Annotation of wikisrc/ports/xen/howto.mdwn, revision 1.200

1.144     maxv        1: [[!meta title="Xen HowTo"]]
                      2: 
1.157     maxv        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.
1.12      gdt        10: 
1.182     gdt        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: 
1.183     gdt        18: # Overview
1.182     gdt        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: 
1.186     gdt        30: There are many choices one can make; the HOWTO recommends the standard
                     31: approach and limits discussion of alternatives in many cases.
                     32: 
1.182     gdt        33: ## Guest Styles
                     34: 
                     35: Xen supports different styles of guests.
1.149     maxv       36: 
                     37: [[!table data="""
                     38: Style of guest |Supported by NetBSD
1.150     maxv       39: PV             |Yes (dom0, domU)
                     40: HVM            |Yes (domU)
1.171     gdt        41: PVHVM          |current-only (domU)
1.180     gdt        42: PVH            |current-only (domU, dom0 not yet)
1.149     maxv       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
1.170     gdt        47: guests must be specifically coded for Xen.
1.182     gdt        48: See [PV](https://wiki.xen.org/wiki/Paravirtualization_(PV\)).
1.170     gdt        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.
1.199     gdt        52: The dom0 runs qemu to emulate hardware.  It is therefore non-sensical
                     53: to have an HVM dom0.
1.170     gdt        54: 
1.176     gdt        55: In PVHVM mode, the guest runs as HVM, but additionally can use PV
1.199     gdt        56: drivers for efficiency.  Therefore it is non-sensical for to have a
                     57: PVHVM dom0.  See [PV on HVM](https://wiki.xen.org/wiki/PV_on_HVM).
1.170     gdt        58: 
1.180     gdt        59: There have been two PVH modes: original PVH and PVHv2.  Original PVH
                     60: was based on PV mode and is no longer relevant at all.  PVHv2 is
                     61: basically lightweight HVM with PV drivers.  A critical feature of it
                     62: is that qemu is not needed; the hypervisor can do the emulation that
                     63: is required.  Thus, a dom0 can be PVHv2.
1.182     gdt        64: The source code uses PVH and config files use pvh; this refers to PVHv2.
                     65: See [PVH(v2)](https://wiki.xenproject.org/wiki/PVH_(v2\)_Domu).
1.180     gdt        66: 
1.182     gdt        67: At system boot, the dom0 kernel is loaded as a module with Xen as the kernel.
1.12      gdt        68: The dom0 can start one or more domUs.  (Booting is explained in detail
                     69: in the dom0 section.)
                     70: 
1.182     gdt        71: ## CPU Architecture
                     72: 
                     73: Xen runs on x86_64 hardware (the NetBSD amd64 port).
                     74: 
                     75: There is a concept of Xen running on ARM, but there are no reports of this working with NetBSD.
                     76: 
                     77: The dom0 system should be amd64.  (Instructions for i386PAE dom0 have been removed from the HOWTO.)
1.1       mspo       78: 
1.182     gdt        79: The domU can be i386PAE or amd64.
                     80: i386PAE at one point was considered as [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html) than amd64.
1.150     maxv       81: 
1.182     gdt        82: ## Xen Versions
1.15      gdt        83: 
1.111     wiz        84: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
1.15      gdt        85: xenkernel and xentools.  We will refer only to the kernel versions,
                     86: but note that both packages must be installed together and must have
                     87: matching versions.
                     88: 
1.145     maxv       89: Versions available in pkgsrc:
1.85      gdt        90: 
1.145     maxv       91: [[!table data="""
1.184     gdt        92: Xen Version    |Package Name   |Xen CPU Support        |EOL'ed By Upstream
                     93: 4.11           |xenkernel411   |x86_64                 |No
                     94: 4.13           |xenkernel413   |x86_64                 |No
1.145     maxv       95: """]]
1.113     gdt        96: 
1.96      gdt        97: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
                     98: 
1.166     gdt        99: Older Xen had a python-based management tool called xm, now replaced
1.176     gdt       100: by xl.
1.166     gdt       101: 
1.183     gdt       102: ## NetBSD versions
                    103: 
                    104: Xen has been supported in NetBSD for a long time, at least since 2005.
                    105: Initially Xen was PV only.
                    106: 
                    107: NetBSD 8 and up support PV and HVM modes.
                    108: 
                    109: Support for PVHVM and PVH is available only in NetBSD-current.
                    110: 
1.191     gdt       111: NetBSD up to and including NetBSD 9 as a dom0 does not run SMP,
                    112: because some drivers are not yet safe for this.  NetBSD-current
                    113: supports SMP in dom0.
1.185     gdt       114: 
                    115: NetBSD, when run as a domU, can and does typically run SMP.
                    116: 
1.191     gdt       117: Note that while Xen 4.13 is current, the kernel support is still
                    118: called XEN3, because the hypercall interface has not changed
                    119: significantly.
1.185     gdt       120: 
1.187     gdt       121: # Creating a NetBSD dom0
1.15      gdt       122: 
1.191     gdt       123: In order to install a NetBSD as a dom0, one first installs a normal
                    124: NetBSD system, and then pivot the install to a dom0 install by
                    125: changing the kernel and boot configuration.
