[RISCV] Use hasStdExtCOrZca instead of hasStdExtC in estimateFunctionSizeInBytes.

[topperc]

I'm not sure how to test this. I commented out the existing hasStdC check and no tests failed.

[llvmbot]

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

Author: Craig Topper (topperc)

Changes

I'm not sure how to test this. I commented out the existing hasStdC check and no tests failed.

---

Full diff: https://github.com/llvm/llvm-project/pull/80905.diff

1 Files Affected:

• (modified) llvm/lib/Target/RISCV/RISCVFrameLowering.cpp (+1-1)

diff --git a/llvm/lib/Target/RISCV/RISCVFrameLowering.cpp b/llvm/lib/Target/RISCV/RISCVFrameLowering.cpp
index d793c0b7377ba..745ae68123b08 100644
--- a/llvm/lib/Target/RISCV/RISCVFrameLowering.cpp
+++ b/llvm/lib/Target/RISCV/RISCVFrameLowering.cpp
@@ -1172,7 +1172,7 @@ static unsigned estimateFunctionSizeInBytes(const MachineFunction &MF,
       if (MI.isConditionalBranch())
         FnSize += TII.getInstSizeInBytes(MI);
       if (MI.isConditionalBranch() || MI.isUnconditionalBranch()) {
-        if (MF.getSubtarget<RISCVSubtarget>().hasStdExtC())
+        if (MF.getSubtarget<RISCVSubtarget>().hasStdExtCOrZca())
           FnSize += 2 + 8 + 2 + 2;
         else
           FnSize += 4 + 8 + 4 + 4;

[preames]

LGTM

With the assembler changes from a while back, this should not be simply a heuristic.