File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.198: download - view: text, annotated - select for diffs
Fri Mar 5 21:52:36 2021 UTC (6 weeks, 5 days ago) by gdt
Branches: MAIN
CVS tags: HEAD
xen howto: caution about root=

    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: This HOWTO presumes a basic familiarity with the Xen system
   12: architecture, with installing NetBSD on amd64 hardware, and with
   13: installing software from pkgsrc.  See also the [Xen
   14: website](http://www.xenproject.org/).
   15: 
   16: [[!toc]]
   17: 
   18: # Overview
   19: 
   20: The basic concept of Xen is that the hypervisor (xenkernel) runs on
   21: the hardware, and runs a privileged domain ("dom0") that can access
   22: disks/networking/etc.  One then runs additonal unprivileged domains
   23: (each a "domU"), presumably to do something useful.
   24: 
   25: This HOWTO addresses how to run a NetBSD dom0 (and hence also build
   26: xen itself).  It also addresses how to run domUs in that environment,
   27: and how to deal with having a domU in a Xen environment run by someone
   28: else and/or not running NetBSD.
   29: 
   30: There are many choices one can make; the HOWTO recommends the standard
   31: approach and limits discussion of alternatives in many cases.
   32: 
   33: ## Guest Styles
   34: 
   35: Xen supports different styles of guests.
   36: 
   37: [[!table data="""
   38: Style of guest	|Supported by NetBSD
   39: PV		|Yes (dom0, domU)
   40: HVM		|Yes (domU)
   41: PVHVM		|current-only (domU)
   42: PVH		|current-only (domU, dom0 not yet)
   43: """]]
   44: 
   45: In Para-Virtualized (PV) mode, the guest OS does not attempt to access
   46: hardware directly, but instead makes hypercalls to the hypervisor; PV
   47: guests must be specifically coded for Xen.
   48: See [PV](https://wiki.xen.org/wiki/Paravirtualization_(PV\)).
   49: 
   50: In HVM mode, no guest modification is required; however, hardware
   51: support is required, such as VT-x on Intel CPUs and SVM on AMD CPUs.
   52: The dom0 runs qemu to emulate hardware.
   53: 
   54: In PVHVM mode, the guest runs as HVM, but additionally can use PV
   55: drivers for efficiency.
   56: See [PV on HVM](https://wiki.xen.org/wiki/PV_on_HVM).
   57: 
   58: There have been two PVH modes: original PVH and PVHv2.  Original PVH
   59: was based on PV mode and is no longer relevant at all.  PVHv2 is
   60: basically lightweight HVM with PV drivers.  A critical feature of it
   61: is that qemu is not needed; the hypervisor can do the emulation that
   62: is required.  Thus, a dom0 can be PVHv2.
   63: The source code uses PVH and config files use pvh; this refers to PVHv2.
   64: See [PVH(v2)](https://wiki.xenproject.org/wiki/PVH_(v2\)_Domu).
   65: 
   66: At system boot, the dom0 kernel is loaded as a module with Xen as the kernel.
   67: The dom0 can start one or more domUs.  (Booting is explained in detail
   68: in the dom0 section.)
   69: 
   70: ## CPU Architecture
   71: 
   72: Xen runs on x86_64 hardware (the NetBSD amd64 port).
   73: 
   74: There is a concept of Xen running on ARM, but there are no reports of this working with NetBSD.
   75: 
   76: The dom0 system should be amd64.  (Instructions for i386PAE dom0 have been removed from the HOWTO.)
   77: 
   78: The domU can be i386PAE or amd64.
   79: i386PAE at one point was considered as [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html) than amd64.
   80: 
   81: ## Xen Versions
   82: 
   83: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
   84: xenkernel and xentools.  We will refer only to the kernel versions,
   85: but note that both packages must be installed together and must have
   86: matching versions.
   87: 
   88: Versions available in pkgsrc:
   89: 
   90: [[!table data="""
   91: Xen Version	|Package Name	|Xen CPU Support	|EOL'ed By Upstream
   92: 4.11		|xenkernel411	|x86_64			|No
   93: 4.13		|xenkernel413	|x86_64			|No
   94: """]]
   95: 
   96: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
   97: 
   98: Older Xen had a python-based management tool called xm, now replaced
   99: by xl.
