rcu: call_rcu() power improvements
Subsystems (8)
v8 cover rcu: call_rcu() power improvements
v8 version of RCU lazy patches based on rcu/next branch. Very small mostly cosmetic changes since v7, the one exception being the rxrpc patch.
I will post the tracing patches later along with rcutop as I need to add new tracepoints that Frederic suggested.
Main recent changes:
1. rcu_barrier() wake up only for lazy bypass list.
2. Make all call_rcu() default-lazy and add call_rcu_flush() API.
3. Take care of some callers using call_rcu_flush() API.
4. Several refactorings suggested by Paul/Frederic.
5. New call_rcu() to call_rcu_flush() conversions by Joel/Vlad/Paul.
6. New scripts in cover-letter to check for callbacks doing wake-ups.
I am seeing good performance and power with these patches on real ChromeOS x86 asymmetric hardware.
Earlier cover letter with lots of details is here:
https://lore.kernel.org/all/[email protected]/
List of recent changes:
[ Frederic Weisbec: Program the lazy timer only if WAKE_NOT, since other
deferral levels wake much earlier so for those it is not needed. ]
[ Frederic Weisbec: Use flush flags to keep bypass API code clean. ]
[ Frederic Weisbec: Make rcu_barrier() wake up only if main list empty. ]
[ Frederic Weisbec: Remove extra 'else if' branch in rcu_nocb_try_bypass(). ]
[ Joel: Fix issue where I was not resetting lazy_len after moving it to rdp ]
[ Paul/Thomas/Joel: Make call_rcu() default lazy so users don't mess up. ]
[ Paul/Frederic : Cosmetic changes, split out wakeup of nocb thread. ]
[ Vlad/Joel : More call_rcu -> flush conversions ]
[ debug code for detecting "wake" in kernel's call_rcu() callbacks. ]
The following 2 scripts can be used to check if any callbacks in the kernel are doing a wake up (it is best effort and may miss some things, but we found issues using it)
1. Script to search for call_rcu() references and dump the callback list to a file:
#!/bin/bash
rm func-list
touch func-list
for f in $(find . \( -name "*.c" -o -name "*.h" \) | grep -v rcu); do
funcs=$(perl -0777 -ne 'while(m/call_rcu\([&]?.+,\s?(.+)\).*;/g){print "$1\n";}' $f)
if [ "x$funcs" != "x" ]; then
for func in $funcs; do
echo "$f $func" >> func-list
echo "$f $func"
done
fidone
cat func-list | sort | uniq | tee func-list-sorted
2. Script to search "wake" after callback references:
#!/bin/bash
while read fl; do
file=$(echo $fl | cut -d " " -f1) func=$(echo $fl | cut -d " " -f2) grep -A 30 $func $file | grep wake > /dev/null if [ $? -eq 0 ]; then echo "keyword wake found after function reference $func in $file" echo "Output:" grep -A 30 $func $file echo "===========================================================" fi
done < func-list-sorted
Frederic Weisbecker (1):
rcu: Fix missing nocb gp wake on rcu_barrier()
Joel Fernandes (Google) (9):
rcu: Make call_rcu() lazy to save power
rcu: Refactor code a bit in rcu_nocb_do_flush_bypass()
rcuscale: Add laziness and kfree tests
percpu-refcount: Use call_rcu_flush() for atomic switch rcu/sync: Use call_rcu_flush() instead of call_rcu rcu/rcuscale: Use call_rcu_flush() for async reader test rcu/rcutorture: Use call_rcu_flush() where needed
rxrpc: Use call_rcu_flush() instead of call_rcu() rcu/debug: Add wake-up debugging for lazy callbacks
Uladzislau Rezki (2):
scsi/scsi_error: Use call_rcu_flush() instead of call_rcu()
workqueue: Make queue_rcu_work() use call_rcu_flush()
Vineeth Pillai (1):
rcu: shrinker for lazy rcu
drivers/scsi/scsi_error.c | 2 +- include/linux/rcupdate.h | 7 ++ kernel/rcu/Kconfig | 15 +++ kernel/rcu/lazy-debug.h | 154 +++++++++++++++++++++++++++ kernel/rcu/rcu.h | 8 ++ kernel/rcu/rcuscale.c | 70 ++++++++++++- kernel/rcu/rcutorture.c | 16 +-- kernel/rcu/sync.c | 2 +- kernel/rcu/tiny.c | 2 +- kernel/rcu/tree.c | 144 +++++++++++++++++-------- kernel/rcu/tree.h | 12 ++- kernel/rcu/tree_exp.h | 2 +- kernel/rcu/tree_nocb.h | 215 +++++++++++++++++++++++++++++++++----- kernel/workqueue.c | 2 +- lib/percpu-refcount.c | 3 +- net/rxrpc/conn_object.c | 2 +- 16 files changed, 559 insertions(+), 97 deletions(-)
create mode 100644 kernel/rcu/lazy-debug.h
--
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 1/13 rcu: Fix missing nocb gp wake on rcu_barrier()
From: Frederic Weisbecker <[email protected]>
Upon entraining a callback to a NOCB CPU, no further wake up is issued on the corresponding nocb_gp kthread. As a result, the callback and all the subsequent ones on that CPU may be ignored, at least until an RCU_NOCB_WAKE_FORCE timer is ever armed or another NOCB CPU belonging to the same group enqueues a callback on an empty queue.
Here is a possible bad scenario:
1) CPU 0 is NOCB unlike all other CPUs.
2) CPU 0 queues a callback
2) The grace period related to that callback elapses
3) The callback is moved to the done list (but is not invoked yet), there are no more pending callbacks for CPU 0
4) CPU 1 calls rcu_barrier() and sends an IPI to CPU 0
5) CPU 0 entrains the callback but doesn't wake up nocb_gp
6) CPU 1 blocks forever, unless CPU 0 ever queues enough further callbacks to arm an RCU_NOCB_WAKE_FORCE timer.
Make sure the necessary wake up is produced whenever necessary.
This is also required to make sure lazy callbacks in future patches don't end up making rcu_barrier() wait for multiple seconds.
--- kernel/rcu/tree.c | 6 ++++++ kernel/rcu/tree.h | 1 + kernel/rcu/tree_nocb.h | 5 +++++ 3 files changed, 12 insertions(+)
kernel/rcu/tree.c +6 -0
@@ -3894,6 +3894,8 @@ static void rcu_barrier_entrain(struct rcu_data *rdp) {unsigned long gseq =READ_ONCE (rcu_state.barrier_sequence);unsigned long lseq =READ_ONCE (rdp->barrier_seq_snap); +bool wake_nocb = false; +bool was_alldone = false;lockdep_assert_held (&rcu_state.barrier_lock);if (rcu_seq_state (lseq) || !rcu_seq_state (gseq) ||rcu_seq_ctr (lseq) !=rcu_seq_ctr (gseq))
@@ -3902,6 +3904,7 @@ static void rcu_barrier_entrain(struct rcu_data *rdp) rdp->barrier_head.func = rcu_barrier_callback;debug_rcu_head_queue (&rdp->barrier_head);rcu_nocb_lock (rdp); + was_alldone =rcu_rdp_is_offloaded (rdp) && !rcu_segcblist_pend_cbs (&rdp->cblist);WARN_ON_ONCE (!rcu_nocb_flush_bypass (rdp,NULL , jiffies));if (rcu_segcblist_entrain (&rdp->cblist, &rdp->barrier_head)) {atomic_inc (&rcu_state.barrier_cpu_count);
@@ -3909,7 +3912,10 @@ static void rcu_barrier_entrain(struct rcu_data *rdp)debug_rcu_head_unqueue (&rdp->barrier_head);rcu_barrier_trace (TPS ("IRQNQ" ),-1 , rcu_state.barrier_sequence); } + wake_nocb = was_alldone &&rcu_segcblist_pend_cbs (&rdp->cblist);rcu_nocb_unlock (rdp); +if (wake_nocb) +wake_nocb_gp (rdp, false);smp_store_release (&rdp->barrier_seq_snap, gseq); }
kernel/rcu/tree.h +1 -0
@@ -439,6 +439,7 @@ static void zero_cpu_stall_ticks(struct rcu_data *rdp);static struct swait_queue_head *rcu_nocb_gp_get (struct rcu_node *rnp);static void rcu_nocb_gp_cleanup (struct swait_queue_head *sq);static void rcu_init_one_nocb (struct rcu_node *rnp); +static bool wake_nocb_gp (struct rcu_data *rdp,bool force);static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp,unsigned long j); static bool rcu_nocb_try_bypass(struct rcu_data *rdp, struct rcu_head *rhp,
kernel/rcu/tree_nocb.h +5 -0
@@ -1558,6 +1558,11 @@ static void rcu_init_one_nocb(struct rcu_node *rnp) { } +static bool wake_nocb_gp (struct rcu_data *rdp,bool force) +{ +return false; +} +static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp,unsigned long j) {
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
Paul E. McKenney
maintainerRe: [PATCH v8 01/13] rcu: Fix missing nocb gp wake on rcu_barrier()
Quoted 8 lines · Joel Fernandes (Google)
From: Frederic Weisbecker <[email protected]>
Upon entraining a callback to a NOCB CPU, no further wake up is issued on the corresponding
nocb_gpkthread. As a result, the callback and all the subsequent ones on that CPU may be ignored, at least until anRCU_NOCB_WAKE_FORCEtimer is ever armed or another NOCB CPU belonging to the same group enqueues a callback on an empty queue.
