Skip to contentSkip to conversation

Fix local storage helper OOPs

PATCH Superseded KP Singh 2 participants 8 messages Jan 11, 2021 – Jan 12, 2021

Subsystems (3)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

KP Singh

author
#0 lore raw

v2 cover Fix local storage helper OOPs

It was noted in
https://lore.kernel.org/bpf/CACYkzJ55X8Tp2q4+EFf2hOM_Lysoim1xJY1YdA3k=T3woMW6mg@mail.gmail.com/T/#t
that the local storage helpers do not handle null owner pointers correctly. This patch fixes the task and inode storage helpers with a null check. In order to keep the check explicit, it's kept in the body of the helpers similar to sk_storage and also fixes a minor typo in bpf_inode_storage.c [I did not add a fixes and reported tag to the patch that fixes the typo since it's a non-functional change].

The series also incorporates the example posted by Gilad into the selftest. The selftest, without the fix reproduces the oops and the subsequent patch fixes it.

KP Singh (3):
bpf: update local storage test to check handling of null ptrs
bpf: local storage helpers should check nullness of owner ptr passed
bpf: Fix typo in bpf_inode_storage.c

 kernel/bpf/bpf_inode_storage.c                |  9 +-
 kernel/bpf/bpf_task_storage.c                 |  5 +-
 .../bpf/prog_tests/test_local_storage.c       | 96 +++++--------------
 .../selftests/bpf/progs/local_storage.c       | 62 ++++++------
 4 files changed, 71 insertions(+), 101 deletions(-)
Signature · 1 line
2.30.0.284.gd98b1dd5eaa7-goog
author
#1 lore raw

PATCH v2 1/3 bpf: update local storage test to check handling of null ptrs

It was found in [1] that bpf_inode_storage_get helper did not check the nullness of the passed owner ptr which caused an oops when dereferenced. This change incorporates the example suggested in [1] into the local storage selftest.

The test is updated to create a temporary directory instead of just using a tempfile. In order to replicate the issue this copied rm binary is renamed tiggering the inode_rename with a null pointer for the new_inode. The logic to verify the setting and deletion of the inode local storage of the old inode is also moved to this LSM hook.

The change also removes the copy_rm function and simply shells out to copy files and recursively delete directories and consolidates the logic of setting the initial inode storage to the bprm_committed_creds hook and removes the file_open hook.

[1]: https://lore.kernel.org/bpf/CANaYP3HWkH91SN=wTNO9FL_2ztHfqcXKX38SSE-JJ2voh+vssw@mail.gmail.com

---
 .../bpf/prog_tests/test_local_storage.c       | 96 +++++--------------
 .../selftests/bpf/progs/local_storage.c       | 62 ++++++------
 2 files changed, 61 insertions(+), 97 deletions(-)
tools/testing/selftests/bpf/prog_tests/test_local_storage.c +26 -70
index c0fe73a17ed1..338475fe9ffb 100644
@@ -34,61 +34,6 @@ struct storage {
 	struct bpf_spin_lock lock;
 };
 
-/* Copies an rm binary to a temp file. dest is a mkstemp template */
-static int copy_rm(char *dest)
-{
-	int fd_in, fd_out = -1, ret = 0;
-	struct stat stat;
-	char *buf = NULL;
-
-	fd_in = open("/bin/rm", O_RDONLY);
-	if (fd_in < 0)
-		return -errno;
-
-	fd_out = mkstemp(dest);
-	if (fd_out < 0) {
-		ret = -errno;
-		goto out;
-	}
-
-	ret = fstat(fd_in, &stat);
-	if (ret == -1) {
-		ret = -errno;
-		goto out;
-	}
-
-	buf = malloc(stat.st_blksize);
-	if (!buf) {
-		ret = -errno;
-		goto out;
-	}
-
-	while (ret = read(fd_in, buf, stat.st_blksize), ret > 0) {
-		ret = write(fd_out, buf, ret);
-		if (ret < 0) {
-			ret = -errno;
-			goto out;
-
-		}
-	}
-	if (ret < 0) {
-		ret = -errno;
-		goto out;
-
-	}
-
-	/* Set executable permission on the copied file */
-	ret = chmod(dest, 0100);
-	if (ret == -1)
-		ret = -errno;
-
-out:
-	free(buf);
-	close(fd_in);
-	close(fd_out);
-	return ret;
-}
-
 /* Fork and exec the provided rm binary and return the exit code of the
  * forked process and its pid.
  */
@@ -168,9 +113,11 @@ static bool check_syscall_operations(int map_fd, int obj_fd)
 
