[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <81f27dee-345d-17d6-7aff-1107390800ce@linux.intel.com>
Date: Tue, 9 Jan 2018 14:29:06 -0800
From: Dave Hansen <dave.hansen@...ux.intel.com>
To: Borislav Petkov <bp@...en8.de>, Willy Tarreau <w@....eu>
Cc: Andy Lutomirski <luto@...nel.org>,
LKML <linux-kernel@...r.kernel.org>, X86 ML <x86@...nel.org>,
Brian Gerst <brgerst@...il.com>,
Ingo Molnar <mingo@...nel.org>,
Linus Torvalds <torvalds@...ux-foundation.org>,
Peter Zijlstra <peterz@...radead.org>,
Thomas Gleixner <tglx@...utronix.de>,
Josh Poimboeuf <jpoimboe@...hat.com>,
"H. Peter Anvin" <hpa@...or.com>, Kees Cook <keescook@...omium.org>
Subject: Re: [RFC PATCH v2 2/6] x86/arch_prctl: add ARCH_GET_NOPTI and
ARCH_SET_NOPTI to enable/disable PTI
Has anyone else noticed that CR3 in oopses is now basically a lie with
PTI? It shows CR3 at the time of the printing of the oops, not at the
time of the fault like the normal registers. That greatly limits its
usefulness in the dumps with PTI.
Should we carve out some space somewhere to stash it at entry so we can
dump it in oopses? Any preferences on where? We need somewhere to
write that's before the SWITCH_TO_KERNEL_CR3, most likely so
cpu_entry_area is a candidate. The trampoline stack should also have space.
Powered by blists - more mailing lists