1.15      gdt       126: 
1.142     gdt       127: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
                    128: 512M it worked reliably.  This does not make sense, but if you see
                    129: "not ELF" after Xen boots, try increasing dom0 RAM.
                    130: 
1.181     gdt       131: ## Installation of NetBSD
1.13      gdt       132: 
1.191     gdt       133: [Install NetBSD/amd64](/guide/inst/) just as you would if you were not
                    134: using Xen.  Therefore, use the most recent release, or a build from
                    135: the most recent stable branch.  Alternatively, use -current, being
                    136: mindful of all the usual caveats of lower stability of current, and
                    137: likely a bit more so.
1.1       mspo      138: 
1.181     gdt       139: ## Installation of Xen
1.1       mspo      140: 
1.191     gdt       141: ### Building Xen
                    142: 
1.187     gdt       143: Use the most recent version of Xen in pkgsrc, unless the DESCR says that it is not suitable.
1.186     gdt       144: Therefore, choose 4.13.
                    145: In the dom0, install xenkernel413 and xentools413 from pkgsrc.
1.155     maxv      146: 
1.186     gdt       147: Once this is done, copy the Xen kernel from where pkgsrc puts it to
                    148: where the boot process will be able to find it:
1.155     maxv      149: 
                    150: [[!template id=programlisting text="""
1.186     gdt       151: # cp -p /usr/pkg/xen413-kernel/xen.gz /
1.155     maxv      152: """]]
                    153: 
1.191     gdt       154: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such
                    155: kernel can either be taken from a local release build.sh run, compiled
                    156: manually, or downloaded from the NetBSD FTP, for example at:
1.159     maxv      157: 
                    158: [[!template id=programlisting text="""
1.186     gdt       159: ftp.netbsd.org/pub/NetBSD/NetBSD-9.1/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
1.159     maxv      160: """]]
1.20      gdt       161: 
1.191     gdt       162: ### Configuring booting
                    163: 
1.198     gdt       164: Read boot.cfg(8) carefully.  Add lines to /boot.cfg to boot Xen,
                    165: adjusting for your root filesystem:
1.20      gdt       166: 
1.157     maxv      167: [[!template id=filecontent name="/boot.cfg" text="""
1.198     gdt       168: menu=Xen:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc;multiboot /xen.gz dom0_mem=512M
                    169: menu=Xen single user:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc -s;multiboot /xen.gz dom0_mem=512M
1.152     maxv      170: """]]
1.20      gdt       171: 
1.159     maxv      172: This specifies that the dom0 should have 512MB of ram, leaving the rest
1.187     gdt       173: to be allocated for domUs.
                    174: 
1.191     gdt       175: NB: This says add, not replace, so that you will be able to more
                    176: easily boot a NetBSD kernel without Xen.  Once Xen boots ok, you may
                    177: want to set it as default.  It is highly likely that you will have
                    178: trouble at some point, and keeping an up-to-date GENERIC for use in
                    179: fixing problems is the standard prudent approach.
                    180: 
1.194     gdt       181: \todo Explain why rndseed is not set with Xen as part of the dom0
                    182: subconfiguration.
                    183: 
1.198     gdt       184: Note that you are likely to have to set root= because the boot device
                    185: from /boot is not passed via Xen to the dom0 kernel.  With one disk,
                    186: it will work, but e.g. plugging in USB disk to a machine with root on
                    187: wd0a causes boot to fail.
                    188: 
1.195     gdt       189: Beware that userconf statements must be attached to the dom0 load, and
                    190: may not be at top-level, because then they would try to configure the
                    191: hypervisor, if there is a way to pass them via multiboot .  It appears
                    192: that adding `userconf=pckbc` to `/boot.cfg` causes Xen to crash very
                    193: early with a heap overflow.
                    194: 
1.191     gdt       195: ### Console selection
                    196: 
                    197: See boot_console(8).  Understand that you should start from a place of
                    198: having console setup correct for booting GENERIC before trying to
                    199: configure Xen.
                    200: 
1.193     gdt       201: Generally for GENERIC, one sets the console in bootxx_ffsv1 or
                    202: equivalent, and this is passed on to /boot (where one typically does
                    203: not set the console).  This configuration of bootxx_ffsv1 should also
                    204: be in place for Xen systems, to allow seeing messages from /boot and
                    205: use of a keyboard to select a line from the menu.  And, one should
                    206: have a working boot path to GENERIC for rescue situations.
                    207: 
                    208: With GENERIC, the boot options are passed on to /netbsd, but there is
                    209: currently no mechanism to pass these via multiboot to the hypervisor.
                    210: Thus, in addition to configuring the console in the boot blocks, one
                    211: must also configure it for Xen.
                    212: 
                    213: By default, the hypervisor (Xen itself) will use some sort of vga
                    214: device as the console, much like GENERIC uses by default.  The vga
                    215: console is relinquished at the conclusion of hypervisor boot, before
1.194     gdt       216: the dom0 is started.  Xen when using a vga console does not process
                    217: console input.
