[PATCH 0/7] RCU torture-test updates for v6.18

[PATCH 0/7] RCU torture-test updates for v6.18

[Paul E. McKenney]

Hello!

This series contains torture-test updates for v6.18:

1.	Fix jitter.sh spin time.

2.	Add --do-normal parameter to torture.sh help text.

3.	Announce kernel boot status at torture-test startup.

4.	Suppress "Writer stall state" reports during boot.

5.	Delay rcutorture readers and writers until boot completes.

6.	Delay CPU-hotplug operations until boot completes.

7.	Delay forward-progress testing until boot completes.

						Thanx, Paul

------------------------------------------------------------------------

 b/kernel/rcu/rcutorture.c                           |    3 +-
 b/kernel/torture.c                                  |    5 ++-
 b/tools/testing/selftests/rcutorture/bin/jitter.sh  |   27 +++++++++++++++++---
 b/tools/testing/selftests/rcutorture/bin/torture.sh |    1 
 kernel/rcu/rcutorture.c                             |   20 +++++++++++---
 kernel/torture.c                                    |    2 +
 6 files changed, 48 insertions(+), 10 deletions(-)

[PATCH 1/7] rcutorture: Fix jitter.sh spin time

[Paul E. McKenney]

An embarrassing syntax error in jitter.sh makes for fixed spin time.
This commit therefore makes it be variable, as intended, albeit with
very coarse-grained adjustment.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
---
 .../selftests/rcutorture/bin/jitter.sh        | 27 ++++++++++++++++---
 1 file changed, 24 insertions(+), 3 deletions(-)

diff --git a/tools/testing/selftests/rcutorture/bin/jitter.sh b/tools/testing/selftests/rcutorture/bin/jitter.sh
index fd1ffaa5a1358e..3c1e5d3f88058c 100755
--- a/tools/testing/selftests/rcutorture/bin/jitter.sh
+++ b/tools/testing/selftests/rcutorture/bin/jitter.sh
@@ -39,6 +39,22 @@ do
 	fi
 done
 
+# Uses global variables startsecs, startns, endsecs, endns, and limit.
+# Exit code is success for time not yet elapsed and failure otherwise.
+function timecheck {
+	local done=`awk -v limit=$limit \
+			-v startsecs=$startsecs \
+			-v startns=$startns \
+			-v endsecs=$endsecs \
+			-v endns=$endns < /dev/null '
+		BEGIN {
+			delta = (endsecs - startsecs) * 1000 * 1000;
+			delta += int((endns - startns) / 1000);
+			print delta >= limit;
+		}'`
+	return $done
+}
+
 while :
 do
 	# Check for done.
@@ -85,15 +101,20 @@ do
 	n=$(($n+1))
 	sleep .$sleeptime
 
-	# Spin a random duration
+	# Spin a random duration, but with rather coarse granularity.
 	limit=`awk -v me=$me -v n=$n -v spinmax=$spinmax 'BEGIN {
 		srand(n + me + systime());
 		printf("%06d", int(rand() * spinmax));
 	}' < /dev/null`
 	n=$(($n+1))
-	for i in {1..$limit}
+	startsecs=`date +%s`
+	startns=`date +%N`
+	endsecs=$startns
+	endns=$endns
+	while timecheck
 	do
-		echo > /dev/null
+		endsecs=`date +%s`
+		endns=`date +%N`
 	done
 done
 
-- 
2.40.1

[PATCH 2/7] torture: Add --do-normal parameter to torture.sh help text

[Paul E. McKenney]

The --do-normal parameter was missing from the torture.sh script's help
text, so this commit adds it.  Hopefully better late than never!

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
---
 tools/testing/selftests/rcutorture/bin/torture.sh | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tools/testing/selftests/rcutorture/bin/torture.sh b/tools/testing/selftests/rcutorture/bin/torture.sh
index 611bc03a8dc705..a33ba109ef0b7f 100755
--- a/tools/testing/selftests/rcutorture/bin/torture.sh
+++ b/tools/testing/selftests/rcutorture/bin/torture.sh
@@ -94,6 +94,7 @@ usage () {
 	echo "       --do-kvfree / --do-no-kvfree / --no-kvfree"
 	echo "       --do-locktorture / --do-no-locktorture / --no-locktorture"
 	echo "       --do-none"
+	echo "       --do-normal / --do-no-normal / --no-normal"
 	echo "       --do-rcuscale / --do-no-rcuscale / --no-rcuscale"
 	echo "       --do-rcutasksflavors / --do-no-rcutasksflavors / --no-rcutasksflavors"
 	echo "       --do-rcutorture / --do-no-rcutorture / --no-rcutorture"
-- 
2.40.1

[PATCH 3/7] torture: Announce kernel boot status at torture-test startup

[Paul E. McKenney]

Sometimes a given system takes surprisingly long to boot, for example,
in one recent case, 70 seconds instead of three seconds.  It would
be good to fix these slow-boot issues, but it would also be good for
the torture tests to announce that the system was still booting at the
start of the test.  Especially for tests that have a greater probability
of false positives when run in the single-CPU boot-time environment.
Yes, those tests should defend themselves, but we should also make this
situation easier to diagnose.

This commit therefore causes torture_print_module_parms() to print
"still booting" at the end of its printk() that dumps out the values of
its module parameters.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
---
 kernel/torture.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/kernel/torture.c b/kernel/torture.c
index 3a0a8cc604010a..5abb4b25d971cc 100644
--- a/kernel/torture.c
+++ b/kernel/torture.c
@@ -797,8 +797,9 @@ static unsigned long torture_init_jiffies;
 static void
 torture_print_module_parms(void)
 {
-	pr_alert("torture module --- %s:  disable_onoff_at_boot=%d ftrace_dump_at_shutdown=%d verbose_sleep_frequency=%d verbose_sleep_duration=%d random_shuffle=%d\n",
-		 torture_type, disable_onoff_at_boot, ftrace_dump_at_shutdown, verbose_sleep_frequency, verbose_sleep_duration, random_shuffle);
+	pr_alert("torture module --- %s:  disable_onoff_at_boot=%d ftrace_dump_at_shutdown=%d verbose_sleep_frequency=%d verbose_sleep_duration=%d random_shuffle=%d%s\n",
+		 torture_type, disable_onoff_at_boot, ftrace_dump_at_shutdown, verbose_sleep_frequency, verbose_sleep_duration, random_shuffle,
+		 rcu_inkernel_boot_has_ended() ? "" : " still booting");
 }
 
 /*
-- 
2.40.1

[PATCH 4/7] rcutorture: Suppress "Writer stall state" reports during boot

[Paul E. McKenney]

When rcutorture is running on only the one boot-time CPU while that CPU
is busy invoking initcall() functions, the added load is quite likely to
unduly delay the RCU grace-period kthread, rcutorture readers, and much
else besides.  This can result in rcu_torture_stats_print() reporting
rcutorture writer stalls, which are not really a bug in that environment.
After all, one CPU can only do so much.

This commit therefore suppresses rcutorture writer stalls while the
kernel is booting, that is, while rcu_inkernel_boot_has_ended() continues
returning false.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
---
 kernel/rcu/rcutorture.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c
index 7a893d51d02b6a..49e048da4f6810 100644
--- a/kernel/rcu/rcutorture.c
+++ b/kernel/rcu/rcutorture.c
@@ -2756,7 +2756,8 @@ rcu_torture_stats_print(void)
 		cur_ops->stats();
 	if (rtcv_snap == rcu_torture_current_version &&
 	    rcu_access_pointer(rcu_torture_current) &&
-	    !rcu_stall_is_suppressed()) {
+	    !rcu_stall_is_suppressed() &&
+	    rcu_inkernel_boot_has_ended()) {
 		int __maybe_unused flags = 0;
 		unsigned long __maybe_unused gp_seq = 0;
 
-- 
2.40.1

[PATCH 5/7] rcutorture: Delay rcutorture readers and writers until boot completes

[Paul E. McKenney]

The rcutorture writers and (especially) readers are the biggest CPU
hogs of the bunch, so this commit therefore makes them wait until boot
has completed.

This makes the current setting of the boot_ended local variable dead code,
so while in the area, this commit removes that as well.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
---
 kernel/rcu/rcutorture.c | 18 ++++++++++++++----
 1 file changed, 14 insertions(+), 4 deletions(-)

diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c
index 49e048da4f6810..1578d330565752 100644
--- a/kernel/rcu/rcutorture.c
+++ b/kernel/rcu/rcutorture.c
@@ -1528,7 +1528,7 @@ static void do_rtws_sync(struct torture_random_state *trsp, void (*sync)(void))
 static int
 rcu_torture_writer(void *arg)
 {
-	bool boot_ended;
+	bool booting_still = false;
 	bool can_expedite = !rcu_gp_is_expedited() && !rcu_gp_is_normal();
 	unsigned long cookie;
 	struct rcu_gp_oldstate cookie_full;
@@ -1539,6 +1539,7 @@ rcu_torture_writer(void *arg)
 	struct rcu_gp_oldstate gp_snap1_full;
 	int i;
 	int idx;
+	unsigned long j;
 	int oldnice = task_nice(current);
 	struct rcu_gp_oldstate *rgo = NULL;
 	int rgo_size = 0;
@@ -1581,6 +1582,16 @@ rcu_torture_writer(void *arg)
 			rgo_size = cur_ops->poll_active_full;
 	}
 
+	// If the system is still booting, let it finish.
+	j = jiffies;
+	while (!torture_must_stop() && !rcu_inkernel_boot_has_ended()) {
+		booting_still = true;
+		schedule_timeout_interruptible(HZ);
+	}
+	if (booting_still)
+		pr_alert("%s" TORTURE_FLAG " Waited %lu jiffies for boot to complete.\n",
+			 torture_type, jiffies - j);
+
 	do {
 		rcu_torture_writer_state = RTWS_FIXED_DELAY;
 		torture_hrtimeout_us(500, 1000, &rand);
@@ -1769,13 +1780,11 @@ rcu_torture_writer(void *arg)
 				       !rcu_gp_is_normal();
 		}
 		rcu_torture_writer_state = RTWS_STUTTER;
-		boot_ended = rcu_inkernel_boot_has_ended();
 		stutter_waited = stutter_wait("rcu_torture_writer");
 		if (stutter_waited &&
 		    !atomic_read(&rcu_fwd_cb_nodelay) &&
 		    !cur_ops->slow_gps &&
 		    !torture_must_stop() &&
-		    boot_ended &&
 		    time_after(jiffies, stallsdone))
 			for (i = 0; i < ARRAY_SIZE(rcu_tortures); i++)
 				if (list_empty(&rcu_tortures[i].rtort_free) &&
@@ -2437,7 +2446,8 @@ rcu_torture_reader(void *arg)
 			torture_hrtimeout_us(500, 1000, &rand);
 			lastsleep = jiffies + 10;
 		}
-		while (torture_num_online_cpus() < mynumonline && !torture_must_stop())
+		while (!torture_must_stop() &&
+		       (torture_num_online_cpus() < mynumonline || !rcu_inkernel_boot_has_ended()))
 			schedule_timeout_interruptible(HZ / 5);
 		stutter_wait("rcu_torture_reader");
 	} while (!torture_must_stop());
-- 
2.40.1

[PATCH 6/7] torture: Delay CPU-hotplug operations until boot completes

[Paul E. McKenney]

CPU-hotplug operations invoke stop-machine, which can hog CPUs, which is
not a great thing to do before boot has completed.  This commit therefore
makes the CPU-hotplug operations hold off until boot has completed.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
---
 kernel/torture.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/torture.c b/kernel/torture.c
index 5abb4b25d971cc..1ea9f67953a766 100644
--- a/kernel/torture.c
+++ b/kernel/torture.c
@@ -359,6 +359,8 @@ torture_onoff(void *arg)
 		torture_hrtimeout_jiffies(onoff_holdoff, &rand);
 		VERBOSE_TOROUT_STRING("torture_onoff end holdoff");
 	}
+	while (!rcu_inkernel_boot_has_ended())
+		schedule_timeout_interruptible(HZ / 10);
 	while (!torture_must_stop()) {
 		if (disable_onoff_at_boot && !rcu_inkernel_boot_has_ended()) {
 			torture_hrtimeout_jiffies(HZ / 10, &rand);
-- 
2.40.1

[PATCH 7/7] rcutorture: Delay forward-progress testing until boot completes

[Paul E. McKenney]

Forward-progress testing can hog CPUs, which is not a great thing to do
before boot has completed.  This commit therefore makes the CPU-hotplug
operations hold off until boot has completed.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
---
 kernel/rcu/rcutorture.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c
index 1578d330565752..b8a684459381f2 100644
--- a/kernel/rcu/rcutorture.c
+++ b/kernel/rcu/rcutorture.c
@@ -3457,6 +3457,8 @@ static int rcu_torture_fwd_prog(void *args)
 	int tested_tries = 0;
 
 	VERBOSE_TOROUT_STRING("rcu_torture_fwd_progress task started");
+	while (!rcu_inkernel_boot_has_ended())
+		schedule_timeout_interruptible(HZ / 10);
 	rcu_bind_current_to_nocb();
 	if (!IS_ENABLED(CONFIG_SMP) || !IS_ENABLED(CONFIG_RCU_BOOST))
 		set_user_nice(current, MAX_NICE);
-- 
2.40.1