Quoted 4 lines
Here is a possible bad scenario:
1) CPU 0 is NOCB unlike all other CPUs.
2) CPU 0 queues a callback
Call it CB1.
Quoted 3 lines
2) The grace period related to that callback elapses
3) The callback is moved to the done list (but is not invoked yet), there are no more pending callbacks for CPU 0
So CB1 is on ->cblist waiting to be invoked, correct?
Quoted 2 lines
4) CPU 1 calls
rcu_barrier()and sends an IPI to CPU 0
5) CPU 0 entrains the callback but doesn't wake upnocb_gp
And CB1 must still be there because otherwise the IPI handler would not have entrained the callback, correct? If so, we have both CB1 and the rcu_barrier() callback (call it CB2) in ->cblist, but on the done list.
Quoted 2 lines
6) CPU 1 blocks forever, unless CPU 0 ever queues enough further callbacks to arm an
RCU_NOCB_WAKE_FORCEtimer.
Except that -something- must have already been prepared to wake up in order to invoke CB1. And that something would invoke CB2 along with CB1, given that they are both on the done list. If there is no such wakeup already, then the hang could occur with just CB1, without the help of CB2.
Quoted 1 line
Make sure the necessary wake up is produced whenever necessary.
I am not seeing that the wakeup is needed in this case.
So what am I missing here?
Quoted 2 lines
This is also required to make sure lazy callbacks in future patches don't end up making
rcu_barrier()wait for multiple seconds.
But I do see that the wakeup is needed in the lazy case, and if I remember correctly, the ten-second rcu_barrier() delay really did happen. If I understand correctly, for this to happen, all of the callbacks must be in the bypass list, that is, ->cblist must be empty.
So has the scenario steps 1-6 called out above actually happened in the absence of lazy callbacks?
Thanx, Paul
Quoted 73 lines
Reported-by Joel Fernandes (Google) <[email protected]>Fixes 5d6742b37727 ("rcu/nocb: Use rcu_segcblist for no-CBs CPUs")Signed-off-by Frederic Weisbecker <[email protected]>Signed-off-by Joel Fernandes (Google) <[email protected]>--- kernel/rcu/tree.c | 6 ++++++ kernel/rcu/tree.h | 1 + kernel/rcu/tree_nocb.h | 5 +++++ 3 files changed, 12 insertions(+)kernel/rcu/tree.c +6
-0@@ -3894,6 +3894,8 @@ static void rcu_barrier_entrain(struct rcu_data *rdp) {unsigned long gseq =READ_ONCE (rcu_state.barrier_sequence);unsigned long lseq =READ_ONCE (rdp->barrier_seq_snap); +bool wake_nocb = false; +bool was_alldone = false;lockdep_assert_held (&rcu_state.barrier_lock);if (rcu_seq_state (lseq) || !rcu_seq_state (gseq) ||rcu_seq_ctr (lseq) !=rcu_seq_ctr (gseq))@@ -3902,6 +3904,7 @@ static void rcu_barrier_entrain(struct rcu_data *rdp) rdp->barrier_head.func = rcu_barrier_callback;debug_rcu_head_queue (&rdp->barrier_head);rcu_nocb_lock (rdp); + was_alldone =rcu_rdp_is_offloaded (rdp) && !rcu_segcblist_pend_cbs (&rdp->cblist);WARN_ON_ONCE (!rcu_nocb_flush_bypass (rdp,NULL , jiffies));if (rcu_segcblist_entrain (&rdp->cblist, &rdp->barrier_head)) {atomic_inc (&rcu_state.barrier_cpu_count);@@ -3909,7 +3912,10 @@ static void rcu_barrier_entrain(struct rcu_data *rdp)debug_rcu_head_unqueue (&rdp->barrier_head);rcu_barrier_trace (TPS ("IRQNQ" ),-1 , rcu_state.barrier_sequence); } + wake_nocb = was_alldone &&rcu_segcblist_pend_cbs (&rdp->cblist);rcu_nocb_unlock (rdp); +if (wake_nocb) +wake_nocb_gp (rdp, false);smp_store_release (&rdp->barrier_seq_snap, gseq); }kernel/rcu/tree.h +1
-0@@ -439,6 +439,7 @@ static void zero_cpu_stall_ticks(struct rcu_data *rdp);static struct swait_queue_head *rcu_nocb_gp_get (struct rcu_node *rnp);static void rcu_nocb_gp_cleanup (struct swait_queue_head *sq);static void rcu_init_one_nocb (struct rcu_node *rnp); +static bool wake_nocb_gp (struct rcu_data *rdp,bool force);static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp,unsigned long j); static bool rcu_nocb_try_bypass(struct rcu_data *rdp, struct rcu_head *rhp,kernel/rcu/tree_nocb.h +5
-0@@ -1558,6 +1558,11 @@ static void rcu_init_one_nocb(struct rcu_node *rnp) { } +static bool wake_nocb_gp (struct rcu_data *rdp,bool force) +{ +return false; +} +static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp,unsigned long j) {Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
Frederic Weisbecker
maintainerRe: [PATCH v8 01/13] rcu: Fix missing nocb gp wake on rcu_barrier()
Quoted 29 lines · Paul E. McKenney
Quoted 8 lines · Joel Fernandes (Google)
From: Frederic Weisbecker <[email protected]>
Upon entraining a callback to a NOCB CPU, no further wake up is issued on the corresponding
nocb_gpkthread. As a result, the callback and all the subsequent ones on that CPU may be ignored, at least until anRCU_NOCB_WAKE_FORCEtimer is ever armed or another NOCB CPU belonging to the same group enqueues a callback on an empty queue.Quoted 4 lines
Here is a possible bad scenario:
1) CPU 0 is NOCB unlike all other CPUs.
2) CPU 0 queues a callbackCall it CB1.
Quoted 3 lines
2) The grace period related to that callback elapses
3) The callback is moved to the done list (but is not invoked yet), there are no more pending callbacks for CPU 0So CB1 is on ->cblist waiting to be invoked, correct?
Quoted 2 lines
4) CPU 1 calls
rcu_barrier()and sends an IPI to CPU 0
5) CPU 0 entrains the callback but doesn't wake upnocb_gpAnd CB1 must still be there because otherwise the IPI handler would not have entrained the callback, correct? If so, we have both CB1 and the
rcu_barrier()callback (call it CB2) in ->cblist, but on the done list.
Quoted 7 lines
Quoted 2 lines
6) CPU 1 blocks forever, unless CPU 0 ever queues enough further callbacks to arm an
RCU_NOCB_WAKE_FORCEtimer.Except that -something- must have already been prepared to wake up in order to invoke CB1. And that something would invoke CB2 along with CB1, given that they are both on the done list. If there is no such wakeup already, then the hang could occur with just CB1, without the help of CB2.
Heh good point. I was confused with CB1 on RCU_DONE_TAIL and the possibility for CB2 to be entrained on RCU_WAIT_TAIL. But that's indeed not supposed to happen. Ok so this patch indeed doesn't make sense outside lazy.
Quoted 10 lines
Quoted 2 lines
This is also required to make sure lazy callbacks in future patches don't end up making
rcu_barrier()wait for multiple seconds.But I do see that the wakeup is needed in the lazy case, and if I remember correctly, the ten-second
rcu_barrier()delay really did happen. If I understand correctly, for this to happen, all of the callbacks must be in the bypass list, that is, ->cblist must be empty.So has the scenario steps 1-6 called out above actually happened in the absence of lazy callbacks?
