From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 26471 invoked by alias); 15 Mar 2012 20:50:20 -0000 Received: (qmail 26462 invoked by uid 22791); 15 Mar 2012 20:50:19 -0000 X-SWARE-Spam-Status: No, hits=-6.9 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_HI,SPF_HELO_PASS,T_RP_MATCHES_RCVD X-Spam-Check-By: sourceware.org Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 15 Mar 2012 20:50:07 +0000 Received: from int-mx09.intmail.prod.int.phx2.redhat.com (int-mx09.intmail.prod.int.phx2.redhat.com [10.5.11.22]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id q2FKo6Dd019628 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Thu, 15 Mar 2012 16:50:06 -0400 Received: from barimba (ovpn01.gateway.prod.ext.phx2.redhat.com [10.5.9.1]) by int-mx09.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id q2FKo5R6028669 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NO); Thu, 15 Mar 2012 16:50:06 -0400 From: Tom Tromey To: Pedro Alves Cc: Sergio Durigan Junior , gdb-patches@sourceware.org Subject: Re: [PATCH 2/3] Implement new features needed for handling SystemTap probes References: <4F620C71.8060501@redhat.com> Date: Thu, 15 Mar 2012 20:50:00 -0000 In-Reply-To: <4F620C71.8060501@redhat.com> (Pedro Alves's message of "Thu, 15 Mar 2012 15:36:17 +0000") Message-ID: <87ehsty4g2.fsf@fleche.redhat.com> User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.0.94 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2012-03/txt/msg00580.txt.bz2 >>>>> "Pedro" == Pedro Alves writes: Pedro> I don't understand why we'd now skip the test when we don't have Pedro> the unwinder stap probe, I think the patch checks for the probe and for the unwinder debug hook, and bails if they are both missing. If either one exists then the test should work. Pedro> This made me wonder about something else with this semaphore Pedro> handling: the target can itself stop tracing, without GDB requesting Pedro> it. E.g., if the trace buffer is full. If so, then you'll miss Pedro> decrementing the semaphore count... Even worse with disconnected Pedro> tracing; GDB might not even be connected when the tracing stops, Pedro> and when you reconnect, you have no clue whether to decrement Pedro> the counts or not... Leaving the semaphore enabled in corner cases is not terrible. It may affect performance a little, but that is all. Anyone affected will already be doing reasonably sophisticated debugging -- running disconnected trace experiments. If it is important they could also just reset the semaphore to 0 manually. Tom