  100: 
  101: ## NetBSD versions
  102: 
  103: Xen has been supported in NetBSD for a long time, at least since 2005.
  104: Initially Xen was PV only.
  105: 
  106: NetBSD 8 and up support PV and HVM modes.
  107: 
  108: Support for PVHVM and PVH is available only in NetBSD-current.
  109: 
  110: NetBSD up to and including NetBSD 9 as a dom0 does not run SMP,
  111: because some drivers are not yet safe for this.  NetBSD-current
  112: supports SMP in dom0.
  113: 
  114: NetBSD, when run as a domU, can and does typically run SMP.
  115: 
  116: Note that while Xen 4.13 is current, the kernel support is still
  117: called XEN3, because the hypercall interface has not changed
  118: significantly.
  119: 
  120: # Creating a NetBSD dom0
  121: 
  122: In order to install a NetBSD as a dom0, one first installs a normal
  123: NetBSD system, and then pivot the install to a dom0 install by
  124: changing the kernel and boot configuration.
  125: 
  126: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
  127: 512M it worked reliably.  This does not make sense, but if you see
  128: "not ELF" after Xen boots, try increasing dom0 RAM.
  129: 
  130: ## Installation of NetBSD
  131: 
  132: [Install NetBSD/amd64](/guide/inst/) just as you would if you were not
  133: using Xen.  Therefore, use the most recent release, or a build from
  134: the most recent stable branch.  Alternatively, use -current, being
  135: mindful of all the usual caveats of lower stability of current, and
  136: likely a bit more so.
  137: 
  138: ## Installation of Xen
  139: 
  140: ### Building Xen
  141: 
  142: Use the most recent version of Xen in pkgsrc, unless the DESCR says that it is not suitable.
  143: Therefore, choose 4.13.
  144: In the dom0, install xenkernel413 and xentools413 from pkgsrc.
  145: 
  146: Once this is done, copy the Xen kernel from where pkgsrc puts it to
  147: where the boot process will be able to find it:
  148: 
  149: [[!template id=programlisting text="""
  150: # cp -p /usr/pkg/xen413-kernel/xen.gz /
  151: """]]
  152: 
  153: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such
  154: kernel can either be taken from a local release build.sh run, compiled
  155: manually, or downloaded from the NetBSD FTP, for example at:
  156: 
  157: [[!template id=programlisting text="""
  158: ftp.netbsd.org/pub/NetBSD/NetBSD-9.1/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
  159: """]]
  160: 
  161: ### Configuring booting
  162: 
  163: Read boot.cfg(8) carefully.  Add lines to /boot.cfg to boot Xen,
  164: adjusting for your root filesystem:
  165: 
  166: [[!template id=filecontent name="/boot.cfg" text="""
  167: menu=Xen:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc;multiboot /xen.gz dom0_mem=512M
  168: menu=Xen single user:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc -s;multiboot /xen.gz dom0_mem=512M
  169: """]]
  170: 
  171: This specifies that the dom0 should have 512MB of ram, leaving the rest
  172: to be allocated for domUs.
  173: 
  174: NB: This says add, not replace, so that you will be able to more
  175: easily boot a NetBSD kernel without Xen.  Once Xen boots ok, you may
  176: want to set it as default.  It is highly likely that you will have
  177: trouble at some point, and keeping an up-to-date GENERIC for use in
  178: fixing problems is the standard prudent approach.
  179: 
  180: \todo Explain why rndseed is not set with Xen as part of the dom0
  181: subconfiguration.
  182: 
  183: Note that you are likely to have to set root= because the boot device
  184: from /boot is not passed via Xen to the dom0 kernel.  With one disk,
  185: it will work, but e.g. plugging in USB disk to a machine with root on
  186: wd0a causes boot to fail.
  187: 
  188: Beware that userconf statements must be attached to the dom0 load, and
  189: may not be at top-level, because then they would try to configure the
  190: hypervisor, if there is a way to pass them via multiboot .  It appears
  191: that adding `userconf=pckbc` to `/boot.cfg` causes Xen to crash very
  192: early with a heap overflow.
  193: 
  194: ### Console selection
  195: 
  196: See boot_console(8).  Understand that you should start from a place of
  197: having console setup correct for booting GENERIC before trying to
  198: configure Xen.
