Code

Add new overflow-checking primitives#273

Code

Merged

MatthewFluet merged 57 commits intoMLton:masterMLton/mlton:masterfrom daemanos:new-overflowdaemanos/mlton:new-overflowCopy head branch name to clipboardOct 15, 2018

+5,399-546

Commits

Commits on May 23, 2018

• Ignore vim configuration files

  daemanos

  committed
  d4c412e

  Copy full SHA for d4c412e

• Flip semantics of ! and ? primitive operators

  Show description for c938d1b
  daemanos

  committed
  c938d1b

  Copy full SHA for c938d1b

• Introduce $ operators and make ? predicates

  Show description for 8b3cef1
  daemanos

  committed
  8b3cef1

  Copy full SHA for 8b3cef1

Commits on May 31, 2018

• Make new primitives backwards-compatible

  Show description for cb4f3fc
  daemanos

  committed
  cb4f3fc

  Copy full SHA for cb4f3fc

• Add minimal backend support for new primitives

  daemanos

  committed
  f9e1cb6

  Copy full SHA for f9e1cb6

• Remove unnecessary cases of Arith transfers

  daemanos

  committed
  7bb41e5

  Copy full SHA for 7bb41e5

Commits on Jun 12, 2018

• Add support for new primitives in amd64 codegen

  daemanos

  committed
  81d8cf8

  Copy full SHA for 81d8cf8

Commits on Jun 25, 2018

• Use GCC builtins for overflow checking

  daemanos

  committed
  efef5dc

  Copy full SHA for efef5dc

• Add support for LLVM codegen

  daemanos

  committed
  b05ba13

  Copy full SHA for b05ba13

Commits on Jun 29, 2018

• Refactor checked word primitives for C codegen

  Show description for 819ddff
  daemanos

  committed
  819ddff

  Copy full SHA for 819ddff

• Correct typo in neg_overflow_b application

  daemanos

  committed
  8de2d7c

  Copy full SHA for 8de2d7c

Commits on Jul 5, 2018

• Add overflowCheckTemp storage for x86 codegen

  daemanos

  committed
  edafb83

  Copy full SHA for edafb83

• Make negCheckP use 0-x for bounds checking

  daemanos

  committed
  87ea212

  Copy full SHA for 87ea212

Commits on Jul 6, 2018

• Add support for x86 codegen

  daemanos

  committed
  4a32966

  Copy full SHA for 4a32966

Commits on Jul 9, 2018

• Use GCC builtins in implementation of word prims

  daemanos

  committed
  a68f209

  Copy full SHA for a68f209

Commits on Jul 10, 2018

• Add support for new primitives in redundantTests

  daemanos

  committed
  1822eb4

  Copy full SHA for 1822eb4

• Add arithmetic identity support for newOverflow

  daemanos

  committed
  267504c

  Copy full SHA for 267504c

• Re-arrange newOverflow operation order

  Show description for 58eae83
  daemanos

  committed
  58eae83

  Copy full SHA for 58eae83

• Use 0-x for negation overflow checking in amd64

  daemanos

  committed
  e6abe7f

  Copy full SHA for e6abe7f

• Remove unused variables

  daemanos

  committed
  72cb10b

  Copy full SHA for 72cb10b

Commits on Jul 12, 2018

• Correct implementation of mulCheckP

  Show description for 3dffc90
  daemanos

  committed
  3dffc90

  Copy full SHA for 3dffc90

• Rename negal(64)cc to neg(64)cc

  daemanos

  committed
  844d11f

  Copy full SHA for 844d11f

• Use a single statement map in redundantTests

  Show description for 8f2dfab
  daemanos

  committed
  8f2dfab

  Copy full SHA for 8f2dfab

Commits on Jul 16, 2018

• Add peephole optimization for new primitives

  Show description for afd428d
  daemanos

  committed
  afd428d

  Copy full SHA for afd428d

Commits on Jul 27, 2018

• Add layout for amd64 chunks/blocks

  daemanos

  committed
  01c9f12

  Copy full SHA for 01c9f12

• Perform elimALRedundant before moveHoist

  Show description for a5d8e79
  daemanos

  committed
  a5d8e79

  Copy full SHA for a5d8e79

Commits on Jul 30, 2018

• Add elimALRedundant pass to x86 codegen

  daemanos

  committed
  81f2c6e

  Copy full SHA for 81f2c6e

Commits on Jul 31, 2018

• Add more peephole optimization coverage

  Show description for ff6b544
  daemanos

  committed
  ff6b544

  Copy full SHA for ff6b544

Commits on Aug 1, 2018

• Fix undefined reference in amd64-simplify.fun