 void test_test_local_storage(void)
 {
-	char tmp_exec_path[PATH_MAX] = "/tmp/copy_of_rmXXXXXX";
+	char tmp_dir_path[64] = "/tmp/local_storageXXXXXX";
 	int err, serv_sk = -1, task_fd = -1, rm_fd = -1;
 	struct local_storage *skel = NULL;
+	char tmp_exec_path[64];
+	char cmd[256];
 
 	skel = local_storage__open_and_load();
 	if (CHECK(!skel, "skel_load", "lsm skeleton failed\n"))
@@ -189,18 +136,24 @@ void test_test_local_storage(void)
 				      task_fd))
 		goto close_prog;
 
-	err = copy_rm(tmp_exec_path);
-	if (CHECK(err < 0, "copy_rm", "err %d errno %d\n", err, errno))
+	mkdtemp(tmp_dir_path);
+	if (CHECK(errno < 0, "mkdtemp", "unable to create tmpdir: %d\n", errno))
 		goto close_prog;
 
+	snprintf(tmp_exec_path, sizeof(tmp_exec_path), "%s/copy_of_rm",
+		 tmp_dir_path);
+	snprintf(cmd, sizeof(cmd), "cp /bin/rm %s", tmp_exec_path);
+	if (CHECK_FAIL(system(cmd)))
+		goto close_prog_rmdir;
+
 	rm_fd = open(tmp_exec_path, O_RDONLY);
 	if (CHECK(rm_fd < 0, "open", "failed to open %s err:%d, errno:%d",
 		  tmp_exec_path, rm_fd, errno))
-		goto close_prog;
+		goto close_prog_rmdir;
 
 	if (!check_syscall_operations(bpf_map__fd(skel->maps.inode_storage_map),
 				      rm_fd))
-		goto close_prog;
+		goto close_prog_rmdir;
 
 	/* Sets skel->bss->monitored_pid to the pid of the forked child
 	 * forks a child process that executes tmp_exec_path and tries to
@@ -209,33 +162,36 @@ void test_test_local_storage(void)
 	 */
 	err = run_self_unlink(&skel->bss->monitored_pid, tmp_exec_path);
 	if (CHECK(err != EPERM, "run_self_unlink", "err %d want EPERM\n", err))
-		goto close_prog_unlink;
+		goto close_prog_rmdir;
 
 	/* Set the process being monitored to be the current process */
 	skel->bss->monitored_pid = getpid();
 
-	/* Remove the temporary created executable */
-	err = unlink(tmp_exec_path);
-	if (CHECK(err != 0, "unlink", "unable to unlink %s: %d", tmp_exec_path,
-		  errno))
-		goto close_prog_unlink;
+	/* Move copy_of_rm to a new location so that it triggers the
+	 * inode_rename LSM hook with a new_dentry that has a NULL inode ptr.
+	 */
+	snprintf(cmd, sizeof(cmd), "mv %s/copy_of_rm %s/check_null_ptr",
+		 tmp_dir_path, tmp_dir_path);
+	if (CHECK_FAIL(system(cmd)))
+		goto close_prog_rmdir;
 
 	CHECK(skel->data->inode_storage_result != 0, "inode_storage_result",
 	      "inode_local_storage not set\n");
 
 	serv_sk = start_server(AF_INET6, SOCK_STREAM, NULL, 0, 0);
 	if (CHECK(serv_sk < 0, "start_server", "failed to start server\n"))
-		goto close_prog;
+		goto close_prog_rmdir;
 
 	CHECK(skel->data->sk_storage_result != 0, "sk_storage_result",
 	      "sk_local_storage not set\n");
 
 	if (!check_syscall_operations(bpf_map__fd(skel->maps.sk_storage_map),
 				      serv_sk))
-		goto close_prog;
+		goto close_prog_rmdir;
 
-close_prog_unlink:
-	unlink(tmp_exec_path);
+close_prog_rmdir:
+	snprintf(cmd, sizeof(cmd), "rm -rf %s", tmp_dir_path);
+	system(cmd);
 close_prog:
 	close(serv_sk);
 	close(rm_fd);
tools/testing/selftests/bpf/progs/local_storage.c +35 -27

98 lines not on this page · raw message

Signature · 1 line
2.30.0.284.gd98b1dd5eaa7-goog
Acked-by on v2 1/3
#2 lore raw

Re: [PATCH bpf v2 1/3] bpf: update local storage test to check handling of null ptrs

Quoted 16 lines · KP Singh

It was found in [1] that bpf_inode_storage_get helper did not check the nullness of the passed owner ptr which caused an oops when dereferenced. This change incorporates the example suggested in [1] into the local storage selftest.

The test is updated to create a temporary directory instead of just using a tempfile. In order to replicate the issue this copied rm binary is renamed tiggering the inode_rename with a null pointer for the new_inode. The logic to verify the setting and deletion of the inode local storage of the old inode is also moved to this LSM hook.

The change also removes the copy_rm function and simply shells out to copy files and recursively delete directories and consolidates the logic of setting the initial inode storage to the bprm_committed_creds hook and removes the file_open hook.

Quoted 4 lines

[1]: https://lore.kernel.org/bpf/CANaYP3HWkH91SN=wTNO9FL_2ztHfqcXKX38SSE-JJ2voh+vssw@mail.gmail.com

Ack with one nit below.

Quoted 85 lines
---
  .../bpf/prog_tests/test_local_storage.c       | 96 +++++--------------
  .../selftests/bpf/progs/local_storage.c       | 62 ++++++------
  2 files changed, 61 insertions(+), 97 deletions(-)
tools/testing/selftests/bpf/prog_tests/test_local_storage.c +3 -56
index c0fe73a17ed1..338475fe9ffb 100644
@@ -34,61 +34,6 @@ struct storage {
  	struct bpf_spin_lock lock;
  };
 
-/* Copies an rm binary to a temp file. dest is a mkstemp template */
-static int copy_rm(char *dest)
-{
-	int fd_in, fd_out = -1, ret = 0;
-	struct stat stat;
-	char *buf = NULL;
-
-	fd_in = open("/bin/rm", O_RDONLY);
-	if (fd_in < 0)
-		return -errno;
-
-	fd_out = mkstemp(dest);
-	if (fd_out < 0) {
-		ret = -errno;
-		goto out;
-	}
-
-	ret = fstat(fd_in, &stat);
-	if (ret == -1) {
-		ret = -errno;
-		goto out;
-	}
-
-	buf = malloc(stat.st_blksize);
-	if (!buf) {
-		ret = -errno;
-		goto out;
-	}
-
-	while (ret = read(fd_in, buf, stat.st_blksize), ret > 0) {
-		ret = write(fd_out, buf, ret);
-		if (ret < 0) {
-			ret = -errno;
-			goto out;
-
-		}
-	}
-	if (ret < 0) {
-		ret = -errno;
-		goto out;
-
-	}
-
-	/* Set executable permission on the copied file */
-	ret = chmod(dest, 0100);
-	if (ret == -1)
-		ret = -errno;
-
-out:
-	free(buf);
-	close(fd_in);
-	close(fd_out);
-	return ret;
-}
-
  /* Fork and exec the provided rm binary and return the exit code of the
   * forked process and its pid.
   */
@@ -168,9 +113,11 @@ static bool check_syscall_operations(int map_fd, int obj_fd)
 
  void test_test_local_storage(void)
  {
-	char tmp_exec_path[PATH_MAX] = "/tmp/copy_of_rmXXXXXX";
+	char tmp_dir_path[64] = "/tmp/local_storageXXXXXX";
  	int err, serv_sk = -1, task_fd = -1, rm_fd = -1;
  	struct local_storage *skel = NULL;
+	char tmp_exec_path[64];
+	char cmd[256];
 
  	skel = local_storage__open_and_load();
  	if (CHECK(!skel, "skel_load", "lsm skeleton failed\n"))

tools/testing/selftests/bpf/prog_tests/test_local_storage.c void test_test_local_storage(void) · line 140

Quoted 7 lines
tools/testing/selftests/bpf/prog_tests/test_local_storage.c +2 -2
@@ -189,18 +136,24 @@ void test_test_local_storage(void)
  				      task_fd))
  		goto close_prog;
 
-	err = copy_rm(tmp_exec_path);
-	if (CHECK(err < 0, "copy_rm", "err %d errno %d\n", err, errno))
+	mkdtemp(tmp_dir_path);
+	if (CHECK(errno < 0, "mkdtemp", "unable to create tmpdir: %d\n", errno))

I think checking mkdtemp return value is more reliable than checking errno. It is possible mkdtemp returns 0 and errno is not 0 (inheritted from previous syscall).

