[PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[Tio Zhang]

Switch the order of prev_comm and next_comm in sched_switch's code to
align with its printing order.

Signed-off-by: Tio Zhang <tiozhang@didiglobal.com>
---
 include/trace/events/sched.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
index dbb01b4b7451..a4bd4330db4c 100644
--- a/include/trace/events/sched.h
+++ b/include/trace/events/sched.h
@@ -239,11 +239,11 @@ TRACE_EVENT(sched_switch,
 	),
 
 	TP_fast_assign(
-		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
+		memcpy(__entry->prev_comm, prev->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);
-		memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
+		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
 		__entry->next_pid	= next->pid;
 		__entry->next_prio	= next->prio;
 		/* XXX SCHED_DEADLINE */
-- 
2.17.1

Re: [PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[Madadi Vineeth Reddy]

Hi Tio zhang,

On 17/04/24 15:59, Tio Zhang wrote:
> Switch the order of prev_comm and next_comm in sched_switch's code to
> align with its printing order.
> 
> Signed-off-by: Tio Zhang <tiozhang@didiglobal.com>
> ---
>  include/trace/events/sched.h | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
> index dbb01b4b7451..a4bd4330db4c 100644
> --- a/include/trace/events/sched.h
> +++ b/include/trace/events/sched.h
> @@ -239,11 +239,11 @@ TRACE_EVENT(sched_switch,
>  	),
>  
>  	TP_fast_assign(
> -		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
> +		memcpy(__entry->prev_comm, prev->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);
> -		memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
> +		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
>  		__entry->next_pid	= next->pid;
>  		__entry->next_prio	= next->prio;
>  		/* XXX SCHED_DEADLINE */

Yes, makes sense to have all previous and then the next for increased readability.

Reviewed-by: Madadi Vineeth Reddy <vineethr@linux.ibm.com>

Re: [PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[张元瀚 Tio Zhang]

Hi Steven, 

Ping :) 

On 4/23/24 下午6:24, "Madadi Vineeth Reddy" <vineethr@linux.ibm.com <mailto:vineethr@linux.ibm.com>> wrote:


Hi Tio zhang,


On 17/04/24 15:59, Tio Zhang wrote:
> Switch the order of prev_comm and next_comm in sched_switch's code to
> align with its printing order.
> 
> Signed-off-by: Tio Zhang <tiozhang@didiglobal.com <mailto:tiozhang@didiglobal.com>>
> ---
> include/trace/events/sched.h | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
> index dbb01b4b7451..a4bd4330db4c 100644
> --- a/include/trace/events/sched.h
> +++ b/include/trace/events/sched.h
> @@ -239,11 +239,11 @@ TRACE_EVENT(sched_switch,
> ),
> 
> TP_fast_assign(
> - memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
> + memcpy(__entry->prev_comm, prev->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);
> - memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
> + memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
> __entry->next_pid = next->pid;
> __entry->next_prio = next->prio;
> /* XXX SCHED_DEADLINE */


Yes, makes sense to have all previous and then the next for increased readability.


Reviewed-by: Madadi Vineeth Reddy <vineethr@linux.ibm.com <mailto:vineethr@linux.ibm.com>>

Re: [PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[Steven Rostedt]

On Tue, 2 Jul 2024 07:14:33 +0000
张元瀚 Tio Zhang <tiozhang@didiglobal.com> wrote:

> Hi Steven, 
> 
> Ping :) 

If you want me to take this, please resend it and add
linux-trace-kernel@vger.kernel.org to the Cc so that my patchwork can
pick it up.

Also add anyone's tags to it as well.

-- Steve

[PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[Tio Zhang]

Switch the order of prev_comm and next_comm in sched_switch's code to
align with its printing order.

Signed-off-by: Tio Zhang <tiozhang@didiglobal.com>
Reviewed-by: Madadi Vineeth Reddy <vineethr@linux.ibm.com>
---
 include/trace/events/sched.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
index dbb01b4b7451..a4bd4330db4c 100644
--- a/include/trace/events/sched.h
+++ b/include/trace/events/sched.h
@@ -239,11 +239,11 @@ TRACE_EVENT(sched_switch,
 	),
 
 	TP_fast_assign(
-		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
+		memcpy(__entry->prev_comm, prev->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);
-		memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
+		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
 		__entry->next_pid	= next->pid;
 		__entry->next_prio	= next->prio;
 		/* XXX SCHED_DEADLINE */
-- 
2.17.1

Re: [PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[张元瀚 Tio Zhang]

Hi Steven, 

Thank you, I resend this patch with linux-trace-kernel@vger.kernel.org on the cc list, please see https://lore.kernel.org/lkml/20240703033353.GA2833@didi-ThinkCentre-M930t-N000/#t

On 7/3/24 上午12:29, "Steven Rostedt" <rostedt@goodmis.org <mailto:rostedt@goodmis.org>> wrote:


On Tue, 2 Jul 2024 07:14:33 +0000
张元瀚 Tio Zhang <tiozhang@didiglobal.com <mailto:tiozhang@didiglobal.com>> wrote:


> Hi Steven, 
> 
> Ping :) 


If you want me to take this, please resend it and add
linux-trace-kernel@vger.kernel.org <mailto:linux-trace-kernel@vger.kernel.org> to the Cc so that my patchwork can
pick it up.


Also add anyone's tags to it as well.


-- Steve

Re: [PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[Steven Rostedt]

[ Adding sched maintainers, as this is a scheduling trace event ]

On Wed, 3 Jul 2024 11:33:53 +0800
Tio Zhang <tiozhang@didiglobal.com> wrote:

> Switch the order of prev_comm and next_comm in sched_switch's code to
> align with its printing order.

I'm going to pick this up in my tree, as it is pretty much a nop. It's
just changing the ordering of the assignments of the comm names so that
the copy of prev_comm is with the updates to the prev_fields, and the
next_comm is with the updates to the next fields.

-- Steve


> 
> Signed-off-by: Tio Zhang <tiozhang@didiglobal.com>
> Reviewed-by: Madadi Vineeth Reddy <vineethr@linux.ibm.com>
> ---
>  include/trace/events/sched.h | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
> index dbb01b4b7451..a4bd4330db4c 100644
> --- a/include/trace/events/sched.h
> +++ b/include/trace/events/sched.h
> @@ -239,11 +239,11 @@ TRACE_EVENT(sched_switch,
>  	),
>  
>  	TP_fast_assign(
> -		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
> +		memcpy(__entry->prev_comm, prev->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);
> -		memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
> +		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
>  		__entry->next_pid	= next->pid;
>  		__entry->next_prio	= next->prio;
>  		/* XXX SCHED_DEADLINE */

Re: [PATCH 1/1] tracing/sched: sched_switch: place prev_comm and next_comm in right order

[Peter Zijlstra]

On Mon, Jul 15, 2024 at 03:04:12PM -0400, Steven Rostedt wrote:
> 
> [ Adding sched maintainers, as this is a scheduling trace event ]
> 
> On Wed, 3 Jul 2024 11:33:53 +0800
> Tio Zhang <tiozhang@didiglobal.com> wrote:
> 
> > Switch the order of prev_comm and next_comm in sched_switch's code to
> > align with its printing order.
> 
> I'm going to pick this up in my tree, as it is pretty much a nop. It's
> just changing the ordering of the assignments of the comm names so that
> the copy of prev_comm is with the updates to the prev_fields, and the
> next_comm is with the updates to the next fields.

Right, have at. very close to a whitespace patch.