3 Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
6 Redistribution and use in source and binary forms, with or without
7 modification, are permitted provided that the following conditions
10 * Redistributions of source code must retain the above copyright
11 notice, this list of conditions and the following disclaimer.
12 * Redistributions in binary form must reproduce the above copyright
13 notice, this list of conditions and the following disclaimer in
14 the documentation and/or other materials provided with the
16 * Neither the name of Intel Corporation nor the names of its
17 contributors may be used to endorse or promote products derived
18 from this software without specific prior written permission.
20 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
23 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
24 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
25 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
26 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
27 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
28 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
29 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
30 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
33 Vhost Sample Application
34 ========================
36 The vhost sample application demonstrates integration of the Data Plane Development Kit (DPDK)
37 with the Linux* KVM hypervisor by implementing the vhost-net offload API.
38 The sample application performs simple packet switching between virtual machines based on Media Access Control
39 (MAC) address or Virtual Local Area Network (VLAN) tag.
40 The splitting of Ethernet traffic from an external switch is performed in hardware by the Virtual Machine Device Queues
41 (VMDQ) and Data Center Bridging (DCB) features of the IntelĀ® 82599 10 Gigabit Ethernet Controller.
46 Virtio networking (virtio-net) was developed as the Linux* KVM para-virtualized method for communicating network packets
47 between host and guest.
48 It was found that virtio-net performance was poor due to context switching and packet copying between host, guest, and QEMU.
49 The following figure shows the system architecture for a virtio-based networking (virtio-net).
51 .. _figure_qemu_virtio_net:
53 .. figure:: img/qemu_virtio_net.*
55 System Architecture for Virtio-based Networking (virtio-net).
58 The Linux* Kernel vhost-net module was developed as an offload mechanism for virtio-net.
59 The vhost-net module enables KVM (QEMU) to offload the servicing of virtio-net devices to the vhost-net kernel module,
60 reducing the context switching and packet copies in the virtual dataplane.
62 This is achieved by QEMU sharing the following information with the vhost-net module through the vhost-net API:
64 * The layout of the guest memory space, to enable the vhost-net module to translate addresses.
66 * The locations of virtual queues in QEMU virtual address space,
67 to enable the vhost module to read/write directly to and from the virtqueues.
69 * An event file descriptor (eventfd) configured in KVM to send interrupts to the virtio- net device driver in the guest.
70 This enables the vhost-net module to notify (call) the guest.
72 * An eventfd configured in KVM to be triggered on writes to the virtio-net device's
73 Peripheral Component Interconnect (PCI) config space.
74 This enables the vhost-net module to receive notifications (kicks) from the guest.
76 The following figure shows the system architecture for virtio-net networking with vhost-net offload.
78 .. _figure_virtio_linux_vhost:
80 .. figure:: img/virtio_linux_vhost.*
88 The DPDK vhost-net sample code demonstrates KVM (QEMU) offloading the servicing of a Virtual Machine's (VM's)
89 virtio-net devices to a DPDK-based application in place of the kernel's vhost-net module.
91 The DPDK vhost-net sample code is based on vhost library. Vhost library is developed for user space Ethernet switch to
92 easily integrate with vhost functionality.
94 The vhost library implements the following features:
96 * Management of virtio-net device creation/destruction events.
98 * Mapping of the VM's physical memory into the DPDK vhost-net's address space.
100 * Triggering/receiving notifications to/from VMs via eventfds.
102 * A virtio-net back-end implementation providing a subset of virtio-net features.
104 There are two vhost implementations in vhost library, vhost cuse and vhost user. In vhost cuse, a character device driver is implemented to
105 receive and process vhost requests through ioctl messages. In vhost user, a socket server is created to received vhost requests through
106 socket messages. Most of the messages share the same handler routine.
109 **Any vhost cuse specific requirement in the following sections will be emphasized**.
111 Two implementations are turned on and off statically through configure file. Only one implementation could be turned on. They don't co-exist in current implementation.
113 The vhost sample code application is a simple packet switching application with the following feature:
115 * Packet switching between virtio-net devices and the network interface card,
116 including using VMDQs to reduce the switching that needs to be performed in software.
