From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754281Ab2HASqy (ORCPT ); Wed, 1 Aug 2012 14:46:54 -0400 Received: from www.linutronix.de ([62.245.132.108]:52696 "EHLO Galois.linutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753601Ab2HASqw (ORCPT ); Wed, 1 Aug 2012 14:46:52 -0400 Message-ID: <50197994.4090800@linutronix.de> Date: Wed, 01 Aug 2012 20:46:44 +0200 From: Sebastian Andrzej Siewior User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:10.0.5) Gecko/20120624 Icedove/10.0.5 MIME-Version: 1.0 To: Oleg Nesterov CC: Roland McGrath , "H. Peter Anvin" , linux-kernel@vger.kernel.org, ananth@in.ibm.com, a.p.zijlstra@chello.nl, mingo@redhat.com, srikar@linux.vnet.ibm.com Subject: Re: Q: user_enable_single_step() && update_debugctlmsr() References: <20120801130118.GA2386@redhat.com> <50192FF5.1060208@linutronix.de> <20120801134652.GA4707@redhat.com> <50193528.3070503@linutronix.de> <20120801140121.GA5333@redhat.com> <50193B5C.90404@linutronix.de> <20120801143132.GA7550@redhat.com> <5019427B.1070002@linutronix.de> <20120801150119.GA8845@redhat.com> <50194750.2050601@linutronix.de> <20120801151442.GA9697@redhat.com> In-Reply-To: <20120801151442.GA9697@redhat.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Linutronix-Spam-Score: -1.0 X-Linutronix-Spam-Level: - X-Linutronix-Spam-Status: No , -1.0 points, 5.0 required, ALL_TRUSTED=-1,SHORTCIRCUIT=-0.0001 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 08/01/2012 05:14 PM, Oleg Nesterov wrote: > On 08/01, Sebastian Andrzej Siewior wrote: >> >> On 08/01/2012 05:01 PM, Oleg Nesterov wrote: >>> On 08/01, Sebastian Andrzej Siewior wrote: >>>> So a patch like >>>> --- a/arch/x86/kernel/step.c >>>> +++ b/arch/x86/kernel/step.c >>>> @@ -173,8 +173,8 @@ static void enable_step(struct task_struct *child, >>>> bool block) >>>> unsigned long debugctl = get_debugctlmsr(); >>>> >>>> debugctl |= DEBUGCTLMSR_BTF; >>>> - update_debugctlmsr(debugctl); >>>> set_tsk_thread_flag(child, TIF_BLOCKSTEP); >>>> + update_debugctlmsr(debugctl); >>>> } else if (test_tsk_thread_flag(child, TIF_BLOCKSTEP)) { >>>> unsigned long debugctl = get_debugctlmsr(); >>>> >>>> should fix the race >>> >>> No, I don't think it can fix something ;) or make any difference. >> >> Why? You _first_ set the task flag > > Yes, and this task is "child". > >> followed by the CPU register. Now >> switch_to() would see the bit set and act. > > child sleeps and doesn't participate in switch_to(). Debugger and another > (unrelated) task do. This is confusing. In order to allow the debugger to ptrace()->enable_blockstep() the child has to be stopped/traced. We switch X86_EFLAGS_TF in child's regs and enable DEBUGCTLMSR_BTF for the debugger which is wrong. If we quit to userspace then the CPU on which the debugger runs has DEBUGCTLMSR_BTF. If the tracee task runs on the same then nothing happens, the bit remains set. If the tracee happens to run on a different CPU then switch_to() will enable the DEBUGCTLMSR_BTF bit for the debugger's CPU and switch_to() will enable it also on the other CPU. I added a few printks in the source and I see output that __switch_to_xtra() enables the bit as well as in enable_single() for debugger's CPU. I didn't find where the single step is disabled for the tracee. I haven't notice this in __switch_to_xtra() nor or in disable_single_step(). > Oleg. > Sebastian