openwrt/staging/svanheule.git
8 years agoramips: Fix alphabetical sorting for 02_network
Vittorio Gambaletta (VittGam) [Wed, 4 May 2016 20:04:24 +0000 (22:04 +0200)]
ramips: Fix alphabetical sorting for 02_network

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
8 years agoramips: Don't use a VLAN for the single ethernet port of the A5-V11.
Vittorio Gambaletta (VittGam) [Wed, 4 May 2016 19:22:19 +0000 (21:22 +0200)]
ramips: Don't use a VLAN for the single ethernet port of the A5-V11.

This router only has one ethernet port, so a VLAN is useless here, now that the rt3050 TCP bug that happened without VLANs has been fixed for a very long time.

Add this router to the VLAN-less config that is used by other single-port routers.

Also fix MAC address detection code since this router has no WAN port.

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
8 years agoramips: Fix network for routers without VLANs on eth0.
Vittorio Gambaletta (VittGam) [Wed, 4 May 2016 19:21:41 +0000 (21:21 +0200)]
ramips: Fix network for routers without VLANs on eth0.

Some routers only have one port, so eth0 is used without VLANs for them.

Revision r47720 introduced some changes, but wrongly confused "enable" with "reset".

VLANs need to be disabled for those routers, and the switch may be reset.

Fix this, by explicitly disabling VLANs instead of resetting the switch for these routers.

Also merge duplicate configuration for the "m2m".

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
8 years agoramips: Drop hacky switch workaround for failsafe on rt3x5x and rt5350.
blogic [Mon, 2 May 2016 18:51:20 +0000 (18:51 +0000)]
ramips: Drop hacky switch workaround for failsafe on rt3x5x and rt5350.

The new rt3050 switch driver doesn't have problems with TCP when not
using VLANs.

This piece of code also broke failsafe for all routers where the LAN
port is not wired to port 0 of the internal switch.

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49293

8 years agoramips: Disable all ethernet ports except port 0 on A5-V11.
blogic [Mon, 2 May 2016 18:51:13 +0000 (18:51 +0000)]
ramips: Disable all ethernet ports except port 0 on A5-V11.

Port 0 is the only ethernet port on this router, so disable all other PHYs
in order to save power.

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49292

8 years agoramips: Disable all ethernet ports except port 4 on HT-TM02.
blogic [Mon, 2 May 2016 18:51:09 +0000 (18:51 +0000)]
ramips: Disable all ethernet ports except port 4 on HT-TM02.

Port 4 is the only ethernet port on this router, so disable all other PHYs
in order to save power.

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49291

8 years agoramips: Get rt3050 ethernet ports to be disabled from the device tree.
blogic [Mon, 2 May 2016 18:51:05 +0000 (18:51 +0000)]
ramips: Get rt3050 ethernet ports to be disabled from the device tree.

This patch allows configuring ports to be disabled in the device tree; this
saves power, since disabling ports here actually disables power to ethernet
PHYs.

Line 444 enables all ethernet ports, so line 487 is getting zero ports to be
disabled, except for port 5 in SoCs where this is not implemented as it will
be sticky disabled in register POC0. Because of this, the code will still read
the switch configuration and OR it to the device tree setting.

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49290

8 years agoramips: Fix comment in rt3050 ethernet switch driver.
blogic [Mon, 2 May 2016 18:51:01 +0000 (18:51 +0000)]
ramips: Fix comment in rt3050 ethernet switch driver.

Line 444 is actually enabling all switch ports by setting the disable bits
to 0. This needs to be done because the bootloader sets all ports to disabled
by default (which is the case for at least one router based on RT5350).

So, this patch fixes the comment in line 443.

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49289

8 years agoramips: Fix documentation for the rt3050 switch driver.
blogic [Mon, 2 May 2016 18:50:56 +0000 (18:50 +0000)]
ramips: Fix documentation for the rt3050 switch driver.

The prefix used in the driver is now "mediatek" instead of "ralink".

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49288

8 years agoramips: Fix multicast ICMPv6 for the rt3050 ethernet switch.
blogic [Mon, 2 May 2016 18:50:52 +0000 (18:50 +0000)]
ramips: Fix multicast ICMPv6 for the rt3050 ethernet switch.

The FCT2 esw register should be set to 0x2500C to have "unknown IPv6
multicast" packets broadcasted to every port, instead of dropped.
The previous value only let those packets go through ports 1 and 3.

"Unknown IPv6 multicast" packets include packets needed by ICMPv6 echo
requests addressed to well-known addresses, such as ff02::1 (MAC address
is 33:33:00:00:00:01 in this case).

