Skip to contentSkip to conversation

(no subject)

LTP In review Joerg Vehlow 2 participants 9 messages Nov 15, 2021

lore
Share & tools
lore
#0 lore raw

Subject: [PATCH 0/3] realtime/matrix_mult: Fix test execution

Hi,

this fixes the test execution of matrix_mult.
The workload function was optimized away, because the variables used for calculations were unreachable outside of the function.

Joerg

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#3 lore raw

PATCH 2/3 realtime/librttest: Fix memory leaks

From: Joerg Vehlow <[email protected]>

---
 testcases/realtime/lib/librttest.c | 4 ++++
 1 file changed, 4 insertions(+)
testcases/realtime/lib/librttest.c +4 -0
index 89e2bfc6e..eaa623b72 100644
@@ -226,6 +226,8 @@ int rt_init_long(const char *options, const struct option *longopts,
 
 	calibrate_busyloop();
 
+	free(all_options);
+
 	/*
 	 * atexit() order matters here - buffer_print() will be called before
 	 * buffer_fini().
@@ -378,6 +380,7 @@ void join_thread(int i)
 		if (t->pthread)
 			pthread_join(t->pthread, NULL);
 		list_del(&t->_threads);
+		free(t);
 	}
 }
 
@@ -397,6 +400,7 @@ void join_threads(void)
 		if (p->pthread)
 			pthread_join(p->pthread, NULL);
 		list_del(&p->_threads);
+		free(p);
 	}
 }
 
Signature · 5 lines
2.25.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#5 lore raw

PATCH 3/3 realtime/matrix_mult: Fix test optimization

In reply to Joerg Vehlow

From: Joerg Vehlow <[email protected]>

The actual load of the test was optimized away, because there was no way to reach the memory used for calculation from the outside of the function.

To fix this, the memory used for the matrices is now allocated in the main thread and passed to the workload function as a pointer. This should prevent optimization and also allows for the matrices to be bigger, without overflowing the stack.

---
 .../realtime/func/matrix_mult/matrix_mult.c   | 115 ++++++++----------
 1 file changed, 52 insertions(+), 63 deletions(-)
testcases/realtime/func/matrix_mult/matrix_mult.c +52 -63
index e702c0ff9..6d4da89bd 100644
@@ -1,40 +1,17 @@
-/******************************************************************************
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) International Business Machines  Corp., 2007, 2008
+ * 
+ * Authors: Darren Hart <[email protected]>
+ *          Dinakar Guniguntala <[email protected]>
+ */
+/*\
+ * [Description]
  *
- *   Copyright © International Business Machines  Corp., 2007, 2008
- *
- *   This program is free software;  you can redistribute it and/or modify
- *   it under the terms of the GNU General Public License as published by
- *   the Free Software Foundation; either version 2 of the License, or
- *   (at your option) any later version.
- *
- *   This program is distributed in the hope that it will be useful,
- *   but WITHOUT ANY WARRANTY;  without even the implied warranty of
- *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See
- *   the GNU General Public License for more details.
- *
- *   You should have received a copy of the GNU General Public License
- *   along with this program;  if not, write to the Free Software
- *   Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- *
- * NAME
- *      matrix_mult.c
- *
- * DESCRIPTION
- *      Compare running sequential matrix multiplication routines
- *      to running them in parallel to judge mutliprocessor
- *      performance
- *
- * USAGE:
- *      Use run_auto.sh script in current directory to build and run test.
- *
- * AUTHOR
- *      Darren Hart <[email protected]>
- *
- * HISTORY
- *      2007-Mar-09:  Initial version by Darren Hart <[email protected]>
- *      2008-Feb-26:  Closely emulate jvm Dinakar Guniguntala <[email protected]>
- *
- *****************************************************************************/
+ * Compare running sequential matrix multiplication routines
+ * to running them in parallel to judge multiprocessor
+ * performance
+ */
 
 #include <stdio.h>
 #include <stdlib.h>
@@ -69,9 +46,14 @@ static int iterations_percpu;
 stats_container_t sdat, cdat, *curdat;
 stats_container_t shist, chist;
 static pthread_barrier_t mult_start;
-static pthread_mutex_t mutex_cpu;
 
