File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.143: download - view: text, annotated - select for diffs
Sun Jun 17 21:11:00 2018 UTC (17 months, 2 weeks ago) by wiki
Branches: MAIN
CVS tags: HEAD
web commit by jdolecek

    1: Introduction
    2: ============
    3: 
    4: [![[Xen
    5: screenshot]](https://www.netbsd.org/gallery/in-Action/hubertf-xens.png)](https://www.netbsd.org/gallery/in-Action/hubertf-xen.png)
    6: 
    7: Xen is a hypervisor (or virtual machine monitor) for x86 hardware
    8: (i686-class or higher), which supports running multiple guest
    9: operating systems on a single physical machine.  Xen is a Type 1 or
   10: bare-metal hypervisor; one uses the Xen kernel to control the CPU,
   11: memory and console, a dom0 operating system which mediates access to
   12: other hardware (e.g., disks, network, USB), and one or more domU
   13: operating systems which operate in an unprivileged virtualized
   14: environment.  IO requests from the domU systems are forwarded by the
   15: hypervisor (Xen) to the dom0 to be fulfilled.
   16: 
   17: Xen supports two styles of guests.  The original is Para-Virtualized
   18: (PV) which means that the guest OS does not attempt to access hardware
   19: directly, but instead makes hypercalls to the hypervisor.  This is
   20: analogous to a user-space program making system calls.  (The dom0
   21: operating system uses PV calls for some functions, such as updating
   22: memory mapping page tables, but has direct hardware access for disk
   23: and network.)   PV guests must be specifically coded for Xen.
   24: 
   25: The more recent style is HVM, which means that the guest does not have
   26: code for Xen and need not be aware that it is running under Xen.
   27: Attempts to access hardware registers are trapped and emulated.  This
   28: style is less efficient but can run unmodified guests.
   29: 
   30: Generally any machine that runs NetBSD/amd64 will work with Xen and PV
   31: guests.  In theory i386 computers (without x86_64/amd64 support) can
   32: be used for Xen <= 4.2, but we have no recent reports of this working
   33: (this is a hint).  For HVM guests, hardware support is needed, but it
   34: is common on recent machines.  For Intel CPUs, one needs the VT-x
   35: extension, shown in "cpuctl identify 0" as VMX.  For AMD CPUs, one
   36: needs the AMD-V extensions, shown in "cpuctl identify 0" as SVM.
   37: There are further features for IOMMU virtualization, Intel's VT-d and
   38: AMD's AMD-Vi.  TODO: Explain whether Xen on NetBSD makes use of these
   39: features.  TODO: Review by someone who really understands this.
   40: 
   41: Note that a FreeBSD dom0 requires VT-x and VT-d (or equivalent); this
   42: is because the FreeBSD dom0 does not run in PV mode.
   43: 
   44: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
   45: The dom0 can start one or more domUs.  (Booting is explained in detail
   46: in the dom0 section.)
   47: 
   48: NetBSD supports Xen in that it can serve as dom0, be used as a domU,
   49: and that Xen kernels and tools are available in pkgsrc.  This HOWTO
   50: attempts to address both the case of running a NetBSD dom0 on hardware
   51: and running domUs under it (NetBSD and other), and also running NetBSD
   52: as a domU in a VPS.
   53: 
   54: Xen 3.1 in pkgsrc used to support "PCI passthrough", which means that
   55: specific PCI devices can be made available to a specific domU instead
   56: of the dom0.  This can be useful to let a domU run X11, or access some
   57: network interface or other peripheral.
   58: 
   59: NetBSD 6 and earlier supported Xen 2; support was removed from NetBSD
   60: 7.  Xen 2 has been removed from pkgsrc.
   61: 
   62: Prerequisites
   63: -------------
   64: 
   65: Installing NetBSD/Xen is not extremely difficult, but it is more
   66: complex than a normal installation of NetBSD.
   67: In general, this HOWTO is occasionally overly restrictive about how
   68: things must be done, guiding the reader to stay on the established
   69: path when there are no known good reasons to stray.
   70: 
   71: This HOWTO presumes a basic familiarity with the Xen system
   72: architecture, with installing NetBSD on i386/amd64 hardware, and with
   73: installing software from pkgsrc.  See also the [Xen
   74: website](http://www.xenproject.org/).
   75: 
   76: Versions of Xen and NetBSD
   77: ==========================
   78: 
   79: Most of the installation concepts and instructions are independent
   80: of Xen version and NetBSD version.  This section gives advice on
   81: which version to choose.  Versions not in pkgsrc and older unsupported
   82: versions of NetBSD are intentionally ignored.
   83: 
   84: The term "amd64" is used to refer to both the NetBSD port and to the
   85: hardware architecture on which it runs.  (Such hardware is made by
   86: both Intel and AMD, and in 2016 a normal PC has this CPU
   87: architecture.)
   88: 
   89: Xen
   90: ---
   91: 
   92: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
   93: xenkernel and xentools.  We will refer only to the kernel versions,
   94: but note that both packages must be installed together and must have
   95: matching versions.
   96: 
   97: xenkernel3 provides Xen 3.1.  It is no longer maintained by Xen, and
   98: the last applied security patch was in 2011. Thus, it should not be
   99: used.  It supports PCI passthrough, which is why people use it anyway.
  100: Xen 3.1 runs on i386 (both non-PAE and PAE) and amd64 hardware.
  101: 
  102: xenkernel33 provides Xen 3.3.  It is no longer maintained by Xen, and
  103: the last applied security patch was in 2012.  Thus, it should not be
  104: used.  Xen 3.3 runs on i386 PAE and amd64 hardware.  There are no good
  105: reasons to run this version.
  106: 
  107: xenkernel41 provides Xen 4.1.  It is no longer maintained by Xen, but
  108: as of 2016-12 received backported security patches.  Xen 4.1 runs on
  109: i386 PAE and amd64 hardware.  There are no good reasons to run this
  110: version.
  111: 
  112: Note that 3.1, 3.3 and 4.1 have been removed from pkgsrc-current, but
  113: are in 2016Q4.  They will be removed from this HOWTO sometime after
  114: 2017Q1.
  115: 
  116: xenkernel42 provides Xen 4.2.  It is no longer maintained by Xen, but
  117: as of 2016-12 received backported security patches.  Xen 4.2 runs on
  118: i386 PAE and amd64 hardware.  The only reason to run this is if you
  119: need to use xm instead of xl, or if you need to run on hardware that
  120: supports i386 but not amd64.  (This might also be useful if you need
  121: an i386 dom0, if it turns out that an amd64 Xen kernel and an i386
  122: dom0 is problematic.)
  123: 
  124: xenkernel45 provides Xen 4.5.  Security advisories released in 2018-05
  125: did not include support for 4.5.  Xen 4.5 and newer runs on amd64
  126: hardware only.  While slightly old, 4.5 has been tested and run by
  127: others, so it is a very conservative choice.
  128: 
  129: xenkernel46 provides Xen 4.6, and was added to pkgsrc as of 2016-05.
