rcu: remove unnecessary check cpu_no_qs.norm on rcu_report_qs_rdp
PATCH 1/1 rcu: remove unnecessary check cpu_no_qs.norm on rcu_report_qs_rdp
Wherever rcu_report_qs_rdp is called, cpu_no_qs.norm value is false. Therefore, Remove unnecessary check in rcu_report_qs_rdp.
--- kernel/rcu/tree.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
kernel/rcu/tree.c +1 -2
@@ -1962,8 +1962,7 @@ rcu_report_qs_rdp(struct rcu_data *rdp)WARN_ON_ONCE (rdp->cpu !=smp_processor_id ()); rnp = rdp->mynode;raw_spin_lock_irqsave_rcu_node (rnp, flags); -if (rdp->cpu_no_qs.b.norm || rdp->gp_seq != rnp->gp_seq || - rdp->gpwrap) { +if (rdp->gp_seq != rnp->gp_seq || rdp->gpwrap) {/* * The grace period in which this quiescent state was
--
2.37.2
Paul E. McKenney
maintainerQuoted 4 lines · Levi Yun
Wherever
rcu_report_qs_rdpis called, cpu_no_qs.norm value is false. Therefore, Remove unnecessary check inrcu_report_qs_rdp.Signed-off-by Levi Yun <[email protected]>
Why not start with something like this?
if (!WARN_ON_ONCE(!rdp->cpu_no_qs.b.norm) ||
rdp->gp_seq != rnp->gp_seq || rdp->gpwrap) {Except that rcu_report_qs_rdp() is invoked with interrupts enabled, which means that there is some possibility of state changes up to the raw_spin_lock_irqsave_rcu_node(rnp, flags) statement.
So, did you check whether RCU's interrupt paths change this state?
Thanx, Paul
Quoted 20 lines
--- kernel/rcu/tree.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)kernel/rcu/tree.c +1
-3@@ -1962,8 +1962,7 @@ rcu_report_qs_rdp(struct rcu_data *rdp)WARN_ON_ONCE (rdp->cpu !=smp_processor_id ()); rnp = rdp->mynode;raw_spin_lock_irqsave_rcu_node (rnp, flags); -if (rdp->cpu_no_qs.b.norm || rdp->gp_seq != rnp->gp_seq || - rdp->gpwrap) { +if (rdp->gp_seq != rnp->gp_seq || rdp->gpwrap) {/* * The grace period in which this quiescent state was -- 2.37.2
Hi, Paul.
Thanks for looking into this :)
Quoted 5 lines
Except that
rcu_report_qs_rdp()is invoked with interrupts enabled, which means that there is some possibility of state changes up to theraw_spin_lock_irqsave_rcu_node(rnp, flags)statement.So, did you check whether RCU's interrupt paths change this state?
In my narrow view,
only a new gp started, cpu_no_qs.b.norm changes as true in the path of rcu_sched_clock_irq.
But in that case, rcu_report_qs_rdp isn't called.
Did I understand your question well and are there any missed paths I didn't see?
Quoted 5 lines
Why not start with something like this?
if (!WARN_ON_ONCE(!rdp->cpu_no_qs.b.norm) || rdp->gp_seq != rnp->gp_seq || rdp->gpwrap) {
Yes. but with different message
Paul E. McKenney
maintainerQuoted 9 lines · Yun Levi
Hi, Paul.
Thanks for looking into this :)
Quoted 4 lines
Except that
rcu_report_qs_rdp()is invoked with interrupts enabled, which means that there is some possibility of state changes up to theraw_spin_lock_irqsave_rcu_node(rnp, flags)statement.
Quoted 8 lines
Quoted 1 line
So, did you check whether RCU's interrupt paths change this state?
In my narrow view,
only a new gp started, cpu_no_qs.b.norm changes as true in the path ofrcu_sched_clock_irq.
But in that case,rcu_report_qs_rdpisn't called.Did I understand your question well and are there any missed paths I didn't see?
