openwrt/staging/aparcar.git
6 years agoar71xx: fix QCA955X SGMII link loss
David Bauer [Mon, 6 Aug 2018 14:15:04 +0000 (16:15 +0200)]
ar71xx: fix QCA955X SGMII link loss

The QCA955X is affected by a hardware bug which causes link-loss of the
SGMII link between SoC and PHY. This happens on change of link-state or
speed.

It is not really known what causes this bug. It definitely occurs when
using a AR8033 Gigabit Ethernet PHY.

Qualcomm solves this Bug in a similar fashion. We need to apply the fix
on a per-device base via platform-data as performing the fixup work will
break connectivity in case the SGMII interface is connected to a Switch.

This bug was first proposed to be fixed by Sven Eckelmann in 2016.
 https://patchwork.ozlabs.org/patch/604782/

Based-on-patch-by: Sven Eckelmann <sven.eckelmann@open-mesh.com>
Signed-off-by: David Bauer <mail@david-bauer.net>
6 years agokernel: allow device-tree configuration of at803x
David Bauer [Mon, 6 Aug 2018 14:21:03 +0000 (16:21 +0200)]
kernel: allow device-tree configuration of at803x

This commit adds the ability to configure specific functions of the
at803x series ethernet-PHYs, which were previously configured
exclusively with the help of platform-data, via device-tree.

This is needed to fully support existing boards of the ar71xx platform.

Signed-off-by: David Bauer <mail@david-bauer.net>
6 years agokernel: backport upstream fix for CVE-2018-5390
Jo-Philipp Wich [Wed, 8 Aug 2018 09:12:18 +0000 (11:12 +0200)]
kernel: backport upstream fix for CVE-2018-5390

Backport an upstream fix for a remotely exploitable TCP denial of service
flaw in Linux 4.9+.

The fixes are included in Linux 4.14.59 and later but did not yet end up in
version 4.9.118.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoiproute2: add libutil to InstallDev section
Hans Dedecker [Wed, 8 Aug 2018 12:22:11 +0000 (14:22 +0200)]
iproute2: add libutil to InstallDev section

In iproute2 v4.17 ll_map has been moved from the libnetlink to the libutil
library; add libutil as well to the staging dir in order to keep support
for ll_map

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agokernel: remove linux 4.4 support
Koen Vandeputte [Wed, 8 Aug 2018 07:51:23 +0000 (09:51 +0200)]
kernel: remove linux 4.4 support

No targets are using this one anymore

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agoath79: fix QCA9557 eth PLL settings
David Bauer [Mon, 6 Aug 2018 14:21:02 +0000 (16:21 +0200)]
ath79: fix QCA9557 eth PLL settings

The QCA9557 dtsi is currently missing pll-handle and pll-regs for both
eth0 and eth1, therefore PLL settings won't be applied. This commit
fixes this behavior.

Signed-off-by: David Bauer <mail@david-bauer.net>
6 years agouboot-envtools: add ath79 target
David Bauer [Mon, 6 Aug 2018 14:21:01 +0000 (16:21 +0200)]
uboot-envtools: add ath79 target

This adds uci entries for all ath79 devices for which this already was
the case on ar71xx. Additionally we add the OCEDO Koala as there was no
support in OpenWRT yet.

Signed-off-by: David Bauer <mail@david-bauer.net>
6 years agomediatek: Fix amount of memory on U7623
Kristian Evensen [Wed, 8 Aug 2018 04:38:32 +0000 (06:38 +0200)]
mediatek: Fix amount of memory on U7623

While finalizing support for the U7623 with 512MB, I made an embarresing
error and configured 1GB RAM for the board. I also forgot to move memory
from the dtsi and to the dts. This commit takes care of my mistakes.

While I am confessing my mistakes, I also note that I made a mistake in
the commit message of the initial U7623 commit. It is the .bin-file, and
not the .gz file that shall be sent to the device via tftp.

v1->v2:
* Remove redundant memory node (thanks Jonas Gorski)

Signed-off-by: Kristian Evensen <kristian.evensen@gmail.com>
6 years agoath79: don't include tl-wdr3600 image build code for tl-wdr4300
Mathias Kresin [Wed, 8 Aug 2018 05:19:30 +0000 (07:19 +0200)]
ath79: don't include tl-wdr3600 image build code for tl-wdr4300

Including the tl-wdr3600 image build code just to overwrite most of it
doesn't make much sense and only makes it hard to read.

Furthermore, the tl-wdr4300 image will be marked as compatible with the
tl-wdr3600 this way.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoath79: cleanup netgear wnr612-v2 supported devices
Mathias Kresin [Wed, 8 Aug 2018 05:16:40 +0000 (07:16 +0200)]
ath79: cleanup netgear wnr612-v2 supported devices

The netgear,wnr612-v2 is included by default based on the device
define.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoath79: mark netgear variables as device specific
Mathias Kresin [Mon, 6 Aug 2018 15:47:41 +0000 (17:47 +0200)]
ath79: mark netgear variables as device specific

The variables are used in image build recipes and need to be marked as
per devices vars to be stored individual per image define. Otherwise
the last defined variable will be used for all boards.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoath79: Fix led nodes for TL-WR740N v2 and add its clones
Chuanhong Guo [Thu, 2 Aug 2018 05:17:44 +0000 (13:17 +0800)]
ath79: Fix led nodes for TL-WR740N v2 and add its clones

This patch did the following things:
1. Separate ath9k-leds out of gpio leds so that all other leds will work
   before ath9k loded (e.g. during preinit/init stage).
2. Rename wps led to qss since that's how TP-Link mark it.
3. Rename LED prefix to tp-link because that dts is shared by many devices.
4. Rename to wr740n-v1 because v1 is the first and v2 just use the fw of v1.
   (This will require a forced sysupgrade if you comes from
   the previous wr740n v2 image.)
5. Remove SUPPORTED_DEVICES.
   (tl-wr740n-v2 doesn't exist anywhere so it's useless.)
6. Add all WR741ND v1 clones found in ar71xx.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agoath79: fix dts warnings
Mathias Kresin [Mon, 6 Aug 2018 05:11:13 +0000 (07:11 +0200)]
ath79: fix dts warnings

Fix all issues found by the devicetree compiler like wrong address/size
cells as well as wrong/missing/superfluous unit addresses.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoath79: fix node names
Mathias Kresin [Mon, 6 Aug 2018 06:17:24 +0000 (08:17 +0200)]
ath79: fix node names

Use the standardized node names from the devicetree specification.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoath79: fix compatible strings
Mathias Kresin [Mon, 6 Aug 2018 06:03:30 +0000 (08:03 +0200)]
ath79: fix compatible strings