118 The following figure shows the architecture of the Vhost sample application based on vhost-cuse.
120 .. _figure_vhost_net_arch:
122 .. figure:: img/vhost_net_arch.*
124 Vhost-net Architectural Overview
127 The following figure shows the flow of packets through the vhost-net sample application.
129 .. _figure_vhost_net_sample_app:
131 .. figure:: img/vhost_net_sample_app.*
133 Packet Flow Through the vhost-net Sample Application
136 Supported Distributions
137 -----------------------
139 The example in this section have been validated with the following distributions:
150 This section lists prerequisite packages that must be installed.
152 Installing Packages on the Host(vhost cuse required)
153 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
155 The vhost cuse code uses the following packages; fuse, fuse-devel, and kernel-modules-extra.
156 The vhost user code don't rely on those modules as eventfds are already installed into vhost process through
159 #. Install Fuse Development Libraries and headers:
161 .. code-block:: console
163 yum -y install fuse fuse-devel
165 #. Install the Cuse Kernel Module:
167 .. code-block:: console
169 yum -y install kernel-modules-extra
174 For vhost user, qemu 2.2 is required.
176 Setting up the Execution Environment
177 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
179 The vhost sample code requires that QEMU allocates a VM's memory on the hugetlbfs file system.
180 As the vhost sample code requires hugepages,
181 the best practice is to partition the system into separate hugepage mount points for the VMs and the vhost sample code.
185 This is best-practice only and is not mandatory.
186 For systems that only support 2 MB page sizes,
187 both QEMU and vhost sample code can use the same hugetlbfs mount point without issue.
191 VMs with gigabytes of memory can benefit from having QEMU allocate their memory from 1 GB huge pages.
192 1 GB huge pages must be allocated at boot time by passing kernel parameters through the grub boot loader.
194 #. Calculate the maximum memory usage of all VMs to be run on the system.
195 Then, round this value up to the nearest Gigabyte the execution environment will require.
197 #. Edit the /etc/default/grub file, and add the following to the GRUB_CMDLINE_LINUX entry:
199 .. code-block:: console
201 GRUB_CMDLINE_LINUX="... hugepagesz=1G hugepages=<Number of hugepages required> default_hugepagesz=1G"
203 #. Update the grub boot loader:
205 .. code-block:: console
207 grub2-mkconfig -o /boot/grub2/grub.cfg
209 #. Reboot the system.
211 #. The hugetlbfs mount point (/dev/hugepages) should now default to allocating gigabyte pages.
215 Making the above modification will change the system default hugepage size to 1 GB for all applications.
217 **Vhost Sample Code**
219 In this section, we create a second hugetlbs mount point to allocate hugepages for the DPDK vhost sample code.
221 #. Allocate sufficient 2 MB pages for the DPDK vhost sample code:
223 .. code-block:: console
225 echo 256 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages
227 #. Mount hugetlbs at a separate mount point for 2 MB pages:
229 .. code-block:: console
231 mount -t hugetlbfs nodev /mnt/huge -o pagesize=2M
233 The above steps can be automated by doing the following:
235 #. Edit /etc/fstab to add an entry to automatically mount the second hugetlbfs mount point:
239 hugetlbfs <tab> /mnt/huge <tab> hugetlbfs defaults,pagesize=1G 0 0
241 #. Edit the /etc/default/grub file, and add the following to the GRUB_CMDLINE_LINUX entry:
245 GRUB_CMDLINE_LINUX="... hugepagesz=2M hugepages=256 ... default_hugepagesz=1G"
247 #. Update the grub bootloader:
249 .. code-block:: console
251 grub2-mkconfig -o /boot/grub2/grub.cfg
253 #. Reboot the system.
257 Ensure that the default hugepage size after this setup is 1 GB.
259 Setting up the Guest Execution Environment
260 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
262 It is recommended for testing purposes that the DPDK testpmd sample application is used in the guest to forward packets,
263 the reasons for this are discussed in Section 22.7, "Running the Virtual Machine (QEMU)".
265 The testpmd application forwards packets between pairs of Ethernet devices,
266 it requires an even number of Ethernet devices (virtio or otherwise) to execute.
267 It is therefore recommended to create multiples of two virtio-net devices for each Virtual Machine either through libvirt or
268 at the command line as follows.
