Re: [PATCH tip/core/rcu 01/10] kcsan: Add pointer to access-marking.txt to data_race() bullet

[Marco Elver]

On Thu, 13 May 2021 at 12:47, Akira Yokosawa <akiyks@xxxxxxxxx> wrote:
>
> Hi Paul,
>
> On Tue, 11 May 2021 16:23:52 -0700, Paul E. McKenney wrote:
> > This commit references tools/memory-model/Documentation/access-marking.txt
> > in the bullet introducing data_race().  The access-marking.txt file
> > gives advice on when data_race() should and should not be used.
> >
> > Suggested-by: Akira Yokosawa <akiyks@xxxxxxxxx>
> > Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>
> > ---
> >  Documentation/dev-tools/kcsan.rst | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/Documentation/dev-tools/kcsan.rst b/Documentation/dev-tools/kcsan.rst
> > index d85ce238ace7..80894664a44c 100644
> > --- a/Documentation/dev-tools/kcsan.rst
> > +++ b/Documentation/dev-tools/kcsan.rst
> > @@ -106,7 +106,9 @@ the below options are available:
> >
> >  * KCSAN understands the ``data_race(expr)`` annotation, which tells KCSAN that
> >    any data races due to accesses in ``expr`` should be ignored and resulting
> > -  behaviour when encountering a data race is deemed safe.
> > +  behaviour when encountering a data race is deemed safe.  Please see
> > +  ``tools/memory-model/Documentation/access-marking.txt`` in the kernel source
> > +  tree for more information.
> >
> >  * Disabling data race detection for entire functions can be accomplished by
> >    using the function attribute ``__no_kcsan``::
> >
>
> I think this needs some adjustment for overall consistency.
> A possible follow-up patch (relative to the change above) would look
> like the following.
>
> Thoughts?
>
>         Thanks, Akira
>
> -------8<--------
> From: Akira Yokosawa <akiyks@xxxxxxxxx>
> Subject: [PATCH] kcsan: Use URL link for pointing access-marking.txt
>
> For consistency within kcsan.rst, use a URL link as the same as in
> section "Data Races".
>
> Signed-off-by: Akira Yokosawa <akiyks@xxxxxxxxx>
> Cc: Paul E. McKenney <paulmck@xxxxxxxxxx>

Good catch. I'd be in favour of this change, as it makes it simpler to
just follow the link. Because in most cases I usually just point folks
at the rendered version of this:
https://www.kernel.org/doc/html/latest/dev-tools/kcsan.html

Acked-by: Marco Elver <elver@xxxxxxxxxx>

> ---
>  Documentation/dev-tools/kcsan.rst | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/Documentation/dev-tools/kcsan.rst b/Documentation/dev-tools/kcsan.rst
> index 80894664a44c..151f96b7fef0 100644
> --- a/Documentation/dev-tools/kcsan.rst
> +++ b/Documentation/dev-tools/kcsan.rst
> @@ -107,8 +107,7 @@ the below options are available:
>  * KCSAN understands the ``data_race(expr)`` annotation, which tells KCSAN that
>    any data races due to accesses in ``expr`` should be ignored and resulting
>    behaviour when encountering a data race is deemed safe.  Please see
> -  ``tools/memory-model/Documentation/access-marking.txt`` in the kernel source
> -  tree for more information.
> +  `"Marking Shared-Memory Accesses" in the LKMM`_ for more information.
>
>  * Disabling data race detection for entire functions can be accomplished by
>    using the function attribute ``__no_kcsan``::
> @@ -130,6 +129,8 @@ the below options are available:
>
>      KCSAN_SANITIZE := n
>
> +.. _"Marking Shared-Memory Accesses" in the LKMM: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/memory-model/Documentation/access-marking.txt
> +
>  Furthermore, it is possible to tell KCSAN to show or hide entire classes of
>  data races, depending on preferences. These can be changed via the following
>  Kconfig options:
> --
> 2.17.1