File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.158: download - view: text, annotated - select for diffs
Wed Sep 5 09:16:10 2018 UTC (15 months, 1 week ago) by maxv
Branches: MAIN
CVS tags: HEAD
noise reduction police, and clarify a few things

    1: [[!meta title="Xen HowTo"]]
    2: 
    3: Xen is a Type 1 hypervisor which supports running multiple guest operating
    4: systems on a single physical machine. One uses the Xen kernel to control the
    5: CPU, memory and console, a dom0 operating system which mediates access to
    6: other hardware (e.g., disks, network, USB), and one or more domU operating
    7: systems which operate in an unprivileged virtualized environment. IO requests
    8: from the domU systems are forwarded by the Xen hypervisor to the dom0 to be
    9: fulfilled.
   10: 
   11: Xen supports different styles of guest:
   12: 
   13: [[!table data="""
   14: Style of guest	|Supported by NetBSD
   15: PV		|Yes (dom0, domU)
   16: HVM		|Yes (domU)
   17: PVHVM		|No
   18: PVH		|No
   19: """]]
   20: 
   21: In Para-Virtualized (PV) mode, the guest OS does not attempt to access
   22: hardware directly, but instead makes hypercalls to the hypervisor; PV
   23: guests must be specifically coded for Xen. In HVM mode, no guest
   24: modification is required; however, hardware support is required, such
   25: as VT-x on Intel CPUs and SVM on AMD CPUs.
   26: 
   27: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
   28: The dom0 can start one or more domUs.  (Booting is explained in detail
   29: in the dom0 section.)
   30: 
   31: This HOWTO presumes a basic familiarity with the Xen system
   32: architecture, with installing NetBSD on i386/amd64 hardware, and with
   33: installing software from pkgsrc.  See also the [Xen
   34: website](http://www.xenproject.org/).
   35: 
   36: This HOWTO attempts to address both the case of running a NetBSD dom0
   37: on hardware and running domUs under it (NetBSD and other), and also
   38: running NetBSD as a domU in a VPS.
   39: 
   40: Versions and Support
   41: ====================
   42: 
   43: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
   44: xenkernel and xentools.  We will refer only to the kernel versions,
   45: but note that both packages must be installed together and must have
   46: matching versions.
   47: 
   48: Versions available in pkgsrc:
   49: 
   50: [[!table data="""
   51: Xen Version	|Package Name	|Xen CPU Support	|EOL'ed By Upstream
   52: 4.2		|xenkernel42	|32bit, 64bit		|Yes
   53: 4.5		|xenkernel45	|64bit			|Yes
   54: 4.6		|xenkernel46	|64bit			|Partially
   55: 4.8		|xenkernel48	|64bit			|No
   56: 4.11		|xenkernel411	|64bit			|No
   57: """]]
   58: 
   59: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
   60: 
   61: Multiprocessor (SMP) support in NetBSD differs depending on the domain:
   62: 
   63: [[!table data="""
   64: Domain		|Supports SMP
   65: dom0		|No
   66: domU		|Yes
   67: """]]
   68: 
   69: Note: NetBSD support is called XEN3. However, it does support Xen 4,
   70: because the hypercall interface has remained identical.
   71: 
   72: Architecture
   73: ------------
   74: 
   75: Xen itself runs on x86_64 hardware.
   76: 
   77: The dom0 system, plus each domU, can be either i386PAE or amd64.
   78: i386 without PAE is not supported.
   79: 
   80: The standard approach is to use NetBSD/amd64 for the dom0.
   81: 
   82: To use an i386PAE dom0, one must build or obtain a 64bit Xen kernel and
   83: install it on the system.
   84: 
   85: For domUs, i386PAE is considered as
   86: [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html)
   87: than amd64.
   88: 
   89: NetBSD as a dom0
   90: ================
   91: 
   92: In order to install a NetBSD as a dom0, one must first install a normal
   93: NetBSD system, and then pivot the install to a dom0 install by changing
   94: the kernel and boot configuration.
   95: 
   96: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
   97: 512M it worked reliably.  This does not make sense, but if you see
   98: "not ELF" after Xen boots, try increasing dom0 RAM.
   99: 
  100: Installation of NetBSD
  101: ----------------------
  102: 
  103: [Install NetBSD/amd64](/guide/inst/)
  104: just as you would if you were not using Xen.
  105: 
  106: Installation of Xen
  107: -------------------
  108: 
  109: We will consider that you chose to use Xen 4.8, with NetBSD/amd64 as
  110: dom0. In the dom0, install xenkernel48 and xentools48 from pkgsrc.
