[REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Minfei Huang]

From: Minfei Huang <mnfhuang@gmail.com>

Since the patch "ftrace: remove daemon(cb7be3b)" remove the function
ftraced, the variant ftrace_update_time never be used any more.

Remove the unused variant ftrace_update_time.

Signed-off-by: Minfei Huang <mnfhuang@gmail.com>
---
 kernel/trace/ftrace.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index b0623ac..d790f28 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -2766,7 +2766,6 @@ static void ftrace_shutdown_sysctl(void)
 	}
 }
 
-static cycle_t		ftrace_update_time;
 unsigned long		ftrace_update_tot_cnt;
 
 static inline int ops_traces_mod(struct ftrace_ops *ops)
@@ -2826,7 +2825,6 @@ static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
 {
 	struct ftrace_page *pg;
 	struct dyn_ftrace *p;
-	cycle_t start, stop;
 	unsigned long update_cnt = 0;
 	unsigned long ref = 0;
 	bool test = false;
@@ -2852,8 +2850,6 @@ static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
 		}
 	}
 
-	start = ftrace_now(raw_smp_processor_id());
-
 	for (pg = new_pgs; pg; pg = pg->next) {
 
 		for (i = 0; i < pg->index; i++) {
@@ -2894,8 +2890,6 @@ static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
 		}
 	}
 
-	stop = ftrace_now(raw_smp_processor_id());
-	ftrace_update_time = stop - start;
 	ftrace_update_tot_cnt += update_cnt;
 
 	return 0;
-- 
2.1.0

Re: [REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Steven Rostedt]

On Tue, 15 Sep 2015 15:10:43 +0800
Minfei Huang <mhuang@redhat.com> wrote:

> From: Minfei Huang <mnfhuang@gmail.com>
> 
> Since the patch "ftrace: remove daemon(cb7be3b)" remove the function
> ftraced, the variant ftrace_update_time never be used any more.
> 
> Remove the unused variant ftrace_update_time.

Actually, the patch I would like to see is this variable being used.
I'd like to know this statistic.

Maybe add it to the dyn_ftrace_total_info file?

-- Steve

Re: [REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Minfei Huang]

On 09/15/15 at 09:35am, Steven Rostedt wrote:
> On Tue, 15 Sep 2015 15:10:43 +0800
> Minfei Huang <mhuang@redhat.com> wrote:
> 
> > From: Minfei Huang <mnfhuang@gmail.com>
> > 
> > Since the patch "ftrace: remove daemon(cb7be3b)" remove the function
> > ftraced, the variant ftrace_update_time never be used any more.
> > 
> > Remove the unused variant ftrace_update_time.
> 
> Actually, the patch I would like to see is this variable being used.
> I'd like to know this statistic.

Hi, Steven.

I think the variant ftrace_update_time is not used any more. Previously,
it is used to print the time how long ftrace will spend to update the
code in fucntion ftraced. 

> 
> Maybe add it to the dyn_ftrace_total_info file?
> 

This file only uses the variant ftrace_update_tot_cnt which is stored
the updated function number.

Thanks
Minfei

Re: [REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Steven Rostedt]

On Tue, 15 Sep 2015 23:37:39 +0800
Minfei Huang <mnfhuang@gmail.com> wrote:

 
> I think the variant ftrace_update_time is not used any more. Previously,
> it is used to print the time how long ftrace will spend to update the
> code in fucntion ftraced. 

Yes, I know what it measures. It's been on my todo list to export that
out to userspace, as it is an interesting metric.

> 
> > 
> > Maybe add it to the dyn_ftrace_total_info file?
> > 
> 
> This file only uses the variant ftrace_update_tot_cnt which is stored
> the updated function number.

But we can add to that file with a space delimited output.

