openwrt/staging/rmilecki.git
6 years agokernel: bump 4.14 to 4.14.70
Koen Vandeputte [Mon, 17 Sep 2018 09:52:06 +0000 (11:52 +0200)]
kernel: bump 4.14 to 4.14.70

Refreshed all patches.

Added new patch:
- 192-Revert-ubifs-xattr-Don-t-operate-on-deleted-inodes.patch

This fixes a bug introduced in upstream 4.14.68 which caused targets using
ubifs to produce file-system errors on boot, rendering them useless.

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 4.9.127
Koen Vandeputte [Mon, 17 Sep 2018 09:48:49 +0000 (11:48 +0200)]
kernel: bump 4.9 to 4.9.127

Refreshed all patches.

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

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agokernel: bump 3.18 to 3.18.122
Koen Vandeputte [Mon, 17 Sep 2018 09:47:42 +0000 (11:47 +0200)]
kernel: bump 3.18 to 3.18.122

Refreshed all patches.

Compile-tested on: adm5120
Runtime-tested on: none

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agouqmi: pass-through ipXtable to child interfaces
Daniel Golle [Sat, 15 Sep 2018 17:16:52 +0000 (19:16 +0200)]
uqmi: pass-through ipXtable to child interfaces

Allow setting specific routing tables via the ip4table and ip6table
options also when ${ifname}_4 and ${ifname}_6 child interfaces are
being created.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 years agotools/e2fsprogs: fix build with clang
Felix Fietkau [Sat, 15 Sep 2018 13:49:31 +0000 (15:49 +0200)]
tools/e2fsprogs: fix build with clang

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agoiproute2: q_cake: Also print nonat, nowash and no-ack-filter keywords
Kevin Darbyshire-Bryant [Fri, 14 Sep 2018 20:52:39 +0000 (21:52 +0100)]
iproute2: q_cake: Also print nonat, nowash and no-ack-filter keywords

Pull in latest upstream tweaks:
Similar to the previous patch for no-split-gso, the negative keywords for
'nat', 'wash' and 'ack-filter' were not printed either. Add those as well.

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
6 years agobusybox: update to 1.29.3
Hannu Nyman [Fri, 14 Sep 2018 19:07:42 +0000 (22:07 +0300)]
busybox: update to 1.29.3

Update busybox to 1.29.3, minor bugfix release

https://git.busybox.net/busybox/log/?h=1_29_3

Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
6 years agotoolchain/glibc: update to latest 2.26 commit
Hans Dedecker [Thu, 13 Sep 2018 07:26:44 +0000 (09:26 +0200)]
toolchain/glibc: update to latest 2.26 commit

c5c90b480e Fix segfault in maybe_script_execute.
174709d879 pthread_cond_broadcast: Fix waiters-after-spinning case [BZ #23538]

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agoodhcpd: enable ipv6 server mode only when it is supported
Rosy Song [Wed, 12 Sep 2018 01:21:19 +0000 (09:21 +0800)]
odhcpd: enable ipv6 server mode only when it is supported

Signed-off-by: Rosy Song <rosysong@rosinson.com>
6 years agoiproute2: q_cake: Add printing of no-split-gso option
Kevin Darbyshire-Bryant [Wed, 12 Sep 2018 07:24:37 +0000 (08:24 +0100)]
iproute2: q_cake: Add printing of no-split-gso option

When the GSO splitting was turned into dual split-gso/no-split-gso options,
the printing of the latter was left out. Add that, so output is consistent
with the options passed

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
6 years agomac80211: brcmfmac: backport CYW89342 support & fixes from 4.20
Rafał Miłecki [Wed, 12 Sep 2018 06:40:03 +0000 (08:40 +0200)]
mac80211: brcmfmac: backport CYW89342 support & fixes from 4.20

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
6 years agokmod-sched-cake: fix 6in4/gso performance issue
Kevin Darbyshire-Bryant [Wed, 12 Sep 2018 04:10:55 +0000 (05:10 +0100)]
kmod-sched-cake: fix 6in4/gso performance issue

Bump to latest upstream cake:

Add workaround for wrong skb->mac_len values after splitting GSO

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
6 years agonetifd: update to latest HEAD
Florian Fainelli [Wed, 12 Sep 2018 00:19:51 +0000 (17:19 -0700)]
netifd: update to latest HEAD

0059335c5b60 CMakeList: Check that compiler supports -Wimplicit-fallthrough

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
6 years agotools/expat: fix docbook2man error on some systems
Marko Ratkaj [Mon, 10 Sep 2018 09:04:36 +0000 (11:04 +0200)]
tools/expat: fix docbook2man error on some systems

On some systems (Gentoo) configure stage fails because of docbook2man
working with SGML rather than with XML. We don't need xmlwf man pages so
we disable this.

Signed-off-by: Marko Ratkaj <marko.ratkaj@sartura.hr>
6 years agowireguard: bump to 0.0.20180910
Jason A. Donenfeld [Tue, 11 Sep 2018 00:59:16 +0000 (18:59 -0600)]
wireguard: bump to 0.0.20180910

* curve25519: arm: do not modify sp directly
* compat: support neon.h on old kernels
* compat: arch-namespace certain includes
* compat: move simd.h from crypto to compat since it's going upstream

This fixes a decent amount of compat breakage and thumb2-mode breakage
introduced by our move to Zinc.

* crypto: use CRYPTOGAMS license

Rather than using code from OpenSSL, use code directly from AndyP.

* poly1305: rewrite self tests from scratch
* poly1305: switch to donna

This makes our C Poly1305 implementation a bit more intensely tested and also
faster, especially on 64-bit systems. It also sets the stage for moving to a
HACL* implementation when that's ready.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
6 years agokernel: add missing symbol
John Crispin [Mon, 10 Sep 2018 15:50:40 +0000 (17:50 +0200)]
kernel: add missing symbol

Signed-off-by: John Crispin <john@phrozen.org>
6 years agobase-files: /etc/services: add missing 'rpcbind' alias
Andy Walsh [Mon, 27 Aug 2018 12:19:55 +0000 (14:19 +0200)]
base-files: /etc/services: add missing 'rpcbind' alias

* add missing 'rpcbind' alias to /etc/services

Allows rpcbind to open its 111 port and be reachable via lan, this is the default behaviour.

Signed-off-by: Andy Walsh <andy.walsh44+github@gmail.com>
6 years agoramips: Fix early memory calculation for certain MIPS platforms
Tobias Wolf [Sun, 19 Aug 2018 11:04:02 +0000 (13:04 +0200)]
ramips: Fix early memory calculation for certain MIPS platforms

Kernel upstream commit 67a3ba25aa95 ("MIPS: Fix incorrect mem=X@Y handling") introduced a new issue for rt288x where "PHYS_OFFSET" is 0x0 but the calculated "ramstart" is not. As the prerequisite of custom memory map has been removed, this results in the full memory range of 0x0 - 0x8000000 to be marked as reserved
for this platform.

