openwrt/staging/dangole.git
8 years agocns3xxx: remove obsolete file
Felix Fietkau [Mon, 18 Jan 2016 22:37:57 +0000 (22:37 +0000)]
cns3xxx: remove obsolete file

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48330

8 years agocns3xxx: remove unused define and use of a deprecated function
Felix Fietkau [Mon, 18 Jan 2016 22:37:51 +0000 (22:37 +0000)]
cns3xxx: remove unused define and use of a deprecated function

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48329

8 years agolantiq: add support for TP-Link VR200v
Hauke Mehrtens [Mon, 18 Jan 2016 20:40:03 +0000 (20:40 +0000)]
lantiq: add support for TP-Link VR200v

This adds basic support for TP-Link VR200v.
Currently the following parts are not working: FXO, Voice, DECT, WIFI (both)

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 48328

8 years agokernel: fix uninitialized variable in bridge multicast-to-unicast patch on 4.4
Felix Fietkau [Mon, 18 Jan 2016 20:16:37 +0000 (20:16 +0000)]
kernel: fix uninitialized variable in bridge multicast-to-unicast patch on 4.4

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48327

8 years agoarc: clean-up and move CFLAGS to include/target.mk
Felix Fietkau [Mon, 18 Jan 2016 17:52:03 +0000 (17:52 +0000)]
arc: clean-up and move CFLAGS to include/target.mk

Most of currently mentioned CFLAGS in arc770/Makefile
are not really required because:
 [1] "-Os -pipe" are set by default in include/target.mk
 [2] "-fno-caller-saves" gets enabled via menuconfig
     as an extra compiler flag for developers

So the only one that makes sense is "-matomic" and
that one is really essential. Without it many software
packges won't build complainin on unresolved atomic ops.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Felix Fietkau <nbd@openwrt.org>
Cc: Jo-Philipp Wich <jow@openwrt.org>
Cc: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 48326

8 years agoarc770/axs101: fix console output
Felix Fietkau [Mon, 18 Jan 2016 17:51:54 +0000 (17:51 +0000)]
arc770/axs101: fix console output

While bumping kernel version kernel command line was
unintentionally modified in attempt to make it closer
to upstream version.

In case of AXS that has not only serial port but HDMI/USB
both capable of being debug console we have 2 entries in
kernel's command line:
------------------->8-----------------
console=tty0 console=ttyS3,115200n8
------------------->8-----------------

But as it turned out OpenWRT uses procd as init instead of
Busybox. And in its turn procd gets the first "console"
entry from kernel command line (/proc/cmdline) and uses it
if default inittab is used:
------------------->8-----------------
...
::askconsole:/bin/ash --login
------------------->8-----------------

So what we got is non-functional serial console.
That change removes "console=tty0" which brings serial
console back to life.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Felix Fietkau <nbd@openwrt.org>
Cc: Jo-Philipp Wich <jow@openwrt.org>
Cc: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 48325

8 years agokernel: fix a build regression in the ld-version script fix
Felix Fietkau [Mon, 18 Jan 2016 17:40:40 +0000 (17:40 +0000)]
kernel: fix a build regression in the ld-version script fix

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48324

8 years agolantiq: drop linux 4.1 support
Felix Fietkau [Mon, 18 Jan 2016 15:54:23 +0000 (15:54 +0000)]
lantiq: drop linux 4.1 support

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48323

8 years agolantiq: switch to linux 4.4
Felix Fietkau [Mon, 18 Jan 2016 15:54:18 +0000 (15:54 +0000)]
lantiq: switch to linux 4.4

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48322

8 years agobr2684ctl: resolve a boot time race condition with nas0 bringup by using explicit...
Felix Fietkau [Mon, 18 Jan 2016 15:35:30 +0000 (15:35 +0000)]
br2684ctl: resolve a boot time race condition with nas0 bringup by using explicit notification when init is done

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48321

8 years agobr2684ctl: add support for notifying nas* bringup via a script
Felix Fietkau [Mon, 18 Jan 2016 15:35:24 +0000 (15:35 +0000)]
br2684ctl: add support for notifying nas* bringup via a script

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48320

8 years agolantiq: fix VDSL device detection with Linux 4.4
Felix Fietkau [Mon, 18 Jan 2016 15:35:17 +0000 (15:35 +0000)]
lantiq: fix VDSL device detection with Linux 4.4

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48319

8 years agoramips: delete the broken sound driver
Felix Fietkau [Mon, 18 Jan 2016 13:40:15 +0000 (13:40 +0000)]
ramips: delete the broken sound driver

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48318

8 years agoramips: mark kmod-sound-mtk as broken, it does not compile properly
Felix Fietkau [Mon, 18 Jan 2016 13:22:38 +0000 (13:22 +0000)]
ramips: mark kmod-sound-mtk as broken, it does not compile properly

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48317

8 years agoramips: fix kernel config handling for mt7620/mt7628 sound module
Felix Fietkau [Mon, 18 Jan 2016 13:22:34 +0000 (13:22 +0000)]
ramips: fix kernel config handling for mt7620/mt7628 sound module

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48316

8 years agofirewall: add CONFIG_IPV6 to PKG_CONFIG_DEPENDS to fix a rebuild error
Felix Fietkau [Mon, 18 Jan 2016 13:21:37 +0000 (13:21 +0000)]
firewall: add CONFIG_IPV6 to PKG_CONFIG_DEPENDS to fix a rebuild error

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48315

8 years agoiptables: fix rebuild errors on configuration changes
Felix Fietkau [Mon, 18 Jan 2016 13:21:32 +0000 (13:21 +0000)]
iptables: fix rebuild errors on configuration changes

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48314

8 years agobuild: use sstrip by default for musl
Felix Fietkau [Mon, 18 Jan 2016 12:47:36 +0000 (12:47 +0000)]
build: use sstrip by default for musl

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48313

8 years agoopenssl: remove the separate configuration menu, use the implicit one (via MENU:=1)
Felix Fietkau [Mon, 18 Jan 2016 12:42:08 +0000 (12:42 +0000)]
openssl: remove the separate configuration menu, use the implicit one (via MENU:=1)

Fixes warning on selecting OPENSSL_ENGINE_CRYPTO if openssl is not selected

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48312

8 years agobrcm2708: fix RPi model B plus support
Felix Fietkau [Mon, 18 Jan 2016 12:42:03 +0000 (12:42 +0000)]
brcm2708: fix RPi model B plus support

Not sure if B+ is used or not (mine has B Plus) so leave both

Signed-off-by: Roman Yeryomin <roman@advem.lv>
SVN-Revision: 48311

8 years agokernel: fix MIPS linux-4.4 build on non-GNU systems
Felix Fietkau [Mon, 18 Jan 2016 11:40:49 +0000 (11:40 +0000)]
kernel: fix MIPS linux-4.4 build on non-GNU systems

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48310

8 years agowpa_supplicant: improve generating phase2 config line for WPA-EAP
Felix Fietkau [Mon, 18 Jan 2016 11:40:44 +0000 (11:40 +0000)]
wpa_supplicant: improve generating phase2 config line for WPA-EAP

WPA-EAP supports several phase2 (=inner) authentication methods when
using EAP-TTLS, EAP-PEAP or EAP-FAST (the latter is added as a first
step towards the UCI model supporting EAP-FAST by this commit)
The value of the auth config variable was previously expected to be
directly parseable as the content of the 'phase2' option of
wpa_supplicant.
This exposed wpa_supplicant's internals, leaving it to view-level to
set the value properly. Unfortunately, this is currently not the case,
as LuCI currently allows values like 'PAP', 'CHAP', 'MSCHAPV2'.
Users thus probably diverged and set auth to values like
'auth=MSCHAPV2' as a work-around.
This behaviour isn't explicitely documented anywhere and is not quite
intuitive...

The phase2-string is now generated according to $eap_type and $auth,
following the scheme also found in hostap's test-cases:
http://w1.fi/cgit/hostap/tree/tests/hwsim/test_ap_eap.py
The old behaviour is also still supported for the sake of not breaking
existing, working configurations.

Examples:
  eap_type   auth
  'ttls'     'EAP-MSCHAPV2'     -> phase2="autheap=MSCHAPV2"
  'ttls'     'MSCHAPV2'         -> phase2="auth=MSCHAPV2"
  'peap'     'EAP-GTC'          -> phase2="auth=GTC"

Deprecated syntax supported for compatibility:
  'ttls'     'autheap=MSCHAPV2' -> phase2="autheap=MSCHAPV2"

I will suggest a patch to LuCI adding EAP-MSCHAPV2, EAP-GTC, ... to
the list of Authentication methods available.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
SVN-Revision: 48309

8 years agolantiq: ltq-deu: Remove the "DEU test manager"
Felix Fietkau [Mon, 18 Jan 2016 11:40:40 +0000 (11:40 +0000)]
lantiq: ltq-deu: Remove the "DEU test manager"

Remove the "DEU test manager" code which has not been used for more than
two years (as the kernel module is not installed anymore since r38731).
This fixes compilation on kernel 4.3, which removes
aead_request_set_assoc (and newer kernels).

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48308

8 years agolantiq: Add support for linux 4.4
Felix Fietkau [Mon, 18 Jan 2016 11:40:19 +0000 (11:40 +0000)]
lantiq: Add support for linux 4.4

The following patches were dropped because they are already applied
upstream:
- 0038-MIPS-lantiq-fpi-on-ar9.patch
- 0039-MIPS-lantiq-initialize-usb-on-boot.patch
- 0042-USB-DWC2-big-endian-support.patch
- 0043-gpio-stp-xway-fix-phy-mask.patch

All other patches were simply refreshed, except the following:
- 0001-MIPS-lantiq-add-pcie-driver.patch
  Changes to arch/mips/lantiq/xway/sysctrl.c (these changes disabled
  some PMU gates for the vrx200 / VR9 SoCs) were removed since the
  upstream kernel disables unused PMU gates automatically (since
  95135bfa7ead1becc2879230f72583dde2b71a0c
  "MIPS: Lantiq: Deactivate most of the devices by default").
- 0025-NET-MIPS-lantiq-adds-xrx200-net.patch
  Since OpenWrt commit 55ba20afcc2fe785146316e5be2c2473cb329885 drivers
  should use of_get_mac_address(). of_get_mac_address_mtd is not
  available for drivers anymore since it's called automatically within
  of_get_mac_address().
- 0028-NET-lantiq-various-etop-fixes.patch
  Same changes as in 0025-NET-MIPS-lantiq-adds-xrx200-net.patch