-void usage(void)
+struct matrices {
+	double A[MATRIX_SIZE][MATRIX_SIZE];
+	double B[MATRIX_SIZE][MATRIX_SIZE];
+	double C[MATRIX_SIZE][MATRIX_SIZE];
+};
+
+static void usage(void)
 {
 	rt_help();
 	printf("matrix_mult specific options:\n");
@@ -80,7 +62,7 @@ void usage(void)
 	printf("  -i#	   #: number of iterations\n");
 }
 
-int parse_args(int c, char *v)
+static int parse_args(int c, char *v)
 {
 	int handled = 1;
 	switch (c) {
@@ -100,7 +82,7 @@ int parse_args(int c, char *v)
 	return handled;
 }
 
-void matrix_init(double A[MATRIX_SIZE][MATRIX_SIZE],
+static void matrix_init(double A[MATRIX_SIZE][MATRIX_SIZE],
 		 double B[MATRIX_SIZE][MATRIX_SIZE])
 {
 	int i, j;
@@ -112,41 +94,39 @@ void matrix_init(double A[MATRIX_SIZE][MATRIX_SIZE],
 	}
 }
 
-void matrix_mult(int m_size)
+static void matrix_mult(struct matrices *matrices)
 {
-	double A[m_size][m_size];
-	double B[m_size][m_size];
-	double C[m_size][m_size];
 	int i, j, k;
 
-	matrix_init(A, B);
-	for (i = 0; i < m_size; i++) {
-		int i_m = m_size - i;
-		for (j = 0; j < m_size; j++) {
-			double sum = A[i_m][j] * B[j][i];
-			for (k = 0; k < m_size; k++)
-				sum += A[i_m][k] * B[k][j];
-			C[i][j] = sum;
+	matrix_init(matrices->A, matrices->B);
+	for (i = 0; i < MATRIX_SIZE; i++) {
+		int i_m = MATRIX_SIZE - i;
+		for (j = 0; j < MATRIX_SIZE; j++) {
+			double sum = matrices->A[i_m][j] *  matrices->B[j][i];
+			for (k = 0; k < MATRIX_SIZE; k++)
+				sum +=  matrices->A[i_m][k] *  matrices->B[k][j];
+			 matrices->C[i][j] = sum;
 		}
 	}
 }
 
-void matrix_mult_record(int m_size, int index)
+static void matrix_mult_record(struct matrices *matrices, int index)
 {
 	nsec_t start, end, delta;
 	int i;
 
 	start = rt_gettime();
 	for (i = 0; i < ops; i++)
-		matrix_mult(MATRIX_SIZE);
+		matrix_mult(matrices);
 	end = rt_gettime();
 	delta = (long)((end - start) / NS_PER_US);
 	curdat->records[index].x = index;
 	curdat->records[index].y = delta;
 }
 
-int set_affinity(void)
+static int set_affinity(void)
 {
+	static pthread_mutex_t mutex_cpu = PTHREAD_MUTEX_INITIALIZER;
 	cpu_set_t mask;
 	int cpuid;
 
@@ -166,9 +146,10 @@ int set_affinity(void)
 	return -1;
 }
 
-void *concurrent_thread(void *thread)
+static void *concurrent_thread(void *thread)
 {
 	struct thread *t = (struct thread *)thread;
+	struct matrices *matrices = (struct matrices *) t->arg;
 	int thread_id = (intptr_t) t->id;
 	int cpuid;
 	int i;
@@ -183,18 +164,23 @@ void *concurrent_thread(void *thread)
 	index = iterations_percpu * thread_id;	/* To avoid stats overlapping */
 	pthread_barrier_wait(&mult_start);
 	for (i = 0; i < iterations_percpu; i++)
-		matrix_mult_record(MATRIX_SIZE, index++);
+		matrix_mult_record(matrices, index++);
 
 	return NULL;
 }
 
-int main_thread(void)
+static int main_thread(void)
 {
 	int ret, i, j;
 	nsec_t start, end;
 	long smin = 0, smax = 0, cmin = 0, cmax = 0, delta = 0;
 	float savg, cavg;
 	int cpuid;
+	struct matrices *matrices[numcpus];
+
+	for (i = 0; i < numcpus; ++i) {
+		matrices[i] = malloc(sizeof(struct matrices));
+	}
 
 	if (stats_container_init(&sdat, iterations) ||
 	    stats_container_init(&shist, HIST_BUCKETS) ||
@@ -205,12 +191,11 @@ int main_thread(void)
 		exit(1);
 	}
 
-	tids = malloc(sizeof(int) * numcpus);
+	tids = calloc(numcpus, sizeof(int));
 	if (!tids) {
 		perror("malloc");
 		exit(1);
 	}
-	memset(tids, 0, numcpus);
 
 	cpuid = set_affinity();
 	if (cpuid == -1) {
@@ -223,8 +208,9 @@ int main_thread(void)
 	curdat->index = iterations - 1;
 	printf("\nRunning sequential operations\n");
 	start = rt_gettime();
-	for (i = 0; i < iterations; i++)
-		matrix_mult_record(MATRIX_SIZE, i);
+	for (i = 0; i < iterations; i++) {
+		matrix_mult_record(matrices[0], i);
+	}
 	end = rt_gettime();
 	delta = (long)((end - start) / NS_PER_US);
 
@@ -256,7 +242,7 @@ int main_thread(void)
 	online_cpu_id = -1;	/* Redispatch cpus */
 	/* Create numcpus-1 concurrent threads */
 	for (j = 0; j < numcpus; j++) {
-		tids[j] = create_fifo_thread(concurrent_thread, NULL, PRIO);
+		tids[j] = create_fifo_thread(concurrent_thread, matrices[j], PRIO);
 		if (tids[j] == -1) {
 			printf
 			    ("Thread creation failed (max threads exceeded?)\n");
@@ -308,6 +294,9 @@ int main_thread(void)
 	     criteria);
 	printf("Result: %s\n", ret ? "FAIL" : "PASS");
 
+	for (i = 0; i < numcpus; i++)
+		free(matrices[i]);
+
 	return ret;
 }
 
Signature · 5 lines
2.25.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#1 lore raw

PATCH 1/3 realtime/librttest: Fix functions with empty parameters

In reply to Joerg Vehlow

From: Joerg Vehlow <[email protected]>

---
 testcases/realtime/include/librttest.h | 16 ++++++++--------
 testcases/realtime/lib/librttest.c     |  2 +-
 2 files changed, 9 insertions(+), 9 deletions(-)
testcases/realtime/include/librttest.h +8 -8
index b829f4089..8e3421869 100644
@@ -58,8 +58,8 @@
 #include "list.h"
 #include "realtime_config.h"
 
-extern void setup();
-extern void cleanup();
+extern void setup(void);
+extern void cleanup(int i);
 
 extern int optind, opterr, optopt;
 extern char *optarg;
@@ -147,15 +147,15 @@ static inline void atomic_set(int i, atomic_t *v)
 
 /* buffer_init: initialize the buffered printing system
  */
-void buffer_init();
+void buffer_init(void);
 
 /* buffer_print: prints the contents of the buffer
  */
-void buffer_print();
+void buffer_print(void);
 
 /* buffer_fini: destroy the buffer
  */
-void buffer_fini();
+void buffer_fini(void);
 
 /* debug: do debug prints at level L (see DBG_* below).  If buffer_init
  * has been called previously, this will print to the internal memory
@@ -185,7 +185,7 @@ static volatile int _debug_count = 0;
 #define DBG_DEBUG 4
 
 /* rt_help: print help for standard args */
-void rt_help();
+void rt_help(void);
 
 /* rt_init_long: initialize librttest
  * options: pass in an getopt style string of options -- e.g. "ab:cd::e:"
@@ -269,7 +269,7 @@ void all_threads_quit(void);
 /* join_threads: wait for all threads to finish
  * (calls all_threads_quit interally)
  */
-void join_threads();
+void join_threads(void);
 
 /* get_thread: return a struct thread pointer from the list */
 struct thread * get_thread(int i);
@@ -317,7 +317,7 @@ void rt_nanosleep_until(nsec_t ns);
 
 /* rt_gettime: get CLOCK_MONOTONIC time in nanoseconds
  */
-nsec_t rt_gettime();
+nsec_t rt_gettime(void);
 
 /* busy_work_ms: do busy work for ms milliseconds
  */
testcases/realtime/lib/librttest.c +1 -1
index 722d99d50..89e2bfc6e 100644
@@ -276,7 +276,7 @@ void cleanup(int i)
 		exit(i);
 }
 
-void setup()
+void setup(void)
 {
 	signal(SIGINT, cleanup);
 	signal(SIGQUIT, cleanup);
Signature · 5 lines
2.25.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#2 lore raw

Re: [LTP] [PATCH 1/3] realtime/librttest: Fix functions with empty parameters

Hi!
Pushed, thanks.

Signature · 5 lines
Cyril Hrubis
[email protected]

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#6 lore raw

Re: [LTP] [PATCH 3/3] realtime/matrix_mult: Fix test optimization

In reply to Joerg Vehlow

Hi!

Quoted 39 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +9 -29
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) International Business Machines  Corp., 2007, 2008
+ * 
+ * Authors: Darren Hart <[email protected]>
+ *          Dinakar Guniguntala <[email protected]>
+ */
+/*\
+ * [Description]
  *
- *   Copyright ?? International Business Machines  Corp., 2007, 2008
- *
- *   This program is free software;  you can redistribute it and/or modify
- *   it under the terms of the GNU General Public License as published by
- *   the Free Software Foundation; either version 2 of the License, or
- *   (at your option) any later version.
- *
- *   This program is distributed in the hope that it will be useful,
- *   but WITHOUT ANY WARRANTY;  without even the implied warranty of
- *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See
- *   the GNU General Public License for more details.
- *
- *   You should have received a copy of the GNU General Public License
- *   along with this program;  if not, write to the Free Software
- *   Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- *
- * NAME
- *      matrix_mult.c
- *
- * DESCRIPTION
- *      Compare running sequential matrix multiplication routines
- *      to running them in parallel to judge mutliprocessor
- *      performance
- *
- * USAGE:
- *      Use run_auto.sh script in current directory to build and run test.
- *
- * AUTHOR
- *      Darren Hart <[email protected]>

testcases/realtime/func/matrix_mult/matrix_mult.c line 14

Quoted 10 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +4 -6
- *
- * HISTORY
- *      2007-Mar-09:  Initial version by Darren Hart <[email protected]>
- *      2008-Feb-26:  Closely emulate jvm Dinakar Guniguntala <[email protected]>
- *
- *****************************************************************************/
+ * Compare running sequential matrix multiplication routines
+ * to running them in parallel to judge multiprocessor
+ * performance
+ */

Ideally the comment changes should be in a separate patch...

Quoted 29 lines

#include <stdio.h>
#include <stdlib.h>

testcases/realtime/func/matrix_mult/matrix_mult.c +8 -3
@@ -69,9 +46,14 @@ static int iterations_percpu;
 stats_container_t sdat, cdat, *curdat;
 stats_container_t shist, chist;
 static pthread_barrier_t mult_start;
-static pthread_mutex_t mutex_cpu;
 
-void usage(void)
+struct matrices {
+	double A[MATRIX_SIZE][MATRIX_SIZE];
+	double B[MATRIX_SIZE][MATRIX_SIZE];
+	double C[MATRIX_SIZE][MATRIX_SIZE];
+};
+
+static void usage(void)
 {
 	rt_help();
 	printf("matrix_mult specific options:\n");
@@ -80,7 +62,7 @@ void usage(void)
 	printf("  -i#	   #: number of iterations\n");
 }
 
-int parse_args(int c, char *v)
+static int parse_args(int c, char *v)
 {
 	int handled = 1;
 	switch (c) {

testcases/realtime/func/matrix_mult/matrix_mult.c int parse_args(int c, char *v) · line 86

Quoted 6 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +1 -1
@@ -100,7 +82,7 @@ int parse_args(int c, char *v)
 	return handled;
 }
 
-void matrix_init(double A[MATRIX_SIZE][MATRIX_SIZE],
+static void matrix_init(double A[MATRIX_SIZE][MATRIX_SIZE],
 		 double B[MATRIX_SIZE][MATRIX_SIZE])

These static function conversions should ideally be in a separate patch as well. At least the patch description does not mention either of these changes.

Quoted 44 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +11 -14
 {
 	int i, j;
@@ -112,41 +94,39 @@ void matrix_init(double A[MATRIX_SIZE][MATRIX_SIZE],
 	}
 }
 
-void matrix_mult(int m_size)
+static void matrix_mult(struct matrices *matrices)
 {
-	double A[m_size][m_size];
-	double B[m_size][m_size];
-	double C[m_size][m_size];
 	int i, j, k;
 
-	matrix_init(A, B);
-	for (i = 0; i < m_size; i++) {
-		int i_m = m_size - i;
-		for (j = 0; j < m_size; j++) {
-			double sum = A[i_m][j] * B[j][i];
-			for (k = 0; k < m_size; k++)
-				sum += A[i_m][k] * B[k][j];
-			C[i][j] = sum;
+	matrix_init(matrices->A, matrices->B);
+	for (i = 0; i < MATRIX_SIZE; i++) {
+		int i_m = MATRIX_SIZE - i;
+		for (j = 0; j < MATRIX_SIZE; j++) {
+			double sum = matrices->A[i_m][j] *  matrices->B[j][i];
+			for (k = 0; k < MATRIX_SIZE; k++)
+				sum +=  matrices->A[i_m][k] *  matrices->B[k][j];
+			 matrices->C[i][j] = sum;
 		}
 	}
 }
 
-void matrix_mult_record(int m_size, int index)
+static void matrix_mult_record(struct matrices *matrices, int index)
 {
 	nsec_t start, end, delta;
 	int i;
 
 	start = rt_gettime();
 	for (i = 0; i < ops; i++)
-		matrix_mult(MATRIX_SIZE);
+		matrix_mult(matrices);

testcases/realtime/func/matrix_mult/matrix_mult.c void matrix_init(double A[MATRIX_SIZE][MATRIX_SIZE], · line 129

Quoted 10 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +2 -1
 	end = rt_gettime();
 	delta = (long)((end - start) / NS_PER_US);
 	curdat->records[index].x = index;
 	curdat->records[index].y = delta;
 }
 
-int set_affinity(void)
+static int set_affinity(void)
 {
+	static pthread_mutex_t mutex_cpu = PTHREAD_MUTEX_INITIALIZER;

As well as this change, it looks okay, but it's not listed in the patch description.

Quoted 56 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +10 -5
 	cpu_set_t mask;
 	int cpuid;
 
@@ -166,9 +146,10 @@ int set_affinity(void)
 	return -1;
 }
 
-void *concurrent_thread(void *thread)
+static void *concurrent_thread(void *thread)
 {
 	struct thread *t = (struct thread *)thread;
+	struct matrices *matrices = (struct matrices *) t->arg;
 	int thread_id = (intptr_t) t->id;
 	int cpuid;
 	int i;
@@ -183,18 +164,23 @@ void *concurrent_thread(void *thread)
 	index = iterations_percpu * thread_id;	/* To avoid stats overlapping */
 	pthread_barrier_wait(&mult_start);
 	for (i = 0; i < iterations_percpu; i++)
-		matrix_mult_record(MATRIX_SIZE, index++);
+		matrix_mult_record(matrices, index++);
 
 	return NULL;
 }
 
-int main_thread(void)
+static int main_thread(void)
 {
 	int ret, i, j;
 	nsec_t start, end;
 	long smin = 0, smax = 0, cmin = 0, cmax = 0, delta = 0;
 	float savg, cavg;
 	int cpuid;
+	struct matrices *matrices[numcpus];
+
+	for (i = 0; i < numcpus; ++i) {
+		matrices[i] = malloc(sizeof(struct matrices));
+	}
 
 	if (stats_container_init(&sdat, iterations) ||
 	    stats_container_init(&shist, HIST_BUCKETS) ||
@@ -205,12 +191,11 @@ int main_thread(void)
 		exit(1);
 	}
 
