From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from simark.ca (simark.ca [158.69.221.121]) by sourceware.org (Postfix) with ESMTPS id 61E1C386EC42 for ; Thu, 9 Jul 2020 14:16:43 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.3.2 sourceware.org 61E1C386EC42 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=simark.ca Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=simark@simark.ca Received: from [10.0.0.11] (173-246-6-90.qc.cable.ebox.net [173.246.6.90]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by simark.ca (Postfix) with ESMTPSA id D8E6F1E5F9; Thu, 9 Jul 2020 10:16:42 -0400 (EDT) Subject: Re: [PATCH 2/3] Fix crash if connection drops in scoped_restore_current_thread's ctor, part 2 To: Pedro Alves , gdb-patches@sourceware.org References: <20200708233125.1030-1-pedro@palves.net> <20200708233125.1030-3-pedro@palves.net> <1809ad79-7542-776d-cb1f-8a1ce0ed6825@palves.net> From: Simon Marchi Message-ID: <482b9730-707c-adf4-2fd0-c430fb101402@simark.ca> Date: Thu, 9 Jul 2020 10:16:41 -0400 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.10.0 MIME-Version: 1.0 In-Reply-To: <1809ad79-7542-776d-cb1f-8a1ce0ed6825@palves.net> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit X-Spam-Status: No, score=-3.6 required=5.0 tests=BAYES_00, KAM_DMARC_STATUS, KAM_NUMSUBJECT, SPF_HELO_PASS, SPF_PASS, TXREP autolearn=no autolearn_force=no version=3.4.2 X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on server2.sourceware.org X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 09 Jul 2020 14:16:44 -0000 On 2020-07-09 7:12 a.m., Pedro Alves wrote: > Aw, no, I got confused and misremembered how exceptions in ctors work. > The dtor for scoped_restore_current_thread isn't called, and I assumed > it was called. We need to decref the inferior and thread before letting > the exception propagate, otherwise we leak them. I'm testing the updated > version of the patch below, which does that: > > /* Better let this propagate. */ > if (ex.error == TARGET_CLOSE_ERROR) > { > m_thread->decref (); > m_inf->decref (); > throw; > } > > It passes multi-target.exp with Asan-enabled GDB. Running the full > testsuite now. Ok, I also had the reasoning about exceptions thrown by constructors, and thought it might be on purpose that it wasn't decref'ed. If the references were kept by fields of scoped_restore_current_thread, then it would work automatically, because their destructor would get called. Simon