Re: [PATCH 3/6] x86/entry/64/compat: Fix Xen PV SYSENTER frame setup

From: Andy Lutomirski <luto@kernel.org>
To: Brian Gerst <brgerst@gmail.com>
Cc: Andy Lutomirski <luto@kernel.org>,
	"the arch/x86 maintainers" <x86@kernel.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Andrew Cooper <andrew.cooper3@citrix.com>,
	Juergen Gross <jgross@suse.com>,
	Boris Ostrovsky <boris.ostrovsky@oracle.com>,
	Stefano Stabellini <sstabellini@kernel.org>,
	xen-devel <xen-devel@lists.xenproject.org>
Subject: Re: [PATCH 3/6] x86/entry/64/compat: Fix Xen PV SYSENTER frame setup
Date: Wed, 1 Jul 2020 11:39:40 -0700	[thread overview]
Message-ID: <CALCETrVy-Q4K04wmEPe5VeU=at2BL4b-bSFkoSU-BPbTaTB2Yg@mail.gmail.com> (raw)
In-Reply-To: <CAMzpN2iW4XD1Gsgq0ZeeH2eewLO+9Mk6eyk0LnbF-kP3v=smLg@mail.gmail.com>

On Wed, Jul 1, 2020 at 8:42 AM Brian Gerst <brgerst@gmail.com> wrote:
>
> On Fri, Jun 26, 2020 at 1:30 PM Andy Lutomirski <luto@kernel.org> wrote:
> >
> > The SYSENTER frame setup was nonsense.  It worked by accident
> > because the normal code into which the Xen asm jumped
> > (entry_SYSENTER_32/compat) threw away SP without touching the stack.
> > entry_SYSENTER_compat was recently modified such that it relied on
> > having a valid stack pointer, so now the Xen asm needs to invoke it
> > with a valid stack.
> >
> > Fix it up like SYSCALL: use the Xen-provided frame and skip the bare
> > metal prologue.
> >
> > Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
> > Cc: Juergen Gross <jgross@suse.com>
> > Cc: Stefano Stabellini <sstabellini@kernel.org>
> > Cc: xen-devel@lists.xenproject.org
> > Fixes: 1c3e5d3f60e2 ("x86/entry: Make entry_64_compat.S objtool clean")
> > Signed-off-by: Andy Lutomirski <luto@kernel.org>
> > ---
> >  arch/x86/entry/entry_64_compat.S |  1 +
> >  arch/x86/xen/xen-asm_64.S        | 20 ++++++++++++++++----
> >  2 files changed, 17 insertions(+), 4 deletions(-)
> >
> > diff --git a/arch/x86/entry/entry_64_compat.S b/arch/x86/entry/entry_64_compat.S
> > index 7b9d8150f652..381a6de7de9c 100644
> > --- a/arch/x86/entry/entry_64_compat.S
> > +++ b/arch/x86/entry/entry_64_compat.S
> > @@ -79,6 +79,7 @@ SYM_CODE_START(entry_SYSENTER_compat)
> >         pushfq                          /* pt_regs->flags (except IF = 0) */
> >         pushq   $__USER32_CS            /* pt_regs->cs */
> >         pushq   $0                      /* pt_regs->ip = 0 (placeholder) */
> > +SYM_INNER_LABEL(entry_SYSENTER_compat_after_hwframe, SYM_L_GLOBAL)
>
> This skips over the section that truncates the syscall number to
> 32-bits.  The comments present some doubt that it is actually
> necessary, but the Xen path shouldn't differ from native.  That code
> should be moved after this new label.

Whoops.  I thought I caught that myself, but apparently not.  I'll fix it.

>
> --
> Brian Gerst