From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Cooper Subject: Re: [PATCH v6 05/10] Clear AC bit in RFLAGS to protect Xen itself by SMAP Date: Thu, 08 May 2014 02:57:55 +0100 Message-ID: <536AE4A3.6020905@citrix.com> References: <1399450782-14735-1-git-send-email-feng.wu@intel.com> <1399450782-14735-6-git-send-email-feng.wu@intel.com> <536A009B.20802@citrix.com> <536A37DD020000780000FE3D@mail.emea.novell.com> <536A1EC4.40304@citrix.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: "Wu, Feng" , Jan Beulich Cc: "Tian, Kevin" , "Dong, Eddie" , "ian.campbell@citrix.com" , "Nakajima, Jun" , "xen-devel@lists.xen.org" List-Id: xen-devel@lists.xenproject.org On 08/05/2014 02:41, Wu, Feng wrote: > >> -----Original Message----- >> From: Andrew Cooper [mailto:andrew.cooper3@citrix.com] >> Sent: Wednesday, May 07, 2014 7:54 PM >> To: Jan Beulich >> Cc: Wu, Feng; ian.campbell@citrix.com; Dong, Eddie; Nakajima, Jun; Tian, >> Kevin; xen-devel@lists.xen.org >> Subject: Re: [PATCH v6 05/10] Clear AC bit in RFLAGS to protect Xen itself by >> SMAP >> >> On 07/05/14 12:40, Jan Beulich wrote: >>>>>> On 07.05.14 at 11:44, wrote: >>>> On 07/05/14 09:19, Feng Wu wrote: >>>>> @@ -673,6 +675,7 @@ ENTRY(nmi_crash) >>>>> ud2 >>>>> >>>>> ENTRY(machine_check) >>>>> + ASM_CLAC >>>> This is not needed. the start of handle_ist_exception has a SAVE_ALL, >>>> which also covers the nmi entry point. >>>> >>>> On the subject of IST exceptions, perhaps the double fault explicitly >>>> wants a STAC to reduce the likelihood of taking a further fault while >>>> trying to dump state. ? >>> I agree. And perhaps along with do_double_fault(), fatal_trap() >>> should then also get a stac() added? >>> >>> Jan >>> >> With doubt_fault: being sole caller of do_double_fault(), editing the >> entry point in entry.S to "ASM_STAC; SAVE_ALL 0" is sufficient to avoid >> stac() in do_doube_fault() itself. > I think it's better to add "ASM_STAC" just before " call do_double_fault". > Do you think this is okay, Andrew? Thanks! I am not fussed where exactly the STAC goes in the entry point, but don't leave a CLAC in the SAVE_ALL. ~Andrew