Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 13 Jul 2021 16:20:04 +0000
Message-Id: <20210713162008.1056986-1-kaleshsingh@google.com>
Mime-Version: 1.0
Subject: [PATCH v2] procfs: Prevent unpriveleged processes accessing fdinfo dir
From:   Kalesh Singh <kaleshsingh@google.com>
Cc:     hridya@google.com, surenb@google.com, ebiederm@xmission.com,
        christian.brauner@ubuntu.com, torvalds@linux-foundation.org,
        christian.koenig@amd.com, kernel-team@android.com,
        Kalesh Singh <kaleshsingh@google.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
To:     unlisted-recipients:; (no To-header on input)
Precedence: bulk

The file permissions on the fdinfo dir from were changed from
S_IRUSR|S_IXUSR to S_IRUGO|S_IXUGO, and a PTRACE_MODE_READ check was
added for opening the fdinfo files [1]. However, the ptrace permission
check was not added to the directory, allowing anyone to get the open FD
numbers by reading the fdinfo directory.

Add the missing ptrace permission check for opening the fdinfo directory.

[1] https://lkml.kernel.org/r/20210308170651.919148-1-kaleshsingh@google.com

Fixes: 7bc3fa0172a4 ("procfs: allow reading fdinfo with PTRACE_MODE_READ")
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
---
v1 of this patch was posted at:
https://lore.kernel.org/r/20210708155647.44208-1-kaleshsingh@google.com/

Changes in v2:
 - Drop the ptrace checks from read and lseek ops. The problem of accessing
   already opened files after a suid exec is pre-existing and not unique to
   fdinfo. Needs to be addressed separately in a more generic way.

 fs/proc/fd.c | 23 ++++++++++++++++++++++-
 1 file changed, 22 insertions(+), 1 deletion(-)

diff --git a/fs/proc/fd.c b/fs/proc/fd.c
index 172c86270b31..913bef0d2a36 100644
--- a/fs/proc/fd.c
+++ b/fs/proc/fd.c
@@ -72,7 +72,7 @@ static int seq_show(struct seq_file *m, void *v)
 	return 0;
 }
 
-static int seq_fdinfo_open(struct inode *inode, struct file *file)
+static int proc_fdinfo_access_allowed(struct inode *inode)
 {
 	bool allowed = false;
 	struct task_struct *task = get_proc_task(inode);
@@ -86,6 +86,16 @@ static int seq_fdinfo_open(struct inode *inode, struct file *file)
 	if (!allowed)
 		return -EACCES;
 
+	return 0;
+}
+
+static int seq_fdinfo_open(struct inode *inode, struct file *file)
+{
+	int ret = proc_fdinfo_access_allowed(inode);
+
+	if (ret)
+		return ret;
+
 	return single_open(file, seq_show, inode);
 }
 
@@ -348,12 +358,23 @@ static int proc_readfdinfo(struct file *file, struct dir_context *ctx)
 				  proc_fdinfo_instantiate);
 }
 
+static int proc_open_fdinfo(struct inode *inode, struct file *file)
+{
+	int ret = proc_fdinfo_access_allowed(inode);
+
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
 const struct inode_operations proc_fdinfo_inode_operations = {
 	.lookup		= proc_lookupfdinfo,
 	.setattr	= proc_setattr,
 };
 
 const struct file_operations proc_fdinfo_operations = {
+	.open		= proc_open_fdinfo,
 	.read		= generic_read_dir,
 	.iterate_shared	= proc_readfdinfo,
 	.llseek		= generic_file_llseek,

base-commit: 7fef2edf7cc753b51f7ccc74993971b0a9c81eca
-- 
2.32.0.93.g670b81a890-goog