[LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Tudor Cretu]

Accessing elements in an empty va_list results in undefined behaviour[0]
that can include accessing arbitrary stack memory. While typically this
doesn't raise a fault, some new more security-oriented architectures
(e.g. CHERI[1] or Morello[2]) don't allow it.

Therefore, remove the variadicness from safe_* wrappers that always call
the functions with the optional argument included.

Adapt the respective SAFE_* macros to handle the change by passing a
default argument if they're omitted.

[0]: [ISO/IEC 9899:2011] Programming Languages—C, 3rd ed, paragraph 7.16.1.1
[1]: https://www.cl.cam.ac.uk/research/security/ctsrd/cheri/
[2]: https://www.morello-project.org/

Tudor Cretu (3):
  safe_open: Fix undefined behaviour in vararg handling
  safe_openat: Fix undefined behaviour in vararg handling
  safe_semctl: Fix undefined behaviour in vararg handling

 include/old/safe_macros.h   |  6 ++++--
 include/safe_macros_fn.h    |  3 ++-
 include/tst_safe_file_at.h  | 10 ++++++----
 include/tst_safe_macros.h   |  6 ++++--
 include/tst_safe_sysv_ipc.h | 14 +++++++++-----
 lib/safe_macros.c           | 13 +------------
 lib/tst_cgroup.c            |  2 +-
 lib/tst_safe_file_at.c      | 11 +++--------
 lib/tst_safe_sysv_ipc.c     | 10 +---------
 9 files changed, 31 insertions(+), 44 deletions(-)

-- 
2.25.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

[LTP] [PATCH 1/3] safe_open: Fix undefined behaviour in vararg handling

[Tudor Cretu]

Accessing elements in an empty va_list is undefined behaviour.
Therefore, remove the variadicness from safe_open as it always calls
open with the mode argument included.

Adapt the SAFE_OPEN macro to handle the change by passing a default
argument of 0 to mode if it's omitted.

Signed-off-by: Tudor Cretu <tudor.cretu@arm.com>
---
 include/old/safe_macros.h |  6 ++++--
 include/safe_macros_fn.h  |  3 ++-
 include/tst_safe_macros.h |  6 ++++--
 lib/safe_macros.c         | 13 +------------
 4 files changed, 11 insertions(+), 17 deletions(-)

diff --git a/include/old/safe_macros.h b/include/old/safe_macros.h
index fb1d7a110..d16540d63 100644
--- a/include/old/safe_macros.h
+++ b/include/old/safe_macros.h
@@ -59,9 +59,11 @@
 #define SAFE_MUNMAP(cleanup_fn, addr, length)	\
 	safe_munmap(__FILE__, __LINE__, (cleanup_fn), (addr), (length))
 
+#define __SAFE_OPEN(cleanup_fn, pathname, oflags, mode, ...)	\
+	safe_open(__FILE__, __LINE__, (cleanup_fn), (pathname), (oflags), (mode))
+
 #define SAFE_OPEN(cleanup_fn, pathname, oflags, ...)	\
-	safe_open(__FILE__, __LINE__, (cleanup_fn), (pathname), (oflags), \
-	    ##__VA_ARGS__)
+	__SAFE_OPEN((cleanup_fn), (pathname), (oflags), ##__VA_ARGS__, 0)
 
 #define SAFE_PIPE(cleanup_fn, fildes)	\
 	safe_pipe(__FILE__, __LINE__, cleanup_fn, (fildes))
diff --git a/include/safe_macros_fn.h b/include/safe_macros_fn.h
index 114d8fd43..d143079c3 100644
--- a/include/safe_macros_fn.h
+++ b/include/safe_macros_fn.h
@@ -74,7 +74,8 @@ int safe_munmap(const char *file, const int lineno,
                 void (*cleanup_fn)(void), void *addr, size_t length);
 
 int safe_open(const char *file, const int lineno,
-              void (*cleanup_fn)(void), const char *pathname, int oflags, ...);
+              void (*cleanup_fn)(void), const char *pathname, int oflags,
+              mode_t mode);
 
 int safe_pipe(const char *file, const int lineno,
               void (*cleanup_fn)(void), int fildes[2]);
diff --git a/include/tst_safe_macros.h b/include/tst_safe_macros.h
index 81c4b0844..d53555c88 100644
--- a/include/tst_safe_macros.h
+++ b/include/tst_safe_macros.h
@@ -86,9 +86,11 @@ void *safe_realloc(const char *file, const int lineno, void *ptr, size_t size);
 #define SAFE_MUNMAP(addr, length) \
 	safe_munmap(__FILE__, __LINE__, NULL, (addr), (length))
 
+#define __SAFE_OPEN(pathname, oflags, mode, ...) \
+	safe_open(__FILE__, __LINE__, NULL, (pathname), (oflags), (mode))
+
 #define SAFE_OPEN(pathname, oflags, ...) \
-	safe_open(__FILE__, __LINE__, NULL, (pathname), (oflags), \
-	    ##__VA_ARGS__)
+	__SAFE_OPEN((pathname), (oflags), ##__VA_ARGS__, 0)
 
