Skip to contentSkip to conversation

rcutorture: Allow a negative value for nfakewriters

PATCH Superseded Uladzislau Rezki (Sony) 2 participants 32 messages Jan 23, 2025 – Jan 28, 2025

Subsystems (4)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Uladzislau Rezki (Sony)

author maintainer
#0 lore raw

PATCH 1/4 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_possible_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 4 ++++
 1 file changed, 4 insertions(+)
kernel/rcu/rcutorture.c +4 -0
index d26fb1d33ed9..6bc161e1e8ac 100644
@@ -4050,6 +4050,10 @@ rcu_torture_init(void)
 					  writer_task);
 	if (torture_init_error(firsterr))
 		goto unwind;
+
+	if (nfakewriters < 0)
+		nfakewriters = (int) num_possible_cpus();
+
 	if (nfakewriters > 0) {
 		fakewriter_tasks = kcalloc(nfakewriters,
 					   sizeof(fakewriter_tasks[0]),
Signature · 1 line
2.39.5
author maintainer
#1 lore raw

PATCH 2/4 torture: Remove CONFIG_NR_CPUS configuration

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

---
 tools/testing/selftests/rcutorture/configs/rcu/TREE05 | 1 -
 1 file changed, 1 deletion(-)
tools/testing/selftests/rcutorture/configs/rcu/TREE05 +0 -1
index 9f48c73709ec..d6fbb82e3e6d 100644
@@ -1,5 +1,4 @@
 CONFIG_SMP=y
-CONFIG_NR_CPUS=8
 CONFIG_PREEMPT_NONE=y
 CONFIG_PREEMPT_VOLUNTARY=n
 CONFIG_PREEMPT=n
Signature · 1 line
2.39.5
maintainer
#2 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Or am I missing something here?

Thanx, Paul

Quoted 17 lines
---
 tools/testing/selftests/rcutorture/configs/rcu/TREE05 | 1 -
 1 file changed, 1 deletion(-)
tools/testing/selftests/rcutorture/configs/rcu/TREE05 +0 -1
index 9f48c73709ec..d6fbb82e3e6d 100644
@@ -1,5 +1,4 @@
 CONFIG_SMP=y
-CONFIG_NR_CPUS=8
 CONFIG_PREEMPT_NONE=y
 CONFIG_PREEMPT_VOLUNTARY=n
 CONFIG_PREEMPT=n
Signature · 1 line
2.39.5
author maintainer
#3 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

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

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

Quoted 9 lines
Quoted 1 line

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

--
Uladzislau Rezki

maintainer
#4 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

Thanx, Paul

author maintainer
#5 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

I will not resist, we just drop this patch :)

--
Uladzislau Rezki

maintainer
#6 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 35 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 8 lines
Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thoughts?

Thanx, Paul

author maintainer
#7 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 3 lines

Thoughts?

If you mean below splat:

<snip>

[   32.107748] =============================
[   32.108512] WARNING: suspicious RCU usage
[   32.109232] 6.12.0-rc4-dirty #66 Not tainted
[   32.110058] -----------------------------
[   32.110817] kernel/events/core.c:13962 RCU-list traversed in non-reader section!!
[   32.111221] kworker/u34:2 (251) used greatest stack depth: 12112 bytes left
[   32.112125]
[   32.112125] other info that might help us debug this:
[   32.112125]
[   32.112130]
[   32.112130] rcu_scheduler_active = 2, debug_locks = 1
[   32.116039] 3 locks held by cpuhp/1/20:
[   32.116758]  #0: ffffffff93a6a750 (cpu_hotplug_lock){++++}-{0:0}, at: cpuhp_thread_fun+0x50/0x220
[   32.118410]  #1: ffffffff93a6ce00 (cpuhp_state-down){+.+.}-{0:0}, at: cpuhp_thread_fun+0x50/0x220
[   32.120091]  #2: ffffffff93b7eb68 (pmus_lock){+.+.}-{3:3}, at: perf_event_exit_cpu_context+0x32/0x2d0
[   32.121723]
[   32.121723] stack backtrace:
[   32.122413] CPU: 1 UID: 0 PID: 20 Comm: cpuhp/1 Not tainted 6.12.0-rc4-dirty #66
[   32.123666] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1ubuntu1.1 04/01/2014
[   32.125302] Call Trace:
[   32.125769]  <TASK>
[   32.126148]  dump_stack_lvl+0x83/0xa0
[   32.126823]  lockdep_rcu_suspicious+0x113/0x180
[   32.127652]  perf_event_exit_cpu_context+0x2c4/0x2d0
[   32.128593]  ? __pfx_perf_event_exit_cpu+0x10/0x10
[   32.129489]  perf_event_exit_cpu+0x9/0x10
[   32.130243]  cpuhp_invoke_callback+0x187/0x6e0
[   32.131065]  ? cpuhp_thread_fun+0x50/0x220
[   32.131800]  cpuhp_thread_fun+0x185/0x220
[   32.132560]  ? __pfx_smpboot_thread_fn+0x10/0x10                                                                                                                           [   32.133394]  smpboot_thread_fn+0xd8/0x1d0
[   32.134050]  kthread+0xd0/0x100
[   32.134592]  ? __pfx_kthread+0x10/0x10
[   32.135270]  ret_from_fork+0x2f/0x50
[   32.135896]  ? __pfx_kthread+0x10/0x10
[   32.136610]  ret_from_fork_asm+0x1a/0x30
[   32.137356]  </TASK>
[   32.140997] smpboot: CPU 1 is now offline

<snip>

I reproduced that using:

+rcutorture.nfakewriters=128
+rcutorture.gp_sync=1
+rcupdate.rcu_expedited=0
+rcupdate.rcu_normal=1
+rcutree.rcu_normal_wake_from_gp=1
<snip>

The test script:

for (( i=0; i<$LOOPS; i++ )); do

        tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 64 --configs \
        '100*TREE05' --memory 20G --bootargs 'rcutorture.fwd_progress=1'
        echo "Done $i"

done

i.e. with more nfakewriters.

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

--
Uladzislau Rezki

maintainer
#8 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 77 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 13 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128

Do this:

	--configs "16*TREE05"

Or maybe even this:

Quoted 7 lines
Quoted 1 line
	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 3 lines

Thoughts?

If you mean below splat:

No, instead the one reported by cheung wall <[email protected]>.

Quoted 51 lines

<snip>

[   32.107748] =============================
[   32.108512] WARNING: suspicious RCU usage
[   32.109232] 6.12.0-rc4-dirty #66 Not tainted
[   32.110058] -----------------------------
[   32.110817] kernel/events/core.c:13962 RCU-list traversed in non-reader section!!
[   32.111221] kworker/u34:2 (251) used greatest stack depth: 12112 bytes left
[   32.112125]
[   32.112125] other info that might help us debug this:
[   32.112125]
[   32.112130]
[   32.112130] rcu_scheduler_active = 2, debug_locks = 1
[   32.116039] 3 locks held by cpuhp/1/20:
[   32.116758]  #0: ffffffff93a6a750 (cpu_hotplug_lock){++++}-{0:0}, at: cpuhp_thread_fun+0x50/0x220
[   32.118410]  #1: ffffffff93a6ce00 (cpuhp_state-down){+.+.}-{0:0}, at: cpuhp_thread_fun+0x50/0x220
[   32.120091]  #2: ffffffff93b7eb68 (pmus_lock){+.+.}-{3:3}, at: perf_event_exit_cpu_context+0x32/0x2d0
[   32.121723]
[   32.121723] stack backtrace:
[   32.122413] CPU: 1 UID: 0 PID: 20 Comm: cpuhp/1 Not tainted 6.12.0-rc4-dirty #66
[   32.123666] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1ubuntu1.1 04/01/2014
[   32.125302] Call Trace:
[   32.125769]  <TASK>
[   32.126148]  dump_stack_lvl+0x83/0xa0
[   32.126823]  lockdep_rcu_suspicious+0x113/0x180
[   32.127652]  perf_event_exit_cpu_context+0x2c4/0x2d0
[   32.128593]  ? __pfx_perf_event_exit_cpu+0x10/0x10
[   32.129489]  perf_event_exit_cpu+0x9/0x10
[   32.130243]  cpuhp_invoke_callback+0x187/0x6e0
[   32.131065]  ? cpuhp_thread_fun+0x50/0x220
[   32.131800]  cpuhp_thread_fun+0x185/0x220
[   32.132560]  ? __pfx_smpboot_thread_fn+0x10/0x10                                                                                                                           [   32.133394]  smpboot_thread_fn+0xd8/0x1d0
[   32.134050]  kthread+0xd0/0x100
[   32.134592]  ? __pfx_kthread+0x10/0x10
[   32.135270]  ret_from_fork+0x2f/0x50
[   32.135896]  ? __pfx_kthread+0x10/0x10
[   32.136610]  ret_from_fork_asm+0x1a/0x30
[   32.137356]  </TASK>
[   32.140997] smpboot: CPU 1 is now offline

