From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756183AbYELWOT (ORCPT ); Mon, 12 May 2008 18:14:19 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752954AbYELWOL (ORCPT ); Mon, 12 May 2008 18:14:11 -0400 Received: from vena.lwn.net ([206.168.112.25]:37977 "EHLO vena.lwn.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751546AbYELWOK (ORCPT ); Mon, 12 May 2008 18:14:10 -0400 To: linux-kernel@vger.kernel.org Subject: V4L2-based build error cc: mchehab@infradead.org From: Jonathan Corbet Date: Mon, 12 May 2008 16:14:09 -0600 Message-ID: <18844.1210630449@vena.lwn.net> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org If you configure V4L2 as a module, the config system will still let you configure DVB_CORE built-in. This yields: drivers/built-in.o: In function `v4l2_i2c_drv_attach_legacy': tuner-core.c:(.text+0x901d5): undefined reference to `v4l2_i2c_attach' drivers/built-in.o: In function `tuner_command': tuner-core.c:(.text+0x91aed): undefined reference to `v4l_printk_ioctl' I never had trouble with this before 2.6.26-rc2. I'm happy making DVB_CORE modular, but it would be nice if the config system didn't allow inconsistent setups. jon P.S. As I was digging into this, I ran into , which takes an interesting approach - function definitions and module_init() within a header file? This thing is included in a number of places; it seems like there should be a better way...