[PATCH] D88256: [NFC][regalloc] Separate iteration from AllocationOrder

[Mircea Trofin via Phabricator via llvm-commits]

mtrofin created this revision.
mtrofin added a reviewer: wmi.
Herald added subscribers: llvm-commits, hiraditya, qcolombet, MatzeB.
Herald added a project: LLVM.
mtrofin requested review of this revision.

This separates the two concerns - encapsulation of traversal order; and
iteration. As a deviation from the original design, setting a limit to
the number of Order registers is done upfront - this is the way the API
was used previously anyway, albeit it would have allowed changing the
limit at each call to next(). That's unnecessary complexity.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D88256

Files:
  llvm/lib/CodeGen/AllocationOrder.cpp
  llvm/lib/CodeGen/AllocationOrder.h
  llvm/lib/CodeGen/RegAllocBasic.cpp
  llvm/lib/CodeGen/RegAllocGreedy.cpp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D88256.294139.patch
Type: text/x-patch
Size: 7844 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200924/e0d7dbd4/attachment.bin>