Annotation of wikisrc/ports/xen/howto.mdwn, revision 1.206

1.203     gdt         1: [[!meta title="Xen Status and HowTo"]]
1.144     maxv        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: 
1.203     gdt        11: This document provides status on what Xen things work on NetBSD
                     12: (upstream documentation might say something works if it works on some
                     13: particular Linux system).
                     15: This document is also a HOWTO that presumes a basic familiarity with
                     16: the Xen system architecture, with installing NetBSD on amd64 hardware,
                     17: and with installing software from pkgsrc.  See also the [Xen
1.182     gdt        18: website](
1.203     gdt        20: If this document says that something works, and you find that it does
                     21: not, it is best to ask on port-xen and if you are correct to file a
                     22: PR.
1.201     gdt        23: 
1.182     gdt        24: [[!toc]]
1.183     gdt        26: # Overview
1.182     gdt        27: 
                     28: The basic concept of Xen is that the hypervisor (xenkernel) runs on
                     29: the hardware, and runs a privileged domain ("dom0") that can access
1.204     gdt        30: disks/networking/etc.  One then runs additional unprivileged domains
1.182     gdt        31: (each a "domU"), presumably to do something useful.
                     33: This HOWTO addresses how to run a NetBSD dom0 (and hence also build
                     34: xen itself).  It also addresses how to run domUs in that environment,
                     35: and how to deal with having a domU in a Xen environment run by someone
                     36: else and/or not running NetBSD.
1.186     gdt        38: There are many choices one can make; the HOWTO recommends the standard
                     39: approach and limits discussion of alternatives in many cases.
1.182     gdt        41: ## Guest Styles
1.205     gdt        43: Xen supports different styles of guests.  See
                     44: for a
                     45: discussion.
                     47: This table shows the styles, and if a NetBSD dom0 can run in that
                     48: style, if a NetBSD dom0 can sypport that style of guest in a domU, and
                     49: if NetBSD as a domU can support that style.
1.149     maxv       50: 
                     51: [[!table data="""
1.205     gdt        52: Style of guest |dom0 can be?   |dom0 can support?      |domU can be?
                     53: PV             |yes            |yes                    |yes
                     54: HVM            |N/A            |yes                    |yes
                     55: PVHVM          |N/A            |yes                    |current only
                     56: PVH            |not yet        |current only           |current only
1.149     maxv       57: """]]
1.205     gdt        59: In PV (paravirtualized) mode, the guest OS does not attempt to access
1.149     maxv       60: hardware directly, but instead makes hypercalls to the hypervisor; PV
1.205     gdt        61: guests must be specifically coded for Xen.  See
                     62: [PV](\)).
1.170     gdt        63: 
1.205     gdt        64: In HVM (Hardware Virtual Machine) mode, no guest modification is
                     65: required.  However, hardware support is required, such as VT-x on
                     66: Intel CPUs and SVM on AMD CPUs to assist with the processor emulation.
                     67: The dom0 runs qemu to emulate hardware other than the processor.  It
                     68: is therefore non-sensical to have an HVM dom0, because there is no
                     69: underlying system to provide emulation.
1.170     gdt        70: 
1.205     gdt        71: In PVHVM mode, the guest runs as HVM, but additionally uses PV
1.199     gdt        72: drivers for efficiency.  Therefore it is non-sensical for to have a
                     73: PVHVM dom0.  See [PV on HVM](
1.170     gdt        74: 
1.180     gdt        75: There have been two PVH modes: original PVH and PVHv2.  Original PVH
1.203     gdt        76: was based on PV mode and is no longer relevant at all.  Therefore
1.205     gdt        77: PVHv2 is written as PVH, here and elsewhere.  PVH is basically
                     78: lightweight HVM with PV drivers.  A critical feature of it is that
                     79: qemu is not needed; the hypervisor can do the emulation that is
                     80: required.  Thus, a dom0 can be PVH.  The source code uses PVH and
                     81: config files use pvh, but NB that this refers to PVHv2.  See
1.203     gdt        82: [PVH(v2)](\)_Domu).
1.180     gdt        83: 
1.205     gdt        84: At system boot, the dom0 kernel is loaded as a module with Xen as the
                     85: kernel.  The dom0 can start one or more domUs.  (Booting is explained
                     86: in detail in the dom0 section.)
1.12      gdt        87: 
1.182     gdt        88: ## CPU Architecture
                     90: Xen runs on x86_64 hardware (the NetBSD amd64 port).
                     92: There is a concept of Xen running on ARM, but there are no reports of this working with NetBSD.
                     94: The dom0 system should be amd64.  (Instructions for i386PAE dom0 have been removed from the HOWTO.)
1.1       mspo       95: 
1.204     gdt        96: The domU can be i386 PAE or amd64.
                     97: i386 PAE at one point was considered as [faster]( than amd64.
                     98: However, as of 2021 it is normal to use amd64 as the domU architecture, and use of i386 is dwindling.
1.150     maxv       99: 
1.182     gdt       100: ## Xen Versions
1.15      gdt       101: 
1.111     wiz       102: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
1.15      gdt       103: xenkernel and xentools.  We will refer only to the kernel versions,
                    104: but note that both packages must be installed together and must have
                    105: matching versions.
1.145     maxv      107: Versions available in pkgsrc:
1.85      gdt       108: 
1.145     maxv      109: [[!table data="""
1.184     gdt       110: Xen Version    |Package Name   |Xen CPU Support        |EOL'ed By Upstream
                    111: 4.11           |xenkernel411   |x86_64                 |No
                    112: 4.13           |xenkernel413   |x86_64                 |No
1.145     maxv      113: """]]
1.113     gdt       114: 
1.96      gdt       115: See also the [Xen Security Advisory page](
1.204     gdt       117: Older Xen had a python-based management tool called xm; this has been
                    118: replaced by xl.