  199: 
  200: Generally for GENERIC, one sets the console in bootxx_ffsv1 or
  201: equivalent, and this is passed on to /boot (where one typically does
  202: not set the console).  This configuration of bootxx_ffsv1 should also
  203: be in place for Xen systems, to allow seeing messages from /boot and
  204: use of a keyboard to select a line from the menu.  And, one should
  205: have a working boot path to GENERIC for rescue situations.
  206: 
  207: With GENERIC, the boot options are passed on to /netbsd, but there is
  208: currently no mechanism to pass these via multiboot to the hypervisor.
  209: Thus, in addition to configuring the console in the boot blocks, one
  210: must also configure it for Xen.
  211: 
  212: By default, the hypervisor (Xen itself) will use some sort of vga
  213: device as the console, much like GENERIC uses by default.  The vga
  214: console is relinquished at the conclusion of hypervisor boot, before
  215: the dom0 is started.  Xen when using a vga console does not process
  216: console input.
  217: 
  218: The hypervisor can be configured to use a serial port console, e.g.
  219: [[!template id=filecontent name="/boot.cfg" text="""
  220: menu=Xen:losad /netbsd-XEN3_DOM0.gz console=com0;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
  221: """]]
  222: This example uses the first serial port (Xen counts from 1; this is
  223: what NetBSD would call com0), and sets speed and parity.  (The dom0 is
  224: then configured to use the same serial port in this example.)
  225: 
  226: With the hypervisor configured for a serial console, it can get input,
  227: and there is a notion of passing this input to the dom0.  \todo
  228: Explain why, if Xen has a serial console, the dom0 console is
  229: typically also configured to open that same serial port, instead of
  230: getting the passthrough input via the xen console.
  231: 
  232: One also configures the console for the dom0.  While one might expect
  233: console=pc to be default, following behavior of GENERIC, a hasty read
  234: of the code suggests there is no default and booting without a
  235: selected console might lead to a panic.  Also, there is merit in
  236: explicit configuration.  Therefore the standard approach is to place
  237: console=pc as part of the load statement for the dom0 kernel, or
  238: alternatively console=com0.
  239: 
  240: The NetBSD dom0 kernel will attach xencons(4) (the man page does not
  241: exist), but this is not used as a console.  It is used to obtain the
  242: messages from the hypervisor's console; run `xl dmesg` to see them.
  243: 
  244: ### Tuning
  245: 
  246: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`,
  247: to force only one vcpu to be provided (since NetBSD dom0 can't use
  248: more) and to pin that vcpu to a physical CPU. Xen has
  249: [many boot options](http://xenbits.xenproject.org/docs/4.13-testing/misc/xen-command-line.html),
  250: and other than dom0 memory and max_vcpus, they are generally not
  251: necessary.
  252: \todo Revisit this advice with current.
  253: \todo Explain if anyone has ever actually measured that this helps.
  254: 
  255: ### rc.conf
  256: 
  257: Ensure that the boot scripts installed in
  258: `/usr/pkg/share/examples/rc.d` are in `/etc/rc.d`, either because you
  259: have `PKG_RCD_SCRIPTS=yes`, or manually.  (This is not special to Xen,
  260: but a normal part of pkgsrc usage.)
  261: 
  262: Set `xencommons=YES` in rc.conf:
  263: 
  264: [[!template id=filecontent name="/etc/rc.conf" text="""
  265: xencommons=YES
  266: """]]
  267: 
  268: \todo Recommend for/against xen-watchdog.
  269: 
  270: ### Testing
  271: 
  272: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  273: than GENERIC without Xen.
  274: 
  275: Once the reboot is done, use `xl` to inspect Xen's boot messages,
  276: available resources, and running domains.  For example:
  277: 
  278: [[!template id=programlisting text="""
  279: # xl dmesg
  280: ... xen's boot info ...
  281: # xl info
  282: ... available memory, etc ...
  283: # xl list
  284: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  285: Domain-0           0       64    0  r----     58.1
  286: """]]
  287: 
  288: Xen logs will be in /var/log/xen.
  289: 
  290: ### Issues with xencommons
  291: 
  292: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
  293: domUs.  It does not currently work to stop and start xenstored.
  294: Certainly all domUs should be shutdown first, following the sort order
  295: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
  296: and is not notified when xenstored exits, leading to not recreating
  297: the state when the new xenstored starts.  Until there's a mechanism to
  298: make this work, one should not expect to be able to restart xenstored
  299: (and thus xencommons).  There is currently no reason to expect that
  300: this will get fixed any time soon.
