The return value of the *nl_send_simple() functions is
inconsistent and not according to the documentation.
nl_send_simple() is document to return the number of bytes sent.
Other *nl_send_simple() functions are documented to return 0 on
success -- for the most part.
See also commit b70174668b which
changed behavior of nl_rtgen_request() to be according to documenation.
Don't change behavior again, only adjust the documentation.
http://lists.infradead.org/pipermail/libnl/2015-May/001872.html
Reported-by: Xiao Jia <stfairy@gmail.com>
Signed-off-by: Thomas Haller <thaller@redhat.com>
The cache pickup does a check for duplicates for some time now.
However, the pickup in nl_cache_refill() explicitely clears the
cache, so the pickup doesn't need to actually do the dupe check,
as uniqueness is already guaranteed be Netlink subsystem.
And avoiding the dup check is beneficitial for performance
reasons, as the current algorithm is O(n^2).
http://lists.infradead.org/pipermail/libnl/2014-October/001680.html
Fixes: 96bb7c9a4c
Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Thomas Haller <thaller@redhat.com>
This patch adds new cache find api
nl_cache_find api was suggested by Thomas.
Unlike nl_cache_search, this patch uses
nl_object_match_filter() to look for an
object match.
Am not sure this matches what was decided
on the list few weeks back. I will be happy
to make any changes.
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
In cache_include(), the user callback is used even though one might not
have been specified, leading to a crash. This happens, for example, when
NetworkManager calls nl_cache_include() with the change_cb argument set
to NULL. Fix it by making sure the callback points to a valid address.
Signed-off-by: Mihai Dontu <mihai.dontu@gmail.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
This patch adds search and replace/update functionality
to cache pickup_cb. This change is required to avoid
duplicates getting into the cache during pickup. Also
helps updating existing objects during cache pickup.
We have seen this in cases of ipv6 equal cost multipath
routes
changes v1 to v2:
- Updated documentation
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
The current code does a rtnl_link_put on new object instead of
old object. This patch fixes it. None of the caches have support
for object update, so this should not have affected anyone yet.
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
A co_groups == NULL must enter the loop to trigger the initial
fill of the cache but may never bump the grp pointer as doing
so prevents the loop from being escaped correctly.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
This makes runtime removal of cache operations possible if non-safe
API is not in use by application. The non-safe API will be removed
in the next major version.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
This patch adds support to iter over all supported families
during cache fill and resync.
The motivation for this was previously introduced at
http://lists.infradead.org/pipermail/libnl/2012-November/000734.html
In short, this patch allows caches to request dump on all supported
families instead of only AF_UNSPEC as done today.
With feedback from thomas this patch makes the iter over all families
conditional on per cache flag NL_CACHE_AF_ITER
Signed-off-by: Thomas Graf <tgraf@suug.ch>
This patch adds support for per cache flags
and adds a flag NL_CACHE_AF_ITER to iter over all
supported families when filling the cache.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
This patch adds support to create, delete modify hash table for a cache
Signed-off-by: Shrijeet Mukherjee <shm@cumulusnetworks.com>
Signed-off-by: Nolan Leake <nolan@cumulusnetworks.com>
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Reviewed-by: Wilson Kok <wkok@cumulusnetworks.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
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>
Attached is a patch to fix two problems with dumping objects to a buffer in=
stead of a file descriptor.
One was a problem in detecting the end of the buffer in the newline code.
The other was a problem with clearing the whole buffer before printing each=
object.
- changes the modules hierarchy to better represent the set of libaries
- list the header file that needs to be included
- remove examples/doc from api ref that is included in the guide
- add references to the guide
- fix doxygen api linking for version 1.8.0
- readd doxygen mainpage to config file
- fix a couple of doxygen doc bugs
The function can be used to make a copy of an existing cache. It is very
similar to nl_cache_subset() except that it allows no filtering but
copies every object.
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Thomas Graf <tgraf@redhat.com>
the patch below adds the possibility to
pass user data to callbacks of type
change_func_t when using the nl_cache_mngr_*
family of functions.
If there is any better way to do this,
without duplicating the code in
cache_mngr.c please let me know.
When resyncing a cache, there are no delete messages, so they need to
be synthesized for deleted objects.
Signed-off-by: Patrick McHardy <kaber@trash.net>
The idea of a common handle is long revised and only misleading,
nl_handle really represents a socket with some additional
action handlers assigned to it.
Alias for nl_handle is kept for backwards compatibility.
In order for the interface to become more thread safe, the error
handling was revised to no longer depend on a static errno and
error string buffer.
This patch converts all error paths to return a libnl specific
error code which can be translated to a error message using
nl_geterror(int error). The functions nl_error() and
nl_get_errno() are therefore obsolete.
This change required various sets of function prototypes to be
changed in order to return an error code, the most prominent
are:
struct nl_cache *foo_alloc_cache(...);
changed to:
int foo_alloc_cache(..., struct nl_cache **);
struct nl_msg *foo_build_request(...);
changed to:
int foo_build_request(..., struct nl_msg **);
struct foo *foo_parse(...);
changed to:
int foo_parse(..., struct foo **);
This pretty much only leaves trivial allocation functions to
still return a pointer object which can still return NULL to
signal out of memory.
This change is a serious API and ABI breaker, sorry!
