2 .\" Copyright (c) 2010 The FreeBSD Foundation
3 .\" All rights reserved.
5 .\" This documentation was written by CK Software GmbH under sponsorship from
6 .\" the FreeBSD Foundation.
8 .\" Redistribution and use in source and binary forms, with or without
9 .\" modification, are permitted provided that the following conditions
11 .\" 1. Redistributions of source code must retain the above copyright
12 .\" notice, this list of conditions and the following disclaimer.
13 .\" 2. Redistributions in binary form must reproduce the above copyright
14 .\" notice, this list of conditions and the following disclaimer in the
15 .\" documentation and/or other materials provided with the distribution.
17 .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36 .Nd "network subsystem virtualization infrastructure"
39 .Cd "options VNET_DEBUG"
43 .\"------------------------------------------------------------
44 .Ss "Constants and Global Variables"
50 .Vt extern struct vnet *vnet0;
51 .\"------------------------------------------------------------
52 .Ss "Variable Declaration"
70 #define V_name VNET(name)
72 .\" ------------------------------------------------------------
73 .Ss "Virtual Instance Selection"
106 .Fo VNET_ITERATOR_DECL
113 .\" ------------------------------------------------------------
118 .Fo VNET_LIST_RUNLOCK
120 .Fo VNET_LIST_RLOCK_NOSLEEP
122 .Fo VNET_LIST_RUNLOCK_NOSLEEP
124 .\" ------------------------------------------------------------
125 .Ss "Startup and Teardown Functions"
139 .Fa "enum sysinit_sub_id subsystem"
140 .Fa "enum sysinit_elem_order order"
141 .Fa "sysinit_cfunc_t func"
142 .Fa "const void *arg"
147 .Fa "enum sysinit_sub_id subsystem"
148 .Fa "enum sysinit_elem_order order"
149 .Fa "sysinit_cfunc_t func"
150 .Fa "const void *arg"
152 .\" ------------------------------------------------------------
155 .Fo VNET_GLOBAL_EVENTHANDLER_REGISTER
156 .Fa "const char *name"
162 .Fo VNET_GLOBAL_EVENTHANDLER_REGISTER_TAG
163 .Fa "eventhandler_tag tag"
164 .Fa "const char *name"
169 .\" ------------------------------------------------------------
170 .Ss "Sysctl Handling"
172 .Fa parent nbr name access ptr val descr
175 .Fa parent nbr name access ptr arg handler fmt descr
177 .Fo SYSCTL_VNET_STRING
178 .Fa parent nbr name access arg len descr
180 .Fo SYSCTL_VNET_STRUCT
181 .Fa parent nbr name access ptr type descr
184 .Fa parent nbr name access ptr val descr
189 .\" ------------------------------------------------------------
192 is the name of a technique to virtualize the network stack.
193 The basic idea is to change global resources most notably variables into
194 per network stack resources and have functions, sysctls, eventhandlers,
195 etc. access and handle them in the context of the correct instance.
196 Each (virtual) network stack is attached to a
200 being the unrestricted default network stack of the base system.
202 The global defines for
208 to access internals for debugging reasons.
209 .\" ------------------------------------------------------------
210 .Ss "Variable Declaration"
212 Variables are virtualized by using the
214 macro rather than writing them out as
216 One can still use static initialization or storage class specifiers, e.g.,
218 .Dl Li static VNET_DEFINE(int, foo) = 1;
220 .Dl Li static VNET_DEFINE(SLIST_HEAD(, bar), bars);
222 Static initialization is not possible when the virtualized variable
223 would need to be referenced, e.g., with
224 .Dq TAILQ_HEAD_INITIALIZER() .
227 based initialization function must be used.
229 External variables have to be declared using the
232 In either case the convention is to define another macro,
233 that is then used throughout the implementation to access that variable.
234 The variable name is usually prefixed by
236 to express that it is virtualized.
239 macro will then translate accesses to that variable to the copy of the
240 currently selected instance (see the
241 .Sx "Virtual instance selection"
244 .Dl Li #define V_name VNET(name)
247 Do not confuse this with the convention used by
252 macro returns the offset within the memory region of the virtual network
254 It is usually only used with
257 .\" ------------------------------------------------------------
258 .Ss "Virtual Instance Selection"
260 There are three different places where the current virtual
261 network stack pointer is stored and can be taken from:
262 .Bl -enum -offset indent
266 .Dl "(struct prison *)->pr_vnet"
268 For convenience the following macros are provided:
269 .Bd -literal -compact -offset indent
270 .Fn CRED_TO_VNET "struct ucred *"
271 .Fn TD_TO_VNET "struct thread *"
272 .Fn P_TO_VNET "struct proc *"
277 .Dl "(struct socket *)->so_vnet"
281 .Dl "(struct ifnet *)->if_vnet"
285 In addition the currently active instance is cached in
286 .Dq "curthread->td_vnet"
287 which is usually only accessed through the
292 To set the correct context of the current virtual network instance, use the
295 .Fn CURVNET_SET_QUIET
298 .Fn CURVNET_SET_QUIET
299 version will not record vnet recursions in case the kernel was compiled
301 .Cd "options VNET_DEBUG"
302 and should thus only be used in well known cases, where recursion is
304 Both macros will save the previous state on the stack and it must be restored
310 As the previous state is saved on the stack, you cannot have multiple
312 calls in the same block.
315 As the previous state is saved on the stack, a
317 call has to be in the same block as the
319 call or in a subblock with the same idea of the saved instances as the
323 As each macro is a set of operations and, as previously explained, cannot
324 be put into its own block when defined, one cannot conditionally set
325 the current vnet context.
329 .Bd -literal -offset indent
335 .Bd -literal -offset indent
343 Sometimes one needs to loop over all virtual instances, for example to update
344 virtual from global state, to run a function from a
346 for each instance, etc.
348 .Fn VNET_ITERATOR_DECL
352 The former macro defines the variable that iterates over the loop,
353 and the latter loops over all of the virtual network stack instances.
356 for how to savely traverse the list of all virtual instances.
361 macro provides a safe way to check whether the currently active instance is the
362 unrestricted default network stack of the base system
368 macro provides a way to conditionally add assertions that are only active with
370 compiled in and either
371 .Cd "options VNET_DEBUG"
373 .Cd "options INVARIANTS"
375 It uses the same semantics as
377 .\" ------------------------------------------------------------
380 For public access to the list of virtual network stack instances
383 macro, read locks are provided.
384 Macros are used to abstract from the actual type of the locks.
385 If a caller may sleep while traversing the list, it must use the
388 .Fn VNET_LIST_RUNLOCK
390 Otherwise, the caller can use
391 .Fn VNET_LIST_RLOCK_NOSLEEP
393 .Fn VNET_LIST_RUNLOCK_NOSLEEP .
394 .\" ------------------------------------------------------------
395 .Ss "Startup and Teardown Functions"
397 To start or tear down a virtual network stack instance the internal
402 are provided and called from the jail framework.
403 They run the publicly provided methods to handle network stack
404 startup and teardown.
406 For public control, the system startup interface has been enhanced
407 to not only handle a system boot but to also handle a virtual
408 network stack startup and teardown.
409 To the base system the
413 macros look exactly as if there were no virtual network stack.
416 is not compiled in they are compiled to the standard
419 In addition to that they are run for each virtual network stack
420 when starting or, in reverse order, when shutting down.
421 .\" ------------------------------------------------------------
424 Eventhandlers can be handled in two ways:
426 .Bl -enum -offset indent -compact
430 returned in each virtual instance and properly free the eventhandlers
431 on teardown using those, or
433 use one eventhandler that will iterate over all virtual network
437 For the first case one can just use the normal
439 functions, while for the second case the
440 .Fn VNET_GLOBAL_EVENTHANDLER_REGISTER
442 .Fn VNET_GLOBAL_EVENTHANDLER_REGISTER_TAG
445 .Fn VNET_GLOBAL_EVENTHANDLER_REGISTER_TAG
446 takes an extra first argument that will carry the
449 Eventhandlers registered with either of these will not run
453 will be called from an internal iterator function for each vnet.
454 Both macros can only be used for eventhandlers that do not take
455 additional arguments, as the variadic arguments from an
456 .Xr EVENTHANDLER_INVOKE 9
457 call will be ignored.
458 .\" ------------------------------------------------------------
459 .Ss "Sysctl Handling"
463 can be virtualized by using one of the
467 They take the same arguments as the standard
469 functions, with the only difference, that the
471 argument has to be passed as
475 so that the variable can be selected from the correct memory
476 region of the virtual network stack instance of the caller.
478 For the very rare case a sysctl handler function would want to
482 .Fn VNET_SYSCTL_ARG req arg1
483 is provided that will translate the
485 argument to the correct memory address in the virtual network stack
486 context of the caller.
487 .\" ------------------------------------------------------------
497 The virtual network stack implementation first appeared in
500 This manual page was written by
501 .An Bjoern A. Zeeb, CK Software GmbH,
502 under sponsorship from the FreeBSD Foundation.