Re: [PATCH v2 19/21] gcc-plugins: test GCC plugin support in Kconfig

[Kees Cook]

On Mon, Mar 26, 2018 at 10:29 PM, Masahiro Yamada
<yamada.masahiro@xxxxxxxxxxxxx> wrote:
> Run scripts/gcc-plugin.sh from Kconfig.  Users can enable GCC_PLUGINS
> only when it is supported.
>
> Signed-off-by: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>
> ---
>
> Changes in v2: None
>
>  arch/Kconfig                 |  4 +++
>  scripts/Makefile.gcc-plugins | 82 ++++++++++++++++----------------------------
>  scripts/gcc-plugin.sh        |  1 -
>  3 files changed, 33 insertions(+), 54 deletions(-)
>
> diff --git a/arch/Kconfig b/arch/Kconfig
> index b42378d..88cc925 100644
> --- a/arch/Kconfig
> +++ b/arch/Kconfig
> @@ -407,9 +407,13 @@ config HAVE_GCC_PLUGINS
>           An arch should select this symbol if it supports building with
>           GCC plugins.
>
> +config CC_HAS_GCC_PLUGINS
> +       bool
> +

This doesn't seem used anywhere?

>  menuconfig GCC_PLUGINS
>         bool "GCC plugins"
>         depends on HAVE_GCC_PLUGINS
> +       depends on $(success $srctree/scripts/gcc-plugin.sh $HOSTCXX $CC)
>         depends on !COMPILE_TEST
>         help
>           GCC plugins are loadable modules that provide extra features to the
> diff --git a/scripts/Makefile.gcc-plugins b/scripts/Makefile.gcc-plugins
> index 25da4c0..19d0d5b 100644
> --- a/scripts/Makefile.gcc-plugins
> +++ b/scripts/Makefile.gcc-plugins
> [...]
> -# If plugins aren't supported, abort the build before hard-to-read compiler
> -# errors start getting spewed by the main build.
> -PHONY += gcc-plugins-check
> -gcc-plugins-check: FORCE
> -ifdef CONFIG_GCC_PLUGINS
> -  ifeq ($(PLUGINCC),)
> -    ifneq ($(GCC_PLUGINS_CFLAGS),)
> -      $(Q)$(srctree)/scripts/gcc-plugin.sh --show-error $(HOSTCXX) $(CC) || true
> -      @echo "Cannot use CONFIG_GCC_PLUGINS: your gcc installation does not support plugins, perhaps the necessary headers are missing?" >&2 && exit 1

As mentioned in the other email, we lose the error reporting. Now the
lack of plugins is just a silent =n in menuconfig. Keeping
--show-error in the Kconfig call and retaining stderr would be nice.

I need to do some further testing with SANCOV, but otherwise this all
looks correct, and my testing shows it behaving correctly.

-Kees

-- 
Kees Cook
Pixel Security