2006-05-26 07:00:12 +08:00
|
|
|
/*
|
|
|
|
* Bridge netlink control interface
|
|
|
|
*
|
|
|
|
* Authors:
|
|
|
|
* Stephen Hemminger <shemminger@osdl.org>
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public License
|
|
|
|
* as published by the Free Software Foundation; either version
|
|
|
|
* 2 of the License, or (at your option) any later version.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/kernel.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
|
|
|
#include <linux/slab.h>
|
2011-04-04 22:03:32 +08:00
|
|
|
#include <linux/etherdevice.h>
|
2007-03-23 02:59:03 +08:00
|
|
|
#include <net/rtnetlink.h>
|
2007-09-18 02:56:21 +08:00
|
|
|
#include <net/net_namespace.h>
|
2007-11-30 21:21:31 +08:00
|
|
|
#include <net/sock.h>
|
2013-02-13 20:00:12 +08:00
|
|
|
#include <uapi/linux/if_bridge.h>
|
2011-04-04 22:03:32 +08:00
|
|
|
|
2006-05-26 07:00:12 +08:00
|
|
|
#include "br_private.h"
|
2011-11-25 08:16:37 +08:00
|
|
|
#include "br_private_stp.h"
|
2006-05-26 07:00:12 +08:00
|
|
|
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
static int __get_num_vlan_infos(struct net_bridge_vlan_group *vg,
|
2015-10-01 02:16:53 +08:00
|
|
|
u32 filter_mask)
|
2015-02-26 15:55:40 +08:00
|
|
|
{
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
struct net_bridge_vlan *v;
|
|
|
|
u16 vid_range_start = 0, vid_range_end = 0, vid_range_flags = 0;
|
2015-10-01 02:16:53 +08:00
|
|
|
u16 flags, pvid;
|
2015-02-26 15:55:40 +08:00
|
|
|
int num_vlans = 0;
|
|
|
|
|
|
|
|
if (!(filter_mask & RTEXT_FILTER_BRVLAN_COMPRESSED))
|
|
|
|
return 0;
|
|
|
|
|
2015-10-01 02:16:53 +08:00
|
|
|
pvid = br_get_pvid(vg);
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
/* Count number of vlan infos */
|
2015-10-02 21:05:10 +08:00
|
|
|
list_for_each_entry_rcu(v, &vg->vlan_list, vlist) {
|
2015-02-26 15:55:40 +08:00
|
|
|
flags = 0;
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
/* only a context, bridge vlan not activated */
|
|
|
|
if (!br_vlan_should_use(v))
|
|
|
|
continue;
|
|
|
|
if (v->vid == pvid)
|
2015-02-26 15:55:40 +08:00
|
|
|
flags |= BRIDGE_VLAN_INFO_PVID;
|
|
|
|
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (v->flags & BRIDGE_VLAN_INFO_UNTAGGED)
|
2015-02-26 15:55:40 +08:00
|
|
|
flags |= BRIDGE_VLAN_INFO_UNTAGGED;
|
|
|
|
|
|
|
|
if (vid_range_start == 0) {
|
|
|
|
goto initvars;
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
} else if ((v->vid - vid_range_end) == 1 &&
|
2015-02-26 15:55:40 +08:00
|
|
|
flags == vid_range_flags) {
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
vid_range_end = v->vid;
|
2015-02-26 15:55:40 +08:00
|
|
|
continue;
|
|
|
|
} else {
|
|
|
|
if ((vid_range_end - vid_range_start) > 0)
|
|
|
|
num_vlans += 2;
|
|
|
|
else
|
|
|
|
num_vlans += 1;
|
|
|
|
}
|
|
|
|
initvars:
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
vid_range_start = v->vid;
|
|
|
|
vid_range_end = v->vid;
|
2015-02-26 15:55:40 +08:00
|
|
|
vid_range_flags = flags;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (vid_range_start != 0) {
|
|
|
|
if ((vid_range_end - vid_range_start) > 0)
|
|
|
|
num_vlans += 2;
|
|
|
|
else
|
|
|
|
num_vlans += 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return num_vlans;
|
|
|
|
}
|
|
|
|
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
static int br_get_num_vlan_infos(struct net_bridge_vlan_group *vg,
|
2015-10-01 02:16:53 +08:00
|
|
|
u32 filter_mask)
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
{
|
2015-10-02 21:05:10 +08:00
|
|
|
int num_vlans;
|
|
|
|
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (!vg)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (filter_mask & RTEXT_FILTER_BRVLAN)
|
|
|
|
return vg->num_vlans;
|
|
|
|
|
2015-10-02 21:05:10 +08:00
|
|
|
rcu_read_lock();
|
|
|
|
num_vlans = __get_num_vlan_infos(vg, filter_mask);
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
|
|
|
return num_vlans;
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
}
|
|
|
|
|
2015-02-26 15:55:40 +08:00
|
|
|
static size_t br_get_link_af_size_filtered(const struct net_device *dev,
|
|
|
|
u32 filter_mask)
|
2015-02-22 12:21:51 +08:00
|
|
|
{
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
struct net_bridge_vlan_group *vg = NULL;
|
|
|
|
struct net_bridge_port *p;
|
|
|
|
struct net_bridge *br;
|
2015-02-26 15:55:40 +08:00
|
|
|
int num_vlan_infos;
|
2015-02-22 12:21:51 +08:00
|
|
|
|
2015-03-03 23:02:16 +08:00
|
|
|
rcu_read_lock();
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (br_port_exists(dev)) {
|
|
|
|
p = br_port_get_rcu(dev);
|
|
|
|
vg = nbp_vlan_group(p);
|
|
|
|
} else if (dev->priv_flags & IFF_EBRIDGE) {
|
|
|
|
br = netdev_priv(dev);
|
|
|
|
vg = br_vlan_group(br);
|
|
|
|
}
|
2015-10-01 02:16:53 +08:00
|
|
|
num_vlan_infos = br_get_num_vlan_infos(vg, filter_mask);
|
2015-03-03 23:02:16 +08:00
|
|
|
rcu_read_unlock();
|
2015-02-22 12:21:51 +08:00
|
|
|
|
|
|
|
/* Each VLAN is returned in bridge_vlan_info along with flags */
|
2015-02-26 15:55:40 +08:00
|
|
|
return num_vlan_infos * nla_total_size(sizeof(struct bridge_vlan_info));
|
2015-02-22 12:21:51 +08:00
|
|
|
}
|
|
|
|
|
2012-11-13 15:53:05 +08:00
|
|
|
static inline size_t br_port_info_size(void)
|
|
|
|
{
|
|
|
|
return nla_total_size(1) /* IFLA_BRPORT_STATE */
|
|
|
|
+ nla_total_size(2) /* IFLA_BRPORT_PRIORITY */
|
|
|
|
+ nla_total_size(4) /* IFLA_BRPORT_COST */
|
|
|
|
+ nla_total_size(1) /* IFLA_BRPORT_MODE */
|
2012-11-13 15:53:07 +08:00
|
|
|
+ nla_total_size(1) /* IFLA_BRPORT_GUARD */
|
2012-11-13 15:53:08 +08:00
|
|
|
+ nla_total_size(1) /* IFLA_BRPORT_PROTECT */
|
2013-03-11 21:52:17 +08:00
|
|
|
+ nla_total_size(1) /* IFLA_BRPORT_FAST_LEAVE */
|
2013-06-05 22:08:00 +08:00
|
|
|
+ nla_total_size(1) /* IFLA_BRPORT_LEARNING */
|
2013-06-05 22:08:01 +08:00
|
|
|
+ nla_total_size(1) /* IFLA_BRPORT_UNICAST_FLOOD */
|
2015-08-05 01:06:32 +08:00
|
|
|
+ nla_total_size(1) /* IFLA_BRPORT_PROXYARP */
|
2015-08-05 01:06:33 +08:00
|
|
|
+ nla_total_size(1) /* IFLA_BRPORT_PROXYARP_WIFI */
|
2012-11-13 15:53:05 +08:00
|
|
|
+ 0;
|
|
|
|
}
|
|
|
|
|
2015-02-26 15:55:40 +08:00
|
|
|
static inline size_t br_nlmsg_size(struct net_device *dev, u32 filter_mask)
|
2006-11-11 06:10:15 +08:00
|
|
|
{
|
|
|
|
return NLMSG_ALIGN(sizeof(struct ifinfomsg))
|
2012-11-13 15:53:05 +08:00
|
|
|
+ nla_total_size(IFNAMSIZ) /* IFLA_IFNAME */
|
|
|
|
+ nla_total_size(MAX_ADDR_LEN) /* IFLA_ADDRESS */
|
|
|
|