<snip>

I reproduced that using:

(quoted diff) +5 -0
+rcutorture.nfakewriters=128
+rcutorture.gp_sync=1
+rcupdate.rcu_expedited=0
+rcupdate.rcu_normal=1
+rcutree.rcu_normal_wake_from_gp=1
<snip>

The test script:

Quoted 7 lines

for (( i=0; i<$LOOPS; i++ )); do

        tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 64 --configs \
        '100*TREE05' --memory 20G --bootargs 'rcutorture.fwd_progress=1'
        echo "Done $i"

done

i.e. with more nfakewriters.

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

Thanx, Paul

author maintainer
#9 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 89 lines · Paul E. McKenney
Quoted 81 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

--
Uladzislau Rezki

maintainer
#10 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 99 lines · Uladzislau Rezki
Quoted 89 lines · Paul E. McKenney
Quoted 81 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

Quoted 7 lines

OK. I will give a try this:

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

Thanx, Paul

author maintainer
#11 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 102 lines · Paul E. McKenney
Quoted 101 lines · Uladzislau Rezki
Quoted 89 lines · Paul E. McKenney
Quoted 81 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

Quoted 9 lines
Quoted 5 lines

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

By default we have NR_CPUS=8, we we discussed. Providing to kvm "--cpus 5" parameter will just set number of CPUs for a VM to 5:

<snip>
...

[    0.060672] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=5, Nodes=1

...
<snip>

so, for my test i do not see why i need to set --kconfig CONFIG_NR_CPUS=4.

Am i missing something? :)

--
Uladzislau Rezki

maintainer
#12 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 118 lines · Uladzislau Rezki
Quoted 102 lines · Paul E. McKenney
Quoted 101 lines · Uladzislau Rezki
Quoted 89 lines · Paul E. McKenney
Quoted 81 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

Quoted 9 lines
Quoted 5 lines

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

By default we have NR_CPUS=8, we we discussed. Providing to kvm "--cpus 5" parameter will just set number of CPUs for a VM to 5:

<snip>
...

Quoted 7 lines
[    0.060672] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=5, Nodes=1

...
<snip>

so, for my test i do not see why i need to set --kconfig CONFIG_NR_CPUS=4.

Am i missing something? :)

Because that gets you more guest OSes running on your system, each with one RCU-update kthread that is being checked by RCU reader kthreads. Therefore, it might double the rate at which you are able to reproduce this issue.

Thanx, Paul

author maintainer
#13 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 123 lines · Paul E. McKenney
Quoted 122 lines · Uladzislau Rezki
Quoted 102 lines · Paul E. McKenney
Quoted 101 lines · Uladzislau Rezki
Quoted 89 lines · Paul E. McKenney
Quoted 81 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which is set to 8. The problem is that it can not be overwritten for something higher.

Remove that configuration for TREE05, so it is possible to run the torture test on as many CPUs as many system has.

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

Quoted 9 lines
Quoted 5 lines

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

By default we have NR_CPUS=8, we we discussed. Providing to kvm "--cpus 5" parameter will just set number of CPUs for a VM to 5:

<snip>
...

[    0.060672] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=5, Nodes=1

...
<snip>

Quoted 9 lines
Quoted 3 lines

so, for my test i do not see why i need to set --kconfig CONFIG_NR_CPUS=4.

Am i missing something? :)

Because that gets you more guest OSes running on your system, each with one RCU-update kthread that is being checked by RCU reader kthreads. Therefore, it might double the rate at which you are able to reproduce this issue.

You mean that setting --kconfig CONFIG_NR_CPUS=4 and 16*TREE05 will run 4 separate KVM instances?

--
Uladzislau Rezki

maintainer
#14 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 128 lines · Uladzislau Rezki
Quoted 127 lines · Paul E. McKenney
Quoted 125 lines · Uladzislau Rezki
Quoted 102 lines · Paul E. McKenney
Quoted 101 lines · Uladzislau Rezki
Quoted 89 lines · Paul E. McKenney
Quoted 81 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney
Quoted 8 lines · Uladzislau Rezki (Sony)

This configuration specifies the maximum number of CPUs which
is set to 8. The problem is that it can not be overwritten for
something higher.

Remove that configuration for TREE05, so it is possible to run
the torture test on as many CPUs as many system has.

Signed-off-by: Uladzislau Rezki (Sony) <[email protected]>

You should be able to override this on the kvm.sh command line by specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish. For example, see the torture.sh querying the system's number of CPUs and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

Quoted 9 lines
Quoted 5 lines

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

By default we have NR_CPUS=8, we we discussed. Providing to kvm "--cpus 5" parameter will just set number of CPUs for a VM to 5:

<snip>
...

[    0.060672] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=5, Nodes=1

...
<snip>

so, for my test i do not see why i need to set --kconfig CONFIG_NR_CPUS=4.

Am i missing something? :)

Quoted 7 lines
Quoted 5 lines

Because that gets you more guest OSes running on your system, each with one RCU-update kthread that is being checked by RCU reader kthreads. Therefore, it might double the rate at which you are able to reproduce this issue.

You mean that setting --kconfig CONFIG_NR_CPUS=4 and 16*TREE05 will run 4 separate KVM instances?

Almost but not quite.

I am assuming that you have a system with a multiple of eight CPUs.

If so, and assuming that Cheung's bug is an interaction between a fast synchronize_rcu() grace period and a reader task that this grace period is waiting on, having more and smaller guest OSes might make the problem happen faster. So instead of your:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'

You might be able to double the number of reproductions of the bug per unit time by instead using:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'32*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

Does that seem reasonable to you?

Thanx, Paul

author maintainer
#15 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 149 lines · Paul E. McKenney
Quoted 128 lines · Uladzislau Rezki
Quoted 127 lines · Paul E. McKenney
Quoted 125 lines · Uladzislau Rezki
Quoted 102 lines · Paul E. McKenney
Quoted 101 lines · Uladzislau Rezki
Quoted 89 lines · Paul E. McKenney
Quoted 81 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney

On Thu, Jan 23, 2025 at 07:58:26PM +0100, Uladzislau Rezki (Sony) wrote:
This configuration specifies the maximum number of CPUs which
is set to 8. The problem is that it can not be overwritten for
something higher.

Remove that configuration for TREE05, so it is possible to run
the torture test on as many CPUs as many system has.

Signed-off-by: Uladzislau Rezki (Sony) <[email protected]>

You should be able to override this on the kvm.sh command line by
specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish.
For example, see the torture.sh querying the system's number of CPUs
and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

Quoted 9 lines
Quoted 5 lines

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

By default we have NR_CPUS=8, we we discussed. Providing to kvm "--cpus 5" parameter will just set number of CPUs for a VM to 5:

<snip>
...

[    0.060672] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=5, Nodes=1

...
<snip>

so, for my test i do not see why i need to set --kconfig CONFIG_NR_CPUS=4.

