Messages in this thread | | | From | Dirk Gouders <> | Subject | Re: tools/lib/lk: redefinition of _FORTIFY_SOURCE (gcc-4.7.2) | Date | Thu, 23 May 2013 22:32:56 +0200 |
| |
Hi Borislav,
Borislav Petkov <bp@suse.de> writes:
> Hi, > > On Wed, May 22, 2013 at 12:00:21PM +0200, Dirk Gouders wrote: >> CFLAGS = -ggdb3 -Wall -Wextra -std=gnu99 -Werror -O6 -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2 $(EXTRA_WARNINGS) $(EXTRA_CFLAGS) -fPIC > > it seems someone else hit this already and fixed it too: > > commit d2f32479e5526a1ab3b4e43910fcb279871524ce > Author: Marcin Slusarz <marcin.slusarz@gmail.com> > Date: Sun Feb 17 16:03:36 2013 +0100 > > perf tools: check if -DFORTIFY_SOURCE=2 is allowed
I thought about this _FORTIFY_SOURCE test and how the above commit could/should be adopted to lib/lk/Makefile, and I thought that if it were true that recent versions of gcc define _FORTIFY_SOURCE by default, that test could probably be modified and just check gcc's builtin macros to find out if _FORTIFY_SOURCE has to be defined explicitely and I tried to find out when gcc started to use _FORTIFY_SOURCE builtin definitions...
In short: all what I said in my initial post was tested with gcc versions on gentoo machines and it is gentoo that patches gcc so that _FORTIFY_SOURCE becomes a builtin definition. Unfortunately I don't have access to machines running other distributions and can only report about gcc on gentoo, but even with this limited information I would say it depends on the distribution in use if -D_FORTIFY_SOURCE=2 is needed and not on the gcc version.
Sorry for the noise if you already noticed my fault, I felt I should correct my initial misleading information.
Dirk
| |