  130: As of 2018-05, security patches were released by Xen and are expected
  131: to be applied to pkgsrc.  Xen 4.6 runs on amd64 hardware only.  (If
  132: using Ubuntu guests, be sure to have the xentools46 from December,
  133: 2016).  4.6 is perhaps an old choice, or perhaps the standard
  134: approach.
  135: 
  136: Xen 4.7 was released in 2016-06 and is not in pkgsrc.
  137: 
  138: xenkernel48 provides Xen 4.8, and was added to pkgsrc in 2017-03.  As
  139: of 2018-05, security patches were released by Xen and are expected to
  140: be applied to pkgsrc.  4.8 is perhaps the standard choice, or perhaps
  141: slightly new.
  142: 
  143: Xen 4.9 and 4.10 are not in pkgsrc.
  144: 
  145: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
  146: 
  147: Note that NetBSD support is called XEN3.  It works with Xen 3 and Xen
  148: 4 because the hypercall interface has been stable.
  149: 
  150: Xen command program
  151: -------------------
  152: 
  153: Early Xen used a program called xm to manipulate the system from the
  154: dom0.  Starting in 4.1, a replacement program with similar behavior
  155: called xl is provided, but it does not work well in 4.1.  In 4.2, both
  156: xm and xl work fine.  4.4 is the last version that has xm.
  157: 
  158: You must make a global choice to use xm or xl, because it affects not
  159: only which command you use, but the command used by rc.d scripts
  160: (specifically xendomains) and which daemons should be run.  The
  161: xentools packages provide xm for 3.1, 3.3 and 4.1 and xl for 4.2 and up.
  162: 
  163: In 4.2, you can choose to use xm by simply changing the ctl_command
  164: variable and setting xend=YES in rc.conf.
  165: 
  166: With xl, virtual devices are configured in parallel, which can cause
  167: problems if they are written assuming serial operation (e.g., updating
  168: firewall rules without explicit locking).  There is now locking for
  169: the provided scripts, which works for normal casses (e.g, file-backed
  170: xbd, where a vnd must be allocated).  But, as of 201612, it has not
  171: been adequately tested for a complex custom setup with a large number
  172: of interfaces.
  173: 
  174: NetBSD
  175: ------
  176: 
  177: The netbsd-7, netbsd-8, and -current branches are all reasonable
  178: choices, with more or less the same considerations for non-Xen use.
  179: Therefore, netbsd-7 is recommended as the stable version of the most
  180: recent release for production use.  (Note that netbsd-7 (and therefore
  181: 8/current) have a important scheduler fix (in November of 2015)
  182: affecting contention between dom0 and domUs; see
  183: https://releng.netbsd.org/cgi-bin/req-7.cgi?show=1040 for a
  184: description.)  For production, netbsd-7 is appropriate.  For learning,
  185: netbsd-8 is appropriate.  For developing Xen, netbsd-current may be
  186: appropriate.
  187: 
  188: As of NetBSD 6, a NetBSD domU will support multiple vcpus.  There is
  189: no SMP support for NetBSD as dom0.  (The dom0 itself doesn't really
  190: need SMP for dom0 functions; the lack of support is really a problem
  191: when using a dom0 as a normal computer.)
  192: 
  193: Architecture
  194: ------------
  195: 
  196: Xen itself can run on i386 (Xen < 4.2) or amd64 hardware (all Xen
  197: versions).  (Practically, almost any computer where one would want to
  198: run Xen today supports amd64.)
  199: 
  200: Xen, the dom0 system, and each domU system can be either i386 or
  201: amd64.  When building a xenkernel package, one obtains an i386 Xen
  202: kernel on an i386 host, and an amd64 Xen kernel on an amd64 host.  If
  203: the Xen kernel is i386, then the dom0 kernel and all domU kernels must
  204: be i386.  With an amd64 Xen kernel, an amd64 dom0 kernel is known to
  205: work, and an i386 dom0 kernel should in theory work.  An amd64
  206: Xen/dom0 is known to support both i386 and amd64 domUs.
  207: 
  208: i386 dom0 and domU kernels must be PAE (except for an i386 Xen 3.1
  209: kernel, where one can use non-PAE for dom0 and all domUs); PAE kernels
  210: are included in the NetBSD default build.  (Note that emacs (at least)
  211: fails if run on i386 with PAE when built without, and vice versa,
  212: presumably due to bugs in the undump code.)
  213: 
  214: Because of the above, the standard approach is to use an amd64 Xen
  215: kernel and NetBSD/amd64 for the dom0.  For domUs, NetBSD/i386 (with
  216: the PAE kernel) and NetBSD/amd64 are in widespread use, and there is
  217: little to no Xen-specific reason to prefer one over the other.
  218: 
  219: Note that to use an i386 dom0 with Xen 4.5 or higher, one must build
  220: (or obtain from pre-built packages) an amd64 Xen kernel and install
  221: that on the system.  (One must also use a PAE i386 kernel, but this is
  222: also required with an i386 Xen kernel.).  Almost no one in the
  223: NetBSD/Xen community does this, and the standard, well-tested,
  224: approach is to use an amd64 dom0.
  225: 
  226: A [posting on
  227: xen-devel](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html)
  228: explained that PV system call overhead was higher on amd64, and thus
  229: there is some notion that i386 guests are faster.  It goes on to
  230: caution that the total situation is complex and not entirely
  231: understood. On top of that caution, the post is about Linux, not
  232: NetBSD.  TODO: Include link to benchmarks, if someone posts them.
  233: 
  234: Stability
  235: ---------
  236: 
  237: Mostly, NetBSD as a dom0 or domU is quite stable.
  238: However, there are some open PRs indicating problems.
  239: 
  240:  - [PR 48125](http://gnats.netbsd.org/48125)
  241: 
  242: Note also that there are issues with sparse vnd(4) instances, but
  243: these are not about Xen -- they just are noticed with sparse vnd(4)
  244: instances in support of virtual disks in a dom0.
  245: 
  246: Recommendation
  247: --------------
  248: 
  249: Therefore, this HOWTO recommends running xenkernel46, xl, the NetBSD 7
  250: stable branch, and therefore to use an amd64 kernel as the dom0.
  251: Either the i386PAE or amd64 version of NetBSD may be used as domUs.
  252: 
  253: A tentative replacement recommendation is xenkernel48, xl, and NetBSD
  254: 8.
  255: 
  256: Because bugs are fixed quite often, and because of Xen security
  257: advisories, it is good to stay up to date with NetBSD (tracking a
  258: stable branch), with the Xen kernel (tracking a Xen version via
  259: pkgsrc), and with the Xen tools.  Specifically, NetBSD (-7 and
  260: -current) got an important fix affecting dom0/domU timesharing in
  261: November, 2015, and xentools46 got a fix to enable Ubuntu guests to
  262: boot in December, 2016.
  263: 
  264: Status
  265: ------
  266: 
  267: Ideally, all versions of Xen in pkgsrc would build on all supported
  268: versions of NetBSD/amd64, to the point where this section would be
  269: silly.  However, that has not always been the case.  Besides aging
  270: code and aging compilers, qemu (included in xentools for HVM support)
  271: is difficult to build.  Note that there is intentionally no data for
  272: 4.5+ up for i386, and often omits xentools info if the corresponding
  273: kernel fails.
  274: 
  275: The following table gives status, with the date last checked
  276: (generally on the most recent quarterly branch).  The first code is
  277: "builds" if it builds ok, and "FAIL" for a failure to build.  The
  278: second code/date only appears for xenkernel* and is "works" if it runs
  279: ok as a dom0 and can support a domU, and "FAIL" if it won't boot or
  280: run a domU.
  281: 
  282: 	xenkernel3 netbsd-6 i386 FAIL 201612
  283: 	xenkernel33 netbsd-6 i386 FAIL 201612
  284: 	xenkernel41 netbsd-6 i386 builds 201612
  285: 	xenkernel42 netbsd-6 i386 builds 201612
  286: 	xentools3 netbsd-6 i386 FAIL 201612
  287: 	xentools33 netbsd-6 i386 FAIL 201612
  288: 	xentools41 netbsd-6 i386 builds 201612
  289: 	xentools42 netbsd-6 i386 FAIL 201612
  290: 
  291: 	xenkernel3 netbsd-7 i386 FAIL 201412
  292: 	xenkernel33 netbsd-7 i386 FAIL 201412
  293: 	xenkernel41 netbsd-7 i386 builds 201412
  294: 	xenkernel42 netbsd-7 i386 builds 201412
  295: 	xentools41 netbsd-7 i386 builds 201412
  296: 	xentools42 netbsd-7 i386 ??FAIL 201412
  297: 
  298: 	xenkernel3 netbsd-6 amd64 FAIL 201612
  299: 	xenkernel33 netbsd-6 amd64 FAIL 201612
  300: 	xenkernel41 netbsd-6 amd64 builds 201612 works 201612
  301: 	xenkernel42 netbsd-6 amd64 builds 201612 works 201612
  302: 	xenkernel45 netbsd-6 amd64 builds 201612
  303: 	xenkernel46 netbsd-6 amd64 builds 201612
  304: 	xentools41 netbsd-6 amd64 builds 201612
  305: 	xentools42 netbsd-6 amd64 builds 201612
  306: 	xentools45 netbsd-6 amd64 builds 201612
  307: 	xentools46 netbsd-6 amd64 FAIL 201612
  308: 
  309: 	xenkernel3 netbsd-7 amd64 builds 201612
  310: 	xenkernel33 netbsd-7 amd64 builds 201612
  311: 	xenkernel41 netbsd-7 amd64 builds 201612
  312: 	xenkernel42 netbsd-7 amd64 builds 201612
  313: 	xenkernel45 netbsd-7 amd64 builds 201612
  314: 	xenkernel46 netbsd-7 amd64 builds 201612
  315: 	xentools3 netbsd-7 amd64 builds 201612
  316: 	xentools3-hvm netbsd-7 amd64 builds 201612
  317: 	xentools33 netbsd-7 amd64 FAIL 201612
  318: 	xentools41 netbsd-7 amd64 builds 201612
  319: 	xentools42 netbsd-7 amd64 builds 201612
  320: 	xentools45 netbsd-7 amd64 builds 201612
  321: 	xentools46 netbsd-7 amd64 builds 201612
  322: 
  323: NetBSD as a dom0
  324: ================
  325: 
  326: NetBSD can be used as a dom0 and works very well.  The following
  327: sections address installation, updating NetBSD, and updating Xen.
  328: Note that it doesn't make sense to talk about installing a dom0 OS
  329: without also installing Xen itself.  We first address installing
  330: NetBSD, which is not yet a dom0, and then adding Xen, pivoting the
  331: NetBSD install to a dom0 install by just changing the kernel and boot
  332: configuration.
  333: 
  334: For experimenting with Xen, a machine with as little as 1G of RAM and
  335: 100G of disk can work.  For running many domUs in productions, far
  336: more will be needed; e.g. 4-8G and 1T of disk is reasonable for a
  337: half-dozen domUs of 512M and 32G each.  Basically, the RAM and disk
  338: have to be bigger than the sum of the RAM/disk needs of the dom0 and
  339: all the domUs.
  340: 
  341: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
  342: 512M it worked reliably.  This does not make sense, but if you see
  343: "not ELF" after Xen boots, try increasing dom0 RAM.
  344: 
  345: Styles of dom0 operation
  346: ------------------------
  347: 
  348: There are two basic ways to use Xen.  The traditional method is for
  349: the dom0 to do absolutely nothing other than providing support to some
  350: number of domUs.  Such a system was probably installed for the sole
  351: purpose of hosting domUs, and sits in a server room on a UPS.
  352: 
  353: The other way is to put Xen under a normal-usage computer, so that the
  354: dom0 is what the computer would have been without Xen, perhaps a
  355: desktop or laptop.  Then, one can run domUs at will.  Purists will
  356: deride this as less secure than the previous approach, and for a
  357: computer whose purpose is to run domUs, they are right.  But Xen and a
  358: dom0 (without domUs) is not meaningfully less secure than the same
  359: things running without Xen.  One can boot Xen or boot regular NetBSD
  360: alternately with little problems, simply refraining from starting the
  361: Xen daemons when not running Xen.
  362: 
  363: Note that NetBSD as dom0 does not support multiple CPUs.  This will
  364: limit the performance of the Xen/dom0 workstation approach.  In theory
  365: the only issue is that the "backend drivers" are not yet MPSAFE:
  366:   https://mail-index.netbsd.org/netbsd-users/2014/08/29/msg015195.html
  367: 
  368: Installation of NetBSD
  369: ----------------------
  370: 
  371: First,
  372: [install NetBSD/amd64](/guide/inst/)
  373: just as you would if you were not using Xen.
  374: However, the partitioning approach is very important.
  375: 
  376: If you want to use RAIDframe for the dom0, there are no special issues
  377: for Xen.  Typically one provides RAID storage for the dom0, and the
  378: domU systems are unaware of RAID.  The 2nd-stage loader bootxx_* skips
  379: over a RAID1 header to find /boot from a file system within a RAID
  380: partition; this is no different when booting Xen.
  381: 
  382: There are 4 styles of providing backing storage for the virtual disks
  383: used by domUs: raw partitions, LVM, file-backed vnd(4), and SAN.
  384: 
  385: With raw partitions, one has a disklabel (or gpt) partition sized for
  386: each virtual disk to be used by the domU.  (If you are able to predict
  387: how domU usage will evolve, please add an explanation to the HOWTO.
  388: Seriously, needs tend to change over time.)
  389: 
  390: One can use [lvm(8)](/guide/lvm/) to create logical devices to use
  391: for domU disks.  This is almost as efficient as raw disk partitions
  392: and more flexible.  Hence raw disk partitions should typically not
  393: be used.
  394: 
  395: One can use files in the dom0 file system, typically created by dd'ing
  396: /dev/zero to create a specific size.  This is somewhat less efficient,
  397: but very convenient, as one can cp the files for backup, or move them
  398: between dom0 hosts.
  399: 
  400: Finally, in theory one can place the files backing the domU disks in a
  401: SAN.  (This is an invitation for someone who has done this to add a
  402: HOWTO page.)
