This contains two sets of fixes for the 8139cp driver.
For all kernel versions older than 4.3, we can apply the fixes from the
4.3-rc4 kernel. In particular, these fix the TX timeout recovery which
is causing my Geos to lock up until the hardware watchdog kicks in.
For 4.0 and later kernels, we can also apply the additional improvements
which are going into 4.4 to fix and enable hardware checksum/TSO
offload. Backporting those to older kernels is non-trivial.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
SVN-Revision: 47220
the lantiq dsl hack to struct atm_trafprm broke the ABI fix this by moving
the variables to the end of the struct
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 46964
commit r30917 ("kernel: bypass all netfilter hooks if the sysctls for that
functionality have been disabled - eliminates the overhead of enabling
CONFIG_BRIDGE_NETFILTER in the kernel config") introduced an optimization
which should reduce/eliminate the overhead for traffic send over bridges on
kernels compiled with CONFIG_BRIDGE_NETFILTER=y. But this optimization
breaks the nf_call_iptables per bridge setting which is more fine grained
than the global sysctl net.bridge.bridge-nf-call-iptables setting.
A test reflecting a real world setup was created to identify if this really
eliminates the overhead and if per-bridge nf_call_iptables could be used in
some setups to increase the throughput. A Qualcomm Atheros QCA9558 based
system with one ethernet and an ath9k wifi 3x3 in HT40 mode was used.
Cables from the AP to the wifi station were used to reduce interference
problems during the tests.
The wlan interface was put in one bridge interface called br-wlan. This
bridge usually contains some more wlan interfaces. The eth0 was put in a
second bridge called br-lan. This usually contains some other privileged
wlan or mesh interfaces. Routing was added between br-lan and br-wlan.
Three kernels were tested:
* (default) OpenWrt kernel for this device
* (brfilter-global) OpenWrt kernel with CONFIG_BRIDGE_NETFILTER=y
* (brfilter-local) OpenWrt kernel with CONFIG_BRIDGE_NETFILTER=y and
without 644-bridge_optimize_netfilter_hooks.patch
The changes to the the netfilter settings of the bridge were done via:
* (brfilter-global) /sbin/sysctl -w net.bridge.bridge-nf-call-iptables=1
* (brfilter-lobal) echo 1 > /sys/class/net/br-lan/bridge/nf_call_iptables
and/or echo 1 > /sys/class/net/br-wan/bridge/nf_call_iptables
A station connected to the wlan0 (AP) interface was used to send traffic to
a PC connected via ethernet. iperf with 3 concurrent transmissions was used
to generate the traffic.
| kernel | br-nf-* global | nf-call* iface | download | upload |
|-----------------|----------------|----------------|----------|----------|
| default | 0 | - | 209 | 268 |
| brfilter-global | 0 | - | 185 | 243 |
| brfilter-local | 0 | - | 187 | 243 |
| brfilter-local | 0 | br-lan | 157 | 226 |
| brfilter-local | 0 | br-lan br-wlan | 139 | 161 |
| brfilter-global | 1 | - | 136 | 162 |
Download/upload results in Mibit/s
It can be seen that the patch doesn't eliminate the overhead. It can also
be seen that the throughput of brfilter-global and brfilter-local with
disabled filtering is the roughly the same. Also the throughput for
brfilter-global and brfilter-local for enabled filtering on all bridges is
roughly the same.
But also the brfilter-local throughput is higher when only br-lan requires
the filtering. This setting would not be possible with
644-bridge_optimize_netfilter_hooks.patch applied and thus can only be
compared with brfilter-global and filtering enabled for all interfaces.
Signed-off-by: Sven Eckelmann <sven@open-mesh.com>
SVN-Revision: 46835
It was corrupted in r38528. The most obvious symptom is repeated messages like this:
Tue Sep 8 08:25:18 2015 kern.warn kernel: [77141.972226] br-lan: received packet on wlan0 with own address as source address
Signed-off-by: Dmitry Ivanov <dima@ubnt.com>
SVN-Revision: 46821
Currently, multicast packets from an STA are sent to any according
multicast listener directly through the bridge multicast-to-unicast
feature. Unfortunately, so far this includes the originating STA, too,
resulting in multicast packets being echo'ed back to the originating STA
if it itself is a multicast listener for that group.