Please note that by default ICMPv6 echo requests to ff02::1 are not replied
to by the router because of ip6tables considering those packets to be invalid.
But this is another bug/patch. ;)

Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49287

8 years agoar71xx: add TP-Link TL-WR810N support
blogic [Mon, 2 May 2016 18:50:48 +0000 (18:50 +0000)]
ar71xx: add TP-Link TL-WR810N support

This patch adds support for the TP-Link TL-WR810N.
https://wiki.openwrt.org/toh/tp-link/tl-wr810n

Signed-off-by: Jens Steinhauser <jens.steinhauser@gmail.com>
SVN-Revision: 49286

8 years agoubox: turn logd into a separate package
blogic [Mon, 2 May 2016 18:50:40 +0000 (18:50 +0000)]
ubox: turn logd into a separate package

Currently system log is always included as a part of ubox.
Add logd as a seperate package and add it to default packages list.

Signed-off-by: Andrej Vlasic <andrej.vlasic@sartura.hr>
SVN-Revision: 49285

8 years agolantiq: VGV7510KW22BRN - set the phy clock source
blogic [Mon, 2 May 2016 18:50:35 +0000 (18:50 +0000)]
lantiq: VGV7510KW22BRN - set the phy clock source

VGV7510KW2 with VRX288 v1.2 has brnboot 1.8 installed. Starting with
this brnboot version, the "GPHY Clock Source" isn't set anymore by
brnboot, with the result that xrx200-net fails to probe/initialize the
phys.

Use the phy clock source device tree binding to specify the clock source.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49284

8 years agolantiq: add device tree binding for the phy clock source
blogic [Mon, 2 May 2016 18:50:31 +0000 (18:50 +0000)]
lantiq: add device tree binding for the phy clock source

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49283

8 years agolantiq: VGV7510KW22BRN - support dual-image
blogic [Mon, 2 May 2016 18:50:27 +0000 (18:50 +0000)]
lantiq: VGV7510KW22BRN - support dual-image

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49282

8 years agolantiq: handle the dual-firmware layout of brnboot
blogic [Mon, 2 May 2016 18:50:23 +0000 (18:50 +0000)]
lantiq: handle the dual-firmware layout of brnboot

brnboot based devices can have two Image partitions. When flashing
images via the brnboot recovery web interface, the Image partitions are
written alternating.

The current active Image partition is stored in the first byte of the
Primary_Setting partition by using 0x00 for Code_Image_0 and 0x01 for
Code_Image_1.

By using the information about the active "Code Image", it is possible
to ensure that the rootfs belongs to the current booted Image/Kernel.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49281

8 years agolantiq: move partitions into partion table node
blogic [Mon, 2 May 2016 18:50:18 +0000 (18:50 +0000)]
lantiq: move partitions into partion table node

Starting with kernel 4.4, the use of partitions as direct subnodes of the
mtd device is discouraged and only supported for backward compatiblity
reasons.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49280

8 years agolantiq: use sysupgrade.bin postfix for tplink images
blogic [Mon, 2 May 2016 18:50:10 +0000 (18:50 +0000)]
lantiq: use sysupgrade.bin postfix for tplink images

Use the same name for TP-Link images as it was with the old image build
code.

Move the BOARD_ID export to the TP-Link image build recipe, to indicate
that the variable is only related in this context.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49279

8 years agolantiq: VG3503J - use the same PHY led functionality as the OEM firmware
blogic [Mon, 2 May 2016 18:50:06 +0000 (18:50 +0000)]
lantiq: VG3503J - use the same PHY led functionality as the OEM firmware

Based on the vg3503j_gphy_led.sh script published in the VG3503J wiki
article, the OEM Firmware uses the following PHY led functionality:

    gphy led 0: LINK/ACTIVITY
    gphy led 1: LINK
    gphy led 2: ACTIVITY

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49278

8 years agoar71xx: Fix eth0 support for Ubiquiti UniFi AP AC
blogic [Mon, 2 May 2016 18:50:00 +0000 (18:50 +0000)]
ar71xx: Fix eth0 support for Ubiquiti UniFi AP AC

Fix eth0 support for the Ubiquiti UniFi AP AC
Signed-off-by: Paul Wassi <p.wassi at gmx.at>
SVN-Revision: 49277

8 years agolantiq: use the same functionality for all ethernet phys led
blogic [Fri, 29 Apr 2016 11:35:05 +0000 (11:35 +0000)]
lantiq: use the same functionality for all ethernet phys led

The VGV7510KW22 has the leds for LAN1-3 connected to pin1 of the phys
and the led for LAN4 connect to pin0 of the phy. This results with the
current configuration in a fast flashing LAN4 led as soon as a network
cable is connected. Something similar was reported on the forum[1] for
the VGV7519 as well.

Since it isn't predicable to which pin a (single) phy led is connected,
use the (default) pin1 functionality

    Constant On: 10/100/1000MBit
    Blink Fast: None
    Blink Slow: None
    Pulse: TX/RX

for all ethernet phy leds.

After checking pictures of all vr9 boards, it looks like only the VG3503J
has more than one led connected per phy. Using the phy led device tree
bindings to assign the functionality to the "additional" leds, the
VG3503J phy leds should behave as before.

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
[1] https://forum.openwrt.org/viewtopic.php?pid=321523

SVN-Revision: 49270

8 years agolantiq: fix minor typos in 11G/22F phy driver
blogic [Fri, 29 Apr 2016 11:35:01 +0000 (11:35 +0000)]
lantiq: fix minor typos in 11G/22F phy driver

Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49269

8 years agoramips: add support for Planex MZK-EX750NP.
blogic [Fri, 29 Apr 2016 11:34:56 +0000 (11:34 +0000)]
ramips: add support for Planex MZK-EX750NP.

add support for Planex MZK-EX750NP.
MZK-EX750NP is MT7620A and MT7610E based 11ac wifi repeater.
Built-in power supply.
64MiB RAM, 8MiB SPI Flash, non Wired Ethernet.

Signed-off-by: YuheiOKAWA <tochiro.srchack@gmail.com>
SVN-Revision: 49268

8 years agomediatek: sync patches and add more ethernet stability fixes
blogic [Fri, 29 Apr 2016 11:34:31 +0000 (11:34 +0000)]
mediatek: sync patches and add more ethernet stability fixes

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 49265

8 years agoncurses: install pkg-config files to fix util-linux build breakage
Felix Fietkau [Tue, 10 May 2016 08:05:22 +0000 (10:05 +0200)]
ncurses: install pkg-config files to fix util-linux build breakage

Signed-off-by: Felix Fietkau <nbd@nbd.name>
8 years agomtd: add -c option for specifying amount of data to be used for checksum
Rafał Miłecki [Fri, 6 May 2016 12:37:05 +0000 (14:37 +0200)]
mtd: add -c option for specifying amount of data to be used for checksum

So far fixtrx was calculating checksum over amount of data matching
partition erase size. It was mostly a workaround of checksum problem
after changing anything in initial TRX content (e.g. formatting JFFS2).
Its main purpose was to make bootloader accept modified TRX. This didn't
provide much protection of flash data against corruption.

This new option lets caller request calculating checksum over a bigger
amount of data. It may be used e.g. to include whole kernel data for
checksum and hopefully make bootloader go info failsafe mode if
something goes wrong.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: trx: use separated buffer for TRX header
Rafał Miłecki [Mon, 9 May 2016 17:00:53 +0000 (19:00 +0200)]
mtd: trx: use separated buffer for TRX header

We plan to adjust usage of the main buffer to allow reading custom
amount of data for CRC32. This means we need another buffer that will be
always block aligned.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: seama: exit with error if Seama header wasn't found
rmilecki [Mon, 9 May 2016 15:33:42 +0000 (15:33 +0000)]
mtd: seama: exit with error if Seama header wasn't found

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: seama: fix image data handling
rmilecki [Mon, 9 May 2016 12:53:39 +0000 (12:53 +0000)]
mtd: seama: fix image data handling

1) Put sanity checks in one place
2) Respect provided offset
3) Read only as much data as needed for MD5 calculation

Thanks to the last change this is a great speedup and memory saver. On
devices with NAND flash we were allocating & reading about 128 MiB while
something about 8 MiB is enough.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: seama: move buf allocation to the MD5 function
rmilecki [Mon, 9 May 2016 12:53:36 +0000 (12:53 +0000)]
mtd: seama: move buf allocation to the MD5 function

This buf is only used in this function now, so lets move it there.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: seama: update MD5 using header in the first block buffer
rmilecki [Mon, 9 May 2016 12:53:33 +0000 (12:53 +0000)]
mtd: seama: update MD5 using header in the first block buffer

This will allow separating first block buffer from a buffer used for MD5
calculation.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: seama: add md5 to header struct
rmilecki [Mon, 9 May 2016 12:53:29 +0000 (12:53 +0000)]
mtd: seama: add md5 to header struct

