forked from OSchip/llvm-project
95d9ccb2a0
In r263618, JumpThreading learned to look trough simple cast instructions, but only if the source of those cast instructions was a phi/cmp i1 (in an effort to limit compile time effects). I think this condition is too restrictive. For switches with limited value range, InstCombine will readily introduce an extra trunc instruction to a smaller integer type (e.g. from i8 to i2), leaving us in the somewhat perverse situation that jump-threading would work before running instcombine, but not after. Since instcombine produces this pattern, I think we need to consider it canonical and support it in JumpThreading. In general, for limiting recursion, I think the existing restriction to phi and cmp nodes should be sufficient to avoid looking through unprofitable chains of instructions. Patch by Keno Fischer! Differential Revision: https://reviews.llvm.org/D42262 llvm-svn: 327150 |
||
|---|---|---|
| .. | ||
| 2008-11-27-EntryMunge.ll | ||
| 2010-08-26-and.ll | ||
| 2011-04-02-SimplifyDeadBlock.ll | ||
| 2011-04-14-InfLoop.ll | ||
| 2012-07-19-NoSuccessorIndirectBr.ll | ||
| and-and-cond.ll | ||
| and-cond.ll | ||
| assume-edge-dom.ll | ||
| assume.ll | ||
| basic.ll | ||
| branch-no-const.ll | ||
| compare.ll | ||
| conservative-lvi.ll | ||
| crash.ll | ||
| ddt-crash.ll | ||
| ddt-crash2.ll | ||
| ddt-crash3.ll | ||
| ddt-crash4.ll | ||
| degenerate-phi.ll | ||
| fold-not-thread.ll | ||
| guards.ll | ||
| implied-cond.ll | ||
| indirectbr.ll | ||
| induction.ll | ||
| landing-pad.ll | ||
| lvi-load.ll | ||
| lvi-tristate.ll | ||
| no-irreducible-loops.ll | ||
| or-undef.ll | ||
| phi-eq.ll | ||
| phi-known.ll | ||
| pr9331.ll | ||
| pr15851_hang.ll | ||
| pr22086.ll | ||
| pr26096.ll | ||
| pr27840.ll | ||
| pr33605.ll | ||
| pr33917.ll | ||
| pr36133.ll | ||
| range-compare.ll | ||
| select.ll | ||
| static-profile.ll | ||
| thread-cmp.ll | ||
| thread-loads.ll | ||
| threading_prof1.ll | ||
| threading_prof2.ll | ||
| update-edge-weight.ll | ||