Suppose that the scheduler-clock interrupt invoking rcu_sched_clock_irq() happened just before the lock was acquired in rcu_report_qs_rdp(). Suppose further that the RCU grace-period kthread started a new grace period just before that interrupt occurred. Then mightn't that interrupt notice the new grace period and set ->cpu_no_qs.b.norm to true before fully returning?
Thanx, Paul
Quoted 7 lines
Quoted 5 lines
Why not start with something like this?
if (!WARN_ON_ONCE(!rdp->cpu_no_qs.b.norm) || rdp->gp_seq != rnp->gp_seq || rdp->gpwrap) {Yes. but with different message
Hi Paul.
Quoted 6 lines
Suppose that the scheduler-clock interrupt invoking
rcu_sched_clock_irq()happened just before the lock was acquired inrcu_report_qs_rdp(). Suppose further that the RCU grace-period kthread started a new grace period just before that interrupt occurred. Then mightn't that interrupt notice the new grace period and set ->cpu_no_qs.b.norm to true before fully returning?
IIUC, RCU grace-period kthread couldn't start new grace period because the interrupted cpu don't report qs to rnp via rcu_report_qs_rdp. That situation is listened like new gp could be started thou all cpus doesn't enter yet.
That's is the reason why it's better to use WARN_ON_ONCE as you suggest to notice if the buggy situation happens
Am I missing something or wrong?
Thanks.
--------
Sincerely,
Levi.
Paul E. McKenney
maintainerQuoted 9 lines · Yun Levi
Hi Paul.
Quoted 6 lines
Suppose that the scheduler-clock interrupt invoking
rcu_sched_clock_irq()happened just before the lock was acquired inrcu_report_qs_rdp(). Suppose further that the RCU grace-period kthread started a new grace period just before that interrupt occurred. Then mightn't that interrupt notice the new grace period and set ->cpu_no_qs.b.norm to true before fully returning?
Quoted 6 lines
IIUC, RCU grace-period kthread couldn't start new grace period because the interrupted cpu don't report qs to rnp via
rcu_report_qs_rdp. That situation is listened like new gp could be started thou all cpus doesn't enter yet.
That's is the reason why it's better to useWARN_ON_ONCEas you suggest to notice if the buggy situation happens
And try testing with CONFIG_RCU_STRICT_GRACE_PERIOD=y and CONFIG_PREEMPT=n. Though there might be better Kconfig options to use. Those two come immediately to mind.
Quoted 1 line
Am I missing something or wrong?
I cannot see into your head, so I cannot say.
But one critical piece is that softirq handlers, including the RCU_SOFTIRQ handler rcu_core_si(), can be invoked upon return from interrupts. Another critical piece is that if a CPU is idle during any part of a grace period, the grace-period kthread can report a quiescent state on its behalf.
Does that help?
Thanx, Paul
Thanks for replying to reply Paul :)
Quoted 3 lines
And try testing with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=n. Though there might be betterKconfigoptions to use. Those two come immediately to mind.
I've tested with this option via rcu torture.
and it doesn't report any problems.
and after commit 6d60ea03ac2d3 ("rcu: Report QS for outermost PREEMPT=n rcu_read_unlock() for strict GPs")
it always makes cpu_no_qs.b.norm false whenever it calls rcu_report_qs_rdp in rcu_read_unlock.
Quoted 2 lines
But one critical piece is that softirq handlers, including the
RCU_SOFTIRQhandlerrcu_core_si(), can be invoked upon return from interrupts.
I think in that case, no problem because if it reports qs already, rcu_check_quiescent_state wouldn't call rcu_report_qs_rdp again. And if RCU_SOFTIRQ is called as soon as RCU interrupt is finished, it seems the fastest one to notify qs to related tree.
Quoted 3 lines
Another critical piece is that if a CPU is idle during any part of a grace period, the grace-period kthread can report a quiescent state on its behalf.
I think
1) If timer interrupt is still programed,
- when rcu_sched_clock_irq first reports qs, no problem
- If qs is reported via grace period thread first,note_gp_chagned in rcu interrupt
will recognize this situation by setting core_needs_qs as false, it doesn't call rcu_report_qs_rdp thou cpu_no_qs.b.norm is true.
2) If the timer interrupt isn't programmed,
- rcu_gp_kthreaad reports its qs, no problem.