Nope, so I guess we can have the pending check around rcu_nocb_flush_bypass() only...
Thanks!
Paul E. McKenney
maintainerRe: [PATCH v8 01/13] rcu: Fix missing nocb gp wake on rcu_barrier()
Quoted 33 lines · Frederic Weisbecker
Quoted 32 lines · Paul E. McKenney
Quoted 8 lines · Joel Fernandes (Google)
From: Frederic Weisbecker <[email protected]>
Upon entraining a callback to a NOCB CPU, no further wake up is issued on the corresponding
nocb_gpkthread. As a result, the callback and all the subsequent ones on that CPU may be ignored, at least until anRCU_NOCB_WAKE_FORCEtimer is ever armed or another NOCB CPU belonging to the same group enqueues a callback on an empty queue.Quoted 4 lines
Here is a possible bad scenario:
1) CPU 0 is NOCB unlike all other CPUs.
2) CPU 0 queues a callbackCall it CB1.
Quoted 3 lines
2) The grace period related to that callback elapses
3) The callback is moved to the done list (but is not invoked yet), there are no more pending callbacks for CPU 0So CB1 is on ->cblist waiting to be invoked, correct?
Quoted 2 lines
4) CPU 1 calls
rcu_barrier()and sends an IPI to CPU 0
5) CPU 0 entrains the callback but doesn't wake upnocb_gpAnd CB1 must still be there because otherwise the IPI handler would not have entrained the callback, correct? If so, we have both CB1 and the
rcu_barrier()callback (call it CB2) in ->cblist, but on the done list.Quoted 2 lines
6) CPU 1 blocks forever, unless CPU 0 ever queues enough further callbacks to arm an
RCU_NOCB_WAKE_FORCEtimer.
Quoted 8 lines
Quoted 4 lines
Except that -something- must have already been prepared to wake up in order to invoke CB1. And that something would invoke CB2 along with CB1, given that they are both on the done list. If there is no such wakeup already, then the hang could occur with just CB1, without the help of CB2.
Heh good point. I was confused with CB1 on
RCU_DONE_TAILand the possibility for CB2 to be entrained onRCU_WAIT_TAIL. But that's indeed not supposed to happen. Ok so this patch indeed doesn't make sense outside lazy.
Whew!!! ;-)
Quoted 8 lines
Quoted 8 lines
Quoted 2 lines
This is also required to make sure lazy callbacks in future patches don't end up making
rcu_barrier()wait for multiple seconds.But I do see that the wakeup is needed in the lazy case, and if I remember correctly, the ten-second
rcu_barrier()delay really did happen. If I understand correctly, for this to happen, all of the callbacks must be in the bypass list, that is, ->cblist must be empty.
Quoted 5 lines
Quoted 2 lines
So has the scenario steps 1-6 called out above actually happened in the absence of lazy callbacks?
Nope, so I guess we can have the pending check around
rcu_nocb_flush_bypass()only...
OK, sounds good.
I have put this series on branch lazy.2022.10.14a and am testing it.
Thanx, Paul
Re: [PATCH v8 01/13] rcu: Fix missing nocb gp wake on rcu_barrier()
Quoted 40 lines · Paul E. McKenney
Quoted 38 lines · Frederic Weisbecker
Quoted 36 lines · Paul E. McKenney
Quoted 8 lines · Joel Fernandes (Google)
From: Frederic Weisbecker <[email protected]>
Upon entraining a callback to a NOCB CPU, no further wake up is issued on the corresponding
nocb_gpkthread. As a result, the callback and all the subsequent ones on that CPU may be ignored, at least until anRCU_NOCB_WAKE_FORCEtimer is ever armed or another NOCB CPU belonging to the same group enqueues a callback on an empty queue.Quoted 4 lines
Here is a possible bad scenario:
1) CPU 0 is NOCB unlike all other CPUs.
2) CPU 0 queues a callbackCall it CB1.
Quoted 3 lines
2) The grace period related to that callback elapses
3) The callback is moved to the done list (but is not invoked yet), there are no more pending callbacks for CPU 0So CB1 is on ->cblist waiting to be invoked, correct?
Quoted 2 lines
4) CPU 1 calls
rcu_barrier()and sends an IPI to CPU 0
5) CPU 0 entrains the callback but doesn't wake upnocb_gpAnd CB1 must still be there because otherwise the IPI handler would not have entrained the callback, correct? If so, we have both CB1 and the
rcu_barrier()callback (call it CB2) in ->cblist, but on the done list.Quoted 2 lines
6) CPU 1 blocks forever, unless CPU 0 ever queues enough further callbacks to arm an
RCU_NOCB_WAKE_FORCEtimer.Except that -something- must have already been prepared to wake up in order to invoke CB1. And that something would invoke CB2 along with CB1, given that they are both on the done list. If there is no such wakeup already, then the hang could occur with just CB1, without the help of CB2.
Quoted 11 lines
Quoted 3 lines
Heh good point. I was confused with CB1 on
RCU_DONE_TAILand the possibility for CB2 to be entrained onRCU_WAIT_TAIL. But that's indeed not supposed to happen. Ok so this patch indeed doesn't make sense outside lazy.Whew!!! ;-)
Quoted 5 lines
Quoted 5 lines
Quoted 2 lines
This is also required to make sure lazy callbacks in future patches don't end up making
rcu_barrier()wait for multiple seconds.But I do see that the wakeup is needed in the lazy case, and if I remember correctly, the ten-second
rcu_barrier()delay really did happen. If I
Yes it did happen. Real world device testing confirmed it.
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines
understand correctly, for this to happen, all of the callbacks must be in the bypass list, that is, ->cblist must be empty.
So has the scenario steps 1-6 called out above actually happened in the absence of lazy callbacks?
Quoted 6 lines
Quoted 2 lines
Nope, so I guess we can have the pending check around
rcu_nocb_flush_bypass()only...OK, sounds good.
I have put this series on branch lazy.2022.10.14a and am testing it.
I agree with the discussion, though if all CBs are in the bypass list, the patch will also save 2 jiffies.
So just commit messages that need rework then? This one can be taken instead:
https://lore.kernel.org/rcu/[email protected]/T/#m14d21fbce23539a521693a4184b28ddc55d7d2c5
Thanks!
- Joel
Paul E. McKenney
maintainerRe: [PATCH v8 01/13] rcu: Fix missing nocb gp wake on rcu_barrier()
Quoted 47 lines · Joel Fernandes
Quoted 46 lines · Paul E. McKenney
Quoted 33 lines · Frederic Weisbecker
Quoted 32 lines · Paul E. McKenney
Quoted 8 lines · Joel Fernandes (Google)
From: Frederic Weisbecker <[email protected]>
Upon entraining a callback to a NOCB CPU, no further wake up is issued on the corresponding
nocb_gpkthread. As a result, the callback and all the subsequent ones on that CPU may be ignored, at least until anRCU_NOCB_WAKE_FORCEtimer is ever armed or another NOCB CPU belonging to the same group enqueues a callback on an empty queue.Quoted 4 lines
Here is a possible bad scenario:
1) CPU 0 is NOCB unlike all other CPUs.
2) CPU 0 queues a callbackCall it CB1.
Quoted 3 lines
2) The grace period related to that callback elapses
3) The callback is moved to the done list (but is not invoked yet), there are no more pending callbacks for CPU 0So CB1 is on ->cblist waiting to be invoked, correct?
Quoted 2 lines
4) CPU 1 calls
rcu_barrier()and sends an IPI to CPU 0
5) CPU 0 entrains the callback but doesn't wake upnocb_gpAnd CB1 must still be there because otherwise the IPI handler would not have entrained the callback, correct? If so, we have both CB1 and the
rcu_barrier()callback (call it CB2) in ->cblist, but on the done list.Quoted 2 lines
6) CPU 1 blocks forever, unless CPU 0 ever queues enough further callbacks to arm an
RCU_NOCB_WAKE_FORCEtimer.Quoted 8 lines
Quoted 4 lines
Except that -something- must have already been prepared to wake up in order to invoke CB1. And that something would invoke CB2 along with CB1, given that they are both on the done list. If there is no such wakeup already, then the hang could occur with just CB1, without the help of CB2.
