dect
/
libnl
Archived
13
0
Fork 0
This repository has been archived on 2022-02-17. You can view files and clone it, but cannot push or open issues or pull requests.
libnl/include/netlink-private/object-api.h

377 lines
10 KiB
C
Raw Normal View History

/*
* netlink-private/object-api.c Object API
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation version 2.1
* of the License.
*
* Copyright (c) 2003-2013 Thomas Graf <tgraf@suug.ch>
*/
#ifndef NETLINK_OBJECT_API_H_
#define NETLINK_OBJECT_API_H_
#include <netlink/netlink.h>
#include <netlink/utils.h>
#include <netlink/object.h>
#ifdef __cplusplus
extern "C" {
#endif
/**
* @ingroup object
* @defgroup object_api Object API
* @brief
*
* @par 1) Object Definition
* @code
* // Define your object starting with the common object header
* struct my_obj {
* NLHDR_COMMON
* int my_data;
* };
*
* // Fill out the object operations structure
* struct nl_object_ops my_ops = {
* .oo_name = "my_obj",
* .oo_size = sizeof(struct my_obj),
* };
*
* // At this point the object can be allocated, you may want to provide a
* // separate _alloc() function to ease allocting objects of this kind.
* struct nl_object *obj = nl_object_alloc(&my_ops);
*
* // And release it again...
* nl_object_put(obj);
* @endcode
*
* @par 2) Allocating additional data
* @code
* // You may require to allocate additional data and store it inside
* // object, f.e. assuming there is a field `ptr'.
* struct my_obj {
* NLHDR_COMMON
* void * ptr;
* };
*
* // And at some point you may assign allocated data to this field:
* my_obj->ptr = calloc(1, ...);
*
* // In order to not introduce any memory leaks you have to release
* // this data again when the last reference is given back.
* static void my_obj_free_data(struct nl_object *obj)
* {
* struct my_obj *my_obj = nl_object_priv(obj);
*
* free(my_obj->ptr);
* }
*
* // Also when the object is cloned, you must ensure for your pointer
* // stay valid even if one of the clones is freed by either making
* // a clone as well or increase the reference count.
* static int my_obj_clone(struct nl_object *src, struct nl_object *dst)
* {
* struct my_obj *my_src = nl_object_priv(src);
* struct my_obj *my_dst = nl_object_priv(dst);
*
* if (src->ptr) {
* dst->ptr = calloc(1, ...);
* memcpy(dst->ptr, src->ptr, ...);
* }
* }
*
* struct nl_object_ops my_ops = {
* ...
* .oo_free_data = my_obj_free_data,
* .oo_clone = my_obj_clone,
* };
* @endcode
*
* @par 3) Object Dumping
* @code
* static int my_obj_dump_detailed(struct nl_object *obj,
* struct nl_dump_params *params)
* {
* struct my_obj *my_obj = nl_object_priv(obj);
*
* // It is absolutely essential to use nl_dump() when printing
* // any text to make sure the dumping parameters are respected.
* nl_dump(params, "Obj Integer: %d\n", my_obj->my_int);
*
* // Before we can dump the next line, make sure to prefix
* // this line correctly.
2008-05-23 21:45:14 +00:00
* nl_new_line(params);
*
* // You may also split a line into multiple nl_dump() calls.
* nl_dump(params, "String: %s ", my_obj->my_string);
* nl_dump(params, "String-2: %s\n", my_obj->another_string);
* }
*
* struct nl_object_ops my_ops = {
* ...
* .oo_dump[NL_DUMP_FULL] = my_obj_dump_detailed,
* };
* @endcode
*
* @par 4) Object Attributes
* @code
* // The concept of object attributes is optional but can ease the typical
* // case of objects that have optional attributes, e.g. a route may have a
* // nexthop assigned but it is not required to.
*
* // The first step to define your object specific bitmask listing all
* // attributes
* #define MY_ATTR_FOO (1<<0)
* #define MY_ATTR_BAR (1<<1)
*
* // When assigning an optional attribute to the object, make sure
* // to mark its availability.
