File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.161: download - view: text, annotated - select for diffs
Wed Sep 5 09:39:39 2018 UTC (13 months, 1 week ago) by maxv
Branches: MAIN
CVS tags: HEAD
rename one section, for consistency

    1: [[!meta title="Xen HowTo"]]
    2: 
    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.
   10: 
   11: Xen supports different styles of guest:
   12: 
   13: [[!table data="""
   14: Style of guest	|Supported by NetBSD
   15: PV		|Yes (dom0, domU)
   16: HVM		|Yes (domU)
   17: PVHVM		|No
   18: PVH		|No
   19: """]]
   20: 
   21: In Para-Virtualized (PV) mode, the guest OS does not attempt to access
   22: hardware directly, but instead makes hypercalls to the hypervisor; PV
   23: guests must be specifically coded for Xen. In HVM mode, no guest
   24: modification is required; however, hardware support is required, such
   25: as VT-x on Intel CPUs and SVM on AMD CPUs.
   26: 
   27: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
   28: The dom0 can start one or more domUs.  (Booting is explained in detail
   29: in the dom0 section.)
   30: 
   31: This HOWTO presumes a basic familiarity with the Xen system
   32: architecture, with installing NetBSD on i386/amd64 hardware, and with
   33: installing software from pkgsrc.  See also the [Xen
   34: website](http://www.xenproject.org/).
   35: 
   36: [[!toc]]
   37: 
   38: #Versions and Support
   39: 
   40: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
   41: xenkernel and xentools.  We will refer only to the kernel versions,
   42: but note that both packages must be installed together and must have
   43: matching versions.
   44: 
   45: Versions available in pkgsrc:
   46: 
   47: [[!table data="""
   48: Xen Version	|Package Name	|Xen CPU Support	|EOL'ed By Upstream
   49: 4.2		|xenkernel42	|32bit, 64bit		|Yes
   50: 4.5		|xenkernel45	|64bit			|Yes
   51: 4.6		|xenkernel46	|64bit			|Partially
   52: 4.8		|xenkernel48	|64bit			|No
   53: 4.11		|xenkernel411	|64bit			|No
   54: """]]
   55: 
   56: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
   57: 
   58: Multiprocessor (SMP) support in NetBSD differs depending on the domain:
   59: 
   60: [[!table data="""
   61: Domain		|Supports SMP
   62: dom0		|No
   63: domU		|Yes
   64: """]]
   65: 
   66: Note: NetBSD support is called XEN3. However, it does support Xen 4,
   67: because the hypercall interface has remained identical.
   68: 
   69: Architecture
   70: ------------
   71: 
   72: Xen itself runs on x86_64 hardware.
   73: 
   74: The dom0 system, plus each domU, can be either i386PAE or amd64.
   75: i386 without PAE is not supported.
   76: 
   77: The standard approach is to use NetBSD/amd64 for the dom0.
   78: 
   79: To use an i386PAE dom0, one must build or obtain a 64bit Xen kernel and
   80: install it on the system.
   81: 
   82: For domUs, i386PAE is considered as
   83: [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html)
   84: than amd64.
   85: 
   86: #Creating a dom0
   87: 
   88: In order to install a NetBSD as a dom0, one must first install a normal
   89: NetBSD system, and then pivot the install to a dom0 install by changing
   90: the kernel and boot configuration.
   91: 
   92: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
   93: 512M it worked reliably.  This does not make sense, but if you see
   94: "not ELF" after Xen boots, try increasing dom0 RAM.
   95: 
   96: Installation of NetBSD
   97: ----------------------
   98: 
   99: [Install NetBSD/amd64](/guide/inst/)
  100: just as you would if you were not using Xen.
  101: 
  102: Installation of Xen
  103: -------------------
  104: 
  105: We will consider that you chose to use Xen 4.8, with NetBSD/amd64 as
  106: dom0. In the dom0, install xenkernel48 and xentools48 from pkgsrc.