Am i missing something? :)

Quoted 7 lines
Quoted 5 lines

Because that gets you more guest OSes running on your system, each with one RCU-update kthread that is being checked by RCU reader kthreads. Therefore, it might double the rate at which you are able to reproduce this issue.

You mean that setting --kconfig CONFIG_NR_CPUS=4 and 16*TREE05 will run 4 separate KVM instances?

Almost but not quite.

I am assuming that you have a system with a multiple of eight CPUs.

If so, and assuming that Cheung's bug is an interaction between a fast synchronize_rcu() grace period and a reader task that this grace period is waiting on, having more and smaller guest OSes might make the problem happen faster. So instead of your:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'

Quoted 9 lines

You might be able to double the number of reproductions of the bug per unit time by instead using:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'32*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

Does that seem reasonable to you?

I was confused by the: how CONFIG_NR_CPUS can influence on number of instances kvm.sh runs.

It is obvious, that as much parallel setups you run as faster you can reproduce it. Of course if there are enough resources a system runs the test.

--
Uladzislau Rezki

author maintainer
#16 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

In reply to Paul E. McKenney
Quoted 149 lines · Paul E. McKenney
Quoted 128 lines · Uladzislau Rezki
Quoted 127 lines · Paul E. McKenney
Quoted 125 lines · Uladzislau Rezki
Quoted 102 lines · Paul E. McKenney
Quoted 101 lines · Uladzislau Rezki
Quoted 89 lines · Paul E. McKenney
Quoted 81 lines · Uladzislau Rezki
Quoted 43 lines · Paul E. McKenney
Quoted 42 lines · Uladzislau Rezki
Quoted 34 lines · Paul E. McKenney
Quoted 16 lines · Uladzislau Rezki
Quoted 15 lines · Paul E. McKenney

On Thu, Jan 23, 2025 at 07:58:26PM +0100, Uladzislau Rezki (Sony) wrote:
This configuration specifies the maximum number of CPUs which
is set to 8. The problem is that it can not be overwritten for
something higher.

Remove that configuration for TREE05, so it is possible to run
the torture test on as many CPUs as many system has.

Signed-off-by: Uladzislau Rezki (Sony) <[email protected]>

You should be able to override this on the kvm.sh command line by
specifying "--kconfig CONFIG_NR_CPUS=128" or whatever number you wish.
For example, see the torture.sh querying the system's number of CPUs
and then specifying it to a number of tests.

Quoted 8 lines
Quoted 2 lines

Or am I missing something here?

It took me a while to understand what happens. Apparently there is this 8 CPUs limitation. Yes, i can do it manually by passing --kconfig but you need to know about that. I have not expected that.

Therefore i removed it from the configuration because i have not found a good explanation why we need. It is confusing instead :)

Right now, if I do a run with --configs "TREE10 14*CFLIST", this will make use of 20 systems with 80 CPUs each. If you remove that line from TREE05, won't each instance of TREE05 consume a full system, for a total of 33 systems? Yes, I could use "--kconfig CONFIG_NR_CPUS=8" on the command line, but that would affect all the scenarios, not just TREE05. Including (say) TINY01, where I believe that it would cause kvm.sh to complain about a Kconfig conflict.

Quoted 4 lines

Hence me not being in favor of this change. ;-)

Is there another way to make things work for both situations?

OK, i see. Well. I will just go with --kconfig CONFIG_NR_CPUS=foo if i need more CPUs for TREE05.

Quoted 7 lines
Quoted 1 line

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

Quoted 9 lines
Quoted 5 lines

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

By default we have NR_CPUS=8, we we discussed. Providing to kvm "--cpus 5" parameter will just set number of CPUs for a VM to 5:

<snip>
...

[    0.060672] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=5, Nodes=1

...
<snip>

so, for my test i do not see why i need to set --kconfig CONFIG_NR_CPUS=4.

Am i missing something? :)

Quoted 7 lines
Quoted 5 lines

Because that gets you more guest OSes running on your system, each with one RCU-update kthread that is being checked by RCU reader kthreads. Therefore, it might double the rate at which you are able to reproduce this issue.

You mean that setting --kconfig CONFIG_NR_CPUS=4 and 16*TREE05 will run 4 separate KVM instances?

Almost but not quite.

I am assuming that you have a system with a multiple of eight CPUs.

If so, and assuming that Cheung's bug is an interaction between a fast synchronize_rcu() grace period and a reader task that this grace period is waiting on, having more and smaller guest OSes might make the problem happen faster. So instead of your:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'

Quoted 9 lines

You might be able to double the number of reproductions of the bug per unit time by instead using:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'32*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

Does that seem reasonable to you?

It only runs one instance for me:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs 32*TREE05 --memory 10G --bootargs rcutorture.fwd_progress=1 --kconfig CONFIG_NR_CPUS=4 ----Start batch 1: Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Starting build. Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Waiting for build to complete. Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Build complete. Mon Jan 27 08:21:26 PM CET 2025
---- TREE05 4: Kernel present. Mon Jan 27 08:21:26 PM CET 2025
---- Starting kernels. Mon Jan 27 08:21:26 PM CET 2025

with 4 CPUs inside VM :)

--
Uladzislau Rezki

author maintainer
#17 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 119 lines
Quoted 100 lines
Quoted 80 lines
Quoted 80 lines
Quoted 79 lines
Quoted 57 lines
Quoted 57 lines
Quoted 46 lines
Quoted 39 lines
Quoted 9 lines
Quoted 3 lines

need more CPUs for TREE05.

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

Quoted 9 lines
Quoted 5 lines

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

By default we have NR_CPUS=8, we we discussed. Providing to kvm "--cpus 5" parameter will just set number of CPUs for a VM to 5:

<snip>
...

[    0.060672] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=5, Nodes=1

...
<snip>

so, for my test i do not see why i need to set --kconfig CONFIG_NR_CPUS=4.

Am i missing something? :)

Quoted 7 lines
Quoted 5 lines

Because that gets you more guest OSes running on your system, each with one RCU-update kthread that is being checked by RCU reader kthreads. Therefore, it might double the rate at which you are able to reproduce this issue.

You mean that setting --kconfig CONFIG_NR_CPUS=4 and 16*TREE05 will run 4 separate KVM instances?

Almost but not quite.

I am assuming that you have a system with a multiple of eight CPUs.

If so, and assuming that Cheung's bug is an interaction between a fast synchronize_rcu() grace period and a reader task that this grace period is waiting on, having more and smaller guest OSes might make the problem happen faster. So instead of your:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'

Quoted 9 lines

You might be able to double the number of reproductions of the bug per unit time by instead using:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'32*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

Does that seem reasonable to you?

It only runs one instance for me:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs 32*TREE05 --memory 10G --bootargs rcutorture.fwd_progress=1 --kconfig CONFIG_NR_CPUS=4 ----Start batch 1: Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Starting build. Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Waiting for build to complete. Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Build complete. Mon Jan 27 08:21:26 PM CET 2025
---- TREE05 4: Kernel present. Mon Jan 27 08:21:26 PM CET 2025
---- Starting kernels. Mon Jan 27 08:21:26 PM CET 2025

Quoted 2 lines

with 4 CPUs inside VM :)

And when running 16 instances with 4 CPUs each i can reproduce the splat which has been reported:

tools/testing/selftests/rcutorture/bin/kvm.sh --allcpus --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

<snip>
...

