[PATCH] D124631: [TwoAddressInstructionPass] Special processing of STATEPOINT instruction.

[Matthias Braun via Phabricator via llvm-commits]

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

Well I don't have too much experience with `STATEPOINT`s, but the change makes sense to me so far.

Please take another look if you can simplify/shrink the `statepoint-vreg-twoaddr.mir` test so future readers have an easier time.

Either way LGTM.



================
Comment at: llvm/test/CodeGen/X86/statepoint-vreg-twoaddr.mir:6-35
+--- |
+  ; ModuleID = 'statepoint-vreg-twoaddr.ll'
+  source_filename = "statepoint-vreg-twoaddr.ll"
+  target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+  target triple = "x86_64-pc-linux-gnu"
+
+  declare i1 @return_i1()
----------------
So you need the LLVM-IR here? It feels to me like we can drop it without affecting the test.


================
Comment at: llvm/test/CodeGen/X86/statepoint-vreg-twoaddr.mir:60-70
+    %0:gr64 = COPY killed $rdi
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %1:gr64 = STATEPOINT 0, 0, 0, target-flags(x86-plt) @return_i1, 2, 0, 2, 0, 2, 0, 2, 1, killed %0(tied-def 0), 2, 0, 2, 1, 0, 0, csr_64, implicit-def $rsp, implicit-def $ssp, implicit-def $al
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %2:gr8 = COPY killed $al
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    $rdi = COPY killed %1
----------------
Is it possible to further simplify this to focus on the thing being tested? Does the following change work:


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D124631/new/

https://reviews.llvm.org/D124631