All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] selftests/exec: Avoid future NULL argv execve warning
@ 2022-02-01  0:08 Kees Cook
  2022-02-02 15:13 ` Alexey Dobriyan
  0 siblings, 1 reply; 5+ messages in thread
From: Kees Cook @ 2022-02-01  0:08 UTC (permalink / raw)
  To: Eric Biederman
  Cc: Kees Cook, Alexey Dobriyan, Shuah Khan, linux-kselftest,
	Ariadne Conill, Michael Kerrisk, Matthew Wilcox,
	Christian Brauner, Rich Felker, Alexander Viro, linux-fsdevel,
	linux-kernel, linux-hardening

Build actual argv for launching recursion test to avoid future warning
about using an empty argv in execve().

Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: linux-kselftest@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
---
 tools/testing/selftests/exec/recursion-depth.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/tools/testing/selftests/exec/recursion-depth.c b/tools/testing/selftests/exec/recursion-depth.c
index 2dbd5bc45b3e..35348db00c52 100644
--- a/tools/testing/selftests/exec/recursion-depth.c
+++ b/tools/testing/selftests/exec/recursion-depth.c
@@ -24,8 +24,14 @@
 #include <sys/mount.h>
 #include <unistd.h>
 
+#define FILENAME "/tmp/1"
+#define HASHBANG "#!" FILENAME "\n"
+
 int main(void)
 {
+	char * const argv[] = { FILENAME, NULL };
+	int rv;
+
 	if (unshare(CLONE_NEWNS) == -1) {
 		if (errno == ENOSYS || errno == EPERM) {
 			fprintf(stderr, "error: unshare, errno %d\n", errno);
@@ -44,21 +50,19 @@ int main(void)
 		return 1;
 	}
 
-#define FILENAME "/tmp/1"
 
 	int fd = creat(FILENAME, 0700);
 	if (fd == -1) {
 		fprintf(stderr, "error: creat, errno %d\n", errno);
 		return 1;
 	}
-#define S "#!" FILENAME "\n"
-	if (write(fd, S, strlen(S)) != strlen(S)) {
+	if (write(fd, HASHBANG, strlen(HASHBANG)) != strlen(HASHBANG)) {
 		fprintf(stderr, "error: write, errno %d\n", errno);
 		return 1;
 	}
 	close(fd);
 
-	int rv = execve(FILENAME, NULL, NULL);
+	rv = execve(FILENAME, argv, NULL);
 	if (rv == -1 && errno == ELOOP) {
 		return 0;
 	}
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH] selftests/exec: Avoid future NULL argv execve warning
  2022-02-01  0:08 [PATCH] selftests/exec: Avoid future NULL argv execve warning Kees Cook
@ 2022-02-02 15:13 ` Alexey Dobriyan
  2022-02-02 17:38   ` Shuah Khan
  0 siblings, 1 reply; 5+ messages in thread
From: Alexey Dobriyan @ 2022-02-02 15:13 UTC (permalink / raw)
  To: Kees Cook
  Cc: Eric Biederman, Shuah Khan, linux-kselftest, Ariadne Conill,
	Michael Kerrisk, Matthew Wilcox, Christian Brauner, Rich Felker,
	Alexander Viro, linux-fsdevel, linux-kernel, linux-hardening

On Mon, Jan 31, 2022 at 04:08:07PM -0800, Kees Cook wrote:
> Build actual argv for launching recursion test to avoid future warning
> about using an empty argv in execve().

> --- a/tools/testing/selftests/exec/recursion-depth.c
> +++ b/tools/testing/selftests/exec/recursion-depth.c
> @@ -24,8 +24,14 @@
>  #include <sys/mount.h>
>  #include <unistd.h>
>  
> +#define FILENAME "/tmp/1"
> +#define HASHBANG "#!" FILENAME "\n"
> +
>  int main(void)
>  {
> +	char * const argv[] = { FILENAME, NULL };
> +	int rv;

Can we move out of -Wdeclaration-after-statement mentality in tests at least?

> -	int rv = execve(FILENAME, NULL, NULL);
> +	rv = execve(FILENAME, argv, NULL);

	int rv = execve(FILENAME, (char*[]){FILENAME, NULL}, NULL);

is cleaner (and modern)!

>  	if (rv == -1 && errno == ELOOP) {
>  		return 0;
>  	}

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] selftests/exec: Avoid future NULL argv execve warning
  2022-02-02 15:13 ` Alexey Dobriyan
@ 2022-02-02 17:38   ` Shuah Khan
  2022-02-02 21:00     ` Alexey Dobriyan
  0 siblings, 1 reply; 5+ messages in thread
From: Shuah Khan @ 2022-02-02 17:38 UTC (permalink / raw)
  To: Alexey Dobriyan, Kees Cook
  Cc: Eric Biederman, Shuah Khan, linux-kselftest, Ariadne Conill,
	Michael Kerrisk, Matthew Wilcox, Christian Brauner, Rich Felker,
	Alexander Viro, linux-fsdevel, linux-kernel, linux-hardening,
	Shuah Khan

On 2/2/22 8:13 AM, Alexey Dobriyan wrote:
> On Mon, Jan 31, 2022 at 04:08:07PM -0800, Kees Cook wrote:
>> Build actual argv for launching recursion test to avoid future warning
>> about using an empty argv in execve().
> 
>> --- a/tools/testing/selftests/exec/recursion-depth.c
>> +++ b/tools/testing/selftests/exec/recursion-depth.c
>> @@ -24,8 +24,14 @@
>>   #include <sys/mount.h>
>>   #include <unistd.h>
>>   
>> +#define FILENAME "/tmp/1"
>> +#define HASHBANG "#!" FILENAME "\n"
>> +
>>   int main(void)
>>   {
>> +	char * const argv[] = { FILENAME, NULL };
>> +	int rv;
> 
> Can we move out of -Wdeclaration-after-statement mentality in tests at least?

selftest like the rest of the kernel follows the same coding guidelines.
It will follow the moving "-Wdeclaration-after-statement mentality" when
the rest of the kernel does.

