Re: [powerpc] ftrace warning kernel/trace/ftrace.c:2068 with code-patching selftests

From: Kees Cook <keescook@chromium.org>
To: Mark Rutland <mark.rutland@arm.com>
Cc: Yinan Liu <yinan@linux.alibaba.com>,
	Steven Rostedt <rostedt@goodmis.org>,
	linuxppc-dev@lists.ozlabs.org,
	Sachin Sant <sachinp@linux.ibm.com>,
	linux-kernel@vger.kernel.org, ardb@kernel.org
Subject: Re: [powerpc] ftrace warning kernel/trace/ftrace.c:2068 with code-patching selftests
Date: Thu, 27 Jan 2022 08:41:38 -0800	[thread overview]
Message-ID: <202201270834.4420E21F6@keescook> (raw)
In-Reply-To: <YfKGKWW5UfZ15kCW@FVFF77S0Q05N>

On Thu, Jan 27, 2022 at 11:46:49AM +0000, Mark Rutland wrote:
> I'm not sure how x86 works here; AFAICT the relocations are performed during
> decompression, but it looks like there's some special build-time processing
> associated with that, and the vmlinux doesn't contain standard ELF relocations.
> 
> Kees, IIUC you added the x86_64 support there, can you shed any light on if/how
> this works on x86?

I think Sven beat me to it, and this was answered in
https://lore.kernel.org/lkml/yt9dy231gzae.fsf@linux.ibm.com
but let me know if anything needs further info.

An additional note is that x86 is built with "-2G addressing"
(-mcmodel=kernel). There was some work done to make it actually
PIE, which would allow the KASLR base to move further:
https://github.com/KSPP/linux/issues/38

-Kees

-- 
Kees Cook