libunwind-devel
[Top][All Lists]
Advanced

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

Re: [Libunwind-devel] Power Libunwind


From: Jose Flavio Aguilar Paulino
Subject: Re: [Libunwind-devel] Power Libunwind
Date: Fri, 03 Aug 2007 03:56:25 -0300

Well, you need to set the CFLAGS, or the CC with -m64
do like this:
go to your build directory,
CC='gcc -m64 -maltivec' ../libunwind/configure && make -j2

--
Flavio

ps: I havent tested with cell yet, but hope that there will be no
problems in ppu

On Thu, 2007-08-02 at 23:28 -0700, address@hidden wrote:
> I did as you said and it tried to look for libunwind-ppc32.h! This is
> on a cell running Fedora Core 5 Linux ppc64
> 
> On 8/2/07, Jose Flavio Aguilar Paulino <address@hidden> wrote:
> > The problem is that the new configure.sh is not available updated in
> > Git.
> > To make it build easily just
> > delete the files:
> > Makefile.in
> > src/Makefile.in
> > tests/Makefile.in
> > after that run autoconf and automake in libuniwnd root dir.
> > then you can run configure... the correct arch to detect would be
> > "ppc64"!
> >
> > any doubts, just say, I would be glad to help,
> > Jose Flavio A. Paulino
> > IBM Ltc Brazil Toolchain Team
> >
> > On Thu, 2007-08-02 at 17:56 -0700, address@hidden wrote:
> > > I'm facing problems with my build because the configure detects the
> > > platform as powerpc64 instead of ppc64 and looks for files named
> > > *powerpc64* instead of *ppc64*
> > >
> > > Is there a way to configure this?
> > >
> > > Also, does this patch apply to 32-bit Power3 AIX too?
> > >
> > > On 8/2/07, Jose Flavio Aguilar Paulino <address@hidden> wrote:
> > > > Hi David!
> > > >
> > > > Its very good to have power main stream!
> > > > Now I will do more complex tests and maybe incremental fixes!
> > > >
> > > > The only last problem is that you didnt run autoconf and automake, or
> > > > didnt have removed Makefile.in, /src/Makefile.in and /tests/Makefile.in
> > > >
> > > > If you could do that and commit this files, I would be glad!
> > > >
> > > > Thanks,
> > > > Flavio
> > > >
> > > > On Thu, 2007-08-02 at 10:01 -0600, David Mosberger-Tang wrote:
> > > > > Jose,
> > > > >
> > > > > This patch looks good to me.  I merged it and pushed it into the git
> > > > > tree.  I only did a build on x86 so no real testing on my end, but I
> > > > > figured you were eager to see it in GIT so we can start doing
> > > > > incremental fixes. ;-)
> > > > >
> > > > > Documentation updates are definitely welcome, too!
> > > > >
> > > > > Please let me know if you see any issues.
> > > > >
> > > > > Thanks!
> > > > >
> > > > >   --david
> > > > >
> > > > > On 8/2/07, Jose Flavio Aguilar Paulino <address@hidden> wrote:
> > > > >         Well, if the symbol has the the address of the function
> > > > >         pointer, it is
> > > > >         not a problem, I removed all get_function_address!
> > > > >
> > > > >         Well, remember to do autoconf and automake.
> > > > >         And if possible, in libunwind page write this build
> > > > >         instructions:
> > > > >         CC='gcc -m64 -maltivec' ../libunwind/configure && make -j2
> > > > >
> > > > >         regards,
> > > > >         Flávio
> > > > >
> > > > >
> > > > >         On Thu, 2007-08-02 at 07:27 -0600, David Mosberger-Tang wrote:
> > > > >         > On 8/2/07, Jose Flavio Aguilar Paulino <address@hidden>
> > > > >         wrote:
> > > > >         >         because when the symbol is a function, and you are
> > > > >         looking for
> > > > >         >         it, we
> > > > >         >         have the following:
> > > > >         >         val = sym->st_value;
> > > > >         >
> > > > >         >         but, in Power (I don't know how it works in IA64),
> > > > >         this value
> > > > >         >         would be
> > > > >         >         the address of entry in the table, not the address
> > > > >         of the
> > > > >         >         symbol, so the
> > > > >         >         look up will fail.
> > > > >         >
> > > > >         > Why is that a problem?  On ia64, we also return the address
> > > > >         of the
> > > > >         > function descriptor.  That *is* the value of a function
> > > > >         pointer, after
> > > > >         > all.  I guess I'm still not following.
> > > > >         >
> > > > >         >
> > > > >         >
> > > > >         >   --david
> > > > >         > --
> > > > >         > Mosberger Consulting LLC,
> > > > >         http://www.mosberger-consulting.com/
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Mosberger Consulting LLC, http://www.mosberger-consulting.com/
> > > >
> > > >
> > > >
> > > > _______________________________________________
> > > > Libunwind-devel mailing list
> > > > address@hidden
> > > > http://lists.nongnu.org/mailman/listinfo/libunwind-devel
> > > >
> > >
> > >
> >
> >
> 
> 





reply via email to

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