wikisrc/ports/xen/howto.mdwn - view

File: [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.185: download - view: text, annotated - select for diffs
Sun Nov 15 14:59:54 2020 UTC (5 months, 3 weeks ago) by gdt
Branches: MAIN
CVS tags: HEAD

xen: move netbsd info to netbsd section

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: This HOWTO presumes a basic familiarity with the Xen system 12: architecture, with installing NetBSD on amd64 hardware, and with 13: installing software from pkgsrc. See also the [Xen 14: website](http://www.xenproject.org/). 15: 16: [[!toc]] 17: 18: # Overview 19: 20: The basic concept of Xen is that the hypervisor (xenkernel) runs on 21: the hardware, and runs a privileged domain ("dom0") that can access 22: disks/networking/etc. One then runs additonal unprivileged domains 23: (each a "domU"), presumably to do something useful. 24: 25: This HOWTO addresses how to run a NetBSD dom0 (and hence also build 26: xen itself). It also addresses how to run domUs in that environment, 27: and how to deal with having a domU in a Xen environment run by someone 28: else and/or not running NetBSD. 29: 30: ## Guest Styles 31: 32: Xen supports different styles of guests. 33: 34: [[!table data=""" 35: Style of guest |Supported by NetBSD 36: PV |Yes (dom0, domU) 37: HVM |Yes (domU) 38: PVHVM |current-only (domU) 39: PVH |current-only (domU, dom0 not yet) 40: """]] 41: 42: In Para-Virtualized (PV) mode, the guest OS does not attempt to access 43: hardware directly, but instead makes hypercalls to the hypervisor; PV 44: guests must be specifically coded for Xen. 45: See [PV](https://wiki.xen.org/wiki/Paravirtualization_(PV\)). 46: 47: In HVM mode, no guest modification is required; however, hardware 48: support is required, such as VT-x on Intel CPUs and SVM on AMD CPUs. 49: The dom0 runs qemu to emulate hardware. 50: 51: In PVHVM mode, the guest runs as HVM, but additionally can use PV 52: drivers for efficiency. 53: See [PV on HVM](https://wiki.xen.org/wiki/PV_on_HVM). 54: 55: There have been two PVH modes: original PVH and PVHv2. Original PVH 56: was based on PV mode and is no longer relevant at all. PVHv2 is 57: basically lightweight HVM with PV drivers. A critical feature of it 58: is that qemu is not needed; the hypervisor can do the emulation that 59: is required. Thus, a dom0 can be PVHv2. 60: The source code uses PVH and config files use pvh; this refers to PVHv2. 61: See [PVH(v2)](https://wiki.xenproject.org/wiki/PVH_(v2\)_Domu). 62: 63: At system boot, the dom0 kernel is loaded as a module with Xen as the kernel. 64: The dom0 can start one or more domUs. (Booting is explained in detail 65: in the dom0 section.) 66: 67: ## CPU Architecture 68: 69: Xen runs on x86_64 hardware (the NetBSD amd64 port). 70: 71: There is a concept of Xen running on ARM, but there are no reports of this working with NetBSD. 72: 73: The dom0 system should be amd64. (Instructions for i386PAE dom0 have been removed from the HOWTO.) 74: 75: The domU can be i386PAE or amd64. 76: i386PAE at one point was considered as [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html) than amd64. 77: 78: ## Xen Versions 79: 80: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages 81: xenkernel and xentools. We will refer only to the kernel versions, 82: but note that both packages must be installed together and must have 83: matching versions. 84: 85: Versions available in pkgsrc: 86: 87: [[!table data=""" 88: Xen Version |Package Name |Xen CPU Support |EOL'ed By Upstream 89: 4.11 |xenkernel411 |x86_64 |No 90: 4.13 |xenkernel413 |x86_64 |No 91: """]] 92: 93: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/). 94: 95: Older Xen had a python-based management tool called xm, now replaced 96: by xl. 97: 98: ## NetBSD versions 99: 100: Xen has been supported in NetBSD for a long time, at least since 2005. 101: Initially Xen was PV only. 102: 103: NetBSD 8 and up support PV and HVM modes. 104: 105: Support for PVHVM and PVH is available only in NetBSD-current. 106: 107: NetBSD as a dom0 does not run SMP, because some drivers are not yet 108: safe for this. \todo Link to more information about what needs work. 109: 110: NetBSD, when run as a domU, can and does typically run SMP. 111: 112: Note: NetBSD support is called XEN3. However, it does support Xen 4, 113: because the hypercall interface has remained identical. 114: 115: # Creating a dom0 116: 117: In order to install a NetBSD as a dom0, one must first install a normal 118: NetBSD system, and then pivot the install to a dom0 install by changing 119: the kernel and boot configuration. 