  403: 
  404: Installation of Xen
  405: -------------------
  406: 
  407: In the dom0, install sysutils/xenkernel42 and sysutils/xentools42 from
  408: pkgsrc (or another matching pair).  See [the pkgsrc
  409: documentation](https://www.NetBSD.org/docs/pkgsrc/) for help with
  410: pkgsrc.  Ensure that your packages are recent; the HOWTO does not
  411: contemplate old builds.
  412: 
  413: 
  414: For Xen 3.1, support for HVM guests is in sysutils/xentool3-hvm.  More
  415: recent versions have HVM support integrated in the main xentools
  416: package.  It is entirely reasonable to run only PV guests.
  417: 
  418: Next you need to install the selected Xen kernel itself, which is
  419: installed by pkgsrc as "/usr/pkg/xen*-kernel/xen.gz".  Copy it to /.
  420: For debugging, one may copy xen-debug.gz; this is conceptually similar
  421: to DIAGNOSTIC and DEBUG in NetBSD.  xen-debug.gz is basically only
  422: useful with a serial console.  Then, place a NetBSD XEN3_DOM0 kernel
  423: in /, copied from releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
  424: of a NetBSD build.  If using i386, use
  425: releasedir/i386/binary/kernel/netbsd-XEN3PAE_DOM0.gz.  (If using Xen
  426: 3.1 and i386, you may use XEN3_DOM0 with the non-PAE Xen.  But you
  427: should not use Xen 3.1.)  Both xen and the NetBSD kernel may be (and
  428: typically are) left compressed.
  429: 
  430: In a dom0, kernfs is mandatory for xend to communicate with the
  431: kernel, so ensure that /kern is in fstab.  (A standard NetBSD install
  432: should already mount /kern.)
  433: 
  434: Because you already installed NetBSD, you have a working boot setup
  435: with an MBR bootblock, either bootxx_ffsv1 or bootxx_ffsv2 at the
  436: beginning of your root file system, have /boot, and likely also
  437: /boot.cfg.  (If not, fix before continuing!)
  438: 
  439: Add a line to to /boot.cfg to boot Xen.  See boot.cfg(5) for an
  440: example.  The basic line is
  441: 
  442:         menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
  443: 
  444: which specifies that the dom0 should have 512M, leaving the rest to be
  445: allocated for domUs.  To use a serial console, use
  446: 
  447:         menu=Xen:load /netbsd-XEN3_DOM0.gz;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
  448: 
  449: which will use the first serial port for Xen (which counts starting
  450: from 1, unlike NetBSD which counts starting from 0), forcing
  451: speed/parity.  Because the NetBSD command line lacks a
  452: "console=pc" argument, it will use the default "xencons" console device,
  453: which directs the console I/O through Xen to the same console device Xen
  454: itself uses (in this case, the serial port).
  455: 
  456: In an attempt to add performance, one can also add
  457: 
  458:         dom0_max_vcpus=1 dom0_vcpus_pin
  459: 
  460: to force only one vcpu to be provided (since NetBSD dom0 can't use
  461: more) and to pin that vcpu to a physical CPU.  TODO: benchmark this.
  462: 
  463: Xen has [many boot
  464: options](http://xenbits.xenproject.org/docs/4.5-testing/misc/xen-command-line.html),
  465: and other than dom0 memory and max_vcpus, they are generally not
  466: necessary.
  467: 
  468: As with non-Xen systems, you should have a line to boot /netbsd (a
  469: kernel that works without Xen).  Consider a line to boot /netbsd.ok (a
  470: fallback version of the non-Xen kernel, updated manually when you are
  471: sure /netbsd is ok).  Consider also a line to boot fallback versions
  472: of Xen and the dom0 kernel, but note that non-Xen NetBSD can be used
  473: to resolve Xen booting issues.
  474: 
  475: Probably you want a default=N line to choose Xen in the absence of
  476: intervention.
  477: 
  478: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  479: than GENERIC without Xen.
  480: 
  481: Using grub (historic)
  482: ---------------------
  483: 
  484: Before NetBSD's native bootloader could support Xen, the use of
  485: grub was recommended.  If necessary, see the
  486: [old grub information](/ports/xen/howto-grub).
  487: 
  488: The [HowTo on Installing into
  489: RAID-1](https://mail-index.NetBSD.org/port-xen/2006/03/01/0010.html)
  490: explains how to set up booting a dom0 with Xen using grub with
  491: NetBSD's RAIDframe.  (This is obsolete with the use of NetBSD's native
  492: boot.  Now, just create a system with RAID-1, and alter /boot.cfg as
  493: described above.)
  494: 
  495: Configuring Xen
  496: ---------------
  497: 
  498: Xen logs will be in /var/log/xen.
  499: 
  500: Now, you have a system that will boot Xen and the dom0 kernel, but not
  501: do anything else special.  Make sure that you have rebooted into Xen.
  502: There will be no domUs, and none can be started because you still have
  503: to configure the dom0 daemons.
  504: 
  505: The daemons which should be run vary with Xen version and with whether
  506: one is using xm or xl.  The Xen 3.1, 3.3 and 4.1 packages use xm.  Xen
  507: 4.2 and up packages use xl.  To use xm with 4.2, edit xendomains to
  508: use xm instead.
  509: 
  510: For 3.1 and 3.3, you should enable xend and xenbackendd:
  511: 
  512:         xend=YES
  513:         xenbackendd=YES
  514: 
  515: For 4.1 and up, you should enable xencommons.  Not enabling xencommons
  516: will result in a hang; it is necessary to hit ^C on the console to let
  517: the machine finish booting.  If you are using xm (default in 4.1, or
  518: if you changed xendomains in 4.2), you should also enable xend:
  519: 
  520:         xend=YES # only if using xm, and only installed <= 4.2
  521:         xencommons=YES
  522: 
  523: TODO: Recommend for/against xen-watchdog.
  524: 
  525: After you have configured the daemons and either started them (in the
  526: order given) or rebooted, use xm or xl to inspect Xen's boot messages,
  527: available resources, and running domains.  An example with xl follows:
  528: 
  529:         # xl dmesg
  530: 	[xen's boot info]
  531:         # xl info
  532: 	[available memory, etc.]
  533:         # xl list
  534:         Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  535:         Domain-0           0       64    0  r----     58.1
  536: 
  537: ### Issues with xencommons
  538: 
  539: xencommons starts xenstored, which stores data on behalf of dom0 and
  540: domUs.  It does not currently work to stop and start xenstored.
  541: Certainly all domUs should be shutdown first, following the sort order
  542: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
  543: and is not notified when xenstored exits, leading to not recreating
  544: the state when the new xenstored starts.  Until there's a mechanism to
  545: make this work, one should not expect to be able to restart xenstored
  546: (and thus xencommons).  There is currently no reason to expect that
  547: this will get fixed any time soon.
