bug-binutils
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Bug gold/16781] New: gold breaks alignment of global variable


From: daniel.oertwig+sourcewarebug at gmail dot com
Subject: [Bug gold/16781] New: gold breaks alignment of global variable
Date: Mon, 31 Mar 2014 10:21:02 +0000

https://sourceware.org/bugzilla/show_bug.cgi?id=16781

            Bug ID: 16781
           Summary: gold breaks alignment of global variable
           Product: binutils
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: gold
          Assignee: ian at airs dot com
          Reporter: daniel.oertwig+sourcewarebug at gmail dot com
                CC: ccoutant at google dot com

Created attachment 7506
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7506&action=edit
Map produced when linking with gold, broken variable alignment

Linking with gold breaks the alignment requirement I set on a global variable,
using the C++11 alignas specifier. Linking the same with the default ld works.

I suspect this only happens in combination with memory regions in the linker
script, but I could not verify this.

I am using clang++ to compile, and clang++ (which defaults to gold on my
system) to link. I attached the map file produced as bug.map, where you can see
that the variable bug does not start at a correctly aligned address.

Using g++ (which I guess is using the "old" linker) to link and still clang++
to compile works fine, as seen in the attached good.map.
Using g++ both to link and compile fails, too.
But this is not linker related, as the alignemnt requirement is already broken
in the *.o file.

I tried to strip down a minimal example, seen in goldbug.cpp. The linker script
used is in the attached goldbug.ld.

Let me know if I can help any further!

-- 
You are receiving this mail because:
You are on the CC list for the bug.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]