From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755080Ab1IMIEv (ORCPT ); Tue, 13 Sep 2011 04:04:51 -0400 Received: from smarthost1.greenhost.nl ([195.190.28.78]:36475 "EHLO smarthost1.greenhost.nl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753596Ab1IMIEs (ORCPT ); Tue, 13 Sep 2011 04:04:48 -0400 Message-ID: <9b86bd67bedfcd3d4e592a419d190b90.squirrel@webmail.greenhost.nl> In-Reply-To: <201109101312.39256.pedro@codesourcery.com> References: <1315506333.18043.49.camel@dhcp-25-63.brq.redhat.com> <201109101219.30942.pedro@codesourcery.com> <201109101340.16506.vda.linux@googlemail.com> <201109101312.39256.pedro@codesourcery.com> Date: Tue, 13 Sep 2011 10:04:46 +0200 Subject: Re: [PATCH v3] Make PTRACE_SEIZE set ptrace options specified in 'data' From: "Indan Zupancic" To: "Pedro Alves" Cc: "Denys Vlasenko" , "Denys Vlasenko" , "Oleg Nesterov" , "Tejun Heo" , linux-kernel@vger.kernel.org User-Agent: SquirrelMail/1.4.22 MIME-Version: 1.0 Content-Type: text/plain;charset=UTF-8 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal X-Spam-Score: 0.0 X-Scan-Signature: 4052b6d1c9976086d5ab5ce040fcf5b8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hello, On Sat, September 10, 2011 14:12, Pedro Alves wrote: > > I happen to know gdb's code quite well. ;-) In that case, could you explain why gdb isn't using WCONTINUED notifications for selective thread wakeup instead of trying to achieve the same at group stop time? Background thread: https://lkml.org/lkml/2011/9/7/294 Basically the only argument against making ptrace not automatically continue group stopped tasks was that gdb uses that to resume one thread instead of all. But it does this by letting the tasks hang in trapped state at group stop notification time, breaking normal group stops. If it would instead use the group continuation notification, gdb would never need to directly interfere with group stops. I find PTRACE_LISTEN an ugly solution to this problem and am searching for something nicer. Greetings, Indan