  548: 
  549: ### No-longer needed advice about devices
  550: 
  551: The installation of NetBSD should already have created devices for xen
  552: (xencons, xenevt, xsd_kva), but if they are not present, create them:
  553: 
  554:         cd /dev && sh MAKEDEV xen
  555: 
  556: anita (for testing NetBSD)
  557: --------------------------
  558: 
  559: With the setup so far (assuming 4.2/xl), one should be able to run
  560: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
  561: root, because anita must create a domU):
  562: 
  563:         anita --vmm=xl test file:///usr/obj/i386/
  564: 
  565: Alternatively, one can use --vmm=xm to use xm-based domU creation
  566: instead (and must, on Xen <= 4.1).   TODO: confirm that anita xl really works.
  567:     
  568: Xen-specific NetBSD issues
  569: --------------------------
  570: 
  571: There are (at least) two additional things different about NetBSD as a
  572: dom0 kernel compared to hardware.
  573: 
  574: One is that the module ABI is different because some of the #defines
  575: change, so one must build modules for Xen.  As of netbsd-7, the build
  576: system does this automatically.  TODO: check this.  (Before building
  577: Xen modules was added, it was awkward to use modules to the point
  578: where it was considered that it did not work.)
  579: 
  580: The other difference is that XEN3_DOM0 does not have exactly the same
  581: options as GENERIC.  While it is debatable whether or not this is a
  582: bug, users should be aware of this and can simply add missing config
  583: items if desired.
  584: 
  585: Updating NetBSD in a dom0
  586: -------------------------
  587: 
  588: This is just like updating NetBSD on bare hardware, assuming the new
  589: version supports the version of Xen you are running.  Generally, one
  590: replaces the kernel and reboots, and then overlays userland binaries
  591: and adjusts /etc.
  592: 
  593: Note that one must update both the non-Xen kernel typically used for
  594: rescue purposes and the DOM0 kernel used with Xen.
  595: 
  596: Converting from grub to /boot
  597: -----------------------------
  598: 
  599: These instructions were [TODO: will be] used to convert a system from
  600: grub to /boot.  The system was originally installed in February of
  601: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
  602: over time.  Before these commands, it was running NetBSD 6 i386, Xen
  603: 4.1 and grub, much like the message linked earlier in the grub
  604: section.
  605: 
  606:         # Install MBR bootblocks on both disks. 
  607:         fdisk -i /dev/rwd0d
  608:         fdisk -i /dev/rwd1d
  609:         # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
  610:         installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
  611:         installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
  612:         # Install secondary boot loader
  613:         cp -p /usr/mdec/boot /
  614:         # Create boot.cfg following earlier guidance:
  615:         menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
  616:         menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=512M
  617:         menu=GENERIC:boot
  618:         menu=GENERIC single-user:boot -s
  619:         menu=GENERIC.ok:boot netbsd.ok
  620:         menu=GENERIC.ok single-user:boot netbsd.ok -s
  621:         menu=Drop to boot prompt:prompt
  622:         default=1
  623:         timeout=30
  624: 
  625: TODO: actually do this and fix it if necessary.
  626: 
  627: Upgrading Xen versions
  628: ---------------------
  629: 
  630: Minor version upgrades are trivial.  Just rebuild/replace the
  631: xenkernel version and copy the new xen.gz to / (where /boot.cfg
  632: references it), and reboot.
  633: 
  634: Major version upgrades are conceptually not difficult, but can run
  635: into all the issues found when installing Xen.  Assuming migration
  636: from 4.1 to 4.2, remove the xenkernel41 and xentools41 packages and
  637: install the xenkernel42 and xentools42 packages.  Copy the 4.2 xen.gz
  638: to /.
  639: 
  640: Ensure that the contents of /etc/rc.d/xen* are correct.  Specifically,
  641: they must match the package you just installed and not be left over
  642: from some previous installation.
  643: 
  644: Enable the correct set of daemons; see the configuring section above.
  645: (Upgrading from 3.x to 4.x without doing this will result in a hang.)
  646: 
  647: Ensure that the domU config files are valid for the new version.
  648: Specifically, for 4.x remove autorestart=True, and ensure that disks
  649: are specified with numbers as the second argument, as the examples
  650: above show, and not NetBSD device names.
  651: 
  652: Hardware known to work
  653: ----------------------
  654: 
  655: Arguably, this section is misplaced, and there should be a page of
  656: hardware that runs NetBSD/amd64 well, with the mostly-well-founded
  657: assumption that NetBSD/xen runs fine on any modern hardware that
  658: NetBSD/amd64 runs well on.  Until then, we give motherboard/CPU (and
  659: sometimes RAM) pairs/triples to aid those choosing a motherboard.
  660: Note that Xen systems usually do not run X, so a listing here does not
  661: imply that X works at all.
  662: 
  663:         Supermicro X9SRL-F, Xeon E5-1650 v2, 96 GiB ECC
  664:         Supermicro ??, Atom C2758 (8 core), 32 GiB ECC
  665:         ASUS M5A78L-M/USB3 AM3+ microATX, AMD Piledriver X8 4000MHz, 16 GiB ECC
  666: 
  667: Older hardware:
  668: 
  669:         Intel D915GEV, Pentium4 CPU 3.40GHz, 4GB 533MHz Synchronous DDR2
  670:         INTEL DG33FB, "Intel(R) Core(TM)2 Duo CPU     E6850  @ 3.00GHz"
  671:         INTEL DG33FB, "Intel(R) Core(TM)2 Duo CPU     E8400  @ 3.00GHz"
  672: 
  673: Running Xen under qemu
  674: ----------------------
  675: 
  676: The astute reader will note that this section is somewhat twisted.
  677: However, it can be useful to run Xen under qemu either because the
  678: version of NetBSD as a dom0 does not run on the hardware in use, or to
  679: generate automated test cases involving Xen.
  680: 
  681: In 2015-01, the following combination was reported to mostly work:
  682: 
  683:         host OS: NetBSD/amd64 6.1.4
  684:         qemu: 2.2.0 from pkgsrc
  685:         Xen kernel: xenkernel42-4.2.5nb1 from pkgsrc
  686:         dom0 kernel: NetBSD/amd64 6.1.5
  687:         Xen tools: xentools42-4.2.5 from pkgsrc
  688: 
  689: See [PR 47720](https://gnats.netbsd.org/47720) for a problem with dom0
  690: shutdown.
  691: 
  692: Unprivileged domains (domU)
  693: ===========================
  694: 
  695: This section describes general concepts about domUs.  It does not
  696: address specific domU operating systems or how to install them.  The
  697: config files for domUs are typically in /usr/pkg/etc/xen, and are
  698: typically named so that the file name, domU name and the domU's host
  699: name match.
  700: 
  701: The domU is provided with CPU and memory by Xen, configured by the
  702: dom0.  The domU is provided with disk and network by the dom0,
  703: mediated by Xen, and configured in the dom0.
  704: 
  705: Entropy in domUs can be an issue; physical disks and network are on
  706: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  707: 
  708: Config files
  709: ------------
  710: 
  711: There is no good order to present config files and the concepts
  712: surrounding what is being configured.  We first show an example config
  713: file, and then in the various sections give details.
  714: 
  715: See (at least in xentools41) /usr/pkg/share/examples/xen/xmexample*,
  716: for a large number of well-commented examples, mostly for running
  717: GNU/Linux.
