[llvm] [DirectX] Implement the ForwardHandleAccesses pass (PR #135378)

Helena Kotas via llvm-commits llvm-commits at lists.llvm.org
Fri Apr 11 16:31:58 PDT 2025


================
@@ -0,0 +1,21 @@
+; RUN: not opt -S -dxil-forward-handle-accesses -mtriple=dxil--shadermodel6.3-library %s 2>&1 | FileCheck %s
+
+; CHECK: error: Handle at "h2" overwrites handle at "h1"
+
+%"class.hlsl::RWStructuredBuffer" = type { target("dx.RawBuffer", <4 x float>, 1, 0) }
+ at Buf = internal global %"class.hlsl::RWStructuredBuffer" poison, align 4
+
+define float @f() {
+entry:
+  %h1 = call target("dx.RawBuffer", <4 x float>, 1, 0) @llvm.dx.resource.handlefrombinding(i32 0, i32 0, i32 1, i32 0, i1 false)
+  store target("dx.RawBuffer", <4 x float>, 1, 0) %h1, ptr @Buf, align 4
+  %h2 = call target("dx.RawBuffer", <4 x float>, 1, 0) @llvm.dx.resource.handlefrombinding(i32 0, i32 1, i32 1, i32 0, i1 false)
+  store target("dx.RawBuffer", <4 x float>, 1, 0) %h2, ptr @Buf, align 4
+
+  %b = load target("dx.RawBuffer", <4 x float>, 1, 0), ptr @Buf, align 4
+  %l = call { <4 x float>, i1 } @llvm.dx.resource.load.rawbuffer(target("dx.RawBuffer", <4 x float>, 1, 0) %b, i32 0, i32 0)
+  %x = extractvalue { <4 x float>, i1 } %l, 0
+  %v = extractelement <4 x float> %x, i32 0
----------------
hekota wrote:

Nit - not really related to what this is testing. If the type of the "dx.RawBuffer" is `<4 x float>`, should it be returning the vector `<4 x float>` instead of struct `{<4 x float>}`, and then the `extractvalue` is not needed? Or am I missing something here?

https://github.com/llvm/llvm-project/pull/135378


More information about the llvm-commits mailing list