diff options
| author | Richard Henderson | 2021-05-26 00:58:10 +0200 |
|---|---|---|
| committer | Peter Maydell | 2021-06-03 17:43:26 +0200 |
| commit | 60c8f7265d7eb51dfb38ea6701d10cbe2d7c7a64 (patch) | |
| tree | 643b8806c9ba6e8d716708c8f7011d9904dbb46d /fpu/softfloat-parts.c.inc | |
| parent | target/arm: Implement vector float32 to bfloat16 conversion (diff) | |
| download | qemu-60c8f7265d7eb51dfb38ea6701d10cbe2d7c7a64.tar.gz qemu-60c8f7265d7eb51dfb38ea6701d10cbe2d7c7a64.tar.xz qemu-60c8f7265d7eb51dfb38ea6701d10cbe2d7c7a64.zip | |
softfpu: Add float_round_to_odd_inf
For Arm BFDOT and BFMMLA, we need a version of round-to-odd
that overflows to infinity, instead of the max normal number.
Cc: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20210525225817.400336-6-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Diffstat (limited to 'fpu/softfloat-parts.c.inc')
| -rw-r--r-- | fpu/softfloat-parts.c.inc | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index a897a5a743..7f69da1d8f 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -176,13 +176,12 @@ static void partsN(uncanon)(FloatPartsN *p, float_status *s, g_assert_not_reached(); } + overflow_norm = false; switch (s->float_rounding_mode) { case float_round_nearest_even: - overflow_norm = false; inc = ((p->frac_lo & roundeven_mask) != frac_lsbm1 ? frac_lsbm1 : 0); break; case float_round_ties_away: - overflow_norm = false; inc = frac_lsbm1; break; case float_round_to_zero: @@ -199,6 +198,8 @@ static void partsN(uncanon)(FloatPartsN *p, float_status *s, break; case float_round_to_odd: overflow_norm = true; + /* fall through */ + case float_round_to_odd_inf: inc = p->frac_lo & frac_lsb ? 0 : round_mask; break; default: @@ -259,6 +260,7 @@ static void partsN(uncanon)(FloatPartsN *p, float_status *s, ? frac_lsbm1 : 0); break; case float_round_to_odd: + case float_round_to_odd_inf: inc = p->frac_lo & frac_lsb ? 0 : round_mask; break; default: |