File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.203: download - view: text, annotated - select for diffs
Mon Mar 15 00:11:52 2021 UTC (3 months ago) by gdt
Branches: MAIN
CVS tags: HEAD
xen: rototill SMP and PV/HVM/etc. sections

    1: [[!meta title="Xen Status and 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 document provides status on what Xen things work on NetBSD
   12: (upstream documentation might say something works if it works on some
   13: particular Linux system).
   14: 
   15: This document is also a HOWTO that presumes a basic familiarity with
   16: the Xen system architecture, with installing NetBSD on amd64 hardware,
   17: and with installing software from pkgsrc.  See also the [Xen
   18: website](http://www.xenproject.org/).
   19: 
   20: If this document says that something works, and you find that it does
   21: not, it is best to ask on port-xen and if you are correct to file a
   22: PR.
   23: 
   24: [[!toc]]
   25: 
   26: # Overview
   27: 
   28: The basic concept of Xen is that the hypervisor (xenkernel) runs on
   29: the hardware, and runs a privileged domain ("dom0") that can access
   30: disks/networking/etc.  One then runs additonal unprivileged domains
   31: (each a "domU"), presumably to do something useful.
   32: 
   33: This HOWTO addresses how to run a NetBSD dom0 (and hence also build
   34: xen itself).  It also addresses how to run domUs in that environment,
   35: and how to deal with having a domU in a Xen environment run by someone
   36: else and/or not running NetBSD.
   37: 
   38: There are many choices one can make; the HOWTO recommends the standard
   39: approach and limits discussion of alternatives in many cases.
   40: 
   41: ## Guest Styles
   42: 
   43: Xen supports different styles of guests.
   44: 
   45: [[!table data="""
   46: Style of guest	|description				|NetBSD dom0?	|NetBSD domU?
   47: PV		|Paravirtualization (hypercalls)   	|yes		|yes
   48: HVM		|unmodified guest - domU perceives HW	|N/A		|yes
   49: PVHVM		|Like HVM but also PV drivers		|N/A		|current only
   50: PVH		|					|not yet	|current only
   51: """]]
   52: 
   53: In Para-Virtualized (PV) mode, the guest OS does not attempt to access
   54: hardware directly, but instead makes hypercalls to the hypervisor; PV
   55: guests must be specifically coded for Xen.
   56: See [PV](https://wiki.xen.org/wiki/Paravirtualization_(PV\)).
   57: 
   58: In HVM mode, no guest modification is required; however, hardware
   59: support is required, such as VT-x on Intel CPUs and SVM on AMD CPUs.
   60: The dom0 runs qemu to emulate hardware.  It is therefore non-sensical
   61: to have an HVM dom0.
   62: 
   63: In PVHVM mode, the guest runs as HVM, but additionally can use PV
   64: drivers for efficiency.  Therefore it is non-sensical for to have a
   65: PVHVM dom0.  See [PV on HVM](https://wiki.xen.org/wiki/PV_on_HVM).
   66: 
   67: There have been two PVH modes: original PVH and PVHv2.  Original PVH
   68: was based on PV mode and is no longer relevant at all.  Therefore
   69: PVHv2 is abreviated PVH.  PVHv2 is basically lightweight HVM with PV
   70: drivers.  A critical feature of it is that qemu is not needed; the
   71: hypervisor can do the emulation that is required.  Thus, a dom0 can be
   72: PVHv2.  The source code uses PVH and config files use pvh, but NB that
   73: this refers to PVHv2.  See
   74: [PVH(v2)](https://wiki.xenproject.org/wiki/PVH_(v2\)_Domu).
   75: 
   76: At system boot, the dom0 kernel is loaded as a module with Xen as the kernel.
   77: The dom0 can start one or more domUs.  (Booting is explained in detail
   78: in the dom0 section.)
   79: 
   80: ## CPU Architecture
   81: 
   82: Xen runs on x86_64 hardware (the NetBSD amd64 port).
   83: 
   84: There is a concept of Xen running on ARM, but there are no reports of this working with NetBSD.
   85: 
   86: The dom0 system should be amd64.  (Instructions for i386PAE dom0 have been removed from the HOWTO.)
   87: 
   88: The domU can be i386PAE or amd64.
   89: i386PAE at one point was considered as [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html) than amd64.
   90: 
   91: ## Xen Versions
   92: 
   93: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
   94: xenkernel and xentools.  We will refer only to the kernel versions,
   95: but note that both packages must be installed together and must have
   96: matching versions.
