Re: [Qemu-devel] [PATCH 1/9] tests: cris: force inlining

[Peter Maydell]

On 5 September 2016 at 12:54, Rabin Vincent <address@hidden> wrote:
> From: Rabin Vincent <address@hidden>
>
> The CRIS tests expect that functions marked inline are always inline.
> With newer versions of GCC, building them results warnings like the
> following and spurious failures when they are run.
>
> In file included from tests/tcg/cris/check_moveq.c:5:0:
> tests/tcg/cris/crisutils.h:66:20: warning: inlining failed in call to
> 'cris_tst_cc.constprop.0': call is unlikely and code size would grow 
> [-Winline]
> tests/tcg/cris/check_moveq.c:28:13: warning: called from here [-Winline]
>
> Use the always_inline attribute when building them to fix this.

This test code is pretty horrific; it relies on a whole bunch
of stuff that the compiler doesn't actually guarantee you,
and using always-inline is just a bandaid over the real
problem (which is that it should really be using multi-insn
asm statements when it cares about order, and not assuming
it can put two asm statements next to each other and set
cflags in one and read them in another).

That said, you probably aren't interested in rewriting it, so:

> Signed-off-by: Rabin Vincent <address@hidden>
> ---
>  tests/tcg/cris/sys.h | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/tests/tcg/cris/sys.h b/tests/tcg/cris/sys.h
> index c5f88e1..b1bf4c5 100644
> --- a/tests/tcg/cris/sys.h
> +++ b/tests/tcg/cris/sys.h
> @@ -3,6 +3,8 @@
>  #define STRINGIFY(x) #x
>  #define TOSTRING(x) STRINGIFY(x)
>
> +#define inline inline __attribute__((always_inline))
> +

I think redefining C keywords is generally a bad idea.
Can you instead define an "always_inline" macro and
use it where necessary?

thanks
-- PMM