Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752342AbbKIIwa (ORCPT ); Mon, 9 Nov 2015 03:52:30 -0500 Received: from mail-wm0-f47.google.com ([74.125.82.47]:33705 "EHLO mail-wm0-f47.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751353AbbKIIw1 (ORCPT ); Mon, 9 Nov 2015 03:52:27 -0500 Date: Mon, 9 Nov 2015 09:52:23 +0100 From: Ingo Molnar To: Andy Lutomirski Cc: x86@kernel.org, linux-kernel@vger.kernel.org, Brian Gerst , Linus Torvalds , Borislav Petkov , =?iso-8859-1?Q?Fr=E9d=E9ric?= Weisbecker , Peter Zijlstra Subject: Re: [PATCH 4/4] x86/entry/64: Bypass enter_from_user_mode on non-context-tracking boots Message-ID: <20151109085223.GA12356@gmail.com> References: <32bb09004dc9b005d5c14f021e42eb1f83ce3db3.1446849780.git.luto@kernel.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <32bb09004dc9b005d5c14f021e42eb1f83ce3db3.1446849780.git.luto@kernel.org> User-Agent: Mutt/1.5.23 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2769 Lines: 90 * Andy Lutomirski wrote: > On CONFIG_CONTEXT_TRACKING kernels that have context tracking > disabled at runtime (which includes most distro kernels), we still > have the overhead of a call to enter_from_user_mode in interrupt and > exception entries. > > If jump labels are available, this uses the jump label > infrastructure to skip the call. > > Signed-off-by: Andy Lutomirski > --- > arch/x86/entry/calling.h | 15 +++++++++++++++ > arch/x86/entry/entry_64.S | 8 ++------ > 2 files changed, 17 insertions(+), 6 deletions(-) > > diff --git a/arch/x86/entry/calling.h b/arch/x86/entry/calling.h > index 3c71dd947c7b..271e30c585bc 100644 > --- a/arch/x86/entry/calling.h > +++ b/arch/x86/entry/calling.h > @@ -1,3 +1,5 @@ > +#include > + > /* > > x86 function call convention, 64-bit: > @@ -232,3 +234,16 @@ For 32-bit we have the following conventions - kernel is built with > > #endif /* CONFIG_X86_64 */ > > +/* > + * This does 'call enter_from_user_mode' unless we can avoid it based on > + * kernel config or using the static jump infrastructure. > + */ > +.macro CALL_ENTER_FROM_USER_MODE > +#ifdef CONFIG_CONTEXT_TRACKING > +#ifdef HAVE_JUMP_LABEL > + STATIC_JUMP_IF_FALSE .Lafter_call_\@, context_tracking_enabled, def=0 > +#endif > + call enter_from_user_mode > +.Lafter_call_\@: > +#endif > +.endm > diff --git a/arch/x86/entry/entry_64.S b/arch/x86/entry/entry_64.S > index f585df24ab3d..9b49b56efa29 100644 > --- a/arch/x86/entry/entry_64.S > +++ b/arch/x86/entry/entry_64.S > @@ -517,9 +517,7 @@ END(irq_entries_start) > */ > TRACE_IRQS_OFF > > -#ifdef CONFIG_CONTEXT_TRACKING > - call enter_from_user_mode > -#endif > + CALL_ENTER_FROM_USER_MODE > > 1: > /* > @@ -1058,9 +1056,7 @@ ENTRY(error_entry) > > .Lerror_entry_from_usermode_after_swapgs: > TRACE_IRQS_OFF > -#ifdef CONFIG_CONTEXT_TRACKING > - call enter_from_user_mode > -#endif > + CALL_ENTER_FROM_USER_MODE > ret Yeah, so I only have a really minor syntactic nit abou tthis: it's OK to capitalize things when doing macros, but here I don't think capitalizing the called function name is good - I think the following would be more obvious: CALL_enter_from_user_mode this makes it really, really obvious (to me!) that this macro is a shortcut for: call enter_from_user_mode with some glue around it, and I could grep for 'enter_from_user_mode' immediately - while grepping for ENTER_FROM_USER_MODE would miss the called function. Thanks, Ingo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/