openwrt/staging/xback.git
9 years agompc85xx: disable SPE FPU support until musl gets support for it (fixes #19896)
Felix Fietkau [Wed, 8 Jul 2015 12:06:16 +0000 (12:06 +0000)]
mpc85xx: disable SPE FPU support until musl gets support for it (fixes #19896)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46268

9 years agomusl: add powerpc soft-float support
Felix Fietkau [Wed, 8 Jul 2015 12:06:03 +0000 (12:06 +0000)]
musl: add powerpc soft-float support

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46267

9 years agomusl: call a_crash() inside __stack_chk_fail_local instead of calling __stack_chk_fail
Felix Fietkau [Wed, 8 Jul 2015 12:05:55 +0000 (12:05 +0000)]
musl: call a_crash() inside __stack_chk_fail_local instead of calling __stack_chk_fail

Fixes relocation errors on PowerPC

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46266

9 years agokernel: add missing config symbol (fixes #19880)
Felix Fietkau [Wed, 8 Jul 2015 08:20:44 +0000 (08:20 +0000)]
kernel: add missing config symbol (fixes #19880)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46265

9 years agokernel: add missing 4.1 config symbol
Felix Fietkau [Wed, 8 Jul 2015 08:20:38 +0000 (08:20 +0000)]
kernel: add missing 4.1 config symbol

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46264

9 years agonvram: get rid of NVRAM_SPACE
Rafał Miłecki [Wed, 8 Jul 2015 06:05:43 +0000 (06:05 +0000)]
nvram: get rid of NVRAM_SPACE

Now we support NVRAM whatever its size is.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 46262

9 years agonvram: fix displayed info about NVRAM size
Rafał Miłecki [Wed, 8 Jul 2015 05:42:30 +0000 (05:42 +0000)]
nvram: fix displayed info about NVRAM size

Use actual partition size and content offset to calculate it.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 46260

9 years agoRevert "ar71xx: Allow to use ath79_gpio_output_select on QCA955x" (r46240)
Felix Fietkau [Tue, 7 Jul 2015 22:44:28 +0000 (22:44 +0000)]
Revert "ar71xx: Allow to use ath79_gpio_output_select on QCA955x" (r46240)

It breaks the build

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46256

9 years agoscripts/feeds: print the source of packages installed
Steven Barth [Tue, 7 Jul 2015 22:32:38 +0000 (22:32 +0000)]
scripts/feeds: print the source of packages installed

Can be helpful in diagnosing unexpected package overlaps between
feeds/core.

Signed-off-by: Karl Palsson <karlp@remake.is>
SVN-Revision: 46255

9 years agomwlwifi: fix compile error after mac80211 update
Felix Fietkau [Tue, 7 Jul 2015 22:07:22 +0000 (22:07 +0000)]
mwlwifi: fix compile error after mac80211 update

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46254

9 years agoRevert "Revert "nvram: increase NVRAM size to 64 KiB""
Rafał Miłecki [Tue, 7 Jul 2015 22:06:04 +0000 (22:06 +0000)]
Revert "Revert "nvram: increase NVRAM size to 64 KiB""

This reverts commit r45579.

With the latest change r46251 ("nvram: fix "Segmentation fault" caused
by setting memory out of buffer") nvram utility shouldn't crash anymore.

It was tested on 3 brcm47xx devices:
1) Unknown with 0x10000 NVRAM size (0x8000 offset)
2) Linksys E1000 V2.1 with 0x10000 (0x8000 offset)
3) Linksys WRT300N V1 with 0x10000 (0x8000 offset)

And 3 bcm53xx devices:
1) Buffalo WZR-600DHP2 with 0x160000 NVRAM size
2) Buffalo WZR-1750DHP with 0x10000 NVRAM size
3) Netgear R6250 V1 with 0x180000 NVRAM size
(all using 0 offset)

This is an important change as it allows reading whole NVRAM. This may
critical when reading some basic configuration (e.g. switch ports).

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 46253

9 years agonvram: fix "Segmentation fault" caused by setting memory out of buffer
Rafał Miłecki [Tue, 7 Jul 2015 21:21:49 +0000 (21:21 +0000)]
nvram: fix "Segmentation fault" caused by setting memory out of buffer

Some MTD partitions with NVRAM have content starting in the middle. In
such case offset is set and nvram_header returns pointer to the middle.
It means we have to respect offset when calculating remaining space.

By the way use real MTD partition size (nvram_part_size variable) as we
may want to bump NVRAM_SPACE in the (very near) future.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 46251

9 years agoar71xx: Add support for Cisco-Linksys WAP4410N
John Crispin [Tue, 7 Jul 2015 18:06:39 +0000 (18:06 +0000)]
ar71xx: Add support for Cisco-Linksys WAP4410N

This patch adds support for the Cisco WAP4410N, an access point that uses the
AR9132 SoC. Web upgrades from stock are not yet possible, UART access required
for the initial flash.

Signed-off-by: Ryan A Young <rayoung@utexas.edu>
SVN-Revision: 46250

9 years agogdb: fix compilation for x86_64 with musl
Steven Barth [Tue, 7 Jul 2015 17:30:08 +0000 (17:30 +0000)]
gdb: fix compilation for x86_64 with musl

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 46249

9 years agoar71xx: add support for D-Link DIR-615 rev. Ix board
John Crispin [Tue, 7 Jul 2015 13:48:39 +0000 (13:48 +0000)]
ar71xx: add support for D-Link DIR-615 rev. Ix board

Based on trunk r46212.

Signed-off-by: Jaehoon You <teslamint@gmail.com>
SVN-Revision: 46248

9 years agomips: Free memory when load_module fails (#14453)
John Crispin [Tue, 7 Jul 2015 13:48:27 +0000 (13:48 +0000)]
mips: Free memory when load_module fails (#14453)

The mips reloc patch introduced new allocations which were done before
add_unformed_module but never freed them in case of an error. A new hook in
Linux 3.19 called module_arch_freeing_init can be used for freeing memory
which were allocated during this init phase.

The problem can be seen when trying to load a module (via busybox insmod)
when it was already loaded.

    free -m
    for i in `seq 1 100`; do
        /sbin/insmod /lib/modules/*/ath9k.ko >& /dev/null
    done
    free -m

This simple loop would leak ~3.2 MB.

Signed-off-by: Sven Eckelmann <sven@open-mesh.com>
SVN-Revision: 46247

9 years agoubus: bump to latest git HEAD
John Crispin [Tue, 7 Jul 2015 13:48:17 +0000 (13:48 +0000)]
ubus: bump to latest git HEAD

fixes a bug where gid was checked twice instead of the uid/gid combo

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 46246

9 years agomac80211: Drop ath10k workaround for providing calibration data
John Crispin [Tue, 7 Jul 2015 13:48:06 +0000 (13:48 +0000)]
mac80211: Drop ath10k workaround for providing calibration data

Ath10k has now a proper method of providing calibration data via
the kernel firmware API. This patch can be dropped as all boards
now use the proper method.

Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 46245

9 years agoar71xx: Unify ath10k calibration data loading
John Crispin [Tue, 7 Jul 2015 13:47:57 +0000 (13:47 +0000)]
ar71xx: Unify ath10k calibration data loading

Currently one board uses the "proper" way of providing firmware
with the name ath10k/cal-pci-<pci address>.bin, whereas other
boards patch board.bin template in preinit and rely on a custom
ath10k patch to use calibration data from the template.

This patch makes all boards use the first method.

Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 46244

9 years agomac80211: ath9k: fix qca956x name also
John Crispin [Tue, 7 Jul 2015 13:47:51 +0000 (13:47 +0000)]
mac80211: ath9k: fix qca956x name also

Signed-off-by: Roman Yeryomin <roman@advem.lv>
SVN-Revision: 46243

9 years agokernel: move wpan to separate menu
John Crispin [Tue, 7 Jul 2015 13:47:45 +0000 (13:47 +0000)]
kernel: move wpan to separate menu

bluetooth is separate from 802.11 entries too

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 46242

9 years agoar71xx: Use *_eth_cfg helper for Open Mesh MR900 boards
John Crispin [Tue, 7 Jul 2015 13:47:39 +0000 (13:47 +0000)]
ar71xx: Use *_eth_cfg helper for Open Mesh MR900 boards

Signed-off-by: Sven Eckelmann <sven@open-mesh.com>
SVN-Revision: 46241

9 years agoar71xx: Allow to use ath79_gpio_output_select on QCA955x
John Crispin [Tue, 7 Jul 2015 13:47:30 +0000 (13:47 +0000)]
ar71xx: Allow to use ath79_gpio_output_select on QCA955x

Signed-off-by: Sven Eckelmann <sven@open-mesh.com>
SVN-Revision: 46240

9 years agosysctl: read settings from /etc/sysctl.d/*.conf
John Crispin [Tue, 7 Jul 2015 13:47:23 +0000 (13:47 +0000)]
sysctl: read settings from /etc/sysctl.d/*.conf

This changes makes it possible to store custom settings
in individual files inside the directory /etc/sysctl.d/.

Signed-off-by: Stefan Tomanek <stefan.tomanek+openwrt@wertarbyte.de>
SVN-Revision: 46239

9 years agoar71xx: add support for TL-WR741ND v5
John Crispin [Tue, 7 Jul 2015 13:47:17 +0000 (13:47 +0000)]
ar71xx: add support for TL-WR741ND v5

This device is identical to the TL-WR740N v5, it even uses the same HWID (which
wasn't the case for older TL-WR741ND revisions).

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 46238

9 years agogeneric: mtd backport for Spansion S25FL164K
John Crispin [Tue, 7 Jul 2015 13:47:08 +0000 (13:47 +0000)]
generic: mtd backport for Spansion S25FL164K

Backport for the Spansion S25FL164K
It's a 8 MiB flash chip with 4 KiB erase sectors.

Signed-off-by: L. D. Pinney <ldpinney@gmail.com>
SVN-Revision: 46237

9 years agoar71xx: Add Support for the Bitmain Antminer S3
John Crispin [Tue, 7 Jul 2015 13:46:54 +0000 (13:46 +0000)]
ar71xx: Add Support for the Bitmain Antminer S3

This patch adds support for the Bitmain Antminer S3 Cryptocurrency Miner

http://wiki.openwrt.org/toh/bitmain/s3

Signed-off-by: L. D. Pinney <ldpinney@gmail.com>
SVN-Revision: 46236

9 years agoar71xx: allow pci calibration fixup to work with ar9344
John Crispin [Tue, 7 Jul 2015 13:46:44 +0000 (13:46 +0000)]
ar71xx: allow pci calibration fixup to work with ar9344

Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
SVN-Revision: 46235

9 years agoar71xx: add support for ubnt rocket-m ti
John Crispin [Tue, 7 Jul 2015 13:46:38 +0000 (13:46 +0000)]
ar71xx: add support for ubnt rocket-m ti

rocket-m titanium is a device based on ar9344 with 802.11af poe.
It doesn't use the SoC wifi, instead it's using an ar9280 connected to
the pci bus. The gps version of the rocket-m ti is working, but
gps is untested. The gps is connected to the first serial port.

Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
SVN-Revision: 46234

9 years agox86/xen_domu: enable image generation
John Crispin [Tue, 7 Jul 2015 13:46:26 +0000 (13:46 +0000)]
x86/xen_domu: enable image generation

Add features ext4 targz to target x86/xen_domu in order to
generate images in defconfig.

This fixes #18074.

Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
SVN-Revision: 46233

9 years agoswconfig: swlib.c: free name and description of attributes
John Crispin [Tue, 7 Jul 2015 13:46:16 +0000 (13:46 +0000)]
swconfig: swlib.c: free name and description of attributes

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
SVN-Revision: 46232

9 years agoswconfig: swlib.c: free portmaps in swlib_free()
John Crispin [Tue, 7 Jul 2015 13:46:05 +0000 (13:46 +0000)]
swconfig: swlib.c: free portmaps in swlib_free()

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
SVN-Revision: 46231

9 years agoswconfig: swlib.c: remove const qualifier for val.s since this is supposed to be...
John Crispin [Tue, 7 Jul 2015 13:45:56 +0000 (13:45 +0000)]
swconfig: swlib.c: remove const qualifier for val.s since this is supposed to be free'd

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
SVN-Revision: 46230

9 years agoswconfig: swlib.c: free device name and alias
John Crispin [Tue, 7 Jul 2015 13:45:44 +0000 (13:45 +0000)]
swconfig: swlib.c: free device name and alias

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
SVN-Revision: 46229

9 years agoar71xx: Enable sysupgrade on the Bitmain Antminer S1
John Crispin [Tue, 7 Jul 2015 13:45:37 +0000 (13:45 +0000)]
ar71xx: Enable sysupgrade on the Bitmain Antminer S1

This patch enables sysupgrade for the Bitmain Antminer S1

Signed-off-by: L. D. Pinney <ldpinney@gmail.com>
SVN-Revision: 46228

9 years agoar71xx: Fix WAN port location on EnGenius ESR1750 and EPG5000.
John Crispin [Tue, 7 Jul 2015 13:45:30 +0000 (13:45 +0000)]
ar71xx: Fix WAN port location on EnGenius ESR1750 and EPG5000.

The designated blue WAN port on both devices has index 5, not 1.

Signed-off-by: Christian Beier <cb@shoutrlabs.com>
SVN-Revision: 46227

9 years agoar71xx: add LED defaults for the EnGenius EPG5000
John Crispin [Tue, 7 Jul 2015 13:45:20 +0000 (13:45 +0000)]
ar71xx: add LED defaults for the EnGenius EPG5000

NB that wlan5g is phy0tpt and wlan2g phy1tpt.

Signed-off-by: Christian Beier <cb@shoutrlabs.com>
SVN-Revision: 46226

9 years agouboot-lantiq: Add support for Arcadyan ARV8539PW22 (Speedport W 504V)
John Crispin [Tue, 7 Jul 2015 13:45:10 +0000 (13:45 +0000)]
uboot-lantiq: Add support for Arcadyan ARV8539PW22 (Speedport W 504V)

uboot-lantiq: Add support for Arcadyan ARV8539PW22 (Speedport W 504V)

Signed-off-by: Jannis Pinter <jannis@pinterjann.is>
SVN-Revision: 46225

9 years agolantiq: Add support for Arcadyan ARV8539PW22 (Speedport W 504V)
John Crispin [Tue, 7 Jul 2015 13:44:58 +0000 (13:44 +0000)]
lantiq: Add support for Arcadyan ARV8539PW22 (Speedport W 504V)

lantiq: Add support for Arcadyan ARV8539PW22 (Speedport W 504V)

Signed-off-by: Jannis Pinter <jannis@pinterjann.is>
SVN-Revision: 46224

9 years agolantiq: Add support for the BT Home Hub 5A
John Crispin [Tue, 7 Jul 2015 13:44:53 +0000 (13:44 +0000)]
lantiq: Add support for the BT Home Hub 5A

u-boot support depends on the next "upstream" version ([0]) from
Daniel Schwierzeck.
Since the installation process is quite complicated a "how to" was
added to the wiki: [1]

[0] https://github.com/danielschwierzeck/u-boot-lantiq/tree/openwrt/v2014.01-next
[1] http://wiki.openwrt.org/toh/bt/homehub_v5a

V2: Use the correct PCI interrupt (fixes 2.4GHz wifi)
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46223

9 years agolantiq: Silence the printk when a PCIe error interrupt arrives
John Crispin [Tue, 7 Jul 2015 13:44:37 +0000 (13:44 +0000)]
lantiq: Silence the printk when a PCIe error interrupt arrives

Basically the only error I am seeing is "Correctable Error". Also newer
lantiq PCIe drivers have this message wrapped in a "if debug enabled"
block. So it should be safe to disable this warning.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46222

9 years agolantiq: Make the MEI address available for kernel drivers
John Crispin [Tue, 7 Jul 2015 13:44:32 +0000 (13:44 +0000)]
lantiq: Make the MEI address available for kernel drivers

Newer DSL driver versions depend on the address information.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46221

9 years agogeneric: Allow configuring an increment with mtd-mac-address
John Crispin [Tue, 7 Jul 2015 13:44:27 +0000 (13:44 +0000)]
generic: Allow configuring an increment with mtd-mac-address

This is useful if the device also has an ethernet WAN interface with a
separate mac address (that is derived from the LAN mac address).

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46220

9 years agolantiq: Change the data-type of mac address in ETOP driver
John Crispin [Tue, 7 Jul 2015 13:44:12 +0000 (13:44 +0000)]
lantiq: Change the data-type of mac address in ETOP driver

Make it consistent with the net_device struct and the xrx200 driver

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46219

9 years agolantiq: Make ATM and PTM drivers compatible with kernel 4.1
John Crispin [Tue, 7 Jul 2015 13:44:04 +0000 (13:44 +0000)]
lantiq: Make ATM and PTM drivers compatible with kernel 4.1

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46218

9 years agolantiq: Make ltq-ifxos compile with kernel 4.1
John Crispin [Tue, 7 Jul 2015 13:43:54 +0000 (13:43 +0000)]
lantiq: Make ltq-ifxos compile with kernel 4.1

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46217

9 years agolantiq: Add support for linux 4.1
John Crispin [Tue, 7 Jul 2015 13:43:47 +0000 (13:43 +0000)]
lantiq: Add support for linux 4.1

All (still relevant) patches were refresh.
The following patches were dropped because they are applied upstream:
- 0003-MIPS-lantiq-handle-vmmc-memory-reservation.patch
- 0005-MIPS-lantiq-add-reset-controller-api-support.patch
- 0006-MIPS-lantiq-reboot-gphy-on-restart.patch
- 0009-MIPS-lantiq-command-line-work-around.patch
- 0010-MIPS-lantiq-export-soc-type.patch
- 0011-lantiq-add-support-for-xrx200-firmware-depending-on-.patch
- 0037-MIPS-lantiq-move-eiu-init-after-irq_domain-register.patch

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 46216

9 years agoramips: a5-v11 board fix power led blinking at boot
John Crispin [Tue, 7 Jul 2015 13:43:37 +0000 (13:43 +0000)]
ramips: a5-v11 board fix power led blinking at boot

Signed-off-by: Konrad Kotowicz <skoti1@go2.pl>
SVN-Revision: 46215

9 years agoramips: fix typo in WHR1166D mtd size variable
John Crispin [Tue, 7 Jul 2015 13:43:28 +0000 (13:43 +0000)]
ramips: fix typo in WHR1166D mtd size variable

Fix typo in WHR1166D mtd size variable

Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi83@gmail.com>
SVN-Revision: 46214

9 years agoprocd: update to the latest version, fix stdout logging
Rafał Miłecki [Tue, 7 Jul 2015 10:22:25 +0000 (10:22 +0000)]
procd: update to the latest version, fix stdout logging

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 46212

9 years agodnsmasq: Add sequential_ip UCI parameter
Steven Barth [Tue, 7 Jul 2015 08:13:22 +0000 (08:13 +0000)]
dnsmasq: Add sequential_ip UCI parameter

When enabled the dnsmasq DHCP server allocates the IP addresses sequentially
starting from the lowest available IP address.

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
SVN-Revision: 46211

9 years agodnsmasq: enable extra tracing by default when UCI parameter logqueries is set
Steven Barth [Tue, 7 Jul 2015 08:13:16 +0000 (08:13 +0000)]
dnsmasq: enable extra tracing by default when UCI parameter logqueries is set

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
SVN-Revision: 46210

9 years agomac80211: update regulatory database to master-2015-06-05
Felix Fietkau [Tue, 7 Jul 2015 08:06:10 +0000 (08:06 +0000)]
mac80211: update regulatory database to master-2015-06-05

Signed-off-by: Petko Bordjukov <bordjukov@gmail.com>
SVN-Revision: 46209

9 years agoar71xx: add support for ap143
Felix Fietkau [Tue, 7 Jul 2015 08:06:05 +0000 (08:06 +0000)]
ar71xx: add support for ap143

Signed-off-by: Miaoqing Pan <miaoqing@codeaurora.org>
SVN-Revision: 46208

9 years agoar71xx: rework patch for qca953x/956x
Felix Fietkau [Tue, 7 Jul 2015 08:05:55 +0000 (08:05 +0000)]
ar71xx: rework patch for qca953x/956x

Patch cherry-picked from the following location:
https://www.codeaurora.org/cgit/quic/qsdk/oss/system/openwrt/commit/?h=release/coconut_ioe4531_2.0&id=5c357bf6c763e4140dddcc9a3bc5f005525a9c0e

Changelist,
    - add more register defines
    - add EHCI support
    - fix GPIO pin count to 18
    - fix chained irq disabled
    - fix GMAC0/GMAC1 initial
    - fix WMAC irq number to 47
    - merge the changes of dev-eth.c from the patch to file.

Signed-off-by: Miaoqing Pan <miaoqing@codeaurora.org>
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 46207

9 years agoar71xx: fix ethernet initialization on QCA953x-based boards (TP-Link TL-WR841N/ND...
Felix Fietkau [Tue, 7 Jul 2015 08:05:50 +0000 (08:05 +0000)]
ar71xx: fix ethernet initialization on QCA953x-based boards (TP-Link TL-WR841N/ND v9, Compex WPJ531)

The initialization routines for these boards were relying on some (wrong)
defaults for the QCA953x ethernet. Make these defaults explicit to prevent
breaking them when the QCA953x defaults are fixed.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 46206

9 years agomac80211: fix compile warning in rt2800lib.c
Hauke Mehrtens [Mon, 6 Jul 2015 21:43:48 +0000 (21:43 +0000)]
mac80211: fix compile warning in rt2800lib.c

reorder some parts of the patch to remove the declaration-after-
statement warning.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46205

9 years agomac80211: add kernel 4.1 support again
Hauke Mehrtens [Mon, 6 Jul 2015 21:32:17 +0000 (21:32 +0000)]
mac80211: add kernel 4.1 support again

The new version now works with kernel 4.1

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46204

9 years agomac80211: remove useless patch
Hauke Mehrtens [Mon, 6 Jul 2015 21:31:40 +0000 (21:31 +0000)]
mac80211: remove useless patch

unregister_netdevice_many() was extended in this commit to also delete unreg_list:
https://git.kernel.org/cgit/linux/kernel/git/backports/backports.git/commit/?id=97666f936e0656a1b28c40cc33981a8db65e1a11

This patch is no longer needed.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46203

9 years agomac80211: remove useless patch
Hauke Mehrtens [Mon, 6 Jul 2015 21:31:13 +0000 (21:31 +0000)]
mac80211: remove useless patch

This patch was already applied upstream, linux/device.h gets included 2
lines above this change.

This was done in this commit:
https://git.kernel.org/cgit/linux/kernel/git/backports/backports.git/commit/?id=fd6bb301116d48c50722b0a4399e26ab0237e563

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46202

9 years agomac80211: add missing patch chunk
Felix Fietkau [Mon, 6 Jul 2015 18:41:17 +0000 (18:41 +0000)]
mac80211: add missing patch chunk

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46201

9 years agoar71xx: fix kernel Oops in at803x_link_change_notify
Felix Fietkau [Mon, 6 Jul 2015 17:27:29 +0000 (17:27 +0000)]
ar71xx: fix kernel Oops in at803x_link_change_notify

r45954 ("ar71xx: fix 100/10mbps ethernet link issues on mynet range
extender") introduced a pdata based modification of the tx_clk_dly. But it
was not checked if pdata actually existed. This caused a page fault on all
devices which didn't have at803x_platform_data specified for an at803x
based device.

Signed-off-by: Sven Eckelmann <sven@open-mesh.com>
SVN-Revision: 46200

9 years agoubus: update to the latest version, fixes a segfault on message retry
Felix Fietkau [Mon, 6 Jul 2015 16:44:45 +0000 (16:44 +0000)]
ubus: update to the latest version, fixes a segfault on message retry

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46199

9 years agomac80211: update to version 2015-06-22
Felix Fietkau [Mon, 6 Jul 2015 16:26:34 +0000 (16:26 +0000)]
mac80211: update to version 2015-06-22

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46198

9 years agoath10k: remove support for the obsolete STA firmware
Felix Fietkau [Mon, 6 Jul 2015 10:21:26 +0000 (10:21 +0000)]
ath10k: remove support for the obsolete STA firmware

Current firmware should work for station mode as well, and the older
firmware will usually behave worse and become unsupported over time.
Also, the MAC address handling on ar71xx is only implemented on one
firmware version.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46197

9 years agoar71xx: update preinit script for new ath10k firmware
Felix Fietkau [Mon, 6 Jul 2015 10:19:21 +0000 (10:19 +0000)]
ar71xx: update preinit script for new ath10k firmware

Update firmware name and md5sum. MAC-address location remains
unchanged from previous version.

Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 46196

9 years agomac80211: Update QCA9880 firmware to 10.2.4.70-2
Felix Fietkau [Mon, 6 Jul 2015 10:19:08 +0000 (10:19 +0000)]
mac80211: Update QCA9880 firmware to 10.2.4.70-2

This firmware adds support for management frames over normal
packet transport interface instead of over the WMI management
interface.

Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 46195

9 years agomac80211: Backport ath10k firmware API v5 support
Felix Fietkau [Mon, 6 Jul 2015 10:19:02 +0000 (10:19 +0000)]
mac80211: Backport ath10k firmware API v5 support

These patches add support for the new version 5 firmware API.

Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 46194

9 years agoar71xx: fix UAP-PRO images
Felix Fietkau [Mon, 6 Jul 2015 10:18:50 +0000 (10:18 +0000)]
ar71xx: fix UAP-PRO images

- it should be BOARDNAME instead of BOARD_NAME
- kernel partition should be padded to 1536k (somehow this padding was lost in translation to BuildCode in bc797c73f6328941b2194d144385655ad9297e7c)

Signed-off-by: Roman Yeryomin <roman@advem.lv>
SVN-Revision: 46193

9 years agouclibc++: link libssp_nonshared only for musl
Steven Barth [Mon, 6 Jul 2015 08:55:28 +0000 (08:55 +0000)]
uclibc++: link libssp_nonshared only for musl

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 46192

9 years agoar71xx: microtik: disable all non-microtik boards
Jonas Gorski [Sun, 5 Jul 2015 21:33:31 +0000 (21:33 +0000)]
ar71xx: microtik: disable all non-microtik boards

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46191

9 years agoar71xx: EPG5000 requires ATH79_NVRAM
Jonas Gorski [Sun, 5 Jul 2015 21:33:28 +0000 (21:33 +0000)]
ar71xx: EPG5000 requires ATH79_NVRAM

EPG5000 uses ath79 nvram functions, but does not select it. This breaks
build if no other board support enabling nvram support is selected.

Closes #19952.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46190

9 years agoixp4xx: add linux 4.1 support
Jonas Gorski [Sun, 5 Jul 2015 21:33:21 +0000 (21:33 +0000)]
ixp4xx: add linux 4.1 support

Compile tested only.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46189

9 years agokernel: add a few more missing config symbols
Jonas Gorski [Sun, 5 Jul 2015 21:32:55 +0000 (21:32 +0000)]
kernel: add a few more missing config symbols

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46188

9 years agokernel: generic: add missing arm machtypes
Jonas Gorski [Sun, 5 Jul 2015 21:32:49 +0000 (21:32 +0000)]
kernel: generic: add missing arm machtypes

Some machtypes were dropped when 4.0 support was added, and the
incomplete patch was taken over to 4.1.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46187

9 years agoar7: add linux 4.1 support
Jonas Gorski [Sun, 5 Jul 2015 21:32:44 +0000 (21:32 +0000)]
ar7: add linux 4.1 support

Compile tested only.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46186

9 years agopackage: ar7-atm: depend on !LINUX_4_1
Jonas Gorski [Sun, 5 Jul 2015 21:31:57 +0000 (21:31 +0000)]
package: ar7-atm: depend on !LINUX_4_1

ar7-atm will require modifications to build on 4.1.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46185

9 years agokernel: mips_disable_fpu: initialize fault_addr in fpu_emulator_cop1Handler
Jonas Gorski [Sun, 5 Jul 2015 21:31:54 +0000 (21:31 +0000)]
kernel: mips_disable_fpu: initialize fault_addr in fpu_emulator_cop1Handler

Code calling fpu_emulator_cop1Handler will pass on fault_addr, making gcc
complain about it not being initialized when the FPU emulator is disabled.

Fixes:

arch/mips/kernel/traps.c: In function 'do_fpe':
arch/mips/kernel/traps.c:864:22: error: 'fault_addr' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  process_fpemu_return(sig, fault_addr, fcr31);
                      ^
arch/mips/kernel/traps.c: In function 'do_ri':
arch/mips/kernel/traps.c:806:22: error: 'fault_addr' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  process_fpemu_return(sig, fault_addr, fcr31);
                      ^
arch/mips/kernel/traps.c:763:15: note: 'fault_addr' was declared here
  void __user *fault_addr;
               ^
arch/mips/kernel/traps.c: In function 'do_cpu':
arch/mips/kernel/traps.c:1421:28: error: 'fault_addr' may be used uninitialized in this function [-Werror=maybe-uninitialized]
   if (!process_fpemu_return(sig, fault_addr, fcr31) && !err)
                            ^
cc1: all warnings being treated as errors
make[7]: *** [arch/mips/kernel/traps.o] Error 1

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46184

9 years agobrcm63xx: use IRQCHIP_DECLARE instead of self-baked version
Jonas Gorski [Sun, 5 Jul 2015 21:31:52 +0000 (21:31 +0000)]
brcm63xx: use IRQCHIP_DECLARE instead of self-baked version

Backport a commit from 4.2 making IRQCHIP_DECLARE available outside of
drivers/irqchip and use it.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46183

9 years agokernel: assume modules.builtin is always present
Jonas Gorski [Sun, 5 Jul 2015 21:31:34 +0000 (21:31 +0000)]
kernel: assume modules.builtin is always present

We do not support old kernel versions not generating modules.builtin
anymore, so assume it will always be present and check for modules
to be built in first.

This prevents old modules being packages up after changing the kernel
config to include them in the kernel without cleaning the kernel tree.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46182

9 years agobrcm63xx: 4.1: switch to upstream UHI interface
Jonas Gorski [Sun, 5 Jul 2015 11:28:46 +0000 (11:28 +0000)]
brcm63xx: 4.1: switch to upstream UHI interface

Replace the appended dtb patch with the upstream accepted version and
update the code to work with the changed interface.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46181

9 years agokernel: add crypto-cmac
Hauke Mehrtens [Sat, 4 Jul 2015 22:33:00 +0000 (22:33 +0000)]
kernel: add crypto-cmac

The bluetooth driver needs this new crypto lib in kernel 4.1

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46180

9 years agokernel: fix loading bluetooth_6lowpan
Hauke Mehrtens [Sat, 4 Jul 2015 22:32:04 +0000 (22:32 +0000)]
kernel: fix loading bluetooth_6lowpan

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46179

9 years agokernel: kmod-scsi-core: fix load on install
Jonas Gorski [Sat, 4 Jul 2015 12:52:57 +0000 (12:52 +0000)]
kernel: kmod-scsi-core: fix load on install

sd_mod depends on scsi_mod, but due to it being an AutoLoad and not
AutoProbe module, it was not loading when installing the package,
causing unknown symbol errors for sd_mod and anything depending on it.

Closes #14927, #18293, #19351.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46176

9 years agoscons: change md5sum to new release
Hauke Mehrtens [Sat, 4 Jul 2015 08:54:20 +0000 (08:54 +0000)]
scons: change md5sum to new release

The developers of scons found a bug in the packaging of their release
and just replaced the already released version with a new one without
changing the version number. This breaks the download script because it
also has a new md5sum.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46175

9 years agoat91: add Device Tree based board detection
Florian Fainelli [Sat, 4 Jul 2015 01:25:15 +0000 (01:25 +0000)]
at91: add Device Tree based board detection

Set up eth0 as LAN interface by default.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
SVN-Revision: 46174

9 years agoat91: sama5d3: enable USB2 by default
Florian Fainelli [Sat, 4 Jul 2015 01:24:48 +0000 (01:24 +0000)]
at91: sama5d3: enable USB2 by default

Signed-off-by: Florian Fainelli <florian@openwrt.org>
SVN-Revision: 46173

9 years agokernel: fix OHCI and EHCI modules for AT91
Florian Fainelli [Sat, 4 Jul 2015 01:24:37 +0000 (01:24 +0000)]
kernel: fix OHCI and EHCI modules for AT91

We were not copying the additional files needed to get USB to work on
the AT91 platforms, make sure we do this.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
SVN-Revision: 46172

9 years agoat91: enable Micrel PHY driver
Florian Fainelli [Sat, 4 Jul 2015 01:24:27 +0000 (01:24 +0000)]
at91: enable Micrel PHY driver

The AT91 SAMA5D3XPLAINED uses a KSZ9031 and a KSZ8081 that are supported
by this PHY driver.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
SVN-Revision: 46171

9 years agokernel: b53: fix build with brcm47xx
Hauke Mehrtens [Fri, 3 Jul 2015 23:27:21 +0000 (23:27 +0000)]
kernel: b53: fix build with brcm47xx

The position of the nvram header file on brcm47xx changed with kernel
version 4.1.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46170

9 years agocurl: update curl to version 7.43.0
Hauke Mehrtens [Fri, 3 Jul 2015 23:21:01 +0000 (23:21 +0000)]
curl: update curl to version 7.43.0

This brings curl to version 7.43.0 and contains fixes for the following
security vulnerabilities:

CVE-2015-3236: lingering HTTP credentials in connection re-use
http://curl.haxx.se/docs/adv_20150617A.html

CVE-2015-3237: SMB send off unrelated memory contents
http://curl.haxx.se/docs/adv_20150617B.html

The 100-check_long_long patch is not needed any more, because the
upstream autoconf script already checks for long long when cyassl is
selected.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46169

9 years agocyassl: update to wolfssl 3.6.0
Hauke Mehrtens [Fri, 3 Jul 2015 23:20:36 +0000 (23:20 +0000)]
cyassl: update to wolfssl 3.6.0

Upstream wolfssl already has better checks to detect broken ssl v2
ClientHellos, we can remove our hack.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46168

9 years agocyassl: version bump to 3.4.6
Hauke Mehrtens [Fri, 3 Jul 2015 23:20:01 +0000 (23:20 +0000)]
cyassl: version bump to 3.4.6

This patch introduces a new build error into coova-chilli, but
coova-chilli already fails to build even without it anyway. CyaSSL is
now called wolfSSL, and all the API's have been renamed, and
backward-compatibility headers added.

Signed-off-by: Daniel Gimpelevich <daniel@gimpelevich.san-francisco.ca.us>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46167

9 years agobrcm47xx: add support for kernel 4.1
Hauke Mehrtens [Fri, 3 Jul 2015 23:18:33 +0000 (23:18 +0000)]
brcm47xx: add support for kernel 4.1

This adds support for kernel 4.1 and removes kernel 4.0.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46166

9 years agoar71xx: make 902-unaligned_access_hacks.patch apply again
Jonas Gorski [Fri, 3 Jul 2015 16:47:44 +0000 (16:47 +0000)]
ar71xx: make 902-unaligned_access_hacks.patch apply again

r46161 updated headers also touched by it, so it needed to be refreshed
as well.

Closes #20033.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46165

9 years agouclibc++: make g++-uc* wrappers relocatable
Jo-Philipp Wich [Fri, 3 Jul 2015 13:33:05 +0000 (13:33 +0000)]
uclibc++: make g++-uc* wrappers relocatable

The g++-uc wrapper hardcodes $(STAGING_DIR) and $(TOOLCHAIN_DIR) paths which
will not work outside of the original build environment.

Replace the hardcoded staging_dir occurences with paths relative to the
$STAGING_DIR environment variable to make the g++-uc* wrappers usable in an
SDK environment.

Fixes the libdb47 build failure reported at
  https://lists.openwrt.org/pipermail/openwrt-devel/2015-April/032455.html

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 46162

9 years agokernel: improve uapi headers coexistence with musl
Jonas Gorski [Fri, 3 Jul 2015 11:48:45 +0000 (11:48 +0000)]
kernel: improve uapi headers coexistence with musl

Make some network uapi headers detect if they are included after
not only glibc but also musl headers.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46161

9 years agonf-conntrack: enable netlink events
Steven Barth [Thu, 2 Jul 2015 14:11:15 +0000 (14:11 +0000)]
nf-conntrack: enable netlink events

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 46160

9 years agoath9k: make DMA stop related messages debug-only (#11862, #9654)
Felix Fietkau [Thu, 2 Jul 2015 11:45:08 +0000 (11:45 +0000)]
ath9k: make DMA stop related messages debug-only (#11862, #9654)

A long time ago, ath9k had issues during reset where the DMA engine
would stay active and could potentially corrupt memory.
To debug those issues, the driver would print warnings whenever they
occur.

Nowadays, these issues are gone and the primary cause of these messages
is if the MAC is stuck during reset or busy processing a long
transmission. This is fairly harmless, yet these messages continue to
worry users.

To reduce the number of bogus bug reports, turn these messages into
debug messages and count their occurence in the "reset" debugfs file.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46158

9 years agomdns: fix ubus wait_for command
Felix Fietkau [Thu, 2 Jul 2015 11:26:12 +0000 (11:26 +0000)]
mdns: fix ubus wait_for command

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 46156