openwrt/staging/stintel.git
8 years agoimage.mk: use new pure make functions for sanitizing image filenames
Felix Fietkau [Sun, 3 Jan 2016 20:58:02 +0000 (20:58 +0000)]
image.mk: use new pure make functions for sanitizing image filenames

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48106

8 years agobuild: add pure make tolower/toupper funtions that do not require shell calls
Felix Fietkau [Sun, 3 Jan 2016 20:57:53 +0000 (20:57 +0000)]
build: add pure make tolower/toupper funtions that do not require shell calls

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48105

8 years agomac80211: add rtl8xxxu
Felix Fietkau [Sun, 3 Jan 2016 20:57:38 +0000 (20:57 +0000)]
mac80211: add rtl8xxxu

OpenWrt is often used to develop or test new devices
Some users might want to test and help to improve
this new driver

upstream commit notice 26f1fad29ad973b0fb26a9ca3dcb2a73dde781aa

New driver: rtl8xxxu (mac80211)
This is an alternate driver for a number of Realtek WiFi USB devices,
including RTL8723AU, RTL8188CU, RTL8188RU, RTL8191CU, and RTL8192CU.
It was written from scratch utilizing the Linux mac80211 stack.

After spending months cleaning up the vendor provided rtl8723au
driver, which comes with it's own 802.11 stack included, I decided to
rewrite this driver from the bottom up.

Many thanks to Johannes Berg for 802.11 insights and help and Larry
Finger for help with the vendor driver.

The full git log for the development of this driver can be found here:
git git://git.kernel.org/pub/scm/linux/kernel/git/jes/linux.git
    branch rtl8723au-mac80211

This driver is still under development, but has proven to be very
stable for me. It currently supports station mode only. It has support
for OFDM and CCK rates. It does lack certain features found in the
staging driver, such as power management, AMPDU, and 40MHz channel
support. In addition it does not support AD-HOC, AP, and monitor mode
support at this point.

The driver is known to work with the following devices:
Lenovo Yoga (rtl8723au)
TP-Link TL-WN823N (rtl8192cu)
Etekcity 6R (rtl8188cu)
Daffodil LAN03 (rtl8188cu)
Alfa AWUS036NHR (rtl8188ru)

Compile tested only

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 48104

8 years agomac80211: iwlwifi: update firmware to ucode -16 version
Felix Fietkau [Sun, 3 Jan 2016 20:57:31 +0000 (20:57 +0000)]
mac80211: iwlwifi: update firmware to ucode -16 version

We don't use Kernel drivers but ucode -16 is
usable on Kernel 4.3+ - with backports and mac80211
this should work on older Kernel versions too.

Intel does not provide a changelog.

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 48103

8 years agomac80211: mt7601u use linux-firmware
Felix Fietkau [Sun, 3 Jan 2016 20:57:23 +0000 (20:57 +0000)]
mac80211: mt7601u use linux-firmware

upstream added mt7601u firmware with commit
9df5430a3c1bf4fa92d9198b6d7d8713408ff6bc Add firmware for mt7601u. version 34

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 48102

8 years agoar71xx: ensure that LOADADDR is passed to lzma-loader
Felix Fietkau [Sun, 3 Jan 2016 20:57:16 +0000 (20:57 +0000)]
ar71xx: ensure that LOADADDR is passed to lzma-loader

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48101

8 years agopackage/network/config/gre: ipv6 gre kmod package name was wrong
Felix Fietkau [Sun, 3 Jan 2016 20:57:06 +0000 (20:57 +0000)]
package/network/config/gre: ipv6 gre kmod package name was wrong

Source package gre was depending on kmod-ip6-gre, however the actual
kernel module package that is created is kmod-gre6.  Therefore
update (source) package gre for ipv6 gre support.

Signed-off-by: Daniel Dickinson <openwrt@daniel.thecshore.com>
SVN-Revision: 48100

8 years agowpa_supplicant: set regulatory domain the same way as hostapd
Felix Fietkau [Sun, 3 Jan 2016 20:56:57 +0000 (20:56 +0000)]
wpa_supplicant: set regulatory domain the same way as hostapd

In sta-only configuration, wpa_supplicant needs correct regulatory
domain because otherwise it may skip channel of its AP during scan.

Another alternative is to fix "iw reg set" in mac80211 netifd script.
Currently it fails if some phy has private regulatory domain which
matches configured one.

Signed-off-by: Dmitry Ivanov <dima@ubnt.com>
SVN-Revision: 48099

8 years agoiproute2: update to 4.3.0
Felix Fietkau [Sun, 3 Jan 2016 20:56:45 +0000 (20:56 +0000)]
iproute2: update to 4.3.0

iproute2-4.0 had connmark support added by nbd.  This does not work
with 4.x kernels.  iproute2-4.3 is the latest version and has his
changes mainlined.  This patch updates the package to iproute2-4.3
and fixes the patches so that it compiles.  This should resolve
ticket #21374.

