tracing: ftrace_lib: check if tracing_max_latency exists.

Message ID 20180808234019.241238-1-sspatil@google.com
State Accepted
Headers show
Series
  • tracing: ftrace_lib: check if tracing_max_latency exists.
Related show

Commit Message

Sandeep Patil Aug. 8, 2018, 11:40 p.m.
This avoids unnecessary error in ftrace stress test when the
kernel being tested does not have CONFIG_TRACER_MAX_TRACE defined.

Signed-off-by: Sandeep Patil <sspatil@google.com>
---
 testcases/kernel/tracing/ftrace_test/ftrace_lib.sh | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

Comments

Cyril Hrubis Aug. 10, 2018, 12:57 p.m. | #1
Hi!
> This avoids unnecessary error in ftrace stress test when the
> kernel being tested does not have CONFIG_TRACER_MAX_TRACE defined.
> 
> Signed-off-by: Sandeep Patil <sspatil@google.com>
> ---
>  testcases/kernel/tracing/ftrace_test/ftrace_lib.sh | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh b/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
> index 20cf5e760..ec4f7a557 100755
> --- a/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
> +++ b/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
> @@ -100,7 +100,9 @@ restore_old_setting()
>  
>  	echo nop > current_tracer
>  	echo 0 > events/enable
> -	echo 0 > tracing_max_latency 2> /dev/null
> +	if [ -e tracing_max_latency ]; then
> +		echo 0 > tracing_max_latency 2> /dev/null
> +	fi

I guess that we can remove the 2> /dev/null that I suppose was supposed
to silence the error. Does it work okay for you without the redirection?
Sandeep Patil Aug. 13, 2018, 3:46 p.m. | #2
On Fri, Aug 10, 2018 at 02:57:41PM +0200, Cyril Hrubis wrote:
> Hi!
> > This avoids unnecessary error in ftrace stress test when the
> > kernel being tested does not have CONFIG_TRACER_MAX_TRACE defined.
> > 
> > Signed-off-by: Sandeep Patil <sspatil@google.com>
> > ---
> >  testcases/kernel/tracing/ftrace_test/ftrace_lib.sh | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> > 
> > diff --git a/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh b/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
> > index 20cf5e760..ec4f7a557 100755
> > --- a/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
> > +++ b/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
> > @@ -100,7 +100,9 @@ restore_old_setting()
> >  
> >  	echo nop > current_tracer
> >  	echo 0 > events/enable
> > -	echo 0 > tracing_max_latency 2> /dev/null
> > +	if [ -e tracing_max_latency ]; then
> > +		echo 0 > tracing_max_latency 2> /dev/null
> > +	fi
> 
> I guess that we can remove the 2> /dev/null that I suppose was supposed
> to silence the error. Does it work okay for you without the redirection?

Yes, just tested it. Works for me. Want me to send the patch removing it?

- ssp
> 
> -- 
> Cyril Hrubis
> chrubis@suse.cz
Cyril Hrubis Aug. 13, 2018, 3:48 p.m. | #3
Hi!
> > I guess that we can remove the 2> /dev/null that I suppose was supposed
> > to silence the error. Does it work okay for you without the redirection?
> 
> Yes, just tested it. Works for me. Want me to send the patch removing it?

No need, I will edit the patch you send before applying.

Patch

diff --git a/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh b/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
index 20cf5e760..ec4f7a557 100755
--- a/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
+++ b/testcases/kernel/tracing/ftrace_test/ftrace_lib.sh
@@ -100,7 +100,9 @@  restore_old_setting()
 
 	echo nop > current_tracer
 	echo 0 > events/enable
-	echo 0 > tracing_max_latency 2> /dev/null
+	if [ -e tracing_max_latency ]; then
+		echo 0 > tracing_max_latency 2> /dev/null
+	fi
 
 	if [ -e tracing_cpumask ]; then
 		echo $old_tracing_cpumask > tracing_cpumask