[llvm-bugs] [Bug 37119] New: Compiler miscompiling code after r329525 when optimizations are enabled

via llvm-bugs llvm-bugs at lists.llvm.org
Fri Apr 13 01:00:16 PDT 2018


https://bugs.llvm.org/show_bug.cgi?id=37119

            Bug ID: 37119
           Summary: Compiler miscompiling code after r329525 when
                    optimizations are enabled
           Product: new-bugs
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: douglas_yung at playstation.sony.com
                CC: llvm-bugs at lists.llvm.org

Created attachment 20186
  --> https://bugs.llvm.org/attachment.cgi?id=20186&action=edit
repro code

One of our internal tests started to fail after r329525 was committed. The test
was automagically generated, so it is a bit messy, but I have cut it down to a
somewhat smaller sample that I am able to reproduce the issue with on linux
independent of our test framework.

To reproduce the problem, build a compiler that is r329525 or newer on x64
linux, and the build/run the attached test with (-O2) and without (-O0)
optimizations and compare the output of the line for id11513.

Steps:
clang++ -O0 repro3.cpp -o repro3.good
clang++ -O2 repro3.cpp -o repro3.bad

Output of produced executables on my machine:

$ ./repro3.good
{BEGIN}
id11095:-33.831909 -33.831909 -33.831909 -33.831909
id11513:0 ffffffff 0 ffffffff 0 ffffffff ffffffff ffffffff ffffffff 0 0 0
ffffffff ffffffff 0 0
{END}
$ ./repro3.bad
{BEGIN}
id11095:-33.831909 -33.831909 -33.831909 -33.831909
id11513:0 0 ffffffff ffffffff 0 ffffffff 0 ffffffff 0 ffffffff 0 0 0 ffffffff 0
0
{END}

Note that in the optimized case (repro3.bad), in id11513, elements 2, 3, 7, 9,
10 and 13 differ from the non-optimized case (repro3.good).

-- 
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/20180413/ce0260c7/attachment.html>


More information about the llvm-bugs mailing list