Ahmed Bougacha 51eefd689e [CodeGen] Don't assume fp_to_fp16 produces i16 when legalizing it. ...

Since r230276, we support an improved legalization for f64->f16,
which goes through a temporary f32, improving codegen when
f32->f16 is legal but not f64->f16. This requires unsafe-fp-math.

However, that legalization assumed that the second step, producing
a pseudo-softened f16, had type i16. That's not true on targets
with illegal i16, such as ARM.

Use the initial f64->f16 result type instead.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257794 91177308-0d34-0410-b5e6-96231b3b80d8

2016-01-14 19:45:36 +00:00

..

AArch64

[AArch64] Don't assume extractelt constant index when matching shuffle.

2016-01-14 02:12:30 +00:00

AMDGPU

AMDGPU/SI: Update ISA version for FIJI

2016-01-13 20:39:25 +00:00

ARM

[CodeGen] Don't assume fp_to_fp16 produces i16 when legalizing it.

2016-01-14 19:45:36 +00:00

BPF

Revert "Change memcpy/memset/memmove to have dest and source alignments."

2015-11-19 05:56:52 +00:00

CPP

Fix CPP Backend for GEP API changes for opaque pointer types

2015-09-08 18:42:29 +00:00

Generic

Unxfail passing testcase on Hexagon

2016-01-14 15:24:15 +00:00

Hexagon

[Hexagon] Expand pseudo instruction Insert4

2016-01-14 15:37:16 +00:00

Inputs

DI: Reverse direction of subprogram -> function edge.

2015-11-05 22:03:56 +00:00

Mips

[mips] Correct operand order in DSP's mthi/mtlo

2016-01-12 15:15:14 +00:00

MIR

Re-Revert r257105 (Verifier debug info changes)

2016-01-13 02:31:14 +00:00

MSP430

Revert "Change memcpy/memset/memmove to have dest and source alignments."

2015-11-19 05:56:52 +00:00

NVPTX

Have a single way for creating unique value names.

2015-11-22 00:16:24 +00:00

PowerPC

[PowerPC] Fix large code model with the ELFv2 ABI

2016-01-13 13:12:23 +00:00

SPARC

[SPARC] Revamp AnalyzeBranch and add ReverseBranchCondition.

2016-01-13 04:44:14 +00:00

SystemZ

[SystemZ] Fix assertion failure in adjustSubwordCmp

2015-12-16 18:04:06 +00:00

Thumb

[Thumb] Fix assembler error 'cannot honor width suffix pop {lr}'

2015-12-28 21:40:45 +00:00

Thumb2

Fix test case label check

2015-11-20 20:24:49 +00:00

WebAssembly

[WebAssembly] MCFixupKindInfo's TargetSize is in bits rather than bytes.

2016-01-13 19:29:37 +00:00

WinEH

[WinEH] Disallow cyclic unwinds

2016-01-10 04:31:05 +00:00

X86

AVX512: VMOVDQA32/64 (load) intrinsic implementation.

2016-01-14 07:56:04 +00:00

XCore

Revert "Change memcpy/memset/memmove to have dest and source alignments."

2015-11-19 05:56:52 +00:00