On Thu, Aug 29, 2019 at 12:55 PM Thomas Garnier [off-list ref] wrote:

On Tue, Aug 6, 2019 at 8:51 AM Peter Zijlstra [off-list ref] wrote:

quoted

On Tue, Aug 06, 2019 at 05:43:47PM +0200, Borislav Petkov wrote:

quoted

On Tue, Jul 30, 2019 at 12:12:44PM -0700, Thomas Garnier wrote:

quoted

These patches make some of the changes necessary to build the kernel as
Position Independent Executable (PIE) on x86_64. Another patchset will
add the PIE option and larger architecture changes.

Yeah, about this: do we have a longer writeup about the actual benefits
of all this and why we should take this all? After all, after looking
at the first couple of asm patches, it is posing restrictions to how
we deal with virtual addresses in asm (only RIP-relative addressing in
64-bit mode, MOVs with 64-bit immediates, etc, for example) and I'm
willing to bet money that some future unrelated change will break PIE
sooner or later.

The goal is being able to extend the range of addresses where the
kernel can be placed with KASLR. I will look at clarifying that in the
future.

quoted

Possibly objtool can help here; it should be possible to teach it about
these rules, and then it will yell when violated. That should avoid
regressions.

I will look into that as well.

Following a discussion with Kees. I will explore objtool in the
follow-up patchset as we still have more elaborate pie changes in the
second set. I like the idea overall and I think it would be great if
it works.