openwrt/staging/florian.git
10 years agoath9k: add workaround for AR9331 USB host controller hang with USB1.1 devices
Gabor Juhos [Fri, 10 Jan 2014 09:05:31 +0000 (09:05 +0000)]
ath9k: add workaround for AR9331 USB host controller hang with USB1.1 devices

When wifi radio is reset, some bits in (undocumented) USB registers may
flip. Patch adds setting them after radio reset (channel change, up/down).
With this patch stability of USB1.1 full-speed devices, connected directly
to AR9331 host port (no USB2.0 hub) is improved.

See "ar9331's usb stability issue" OpenWrt forum thread for details:
  <https://forum.openwrt.org/viewtopic.php?id=39956>

Based on QCA's Open HAL WAR_USB_DISABLE_PLL_LOCK_DETECT() macro:
  <https://github.com/qca/qcamain_open_hal_public/blob/master/hal/ar9300/ar9300_reset.c#L74>

Adapted to Linux mac80211 driver for 8Devices Carambola2 board by Mantas Pucka:
  <https://github.com/8devices/carambola2/commit/b08b59ceb4e7f0bad4c0b17e703e1e5c4e168923>

Ported to latest OpenWrt BB branch and tested on TP-Link TL-WR703N by
Michel Stempin <michel.stempin@wanadoo.fr>.

Signed-off-by: Michel Stempin <michel.stempin@wanadoo.fr>
[juhosg: refresh and clean up the ath9k patch]
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39212

10 years agojsonpath: make output in non-export mode more shell friendly
Jo-Philipp Wich [Wed, 8 Jan 2014 22:01:55 +0000 (22:01 +0000)]
jsonpath: make output in non-export mode more shell friendly

SVN-Revision: 39211

10 years agoupgrade the Linaro binutils to 2.24-2013.12
Imre Kaloz [Tue, 7 Jan 2014 15:03:55 +0000 (15:03 +0000)]
upgrade the Linaro binutils to 2.24-2013.12

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39210

10 years agoupgrade Linaro GCC 4.8 to 4.8-2013.12
Imre Kaloz [Tue, 7 Jan 2014 15:00:33 +0000 (15:00 +0000)]
upgrade Linaro GCC 4.8 to 4.8-2013.12

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39209

10 years agoadd x86_64 target support
Imre Kaloz [Tue, 7 Jan 2014 12:23:35 +0000 (12:23 +0000)]
add x86_64 target support

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39208

10 years agovirtualbox emulates the e1000 for a long time now
Imre Kaloz [Tue, 7 Jan 2014 10:21:31 +0000 (10:21 +0000)]
virtualbox emulates the e1000 for a long time now

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39207

10 years agomove most of the x86 image generation options to the main file
Imre Kaloz [Tue, 7 Jan 2014 10:18:55 +0000 (10:18 +0000)]
move most of the x86 image generation options to the main file

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39206

10 years agosunxi: refresh usb vbus patch
Zoltan Herpai [Mon, 6 Jan 2014 23:34:31 +0000 (23:34 +0000)]
sunxi: refresh usb vbus patch

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39205

10 years agosunxi: various fixes - DT: update USB vbus from 3.3v to 5v - emac: add missing free_i...
Zoltan Herpai [Mon, 6 Jan 2014 21:51:39 +0000 (21:51 +0000)]
sunxi: various fixes - DT: update USB vbus from 3.3v to 5v - emac: add missing free_irq - DT: add emac aliases - DT: fix interrupts on A20 - modules: add watchdog module - don't compile in mac80211 - don't compile in usbnet

Thanks to Hans de Goede, Zalan Blenessy et al.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39204

10 years agobcm53xx: make the l2x0 L2 cache controller work
Hauke Mehrtens [Sat, 4 Jan 2014 23:46:44 +0000 (23:46 +0000)]
bcm53xx: make the l2x0 L2 cache controller work

This fixes the DMA problems with the Ethernet driver.
This also updates some other parts of the patches.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 39203

10 years agobcm53xx: update the ethernet core hacks
Hauke Mehrtens [Sat, 4 Jan 2014 18:27:09 +0000 (18:27 +0000)]
bcm53xx: update the ethernet core hacks

The PHY always says there is no link
Instead of doing this dma sync a mdelay does the same trick and it is as unreliable.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 39202

10 years agokernel: bgmac: update the driver patches to a version submitted for mainline kernel.
Hauke Mehrtens [Sat, 4 Jan 2014 18:25:05 +0000 (18:25 +0000)]
kernel: bgmac: update the driver patches to a version submitted for mainline kernel.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 39201

10 years agosunxi: emac changes - move emac into module - add into profiles where appropriate
Zoltan Herpai [Sat, 4 Jan 2014 17:33:44 +0000 (17:33 +0000)]
sunxi: emac changes - move emac into module - add into profiles where appropriate

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39200

10 years agosunxi: add sata driver to profiles where appropriate
Zoltan Herpai [Sat, 4 Jan 2014 16:39:31 +0000 (16:39 +0000)]
sunxi: add sata driver to profiles where appropriate

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39198

10 years agosunxi: add sata driver for sun[47]i
Zoltan Herpai [Sat, 4 Jan 2014 16:30:56 +0000 (16:30 +0000)]
sunxi: add sata driver for sun[47]i

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39197

10 years agosunxi: move sunxi-sid (security ID) off to module
Zoltan Herpai [Sat, 4 Jan 2014 13:46:43 +0000 (13:46 +0000)]
sunxi: move sunxi-sid (security ID) off to module

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39196

10 years agosunxi: move rtc off to module
Zoltan Herpai [Sat, 4 Jan 2014 12:47:51 +0000 (12:47 +0000)]
sunxi: move rtc off to module

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39195

10 years agosunxi: add the actual mmc driver
Zoltan Herpai [Sat, 4 Jan 2014 11:08:27 +0000 (11:08 +0000)]
sunxi: add the actual mmc driver

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39194

10 years agosunxi: delete a31 reset for now in lack of the mach split
Zoltan Herpai [Sat, 4 Jan 2014 11:02:16 +0000 (11:02 +0000)]
sunxi: delete a31 reset for now in lack of the mach split

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39193

10 years agoadd support for pcDuino board
Zoltan Herpai [Sat, 4 Jan 2014 10:09:52 +0000 (10:09 +0000)]
add support for pcDuino board

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39192

10 years agoupdate version
Zoltan Herpai [Sat, 4 Jan 2014 10:08:31 +0000 (10:08 +0000)]
update version

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39191

10 years agosunxi: add support for pcDuino board
Zoltan Herpai [Sat, 4 Jan 2014 10:06:17 +0000 (10:06 +0000)]
sunxi: add support for pcDuino board

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39190

10 years agosunxi: add mmc support - add MMC support for sun457i - update kernel config to compil...
Zoltan Herpai [Sat, 4 Jan 2014 10:02:01 +0000 (10:02 +0000)]
sunxi: add mmc support - add MMC support for sun457i - update kernel config to compile in MMC and ext4 - update kernel cmdline

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39189

10 years agosunxi: various fixes - have Kconfig options ordered - add new olinuxino a13-micro...
Zoltan Herpai [Sat, 4 Jan 2014 09:55:41 +0000 (09:55 +0000)]
sunxi: various fixes - have Kconfig options ordered - add new olinuxino a13-micro board - various fixes for EHCI

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39188

10 years agosunxi: remove sun6i/7i split for now
Zoltan Herpai [Sat, 4 Jan 2014 09:49:44 +0000 (09:49 +0000)]
sunxi: remove sun6i/7i split for now

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39187

10 years agosunxi: refresh clock framework
Zoltan Herpai [Sat, 4 Jan 2014 09:47:50 +0000 (09:47 +0000)]
sunxi: refresh clock framework

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
SVN-Revision: 39186

10 years agoAdd jsonpath - a command line utility to extract values from JSON data using XPath...
Jo-Philipp Wich [Thu, 2 Jan 2014 11:55:16 +0000 (11:55 +0000)]
Add jsonpath - a command line utility to extract values from JSON data using XPath-like expressions

SVN-Revision: 39185

10 years ago6relayd: try to workaround race condition at startup
Steven Barth [Thu, 2 Jan 2014 11:38:36 +0000 (11:38 +0000)]
6relayd: try to workaround race condition at startup

SVN-Revision: 39184

10 years agonetifd: initialize the switch early at start time and on reload (fixes #13015)
Felix Fietkau [Tue, 31 Dec 2013 13:09:20 +0000 (13:09 +0000)]
netifd: initialize the switch early at start time and on reload (fixes #13015)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39183

10 years agoixp4xx: upgrade: don't copy extra binaries
Luka Perkov [Mon, 30 Dec 2013 23:49:25 +0000 (23:49 +0000)]
ixp4xx: upgrade: don't copy extra binaries

hexdump is already added to new temporary file system while less is not used at
all. While at it, remove some trailing whitespaces.

Signed-off-by: Luka Perkov <luka@openwrt.org>
CC: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39181

10 years agonetifd: update to the latest version, fixes wireless related segfaults on arm
Felix Fietkau [Sun, 29 Dec 2013 02:26:34 +0000 (02:26 +0000)]
netifd: update to the latest version, fixes wireless related segfaults on arm

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39180

10 years agouClibc: Fix lookup with DNS search in multi-threaded application.
Felix Fietkau [Sat, 28 Dec 2013 15:06:14 +0000 (15:06 +0000)]
uClibc: Fix lookup with DNS search in multi-threaded application.

Signed-off-by: Jiri Slachta <slachta@cesnet.cz>
SVN-Revision: 39179

10 years agouClibc: make res_init() thread safe
Felix Fietkau [Sat, 28 Dec 2013 15:06:04 +0000 (15:06 +0000)]
uClibc: make res_init() thread safe

Signed-off-by: Jiri Slachta <slachta@cesnet.cz>
SVN-Revision: 39178

10 years agouClibc: Fix threaded use of res_ functions.
Felix Fietkau [Sat, 28 Dec 2013 15:05:56 +0000 (15:05 +0000)]
uClibc: Fix threaded use of res_ functions.

This patch moves res_init() back above #undef _res. It fixes dns resolving issue in OpenWrt
(uClibc related - OpenWrt ticket #11929). It is a backport from uClibc master.

Further details are there:
http://git.uclibc.org/uClibc/commit/libc/inet/resolv.c?id=20b69920b299585265eb100d0b67e1097ccb1092

Signed-off-by: Jiri Slachta <slachta@cesnet.cz>
SVN-Revision: 39177

10 years agonetifd: update to the latest version, fixes wifi related segfaults
Felix Fietkau [Sat, 28 Dec 2013 14:19:54 +0000 (14:19 +0000)]
netifd: update to the latest version, fixes wifi related segfaults

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39176

10 years agohostapd: fix mixed wep/wpa with netifd
Felix Fietkau [Sat, 28 Dec 2013 11:07:37 +0000 (11:07 +0000)]
hostapd: fix mixed wep/wpa with netifd

Signed-off-by: Catalin Patulea <cat@vv.carleton.ca>
SVN-Revision: 39174

10 years agohostapd: fix the uci option name for ap isolate
Felix Fietkau [Sat, 28 Dec 2013 11:07:28 +0000 (11:07 +0000)]
hostapd: fix the uci option name for ap isolate

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39173

10 years agofirmware-utils: check_magic() in buffalo-lib.c always return 0
John Crispin [Fri, 27 Dec 2013 21:15:31 +0000 (21:15 +0000)]
firmware-utils: check_magic() in buffalo-lib.c always return 0

spotted with cppcheck

Signed-off-by: Etienne CHAMPETIER <etienne.champetier@free.fr>
SVN-Revision: 39171

10 years agoswconfig: remove useless variables, return -1 on errors
John Crispin [Fri, 27 Dec 2013 21:15:20 +0000 (21:15 +0000)]
swconfig: remove useless variables, return -1 on errors

spotted with cppcheck

Signed-off-by: Etienne CHAMPETIER <etienne.champetier@free.fr>
SVN-Revision: 39170

10 years agobase-files: fix mount-utils incompatibility
John Crispin [Fri, 27 Dec 2013 21:15:09 +0000 (21:15 +0000)]
base-files: fix mount-utils incompatibility

Signed-off-by: Toke Høiland-Jørgensen <toke@toke.dk>
SVN-Revision: 39169

10 years agoralink: fix tso regression in the ethernet driver
John Crispin [Fri, 27 Dec 2013 21:14:49 +0000 (21:14 +0000)]
ralink: fix tso regression in the ethernet driver

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 39168

10 years agobrcm47xx: fix cpu wait for BCM4706
Hauke Mehrtens [Thu, 26 Dec 2013 23:30:10 +0000 (23:30 +0000)]
brcm47xx: fix cpu wait for BCM4706

This was called to early.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 39167

10 years agobrcm47xx: b44: This updates the phylib patches to the version send upstream
Hauke Mehrtens [Thu, 26 Dec 2013 23:27:52 +0000 (23:27 +0000)]
brcm47xx: b44: This updates the phylib patches to the version send upstream

This uses a fixed phy instead of a dummy one.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 39166

10 years agokernel: bgmac: update patches
Hauke Mehrtens [Thu, 26 Dec 2013 23:22:31 +0000 (23:22 +0000)]
kernel: bgmac: update patches

This updates the bgmac backport patch to the current upstream version
and adds some more patches mostly for BCM4707.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 39165

10 years agohandle the Linaro version as well
Imre Kaloz [Thu, 26 Dec 2013 22:19:29 +0000 (22:19 +0000)]
handle the Linaro version as well

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39164

10 years agoramips: add support for RT-N14U (mt7620n based) board
John Crispin [Wed, 25 Dec 2013 17:04:50 +0000 (17:04 +0000)]
ramips: add support for RT-N14U (mt7620n based) board

added dts file and userspace scripts modified to support Asus RT-N14U board

current support status:
usb        works
ethernet   works
buttons    reset, wps
leds       asus:blue:[usb|lan|wan|air|power]
i2c        not tested
uart       not tested
wifi       not yet

Signed-off-by: Pavel Löbl <lobl.pavel@gmail.com>
SVN-Revision: 39163

10 years agoramips: add gpio pin 72 in mt7620 chips to dtsi files
John Crispin [Wed, 25 Dec 2013 17:04:34 +0000 (17:04 +0000)]
ramips: add gpio pin 72 in mt7620 chips to dtsi files

describes register set to control last gpio pin on mt7620 platfrom

Signed-off-by: Pavel Löbl <lobl.pavel@gmail.com>
SVN-Revision: 39162

10 years agoar71xx: fix max frame length of the QCA955x SoCs
Gabor Juhos [Mon, 23 Dec 2013 17:05:23 +0000 (17:05 +0000)]
ar71xx: fix max frame length of the QCA955x SoCs

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39161

10 years agoar71xx: don't set builtin_switch flag for QCA9558
Gabor Juhos [Mon, 23 Dec 2013 17:05:21 +0000 (17:05 +0000)]
ar71xx: don't set builtin_switch flag for QCA9558

It makes no sense, the SoC has no built-in switch.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39160

10 years agox86/kvm_guest: remove 3.7 support
Gabor Juhos [Mon, 23 Dec 2013 17:05:19 +0000 (17:05 +0000)]
x86/kvm_guest: remove 3.7 support

Patch-by: Stijn Tintel <stijn@linux-ipv6.be>
Patchwork: http://patchwork.openwrt.org/patch/4637/
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39159

10 years agox86/kvm_guest: remove 3.3 support
Gabor Juhos [Mon, 23 Dec 2013 17:05:18 +0000 (17:05 +0000)]
x86/kvm_guest: remove 3.3 support

Patch-by: Stijn Tintel <stijn@linux-ipv6.be>
Patchwork: http://patchwork.openwrt.org/patch/46366/
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39158

10 years agox86/kvm_guest: add support for 3.10 and switch to it
Gabor Juhos [Mon, 23 Dec 2013 17:05:16 +0000 (17:05 +0000)]
x86/kvm_guest: add support for 3.10 and switch to it

Patch-by: Stijn Tintel <stijn@linux-ipv6.be>
Patchwork: http://patchwork.openwrt.org/patch/4635/
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39157

10 years agohostapd: fix wep with netifd
Felix Fietkau [Mon, 23 Dec 2013 12:11:28 +0000 (12:11 +0000)]
hostapd: fix wep with netifd

Signed-off-by: Catalin Patulea <cat@vv.carleton.ca>
SVN-Revision: 39156

10 years agowpa_supplicant: fix interface combination parsing issues
Felix Fietkau [Sat, 21 Dec 2013 19:42:45 +0000 (19:42 +0000)]
wpa_supplicant: fix interface combination parsing issues

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39155

10 years agoopkg: switch to git repository (#14655)
Jo-Philipp Wich [Sat, 21 Dec 2013 13:47:46 +0000 (13:47 +0000)]
opkg: switch to git repository (#14655)

Upstream development now takes place in a git repository which was clone
from the old svn repository. Switch the source location to the git tree
using the commit that corresponds to the old svn r618 revision.

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 39153

10 years agodnsmasq: rework init procedure
Jo-Philipp Wich [Sat, 21 Dec 2013 13:31:28 +0000 (13:31 +0000)]
dnsmasq: rework init procedure

- cache udhcp check results to speed up subsequent reloads
- enable procd file tracking for /var/etc/dnsmasq.conf to only reload service if needed
- implement reload action to only restart dnsmasq if /var/etc/dnsmasq.conf actually changed
- launch dnsmasq from interface hotplug to avoid race conditions with network bringup

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 39152

10 years agochange fixup method and fix CFLAGS handling
Imre Kaloz [Fri, 20 Dec 2013 20:25:44 +0000 (20:25 +0000)]
change fixup method and fix CFLAGS handling

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39151

10 years agoath9k: merge a timer handling fixes
Felix Fietkau [Fri, 20 Dec 2013 16:02:32 +0000 (16:02 +0000)]
ath9k: merge a timer handling fixes

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39150

10 years agoar71xx: allow to use large ethernet frames on AR934x SoCs
Gabor Juhos [Fri, 20 Dec 2013 11:41:23 +0000 (11:41 +0000)]
ar71xx: allow to use large ethernet frames on AR934x SoCs

The hardware supports large ethernet frames. Override
the maximum frame length and packet lenght mask in the
platform data to allow to use large MTU on the ethernet
interfaces.

Limit the feature to AR934x SoCs for now. It should work
on some other SoCs as well, but those has not been tested
yet.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39149

10 years agoar71xx: ag71xx: fix max frame length setup of the built-in switches
Gabor Juhos [Fri, 20 Dec 2013 11:41:22 +0000 (11:41 +0000)]
ar71xx: ag71xx: fix max frame length setup of the built-in switches

The currently used bitmask of the maximum frame length field
is wrong for both models. On AR724x/AR933x the largest frame
size is 2047 bytes, on the AR934x it is 16383 bytes.

Make the MTU setup code model specific, and use the correct
bitmask for both models. Also change the value to the maximum.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39148

10 years agoar71xx: ag71xx: calculate max frame len register value from the MTU
Gabor Juhos [Fri, 20 Dec 2013 11:41:20 +0000 (11:41 +0000)]
ar71xx: ag71xx: calculate max frame len register value from the MTU

Set the MAX_FRAME_LEN register to zero in ag71xx_hw_init()
and write the correct value into that from the ag71xx_open()
and ag71xx_fast_reset() functions.

Also recalculate the RX buffer size based on the actual
maximum frame length value to optimize memory allocation.
Additionaly, disallow to change the MTU value while the
interface it running.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39147

10 years agoar71xx: ag71xx: add ag71xx_max_frame_len() helper
Gabor Juhos [Fri, 20 Dec 2013 11:41:18 +0000 (11:41 +0000)]
ar71xx: ag71xx: add ag71xx_max_frame_len() helper

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39146

10 years agoar71xx: ag71xx: get max_frame_len and desc_pktlen_mask from platform data
Gabor Juhos [Fri, 20 Dec 2013 11:41:17 +0000 (11:41 +0000)]
ar71xx: ag71xx: get max_frame_len and desc_pktlen_mask from platform data

This will allow to use SoC specific values for both.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39145

10 years agoar71xx: ag71xx: store descriptor packet length mask in ag71xx struct
Gabor Juhos [Fri, 20 Dec 2013 11:41:16 +0000 (11:41 +0000)]
ar71xx: ag71xx: store descriptor packet length mask in ag71xx struct

The currently used bitmask is not correct for all SoCs.
Introduce a new field in struct ag71xx and store the
bitmask in that. Use the current value for now, it will
be adjusted for each SoCs in further patches.

Aslo use the new field directly in the ag71xx_rx_packets
and ag71xx_hard_start_xmit() functions and remove the
ag71xx_desc_pktlen() helper.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39144

10 years agohandle new libs required for new kernels
Imre Kaloz [Fri, 20 Dec 2013 06:09:41 +0000 (06:09 +0000)]
handle new libs required for new kernels

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39142

10 years agodon't fail if there are no dtbs
Imre Kaloz [Fri, 20 Dec 2013 05:50:24 +0000 (05:50 +0000)]
don't fail if there are no dtbs

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 39141

10 years agoar71xx: add missing tx gain table flag for buffalo wzr-hp-g450h
Felix Fietkau [Thu, 19 Dec 2013 17:18:12 +0000 (17:18 +0000)]
ar71xx: add missing tx gain table flag for buffalo wzr-hp-g450h

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39139

10 years agoath9k: merge another round of upstream (or -pending) fixes and cleanups
Felix Fietkau [Thu, 19 Dec 2013 17:18:07 +0000 (17:18 +0000)]
ath9k: merge another round of upstream (or -pending) fixes and cleanups

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39138

10 years agoimx6: refresh patches
Luka Perkov [Thu, 19 Dec 2013 02:03:26 +0000 (02:03 +0000)]
imx6: refresh patches

Signed-off-by: Luka Perkov <luka@openwrt.org>
SVN-Revision: 39135

10 years agokernel: backport upstream inet fix
Luka Perkov [Thu, 19 Dec 2013 02:03:25 +0000 (02:03 +0000)]
kernel: backport upstream inet fix

Without it in some occasions oops is shown on network restart. Problem was
noticed on imx6 targed.

Signed-off-by: Luka Perkov <luka@openwrt.org>
SVN-Revision: 39134

10 years agoimx6: proper fix for io-remap
Luka Perkov [Thu, 19 Dec 2013 02:03:16 +0000 (02:03 +0000)]
imx6: proper fix for io-remap

The previous patch was not the proper fix for PCI devices that require io
resources.  The new patch is the proper fix backported from mainline.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 39133

10 years agoprocd: fix service file tracking
Jo-Philipp Wich [Wed, 18 Dec 2013 15:38:28 +0000 (15:38 +0000)]
procd: fix service file tracking

The instance_config_move() function in the procd instance managing code is
missing a blobmsg_list_move() call to update the tracked file items with the
newly calculated checksums in case of a file change.

Without that change, services are always reloaded after the first change to
a tracked file regardless of whether there are any subsequent changes or not.

Add a missing instance of blobmsg_list_free() as well.

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 39132

10 years agonetifd: add reload trigger for /etc/config/wireless as well
Jo-Philipp Wich [Wed, 18 Dec 2013 12:38:29 +0000 (12:38 +0000)]
netifd: add reload trigger for /etc/config/wireless as well

SVN-Revision: 39131

10 years agoar71xx: enable images for WD My Net N750
Gabor Juhos [Tue, 17 Dec 2013 22:14:13 +0000 (22:14 +0000)]
ar71xx: enable images for WD My Net N750

Now that the switch works correctly we can enable the images for this
board.

Flashing from the factory firmware, the factory failsafe loader and
sysupgrade was tested successfully.

Signed-off-by: Felix Kaechele <heffer@fedoraproject.org>
Patchwork: http://patchwork.openwrt.org/patch/4617/
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39130

10 years agoar71xx: fix WD My Net N750 switch led cfg
Gabor Juhos [Tue, 17 Dec 2013 22:14:12 +0000 (22:14 +0000)]
ar71xx: fix WD My Net N750 switch led cfg

Now that the switch is working correctly I had the chance to actually
test the LED config.

Signed-off-by: Felix Kaechele <heffer@fedoraproject.org>
Patchwork: http://patchwork.openwrt.org/patch/4616/
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39129

10 years agoar71xx: implement switch fix for WD My Net N750
Gabor Juhos [Tue, 17 Dec 2013 22:14:11 +0000 (22:14 +0000)]
ar71xx: implement switch fix for WD My Net N750

The bootloader on the WD My Net N750 disables the ports on it's internal
AR8327N switch by powering them down. The stock firmware then brings the
ports back up again by starting the auto negotiation process on each
port.

This fix implements just that.

Signed-off-by: Felix Kaechele <heffer@fedoraproject.org>
Patchwork: http://patchwork.openwrt.org/patch/4615/
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39128

10 years agoar71xx: implement callback in mdio reset
Gabor Juhos [Tue, 17 Dec 2013 22:14:10 +0000 (22:14 +0000)]
ar71xx: implement callback in mdio reset

This enables us to add fixups to the board specific code for boards that
require special treatment of PHYs on mdio bus reset.

Signed-off-by: Felix Kaechele <heffer@fedoraproject.org>
http://patchwork.openwrt.org/patch/4614/
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39127

10 years agoar71xx: make ag71xx_mdio_platform_data visible
Gabor Juhos [Tue, 17 Dec 2013 22:14:07 +0000 (22:14 +0000)]
ar71xx: make ag71xx_mdio_platform_data visible

This enables us to modify the ag71xx_mdio_platform_data from within the
board support files.

Signed-off-by: Felix Kaechele <heffer@fedoraproject.org>
Patchwork: http://patchwork.openwrt.org/patch/4613/
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39126

10 years agobusybox: add back support for top-level menuconfig based config overrides (can be...
Felix Fietkau [Tue, 17 Dec 2013 20:37:27 +0000 (20:37 +0000)]
busybox: add back support for top-level menuconfig based config overrides (can be used by custom packages to alter the config)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39125

10 years agogpio-button-hotplug: fix crash on remove
Jonas Gorski [Tue, 17 Dec 2013 19:05:03 +0000 (19:05 +0000)]
gpio-button-hotplug: fix crash on remove

Don't call gpio_keys_remove recursively. Setting the platform
data to NULL triggered an oops on the second iteration, so there was
no infinate loop.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 39124

10 years agobusybox: restore init scripts
Jo-Philipp Wich [Tue, 17 Dec 2013 17:17:00 +0000 (17:17 +0000)]
busybox: restore init scripts

Since the removal of the busybox menuconfig entries, the init scripts
for cron, telnet and ntp are not packaged anymore. Unconditionally
ship them from now on.

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 39123

10 years agofirewall: fix handling of tcp_ecn parameter
Jo-Philipp Wich [Tue, 17 Dec 2013 16:59:47 +0000 (16:59 +0000)]
firewall: fix handling of tcp_ecn parameter

The firewall3 implementation as well as the shell implementation predating it
used to process the tcp_ecnoption as boolean while it actually is an integer.

Change the code to parse tcp_ecn as integer.

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 39122

10 years agoar71xx: ag71xx: compute the RX buffer size from the maximum frame size
Gabor Juhos [Tue, 17 Dec 2013 16:27:46 +0000 (16:27 +0000)]
ar71xx: ag71xx: compute the RX buffer size from the maximum frame size

Currently, the AG71XX_RX_PKT_SIZE value limits the received
frame size to 1514/1516 bytes with/without a VLAN header
respectively. However the hardware limit is controlled by
the value the AG71XX_REG_MAC_MFL register which contains
the value of the max_frame_len field.

Compute the RX buffer size from the max_frame_len field
to get rid of the 1514/1516 byte limitation. Also remove
the unused AG71XX_RX_PKT_SIZE definition.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39121

10 years agoar71xx: ag71xx: store RX buffer size in the ag71xx struct
Gabor Juhos [Tue, 17 Dec 2013 16:27:44 +0000 (16:27 +0000)]
ar71xx: ag71xx: store RX buffer size in the ag71xx struct

This allows to change the value dynamically.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39120

10 years agoiproute2: reduce .ipk package size from ~100k down to ~60k
Felix Fietkau [Tue, 17 Dec 2013 15:59:29 +0000 (15:59 +0000)]
iproute2: reduce .ipk package size from ~100k down to ~60k

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39119

10 years agobusybox: disable module utils, kmodloader provides them
Felix Fietkau [Tue, 17 Dec 2013 15:59:25 +0000 (15:59 +0000)]
busybox: disable module utils, kmodloader provides them

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39118

10 years agoar71xx: ag71xx: store maximum frame length in the ag71xx struct
Gabor Juhos [Tue, 17 Dec 2013 15:18:16 +0000 (15:18 +0000)]
ar71xx: ag71xx: store maximum frame length in the ag71xx struct

This will allow to use different values for the
different SoCs.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39117

10 years agoar71xx: ag71xx: use mdio bus name in ar7240_probe messages
Gabor Juhos [Tue, 17 Dec 2013 15:12:52 +0000 (15:12 +0000)]
ar71xx: ag71xx: use mdio bus name in ar7240_probe messages

The ar7240_probe function uses the network device name
in the kernel log messages, however the name is not yet
initialized when the ar7240_probe function is called.
Use the mdio bus name in the messages to avoid ugly
log lines like the following one:

  eth%d: Found an AR7240/AR9330 built-in switch

Reported-by: Ronald Wahl <ronald.wahl@raritan.com>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39116

10 years agoag71xx: ag71xx: use device name for debugfs entry
Gabor Juhos [Tue, 17 Dec 2013 15:12:50 +0000 (15:12 +0000)]
ag71xx: ag71xx: use device name for debugfs entry

The ag71xx debugfs code uses the network device name
for the device specific debugfs directory. Since r38689
'ar71xx: ag71xx: fix a race involving netdev registration'
the debugfs initialization happens before the ethernet
device gets registered and the network device name contains
'eth%d' at this point. If the board setup code registers
multiple ag71xx devices, the debugfs code tries to create
the device specific dir with the same name which causes
an error like this:

  eth0: Atheros AG71xx at 0xba000000, irq 5, mode:GMII
  ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.1:04 [uid=004dd041, driver=Generic PHY]
  ag71xx: probe of ag71xx.0 failed with error -12

Use the device name for the debugfs directory to avoid the
collisions. Also add an error message and change the return
code if the debugfs_create_dir call fails.

Reported-by: Ronald Wahl <ronald.wahl@raritan.com>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 39115

10 years agobusybox: fix typo in Config.in (#14628)
Felix Fietkau [Tue, 17 Dec 2013 07:17:00 +0000 (07:17 +0000)]
busybox: fix typo in Config.in (#14628)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39114

10 years agoimx6: ventana: add canbus utils
Luka Perkov [Tue, 17 Dec 2013 02:47:20 +0000 (02:47 +0000)]
imx6: ventana: add canbus utils

Some of the Gateworks Ventana boards have CAN bus support, therefore
we will install the userspace utils.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 39113

10 years agoimx6: ventana: add default kernel modules for GSC
Luka Perkov [Tue, 17 Dec 2013 02:47:18 +0000 (02:47 +0000)]
imx6: ventana: add default kernel modules for GSC

All Gateworks Ventana boards have a Gateworks System Controller
which emulates the following:
 - ds1672 RTC
 - pca953x GPIO expander
 - hardware monitor
 - at24 eeprom

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 39112

10 years agoimx6: add rtc feature support
Luka Perkov [Tue, 17 Dec 2013 02:47:17 +0000 (02:47 +0000)]
imx6: add rtc feature support

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Luka Perkov <luka@openwrt.org>
SVN-Revision: 39111

10 years agokernel: add rtc-ds1672 module support
Luka Perkov [Tue, 17 Dec 2013 02:47:16 +0000 (02:47 +0000)]
kernel: add rtc-ds1672 module support

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 39110

10 years agoimx6: drop upstreamed patch
Luka Perkov [Tue, 17 Dec 2013 02:47:15 +0000 (02:47 +0000)]
imx6: drop upstreamed patch

Signed-off-by: Luka Perkov <luka@openwrt.org>
SVN-Revision: 39109

10 years agoautomake: update to 1.11.6
Luka Perkov [Tue, 17 Dec 2013 02:47:13 +0000 (02:47 +0000)]
automake: update to 1.11.6

Signed-off-by: Luka Perkov <luka@openwrt.org>
SVN-Revision: 39108

10 years agobusybox: add support for config overrides via env/busybox-config
Felix Fietkau [Mon, 16 Dec 2013 19:27:35 +0000 (19:27 +0000)]
busybox: add support for config overrides via env/busybox-config

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39107

10 years agobusybox: remove busybox menuconfig from top level menuconfig
Felix Fietkau [Mon, 16 Dec 2013 19:09:57 +0000 (19:09 +0000)]
busybox: remove busybox menuconfig from top level menuconfig

This will help with avoiding bogus bug reports caused by stale configs
and bad cases of user tinkering :)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39106

10 years agouClibc: disable sha256 and sha512 for libcrypt - saves about 13k uncompressed in...
Felix Fietkau [Mon, 16 Dec 2013 18:49:10 +0000 (18:49 +0000)]
uClibc: disable sha256 and sha512 for libcrypt - saves about 13k uncompressed in libcrypt

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39105

10 years agohostapd: move old wifi setup scripts to hostapd-common-old
Felix Fietkau [Mon, 16 Dec 2013 18:42:43 +0000 (18:42 +0000)]
hostapd: move old wifi setup scripts to hostapd-common-old

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 39103