1.77      gdt       218: 
1.193     gdt       219: The hypervisor can be configured to use a serial port console, e.g.
1.157     maxv      220: [[!template id=filecontent name="/boot.cfg" text="""
1.193     gdt       221: menu=Xen:losad /netbsd-XEN3_DOM0.gz console=com0;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
1.152     maxv      222: """]]
1.194     gdt       223: This example uses the first serial port (Xen counts from 1; this is
1.193     gdt       224: what NetBSD would call com0), and sets speed and parity.  (The dom0 is
                    225: then configured to use the same serial port in this example.)
                    226: 
1.194     gdt       227: With the hypervisor configured for a serial console, it can get input,
                    228: and there is a notion of passing this input to the dom0.  \todo
                    229: Explain why, if Xen has a serial console, the dom0 console is
                    230: typically also configured to open that same serial port, instead of
                    231: getting the passthrough input via the xen console.
                    232: 
1.193     gdt       233: One also configures the console for the dom0.  While one might expect
                    234: console=pc to be default, following behavior of GENERIC, a hasty read
                    235: of the code suggests there is no default and booting without a
                    236: selected console might lead to a panic.  Also, there is merit in
                    237: explicit configuration.  Therefore the standard approach is to place
                    238: console=pc as part of the load statement for the dom0 kernel, or
                    239: alternatively console=com0.
                    240: 
                    241: The NetBSD dom0 kernel will attach xencons(4) (the man page does not
                    242: exist), but this is not used as a console.  It is used to obtain the
                    243: messages from the hypervisor's console; run `xl dmesg` to see them.
1.191     gdt       244: 
                    245: ### Tuning
                    246: 
1.157     maxv      247: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`,
1.37      gdt       248: to force only one vcpu to be provided (since NetBSD dom0 can't use
1.158     maxv      249: more) and to pin that vcpu to a physical CPU. Xen has
1.178     gdt       250: [many boot options](http://xenbits.xenproject.org/docs/4.13-testing/misc/xen-command-line.html),
1.111     wiz       251: and other than dom0 memory and max_vcpus, they are generally not
1.93      gdt       252: necessary.
1.191     gdt       253: \todo Revisit this advice with current.
                    254: \todo Explain if anyone has ever actually measured that this helps.
                    255: 
                    256: ### rc.conf
1.93      gdt       257: 
1.186     gdt       258: Ensure that the boot scripts installed in
                    259: `/usr/pkg/share/examples/rc.d` are in `/etc/rc.d`, either because you
                    260: have `PKG_RCD_SCRIPTS=yes`, or manually.  (This is not special to Xen,
                    261: but a normal part of pkgsrc usage.)
1.159     maxv      262: 
1.186     gdt       263: Set `xencommons=YES` in rc.conf:
1.31      gdt       264: 
1.157     maxv      265: [[!template id=filecontent name="/etc/rc.conf" text="""
1.152     maxv      266: xencommons=YES
                    267: """]]
1.31      gdt       268: 
1.187     gdt       269: \todo Recommend for/against xen-watchdog.
                    270: 
1.191     gdt       271: ### Testing
                    272: 
1.158     maxv      273: Now, reboot so that you are running a DOM0 kernel under Xen, rather
                    274: than GENERIC without Xen.
1.157     maxv      275: 
1.158     maxv      276: Once the reboot is done, use `xl` to inspect Xen's boot messages,
1.157     maxv      277: available resources, and running domains.  For example:
1.34      gdt       278: 
1.153     maxv      279: [[!template id=programlisting text="""
                    280: # xl dmesg
                    281: ... xen's boot info ...
                    282: # xl info
                    283: ... available memory, etc ...
                    284: # xl list
                    285: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
                    286: Domain-0           0       64    0  r----     58.1
                    287: """]]
                    288: 
                    289: Xen logs will be in /var/log/xen.
1.33      gdt       290: 
1.88      gdt       291: ### Issues with xencommons
                    292: 
1.157     maxv      293: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
1.88      gdt       294: domUs.  It does not currently work to stop and start xenstored.
                    295: Certainly all domUs should be shutdown first, following the sort order
                    296: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
                    297: and is not notified when xenstored exits, leading to not recreating
                    298: the state when the new xenstored starts.  Until there's a mechanism to
                    299: make this work, one should not expect to be able to restart xenstored
                    300: (and thus xencommons).  There is currently no reason to expect that
                    301: this will get fixed any time soon.
1.187     gdt       302: \todo Confirm if this is still true in 2020.
1.82      gdt       303: 
1.181     gdt       304: ## Xen-specific NetBSD issues
1.40      gdt       305: 
                    306: There are (at least) two additional things different about NetBSD as a
                    307: dom0 kernel compared to hardware.
                    308: 
1.191     gdt       309: One is that through NetBSD 9 the module ABI is different because some
                    310: of the #defines change, so there are separate sets of modules in
                    311: /stand.  In NetBSD-current, there is only one set of modules.
