Annotation of wikisrc/ports/xen/howto.mdwn, revision 1.182

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: 
        !            18: # Overview: Versions of Xen, Styles, NetBSD
        !            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: 
        !            30: ## Guest Styles
        !            31: 
        !            32: Xen supports different styles of guests.
1.149     maxv       33: 
                     34: [[!table data="""
                     35: Style of guest |Supported by NetBSD
1.150     maxv       36: PV             |Yes (dom0, domU)
                     37: HVM            |Yes (domU)
1.171     gdt        38: PVHVM          |current-only (domU)
1.180     gdt        39: PVH            |current-only (domU, dom0 not yet)
1.149     maxv       40: """]]
                     41: 
                     42: In Para-Virtualized (PV) mode, the guest OS does not attempt to access
                     43: hardware directly, but instead makes hypercalls to the hypervisor; PV
1.170     gdt        44: guests must be specifically coded for Xen.
1.182   ! gdt        45: See [PV](https://wiki.xen.org/wiki/Paravirtualization_(PV\)).
1.170     gdt        46: 
                     47: In HVM mode, no guest modification is required; however, hardware
                     48: support is required, such as VT-x on Intel CPUs and SVM on AMD CPUs.
1.177     gdt        49: The dom0 runs qemu to emulate hardware.
1.170     gdt        50: 
1.176     gdt        51: In PVHVM mode, the guest runs as HVM, but additionally can use PV
                     52: drivers for efficiency.
1.182   ! gdt        53: See [PV on HVM](https://wiki.xen.org/wiki/PV_on_HVM).
1.170     gdt        54: 
1.180     gdt        55: There have been two PVH modes: original PVH and PVHv2.  Original PVH
                     56: was based on PV mode and is no longer relevant at all.  PVHv2 is
                     57: basically lightweight HVM with PV drivers.  A critical feature of it
                     58: is that qemu is not needed; the hypervisor can do the emulation that
                     59: is required.  Thus, a dom0 can be PVHv2.
1.182   ! gdt        60: The source code uses PVH and config files use pvh; this refers to PVHv2.
        !            61: See [PVH(v2)](https://wiki.xenproject.org/wiki/PVH_(v2\)_Domu).
1.180     gdt        62: 
1.149     maxv       63: 
1.182   ! gdt        64: At system boot, the dom0 kernel is loaded as a module with Xen as the kernel.
1.12      gdt        65: The dom0 can start one or more domUs.  (Booting is explained in detail
                     66: in the dom0 section.)
                     67: 
1.182   ! gdt        68: ## CPU Architecture
        !            69: 
        !            70: Xen runs on x86_64 hardware (the NetBSD amd64 port).
        !            71: 
        !            72: There is a concept of Xen running on ARM, but there are no reports of this working with NetBSD.
        !            73: 
        !            74: The dom0 system should be amd64.  (Instructions for i386PAE dom0 have been removed from the HOWTO.)
1.1       mspo       75: 
1.182   ! gdt        76: The domU can be i386PAE or amd64.
        !            77: 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       78: 
1.182   ! gdt        79: ## Xen Versions
1.15      gdt        80: 
1.111     wiz        81: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
1.15      gdt        82: xenkernel and xentools.  We will refer only to the kernel versions,
                     83: but note that both packages must be installed together and must have
                     84: matching versions.
                     85: 
1.145     maxv       86: Versions available in pkgsrc:
1.85      gdt        87: 
1.145     maxv       88: [[!table data="""
1.166     gdt        89: Xen Version    |Package Name   |Xen CPU Support        |xm?    |EOL'ed By Upstream
1.167     gdt        90: 4.11           |xenkernel411   |x86_64                 |       |No
1.176     gdt        91: 4.13           |xenkernel413   |x86_64                 |       |No
1.145     maxv       92: """]]
1.113     gdt        93: 
1.96      gdt        94: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
                     95: 
1.157     maxv       96: Multiprocessor (SMP) support in NetBSD differs depending on the domain:
1.145     maxv       97: 
1.157     maxv       98: [[!table data="""
                     99: Domain         |Supports SMP
                    100: dom0           |No
                    101: domU           |Yes
                    102: """]]
1.15      gdt       103: 
1.167     gdt       104: Note: NetBSD support is called XEN3.  However, it does support Xen 4,
1.147     maxv      105: because the hypercall interface has remained identical.
                    106: 
1.166     gdt       107: Older Xen had a python-based management tool called xm, now replaced
1.176     gdt       108: by xl.