Use only the jedec,spi-nor compatible string. Everything else either
never worked or is only support to keep compatibility.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoath79: fix whitespace issue in dts files
Mathias Kresin [Mon, 6 Aug 2018 05:55:13 +0000 (07:55 +0200)]
ath79: fix whitespace issue in dts files

Fixes spaces vs. tabs issues and remove indentation on blank lines.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: use #include syntax for dtsi files
Mathias Kresin [Mon, 6 Aug 2018 18:02:00 +0000 (20:02 +0200)]
ramips: use #include syntax for dtsi files

Use the same syntax for including dtsi for all dts files.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agobase-files: introduce sysupgrade signature chain verification
Daniel Golle [Mon, 15 Jan 2018 02:37:17 +0000 (03:37 +0100)]
base-files: introduce sysupgrade signature chain verification

Verify ucert signature chains in sysupgrade images in case ucert is
installed and $CHECK_IMAGE_SIGNARURE = 1.
Also make sure ucert host binary is present and generate a self-signed
ucert in case $TOPDIR/key-build.ucert is missing.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agoimage: fix build without ucert
Daniel Golle [Tue, 7 Aug 2018 22:47:12 +0000 (00:47 +0200)]
image: fix build without ucert

Make sure the Shell-expression returns true also in case of
key-build.ucert being absent.

Fixes commit 848b455d2e ("image: use ucert to append signature")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agoimage: use ucert to append signature
Daniel Golle [Mon, 6 Aug 2018 19:20:57 +0000 (21:20 +0200)]
image: use ucert to append signature

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agoucert: update source
Daniel Golle [Mon, 6 Aug 2018 16:00:45 +0000 (18:00 +0200)]
ucert: update source

 ad816fc set rpath to make bundle-libraries.sh happy
 63ad591 blob_buf needs to be zero'd

Now that libubox, libjson-c and libblobms_json are installed into
STAGING_DIR_HOST we can properly bundle ucert in the ImageBuilder.
Follow-up commits will make use of it to include a signature-chain in
sysupgrade images using fwtool.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agolibubox: set HOST_BUILD_PREFIX
Daniel Golle [Mon, 6 Aug 2018 16:00:15 +0000 (18:00 +0200)]
libubox: set HOST_BUILD_PREFIX

Install into STAGING_DIR_HOST rather than STAGING_DIR_HOSTPKG to make
bundle-libraries.sh happy.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agolibjson-c: set HOST_BUILD_PREFIX
Daniel Golle [Mon, 6 Aug 2018 15:58:32 +0000 (17:58 +0200)]
libjson-c: set HOST_BUILD_PREFIX

Install into STAGING_DIR_HOST rather than STAGING_DIR_HOSTPKG to make
bundle-libraries.sh happy.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agolibubox: fix source version date
Jo-Philipp Wich [Tue, 7 Aug 2018 15:24:30 +0000 (17:24 +0200)]
libubox: fix source version date

The referenced Git commit was made on the 25th of July, not June.

Fixes 432eaa940f ("libubox: fix mirror hash")
Fixes 5dc32620c4 ("libubox: update to latest git HEAD")
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agolibubox: fix mirror hash
Jo-Philipp Wich [Tue, 7 Aug 2018 14:30:20 +0000 (16:30 +0200)]
libubox: fix mirror hash

Correct the mirror hash to reflect whats on the download server.

A locally produced libubox SCM tarball was also verified to yield an identical
checksum compared to the one currently on the download server.

Fixes FS#1707.
Fixes 5dc32620c4 ("libubox: update to latest git HEAD")
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agotoolchain/glibc: update to latest 2.26 commit
Hans Dedecker [Tue, 7 Aug 2018 11:29:29 +0000 (13:29 +0200)]
toolchain/glibc: update to latest 2.26 commit

c9570bd2f5 x86: Populate COMMON_CPUID_INDEX_80000001 for Intel CPUs [BZ #23459]
86e0996b1a x86: Correct index_cpu_LZCNT [BZ #23456]
cf6deb084b conform/conformtest.pl: Escape literal braces in regular expressions
b12bed3e06 stdio-common/tst-printf.c: Remove part under a non-free license [BZ #23363]
20dc7a909a libio: Add tst-vtables, tst-vtables-interposed
4b10e69b1f Synchronize support/ infrastructure with master
762e9d63d5 NEWS: Reorder out-of-order bugs
2781bd5a86 libio: Disable vtable validation in case of interposition [BZ #23313]

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agoustream-ssl: update to latest git HEAD
Eneas U de Queiroz [Mon, 6 Aug 2018 18:19:20 +0000 (18:19 +0000)]
ustream-ssl: update to latest git HEAD

23a3f28 openssl, wolfssl: match mbedTLS ciphersuite list
450ada0 ustream-ssl: Revised security on mbedtls
34b0b80 ustream-ssl: add openssl-1.1.0 compatibility

Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
6 years agocns3xxx: ethernet: use circular queue checks consistently
Koen Vandeputte [Tue, 7 Aug 2018 09:44:36 +0000 (11:44 +0200)]
cns3xxx: ethernet: use circular queue checks consistently

Use the same method for setting queue index pointers consistenly
throughout the source file.

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agocns3xxx: ethernet: cleanup code
Koen Vandeputte [Tue, 7 Aug 2018 09:18:08 +0000 (11:18 +0200)]
cns3xxx: ethernet: cleanup code

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agocns3xxx: ethernet: fix signed/unsigned comparison
Koen Vandeputte [Tue, 7 Aug 2018 08:52:53 +0000 (10:52 +0200)]
cns3xxx: ethernet: fix signed/unsigned comparison

Fixes a compiler warning

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agocns3xxx: add myself as maintainer
Koen Vandeputte [Tue, 7 Aug 2018 10:43:53 +0000 (12:43 +0200)]
cns3xxx: add myself as maintainer

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agokernel: bump 4.14 to 4.14.61
Koen Vandeputte [Tue, 7 Aug 2018 08:34:37 +0000 (10:34 +0200)]
kernel: bump 4.14 to 4.14.61

Refreshed all patches.

Compile-tested on: cns3xxx, imx6, x86_64
Runtime-tested on: cns3xxx, imx6, x86_64

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agokernel: bump 4.9 to 4.9.118
Koen Vandeputte [Tue, 7 Aug 2018 08:33:52 +0000 (10:33 +0200)]
kernel: bump 4.9 to 4.9.118

Refreshed all patches.