 #define SAFE_PIPE(fildes) \
 	safe_pipe(__FILE__, __LINE__, NULL, (fildes))
diff --git a/lib/safe_macros.c b/lib/safe_macros.c
index d8816631f..a92b58347 100644
--- a/lib/safe_macros.c
+++ b/lib/safe_macros.c
@@ -234,20 +234,9 @@ int safe_munmap(const char *file, const int lineno, void (*cleanup_fn) (void),
 }
 
 int safe_open(const char *file, const int lineno, void (*cleanup_fn) (void),
-              const char *pathname, int oflags, ...)
+              const char *pathname, int oflags, mode_t mode)
 {
-	va_list ap;
 	int rval;
-	mode_t mode;
-
-	va_start(ap, oflags);
-
-	/* Android's NDK's mode_t is smaller than an int, which results in
-	 * SIGILL here when passing the mode_t type.
-	 */
-	mode = va_arg(ap, int);
-
-	va_end(ap);
 
 	rval = open(pathname, oflags, mode);
 
-- 
2.25.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

[LTP] [PATCH 2/3] safe_openat: Fix undefined behaviour in vararg handling

[Tudor Cretu]

Accessing elements in an empty va_list is undefined behaviour.
Therefore, remove the variadicness from safe_openat as it always calls
openat with the mode argument included.

Adapt the SAFE_OPENAT macro to handle the change by passing a default
argument of 0 to mode if it's omitted.

Signed-off-by: Tudor Cretu <tudor.cretu@arm.com>
---
 include/tst_safe_file_at.h | 10 ++++++----
 lib/tst_cgroup.c           |  2 +-
 lib/tst_safe_file_at.c     | 11 +++--------
 3 files changed, 10 insertions(+), 13 deletions(-)

diff --git a/include/tst_safe_file_at.h b/include/tst_safe_file_at.h
index a1aa19fad..dd43d8f65 100644
--- a/include/tst_safe_file_at.h
+++ b/include/tst_safe_file_at.h
@@ -11,9 +11,11 @@
 #include <unistd.h>
 #include <stdarg.h>
 
-#define SAFE_OPENAT(dirfd, path, oflags, ...)			\
-	safe_openat(__FILE__, __LINE__,				\
-		    (dirfd), (path), (oflags), ## __VA_ARGS__)
+#define __SAFE_OPENAT(dirfd, path, oflags, mode, ...)			\
+	safe_openat(__FILE__, __LINE__, (dirfd), (path), (oflags), (mode))
+
+#define SAFE_OPENAT(dirfd, path, oflags, ...)				\
+	__SAFE_OPENAT((dirfd), (path), (oflags), ##__VA_ARGS__, 0)
 
 #define SAFE_FILE_READAT(dirfd, path, buf, nbyte)			\
 	safe_file_readat(__FILE__, __LINE__,				\
@@ -38,7 +40,7 @@ const char *tst_decode_fd(const int fd)
 			  __attribute__((warn_unused_result));
 
 int safe_openat(const char *const file, const int lineno, const int dirfd,
-                const char *const path, const int oflags, ...)
+                const char *const path, const int oflags, const mode_t mode)
 		__attribute__((nonnull, warn_unused_result));
 
 ssize_t safe_file_readat(const char *const file, const int lineno,
diff --git a/lib/tst_cgroup.c b/lib/tst_cgroup.c
index 50699bc63..9831bc336 100644
--- a/lib/tst_cgroup.c
+++ b/lib/tst_cgroup.c
@@ -1345,7 +1345,7 @@ int safe_cg_open(const char *const file, const int lineno,
 		if (!alias)
 			continue;
 
-		fds[i++] = safe_openat(file, lineno, (*dir)->dir_fd, alias, flags);
+		fds[i++] = safe_openat(file, lineno, (*dir)->dir_fd, alias, flags, 0);
 	}
 
 	return i;
diff --git a/lib/tst_safe_file_at.c b/lib/tst_safe_file_at.c
index f530dc349..9b8944f01 100644
--- a/lib/tst_safe_file_at.c
+++ b/lib/tst_safe_file_at.c
@@ -33,15 +33,10 @@ const char *tst_decode_fd(const int fd)
 }
 
 int safe_openat(const char *const file, const int lineno,
-		const int dirfd, const char *const path, const int oflags, ...)
+		const int dirfd, const char *const path, const int oflags,
+		const mode_t mode)
 {
-	va_list ap;
 	int fd;
-	mode_t mode;
-
-	va_start(ap, oflags);
-	mode = va_arg(ap, int);
-	va_end(ap);
 
 	fd = openat(dirfd, path, oflags, mode);
 	if (fd > -1)
@@ -58,7 +53,7 @@ ssize_t safe_file_readat(const char *const file, const int lineno,
 			 const int dirfd, const char *const path,
 			 char *const buf, const size_t nbyte)
 {
-	int fd = safe_openat(file, lineno, dirfd, path, O_RDONLY);
+	int fd = safe_openat(file, lineno, dirfd, path, O_RDONLY, 0);
 	ssize_t rval;
 
 	if (fd < 0)
-- 
2.25.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

[LTP] [PATCH 3/3] safe_semctl: Fix undefined behaviour in vararg handling

[Tudor Cretu]

Accessing elements in an empty va_list is undefined behaviour.
Therefore, remove the variadicness from safe_semctl as it always calls
semctl with the union semun argument included.