This patch adds the originally intended prerequisite again.

Signed-off-by: Tobias Wolf <dev-NTEO@vplace.de>
6 years agoar71xx: Skip more hashed blocks for OM2P(-HS) 64k variant
Sven Eckelmann [Mon, 3 Sep 2018 07:58:42 +0000 (09:58 +0200)]
ar71xx: Skip more hashed blocks for OM2P(-HS) 64k variant

The OM2P(-HS)v4 got a variant which uses a slightly different flash. The
standard versions used a flash with 256KB blocks which is no longer
available. The replacement flash uses a flash with 64K blocks.

The padding for the image rootfs is already for 64K and 256K and thus can
be flashed on the device without any problems. Unfortunately, the
bootloader will check $rootfs_size (rounded down to the nearest 64k block)
minus 1x 64k. But it is now possible that the new JFFS2 rootfs_data starts
even earlier and modifies the checked region. The check will then fail and
the backup image (when available) will be booted.

Just setting it to the same number of skipped blocks as other 64K models
avoids this problem.

Signed-off-by: Sven Eckelmann <sven.eckelmann@openmesh.com>
6 years agotools/expat: Update to 2.2.6
Daniel Engberg [Wed, 29 Aug 2018 06:58:44 +0000 (08:58 +0200)]
tools/expat: Update to 2.2.6

Update (lib)expat to 2.2.6

Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
6 years agotools/e2fsprogs: Update to 1.44.4
Daniel Engberg [Wed, 29 Aug 2018 07:01:54 +0000 (09:01 +0200)]
tools/e2fsprogs: Update to 1.44.4

Update e2fsprogs to 1.44.4

Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
6 years agoath79: ubnt-xm: add rssileds package
Lech Perczak [Wed, 5 Sep 2018 17:40:16 +0000 (19:40 +0200)]
ath79: ubnt-xm: add rssileds package

In order to make RSSI indicator on the device work out of box,
include "rssileds" package in per-device rootfs image by default
for Ubiquiti XM family.

Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
6 years agoath79: ubnt-xm: create RSSI monitor on wlan0
Lech Perczak [Wed, 5 Sep 2018 17:40:06 +0000 (19:40 +0200)]
ath79: ubnt-xm: create RSSI monitor on wlan0

When mapping for RSSI LEDs was defined for interface wlan0 on
Ubiquiti XM family, the mapping for rssileds monitor was omitted
by mistake. Therefore create the mapping, so RSSI LEDs work without
additional configuration, after starting rssileds service.

Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
6 years agoath79: add support for Buffalo WZR-HP-AG300H
Bernhard Frauendienst [Sat, 18 Aug 2018 16:28:02 +0000 (18:28 +0200)]
ath79: add support for Buffalo WZR-HP-AG300H

Buffalo WZR-HP-AG300H is a dual band router based on
Qualcom Atheros AR7161 rev 2

Specification:
- 680 MHz CPU (Qualcomm Atheros AR7161)
- 128 MiB RAM (2x Samsung K4H511638G-LCCC)
- 32 MiB Flash (2x Winbond 25Q128BVFG)
- WiFi 5 GHz a/n (Atheros AR9220)
- WiFi 2.4 GHz b/g/n (Atheros AR9223)
- 1000Base-T WAN (Atheros AR7161)
- 4x 1000Base-T Switch (Atheros AR8316)
- 1x USB 2.0
- 3 Buttons (AOSS/WPS, Reset, USB Eject)
- 2 Slide switches (Router (on/off/auto), Movie Engine (on/off))
- 9 LEDs (Power green, WLAN 2GHz green, WLAN 2GHz amber,
    WLAN 5GHz green, WLAN 5GHz LED amber, Router green,
    Diag red, Movie Engine blue, USB green)

It is already supported by the ar71xx target.

For more information on the device visit the wiki:
<https://openwrt.org/toh/buffalo/wzr-hp-ag300h>

Serial console:
- The UART Header is next to Movie Engine Switch.
- Pinout is RX - TX - GND - 3.3V (Square Pad is 3.3V)
- The Serial setting is 115200-8-N-1.

Installation of OpenWRT from vendor firmware:
- Connect to the Web-interface at http://192.168.11.1
- Go to “Administration” → “Firmware Upgrade”
- Upload the OpenWrt factory image

Tested:
- Ethernet (LAN, WAN)
- WiFi
- Installation
  - via TFTP rescue
  - via factory image
    - on firmware v1.77 (28-05-2012)
    - on pro firmware v24SP2 r30356 (26-03-2018)
  - via sysupgrade from ar71xx
    (wlan devices don't work because of new names)
  - via sysupgrade from itself
- Buttons
- LEDS
- USB (Power control and device recognition)

Signed-off-by: Bernhard Frauendienst <openwrt@nospam.obeliks.de>
6 years agokernel: add driver for virtual mtd_concat devices
Bernhard Frauendienst [Sat, 25 Aug 2018 10:48:54 +0000 (12:48 +0200)]
kernel: add driver for virtual mtd_concat devices

Some systems require multiple flash chips to be concatenated and read as
a single mtd device. The ar71xx target provides custom code to create
such mtdconcat devices. When porting devices to ath79, however, there is
no way to create such devices from within the device tree.

This commit adds a driver for creating virtual mtd-concat devices to the
ath79 target. Nodes must have a compatible = "virtual,mtd-concat" line,
and define a list of devices to concat in the 'devices' property,
for example:

flash {
compatible = "virtual,mtd-concat";

devices = <&flash0 &flash1>;
};

The driver is added to the very end of the mtd Makefile to increase the
likelyhood of all child devices already being loaded at the time of
probing, preventing unnecessary deferred probes which might in turn
cause other problems (like failure to load MAC addresses from art because
the partitions are not loaded yet).

Signed-off-by: Bernhard Frauendienst <openwrt@nospam.obeliks.de>
6 years agoath79: add support for ELECOM WRC-300GHBK2-I
INAGAKI Hiroshi [Mon, 10 Sep 2018 07:30:28 +0000 (09:30 +0200)]
ath79: add support for ELECOM WRC-300GHBK2-I

ELECOM WRC-300GHBK2-I is a 2.4 GHz wireless router, based on Qualcomm
Atheros QCA9563.

