Skip to contentSkip to conversation

rcutorture: Allow a negative value for nfakewriters

PATCH Superseded Uladzislau Rezki (Sony) 4 participants 15 messages Feb 25, 2025 – Feb 26, 2025

Subsystems (4)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Uladzislau Rezki (Sony)

author maintainer
#0 lore raw

PATCH v3 1/3 rcutorture: Allow a negative value for nfakewriters

Currently "nfakewriters" parameter can be set to any value but there is no possibility to adjust it automatically based on how many CPUs a system has where a test is run on.

To address this, if the "nfakewriters" is set to negative it will be adjusted to num_online_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
kernel/rcu/rcutorture.c +16 -6
index d98b3bd6d91f..f376262532ce 100644
@@ -148,6 +148,7 @@ MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");
 
 static int nrealnocbers;
 static int nrealreaders;
+static int nrealfakewriters;
 static struct task_struct *writer_task;
 static struct task_struct **fakewriter_tasks;
 static struct task_struct **reader_tasks;
@@ -1763,7 +1764,7 @@ rcu_torture_fakewriter(void *arg)
 	do {
 		torture_hrtimeout_jiffies(torture_random(&rand) % 10, &rand);
 		if (cur_ops->cb_barrier != NULL &&
-		    torture_random(&rand) % (nfakewriters * 8) == 0) {
+		    torture_random(&rand) % (nrealfakewriters * 8) == 0) {
 			cur_ops->cb_barrier();
 		} else {
 			switch (synctype[torture_random(&rand) % nsynctypes]) {
@@ -2568,7 +2569,7 @@ rcu_torture_print_module_parms(struct rcu_torture_ops *cur_ops, const char *tag)
 		 "nocbs_nthreads=%d nocbs_toggle=%d "
 		 "test_nmis=%d "
 		 "preempt_duration=%d preempt_interval=%d\n",
-		 torture_type, tag, nrealreaders, nfakewriters,
+		 torture_type, tag, nrealreaders, nrealfakewriters,
 		 stat_interval, verbose, test_no_idle_hz, shuffle_interval,
 		 stutter, irqreader, fqs_duration, fqs_holdoff, fqs_stutter,
 		 test_boost, cur_ops->can_boost,
@@ -3644,7 +3645,7 @@ rcu_torture_cleanup(void)
 	rcu_torture_reader_mbchk = NULL;
 
 	if (fakewriter_tasks) {
-		for (i = 0; i < nfakewriters; i++)
+		for (i = 0; i < nrealfakewriters; i++)
 			torture_stop_kthread(rcu_torture_fakewriter,
 					     fakewriter_tasks[i]);
 		kfree(fakewriter_tasks);
@@ -4066,6 +4067,14 @@ rcu_torture_init(void)
 
 	rcu_torture_init_srcu_lockdep();
 
+	if (nfakewriters >= 0) {
+		nrealfakewriters = nfakewriters;
+	} else {
+		nrealfakewriters = num_online_cpus() - 2 - nfakewriters;
+		if (nrealfakewriters <= 0)
+			nrealfakewriters = 1;
+	}
+
 	if (nreaders >= 0) {
 		nrealreaders = nreaders;
 	} else {
@@ -4122,8 +4131,9 @@ rcu_torture_init(void)
 					  writer_task);
 	if (torture_init_error(firsterr))
 		goto unwind;
-	if (nfakewriters > 0) {
-		fakewriter_tasks = kcalloc(nfakewriters,
+
+	if (nrealfakewriters > 0) {
+		fakewriter_tasks = kcalloc(nrealfakewriters,
 					   sizeof(fakewriter_tasks[0]),
 					   GFP_KERNEL);
 		if (fakewriter_tasks == NULL) {
@@ -4132,7 +4142,7 @@ rcu_torture_init(void)
 			goto unwind;
 		}
 	}
-	for (i = 0; i < nfakewriters; i++) {
+	for (i = 0; i < nrealfakewriters; i++) {
 		firsterr = torture_create_kthread(rcu_torture_fakewriter,
 						  NULL, fakewriter_tasks[i]);
 		if (torture_init_error(firsterr))
Signature · 1 line
2.39.5
author maintainer
#1 lore raw

PATCH v3 2/3 rcu: Update TREE05.boot to test normal synchronize_rcu()

Add extra parameters for rcutorture module. One is the "nfakewriters" which is set -1. There will be created number of test-kthreads which correspond to number of CPUs in a test system. Those threads randomly invoke synchronize_rcu() call.

