File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.206: download - view: text, annotated - select for diffs
Thu Apr 15 15:21:25 2021 UTC (3 months, 1 week ago) by gdt
Branches: MAIN
CVS tags: HEAD
xen howto: BIOS/EFI and ugprade improvements

Note that one needs to use BIOS boot, not EFI, for a dom0, for now.

Thanks to Greg Woods for details about updating Xen in a dom0; this
text is summarized from on and off list comments.

    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 additional 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.  See
   44: https://wiki.xenproject.org/wiki/Virtualization_Spectrum for a
   45: discussion.
   46: 
   47: This table shows the styles, and if a NetBSD dom0 can run in that
   48: style, if a NetBSD dom0 can sypport that style of guest in a domU, and
   49: if NetBSD as a domU can support that style.
   50: 
   51: [[!table data="""
   52: Style of guest	|dom0 can be?	|dom0 can support?	|domU can be?
   53: PV		|yes		|yes			|yes
   54: HVM		|N/A		|yes			|yes
   55: PVHVM		|N/A		|yes			|current only
   56: PVH		|not yet	|current only		|current only
   57: """]]
   58: 
   59: In PV (paravirtualized) mode, the guest OS does not attempt to access
   60: hardware directly, but instead makes hypercalls to the hypervisor; PV
   61: guests must be specifically coded for Xen.  See
   62: [PV](https://wiki.xen.org/wiki/Paravirtualization_(PV\)).
   63: 
   64: In HVM (Hardware Virtual Machine) mode, no guest modification is
   65: required.  However, hardware support is required, such as VT-x on
   66: Intel CPUs and SVM on AMD CPUs to assist with the processor emulation.
   67: The dom0 runs qemu to emulate hardware other than the processor.  It
   68: is therefore non-sensical to have an HVM dom0, because there is no
   69: underlying system to provide emulation.
   70: 
   71: In PVHVM mode, the guest runs as HVM, but additionally uses PV
   72: drivers for efficiency.  Therefore it is non-sensical for to have a
   73: PVHVM dom0.  See [PV on HVM](https://wiki.xen.org/wiki/PV_on_HVM).
   74: 
   75: There have been two PVH modes: original PVH and PVHv2.  Original PVH
   76: was based on PV mode and is no longer relevant at all.  Therefore
   77: PVHv2 is written as PVH, here and elsewhere.  PVH is basically
   78: lightweight HVM with PV drivers.  A critical feature of it is that
   79: qemu is not needed; the hypervisor can do the emulation that is
   80: required.  Thus, a dom0 can be PVH.  The source code uses PVH and
   81: config files use pvh, but NB that this refers to PVHv2.  See
   82: [PVH(v2)](https://wiki.xenproject.org/wiki/PVH_(v2\)_Domu).
   83: 
   84: At system boot, the dom0 kernel is loaded as a module with Xen as the
   85: kernel.  The dom0 can start one or more domUs.  (Booting is explained
   86: in detail in the dom0 section.)
   87: 
   88: ## CPU Architecture
   89: 
   90: Xen runs on x86_64 hardware (the NetBSD amd64 port).
   91: 
   92: There is a concept of Xen running on ARM, but there are no reports of this working with NetBSD.
   93: 
   94: The dom0 system should be amd64.  (Instructions for i386PAE dom0 have been removed from the HOWTO.)
   95: 
   96: The domU can be i386 PAE or amd64.
   97: i386 PAE at one point was considered as [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html) than amd64.
   98: However, as of 2021 it is normal to use amd64 as the domU architecture, and use of i386 is dwindling.
   99: 
  100: ## Xen Versions
  101: 
  102: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
  103: xenkernel and xentools.  We will refer only to the kernel versions,
  104: but note that both packages must be installed together and must have
  105: matching versions.
  106: 
  107: Versions available in pkgsrc:
  108: 
  109: [[!table data="""
  110: Xen Version	|Package Name	|Xen CPU Support	|EOL'ed By Upstream
  111: 4.11		|xenkernel411	|x86_64			|No
  112: 4.13		|xenkernel413	|x86_64			|No
  113: """]]
  114: 
  115: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
  116: 
  117: Older Xen had a python-based management tool called xm; this has been
  118: replaced by xl.
  119: 
  120: ## NetBSD versions
  121: 
  122: Xen has been supported in NetBSD for a long time, at least since 2005.
  123: Initially Xen was PV only.
  124: 
  125: NetBSD Xen has always supported PV, in both dom0 and domU; for a long
  126: time this was the only way.  NetBSD >=8 as a dom0 supports HVM mode in
  127: domUs.
  128: 
  129: Support for PVHVM and PVH is available only in NetBSD-current; this is
  130: currently somewhat experimental, although PVHVM appears reasonably
  131: solid.
