man/man9/sysctl_ctx_init.9

.\"
.\" Copyright (c) 2000, Andrzej Bialecki <abial@FreeBSD.org>
.\" 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.
.\" 3. The name of the author may not be used to endorse or promote products
.\"    derived from this software without specific prior written permission.
.\"
.\" 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 July 31, 2014
.Dt SYSCTL_CTX_INIT 9
.Os
.Sh NAME
.Nm sysctl_ctx_init ,
.Nm sysctl_ctx_free ,
.Nm sysctl_ctx_entry_add ,
.Nm sysctl_ctx_entry_find ,
.Nm sysctl_ctx_entry_del
.Nd "sysctl context for managing dynamically created sysctl OIDs"
.Sh SYNOPSIS
.In sys/types.h
.In sys/sysctl.h
.Ft int
.Fo sysctl_ctx_init
.Fa "struct sysctl_ctx_list *clist"
.Fc
.Ft int
.Fo sysctl_ctx_free
.Fa "struct sysctl_ctx_list *clist"
.Fc
.Ft struct sysctl_ctx_entry *
.Fo sysctl_ctx_entry_add
.Fa "struct sysctl_ctx_list *clist"
.Fa "struct sysctl_oid *oidp"
.Fc
.Ft struct sysctl_ctx_entry *
.Fo sysctl_ctx_entry_find
.Fa "struct sysctl_ctx_list *clist"
.Fa "struct sysctl_oid *oidp"
.Fc
.Ft int
.Fo sysctl_ctx_entry_del
.Fa "struct sysctl_ctx_list *clist"
.Fa "struct sysctl_oid *oidp"
.Fc
.Sh DESCRIPTION
These functions provide an interface
for managing dynamically created OIDs.
The sysctl context is responsible for keeping track of created OIDs,
as well as their proper removal when needed.
It adds a simple transactional aspect to OID removal operations;
i.e., if a removal operation fails part way,
it is possible to roll back the sysctl tree
to its previous state.
.Pp
The
.Fn sysctl_ctx_init
function initializes a sysctl context.
The
.Fa clist
argument must point to an already allocated variable.
A context
.Em must
be initialized before use.
Once it is initialized,
a pointer to the context can be passed as an argument to all the
.Fa SYSCTL_ADD_*
macros (see
.Xr sysctl_add_oid 9 ) ,
and it will be updated with entries pointing to newly created OIDS.
.Pp
Internally, the context is represented as a
.Xr queue 3
TAILQ linked list.
The list consists of
.Li struct sysctl_ctx_entry
entries:
.Bd -literal -offset indent
struct sysctl_ctx_entry {
	struct sysctl_oid *entry;
	TAILQ_ENTRY(sysctl_ctx_entry) link;
};

TAILQ_HEAD(sysctl_ctx_list, sysctl_ctx_entry);
.Ed
.Pp
Each context entry points to one dynamic OID that it manages.
Newly created OIDs are always inserted in the front of the list.
.Pp
The
.Fn sysctl_ctx_free
function removes the context and associated OIDs it manages.
If the function completes successfully,
all managed OIDs have been unregistered
(removed from the tree)
and freed,
together with all their allocated memory,
and the entries of the context have been freed as well.
.Pp
The removal operation is performed in two steps.
First, for each context entry, the function
.Xr sysctl_remove_oid 9
is executed, with parameter
.Fa del
set to 0, which inhibits the freeing of resources.
If there are no errors during this step,
.Fn sysctl_ctx_free
proceeds to the next step.
If the first step fails,
all unregistered OIDs associated with the context are registered again.
.Pp
.Em Note :
in most cases, the programmer specifies
.Dv OID_AUTO
as the OID number when creating an OID.
However, during registration of the OID in the tree,
this number is changed to the first available number
greater than or equal to
.Dv CTL_AUTO_START .
If the first step of context deletion fails,
re-registration of the OID does not change the already assigned OID number
(which is different from OID_AUTO).
This ensures that re-registered entries
maintain their original positions in the tree.
.Pp
The second step actually performs the deletion of the dynamic OIDs.
.Xr sysctl_remove_oid 9
iterates through the context list,
starting from beginning (i.e., the newest entries).
.Em Important :
this time, the function not only deletes the OIDs from the tree,
but also frees their memory (provided that oid_refcnt == 0),
as well as the memory of all context entries.
.Pp
The
.Fn sysctl_ctx_entry_add
function allows the addition of an existing dynamic OID to a context.
.Pp
The
.Fn sysctl_ctx_entry_del
function removes an entry from the context.
.Em Important :
in this case, only the corresponding
.Li struct sysctl_ctx_entry
is freed, but the
.Fa oidp
pointer remains intact.
Thereafter, the programmer is responsible for managing the resources
allocated to this OID.
.Pp
The
.Fn sysctl_ctx_entry_find
function searches for a given
.Fa oidp
within a context list,
either returning a pointer to the
.Fa struct sysctl_ctx_entry
found,
or
.Dv NULL .
.Sh EXAMPLES
The following is an example of how to create a new top-level category
and how to hook up another subtree to an existing static node.
This example uses contexts to keep track of the OIDs.
.Bd -literal
#include <sys/sysctl.h>
 ...
static struct sysctl_ctx_list clist;
static struct sysctl_oid *oidp;
static int a_int;
static const char *string = "dynamic sysctl";
 ...

sysctl_ctx_init(&clist);
oidp = SYSCTL_ADD_ROOT_NODE(&clist,
	OID_AUTO, "newtree", CTLFLAG_RW, 0, "new top level tree");
oidp = SYSCTL_ADD_INT(&clist, SYSCTL_CHILDREN(oidp),
	OID_AUTO, "newint", CTLFLAG_RW, &a_int, 0, "new int leaf");
 ...
oidp = SYSCTL_ADD_NODE(&clist, SYSCTL_STATIC_CHILDREN(_debug),
	OID_AUTO, "newtree", CTLFLAG_RW, 0, "new tree under debug");
oidp = SYSCTL_ADD_STRING(&clist, SYSCTL_CHILDREN(oidp),
	OID_AUTO, "newstring", CTLFLAG_RD, string, 0, "new string leaf");
 ...
/* Now we can free up the OIDs */
if (sysctl_ctx_free(&clist)) {
	printf("can't free this context - other OIDs depend on it");
	return (ENOTEMPTY);
} else {
	printf("Success!\\n");
	return (0);
}
.Ed
.Pp
This example creates the following subtrees:
.Bd -literal -offset indent
debug.newtree.newstring
newtree.newint
.Ed
.Pp
Note that both trees are removed, and their resources freed,
through one
.Fn sysctl_ctx_free
call, which starts by freeing the newest entries (leaves)
and then proceeds to free the older entries (in this case the nodes).
.Sh SEE ALSO
.Xr queue 3 ,
.Xr sysctl 8 ,
.Xr sysctl 9 ,
.Xr sysctl_add_oid 9 ,
.Xr sysctl_remove_oid 9
.Sh HISTORY
These functions first appeared in
.Fx 4.2 .
.Sh AUTHORS
.An Andrzej Bialecki Aq Mt abial@FreeBSD.org
.Sh BUGS
The current removal algorithm is somewhat heavy.
In the worst case,
all OIDs need to be unregistered, registered again,
and then unregistered and deleted.
However, the algorithm does guarantee transactional properties
for removal operations.
.Pp
All operations on contexts involve linked list traversal.
For this reason,
creation and removal of entries is relatively costly.