[PATCH] kcsan: Prefer '__no_kcsan inline' in test

[PATCH] kcsan: Prefer '__no_kcsan inline' in test

[Marco Elver]

Instead of __no_kcsan_or_inline, prefer '__no_kcsan inline' in test --
this is in case we decide to remove __no_kcsan_or_inline.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Marco Elver <elver@google.com>
---

Hi Paul,

This is to prepare eventual removal of __no_kcsan_or_inline, and avoid a
series that doesn't apply to anything other than -next (because some
bits are in -tip and the test only in -rcu; although this problem might
be solved in 2 weeks). This patch is to make sure in case the
__kcsan_or_inline series is based on -tip, integration in -next doesn't
cause problems.

This came up in
https://lkml.kernel.org/r/20200529185923.GO706495@hirez.programming.kicks-ass.net

Thanks,
-- Marco

---
 kernel/kcsan/kcsan-test.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/kcsan/kcsan-test.c b/kernel/kcsan/kcsan-test.c
index a8c11506dd2a..3af420ad6ee7 100644
--- a/kernel/kcsan/kcsan-test.c
+++ b/kernel/kcsan/kcsan-test.c
@@ -43,7 +43,7 @@ static struct {
 };
 
 /* Setup test checking loop. */
-static __no_kcsan_or_inline void
+static __no_kcsan inline void
 begin_test_checks(void (*func1)(void), void (*func2)(void))
 {
 	kcsan_disable_current();
@@ -60,7 +60,7 @@ begin_test_checks(void (*func1)(void), void (*func2)(void))
 }
 
 /* End test checking loop. */
-static __no_kcsan_or_inline bool
+static __no_kcsan inline bool
 end_test_checks(bool stop)
 {
 	if (!stop && time_before(jiffies, end_time)) {
-- 
2.27.0.rc2.251.g90737beb825-goog

Re: [PATCH] kcsan: Prefer '__no_kcsan inline' in test

[Peter Zijlstra]

On Tue, Jun 02, 2020 at 04:36:33PM +0200, Marco Elver wrote:
> Instead of __no_kcsan_or_inline, prefer '__no_kcsan inline' in test --
> this is in case we decide to remove __no_kcsan_or_inline.
> 
> Suggested-by: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Marco Elver <elver@google.com>
> ---
> 
> Hi Paul,
> 
> This is to prepare eventual removal of __no_kcsan_or_inline, and avoid a
> series that doesn't apply to anything other than -next (because some
> bits are in -tip and the test only in -rcu; although this problem might
> be solved in 2 weeks). This patch is to make sure in case the
> __kcsan_or_inline series is based on -tip, integration in -next doesn't
> cause problems.
> 
> This came up in
> https://lkml.kernel.org/r/20200529185923.GO706495@hirez.programming.kicks-ass.net

Thanks Marco!

I just sent the rest of that patch here:

  https://lkml.kernel.org/r/20200602173103.931412766@infradead.org

Re: [PATCH] kcsan: Prefer '__no_kcsan inline' in test

[Paul E. McKenney]

On Tue, Jun 02, 2020 at 04:36:33PM +0200, Marco Elver wrote:
> Instead of __no_kcsan_or_inline, prefer '__no_kcsan inline' in test --
> this is in case we decide to remove __no_kcsan_or_inline.
> 
> Suggested-by: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Marco Elver <elver@google.com>
> ---
> 
> Hi Paul,
> 
> This is to prepare eventual removal of __no_kcsan_or_inline, and avoid a
> series that doesn't apply to anything other than -next (because some
> bits are in -tip and the test only in -rcu; although this problem might
> be solved in 2 weeks). This patch is to make sure in case the
> __kcsan_or_inline series is based on -tip, integration in -next doesn't
> cause problems.
> 
> This came up in
> https://lkml.kernel.org/r/20200529185923.GO706495@hirez.programming.kicks-ass.net

Applied and pushed, thank you!

Please note that unless you would like this pushed into the current
merge window, it will not be visible in -next until v5.8-rc1 comes out.
Which sounds like you are aware of already, just want to be sure.  ;-)

							Thanx, Paul

> Thanks,
> -- Marco
> 
> ---
>  kernel/kcsan/kcsan-test.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/kcsan/kcsan-test.c b/kernel/kcsan/kcsan-test.c
> index a8c11506dd2a..3af420ad6ee7 100644
> --- a/kernel/kcsan/kcsan-test.c
> +++ b/kernel/kcsan/kcsan-test.c
> @@ -43,7 +43,7 @@ static struct {
>  };
>  
>  /* Setup test checking loop. */
> -static __no_kcsan_or_inline void
> +static __no_kcsan inline void
>  begin_test_checks(void (*func1)(void), void (*func2)(void))
>  {
>  	kcsan_disable_current();
> @@ -60,7 +60,7 @@ begin_test_checks(void (*func1)(void), void (*func2)(void))
>  }
>  
>  /* End test checking loop. */
> -static __no_kcsan_or_inline bool
> +static __no_kcsan inline bool
>  end_test_checks(bool stop)
>  {
>  	if (!stop && time_before(jiffies, end_time)) {
> -- 
> 2.27.0.rc2.251.g90737beb825-goog
> 

Re: [PATCH] kcsan: Prefer '__no_kcsan inline' in test

[Marco Elver]

On Tue, 2 Jun 2020 at 22:48, Paul E. McKenney <paulmck@kernel.org> wrote:
>
> On Tue, Jun 02, 2020 at 04:36:33PM +0200, Marco Elver wrote:
> > Instead of __no_kcsan_or_inline, prefer '__no_kcsan inline' in test --
> > this is in case we decide to remove __no_kcsan_or_inline.
> >
> > Suggested-by: Peter Zijlstra <peterz@infradead.org>
> > Signed-off-by: Marco Elver <elver@google.com>
> > ---
> >
> > Hi Paul,
> >
> > This is to prepare eventual removal of __no_kcsan_or_inline, and avoid a
> > series that doesn't apply to anything other than -next (because some
> > bits are in -tip and the test only in -rcu; although this problem might
> > be solved in 2 weeks). This patch is to make sure in case the
> > __kcsan_or_inline series is based on -tip, integration in -next doesn't
> > cause problems.
> >
> > This came up in
> > https://lkml.kernel.org/r/20200529185923.GO706495@hirez.programming.kicks-ass.net
>
> Applied and pushed, thank you!
>
> Please note that unless you would like this pushed into the current
> merge window, it will not be visible in -next until v5.8-rc1 comes out.
> Which sounds like you are aware of already, just want to be sure.  ;-)

Thank you! Yes, that's fine. The test and this patch I'd expect will
go into 5.9 earliest.

Thanks,
-- Marco