  132: 
  133: NetBSD up to and including NetBSD 9 as a dom0 cannot safely run SMP.
  134: Even if one added "options MULTIPROCESSOR" and configured multiple
  135: vcpus, the kernel is likely to crash because of drivers without
  136: adequate locking.
  137: 
  138: NetBSD-current supports SMP in dom0, and XEN3_DOM0 includes "options
  139: MULTIPROCESSOR".
  140: 
  141: NetBSD (since NetBSD 6), when run as a domU, can run SMP, using
  142: multiple CPUs if provided.  The XEN3_DOMU kernel is built
  143: with "options MULITPROCESSOR".
  144: 
  145: Note that while Xen 4.13 is current, the kernel support is still
  146: called XEN3, because the hypercall interface has not changed
  147: significantly.
  148: 
  149: # Creating a NetBSD dom0
  150: 
  151: In order to install a NetBSD as a dom0, one first installs a normal
  152: NetBSD system, and then pivot the install to a dom0 install by
  153: changing the kernel and boot configuration.
  154: 
  155: NB: As of 2021-04, you must arrange to have the system use BIOS boot,
  156: not EFI boot.  (Perhaps, with recent current and using multiboot
  157: instead of NetBSD boot, this might not be true.)
  158: 
  159: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
  160: 512M it worked reliably.  This does not make sense, but if you see
  161: "not ELF" after Xen boots, try increasing dom0 RAM.
  162: 
  163: ## Installation of NetBSD
  164: 
  165: [Install NetBSD/amd64](/guide/inst/) just as you would if you were not
  166: using Xen.  Therefore, use the most recent release, or a build from
  167: the most recent stable branch.  Alternatively, use -current, being
  168: mindful of all the usual caveats of lower stability of current, and
  169: likely a bit more so.  Think about how you will provide storage for
  170: disk images.
  171: 
  172: ## Installation of Xen
  173: 
  174: ### Building Xen
  175: 
  176: Use the most recent version of Xen in pkgsrc, unless the DESCR says
  177: that it is not suitable.  Therefore, choose 4.13.  In the dom0,
  178: install xenkernel413 and xentools413 from pkgsrc.
  179: 
  180: Once this is done, copy the Xen kernel from where pkgsrc puts it to
  181: where the boot process will be able to find it:
  182: 
  183: [[!template id=programlisting text="""
  184: # cp -p /usr/pkg/xen413-kernel/xen.gz /
  185: """]]
  186: 
  187: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such
  188: kernel can either be taken from a local release build.sh run, compiled
  189: manually, or downloaded from the NetBSD FTP, for example at:
  190: 
  191: [[!template id=programlisting text="""
  192: ftp.netbsd.org/pub/NetBSD/NetBSD-9.1/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
  193: """]]
  194: 
  195: ### Configuring booting
  196: 
  197: Read boot.cfg(8) carefully.  Add lines to /boot.cfg to boot Xen,
  198: adjusting for your root filesystem:
  199: 
  200: [[!template id=filecontent name="/boot.cfg" text="""
  201: menu=Xen:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc;multiboot /xen.gz dom0_mem=512M
  202: menu=Xen single user:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc -s;multiboot /xen.gz dom0_mem=512M
  203: """]]
  204: 
  205: This specifies that the dom0 should have 512MB of ram, leaving the rest
  206: to be allocated for domUs.
  207: 
  208: NB: This says add, not replace, so that you will be able to more
  209: easily boot a NetBSD kernel without Xen.  Once Xen boots ok, you may
  210: want to set it as default.  It is highly likely that you will have
  211: trouble at some point, and keeping an up-to-date GENERIC for use in
  212: fixing problems is the standard prudent approach.
  213: 
  214: \todo Explain why rndseed is not set with Xen as part of the dom0
  215: subconfiguration.
  216: 
  217: Note that you are likely to have to set root= because the boot device
  218: from /boot is not passed via Xen to the dom0 kernel.  With one disk,
  219: it will work, but e.g. plugging in USB disk to a machine with root on
  220: wd0a causes boot to fail.
  221: 
  222: Beware that userconf statements must be attached to the dom0 load, and
  223: may not be at top-level, because then they would try to configure the
  224: hypervisor, if there is a way to pass them via multiboot.  It appears
  225: that adding `userconf=pckbc` to `/boot.cfg` causes Xen to crash very
  226: early with a heap overflow.
  227: 
  228: ### Console selection
  229: 
  230: See boot_console(8).  Understand that you should start from a place of
  231: having console setup correct for booting GENERIC before trying to
  232: configure Xen.