272 Observe that in the example, "-device" and "-netdev" are repeated for two virtio-net devices.
276 .. code-block:: console
278 user@target:~$ qemu-system-x86_64 ... \
279 -netdev tap,id=hostnet1,vhost=on,vhostfd=<open fd> \
280 -device virtio-net-pci, netdev=hostnet1,id=net1 \
281 -netdev tap,id=hostnet2,vhost=on,vhostfd=<open fd> \
282 -device virtio-net-pci, netdev=hostnet2,id=net1
286 .. code-block:: console
288 user@target:~$ qemu-system-x86_64 ... \
289 -chardev socket,id=char1,path=<sock_path> \
290 -netdev type=vhost-user,id=hostnet1,chardev=char1 \
291 -device virtio-net-pci,netdev=hostnet1,id=net1 \
292 -chardev socket,id=char2,path=<sock_path> \
293 -netdev type=vhost-user,id=hostnet2,chardev=char2 \
294 -device virtio-net-pci,netdev=hostnet2,id=net2
296 sock_path is the path for the socket file created by vhost.
298 Compiling the Sample Code
299 -------------------------
300 #. Compile vhost lib:
302 To enable vhost, turn on vhost library in the configure file config/common_linuxapp.
304 .. code-block:: console
306 CONFIG_RTE_LIBRTE_VHOST=n
308 vhost user is turned on by default in the configure file config/common_linuxapp.
309 To enable vhost cuse, disable vhost user.
311 .. code-block:: console
313 CONFIG_RTE_LIBRTE_VHOST_USER=y
315 After vhost is enabled and the implementation is selected, build the vhost library.
317 #. Go to the examples directory:
319 .. code-block:: console
321 export RTE_SDK=/path/to/rte_sdk
322 cd ${RTE_SDK}/examples/vhost
324 #. Set the target (a default target is used if not specified). For example:
326 .. code-block:: console
328 export RTE_TARGET=x86_64-native-linuxapp-gcc
330 See the DPDK Getting Started Guide for possible RTE_TARGET values.
332 #. Build the application:
334 .. code-block:: console
337 make config ${RTE_TARGET}
338 make install ${RTE_TARGET}
339 cd ${RTE_SDK}/examples/vhost
342 #. Go to the eventfd_link directory(vhost cuse required):
344 .. code-block:: console
346 cd ${RTE_SDK}/lib/librte_vhost/eventfd_link
348 #. Build the eventfd_link kernel module(vhost cuse required):
350 .. code-block:: console
354 Running the Sample Code
355 -----------------------
357 #. Install the cuse kernel module(vhost cuse required):
359 .. code-block:: console
363 #. Go to the eventfd_link directory(vhost cuse required):
365 .. code-block:: console
367 export RTE_SDK=/path/to/rte_sdk
368 cd ${RTE_SDK}/lib/librte_vhost/eventfd_link
370 #. Install the eventfd_link module(vhost cuse required):
372 .. code-block:: console
374 insmod ./eventfd_link.ko
376 #. Go to the examples directory:
378 .. code-block:: console
380 export RTE_SDK=/path/to/rte_sdk
381 cd ${RTE_SDK}/examples/vhost
383 #. Run the vhost-switch sample code:
387 .. code-block:: console
389 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir / mnt/huge -- -p 0x1 --dev-basename usvhost --dev-index 1
391 vhost user: a socket file named usvhost will be created under current directory. Use its path as the socket path in guest's qemu commandline.
393 .. code-block:: console
395 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir / mnt/huge -- -p 0x1 --dev-basename usvhost
399 Please note the huge-dir parameter instructs the DPDK to allocate its memory from the 2 MB page hugetlbfs.
404 **Basename and Index.**
405 vhost cuse uses a Linux* character device to communicate with QEMU.
406 The basename and the index are used to generate the character devices name.
408 /dev/<basename>-<index>
410 The index parameter is provided for a situation where multiple instances of the virtual switch is required.
412 For compatibility with the QEMU wrapper script, a base name of "usvhost" and an index of "1" should be used:
414 .. code-block:: console
416 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir / mnt/huge -- -p 0x1 --dev-basename usvhost --dev-index 1
419 The vm2vm parameter disable/set mode of packet switching between guests in the host.
420 Value of "0" means disabling vm2vm implies that on virtual machine packet transmission will always go to the Ethernet port;
421 Value of "1" means software mode packet forwarding between guests, it needs packets copy in vHOST,
422 so valid only in one-copy implementation, and invalid for zero copy implementation;
423 value of "2" means hardware mode packet forwarding between guests, it allows packets go to the Ethernet port,
424 hardware L2 switch will determine which guest the packet should forward to or need send to external,
425 which bases on the packet destination MAC address and VLAN tag.
427 .. code-block:: console
429 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir /mnt/huge -- --vm2vm [0,1,2]
431 **Mergeable Buffers.**
432 The mergeable buffers parameter controls how virtio-net descriptors are used for virtio-net headers.
433 In a disabled state, one virtio-net header is used per packet buffer;
434 in an enabled state one virtio-net header is used for multiple packets.
435 The default value is 0 or disabled since recent kernels virtio-net drivers show performance degradation with this feature is enabled.
437 .. code-block:: console
439 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir / mnt/huge -- --mergeable [0,1]
442 The stats parameter controls the printing of virtio-net device statistics.
443 The parameter specifies an interval second to print statistics, with an interval of 0 seconds disabling statistics.
445 .. code-block:: console
447 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir / mnt/huge -- --stats [0,n]
450 The rx-retry option enables/disables enqueue retries when the guests RX queue is full.
451 This feature resolves a packet loss that is observed at high data-rates,
452 by allowing it to delay and retry in the receive path.
453 This option is enabled by default.
455 .. code-block:: console
457 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir / mnt/huge -- --rx-retry [0,1]
460 The rx-retry-num option specifies the number of retries on an RX burst,
461 it takes effect only when rx retry is enabled.
462 The default value is 4.
464 .. code-block:: console
466 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir / mnt/huge -- --rx-retry 1 --rx-retry-num 5
468 **RX Retry Delay Time.**
469 The rx-retry-delay option specifies the timeout (in micro seconds) between retries on an RX burst,
470 it takes effect only when rx retry is enabled.
471 The default value is 15.
473 .. code-block:: console
475 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir / mnt/huge -- --rx-retry 1 --rx-retry-delay 20
478 The zero copy option enables/disables the zero copy mode for RX/TX packet,
479 in the zero copy mode the packet buffer address from guest translate into host physical address
480 and then set directly as DMA address.
481 If the zero copy mode is disabled, then one copy mode is utilized in the sample.
482 This option is disabled by default.
484 .. code-block:: console
486 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir /mnt/huge -- --zero-copy [0,1]
488 **RX descriptor number.**
489 The RX descriptor number option specify the Ethernet RX descriptor number,
490 Linux legacy virtio-net has different behavior in how to use the vring descriptor from DPDK based virtio-net PMD,
491 the former likely allocate half for virtio header, another half for frame buffer,
492 while the latter allocate all for frame buffer,
493 this lead to different number for available frame buffer in vring,
494 and then lead to different Ethernet RX descriptor number could be used in zero copy mode.
495 So it is valid only in zero copy mode is enabled. The value is 32 by default.
497 .. code-block:: console
499 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir /mnt/huge -- --zero-copy 1 --rx-desc-num [0, n]
501 **TX descriptor number.**
502 The TX descriptor number option specify the Ethernet TX descriptor number, it is valid only in zero copy mode is enabled.
503 The value is 64 by default.
505 .. code-block:: console
507 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir /mnt/huge -- --zero-copy 1 --tx-desc-num [0, n]
510 The VLAN strip option enable/disable the VLAN strip on host, if disabled, the guest will receive the packets with VLAN tag.
511 It is enabled by default.
513 .. code-block:: console
515 user@target:~$ ./build/app/vhost-switch -c f -n 4 --huge-dir /mnt/huge -- --vlan-strip [0, 1]
517 Running the Virtual Machine (QEMU)
518 ----------------------------------
520 QEMU must be executed with specific parameters to:
522 * Ensure the guest is configured to use virtio-net network adapters.
524 .. code-block:: console
526 user@target:~$ qemu-system-x86_64 ... -device virtio-net-pci,netdev=hostnet1,id=net1 ...
528 * Ensure the guest's virtio-net network adapter is configured with offloads disabled.
530 .. code-block:: console
532 user@target:~$ qemu-system-x86_64 ... -device virtio-net-pci,netdev=hostnet1,id=net1,csum=off,gso=off,guest_tso4=off,guest_tso6=off,guest_ecn=off
534 * Redirect QEMU to communicate with the DPDK vhost-net sample code in place of the vhost-net kernel module(vhost cuse).
536 .. code-block:: console
538 user@target:~$ qemu-system-x86_64 ... -netdev tap,id=hostnet1,vhost=on,vhostfd=<open fd> ...
540 * Enable the vhost-net sample code to map the VM's memory into its own process address space.
542 .. code-block:: console
544 user@target:~$ qemu-system-x86_64 ... -mem-prealloc -mem-path / dev/hugepages ...
548 The QEMU wrapper (qemu-wrap.py) is a Python script designed to automate the QEMU configuration described above.
549 It also facilitates integration with libvirt, although the script may also be used standalone without libvirt.
551 Redirecting QEMU to vhost-net Sample Code(vhost cuse)
552 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
554 To redirect QEMU to the vhost-net sample code implementation of the vhost-net API,
555 an open file descriptor must be passed to QEMU running as a child process.
557 .. code-block:: python
560 fd = os.open("/dev/usvhost-1", os.O_RDWR)
561 subprocess.call("qemu-system-x86_64 ... . -netdev tap,id=vhostnet0,vhost=on,vhostfd=" + fd +"...", shell=True)
565 This process is automated in the QEMU wrapper script discussed in Section 24.7.3.
567 Mapping the Virtual Machine's Memory
568 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
570 For the DPDK vhost-net sample code to be run correctly, QEMU must allocate the VM's memory on hugetlbfs.
571 This is done by specifying mem-prealloc and mem-path when executing QEMU.
572 The vhost-net sample code accesses the virtio-net device's virtual rings and packet buffers
573 by finding and mapping the VM's physical memory on hugetlbfs.
574 In this case, the path passed to the guest should be that of the 1 GB page hugetlbfs:
576 .. code-block:: console
578 user@target:~$ qemu-system-x86_64 ... -mem-prealloc -mem-path / dev/hugepages ...
582 This process is automated in the QEMU wrapper script discussed in Section 24.7.3.
583 The following two sections only applies to vhost cuse. For vhost-user, please make corresponding changes to qemu-wrapper script and guest XML file.
588 The QEMU wrapper script automatically detects and calls QEMU with the necessary parameters required
589 to integrate with the vhost sample code.
590 It performs the following actions:
592 * Automatically detects the location of the hugetlbfs and inserts this into the command line parameters.
594 * Automatically open file descriptors for each virtio-net device and inserts this into the command line parameters.
596 * Disables offloads on each virtio-net device.
598 * Calls Qemu passing both the command line parameters passed to the script itself and those it has auto-detected.
600 The QEMU wrapper script will automatically configure calls to QEMU:
602 .. code-block:: console
604 user@target:~$ qemu-wrap.py -machine pc-i440fx-1.4,accel=kvm,usb=off -cpu SandyBridge -smp 4,sockets=4,cores=1,threads=1
605 -netdev tap,id=hostnet1,vhost=on -device virtio-net-pci,netdev=hostnet1,id=net1 -hda <disk img> -m 4096
607 which will become the following call to QEMU:
609 .. code-block:: console
611 /usr/local/bin/qemu-system-x86_64 -machine pc-i440fx-1.4,accel=kvm,usb=off -cpu SandyBridge -smp 4,sockets=4,cores=1,threads=1
612 -netdev tap,id=hostnet1,vhost=on,vhostfd=<open fd> -device virtio-net-pci,netdev=hostnet1,id=net1,
613 csum=off,gso=off,guest_tso4=off,guest_tso6=off,guest_ecn=off -hda <disk img> -m 4096 -mem-path /dev/hugepages -mem-prealloc
618 The QEMU wrapper script (qemu-wrap.py) "wraps" libvirt calls to QEMU,
619 such that QEMU is called with the correct parameters described above.
