This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH: -save-temps still produces different .s and .o


On Jan 24, 2004, Per Bothner <per@bothner.com> wrote:

> Alexandre Oliva wrote:
>> FWIW, I can't fully test this: most (all?) PCH tests fail for me
>> (Fedora Core 1 i686 and athlon boxes with all updates installed) if I
>> bootstrap a pristine source tree, and they fail similarly after
>> earlier versions of our patches.  I still haven't tried this one.  Is
>> anyone else seeing this?

> The thread starting here provides a couple of suggestions
> for how to turn of Position Independent Executables:

> http://gcc.gnu.org/ml/gcc/2003-12/msg00747.html

FWIW, the issue is not about PIE.  We're not creating PIE.

The issue is about the randomized heap and stack addresses.  This can
be disabled with:

echo 0 > /proc/sys/kernel/exec-shield-randomize

i.e., you don't have to disable the randomization of PIEs in order to
get PCH to work.  You give up some of the security benefits of this in
order to use PCH, though, so it would be ideal to fix PCH such that it
can be used in such settings.  It would also make PCH usable on
Windows.

Anyhow, PCH shouldn't crash in case exec-shield-randomize is broken,
like it does now.  Any chance someone more familiar with the PCH
implementation could make it more robust in this regard?

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Happy GNU Year!                     oliva@{lsd.ic.unicamp.br, gnu.org}
Red Hat GCC Developer                 aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist                Professional serial bug killer


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]