1.166     gdt       119: 
1.183     gdt       120: ## NetBSD versions
                    122: Xen has been supported in NetBSD for a long time, at least since 2005.
                    123: Initially Xen was PV only.
1.204     gdt       125: NetBSD Xen has always supported PV, in both dom0 and domU; for a long
                    126: time this was the only way.  NetBSD >=8 as a dom0 supports HVM mode in
1.205     gdt       127: domUs.
1.203     gdt       128: 
                    129: Support for PVHVM and PVH is available only in NetBSD-current; this is
                    130: currently somewhat experimental, although PVHVM appears reasonably
                    131: solid.
                    133: NetBSD up to and including NetBSD 9 as a dom0 cannot safely run SMP.
                    134: Even if one added "options MULTIPROCESSOR" and configured multiple
                    135: vcpus, the kernel is likely to crash because of drivers without
                    136: adequate locking.
                    138: NetBSD-current supports SMP in dom0, and XEN3_DOM0 includes "options
                    139: MULTIPROCESSOR".
                    141: NetBSD (since NetBSD 6), when run as a domU, can run SMP, using
                    142: multiple CPUs if provided.  The XEN3_DOMU kernel is built
                    143: with "options MULITPROCESSOR".
1.185     gdt       144: 
1.191     gdt       145: Note that while Xen 4.13 is current, the kernel support is still
                    146: called XEN3, because the hypercall interface has not changed
                    147: significantly.
1.185     gdt       148: 
1.187     gdt       149: # Creating a NetBSD dom0
1.15      gdt       150: 
1.191     gdt       151: In order to install a NetBSD as a dom0, one first installs a normal
                    152: NetBSD system, and then pivot the install to a dom0 install by
                    153: changing the kernel and boot configuration.
1.15      gdt       154: 
1.206   ! gdt       155: NB: As of 2021-04, you must arrange to have the system use BIOS boot,
        !           156: not EFI boot.  (Perhaps, with recent current and using multiboot
        !           157: instead of NetBSD boot, this might not be true.)
        !           158: 
1.142     gdt       159: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
                    160: 512M it worked reliably.  This does not make sense, but if you see
                    161: "not ELF" after Xen boots, try increasing dom0 RAM.
1.181     gdt       163: ## Installation of NetBSD
1.13      gdt       164: 
1.191     gdt       165: [Install NetBSD/amd64](/guide/inst/) just as you would if you were not
                    166: using Xen.  Therefore, use the most recent release, or a build from
                    167: the most recent stable branch.  Alternatively, use -current, being
                    168: mindful of all the usual caveats of lower stability of current, and
1.205     gdt       169: likely a bit more so.  Think about how you will provide storage for
                    170: disk images.
1.1       mspo      171: 
1.181     gdt       172: ## Installation of Xen
1.1       mspo      173: 
1.191     gdt       174: ### Building Xen
1.205     gdt       176: Use the most recent version of Xen in pkgsrc, unless the DESCR says
                    177: that it is not suitable.  Therefore, choose 4.13.  In the dom0,
                    178: install xenkernel413 and xentools413 from pkgsrc.
1.155     maxv      179: 
1.186     gdt       180: Once this is done, copy the Xen kernel from where pkgsrc puts it to
                    181: where the boot process will be able to find it:
1.155     maxv      182: 
                    183: [[!template id=programlisting text="""
1.186     gdt       184: # cp -p /usr/pkg/xen413-kernel/xen.gz /
1.155     maxv      185: """]]
1.191     gdt       187: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such
                    188: kernel can either be taken from a local release run, compiled
                    189: manually, or downloaded from the NetBSD FTP, for example at:
1.159     maxv      190: 
                    191: [[!template id=programlisting text="""
1.186     gdt       192:
1.159     maxv      193: """]]
1.20      gdt       194: 
1.191     gdt       195: ### Configuring booting
1.198     gdt       197: Read boot.cfg(8) carefully.  Add lines to /boot.cfg to boot Xen,
                    198: adjusting for your root filesystem:
1.20      gdt       199: 
1.157     maxv      200: [[!template id=filecontent name="/boot.cfg" text="""
1.198     gdt       201: menu=Xen:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc;multiboot /xen.gz dom0_mem=512M
                    202: menu=Xen single user:load /netbsd-XEN3_DOM0.gz root=wd0a console=pc -s;multiboot /xen.gz dom0_mem=512M
1.152     maxv      203: """]]
1.20      gdt       204: 
1.159     maxv      205: This specifies that the dom0 should have 512MB of ram, leaving the rest
1.187     gdt       206: to be allocated for domUs.
1.191     gdt       208: NB: This says add, not replace, so that you will be able to more
                    209: easily boot a NetBSD kernel without Xen.  Once Xen boots ok, you may
                    210: want to set it as default.  It is highly likely that you will have
                    211: trouble at some point, and keeping an up-to-date GENERIC for use in
                    212: fixing problems is the standard prudent approach.
1.194     gdt       214: \todo Explain why rndseed is not set with Xen as part of the dom0
                    215: subconfiguration.
1.198     gdt       217: Note that you are likely to have to set root= because the boot device
                    218: from /boot is not passed via Xen to the dom0 kernel.  With one disk,
                    219: it will work, but e.g. plugging in USB disk to a machine with root on
                    220: wd0a causes boot to fail.
1.195     gdt       222: Beware that userconf statements must be attached to the dom0 load, and
                    223: may not be at top-level, because then they would try to configure the
1.205     gdt       224: hypervisor, if there is a way to pass them via multiboot.  It appears
1.195     gdt       225: that adding `userconf=pckbc` to `/boot.cfg` causes Xen to crash very
                    226: early with a heap overflow.
1.191     gdt       228: ### Console selection
                    230: See boot_console(8).  Understand that you should start from a place of
                    231: having console setup correct for booting GENERIC before trying to
                    232: configure Xen.
1.193     gdt       234: Generally for GENERIC, one sets the console in bootxx_ffsv1 or
                    235: equivalent, and this is passed on to /boot (where one typically does
                    236: not set the console).  This configuration of bootxx_ffsv1 should also
                    237: be in place for Xen systems, to allow seeing messages from /boot and
                    238: use of a keyboard to select a line from the menu.  And, one should
                    239: have a working boot path to GENERIC for rescue situations.
                    241: With GENERIC, the boot options are passed on to /netbsd, but there is
                    242: currently no mechanism to pass these via multiboot to the hypervisor.
                    243: Thus, in addition to configuring the console in the boot blocks, one
                    244: must also configure it for Xen.
                    246: By default, the hypervisor (Xen itself) will use some sort of vga
                    247: device as the console, much like GENERIC uses by default.  The vga
                    248: console is relinquished at the conclusion of hypervisor boot, before
1.194     gdt       249: the dom0 is started.  Xen when using a vga console does not process
                    250: console input.
1.77      gdt       251: 
1.193     gdt       252: The hypervisor can be configured to use a serial port console, e.g.
1.157     maxv      253: [[!template id=filecontent name="/boot.cfg" text="""
1.202     gson      254: menu=Xen:load /netbsd-XEN3_DOM0.gz console=com0;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
1.152     maxv      255: """]]
1.194     gdt       256: This example uses the first serial port (Xen counts from 1; this is
1.193     gdt       257: what NetBSD would call com0), and sets speed and parity.  (The dom0 is
                    258: then configured to use the same serial port in this example.)
1.194     gdt       260: With the hypervisor configured for a serial console, it can get input,
                    261: and there is a notion of passing this input to the dom0.  \todo
                    262: Explain why, if Xen has a serial console, the dom0 console is
                    263: typically also configured to open that same serial port, instead of
                    264: getting the passthrough input via the xen console.
1.193     gdt       266: One also configures the console for the dom0.  While one might expect
                    267: console=pc to be default, following behavior of GENERIC, a hasty read
                    268: of the code suggests there is no default and booting without a
                    269: selected console might lead to a panic.  Also, there is merit in
                    270: explicit configuration.  Therefore the standard approach is to place
                    271: console=pc as part of the load statement for the dom0 kernel, or
                    272: alternatively console=com0.
                    274: The NetBSD dom0 kernel will attach xencons(4) (the man page does not
                    275: exist), but this is not used as a console.  It is used to obtain the
                    276: messages from the hypervisor's console; run `xl dmesg` to see them.
1.191     gdt       277: 
                    278: ### Tuning
1.205     gdt       280: In an attempt to add performance, one can also add `dom0_max_vcpus=1
                    281: dom0_vcpus_pin`, to force only one vcpu to be provided (since NetBSD
                    282: dom0 can't use more) and to pin that vcpu to a physical CPU. Xen has
                    283: [many boot
                    284: options](,
1.111     wiz       285: and other than dom0 memory and max_vcpus, they are generally not
1.93      gdt       286: necessary.
1.205     gdt       287: 
1.191     gdt       288: \todo Revisit this advice with current.
                    289: \todo Explain if anyone has ever actually measured that this helps.
                    291: ### rc.conf
1.93      gdt       292: 
1.186     gdt       293: Ensure that the boot scripts installed in
                    294: `/usr/pkg/share/examples/rc.d` are in `/etc/rc.d`, either because you
                    295: have `PKG_RCD_SCRIPTS=yes`, or manually.  (This is not special to Xen,
                    296: but a normal part of pkgsrc usage.)
1.159     maxv      297: 
1.186     gdt       298: Set `xencommons=YES` in rc.conf:
1.31      gdt       299: 
1.157     maxv      300: [[!template id=filecontent name="/etc/rc.conf" text="""
1.152     maxv      301: xencommons=YES
                    302: """]]
1.31      gdt       303: 
1.187     gdt       304: \todo Recommend for/against xen-watchdog.
1.191     gdt       306: ### Testing
1.158     maxv      308: Now, reboot so that you are running a DOM0 kernel under Xen, rather
                    309: than GENERIC without Xen.
1.157     maxv      310: 
1.158     maxv      311: Once the reboot is done, use `xl` to inspect Xen's boot messages,
1.157     maxv      312: available resources, and running domains.  For example:
1.34      gdt       313: 
1.153     maxv      314: [[!template id=programlisting text="""
                    315: # xl dmesg
                    316: ... xen's boot info ...
                    317: # xl info
                    318: ... available memory, etc ...
                    319: # xl list
                    320: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
                    321: Domain-0           0       64    0  r----     58.1
                    322: """]]
                    324: Xen logs will be in /var/log/xen.
1.33      gdt       325: 
1.88      gdt       326: ### Issues with xencommons
1.157     maxv      328: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
1.88      gdt       329: domUs.  It does not currently work to stop and start xenstored.
                    330: Certainly all domUs should be shutdown first, following the sort order
                    331: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
                    332: and is not notified when xenstored exits, leading to not recreating
                    333: the state when the new xenstored starts.  Until there's a mechanism to
                    334: make this work, one should not expect to be able to restart xenstored
                    335: (and thus xencommons).  There is currently no reason to expect that
                    336: this will get fixed any time soon.
