================
@@ -1660,7 +2003,24 @@ static bool runImpl(Module &M, AnalysisGetter &AG,
TargetMachine &TM,
}
}
- return A.run() == ChangeStatus::CHANGED;
+ bool Changed = A.run() == ChangeStatus::CHANGED;
+
+ // Kernel attribute lowering (merged from AMDGPULowerKernelAttributesPass)
----------------
arsenm wrote:
This is just moving the code into the pass, which is still functionally a
different pass. This isn't really using the frame work. The framework should be
used to identify the call sites and do the rewrite instead of the old manual
way
https://github.com/llvm/llvm-project/pull/177432
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
