load_complex and store_complex instructions (#309)

* Start adding the load_complex and store_complex instructions.

N.b.:
The text format is not correct yet. Requires changes to the lexer and parser.
I'm not sure why I needed to change the RuntimeError to Exception yet. Will fix.

* Get first few encodings of load_complex working. Still needs var args type checking.

* Clean up ModRM helper functions in binemit.

* Implement 32-bit displace for load_complex

* Use encoding helpers instead of doing them all by hand

* Initial implementation of store_complex

* Parse value list for load/store_complex with + as delimiter. Looks nice.

* Add sign/zero-extension and size variants for load_complex.

* Add size variants of store_complex.

* Add asm helper lines to load/store complex bin tests.

* Example of length-checking the instruction ValueList for an encoding. Extremely questionable implementation.

* Fix Python linting issues

* First draft of postopt pass to fold adds and loads into load_complex. Just simple loads for now.

* Optimization pass now works with all types of loads.

* Add store+add -> store_complex to postopt pass

* Put complex address optimization behind ISA flag.

* Add load/store complex for f32 and f64

* Fixes changes to lexer that broke NaN parsing.

Abstracts away the repeated checks for whether or not the characters
following a + or - are going to be parsed as a number or not.

* Fix formatting issues

* Fix register restrictions for complex addresses.

* Encoding tests for x86-32.

* Add documentation for newly added instructions, recipes, and cdsl changes.

* Fix python formatting again

* Apply value-list length predicates to all LoadComplex and StoreComplex instructions.

* Add predicate types to new encoding helpers for mypy.

* Import FieldPredicate to satisfy mypy.

* Add and fix some "asm" strings in the encoding tests.

* Line-up 'bin' comments in x86/binary64 test

* Test parsing of offset-less store_complex instruction.

* 'sNaN' not 'sNan'

* Bounds check the lookup for polymorphic typevar operand.

* Fix encodings for istore16_complex.
This commit is contained in:
Tyler McMullen
2018-05-09 12:07:00 -07:00
committed by Dan Gohman
parent 5aa84a744b
commit f636d795c5
25 changed files with 1127 additions and 21 deletions

View File

@@ -5,9 +5,9 @@
use cursor::{Cursor, EncCursor};
use ir::condcodes::{CondCode, FloatCC, IntCC};
use ir::dfg::ValueDef;
use ir::immediates::Imm64;
use ir::immediates::{Imm64, Offset32};
use ir::instructions::{Opcode, ValueList};
use ir::{Ebb, Function, Inst, InstBuilder, InstructionData, Value};
use ir::{Ebb, Function, Inst, InstBuilder, InstructionData, Value, Type, MemFlags};
use isa::TargetIsa;
use timing;
@@ -173,6 +173,158 @@ fn optimize_cpu_flags(
pos.func.update_encoding(info.br_inst, isa).is_ok();
}
struct MemOpInfo {
opcode: Opcode,
inst: Inst,
itype: Type,
arg: Value,
st_arg: Option<Value>,
flags: MemFlags,
offset: Offset32,
add_args: Option<[Value; 2]>,
}
fn optimize_complex_addresses(pos: &mut EncCursor, inst: Inst, isa: &TargetIsa) {
let mut info = match pos.func.dfg[inst] {
InstructionData::Load {
opcode,
arg,
flags,
offset,
} => MemOpInfo {
opcode: opcode,
inst: inst,
itype: pos.func.dfg.ctrl_typevar(inst),
arg: arg,
st_arg: None,
flags: flags,
offset: offset,
add_args: None,
},
InstructionData::Store {
opcode,
args,
flags,
offset,
} => MemOpInfo {
opcode: opcode,
inst: inst,
itype: pos.func.dfg.ctrl_typevar(inst),
arg: args[1],
st_arg: Some(args[0]),
flags: flags,
offset: offset,
add_args: None,
},
_ => return,
};
if let ValueDef::Result(result_inst, _) = pos.func.dfg.value_def(info.arg) {
match pos.func.dfg[result_inst] {
InstructionData::Binary { opcode, args } if opcode == Opcode::Iadd => {
info.add_args = Some(args.clone());
}
_ => return,
}
} else {
return;
}
match info.opcode {
Opcode::Load => {
pos.func.dfg.replace(info.inst).load_complex(
info.itype,
info.flags,
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Uload8 => {
pos.func.dfg.replace(info.inst).uload8_complex(
info.itype,
info.flags,
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Sload8 => {
pos.func.dfg.replace(info.inst).sload8_complex(
info.itype,
info.flags,
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Uload16 => {
pos.func.dfg.replace(info.inst).uload16_complex(
info.itype,
info.flags,
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Sload16 => {
pos.func.dfg.replace(info.inst).sload16_complex(
info.itype,
info.flags,
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Uload32 => {
pos.func.dfg.replace(info.inst).uload32_complex(
info.flags,
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Sload32 => {
pos.func.dfg.replace(info.inst).sload32_complex(
info.flags,
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Store => {
pos.func.dfg.replace(info.inst).store_complex(
info.flags,
info.st_arg.unwrap(),
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Istore8 => {
pos.func.dfg.replace(info.inst).istore8_complex(
info.flags,
info.st_arg.unwrap(),
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Istore16 => {
pos.func.dfg.replace(info.inst).istore16_complex(
info.flags,
info.st_arg.unwrap(),
&info.add_args.unwrap(),
info.offset,
);
}
Opcode::Istore32 => {
pos.func.dfg.replace(info.inst).istore32_complex(
info.flags,
info.st_arg.unwrap(),
&info.add_args.unwrap(),
info.offset,
);
}
_ => return,
}
pos.func.update_encoding(info.inst, isa).is_ok();
}
//----------------------------------------------------------------------
//
// The main post-opt pass.
@@ -198,6 +350,10 @@ pub fn do_postopt(func: &mut Function, isa: &TargetIsa) {
}
}
}
if isa.uses_complex_addresses() {
optimize_complex_addresses(&mut pos, inst, isa);
}
}
}
}