Compile-tested on: ar71xx
Runtime-tested on: ar71xx

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agoprocd: update to latest git HEAD
Hans Dedecker [Mon, 6 Aug 2018 20:19:12 +0000 (22:19 +0200)]
procd: update to latest git HEAD

e29966f Allow disabling seccomp or changing the whitelist
5f57223 trace: Use properly sized type for PTRACE_GETEVENTMSG
747efb6 procd: fix ustream deadlock when there are 0 bytes or no newlines

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agoRevert "netfilter: separate IPv6 relevant kernel modules from IPv4"
Jo-Philipp Wich [Mon, 6 Aug 2018 17:46:31 +0000 (19:46 +0200)]
Revert "netfilter: separate IPv6 relevant kernel modules from IPv4"

This reverts commit 42a3c6465a230a4e03f2a185f4db5ac57b89f673.

The change was apparently never build-tested with all kmods enabled. I took
a brief look but found no simple way to untangle this, so revert it.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agonetfilter: separate IPv6 relevant kernel modules from IPv4
Rosy Song [Sat, 4 Aug 2018 02:25:54 +0000 (10:25 +0800)]
netfilter: separate IPv6 relevant kernel modules from IPv4

Signed-off-by: Rosy Song <rosysong@rosinson.com>
6 years agobase-files: do not add relevant sections & options except when ipv6 is support in...
Rosy Song [Sat, 4 Aug 2018 11:01:18 +0000 (19:01 +0800)]
base-files: do not add relevant sections & options except when ipv6 is support in kernel

Signed-off-by: Rosy Song <rosysong@rosinson.com>
6 years agoath79: rename tl-archer-c7 to archer-c7
Chuanhong Guo [Thu, 2 Aug 2018 07:14:10 +0000 (15:14 +0800)]
ath79: rename tl-archer-c7 to archer-c7

This router is called Archer C7 and the tl was used to identify
TP-LINK. Since we have added tplink in dts/board name, the tl
prefix is useless now.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agoath79: alphabetical order image/generic.mk
Chuanhong Guo [Thu, 2 Aug 2018 07:14:10 +0000 (15:14 +0800)]
ath79: alphabetical order image/generic.mk

Move the wzr-hp-g450h in image/generic.mk to keep alphabetical order.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agoath79: rename dts/image using manufacturer_board scheme
Chuanhong Guo [Thu, 2 Aug 2018 07:14:10 +0000 (15:14 +0800)]
ath79: rename dts/image using manufacturer_board scheme

This helps getting rid of SUPPORTED_DEVICES which can be
auto-generated in image/Makefile.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agoath79: remove quotes for boardname in scripts
Chuanhong Guo [Thu, 2 Aug 2018 07:14:10 +0000 (15:14 +0800)]
ath79: remove quotes for boardname in scripts

Quotes are not required in case statements.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agoath79: put all devices in alphabetical order in scripts
Chuanhong Guo [Thu, 2 Aug 2018 07:14:10 +0000 (15:14 +0800)]
ath79: put all devices in alphabetical order in scripts

And remove specifying lan interfaces as eth1.1 because this is
handled by ucidef_add_switch.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agoath79: fix support for TL-MR3020 v1
Chuanhong Guo [Thu, 2 Aug 2018 07:14:10 +0000 (15:14 +0800)]
ath79: fix support for TL-MR3020 v1

Change lan and it's LED to eth0
It's broken since c7c807cb8c3fd6538101de885f66d4681785defe
where I changed the dts but forgot to change default configurations.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agoath79: fix support for TL-WR740N/WR741N v4
Chuanhong Guo [Thu, 2 Aug 2018 07:14:10 +0000 (15:14 +0800)]
ath79: fix support for TL-WR740N/WR741N v4

1. Swap eth0/eth1
 Both devices are using AR9331, the builtin switch on AR9331 is
 connected to gmac1 and gmac1 is named as eth1 in ath79.
 PS: gmac1 is eth0 and gmac0 is eth1 in ar71xx because of the
 reversed initialization order.
2. Fix the incorrect compatible string in dts

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agoath79: remove useless phy-handle nodes
Chuanhong Guo [Thu, 2 Aug 2018 07:14:10 +0000 (15:14 +0800)]
ath79: remove useless phy-handle nodes

phy-handle is used to poll link status. They are useless when
we need fixed-link on these interfaces.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
6 years agokernel: modules: fix kmod-regmap redux
Christian Lamparter [Fri, 3 Aug 2018 21:49:22 +0000 (23:49 +0200)]
kernel: modules: fix kmod-regmap redux

Jonas Gorski commented on the previous patch:
|This is actually the wrong fix and papers over an issue in one of our
|local patches.
|
|We intentionally allow regmap to be built as a module, see
|
|/target/linux/generic/hack-4.14/259-regmap_dynamic.patch
|[...]
|[The regulator code] optionally supports regmap thanks to the stubs
|provided if regmap is disabled - which breaks if you compile regmap
|as a module.

In order to mitigate this issue, this patch reverts the previous patch
and replaces the existing IS_ENABLED(CONFIG_REGMAP) with
IS_REACHABLE(CONFIG_REGMAP). This solves this particular issue as the
regulator code will now automatically fallback to the regmap stubs in
case the kmod-regmap module is enabled, but nothing else sets
CONFIG_REGMAP=y.

Note: There's still a potential issue that this patch doesn't solve:
If someone ever wants to make a OpenWrt kernel package for a
regulator module that requires the REGMAP feature for a target that
doesn't set CONFIG_REGMAP=y but has CONFIG_REGULATOR=y, the resulting
kmod-regulator-xyz package will not work on the target.
Luckily, there aren't any in-tree OpenWrt kernel module packages for
regulators at the moment. On the bright side: regmap is a critical
part nowadays and all new and upcoming architectures require it by
default. This will likely only ever be a problem for legacy targets
and devices that cannot afford to enable REGMAP.

Cc: Jonas Gorski <jonas.gorski@gmail.com>
Cc: John Crispin <john@phrozen.org>
Fixes: d00913d1215b ("kernel: modules: fix kmod-regmap")
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
6 years agomt7620: gsw: make IntPHY and ExtPHY share mdio addr 4 possible
Chen Minqiang [Fri, 3 Aug 2018 17:14:07 +0000 (01:14 +0800)]
mt7620: gsw: make IntPHY and ExtPHY share mdio addr 4 possible

To share mdio addr for IntPHY and ExtPHY,
as described in the documentation (MT7620_ProgrammingGuide.pdf).
(refer: http://download.villagetelco.org/hardware/MT7620/MT7620_ProgrammingGuide.pdf)

when port4 setup to work as gmac mode, dts like:

&gsw {
    mediatek,port4 = "gmac";
};

we should set SYSCFG1.GE2_MODE==0x0 (RGMII).
but SYSCFG1.GE2_MODE may have been set to 3(RJ-45) by uboot/default
so we need to re-set it to 0x0

before this changes:
gsw: 4FE + 2GE may not work correctly and MDIO addr 4 cannot be used by ExtPHY

after this changes:
gsw: 4FE + 2GE works and MDIO addr 4 can be used by ExtPHY

Signed-off-by: Chen Minqiang <ptpt52@gmail.com>
6 years agoramips: fix gigabit switch PHY access on MDIO
Daniel Gimpelevich [Wed, 1 Aug 2018 14:51:47 +0000 (07:51 -0700)]
ramips: fix gigabit switch PHY access on MDIO

When PHY's are defined on the MDIO bus in the DTS, gigabit support was
being masked out for no apparent reason, pegging all such ports to 10/100.
If gigabit support must be disabled for some reason, there should be a
"max-speed" property in the DTS.

