All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Joel Fernandes (Google)" <joel@joelfernandes.org>
To: linux-kernel@vger.kernel.org, Shuah Khan <shuah@kernel.org>,
	Nathan Chancellor <nathan@kernel.org>,
	Nick Desaulniers <ndesaulniers@google.com>,
	Bill Wendling <morbo@google.com>,
	Justin Stitt <justinstitt@google.com>
Cc: Suleiman Souhlal <suleiman@google.com>,
	Youssef Esmat <youssefesmat@google.com>,
	Steven Rostedt <rostedt@goodmis.org>,
	David Vernet <void@manifault.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	"Paul E . McKenney" <paulmck@kernel.org>,
	joseph.salisbury@canonical.com,
	Dietmar Eggemann <dietmar.eggemann@arm.com>,
	Ben Segall <bsegall@google.com>, Mel Gorman <mgorman@suse.de>,
	Daniel Bristot de Oliveira <bristot@redhat.com>,
	Valentin Schneider <vschneid@redhat.com>,
	Luca Abeni <luca.abeni@santannapisa.it>,
	Tommaso Cucinotta <tommaso.cucinotta@santannapisa.it>,
	Vineeth Pillai <vineeth@bitbyteword.org>,
	Shuah Khan <skhan@linuxfoundation.org>,
	Phil Auld <pauld@redhat.com>,
	"Joel Fernandes (Google)" <joel@joelfernandes.org>,
	linux-kselftest@vger.kernel.org, llvm@lists.linux.dev
Subject: [PATCH v2 07/15] selftests/sched: Add a test to verify that DL server works with core scheduling
Date: Tue, 12 Mar 2024 21:24:43 -0400	[thread overview]
Message-ID: <20240313012451.1693807-8-joel@joelfernandes.org> (raw)
In-Reply-To: <20240313012451.1693807-1-joel@joelfernandes.org>

This test verifies that DL server infrastructure gives CFS tasks a fixed
bandwidth even when RT tasks are being "core scheduled" on a core.
Verify that they are getting the expected bandwidth (and thus not being
starved).

Also verified that not having core scheduling fixes makes the test fail
as the CFS task gets no bandwidth.

Sample output:

 # Runtime of PID 97 is 4.440000 seconds
 # Runtime of PID 98 is 4.560000 seconds
 # Runtime of PID 99 is 4.550000 seconds
 ok 1 PASS

Notes about test that generated the sample output:

The test runs for 12 seconds. We check the runtimes at 9 seconds. We
expect the CFS task (PID 7) to get ~50% of the 9 seconds. The DL server
is configured for 50% bandwidth.

The RT tasks (PID 98, 99) each get 50% as well, because they run
concurrently on 2 hyperthreads of a core.

Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
---
 tools/testing/selftests/sched/Makefile        |  13 +-
 tools/testing/selftests/sched/common.c        |  24 ++
 tools/testing/selftests/sched/common.h        |   8 +
 .../selftests/sched/cs_dlserver_test.c        | 254 ++++++++++++++++++
 4 files changed, 290 insertions(+), 9 deletions(-)
 create mode 100644 tools/testing/selftests/sched/common.c
 create mode 100644 tools/testing/selftests/sched/common.h
 create mode 100644 tools/testing/selftests/sched/cs_dlserver_test.c

diff --git a/tools/testing/selftests/sched/Makefile b/tools/testing/selftests/sched/Makefile
index 099ee9213557..f491d741cb45 100644
--- a/tools/testing/selftests/sched/Makefile
+++ b/tools/testing/selftests/sched/Makefile
@@ -1,14 +1,9 @@
 # SPDX-License-Identifier: GPL-2.0+
+TEST_GEN_PROGS := cs_dlserver_test
 
-ifneq ($(shell $(CC) --version 2>&1 | head -n 1 | grep clang),)
-CLANG_FLAGS += -no-integrated-as
-endif
+cs_dlserver_test: cs_dlserver_test.c common.c
 
-CFLAGS += -O2 -Wall -g -I./ $(KHDR_INCLUDES) -Wl,-rpath=./ \
-	  $(CLANG_FLAGS)
-LDLIBS += -lpthread
-
-TEST_GEN_FILES := cs_prctl_test
-TEST_PROGS := cs_prctl_test
+CFLAGS += $(KHDR_INCLUDES)
+CFLAGS += -Wall
 
 include ../lib.mk
diff --git a/tools/testing/selftests/sched/common.c b/tools/testing/selftests/sched/common.c
new file mode 100644
index 000000000000..5cf0022acc8d
--- /dev/null
+++ b/tools/testing/selftests/sched/common.c
@@ -0,0 +1,24 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include "common.h"
+
+bool hyperthreading_enabled(void)
+{
+	FILE *file = fopen("/sys/devices/system/cpu/smt/active", "r");
+	char smt_active[2];
+
+	if (file == NULL) {
+		ksft_print_msg("Could not determine if hyperthreading is enabled\n");
+		return false;
+	}
+
+	if (fgets(smt_active, sizeof(smt_active), file)	== NULL) {
+		perror("Failed to read smt_active");
+		return false;
+	}
+	fclose(file);
+
+	if (smt_active[0] != '1')
+		return false;
+	return true;
+}
diff --git a/tools/testing/selftests/sched/common.h b/tools/testing/selftests/sched/common.h
new file mode 100644
index 000000000000..7bcedbd0ed99
--- /dev/null
+++ b/tools/testing/selftests/sched/common.h
@@ -0,0 +1,8 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#include <stdio.h>
+#include <stdbool.h>
+#include <dirent.h>
+#include "../kselftest.h"
+
+bool hyperthreading_enabled(void);
diff --git a/tools/testing/selftests/sched/cs_dlserver_test.c b/tools/testing/selftests/sched/cs_dlserver_test.c
new file mode 100644
index 000000000000..9f2a74a25686
--- /dev/null
+++ b/tools/testing/selftests/sched/cs_dlserver_test.c
@@ -0,0 +1,254 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Use the DL server infrastructure to give CFS tasks a fixed bandwidth
+ * even when RT tasks are being "core scheduled" on a core. Verify that
+ * they are getting the expected bandwidth (and thus not being starved).
+ *
+ * Copyright (c) 2024 Google.
+ * Author: Joel Fernandes <joel@joelfernandes.org>
+ *
+ * This library is free software; you can redistribute it and/or modify it
+ * under the terms of version 2.1 of the GNU Lesser General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This library 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 Lesser General Public License
+ * for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public License
+ * along with this library; if not, see <http://www.gnu.org/licenses>.
+ */
+
+#define _GNU_SOURCE
+
+#include <stdlib.h>
+#include <unistd.h>
+#include <sched.h>
+#include <time.h>
+#include <sys/wait.h>
+#include <sys/types.h>
+#include <sys/prctl.h>
+#include <fcntl.h>
+#include <string.h>
+
+#include "common.h"
+
+enum pid_type {PIDTYPE_PID = 0, PIDTYPE_TGID, PIDTYPE_PGID};
+
+#define RUN_TIME 12 // Running time of the test in seconds
+#define CORE_ID 0 // Assuming we're pinning processes to the first core
+#define DL_SERVER_DEBUGFS "/sys/kernel/debug/sched/fair_server"
+
+void write_server_debugfs(char *file, char *type, unsigned long value)
+{
+	char path[1024], buf[1024];
+	int fd, n;
+
+	snprintf(path, sizeof(path), "%s/%s/%s", DL_SERVER_DEBUGFS, file, type);
+	fd = open(path,	O_WRONLY);
+	if (fd == -1) {
+		perror("Failed to open file for writing");
+		return;
+	}
+	n = snprintf(buf, sizeof(buf), "%lu\n", value);
+	n = write(fd, buf, n);
+	if (n == -1)
+		perror("Failed to write file");
+
+	close(fd);
+}
+
+void write_dl_server_params(void)
+{
+	DIR *dir;
+	struct dirent *entry;
+
+	if (access(DL_SERVER_DEBUGFS, F_OK) == -1) {
+		perror("DL server debugfs not found, cannot set DL parameters.");
+		exit(EXIT_FAILURE);
+	}
+
+	dir = opendir(DL_SERVER_DEBUGFS);
+	if (dir	== NULL) {
+		perror("Failed to open directory");
+		exit(EXIT_FAILURE);
+	}
+
+	while ((entry = readdir(dir)) != NULL) {
+		if (strcmp(entry->d_name, ".") == 0 || strcmp(entry->d_name, "..") == 0)
+			continue;
+
+		write_server_debugfs(entry->d_name, "period", 100000000);
+		write_server_debugfs(entry->d_name, "runtime", 50000000);
+	}
+	closedir(dir);
+}
+
+void process_func(void)
+{
+	unsigned long long count = 0;
+	time_t end;
+
+	// Busy loop for RUN_TIME seconds
+	end = time(NULL) + RUN_TIME;
+	while (time(NULL) < end) {
+		count++; // Just a dummy operation
+	}
+}
+
+void set_affinity(int cpu_id)
+{
+	cpu_set_t cpuset;
+
+	CPU_ZERO(&cpuset);
+	CPU_SET(cpu_id, &cpuset);
+	CPU_SET(cpu_id + 1, &cpuset);
+
+	if (sched_setaffinity(0, sizeof(cpu_set_t), &cpuset) != 0) {
+		perror("sched_setaffinity");
+		exit(EXIT_FAILURE);
+	}
+}
+
+void set_sched(int policy, int priority)
+{
+	struct sched_param param;
+
+	param.sched_priority = priority;
+	if (sched_setscheduler(0, policy, &param) != 0) {
+		perror("sched_setscheduler");
+		exit(EXIT_FAILURE);
+	}
+}
+
+float get_process_runtime(int pid)
+{
+	char path[256];
+	FILE *file;
+	long utime, stime;
+	int fields;
+
+	snprintf(path, sizeof(path), "/proc/%d/stat", pid);
+	file = fopen(path, "r");
+	if (file == NULL) {
+		perror("Failed to open stat file");
+		return -1; // Indicate failure
+	}
+
+	// Skip the first 13 fields and read the 14th and 15th
+	fields = fscanf(file,
+					"%*d %*s %*c %*d %*d %*d %*d %*d %*u %*u %*u %*u %*u %lu %lu",
+					&utime, &stime);
+	fclose(file);
+
+	if (fields != 2) {
+		fprintf(stderr, "Failed to read stat file\n");
+		return -1; // Indicate failure
+	}
+
+	// Calculate the total time spent in the process
+	long total_time = utime + stime;
+	long ticks_per_second = sysconf(_SC_CLK_TCK);
+	float runtime_seconds = total_time * 1.0 / ticks_per_second;
+
+	return runtime_seconds;
+}
+
+int main(void)
+{
+	float runtime1, runtime2, runtime3;
+	int pid1, pid2, pid3;
+
+	if (!hyperthreading_enabled())
+		ksft_test_result_skip("This test requires hyperthreading to be enabled\n");
+
+	write_dl_server_params();
+
+	ksft_print_header();
+	ksft_set_plan(1);
+
+	// Create and set up a CFS task
+	pid1 = fork();
+	if (pid1 == 0) {
+		set_affinity(CORE_ID);
+		process_func();
+		exit(0);
+	} else if (pid1 < 0) {
+		perror("fork for p1");
+		ksft_exit_fail();
+	}
+
+	// Create a new unique cookie for the CFS task
+	if (prctl(PR_SCHED_CORE, PR_SCHED_CORE_CREATE, pid1, PIDTYPE_TGID, 0) < 0) {
+		perror("prctl for pid1");
+		ksft_exit_fail();
+	}
+
+	// Create a new unique cookie for the current process. Future
+	// forks will inherit this cookie.
+	if (prctl(PR_SCHED_CORE, PR_SCHED_CORE_CREATE, 0, PIDTYPE_TGID, 0) < 0) {
+		perror("prctl for current process");
+		ksft_exit_fail();
+	}
+
+	// Create an RT task which inherits the parent's cookie
+	pid2 = fork();
+	if (pid2 == 0) {
+		set_affinity(CORE_ID);
+		set_sched(SCHED_FIFO, 50);
+		process_func();
+		exit(0);
+	} else if (pid2 < 0) {
+		perror("fork for p2");
+		ksft_exit_fail();
+	}
+
+	// Create another RT task which inherits the parent's cookie
+	pid3 = fork();
+	if (pid3 == 0) {
+		set_affinity(CORE_ID);
+		set_sched(SCHED_FIFO, 50);
+		process_func();
+		exit(0);
+	} else if (pid3 < 0) {
+		perror("fork for p3");
+		ksft_exit_fail();
+	}
+
+	sleep(RUN_TIME * 3 / 4);
+	runtime1 = get_process_runtime(pid1);
+	if (runtime1 != -1)
+		ksft_print_msg("Runtime of PID %d is %f seconds\n", pid1, runtime1);
+	else
+		ksft_exit_fail_msg("Error getting runtime for PID %d\n", pid1);
+
+	runtime2 = get_process_runtime(pid2);
+	if (runtime2 != -1)
+		ksft_print_msg("Runtime of PID %d is %f seconds\n", pid2, runtime2);
+	else
+		ksft_exit_fail_msg("Error getting runtime for PID %d\n", pid2);
+
+	runtime3 = get_process_runtime(pid3);
+	if (runtime3 != -1)
+		ksft_print_msg("Runtime of PID %d is %f seconds\n", pid3, runtime3);
+	else
+		ksft_exit_fail_msg("Error getting runtime for PID %d\n", pid3);
+
+	// Make sure runtime1 is within 30% of runtime2
+	if (runtime1 < 0.7 * runtime2 || runtime1 > 1.3	* runtime2)
+		ksft_exit_fail_msg("Runtime of PID %d is not within 30%% of runtime of PID %d\n",
+						   pid1, pid2);
+
+	// Make	sure runtime1 is within 30% of runtime3
+	if (runtime1 < 0.7 * runtime3 || runtime1 > 1.3 * runtime3)
+		ksft_exit_fail_msg("Runtime of PID %d is not within 30%% of runtime of PID %d\n",
+						   pid1, pid3);
+
+	waitpid(pid1, NULL, 0);
+	waitpid(pid2, NULL, 0);
+	waitpid(pid3, NULL, 0);
+
+	ksft_test_result_pass("PASS\n");
+	return 0;
+}
-- 
2.34.1


  parent reply	other threads:[~2024-03-13  1:25 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-13  1:24 [PATCH v2 00/15] Fair scheduling deadline server fixes Joel Fernandes (Google)
2024-03-13  1:24 ` [PATCH v2 01/15] sched/core: Add clearing of ->dl_server in put_prev_task_balance() Joel Fernandes (Google)
2024-04-04 17:46   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` [PATCH v2 02/15] sched/core: Clear prev->dl_server in CFS pick fast path Joel Fernandes (Google)
2024-04-04 17:52   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` [PATCH v2 03/15] sched/core: Fix priority checking for DL server picks Joel Fernandes (Google)
2024-03-13  1:24 ` [PATCH v2 04/15] sched/core: Fix picking of tasks for core scheduling with DL server Joel Fernandes (Google)
2024-04-05  9:37   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` [PATCH v2 05/15] sched/debug: Use unsigned long for cpu variable to prevent cast errors Joel Fernandes (Google)
2024-03-13 20:02   ` Chris Hyser
2024-03-13  1:24 ` [PATCH v2 06/15] sched: server: Don't start hrtick for DL server tasks Joel Fernandes (Google)
2024-04-05  8:49   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` Joel Fernandes (Google) [this message]
2024-03-13  1:24 ` [PATCH v2 08/15] selftests/sched: Migrate cs_prctl_test to kselfttest Joel Fernandes (Google)
2024-03-13 18:44   ` Chris Hyser
2024-03-13  1:24 ` [PATCH v2 09/15] admin-guide/hw-vuln: Correct prctl() argument description Joel Fernandes (Google)
2024-03-13 19:14   ` Chris Hyser
2024-03-13 19:26     ` Chris Hyser
2024-04-05  9:32   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` [PATCH v2 10/15] sched: Fix build error in "sched/rt: Remove default bandwidth control" Joel Fernandes (Google)
2024-03-13 20:06   ` Chris Hyser
2024-03-13  1:24 ` [PATCH v2 11/15] sched/deadline: Mark DL server as unthrottled before enqueue Joel Fernandes (Google)
2024-04-05  8:54   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` [PATCH v2 12/15] sched/deadline: Reverse args to dl_time_before in replenish Joel Fernandes (Google)
2024-04-05  8:54   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` [PATCH v2 13/15] sched/deadline: Make start_dl_timer callers more robust Joel Fernandes (Google)
2024-04-05  9:10   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` [PATCH v2 14/15] sched/deadline: Do not restart the DL server on replenish from timer Joel Fernandes (Google)
2024-04-05  9:11   ` Daniel Bristot de Oliveira
2024-03-13  1:24 ` [PATCH v2 15/15] sched/deadline: Always start a new period if CFS exceeded DL runtime Joel Fernandes (Google)
2024-04-05  9:19   ` Daniel Bristot de Oliveira

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20240313012451.1693807-8-joel@joelfernandes.org \
    --to=joel@joelfernandes.org \
    --cc=bristot@redhat.com \
    --cc=bsegall@google.com \
    --cc=dietmar.eggemann@arm.com \
    --cc=joseph.salisbury@canonical.com \
    --cc=justinstitt@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=llvm@lists.linux.dev \
    --cc=luca.abeni@santannapisa.it \
    --cc=mgorman@suse.de \
    --cc=morbo@google.com \
    --cc=nathan@kernel.org \
    --cc=ndesaulniers@google.com \
    --cc=pauld@redhat.com \
    --cc=paulmck@kernel.org \
    --cc=rostedt@goodmis.org \
    --cc=shuah@kernel.org \
    --cc=skhan@linuxfoundation.org \
    --cc=suleiman@google.com \
    --cc=tglx@linutronix.de \
    --cc=tommaso.cucinotta@santannapisa.it \
    --cc=vineeth@bitbyteword.org \
    --cc=void@manifault.com \
    --cc=vschneid@redhat.com \
    --cc=youssefesmat@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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.