Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[RISCV] Make sure ADDI replacement in optimizeCondBranch has a virtual reg destination. #81938

Merged
merged 2 commits into from
Feb 16, 2024

Conversation

topperc
Copy link
Collaborator

@topperc topperc commented Feb 15, 2024

If it isn't virtual, we may extend the live range of the physical register past were it is valid. For example, across a call.

Found while trying to enable -riscv-enable-sink-fold which enables some copy propagation in machine sink that led to ADDIs with physical register destinations.

We're can replace a virtual register ADDI with an ADDI with a
physical register destination. We don't check if its safe to extend
the live range of a phyical register so we shouldn't do this.
@llvmbot
Copy link
Collaborator

llvmbot commented Feb 15, 2024

@llvm/pr-subscribers-backend-risc-v

Author: Craig Topper (topperc)

Changes

If it isn't virtual, we may extend the live range of the physical register past were it is valid. For example, across a call.


Full diff: https://github.com/llvm/llvm-project/pull/81938.diff

2 Files Affected:

  • (modified) llvm/lib/Target/RISCV/RISCVInstrInfo.cpp (+2-1)
  • (added) llvm/test/CodeGen/RISCV/branch-opt.mir (+68)
diff --git a/llvm/lib/Target/RISCV/RISCVInstrInfo.cpp b/llvm/lib/Target/RISCV/RISCVInstrInfo.cpp
index 225a9db8f3ee11..af7c40d0ca1ec6 100644
--- a/llvm/lib/Target/RISCV/RISCVInstrInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVInstrInfo.cpp
@@ -1228,7 +1228,8 @@ bool RISCVInstrInfo::optimizeCondBranch(MachineInstr &MI) const {
     MachineBasicBlock::reverse_iterator II(&MI), E = MBB->rend();
     auto DefC1 = std::find_if(++II, E, [&](const MachineInstr &I) -> bool {
       int64_t Imm;
-      return isLoadImm(&I, Imm) && Imm == C1;
+      return isLoadImm(&I, Imm) && Imm == C1 &&
+             I.getOperand(0).getReg().isVirtual();
     });
     if (DefC1 != E)
       return DefC1->getOperand(0).getReg();
diff --git a/llvm/test/CodeGen/RISCV/branch-opt.mir b/llvm/test/CodeGen/RISCV/branch-opt.mir
new file mode 100644
index 00000000000000..ba3a20f2fbfcd3
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/branch-opt.mir
@@ -0,0 +1,68 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 4
+# RUN: llc %s -mtriple=riscv64 -run-pass=peephole-opt -o - | FileCheck %s
+
+# Make sure we shouldn't replace the %2 ADDI with the $x10 ADDI since it has a
+# physical register destination.
+
+--- |
+  define void @foo(i32 signext %0) {
+    tail call void @bar(i32 1)
+    %2 = icmp ugt i32 %0, 1
+    br i1 %2, label %3, label %4
+
+  3:                                                ; preds = %1
+    tail call void @bar(i32 3)
+    ret void
+
+  4:                                                ; preds = %1
+    ret void
+  }
+
+  declare void @bar(...)
+
+...
+---
+name:            foo
+tracksRegLiveness: true
+body:             |
+  ; CHECK-LABEL: name: foo
+  ; CHECK: bb.0 (%ir-block.1):
+  ; CHECK-NEXT:   successors: %bb.1(0x40000000), %bb.2(0x40000000)
+  ; CHECK-NEXT:   liveins: $x10
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:gpr = COPY $x10
+  ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def dead $x2, implicit $x2
+  ; CHECK-NEXT:   $x10 = ADDI $x0, 1
+  ; CHECK-NEXT:   PseudoCALL target-flags(riscv-call) @bar, csr_ilp32_lp64, implicit-def dead $x1, implicit $x10, implicit-def $x2
+  ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def dead $x2, implicit $x2
+  ; CHECK-NEXT:   [[ADDI:%[0-9]+]]:gpr = ADDI $x0, 2
+  ; CHECK-NEXT:   BLTU [[COPY]], killed [[ADDI]], %bb.2
+  ; CHECK-NEXT:   PseudoBR %bb.1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.1 (%ir-block.3):
+  ; CHECK-NEXT:   $x10 = ADDI $x0, 3
+  ; CHECK-NEXT:   PseudoTAIL target-flags(riscv-call) @bar, implicit $x2, implicit $x10
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.2 (%ir-block.4):
+  ; CHECK-NEXT:   PseudoRET
+  bb.0 (%ir-block.1):
+    successors: %bb.1, %bb.2
+    liveins: $x10
+
+    %0:gpr = COPY $x10
+    ADJCALLSTACKDOWN 0, 0, implicit-def dead $x2, implicit $x2
+    $x10 = ADDI $x0, 1
+    PseudoCALL target-flags(riscv-call) @bar, csr_ilp32_lp64, implicit-def dead $x1, implicit $x10, implicit-def $x2
+    ADJCALLSTACKUP 0, 0, implicit-def dead $x2, implicit $x2
+    %2:gpr = ADDI $x0, 2
+    BLTU %0, killed %2, %bb.2
+    PseudoBR %bb.1
+
+  bb.1 (%ir-block.3):
+    $x10 = ADDI $x0, 3
+    PseudoTAIL target-flags(riscv-call) @bar, implicit $x2, implicit $x10
+
+  bb.2 (%ir-block.4):
+    PseudoRET
+
+...

Copy link
Member

@mshockwave mshockwave left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@topperc topperc merged commit feee627 into llvm:main Feb 16, 2024
5 of 6 checks passed
@topperc topperc deleted the pr/condbranch-bug branch February 16, 2024 00:34
llvmbot pushed a commit to llvmbot/llvm-project that referenced this pull request Feb 16, 2024
…l reg destination. (llvm#81938)

If it isn't virtual, we may extend the live range of the physical
register past were it is valid. For example, across a call.

Found while trying to enable -riscv-enable-sink-fold which enables some
copy propagation in machine sink that led to ADDIs with physical
register destinations.

(cherry picked from commit feee627)
llvmbot pushed a commit to llvmbot/llvm-project that referenced this pull request Feb 16, 2024
…l reg destination. (llvm#81938)

If it isn't virtual, we may extend the live range of the physical
register past were it is valid. For example, across a call.

Found while trying to enable -riscv-enable-sink-fold which enables some
copy propagation in machine sink that led to ADDIs with physical
register destinations.

(cherry picked from commit feee627)
@pointhex pointhex mentioned this pull request May 7, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants