From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH] mmc: sdhci-brcmstb: Add SDHCI_QUIRK2_BROKEN_64_BIT_DMA Date: Mon, 29 Aug 2016 11:44:32 +0200 Message-ID: <5229438.Oa2AiqCmJN@wuerfel> References: <20160808015803.2528-1-jaedon.shin@gmail.com> <501206D1-D60B-4BA7-99C0-C38065C3AE9E@gmail.com> <7edcef69-d364-8845-a677-3827a50f4da1@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7Bit Return-path: Received: from mout.kundenserver.de ([217.72.192.75]:52555 "EHLO mout.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756125AbcH2Jot (ORCPT ); Mon, 29 Aug 2016 05:44:49 -0400 In-Reply-To: <7edcef69-d364-8845-a677-3827a50f4da1@gmail.com> Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Florian Fainelli Cc: Jaedon Shin , Alan Cooper , Ulf Hansson , Adrian Hunter , bcm-kernel-feedback-list@broadcom.com, linux-mmc@vger.kernel.org On Saturday, August 27, 2016 12:56:10 PM CEST Florian Fainelli wrote: > > Finally, Arnd's suggestions of using "dma-ranges" is fine, but I do not > think we quite need this here because we really need to advertise the > right set of capabilities based on the generation/version of the > controller deployed in specific chips. To be more specific here, I think that without the dma-ranges property you should never be able to set a dma-mask larger than the 32-bit mask, so if you have machines that are capable of high DMA, you should definitely add the property in the bus, even if that is currently ignored. I've suggested a patch before, but I believe both ARM and MIPS ignore this at the moment, and just allow drivers to set arbitrary masks even when the bus does not have a dma-ranges property, and that is a bug. Arnd