user0/rust
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update x86 codegen

Browse source
This commit is contained in:
gnzlbg 2018-11-27 10:49:58 +01:00 • committed by gnzlbg
parent 4a582fe37e
commit 900fa0261a
2 changed files with 13 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

1
library/stdarch/coresimd/macros.rs
View file

@ -1,5 +1,6 @@
//! Utility macros.
#[allow(unused)]
macro_rules! constify_imm8 {
($imm8:expr, $expand:ident) => {
#[allow(overflowing_literals)]

26
library/stdarch/coresimd/x86/sse.rs
View file

@ -1144,18 +1144,20 @@ pub unsafe fn _mm_movemask_ps(a: __m128) -> i32 {
/// ```
#[inline]
#[target_feature(enable = "sse")]
// TODO: generates MOVHPD if the CPU supports SSE2.
// #[cfg_attr(test, assert_instr(movhps))]
#[cfg_attr(all(test, target_arch = "x86_64"), assert_instr(movhpd))]
// 32-bit codegen does not generate `movhps` or `movhpd`, but instead
// `movsd` followed by `unpcklpd` (or `movss'/`unpcklps` if there's no SSE2).
#[cfg_attr(
all(test, target_arch = "x86", target_feature = "sse2"),
assert_instr(movlhps)
all(
test,
any(
target_arch = "x86_64",
all(target_arch = "x86", target_feature = "sse2")
)
),
assert_instr(movhpd)
)]
// FIXME: 32-bit codegen without SSE2 generates two `shufps` instead of `movhps`
#[cfg_attr(
all(test, target_arch = "x86", not(target_feature = "sse2")),
assert_instr(unpcklps)
assert_instr(shufps)
)]
// TODO: This function is actually not limited to floats, but that's what
// what matches the C type most closely: (__m128, *const __m64) -> __m128
@ -1202,20 +1204,16 @@ pub unsafe fn _mm_loadh_pi(a: __m128, p: *const __m64) -> __m128 {
/// ```
#[inline]
#[target_feature(enable = "sse")]
// TODO: generates MOVLPD if the CPU supports SSE2.
// #[cfg_attr(test, assert_instr(movlps))]
#[cfg_attr(all(test, target_arch = "x86_64"), assert_instr(movlpd))]
// On 32-bit targets with SSE2, it just generates two `movsd`.
#[cfg_attr(
all(test, target_arch = "x86", target_feature = "sse2"),
assert_instr(movsd)
assert_instr(movlpd)
)]
// It should really generate "movlps", but oh well...
// FIXME: On 32-bit targets without SSE2, it just generates two `movss`...
#[cfg_attr(
all(test, target_arch = "x86", not(target_feature = "sse2")),
assert_instr(movss)
)]
// TODO: Like _mm_loadh_pi, this also isn't limited to floats.
pub unsafe fn _mm_loadl_pi(a: __m128, p: *const __m64) -> __m128 {
let q = p as *const f32x2;
let b: f32x2 = *q;