Adapt the SAFE_SEMCTL macro to handle the change by passing a
zero-initialised union semun if it's omitted.

Signed-off-by: Tudor Cretu <tudor.cretu@arm.com>
---
 include/tst_safe_sysv_ipc.h | 14 +++++++++-----
 lib/tst_safe_sysv_ipc.c     | 10 +---------
 2 files changed, 10 insertions(+), 14 deletions(-)

diff --git a/include/tst_safe_sysv_ipc.h b/include/tst_safe_sysv_ipc.h
index 7804ce192..976a30409 100644
--- a/include/tst_safe_sysv_ipc.h
+++ b/include/tst_safe_sysv_ipc.h
@@ -10,6 +10,7 @@
 #include <sys/msg.h>
 #include <sys/shm.h>
 #include <sys/sem.h>
+#include "lapi/sem.h"
 
 int safe_msgget(const char *file, const int lineno, key_t key, int msgflg);
 #define SAFE_MSGGET(key, msgflg) \
@@ -58,11 +59,14 @@ int safe_semget(const char *file, const int lineno, key_t key, int nsems,
 	safe_semget(__FILE__, __LINE__, (key), (nsems), (semflg))
 
 int safe_semctl(const char *file, const int lineno, int semid, int semnum,
-		int cmd, ...);
-#define SAFE_SEMCTL(semid, semnum, cmd, ...) ({ \
-	int tst_ret_ = safe_semctl(__FILE__, __LINE__, (semid), (semnum), \
-				(cmd), ##__VA_ARGS__); \
-	(semid) = ((cmd) == IPC_RMID ? -1 : (semid)); \
+		int cmd, union semun un);
+#define __SAFE_SEMCTL(semid, semnum, cmd, un, ...) \
+	safe_semctl(__FILE__, __LINE__, (semid), (semnum), (cmd), (un))
+
+#define SAFE_SEMCTL(semid, semnum, cmd, ...) ({					\
+	int tst_ret_ = __SAFE_SEMCTL((semid), (semnum), (cmd), ##__VA_ARGS__,	\
+				     (union semun){0});				\
+	(semid) = ((cmd) == IPC_RMID ? -1 : (semid));				\
 	tst_ret_; })
 
 int safe_semop(const char *file, const int lineno, int semid, struct sembuf *sops,
diff --git a/lib/tst_safe_sysv_ipc.c b/lib/tst_safe_sysv_ipc.c
index 5eaa82539..f99f6db5e 100644
--- a/lib/tst_safe_sysv_ipc.c
+++ b/lib/tst_safe_sysv_ipc.c
@@ -228,17 +228,9 @@ int safe_semget(const char *file, const int lineno, key_t key, int nsems,
 }
 
 int safe_semctl(const char *file, const int lineno, int semid, int semnum,
-		int cmd, ...)
+		int cmd, union semun un)
 {
 	int rval;
-	va_list va;
-	union semun un;
-
-	va_start(va, cmd);
-
-	un = va_arg(va, union semun);
-
-	va_end(va);
 
 	rval = semctl(semid, semnum, cmd, un);
 
-- 
2.25.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Petr Vorel]

Hi Tudor,

> Accessing elements in an empty va_list results in undefined behaviour[0]
> that can include accessing arbitrary stack memory. While typically this
> doesn't raise a fault, some new more security-oriented architectures
> (e.g. CHERI[1] or Morello[2]) don't allow it.

> Therefore, remove the variadicness from safe_* wrappers that always call
> the functions with the optional argument included.

> Adapt the respective SAFE_* macros to handle the change by passing a
> default argument if they're omitted.

Thanks for an interesting patchset!

I wonder if it's a correct approach as C API allows
int open(const char *pathname, int flags);
we will replace it 
int open(const char *pathname, int flags, mode_t mode);
where mode is 0.
But as it's only in safe_* it should be ok.
We still have some open() tests without mode, i.e.
testcases/kernel/syscalls/open/open06.c

Unfortunately some tests need to be adjusted, at least all tests in
testcases/kernel/syscalls/fgetxattr will fail due int-conversion,
as they use NULL as the third parameter:

$ export CFLAGS="-Werror=conversion"
$ ./configure
$ cd testcases/kernel/syscalls/fgetxattr
$ make clean
$ make V=1
gcc -I../../../../include -I../../../../include -I../../../../include/old/ -Werror=int-conversion -g -O2 -fno-strict-aliasing -pipe -Wall -W -Wold-style-definition -std=gnu99 -L../../../../lib fgetxattr01.c  -lltp -o fgetxattr01
In file included from ../../../../include/tst_test.h:98,
                 from fgetxattr01.c:34:
fgetxattr01.c: In function ‘setup’:
../../../../include/tst_safe_macros.h:90:67: error: passing argument 6 of ‘safe_open’ makes integer from pointer without a cast [-Werror=int-conversion]
   90 |         safe_open(__FILE__, __LINE__, NULL, (pathname), (oflags), (mode))
      |                                                                   ^~~~~~
      |                                                                   |
      |                                                                   void *
../../../../include/tst_safe_macros.h:93:9: note: in expansion of macro ‘__SAFE_OPEN’
   93 |         __SAFE_OPEN((pathname), (oflags), ##__VA_ARGS__, 0)
      |         ^~~~~~~~~~~
fgetxattr01.c:114:14: note: in expansion of macro ‘SAFE_OPEN’
  114 |         fd = SAFE_OPEN(FNAME, O_RDONLY, NULL);
      |              ^~~~~~~~~
In file included from ../../../../include/tst_safe_macros.h:24:
../../../../include/safe_macros_fn.h:78:22: note: expected ‘mode_t’ {aka ‘unsigned int’} but argument is of type ‘void *’
   78 |               mode_t mode);
      |               ~~~~~~~^~~~
cc1: some warnings being treated as errors
make: *** [../../../../include/mk/rules.mk:43: fgetxattr01] Error 1

Found on recent openSUSE Tumbleweed which has 2.36. Also our CI has caught it on
Fedora (which also uses 2.36):
https://github.com/pevik/ltp/actions/runs/3573215390/jobs/6007016572

Kind regards,
Petr

> [0]: [ISO/IEC 9899:2011] Programming Languages—C, 3rd ed, paragraph 7.16.1.1
> [1]: https://www.cl.cam.ac.uk/research/security/ctsrd/cheri/
> [2]: https://www.morello-project.org/

> Tudor Cretu (3):
>   safe_open: Fix undefined behaviour in vararg handling
>   safe_openat: Fix undefined behaviour in vararg handling
>   safe_semctl: Fix undefined behaviour in vararg handling

>  include/old/safe_macros.h   |  6 ++++--
>  include/safe_macros_fn.h    |  3 ++-
>  include/tst_safe_file_at.h  | 10 ++++++----
>  include/tst_safe_macros.h   |  6 ++++--
>  include/tst_safe_sysv_ipc.h | 14 +++++++++-----
>  lib/safe_macros.c           | 13 +------------
>  lib/tst_cgroup.c            |  2 +-
>  lib/tst_safe_file_at.c      | 11 +++--------
>  lib/tst_safe_sysv_ipc.c     | 10 +---------
>  9 files changed, 31 insertions(+), 44 deletions(-)

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Petr Vorel]

> Hi Tudor,

> > Accessing elements in an empty va_list results in undefined behaviour[0]
> > that can include accessing arbitrary stack memory. While typically this
> > doesn't raise a fault, some new more security-oriented architectures
> > (e.g. CHERI[1] or Morello[2]) don't allow it.

> > Therefore, remove the variadicness from safe_* wrappers that always call
> > the functions with the optional argument included.

> > Adapt the respective SAFE_* macros to handle the change by passing a
> > default argument if they're omitted.

> Thanks for an interesting patchset!

> I wonder if it's a correct approach as C API allows
> int open(const char *pathname, int flags);
> we will replace it 
> int open(const char *pathname, int flags, mode_t mode);
> where mode is 0.
> But as it's only in safe_* it should be ok.
> We still have some open() tests without mode, i.e.
> testcases/kernel/syscalls/open/open06.c

> Unfortunately some tests need to be adjusted, at least all tests in
> testcases/kernel/syscalls/fgetxattr will fail due int-conversion,
> as they use NULL as the third parameter:

> $ export CFLAGS="-Werror=conversion"
> $ ./configure
> $ cd testcases/kernel/syscalls/fgetxattr
> $ make clean
> $ make V=1
> gcc -I../../../../include -I../../../../include -I../../../../include/old/ -Werror=int-conversion -g -O2 -fno-strict-aliasing -pipe -Wall -W -Wold-style-definition -std=gnu99 -L../../../../lib fgetxattr01.c  -lltp -o fgetxattr01
> In file included from ../../../../include/tst_test.h:98,
>                  from fgetxattr01.c:34:
> fgetxattr01.c: In function ‘setup’:
> ../../../../include/tst_safe_macros.h:90:67: error: passing argument 6 of ‘safe_open’ makes integer from pointer without a cast [-Werror=int-conversion]
>    90 |         safe_open(__FILE__, __LINE__, NULL, (pathname), (oflags), (mode))
>       |                                                                   ^~~~~~

>       |                                                                   void *
> ../../../../include/tst_safe_macros.h:93:9: note: in expansion of macro ‘__SAFE_OPEN’
>    93 |         __SAFE_OPEN((pathname), (oflags), ##__VA_ARGS__, 0)
>       |         ^~~~~~~~~~~
> fgetxattr01.c:114:14: note: in expansion of macro ‘SAFE_OPEN’
>   114 |         fd = SAFE_OPEN(FNAME, O_RDONLY, NULL);
>       |              ^~~~~~~~~
> In file included from ../../../../include/tst_safe_macros.h:24:
> ../../../../include/safe_macros_fn.h:78:22: note: expected ‘mode_t’ {aka ‘unsigned int’} but argument is of type ‘void *’
>    78 |               mode_t mode);
>       |               ~~~~~~~^~~~
> cc1: some warnings being treated as errors
> make: *** [../../../../include/mk/rules.mk:43: fgetxattr01] Error 1

NOTE: this is from gcc, but obviously also clang has the same problem,
even without -Werror=int-conversion in CFLAGS, obviously it's the default.
That's why it was found on Fedora, which we test with clang.