While refreshing the kernel configuration SPI support had to be moved to
config-4.4 because otherwise M25P80 was disabled.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48307

8 years agobusybox: fix broken IPv6 address displaying in ifconfig
Felix Fietkau [Mon, 18 Jan 2016 11:11:26 +0000 (11:11 +0000)]
busybox: fix broken IPv6 address displaying in ifconfig

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48306

8 years agobrcm63xx: add support for Huawei HG622
Jonas Gorski [Mon, 18 Jan 2016 10:47:23 +0000 (10:47 +0000)]
brcm63xx: add support for Huawei HG622

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 48305

8 years agoipq806x: fix sysupgrade for AP148
Jonas Gorski [Mon, 18 Jan 2016 10:46:58 +0000 (10:46 +0000)]
ipq806x: fix sysupgrade for AP148

AP148 uses lowercase name as boardname, so we need to use this in
platform.sh as well.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 48304

8 years agobrcm63xx: fix platform gpio lookups for gpios < 32
Jonas Gorski [Mon, 18 Jan 2016 10:46:50 +0000 (10:46 +0000)]
brcm63xx: fix platform gpio lookups for gpios < 32

Overwriting static strings is never a good idea, especially expecting
identical strings to be stored in different memory locations.

This caused the lookups to always return the second chip's name. Fix
this by just initializing the lookup with the right values, so we
don't need to modify the strings at all.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 48303

8 years agob53: update header register difinitions
Jonas Gorski [Mon, 18 Jan 2016 10:46:47 +0000 (10:46 +0000)]
b53: update header register difinitions

BCM531x5 has two pontential cpu ports, and header mode can be enabled
independently on both.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 48302

8 years agopackage/libs/libiconv: function names
Jo-Philipp Wich [Mon, 18 Jan 2016 09:23:17 +0000 (09:23 +0000)]
package/libs/libiconv: function names

Currently libiconv-stub and libiconv-full use different names
for functions iconv, iconv_open, and iconv_close.

This may lead to failures when building modules, e.g. with
apr-util when NLS is not activated.

The two modules libiconv-stub and libiconv-full should be
interchangeable, so we need the same function names.

cf.
http://git.savannah.gnu.org/cgit/libiconv.git/tree/include/iconv.h.in

After applying this patch execute

make distclean

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
SVN-Revision: 48301

8 years agoimx6: drop 4.1 support
Luka Perkov [Mon, 18 Jan 2016 06:43:50 +0000 (06:43 +0000)]
imx6: drop 4.1 support

Signed-off-by: Luka Perkov <luka@openwrt.org>
SVN-Revision: 48300

8 years agoimx6: drop 4.3 support
Luka Perkov [Mon, 18 Jan 2016 06:43:48 +0000 (06:43 +0000)]
imx6: drop 4.3 support

Signed-off-by: Luka Perkov <luka@openwrt.org>
SVN-Revision: 48299

8 years agoimx6: move to 4.4 kernel
Luka Perkov [Mon, 18 Jan 2016 06:40:16 +0000 (06:40 +0000)]
imx6: move to 4.4 kernel

Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
SVN-Revision: 48298

8 years agompc85xx: update to linux 4.4, remove unnecessary patches
Felix Fietkau [Sun, 17 Jan 2016 21:05:19 +0000 (21:05 +0000)]
mpc85xx: update to linux 4.4, remove unnecessary patches

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48297

8 years agoiproute2: remove odd conffiles generation
Felix Fietkau [Sun, 17 Jan 2016 20:41:09 +0000 (20:41 +0000)]
iproute2: remove odd conffiles generation

This was generating a conffiles list that included the binary
and CONTROL/ files.

Signed-off-by: Rob Mosher <nyt-openwrt@countercultured.net>
SVN-Revision: 48296

8 years agolantiq: Remove incorrect PCIe compatible strings
Felix Fietkau [Sun, 17 Jan 2016 19:56:16 +0000 (19:56 +0000)]
lantiq: Remove incorrect PCIe compatible strings

Re-defining the compatible property is not required since the correct
value is inherited from vr9.dtsi.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48295

8 years agolantiq: Remove the old SPI driver patch
Felix Fietkau [Sun, 17 Jan 2016 19:56:10 +0000 (19:56 +0000)]
lantiq: Remove the old SPI driver patch

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48294

8 years agolantiq: Switch to the new SPI driver
Felix Fietkau [Sun, 17 Jan 2016 19:56:03 +0000 (19:56 +0000)]
lantiq: Switch to the new SPI driver

Compared to the "old" driver:
- Each device must assign a pinctrl setting to the SPI node to allow the
  new SPI driver to configure the SPI pins.
  While here we are also using separate input and output settings so we
  are independent of whether the bootloader configures the pins correctly.
- We use the new "compatible" strings to make the driver choose the
  correct number of chip-selects for each SoC.
- The new driver starts counting the chip-selects at 1 (instead of 0, like
  the old one did). Thus we have to adjust the devices accordingly.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48293

8 years agolantiq: Add a rewritten version of the SPI driver
Felix Fietkau [Sun, 17 Jan 2016 19:55:55 +0000 (19:55 +0000)]
lantiq: Add a rewritten version of the SPI driver

The new driver provides a few improvements over the old one:
- Separate compatible strings per SoC type (this allows removing some
  hardcoded of_device_is_compatible() checks)
- It does not rely upon spi-bitbang anymore
- chip-selects are numbered as in the datasheet (= starting at 1 instead
  of 0)

