From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751635AbcGPMPv (ORCPT ); Sat, 16 Jul 2016 08:15:51 -0400 Received: from ec2-52-27-115-49.us-west-2.compute.amazonaws.com ([52.27.115.49]:46780 "EHLO s-opensource.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751447AbcGPMPu (ORCPT ); Sat, 16 Jul 2016 08:15:50 -0400 Message-ID: <578A256F.7040706@osg.samsung.com> Date: Sat, 16 Jul 2016 13:15:43 +0100 From: Luis de Bethencourt User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Icedove/31.8.0 MIME-Version: 1.0 To: Javier Martinez Canillas , linux-kernel@vger.kernel.org CC: Marek Szyprowski , Mauro Carvalho Chehab , Kyungmin Park , Pawel Osciak , linux-media@vger.kernel.org, Hans Verkuil , Shuah Khan , Nicolas Dufresne Subject: Re: [PATCH] [media] vb2: map dmabuf for planes on driver queue instead of vidioc_qbuf References: <1468599966-31988-1-git-send-email-javier@osg.samsung.com> In-Reply-To: <1468599966-31988-1-git-send-email-javier@osg.samsung.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 15/07/16 17:26, Javier Martinez Canillas wrote: > The buffer planes' dma-buf are currently mapped when buffers are queued > from userspace but it's more appropriate to do the mapping when buffers > are queued in the driver since that's when the actual DMA operation are > going to happen. > > Suggested-by: Nicolas Dufresne > Signed-off-by: Javier Martinez Canillas > > --- > > Hello, > > A side effect of this change is that if the dmabuf map fails for some > reasons (i.e: a driver using the DMA contig memory allocator but CMA > not being enabled), the fail will no longer happen on VIDIOC_QBUF but > later (i.e: in VIDIOC_STREAMON). > > I don't know if that's an issue though but I think is worth mentioning. > > Best regards, > Javier > Just run this path on the ODROID using GStreamer and the vivid driver. It worked nicely. Tested-by: Luis de Bethencourt Thanks Javier, Luis