Heh good point. I was confused with CB1 on
RCU_DONE_TAILand the possibility for CB2 to be entrained onRCU_WAIT_TAIL. But that's indeed not supposed to happen. Ok so this patch indeed doesn't make sense outside lazy.Whew!!! ;-)
Quoted 7 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 2 lines
This is also required to make sure lazy callbacks in future patches don't end up making
rcu_barrier()wait for multiple seconds.But I do see that the wakeup is needed in the lazy case, and if I remember correctly, the ten-second
rcu_barrier()delay really did happen. If IYes it did happen. Real world device testing confirmed it.
Very good, thank you!
Quoted 11 lines
Quoted 11 lines
Quoted 8 lines
Quoted 5 lines
understand correctly, for this to happen, all of the callbacks must be in the bypass list, that is, ->cblist must be empty.
So has the scenario steps 1-6 called out above actually happened in the absence of lazy callbacks?
Nope, so I guess we can have the pending check around
rcu_nocb_flush_bypass()only...OK, sounds good.
Quoted 7 lines
Quoted 1 line
I have put this series on branch lazy.2022.10.14a and am testing it.
I agree with the discussion, though if all CBs are in the bypass list, the patch will also save 2 jiffies.
So just commit messages that need rework then? This one can be taken instead:
https://lore.kernel.org/rcu/[email protected]/T/#m14d21fbce23539a521693a4184b28ddc55d7d2c5
This one looks plausible to me.
Thanx, Paul
Frederic Weisbecker
maintainerRe: [PATCH v8 01/13] rcu: Fix missing nocb gp wake on rcu_barrier()
Quoted 8 lines · Paul E. McKenney
Quoted 5 lines · Joel Fernandes
I agree with the discussion, though if all CBs are in the bypass list, the patch will also save 2 jiffies.
So just commit messages that need rework then? This one can be taken instead:
https://lore.kernel.org/rcu/[email protected]/T/#m14d21fbce23539a521693a4184b28ddc55d7d2c5This one looks plausible to me.
With the following modified diff (passed 25 hours of TREE01):
kernel/rcu/tree.c +11 -0
@@ -3914,6 +3914,8 @@ static void rcu_barrier_entrain(struct rcu_data *rdp) {unsigned long gseq =READ_ONCE (rcu_state.barrier_sequence);unsigned long lseq =READ_ONCE (rdp->barrier_seq_snap); +bool wake_nocb = false; +bool was_alldone = false;lockdep_assert_held (&rcu_state.barrier_lock);if (rcu_seq_state (lseq) || !rcu_seq_state (gseq) ||rcu_seq_ctr (lseq) !=rcu_seq_ctr (gseq))
@@ -3922,7 +3924,14 @@ static void rcu_barrier_entrain(struct rcu_data *rdp) rdp->barrier_head.func = rcu_barrier_callback;debug_rcu_head_queue (&rdp->barrier_head);rcu_nocb_lock (rdp); +/* +* Flush bypass and wakeup rcuog if we add callbacks to an empty regular +* queue. This way we don't wait for bypass timer that can reach seconds +* if it's fully lazy. +*/ + was_alldone =rcu_rdp_is_offloaded (rdp) && !rcu_segcblist_pend_cbs (&rdp->cblist);WARN_ON_ONCE (!rcu_nocb_flush_bypass (rdp,NULL , jiffies)); + wake_nocb = was_alldone &&rcu_segcblist_pend_cbs (&rdp->cblist);if (rcu_segcblist_entrain (&rdp->cblist, &rdp->barrier_head)) {atomic_inc (&rcu_state.barrier_cpu_count); }else {
@@ -3930,6 +3939,8 @@ static void rcu_barrier_entrain(struct rcu_data *rdp)rcu_barrier_trace (TPS ("IRQNQ" ),-1 , rcu_state.barrier_sequence); }rcu_nocb_unlock (rdp); +if (wake_nocb) +wake_nocb_gp (rdp, false);smp_store_release (&rdp->barrier_seq_snap, gseq); }
kernel/rcu/tree.h +1 -0
@@ -439,6 +439,7 @@ static void zero_cpu_stall_ticks(struct rcu_data *rdp);static struct swait_queue_head *rcu_nocb_gp_get (struct rcu_node *rnp);static void rcu_nocb_gp_cleanup (struct swait_queue_head *sq);static void rcu_init_one_nocb (struct rcu_node *rnp); +static bool wake_nocb_gp (struct rcu_data *rdp,bool force);static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp,unsigned long j); static bool rcu_nocb_try_bypass(struct rcu_data *rdp, struct rcu_head *rhp,
kernel/rcu/tree_nocb.h +5 -0
@@ -1558,6 +1558,11 @@ static void rcu_init_one_nocb(struct rcu_node *rnp) { } +static bool wake_nocb_gp (struct rcu_data *rdp,bool force) +{ +return false; +} +static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp,unsigned long j) {
Paul E. McKenney
maintainerRe: [PATCH v8 01/13] rcu: Fix missing nocb gp wake on rcu_barrier()
Quoted 11 lines · Frederic Weisbecker
Quoted 8 lines · Paul E. McKenney
Quoted 5 lines · Joel Fernandes
I agree with the discussion, though if all CBs are in the bypass list, the patch will also save 2 jiffies.
So just commit messages that need rework then? This one can be taken instead:
https://lore.kernel.org/rcu/[email protected]/T/#m14d21fbce23539a521693a4184b28ddc55d7d2c5This one looks plausible to me.
With the following modified diff (passed 25 hours of TREE01):
Very good!
Could one of you (presumably Joel) please send v9?
Just to avoid me getting the wrong patch in the wrong place or similar. Or mangling the required rebase following a pull (otherwise, as soon as I create branches, Stephen Rothwell notes the lack of a committer signoff.)
Thanx, Paul
Quoted 65 lines
kernel/rcu/tree.c +11
-0@@ -3914,6 +3914,8 @@ static void rcu_barrier_entrain(struct rcu_data *rdp) {unsigned long gseq =READ_ONCE (rcu_state.barrier_sequence);unsigned long lseq =READ_ONCE (rdp->barrier_seq_snap); +bool wake_nocb = false; +bool was_alldone = false;lockdep_assert_held (&rcu_state.barrier_lock);if (rcu_seq_state (lseq) || !rcu_seq_state (gseq) ||rcu_seq_ctr (lseq) !=rcu_seq_ctr (gseq))@@ -3922,7 +3924,14 @@ static void rcu_barrier_entrain(struct rcu_data *rdp) rdp->barrier_head.func = rcu_barrier_callback;debug_rcu_head_queue (&rdp->barrier_head);rcu_nocb_lock (rdp); +/* +* Flush bypass and wakeup rcuog if we add callbacks to an empty regular +* queue. This way we don't wait for bypass timer that can reach seconds +* if it's fully lazy. +*/ + was_alldone =rcu_rdp_is_offloaded (rdp) && !rcu_segcblist_pend_cbs (&rdp->cblist);WARN_ON_ONCE (!rcu_nocb_flush_bypass (rdp,NULL , jiffies)); + wake_nocb = was_alldone &&rcu_segcblist_pend_cbs (&rdp->cblist);if (rcu_segcblist_entrain (&rdp->cblist, &rdp->barrier_head)) {atomic_inc (&rcu_state.barrier_cpu_count); }else {@@ -3930,6 +3939,8 @@ static void rcu_barrier_entrain(struct rcu_data *rdp)rcu_barrier_trace (TPS ("IRQNQ" ),-1 , rcu_state.barrier_sequence); }rcu_nocb_unlock (rdp); +if (wake_nocb) +wake_nocb_gp (rdp, false);smp_store_release (&rdp->barrier_seq_snap, gseq); }kernel/rcu/tree.h +1
-0@@ -439,6 +439,7 @@ static void zero_cpu_stall_ticks(struct rcu_data *rdp);static struct swait_queue_head *rcu_nocb_gp_get (struct rcu_node *rnp);static void rcu_nocb_gp_cleanup (struct swait_queue_head *sq);static void rcu_init_one_nocb (struct rcu_node *rnp); +static bool wake_nocb_gp (struct rcu_data *rdp,bool force);static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp,unsigned long j); static bool rcu_nocb_try_bypass(struct rcu_data *rdp, struct rcu_head *rhp,kernel/rcu/tree_nocb.h +5
-0@@ -1558,6 +1558,11 @@ static void rcu_init_one_nocb(struct rcu_node *rnp) { } +static bool wake_nocb_gp (struct rcu_data *rdp,bool force) +{ +return false; +} +static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp,unsigned long j) {
Re: [PATCH v8 01/13] rcu: Fix missing nocb gp wake on rcu_barrier()
Quoted 14 lines · Paul E. McKenney
Quoted 11 lines · Frederic Weisbecker
Quoted 8 lines · Paul E. McKenney
Quoted 5 lines · Joel Fernandes
I agree with the discussion, though if all CBs are in the bypass list, the patch will also save 2 jiffies.