   97: 
   98: Versions available in pkgsrc:
   99: 
  100: [[!table data="""
  101: Xen Version	|Package Name	|Xen CPU Support	|EOL'ed By Upstream
  102: 4.11		|xenkernel411	|x86_64			|No
  103: 4.13		|xenkernel413	|x86_64			|No
  104: """]]
  105: 
  106: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
  107: 
  108: Older Xen had a python-based management tool called xm, now replaced
  109: by xl.
  110: 
  111: ## NetBSD versions
  112: 
  113: Xen has been supported in NetBSD for a long time, at least since 2005.
  114: Initially Xen was PV only.
  115: 
  116: NetBSD has supported PV for a very long time, in both dom0 and domU.
  117: NetBSD >=8 supports HVM mode as a domU (HVM as a dom0 does not make
  118: sense).
  119: 
  120: Support for PVHVM and PVH is available only in NetBSD-current; this is
  121: currently somewhat experimental, although PVHVM appears reasonably
  122: solid.
  123: 
  124: NetBSD up to and including NetBSD 9 as a dom0 cannot safely run SMP.
  125: Even if one added "options MULTIPROCESSOR" and configured multiple
  126: vcpus, the kernel is likely to crash because of drivers without
  127: adequate locking.
  128: 
  129: NetBSD-current supports SMP in dom0, and XEN3_DOM0 includes "options
  130: MULTIPROCESSOR".
  131: 
  132: NetBSD (since NetBSD 6), when run as a domU, can run SMP, using
  133: multiple CPUs if provided.  The XEN3_DOMU kernel is built
  134: with "options MULITPROCESSOR".
  135: 
  136: Note that while Xen 4.13 is current, the kernel support is still
  137: called XEN3, because the hypercall interface has not changed
  138: significantly.
  139: 
  140: # Creating a NetBSD dom0
  141: 
  142: In order to install a NetBSD as a dom0, one first installs a normal
  143: NetBSD system, and then pivot the install to a dom0 install by
  144: changing the kernel and boot configuration.
  145: 
  146: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
  147: 512M it worked reliably.  This does not make sense, but if you see
  148: "not ELF" after Xen boots, try increasing dom0 RAM.
  149: 
  150: ## Installation of NetBSD
  151: 
  152: [Install NetBSD/amd64](/guide/inst/) just as you would if you were not
  153: using Xen.  Therefore, use the most recent release, or a build from
  154: the most recent stable branch.  Alternatively, use -current, being
  155: mindful of all the usual caveats of lower stability of current, and
  156: likely a bit more so.
  157: 
  158: ## Installation of Xen
  159: 
  160: ### Building Xen
  161: 
  162: Use the most recent version of Xen in pkgsrc, unless the DESCR says that it is not suitable.
  163: Therefore, choose 4.13.
  164: In the dom0, install xenkernel413 and xentools413 from pkgsrc.
  165: 
  166: Once this is done, copy the Xen kernel from where pkgsrc puts it to
  167: where the boot process will be able to find it:
  168: 
  169: [[!template id=programlisting text="""
  170: # cp -p /usr/pkg/xen413-kernel/xen.gz /
  171: """]]
  172: 
  173: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such
  174: kernel can either be taken from a local release build.sh run, compiled
  175: manually, or downloaded from the NetBSD FTP, for example at:
  176: 
  177: [[!template id=programlisting text="""
  178: ftp.netbsd.org/pub/NetBSD/NetBSD-9.1/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
  179: """]]
  180: 
  181: ### Configuring booting
  182: 
  183: Read boot.cfg(8) carefully.  Add lines to /boot.cfg to boot Xen,
  184: adjusting for your root filesystem:
  185: 
  186: [[!template id=filecontent name="/boot.cfg" text="""
  187: menu=Xen:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc;multiboot /xen.gz dom0_mem=512M
  188: menu=Xen single user:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc -s;multiboot /xen.gz dom0_mem=512M
  189: """]]
  190: 
  191: This specifies that the dom0 should have 512MB of ram, leaving the rest
  192: to be allocated for domUs.
  193: 
  194: NB: This says add, not replace, so that you will be able to more
  195: easily boot a NetBSD kernel without Xen.  Once Xen boots ok, you may
  196: want to set it as default.  It is highly likely that you will have
  197: trouble at some point, and keeping an up-to-date GENERIC for use in
  198: fixing problems is the standard prudent approach.
  199: 
  200: \todo Explain why rndseed is not set with Xen as part of the dom0
  201: subconfiguration.