Specification:

- Qualcomm Atheros QCA9563
- 64 MB of RAM (DDR2)
- 8 MB of Flash (SPI-NOR)
- 2T2R 2.4 GHz wifi
  - SoC internal
- 5x 10/100/1000 Mbps Ethernet
- 3x LEDs, 4x keys(connected to GPIO: 3x)
- UART header on PCB
  - TX, GND, RX, Vcc from ethernet port side
  - 115200n8

Flash instruction using factory image:

1. Boot the WRC-300GHBK2-I normaly and connect the computer to its
LAN port
2. Access to "http://192.168.2.1/" and open firmware update page
("ファームウェア更新 手動更新(アップデート)")
3. Select the OpenWrt factory image and click apply ("適用") button
to perform firmware update
4. On the (initramfs) factory image, execute "mtd erase firmware" to
erase stock firmware and execute sysupgrade with squashfs-sysupgrade
image for WRC-300GHBK2-I
5. Wait ~150 seconds to complete flashing

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
6 years agofirmware-utils: mkedimaximg: add "big-endianness mode" option
INAGAKI Hiroshi [Mon, 10 Sep 2018 07:29:39 +0000 (09:29 +0200)]
firmware-utils: mkedimaximg: add "big-endianness mode" option

In addition to the default little-endianness format, I added a mode
so that we can generate firmware with big-endianness format.

example: ELECOM WRC-300GHBK2-I (QCA9563)

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
6 years agokernel: backport i2c-gpio working over slow can_sleep GPIOs
Martin Schiller [Wed, 29 Aug 2018 13:06:21 +0000 (15:06 +0200)]
kernel: backport i2c-gpio working over slow can_sleep GPIOs

Signed-off-by: Martin Schiller <ms@dev.tdt.de>
6 years agokernel: backport rtc ds1308 support to 4.9
Martin Schiller [Wed, 29 Aug 2018 09:13:42 +0000 (11:13 +0200)]
kernel: backport rtc ds1308 support to 4.9

Signed-off-by: Martin Schiller <ms@dev.tdt.de>
6 years agolibpcap: patch to add limits.h to pcap-usb-linux.c
Eneas U de Queiroz [Wed, 1 Aug 2018 12:49:35 +0000 (12:49 +0000)]
libpcap: patch to add limits.h to pcap-usb-linux.c

The sender domain has a DMARC Reject/Quarantine policy which disallows
sending mailing list messages using the original "From" header.

To mitigate this problem, the original message has been wrapped
automatically by the mailing list software.
This is an upstream-applied patch that fixes 'PATH_MAX' and 'NAME_MAX'
undeclared when compiling on musl with CONFIG_PCAP_HAS_USB.

[aafa351] pcap-usb-linux.c: add missing limits.h for musl systems.

Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
6 years agomvebu: Replace RTC initialization patch with upstreamed version
Rosen Penev [Sun, 9 Sep 2018 04:23:57 +0000 (21:23 -0700)]
mvebu: Replace RTC initialization patch with upstreamed version

While we're at it, rename the patches to their proper git format-patch
name.

Tested on a Turris Omnia.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
6 years agousbutils: Update usb.ids to 0.315
Rosen Penev [Mon, 10 Sep 2018 01:55:43 +0000 (18:55 -0700)]
usbutils: Update usb.ids to 0.315

Referencing the version instead of revision should fix uscan.

Tested on Turria Omnia.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
6 years agoath79: add support for WD My Net Wi-Fi Range Extender
Christian Lamparter [Tue, 4 Sep 2018 21:27:02 +0000 (23:27 +0200)]
ath79: add support for WD My Net Wi-Fi Range Extender

This patch ports over support for the device from ar71xx.

SOC: AR7370 (Wasp - AR9344 rev2 0001974c)
RAM: Winbond W9725G6KB-25 32MiB
FLASH: Winbond 25Q64FVSIG 8MiB
WLAN: AR9380 Dual-Band 802.11abgn 3x3:3
INPUT: WPS, RESET button (hardware on/off toggle button)
LED:    Power, LAN, WiFi, 3 RSSI-Leds (low, medium, high)
Serial: Header Next to the winbond flash chip (labeld JP1)
Pinout is GND - NC - RX - TX - 3V3 (JP1)
The Serial setting is 115200-8-N-1.

- Installation via uboot's upgrade command
  0. attach serial cable
  1. interrupt uboot and enter "upgrade code.bin" into
      the u-boot prompt
ar7240> upgrade code.bin

  2. rename openwrt...sysupgrade.bin to code.bin on PC
  3. run a tftp-client on the PC
 (shell)$ tftp 192.168.1.230
 binary
 put code.bin
  4. wait for the device to finish

[...]
Copy buff to Flash from 0x9f040000 length 0x79f000
Copy to Flash... write addr: 9f040000
done

  5.  enter "go" in the u-boot prompt
ar7240> go

- TFTP ramdisk image boot from the uboot prompt
  (tftp server defaults to serverip 192.168.1.254)
=> tftpboot 81000000 initramfs.bin
=> bootm

Tested and working:
- LEDs
- Buttons
- Ethernet
- Wi-Fi
- OpenWRT sysupgrade

For flashing and debricking information see:
<https://openwrt.org/toh/wd/rext>

Users coming from ar71xx can use sysupgrade too. But I highly
advise to no save the old configuration and start from a clean
state.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
6 years agoath79: port cybertan_part from ar71xx
Christian Lamparter [Tue, 4 Sep 2018 21:27:01 +0000 (23:27 +0200)]
ath79: port cybertan_part from ar71xx

This patch ports the cybertan_part code from ar71xx and converts the
driver to a DT-supported mtd parser. As a result, it will no longer
add the u-boot, nvram and art partitions, which were never part of
the special Cybertan header.

Instead these partitions have to be specified in the DT, which has the
upside of making it possible to add properties (i.e.: read-only), labels
and references to these important partitions.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
6 years agomac80211: add iw command wrapper with error logging
Rafał Miłecki [Tue, 4 Sep 2018 13:20:34 +0000 (15:20 +0200)]
mac80211: add iw command wrapper with error logging

Currently it's close to impossible to tell what part of mac80211 setup
went wrong. Errors logged into system log look like this:
radio0 (6155): command failed: No error information (-524)
radio0 (6155): command failed: Not supported (-95)
radio0 (6155): command failed: I/O error (-5)
radio0 (6155): command failed: Too many open files in system (-23)

With this commit change it's getting clear:
command failed: No error information (-524)
Failed command: iw dev wlan0 del
command failed: Not supported (-95)
Failed command: iw phy phy0 set antenna_gain 0
command failed: I/O error (-5)
Failed command: iw phy phy0 set distance 0
command failed: Too many open files in system (-23)
Failed command: iw phy phy0 interface add wlan0 type __ap

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
6 years agohostapd: Fix compile with OpenSSL 1.1.0 + no deprecated APIs
Rosen Penev [Sun, 2 Sep 2018 02:06:54 +0000 (19:06 -0700)]
hostapd: Fix compile with OpenSSL 1.1.0 + no deprecated APIs