This behaviour breaks IPv6 duplicate address detection: An IPv6 Neighbor
Solicitation for IPv6 Duplicate Address Detection is being echo'ed back,
resulting in the host falsely detecting an address collision, which
makes the node unable to claim an IPv6 address and use IPv6 in general.
Mac80211 unfortunately only prevents the echoes for us for multicast
frames. For the multicast frames cast to a unicast destination we'll
need to take care of excluding the originator ourselves.
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
SVN-Revision: 46765
A call to pskb_may_pull() might reallocate skb->data. Therefore we
should only assign the src-pointer after any potential reallocations.
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46721
This reverts commit a080e8e1943156168913d0353a2e99d1151102aa.
It did not fix the problem but just hid some symptom. The real issue was
that IGMP/MLD report suppression was not considered for the
multicast-to-unicast feature. A recent netifd which isolates IGMP/MLD
reports between STAs by utilizing AP-isolation and bridge-hairpinning
should have fixed this.
It is perfectly fine to apply multicast-to-unicast to IPv6 Neighbor
Solicitations, too (once that feature is configured correctly).
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46720
Properly treat -ENOSYS as no PHY, else ehci-orion won't work without
generic phy support.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46711
The "linux,part-probe" dts parsing is a pretty neat generic feature.
It has been posted to kernel.org and could easily be reused by all
targets.
This change moves the patch to the 3.18 and 4.1 generic folders, and
makes the feature available to all platforms who may want to use it.
Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46654
In addition to the update this also fixes compile problems with kernel 4.1.
This closes#20323.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46609
This patch was missing for kernel 4.0 and 4.1 in r46464. I also
replaced the version in OpenWrt with the version which went upstream
into the Linux mainline kernel.
This closes#20193 and #20192
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46493
REGMAP_MMIO is a hidden symbol, so it cannot be directly selected. If nothing
selects it, it will fail the build with:
ERROR: module '/store/buildbot/slave/slave/rb532/build/build_dir/target-mipsel_mips32_musl-1.1.10/linux-rb532/linux-3.18.19/drivers/base/regmap/regmap-mmio.ko' is missing.
make[2]: *** [/store/buildbot/slave/slave/rb532/build/bin/rb532/packages/base/kmod-regmap_3.18.19-1_rb532.ipk] Error 1
Fix this by adding a description to allow selecting it manually.
Fixes: r46451 ("kmod-regmap: enable MMIO support")
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46485
NCM specs are not actually mandating a specific position in the frame for
the NDP (Network Datagram Pointer). However, some Huawei devices will
ignore our aggregates if it is not placed after the datagrams it points
to. Add support for doing just this, in a per-device configurable way.
While at it, update NCM subdrivers, disabling this functionality in all of
them, except in huawei_cdc_ncm where it is enabled instead.
We aren't making any distinction between different Huawei NCM devices,
based on what the vendor driver does. Standard NCM devices are left
unaffected: if they are compliant, they should be always usable, still
stay on the safe side.
This change has been tested and working with a Huawei E3131 device (which
works regardless of NDP position) and an E3372 device (which mandates NDP
to be after indexed datagrams).
Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 46464
The mips reloc patch introduced new allocations which were done before
add_unformed_module but never freed them in case of an error. A new hook in
Linux 3.19 called module_arch_freeing_init can be used for freeing memory
which were allocated during this init phase.
The problem can be seen when trying to load a module (via busybox insmod)
when it was already loaded.
free -m
for i in `seq 1 100`; do
/sbin/insmod /lib/modules/*/ath9k.ko >& /dev/null
done
free -m
This simple loop would leak ~3.2 MB.
Signed-off-by: Sven Eckelmann <sven@open-mesh.com>
SVN-Revision: 46247
Backport for the Spansion S25FL164K
It's a 8 MiB flash chip with 4 KiB erase sectors.
Signed-off-by: L. D. Pinney <ldpinney@gmail.com>
SVN-Revision: 46237
This is useful if the device also has an ethernet WAN interface with a
separate mac address (that is derived from the LAN mac address).
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46220
Make some network uapi headers detect if they are included after
not only glibc but also musl headers.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46161
With this patch the following two patches are backported:
* bridge: fix parsing of MLDv2 reports
* bridge: allow setting hash_max + multicast_router if interface is down
The former one is an important fix which got just applied to the net-tree
and is queued for stable. The latter is a patch which is needed to make
the hash_max and multicast_router attributes configurable through
netifd.
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
SVN-Revision: 45783