  718: 
  719: The following is an example minimal domain configuration file
  720: "/usr/pkg/etc/xen/foo".  It is (with only a name change) an actual
  721: known working config file on Xen 4.1 (NetBSD 5 amd64 dom0 and NetBSD 5
  722: i386 domU).  The domU serves as a network file server.
  723: 
  724:         # -*- mode: python; -*-
  725: 
  726:         kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  727:         memory = 1024
  728:         vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  729:         disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  730:                  'file:/n0/xen/foo-wd1,0x1,w' ]
  731: 
  732: The domain will have the same name as the file.  The kernel has the
  733: host/domU name in it, so that on the dom0 one can update the various
  734: domUs independently.  The vif line causes an interface to be provided,
  735: with a specific mac address (do not reuse MAC addresses!), in bridge
  736: mode.  Two disks are provided, and they are both writable; the bits
  737: are stored in files and Xen attaches them to a vnd(4) device in the
  738: dom0 on domain creation.  The system treats xbd0 as the boot device
  739: without needing explicit configuration.
  740: 
  741: By default xm looks for domain config files in /usr/pkg/etc/xen.  Note
  742: that "xm create" takes the name of a config file, while other commands
  743: take the name of a domain.  To create the domain, connect to the
  744: console, create the domain while attaching the console, shutdown the
  745: domain, and see if it has finished stopping, do (or xl with Xen >=
  746: 4.2):
  747: 
  748:         xm create foo
  749:         xm console foo
  750:         xm create -c foo
  751:         xm shutdown foo
  752:         xm list
  753: 
  754: Typing ^] will exit the console session.  Shutting down a domain is
  755: equivalent to pushing the power button; a NetBSD domU will receive a
  756: power-press event and do a clean shutdown.  Shutting down the dom0
  757: will trigger controlled shutdowns of all configured domUs.
  758: 
  759: domU kernels
  760: ------------
  761: 
  762: On a physical computer, the BIOS reads sector 0, and a chain of boot
  763: loaders finds and loads a kernel.  Normally this comes from the root
  764: file system.  With Xen domUs, the process is totally different.  The
  765: normal path is for the domU kernel to be a file in the dom0's
  766: file system.  At the request of the dom0, Xen loads that kernel into a
  767: new domU instance and starts execution.  While domU kernels can be
  768: anyplace, reasonable places to store domU kernels on the dom0 are in /
  769: (so they are near the dom0 kernel), in /usr/pkg/etc/xen (near the
  770: config files), or in /u0/xen (where the vdisks are).
  771: 
  772: Note that loading the domU kernel from the dom0 implies that boot
  773: blocks, /boot, /boot.cfg, and so on are all ignored in the domU.
  774: See the VPS section near the end for discussion of alternate ways to
  775: obtain domU kernels.
  776: 
  777: CPU and memory
  778: --------------
  779: 
  780: A domain is provided with some number of vcpus, less than the number
  781: of CPUs seen by the hypervisor.  (For a dom0, this is controlled by
  782: the boot argument "dom0_max_vcpus=1".)  For a domU, it is controlled
  783: from the config file by the "vcpus = N" directive.
  784: 
  785: A domain is provided with memory; this is controlled in the config
  786: file by "memory = N" (in megabytes).  In the straightforward case, the
  787: sum of the the memory allocated to the dom0 and all domUs must be less
  788: than the available memory.
  789: 
  790: Xen also provides a "balloon" driver, which can be used to let domains
  791: use more memory temporarily.  TODO: Explain better, and explain how
  792: well it works with NetBSD.
  793: 
  794: Virtual disks
  795: -------------
  796: 
  797: With the file/vnd style, typically one creates a directory,
  798: e.g. /u0/xen, on a disk large enough to hold virtual disks for all
  799: domUs.  Then, for each domU disk, one writes zeros to a file that then
  800: serves to hold the virtual disk's bits; a suggested name is foo-xbd0
  801: for the first virtual disk for the domU called foo.  Writing zeros to
  802: the file serves two purposes.  One is that preallocating the contents
  803: improves performance.  The other is that vnd on sparse files has
  804: failed to work.  TODO: give working/notworking NetBSD versions for
  805: sparse vnd and gnats reference.  Note that the use of file/vnd for Xen
  806: is not really different than creating a file-backed virtual disk for
  807: some other purpose, except that xentools handles the vnconfig
  808: commands.  To create an empty 4G virtual disk, simply do
  809: 
  810:         dd if=/dev/zero of=foo-xbd0 bs=1m count=4096
  811: 
  812: Do not use qemu-img-xen, because this will create sparse file.  There
  813: have been recent (2015) reports of sparse vnd(4) devices causing
  814: lockups, but there is apparently no PR.
  815: 
  816: With the lvm style, one creates logical devices.  They are then used
  817: similarly to vnds.  TODO: Add an example with lvm.
  818: 
  819: In domU config files, the disks are defined as a sequence of 3-tuples.
  820: The first element is "method:/path/to/disk".  Common methods are
  821: "file:" for file-backed vnd. and "phy:" for something that is already
  822: a (TODO: character or block) device.
  823: 
  824: The second element is an artifact of how virtual disks are passed to
  825: Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  826: are given a device name to associate with the disk, and values like
  827: "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  828: as xbd0, the second as xbd1, and so on.  However, xm/xl demand a
  829: second argument.  The name given is converted to a major/minor by
  830: calling stat(2) on the name in /dev and this is passed to the domU.
  831: In the general case, the dom0 and domU can be different operating
  832: systems, and it is an unwarranted assumption that they have consistent
  833: numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  834: as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  835: for the second works fine and avoids this issue.  For a GNU/Linux
  836: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  837: /dev/hda1.
  838: 
  839: The third element is "w" for writable disks, and "r" for read-only
  840: disks.
  841: 
  842: Note that NetBSD by default creates only vnd[0123].  If you need more
  843: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
  844: dom0.
  845: 
  846: Note that NetBSD by default creates only xbd[0123].  If you need more
  847: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
  848: 
  849: Virtual Networking
  850: ------------------
  851: 
  852: Xen provides virtual Ethernets, each of which connects the dom0 and a
  853: domU.  For each virtual network, there is an interface "xvifN.M" in
  854: the dom0, and in domU index N, a matching interface xennetM (NetBSD
  855: name).  The interfaces behave as if there is an Ethernet with two
  856: adapters connected.  From this primitive, one can construct various
  857: configurations.  We focus on two common and useful cases for which
  858: there are existing scripts: bridging and NAT.
  859: 
  860: With bridging (in the example above), the domU perceives itself to be
  861: on the same network as the dom0.  For server virtualization, this is
  862: usually best.  Bridging is accomplished by creating a bridge(4) device
  863: and adding the dom0's physical interface and the various xvifN.0
  864: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  865: config file.  The bridge must be set up already in the dom0; an
  866: example /etc/ifconfig.bridge0 is:
  867: 
  868:         create
  869:         up
  870:         !brconfig bridge0 add wm0
  871: 
  872: With NAT, the domU perceives itself to be behind a NAT running on the
  873: dom0.  This is often appropriate when running Xen on a workstation.
