public inbox for libc-stable@sourceware.org
 help / color / mirror / Atom feed
From: Aurelien Jarno <aurelien@aurel32.net>
To: libc-stable@sourceware.org
Cc: Siddhesh Poyarekar <siddhesh@sourceware.org>,
	Andreas Schwab <schwab@linux-m68k.org>,
	Adhemerval Zanella <adhemerval.zanella@linaro.org>,
	Qualys Security Advisory <qsa@qualys.com>
Subject: [COMMITTED 2.33 5/7] getcwd: Set errno to ERANGE for size == 1 (CVE-2021-3999)
Date: Mon, 24 Jan 2022 23:48:55 +0100	[thread overview]
Message-ID: <20220124224857.3799610-5-aurelien@aurel32.net> (raw)
In-Reply-To: <20220124224857.3799610-1-aurelien@aurel32.net>

From: Siddhesh Poyarekar <siddhesh@sourceware.org>

No valid path returned by getcwd would fit into 1 byte, so reject the
size early and return NULL with errno set to ERANGE.  This change is
prompted by CVE-2021-3999, which describes a single byte buffer
underflow and overflow when all of the following conditions are met:

- The buffer size (i.e. the second argument of getcwd) is 1 byte
- The current working directory is too long
- '/' is also mounted on the current working directory

Sequence of events:

- In sysdeps/unix/sysv/linux/getcwd.c, the syscall returns ENAMETOOLONG
  because the linux kernel checks for name length before it checks
  buffer size

- The code falls back to the generic getcwd in sysdeps/posix

- In the generic func, the buf[0] is set to '\0' on line 250

- this while loop on line 262 is bypassed:

    while (!(thisdev == rootdev && thisino == rootino))

  since the rootfs (/) is bind mounted onto the directory and the flow
  goes on to line 449, where it puts a '/' in the byte before the
  buffer.

- Finally on line 458, it moves 2 bytes (the underflowed byte and the
  '\0') to the buf[0] and buf[1], resulting in a 1 byte buffer overflow.

- buf is returned on line 469 and errno is not set.

This resolves BZ #28769.

Reviewed-by: Andreas Schwab <schwab@linux-m68k.org>
Reviewed-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>
Signed-off-by: Qualys Security Advisory <qsa@qualys.com>
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
(cherry picked from commit 23e0e8f5f1fb5ed150253d986ecccdc90c2dcd5e)
---
 NEWS                                          |   7 +
 sysdeps/posix/getcwd.c                        |   7 +
 sysdeps/unix/sysv/linux/Makefile              |   7 +-
 .../unix/sysv/linux/tst-getcwd-smallbuff.c    | 241 ++++++++++++++++++
 4 files changed, 261 insertions(+), 1 deletion(-)
 create mode 100644 sysdeps/unix/sysv/linux/tst-getcwd-smallbuff.c

diff --git a/NEWS b/NEWS
index 5352599d7f..cc026f03b1 100644
--- a/NEWS
+++ b/NEWS
@@ -30,6 +30,12 @@ Security related changes:
   function could result in a memory leak and potential access of
   uninitialized memory.  Reported by Qualys.
 
+  CVE-2021-3999: Passing a buffer of size exactly 1 byte to the getcwd
+  function may result in an off-by-one buffer underflow and overflow
+  when the current working directory is longer than PATH_MAX and also
+  corresponds to the / directory through an unprivileged mount
+  namespace.  Reported by Qualys.
+
 The following bugs are resolved with this release:
 
   [15271] dlfcn function failure after dlmopen terminates process
@@ -50,6 +56,7 @@ The following bugs are resolved with this release:
   [28524] Conversion from ISO-2022-JP-3 with iconv may emit spurious NULs
   [28532] powerpc64[le]: CFI for assembly templated syscalls is incorrect
   [28768] CVE-2022-23218: Buffer overflow in sunrpc svcunix_create
+  [28769] CVE-2021-3999: Off-by-one buffer overflow/underflow in getcwd()
   [28770] CVE-2021-3998: Unexpected return value from realpath() for too long results
 \f
 Version 2.33
diff --git a/sysdeps/posix/getcwd.c b/sysdeps/posix/getcwd.c
index f11644aae7..242e4dbe45 100644
--- a/sysdeps/posix/getcwd.c
+++ b/sysdeps/posix/getcwd.c
@@ -187,6 +187,13 @@ __getcwd_generic (char *buf, size_t size)
   size_t allocated = size;
   size_t used;
 