Apart of that "rcu_normal" is set to 1, because it is specifically for a normal synchronize_rcu() testing, also a newly added parameter which is "rcu_normal_wake_from_gp" is set to 1 also. That prevents interaction with other callbacks in a system.

---
 tools/testing/selftests/rcutorture/configs/rcu/TREE05.boot | 6 ++++++
 1 file changed, 6 insertions(+)
tools/testing/selftests/rcutorture/configs/rcu/TREE05.boot +6 -0
index c419cac233ee..54f5c9053474 100644
@@ -2,3 +2,9 @@ rcutree.gp_preinit_delay=3
 rcutree.gp_init_delay=3
 rcutree.gp_cleanup_delay=3
 rcupdate.rcu_self_test=1
+
+# This part is for synchronize_rcu() testing
+rcutorture.nfakewriters=-1
+rcutorture.gp_sync=1
+rcupdate.rcu_normal=1
+rcutree.rcu_normal_wake_from_gp=1
Signature · 1 line
2.39.5
author maintainer
#2 lore raw

PATCH v3 3/3 rcu: Use _full() API to debug synchronize_rcu()

In reply to Uladzislau Rezki (Sony)

Switch for using of get_state_synchronize_rcu_full() and poll_state_synchronize_rcu_full() pair for debug a normal synchronize_rcu() call.

Just using "not" full APIs to identify if a grace period is passed or not might lead to a false kernel splat.

---
 include/linux/rcupdate_wait.h | 3 +++
 kernel/rcu/tree.c             | 8 +++-----
 2 files changed, 6 insertions(+), 5 deletions(-)
include/linux/rcupdate_wait.h +3 -0
index f9bed3d3f78d..4c92d4291cce 100644
@@ -16,6 +16,9 @@
 struct rcu_synchronize {
 	struct rcu_head head;
 	struct completion completion;
+
+	/* This is for debugging. */
+	struct rcu_gp_oldstate oldstate;
 };
 void wakeme_after_rcu(struct rcu_head *head);
 
kernel/rcu/tree.c +3 -5
index 8625f616c65a..48384fa2eaeb 100644
@@ -1632,12 +1632,10 @@ static void rcu_sr_normal_complete(struct llist_node *node)
 {
 	struct rcu_synchronize *rs = container_of(
 		(struct rcu_head *) node, struct rcu_synchronize, head);
-	unsigned long oldstate = (unsigned long) rs->head.func;
 
 	WARN_ONCE(IS_ENABLED(CONFIG_PROVE_RCU) &&
-		!poll_state_synchronize_rcu(oldstate),
-		"A full grace period is not passed yet: %lu",
-		rcu_seq_diff(get_state_synchronize_rcu(), oldstate));
+		!poll_state_synchronize_rcu_full(&rs->oldstate),
+		"A full grace period is not passed yet!\n");
 
 	/* Finally. */
 	complete(&rs->completion);
@@ -3247,7 +3245,7 @@ static void synchronize_rcu_normal(void)
 	 * snapshot before adding a request.
 	 */
 	if (IS_ENABLED(CONFIG_PROVE_RCU))
-		rs.head.func = (void *) get_state_synchronize_rcu();
+		get_state_synchronize_rcu_full(&rs.oldstate);
 
 	rcu_sr_normal_add_req(&rs);
 
Signature · 1 line
2.39.5
maintainer Reviewed-by on v3 3/3
#3 lore raw

Re: [PATCH v3 3/3] rcu: Use _full() API to debug synchronize_rcu()

Quoted 11 lines · Uladzislau Rezki (Sony)

Switch for using of get_state_synchronize_rcu_full() and poll_state_synchronize_rcu_full() pair for debug a normal synchronize_rcu() call.

