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

load_complex and store_complex instructions (#309)

Browse Source 
* 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
Download Patch File Download Diff File Expand all files Collapse all files

74
cranelift/filetests/isa/x86/binary64.cton
View File

@@ -594,6 +594,80 @@ ebb0:
[-,%r8] v520 = ushr_imm v4, 63 ; bin: 49 c1 e8 3f
; Load Complex
[-,%rax] v521 = iconst.i64 1
[-,%rbx] v522 = iconst.i64 1
[-,%rdi] v523 = iconst.i32 1
[-,%rsi] v524 = iconst.i32 1
; asm: movq (%rax,%rbx,1), %rcx
[-,%rcx] v525 = load_complex.i64 v521+v522 ; bin: heap_oob 48 8b 0c 18
; asm: movl (%rax,%rbx,1), %ecx
[-,%rcx] v526 = load_complex.i32 v521+v522 ; bin: heap_oob 8b 0c 18
; asm: movq 1(%rax,%rbx,1), %rcx
[-,%rcx] v527 = load_complex.i64 v521+v522+1 ; bin: heap_oob 48 8b 4c 18 01
; asm: movl 1(%rax,%rbx,1), %ecx
[-,%rcx] v528 = load_complex.i32 v521+v522+1 ; bin: heap_oob 8b 4c 18 01
; asm: mov 0x100000(%rax,%rbx,1),%rcx
[-,%rcx] v529 = load_complex.i64 v521+v522+0x1000 ; bin: heap_oob 48 8b 8c 18 00001000
; asm: mov 0x100000(%rax,%rbx,1),%ecx
[-,%rcx] v530 = load_complex.i32 v521+v522+0x1000 ; bin: heap_oob 8b 8c 18 00001000
; asm: movzbq (%rax,%rbx,1),%rcx
[-,%rcx] v531 = uload8_complex.i64 v521+v522 ; bin: heap_oob 48 0f b6 0c 18
; asm: movzbl (%rax,%rbx,1),%ecx
[-,%rcx] v532 = uload8_complex.i32 v521+v522 ; bin: heap_oob 0f b6 0c 18
; asm: movsbq (%rax,%rbx,1),%rcx
[-,%rcx] v533 = sload8_complex.i64 v521+v522 ; bin: heap_oob 48 0f be 0c 18
; asm: movsbl (%rax,%rbx,1),%ecx
[-,%rcx] v534 = sload8_complex.i32 v521+v522 ; bin: heap_oob 0f be 0c 18
; asm: movzwq (%rax,%rbx,1),%rcx
[-,%rcx] v535 = uload16_complex.i64 v521+v522 ; bin: heap_oob 48 0f b7 0c 18
; asm: movzwl (%rax,%rbx,1),%ecx
[-,%rcx] v536 = uload16_complex.i32 v521+v522 ; bin: heap_oob 0f b7 0c 18
; asm: movswq (%rax,%rbx,1),%rcx
[-,%rcx] v537 = sload16_complex.i64 v521+v522 ; bin: heap_oob 48 0f bf 0c 18
; asm: movswl (%rax,%rbx,1),%ecx
[-,%rcx] v538 = sload16_complex.i32 v521+v522 ; bin: heap_oob 0f bf 0c 18
; asm: mov (%rax,%rbx,1),%ecx
[-,%rcx] v539 = uload32_complex v521+v522 ; bin: heap_oob 8b 0c 18
; asm: movslq (%rax,%rbx,1),%rcx
[-,%rcx] v540 = sload32_complex v521+v522 ; bin: heap_oob 48 63 0c 18
[-,%r13] v550 = iconst.i64 1
[-,%r14] v551 = iconst.i64 1
; asm: mov 0x0(%r13,%r14,1),%r12d
[-,%r12] v552 = load_complex.i32 v550+v551 ; bin: heap_oob 47 8b 64 35 00
; Store Complex
[-,%rcx] v600 = iconst.i64 1
[-,%rcx] v601 = iconst.i32 1
[-,%r10] v602 = iconst.i64 1
[-,%r11] v603 = iconst.i32 1
; asm: mov %rcx,(%rax,%rbx,1)
store_complex v600, v521+v522 ; bin: heap_oob 48 89 0c 18
; asm: mov %rcx,0x1(%rax,%rbx,1)
store_complex v600, v521+v522+1 ; bin: heap_oob 48 89 4c 18 01
; asm: mov %rcx,0x100000(%rax,%rbx,1)
store_complex v600, v521+v522+0x1000 ; bin: heap_oob 48 89 8c 18 00001000
; asm: mov %ecx,(%rax,%rbx,1)
store_complex v601, v521+v522 ; bin: heap_oob 89 0c 18
; asm: mov %ecx,0x1(%rax,%rbx,1)
store_complex v601, v521+v522+1 ; bin: heap_oob 89 4c 18 01
; asm: mov %ecx,0x100000(%rax,%rbx,1)
store_complex v601, v521+v522+0x1000 ; bin: heap_oob 89 8c 18 00001000
; asm: mov %ecx,(%rax,%rbx,1)
istore32_complex v600, v521+v522 ; bin: heap_oob 89 0c 18
; asm: mov %cx,(%rax,%rbx,1)
istore16_complex v600, v521+v522 ; bin: heap_oob 66 89 0c 18
; asm: mov %cx,(%rax,%rbx,1)
istore16_complex v601, v521+v522 ; bin: heap_oob 66 89 0c 18
; asm: mov %r10w,(%rax,%rbx,1)
istore16_complex v602, v521+v522 ; bin: heap_oob 66 44 89 14 18
; asm: mov %r11w,(%rax,%rbx,1)
istore16_complex v603, v521+v522 ; bin: heap_oob 66 44 89 1c 18
; asm: mov %cl,(%rax,%rbx,1)
istore8_complex v600, v521+v522 ; bin: heap_oob 88 0c 18
; asm: mov %cl,(%rax,%rbx,1)
istore8_complex v601, v521+v522 ; bin: heap_oob 88 0c 18
; asm: testq %rcx, %rcx
; asm: je ebb1
brz v1, ebb1 ; bin: 48 85 c9 74 1b