* [PATCH] block,bfq: Skip tracing hooks if possible
@ 2019-11-01 13:11 Dmitry Monakhov
2019-11-15 6:47 ` Paolo Valente
2019-11-20 23:11 ` Jens Axboe
0 siblings, 2 replies; 3+ messages in thread
From: Dmitry Monakhov @ 2019-11-01 13:11 UTC (permalink / raw)
To: linux-kernel; +Cc: axboe, paolo.valente, Dmitry Monakhov
In most cases blk_tracing is not active, but bfq_log_bfqq macro
generate pid_str unconditionally, which result in significant overhead.
## Test
modprobe null_blk
echo bfq > /sys/block/nullb0/queue/scheduler
fio --name=t --ioengine=libaio --direct=1 --filename=/dev/nullb0 \
--runtime=30 --time_based=1 --rw=write --iodepth=128 --bs=4k
# Results
| | baseline | w/ patch | gain |
| iops | 113.19K | 126.42K | +11% |
Signed-off-by: Dmitry Monakhov <dmonakhov@gmail.com>
---
block/bfq-iosched.h | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/block/bfq-iosched.h b/block/bfq-iosched.h
index 5d1a519..b320fe9 100644
--- a/block/bfq-iosched.h
+++ b/block/bfq-iosched.h
@@ -1062,6 +1062,8 @@ struct bfq_group *bfqq_group(struct bfq_queue *bfqq);
#define bfq_log_bfqq(bfqd, bfqq, fmt, args...) do { \
char pid_str[MAX_PID_STR_LENGTH]; \
+ if (likely(!blk_trace_note_message_enabled((bfqd)->queue))) \
+ break; \
bfq_pid_to_str((bfqq)->pid, pid_str, MAX_PID_STR_LENGTH); \
blk_add_cgroup_trace_msg((bfqd)->queue, \
bfqg_to_blkg(bfqq_group(bfqq))->blkcg, \
@@ -1078,6 +1080,8 @@ struct bfq_group *bfqq_group(struct bfq_queue *bfqq);
#define bfq_log_bfqq(bfqd, bfqq, fmt, args...) do { \
char pid_str[MAX_PID_STR_LENGTH]; \
+ if (likely(!blk_trace_note_message_enabled((bfqd)->queue))) \
+ break; \
bfq_pid_to_str((bfqq)->pid, pid_str, MAX_PID_STR_LENGTH); \
blk_add_trace_msg((bfqd)->queue, "bfq%s%c " fmt, pid_str, \
bfq_bfqq_sync((bfqq)) ? 'S' : 'A', \
--
2.7.4
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] block,bfq: Skip tracing hooks if possible
2019-11-01 13:11 [PATCH] block,bfq: Skip tracing hooks if possible Dmitry Monakhov
@ 2019-11-15 6:47 ` Paolo Valente
2019-11-20 23:11 ` Jens Axboe
1 sibling, 0 replies; 3+ messages in thread
From: Paolo Valente @ 2019-11-15 6:47 UTC (permalink / raw)
To: Dmitry Monakhov; +Cc: linux-kernel, axboe
> Il giorno 1 nov 2019, alle ore 14:11, Dmitry Monakhov <dmonakhov@gmail.com> ha scritto:
>
> In most cases blk_tracing is not active, but bfq_log_bfqq macro
> generate pid_str unconditionally, which result in significant overhead.
>
> ## Test
> modprobe null_blk
> echo bfq > /sys/block/nullb0/queue/scheduler
> fio --name=t --ioengine=libaio --direct=1 --filename=/dev/nullb0 \
> --runtime=30 --time_based=1 --rw=write --iodepth=128 --bs=4k
>
> # Results
> | | baseline | w/ patch | gain |
> | iops | 113.19K | 126.42K | +11% |
>
Thank you very much for this simple yet effective improvement.
Acked-by: Paolo Valente <paolo.valente@linaro.org>
> Signed-off-by: Dmitry Monakhov <dmonakhov@gmail.com>
> ---
> block/bfq-iosched.h | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/block/bfq-iosched.h b/block/bfq-iosched.h
> index 5d1a519..b320fe9 100644
> --- a/block/bfq-iosched.h
> +++ b/block/bfq-iosched.h
> @@ -1062,6 +1062,8 @@ struct bfq_group *bfqq_group(struct bfq_queue *bfqq);
>
> #define bfq_log_bfqq(bfqd, bfqq, fmt, args...) do { \
> char pid_str[MAX_PID_STR_LENGTH]; \
> + if (likely(!blk_trace_note_message_enabled((bfqd)->queue))) \
> + break; \
> bfq_pid_to_str((bfqq)->pid, pid_str, MAX_PID_STR_LENGTH); \
> blk_add_cgroup_trace_msg((bfqd)->queue, \
> bfqg_to_blkg(bfqq_group(bfqq))->blkcg, \
> @@ -1078,6 +1080,8 @@ struct bfq_group *bfqq_group(struct bfq_queue *bfqq);
>
> #define bfq_log_bfqq(bfqd, bfqq, fmt, args...) do { \
> char pid_str[MAX_PID_STR_LENGTH]; \
> + if (likely(!blk_trace_note_message_enabled((bfqd)->queue))) \
> + break; \
> bfq_pid_to_str((bfqq)->pid, pid_str, MAX_PID_STR_LENGTH); \
> blk_add_trace_msg((bfqd)->queue, "bfq%s%c " fmt, pid_str, \
> bfq_bfqq_sync((bfqq)) ? 'S' : 'A', \
> --
> 2.7.4
>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] block,bfq: Skip tracing hooks if possible
2019-11-01 13:11 [PATCH] block,bfq: Skip tracing hooks if possible Dmitry Monakhov
2019-11-15 6:47 ` Paolo Valente
@ 2019-11-20 23:11 ` Jens Axboe
1 sibling, 0 replies; 3+ messages in thread
From: Jens Axboe @ 2019-11-20 23:11 UTC (permalink / raw)
To: Dmitry Monakhov, linux-kernel; +Cc: paolo.valente
On 11/1/19 7:11 AM, Dmitry Monakhov wrote:
> In most cases blk_tracing is not active, but bfq_log_bfqq macro
> generate pid_str unconditionally, which result in significant overhead.
>
> ## Test
> modprobe null_blk
> echo bfq > /sys/block/nullb0/queue/scheduler
> fio --name=t --ioengine=libaio --direct=1 --filename=/dev/nullb0 \
> --runtime=30 --time_based=1 --rw=write --iodepth=128 --bs=4k
>
> # Results
> | | baseline | w/ patch | gain |
> | iops | 113.19K | 126.42K | +11% |
Applied, thanks.
--
Jens Axboe
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2019-11-21 1:59 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-11-01 13:11 [PATCH] block,bfq: Skip tracing hooks if possible Dmitry Monakhov
2019-11-15 6:47 ` Paolo Valente
2019-11-20 23:11 ` Jens Axboe
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).