[llvm] [AMDGPU][NFC] Cleanly make 32-bit abs legal (PR #164837)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-amdgpu

Author: None (LU-JOHN)

<details>
<summary>Changes</summary>

Cleanly make 32-bit abs legal only in SIISelLowering.cpp

---
Full diff: https://github.com/llvm/llvm-project/pull/164837.diff


3 Files Affected:

- (modified) llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp (+2-2) 
- (modified) llvm/lib/Target/AMDGPU/R600ISelLowering.cpp (-3) 
- (modified) llvm/lib/Target/AMDGPU/SIISelLowering.cpp (+1-1) 


``````````diff
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp b/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
index 8ed4062e43946..1b559a628be08 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
@@ -514,8 +514,8 @@ AMDGPUTargetLowering::AMDGPUTargetLowering(const TargetMachine &TM,
       MVT::i64, Custom);
   setOperationAction(ISD::SELECT_CC, MVT::i64, Expand);
 
-  setOperationAction({ISD::ABS, ISD::SMIN, ISD::UMIN, ISD::SMAX, ISD::UMAX},
-                     MVT::i32, Legal);
+  setOperationAction({ISD::SMIN, ISD::UMIN, ISD::SMAX, ISD::UMAX}, MVT::i32,
+                     Legal);
 
   setOperationAction(
       {ISD::CTTZ, ISD::CTTZ_ZERO_UNDEF, ISD::CTLZ, ISD::CTLZ_ZERO_UNDEF},
diff --git a/llvm/lib/Target/AMDGPU/R600ISelLowering.cpp b/llvm/lib/Target/AMDGPU/R600ISelLowering.cpp
index 09ef6ac7bcdf2..2aa54c920a046 100644
--- a/llvm/lib/Target/AMDGPU/R600ISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/R600ISelLowering.cpp
@@ -45,9 +45,6 @@ R600TargetLowering::R600TargetLowering(const TargetMachine &TM,
   // Legalize loads and stores to the private address space.
   setOperationAction(ISD::LOAD, {MVT::i32, MVT::v2i32, MVT::v4i32}, Custom);
 
-  // 32-bit ABS is legal for AMDGPU except for R600
-  setOperationAction(ISD::ABS, MVT::i32, Expand);
-
   // EXTLOAD should be the same as ZEXTLOAD. It is legal for some address
   // spaces, so it is custom lowered to handle those where it isn't.
   for (auto Op : {ISD::SEXTLOAD, ISD::ZEXTLOAD, ISD::EXTLOAD})
diff --git a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
index a7574213c2907..be4229155c983 100644
--- a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
@@ -298,7 +298,7 @@ SITargetLowering::SITargetLowering(const TargetMachine &TM,
   setOperationAction(ISD::BR_CC,
                      {MVT::i1, MVT::i32, MVT::i64, MVT::f32, MVT::f64}, Expand);
 
-  setOperationAction({ISD::UADDO, ISD::USUBO}, MVT::i32, Legal);
+  setOperationAction({ISD::ABS, ISD::UADDO, ISD::USUBO}, MVT::i32, Legal);
 
   setOperationAction({ISD::UADDO_CARRY, ISD::USUBO_CARRY}, MVT::i32, Legal);
 

``````````

</details>


https://github.com/llvm/llvm-project/pull/164837