[all-commits] [llvm/llvm-project] 07de5d: [SCEV] Remember blocks for which we know symbolic ...

[Max Kazantsev via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 07de5d18c90182f5d38fd9b9b3911375d045ff2a
      https://github.com/llvm/llvm-project/commit/07de5d18c90182f5d38fd9b9b3911375d045ff2a
  Author: Max Kazantsev <mkazantsev at azul.com>
  Date:   2022-12-07 (Wed, 07 Dec 2022)

  Changed paths:
    M llvm/lib/Analysis/ScalarEvolution.cpp
    M llvm/test/Analysis/ScalarEvolution/addrec-computed-during-addrec-calculation.ll
    M llvm/test/Analysis/ScalarEvolution/symbolic_max_exit_count.ll
    M llvm/test/Analysis/ScalarEvolution/trip-count10.ll
    M llvm/test/Analysis/ScalarEvolution/trip-count7.ll
    M llvm/test/Analysis/ScalarEvolution/widenable-condition.ll

  Log Message:
  -----------
  [SCEV] Remember blocks for which we know symbolic exit count but not exact

The old code didn't bother to memoize blocks for which exact exit count is not
known. As result, in situation when exact isn't known but symbolic is known, this
info was lost. This patch fixes the situation: now we memoize when symbolic is
known (exact always implies symbolic, so this is a strict superset of what was before).

Differential Revision: https://reviews.llvm.org/D139515
Reviewed By: nikic