[LLVMbugs] [Bug 23097] New: Invalid code generation (regression) in pass-by-value on PowerPC

[bugzilla-daemon at llvm.org]

https://llvm.org/bugs/show_bug.cgi?id=23097

            Bug ID: 23097
           Summary: Invalid code generation (regression) in pass-by-value
                    on PowerPC
           Product: clang
           Version: 3.6
          Hardware: Other
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
          Assignee: unassignedclangbugs at nondot.org
          Reporter: andyg1001 at hotmail.co.uk
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Created attachment 14132
  --> https://llvm.org/bugs/attachment.cgi?id=14132&action=edit
Simplified test case

I've spotted a regression between clang 3.4.1 and 3.6.0 when compiling for
powerpc (compiling for x86 is ok).  A simplified test case is attached.  The
problem occurs on return from test() where the copies of A are destructed.

Compiling the test case with clang 3.4.1 produces this output:

1 [0xbf896c08::0x10012008] 1
2 [0xbf896c00::0x10012018] 1
1 [0xbf896bf0::0x10012008] 2
2 [0xbf896be8::0x10012018] 2
1 [0xbf896bf0::0x10012018] 3
2 [0xbf896be8::0x10012018] 3
1 [0xbf896c08::0x10012008] 1
2 [0xbf896c00::0x10012018] 1

Compiling with clang 3.6.0 produces this output:

1 [0xbf969b38::0x10012008] 1
2 [0xbf969b30::0x10012018] 1
1 [0xbf969b08::0x10012008] 2
2 [0xbf969b0c::0x10012018] 2
1 [0xbf969b08::0x10012018] 3
2 [0xbf969b0c::0x10012018] 3
1 [0xbf969b38::0x10012008] 0
2 [0xbf969b30::0x10012018] 2

Also attached are the outputs from the compiler in both llvm and ppc assembly
code, in case this is useful.

-- 
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/20150401/fd4e985c/attachment.html>