  874: TODO: NAT appears to be configured by "vif = [ '' ]".
  875: 
  876: The MAC address specified is the one used for the interface in the new
  877: domain.  The interface in dom0 will use this address XOR'd with
  878: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  879: 
  880: Sizing domains
  881: --------------
  882: 
  883: Modern x86 hardware has vast amounts of resources.  However, many
  884: virtual servers can function just fine on far less.  A system with
  885: 512M of RAM and a 4G disk can be a reasonable choice.  Note that it is
  886: far easier to adjust virtual resources than physical ones.  For
  887: memory, it's just a config file edit and a reboot.  For disk, one can
  888: create a new file and vnconfig it (or lvm), and then dump/restore,
  889: just like updating physical disks, but without having to be there and
  890: without those pesky connectors.
  891: 
  892: Starting domains automatically
  893: ------------------------------
  894: 
  895: To start domains foo at bar at boot and shut them down cleanly on dom0
  896: shutdown, in rc.conf add:
  897: 
  898:         xendomains="foo bar"
  899: 
  900: Note that earlier versions of the xentools41 xendomains rc.d script
  901: used xl, when one should use xm with 4.1.
  902: 
  903: Creating specific unprivileged domains (domU)
  904: =============================================
  905: 
  906: Creating domUs is almost entirely independent of operating system.  We
  907: have already presented the basics of config files.  Note that you must
  908: have already completed the dom0 setup so that "xl list" (or "xm list")
  909: works.
  910: 
  911: Creating an unprivileged NetBSD domain (domU)
  912: ---------------------------------------------
  913: 
  914: See the earlier config file, and adjust memory.  Decide on how much
  915: storage you will provide, and prepare it (file or lvm).
  916: 
  917: While the kernel will be obtained from the dom0 file system, the same
  918: file should be present in the domU as /netbsd so that tools like
  919: savecore(8) can work.   (This is helpful but not necessary.)
  920: 
  921: The kernel must be specifically for Xen and for use as a domU.  The
  922: i386 and amd64 provide the following kernels:
  923: 
  924:         i386 XEN3_DOMU
  925:         i386 XEN3PAE_DOMU
  926:         amd64 XEN3_DOMU
  927: 
  928: Unless using Xen 3.1 (and you shouldn't) with i386-mode Xen, you must
  929: use the PAE version of the i386 kernel.
  930: 
  931: This will boot NetBSD, but this is not that useful if the disk is
  932: empty.  One approach is to unpack sets onto the disk outside of xen
  933: (by mounting it, just as you would prepare a physical disk for a
  934: system you can't run the installer on).
  935: 
  936: A second approach is to run an INSTALL kernel, which has a miniroot
  937: and can load sets from the network.  To do this, copy the INSTALL
  938: kernel to / and change the kernel line in the config file to:
  939: 
  940:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  941: 
  942: Then, start the domain as "xl create -c configname".
  943: 
  944: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  945: line should be used in the config file.
  946: 
  947:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  948: 
  949: After booting the domain, the option to install via CDROM may be
  950: selected.  The CDROM device should be changed to `xbd1d`.
  951: 
  952: Once done installing, "halt -p" the new domain (don't reboot or halt,
  953: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  954: config file), switch the config file back to the XEN3_DOMU kernel,
  955: and start the new domain again. Now it should be able to use "root on
  956: xbd0a" and you should have a, functional NetBSD domU.
  957: 
  958: TODO: check if this is still accurate.
  959: When the new domain is booting you'll see some warnings about *wscons*
  960: and the pseudo-terminals. These can be fixed by editing the files
  961: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  962: `/etc/ttys`, except *console*, like this:
  963: 
  964:     console "/usr/libexec/getty Pc"         vt100   on secure
  965:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  966:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  967:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  968:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  969: 
  970: Finally, all screens must be commented out from `/etc/wscons.conf`.
  971: 
  972: It is also desirable to add
  973: 
  974:         powerd=YES
  975: 
  976: in rc.conf. This way, the domain will be properly shut down if
  977: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
  978: 
  979: It is not strictly necessary to have a kernel (as /netbsd) in the domU
  980: file system.  However, various programs (e.g. netstat) will use that
  981: kernel to look up symbols to read from kernel virtual memory.  If
  982: /netbsd is not the running kernel, those lookups will fail.  (This is
  983: not really a Xen-specific issue, but because the domU kernel is
  984: obtained from the dom0, it is far more likely to be out of sync or
  985: missing with Xen.)
  986: 
  987: Creating an unprivileged Linux domain (domU)
  988: --------------------------------------------
  989: 
  990: Creating unprivileged Linux domains isn't much different from
  991: unprivileged NetBSD domains, but there are some details to know.
  992: 
  993: First, the second parameter passed to the disk declaration (the '0x1' in
  994: the example below)
  995: 
  996:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  997: 
  998: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  999: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
 1000: So, hda1 which has major 3 and minor 1 on a Linux system will have
 1001: device number 0x301.  Alternatively, devices names can be used (hda,
 1002: hdb, ...)  as xentools has a table to map these names to devices
 1003: numbers.  To export a partition to a Linux guest we can use:
 1004: 
 1005:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
 1006:         root = "/dev/hda1 ro"
 1007: 
 1008: and it will appear as /dev/hda on the Linux system, and be used as root
 1009: partition.
 1010: 
 1011: To install the Linux system on the partition to be exported to the
 1012: guest domain, the following method can be used: install
 1013: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
 1014: that will be the root partition of your Linux domain, and mount it.
 1015: Then copy the files from a working Linux system, make adjustments in
 1016: `/etc` (fstab, network config).  It should also be possible to extract
 1017: binary packages such as .rpm or .deb directly to the mounted partition
 1018: using the appropriate tool, possibly running under NetBSD's Linux
 1019: emulation.  Once the file system has been populated, umount it.  If
 1020: desirable, the file system can be converted to ext3 using tune2fs -j.
 1021: It should now be possible to boot the Linux guest domain, using one of
 1022: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
 1023: 
 1024: To get the Linux console right, you need to add:
 1025: 
 1026:     extra = "xencons=tty1"
 1027: 
 1028: to your configuration since not all Linux distributions auto-attach a
 1029: tty to the xen console.
 1030: 
 1031: Creating an unprivileged Solaris domain (domU)
 1032: ----------------------------------------------
 1033: 
 1034: See possibly outdated
 1035: [Solaris domU instructions](/ports/xen/howto-solaris/).
 1036: 
 1037: 
 1038: PCI passthrough: Using PCI devices in guest domains
 1039: ---------------------------------------------------
 1040: 
 1041: The dom0 can give other domains access to selected PCI
 1042: devices. This can allow, for example, a non-privileged domain to have
 1043: access to a physical network interface or disk controller.  However,
 1044: keep in mind that giving a domain access to a PCI device most likely
 1045: will give the domain read/write access to the whole physical memory,
 1046: as PCs don't have an IOMMU to restrict memory access to DMA-capable
 1047: device.  Also, it's not possible to export ISA devices to non-dom0
 1048: domains, which means that the primary VGA adapter can't be exported.