  301: \todo Confirm if this is still true in 2020.
  302: 
  303: ## Xen-specific NetBSD issues
  304: 
  305: There are (at least) two additional things different about NetBSD as a
  306: dom0 kernel compared to hardware.
  307: 
  308: One is that through NetBSD 9 the module ABI is different because some
  309: of the #defines change, so there are separate sets of modules in
  310: /stand.  In NetBSD-current, there is only one set of modules.
  311: 
  312: The other difference is that XEN3_DOM0 does not have exactly the same
  313: options as GENERIC.  While it is debatable whether or not this is a
  314: bug, users should be aware of this and can simply add missing config
  315: items if desired.
  316: 
  317: Finally, there have been occasional reports of trouble with X11
  318: servers in NetBSD as a dom0.
  319: 
  320: ## Updating Xen in a dom0
  321: 
  322: Basically, update the xenkernel and xentools packages and copy the new
  323: Xen kernel into place, and reboot.  This procedure should be usable to
  324: update to a new Xen release, but the reader is reminded that having a
  325: non-Xen boot methods was recommended earlier.
  326: 
  327: ## Updating NetBSD in a dom0
  328: 
  329: This is just like updating NetBSD on bare hardware, assuming the new
  330: version supports the version of Xen you are running.  Generally, one
  331: replaces the kernel and reboots, and then overlays userland binaries
  332: and adjusts `/etc`.
  333: 
  334: Note that one should update both the non-Xen kernel typically used for
  335: rescue purposes, as well as the DOM0 kernel used with Xen.
  336: 
  337: ## anita (for testing NetBSD)
  338: 
  339: With a NetBSD dom0, even without any domUs, one should be able to run
  340: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
  341: root, because anita must create a domU):
  342: 
  343: [[!template id=programlisting text="""
  344: anita --vmm=xl test file:///usr/obj/i386/
  345: """]]
  346: 
  347: # Unprivileged domains (domU)
  348: 
  349: This section describes general concepts about domUs.  It does not
  350: address specific domU operating systems or how to install them.  The
  351: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
  352: typically named so that the file name, domU name and the domU's host
  353: name match.
  354: 
  355: The domU is provided with CPU and memory by Xen, configured by the
  356: dom0.  The domU is provided with disk and network by the dom0,
  357: mediated by Xen, and configured in the dom0.
  358: 
  359: Entropy in domUs can be an issue; physical disks and network are on
  360: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  361: 
  362: ## Config files
  363: 
  364: See /usr/pkg/share/examples/xen/xlexample*
  365: for a small number of well-commented examples, mostly for running
  366: GNU/Linux.
  367: 
  368: The following is an example minimal domain configuration file. The domU
  369: serves as a network file server.
  370: 
  371: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  372: name = "domU-id"
  373: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  374: memory = 1024
  375: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  376: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  377:          'file:/n0/xen/foo-wd1,0x1,w' ]
  378: """]]
  379: 
  380: The domain will have name given in the `name` setting.  The kernel has the
  381: host/domU name in it, so that on the dom0 one can update the various
  382: domUs independently.  The `vif` line causes an interface to be provided,
  383: with a specific mac address (do not reuse MAC addresses!), in bridge
  384: mode.  Two disks are provided, and they are both writable; the bits
  385: are stored in files and Xen attaches them to a vnd(4) device in the
  386: dom0 on domain creation.  The system treats xbd0 as the boot device
  387: without needing explicit configuration.
  388: 
  389: There is not type line; that implicitly defines a pv domU.
  390: 
  391: By convention, domain config files are kept in `/usr/pkg/etc/xen`.  Note
  392: that "xl create" takes the name of a config file, while other commands
  393: take the name of a domain.
  394: 
  395: Examples of commands:
  396: 
  397: [[!template id=programlisting text="""
  398: xl create /usr/pkg/etc/xen/foo
  399: xl console domU-id
  400: xl create -c /usr/pkg/etc/xen/foo
  401: xl shutdown domU-id
  402: xl list
  403: """]]
  404: 
  405: Typing `^]` will exit the console session.  Shutting down a domain is
  406: equivalent to pushing the power button; a NetBSD domU will receive a
  407: power-press event and do a clean shutdown.  Shutting down the dom0
  408: will trigger controlled shutdowns of all configured domUs.