  233: 
  234: Generally for GENERIC, one sets the console in bootxx_ffsv1 or
  235: equivalent, and this is passed on to /boot (where one typically does
  236: not set the console).  This configuration of bootxx_ffsv1 should also
  237: be in place for Xen systems, to allow seeing messages from /boot and
  238: use of a keyboard to select a line from the menu.  And, one should
  239: have a working boot path to GENERIC for rescue situations.
  240: 
  241: With GENERIC, the boot options are passed on to /netbsd, but there is
  242: currently no mechanism to pass these via multiboot to the hypervisor.
  243: Thus, in addition to configuring the console in the boot blocks, one
  244: must also configure it for Xen.
  245: 
  246: By default, the hypervisor (Xen itself) will use some sort of vga
  247: device as the console, much like GENERIC uses by default.  The vga
  248: console is relinquished at the conclusion of hypervisor boot, before
  249: the dom0 is started.  Xen when using a vga console does not process
  250: console input.
  251: 
  252: The hypervisor can be configured to use a serial port console, e.g.
  253: [[!template id=filecontent name="/boot.cfg" text="""
  254: menu=Xen:load /netbsd-XEN3_DOM0.gz console=com0;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
  255: """]]
  256: This example uses the first serial port (Xen counts from 1; this is
  257: what NetBSD would call com0), and sets speed and parity.  (The dom0 is
  258: then configured to use the same serial port in this example.)
  259: 
  260: With the hypervisor configured for a serial console, it can get input,
  261: and there is a notion of passing this input to the dom0.  \todo
  262: Explain why, if Xen has a serial console, the dom0 console is
  263: typically also configured to open that same serial port, instead of
  264: getting the passthrough input via the xen console.
  265: 
  266: One also configures the console for the dom0.  While one might expect
  267: console=pc to be default, following behavior of GENERIC, a hasty read
  268: of the code suggests there is no default and booting without a
  269: selected console might lead to a panic.  Also, there is merit in
  270: explicit configuration.  Therefore the standard approach is to place
  271: console=pc as part of the load statement for the dom0 kernel, or
  272: alternatively console=com0.
  273: 
  274: The NetBSD dom0 kernel will attach xencons(4) (the man page does not
  275: exist), but this is not used as a console.  It is used to obtain the
  276: messages from the hypervisor's console; run `xl dmesg` to see them.
  277: 
  278: ### Tuning
  279: 
  280: In an attempt to add performance, one can also add `dom0_max_vcpus=1
  281: dom0_vcpus_pin`, to force only one vcpu to be provided (since NetBSD
  282: dom0 can't use more) and to pin that vcpu to a physical CPU. Xen has
  283: [many boot
  284: options](http://xenbits.xenproject.org/docs/4.13-testing/misc/xen-command-line.html),
  285: and other than dom0 memory and max_vcpus, they are generally not
  286: necessary.
  287: 
  288: \todo Revisit this advice with current.
  289: \todo Explain if anyone has ever actually measured that this helps.
  290: 
  291: ### rc.conf
  292: 
  293: Ensure that the boot scripts installed in
  294: `/usr/pkg/share/examples/rc.d` are in `/etc/rc.d`, either because you
  295: have `PKG_RCD_SCRIPTS=yes`, or manually.  (This is not special to Xen,
  296: but a normal part of pkgsrc usage.)
  297: 
  298: Set `xencommons=YES` in rc.conf:
  299: 
  300: [[!template id=filecontent name="/etc/rc.conf" text="""
  301: xencommons=YES
  302: """]]
  303: 
  304: \todo Recommend for/against xen-watchdog.
  305: 
  306: ### Testing
  307: 
  308: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  309: than GENERIC without Xen.
  310: 
  311: Once the reboot is done, use `xl` to inspect Xen's boot messages,
  312: available resources, and running domains.  For example:
  313: 
  314: [[!template id=programlisting text="""
  315: # xl dmesg
  316: ... xen's boot info ...
  317: # xl info
  318: ... available memory, etc ...
  319: # xl list
  320: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  321: Domain-0           0       64    0  r----     58.1
  322: """]]
  323: 
  324: Xen logs will be in /var/log/xen.
  325: 
  326: ### Issues with xencommons
  327: 
  328: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
  329: domUs.  It does not currently work to stop and start xenstored.
  330: Certainly all domUs should be shutdown first, following the sort order
  331: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
  332: and is not notified when xenstored exits, leading to not recreating
  333: the state when the new xenstored starts.  Until there's a mechanism to
  334: make this work, one should not expect to be able to restart xenstored
  335: (and thus xencommons).  There is currently no reason to expect that
  336: this will get fixed any time soon.
  337: \todo Confirm if this is still true in 2020.
  338: 
  339: ## Xen-specific NetBSD issues
  340: 
  341: There are (at least) two additional things different about NetBSD as a
  342: dom0 kernel compared to hardware.