Patch was accepted upsteam:

https://w1.fi/cgit/hostap/commit/?id=373c796948599a509bad71695b5b72eef003f661

Signed-off-by: Rosen Penev <rosenp@gmail.com>
6 years agozynq: rename zedboard with correct manufacturer
Luis Araneda [Sat, 1 Sep 2018 05:36:09 +0000 (02:36 -0300)]
zynq: rename zedboard with correct manufacturer

The board is made by Avnet, so rename it accordingly as
upstream has done it

Also move the device to maintain alphabetic order

Signed-off-by: Luis Araneda <luaraneda@gmail.com>
6 years agozynq: add sdcard image support
Luis Araneda [Tue, 4 Sep 2018 01:40:02 +0000 (22:40 -0300)]
zynq: add sdcard image support

Implement the generation of sdcard images with ext4 or
squashfs + f2fs overlay, but only enable the latter
automatically

Additionally, add mkf2fs and e2fsprogs to default packages
to manipulate ext4 and f2fs filesystems

Finally, disable the automatic generation of initramfs
and rootfs.tar.gz images, as they are no longer required
(they can still be selected in menuconfig)

Signed-off-by: Luis Araneda <luaraneda@gmail.com>
6 years agozynq: enable kernel overlay and f2fs support
Luis Araneda [Tue, 4 Sep 2018 01:40:01 +0000 (22:40 -0300)]
zynq: enable kernel overlay and f2fs support

Select configs to enable build-in F2FS support
and use the overlayfs provided by generic config

Signed-off-by: Luis Araneda <luaraneda@gmail.com>
6 years agozynq: remove bootargs from kernel config
Luis Araneda [Tue, 4 Sep 2018 01:40:00 +0000 (22:40 -0300)]
zynq: remove bootargs from kernel config

Use options from generic instead of target-specific,
as the bootargs are now passed by the bootloader
and can be device-specific

Signed-off-by: Luis Araneda <luaraneda@gmail.com>
6 years agouboot-zynq: use a file to modify the default environment
Luis Araneda [Tue, 4 Sep 2018 01:39:59 +0000 (22:39 -0300)]
uboot-zynq: use a file to modify the default environment

Follow the strategy of other targets and create a
default environment file, uEnv.txt, to configure the
behavior of U-Boot.
For now, use it to pass bootargs to the kernel

Signed-off-by: Luis Araneda <luaraneda@gmail.com>
6 years agouboot-zynq: copy U-Boot images to STAGING_DIR
Luis Araneda [Tue, 4 Sep 2018 01:39:58 +0000 (22:39 -0300)]
uboot-zynq: copy U-Boot images to STAGING_DIR

Create a directory inside STAGING_DIR and copy U-Boot
output images, so they can be used later when creating the
sdcard image

Additionally, like others targets, override the default
install method to avoid copying the images to bin directory

Signed-off-by: Luis Araneda <luaraneda@gmail.com>
6 years agouboot-zynq: automatically select the appropriate variant
Luis Araneda [Tue, 4 Sep 2018 01:39:57 +0000 (22:39 -0300)]
uboot-zynq: automatically select the appropriate variant

Select the U-Boot variant automatically based on the
current selected device, and hide the package from
menuconfig

Signed-off-by: Luis Araneda <luaraneda@gmail.com>
6 years agouboot-zynq: remove ZC706 board
Luis Araneda [Tue, 4 Sep 2018 01:39:56 +0000 (22:39 -0300)]
uboot-zynq: remove ZC706 board

The board was added when creating the target, but the
corresponding device was never defined inside the target

Signed-off-by: Luis Araneda <luaraneda@gmail.com>
6 years agodnsmasq: bump to dnsmasq 2.80test6
Hans Dedecker [Sun, 9 Sep 2018 19:46:43 +0000 (21:46 +0200)]
dnsmasq: bump to dnsmasq 2.80test6

Refresh patches

Changes since latest bump:

af3bd07 Man page typo.
d682099 Picky changes to 47b45b2967c931fed3c89a2e6a8df9f9183a5789
47b45b2 Fix lengths of interface names
2b38e38 Minor improvements in lease-tools
282eab7 Mark die function as never returning
c346f61 Handle ANY queries in context of da8b6517decdac593e7ce24bde2824dd841725c8
03212e5 Manpage typo.

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agotools/cmake: Update to 3.12.2
Daniel Engberg [Sun, 9 Sep 2018 11:17:28 +0000 (13:17 +0200)]
tools/cmake: Update to 3.12.2

Update CMake to 3.12.2

Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
6 years agocurl: Update to 7.61.1
Daniel Engberg [Sun, 9 Sep 2018 11:21:00 +0000 (13:21 +0200)]
curl: Update to 7.61.1

Update curl to 7.61.1

Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
6 years agoscripts/ipkg-build: quash error messages when conffiles.resolved is empty
Yousong Zhou [Thu, 6 Sep 2018 08:00:24 +0000 (08:00 +0000)]
scripts/ipkg-build: quash error messages when conffiles.resolved is empty

When Package/xx/conffiles only contains directories that are empty at
package time, conffiles.resolved will be missing and the following
error messages will appear in the build log.

/home/yousong/git-repo/openwrt/openwrt/scripts/ipkg-build -c -o 0 -g 0 /home/yousong/git-repo/openwrt/openwrt/build_dir/target-mips_24kc_musl/linux-malta_be/openvswitch-2.10.0/ipkg-mips_24kc/openvswitch-common /home/yousong/git-repo/openwrt/openwrt/bin/packages/mips_24kc/packages
mv: cannot stat 'CONTROL/conffiles.resolved': No such file or directory
chmod: cannot access 'CONTROL/conffiles': No such file or directory

It will not break the ipkg-build process.  The change is merely cosmetic
to not cause confusion when reading logs

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
6 years agoscripts/metadata.pm: avoid adding dup names in provides list
Yousong Zhou [Wed, 5 Sep 2018 15:29:25 +0000 (15:29 +0000)]
scripts/metadata.pm: avoid adding dup names in provides list

The need arises from building Open vSwitch kernel datapath modules, e.g.

 - kmod-openvswitch from Linux upstream
 - kmod-openvswitch-intree from openvswitch source code

where both provides virtual package "kmod-openvswitch" for userspace
packages to select and depend on

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
6 years agoodhcp6c: add client fqdn and reconfigure options
pacien [Thu, 6 Sep 2018 13:56:17 +0000 (15:56 +0200)]
odhcp6c: add client fqdn and reconfigure options

Allowing DHCPV6_CLIENT_FQDN and DHCPV6_ACCEPT_RECONFIGURE to be turned off.
Defaulting to false, former behavior remains unchanged.