  202: 
  203: Note that you are likely to have to set root= because the boot device
  204: from /boot is not passed via Xen to the dom0 kernel.  With one disk,
  205: it will work, but e.g. plugging in USB disk to a machine with root on
  206: wd0a causes boot to fail.
  207: 
  208: Beware that userconf statements must be attached to the dom0 load, and
  209: may not be at top-level, because then they would try to configure the
  210: hypervisor, if there is a way to pass them via multiboot .  It appears
  211: that adding `userconf=pckbc` to `/boot.cfg` causes Xen to crash very
  212: early with a heap overflow.
  213: 
  214: ### Console selection
  215: 
  216: See boot_console(8).  Understand that you should start from a place of
  217: having console setup correct for booting GENERIC before trying to
  218: configure Xen.
  219: 
  220: Generally for GENERIC, one sets the console in bootxx_ffsv1 or
  221: equivalent, and this is passed on to /boot (where one typically does
  222: not set the console).  This configuration of bootxx_ffsv1 should also
  223: be in place for Xen systems, to allow seeing messages from /boot and
  224: use of a keyboard to select a line from the menu.  And, one should
  225: have a working boot path to GENERIC for rescue situations.
  226: 
  227: With GENERIC, the boot options are passed on to /netbsd, but there is
  228: currently no mechanism to pass these via multiboot to the hypervisor.
  229: Thus, in addition to configuring the console in the boot blocks, one
  230: must also configure it for Xen.
  231: 
  232: By default, the hypervisor (Xen itself) will use some sort of vga
  233: device as the console, much like GENERIC uses by default.  The vga
  234: console is relinquished at the conclusion of hypervisor boot, before
  235: the dom0 is started.  Xen when using a vga console does not process
  236: console input.
  237: 
  238: The hypervisor can be configured to use a serial port console, e.g.
  239: [[!template id=filecontent name="/boot.cfg" text="""
  240: menu=Xen:load /netbsd-XEN3_DOM0.gz console=com0;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
  241: """]]
  242: This example uses the first serial port (Xen counts from 1; this is
  243: what NetBSD would call com0), and sets speed and parity.  (The dom0 is
  244: then configured to use the same serial port in this example.)
  245: 
  246: With the hypervisor configured for a serial console, it can get input,
  247: and there is a notion of passing this input to the dom0.  \todo
  248: Explain why, if Xen has a serial console, the dom0 console is
  249: typically also configured to open that same serial port, instead of
  250: getting the passthrough input via the xen console.
  251: 
  252: One also configures the console for the dom0.  While one might expect
  253: console=pc to be default, following behavior of GENERIC, a hasty read
  254: of the code suggests there is no default and booting without a
  255: selected console might lead to a panic.  Also, there is merit in
  256: explicit configuration.  Therefore the standard approach is to place
  257: console=pc as part of the load statement for the dom0 kernel, or
  258: alternatively console=com0.
  259: 
  260: The NetBSD dom0 kernel will attach xencons(4) (the man page does not
  261: exist), but this is not used as a console.  It is used to obtain the
  262: messages from the hypervisor's console; run `xl dmesg` to see them.
  263: 
  264: ### Tuning
  265: 
  266: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`,
  267: to force only one vcpu to be provided (since NetBSD dom0 can't use
  268: more) and to pin that vcpu to a physical CPU. Xen has
  269: [many boot options](http://xenbits.xenproject.org/docs/4.13-testing/misc/xen-command-line.html),
  270: and other than dom0 memory and max_vcpus, they are generally not
  271: necessary.
  272: \todo Revisit this advice with current.
  273: \todo Explain if anyone has ever actually measured that this helps.
  274: 
  275: ### rc.conf
  276: 
  277: Ensure that the boot scripts installed in
  278: `/usr/pkg/share/examples/rc.d` are in `/etc/rc.d`, either because you
  279: have `PKG_RCD_SCRIPTS=yes`, or manually.  (This is not special to Xen,
  280: but a normal part of pkgsrc usage.)
  281: 
  282: Set `xencommons=YES` in rc.conf:
  283: 
  284: [[!template id=filecontent name="/etc/rc.conf" text="""
  285: xencommons=YES
  286: """]]
  287: 
  288: \todo Recommend for/against xen-watchdog.
  289: 
  290: ### Testing
  291: 
  292: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  293: than GENERIC without Xen.
