - update a named group, to an unnamed group

It appears that the whole purpose of sysfs_update_group() was to
allow changing the permissions or visibility of attributes and not
the names. So make it clear in the comments, and allow it to update
an existing named group.

Signed-off-by: Rajat Jain <rajatja@google.com>
---
v2: Use pr_warn() instead of WARN()
v3: drop the extra reference taken by kernfs_find_and_get()

 fs/sysfs/group.c | 29 ++++++++++++++++++++++-------
 1 file changed, 22 insertions(+), 7 deletions(-)

diff --git a/fs/sysfs/group.c b/fs/sysfs/group.c
index 4802ec0e1e3a..38240410f831 100644
--- a/fs/sysfs/group.c
+++ b/fs/sysfs/group.c
@@ -119,12 +119,22 @@ static int internal_create_group(struct kobject *kobj, int update,
 		return -EINVAL;
 	}
 	if (grp->name) {
-		kn = kernfs_create_dir(kobj->sd, grp->name,
-				       S_IRWXU | S_IRUGO | S_IXUGO, kobj);
-		if (IS_ERR(kn)) {
-			if (PTR_ERR(kn) == -EEXIST)
-				sysfs_warn_dup(kobj->sd, grp->name);
-			return PTR_ERR(kn);
+		if (update) {
+			kn = kernfs_find_and_get(kobj->sd, grp->name);
+			if (!kn) {
+				pr_warn("Can't update unknown attr grp name: %s/%s\n",
+					kobj->name, grp->name);
+				return -EINVAL;
+			}
+		} else {
+			kn = kernfs_create_dir(kobj->sd, grp->name,
+					       S_IRWXU | S_IRUGO | S_IXUGO,
+					       kobj);
+			if (IS_ERR(kn)) {
+				if (PTR_ERR(kn) == -EEXIST)
+					sysfs_warn_dup(kobj->sd, grp->name);
+				return PTR_ERR(kn);
+			}
 		}
 	} else
 		kn = kobj->sd;
@@ -135,6 +145,10 @@ static int internal_create_group(struct kobject *kobj, int update,
 			kernfs_remove(kn);
 	}
 	kernfs_put(kn);
+
+	if (grp->name && update)
+		kernfs_put(kn);
+
 	return error;
 }
 
@@ -199,7 +213,8 @@ EXPORT_SYMBOL_GPL(sysfs_create_groups);
  * of the attribute files being created already exist.  Furthermore,
  * if the visibility of the files has changed through the is_visible()
  * callback, it will update the permissions and add or remove the
- * relevant files.
+ * relevant files. Changing a group's name (subdirectory name under
+ * kobj's directory in sysfs) is not allowed.
  *
  * The primary use for this function is to call it after making a change
  * that affects group visibility.
-- 
2.18.0.rc1.244.gcf134e6275-goog