[PATCH] D159221: [RFC] Alterative to https://reviews.llvm.org/D159169

[Lei Wang via Phabricator via llvm-commits]

wlei added a comment.

In D159221#4632279 <https://reviews.llvm.org/D159221#4632279>, @wenlei wrote:

> Thanks for attempting the clean up. I think the original version is probably more readable. I left comments there. wdyt?

Sounds good, abandoning this, will work on the original patch.



================
Comment at: llvm/lib/Transforms/IPO/SampleProfile.cpp:2131
+    } else if (const auto *CB = dyn_cast<CallBase>(&I)) {
+      if (isa<IntrinsicInst>(&I))
+        return CalleeName;
----------------
wenlei wrote:
> I guess you want to exclude debug or probe instrinsics, but I'm still not sure if we should exclude all intrinsics. `llvm.memcpy` is also an intrinsic, but is probably still a good anchor?  
It should be ok to exclude all the IntrinsicInst because this is consistent with the sample loader/inliner, I did a search, it looks in inliner we all have the `  if (isa<IntrinsicInst>(CB)) continue;` right after the `auto *CB = dyn_cast<CallBase>(&I);`. 


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D159221/new/

https://reviews.llvm.org/D159221