Signed-off-by: Rob Mosher <nyt-openwrt@countercultured.net>
SVN-Revision: 48098

8 years agoar71xx: fix factory images for WNR2200 (#19807)
Felix Fietkau [Sun, 3 Jan 2016 14:45:26 +0000 (14:45 +0000)]
ar71xx: fix factory images for WNR2200 (#19807)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48085

8 years agotarget/sdk: Fix filename with external toolchain
Felix Fietkau [Sun, 3 Jan 2016 14:45:18 +0000 (14:45 +0000)]
target/sdk: Fix filename with external toolchain

When building SDK with external toolchain, don't
include _gcc in the filename

Signed-off-by: Daniel Dickinson <openwrt@daniel.thecshore.com>
SVN-Revision: 48084

8 years agoimages imagebuilder: Allow to add sanitized extra name
Felix Fietkau [Sun, 3 Jan 2016 14:44:56 +0000 (14:44 +0000)]
images imagebuilder: Allow to add sanitized extra name

For final output image names allow user to add an
extra string (which is sanitized).  This is particularly
useful with ImageBuilder where you may generate multiple
images from the same base and for the same board,
with different package selections and additional files
(via FILES=).

Signed-off-by: Daniel Dickinson <openwrt@daniel.thecshore.com>
SVN-Revision: 48083

8 years agoimages: use configured distribution name for output image names
Felix Fietkau [Sun, 3 Jan 2016 14:44:48 +0000 (14:44 +0000)]
images: use configured distribution name for output image names

We allow to configure the version distribution name;
let's also use it for the final image filenames.

Signed-off-by: Daniel Dickinson <openwrt@daniel.thecshore.com>
SVN-Revision: 48082

8 years agotargets: Use configured distribution name for SDK, IB, and Toolchain tarball names
Felix Fietkau [Sun, 3 Jan 2016 14:44:42 +0000 (14:44 +0000)]
targets: Use configured distribution name for SDK, IB, and Toolchain tarball names

We allow to configure the version distribution name;
let's also use it for the tarballs (SDK, ImageBuilder,
and SDK).

Signed-off-by: Daniel Dickinson <openwrt@daniel.thecshore.com>
SVN-Revision: 48081

8 years agotarget/toolchain: Fix tarball filename
Felix Fietkau [Sun, 3 Jan 2016 14:44:36 +0000 (14:44 +0000)]
target/toolchain: Fix tarball filename

Toolchain makefile was missing include of version.mk
so the openwrt version was missing from the filename.
This patch fixes that.

Signed-off-by: Daniel Dickinson <openwrt@daniel.thecshore.com>
SVN-Revision: 48080

8 years agobuild: for cmake, always use ccache from staging dir instead of relying on the host
Felix Fietkau [Sun, 3 Jan 2016 14:44:30 +0000 (14:44 +0000)]
build: for cmake, always use ccache from staging dir instead of relying on the host

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48079

8 years agobusybox: adjust convert_menuconfig.pl to emit relative path references for Config...
Felix Fietkau [Sun, 3 Jan 2016 11:38:31 +0000 (11:38 +0000)]
busybox: adjust convert_menuconfig.pl to emit relative path references for Config.in files and refresh generated files

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48078

8 years agobusybox: fix handling of special characters in path when converting menuconfig
Felix Fietkau [Sun, 3 Jan 2016 11:36:38 +0000 (11:36 +0000)]
busybox: fix handling of special characters in path when converting menuconfig

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48077

8 years agomalta: fix generating network configuration
Felix Fietkau [Sun, 3 Jan 2016 11:36:26 +0000 (11:36 +0000)]
malta: fix generating network configuration

With the default network config provided by package base-files present,
board_detect won't create a new config from board.json.

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
SVN-Revision: 48076

8 years agoar71xx: another fix to wndr3700_board_detect
Felix Fietkau [Sun, 3 Jan 2016 11:09:45 +0000 (11:09 +0000)]
ar71xx: another fix to wndr3700_board_detect

My last bugfix (r47538) introduced a new bug in wndr3700_board_detect
(again...).

Assigning the result of ar71xx_get_mtd_offset_size_format to the model
variable before stripping of garbage using awk will cause all NUL bytes to
be removed before awk is applied, leading to model strings like
"NETGEAR WNDRMACv2NETGEAR", where a NUL byte after the v2 is supposed to
terminate the string.

