diff --git a/share/man/man9/unr.9 b/share/man/man9/unr.9 index c0cf44b9033e..201717ccdfd2 100644 --- a/share/man/man9/unr.9 +++ b/share/man/man9/unr.9 @@ -1,149 +1,193 @@ .\" Copyright (c) 2005 Gleb Smirnoff .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .\" $FreeBSD$ .\" .Dd April 21, 2022 .Dt UNR 9 .Os .Sh NAME .Nm new_unrhdr , .Nm clean_unrhdr , .Nm clear_unrhdr , .Nm delete_unrhdr , .Nm alloc_unr , .Nm alloc_unr_specific , -.Nm free_unr +.Nm free_unr , +.Nm create_iter_unr , +.Nm next_iter_unr , +.Nm free_iter_unr .Nd "kernel unit number allocator" .Sh SYNOPSIS .In sys/systm.h .Ft "struct unrhdr *" .Fn new_unrhdr "int low" "int high" "struct mtx *mutex" .Ft void .Fn clean_unrhdr "struct unrhdr *uh" .Ft void .Fn clean_unrhdrl "struct unrhdr *uh" .Ft void .Fn clear_unrhdr "struct unrhdr *uh" .Ft void .Fn delete_unrhdr "struct unrhdr *uh" .Ft int .Fn alloc_unr "struct unrhdr *uh" .Ft int .Fn alloc_unrl "struct unrhdr *uh" .Ft int .Fn alloc_unr_specific "struct unrhdr *uh" "u_int item" .Ft void .Fn free_unr "struct unrhdr *uh" "u_int item" +.Ft void * +.Fn create_iter_unr "struct unrhdr *uh" +.Ft int +.Fn next_iter_unr "void *handle" +.Ft void +.Fn free_iter_unr "void *handle" .Sh DESCRIPTION The kernel unit number allocator is a generic facility, which allows to allocate unit numbers within a specified range. .Bl -tag -width indent .It Fn new_unrhdr low high mutex Initialize a new unit number allocator entity. The .Fa low and .Fa high arguments specify minimum and maximum number of unit numbers. There is no cost associated with the range of unit numbers, so unless the resource really is finite, .Dv INT_MAX can be used. If .Fa mutex is not .Dv NULL , it is used for locking when allocating and freeing units. If the passed value is the token .Va UNR_NO_MTX , then no locking is applied internally. Otherwise, internal mutex is used. .It Fn clear_unrhdr uh Clear all units from the specified unit number allocator entity. This function resets the entity as if it were just initialized with .Fn new_unrhdr . .It Fn delete_unrhdr uh Delete specified unit number allocator entity. This function frees the memory associated with the entity, it does not free any units. To free all units use .Fn clear_unrhdr . .It Fn clean_unrhdr uh Freeing unit numbers might result in some internal memory becoming unused. There are .Nm unit allocator consumers that cannot tolerate taking .Xr malloc 9 locks to free the memory, while having their unit mutex locked. For this reason, free of the unused memory after delete is postponed until the consumer can afford calling into the .Xr malloc 9 subsystem. Call .Fn clean_unrhdr uh to do the cleanup. In particular, this needs to be done before freeing a unr, if a deletion of units could have been performed. .It Fn clean_unrhdrl Same as .Fn clean_unrhdr , but assumes that the unr mutex is already owned, if any. .It Fn alloc_unr uh Return a new unit number. The lowest free number is always allocated. This function does not allocate memory and never sleeps, however it may block on a mutex. If no free unit numbers are left, .Li \-1 is returned. .It Fn alloc_unrl uh Same as .Fn alloc_unr except that mutex is assumed to be already locked and thus is not used. .It Fn alloc_unr_specific uh item Allocate a specific unit number. This function allocates memory and thus may sleep. The allocated unit number is returned on success. If the specified number is already allocated or out of the range, .Li \-1 is returned. .It Fn free_unr uh item Free a previously allocated unit number. This function may require allocating memory, and thus it can sleep. There is no pre-locked variant. .El +.Sh ITERATOR INTERFACE +The +.Nm unr +facility provides an interface to iterate over all allocated units +for the given +.Dv unrhdr . +Iterators are identified by an opaque handle. +More than one iterators can operate simultaneously; the iterator position +data is recorded only in the iterator handle. +.Pp +Consumers must ensure that the unit allocator is not modified between +calls to the iterator functions. +In particular, the internal allocator mutex cannot provide consistency, +because it is acquired and dropped inside the +.Fn next_iter_unr +function. +If the allocator was modified, it is safe to free the iterator with +.Fn free_iter_unr +method nevertheless. +.Bl -tag -width indent +.It Fn create_iter_unr uh +Create an iterator. +Return the handle that should be passed to other iterator functions. +.It Fn next_iter_unr handle +Return the value of the next unit. +Units are returned in ascending order. +A return value of +.Li \-1 +indicates the end of iteration, in which +case +.Li \-1 +is returned for all future calls. +.It Fn free_iter_unr handle +Free the iterator, handle is no longer valid. +.El .Sh CODE REFERENCES The above functions are implemented in .Pa sys/kern/subr_unit.c . .Sh HISTORY Kernel unit number allocator first appeared in .Fx 6.0 . .Sh AUTHORS .An -nosplit Kernel unit number allocator was written by .An Poul-Henning Kamp . This manpage was written by .An Gleb Smirnoff .