From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:57248) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1amlPR-0003uD-1T for qemu-devel@nongnu.org; Sun, 03 Apr 2016 13:01:57 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1amlPQ-0001bo-2j for qemu-devel@nongnu.org; Sun, 03 Apr 2016 13:01:56 -0400 Received: from mail-vk0-x229.google.com ([2607:f8b0:400c:c05::229]:36834) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1amlPP-0001bf-Tp for qemu-devel@nongnu.org; Sun, 03 Apr 2016 13:01:56 -0400 Received: by mail-vk0-x229.google.com with SMTP id c4so29366051vkb.3 for ; Sun, 03 Apr 2016 10:01:55 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <0F59BEF3-8BDD-434B-9EC4-3DECF9465A44@livius.net> References: <0F59BEF3-8BDD-434B-9EC4-3DECF9465A44@livius.net> From: Peter Maydell Date: Sun, 3 Apr 2016 18:01:35 +0100 Message-ID: Content-Type: text/plain; charset=UTF-8 Subject: Re: [Qemu-devel] segmentation fault in object.c:type_initialize_interface() if interface not defined List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Liviu Ionescu Cc: QEMU Developers On 3 April 2016 at 17:57, Liviu Ionescu wrote: > >> On 03 Apr 2016, at 19:43, Peter Maydell wrote: >> >> On 3 April 2016 at 16:40, Liviu Ionescu wrote: >>> >>>> On 03 Apr 2016, at 15:28, Peter Maydell wrote: >>> >>>> since hw/arm/boot.o is in obj-y it should always be built, >>> >>> not necessarily, in my build configuration I have if's that >>> exclude most of the files non related to Cortex-M. >> >> boot.o is in obj-y; it is (I think) impossible to >> build a v7M supporting QEMU without boot.o unless you've modified >> the makefiles for some reason. > > that's nothing more than the usual make magic, most of the makefiles look like this: > > > ifeq ($(CONFIG_GNU_ARM_ECLIPSE),n) This is you changing QEMU to not compile a file that was previously always compiled. If you do that then it's not surprising if things might break when you move to an upstream version, that's all. Ideally we'd have more CONFIG_SOME_DEVICE and CONFIG_SOME_BOARD checks here the way we do for the more recent boards like the Xilinx and iMX ones, then you could disable boards you didn't want in the config file. thanks -- PMM