Annotation of wikisrc/ports/xen/howto.mdwn, revision 1.193

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

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