1.\" 2.\" Copyright (c) 2006, 2007 3.\" The DragonFly Project. All rights reserved. 4.\" 5.\" Redistribution and use in source and binary forms, with or without 6.\" modification, are permitted provided that the following conditions 7.\" are met: 8.\" 9.\" 1. Redistributions of source code must retain the above copyright 10.\" notice, this list of conditions and the following disclaimer. 11.\" 2. Redistributions in binary form must reproduce the above copyright 12.\" notice, this list of conditions and the following disclaimer in 13.\" the documentation and/or other materials provided with the 14.\" distribution. 15.\" 3. Neither the name of The DragonFly Project nor the names of its 16.\" contributors may be used to endorse or promote products derived 17.\" from this software without specific, prior written permission. 18.\" 19.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 20.\" ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 21.\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS 22.\" FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 23.\" COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 24.\" INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING, 25.\" BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; 26.\" LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED 27.\" AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, 28.\" OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT 29.\" OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 30.\" SUCH DAMAGE. 31.\" 32.Dd March 6, 2013 33.Dt VKERNEL 7 34.Os 35.Sh NAME 36.Nm vkernel , 37.Nm vcd , 38.Nm vkd , 39.Nm vke 40.Nd virtual kernel architecture 41.Sh SYNOPSIS 42.Cd "platform vkernel # for 32 bit vkernels" 43.Cd "platform vkernel64 # for 64 bit vkernels" 44.Cd "device vcd" 45.Cd "device vkd" 46.Cd "device vke" 47.Pp 48.Pa /var/vkernel/boot/kernel/kernel 49.Op Fl hsUv 50.Op Fl c Ar file 51.Op Fl e Ar name Ns = Ns Li value : Ns Ar name Ns = Ns Li value : Ns ... 52.Op Fl i Ar file 53.Op Fl I Ar interface Ns Op Ar :address1 Ns Oo Ar :address2 Oc Ns Oo Ar /netmask Oc Ns Oo Ar =mac Oc 54.Op Fl l Ar cpulock 55.Op Fl m Ar size 56.Op Fl n Ar numcpus Ns Op Ar :lbits Ns Oo Ar :cbits Oc 57.Op Fl p Ar pidfile 58.Op Fl r Ar file Ns Op Ar :serno 59.Sh DESCRIPTION 60The 61.Nm 62architecture allows for running 63.Dx 64kernels in userland. 65.Pp 66The following options are available: 67.Bl -tag -width ".Fl m Ar size" 68.It Fl c Ar file 69Specify a readonly CD-ROM image 70.Ar file 71to be used by the kernel, with the first 72.Fl c 73option defining 74.Li vcd0 , 75the second one 76.Li vcd1 , 77and so on. 78The first 79.Fl r 80or 81.Fl c 82option specified on the command line will be the boot disk. 83The CD9660 filesystem is assumed when booting from this media. 84.It Fl e Ar name Ns = Ns Li value : Ns Ar name Ns = Ns Li value : Ns ... 85Specify an environment to be used by the kernel. 86This option can be specified more than once. 87.It Fl h 88Shows a list of available options, each with a short description. 89.It Fl i Ar file 90Specify a memory image 91.Ar file 92to be used by the virtual kernel. 93If no 94.Fl i 95option is given, the kernel will generate a name of the form 96.Pa /var/vkernel/memimg.XXXXXX , 97with the trailing 98.Ql X Ns s 99being replaced by a sequential number, e.g.\& 100.Pa memimg.000001 . 101.It Fl I Ar interface Ns Op Ar :address1 Ns Oo Ar :address2 Oc Ns Oo Ar /netmask Oc Ns Oo Ar =MAC Oc 102Create a virtual network device, with the first 103.Fl I 104option defining 105.Li vke0 , 106the second one 107.Li vke1 , 108and so on. 109.Pp 110The 111.Ar interface 112argument is the name of a 113.Xr tap 4 114device node or the path to a 115.Xr vknetd 8 116socket. 117The 118.Pa /dev/ 119path prefix does not have to be specified and will be automatically prepended 120for a device node. 121Specifying 122.Cm auto 123will pick the first unused 124.Xr tap 4 125device. 