[LTP] [PATCH] sched_getaffinity: fix possible fail in negative case

[LTP] [PATCH] sched_getaffinity: fix possible fail in negative case

[Caspar Zhang]

[-- Attachment #1: Type: text/plain, Size: 559 bytes --]


in negative case:

    QUICK_TEST(sched_getaffinity(getpid()+1, len, mask));

sometimes getpid()+1 will exist if there're other processes running.
This patch uses a different way to make sure the required pid doesn't
exist by forking a child and terminating it. The pid of the terminated
child will not be used for a short time and it's enough for the test to
continue.

Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
---
 .../sched_getaffinity/sched_getaffinity01.c        |   18 ++++++++++++++++--
 1 files changed, 16 insertions(+), 2 deletions(-)


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-sched_getaffinity-fix-possible-fail-in-negative-case.patch --]
[-- Type: text/x-patch; name="0001-sched_getaffinity-fix-possible-fail-in-negative-case.patch", Size: 1125 bytes --]

diff --git a/testcases/kernel/syscalls/sched_getaffinity/sched_getaffinity01.c b/testcases/kernel/syscalls/sched_getaffinity/sched_getaffinity01.c
index 97cf90f..80dea9a 100644
--- a/testcases/kernel/syscalls/sched_getaffinity/sched_getaffinity01.c
+++ b/testcases/kernel/syscalls/sched_getaffinity/sched_getaffinity01.c
@@ -94,7 +94,7 @@ int main(int ac, char **av)
 
 static void do_test(void)
 {
-	int i;
+	int i, status;
 	cpu_set_t *mask;
 	int nrcpus = 1024;
 	pid_t pid;
@@ -153,10 +153,24 @@ realloc:
 #else
 	CPU_ZERO(mask);
 #endif
+
 	/* negative tests */
+
 	QUICK_TEST(sched_getaffinity(0, len, (cpu_set_t *)-1));
+
 	QUICK_TEST(sched_getaffinity(0, 0, mask));
-	QUICK_TEST(sched_getaffinity(getpid() + 1, len, mask));
+
+	switch (pid = fork()) {
+	case -1:
+		tst_brkm(TBROK|TERRNO, cleanup, "fork");
+	case 0:
+		exit(0);
+	default:
+		if (waitpid(pid, &status, WUNTRACED|WCONTINUED) == -1)
+			tst_brkm(TBROK|TERRNO, cleanup, "waitpid");
+	}
+	QUICK_TEST(sched_getaffinity(pid, len, mask));
+
 	/*
 	 * pid_t -> int -- the actual kernel limit is lower
 	 * though, but this is a negative test, not a positive

[-- Attachment #3: Type: text/plain, Size: 317 bytes --]

------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/

[-- Attachment #4: Type: text/plain, Size: 155 bytes --]

_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] sched_getaffinity: fix possible fail in negative case

[Caspar Zhang]

On 03/09/2012 01:28 PM, Caspar Zhang wrote:
> 
> in negative case:
> 
>     QUICK_TEST(sched_getaffinity(getpid()+1, len, mask));
> 
> sometimes getpid()+1 will exist if there're other processes running.
> This patch uses a different way to make sure the required pid doesn't
> exist by forking a child and terminating it. The pid of the terminated
> child will not be used for a short time and it's enough for the test to
> continue.
> 
> Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
> ---
>  .../sched_getaffinity/sched_getaffinity01.c        |   18 ++++++++++++++++--
>  1 files changed, 16 insertions(+), 2 deletions(-)
> 

Hmmm, I took a look at commit c8319c8 and found that Garrett added a
similar non-exist-pid case already. I guess we can simply remove the one
I tried to fix...