+ nla_total_size(4) /* IFLA_MASTER */
|
|
|
|
+ nla_total_size(4) /* IFLA_MTU */
|
|
|
|
+ nla_total_size(4) /* IFLA_LINK */
|
|
|
|
+ nla_total_size(1) /* IFLA_OPERSTATE */
|
2015-02-22 12:21:51 +08:00
|
|
|
+ nla_total_size(br_port_info_size()) /* IFLA_PROTINFO */
|
2015-02-26 15:55:40 +08:00
|
|
|
+ nla_total_size(br_get_link_af_size_filtered(dev,
|
|
|
|
filter_mask)); /* IFLA_AF_SPEC */
|
2012-11-13 15:53:05 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static int br_port_fill_attrs(struct sk_buff *skb,
|
|
|
|
const struct net_bridge_port *p)
|
|
|
|
{
|
|
|
|
u8 mode = !!(p->flags & BR_HAIRPIN_MODE);
|
|
|
|
|
|
|
|
if (nla_put_u8(skb, IFLA_BRPORT_STATE, p->state) ||
|
|
|
|
nla_put_u16(skb, IFLA_BRPORT_PRIORITY, p->priority) ||
|
|
|
|
nla_put_u32(skb, IFLA_BRPORT_COST, p->path_cost) ||
|
2012-11-13 15:53:07 +08:00
|
|
|
nla_put_u8(skb, IFLA_BRPORT_MODE, mode) ||
|
2012-11-13 15:53:08 +08:00
|
|
|
nla_put_u8(skb, IFLA_BRPORT_GUARD, !!(p->flags & BR_BPDU_GUARD)) ||
|
2012-12-06 05:24:45 +08:00
|
|
|
nla_put_u8(skb, IFLA_BRPORT_PROTECT, !!(p->flags & BR_ROOT_BLOCK)) ||
|
2013-06-05 22:08:00 +08:00
|
|
|
nla_put_u8(skb, IFLA_BRPORT_FAST_LEAVE, !!(p->flags & BR_MULTICAST_FAST_LEAVE)) ||
|
2013-06-05 22:08:01 +08:00
|
|
|
nla_put_u8(skb, IFLA_BRPORT_LEARNING, !!(p->flags & BR_LEARNING)) ||
|
2014-10-24 05:49:17 +08:00
|
|
|
nla_put_u8(skb, IFLA_BRPORT_UNICAST_FLOOD, !!(p->flags & BR_FLOOD)) ||
|
2015-03-04 18:54:21 +08:00
|
|
|
nla_put_u8(skb, IFLA_BRPORT_PROXYARP, !!(p->flags & BR_PROXYARP)) ||
|
|
|
|
nla_put_u8(skb, IFLA_BRPORT_PROXYARP_WIFI,
|
|
|
|
!!(p->flags & BR_PROXYARP_WIFI)))
|
2012-11-13 15:53:05 +08:00
|
|
|
return -EMSGSIZE;
|
|
|
|
|
|
|
|
return 0;
|
2006-11-11 06:10:15 +08:00
|
|
|
}
|
|
|
|
|
2015-01-10 23:31:14 +08:00
|
|
|
static int br_fill_ifvlaninfo_range(struct sk_buff *skb, u16 vid_start,
|
|
|
|
u16 vid_end, u16 flags)
|
|
|
|
{
|
|
|
|
struct bridge_vlan_info vinfo;
|
|
|
|
|
|
|
|
if ((vid_end - vid_start) > 0) {
|
|
|
|
/* add range to skb */
|
|
|
|
vinfo.vid = vid_start;
|
|
|
|
vinfo.flags = flags | BRIDGE_VLAN_INFO_RANGE_BEGIN;
|
|
|
|
if (nla_put(skb, IFLA_BRIDGE_VLAN_INFO,
|
|
|
|
sizeof(vinfo), &vinfo))
|
|
|
|
goto nla_put_failure;
|
|
|
|
|
|
|
|
vinfo.vid = vid_end;
|
|
|
|
vinfo.flags = flags | BRIDGE_VLAN_INFO_RANGE_END;
|
|
|
|
if (nla_put(skb, IFLA_BRIDGE_VLAN_INFO,
|
|
|
|
sizeof(vinfo), &vinfo))
|
|
|
|
goto nla_put_failure;
|
|
|
|
} else {
|
|
|
|
vinfo.vid = vid_start;
|
|
|
|
vinfo.flags = flags;
|
|
|
|
if (nla_put(skb, IFLA_BRIDGE_VLAN_INFO,
|
|
|
|
sizeof(vinfo), &vinfo))
|
|
|
|
goto nla_put_failure;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
nla_put_failure:
|
|
|
|
return -EMSGSIZE;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int br_fill_ifvlaninfo_compressed(struct sk_buff *skb,
|
2015-10-01 02:16:53 +08:00
|
|
|
struct net_bridge_vlan_group *vg)
|
2015-01-10 23:31:14 +08:00
|
|
|
{
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
struct net_bridge_vlan *v;
|
|
|
|
u16 vid_range_start = 0, vid_range_end = 0, vid_range_flags = 0;
|
2015-10-01 02:16:53 +08:00
|
|
|
u16 flags, pvid;
|
2015-01-10 23:31:14 +08:00
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
/* Pack IFLA_BRIDGE_VLAN_INFO's for every vlan
|
|
|
|
* and mark vlan info with begin and end flags
|
|
|
|
* if vlaninfo represents a range
|
|
|
|
*/
|
2015-10-01 02:16:53 +08:00
|
|
|
pvid = br_get_pvid(vg);
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
list_for_each_entry(v, &vg->vlan_list, vlist) {
|
2015-01-10 23:31:14 +08:00
|
|
|
flags = 0;
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (!br_vlan_should_use(v))
|
|
|
|
continue;
|
|
|
|
if (v->vid == pvid)
|
2015-01-10 23:31:14 +08:00
|
|
|
flags |= BRIDGE_VLAN_INFO_PVID;
|
|
|
|
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (v->flags & BRIDGE_VLAN_INFO_UNTAGGED)
|
2015-01-10 23:31:14 +08:00
|
|
|
flags |= BRIDGE_VLAN_INFO_UNTAGGED;
|
|
|
|
|
|
|
|
if (vid_range_start == 0) {
|
|
|
|
goto initvars;
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
} else if ((v->vid - vid_range_end) == 1 &&
|
2015-01-10 23:31:14 +08:00
|
|
|
flags == vid_range_flags) {
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
vid_range_end = v->vid;
|
2015-01-10 23:31:14 +08:00
|
|
|
continue;
|
|
|
|
} else {
|
|
|
|
err = br_fill_ifvlaninfo_range(skb, vid_range_start,
|
|
|
|
vid_range_end,
|
|
|
|
vid_range_flags);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
initvars:
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
vid_range_start = v->vid;
|
|
|
|
vid_range_end = v->vid;
|
2015-01-10 23:31:14 +08:00
|
|
|
vid_range_flags = flags;
|
|
|
|
}
|
|
|
|
|
2015-01-13 08:25:28 +08:00
|
|
|
if (vid_range_start != 0) {
|
|
|
|
/* Call it once more to send any left over vlans */
|
|
|
|
err = br_fill_ifvlaninfo_range(skb, vid_range_start,
|
|
|
|
vid_range_end,
|
|
|
|
vid_range_flags);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
2015-01-10 23:31:14 +08:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int br_fill_ifvlaninfo(struct sk_buff *skb,
|
2015-10-01 02:16:53 +08:00
|
|
|
struct net_bridge_vlan_group *vg)
|
2015-01-10 23:31:14 +08:00
|
|
|
{
|
|
|
|
struct bridge_vlan_info vinfo;
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
struct net_bridge_vlan *v;
|
2015-10-01 02:16:53 +08:00
|
|
|
u16 pvid;
|
2015-01-10 23:31:14 +08:00
|
|
|
|
2015-10-01 02:16:53 +08:00
|
|
|
pvid = br_get_pvid(vg);
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
list_for_each_entry(v, &vg->vlan_list, vlist) {
|
|
|
|
if (!br_vlan_should_use(v))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
vinfo.vid = v->vid;
|
2015-01-10 23:31:14 +08:00
|
|
|
vinfo.flags = 0;
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (v->vid == pvid)
|
2015-01-10 23:31:14 +08:00
|
|
|
vinfo.flags |= BRIDGE_VLAN_INFO_PVID;
|
|
|
|
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (v->flags & BRIDGE_VLAN_INFO_UNTAGGED)
|
2015-01-10 23:31:14 +08:00
|
|
|
vinfo.flags |= BRIDGE_VLAN_INFO_UNTAGGED;
|
|
|
|
|
|
|
|
if (nla_put(skb, IFLA_BRIDGE_VLAN_INFO,
|
|
|
|
sizeof(vinfo), &vinfo))
|
|
|
|
goto nla_put_failure;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
nla_put_failure:
|
|
|
|
return -EMSGSIZE;
|
|
|
|
}
|
|
|
|
|
2006-05-26 07:00:12 +08:00
|
|
|
/*
|
|
|
|
* Create one netlink message for one interface
|
|
|
|
* Contains port and master info as well as carrier and bridge state.