  294: 
  295: Once the reboot is done, use `xl` to inspect Xen's boot messages,
  296: available resources, and running domains.  For example:
  297: 
  298: [[!template id=programlisting text="""
  299: # xl dmesg
  300: ... xen's boot info ...
  301: # xl info
  302: ... available memory, etc ...
  303: # xl list
  304: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  305: Domain-0           0       64    0  r----     58.1
  306: """]]
  307: 
  308: Xen logs will be in /var/log/xen.
  309: 
  310: ### Issues with xencommons
  311: 
  312: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
  313: domUs.  It does not currently work to stop and start xenstored.
  314: Certainly all domUs should be shutdown first, following the sort order
  315: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
  316: and is not notified when xenstored exits, leading to not recreating
  317: the state when the new xenstored starts.  Until there's a mechanism to
  318: make this work, one should not expect to be able to restart xenstored
  319: (and thus xencommons).  There is currently no reason to expect that
  320: this will get fixed any time soon.
  321: \todo Confirm if this is still true in 2020.
  322: 
  323: ## Xen-specific NetBSD issues
  324: 
  325: There are (at least) two additional things different about NetBSD as a
  326: dom0 kernel compared to hardware.
  327: 
  328: One is that through NetBSD 9 the module ABI is different because some
  329: of the #defines change, so there are separate sets of modules in
  330: /stand.  In NetBSD-current, there is only one set of modules.
  331: 
  332: The other difference is that XEN3_DOM0 does not have exactly the same
  333: options as GENERIC.  While it is debatable whether or not this is a
  334: bug, users should be aware of this and can simply add missing config
  335: items if desired.
  336: 
  337: Finally, there have been occasional reports of trouble with X11
  338: servers in NetBSD as a dom0.
  339: 
  340: ## Updating Xen in a dom0
  341: 
  342: Basically, update the xenkernel and xentools packages and copy the new
  343: Xen kernel into place, and reboot.  This procedure should be usable to
  344: update to a new Xen release, but the reader is reminded that having a
  345: non-Xen boot methods was recommended earlier.
  346: 
  347: ## Updating NetBSD in a dom0
  348: 
  349: This is just like updating NetBSD on bare hardware, assuming the new
  350: version supports the version of Xen you are running.  Generally, one
  351: replaces the kernel and reboots, and then overlays userland binaries
  352: and adjusts `/etc`.
  353: 
  354: Note that one should update both the non-Xen kernel typically used for
  355: rescue purposes, as well as the DOM0 kernel used with Xen.
  356: 
  357: ## anita (for testing NetBSD)
  358: 
  359: With a NetBSD dom0, even without any domUs, one should be able to run
  360: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
  361: root, because anita must create a domU):
  362: 
  363: [[!template id=programlisting text="""
  364: anita --vmm=xl test file:///usr/obj/i386/
  365: """]]
  366: 
  367: # Unprivileged domains (domU)
  368: 
  369: This section describes general concepts about domUs.  It does not
  370: address specific domU operating systems or how to install them.  The
  371: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
  372: typically named so that the file name, domU name and the domU's host
  373: name match.
  374: 
  375: The domU is provided with CPU and memory by Xen, configured by the
  376: dom0.  The domU is provided with disk and network by the dom0,
  377: mediated by Xen, and configured in the dom0.
  378: 
  379: Entropy in domUs can be an issue; physical disks and network are on
  380: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  381: 
  382: ## Config files
  383: 
  384: See /usr/pkg/share/examples/xen/xlexample* for a very small number of
  385: examples for running GNU/Linux.
  386: 
  387: The following is an example minimal domain configuration file. The domU
  388: serves as a network file server.
  389: 
  390: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  391: name = "domU-id"
  392: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  393: memory = 1024
  394: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  395: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  396:          'file:/n0/xen/foo-wd1,0x1,w' ]
  397: """]]
  398: 
  399: The domain will have name given in the `name` setting.  The kernel has the
  400: host/domU name in it, so that on the dom0 one can update the various
  401: domUs independently.  The `vif` line causes an interface to be provided,
  402: with a specific mac address (do not reuse MAC addresses!), in bridge
  403: mode.  Two disks are provided, and they are both writable; the bits
  404: are stored in files and Xen attaches them to a vnd(4) device in the
  405: dom0 on domain creation.  The system treats xbd0 as the boot device
  406: without needing explicit configuration.
  407: 
  408: There is not type line; that implicitly defines a pv domU.  Otherwise,
  409: one sets type to the lower-case version of the domU type in the table
  410: above, e.g. `type = "hvm"`.
