Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1169978AbdDXPxZ (ORCPT ); Mon, 24 Apr 2017 11:53:25 -0400 Received: from mail-wm0-f65.google.com ([74.125.82.65]:35915 "EHLO mail-wm0-f65.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S965296AbdDXPxS (ORCPT ); Mon, 24 Apr 2017 11:53:18 -0400 MIME-Version: 1.0 In-Reply-To: <20170424144429.GI12323@arm.com> References: <1492430246-17267-1-git-send-email-sunil.kovvuri@gmail.com> <20170424144429.GI12323@arm.com> From: Sunil Kovvuri Date: Mon, 24 Apr 2017 21:23:16 +0530 Message-ID: Subject: Re: [PATCH] iommu/arm-smmu: Return IOVA in iova_to_phys when SMMU is bypassed To: Will Deacon Cc: iommu@lists.linux-foundation.org, LAKML , LKML , robert.richter@cavium.com, Sunil Goutham Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2476 Lines: 69 On Mon, Apr 24, 2017 at 8:14 PM, Will Deacon wrote: > On Mon, Apr 17, 2017 at 05:27:26PM +0530, sunil.kovvuri@gmail.com wrote: >> From: Sunil Goutham >> >> For software initiated address translation, when domain type is >> IOMMU_DOMAIN_IDENTITY i.e SMMU is bypassed, mimic HW behavior >> i.e return the same IOVA as translated address. >> >> This patch is an extension to Will Deacon's patchset >> "Implement SMMU passthrough using the default domain". > > Are you actually seeing an issue here? If so, why isn't SMMUv3 affected too? Yes and SMMUv3 should also be effected but as of now I don't see any use case. If needed, i can re-submit the patch with changes in SMMUv3 as well. > >> Signed-off-by: Sunil Goutham >> --- >> drivers/iommu/arm-smmu.c | 3 +++ >> 1 file changed, 3 insertions(+) >> >> diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c >> index 41afb07..2f4a130 100644 >> --- a/drivers/iommu/arm-smmu.c >> +++ b/drivers/iommu/arm-smmu.c >> @@ -1405,6 +1405,9 @@ static phys_addr_t arm_smmu_iova_to_phys(struct iommu_domain *domain, >> struct arm_smmu_domain *smmu_domain = to_smmu_domain(domain); >> struct io_pgtable_ops *ops= smmu_domain->pgtbl_ops; >> >> + if (domain->type == IOMMU_DOMAIN_IDENTITY) >> + return iova; >> + >> if (!ops) >> return 0; > > I'd have thought ops would be NULL, since arm_smmu_init_domain_context > doesn't allocate them for an identity domain. Yes ops is set to NULL. > > I don't understand this patch. Please can you explain the problem more > clearly? AFAIK for any driver outside IOMMU there is only one way to identify if device is attached to IOMMU or not and that is by checking iommu_domain. And I don't think it would be appropriate for the driver to check domain->type before calling 'iommu_iova_to_phys()' API. The difference between IOMMU disabled and IOMMU being in passthrough mode is that, in the later case device is still attached to default domain but in former's case it's NULL. So there is no way to differentiate for the external driver whether IOMMU is in passthrough mode or DMA mode. And since ops is NULL in passthrough mode, 'iommu_iova_to_phys()' will return zero. Use case for your reference https://lkml.org/lkml/2017/3/7/299 This driver is for a NIC interface on platform which supports SMMUv2. Let me know if any more details are needed. Thanks, Sunil. > > Will