[PATCH] D95136: [X86] Fix tile config register spill issue.

[Pengfei Wang via Phabricator via llvm-commits]

pengfei marked 3 inline comments as done.
pengfei added inline comments.


================
Comment at: llvm/lib/Target/X86/X86PreTileConfig.cpp:283
+    // But iterating all BBs isn't too expensive here.
+    for (auto I : BBVisitedInfo) {
+      if (I.second.NeedUpdatePred)
----------------
LuoYuanke wrote:
> xiangzhangllvm wrote:
> > The logic here is complex, better to add clear comment to show how we update the Status for BBs.
> WorkList should be clear before this loop?
Not necessary. But adding it seems more robust.


================
Comment at: llvm/lib/Target/X86/X86PreTileConfig.cpp:302
+            // If BB has AMX, its predecessors don't need to be updated again.
+            if (!BBVisitedInfo[*I].HasAMX)
+              BBVisitedInfo[*I].NeedUpdatePred = true;
----------------
LuoYuanke wrote:
> Don't understand this. Can you add a diagram to illustrate it?
Add some comments on line 290.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D95136/new/

https://reviews.llvm.org/D95136