Just using "not" full APIs to identify if a grace period is passed or not might lead to a false kernel splat.

This one passes light testing, so:

Quoted 50 lines
---
 include/linux/rcupdate_wait.h | 3 +++
 kernel/rcu/tree.c             | 8 +++-----
 2 files changed, 6 insertions(+), 5 deletions(-)
include/linux/rcupdate_wait.h +3 -0
index f9bed3d3f78d..4c92d4291cce 100644
@@ -16,6 +16,9 @@
 struct rcu_synchronize {
 	struct rcu_head head;
 	struct completion completion;
+
+	/* This is for debugging. */
+	struct rcu_gp_oldstate oldstate;
 };
 void wakeme_after_rcu(struct rcu_head *head);
 
kernel/rcu/tree.c +3 -5
index 8625f616c65a..48384fa2eaeb 100644
@@ -1632,12 +1632,10 @@ static void rcu_sr_normal_complete(struct llist_node *node)
 {
 	struct rcu_synchronize *rs = container_of(
 		(struct rcu_head *) node, struct rcu_synchronize, head);
-	unsigned long oldstate = (unsigned long) rs->head.func;
 
 	WARN_ONCE(IS_ENABLED(CONFIG_PROVE_RCU) &&
-		!poll_state_synchronize_rcu(oldstate),
-		"A full grace period is not passed yet: %lu",
-		rcu_seq_diff(get_state_synchronize_rcu(), oldstate));
+		!poll_state_synchronize_rcu_full(&rs->oldstate),
+		"A full grace period is not passed yet!\n");
 
 	/* Finally. */
 	complete(&rs->completion);
@@ -3247,7 +3245,7 @@ static void synchronize_rcu_normal(void)
 	 * snapshot before adding a request.
 	 */
 	if (IS_ENABLED(CONFIG_PROVE_RCU))
-		rs.head.func = (void *) get_state_synchronize_rcu();
+		get_state_synchronize_rcu_full(&rs.oldstate);
 
 	rcu_sr_normal_add_req(&rs);
 
Signature · 1 line
2.39.5
author maintainer
#4 lore raw

Re: [PATCH v3 3/3] rcu: Use _full() API to debug synchronize_rcu()

Quoted 8 lines · Paul E. McKenney
Quoted 7 lines · Uladzislau Rezki (Sony)

Switch for using of get_state_synchronize_rcu_full() and poll_state_synchronize_rcu_full() pair for debug a normal synchronize_rcu() call.

Just using "not" full APIs to identify if a grace period is passed or not might lead to a false kernel splat.

Quoted 9 lines
Quoted 4 lines

This one passes light testing, so:

Thank you!

--
Uladzislau Rezki

#5 lore raw

Re: [PATCH v3 3/3] rcu: Use _full() API to debug synchronize_rcu()

In reply to Uladzislau Rezki (Sony)

Hi Ulad,

Quoted 7 lines · Uladzislau Rezki (Sony)

Switch for using of get_state_synchronize_rcu_full() and poll_state_synchronize_rcu_full() pair for debug a normal synchronize_rcu() call.

Just using "not" full APIs to identify if a grace period is passed or not might lead to a false kernel splat.

Could you provide detailed explanation on this? I.e. why is _full() is needed? I find the current commit message is a bit vague.

