On 09/04/2019 14.51, Andrew Jones wrote:
> We're being a bit tricky when we change the prototype of halt() to
> take 'code' as an argument. Doing this gives us a last chance to
> see it when inspecting the halted unit test state (it'll be in
> r0/x0 for arm/arm64 and r3 for powerpc.) Let's comment this trick
> to avoid confusion.
>
> Suggested-by: Alexandru Elisei <alexandru.elisei@xxxxxxx>
> Signed-off-by: Andrew Jones <drjones@xxxxxxxxxx>
> ---
> lib/arm/io.c | 10 ++++++++--
> lib/powerpc/io.c | 9 +++++++--
> 2 files changed, 15 insertions(+), 4 deletions(-)
>
> diff --git a/lib/arm/io.c b/lib/arm/io.c
> index 8226b765bdc5..99fd31560084 100644
> --- a/lib/arm/io.c
> +++ b/lib/arm/io.c
> @@ -18,8 +18,6 @@
>
> #include "io.h"
>
> -extern void halt(int code);
> -
> static struct spinlock uart_lock;
> /*
> * Use this guess for the uart base in order to make an attempt at
> @@ -89,6 +87,14 @@ void puts(const char *s)
> spin_unlock(&uart_lock);
> }
>
> +
> +/*
> + * Defining halt to take 'code' as an argument guarantees that it will
> + * be in x0/r0 when we halt. That gives us a final chance to see the exit
> + * status while inspecting the halted unit test state.
> + */
> +extern void halt(int code);
> +
> void exit(int code)
> {
> chr_testdev_exit(code);
> diff --git a/lib/powerpc/io.c b/lib/powerpc/io.c
> index 217eb0770ef5..a381688bce51 100644
> --- a/lib/powerpc/io.c
> +++ b/lib/powerpc/io.c
> @@ -11,8 +11,6 @@
> #include <asm/setup.h>
> #include "io.h"
>
> -extern void halt(int code);
> -
> static struct spinlock print_lock;
>
> void io_init(void)
> @@ -28,6 +26,13 @@ void puts(const char *s)
> spin_unlock(&print_lock);
> }
>
> +/*
> + * Defining halt to take 'code' as an argument guarantees that it will
> + * be in r3 when we halt. That gives us a final chance to see the exit
> + * status while inspecting the halted unit test state.
> + */
> +extern void halt(int code);
Acked-by: Thomas Huth <thuth@xxxxxxxxxx>