+  /* A size of 1 byte is never useful.  */
+  if (allocated == 1)
+    {
+      __set_errno (ERANGE);
+      return NULL;
+    }
+
 #if HAVE_MINIMALLY_WORKING_GETCWD
   /* If AT_FDCWD is not defined, the algorithm below is O(N**2) and
      this is much slower than the system getcwd (at least on
diff --git a/sysdeps/unix/sysv/linux/Makefile b/sysdeps/unix/sysv/linux/Makefile
index 472eab700d..9531641f82 100644
--- a/sysdeps/unix/sysv/linux/Makefile
+++ b/sysdeps/unix/sysv/linux/Makefile
@@ -282,7 +282,12 @@ sysdep_routines += xstatconv internal_statvfs internal_statvfs64 \
 
 sysdep_headers += bits/fcntl-linux.h
 
-tests += tst-fallocate tst-fallocate64 tst-o_path-locks
+tests += \
+  tst-fallocate \
+  tst-fallocate64 \
+  tst-getcwd-smallbuff \
+  tst-o_path-locks \
+# tests
 endif
 
 ifeq ($(subdir),elf)
diff --git a/sysdeps/unix/sysv/linux/tst-getcwd-smallbuff.c b/sysdeps/unix/sysv/linux/tst-getcwd-smallbuff.c
new file mode 100644
index 0000000000..d460d6e766
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/tst-getcwd-smallbuff.c
@@ -0,0 +1,241 @@
+/* Verify that getcwd returns ERANGE for size 1 byte and does not underflow
+   buffer when the CWD is too long and is also a mount target of /.  See bug
+   #28769 or CVE-2021-3999 for more context.
+   Copyright The GNU Toolchain Authors.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <errno.h>
+#include <fcntl.h>
+#include <intprops.h>
+#include <limits.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mount.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+
+#include <sys/socket.h>
+#include <sys/un.h>
+#include <support/check.h>
+#include <support/temp_file.h>
+#include <support/xsched.h>
+#include <support/xunistd.h>
+
+static char *base;
+#define BASENAME "tst-getcwd-smallbuff"
+#define MOUNT_NAME "mpoint"
+static int sockfd[2];
+
+static void
+do_cleanup (void)
+{
+  support_chdir_toolong_temp_directory (base);
+  TEST_VERIFY_EXIT (rmdir (MOUNT_NAME) == 0);
+  free (base);
+}
+
+static void
+send_fd (const int sock, const int fd)
+{
+  struct msghdr msg = {0};
+  union
+    {
+      struct cmsghdr hdr;
+      char buf[CMSG_SPACE (sizeof (int))];
+    } cmsgbuf = {0};
+  struct cmsghdr *cmsg;
+  struct iovec vec;
+  char ch = 'A';
+  ssize_t n;
+
+  msg.msg_control = &cmsgbuf.buf;
+  msg.msg_controllen = sizeof (cmsgbuf.buf);
+
+  cmsg = CMSG_FIRSTHDR (&msg);
+  cmsg->cmsg_len = CMSG_LEN (sizeof (int));
+  cmsg->cmsg_level = SOL_SOCKET;
+  cmsg->cmsg_type = SCM_RIGHTS;
+  memcpy (CMSG_DATA (cmsg), &fd, sizeof (fd));
+
+  vec.iov_base = &ch;
+  vec.iov_len = 1;
+  msg.msg_iov = &vec;
+  msg.msg_iovlen = 1;
+
+  while ((n = sendmsg (sock, &msg, 0)) == -1 && errno == EINTR);
+
+  TEST_VERIFY_EXIT (n == 1);
+}
+
+static int
+recv_fd (const int sock)
+{
+  struct msghdr msg = {0};
+  union
+    {
+      struct cmsghdr hdr;
+      char buf[CMSG_SPACE(sizeof(int))];
+    } cmsgbuf = {0};
+  struct cmsghdr *cmsg;
+  struct iovec vec;
+  ssize_t n;
+  char ch = '\0';
+  int fd = -1;
+
+  vec.iov_base = &ch;
+  vec.iov_len = 1;
+  msg.msg_iov = &vec;
+  msg.msg_iovlen = 1;
+
+  msg.msg_control = &cmsgbuf.buf;
+  msg.msg_controllen = sizeof (cmsgbuf.buf);
+
+  while ((n = recvmsg (sock, &msg, 0)) == -1 && errno == EINTR);
+  if (n != 1 || ch != 'A')
+    return -1;
+
+  cmsg = CMSG_FIRSTHDR (&msg);
+  if (cmsg == NULL)
+    return -1;
+  if (cmsg->cmsg_type != SCM_RIGHTS)
+    return -1;
+  memcpy (&fd, CMSG_DATA (cmsg), sizeof (fd));
+  if (fd < 0)
+    return -1;
+  return fd;
+}
+
+static int
+child_func (void * const arg)
+{
+  xclose (sockfd[0]);
+  const int sock = sockfd[1];
+  char ch;
+
+  TEST_VERIFY_EXIT (read (sock, &ch, 1) == 1);
+  TEST_VERIFY_EXIT (ch == '1');
+
+  if (mount ("/", MOUNT_NAME, NULL, MS_BIND | MS_REC, NULL))
+    FAIL_EXIT1 ("mount failed: %m\n");
+  const int fd = xopen ("mpoint",
+			O_RDONLY | O_PATH | O_DIRECTORY | O_NOFOLLOW, 0);
+
+  send_fd (sock, fd);
+  xclose (fd);
+
+  TEST_VERIFY_EXIT (read (sock, &ch, 1) == 1);
+  TEST_VERIFY_EXIT (ch == 'a');
+
+  xclose (sock);
+  return 0;
+}
+
+static void
+update_map (char * const mapping, const char * const map_file)
+{
+  const size_t map_len = strlen (mapping);
+
+  const int fd = xopen (map_file, O_WRONLY, 0);
+  xwrite (fd, mapping, map_len);
+  xclose (fd);
+}
+
+static void
+proc_setgroups_write (const long child_pid, const char * const str)
+{
+  const size_t str_len = strlen(str);
+
+  char setgroups_path[sizeof ("/proc//setgroups") + INT_STRLEN_BOUND (long)];
+
+  snprintf (setgroups_path, sizeof (setgroups_path),
+	    "/proc/%ld/setgroups", child_pid);
+
+  const int fd = open (setgroups_path, O_WRONLY);
+
+  if (fd < 0)
+    {
+      TEST_VERIFY_EXIT (errno == ENOENT);
+      FAIL_UNSUPPORTED ("/proc/%ld/setgroups not found\n", child_pid);
+    }
+
+  xwrite (fd, str, str_len);
+  xclose(fd);
+}
+
+static char child_stack[1024 * 1024];
+
+int
+do_test (void)
+{
+  base = support_create_and_chdir_toolong_temp_directory (BASENAME);
+
+  xmkdir (MOUNT_NAME, S_IRWXU);
+  atexit (do_cleanup);
+
+  TEST_VERIFY_EXIT (socketpair (AF_UNIX, SOCK_STREAM, 0, sockfd) == 0);
+  pid_t child_pid = xclone (child_func, NULL, child_stack,
+			    sizeof (child_stack),
+			    CLONE_NEWUSER | CLONE_NEWNS | SIGCHLD);
+
+  xclose (sockfd[1]);
+  const int sock = sockfd[0];
+
+  char map_path[sizeof ("/proc//uid_map") + INT_STRLEN_BOUND (long)];
+  char map_buf[sizeof ("0  1") + INT_STRLEN_BOUND (long)];
+
+  snprintf (map_path, sizeof (map_path), "/proc/%ld/uid_map",
+	    (long) child_pid);
+  snprintf (map_buf, sizeof (map_buf), "0 %ld 1", (long) getuid());
+  update_map (map_buf, map_path);
+
+  proc_setgroups_write ((long) child_pid, "deny");
+  snprintf (map_path, sizeof (map_path), "/proc/%ld/gid_map",
+	    (long) child_pid);
+  snprintf (map_buf, sizeof (map_buf), "0 %ld 1", (long) getgid());
+  update_map (map_buf, map_path);
+
+  TEST_VERIFY_EXIT (send (sock, "1", 1, MSG_NOSIGNAL) == 1);
+  const int fd = recv_fd (sock);
+  TEST_VERIFY_EXIT (fd >= 0);
+  TEST_VERIFY_EXIT (fchdir (fd) == 0);
+
+  static char buf[2 * 10 + 1];
+  memset (buf, 'A', sizeof (buf));
+
+  /* Finally, call getcwd and check if it resulted in a buffer underflow.  */
+  char * cwd = getcwd (buf + sizeof (buf) / 2, 1);
+  TEST_VERIFY (cwd == NULL);
+  TEST_VERIFY (errno == ERANGE);
+
+  for (int i = 0; i < sizeof (buf); i++)
+    if (buf[i] != 'A')
+      {
+	printf ("buf[%d] = %02x\n", i, (unsigned int) buf[i]);
+	support_record_failure ();
+      }
+
+  TEST_VERIFY_EXIT (send (sock, "a", 1, MSG_NOSIGNAL) == 1);
+  xclose (sock);
+  TEST_VERIFY_EXIT (xwaitpid (child_pid, NULL, 0) == child_pid);
+
+  return 0;
+}
+
+#define CLEANUP_HANDLER do_cleanup
+#include <support/test-driver.c>
-- 
2.34.1