  111: 
  112: Once this is done, install the Xen kernel itself:
  113: 
  114: [[!template id=programlisting text="""
  115: # cp /usr/pkg/xen48-kernel/xen.gz /
  116: """]]
  117: 
  118: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory, copied from
  119: `releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz` of a NetBSD build.
  120: 
  121: Add a line to /boot.cfg to boot Xen:
  122: 
  123: [[!template id=filecontent name="/boot.cfg" text="""
  124: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
  125: """]]
  126: 
  127: which specifies that the dom0 should have 512MB of ram, leaving the rest
  128: to be allocated for domUs.  To use a serial console, use
  129: 
  130: [[!template id=filecontent name="/boot.cfg" text="""
  131: menu=Xen:load /netbsd-XEN3_DOM0.gz;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
  132: """]]
  133: 
  134: which will use the first serial port for Xen (which counts starting
  135: from 1, unlike NetBSD which counts starting from 0), forcing
  136: speed/parity.  Because the NetBSD command line lacks a
  137: "console=pc" argument, it will use the default "xencons" console device,
  138: which directs the console I/O through Xen to the same console device Xen
  139: itself uses (in this case, the serial port).
  140: 
  141: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`,
  142: to force only one vcpu to be provided (since NetBSD dom0 can't use
  143: more) and to pin that vcpu to a physical CPU. Xen has
  144: [many boot options](http://xenbits.xenproject.org/docs/4.8-testing/misc/xen-command-line.html),
  145: and other than dom0 memory and max_vcpus, they are generally not
  146: necessary.
  147: 
  148: Enable `xencommons`:
  149: 
  150: [[!template id=filecontent name="/etc/rc.conf" text="""
  151: xencommons=YES
  152: """]]
  153: 
  154: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  155: than GENERIC without Xen.
  156: 
  157: TODO: Recommend for/against xen-watchdog.
  158: 
  159: Once the reboot is done, use `xl` to inspect Xen's boot messages,
  160: available resources, and running domains.  For example:
  161: 
  162: [[!template id=programlisting text="""
  163: # xl dmesg
  164: ... xen's boot info ...
  165: # xl info
  166: ... available memory, etc ...
  167: # xl list
  168: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  169: Domain-0           0       64    0  r----     58.1
  170: """]]
  171: 
  172: Xen logs will be in /var/log/xen.
  173: 
  174: ### Issues with xencommons
  175: 
  176: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
  177: domUs.  It does not currently work to stop and start xenstored.
  178: Certainly all domUs should be shutdown first, following the sort order
  179: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
  180: and is not notified when xenstored exits, leading to not recreating
  181: the state when the new xenstored starts.  Until there's a mechanism to
  182: make this work, one should not expect to be able to restart xenstored
  183: (and thus xencommons).  There is currently no reason to expect that
  184: this will get fixed any time soon.
  185: 
  186: anita (for testing NetBSD)
  187: --------------------------
  188: 
  189: With the setup so far (assuming 4.8/xl), one should be able to run
  190: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
  191: root, because anita must create a domU):
  192: 
  193: [[!template id=programlisting text="""
  194: anita --vmm=xl test file:///usr/obj/i386/
  195: """]]
  196: 
  197: Xen-specific NetBSD issues
  198: --------------------------
  199: 
  200: There are (at least) two additional things different about NetBSD as a
  201: dom0 kernel compared to hardware.
  202: 
  203: One is that the module ABI is different because some of the #defines
  204: change, so one must build modules for Xen.  As of netbsd-7, the build
  205: system does this automatically.
  206: 
  207: The other difference is that XEN3_DOM0 does not have exactly the same
  208: options as GENERIC.  While it is debatable whether or not this is a
  209: bug, users should be aware of this and can simply add missing config
  210: items if desired.
  211: 
  212: Updating NetBSD in a dom0
  213: -------------------------
  214: 
  215: This is just like updating NetBSD on bare hardware, assuming the new
  216: version supports the version of Xen you are running.  Generally, one
  217: replaces the kernel and reboots, and then overlays userland binaries
  218: and adjusts `/etc`.
  219: 
  220: Note that one must update both the non-Xen kernel typically used for
  221: rescue purposes and the DOM0 kernel used with Xen.