126.Pp 127The 128.Ar address1 129and 130.Ar address2 131arguments are the IP addresses of the 132.Xr tap 4 133and 134.Nm vke 135interfaces. 136Optionally, 137.Ar address1 138may be of the form 139.Li bridge Ns Em X 140in which case the 141.Xr tap 4 142interface is added to the specified 143.Xr bridge 4 144interface. 145The 146.Nm vke 147address is not assigned until the interface is brought up in the guest. 148.Pp 149The 150.Ar netmask 151argument applies to all interfaces for which an address is specified. 152.Pp 153The 154.Ar MAC 155argument is the MAC address of the 156.Xr vke 4 157interface. 158If not specified, a pseudo-random one will be generated. 159.Pp 160When running multiple vkernels it is often more convenient to simply 161connect to a 162.Xr vknetd 8 163socket and let vknetd deal with the tap and/or bridge. 164An example of this would be 165.Pa /var/run/vknet:0.0.0.0:10.2.0.2/16 . 166.It Fl l Ar cpulock 167Specify which, if any, real CPUs to lock virtual CPUs to. 168.Ar cpulock 169is one of 170.Cm any , 171.Cm map Ns Op Ns , Ns Ar startCPU , 172or 173.Ar CPU . 174.Pp 175.Cm any 176does not map virtual CPUs to real CPUs. 177This is the default. 178.Pp 179.Cm map Ns Op Ns , Ns Ar startCPU 180maps each virtual CPU to a real CPU starting with real CPU 0 or 181.Ar startCPU 182if specified. 183.Pp 184.Ar CPU 185locks all virtual CPUs to the real CPU specified by 186.Ar CPU . 187.It Fl m Ar size 188Specify the amount of memory to be used by the kernel in bytes, 189.Cm K 190.Pq kilobytes , 191.Cm M 192.Pq megabytes 193or 194.Cm G 195.Pq gigabytes . 196Lowercase versions of 197.Cm K , M , 198and 199.Cm G 200are allowed. 201.It Fl n Ar numcpus Ns Op Ar :lbits Ns Oo Ar :cbits Oc 202.Ar numcpus 203specifies the number of CPUs you wish to emulate. 204Up to 16 CPUs are supported with 2 being the default unless otherwise 205specified. 206.Ar lbits 207specifies the number of bits within APICID(=CPUID) needed for representing 208the logical ID. 209Controls the number of threads/core (0bits - 1 thread, 1bit - 2 threads). 210This parameter is optional (mandatory only if 211.Ar cbits 212is specified). 213.Ar cbits 214specifies the number of bits within APICID(=CPUID) needed for representing 215the core ID. 216Controls the number of core/package (0bits - 1 core, 1bit - 2 cores). 217This parameter is optional. 218.It Fl p Ar pidfile 219Specify a pidfile in which to store the process ID. 220Scripts can use this file to locate the vkernel pid for the purpose of 221shutting down or killing it. 222.Pp 223The vkernel will hold a lock on the pidfile while running. 224Scripts may test for the lock to determine if the pidfile is valid or 225stale so as to avoid accidentally killing a random process. 226Something like '/usr/bin/lockf -ks -t 0 pidfile echo -n' may be used 227to test the lock. 228A non-zero exit code indicates that the pidfile represents a running 229vkernel. 230.Pp 231An error is issued and the vkernel exits if this file cannot be opened for 232writing or if it is already locked by an active vkernel process. 233.It Fl r Ar file Ns Op Ar :serno 234Specify a R/W disk image 235.Ar file 236to be used by the kernel, with the first 237.Fl r 238option defining 239.Li vkd0 , 240the second one 241.Li vkd1 , 242and so on. 243A serial number for the virtual disk can be specified in 244.Ar serno . 245.Pp 246The first 247.Fl r 248or 249.Fl c 250option specified on the command line will be the boot disk. 251.It Fl s 252Boot into single-user mode. 253.It Fl U 254Enable writing to kernel memory and module loading. 255By default, those are disabled for security reasons. 256.It Fl v 257Turn on verbose booting. 258.El 259.Sh DEVICES 260A number of virtual device drivers exist to supplement the virtual kernel. 261.Ss Disk device 262The 263.Nm vkd 264driver allows for up to 16 265.Xr vn 4 266based disk devices. 267The root device will be 268.Li vkd0 269(see 270.Sx EXAMPLES 271for further information on how to prepare a root image). 