-	tids = malloc(sizeof(int) * numcpus);
+	tids = calloc(numcpus, sizeof(int));
 	if (!tids) {
 		perror("malloc");
 		exit(1);
 	}
-	memset(tids, 0, numcpus);
 
 	cpuid = set_affinity();
 	if (cpuid == -1) {

testcases/realtime/func/matrix_mult/matrix_mult.c int main_thread(void) · line 213

Quoted 8 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +3 -2
@@ -223,8 +208,9 @@ int main_thread(void)
 	curdat->index = iterations - 1;
 	printf("\nRunning sequential operations\n");
 	start = rt_gettime();
-	for (i = 0; i < iterations; i++)
-		matrix_mult_record(MATRIX_SIZE, i);
+	for (i = 0; i < iterations; i++) {
+		matrix_mult_record(matrices[0], i);
+	}

And here LMKL coding style prefers to avoid the curly braces, but that is very minor.

Quoted 28 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +4 -1
 	end = rt_gettime();
 	delta = (long)((end - start) / NS_PER_US);
 
@@ -256,7 +242,7 @@ int main_thread(void)
 	online_cpu_id = -1;	/* Redispatch cpus */
 	/* Create numcpus-1 concurrent threads */
 	for (j = 0; j < numcpus; j++) {
-		tids[j] = create_fifo_thread(concurrent_thread, NULL, PRIO);
+		tids[j] = create_fifo_thread(concurrent_thread, matrices[j], PRIO);
 		if (tids[j] == -1) {
 			printf
 			    ("Thread creation failed (max threads exceeded?)\n");
@@ -308,6 +294,9 @@ int main_thread(void)
 	     criteria);
 	printf("Result: %s\n", ret ? "FAIL" : "PASS");
 
+	for (i = 0; i < numcpus; i++)
+		free(matrices[i]);
+
 	return ret;
 }
 
Signature · 5 lines
2.25.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
Signature · 5 lines
Cyril Hrubis
[email protected]

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#7 lore raw

Re: [LTP] [PATCH 3/3] realtime/matrix_mult: Fix test optimization

Hi Cyril,

Quoted 7 lines

Ideally the comment changes should be in a separate patch...

These static function conversions should ideally be in a separate patch as well. At least the patch description does not mention either of these changes.

I just did some cleanup here. Would it be ok, to do all cleanup in one commit and the fix in a second one?

Quoted 7 lines

testcases/realtime/func/matrix_mult/matrix_mult.c void matrix_init(double A[MATRIX_SIZE][MATRIX_SIZE], · line 129

Quoted 5 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +2 -1
-int set_affinity(void)
+static int set_affinity(void)
  {
+	static pthread_mutex_t mutex_cpu = PTHREAD_MUTEX_INITIALIZER;

As well as this change, it looks okay, but it's not listed in the patch description.

Right, that is also just cleanup. The mutex is only used inside of this function, and I figured using the static initializer is  better, than 0-initialization.

Quoted 5 lines

testcases/realtime/func/matrix_mult/matrix_mult.c int main_thread(void) · line 213

Quoted 3 lines
testcases/realtime/func/matrix_mult/matrix_mult.c +3 -0
+	for (i = 0; i < iterations; i++) {
+		matrix_mult_record(matrices[0], i);
+	}

And here LMKL coding style prefers to avoid the curly braces, but that is very minor.

Yes sorry, I had this without curly braces, than added some debug code and forgot to remove them again. This btw. is the reason, why I don't like this rule (but I will not try to argue about it here)

Joerg

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#8 lore raw

Re: [LTP] [PATCH 3/3] realtime/matrix_mult: Fix test optimization

Hi!

Quoted 5 lines
Quoted 3 lines

These static function conversions should ideally be in a separate patch as well. At least the patch description does not mention either of these changes.

I just did some cleanup here. Would it be ok, to do all cleanup in one commit and the fix in a second one?

Yes, that would be perfect.

Signature · 5 lines
Cyril Hrubis
[email protected]

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