Received-SPF: pass (google.com: domain of linux-kernel+bounces-152181-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Precedence: bulk
MIME-Version: 1.0
References: <20240416121838.95427-1-masahiroy@kernel.org> <20240416121838.95427-5-masahiroy@kernel.org>
 <ZiKJNrei7zUVaLkb@bergen.fjasle.eu>
In-Reply-To: <ZiKJNrei7zUVaLkb@bergen.fjasle.eu>
From: Masahiro Yamada <masahiroy@kernel.org>
Date: Sat, 20 Apr 2024 17:56:47 +0900
Message-ID: <CAK7LNATQBZyKArGAO9GyjiN26cUgjmxaczStxsJG-Q0b7RW+Yg@mail.gmail.com>
Subject: Re: [PATCH 4/4] kbuild: use $(src) instead of $(srctree)/$(src) for
 source directory
To: Nicolas Schier <nicolas@fjasle.eu>
Cc: linux-kbuild@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Sat, Apr 20, 2024 at 12:09=E2=80=AFAM Nicolas Schier <nicolas@fjasle.eu>=
 wrote:
>
> On Tue 16 Apr 2024 21:18:38 GMT, Masahiro Yamada wrote:
> > Kbuild conventionally uses $(obj)/ for generated files, and $(src)/ for
> > checked-in source files. It is merely a convention without any function=
al
> > difference. In fact, $(obj) and $(src) are exactly the same, as defined
> > in scripts/Makefile.build:
> >
> >     src :=3D $(obj)
> >
> > When the kernel is built in a separate output directory, $(src) does
> > not match the directory in the source tree. It still works because the
> > top Makefile sets VPATH to $(srctree). However, VPATH does not cover
> > all cases. There are many places where the exact path to the source
> > directory is needed. For example, when adding a header search path for
> > local headers, it is typical to pass -I$(srctree)/$(src) to the compile=
r.
> >
> > However, it introduces inconsistency between upstream Makefiles and
> > downstream Makefiles; when the source directory path is needed,
> > $(srctree)/$(src) is used in the upstream kernel, while $(src) is used
> > in external modules.
> >
> > To resolve the distortion, this commit changes the semantics of $(src)
> > so that it always points to the directory in the source tree.
> >
> > Going forward, the variables used in Makefiles will work as follows:
> >
> >   $(obj)     - directory in the object tree
> >   $(src)     - directory in the source tree  (changed by this commit!)
> >   $(objtree) - the top of the kernel object tree
> >   $(srctree) - the top of the kernel source tree
> >
> > This is a tweewide Makefile conversion for replacing $(srctree)/$(src)
> > with $(src).
>
> tweewide -> treewide
>
> [...]
> > diff --git a/drivers/Makefile b/drivers/Makefile
> > index 3bf5cab4b451..fe9ceb0d2288 100644
> > --- a/drivers/Makefile
> > +++ b/drivers/Makefile
> > @@ -6,11 +6,6 @@
> >  # Rewritten to use lists instead of if-statements.
> >  #
> >
> > -# Some driver Makefiles miss $(srctree)/ for include directive.
> > -ifdef building_out_of_srctree
> > -MAKEFLAGS +=3D --include-dir=3D$(srctree)
> > -endif
> > -
>
> Ah, it's nice to get rid of this.
>
> >  obj-y                                +=3D cache/
> >  obj-y                                +=3D irqchip/
> >  obj-y                                +=3D bus/
> [...]
> > diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib
> > index 3179747cbd2c..c53d2f168801 100644
> > --- a/scripts/Makefile.lib
> > +++ b/scripts/Makefile.lib
> > @@ -209,13 +209,13 @@ _c_flags +=3D $(if $(patsubst n%,, \
> >       -D__KCSAN_INSTRUMENT_BARRIERS__)
> >  endif
> >
> > -# $(srctree)/$(src) for including checkin headers from generated sourc=
e files
> > -# $(objtree)/$(obj) for including generated headers from checkin sourc=
e files
> > +# $(src) for including checkin headers from generated source files
> > +# $(obj) for including generated headers from checkin source files
>
> this is a funny criss-cross of words; but might the following be more
> accurate?
>
> # $(src) for including headers from checked-in source files
> # $(obj) for including generated headers


These are added only when building_out_of_srctree is defined.

When the kernel is built in-tree,
the include directive with the double-quote form
should work without the header inclusion path.


Anyway, this does not seem to be a good idea.
I want to get rid of these entirely.


>
>
>
> Anyway, thanks for the patch.  I think it is quite a good way forward!
>
> Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>


--=20
Best Regards
Masahiro Yamada