This allows us to use the bridge as a managed switch and gracefully
handle mixed tagged and untagged frames. Prior to this, the only
alternative was creating one bridge per vlan which quickly becomes a
nightmare and still won't let you mix both tagged and untagged frames on
the physical port without some complex ebtables magic.
This is in line with the notion that OpenWRT is the network go-to swiss
army knife when you need a nice set-and-forget, low maintenance box to
handle a specific task.
Current builds of the ip-bridge package already fully support this
feature so the only requirement is enabling the kernel config.
This is disabled by default so existing bridge configurations will not
be affected. This patch only gives the ability to turn it on with an
'ip link' command. If there is interest, I could look into making the
feature accessible via uci configuration.
It causes about 3.1% hit on raw bridging speed, which is relatively
trivial considering that I had to use 300 byte packets to strain the CPU
enough to notice a slowdown at all. The ER8 would chug along at wire
speed otherwise, and that's using only one core. Since the typical
bridge use case on OpenWRT is wireless, I doubt it would be noticeable
at all.
With BRIDGE_VLAN_FILTERING
iperf -u -c 192.168.1.105 -b 1G -l 300
------------------------------------------------------------
Client connecting to 192.168.1.105, UDP port 5001
Sending 300 byte datagrams, IPG target: 2.24 us (kalman adjust)
UDP buffer size: 208 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.1.12 port 58045 connected with 192.168.1.105 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 977 MBytes 820 Mbits/sec
[ 3] Sent 3414986 datagrams
[ 3] Server Report:
[ 3] 0.0-10.0 sec 811 MBytes 680 Mbits/sec 0.000 ms
581210/3414986 (0%)
Without BRIDGE_VLAN_FILTERING
iperf -u -c 192.168.1.105 -b 1G -l 300
------------------------------------------------------------
Client connecting to 192.168.1.105, UDP port 5001
Sending 300 byte datagrams, IPG target: 2.24 us (kalman adjust)
UDP buffer size: 208 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.1.12 port 36645 connected with 192.168.1.105 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 977 MBytes 820 Mbits/sec
[ 3] Sent 3414990 datagrams
[ 3] Server Report:
[ 3] 0.0-10.0 sec 836 MBytes 701 Mbits/sec 0.000 ms
493950/3414990 (0%)
In terms of kernel size, it uses 16KB (6753K vs 6737K on ER8) so a
0.002% hit. The exact 16KB is probably just due to how the kernel is
compressed.
Suggested-by: Jonathan Thibault <jonathan@navigue.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
No single target/arch uses it and most likely there is no need to make
such a potential code target/arch specific.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
If "compatible" is being used that should trigger a proper parser
directly. It's more reliable thanks to not trying parsers one by one. In
such case partition shouldn't be split automatically to avoid parsing it
twice.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Currently, the b53 MDIO switch driver registers the switch on
config-init and not on device probe. Because of this, the switch
gets added every time the associated interface comes up.
This commit fixes this behavior by registering the switch on device
probe.
Compile- and run-tested on OCEDO Koala.
Signed-off-by: David Bauer <mail@david-bauer.net>
It makes more sense to add run_parsers_by_type() in a patch that
introduces parser types. That makes the other one just add a code using
it.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
There was a bug in ubifs related to the O_TMPFILE. When reapplying
changes after power cut data could be lost. This problem was exposed by
overlayfs and the upstream commit 3a1e819b4e80 ("ovl: store file handle
of lower inode on copy up").
This fixes a regression introduced when switching from 4.9 to 4.14.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
It has been rejected upstream and instead a nice/more generic solution
has been implemented. It's possible now to describe partitions format
using "compatible" DT string.
No OpenWrt target uses "linux,part-probe" anymore, leave it only in case
some forks need it. It will be dropped with support for new kernels.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Newer batches of several Mikrotik boards contain this yet-unsupported
flash chip, for instance:
- rb941-2nd (hAP lite)
- rb952ui-5ac2nd (hAP ac lite)
- RBM33G
and probably other Mikrotik boards need this patch as well.
The patch was submitted upstream by Robert Marko: https://patchwork.ozlabs.org/patch/934181/
Closes: FS#1715
Signed-off-by: Baptiste Jonglez <git@bitsofnetworks.org>
Cc: Robert Marko <robimarko@gmail.com>
Refreshed all patches.
Altered patches:
- 666-Add-support-for-MAP-E-FMRs-mesh-mode.patch
New symbol for arm targets:
- HARDEN_BRANCH_PREDICTOR
Compile-tested on: ar71xx, cns3xxx, imx6
Runtime-tested on: ar71xx, cns3xxx, imx6
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Compaction is the only memory management component to form high order (larger
physically contiguous) memory blocks reliably. The page allocator relies on
compaction heavily and the lack of the feature can lead to unexpected OOM
killer invocations for high order memory requests. You shouldn't disable this
option unless there really is a strong reason for it.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Michal Hrusecky <michal.hrusecky@nic.cz>
Optimized inlining was disabled by default when gcc 4 was still
relatively new. By now, all gcc versions handle this well and there
seems to be no real reason to keep it x86-only.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The cache coloring problem on MIPS CPUs was fixed with kernel 4.9.129 of
the kernel 4.9 branch. Activate VDSO support for MIPS again.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This adds some configuration options which are selectable when the
layerscape armv7 target is compiled.
This was found by build bot.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This adds some configuration options which are selectable when the
samsung target is compiled.
This was found by build bot.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Backport an additional patch from 4.16 for nftables.
This fixes a build problem recently introduced.
Fixes: f57806b56e ("kernel: generic: Fix nftables inet table breakage")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Commit b7265c59ab ("kernel: backport a series of netfilter cleanup
patches to 4.14") added patch 302-netfilter-nf_tables_inet-don-t-use-
multihook-infrast.patch. That patch switches the netfilter core in the
kernel to use the new native NFPROTO_INET support. Unfortunately, the
new native NFPROTO_INET support does not exist in 4.14 and was not
backported along with this patchset. As such, nftables inet tables never
see any traffic.
As an example the following nft counter rule should increment for every
packet coming into the box, but never will:
nft add table inet foo
nft add chain inet foo bar { type filter hook input priority 0\; }
nft add rule inet foo bar counter
This commit pulls in the required backport patches to add the new
native NFPROTO_INET support, and thus restore nftables inet table
functionality.
Tested on Turris Omnia (mvebu)
Fixes: b7265c59ab ("kernel: backport a series of netfilter cleanup ...")
Signed-off-by: Brett Mastbergen <bmastbergen@untangle.com>
kernel upstream commit 9efcaa7c4afba5628f2650a76f69c798f47eeb18 to 4.14
itself a backport of 0f02cfbc3d9e413d450d8d0fd660077c23f67eff has
resolved the cache line issues that led to us disabling VDSO by default
on MIPS.
Remove our force disable patch:
pending-4.14/206-mips-disable-vdso.patch
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
Refreshed all patches.
Added new patch:
- 192-Revert-ubifs-xattr-Don-t-operate-on-deleted-inodes.patch
This fixes a bug introduced in upstream 4.14.68 which caused targets using
ubifs to produce file-system errors on boot, rendering them useless.
Compile-tested on: cns3xxx, imx6, x86_64
Runtime-tested on: cns3xxx, imx6, x86_64
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Some systems require multiple flash chips to be concatenated and read as
a single mtd device. The ar71xx target provides custom code to create
such mtdconcat devices. When porting devices to ath79, however, there is
no way to create such devices from within the device tree.
This commit adds a driver for creating virtual mtd-concat devices to the
ath79 target. Nodes must have a compatible = "virtual,mtd-concat" line,
and define a list of devices to concat in the 'devices' property,
for example:
flash {
compatible = "virtual,mtd-concat";
devices = <&flash0 &flash1>;
};
The driver is added to the very end of the mtd Makefile to increase the
likelyhood of all child devices already being loaded at the time of
probing, preventing unnecessary deferred probes which might in turn
cause other problems (like failure to load MAC addresses from art because
the partitions are not loaded yet).
Signed-off-by: Bernhard Frauendienst <openwrt@nospam.obeliks.de>
