[llvm-bugs] [Bug 39783] New: [LoopSimplifyCFG] -enable-loop-simplifycfg-term-folding causes many assertion failures

via llvm-bugs llvm-bugs at lists.llvm.org
Mon Nov 26 02:00:17 PST 2018


https://bugs.llvm.org/show_bug.cgi?id=39783

            Bug ID: 39783
           Summary: [LoopSimplifyCFG]
                    -enable-loop-simplifycfg-term-folding causes many
                    assertion failures
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Loop Optimizer
          Assignee: unassignedbugs at nondot.org
          Reporter: paulsson at linux.vnet.ibm.com
                CC: llvm-bugs at lists.llvm.org

I got a lot of assertions from the previous weeks test runs (csmith), to find
that they were all "hidden" with a later revision. Since I did not find these
asserts here on Bugzilla, I continued with a git bisect and found out that
there has been a new option '-enable-loop-simplifycfg-term-folding' addedwhich
defaults to false, which disables the(se) bug(s).

I continued to make three reduced test cases for the different asserts I saw,
which I will post here in a minute.

And then, after all this, I found that this is actually a known issue with
XFAILing test case using the option:
test/Transforms/LoopSimplifyCFG/phi_with_duplicating_inputs.ll :-/

I wish that this would have been posted here on Bugzilla, or at least explained
as a comment around that option, or in the commit messages. There appears to
not even have been any review made, so I am a bit confused. Perhaps I just
missed this information somewhere?

I wonder now if my test cases are redundant with the XFAILing ones? Could the
committer perhaps make further comments on this, please?

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20181126/1e365f63/attachment.html>


More information about the llvm-bugs mailing list