Re: [RESEND PATCH v6 00/12] x86: Enable FSGSBASE instructions

From: Andy Lutomirski
Date: Mon Mar 25 2019 - 20:44:11 EST


On Fri, Mar 15, 2019 at 1:07 PM Chang S. Bae <chang.seok.bae@xxxxxxxxx> wrote:
>
> Updates from v5 [5]:
> * Drop the new tain flag (TAINT_INSECURE)
> * Cleanup copy_thread_tls(), some changelog, and unnecessary comments on
> assembly macros
> * Rearrange some helper updates appropriately (from patch 4 to 6)

I think this stuff is in generally decent shape, but I have two big
broad comments:

What's the status of the stuff hpa was working on to make the behavior
of modify_ldt() predictable?

Can we please have a test case that explicitly exercises the way that
ptrace reads and writes the base registers?