On 2020/6/13 8:41, Fenghua Yu wrote:
From: Ashok Raj <ashok....@intel.com>
ENQCMD and Data Streaming Accelerator (DSA) and all of their associated
features are a complicated stack with lots of interconnected pieces.
This documentation provides a big picture overview for all of the
Signed-off-by: Ashok Raj <ashok....@intel.com>
Co-developed-by: Fenghua Yu <fenghua...@intel.com>
Signed-off-by: Fenghua Yu <fenghua...@intel.com>
Reviewed-by: Tony Luck <tony.l...@intel.com>
- Fix the doc format and add the doc in toctree (Thomas)
- Modify the doc for better description (Thomas, Tony, Dave)
Documentation/x86/index.rst | 1 +
Documentation/x86/sva.rst | 287 ++++++++++++++++++++++++++++++++++++
2 files changed, 288 insertions(+)
create mode 100644 Documentation/x86/sva.rst
diff --git a/Documentation/x86/index.rst b/Documentation/x86/index.rst
index 265d9e9a093b..e5d5ff096685 100644
@@ -30,3 +30,4 @@ x86-specific Documentation
diff --git a/Documentation/x86/sva.rst b/Documentation/x86/sva.rst
new file mode 100644
@@ -0,0 +1,287 @@
+.. SPDX-License-Identifier: GPL-2.0
+Shared Virtual Addressing (SVA) with ENQCMD
+Shared Virtual Addressing (SVA) allows the processor and device to use the
+same virtual addresses avoiding the need for software to translate virtual
+addresses to physical addresses. SVA is what PCIe calls Shared Virtual
+In addition to the convenience of using application virtual addresses
+by the device, it also doesn't require pinning pages for DMA.
+PCIe Address Translation Services (ATS) along with Page Request Interface
+(PRI) allow devices to function much the same way as the CPU handling
+application page-faults. For more information please refer to PCIe
+specification Chapter 10: ATS Specification.
+Use of SVA requires IOMMU support in the platform. IOMMU also is required
+to support PCIe features ATS and PRI. ATS allows devices to cache
+translations for the virtual address. IOMMU driver uses the mmu_notifier()
+support to keep the device tlb cache and the CPU cache in sync. PRI allows
+the device to request paging the virtual address before using if they are
+not paged in the CPU page tables.
+Shared Hardware Workqueues
+Unlike Single Root I/O Virtualization (SRIOV), Scalable IOV (SIOV) permits
+the use of Shared Work Queues (SWQ) by both applications and Virtual
+Machines (VM's). This allows better hardware utilization vs. hard
+partitioning resources that could result in under utilization. In order to
+allow the hardware to distinguish the context for which work is being
+executed in the hardware by SWQ interface, SIOV uses Process Address Space
+ID (PASID), which is a 20bit number defined by the PCIe SIG.
+PASID value is encoded in all transactions from the device. This allows the
+IOMMU to track I/O on a per-PASID granularity in addition to using the PCIe
+Resource Identifier (RID) which is the Bus/Device/Function.
+ENQCMD is a new instruction on Intel platforms that atomically submits a
+work descriptor to a device. The descriptor includes the operation to be
+performed, virtual addresses of all parameters, virtual address of a completion
+record, and the PASID (process address space ID) of the current process.
+ENQCMD works with non-posted semantics and carries a status back if the
+command was accepted by hardware. This allows the submitter to know if the
+submission needs to be retried or other device specific mechanisms to
+implement implement fairness or ensure forward progress can be made.
+ENQCMD is the glue that ensures applications can directly submit commands
+to the hardware and also permit hardware to be aware of application context
+to perform I/O operations via use of PASID.
+Process Address Space Tagging
+A new thread scoped MSR (IA32_PASID) provides the connection between
+user processes and the rest of the hardware. When an application first
+accesses an SVA capable device this MSR is initialized with a newly
+allocated PASID. The driver for the device calls an IOMMU specific api
+that sets up the routing for DMA and page-requests.
+For example, the Intel Data Streaming Accelerator (DSA) uses
+intel_svm_bind_mm(), which will do the following.
The Intel SVM APIs have been deprecated. Drivers should use
iommu_sva_bind_device() instead. Please also update other places in
+- Allocate the PASID, and program the process page-table (cr3) in the PASID
+ context entries.
+- Register for mmu_notifier() to track any page-table invalidations to keep
+ the device tlb in sync. For example, when a page-table entry is invalidated,
+ IOMMU propagates the invalidation to device tlb. This will force any
+ future access by the device to this virtual address to participate in
+ ATS. If the IOMMU responds with proper response that a page is not
+ present, the device would request the page to be paged in via the PCIe PRI
+ protocol before performing I/O.
+This MSR is managed with the XSAVE feature set as "supervisor state" to
+ensure the MSR is updated during context switch.
+The kernel must allocate a PASID on behalf of each process and program it
+into the new MSR to communicate the process identity to platform hardware.
+ENQCMD uses the PASID stored in this MSR to tag requests from this process.
+When a user submits a work descriptor to a device using the ENQCMD
+instruction, the PASID field in the descriptor is auto-filled with the
+value from MSR_IA32_PASID. Requests for DMA from the device are also tagged
+with the same PASID. The platform IOMMU uses the PASID in the transaction to
+perform address translation. The IOMMU api's setup the corresponding PASID
+entry in IOMMU with the process address used by the CPU (for e.g cr3 in x86).
+The MSR must be configured on each logical CPU before any application
+thread can interact with a device. Threads that belong to the same
+process share the same page tables, thus the same MSR value.
+PASID is cleared when a process is created. The PASID allocation and MSR
+programming may occur long after a process and its threads have been created.
+One thread must call bind() to allocate the PASID for the process. If a
+thread uses ENQCMD without the MSR first being populated, it will cause #GP.
+The kernel will fix up the #GP by writing the process-wide PASID into the
+thread that took the #GP. A single process PASID can be used simultaneously
+with multiple devices since they all share the same address space.
+New threads could inherit the MSR value from the parent. But this would
+involve additional state management for those threads which may never use
+ENQCMD. Clearing the MSR at thread creation permits all threads to have a
+consistent behavior; the PASID is only programmed when the thread calls
+the bind() api (intel_svm_bind_mm()()), or when a thread calls ENQCMD for
+the first time.
+PASID Lifecycle Management
+Only processes that access SVA capable devices need to have a PASID
+allocated. This allocation happens when a process first opens an SVA
+capable device (subsequent opens of the same, or other devices will
+share the same PASID).
+Although the PASID is allocated to the process by opening a device,
+it is not active in any of the threads of that process. Activation is
+done lazily when a thread tries to submit a work descriptor to a device
+using the ENQCMD.
+That first access will trigger a #GP fault because the IA32_PASID MSR
+has not been initialized with the PASID value assigned to the process
+when the device was opened. The Linux #GP handler notes that a PASID as
+been allocated for the process, and so initializes the IA32_PASID MSR
+and returns so that the ENQCMD instruction is re-executed.
+On fork(2) or exec(2) the PASID is removed from the process as it no
+longer has the same address space that it had when the device was opened.
+On clone(2) the new task shares the same address space, so will be
+able to use the PASID allocated to the process. The IA32_PASID is not
+preemptively initialized as the kernel does not know whether this thread
+is going to access the device.
+On exit(2) the PASID is freed. The device driver ensures that any pending
+operations queued to the device are either completed or aborted before
+allowing the PASID to be re-allocated.
+ * Each process has many threads, but only one PASID
+ * Devices have a limited number (~10's to 1000's) of hardware
+ workqueues and each portal maps down to a single workqueue.
+ The device driver manages allocating hardware workqueues.
+ * A single mmap() maps a single hardware workqueue as a "portal"
+ * For each device with which a process interacts, there must be
+ one or more mmap()'d portals.
+ * Many threads within a process can share a single portal to access
+ a single device.
+ * Multiple processes can separately mmap() the same portal, in
+ which case they still share one device hardware workqueue.
+ * The single process-wide PASID is used by all threads to interact
+ with all devices. There is not, for instance, a PASID for each
+ thread or each thread<->device pair.
+* What is SVA/SVM?