  222: 
  223: Converting from grub to /boot
  224: -----------------------------
  225: 
  226: These instructions were used to convert a system from
  227: grub to /boot.  The system was originally installed in February of
  228: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
  229: over time.  Before these commands, it was running NetBSD 6 i386, Xen
  230: 4.1 and grub, much like the message linked earlier in the grub
  231: section.
  232: 
  233: [[!template id=programlisting text="""
  234: # Install MBR bootblocks on both disks.
  235: fdisk -i /dev/rwd0d
  236: fdisk -i /dev/rwd1d
  237: # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
  238: installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
  239: installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
  240: # Install secondary boot loader
  241: cp -p /usr/mdec/boot /
  242: # Create boot.cfg following earlier guidance:
  243: menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
  244: menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=512M
  245: menu=GENERIC:boot
  246: menu=GENERIC single-user:boot -s
  247: menu=GENERIC.ok:boot netbsd.ok
  248: menu=GENERIC.ok single-user:boot netbsd.ok -s
  249: menu=Drop to boot prompt:prompt
  250: default=1
  251: timeout=30
  252: """]]
  253: 
  254: Upgrading Xen versions
  255: ---------------------
  256: 
  257: Minor version upgrades are trivial.  Just rebuild/replace the
  258: xenkernel version and copy the new xen.gz to `/` (where `/boot.cfg`
  259: references it), and reboot.
  260: 
  261: Unprivileged domains (domU)
  262: ===========================
  263: 
  264: This section describes general concepts about domUs.  It does not
  265: address specific domU operating systems or how to install them.  The
  266: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
  267: typically named so that the file name, domU name and the domU's host
  268: name match.
  269: 
  270: The domU is provided with CPU and memory by Xen, configured by the
  271: dom0.  The domU is provided with disk and network by the dom0,
  272: mediated by Xen, and configured in the dom0.
  273: 
  274: Entropy in domUs can be an issue; physical disks and network are on
  275: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  276: 
  277: Config files
  278: ------------
  279: 
  280: See /usr/pkg/share/examples/xen/xlexample*,
  281: for a large number of well-commented examples, mostly for running
  282: GNU/Linux.
  283: 
  284: The following is an example minimal domain configuration file. The domU
  285: serves as a network file server.
  286: 
  287: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  288: name = "domU-id"
  289: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  290: memory = 1024
  291: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  292: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  293:          'file:/n0/xen/foo-wd1,0x1,w' ]
  294: """]]
  295: 
  296: The domain will have name given in the `name` setting.  The kernel has the
  297: host/domU name in it, so that on the dom0 one can update the various
  298: domUs independently.  The `vif` line causes an interface to be provided,
  299: with a specific mac address (do not reuse MAC addresses!), in bridge
  300: mode.  Two disks are provided, and they are both writable; the bits
  301: are stored in files and Xen attaches them to a vnd(4) device in the
  302: dom0 on domain creation.  The system treats xbd0 as the boot device
  303: without needing explicit configuration.
  304: 
  305: By default, `xl` looks for domain config files in `/usr/pkg/etc/xen`.  Note
  306: that "xl create" takes the name of a config file, while other commands
  307: take the name of a domain.
  308: 
  309: Examples of commands:
  310: 
  311: [[!template id=programlisting text="""
  312: xl create foo
  313: xl console foo
  314: xl create -c foo
  315: xl shutdown foo
  316: xl list
  317: """]]
  318: 
  319: Typing `^]` will exit the console session.  Shutting down a domain is
  320: equivalent to pushing the power button; a NetBSD domU will receive a
  321: power-press event and do a clean shutdown.  Shutting down the dom0
  322: will trigger controlled shutdowns of all configured domUs.
  323: 
  324: CPU and memory
  325: --------------
  326: 
  327: A domain is provided with some number of vcpus, less than the number
  328: of CPUs seen by the hypervisor. For a domU, it is controlled
  329: from the config file by the "vcpus = N" directive.
  330: 
  331: A domain is provided with memory; this is controlled in the config
  332: file by "memory = N" (in megabytes).  In the straightforward case, the
  333: sum of the the memory allocated to the dom0 and all domUs must be less
  334: than the available memory.
  335: 
  336: Xen also provides a "balloon" driver, which can be used to let domains
  337: use more memory temporarily.
  338: 
  339: Virtual disks
  340: -------------
  341: 
  342: In domU config files, the disks are defined as a sequence of 3-tuples:
  343: 
  344:  * The first element is "method:/path/to/disk". Common methods are
  345:    "file:" for file-backed vnd, and "phy:" for something that is already
  346:    a device.