120: 121: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with 122: 512M it worked reliably. This does not make sense, but if you see 123: "not ELF" after Xen boots, try increasing dom0 RAM. 124: 125: ## Installation of NetBSD 126: 127: [Install NetBSD/amd64](/guide/inst/) 128: just as you would if you were not using Xen. 129: 130: ## Installation of Xen 131: 132: We will consider that you chose to use Xen 4.13, with NetBSD/amd64 as 133: dom0. In the dom0, install xenkernel48 and xentools48 from pkgsrc. 134: 135: Once this is done, install the Xen kernel itself: 136: 137: [[!template id=programlisting text=""" 138: # cp /usr/pkg/xen48-kernel/xen.gz / 139: """]] 140: 141: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory. Such kernel 142: can either be compiled manually, or downloaded from the NetBSD FTP, for 143: example at: 144: 145: [[!template id=programlisting text=""" 146: ftp.netbsd.org/pub/NetBSD/NetBSD-8.0/amd64/binary/kernel/netbsd-XEN3_DOM0.gz 147: """]] 148: 149: Add a line to /boot.cfg to boot Xen: 150: 151: [[!template id=filecontent name="/boot.cfg" text=""" 152: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M 153: """]] 154: 155: This specifies that the dom0 should have 512MB of ram, leaving the rest 156: to be allocated for domUs. To use a serial console, use: 157: 158: [[!template id=filecontent name="/boot.cfg" text=""" 159: menu=Xen:load /netbsd-XEN3_DOM0.gz;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1 160: """]] 161: 162: which will use the first serial port for Xen (which counts starting 163: from 1, unlike NetBSD which counts starting from 0), forcing 164: speed/parity. Because the NetBSD command line lacks a 165: "console=pc" argument, it will use the default "xencons" console device, 166: which directs the console I/O through Xen to the same console device Xen 167: itself uses (in this case, the serial port). 168: 169: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`, 170: to force only one vcpu to be provided (since NetBSD dom0 can't use 171: more) and to pin that vcpu to a physical CPU. Xen has 172: [many boot options](http://xenbits.xenproject.org/docs/4.13-testing/misc/xen-command-line.html), 173: and other than dom0 memory and max_vcpus, they are generally not 174: necessary. 175: 176: Copy the boot scripts into `/etc/rc.d`: 177: 178: [[!template id=programlisting text=""" 179: # cp /usr/pkg/share/examples/rc.d/xen* /etc/rc.d/ 180: """]] 181: 182: Enable `xencommons`: 183: 184: [[!template id=filecontent name="/etc/rc.conf" text=""" 185: xencommons=YES 186: """]] 187: 188: Now, reboot so that you are running a DOM0 kernel under Xen, rather 189: than GENERIC without Xen. 190: 191: TODO: Recommend for/against xen-watchdog. 192: 193: Once the reboot is done, use `xl` to inspect Xen's boot messages, 194: available resources, and running domains. For example: 195: 196: [[!template id=programlisting text=""" 197: # xl dmesg 198: ... xen's boot info ... 199: # xl info 200: ... available memory, etc ... 201: # xl list 202: Name Id Mem(MB) CPU State Time(s) Console 203: Domain-0 0 64 0 r---- 58.1 204: """]] 205: 206: Xen logs will be in /var/log/xen. 207: 208: ### Issues with xencommons 209: 210: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and 211: domUs. It does not currently work to stop and start xenstored. 212: Certainly all domUs should be shutdown first, following the sort order 213: of the rc.d scripts. However, the dom0 sets up state with xenstored, 214: and is not notified when xenstored exits, leading to not recreating 215: the state when the new xenstored starts. Until there's a mechanism to 216: make this work, one should not expect to be able to restart xenstored 217: (and thus xencommons). There is currently no reason to expect that 218: this will get fixed any time soon. 219: 220: ## anita (for testing NetBSD) 221: 222: With the setup so far, one should be able to run 223: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as 224: root, because anita must create a domU): 225: 226: [[!template id=programlisting text=""" 227: anita --vmm=xl test file:///usr/obj/i386/ 228: """]] 229: 230: ## Xen-specific NetBSD issues 231: 232: There are (at least) two additional things different about NetBSD as a 233: dom0 kernel compared to hardware. 234: 235: One is that the module ABI is different because some of the #defines 236: change, so one must build modules for Xen. As of netbsd-7, the build 237: system does this automatically. 