1.187     gdt       337: \todo Confirm if this is still true in 2020.
1.82      gdt       338: 
1.181     gdt       339: ## Xen-specific NetBSD issues
1.40      gdt       340: 
                    341: There are (at least) two additional things different about NetBSD as a
                    342: dom0 kernel compared to hardware.
1.191     gdt       344: One is that through NetBSD 9 the module ABI is different because some
                    345: of the #defines change, so there are separate sets of modules in
1.205     gdt       346: /stand.  (Further, zfs in Xen is troubled because of differing
                    347: MAXPHYS; see the zfs howto for more.)  In NetBSD-current, there is
                    348: only one set of modules.
1.40      gdt       349: 
                    350: The other difference is that XEN3_DOM0 does not have exactly the same
1.205     gdt       351: options as GENERIC.  While this is roughly agreed to be in large part
                    352: a bug, users should be aware of this and can simply add missing config
1.40      gdt       353: items if desired.
1.187     gdt       355: Finally, there have been occasional reports of trouble with X11
1.205     gdt       356: servers in NetBSD as a dom0.  Some hardware support is intentionally
                    357: disabled in XEN3_DOM0.
1.187     gdt       358: 
                    359: ## Updating Xen in a dom0
1.206   ! gdt       361: Note the previous advice to maintain a working and tested boot config
        !           362: into GENERIC without Xen.
        !           363: 
        !           364: Updating Xen in a dom0 consists of updating the xnekernel and xentools
        !           365: packages, along with copying the xen.gz into place, and of course
        !           366: rebooting.
        !           367: 
        !           368: If updating along a Xen minor version, e.g. from 4.13.1 to 4.13.2, or
        !           369: from 4.13.2nb1 to 4.13.2nb3, it is very likely that this can be done
        !           370: on a running system.  The point is that the xentools programs will be
        !           371: replaced, and you will be using "xl" from the new installation to talk
        !           372: to the older programs which are still running.  Problems from this
        !           373: update path should be reported.
        !           374: 
        !           375: For added safety, shutdown all domUs before updating, to remove the
        !           376: need for new xl to talk to old xenstored.  Note that Xen does not
        !           377: guarantee stability of internal ABIs.
        !           378: 
        !           379: If updating across Xen minor versions, e.g. from 4.11 to 4.13, the
        !           380: likelihood of trouble is increased.  Therefore, 'make replace' of
        !           381: xentools on a dom0 with running domUs is not recommended.  A shutdown
        !           382: on all domUs before replacing xentools is likely sufficient.  A safer
        !           383: appraoch is to boot into GENERIC to replace the packages, as then no
        !           384: Xen code will be running.  Single user is another option.
1.187     gdt       385: 
1.181     gdt       386: ## Updating NetBSD in a dom0
1.15      gdt       387: 
                    388: This is just like updating NetBSD on bare hardware, assuming the new
                    389: version supports the version of Xen you are running.  Generally, one
                    390: replaces the kernel and reboots, and then overlays userland binaries
1.157     maxv      391: and adjusts `/etc`.
1.15      gdt       392: 
1.191     gdt       393: Note that one should update both the non-Xen kernel typically used for
                    394: rescue purposes, as well as the DOM0 kernel used with Xen.
1.15      gdt       395: 
1.187     gdt       396: ## anita (for testing NetBSD)
1.205     gdt       398: With a NetBSD dom0, even without any domUs, one can run anita (see
                    399: pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as root,
                    400: because anita must create a domU):
1.187     gdt       401: 
                    402: [[!template id=programlisting text="""
                    403: anita --vmm=xl test file:///usr/obj/i386/
                    404: """]]
                    406: # Unprivileged domains (domU)
1.28      gdt       407: 
                    408: This section describes general concepts about domUs.  It does not
1.33      gdt       409: address specific domU operating systems or how to install them.  The
1.157     maxv      410: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
1.60      wiki      411: typically named so that the file name, domU name and the domU's host
1.33      gdt       412: name match.
1.111     wiz       414: The domU is provided with CPU and memory by Xen, configured by the
1.33      gdt       415: dom0.  The domU is provided with disk and network by the dom0,
                    416: mediated by Xen, and configured in the dom0.
                    418: Entropy in domUs can be an issue; physical disks and network are on
                    419: the dom0.  NetBSD's /dev/random system works, but is often challenged.
1.181     gdt       421: ## Config files
1.48      gdt       422: 
1.200     gdt       423: See /usr/pkg/share/examples/xen/xlexample* for a very small number of
                    424: examples for running GNU/Linux.
1.48      gdt       425: 
1.205     gdt       426: The following is an example minimal domain configuration file.  The
                    427: domU serves as a network file server.
1.157     maxv      428: 
                    429: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
                    430: name = "domU-id"
                    431: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
                    432: memory = 1024
                    433: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
                    434: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
                    435:          'file:/n0/xen/foo-wd1,0x1,w' ]
                    436: """]]
1.48      gdt       437: 
1.157     maxv      438: The domain will have name given in the `name` setting.  The kernel has the
1.48      gdt       439: host/domU name in it, so that on the dom0 one can update the various
1.157     maxv      440: domUs independently.  The `vif` line causes an interface to be provided,
1.48      gdt       441: with a specific mac address (do not reuse MAC addresses!), in bridge
                    442: mode.  Two disks are provided, and they are both writable; the bits
                    443: are stored in files and Xen attaches them to a vnd(4) device in the
1.111     wiz       444: dom0 on domain creation.  The system treats xbd0 as the boot device
1.48      gdt       445: without needing explicit configuration.