Reported-by: James McKenzie <openwrt@madingley.org>
Signed-off-by: Daniel Gimpelevich <daniel@gimpelevich.san-francisco.ca.us>
6 years agoramips: remove superfluous & confusing DT binding
Daniel Gimpelevich [Wed, 1 Aug 2018 14:44:20 +0000 (07:44 -0700)]
ramips: remove superfluous & confusing DT binding

Mediatek has a reference platform that pairs an MT7620A with an MT7530W,
where the latter responds on MDIO address 0x1f while both chips respond on
0x0 to 0x4. The driver special-cases this arrangement to make sure it's
talking to the right chip, but two different ways in two different places.
This patch consolidates the detection without the current requirement of
both tests to be separately satisfied in the DTS.

Signed-off-by: Daniel Gimpelevich <daniel@gimpelevich.san-francisco.ca.us>
6 years agoramips: add support for Edimax BR-6478AC v2
Daniel Gimpelevich [Mon, 16 Jul 2018 14:19:19 +0000 (07:19 -0700)]
ramips: add support for Edimax BR-6478AC v2

Roll-up of patches by Rohan Murch, Hans Ulli Kroll, and James McKenzie.
Taken from https://forum.openwrt.org/viewtopic.php?id=67192 and updated.

Specification:
- System-On-Chip: MT7620A
- CPU/Speed: 580 MHz
- Flash-Chip: Macronix MX25L6405D
- Flash size: 8192 KiB
- RAM: 64 MiB
- Wireless No1: SoC-integrated: MT7620A 2.4GHz 802.11bgn
- Wireless No2: On-board chip: MT7612E 5GHz 802.11ac
- Switch: Mediatek MT7530W Gigabit Switch
- USB: Yes 1 x 2.0

Installation:
1. Download sysupgrade.bin
2. Open vendor web interface
3. Choose to upgrade firmware
4. After reboot connect via ethernet at 192.168.1.1

Signed-off-by: Daniel Gimpelevich <daniel@gimpelevich.san-francisco.ca.us>
6 years agoimagebuilder: manifest function show stderr
Paul Spooren [Sun, 5 Aug 2018 03:17:29 +0000 (12:17 +0900)]
imagebuilder: manifest function show stderr

This really simplifies debugging, if a package is not found or a feed is
not reachable, a proper stderr is printed. Currently it would only say
`_call_manifest` failed.

Signed-off-by: Paul Spooren <mail@aparcar.org>
6 years agobase-files: sysupgrade: abort if config backup fails
Andreas Ziegler [Fri, 3 Aug 2018 23:56:39 +0000 (01:56 +0200)]
base-files: sysupgrade: abort if config backup fails

Sysupgrade shouldn't proceed, if the backup of the configuration
fails because tar (or gzip) exit with a non-zero code.

Signed-off-by: Andreas Ziegler <dev@andreas-ziegler.de>
6 years agoscripts: time.pl: avoid hard Time::HiRes dependency
Jo-Philipp Wich [Sun, 5 Aug 2018 10:24:01 +0000 (12:24 +0200)]
scripts: time.pl: avoid hard Time::HiRes dependency

Use Time::HiRes when available and fallback to raw syscall interface
when not. If that fails too, simply report 0, 0 as real time.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agobuild: remove GNU time dependency
Jo-Philipp Wich [Sat, 4 Aug 2018 22:06:27 +0000 (00:06 +0200)]
build: remove GNU time dependency

Replace the GNU time program invocation with a simple Perl script reporting
the timing values. Since we require Perl anyway for the build system, we can
as well use that instead of requiring a random GNU utility rarely installed
by default.

Fixes: ff6e62b288 ("build: log time taken by each packages/steps")
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoramips: move partitions into partition table node
Alex Maclean [Wed, 1 Aug 2018 20:32:34 +0000 (21:32 +0100)]
ramips: move partitions into partition table node

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

Signed-off-by: Alex Maclean <monkeh@monkeh.net>
6 years agoramips: unify partition node names in dts files
Mathias Kresin [Fri, 3 Aug 2018 18:14:49 +0000 (20:14 +0200)]
ramips: unify partition node names in dts files

Use partition@ as name for all partition nodes. Add a label where
necessary.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix whitespace and comment issues in dts
Mathias Kresin [Sun, 22 Jul 2018 08:33:42 +0000 (10:33 +0200)]
ramips: fix whitespace and comment issues in dts

Fix space vs. tabs issue and trainling whitespaces. Use C style
comments or drop the comments if they explain what is already to see in
the devicetree parameters.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix dtc warnings
Mathias Kresin [Sat, 21 Jul 2018 21:00:20 +0000 (23:00 +0200)]
ramips: fix dtc warnings

Fix individual boards dtc warnings or obvious mistakes.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix hnat dtc warning
Mathias Kresin [Sat, 21 Jul 2018 16:20:59 +0000 (18:20 +0200)]
ramips: fix hnat dtc warning

The hardware NAT node has the same reg/unit as the ethernet node. One
of them need to be a child of the other.

Make the hardware NAT node a child of the ethernet node since the it
"reference" the netdev in its properties.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix pci/pcie related dtc warnings
Mathias Kresin [Sat, 21 Jul 2018 14:19:46 +0000 (16:19 +0200)]
ramips: fix pci/pcie related dtc warnings

Add the ranges property to the PCI bridges where missing. Add the unit
address to PCI bridge where missing.

Rework the complete rt3883 pci node. Drop the PCI unit nodes from the
dtsi. They are not used by any dts file and should be rather in the dts
than in the SoC dtsi. Express the PCI-PCI bridge in a clean devicetree
syntax. The ralink,pci-slot isn't used by any driver, drop it. Move the
pci interrupt controller out of the pci node. It doesn't share the same
reg and therefore should be an independent/SoC child node.

Move the pci related rt3883 pinctrl setting to the dtsi instead of
defining the very same for each rt3883 board.

If the device_type property is used for PCI units, the unit is treated
as pci bridge which it isn't. Drop it for PCI units.