+Shared Virtual Addressing (SVA) permits I/O hardware and the processor to
+work in the same address space. In short, sharing the address space. Some
+call it Shared Virtual Memory (SVM), but Linux community wanted to avoid
+it with Posix Shared Memory and Secure Virtual Machines which were terms
+already in circulation.
+* What is a PASID?
+A Process Address Space ID (PASID) is a PCIe-defined TLP Prefix. A PASID is
+a 20 bit number allocated and managed by the OS. PASID is included in all
+transactions between the platform and the device.
+* How are shared work queues different?
+Traditionally to allow user space applications interact with hardware,
+there is a separate instance required per process. For example, consider
+doorbells as a mechanism of informing hardware about work to process. Each
+doorbell is required to be spaced 4k (or page-size) apart for process
+isolation. This requires hardware to provision that space and reserve in
+MMIO. This doesn't scale as the number of threads becomes quite large. The
+hardware also manages the queue depth for Shared Work Queues (SWQ), and
+consumers don't need to track queue depth. If there is no space to accept
+a command, the device will return an error indicating retry. Also
+submitting a command to an MMIO address that can't accept ENQCMD will
+return retry in response. In the new DMWr PCIe terminology, devices need to
+support DMWr completer capability. In addition it requires all switch ports
+to support DMWr routing and must be enabled by the PCIe subsystem, much
+like how PCIe Atomics() are managed for instance.
+SWQ allows hardware to provision just a single address in the device. When
+used with ENQCMD to submit work, the device can distinguish the process
+submitting the work since it will include the PASID assigned to that
+process. This decreases the pressure of hardware requiring to support
+hardware to scale to a large number of processes.
+* Is this the same as a user space device driver?
+Communicating with the device via the shared work queue is much simpler
+than a full blown user space driver. The kernel driver does all the
+initialization of the hardware. User space only needs to worry about
+submitting work and processing completions.
+* Is this the same as SR-IOV?
+Single Root I/O Virtualization (SR-IOV) focuses on providing independent
+hardware interfaces for virtualizing hardware. Hence its required to be
+almost fully functional interface to software supporting the traditional
+BAR's, space for interrupts via MSI-x, its own register layout.
+Virtual Functions (VFs) are assisted by the Physical Function (PF)
+Scalable I/O Virtualization builds on the PASID concept to create device
+instances for virtualization. SIOV requires host software to assist in
+creating virtual devices, each virtual device is represented by a PASID
+along with the BDF of the device. This allows device hardware to optimize
+device resource creation and can grow dynamically on demand. SR-IOV creation
+and management is very static in nature. Consult references below for more
+* Why not just create a virtual function for each app?
+Creating PCIe SRIOV type virtual functions (VF) are expensive. They create
+duplicated hardware for PCI config space requirements, Interrupts such as
+MSIx for instance. Resources such as interrupts have to be hard partitioned
+between VF's at creation time, and cannot scale dynamically on demand. The
+VF's are not completely independent from the Physical function (PF). Most
+VF's require some communication and assistance from the PF driver. SIOV
+creates a software defined device. Where all the configuration and control
+aspects are mediated via the slow path. The work submission and completion
+happen without any mediation.
+* Does this support virtualization?
+ENQCMD can be used from within a guest VM. In these cases the VMM helps
+with setting up a translation table to translate from Guest PASID to Host
+PASID. Please consult the ENQCMD instruction set reference for more
+* Does memory need to be pinned?
+When devices support SVA, along with platform hardware such as IOMMU
+supporting such devices, there is no need to pin memory for DMA purposes.
+Devices that support SVA also support other PCIe features that remove the
+pinning requirement for memory.
+Device TLB support - Device requests the IOMMU to lookup an address before
+use via Address Translation Service (ATS) requests. If the mapping exists
+but there is no page allocated by the OS, IOMMU hardware returns that no
+Device requests that virtual address to be mapped via Page Request
+Interface (PRI). Once the OS has successfully completed the mapping, it
+returns the response back to the device. The device continues again to
+request for a translation and continues.
+IOMMU works with the OS in managing consistency of page-tables with the
+device. When removing pages, it interacts with the device to remove any
+device-tlb that might have been cached before removing the mappings from
+ENQCMD in ISE: