diff options
| author | Sanjay Patel <spatel@rotateright.com> | 2017-02-21 19:33:53 +0000 |
|---|---|---|
| committer | Sanjay Patel <spatel@rotateright.com> | 2017-02-21 19:33:53 +0000 |
| commit | cb731f1538835f874bfa3ab3f0e57bb93a490a32 (patch) | |
| tree | 9f795bcc9f84a5f99baed76eb521daae4da2b484 /llvm/lib/Analysis/LoopAccessAnalysis.cpp | |
| parent | f3ffe75a1be894c3259e37ecc59724328cdab227 (diff) | |
| download | bcm5719-llvm-cb731f1538835f874bfa3ab3f0e57bb93a490a32.tar.gz bcm5719-llvm-cb731f1538835f874bfa3ab3f0e57bb93a490a32.zip | |
[InstCombine] canonicalize non-obivous forms of integer min/max
This is part of trying to clean up our handling of min/max patterns in IR.
By converting these to canonical form, we're more likely to recognize them
because there are various places in InstCombine that don't use
matchSelectPattern or m_SMax and friends.
The backend fixups referenced in the now deleted TODO comment were added with:
https://reviews.llvm.org/rL291392
https://reviews.llvm.org/rL289738
If there's any codegen fallout from this change, we should be able to address
it in DAGCombiner or target-specific lowering.
llvm-svn: 295758
Diffstat (limited to 'llvm/lib/Analysis/LoopAccessAnalysis.cpp')
0 files changed, 0 insertions, 0 deletions