Kind regards,
Petr

> Found on recent openSUSE Tumbleweed which has 2.36. Also our CI has caught it on
> Fedora (which also uses 2.36):
> https://github.com/pevik/ltp/actions/runs/3573215390/jobs/6007016572

> Kind regards,
> Petr

> > [0]: [ISO/IEC 9899:2011] Programming Languages—C, 3rd ed, paragraph 7.16.1.1
> > [1]: https://www.cl.cam.ac.uk/research/security/ctsrd/cheri/
> > [2]: https://www.morello-project.org/

> > Tudor Cretu (3):
> >   safe_open: Fix undefined behaviour in vararg handling
> >   safe_openat: Fix undefined behaviour in vararg handling
> >   safe_semctl: Fix undefined behaviour in vararg handling

> >  include/old/safe_macros.h   |  6 ++++--
> >  include/safe_macros_fn.h    |  3 ++-
> >  include/tst_safe_file_at.h  | 10 ++++++----
> >  include/tst_safe_macros.h   |  6 ++++--
> >  include/tst_safe_sysv_ipc.h | 14 +++++++++-----
> >  lib/safe_macros.c           | 13 +------------
> >  lib/tst_cgroup.c            |  2 +-
> >  lib/tst_safe_file_at.c      | 11 +++--------
> >  lib/tst_safe_sysv_ipc.c     | 10 +---------
> >  9 files changed, 31 insertions(+), 44 deletions(-)

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Richard Palethorpe]

Hello,

Petr Vorel <pvorel@suse.cz> writes:

>> Hi Tudor,
>
>> > Accessing elements in an empty va_list results in undefined behaviour[0]
>> > that can include accessing arbitrary stack memory. While typically this
>> > doesn't raise a fault, some new more security-oriented architectures
>> > (e.g. CHERI[1] or Morello[2]) don't allow it.
>
>> > Therefore, remove the variadicness from safe_* wrappers that always call
>> > the functions with the optional argument included.
>
>> > Adapt the respective SAFE_* macros to handle the change by passing a
>> > default argument if they're omitted.
>
>> Thanks for an interesting patchset!
>
>> I wonder if it's a correct approach as C API allows

Perhaps, muslc does the following for open()

	if ((flags & O_CREAT) || (flags & O_TMPFILE) == O_TMPFILE) {
		va_list ap;
		va_start(ap, flags);
		mode = va_arg(ap, mode_t);
		va_end(ap);
	}

So it's only accessed if we need the mode. If the error below can be
fixed with the current approach I'd also be happy.


>> int open(const char *pathname, int flags);
>> we will replace it 
>> int open(const char *pathname, int flags, mode_t mode);
>> where mode is 0.
>> But as it's only in safe_* it should be ok.
>> We still have some open() tests without mode, i.e.
>> testcases/kernel/syscalls/open/open06.c
>
>> Unfortunately some tests need to be adjusted, at least all tests in
>> testcases/kernel/syscalls/fgetxattr will fail due int-conversion,
>> as they use NULL as the third parameter:
>
>> $ export CFLAGS="-Werror=conversion"
>> $ ./configure
>> $ cd testcases/kernel/syscalls/fgetxattr
>> $ make clean
>> $ make V=1
>> gcc -I../../../../include -I../../../../include -I../../../../include/old/ -Werror=int-conversion -g -O2 -fno-strict-aliasing -pipe -Wall -W -Wold-style-definition -std=gnu99 -L../../../../lib fgetxattr01.c  -lltp -o fgetxattr01
>> In file included from ../../../../include/tst_test.h:98,
>>                  from fgetxattr01.c:34:
>> fgetxattr01.c: In function ‘setup’:
>> ../../../../include/tst_safe_macros.h:90:67: error: passing argument 6 of ‘safe_open’ makes integer from pointer without a cast [-Werror=int-conversion]
>>    90 |         safe_open(__FILE__, __LINE__, NULL, (pathname), (oflags), (mode))
>>       |                                                                   ^~~~~~
>
>>       |                                                                   void *
>> ../../../../include/tst_safe_macros.h:93:9: note: in expansion of macro ‘__SAFE_OPEN’
>>    93 |         __SAFE_OPEN((pathname), (oflags), ##__VA_ARGS__, 0)
>>       |         ^~~~~~~~~~~
>> fgetxattr01.c:114:14: note: in expansion of macro ‘SAFE_OPEN’
>>   114 |         fd = SAFE_OPEN(FNAME, O_RDONLY, NULL);
>>       |              ^~~~~~~~~
>> In file included from ../../../../include/tst_safe_macros.h:24:
>> ../../../../include/safe_macros_fn.h:78:22: note: expected ‘mode_t’ {aka ‘unsigned int’} but argument is of type ‘void *’
>>    78 |               mode_t mode);
>>       |               ~~~~~~~^~~~
>> cc1: some warnings being treated as errors
>> make: *** [../../../../include/mk/rules.mk:43: fgetxattr01] Error 1
>
> NOTE: this is from gcc, but obviously also clang has the same problem,
> even without -Werror=int-conversion in CFLAGS, obviously it's the default.
> That's why it was found on Fedora, which we test with clang.
>
> Kind regards,
> Petr
>
>> Found on recent openSUSE Tumbleweed which has 2.36. Also our CI has caught it on
>> Fedora (which also uses 2.36):
>> https://github.com/pevik/ltp/actions/runs/3573215390/jobs/6007016572
>
>> Kind regards,
>> Petr
>
>> > [0]: [ISO/IEC 9899:2011] Programming Languages—C, 3rd ed, paragraph 7.16.1.1
>> > [1]: https://www.cl.cam.ac.uk/research/security/ctsrd/cheri/
>> > [2]: https://www.morello-project.org/
>
>> > Tudor Cretu (3):
>> >   safe_open: Fix undefined behaviour in vararg handling
>> >   safe_openat: Fix undefined behaviour in vararg handling
>> >   safe_semctl: Fix undefined behaviour in vararg handling
>
>> >  include/old/safe_macros.h   |  6 ++++--
>> >  include/safe_macros_fn.h    |  3 ++-
>> >  include/tst_safe_file_at.h  | 10 ++++++----
>> >  include/tst_safe_macros.h   |  6 ++++--
>> >  include/tst_safe_sysv_ipc.h | 14 +++++++++-----
>> >  lib/safe_macros.c           | 13 +------------
>> >  lib/tst_cgroup.c            |  2 +-
>> >  lib/tst_safe_file_at.c      | 11 +++--------
>> >  lib/tst_safe_sysv_ipc.c     | 10 +---------
>> >  9 files changed, 31 insertions(+), 44 deletions(-)