Looks like this topic was discussed in the following:
https://patchwork.kernel.org/project/linux-kbuild/patch/c6fda26e8d134264b04fadc3386d6c32@gmail.com/

thanks,
-- Shuah

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] selftests/exec: Avoid future NULL argv execve warning
  2022-02-02 17:38   ` Shuah Khan
@ 2022-02-02 21:00     ` Alexey Dobriyan
  2022-02-04 16:11       ` Shuah Khan
  0 siblings, 1 reply; 5+ messages in thread
From: Alexey Dobriyan @ 2022-02-02 21:00 UTC (permalink / raw)
  To: Shuah Khan
  Cc: Kees Cook, Eric Biederman, Shuah Khan, linux-kselftest,
	Ariadne Conill, Michael Kerrisk, Matthew Wilcox,
	Christian Brauner, Rich Felker, Alexander Viro, linux-fsdevel,
	linux-kernel, linux-hardening

On Wed, Feb 02, 2022 at 10:38:57AM -0700, Shuah Khan wrote:
> On 2/2/22 8:13 AM, Alexey Dobriyan wrote:
> > On Mon, Jan 31, 2022 at 04:08:07PM -0800, Kees Cook wrote:
> > > Build actual argv for launching recursion test to avoid future warning
> > > about using an empty argv in execve().
> > 
> > > --- a/tools/testing/selftests/exec/recursion-depth.c
> > > +++ b/tools/testing/selftests/exec/recursion-depth.c
> > > @@ -24,8 +24,14 @@
> > >   #include <sys/mount.h>
> > >   #include <unistd.h>
> > > +#define FILENAME "/tmp/1"
> > > +#define HASHBANG "#!" FILENAME "\n"
> > > +
> > >   int main(void)
> > >   {
> > > +	char * const argv[] = { FILENAME, NULL };
> > > +	int rv;
> > 
> > Can we move out of -Wdeclaration-after-statement mentality in tests at least?
> 
> selftest like the rest of the kernel follows the same coding guidelines.
> It will follow the moving "-Wdeclaration-after-statement mentality" when
> the rest of the kernel does.
> 
> Looks like this topic was discussed in the following:
> https://patchwork.kernel.org/project/linux-kbuild/patch/c6fda26e8d134264b04fadc3386d6c32@gmail.com/

The only real argument is "gcc miscompiles /proc" to which adding -Wdeclaration-after-statement
looks like a too big hammer.

Why can't we have nice things?

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] selftests/exec: Avoid future NULL argv execve warning
  2022-02-02 21:00     ` Alexey Dobriyan
@ 2022-02-04 16:11       ` Shuah Khan
  0 siblings, 0 replies; 5+ messages in thread
From: Shuah Khan @ 2022-02-04 16:11 UTC (permalink / raw)
  To: Alexey Dobriyan
  Cc: Kees Cook, Eric Biederman, Shuah Khan, linux-kselftest,
	Ariadne Conill, Michael Kerrisk, Matthew Wilcox,
	Christian Brauner, Rich Felker, Alexander Viro, linux-fsdevel,
	linux-kernel, linux-hardening, Shuah Khan

On 2/2/22 2:00 PM, Alexey Dobriyan wrote:
> On Wed, Feb 02, 2022 at 10:38:57AM -0700, Shuah Khan wrote:
>> On 2/2/22 8:13 AM, Alexey Dobriyan wrote:
>>> On Mon, Jan 31, 2022 at 04:08:07PM -0800, Kees Cook wrote:
>>>> Build actual argv for launching recursion test to avoid future warning
>>>> about using an empty argv in execve().
>>>
>>>> --- a/tools/testing/selftests/exec/recursion-depth.c
>>>> +++ b/tools/testing/selftests/exec/recursion-depth.c
>>>> @@ -24,8 +24,14 @@
>>>>    #include <sys/mount.h>
>>>>    #include <unistd.h>
>>>> +#define FILENAME "/tmp/1"
>>>> +#define HASHBANG "#!" FILENAME "\n"
>>>> +
>>>>    int main(void)
>>>>    {
>>>> +	char * const argv[] = { FILENAME, NULL };
>>>> +	int rv;
>>>
>>> Can we move out of -Wdeclaration-after-statement mentality in tests at least?
>>
>> selftest like the rest of the kernel follows the same coding guidelines.
>> It will follow the moving "-Wdeclaration-after-statement mentality" when
>> the rest of the kernel does.
>>
>> Looks like this topic was discussed in the following:
>> https://patchwork.kernel.org/project/linux-kbuild/patch/c6fda26e8d134264b04fadc3386d6c32@gmail.com/
> 
> The only real argument is "gcc miscompiles /proc" to which adding -Wdeclaration-after-statement
> looks like a too big hammer.
> 

Either way - selftest will stay in sync with the kernel coding standards
for good reasons. Doing its own thing confuses developers and makes it
hard for maintainers.

thanks,
-- Shuah

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2022-02-04 16:11 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-02-01  0:08 [PATCH] selftests/exec: Avoid future NULL argv execve warning Kees Cook
2022-02-02 15:13 ` Alexey Dobriyan
2022-02-02 17:38   ` Shuah Khan
2022-02-02 21:00     ` Alexey Dobriyan
2022-02-04 16:11       ` Shuah Khan

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.