  347: 
  348:  * The second element is an artifact of how virtual disks are passed to
  349:    Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  350:    are given a device name to associate with the disk, and values like
  351:    "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  352:    as xbd0, the second as xbd1, and so on.  However, xl demands a
  353:    second argument.  The name given is converted to a major/minor by
  354:    calling stat(2) on the name in /dev and this is passed to the domU.
  355:    In the general case, the dom0 and domU can be different operating
  356:    systems, and it is an unwarranted assumption that they have consistent
  357:    numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  358:    as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  359:    for the second works fine and avoids this issue.  For a GNU/Linux
  360:    guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  361:    /dev/hda1.
  362: 
  363:  * The third element is "w" for writable disks, and "r" for read-only
  364:    disks.
  365: 
  366: Example:
  367: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  368: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ]
  369: """]]
  370: 
  371: Note that NetBSD by default creates only vnd[0123].  If you need more
  372: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
  373: dom0.
  374: 
  375: Note that NetBSD by default creates only xbd[0123].  If you need more
  376: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
  377: 
  378: Virtual Networking
  379: ------------------
  380: 
  381: Xen provides virtual Ethernets, each of which connects the dom0 and a
  382: domU.  For each virtual network, there is an interface "xvifN.M" in
  383: the dom0, and a matching interface xennetM (NetBSD name) in domU index N.
  384: The interfaces behave as if there is an Ethernet with two
  385: adapters connected.  From this primitive, one can construct various
  386: configurations.  We focus on two common and useful cases for which
  387: there are existing scripts: bridging and NAT.
  388: 
  389: With bridging (in the example above), the domU perceives itself to be
  390: on the same network as the dom0.  For server virtualization, this is
  391: usually best.  Bridging is accomplished by creating a bridge(4) device
  392: and adding the dom0's physical interface and the various xvifN.0
  393: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  394: config file.  The bridge must be set up already in the dom0; an
  395: example /etc/ifconfig.bridge0 is:
  396: 
  397: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
  398: create
  399: up
  400: !brconfig bridge0 add wm0
  401: """]]
  402: 
  403: With NAT, the domU perceives itself to be behind a NAT running on the
  404: dom0.  This is often appropriate when running Xen on a workstation.
  405: TODO: NAT appears to be configured by "vif = [ '' ]".
  406: 
  407: The MAC address specified is the one used for the interface in the new
  408: domain.  The interface in dom0 will use this address XOR'd with
  409: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  410: 
  411: Starting domains automatically
  412: ------------------------------
  413: 
  414: To start domains `domU-netbsd` and `domU-linux` at boot and shut them
  415: down cleanly on dom0 shutdown, add the following in rc.conf:
  416: 
  417: [[!template id=filecontent name="/etc/rc.conf" text="""
  418: xendomains="domU-netbsd domU-linux"
  419: """]]
  420: 
  421: Creating a domU
  422: ===============
  423: 
  424: Creating domUs is almost entirely independent of operating system.  We
  425: have already presented the basics of config files.  Note that you must
  426: have already completed the dom0 setup so that "xl list" works.
  427: 
  428: Creating a NetBSD domU
  429: ----------------------
  430: 
  431: See the earlier config file, and adjust memory.  Decide on how much
  432: storage you will provide, and prepare it (file or lvm).
  433: 
  434: While the kernel will be obtained from the dom0 file system, the same
  435: file should be present in the domU as /netbsd so that tools like
  436: savecore(8) can work.   (This is helpful but not necessary.)
  437: 
  438: The kernel must be specifically for Xen and for use as a domU.  The
  439: i386 and amd64 provide the following kernels:
  440: 
  441:         i386 XEN3PAE_DOMU
  442:         amd64 XEN3_DOMU
  443: 
  444: This will boot NetBSD, but this is not that useful if the disk is
  445: empty.  One approach is to unpack sets onto the disk outside of xen
  446: (by mounting it, just as you would prepare a physical disk for a
  447: system you can't run the installer on).
  448: 
  449: A second approach is to run an INSTALL kernel, which has a miniroot
  450: and can load sets from the network.  To do this, copy the INSTALL
  451: kernel to / and change the kernel line in the config file to:
  452: 
  453:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  454: 
  455: Then, start the domain as "xl create -c configname".
  456: 
  457: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  458: line should be used in the config file.
  459: 
  460:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  461: 
  462: After booting the domain, the option to install via CDROM may be
  463: selected.  The CDROM device should be changed to `xbd1d`.