Fix by calling ar71xx_get_mtd_offset_size_format twice, once piping to awk
directly.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 48075

8 years agoledtrig-netdev: don't stop timer on events for different interfaces
Rafał Miłecki [Sat, 2 Jan 2016 20:38:56 +0000 (20:38 +0000)]
ledtrig-netdev: don't stop timer on events for different interfaces

This fixes regression introduced in my recent ledtrig-netdev commit.
Events triggered by different interfaces were stopping timer so it
wasn't working for tx/rx mode.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 48072

8 years agoperf: select objdump instead of the full binutils package
Felix Fietkau [Sat, 2 Jan 2016 16:19:39 +0000 (16:19 +0000)]
perf: select objdump instead of the full binutils package

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48071

8 years agoramips: clean up and refresh kernel patches
Felix Fietkau [Sat, 2 Jan 2016 14:48:27 +0000 (14:48 +0000)]
ramips: clean up and refresh kernel patches

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48070

8 years agoramips: fix performance counters on mt7621
Felix Fietkau [Sat, 2 Jan 2016 14:48:11 +0000 (14:48 +0000)]
ramips: fix performance counters on mt7621

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48069

8 years agoperf: add myself as a maintainer
Felix Fietkau [Sat, 2 Jan 2016 14:48:04 +0000 (14:48 +0000)]
perf: add myself as a maintainer

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48068

8 years agoperf: fix musl compatibility
Felix Fietkau [Sat, 2 Jan 2016 14:47:57 +0000 (14:47 +0000)]
perf: fix musl compatibility

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48067

8 years agoperf: fix parallel build support, explicitly disable more library dependencies
Felix Fietkau [Sat, 2 Jan 2016 14:47:48 +0000 (14:47 +0000)]
perf: fix parallel build support, explicitly disable more library dependencies

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48066

8 years agoperf: build in a copy of the source dir instead of relying on make clean
Felix Fietkau [Sat, 2 Jan 2016 14:47:42 +0000 (14:47 +0000)]
perf: build in a copy of the source dir instead of relying on make clean

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48065

8 years agoralink: more esw_rt3050 fixes
John Crispin [Sat, 2 Jan 2016 11:44:08 +0000 (11:44 +0000)]
ralink: more esw_rt3050 fixes

the IRQ was not setup properly causing an oops

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 48064

8 years agobcm53xx: switch to simpler devm_gpiod_get in USB driver
Rafał Miłecki [Sat, 2 Jan 2016 10:49:46 +0000 (10:49 +0000)]
bcm53xx: switch to simpler devm_gpiod_get in USB driver

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 48062

8 years agokernel: make bcm47xxpart work with TRX flashed at bad block(s)
Rafał Miłecki [Sat, 2 Jan 2016 00:33:48 +0000 (00:33 +0000)]
kernel: make bcm47xxpart work with TRX flashed at bad block(s)

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 48061

8 years agokernel: drop bcm47xxpart patch for Xiaomi
Rafał Miłecki [Sat, 2 Jan 2016 00:33:31 +0000 (00:33 +0000)]
kernel: drop bcm47xxpart patch for Xiaomi

It was never tested, most likely not working (because of le32_to_cpu)
and not upstreamed.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 48060

8 years agogeneric/4.4: update to 4.4-rc7
Imre Kaloz [Fri, 1 Jan 2016 22:02:47 +0000 (22:02 +0000)]
generic/4.4: update to 4.4-rc7

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 48059

8 years agoV2 ar71xx : Add Support for the Bitmain Antrouter R1
John Crispin [Fri, 1 Jan 2016 21:21:13 +0000 (21:21 +0000)]
V2 ar71xx : Add Support for the Bitmain Antrouter R1

This patch adds support for the Bitmain Antrouter R1

http://wiki.openwrt.org/toh/bitmain/r1

Signed-off-by: L. D. Pinney <ldpinney@gmail.com>
SVN-Revision: 48058

8 years agolantiq: ltq-vmmc: Fix compilation with linux 4.4
John Crispin [Fri, 1 Jan 2016 21:21:04 +0000 (21:21 +0000)]
lantiq: ltq-vmmc: Fix compilation with linux 4.4

gpio.h needs the linux/ prefix with linux 4.4.

SVN-Revision: 48057

8 years agolantiq: Add the xbar to vr9.dts
John Crispin [Fri, 1 Jan 2016 21:21:00 +0000 (21:21 +0000)]
lantiq: Add the xbar to vr9.dts

linux 4.4 (since commit 08b3c894e56580b8ed3e601212a25bda974c3cc2
"MIPS: lantiq: Disable xbar fpi burst mode") requires that the xbar is
defined in the .dts of vrx200 (VR9) SoCs.

SVN-Revision: 48056

8 years agoramips: Remove LED GPIO definitions in MQmaker WiTi device tree
John Crispin [Fri, 1 Jan 2016 21:20:54 +0000 (21:20 +0000)]
ramips: Remove LED GPIO definitions in MQmaker WiTi device tree

LED's were defined wrong in the device tree file, they are hardware driven because they are connected directly to the switch chip and wireless chips respectively, thus no GPIO addresses are assigned to them. It is safe to remove them from the device tree file to stop confusion.

Signed-off-by: Sebastian Careba <nitroshift@yahoo.com>
SVN-Revision: 48055

8 years agoltq-vdsl-app: re-add lowlevel settings
John Crispin [Fri, 1 Jan 2016 21:20:24 +0000 (21:20 +0000)]
ltq-vdsl-app: re-add lowlevel settings

Add back a slightly modified version of the lowlevel settings which
where removed with r46920.

In compare to the old lowlevel settings, the B43c tone is added to
tone_adsl_b and tone_adsl_bv.

If an unsupported tone value is used, the auto probing mode is used, in
compare to the fallback to tone_adsl_av and tone_vdsl_av with the old
lowlevel settings.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48054

8 years agoltq-vdsl-app: enable G.993.5 XTSE bit by default
John Crispin [Fri, 1 Jan 2016 21:20:16 +0000 (21:20 +0000)]
ltq-vdsl-app: enable G.993.5 XTSE bit by default

According to ITU-T G.997.1 Amendment 2 (04/2013) section 2.1, bit 3 of
XTSE octet 8 either allow or denies the initialization of G.993.5.

Even if the current redistributable xDSL firmware doesn't include
G.993.5 vectoring support, enable this bit by default to allow people to
get their G.993.5 line working using a custom xDSL firmware.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48053

8 years agoltq-vdsl-app: let the driver/app probe the xtse on missing annex
John Crispin [Fri, 1 Jan 2016 21:20:08 +0000 (21:20 +0000)]
ltq-vdsl-app: let the driver/app probe the xtse on missing annex

r47933 revealed that the driver/app in combination with the chosen
firmware does a good job in selecting a working xtse.

Use this probing mode if no annex is specified.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48052

8 years agoltq-vdsl-app: add/enable missing G.993.2 XTSE bits
John Crispin [Fri, 1 Jan 2016 21:20:02 +0000 (21:20 +0000)]
ltq-vdsl-app: add/enable missing G.993.2 XTSE bits

This patch adds the missing VDSL2 bits to the annex specific XTSE (like
it should be according to the comments above the XTSE bits).

Since r47933 it's mandatory to remove the annex option to switch to
VDSL2 (only) operation mode.

As shown by ticket #21436 and a few mails I received personally, even
experienced users are not aware that they have to remove the annex
option to get their VDSL2 line working and as shown by this patch it
doesn't need to be that "complicated".

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48051

8 years agoltq-vdsl-app: use the final xtse format
John Crispin [Fri, 1 Jan 2016 21:19:55 +0000 (21:19 +0000)]
ltq-vdsl-app: use the final xtse format

This way we can drop the call to sed.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48050

8 years agoralink: add WRTnode2 platform and led support
John Crispin [Fri, 1 Jan 2016 21:19:50 +0000 (21:19 +0000)]
ralink: add WRTnode2 platform and led support

Signed-off-by: xzmu <xzmu@wrtnode.com>
SVN-Revision: 48049

8 years agoledtrig-netdev: add support for renamed devices
John Crispin [Fri, 1 Jan 2016 21:19:41 +0000 (21:19 +0000)]
ledtrig-netdev: add support for renamed devices

The ppp0 interface is renamed after the connection is established. Due
to a missing NETDEV_REGISTER event, the ledtrig-netdev isn't aware of
the renamed interface and literally ignores the device
(no tx/rx indication, led isn't switched off with 'ifdown wan').

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48048

8 years agolantiq: match default xdsl mode and default wan interface
John Crispin [Fri, 1 Jan 2016 21:19:32 +0000 (21:19 +0000)]
lantiq: match default xdsl mode and default wan interface

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48047

8 years agolantiq: convert vendorid into human readable form
John Crispin [Fri, 1 Jan 2016 21:19:27 +0000 (21:19 +0000)]
lantiq: convert vendorid into human readable form

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48046

8 years agolantiq: add the current profile to the stats output
John Crispin [Fri, 1 Jan 2016 21:19:21 +0000 (21:19 +0000)]
lantiq: add the current profile to the stats output

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48045

8 years agolantiq: TDW8980 - use devicename:colour:function led naming scheme
John Crispin [Fri, 1 Jan 2016 21:19:12 +0000 (21:19 +0000)]
lantiq: TDW8980 - use devicename:colour:function led naming scheme

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48044

8 years agolantiq: P2812HNUFX - move usb leds to P2812HNU-F1
John Crispin [Fri, 1 Jan 2016 21:18:59 +0000 (21:18 +0000)]
lantiq: P2812HNUFX - move usb leds to P2812HNU-F1

The P2812HNU-F3 doesn't have usb leds. Only the P2812HNU-F1 has those
leds.