  409: 
  410: ## CPU and memory
  411: 
  412: A domain is provided with some number of vcpus, up to the number
  413: of CPUs seen by the hypervisor. For a domU, it is controlled
  414: from the config file by the "vcpus = N" directive.
  415: 
  416: A domain is provided with memory; this is controlled in the config
  417: file by "memory = N" (in megabytes).  In the straightforward case, the
  418: sum of the the memory allocated to the dom0 and all domUs must be less
  419: than the available memory.
  420: 
  421: ## Balloon driver
  422: 
  423: Xen provides a `balloon` driver, which can be used to let domains use
  424: more memory temporarily.
  425: 
  426: \todo Explain how to set up a aystem to use the balloon scheme in a
  427: useful manner.
  428: 
  429: ## Virtual disks
  430: 
  431: In domU config files, the disks are defined as a sequence of 3-tuples:
  432: 
  433:  * The first element is "method:/path/to/disk". Common methods are
  434:    "file:" for a file-backed vnd, and "phy:" for something that is already
  435:    a device, such as an LVM logical volume.
  436: 
  437:  * The second element is an artifact of how virtual disks are passed to
  438:    Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  439:    are given a device name to associate with the disk, and values like
  440:    "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  441:    as xbd0, the second as xbd1, and so on.  However, xl demands a
  442:    second argument.  The name given is converted to a major/minor by
  443:    calling stat(2) on the name in /dev and this is passed to the domU.
  444:    In the general case, the dom0 and domU can be different operating
  445:    systems, and it is an unwarranted assumption that they have consistent
  446:    numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  447:    as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  448:    for the second works fine and avoids this issue.  For a GNU/Linux
  449:    guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  450:    /dev/hda1.
  451: 
  452:  * The third element is "w" for writable disks, and "r" for read-only
  453:    disks.
  454: 
  455: Example:
  456: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  457: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ]
  458: """]]
  459: 
  460: Note that NetBSD by default creates only vnd[0123].  If you need more
  461: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
  462: dom0.
  463: 
  464: ## Virtual Networking
  465: 
  466: Xen provides virtual Ethernets, each of which connects the dom0 and a
  467: domU.  For each virtual network, there is an interface "xvifN.M" in
  468: the dom0, and a matching interface xennetM (NetBSD name) in domU index N.
  469: The interfaces behave as if there is an Ethernet with two
  470: adapters connected.  From this primitive, one can construct various
  471: configurations.  We focus on two common and useful cases for which
  472: there are existing scripts: bridging and NAT.
  473: 
  474: With bridging (in the example above), the domU perceives itself to be
  475: on the same network as the dom0.  For server virtualization, this is
  476: usually best.  Bridging is accomplished by creating a bridge(4) device
  477: and adding the dom0's physical interface and the various xvifN.0
  478: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  479: config file.  The bridge must be set up already in the dom0; an
  480: example /etc/ifconfig.bridge0 is:
  481: 
  482: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
  483: create
  484: up
  485: !brconfig bridge0 add wm0
  486: """]]
  487: 
  488: With NAT, the domU perceives itself to be behind a NAT running on the
  489: dom0.  This is often appropriate when running Xen on a workstation.
  490: TODO: NAT appears to be configured by "vif = [ '' ]".
  491: 
  492: The MAC address specified is the one used for the interface in the new
  493: domain.  The interface in dom0 will use this address XOR'd with
  494: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  495: 
  496: ## Starting domains automatically
  497: 
  498: To start domains `domU-netbsd` and `domU-linux` at boot and shut them
  499: down cleanly on dom0 shutdown, add the following in rc.conf:
  500: 
  501: [[!template id=filecontent name="/etc/rc.conf" text="""
  502: xendomains="domU-netbsd domU-linux"
  503: """]]
  504: 
  505: # domU setup for specific systems
  506: 
  507: Creating domUs is almost entirely independent of operating system.  We
  508: have already presented the basics of config files in the previous system.
  509: 
  510: Of course, this section presumes that you have a working dom0.
  511: 
  512: ## Creating a NetBSD PV domU
  513: 
  514: See the earlier config file, and adjust memory.  Decide on how much
  515: storage you will provide, and prepare it (file or LVM).