1.166     gdt       109: 
                    110: # Creating a dom0
1.15      gdt       111: 
1.158     maxv      112: In order to install a NetBSD as a dom0, one must first install a normal
                    113: NetBSD system, and then pivot the install to a dom0 install by changing
                    114: the kernel and boot configuration.
1.15      gdt       115: 
1.142     gdt       116: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
                    117: 512M it worked reliably.  This does not make sense, but if you see
                    118: "not ELF" after Xen boots, try increasing dom0 RAM.
                    119: 
1.181     gdt       120: ## Installation of NetBSD
1.13      gdt       121: 
1.157     maxv      122: [Install NetBSD/amd64](/guide/inst/)
1.19      gdt       123: just as you would if you were not using Xen.
1.1       mspo      124: 
1.181     gdt       125: ## Installation of Xen
1.1       mspo      126: 
1.178     gdt       127: We will consider that you chose to use Xen 4.13, with NetBSD/amd64 as
1.155     maxv      128: dom0. In the dom0, install xenkernel48 and xentools48 from pkgsrc.
                    129: 
                    130: Once this is done, install the Xen kernel itself:
                    131: 
                    132: [[!template id=programlisting text="""
                    133: # cp /usr/pkg/xen48-kernel/xen.gz /
                    134: """]]
                    135: 
1.159     maxv      136: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such kernel
                    137: can either be compiled manually, or downloaded from the NetBSD FTP, for
                    138: example at:
                    139: 
                    140: [[!template id=programlisting text="""
                    141: ftp.netbsd.org/pub/NetBSD/NetBSD-8.0/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
                    142: """]]
1.20      gdt       143: 
1.155     maxv      144: Add a line to /boot.cfg to boot Xen:
1.20      gdt       145: 
1.157     maxv      146: [[!template id=filecontent name="/boot.cfg" text="""
1.152     maxv      147: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
                    148: """]]
1.20      gdt       149: 
1.159     maxv      150: This specifies that the dom0 should have 512MB of ram, leaving the rest
                    151: to be allocated for domUs.  To use a serial console, use:
1.77      gdt       152: 
1.157     maxv      153: [[!template id=filecontent name="/boot.cfg" text="""
1.152     maxv      154: menu=Xen:load /netbsd-XEN3_DOM0.gz;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
                    155: """]]
1.77      gdt       156: 
                    157: which will use the first serial port for Xen (which counts starting
1.139     gson      158: from 1, unlike NetBSD which counts starting from 0), forcing
                    159: speed/parity.  Because the NetBSD command line lacks a
                    160: "console=pc" argument, it will use the default "xencons" console device,
                    161: which directs the console I/O through Xen to the same console device Xen
                    162: itself uses (in this case, the serial port).
                    163: 
1.157     maxv      164: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`,
1.37      gdt       165: to force only one vcpu to be provided (since NetBSD dom0 can't use
1.158     maxv      166: more) and to pin that vcpu to a physical CPU. Xen has
1.178     gdt       167: [many boot options](http://xenbits.xenproject.org/docs/4.13-testing/misc/xen-command-line.html),
1.111     wiz       168: and other than dom0 memory and max_vcpus, they are generally not
1.93      gdt       169: necessary.
                    170: 
1.159     maxv      171: Copy the boot scripts into `/etc/rc.d`:
                    172: 
                    173: [[!template id=programlisting text="""
                    174: # cp /usr/pkg/share/examples/rc.d/xen* /etc/rc.d/
                    175: """]]
                    176: 
1.158     maxv      177: Enable `xencommons`:
1.31      gdt       178: 
1.157     maxv      179: [[!template id=filecontent name="/etc/rc.conf" text="""
1.152     maxv      180: xencommons=YES
                    181: """]]
1.31      gdt       182: 
1.158     maxv      183: Now, reboot so that you are running a DOM0 kernel under Xen, rather
                    184: than GENERIC without Xen.
1.157     maxv      185: 
1.31      gdt       186: TODO: Recommend for/against xen-watchdog.
1.27      jnemeth   187: 
1.158     maxv      188: Once the reboot is done, use `xl` to inspect Xen's boot messages,
1.157     maxv      189: available resources, and running domains.  For example:
1.34      gdt       190: 
1.153     maxv      191: [[!template id=programlisting text="""
                    192: # xl dmesg
                    193: ... xen's boot info ...
                    194: # xl info
                    195: ... available memory, etc ...