Signed-off-by: pacien <pacien.trangirard@pacien.net>
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com> [PKG_RELEASE increase]
6 years agoar71xx: fix switch probing on kernel 4.14
Koen Vandeputte [Thu, 30 Aug 2018 17:09:46 +0000 (19:09 +0200)]
ar71xx: fix switch probing on kernel 4.14

The bump to 4.14 changed the way mdio probes behind switches.

While the board_info is added to the list, the code that actually inserted
the list info into the phydev structure was missing.

This resulted in non-working ethernet ports.

Re-add it to fix switch probing.
This mimics the exact behaviour as it was in kernel 4.9.

Before:

[    1.066007] switch0: Atheros AR8327 rev. 4 switch registered on ag71xx-mdio.0
[    1.073409] Atheros AR8216/AR8236/AR8316: probe of ag71xx-mdio.0:00 failed with error -22
[    1.102455] libphy: ag71xx_mdio: probed
[    1.737938] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00 [uid=004dd034, driver=Generic PHY]
[    1.747994] eth0: Atheros AG71xx at 0xb9000000, irq 4, mode:RGMII
[    2.377642] ag71xx-mdio.1: Found an AR934X built-in switch
[    2.429938] eth1: Atheros AG71xx at 0xba000000, irq 5, mode:GMII

After:

[   11.163357] libphy: Fixed MDIO Bus: probed
[   11.319898] libphy: ag71xx_mdio: probed
[   11.360844] switch0: Atheros AR8327 rev. 4 switch registered on ag71xx-mdio.0
[   12.447398] libphy: ag71xx_mdio: probed
[   13.077402] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00 [uid=004dd034, driver=Atheros AR8216/AR8236/AR8316]
[   13.088989] eth0: Atheros AG71xx at 0xb9000000, irq 4, mode:RGMII
[   13.717716] ag71xx-mdio.1: Found an AR934X built-in switch
[   13.769990] eth1: Atheros AG71xx at 0xba000000, irq 5, mode:GMII

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agogdb: bump to 8.2
Koen Vandeputte [Thu, 6 Sep 2018 12:44:17 +0000 (14:44 +0200)]
gdb: bump to 8.2

*** Changes in GDB 8.2

 Support for the following target has been added:

    RiscV ELF (riscv*-*-elf)

 Support for following targets and native configurations has been removed:

    m88k running OpenBSD (m88*-*-openbsd*)
    SH-5/SH64 ELF (sh64-*-elf*)
    SH-5/SH64 (sh*)
    SH-5/SH64 running GNU/Linux (sh*-*-linux*)
    SH-5/SH64 running OpenBSD (sh*-*-openbsd*)

 Various Python API enhancements
 Aarch64/Linux enhancements:

    SVE support.
    Hardware watchpoints improvements for entities stored at unaligned addresses.
        New "c" response to disable the pager for the rest of the current command.
        C expressions can now use _Alignof, and C++ expressions can now use alignof.
        Improved flexibility for loading symbol files.
        The 'info proc' command nows works on running processes on FreeBSD systems as well as core files created on FreeBSD systems.
        A new --enable-codesign=CERT configure option to automatically codesign GDB after build (useful on MacOS X).

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agotoolchain/gdb: bump to 8.2
Koen Vandeputte [Thu, 6 Sep 2018 12:43:56 +0000 (14:43 +0200)]
toolchain/gdb: bump to 8.2

*** Changes in GDB 8.2

 Support for the following target has been added:

    RiscV ELF (riscv*-*-elf)

 Support for following targets and native configurations has been removed:

    m88k running OpenBSD (m88*-*-openbsd*)
    SH-5/SH64 ELF (sh64-*-elf*)
    SH-5/SH64 (sh*)
    SH-5/SH64 running GNU/Linux (sh*-*-linux*)
    SH-5/SH64 running OpenBSD (sh*-*-openbsd*)

 Various Python API enhancements
 Aarch64/Linux enhancements:

    SVE support.
    Hardware watchpoints improvements for entities stored at unaligned addresses.
        New "c" response to disable the pager for the rest of the current command.
        C expressions can now use _Alignof, and C++ expressions can now use alignof.
        Improved flexibility for loading symbol files.
        The 'info proc' command nows works on running processes on FreeBSD systems as well as core files created on FreeBSD systems.
        A new --enable-codesign=CERT configure option to automatically codesign GDB after build (useful on MacOS X).

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agomac80211: backport upstream fixes
Koen Vandeputte [Tue, 4 Sep 2018 13:09:48 +0000 (15:09 +0200)]
mac80211: backport upstream fixes

Backport most significant upstream fixes (excl. hwsim fixes)
Refreshed all patches.

Contains important fixes for CSA (Channel Switch Announcement)
and A-MSDU frames.

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agokernel: bump 4.14 to 4.14.68
Koen Vandeputte [Thu, 6 Sep 2018 12:25:15 +0000 (14:25 +0200)]
kernel: bump 4.14 to 4.14.68

Refreshed all patches.

Remove upstream accepted:
- 330-Revert-MIPS-BCM47XX-Enable-74K-Core-ExternalSync-for.patch

Altered:
- 303-v4.16-netfilter-nf_tables-remove-multihook-chains-and-fami.patch
- 308-mips32r2_tune.patch

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

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agokernel: bump 4.9 to 4.9.125
Koen Vandeputte [Thu, 6 Sep 2018 12:23:27 +0000 (14:23 +0200)]
kernel: bump 4.9 to 4.9.125

Refreshed all patches.

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

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agokernel: bump 3.18 to 3.18.121
Koen Vandeputte [Thu, 6 Sep 2018 12:21:05 +0000 (14:21 +0200)]
kernel: bump 3.18 to 3.18.121

Refreshed all patches.

Compile-tested on: adm5120
Runtime-tested on: none

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
6 years agoramips: add support for TP-Link TL-MR3020 v3
Carlo Nel [Sun, 26 Aug 2018 19:11:59 +0000 (21:11 +0200)]
ramips: add support for TP-Link TL-MR3020 v3

TP-Link TL-MR3020 v3 is a pocket-size router based on MediaTek MT7628N.

This PR is based on the work of @meyergru[1], with his permission.

Specification:
- MediaTek MT7628N/N (575 Mhz)
- 64 MB of RAM
- 8 MB of FLASH
- 2T2R 2.4 GHz
- 1x 10/100 Mbps Ethernet

Flash instruction:

The only way to flash the image in TL-MR3020 v3 is to use
tftp recovery mode in U-Boot:

1. Configure PC with static IP 192.168.0.225/24 and tftp server.
2. Rename "openwrt-ramips-mt76x8-tplink_tl-mr3020-v3-squashfs-tftp-recovery.bin"
   to "tp_recovery.bin" and place it in tftp server directory.
3. Connect PC with the LAN port, press the reset button, power up
   the router and keep button pressed for around 6-7 seconds, until
   device starts downloading the file.
4. Router will download file from server, write it to flash and reboot.