  411: 
  412: By convention, domain config files are kept in `/usr/pkg/etc/xen`.  Note
  413: that "xl create" takes the name of a config file, while other commands
  414: take the name of a domain.
  415: 
  416: Examples of commands:
  417: 
  418: [[!template id=programlisting text="""
  419: xl create /usr/pkg/etc/xen/foo
  420: xl console domU-id
  421: xl create -c /usr/pkg/etc/xen/foo
  422: xl shutdown domU-id
  423: xl list
  424: """]]
  425: 
  426: Typing `^]` will exit the console session.  Shutting down a domain is
  427: equivalent to pushing the power button; a NetBSD domU will receive a
  428: power-press event and do a clean shutdown.  Shutting down the dom0
  429: will trigger controlled shutdowns of all configured domUs.
  430: 
  431: ## CPU and memory
  432: 
  433: A domain is provided with some number of vcpus, up to the number
  434: of CPUs seen by the hypervisor. For a domU, it is controlled
  435: from the config file by the "vcpus = N" directive.
  436: 
  437: A domain is provided with memory; this is controlled in the config
  438: file by "memory = N" (in megabytes).  In the straightforward case, the
  439: sum of the the memory allocated to the dom0 and all domUs must be less
  440: than the available memory.
  441: 
  442: ## Balloon driver
  443: 
  444: Xen provides a `balloon` driver, which can be used to let domains use
  445: more memory temporarily.
  446: 
  447: \todo Explain how to set up a aystem to use the balloon scheme in a
  448: useful manner.
  449: 
  450: ## Virtual disks
  451: 
  452: In domU config files, the disks are defined as a sequence of 3-tuples:
  453: 
  454:  * The first element is "method:/path/to/disk". Common methods are
  455:    "file:" for a file-backed vnd, and "phy:" for something that is already
  456:    a device, such as an LVM logical volume.
  457: 
  458:  * The second element is an artifact of how virtual disks are passed to
  459:    Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  460:    are given a device name to associate with the disk, and values like
  461:    "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  462:    as xbd0, the second as xbd1, and so on.  However, xl demands a
  463:    second argument.  The name given is converted to a major/minor by
  464:    calling stat(2) on the name in /dev and this is passed to the domU.
  465:    In the general case, the dom0 and domU can be different operating
  466:    systems, and it is an unwarranted assumption that they have consistent
  467:    numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  468:    as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  469:    for the second works fine and avoids this issue.  For a GNU/Linux
  470:    guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  471:    /dev/hda1.
  472: 
  473:  * The third element is "w" for writable disks, and "r" for read-only
  474:    disks.
  475: 
  476: Example:
  477: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  478: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ]
  479: """]]
  480: 
  481: Note that NetBSD by default creates only vnd[0123].  If you need more
  482: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
  483: dom0.
  484: 
  485: ## Virtual Networking
  486: 
  487: Xen provides virtual Ethernets, each of which connects the dom0 and a
  488: domU.  For each virtual network, there is an interface "xvifN.M" in
  489: the dom0, and a matching interface xennetM (NetBSD name) in domU index N.
  490: The interfaces behave as if there is an Ethernet with two
  491: adapters connected.  From this primitive, one can construct various
  492: configurations.  We focus on two common and useful cases for which
  493: there are existing scripts: bridging and NAT.
  494: 
  495: With bridging (in the example above), the domU perceives itself to be
  496: on the same network as the dom0.  For server virtualization, this is
  497: usually best.  Bridging is accomplished by creating a bridge(4) device
  498: and adding the dom0's physical interface and the various xvifN.0
  499: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  500: config file.  The bridge must be set up already in the dom0; an
  501: example /etc/ifconfig.bridge0 is:
  502: 
  503: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
  504: create
  505: up
  506: !brconfig bridge0 add wm0
  507: """]]
  508: 
  509: With NAT, the domU perceives itself to be behind a NAT running on the
  510: dom0.  This is often appropriate when running Xen on a workstation.
  511: TODO: NAT appears to be configured by "vif = [ '' ]".
  512: 
  513: The MAC address specified is the one used for the interface in the new
  514: domain.  The interface in dom0 will use this address XOR'd with
  515: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  516: 
  517: ## Starting domains automatically
  518: 
  519: To start domains `domU-netbsd` and `domU-linux` at boot and shut them
  520: down cleanly on dom0 shutdown, add the following in rc.conf:
  521: 
  522: [[!template id=filecontent name="/etc/rc.conf" text="""
  523: xendomains="domU-netbsd domU-linux"
  524: """]]
  525: 
  526: # domU setup for specific systems
  527: 
  528: Creating domUs is almost entirely independent of operating system.  We
  529: have already presented the basics of config files in the previous system.
