[PATCH 1/2] softmmu/cpus: Check if the cpu work list is empty atomically

[PATCH 1/2] softmmu/cpus: Check if the cpu work list is empty atomically

[Idan Horowitz]

Instead of taking the lock of the cpu work list in order to check if it's
empty, we can just read the head pointer atomically. This decreases
cpu_work_list_empty's share from 5% to 1.3% in a profile of icount-enabled
aarch64-softmmu.

Signed-off-by: Idan Horowitz <idan.horowitz@gmail.com>
---
 softmmu/cpus.c | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/softmmu/cpus.c b/softmmu/cpus.c
index 23bca46b07..035395ae13 100644
--- a/softmmu/cpus.c
+++ b/softmmu/cpus.c
@@ -73,12 +73,7 @@ bool cpu_is_stopped(CPUState *cpu)
 
 bool cpu_work_list_empty(CPUState *cpu)
 {
-    bool ret;
-
-    qemu_mutex_lock(&cpu->work_mutex);
-    ret = QSIMPLEQ_EMPTY(&cpu->work_list);
-    qemu_mutex_unlock(&cpu->work_mutex);
-    return ret;
+    return QSIMPLEQ_EMPTY_ATOMIC(&cpu->work_list);
 }
 
 bool cpu_thread_is_idle(CPUState *cpu)
-- 
2.34.1

[PATCH 2/2] qemu-timer: Skip empty timer lists before locking in qemu_clock_deadline_ns_all

[Idan Horowitz]

This decreases qemu_clock_deadline_ns_all's share from 23.2% to 13% in a
profile of icount-enabled aarch64-softmmu.

Signed-off-by: Idan Horowitz <idan.horowitz@gmail.com>
---
 util/qemu-timer.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/util/qemu-timer.c b/util/qemu-timer.c
index f36c75e594..e56895ef55 100644
--- a/util/qemu-timer.c
+++ b/util/qemu-timer.c
@@ -261,6 +261,9 @@ int64_t qemu_clock_deadline_ns_all(QEMUClockType type, int attr_mask)
     }
 
     QLIST_FOREACH(timer_list, &clock->timerlists, list) {
+        if (!qatomic_read(&timer_list->active_timers)) {
+            continue;
+        }
         qemu_mutex_lock(&timer_list->active_timers_lock);
         ts = timer_list->active_timers;
         /* Skip all external timers */
-- 
2.34.1

Re: [PATCH 2/2] qemu-timer: Skip empty timer lists before locking in qemu_clock_deadline_ns_all

[Richard Henderson]

On 1/14/22 11:43 AM, Idan Horowitz wrote:
> This decreases qemu_clock_deadline_ns_all's share from 23.2% to 13% in a
> profile of icount-enabled aarch64-softmmu.
> 
> Signed-off-by: Idan Horowitz<idan.horowitz@gmail.com>
> ---
>   util/qemu-timer.c | 3 +++
>   1 file changed, 3 insertions(+)

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

r~

Re: [PATCH 2/2] qemu-timer: Skip empty timer lists before locking in qemu_clock_deadline_ns_all

[Idan Horowitz]

On Wed, 26 Jan 2022 at 23:43, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
>
> r~

Hey, I believe this patch might have accidentally been missed?

Thanks,
Idan Horowitz

Re: [PATCH 2/2] qemu-timer: Skip empty timer lists before locking in qemu_clock_deadline_ns_all

[Richard Henderson]

On 6/19/22 01:54, Idan Horowitz wrote:
> On Wed, 26 Jan 2022 at 23:43, Richard Henderson
> <richard.henderson@linaro.org> wrote:
>>
>> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
>>
>> r~
> 
> Hey, I believe this patch might have accidentally been missed?

Oops, yes.  Got 1/2 but not 2/2.  Queued to tcg-next.

r~

Re: [PATCH 1/2] softmmu/cpus: Check if the cpu work list is empty atomically

[Richard Henderson]

On 1/14/22 11:43 AM, Idan Horowitz wrote:
> Instead of taking the lock of the cpu work list in order to check if it's
> empty, we can just read the head pointer atomically. This decreases
> cpu_work_list_empty's share from 5% to 1.3% in a profile of icount-enabled
> aarch64-softmmu.
> 
> Signed-off-by: Idan Horowitz<idan.horowitz@gmail.com>
> ---
>   softmmu/cpus.c | 7 +------
>   1 file changed, 1 insertion(+), 6 deletions(-)

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

Queued to tcg-next.


r~