  parent reply	other threads:[~2022-01-24 22:49 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-24 22:48 [COMMITTED 2.33 1/7] support: Add helpers to create paths longer than PATH_MAX Aurelien Jarno
2022-01-24 22:48 ` [COMMITTED 2.33 2/7] realpath: Set errno to ENAMETOOLONG for result larger than PATH_MAX [BZ #28770] Aurelien Jarno
2022-01-24 22:48 ` [COMMITTED 2.33 3/7] tst-realpath-toolong: Fix hurd build Aurelien Jarno
2022-01-24 22:48 ` [COMMITTED 2.33 4/7] support: Add xclone Aurelien Jarno
2022-01-24 22:48 ` Aurelien Jarno [this message]
2022-01-24 22:48 ` [COMMITTED 2.33 6/7] realpath: Avoid overwriting preexisting error (CVE-2021-3998) Aurelien Jarno
2022-01-24 22:48 ` [COMMITTED 2.33 7/7] Linux: Detect user namespace support in io/tst-getcwd-smallbuff Aurelien Jarno

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220124224857.3799610-5-aurelien@aurel32.net \
    --to=aurelien@aurel32.net \
    --cc=adhemerval.zanella@linaro.org \
    --cc=libc-stable@sourceware.org \
    --cc=qsa@qualys.com \
    --cc=schwab@linux-m68k.org \
    --cc=siddhesh@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).