On 01/10/2024 00.07, Philippe Mathieu-Daudé wrote:
> Replace the target-specific tswap32() calls by stl_endian_p()
> which does the same but takes the endianness as argument, thus
> is target-agnostic.
>
> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> ---
> Based-on: <20240930073450.33195-2-philmd@linaro.org>
> "qemu/bswap: Introduce ld/st_endian_p() API"
> ---
> hw/sh4/r2d.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/hw/sh4/r2d.c b/hw/sh4/r2d.c
> index e5ac6751bd..6739ac435a 100644
> --- a/hw/sh4/r2d.c
> +++ b/hw/sh4/r2d.c
> @@ -353,9 +353,11 @@ static void r2d_init(MachineState *machine)
> }
>
> /* initialization which should be done by firmware */
> - boot_params.loader_type = tswap32(1);
> - boot_params.initrd_start = tswap32(INITRD_LOAD_OFFSET);
> - boot_params.initrd_size = tswap32(initrd_size);
> + stl_endian_p(TARGET_BIG_ENDIAN, &boot_params.loader_type, 1);
> + stl_endian_p(TARGET_BIG_ENDIAN, &boot_params.initrd_start,
> + INITRD_LOAD_OFFSET);
> + stl_endian_p(TARGET_BIG_ENDIAN, &boot_params.initrd_size,
> + initrd_size);
> }
I still don't quite get it - the tswap functions already use
target_words_bigendian() when being used in common code, so what do you
really gain by these changes?
Thomas