So, I think there's no problem removing
"rdp->cpu_no_qs.b.norm" check in rcu_report_qs_rdp.
or wrap this condition check as WARN_ON_ONCE.
Quoted 1 line
Does that help?
Many thanks always :)
--------
SIncerely,
Levi.
Paul E. McKenney
maintainerQuoted 6 lines · Yun Levi
Thanks for replying to reply Paul :)
Quoted 3 lines
And try testing with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=n. Though there might be betterKconfigoptions to use. Those two come immediately to mind.
Quoted 6 lines
I've tested with this option via rcu torture.
and it doesn't report any problems.
and after commit6d60ea03ac2d3("rcu: Report QS for outermost PREEMPT=nrcu_read_unlock()for strict GPs")
it always makes cpu_no_qs.b.norm false whenever it callsrcu_report_qs_rdpinrcu_read_unlock.
Again, the concern is that an interrupt and softirq handler at that point changes the grace period. What else could you do to check that this sequence of events actually occurred? That is, that your testing didn't just get lucky and not actually hit this condition?
Interrupts are enabled at that point, so there is nothing that prevents that condition, after all.
Quoted 20 lines
Quoted 2 lines
But one critical piece is that softirq handlers, including the
RCU_SOFTIRQhandlerrcu_core_si(), can be invoked upon return from interrupts.I think in that case, no problem because if it reports qs already,
rcu_check_quiescent_statewouldn't callrcu_report_qs_rdpagain. And ifRCU_SOFTIRQis called as soon as RCU interrupt is finished, it seems the fastest one to notify qs to related tree.Quoted 3 lines
Another critical piece is that if a CPU is idle during any part of a grace period, the grace-period kthread can report a quiescent state on its behalf.
I think
1) If timer interrupt is still programed,
- whenrcu_sched_clock_irqfirst reports qs, no problem
- If qs is reported via grace period thread first,note_gp_chagnedin rcu interrupt
will recognize this situation by settingcore_needs_qsas false, it doesn't callrcu_report_qs_rdpthou cpu_no_qs.b.norm is true.
Quoted 6 lines
2) If the timer interrupt isn't programmed,
-rcu_gp_kthreaadreports its qs, no problem.So, I think there's no problem removing
"rdp->cpu_no_qs.b.norm" check in rcu_report_qs_rdp.or wrap this condition check as
WARN_ON_ONCE.
One additional concern is that a quiescent state detected in an old grace period will be incorrectly reported for a new grace period.
Thanx, Paul
Quoted 6 lines
Quoted 1 line
Does that help?
Many thanks always :)
--------
SIncerely,
Levi.
Quoted 37 lines
Thanks for replying to reply Paul :)
Quoted 3 lines
And try testing with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=n. Though there might be betterKconfigoptions to use. Those two come immediately to mind.I've tested with this option via rcu torture.
and it doesn't report any problems.
and after commit6d60ea03ac2d3("rcu: Report QS for outermost PREEMPT=nrcu_read_unlock()for strict GPs")
it always makes cpu_no_qs.b.norm false whenever it callsrcu_report_qs_rdpinrcu_read_unlock.Quoted 2 lines
But one critical piece is that softirq handlers, including the
RCU_SOFTIRQhandlerrcu_core_si(), can be invoked upon return from interrupts.I think in that case, no problem because if it reports qs already,
rcu_check_quiescent_statewouldn't callrcu_report_qs_rdpagain. And ifRCU_SOFTIRQis called as soon as RCU interrupt is finished, it seems the fastest one to notify qs to related tree.Quoted 3 lines
Another critical piece is that if a CPU is idle during any part of a grace period, the grace-period kthread can report a quiescent state on its behalf.
I think
1) If timer interrupt is still programed,
- whenrcu_sched_clock_irqfirst reports qs, no problem
- If qs is reported via grace period thread first,note_gp_chagnedin rcu interrupt
will recognize this situation by settingcore_needs_qsas false, it doesn't callrcu_report_qs_rdpthou cpu_no_qs.b.norm is true.2) If the timer interrupt isn't programmed,
-rcu_gp_kthreaadreports its qs, no problem.
Quoted 7 lines
So, I think there's no problem removing
"rdp->cpu_no_qs.b.norm" check in rcu_report_qs_rdp.or wrap this condition check as
WARN_ON_ONCE.Quoted 1 line
Does that help?
Many thanks always :)
Hi Levi
For built with CONFIG_RCU_STRICT_GRACE_PERIOD=y and CONFIG_PREEMPT=n kernels Consider the following scenario:
__rcu_read_unlock()
-> rcu_read_unlock_strict()
->rdp = this_cpu_ptr(&rcu_data);
->rdp->cpu_no_qs.b.norm = false;
by interrupt and return invoke rcu_core():
->rcu_check_quiescent_state()
->rdp = raw_cpu_ptr(&rcu_data);
-> rcu_check_quiescent_state(rdp);
->note_gp_changes(rdp);
-> __note_gp_changes(rnp, rdp)
start new gp
->rdp->cpu_no_qs.b.norm = true;
->rcu_report_qs_rdp(rdp);
->if (rdp->cpu_no_qs.b.norm || ...)Thanks
Zqiang
Quoted 4 lines
--------
SIncerely,
Levi.
Hi Z qiang!
Thanks for replying. But I'm pinned on something wrong..!
Quoted 5 lines
For built with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=nkernels Consider the following scenario:
__rcu_read_unlock()
->rcu_read_unlock_strict()
Quoted 14 lines
->rdp = this_cpu_ptr(&rcu_data);
->rdp->cpu_no_qs.b.norm = false;
by interrupt and return invoke rcu_core():
->rcu_check_quiescent_state()
->rdp = raw_cpu_ptr(&rcu_data);
-> rcu_check_quiescent_state(rdp);
->note_gp_changes(rdp);
-> __note_gp_changes(rnp, rdp)
start new gp
->rdp->cpu_no_qs.b.norm = true;
->rcu_report_qs_rdp(rdp);
->if (rdp->cpu_no_qs.b.norm || ...)I've already seen this scenario, But I think something is missing in my view. What I couldn't catch is
->rcu_check_quiescent_state()
->rdp = raw_cpu_ptr(&rcu_data);
-> rcu_check_quiescent_state(rdp);
->note_gp_changes(rdp);
-> __note_gp_changes(rnp, rdp)
start new gpthe new gp is started.
to set cpu_no_qs.b.norm as true, below condition should be true
1201 >---if (rcu_seq_new_gp(rdp->gp_seq, rnp->gp_seq) ||
1202 >--- unlikely(READ_ONCE(rdp->gpwrap))) {
Here,
How rcu_seq_new_gp could return true and new gp already started via rcu_gp_kthread.
IIUC, because rcu_gp_fqs_loop couldn't see the root rnp->qsmask is zero, it couldn't call rcu_gp_init.
Sorry to make noise, but would you correct me what I'm thinking wrong?
Many thanks..!
-----------
Sincerely,
Levi.
Quoted 78 lines · Z qiang
Quoted 38 lines
Thanks for replying to reply Paul :)
Quoted 3 lines
And try testing with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=n. Though there might be betterKconfigoptions to use. Those two come immediately to mind.I've tested with this option via rcu torture.
and it doesn't report any problems.
and after commit6d60ea03ac2d3("rcu: Report QS for outermost PREEMPT=nrcu_read_unlock()for strict GPs")
it always makes cpu_no_qs.b.norm false whenever it callsrcu_report_qs_rdpinrcu_read_unlock.Quoted 2 lines
But one critical piece is that softirq handlers, including the
RCU_SOFTIRQhandlerrcu_core_si(), can be invoked upon return from interrupts.I think in that case, no problem because if it reports qs already,
rcu_check_quiescent_statewouldn't callrcu_report_qs_rdpagain. And ifRCU_SOFTIRQis called as soon as RCU interrupt is finished, it seems the fastest one to notify qs to related tree.Quoted 3 lines
Another critical piece is that if a CPU is idle during any part of a grace period, the grace-period kthread can report a quiescent state on its behalf.
