[LLVMbugs] [Bug 21394] New: Miscompiled x86_64 floating-point code at -O2

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=21394

            Bug ID: 21394
           Summary: Miscompiled x86_64 floating-point code at -O2
           Product: new-bugs
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: cb at cebix.net
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Created attachment 13255
  --> http://llvm.org/bugs/attachment.cgi?id=13255&action=edit
Test case

Clang 3.5.0 appears to miscompile the attached program on x86_64 at -O2:

$ clang++ --version
clang version 3.5.0 (tags/RELEASE_350/final)
Target: x86_64-unknown-linux-gnu
Thread model: posix
$ clang++ -O2 -o test -lQtCore test.cpp && ./test
QPointF(1, -1) 
QPointF(0, -1) 
$ clang++ -O1 -o test -lQtCore test.cpp && ./test
QPointF(1, -1) 
QPointF(0, 0) 

The result with -O1 is the expected output.

The problem does not appear with Clang 3.4.1.

Tested on SUSE Linux Enterprise Server 11 SP3 (x86_64) with Qt 4.6.3 and a
manually built Clang 3.5.0 obtained from the LLVM Download Page.

-- 
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/20141027/5a3e93e8/attachment.html>