Re: Porting issue from 32-bit to 64 bit



Thanks guys.

We verified the compiler time flags. It seems OK. We created a
separate program and complie with the same flags then the behavior is
as expected i.e., Proper memory is allocated (shown outside the
class) and same is shown for this pointer(inside the class).
We have used "ldd" to verify the libraries which seems OK to us.

Can you please suggest any other way to verify the wrong libraries
being used? Is there any flag which must be used for 64 bit
compilation?

This is the GCC version, we are using.

$ g++ -v
Using built-in specs.
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--enable-checking=release --with-system-zlib --enable-__cxa_atexit
--disable-libunwind-exceptions --enable-libgcj-multifile
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada
--enable-java-awt=gtk --disable-dssi --disable-plugin
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre
--with-cpu=generic --host=x86_64-redhat-linux
Thread model: posix
gcc version 4.1.2 20080704 (Red Hat 4.1.2-51)



On 1/5/12, Colin Law <clanlaw@xxxxxxxxxxxxxx> wrote:
On 5 January 2012 11:00, Shashank Ghoti <shashank.ghoti@xxxxxxxxx> wrote:
Hi All,

We are porting the application from 32-Bit to 64 Bit and OS is also 64-Bit
.
All the processes are highly generating core dump because of align problem
when we go to 64 bit.

We have run processes using GDB and found the below problem.

Let's say the class
class demo
{
private:
int a;
char *ptr;
int b;
};

GDB shows the object size created is 16 bytes ( 16 bytes memory creation
in
HEAP)while we go inside the class constructor, the size is shown of the
"This pointer" is 24 for this kind of class defined. We think that the
problem is due to alignment.

The problem has gone away if we re-define the class as:
class demo
{
private:
char *ptr;
int a;
int b;
};

But we have many processes and there are lot of files. So it would be
taking
a lot of time to manually fix it.
Are we missing any compiler flag? OR how to fix it? We have been trying
find out since last two weeks? We don't know.

Are you sure that you have compiled everything with the same flags? I
have seen that sort of problem when different files have been compiled
with different options. Another possibility is that the wrong
versions of libraries are being used.

Are you able to replicate the problem you describe above with a very
simple application compiled in the same way.

Colin

--
ubuntu-users mailing list
ubuntu-users@xxxxxxxxxxxxxxxx
Modify settings or unsubscribe at:
https://lists.ubuntu.com/mailman/listinfo/ubuntu-users


--
ubuntu-users mailing list
ubuntu-users@xxxxxxxxxxxxxxxx
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users



Relevant Pages

  • Re: Old question?
    ... int f ... A statement or declaration is terminated by a ";". ... the way the compiler handles the C++ comments depends on the ... flags used during the compilation. ...
    (comp.lang.c)
  • Re: -mm merge plans for 2.6.21
    ... anyway) then the compiler should be able to not allocate store for these. ... That they're const might help too. ... int flags = SA_INTERRUPT; ...
    (Linux-Kernel)
  • Re: Old question?
    ... int f ... A statement or declaration is terminated by a ";". ... the way the compiler handles the C++ comments depends on the ... flags used during the compilation. ...
    (comp.lang.c)
  • Re: amd64 bitops fix for -Os
    ... > kernel build. ... because its inline asm assumes at least one iteration ... the generated code is such that the flags ... although in a perfect world the compiler would be ...
    (Linux-Kernel)
  • Re: [PATCH 3/3] Enforce "unsigned long flags;" when spinlocking
    ... something different from "unsigned long flags;". ... This is a pretty ugly-looking patch. ... I want build system, compiler, headers etc to do everything against ... sparse spits warning ...
    (Linux-Kernel)