|
|
|
|
*/
|
2013-02-13 20:00:13 +08:00
|
|
|
static int br_fill_ifinfo(struct sk_buff *skb,
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
struct net_bridge_port *port,
|
2013-02-13 20:00:13 +08:00
|
|
|
u32 pid, u32 seq, int event, unsigned int flags,
|
|
|
|
u32 filter_mask, const struct net_device *dev)
|
2006-05-26 07:00:12 +08:00
|
|
|
{
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
struct net_bridge *br;
|
2006-11-21 08:20:22 +08:00
|
|
|
struct ifinfomsg *hdr;
|
2006-05-26 07:00:12 +08:00
|
|
|
struct nlmsghdr *nlh;
|
|
|
|
u8 operstate = netif_running(dev) ? dev->operstate : IF_OPER_DOWN;
|
|
|
|
|
2013-02-13 20:00:13 +08:00
|
|
|
if (port)
|
|
|
|
br = port->br;
|
|
|
|
else
|
|
|
|
br = netdev_priv(dev);
|
|
|
|
|
2010-05-10 17:31:09 +08:00
|
|
|
br_debug(br, "br_fill_info event %d port %s master %s\n",
|
|
|
|
event, dev->name, br->dev->name);
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2006-11-21 08:20:22 +08:00
|
|
|
nlh = nlmsg_put(skb, pid, seq, event, sizeof(*hdr), flags);
|
|
|
|
if (nlh == NULL)
|
2007-02-01 15:16:40 +08:00
|
|
|
return -EMSGSIZE;
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2006-11-21 08:20:22 +08:00
|
|
|
hdr = nlmsg_data(nlh);
|
|
|
|
hdr->ifi_family = AF_BRIDGE;
|
|
|
|
hdr->__ifi_pad = 0;
|
|
|
|
hdr->ifi_type = dev->type;
|
|
|
|
hdr->ifi_index = dev->ifindex;
|
|
|
|
hdr->ifi_flags = dev_get_flags(dev);
|
|
|
|
hdr->ifi_change = 0;
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2012-04-02 08:49:54 +08:00
|
|
|
if (nla_put_string(skb, IFLA_IFNAME, dev->name) ||
|
|
|
|
nla_put_u32(skb, IFLA_MASTER, br->dev->ifindex) ||
|
|
|
|
nla_put_u32(skb, IFLA_MTU, dev->mtu) ||
|
|
|
|
nla_put_u8(skb, IFLA_OPERSTATE, operstate) ||
|
|
|
|
(dev->addr_len &&
|
|
|
|
nla_put(skb, IFLA_ADDRESS, dev->addr_len, dev->dev_addr)) ||
|
2015-04-02 23:07:00 +08:00
|
|
|
(dev->ifindex != dev_get_iflink(dev) &&
|
|
|
|
nla_put_u32(skb, IFLA_LINK, dev_get_iflink(dev))))
|
2012-04-02 08:49:54 +08:00
|
|
|
goto nla_put_failure;
|
2012-11-13 15:53:05 +08:00
|
|
|
|
2013-02-13 20:00:13 +08:00
|
|
|
if (event == RTM_NEWLINK && port) {
|
2012-11-13 15:53:05 +08:00
|
|
|
struct nlattr *nest
|
|
|
|
= nla_nest_start(skb, IFLA_PROTINFO | NLA_F_NESTED);
|
|
|
|
|
|
|
|
if (nest == NULL || br_port_fill_attrs(skb, port) < 0)
|
|
|
|
goto nla_put_failure;
|
|
|
|
nla_nest_end(skb, nest);
|
|
|
|
}
|
|
|
|
|
2013-02-13 20:00:13 +08:00
|
|
|
/* Check if the VID information is requested */
|
2015-01-10 23:31:14 +08:00
|
|
|
if ((filter_mask & RTEXT_FILTER_BRVLAN) ||
|
|
|
|
(filter_mask & RTEXT_FILTER_BRVLAN_COMPRESSED)) {
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
struct net_bridge_vlan_group *vg;
|
2015-01-10 23:31:14 +08:00
|
|
|
struct nlattr *af;
|
|
|
|
int err;
|
2013-02-13 20:00:13 +08:00
|
|
|
|
2015-10-01 02:16:53 +08:00
|
|
|
if (port)
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
vg = nbp_vlan_group(port);
|
2015-10-01 02:16:53 +08:00
|
|
|
else
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
vg = br_vlan_group(br);
|
2013-02-13 20:00:13 +08:00
|
|
|
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (!vg || !vg->num_vlans)
|
2013-02-13 20:00:13 +08:00
|
|
|
goto done;
|
|
|
|
|
|
|
|
af = nla_nest_start(skb, IFLA_AF_SPEC);
|
|
|
|
if (!af)
|
|
|
|
goto nla_put_failure;
|
|
|
|
|
2015-01-10 23:31:14 +08:00
|
|
|
if (filter_mask & RTEXT_FILTER_BRVLAN_COMPRESSED)
|
2015-10-01 02:16:53 +08:00
|
|
|
err = br_fill_ifvlaninfo_compressed(skb, vg);
|
2015-01-10 23:31:14 +08:00
|
|
|
else
|
2015-10-01 02:16:53 +08:00
|
|
|
err = br_fill_ifvlaninfo(skb, vg);
|
2015-01-10 23:31:14 +08:00
|
|
|
if (err)
|
|
|
|
goto nla_put_failure;
|
2013-02-13 20:00:13 +08:00
|
|
|
nla_nest_end(skb, af);
|
|
|
|
}
|
|
|
|
|
|
|
|
done:
|
2015-01-17 05:09:00 +08:00
|
|
|
nlmsg_end(skb, nlh);
|
|
|
|
return 0;
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2006-11-21 08:20:22 +08:00
|
|
|
nla_put_failure:
|
2007-02-01 15:16:40 +08:00
|
|
|
nlmsg_cancel(skb, nlh);
|
|
|
|
return -EMSGSIZE;
|
2006-05-26 07:00:12 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Notify listeners of a change in port information
|
|
|
|
*/
|
|
|
|
void br_ifinfo_notify(int event, struct net_bridge_port *port)
|
|
|
|
{
|
2013-02-13 20:00:12 +08:00
|
|
|
struct net *net;
|
2006-05-26 07:00:12 +08:00
|
|
|
struct sk_buff *skb;
|
2006-08-15 15:36:28 +08:00
|
|
|
int err = -ENOBUFS;
|
2015-02-26 15:55:40 +08:00
|
|
|
u32 filter = RTEXT_FILTER_BRVLAN_COMPRESSED;
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2013-02-13 20:00:12 +08:00
|
|
|
if (!port)
|
|
|
|
return;
|
|
|
|
|
|
|
|
net = dev_net(port->dev);
|
2010-05-10 17:31:09 +08:00
|
|
|
br_debug(port->br, "port %u(%s) event %d\n",
|
2012-04-15 13:58:06 +08:00
|
|
|
(unsigned int)port->port_no, port->dev->name, event);
|
2010-05-10 17:31:09 +08:00
|
|
|
|
2015-02-26 15:55:40 +08:00
|
|
|
skb = nlmsg_new(br_nlmsg_size(port->dev, filter), GFP_ATOMIC);
|
2006-08-15 15:36:28 +08:00
|
|
|
if (skb == NULL)
|
|
|
|
goto errout;
|
|
|
|
|
2015-02-26 15:55:40 +08:00
|
|
|
err = br_fill_ifinfo(skb, port, 0, 0, event, 0, filter, port->dev);
|
2007-02-01 15:16:40 +08:00
|
|
|
if (err < 0) {
|
|
|
|
/* -EMSGSIZE implies BUG in br_nlmsg_size() */
|
|
|
|
WARN_ON(err == -EMSGSIZE);
|
|
|
|
kfree_skb(skb);
|
|
|
|
goto errout;
|
|
|
|
}
|
2009-02-25 15:18:28 +08:00
|
|
|
rtnl_notify(skb, net, 0, RTNLGRP_LINK, NULL, GFP_ATOMIC);
|
|
|
|
return;
|
2006-08-15 15:36:28 +08:00
|
|
|
errout:
|
2013-12-19 13:28:10 +08:00
|
|
|
rtnl_set_sk_err(net, RTNLGRP_LINK, err);
|
2006-05-26 07:00:12 +08:00
|
|
|
}
|
|
|
|
|
2013-02-13 20:00:12 +08:00
|
|
|
|
2006-05-26 07:00:12 +08:00
|
|
|
/*
|
|
|
|
* Dump information about all ports, in response to GETLINK
|
|
|
|
*/
|
2012-10-24 16:12:57 +08:00
|
|
|
int br_getlink(struct sk_buff *skb, u32 pid, u32 seq,
|
2015-04-29 00:33:49 +08:00
|
|
|
struct net_device *dev, u32 filter_mask, int nlflags)
|
2006-05-26 07:00:12 +08:00
|
|
|
{
|
2013-09-14 22:42:27 +08:00
|
|
|
struct net_bridge_port *port = br_port_get_rtnl(dev);
|
2012-10-24 16:12:57 +08:00
|
|
|
|
2015-01-10 23:31:14 +08:00
|
|
|
if (!port && !(filter_mask & RTEXT_FILTER_BRVLAN) &&
|
|
|
|
!(filter_mask & RTEXT_FILTER_BRVLAN_COMPRESSED))
|
2015-01-21 17:22:35 +08:00
|