  464: 
  465: Once done installing, "halt -p" the new domain (don't reboot or halt,
  466: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  467: config file), switch the config file back to the XEN3_DOMU kernel,
  468: and start the new domain again. Now it should be able to use "root on
  469: xbd0a" and you should have a, functional NetBSD domU.
  470: 
  471: TODO: check if this is still accurate.
  472: When the new domain is booting you'll see some warnings about *wscons*
  473: and the pseudo-terminals. These can be fixed by editing the files
  474: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  475: `/etc/ttys`, except *console*, like this:
  476: 
  477:     console "/usr/libexec/getty Pc"         vt100   on secure
  478:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  479:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  480:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  481:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  482: 
  483: Finally, all screens must be commented out from `/etc/wscons.conf`.
  484: 
  485: It is also desirable to add
  486: 
  487:         powerd=YES
  488: 
  489: in rc.conf. This way, the domain will be properly shut down if
  490: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
  491: 
  492: It is not strictly necessary to have a kernel (as /netbsd) in the domU
  493: file system.  However, various programs (e.g. netstat) will use that
  494: kernel to look up symbols to read from kernel virtual memory.  If
  495: /netbsd is not the running kernel, those lookups will fail.  (This is
  496: not really a Xen-specific issue, but because the domU kernel is
  497: obtained from the dom0, it is far more likely to be out of sync or
  498: missing with Xen.)
  499: 
  500: Creating a Linux domU
  501: ---------------------
  502: 
  503: Creating unprivileged Linux domains isn't much different from
  504: unprivileged NetBSD domains, but there are some details to know.
  505: 
  506: First, the second parameter passed to the disk declaration (the '0x1' in
  507: the example below)
  508: 
  509:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  510: 
  511: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  512: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
  513: So, hda1 which has major 3 and minor 1 on a Linux system will have
  514: device number 0x301.  Alternatively, devices names can be used (hda,
  515: hdb, ...)  as xentools has a table to map these names to devices
  516: numbers.  To export a partition to a Linux guest we can use:
  517: 
  518:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
  519:         root = "/dev/hda1 ro"
  520: 
  521: and it will appear as /dev/hda on the Linux system, and be used as root
  522: partition.
  523: 
  524: To install the Linux system on the partition to be exported to the
  525: guest domain, the following method can be used: install
  526: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
  527: that will be the root partition of your Linux domain, and mount it.
  528: Then copy the files from a working Linux system, make adjustments in
  529: `/etc` (fstab, network config).  It should also be possible to extract
  530: binary packages such as .rpm or .deb directly to the mounted partition
  531: using the appropriate tool, possibly running under NetBSD's Linux
  532: emulation.  Once the file system has been populated, umount it.  If
  533: desirable, the file system can be converted to ext3 using tune2fs -j.
  534: It should now be possible to boot the Linux guest domain, using one of
  535: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
  536: 
  537: To get the Linux console right, you need to add:
  538: 
  539:     extra = "xencons=tty1"
  540: 
  541: to your configuration since not all Linux distributions auto-attach a
  542: tty to the xen console.
  543: 
  544: Creating a Solaris domU
  545: -----------------------
  546: 
  547: See possibly outdated
  548: [Solaris domU instructions](/ports/xen/howto-solaris/).
  549: 
  550: 
  551: PCI passthrough: Using PCI devices in guest domains
  552: ---------------------------------------------------
  553: 
  554: The dom0 can give other domains access to selected PCI
  555: devices. This can allow, for example, a non-privileged domain to have
  556: access to a physical network interface or disk controller.  However,
  557: keep in mind that giving a domain access to a PCI device most likely
  558: will give the domain read/write access to the whole physical memory,
  559: as PCs don't have an IOMMU to restrict memory access to DMA-capable
  560: device.  Also, it's not possible to export ISA devices to non-dom0
  561: domains, which means that the primary VGA adapter can't be exported.
  562: A guest domain trying to access the VGA registers will panic.
  563: 
  564: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
  565: not been ported to later versions at this time.
  566: 
  567: For a PCI device to be exported to a domU, is has to be attached to
  568: the "pciback" driver in dom0.  Devices passed to the dom0 via the
  569: pciback.hide boot parameter will attach to "pciback" instead of the
  570: usual driver.  The list of devices is specified as "(bus:dev.func)",
  571: where bus and dev are 2-digit hexadecimal numbers, and func a
  572: single-digit number:
  573: 
  574:         pciback.hide=(00:0a.0)(00:06.0)
  575: 
  576: pciback devices should show up in the dom0's boot messages, and the
  577: devices should be listed in the `/kern/xen/pci` directory.