                    196: # xl list
                    197: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
                    198: Domain-0           0       64    0  r----     58.1
                    199: """]]
                    200: 
                    201: Xen logs will be in /var/log/xen.
1.33      gdt       202: 
1.88      gdt       203: ### Issues with xencommons
                    204: 
1.157     maxv      205: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
1.88      gdt       206: domUs.  It does not currently work to stop and start xenstored.
                    207: Certainly all domUs should be shutdown first, following the sort order
                    208: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
                    209: and is not notified when xenstored exits, leading to not recreating
                    210: the state when the new xenstored starts.  Until there's a mechanism to
                    211: make this work, one should not expect to be able to restart xenstored
                    212: (and thus xencommons).  There is currently no reason to expect that
                    213: this will get fixed any time soon.
                    214: 
1.181     gdt       215: ## anita (for testing NetBSD)
1.41      gdt       216: 
1.178     gdt       217: With the setup so far, one should be able to run
1.82      gdt       218: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
                    219: root, because anita must create a domU):
                    220: 
1.153     maxv      221: [[!template id=programlisting text="""
                    222: anita --vmm=xl test file:///usr/obj/i386/
                    223: """]]
1.82      gdt       224: 
1.181     gdt       225: ## Xen-specific NetBSD issues
1.40      gdt       226: 
                    227: There are (at least) two additional things different about NetBSD as a
                    228: dom0 kernel compared to hardware.
                    229: 
1.111     wiz       230: One is that the module ABI is different because some of the #defines
1.109     gdt       231: change, so one must build modules for Xen.  As of netbsd-7, the build
1.157     maxv      232: system does this automatically.
1.40      gdt       233: 
                    234: The other difference is that XEN3_DOM0 does not have exactly the same
                    235: options as GENERIC.  While it is debatable whether or not this is a
                    236: bug, users should be aware of this and can simply add missing config
                    237: items if desired.
                    238: 
1.181     gdt       239: ## Updating NetBSD in a dom0
1.15      gdt       240: 
                    241: This is just like updating NetBSD on bare hardware, assuming the new
                    242: version supports the version of Xen you are running.  Generally, one
                    243: replaces the kernel and reboots, and then overlays userland binaries
1.157     maxv      244: and adjusts `/etc`.
1.15      gdt       245: 
                    246: Note that one must update both the non-Xen kernel typically used for
                    247: rescue purposes and the DOM0 kernel used with Xen.
                    248: 
1.181     gdt       249: ## Converting from grub to /boot
1.55      gdt       250: 
1.157     maxv      251: These instructions were used to convert a system from
1.55      gdt       252: grub to /boot.  The system was originally installed in February of
                    253: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
                    254: over time.  Before these commands, it was running NetBSD 6 i386, Xen
                    255: 4.1 and grub, much like the message linked earlier in the grub
                    256: section.
                    257: 
1.157     maxv      258: [[!template id=programlisting text="""
                    259: # Install MBR bootblocks on both disks.
                    260: fdisk -i /dev/rwd0d
                    261: fdisk -i /dev/rwd1d
                    262: # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
                    263: installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
                    264: installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
                    265: # Install secondary boot loader
                    266: cp -p /usr/mdec/boot /
                    267: # Create boot.cfg following earlier guidance:
                    268: menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
                    269: menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=512M
                    270: menu=GENERIC:boot
                    271: menu=GENERIC single-user:boot -s
                    272: menu=GENERIC.ok:boot netbsd.ok
                    273: menu=GENERIC.ok single-user:boot netbsd.ok -s
                    274: menu=Drop to boot prompt:prompt
                    275: default=1
                    276: timeout=30
                    277: """]]
1.22      gdt       278: 
1.181     gdt       279: ## Upgrading Xen versions
1.15      gdt       280: 
1.110     gdt       281: Minor version upgrades are trivial.  Just rebuild/replace the
1.157     maxv      282: xenkernel version and copy the new xen.gz to `/` (where `/boot.cfg`
1.110     gdt       283: references it), and reboot.
                    284: 
1.160     maxv      285: #Unprivileged domains (domU)
1.28      gdt       286: 
                    287: This section describes general concepts about domUs.  It does not
1.33      gdt       288: address specific domU operating systems or how to install them.  The
1.157     maxv      289: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
1.60      wiki      290: typically named so that the file name, domU name and the domU's host
1.33      gdt       291: name match.
                    292: 
1.111     wiz       293: The domU is provided with CPU and memory by Xen, configured by the
1.33      gdt       294: dom0.  The domU is provided with disk and network by the dom0,
                    295: mediated by Xen, and configured in the dom0.
