From: Christoph Hellwig Subject: Re: [PATCH 07/21] dma-iommu: move the arm64 wrappers to common code Date: Tue, 9 Apr 2019 19:15:11 +0200 Message-ID: <20190409171511.GC14679@lst.de> References: <20190327080448.5500-1-hch@lst.de> <20190327080448.5500-8-hch@lst.de> <67573dd3-72c7-692d-bc1a-7edb49ff9551@arm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <67573dd3-72c7-692d-bc1a-7edb49ff9551@arm.com> Sender: linux-kernel-owner@vger.kernel.org To: Robin Murphy Cc: Christoph Hellwig , Joerg Roedel , Catalin Marinas , Will Deacon , Tom Lendacky , iommu@lists.linux-foundation.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org List-Id: iommu@lists.linux-foundation.org On Tue, Apr 09, 2019 at 04:07:02PM +0100, Robin Murphy wrote: >> +static void iommu_dma_unmap_page(struct device *dev, dma_addr_t dma_handle, >> + size_t size, enum dma_data_direction dir, unsigned long attrs) >> +{ >> + if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC)) >> + iommu_dma_sync_single_for_cpu(dev, dma_handle, size, dir); >> + __iommu_dma_unmap(iommu_get_domain_for_dev(dev), dma_handle, size); > > That wants to be iommu_get_dma_domain() there to minimise the overhead. In > fact, I guess this could now be streamlined a bit in the style of > iommu_dma_map_page() above - i.e. avoid doing a second domain lookup in the > sync case - but that can happen later (if indeed you haven't already). Yes, this should be iommu_get_dma_domain - this got lost during a rebase to the kernel version that changed to the iommu_get_dma_domain calls. I don't think I've optimized to remove the additional call, but I can easily throw in another patch to do that.