1.40      gdt       312: 
                    313: The other difference is that XEN3_DOM0 does not have exactly the same
                    314: options as GENERIC.  While it is debatable whether or not this is a
                    315: bug, users should be aware of this and can simply add missing config
                    316: items if desired.
                    317: 
1.187     gdt       318: Finally, there have been occasional reports of trouble with X11
                    319: servers in NetBSD as a dom0.
                    320: 
                    321: ## Updating Xen in a dom0
                    322: 
                    323: Basically, update the xenkernel and xentools packages and copy the new
                    324: Xen kernel into place, and reboot.  This procedure should be usable to
                    325: update to a new Xen release, but the reader is reminded that having a
                    326: non-Xen boot methods was recommended earlier.
                    327: 
1.181     gdt       328: ## Updating NetBSD in a dom0
1.15      gdt       329: 
                    330: This is just like updating NetBSD on bare hardware, assuming the new
                    331: version supports the version of Xen you are running.  Generally, one
                    332: replaces the kernel and reboots, and then overlays userland binaries
1.157     maxv      333: and adjusts `/etc`.
1.15      gdt       334: 
1.191     gdt       335: Note that one should update both the non-Xen kernel typically used for
                    336: rescue purposes, as well as the DOM0 kernel used with Xen.
1.15      gdt       337: 
1.187     gdt       338: ## anita (for testing NetBSD)
                    339: 
                    340: With a NetBSD dom0, even without any domUs, one should be able to run
                    341: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
                    342: root, because anita must create a domU):
                    343: 
                    344: [[!template id=programlisting text="""
                    345: anita --vmm=xl test file:///usr/obj/i386/
                    346: """]]
                    347: 
                    348: # Unprivileged domains (domU)
1.28      gdt       349: 
                    350: This section describes general concepts about domUs.  It does not
1.33      gdt       351: address specific domU operating systems or how to install them.  The
1.157     maxv      352: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
1.60      wiki      353: typically named so that the file name, domU name and the domU's host
1.33      gdt       354: name match.
                    355: 
1.111     wiz       356: The domU is provided with CPU and memory by Xen, configured by the
1.33      gdt       357: dom0.  The domU is provided with disk and network by the dom0,
                    358: mediated by Xen, and configured in the dom0.
                    359: 
                    360: Entropy in domUs can be an issue; physical disks and network are on
                    361: the dom0.  NetBSD's /dev/random system works, but is often challenged.
                    362: 
1.181     gdt       363: ## Config files
1.48      gdt       364: 
1.200   ! gdt       365: See /usr/pkg/share/examples/xen/xlexample* for a very small number of
        !           366: examples for running GNU/Linux.
1.48      gdt       367: 
1.157     maxv      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: """]]
1.48      gdt       379: 
1.157     maxv      380: The domain will have name given in the `name` setting.  The kernel has the
1.48      gdt       381: host/domU name in it, so that on the dom0 one can update the various
1.157     maxv      382: domUs independently.  The `vif` line causes an interface to be provided,
1.48      gdt       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
1.111     wiz       386: dom0 on domain creation.  The system treats xbd0 as the boot device
1.48      gdt       387: without needing explicit configuration.
                    388: 
1.200   ! gdt       389: There is not type line; that implicitly defines a pv domU.  Otherwise,
        !           390: one sets type to the lower-case version of the domU type in the table
        !           391: above, e.g.  `type = "hvm"`.
1.187     gdt       392: 
1.163     gson      393: By convention, domain config files are kept in `/usr/pkg/etc/xen`.  Note
1.157     maxv      394: that "xl create" takes the name of a config file, while other commands
                    395: take the name of a domain.
                    396: 
                    397: Examples of commands:
1.48      gdt       398: 
1.157     maxv      399: [[!template id=programlisting text="""
1.163     gson      400: xl create /usr/pkg/etc/xen/foo
                    401: xl console domU-id
                    402: xl create -c /usr/pkg/etc/xen/foo
                    403: xl shutdown domU-id
1.157     maxv      404: xl list
                    405: """]]
                    406: 
                    407: Typing `^]` will exit the console session.  Shutting down a domain is
1.48      gdt       408: equivalent to pushing the power button; a NetBSD domU will receive a
                    409: power-press event and do a clean shutdown.  Shutting down the dom0
                    410: will trigger controlled shutdowns of all configured domUs.
                    411: 
1.181     gdt       412: ## CPU and memory
1.33      gdt       413: 
1.169     plunky    414: A domain is provided with some number of vcpus, up to the number
1.157     maxv      415: of CPUs seen by the hypervisor. For a domU, it is controlled
1.48      gdt       416: from the config file by the "vcpus = N" directive.
                    417: 
                    418: A domain is provided with memory; this is controlled in the config
                    419: file by "memory = N" (in megabytes).  In the straightforward case, the
                    420: sum of the the memory allocated to the dom0 and all domUs must be less
1.33      gdt       421: than the available memory.
                    422: 
1.197     gdt       423: ## Balloon driver
                    424: 
                    425: Xen provides a `balloon` driver, which can be used to let domains use
                    426: more memory temporarily.
                    427: 
                    428: \todo Explain how to set up a aystem to use the balloon scheme in a
                    429: useful manner.
