On Tue, May 26, 2020 at 2:38 PM Masahiro Yamada <masahi...@kernel.org> wrote:
> arch/m68k/Makefile computes lots of unneeded cc-option calls.
>
> For example, if CONFIG_M5441x is not defined, there is not point in
> evaluating the following compiler flag.
>
>  cpuflags-$(CONFIG_M5441x)      := $(call cc-option,-mcpu=54455,-mcfv4e)
>
> The result is set to cpuflags-, then thrown away.
>
> The right hand side of ':=' is immediately expanded. Hence, all of the
> 16 calls for cc-option are evaluated. This is expensive since cc-option
> invokes the compiler. This occurs even if you are not attempting to
> build anything, like 'make ARCH=m68k help'.
>
> Use '=' to expand the value _lazily_. The evaluation for cc-option is
> delayed until $(cpuflags-y) is expanded. So, the cc-option test happens
> just once at most.
>
> This commit mimics tune-y of arch/arm/Makefile.
>
> Signed-off-by: Masahiro Yamada <masahi...@kernel.org>

Reviewed-by: Geert Uytterhoeven <ge...@linux-m68k.org>
i.e. will queue in the m68k for-v5.9 branch.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Reply via email to