[PATCH] D21437: AMDGPU: Fold more custom nodes to undef

Matt Arsenault via llvm-commits llvm-commits at lists.llvm.org
Fri Jun 17 22:11:36 PDT 2016


arsenm added inline comments.

================
Comment at: test/CodeGen/AMDGPU/vi-removed-intrinsics.ll:16-24
@@ +15,10 @@
+
+!llvm.dbg.cu = !{!0}
+!llvm.module.flags = !{!2, !3}
+
+!0 = distinct !DICompileUnit(language: DW_LANG_OpenCL, file: !1, isOptimized: false, runtimeVersion: 0, emissionKind: NoDebug)
+!1 = !DIFile(filename: "foo.cl", directory: "/dev/null")
+!2 = !{i32 2, !"Dwarf Version", i32 4}
+!3 = !{i32 2, !"Debug Info Version", i32 3}
+!4 = !DILocation(line: 1, column: 42, scope: !5)
+!5 = distinct !DISubprogram(name: "rsq_legacy_f32", scope: null, line: 1, isLocal: false, isDefinition: true, scopeLine: 2, isOptimized: false, unit: !0)
----------------
tstellarAMD wrote:
> Is all this necessary?
Yes. Most of these things are surprisingly required


http://reviews.llvm.org/D21437





More information about the llvm-commits mailing list