-- 
Thank you,
Richard.

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Richard Palethorpe]

Hello,

This looks pretty trivial to fix actually. We shouldn't pass NULL as
mode. If it works I'll add the fix and push.

Richard Palethorpe <rpalethorpe@suse.de> writes:

> Hello,
>
> Petr Vorel <pvorel@suse.cz> writes:
>
>>> Hi Tudor,
>>
>>> > Accessing elements in an empty va_list results in undefined behaviour[0]
>>> > that can include accessing arbitrary stack memory. While typically this
>>> > doesn't raise a fault, some new more security-oriented architectures
>>> > (e.g. CHERI[1] or Morello[2]) don't allow it.
>>
>>> > Therefore, remove the variadicness from safe_* wrappers that always call
>>> > the functions with the optional argument included.
>>
>>> > Adapt the respective SAFE_* macros to handle the change by passing a
>>> > default argument if they're omitted.
>>
>>> Thanks for an interesting patchset!
>>
>>> I wonder if it's a correct approach as C API allows
>
> Perhaps, muslc does the following for open()
>
> 	if ((flags & O_CREAT) || (flags & O_TMPFILE) == O_TMPFILE) {
> 		va_list ap;
> 		va_start(ap, flags);
> 		mode = va_arg(ap, mode_t);
> 		va_end(ap);
> 	}
>
> So it's only accessed if we need the mode. If the error below can be
> fixed with the current approach I'd also be happy.
>
>
>>> int open(const char *pathname, int flags);
>>> we will replace it 
>>> int open(const char *pathname, int flags, mode_t mode);
>>> where mode is 0.
>>> But as it's only in safe_* it should be ok.
>>> We still have some open() tests without mode, i.e.
>>> testcases/kernel/syscalls/open/open06.c
>>
>>> Unfortunately some tests need to be adjusted, at least all tests in
>>> testcases/kernel/syscalls/fgetxattr will fail due int-conversion,
>>> as they use NULL as the third parameter:
>>
>>> $ export CFLAGS="-Werror=conversion"
>>> $ ./configure
>>> $ cd testcases/kernel/syscalls/fgetxattr
>>> $ make clean
>>> $ make V=1
>>> gcc -I../../../../include -I../../../../include -I../../../../include/old/ -Werror=int-conversion -g -O2 -fno-strict-aliasing -pipe -Wall -W -Wold-style-definition -std=gnu99 -L../../../../lib fgetxattr01.c  -lltp -o fgetxattr01
>>> In file included from ../../../../include/tst_test.h:98,
>>>                  from fgetxattr01.c:34:
>>> fgetxattr01.c: In function ‘setup’:
>>> ../../../../include/tst_safe_macros.h:90:67: error: passing argument 6 of ‘safe_open’ makes integer from pointer without a cast [-Werror=int-conversion]
>>>    90 |         safe_open(__FILE__, __LINE__, NULL, (pathname), (oflags), (mode))
>>>       |                                                                   ^~~~~~
>>
>>>       |                                                                   void *
>>> ../../../../include/tst_safe_macros.h:93:9: note: in expansion of macro ‘__SAFE_OPEN’
>>>    93 |         __SAFE_OPEN((pathname), (oflags), ##__VA_ARGS__, 0)
>>>       |         ^~~~~~~~~~~
>>> fgetxattr01.c:114:14: note: in expansion of macro ‘SAFE_OPEN’
>>>   114 |         fd = SAFE_OPEN(FNAME, O_RDONLY, NULL);
>>>       |              ^~~~~~~~~
>>> In file included from ../../../../include/tst_safe_macros.h:24:
>>> ../../../../include/safe_macros_fn.h:78:22: note: expected ‘mode_t’ {aka ‘unsigned int’} but argument is of type ‘void *’
>>>    78 |               mode_t mode);
>>>       |               ~~~~~~~^~~~
>>> cc1: some warnings being treated as errors
>>> make: *** [../../../../include/mk/rules.mk:43: fgetxattr01] Error 1
>>
>> NOTE: this is from gcc, but obviously also clang has the same problem,
>> even without -Werror=int-conversion in CFLAGS, obviously it's the default.
>> That's why it was found on Fedora, which we test with clang.
>>
>> Kind regards,
>> Petr
>>
>>> Found on recent openSUSE Tumbleweed which has 2.36. Also our CI has caught it on
>>> Fedora (which also uses 2.36):
>>> https://github.com/pevik/ltp/actions/runs/3573215390/jobs/6007016572
>>
>>> Kind regards,
>>> Petr
>>
>>> > [0]: [ISO/IEC 9899:2011] Programming Languages—C, 3rd ed, paragraph 7.16.1.1
>>> > [1]: https://www.cl.cam.ac.uk/research/security/ctsrd/cheri/
>>> > [2]: https://www.morello-project.org/
>>
>>> > Tudor Cretu (3):
>>> >   safe_open: Fix undefined behaviour in vararg handling
>>> >   safe_openat: Fix undefined behaviour in vararg handling
>>> >   safe_semctl: Fix undefined behaviour in vararg handling
>>
>>> >  include/old/safe_macros.h   |  6 ++++--
>>> >  include/safe_macros_fn.h    |  3 ++-
>>> >  include/tst_safe_file_at.h  | 10 ++++++----
>>> >  include/tst_safe_macros.h   |  6 ++++--
>>> >  include/tst_safe_sysv_ipc.h | 14 +++++++++-----
>>> >  lib/safe_macros.c           | 13 +------------
>>> >  lib/tst_cgroup.c            |  2 +-
>>> >  lib/tst_safe_file_at.c      | 11 +++--------
>>> >  lib/tst_safe_sysv_ipc.c     | 10 +---------
>>> >  9 files changed, 31 insertions(+), 44 deletions(-)


