From mboxrd@z Thu Jan 1 00:00:00 1970 From: Harvey Harrison Subject: Re: [PATCH 1/8] kernel: add common infrastructure for unaligned access Date: Thu, 10 Apr 2008 15:37:59 -0700 Message-ID: <1207867079.22001.60.camel@brick> References: <1207866046.22001.53.camel@brick> <1207865183.22001.49.camel@brick> <1207864537.22001.47.camel@brick> <1207856646.22001.25.camel@brick> <11527.1207863801@redhat.com> <11814.1207864864@redhat.com> <11907.1207865758@redhat.com> <11989.1207866793@redhat.com> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <11989.1207866793-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> Sender: linux-arch-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: To: David Howells Cc: Andrew Morton , linux-arch On Thu, 2008-04-10 at 23:33 +0100, David Howells wrote: > Harvey Harrison wrote: > > > Well, for the u16 case, won't the compiler warn about truncating the > > return if I return an int when the function returns u16? > > It doesn't for me for: > > u16 test(const u8 *p) > { > return (u16) (p[0] | p[1] << 8); > } > > > In the u64 case, I need the cast to ensure it expands to u64 rather than > > int from the shift. Agreed in the u32 case, but then it looks different > > than the u16/u64 case (cargo-cult I know, but nice for consistency). > > Agreed, the u64 cast is necessary, but I was talking about casts generally of: > > function(...) > { > return () (...); > } > OK, I'll see what other comments come in and keep it in mind for the next iteration. I'll also move frv over to the generic C code in that submission as we discussed. Harvey From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from wr-out-0506.google.com ([64.233.184.231]:41144 "EHLO wr-out-0506.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753477AbYDJWiK (ORCPT ); Thu, 10 Apr 2008 18:38:10 -0400 Received: by wr-out-0506.google.com with SMTP id c48so232187wra.1 for ; Thu, 10 Apr 2008 15:38:03 -0700 (PDT) Subject: Re: [PATCH 1/8] kernel: add common infrastructure for unaligned access From: Harvey Harrison In-Reply-To: <11989.1207866793@redhat.com> References: <1207866046.22001.53.camel@brick> <1207865183.22001.49.camel@brick> <1207864537.22001.47.camel@brick> <1207856646.22001.25.camel@brick> <11527.1207863801@redhat.com> <11814.1207864864@redhat.com> <11907.1207865758@redhat.com> <11989.1207866793@redhat.com> Content-Type: text/plain Date: Thu, 10 Apr 2008 15:37:59 -0700 Message-ID: <1207867079.22001.60.camel@brick> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: linux-arch-owner@vger.kernel.org List-ID: To: David Howells Cc: Andrew Morton , linux-arch Message-ID: <20080410223759.GMGnxlQaOL23I7qx3wEPzqclm_6J2x-i8BbMuLqXrRU@z> On Thu, 2008-04-10 at 23:33 +0100, David Howells wrote: > Harvey Harrison wrote: > > > Well, for the u16 case, won't the compiler warn about truncating the > > return if I return an int when the function returns u16? > > It doesn't for me for: > > u16 test(const u8 *p) > { > return (u16) (p[0] | p[1] << 8); > } > > > In the u64 case, I need the cast to ensure it expands to u64 rather than > > int from the shift. Agreed in the u32 case, but then it looks different > > than the u16/u64 case (cargo-cult I know, but nice for consistency). > > Agreed, the u64 cast is necessary, but I was talking about casts generally of: > > function(...) > { > return () (...); > } > OK, I'll see what other comments come in and keep it in mind for the next iteration. I'll also move frv over to the generic C code in that submission as we discussed. Harvey