From: Jacob Pan <jacob.jun.pan@linux.intel.com> To: Jason Gunthorpe <jgg@nvidia.com> Cc: LKML <linux-kernel@vger.kernel.org>, iommu@lists.linux-foundation.org, Joerg Roedel <joro@8bytes.org>, Lu Baolu <baolu.lu@linux.intel.com>, Jean-Philippe Brucker <jean-philippe@linaro.com>, Christoph Hellwig <hch@infradead.org>, Yi Liu <yi.l.liu@intel.com>, Raj Ashok <ashok.raj@intel.com>, "Tian, Kevin" <kevin.tian@intel.com>, Dave Jiang <dave.jiang@intel.com>, wangzhou1@hisilicon.com, zhangfei.gao@linaro.org, vkoul@kernel.org, David Woodhouse <dwmw2@infradead.org>, jacob.jun.pan@linux.intel.com Subject: Re: [PATCH v4 1/2] iommu/sva: Tighten SVA bind API with explicit flags Date: Tue, 11 May 2021 11:05:50 -0700 [thread overview] Message-ID: <20210511110550.477a434f@jacob-builder> (raw) In-Reply-To: <20210511163521.GN1002214@nvidia.com> Hi Jason, On Tue, 11 May 2021 13:35:21 -0300, Jason Gunthorpe <jgg@nvidia.com> wrote: > On Tue, May 11, 2021 at 09:14:52AM -0700, Jacob Pan wrote: > > > > Honestly, I'm not convinced we should have "kernel SVA" at all.. Why > > > does IDXD use normal DMA on the RID for kernel controlled accesses? > > > > Using SVA simplifies the work submission, there is no need to do > > map/unmap. Just bind PASID with init_mm, then submit work directly > > either with ENQCMDS (supervisor version of ENQCMD) to a shared > > workqueue or put the supervisor PASID in the descriptor for dedicated > > workqueue. > > That is not OK, protable drivers in Linux have to sue dma map/unmap > calls to manage cache coherence. PASID does not opt out of any of > that. > Let me try to break down your concerns: 1. portability - driver uses DMA APIs can function w/ and w/o IOMMU. is that your concern? But PASID is intrinsically tied with IOMMU and if the drivers are using a generic sva-lib API, why they are not portable? SVA by its definition is to avoid map/unmap every time. 2. cache coherence - as you suggested if we name the flag "direct_map", there is no mapping change, then there is no need for mmu_notifier like tlb flush calls, right? it is caller's responsibility to make sure vmalloc are not used. > I dislike this whole idea a lot. A single driver should not opt itself > out of IOMMU based security "just because" > Perhaps I missed your point here. This is NOT a single driver, privileged access is a PCIe spec defined feature. We are using IOMMU sva-lib APIs, not sure why it is by-passing. Perhaps we should add check for struct pci_dev->untrusted && rciep to address security? > Jason Thanks, Jacob
WARNING: multiple messages have this Message-ID (diff)
From: Jacob Pan <jacob.jun.pan@linux.intel.com> To: Jason Gunthorpe <jgg@nvidia.com> Cc: vkoul@kernel.org, "Tian, Kevin" <kevin.tian@intel.com>, Dave Jiang <dave.jiang@intel.com>, Raj Ashok <ashok.raj@intel.com>, David Woodhouse <dwmw2@infradead.org>, LKML <linux-kernel@vger.kernel.org>, Christoph Hellwig <hch@infradead.org>, iommu@lists.linux-foundation.org, zhangfei.gao@linaro.org, Jean-Philippe Brucker <jean-philippe@linaro.com> Subject: Re: [PATCH v4 1/2] iommu/sva: Tighten SVA bind API with explicit flags Date: Tue, 11 May 2021 11:05:50 -0700 [thread overview] Message-ID: <20210511110550.477a434f@jacob-builder> (raw) In-Reply-To: <20210511163521.GN1002214@nvidia.com> Hi Jason, On Tue, 11 May 2021 13:35:21 -0300, Jason Gunthorpe <jgg@nvidia.com> wrote: > On Tue, May 11, 2021 at 09:14:52AM -0700, Jacob Pan wrote: > > > > Honestly, I'm not convinced we should have "kernel SVA" at all.. Why > > > does IDXD use normal DMA on the RID for kernel controlled accesses? > > > > Using SVA simplifies the work submission, there is no need to do > > map/unmap. Just bind PASID with init_mm, then submit work directly > > either with ENQCMDS (supervisor version of ENQCMD) to a shared > > workqueue or put the supervisor PASID in the descriptor for dedicated > > workqueue. > > That is not OK, protable drivers in