-- 
Thank you,
Richard.

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Petr Vorel]

> Hello,

> This looks pretty trivial to fix actually. We shouldn't pass NULL as
> mode. If it works I'll add the fix and push.
Yes, it fixes it, good point. I was also surprised by NULL.

to whole patchset:
Reviewed-by: Petr Vorel <pvorel@suse.cz>


Kind regards,
Petr

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Tudor Cretu]

Hi Petr,

On 29-11-2022 11:01, Petr Vorel wrote:
> Hi Tudor,
> 
>> Accessing elements in an empty va_list results in undefined behaviour[0]
>> that can include accessing arbitrary stack memory. While typically this
>> doesn't raise a fault, some new more security-oriented architectures
>> (e.g. CHERI[1] or Morello[2]) don't allow it.
> 
>> Therefore, remove the variadicness from safe_* wrappers that always call
>> the functions with the optional argument included.
> 
>> Adapt the respective SAFE_* macros to handle the change by passing a
>> default argument if they're omitted.
> 
> Thanks for an interesting patchset!

Thank you for having a look!

> 
> I wonder if it's a correct approach as C API allows
> int open(const char *pathname, int flags);
> we will replace it
> int open(const char *pathname, int flags, mode_t mode);
> where mode is 0.

The man page for open specifies that in the cases where mode is ignored, 
it can either be omitted or specified as 0. safe_open was reading random 
bytes from the stack, while now it ensures it's 0 in those cases. I 
agree it's not a crystal clear approach for the same reasons you 
mention, but I find this approach to efficiently remove the undefined 
behaviour, while keeping the code clean.

> But as it's only in safe_* it should be ok.
> We still have some open() tests without mode, i.e.
> testcases/kernel/syscalls/open/open06.c

I don't think they need to be modified as a result of this patch series. 
It's up to the libcs to avoid the undefined behaviour in that case.

