Re: [RFC PATCH 00/16] PTI support for x86-32

[Joerg Roedel]

Hi Linus,

On Tue, Jan 16, 2018 at 10:59:01AM -0800, Linus Torvalds wrote:
> Yes, I'm very happy to see that this is actually not nearly as bad as
> I feared it might be,

Yeah, I was looking at the original PTI patches and my impression was
that a lot of the complicated stuff (like setting up the cpu_entry_area)
was already in there for 32 bit too. So it was mostly about the entry
code and some changes to the 32bit page-table code.

> Some of those #ifdef's in the PTI code you added might want more
> commentary about what the exact differences are. And maybe they could
> be done more cleanly with some abstraction. But nothing looked
> _horrible_.

I'll add more comments and better abstraction, Dave has already
suggested some improvements here. Reading some of my comments again,
they need a rework anyway.

> .. and please run all the segment and syscall selfchecks that Andy has written.

Didn't know about them yet, thanks. I will run them too in my testing

> Xen PV and PTI don't work together even on x86-64 afaik, the Xen
> people apparently felt it wasn't worth it.  See the
> 
>         if (hypervisor_is_type(X86_HYPER_XEN_PV)) {
>                 pti_print_if_insecure("disabled on XEN PV.");
>                 return;
>         }
> 
> in pti_check_boottime_disable().

But I might have broken something for them anyway, honestly I didn't pay
much attention to the XEN_PV case as I was trying to get it running
here. My hope is that someone who knows Xen better than I do will help
out :)


Regards,

	Joerg