Thanks,
Caspar

------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] sched_getaffinity: fix possible fail in negative case

[Wanlong Gao]

On 03/09/2012 01:48 PM, Caspar Zhang wrote:

> On 03/09/2012 01:28 PM, Caspar Zhang wrote:
>>
>> in negative case:
>>
>>     QUICK_TEST(sched_getaffinity(getpid()+1, len, mask));
>>
>> sometimes getpid()+1 will exist if there're other processes running.
>> This patch uses a different way to make sure the required pid doesn't
>> exist by forking a child and terminating it. The pid of the terminated
>> child will not be used for a short time and it's enough for the test to
>> continue.
>>
>> Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
>> ---
>>  .../sched_getaffinity/sched_getaffinity01.c        |   18 ++++++++++++++++--
>>  1 files changed, 16 insertions(+), 2 deletions(-)
>>
> 
> Hmmm, I took a look at commit c8319c8 and found that Garrett added a
> similar non-exist-pid case already. I guess we can simply remove the one
> I tried to fix...


Yes, agree, please remove this one.

Thanks,
Wanlong Gao

> 
> Thanks,
> Caspar
> 
> ------------------------------------------------------------------------------
> Virtualization & Cloud Management Using Capacity Planning
> Cloud computing makes use of virtualization - but cloud computing 
> also focuses on allowing computing to be delivered as a service.
> http://www.accelacomm.com/jaw/sfnl/114/51521223/
> _______________________________________________
> Ltp-list mailing list
> Ltp-list@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/ltp-list
> 



------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] sched_getaffinity: fix possible fail in negative case

[Garrett Cooper]

On Mar 8, 2012, at 9:28 PM, Caspar Zhang wrote:

> 
> in negative case:
> 
>    QUICK_TEST(sched_getaffinity(getpid()+1, len, mask));
> 
> sometimes getpid()+1 will exist if there're other processes running.
> This patch uses a different way to make sure the required pid doesn't
> exist by forking a child and terminating it. The pid of the terminated
> child will not be used for a short time and it's enough for the test to
> continue.

	pid_t is signed on FreeBSD and Linux, so it's safe to assume that a pid of -1 won't work. It all depends on what you're trying to achieve (the EINVAL or the ESRCH requirement).
	Testing to make sure that a pid isn't running via kill(pid, 0) is another semi-viable method, but it's also racy.
Cheers,
-Garrett
------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] sched_getaffinity: fix possible fail in negative case

[Caspar Zhang]

On 03/09/2012 03:19 PM, Garrett Cooper wrote:
> On Mar 8, 2012, at 9:28 PM, Caspar Zhang wrote:
> 
>>
>> in negative case:
>>
>>    QUICK_TEST(sched_getaffinity(getpid()+1, len, mask));
>>
>> sometimes getpid()+1 will exist if there're other processes running.
>> This patch uses a different way to make sure the required pid doesn't
>> exist by forking a child and terminating it. The pid of the terminated
>> child will not be used for a short time and it's enough for the test to
>> continue.
> 
> 	pid_t is signed on FreeBSD and Linux, so it's safe to assume that a pid of -1 won't work. It all depends on what you're trying to achieve (the EINVAL or the ESRCH requirement).
> 	Testing to make sure that a pid isn't running via kill(pid, 0) is another semi-viable method, but it's also racy.
> Cheers,
> -Garrett

I'll keep your test case and remove the getpid()+1 one.

Thanks,
Caspar

------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] sched_getaffinity: fix possible fail in negative case

[Cyril Hrubis]

Hi!
> > in negative case:
> > 
> >    QUICK_TEST(sched_getaffinity(getpid()+1, len, mask));
> > 
> > sometimes getpid()+1 will exist if there're other processes running.
> > This patch uses a different way to make sure the required pid doesn't
> > exist by forking a child and terminating it. The pid of the terminated
> > child will not be used for a short time and it's enough for the test to
> > continue.
> 
> 	pid_t is signed on FreeBSD and Linux, so it's safe to assume that a pid of -1 won't work. It all depends on what you're trying to achieve (the EINVAL or the ESRCH requirement).
> 	Testing to make sure that a pid isn't running via kill(pid, 0) is another semi-viable method, but it's also racy.

On Linux we could use /proc/sys/kernel/pid_max value + 1 value too. But
that would probably lead to EINVAL as well as -1.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list