From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Leizhen (ThunderTown)" Subject: Re: [PATCH v4 4/5] iommu/arm-smmu-v3: add support for non-strict mode Date: Tue, 14 Aug 2018 09:49:11 +0800 Message-ID: <5B723517.3010603@huawei.com> References: <1533558424-16748-1-git-send-email-thunder.leizhen@huawei.com> <1533558424-16748-5-git-send-email-thunder.leizhen@huawei.com> <94c88d9e-818e-f893-9ee2-7a0cf425bd9a@arm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <94c88d9e-818e-f893-9ee2-7a0cf425bd9a@arm.com> Sender: linux-kernel-owner@vger.kernel.org To: Robin Murphy , Will Deacon , Joerg Roedel , linux-arm-kernel , iommu , linux-kernel Cc: LinuxArm , Hanjun Guo , Libin List-Id: iommu@lists.linux-foundation.org On 2018/8/9 19:06, Robin Murphy wrote: > On 06/08/18 13:27, Zhen Lei wrote: >> Dynamically choose strict or non-strict mode for page table config based >> on the iommu domain type. >> >> Signed-off-by: Zhen Lei >> --- >> drivers/iommu/arm-smmu-v3.c | 7 ++++++- >> 1 file changed, 6 insertions(+), 1 deletion(-) >> >> diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c >> index 2f1304b..904bc1e 100644 >> --- a/drivers/iommu/arm-smmu-v3.c >> +++ b/drivers/iommu/arm-smmu-v3.c >> @@ -1622,6 +1622,11 @@ static int arm_smmu_domain_finalise(struct iommu_domain *domain) >> if (smmu->features & ARM_SMMU_FEAT_COHERENCY) >> pgtbl_cfg.quirks = IO_PGTABLE_QUIRK_NO_DMA; >> >> + if (domain->type == IOMMU_DOMAIN_DMA) { >> + domain->non_strict = 1; >> + pgtbl_cfg.quirks |= IO_PGTABLE_QUIRK_NON_STRICT; >> + } >> + >> pgtbl_ops = alloc_io_pgtable_ops(fmt, &pgtbl_cfg, smmu_domain); >> if (!pgtbl_ops) >> return -ENOMEM; >> @@ -1782,7 +1787,7 @@ static void arm_smmu_iotlb_sync(struct iommu_domain *domain) >> { >> struct arm_smmu_device *smmu = to_smmu_domain(domain)->smmu; >> >> - if (smmu) >> + if (smmu && !domain->non_strict) > > That doesn't smell right - even in non-strict domains we still need stuff like walk cache invalidations for non-leaf unmaps to be synchronous, so we can't just ignore all sync operations at the driver level. I think the right thing to do to elide the "normal" sync on unmap is to first convert __iommu_dma_unmap to use iommu_unmap_fast()/iommu_tlb_sync(), then make it not issue that sync at all for non-strict domains. OK, I will try it. > > Robin. > >> __arm_smmu_tlb_sync(smmu); >> } >> >> -- >> 1.8.3 >> >> > > . > -- Thanks! BestRegards From mboxrd@z Thu Jan 1 00:00:00 1970 From: thunder.leizhen@huawei.com (Leizhen (ThunderTown)) Date: Tue, 14 Aug 2018 09:49:11 +0800 Subject: [PATCH v4 4/5] iommu/arm-smmu-v3: add support for non-strict mode In-Reply-To: <94c88d9e-818e-f893-9ee2-7a0cf425bd9a@arm.com> References: <1533558424-16748-1-git-send-email-thunder.leizhen@huawei.com> <1533558424-16748-5-git-send-email-thunder.leizhen@huawei.com> <94c88d9e-818e-f893-9ee2-7a0cf425bd9a@arm.com> Message-ID: <5B723517.3010603@huawei.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 2018/8/9 19:06, Robin Murphy wrote: > On 06/08/18 13:27, Zhen Lei wrote: >> Dynamically choose strict or non-strict mode for page table config based >> on the iommu domain type. >> >> Signed-off-by: Zhen Lei >> --- >> drivers/iommu/arm-smmu-v3.c | 7 ++++++- >> 1 file changed, 6 insertions(+), 1 deletion(-) >> >> diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c >> index 2f1304b..904bc1e 100644 >> --- a/drivers/iommu/arm-smmu-v3.c >> +++ b/drivers/iommu/arm-smmu-v3.c >> @@ -1622,6 +1622,11 @@ static int arm_smmu_domain_finalise(struct iommu_domain *domain) >> if (smmu->features & ARM_SMMU_FEAT_COHERENCY) >> pgtbl_cfg.quirks = IO_PGTABLE_QUIRK_NO_DMA; >> >> + if (domain->type == IOMMU_DOMAIN_DMA) { >> + domain->non_strict = 1; >> + pgtbl_cfg.quirks |= IO_PGTABLE_QUIRK_NON_STRICT; >> + } >> + >> pgtbl_ops = alloc_io_pgtable_ops(fmt, &pgtbl_cfg, smmu_domain); >> if (!pgtbl_ops) >> return -ENOMEM; >> @@ -1782,7 +1787,7 @@ static void arm_smmu_iotlb_sync(struct iommu_domain *domain) >> { >> struct arm_smmu_device *smmu = to_smmu_domain(domain)->smmu; >> >> - if (smmu) >> + if (smmu && !domain->non_strict) > > That doesn't smell right - even in non-strict domains we still need stuff like walk cache invalidations for non-leaf unmaps to be synchronous, so we can't just ignore all sync operations at the driver level. I think the right thing to do to elide the "normal" sync on unmap is to first convert __iommu_dma_unmap to use iommu_unmap_fast()/iommu_tlb_sync(), then make it not issue that sync at all for non-strict domains. OK, I will try it. > > Robin. > >> __arm_smmu_tlb_sync(smmu); >> } >> >> -- >> 1.8.3 >> >> > > . > -- Thanks! BestRegards