From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1765375AbXGKPPw (ORCPT ); Wed, 11 Jul 2007 11:15:52 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1765195AbXGKPPh (ORCPT ); Wed, 11 Jul 2007 11:15:37 -0400 Received: from mailout.stusta.mhn.de ([141.84.69.5]:58036 "EHLO mailhub.stusta.mhn.de" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1765098AbXGKPPg (ORCPT ); Wed, 11 Jul 2007 11:15:36 -0400 Date: Wed, 11 Jul 2007 17:16:08 +0200 From: Adrian Bunk To: Arnd Bergmann Cc: Miles Bader , linux-kernel@vger.kernel.org Subject: Re: [2.6 patch] re-add required code to include/asm-v850/unistd.h Message-ID: <20070711151608.GG3492@stusta.de> References: <20070707211124.GN3492@stusta.de> <200707072341.16389.arnd@arndb.de> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <200707072341.16389.arnd@arndb.de> User-Agent: Mutt/1.5.16 (2007-06-11) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sat, Jul 07, 2007 at 11:41:15PM +0200, Arnd Bergmann wrote: > On Saturday 07 July 2007, Adrian Bunk wrote: > > > > Commit f5738ceed46782aea7663d62cb6398eb05fc4ce0 removed a bit too much, > > resulting in the following compile error: > > Sorry about the trouble this has caused you. No problem, I only ran into it when doing crosscompile tests. > >  #ifdef __KERNEL__ > >   > > +/* Syscall protocol: > > +   Syscall number in r12, args in r6-r9, r13-r14 > > +   Return value in r10 > > +   Trap 0 for `short' syscalls, where all the args can fit in function > > +   call argument registers, and trap 1 when there are additional args in > > +   r13-r14.  */ > > + > > +#define SYSCALL_NUM    "r12" > > +#define SYSCALL_ARG0   "r6" > > +#define SYSCALL_ARG1   "r7" > > +#define SYSCALL_ARG2   "r8" > > +#define SYSCALL_ARG3   "r9" > > +#define SYSCALL_ARG4   "r13" > > +#define SYSCALL_ARG5   "r14" > > +#define SYSCALL_RET    "r10" > > + > > +#define SYSCALL_SHORT_TRAP     "0" > > +#define SYSCALL_LONG_TRAP      "1" > > + > > +/* Registers clobbered by any syscall.  This _doesn't_ include the syscall > > +   number (r12) or the `extended arg' registers (r13, r14), even though > > +   they are actually clobbered too (this is because gcc's `asm' statement > > +   doesn't allow a clobber to be used as an input or output).  */ > > +#define SYSCALL_CLOBBERS       "r1", "r5", "r11", "r15", "r16", \ > > +                               "r17", "r18", "r19" > > + > > +/* Registers clobbered by a `short' syscall.  This includes all clobbers > > +   except the syscall number (r12).  */ > > +#define SYSCALL_SHORT_CLOBBERS SYSCALL_CLOBBERS, "r13", "r14" > > + > > Wouldn't it be more appropriate to just open-code the contents of these > macros in the place where they are used? > > Even better probably would be to convert the v850 kernel_thread code > to call do_fork directly, like most other architectures do. The bigger problem seems to be that v850 is not that near at compiling at all... :-( > Arnd <>< cu Adrian -- "Is there not promise of rain?" Ling Tan asked suddenly out of the darkness. There had been need of rain for many days. "Only a promise," Lao Er said. Pearl S. Buck - Dragon Seed