[llvm] [WebAssembly] Fix f16x8.demote_f32x4_zero opcode. (PR #194722)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-webassembly

Author: Brendan Dahl (brendandahl)

<details>
<summary>Changes</summary>

Use the correct opcode from:
https://github.com/WebAssembly/half-precision/blob/f267a3d54432e5723dcc13ad4530c3581a0cc4b3/proposals/half-precision/Overview.md

---
Full diff: https://github.com/llvm/llvm-project/pull/194722.diff


1 Files Affected:

- (modified) llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td (+1-1) 


``````````diff
diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td b/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
index 38487d6ac7e0f..b6e69d617cd07 100644
--- a/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
+++ b/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
@@ -1654,7 +1654,7 @@ def demote_t : SDTypeProfile<1, 1, [SDTCisVec<0>, SDTCisVec<1>]>;
 def demote_zero : SDNode<"WebAssemblyISD::DEMOTE_ZERO", demote_t>;
 defm "" : SIMDConvert<F32x4, F64x2, demote_zero,
                       "demote_f64x2_zero", 0x5e>;
-defm "" : HalfPrecisionConvert<F16x8, F32x4, demote_zero, "demote_f32x4_zero", 0x14c>;
+defm "" : HalfPrecisionConvert<F16x8, F32x4, demote_zero, "demote_f32x4_zero", 0x149>;
 
 def promote_t : SDTypeProfile<1, 1, [SDTCisVec<0>, SDTCisVec<1>]>;
 def promote_low : SDNode<"WebAssemblyISD::PROMOTE_LOW", promote_t>;

``````````

</details>


https://github.com/llvm/llvm-project/pull/194722