Richard Henderson <richard.hender...@linaro.org> writes:
> On 4/19/22 02:10, Alex Bennée wrote:
>> This is a simple transliteration of the float_convs test but this time
>> working with doubles. I'm used it to test the handling of vector
>> registers in gdbstub but wasn't able to find a non-ugly way to
>> automate it.
>> Signed-off-by: Alex Bennée<alex.ben...@linaro.org>
>> ---
>> tests/tcg/multiarch/float_convd.c | 106 ++++
>> tests/tcg/aarch64/float_convd.ref | 988 ++++++++++++++++++++++++++++++
>> tests/tcg/arm/float_convd.ref | 988 ++++++++++++++++++++++++++++++
>> tests/tcg/i386/float_convd.conf | 988 ++++++++++++++++++++++++++++++
>> tests/tcg/x86_64/float_convd.ref | 988 ++++++++++++++++++++++++++++++
>> 5 files changed, 4058 insertions(+)
>> create mode 100644 tests/tcg/multiarch/float_convd.c
>> create mode 100644 tests/tcg/aarch64/float_convd.ref
>> create mode 100644 tests/tcg/arm/float_convd.ref
>> create mode 100644 tests/tcg/i386/float_convd.conf
>> create mode 100644 tests/tcg/x86_64/float_convd.ref
>
> Is it worth adjusting the current file to float_conv.c.inc, with
> #define FLOAT {float,double} in float_conv{s,d}.c before #include?
>
> But I suppose this is ok as-is.
I did consider pushing more into libs/float_helpers.c but the whole test file
is fairly simple as is.
> Reviewed-by: Richard Henderson <richard.hender...@linaro.org>
>
>
> r~
--
Alex Bennée