1.28      gdt       430: 
1.181     gdt       431: ## Virtual disks
1.28      gdt       432: 
1.158     maxv      433: In domU config files, the disks are defined as a sequence of 3-tuples:
                    434: 
                    435:  * The first element is "method:/path/to/disk". Common methods are
1.162     gutterid  436:    "file:" for a file-backed vnd, and "phy:" for something that is already
                    437:    a device, such as an LVM logical volume.
1.158     maxv      438: 
                    439:  * The second element is an artifact of how virtual disks are passed to
                    440:    Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
                    441:    are given a device name to associate with the disk, and values like
                    442:    "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
                    443:    as xbd0, the second as xbd1, and so on.  However, xl demands a
                    444:    second argument.  The name given is converted to a major/minor by
                    445:    calling stat(2) on the name in /dev and this is passed to the domU.
                    446:    In the general case, the dom0 and domU can be different operating
                    447:    systems, and it is an unwarranted assumption that they have consistent
                    448:    numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
                    449:    as both dom0 and domU, using values of 0x0 for the first disk and 0x1
                    450:    for the second works fine and avoids this issue.  For a GNU/Linux
                    451:    guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
                    452:    /dev/hda1.
1.48      gdt       453: 
1.158     maxv      454:  * The third element is "w" for writable disks, and "r" for read-only
                    455:    disks.
                    456: 
                    457: Example:
                    458: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
                    459: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ]
                    460: """]]
1.28      gdt       461: 
1.127     gdt       462: Note that NetBSD by default creates only vnd[0123].  If you need more
                    463: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
                    464: dom0.
                    465: 
1.187     gdt       466: ## Virtual Networking
1.28      gdt       467: 
1.111     wiz       468: Xen provides virtual Ethernets, each of which connects the dom0 and a
1.46      gdt       469: domU.  For each virtual network, there is an interface "xvifN.M" in
1.158     maxv      470: the dom0, and a matching interface xennetM (NetBSD name) in domU index N.
                    471: The interfaces behave as if there is an Ethernet with two
1.111     wiz       472: adapters connected.  From this primitive, one can construct various
1.46      gdt       473: configurations.  We focus on two common and useful cases for which
                    474: there are existing scripts: bridging and NAT.
1.28      gdt       475: 
1.48      gdt       476: With bridging (in the example above), the domU perceives itself to be
                    477: on the same network as the dom0.  For server virtualization, this is
                    478: usually best.  Bridging is accomplished by creating a bridge(4) device
                    479: and adding the dom0's physical interface and the various xvifN.0
                    480: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
                    481: config file.  The bridge must be set up already in the dom0; an
                    482: example /etc/ifconfig.bridge0 is:
1.46      gdt       483: 
1.157     maxv      484: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
                    485: create
                    486: up
                    487: !brconfig bridge0 add wm0
                    488: """]]
1.28      gdt       489: 
                    490: With NAT, the domU perceives itself to be behind a NAT running on the
                    491: dom0.  This is often appropriate when running Xen on a workstation.
1.48      gdt       492: TODO: NAT appears to be configured by "vif = [ '' ]".
1.28      gdt       493: 
1.49      gdt       494: The MAC address specified is the one used for the interface in the new
1.53      gdt       495: domain.  The interface in dom0 will use this address XOR'd with
1.49      gdt       496: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
                    497: 
1.187     gdt       498: ## Starting domains automatically
1.28      gdt       499: 
1.158     maxv      500: To start domains `domU-netbsd` and `domU-linux` at boot and shut them
                    501: down cleanly on dom0 shutdown, add the following in rc.conf:
1.28      gdt       502: 
1.158     maxv      503: [[!template id=filecontent name="/etc/rc.conf" text="""
                    504: xendomains="domU-netbsd domU-linux"
                    505: """]]
1.28      gdt       506: 
1.188     gdt       507: # domU setup for specific systems
1.14      gdt       508: 
                    509: Creating domUs is almost entirely independent of operating system.  We
1.188     gdt       510: have already presented the basics of config files in the previous system.
                    511: 
                    512: Of course, this section presumes that you have a working dom0.
1.14      gdt       513: 
1.187     gdt       514: ## Creating a NetBSD PV domU
                    515: 
1.49      gdt       516: See the earlier config file, and adjust memory.  Decide on how much
1.162     gutterid  517: storage you will provide, and prepare it (file or LVM).
1.49      gdt       518: 
1.111     wiz       519: While the kernel will be obtained from the dom0 file system, the same
1.49      gdt       520: file should be present in the domU as /netbsd so that tools like
                    521: savecore(8) can work.   (This is helpful but not necessary.)
                    522: 
1.188     gdt       523: The kernel must be specifically built for Xen, to use PV interfacesas
                    524: a domU.  NetBSD release builds provide the following kernels:
1.49      gdt       525: 
                    526:         i386 XEN3PAE_DOMU
1.95      gdt       527:         amd64 XEN3_DOMU
1.5       mspo      528: 
1.49      gdt       529: This will boot NetBSD, but this is not that useful if the disk is
1.188     gdt       530: empty.  One approach is to unpack sets onto the disk outside of Xen
1.49      gdt       531: (by mounting it, just as you would prepare a physical disk for a
                    532: system you can't run the installer on).
                    533: 
                    534: A second approach is to run an INSTALL kernel, which has a miniroot
                    535: and can load sets from the network.  To do this, copy the INSTALL
                    536: kernel to / and change the kernel line in the config file to:
1.5       mspo      537: 
1.49      gdt       538:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
1.5       mspo      539: 
1.163     gson      540: Then, start the domain as "xl create -c configfile".
1.1       mspo      541: 
1.49      gdt       542: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
                    543: line should be used in the config file.
1.1       mspo      544: 
1.3       mspo      545:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
1.1       mspo      546: 
                    547: After booting the domain, the option to install via CDROM may be
1.49      gdt       548: selected.  The CDROM device should be changed to `xbd1d`.
1.1       mspo      549: 
1.188     gdt       550: Once done installing, "halt -p" the new domain (don't reboot or halt:
1.49      gdt       551: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
                    552: config file), switch the config file back to the XEN3_DOMU kernel,
                    553: and start the new domain again. Now it should be able to use "root on
1.188     gdt       554: xbd0a" and you should have a functional NetBSD domU.
1.1       mspo      555: 
1.49      gdt       556: TODO: check if this is still accurate.
1.1       mspo      557: When the new domain is booting you'll see some warnings about *wscons*
                    558: and the pseudo-terminals. These can be fixed by editing the files
1.5       mspo      559: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
                    560: `/etc/ttys`, except *console*, like this:
1.1       mspo      561: 
1.3       mspo      562:     console "/usr/libexec/getty Pc"         vt100   on secure
                    563:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
                    564:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
                    565:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
                    566:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
1.1       mspo      567: 
1.5       mspo      568: Finally, all screens must be commented out from `/etc/wscons.conf`.
1.1       mspo      569: 
1.192     gdt       570: One should also run `powerd` in a domU, but this should not need
                    571: configuring.  With powerd, the domain will run a controlled shutdown
                    572: if `xl shutdown -R` or `xl shutdown -H` is used on the dom0, via
                    573: receiving a synthetic `power button pressed` signal.  In 9 and
                    574: current, `powerd` is run by default under Xen kernels (or if ACPI is
                    575: present), and it can be added to rc.conf if not.
1.1       mspo      576: 
1.92      gdt       577: It is not strictly necessary to have a kernel (as /netbsd) in the domU
1.111     wiz       578: file system.  However, various programs (e.g. netstat) will use that
1.92      gdt       579: kernel to look up symbols to read from kernel virtual memory.  If
                    580: /netbsd is not the running kernel, those lookups will fail.  (This is
                    581: not really a Xen-specific issue, but because the domU kernel is
                    582: obtained from the dom0, it is far more likely to be out of sync or
                    583: missing with Xen.)
                    584: 
1.187     gdt       585: Note that NetBSD by default creates only xbd[0123].  If you need more
                    586: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
                    587: 
1.196     gdt       588: ## Creating a Linux PV domU
1.1       mspo      589: 
                    590: Creating unprivileged Linux domains isn't much different from
                    591: unprivileged NetBSD domains, but there are some details to know.
                    592: 
                    593: First, the second parameter passed to the disk declaration (the '0x1' in
                    594: the example below)
                    595: 
1.3       mspo      596:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
1.1       mspo      597: 
                    598: does matter to Linux. It wants a Linux device number here (e.g. 0x300
1.49      gdt       599: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
                    600: So, hda1 which has major 3 and minor 1 on a Linux system will have
                    601: device number 0x301.  Alternatively, devices names can be used (hda,
                    602: hdb, ...)  as xentools has a table to map these names to devices
                    603: numbers.  To export a partition to a Linux guest we can use:
1.1       mspo      604: 
1.49      gdt       605:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
                    606:         root = "/dev/hda1 ro"
1.1       mspo      607: 
                    608: and it will appear as /dev/hda on the Linux system, and be used as root
                    609: partition.
                    610: 
1.49      gdt       611: To install the Linux system on the partition to be exported to the
                    612: guest domain, the following method can be used: install
                    613: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
                    614: that will be the root partition of your Linux domain, and mount it.
                    615: Then copy the files from a working Linux system, make adjustments in
                    616: `/etc` (fstab, network config).  It should also be possible to extract
                    617: binary packages such as .rpm or .deb directly to the mounted partition
                    618: using the appropriate tool, possibly running under NetBSD's Linux
1.111     wiz       619: emulation.  Once the file system has been populated, umount it.  If
                    620: desirable, the file system can be converted to ext3 using tune2fs -j.
1.49      gdt       621: It should now be possible to boot the Linux guest domain, using one of
                    622: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
1.1       mspo      623: 
1.111     wiz       624: To get the Linux console right, you need to add:
1.1       mspo      625: 
1.3       mspo      626:     extra = "xencons=tty1"
1.1       mspo      627: 
1.111     wiz       628: to your configuration since not all Linux distributions auto-attach a
1.1       mspo      629: tty to the xen console.
                    630: 
1.180     gdt       631: ## Creating a NetBSD HVM domU
                    632: 
1.196     gdt       633: Use type='hvm', probably.  Use a GENERIC kernel within the disk image.
1.180     gdt       634: 
                    635: ## Creating a NetBSD PVH domU
                    636: 
1.196     gdt       637: This only works with a current kernel in the domU.
1.180     gdt       638: 
1.196     gdt       639: Use type='pvh'.  Probably, use a GENERIC kernel within the disk image,
                    640: which in current has PV support.
                    641: 
                    642: \todo Verify.
                    643: 
                    644: \todo Verify if one can have current PVH domU on a 9 dom0.
1.180     gdt       645: 
1.187     gdt       646: ## Creating a Solaris domU
1.1       mspo      647: 
1.50      gdt       648: See possibly outdated
                    649: [Solaris domU instructions](/ports/xen/howto-solaris/).
1.5       mspo      650: 
1.187     gdt       651: ## PCI passthrough: Using PCI devices in guest domains
1.1       mspo      652: 
1.180     gdt       653: NB: PCI passthrough only works on some Xen versions and as of 2020 it
1.196     gdt       654: is not clear that it works on any version in pkgsrc.  \todo Reports
1.180     gdt       655: confirming or denying this notion should be sent to port-xen@.
                    656: 
1.53      gdt       657: The dom0 can give other domains access to selected PCI
1.52      gdt       658: devices. This can allow, for example, a non-privileged domain to have
                    659: access to a physical network interface or disk controller.  However,
                    660: keep in mind that giving a domain access to a PCI device most likely
                    661: will give the domain read/write access to the whole physical memory,
                    662: as PCs don't have an IOMMU to restrict memory access to DMA-capable
1.53      gdt       663: device.  Also, it's not possible to export ISA devices to non-dom0
1.52      gdt       664: domains, which means that the primary VGA adapter can't be exported.
                    665: A guest domain trying to access the VGA registers will panic.
                    666: 
1.53      gdt       667: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
1.52      gdt       668: not been ported to later versions at this time.
                    669: 
                    670: For a PCI device to be exported to a domU, is has to be attached to
                    671: the "pciback" driver in dom0.  Devices passed to the dom0 via the
                    672: pciback.hide boot parameter will attach to "pciback" instead of the
                    673: usual driver.  The list of devices is specified as "(bus:dev.func)",
1.5       mspo      674: where bus and dev are 2-digit hexadecimal numbers, and func a
                    675: single-digit number:
1.1       mspo      676: 
1.52      gdt       677:         pciback.hide=(00:0a.0)(00:06.0)
1.1       mspo      678: 
1.52      gdt       679: pciback devices should show up in the dom0's boot messages, and the
1.5       mspo      680: devices should be listed in the `/kern/xen/pci` directory.
1.1       mspo      681: 
1.52      gdt       682: PCI devices to be exported to a domU are listed in the "pci" array of
                    683: the domU's config file, with the format "0000:bus:dev.func".
1.1       mspo      684: 
1.52      gdt       685:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
1.1       mspo      686: 
1.52      gdt       687: In the domU an "xpci" device will show up, to which one or more pci
1.111     wiz       688: buses will attach.  Then the PCI drivers will attach to PCI buses as
1.52      gdt       689: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
                    690: or any PCI drivers built in by default; you have to build your own
                    691: kernel to use PCI devices in a domU.  Here's a kernel config example;
                    692: note that only the "xpci" lines are unusual.
                    693: 
                    694:         include         "arch/i386/conf/XEN3_DOMU"
                    695: 
1.111     wiz       696:         # Add support for PCI buses to the XEN3_DOMU kernel
1.52      gdt       697:         xpci* at xenbus ?
                    698:         pci* at xpci ?
                    699: 
                    700:         # PCI USB controllers
                    701:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
                    702: 
                    703:         # USB bus support
                    704:         usb*    at uhci?
                    705: 
                    706:         # USB Hubs
                    707:         uhub*   at usb?
                    708:         uhub*   at uhub? port ? configuration ? interface ?
                    709: 
                    710:         # USB Mass Storage
                    711:         umass*  at uhub? port ? configuration ? interface ?
                    712:         wd*     at umass?
                    713:         # SCSI controllers
                    714:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
                    715: 
                    716:         # SCSI bus support (for both ahc and umass)
                    717:         scsibus* at scsi?
                    718: 
                    719:         # SCSI devices
                    720:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
                    721:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
1.1       mspo      722: 
                    723: 
1.189     gdt       724: # Miscellaneous Information
1.173     gdt       725: 
1.189     gdt       726: ## Nesting under Linux KVM
                    727: 
                    728: It is possible to run a Xen and a NetBSD dom0 under Linux KVM.  One
                    729: can enable virtio in the dom0 for greater speed.
                    730: 
                    731: ## Other nesting
                    732: 
                    733: In theory, any full emulation should be able to run Xen and a NetBSD
                    734: dom0.  The HOWTO does not currently have information about Xen XVM
                    735: mode, nvmm, qemu, Virtualbox, etc.
                    736: 
                    737: ## NetBSD 5 as domU
1.173     gdt       738: 
1.176     gdt       739: [NetBSD 5 is known to panic.](http://mail-index.netbsd.org/port-xen/2018/04/17/msg009181.html)
1.175     gdt       740: (However, NetBSD 5 systems should be updated to a supported version.)
1.173     gdt       741: 
                    742: # NetBSD as a domU in a VPS
1.28      gdt       743: 
                    744: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
                    745: hardware.  This section explains how to deal with Xen in a domU as a
                    746: virtual private server where you do not control or have access to the
1.70      gdt       747: dom0.  This is not intended to be an exhaustive list of VPS providers;
                    748: only a few are mentioned that specifically support NetBSD.
1.28      gdt       749: 
1.52      gdt       750: VPS operators provide varying degrees of access and mechanisms for
                    751: configuration.  The big issue is usually how one controls which kernel
1.111     wiz       752: is booted, because the kernel is nominally in the dom0 file system (to
                    753: which VPS users do not normally have access).  A second issue is how
1.70      gdt       754: to install NetBSD.
1.52      gdt       755: A VPS user may want to compile a kernel for security updates, to run
                    756: npf, run IPsec, or any other reason why someone would want to change
                    757: their kernel.
                    758: 
1.111     wiz       759: One approach is to have an administrative interface to upload a kernel,
1.68      gdt       760: or to select from a prepopulated list.  Other approaches are pygrub
1.59      gdt       761: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
1.111     wiz       762: kernel from the domU file system.  This is closer to a regular physical
1.59      gdt       763: computer, where someone who controls a machine can replace the kernel.
1.52      gdt       764: 
1.74      gdt       765: A second issue is multiple CPUs.  With NetBSD 6, domUs support
                    766: multiple vcpus, and it is typical for VPS providers to enable multiple
                    767: CPUs for NetBSD domUs.
                    768: 
1.180     gdt       769: ## Complexities due to Xen changes
                    770: 
                    771: Xen has many security advisories and people running Xen systems make
                    772: different choices.
                    773: 
                    774: ### stub domains
                    775: 
                    776: Some (Linux only?) dom0 systems use something called "stub domains" to
                    777: isolate qemu from the dom0 system, as a security and reliabilty
                    778: mechanism when running HVM domUs.  Somehow, NetBSD's GENERIC kernel
                    779: ends up using PIO for disks rather than DMA.  Of course, all of this
                    780: is emulated, but emulated PIO is unusably slow.  This problem is not
                    781: currently understood.
                    782: 
                    783: ### Grant tables
                    784: 
                    785: There are multiple versions of using grant tables, and some security
                    786: advisories have suggested disabling some versions.  Some versions of
                    787: NetBSD apparently only use specific versions and this can lead to
                    788: "NetBSD current doesn't run on hosting provider X" situations.
                    789: 
                    790: \todo Explain better.
                    791: 
1.181     gdt       792: ## Boot methods
                    793: 
                    794: ### pvgrub
1.59      gdt       795: 
                    796: pvgrub is a version of grub that uses PV operations instead of BIOS
                    797: calls.  It is booted from the dom0 as the domU kernel, and then reads
1.111     wiz       798: /grub/menu.lst and loads a kernel from the domU file system.
1.59      gdt       799: 
1.70      gdt       800: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
1.71      gdt       801: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
                    802: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
1.70      gdt       803: page](http://www.panix.com/v-colo/grub.html), which describes only
1.74      gdt       804: Linux but should be updated to cover NetBSD :-).
1.70      gdt       805: 
                    806: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
                    807: their own kernel.  See then [prgmr.com NetBSD
1.74      gdt       808: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
                    809: (which is in need of updating).
1.59      gdt       810: 
1.70      gdt       811: It appears that [grub's FFS
                    812: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
                    813: does not support all aspects of modern FFS, but there are also reports
1.72      gdt       814: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
1.70      gdt       815: partition for the kernel with the intent that grub can understand it,
                    816: which leads to /netbsd not being the actual kernel.  One must remember
1.111     wiz       817: to update the special boot partition.
1.59      gdt       818: 
1.181     gdt       819: ### pygrub
1.168     wiki      820: 
                    821: pygrub runs in the dom0 and looks into the domU file system.  This
                    822: implies that the domU must have a kernel in a file system in a format
                    823: known to pygrub.
                    824: 
                    825: pygrub doesn't seem to work to load Linux images under NetBSD dom0,
                    826: and is inherently less secure than pvgrub due to running inside dom0. For both these
                    827: reasons, pygrub should not be used, and is only still present so that
                    828: historical DomU images using it still work.
                    829: 
                    830: As of 2014, pygrub seems to be of mostly historical
                    831: interest. New DomUs should use pvgrub.
                    832: 
1.181     gdt       833: ## Specific Providers
                    834: 
                    835: ### Amazon
1.59      gdt       836: 
1.143     wiki      837: See the [Amazon EC2 page](/amazon_ec2/).

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