Reported-by: Sylwester Petela <sscapi@gmail.com>
Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48043

8 years agolantiq: use devicename:colour:function led naming scheme
John Crispin [Fri, 1 Jan 2016 21:18:54 +0000 (21:18 +0000)]
lantiq: use devicename:colour:function led naming scheme

The leds of the following boards are not renamed due to lack of
manuals/informations:

- ARV7519PW
- ARV7510PW22
- ARV4510PW

The leds of the ARV4518PWR01* boards are unchanged, since the leds doesn't
match the leds from the manual or pictures (e.g. there shouldn't be a wps led).

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48042

8 years agolantiq: add support for indicating the boot state using three leds
John Crispin [Fri, 1 Jan 2016 21:18:47 +0000 (21:18 +0000)]
lantiq: add support for indicating the boot state using three leds

The BTHOMEHUBV5A has a RGB power led, where every colour is perfect to
indicate the current boot state. This patch adds support for such cases.

The existing led sequences should be the same as before.

Boards which are using a led different from power (like TDW89x0) are
changed to switch of the led after boot

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48041

8 years agolantiq: use dsl led defined in DTS
John Crispin [Fri, 1 Jan 2016 21:18:38 +0000 (21:18 +0000)]
lantiq: use dsl led defined in DTS

dsl_control (dsl_notify.sh) is the only process which is aware of the
state of the atm/ptm interface. Use the dsl led exclusive for the dsl
line state.

On boards which don't have a distinct internet and a dsl led, let the
netdev status of the atm interface trigger the shared led.

Triggering the shared led according to the status of the ppp interface
isn't suitable, since the led would be switched of if the ppp
connection goes down, but the line is still in sync.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48040

8 years agolantiq: use power leds defined in DTS
John Crispin [Fri, 1 Jan 2016 21:18:30 +0000 (21:18 +0000)]
lantiq: use power leds defined in DTS

Use the power leds defined in the dts file instead of hardcoded led names.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48039

8 years agolantiq: use default leds defined in DTS
John Crispin [Fri, 1 Jan 2016 21:18:21 +0000 (21:18 +0000)]
lantiq: use default leds defined in DTS

Remove all now double defined leds from the led board file. Use pppoe
as default for all broadband connections, since it's the default in
OpenWrt now.

Rename the the wifi leds to make sure, the not applicable default
values get overwritten.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48038

8 years agolantiq: define default leds in dts files
John Crispin [Fri, 1 Jan 2016 21:18:08 +0000 (21:18 +0000)]
lantiq: define default leds in dts files

- ARV7525PW: use the power led as dsl led as done by the stock firmware
- FRITZ3370: use the info led as internet led
- FRITZ7320: use the power led as dsl led as done by the stock firmware

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48037

8 years agolantiq: ARV752DPW - fix dts file
John Crispin [Fri, 1 Jan 2016 21:18:00 +0000 (21:18 +0000)]
lantiq: ARV752DPW - fix dts file

Use the same led logic and labels as the OEM firmware (red = okay,
blue = failure).

Add the red internet led.

Remove missing usb led workaround. The workaround shouldn't be in the
default configuration.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48036

8 years agolantiq: ARV752DPW22 - fix dts file
John Crispin [Fri, 1 Jan 2016 21:17:55 +0000 (21:17 +0000)]
lantiq: ARV752DPW22 - fix dts file

No need to switch (and keep) on all leds at boot. Use the same led
logic and labels as the OEM firmware (red = okay, blue = failure).

Add the red internet led.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48035

8 years agolantiq: DGN1000B - fix typo in dts file
John Crispin [Fri, 1 Jan 2016 21:17:50 +0000 (21:17 +0000)]
lantiq: DGN1000B - fix typo in dts file

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48034

8 years agolantiq: ARV4518PWR01* move redundant parts to dtsi
John Crispin [Fri, 1 Jan 2016 21:17:42 +0000 (21:17 +0000)]
lantiq: ARV4518PWR01* move redundant parts to dtsi

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48033

8 years agolantiq: BTHOMEHUBV2B - remove missing usb led workaround
John Crispin [Fri, 1 Jan 2016 21:17:34 +0000 (21:17 +0000)]
lantiq: BTHOMEHUBV2B - remove missing usb led workaround

The workaround shouldn't be in the default configuration.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48032

8 years agolantiq: P2812HNUFX - move leds to dtsi
John Crispin [Fri, 1 Jan 2016 21:17:26 +0000 (21:17 +0000)]
lantiq: P2812HNUFX - move leds to dtsi

Beside the used labels, they are the same on both boards.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48031

8 years agoralink: fix rt3050/5350 ethernet switch support
John Crispin [Fri, 1 Jan 2016 21:17:21 +0000 (21:17 +0000)]
ralink: fix rt3050/5350 ethernet switch support

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 48030

8 years agoramips: set the wifi mac address to the same as LAN on witi
Felix Fietkau [Fri, 1 Jan 2016 18:58:02 +0000 (18:58 +0000)]
ramips: set the wifi mac address to the same as LAN on witi

