Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751511AbdF1Hdk (ORCPT ); Wed, 28 Jun 2017 03:33:40 -0400 Received: from mail-qt0-f172.google.com ([209.85.216.172]:33809 "EHLO mail-qt0-f172.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751084AbdF1Hdf (ORCPT ); Wed, 28 Jun 2017 03:33:35 -0400 MIME-Version: 1.0 In-Reply-To: <20170627173640.GA375@lst.de> References: <1498468741-12020-1-git-send-email-vladimir.murzin@arm.com> <20170627152753.GA29186@lst.de> <20170627165416.GP4902@n2100.armlinux.org.uk> <20170627173640.GA375@lst.de> From: Benjamin Gaignard Date: Wed, 28 Jun 2017 09:33:33 +0200 Message-ID: Subject: Re: [PATCH v6 0/7] ARM: Fix dma_alloc_coherent() and friends for NOMMU To: Christoph Hellwig Cc: Russell King - ARM Linux , Vladimir Murzin , linux-arm-kernel@lists.infradead.org, Linux Kernel Mailing List , =?UTF-8?B?U3plbXrFkSBBbmRyw6Fz?= , Arnd Bergmann , Greg Kroah-Hartman , Andrew Morton , Alexandre Torgue , robin.murphy@arm.com, kbuild-all@01.org, Marek Szyprowski , Joerg Roedel , Christian Borntraeger , Michal Nazarewicz , Alan Stern , Yoshinori Sato , Rich Felker , Roger Quadros , Rob Herring , Mark Rutland , Doug Ledford Content-Type: text/plain; charset="UTF-8" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1113 Lines: 24 2017-06-27 19:36 GMT+02:00 Christoph Hellwig : > On Tue, Jun 27, 2017 at 05:54:17PM +0100, Russell King - ARM Linux wrote: >> On Tue, Jun 27, 2017 at 05:27:53PM +0200, Christoph Hellwig wrote: >> > I've applied patches 1, 3 and 4 to the dma-mapping tree. 2 will need >> > a respin for less code duplication. I'd be happy to take 5,6 and 7 if >> > Russell is fine with that. >> >> If Vladimir is okay with those going in without patch 2, then I'm >> happy - I seem to remember a mail from Vladimir stating that they >> depended on some of the previous patches, but I don't know which. >> >> Acked-by: Russell King > > Thanks, I'll add them. Patch 2 only adds optional mmap functionality, > so it's not really required. Although I'd really like to see a fixed > up version that I can merge anyway. Without mmap we will have a regression for nommu platform which are using dma_alloc_coherent and specialy in drm/kms where dumb buffer are allocated like that. Even with this problem I think the patches should be merged but we should have a quick fix for mmap. Benjamin