|
|
return 0;
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2015-04-29 00:33:49 +08:00
|
|
|
return br_fill_ifinfo(skb, port, pid, seq, RTM_NEWLINK, nlflags,
|
2015-01-21 17:22:35 +08:00
|
|
|
filter_mask, dev);
|
2006-05-26 07:00:12 +08:00
|
|
|
}
|
|
|
|
|
2015-01-10 23:31:12 +08:00
|
|
|
static int br_vlan_info(struct net_bridge *br, struct net_bridge_port *p,
|
|
|
|
int cmd, struct bridge_vlan_info *vinfo)
|
|
|
|
{
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
switch (cmd) {
|
|
|
|
case RTM_SETLINK:
|
|
|
|
if (p) {
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
/* if the MASTER flag is set this will act on the global
|
|
|
|
* per-VLAN entry as well
|
|
|
|
*/
|
2015-01-10 23:31:12 +08:00
|
|
|
err = nbp_vlan_add(p, vinfo->vid, vinfo->flags);
|
|
|
|
if (err)
|
|
|
|
break;
|
|
|
|
} else {
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
vinfo->flags |= BRIDGE_VLAN_INFO_BRENTRY;
|
2015-01-10 23:31:12 +08:00
|
|
|
err = br_vlan_add(br, vinfo->vid, vinfo->flags);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
case RTM_DELLINK:
|
|
|
|
if (p) {
|
|
|
|
nbp_vlan_delete(p, vinfo->vid);
|
|
|
|
if (vinfo->flags & BRIDGE_VLAN_INFO_MASTER)
|
|
|
|
br_vlan_delete(p->br, vinfo->vid);
|
|
|
|
} else {
|
|
|
|
br_vlan_delete(br, vinfo->vid);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
2013-02-13 20:00:12 +08:00
|
|
|
|
|
|
|
static int br_afspec(struct net_bridge *br,
|
|
|
|
struct net_bridge_port *p,
|
|
|
|
struct nlattr *af_spec,
|
|
|
|
int cmd)
|
|
|
|
{
|
2015-01-10 23:31:12 +08:00
|
|
|
struct bridge_vlan_info *vinfo_start = NULL;
|
|
|
|
struct bridge_vlan_info *vinfo = NULL;
|
|
|
|
struct nlattr *attr;
|
2013-02-13 20:00:12 +08:00
|
|
|
int err = 0;
|
2015-01-10 23:31:12 +08:00
|
|
|
int rem;
|
2013-02-13 20:00:12 +08:00
|
|
|
|
2015-01-10 23:31:12 +08:00
|
|
|
nla_for_each_nested(attr, af_spec, rem) {
|
|
|
|
if (nla_type(attr) != IFLA_BRIDGE_VLAN_INFO)
|
|
|
|
continue;
|
|
|
|
if (nla_len(attr) != sizeof(struct bridge_vlan_info))
|
|
|
|
return -EINVAL;
|
|
|
|
vinfo = nla_data(attr);
|
2015-07-02 20:48:17 +08:00
|
|
|
if (!vinfo->vid || vinfo->vid >= VLAN_VID_MASK)
|
|
|
|
return -EINVAL;
|
2015-01-10 23:31:12 +08:00
|
|
|
if (vinfo->flags & BRIDGE_VLAN_INFO_RANGE_BEGIN) {
|
|
|
|
if (vinfo_start)
|
|
|
|
return -EINVAL;
|
|
|
|
vinfo_start = vinfo;
|
|
|
|
continue;
|
|
|
|
}
|
2013-02-13 20:00:12 +08:00
|
|
|
|
2015-01-10 23:31:12 +08:00
|
|
|
if (vinfo_start) {
|
|
|
|
struct bridge_vlan_info tmp_vinfo;
|
|
|
|
int v;
|
2013-02-13 20:00:12 +08:00
|
|
|
|
2015-01-10 23:31:12 +08:00
|
|
|
if (!(vinfo->flags & BRIDGE_VLAN_INFO_RANGE_END))
|
|
|
|
return -EINVAL;
|
2013-02-13 20:00:12 +08:00
|
|
|
|
2015-01-10 23:31:12 +08:00
|
|
|
if (vinfo->vid <= vinfo_start->vid)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
memcpy(&tmp_vinfo, vinfo_start,
|
|
|
|
sizeof(struct bridge_vlan_info));
|
2013-02-13 20:00:12 +08:00
|
|
|
|
2015-01-10 23:31:12 +08:00
|
|
|
for (v = vinfo_start->vid; v <= vinfo->vid; v++) {
|
|
|
|
tmp_vinfo.vid = v;
|
|
|
|
err = br_vlan_info(br, p, cmd, &tmp_vinfo);
|
2013-02-13 20:00:12 +08:00
|
|
|
if (err)
|
|
|
|
break;
|
2015-01-10 23:31:12 +08:00
|
|
|
}
|
|
|
|
vinfo_start = NULL;
|
|
|
|
} else {
|
|
|
|
err = br_vlan_info(br, p, cmd, vinfo);
|
2013-02-13 20:00:12 +08:00
|
|
|
}
|
2015-01-10 23:31:12 +08:00
|
|
|
if (err)
|
|
|
|
break;
|
2013-02-13 20:00:12 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2014-09-05 21:51:30 +08:00
|
|
|
static const struct nla_policy br_port_policy[IFLA_BRPORT_MAX + 1] = {
|
2012-11-13 15:53:05 +08:00
|
|
|
[IFLA_BRPORT_STATE] = { .type = NLA_U8 },
|
|
|
|
[IFLA_BRPORT_COST] = { .type = NLA_U32 },
|
|
|
|
[IFLA_BRPORT_PRIORITY] = { .type = NLA_U16 },
|
|
|
|
[IFLA_BRPORT_MODE] = { .type = NLA_U8 },
|
2012-11-13 15:53:07 +08:00
|
|
|
[IFLA_BRPORT_GUARD] = { .type = NLA_U8 },
|
2012-11-13 15:53:08 +08:00
|
|
|
[IFLA_BRPORT_PROTECT] = { .type = NLA_U8 },
|
2014-11-26 20:42:19 +08:00
|
|
|
[IFLA_BRPORT_FAST_LEAVE]= { .type = NLA_U8 },
|
2013-06-05 22:08:00 +08:00
|
|
|
[IFLA_BRPORT_LEARNING] = { .type = NLA_U8 },
|
2013-06-05 22:08:01 +08:00
|
|
|
[IFLA_BRPORT_UNICAST_FLOOD] = { .type = NLA_U8 },
|
2015-08-05 01:06:32 +08:00
|
|
|
[IFLA_BRPORT_PROXYARP] = { .type = NLA_U8 },
|
2015-08-05 01:06:33 +08:00
|
|
|
[IFLA_BRPORT_PROXYARP_WIFI] = { .type = NLA_U8 },
|
2012-11-13 15:53:05 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
/* Change the state of the port and notify spanning tree */
|
|
|
|
static int br_set_port_state(struct net_bridge_port *p, u8 state)
|
|
|
|
{
|
|
|
|
if (state > BR_STATE_BLOCKING)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/* if kernel STP is running, don't allow changes */
|
|
|
|
if (p->br->stp_enabled == BR_KERNEL_STP)
|
|
|
|
return -EBUSY;
|
|
|
|
|
2012-12-29 02:15:22 +08:00
|
|
|
/* if device is not up, change is not allowed
|
|
|
|
* if link is not present, only allowable state is disabled
|
|
|
|
*/
|
2012-11-13 15:53:05 +08:00
|
|
|
if (!netif_running(p->dev) ||
|
2012-12-29 02:15:22 +08:00
|
|
|
(!netif_oper_up(p->dev) && state != BR_STATE_DISABLED))
|
2012-11-13 15:53:05 +08:00
|
|
|
return -ENETDOWN;
|
|
|
|
|
2014-10-01 07:13:19 +08:00
|
|
|
br_set_state(p, state);
|
2012-11-13 15:53:05 +08:00
|
|
|
br_log_state(p);
|
|
|
|
br_port_state_selection(p->br);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Set/clear or port flags based on attribute */
|
|
|
|
static void br_set_port_flag(struct net_bridge_port *p, struct nlattr *tb[],
|
|
|
|
int attrtype, unsigned long mask)
|
|
|
|
{
|
|
|
|
if (tb[attrtype]) {
|
|
|
|
u8 flag = nla_get_u8(tb[attrtype]);
|
|
|
|
if (flag)
|
|
|
|
p->flags |= mask;
|
|
|
|
else
|
|
|
|
p->flags &= ~mask;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Process bridge protocol info on port */
|
|
|
|
static int br_setport(struct net_bridge_port *p, struct nlattr *tb[])
|
|
|
|
{
|
|
|
|
int err;
|
2014-05-16 21:59:16 +08:00
|
|
|
unsigned long old_flags = p->flags;
|
2012-11-13 15:53:05 +08:00
|
|
|
|
|
|
|
br_set_port_flag(p, tb, IFLA_BRPORT_MODE, BR_HAIRPIN_MODE);
|
2012-11-13 15:53:07 +08:00
|
|
|
br_set_port_flag(p, tb, IFLA_BRPORT_GUARD, BR_BPDU_GUARD);
|
2012-12-06 05:24:45 +08:00
|
|
|
br_set_port_flag(p, tb, IFLA_BRPORT_FAST_LEAVE, BR_MULTICAST_FAST_LEAVE);
|
2013-03-15 14:39:12 +08:00
|
|
|
br_set_port_flag(p, tb, IFLA_BRPORT_PROTECT, BR_ROOT_BLOCK);
|
2013-06-05 22:08:00 +08:00
|
|
|
br_set_port_flag(p, tb, IFLA_BRPORT_LEARNING, BR_LEARNING);
|
2013-06-05 22:08:01 +08:00
|
|
|
br_set_port_flag(p, tb, IFLA_BRPORT_UNICAST_FLOOD, BR_FLOOD);
|
2014-10-24 05:49:17 +08:00
|
|
|
br_set_port_flag(p, tb, IFLA_BRPORT_PROXYARP, BR_PROXYARP);
|
2015-03-04 18:54:21 +08:00
|
|
|
br_set_port_flag(p, tb, IFLA_BRPORT_PROXYARP_WIFI, BR_PROXYARP_WIFI);
|
2012-11-13 15:53:05 +08:00
|
|
|
|
|
|
|
if (tb[IFLA_BRPORT_COST]) {
|
|
|
|
err = br_stp_set_path_cost(p, nla_get_u32(tb[IFLA_BRPORT_COST]));
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (tb[IFLA_BRPORT_PRIORITY]) {
|
|
|
|
err = br_stp_set_port_priority(p, nla_get_u16(tb[IFLA_BRPORT_PRIORITY]));
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (tb[IFLA_BRPORT_STATE]) {
|
|
|
|
err = br_set_port_state(p, nla_get_u8(tb[IFLA_BRPORT_STATE]));
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
2014-05-16 21:59:16 +08:00
|
|
|
|
|
|
|
br_port_flags_change(p, old_flags ^ p->flags);
|
2012-11-13 15:53:05 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Change state and parameters on port. */
|
2015-01-30 14:40:12 +08:00
|
|
|
int br_setlink(struct net_device *dev, struct nlmsghdr *nlh, u16 flags)
|
2006-05-26 07:00:12 +08:00
|
|
|
{
|
2006-11-21 08:20:22 +08:00
|
|
|
struct nlattr *protinfo;
|
2013-02-13 20:00:12 +08:00
|
|
|
struct nlattr *afspec;
|
2006-05-26 07:00:12 +08:00
|
|
|
struct net_bridge_port *p;
|
2012-12-07 09:10:46 +08:00
|
|
|
struct nlattr *tb[IFLA_BRPORT_MAX + 1];
|
2015-05-11 00:47:59 +08:00
|
|
|
int err = 0;
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2013-03-28 14:21:22 +08:00
|
|
|
protinfo = nlmsg_find_attr(nlh, sizeof(struct ifinfomsg), IFLA_PROTINFO);
|
|
|
|
afspec = nlmsg_find_attr(nlh, sizeof(struct ifinfomsg), IFLA_AF_SPEC);
|
2013-02-13 20:00:12 +08:00
|
|
|
if (!protinfo && !afspec)
|
2012-11-13 15:53:05 +08:00
|
|
|
return 0;
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2010-11-15 14:38:14 +08:00
|
|
|
p = br_port_get_rtnl(dev);
|
2013-02-13 20:00:12 +08:00
|
|
|
/* We want to accept dev as bridge itself if the AF_SPEC
|
2013-12-09 04:15:44 +08:00
|
|
|
* is set to see if someone is setting vlan info on the bridge
|
2013-02-13 20:00:12 +08:00
|
|
|
*/
|
2013-03-24 11:26:47 +08:00
|
|
|
if (!p && !afspec)
|
2010-11-15 14:38:13 +08:00
|
|
|
return -EINVAL;
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2013-02-13 20:00:12 +08:00
|
|
|
if (p && protinfo) {
|
|
|
|
if (protinfo->nla_type & NLA_F_NESTED) {
|
|
|
|
err = nla_parse_nested(tb, IFLA_BRPORT_MAX,
|
2014-09-05 21:51:30 +08:00
|
|
|
protinfo, br_port_policy);
|
2013-02-13 20:00:12 +08:00
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
spin_lock_bh(&p->br->lock);
|
|
|
|
err = br_setport(p, tb);
|
|
|
|
spin_unlock_bh(&p->br->lock);
|
|
|
|
} else {
|
2013-12-09 04:15:44 +08:00
|
|
|
/* Binary compatibility with old RSTP */
|
2013-02-13 20:00:12 +08:00
|
|
|
if (nla_len(protinfo) < sizeof(u8))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
spin_lock_bh(&p->br->lock);
|
|
|
|
err = br_set_port_state(p, nla_get_u8(protinfo));
|
|
|
|
spin_unlock_bh(&p->br->lock);
|
|
|
|
}
|
2012-11-13 15:53:05 +08:00
|
|
|
if (err)
|
2013-02-13 20:00:12 +08:00
|
|
|
goto out;
|
|
|
|
}
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2013-02-13 20:00:12 +08:00
|
|
|
if (afspec) {
|
|
|
|
err = br_afspec((struct net_bridge *)netdev_priv(dev), p,
|
|
|
|
afspec, RTM_SETLINK);
|
2012-11-13 15:53:05 +08:00
|
|
|
}
|
2011-11-25 08:16:37 +08:00
|
|
|
|
2012-11-13 15:53:05 +08:00
|
|
|
if (err == 0)
|
|
|
|
br_ifinfo_notify(RTM_NEWLINK, p);
|
2013-02-13 20:00:12 +08:00
|
|
|
out:
|
2012-11-13 15:53:05 +08:00
|
|
|
return err;
|
2006-05-26 07:00:12 +08:00
|
|
|
}
|
|
|
|
|
2013-02-13 20:00:12 +08:00
|
|
|
/* Delete port information */
|
2015-01-30 14:40:12 +08:00
|
|
|
int br_dellink(struct net_device *dev, struct nlmsghdr *nlh, u16 flags)
|
2013-02-13 20:00:12 +08:00
|
|
|
{
|
|
|
|
struct nlattr *afspec;
|
|
|
|
struct net_bridge_port *p;
|
2015-05-11 00:48:03 +08:00
|
|
|
int err = 0;
|
2013-02-13 20:00:12 +08:00
|
|
|
|
2013-03-28 14:21:22 +08:00
|
|
|
afspec = nlmsg_find_attr(nlh, sizeof(struct ifinfomsg), IFLA_AF_SPEC);
|
2013-02-13 20:00:12 +08:00
|
|
|
if (!afspec)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
p = br_port_get_rtnl(dev);
|
|
|
|
/* We want to accept dev as bridge itself as well */
|
|
|
|
if (!p && !(dev->priv_flags & IFF_EBRIDGE))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
err = br_afspec((struct net_bridge *)netdev_priv(dev), p,
|
|
|
|
afspec, RTM_DELLINK);
|
bridge: fix setlink/dellink notifications
problems with bridge getlink/setlink notifications today:
- bridge setlink generates two notifications to userspace
- one from the bridge driver
- one from rtnetlink.c (rtnl_bridge_notify)
- dellink generates one notification from rtnetlink.c. Which
means bridge setlink and dellink notifications are not
consistent
- Looking at the code it appears,
If both BRIDGE_FLAGS_MASTER and BRIDGE_FLAGS_SELF were set,
the size calculation in rtnl_bridge_notify can be wrong.
Example: if you set both BRIDGE_FLAGS_MASTER and BRIDGE_FLAGS_SELF
in a setlink request to rocker dev, rtnl_bridge_notify will
allocate skb for one set of bridge attributes, but,
both the bridge driver and rocker dev will try to add
attributes resulting in twice the number of attributes
being added to the skb. (rocker dev calls ndo_dflt_bridge_getlink)
There are multiple options:
1) Generate one notification including all attributes from master and self:
But, I don't think it will work, because both master and self may use
the same attributes/policy. Cannot pack the same set of attributes in a
single notification from both master and slave (duplicate attributes).
2) Generate one notification from master and the other notification from
self (This seems to be ideal):
For master: the master driver will send notification (bridge in this
example)
For self: the self driver will send notification (rocker in the above
example. It can use helpers from rtnetlink.c to do so. Like the
ndo_dflt_bridge_getlink api).
This patch implements 2) (leaving the 'rtnl_bridge_notify' around to be used
with 'self').
v1->v2 :
- rtnl_bridge_notify is now called only for self,
so, remove 'BRIDGE_FLAGS_SELF' check and cleanup a few things
- rtnl_bridge_dellink used to always send a RTM_NEWLINK msg
earlier. So, I have changed the notification from br_dellink to
go as RTM_NEWLINK
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-15 12:02:25 +08:00
|
|
|
if (err == 0)
|
|
|
|
/* Send RTM_NEWLINK because userspace
|
|
|
|
* expects RTM_NEWLINK for vlan dels
|
|
|
|
*/
|
|
|
|
br_ifinfo_notify(RTM_NEWLINK, p);
|
2013-02-13 20:00:12 +08:00
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
2011-04-04 22:03:32 +08:00
|
|
|
static int br_validate(struct nlattr *tb[], struct nlattr *data[])
|
|
|
|
{
|
|
|
|
if (tb[IFLA_ADDRESS]) {
|
|
|
|
if (nla_len(tb[IFLA_ADDRESS]) != ETH_ALEN)
|
|
|
|
return -EINVAL;
|
|
|
|
if (!is_valid_ether_addr(nla_data(tb[IFLA_ADDRESS])))
|
|
|
|
return -EADDRNOTAVAIL;
|
|
|
|
}
|
|
|
|
|
2015-08-27 14:32:26 +08:00
|
|
|
if (!data)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
#ifdef CONFIG_BRIDGE_VLAN_FILTERING
|
|
|
|
if (data[IFLA_BR_VLAN_PROTOCOL]) {
|
|
|
|
switch (nla_get_be16(data[IFLA_BR_VLAN_PROTOCOL])) {
|
|
|
|
case htons(ETH_P_8021Q):
|
|
|
|
case htons(ETH_P_8021AD):
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return -EPROTONOSUPPORT;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2011-04-04 22:03:32 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-04-25 16:01:18 +08:00
|
|
|
static int br_dev_newlink(struct net *src_net, struct net_device *dev,
|
|
|
|
struct nlattr *tb[], struct nlattr *data[])
|
|
|
|
{
|
|
|
|
struct net_bridge *br = netdev_priv(dev);
|
|
|
|
|
|
|
|
if (tb[IFLA_ADDRESS]) {
|
|
|
|
spin_lock_bh(&br->lock);
|
|
|
|
br_stp_change_bridge_id(br, nla_data(tb[IFLA_ADDRESS]));
|
|
|
|
spin_unlock_bh(&br->lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
return register_netdevice(dev);
|
|
|
|
}
|
|
|
|
|
2014-09-05 21:51:30 +08:00
|
|
|
static int br_port_slave_changelink(struct net_device *brdev,
|
|
|
|
struct net_device *dev,
|
|
|
|
struct nlattr *tb[],
|
|
|
|
struct nlattr *data[])
|
|
|
|
{
|
2015-07-22 19:03:40 +08:00
|
|
|
struct net_bridge *br = netdev_priv(brdev);
|
|
|
|
int ret;
|
|
|
|
|
2014-09-05 21:51:30 +08:00
|
|
|
if (!data)
|
|
|
|
return 0;
|
2015-07-22 19:03:40 +08:00
|
|
|
|
|
|
|
spin_lock_bh(&br->lock);
|
|
|
|
ret = br_setport(br_port_get_rtnl(dev), data);
|
|
|
|
spin_unlock_bh(&br->lock);
|
|
|
|
|
|
|
|
return ret;
|
2014-09-05 21:51:30 +08:00
|
|
|
}
|
|
|
|
|
2014-09-05 21:51:29 +08:00
|
|
|
static int br_port_fill_slave_info(struct sk_buff *skb,
|
|
|
|
const struct net_device *brdev,
|
|
|
|
const struct net_device *dev)
|
|
|
|
{
|
|
|
|
return br_port_fill_attrs(skb, br_port_get_rtnl(dev));
|
|
|
|
}
|
|
|
|
|
|
|
|
static size_t br_port_get_slave_size(const struct net_device *brdev,
|
|
|
|
const struct net_device *dev)
|
|
|
|
{
|
|
|
|
return br_port_info_size();
|
|
|
|
}
|
|
|
|
|
2014-09-05 21:51:32 +08:00
|
|
|
static const struct nla_policy br_policy[IFLA_BR_MAX + 1] = {
|
|
|
|
[IFLA_BR_FORWARD_DELAY] = { .type = NLA_U32 },
|
|
|
|
[IFLA_BR_HELLO_TIME] = { .type = NLA_U32 },
|
|
|
|
[IFLA_BR_MAX_AGE] = { .type = NLA_U32 },
|
2015-03-18 17:06:58 +08:00
|
|
|
[IFLA_BR_AGEING_TIME] = { .type = NLA_U32 },
|
|
|
|
[IFLA_BR_STP_STATE] = { .type = NLA_U32 },
|
|
|
|
[IFLA_BR_PRIORITY] = { .type = NLA_U16 },
|
2015-08-08 00:40:45 +08:00
|
|
|
[IFLA_BR_VLAN_FILTERING] = { .type = NLA_U8 },
|
2015-08-27 14:32:26 +08:00
|
|
|
[IFLA_BR_VLAN_PROTOCOL] = { .type = NLA_U16 },
|
2015-10-04 20:23:28 +08:00
|
|
|
[IFLA_BR_GROUP_FWD_MASK] = { .type = NLA_U16 },
|
2015-10-04 20:23:35 +08:00
|
|
|
[IFLA_BR_GROUP_ADDR] = { .type = NLA_BINARY,
|
|
|
|
.len = ETH_ALEN },
|
2015-10-04 20:23:37 +08:00
|
|
|
[IFLA_BR_MCAST_ROUTER] = { .type = NLA_U8 },
|
2014-09-05 21:51:32 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static int br_changelink(struct net_device *brdev, struct nlattr *tb[],
|
|
|
|
struct nlattr *data[])
|
|
|
|
{
|
|
|
|
struct net_bridge *br = netdev_priv(brdev);
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (!data)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (data[IFLA_BR_FORWARD_DELAY]) {
|
|
|
|
err = br_set_forward_delay(br, nla_get_u32(data[IFLA_BR_FORWARD_DELAY]));
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (data[IFLA_BR_HELLO_TIME]) {
|
|
|
|
err = br_set_hello_time(br, nla_get_u32(data[IFLA_BR_HELLO_TIME]));
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (data[IFLA_BR_MAX_AGE]) {
|
|
|
|
err = br_set_max_age(br, nla_get_u32(data[IFLA_BR_MAX_AGE]));
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2015-03-18 17:06:58 +08:00
|
|
|
if (data[IFLA_BR_AGEING_TIME]) {
|
|
|
|
u32 ageing_time = nla_get_u32(data[IFLA_BR_AGEING_TIME]);
|
|
|
|
|
|
|
|
br->ageing_time = clock_t_to_jiffies(ageing_time);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (data[IFLA_BR_STP_STATE]) {
|
|
|
|
u32 stp_enabled = nla_get_u32(data[IFLA_BR_STP_STATE]);
|
|
|
|
|
|
|
|
br_stp_set_enabled(br, stp_enabled);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (data[IFLA_BR_PRIORITY]) {
|
|
|
|
u32 priority = nla_get_u16(data[IFLA_BR_PRIORITY]);
|
|
|
|
|
|
|
|
br_stp_set_bridge_priority(br, priority);
|
|
|
|
}
|
|
|
|
|
2015-08-08 00:40:45 +08:00
|
|
|
if (data[IFLA_BR_VLAN_FILTERING]) {
|
|
|
|
u8 vlan_filter = nla_get_u8(data[IFLA_BR_VLAN_FILTERING]);
|
|
|
|
|
|
|
|
err = __br_vlan_filter_toggle(br, vlan_filter);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2015-08-27 14:32:26 +08:00
|
|
|
#ifdef CONFIG_BRIDGE_VLAN_FILTERING
|
|
|
|
if (data[IFLA_BR_VLAN_PROTOCOL]) {
|
|
|
|
__be16 vlan_proto = nla_get_be16(data[IFLA_BR_VLAN_PROTOCOL]);
|
|
|
|
|
|
|
|
err = __br_vlan_set_proto(br, vlan_proto);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2015-10-04 20:23:28 +08:00
|
|
|
if (data[IFLA_BR_GROUP_FWD_MASK]) {
|
|
|
|
u16 fwd_mask = nla_get_u16(data[IFLA_BR_GROUP_FWD_MASK]);
|
|
|
|
|
|
|
|
if (fwd_mask & BR_GROUPFWD_RESTRICTED)
|
|
|
|
return -EINVAL;
|
|
|
|
br->group_fwd_mask = fwd_mask;
|
|
|
|
}
|
|
|
|
|
2015-10-04 20:23:35 +08:00
|
|
|
if (data[IFLA_BR_GROUP_ADDR]) {
|
|
|
|
u8 new_addr[ETH_ALEN];
|
|
|
|
|
|
|
|
if (nla_len(data[IFLA_BR_GROUP_ADDR]) != ETH_ALEN)
|
|
|
|
return -EINVAL;
|
|
|
|
memcpy(new_addr, nla_data(data[IFLA_BR_GROUP_ADDR]), ETH_ALEN);
|
|
|
|
if (!is_link_local_ether_addr(new_addr))
|
|
|
|
return -EINVAL;
|
|
|
|
if (new_addr[5] == 1 || /* 802.3x Pause address */
|
|
|
|
new_addr[5] == 2 || /* 802.3ad Slow protocols */
|
|
|
|
new_addr[5] == 3) /* 802.1X PAE address */
|
|
|
|
return -EINVAL;
|
|
|
|
spin_lock_bh(&br->lock);
|
|
|
|
memcpy(br->group_addr, new_addr, sizeof(br->group_addr));
|
|
|
|
spin_unlock_bh(&br->lock);
|
|
|
|
br->group_addr_set = true;
|
|
|
|
br_recalculate_fwd_mask(br);
|
|
|
|
}
|
|
|
|
|
2015-10-04 20:23:36 +08:00
|
|
|
if (data[IFLA_BR_FDB_FLUSH])
|
|
|
|
br_fdb_flush(br);
|
|
|
|
|
2015-10-04 20:23:37 +08:00
|
|
|
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
|
|
|
|
if (data[IFLA_BR_MCAST_ROUTER]) {
|
|
|
|
u8 multicast_router = nla_get_u8(data[IFLA_BR_MCAST_ROUTER]);
|
|
|
|
|
|
|
|
err = br_multicast_set_router(br, multicast_router);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2014-09-05 21:51:32 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-09-05 21:51:31 +08:00
|
|
|
static size_t br_get_size(const struct net_device *brdev)
|
|
|
|
{
|
|
|
|
return nla_total_size(sizeof(u32)) + /* IFLA_BR_FORWARD_DELAY */
|
|
|
|
nla_total_size(sizeof(u32)) + /* IFLA_BR_HELLO_TIME */
|
|
|
|
nla_total_size(sizeof(u32)) + /* IFLA_BR_MAX_AGE */
|
2015-03-18 17:06:58 +08:00
|
|
|
nla_total_size(sizeof(u32)) + /* IFLA_BR_AGEING_TIME */
|
|
|
|
nla_total_size(sizeof(u32)) + /* IFLA_BR_STP_STATE */
|
|
|
|
nla_total_size(sizeof(u16)) + /* IFLA_BR_PRIORITY */
|
2015-08-08 00:40:45 +08:00
|
|
|
nla_total_size(sizeof(u8)) + /* IFLA_BR_VLAN_FILTERING */
|
2015-08-27 14:32:26 +08:00
|
|
|
#ifdef CONFIG_BRIDGE_VLAN_FILTERING
|
|
|
|
nla_total_size(sizeof(__be16)) + /* IFLA_BR_VLAN_PROTOCOL */
|
|
|
|
#endif
|
2015-10-04 20:23:28 +08:00
|
|
|
nla_total_size(sizeof(u16)) + /* IFLA_BR_GROUP_FWD_MASK */
|
2015-10-04 20:23:29 +08:00
|
|
|
nla_total_size(sizeof(struct ifla_bridge_id)) + /* IFLA_BR_ROOT_ID */
|
2015-10-04 20:23:30 +08:00
|
|
|
nla_total_size(sizeof(struct ifla_bridge_id)) + /* IFLA_BR_BRIDGE_ID */
|
2015-10-04 20:23:31 +08:00
|
|
|
nla_total_size(sizeof(u16)) + /* IFLA_BR_ROOT_PORT */
|
2015-10-04 20:23:32 +08:00
|
|
|
nla_total_size(sizeof(u32)) + /* IFLA_BR_ROOT_PATH_COST */
|
2015-10-04 20:23:33 +08:00
|
|
|
nla_total_size(sizeof(u8)) + /* IFLA_BR_TOPOLOGY_CHANGE */
|
|
|
|
nla_total_size(sizeof(u8)) + /* IFLA_BR_TOPOLOGY_CHANGE_DETECTED */
|
2015-10-04 20:23:34 +08:00
|
|
|
nla_total_size(sizeof(u64)) + /* IFLA_BR_HELLO_TIMER */
|
|
|
|
nla_total_size(sizeof(u64)) + /* IFLA_BR_TCN_TIMER */
|
|
|
|
nla_total_size(sizeof(u64)) + /* IFLA_BR_TOPOLOGY_CHANGE_TIMER */
|
|
|
|
nla_total_size(sizeof(u64)) + /* IFLA_BR_GC_TIMER */
|
2015-10-04 20:23:35 +08:00
|
|
|
nla_total_size(ETH_ALEN) + /* IFLA_BR_GROUP_ADDR */
|
2015-10-04 20:23:37 +08:00
|
|
|
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
|
|
|
|
nla_total_size(sizeof(u8)) + /* IFLA_BR_MCAST_ROUTER */
|
|
|
|
#endif
|
2014-09-05 21:51:31 +08:00
|
|
|
0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int br_fill_info(struct sk_buff *skb, const struct net_device *brdev)
|
|
|
|
{
|
|
|
|
struct net_bridge *br = netdev_priv(brdev);
|
2015-10-04 20:23:34 +08:00
|
|
|
u64 hello_timer, tcn_timer, topology_change_timer, gc_timer;
|
2014-09-05 21:51:31 +08:00
|
|
|
u32 forward_delay = jiffies_to_clock_t(br->forward_delay);
|
|
|
|
u32 hello_time = jiffies_to_clock_t(br->hello_time);
|
|
|
|
u32 age_time = jiffies_to_clock_t(br->max_age);
|
2015-03-18 17:06:58 +08:00
|
|
|
u32 ageing_time = jiffies_to_clock_t(br->ageing_time);
|
|
|
|
u32 stp_enabled = br->stp_enabled;
|
|
|
|
u16 priority = (br->bridge_id.prio[0] << 8) | br->bridge_id.prio[1];
|
2015-10-04 20:23:28 +08:00
|
|
|
u16 group_fwd_mask = br->group_fwd_mask;
|
2015-08-08 00:40:45 +08:00
|
|
|
u8 vlan_enabled = br_vlan_enabled(br);
|
2015-10-04 20:23:30 +08:00
|
|
|
struct ifla_bridge_id root_id, bridge_id;
|
2015-10-04 20:23:29 +08:00
|
|
|
|
2015-10-04 20:23:30 +08:00
|
|
|
memset(&bridge_id, 0, sizeof(bridge_id));
|
2015-10-04 20:23:29 +08:00
|
|
|
memset(&root_id, 0, sizeof(root_id));
|
|
|
|
memcpy(root_id.prio, br->designated_root.prio, sizeof(root_id.prio));
|
|
|
|
memcpy(root_id.addr, br->designated_root.addr, sizeof(root_id.addr));
|
2015-10-04 20:23:30 +08:00
|
|
|
memcpy(bridge_id.prio, br->bridge_id.prio, sizeof(bridge_id.prio));
|
|
|
|
memcpy(bridge_id.addr, br->bridge_id.addr, sizeof(bridge_id.addr));
|
2015-10-04 20:23:34 +08:00
|
|
|
hello_timer = br_timer_value(&br->hello_timer);
|
|
|
|
tcn_timer = br_timer_value(&br->tcn_timer);
|
|
|
|
topology_change_timer = br_timer_value(&br->topology_change_timer);
|
|
|
|
gc_timer = br_timer_value(&br->gc_timer);
|
2014-09-05 21:51:31 +08:00
|
|
|
|
|
|
|
if (nla_put_u32(skb, IFLA_BR_FORWARD_DELAY, forward_delay) ||
|
|
|
|
nla_put_u32(skb, IFLA_BR_HELLO_TIME, hello_time) ||
|
2015-03-18 17:06:58 +08:00
|
|
|
nla_put_u32(skb, IFLA_BR_MAX_AGE, age_time) ||
|
|
|
|
nla_put_u32(skb, IFLA_BR_AGEING_TIME, ageing_time) ||
|
|
|
|
nla_put_u32(skb, IFLA_BR_STP_STATE, stp_enabled) ||
|
2015-08-08 00:40:45 +08:00
|
|
|
nla_put_u16(skb, IFLA_BR_PRIORITY, priority) ||
|
2015-10-04 20:23:28 +08:00
|
|
|
nla_put_u8(skb, IFLA_BR_VLAN_FILTERING, vlan_enabled) ||
|
2015-10-04 20:23:29 +08:00
|
|
|
nla_put_u16(skb, IFLA_BR_GROUP_FWD_MASK, group_fwd_mask) ||
|
2015-10-04 20:23:30 +08:00
|
|
|
nla_put(skb, IFLA_BR_ROOT_ID, sizeof(root_id), &root_id) ||
|
2015-10-04 20:23:31 +08:00
|
|
|
nla_put(skb, IFLA_BR_BRIDGE_ID, sizeof(bridge_id), &bridge_id) ||
|
2015-10-04 20:23:32 +08:00
|
|
|
nla_put_u16(skb, IFLA_BR_ROOT_PORT, br->root_port) ||
|
2015-10-04 20:23:33 +08:00
|
|
|
nla_put_u32(skb, IFLA_BR_ROOT_PATH_COST, br->root_path_cost) ||
|
|
|
|
nla_put_u8(skb, IFLA_BR_TOPOLOGY_CHANGE, br->topology_change) ||
|
|
|
|
nla_put_u8(skb, IFLA_BR_TOPOLOGY_CHANGE_DETECTED,
|
2015-10-04 20:23:34 +08:00
|
|
|
br->topology_change_detected) ||
|
|
|
|
nla_put_u64(skb, IFLA_BR_HELLO_TIMER, hello_timer) ||
|
|
|
|
nla_put_u64(skb, IFLA_BR_TCN_TIMER, tcn_timer) ||
|
|
|
|
nla_put_u64(skb, IFLA_BR_TOPOLOGY_CHANGE_TIMER,
|
|
|
|
topology_change_timer) ||
|
2015-10-04 20:23:35 +08:00
|
|
|
nla_put_u64(skb, IFLA_BR_GC_TIMER, gc_timer) ||
|
|
|
|
nla_put(skb, IFLA_BR_GROUP_ADDR, ETH_ALEN, br->group_addr))
|
2014-09-05 21:51:31 +08:00
|
|
|
return -EMSGSIZE;
|
|
|
|
|
2015-08-27 14:32:26 +08:00
|
|
|
#ifdef CONFIG_BRIDGE_VLAN_FILTERING
|
|
|
|
if (nla_put_be16(skb, IFLA_BR_VLAN_PROTOCOL, br->vlan_proto))
|
|
|
|
return -EMSGSIZE;
|
|
|
|
#endif
|
|
|
|
|
2015-10-04 20:23:37 +08:00
|
|
|
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
|
|
|
|
if (nla_put_u8(skb, IFLA_BR_MCAST_ROUTER, br->multicast_router))
|
|
|
|
return -EMSGSIZE;
|
|
|
|
#endif
|
|
|
|
|
2014-09-05 21:51:31 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-02-26 15:55:40 +08:00
|
|
|
static size_t br_get_link_af_size(const struct net_device *dev)
|
|
|
|
{
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
struct net_bridge_port *p;
|
|
|
|
struct net_bridge *br;
|
|
|
|
int num_vlans = 0;
|
2015-02-26 15:55:40 +08:00
|
|
|
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
if (br_port_exists(dev)) {
|
|
|
|
p = br_port_get_rtnl(dev);
|
|
|
|
num_vlans = br_get_num_vlan_infos(nbp_vlan_group(p),
|
2015-10-01 02:16:53 +08:00
|
|
|
RTEXT_FILTER_BRVLAN);
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
} else if (dev->priv_flags & IFF_EBRIDGE) {
|
|
|
|
br = netdev_priv(dev);
|
|
|
|
num_vlans = br_get_num_vlan_infos(br_vlan_group(br),
|
2015-10-01 02:16:53 +08:00
|
|
|
RTEXT_FILTER_BRVLAN);
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
}
|
2015-02-26 15:55:40 +08:00
|
|
|
|
|
|
|
/* Each VLAN is returned in bridge_vlan_info along with flags */
|
bridge: vlan: add per-vlan struct and move to rhashtables
This patch changes the bridge vlan implementation to use rhashtables
instead of bitmaps. The main motivation behind this change is that we
need extensible per-vlan structures (both per-port and global) so more
advanced features can be introduced and the vlan support can be
extended. I've tried to break this up but the moment net_port_vlans is
changed and the whole API goes away, thus this is a larger patch.
A few short goals of this patch are:
- Extensible per-vlan structs stored in rhashtables and a sorted list
- Keep user-visible behaviour (compressed vlans etc)
- Keep fastpath ingress/egress logic the same (optimizations to come
later)
Here's a brief list of some of the new features we'd like to introduce:
- per-vlan counters
- vlan ingress/egress mapping
- per-vlan igmp configuration
- vlan priorities
- avoid fdb entries replication (e.g. local fdb scaling issues)
The structure is kept single for both global and per-port entries so to
avoid code duplication where possible and also because we'll soon introduce
"port0 / aka bridge as port" which should simplify things further
(thanks to Vlad for the suggestion!).
Now we have per-vlan global rhashtable (bridge-wide) and per-vlan port
rhashtable, if an entry is added to a port it'll get a pointer to its
global context so it can be quickly accessed later. There's also a
sorted vlan list which is used for stable walks and some user-visible
behaviour such as the vlan ranges, also for error paths.
VLANs are stored in a "vlan group" which currently contains the
rhashtable, sorted vlan list and the number of "real" vlan entries.
A good side-effect of this change is that it resembles how hw keeps
per-vlan data.
One important note after this change is that if a VLAN is being looked up
in the bridge's rhashtable for filtering purposes (or to check if it's an
existing usable entry, not just a global context) then the new helper
br_vlan_should_use() needs to be used if the vlan is found. In case the
lookup is done only with a port's vlan group, then this check can be
skipped.
Things tested so far:
- basic vlan ingress/egress
- pvids
- untagged vlans
- undef CONFIG_BRIDGE_VLAN_FILTERING
- adding/deleting vlans in different scenarios (with/without global ctx,
while transmitting traffic, in ranges etc)
- loading/removing the module while having/adding/deleting vlans
- extracting bridge vlan information (user ABI), compressed requests
- adding/deleting fdbs on vlans
- bridge mac change, promisc mode
- default pvid change
- kmemleak ON during the whole time
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-26 01:00:11 +08:00
|
|
|
return num_vlans * nla_total_size(sizeof(struct bridge_vlan_info));
|
2015-02-26 15:55:40 +08:00
|
|
|
}
|
|
|
|
|
2015-01-29 19:15:03 +08:00
|
|
|
static struct rtnl_af_ops br_af_ops __read_mostly = {
|
2013-02-13 20:00:13 +08:00
|
|
|
.family = AF_BRIDGE,
|
|
|
|
.get_link_af_size = br_get_link_af_size,
|
|
|
|
};
|
|
|
|
|
2012-06-26 13:48:45 +08:00
|
|
|
struct rtnl_link_ops br_link_ops __read_mostly = {
|
2014-09-05 21:51:29 +08:00
|
|
|
.kind = "bridge",
|
|
|
|
.priv_size = sizeof(struct net_bridge),
|
|
|
|
.setup = br_dev_setup,
|
2015-08-20 02:29:35 +08:00
|
|
|
.maxtype = IFLA_BR_MAX,
|
2014-09-05 21:51:32 +08:00
|
|
|
.policy = br_policy,
|
2014-09-05 21:51:29 +08:00
|
|
|
.validate = br_validate,
|
|
|
|
.newlink = br_dev_newlink,
|
2014-09-05 21:51:32 +08:00
|
|
|
.changelink = br_changelink,
|
2014-09-05 21:51:29 +08:00
|
|
|
.dellink = br_dev_delete,
|
2014-09-05 21:51:31 +08:00
|
|
|
.get_size = br_get_size,
|
|
|
|
.fill_info = br_fill_info,
|
2014-09-05 21:51:30 +08:00
|
|
|
|
|
|
|
.slave_maxtype = IFLA_BRPORT_MAX,
|
|
|
|
.slave_policy = br_port_policy,
|
|
|
|
.slave_changelink = br_port_slave_changelink,
|
2014-09-05 21:51:29 +08:00
|
|
|
.get_slave_size = br_port_get_slave_size,
|
|
|
|
.fill_slave_info = br_port_fill_slave_info,
|
2011-04-04 22:03:32 +08:00
|
|
|
};
|
2006-05-26 07:00:12 +08:00
|
|
|
|
2007-03-23 02:59:03 +08:00
|
|
|
int __init br_netlink_init(void)
|
2006-05-26 07:00:12 +08:00
|
|
|
{
|
2013-01-02 17:41:25 +08:00
|
|
|
int err;
|
|
|
|
|
|
|
|
br_mdb_init();
|
2013-12-31 02:41:32 +08:00
|
|
|
rtnl_af_register(&br_af_ops);
|
2013-01-02 17:41:25 +08:00
|
|
|
|
2013-02-13 20:00:13 +08:00
|
|
|
err = rtnl_link_register(&br_link_ops);
|
|
|
|
if (err)
|
|
|
|
goto out_af;
|
|
|
|
|
2013-01-02 17:41:25 +08:00
|
|
|
return 0;
|
2013-02-13 20:00:13 +08:00
|
|
|
|
|
|
|
out_af:
|
|
|
|
rtnl_af_unregister(&br_af_ops);
|
2013-01-02 17:41:25 +08:00
|
|
|
br_mdb_uninit();
|
|
|
|
return err;
|
2006-05-26 07:00:12 +08:00
|
|
|
}
|
|
|
|
|
2014-09-18 17:29:03 +08:00
|
|
|
void br_netlink_fini(void)
|
2006-05-26 07:00:12 +08:00
|
|
|
{
|
2013-01-02 17:41:25 +08:00
|
|
|
br_mdb_uninit();
|
2013-02-13 20:00:13 +08:00
|
|
|
rtnl_af_unregister(&br_af_ops);
|
2011-04-04 22:03:32 +08:00
|
|
|
rtnl_link_unregister(&br_link_ops);
|
2006-05-26 07:00:12 +08:00
|
|
|
}
|