620 To call the QEMU wrapper automatically from libvirt, the following configuration changes must be made:
622 * Place the QEMU wrapper script in libvirt's binary search PATH ($PATH).
623 A good location is in the directory that contains the QEMU binary.
625 * Ensure that the script has the same owner/group and file permissions as the QEMU binary.
627 * Update the VM xml file using virsh edit <vm name>:
629 * Set the VM to use the launch script
631 * Set the emulator path contained in the #<emulator><emulator/> tags For example,
632 replace <emulator>/usr/bin/qemu-kvm<emulator/> with <emulator>/usr/bin/qemu-wrap.py<emulator/>
634 * Set the VM's virtio-net device's to use vhost-net offload:
638 <interface type="network">
639 <model type="virtio"/>
640 <driver name="vhost"/>
643 * Enable libvirt to access the DPDK Vhost sample code's character device file by adding it
644 to controllers cgroup for libvirtd using the following steps:
648 cgroup_controllers = [ ... "devices", ... ] clear_emulator_capabilities = 0
649 user = "root" group = "root"
650 cgroup_device_acl = [
651 "/dev/null", "/dev/full", "/dev/zero",
652 "/dev/random", "/dev/urandom",
653 "/dev/ptmx", "/dev/kvm", "/dev/kqemu",
654 "/dev/rtc", "/dev/hpet", "/dev/net/tun",
655 "/dev/<devbase-name>-<index>",
658 * Disable SELinux or set to permissive mode.
661 * Mount cgroup device controller:
663 .. code-block:: console
665 user@target:~$ mkdir /dev/cgroup
666 user@target:~$ mount -t cgroup none /dev/cgroup -o devices
668 * Restart the libvirtd system process
670 For example, on Fedora* "systemctl restart libvirtd.service"
672 * Edit the configuration parameters section of the script:
674 * Configure the "emul_path" variable to point to the QEMU emulator.
678 emul_path = "/usr/local/bin/qemu-system-x86_64"
680 * Configure the "us_vhost_path" variable to point to the DPDK vhost-net sample code's character devices name.
681 DPDK vhost-net sample code's character device will be in the format "/dev/<basename>-<index>".
685 us_vhost_path = "/dev/usvhost-1"
690 * QEMU failing to allocate memory on hugetlbfs, with an error like the following::
692 file_ram_alloc: can't mmap RAM pages: Cannot allocate memory
694 When running QEMU the above error indicates that it has failed to allocate memory for the Virtual Machine on
695 the hugetlbfs. This is typically due to insufficient hugepages being free to support the allocation request.
696 The number of free hugepages can be checked as follows:
698 .. code-block:: console
700 cat /sys/kernel/mm/hugepages/hugepages-<pagesize>/nr_hugepages
702 The command above indicates how many hugepages are free to support QEMU's allocation request.
704 * User space VHOST when the guest has 2MB sized huge pages:
706 The guest may have 2MB or 1GB sized huge pages. The user space VHOST should work properly in both cases.
708 * User space VHOST will not work with QEMU without the ``-mem-prealloc`` option:
710 The current implementation works properly only when the guest memory is pre-allocated, so it is required to
711 use a QEMU version (e.g. 1.6) which supports ``-mem-prealloc``. The ``-mem-prealloc`` option must be
712 specified explicitly in the QEMU command line.
714 * User space VHOST will not work with a QEMU version without shared memory mapping:
716 As shared memory mapping is mandatory for user space VHOST to work properly with the guest, user space VHOST
717 needs access to the shared memory from the guest to receive and transmit packets. It is important to make sure
718 the QEMU version supports shared memory mapping.
720 * Issues with ``virsh destroy`` not destroying the VM:
722 Using libvirt ``virsh create`` the ``qemu-wrap.py`` spawns a new process to run ``qemu-kvm``. This impacts the behavior
723 of ``virsh destroy`` which kills the process running ``qemu-wrap.py`` without actually destroying the VM (it leaves
724 the ``qemu-kvm`` process running):
726 This following patch should fix this issue:
727 http://dpdk.org/ml/archives/dev/2014-June/003607.html
729 * In an Ubuntu environment, QEMU fails to start a new guest normally with user space VHOST due to not being able
730 to allocate huge pages for the new guest:
732 The solution for this issue is to add ``-boot c`` into the QEMU command line to make sure the huge pages are
733 allocated properly and then the guest should start normally.
