2023-08-02 13:32:51

by Ze Gao

Subject: [RFC PATCH v4 4/7] sched, tracing: reorganize fields of switch event struct

From: Ze Gao <[email protected]>

Report priorities in 'short' and prev_state in 'int' to save
some buffer space. And also reorder the fields so that we take
struct alignment into consideration to make the record compact.

Suggested-by: Steven Rostedt (Google) <[email protected]>
Signed-off-by: Ze Gao <[email protected]>
---
include/trace/events/sched.h | 24 ++++++++++++------------
1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
index fbb99a61f714..7d34db20b2c6 100644
--- a/include/trace/events/sched.h
+++ b/include/trace/events/sched.h
@@ -187,7 +187,7 @@ DEFINE_EVENT(sched_wakeup_template, sched_wakeup_new,
TP_ARGS(p));

#ifdef CREATE_TRACE_POINTS
-static inline long __trace_sched_switch_state(bool preempt,
+static inline int __trace_sched_switch_state(bool preempt,
unsigned int prev_state,
struct task_struct *p)
{
@@ -229,23 +229,23 @@ TRACE_EVENT(sched_switch,
TP_ARGS(preempt, prev, next, prev_state),

TP_STRUCT__entry(
- __array( char, prev_comm, TASK_COMM_LEN )
__field( pid_t, prev_pid )
- __field( int, prev_prio )
- __field( long, prev_state )
- __array( char, next_comm, TASK_COMM_LEN )
__field( pid_t, next_pid )
- __field( int, next_prio )
+ __field( short, prev_prio )
+ __field( short, next_prio )
+ __field( int, prev_state )
+ __array( char, prev_comm, TASK_COMM_LEN )
+ __array( char, next_comm, TASK_COMM_LEN )
),

TP_fast_assign(
- memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
- __entry->prev_pid = prev->pid;
- __entry->prev_prio = prev->prio;
- __entry->prev_state = __trace_sched_switch_state(preempt, prev_state, prev);
+ __entry->prev_pid = prev->pid;
+ __entry->next_pid = next->pid;
+ __entry->prev_prio = (short) prev->prio;
+ __entry->next_prio = (short) next->prio;
+ __entry->prev_state = __trace_sched_switch_state(preempt, prev_state, prev);
memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
- __entry->next_pid = next->pid;
- __entry->next_prio = next->prio;
+ memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
/* XXX SCHED_DEADLINE */
),

--
2.41.0

2023-08-02 15:16:15

by Steven Rostedt

Subject: Re: [RFC PATCH v4 4/7] sched, tracing: reorganize fields of switch event struct

On Wed, 2 Aug 2023 08:09:59 -0400
Ze Gao <[email protected]> wrote:

> From: Ze Gao <[email protected]>
>
> Report priorities in 'short' and prev_state in 'int' to save
> some buffer space. And also reorder the fields so that we take
> struct alignment into consideration to make the record compact.
>
> Suggested-by: Steven Rostedt (Google) <[email protected]>
> Signed-off-by: Ze Gao <[email protected]>
> ---
> include/trace/events/sched.h | 24 ++++++++++++------------
> 1 file changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
> index fbb99a61f714..7d34db20b2c6 100644
> --- a/include/trace/events/sched.h
> +++ b/include/trace/events/sched.h
> @@ -187,7 +187,7 @@ DEFINE_EVENT(sched_wakeup_template, sched_wakeup_new,
> TP_ARGS(p));
>
> #ifdef CREATE_TRACE_POINTS
> -static inline long __trace_sched_switch_state(bool preempt,
> +static inline int __trace_sched_switch_state(bool preempt,
> unsigned int prev_state,
> struct task_struct *p)
> {
> @@ -229,23 +229,23 @@ TRACE_EVENT(sched_switch,
> TP_ARGS(preempt, prev, next, prev_state),
>
> TP_STRUCT__entry(
> - __array( char, prev_comm, TASK_COMM_LEN )
> __field( pid_t, prev_pid )
> - __field( int, prev_prio )
> - __field( long, prev_state )
> - __array( char, next_comm, TASK_COMM_LEN )
> __field( pid_t, next_pid )
> - __field( int, next_prio )
> + __field( short, prev_prio )
> + __field( short, next_prio )
> + __field( int, prev_state )

I was talking with Peter on IRC and since the biggest number that
prev_state can be is TASK_REPORT_MASK which is 0x100, I think we can make
prev_state into a short as well.

-- Steve


> + __array( char, prev_comm, TASK_COMM_LEN )
> + __array( char, next_comm, TASK_COMM_LEN )
> ),
>
> TP_fast_assign(
> - memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
> - __entry->prev_pid = prev->pid;
> - __entry->prev_prio = prev->prio;
> - __entry->prev_state = __trace_sched_switch_state(preempt, prev_state, prev);
> + __entry->prev_pid = prev->pid;
> + __entry->next_pid = next->pid;
> + __entry->prev_prio = (short) prev->prio;
> + __entry->next_prio = (short) next->prio;
> + __entry->prev_state = __trace_sched_switch_state(preempt, prev_state, prev);
> memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
> - __entry->next_pid = next->pid;
> - __entry->next_prio = next->prio;
> + memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
> /* XXX SCHED_DEADLINE */
> ),
>

2023-08-03 03:57:48

by Ze Gao

Subject: Re: [RFC PATCH v4 4/7] sched, tracing: reorganize fields of switch event struct

On Wed, Aug 2, 2023 at 11:05 PM Steven Rostedt <[email protected]> wrote:
>
> On Wed, 2 Aug 2023 08:09:59 -0400
> Ze Gao <[email protected]> wrote:
>
> > From: Ze Gao <[email protected]>
> >
> > Report priorities in 'short' and prev_state in 'int' to save
> > some buffer space. And also reorder the fields so that we take
> > struct alignment into consideration to make the record compact.
> >
> > Suggested-by: Steven Rostedt (Google) <[email protected]>
> > Signed-off-by: Ze Gao <[email protected]>
> > ---
> > include/trace/events/sched.h | 24 ++++++++++++------------
> > 1 file changed, 12 insertions(+), 12 deletions(-)
> >
> > diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
> > index fbb99a61f714..7d34db20b2c6 100644
> > --- a/include/trace/events/sched.h
> > +++ b/include/trace/events/sched.h
> > @@ -187,7 +187,7 @@ DEFINE_EVENT(sched_wakeup_template, sched_wakeup_new,
> > TP_ARGS(p));
> >
> > #ifdef CREATE_TRACE_POINTS
> > -static inline long __trace_sched_switch_state(bool preempt,
> > +static inline int __trace_sched_switch_state(bool preempt,
> > unsigned int prev_state,
> > struct task_struct *p)
> > {
> > @@ -229,23 +229,23 @@ TRACE_EVENT(sched_switch,
> > TP_ARGS(preempt, prev, next, prev_state),
> >
> > TP_STRUCT__entry(
> > - __array( char, prev_comm, TASK_COMM_LEN )
> > __field( pid_t, prev_pid )
> > - __field( int, prev_prio )
> > - __field( long, prev_state )
> > - __array( char, next_comm, TASK_COMM_LEN )
> > __field( pid_t, next_pid )
> > - __field( int, next_prio )
> > + __field( short, prev_prio )
> > + __field( short, next_prio )
> > + __field( int, prev_state )
>
> I was talking with Peter on IRC and since the biggest number that
> prev_state can be is TASK_REPORT_MASK which is 0x100, I think we can make
> prev_state into a short as well.

Make sense, let's change it to short.

Regards,
Ze