From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jaehoon Chung Subject: Re: [PATCHv10 0/5] ARM: remove the sub-node and deprecate supports-highspeed property for dwmmc. Date: Mon, 25 Aug 2014 21:11:51 +0900 Message-ID: <53FB2807.7070503@samsung.com> References: <1407397082-32495-1-git-send-email-jh80.chung@samsung.com> <20140825112151.GA21336@amd> <53FB1DD5.3050207@samsung.com> <20140825113719.GA31960@amd> <53FB278B.7010509@samsung.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from mailout4.samsung.com ([203.254.224.34]:55926 "EHLO mailout4.samsung.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753027AbaHYMLx (ORCPT ); Mon, 25 Aug 2014 08:11:53 -0400 In-reply-to: <53FB278B.7010509@samsung.com> Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Pavel Machek Cc: Mark Rutland , devicetree@vger.kernel.org, Ulf Hansson , kgene.kim@samsung.com, heiko@sntech.de, tgih.jun@samsung.com, linux-mmc , Chris Ball , linux-samsung-soc@vger.kernel.org, dinguyen@altera.com, linux-arm-kernel@lists.infradead.org On 08/25/2014 09:09 PM, Jaehoon Chung wrote: > On 08/25/2014 08:37 PM, Pavel Machek wrote: >> On Mon 2014-08-25 20:28:21, Jaehoon Chung wrote: >>> On 08/25/2014 08:21 PM, Pavel Machek wrote: >>>> On Thu 2014-08-07 16:37:57, Jaehoon Chung wrote: >>>>> Since used the mmc_of_parse(), didn't parse the sub-node. >>>>> So we can remove the sub-node, because almost SoC used the only one card per a host. >>>>> And supports-highspeed can be replaced with "cap-mmc/sd-highspeed" >>>>> property. >>>> >>>> Would it be better to fix parsing of the device tree, and not to >>>> change all the device trees? >>>> >>>> Someone will want to do two slots sooner or later... >>>> >>> >>> First, I had considered that controller can be supported the multiple slot. >>> But MMC maintainers and other people suggested that consider the only one card per a host. >>> Two slots or more don't have any benefit, (power or performance, etc). >> >> Would you elaborate? >> >> If I have a device like a phone, I may want to put one "slot" inside >> phone for basic system, and offer second slot for user expansion >> (initially empty). > > if multiple slot is supported, then a mmcqd should be processing for multiple slots. > It's too inefficient, and affect the whole performance reduction. Sorry, Discard this comment. it means dwmci, not mmcqd. > > If want to offer the second slot for user expansion, add the host for expansion slot. > Almost All SoC didn't use the multiple slot per a host controller for eMMC/SD/SDIO. > > If Some device(Phone) supports the SD-card and eMMC, then there are two Host IP. > One Host IP is used for eMMC, other is used for SD-card. > > this is H/W design issue. > > a) You means the below, > > One Host IP -------- eMMC > | > ---- SD > | > ---- SDIO > > b) We means the below > One Host IP -------- eMMC > One Host IP -------- SD > One Host IP -------- SDIO > > In now, I knew every SoC have used like b) type. I didn't see a) type (especially, dwmmc's case). > > If i missed something, let me know, plz. > > Best Regards, > Jaehoon Chung > >> >> Or I may want to have internal slot with a card to boot from and have >> external slot (initially empty) for system update for embedded system. >> >> I see quite an obvious benefit there. >> Pavel >> >