735 Use ``cat /proc/meminfo`` to check if there is any changes in the value of ``HugePages_Total`` and ``HugePages_Free``
736 after the guest startup.
738 * Log message: ``eventfd_link: module verification failed: signature and/or required key missing - tainting kernel``:
740 This log message may be ignored. The message occurs due to the kernel module ``eventfd_link``, which is not a standard
741 Linux module but which is necessary for the user space VHOST current implementation (CUSE-based) to communicate with
745 Running DPDK in the Virtual Machine
746 -----------------------------------
748 For the DPDK vhost-net sample code to switch packets into the VM,
749 the sample code must first learn the MAC address of the VM's virtio-net device.
750 The sample code detects the address from packets being transmitted from the VM, similar to a learning switch.
752 This behavior requires no special action or configuration with the Linux* virtio-net driver in the VM
753 as the Linux* Kernel will automatically transmit packets during device initialization.
754 However, DPDK-based applications must be modified to automatically transmit packets during initialization
755 to facilitate the DPDK vhost- net sample code's MAC learning.
757 The DPDK testpmd application can be configured to automatically transmit packets during initialization
758 and to act as an L2 forwarding switch.
760 Testpmd MAC Forwarding
761 ~~~~~~~~~~~~~~~~~~~~~~
763 At high packet rates, a minor packet loss may be observed.
764 To resolve this issue, a "wait and retry" mode is implemented in the testpmd and vhost sample code.
765 In the "wait and retry" mode if the virtqueue is found to be full, then testpmd waits for a period of time before retrying to enqueue packets.
767 The "wait and retry" algorithm is implemented in DPDK testpmd as a forwarding method call "mac_retry".
768 The following sequence diagram describes the algorithm in detail.
770 .. _figure_tx_dpdk_testpmd:
772 .. figure:: img/tx_dpdk_testpmd.*
774 Packet Flow on TX in DPDK-testpmd
780 The testpmd application is automatically built when DPDK is installed.
781 Run the testpmd application as follows:
783 .. code-block:: console
785 user@target:~$ x86_64-native-linuxapp-gcc/app/testpmd -c 0x3 -- n 4 -socket-mem 128 -- --burst=64 -i
787 The destination MAC address for packets transmitted on each port can be set at the command line:
789 .. code-block:: console
791 user@target:~$ x86_64-native-linuxapp-gcc/app/testpmd -c 0x3 -- n 4 -socket-mem 128 -- --burst=64 -i --eth- peer=0,aa:bb:cc:dd:ee:ff --eth-peer=1,ff,ee,dd,cc,bb,aa
793 * Packets received on port 1 will be forwarded on port 0 to MAC address
797 * Packets received on port 0 will be forwarded on port 1 to MAC address
801 The testpmd application can then be configured to act as an L2 forwarding application:
803 .. code-block:: console
805 testpmd> set fwd mac_retry
807 The testpmd can then be configured to start processing packets,
808 transmitting packets first so the DPDK vhost sample code on the host can learn the MAC address:
810 .. code-block:: console
812 testpmd> start tx_first
816 Please note "set fwd mac_retry" is used in place of "set fwd mac_fwd" to ensure the retry feature is activated.
818 Passing Traffic to the Virtual Machine Device
819 ---------------------------------------------
821 For a virtio-net device to receive traffic,
822 the traffic's Layer 2 header must include both the virtio-net device's MAC address and VLAN tag.
823 The DPDK sample code behaves in a similar manner to a learning switch in that
824 it learns the MAC address of the virtio-net devices from the first transmitted packet.
825 On learning the MAC address,
826 the DPDK vhost sample code prints a message with the MAC address and VLAN tag virtio-net device.
829 .. code-block:: console
831 DATA: (0) MAC_ADDRESS cc:bb:bb:bb:bb:bb and VLAN_TAG 1000 registered
833 The above message indicates that device 0 has been registered with MAC address cc:bb:bb:bb:bb:bb and VLAN tag 1000.
834 Any packets received on the NIC with these values is placed on the devices receive queue.
835 When a virtio-net device transmits packets, the VLAN tag is added to the packet by the DPDK vhost sample code.