238: 239: The other difference is that XEN3_DOM0 does not have exactly the same 240: options as GENERIC. While it is debatable whether or not this is a 241: bug, users should be aware of this and can simply add missing config 242: items if desired. 243: 244: ## Updating NetBSD in a dom0 245: 246: This is just like updating NetBSD on bare hardware, assuming the new 247: version supports the version of Xen you are running. Generally, one 248: replaces the kernel and reboots, and then overlays userland binaries 249: and adjusts `/etc`. 250: 251: Note that one must update both the non-Xen kernel typically used for 252: rescue purposes and the DOM0 kernel used with Xen. 253: 254: ## Converting from grub to /boot 255: 256: These instructions were used to convert a system from 257: grub to /boot. The system was originally installed in February of 258: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated 259: over time. Before these commands, it was running NetBSD 6 i386, Xen 260: 4.1 and grub, much like the message linked earlier in the grub 261: section. 262: 263: [[!template id=programlisting text=""" 264: # Install MBR bootblocks on both disks. 265: fdisk -i /dev/rwd0d 266: fdisk -i /dev/rwd1d 267: # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components. 268: installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1 269: installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1 270: # Install secondary boot loader 271: cp -p /usr/mdec/boot / 272: # Create boot.cfg following earlier guidance: 273: menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M 274: menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=512M 275: menu=GENERIC:boot 276: menu=GENERIC single-user:boot -s 277: menu=GENERIC.ok:boot netbsd.ok 278: menu=GENERIC.ok single-user:boot netbsd.ok -s 279: menu=Drop to boot prompt:prompt 280: default=1 281: timeout=30 282: """]] 283: 284: ## Upgrading Xen versions 285: 286: Minor version upgrades are trivial. Just rebuild/replace the 287: xenkernel version and copy the new xen.gz to `/` (where `/boot.cfg` 288: references it), and reboot. 289: 290: #Unprivileged domains (domU) 291: 292: This section describes general concepts about domUs. It does not 293: address specific domU operating systems or how to install them. The 294: config files for domUs are typically in `/usr/pkg/etc/xen`, and are 295: typically named so that the file name, domU name and the domU's host 296: name match. 297: 298: The domU is provided with CPU and memory by Xen, configured by the 299: dom0. The domU is provided with disk and network by the dom0, 300: mediated by Xen, and configured in the dom0. 301: 302: Entropy in domUs can be an issue; physical disks and network are on 303: the dom0. NetBSD's /dev/random system works, but is often challenged. 304: 305: ## Config files 306: 307: See /usr/pkg/share/examples/xen/xlexample* 308: for a small number of well-commented examples, mostly for running 309: GNU/Linux. 310: 311: The following is an example minimal domain configuration file. The domU 312: serves as a network file server. 313: 314: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text=""" 315: name = "domU-id" 316: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz" 317: memory = 1024 318: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ] 319: disk = [ 'file:/n0/xen/foo-wd0,0x0,w', 320: 'file:/n0/xen/foo-wd1,0x1,w' ] 321: """]] 322: 323: The domain will have name given in the `name` setting. The kernel has the 324: host/domU name in it, so that on the dom0 one can update the various 325: domUs independently. The `vif` line causes an interface to be provided, 326: with a specific mac address (do not reuse MAC addresses!), in bridge 327: mode. Two disks are provided, and they are both writable; the bits 328: are stored in files and Xen attaches them to a vnd(4) device in the 329: dom0 on domain creation. The system treats xbd0 as the boot device 330: without needing explicit configuration. 331: 332: By convention, domain config files are kept in `/usr/pkg/etc/xen`. Note 333: that "xl create" takes the name of a config file, while other commands 334: take the name of a domain. 335: 336: Examples of commands: 337: 338: [[!template id=programlisting text=""" 339: xl create /usr/pkg/etc/xen/foo 340: xl console domU-id 341: xl create -c /usr/pkg/etc/xen/foo 342: xl shutdown domU-id 343: xl list 344: """]] 345: 346: Typing `^]` will exit the console session. Shutting down a domain is 347: equivalent to pushing the power button; a NetBSD domU will receive a 348: power-press event and do a clean shutdown. Shutting down the dom0 349: will trigger controlled shutdowns of all configured domUs. 