  516: 
  517: While the kernel will be obtained from the dom0 file system, the same
  518: file should be present in the domU as /netbsd so that tools like
  519: savecore(8) can work.   (This is helpful but not necessary.)
  520: 
  521: The kernel must be specifically built for Xen, to use PV interfacesas
  522: a domU.  NetBSD release builds provide the following kernels:
  523: 
  524:         i386 XEN3PAE_DOMU
  525:         amd64 XEN3_DOMU
  526: 
  527: This will boot NetBSD, but this is not that useful if the disk is
  528: empty.  One approach is to unpack sets onto the disk outside of Xen
  529: (by mounting it, just as you would prepare a physical disk for a
  530: system you can't run the installer on).
  531: 
  532: A second approach is to run an INSTALL kernel, which has a miniroot
  533: and can load sets from the network.  To do this, copy the INSTALL
  534: kernel to / and change the kernel line in the config file to:
  535: 
  536:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  537: 
  538: Then, start the domain as "xl create -c configfile".
  539: 
  540: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  541: line should be used in the config file.
  542: 
  543:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  544: 
  545: After booting the domain, the option to install via CDROM may be
  546: selected.  The CDROM device should be changed to `xbd1d`.
  547: 
  548: Once done installing, "halt -p" the new domain (don't reboot or halt:
  549: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  550: config file), switch the config file back to the XEN3_DOMU kernel,
  551: and start the new domain again. Now it should be able to use "root on
  552: xbd0a" and you should have a functional NetBSD domU.
  553: 
  554: TODO: check if this is still accurate.
  555: When the new domain is booting you'll see some warnings about *wscons*
  556: and the pseudo-terminals. These can be fixed by editing the files
  557: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  558: `/etc/ttys`, except *console*, like this:
  559: 
  560:     console "/usr/libexec/getty Pc"         vt100   on secure
  561:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  562:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  563:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  564:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  565: 
  566: Finally, all screens must be commented out from `/etc/wscons.conf`.
  567: 
  568: One should also run `powerd` in a domU, but this should not need
  569: configuring.  With powerd, the domain will run a controlled shutdown
  570: if `xl shutdown -R` or `xl shutdown -H` is used on the dom0, via
  571: receiving a synthetic `power button pressed` signal.  In 9 and
  572: current, `powerd` is run by default under Xen kernels (or if ACPI is
  573: present), and it can be added to rc.conf if not.
  574: 
  575: It is not strictly necessary to have a kernel (as /netbsd) in the domU
  576: file system.  However, various programs (e.g. netstat) will use that
  577: kernel to look up symbols to read from kernel virtual memory.  If
  578: /netbsd is not the running kernel, those lookups will fail.  (This is
  579: not really a Xen-specific issue, but because the domU kernel is
  580: obtained from the dom0, it is far more likely to be out of sync or
  581: missing with Xen.)
  582: 
  583: Note that NetBSD by default creates only xbd[0123].  If you need more
  584: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
  585: 
  586: ## Creating a Linux PV domU
  587: 
  588: Creating unprivileged Linux domains isn't much different from
  589: unprivileged NetBSD domains, but there are some details to know.
  590: 
  591: First, the second parameter passed to the disk declaration (the '0x1' in
  592: the example below)
  593: 
  594:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  595: 
  596: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  597: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
  598: So, hda1 which has major 3 and minor 1 on a Linux system will have
  599: device number 0x301.  Alternatively, devices names can be used (hda,
  600: hdb, ...)  as xentools has a table to map these names to devices
  601: numbers.  To export a partition to a Linux guest we can use:
  602: 
  603:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
  604:         root = "/dev/hda1 ro"
  605: 
  606: and it will appear as /dev/hda on the Linux system, and be used as root
  607: partition.
  608: 
  609: To install the Linux system on the partition to be exported to the
  610: guest domain, the following method can be used: install
  611: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
  612: that will be the root partition of your Linux domain, and mount it.
  613: Then copy the files from a working Linux system, make adjustments in
  614: `/etc` (fstab, network config).  It should also be possible to extract
  615: binary packages such as .rpm or .deb directly to the mounted partition
  616: using the appropriate tool, possibly running under NetBSD's Linux
  617: emulation.  Once the file system has been populated, umount it.  If
  618: desirable, the file system can be converted to ext3 using tune2fs -j.