[    0.595251] ------------[ cut here ]------------
[    0.595867] A full grace period is not passed yet: 0
[    0.595875] WARNING: CPU: 1 PID: 16 at kernel/rcu/tree.c:1617 rcu_sr_normal_complete+0xa9/0xc0
[    0.598248] Modules linked in:
[    0.598649] CPU: 1 UID: 0 PID: 16 Comm: rcu_preempt Not tainted 6.13.0-02530-g8950af6a11ff #261
[    0.599248] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[    0.600248] RIP: 0010:rcu_sr_normal_complete+0xa9/0xc0
[    0.600913] Code: 48 29 c2 48 8d 04 0a ba 03 00 00 00 48 39 c2 79 0c 48 83 e8 04 48 c1 e8 02 48 8d 70 02 48 c7 c7 20 e9 33 b5 e8 d8 03 f4 ff 90 <0f> 0b 90 90 48 8d 7b 10 5b e9 f9 38 fb ff 66 0f 1f 84 00 00 00 00
[    0.603249] RSP: 0018:ffffadad0008be60 EFLAGS: 00010282
[    0.603925] RAX: 0000000000000000 RBX: ffffadad00013d10 RCX: 00000000ffffdfff
[    0.605247] RDX: 0000000000000000 RSI: ffffadad0008bd10 RDI: 0000000000000001
[    0.606247] RBP: 0000000000000000 R08: 0000000000009ffb R09: 00000000ffffdfff
[    0.607248] R10: 00000000ffffdfff R11: ffffffffb56789a0 R12: 0000000000000005
[    0.608247] R13: 0000000000031a40 R14: fffffffffffffb74 R15: 0000000000000000
[    0.609250] FS:  0000000000000000(0000) GS:ffff9081f5c80000(0000) knlGS:0000000000000000
[    0.610249] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.611248] CR2: 0000000000000000 CR3: 00000002f024a000 CR4: 00000000000006f0
[    0.612249] Call Trace:
[    0.612574]  <TASK>
[    0.612854]  ? __warn+0x8c/0x190
[    0.613248]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.613840]  ? report_bug+0x164/0x190
[    0.614248]  ? handle_bug+0x54/0x90
[    0.614705]  ? exc_invalid_op+0x17/0x70
[    0.615248]  ? asm_exc_invalid_op+0x1a/0x20
[    0.615797]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.616248]  rcu_gp_cleanup+0x403/0x5a0
[    0.616248]  ? __pfx_rcu_gp_kthread+0x10/0x10
[    0.616818]  rcu_gp_kthread+0x136/0x1c0
[    0.617249]  kthread+0xec/0x1f0
[    0.617664]  ? __pfx_kthread+0x10/0x10
[    0.618156]  ret_from_fork+0x2f/0x50
[    0.618728]  ? __pfx_kthread+0x10/0x10
[    0.619216]  ret_from_fork_asm+0x1a/0x30
[    0.620251]  </TASK>

...
<snip>

Linus tip-tree, HEAD is c4b9570cfb63501638db720f3bee9f6dfd044b82

--
Uladzislau Rezki

maintainer
#18 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 165 lines · Uladzislau Rezki
Quoted 119 lines
Quoted 100 lines
Quoted 80 lines
Quoted 80 lines
Quoted 79 lines
Quoted 57 lines
Quoted 57 lines
Quoted 46 lines
Quoted 39 lines
Quoted 9 lines
Quoted 3 lines

need more CPUs for TREE05.

I will not resist, we just drop this patch :)

Thank you!

The bug you are chasing happens when a given synchonize_rcu() interacts with RCU readers, correct?

Below one:

<snip>
/*
* RCU torture fake writer kthread. Repeatedly calls sync, with a random
* delay between calls.
*/
static int
rcu_torture_fakewriter(void *arg)
{
...
<snip>

Quoted 7 lines

In rcutorture, only the rcu_torture_writer() call to synchronize_rcu() interacts with rcu_torture_reader(). So my guess is that running many small TREE05 guest OSes would reproduce this bug more quickly. So instead of this:

	--kconfig CONFIG_NR_CPUS=128
Quoted 7 lines

Do this:

	--configs "16*TREE05"

Or maybe even this:

	--configs "16*TREE05" --kconfig CONFIG_NR_CPUS=4

Thanks for input.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Thoughts?

If you mean below splat:

i.e. with more nfakewriters.

Quoted 9 lines

Right, and large nfakewriters would help push the synchronize_rcu() wakeups off of the grace-period kthread.

Quoted 2 lines

If you mean the one that has recently reported, i am not able to reproduce it anyhow :)

Using larger numbers of smaller rcutorture guest OSes might help to reproduce it. Maybe as small as three CPUs each. ;-)

OK. I will give a try this:

Quoted 9 lines
Quoted 5 lines

for (( i=0; i<$LOOPS; i++ )); do

	tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
	'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'
	echo "Done $i"

done

Making each guest OS smaller needs "--kconfig CONFIG_NR_CPUS=4" (or whatever) as well, perhaps also increasing the "16*TREE05".

By default we have NR_CPUS=8, we we discussed. Providing to kvm "--cpus 5" parameter will just set number of CPUs for a VM to 5:

<snip>
...

[    0.060672] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=5, Nodes=1

...
<snip>

so, for my test i do not see why i need to set --kconfig CONFIG_NR_CPUS=4.

Am i missing something? :)

Quoted 7 lines
Quoted 5 lines

Because that gets you more guest OSes running on your system, each with one RCU-update kthread that is being checked by RCU reader kthreads. Therefore, it might double the rate at which you are able to reproduce this issue.

You mean that setting --kconfig CONFIG_NR_CPUS=4 and 16*TREE05 will run 4 separate KVM instances?

Almost but not quite.

I am assuming that you have a system with a multiple of eight CPUs.

If so, and assuming that Cheung's bug is an interaction between a fast synchronize_rcu() grace period and a reader task that this grace period is waiting on, having more and smaller guest OSes might make the problem happen faster. So instead of your:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1'

Quoted 9 lines

You might be able to double the number of reproductions of the bug per unit time by instead using:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs \
'32*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

Does that seem reasonable to you?

It only runs one instance for me:

tools/testing/selftests/rcutorture/bin/kvm.sh --cpus 5 --configs 32*TREE05 --memory 10G --bootargs rcutorture.fwd_progress=1 --kconfig CONFIG_NR_CPUS=4 ----Start batch 1: Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Starting build. Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Waiting for build to complete. Mon Jan 27 08:20:17 PM CET 2025
TREE05 4: Build complete. Mon Jan 27 08:21:26 PM CET 2025
---- TREE05 4: Kernel present. Mon Jan 27 08:21:26 PM CET 2025
---- Starting kernels. Mon Jan 27 08:21:26 PM CET 2025

Quoted 2 lines

with 4 CPUs inside VM :)

And when running 16 instances with 4 CPUs each i can reproduce the splat which has been reported:

tools/testing/selftests/rcutorture/bin/kvm.sh --allcpus --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

<snip>
...

[    0.595251] ------------[ cut here ]------------
[    0.595867] A full grace period is not passed yet: 0
[    0.595875] WARNING: CPU: 1 PID: 16 at kernel/rcu/tree.c:1617 rcu_sr_normal_complete+0xa9/0xc0
[    0.598248] Modules linked in:
[    0.598649] CPU: 1 UID: 0 PID: 16 Comm: rcu_preempt Not tainted 6.13.0-02530-g8950af6a11ff #261
[    0.599248] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[    0.600248] RIP: 0010:rcu_sr_normal_complete+0xa9/0xc0
[    0.600913] Code: 48 29 c2 48 8d 04 0a ba 03 00 00 00 48 39 c2 79 0c 48 83 e8 04 48 c1 e8 02 48 8d 70 02 48 c7 c7 20 e9 33 b5 e8 d8 03 f4 ff 90 <0f> 0b 90 90 48 8d 7b 10 5b e9 f9 38 fb ff 66 0f 1f 84 00 00 00 00
[    0.603249] RSP: 0018:ffffadad0008be60 EFLAGS: 00010282
[    0.603925] RAX: 0000000000000000 RBX: ffffadad00013d10 RCX: 00000000ffffdfff
[    0.605247] RDX: 0000000000000000 RSI: ffffadad0008bd10 RDI: 0000000000000001
[    0.606247] RBP: 0000000000000000 R08: 0000000000009ffb R09: 00000000ffffdfff
[    0.607248] R10: 00000000ffffdfff R11: ffffffffb56789a0 R12: 0000000000000005
[    0.608247] R13: 0000000000031a40 R14: fffffffffffffb74 R15: 0000000000000000
[    0.609250] FS:  0000000000000000(0000) GS:ffff9081f5c80000(0000) knlGS:0000000000000000
[    0.610249] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.611248] CR2: 0000000000000000 CR3: 00000002f024a000 CR4: 00000000000006f0
[    0.612249] Call Trace:
[    0.612574]  <TASK>
[    0.612854]  ? __warn+0x8c/0x190
[    0.613248]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.613840]  ? report_bug+0x164/0x190
[    0.614248]  ? handle_bug+0x54/0x90
[    0.614705]  ? exc_invalid_op+0x17/0x70
[    0.615248]  ? asm_exc_invalid_op+0x1a/0x20
[    0.615797]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.616248]  rcu_gp_cleanup+0x403/0x5a0
[    0.616248]  ? __pfx_rcu_gp_kthread+0x10/0x10
[    0.616818]  rcu_gp_kthread+0x136/0x1c0
[    0.617249]  kthread+0xec/0x1f0
[    0.617664]  ? __pfx_kthread+0x10/0x10
[    0.618156]  ret_from_fork+0x2f/0x50
[    0.618728]  ? __pfx_kthread+0x10/0x10
[    0.619216]  ret_from_fork_asm+0x1a/0x30
[    0.620251]  </TASK>
Quoted 4 lines

...
<snip>

Linus tip-tree, HEAD is c4b9570cfb63501638db720f3bee9f6dfd044b82

Very good! And of course, the next question is "does going to _full() make the problem go away?" ;-)

Thanx, Paul

author maintainer
#19 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 49 lines
Quoted 49 lines
Quoted 2 lines

with 4 CPUs inside VM :)

And when running 16 instances with 4 CPUs each i can reproduce the splat which has been reported:

tools/testing/selftests/rcutorture/bin/kvm.sh --allcpus --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

<snip>
...

[    0.595251] ------------[ cut here ]------------
[    0.595867] A full grace period is not passed yet: 0
[    0.595875] WARNING: CPU: 1 PID: 16 at kernel/rcu/tree.c:1617 rcu_sr_normal_complete+0xa9/0xc0
[    0.598248] Modules linked in:
[    0.598649] CPU: 1 UID: 0 PID: 16 Comm: rcu_preempt Not tainted 6.13.0-02530-g8950af6a11ff #261
[    0.599248] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[    0.600248] RIP: 0010:rcu_sr_normal_complete+0xa9/0xc0
[    0.600913] Code: 48 29 c2 48 8d 04 0a ba 03 00 00 00 48 39 c2 79 0c 48 83 e8 04 48 c1 e8 02 48 8d 70 02 48 c7 c7 20 e9 33 b5 e8 d8 03 f4 ff 90 <0f> 0b 90 90 48 8d 7b 10 5b e9 f9 38 fb ff 66 0f 1f 84 00 00 00 00
[    0.603249] RSP: 0018:ffffadad0008be60 EFLAGS: 00010282
[    0.603925] RAX: 0000000000000000 RBX: ffffadad00013d10 RCX: 00000000ffffdfff
[    0.605247] RDX: 0000000000000000 RSI: ffffadad0008bd10 RDI: 0000000000000001
[    0.606247] RBP: 0000000000000000 R08: 0000000000009ffb R09: 00000000ffffdfff
[    0.607248] R10: 00000000ffffdfff R11: ffffffffb56789a0 R12: 0000000000000005
[    0.608247] R13: 0000000000031a40 R14: fffffffffffffb74 R15: 0000000000000000
[    0.609250] FS:  0000000000000000(0000) GS:ffff9081f5c80000(0000) knlGS:0000000000000000
[    0.610249] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.611248] CR2: 0000000000000000 CR3: 00000002f024a000 CR4: 00000000000006f0
[    0.612249] Call Trace:
[    0.612574]  <TASK>
[    0.612854]  ? __warn+0x8c/0x190
[    0.613248]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.613840]  ? report_bug+0x164/0x190
[    0.614248]  ? handle_bug+0x54/0x90
[    0.614705]  ? exc_invalid_op+0x17/0x70
[    0.615248]  ? asm_exc_invalid_op+0x1a/0x20
[    0.615797]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.616248]  rcu_gp_cleanup+0x403/0x5a0
[    0.616248]  ? __pfx_rcu_gp_kthread+0x10/0x10
[    0.616818]  rcu_gp_kthread+0x136/0x1c0
[    0.617249]  kthread+0xec/0x1f0
[    0.617664]  ? __pfx_kthread+0x10/0x10
[    0.618156]  ret_from_fork+0x2f/0x50
[    0.618728]  ? __pfx_kthread+0x10/0x10
[    0.619216]  ret_from_fork_asm+0x1a/0x30
[    0.620251]  </TASK>

...
<snip>

Quoted 5 lines
Quoted 1 line

Linus tip-tree, HEAD is c4b9570cfb63501638db720f3bee9f6dfd044b82

Very good! And of course, the next question is "does going to _full() make the problem go away?" ;-)

Yes does its job if i apply:

https://lore.kernel.org/rcu/00900afe-ac4e-4362-a3f9-d65f2c9dcd9a@paulmck-laptop/T/#m5d9263f3825d3170c044beedbae741717702d4aa

after that i am not able to reproduce the warning anymore. Tested over night. Without it, i can reproduce it pretty easy :)

--
Uladzislau Rezki

maintainer
#20 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 54 lines · Uladzislau Rezki
Quoted 49 lines
Quoted 49 lines
Quoted 2 lines

with 4 CPUs inside VM :)

And when running 16 instances with 4 CPUs each i can reproduce the splat which has been reported:

tools/testing/selftests/rcutorture/bin/kvm.sh --allcpus --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

<snip>
...

[    0.595251] ------------[ cut here ]------------
[    0.595867] A full grace period is not passed yet: 0
[    0.595875] WARNING: CPU: 1 PID: 16 at kernel/rcu/tree.c:1617 rcu_sr_normal_complete+0xa9/0xc0
[    0.598248] Modules linked in:
[    0.598649] CPU: 1 UID: 0 PID: 16 Comm: rcu_preempt Not tainted 6.13.0-02530-g8950af6a11ff #261
[    0.599248] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[    0.600248] RIP: 0010:rcu_sr_normal_complete+0xa9/0xc0
[    0.600913] Code: 48 29 c2 48 8d 04 0a ba 03 00 00 00 48 39 c2 79 0c 48 83 e8 04 48 c1 e8 02 48 8d 70 02 48 c7 c7 20 e9 33 b5 e8 d8 03 f4 ff 90 <0f> 0b 90 90 48 8d 7b 10 5b e9 f9 38 fb ff 66 0f 1f 84 00 00 00 00
[    0.603249] RSP: 0018:ffffadad0008be60 EFLAGS: 00010282
[    0.603925] RAX: 0000000000000000 RBX: ffffadad00013d10 RCX: 00000000ffffdfff
[    0.605247] RDX: 0000000000000000 RSI: ffffadad0008bd10 RDI: 0000000000000001
[    0.606247] RBP: 0000000000000000 R08: 0000000000009ffb R09: 00000000ffffdfff
[    0.607248] R10: 00000000ffffdfff R11: ffffffffb56789a0 R12: 0000000000000005
[    0.608247] R13: 0000000000031a40 R14: fffffffffffffb74 R15: 0000000000000000
[    0.609250] FS:  0000000000000000(0000) GS:ffff9081f5c80000(0000) knlGS:0000000000000000
[    0.610249] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.611248] CR2: 0000000000000000 CR3: 00000002f024a000 CR4: 00000000000006f0
[    0.612249] Call Trace:
[    0.612574]  <TASK>
[    0.612854]  ? __warn+0x8c/0x190
[    0.613248]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.613840]  ? report_bug+0x164/0x190
[    0.614248]  ? handle_bug+0x54/0x90
[    0.614705]  ? exc_invalid_op+0x17/0x70
[    0.615248]  ? asm_exc_invalid_op+0x1a/0x20
[    0.615797]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.616248]  rcu_gp_cleanup+0x403/0x5a0
[    0.616248]  ? __pfx_rcu_gp_kthread+0x10/0x10
[    0.616818]  rcu_gp_kthread+0x136/0x1c0
[    0.617249]  kthread+0xec/0x1f0
[    0.617664]  ? __pfx_kthread+0x10/0x10
[    0.618156]  ret_from_fork+0x2f/0x50
[    0.618728]  ? __pfx_kthread+0x10/0x10
[    0.619216]  ret_from_fork_asm+0x1a/0x30
[    0.620251]  </TASK>

...
<snip>

Quoted 5 lines
Quoted 1 line

Linus tip-tree, HEAD is c4b9570cfb63501638db720f3bee9f6dfd044b82

Very good! And of course, the next question is "does going to _full() make the problem go away?" ;-)

Quoted 6 lines

Yes does its job if i apply:

https://lore.kernel.org/rcu/00900afe-ac4e-4362-a3f9-d65f2c9dcd9a@paulmck-laptop/T/#m5d9263f3825d3170c044beedbae741717702d4aa

after that i am not able to reproduce the warning anymore. Tested over night. Without it, i can reproduce it pretty easy :)

Thank you, and good to hear!!!

May I add your Tested-by to that patch?

Thanx, Paul

author maintainer Tested-by
#21 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 57 lines · Paul E. McKenney
Quoted 56 lines · Uladzislau Rezki
Quoted 49 lines
Quoted 49 lines
Quoted 2 lines

with 4 CPUs inside VM :)

And when running 16 instances with 4 CPUs each i can reproduce the splat which has been reported:

tools/testing/selftests/rcutorture/bin/kvm.sh --allcpus --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

<snip>
...

[    0.595251] ------------[ cut here ]------------
[    0.595867] A full grace period is not passed yet: 0
[    0.595875] WARNING: CPU: 1 PID: 16 at kernel/rcu/tree.c:1617 rcu_sr_normal_complete+0xa9/0xc0
[    0.598248] Modules linked in:
[    0.598649] CPU: 1 UID: 0 PID: 16 Comm: rcu_preempt Not tainted 6.13.0-02530-g8950af6a11ff #261
[    0.599248] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[    0.600248] RIP: 0010:rcu_sr_normal_complete+0xa9/0xc0
[    0.600913] Code: 48 29 c2 48 8d 04 0a ba 03 00 00 00 48 39 c2 79 0c 48 83 e8 04 48 c1 e8 02 48 8d 70 02 48 c7 c7 20 e9 33 b5 e8 d8 03 f4 ff 90 <0f> 0b 90 90 48 8d 7b 10 5b e9 f9 38 fb ff 66 0f 1f 84 00 00 00 00
[    0.603249] RSP: 0018:ffffadad0008be60 EFLAGS: 00010282
[    0.603925] RAX: 0000000000000000 RBX: ffffadad00013d10 RCX: 00000000ffffdfff
[    0.605247] RDX: 0000000000000000 RSI: ffffadad0008bd10 RDI: 0000000000000001
[    0.606247] RBP: 0000000000000000 R08: 0000000000009ffb R09: 00000000ffffdfff
[    0.607248] R10: 00000000ffffdfff R11: ffffffffb56789a0 R12: 0000000000000005
[    0.608247] R13: 0000000000031a40 R14: fffffffffffffb74 R15: 0000000000000000
[    0.609250] FS:  0000000000000000(0000) GS:ffff9081f5c80000(0000) knlGS:0000000000000000
[    0.610249] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.611248] CR2: 0000000000000000 CR3: 00000002f024a000 CR4: 00000000000006f0
[    0.612249] Call Trace:
[    0.612574]  <TASK>
[    0.612854]  ? __warn+0x8c/0x190
[    0.613248]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.613840]  ? report_bug+0x164/0x190
[    0.614248]  ? handle_bug+0x54/0x90
[    0.614705]  ? exc_invalid_op+0x17/0x70
[    0.615248]  ? asm_exc_invalid_op+0x1a/0x20
[    0.615797]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.616248]  rcu_gp_cleanup+0x403/0x5a0
[    0.616248]  ? __pfx_rcu_gp_kthread+0x10/0x10
[    0.616818]  rcu_gp_kthread+0x136/0x1c0
[    0.617249]  kthread+0xec/0x1f0
[    0.617664]  ? __pfx_kthread+0x10/0x10
[    0.618156]  ret_from_fork+0x2f/0x50
[    0.618728]  ? __pfx_kthread+0x10/0x10
[    0.619216]  ret_from_fork_asm+0x1a/0x30
[    0.620251]  </TASK>

...
<snip>

Quoted 5 lines
Quoted 1 line

Linus tip-tree, HEAD is c4b9570cfb63501638db720f3bee9f6dfd044b82

Very good! And of course, the next question is "does going to _full() make the problem go away?" ;-)

Yes does its job if i apply:

Quoted 9 lines
Quoted 4 lines

https://lore.kernel.org/rcu/00900afe-ac4e-4362-a3f9-d65f2c9dcd9a@paulmck-laptop/T/#m5d9263f3825d3170c044beedbae741717702d4aa

after that i am not able to reproduce the warning anymore. Tested over night. Without it, i can reproduce it pretty easy :)

Thank you, and good to hear!!!

May I add your Tested-by to that patch?

Sure.

One question, we discussed that it is worth to print seq-delta in case of warning. Whereas a newly patch does do it and just emits a plain text.

I can send a separate patch or modify this one?

--
Uladzislau Rezki

maintainer
#22 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 63 lines · Uladzislau Rezki
Quoted 62 lines · Paul E. McKenney
Quoted 60 lines · Uladzislau Rezki
Quoted 49 lines
Quoted 49 lines
Quoted 2 lines

with 4 CPUs inside VM :)

And when running 16 instances with 4 CPUs each i can reproduce the splat which has been reported:

tools/testing/selftests/rcutorture/bin/kvm.sh --allcpus --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

<snip>
...

[    0.595251] ------------[ cut here ]------------
[    0.595867] A full grace period is not passed yet: 0
[    0.595875] WARNING: CPU: 1 PID: 16 at kernel/rcu/tree.c:1617 rcu_sr_normal_complete+0xa9/0xc0
[    0.598248] Modules linked in:
[    0.598649] CPU: 1 UID: 0 PID: 16 Comm: rcu_preempt Not tainted 6.13.0-02530-g8950af6a11ff #261
[    0.599248] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[    0.600248] RIP: 0010:rcu_sr_normal_complete+0xa9/0xc0
[    0.600913] Code: 48 29 c2 48 8d 04 0a ba 03 00 00 00 48 39 c2 79 0c 48 83 e8 04 48 c1 e8 02 48 8d 70 02 48 c7 c7 20 e9 33 b5 e8 d8 03 f4 ff 90 <0f> 0b 90 90 48 8d 7b 10 5b e9 f9 38 fb ff 66 0f 1f 84 00 00 00 00
[    0.603249] RSP: 0018:ffffadad0008be60 EFLAGS: 00010282
[    0.603925] RAX: 0000000000000000 RBX: ffffadad00013d10 RCX: 00000000ffffdfff
[    0.605247] RDX: 0000000000000000 RSI: ffffadad0008bd10 RDI: 0000000000000001
[    0.606247] RBP: 0000000000000000 R08: 0000000000009ffb R09: 00000000ffffdfff
[    0.607248] R10: 00000000ffffdfff R11: ffffffffb56789a0 R12: 0000000000000005
[    0.608247] R13: 0000000000031a40 R14: fffffffffffffb74 R15: 0000000000000000
[    0.609250] FS:  0000000000000000(0000) GS:ffff9081f5c80000(0000) knlGS:0000000000000000
[    0.610249] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.611248] CR2: 0000000000000000 CR3: 00000002f024a000 CR4: 00000000000006f0
[    0.612249] Call Trace:
[    0.612574]  <TASK>
[    0.612854]  ? __warn+0x8c/0x190
[    0.613248]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.613840]  ? report_bug+0x164/0x190
[    0.614248]  ? handle_bug+0x54/0x90
[    0.614705]  ? exc_invalid_op+0x17/0x70
[    0.615248]  ? asm_exc_invalid_op+0x1a/0x20
[    0.615797]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.616248]  rcu_gp_cleanup+0x403/0x5a0
[    0.616248]  ? __pfx_rcu_gp_kthread+0x10/0x10
[    0.616818]  rcu_gp_kthread+0x136/0x1c0
[    0.617249]  kthread+0xec/0x1f0
[    0.617664]  ? __pfx_kthread+0x10/0x10
[    0.618156]  ret_from_fork+0x2f/0x50
[    0.618728]  ? __pfx_kthread+0x10/0x10
[    0.619216]  ret_from_fork_asm+0x1a/0x30
[    0.620251]  </TASK>

...
<snip>

Quoted 5 lines
Quoted 1 line

Linus tip-tree, HEAD is c4b9570cfb63501638db720f3bee9f6dfd044b82

Very good! And of course, the next question is "does going to _full() make the problem go away?" ;-)

Yes does its job if i apply:

https://lore.kernel.org/rcu/00900afe-ac4e-4362-a3f9-d65f2c9dcd9a@paulmck-laptop/T/#m5d9263f3825d3170c044beedbae741717702d4aa

after that i am not able to reproduce the warning anymore. Tested over night. Without it, i can reproduce it pretty easy :)

Quoted 7 lines
Quoted 4 lines

Thank you, and good to hear!!!

May I add your Tested-by to that patch?

Sure.

Thank you! I will apply this on my next rebase.

Quoted 5 lines

One question, we discussed that it is worth to print seq-delta in case of warning. Whereas a newly patch does do it and just emits a plain text.

I can send a separate patch or modify this one?

A separate patch would be best.

If it helps, one possible set of functions to model this on is rcutorture_format_gp_seqs() on the "dev" branch of -rcu:

9357e5aecb63 ("rcutorture: Include grace-period sequence numbers in failure/close-call")

This has the needed #ifdefs and the different implementations for Tree and Tiny RCU.

Thanx, Paul

author maintainer
#23 lore raw

Re: [PATCH 2/4] torture: Remove CONFIG_NR_CPUS configuration

Quoted 74 lines · Paul E. McKenney
Quoted 63 lines · Uladzislau Rezki
Quoted 62 lines · Paul E. McKenney
Quoted 60 lines · Uladzislau Rezki
Quoted 49 lines
Quoted 49 lines
Quoted 2 lines

with 4 CPUs inside VM :)

And when running 16 instances with 4 CPUs each i can reproduce the splat which has been reported:

tools/testing/selftests/rcutorture/bin/kvm.sh --allcpus --configs \
'16*TREE05' --memory 10G --bootargs 'rcutorture.fwd_progress=1' \
--kconfig "CONFIG_NR_CPUS=4"

<snip>
...

[    0.595251] ------------[ cut here ]------------
[    0.595867] A full grace period is not passed yet: 0
[    0.595875] WARNING: CPU: 1 PID: 16 at kernel/rcu/tree.c:1617 rcu_sr_normal_complete+0xa9/0xc0
[    0.598248] Modules linked in:
[    0.598649] CPU: 1 UID: 0 PID: 16 Comm: rcu_preempt Not tainted 6.13.0-02530-g8950af6a11ff #261
[    0.599248] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[    0.600248] RIP: 0010:rcu_sr_normal_complete+0xa9/0xc0
[    0.600913] Code: 48 29 c2 48 8d 04 0a ba 03 00 00 00 48 39 c2 79 0c 48 83 e8 04 48 c1 e8 02 48 8d 70 02 48 c7 c7 20 e9 33 b5 e8 d8 03 f4 ff 90 <0f> 0b 90 90 48 8d 7b 10 5b e9 f9 38 fb ff 66 0f 1f 84 00 00 00 00
[    0.603249] RSP: 0018:ffffadad0008be60 EFLAGS: 00010282
[    0.603925] RAX: 0000000000000000 RBX: ffffadad00013d10 RCX: 00000000ffffdfff
[    0.605247] RDX: 0000000000000000 RSI: ffffadad0008bd10 RDI: 0000000000000001
[    0.606247] RBP: 0000000000000000 R08: 0000000000009ffb R09: 00000000ffffdfff
[    0.607248] R10: 00000000ffffdfff R11: ffffffffb56789a0 R12: 0000000000000005
[    0.608247] R13: 0000000000031a40 R14: fffffffffffffb74 R15: 0000000000000000
[    0.609250] FS:  0000000000000000(0000) GS:ffff9081f5c80000(0000) knlGS:0000000000000000
[    0.610249] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.611248] CR2: 0000000000000000 CR3: 00000002f024a000 CR4: 00000000000006f0
[    0.612249] Call Trace:
[    0.612574]  <TASK>
[    0.612854]  ? __warn+0x8c/0x190
[    0.613248]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.613840]  ? report_bug+0x164/0x190
[    0.614248]  ? handle_bug+0x54/0x90
[    0.614705]  ? exc_invalid_op+0x17/0x70
[    0.615248]  ? asm_exc_invalid_op+0x1a/0x20
[    0.615797]  ? rcu_sr_normal_complete+0xa9/0xc0
[    0.616248]  rcu_gp_cleanup+0x403/0x5a0
[    0.616248]  ? __pfx_rcu_gp_kthread+0x10/0x10
[    0.616818]  rcu_gp_kthread+0x136/0x1c0
[    0.617249]  kthread+0xec/0x1f0
[    0.617664]  ? __pfx_kthread+0x10/0x10
[    0.618156]  ret_from_fork+0x2f/0x50
[    0.618728]  ? __pfx_kthread+0x10/0x10
[    0.619216]  ret_from_fork_asm+0x1a/0x30
[    0.620251]  </TASK>

...
<snip>

Quoted 5 lines
Quoted 1 line

Linus tip-tree, HEAD is c4b9570cfb63501638db720f3bee9f6dfd044b82

Very good! And of course, the next question is "does going to _full() make the problem go away?" ;-)

Yes does its job if i apply:

https://lore.kernel.org/rcu/00900afe-ac4e-4362-a3f9-d65f2c9dcd9a@paulmck-laptop/T/#m5d9263f3825d3170c044beedbae741717702d4aa

after that i am not able to reproduce the warning anymore. Tested over night. Without it, i can reproduce it pretty easy :)

Quoted 7 lines
Quoted 4 lines

Thank you, and good to hear!!!

May I add your Tested-by to that patch?

Sure.

Thank you! I will apply this on my next rebase.

Quoted 8 lines
Quoted 5 lines

One question, we discussed that it is worth to print seq-delta in case of warning. Whereas a newly patch does do it and just emits a plain text.

I can send a separate patch or modify this one?

A separate patch would be best.

Sounds good :)

Quoted 8 lines

If it helps, one possible set of functions to model this on is rcutorture_format_gp_seqs() on the "dev" branch of -rcu:

9357e5aecb63 ("rcutorture: Include grace-period sequence numbers in failure/close-call")

This has the needed #ifdefs and the different implementations for Tree and Tiny RCU.

I will have a look.

--
Uladzislau Rezki

author maintainer
#24 lore raw

PATCH 3/4 rcu: Update TREE05.boot to test normal synchronize_rcu()

In reply to Uladzislau Rezki (Sony)

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
maintainer Reviewed-by on v1 3/4
#25 lore raw

Re: [PATCH 3/4] rcu: Update TREE05.boot to test normal synchronize_rcu()

Quoted 11 lines · Uladzislau Rezki (Sony)

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.

This one looks fine to me.

Thanx, Paul

Quoted 21 lines
---
 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
#26 lore raw

PATCH 4/4 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 | 4 ++++
 kernel/rcu/tree.c             | 8 +++-----
 2 files changed, 7 insertions(+), 5 deletions(-)
include/linux/rcupdate_wait.h +4 -0
index f9bed3d3f78d..a16fc2a9a7d7 100644
@@ -16,6 +16,10 @@
 struct rcu_synchronize {
 	struct rcu_head head;
 	struct completion completion;
+#ifdef CONFIG_PROVE_RCU
+	/* This is for testing. */
+	struct rcu_gp_oldstate oldstate;
+#endif
 };
 void wakeme_after_rcu(struct rcu_head *head);
 
kernel/rcu/tree.c +3 -5
index 2795d6b5109c..0ae90089ef09 100644
@@ -1612,12 +1612,10 @@ 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);
@@ -3214,7 +3212,7 @@ 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
#27 lore raw

Re: [PATCH 4/4] rcu: Use _full() API to debug synchronize_rcu()

Quoted 34 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.

---
 include/linux/rcupdate_wait.h | 4 ++++
 kernel/rcu/tree.c             | 8 +++-----
 2 files changed, 7 insertions(+), 5 deletions(-)
include/linux/rcupdate_wait.h +4 -0
index f9bed3d3f78d..a16fc2a9a7d7 100644
@@ -16,6 +16,10 @@
 struct rcu_synchronize {
 	struct rcu_head head;
 	struct completion completion;
+#ifdef CONFIG_PROVE_RCU
+	/* This is for testing. */
+	struct rcu_gp_oldstate oldstate;
+#endif
 };
 void wakeme_after_rcu(struct rcu_head *head);
 
kernel/rcu/tree.c +0 -0
index 2795d6b5109c..0ae90089ef09 100644
@@ -1612,12 +1612,10 @@ static void rcu_sr_normal_complete(struct llist_node *node)
 {

kernel/rcu/tree.c static void rcu_sr_normal_complete(struct llist_node *node) · line 1618

Quoted 10 lines
kernel/rcu/tree.c +2 -4
 	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");

Looks good, but why not also continue printing out the required grace-period sequence number? Yes, there would need to be helper sprintf()-style functions to paper over the difference between Tiny RCU and Tree RCU. ;-)

Thanx, Paul

Quoted 14 lines
kernel/rcu/tree.c +1 -1
 	/* Finally. */
 	complete(&rs->completion);
@@ -3214,7 +3212,7 @@ 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
#28 lore raw

Re: [PATCH 4/4] rcu: Use _full() API to debug synchronize_rcu()

Quoted 39 lines · Paul E. McKenney
Quoted 38 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.

---
 include/linux/rcupdate_wait.h | 4 ++++
 kernel/rcu/tree.c             | 8 +++-----
 2 files changed, 7 insertions(+), 5 deletions(-)
include/linux/rcupdate_wait.h +4 -0
index f9bed3d3f78d..a16fc2a9a7d7 100644
@@ -16,6 +16,10 @@
 struct rcu_synchronize {
 	struct rcu_head head;
 	struct completion completion;
+#ifdef CONFIG_PROVE_RCU
+	/* This is for testing. */
+	struct rcu_gp_oldstate oldstate;
+#endif
 };
 void wakeme_after_rcu(struct rcu_head *head);
 
kernel/rcu/tree.c +0 -1
index 2795d6b5109c..0ae90089ef09 100644
@@ -1612,12 +1612,10 @@ 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;
 
Quoted 12 lines

kernel/rcu/tree.c static void rcu_sr_normal_complete(struct llist_node *node) · line 1618

Quoted 6 lines
kernel/rcu/tree.c +2 -3
 	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");

Looks good, but why not also continue printing out the required grace-period sequence number? Yes, there would need to be helper sprintf()-style functions to paper over the difference between Tiny RCU and Tree RCU. ;-)

Uhh :) Do we have rcu_seq_diff() for a _full() API? Looks like not :)

It contains both, rgos_norm and rgos_exp! Take a delta of both?

--
Uladzislau Rezki

maintainer
#29 lore raw

Re: [PATCH 4/4] rcu: Use _full() API to debug synchronize_rcu()

Quoted 52 lines · Uladzislau Rezki
Quoted 39 lines · Paul E. McKenney
Quoted 38 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.

---
 include/linux/rcupdate_wait.h | 4 ++++
 kernel/rcu/tree.c             | 8 +++-----
 2 files changed, 7 insertions(+), 5 deletions(-)
include/linux/rcupdate_wait.h +4 -0
index f9bed3d3f78d..a16fc2a9a7d7 100644
@@ -16,6 +16,10 @@
 struct rcu_synchronize {
 	struct rcu_head head;
 	struct completion completion;
+#ifdef CONFIG_PROVE_RCU
+	/* This is for testing. */
+	struct rcu_gp_oldstate oldstate;
+#endif
 };
 void wakeme_after_rcu(struct rcu_head *head);
 
kernel/rcu/tree.c +0 -1
index 2795d6b5109c..0ae90089ef09 100644
@@ -1612,12 +1612,10 @@ 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;
 
Quoted 12 lines

kernel/rcu/tree.c static void rcu_sr_normal_complete(struct llist_node *node) · line 1618

Quoted 6 lines
kernel/rcu/tree.c +2 -3
 	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");

Looks good, but why not also continue printing out the required grace-period sequence number? Yes, there would need to be helper sprintf()-style functions to paper over the difference between Tiny RCU and Tree RCU. ;-)

Quoted 3 lines

Uhh :) Do we have rcu_seq_diff() for a _full() API? Looks like not :)

It contains both, rgos_norm and rgos_exp! Take a delta of both?

Why not? Maybe separate the two differences with a colon.

Or maybe make a variant of poll_state_synchronize_rcu_full() that take a char* argument, which uses the same value for the check and the string to be output.

Thanx, Paul

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

Hello, Paul!

Quoted 30 lines

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_possible_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 4 ++++
 1 file changed, 4 insertions(+)
kernel/rcu/rcutorture.c +4 -0
index d26fb1d33ed9..6bc161e1e8ac 100644
@@ -4050,6 +4050,10 @@ rcu_torture_init(void)
 					  writer_task);
 	if (torture_init_error(firsterr))
 		goto unwind;
+
+	if (nfakewriters < 0)
+		nfakewriters = (int) num_possible_cpus();
+
 	if (nfakewriters > 0) {
 		fakewriter_tasks = kcalloc(nfakewriters,
 					   sizeof(fakewriter_tasks[0]),
Signature · 1 line
2.39.5

Don't you mind to take this as well? It is needed for:

rcu: Update TREE05.boot to test normal synchronize_rcu()

--
Uladzislau Rezki

maintainer
#31 lore raw
Quoted 35 lines · Uladzislau Rezki

Hello, Paul!

Quoted 30 lines

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_possible_cpus() during torture initialization.

---
 kernel/rcu/rcutorture.c | 4 ++++
 1 file changed, 4 insertions(+)
kernel/rcu/rcutorture.c +4 -0
index d26fb1d33ed9..6bc161e1e8ac 100644
@@ -4050,6 +4050,10 @@ rcu_torture_init(void)
 					  writer_task);
 	if (torture_init_error(firsterr))
 		goto unwind;
+
+	if (nfakewriters < 0)
+		nfakewriters = (int) num_possible_cpus();
+
 	if (nfakewriters > 0) {
 		fakewriter_tasks = kcalloc(nfakewriters,
 					   sizeof(fakewriter_tasks[0]),
Signature · 1 line
2.39.5

Don't you mind to take this as well? It is needed for:

Quoted 1 line

rcu: Update TREE05.boot to test normal synchronize_rcu()

I would, but could you please set something up like we have for nreaders (the module parameter) and nrealreaders (the value actually used throughout). I freely admit that nrealfakereaders sounds a bit strange, so please feel free to either embrace the strangeness or propose an alternative. ;-)

The reason for this is so that, on a system with 128 CPUs, the user can distinguish between having specified (say) nfakewriters=128 on the one hand or nfakewriters=-1 on the other.

Thanx, Paul