                    296: 
                    297: Entropy in domUs can be an issue; physical disks and network are on
                    298: the dom0.  NetBSD's /dev/random system works, but is often challenged.
                    299: 
1.181     gdt       300: ## Config files
1.48      gdt       301: 
1.165     gson      302: See /usr/pkg/share/examples/xen/xlexample*
1.164     gson      303: for a small number of well-commented examples, mostly for running
1.48      gdt       304: GNU/Linux.
                    305: 
1.157     maxv      306: The following is an example minimal domain configuration file. The domU
                    307: serves as a network file server.
                    308: 
                    309: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
                    310: name = "domU-id"
                    311: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
                    312: memory = 1024
                    313: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
                    314: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
                    315:          'file:/n0/xen/foo-wd1,0x1,w' ]
                    316: """]]
1.48      gdt       317: 
1.157     maxv      318: The domain will have name given in the `name` setting.  The kernel has the
1.48      gdt       319: host/domU name in it, so that on the dom0 one can update the various
1.157     maxv      320: domUs independently.  The `vif` line causes an interface to be provided,
1.48      gdt       321: with a specific mac address (do not reuse MAC addresses!), in bridge
                    322: mode.  Two disks are provided, and they are both writable; the bits
                    323: are stored in files and Xen attaches them to a vnd(4) device in the
1.111     wiz       324: dom0 on domain creation.  The system treats xbd0 as the boot device
1.48      gdt       325: without needing explicit configuration.
                    326: 
1.163     gson      327: By convention, domain config files are kept in `/usr/pkg/etc/xen`.  Note
1.157     maxv      328: that "xl create" takes the name of a config file, while other commands
                    329: take the name of a domain.
                    330: 
                    331: Examples of commands:
1.48      gdt       332: 
1.157     maxv      333: [[!template id=programlisting text="""
1.163     gson      334: xl create /usr/pkg/etc/xen/foo
                    335: xl console domU-id
                    336: xl create -c /usr/pkg/etc/xen/foo
                    337: xl shutdown domU-id
1.157     maxv      338: xl list
                    339: """]]
                    340: 
                    341: Typing `^]` will exit the console session.  Shutting down a domain is
1.48      gdt       342: equivalent to pushing the power button; a NetBSD domU will receive a
                    343: power-press event and do a clean shutdown.  Shutting down the dom0
                    344: will trigger controlled shutdowns of all configured domUs.
                    345: 
1.181     gdt       346: ## CPU and memory
1.33      gdt       347: 
1.169     plunky    348: A domain is provided with some number of vcpus, up to the number
1.157     maxv      349: of CPUs seen by the hypervisor. For a domU, it is controlled
1.48      gdt       350: from the config file by the "vcpus = N" directive.
                    351: 
                    352: A domain is provided with memory; this is controlled in the config
                    353: file by "memory = N" (in megabytes).  In the straightforward case, the
                    354: sum of the the memory allocated to the dom0 and all domUs must be less
1.33      gdt       355: than the available memory.
                    356: 
                    357: Xen also provides a "balloon" driver, which can be used to let domains
1.157     maxv      358: use more memory temporarily.
1.28      gdt       359: 
1.181     gdt       360: ## Virtual disks
1.28      gdt       361: 
1.158     maxv      362: In domU config files, the disks are defined as a sequence of 3-tuples:
                    363: 
                    364:  * The first element is "method:/path/to/disk". Common methods are
1.162     gutterid  365:    "file:" for a file-backed vnd, and "phy:" for something that is already
                    366:    a device, such as an LVM logical volume.
1.158     maxv      367: 
                    368:  * The second element is an artifact of how virtual disks are passed to
                    369:    Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
                    370:    are given a device name to associate with the disk, and values like
                    371:    "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
                    372:    as xbd0, the second as xbd1, and so on.  However, xl demands a
                    373:    second argument.  The name given is converted to a major/minor by
                    374:    calling stat(2) on the name in /dev and this is passed to the domU.
                    375:    In the general case, the dom0 and domU can be different operating
                    376:    systems, and it is an unwarranted assumption that they have consistent
                    377:    numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
                    378:    as both dom0 and domU, using values of 0x0 for the first disk and 0x1
                    379:    for the second works fine and avoids this issue.  For a GNU/Linux
                    380:    guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
                    381:    /dev/hda1.
1.48      gdt       382: 
1.158     maxv      383:  * The third element is "w" for writable disks, and "r" for read-only
                    384:    disks.
                    385: 
                    386: Example:
                    387: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
                    388: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ]
                    389: """]]
1.28      gdt       390: 
1.127     gdt       391: Note that NetBSD by default creates only vnd[0123].  If you need more
                    392: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
                    393: dom0.
                    394: 
                    395: Note that NetBSD by default creates only xbd[0123].  If you need more
                    396: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
                    397: 
1.28      gdt       398: Virtual Networking
                    399: ------------------
                    400: 
1.111     wiz       401: Xen provides virtual Ethernets, each of which connects the dom0 and a
1.46      gdt       402: domU.  For each virtual network, there is an interface "xvifN.M" in
1.158     maxv      403: the dom0, and a matching interface xennetM (NetBSD name) in domU index N.
                    404: The interfaces behave as if there is an Ethernet with two
1.111     wiz       405: adapters connected.  From this primitive, one can construct various
1.46      gdt       406: configurations.  We focus on two common and useful cases for which
                    407: there are existing scripts: bridging and NAT.
1.28      gdt       408: 
1.48      gdt       409: With bridging (in the example above), the domU perceives itself to be
                    410: on the same network as the dom0.  For server virtualization, this is
                    411: usually best.  Bridging is accomplished by creating a bridge(4) device
                    412: and adding the dom0's physical interface and the various xvifN.0
                    413: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
                    414: config file.  The bridge must be set up already in the dom0; an
                    415: example /etc/ifconfig.bridge0 is:
1.46      gdt       416: 
1.157     maxv      417: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
                    418: create
                    419: up
                    420: !brconfig bridge0 add wm0
                    421: """]]
1.28      gdt       422: 
                    423: With NAT, the domU perceives itself to be behind a NAT running on the
                    424: dom0.  This is often appropriate when running Xen on a workstation.
1.48      gdt       425: TODO: NAT appears to be configured by "vif = [ '' ]".
1.28      gdt       426: 
1.49      gdt       427: The MAC address specified is the one used for the interface in the new
1.53      gdt       428: domain.  The interface in dom0 will use this address XOR'd with
1.49      gdt       429: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
                    430: 
1.48      gdt       431: Starting domains automatically
                    432: ------------------------------
1.28      gdt       433: 
1.158     maxv      434: To start domains `domU-netbsd` and `domU-linux` at boot and shut them
                    435: down cleanly on dom0 shutdown, add the following in rc.conf:
1.28      gdt       436: 
1.158     maxv      437: [[!template id=filecontent name="/etc/rc.conf" text="""
                    438: xendomains="domU-netbsd domU-linux"
                    439: """]]
1.28      gdt       440: 
1.180     gdt       441: # Creating a domU
1.14      gdt       442: 
                    443: Creating domUs is almost entirely independent of operating system.  We
1.49      gdt       444: have already presented the basics of config files.  Note that you must
1.158     maxv      445: have already completed the dom0 setup so that "xl list" works.
1.14      gdt       446: 
1.180     gdt       447: Creating a NetBSD PV domU
                    448: --------------------------
1.1       mspo      449: 
1.49      gdt       450: See the earlier config file, and adjust memory.  Decide on how much
1.162     gutterid  451: storage you will provide, and prepare it (file or LVM).
1.49      gdt       452: 
1.111     wiz       453: While the kernel will be obtained from the dom0 file system, the same
1.49      gdt       454: file should be present in the domU as /netbsd so that tools like
                    455: savecore(8) can work.   (This is helpful but not necessary.)
                    456: 
                    457: The kernel must be specifically for Xen and for use as a domU.  The
                    458: i386 and amd64 provide the following kernels:
                    459: 
                    460:         i386 XEN3PAE_DOMU
1.95      gdt       461:         amd64 XEN3_DOMU
1.5       mspo      462: 
1.49      gdt       463: This will boot NetBSD, but this is not that useful if the disk is
                    464: empty.  One approach is to unpack sets onto the disk outside of xen
                    465: (by mounting it, just as you would prepare a physical disk for a
                    466: system you can't run the installer on).
                    467: 
                    468: A second approach is to run an INSTALL kernel, which has a miniroot
                    469: and can load sets from the network.  To do this, copy the INSTALL
                    470: kernel to / and change the kernel line in the config file to:
1.5       mspo      471: 
1.49      gdt       472:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
1.5       mspo      473: 
1.163     gson      474: Then, start the domain as "xl create -c configfile".
1.1       mspo      475: 
1.49      gdt       476: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
                    477: line should be used in the config file.