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48292

8 years agolantiq: Disable CONFIG_SPI_GPIO
Felix Fietkau [Sun, 17 Jan 2016 19:55:48 +0000 (19:55 +0000)]
lantiq: Disable CONFIG_SPI_GPIO

All devices are now using the HW SPI driver, so this is not necessary
anymore.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48291

8 years agolantiq: Enable the hardware SPI driver on the DGN3500/DGN3500B
Felix Fietkau [Sun, 17 Jan 2016 19:55:42 +0000 (19:55 +0000)]
lantiq: Enable the hardware SPI driver on the DGN3500/DGN3500B

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48290

8 years agolantiq: Enable SPI for the EASY80920 board again
Felix Fietkau [Sun, 17 Jan 2016 19:55:37 +0000 (19:55 +0000)]
lantiq: Enable SPI for the EASY80920 board again

Also switch to the SPI definition provided by vr9.dtsi

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48289

8 years agolantiq: Switch FRITZ3370 from spi-gpio to the hardware SPI driver
Felix Fietkau [Sun, 17 Jan 2016 19:55:31 +0000 (19:55 +0000)]
lantiq: Switch FRITZ3370 from spi-gpio to the hardware SPI driver

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48288

8 years agolantiq: Re-use the SPI node from vr9.dtsi in TDW89X0.dtsi
Felix Fietkau [Sun, 17 Jan 2016 19:55:25 +0000 (19:55 +0000)]
lantiq: Re-use the SPI node from vr9.dtsi in TDW89X0.dtsi

This removes the duplicate SPI register definition.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48287

8 years agolantiq: Add the SPI node to ar9.dtsi and vr9.dtsi
Felix Fietkau [Sun, 17 Jan 2016 19:55:17 +0000 (19:55 +0000)]
lantiq: Add the SPI node to ar9.dtsi and vr9.dtsi

This allows devices to use SPI without having to re-define (and thus
duplicating) the whole SPI node.
By default SPI is disabled (as before) because only few devices need it.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48286

8 years agolantiq: Configure the PCIe reset GPIO using OF
Felix Fietkau [Sun, 17 Jan 2016 19:55:10 +0000 (19:55 +0000)]
lantiq: Configure the PCIe reset GPIO using OF

After the latest pinctrl backports there are only 50 (instead of 56 as
before) GPIOs/pins exported (thus the first GPIO on VRX200 SoCs is now
462, before it was 456). This means that any hardcoded GPIOs have to be
adjusted.
This broke the PCIe driver (which seems to be the only driver which uses
hardcoded GPIO numbers), it only reports:
ifx_pcie_wait_phy_link_up timeout
ifx_pcie_wait_phy_link_up timeout
ifx_pcie_wait_phy_link_up timeout
ifx_pcie_wait_phy_link_up timeout
ifx_pcie_wait_phy_link_up timeout
pcie_rc_initialize link up failed!!!!!

To prevent more of these issues in the future we remove the hardcoded
PCIe reset GPIO definition and simply pass it via device-tree (like the
PCI driver does).

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48285

8 years agolantiq: Use the new pinctrl compatible strings
Felix Fietkau [Sun, 17 Jan 2016 19:55:04 +0000 (19:55 +0000)]
lantiq: Use the new pinctrl compatible strings

These were introduced in upstream commit
be14811c03cf "pinctrl/lantiq: introduce new dedicated devicetree
bindings" and finally allow us to use the individual pins within our dts
(for example spi_clk, etc.).
Please note that this changes the number of GPIOs which are available for
some SoCs. VRX200 SoCs for example only have 50 pins, but previously 56
pins were exposed. This means that all places which are using hardcoded
GPIO numbers (which are not passed via device-tree) need to be adjusted
(because the first GPIO number is now 462, instead of 456).

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48284

8 years agolantiq: Backport upstream pinctrl-xway patches from 4.5
Felix Fietkau [Sun, 17 Jan 2016 19:54:59 +0000 (19:54 +0000)]
lantiq: Backport upstream pinctrl-xway patches from 4.5

Upstream commit be14811c03cf2 "pinctrl/lantiq: introduce new dedicated
devicetree bindings" allows us to use each pin in the pinmux. This is
useful for example in the "spi" group which contains some pins which
are inputs, and some which are outputs.
These can only be used once the new compatible strings for the pinctrl
node are used.
Additionally 0150-lantiq-pinctrl-xway.patch and the "GPIO PORT3 fix"
(which was part of 0012-pinctrl-lantiq-fix-up-pinmux.patch) were
replaced with their upstream variants which are also in 4.5.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48283

8 years agolantiq: ltq-deu: Make the DEU driver compatible with linux 4.2 and newer
Felix Fietkau [Sun, 17 Jan 2016 19:23:14 +0000 (19:23 +0000)]
lantiq: ltq-deu: Make the DEU driver compatible with linux 4.2 and newer

Upstream linux 4.2 commit 84be456f883c4685680fba8e5154b5f72e92957e
"remove <asm/scatterlist.h>" requires us to include linux/scatterlist.h
instead. This also works with older kernels (at least 4.1, thanks to
Hauke Mehrtens for testing).