I think
1) If timer interrupt is still programed,
- whenrcu_sched_clock_irqfirst reports qs, no problem
- If qs is reported via grace period thread first,note_gp_chagnedin rcu interrupt
will recognize this situation by settingcore_needs_qsas false, it doesn't callrcu_report_qs_rdpthou cpu_no_qs.b.norm is true.2) If the timer interrupt isn't programmed,
-rcu_gp_kthreaadreports its qs, no problem.Quoted 7 lines
So, I think there's no problem removing
"rdp->cpu_no_qs.b.norm" check in rcu_report_qs_rdp.or wrap this condition check as
WARN_ON_ONCE.Quoted 1 line
Does that help?
Many thanks always :)
Hi Levi
For built with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=nkernels Consider the following scenario:
__rcu_read_unlock()
->rcu_read_unlock_strict()->rdp = this_cpu_ptr(&rcu_data); ->rdp->cpu_no_qs.b.norm = false; by interrupt and return invoke rcu_core(): ->rcu_check_quiescent_state() ->rdp = raw_cpu_ptr(&rcu_data); -> rcu_check_quiescent_state(rdp); ->note_gp_changes(rdp); -> __note_gp_changes(rnp, rdp) start new gp ->rdp->cpu_no_qs.b.norm = true; ->rcu_report_qs_rdp(rdp); ->if (rdp->cpu_no_qs.b.norm || ...)Thanks
ZqiangQuoted 4 lines
--------
SIncerely,
Levi.
Quoted 10 lines
Hi Z qiang!
Thanks for replying. But I'm pinned on something wrong..!
Quoted 5 lines
For built with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=nkernels Consider the following scenario:
__rcu_read_unlock()
->rcu_read_unlock_strict()
note that as soon as we exit the rcu read critical section the rcu_read_unlock_strict() will be called, even if the RCU is idle at this time
Quoted 18 lines
Quoted 6 lines
->rdp = this_cpu_ptr(&rcu_data); ->rdp->cpu_no_qs.b.norm = false; by interrupt and return invoke rcu_core(): ->rcu_check_quiescent_state() ->rdp = raw_cpu_ptr(&rcu_data);Quoted 8 lines
-> rcu_check_quiescent_state(rdp); ->note_gp_changes(rdp); -> __note_gp_changes(rnp, rdp) start new gp ->rdp->cpu_no_qs.b.norm = true; ->rcu_report_qs_rdp(rdp); ->if (rdp->cpu_no_qs.b.norm || ...)I've already seen this scenario, But I think something is missing in my view. What I couldn't catch is
Quoted 10 lines
->rcu_check_quiescent_state() ->rdp = raw_cpu_ptr(&rcu_data); -> rcu_check_quiescent_state(rdp); ->note_gp_changes(rdp); -> __note_gp_changes(rnp, rdp) start new gpthe new gp is started.
Maybe this "start new gp" note misunderstood you.
For built with CONFIG_RCU_STRICT_GRACE_PERIOD=y and CONFIG_PREEMPT=n kernels, if the gp kthread start a new GP before we exit the RCU read critical section, and just before we call rcu_report_qs_rdp() in rcu_read_unlock_strict(), at this time if the clock irq
happens and find that the "rcu_seq_current(&rnp->gp_seq) != rdp->gp_seq" is true in rcu_pening(),
will trigger RCU softirq and find that the rcu_seq_new_gp(rdp->gp_seq, rnp->gp_seq) is true,
will set rdp->cpu_no_qs.b.norm is true. when we return from the softirq and call rcu_report_qs_rdp()
in rcu_read_unlock_strict(), find that the rdp->cpu_no_qs.b.norm is true. so there is a situation where the rdp->cpu_no_qs.b.norm is true.
Quoted 100 lines
to set cpu_no_qs.b.norm as true, below condition should be true
1201 >---if (
rcu_seq_new_gp(rdp->gp_seq, rnp->gp_seq)||
1202 >--- unlikely(READ_ONCE(rdp->gpwrap))) {Here,
Howrcu_seq_new_gpcould return true and new gp already started viarcu_gp_kthread.