1.1       mspo      478: 
1.3       mspo      479:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
1.1       mspo      480: 
                    481: After booting the domain, the option to install via CDROM may be
1.49      gdt       482: selected.  The CDROM device should be changed to `xbd1d`.
1.1       mspo      483: 
1.49      gdt       484: Once done installing, "halt -p" the new domain (don't reboot or halt,
                    485: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
                    486: config file), switch the config file back to the XEN3_DOMU kernel,
                    487: and start the new domain again. Now it should be able to use "root on
                    488: xbd0a" and you should have a, functional NetBSD domU.
1.1       mspo      489: 
1.49      gdt       490: TODO: check if this is still accurate.
1.1       mspo      491: When the new domain is booting you'll see some warnings about *wscons*
                    492: and the pseudo-terminals. These can be fixed by editing the files
1.5       mspo      493: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
                    494: `/etc/ttys`, except *console*, like this:
1.1       mspo      495: 
1.3       mspo      496:     console "/usr/libexec/getty Pc"         vt100   on secure
                    497:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
                    498:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
                    499:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
                    500:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
1.1       mspo      501: 
1.5       mspo      502: Finally, all screens must be commented out from `/etc/wscons.conf`.
1.1       mspo      503: 
                    504: It is also desirable to add
                    505: 
1.49      gdt       506:         powerd=YES
1.1       mspo      507: 
1.5       mspo      508: in rc.conf. This way, the domain will be properly shut down if
1.53      gdt       509: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
1.1       mspo      510: 
1.92      gdt       511: It is not strictly necessary to have a kernel (as /netbsd) in the domU
1.111     wiz       512: file system.  However, various programs (e.g. netstat) will use that
1.92      gdt       513: kernel to look up symbols to read from kernel virtual memory.  If
                    514: /netbsd is not the running kernel, those lookups will fail.  (This is
                    515: not really a Xen-specific issue, but because the domU kernel is
                    516: obtained from the dom0, it is far more likely to be out of sync or
                    517: missing with Xen.)
                    518: 
1.158     maxv      519: Creating a Linux domU
                    520: ---------------------
1.1       mspo      521: 
                    522: Creating unprivileged Linux domains isn't much different from
                    523: unprivileged NetBSD domains, but there are some details to know.
                    524: 
                    525: First, the second parameter passed to the disk declaration (the '0x1' in
                    526: the example below)
                    527: 
1.3       mspo      528:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
1.1       mspo      529: 
                    530: does matter to Linux. It wants a Linux device number here (e.g. 0x300
1.49      gdt       531: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
                    532: So, hda1 which has major 3 and minor 1 on a Linux system will have
                    533: device number 0x301.  Alternatively, devices names can be used (hda,
                    534: hdb, ...)  as xentools has a table to map these names to devices
                    535: numbers.  To export a partition to a Linux guest we can use:
1.1       mspo      536: 
1.49      gdt       537:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
                    538:         root = "/dev/hda1 ro"
1.1       mspo      539: 
                    540: and it will appear as /dev/hda on the Linux system, and be used as root
                    541: partition.
                    542: 
1.49      gdt       543: To install the Linux system on the partition to be exported to the
                    544: guest domain, the following method can be used: install
                    545: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
                    546: that will be the root partition of your Linux domain, and mount it.
                    547: Then copy the files from a working Linux system, make adjustments in
                    548: `/etc` (fstab, network config).  It should also be possible to extract
                    549: binary packages such as .rpm or .deb directly to the mounted partition
                    550: using the appropriate tool, possibly running under NetBSD's Linux
1.111     wiz       551: emulation.  Once the file system has been populated, umount it.  If
                    552: desirable, the file system can be converted to ext3 using tune2fs -j.
1.49      gdt       553: It should now be possible to boot the Linux guest domain, using one of
                    554: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
1.1       mspo      555: 
1.111     wiz       556: To get the Linux console right, you need to add:
1.1       mspo      557: 
1.3       mspo      558:     extra = "xencons=tty1"
1.1       mspo      559: 
1.111     wiz       560: to your configuration since not all Linux distributions auto-attach a
1.1       mspo      561: tty to the xen console.
                    562: 
1.180     gdt       563: ## Creating a NetBSD HVM domU
                    564: 
                    565: Use type='hmv', probably.  Use a GENERIC kernel within the disk image.
                    566: 
                    567: ## Creating a NetBSD PVH domU
                    568: 
                    569: Use type='pvh'.
