This is an archive of the discontinued LLVM Phabricator instance.

Differential D52959

[WebAssembly] Handle V128 register class in explicit locals pass
ClosedPublic

Authored by tlively on Oct 5 2018, 8:37 PM.

Details

Reviewers
aheejin
dschuff
Commits
rG409f5840a7b6: [WebAssembly] Handle V128 register class in explicit locals pass
rL344094: [WebAssembly] Handle V128 register class in explicit locals pass
Summary

Also add tests to catch crashes in passes that are not normally run in
tests. Depends on D52948.

Diff Detail

Repository
rL LLVM

Event Timeline

tlively created this revision.Oct 5 2018, 8:37 PM
Herald added subscribers: llvm-commits, sunfish, jgravelle-google, sbc100. · View Herald TranscriptOct 5 2018, 8:37 PM
Harbormaster completed remote builds in B23541: Diff 168560.Oct 5 2018, 8:38 PM
tlively added a child revision: D53009: [WebAssembly] Saturating float-to-int builtins.Oct 8 2018, 5:32 PM
tlively removed a child revision: D53009: [WebAssembly] Saturating float-to-int builtins.Oct 8 2018, 5:32 PM
aheejin accepted this revision.Oct 9 2018, 8:09 AM
aheejin added inline comments.
test/CodeGen/WebAssembly/simd-arith.ll
8 ↗(On Diff #168560)

Maybe elaborate in the comment that these commands do not have -wasm-disable-explicit-locals

This revision is now accepted and ready to land.Oct 9 2018, 8:09 AM
tlively updated this revision to Diff 168903.Oct 9 2018, 4:34 PM
tlively marked an inline comment as done.
  • augment comment
tlively removed a parent revision: D52948: [WebAssembly] Fix fneg lowering.Oct 9 2018, 4:34 PM
Harbormaster completed remote builds in B23618: Diff 168903.Oct 9 2018, 4:35 PM
Closed by commit rL344094: [WebAssembly] Handle V128 register class in explicit locals pass (authored by tlively). · Explain WhyOct 9 2018, 4:35 PM
This revision was automatically updated to reflect the committed changes.

Revision Contents

PathSize
llvm/
trunk/
lib/
Target/
WebAssembly/
2 lines
test/
CodeGen/
WebAssembly/
4 lines

Diff 168904

llvm/trunk/lib/Target/WebAssembly/WebAssemblyExplicitLocals.cpp

Show First 20 Lines • Show All 151 Lines • ▼ Show 20 Linesstatic MVT typeForRegClass(const TargetRegisterClass *RC) {
if (RC == &WebAssembly::I32RegClass) if (RC == &WebAssembly::I32RegClass)
return MVT::i32; return MVT::i32;
if (RC == &WebAssembly::I64RegClass) if (RC == &WebAssembly::I64RegClass)
return MVT::i64; return MVT::i64;
if (RC == &WebAssembly::F32RegClass) if (RC == &WebAssembly::F32RegClass)
return MVT::f32; return MVT::f32;
if (RC == &WebAssembly::F64RegClass) if (RC == &WebAssembly::F64RegClass)
return MVT::f64; return MVT::f64;
if (RC == &WebAssembly::V128RegClass)
return MVT::v16i8;
if (RC == &WebAssembly::EXCEPT_REFRegClass) if (RC == &WebAssembly::EXCEPT_REFRegClass)
return MVT::ExceptRef; return MVT::ExceptRef;
llvm_unreachable("unrecognized register class"); llvm_unreachable("unrecognized register class");
} }
/// Given a MachineOperand of a stackified vreg, return the instruction at the /// Given a MachineOperand of a stackified vreg, return the instruction at the
/// start of the expression tree. /// start of the expression tree.
static MachineInstr *findStartOfTree(MachineOperand &MO, static MachineInstr *findStartOfTree(MachineOperand &MO,
▲ Show 20 Lines • Show All 230 Lines • Show Last 20 Lines

llvm/trunk/test/CodeGen/WebAssembly/simd-arith.ll

; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -wasm-enable-unimplemented-simd -mattr=+simd128 | FileCheck %s --check-prefixes CHECK,SIMD128,SIMD128-SLOW ; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -wasm-enable-unimplemented-simd -mattr=+simd128 | FileCheck %s --check-prefixes CHECK,SIMD128,SIMD128-SLOW
; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -wasm-enable-unimplemented-simd -mattr=+simd128 -fast-isel | FileCheck %s --check-prefixes CHECK,SIMD128,SIMD128-FAST ; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -wasm-enable-unimplemented-simd -mattr=+simd128 -fast-isel | FileCheck %s --check-prefixes CHECK,SIMD128,SIMD128-FAST
; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+simd128 | FileCheck %s --check-prefixes CHECK,SIMD128-VM ; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+simd128 | FileCheck %s --check-prefixes CHECK,SIMD128-VM
; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+simd128 -fast-isel | FileCheck %s --check-prefixes CHECK,SIMD128-VM ; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+simd128 -fast-isel | FileCheck %s --check-prefixes CHECK,SIMD128-VM
; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=-simd128 | FileCheck %s --check-prefixes CHECK,NO-SIMD128 ; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=-simd128 | FileCheck %s --check-prefixes CHECK,NO-SIMD128
; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=-simd128 -fast-isel | FileCheck %s --check-prefixes CHECK,NO-SIMD128 ; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=-simd128 -fast-isel | FileCheck %s --check-prefixes CHECK,NO-SIMD128
; check that a non-test run (including explicit locals pass) at least finishes
; RUN: llc < %s -O0 -wasm-enable-unimplemented-simd -mattr=+simd128,+sign-ext
; RUN: llc < %s -O2 -wasm-enable-unimplemented-simd -mattr=+simd128,+sign-ext
; Test that basic SIMD128 arithmetic operations assemble as expected. ; Test that basic SIMD128 arithmetic operations assemble as expected.
target datalayout = "e-m:e-p:32:32-i64:64-n32:64-S128" target datalayout = "e-m:e-p:32:32-i64:64-n32:64-S128"
target triple = "wasm32-unknown-unknown" target triple = "wasm32-unknown-unknown"
; ============================================================================== ; ==============================================================================
; 16 x i8 ; 16 x i8
; ============================================================================== ; ==============================================================================
▲ Show 20 Lines • Show All 888 Lines • Show Last 20 Lines