From mboxrd@z Thu Jan 1 00:00:00 1970 From: Yu-cheng Yu Subject: Re: [PATCH 03/10] x86/cet: Signal handling for shadow stack Date: Thu, 07 Jun 2018 13:12:30 -0700 Message-ID: <1528402350.5265.21.camel@2b52.sc.intel.com> References: <20180607143807.3611-1-yu-cheng.yu@intel.com> <20180607143807.3611-4-yu-cheng.yu@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Andy Lutomirski Cc: Florian Weimer , Dmitry Safonov , Cyrill Gorcunov , LKML , linux-doc@vger.kernel.org, Linux-MM , linux-arch , X86 ML , "H. Peter Anvin" , Thomas Gleixner , Ingo Molnar , "H. J. Lu" , "Shanbhogue, Vedvyas" , "Ravi V. Shankar" , Dave Hansen , Jonathan Corbet , Oleg Nesterov , Arnd Bergmann , mike.kravetz@oracle.com List-Id: linux-arch.vger.kernel.org On Thu, 2018-06-07 at 11:30 -0700, Andy Lutomirski wrote: > On Thu, Jun 7, 2018 at 7:41 AM Yu-cheng Yu wrote: > > > > Set and restore shadow stack pointer for signals. > > How does this interact with siglongjmp()? > > This patch makes me extremely nervous due to the possibility of ABI > issues and CRIU breakage. Longjmp/Siglongjmp is handled in GLIBC and basically the shadow stack pointer is unwound. There could be some unexpected conditions. However, we run all GLIBC tests. > > > diff --git a/arch/x86/include/uapi/asm/sigcontext.h b/arch/x86/include/uapi/asm/sigcontext.h > > index 844d60eb1882..6c8997a0156a 100644 > > --- a/arch/x86/include/uapi/asm/sigcontext.h > > +++ b/arch/x86/include/uapi/asm/sigcontext.h > > @@ -230,6 +230,7 @@ struct sigcontext_32 { > > __u32 fpstate; /* Zero when no FPU/extended context */ > > __u32 oldmask; > > __u32 cr2; > > + __u32 ssp; > > }; > > > > /* > > @@ -262,6 +263,7 @@ struct sigcontext_64 { > > __u64 trapno; > > __u64 oldmask; > > __u64 cr2; > > + __u64 ssp; > > > > /* > > * fpstate is really (struct _fpstate *) or (struct _xstate *) > > @@ -320,6 +322,7 @@ struct sigcontext { > > struct _fpstate __user *fpstate; > > __u32 oldmask; > > __u32 cr2; > > + __u32 ssp; > > Is it actually okay to modify these structures like this? They're > part of the user ABI, and I don't know whether any user code relies on > the size being constant. > > > +int cet_push_shstk(int ia32, unsigned long ssp, unsigned long val) > > +{ > > + if (val >= TASK_SIZE) > > + return -EINVAL; > > TASK_SIZE_MAX. But I'm a bit unsure why you need this check at all. If an invalid address is put on the shadow stack, the task will get a control protection fault. I will change it to TASK_SIZE_MAX. > > > +int cet_restore_signal(unsigned long ssp) > > +{ > > + if (!current->thread.cet.shstk_enabled) > > + return 0; > > + return cet_set_shstk_ptr(ssp); > > +} > > This will blow up if the shadow stack enabled state changes in a > signal handler. Maybe we don't care. Yes, the task will get a control protection fault. From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga02.intel.com ([134.134.136.20]:14891 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752602AbeFGUPn (ORCPT ); Thu, 7 Jun 2018 16:15:43 -0400 Message-ID: <1528402350.5265.21.camel@2b52.sc.intel.com> Subject: Re: [PATCH 03/10] x86/cet: Signal handling for shadow stack From: Yu-cheng Yu Date: Thu, 07 Jun 2018 13:12:30 -0700 In-Reply-To: References: <20180607143807.3611-1-yu-cheng.yu@intel.com> <20180607143807.3611-4-yu-cheng.yu@intel.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: linux-arch-owner@vger.kernel.org List-ID: To: Andy Lutomirski Cc: Florian Weimer , Dmitry Safonov , Cyrill Gorcunov , LKML , linux-doc@vger.kernel.org, Linux-MM , linux-arch , X86 ML , "H. Peter Anvin" , Thomas Gleixner , Ingo Molnar , "H. J. Lu" , "Shanbhogue, Vedvyas" , "Ravi V. Shankar" , Dave Hansen , Jonathan Corbet , Oleg Nesterov , Arnd Bergmann , mike.kravetz@oracle.com Message-ID: <20180607201230.ON8zvcfg6NwHRv6hpQSz6b0cIQJgdlWJNeQO2_WMTqA@z> On Thu, 2018-06-07 at 11:30 -0700, Andy Lutomirski wrote: > On Thu, Jun 7, 2018 at 7:41 AM Yu-cheng Yu wrote: > > > > Set and restore shadow stack pointer for signals. > > How does this interact with siglongjmp()? > > This patch makes me extremely nervous due to the possibility of ABI > issues and CRIU breakage. Longjmp/Siglongjmp is handled in GLIBC and basically the shadow stack pointer is unwound. There could be some unexpected conditions. However, we run all GLIBC tests. > > > diff --git a/arch/x86/include/uapi/asm/sigcontext.h b/arch/x86/include/uapi/asm/sigcontext.h > > index 844d60eb1882..6c8997a0156a 100644 > > --- a/arch/x86/include/uapi/asm/sigcontext.h > > +++ b/arch/x86/include/uapi/asm/sigcontext.h > > @@ -230,6 +230,7 @@ struct sigcontext_32 { > > __u32 fpstate; /* Zero when no FPU/extended context */ > > __u32 oldmask; > > __u32 cr2; > > + __u32 ssp; > > }; > > > > /* > > @@ -262,6 +263,7 @@ struct sigcontext_64 { > > __u64 trapno; > > __u64 oldmask; > > __u64 cr2; > > + __u64 ssp; > > > > /* > > * fpstate is really (struct _fpstate *) or (struct _xstate *) > > @@ -320,6 +322,7 @@ struct sigcontext { > > struct _fpstate __user *fpstate; > > __u32 oldmask; > > __u32 cr2; > > + __u32 ssp; > > Is it actually okay to modify these structures like this? They're > part of the user ABI, and I don't know whether any user code relies on > the size being constant. > > > +int cet_push_shstk(int ia32, unsigned long ssp, unsigned long val) > > +{ > > + if (val >= TASK_SIZE) > > + return -EINVAL; > > TASK_SIZE_MAX. But I'm a bit unsure why you need this check at all. If an invalid address is put on the shadow stack, the task will get a control protection fault. I will change it to TASK_SIZE_MAX. > > > +int cet_restore_signal(unsigned long ssp) > > +{ > > + if (!current->thread.cet.shstk_enabled) > > + return 0; > > + return cet_set_shstk_ptr(ssp); > > +} > > This will blow up if the shadow stack enabled state changes in a > signal handler. Maybe we don't care. Yes, the task will get a control protection fault.