[1] https://github.com/meyergru/lede-source/commits/TL-MR3020-V3

Signed-off-by: Carlo Nel <carlojnel@gmail.com>
6 years agoath79: fix unaligned access panic in ag71xx_mdio_probe
Matt Merhar [Thu, 30 Aug 2018 12:58:26 +0000 (08:58 -0400)]
ath79: fix unaligned access panic in ag71xx_mdio_probe

Without "syscon" being present in an ag71xx ethernet DT node's
compatible property, a panic occurs at boot during probe citing
"Unhandled kernel unaligned access".

With this modification, the panic no longer occurs and instead the probe
simply fails, allowing the boot process to continue.

Signed-off-by: Matt Merhar <mattmerhar@protonmail.com>
6 years agoath79: add syscon compatible property to ar7100 ethernet nodes
Matt Merhar [Thu, 30 Aug 2018 00:14:19 +0000 (20:14 -0400)]
ath79: add syscon compatible property to ar7100 ethernet nodes

This adds "syscon" to the compatible properties for the eth0/eth1 nodes
in ar7100.dtsi.

Without this, a kernel panic is encountered on boot with some ar7100
boards. This for some reason wasn't an issue for the WNDR3800, which
uses a Realtek switch chipset, but the panic was encountered on the
RouterStation Pro (using an AR8216 switch) and some other boards that
haven't yet been merged.

The panic message mentions an unaligned access and happens in
ag71xx_mdio_probe in drivers/net/ethernet/atheros/ag71xx/ag71xx_mdio.c.

Even if the unaligned access is fixed, the ag71xx_mdio probe still fails
without the "syscon" property.

This was already being worked around in
ar7161_ubnt_routerstation-pro.dts by overriding the compatible property,
so this commit removes that as well.

All of the other ath79 .dtsi already have this property, so no changes
are needed elsewhere.

Signed-off-by: Matt Merhar <mattmerhar@protonmail.com>
6 years agoramips: drop obsolete sd card driver code
Mathias Kresin [Thu, 30 Aug 2018 17:16:53 +0000 (19:16 +0200)]
ramips: drop obsolete sd card driver code

The pinmux for all SoCs using this driver is now set via the pinmux. It
makes this code obsolete.

Some of the code targeting the mt76x8 SoCs is still required. The sd
card pins share the pads with the EPHY. These pads need to be switched
to digital mode if the pins are used for sd cards.

The eMMC 8-bit mode has to be enabled via pinmux instead of a kernel
option. The uart2 group need to be set to function "sdxc d5 d4", pwm1
to "sdxc d6" and pwm0 to "sdxc d7" to do so. It can't be done by as
part of a default pinmux, as it would break the normal operation of
uart2.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: add mt7620/1 sdhci pinmux
Mathias Kresin [Thu, 30 Aug 2018 17:13:20 +0000 (19:13 +0200)]
ramips: add mt7620/1 sdhci pinmux

Set the pins to the required mode via the pinmux driver. It allows to
get rid of the pinmux related code in the sd card driver.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix mt7620a ND/SD pins pinmuxes
Mathias Kresin [Fri, 31 Aug 2018 05:53:42 +0000 (07:53 +0200)]
ramips: fix mt7620a ND/SD pins pinmuxes

Drop the nd_sd gpio pinmux in case sdcard is used. They're mutually
exclusive and for most of the boards not even used as GPIOs.

If the pins are in sdcard mode, the pins ND_WE_N and ND_CS_N are still
GPIOs (#45 and #46).

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: reference node by label
Mathias Kresin [Fri, 31 Aug 2018 07:01:34 +0000 (09:01 +0200)]
ramips: reference node by label

Reference the HC5661A sdhci node by label instead of by the full path.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: add rt3352 SPI_CS1 pinmux
Mathias Kresin [Wed, 22 Aug 2018 06:02:20 +0000 (08:02 +0200)]
ramips: add rt3352 SPI_CS1 pinmux

The rt3352 has a pin that can be used as second spi chip select,
watchdog reset or GPIO. The pinmux setup was missing the definition of
said pin but it is already used in the SoC dtsi.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agodnsmasq: allow dnsmasq variants to be included in image
Henrique de Moraes Holschuh [Mon, 27 Aug 2018 11:31:29 +0000 (08:31 -0300)]
dnsmasq: allow dnsmasq variants to be included in image

The dnsmasq variants should provide dnsmasq, otherwise it is impossible
to include them in the image.

This change allows one to have CONFIG_PACKAGE_dnsmasq=m and
CONFIG_PACKAGE_dnsmasq-full=y, e.g. because you want DNSSEC support, or
IPSETs suport on your 3000-devices fleet ;-)

Signed-off-by: Henrique de Moraes Holschuh <henrique@nic.br>
6 years agodnsmasq: bump to dnsmasq v2.80test5
Hans Dedecker [Wed, 5 Sep 2018 13:39:39 +0000 (15:39 +0200)]
dnsmasq: bump to dnsmasq v2.80test5

Refresh patches
Remove 240-ubus patch as upstream accepted.
Add uci option ubus which allows to enable/disable ubus support (enabled
by default)

Upstream commits since last bump:

da8b651 Implement --address=/example.com/#
c5db8f9 Tidy 7f876b64c22b2b18412e2e3d8506ee33e42db7c
974a6d0 Add --caa-record
b758b67 Improve logging of RRs from --dns-rr.
9bafdc6 Tidy up file parsing code.
97f876b Properly deal with unaligned addresses in DHCPv6 packets.
cbfbd17 Fix broken DNSSEC records in previous.
b6f926f Don't return NXDOMAIN to empty non-terminals.
c822620 Add --dhcp-name-match
397c050 Handle case of --auth-zone but no --auth-server.
1682d15 Add missing EDNS0 section. EDNS0 section missing in replies to EDNS0-containing queries where answer generated from --local=/<domain>/
dd33e98 Fix crash parsing a --synth-domain with no prefix. Problem introduced in 2.79/6b2b564ac34cb3c862f168e6b1457f9f0b9ca69c
c16d966 Add copyright to src/metrics.h
1dfed16 Remove C99 only code.
6f835ed Format fixes - ubus.c
9d6fd17 dnsmasq.c fix OPT_UBUS option usage
8c1b6a5 New metrics and ubus files.
8dcdb33 Add --enable-ubus option.
aba8bbb Add collection of metrics
caf4d57 Add OpenWRT ubus patch

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agostrace: fix build on aarch64
Stijn Tintel [Thu, 28 Jun 2018 10:00:10 +0000 (13:00 +0300)]
strace: fix build on aarch64

As of version 4.21, strace enforces mpers by default. The current
implementation of aarch64 compat in strace assumes it's identical to
ARMv7 EABI and therefore tries to enable m32 personality support. As
there is no -m32 support on aarch64, this causes the build to fail.

Restore previous strace behavior to fix build on aarch64.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Tested-by: Karl Palsson <karlp@tweak.net.au>
6 years agoodhcpd: bump to git HEAD (detect broken hostnames)
Hans Dedecker [Wed, 5 Sep 2018 08:59:23 +0000 (10:59 +0200)]
odhcpd: bump to git HEAD (detect broken hostnames)

881f66b odhcpd: detect broken hostnames
3e17fd9 config: fix odhcpd_attrs array size

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agoimx6: use BUILD_DEVICES in uboot-imx6 for mx6cuboxi
Felix Fietkau [Wed, 5 Sep 2018 07:56:52 +0000 (09:56 +0200)]
imx6: use BUILD_DEVICES in uboot-imx6 for mx6cuboxi

Fixes build with the default profile

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agoimx6: rename Generic profile to Default, fixes u-boot image build issue
Felix Fietkau [Wed, 5 Sep 2018 07:52:57 +0000 (09:52 +0200)]
imx6: rename Generic profile to Default, fixes u-boot image build issue

u-boot.mk checks for the Default profile to build images for all targets.
This brings the target default profile in sync with other targets

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agowireguard: bump to 0.0.20180904
Jason A. Donenfeld [Tue, 4 Sep 2018 18:33:46 +0000 (12:33 -0600)]
wireguard: bump to 0.0.20180904

* Kconfig: use new-style help marker
* global: run through clang-format
* uapi: reformat
* global: satisfy check_patch.pl errors
* global: prefer sizeof(*pointer) when possible
* global: always find OOM unlikely

Tons of style cleanups.

* crypto: use unaligned helpers

We now avoid unaligned accesses for generic users of the crypto API.

* crypto: import zinc

More style cleanups and a rearrangement of the crypto routines to fit how this
is going to work upstream. This required some fairly big changes to our build
system, so there may be some build errors we'll have to address in subsequent
snapshots.

* compat: rng_is_initialized made it into 4.19

We therefore don't need it in the compat layer anymore.

* curve25519-hacl64: use formally verified C for comparisons

The previous code had been proved in Z3, but this new code from upstream
KreMLin is directly generated from the F*, which is preferable. The
assembly generated is identical.

* curve25519-x86_64: let the compiler decide when/how to load constants

Small performance boost.

* curve25519-arm: reformat
* curve25519-arm: cleanups from lkml
* curve25519-arm: add spaces after commas
* curve25519-arm: use ordinary prolog and epilogue
* curve25519-arm: do not waste 32 bytes of stack
* curve25519-arm: prefix immediates with #

This incorporates ASM nits from upstream review.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
6 years agompc85xx: add migration script for TP-Link TL-WDR4900 v1 WLAN PCI paths
Matthias Schiffer [Tue, 4 Sep 2018 19:27:27 +0000 (21:27 +0200)]
mpc85xx: add migration script for TP-Link TL-WDR4900 v1 WLAN PCI paths

PCI paths of the WLAN devices have changed between kernel 4.4 and 4.9;
migrate config so existing wifi-iface definitions don't break.

This is implemented as a hotplug handler rather than a uci-defaults script
as the migration script must run before the 10-wifi-detect hotplug handler.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
6 years agohostapd: fix build of wpa-supplicant-p2p
Alexander Couzens [Mon, 3 Sep 2018 19:42:02 +0000 (21:42 +0200)]
hostapd: fix build of wpa-supplicant-p2p

VARIANT:= got removed by accident.

Fixes: 3838b16943c6 ("hostapd: fix conflicts hell")
Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
6 years agoimx6: extend cubox support to hummingboard, add support for building full images
Felix Fietkau [Sun, 26 Aug 2018 18:04:08 +0000 (20:04 +0200)]
imx6: extend cubox support to hummingboard, add support for building full images

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agoimx6: enable f2fs and loopback support in preparation for squashfs with overlay support
Felix Fietkau [Sun, 26 Aug 2018 18:03:00 +0000 (20:03 +0200)]
imx6: enable f2fs and loopback support in preparation for squashfs with overlay support

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agobuild: add support for enabling the rootfs/boot partition size option via target...
Felix Fietkau [Sun, 26 Aug 2018 17:36:33 +0000 (19:36 +0200)]
build: add support for enabling the rootfs/boot partition size option via target feature

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agouboot-imx6: fall back to MMC for SPL if boot mode is invalid
Felix Fietkau [Sun, 26 Aug 2018 17:34:41 +0000 (19:34 +0200)]
uboot-imx6: fall back to MMC for SPL if boot mode is invalid

Fixes boot on Hummingboard

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agouboot-imx6: remove obsolete patch
Felix Fietkau [Sun, 26 Aug 2018 17:34:17 +0000 (19:34 +0200)]
uboot-imx6: remove obsolete patch

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agouboot-imx6: install images into STAGING_DIR_IMAGE
Felix Fietkau [Sun, 26 Aug 2018 11:47:23 +0000 (13:47 +0200)]
uboot-imx6: install images into STAGING_DIR_IMAGE

Will be used by a new combined image for cubox

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agoramips: ethernet: unify tx descriptor buffer splitting
Felix Fietkau [Sun, 22 Jul 2018 11:07:29 +0000 (13:07 +0200)]
ramips: ethernet: unify tx descriptor buffer splitting

A buffer is split into multiple descriptors if it exceeds 16 KB.
Apply the same split for the skb head as well (to deal with corner cases
on fraglist support)

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 years agonghttp2: bump to 1.33.0
Hans Dedecker [Mon, 3 Sep 2018 08:29:56 +0000 (10:29 +0200)]
nghttp2: bump to 1.33.0

