From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexey Kardashevskiy Subject: Re: [PATCH] vfio: Fix endianness handling for emulated BARs Date: Wed, 25 Jun 2014 12:43:17 +1000 Message-ID: <53AA3745.7070108@ozlabs.ru> References: <1403091391-31780-1-git-send-email-aik@ozlabs.ru> <1403116512.3707.175.camel@ul30vt.home> <53A233E9.6030006@ozlabs.ru> <53A241F6.9010307@ozlabs.ru> <53A25D74.5000804@ozlabs.ru> <1403234514.3707.278.camel@ul30vt.home> <1403305961.4587.66.camel@pasglop> <53A94EBD.101@ozlabs.ru> <53A955F5.6050801@suse.de> <53A9741B.1040500@ozlabs.ru> <53A97486.4070604@suse.de> <53A976B7.3070709@ozlabs.ru> <53A97BA9.4020702@suse.de> <1403619685.16078.15.camel@ul30vt.home> <53A98C38.8020205@ozlabs.ru> <1403646877.4587.184.camel@pasglop> Mime-Version: 1.0 Content-Type: text/plain; charset=KOI8-R Content-Transfer-Encoding: 7bit Cc: Alex Williamson , Alexander Graf , linuxppc-dev@lists.ozlabs.org, kvm@vger.kernel.org, linux-kernel@vger.kernel.org, Nikunj A Dadhania To: Benjamin Herrenschmidt Return-path: Received: from mail-pa0-f45.google.com ([209.85.220.45]:50063 "EHLO mail-pa0-f45.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753629AbaFYCnX (ORCPT ); Tue, 24 Jun 2014 22:43:23 -0400 Received: by mail-pa0-f45.google.com with SMTP id rd3so1064103pab.18 for ; Tue, 24 Jun 2014 19:43:23 -0700 (PDT) In-Reply-To: <1403646877.4587.184.camel@pasglop> Sender: kvm-owner@vger.kernel.org List-ID: On 06/25/2014 07:54 AM, Benjamin Herrenschmidt wrote: > On Wed, 2014-06-25 at 00:33 +1000, Alexey Kardashevskiy wrote: >> >> I do not understand why @val is considered LE here and need to be >> converted >> to CPU. Really. I truly believe it should be cpu_to_le32(). > > No. Both are slightly wrong semantically but le32_to_cpu() is less > wrong :-) > > iowrite32 supposedly takes a "cpu" value as argument and writes an "le" > value. So if anything, you need something that converts to a "cpu" value > before you call iowrite32. > > Now it's still slightly wrong because the "input" to le32_to_cpu() is > supposed to be an "LE" value but of course here it's not, it's a "cpu" > value too :-) > > But yes, I agree with aw here, either do nothing or stick a set of > iowriteXX_native or something equivalent in the generic iomap header, > define them in term of iowrite32be/iowrite32 based on the compile time > endian of the arch. Ok. I'll do nothing. > Hitting asm-generic/iomap.h I think will cover all archs except ARM. > For ARM, just hack arch/arm/asm/io.h -- Alexey