[llvm-bugs] [Bug 35469] New: [InlineCost] CallAnalyzer crashes when analyzing bitcasted callee

via llvm-bugs llvm-bugs at lists.llvm.org
Wed Nov 29 12:24:19 PST 2017


https://bugs.llvm.org/show_bug.cgi?id=35469

            Bug ID: 35469
           Summary: [InlineCost] CallAnalyzer crashes when analyzing
                    bitcasted callee
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Interprocedural Analyses
          Assignee: unassignedbugs at nondot.org
          Reporter: kavon.farvardin at gmail.com
                CC: llvm-bugs at lists.llvm.org

Created attachment 19489
  --> https://bugs.llvm.org/attachment.cgi?id=19489&action=edit
backtrace of infinite recursion crash

The following IR crashes LLVM when running -inline -early-cse. It is not
specific to early-cse, since other passes following inline also crash it, such
as instcombine, gvn, etc.


define void @func1() {
  %t = bitcast void ()* @func2 to void ()*
  tail call void %t()
  ret void
}

define void @func2() {
  %t = bitcast void ()* @func2 to void ()*
  tail call void %t()
  ret void
}


It appears to be the bitcast that is throwing off CallAnalyzer, and sends it
into an infinite loop. My best guess is that somewhere it assumes it is an
indirect call, but somewhere else it sees that the callee is known, and so it
recurses forever trying to analyze @func2. I have attached a backtrace to aid
in debugging.

-- 
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/20171129/49875763/attachment.html>


More information about the llvm-bugs mailing list