  619: It should now be possible to boot the Linux guest domain, using one of
  620: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
  621: 
  622: To get the Linux console right, you need to add:
  623: 
  624:     extra = "xencons=tty1"
  625: 
  626: to your configuration since not all Linux distributions auto-attach a
  627: tty to the xen console.
  628: 
  629: ## Creating a NetBSD HVM domU
  630: 
  631: Use type='hvm', probably.  Use a GENERIC kernel within the disk image.
  632: 
  633: ## Creating a NetBSD PVH domU
  634: 
  635: This only works with a current kernel in the domU.
  636: 
  637: Use type='pvh'.  Probably, use a GENERIC kernel within the disk image,
  638: which in current has PV support.
  639: 
  640: \todo Verify.
  641: 
  642: \todo Verify if one can have current PVH domU on a 9 dom0.
  643: 
  644: ## Creating a Solaris domU
  645: 
  646: See possibly outdated
  647: [Solaris domU instructions](/ports/xen/howto-solaris/).
  648: 
  649: ## PCI passthrough: Using PCI devices in guest domains
  650: 
  651: NB: PCI passthrough only works on some Xen versions and as of 2020 it
  652: is not clear that it works on any version in pkgsrc.  \todo Reports
  653: confirming or denying this notion should be sent to port-xen@.
  654: 
  655: The dom0 can give other domains access to selected PCI
  656: devices. This can allow, for example, a non-privileged domain to have
  657: access to a physical network interface or disk controller.  However,
  658: keep in mind that giving a domain access to a PCI device most likely
  659: will give the domain read/write access to the whole physical memory,
  660: as PCs don't have an IOMMU to restrict memory access to DMA-capable
  661: device.  Also, it's not possible to export ISA devices to non-dom0
  662: domains, which means that the primary VGA adapter can't be exported.
  663: A guest domain trying to access the VGA registers will panic.
  664: 
  665: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
  666: not been ported to later versions at this time.
  667: 
  668: For a PCI device to be exported to a domU, is has to be attached to
  669: the "pciback" driver in dom0.  Devices passed to the dom0 via the
  670: pciback.hide boot parameter will attach to "pciback" instead of the
  671: usual driver.  The list of devices is specified as "(bus:dev.func)",
  672: where bus and dev are 2-digit hexadecimal numbers, and func a
  673: single-digit number:
  674: 
  675:         pciback.hide=(00:0a.0)(00:06.0)
  676: 
  677: pciback devices should show up in the dom0's boot messages, and the
  678: devices should be listed in the `/kern/xen/pci` directory.
  679: 
  680: PCI devices to be exported to a domU are listed in the "pci" array of
  681: the domU's config file, with the format "0000:bus:dev.func".
  682: 
  683:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
  684: 
  685: In the domU an "xpci" device will show up, to which one or more pci
  686: buses will attach.  Then the PCI drivers will attach to PCI buses as
  687: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
  688: or any PCI drivers built in by default; you have to build your own
  689: kernel to use PCI devices in a domU.  Here's a kernel config example;
  690: note that only the "xpci" lines are unusual.
  691: 
  692:         include         "arch/i386/conf/XEN3_DOMU"
  693: 
  694:         # Add support for PCI buses to the XEN3_DOMU kernel
  695:         xpci* at xenbus ?
  696:         pci* at xpci ?
  697: 
  698:         # PCI USB controllers
  699:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
  700: 
  701:         # USB bus support
  702:         usb*    at uhci?
  703: 
  704:         # USB Hubs
  705:         uhub*   at usb?
  706:         uhub*   at uhub? port ? configuration ? interface ?
  707: 
  708:         # USB Mass Storage
  709:         umass*  at uhub? port ? configuration ? interface ?
  710:         wd*     at umass?
  711:         # SCSI controllers
  712:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
  713: 
  714:         # SCSI bus support (for both ahc and umass)
  715:         scsibus* at scsi?
  716: 
  717:         # SCSI devices
  718:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
  719:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
  720: 
  721: 
  722: # Miscellaneous Information
  723: 
  724: ## Nesting under Linux KVM
  725: 
  726: It is possible to run a Xen and a NetBSD dom0 under Linux KVM.  One
  727: can enable virtio in the dom0 for greater speed.
