From cba9f302ef1234935521d86bba4086ac8de3a864 Mon Sep 17 00:00:00 2001 From: Aaron Smith Date: Fri, 25 Sep 2020 21:53:49 -0400 Subject: [PATCH] Add support for running inside a VM with vfio-noiommu Fixed link path Check for missing name fix name missing fix err name Add dual return Additional Documentation --- README.md | 31 +++++++++++ docs/dpdk/README-virt.md | 102 ++++++++++++++++++++++++++++++++++ pkg/resources/vfioResource.go | 8 +-- pkg/utils/utils.go | 21 ++++++- pkg/utils/utils_test.go | 5 +- 5 files changed, 158 insertions(+), 9 deletions(-) create mode 100644 docs/dpdk/README-virt.md diff --git a/README.md b/README.md index 747dd323a..b58057e4a 100644 --- a/README.md +++ b/README.md @@ -51,6 +51,7 @@ The SR-IOV network device plugin is Kubernetes device plugin for discovering and - Detects Kubelet restarts and auto-re-register - Detects Link status (for Linux network devices) and updates associated VFs health accordingly - Extensible to support new device types with minimal effort if not already supported +- Works within virtual deployments of Kubernetes that do not have virtualized-iommu support (VFIO No-IOMMU support) To deploy workloads with SR-IOV VF this plugin needs to work together with the following two CNI components: @@ -367,6 +368,36 @@ $ kubectl get node node1 -o json | jq '.status.allocatable' ``` +### Virtual environments with no iommu + +In virtual deployments of Kubernetes where the underlying virtualization platform does not support a virtualized iommu, the VFIO driver needs to be loaded with a special +flag. The file **/etc/modprobe.d/vfio-noiommu.conf** must be created with the contents: + +```` +# cat /etc/modprobe.d/vfio-noiommu.conf +options vfio enable_unsafe_noiommu_mode=1 +```` + +With the above option, vfio devices will be created with the form on the virtual host (VM): + +```` +/dev/vfio/noiommu-0 +/dev/vfio/noiommu-1 +... +```` + +The presence of noiommu-* devices will automatically be detected by the sriov-device-plugin. The noiommu-N devices will be mounted **inside** the pod in their expected/normal location; + +```` +/dev/vfio/0 +/dev/vfio/1 +... +```` +It should be noted that with no IOMMU, there is no way to ensure safe use of DMA. When *enable_unsafe_noiommu_mode* is used, CAP_SYS_RAWIO privileges are necessary to work with groups and +containers using this mode. Use of this mode, specifically +binding a device without a native IOMMU group to a VFIO bus driver will taint the kernel. Only no-iommu support for the vfio-pci bus is provided. However, there are still those users +that want userspace drivers even under those conditions. + ## Example deployments We assume that you have working K8s cluster configured with one of the supported meta plugins for multi-network support. Please see [Features](#features) and [Quick Start](#quick-start) sections for more information on required CNI plugins. diff --git a/docs/dpdk/README-virt.md b/docs/dpdk/README-virt.md new file mode 100644 index 000000000..8c1572fdb --- /dev/null +++ b/docs/dpdk/README-virt.md @@ -0,0 +1,102 @@ +# Running DPDK applications in a Kubernetes virtual environment + +## Pre-requisites + +### Hugepages +DPDK applications require Hugepages memory. Please refer to the [Hugepages section](http://doc.dpdk.org/guides/linux_gsg/sys_reqs.html#use-of-hugepages-in-the-linux-environment) in DPDK getting started guide on hugespages in DPDK. + +Make sure that the virtual environment is enabled for creating VMs with hugepage support. + +Kubernetes nodes can only advertise a single size pre-allocated hugepages. Which means even though one can have both 2M and 1G hugepages in a system, Kubernetes will only recognize the default hugepages as schedulable resources. Workload can request for hugepages using resource requests and limits specifying `hugepage-2Mi` or `hugepage-1Gi` resource references. + +> One important thing to note here is that when requesting hugepage resources, either memory or CPU resource requests need to be specified. + +For more information on hugepage support in Kubernetes please see [here](https://kubernetes.io/docs/tasks/manage-hugepages/scheduling-hugepages/). + + +### VF drivers +DPDK applications require devices to be attached with supported dpdk backend driver. +* For Intel® x700 series NICs `vfio-pci` is required. +* For Mellanox ConnectX®-4 Lx, ConnectX®-5 Adapters `mlx5_core` or `mlx5_ib` is required. + +Native-bifurcating devices/drivers (i.e. Mellanox/mlx5_*) do not need to run with privilege. Non-bifurcating devices/drivers (i.e. Intel/vfio-pci) the PODs need to run with privilege. + +### Privileges +Certain privileges are required for dpdk application to function properly in Kubernetes Pod. The level of privileges depend on the application and the host device driver attached (as mentioned above). When running in an environment without a fully virtualized IOMMU, the *enable_unsafe_noiommu_mode* of vfio must be used by creating a modprobe.d file. + +```` +# cat /etc/modprobe.d/vfio-noiommu.conf +options vfio enable_unsafe_noiommu_mode=1 +```` + +With `vfio-pci` an application must run privilege Pod with **IPC_LOCK** and **CAP_SYS_RAWIO** capability. + +# Example deployment +This directory includes sample deployment yaml files showing how to deploy a dpdk application in Kubernetes with in privileged Pod with SR-IOV VF attached to vfio-pci driver in the case of non-bifurcating nic devices/drivers or the VF attached to the default driver in the case of native-bifurcating devices/drivers (non-priveleged). See [this](https://doc.dpdk.org/guides/howto/flow_bifurcation.html) for more information. + +## Deploy Virtual machines with attached VFs + +1. Depending on the virtualization environment, create a network that supports SR-IOV. Configure the VF as per your requirements: +- Trusted On/Off +- Spoof-Checking On/Off + +In a virtual environment, some VF characteristics are set by the underlying virtualization platform and are used 'as is' inside the VM. A virtual deployment does not have access the VFs associated PF. + +2. Attach the VFs or associated ports to the VM + +## Check that environment supports VFIO and hugepages memory + +1. After deployment of the VM, confirm that your hugepagesz parameter is present. +```` +sh-4.4# cat /proc/cmdline +BOOT_IMAGE=(hd0,gpt1)/ostree/rhcos-92d66d9df4cafad87abd888fd1b22fd1d890e86bc2ad8b9009bb9faa4f403a95/vmlinuz-4.18.0-193.24.1.el8_2.dt1.x86_64 rhcos.root=crypt_rootfs random.trust_cpu=on console=tty0 console=ttyS0,115200n8 rd.luks.options=discard ostree=/ostree/boot.1/rhcos/92d66d9df4cafad87abd888fd1b22fd1d890e86bc2ad8b9009bb9faa4f403a95/0 ignition.platform.id=openstack nohz=on nosoftlockup skew_tick=1 intel_pstate=disable intel_iommu=on iommu=pt rcu_nocbs=2-3 tuned.non_isolcpus=00000003 default_hugepagesz=1G nmi_watchdog=0 audit=0 mce=off processor.max_cstate=1 idle=poll intel_idle.max_cstate=0 +```` +2. On the desired worker node, + +```` +cat /proc/meminfo | grep -i hugepage +AnonHugePages: 245760 kB +ShmemHugePages: 0 kB +HugePages_Total: 8 +HugePages_Free: 8 +HugePages_Rsvd: 0 +HugePages_Surp: 0 +Hugepagesize: 1048576 kB +```` +You should see your requested hugepage size and a non-zero HugePages_Total. + +3. Confirm that Hugepages memory are allocated and mounted +``` +# cat /proc/meminfo | grep -i hugepage +HugePages_Total: 16 +HugePages_Free: 16 +HugePages_Rsvd: 0 +HugePages_Surp: 0 +Hugepagesize: 1048576 kB + +# mount | grep hugetlbfs +hugetlbfs on /dev/hugepages type hugetlbfs (rw,relatime) + +``` + +5. Load vfio-pci module + +```` +# echo "options vfio enable_unsafe_noiommu_mode=1" > /etc/modprobe.d/vfio-noiommu.conf +```` + +``` +modprobe vfio-pci +``` + +7. For non-bifurcating devices/drivers, bind the appropriate interfaces (VF) to the vfio-pci driver. You can use or `driverctl` or [`dpdk-devbind.py`](https://github.com/DPDK/dpdk/blob/master/usertools/dpdk-devbind.py) to bind/unbind drivers using devices PCI addresses. Please see [here](https://dpdk-guide.gitlab.io/dpdk-guide/setup/binding.html) more information on NIC driver bindings. + +Native-bifurcating devices/drivers can stay with the default binding. + +# Performance +It is worth mentioning that to achieve maximum performance from a dpdk application the followings are required: + +1. Application process needs to be pinned to some dedicated isolated CPUs. Detailing how to achieve this is out of scope of this document. You can refer to [CPU Manager for Kubernetes](https://github.com/intel/CPU-Manager-for-Kubernetes) that provides such functionality in Kubernetes. In the virtualized case, cpu pinning and isolation must be considered at the phyiscal layer as well as the virtual layer. + +2. All application resources(CPUs, devices and memory) are from same NUMA locality. In the virtualized case, NUMA locality is controlled by the underlying virtualized platform for the VM. + diff --git a/pkg/resources/vfioResource.go b/pkg/resources/vfioResource.go index 0541e0855..1369baf59 100644 --- a/pkg/resources/vfioResource.go +++ b/pkg/resources/vfioResource.go @@ -50,13 +50,13 @@ func (rp *vfioResource) GetDeviceSpecs(pciAddr string) []*pluginapi.DeviceSpec { Permissions: "mrw", }) - vfioDev, err := utils.GetVFIODeviceFile(pciAddr) + vfioDevHost, vfioDevContainer, err := utils.GetVFIODeviceFile(pciAddr) if err != nil { - glog.Errorf("GetDeviceSpecs(): error getting vfio device file for device: %s", pciAddr) + glog.Errorf("GetDeviceSpecs(): error getting vfio device file for device: %s, %s", pciAddr, err.Error()) } else { devSpecs = append(devSpecs, &pluginapi.DeviceSpec{ - HostPath: vfioDev, - ContainerPath: vfioDev, + HostPath: vfioDevHost, + ContainerPath: vfioDevContainer, Permissions: "mrw", }) } diff --git a/pkg/utils/utils.go b/pkg/utils/utils.go index 7c96ecbef..aa4168c9e 100644 --- a/pkg/utils/utils.go +++ b/pkg/utils/utils.go @@ -259,7 +259,7 @@ func ValidResourceName(name string) bool { } // GetVFIODeviceFile returns a vfio device files for vfio-pci bound PCI device's PCI address -func GetVFIODeviceFile(dev string) (devFile string, err error) { +func GetVFIODeviceFile(dev string) (devFileHost string, devFileContainer string, err error) { // Get iommu group for this device devPath := filepath.Join(sysBusPci, dev) _, err = os.Lstat(devPath) @@ -290,8 +290,23 @@ func GetVFIODeviceFile(dev string) (devFile string, err error) { err = fmt.Errorf("GetVFIODeviceFile(): error reading symlink to iommu_group %v", err) return } - - devFile = filepath.Join("/dev/vfio", filepath.Base(linkName)) + devFileContainer = filepath.Join("/dev/vfio", filepath.Base(linkName)) + devFileHost = devFileContainer + + // Get a file path to the iommu group name + namePath := filepath.Join(linkName, "name") + // Read the iommu group name + // The name file will not exist on baremetal + vfioName, errName := ioutil.ReadFile(namePath) + if errName == nil { + vName := strings.TrimSpace(string(vfioName)) + + // if the iommu group name == vfio-noiommu then we are in a VM, adjust path to vfio device + if vName == "vfio-noiommu" { + linkName = filepath.Join(filepath.Dir(linkName), "noiommu-"+filepath.Base(linkName)) + devFileHost = filepath.Join("/dev/vfio", filepath.Base(linkName)) + } + } return } diff --git a/pkg/utils/utils_test.go b/pkg/utils/utils_test.go index 2046fddde..a45c33662 100644 --- a/pkg/utils/utils_test.go +++ b/pkg/utils/utils_test.go @@ -296,8 +296,9 @@ var _ = Describe("In the utils package", func() { DescribeTable("getting VFIO device file", func(fs *FakeFilesystem, device, expected string, shouldFail bool) { defer fs.Use()() - actual, err := GetVFIODeviceFile(device) - Expect(actual).To(Equal(expected)) + //TODO: adapt test to running in a virtual environment + actualHost, _, err := GetVFIODeviceFile(device) + Expect(actualHost).To(Equal(expected)) assertShouldFail(err, shouldFail) }, Entry("could not get directory information for device",