[AMDGPU] Only allow global fp atomics with unsafe option

[rampitec]

Summary

Previously we were allowing to use FP atomics without
-amdgpu-unsafe-fp-atomics option if a scope is less then
system. This is not safe just as well if we have UC memory.

This change only allows global and flat FP atomics with
the unsafe option. Consequentially that makes a check for
denorm mode redundant since we skip it with the unsafe
option and do not have a way to produce these instructions
without it anyway.

[t-tye]

llvm/lib/Target/AMDGPU/SIISelLowering.cpp

12147–12153

Should the amdgpu-unsafe-fp-atomics also ally here too? If not should the attribute be named to reflect it only allies to global fp atomics?

[rampitec]

llvm/lib/Target/AMDGPU/SIISelLowering.cpp

12147–12153

This is LDS, so does not have the same problem with mtypes. We can skip denorms check, but that would be a separate patch.

[rampitec]

Added tests for safe case.

[rampitec]

llvm/lib/Target/AMDGPU/SIISelLowering.cpp

12147–12153

D102366

[t-tye]

LGTM

Added @b-sumner to review.

[b-sumner]

This looks good to me.

llvm/lib/Target/AMDGPU/SIISelLowering.cpp

case AtomicRMWInst::FAdd: {
     unsigned AS = RMW->getPointerAddressSpace();
 
     if ((AS == AMDGPUAS::GLOBAL_ADDRESS || AS == AMDGPUAS::FLAT_ADDRESS) &&
          Subtarget->hasAtomicFaddInsts()) {
       // The amdgpu-unsafe-fp-atomics attribute enables generation of unsafe
       // floating point atomic instructions. May generate more efficient code,
       // but may not respect rounding and denormal modes, and may give incorrect
       // results for certain memory destinations.
-      if (!fpModeMatchesGlobalFPAtomicMode(RMW) &&
-          RMW->getFunction()
-                  ->getFnAttribute("amdgpu-unsafe-fp-atomics")
-                  .getValueAsString() != "true")
+      if (RMW->getFunction()
         return AtomicExpansionKind::CmpXChg;
 
       if (Subtarget->hasGFX90AInsts()) {
         auto SSID = RMW->getSyncScopeID();
         if (SSID == SyncScope::System ||
             SSID == RMW->getContext().getOrInsertSyncScopeID("one-as"))
           return AtomicExpansionKind::CmpXChg;
 
         return (Ty->isFloatTy() && AS == AMDGPUAS::FLAT_ADDRESS) ?
           AtomicExpansionKind::CmpXChg : AtomicExpansionKind::None;
       }
 
       if (!Subtarget->hasGFX90AInsts() && AS != AMDGPUAS::GLOBAL_ADDRESS)
         return AtomicExpansionKind::CmpXChg;
 
       return RMW->use_empty() ? AtomicExpansionKind::None :
                                 AtomicExpansionKind::CmpXChg;
     }
 
     // DS FP atomics do repect the denormal mode, but the rounding mode is fixed
     // to round-to-nearest-even.
     // The only exception is DS_ADD_F64 which never flushes regardless of mode.
     if (AS == AMDGPUAS::LOCAL_ADDRESS && Subtarget->hasLDSFPAtomics()) {
       return (Ty->isDoubleTy() && !fpModeMatchesGlobalFPAtomicMode(RMW)) ?
         AtomicExpansionKind::CmpXChg : AtomicExpansionKind::None;
     }
 
     return AtomicExpansionKind::CmpXChg;
   }
   default:
     break;
   }
 
   return AMDGPUTargetLowering::shouldExpandAtomicRMWInIR(RMW);