[clang] [llvm] [BPF] Do atomic_fetch_*() pattern matching with memory ordering (PR #107343)
via cfe-commits
cfe-commits at lists.llvm.org
Tue Sep 24 01:52:07 PDT 2024
================
@@ -152,22 +152,93 @@ static bool hasLiveDefs(const MachineInstr &MI, const TargetRegisterInfo *TRI) {
return false;
}
-void BPFMIPreEmitChecking::processAtomicInsts() {
+bool BPFMIPreEmitChecking::processAtomicInsts() {
+ if (!MF->getSubtarget<BPFSubtarget>().getHasJmp32()) {
+ // Only check for cpu version 1 and 2.
+ for (MachineBasicBlock &MBB : *MF) {
+ for (MachineInstr &MI : MBB) {
+ if (MI.getOpcode() != BPF::XADDW && MI.getOpcode() != BPF::XADDD)
+ continue;
+
+ LLVM_DEBUG(MI.dump());
+ if (hasLiveDefs(MI, TRI)) {
+ DebugLoc Empty;
+ const DebugLoc &DL = MI.getDebugLoc();
+ const Function &F = MF->getFunction();
+ F.getContext().diagnose(DiagnosticInfoUnsupported{
+ F, "Invalid usage of the XADD return value", DL});
+ }
+ }
+ }
+ }
----------------
eddyz87 wrote:
In BPFISelLowering.cpp we have function `fail()`, which reports error message/location, prints a backtrace and does `exit(1)` (via default diagnostic handler). Which is not ideal either. Oh, well.
https://github.com/llvm/llvm-project/pull/107343
More information about the cfe-commits
mailing list