IIUC, becausercu_gp_fqs_loopcouldn't see the rootrnp->qsmaskis zero, it couldn't callrcu_gp_init.Sorry to make noise, but would you correct me what I'm thinking wrong?
Many thanks..!
-----------
Sincerely,
Levi.Quoted 78 lines · Z qiang
Quoted 38 lines
Thanks for replying to reply Paul :)
Quoted 3 lines
And try testing with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=n. Though there might be betterKconfigoptions to use. Those two come immediately to mind.I've tested with this option via rcu torture.
and it doesn't report any problems.
and after commit6d60ea03ac2d3("rcu: Report QS for outermost PREEMPT=nrcu_read_unlock()for strict GPs")
it always makes cpu_no_qs.b.norm false whenever it callsrcu_report_qs_rdpinrcu_read_unlock.Quoted 2 lines
But one critical piece is that softirq handlers, including the
RCU_SOFTIRQhandlerrcu_core_si(), can be invoked upon return from interrupts.I think in that case, no problem because if it reports qs already,
rcu_check_quiescent_statewouldn't callrcu_report_qs_rdpagain. And ifRCU_SOFTIRQis called as soon as RCU interrupt is finished, it seems the fastest one to notify qs to related tree.Quoted 3 lines
Another critical piece is that if a CPU is idle during any part of a grace period, the grace-period kthread can report a quiescent state on its behalf.
I think
1) If timer interrupt is still programed,
- whenrcu_sched_clock_irqfirst reports qs, no problem
- If qs is reported via grace period thread first,note_gp_chagnedin rcu interrupt
will recognize this situation by settingcore_needs_qsas false, it doesn't callrcu_report_qs_rdpthou cpu_no_qs.b.norm is true.2) If the timer interrupt isn't programmed,
-rcu_gp_kthreaadreports its qs, no problem.Quoted 7 lines
So, I think there's no problem removing
"rdp->cpu_no_qs.b.norm" check in rcu_report_qs_rdp.or wrap this condition check as
WARN_ON_ONCE.Quoted 1 line
Does that help?
Many thanks always :)
Hi Levi
For built with
CONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=nkernels Consider the following scenario:
__rcu_read_unlock()
->rcu_read_unlock_strict()->rdp = this_cpu_ptr(&rcu_data); ->rdp->cpu_no_qs.b.norm = false; by interrupt and return invoke rcu_core(): ->rcu_check_quiescent_state() ->rdp = raw_cpu_ptr(&rcu_data); -> rcu_check_quiescent_state(rdp); ->note_gp_changes(rdp); -> __note_gp_changes(rnp, rdp) start new gp ->rdp->cpu_no_qs.b.norm = true; ->rcu_report_qs_rdp(rdp); ->if (rdp->cpu_no_qs.b.norm || ...)Thanks
ZqiangQuoted 4 lines
--------
SIncerely,
Levi.
Hi, Z qiang.
Quoted 5 lines
Maybe this "start new gp" note misunderstood you.
For built withCONFIG_RCU_STRICT_GRACE_PERIOD=yandCONFIG_PREEMPT=nkernels, if the gp kthread start a new GP before we exit the RCU read critical section, and just before we callrcu_report_qs_rdp()inrcu_read_unlock_strict(), at this time if the clock irq
Quoted 8 lines
happens and find that the "
rcu_seq_current(&rnp->gp_seq)!=rdp->gp_seq" is true inrcu_pening(),
will trigger RCU softirq and find that thercu_seq_new_gp(rdp->gp_seq, rnp->gp_seq)is true,
will setrdp->cpu_no_qs.b.normis true. when we return from the softirq and callrcu_report_qs_rdp()
inrcu_read_unlock_strict(), find that therdp->cpu_no_qs.b.normis true. so there is a situation where therdp->cpu_no_qs.b.normis true.
Thanks for making me out from my misunderstanding..!
and even the not in STRICT_GRACE_PERIOD,
old grace period's qs could be reported though rcu_gp_kthread start new gp by first rcu_report_qs_rdp, will be clear soon by continuous rcu_report_qs_rdp which is triggered by interrupting before the first rcu_report_qs_rdp.
Many thanks..!
And sorry to make noise...!