From mboxrd@z Thu Jan 1 00:00:00 1970 From: will.deacon@arm.com (Will Deacon) Date: Wed, 26 Nov 2014 13:02:52 +0000 Subject: [PATCH v9 2/6] arm64: ptrace: allow tracer to skip a system call In-Reply-To: <1416977391-24231-3-git-send-email-takahiro.akashi@linaro.org> References: <1416977391-24231-1-git-send-email-takahiro.akashi@linaro.org> <1416977391-24231-3-git-send-email-takahiro.akashi@linaro.org> Message-ID: <20141126130252.GI14866@arm.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Nov 26, 2014 at 04:49:47AM +0000, AKASHI Takahiro wrote: > If tracer modifies a syscall number to -1, this traced system call should > be skipped with a return value specified in x0. > This patch implements this semantics. > > Please note: > * syscall entry tracing and syscall exit tracing (ftrace tracepoint and > audit) are always executed, if enabled, even when skipping a system call > (that is, -1). > In this way, we can avoid a potential bug where audit_syscall_entry() > might be called without audit_syscall_exit() at the previous system call > being called, that would cause OOPs in audit_syscall_entry(). > > Signed-off-by: AKASHI Takahiro > --- > arch/arm64/kernel/entry.S | 10 +++++++++- > 1 file changed, 9 insertions(+), 1 deletion(-) > > diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S > index 726b910..946ec52 100644 > --- a/arch/arm64/kernel/entry.S > +++ b/arch/arm64/kernel/entry.S > @@ -161,6 +161,7 @@ > */ > sc_nr .req x25 // number of system calls > scno .req x26 // syscall number > +scno_w .req w26 // syscall number (lower 32 bits) > stbl .req x27 // syscall table pointer > tsk .req x28 // current thread_info > > @@ -668,8 +669,14 @@ ENDPROC(el0_svc) > * switches, and waiting for our parent to respond. > */ > __sys_trace: > - mov x0, sp > + cmp scno_w, #-1 // set default errno for I hate that we have to use scno_w, but the only alternative I can think of is using w8 directly, which isn't any better and doesn't work for compat. Ho-hum, I guess we'll stick with what you have. > + b.ne 1f // user-issued syscall(-1) > + mov x0, #-ENOSYS > + str x0, [sp] Can you use #S_X0 here for clarity, please? > +1: mov x0, sp > bl syscall_trace_enter > + cmp w0, #-1 // skip the syscall? > + b.eq __sys_trace_return_skipped > adr lr, __sys_trace_return // return address > uxtw scno, w0 // syscall number (possibly new) > mov x1, sp // pointer to regs > @@ -684,6 +691,7 @@ __sys_trace: > > __sys_trace_return: > str x0, [sp] // save returned x0 and update this guy too. Will From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751056AbaKZNCw (ORCPT ); Wed, 26 Nov 2014 08:02:52 -0500 Received: from foss-mx-na.foss.arm.com ([217.140.108.86]:39721 "EHLO foss-mx-na.foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750816AbaKZNCv (ORCPT ); Wed, 26 Nov 2014 08:02:51 -0500 Date: Wed, 26 Nov 2014 13:02:52 +0000 From: Will Deacon To: AKASHI Takahiro Cc: "keescook@chromium.org" , Catalin Marinas , "dsaxena@linaro.org" , "arndb@arndb.de" , "linux-arm-kernel@lists.infradead.org" , "linaro-kernel@lists.linaro.org" , "linux-kernel@vger.kernel.org" Subject: Re: [PATCH v9 2/6] arm64: ptrace: allow tracer to skip a system call Message-ID: <20141126130252.GI14866@arm.com> References: <1416977391-24231-1-git-send-email-takahiro.akashi@linaro.org> <1416977391-24231-3-git-send-email-takahiro.akashi@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1416977391-24231-3-git-send-email-takahiro.akashi@linaro.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 On Wed, Nov 26, 2014 at 04:49:47AM +0000, AKASHI Takahiro wrote: > If tracer modifies a syscall number to -1, this traced system call should > be skipped with a return value specified in x0. > This patch implements this semantics. > > Please note: > * syscall entry tracing and syscall exit tracing (ftrace tracepoint and > audit) are always executed, if enabled, even when skipping a system call > (that is, -1). > In this way, we can avoid a potential bug where audit_syscall_entry() > might be called without audit_syscall_exit() at the previous system call > being called, that would cause OOPs in audit_syscall_entry(). > > Signed-off-by: AKASHI Takahiro > --- > arch/arm64/kernel/entry.S | 10 +++++++++- > 1 file changed, 9 insertions(+), 1 deletion(-) > > diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S > index 726b910..946ec52 100644 > --- a/arch/arm64/kernel/entry.S > +++ b/arch/arm64/kernel/entry.S > @@ -161,6 +161,7 @@ > */ > sc_nr .req x25 // number of system calls > scno .req x26 // syscall number > +scno_w .req w26 // syscall number (lower 32 bits) > stbl .req x27 // syscall table pointer > tsk .req x28 // current thread_info > > @@ -668,8 +669,14 @@ ENDPROC(el0_svc) > * switches, and waiting for our parent to respond. > */ > __sys_trace: > - mov x0, sp > + cmp scno_w, #-1 // set default errno for I hate that we have to use scno_w, but the only alternative I can think of is using w8 directly, which isn't any better and doesn't work for compat. Ho-hum, I guess we'll stick with what you have. > + b.ne 1f // user-issued syscall(-1) > + mov x0, #-ENOSYS > + str x0, [sp] Can you use #S_X0 here for clarity, please? > +1: mov x0, sp > bl syscall_trace_enter > + cmp w0, #-1 // skip the syscall? > + b.eq __sys_trace_return_skipped > adr lr, __sys_trace_return // return address > uxtw scno, w0 // syscall number (possibly new) > mov x1, sp // pointer to regs > @@ -684,6 +691,7 @@ __sys_trace: > > __sys_trace_return: > str x0, [sp] // save returned x0 and update this guy too. Will