[PATCH v2] lib/raid6: Reduce high latency by using migrate instead of preempt

We found an abnormally high latency when executing modprobe raid6_pq, the
latency is greater than 1.2s when CONFIG_PREEMPT_VOLUNTARY=y, greater than
67ms when CONFIG_PREEMPT=y, and greater than 16ms when CONFIG_PREEMPT_RT=y.

This is caused by ksoftirqd fail to scheduled due to disable preemption,
this time is too long and unreasonable.

Reduce high latency by using migrate_disabl()/emigrate_enable() instead of
preempt_disable()/preempt_enable().

How to reproduce:
 - Install cyclictest
     sudo apt install rt-tests
 - Run cyclictest example in one terminal
     sudo cyclictest -S -p 95 -d 0 -i 1000 -D 24h -m
 - Modprobe raid6_pq in another terminal
      sudo modprobe raid6_pq

This patch beneficial for CONFIG_PREEMPT=y and CONFIG_PREEMPT_RT=y, but
no effect for CONFIG_PREEMPT_VOLUNTARY=y.

Fixes: fe5cbc6e06c7 ("md/raid6 algorithms: delta syndrome functions")
Fixes: cc4589ebfae6 ("Rename raid6 files now they're in a 'raid6' directory.")
Link: https://lore.kernel.org/linux-raid/b06c5e3ef3413f12a2c2b2a241005af9@linux.dev/T/#t
Signed-off-by: Yajun Deng <redacted>
---
 lib/raid6/algos.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/raid6/algos.c b/lib/raid6/algos.c
index 6d5e5000fdd7..21611d05c34c 100644
--- a/lib/raid6/algos.c
+++ b/lib/raid6/algos.c

@@ -162,7 +162,7 @@ static inline const struct raid6_calls *raid6_choose_gen(
 
 			perf = 0;
 
-			preempt_disable();
+			migrate_disable();
 			j0 = jiffies;
 			while ((j1 = jiffies) == j0)
 				cpu_relax();

@@ -171,7 +171,7 @@ static inline const struct raid6_calls *raid6_choose_gen(
 				(*algo)->gen_syndrome(disks, PAGE_SIZE, *dptrs);
 				perf++;
 			}
-			preempt_enable();
+			migrate_enable();
 
 			if (perf > bestgenperf) {
 				bestgenperf = perf;

@@ -186,7 +186,7 @@ static inline const struct raid6_calls *raid6_choose_gen(
 
 			perf = 0;
 
-			preempt_disable();
+			migrate_disable();
 			j0 = jiffies;
 			while ((j1 = jiffies) == j0)
 				cpu_relax();

@@ -196,7 +196,7 @@ static inline const struct raid6_calls *raid6_choose_gen(
 						      PAGE_SIZE, *dptrs);
 				perf++;
 			}
-			preempt_enable();
+			migrate_enable();
 
 			if (best == *algo)
 				bestxorperf = perf;

-- 
2.32.0

On Thu, Dec 16, 2021 at 1:27 AM Yajun Deng [off-list ref] wrote:

We found an abnormally high latency when executing modprobe raid6_pq, the
latency is greater than 1.2s when CONFIG_PREEMPT_VOLUNTARY=y, greater than
67ms when CONFIG_PREEMPT=y, and greater than 16ms when CONFIG_PREEMPT_RT=y.

This is caused by ksoftirqd fail to scheduled due to disable preemption,
this time is too long and unreasonable.

Reduce high latency by using migrate_disabl()/emigrate_enable() instead of
preempt_disable()/preempt_enable().

How to reproduce:
 - Install cyclictest
     sudo apt install rt-tests
 - Run cyclictest example in one terminal
     sudo cyclictest -S -p 95 -d 0 -i 1000 -D 24h -m
 - Modprobe raid6_pq in another terminal
      sudo modprobe raid6_pq

This patch beneficial for CONFIG_PREEMPT=y and CONFIG_PREEMPT_RT=y, but
no effect for CONFIG_PREEMPT_VOLUNTARY=y.

Fixes: fe5cbc6e06c7 ("md/raid6 algorithms: delta syndrome functions")
Fixes: cc4589ebfae6 ("Rename raid6 files now they're in a 'raid6' directory.")
Link: https://lore.kernel.org/linux-raid/b06c5e3ef3413f12a2c2b2a241005af9@linux.dev/T/#t
Signed-off-by: Yajun Deng <redacted>

Updated the patch and CC'ed stable. It should make to LTS kernels.

Thanks,
Song

NACK. This is a BUG.

--nX

On Fri, Dec 17, 2021 at 1:02 AM Song Liu [off-list ref] wrote:

On Thu, Dec 16, 2021 at 1:27 AM Yajun Deng [off-list ref] wrote:

quoted

We found an abnormally high latency when executing modprobe raid6_pq, the
latency is greater than 1.2s when CONFIG_PREEMPT_VOLUNTARY=y, greater than
67ms when CONFIG_PREEMPT=y, and greater than 16ms when CONFIG_PREEMPT_RT=y.

This is caused by ksoftirqd fail to scheduled due to disable preemption,
this time is too long and unreasonable.

Reduce high latency by using migrate_disabl()/emigrate_enable() instead of
preempt_disable()/preempt_enable().

How to reproduce:
 - Install cyclictest
     sudo apt install rt-tests
 - Run cyclictest example in one terminal
     sudo cyclictest -S -p 95 -d 0 -i 1000 -D 24h -m
 - Modprobe raid6_pq in another terminal
      sudo modprobe raid6_pq

This patch beneficial for CONFIG_PREEMPT=y and CONFIG_PREEMPT_RT=y, but
no effect for CONFIG_PREEMPT_VOLUNTARY=y.

Fixes: fe5cbc6e06c7 ("md/raid6 algorithms: delta syndrome functions")
Fixes: cc4589ebfae6 ("Rename raid6 files now they're in a 'raid6' directory.")
Link: https://lore.kernel.org/linux-raid/b06c5e3ef3413f12a2c2b2a241005af9@linux.dev/T/#t
Signed-off-by: Yajun Deng <redacted>

Updated the patch and CC'ed stable. It should make to LTS kernels.

Thanks,
Song