From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S264522AbUENJhv (ORCPT ); Fri, 14 May 2004 05:37:51 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S264904AbUENJhv (ORCPT ); Fri, 14 May 2004 05:37:51 -0400 Received: from mailout2.samsung.com ([203.254.224.25]:4802 "EHLO mailout2.samsung.com") by vger.kernel.org with ESMTP id S264522AbUENJhs (ORCPT ); Fri, 14 May 2004 05:37:48 -0400 Date: Fri, 14 May 2004 18:39:22 +0900 From: "Hyok S. Choi" Subject: Re: MMUless CPU support? In-reply-to: <20040513174521.A10776@flint.arm.linux.org.uk> To: Russell King - ARM Linux Cc: linux-kernel mailing-list Message-id: <40A493CA.7030702@samsung.com> MIME-version: 1.0 Content-type: text/plain; format=flowed; charset=us-ascii Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7b) Gecko/20040316 References: <200405040617.42342.jeffs@fairwayacademy.org> <409AC200.7020106@mailcan.com> <20040507084536.A14179@flint.arm.linux.org.uk> <409B97E4.8010100@snapgear.com> <20040507194740.A5778@flint.arm.linux.org.uk> <409E3752.3050102@snapgear.com> <20040509152414.C17714@flint.arm.linux.org.uk> <409EC97D.7030105@samsung.com> <20040510094435.B27722@flint.arm.linux.org.uk> <409F62D5.6080500@samsung.com> <20040510123124.C27722@flint.arm.linux.org.uk> <409F7341.4090207@samsung.com> <042601c43905$beed50e0$0100a8c0@SHUTTLE> <20040513174521.A10776@flint.arm.linux.org.uk> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Russell King - ARM Linux wrote: >On Thu, May 13, 2004 at 06:17:09PM +0200, Vadim Lebedev wrote: > > >>Hyok, >> >>about fork and mmap syscalls on uClinux: >> >>actually i believe it would be better to leave calls.S untouched and return >>EINVAL or similar in the C level code... >>BTW, if i remember correctly (it was in 97 when i've ported uClinux 2.0.xx >>from 68000 to ARM7) mmap call was partially functionnal on ucLinux >>(depending >>on arguments) so >> there is no need to scrap it from calls.S too.... >> >>the less we touch .S file the eathier it'll be to maintain the NO-MMU >>version >> >> > >And you can use assembler/linker magic to alias sys_fork to >sys_ni_syscall. > >Since Hyok seems to be 100% against any kind of merge, it's useless >even talking about this though. > > Hmm, I think about the clean and well structured kind of merge. His comment was useful for me. :-) I always think almost all of tricky method to cross among codes is NO good, for maintainer, code readers, porting guys and all. I like Vadim's method. Regards, Hyok