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48029

8 years agomt76: update to the latest version from the mt7603 branch, adds support for overridin...
Felix Fietkau [Fri, 1 Jan 2016 18:57:54 +0000 (18:57 +0000)]
mt76: update to the latest version from the mt7603 branch, adds support for overriding the mac address via OF

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48028

8 years agoramips: add a script to fix up the broken mac address on the witi board
Felix Fietkau [Fri, 1 Jan 2016 18:57:39 +0000 (18:57 +0000)]
ramips: add a script to fix up the broken mac address on the witi board

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48027

8 years agokernel: fold of_get_mac_address_mtd into of_get_mac_address
Felix Fietkau [Fri, 1 Jan 2016 18:57:29 +0000 (18:57 +0000)]
kernel: fold of_get_mac_address_mtd into of_get_mac_address

This makes it easier to keep drivers in sync with upstream

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48026

8 years agoramips: fix kernel configuration for mt7628
Felix Fietkau [Fri, 1 Jan 2016 15:20:38 +0000 (15:20 +0000)]
ramips: fix kernel configuration for mt7628

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48025

8 years agoramips: add missing clk_round_rate function to fix buildbot errors
Felix Fietkau [Fri, 1 Jan 2016 14:33:09 +0000 (14:33 +0000)]
ramips: add missing clk_round_rate function to fix buildbot errors

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48024

8 years agoramips: fix inverted LEDs for RE6500
Felix Fietkau [Fri, 1 Jan 2016 12:27:45 +0000 (12:27 +0000)]
ramips: fix inverted LEDs for RE6500

Signed-off-by: Bjoern Dobe <bjoern@dobecom.de>
SVN-Revision: 48022

8 years agoramips: add mac address assignment for eth0 on RE6500
Felix Fietkau [Fri, 1 Jan 2016 12:26:38 +0000 (12:26 +0000)]
ramips: add mac address assignment for eth0 on RE6500

Signed-off-by: Bjoern Dobe <bjoern@dobecom.de>
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48021

8 years agoramips: fix firmware partition size
Felix Fietkau [Fri, 1 Jan 2016 12:26:30 +0000 (12:26 +0000)]
ramips: fix firmware partition size

Signed-off-by: Bjoern Dobe <bjoern@dobecom.de>
SVN-Revision: 48020

8 years agoramips: fix memory size for RE6500
Felix Fietkau [Fri, 1 Jan 2016 12:26:26 +0000 (12:26 +0000)]
ramips: fix memory size for RE6500

Signed-off-by: Bjoern Dobe <bjoern@dobecom.de>
SVN-Revision: 48019

8 years agokernel: fix duplication of sch_netem in multiple packages (#21479)
Felix Fietkau [Fri, 1 Jan 2016 12:26:09 +0000 (12:26 +0000)]
kernel: fix duplication of sch_netem in multiple packages (#21479)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48018

8 years agodante: fix MD5SUM
Nicolas Thill [Thu, 31 Dec 2015 09:20:59 +0000 (09:20 +0000)]
dante: fix MD5SUM

MD5SUM is wrong, it was not updated during last update to v1.4.1.

Thanks to Daniel Dickinson <openwrt@daniel.thecshore.com> for reporting it.

Signed-off-by: Nicolas Thill <nico@openwrt.org>
SVN-Revision: 48017

8 years agoimx6: images: ventana: change name of ubi output
Luka Perkov [Tue, 29 Dec 2015 23:40:41 +0000 (23:40 +0000)]
imx6: images: ventana: change name of ubi output

Change the name of the .ubi produced to strip out the word 'factory'. This is
mainly due to the fact that there is no difference between the Ventana 'factory'
image vs the standard image.

Name changes from:
openwrt-imx6-ventana-squashfs.nand-factory_<size>.ubi to
openwrt-imx6-ventana-squashfs-nand_<size>.ubi

Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
SVN-Revision: 48016

8 years agouboot-imx6: add support for i.MX6q SABRE board
Luka Perkov [Mon, 28 Dec 2015 04:56:33 +0000 (04:56 +0000)]
uboot-imx6: add support for i.MX6q SABRE board

U-boot ships with the `mx6qsabresd` configuration file to configure a
bootloader suitable for booting the Freescale i.MX6q SABRE evaluation board
from SD card. Expose this option in the uboot-imx6 Makefile.

Signed-off-by: Keno Fischer <keno@juliacomputing.com>
SVN-Revision: 48014

8 years agoimx6: fix device-tree mac address assignment for ventana sky2/eth1
Luka Perkov [Mon, 28 Dec 2015 04:56:27 +0000 (04:56 +0000)]
imx6: fix device-tree mac address assignment for ventana sky2/eth1

Some Ventana boards have a Marvell sky2 GigE controller as eth1 however
assigning the mac address through device-tree is difficult because the
PCI slot can move around depending on board configuration and slot population.

To work around this we add a patch to the sky2 driver to allow accessing its
mac address via a device-tree alias.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
SVN-Revision: 48013

8 years agoimx6: add i2c retries for Ventana /dev/i2c-0
Luka Perkov [Mon, 28 Dec 2015 04:56:13 +0000 (04:56 +0000)]
imx6: add i2c retries for Ventana /dev/i2c-0

On Ventana boards the Gateworks System Controller is the only device on I2C1
(/dev/i2c-0) and it can NAK transfers if it is busy in an ADC loop. Because
this is a multi-function device with several slave addresses it is best to
add retries at the controller level instead of within each slave driver. This
adds a patch that adds 3 retries for i2c transactions only for Ventana boards
and only for I2C1

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
SVN-Revision: 48012

8 years agoimx6: imx: thermal: use CPU temperature grade info for thresholds
Luka Perkov [Mon, 28 Dec 2015 04:56:07 +0000 (04:56 +0000)]
imx6: imx: thermal: use CPU temperature grade info for thresholds

This is a backport of a2291badc355d58ead5c19ae0609468947416040 from thermal-soc
accepted upstream.

The IMX6Q/IMX6DL SoC's have a 2-bit temperature grade stored in OTP. Instead
of assuming 85C for passive cooling threshold and 100C for critical base
these thresholds off the thermal gade max CPU temperature:
 - passive threshold: max - 10C
 - critical threshold: max - 5C

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 48011

8 years agoimx6: Add pwm nodes
Luka Perkov [Mon, 28 Dec 2015 04:55:54 +0000 (04:55 +0000)]
imx6: Add pwm nodes

Upstream status: Accepted

Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
SVN-Revision: 48010

8 years agoimx6: ventana: add ecspi3 host controller for GW52xx
Luka Perkov [Mon, 28 Dec 2015 04:55:48 +0000 (04:55 +0000)]
imx6: ventana: add ecspi3 host controller for GW52xx

Certain board revisions of the GW52xx support an SPI host controller with
a single chip-select going to an off board connector.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
SVN-Revision: 48009

8 years agoimx6: Add 4.3 support
Luka Perkov [Mon, 28 Dec 2015 04:55:45 +0000 (04:55 +0000)]
imx6: Add 4.3 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: 48008

8 years agoimx6: drop lingering files
Luka Perkov [Mon, 28 Dec 2015 04:55:33 +0000 (04:55 +0000)]
imx6: drop lingering files

Since 3.18 kernel support was dropped, remove lingering files.

Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
SVN-Revision: 48007

8 years ago6in4: Corrected tunnelbroker tunnel update URL
Hauke Mehrtens [Sun, 27 Dec 2015 20:42:26 +0000 (20:42 +0000)]
6in4: Corrected tunnelbroker tunnel update URL

Changed the tunnel update URL into format tunnelbrokers
example has, that made it work again. Current method gives "Username/Password
Authentication Failed." when I tried the wget line manually and logread
eventually says also "6in4: update failed". With corrected URL it works fine:
"good 111.222.333.444" or "nochg 111.222.333.444" and logread concurs with
success, and tunnel actually updates.

Tested-by: Vaasa Hacklab <info@vaasa.hacklab.fi>
Signed-off-by: Sami Olmari <sami@olmari.fi>
SVN-Revision: 48006

8 years agoar71xx: fix sysupgrade for wnr2200
Felix Fietkau [Wed, 23 Dec 2015 23:42:18 +0000 (23:42 +0000)]
ar71xx: fix sysupgrade for wnr2200

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 48005

8 years agolibnl: fix warning with poll.h include on musl
John Crispin [Wed, 23 Dec 2015 19:26:26 +0000 (19:26 +0000)]
libnl: fix warning with poll.h include on musl

Warning is:
  #warning redirecting incorrect #include <sys/poll.h> to <poll.

Not a big issue.
But it can be annoying when building with -Werror set.

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
SVN-Revision: 48004

8 years agobinutils: upgrade to version 2.25
John Crispin [Wed, 23 Dec 2015 19:26:11 +0000 (19:26 +0000)]
binutils: upgrade to version 2.25

Fixes build:
config/tc-i386.c: In function 'build_modrm_byte':
config/tc-i386.c:6143:31: error: logical not is only applied to the left hand side of comparison [-Werror=logical-not-parentheses]
         && !op.bitfield.reg64 != 1

Contains this fix:
   https://sourceware.org/ml/binutils/2014-09/msg00095.html

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
SVN-Revision: 48003

8 years agolantiq: fix Image Builder
John Crispin [Wed, 23 Dec 2015 19:25:55 +0000 (19:25 +0000)]
lantiq: fix Image Builder

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 48002

8 years agofstools: update to latest git HEAD
John Crispin [Wed, 23 Dec 2015 19:25:46 +0000 (19:25 +0000)]
fstools: update to latest git HEAD

adds stacked ram overlays.

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 48001

8 years agoca-certificates: update to version 20151214
John Crispin [Wed, 23 Dec 2015 19:25:33 +0000 (19:25 +0000)]
ca-certificates: update to version 20151214

update to version 20151214

Signed-off-by: Christian Schoenebeck <christian.schoenebeck@gmail.com>
SVN-Revision: 48000

8 years agoswconfig: switch kernel PORT_LINK support to SWITCH_TYPE_LINK
John Crispin [Wed, 23 Dec 2015 19:25:02 +0000 (19:25 +0000)]
swconfig: switch kernel PORT_LINK support to SWITCH_TYPE_LINK

As explained earlier, using SWITCH_TYPE_LINK gives more flexibility,
it doesn't require e.g. string parsing to read some data.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 47999

8 years agoswconfig: support receiving SWITCH_TYPE_LINK from kernel
John Crispin [Wed, 23 Dec 2015 19:24:45 +0000 (19:24 +0000)]
swconfig: support receiving SWITCH_TYPE_LINK from kernel

When using cli, print link state the same way kernel used to do it.
This will allow kernel switching PORT_LINK from SWITCH_TYPE_STRING.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 47998

8 years agoswconfig: add SWITCH_TYPE_LINK and support sending link info to user space
John Crispin [Wed, 23 Dec 2015 19:24:30 +0000 (19:24 +0000)]
swconfig: add SWITCH_TYPE_LINK and support sending link info to user space

So far we were sending link data as a string. It got some drawbacks:
1) Didn't allow writing clean user space apps reading link state. It was
   needed to do some screen scraping.
2) Forced whole PORT_LINK communication to be string based. Adding
   support for *setting* port link required passing string and parting
   it in the kernel space.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 47997

8 years agolantiq: fix PCI_DEVICE_ID_LANTIQ_PCIE
John Crispin [Wed, 23 Dec 2015 18:25:22 +0000 (18:25 +0000)]
lantiq: fix PCI_DEVICE_ID_LANTIQ_PCIE

It's not necessary to define PCI_* if pci_ids.h is included a few
lines above.

The change to pci_ids.h doesn't look intentional to me, especially
since the former value is added to the top of ifxmips_fixup_pcie.c.

Both changes were introduced with the kernel 4.1 support patches and
were not present in the 3.18 patches.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 47996

8 years agolantiq: ltq-adsl-mei: fix typo
John Crispin [Wed, 23 Dec 2015 18:25:03 +0000 (18:25 +0000)]
lantiq: ltq-adsl-mei: fix typo

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 47995

8 years agolantiq: TDW89x0 - increase spi frequency
John Crispin [Wed, 23 Dec 2015 18:24:56 +0000 (18:24 +0000)]
lantiq: TDW89x0 - increase spi frequency

Use the same max spi frequency as set in u-boot.

According to the datasheets, the Q64-104HIP as well as the Winbond
25Q64FVSIG support spi frequencies up to 50 MHz. During my tests, the
Q64-104HIP couldn't be recognized/initialized if the frequency
was > 40MHz.

Both chips do support fast read as well.

While touching the dts file, I fixed the dtc compiler warnings.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 47994

8 years agolantiq: disable SW_PORTMAP usage in the ethernet driver
John Crispin [Wed, 23 Dec 2015 18:24:43 +0000 (18:24 +0000)]
lantiq: disable SW_PORTMAP usage in the ethernet driver

using the special tag in this way lead to port mirroring for certain types of traffic.
fallback to using th PMAC_EWAN register for the wan portmap.

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 47993

8 years agoledtrig-netdev: reset link status & stats after changing device_name
Rafał Miłecki [Wed, 23 Dec 2015 17:11:49 +0000 (17:11 +0000)]
ledtrig-netdev: reset link status & stats after changing device_name

Previously switching to non-existing device (interface) could result in
leaving LED on.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 47990

8 years agoledtrig-netdev: drop locking from timer callback function
Rafał Miłecki [Wed, 23 Dec 2015 15:45:43 +0000 (15:45 +0000)]
ledtrig-netdev: drop locking from timer callback function

We may just delete timer on every trigger update and then start it again
if needed. This will let us avoid both: races and locking in frequently
called timer callback.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 47987

8 years agoledtrig-netdev: switch rwlock to spinlock
Rafał Miłecki [Wed, 23 Dec 2015 15:45:36 +0000 (15:45 +0000)]
ledtrig-netdev: switch rwlock to spinlock

Read/write lock was adding useless complexity, there wasn't any real
gain in case of this driver.
Also switch to _bh variants to avoid deadlocks.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 47986