Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 21 Mar 2019 19:33:28 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     Andy Lutomirski <luto@kernel.org>
Cc:     Steven Rostedt <rostedt@goodmis.org>,
        Juergen Gross <jgross@suse.com>,
        LKML <linux-kernel@vger.kernel.org>,
        "H. Peter Anvin" <hpa@zytor.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@kernel.org>, Borislav Petkov <bp@alien8.de>,
        Joel Fernandes <joel@joelfernandes.org>,
        He Zhe <zhe.he@windriver.com>,
        Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [RFC][PATCH] tracing/x86: Save CR2 before tracing irqsoff on
 error_entry
Message-ID: <20190321183327.GW5996@hirez.programming.kicks-ass.net>
References: <20190320221534.165ab87b@oasis.local.home>
 <20190321083317.GL6058@hirez.programming.kicks-ass.net>
 <20190321090241.GL6521@hirez.programming.kicks-ass.net>
 <20190321104517.GM6521@hirez.programming.kicks-ass.net>
 <20190321093242.4a948198@gandalf.local.home>
 <20190321172203.GS5996@hirez.programming.kicks-ass.net>
 <CALCETrXqmjJt7azNoEtxs2eBkDd-vjUAYvrjNh_RD30DAvksBw@mail.gmail.com>
 <20190321182745.GU5996@hirez.programming.kicks-ass.net>
 <CALCETrVd8i9xmGT9GPXU6tuvte+M9e3otYa8YNLEgPE+W81E2w@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CALCETrVd8i9xmGT9GPXU6tuvte+M9e3otYa8YNLEgPE+W81E2w@mail.gmail.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Mar 21, 2019 at 11:28:50AM -0700, Andy Lutomirski wrote:
> On Thu, Mar 21, 2019 at 11:27 AM Peter Zijlstra <peterz@infradead.org> wrote:
> >
> > On Thu, Mar 21, 2019 at 11:05:06AM -0700, Andy Lutomirski wrote:
> >
> > > Ugh.
> > >
> > > I certainly agree in principle that sticking the CR2 read into the asm
> > > is the right solution.  But this patch makes the spaghetti even more
> > > tangled.  Maybe we can rearrange the code a bit so that the entry
> > > sequence saves at least one register before calling error_entry, so we
> > > can do it the obvious way.
> >
> > What is the obvious way? Note that with all the Xen/paravirt-me-harder
> > bits on, we need to clobber at least RAX to get CR2.
> >
> > And most of the patch was munging the paravirt bits to make that happen.
> >
> > Another thing I did notices is that CLAC and CLD are not next to one
> > another in these paths.
> 
> The thing I really don't like is all the EBX crud to try to get
> everything in to the right order.

Ah yes, granted. I hate that part too but didn't see a way around it :/