  343: 
  344: One is that through NetBSD 9 the module ABI is different because some
  345: of the #defines change, so there are separate sets of modules in
  346: /stand.  (Further, zfs in Xen is troubled because of differing
  347: MAXPHYS; see the zfs howto for more.)  In NetBSD-current, there is
  348: only one set of modules.
  349: 
  350: The other difference is that XEN3_DOM0 does not have exactly the same
  351: options as GENERIC.  While this is roughly agreed to be in large part
  352: a bug, users should be aware of this and can simply add missing config
  353: items if desired.
  354: 
  355: Finally, there have been occasional reports of trouble with X11
  356: servers in NetBSD as a dom0.  Some hardware support is intentionally
  357: disabled in XEN3_DOM0.
  358: 
  359: ## Updating Xen in a dom0
  360: 
  361: Note the previous advice to maintain a working and tested boot config
  362: into GENERIC without Xen.
  363: 
  364: Updating Xen in a dom0 consists of updating the xnekernel and xentools
  365: packages, along with copying the xen.gz into place, and of course
  366: rebooting.
  367: 
  368: If updating along a Xen minor version, e.g. from 4.13.1 to 4.13.2, or
  369: from 4.13.2nb1 to 4.13.2nb3, it is very likely that this can be done
  370: on a running system.  The point is that the xentools programs will be
  371: replaced, and you will be using "xl" from the new installation to talk
  372: to the older programs which are still running.  Problems from this
  373: update path should be reported.
  374: 
  375: For added safety, shutdown all domUs before updating, to remove the
  376: need for new xl to talk to old xenstored.  Note that Xen does not
  377: guarantee stability of internal ABIs.
  378: 
  379: If updating across Xen minor versions, e.g. from 4.11 to 4.13, the
  380: likelihood of trouble is increased.  Therefore, 'make replace' of
  381: xentools on a dom0 with running domUs is not recommended.  A shutdown
  382: on all domUs before replacing xentools is likely sufficient.  A safer
  383: appraoch is to boot into GENERIC to replace the packages, as then no
  384: Xen code will be running.  Single user is another option.
  385: 
  386: ## Updating NetBSD in a dom0
  387: 
  388: This is just like updating NetBSD on bare hardware, assuming the new
  389: version supports the version of Xen you are running.  Generally, one
  390: replaces the kernel and reboots, and then overlays userland binaries
  391: and adjusts `/etc`.
  392: 
  393: Note that one should update both the non-Xen kernel typically used for
  394: rescue purposes, as well as the DOM0 kernel used with Xen.
  395: 
  396: ## anita (for testing NetBSD)
  397: 
  398: With a NetBSD dom0, even without any domUs, one can run anita (see
  399: pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as root,
  400: because anita must create a domU):
  401: 
  402: [[!template id=programlisting text="""
  403: anita --vmm=xl test file:///usr/obj/i386/
  404: """]]
  405: 
  406: # Unprivileged domains (domU)
  407: 
  408: This section describes general concepts about domUs.  It does not
  409: address specific domU operating systems or how to install them.  The
  410: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
  411: typically named so that the file name, domU name and the domU's host
  412: name match.
  413: 
  414: The domU is provided with CPU and memory by Xen, configured by the
  415: dom0.  The domU is provided with disk and network by the dom0,
  416: mediated by Xen, and configured in the dom0.
  417: 
  418: Entropy in domUs can be an issue; physical disks and network are on
  419: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  420: 
  421: ## Config files
  422: 
  423: See /usr/pkg/share/examples/xen/xlexample* for a very small number of
  424: examples for running GNU/Linux.
  425: 
  426: The following is an example minimal domain configuration file.  The
  427: domU serves as a network file server.
  428: 
  429: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  430: name = "domU-id"
  431: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  432: memory = 1024
  433: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  434: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  435:          'file:/n0/xen/foo-wd1,0x1,w' ]
  436: """]]
  437: 
  438: The domain will have name given in the `name` setting.  The kernel has the
  439: host/domU name in it, so that on the dom0 one can update the various
  440: domUs independently.  The `vif` line causes an interface to be provided,
  441: with a specific mac address (do not reuse MAC addresses!), in bridge
  442: mode.  Two disks are provided, and they are both writable; the bits
  443: are stored in files and Xen attaches them to a vnd(4) device in the
  444: dom0 on domain creation.  The system treats xbd0 as the boot device
  445: without needing explicit configuration.
  446: 
  447: There is not a type line; that implicitly defines a pv domU.
  448: Otherwise, one sets type to the lower-case version of the domU type in
  449: the table above; see later sections.
