[llvm] [AMDGPU][AsmParser][NFC] Remove a misleading comment. (PR #102604)
Ivan Kosarev via llvm-commits
llvm-commits at lists.llvm.org
Fri Aug 9 05:12:54 PDT 2024
https://github.com/kosarev created https://github.com/llvm/llvm-project/pull/102604
The work of ParseRegularReg() should remain to be parsing the register as it was specified, and not to try translate it to anything else.
It's up to operand predicates to decide on what is and is not an acceptable register for an operand, including considering its expected register class, and for the rest of the AsmParser infrastructure to handle it respectively from there on.
>From 4ceb4f8f09680c8994186207f161aee104ffdb3f Mon Sep 17 00:00:00 2001
From: Ivan Kosarev <ivan.kosarev at amd.com>
Date: Fri, 9 Aug 2024 12:56:46 +0100
Subject: [PATCH] [AMDGPU][AsmParser][NFC] Remove a misleading comment.
The work of ParseRegularReg() should remain to be parsing the register as
it was specified, and not to try translate it to anything else.
It's up to operand predicates to decide on what is and is not an
acceptable register for an operand, including considering its expected
register class, and for the rest of the AsmParser infrastructure to
handle it respectively from there on.
---
llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp | 4 ----
1 file changed, 4 deletions(-)
diff --git a/llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp b/llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
index bbd9d75aac0e91..fc4cf57281d9b0 100644
--- a/llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
+++ b/llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
@@ -2937,10 +2937,6 @@ unsigned AMDGPUAsmParser::ParseRegularReg(RegisterKind &RegKind,
StringRef RegSuffix = RegName.substr(RI->Name.size());
unsigned SubReg = NoSubRegister;
if (!RegSuffix.empty()) {
- // We don't know the opcode till we are done parsing, so we don't know if
- // registers should be 16 or 32 bit. It is therefore mandatory to put .l or
- // .h to correctly specify 16 bit registers. We also can't determine class
- // VGPR_16_Lo128 or VGPR_16, so always parse them as VGPR_16.
if (RegSuffix.consume_back(".l"))
SubReg = AMDGPU::lo16;
else if (RegSuffix.consume_back(".h"))
More information about the llvm-commits
mailing list