This allows us to drop some extra offset calculations and simplifies
code a bit.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: check for Seama magic early when fixing MD5
rmilecki [Mon, 9 May 2016 09:54:48 +0000 (09:54 +0000)]
mtd: check for Seama magic early when fixing MD5

This avoid long (and unneeded) process of reading all data in case of
running on MTD not containig Seama entity.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: add missing breaks in a switch
rmilecki [Mon, 9 May 2016 07:51:50 +0000 (07:51 +0000)]
mtd: add missing breaks in a switch

On platforms supporting both: TRX and Seama calling "fixtrx" was
resulting in trying to fix Seama as well.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agomtd: use tabs for indents
rmilecki [Mon, 9 May 2016 07:51:45 +0000 (07:51 +0000)]
mtd: use tabs for indents

This makes code style consistent across the whole file.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
8 years agoncurses: remove libncurses, provide it via libncursesw
Felix Fietkau [Mon, 9 May 2016 12:49:33 +0000 (14:49 +0200)]
ncurses: remove libncurses, provide it via libncursesw

Signed-off-by: Felix Fietkau <nbd@nbd.name>
8 years agometadata.pl: add support for selecting packages available only via PROVIDES
Felix Fietkau [Mon, 9 May 2016 12:48:49 +0000 (14:48 +0200)]
metadata.pl: add support for selecting packages available only via PROVIDES

Signed-off-by: Felix Fietkau <nbd@nbd.name>
8 years agokernel: add workaround to rebuild vdso-o32.so.dbg
Stijn Tintel [Mon, 9 May 2016 00:30:38 +0000 (02:30 +0200)]
kernel: add workaround to rebuild vdso-o32.so.dbg

Building for octeon fails with

  'arch/mips/vdso/vdso-o32.so.dbg' already contains a '.MIPS.abiflags'
  section

if the file already exists from a prior build.

Use the same workaround as the one for vdso.so.dbg committed in
9eb155353a5f5137ec85a5b5b0287af63c544710.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
8 years agotoolchain: gcc: fix build with GCC 6
Matthias Schiffer [Sun, 8 May 2016 21:10:13 +0000 (23:10 +0200)]
toolchain: gcc: fix build with GCC 6

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
8 years agotools: mkimage: sync include/linux/compiler*.h with u-boot master
Matthias Schiffer [Sun, 8 May 2016 20:06:51 +0000 (22:06 +0200)]
tools: mkimage: sync include/linux/compiler*.h with u-boot master

Fixes build with GCC 6.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
8 years agotools: pkg-config: fix build with GCC 6
Matthias Schiffer [Sun, 8 May 2016 13:53:14 +0000 (15:53 +0200)]
tools: pkg-config: fix build with GCC 6

Fixes the following error:

gdate.c: In function ‘g_date_strftime’:
gdate.c:2497:7: error: format not a string literal, format string not checked [-Werror=format-nonliteral]
       tmplen = strftime (tmpbuf, tmpbufsize, locale_format, &tm);
       ^~~~~~

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
8 years agomt76: update to the latest version
Felix Fietkau [Thu, 5 May 2016 18:42:43 +0000 (19:42 +0100)]
mt76: update to the latest version

Fixes a tx locking error and adds a pci id

Signed-off-by: Felix Fietkau <nbd@nbd.name>
8 years agomt76: fix rebuild on CONFIG_PACKAGE_MAC80211_MESH changes
Felix Fietkau [Thu, 5 May 2016 18:39:21 +0000 (19:39 +0100)]
mt76: fix rebuild on CONFIG_PACKAGE_MAC80211_MESH changes

Signed-off-by: Felix Fietkau <nbd@nbd.name>
8 years agobuild: add support for specifying extra package dependencies for prepared, configured...
Felix Fietkau [Thu, 5 May 2016 18:37:40 +0000 (19:37 +0100)]
build: add support for specifying extra package dependencies for prepared, configured and built

Signed-off-by: Felix Fietkau <nbd@nbd.name>
8 years agotoolchain: rename OpenWrt into LEDE
Alexander Couzens [Thu, 5 May 2016 17:33:03 +0000 (19:33 +0200)]
toolchain: rename OpenWrt into LEDE

Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
8 years agosdk: rename OpenWrt into LEDE
Alexander Couzens [Thu, 5 May 2016 17:33:02 +0000 (19:33 +0200)]
sdk: rename OpenWrt into LEDE

Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
8 years agoimagebuilder: rename OpenWrt into LEDE
Alexander Couzens [Thu, 5 May 2016 17:33:01 +0000 (19:33 +0200)]
imagebuilder: rename OpenWrt into LEDE

Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
8 years agotools: make-ext4fs: fix build regression on mac os x
Felix Fietkau [Thu, 5 May 2016 17:47:55 +0000 (18:47 +0100)]
tools: make-ext4fs: fix build regression on mac os x

Including sys/sysmacros.h is not portable

Signed-off-by: Felix Fietkau <nbd@nbd.name>
8 years agobase-files: evaluate uci-defaults on run-time installations
Daniel Golle [Fri, 29 Apr 2016 11:34:44 +0000 (11:34 +0000)]
base-files: evaluate uci-defaults on run-time installations

Packages may install scripts into /etc/uci-defaults to be executed once
after installation, usually at the first boot of the target. This works
fine if the package was installed to the rootfs during build or using
the ImageBuilder.
If the package is installed using opkg during run-time uci-defaults
were applied only after a reboot of the device. Avoid the need to
reboot by evaluting the package's uci-defaults in default-postinst.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
8 years agoopenssl: Update to version 1.0.2h
Michal Hrusecky [Wed, 4 May 2016 11:53:27 +0000 (13:53 +0200)]
openssl: Update to version 1.0.2h

Bump to the latest version, fixes several security issues:
 * CVE-2016-2107, CVE-2016-2105, CVE-2016-2106, CVE-2016-2109, CVE-2016-2176
More details at https://www.openssl.org/news/openssl-1.0.2-notes.html

Signed-off-by: Michal Hrusecky <Michal.Hrusecky@nic.cz>
8 years agofirewall3: fix mark rules for local traffic, fix race condition
Jo-Philipp Wich [Mon, 2 May 2016 17:46:23 +0000 (18:46 +0100)]
firewall3: fix mark rules for local traffic, fix race condition

Update to latest HEAD in order to fix MARK rule generation for local traffic,
also fix a possible race condition during firewall start.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agolibiconv: add all ASCII aliases
Gergely Kiss [Mon, 2 May 2016 00:46:33 +0000 (02:46 +0200)]
libiconv: add all ASCII aliases

This patch adds missing ASCII aliases to the libiconv stub in order to avoid conversion errors like https://github.com/openwrt/packages/issues/2373

Signed-off-by: Gergely Kiss <mail.gery@gmail.com>
8 years agoxtables-addons: Avoid redefinition of SHRT_MAX in lua packet script
Hans Dedecker [Mon, 2 May 2016 09:57:59 +0000 (11:57 +0200)]
xtables-addons: Avoid redefinition of SHRT_MAX in lua packet script

Patch Lua packet script defines SHRT_MAX which is already defined in <linux/kernel.h> and
is included indirectly by lauxlib.h. Fix the redefintion as it leads to compile failure
on systems which treat macro redefinition as an error

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
8 years agosigning: remove unatteded build key and use current keyring instead
Jo-Philipp Wich [Sat, 30 Apr 2016 18:39:23 +0000 (20:39 +0200)]
signing: remove unatteded build key and use current keyring instead

Remove the public unatteded buildkey from the opkg package to avoid
having hardcoded keys in tree. Use the external keyring package instead
which can be easily updated by users.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agofstools: fix snapshot support
John Crispin [Fri, 29 Apr 2016 01:41:48 +0000 (03:41 +0200)]
fstools: fix snapshot support

Signed-off-by: John Crispin <john@phrozen.org>
8 years agofstools: update to latest git HEAD
John Crispin [Fri, 29 Apr 2016 01:49:45 +0000 (03:49 +0200)]
fstools: update to latest git HEAD

fixes snapshot support

Signed-off-by: John Crispin <john@phrozen.org>
8 years agobase-files: fix group/user settings after sysupgrade
John Crispin [Fri, 29 Apr 2016 01:30:52 +0000 (03:30 +0200)]
base-files: fix group/user settings after sysupgrade

Signed-off-by: John Crispin <john@phrozen.org>
8 years agobase-files: split user/group addition code into a function
John Crispin [Fri, 29 Apr 2016 01:13:40 +0000 (03:13 +0200)]
base-files: split user/group addition code into a function

Signed-off-by: John Crispin <john@phrozen.org>
8 years agoscripts: add "r" to revision
John Crispin [Fri, 29 Apr 2016 01:04:16 +0000 (03:04 +0200)]
scripts: add "r" to revision

343c3be45451e4a11f03109bd6cb1dd9061ec5eb accidentially removed the "r"

Signed-off-by: John Crispin <john@phrozen.org>
8 years agofstools: update to latest git HEAD
John Crispin [Wed, 27 Apr 2016 01:37:00 +0000 (03:37 +0200)]
fstools: update to latest git HEAD

this adds the remount command to the block tool

Signed-off-by: John Crispin <john@phrozen.org>
8 years agobase-files: add new public key used by unattended builds
John Crispin [Wed, 27 Apr 2016 00:20:35 +0000 (02:20 +0200)]
base-files: add new public key used by unattended builds

Signed-off-by: John Crispin <john@phrozen.org>
8 years agobase-files: add a conditional dependency to lede-keyring
John Crispin [Tue, 26 Apr 2016 23:52:02 +0000 (01:52 +0200)]
base-files: add a conditional dependency to lede-keyring

Signed-off-by: John Crispin <john@phrozen.org>
8 years agolede-keyring: add the developer public keyring
John Crispin [Tue, 26 Apr 2016 23:50:04 +0000 (01:50 +0200)]
lede-keyring: add the developer public keyring

Signed-off-by: John Crispin <john@phrozen.org>
8 years agoinclude: add lede git server url
John Crispin [Tue, 26 Apr 2016 23:33:14 +0000 (01:33 +0200)]
include: add lede git server url

Signed-off-by: John Crispin <john@phrozen.org>
8 years agokernel: add a workaround to rebuild vdso.so.dbg after genvdso
Jo-Philipp Wich [Fri, 29 Apr 2016 07:53:09 +0000 (09:53 +0200)]
kernel: add a workaround to rebuild vdso.so.dbg after genvdso

Currently the build fails with

  'arch/mips/vdso/vdso.so.dbg' already contains a '.MIPS.abiflags' section

if the file already exists from a prior build.

Add a makefile rule to force the rebuild of vdso.so.dbg if genvdso has
has been changed to workaround the failure.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agoscripts: avoid hard-coded paths in scripts
Iblis Lin [Thu, 28 Apr 2016 12:34:13 +0000 (20:34 +0800)]
scripts: avoid hard-coded paths in scripts

Signed-off-by: Iblis Lin <e196819@hotmail.com>
8 years agoppp: Add ppp-mod-passwordfd subpackage to ppp
Hans Dedecker [Wed, 27 Apr 2016 09:14:02 +0000 (11:14 +0200)]
ppp: Add ppp-mod-passwordfd subpackage to ppp

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
8 years agodnsmasq: Add conntrack support in the full variant
Hans Dedecker [Wed, 27 Apr 2016 09:14:01 +0000 (11:14 +0200)]
dnsmasq: Add conntrack support in the full variant

Conntrack support reads the connection track mark associated with
incoming DNS queries and sets the same mark value on the upstream
forwarded DNS query. This can be usefull to track traffic generated
by dnsmasq to associate it with the clients who generate the queries,
usefull for bandwidth accouting and firewall.

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
8 years agodropbear: Add procd interface triggers when interface config is specified
Hans Dedecker [Wed, 27 Apr 2016 09:14:00 +0000 (11:14 +0200)]
dropbear: Add procd interface triggers when interface config is specified

A dropbear instance having an interface config won't start if the interface is down as no
IP address is available.
Adding interface triggers for each configured interface executing the dropbear reload script
will start the dropbear instance when the interface is up.

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
8 years agoiproute2: Add package for nstat utility
Hans Dedecker [Wed, 27 Apr 2016 09:13:59 +0000 (11:13 +0200)]
iproute2: Add package for nstat utility

Add support for the command line utility nstat displaying network statistics

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
8 years agodropbear: Make utmp and putuline support configurable via seperate config options
Hans Dedecker [Wed, 27 Apr 2016 09:13:58 +0000 (11:13 +0200)]
dropbear: Make utmp and putuline support configurable via seperate config options

Utmp support tracks who is currenlty logged in by logging info to the file /var/run/utmp (supported by busybox)
Putuline support will use the utmp structure to write to the utmp file

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
8 years agonetifd: Add configurable DHCP release behavior
Hans Dedecker [Thu, 28 Apr 2016 11:40:17 +0000 (13:40 +0200)]
netifd: Add configurable DHCP release behavior

Make sending a DHCP release configurable when the client exits allowing to clean up
IP/mac state info in intermediate devices.

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
8 years agobcm53xx: add m25p80 workaround for SPI flash writing problems
Rafał Miłecki [Wed, 27 Apr 2016 21:20:45 +0000 (21:20 +0000)]
bcm53xx: add m25p80 workaround for SPI flash writing problems

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 49264

8 years agomediatek: sync and patches add support for several boards
John Crispin [Wed, 27 Apr 2016 08:58:15 +0000 (08:58 +0000)]
mediatek: sync and patches add support for several boards

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 49263

8 years agobcm53xx: support SPI-NOR on dual flash devices
Rafał Miłecki [Wed, 27 Apr 2016 08:04:30 +0000 (08:04 +0000)]
bcm53xx: support SPI-NOR on dual flash devices

We need it to access NVRAM.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 49262

8 years agobcm53xx: switch to m25p80 and drop bcm53xxspiflash
Rafał Miłecki [Tue, 26 Apr 2016 17:08:53 +0000 (17:08 +0000)]
bcm53xx: switch to m25p80 and drop bcm53xxspiflash

For some time now m25p80 supports 32 MiB flashes and we just needed to
patch spi-nor to support JEDEC incompatible w25q128. Also by switching
to m25p80 we gain accelerated SPI flash reads.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 49260

8 years agobcm53xx: support JEDEC incompatible w25q128 in spi-nor
Rafał Miłecki [Tue, 26 Apr 2016 12:36:34 +0000 (12:36 +0000)]
bcm53xx: support JEDEC incompatible w25q128 in spi-nor

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 49259

8 years agoxtables-addons: Fix Lua packet script implementation
Hans Dedecker [Thu, 28 Apr 2016 09:18:56 +0000 (11:18 +0200)]
xtables-addons: Fix Lua packet script implementation

lua_packet_segment parameter start has type char pointer; in function lua_tg
it's assigned an uint16 value generating compiler warnings obviously indicating
posssible seg fault problems. Fix the issue by using the correct skb functions
so the parameter points to the position inside the sk_buff

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
Signed-off-by: Stijn Cleynhens <stijn.cleynhens@gmail.com>
8 years agolantiq: fix xway image building
Jo-Philipp Wich [Tue, 26 Apr 2016 23:13:34 +0000 (01:13 +0200)]
lantiq: fix xway image building

Commit 8c130dd (lantiq: use new image build process for some images) broke
image generation for the xway target since the generated macros reference not
existing WBMRA.dts and WBMRB.dts files leading to the following build error:

  mips-openwrt-linux-musl-cpp: error: ../dts/WBMRA.dts: No such file or directory
  mips-openwrt-linux-musl-cpp: warning: '-x assembler-with-cpp' after last input file has no effect
  mips-openwrt-linux-musl-cpp: fatal error: no input files

Previously both the WBMRA and WBMRB profiles shared a common WBMR DT since
the only difference is the embedded firmware kmod package.

Extend the lantiqImage macro to optionally support specifying a different
DTS file and use this facility to let the both WBMR profiles reference the
correct device tree file.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agopackage: remove duplicate lines from otrx and nvram makefiles
Jo-Philipp Wich [Tue, 26 Apr 2016 21:28:17 +0000 (23:28 +0200)]
package: remove duplicate lines from otrx and nvram makefiles

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agopackage: flag further target specific packages as nonshared
Jo-Philipp Wich [Tue, 26 Apr 2016 21:09:12 +0000 (23:09 +0200)]
package: flag further target specific packages as nonshared

Add nonshared flag to package depending on specific targets or subtargets as
there's no guarantee otherwise that they'll be available in the shared repo.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agopackage: mark nvram and otrx nonshared as they're target specific
Jo-Philipp Wich [Tue, 26 Apr 2016 21:09:12 +0000 (23:09 +0200)]
package: mark nvram and otrx nonshared as they're target specific

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agobrcm2708: removes backported patch (linux-4.4.7)
Álvaro Fernández Rojas [Tue, 26 Apr 2016 14:41:41 +0000 (16:41 +0200)]
brcm2708: removes backported patch (linux-4.4.7)

This regression was introduced while rebasing LEDE's commits over the latest
OpenWrt repository.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
8 years agoat91: fix SAMA5D3 subtarget
Jo-Philipp Wich [Mon, 25 Apr 2016 12:54:07 +0000 (14:54 +0200)]
at91: fix SAMA5D3 subtarget

Add required architecture overrides to enable SAMA5 boards in Kconfig, add
resulting missing symbols, reorder config and remove unrelated XEN setting.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agox86: remove old references to kmod-acpi-button
Jo-Philipp Wich [Mon, 25 Apr 2016 12:14:47 +0000 (14:14 +0200)]
x86: remove old references to kmod-acpi-button

The kmod-acpi-button functionality is included in the kernel where needed
and the corresponding kmod package has been deleted long ago, therfor drop
the remaining references now.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agouml: revert accidentally committed change
Jo-Philipp Wich [Mon, 25 Apr 2016 12:09:05 +0000 (14:09 +0200)]
uml: revert accidentally committed change

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agotools: fix make_ext4fs build with recent glibc
Jo-Philipp Wich [Mon, 25 Apr 2016 11:41:43 +0000 (13:41 +0200)]
tools: fix make_ext4fs build with recent glibc

Update to latest make_ext4fs Git HEAD to fix build with recent glibc version
which does not implicitely includes sysmacros.h anymore.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agobrcm2708: update linux 4.4 patches to latest version
Álvaro Fernández Rojas [Sun, 24 Apr 2016 11:03:39 +0000 (13:03 +0200)]
brcm2708: update linux 4.4 patches to latest version

As usual these patches were extracted from the raspberry pi repo:
https://github.com/raspberrypi/linux/tree/rpi-4.4.y

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
8 years agobrcm2708-gpu-fw: update to latest version
Álvaro Fernández Rojas [Sun, 24 Apr 2016 11:00:11 +0000 (13:00 +0200)]
brcm2708-gpu-fw: update to latest version

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
8 years agokernel: fix ip6_tunnel compilation
Álvaro Fernández Rojas [Sun, 24 Apr 2016 10:56:07 +0000 (12:56 +0200)]
kernel: fix ip6_tunnel compilation

Replace undefined iph for ip_hdr(skb)

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
8 years agokernel: update kernel 4.4 to version 4.4.8
Álvaro Fernández Rojas [Sun, 24 Apr 2016 09:32:17 +0000 (11:32 +0200)]
kernel: update kernel 4.4 to version 4.4.8

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
8 years agoopkg: fix use-after-free with duplicate packages on the command line
Jo-Philipp Wich [Sat, 23 Apr 2016 12:03:50 +0000 (14:03 +0200)]
opkg: fix use-after-free with duplicate packages on the command line

When the same package file is specified multiple times on the opkg install
command line, the name pointer on the argv array becomes stale after the
package structures have been merged, leading to invalid memory accesses
upon install.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agoimagebuilder: use correct package directory when bundling kmods and libc
Jo-Philipp Wich [Fri, 22 Apr 2016 10:22:57 +0000 (12:22 +0200)]
imagebuilder: use correct package directory when bundling kmods and libc

The libc and kernel package files moved since the introduction of shared
packages and the changed output directory layout. This causes the generated
ImageBuilder archive to lack the necessary "libc" and "kernel" meta packages,
leading to opkg install errors later on.

Use the FeedPackageDir macro to figure out the proper source directory to use.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agopackage: fix toolchain ipk flags
Jo-Philipp Wich [Fri, 22 Apr 2016 10:20:47 +0000 (12:20 +0200)]
package: fix toolchain ipk flags

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agoinclude: remove unused FeedPackageDir macro arguments
Jo-Philipp Wich [Fri, 22 Apr 2016 10:02:53 +0000 (12:02 +0200)]
include: remove unused FeedPackageDir macro arguments

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agolinux: kmod-e100: use preconverted firmware files
Jo-Philipp Wich [Thu, 21 Apr 2016 14:39:59 +0000 (16:39 +0200)]
linux: kmod-e100: use preconverted firmware files

Instead of converting the firmware files ourselves, use the files
generated during the normal kernel build process. This fixes packaging
kmod-e100 in the SDK environment.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agosdk: bundle required firmware files
Jo-Philipp Wich [Thu, 21 Apr 2016 14:34:39 +0000 (16:34 +0200)]
sdk: bundle required firmware files

Bundle all firmware files which have been built along with the kernel to fix
packaging of kmods in the SDK environment that rely on in-tree firmwares,
specifically kmod-e100.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
8 years agoinclude: bin/ dir was not created
John Crispin [Tue, 19 Apr 2016 14:08:18 +0000 (16:08 +0200)]
include: bin/ dir was not created

explicitly create the bin dir

Signed-off-by: John Crispin <john@phrozen.org>
8 years agobase-files: remove ununsed login.sh
John Crispin [Tue, 19 Apr 2016 18:42:34 +0000 (20:42 +0200)]
base-files: remove ununsed login.sh

Signed-off-by: John Crispin <john@phrozen.org>
8 years agobase-files: Allow to disable failsafe mode
John Crispin [Tue, 19 Apr 2016 08:07:35 +0000 (10:07 +0200)]
base-files: Allow to disable failsafe mode

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