From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from linux.microsoft.com (linux.microsoft.com [13.77.154.182]) by sourceware.org (Postfix) with ESMTP id DC41C3858038 for ; Sat, 5 Dec 2020 02:38:13 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.3.2 sourceware.org DC41C3858038 Received: from [192.168.254.32] (unknown [47.187.219.45]) by linux.microsoft.com (Postfix) with ESMTPSA id E20D920B717A for ; Fri, 4 Dec 2020 18:38:12 -0800 (PST) DKIM-Filter: OpenDKIM Filter v2.11.0 linux.microsoft.com E20D920B717A Subject: Re: [RFC PATCH v1 1/4] Static Trampolines - Quick question From: "Madhavan T. Venkataraman" To: libffi-discuss References: <20201124193206.10289-1-madvenka@linux.microsoft.com> <20201124193206.10289-2-madvenka@linux.microsoft.com> <977a6824-c0b2-b69e-97ba-ce11639b4ba3@linux.microsoft.com> Message-ID: Date: Fri, 4 Dec 2020 20:38:12 -0600 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: <977a6824-c0b2-b69e-97ba-ce11639b4ba3@linux.microsoft.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit X-Spam-Status: No, score=-20.6 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, ENV_AND_HDR_SPF_MATCH, NICE_REPLY_A, SPF_HELO_PASS, SPF_PASS, TXREP, USER_IN_DEF_DKIM_WL, USER_IN_DEF_SPF_WL autolearn=ham autolearn_force=no version=3.4.2 X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on server2.sourceware.org X-BeenThere: libffi-discuss@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Libffi-discuss mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 05 Dec 2020 02:38:15 -0000 Hi, I am preparing version 2. I have a question. Are we allowed to change FFI_TRAMPOLINE_SIZE, the size of closure->tramp[]? As an example, I checked the git history for ARM. I can see that it has changed a couple of times in the past. Just want to know if there would be any compatibility issues if the size is changed. My solution can be more elegant if the trampoline size can be changed. I recall seeing somewhere that in some cases, a libffi consumer manages closure memory and does not call ffi_closure_alloc() and ffi_closure_free(). In such cases, the consumer would have to be rebuilt if the closure size changed. Do we still support such consumers? Is it OK for them to rebuild? When the size of the closure structure changes, does the library version have to be bumped up? Please advise. Thanks. Madhavan