openwrt/openwrt.git
7 years agoimage: add a helper variable for getting kernel/rootfs from within image Build/*...
Felix Fietkau [Sat, 30 Jul 2016 12:19:11 +0000 (14:19 +0200)]
image: add a helper variable for getting kernel/rootfs from within image Build/* templates

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agokernel: remove hostap driver
Felix Fietkau [Sat, 30 Jul 2016 10:53:02 +0000 (12:53 +0200)]
kernel: remove hostap driver

It has been marked as broken for well over a month now and nobody has
complained.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoocteon: increase block2mtd rootfs probe timeout on ER
Felix Fietkau [Sat, 30 Jul 2016 12:07:47 +0000 (14:07 +0200)]
octeon: increase block2mtd rootfs probe timeout on ER

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoimage.mk: add CMDLINE to DEVICE_VARS
Felix Fietkau [Sat, 30 Jul 2016 12:01:47 +0000 (14:01 +0200)]
image.mk: add CMDLINE to DEVICE_VARS

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoocteon: pad squashfs sysupgrade image rootfs
Felix Fietkau [Sat, 30 Jul 2016 11:48:00 +0000 (13:48 +0200)]
octeon: pad squashfs sysupgrade image rootfs

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoimage: add support for overriding kernel/rootfs images in sysupgrade-tar template
Felix Fietkau [Sat, 30 Jul 2016 11:47:36 +0000 (13:47 +0200)]
image: add support for overriding kernel/rootfs images in sysupgrade-tar template

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoocteon: drop unsupported jffs2 feature flag
Felix Fietkau [Sat, 30 Jul 2016 11:45:07 +0000 (13:45 +0200)]
octeon: drop unsupported jffs2 feature flag

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoocteon: fix sysupgrade images
Felix Fietkau [Sat, 30 Jul 2016 10:36:33 +0000 (12:36 +0200)]
octeon: fix sysupgrade images

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agokernel: fix crashlog regression on x86
Felix Fietkau [Sat, 30 Jul 2016 10:18:08 +0000 (12:18 +0200)]
kernel: fix crashlog regression on x86

Check memblock regions for sufficient size before attempting to use
them. Allow checks for multiple memblock regions until a suitable one is
found.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoinclude/cmake.mk: fix host builds
Felix Fietkau [Fri, 29 Jul 2016 20:47:46 +0000 (22:47 +0200)]
include/cmake.mk: fix host builds

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agodnsmasq: drop --interface and --except-interface options when the interface cannot...
Felix Fietkau [Fri, 29 Jul 2016 18:58:14 +0000 (20:58 +0200)]
dnsmasq: drop --interface and --except-interface options when the interface cannot be found

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agocmake: include/cmake.mk add CMAKE_BINARY_SUBDIR to allow out of source tree builds
Luke McKee [Sat, 23 Jul 2016 19:00:24 +0000 (02:00 +0700)]
cmake: include/cmake.mk add CMAKE_BINARY_SUBDIR to allow out of source tree builds

Some packages need out of source tree building with cmake, for example
when building kernel modules.
See an example here:
https://sourceforge.net/p/accel-ppp/code/ci/master/tree/README

Signed-off-by: Luke McKee <hojuruku@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name> [cleanup, rework]
7 years agonetifd: update to the latest version, adds an event handling fix
Felix Fietkau [Fri, 29 Jul 2016 17:51:31 +0000 (19:51 +0200)]
netifd: update to the latest version, adds an event handling fix

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agodnsmasq: remove use of uci state for getting network ifname
Felix Fietkau [Fri, 29 Jul 2016 17:48:12 +0000 (19:48 +0200)]
dnsmasq: remove use of uci state for getting network ifname

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoimage.mk: remove leftover variable from a previous rework
Felix Fietkau [Fri, 29 Jul 2016 15:13:48 +0000 (17:13 +0200)]
image.mk: remove leftover variable from a previous rework

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoath25: rework image building
Jo-Philipp Wich [Fri, 29 Jul 2016 14:47:45 +0000 (16:47 +0200)]
ath25: rework image building

* Remove legacy macros
* Emit generic LZMA and Gzip kernel images from the combined device recipe
* Rename combined device to generic

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agobuild: rename sysupgrade-nand to sysupgrade-tar
Jo-Philipp Wich [Fri, 29 Jul 2016 13:05:51 +0000 (15:05 +0200)]
build: rename sysupgrade-nand to sysupgrade-tar

