[PATCH] D84502: [AArch64][GlobalISel] Implement __builtin_return_address for PAC-RET

[Ahmed Bougacha via Phabricator via llvm-commits]

ab accepted this revision.
ab added a comment.
This revision is now accepted and ready to land.

This looks sensible, modulo a couple inline comments



================
Comment at: llvm/lib/Target/AArch64/GISel/AArch64InstructionSelector.cpp:4720
       EntryBuilder.setInstr(*EntryBlock.begin());
-      EntryBuilder.buildCopy({DstReg}, {Register(AArch64::LR)});
+      if (MF.getFunction().hasFnAttribute("sign-return-address")) {
+        if (STI.hasV8_3aOps()) {
----------------
I don't know what model you have for the attribute and the command line options, and I guess you're probably aware, but this seems unsafe for depths > 0.  I'm not sure there's a better way to deal with this, short of just forcing the strips in any aarch64 code, which is not an option.  Maybe emitting strips when +pa is available?  I imagine that's not at all reliable either.


================
Comment at: llvm/lib/Target/AArch64/GISel/AArch64InstructionSelector.cpp:4721
+      if (MF.getFunction().hasFnAttribute("sign-return-address")) {
+        if (STI.hasV8_3aOps()) {
+          Register TmpReg = MRI.createVirtualRegister(&AArch64::GPR64RegClass);
----------------
-> `hasPA()` ?


================
Comment at: llvm/test/CodeGen/AArch64/GlobalISel/builtin-return-address-pacret.ll:1
+;; RUN: llc -mtriple aarch64               -O0 %s -o - | FileCheck %s --check-prefixes=CHECK,CHECK-NOP
+;; RUN: llc -mtriple aarch64 -mattr=+v8.3a -O0 %s -o - | FileCheck %s --check-prefixes=CHECK,CHECK-V83
----------------
can you explicitly pass `-global-isel` ?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D84502/new/

https://reviews.llvm.org/D84502