Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S938016AbdD0BaQ (ORCPT ); Wed, 26 Apr 2017 21:30:16 -0400 Received: from mga01.intel.com ([192.55.52.88]:40519 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750786AbdD0BaI (ORCPT ); Wed, 26 Apr 2017 21:30:08 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.37,256,1488873600"; d="scan'208";a="961559659" Message-ID: <1493256599.36058.83.camel@ranerica-desktop> Subject: Re: [v6 PATCH 10/21] x86/insn-eval: Do not use R/EBP as base if mod in ModRM is zero From: Ricardo Neri To: Borislav Petkov Cc: Ingo Molnar , Thomas Gleixner , "H. Peter Anvin" , Andy Lutomirski , Peter Zijlstra , Andrew Morton , Brian Gerst , Chris Metcalf , Dave Hansen , Paolo Bonzini , Masami Hiramatsu , Huang Rui , Jiri Slaby , Jonathan Corbet , "Michael S. Tsirkin" , Paul Gortmaker , Vlastimil Babka , Chen Yucong , Alexandre Julliard , Stas Sergeev , Fenghua Yu , "Ravi V. Shankar" , Shuah Khan , linux-kernel@vger.kernel.org, x86@kernel.org, linux-msdos@vger.kernel.org, wine-devel@winehq.org, Adam Buchbinder , Colin Ian King , Lorenzo Stoakes , Qiaowei Ren , Arnaldo Carvalho de Melo , Adrian Hunter , Kees Cook , Thomas Garnier , Dmitry Vyukov Date: Wed, 26 Apr 2017 18:29:59 -0700 In-Reply-To: <20170421105258.q4vqrephtm7t3h4c@pd.tnic> References: <20170308003254.27833-1-ricardo.neri-calderon@linux.intel.com> <20170308003254.27833-11-ricardo.neri-calderon@linux.intel.com> <20170421105258.q4vqrephtm7t3h4c@pd.tnic> Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.10.4-0ubuntu2 Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5248 Lines: 149 On Fri, 2017-04-21 at 12:52 +0200, Borislav Petkov wrote: > On Tue, Mar 07, 2017 at 04:32:43PM -0800, Ricardo Neri wrote: > > Section 2.2.1.3 of the Intel 64 and IA-32 Architectures Software > > Developer's Manual volume 2A states that when the mod part of the ModRM > > byte is zero and R/EBP is specified in the R/M part of such bit, the value > > of the aforementioned register should not be used in the address > > computation. Instead, a 32-bit displacement is expected. The instruction > > decoder takes care of setting the displacement to the expected value. > > Returning -EDOM signals callers that they should ignore the value of such > > register when computing the address encoded in the instruction operands. > > > > Also, callers should exercise care to correctly interpret this particular > > case. In IA-32e 64-bit mode, the address is given by the displacement plus > > the value of the RIP. In IA-32e compatibility mode, the value of EIP is > > ignored. This correction is done for our insn_get_addr_ref. > > > > Cc: Dave Hansen > > Cc: Adam Buchbinder > > Cc: Colin Ian King > > Cc: Lorenzo Stoakes > > Cc: Qiaowei Ren > > Cc: Arnaldo Carvalho de Melo > > Cc: Masami Hiramatsu > > Cc: Adrian Hunter > > Cc: Kees Cook > > Cc: Thomas Garnier > > Cc: Peter Zijlstra > > Cc: Borislav Petkov > > Cc: Dmitry Vyukov > > Cc: Ravi V. Shankar > > Cc: x86@kernel.org > > Signed-off-by: Ricardo Neri > > --- > > arch/x86/lib/insn-eval.c | 25 +++++++++++++++++++++++-- > > 1 file changed, 23 insertions(+), 2 deletions(-) > > > > diff --git a/arch/x86/lib/insn-eval.c b/arch/x86/lib/insn-eval.c > > index cda6c71..ea10b03 100644 > > --- a/arch/x86/lib/insn-eval.c > > +++ b/arch/x86/lib/insn-eval.c > > @@ -250,6 +250,14 @@ static int get_reg_offset(struct insn *insn, struct pt_regs *regs, > > switch (type) { > > case REG_TYPE_RM: > > regno = X86_MODRM_RM(insn->modrm.value); > > + /* if mod=0, register R/EBP is not used in the address > > + * computation. Instead, a 32-bit displacement is expected; > > + * the instruction decoder takes care of reading such > > + * displacement. This is true for both R/EBP and R13, as the > > + * REX.B bit is not decoded. > > + */ > > I'd simply write here: "ModRM.mod == 0 and ModRM.rm == 5 means a 32-bit > displacement is following." I will shorten the comment. > > In addition, kernel comments style is: > > /* > * A sentence ending with a full-stop. > * Another sentence. ... > * More sentences. ... > */ ... and use the correct style. I feel bad I missed this one. > > > + if (regno == 5 && X86_MODRM_MOD(insn->modrm.value) == 0) > > + return -EDOM; > > if (X86_MODRM_MOD(insn->modrm.value) == 0 && > X86_MODRM_RM(insn->modrm.value) == 5) > > looks more understandable to me. Should I go with !(X86_MODRM_MOD(insn->modrm.value)) as you suggested in other patches? > > > if (X86_REX_B(insn->rex_prefix.value)) > > regno += 8; > > break; > > @@ -599,9 +607,22 @@ void __user *insn_get_addr_ref(struct insn *insn, struct pt_regs *regs) > > eff_addr = base + indx * (1 << X86_SIB_SCALE(sib)); > > } else { > > addr_offset = get_reg_offset(insn, regs, REG_TYPE_RM); > > - if (addr_offset < 0) > > + /* -EDOM means that we must ignore the address_offset. > > + * The only case in which we see this value is when > > + * R/M points to R/EBP. In such a case, in 64-bit mode > > + * the effective address is relative to tho RIP. > > s/tho// Will correct. > > > + */ > > Kernel comments style is: > > /* > * A sentence ending with a full-stop. > * Another sentence. ... > * More sentences. ... > */ > Will correct. > > + if (addr_offset == -EDOM) { > > + eff_addr = 0; > > +#ifdef CONFIG_X86_64 > > + if (user_64bit_mode(regs)) > > + eff_addr = (long)regs->ip; > > Is regs->ip the rIP of the *following* insn? No this is a bug. This should be regs->ip + insn.length. > > > +#endif > > You can do this in a prepatch and then get rid of the ifdeffery here: > > diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h > index 2b5d686ea9f3..f6239273c5f1 100644 > --- a/arch/x86/include/asm/ptrace.h > +++ b/arch/x86/include/asm/ptrace.h > @@ -115,9 +115,9 @@ static inline int v8086_mode(struct pt_regs *regs) > #endif > } > > -#ifdef CONFIG_X86_64 > static inline bool user_64bit_mode(struct pt_regs *regs) > { > +#ifdef CONFIG_X86_64 > #ifndef CONFIG_PARAVIRT > /* > * On non-paravirt systems, this is the only long mode CPL 3 > @@ -128,6 +128,9 @@ static inline bool user_64bit_mode(struct pt_regs *regs) > /* Headers are too twisted for this to go in paravirt.h. */ > return regs->cs == __USER_CS || regs->cs == pv_info.extra_user_64bit_cs; > #endif > +#else /* !CONFIG_X86_64 */ > + return false; > +#endif > } This look nice. I will add this pre-patch. Thanks and BR, Ricardo