linux-kernel - Re: [PATCH 3/3] powerpc: enable support for GCC plugins

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <f32f62b8-dc38-6f0b-e7b7-1f85e504df5d@au1.ibm.com>
Date:   Wed, 7 Dec 2016 16:49:41 +1100
From:   Andrew Donnellan <andrew.donnellan@....ibm.com>
To:     Emese Revfy <re.emese@...il.com>
Cc:     spender@...ecurity.net, keescook@...omium.org,
        linux-kbuild@...r.kernel.org, kernel-hardening@...ts.openwall.com,
        linux-kernel@...r.kernel.org, mmarek@...e.com,
        pageexec@...email.hu, linuxppc-dev@...ts.ozlabs.org
Subject: Re: [PATCH 3/3] powerpc: enable support for GCC plugins

On 07/12/16 08:25, Emese Revfy wrote:
> What are these missing headers? Because if they aren't necessary then they can
> be removed from gcc-common.h. There were missing headers on arm/arm64 and these
> archs are supported. I think this version check is unnecessary because
> gcc-plugin.sh also checks the missing headers.

rs6000-cpus.def, included via tm.h - see 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66840

I realise gcc-plugin.sh does detect this, but the point of the 
additional version check is to provide somewhat more helpful advice to 
the user.

> What is the problem on gcc-4.5/gcc-4.6?

On 4.6.4, c-family/c-common.h:

/scratch/ajd/gcc-test-v2/kernel/scripts/gcc-plugins/gcc-common.h:60:31: 
fatal error: c-family/c-common.h: No such file or directory

ajd@ka1:/scratch/ajd/tmp/cross/gcc-4.6.4-nolibc/powerpc64-linux$ find 
-name c-common.*
./lib/gcc/powerpc64-linux/4.6.4/plugin/include/c-common.h
./lib/gcc/powerpc64-linux/4.6.4/plugin/include/c-family/c-common.def

Are we sure the version check in gcc-common.h:59 is correct, or is this 
just a peculiarity of my particular toolchain?

I need to build another 4.5 toolchain, I'll try to do that this week.

-- 
Andrew Donnellan              OzLabs, ADL Canberra
andrew.donnellan@....ibm.com  IBM Australia Limited