Now that the "sysupgrade-nand" step is used by non-NAND targets as well,
rename it to "sysupgrade-tar" to make it more generic.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agoocteon: clean up image build code
Jo-Philipp Wich [Fri, 29 Jul 2016 13:03:19 +0000 (15:03 +0200)]
octeon: clean up image build code

Use the generic "sysupgrade-nand" build step to pack the sysupgrade file,
it performs the same steps as the inline "tar-file" macro.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agodnsmasq: replace the iface hotplug script with a procd trigger
Felix Fietkau [Fri, 29 Jul 2016 12:56:35 +0000 (14:56 +0200)]
dnsmasq: replace the iface hotplug script with a procd trigger

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agodnsmasq: make the check for existing DHCP servers more reliable
Felix Fietkau [Fri, 29 Jul 2016 12:26:55 +0000 (14:26 +0200)]
dnsmasq: make the check for existing DHCP servers more reliable

If there is no carrier yet, wait for 2 seconds (STP forwarding delay)

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agodnsmasq: write atomic config file
Ulrich Weber [Mon, 13 Jun 2016 13:21:47 +0000 (15:21 +0200)]
dnsmasq: write atomic config file

multiple invocation of dnsmasq script (e.g. by procd and hotplugd)
might cause procd to restart dnsmasq with an incomplete config file.
Config file generation might take quite a long time on larger configs
due ubus calls for each listening interface...

Signed-off-by: Ulrich Weber <ulrich.weber@riverbed.com>
7 years agonetifd: update to the latest version
Felix Fietkau [Fri, 29 Jul 2016 14:17:40 +0000 (16:17 +0200)]
netifd: update to the latest version

Emits an initial event after the first link-up of a force_link
interface. This is needed for making the dnsmasq dhcp check more
reliable

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoprocd: add default timeout for reload trigger actions
Felix Fietkau [Fri, 29 Jul 2016 12:19:17 +0000 (14:19 +0200)]
procd: add default timeout for reload trigger actions

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoigmpproxy: remove procd_open_trigger/procd_close_trigger calls
Felix Fietkau [Fri, 29 Jul 2016 11:07:08 +0000 (13:07 +0200)]
igmpproxy: remove procd_open_trigger/procd_close_trigger calls

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agodropbear: remove procd_open_trigger/procd_close_trigger calls
Felix Fietkau [Fri, 29 Jul 2016 11:04:13 +0000 (13:04 +0200)]
dropbear: remove procd_open_trigger/procd_close_trigger calls

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agosysntpd: remove procd_open_trigger/procd_close_trigger calls
Felix Fietkau [Fri, 29 Jul 2016 11:03:57 +0000 (13:03 +0200)]
sysntpd: remove procd_open_trigger/procd_close_trigger calls

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoprocd: rework trigger handling
Felix Fietkau [Fri, 29 Jul 2016 11:00:56 +0000 (13:00 +0200)]
procd: rework trigger handling

Open/close triggers array around service_triggers call to make using
multiple triggers easier to deal with.
The API was quite confusing, because some functions contained implicit
trigger open/close calls and some didn't.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoprocd: update to the latest version
Felix Fietkau [Fri, 29 Jul 2016 12:07:09 +0000 (14:07 +0200)]
procd: update to the latest version

Fixes a long standing design issue in handling of delayed triggers

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agolibubox: update to the latest version, adds a few utility functions
Felix Fietkau [Fri, 29 Jul 2016 12:06:08 +0000 (14:06 +0200)]
libubox: update to the latest version, adds a few utility functions

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: add option to enable all profiles
Felix Fietkau [Fri, 22 Jul 2016 12:29:32 +0000 (14:29 +0200)]
build: add option to enable all profiles

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: add support for per-device rootfs based on device profile packges
Felix Fietkau [Fri, 22 Jul 2016 12:27:52 +0000 (14:27 +0200)]
build: add support for per-device rootfs based on device profile packges

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: minor cleanup of redundant code
Felix Fietkau [Fri, 22 Jul 2016 12:04:16 +0000 (14:04 +0200)]
build: minor cleanup of redundant code

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: set TMPDIR for opkg calls
Felix Fietkau [Fri, 22 Jul 2016 11:04:11 +0000 (13:04 +0200)]
build: set TMPDIR for opkg calls

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: add target_params variable for getting root filesystem image parameters
Felix Fietkau [Fri, 22 Jul 2016 10:43:54 +0000 (12:43 +0200)]
build: add target_params variable for getting root filesystem image parameters

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: add template for getting opkg package files from package names
Felix Fietkau [Fri, 22 Jul 2016 10:22:16 +0000 (12:22 +0200)]
build: add template for getting opkg package files from package names

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: rework opkg command invocation
Felix Fietkau [Fri, 22 Jul 2016 10:10:17 +0000 (12:10 +0200)]
build: rework opkg command invocation

Drop included $(XARGS), add support for passing target dir via parameter

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: remove obsolete variables from opkg command
Felix Fietkau [Fri, 22 Jul 2016 10:04:33 +0000 (12:04 +0200)]
build: remove obsolete variables from opkg command

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: rework prepare_rootfs to pass target dir via parameter
Felix Fietkau [Fri, 22 Jul 2016 09:54:58 +0000 (11:54 +0200)]
build: rework prepare_rootfs to pass target dir via parameter

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoimage: add wrapper variable to get the target dir for mkfs commands
Felix Fietkau [Fri, 22 Jul 2016 08:53:12 +0000 (10:53 +0200)]
image: add wrapper variable to get the target dir for mkfs commands

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoimage: make mkfs template output to $@
Felix Fietkau [Thu, 21 Jul 2016 23:17:35 +0000 (01:17 +0200)]
image: make mkfs template output to $@

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: move rootfs processing code to include/rootfs.mk so it can be reused later
Felix Fietkau [Thu, 21 Jul 2016 16:49:54 +0000 (18:49 +0200)]
build: move rootfs processing code to include/rootfs.mk so it can be reused later

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoapm821xx: only attempt to mount /boot on MyBook Live
Felix Fietkau [Thu, 28 Jul 2016 17:30:03 +0000 (19:30 +0200)]
apm821xx: only attempt to mount /boot on MyBook Live

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agolantiq: fix switch configuration for EASY80920
Hauke Mehrtens [Wed, 27 Jul 2016 19:27:54 +0000 (21:27 +0200)]
lantiq: fix switch configuration for EASY80920

The device tree description misses some Ethernet ports and there was no
model specified for this board. In addition there was no switch
specific default configuration created.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agokernel: fix crashlog issues on various architectures
Felix Fietkau [Thu, 28 Jul 2016 09:59:48 +0000 (11:59 +0200)]
kernel: fix crashlog issues on various architectures

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agolantiq: fix WBMR-300HPD switch port assignment
Felix Fietkau [Thu, 28 Jul 2016 18:52:55 +0000 (20:52 +0200)]
lantiq: fix WBMR-300HPD switch port assignment

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoocteon: fix image build
Jo-Philipp Wich [Thu, 28 Jul 2016 17:21:44 +0000 (19:21 +0200)]
octeon: fix image build

After b47f438 "build: remove image prefix from kernel files in KDIR", the
kernel image in $(KDIR) has a different name and could not be found by the
tar-file build step anymore, leading to the following error on the build
servers:

  cp: cannot stat `.../linux-octeon/lede-octeon-generic-kernel.bin': No such file or directory
  make[4]: *** [.../linux-octeon/tmp/lede-octeon-generic-ext4-sysupgrade.tar] Error 1

Adjust the path to the source kernel image in order to fix the problem.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agoar71xx: don't use D-Link DIR-505 status LED as ethernet indicator
Matthias Schiffer [Thu, 28 Jul 2016 13:59:04 +0000 (15:59 +0200)]
ar71xx: don't use D-Link DIR-505 status LED as ethernet indicator

The stock firmware uses the single LED as status indicator only. Using the
same LED both for status and as ethernet indicator is uncommon, and has
been confusing users who were using the device as a WLAN mesh node (so the
LED was just off, as no ethernet was connected).

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agoar71xx: add revision detection for D-Link DIR-505 A1/A2
Matthias Schiffer [Thu, 28 Jul 2016 13:43:25 +0000 (15:43 +0200)]
ar71xx: add revision detection for D-Link DIR-505 A1/A2

The A1 image also works on the A2, correctly set AR71XX_MODEL in this case.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agobrcmfmac43430-firmware: update to v7.45.41.26
Álvaro Fernández Rojas [Wed, 27 Jul 2016 10:08:50 +0000 (12:08 +0200)]
brcmfmac43430-firmware: update to v7.45.41.26

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
7 years agobcm53xx: enable kernel symbols/drivers needed for BCM53573
Rafał Miłecki [Thu, 28 Jul 2016 12:04:58 +0000 (14:04 +0200)]
bcm53xx: enable kernel symbols/drivers needed for BCM53573

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
7 years agobcm53xx: backport DTS patches for USB 2.0 and Tenda AC9
Rafał Miłecki [Thu, 28 Jul 2016 10:43:42 +0000 (12:43 +0200)]
bcm53xx: backport DTS patches for USB 2.0 and Tenda AC9

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
7 years agobcm53xx: refresh kernel patches
Rafał Miłecki [Thu, 28 Jul 2016 10:44:17 +0000 (12:44 +0200)]
bcm53xx: refresh kernel patches

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
7 years agoar71xx: add a missing ;; to ar71xx.sh
John Crispin [Sun, 24 Jul 2016 09:17:34 +0000 (11:17 +0200)]
ar71xx: add a missing ;; to ar71xx.sh

commit 42305ae24aa7 ("ar71xx: add support for gl-mifi")

was missing a ;; breaking boot up

Signed-off-by: John Crispin <john@phrozen.org>
7 years agonetifd: Use -x hostname:$hostname instead of -H
Merlijn Wajer [Wed, 20 Jul 2016 15:23:52 +0000 (17:23 +0200)]
netifd: Use -x hostname:$hostname instead of -H

Passing the hostname is currently broken in since the shipped busybox includes this commit:
https://git.busybox.net/busybox/commit/networking/udhcp/dhcpc.c?id=2017d48c0d70bef8768efb42909e605ea8eb5a21

Before:

    Sun Jan 31 18:11:32 2016 daemon.notice netifd: Interface 'wan' is now down
    Sun Jan 31 18:11:32 2016 daemon.notice netifd: Interface 'wan' is setting up now
    Sun Jan 31 18:11:32 2016 daemon.notice netifd: wan (18158): udhcpc: option -h NAME is deprecated, use -x hostname:NAME
    Sun Jan 31 18:11:32 2016 daemon.notice netifd: wan (18158): udhcpc: malformed hex string 'WR150'

After:

    Sun Jan 31 18:11:33 2016 daemon.notice netifd: wan (18169): udhcpc (v1.23.2) started
    Sun Jan 31 18:11:33 2016 daemon.notice netifd: wan (18169): Sending discover...
    Sun Jan 31 18:11:33 2016 daemon.notice netifd: wan (18169): Sending select for xxx.yyy.zzz.xyz...
    Sun Jan 31 18:11:33 2016 daemon.notice netifd: wan (18169): Lease of xxx.yyy.zzz.xyz obtained, lease time 600

Signed-off-by: Merlijn Wajer <merlijn@wizzup.org>
7 years agobase-files: sysupgrade: fix pseudobridge upgrades
Conn O'Griofa [Wed, 27 Jul 2016 18:04:06 +0000 (19:04 +0100)]
base-files: sysupgrade: fix pseudobridge upgrades

Treat 'relayd' as an essential service to avoid connection interruptions during sysupgrade on devices configured as a pseudobridge.

Signed-off-by: Conn O'Griofa <connogriofa@gmail.com>
7 years agob53: allow ports with higher numbers than CPU port
Rafał Miłecki [Wed, 27 Jul 2016 12:15:32 +0000 (14:15 +0200)]
b53: allow ports with higher numbers than CPU port

Our code was assuming CPU port uses the highest number. My BCM53573
device has eth0 connected to port 8 and eth1 connected to port 5. While
working on support for it I tried to:
1) Enable all ports (including port 8)
2) Set CPU port to 5

I noticed port 8 is not accessible anymore. It was just a development
process but it seems like something worth fixing anyway.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Acked-by: Jonas Gorski <jonas.gorski@gmail.com>
7 years agobrcm63xx: fix HG556a C button
dani [Fri, 22 Jul 2016 10:45:54 +0000 (12:45 +0200)]
brcm63xx: fix HG556a C button

One button of the HG556a-C has the wrong GPIO pin defined in the dts file,
fix it

Signed-off-by: Daniel Gonzalez <dgcbueu@gmail.com>
7 years agomountd: update to latest git HEAD
John Crispin [Sun, 24 Jul 2016 04:25:47 +0000 (06:25 +0200)]
mountd: update to latest git HEAD

Signed-off-by: John Crispin <john@phrozen.org>
7 years agofirewall3: update to latest git HEAD
John Crispin [Sun, 24 Jul 2016 04:23:26 +0000 (06:23 +0200)]
firewall3: update to latest git HEAD

Signed-off-by: John Crispin <john@phrozen.org>
7 years agougps: update to latest git HEAD
John Crispin [Sun, 24 Jul 2016 04:14:38 +0000 (06:14 +0200)]
ugps: update to latest git HEAD

Signed-off-by: John Crispin <john@phrozen.org>
7 years agoramips: unify etc/board.d/01_leds configuration
P.Wassi [Tue, 26 Jul 2016 17:18:53 +0000 (19:18 +0200)]
ramips: unify etc/board.d/01_leds configuration

Introduce an optional parameter at the local set_usb_led and
set_wifi_led function such that they can take a triggering
device. If no parameter is passed, behaviour is unchanged.
Signed-off-by: P.Wassi <p.wassi@gmx.at>
7 years agoar71xx: add support for gl-mifi
Zhao Gang [Tue, 19 Jul 2016 02:54:11 +0000 (02:54 +0000)]
ar71xx: add support for gl-mifi

This add initial support for gl-mifi.
Product page: http://www.gl-inet.com/mifi/

Signed-off-by: Zhao Gang <gang.zhao.42@gmail.com>
7 years agoramips: dch-m225 don't have ethernet. default enable wifi
Michael Lee [Wed, 20 Apr 2016 06:01:21 +0000 (14:01 +0800)]
ramips: dch-m225 don't have ethernet. default enable wifi

Signed-off-by: Michael Lee <igvtee@gmail.com>
7 years agoramips: remove indentation in etc/board.d/01_leds
Paul Wassi [Sat, 23 Jul 2016 14:25:38 +0000 (16:25 +0200)]
ramips: remove indentation in etc/board.d/01_leds

Remove indentation at end of line in base-files/etc/board.d/01_leds
Signed-off-by: Paul Wassi <p.wassi@gmx.at>
7 years agoramips: add missing i2s alias
John Crispin [Thu, 21 Jul 2016 21:34:25 +0000 (23:34 +0200)]
ramips: add missing i2s alias

Signed-off-by: John Crispin <john@phrozen.org>
7 years agoar71xx: clean up legacy-devices.mk
Matthias Schiffer [Wed, 27 Jul 2016 14:44:16 +0000 (16:44 +0200)]
ar71xx: clean up legacy-devices.mk

Several legacy images were not buildable because of missing profile
definitions in legacy-devices.mk since MultiProfile was removed. Update
legacy-devices.mk to provide all necessary profiles, and change ordering
to match legacy.mk.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agotoolchain/gcc/arc-2016.03: Fix building on hosts with gcc 6.x
Alexey Brodkin [Wed, 27 Jul 2016 12:51:32 +0000 (15:51 +0300)]
toolchain/gcc/arc-2016.03: Fix building on hosts with gcc 6.x

On attempt to build ARC toolchain on the host with gcc 6.1 I saw the
same failure as described in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69959

This change adds a back-port of the fix from upstream gcc, see
https://gcc.gnu.org/viewcvs/gcc?view=revision&revision=233572

Note the fix is already a part of gcc-5.4 and gcc-6.1 so no need for
the same fix for others.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
7 years agohostapd: backport mesh/ibss HT20/HT40 related fix
Felix Fietkau [Tue, 26 Jul 2016 11:47:00 +0000 (13:47 +0200)]
hostapd: backport mesh/ibss HT20/HT40 related fix

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agofstools: update to latest HEAD
Daniel Golle [Thu, 21 Jul 2016 10:13:50 +0000 (12:13 +0200)]
fstools: update to latest HEAD

Fixes duplicate ubiblock entries being listed and improves
find_mount_point to also match against a block device's
major:minor numbers (needed e.g. for /dev/root).

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
7 years agobase-files: use procd init for urandom_seed
Etienne CHAMPETIER [Wed, 20 Jul 2016 12:22:44 +0000 (12:22 +0000)]
base-files: use procd init for urandom_seed

Previous implementation was blocking the init and
breaking halt/reboot/sysupgrade (reported by Daniel Golle)

v2: use procd logging, use set -e + trap for error handling

Signed-off-by: Etienne CHAMPETIER <champetier.etienne@gmail.com>
Tested-by: Daniel Golle <daniel@makrotopia.org>
7 years agomvebu: limit image builds to profile selection
Felix Fietkau [Tue, 26 Jul 2016 06:36:45 +0000 (08:36 +0200)]
mvebu: limit image builds to profile selection

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobase-files: remove support of profile-specific base-files
Felix Fietkau [Tue, 26 Jul 2016 06:14:14 +0000 (08:14 +0200)]
base-files: remove support of profile-specific base-files

It is incompatible with multi-profile builds and has not been used
in-tree anyway.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoath9k: fix warning in client mode (GH#195)
Felix Fietkau [Tue, 26 Jul 2016 06:25:51 +0000 (08:25 +0200)]
ath9k: fix warning in client mode (GH#195)

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: remove image prefix from kernel files in KDIR
Felix Fietkau [Thu, 21 Jul 2016 15:49:19 +0000 (17:49 +0200)]
build: remove image prefix from kernel files in KDIR

This allows the image builder to change EXTRA_IMAGE_NAME at a later
point in time without breaking the build

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: remove cpio.gz and tar.gz from regular filesystem types
Felix Fietkau [Thu, 21 Jul 2016 17:02:59 +0000 (19:02 +0200)]
build: remove cpio.gz and tar.gz from regular filesystem types

They are not used to produce regular firmware images anyway. Instead,
call their build templates directly if enabled in the config

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoimage: remove shell calls from legacy ubi/ubifs image code
Felix Fietkau [Mon, 25 Jul 2016 10:37:01 +0000 (12:37 +0200)]
image: remove shell calls from legacy ubi/ubifs image code

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agokirkwood: fix UBIFS_OPTS variable in image build
Felix Fietkau [Mon, 25 Jul 2016 10:30:13 +0000 (12:30 +0200)]
kirkwood: fix UBIFS_OPTS variable in image build

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoRemove existing old link before creating a new one
Stefan Weil [Mon, 3 Aug 2009 08:54:04 +0000 (10:54 +0200)]
Remove existing old link before creating a new one

This is needed after a kernel update, otherwise the
new link is written as linux/linux-$(LINUX_VERSION).

Signed-off-by: Stefan Weil <sw@weilnetz.de>
7 years agoapm821xx: fix atheros PCIe cards on the MR24
Christian Lamparter [Sat, 23 Jul 2016 12:20:53 +0000 (14:20 +0200)]
apm821xx: fix atheros PCIe cards on the MR24

The consolidation of the MR24 and WNDR4700 subtargets
into the nand subtarget broke MR24's atheros wlan.
This was because the wndr4700's board code used a
pci_fixup routines to supply the ath9k module with
the calibration data.

This is not necessary on the MR24 as it has standard
mini-pcie ports. Hence the two AR9380 mini-pcie cards
do not need any calibration data overrides.

Cc: chrisrblake93@gmail.com
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
7 years agoapm821xx: use lzma compression for the initramfs images
Christian Lamparter [Sat, 23 Jul 2016 12:20:52 +0000 (14:20 +0200)]
apm821xx: use lzma compression for the initramfs images

The MR24's u-boot takes it sweet time decompressing the
LZMA-packed initramfs image. A user reported that
compared to the old gzip method in v2: it "takes a ton
longer to decompress like 4\x the old boot time for
decompression".

This patch also fixes a issue with the WNDR4700's initramfs
image getting to big and causing the following u-boot crash
during the decompression:

"Uncompressing Multi-File Image ... Error: inflate() returned -5
out-of-mem or overwrite error - must RESET board to recover"

This patch fixes both issues by reverting the MR24's initramfs
compression method back to gzip. And choosing to compress the
initramfs within the initramfs image as LZMA by default.

Cc: chrisrblake93@gmail.com
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
7 years agoath9k: remove intermediate queueing patch until it is fixed properly
Felix Fietkau [Sun, 24 Jul 2016 14:45:22 +0000 (16:45 +0200)]
ath9k: remove intermediate queueing patch until it is fixed properly

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoiperf3: update to version 3.1.3
Hauke Mehrtens [Sun, 24 Jul 2016 11:42:44 +0000 (13:42 +0200)]
iperf3: update to version 3.1.3

old size:
iperf3_3.0.11-1_mips_34kc_dsp.ipk       30147

new size:
iperf3_3.1.3-1_mips_34kc_dsp.ipk        33640

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agoiperf: update to version 2.0.9
Hauke Mehrtens [Sun, 24 Jul 2016 11:42:30 +0000 (13:42 +0200)]
iperf: update to version 2.0.9

old size:
iperf_2.0.8-1_mips_34kc_dsp.ipk 27911

new size:
iperf_2.0.9-1_mips_34kc_dsp.ipk 28681

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agocyassl: update to wolfssl version 3.9.6
Hauke Mehrtens [Sun, 24 Jul 2016 11:26:46 +0000 (13:26 +0200)]
cyassl: update to wolfssl version 3.9.6

Changelog: https://www.wolfssl.com/wolfSSL/Docs-wolfssl-changelog.html

old size:
libcyassl_3.9.0-1_mips_34kc_dsp.ipk     147552

new size:
libcyassl_3.9.6-1_mips_34kc_dsp.ipk     150087

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agocurl: update to version 7.50.0
Hauke Mehrtens [Sun, 24 Jul 2016 11:23:02 +0000 (13:23 +0200)]
curl: update to version 7.50.0

Changelog: https://curl.haxx.se/changes.html

old sizes:
libcurl_7.49.0-1_mips_34kc_dsp.ipk      97569
curl_7.49.0-1_mips_34kc_dsp.ipk         37925

new sizes:
libcurl_7.50.0-1_mips_34kc_dsp.ipk      97578
curl_7.50.0-1_mips_34kc_dsp.ipk         38017

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agoopenssl: re-enable NPN by default
Felix Fietkau [Sun, 24 Jul 2016 12:43:42 +0000 (14:43 +0200)]
openssl: re-enable NPN by default

Several packages rely on it

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoopenssl: add back the CAST cipher by default
Felix Fietkau [Sun, 24 Jul 2016 12:41:50 +0000 (14:41 +0200)]
openssl: add back the CAST cipher by default

At least netatalk and some ipsec packages use it

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agotools: bring back genext2fs for apm821xx
Felix Fietkau [Sat, 23 Jul 2016 17:33:28 +0000 (19:33 +0200)]
tools: bring back genext2fs for apm821xx

This reverts commit 8c68c104eaba65273280c7a4727fbb10ab10f5af.
It is used for apm821xx, which needs ext2 (not ext4) images for some
devices.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agolinux: Get rid of 000-keep_initrafs_the_default.patch
Alexey Brodkin [Fri, 22 Jul 2016 10:18:41 +0000 (13:18 +0300)]
linux: Get rid of 000-keep_initrafs_the_default.patch

With that patch in place for initramfs no additional options are
reported for "/" partition. What's really important is missing
info about sizes. Which in its turn makes opkg think that there's
no space on "/" partition to install software.

I understand that's a sort of corner-case, people rarely install
packages on ramfs but anyways why not?

Just in case that's what I see with the patch:
---------------------->8--------------------
root@lede:/# cat /proc/mounts
rootfs / rootfs rw 0 0
proc /proc proc rw,nosuid,nodev,noexec,noatime 0 0
sysfs /sys sysfs rw,nosuid,nodev,noexec,noatime 0 0
tmpfs /tmp tmpfs rw,nosuid,nodev,noatime 0 0
tmpfs /dev tmpfs rw,nosuid,relatime,size=512k,mode=755 0 0
devpts /dev/pts devpts rw,nosuid,noexec,relatime,mode=600 0 0
debugfs /sys/kernel/debug debugfs rw,noatime 0 0
---------------------->8--------------------

And without:
---------------------->8--------------------
root@lede:/# cat /proc/mounts
rootfs / rootfs rw,size=256168k,nr_inodes=32021 0 0
proc /proc proc rw,nosuid,nodev,noexec,noatime 0 0
sysfs /sys sysfs rw,nosuid,nodev,noexec,noatime 0 0
tmpfs /tmp tmpfs rw,nosuid,nodev,noatime 0 0
tmpfs /dev tmpfs rw,nosuid,relatime,size=512k,mode=755 0 0
devpts /dev/pts devpts rw,nosuid,noexec,relatime,mode=600 0 0
debugfs /sys/kernel/debug debugfs rw,noatime 0 0
---------------------->8--------------------

Note how different is entry for rootfs.

And given there's no known rationale for that patch we're
getting rid of it.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Jonas Gorski <jonas.gorski@gmail.com>
Cc: Rafał Miłecki <zajec5@gmail.com>
Cc: John Crispin <john@phrozen.org>
Cc: Felix Fietkau <nbd@nbd.name>
7 years agoopenssl: revert the no-ripemd change, openssh needs that cipher
Felix Fietkau [Sat, 23 Jul 2016 17:03:47 +0000 (19:03 +0200)]
openssl: revert the no-ripemd change, openssh needs that cipher

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoath10k: Support installing CT firmware for QCA9984 NICs.
Ben Greear [Thu, 21 Jul 2016 23:52:42 +0000 (16:52 -0700)]
ath10k: Support installing CT firmware for QCA9984 NICs.

This firmware has been lightly tested on non LEDE system
to date, and will be undergoing further testing and development.
Allow users to easily install this on their LEDE system as
they prefer.

Signed-off-by: Ben Greear <greearb@candelatech.com>
7 years agoath10k-ct-firmware: Update to latest 9880 firmware.
Ben Greear [Thu, 21 Jul 2016 23:52:41 +0000 (16:52 -0700)]
ath10k-ct-firmware: Update to latest 9880 firmware.

Fixes some issues with AES encryption and macbooks.

Signed-off-by: Ben Greear <greearb@candelatech.com>
7 years agoath10k-ct: Update to latest ath10k-ct driver.
Ben Greear [Thu, 21 Jul 2016 23:52:40 +0000 (16:52 -0700)]
ath10k-ct: Update to latest ath10k-ct driver.

Adds support for 9984 NICs, and an upstream bugfix or two.

Signed-off-by: Ben Greear <greearb@candelatech.com>
7 years agosunxi/pcduino3: Remove selection of absent rtl8188eu
Alexey Brodkin [Fri, 22 Jul 2016 10:04:06 +0000 (13:04 +0300)]
sunxi/pcduino3: Remove selection of absent rtl8188eu

Since commit 217a64375547 "wireless: remove rtl8188eu (staging)"
there's no point in selection of the missing module.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: John Crispin <john@openwrt.org>
Cc: Zoltan Herpai <wigyori@uid0.hu>
7 years agoopenssl: add option to disable SRP support
Dirk Feytons [Fri, 10 Jun 2016 10:02:15 +0000 (12:02 +0200)]
openssl: add option to disable SRP support

Signed-off-by: Dirk Feytons <dirk.feytons@gmail.com>
7 years agoopenssl: add --gc-sections
Dirk Feytons [Thu, 2 Jun 2016 11:43:14 +0000 (13:43 +0200)]
openssl: add --gc-sections

Signed-off-by: Dirk Feytons <dirk.feytons@gmail.com>
7 years agoopenssl: remove some unneeded functionality and algorithms
Dirk Feytons [Fri, 20 May 2016 13:34:02 +0000 (15:34 +0200)]
openssl: remove some unneeded functionality and algorithms

The patch needed for this commit has been sent upstream:
https://github.com/openssl/openssl/pull/1155

Signed-off-by: Dirk Feytons <dirk.feytons@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name> [add back bf and srp]
7 years agoopenssl: add option to disable PSK support
Dirk Feytons [Fri, 20 May 2016 15:23:28 +0000 (17:23 +0200)]
openssl: add option to disable PSK support

Signed-off-by: Dirk Feytons <dirk.feytons@gmail.com>
7 years agoopenssl: add option for NPN support
Dirk Feytons [Fri, 20 May 2016 11:39:12 +0000 (13:39 +0200)]
openssl: add option for NPN support

NPN has been superseded by ALPN so NPN is disabled by default
The patch has been sent to OpenSSL for inclusion, see
https://github.com/openssl/openssl/pull/1100

Signed-off-by: Dirk Feytons <dirk.feytons@gmail.com>
7 years agoopenssl: add option to disable compression support
Dirk Feytons [Thu, 19 May 2016 13:59:55 +0000 (15:59 +0200)]
openssl: add option to disable compression support

By default it's disabled. After the CRIME attack it seems the use of
compression is discouraged.

Signed-off-by: Dirk Feytons <dirk.feytons@gmail.com>