openwrt/staging/kaloz.git
9 years agoinclude: clean package staging dir files before configure
Jo-Philipp Wich [Tue, 26 May 2015 09:17:13 +0000 (09:17 +0000)]
include: clean package staging dir files before configure

This is required to allow packages to ship libraries with a different API.
Previously the linker would favor the old shared libraries in STAGING_DIR
instead of the just compiled ones in PKG_BUILD_DIR.

Avoid that problem by clearing the package staging files right after
prepare.

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 45759

9 years agoiptables: reduce binary size
Jo-Philipp Wich [Tue, 26 May 2015 09:16:50 +0000 (09:16 +0000)]
iptables: reduce binary size

 * drop unused lenient restore patch
 * instead of statically linking core extensions, build shared libraries
   for reuse in fw3
 * strip outdated match revisions and aliases to trim down library size

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 45758

9 years agobuild: allow creating empty package feeds
Felix Fietkau [Tue, 26 May 2015 09:09:07 +0000 (09:09 +0000)]
build: allow creating empty package feeds

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45757

9 years agoar71xx: fix tl-wr710n images (#19692)
Felix Fietkau [Tue, 26 May 2015 07:13:40 +0000 (07:13 +0000)]
ar71xx: fix tl-wr710n images (#19692)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45756

9 years agoppp: do not warn if connect() before close() on pppoe terminate fails (fixes #19651)
Felix Fietkau [Tue, 26 May 2015 07:02:49 +0000 (07:02 +0000)]
ppp: do not warn if connect() before close() on pppoe terminate fails (fixes #19651)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45755

9 years agomusl: update to version 1.1.9
Felix Fietkau [Mon, 25 May 2015 23:11:01 +0000 (23:11 +0000)]
musl: update to version 1.1.9

Patch to 2015-05-25 to fix ldso related regressions on PPC and MIPS.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
SVN-Revision: 45754

9 years agokirkwood: disable the bogus CONFIG_IP_ADVANCED_ROUTER override (#19722)
Felix Fietkau [Mon, 25 May 2015 22:59:15 +0000 (22:59 +0000)]
kirkwood: disable the bogus CONFIG_IP_ADVANCED_ROUTER override (#19722)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45753

9 years agouClibc: elf: Add STT_GNU_IFUNC from glibc
Felix Fietkau [Mon, 25 May 2015 22:59:04 +0000 (22:59 +0000)]
uClibc: elf: Add STT_GNU_IFUNC from glibc

perf in upstream Linux kernel 3.17 onwards expects STT_GNU_IFUNC
replicate it from glibc

Patch cherry-picked from:
http://lists.busybox.net/pipermail/uclibc/2015-February/048825.html

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45752

9 years agofstools: update to the latest version, makes filesystem initialization state tracking...
Felix Fietkau [Mon, 25 May 2015 22:00:06 +0000 (22:00 +0000)]
fstools: update to the latest version, makes filesystem initialization state tracking more reliable (#19564)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45751

9 years agoubus: update to the latest version, switch to git.openwrt.org, fixes a lua binding...
Felix Fietkau [Mon, 25 May 2015 21:15:42 +0000 (21:15 +0000)]
ubus: update to the latest version, switch to git.openwrt.org, fixes a lua binding memory leak

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45750

9 years agompc85xx: TL-WDR4900: Fix port 6 being shown as up (10MBit/half) in LUCI/swconfig
Felix Fietkau [Mon, 25 May 2015 21:15:37 +0000 (21:15 +0000)]
mpc85xx: TL-WDR4900: Fix port 6 being shown as up (10MBit/half) in LUCI/swconfig

Currently port 6 is shown as up 10MBit/half in LUCI and swconfig.
Reason is that all bits in the port 6 config are zero.
This means that also the aneg flag is not set and in this case
ar8216_read_port_link hardcodes the link to be up.

This is no real problem but a little annoying.
To fix this initialize port 6 with the aneg bit enabled.
This causes ar8216_read_port_link to evaluate the link status bit which is
always zero for port 6 as no PHY is connected to this port.
And it doesn't hurt as port 6 isn't connected to anything on TL-WDR4900.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
SVN-Revision: 45749

9 years agonetifd: bump to current HEAD
Felix Fietkau [Mon, 25 May 2015 21:15:31 +0000 (21:15 +0000)]
netifd: bump to current HEAD

This introduces a new config parameter "no-proto-task" useful for
xl2tpd.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
SVN-Revision: 45748

9 years agoramips: fix hang on boot on rt305x
Jonas Gorski [Mon, 25 May 2015 18:02:32 +0000 (18:02 +0000)]
ramips: fix hang on boot on rt305x

Fix an upstream regression on 3.18+ that causes rt305x based devices
to hang on boot.

Fixes #19715.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45747

9 years agoimage.mk: make image too big warnings more visible
Jonas Gorski [Mon, 25 May 2015 08:27:55 +0000 (08:27 +0000)]
image.mk: make image too big warnings more visible

Output warnings through stderr to allow them to be easilier spotted
when building with V=w.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45746

9 years agogdb: disable liblzma explictly
Nicolas Thill [Sun, 24 May 2015 19:15:54 +0000 (19:15 +0000)]
gdb: disable liblzma explictly

 * fix build failure when gdb is build after liblzma
 * bump release number

Signed-off-by: Nicolas Thill <nico@openwrt.org>
SVN-Revision: 45745

9 years agofuse: fix exec environment for mount and umount
Nicolas Thill [Sun, 24 May 2015 17:56:26 +0000 (17:56 +0000)]
fuse: fix exec environment for mount and umount

 * add upstream patch (CVE-2015-3202)
 * refresh patches
 * bump release number

Signed-off-by: Nicolas Thill <nico@openwrt.org>
SVN-Revision: 45744

9 years agopackage/kernel: remove reference to ep93xx target
Nicolas Thill [Sun, 24 May 2015 01:55:14 +0000 (01:55 +0000)]
package/kernel: remove reference to ep93xx target

Signed-off-by: Nicolas Thill <nico@openwrt.org>
SVN-Revision: 45743

9 years agopackage/kernel: remove reference to 3.12 kernel version
Nicolas Thill [Sun, 24 May 2015 01:48:52 +0000 (01:48 +0000)]
package/kernel: remove reference to 3.12 kernel version

Signed-off-by: Nicolas Thill <nico@openwrt.org>
SVN-Revision: 45742

9 years agocomgt: the package contained too many files
John Crispin [Sat, 23 May 2015 15:29:46 +0000 (15:29 +0000)]
comgt: the package contained too many files

fixes #19698

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 45741

9 years agofuse: fix build on arm64
John Crispin [Sat, 23 May 2015 15:29:31 +0000 (15:29 +0000)]
fuse: fix build on arm64

error:
../include/fuse_kernel.h:93:15: error: conflicting types for 'int64_t'
../include/fuse_kernel.h:92:15: error: conflicting types for 'uint64_t'

reported by buildbot

fix by adopting upstream patch:
http://git.baserock.org/cgi-bin/cgit.cgi/delta/fuse.git/commit/include/fuse_kernel.h?id=914871b20a901e3e1e981c92bc42b1c93b7ab81b

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 45740

9 years agow1-gpio-custom: Fix uninitialised variable causing 1-wire to not bind to GPIO
John Crispin [Sat, 23 May 2015 15:29:20 +0000 (15:29 +0000)]
w1-gpio-custom: Fix uninitialised variable causing 1-wire to not bind to GPIO

Signed-off-by: Andrew McDonnell <bugs@andrewmcdonnell.net>
SVN-Revision: 45739

9 years agoscripts/feeds: Delete package/feeds folder in the clean procedure to prevent dangling...
John Crispin [Sat, 23 May 2015 15:29:06 +0000 (15:29 +0000)]
scripts/feeds: Delete package/feeds folder in the clean procedure to prevent dangling links.

Hello,

if you run:
./scripts/feeds clean
It removes ./feeds folder but not ./package/feeds/ which is full of dangling links then. This patch fixes it.

Best Regards,
Martin Strbačka

Signed-off-by: Martin Strbacka <martin.strbacka@nic.cz>
SVN-Revision: 45738

9 years agoconntrack-tools: add init script for conntrackd
John Crispin [Sat, 23 May 2015 15:28:56 +0000 (15:28 +0000)]
conntrack-tools: add init script for conntrackd

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
SVN-Revision: 45737

9 years agoMakefile: remove non-existent STAGING_DIR_TOOLCHAIN from dirclean
John Crispin [Sat, 23 May 2015 15:28:42 +0000 (15:28 +0000)]
Makefile: remove non-existent STAGING_DIR_TOOLCHAIN from dirclean

Makefile: remove non-existent STAGING_DIR_TOOLCHAIN from dirclean

Openwrt's top level Makefile uses STAGING_DIR_TOOLCHAIN in the make dirclean
statement.
https://dev.openwrt.org/browser/trunk/Makefile#L55
    rm -rf $(STAGING_DIR) $(STAGING_DIR_HOST) $(STAGING_DIR_TOOLCHAIN)
$(TOOLCHAIN_DIR) $(BUILD_DIR_HOST) $(BUILD_DIR_TOOLCHAIN)

As far as I can determine, no such variable has been defined. I made a search
in Openwrt source repository and the one line in Makefile's dirclean command
is the only place where that variable exists.

The item has been introduced to Makefile by r8362, but even at that time
neither Makefile nor rules.mk defined such a variable. Most likely the goal
has been to set both staging_dir/toolchain and build_dir/toolchain to be
cleaned, but one of the variables has been erroneous. The correct variable
for build_dir/toolchain has been then added by r13494.

References:
https://dev.openwrt.org/changeset/8362/
https://dev.openwrt.org/browser/trunk/Makefile?rev=8362
https://dev.openwrt.org/browser/trunk/rules.mk?rev=8362
https://lists.openwrt.org/pipermail/openwrt-devel/2007-August/001159.html
https://dev.openwrt.org/changeset/13494

In current code,
    TOOLCHAIN_DIR = $(TOPDIR)/staging_dir/$(TOOLCHAIN_DIR_NAME)
    BUILD_DIR_TOOLCHAIN = $(TOPDIR)/build_dir/$(TOOLCHAIN_DIR_NAME)
so the item STAGING_DIR_TOOLCHAIN in the rm command is unnecessary.

signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>

SVN-Revision: 45736

9 years agomodules/sound: cleanup by removing unsupported Kernel cases
John Crispin [Sat, 23 May 2015 15:28:36 +0000 (15:28 +0000)]
modules/sound: cleanup by removing unsupported Kernel cases

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 45735

9 years agomodules/usb: another cleanup: remove ancient module names
John Crispin [Sat, 23 May 2015 15:28:32 +0000 (15:28 +0000)]
modules/usb: another cleanup: remove ancient module names

modules relating to CONFIG_USB_AUDIO
Kernel <2.6.35 is not supported in trunk

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 45734

9 years agomodules/usb: cleanup remove old module names
John Crispin [Sat, 23 May 2015 15:28:26 +0000 (15:28 +0000)]
modules/usb: cleanup remove old module names

delete cases for Kernel <3.10 (nop-usb-xceiv), <3.12 (phy-nop)

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 45733

9 years agoocteon: build squashfs image for UBNT ERLITE
John Crispin [Sat, 23 May 2015 15:28:20 +0000 (15:28 +0000)]
octeon: build squashfs image for UBNT ERLITE

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
SVN-Revision: 45732

9 years agoipq806x: enable cpu idle support in kernel conf
John Crispin [Sat, 23 May 2015 15:28:11 +0000 (15:28 +0000)]
ipq806x: enable cpu idle support in kernel conf

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45731

9 years agoipq806x: add & enable cpufreq support
John Crispin [Sat, 23 May 2015 15:28:02 +0000 (15:28 +0000)]
ipq806x: add & enable cpufreq support

This change set enables frequency scaling on ipq806x, which speeds-up
the CPU and allows it to achieve its max frequency.

These patches are cherry-picked & backported from the following location:
*130-132: linux-next
*133-143: LKML - https://lkml.org/lkml/2015/3/21/15
*144: derived from other qcom similar dts
*145: derived from https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-3.14/drivers/cpufreq/cpufreq-krait.c

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45730

9 years agoipq806x: add power regulators support
John Crispin [Sat, 23 May 2015 15:27:54 +0000 (15:27 +0000)]
ipq806x: add power regulators support

Patches are cherry-picked from linux-next. We're also adding the
corresponding config option to the kernel.

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45729

9 years agoipq806x: disable i2c device on gsbi4
John Crispin [Sat, 23 May 2015 15:27:45 +0000 (15:27 +0000)]
ipq806x: disable i2c device on gsbi4

Patch cherry-picked from the following location:
https://chromium-review.googlesource.com/#/c/269931/

Disable the i2c device on gsbi4 and mark gsbi4_h and gsbi4_qup clks as
unused. If they are enabled, clock framework will turn them off at end
of probe. On ipq806x by design gsbi4_qup, gsbi4_h clks and i2c on gsbi4
are meant for RPM usage. So turning them off in kernel is incorrect.

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45728

9 years agoipq806x: refresh kernel patches
John Crispin [Sat, 23 May 2015 15:27:28 +0000 (15:27 +0000)]
ipq806x: refresh kernel patches

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45727

9 years agoar71xx: Add support for the Meraki MR12 & MR16
John Crispin [Sat, 23 May 2015 15:27:22 +0000 (15:27 +0000)]
ar71xx: Add support for the Meraki MR12 & MR16

This patch is to add support for the Meraki MR12 and MR16 Access Points.
Currently everything is working, minus the 2nd NIC interface on the MR12
which is built into the SoC.

Signed-off-by: Chris R Blake <chrisrblake93 at gmail.com>
SVN-Revision: 45726

9 years agorampis: add support for status led on ASUS RT-N14U
John Crispin [Sat, 23 May 2015 15:27:10 +0000 (15:27 +0000)]
rampis: add support for status led on ASUS RT-N14U

The diag.sh script lacked an entry for the status led on the RT-N14U,
map it to the asus:blue:power led which is also used by the boot loader
to report boot status (eg. TFTP recovery mode VS normal boot)

Signed-off-by: Matteo Panella <m.panella@level28.org>
SVN-Revision: 45725

9 years agoramips: Adding support for AXIMCom MR-102N 3G/4G Travel Mobile Router
John Crispin [Sat, 23 May 2015 15:27:05 +0000 (15:27 +0000)]
ramips: Adding support for AXIMCom MR-102N 3G/4G Travel Mobile Router

MR-102N is a RT3050F based wireless router(32M RAM + 8M NOR flash) with 1 USB
and 1 ethernet port.  The original product information can be found at:
http://www.aximcom.com/en/MR-102N

Signed-off-by: Tai-hwa Liang <atliang@gmail.com>

SVN-Revision: 45724

9 years agoramips: fix switch names for several boards in device tree
John Crispin [Sat, 23 May 2015 15:26:51 +0000 (15:26 +0000)]
ramips: fix switch names for several boards in device tree

Signed-off-by: Roman Yeryomin <roman@advem.lv>
SVN-Revision: 45723

9 years agoramips: add support for Comfast CF-WR800N
John Crispin [Sat, 23 May 2015 15:26:39 +0000 (15:26 +0000)]
ramips: add support for Comfast CF-WR800N

This patch adds support for Comfast CF-WR800N, a wall-plug wireless router
based on the MT7620N SoC with one Ethernet port and a 802.11n 2.4 GHz radio.

Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
SVN-Revision: 45722

9 years agoubox: various logd cleanups
John Crispin [Sat, 23 May 2015 15:26:29 +0000 (15:26 +0000)]
ubox: various logd cleanups

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 45721

9 years agouClibc: add signal info for seccomp related SIGSYS
John Crispin [Sat, 23 May 2015 15:26:19 +0000 (15:26 +0000)]
uClibc: add signal info for seccomp related SIGSYS

uClibc doesn't define signal info for the SIGSYS signal which is issued
in case of hitting a syscall prohibited by seccomp.
This is sad as it makes debugging seccomp filter policies impossible on
some architectures (at least ARM and PowerPC, maybe also others) which
do not coincidentally set si_value.sival_int as the syscall number.

To fix this, import the definitions and macros needed from glibc.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
SVN-Revision: 45720

9 years agoprocd: make the jail code also work on arm
John Crispin [Sat, 23 May 2015 15:26:09 +0000 (15:26 +0000)]
procd: make the jail code also work on arm

Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 45719

9 years agolantiq: Fix initialization of the PCI IRQs when PCIe is also enabled
John Crispin [Sat, 23 May 2015 15:25:59 +0000 (15:25 +0000)]
lantiq: Fix initialization of the PCI IRQs when PCIe is also enabled

There are already ifx_pcie_bios_{map_irq,plat_dev_init} hooks defined in
ifxmips_pcie.c. Instead of defining a new hook we simply re-use the
existing ones (this is basically what the lantiq BSP code does).

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 45718

9 years agolantiq: Fix PCIe bus when PCI is also enabled.
John Crispin [Sat, 23 May 2015 15:25:51 +0000 (15:25 +0000)]
lantiq: Fix PCIe bus when PCI is also enabled.

The PCIe bus seems to require a hack/workaround when PCI is enabled as
well. Unfortunately this is guarded by an CONFIG_IFX_PCI ifdef, which is
only defined in lantiq's BSP code. The config symbol for the upstream
lantiq PCI driver is CONFIG_PCI_LANTIQ.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
SVN-Revision: 45717

9 years agobcm53xx: add OHCI (USB 1.1) support
Hauke Mehrtens [Sat, 23 May 2015 14:54:44 +0000 (14:54 +0000)]
bcm53xx: add OHCI (USB 1.1) support

There is also a OHCI controller, activate it for USB 1.1 support.

This should close #19601.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 45716

9 years agoarm64: enable new errata backported to 3.18
Jonas Gorski [Fri, 22 May 2015 22:51:37 +0000 (22:51 +0000)]
arm64: enable new errata backported to 3.18

3.18.13 introduced a bunch of new errata, enable them to be on the
safe side.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45715

9 years agokernel: add missing generic config symbol
Jonas Gorski [Fri, 22 May 2015 22:51:34 +0000 (22:51 +0000)]
kernel: add missing generic config symbol

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45714

9 years agotools: gengetopt: Disable doc and tests builds
Florian Fainelli [Fri, 22 May 2015 22:33:41 +0000 (22:33 +0000)]
tools: gengetopt: Disable doc and tests builds

Simply disable the build of the "doc" and "tests" directories to speed up the
build and avoid the potential for errors coming from external tools being used
(e.g: source-highlight installed on a non-standard location).

Signed-off-by: Florian Fainelli <florian@openwrt.org>
SVN-Revision: 45713

9 years agods-lite: fix resolve retry for fqdn peeraddrs
Jonas Gorski [Fri, 22 May 2015 10:24:30 +0000 (10:24 +0000)]
ds-lite: fix resolve retry for fqdn peeraddrs

If the first resolveip call will fail, peeraddr will be now empty, and
the subsequent resolveip call will try to resolve an empty string.

Fix this by storing the result in a temporary variable.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45712

9 years agokernel: update 3.18 to 3.18.14
Jonas Gorski [Thu, 21 May 2015 19:32:46 +0000 (19:32 +0000)]
kernel: update 3.18 to 3.18.14

Changelogs:

* https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.12
* https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.13
* https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.14

Build tested on brcm63xx and ipq806x, runtested on brcm63xx.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45711

9 years agomac80211: make it work with 3.18.12+
Jonas Gorski [Thu, 21 May 2015 19:32:16 +0000 (19:32 +0000)]
mac80211: make it work with 3.18.12+

3.18.12 backported 61ada528dea028331e99e8ceaed87c683ad25de2 ("sched/wait:
Provide infrastructure to deal with nested blocking") from 3.19, causing
the following error on load:

[   13.588000] compat: exports duplicate symbol woken_wake_function (owned by kernel)

Fix this by guarding it with a check for 3.18.11 or earlier instead of
3.19.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45710

9 years agobrcm63xx: fix failsafe lan and status led for DSL-274xB rev. F1
Jonas Gorski [Thu, 21 May 2015 19:32:00 +0000 (19:32 +0000)]
brcm63xx: fix failsafe lan and status led for DSL-274xB rev. F1

Add it to the appropriate places so the power led properly works
and ethernet is properly configured for failsafe.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45709

9 years agobrcm63xx: Add Huawei HG655b support
Jonas Gorski [Thu, 21 May 2015 19:31:54 +0000 (19:31 +0000)]
brcm63xx: Add Huawei HG655b support

This patch adds support for the Huawei HG655b.

Nothing much special in this router, it's just another BCM6368 with
a Ralink RT3062 wifi chip and the calibration data embedded in the
main flash chip at offset 0x7c0000. There is also configuration data
used by the OEM firmware before the cal_data partition, this area is
protected by the board_data partition in this patch.

Signed-off-by: Daniel Gonzalez <dgcbueu@gmail.com>
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45708

9 years agoodhcpd: fix DHCPv6 downstream PD
Steven Barth [Thu, 21 May 2015 15:07:54 +0000 (15:07 +0000)]
odhcpd: fix DHCPv6 downstream PD

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 45707

9 years agogeneric/4.0: fix error during kernel patch application
Steven Barth [Wed, 20 May 2015 19:23:33 +0000 (19:23 +0000)]
generic/4.0: fix error during kernel patch application

Commit 5168c9a5702648eb690d32ec821647aca80aeba9 introduced a regression
during patch application on the 4.0 kernel. Some of the patched content
doesn't match the actual code, therefore leading to the following error:
  Applying patch generic/667-ipv6-Fixed-source-specific-default-route-handling.patch
  patching file net/ipv6/ip6_output.c
  Hunk #1 FAILED at 886.
  1 out of 1 hunk FAILED -- rejects in file net/ipv6/ip6_output.c
  patching file net/ipv6/route.c
  Hunk #1 succeeded at 2247 (offset 2 lines).
  Patch generic/667-ipv6-Fixed-source-specific-default-route-handling.patch does not apply (enforce with -f)

This change just adapts the actual patch to fix what is in kernel 4.0
and make it apply cleanly.

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45705

9 years agomvebu: supress useless in-cpu rtc init on the Mamba
Imre Kaloz [Tue, 19 May 2015 22:54:05 +0000 (22:54 +0000)]
mvebu: supress useless in-cpu rtc init on the Mamba

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 45704

9 years agotools/mkimage: backport gcc5 compatibility, patch
Hauke Mehrtens [Tue, 19 May 2015 22:17:16 +0000 (22:17 +0000)]
tools/mkimage: backport gcc5 compatibility, patch

Fedora 22 includes gcc5 by default
build leads to error:
include/linux/compiler-gcc.h:114:30: fatal error: linux/compiler-gcc5.h: No such file or directory

backport upstream patch: "Add linux/compiler-gcc5.h to fix builds with gcc5"
id: 478b02f1a7043b673565075ea5016376f3293b23

Issue was reported at
https://forum.openwrt.org/viewtopic.php?id=56853
& replicated in a VM

created, flashed and tested ar71xx image successfully
as test

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 45703

9 years agobase-files: fix default pre-/postrm for packages with a dot
Jonas Gorski [Tue, 19 May 2015 09:13:17 +0000 (09:13 +0000)]
base-files: fix default pre-/postrm for packages with a dot

Instead of stripping everything after the first dot, strip everything
after the last dot.
This fixes pre-/postrm actions for packages with a dot in their name,
like libusb-1.0.

Fixes #19668.

Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45702

9 years agonetifd: improve IPv6 onlink-route handling
Steven Barth [Tue, 19 May 2015 09:01:34 +0000 (09:01 +0000)]
netifd: improve IPv6 onlink-route handling

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 45701

9 years agoipv6: remove now unneeded source-dest-routing workarounds
Steven Barth [Tue, 19 May 2015 07:53:08 +0000 (07:53 +0000)]
ipv6: remove now unneeded source-dest-routing workarounds

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 45700

9 years agolinux: backport IPv6 SAS fixes for source-specific routes
Steven Barth [Tue, 19 May 2015 07:43:59 +0000 (07:43 +0000)]
linux: backport IPv6 SAS fixes for source-specific routes

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 45699

9 years agorpcd: unload package after uci reverts
Jo-Philipp Wich [Sun, 17 May 2015 12:16:27 +0000 (12:16 +0000)]
rpcd: unload package after uci reverts

Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 45695

9 years agobcm53xx: add Asus RT-AC87U partitioning hacky workaround
Rafał Miłecki [Sun, 17 May 2015 12:10:38 +0000 (12:10 +0000)]
bcm53xx: add Asus RT-AC87U partitioning hacky workaround

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45694

9 years agodnsmasq: bump to dnsmasq2.73rc8 Important.
Steven Barth [Sun, 17 May 2015 08:06:45 +0000 (08:06 +0000)]
dnsmasq: bump to dnsmasq2.73rc8 Important.

Bump dnsmasq to v2.73rc8

Important - fixes remotely exploitable buffer overflow introduced in all v2.73 test/release candidates.

Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
SVN-Revision: 45693

9 years agobcm53xx: try to fix network on Asus RT-AC87U
Rafał Miłecki [Sat, 16 May 2015 17:14:25 +0000 (17:14 +0000)]
bcm53xx: try to fix network on Asus RT-AC87U

This device seems to have switch port 7 connected to the CPU:
vlan1ports=1 2 3 5 7*
vlan2ports=0 7u
it should be handled by eth1 and NVRAM seems to confirm that (no
et0macaddr entry, existing et1macaddr & et1phyaddr entries).

One of the remaining ports (4/8?) may be connected to the Quantenna SoC.
Original firmware boot log contains following messages:
(0x00,0x5d)Port 5 States Override: 0xfb
(0x00,0x5f)Port 7 States Override: 0xfb
(0x00,0x0e)Port 8 States Override: 0x0a
(why does it force port 5 state?!)

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45692

9 years agobcm53xx: use backported Asus RT-AC87U DT patch
Rafał Miłecki [Sat, 16 May 2015 15:44:00 +0000 (15:44 +0000)]
bcm53xx: use backported Asus RT-AC87U DT patch

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45691

9 years agobcm53xx: enable RT-AC87U UART console
Rafał Miłecki [Sat, 16 May 2015 11:45:52 +0000 (11:45 +0000)]
bcm53xx: enable RT-AC87U UART console

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
SVN-Revision: 45690

9 years agomap: set ealen to psidlen for lw4over6
Steven Barth [Sat, 16 May 2015 10:12:15 +0000 (10:12 +0000)]
map: set ealen to psidlen for lw4over6

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 45689

9 years agoodhcp6c: fix parsing of LW4over6 parameters
Steven Barth [Sat, 16 May 2015 07:29:42 +0000 (07:29 +0000)]
odhcp6c: fix parsing of LW4over6 parameters

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 45688

9 years agobcm53xx: add support for SmartRG SR400ac
Rafał Miłecki [Wed, 13 May 2015 20:01:15 +0000 (20:01 +0000)]
bcm53xx: add support for SmartRG SR400ac

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45686

9 years agobcm53xx: replace DT patches with backports from stblinux
Rafał Miłecki [Wed, 13 May 2015 20:01:03 +0000 (20:01 +0000)]
bcm53xx: replace DT patches with backports from stblinux

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45685

9 years agobcm53xx: use of_machine_is_compatible helper in b53 fix
Rafał Miłecki [Wed, 13 May 2015 14:34:23 +0000 (14:34 +0000)]
bcm53xx: use of_machine_is_compatible helper in b53 fix

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45684

9 years agobcm53xx: fix "typo" in b53 patch fixing R8000 CPU port
Rafał Miłecki [Wed, 13 May 2015 13:12:57 +0000 (13:12 +0000)]
bcm53xx: fix "typo" in b53 patch fixing R8000 CPU port

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45683

9 years agobcm53xx: add b53 hacky workaround for Netgear R8000 CPU port
Rafał Miłecki [Wed, 13 May 2015 13:11:17 +0000 (13:11 +0000)]
bcm53xx: add b53 hacky workaround for Netgear R8000 CPU port

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45682

9 years agobcm53xx: fix default network interface on Netgear R8000
Rafał Miłecki [Wed, 13 May 2015 13:11:04 +0000 (13:11 +0000)]
bcm53xx: fix default network interface on Netgear R8000

It has 3 Ethernet interfaces, each of them connected to separated switch
port. Default NVRAM uses switch port 8 as CPU which is connected to the
3rd interface (eth2).

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45681

9 years agobcm53xx: bgmac: add support for the 3rd core
Rafał Miłecki [Wed, 13 May 2015 13:11:00 +0000 (13:11 +0000)]
bcm53xx: bgmac: add support for the 3rd core

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45680

9 years agoodhcpd: ignore /64 on interface when doing PD
Steven Barth [Wed, 13 May 2015 12:31:06 +0000 (12:31 +0000)]
odhcpd: ignore /64 on interface when doing PD

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 45679

9 years agobcm53xx: sync NVRAM and SPROM drivers with upstream version
Rafał Miłecki [Wed, 13 May 2015 08:30:45 +0000 (08:30 +0000)]
bcm53xx: sync NVRAM and SPROM drivers with upstream version

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45678

9 years agokernel: backport struct ssb_sprom changes submitted by MIPS tree
Rafał Miłecki [Wed, 13 May 2015 07:52:43 +0000 (07:52 +0000)]
kernel: backport struct ssb_sprom changes submitted by MIPS tree

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45677

9 years agob53: Allow using all ports on BCM53012
Rafał Miłecki [Tue, 12 May 2015 13:18:53 +0000 (13:18 +0000)]
b53: Allow using all ports on BCM53012

This chipset has at least 8 usable ports, e.g. Netgear R8000 has ports
5, 7 and 8 connected to Ethernet interfaces:
vlan1ports=0 1 2 3 5 7 8*
vlan2ports=4 8u
Port 6 seems to be always disabled.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Acked-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45676

9 years agoodhcpd: remove invalid call to free()
Steven Barth [Mon, 11 May 2015 19:49:03 +0000 (19:49 +0000)]
odhcpd: remove invalid call to free()

Signed-off-by: Steven Barth <steven@midlink.org>
SVN-Revision: 45675

9 years agoimage: disable compile target within image builder
Felix Fietkau [Mon, 11 May 2015 09:17:51 +0000 (09:17 +0000)]
image: disable compile target within image builder

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45674

9 years agobuild: disable initramfs build within image builder
Felix Fietkau [Mon, 11 May 2015 09:17:39 +0000 (09:17 +0000)]
build: disable initramfs build within image builder

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45673

9 years agoar71xx: convert tl-wr741nd v1/v2 to new image building code, add infrastructure for...
Felix Fietkau [Mon, 11 May 2015 09:08:33 +0000 (09:08 +0000)]
ar71xx: convert tl-wr741nd v1/v2 to new image building code, add infrastructure for building the loader

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45672

9 years agobuild: make initramfs kernel builds depend on the image prepare step
Felix Fietkau [Mon, 11 May 2015 09:07:58 +0000 (09:07 +0000)]
build: make initramfs kernel builds depend on the image prepare step

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45671

9 years agobuild: add support for adding a per-device compile step in the new image building...
Felix Fietkau [Mon, 11 May 2015 09:07:21 +0000 (09:07 +0000)]
build: add support for adding a per-device compile step in the new image building code

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45670

9 years agomtd: remove the warning about read-only caused by size vs. block boundary mismatch
Felix Fietkau [Sun, 10 May 2015 16:09:58 +0000 (16:09 +0000)]
mtd: remove the warning about read-only caused by size vs. block boundary mismatch

Since the kernel/rootfs split handling was modified 2 years ago by r37283 (
https://dev.openwrt.org/changeset/37283 ) and by the subsequent checkins,
users have seen rather scary mtd errors in the log at every boot. The message
ends "-- forcing read-only", which looks a bit error-like. That error has
been mentioned in some forum threads, when users have noticed this message
instead of some actual error.

[    2.940000] 0x000000070000-0x000000ff0000 : "firmware"
[    2.970000] 2 netgear-fw partitions found on MTD device firmware
[    2.970000] 0x000000070000-0x000000188440 : "kernel"
[    2.980000] mtd: partition "kernel" must either start or end on erase
block boundary or be smaller than an erase block -- forcing read-only
[    2.990000] 0x000000188440-0x000000ff0000 : "rootfs"

The patch removes the rather useless warning message.

signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>

SVN-Revision: 45669

9 years agoscripts/feeds: add support for optionally using a full clone of git repositories
Felix Fietkau [Sun, 10 May 2015 15:51:19 +0000 (15:51 +0000)]
scripts/feeds: add support for optionally using a full clone of git repositories

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45668

9 years agoar71xx: add platform_pre_upgrade for sysupgrade
Felix Fietkau [Sun, 10 May 2015 13:54:25 +0000 (13:54 +0000)]
ar71xx: add platform_pre_upgrade for sysupgrade

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 45667

9 years agotoolchain: change default emulation for mips64 (octeon) to n64
Felix Fietkau [Sun, 10 May 2015 13:54:20 +0000 (13:54 +0000)]
toolchain: change default emulation for mips64 (octeon) to n64

fixes several ld errors reported by buildbot:
http://buildbot.openwrt.org:8010/broken_packages/octeon/mdnsresponder/compile.txt
http://buildbot.openwrt.org:8010/broken_packages/octeon/luasocket/compile.txt
http://buildbot.openwrt.org:8010/broken_packages/octeon/luasec/compile.txt
http://buildbot.openwrt.org:8010/broken_packages/octeon/openvswitch/compile.txt
http://buildbot.openwrt.org:8010/broken_packages/octeon/prosody/compile.txt

link error:
 ABI is incompatible with that of the selected emulation

found at OE toolchain for bintils 2.22:
http://patchwork.openembedded.org/patch/28033/

Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 45666

9 years agokernel: fix regression in backtrace module info change (#19631)
Felix Fietkau [Sun, 10 May 2015 13:16:40 +0000 (13:16 +0000)]
kernel: fix regression in backtrace module info change (#19631)

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45665

9 years agogcc: add some size optimization patches
Felix Fietkau [Sun, 10 May 2015 13:16:35 +0000 (13:16 +0000)]
gcc: add some size optimization patches

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45664

9 years agoipq806x: add pcie support to ipq806x based platforms
Felix Fietkau [Sun, 10 May 2015 11:47:09 +0000 (11:47 +0000)]
ipq806x: add pcie support to ipq806x based platforms

This change adds PCIe support to IPQ806x based platforms. The driver is
actually cherry-picked from the following LKML thread:
*https://lwn.net/Articles/643086/ (patches 110-111)

We also add here an additional fix to support multiple PCI controllers
on the same platform (patch 112), and to patch the ap148 & dbs149 DTS
files (patch 113).

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45663

9 years agoipq806x: add support for zImage kernel
Felix Fietkau [Sun, 10 May 2015 11:47:02 +0000 (11:47 +0000)]
ipq806x: add support for zImage kernel

This change enable zImage+appended dtb support in ipq806x kernel
options. The zImage will now be generated as part of the kernel
binaries. Platforms which do not have DT support enabled in U-boot
can now make use of it by generating zImage files and appending dtb
to it.

It is not used yet but it is done as a stepping stone for early IPQ806x
platforms, which did not include DT support in U-boot.

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45662

9 years agoipq806x: fix boot freeze on zImage kernel
Felix Fietkau [Sun, 10 May 2015 11:46:56 +0000 (11:46 +0000)]
ipq806x: fix boot freeze on zImage kernel

ARCH_QCOM is using the ARCH_MULTIPLATFORM option, as now recommended
on most ARM architectures. This automatically calculate ZRELADDR by
masking PHYS_OFFSET with 0xf8000000.

On IPQ806x though, the first ~20MB of RAM is reserved for the hardware.
In newer bootloader, when DT is used, this is not a problem, we just
reserve this memory in the device tree. But if the bootloader doesn't
have DT support, then ATAGS have to be used. In this case, the ARM
decompressor will position the kernel in this low mem, which will not be
in the RAM section mapped by the bootloader, which means the kernel will
freeze in the middle of the boot process trying to map the memory.

As a work around, this patch allows disabling AUTO_ZRELADDR when
ARCH_QCOM is selected. It makes the zImage usage possible on bootloaders
which don't support device-tree, which is the case on certain early
IPQ806x based designs.

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45661

9 years agoinclude: make dtb argument optional in MkFIT
Felix Fietkau [Sun, 10 May 2015 11:46:50 +0000 (11:46 +0000)]
include: make dtb argument optional in MkFIT

As mkits.sh makes the -d argument optional, we'll make this same
argument optional in the MkFIT function as well.

With this change, MkFIT can be used both to generate FIT images with DT,
and without DT.

Signed-off-by: Mathieu Olivari <mathieu@qca.qualcomm.com>
SVN-Revision: 45660

9 years agobuild: make device-tree arg optional in mkits.sh
Felix Fietkau [Sun, 10 May 2015 11:46:45 +0000 (11:46 +0000)]
build: make device-tree arg optional in mkits.sh

mkits.sh help currently shows dtb file as optional, but generates an
invalid its file when this option is unspecified.

We're fixing this problem by removing the use of the sed command and
just including variables instead.

Signed-off-by: Mathieu Olivari <mathieu@qca.qualcomm.com>
SVN-Revision: 45659

9 years agogeneric: add CRYPTO_DEV_QCE option to default config
Felix Fietkau [Sun, 10 May 2015 11:46:41 +0000 (11:46 +0000)]
generic: add CRYPTO_DEV_QCE option to default config

This option has been added in kernel 3.17. It shows-up only when both
ARCH_QCOM and CRYPTO are enabled. So we'll disable these two by default
to avoid stalling the build when these conditions are met.

Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
SVN-Revision: 45658

9 years agobuild: don't call prereq for any package/symlinks rules
Felix Fietkau [Sun, 10 May 2015 11:17:29 +0000 (11:17 +0000)]
build: don't call prereq for any package/symlinks rules

Most of the time, we want to make sure OpenWrt has been configured and
setup before start running make. However, in case of package/symlinks,
forcing prereq as a dependency creates multiple issues:
*when executed on a clean workspace, it will prompt for user input
 and open a menuconfig window before executing the feeds command
*the only way around that is to provide a .config. However, the "prereq"
 target would then run a "make defconfig", which will remove all the
 packages in the .config but from external feeds, as feeds have not been
 installed yet.

The only way to currently work around this, is to generate a fake config
by running "make defconfig", then "make package/symlinks", copy the real
config (which at this point disregards the previously generated config),
and run make defconfig again. Something like this:

make defconfig
make package/symlinks
cp real.config .config
make defconfig

This change is removing the need for the first defconfig, making the
process more logical for OpenWrt users using the package/symlinks target.

Signed-off-by: Mathieu Olivari <mathieu@qca.qualcomm.com>
SVN-Revision: 45657

9 years agoar71xx: add status led found on ew-dorin boards
Luka Perkov [Sun, 10 May 2015 07:47:04 +0000 (07:47 +0000)]
ar71xx: add status led found on ew-dorin boards

Signed-off-by: Catrinel Catrinescu <cc@80211.de>
SVN-Revision: 45656

9 years agokernel: when KALLSYMS is disabled, print module address + size for matching backtrace...
Felix Fietkau [Sun, 10 May 2015 00:02:18 +0000 (00:02 +0000)]
kernel: when KALLSYMS is disabled, print module address + size for matching backtrace entries

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 45655