  578: 
  579: PCI devices to be exported to a domU are listed in the "pci" array of
  580: the domU's config file, with the format "0000:bus:dev.func".
  581: 
  582:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
  583: 
  584: In the domU an "xpci" device will show up, to which one or more pci
  585: buses will attach.  Then the PCI drivers will attach to PCI buses as
  586: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
  587: or any PCI drivers built in by default; you have to build your own
  588: kernel to use PCI devices in a domU.  Here's a kernel config example;
  589: note that only the "xpci" lines are unusual.
  590: 
  591:         include         "arch/i386/conf/XEN3_DOMU"
  592: 
  593:         # Add support for PCI buses to the XEN3_DOMU kernel
  594:         xpci* at xenbus ?
  595:         pci* at xpci ?
  596: 
  597:         # PCI USB controllers
  598:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
  599: 
  600:         # USB bus support
  601:         usb*    at uhci?
  602: 
  603:         # USB Hubs
  604:         uhub*   at usb?
  605:         uhub*   at uhub? port ? configuration ? interface ?
  606: 
  607:         # USB Mass Storage
  608:         umass*  at uhub? port ? configuration ? interface ?
  609:         wd*     at umass?
  610:         # SCSI controllers
  611:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
  612: 
  613:         # SCSI bus support (for both ahc and umass)
  614:         scsibus* at scsi?
  615: 
  616:         # SCSI devices
  617:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
  618:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
  619: 
  620: 
  621: NetBSD as a domU in a VPS
  622: =========================
  623: 
  624: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
  625: hardware.  This section explains how to deal with Xen in a domU as a
  626: virtual private server where you do not control or have access to the
  627: dom0.  This is not intended to be an exhaustive list of VPS providers;
  628: only a few are mentioned that specifically support NetBSD.
  629: 
  630: VPS operators provide varying degrees of access and mechanisms for
  631: configuration.  The big issue is usually how one controls which kernel
  632: is booted, because the kernel is nominally in the dom0 file system (to
  633: which VPS users do not normally have access).  A second issue is how
  634: to install NetBSD.
  635: A VPS user may want to compile a kernel for security updates, to run
  636: npf, run IPsec, or any other reason why someone would want to change
  637: their kernel.
  638: 
  639: One approach is to have an administrative interface to upload a kernel,
  640: or to select from a prepopulated list.  Other approaches are pygrub
  641: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
  642: kernel from the domU file system.  This is closer to a regular physical
  643: computer, where someone who controls a machine can replace the kernel.
  644: 
  645: A second issue is multiple CPUs.  With NetBSD 6, domUs support
  646: multiple vcpus, and it is typical for VPS providers to enable multiple
  647: CPUs for NetBSD domUs.
  648: 
  649: pygrub
  650: -------
  651: 
  652: pygrub runs in the dom0 and looks into the domU file system.  This
  653: implies that the domU must have a kernel in a file system in a format
  654: known to pygrub.  As of 2014, pygrub seems to be of mostly historical
  655: interest.
  656: 
  657: pvgrub
  658: ------
  659: 
  660: pvgrub is a version of grub that uses PV operations instead of BIOS
  661: calls.  It is booted from the dom0 as the domU kernel, and then reads
  662: /grub/menu.lst and loads a kernel from the domU file system.
  663: 
  664: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
  665: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
  666: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
  667: page](http://www.panix.com/v-colo/grub.html), which describes only
  668: Linux but should be updated to cover NetBSD :-).
  669: 
  670: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
  671: their own kernel.  See then [prgmr.com NetBSD
  672: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
  673: (which is in need of updating).
  674: 
  675: It appears that [grub's FFS
  676: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
  677: does not support all aspects of modern FFS, but there are also reports
  678: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
  679: partition for the kernel with the intent that grub can understand it,
  680: which leads to /netbsd not being the actual kernel.  One must remember
  681: to update the special boot partition.
  682: 
  683: Amazon
  684: ------
  685: 
  686: See the [Amazon EC2 page](/amazon_ec2/).
  687: 
  688: Random pointers
  689: ===============
  690: 
  691: This section contains links from elsewhere not yet integrated into the
  692: HOWTO, and other guides.
  693: 
  694: * http://www.lumbercartel.ca/library/xen/
  695: * http://pbraun.nethence.com/doc/sysutils/xen_netbsd_dom0.html
  696: * https://gmplib.org/~tege/xen.html

CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb