[llvm-branch-commits] [llvm] [X86] Avoid generating nested CALLSEQ for TLS pointer function arguments (PR #106965)

[Matt Arsenault via llvm-branch-commits]

================
@@ -35971,6 +35971,15 @@ X86TargetLowering::EmitLoweredTLSAddr(MachineInstr &MI,
   // inside MC, therefore without the two markers shrink-wrapping
   // may push the prologue/epilogue pass them.
   const TargetInstrInfo &TII = *Subtarget.getInstrInfo();
+
----------------
arsenm wrote:

I don't see why EmitLoweredTLS needs to be handled by raw MI emission. Why can't this just emit the call sequence during legalization, like for an ordinary call or dynamic_alloca? If you did that, you would avoid appearing in a call sequence in the first place 

https://github.com/llvm/llvm-project/pull/106965