[LTP] [PATCH] sched_setparam/2-1.c and sched_setparam/2-2.c

[LTP] [PATCH] sched_setparam/2-1.c and sched_setparam/2-2.c

[Song.Li]

From: LiSong <Song.Li@windriver.com>

There is no guarantee that the LOOP of the child proecess
can be certainly big enough on any device at any time.
So if set_affinity function is available,
use it to bind the test processes to a single CPU
rather than using the LOOP.

Signed-off-by: Song.Li <Song.Li@windriver.com>
---
 .../conformance/interfaces/sched_setparam/2-1.c    |   26 +++++++++++++++----
 .../conformance/interfaces/sched_setparam/2-2.c    |   26 +++++++++++++++----
 2 files changed, 40 insertions(+), 12 deletions(-)

diff --git a/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-1.c b/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-1.c
index c548dd4..93f6443 100644
--- a/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-1.c
+++ b/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-1.c
@@ -29,6 +29,7 @@
  *
  */
 
+#define _GNU_SOURCE
 #include <sched.h>
 #include <stdio.h>
 #include <signal.h>
@@ -48,6 +49,8 @@
 # include <sys/pstat.h>
 #endif
 
+#include <affinity.h>
+
 #define NB_LOOP         20000000
 #define NB_LOOP_CHILD  200000000 /* shall be much greater than NB_LOOP */
 
@@ -119,11 +122,22 @@ int main() {
 	int *child_pid;
 	float ratio;
 
-	nb_child = get_ncpu();
-	if (nb_child == -1) {
-		printf("Can not get the number of CPUs of your machine.\n");
-		return PTS_UNRESOLVED;
-	}
+        /* Only use a single CPU and one child process when set_affinity is availaible
+           It's because no matter what value of the counter is set to, 
+           There is no guarantee that the LOOP of the child can be certainly big enough on any device at any time.
+        */
+        int rc = set_affinity(0);
+        if (rc) {
+		nb_child = get_ncpu();
+		if (nb_child == -1) {
+			printf("Can not get the number of CPUs of your machine.\n");
+			return PTS_UNRESOLVED;
+		}
+        }
+        else {
+		nb_child = 1;
+        }
+
 	child_pid = malloc(nb_child);
 
 	param.sched_priority = (sched_get_priority_min(SCHED_FIFO) +
@@ -208,4 +222,4 @@ int main() {
 		return PTS_UNRESOLVED;
 	}
 
-}
\ No newline at end of file
+}
diff --git a/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-2.c b/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-2.c
index 2b71893..c17c5b6 100644
--- a/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-2.c
+++ b/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-2.c
@@ -29,6 +29,7 @@
  *
  */
 
+#define _GNU_SOURCE
 #include <sched.h>
 #include <stdio.h>
 #include <signal.h>
@@ -48,6 +49,8 @@
 # include <sys/pstat.h>
 #endif
 
+#include <affinity.h>
+
 #define NB_LOOP         20000000
 #define NB_LOOP_CHILD  200000000 /* shall be much greater than NB_LOOP */
 
@@ -119,16 +122,27 @@ int main() {
 	int *child_pid;
 	float ratio;
 
-	nb_child = get_ncpu();
-	if (nb_child == -1) {
-		printf("Can not get the number of CPUs of your machine.\n");
-		return PTS_UNRESOLVED;
-	}
+        /* Only use a single CPU and one child process when set_affinity is availaible
+           It's because no matter what value of the counter is set to, 
+           There is no guarantee that the LOOP of the child can be certainly big enough on any device at any time.
+        */
+        int rc = set_affinity(0);
+        if (rc) {
+		nb_child = get_ncpu();
+		if (nb_child == -1) {
+			printf("Can not get the number of CPUs of your machine.\n");
+			return PTS_UNRESOLVED;
+		}
+        }
+        else {
+		nb_child = 1;
+        }
 	child_pid = malloc(nb_child);
 	if (child_pid == NULL) {
 		printf("malloc failed\n");
 		return PTS_UNRESOLVED;
 	}
+
 	param.sched_priority = (sched_get_priority_min(SCHED_RR) +
 				 sched_get_priority_max(SCHED_RR)) / 2;
 
@@ -211,4 +225,4 @@ int main() {
 		return PTS_UNRESOLVED;
 	}
 
-}
\ No newline at end of file
+}
-- 
1.7.5.4


------------------------------------------------------------------------------
Write once. Port to many.
Get the SDK and tools to simplify cross-platform app development. Create 
new or port existing apps to sell to consumers worldwide. Explore the 
Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
http://p.sf.net/sfu/intel-appdev
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] sched_setparam/2-1.c and sched_setparam/2-2.c

[Cyril Hrubis]

Hi!
> There is no guarantee that the LOOP of the child proecess
> can be certainly big enough on any device at any time.
> So if set_affinity function is available,
> use it to bind the test processes to a single CPU
> rather than using the LOOP.

The patch logic is looking good, there are some style problems though.

Please use checkpatch.pl (that is packed with linux kernel) and resend.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
infrastructure or vast IT resources to deliver seamless, secure access to
virtual desktops. With this all-in-one solution, easily deploy virtual 
desktops for less than the cost of PCs and save 60% on VDI infrastructure 
costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] sched_setparam/2-1.c and sched_setparam/2-2.c

[Cyril Hrubis]

Hi!
> > There is no guarantee that the LOOP of the child proecess
> > can be certainly big enough on any device at any time.
> > So if set_affinity function is available,
> > use it to bind the test processes to a single CPU
> > rather than using the LOOP.
> 
> The patch logic is looking good, there are some style problems though.
> 
> Please use checkpatch.pl (that is packed with linux kernel) and resend.

And also I would suggest moving the get_ncpu() into common header just
as the affinity is (in another patch).

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
infrastructure or vast IT resources to deliver seamless, secure access to
virtual desktops. With this all-in-one solution, easily deploy virtual 
desktops for less than the cost of PCs and save 60% on VDI infrastructure 
costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

[LTP] [PATCH] sched_setparam/2-1.c and sched_setparam/2-2.c

[Song.Li]

From: LiSong <Song.Li@windriver.com>

There is no guarantee that the LOOP of the child proecess
can be certainly big enough on any device at any time.
So if set_affinity function is available,
use it to bind the test processes to a single CPU
rather than using the LOOP.

Signed-off-by: Song.Li <Song.Li@windriver.com>
---
 .../conformance/interfaces/sched_setparam/2-1.c    |   26 ++++++++++++++++----
 .../conformance/interfaces/sched_setparam/2-2.c    |   26 ++++++++++++++++----
 2 files changed, 42 insertions(+), 10 deletions(-)

diff --git a/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-1.c b/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-1.c
index c548dd4..2d20f9f 100644
--- a/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-1.c
+++ b/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-1.c
@@ -29,6 +29,7 @@
  *
  */
 
+#define _GNU_SOURCE
 #include <sched.h>
 #include <stdio.h>
 #include <signal.h>
@@ -48,6 +49,8 @@
 # include <sys/pstat.h>
 #endif
 
+#include <affinity.h>
+
 #define NB_LOOP         20000000
 #define NB_LOOP_CHILD  200000000 /* shall be much greater than NB_LOOP */
 
@@ -119,11 +122,24 @@ int main() {
 	int *child_pid;
 	float ratio;
 
-	nb_child = get_ncpu();
-	if (nb_child == -1) {
-		printf("Can not get the number of CPUs of your machine.\n");
-		return PTS_UNRESOLVED;
+	/* Only use a single CPU and one child process
+	when set_affinity is availaible.It's because
+	no matter what value of the counter is set to,
+	There is no guarantee that the LOOP of the child
+	can be certainly big enough on any device at any time.
+	*/
+	int rc = set_affinity(0);
+	if (rc) {
+		nb_child = get_ncpu();
+		if (nb_child == -1) {
+			printf("Can not get the number of"
+				"CPUs of your machine.\n");
+			return PTS_UNRESOLVED;
+		}
+	} else {
+		nb_child = 1;
 	}
+
 	child_pid = malloc(nb_child);
 
 	param.sched_priority = (sched_get_priority_min(SCHED_FIFO) +
@@ -208,4 +224,4 @@ int main() {
 		return PTS_UNRESOLVED;
 	}
 
-}
\ No newline at end of file
+}
diff --git a/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-2.c b/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-2.c
index 2b71893..08cdc34 100644
--- a/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-2.c
+++ b/testcases/open_posix_testsuite/conformance/interfaces/sched_setparam/2-2.c
@@ -29,6 +29,7 @@
  *
  */
 
+#define _GNU_SOURCE
 #include <sched.h>
 #include <stdio.h>
 #include <signal.h>
@@ -48,6 +49,8 @@
 # include <sys/pstat.h>
 #endif
 
+#include <affinity.h>
+
 #define NB_LOOP         20000000
 #define NB_LOOP_CHILD  200000000 /* shall be much greater than NB_LOOP */
 
@@ -119,11 +122,24 @@ int main() {
 	int *child_pid;
 	float ratio;
 
-	nb_child = get_ncpu();
-	if (nb_child == -1) {
-		printf("Can not get the number of CPUs of your machine.\n");
-		return PTS_UNRESOLVED;
+	/* Only use a single CPU and one child process
+	when set_affinity is availaible.It's because
+	no matter what value of the counter is set to,
+	There is no guarantee that the LOOP of the child
+	can be certainly big enough on any device at any time.
+	*/
+	int rc = set_affinity(0);
+	if (rc) {
+		nb_child = get_ncpu();
+		if (nb_child == -1) {
+			printf("Can not get the number of"
+				"CPUs of your machine.\n");
+			return PTS_UNRESOLVED;
+		}
+	} else {
+		nb_child = 1;
 	}
+
 	child_pid = malloc(nb_child);
 	if (child_pid == NULL) {
 		printf("malloc failed\n");
@@ -211,4 +227,4 @@ int main() {
 		return PTS_UNRESOLVED;
 	}
 
-}
\ No newline at end of file
+}
-- 
1.7.5.4


------------------------------------------------------------------------------
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
infrastructure or vast IT resources to deliver seamless, secure access to
virtual desktops. With this all-in-one solution, easily deploy virtual 
desktops for less than the cost of PCs and save 60% on VDI infrastructure 
costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] sched_setparam/2-1.c and sched_setparam/2-2.c

[Cyril Hrubis]

Hi!
> There is no guarantee that the LOOP of the child proecess
> can be certainly big enough on any device at any time.
> So if set_affinity function is available,
> use it to bind the test processes to a single CPU
> rather than using the LOOP.
> 
> Signed-off-by: Song.Li <Song.Li@windriver.com>

Commited, thanks.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
infrastructure or vast IT resources to deliver seamless, secure access to
virtual desktops. With this all-in-one solution, easily deploy virtual 
desktops for less than the cost of PCs and save 60% on VDI infrastructure 
costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list