* my_obj->foo = 123123;
* my_obj->ce_mask |= MY_ATTR_FOO;
*
* // At any time you may use this mask to check for the availability
* // of the attribute, e.g. while dumping
* if (my_obj->ce_mask & MY_ATTR_FOO)
* nl_dump(params, "foo %d ", my_obj->foo);
*
* // One of the big advantages of this concept is that it allows for
* // standardized comparisons which make it trivial for caches to
* // identify unique objects by use of unified comparison functions.
* // In order for it to work, your object implementation must provide
* // a comparison function and define a list of attributes which
* // combined together make an object unique.
*
* static int my_obj_compare(struct nl_object *_a, struct nl_object *_b,
* uint32_t attrs, int flags)
* {
* struct my_obj *a = nl_object_priv(_a):
* struct my_obj *b = nl_object_priv(_b):
* int diff = 0;
*
* // We help ourselves in defining our own DIFF macro which will
* // call ATTR_DIFF() on both objects which will make sure to only
* // compare the attributes if required.
* #define MY_DIFF(ATTR, EXPR) ATTR_DIFF(attrs, MY_ATTR_##ATTR, a, b, EXPR)
*
* // Call our own diff macro for each attribute to build a bitmask
* // representing the attributes which mismatch.
* diff |= MY_DIFF(FOO, a->foo != b->foo)
* diff |= MY_DIFF(BAR, strcmp(a->bar, b->bar))
*
* return diff;
* }
*
* // In order to identify identical objects with differing attributes
* // you must specify the attributes required to uniquely identify
* // your object. Make sure to not include too many attributes, this
* // list is used when caches look for an old version of an object.
* struct nl_object_ops my_ops = {
* ...
* .oo_id_attrs = MY_ATTR_FOO,
* .oo_compare = my_obj_compare,
* };
* @endcode
* @{
*/
/**
* Common Object Header
*
* This macro must be included as first member in every object
* definition to allow objects to be cached.
*/
#define NLHDR_COMMON \
int ce_refcnt; \
struct nl_object_ops * ce_ops; \
struct nl_cache * ce_cache; \
struct nl_list_head ce_list; \
int ce_msgtype; \
int ce_flags; \
uint32_t ce_mask;
struct nl_object
{
NLHDR_COMMON
};
/**
* Return true if attribute is available in both objects
* @arg A an object
* @arg B another object
* @arg ATTR attribute bit
*
* @return True if the attribute is available, otherwise false is returned.
*/
#define AVAILABLE(A, B, ATTR) (((A)->ce_mask & (B)->ce_mask) & (ATTR))
/**
* Return true if attribute is available in only one of both objects
* @arg A an object
* @arg B another object
* @arg ATTR attribute bit
*
* @return True if the attribute is available in only one of both objects,
* otherwise false is returned.
*/
#define AVAILABLE_MISMATCH(A, B, ATTR) (((A)->ce_mask ^ (B)->ce_mask) & (ATTR))
/**
* Return true if attributes mismatch
* @arg A an object
* @arg B another object
* @arg ATTR attribute bit
* @arg EXPR Comparison expression
*
* This function will check if the attribute in question is available
* in both objects, if not this will count as a mismatch.
*
* If available the function will execute the expression which must
* return true if the attributes mismatch.
*
* @return True if the attribute mismatch, or false if they match.
*/
#define ATTR_MISMATCH(A, B, ATTR, EXPR) (AVAILABLE_MISMATCH(A, B, ATTR) || \
(AVAILABLE(A, B, ATTR) && (EXPR)))
/**
* Return attribute bit if attribute does not match
* @arg LIST list of attributes to be compared
* @arg ATTR attribute bit
* @arg A an object
* @arg B another object
* @arg EXPR Comparison expression
*
* This function will check if the attribute in question is available
* in both objects, if not this will count as a mismatch.
*
* If available the function will execute the expression which must
* return true if the attributes mismatch.
*
* In case the attributes mismatch, the attribute is returned, otherwise
* 0 is returned.
*
* @code
* diff |= ATTR_DIFF(attrs, MY_ATTR_FOO, a, b, a->foo != b->foo);
* @endcode
*/
#define ATTR_DIFF(LIST, ATTR, A, B, EXPR) \
({ int diff = 0; \
if (((LIST) & (ATTR)) && ATTR_MISMATCH(A, B, ATTR, EXPR)) \
diff = ATTR; \
diff; })
/**
* Object Operations
*/
struct nl_object_ops
{
/**
* Unique name of object type
*
* Must be in the form family/name, e.g. "route/addr"
*/
char * oo_name;
/** Size of object including its header */
size_t oo_size;
/* List of attributes needed to uniquely identify the object */
uint32_t oo_id_attrs;
/**
* Constructor function
*
* Will be called when a new object of this type is allocated.
* Can be used to initialize members such as lists etc.
*/
void (*oo_constructor)(struct nl_object *);
/**
* Destructor function
*
* Will be called when an object is freed. Must free all
* resources which may have been allocated as part of this
* object.
*/
void (*oo_free_data)(struct nl_object *);
/**
* Cloning function
*
* Will be called when an object needs to be cloned. Please
* note that the generic object code will make an exact
* copy of the object first, therefore you only need to take
* care of members which require reference counting etc.
*
* May return a negative error code to abort cloning.
*/
int (*oo_clone)(struct nl_object *, struct nl_object *);
/**
* Dumping functions
*
* Will be called when an object is dumped. The implementations
* have to use nl_dump(), nl_dump_line(), and nl_new_line() to
* dump objects.
*
* The functions must return the number of lines printed.
*/
2008-05-23 21:45:14 +00:00
void (*oo_dump[NL_DUMP_MAX+1])(struct nl_object *,
struct nl_dump_params *);
/**
* Comparison function
*
* Will be called when two objects of the same type are
* compared. It takes the two objects in question, an object
* specific bitmask defining which attributes should be
* compared and flags to control the behaviour.
*
* The function must return a bitmask with the relevant bit
* set for each attribute that mismatches.
*/
int (*oo_compare)(struct nl_object *, struct nl_object *,
uint32_t, int);
Add support for updating objects in the cache This patch adds support to update a cache object during cache_include instead of the current approach of deleting the original object and adding a new one. This operation is conditional on the object implementing the operation. If the update is not successful, cache_include falls back to the existing cache inclusion process of deleting and adding the object. It adds a new object operation called oo_update. oo_update takes two objects as arguments, first being the existing cache object that needs update, the second argument being the new object. Currently it is left to the implementor to use the msg type to decide wether to delete or add the new object attributes to the old one. But the operation type or msg type can be easily made part of the object arguments. The motivation for this change is explained below in the context of including support for AF_BRIDGE objects into the link cache. libnl today deletes an object before it includes an identical object. But for some objects like the AF_BRIDGE objects this does not work well. link cache uses the ifindex as its key in object searches. If link cache were to support AF_BRIDGE family objects, todays implementation, - will replace the original link object with the bridge port link object for add notifications - And a bridge port delete notification from kernel would delete the link object from the cache leaving the cache without the link object until the kernel sends another notification for that link The bridge port link notification contains some base link object attributes plus bridge specific protocol info attributes. In such cases we think an operation to update the existing object in place in cache might be useful. This can be made to work for AF_INET6 link objects too. Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com> Reviewed-by: Nolan Leake <nolan@cumulusnetworks.com> Reviewed-by: Shrijeet Mukherjee <shm@cumulusnetworks.com> Reviewed-by: Wilson Kok <wkok@cumulusnetworks.com>
2012-11-01 21:13:30 +00:00
/**
* update function
*
* Will be called when the object given by first argument
* needs to be updated with the contents of the second object
*
* The function must return 0 for success and error for failure
* to update. In case of failure its assumed that the original
* object is not touched
*/
int (*oo_update)(struct nl_object *, struct nl_object *);
/**
* Hash Key generator function
*
* When called returns a hash key for the object being
* referenced. This key will be used by higher level hash functions
* to build association lists. Each object type gets to specify
* it's own key formulation
*/
void (*oo_keygen)(struct nl_object *, uint32_t *, uint32_t);
char *(*oo_attrs2str)(int, char *, size_t);
/**
* Get key attributes by family function
*/
uint32_t (*oo_id_attrs_get)(struct nl_object *);
};
/** @} */
#ifdef __cplusplus
}
#endif
#endif