  107: 
  108: Once this is done, install the Xen kernel itself:
  109: 
  110: [[!template id=programlisting text="""
  111: # cp /usr/pkg/xen48-kernel/xen.gz /
  112: """]]
  113: 
  114: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such kernel
  115: can either be compiled manually, or downloaded from the NetBSD FTP, for
  116: example at:
  117: 
  118: [[!template id=programlisting text="""
  119: ftp.netbsd.org/pub/NetBSD/NetBSD-8.0/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
  120: """]]
  121: 
  122: Add a line to /boot.cfg to boot Xen:
  123: 
  124: [[!template id=filecontent name="/boot.cfg" text="""
  125: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
  126: """]]
  127: 
  128: This specifies that the dom0 should have 512MB of ram, leaving the rest
  129: to be allocated for domUs.  To use a serial console, use:
  130: 
  131: [[!template id=filecontent name="/boot.cfg" text="""
  132: menu=Xen:load /netbsd-XEN3_DOM0.gz;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
  133: """]]
  134: 
  135: which will use the first serial port for Xen (which counts starting
  136: from 1, unlike NetBSD which counts starting from 0), forcing
  137: speed/parity.  Because the NetBSD command line lacks a
  138: "console=pc" argument, it will use the default "xencons" console device,
  139: which directs the console I/O through Xen to the same console device Xen
  140: itself uses (in this case, the serial port).
  141: 
  142: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`,
  143: to force only one vcpu to be provided (since NetBSD dom0 can't use
  144: more) and to pin that vcpu to a physical CPU. Xen has
  145: [many boot options](http://xenbits.xenproject.org/docs/4.8-testing/misc/xen-command-line.html),
  146: and other than dom0 memory and max_vcpus, they are generally not
  147: necessary.
  148: 
  149: Copy the boot scripts into `/etc/rc.d`:
  150: 
  151: [[!template id=programlisting text="""
  152: # cp /usr/pkg/share/examples/rc.d/xen* /etc/rc.d/
  153: """]]
  154: 
  155: Enable `xencommons`:
  156: 
  157: [[!template id=filecontent name="/etc/rc.conf" text="""
  158: xencommons=YES
  159: """]]
  160: 
  161: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  162: than GENERIC without Xen.
  163: 
  164: TODO: Recommend for/against xen-watchdog.
  165: 
  166: Once the reboot is done, use `xl` to inspect Xen's boot messages,
  167: available resources, and running domains.  For example:
  168: 
  169: [[!template id=programlisting text="""
  170: # xl dmesg
  171: ... xen's boot info ...
  172: # xl info
  173: ... available memory, etc ...
  174: # xl list
  175: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  176: Domain-0           0       64    0  r----     58.1
  177: """]]
  178: 
  179: Xen logs will be in /var/log/xen.
  180: 
  181: ### Issues with xencommons
  182: 
  183: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
  184: domUs.  It does not currently work to stop and start xenstored.
  185: Certainly all domUs should be shutdown first, following the sort order
  186: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
  187: and is not notified when xenstored exits, leading to not recreating
  188: the state when the new xenstored starts.  Until there's a mechanism to
  189: make this work, one should not expect to be able to restart xenstored
  190: (and thus xencommons).  There is currently no reason to expect that
  191: this will get fixed any time soon.
  192: 
  193: anita (for testing NetBSD)
  194: --------------------------
  195: 
  196: With the setup so far (assuming 4.8/xl), one should be able to run
  197: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
  198: root, because anita must create a domU):
  199: 
  200: [[!template id=programlisting text="""
  201: anita --vmm=xl test file:///usr/obj/i386/
  202: """]]
  203: 
  204: Xen-specific NetBSD issues
  205: --------------------------
  206: 
  207: There are (at least) two additional things different about NetBSD as a
  208: dom0 kernel compared to hardware.
  209: 
  210: One is that the module ABI is different because some of the #defines
  211: change, so one must build modules for Xen.  As of netbsd-7, the build
  212: system does this automatically.
  213: 
  214: The other difference is that XEN3_DOM0 does not have exactly the same
  215: options as GENERIC.  While it is debatable whether or not this is a
  216: bug, users should be aware of this and can simply add missing config
  217: items if desired.
  218: 
  219: Updating NetBSD in a dom0
  220: -------------------------
  221: 
  222: This is just like updating NetBSD on bare hardware, assuming the new
  223: version supports the version of Xen you are running.  Generally, one
  224: replaces the kernel and reboots, and then overlays userland binaries
  225: and adjusts `/etc`.
  226: 
  227: Note that one must update both the non-Xen kernel typically used for
  228: rescue purposes and the DOM0 kernel used with Xen.
  229: 
  230: Converting from grub to /boot
  231: -----------------------------
  232: 
  233: These instructions were used to convert a system from
  234: grub to /boot.  The system was originally installed in February of
  235: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
  236: over time.  Before these commands, it was running NetBSD 6 i386, Xen
  237: 4.1 and grub, much like the message linked earlier in the grub
  238: section.
  239: 
  240: [[!template id=programlisting text="""
  241: # Install MBR bootblocks on both disks.
  242: fdisk -i /dev/rwd0d
  243: fdisk -i /dev/rwd1d
  244: # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
  245: installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
  246: installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
  247: # Install secondary boot loader
  248: cp -p /usr/mdec/boot /
  249: # Create boot.cfg following earlier guidance:
  250: menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
  251: menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=512M
  252: menu=GENERIC:boot
  253: menu=GENERIC single-user:boot -s
  254: menu=GENERIC.ok:boot netbsd.ok
  255: menu=GENERIC.ok single-user:boot netbsd.ok -s
  256: menu=Drop to boot prompt:prompt
  257: default=1
  258: timeout=30
  259: """]]
  260: 
  261: Upgrading Xen versions
  262: ---------------------
  263: 
  264: Minor version upgrades are trivial.  Just rebuild/replace the
  265: xenkernel version and copy the new xen.gz to `/` (where `/boot.cfg`
  266: references it), and reboot.
  267: 
  268: #Unprivileged domains (domU)
  269: 
  270: This section describes general concepts about domUs.  It does not
  271: address specific domU operating systems or how to install them.  The
  272: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
  273: typically named so that the file name, domU name and the domU's host
  274: name match.
  275: 
  276: The domU is provided with CPU and memory by Xen, configured by the
  277: dom0.  The domU is provided with disk and network by the dom0,
  278: mediated by Xen, and configured in the dom0.
  279: 
  280: Entropy in domUs can be an issue; physical disks and network are on
  281: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  282: 
  283: Config files
  284: ------------
  285: 
  286: See /usr/pkg/share/examples/xen/xlexample*,
  287: for a large number of well-commented examples, mostly for running
  288: GNU/Linux.
  289: 
  290: The following is an example minimal domain configuration file. The domU
  291: serves as a network file server.
  292: 
  293: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  294: name = "domU-id"
  295: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  296: memory = 1024
  297: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  298: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  299:          'file:/n0/xen/foo-wd1,0x1,w' ]
  300: """]]
  301: 
  302: The domain will have name given in the `name` setting.  The kernel has the
  303: host/domU name in it, so that on the dom0 one can update the various
  304: domUs independently.  The `vif` line causes an interface to be provided,
  305: with a specific mac address (do not reuse MAC addresses!), in bridge
  306: mode.  Two disks are provided, and they are both writable; the bits
  307: are stored in files and Xen attaches them to a vnd(4) device in the
  308: dom0 on domain creation.  The system treats xbd0 as the boot device
  309: without needing explicit configuration.
  310: 
  311: By default, `xl` looks for domain config files in `/usr/pkg/etc/xen`.  Note
  312: that "xl create" takes the name of a config file, while other commands
  313: take the name of a domain.
  314: 
  315: Examples of commands:
  316: 
  317: [[!template id=programlisting text="""
  318: xl create foo
  319: xl console foo
  320: xl create -c foo
  321: xl shutdown foo
  322: xl list
  323: """]]
  324: 
  325: Typing `^]` will exit the console session.  Shutting down a domain is
  326: equivalent to pushing the power button; a NetBSD domU will receive a
  327: power-press event and do a clean shutdown.  Shutting down the dom0
  328: will trigger controlled shutdowns of all configured domUs.
  329: 
  330: CPU and memory
  331: --------------
  332: 
  333: A domain is provided with some number of vcpus, less than the number
  334: of CPUs seen by the hypervisor. For a domU, it is controlled
  335: from the config file by the "vcpus = N" directive.
  336: 
  337: A domain is provided with memory; this is controlled in the config
  338: file by "memory = N" (in megabytes).  In the straightforward case, the
  339: sum of the the memory allocated to the dom0 and all domUs must be less
  340: than the available memory.
  341: 
  342: Xen also provides a "balloon" driver, which can be used to let domains
  343: use more memory temporarily.
  344: 
  345: Virtual disks
  346: -------------
  347: 
  348: In domU config files, the disks are defined as a sequence of 3-tuples:
  349: 
  350:  * The first element is "method:/path/to/disk". Common methods are
  351:    "file:" for file-backed vnd, and "phy:" for something that is already
  352:    a device.
  353: 
  354:  * The second element is an artifact of how virtual disks are passed to
  355:    Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  356:    are given a device name to associate with the disk, and values like
  357:    "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  358:    as xbd0, the second as xbd1, and so on.  However, xl demands a
  359:    second argument.  The name given is converted to a major/minor by
  360:    calling stat(2) on the name in /dev and this is passed to the domU.
  361:    In the general case, the dom0 and domU can be different operating
  362:    systems, and it is an unwarranted assumption that they have consistent
  363:    numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  364:    as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  365:    for the second works fine and avoids this issue.  For a GNU/Linux
  366:    guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  367:    /dev/hda1.
  368: 
  369:  * The third element is "w" for writable disks, and "r" for read-only
  370:    disks.
  371: 
  372: Example:
  373: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
  374: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ]
  375: """]]
  376: 
  377: Note that NetBSD by default creates only vnd[0123].  If you need more
  378: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
  379: dom0.
  380: 
  381: Note that NetBSD by default creates only xbd[0123].  If you need more
  382: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
  383: 
  384: Virtual Networking
  385: ------------------
  386: 
  387: Xen provides virtual Ethernets, each of which connects the dom0 and a
  388: domU.  For each virtual network, there is an interface "xvifN.M" in
  389: the dom0, and a matching interface xennetM (NetBSD name) in domU index N.
  390: The interfaces behave as if there is an Ethernet with two
  391: adapters connected.  From this primitive, one can construct various
  392: configurations.  We focus on two common and useful cases for which
  393: there are existing scripts: bridging and NAT.
  394: 
  395: With bridging (in the example above), the domU perceives itself to be
  396: on the same network as the dom0.  For server virtualization, this is
  397: usually best.  Bridging is accomplished by creating a bridge(4) device
  398: and adding the dom0's physical interface and the various xvifN.0
  399: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  400: config file.  The bridge must be set up already in the dom0; an
  401: example /etc/ifconfig.bridge0 is:
  402: 
  403: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
  404: create
  405: up
  406: !brconfig bridge0 add wm0
  407: """]]
  408: 
  409: With NAT, the domU perceives itself to be behind a NAT running on the
  410: dom0.  This is often appropriate when running Xen on a workstation.
  411: TODO: NAT appears to be configured by "vif = [ '' ]".
  412: 
  413: The MAC address specified is the one used for the interface in the new
  414: domain.  The interface in dom0 will use this address XOR'd with
  415: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  416: 
  417: Starting domains automatically
  418: ------------------------------
  419: 
  420: To start domains `domU-netbsd` and `domU-linux` at boot and shut them
  421: down cleanly on dom0 shutdown, add the following in rc.conf:
  422: 
  423: [[!template id=filecontent name="/etc/rc.conf" text="""
  424: xendomains="domU-netbsd domU-linux"
  425: """]]
  426: 
  427: #Creating a domU
  428: 
  429: Creating domUs is almost entirely independent of operating system.  We
  430: have already presented the basics of config files.  Note that you must
  431: have already completed the dom0 setup so that "xl list" works.
  432: 
  433: Creating a NetBSD domU
  434: ----------------------
  435: 
  436: See the earlier config file, and adjust memory.  Decide on how much
  437: storage you will provide, and prepare it (file or lvm).
  438: 
  439: While the kernel will be obtained from the dom0 file system, the same
  440: file should be present in the domU as /netbsd so that tools like
  441: savecore(8) can work.   (This is helpful but not necessary.)
  442: 
  443: The kernel must be specifically for Xen and for use as a domU.  The
  444: i386 and amd64 provide the following kernels:
  445: 
  446:         i386 XEN3PAE_DOMU
  447:         amd64 XEN3_DOMU
  448: 
  449: This will boot NetBSD, but this is not that useful if the disk is
  450: empty.  One approach is to unpack sets onto the disk outside of xen
  451: (by mounting it, just as you would prepare a physical disk for a
  452: system you can't run the installer on).
  453: 
  454: A second approach is to run an INSTALL kernel, which has a miniroot
  455: and can load sets from the network.  To do this, copy the INSTALL
  456: kernel to / and change the kernel line in the config file to:
  457: 
  458:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  459: 
  460: Then, start the domain as "xl create -c configname".
  461: 
  462: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  463: line should be used in the config file.
  464: 
  465:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  466: 
  467: After booting the domain, the option to install via CDROM may be
  468: selected.  The CDROM device should be changed to `xbd1d`.
  469: 
  470: Once done installing, "halt -p" the new domain (don't reboot or halt,
  471: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  472: config file), switch the config file back to the XEN3_DOMU kernel,
  473: and start the new domain again. Now it should be able to use "root on
  474: xbd0a" and you should have a, functional NetBSD domU.
  475: 
  476: TODO: check if this is still accurate.
  477: When the new domain is booting you'll see some warnings about *wscons*
  478: and the pseudo-terminals. These can be fixed by editing the files
  479: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  480: `/etc/ttys`, except *console*, like this:
  481: 
  482:     console "/usr/libexec/getty Pc"         vt100   on secure
  483:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  484:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  485:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  486:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  487: 
  488: Finally, all screens must be commented out from `/etc/wscons.conf`.
  489: 
  490: It is also desirable to add
  491: 
  492:         powerd=YES
  493: 
  494: in rc.conf. This way, the domain will be properly shut down if
  495: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
  496: 
  497: It is not strictly necessary to have a kernel (as /netbsd) in the domU
  498: file system.  However, various programs (e.g. netstat) will use that
  499: kernel to look up symbols to read from kernel virtual memory.  If
  500: /netbsd is not the running kernel, those lookups will fail.  (This is
  501: not really a Xen-specific issue, but because the domU kernel is
  502: obtained from the dom0, it is far more likely to be out of sync or
  503: missing with Xen.)
  504: 
  505: Creating a Linux domU
  506: ---------------------
  507: 
  508: Creating unprivileged Linux domains isn't much different from
  509: unprivileged NetBSD domains, but there are some details to know.
  510: 
  511: First, the second parameter passed to the disk declaration (the '0x1' in
  512: the example below)
  513: 
  514:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  515: 
  516: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  517: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
  518: So, hda1 which has major 3 and minor 1 on a Linux system will have
  519: device number 0x301.  Alternatively, devices names can be used (hda,
  520: hdb, ...)  as xentools has a table to map these names to devices
  521: numbers.  To export a partition to a Linux guest we can use:
  522: 
  523:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
  524:         root = "/dev/hda1 ro"
  525: 
  526: and it will appear as /dev/hda on the Linux system, and be used as root
  527: partition.
  528: 
  529: To install the Linux system on the partition to be exported to the
  530: guest domain, the following method can be used: install
  531: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
  532: that will be the root partition of your Linux domain, and mount it.
  533: Then copy the files from a working Linux system, make adjustments in
  534: `/etc` (fstab, network config).  It should also be possible to extract
  535: binary packages such as .rpm or .deb directly to the mounted partition
  536: using the appropriate tool, possibly running under NetBSD's Linux
  537: emulation.  Once the file system has been populated, umount it.  If
  538: desirable, the file system can be converted to ext3 using tune2fs -j.
  539: It should now be possible to boot the Linux guest domain, using one of
  540: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
  541: 
  542: To get the Linux console right, you need to add:
  543: 
  544:     extra = "xencons=tty1"
  545: 
  546: to your configuration since not all Linux distributions auto-attach a
  547: tty to the xen console.
  548: 
  549: Creating a Solaris domU
  550: -----------------------
  551: 
  552: See possibly outdated
  553: [Solaris domU instructions](/ports/xen/howto-solaris/).
  554: 
  555: 
  556: PCI passthrough: Using PCI devices in guest domains
  557: ---------------------------------------------------
  558: 
  559: The dom0 can give other domains access to selected PCI
  560: devices. This can allow, for example, a non-privileged domain to have
  561: access to a physical network interface or disk controller.  However,
  562: keep in mind that giving a domain access to a PCI device most likely
  563: will give the domain read/write access to the whole physical memory,
  564: as PCs don't have an IOMMU to restrict memory access to DMA-capable
  565: device.  Also, it's not possible to export ISA devices to non-dom0
  566: domains, which means that the primary VGA adapter can't be exported.
  567: A guest domain trying to access the VGA registers will panic.
  568: 
  569: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
  570: not been ported to later versions at this time.
  571: 
  572: For a PCI device to be exported to a domU, is has to be attached to
  573: the "pciback" driver in dom0.  Devices passed to the dom0 via the
  574: pciback.hide boot parameter will attach to "pciback" instead of the
  575: usual driver.  The list of devices is specified as "(bus:dev.func)",
  576: where bus and dev are 2-digit hexadecimal numbers, and func a
  577: single-digit number:
  578: 
  579:         pciback.hide=(00:0a.0)(00:06.0)
  580: 
  581: pciback devices should show up in the dom0's boot messages, and the
  582: devices should be listed in the `/kern/xen/pci` directory.
  583: 
  584: PCI devices to be exported to a domU are listed in the "pci" array of
  585: the domU's config file, with the format "0000:bus:dev.func".
  586: 
  587:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
  588: 
  589: In the domU an "xpci" device will show up, to which one or more pci
  590: buses will attach.  Then the PCI drivers will attach to PCI buses as
  591: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
  592: or any PCI drivers built in by default; you have to build your own
  593: kernel to use PCI devices in a domU.  Here's a kernel config example;
  594: note that only the "xpci" lines are unusual.
  595: 
  596:         include         "arch/i386/conf/XEN3_DOMU"
  597: 
  598:         # Add support for PCI buses to the XEN3_DOMU kernel
  599:         xpci* at xenbus ?
  600:         pci* at xpci ?
  601: 
  602:         # PCI USB controllers
  603:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
  604: 
  605:         # USB bus support
  606:         usb*    at uhci?
  607: 
  608:         # USB Hubs
  609:         uhub*   at usb?
  610:         uhub*   at uhub? port ? configuration ? interface ?
  611: 
  612:         # USB Mass Storage
  613:         umass*  at uhub? port ? configuration ? interface ?
  614:         wd*     at umass?
  615:         # SCSI controllers
  616:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
  617: 
  618:         # SCSI bus support (for both ahc and umass)
  619:         scsibus* at scsi?
  620: 
  621:         # SCSI devices
  622:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
  623:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
  624: 
  625: 
  626: #NetBSD as a domU in a VPS
  627: 
  628: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
  629: hardware.  This section explains how to deal with Xen in a domU as a
  630: virtual private server where you do not control or have access to the
  631: dom0.  This is not intended to be an exhaustive list of VPS providers;
  632: only a few are mentioned that specifically support NetBSD.
  633: 
  634: VPS operators provide varying degrees of access and mechanisms for
  635: configuration.  The big issue is usually how one controls which kernel
  636: is booted, because the kernel is nominally in the dom0 file system (to
  637: which VPS users do not normally have access).  A second issue is how
  638: to install NetBSD.
  639: A VPS user may want to compile a kernel for security updates, to run
  640: npf, run IPsec, or any other reason why someone would want to change
  641: their kernel.
  642: 
  643: One approach is to have an administrative interface to upload a kernel,
  644: or to select from a prepopulated list.  Other approaches are pygrub
  645: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
  646: kernel from the domU file system.  This is closer to a regular physical
  647: computer, where someone who controls a machine can replace the kernel.
  648: 
  649: A second issue is multiple CPUs.  With NetBSD 6, domUs support
  650: multiple vcpus, and it is typical for VPS providers to enable multiple
  651: CPUs for NetBSD domUs.
  652: 
  653: pygrub
  654: -------
  655: 
  656: pygrub runs in the dom0 and looks into the domU file system.  This
  657: implies that the domU must have a kernel in a file system in a format
  658: known to pygrub.  As of 2014, pygrub seems to be of mostly historical
  659: interest.
  660: 
  661: pvgrub
  662: ------
  663: 
  664: pvgrub is a version of grub that uses PV operations instead of BIOS
  665: calls.  It is booted from the dom0 as the domU kernel, and then reads
  666: /grub/menu.lst and loads a kernel from the domU file system.
  667: 
  668: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
  669: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
  670: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
  671: page](http://www.panix.com/v-colo/grub.html), which describes only
  672: Linux but should be updated to cover NetBSD :-).
  673: 
  674: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
  675: their own kernel.  See then [prgmr.com NetBSD
  676: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
  677: (which is in need of updating).
  678: 
  679: It appears that [grub's FFS
  680: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
  681: does not support all aspects of modern FFS, but there are also reports
  682: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
  683: partition for the kernel with the intent that grub can understand it,
  684: which leads to /netbsd not being the actual kernel.  One must remember
  685: to update the special boot partition.
  686: 
  687: Amazon
  688: ------
  689: 
  690: See the [Amazon EC2 page](/amazon_ec2/).

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