Re: [PULL 02/60] semihosting: Return failure from softmmu-uaccess.h functions

[Philippe Mathieu-Daudé]

On 28/6/22 06:53, Richard Henderson wrote:
> We were reporting unconditional success for these functions;
> pass on any failure from cpu_memory_rw_debug.
>
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
>   include/semihosting/softmmu-uaccess.h | 91 ++++++++++++---------------
>   1 file changed, 39 insertions(+), 52 deletions(-)
>
> diff --git a/include/semihosting/softmmu-uaccess.h 
> b/include/semihosting/softmmu-uaccess.h
> index e69e3c8548..5246a91570 100644
> --- a/include/semihosting/softmmu-uaccess.h
> +++ b/include/semihosting/softmmu-uaccess.h
> @@ -12,82 +12,69 @@
>   
>   #include "cpu.h"
>   
> -static inline uint64_t softmmu_tget64(CPUArchState *env, target_ulong addr)
> -{
> -    uint64_t val;
> +#define get_user_u64(val, addr)                                         \

These macros match their user emulation equivalent, at the price
of loosing the 'env' argument, still used. This confuses meta
transformation tools such Coccinelle spatch. Not a big deal, just FTR.

> +    ({ uint64_t val_ = 0;                                               \
> +       int ret_ = cpu_memory_rw_debug(env_cpu(env), (addr),             \

                                              ^^^

> +                                      &val_, sizeof(val_), 0);          \
> +       (val) = tswap64(val_); ret_; })
>   
> -    cpu_memory_rw_debug(env_cpu(env), addr, (uint8_t *)&val, 8, 0);
> -    return tswap64(val);
> -}