So just commit messages that need rework then? This one can be taken instead:
https://lore.kernel.org/rcu/[email protected]/T/#m14d21fbce23539a521693a4184b28ddc55d7d2c5This one looks plausible to me.
With the following modified diff (passed 25 hours of TREE01):
Quoted 7 lines
Very good!
Could one of you (presumably Joel) please send v9?
Just to avoid me getting the wrong patch in the wrong place or similar. Or mangling the required rebase following a pull (otherwise, as soon as I create branches, Stephen Rothwell notes the lack of a committer signoff.)
Taking a break from raking leaves so doing it right now! Those leaves!
- Joel
PATCH v8 2/13 rcu: Make call_rcu() lazy to save power
In reply to Joel Fernandes (Google)Implement timer-based RCU callback batching (also known as lazy callbacks). With this we save about 5-10% of power consumed due to RCU requests that happen when system is lightly loaded or idle.
By default, all async callbacks (queued via call_rcu) are marked lazy. An alternate API call_rcu_flush() is provided for the few users, for example synchronize_rcu(), that need the old behavior.
The batch is flushed whenever a certain amount of time has passed, or the batch on a particular CPU grows too big. Also memory pressure will flush it in a future patch.
To handle several corner cases automagically (such as rcu_barrier() and hotplug), we re-use bypass lists which were originally introduced to address lock contention, to handle lazy CBs as well. The bypass list length has the lazy CB length included in it. A separate lazy CB length counter is also introduced to keep track of the number of lazy CBs.
--- include/linux/rcupdate.h | 7 ++ kernel/rcu/Kconfig | 8 ++ kernel/rcu/rcu.h | 8 ++ kernel/rcu/tiny.c | 2 +- kernel/rcu/tree.c | 129 ++++++++++++++++++++------------ kernel/rcu/tree.h | 11 ++- kernel/rcu/tree_exp.h | 2 +- kernel/rcu/tree_nocb.h | 157 +++++++++++++++++++++++++++++++-------- 8 files changed, 243 insertions(+), 81 deletions(-)
include/linux/rcupdate.h +7 -0
@@ -108,6 +108,13 @@ static inline int rcu_preempt_depth(void)#endif /* #else #ifdef CONFIG_PREEMPT_RCU */ +#ifdef CONFIG_RCU_LAZY +void call_rcu_flush (struct rcu_head *head,rcu_callback_t func); +#else +static inline void call_rcu_flush (struct rcu_head *head, +rcu_callback_t func) {call_rcu (head, func); } +#endif +/* Internal to kernel */ void rcu_init (void );extern int rcu_scheduler_active;
kernel/rcu/Kconfig +8 -0
@@ -314,4 +314,12 @@ config TASKS_TRACE_RCU_READ_MB Say N here if you hate read-side memory barriers. Take the default if you are unsure. +config RCU_LAZY + bool "RCU callback lazy invocation functionality" + depends on RCU_NOCB_CPU + default n + help + To save power, batch RCU callbacks and flush after delay, memory + pressure or callback list growing too big. + endmenu # "RCU Subsystem"
kernel/rcu/rcu.h +8 -0
@@ -474,6 +474,14 @@ enum rcutorture_type { INVALID_RCU_FLAVOR }; +#if defined(CONFIG_RCU_LAZY ) +unsigned long rcu_lazy_get_jiffies_till_flush (void ); +void rcu_lazy_set_jiffies_till_flush (unsigned long j); +#else +static inline unsigned long rcu_lazy_get_jiffies_till_flush (void ) {return 0 ; } +static inline void rcu_lazy_set_jiffies_till_flush (unsigned long j) { } +#endif +#if defined(CONFIG_TREE_RCU )void rcutorture_get_gp_data (enum rcutorture_type test_type,int *flags,unsigned long *gp_seq);
kernel/rcu/tiny.c +1 -1
@@ -44,7 +44,7 @@ static struct rcu_ctrlblk rcu_ctrlblk = {void rcu_barrier (void ) { -wait_rcu_gp (call_rcu); +wait_rcu_gp (call_rcu_flush); }EXPORT_SYMBOL (rcu_barrier);
kernel/rcu/tree.c +83 -46
166 lines not on this page · raw message
kernel/rcu/tree.h +7 -4
@@ -263,14 +263,16 @@ struct rcu_data {unsigned long last_fqs_resched;/* Time of last rcu_resched(). */ unsigned long last_sched_clock;/* Jiffies of last rcu_sched_clock_irq(). */ +long lazy_len;/* Length of buffered lazy callbacks. */ int cpu; };/* Values for nocb_defer_wakeup field in struct rcu_data. */ #define RCU_NOCB_WAKE_NOT 0#define RCU_NOCB_WAKE_BYPASS 1 -#define RCU_NOCB_WAKE 2 -#define RCU_NOCB_WAKE_FORCE 3 +#define RCU_NOCB_WAKE_LAZY 2 +#define RCU_NOCB_WAKE 3 +#define RCU_NOCB_WAKE_FORCE 4#define RCU_JIFFIES_TILL_FORCE_QS (1 + (HZ > 250) + (HZ > 500))/* For jiffies_till_first_fqs and */
@@ -441,9 +443,10 @@ static void rcu_nocb_gp_cleanup(struct swait_queue_head *sq);static void rcu_init_one_nocb (struct rcu_node *rnp);static bool wake_nocb_gp (struct rcu_data *rdp,bool force);static bool rcu_nocb_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp, -unsigned long j); +unsigned long j,bool lazy);static bool rcu_nocb_try_bypass (struct rcu_data *rdp,struct rcu_head *rhp, -bool *was_alldone,unsigned long flags); +bool *was_alldone,unsigned long flags, +bool lazy);static void __call_rcu_nocb_wake (struct rcu_data *rdp,bool was_empty,unsigned long flags);static int rcu_nocb_need_deferred_wakeup (struct rcu_data *rdp,int level);
kernel/rcu/tree_exp.h +1 -1
@@ -937,7 +937,7 @@ void synchronize_rcu_expedited(void)/* If expedited grace periods are prohibited, fall back to normal. */ if (rcu_gp_is_normal ()) { -wait_rcu_gp (call_rcu); +wait_rcu_gp (call_rcu_flush);return ; }
kernel/rcu/tree_nocb.h +128 -29
316 lines not on this page · raw message
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
Re: [PATCH v8 02/13] rcu: Make call_rcu() lazy to save power
kernel/rcu/tree_nocb.h static void wake_nocb_gp_defer(struct rcu_data *rdp, int waketype, · line 301
Quoted 12 lines · Joel Fernandes (Google)
kernel/rcu/tree_nocb.h +4
-3@@ -269,10 +294,14 @@ static void wake_nocb_gp_defer(struct rcu_data *rdp, int waketype,raw_spin_lock_irqsave (&rdp_gp->nocb_gp_lock, flags);/* -* Bypass wakeup overrides previous deferments. In case -* of callback storm, no need to wake up too early. +* Bypass wakeup overrides previous deferments. In case of +* callback storm, no need to wake up too early. */ -if (waketype ==RCU_NOCB_WAKE_BYPASS ) { +if (waketype ==RCU_NOCB_WAKE_LAZY && +READ_ONCE (rdp->nocb_defer_wakeup) ==RCU_NOCB_WAKE_NOT ) {
No need for READ_ONCE() here.
kernel/rcu/tree_nocb.h static void nocb_gp_wait(struct rcu_data *my_rdp) · line 686
Quoted 12 lines
kernel/rcu/tree_nocb.h +5
-0+mod_timer (&rdp_gp->nocb_timer, jiffies + jiffies_till_flush); +WRITE_ONCE (rdp_gp->nocb_defer_wakeup, waketype); + }else if (waketype ==RCU_NOCB_WAKE_BYPASS ) {mod_timer (&rdp_gp->nocb_timer, jiffies +2 );WRITE_ONCE (rdp_gp->nocb_defer_wakeup, waketype); }else {@@ -611,6 +682,8 @@ static void nocb_gp_wait(struct rcu_data *my_rdp)unsigned long flags;bool gotcbs = false;unsigned long j = jiffies; +bool lazy = false; +long lazy_ncbs;
Looks like lazy_ncbs can be declared withing the loop, as well as bypass_cbs.
Anyway, apart from such boring cosmectics:
Thanks!
Re: [PATCH v8 02/13] rcu: Make call_rcu() lazy to save power
Quoted 32 lines · Frederic Weisbecker
kernel/rcu/tree_nocb.h static void wake_nocb_gp_defer(struct rcu_data *rdp, int waketype, · line 301
Quoted 12 lines · Joel Fernandes (Google)
kernel/rcu/tree_nocb.h +4
-3@@ -269,10 +294,14 @@ static void wake_nocb_gp_defer(struct rcu_data *rdp, int waketype,raw_spin_lock_irqsave (&rdp_gp->nocb_gp_lock, flags);/* -* Bypass wakeup overrides previous deferments. In case -* of callback storm, no need to wake up too early. +* Bypass wakeup overrides previous deferments. In case of +* callback storm, no need to wake up too early. */ -if (waketype ==RCU_NOCB_WAKE_BYPASS ) { +if (waketype ==RCU_NOCB_WAKE_LAZY && +READ_ONCE (rdp->nocb_defer_wakeup) ==RCU_NOCB_WAKE_NOT ) {No need for
READ_ONCE()here.kernel/rcu/tree_nocb.h static void nocb_gp_wait(struct rcu_data *my_rdp) · line 686
Quoted 12 lines
kernel/rcu/tree_nocb.h +5
-0+mod_timer (&rdp_gp->nocb_timer, jiffies + jiffies_till_flush); +WRITE_ONCE (rdp_gp->nocb_defer_wakeup, waketype); + }else if (waketype ==RCU_NOCB_WAKE_BYPASS ) {mod_timer (&rdp_gp->nocb_timer, jiffies +2 );WRITE_ONCE (rdp_gp->nocb_defer_wakeup, waketype); }else {@@ -611,6 +682,8 @@ static void nocb_gp_wait(struct rcu_data *my_rdp)unsigned long flags;bool gotcbs = false;unsigned long j = jiffies; +bool lazy = false; +long lazy_ncbs;Looks like
lazy_ncbscan be declared withing the loop, as well asbypass_cbs.
Quoted 3 lines
Anyway, apart from such boring cosmectics:
Thanks, I folded the cosmetic suggestions with the Ack and pushed it to:
https://github.com/joelagnel/linux-kernel.git (branch rcu-next-oct.12.22)
thanks,
- Joel
Paul E. McKenney
maintainerRe: [PATCH v8 02/13] rcu: Make call_rcu() lazy to save power
Quoted 33 lines · Joel Fernandes
Quoted 32 lines · Frederic Weisbecker
kernel/rcu/tree_nocb.h static void wake_nocb_gp_defer(struct rcu_data *rdp, int waketype, · line 301
Quoted 12 lines · Joel Fernandes (Google)
kernel/rcu/tree_nocb.h +4
-3@@ -269,10 +294,14 @@ static void wake_nocb_gp_defer(struct rcu_data *rdp, int waketype,raw_spin_lock_irqsave (&rdp_gp->nocb_gp_lock, flags);/* -* Bypass wakeup overrides previous deferments. In case -* of callback storm, no need to wake up too early. +* Bypass wakeup overrides previous deferments. In case of +* callback storm, no need to wake up too early. */ -if (waketype ==RCU_NOCB_WAKE_BYPASS ) { +if (waketype ==RCU_NOCB_WAKE_LAZY && +READ_ONCE (rdp->nocb_defer_wakeup) ==RCU_NOCB_WAKE_NOT ) {No need for
READ_ONCE()here.kernel/rcu/tree_nocb.h static void nocb_gp_wait(struct rcu_data *my_rdp) · line 686
Quoted 12 lines
kernel/rcu/tree_nocb.h +5
-0+mod_timer (&rdp_gp->nocb_timer, jiffies + jiffies_till_flush); +WRITE_ONCE (rdp_gp->nocb_defer_wakeup, waketype); + }else if (waketype ==RCU_NOCB_WAKE_BYPASS ) {mod_timer (&rdp_gp->nocb_timer, jiffies +2 );WRITE_ONCE (rdp_gp->nocb_defer_wakeup, waketype); }else {@@ -611,6 +682,8 @@ static void nocb_gp_wait(struct rcu_data *my_rdp)unsigned long flags;bool gotcbs = false;unsigned long j = jiffies; +bool lazy = false; +long lazy_ncbs;Looks like
lazy_ncbscan be declared withing the loop, as well asbypass_cbs.
Quoted 6 lines
Quoted 3 lines
Anyway, apart from such boring cosmectics:
Thanks, I folded the cosmetic suggestions with the Ack and pushed it to:
https://github.com/joelagnel/linux-kernel.git (branch rcu-next-oct.12.22)
Just following back up with Frederic to see if all of your guys' words had the same meanings. ;-)
Thanx, Paul
PATCH v8 3/13 rcu: Refactor code a bit in rcu_nocb_do_flush_bypass()
In reply to Joel Fernandes (Google)This consolidates the code a bit and makes it cleaner. Functionally it is the same.
--- kernel/rcu/tree_nocb.h | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-)
kernel/rcu/tree_nocb.h +9 -8
@@ -327,10 +327,11 @@ static void wake_nocb_gp_defer(struct rcu_data *rdp, int waketype,* * Note that this function always returns true if rhp is NULL. */ -static bool rcu_nocb_do_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp, +static bool rcu_nocb_do_flush_bypass (struct rcu_data *rdp,struct rcu_head *rhp_in,unsigned long j,bool lazy) {struct rcu_cblist rcl; +struct rcu_head *rhp = rhp_in;WARN_ON_ONCE (!rcu_rdp_is_offloaded (rdp));rcu_lockdep_assert_cblist_protected (rdp);
@@ -345,16 +346,16 @@ static bool rcu_nocb_do_flush_bypass(struct rcu_data *rdp, struct rcu_head *rhp,/* * If the new CB requested was a lazy one, queue it onto the main -* ->cblist so we can take advantage of a sooner grade period. +* ->cblist so that we can take advantage of the grace-period that will +* happen regardless. But queue it onto the bypass list first so that +* the lazy CB is ordered with the existing CBs in the bypass list. */ if (lazy && rhp) { -rcu_cblist_flush_enqueue (&rcl, &rdp->nocb_bypass,NULL ); -rcu_cblist_enqueue (&rcl, rhp); -WRITE_ONCE (rdp->lazy_len,0 ); - }else { -rcu_cblist_flush_enqueue (&rcl, &rdp->nocb_bypass, rhp); -WRITE_ONCE (rdp->lazy_len,0 ); +rcu_cblist_enqueue (&rdp->nocb_bypass, rhp); + rhp =NULL ; } +rcu_cblist_flush_enqueue (&rcl, &rdp->nocb_bypass, rhp); +WRITE_ONCE (rdp->lazy_len,0 );rcu_segcblist_insert_pend_cbs (&rdp->cblist, &rcl);WRITE_ONCE (rdp->nocb_bypass_first, j);
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 4/13 rcu: shrinker for lazy rcu
In reply to Joel Fernandes (Google)From: Vineeth Pillai <[email protected]>
The shrinker is used to speed up the free'ing of memory potentially held by RCU lazy callbacks. RCU kernel module test cases show this to be effective. Test is introduced in a later patch.
--- kernel/rcu/tree_nocb.h | 52 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 52 insertions(+)
kernel/rcu/tree_nocb.h +52 -0
@@ -1312,6 +1312,55 @@ int rcu_nocb_cpu_offload(int cpu) }EXPORT_SYMBOL_GPL (rcu_nocb_cpu_offload); +static unsigned long +lazy_rcu_shrink_count (struct shrinker *shrink,struct shrink_control *sc) +{ +int cpu; +unsigned long count =0 ; + +/* Snapshot count of all CPUs */ +for_each_possible_cpu (cpu) { +struct rcu_data *rdp =per_cpu_ptr (&rcu_data, cpu); + + count +=READ_ONCE (rdp->lazy_len); + } + +return count ? count :SHRINK_EMPTY ; +} + +static unsigned long +lazy_rcu_shrink_scan (struct shrinker *shrink,struct shrink_control *sc) +{ +int cpu; +unsigned long flags; +unsigned long count =0 ; + +/* Snapshot count of all CPUs */ +for_each_possible_cpu (cpu) { +struct rcu_data *rdp =per_cpu_ptr (&rcu_data, cpu); +int _count =READ_ONCE (rdp->lazy_len); + +if (_count ==0 ) +continue ; +rcu_nocb_lock_irqsave (rdp, flags); +WRITE_ONCE (rdp->lazy_len,0 ); +rcu_nocb_unlock_irqrestore (rdp, flags); +wake_nocb_gp (rdp, false); + sc->nr_to_scan -= _count; + count += _count; +if (sc->nr_to_scan <=0 ) +break ; + } +return count ? count :SHRINK_STOP ; +} + +static struct shrinker lazy_rcu_shrinker = { + .count_objects = lazy_rcu_shrink_count, + .scan_objects = lazy_rcu_shrink_scan, + .batch =0 , + .seeks =DEFAULT_SEEKS , +}; +void __init rcu_init_nohz (void ) {int cpu;
@@ -1342,6 +1391,9 @@ void __init rcu_init_nohz(void)if (!rcu_state.nocb_is_setup)return ; +if (register_shrinker (&lazy_rcu_shrinker,"rcu-lazy" )) +pr_err ("Failed to register lazy_rcu shrinker!\n" ); +if (!cpumask_subset (rcu_nocb_mask, cpu_possible_mask)) {pr_info ("\tNote: kernel parameter 'rcu_nocbs=', 'nohz_full', or 'isolcpus=' contains nonexistent CPUs.\n" ); cpumask_and(rcu_nocb_mask, cpu_possible_mask,
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 5/13 rcuscale: Add laziness and kfree tests
In reply to Joel Fernandes (Google)We add 2 tests to rcuscale, first one is a startup test to check whether we are not too lazy or too hard working. Two, emulate kfree_rcu() itself to use call_rcu() and check memory pressure. In my testing, the new call_rcu() does well to keep memory pressure under control, similar to kfree_rcu().
--- kernel/rcu/rcuscale.c | 68 +++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 66 insertions(+), 2 deletions(-)
kernel/rcu/rcuscale.c +66 -2
@@ -95,6 +95,7 @@ torture_param(int, verbose, 1, "Enable verbose debugging printk()s");torture_param (int, writer_holdoff,0 ,"Holdoff (us) between GPs, zero to disable" );torture_param (int, kfree_rcu_test,0 ,"Do we run a kfree_rcu() scale test?" );torture_param (int, kfree_mult,1 ,"Multiple of kfree_obj size to allocate." ); +torture_param (int, kfree_by_call_rcu,0 ,"Use call_rcu() to emulate kfree_rcu()?" );static char *scale_type ="rcu" ;module_param (scale_type, charp,0444 );
@@ -659,6 +660,14 @@ struct kfree_obj {struct rcu_head rh; }; +/* Used if doing RCU-kfree'ing via call_rcu(). */ +static void kfree_call_rcu (struct rcu_head *rh) +{ + struct kfree_obj *obj = container_of(rh, struct kfree_obj, rh); + +kfree (obj); +} +static int kfree_scale_thread (void *arg) {
@@ -696,6 +705,11 @@ kfree_scale_thread(void *arg)if (!alloc_ptr)return -ENOMEM ; +if (kfree_by_call_rcu) { +call_rcu (&(alloc_ptr->rh), kfree_call_rcu); +continue ; + } +// By default kfree_rcu_test_single and kfree_rcu_test_double are // initialized to false. If both have the same value (false or true) // both are randomly tested, otherwise only the one with value true
@@ -767,11 +781,59 @@ kfree_scale_shutdown(void *arg)return -EINVAL ; } +// Used if doing RCU-kfree'ing via call_rcu(). +static unsigned long jiffies_at_lazy_cb; +static struct rcu_head lazy_test1_rh; +static int rcu_lazy_test1_cb_called; +static void call_rcu_lazy_test1 (struct rcu_head *rh) +{ + jiffies_at_lazy_cb = jiffies; +WRITE_ONCE (rcu_lazy_test1_cb_called,1 ); +} +static int __init kfree_scale_init (void ) { -long i;int firsterr =0 ; +long i; +unsigned long jif_start; +unsigned long orig_jif; + +// Also, do a quick self-test to ensure laziness is as much as +// expected. +if (kfree_by_call_rcu && !IS_ENABLED (CONFIG_RCU_LAZY )) { +pr_alert ("CONFIG_RCU_LAZY is disabled, falling back to kfree_rcu() " +"for delayed RCU kfree'ing\n" ); + kfree_by_call_rcu =0 ; + } + +if (kfree_by_call_rcu) { +/* do a test to check the timeout. */ + orig_jif =rcu_lazy_get_jiffies_till_flush (); + +rcu_lazy_set_jiffies_till_flush (2 *HZ ); +rcu_barrier (); + + jif_start = jiffies; + jiffies_at_lazy_cb =0 ; +call_rcu (&lazy_test1_rh, call_rcu_lazy_test1); + +smp_cond_load_relaxed (&rcu_lazy_test1_cb_called,VAL ==1 ); + +rcu_lazy_set_jiffies_till_flush (orig_jif); + +if (WARN_ON_ONCE (jiffies_at_lazy_cb - jif_start <2 *HZ )) { +pr_alert ("ERROR: call_rcu() CBs are not being lazy as expected!\n" ); +WARN_ON_ONCE (1 ); +return -1 ; + } + +if (WARN_ON_ONCE (jiffies_at_lazy_cb - jif_start >3 *HZ )) { +pr_alert ("ERROR: call_rcu() CBs are being too lazy!\n" ); +WARN_ON_ONCE (1 ); +return -1 ; + } + } kfree_nrealthreads =compute_real (kfree_nthreads);/* Start up the kthreads. */
@@ -784,7 +846,9 @@ kfree_scale_init(void)schedule_timeout_uninterruptible (1 ); } -pr_alert ("kfree object size=%zu\n" , kfree_mult *sizeof (struct kfree_obj )); +pr_alert ("kfree object size=%zu, kfree_by_call_rcu=%d\n" , + kfree_mult *sizeof (struct kfree_obj ), + kfree_by_call_rcu); kfree_reader_tasks =kcalloc (kfree_nrealthreads,sizeof (kfree_reader_tasks[0 ]),GFP_KERNEL );
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 6/13 percpu-refcount: Use call_rcu_flush() for atomic switch
In reply to Joel Fernandes (Google)call_rcu() changes to save power will slow down the percpu refcounter's "per-CPU to atomic switch" path. The primitive uses RCU when switching to atomic mode. The enqueued async callback wakes up waiters waiting in the percpu_ref_switch_waitq. Due to this, per-CPU refcount users will slow down, such as blk_pre_runtime_suspend().
Use the call_rcu_flush() API instead which reverts to the old behavior.
--- lib/percpu-refcount.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
lib/percpu-refcount.c +2 -1
@@ -230,7 +230,8 @@ static void __percpu_ref_switch_to_atomic(struct percpu_ref *ref, percpu_ref_noop_confirm_switch;percpu_ref_get (ref);/* put after confirmation */ -call_rcu (&ref->data->rcu, percpu_ref_switch_to_atomic_rcu); +call_rcu_flush (&ref->data->rcu, + percpu_ref_switch_to_atomic_rcu); }static void __percpu_ref_switch_to_percpu (struct percpu_ref *ref)
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 7/13 rcu/sync: Use call_rcu_flush() instead of call_rcu
In reply to Joel Fernandes (Google)call_rcu() changes to save power will slow down rcu sync. Use the call_rcu_flush() API instead which reverts to the old behavior.
--- kernel/rcu/sync.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/rcu/sync.c +1 -1
@@ -44,7 +44,7 @@ static void rcu_sync_func(struct rcu_head *rhp);static void rcu_sync_call (struct rcu_sync *rsp) { -call_rcu (&rsp->cb_head, rcu_sync_func); +call_rcu_flush (&rsp->cb_head, rcu_sync_func); }/**
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 8/13 rcu/rcuscale: Use call_rcu_flush() for async reader test
In reply to Joel Fernandes (Google)rcuscale uses call_rcu() to queue async readers. With recent changes to save power, the test will have fewer async readers in flight. Use the call_rcu_flush() API instead to revert to the old behavior.
--- kernel/rcu/rcuscale.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/rcu/rcuscale.c +1 -1
@@ -176,7 +176,7 @@ static struct rcu_scale_ops rcu_ops = { .get_gp_seq = rcu_get_gp_seq, .gp_diff = rcu_seq_diff, .exp_completed = rcu_exp_batches_completed, - .async = call_rcu, + .async = call_rcu_flush, .gp_barrier = rcu_barrier, .sync = synchronize_rcu, .exp_sync = synchronize_rcu_expedited,
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 9/13 rcu/rcutorture: Use call_rcu_flush() where needed
In reply to Joel Fernandes (Google)call_rcu() changes to save power will change the behavior of rcutorture tests. Use the call_rcu_flush() API instead which reverts to the old behavior.
--- kernel/rcu/rcutorture.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)
kernel/rcu/rcutorture.c +8 -8
@@ -514,7 +514,7 @@ static unsigned long rcu_no_completed(void)static void rcu_torture_deferred_free (struct rcu_torture *p) { -call_rcu (&p->rtort_rcu, rcu_torture_cb); +call_rcu_flush (&p->rtort_rcu, rcu_torture_cb); }static void rcu_sync_torture_init (void )
@@ -559,7 +559,7 @@ static struct rcu_torture_ops rcu_ops = { .start_gp_poll_exp_full = start_poll_synchronize_rcu_expedited_full, .poll_gp_state_exp = poll_state_synchronize_rcu, .cond_sync_exp = cond_synchronize_rcu_expedited, - .call = call_rcu, + .call = call_rcu_flush, .cb_barrier = rcu_barrier, .fqs = rcu_force_quiescent_state, .stats = NULL,
@@ -863,7 +863,7 @@ static void rcu_tasks_torture_deferred_free(struct rcu_torture *p)static void synchronize_rcu_mult_test (void ) { -synchronize_rcu_mult (call_rcu_tasks, call_rcu); +synchronize_rcu_mult (call_rcu_tasks, call_rcu_flush); }static struct rcu_torture_ops tasks_ops = {
@@ -3432,13 +3432,13 @@ static void rcu_test_debug_objects(void)/* Try to queue the rh2 pair of callbacks for the same grace period. */ preempt_disable ();/* Prevent preemption from interrupting test. */ rcu_read_lock ();/* Make it impossible to finish a grace period. */ -call_rcu (&rh1, rcu_torture_leak_cb);/* Start grace period. */ +call_rcu_flush (&rh1, rcu_torture_leak_cb);/* Start grace period. */ local_irq_disable ();/* Make it harder to start a new grace period. */ -call_rcu (&rh2, rcu_torture_leak_cb); -call_rcu (&rh2, rcu_torture_err_cb);/* Duplicate callback. */ +call_rcu_flush (&rh2, rcu_torture_leak_cb); +call_rcu_flush (&rh2, rcu_torture_err_cb);/* Duplicate callback. */ if (rhp) { -call_rcu (rhp, rcu_torture_leak_cb); -call_rcu (rhp, rcu_torture_err_cb);/* Another duplicate callback. */ +call_rcu_flush (rhp, rcu_torture_leak_cb); +call_rcu_flush (rhp, rcu_torture_err_cb);/* Another duplicate callback. */ }local_irq_enable ();rcu_read_unlock ();
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 10/13 scsi/scsi_error: Use call_rcu_flush() instead of call_rcu()
In reply to Joel Fernandes (Google)From: Uladzislau Rezki <[email protected]>
Slow boot time is seen on KVM running typical Linux distributions due to SCSI layer calling call_rcu(). Recent changes to save power may be causing this slowness. Using call_rcu_flush() fixes the issue and brings the boot time back to what it originally was. Convert it.
--- drivers/scsi/scsi_error.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/scsi_error.c +1 -1
@@ -312,7 +312,7 @@ void scsi_eh_scmd_add(struct scsi_cmnd *scmd)* Ensure that all tasks observe the host state change before the * host_failed change. */ -call_rcu (&scmd->rcu, scsi_eh_inc_host_failed); +call_rcu_flush (&scmd->rcu, scsi_eh_inc_host_failed); }/**
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 11/13 workqueue: Make queue_rcu_work() use call_rcu_flush()
In reply to Joel Fernandes (Google)From: Uladzislau Rezki <[email protected]>
call_rcu() changes to save power will slow down RCU workqueue items queued via queue_rcu_work(). This may not be an issue, however we cannot assume that workqueue users are OK with long delays. Use call_rcu_flush() API instead which reverts to the old behavior.
--- kernel/workqueue.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/workqueue.c +1 -1
@@ -1771,7 +1771,7 @@ bool queue_rcu_work(struct workqueue_struct *wq, struct rcu_work *rwork)if (!test_and_set_bit (WORK_STRUCT_PENDING_BIT ,work_data_bits (work))) { rwork->wq = wq; -call_rcu (&rwork->rcu, rcu_work_rcufn); +call_rcu_flush (&rwork->rcu, rcu_work_rcufn);return true; }
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 12/13 rxrpc: Use call_rcu_flush() instead of call_rcu()
In reply to Joel Fernandes (Google)call_rcu() changes to save power may cause slowness. Use the call_rcu_flush() API instead which reverts to the old behavior.
We find this via inspection that the RCU callback does a wakeup of a thread. This usually indicates that something is waiting on it. To be safe, let us use call_rcu_flush() here instead.
--- net/rxrpc/conn_object.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
net/rxrpc/conn_object.c +1 -1
@@ -253,7 +253,7 @@ void rxrpc_kill_connection(struct rxrpc_connection *conn)* must carry a ref on the connection to prevent us getting here whilst * it is queued or running. */ -call_rcu (&conn->rcu, rxrpc_destroy_connection); +call_rcu_flush (&conn->rcu, rxrpc_destroy_connection); }/*
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
PATCH v8 13/13 rcu/debug: Add wake-up debugging for lazy callbacks
In reply to Joel Fernandes (Google)This patch adds initial debugging for lazy callback: whether the callback does a wake up or not. We see that callbacks doing wake ups are usually associated with synchronous use cases (SCSI, rcu_sync, synchronize_rcu() etc).
The code is not very intrusive as almost all the logic is in 'lazy-debug.h' with just a few calls from tree.c
In the future, we will add more functionality such as ensuring callbacks execute in bounded time.
--- kernel/rcu/Kconfig | 7 ++ kernel/rcu/lazy-debug.h | 154 ++++++++++++++++++++++++++++++++++++++++ kernel/rcu/tree.c | 9 +++ 3 files changed, 170 insertions(+) create mode 100644 kernel/rcu/lazy-debug.h
kernel/rcu/Kconfig +7 -0
@@ -322,4 +322,11 @@ config RCU_LAZY To save power, batch RCU callbacks and flush after delay, memory pressure or callback list growing too big. +config RCU_LAZY_DEBUG + bool "RCU callback lazy invocation debugging" + depends on RCU_LAZY + default n + help + Debugging to catch issues caused by delayed RCU callbacks. + endmenu # "RCU Subsystem"
kernel/rcu/lazy-debug.h new file +154 -0
154 lines not on this page · raw message
kernel/rcu/tree.c +9 -0
@@ -67,6 +67,7 @@#include "tree.h" #include "rcu.h" +#include "lazy-debug.h" #ifdef MODULE_PARAM_PREFIX #undef MODULE_PARAM_PREFIX
@@ -2245,7 +2246,10 @@ static void rcu_do_batch(struct rcu_data *rdp) f = rhp->func;WRITE_ONCE (rhp->func, (rcu_callback_t )0L ); + +rcu_set_lazy_context (f);f (rhp); +rcu_reset_lazy_context ();rcu_lock_release (&rcu_callback_map);
@@ -2770,6 +2774,10 @@ __call_rcu_common(struct rcu_head *head, rcu_callback_t func, bool lazy) }check_cb_ovld (rdp); + +if (lazy) +lazy_func_add (func); +if (rcu_nocb_try_bypass (rdp, head, &was_alldone, flags, lazy))return ;// Enqueued onto ->nocb_bypass, so just leave. // If no-CBs CPU gets here, rcu_nocb_try_bypass() acquired ->nocb_lock.
@@ -4800,6 +4808,7 @@ void __init rcu_init(void)rcu_early_boot_tests ();kfree_rcu_batch_init (); +rcu_lazy_debug_init ();rcu_bootup_announce ();sanitize_kthread_prio ();rcu_init_geometry ();
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog