wikisrc/ports/xen/howto.mdwn - view

File: [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.180: download - view: text, annotated - select for diffs
Sun Nov 15 14:31:58 2020 UTC (3 months, 2 weeks ago) by gdt
Branches: MAIN
CVS tags: HEAD

Rewrite PVH.  Add sections

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 guests; see [PV on HVM](https://wiki.xen.org/wiki/PV_on_HVM) and [PVH(v2)](https://wiki.xenproject.org/wiki/PVH_(v2\)_Domu) for upstream documentation. 12: 13: [[!table data=""" 14: Style of guest |Supported by NetBSD 15: PV |Yes (dom0, domU) 16: HVM |Yes (domU) 17: PVHVM |current-only (domU) 18: PVH |current-only (domU, dom0 not yet) 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. 24: 25: In HVM mode, no guest modification is required; however, hardware 26: support is required, such as VT-x on Intel CPUs and SVM on AMD CPUs. 27: The dom0 runs qemu to emulate hardware. 28: 29: In PVHVM mode, the guest runs as HVM, but additionally can use PV 30: drivers for efficiency. 31: 32: There have been two PVH modes: original PVH and PVHv2. Original PVH 33: was based on PV mode and is no longer relevant at all. PVHv2 is 34: basically lightweight HVM with PV drivers. A critical feature of it 35: is that qemu is not needed; the hypervisor can do the emulation that 36: is required. Thus, a dom0 can be PVHv2. 37: 38: The source code uses PVH and config files use pvh; this refers to PVHv2. 39: 40: At boot, the dom0 kernel is loaded as a module with Xen as the kernel. 41: The dom0 can start one or more domUs. (Booting is explained in detail 42: in the dom0 section.) 43: 44: This HOWTO presumes a basic familiarity with the Xen system 45: architecture, with installing NetBSD on i386/amd64 hardware, and with 46: installing software from pkgsrc. See also the [Xen 47: website](http://www.xenproject.org/). 48: 49: [[!toc]] 50: 51: # Versions and Support 52: 53: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages 54: xenkernel and xentools. We will refer only to the kernel versions, 55: but note that both packages must be installed together and must have 56: matching versions. 57: 58: Versions available in pkgsrc: 59: 60: [[!table data=""" 61: Xen Version |Package Name |Xen CPU Support |xm? |EOL'ed By Upstream 62: 4.11 |xenkernel411 |x86_64 | |No 63: 4.13 |xenkernel413 |x86_64 | |No 64: """]] 65: 66: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/). 67: 68: Multiprocessor (SMP) support in NetBSD differs depending on the domain: 69: 70: [[!table data=""" 71: Domain |Supports SMP 72: dom0 |No 73: domU |Yes 74: """]] 75: 76: Note: NetBSD support is called XEN3. However, it does support Xen 4, 77: because the hypercall interface has remained identical. 78: 79: Older Xen had a python-based management tool called xm, now replaced 80: by xl. 81: 82: Architecture 83: ------------ 84: 85: Xen 4.5 and later runs on x86_64 hardware (the NetBSD amd64 port). 86: There is a concept of Xen running on ARM, but there are no reports of this working with NetBSD. 87: 88: The dom0 system should be amd64. (Instructions for i386PAE dom0 have been removed from the HOWTO.) 89: 90: The domU can be i386PAE or amd64. 91: i386PAE at one point was considered as [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html) than amd64. 92: 93: # Creating a dom0 94: 95: In order to install a NetBSD as a dom0, one must first install a normal 96: NetBSD system, and then pivot the install to a dom0 install by changing 97: the kernel and boot configuration. 98: 99: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with 100: 512M it worked reliably. This does not make sense, but if you see 101: "not ELF" after Xen boots, try increasing dom0 RAM. 102: 103: Installation of NetBSD 104: ---------------------- 105: 106: [Install NetBSD/amd64](/guide/inst/) 107: just as you would if you were not using Xen. 108: 109: Installation of Xen 110: ------------------- 111: 112: We will consider that you chose to use Xen 4.13, with NetBSD/amd64 as 113: dom0. In the dom0, install xenkernel48 and xentools48 from pkgsrc. 114: 115: Once this is done, install the Xen kernel itself: 116: 117: [[!template id=programlisting text=""" 118: # cp /usr/pkg/xen48-kernel/xen.gz / 119: """]] 120: 121: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such kernel 122: can either be compiled manually, or downloaded from the NetBSD FTP, for 123: example at: 124: 125: [[!template id=programlisting text=""" 126: ftp.netbsd.org/pub/NetBSD/NetBSD-8.0/amd64/binary/kernel/netbsd-XEN3_DOM0.gz 127: """]] 128: 129: Add a line to /boot.cfg to boot Xen: 130: 131: [[!template id=filecontent name="/boot.cfg" text=""" 132: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M 133: """]] 134: 135: This specifies that the dom0 should have 512MB of ram, leaving the rest 136: to be allocated for domUs. To use a serial console, use: 137: 138: [[!template id=filecontent name="/boot.cfg" text=""" 139: menu=Xen:load /netbsd-XEN3_DOM0.gz;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1 140: """]] 141: 142: which will use the first serial port for Xen (which counts starting 143: from 1, unlike NetBSD which counts starting from 0), forcing 144: speed/parity. Because the NetBSD command line lacks a 145: "console=pc" argument, it will use the default "xencons" console device, 146: which directs the console I/O through Xen to the same console device Xen 147: itself uses (in this case, the serial port). 148: 149: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`, 150: to force only one vcpu to be provided (since NetBSD dom0 can't use 151: more) and to pin that vcpu to a physical CPU. Xen has 152: [many boot options](http://xenbits.xenproject.org/docs/4.13-testing/misc/xen-command-line.html), 153: and other than dom0 memory and max_vcpus, they are generally not 154: necessary. 155: 156: Copy the boot scripts into `/etc/rc.d`: 157: 158: [[!template id=programlisting text=""" 159: # cp /usr/pkg/share/examples/rc.d/xen* /etc/rc.d/ 160: """]] 161: 162: Enable `xencommons`: 163: 164: [[!template id=filecontent name="/etc/rc.conf" text=""" 165: xencommons=YES 166: """]] 167: 168: Now, reboot so that you are running a DOM0 kernel under Xen, rather 169: than GENERIC without Xen. 170: 171: TODO: Recommend for/against xen-watchdog. 172: 173: Once the reboot is done, use `xl` to inspect Xen's boot messages, 174: available resources, and running domains. For example: 175: 176: [[!template id=programlisting text=""" 177: # xl dmesg 178: ... xen's boot info ... 179: # xl info 180: ... available memory, etc ... 181: # xl list 182: Name Id Mem(MB) CPU State Time(s) Console 183: Domain-0 0 64 0 r---- 58.1 184: """]] 185: 186: Xen logs will be in /var/log/xen. 187: 188: ### Issues with xencommons 189: 190: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and 191: domUs. It does not currently work to stop and start xenstored. 192: Certainly all domUs should be shutdown first, following the sort order 193: of the rc.d scripts. However, the dom0 sets up state with xenstored, 194: and is not notified when xenstored exits, leading to not recreating 195: the state when the new xenstored starts. Until there's a mechanism to 196: make this work, one should not expect to be able to restart xenstored 197: (and thus xencommons). There is currently no reason to expect that 198: this will get fixed any time soon. 199: 200: anita (for testing NetBSD) 201: -------------------------- 202: 203: With the setup so far, one should be able to run 204: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as 205: root, because anita must create a domU): 206: 207: [[!template id=programlisting text=""" 208: anita --vmm=xl test file:///usr/obj/i386/ 209: """]] 210: 211: Xen-specific NetBSD issues 212: -------------------------- 213: 214: There are (at least) two additional things different about NetBSD as a 215: dom0 kernel compared to hardware. 216: 217: One is that the module ABI is different because some of the #defines 218: change, so one must build modules for Xen. As of netbsd-7, the build 219: system does this automatically. 220: 221: The other difference is that XEN3_DOM0 does not have exactly the same 222: options as GENERIC. While it is debatable whether or not this is a 223: bug, users should be aware of this and can simply add missing config 224: items if desired. 225: 226: Updating NetBSD in a dom0 227: ------------------------- 228: 229: This is just like updating NetBSD on bare hardware, assuming the new 230: version supports the version of Xen you are running. Generally, one 231: replaces the kernel and reboots, and then overlays userland binaries 232: and adjusts `/etc`. 233: 234: Note that one must update both the non-Xen kernel typically used for 235: rescue purposes and the DOM0 kernel used with Xen. 236: 237: Converting from grub to /boot 238: ----------------------------- 239: 240: These instructions were used to convert a system from 241: grub to /boot. The system was originally installed in February of 242: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated 243: over time. Before these commands, it was running NetBSD 6 i386, Xen 244: 4.1 and grub, much like the message linked earlier in the grub 245: section. 246: 247: [[!template id=programlisting text=""" 248: # Install MBR bootblocks on both disks. 249: fdisk -i /dev/rwd0d 250: fdisk -i /dev/rwd1d 251: # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components. 252: installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1 253: installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1 254: # Install secondary boot loader 255: cp -p /usr/mdec/boot / 256: # Create boot.cfg following earlier guidance: 257: menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M 258: menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=512M 259: menu=GENERIC:boot 260: menu=GENERIC single-user:boot -s 261: menu=GENERIC.ok:boot netbsd.ok 262: menu=GENERIC.ok single-user:boot netbsd.ok -s 263: menu=Drop to boot prompt:prompt 264: default=1 265: timeout=30 266: """]] 267: 268: Upgrading Xen versions 269: --------------------- 270: 271: Minor version upgrades are trivial. Just rebuild/replace the 272: xenkernel version and copy the new xen.gz to `/` (where `/boot.cfg` 273: references it), and reboot. 274: 275: #Unprivileged domains (domU) 276: 277: This section describes general concepts about domUs. It does not 278: address specific domU operating systems or how to install them. The 279: config files for domUs are typically in `/usr/pkg/etc/xen`, and are 280: typically named so that the file name, domU name and the domU's host 281: name match. 282: 283: The domU is provided with CPU and memory by Xen, configured by the 284: dom0. The domU is provided with disk and network by the dom0, 285: mediated by Xen, and configured in the dom0. 286: 287: Entropy in domUs can be an issue; physical disks and network are on 288: the dom0. NetBSD's /dev/random system works, but is often challenged. 289: 290: Config files 291: ------------ 292: 293: See /usr/pkg/share/examples/xen/xlexample* 294: for a small number of well-commented examples, mostly for running 295: GNU/Linux. 296: 297: The following is an example minimal domain configuration file. The domU 298: serves as a network file server. 299: 300: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text=""" 301: name = "domU-id" 302: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz" 303: memory = 1024 304: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ] 305: disk = [ 'file:/n0/xen/foo-wd0,0x0,w', 306: 'file:/n0/xen/foo-wd1,0x1,w' ] 307: """]] 308: 309: The domain will have name given in the `name` setting. The kernel has the 310: host/domU name in it, so that on the dom0 one can update the various 311: domUs independently. The `vif` line causes an interface to be provided, 312: with a specific mac address (do not reuse MAC addresses!), in bridge 313: mode. Two disks are provided, and they are both writable; the bits 314: are stored in files and Xen attaches them to a vnd(4) device in the 315: dom0 on domain creation. The system treats xbd0 as the boot device 316: without needing explicit configuration. 317: 318: By convention, domain config files are kept in `/usr/pkg/etc/xen`. Note 319: that "xl create" takes the name of a config file, while other commands 320: take the name of a domain. 321: 322: Examples of commands: 323: 324: [[!template id=programlisting text=""" 325: xl create /usr/pkg/etc/xen/foo 326: xl console domU-id 327: xl create -c /usr/pkg/etc/xen/foo 328: xl shutdown domU-id 329: xl list 330: """]] 331: 332: Typing `^]` will exit the console session. Shutting down a domain is 333: equivalent to pushing the power button; a NetBSD domU will receive a 334: power-press event and do a clean shutdown. Shutting down the dom0 335: will trigger controlled shutdowns of all configured domUs. 336: 337: CPU and memory 338: -------------- 339: 340: A domain is provided with some number of vcpus, up to the number 341: of CPUs seen by the hypervisor. For a domU, it is controlled 342: from the config file by the "vcpus = N" directive. 