[AMDGPU] Remove FlatVariant argument from isLegalFlatAddressingMode. NFC.

[rampitec]

This argument is easily deduced from AS argument.

[llvmbot]

@llvm/pr-subscribers-backend-amdgpu

Author: Stanislav Mekhanoshin (rampitec)

Changes

…NFC.

This argument is easily deduced from AS argument.

---

Full diff: https://github.com/llvm/llvm-project/pull/93938.diff

2 Files Affected:

• (modified) llvm/lib/Target/AMDGPU/SIISelLowering.cpp (+10-10)
• (modified) llvm/lib/Target/AMDGPU/SIISelLowering.h (+1-2)

diff --git a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
index f9948e92862f7..98e9d18fc201f 100644
--- a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
@@ -1484,14 +1484,18 @@ bool SITargetLowering::getAddrModeArguments(IntrinsicInst *II,
 }
 
 bool SITargetLowering::isLegalFlatAddressingMode(const AddrMode &AM,
-                                                 unsigned AddrSpace,
-                                                 uint64_t FlatVariant) const {
+                                                 unsigned AddrSpace) const {
   if (!Subtarget->hasFlatInstOffsets()) {
     // Flat instructions do not have offsets, and only have the register
     // address.
     return AM.BaseOffs == 0 && AM.Scale == 0;
   }
 
+  uint64_t FlatVariant =
+      AddrSpace == AMDGPUAS::GLOBAL_ADDRESS    ? SIInstrFlags::FlatGlobal
+      : AddrSpace == AMDGPUAS::PRIVATE_ADDRESS ? SIInstrFlags::FlatScratch
+                                               : SIInstrFlags::FLAT;
+
   return AM.Scale == 0 &&
          (AM.BaseOffs == 0 || Subtarget->getInstrInfo()->isLegalFLATOffset(
                                   AM.BaseOffs, AddrSpace, FlatVariant));
@@ -1499,8 +1503,7 @@ bool SITargetLowering::isLegalFlatAddressingMode(const AddrMode &AM,
 
 bool SITargetLowering::isLegalGlobalAddressingMode(const AddrMode &AM) const {
   if (Subtarget->hasFlatGlobalInsts())
-    return isLegalFlatAddressingMode(AM, AMDGPUAS::GLOBAL_ADDRESS,
-                                     SIInstrFlags::FlatGlobal);
+    return isLegalFlatAddressingMode(AM, AMDGPUAS::GLOBAL_ADDRESS);
 
   if (!Subtarget->hasAddr64() || Subtarget->useFlatForGlobal()) {
     // Assume the we will use FLAT for all global memory accesses
@@ -1512,8 +1515,7 @@ bool SITargetLowering::isLegalGlobalAddressingMode(const AddrMode &AM) const {
     // by setting the stride value in the resource descriptor which would
     // increase the size limit to (stride * 4GB).  However, this is risky,
     // because it has never been validated.
-    return isLegalFlatAddressingMode(AM, AMDGPUAS::FLAT_ADDRESS,
-                                     SIInstrFlags::FLAT);
+    return isLegalFlatAddressingMode(AM, AMDGPUAS::FLAT_ADDRESS);
   }
 
   return isLegalMUBUFAddressingMode(AM);
@@ -1619,8 +1621,7 @@ bool SITargetLowering::isLegalAddressingMode(const DataLayout &DL,
 
   if (AS == AMDGPUAS::PRIVATE_ADDRESS)
     return Subtarget->enableFlatScratch()
-               ? isLegalFlatAddressingMode(AM, AMDGPUAS::PRIVATE_ADDRESS,
-                                           SIInstrFlags::FlatScratch)
+               ? isLegalFlatAddressingMode(AM, AMDGPUAS::PRIVATE_ADDRESS)
                : isLegalMUBUFAddressingMode(AM);
 
   if (AS == AMDGPUAS::LOCAL_ADDRESS ||
@@ -1647,8 +1648,7 @@ bool SITargetLowering::isLegalAddressingMode(const DataLayout &DL,
     // computation. We don't have instructions that compute pointers with any
     // addressing modes, so treat them as having no offset like flat
     // instructions.
-    return isLegalFlatAddressingMode(AM, AMDGPUAS::FLAT_ADDRESS,
-                                     SIInstrFlags::FLAT);
+    return isLegalFlatAddressingMode(AM, AMDGPUAS::FLAT_ADDRESS);
   }
 
   // Assume a user alias of global for unknown address spaces.
diff --git a/llvm/lib/Target/AMDGPU/SIISelLowering.h b/llvm/lib/Target/AMDGPU/SIISelLowering.h
index 292b17da93583..c5f3a98e69061 100644
--- a/llvm/lib/Target/AMDGPU/SIISelLowering.h
+++ b/llvm/lib/Target/AMDGPU/SIISelLowering.h
@@ -223,8 +223,7 @@ class SITargetLowering final : public AMDGPUTargetLowering {
   SDValue performClampCombine(SDNode *N, DAGCombinerInfo &DCI) const;
   SDValue performRcpCombine(SDNode *N, DAGCombinerInfo &DCI) const;
 
-  bool isLegalFlatAddressingMode(const AddrMode &AM, unsigned AddrSpace,
-                                 uint64_t FlatVariant) const;
+  bool isLegalFlatAddressingMode(const AddrMode &AM, unsigned AddrSpace) const;
   bool isLegalMUBUFAddressingMode(const AddrMode &AM) const;
 
   unsigned isCFIntrinsic(const SDNode *Intr) const;

[arsenm]

It's not a 1:1 mapping though, you can select global addresses to flat instructions (which the patterns do handle)

[rampitec]

It's not a 1:1 mapping though, you can select global addresses to flat instructions (which the patterns do handle)

But every single use is 1:1.

[arsenm]

This isn't used for the selection anyway

[rampitec]

JBTW, I also believe in many cases we are using 'auto' we really mean 'decltype(something)'. I.e. whatever underlying type is. Maybe we should use it more often and less often use 'auto'.