Regards,
Boqun

Quoted 54 lines
---
 include/linux/rcupdate_wait.h | 3 +++
 kernel/rcu/tree.c             | 8 +++-----
 2 files changed, 6 insertions(+), 5 deletions(-)
include/linux/rcupdate_wait.h +3 -0
index f9bed3d3f78d..4c92d4291cce 100644
@@ -16,6 +16,9 @@
 struct rcu_synchronize {
 	struct rcu_head head;
 	struct completion completion;
+
+	/* This is for debugging. */
+	struct rcu_gp_oldstate oldstate;
 };
 void wakeme_after_rcu(struct rcu_head *head);
 
kernel/rcu/tree.c +3 -5
index 8625f616c65a..48384fa2eaeb 100644
@@ -1632,12 +1632,10 @@ static void rcu_sr_normal_complete(struct llist_node *node)
 {
 	struct rcu_synchronize *rs = container_of(
 		(struct rcu_head *) node, struct rcu_synchronize, head);
-	unsigned long oldstate = (unsigned long) rs->head.func;
 
 	WARN_ONCE(IS_ENABLED(CONFIG_PROVE_RCU) &&
-		!poll_state_synchronize_rcu(oldstate),
-		"A full grace period is not passed yet: %lu",
-		rcu_seq_diff(get_state_synchronize_rcu(), oldstate));
+		!poll_state_synchronize_rcu_full(&rs->oldstate),
+		"A full grace period is not passed yet!\n");
 
 	/* Finally. */
 	complete(&rs->completion);
@@ -3247,7 +3245,7 @@ static void synchronize_rcu_normal(void)
 	 * snapshot before adding a request.
 	 */
 	if (IS_ENABLED(CONFIG_PROVE_RCU))
-		rs.head.func = (void *) get_state_synchronize_rcu();
+		get_state_synchronize_rcu_full(&rs.oldstate);
 
 	rcu_sr_normal_add_req(&rs);
 
Signature · 1 line
2.39.5
author maintainer
#6 lore raw

Re: [PATCH v3 3/3] rcu: Use _full() API to debug synchronize_rcu()

Hello, Boqun.

Quoted 7 lines

Hi Ulad,

Quoted 4 lines · Uladzislau Rezki (Sony)

Switch for using of get_state_synchronize_rcu_full() and poll_state_synchronize_rcu_full() pair for debug a normal synchronize_rcu() call.

Quoted 7 lines
Quoted 3 lines

Just using "not" full APIs to identify if a grace period is passed or not might lead to a false kernel splat.

Could you provide detailed explanation on this? I.e. why is _full() is needed? I find the current commit message is a bit vague.

<snip>
rcu: Use _full() API to debug synchronize_rcu()

Switch for using of get_state_synchronize_rcu_full() and poll_state_synchronize_rcu_full() pair to debug a normal synchronize_rcu() call.

Just using "not" full APIs to identify if a grace period is passed or not might lead to a false-positive kernel splat.

It can happen, because get_state_synchronize_rcu() compresses both normal and expedited states into one single unsigned long value, so a poll_state_synchronize_rcu() can miss GP-completion when synchronize_rcu()/synchronize_rcu_expedited() concurrently run.

To address this, switch to poll_state_synchronize_rcu_full() and get_state_synchronize_rcu_full() APIs, which use separate variables for expedited and normal states.
<snip>

Does it look better?

--
Uladzislau Rezki

#7 lore raw

Re: [PATCH v3 3/3] rcu: Use _full() API to debug synchronize_rcu()

Quoted 32 lines · Uladzislau Rezki

Hello, Boqun.

Quoted 7 lines

Hi Ulad,

Quoted 4 lines · Uladzislau Rezki (Sony)

Switch for using of get_state_synchronize_rcu_full() and poll_state_synchronize_rcu_full() pair for debug a normal synchronize_rcu() call.

Quoted 7 lines
Quoted 3 lines

Just using "not" full APIs to identify if a grace period is passed or not might lead to a false kernel splat.

Could you provide detailed explanation on this? I.e. why is _full() is needed? I find the current commit message is a bit vague.

<snip>
rcu: Use _full() API to debug synchronize_rcu()

Switch for using of get_state_synchronize_rcu_full() and poll_state_synchronize_rcu_full() pair to debug a normal synchronize_rcu() call.

