Allow target access to routines using the minmax flags.
Make the existing min/max wrappers inline.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
There is only one NaN fractional encoding for E4M3. Retain the
incoming sign, but force the outgoing fraction to the unique value.
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Introduce a separate control from float_snan_rule
that applies only to the OCP E4M3 format.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Drop FloatFTZDetection and use #defines, like we do for
tininess_before_rounding. Rename get_float_ftz_detection
to get_ftz_before_rounding and move to softfloat.c, as
there are no external users.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Rename get_float_detect_tininess to get_tininess_before_rounding
and move to softfloat.c, as there are no external users.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We have, in the past, used any of uint8_t, uint16_t or int
to hold the set of exception flags. Use a typedef instead.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Avoid exponent overflow as well as checking that we don't lose information with
opposing scaling. Use it in partsN(scalbn) and partsN(round_to_int_normal).
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Consolidate the tests for zero and anynorm.
Add comments for a few cases.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The nan test had been down below because it was unlikely.
But if we have to have one anyway because of denormals,
we might as well take care of them right away.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Check the likely case of normal product and normal or
zero addend first; shift NaN and infinity detection down;
end with zero product + addend.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This requires more complexity to handle const selectors, and
an indirection macro for each function. Easier to just use
the preprocessor.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use parts64_log2 at each call site.
That leaves parts128_log2 unused, so move the whole function back
to softfloat.c and specialize for FloatParts64.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use parts64_float_to_sint_modulo at each call site.
That leaves parts128_float_to_sint_modulo unused,
so move the whole function back to softfloat.c and
specialize for FloatParts64.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Drop the forward declarations and the _Generic macros.
Add partsW() for use by muladd_scalbn.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Infinity is a special case distinct from numeric overflow:
- Numeric overflow: finite value exceeds format's max normal
-> overflow|inexact
- Infinity conversion: input is already infinite
-> no flags
This commit fixes the unexpect exception flags by relocating the float
exception flag update flow to be outside the uncanon_e4m3_overflow.
And raising the overflow|inexact for numeric overflow in uncanon_normal.
Fixes: 27e989f99c ("fpu: Add conversion routines for OCP FP8 E4M3")
Reviewed-by: Chao Liu <chao.liu.zevorn@gmail.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Max Chou <max.chou@sifive.com>
Message-ID: <20260226071817.1417875-3-max.chou@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
In the uncanonical step, the input FloatParts will be repacked to the
target FloatFmt. This commit fixes following issues after calling
uncanon_e4m3_overflow in the uncanon/uncanon_normal functions.
- Add the local exp update after calling uncanon_e4m3_overflow in the
parts_uncanon_normal function.
- Add the fraction shift after calling uncanon_e4m3_overflow in the
parts_uncanon function.
Fixes: 27e989f99c ("fpu: Add conversion routines for OCP FP8 E4M3")
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Chao Liu <chao.liu.zevorn@gmail.com>
Signed-off-by: Max Chou <max.chou@sifive.com>
Message-ID: <20260226071817.1417875-2-max.chou@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