 1049: A guest domain trying to access the VGA registers will panic.
 1050: 
 1051: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
 1052: not been ported to later versions at this time.
 1053: 
 1054: For a PCI device to be exported to a domU, is has to be attached to
 1055: the "pciback" driver in dom0.  Devices passed to the dom0 via the
 1056: pciback.hide boot parameter will attach to "pciback" instead of the
 1057: usual driver.  The list of devices is specified as "(bus:dev.func)",
 1058: where bus and dev are 2-digit hexadecimal numbers, and func a
 1059: single-digit number:
 1060: 
 1061:         pciback.hide=(00:0a.0)(00:06.0)
 1062: 
 1063: pciback devices should show up in the dom0's boot messages, and the
 1064: devices should be listed in the `/kern/xen/pci` directory.
 1065: 
 1066: PCI devices to be exported to a domU are listed in the "pci" array of
 1067: the domU's config file, with the format "0000:bus:dev.func".
 1068: 
 1069:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
 1070: 
 1071: In the domU an "xpci" device will show up, to which one or more pci
 1072: buses will attach.  Then the PCI drivers will attach to PCI buses as
 1073: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
 1074: or any PCI drivers built in by default; you have to build your own
 1075: kernel to use PCI devices in a domU.  Here's a kernel config example;
 1076: note that only the "xpci" lines are unusual.
 1077: 
 1078:         include         "arch/i386/conf/XEN3_DOMU"
 1079: 
 1080:         # Add support for PCI buses to the XEN3_DOMU kernel
 1081:         xpci* at xenbus ?
 1082:         pci* at xpci ?
 1083: 
 1084:         # PCI USB controllers
 1085:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
 1086: 
 1087:         # USB bus support
 1088:         usb*    at uhci?
 1089: 
 1090:         # USB Hubs
 1091:         uhub*   at usb?
 1092:         uhub*   at uhub? port ? configuration ? interface ?
 1093: 
 1094:         # USB Mass Storage
 1095:         umass*  at uhub? port ? configuration ? interface ?
 1096:         wd*     at umass?
 1097:         # SCSI controllers
 1098:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
 1099: 
 1100:         # SCSI bus support (for both ahc and umass)
 1101:         scsibus* at scsi?
 1102: 
 1103:         # SCSI devices
 1104:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
 1105:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
 1106: 
 1107: 
 1108: NetBSD as a domU in a VPS
 1109: =========================
 1110: 
 1111: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
 1112: hardware.  This section explains how to deal with Xen in a domU as a
 1113: virtual private server where you do not control or have access to the
 1114: dom0.  This is not intended to be an exhaustive list of VPS providers;
 1115: only a few are mentioned that specifically support NetBSD.
 1116: 
 1117: VPS operators provide varying degrees of access and mechanisms for
 1118: configuration.  The big issue is usually how one controls which kernel
 1119: is booted, because the kernel is nominally in the dom0 file system (to
 1120: which VPS users do not normally have access).  A second issue is how
 1121: to install NetBSD.
 1122: A VPS user may want to compile a kernel for security updates, to run
 1123: npf, run IPsec, or any other reason why someone would want to change
 1124: their kernel.
 1125: 
 1126: One approach is to have an administrative interface to upload a kernel,
 1127: or to select from a prepopulated list.  Other approaches are pygrub
 1128: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
 1129: kernel from the domU file system.  This is closer to a regular physical
 1130: computer, where someone who controls a machine can replace the kernel.
 1131: 
 1132: A second issue is multiple CPUs.  With NetBSD 6, domUs support
 1133: multiple vcpus, and it is typical for VPS providers to enable multiple
 1134: CPUs for NetBSD domUs.
 1135: 
 1136: pygrub
 1137: -------
 1138: 
 1139: pygrub runs in the dom0 and looks into the domU file system.  This
 1140: implies that the domU must have a kernel in a file system in a format
 1141: known to pygrub.  As of 2014, pygrub seems to be of mostly historical
 1142: interest.
 1143: 
 1144: pvgrub
 1145: ------
 1146: 
 1147: pvgrub is a version of grub that uses PV operations instead of BIOS
 1148: calls.  It is booted from the dom0 as the domU kernel, and then reads
 1149: /grub/menu.lst and loads a kernel from the domU file system.
 1150: 
 1151: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
 1152: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
 1153: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
 1154: page](http://www.panix.com/v-colo/grub.html), which describes only
 1155: Linux but should be updated to cover NetBSD :-).
 1156: 
 1157: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
 1158: their own kernel.  See then [prgmr.com NetBSD
 1159: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
 1160: (which is in need of updating).
 1161: 
 1162: It appears that [grub's FFS
 1163: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
 1164: does not support all aspects of modern FFS, but there are also reports
 1165: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
 1166: partition for the kernel with the intent that grub can understand it,
 1167: which leads to /netbsd not being the actual kernel.  One must remember
 1168: to update the special boot partition.
 1169: 
 1170: Amazon
 1171: ------
 1172: 
 1173: See the [Amazon EC2 page](/amazon_ec2/).
 1174: 
 1175: Using npf
 1176: ---------
 1177: 
 1178: In standard kernels, npf is a module, and thus cannot be loaded in a
 1179: DOMU kernel.
 1180: 
 1181: TODO: Explain how to compile npf into a custom kernel, answering (but
 1182: note that the problem was caused by not booting the right kernel)
 1183: [this email to
 1184: netbsd-users](https://mail-index.netbsd.org/netbsd-users/2014/12/26/msg015576.html).
 1185: 
 1186: TODO items for improving NetBSD/xen
 1187: ===================================
 1188: 
 1189: * Make the NetBSD dom0 kernel work with SMP.
 1190: * Test the Xen 4.5 packages adequately to be able to recommend them as
 1191:   the standard approach.
 1192: * Get PCI passthrough working on Xen 4.5
 1193: * Get pvgrub into pkgsrc, either via xentools or separately.
 1194: * grub
 1195:   * Check/add support to pkgsrc grub2 for UFS2 and arbitrary
 1196:     fragsize/blocksize (UFS2 support may be present; the point is to
 1197:     make it so that with any UFS1/UFS2 file system setup that works
 1198:     with NetBSD grub will also work).
 1199:     See [pkg/40258](https://gnats.netbsd.org/40258).
 1200:   * Push patches upstream.
 1201:   * Get UFS2 patches into pvgrub.
 1202: * Add support for PV ops to a version of /boot, and make it usable as
 1203:   a kernel in Xen, similar to pvgrub.
 1204: * Solve somehow the issue with modules for GENERIC not being loadable
 1205:   in a Xen dom0 or domU kernel.
 1206: 
 1207: Random pointers
 1208: ===============
 1209: 
 1210: This section contains links from elsewhere not yet integrated into the
 1211: HOWTO, and other guides.
 1212: 
 1213: * http://www.lumbercartel.ca/library/xen/
 1214: * http://pbraun.nethence.com/doc/sysutils/xen_netbsd_dom0.html
 1215: * https://gmplib.org/~tege/xen.html

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