343: 344: A domain is provided with memory; this is controlled in the config 345: file by "memory = N" (in megabytes). In the straightforward case, the 346: sum of the the memory allocated to the dom0 and all domUs must be less 347: than the available memory. 348: 349: Xen also provides a "balloon" driver, which can be used to let domains 350: use more memory temporarily. 351: 352: Virtual disks 353: ------------- 354: 355: In domU config files, the disks are defined as a sequence of 3-tuples: 356: 357: * The first element is "method:/path/to/disk". Common methods are 358: "file:" for a file-backed vnd, and "phy:" for something that is already 359: a device, such as an LVM logical volume. 360: 361: * The second element is an artifact of how virtual disks are passed to 362: Linux, and a source of confusion with NetBSD Xen usage. Linux domUs 363: are given a device name to associate with the disk, and values like 364: "hda1" or "sda1" are common. In a NetBSD domU, the first disk appears 365: as xbd0, the second as xbd1, and so on. However, xl demands a 366: second argument. The name given is converted to a major/minor by 367: calling stat(2) on the name in /dev and this is passed to the domU. 368: In the general case, the dom0 and domU can be different operating 369: systems, and it is an unwarranted assumption that they have consistent 370: numbering in /dev, or even that the dom0 OS has a /dev. With NetBSD 371: as both dom0 and domU, using values of 0x0 for the first disk and 0x1 372: for the second works fine and avoids this issue. For a GNU/Linux 373: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for 374: /dev/hda1. 375: 376: * The third element is "w" for writable disks, and "r" for read-only 377: disks. 378: 379: Example: 380: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text=""" 381: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ] 382: """]] 383: 384: Note that NetBSD by default creates only vnd[0123]. If you need more 385: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the 386: dom0. 387: 388: Note that NetBSD by default creates only xbd[0123]. If you need more 389: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU. 390: 391: Virtual Networking 392: ------------------ 393: 394: Xen provides virtual Ethernets, each of which connects the dom0 and a 395: domU. For each virtual network, there is an interface "xvifN.M" in 396: the dom0, and a matching interface xennetM (NetBSD name) in domU index N. 397: The interfaces behave as if there is an Ethernet with two 398: adapters connected. From this primitive, one can construct various 399: configurations. We focus on two common and useful cases for which 400: there are existing scripts: bridging and NAT. 401: 402: With bridging (in the example above), the domU perceives itself to be 403: on the same network as the dom0. For server virtualization, this is 404: usually best. Bridging is accomplished by creating a bridge(4) device 405: and adding the dom0's physical interface and the various xvifN.0 406: interfaces to the bridge. One specifies "bridge=bridge0" in the domU 407: config file. The bridge must be set up already in the dom0; an 408: example /etc/ifconfig.bridge0 is: 409: 410: [[!template id=filecontent name="/etc/ifconfig.bridge0" text=""" 411: create 412: up 413: !brconfig bridge0 add wm0 414: """]] 415: 416: With NAT, the domU perceives itself to be behind a NAT running on the 417: dom0. This is often appropriate when running Xen on a workstation. 418: TODO: NAT appears to be configured by "vif = [ '' ]". 419: 420: The MAC address specified is the one used for the interface in the new 421: domain. The interface in dom0 will use this address XOR'd with 422: 00:00:00:01:00:00. Random MAC addresses are assigned if not given. 423: 424: Starting domains automatically 425: ------------------------------ 426: 427: To start domains `domU-netbsd` and `domU-linux` at boot and shut them 428: down cleanly on dom0 shutdown, add the following in rc.conf: 429: 430: [[!template id=filecontent name="/etc/rc.conf" text=""" 431: xendomains="domU-netbsd domU-linux" 432: """]] 433: 434: # Creating a domU 435: 436: Creating domUs is almost entirely independent of operating system. We 437: have already presented the basics of config files. Note that you must 438: have already completed the dom0 setup so that "xl list" works. 