From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756934Ab2IXQBN (ORCPT ); Mon, 24 Sep 2012 12:01:13 -0400 Received: from mx1.redhat.com ([209.132.183.28]:44576 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756185Ab2IXQBL (ORCPT ); Mon, 24 Sep 2012 12:01:11 -0400 Date: Mon, 24 Sep 2012 18:02:50 +0200 From: Oleg Nesterov To: Ananth N Mavinakayanahalli Cc: Ingo Molnar , Peter Zijlstra , Srikar Dronamraju , Anton Arapov , Sebastian Andrzej Siewior , linux-kernel@vger.kernel.org Subject: Re: [PATCH 1/4] uprobes: Kill set_swbp()->is_swbp_at_addr() Message-ID: <20120924160250.GA15999@redhat.com> References: <20120923201921.GA27424@redhat.com> <20120923201945.GA27446@redhat.com> <20120924090858.GA2248@in.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20120924090858.GA2248@in.ibm.com> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 09/24, Ananth N Mavinakayanahalli wrote: > > On Sun, Sep 23, 2012 at 10:19:45PM +0200, Oleg Nesterov wrote: > > > It doesn't matter if this "int3" was in fact > > installed by gdb or application itself, we are going to "steal" this > > breakpoint anyway and execute the original insn from vm_file even if > > it no longer matches the memory. > > Wouldn't this text make more sense: > > It doesn't matter if this 'breakpoint' was in fact... > > 'int3' is still an x86 artifact. Agreed, will update the changelog. > On powerpc, we don't even get to install_breakpoint() ->set_swbp() > ->is_swbp_at_addr() because arch_uprobe_analyze_insn() would've already > caused install_breakpoint() to return -ENOTSUPP. > > > OTOH, handle_swbp()->find_active_uprobe() uses is_swbp_at_addr() to > > figure out whether we need to send SIGTRAP or not if we can not find > > uprobe, so in this case it should return true for all trap variants, > > not only for UPROBE_SWBP_INSN. > > So, we will need a powerpc specific is_swbp_insn()... ok. I think yes. But again, we need 2 helpers. One for is_swbp_at_addr(), and another for verify_opcode() which only checks UPROBE_SWBP_INSN. IOW, something like the patch below (on top of this series). Do you agree? Oleg. --- x/kernel/events/uprobes.c +++ x/kernel/events/uprobes.c @@ -178,6 +178,11 @@ static int __replace_page(struct vm_area * Default implementation of is_swbp_insn * Returns true if @insn is a breakpoint instruction. */ +static bool is_uprobe_opcode(uprobe_opcode_t *insn) +{ + return *insn == UPROBE_SWBP_INSN; +} + bool __weak is_swbp_insn(uprobe_opcode_t *insn) { return *insn == UPROBE_SWBP_INSN; @@ -196,9 +201,9 @@ static int verify_opcode(struct page *pa bool is_swbp; copy_opcode(page, vaddr, &old_opcode); - is_swbp = is_swbp_insn(&old_opcode); + is_swbp = is_uprobe_opcode(&old_opcode); - if (is_swbp_insn(new_opcode)) { + if (is_uprobe_opcode(new_opcode)) { if (is_swbp) /* register: already installed? */ return 0; } else { @@ -585,7 +590,7 @@ install_breakpoint(struct uprobe *uprobe if (ret) return ret; - if (is_swbp_insn((uprobe_opcode_t *)uprobe->arch.insn)) + if (is_uprobe_opcode((uprobe_opcode_t *)uprobe->arch.insn)) return -ENOTSUPP; ret = arch_uprobe_analyze_insn(&uprobe->arch, mm, vaddr);