Linux have to sue dma map/unmap > calls to manage cache coherence. PASID does not opt out of any of > that. > Let me try to break down your concerns: 1. portability - driver uses DMA APIs can function w/ and w/o IOMMU. is that your concern? But PASID is intrinsically tied with IOMMU and if the drivers are using a generic sva-lib API, why they are not portable? SVA by its definition is to avoid map/unmap every time. 2. cache coherence - as you suggested if we name the flag "direct_map", there is no mapping change, then there is no need for mmu_notifier like tlb flush calls, right? it is caller's responsibility to make sure vmalloc are not used. > I dislike this whole idea a lot. A single driver should not opt itself > out of IOMMU based security "just because" > Perhaps I missed your point here. This is NOT a single driver, privileged access is a PCIe spec defined feature. We are using IOMMU sva-lib APIs, not sure why it is by-passing. Perhaps we should add check for struct pci_dev->untrusted && rciep to address security? > Jason Thanks, Jacob _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu
next prev parent reply other threads:[~2021-05-11 18:03 UTC|newest] Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-05-10 13:25 [PATCH v4 0/2] Simplify and restrict IOMMU SVA APIs Jacob Pan 2021-05-10 13:25 ` Jacob Pan 2021-05-10 13:25 ` [PATCH v4 1/2] iommu/sva: Tighten SVA bind API with explicit flags Jacob Pan 2021-05-10 13:25 ` Jacob Pan 2021-05-10 23:37 ` Jason Gunthorpe 2021-05-10 23:37 ` Jason Gunthorpe 2021-05-11 3:31 ` Jacob Pan 2021-05-11 3:31 ` Jacob Pan 2021-05-11 11:48 ` Jason Gunthorpe 2021-05-11 11:48 ` Jason Gunthorpe 2021-05-11 16:14 ` Jacob Pan 2021-05-11 16:14 ` Jacob Pan 2021-05-11 16:35 ` Jason Gunthorpe 2021-05-11 16:35 ` Jason Gunthorpe 2021-05-11 18:05 ` Jacob Pan [this message] 2021-05-11 18:05 ` Jacob Pan 2021-05-11 19:47 ` Jason Gunthorpe 2021-05-11 19:47 ` Jason Gunthorpe 2021-05-12 6:37 ` Christoph Hellwig 2021-05-12 6:37 ` Christoph Hellwig 2021-05-13 13:00 ` Jacob Pan 2021-05-13 13:00 ` Jacob Pan 2021-05-13 13:38 ` Jason Gunthorpe 2021-05-13 13:38 ` Jason Gunthorpe 2021-05-13 15:10 ` Jacob Pan 2021-05-13 15:10 ` Jacob Pan 2021-05-13 16:44 ` Luck, Tony 2021-05-13 16:44 ` Luck, Tony 2021-05-13 17:33 ` Jason Gunthorpe 2021-05-13 17:33 ` Jason Gunthorpe 2021-05-13 18:53 ` Luck, Tony 2021-05-13 18:53 ` Luck, Tony 2021-05-13 19:00 ` Jason Gunthorpe 2021-05-13 19:00 ` Jason Gunthorpe 2021-05-13 19:14 ` Luck, Tony 2021-05-13 19:14 ` Luck, Tony 2021-05-13 19:20 ` Jason Gunthorpe 2021-05-13 19:20 ` Jason Gunthorpe 2021-05-13 19:46 ` Jacob Pan 2021-05-13 19:46 ` Jacob Pan 2021-05-13 19:57 ` Luck, Tony 2021-05-13 19:57 ` Luck, Tony 2021-05-13 20:22 ` Jacob Pan 2021-05-13 20:22 ` Jacob Pan 2021-05-13 22:31 ` Jason Gunthorpe 2021-05-13 22:31 ` Jason Gunthorpe 2021-05-13 23:40 ` Jacob Pan 2021-05-13 23:40 ` Jacob Pan 2021-05-17 14:37 ` Jason Gunthorpe 2021-05-17 14:37 ` Jason Gunthorpe 2021-05-19 15:46 ` Jacob Pan 2021-05-19 15:46 ` Jacob Pan 2021-05-12 10:18 ` Jean-Philippe Brucker 2021-05-12 10:18 ` Jean-Philippe Brucker 2021-05-10 13:25 ` [PATCH v4 2/2] iommu/sva: Remove mm parameter from SVA bind API Jacob Pan 2021-05-10 13:25 ` Jacob Pan 2021-05-12 10:24 ` Jean-Philippe Brucker 2021-05-12 10:24 ` Jean-Philippe Brucker
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20210511110550.477a434f@jacob-builder \ --to=jacob.jun.pan@linux.intel.com \ --cc=ashok.raj@intel.com \ --cc=baolu.lu@linux.intel.com \ --cc=dave.jiang@intel.com \ --cc=dwmw2@infradead.org \ --cc=hch@infradead.org \ --cc=iommu@lists.linux-foundation.org \ --cc=jean-philippe@linaro.com \ --cc=jgg@nvidia.com \ --cc=joro@8bytes.org \ --cc=kevin.tian@intel.com \ --cc=linux-kernel@vger.kernel.org \ --cc=vkoul@kernel.org \ --cc=wangzhou1@hisilicon.com \ --cc=yi.l.liu@intel.com \ --cc=zhangfei.gao@linaro.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.