439: 440: Creating a NetBSD PV domU 441: -------------------------- 442: 443: See the earlier config file, and adjust memory. Decide on how much 444: storage you will provide, and prepare it (file or LVM). 445: 446: While the kernel will be obtained from the dom0 file system, the same 447: file should be present in the domU as /netbsd so that tools like 448: savecore(8) can work. (This is helpful but not necessary.) 449: 450: The kernel must be specifically for Xen and for use as a domU. The 451: i386 and amd64 provide the following kernels: 452: 453: i386 XEN3PAE_DOMU 454: amd64 XEN3_DOMU 455: 456: This will boot NetBSD, but this is not that useful if the disk is 457: empty. One approach is to unpack sets onto the disk outside of xen 458: (by mounting it, just as you would prepare a physical disk for a 459: system you can't run the installer on). 460: 461: A second approach is to run an INSTALL kernel, which has a miniroot 462: and can load sets from the network. To do this, copy the INSTALL 463: kernel to / and change the kernel line in the config file to: 464: 465: kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU" 466: 467: Then, start the domain as "xl create -c configfile". 468: 469: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following 470: line should be used in the config file. 471: 472: disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ] 473: 474: After booting the domain, the option to install via CDROM may be 475: selected. The CDROM device should be changed to `xbd1d`. 476: 477: Once done installing, "halt -p" the new domain (don't reboot or halt, 478: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the 479: config file), switch the config file back to the XEN3_DOMU kernel, 480: and start the new domain again. Now it should be able to use "root on 481: xbd0a" and you should have a, functional NetBSD domU. 482: 483: TODO: check if this is still accurate. 484: When the new domain is booting you'll see some warnings about *wscons* 485: and the pseudo-terminals. These can be fixed by editing the files 486: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in 487: `/etc/ttys`, except *console*, like this: 488: 489: console "/usr/libexec/getty Pc" vt100 on secure 490: ttyE0 "/usr/libexec/getty Pc" vt220 off secure 491: ttyE1 "/usr/libexec/getty Pc" vt220 off secure 492: ttyE2 "/usr/libexec/getty Pc" vt220 off secure 493: ttyE3 "/usr/libexec/getty Pc" vt220 off secure 494: 495: Finally, all screens must be commented out from `/etc/wscons.conf`. 496: 497: It is also desirable to add 498: 499: powerd=YES 500: 501: in rc.conf. This way, the domain will be properly shut down if 502: `xm shutdown -R` or `xm shutdown -H` is used on the dom0. 503: 504: It is not strictly necessary to have a kernel (as /netbsd) in the domU 505: file system. However, various programs (e.g. netstat) will use that 506: kernel to look up symbols to read from kernel virtual memory. If 507: /netbsd is not the running kernel, those lookups will fail. (This is 508: not really a Xen-specific issue, but because the domU kernel is 509: obtained from the dom0, it is far more likely to be out of sync or 510: missing with Xen.) 511: 512: Creating a Linux domU 513: --------------------- 514: 515: Creating unprivileged Linux domains isn't much different from 516: unprivileged NetBSD domains, but there are some details to know. 517: 518: First, the second parameter passed to the disk declaration (the '0x1' in 519: the example below) 520: 521: disk = [ 'phy:/dev/wd0e,0x1,w' ] 522: 523: does matter to Linux. It wants a Linux device number here (e.g. 0x300 524: for hda). Linux builds device numbers as: (major \<\< 8 + minor). 525: So, hda1 which has major 3 and minor 1 on a Linux system will have 526: device number 0x301. Alternatively, devices names can be used (hda, 527: hdb, ...) as xentools has a table to map these names to devices 528: numbers. To export a partition to a Linux guest we can use: 529: 530: disk = [ 'phy:/dev/wd0e,0x300,w' ] 531: root = "/dev/hda1 ro" 532: 533: and it will appear as /dev/hda on the Linux system, and be used as root 534: partition. 535: 536: To install the Linux system on the partition to be exported to the 537: guest domain, the following method can be used: install 538: sysutils/e2fsprogs from pkgsrc. Use mke2fs to format the partition 539: that will be the root partition of your Linux domain, and mount it. 540: Then copy the files from a working Linux system, make adjustments in 541: `/etc` (fstab, network config). It should also be possible to extract 542: binary packages such as .rpm or .deb directly to the mounted partition 543: using the appropriate tool, possibly running under NetBSD's Linux 544: emulation. Once the file system has been populated, umount it. If 545: desirable, the file system can be converted to ext3 using tune2fs -j. 546: It should now be possible to boot the Linux guest domain, using one of 547: the vmlinuz-\*-xenU kernels available in the Xen binary distribution. 548: 549: To get the Linux console right, you need to add: 550: 551: extra = "xencons=tty1" 552: 553: to your configuration since not all Linux distributions auto-attach a 554: tty to the xen console. 555: 556: ## Creating a NetBSD HVM domU 557: 558: Use type='hmv', probably. Use a GENERIC kernel within the disk image. 559: 560: ## Creating a NetBSD PVH domU 561: 562: Use type='pvh'. 563: 564: \todo Explain where the kernel comes from. 565: 566: 567: Creating a Solaris domU 568: ----------------------- 569: 570: See possibly outdated 571: [Solaris domU instructions](/ports/xen/howto-solaris/). 572: 573: 574: PCI passthrough: Using PCI devices in guest domains 575: --------------------------------------------------- 576: 577: NB: PCI passthrough only works on some Xen versions and as of 2020 it 578: is not clear that it works on any version in pkgsrc. Reports 579: confirming or denying this notion should be sent to port-xen@. 580: 581: The dom0 can give other domains access to selected PCI 582: devices. This can allow, for example, a non-privileged domain to have 583: access to a physical network interface or disk controller. However, 584: keep in mind that giving a domain access to a PCI device most likely 585: will give the domain read/write access to the whole physical memory, 586: as PCs don't have an IOMMU to restrict memory access to DMA-capable 587: device. Also, it's not possible to export ISA devices to non-dom0 588: domains, which means that the primary VGA adapter can't be exported. 589: A guest domain trying to access the VGA registers will panic. 590: 591: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has 592: not been ported to later versions at this time. 593: 594: For a PCI device to be exported to a domU, is has to be attached to 595: the "pciback" driver in dom0. Devices passed to the dom0 via the 596: pciback.hide boot parameter will attach to "pciback" instead of the 597: usual driver. The list of devices is specified as "(bus:dev.func)", 598: where bus and dev are 2-digit hexadecimal numbers, and func a 599: single-digit number: 600: 601: pciback.hide=(00:0a.0)(00:06.0) 602: 603: pciback devices should show up in the dom0's boot messages, and the 604: devices should be listed in the `/kern/xen/pci` directory. 605: 606: PCI devices to be exported to a domU are listed in the "pci" array of 607: the domU's config file, with the format "0000:bus:dev.func". 608: 609: pci = [ '0000:00:06.0', '0000:00:0a.0' ] 610: 611: In the domU an "xpci" device will show up, to which one or more pci 612: buses will attach. Then the PCI drivers will attach to PCI buses as 613: usual. Note that the default NetBSD DOMU kernels do not have "xpci" 614: or any PCI drivers built in by default; you have to build your own 615: kernel to use PCI devices in a domU. Here's a kernel config example; 616: note that only the "xpci" lines are unusual. 617: 618: include "arch/i386/conf/XEN3_DOMU" 619: 620: # Add support for PCI buses to the XEN3_DOMU kernel 621: xpci* at xenbus ? 622: pci* at xpci ? 623: 624: # PCI USB controllers 625: uhci* at pci? dev ? function ? # Universal Host Controller (Intel) 626: 627: # USB bus support 628: usb* at uhci? 629: 630: # USB Hubs 631: uhub* at usb? 632: uhub* at uhub? port ? configuration ? interface ? 633: 634: # USB Mass Storage 635: umass* at uhub? port ? configuration ? interface ? 636: wd* at umass? 637: # SCSI controllers 638: ahc* at pci? dev ? function ? # Adaptec [23]94x, aic78x0 SCSI 639: 640: # SCSI bus support (for both ahc and umass) 641: scsibus* at scsi? 642: 643: # SCSI devices 644: sd* at scsibus? target ? lun ? # SCSI disk drives 645: cd* at scsibus? target ? lun ? # SCSI CD-ROM drives 646: 647: 648: # Specific Issues 649: 650: ## domU 651: 652: [NetBSD 5 is known to panic.](http://mail-index.netbsd.org/port-xen/2018/04/17/msg009181.html) 653: (However, NetBSD 5 systems should be updated to a supported version.) 654: 655: # NetBSD as a domU in a VPS 656: 657: The bulk of the HOWTO is about using NetBSD as a dom0 on your own 658: hardware. This section explains how to deal with Xen in a domU as a 659: virtual private server where you do not control or have access to the 660: dom0. This is not intended to be an exhaustive list of VPS providers; 661: only a few are mentioned that specifically support NetBSD. 662: 663: VPS operators provide varying degrees of access and mechanisms for 664: configuration. The big issue is usually how one controls which kernel 665: is booted, because the kernel is nominally in the dom0 file system (to 666: which VPS users do not normally have access). A second issue is how 667: to install NetBSD. 668: A VPS user may want to compile a kernel for security updates, to run 669: npf, run IPsec, or any other reason why someone would want to change 670: their kernel. 671: 672: One approach is to have an administrative interface to upload a kernel, 673: or to select from a prepopulated list. Other approaches are pygrub 674: (deprecated) and pvgrub, which are ways to have a bootloader obtain a 675: kernel from the domU file system. This is closer to a regular physical 676: computer, where someone who controls a machine can replace the kernel. 677: 678: A second issue is multiple CPUs. With NetBSD 6, domUs support 679: multiple vcpus, and it is typical for VPS providers to enable multiple 680: CPUs for NetBSD domUs. 681: 682: ## Complexities due to Xen changes 683: 684: Xen has many security advisories and people running Xen systems make 685: different choices. 686: 687: ### stub domains 688: 689: Some (Linux only?) dom0 systems use something called "stub domains" to 690: isolate qemu from the dom0 system, as a security and reliabilty 691: mechanism when running HVM domUs. Somehow, NetBSD's GENERIC kernel 692: ends up using PIO for disks rather than DMA. Of course, all of this 693: is emulated, but emulated PIO is unusably slow. This problem is not 694: currently understood. 695: 696: ### Grant tables 697: 698: There are multiple versions of using grant tables, and some security 699: advisories have suggested disabling some versions. Some versions of 700: NetBSD apparently only use specific versions and this can lead to 701: "NetBSD current doesn't run on hosting provider X" situations. 702: 703: \todo Explain better. 704: 705: pvgrub 706: ------ 707: 708: pvgrub is a version of grub that uses PV operations instead of BIOS 709: calls. It is booted from the dom0 as the domU kernel, and then reads 710: /grub/menu.lst and loads a kernel from the domU file system. 711: 712: [Panix](http://www.panix.com/) lets users use pvgrub. Panix reports 713: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes 714: (and hence with defaults from "newfs -O 2"). See [Panix's pvgrub 715: page](http://www.panix.com/v-colo/grub.html), which describes only 716: Linux but should be updated to cover NetBSD :-). 717: 718: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot 719: their own kernel. See then [prgmr.com NetBSD 720: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU) 721: (which is in need of updating). 722: 723: It appears that [grub's FFS 724: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c) 725: does not support all aspects of modern FFS, but there are also reports 726: that FFSv2 works fine. At prgmr, typically one has an ext2 or FAT 727: partition for the kernel with the intent that grub can understand it, 728: which leads to /netbsd not being the actual kernel. One must remember 729: to update the special boot partition. 730: 731: pygrub 732: ------- 733: 734: pygrub runs in the dom0 and looks into the domU file system. This 735: implies that the domU must have a kernel in a file system in a format 736: known to pygrub. 737: 738: pygrub doesn't seem to work to load Linux images under NetBSD dom0, 739: and is inherently less secure than pvgrub due to running inside dom0. For both these 740: reasons, pygrub should not be used, and is only still present so that 741: historical DomU images using it still work. 742: 743: As of 2014, pygrub seems to be of mostly historical 744: interest. New DomUs should use pvgrub. 745: 746: Amazon 747: ------ 748: 749: See the [Amazon EC2 page](/amazon_ec2/).