272.Ss CD-ROM device 273The 274.Nm vcd 275driver allows for up to 16 virtual CD-ROM devices. 276Basically this is a read only 277.Nm vkd 278device with a block size of 2048. 279.Ss Network interface 280The 281.Nm vke 282driver supports up to 16 virtual network interfaces which are associated with 283.Xr tap 4 284devices on the host. 285For each 286.Nm vke 287device, the per-interface read only 288.Xr sysctl 3 289variable 290.Va hw.vke Ns Em X Ns Va .tap_unit 291holds the unit number of the associated 292.Xr tap 4 293device. 294.Pp 295By default, half of the total mbuf clusters available is distributed equally 296among all the vke devices up to 256. 297This can be overriden with the tunable 298.Va hw.vke.max_ringsize . 299Take into account the number passed will be aligned to the lower power of two. 300.Sh SIGNALS 301The virtual kernel only enables 302.Dv SIGQUIT 303and 304.Dv SIGTERM 305while operating in regular console mode. 306Sending 307.Ql \&^\e 308.Pq Dv SIGQUIT 309to the virtual kernel causes the virtual kernel to enter its internal 310.Xr ddb 4 311debugger and re-enable all other terminal signals. 312Sending 313.Dv SIGTERM 314to the virtual kernel triggers a clean shutdown by passing a 315.Dv SIGUSR2 316to the virtual kernel's 317.Xr init 8 318process. 319.Sh DEBUGGING 320It is possible to directly gdb the virtual kernel's process. 321It is recommended that you do a 322.Ql handle SIGSEGV noprint 323to ignore page faults processed by the virtual kernel itself and 324.Ql handle SIGUSR1 noprint 325to ignore signals used for simulating inter-processor interrupts. 326.Sh PROFILING 327To compile a vkernel with profiling support, the 328.Va CONFIGARGS 329variable needs to be used to pass 330.Fl p 331to 332.Xr config 8 . 333.Bd -literal 334cd /usr/src 335make -DNO_MODULES CONFIGARGS=-p buildkernel KERNCONF=VKERNEL 336.Ed 337.Sh FILES 338.Bl -tag -width ".It Pa /sys/config/VKERNEL" -compact 339.It Pa /sys/config/VKERNEL 340.It Pa /sys/config/VKERNEL64 341.El 342.Pp 343Per architecture 344.Nm 345configuration files, for 346.Xr config 8 . 347.Sh CONFIGURATION FILES 348Your virtual kernel is a complete 349.Dx 350system, but you might not want to run all the services a normal kernel runs. 351Here is what a typical virtual kernel's 352.Pa /etc/rc.conf 353file looks like, with some additional possibilities commented out. 354.Bd -literal 355hostname="vkernel" 356network_interfaces="lo0 vke0" 357ifconfig_vke0="DHCP" 358sendmail_enable="NO" 359#syslog_enable="NO" 360blanktime="NO" 361.Ed 362.Sh DISKLESS OPERATION 363To boot a 364.Nm 365from a NFS root, a number of tunables need to be set: 366.Bl -tag -width indent 367.It Va boot.netif.ip 368IP address to be set in the vkernel interface. 369.It Va boot.netif.netmask 370Netmask for the IP to be set. 371.It Va boot.netif.name 372Network interface name inside the vkernel. 373.It Va boot.nfsroot.server 374Host running 375.Xr nfsd 8 . 376.It Va boot.nfsroot.path 377Host path where a world and distribution 378targets are properly installed. 379.El 380.Pp 381See an example on how to boot a diskless 382.Nm 383in the 384.Sx EXAMPLES 385section. 386.Sh EXAMPLES 387A couple of steps are necessary in order to prepare the system to build and 388run a virtual kernel. 389.Ss Setting up the filesystem 390The 391.Nm 392architecture needs a number of files which reside in 393.Pa /var/vkernel . 394Since these files tend to get rather big and the 395.Pa /var 396partition is usually of limited size, we recommend the directory to be 397created in the 398.Pa /home 399partition with a link to it in 400.Pa /var : 401.Bd -literal 402mkdir -p /home/var.vkernel/boot 403ln -s /home/var.vkernel /var/vkernel 404.Ed 405.Pp 406Next, a filesystem image to be used by the virtual kernel has to be 407created and populated (assuming world has been built previously). 408If the image is created on a UFS filesystem you might want to pre-zero it. 409On a HAMMER filesystem you should just truncate-extend to the image size 410as HAMMER does not re-use data blocks already present in the file. 411.Bd -literal 412vnconfig -c -S 2g -T vn0 /var/vkernel/rootimg.01 413disklabel -r -w vn0s0 auto 414disklabel -e vn0s0 # add `a' partition with fstype `4.2BSD' 415newfs /dev/vn0s0a 416mount /dev/vn0s0a /mnt 417cd /usr/src 418make installworld DESTDIR=/mnt 419cd etc 420make distribution DESTDIR=/mnt 421echo '/dev/vkd0s0a / ufs rw 1 1' >/mnt/etc/fstab 422echo 'proc /proc procfs rw 0 0' >>/mnt/etc/fstab 423.Ed 424.Pp 425Edit 426.Pa /mnt/etc/ttys 427and replace the 428.Li console 429entry with the following line and turn off all other gettys. 430.Bd -literal 431console "/usr/libexec/getty Pc" cons25 on secure 432.Ed 433.Pp 434Replace 435.Li \&Pc 436with 437.Li al.Pc 438if you would like to automatically log in as root. 439.Pp 440Then, unmount the disk. 441.Bd -literal 442umount /mnt 443vnconfig -u vn0 444.Ed 445.Ss Compiling the virtual kernel 446In order to compile a virtual kernel use the 447.Li VKERNEL 448kernel configuration file residing in 449.Pa /sys/config 450(or a configuration file derived thereof): 451.Bd -literal 452cd /usr/src 453make -DNO_MODULES buildkernel KERNCONF=VKERNEL 454make -DNO_MODULES installkernel KERNCONF=VKERNEL DESTDIR=/var/vkernel 455.Ed 456.Ss Enabling virtual kernel operation 457A special 458.Xr sysctl 8 , 459.Va vm.vkernel_enable , 460must be set to enable 461.Nm 462operation: 463.Bd -literal 464sysctl vm.vkernel_enable=1 465.Ed 466.Ss Configuring the network on the host system 467In order to access a network interface of the host system from the 468.Nm , 469you must add the interface to a 470.Xr bridge 4 471device which will then be passed to the 472.Fl I 473option: 474.Bd -literal 475kldload if_bridge.ko 476kldload if_tap.ko 477ifconfig bridge0 create 478ifconfig bridge0 addm re0 # assuming re0 is the host's interface 479ifconfig bridge0 up 480.Ed 481.Ss Running the kernel 482Finally, the virtual kernel can be run: 483.Bd -literal 484cd /var/vkernel 485\&./boot/kernel/kernel -m 64m -r rootimg.01 -I auto:bridge0 486.Ed 487.Pp 488You can issue the 489.Xr reboot 8 , 490.Xr halt 8 , 491or 492.Xr shutdown 8 493commands from inside a virtual kernel. 494After doing a clean shutdown the 495.Xr reboot 8 496command will re-exec the virtual kernel binary while the other two will 497cause the virtual kernel to exit. 498.Ss Diskless operation 499Booting a 500.Nm 501with a 502.Xr vknetd 8 503network configuration: 504.Bd -literal 505\&./boot/kernel/kernel -m 64m -m -i memimg.0000 -I /var/run/vknet 506 -e boot.netif.ip=172.1.0.4 507 -e boot.netif.netmask=255.255.0.0 508 -e boot.netif.name=vke0 509 -e boot.nfsroot.server=172.1.0.1 510 -e boot.nfsroot.path=/home/vkernel/vkdiskless 511.Ed 512.Sh BUILDING THE WORLD UNDER A VKERNEL 513The virtual kernel platform does not have all the header files expected 514by a world build, so the easiest thing to do right now is to specify a 515pc32 (in a 32 bit vkernel) or pc64 (in a 64 bit vkernel) target when 516building the world under a virtual kernel, like this: 517.Bd -literal 518vkernel# make MACHINE_PLATFORM=pc32 buildworld 519vkernel# make MACHINE_PLATFORM=pc32 installworld 520.Ed 521.Sh SEE ALSO 522.Xr vknet 1 , 523.Xr bridge 4 , 524.Xr tap 4 , 525.Xr vn 4 , 526.Xr sysctl.conf 5 , 527.Xr build 7 , 528.Xr config 8 , 529.Xr disklabel 8 , 530.Xr ifconfig 8 , 531.Xr vknetd 8 , 532.Xr vnconfig 8 533.Rs 534.%A Aggelos Economopoulos 535.%D March 2007 536.%T "A Peek at the DragonFly Virtual Kernel" 537.Re 538.Sh HISTORY 539Virtual kernels were introduced in 540.Dx 1.7 . 541.Sh AUTHORS 542.An -nosplit 543.An Matt Dillon 544thought up and implemented the 545.Nm 546architecture and wrote the 547.Nm vkd 548device driver. 549.An Sepherosa Ziehau 550wrote the 551.Nm vke 552device driver. 553This manual page was written by 554.An Sascha Wildner . 555