1.205     gdt       447: There is not a type line; that implicitly defines a pv domU.
                    448: Otherwise, one sets type to the lower-case version of the domU type in
                    449: the table above; see later sections.
1.187     gdt       450: 
1.163     gson      451: By convention, domain config files are kept in `/usr/pkg/etc/xen`.  Note
1.157     maxv      452: that "xl create" takes the name of a config file, while other commands
                    453: take the name of a domain.
                    455: Examples of commands:
1.48      gdt       456: 
1.157     maxv      457: [[!template id=programlisting text="""
1.163     gson      458: xl create /usr/pkg/etc/xen/foo
                    459: xl console domU-id
                    460: xl create -c /usr/pkg/etc/xen/foo
                    461: xl shutdown domU-id
1.157     maxv      462: xl list
                    463: """]]
                    465: Typing `^]` will exit the console session.  Shutting down a domain is
1.48      gdt       466: equivalent to pushing the power button; a NetBSD domU will receive a
                    467: power-press event and do a clean shutdown.  Shutting down the dom0
                    468: will trigger controlled shutdowns of all configured domUs.
1.181     gdt       470: ## CPU and memory
1.33      gdt       471: 
1.205     gdt       472: A domain is provided with some number of vcpus; any domain can have up
                    473: to the number of CPUs seen by the hypervisor. For a domU, it is
                    474: controlled from the config file by the "vcpus = N" directive.  It is
                    475: normal to overcommit vcpus; a 4-core machine machine might well provide 4
                    476: vcpus to each domU.  One might also configure fewer vcpus for a domU.
1.48      gdt       477: 
                    478: A domain is provided with memory; this is controlled in the config
                    479: file by "memory = N" (in megabytes).  In the straightforward case, the
                    480: sum of the the memory allocated to the dom0 and all domUs must be less
1.33      gdt       481: than the available memory.
1.197     gdt       483: ## Balloon driver
                    485: Xen provides a `balloon` driver, which can be used to let domains use
                    486: more memory temporarily.
                    488: \todo Explain how to set up a aystem to use the balloon scheme in a
                    489: useful manner.
1.28      gdt       490: 
1.181     gdt       491: ## Virtual disks
1.28      gdt       492: 
1.158     maxv      493: In domU config files, the disks are defined as a sequence of 3-tuples:
                    495:  * The first element is "method:/path/to/disk". Common methods are
1.162     gutterid  496:    "file:" for a file-backed vnd, and "phy:" for something that is already
                    497:    a device, such as an LVM logical volume.
1.158     maxv      498: 
                    499:  * The second element is an artifact of how virtual disks are passed to
                    500:    Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
                    501:    are given a device name to associate with the disk, and values like
                    502:    "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
                    503:    as xbd0, the second as xbd1, and so on.  However, xl demands a
                    504:    second argument.  The name given is converted to a major/minor by
                    505:    calling stat(2) on the name in /dev and this is passed to the domU.
                    506:    In the general case, the dom0 and domU can be different operating
                    507:    systems, and it is an unwarranted assumption that they have consistent
                    508:    numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
                    509:    as both dom0 and domU, using values of 0x0 for the first disk and 0x1
                    510:    for the second works fine and avoids this issue.  For a GNU/Linux
                    511:    guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
                    512:    /dev/hda1.
1.48      gdt       513: 
1.158     maxv      514:  * The third element is "w" for writable disks, and "r" for read-only
                    515:    disks.
                    517: Example:
                    518: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
                    519: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ]
                    520: """]]
1.28      gdt       521: 
1.127     gdt       522: Note that NetBSD by default creates only vnd[0123].  If you need more
                    523: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
                    524: dom0.
1.187     gdt       526: ## Virtual Networking
1.28      gdt       527: 
1.111     wiz       528: Xen provides virtual Ethernets, each of which connects the dom0 and a
1.46      gdt       529: domU.  For each virtual network, there is an interface "xvifN.M" in
1.158     maxv      530: the dom0, and a matching interface xennetM (NetBSD name) in domU index N.
                    531: The interfaces behave as if there is an Ethernet with two
1.111     wiz       532: adapters connected.  From this primitive, one can construct various
1.46      gdt       533: configurations.  We focus on two common and useful cases for which
                    534: there are existing scripts: bridging and NAT.
1.28      gdt       535: 
1.48      gdt       536: With bridging (in the example above), the domU perceives itself to be
                    537: on the same network as the dom0.  For server virtualization, this is
                    538: usually best.  Bridging is accomplished by creating a bridge(4) device
                    539: and adding the dom0's physical interface and the various xvifN.0
                    540: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
                    541: config file.  The bridge must be set up already in the dom0; an
                    542: example /etc/ifconfig.bridge0 is:
1.46      gdt       543: 
1.157     maxv      544: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
                    545: create
                    546: up
                    547: !brconfig bridge0 add wm0
                    548: """]]
1.28      gdt       549: 
                    550: With NAT, the domU perceives itself to be behind a NAT running on the
                    551: dom0.  This is often appropriate when running Xen on a workstation.
1.48      gdt       552: TODO: NAT appears to be configured by "vif = [ '' ]".
1.28      gdt       553: 
1.49      gdt       554: The MAC address specified is the one used for the interface in the new
1.53      gdt       555: domain.  The interface in dom0 will use this address XOR'd with
1.49      gdt       556: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
1.187     gdt       558: ## Starting domains automatically
1.28      gdt       559: 
1.158     maxv      560: To start domains `domU-netbsd` and `domU-linux` at boot and shut them
                    561: down cleanly on dom0 shutdown, add the following in rc.conf:
1.28      gdt       562: 
1.158     maxv      563: [[!template id=filecontent name="/etc/rc.conf" text="""
                    564: xendomains="domU-netbsd domU-linux"
                    565: """]]
1.28      gdt       566: 
1.188     gdt       567: # domU setup for specific systems
1.14      gdt       568: 
                    569: Creating domUs is almost entirely independent of operating system.  We
1.188     gdt       570: have already presented the basics of config files in the previous system.
                    572: Of course, this section presumes that you have a working dom0.
1.14      gdt       573: 
1.187     gdt       574: ## Creating a NetBSD PV domU
1.49      gdt       576: See the earlier config file, and adjust memory.  Decide on how much
1.162     gutterid  577: storage you will provide, and prepare it (file or LVM).
1.49      gdt       578: 
1.111     wiz       579: While the kernel will be obtained from the dom0 file system, the same
1.49      gdt       580: file should be present in the domU as /netbsd so that tools like
                    581: savecore(8) can work.   (This is helpful but not necessary.)
1.188     gdt       583: The kernel must be specifically built for Xen, to use PV interfacesas
                    584: a domU.  NetBSD release builds provide the following kernels:
1.49      gdt       585: 
                    586:         i386 XEN3PAE_DOMU
1.95      gdt       587:         amd64 XEN3_DOMU
1.5       mspo      588: 
1.49      gdt       589: This will boot NetBSD, but this is not that useful if the disk is
1.188     gdt       590: empty.  One approach is to unpack sets onto the disk outside of Xen
1.49      gdt       591: (by mounting it, just as you would prepare a physical disk for a
                    592: system you can't run the installer on).
                    594: A second approach is to run an INSTALL kernel, which has a miniroot
                    595: and can load sets from the network.  To do this, copy the INSTALL
                    596: kernel to / and change the kernel line in the config file to:
1.5       mspo      597: 
1.49      gdt       598:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
1.5       mspo      599: 
1.163     gson      600: Then, start the domain as "xl create -c configfile".
1.1       mspo      601: 
1.49      gdt       602: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
                    603: line should be used in the config file.
1.1       mspo      604: 
1.3       mspo      605:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
1.1       mspo      606: 
                    607: After booting the domain, the option to install via CDROM may be
1.49      gdt       608: selected.  The CDROM device should be changed to `xbd1d`.
1.1       mspo      609: 
1.188     gdt       610: Once done installing, "halt -p" the new domain (don't reboot or halt:
1.49      gdt       611: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
                    612: config file), switch the config file back to the XEN3_DOMU kernel,
                    613: and start the new domain again. Now it should be able to use "root on
1.188     gdt       614: xbd0a" and you should have a functional NetBSD domU.
1.1       mspo      615: 
1.49      gdt       616: TODO: check if this is still accurate.
1.1       mspo      617: When the new domain is booting you'll see some warnings about *wscons*
                    618: and the pseudo-terminals. These can be fixed by editing the files
1.5       mspo      619: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
                    620: `/etc/ttys`, except *console*, like this:
1.1       mspo      621: 
1.3       mspo      622:     console "/usr/libexec/getty Pc"         vt100   on secure
                    623:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
                    624:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
                    625:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
                    626:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
1.1       mspo      627: 
1.5       mspo      628: Finally, all screens must be commented out from `/etc/wscons.conf`.
1.1       mspo      629: 
1.192     gdt       630: One should also run `powerd` in a domU, but this should not need
                    631: configuring.  With powerd, the domain will run a controlled shutdown
                    632: if `xl shutdown -R` or `xl shutdown -H` is used on the dom0, via
                    633: receiving a synthetic `power button pressed` signal.  In 9 and
                    634: current, `powerd` is run by default under Xen kernels (or if ACPI is
                    635: present), and it can be added to rc.conf if not.
1.1       mspo      636: 
1.92      gdt       637: It is not strictly necessary to have a kernel (as /netbsd) in the domU
1.111     wiz       638: file system.  However, various programs (e.g. netstat) will use that
1.92      gdt       639: kernel to look up symbols to read from kernel virtual memory.  If
                    640: /netbsd is not the running kernel, those lookups will fail.  (This is
                    641: not really a Xen-specific issue, but because the domU kernel is
                    642: obtained from the dom0, it is far more likely to be out of sync or
                    643: missing with Xen.)
1.187     gdt       645: Note that NetBSD by default creates only xbd[0123].  If you need more
                    646: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
1.196     gdt       648: ## Creating a Linux PV domU
1.1       mspo      649: 
                    650: Creating unprivileged Linux domains isn't much different from
                    651: unprivileged NetBSD domains, but there are some details to know.
                    653: First, the second parameter passed to the disk declaration (the '0x1' in
                    654: the example below)
1.3       mspo      656:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
1.1       mspo      657: 
                    658: does matter to Linux. It wants a Linux device number here (e.g. 0x300
1.49      gdt       659: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
                    660: So, hda1 which has major 3 and minor 1 on a Linux system will have
                    661: device number 0x301.  Alternatively, devices names can be used (hda,
                    662: hdb, ...)  as xentools has a table to map these names to devices
                    663: numbers.  To export a partition to a Linux guest we can use:
1.1       mspo      664: 
1.49      gdt       665:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
                    666:         root = "/dev/hda1 ro"
1.1       mspo      667: 
                    668: and it will appear as /dev/hda on the Linux system, and be used as root
                    669: partition.
1.49      gdt       671: To install the Linux system on the partition to be exported to the
                    672: guest domain, the following method can be used: install
                    673: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
                    674: that will be the root partition of your Linux domain, and mount it.
                    675: Then copy the files from a working Linux system, make adjustments in
                    676: `/etc` (fstab, network config).  It should also be possible to extract
                    677: binary packages such as .rpm or .deb directly to the mounted partition
                    678: using the appropriate tool, possibly running under NetBSD's Linux
1.111     wiz       679: emulation.  Once the file system has been populated, umount it.  If
                    680: desirable, the file system can be converted to ext3 using tune2fs -j.
1.49      gdt       681: It should now be possible to boot the Linux guest domain, using one of
                    682: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
1.1       mspo      683: 
1.111     wiz       684: To get the Linux console right, you need to add:
1.1       mspo      685: 
1.3       mspo      686:     extra = "xencons=tty1"
1.1       mspo      687: 
1.111     wiz       688: to your configuration since not all Linux distributions auto-attach a
1.1       mspo      689: tty to the xen console.
1.180     gdt       691: ## Creating a NetBSD HVM domU
1.196     gdt       693: Use type='hvm', probably.  Use a GENERIC kernel within the disk image.
1.180     gdt       694: 
                    695: ## Creating a NetBSD PVH domU
1.196     gdt       697: This only works with a current kernel in the domU.
1.180     gdt       698: 
1.196     gdt       699: Use type='pvh'.  Probably, use a GENERIC kernel within the disk image,
                    700: which in current has PV support.
                    702: \todo Verify.
                    704: \todo Verify if one can have current PVH domU on a 9 dom0.
1.180     gdt       705: 
1.187     gdt       706: ## Creating a Solaris domU
1.1       mspo      707: 
1.50      gdt       708: See possibly outdated
                    709: [Solaris domU instructions](/ports/xen/howto-solaris/).
1.5       mspo      710: 
1.187     gdt       711: ## PCI passthrough: Using PCI devices in guest domains
1.1       mspo      712: 
1.180     gdt       713: NB: PCI passthrough only works on some Xen versions and as of 2020 it
1.196     gdt       714: is not clear that it works on any version in pkgsrc.  \todo Reports
1.180     gdt       715: confirming or denying this notion should be sent to port-xen@.
1.53      gdt       717: The dom0 can give other domains access to selected PCI
1.52      gdt       718: devices. This can allow, for example, a non-privileged domain to have
                    719: access to a physical network interface or disk controller.  However,
                    720: keep in mind that giving a domain access to a PCI device most likely
                    721: will give the domain read/write access to the whole physical memory,
                    722: as PCs don't have an IOMMU to restrict memory access to DMA-capable
1.53      gdt       723: device.  Also, it's not possible to export ISA devices to non-dom0
1.52      gdt       724: domains, which means that the primary VGA adapter can't be exported.
                    725: A guest domain trying to access the VGA registers will panic.
1.53      gdt       727: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
1.52      gdt       728: not been ported to later versions at this time.
                    730: For a PCI device to be exported to a domU, is has to be attached to
                    731: the "pciback" driver in dom0.  Devices passed to the dom0 via the
                    732: pciback.hide boot parameter will attach to "pciback" instead of the
                    733: usual driver.  The list of devices is specified as "(bus:dev.func)",
1.5       mspo      734: where bus and dev are 2-digit hexadecimal numbers, and func a
                    735: single-digit number:
1.1       mspo      736: 
1.52      gdt       737:         pciback.hide=(00:0a.0)(00:06.0)
1.1       mspo      738: 
1.52      gdt       739: pciback devices should show up in the dom0's boot messages, and the
1.5       mspo      740: devices should be listed in the `/kern/xen/pci` directory.
1.1       mspo      741: 
1.52      gdt       742: PCI devices to be exported to a domU are listed in the "pci" array of
                    743: the domU's config file, with the format "0000:bus:dev.func".
1.1       mspo      744: 
1.52      gdt       745:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
1.1       mspo      746: 
1.52      gdt       747: In the domU an "xpci" device will show up, to which one or more pci
1.111     wiz       748: buses will attach.  Then the PCI drivers will attach to PCI buses as
1.52      gdt       749: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
                    750: or any PCI drivers built in by default; you have to build your own
                    751: kernel to use PCI devices in a domU.  Here's a kernel config example;
                    752: note that only the "xpci" lines are unusual.
                    754:         include         "arch/i386/conf/XEN3_DOMU"
1.111     wiz       756:         # Add support for PCI buses to the XEN3_DOMU kernel
1.52      gdt       757:         xpci* at xenbus ?
                    758:         pci* at xpci ?
                    760:         # PCI USB controllers
                    761:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
                    763:         # USB bus support
                    764:         usb*    at uhci?
                    766:         # USB Hubs
                    767:         uhub*   at usb?
                    768:         uhub*   at uhub? port ? configuration ? interface ?
                    770:         # USB Mass Storage
                    771:         umass*  at uhub? port ? configuration ? interface ?
                    772:         wd*     at umass?
                    773:         # SCSI controllers
                    774:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
                    776:         # SCSI bus support (for both ahc and umass)
                    777:         scsibus* at scsi?
                    779:         # SCSI devices
                    780:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
                    781:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
1.1       mspo      782: 
1.189     gdt       784: # Miscellaneous Information
1.173     gdt       785: 
1.189     gdt       786: ## Nesting under Linux KVM
1.205     gdt       788: It is possible to run Xen and a NetBSD dom0 under Linux KVM.  One
1.189     gdt       789: can enable virtio in the dom0 for greater speed.
1.205     gdt       791: ## Nesting under qemu
                    793: It is possible to run Xen and a NetBSD dom0 under qemu on NetBSD, and
                    794: also with nvmm.
                    795: \todo Check this.
1.189     gdt       797: ## Other nesting
                    799: In theory, any full emulation should be able to run Xen and a NetBSD
                    800: dom0.  The HOWTO does not currently have information about Xen XVM
1.205     gdt       801: mode, Virtualbox, etc.
1.189     gdt       802: 
                    803: ## NetBSD 5 as domU
1.173     gdt       804: 
1.176     gdt       805: [NetBSD 5 is known to panic.](
1.175     gdt       806: (However, NetBSD 5 systems should be updated to a supported version.)
1.173     gdt       807: 
                    808: # NetBSD as a domU in a VPS
1.28      gdt       809: 
                    810: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
                    811: hardware.  This section explains how to deal with Xen in a domU as a
                    812: virtual private server where you do not control or have access to the
1.70      gdt       813: dom0.  This is not intended to be an exhaustive list of VPS providers;
                    814: only a few are mentioned that specifically support NetBSD.
1.28      gdt       815: 
1.52      gdt       816: VPS operators provide varying degrees of access and mechanisms for
                    817: configuration.  The big issue is usually how one controls which kernel
1.111     wiz       818: is booted, because the kernel is nominally in the dom0 file system (to
                    819: which VPS users do not normally have access).  A second issue is how
1.70      gdt       820: to install NetBSD.
1.52      gdt       821: A VPS user may want to compile a kernel for security updates, to run
                    822: npf, run IPsec, or any other reason why someone would want to change
                    823: their kernel.
1.111     wiz       825: One approach is to have an administrative interface to upload a kernel,
1.68      gdt       826: or to select from a prepopulated list.  Other approaches are pygrub
1.59      gdt       827: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
1.111     wiz       828: kernel from the domU file system.  This is closer to a regular physical
1.59      gdt       829: computer, where someone who controls a machine can replace the kernel.
1.52      gdt       830: 
1.74      gdt       831: A second issue is multiple CPUs.  With NetBSD 6, domUs support
                    832: multiple vcpus, and it is typical for VPS providers to enable multiple
                    833: CPUs for NetBSD domUs.
1.180     gdt       835: ## Complexities due to Xen changes
                    837: Xen has many security advisories and people running Xen systems make
                    838: different choices.
                    840: ### stub domains
1.205     gdt       842: Some (Linux) dom0 systems use something called "stub domains" to
1.180     gdt       843: isolate qemu from the dom0 system, as a security and reliabilty
                    844: mechanism when running HVM domUs.  Somehow, NetBSD's GENERIC kernel
                    845: ends up using PIO for disks rather than DMA.  Of course, all of this
                    846: is emulated, but emulated PIO is unusably slow.  This problem is not
                    847: currently understood.
                    849: ### Grant tables
                    851: There are multiple versions of using grant tables, and some security
1.205     gdt       852: advisories have suggested disabling some versions.  NetBSD through 9
                    853: uses version 1 and NetBSD-current uses version 2.  This can lead to
1.180     gdt       854: "NetBSD current doesn't run on hosting provider X" situations.
                    856: \todo Explain better.
1.181     gdt       858: ## Boot methods
                    860: ### pvgrub
1.59      gdt       861: 
                    862: pvgrub is a version of grub that uses PV operations instead of BIOS
                    863: calls.  It is booted from the dom0 as the domU kernel, and then reads
1.111     wiz       864: /grub/menu.lst and loads a kernel from the domU file system.
1.59      gdt       865: 
1.70      gdt       866: It appears that [grub's FFS
                    867: code](
                    868: does not support all aspects of modern FFS, but there are also reports
1.205     gdt       869: that FFSv2 works fine.
1.59      gdt       870: 
1.181     gdt       871: ### pygrub
1.168     wiki      872: 
1.205     gdt       873: As of 2014, pygrub seems to be of mostly historical interest.  As of
                    874: 2021, the section should perhaps be outright deleted.
1.168     wiki      876: pygrub runs in the dom0 and looks into the domU file system.  This
                    877: implies that the domU must have a kernel in a file system in a format
                    878: known to pygrub.
                    880: pygrub doesn't seem to work to load Linux images under NetBSD dom0,
1.205     gdt       881: and is inherently less secure than pvgrub due to running inside
                    882: dom0. For both these reasons, pygrub should not be used, and is only
                    883: still present so that historical DomU images using it still work.
                    885: ## Specific Providers
                    887: The intent is to list providers only if they document support for
                    888: running NetBSD, and to point to their resources briefly.
1.168     wiki      889: 
1.205     gdt       890: ###
1.168     wiki      891: 
1.205     gdt       892: [Panix]( provides NetBSD as an OS option.  See
                    893: for some information.
                    894: Users can use pvgrub.  Panix reports that pvgrub works with FFsv2 with
                    895: 16K/2K and 32K/4K block/frag sizes (and hence with defaults from
                    896: "newfs -O 2").  See [Panix's pvgrub
                    897: page]( which describes how to
                    898: boot NetBSD.
                    900: ###
                    902: []( provides released versions of
                    903: NetBSD/amd64 as installation options.  Users can use pvgrub to boot
                    904: their own kernel, and a small FAT32 /boot is encouraged.  See the
                    905: [ NetBSD
                    906: HOWTO](
                    907: (which is in need of updating).
1.181     gdt       908: 
                    909: ### Amazon
1.59      gdt       910: 
1.143     wiki      911: See the [Amazon EC2 page](/amazon_ec2/).

CVSweb for NetBSD wikisrc <> software: FreeBSD-CVSweb