Just using "not" full APIs to identify if a grace period is passed or not might lead to a false-positive kernel splat.

It can happen, because get_state_synchronize_rcu() compresses both normal and expedited states into one single unsigned long value, so a poll_state_synchronize_rcu() can miss GP-completion when synchronize_rcu()/synchronize_rcu_expedited() concurrently run.

Quoted 7 lines

To address this, switch to poll_state_synchronize_rcu_full() and get_state_synchronize_rcu_full() APIs, which use separate variables for expedited and normal states.
<snip>

Does it look better?

Yes, that looks good to me. Thanks!

Regards,
Boqun

Quoted 3 lines

--
Uladzislau Rezki

maintainer
#8 lore raw
In reply to Uladzislau Rezki (Sony)

kernel/rcu/rcutorture.c MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

Quoted 18 lines · Uladzislau Rezki (Sony)

Currently "nfakewriters" parameter can be set to any value but there is no possibility to adjust it automatically based on how many CPUs a system has where a test is run on.

To address this, if the "nfakewriters" is set to negative it will be adjusted to num_online_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
kernel/rcu/rcutorture.c +0 -0
index d98b3bd6d91f..f376262532ce 100644
@@ -148,6 +148,7 @@ MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

IMO, this should also be updated to reflect the possibily to set it negative and hence to number CPUs:

torture_param(int, nfakewriters, 4, "Number of RCU fake writer threads");

thanks,

- Joel

Quoted 73 lines
kernel/rcu/rcutorture.c +16 -6
 static int nrealnocbers;
 static int nrealreaders;
+static int nrealfakewriters;
 static struct task_struct *writer_task;
 static struct task_struct **fakewriter_tasks;
 static struct task_struct **reader_tasks;
@@ -1763,7 +1764,7 @@ rcu_torture_fakewriter(void *arg)
 	do {
 		torture_hrtimeout_jiffies(torture_random(&rand) % 10, &rand);
 		if (cur_ops->cb_barrier != NULL &&
-		    torture_random(&rand) % (nfakewriters * 8) == 0) {
+		    torture_random(&rand) % (nrealfakewriters * 8) == 0) {
 			cur_ops->cb_barrier();
 		} else {
 			switch (synctype[torture_random(&rand) % nsynctypes]) {
@@ -2568,7 +2569,7 @@ rcu_torture_print_module_parms(struct rcu_torture_ops *cur_ops, const char *tag)
 		 "nocbs_nthreads=%d nocbs_toggle=%d "
 		 "test_nmis=%d "
 		 "preempt_duration=%d preempt_interval=%d\n",
-		 torture_type, tag, nrealreaders, nfakewriters,
+		 torture_type, tag, nrealreaders, nrealfakewriters,
 		 stat_interval, verbose, test_no_idle_hz, shuffle_interval,
 		 stutter, irqreader, fqs_duration, fqs_holdoff, fqs_stutter,
 		 test_boost, cur_ops->can_boost,
@@ -3644,7 +3645,7 @@ rcu_torture_cleanup(void)
 	rcu_torture_reader_mbchk = NULL;
 
 	if (fakewriter_tasks) {
-		for (i = 0; i < nfakewriters; i++)
+		for (i = 0; i < nrealfakewriters; i++)
 			torture_stop_kthread(rcu_torture_fakewriter,
 					     fakewriter_tasks[i]);
 		kfree(fakewriter_tasks);
@@ -4066,6 +4067,14 @@ rcu_torture_init(void)
 
 	rcu_torture_init_srcu_lockdep();
 
+	if (nfakewriters >= 0) {
+		nrealfakewriters = nfakewriters;
+	} else {
+		nrealfakewriters = num_online_cpus() - 2 - nfakewriters;
+		if (nrealfakewriters <= 0)
+			nrealfakewriters = 1;
+	}
+
 	if (nreaders >= 0) {
 		nrealreaders = nreaders;
 	} else {
@@ -4122,8 +4131,9 @@ rcu_torture_init(void)
 					  writer_task);
 	if (torture_init_error(firsterr))
 		goto unwind;
-	if (nfakewriters > 0) {
-		fakewriter_tasks = kcalloc(nfakewriters,
+
+	if (nrealfakewriters > 0) {
+		fakewriter_tasks = kcalloc(nrealfakewriters,
 					   sizeof(fakewriter_tasks[0]),
 					   GFP_KERNEL);
 		if (fakewriter_tasks == NULL) {
@@ -4132,7 +4142,7 @@ rcu_torture_init(void)
 			goto unwind;
 		}
 	}
-	for (i = 0; i < nfakewriters; i++) {
+	for (i = 0; i < nrealfakewriters; i++) {
 		firsterr = torture_create_kthread(rcu_torture_fakewriter,
 						  NULL, fakewriter_tasks[i]);
 		if (torture_init_error(firsterr))
Signature · 1 line
2.39.5
author maintainer
#9 lore raw
Quoted 23 lines · Joel Fernandes

kernel/rcu/rcutorture.c MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

Quoted 18 lines · Uladzislau Rezki (Sony)

Currently "nfakewriters" parameter can be set to any value but there is no possibility to adjust it automatically based on how many CPUs a system has where a test is run on.

To address this, if the "nfakewriters" is set to negative it will be adjusted to num_online_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
kernel/rcu/rcutorture.c +0 -0
index d98b3bd6d91f..f376262532ce 100644
@@ -148,6 +148,7 @@ MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

IMO, this should also be updated to reflect the possibily to set it negative and hence to number CPUs:

Quoted 2 lines

torture_param(int, nfakewriters, 4, "Number of RCU fake writer threads");

You can set it to a negative as well as to number of CPUs or any other number.

--
Uladzislau Rezki

maintainer
#10 lore raw
Quoted 24 lines · Uladzislau Rezki
Quoted 23 lines · Joel Fernandes

kernel/rcu/rcutorture.c MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

Quoted 18 lines · Uladzislau Rezki (Sony)

Currently "nfakewriters" parameter can be set to any value but there is no possibility to adjust it automatically based on how many CPUs a system has where a test is run on.

To address this, if the "nfakewriters" is set to negative it will be adjusted to num_online_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
kernel/rcu/rcutorture.c +0 -0
index d98b3bd6d91f..f376262532ce 100644
@@ -148,6 +148,7 @@ MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

IMO, this should also be updated to reflect the possibily to set it negative and hence to number CPUs:

Quoted 4 lines
Quoted 2 lines

torture_param(int, nfakewriters, 4, "Number of RCU fake writer threads");

You can set it to a negative as well as to number of CPUs or any other number.

Joel does have a good point, though. Could I interest one of you in updating the Documentation/admin-guide/kernel-parameters.txt entry for rcutorture.nfakewriters? The rcutorture.nreaders entry is a decent guide.

Thanx, Paul

author maintainer
#11 lore raw
Quoted 25 lines · Paul E. McKenney
Quoted 24 lines · Uladzislau Rezki
Quoted 23 lines · Joel Fernandes

kernel/rcu/rcutorture.c MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

Quoted 18 lines · Uladzislau Rezki (Sony)

Currently "nfakewriters" parameter can be set to any value but there is no possibility to adjust it automatically based on how many CPUs a system has where a test is run on.

To address this, if the "nfakewriters" is set to negative it will be adjusted to num_online_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
kernel/rcu/rcutorture.c +0 -0
index d98b3bd6d91f..f376262532ce 100644
@@ -148,6 +148,7 @@ MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

IMO, this should also be updated to reflect the possibily to set it negative and hence to number CPUs:

Quoted 9 lines
Quoted 4 lines
Quoted 2 lines

torture_param(int, nfakewriters, 4, "Number of RCU fake writer threads");

You can set it to a negative as well as to number of CPUs or any other number.

