diff options
| author | Paolo Bonzini <pbonzini@redhat.com> | 2022-10-19 13:22:06 +0200 |
|---|---|---|
| committer | Paolo Bonzini <pbonzini@redhat.com> | 2022-10-20 15:16:18 +0200 |
| commit | cf5ec6641ed456e2748b211b7bbf5103bfc93098 (patch) | |
| tree | 84e5a3c059e7de484c1c7d66714c68e8d60ca9e1 /target/i386/ops_sse_header.h | |
| parent | 314d3eff66f41f39191aaca2e5f6e3dc81480c1b (diff) | |
target/i386: implement F16C instructions
F16C only consists of two instructions, which are a bit peculiar
nevertheless.
First, they access only the low half of an YMM or XMM register for the
packed-half operand; the exact size still depends on the VEX.L flag.
This is similar to the existing avx_movx flag, but not exactly because
avx_movx is hardcoded to affect operand 2. To this end I added a "ph"
format name; it's possible to reuse this approach for the VPMOVSX and
VPMOVZX instructions, though that would also require adding two more
formats for the low-quarter and low-eighth of an operand.
Second, VCVTPS2PH is somewhat weird because it *stores* the result of
the instruction into memory rather than loading it.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Diffstat (limited to 'target/i386/ops_sse_header.h')
| -rw-r--r-- | target/i386/ops_sse_header.h | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/target/i386/ops_sse_header.h b/target/i386/ops_sse_header.h index 2f1f811f9f..c4c41976c0 100644 --- a/target/i386/ops_sse_header.h +++ b/target/i386/ops_sse_header.h @@ -353,6 +353,12 @@ DEF_HELPER_4(glue(aeskeygenassist, SUFFIX), void, env, Reg, Reg, i32) DEF_HELPER_5(glue(pclmulqdq, SUFFIX), void, env, Reg, Reg, Reg, i32) #endif +/* F16C helpers */ +#if SHIFT >= 1 +DEF_HELPER_3(glue(cvtph2ps, SUFFIX), void, env, Reg, Reg) +DEF_HELPER_4(glue(cvtps2ph, SUFFIX), void, env, Reg, Reg, int) +#endif + /* AVX helpers */ #if SHIFT >= 1 DEF_HELPER_4(glue(vpermilpd, SUFFIX), void, env, Reg, Reg, Reg) |