Re: [LTP] [PATCH] clock_settime03: do not set outside address space tp

Re: [LTP] [PATCH] clock_settime03: do not set outside address space tp

[chrubis]

Hi!
> > Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> 
> Any objections?

I will have a look soon enough, sorry it takes so long.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] clock_settime03: do not set outside address space tp

[chrubis]

Hi!
> Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> ---
>  .../kernel/timers/clock_settime/clock_settime03.c    | 20 ++------------------
>  1 file changed, 2 insertions(+), 18 deletions(-)
> 
> diff --git a/testcases/kernel/timers/clock_settime/clock_settime03.c b/testcases/kernel/timers/clock_settime/clock_settime03.c
> index a949303..0b24e6d 100644
> --- a/testcases/kernel/timers/clock_settime/clock_settime03.c
> +++ b/testcases/kernel/timers/clock_settime/clock_settime03.c
> @@ -49,8 +49,8 @@ int testcases[] = {
>  	EINVAL,			/* Invalid timespec     */
>  	EINVAL,			/* NSEC_PER_SEC + 1     */
>  	EPERM,			/* non-root user        */
> -	0,
> -	0,
> +	EINVAL,			/* PROCESS_CPUTIME_ID	*/
> +	EINVAL,			/* THREAD_CPUTIME_ID	*/
>  };
>  
>  char *TCID = "clock_settime03";
> @@ -70,17 +70,6 @@ int main(int ac, char **av)
>  	if ((msg = parse_opts(ac, av, NULL, NULL)) != NULL)
>  		tst_brkm(TBROK, NULL, "OPTION PARSING ERROR - %s", msg);
>  
> -	/* PROCESS_CPUTIME_ID & THREAD_CPUTIME_ID are not supported on
> -	 * kernel versions lower than 2.6.12 and changed back in 2.6.38
> -	 */
> -	if ((tst_kvercmp(2, 6, 12)) < 0 || (tst_kvercmp(2, 6, 38)) >= 0) {
> -		testcases[7] = EINVAL;
> -		testcases[8] = EINVAL;
> -	} else {
> -		testcases[7] = EFAULT;
> -		testcases[8] = EFAULT;
> -	}
> -
>  	setup();
>  
>  	for (lc = 0; TEST_LOOPING(lc); lc++) {
> @@ -168,11 +157,6 @@ static int setup_test(int option)
>  			return -1;
>  		}
>  		break;
> -	case 7:
> -	case 8:
> -		/* Make tp argument bad pointer */
> -		if (tst_kvercmp(2, 6, 12) >= 0)
> -			temp = (struct timespec *)-1;
>  	}
>  	return 0;
>  }

If I'm getting this right the order of checks in kernel was probably
changed (userspace address validation vs timer id validation) and
because of that the errno values were not stable. And as the test passes
invalid pointer in first testcase (with correct timer id) the invalid
pointer codepath in kernel is covered by the test allready.

Acked.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list