  450: 
  451: By convention, domain config files are kept in `/usr/pkg/etc/xen`.  Note
  452: that "xl create" takes the name of a config file, while other commands
  453: take the name of a domain.
  454: 
  455: Examples of commands:
  456: 
  457: [[!template id=programlisting text="""
  458: xl create /usr/pkg/etc/xen/foo
  459: xl console domU-id
  460: xl create -c /usr/pkg/etc/xen/foo
  461: xl shutdown domU-id
  462: xl list
  463: """]]
  464: 
  465: Typing `^]` will exit the console session.  Shutting down a domain is
  466: equivalent to pushing the power button; a NetBSD domU will receive a
  467: power-press event and do a clean shutdown.  Shutting down the dom0
  468: will trigger controlled shutdowns of all configured domUs.
  469: 
  470: ## CPU and memory
  471: 
  472: A domain is provided with some number of vcpus; any domain can have up
  473: to the number of CPUs seen by the hypervisor. For a domU, it is
  474: controlled from the config file by the "vcpus = N" directive.  It is
  475: normal to overcommit vcpus; a 4-core machine machine might well provide 4
  476: vcpus to each domU.  One might also configure fewer vcpus for a domU.
  477: 
  478: A domain is provided with memory; this is controlled in the config
  479: file by "memory = N" (in megabytes).  In the straightforward case, the
  480: sum of the the memory allocated to the dom0 and all domUs must be less
  481: than the available memory.
  482: 
  483: ## Balloon driver
  484: 
  485: Xen provides a `balloon` driver, which can be used to let domains use
  486: more memory temporarily.
  487: 
  488: \todo Explain how to set up a aystem to use the balloon scheme in a
  489: useful manner.
  490: 
  491: ## Virtual disks
  492: 
  493: In domU config files, the disks are defined as a sequence of 3-tuples:
  494: 
  495:  * The first element is "method:/path/to/disk". Common methods are
  496:    "file:" for a file-backed vnd, and "phy:" for something that is already
  497:    a device, such as an LVM logical volume.
  498: 
  499:  * The second element is an artifact of how virtual disks are passed to
  500:    Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  501:    are given a device name to associate with the disk, and values like
  502:    "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  503:    as xbd0, the second as xbd1, and so on.  However, xl demands a
  504:    second argument.  The name given is converted to a major/minor by
  505:    calling stat(2) on the name in /dev and this is passed to the domU.
  506:    In the general case, the dom0 and domU can be different operating
  507:    systems, and it is an unwarranted assumption that they have consistent
  508:    numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  509:    as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  510:    for the second works fine and avoids this issue.  For a GNU/Linux
  511:    guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  512:    /dev/hda1.
  513: 
  514:  * The third element is "w" for writable disks, and "r" for read-only
  515:    disks.
  516: 
  517: Example:
  518: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  519: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ]
  520: """]]
  521: 
  522: Note that NetBSD by default creates only vnd[0123].  If you need more
  523: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
  524: dom0.
  525: 
  526: ## Virtual Networking
  527: 
  528: Xen provides virtual Ethernets, each of which connects the dom0 and a
  529: domU.  For each virtual network, there is an interface "xvifN.M" in
  530: the dom0, and a matching interface xennetM (NetBSD name) in domU index N.
  531: The interfaces behave as if there is an Ethernet with two
  532: adapters connected.  From this primitive, one can construct various
  533: configurations.  We focus on two common and useful cases for which
  534: there are existing scripts: bridging and NAT.
  535: 
  536: With bridging (in the example above), the domU perceives itself to be
  537: on the same network as the dom0.  For server virtualization, this is
  538: usually best.  Bridging is accomplished by creating a bridge(4) device
  539: and adding the dom0's physical interface and the various xvifN.0
  540: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  541: config file.  The bridge must be set up already in the dom0; an
  542: example /etc/ifconfig.bridge0 is:
  543: 
  544: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
  545: create
  546: up
  547: !brconfig bridge0 add wm0
  548: """]]
  549: 
  550: With NAT, the domU perceives itself to be behind a NAT running on the
  551: dom0.  This is often appropriate when running Xen on a workstation.
  552: TODO: NAT appears to be configured by "vif = [ '' ]".
  553: 
  554: The MAC address specified is the one used for the interface in the new
  555: domain.  The interface in dom0 will use this address XOR'd with
  556: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  557: 
  558: ## Starting domains automatically
  559: 
  560: To start domains `domU-netbsd` and `domU-linux` at boot and shut them
  561: down cleanly on dom0 shutdown, add the following in rc.conf:
  562: 
  563: [[!template id=filecontent name="/etc/rc.conf" text="""
  564: xendomains="domU-netbsd domU-linux"
  565: """]]
  566: 
  567: # domU setup for specific systems
  568: 
  569: Creating domUs is almost entirely independent of operating system.  We
  570: have already presented the basics of config files in the previous system.
  571: 
  572: Of course, this section presumes that you have a working dom0.
  573: 
  574: ## Creating a NetBSD PV domU
  575: 
  576: See the earlier config file, and adjust memory.  Decide on how much
  577: storage you will provide, and prepare it (file or LVM).
  578: 
  579: While the kernel will be obtained from the dom0 file system, the same
  580: file should be present in the domU as /netbsd so that tools like
  581: savecore(8) can work.   (This is helpful but not necessary.)
  582: 
  583: The kernel must be specifically built for Xen, to use PV interfacesas
  584: a domU.  NetBSD release builds provide the following kernels:
  585: 
  586:         i386 XEN3PAE_DOMU
  587:         amd64 XEN3_DOMU
  588: 
  589: This will boot NetBSD, but this is not that useful if the disk is
  590: empty.  One approach is to unpack sets onto the disk outside of Xen
  591: (by mounting it, just as you would prepare a physical disk for a
  592: system you can't run the installer on).
  593: 
  594: A second approach is to run an INSTALL kernel, which has a miniroot
  595: and can load sets from the network.  To do this, copy the INSTALL
  596: kernel to / and change the kernel line in the config file to:
  597: 
  598:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  599: 
  600: Then, start the domain as "xl create -c configfile".
  601: 
  602: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  603: line should be used in the config file.
  604: 
  605:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  606: 
  607: After booting the domain, the option to install via CDROM may be
  608: selected.  The CDROM device should be changed to `xbd1d`.
  609: 
  610: Once done installing, "halt -p" the new domain (don't reboot or halt:
  611: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  612: config file), switch the config file back to the XEN3_DOMU kernel,
  613: and start the new domain again. Now it should be able to use "root on
  614: xbd0a" and you should have a functional NetBSD domU.
  615: 
  616: TODO: check if this is still accurate.
  617: When the new domain is booting you'll see some warnings about *wscons*
  618: and the pseudo-terminals. These can be fixed by editing the files
  619: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  620: `/etc/ttys`, except *console*, like this:
  621: 
  622:     console "/usr/libexec/getty Pc"         vt100   on secure
  623:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  624:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  625:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  626:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  627: 
  628: Finally, all screens must be commented out from `/etc/wscons.conf`.
  629: 
  630: One should also run `powerd` in a domU, but this should not need
  631: configuring.  With powerd, the domain will run a controlled shutdown
  632: if `xl shutdown -R` or `xl shutdown -H` is used on the dom0, via
  633: receiving a synthetic `power button pressed` signal.  In 9 and
  634: current, `powerd` is run by default under Xen kernels (or if ACPI is
  635: present), and it can be added to rc.conf if not.
  636: 
  637: It is not strictly necessary to have a kernel (as /netbsd) in the domU
  638: file system.  However, various programs (e.g. netstat) will use that
  639: kernel to look up symbols to read from kernel virtual memory.  If
  640: /netbsd is not the running kernel, those lookups will fail.  (This is
  641: not really a Xen-specific issue, but because the domU kernel is
  642: obtained from the dom0, it is far more likely to be out of sync or
  643: missing with Xen.)
  644: 
  645: Note that NetBSD by default creates only xbd[0123].  If you need more
  646: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
  647: 
  648: ## Creating a Linux PV domU
  649: 
  650: Creating unprivileged Linux domains isn't much different from
  651: unprivileged NetBSD domains, but there are some details to know.
  652: 
  653: First, the second parameter passed to the disk declaration (the '0x1' in
  654: the example below)
  655: 
  656:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  657: 
  658: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  659: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
  660: So, hda1 which has major 3 and minor 1 on a Linux system will have
  661: device number 0x301.  Alternatively, devices names can be used (hda,
  662: hdb, ...)  as xentools has a table to map these names to devices
  663: numbers.  To export a partition to a Linux guest we can use:
  664: 
  665:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
  666:         root = "/dev/hda1 ro"
  667: 
  668: and it will appear as /dev/hda on the Linux system, and be used as root
  669: partition.
  670: 
  671: To install the Linux system on the partition to be exported to the
  672: guest domain, the following method can be used: install
  673: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
  674: that will be the root partition of your Linux domain, and mount it.
  675: Then copy the files from a working Linux system, make adjustments in
  676: `/etc` (fstab, network config).  It should also be possible to extract
  677: binary packages such as .rpm or .deb directly to the mounted partition
  678: using the appropriate tool, possibly running under NetBSD's Linux
  679: emulation.  Once the file system has been populated, umount it.  If
  680: desirable, the file system can be converted to ext3 using tune2fs -j.
  681: It should now be possible to boot the Linux guest domain, using one of
  682: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
  683: 
  684: To get the Linux console right, you need to add:
  685: 
  686:     extra = "xencons=tty1"
  687: 
  688: to your configuration since not all Linux distributions auto-attach a
  689: tty to the xen console.
  690: 
  691: ## Creating a NetBSD HVM domU
  692: 
  693: Use type='hvm', probably.  Use a GENERIC kernel within the disk image.
  694: 
  695: ## Creating a NetBSD PVH domU
  696: 
  697: This only works with a current kernel in the domU.
  698: 
  699: Use type='pvh'.  Probably, use a GENERIC kernel within the disk image,
  700: which in current has PV support.
  701: 
  702: \todo Verify.
  703: 
  704: \todo Verify if one can have current PVH domU on a 9 dom0.
  705: 
  706: ## Creating a Solaris domU
  707: 
  708: See possibly outdated
  709: [Solaris domU instructions](/ports/xen/howto-solaris/).
  710: 
  711: ## PCI passthrough: Using PCI devices in guest domains
  712: 
  713: NB: PCI passthrough only works on some Xen versions and as of 2020 it
  714: is not clear that it works on any version in pkgsrc.  \todo Reports
  715: confirming or denying this notion should be sent to port-xen@.
  716: 
  717: The dom0 can give other domains access to selected PCI
  718: devices. This can allow, for example, a non-privileged domain to have
  719: access to a physical network interface or disk controller.  However,
  720: keep in mind that giving a domain access to a PCI device most likely
  721: will give the domain read/write access to the whole physical memory,
  722: as PCs don't have an IOMMU to restrict memory access to DMA-capable
  723: device.  Also, it's not possible to export ISA devices to non-dom0
  724: domains, which means that the primary VGA adapter can't be exported.
  725: A guest domain trying to access the VGA registers will panic.
  726: 
  727: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
  728: not been ported to later versions at this time.
  729: 
  730: For a PCI device to be exported to a domU, is has to be attached to
  731: the "pciback" driver in dom0.  Devices passed to the dom0 via the
  732: pciback.hide boot parameter will attach to "pciback" instead of the
  733: usual driver.  The list of devices is specified as "(bus:dev.func)",
  734: where bus and dev are 2-digit hexadecimal numbers, and func a
  735: single-digit number:
  736: 
  737:         pciback.hide=(00:0a.0)(00:06.0)
  738: 
  739: pciback devices should show up in the dom0's boot messages, and the
  740: devices should be listed in the `/kern/xen/pci` directory.
  741: 
  742: PCI devices to be exported to a domU are listed in the "pci" array of
  743: the domU's config file, with the format "0000:bus:dev.func".
  744: 
  745:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
  746: 
  747: In the domU an "xpci" device will show up, to which one or more pci
  748: buses will attach.  Then the PCI drivers will attach to PCI buses as
  749: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
  750: or any PCI drivers built in by default; you have to build your own
  751: kernel to use PCI devices in a domU.  Here's a kernel config example;
  752: note that only the "xpci" lines are unusual.
  753: 
  754:         include         "arch/i386/conf/XEN3_DOMU"
  755: 
  756:         # Add support for PCI buses to the XEN3_DOMU kernel
  757:         xpci* at xenbus ?
  758:         pci* at xpci ?
  759: 
  760:         # PCI USB controllers
  761:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
  762: 
  763:         # USB bus support
  764:         usb*    at uhci?
  765: 
  766:         # USB Hubs
  767:         uhub*   at usb?
  768:         uhub*   at uhub? port ? configuration ? interface ?
  769: 
  770:         # USB Mass Storage
  771:         umass*  at uhub? port ? configuration ? interface ?
  772:         wd*     at umass?
  773:         # SCSI controllers
  774:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
  775: 
  776:         # SCSI bus support (for both ahc and umass)
  777:         scsibus* at scsi?
  778: 
  779:         # SCSI devices
  780:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
  781:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
  782: 
  783: 
  784: # Miscellaneous Information
  785: 
  786: ## Nesting under Linux KVM
  787: 
  788: It is possible to run Xen and a NetBSD dom0 under Linux KVM.  One
  789: can enable virtio in the dom0 for greater speed.
  790: 
  791: ## Nesting under qemu
  792: 
  793: It is possible to run Xen and a NetBSD dom0 under qemu on NetBSD, and
  794: also with nvmm.
  795: \todo Check this.
  796: 
  797: ## Other nesting
  798: 
  799: In theory, any full emulation should be able to run Xen and a NetBSD
  800: dom0.  The HOWTO does not currently have information about Xen XVM
  801: mode, Virtualbox, etc.
  802: 
  803: ## NetBSD 5 as domU
  804: 
  805: [NetBSD 5 is known to panic.](http://mail-index.netbsd.org/port-xen/2018/04/17/msg009181.html)
  806: (However, NetBSD 5 systems should be updated to a supported version.)
  807: 
  808: # NetBSD as a domU in a VPS
  809: 
  810: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
  811: hardware.  This section explains how to deal with Xen in a domU as a
  812: virtual private server where you do not control or have access to the
  813: dom0.  This is not intended to be an exhaustive list of VPS providers;
  814: only a few are mentioned that specifically support NetBSD.
  815: 
  816: VPS operators provide varying degrees of access and mechanisms for
  817: configuration.  The big issue is usually how one controls which kernel
  818: is booted, because the kernel is nominally in the dom0 file system (to
  819: which VPS users do not normally have access).  A second issue is how
  820: to install NetBSD.
  821: A VPS user may want to compile a kernel for security updates, to run
  822: npf, run IPsec, or any other reason why someone would want to change
  823: their kernel.
  824: 
  825: One approach is to have an administrative interface to upload a kernel,
  826: or to select from a prepopulated list.  Other approaches are pygrub
  827: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
  828: kernel from the domU file system.  This is closer to a regular physical
  829: computer, where someone who controls a machine can replace the kernel.
  830: 
  831: A second issue is multiple CPUs.  With NetBSD 6, domUs support
  832: multiple vcpus, and it is typical for VPS providers to enable multiple
  833: CPUs for NetBSD domUs.
  834: 
  835: ## Complexities due to Xen changes
  836: 
  837: Xen has many security advisories and people running Xen systems make
  838: different choices.
  839: 
  840: ### stub domains
  841: 
  842: Some (Linux) dom0 systems use something called "stub domains" to
  843: isolate qemu from the dom0 system, as a security and reliabilty
  844: mechanism when running HVM domUs.  Somehow, NetBSD's GENERIC kernel
  845: ends up using PIO for disks rather than DMA.  Of course, all of this
  846: is emulated, but emulated PIO is unusably slow.  This problem is not
  847: currently understood.
  848: 
  849: ### Grant tables
  850: 
  851: There are multiple versions of using grant tables, and some security
  852: advisories have suggested disabling some versions.  NetBSD through 9
  853: uses version 1 and NetBSD-current uses version 2.  This can lead to
  854: "NetBSD current doesn't run on hosting provider X" situations.
  855: 
  856: \todo Explain better.
  857: 
  858: ## Boot methods
  859: 
  860: ### pvgrub
  861: 
  862: pvgrub is a version of grub that uses PV operations instead of BIOS
  863: calls.  It is booted from the dom0 as the domU kernel, and then reads
  864: /grub/menu.lst and loads a kernel from the domU file system.
  865: 
  866: It appears that [grub's FFS
  867: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
  868: does not support all aspects of modern FFS, but there are also reports
  869: that FFSv2 works fine.
  870: 
  871: ### pygrub
  872: 
  873: As of 2014, pygrub seems to be of mostly historical interest.  As of
  874: 2021, the section should perhaps be outright deleted.
  875: 
  876: pygrub runs in the dom0 and looks into the domU file system.  This
  877: implies that the domU must have a kernel in a file system in a format
  878: known to pygrub.
  879: 
  880: pygrub doesn't seem to work to load Linux images under NetBSD dom0,
  881: and is inherently less secure than pvgrub due to running inside
  882: dom0. For both these reasons, pygrub should not be used, and is only
  883: still present so that historical DomU images using it still work.
  884: 
  885: ## Specific Providers
  886: 
  887: The intent is to list providers only if they document support for
  888: running NetBSD, and to point to their resources briefly.
  889: 
  890: ### panix.com
  891: 
  892: [Panix](http://www.panix.com/) provides NetBSD as an OS option.  See
  893: https://www.panix.com/v-colo/nupgrade.html for some information.
  894: Users can use pvgrub.  Panix reports that pvgrub works with FFsv2 with
  895: 16K/2K and 32K/4K block/frag sizes (and hence with defaults from
  896: "newfs -O 2").  See [Panix's pvgrub
  897: page](http://www.panix.com/v-colo/grub.html) which describes how to
  898: boot NetBSD.
  899: 
  900: ### prgmr.com
  901: 
  902: [prgmr.com](http://prgmr.com/) provides released versions of
  903: NetBSD/amd64 as installation options.  Users can use pvgrub to boot
  904: their own kernel, and a small FAT32 /boot is encouraged.  See the
  905: [prgmr.com NetBSD
  906: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
  907: (which is in need of updating).
  908: 
  909: ### Amazon
  910: 
  911: See the [Amazon EC2 page](/amazon_ec2/).

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