  530: 
  531: Of course, this section presumes that you have a working dom0.
  532: 
  533: ## Creating a NetBSD PV domU
  534: 
  535: See the earlier config file, and adjust memory.  Decide on how much
  536: storage you will provide, and prepare it (file or LVM).
  537: 
  538: While the kernel will be obtained from the dom0 file system, the same
  539: file should be present in the domU as /netbsd so that tools like
  540: savecore(8) can work.   (This is helpful but not necessary.)
  541: 
  542: The kernel must be specifically built for Xen, to use PV interfacesas
  543: a domU.  NetBSD release builds provide the following kernels:
  544: 
  545:         i386 XEN3PAE_DOMU
  546:         amd64 XEN3_DOMU
  547: 
  548: This will boot NetBSD, but this is not that useful if the disk is
  549: empty.  One approach is to unpack sets onto the disk outside of Xen
  550: (by mounting it, just as you would prepare a physical disk for a
  551: system you can't run the installer on).
  552: 
  553: A second approach is to run an INSTALL kernel, which has a miniroot
  554: and can load sets from the network.  To do this, copy the INSTALL
  555: kernel to / and change the kernel line in the config file to:
  556: 
  557:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  558: 
  559: Then, start the domain as "xl create -c configfile".
  560: 
  561: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  562: line should be used in the config file.
  563: 
  564:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  565: 
  566: After booting the domain, the option to install via CDROM may be
  567: selected.  The CDROM device should be changed to `xbd1d`.
  568: 
  569: Once done installing, "halt -p" the new domain (don't reboot or halt:
  570: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  571: config file), switch the config file back to the XEN3_DOMU kernel,
  572: and start the new domain again. Now it should be able to use "root on
  573: xbd0a" and you should have a functional NetBSD domU.
  574: 
  575: TODO: check if this is still accurate.
  576: When the new domain is booting you'll see some warnings about *wscons*
  577: and the pseudo-terminals. These can be fixed by editing the files
  578: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  579: `/etc/ttys`, except *console*, like this:
  580: 
  581:     console "/usr/libexec/getty Pc"         vt100   on secure
  582:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  583:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  584:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  585:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  586: 
  587: Finally, all screens must be commented out from `/etc/wscons.conf`.
  588: 
  589: One should also run `powerd` in a domU, but this should not need
  590: configuring.  With powerd, the domain will run a controlled shutdown
  591: if `xl shutdown -R` or `xl shutdown -H` is used on the dom0, via
  592: receiving a synthetic `power button pressed` signal.  In 9 and
  593: current, `powerd` is run by default under Xen kernels (or if ACPI is
  594: present), and it can be added to rc.conf if not.
  595: 
  596: It is not strictly necessary to have a kernel (as /netbsd) in the domU
  597: file system.  However, various programs (e.g. netstat) will use that
  598: kernel to look up symbols to read from kernel virtual memory.  If
  599: /netbsd is not the running kernel, those lookups will fail.  (This is
  600: not really a Xen-specific issue, but because the domU kernel is
  601: obtained from the dom0, it is far more likely to be out of sync or
  602: missing with Xen.)
  603: 
  604: Note that NetBSD by default creates only xbd[0123].  If you need more
  605: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
  606: 
  607: ## Creating a Linux PV domU
  608: 
  609: Creating unprivileged Linux domains isn't much different from
  610: unprivileged NetBSD domains, but there are some details to know.
  611: 
  612: First, the second parameter passed to the disk declaration (the '0x1' in
  613: the example below)
  614: 
  615:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  616: 
  617: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  618: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
  619: So, hda1 which has major 3 and minor 1 on a Linux system will have
  620: device number 0x301.  Alternatively, devices names can be used (hda,
  621: hdb, ...)  as xentools has a table to map these names to devices
  622: numbers.  To export a partition to a Linux guest we can use:
  623: 
  624:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
  625:         root = "/dev/hda1 ro"
  626: 
  627: and it will appear as /dev/hda on the Linux system, and be used as root
  628: partition.
  629: 
  630: To install the Linux system on the partition to be exported to the
  631: guest domain, the following method can be used: install
  632: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
  633: that will be the root partition of your Linux domain, and mount it.
