From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.5 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,USER_AGENT_MUTT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 7C2A3C04EB8 for ; Tue, 4 Dec 2018 16:41:24 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 4C639206B6 for ; Tue, 4 Dec 2018 16:41:24 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 4C639206B6 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=arm.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727008AbeLDQlX (ORCPT ); Tue, 4 Dec 2018 11:41:23 -0500 Received: from foss.arm.com ([217.140.101.70]:36710 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726367AbeLDQlW (ORCPT ); Tue, 4 Dec 2018 11:41:22 -0500 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.72.51.249]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 5A52CA78; Tue, 4 Dec 2018 08:41:22 -0800 (PST) Received: from edgewater-inn.cambridge.arm.com (usa-sjc-imap-foss1.foss.arm.com [10.72.51.249]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPA id 28F673F614; Tue, 4 Dec 2018 08:41:22 -0800 (PST) Received: by edgewater-inn.cambridge.arm.com (Postfix, from userid 1000) id 862C81AE0BA0; Tue, 4 Dec 2018 16:41:42 +0000 (GMT) Date: Tue, 4 Dec 2018 16:41:42 +0000 From: Will Deacon To: Nicolas Boichat Cc: Robin Murphy , Joerg Roedel , linux-arm-kernel@lists.infradead.org, iommu@lists.linux-foundation.org, linux-kernel@vger.kernel.org, Christoph Lameter , Vlastimil Babka , Michal Hocko , linux-mm@kvack.org, Yong Wu , Matthias Brugger , Tomasz Figa , yingjoe.chen@mediatek.com, hch@infradead.org, Matthew Wilcox Subject: Re: [PATCH v3, RFC] iommu/io-pgtable-arm-v7s: Use page_frag to request DMA32 memory Message-ID: <20181204164142.GA8520@arm.com> References: <20181204082300.95106-1-drinkcat@chromium.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20181204082300.95106-1-drinkcat@chromium.org> User-Agent: Mutt/1.5.23 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Dec 04, 2018 at 04:23:00PM +0800, Nicolas Boichat wrote: > IOMMUs using ARMv7 short-descriptor format require page tables > (level 1 and 2) to be allocated within the first 4GB of RAM, even > on 64-bit systems. > > For level 1/2 tables, ensure GFP_DMA32 is used if CONFIG_ZONE_DMA32 > is defined (e.g. on arm64 platforms). > > For level 2 tables (1 KB), we use page_frag to allocate these pages, > as we cannot directly use kmalloc (no slab cache for GFP_DMA32) or > kmem_cache (mm/ code treats GFP_DMA32 as an invalid flag). > > One downside is that we only free the allocated page if all the > 4 fragments (4 IOMMU L2 tables) are freed, but given that we > usually only allocate limited number of IOMMU L2 tables, this > should not have too much impact on memory usage: In the absolute > worst case (4096 L2 page tables, each on their own 4K page), > we would use 16 MB of memory for 4 MB of L2 tables. > > Also, print an error when the physical address does not fit in > 32-bit, to make debugging easier in the future. > > Fixes: ad67f5a6545f ("arm64: replace ZONE_DMA with ZONE_DMA32") > Signed-off-by: Nicolas Boichat > --- > > As an alternative to the series [1], which adds support for GFP_DMA32 > to kmem_cache in mm/. IMHO the solution in [1] is cleaner and more > efficient, as it allows freed fragments (L2 tables) to be reused, but > this approach does not require any core change. > > [1] https://patchwork.kernel.org/cover/10677529/, 3 patches > > drivers/iommu/io-pgtable-arm-v7s.c | 32 ++++++++++++++++-------------- > 1 file changed, 17 insertions(+), 15 deletions(-) > > diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c > index 445c3bde04800c..0de6a51eb6755f 100644 > --- a/drivers/iommu/io-pgtable-arm-v7s.c > +++ b/drivers/iommu/io-pgtable-arm-v7s.c > @@ -161,6 +161,12 @@ > > #define ARM_V7S_TCR_PD1 BIT(5) > > +#ifdef CONFIG_ZONE_DMA32 > +#define ARM_V7S_TABLE_GFP_DMA GFP_DMA32 > +#else > +#define ARM_V7S_TABLE_GFP_DMA GFP_DMA > +#endif We may as well include __GFP_ZERO in here too. Anyway, this looks alright to me: Acked-by: Will Deacon But it sounds like you're still on the fence about this patch, so I won't pick it up unless you ask explicitly. Will