                    570: 
                    571: \todo Explain where the kernel comes from.
                    572: 
                    573: 
1.158     maxv      574: Creating a Solaris domU
                    575: -----------------------
1.1       mspo      576: 
1.50      gdt       577: See possibly outdated
                    578: [Solaris domU instructions](/ports/xen/howto-solaris/).
1.5       mspo      579: 
1.1       mspo      580: 
1.52      gdt       581: PCI passthrough: Using PCI devices in guest domains
                    582: ---------------------------------------------------
1.1       mspo      583: 
1.180     gdt       584: NB: PCI passthrough only works on some Xen versions and as of 2020 it
                    585: is not clear that it works on any version in pkgsrc.  Reports
                    586: confirming or denying this notion should be sent to port-xen@.
                    587: 
1.53      gdt       588: The dom0 can give other domains access to selected PCI
1.52      gdt       589: devices. This can allow, for example, a non-privileged domain to have
                    590: access to a physical network interface or disk controller.  However,
                    591: keep in mind that giving a domain access to a PCI device most likely
                    592: will give the domain read/write access to the whole physical memory,
                    593: as PCs don't have an IOMMU to restrict memory access to DMA-capable
1.53      gdt       594: device.  Also, it's not possible to export ISA devices to non-dom0
1.52      gdt       595: domains, which means that the primary VGA adapter can't be exported.
                    596: A guest domain trying to access the VGA registers will panic.
                    597: 
1.53      gdt       598: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
1.52      gdt       599: not been ported to later versions at this time.
                    600: 
                    601: For a PCI device to be exported to a domU, is has to be attached to
                    602: the "pciback" driver in dom0.  Devices passed to the dom0 via the
                    603: pciback.hide boot parameter will attach to "pciback" instead of the
                    604: usual driver.  The list of devices is specified as "(bus:dev.func)",
1.5       mspo      605: where bus and dev are 2-digit hexadecimal numbers, and func a
                    606: single-digit number:
1.1       mspo      607: 
1.52      gdt       608:         pciback.hide=(00:0a.0)(00:06.0)
1.1       mspo      609: 
1.52      gdt       610: pciback devices should show up in the dom0's boot messages, and the
1.5       mspo      611: devices should be listed in the `/kern/xen/pci` directory.
1.1       mspo      612: 
1.52      gdt       613: PCI devices to be exported to a domU are listed in the "pci" array of
                    614: the domU's config file, with the format "0000:bus:dev.func".
1.1       mspo      615: 
1.52      gdt       616:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
1.1       mspo      617: 
1.52      gdt       618: In the domU an "xpci" device will show up, to which one or more pci
1.111     wiz       619: buses will attach.  Then the PCI drivers will attach to PCI buses as
1.52      gdt       620: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
                    621: or any PCI drivers built in by default; you have to build your own
                    622: kernel to use PCI devices in a domU.  Here's a kernel config example;
                    623: note that only the "xpci" lines are unusual.
                    624: 
                    625:         include         "arch/i386/conf/XEN3_DOMU"
                    626: 
1.111     wiz       627:         # Add support for PCI buses to the XEN3_DOMU kernel
1.52      gdt       628:         xpci* at xenbus ?
                    629:         pci* at xpci ?
                    630: 
                    631:         # PCI USB controllers
                    632:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
                    633: 
                    634:         # USB bus support
                    635:         usb*    at uhci?
                    636: 
                    637:         # USB Hubs
                    638:         uhub*   at usb?
                    639:         uhub*   at uhub? port ? configuration ? interface ?
                    640: 
                    641:         # USB Mass Storage
                    642:         umass*  at uhub? port ? configuration ? interface ?
                    643:         wd*     at umass?
                    644:         # SCSI controllers
                    645:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
                    646: 
                    647:         # SCSI bus support (for both ahc and umass)
                    648:         scsibus* at scsi?
                    649: 
                    650:         # SCSI devices
                    651:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
                    652:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
1.1       mspo      653: 
                    654: 
1.173     gdt       655: # Specific Issues
                    656: 
                    657: ## domU
                    658: 
1.176     gdt       659: [NetBSD 5 is known to panic.](http://mail-index.netbsd.org/port-xen/2018/04/17/msg009181.html)
1.175     gdt       660: (However, NetBSD 5 systems should be updated to a supported version.)
1.173     gdt       661: 
                    662: # NetBSD as a domU in a VPS
1.28      gdt       663: 
                    664: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
                    665: hardware.  This section explains how to deal with Xen in a domU as a
                    666: virtual private server where you do not control or have access to the
1.70      gdt       667: dom0.  This is not intended to be an exhaustive list of VPS providers;
                    668: only a few are mentioned that specifically support NetBSD.
1.28      gdt       669: 
1.52      gdt       670: VPS operators provide varying degrees of access and mechanisms for
                    671: configuration.  The big issue is usually how one controls which kernel
1.111     wiz       672: is booted, because the kernel is nominally in the dom0 file system (to
                    673: which VPS users do not normally have access).  A second issue is how
1.70      gdt       674: to install NetBSD.
1.52      gdt       675: A VPS user may want to compile a kernel for security updates, to run
                    676: npf, run IPsec, or any other reason why someone would want to change
                    677: their kernel.
                    678: 
1.111     wiz       679: One approach is to have an administrative interface to upload a kernel,
1.68      gdt       680: or to select from a prepopulated list.  Other approaches are pygrub
1.59      gdt       681: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
1.111     wiz       682: kernel from the domU file system.  This is closer to a regular physical
1.59      gdt       683: computer, where someone who controls a machine can replace the kernel.
1.52      gdt       684: 
1.74      gdt       685: A second issue is multiple CPUs.  With NetBSD 6, domUs support
                    686: multiple vcpus, and it is typical for VPS providers to enable multiple
                    687: CPUs for NetBSD domUs.
                    688: 
1.180     gdt       689: ## Complexities due to Xen changes
                    690: 
                    691: Xen has many security advisories and people running Xen systems make
                    692: different choices.
                    693: 
                    694: ### stub domains
                    695: 
                    696: Some (Linux only?) dom0 systems use something called "stub domains" to
                    697: isolate qemu from the dom0 system, as a security and reliabilty
                    698: mechanism when running HVM domUs.  Somehow, NetBSD's GENERIC kernel
                    699: ends up using PIO for disks rather than DMA.  Of course, all of this
                    700: is emulated, but emulated PIO is unusably slow.  This problem is not
                    701: currently understood.
                    702: 
                    703: ### Grant tables
                    704: 
                    705: There are multiple versions of using grant tables, and some security
                    706: advisories have suggested disabling some versions.  Some versions of
                    707: NetBSD apparently only use specific versions and this can lead to
                    708: "NetBSD current doesn't run on hosting provider X" situations.
                    709: 
                    710: \todo Explain better.
                    711: 
1.181     gdt       712: ## Boot methods
                    713: 
                    714: ### pvgrub
1.59      gdt       715: 
                    716: pvgrub is a version of grub that uses PV operations instead of BIOS
                    717: calls.  It is booted from the dom0 as the domU kernel, and then reads
1.111     wiz       718: /grub/menu.lst and loads a kernel from the domU file system.
1.59      gdt       719: 
1.70      gdt       720: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
1.71      gdt       721: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
                    722: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
1.70      gdt       723: page](http://www.panix.com/v-colo/grub.html), which describes only
1.74      gdt       724: Linux but should be updated to cover NetBSD :-).
1.70      gdt       725: 
                    726: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
                    727: their own kernel.  See then [prgmr.com NetBSD
1.74      gdt       728: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
                    729: (which is in need of updating).
1.59      gdt       730: 
1.70      gdt       731: It appears that [grub's FFS
                    732: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
                    733: does not support all aspects of modern FFS, but there are also reports
1.72      gdt       734: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
1.70      gdt       735: partition for the kernel with the intent that grub can understand it,
                    736: which leads to /netbsd not being the actual kernel.  One must remember
1.111     wiz       737: to update the special boot partition.
1.59      gdt       738: 
1.181     gdt       739: ### pygrub
1.168     wiki      740: 
                    741: pygrub runs in the dom0 and looks into the domU file system.  This
                    742: implies that the domU must have a kernel in a file system in a format
                    743: known to pygrub.
                    744: 
                    745: pygrub doesn't seem to work to load Linux images under NetBSD dom0,
                    746: and is inherently less secure than pvgrub due to running inside dom0. For both these
                    747: reasons, pygrub should not be used, and is only still present so that
                    748: historical DomU images using it still work.
                    749: 
                    750: As of 2014, pygrub seems to be of mostly historical
                    751: interest. New DomUs should use pvgrub.
                    752: 
1.181     gdt       753: ## Specific Providers
                    754: 
                    755: ### Amazon
1.59      gdt       756: 
1.143     wiki      757: See the [Amazon EC2 page](/amazon_ec2/).

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