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48282

8 years agobase-files: Fix sysupgrade overlay saving
Jo-Philipp Wich [Sun, 17 Jan 2016 18:24:30 +0000 (18:24 +0000)]
base-files: Fix sysupgrade overlay saving

A previous change to sysupgrade moved the overlay files into upper/
The -c switch generates a list of files to backup, but the sed
calls did not take this into consideration.

Signed-off-by: Rob Mosher <nyt-openwrt@countercultured.net>
SVN-Revision: 48281

8 years agonetifd: fix PKG_VERSION (#21630)
Jo-Philipp Wich [Sun, 17 Jan 2016 17:15:01 +0000 (17:15 +0000)]
netifd: fix PKG_VERSION (#21630)

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 48280

8 years agoar71xx/lzma-loader: fix O32 ABI conformance
Gabor Juhos [Sun, 17 Jan 2016 15:29:03 +0000 (15:29 +0000)]
ar71xx/lzma-loader: fix O32 ABI conformance

According to the calling convention of the o32 ABI the caller
function must reserve stack space for $a0-$a3 registers in case
the callee needs to save its arguments.

The assembly code of the loader does not reserve stack space for
these registers thus when the 'loader_main' function needs to save
its arguments, those will be stored in the 'workspace' area instead
of the stack.

Because the workspace area is also used by other part of the code, the
saved register values gets overwritten and this often leads to failed
kernel boots.

Fix the code to reserve stack space for the registers to avoid this
error.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 48279

8 years agouclient: update to the latest version, adds many fixes/features
Felix Fietkau [Sun, 17 Jan 2016 12:41:06 +0000 (12:41 +0000)]
uclient: update to the latest version, adds many fixes/features

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48278

8 years agokernel: add missing kernel config symbol for crypto-rng (fixes #21596)
Felix Fietkau [Sun, 17 Jan 2016 11:17:06 +0000 (11:17 +0000)]
kernel: add missing kernel config symbol for crypto-rng (fixes #21596)

Signed-off-by: Lucian Cristian <luci@createc.ro>
SVN-Revision: 48277

8 years agoar71xx: Use full 128MB flash on Netgear WNDR4300 and WNDR3700v4
Felix Fietkau [Sun, 17 Jan 2016 11:16:56 +0000 (11:16 +0000)]
ar71xx: Use full 128MB flash on Netgear WNDR4300 and WNDR3700v4

Change MTD on WNDR4300 and WNDR3700v4 to fully utilize the 128MB flash.

Credit to @Tuochenlyu on GitHub.

Signed-off-by: Chris Marchesi <chrism@vancluevertech.com>
SVN-Revision: 48276

8 years agompc85xx: Add PTP node for TL-WD4900 in device tree
Felix Fietkau [Sun, 17 Jan 2016 11:16:52 +0000 (11:16 +0000)]
mpc85xx: Add PTP node for TL-WD4900 in device tree

PTP requires at least one timer to be 1PPS so describe it.
For testing, load kernel module gianfar_ptp and use ptp4l
from linuxptp.

Copied from FSL P1010RDB reference design.

Signed-off-by: Wojciech Dubowik <Wojciech.Dubowik@neratec.com>
SVN-Revision: 48275

8 years agonetwork: add virtual tunnel interface (VTI) support
Felix Fietkau [Sun, 17 Jan 2016 11:06:02 +0000 (11:06 +0000)]
network: add virtual tunnel interface (VTI) support

This adds support for configuring VTI interfaces within /etc/config/network.
VTI interfaces are used to create IPsec tunnel interfaces. These interfaces
may be used for routing and other purposes.

Example config:
config interface 'vti1'
option proto 'vti'
option mtu '1500'
option tunlink 'wan'
option peeraddr '192.168.5.16'
option zone 'VPN'
option ikey 2
option okey 2

config interface 'vti1_static'
option proto 'static'
option ifname '@vti1'
option ipaddr '192.168.7.2/24'

The options ikey and okey correspond to the fwmark value of a ipsec policy.
The may be null if you do not want fwmarks.
Also peeraddr may be 0.0.0 if you want all ESP packets go through the
interface.
Example strongswan config:
conn vti
left=%any
leftcert=peer2.test.der
leftid=@peer2.test
right=192.168.5.16
rightid=@peer3.test
leftsubnet=0.0.0.0/0
rightsubnet=0.0.0.0/0
mark=2
auto=route

Signed-off-by: André Valentin <avalentin@marcant.net>
SVN-Revision: 48274

8 years agonetifd: update to the latest version, adds VTI support and a policy routing fix
Felix Fietkau [Sun, 17 Jan 2016 11:05:53 +0000 (11:05 +0000)]
netifd: update to the latest version, adds VTI support and a policy routing fix

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48273

8 years agokernel: add usb-net-sr9700 package
Felix Fietkau [Sun, 17 Jan 2016 11:03:41 +0000 (11:03 +0000)]
kernel: add usb-net-sr9700 package

Add support for
CoreChip-sz SR9700 based USB 1.1 10/100 ethernet devices

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
SVN-Revision: 48272

8 years agoca-certificates: update to version 20160104
Felix Fietkau [Sun, 17 Jan 2016 11:03:36 +0000 (11:03 +0000)]
ca-certificates: update to version 20160104

- update to latest version 20160104
- remove cpu dependency (PKGARCH:=all)
- set myself as package maintainer

Signed-off-by: Christian Schoenebeck <christian.schoenebeck@gmail.com>
SVN-Revision: 48271

8 years agobuild: add @APACHE download facility
Felix Fietkau [Sun, 17 Jan 2016 10:47:32 +0000 (10:47 +0000)]
build: add @APACHE download facility

The Apache Software Foundation offers diverse download mirros.

For packaging Apache software a new alias @APACHE is defined.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
SVN-Revision: 48270

8 years agolantiq: Fixed brnImage signature for the VGV7510KW22BRN images
Felix Fietkau [Sun, 17 Jan 2016 10:47:10 +0000 (10:47 +0000)]
lantiq: Fixed brnImage signature for the VGV7510KW22BRN images

The old signature only worked because brnboot ignores the leading "5" in
the signature. We can see the correct signature when flashing a brnImage
via recovery web-interface, in this case brnboot reports:
[CGI-Signature Check] buf:[BRNDA6431], sigInFlash:[BRNDA6431]

Thanks to Mathias Kresin for reporting this.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48269

8 years agoar8327: add IGMP Snooping support
Felix Fietkau [Sun, 17 Jan 2016 10:42:46 +0000 (10:42 +0000)]
ar8327: add IGMP Snooping support

This add support for IGMP Snooping on atheros switches (disabled by default),
which avoids flooding the network with multicast data.

Tested on TL-WDR4300: disabling IGMP Snooping results in multicast flooding
on each specific port, enabling it back again prevents each port from
receiving all multicast packets.

Partially based on: http://patchwork.ozlabs.org/patch/418122/

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
SVN-Revision: 48268

8 years agoSupport LibreSSL as provider of openssl
Felix Fietkau [Sun, 17 Jan 2016 10:42:32 +0000 (10:42 +0000)]
Support LibreSSL as provider of openssl

The build scripts check for openssl by grepping the string "OpenSSL" in
the output of openssl version command. LibreSSL fails this test as it
outputs something like "LibreSSL 2.2.4". This patch fix the
prereq-bulid.mk file so that it accepts LibreSSL as openssl provider as
well.

Signed-off-by: Marek Behun <kabel@blackhole.sk>
SVN-Revision: 48267

8 years agobrcm2708: add linux 4.4 support
Felix Fietkau [Sun, 17 Jan 2016 10:42:23 +0000 (10:42 +0000)]
brcm2708: add linux 4.4 support

- random-bcm2708 and spi-bcm2708 have been removed.
- sound-soc-bcm2708-i2s has been upstreamed as sound-soc-bcm2835-i2s.

Let's keep linux 4.1 for a while, since linux 4.4 appears to have some issues
with multicast traffic on RPi ethernet:
https://gist.github.com/Noltari/5b1cfdecce5ed4bc08fd

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
SVN-Revision: 48266

8 years agouml: update to linux 4.4
Felix Fietkau [Sun, 17 Jan 2016 10:41:52 +0000 (10:41 +0000)]
uml: update to linux 4.4

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48265

8 years agolantiq: Switch to the generic mtdsplit parser implementations
Felix Fietkau [Sun, 17 Jan 2016 10:41:46 +0000 (10:41 +0000)]
lantiq: Switch to the generic mtdsplit parser implementations

This removes the arch-specific mtdsplit parsers and enables the generic
implementations for brnImage, EVA and TP-Link instead.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48264

8 years agokernel: mtdsplit_squashfs: Align with the erase-block size
Felix Fietkau [Sun, 17 Jan 2016 10:41:30 +0000 (10:41 +0000)]
kernel: mtdsplit_squashfs: Align with the erase-block size

On most image types the rootfs ends at an erase-block. However, at least
with brnImages this is not the case: while the partitions are aligned
with the erase-block size there is a 12 byte footer at the end of the
partition which must not be touched by any filesystem. This lead to a
rootfs_data partition which was not aligned properly (and thus ended up
being readonly):
0x000000480000-0x00000085a800 : "rootfs_data" (128 KiB EB)

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48263

8 years agokernel: mtdsplit: add support for EVA images
Felix Fietkau [Sun, 17 Jan 2016 10:41:08 +0000 (10:41 +0000)]
kernel: mtdsplit: add support for EVA images

This allows splitting EVA images (usually found in fritz devices). The
firmware will be split into a kernel and a separate rootfs partition.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48262

8 years agokernel: mtdsplit: add support for brnImages
Felix Fietkau [Sun, 17 Jan 2016 10:40:51 +0000 (10:40 +0000)]
kernel: mtdsplit: add support for brnImages

This adds brnImage (used with the brnboot bootloader) firmware parsing
support. brnboot verifies the integrity of the firmware stored on the
"Code Image" partitions by looking at the 12 byte footer at the very end
of the partition. This footer contains the checksum of the original
brnImage (kernel + rootfs/squashfs) and must not be touched (by our JFFS2
rootfs_data - otherwise the image will not be bootable anymore).

Big thanks to Mathias Kresin for analyzing the brnImage structure and
finding out the information how to keep images valid even when adding a
nested rootfs_data partition.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 48261

8 years agokernel: disable bogus DEVTMPFS config overrides
Felix Fietkau [Sun, 17 Jan 2016 10:40:36 +0000 (10:40 +0000)]
kernel: disable bogus DEVTMPFS config overrides

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48260

8 years agodownload.pl: fix openssl sha256 digest output
Felix Fietkau [Sat, 16 Jan 2016 10:24:15 +0000 (10:24 +0000)]
download.pl: fix openssl sha256 digest output

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48259

8 years agoustream-ssl: fix copy&paste mistake in mbedtls variant title
Felix Fietkau [Sat, 16 Jan 2016 09:14:03 +0000 (09:14 +0000)]
ustream-ssl: fix copy&paste mistake in mbedtls variant title

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48258

8 years agopolarssl: update to 1.3.16, fixes intermediate certificate validation
Felix Fietkau [Sat, 16 Jan 2016 00:20:05 +0000 (00:20 +0000)]
polarssl: update to 1.3.16, fixes intermediate certificate validation

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48257

8 years agoustream-ssl: update to the latest version, adds mbedtls variant
Felix Fietkau [Sat, 16 Jan 2016 00:20:01 +0000 (00:20 +0000)]
ustream-ssl: update to the latest version, adds mbedtls variant

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48256

8 years agolibubox: update to the latest version, adds lua/ustream fixes and extends usock
Felix Fietkau [Sat, 16 Jan 2016 00:19:54 +0000 (00:19 +0000)]
libubox: update to the latest version, adds lua/ustream fixes and extends usock

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48255

8 years agombedtls: update to version 2.2.1
Felix Fietkau [Sat, 16 Jan 2016 00:19:47 +0000 (00:19 +0000)]
mbedtls: update to version 2.2.1

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48254

8 years agoscripts/download.pl: add SHA-256 support
Felix Fietkau [Sat, 16 Jan 2016 00:19:41 +0000 (00:19 +0000)]
scripts/download.pl: add SHA-256 support

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48253

8 years agopackages: bump uboot-sunxi to 2016.01
Zoltan Herpai [Fri, 15 Jan 2016 20:27:32 +0000 (20:27 +0000)]
packages: bump uboot-sunxi to 2016.01

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 48252

8 years agomwlwifi: fix build with latest mac80211
Felix Fietkau [Fri, 15 Jan 2016 16:58:00 +0000 (16:58 +0000)]
mwlwifi: fix build with latest mac80211

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48251

8 years agomt76: update to the latest version, fixes build with latest mac80211
Felix Fietkau [Fri, 15 Jan 2016 16:53:44 +0000 (16:53 +0000)]
mt76: update to the latest version, fixes build with latest mac80211

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48250

8 years agoimx6: kernel: Add Micrel PHY used on Apalis SOM
Felix Fietkau [Fri, 15 Jan 2016 15:18:23 +0000 (15:18 +0000)]
imx6: kernel: Add Micrel PHY used on Apalis SOM

Signed-off-by: Petr Štetiar <ynezz@true.cz>
Cc: Luka Perkov <luka@openwrt.org>
SVN-Revision: 48249

8 years agoimx6: add 4.4 support
Felix Fietkau [Fri, 15 Jan 2016 15:18:18 +0000 (15:18 +0000)]
imx6: add 4.4 support

Build and boot tested on the following hardware:
 * GW54xx
 * GW53xx
 * GW52xx
 * GW51xx
 * GW552x
 * GW551x

Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
SVN-Revision: 48248

8 years agomac80211: update to wireless-testing 2016-01-10
Felix Fietkau [Fri, 15 Jan 2016 15:03:48 +0000 (15:03 +0000)]
mac80211: update to wireless-testing 2016-01-10

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48247

8 years agodnsmasq: Add option --min-port
Felix Fietkau [Fri, 15 Jan 2016 11:24:15 +0000 (11:24 +0000)]
dnsmasq: Add option --min-port

By default dnsmasq uses random ports for outbound dns queries;
when the minport UCI option is specified the ports used will
always be larger than the specified value.
This is usefull for systems behind firewalls.

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
SVN-Revision: 48244

8 years agodropbear: update version to 2015.71
Felix Fietkau [Fri, 15 Jan 2016 11:24:09 +0000 (11:24 +0000)]
dropbear: update version to 2015.71

Update dropbear to version 2015.71, released on 3 Dec 2015.
Refresh patches.

Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
SVN-Revision: 48243

8 years agokernel: add missing kmod-video-videobuf2's object file in 4.4
Felix Fietkau [Fri, 15 Jan 2016 10:49:21 +0000 (10:49 +0000)]
kernel: add missing kmod-video-videobuf2's object file in 4.4

CONFIG_VIDEOBUF2_CORE produces two object files in Linux >= 4.4.
Conditionally add the new file to FILES of kmod-video-videobuf2.

Signed-off-by: Jan Čermák <jan.cermak@nic.cz>
SVN-Revision: 48242

8 years agoarc770: move arc patches to target/linux/generic
Felix Fietkau [Fri, 15 Jan 2016 10:49:11 +0000 (10:49 +0000)]
arc770: move arc patches to target/linux/generic

Given those patches are relevant to any ARC platform and even
ISA version it makes perfect sense for patches to exist
in one place instead of being duplicated for each new ARC-based ASIC.

Note this is a prerequisite for upstreaming of ARC HS38 support in
OpenWRT.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Felix Fietkau <nbd@openwrt.org>
Cc: Jo-Philipp Wich <jow@openwrt.org>
Cc: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 48241

8 years agoarc770: bump linux kernel from 4.3 to 4.4
Felix Fietkau [Fri, 15 Jan 2016 10:49:01 +0000 (10:49 +0000)]
arc770: bump linux kernel from 4.3 to 4.4

This switch involved:
 [1] Regeneration of config (few options went away)
 [2] Regeneration of patches so they apply cleanly (different offsets)
 [3] Update of .dts files because we now explicitly specify
     memory regions in use as opposed to previously used offset
     from 0x8000_0000

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Felix Fietkau <nbd@openwrt.org>
Cc: Jo-Philipp Wich <jow@openwrt.org>
Cc: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 48240

8 years agokernel: compile out SNMP MIB related counting if debloat-proc is active - improves...
Felix Fietkau [Fri, 15 Jan 2016 10:47:14 +0000 (10:47 +0000)]
kernel: compile out SNMP MIB related counting if debloat-proc is active - improves network stack performance

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48239

8 years agoprocd: align early init PATH with rest of OpenWrt
Jo-Philipp Wich [Thu, 14 Jan 2016 12:52:18 +0000 (12:52 +0000)]
procd: align early init PATH with rest of OpenWrt

This fixes diverging executable search paths observed in programs
launched throughn etifd which in turn inherited the search path from
procd early on boot.

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 48235

8 years agomt76: update to the latest version, fixes signal strength reporting
Felix Fietkau [Thu, 14 Jan 2016 11:59:10 +0000 (11:59 +0000)]
mt76: update to the latest version, fixes signal strength reporting

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48234

8 years agoath9k: add support for a newer tx gain table used in some TP-Link devices
Felix Fietkau [Thu, 14 Jan 2016 02:17:49 +0000 (02:17 +0000)]
ath9k: add support for a newer tx gain table used in some TP-Link devices

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48233

8 years agomt76: update to the latest version, adds mac address fixes and some performance tuning
Felix Fietkau [Thu, 14 Jan 2016 01:00:42 +0000 (01:00 +0000)]
mt76: update to the latest version, adds mac address fixes and some performance tuning

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48232

8 years agokernel: backport perf build fixes to 4.1
Felix Fietkau [Thu, 14 Jan 2016 01:00:38 +0000 (01:00 +0000)]
kernel: backport perf build fixes to 4.1

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48231

8 years agokernel: remove obsolete config symbols from 4.4
Felix Fietkau [Thu, 14 Jan 2016 01:00:13 +0000 (01:00 +0000)]
kernel: remove obsolete config symbols from 4.4

Only remove old symbols for 4.4 since older Kernels might be
removed sooner/later and 4.4 is the next LTS

build tested against:
 omap 4.4 CONFIG_ALL_KMODS
 x86_64

List of enabled & orphaned symbols:

-CONFIG_ARPD=y
-CONFIG_BT_L2CAP=y
-CONFIG_BT_SCO=y
-CONFIG_CAVIUM_OCTEON_HELPER=y
-CONFIG_DEFAULT_MESSAGE_LOGLEVEL=4
-CONFIG_DISABLE_DEV_COREDUMP=y
-CONFIG_EXPERIMENTAL=y
-CONFIG_EXT4_FS_XATTR=y
-CONFIG_GENERIC_HARDIRQS=y
-CONFIG_GENERIC_TIME=y
-CONFIG_HOTPLUG=y
-CONFIG_INITRAMFS_COMPRESSION_NONE=y
-CONFIG_INLINE_SPIN_UNLOCK=y
-CONFIG_LEDS_GPIO_OF=y
-CONFIG_LEDS_GPIO_PLATFORM=y
-CONFIG_MFD_SUPPORT=y
-CONFIG_MISC_DEVICES=y
-CONFIG_MTD_CHAR=y
-CONFIG_NETDEV_1000=y
-CONFIG_NET_ETHERNET=y
-CONFIG_NET_PCI=y
-CONFIG_NET_VENDOR_SILICOM=y
-CONFIG_SCSI_MULTI_LUN=y
-CONFIG_UIDGID_STRICT_TYPE_CHECKS=y
-CONFIG_USB_ARCH_HAS_EHCI=y
-CONFIG_USB_ARCH_HAS_OHCI=y
-CONFIG_USB_DEVICEFS=y
-CONFIG_USB_EZUSB=y

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 48230

8 years agolantiq: remove clk_round_rate() patch
Hauke Mehrtens [Wed, 13 Jan 2016 20:46:12 +0000 (20:46 +0000)]
lantiq: remove clk_round_rate() patch

This patch was merged into upstream Linux 4.1.
This fixes #21587 and was introduced in r48223.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 48229

8 years agoar71xx: fix ethernet MAC reset on DMA hang
Felix Fietkau [Wed, 13 Jan 2016 16:22:16 +0000 (16:22 +0000)]
ar71xx: fix ethernet MAC reset on DMA hang

Fully reset the chip like on a full up/down, but without the PHY
statemachine restart.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48228

8 years agoar71xx: extend ethernet DMA stuck check to all ar724x (and newer) chips
Felix Fietkau [Wed, 13 Jan 2016 16:22:05 +0000 (16:22 +0000)]
ar71xx: extend ethernet DMA stuck check to all ar724x (and newer) chips

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48227

8 years agoubus: update to the latest version, fixes an acl related crash
Felix Fietkau [Wed, 13 Jan 2016 11:32:41 +0000 (11:32 +0000)]
ubus: update to the latest version, fixes an acl related crash

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48226