-- Steve

Re: [REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Minfei Huang]

On 09/15/15 at 11:50am, Steven Rostedt wrote:
> On Tue, 15 Sep 2015 23:37:39 +0800
> Minfei Huang <mnfhuang@gmail.com> wrote:
> 
>  
> > I think the variant ftrace_update_time is not used any more. Previously,
> > it is used to print the time how long ftrace will spend to update the
> > code in fucntion ftraced. 
> 
> Yes, I know what it measures. It's been on my todo list to export that
> out to userspace, as it is an interesting metric.
> 

Export the time which ftrace spent to update the all of the function?
How about exporting the time ftrace spent in latest time?

Thanks
Minfei

> > 
> > > 
> > > Maybe add it to the dyn_ftrace_total_info file?
> > > 
> > 
> > This file only uses the variant ftrace_update_tot_cnt which is stored
> > the updated function number.
> 
> But we can add to that file with a space delimited output.
> 
> -- Steve
> 

Re: [REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Steven Rostedt]

On Wed, 16 Sep 2015 00:00:46 +0800
Minfei Huang <mnfhuang@gmail.com> wrote:

> On 09/15/15 at 11:50am, Steven Rostedt wrote:
> > On Tue, 15 Sep 2015 23:37:39 +0800
> > Minfei Huang <mnfhuang@gmail.com> wrote:
> > 
> >  
> > > I think the variant ftrace_update_time is not used any more. Previously,
> > > it is used to print the time how long ftrace will spend to update the
> > > code in fucntion ftraced. 
> > 
> > Yes, I know what it measures. It's been on my todo list to export that
> > out to userspace, as it is an interesting metric.
> > 
> 
> Export the time which ftrace spent to update the all of the function?
> How about exporting the time ftrace spent in latest time?
> 

I think you misunderstood me. I want to export what it is now. The time
it took to run ftrace_update_code() the last time it ran. Not a total
amount of time that it ran for.

-- Steve

Re: [REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Minfei Huang]

On 09/15/15 at 12:12pm, Steven Rostedt wrote:
> On Wed, 16 Sep 2015 00:00:46 +0800
> Minfei Huang <mnfhuang@gmail.com> wrote:
> 
> > On 09/15/15 at 11:50am, Steven Rostedt wrote:
> > > On Tue, 15 Sep 2015 23:37:39 +0800
> > > Minfei Huang <mnfhuang@gmail.com> wrote:
> > > 
> > >  
> > > > I think the variant ftrace_update_time is not used any more. Previously,
> > > > it is used to print the time how long ftrace will spend to update the
> > > > code in fucntion ftraced. 
> > > 
> > > Yes, I know what it measures. It's been on my todo list to export that
> > > out to userspace, as it is an interesting metric.
> > > 
> > 
> > Export the time which ftrace spent to update the all of the function?
> > How about exporting the time ftrace spent in latest time?
> > 
> 
> I think you misunderstood me. I want to export what it is now. The time
> it took to run ftrace_update_code() the last time it ran. Not a total
> amount of time that it ran for.
> 

Got it.

There is one more confusion. Is it valuable to export such info to
userspace? What does user do, if kernel exports this?

Thanks
Minfei

Re: [REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Steven Rostedt]

On Wed, 16 Sep 2015 00:32:02 +0800
Minfei Huang <mnfhuang@gmail.com> wrote:
 
> There is one more confusion. Is it valuable to export such info to
> userspace? What does user do, if kernel exports this?

Nothing. The dyn_ftrace_total_info is purely for debugging ftrace. It's
something I use.

-- Steve

Re: [REPOST PATCH] ftrace: Remove the unused variant ftrace_update_time

[Minfei Huang]

On 09/15/15 at 01:01pm, Steven Rostedt wrote:
> On Wed, 16 Sep 2015 00:32:02 +0800
> Minfei Huang <mnfhuang@gmail.com> wrote:
>  
> > There is one more confusion. Is it valuable to export such info to
> > userspace? What does user do, if kernel exports this?
> 
> Nothing. The dyn_ftrace_total_info is purely for debugging ftrace. It's
> something I use.
> 

hmmm...

Thanks. I don't insist for this patch, if you want to export this
variant to userspace.

Thanks
Minfei