  728: 
  729: ## Other nesting
  730: 
  731: In theory, any full emulation should be able to run Xen and a NetBSD
  732: dom0.  The HOWTO does not currently have information about Xen XVM
  733: mode, nvmm, qemu, Virtualbox, etc.
  734: 
  735: ## NetBSD 5 as domU
  736: 
  737: [NetBSD 5 is known to panic.](http://mail-index.netbsd.org/port-xen/2018/04/17/msg009181.html)
  738: (However, NetBSD 5 systems should be updated to a supported version.)
  739: 
  740: # NetBSD as a domU in a VPS
  741: 
  742: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
  743: hardware.  This section explains how to deal with Xen in a domU as a
  744: virtual private server where you do not control or have access to the
  745: dom0.  This is not intended to be an exhaustive list of VPS providers;
  746: only a few are mentioned that specifically support NetBSD.
  747: 
  748: VPS operators provide varying degrees of access and mechanisms for
  749: configuration.  The big issue is usually how one controls which kernel
  750: is booted, because the kernel is nominally in the dom0 file system (to
  751: which VPS users do not normally have access).  A second issue is how
  752: to install NetBSD.
  753: A VPS user may want to compile a kernel for security updates, to run
  754: npf, run IPsec, or any other reason why someone would want to change
  755: their kernel.
  756: 
  757: One approach is to have an administrative interface to upload a kernel,
  758: or to select from a prepopulated list.  Other approaches are pygrub
  759: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
  760: kernel from the domU file system.  This is closer to a regular physical
  761: computer, where someone who controls a machine can replace the kernel.
  762: 
  763: A second issue is multiple CPUs.  With NetBSD 6, domUs support
  764: multiple vcpus, and it is typical for VPS providers to enable multiple
  765: CPUs for NetBSD domUs.
  766: 
  767: ## Complexities due to Xen changes
  768: 
  769: Xen has many security advisories and people running Xen systems make
  770: different choices.
  771: 
  772: ### stub domains
  773: 
  774: Some (Linux only?) dom0 systems use something called "stub domains" to
  775: isolate qemu from the dom0 system, as a security and reliabilty
  776: mechanism when running HVM domUs.  Somehow, NetBSD's GENERIC kernel
  777: ends up using PIO for disks rather than DMA.  Of course, all of this
  778: is emulated, but emulated PIO is unusably slow.  This problem is not
  779: currently understood.
  780: 
  781: ### Grant tables
  782: 
  783: There are multiple versions of using grant tables, and some security
  784: advisories have suggested disabling some versions.  Some versions of
  785: NetBSD apparently only use specific versions and this can lead to
  786: "NetBSD current doesn't run on hosting provider X" situations.
  787: 
  788: \todo Explain better.
  789: 
  790: ## Boot methods
  791: 
  792: ### pvgrub
  793: 
  794: pvgrub is a version of grub that uses PV operations instead of BIOS
  795: calls.  It is booted from the dom0 as the domU kernel, and then reads
  796: /grub/menu.lst and loads a kernel from the domU file system.
  797: 
  798: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
  799: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
  800: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
  801: page](http://www.panix.com/v-colo/grub.html), which describes only
  802: Linux but should be updated to cover NetBSD :-).
  803: 
  804: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
  805: their own kernel.  See then [prgmr.com NetBSD
  806: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
  807: (which is in need of updating).
  808: 
  809: It appears that [grub's FFS
  810: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
  811: does not support all aspects of modern FFS, but there are also reports
  812: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
  813: partition for the kernel with the intent that grub can understand it,
  814: which leads to /netbsd not being the actual kernel.  One must remember
  815: to update the special boot partition.
  816: 
  817: ### pygrub
  818: 
  819: pygrub runs in the dom0 and looks into the domU file system.  This
  820: implies that the domU must have a kernel in a file system in a format
  821: known to pygrub.
  822: 
  823: pygrub doesn't seem to work to load Linux images under NetBSD dom0,
  824: and is inherently less secure than pvgrub due to running inside dom0. For both these
  825: reasons, pygrub should not be used, and is only still present so that
  826: historical DomU images using it still work.
  827: 
  828: As of 2014, pygrub seems to be of mostly historical
  829: interest. New DomUs should use pvgrub.
  830: 
  831: ## Specific Providers
  832: 
  833: ### Amazon
  834: 
  835: See the [Amazon EC2 page](/amazon_ec2/).

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