[llvm] [AMDGPU] Add hazard workarounds to insertIndirectBranch (PR #109127)
Matt Arsenault via llvm-commits
llvm-commits at lists.llvm.org
Wed Sep 18 05:46:20 PDT 2024
================
@@ -2867,6 +2867,12 @@ void SIInstrInfo::insertIndirectBranch(MachineBasicBlock &MBB,
MachineRegisterInfo &MRI = MF->getRegInfo();
const SIMachineFunctionInfo *MFI = MF->getInfo<SIMachineFunctionInfo>();
+ // Note: as this is used after hazard recognizer we need to apply some hazard
+ // workarounds directly.
+ const GCNSubtarget &ST = MF->getSubtarget<GCNSubtarget>();
----------------
arsenm wrote:
ST is already a member of SIInstrInfo
https://github.com/llvm/llvm-project/pull/109127
More information about the llvm-commits
mailing list