> 
> Unfortunately some tests need to be adjusted, at least all tests in
> testcases/kernel/syscalls/fgetxattr will fail due int-conversion,
> as they use NULL as the third parameter:
> 
> $ export CFLAGS="-Werror=conversion"
> $ ./configure
> $ cd testcases/kernel/syscalls/fgetxattr
> $ make clean
> $ make V=1
> gcc -I../../../../include -I../../../../include -I../../../../include/old/ -Werror=int-conversion -g -O2 -fno-strict-aliasing -pipe -Wall -W -Wold-style-definition -std=gnu99 -L../../../../lib fgetxattr01.c  -lltp -o fgetxattr01
> In file included from ../../../../include/tst_test.h:98,
>                   from fgetxattr01.c:34:
> fgetxattr01.c: In function ‘setup’:
> ../../../../include/tst_safe_macros.h:90:67: error: passing argument 6 of ‘safe_open’ makes integer from pointer without a cast [-Werror=int-conversion]
>     90 |         safe_open(__FILE__, __LINE__, NULL, (pathname), (oflags), (mode))
>        |                                                                   ^~~~~~
>        |                                                                   |
>        |                                                                   void *
> ../../../../include/tst_safe_macros.h:93:9: note: in expansion of macro ‘__SAFE_OPEN’
>     93 |         __SAFE_OPEN((pathname), (oflags), ##__VA_ARGS__, 0)
>        |         ^~~~~~~~~~~
> fgetxattr01.c:114:14: note: in expansion of macro ‘SAFE_OPEN’
>    114 |         fd = SAFE_OPEN(FNAME, O_RDONLY, NULL);
>        |              ^~~~~~~~~
> In file included from ../../../../include/tst_safe_macros.h:24:
> ../../../../include/safe_macros_fn.h:78:22: note: expected ‘mode_t’ {aka ‘unsigned int’} but argument is of type ‘void *’
>     78 |               mode_t mode);
>        |               ~~~~~~~^~~~
> cc1: some warnings being treated as errors
> make: *** [../../../../include/mk/rules.mk:43: fgetxattr01] Error 1
> 
> Found on recent openSUSE Tumbleweed which has 2.36. Also our CI has caught it on
> Fedora (which also uses 2.36):
> https://github.com/pevik/ltp/actions/runs/3573215390/jobs/6007016572

Many apologies for letting this slip through and many thanks for 
catching it! I'll remove the NULL and have a recheck on all the macros 
modified.

Kind regards,
Tudor

> 
> Kind regards,
> Petr
> 
>> [0]: [ISO/IEC 9899:2011] Programming Languages—C, 3rd ed, paragraph 7.16.1.1
>> [1]: https://www.cl.cam.ac.uk/research/security/ctsrd/cheri/
>> [2]: https://www.morello-project.org/
> 
>> Tudor Cretu (3):
>>    safe_open: Fix undefined behaviour in vararg handling
>>    safe_openat: Fix undefined behaviour in vararg handling
>>    safe_semctl: Fix undefined behaviour in vararg handling
> 
>>   include/old/safe_macros.h   |  6 ++++--
>>   include/safe_macros_fn.h    |  3 ++-
>>   include/tst_safe_file_at.h  | 10 ++++++----
>>   include/tst_safe_macros.h   |  6 ++++--
>>   include/tst_safe_sysv_ipc.h | 14 +++++++++-----
>>   lib/safe_macros.c           | 13 +------------
>>   lib/tst_cgroup.c            |  2 +-
>>   lib/tst_safe_file_at.c      | 11 +++--------
>>   lib/tst_safe_sysv_ipc.c     | 10 +---------
>>   9 files changed, 31 insertions(+), 44 deletions(-)

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Tudor Cretu]

On 29-11-2022 11:32, Petr Vorel wrote:
>> Hello,
> 
>> This looks pretty trivial to fix actually. We shouldn't pass NULL as
>> mode. If it works I'll add the fix and push.
> Yes, it fixes it, good point. I was also surprised by NULL.

Many thanks, both! I'll remove the NULL and re-post a v2 if that's alright.

Kind regards,
Tudor

> 
> to whole patchset:
> Reviewed-by: Petr Vorel <pvorel@suse.cz>
> 
> 
> Kind regards,
> Petr

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Cyril Hrubis]

Hi!
> Accessing elements in an empty va_list results in undefined behaviour[0]
> that can include accessing arbitrary stack memory. While typically this
> doesn't raise a fault, some new more security-oriented architectures
> (e.g. CHERI[1] or Morello[2]) don't allow it.

Looking at how glibc handles this, the code looks like:

	int mode = 0;

	if (__OPEN_NEEDS_MODE(oflag)) {
		..
		mode = va_arg(arg, int);
		..
	}

That sounds much easier than messing with the macros and should avoid
undefined behavior.

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 0/3] safe_macros: Fix undefined behaviour in vararg handling

[Richard Palethorpe]

Hi,

Tudor Cretu <tudor.cretu@arm.com> writes:

> On 29-11-2022 11:32, Petr Vorel wrote:
>>> Hello,
>> 
>>> This looks pretty trivial to fix actually. We shouldn't pass NULL as
>>> mode. If it works I'll add the fix and push.
>> Yes, it fixes it, good point. I was also surprised by NULL.
>
> Many thanks, both! I'll remove the NULL and re-post a v2 if that's
> alright.

OK, please do.

>
> Kind regards,
> Tudor
>
>> to whole patchset:
>> Reviewed-by: Petr Vorel <pvorel@suse.cz>
>> Kind regards,
>> Petr


-- 
Thank you,
Richard.

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp