Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 30 Mar 2001 04:06:18 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 30 Mar 2001 04:06:09 -0500 Received: from mail.zmailer.org ([194.252.70.162]:43021 "EHLO zmailer.org") by vger.kernel.org with ESMTP id ; Fri, 30 Mar 2001 04:05:58 -0500 Date: Fri, 30 Mar 2001 12:05:01 +0300 From: Matti Aarnio To: George Bonser Cc: linux-kernel@vger.kernel.org, Linus Torvalds Subject: Re: 2.4.3 aic7xxx wont compile Message-ID: <20010330120501.P23336@mea-ext.zmailer.org> In-Reply-To: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: ; from george@gator.com on Thu, Mar 29, 2001 at 11:19:22PM -0800 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Mar 29, 2001 at 11:19:22PM -0800, George Bonser wrote: > Just tried to build 2.4.3, got: > > make[6]: Entering directory > `/usr/local/src/linux/drivers/scsi/aic7xxx/aicasm' > gcc -I/usr/include -ldb1 aicasm_gram.c aicasm_scan.c aicasm.c > aicasm_symbol.c -o aicasm > aicasm/aicasm_gram.y:45: ../queue.h: No such file or directory > aicasm/aicasm_gram.y:50: aicasm.h: No such file or directory ... > `/usr/local/src/linux/drivers/scsi/aic7xxx/aicasm' > ... > Looks like something's missing here. Had 2.4.2 patched to 2.4.3-pre7, backed > out pre7 and applied 2.4.3. Yes, "-I." from gcc flags. The sad part is that people have been patching right and left to get that monster utility to compile because the dependencies say that it must be used to remake the AIC sequencer binary image; which image is perfectly ok except of its timestampts due to patching process. Sources from a tarball never get to this, because Linus has suffered the episode of patching it in, and compiling once -> timestamps are such that the resulting binary (ok, hex version of it) is newer than the source. There are two bugs in those aic7xxx/aicasm makefile(s): - Missing that -I. parameter for ${CC} - Having any sort of dependency from sequencer's images hexified binary (which the driver includes) to its source. While mr. Gibbs is right that it makes a lot of sense to supply the entire tool-chain to generate all generated parts of the aic7xxx driver, nobody should be forced to do the compilation, or then 'make mrproper' should throw away the hexified sequencer code... Compiling of the sequencer source should be *some* make target in that driver directory, but not any of which normal users will encounter at any time. That is, not a dependency target ! Very few people need to go poking at the sequencer source, and they should manage to take a bit more complicated approach, than 'make bzImage' or 'make modules' to achieve that. (i.e.: cd drivers/scsi/aic7xxx;make XYZ) /Matti Aarnio -- presenting own opinnions, obviously - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/