T0b1/wasmtime
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

x64: Improve memory support in {insert,extract}lane (#5982)

Browse Source 
* x64: Improve memory support in `{insert,extract}lane`

This commit improves adds support to Cranelift to emit `pextr{b,w,d,q}`
with a memory destination, merging a store-of-extract operation into one
instruction. Additionally AVX support is added for the `pextr*`
instructions.

I've additionally tried to ensure that codegen tests and runtests exist
for all forms of these instructions too.

* Add missing commas

* Fix tests
This commit is contained in:
Alex Crichton
2023-03-13 14:30:44 -05:00
committed by GitHub
parent 5c95e6fbaf
commit 6ecdc2482e
17 changed files with 1066 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
cranelift/codegen/src/isa/x64/inst/mod.rs
View File

@@ -131,6 +131,7 @@ impl Inst {
// These use dynamic SSE opcodes.
Inst::GprToXmm { op, .. }
| Inst::XmmMovRM { op, .. }
| Inst::XmmMovRMImm { op, .. }
| Inst::XmmRmiReg { opcode: op, .. }
| Inst::XmmRmR { op, .. }
| Inst::XmmRmRUnaligned { op, .. }
@@ -153,7 +154,9 @@ impl Inst {
| Inst::XmmVexPinsr { op, .. }
| Inst::XmmUnaryRmRVex { op, .. }
| Inst::XmmUnaryRmRImmVex { op, .. }
| Inst::XmmMovRMVex { op, .. } => op.available_from(),
| Inst::XmmMovRMVex { op, .. }
| Inst::XmmMovRMImmVex { op, .. }
| Inst::XmmToGprImmVex { op, .. } => op.available_from(),
}
}
}
@@ -331,7 +334,7 @@ impl Inst {
debug_assert!(src.class() == RegClass::Float);
Inst::XmmMovRM {
op,
src,
src: Xmm::new(src).unwrap(),
dst: dst.into(),
}
}
@@ -933,17 +936,33 @@ impl PrettyPrint for Inst {
}
Inst::XmmMovRM { op, src, dst, .. } => {
let src = pretty_print_reg(*src, 8, allocs);
let src = pretty_print_reg(src.to_reg(), 8, allocs);
let dst = dst.pretty_print(8, allocs);
format!("{} {}, {}", ljustify(op.to_string()), src, dst)
}
Inst::XmmMovRMVex { op, src, dst, .. } => {
let src = pretty_print_reg(*src, 8, allocs);
let src = pretty_print_reg(src.to_reg(), 8, allocs);
let dst = dst.pretty_print(8, allocs);
format!("{} {}, {}", ljustify(op.to_string()), src, dst)
}
Inst::XmmMovRMImm {
op, src, dst, imm, ..
} => {
let src = pretty_print_reg(src.to_reg(), 8, allocs);
let dst = dst.pretty_print(8, allocs);
format!("{} ${imm}, {}, {}", ljustify(op.to_string()), src, dst)
}
Inst::XmmMovRMImmVex {
op, src, dst, imm, ..
} => {
let src = pretty_print_reg(src.to_reg(), 8, allocs);
let dst = dst.pretty_print(8, allocs);
format!("{} ${imm}, {}, {}", ljustify(op.to_string()), src, dst)
}
Inst::XmmRmR {
op,
src1,
@@ -1023,7 +1042,7 @@ impl PrettyPrint for Inst {
let src1 = pretty_print_reg(src1.to_reg(), 8, allocs);
let src2 = src2.pretty_print(8, allocs);
format!("{} ${imm} {src1}, {src2}, {dst}", ljustify(op.to_string()))
format!("{} ${imm}, {src1}, {src2}, {dst}", ljustify(op.to_string()))
}
Inst::XmmVexPinsr {
@@ -1038,7 +1057,7 @@ impl PrettyPrint for Inst {
let src1 = pretty_print_reg(src1.to_reg(), 8, allocs);
let src2 = src2.pretty_print(8, allocs);
format!("{} ${imm} {src1}, {src2}, {dst}", ljustify(op.to_string()))
format!("{} ${imm}, {src1}, {src2}, {dst}", ljustify(op.to_string()))
}
Inst::XmmRmRVex3 {
@@ -1190,6 +1209,12 @@ impl PrettyPrint for Inst {
format!("{} ${imm}, {}, {}", ljustify(op.to_string()), src, dst)
}
Inst::XmmToGprImmVex { op, src, dst, imm } => {
let src = pretty_print_reg(src.to_reg(), 8, allocs);
let dst = pretty_print_reg(dst.to_reg().to_reg(), 8, allocs);
format!("{} ${imm}, {}, {}", ljustify(op.to_string()), src, dst)
}
Inst::GprToXmm {
op,
src,
@@ -2033,8 +2058,11 @@ fn x64_get_operands<F: Fn(VReg) -> VReg>(inst: &Inst, collector: &mut OperandCol
collector.reg_reuse_def(dst.to_writable_reg(), 0); // Reuse RHS.
src2.get_operands(collector);
}
Inst::XmmMovRM { src, dst, .. } | Inst::XmmMovRMVex { src, dst, .. } => {
collector.reg_use(*src);
Inst::XmmMovRM { src, dst, .. }
| Inst::XmmMovRMVex { src, dst, .. }
| Inst::XmmMovRMImm { src, dst, .. }
| Inst::XmmMovRMImmVex { src, dst, .. } => {
collector.reg_use(src.to_reg());
dst.get_operands(collector);
}
Inst::XmmCmpRmR { src, dst, .. } => {
@@ -2058,7 +2086,9 @@ fn x64_get_operands<F: Fn(VReg) -> VReg>(inst: &Inst, collector: &mut OperandCol
collector.reg_use(src.to_reg());
collector.reg_fixed_nonallocatable(*dst);
}
Inst::XmmToGpr { src, dst, .. } | Inst::XmmToGprImm { src, dst, .. } => {
Inst::XmmToGpr { src, dst, .. }
| Inst::XmmToGprImm { src, dst, .. }
| Inst::XmmToGprImmVex { src, dst, .. } => {
collector.reg_use(src.to_reg());
collector.reg_def(dst.to_writable_reg());
}