Joel does have a good point, though. Could I interest one of you in updating the Documentation/admin-guide/kernel-parameters.txt entry for rcutorture.nfakewriters? The rcutorture.nreaders entry is a decent guide.

OK. Then i misunderstood Joel. It was about updating a documentation about this!

Good point :)

--
Uladzislau Rezki

maintainer
#12 lore raw
In reply to Uladzislau Rezki
Quoted 24 lines · Uladzislau Rezki
Quoted 23 lines · Joel Fernandes

kernel/rcu/rcutorture.c MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

Quoted 18 lines · Uladzislau Rezki (Sony)

Currently "nfakewriters" parameter can be set to any value but there is no possibility to adjust it automatically based on how many CPUs a system has where a test is run on.

To address this, if the "nfakewriters" is set to negative it will be adjusted to num_online_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
kernel/rcu/rcutorture.c +0 -0
index d98b3bd6d91f..f376262532ce 100644
@@ -148,6 +148,7 @@ MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

IMO, this should also be updated to reflect the possibily to set it negative and hence to number CPUs:

Quoted 4 lines
Quoted 2 lines

torture_param(int, nfakewriters, 4, "Number of RCU fake writer threads");

You can set it to a negative as well as to number of CPUs or any other number.

Sorry I just meant amend the description to something like "Number of RCU fake writer threads (or set to -1 for NR_CPUs)", so user does not have to read code to know that (and update the kernel cmdline params document as well).

thanks,

- Joel

maintainer
#13 lore raw
Quoted 27 lines · Joel Fernandes
Quoted 24 lines · Uladzislau Rezki
Quoted 23 lines · Joel Fernandes

kernel/rcu/rcutorture.c MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

Quoted 18 lines · Uladzislau Rezki (Sony)

Currently "nfakewriters" parameter can be set to any value but there is no possibility to adjust it automatically based on how many CPUs a system has where a test is run on.

To address this, if the "nfakewriters" is set to negative it will be adjusted to num_online_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
kernel/rcu/rcutorture.c +0 -0
index d98b3bd6d91f..f376262532ce 100644
@@ -148,6 +148,7 @@ MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

IMO, this should also be updated to reflect the possibily to set it negative and hence to number CPUs:

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

torture_param(int, nfakewriters, 4, "Number of RCU fake writer threads");

You can set it to a negative as well as to number of CPUs or any other number.

Sorry I just meant amend the description to something like "Number of RCU fake writer threads (or set to -1 for NR_CPUs)", so user does not have to read code to know that (and update the kernel cmdline params document as well).

Should we also adjust the string for nreaders?

Thanx, Paul

author maintainer
#14 lore raw
Quoted 30 lines · Paul E. McKenney
Quoted 29 lines · Joel Fernandes
Quoted 26 lines · Uladzislau Rezki
Quoted 25 lines · Joel Fernandes

kernel/rcu/rcutorture.c MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

Quoted 18 lines · Uladzislau Rezki (Sony)

Currently "nfakewriters" parameter can be set to any value but there is no possibility to adjust it automatically based on how many CPUs a system has where a test is run on.

To address this, if the "nfakewriters" is set to negative it will be adjusted to num_online_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
kernel/rcu/rcutorture.c +0 -0
index d98b3bd6d91f..f376262532ce 100644
@@ -148,6 +148,7 @@ MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, srcu, ...)");

IMO, this should also be updated to reflect the possibily to set it negative and hence to number CPUs:

torture_param(int, nfakewriters, 4, "Number of RCU fake writer threads");

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

You can set it to a negative as well as to number of CPUs or any other number.

Sorry I just meant amend the description to something like "Number of RCU fake writer threads (or set to -1 for NR_CPUs)", so user does not have to read code to know that (and update the kernel cmdline params document as well).

Should we also adjust the string for nreaders?

Makes sense. Both options are adjusted in same way.

--
Uladzislau Rezki