[LLVMbugs] [Bug 23763] New: Floating Point Exception with -loop-vectorize and 0-sized type

[bugzilla-daemon at llvm.org]

https://llvm.org/bugs/show_bug.cgi?id=23763

            Bug ID: 23763
           Summary: Floating Point Exception with -loop-vectorize and
                    0-sized type
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Loop Optimizer
          Assignee: unassignedbugs at nondot.org
          Reporter: alex at crichton.co
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Created attachment 14438
  --> https://llvm.org/bugs/attachment.cgi?id=14438&action=edit
bugpoint-reduced IR

When optimized with the `-loop-vectorize` pass via `opt`, the attached IR will
trigger a floating point exception (divide-by-zero) in the optimization passes.
The stack trace on failure is:


$ ./Debug+Asserts/bin/opt bugpoint-reduced-simplified.ll -loop-vectorize -S
0  opt             0x00000000024c281e
llvm::sys::PrintStackTrace(llvm::raw_ostream&) + 46
1  opt             0x00000000024c3859
2  opt             0x00000000024c648d
3  libpthread.so.0 0x00007fc4b82a7340
4  opt             0x0000000001e4d8b8 llvm::isInductionPHI(llvm::PHINode*,
llvm::ScalarEvolution*, llvm::ConstantInt*&) + 488
5  opt             0x0000000001a619c9
6  opt             0x0000000001a5fea0
7  opt             0x0000000001a400a9
8  opt             0x0000000001a3eacc
9  opt             0x0000000001a3e076
10 opt             0x00000000023e9c5b
llvm::FPPassManager::runOnFunction(llvm::Function&) + 427
11 opt             0x00000000023e9f68
llvm::FPPassManager::runOnModule(llvm::Module&) + 104
12 opt             0x00000000023ea64a
13 opt             0x00000000023ea21e
llvm::legacy::PassManagerImpl::run(llvm::Module&) + 302
14 opt             0x00000000023eac11
llvm::legacy::PassManager::run(llvm::Module&) + 33
15 opt             0x000000000084322e main + 7486
16 libc.so.6       0x00007fc4b72a6ec5 __libc_start_main + 245
17 opt             0x000000000081c474
Stack dump:
0.      Program arguments: ./Debug+Asserts/bin/opt
/home/alex/code/rust4/bugpoint-reduced-simplified.ll -loop-vectorize -S 
1.      Running pass 'Function Pass Manager' on module
'/home/alex/code/rust4/bugpoint-reduced-simplified.ll'.
2.      Running pass 'Loop Vectorization' on function '@foo'
zsh: floating point exception (core dumped)  ./Debug+Asserts/bin/opt
~/code/rust4/bugpoint-reduced-simplified.ll  -S

-- 
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/20150605/79d56a78/attachment.html>