350: 351: ## CPU and memory 352: 353: A domain is provided with some number of vcpus, up to the number 354: of CPUs seen by the hypervisor. For a domU, it is controlled 355: from the config file by the "vcpus = N" directive. 356: 357: A domain is provided with memory; this is controlled in the config 358: file by "memory = N" (in megabytes). In the straightforward case, the 359: sum of the the memory allocated to the dom0 and all domUs must be less 360: than the available memory. 361: 362: Xen also provides a "balloon" driver, which can be used to let domains 363: use more memory temporarily. 364: 365: ## Virtual disks 366: 367: In domU config files, the disks are defined as a sequence of 3-tuples: 368: 369: * The first element is "method:/path/to/disk". Common methods are 370: "file:" for a file-backed vnd, and "phy:" for something that is already 371: a device, such as an LVM logical volume. 372: 373: * The second element is an artifact of how virtual disks are passed to 374: Linux, and a source of confusion with NetBSD Xen usage. Linux domUs 375: are given a device name to associate with the disk, and values like 376: "hda1" or "sda1" are common. In a NetBSD domU, the first disk appears 377: as xbd0, the second as xbd1, and so on. However, xl demands a 378: second argument. The name given is converted to a major/minor by 379: calling stat(2) on the name in /dev and this is passed to the domU. 380: In the general case, the dom0 and domU can be different operating 381: systems, and it is an unwarranted assumption that they have consistent 382: numbering in /dev, or even that the dom0 OS has a /dev. With NetBSD 383: as both dom0 and domU, using values of 0x0 for the first disk and 0x1 384: for the second works fine and avoids this issue. For a GNU/Linux 385: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for 386: /dev/hda1. 387: 388: * The third element is "w" for writable disks, and "r" for read-only 389: disks. 390: 391: Example: 392: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text=""" 393: disk = [ 'file:/n0/xen/foo-wd0,0x0,w' ] 394: """]] 395: 396: Note that NetBSD by default creates only vnd[0123]. If you need more 397: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the 398: dom0. 399: 400: Note that NetBSD by default creates only xbd[0123]. If you need more 401: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU. 402: 403: Virtual Networking 404: ------------------ 405: 406: Xen provides virtual Ethernets, each of which connects the dom0 and a 407: domU. For each virtual network, there is an interface "xvifN.M" in 408: the dom0, and a matching interface xennetM (NetBSD name) in domU index N. 409: The interfaces behave as if there is an Ethernet with two 410: adapters connected. From this primitive, one can construct various 411: configurations. We focus on two common and useful cases for which 412: there are existing scripts: bridging and NAT. 413: 414: With bridging (in the example above), the domU perceives itself to be 415: on the same network as the dom0. For server virtualization, this is 416: usually best. Bridging is accomplished by creating a bridge(4) device 417: and adding the dom0's physical interface and the various xvifN.0 418: interfaces to the bridge. One specifies "bridge=bridge0" in the domU 419: config file. The bridge must be set up already in the dom0; an 420: example /etc/ifconfig.bridge0 is: 421: 422: [[!template id=filecontent name="/etc/ifconfig.bridge0" text=""" 423: create 424: up 425: !brconfig bridge0 add wm0 426: """]] 427: 428: With NAT, the domU perceives itself to be behind a NAT running on the 429: dom0. This is often appropriate when running Xen on a workstation. 430: TODO: NAT appears to be configured by "vif = [ '' ]". 431: 432: The MAC address specified is the one used for the interface in the new 433: domain. The interface in dom0 will use this address XOR'd with 434: 00:00:00:01:00:00. Random MAC addresses are assigned if not given. 435: 436: Starting domains automatically 437: ------------------------------ 438: 439: To start domains `domU-netbsd` and `domU-linux` at boot and shut them 440: down cleanly on dom0 shutdown, add the following in rc.conf: 441: 442: [[!template id=filecontent name="/etc/rc.conf" text=""" 443: xendomains="domU-netbsd domU-linux" 444: """]] 445: 446: # Creating a domU 447: 448: Creating domUs is almost entirely independent of operating system. We 449: have already presented the basics of config files. Note that you must 450: have already completed the dom0 setup so that "xl list" works. 