[AArch64] Handle ISD::LRINT and ISD::LLRINT

[zatrazz]

Summary

This patch optimizes ISD::LRINT and ISD::LLRINT to frintx plus
fcvtzs. It currently only handles the scalar version.

This patch depends on https://reviews.llvm.org/D62017.

[zatrazz]

Ping now that both D62017 and D62019 have been approved.

[mstorsjo]

llvm/lib/Target/AArch64/AArch64InstrInfo.td

3159

Please add patterns for returning i32 as well; the original C standard math function, which is turned into the llvm IR intrinsic, returns long, which is 32 bit on windows platforms. See D62108 for a fix for the same for lround.

[SjoerdMeijer]

llvm/lib/Target/AArch64/AArch64InstrInfo.td

3159

And also a pattern for (lrint f16)?

[zatrazz]

llvm/lib/Target/AArch64/AArch64InstrInfo.td

3159

I plan to address lrint and lround for f16 in a subsequent patch.

[zatrazz]

Updated patch based on previous comments. Add a pattern to handle i32 return and a testcase for Windows.

[zatrazz]

Is this patch ok with windows support now?

[zatrazz]

Ping.

[mstorsjo]

In D62018#1513546, @zatrazz wrote:

Is this patch ok with windows support now?

Sorry, I was away for a bit. Yes, it looks good wrt Windows now - thanks for taking the time to fix it!