  daemanos

  committed
  5de7724

  Copy full SHA for 5de7724

Commits on Aug 3, 2018

• Replace 0-x in negCheckP with a NEG instruction

  daemanos

  committed
  47465f0

  Copy full SHA for 47465f0

Commits on Aug 6, 2018

• Modify peephole template for native codegens

  daemanos

  committed
  da8f4ee

  Copy full SHA for da8f4ee

Commits on Aug 13, 2018

• Use only _overflow builtins for the C codegen

  Show description for 7f8f493
  daemanos

  committed
  7f8f493

  Copy full SHA for 7f8f493

• Refactor C overflow primitives and FFI definitions

  daemanos

  committed
  7cd21e9

  Copy full SHA for 7cd21e9

• Use GCC5 for Travis CI

  daemanos

  committed
  bdb9335

  Copy full SHA for bdb9335

Commits on Aug 14, 2018

• Un-ignore vim configuration files

  daemanos

  committed
  4546174

  Copy full SHA for 4546174

Commits on Aug 15, 2018

• Add int-mul-pow2 regression test demonstrating bug

  Show description for a410df4
  MatthewFluet

  committed
  a410df4

  Copy full SHA for a410df4

Commits on Aug 28, 2018

• Fix int-mul-pow2 bug

  daemanos

  committed
  8fea7b6

  Copy full SHA for 8fea7b6

Commits on Sep 4, 2018

• Remove debug printing from amd64-simplify.fun

  daemanos

  committed
  0da9dfe

  Copy full SHA for 0da9dfe

Commits on Sep 5, 2018

• Remove superfluous elimALRedundant_pseudo optimization

  daemanos

  committed
  56d4362

  Copy full SHA for 56d4362

Commits on Sep 6, 2018

• Correct initial refactoring of overflow primitives

  daemanos

  committed
  03e41a0

  Copy full SHA for 03e41a0

Commits on Sep 10, 2018

• Correct initial refactoring of unchecked/checked arithmetic fs

  MatthewFluet

  committed
  724f1d1

  Copy full SHA for 724f1d1

• Alphabetize symbols in include/amd64-main.h

  MatthewFluet

  committed
  83d6955

  Copy full SHA for 83d6955

• Avoid using @llvm.{s,u}mul.with.overflow.i64 on x86entrytransfer

  Show description for 315ccce
  MatthewFluet

  committed
  315ccce

  Copy full SHA for 315ccce

• Remove superfluous elimALRedundant_pseudo optimization

  Show description for a1a4c0f
  MatthewFluet

  committed
  a1a4c0f

  Copy full SHA for a1a4c0f

• Delete trailing whitespace

  MatthewFluet

  committed
  df2e9e1

  Copy full SHA for df2e9e1

• Tweak code formatting and style

  MatthewFluet

  committed
  cf7c3a7

  Copy full SHA for cf7c3a7

• Refactor to share code for simplifying arithmetic check primitives

  MatthewFluet

  committed
  939e827

  Copy full SHA for 939e827

• Share code for type-checking comparisons and overflow-checking prims

  MatthewFluet

  committed
  9666e8a

  Copy full SHA for 9666e8a

• Code formatting

  MatthewFluet

  committed
  a720283

  Copy full SHA for a720283

Commits on Sep 13, 2018

• Fix #undef Word_negCheckP (had been #undef WordS_negCheckP)

  MatthewFluet

  committed
  ce4352d

  Copy full SHA for ce4352d

• Fix translation of Word_*Check primitives to C functions

  Show description for b91bb9a
  MatthewFluet

  committed
  b91bb9a

  Copy full SHA for b91bb9a

Commits on Sep 17, 2018

• Fix x86-codegen translation of Word64_{add,sub,neg}CheckP prims

  Show description for eda6021
  MatthewFluet

  committed
  eda6021

  Copy full SHA for eda6021

Commits on Oct 12, 2018

• Fix elimDeadDsts peephole optimization in native codegens

  Show description for 97315a4
  MatthewFluet

  committed
  97315a4

  Copy full SHA for 97315a4

• Whitespace fixes

  MatthewFluet

  committed
  5b9520e

  Copy full SHA for 5b9520e

• Strengthen elimAddSub1 and elimDeadDsts peephole optimizations

  Show description for 3a0a42d
  MatthewFluet

  committed
  3a0a42d

  Copy full SHA for 3a0a42d

• Fix x86-codegen translation of Word64_{add,neg,sub}CheckP prims

  Show description for 14df405
  MatthewFluet

  committed
  14df405

  Copy full SHA for 14df405

Commits on Oct 15, 2018

• Add CHANGELOG entry for new Overflow-checking primitives

  MatthewFluet

  committed
  2dc3968

  Copy full SHA for 2dc3968