[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20160518125659.a608972488e2c33f9ec3ed7b@gmail.com>
Date: Wed, 18 May 2016 12:56:59 +0200
From: Emese Revfy <re.emese@...il.com>
To: Michal Marek <mmarek@...e.com>
Cc: linux-kbuild@...r.kernel.org, pageexec@...email.hu,
spender@...ecurity.net, kernel-hardening@...ts.openwall.com,
keescook@...omium.org, linux@...musvillemoes.dk,
fengguang.wu@...el.com, dvyukov@...gle.com,
linux-kernel@...r.kernel.org, david.brown@...aro.org,
yamada.masahiro@...ionext.com
Subject: Re: [PATCH v8 2/4] GCC plugin infrastructure
On Tue, 17 May 2016 16:28:05 +0200
Michal Marek <mmarek@...e.com> wrote:
> Dne 13.5.2016 v 01:57 Emese Revfy napsal(a):
> > --- /dev/null
> > +++ b/scripts/Makefile.gcc-plugins
> > @@ -0,0 +1,21 @@
> > +ifdef CONFIG_GCC_PLUGINS
> > + __PLUGINCC := $(call cc-ifversion, -ge, 0408, $(HOSTCXX), $(HOSTCC))
> > + PLUGINCC := $(shell $(CONFIG_SHELL) $(srctree)/scripts/gcc-plugin.sh "$(__PLUGINCC)" "$(HOSTCXX)" "$(CC)")
> > +
> > + GCC_PLUGINS_CFLAGS := $(addprefix -fplugin=$(objtree)/scripts/gcc-plugins/, $(gcc-plugin-y))
> > +
> > + ifeq ($(PLUGINCC),)
> > + ifneq ($(GCC_PLUGINS_CFLAGS),)
> > + ifeq ($(call cc-ifversion, -ge, 0405, y), y)
> > + PLUGINCC := $(shell $(CO, NFIG_SHELL) -x $(srctree)/scripts/gcc-plugin.sh "$(__PLUGINCC)" "$(HOSTCXX)" "$(CC)")
> > + $(error error, your gcc installation does not support plugins, perhaps the necessary headers are missing?)
> > + else
> > + $(error error, your gcc version does not support plugins, you should upgrade it to gcc 4.5 at least)
> > + endif
> > + endif
> > + endif
> > +
> > + KBUILD_CFLAGS += $(GCC_PLUGINS_CFLAGS)
> > + GCC_PLUGIN := $(gcc-plugin-y)
> > +
> > +endif
>
> Hi Emese,
>
> sorry for the late feedback. A problem I ran into now is that if you
> have a compiler that is either too old or built with plugin support
> disabled, scripts/Makefile.gcc-plugins errors out immediately. This is
> going to be an issue with allmodconfig/allyesconfig test builds. A
> solution would either be to make the respective options depend on
> !COMPILE_TEST, or turn the errors into warnings and do nothing if the
> compiler lacks plugin support. We are handling CONFIG_CC_STACKPROTECTOR
> on x86 this way already.
Hi,
I'll change the errors to warnings in the next patch set.
--
Emese
Powered by blists - more mailing lists