Reference pci-bridges or the pci node defined in the dtsi instead of
recreating the whole node hierarchy. It allows to change the referenced
node in the dtsi without the need to touch all dts.

Fix the PCI(e) wireless unit addresses. All our PCI(e) wireless chips
are the first device on the bus. The unit address has to be the bus
address instead of the PCI vendor/device id.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: use ralink,nr-gpio instead of ralink,num-gpios
Mathias Kresin [Sat, 21 Jul 2018 17:32:51 +0000 (19:32 +0200)]
ramips: use ralink,nr-gpio instead of ralink,num-gpios

Since commit c1e7738988f5 ("checks: add gpio binding properties check")
dtc treats any *-gpios and *-gpio property as phandle at least during
checks. The only whitelisted property is nr-gpio.

Use ralink,nr-gpio in favour of ralink,num-gpios to get rid of false
positive warnings.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix cpu interrupt controller dtc warnings
Mathias Kresin [Sat, 21 Jul 2018 14:53:10 +0000 (16:53 +0200)]
ramips: fix cpu interrupt controller dtc warnings

The cpu interrupt controller doesn't have a reg property, hence we
can't use a unit address in the node name.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix cpu related dtc warnings
Mathias Kresin [Sat, 21 Jul 2018 14:17:39 +0000 (16:17 +0200)]
ramips: fix cpu related dtc warnings

We need a reg property if we are using a unit address.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agotreewide: convert gpio-export to platform driver
René van Dorst [Thu, 19 Jul 2018 10:55:55 +0000 (12:55 +0200)]
treewide: convert gpio-export to platform driver

Without this patch you will get an error "gpio-export probe deferral
not supported" when you try to export i2c expander gpio pins.

gpio-export is probed long before i2c-bus and i2c expander are created
and it doesn't retry it so none pins are exported.

Signed-off-by: René van Dorst <opensource@vdorst.com>
apply the change to all instances of the gpio exports patch
Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoath79: really fix TP-Link Archer C7 v2 MAC address
Mathias Kresin [Wed, 25 Jul 2018 07:03:54 +0000 (09:03 +0200)]
ath79: really fix TP-Link Archer C7 v2 MAC address

Revert 290c54473ead ("ath79: fix TP-Link Archer C7 v2 wlan1 MAC address")
which obviously aims to have a distinct MAC address per interface.

Unfortunally it doesn't match what is used by the stock firmware and we
shouldn'z use MAC Adresses not reserverd for/assigned to a particular
board.

The correct MAC adress increments for this board are:

wlan0 (5GHz) : -1
wlan1 (2.4GHz) : 0
eth1 (LAN) : 0
eth0 (WAN) : 1

Fixes: FS#408
Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agowireguard: bump to 0.0.20180802
Jason A. Donenfeld [Fri, 3 Aug 2018 20:13:13 +0000 (22:13 +0200)]
wireguard: bump to 0.0.20180802

Changelog taken from the version announcement

> == Changes ==
>
>   * chacha20poly1305: selftest: split up test vector constants
>
>   The test vectors are encoded as long strings -- really long strings -- and
>   apparently RFC821 doesn't like lines longer than 998.
>   https://cr.yp.to/smtp/message.html
>
>   * queueing: keep reference to peer after setting atomic state bit
>
>   This fixes a regression introduced when preparing the LKML submission.
>
>   * allowedips: prevent double read in kref
>   * allowedips: avoid window of disappeared peer
>   * hashtables: document immediate zeroing semantics
>   * peer: ensure resources are freed when creation fails
>   * queueing: document double-adding and reference conditions
>   * queueing: ensure strictly ordered loads and stores
>   * cookie: returned keypair might disappear if rcu lock not held
>   * noise: free peer references on failure
>   * peer: ensure destruction doesn't race
>
>   Various fixes, as well as lots of code comment documentation, for a
>   small variety of the less obvious aspects of object lifecycles,
>   focused on correctness.
>
>   * allowedips: free root inside of RCU callback
>   * allowedips: use different macro names so as to avoid confusion
>
>   These incorporate two suggestions from LKML.
>
> This snapshot contains commits from: Jason A. Donenfeld and Jann Horn.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
6 years agouclient: update to latest git HEAD
Jo-Philipp Wich [Fri, 3 Aug 2018 21:50:29 +0000 (23:50 +0200)]
uclient: update to latest git HEAD

f2573da uclient-fetch: use package name pattern in message for missing SSL library
9fd8070 uclient-fetch: Check for nullpointer returned by uclient_get_url_filename
f41ff60 uclient-http: basic auth: Handle memory allocation failure
a73b23b uclient-http: auth digest: Handle multiple possible memory allocation failures
66fb58d uclient-http: Handle memory allocation failure
2ac991b uclient: Handle memory allocation failure for url
63beea4 uclient-http: Implement error handling for header-sending
eb850df uclient-utils: Handle memory allocation failure for url file name
ae1c656 uclient-http: Close ustream file handle only if allocated

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agokernel: bump 4.14 to 4.14.60
Koen Vandeputte [Fri, 3 Aug 2018 08:54:16 +0000 (10:54 +0200)]
kernel: bump 4.14 to 4.14.60

Refreshed all patches

Removed upstreamed patches:
- 500-ext4-fix-check-to-prevent-initializing-reserved-inod.patch

Compile-tested on: cns3xxx, imx6, x86_64
Runtime-tested on: cns3xxx, imx6, x86_64

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agokernel: bump 4.9 to 4.9.117
Koen Vandeputte [Fri, 3 Aug 2018 08:52:28 +0000 (10:52 +0200)]
kernel: bump 4.9 to 4.9.117

Refreshed all patches

Compile-tested on: ar71xx
Runtime-tested on: ar71xx

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agosdk: include arch/arm/ Linux includes along with arch/arm64/ ones
Jo-Philipp Wich [Fri, 3 Aug 2018 11:45:27 +0000 (13:45 +0200)]
sdk: include arch/arm/ Linux includes along with arch/arm64/ ones

The Linux headers on arm64 architectures contain references to common
arch/arm/ headers which were not bundled by the SDK so far.

Check if we're packing the SDK for an arm64 target and if we do, also
include arch/arm headers as well.

Fixes FS#1725.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoiproute2: cake: make gso/gro splitting configurable
Kevin Darbyshire-Bryant [Sun, 29 Jul 2018 13:56:03 +0000 (14:56 +0100)]
iproute2: cake: make gso/gro splitting configurable

This patch makes sch_cake's gso/gro splitting configurable
from userspace.

