On 07/22/14 00:45, Benjamin Marzinski wrote: > On Mon, Jul 21, 2014 at 05:17:32PM +0200, Sebastian Herbszt wrote: >> Hello Benjamin, >> >> your patch "enable gcc format-security check" (9b7741c) not only added >> -Wformat-security, but also changed fstack-protector to fstack-protector-strong. >> Was this change intentional? It is not mentioned in the commit log and seems >> to require a recent gcc: >> >> cc: error: unrecognized command line option ‘-fstack-protector-strong’ >> with >> gcc version 4.8.2 20140324 [gcc-4_8-branch revision 208789] (SUSE Linux) > > I do think switching from fstack-protector to fstack-protector-strong is > a good idea, and I use if for the redhat builds, but I didn't mean to > include that change in my commit. I can send a commit to back it out if > it's causing an annoyance. Once gcc 4.9 has been around for a while > longer, I'll resend a commit to include it. How about only using fstack-protector-strong if it is supported ? It could be interesting to have a look at the Linux kernel build system. A Makefile function called "cc-option" has been defined in scripts/Kbuild.include that is used to test whether or not the compiler supports a certain compiler flag. Bart. -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