Quoted 8 lines
tools/testing/selftests/bpf/prog_tests/test_local_storage.c +6 -0
  		goto close_prog;
 
+	snprintf(tmp_exec_path, sizeof(tmp_exec_path), "%s/copy_of_rm",
+		 tmp_dir_path);
+	snprintf(cmd, sizeof(cmd), "cp /bin/rm %s", tmp_exec_path);
+	if (CHECK_FAIL(system(cmd)))
+		goto close_prog_rmdir;
+

[...]

author
#3 lore raw

Re: [PATCH bpf v2 1/3] bpf: update local storage test to check handling of null ptrs

Quoted 30 lines · Yonghong Song
Quoted 16 lines · KP Singh

It was found in [1] that bpf_inode_storage_get helper did not check the nullness of the passed owner ptr which caused an oops when dereferenced. This change incorporates the example suggested in [1] into the local storage selftest.

The test is updated to create a temporary directory instead of just using a tempfile. In order to replicate the issue this copied rm binary is renamed tiggering the inode_rename with a null pointer for the new_inode. The logic to verify the setting and deletion of the inode local storage of the old inode is also moved to this LSM hook.

The change also removes the copy_rm function and simply shells out to copy files and recursively delete directories and consolidates the logic of setting the initial inode storage to the bprm_committed_creds hook and removes the file_open hook.

Quoted 4 lines

[1]: https://lore.kernel.org/bpf/CANaYP3HWkH91SN=wTNO9FL_2ztHfqcXKX38SSE-JJ2voh+vssw@mail.gmail.com

Ack with one nit below.

Quoted 1 line

---

[...]

Quoted 4 lines
Quoted 4 lines
tools/testing/selftests/bpf/prog_tests/test_local_storage.c +0 -0
@@ -189,18 +136,24 @@ void test_test_local_storage(void)
                                    task_fd))
              goto close_prog;
 
Quoted 8 lines

tools/testing/selftests/bpf/prog_tests/test_local_storage.c void test_test_local_storage(void) · line 140

Quoted 4 lines
tools/testing/selftests/bpf/prog_tests/test_local_storage.c +2 -2
-     err = copy_rm(tmp_exec_path);
-     if (CHECK(err < 0, "copy_rm", "err %d errno %d\n", err, errno))
+     mkdtemp(tmp_dir_path);
+     if (CHECK(errno < 0, "mkdtemp", "unable to create tmpdir: %d\n", errno))

I think checking mkdtemp return value is more reliable than checking errno. It is possible mkdtemp returns 0 and errno is not 0 (inheritted from previous syscall).

You are right, I will send in a v3 with this fixed and also add your Acks. Thanks!

Quoted 10 lines
Quoted 9 lines
tools/testing/selftests/bpf/prog_tests/test_local_storage.c +6 -0
              goto close_prog;
 
+     snprintf(tmp_exec_path, sizeof(tmp_exec_path), "%s/copy_of_rm",
+              tmp_dir_path);
+     snprintf(cmd, sizeof(cmd), "cp /bin/rm %s", tmp_exec_path);
+     if (CHECK_FAIL(system(cmd)))
+             goto close_prog_rmdir;
+

[...]

author
#4 lore raw

PATCH v2 2/3 bpf: local storage helpers should check nullness of owner ptr passed

In reply to KP Singh

The verifier allows ARG_PTR_TO_BTF_ID helper arguments to be NULL, so helper implementations need to check this before dereferencing them. This was already fixed for the socket storage helpers but not for task and inode.

The issue can be reproduced by attaching an LSM program to inode_rename hook (called when moving files) which tries to get the inode of the new file without checking for its nullness and then trying to move an existing file to a new path:

mv existing_file new_file_does_not_exist

The report including the sample program and the steps for reproducing the bug:

https://lore.kernel.org/bpf/CANaYP3HWkH91SN=wTNO9FL_2ztHfqcXKX38SSE-JJ2voh+vssw@mail.gmail.com

Fixes 4cf1bc1f1045 ("bpf: Implement task local storage")
Fixes 8ea636848aca ("bpf: Implement bpf_local_storage for inodes")
---
 kernel/bpf/bpf_inode_storage.c | 5 ++++-
 kernel/bpf/bpf_task_storage.c  | 5 ++++-
 2 files changed, 8 insertions(+), 2 deletions(-)
kernel/bpf/bpf_inode_storage.c +4 -1
index 6edff97ad594..dbc1dbdd2cbf 100644
@@ -176,7 +176,7 @@ BPF_CALL_4(bpf_inode_storage_get, struct bpf_map *, map, struct inode *, inode,
 	 * bpf_local_storage_update expects the owner to have a
 	 * valid storage pointer.
 	 */
-	if (!inode_storage_ptr(inode))
+	if (!inode || !inode_storage_ptr(inode))
 		return (unsigned long)NULL;
 
 	sdata = inode_storage_lookup(inode, map, true);
@@ -200,6 +200,9 @@ BPF_CALL_4(bpf_inode_storage_get, struct bpf_map *, map, struct inode *, inode,
 BPF_CALL_2(bpf_inode_storage_delete,
 	   struct bpf_map *, map, struct inode *, inode)
 {
+	if (!inode)
+		return -EINVAL;
+
 	/* This helper must only called from where the inode is gurranteed
 	 * to have a refcount and cannot be freed.
 	 */
kernel/bpf/bpf_task_storage.c +4 -1
index 4ef1959a78f2..e0da0258b732 100644
@@ -218,7 +218,7 @@ BPF_CALL_4(bpf_task_storage_get, struct bpf_map *, map, struct task_struct *,
 	 * bpf_local_storage_update expects the owner to have a
 	 * valid storage pointer.
 	 */
-	if (!task_storage_ptr(task))
+	if (!task || !task_storage_ptr(task))
 		return (unsigned long)NULL;
 
 	sdata = task_storage_lookup(task, map, true);
@@ -243,6 +243,9 @@ BPF_CALL_4(bpf_task_storage_get, struct bpf_map *, map, struct task_struct *,
 BPF_CALL_2(bpf_task_storage_delete, struct bpf_map *, map, struct task_struct *,
 	   task)
 {
+	if (!task)
+		return -EINVAL;
+
 	/* This helper must only be called from places where the lifetime of the task
 	 * is guaranteed. Either by being refcounted or by being protected
 	 * by an RCU read-side critical section.
Signature · 1 line
2.30.0.284.gd98b1dd5eaa7-goog
Acked-by on v2 2/3
#5 lore raw

Re: [PATCH bpf v2 2/3] bpf: local storage helpers should check nullness of owner ptr passed

Quoted 15 lines · KP Singh

The verifier allows ARG_PTR_TO_BTF_ID helper arguments to be NULL, so helper implementations need to check this before dereferencing them. This was already fixed for the socket storage helpers but not for task and inode.

The issue can be reproduced by attaching an LSM program to inode_rename hook (called when moving files) which tries to get the inode of the new file without checking for its nullness and then trying to move an existing file to a new path:

mv existing_file new_file_does_not_exist

The report including the sample program and the steps for reproducing the bug:

Quoted 7 lines

https://lore.kernel.org/bpf/CANaYP3HWkH91SN=wTNO9FL_2ztHfqcXKX38SSE-JJ2voh+vssw@mail.gmail.com

Fixes 4cf1bc1f1045 ("bpf: Implement task local storage")
Fixes 8ea636848aca ("bpf: Implement bpf_local_storage for inodes")
author
#6 lore raw

PATCH v2 3/3 bpf: Fix typo in bpf_inode_storage.c

In reply to KP Singh

Fix "gurranteed" -> "guaranteed" in bpf_inode_storage.c

---
 kernel/bpf/bpf_inode_storage.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
kernel/bpf/bpf_inode_storage.c +2 -2
index dbc1dbdd2cbf..2f0597320b6d 100644
@@ -183,7 +183,7 @@ BPF_CALL_4(bpf_inode_storage_get, struct bpf_map *, map, struct inode *, inode,
 	if (sdata)
 		return (unsigned long)sdata->data;
 
-	/* This helper must only called from where the inode is gurranteed
+	/* This helper must only called from where the inode is guaranteed
 	 * to have a refcount and cannot be freed.
 	 */
 	if (flags & BPF_LOCAL_STORAGE_GET_F_CREATE) {
@@ -203,7 +203,7 @@ BPF_CALL_2(bpf_inode_storage_delete,
 	if (!inode)
 		return -EINVAL;
 
-	/* This helper must only called from where the inode is gurranteed
+	/* This helper must only called from where the inode is guaranteed
 	 * to have a refcount and cannot be freed.
 	 */
 	return inode_storage_delete(inode, map);
Signature · 1 line
2.30.0.284.gd98b1dd5eaa7-goog