To disable breaking apart superpackets in sch_cake:

tc qdisc replace dev whatever root cake no-split-gso

to enable:

tc qdisc replace dev whatever root cake split-gso

Signed-off-by: Toke Høiland-Jørgensen <toke@toke.dk>
Signed-off-by: Dave Taht <dave.taht@gmail.com>
[pulled from netdev list - no API/ABI change]
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
6 years agokmod-sched-cake: bump to 20180728 optional gso split
Kevin Darbyshire-Bryant [Sat, 28 Jul 2018 20:11:14 +0000 (21:11 +0100)]
kmod-sched-cake: bump to 20180728 optional gso split

Follow upstream kernel patch that restores always splitting gso packets
by default whilst making the option configurable from (tc) userspace.

No ABI/API change

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
6 years agobusybox: update to 1.29.2
Hannu Nyman [Tue, 31 Jul 2018 13:55:42 +0000 (16:55 +0300)]
busybox: update to 1.29.2

* Update busybox to 1.29.2
* refresh default config
* remove upstreamed patches

Config refreshed with
  cd config/
  ../convert_menuconfig.pl ../../../../build_dir/target-arm_cortex-a9+vfpv3_musl_eabi/busybox-1.29.2
  cd ..
  ./convert_defaults.pl < ../../../build_dir/target-arm_cortex-a9+vfpv3_musl_eabi/busybox-1.29.2/.config > Config-defaults.in

Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
Tested-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agoiperf: bump to 2.0.12
Koen Vandeputte [Thu, 2 Aug 2018 12:40:03 +0000 (14:40 +0200)]
iperf: bump to 2.0.12

Fixes the annoying 'feature' were TTL was set to "1" by default ..
Users had to specify -T manually to test outside the own network.

2.0.12 change set (as of June 25th 2018)

o Change the unicast TTL default value from 1 to the system default (to be compatable with previous versions.) Multicast still defaults to 1.
o adpative formatting bug fix: crash occurs when values exceed 1 Tera. Add support for Tera and Peta and eliminate the potential crash condition
o configure default compile to include isochronous support (use configure --disable-isochronous to remove support)
o replace 2.0.11's --vary-load option with a more general -b option to include <mean>,<stdev>, e.g. -b 100m,40m, which will pull from a log normal distribution every 0.1 seconds
o fixes for windows cross compile (using mingw32)
o compile flags of -fPIE for android
o configure --enable-checkprograms to compile ancillary binaries used to test things such as delay, isoch, pdf generation
o compile tests when trying to use 64b seq numbers on a 32b platform
o Fix GCC ver 8 warnings

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agobrcm63xx: switch to 4.14
Jonas Gorski [Wed, 1 Aug 2018 10:18:13 +0000 (12:18 +0200)]
brcm63xx: switch to 4.14

Kernel 4.14 support has been present for quite some time, so let's give
it wider testing.

Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
6 years agobrcm63xx: drop b43 from devices with unsupported wifi
Jonas Gorski [Wed, 1 Aug 2018 10:14:39 +0000 (12:14 +0200)]
brcm63xx: drop b43 from devices with unsupported wifi

The internal wifi isn't supported, so drop b43 from devices that only
have it.

Fixes FS#1698
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
6 years agobrcm63xx: copy SR102 support patch to 4.14 as well
Jonas Gorski [Wed, 1 Aug 2018 10:04:39 +0000 (12:04 +0200)]
brcm63xx: copy SR102 support patch to 4.14 as well

It seems to have been missed when the patch was accepted.

Fixes: d59126040701 ("brcm63xx: initial support for Sky SR102 router")
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
6 years agobrcm63xx: drop bogus SPROM section from SR102
Jonas Gorski [Wed, 1 Aug 2018 09:39:57 +0000 (11:39 +0200)]
brcm63xx: drop bogus SPROM section from SR102

The integrated SoC wifi will likely never be supported, and if,
it will go through DT with the sprom contents there.

Fixes: d59126040701 ("brcm63xx: initial support for Sky SR102 router")
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
6 years agobase-files: fix HOME_URL replace
Chen Minqiang [Thu, 2 Aug 2018 01:10:15 +0000 (09:10 +0800)]
base-files: fix HOME_URL replace

Signed-off-by: Chen Minqiang <ptpt52@gmail.com>
6 years agoar71xx: add support for TP-LINK CPE510 V2.0
Andrew Cameron [Wed, 27 Jun 2018 03:09:03 +0000 (22:09 -0500)]
ar71xx: add support for TP-LINK CPE510 V2.0

Adds Support for the TP-LINK CPE510 V2.0 by TP-Link.
The hardware is almost the same as the CPE510 V1.0
Follow the same processes as for the CPE510 V1.0

Signed-off-by: Andrew Cameron <apcameron@softhome.net>
6 years agoRevert "ar71xx: ag71xx: Prevent kernel oops for board def"
Jo-Philipp Wich [Wed, 1 Aug 2018 20:33:49 +0000 (22:33 +0200)]
Revert "ar71xx: ag71xx: Prevent kernel oops for board def"

This reverts commit 7a3e133751870a3431e2430386a77a002eb24117.

This change reportedly breaks connectivity on some ar71xx devices, so
revert it for now.

Ref: https://github.com/openwrt/openwrt/pull/1217#issuecomment-409708087
Reported-by: Hannu Nyman <hannu.nyman@iki.fi>
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoRevert "ar71xx: ag71xx: Add connect message: fixed phy"
Jo-Philipp Wich [Wed, 1 Aug 2018 20:31:34 +0000 (22:31 +0200)]
Revert "ar71xx: ag71xx: Add connect message: fixed phy"

This reverts commit 0b9f4e880807e3cfd22d12b929202e1edcdc577c.

This change reportedly breaks connectivity on some ar71xx devices, so
revert it for now.

Ref: https://github.com/openwrt/openwrt/pull/1217#issuecomment-409708087
Reported-by: Hannu Nyman <hannu.nyman@iki.fi>
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoRevert "ar71xx: ag71xx_phy: Fix compilation for debug messages"
Jo-Philipp Wich [Wed, 1 Aug 2018 20:31:04 +0000 (22:31 +0200)]
Revert "ar71xx: ag71xx_phy: Fix compilation for debug messages"

This reverts commit 2655fbe8efb2a596aa8db92cffa817e3bf0c64ea.

The patch introduces syntax errors, revert it for now.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agogdb: bump to 8.1.1
Koen Vandeputte [Wed, 1 Aug 2018 11:32:22 +0000 (13:32 +0200)]
gdb: bump to 8.1.1

GDB 8.1.1 brings the following fixes and enhancements over GDB 8.1:

 * PR gdb/22824 (misleading description of new rbreak Python function in GDB 8.1 NEWS file)
 * PR gdb/22849 (ctrl-c doesn't work in extended-remote)
 * PR gdb/22907 ([Regression] gdbserver doesn't work with filename-only binaries)
 * PR gdb/23028 (inconsistent disassemble of vcvtpd2dq)
 * PR gdb/23053 (Fix -D_GLIBCXX_DEBUG gdb-add-index regression)
 * PR gdb/23127 ([AArch64] GDB cannot be used for debugging software that uses high Virtual Addresses)
 * PR server/23158 (gdbserver no longer functional on Windows)
 * PR breakpoints/23210 ([8.1/8.2 Regression] Bogus Breakpoint address adjusted from 0xf7fe7dd3 to 0xfffffffff7fe7dd3)

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agotoolchain/gdb: bump to 8.1.1
Koen Vandeputte [Wed, 1 Aug 2018 11:39:03 +0000 (13:39 +0200)]
toolchain/gdb: bump to 8.1.1

GDB 8.1.1 brings the following fixes and enhancements over GDB 8.1:

 * PR gdb/22824 (misleading description of new rbreak Python function in GDB 8.1 NEWS file)
 * PR gdb/22849 (ctrl-c doesn't work in extended-remote)
 * PR gdb/22907 ([Regression] gdbserver doesn't work with filename-only binaries)
 * PR gdb/23028 (inconsistent disassemble of vcvtpd2dq)
 * PR gdb/23053 (Fix -D_GLIBCXX_DEBUG gdb-add-index regression)
 * PR gdb/23127 ([AArch64] GDB cannot be used for debugging software that uses high Virtual Addresses)
 * PR server/23158 (gdbserver no longer functional on Windows)
 * PR breakpoints/23210 ([8.1/8.2 Regression] Bogus Breakpoint address adjusted from 0xf7fe7dd3 to 0xfffffffff7fe7dd3)

GDB 8.1 includes the following changes and enhancements:

  * Breakpoints on C++ functions are now set on all scopes by default
    ("wild" matching);

  * Support for inserting breakpoints on functions marked with C++ ABI tags;

  * Target floating-point arithmetic emulation during expression evaluation
    (requires MPFR 3.1 or later);

  * Various Python Scripting enhancements;

  * Improved Rust support; in particular, Trait objects can now be inspected
    when debugging Rust code;

  * GDB no longer makes assumptions about the type of symbols without
    debugging information to avoid producing erroneous and often confusing
    results;

  * The 'enable' and 'disable' commands now accept a range of breakpoint
    locations;

  * New 'starti' command to start the program at the first instruction;

  * New 'rbreak' command to insert a number of breakpoints via a regular
    expression pattern (requires Python);

  * The 'ptype' command now supports printing the offset and size of
    the fields in a struct;

  * The 'gcore' command now supports dumping all the memory mappings
    ('-a' command-line option);

  * New shortcuts for TUI Single-Key mode: 'i' for stepi, and 'o' for nexti;

  * GDBserver enhancements:
    ** Support for transmitting environment variables to GDBserver;
    ** Support for starting inferior processes with a specified initial
       working directory;
    ** On Unix systems, support for globbing expansion and variable
       substitution of inferior command-line arguments;

  * Various completion enhancements;

  * The command used to compile and inject code with the 'compile' command
    is now configurable;

  * New '--readnever' command-line option to speed the GDB startup when
    debugging information is not needed;

  * Support for the following new native configurations:
    ** FreeBSD/aarch64 (aarch64*-*-freebsd*);
    ** FreeBSD/arm (arm*-*-freebsd*);

  * Support for the following new targets:
    ** FreeBSD/aarch64 (aarch64*-*-freebsd*);
    ** FreeBSD/arm (arm*-*-freebsd*);
    ** OpenRISC ELF (or1k*-*-elf)

  * Removed support for the following targets and native configurations:
    ** Solaris2/x86 (i?86-*-solaris2.[0-9]);
    ** Solaris2/sparc (sparc*-*-solaris2.[0-9]);

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agoadb: added patch for openssl 1.1.0 compatibility
Eneas U de Queiroz [Fri, 25 May 2018 14:43:32 +0000 (11:43 -0300)]
adb: added patch for openssl 1.1.0 compatibility

Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
6 years agonftables: allow to build with json support
Rosy Song [Tue, 31 Jul 2018 09:09:34 +0000 (17:09 +0800)]
nftables: allow to build with json support

Signed-off-by: Rosy Song <rosysong@rosinson.com>
6 years agoar71xx: cr3000: cleanup board definition
Daniel F. Dickinson [Sun, 29 Jul 2018 11:40:01 +0000 (07:40 -0400)]
ar71xx: cr3000: cleanup board definition

1) Add comments so it's clear why we did things; this may prevent
   someone (e.g. me) from sinking time into fixing things that
   aren't broken and/or were done for reason.

2) Drop mdio 0 probe/register; we don't use ag1xx mdio bus 0.

3) Cosmetic reording of some code (tested) that makes the defintion
   more clear.

Signed-off-by: Daniel F. Dickinson <cshored@thecshore.com>
6 years agoar71xx: cr3000: Use correct company name
Daniel F. Dickinson [Sun, 29 Jul 2018 11:17:19 +0000 (07:17 -0400)]
ar71xx: cr3000: Use correct company name

It's 'PowerCloud Systems' not 'PowerCloud'.

Signed-off-by: Daniel F. Dickinson <cshored@thecshore.com>
6 years agoar71xx: cr3000: Drop support for defunct cloud
Daniel F. Dickinson [Wed, 25 Jul 2018 23:18:07 +0000 (19:18 -0400)]
ar71xx: cr3000: Drop support for defunct cloud

The CR3000 stock firmware is now irrelevant as it required a now defunct
cloud service.  Therefore only build images that use the entire flash
(overwriting stock firwmare-specific partitions that no longer matter),
previously called 'nocloud' images.

Signed-off-by: Daniel F. Dickinson <cshored@thecshore.com>
6 years agoath79: Forward port support for CR3000
Daniel F. Dickinson [Sat, 28 Jul 2018 12:57:22 +0000 (08:57 -0400)]
ath79: Forward port support for CR3000

The PowerCloud Systems CR3000 was a cloud-managed CPE for a now defunct
NaaS offering.  It was previously supported under the ar71xx branch and
this forward ports that support with some notable differences:

1) Since reverting to stock firmware is now irrelevant there is is only
a single openwrt image generated which uses the entire flash rather than
preserving PowerCloud-specific partitions that are unneeded to openwrt--
those partitions will be erased and used by the openwrt image.

2) Rather than use a non-standard probe order for the ethernet devices,
this image uses a set of 'ip link set ethX name ethY' commands very early
in preinit (before the network is used at all), in order to have the the
switch and Wan use the same ethernet names as in previous images.

3) /etc/config/wireless will need to be regenerated as the path to the
wireless device has changed due to differences in ath79 DT for ar93x
compared to ar71xx images.

4) eth0 is wan and eth1 is lan (switch)

Signed-off-by: Daniel F. Dickinson <cshored@thecshore.com>
6 years agokernel: bump 4.9 to 4.9.116
Koen Vandeputte [Tue, 31 Jul 2018 11:30:14 +0000 (13:30 +0200)]
kernel: bump 4.9 to 4.9.116

Refreshed all patches

Remove upstreamed patches.
- 103-MIPS-ath79-fix-register-address-in-ath79_ddr_wb_flus.patch
- 403-mtd_fix_cfi_cmdset_0002_status_check.patch
- 001-4.11-01-mtd-m25p80-consider-max-message-size-in-m25p80_read.patch
- 001-4.15-08-bcm63xx_enet-correct-clock-usage.patch
- 001-4.15-09-bcm63xx_enet-do-not-write-to-random-DMA-channel-on-B.patch
- 900-gen_stats-fix-netlink-stats-padding.patch

Introduce a new backported patch to address ext4 breakage, introduced in 4.9.112
- backport-4.9/500-ext4-fix-check-to-prevent-initializing-reserved-inod.patch

This patch has been slightly altered to compensate for a new helper function
introduced in later kernels.

Also add ARM64_SSBD symbol to ARM64 targets still running kernel 4.9

Compile-tested on: ar71xx, bcm2710
Runtime-tested on: ar71xx

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agosdk: bundle usbip userspace sources
Jo-Philipp Wich [Wed, 1 Aug 2018 07:11:17 +0000 (09:11 +0200)]
sdk: bundle usbip userspace sources

Bundle the usbip utility sources shipped with the Linux kernel tree in
order to allow the usbip packages from the package feed to build within
the OpenWrt SDK.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoiwinfo: update to latest Git HEAD
Jo-Philipp Wich [Tue, 31 Jul 2018 18:39:13 +0000 (20:39 +0200)]
iwinfo: update to latest Git HEAD

a514139 build: compile with -ffunction-sections, -fdata-sections and LTO
3c30b17 wl: only invoke nvram executable if it exists
65b8333 Revert "build: compile with -ffunction-sections, -fdata-sections and LTO"

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoinclude/feeds.mk: fix distfeeds.conf without per-feed repos
Bjørn Mork [Tue, 31 Jul 2018 12:01:12 +0000 (14:01 +0200)]
include/feeds.mk: fix distfeeds.conf without per-feed repos

commit 514a4b3e1b4e4 ("include/feeds.mk: rework generation of opkg
distfeeds.conf") made the per-feed "base" repo unconditional, making
the default configuration fail when PER_FEED_REPO is disabled:

 root@wrt1900ac-1:~# cat /etc/opkg/distfeeds.conf
 src/gz openwrt_core http://openwrt.mork.no/18.06.0/targets/mvebu/cortexa9/packages
 src/gz openwrt_base http://openwrt.mork.no/18.06.0/packages/arm_cortex-a9_vfpv3/base
 root@wrt1900ac-1:~# opkg update
 Downloading http://openwrt.mork.no/18.06.0/targets/mvebu/cortexa9/packages/Packages.gz
 Updated list of available packages in /var/opkg-lists/openwrt_core
 Downloading http://openwrt.mork.no/18.06.0/targets/mvebu/cortexa9/packages/Packages.sig
 Signature check passed.
 Downloading http://openwrt.mork.no/18.06.0/packages/arm_cortex-a9_vfpv3/base/Packages.gz
 *** Failed to download the package list from http://openwrt.mork.no/18.06.0/packages/arm_cortex-a9_vfpv3/base/Packages.gz

 Collected errors:
  * opkg_download: Failed to download http://openwrt.mork.no/18.06.0/packages/arm_cortex-a9_vfpv3/base/Packages.gz, wget returned 8.

Cc: Matthias Schiffer <mschiffer@universe-factory.net>
Fixes: 514a4b3e1b4e ("include/feeds.mk: rework generation of opkg distfeeds.conf")
Signed-off-by: Bjørn Mork <bjorn@mork.no>
[whitespace/indentation fix]
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agokernel: add missing ARM64_SSBD symbol
Stijn Tintel [Tue, 31 Jul 2018 09:19:18 +0000 (12:19 +0300)]
kernel: add missing ARM64_SSBD symbol

In 4.14.57, a new symbol for Spectre v4 mitigation was introduced for
ARM64. Add this symbol to all ARM64 targets using kernel 4.14.

This mitigates CVE-2018-3639 on ARM64.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
6 years agomac80211: brcmfmac: backport patch for per-firmware features
Rafał Miłecki [Tue, 31 Jul 2018 07:44:19 +0000 (09:44 +0200)]
mac80211: brcmfmac: backport patch for per-firmware features

This allows driver to support features that can't be dynamically
discovered.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
6 years agokernel: remove duplicate #define's in at803x Ethernet PHY driver
Daniel Golle [Tue, 31 Jul 2018 03:17:52 +0000 (05:17 +0200)]
kernel: remove duplicate #define's in at803x Ethernet PHY driver

AT803X_REG_CHIP_CONFIG and AT803X_BT_BX_REG_SEL have been defined
upstream by commit f62265b53ef3 ("at803x: double check SGMII side autoneg")
An existing local patch then added those exact same defines again which
isn't necessary, so remove them.

Fixes: f791fb4af450 ("kernel: add linux 4.9 support")
Fixes: b3f95490b9be ("kernel: generic: Add kernel 4.14 support")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agokernel: re-add patch for AT8032 Ethernet PHY
Daniel Golle [Tue, 31 Jul 2018 02:50:38 +0000 (04:50 +0200)]
kernel: re-add patch for AT8032 Ethernet PHY

The patch was wrongly removed by a kernel version bump to 4.9.105 in
the believe that it was merged upstream thow it wasn't. This lead to
unrecoverable link losses on devices which use those PHYs such as
many ubnt single-port CPEs.

Fixes: 7dca1bae82 (kernel: bump to 4.9.105)
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agokernel: add pending e1000e fixes
Stijn Tintel [Thu, 28 Jun 2018 09:44:10 +0000 (11:44 +0200)]
kernel: add pending e1000e fixes

The previous round of fixes for the 82574 chip cause an issue with
emulated e1000e devices in VMware ESXi 6.5. It also contains changes
that are not strictly necessary. These patches fix the issues introduced
in the previous series, revert the unnecessary changes to avoid
unforeseen fallout, and avoid a case where interrupts can be missed.

The final two patches of this series are already in the kernel, so no
need to include them here.

Patchwork: https://patchwork.ozlabs.org/cover/881776/

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