  634: Then copy the files from a working Linux system, make adjustments in
  635: `/etc` (fstab, network config).  It should also be possible to extract
  636: binary packages such as .rpm or .deb directly to the mounted partition
  637: using the appropriate tool, possibly running under NetBSD's Linux
  638: emulation.  Once the file system has been populated, umount it.  If
  639: desirable, the file system can be converted to ext3 using tune2fs -j.
  640: It should now be possible to boot the Linux guest domain, using one of
  641: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
  642: 
  643: To get the Linux console right, you need to add:
  644: 
  645:     extra = "xencons=tty1"
  646: 
  647: to your configuration since not all Linux distributions auto-attach a
  648: tty to the xen console.
  649: 
  650: ## Creating a NetBSD HVM domU
  651: 
  652: Use type='hvm', probably.  Use a GENERIC kernel within the disk image.
  653: 
  654: ## Creating a NetBSD PVH domU
  655: 
  656: This only works with a current kernel in the domU.
  657: 
  658: Use type='pvh'.  Probably, use a GENERIC kernel within the disk image,
  659: which in current has PV support.
  660: 
  661: \todo Verify.
  662: 
  663: \todo Verify if one can have current PVH domU on a 9 dom0.
  664: 
  665: ## Creating a Solaris domU
  666: 
  667: See possibly outdated
  668: [Solaris domU instructions](/ports/xen/howto-solaris/).
  669: 
  670: ## PCI passthrough: Using PCI devices in guest domains
  671: 
  672: NB: PCI passthrough only works on some Xen versions and as of 2020 it
  673: is not clear that it works on any version in pkgsrc.  \todo Reports
  674: confirming or denying this notion should be sent to port-xen@.
  675: 
  676: The dom0 can give other domains access to selected PCI
  677: devices. This can allow, for example, a non-privileged domain to have
  678: access to a physical network interface or disk controller.  However,
  679: keep in mind that giving a domain access to a PCI device most likely
  680: will give the domain read/write access to the whole physical memory,
  681: as PCs don't have an IOMMU to restrict memory access to DMA-capable
  682: device.  Also, it's not possible to export ISA devices to non-dom0
  683: domains, which means that the primary VGA adapter can't be exported.
  684: A guest domain trying to access the VGA registers will panic.
  685: 
  686: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
  687: not been ported to later versions at this time.
  688: 
  689: For a PCI device to be exported to a domU, is has to be attached to
  690: the "pciback" driver in dom0.  Devices passed to the dom0 via the
  691: pciback.hide boot parameter will attach to "pciback" instead of the
  692: usual driver.  The list of devices is specified as "(bus:dev.func)",
  693: where bus and dev are 2-digit hexadecimal numbers, and func a
  694: single-digit number:
  695: 
  696:         pciback.hide=(00:0a.0)(00:06.0)
  697: 
  698: pciback devices should show up in the dom0's boot messages, and the
  699: devices should be listed in the `/kern/xen/pci` directory.
  700: 
  701: PCI devices to be exported to a domU are listed in the "pci" array of
  702: the domU's config file, with the format "0000:bus:dev.func".
  703: 
  704:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
  705: 
  706: In the domU an "xpci" device will show up, to which one or more pci
  707: buses will attach.  Then the PCI drivers will attach to PCI buses as
  708: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
  709: or any PCI drivers built in by default; you have to build your own
  710: kernel to use PCI devices in a domU.  Here's a kernel config example;
  711: note that only the "xpci" lines are unusual.
  712: 
  713:         include         "arch/i386/conf/XEN3_DOMU"
  714: 
  715:         # Add support for PCI buses to the XEN3_DOMU kernel
  716:         xpci* at xenbus ?
  717:         pci* at xpci ?
  718: 
  719:         # PCI USB controllers
  720:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
  721: 
  722:         # USB bus support
  723:         usb*    at uhci?
  724: 
  725:         # USB Hubs
  726:         uhub*   at usb?
  727:         uhub*   at uhub? port ? configuration ? interface ?
  728: 
  729:         # USB Mass Storage
  730:         umass*  at uhub? port ? configuration ? interface ?
  731:         wd*     at umass?
  732:         # SCSI controllers
  733:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
  734: 
  735:         # SCSI bus support (for both ahc and umass)
  736:         scsibus* at scsi?
  737: 
  738:         # SCSI devices
  739:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
  740:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
  741: 
  742: 
  743: # Miscellaneous Information
  744: 
  745: ## Nesting under Linux KVM
  746: 
  747: It is possible to run a Xen and a NetBSD dom0 under Linux KVM.  One
  748: can enable virtio in the dom0 for greater speed.
  749: 
  750: ## Other nesting
  751: 
  752: In theory, any full emulation should be able to run Xen and a NetBSD
  753: dom0.  The HOWTO does not currently have information about Xen XVM
  754: mode, nvmm, qemu, Virtualbox, etc.
  755: 
  756: ## NetBSD 5 as domU
  757: 
  758: [NetBSD 5 is known to panic.](http://mail-index.netbsd.org/port-xen/2018/04/17/msg009181.html)
  759: (However, NetBSD 5 systems should be updated to a supported version.)
  760: 
  761: # NetBSD as a domU in a VPS
  762: 
  763: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
  764: hardware.  This section explains how to deal with Xen in a domU as a
  765: virtual private server where you do not control or have access to the
  766: dom0.  This is not intended to be an exhaustive list of VPS providers;
  767: only a few are mentioned that specifically support NetBSD.
  768: 
  769: VPS operators provide varying degrees of access and mechanisms for
  770: configuration.  The big issue is usually how one controls which kernel
  771: is booted, because the kernel is nominally in the dom0 file system (to
  772: which VPS users do not normally have access).  A second issue is how
  773: to install NetBSD.
  774: A VPS user may want to compile a kernel for security updates, to run
  775: npf, run IPsec, or any other reason why someone would want to change
  776: their kernel.
  777: 
  778: One approach is to have an administrative interface to upload a kernel,
  779: or to select from a prepopulated list.  Other approaches are pygrub
  780: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
  781: kernel from the domU file system.  This is closer to a regular physical
  782: computer, where someone who controls a machine can replace the kernel.
  783: 
  784: A second issue is multiple CPUs.  With NetBSD 6, domUs support
  785: multiple vcpus, and it is typical for VPS providers to enable multiple
  786: CPUs for NetBSD domUs.
  787: 
  788: ## Complexities due to Xen changes
  789: 
  790: Xen has many security advisories and people running Xen systems make
  791: different choices.
  792: 
  793: ### stub domains
  794: 
  795: Some (Linux only?) dom0 systems use something called "stub domains" to
  796: isolate qemu from the dom0 system, as a security and reliabilty
  797: mechanism when running HVM domUs.  Somehow, NetBSD's GENERIC kernel
  798: ends up using PIO for disks rather than DMA.  Of course, all of this
  799: is emulated, but emulated PIO is unusably slow.  This problem is not
  800: currently understood.
  801: 
  802: ### Grant tables
  803: 
  804: There are multiple versions of using grant tables, and some security
  805: advisories have suggested disabling some versions.  Some versions of
  806: NetBSD apparently only use specific versions and this can lead to
  807: "NetBSD current doesn't run on hosting provider X" situations.
  808: 
  809: \todo Explain better.
  810: 
  811: ## Boot methods
  812: 
  813: ### pvgrub
  814: 
  815: pvgrub is a version of grub that uses PV operations instead of BIOS
  816: calls.  It is booted from the dom0 as the domU kernel, and then reads
  817: /grub/menu.lst and loads a kernel from the domU file system.
  818: 
  819: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
  820: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
  821: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
  822: page](http://www.panix.com/v-colo/grub.html), which describes only
  823: Linux but should be updated to cover NetBSD :-).
  824: 
  825: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
  826: their own kernel.  See then [prgmr.com NetBSD
  827: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
  828: (which is in need of updating).
  829: 
  830: It appears that [grub's FFS
  831: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
  832: does not support all aspects of modern FFS, but there are also reports
  833: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
  834: partition for the kernel with the intent that grub can understand it,
  835: which leads to /netbsd not being the actual kernel.  One must remember
  836: to update the special boot partition.
  837: 
  838: ### pygrub
  839: 
  840: pygrub runs in the dom0 and looks into the domU file system.  This
  841: implies that the domU must have a kernel in a file system in a format
  842: known to pygrub.
  843: 
  844: pygrub doesn't seem to work to load Linux images under NetBSD dom0,
  845: and is inherently less secure than pvgrub due to running inside dom0. For both these
  846: reasons, pygrub should not be used, and is only still present so that
  847: historical DomU images using it still work.
  848: 
  849: As of 2014, pygrub seems to be of mostly historical
  850: interest. New DomUs should use pvgrub.
  851: 
  852: ## Specific Providers
  853: 
  854: ### Amazon
  855: 
  856: See the [Amazon EC2 page](/amazon_ec2/).

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