451: 452: Creating a NetBSD PV domU 453: -------------------------- 454: 455: See the earlier config file, and adjust memory. Decide on how much 456: storage you will provide, and prepare it (file or LVM). 457: 458: While the kernel will be obtained from the dom0 file system, the same 459: file should be present in the domU as /netbsd so that tools like 460: savecore(8) can work. (This is helpful but not necessary.) 461: 462: The kernel must be specifically for Xen and for use as a domU. The 463: i386 and amd64 provide the following kernels: 464: 465: i386 XEN3PAE_DOMU 466: amd64 XEN3_DOMU 467: 468: This will boot NetBSD, but this is not that useful if the disk is 469: empty. One approach is to unpack sets onto the disk outside of xen 470: (by mounting it, just as you would prepare a physical disk for a 471: system you can't run the installer on). 472: 473: A second approach is to run an INSTALL kernel, which has a miniroot 474: and can load sets from the network. To do this, copy the INSTALL 475: kernel to / and change the kernel line in the config file to: 476: 477: kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU" 478: 479: Then, start the domain as "xl create -c configfile". 480: 481: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following 482: line should be used in the config file. 483: 484: disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ] 485: 486: After booting the domain, the option to install via CDROM may be 487: selected. The CDROM device should be changed to `xbd1d`. 488: 489: Once done installing, "halt -p" the new domain (don't reboot or halt, 490: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the 491: config file), switch the config file back to the XEN3_DOMU kernel, 492: and start the new domain again. Now it should be able to use "root on 493: xbd0a" and you should have a, functional NetBSD domU. 494: 495: TODO: check if this is still accurate. 496: When the new domain is booting you'll see some warnings about *wscons* 497: and the pseudo-terminals. These can be fixed by editing the files 498: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in 499: `/etc/ttys`, except *console*, like this: 500: 501: console "/usr/libexec/getty Pc" vt100 on secure 502: ttyE0 "/usr/libexec/getty Pc" vt220 off secure 503: ttyE1 "/usr/libexec/getty Pc" vt220 off secure 504: ttyE2 "/usr/libexec/getty Pc" vt220 off secure 505: ttyE3 "/usr/libexec/getty Pc" vt220 off secure 506: 507: Finally, all screens must be commented out from `/etc/wscons.conf`. 508: 509: It is also desirable to add 510: 511: powerd=YES 512: 513: in rc.conf. This way, the domain will be properly shut down if 514: `xl shutdown -R` or `xl shutdown -H` is used on the dom0. 515: \todo Check the translation to xl. 516: 517: It is not strictly necessary to have a kernel (as /netbsd) in the domU 518: file system. However, various programs (e.g. netstat) will use that 519: kernel to look up symbols to read from kernel virtual memory. If 520: /netbsd is not the running kernel, those lookups will fail. (This is 521: not really a Xen-specific issue, but because the domU kernel is 522: obtained from the dom0, it is far more likely to be out of sync or 523: missing with Xen.) 524: 525: Creating a Linux domU 526: --------------------- 527: 528: Creating unprivileged Linux domains isn't much different from 529: unprivileged NetBSD domains, but there are some details to know. 530: 531: First, the second parameter passed to the disk declaration (the '0x1' in 532: the example below) 533: 534: disk = [ 'phy:/dev/wd0e,0x1,w' ] 535: 536: does matter to Linux. It wants a Linux device number here (e.g. 0x300 537: for hda). Linux builds device numbers as: (major \<\< 8 + minor). 538: So, hda1 which has major 3 and minor 1 on a Linux system will have 539: device number 0x301. Alternatively, devices names can be used (hda, 540: hdb, ...) as xentools has a table to map these names to devices 541: numbers. To export a partition to a Linux guest we can use: 542: 543: disk = [ 'phy:/dev/wd0e,0x300,w' ] 544: root = "/dev/hda1 ro" 545: 546: and it will appear as /dev/hda on the Linux system, and be used as root 547: partition. 548: 549: To install the Linux system on the partition to be exported to the 550: guest domain, the following method can be used: install 551: sysutils/e2fsprogs from pkgsrc. Use mke2fs to format the partition 552: that will be the root partition of your Linux domain, and mount it. 553: Then copy the files from a working Linux system, make adjustments in 554: `/etc` (fstab, network config). It should also be possible to extract 555: binary packages such as .rpm or .deb directly to the mounted partition 556: using the appropriate tool, possibly running under NetBSD's Linux 557: emulation. Once the file system has been populated, umount it. If 558: desirable, the file system can be converted to ext3 using tune2fs -j. 559: It should now be possible to boot the Linux guest domain, using one of 560: the vmlinuz-\*-xenU kernels available in the Xen binary distribution. 561: 562: To get the Linux console right, you need to add: 563: 564: extra = "xencons=tty1" 565: 566: to your configuration since not all Linux distributions auto-attach a 567: tty to the xen console. 568: 569: ## Creating a NetBSD HVM domU 570: 571: Use type='hmv', probably. Use a GENERIC kernel within the disk image. 572: 573: ## Creating a NetBSD PVH domU 574: 575: Use type='pvh'. 576: 577: \todo Explain where the kernel comes from. 578: 579: 580: Creating a Solaris domU 581: ----------------------- 582: 583: See possibly outdated 584: [Solaris domU instructions](/ports/xen/howto-solaris/). 585: 586: 587: PCI passthrough: Using PCI devices in guest domains 588: --------------------------------------------------- 589: 590: NB: PCI passthrough only works on some Xen versions and as of 2020 it 591: is not clear that it works on any version in pkgsrc. Reports 592: confirming or denying this notion should be sent to port-xen@. 593: 594: The dom0 can give other domains access to selected PCI 595: devices. This can allow, for example, a non-privileged domain to have 596: access to a physical network interface or disk controller. However, 597: keep in mind that giving a domain access to a PCI device most likely 598: will give the domain read/write access to the whole physical memory, 599: as PCs don't have an IOMMU to restrict memory access to DMA-capable 600: device. Also, it's not possible to export ISA devices to non-dom0 601: domains, which means that the primary VGA adapter can't be exported. 602: A guest domain trying to access the VGA registers will panic. 603: 604: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has 605: not been ported to later versions at this time. 606: 607: For a PCI device to be exported to a domU, is has to be attached to 608: the "pciback" driver in dom0. Devices passed to the dom0 via the 609: pciback.hide boot parameter will attach to "pciback" instead of the 610: usual driver. The list of devices is specified as "(bus:dev.func)", 611: where bus and dev are 2-digit hexadecimal numbers, and func a 612: single-digit number: 613: 614: pciback.hide=(00:0a.0)(00:06.0) 615: 616: pciback devices should show up in the dom0's boot messages, and the 617: devices should be listed in the `/kern/xen/pci` directory. 618: 619: PCI devices to be exported to a domU are listed in the "pci" array of 620: the domU's config file, with the format "0000:bus:dev.func". 621: 622: pci = [ '0000:00:06.0', '0000:00:0a.0' ] 623: 624: In the domU an "xpci" device will show up, to which one or more pci 625: buses will attach. Then the PCI drivers will attach to PCI buses as 626: usual. Note that the default NetBSD DOMU kernels do not have "xpci" 627: or any PCI drivers built in by default; you have to build your own 628: kernel to use PCI devices in a domU. Here's a kernel config example; 629: note that only the "xpci" lines are unusual. 630: 631: include "arch/i386/conf/XEN3_DOMU" 632: 633: # Add support for PCI buses to the XEN3_DOMU kernel 634: xpci* at xenbus ? 635: pci* at xpci ? 636: 637: # PCI USB controllers 638: uhci* at pci? dev ? function ? # Universal Host Controller (Intel) 639: 640: # USB bus support 641: usb* at uhci? 642: 643: # USB Hubs 644: uhub* at usb? 645: uhub* at uhub? port ? configuration ? interface ? 646: 647: # USB Mass Storage 648: umass* at uhub? port ? configuration ? interface ? 649: wd* at umass? 650: # SCSI controllers 651: ahc* at pci? dev ? function ? # Adaptec [23]94x, aic78x0 SCSI 652: 653: # SCSI bus support (for both ahc and umass) 654: scsibus* at scsi? 655: 656: # SCSI devices 657: sd* at scsibus? target ? lun ? # SCSI disk drives 658: cd* at scsibus? target ? lun ? # SCSI CD-ROM drives 659: 660: 661: # Specific Issues 662: 663: ## domU 664: 665: [NetBSD 5 is known to panic.](http://mail-index.netbsd.org/port-xen/2018/04/17/msg009181.html) 666: (However, NetBSD 5 systems should be updated to a supported version.) 667: 668: # NetBSD as a domU in a VPS 669: 670: The bulk of the HOWTO is about using NetBSD as a dom0 on your own 671: hardware. This section explains how to deal with Xen in a domU as a 672: virtual private server where you do not control or have access to the 673: dom0. This is not intended to be an exhaustive list of VPS providers; 674: only a few are mentioned that specifically support NetBSD. 675: 676: VPS operators provide varying degrees of access and mechanisms for 677: configuration. The big issue is usually how one controls which kernel 678: is booted, because the kernel is nominally in the dom0 file system (to 679: which VPS users do not normally have access). A second issue is how 680: to install NetBSD. 681: A VPS user may want to compile a kernel for security updates, to run 682: npf, run IPsec, or any other reason why someone would want to change 683: their kernel. 684: 685: One approach is to have an administrative interface to upload a kernel, 686: or to select from a prepopulated list. Other approaches are pygrub 687: (deprecated) and pvgrub, which are ways to have a bootloader obtain a 688: kernel from the domU file system. This is closer to a regular physical 689: computer, where someone who controls a machine can replace the kernel. 690: 691: A second issue is multiple CPUs. With NetBSD 6, domUs support 692: multiple vcpus, and it is typical for VPS providers to enable multiple 693: CPUs for NetBSD domUs. 694: 695: ## Complexities due to Xen changes 696: 697: Xen has many security advisories and people running Xen systems make 698: different choices. 699: 700: ### stub domains 701: 702: Some (Linux only?) dom0 systems use something called "stub domains" to 703: isolate qemu from the dom0 system, as a security and reliabilty 704: mechanism when running HVM domUs. Somehow, NetBSD's GENERIC kernel 705: ends up using PIO for disks rather than DMA. Of course, all of this 706: is emulated, but emulated PIO is unusably slow. This problem is not 707: currently understood. 708: 709: ### Grant tables 710: 711: There are multiple versions of using grant tables, and some security 712: advisories have suggested disabling some versions. Some versions of 713: NetBSD apparently only use specific versions and this can lead to 714: "NetBSD current doesn't run on hosting provider X" situations. 715: 716: \todo Explain better. 717: 718: ## Boot methods 719: 720: ### pvgrub 721: 722: pvgrub is a version of grub that uses PV operations instead of BIOS 723: calls. It is booted from the dom0 as the domU kernel, and then reads 724: /grub/menu.lst and loads a kernel from the domU file system. 725: 726: [Panix](http://www.panix.com/) lets users use pvgrub. Panix reports 727: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes 728: (and hence with defaults from "newfs -O 2"). See [Panix's pvgrub 729: page](http://www.panix.com/v-colo/grub.html), which describes only 730: Linux but should be updated to cover NetBSD :-). 731: 732: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot 733: their own kernel. See then [prgmr.com NetBSD 734: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU) 735: (which is in need of updating). 736: 737: It appears that [grub's FFS 738: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c) 739: does not support all aspects of modern FFS, but there are also reports 740: that FFSv2 works fine. At prgmr, typically one has an ext2 or FAT 741: partition for the kernel with the intent that grub can understand it, 742: which leads to /netbsd not being the actual kernel. One must remember 743: to update the special boot partition. 744: 745: ### pygrub 746: 747: pygrub runs in the dom0 and looks into the domU file system. This 748: implies that the domU must have a kernel in a file system in a format 749: known to pygrub. 750: 751: pygrub doesn't seem to work to load Linux images under NetBSD dom0, 752: and is inherently less secure than pvgrub due to running inside dom0. For both these 753: reasons, pygrub should not be used, and is only still present so that 754: historical DomU images using it still work. 755: 756: As of 2014, pygrub seems to be of mostly historical 757: interest. New DomUs should use pvgrub. 758: 759: ## Specific Providers 760: 761: ### Amazon 762: 763: See the [Amazon EC2 page](/amazon_ec2/).