From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jon Hunter Subject: Re: [PATCH v3 2/2] dmaengine: tegra210-adma: Add memcpy support Date: Thu, 8 Sep 2016 15:19:57 +0100 Message-ID: <37d4645c-6318-78bd-79bb-844fb6764a1b@nvidia.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-tegra-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Nicolin Chen , vinod.koul-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-tegra-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, dmaengine-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, gnurou-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org, thierry.reding-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org, swarren-3lzwWm7+Weoh9ZMKESR00Q@public.gmane.org, ldewangan-DDmLM1+adcrQT0dZR+AlfA@public.gmane.org List-Id: linux-tegra@vger.kernel.org On 08/09/16 15:08, Jon Hunter wrote: > > On 06/09/16 19:42, Nicolin Chen wrote: >> ADMA supports non-flow controlled Memory-to-Memory direction >> transactions. So this patch just adds an initial support for >> that. It passed a simple dmatest: >> echo dma1chan0 > /sys/module/dmatest/parameters/channel >> echo 1024 > /sys/module/dmatest/parameters/iterations >> echo 0 > /sys/module/dmatest/parameters/dmatest >> echo 1 > /sys/module/dmatest/parameters/run >> dmesg | grep dmatest >> Started 1 threads using dma1chan0 >> dma1chan0-copy0: summary 1024 tests, 0 failures 2054 iops 16520 KB/s (0) > > What board and kernel did you try this on? > > I have tried this on a tegra210-jetson-tx1 and I get : > > [ 202.569204] dmatest: Started 1 threads using dma1chan0 > [ 205.620318] dmatest: dma1chan0-copy0: result #1: 'test timed out' with src_off=0x86c dst_off=0xc80 len=0x307c (0) > [ 208.692315] dmatest: dma1chan0-copy0: result #2: 'test timed out' with src_off=0x3288 dst_off=0x2720 len=0xa1c (0) > [ 211.764323] dmatest: dma1chan0-copy0: result #3: 'test timed out' with src_off=0x2f44 dst_off=0x3164 len=0x3a4 (0) > [ 214.836309] dmatest: dma1chan0-copy0: result #4: 'test timed out' with src_off=0x17d0 dst_off=0x2b10 len=0xe2c (0) > [ 217.908305] dmatest: dma1chan0-copy0: result #5: 'test timed out' with src_off=0x23d8 dst_off=0xe90 len=0xb7c (0) > ... > > I also tried a tegra210-smaug and I get: > > [ 167.508828] dmatest: Started 1 threads using dma1chan0 > [ 167.508870] dmatest: dma1chan0-copy0: dstbuf[0x0] not copied! Expected db, got 09 > [ 167.508873] dmatest: dma1chan0-copy0: dstbuf[0x1] not copied! Expected da, got 05 > [ 167.508875] dmatest: dma1chan0-copy0: dstbuf[0x2] not copied! Expected d9, got 26 > [ 167.508876] dmatest: dma1chan0-copy0: dstbuf[0x3] not copied! Expected d8, got 16 > ... I tried this again on my audio testing branch for tegra210 and it worked ... [ 36.427210] dmatest: Started 1 threads using dma1chan0 [ 37.036948] dmatest: dma1chan0-copy0: summary 1024 tests, 0 failures 2410 iops 19419 KB/s (0) Do you have other patches in your branch? If so it would be good to understand the dependency. May be we are missing a clock somewhere? Cheers Jon -- nvpublic From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758500AbcIHOUN (ORCPT ); Thu, 8 Sep 2016 10:20:13 -0400 Received: from hqemgate14.nvidia.com ([216.228.121.143]:8120 "EHLO hqemgate14.nvidia.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758225AbcIHOUL (ORCPT ); Thu, 8 Sep 2016 10:20:11 -0400 X-PGP-Universal: processed; by hqnvupgp07.nvidia.com on Thu, 08 Sep 2016 07:15:16 -0700 Subject: Re: [PATCH v3 2/2] dmaengine: tegra210-adma: Add memcpy support To: Nicolin Chen , References: CC: , , , , , , From: Jon Hunter Message-ID: <37d4645c-6318-78bd-79bb-844fb6764a1b@nvidia.com> Date: Thu, 8 Sep 2016 15:19:57 +0100 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.2.0 MIME-Version: 1.0 In-Reply-To: X-Originating-IP: [10.21.132.106] X-ClientProxiedBy: DRUKMAIL102.nvidia.com (10.25.59.20) To UKMAIL101.nvidia.com (10.26.138.13) 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 08/09/16 15:08, Jon Hunter wrote: > > On 06/09/16 19:42, Nicolin Chen wrote: >> ADMA supports non-flow controlled Memory-to-Memory direction >> transactions. So this patch just adds an initial support for >> that. It passed a simple dmatest: >> echo dma1chan0 > /sys/module/dmatest/parameters/channel >> echo 1024 > /sys/module/dmatest/parameters/iterations >> echo 0 > /sys/module/dmatest/parameters/dmatest >> echo 1 > /sys/module/dmatest/parameters/run >> dmesg | grep dmatest >> Started 1 threads using dma1chan0 >> dma1chan0-copy0: summary 1024 tests, 0 failures 2054 iops 16520 KB/s (0) > > What board and kernel did you try this on? > > I have tried this on a tegra210-jetson-tx1 and I get : > > [ 202.569204] dmatest: Started 1 threads using dma1chan0 > [ 205.620318] dmatest: dma1chan0-copy0: result #1: 'test timed out' with src_off=0x86c dst_off=0xc80 len=0x307c (0) > [ 208.692315] dmatest: dma1chan0-copy0: result #2: 'test timed out' with src_off=0x3288 dst_off=0x2720 len=0xa1c (0) > [ 211.764323] dmatest: dma1chan0-copy0: result #3: 'test timed out' with src_off=0x2f44 dst_off=0x3164 len=0x3a4 (0) > [ 214.836309] dmatest: dma1chan0-copy0: result #4: 'test timed out' with src_off=0x17d0 dst_off=0x2b10 len=0xe2c (0) > [ 217.908305] dmatest: dma1chan0-copy0: result #5: 'test timed out' with src_off=0x23d8 dst_off=0xe90 len=0xb7c (0) > ... > > I also tried a tegra210-smaug and I get: > > [ 167.508828] dmatest: Started 1 threads using dma1chan0 > [ 167.508870] dmatest: dma1chan0-copy0: dstbuf[0x0] not copied! Expected db, got 09 > [ 167.508873] dmatest: dma1chan0-copy0: dstbuf[0x1] not copied! Expected da, got 05 > [ 167.508875] dmatest: dma1chan0-copy0: dstbuf[0x2] not copied! Expected d9, got 26 > [ 167.508876] dmatest: dma1chan0-copy0: dstbuf[0x3] not copied! Expected d8, got 16 > ... I tried this again on my audio testing branch for tegra210 and it worked ... [ 36.427210] dmatest: Started 1 threads using dma1chan0 [ 37.036948] dmatest: dma1chan0-copy0: summary 1024 tests, 0 failures 2410 iops 19419 KB/s (0) Do you have other patches in your branch? If so it would be good to understand the dependency. May be we are missing a clock somewhere? Cheers Jon -- nvpublic