9d843334 Update bash_completion
23cb3f38 Update manual pages
1d682dcd Bump up version number to 1.33.0, LT revision to 31:0:17
601fbbb4 Update doc
f44aa246 Update AUTHORS
dd74a6dd Update manual pages
e959e733 src: Refactor utos
fb9a204d nghttpx: Fix compile error without mruby
cd096802 Update doc
7417fd71 nghttpx: Per-pattern not per-backend
2d1a981c Merge branch 'akonskarm-master'
45acc922 clang-format
214d0899 Merge branch 'master' of https://github.com/akonskarm/nghttp2 into akonskarm-master
31fd707d nghttpx: Fix broken healthmon frontend
9a2e38e0 fix code for reuse addr on asio client
d24527e7 Bump up LT revision due to v1.32.1 release
6195d747 nghttpx: Share mruby context if it is compiled from same file
fb97f596 nghttpx: Allocate mruby file because fopen requires NULL terminated string
0ccc7a77 nghttpx: Move blocked request data to request buffer for API request
32826466 nghttpx: Fix crash with API request
0422f8a8 nghttpx: Fix worker process crash with neverbleed write error
e329479a Merge pull request #1215 from nghttp2/mruby-per-backend
f80a7873 Merge branch 'akonskarm-reuse_addr'
866ac6ab add option reuse addr in local endpoint configuration of asio client
b574ae6a nghttpx: Support per-backend mruby script
de4fd7cd doc: Update doc
32d7883c nghttpx: Downstream::request_buf_full: take into account blocked_request_buf_
9b24e197 nghttpx: Choose h1 protocol if headers have been sent to backend on retry
13ffece1 Merge pull request #1214 from nghttp2/fix-rst-without-dconn
9d5b781d Fix stream reset if data from client is arrived before dconn is attached

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
6 years agoiproute2: update cake man page
Kevin Darbyshire-Bryant [Mon, 27 Aug 2018 08:52:55 +0000 (15:52 +0700)]
iproute2: update cake man page

CAKE supports overriding of its internal classification of
packets through the tc filter mechanism.

Update the man page in our package, even though we don't
build them.  Someone may find the documentation useful.

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit 30598a05385b0ac2380dd4f30037a9f9d0318cf2)

6 years agokmod-sched-cake: bump to 20180827
Kevin Darbyshire-Bryant [Mon, 13 Aug 2018 12:49:19 +0000 (13:49 +0100)]
kmod-sched-cake: bump to 20180827

Expand filter flow mapping to include hosts as well

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit d14ffdc307d36bd9abe908b46ff7baece54c9551)

6 years agoppp: remove hardcoded lcp-echo-failure, lcp-echo-interval values
Jo-Philipp Wich [Thu, 30 Aug 2018 13:08:20 +0000 (15:08 +0200)]
ppp: remove hardcoded lcp-echo-failure, lcp-echo-interval values

OpenWrt used to ship hardcoded defaults for lcp-echo-failure and
lcp-echo-interval in the non-uci /etc/ppp/options file.

These values break uci support for *disabling* LCP echos through
the use of "option keepalive 0" as either omitting the keepalive
option or setting it to 0 will result in no lcp-echo-* flags
getting passed to the pppd cmdline, causing the pppd process to
revert to the defaults in /etc/ppp/options.

Address this issue by letting the uci "keepalive" option default
to the former hardcoded values "5, 1" and by removing the fixed
lcp-echo-failure and lcp-echo-interval settings from the
/etc/ppp/options files.

Ref: https://github.com/openwrt/luci/issues/2112
Ref: https://dev.archive.openwrt.org/ticket/2373.html
Ref: https://bugs.openwrt.org/index.php?do=details&task_id=854
Ref: https://bugs.openwrt.org/index.php?do=details&task_id=1259
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoRevert "ramips: mmc: Fix init for MT7628AN"
Jo-Philipp Wich [Thu, 30 Aug 2018 11:19:18 +0000 (13:19 +0200)]
Revert "ramips: mmc: Fix init for MT7628AN"

This reverts commit 3a8efaef00d6b7317c46fe6b7e923912d8bf3500.

The change reportedly breaks UART2 on some boards. Furthermore it uses
bitwise logic on an uninitialized variable and fails to explain what it
is fixing exactly.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoramips: only limit lzma dictionary size on mt7621
Jo-Philipp Wich [Thu, 30 Aug 2018 08:51:09 +0000 (10:51 +0200)]
ramips: only limit lzma dictionary size on mt7621

The changed dictionary size leads to a different LZMA header which breaks
sysupgrade image magic checkibng on at least some RT288x boards.

Since the commit message only mentions testing on MT7621 and since the
change appears to break at least one other ramips subtarget, do not take
any chances and restrict the size limitation to only MT7621.

Fixes FS#1797
Fixes 09b6755946 ("ramips: limit dictionary size for lzma compression")
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agoramips: fix rt3883 pinmux for second SPI
Mathias Kresin [Wed, 22 Aug 2018 05:40:17 +0000 (07:40 +0200)]
ramips: fix rt3883 pinmux for second SPI

The rt3883 doesn't have a pinmux group named spi_cs1. The cs1 is part
of the pci group. The function pci-func enables the second chip select.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix rt3883 pci pinmux
Mathias Kresin [Wed, 22 Aug 2018 05:26:39 +0000 (07:26 +0200)]
ramips: fix rt3883 pci pinmux

The PCI pins need to be set to "PCI Host support one device" to allow
the use of one PCI device and flash memory.

The pci-fnc function is intended to be used if no PCI is used but
flash, nand or the codec functionality is.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoscripts: bundle-libraries: fix logic flaw
Jo-Philipp Wich [Wed, 29 Aug 2018 11:16:34 +0000 (13:16 +0200)]
scripts: bundle-libraries: fix logic flaw

Previous refactoring of the script moved the LDSO detection into a
file-not-exists condition, causing onyl the very first executable to
get bundled.

Solve the problem by unconditionally checking for LDSO again.

Fixes: 9030a78a71 ("scripts: bundle-libraries: prevent loading host locales")
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
6 years agouqmi: wait for the control device too
Thomas Equeter [Thu, 16 Aug 2018 19:39:05 +0000 (21:39 +0200)]
uqmi: wait for the control device too

The control device /dev/cdc-wdm0 is not available immediately on the
D-Link DWR-921 Rev.C3, therefore the wwan interface fails to start at
boot with a "The specified control device does not exist" error.

This patch alters /lib/netifd/proto/qmi.sh to wait for
network.wwan.delay earlier, before checking for the control device,
instead of just before interacting with the modem.

One still has to use network.wwan.proto='qmi', as the "wwan" proto
performs that sort of check before any delay is possible, failing with a
"No valid device was found" error.

Signed-off-by: Thomas Equeter <tequeter@users.noreply.github.com>
6 years agoapm821xx: MBL: load kernel/dtb from SATA 0:1 first
Freddy Leitner [Tue, 28 Aug 2018 18:47:58 +0000 (20:47 +0200)]
apm821xx: MBL: load kernel/dtb from SATA 0:1 first

This remedies an issue with the MBL Duo if both disks are inserted
and contain OpenWrt. kernel and dtb would be loaded from SATA 1:1
while rootfs (/dev/sda2) would be mounted on SATA 0:1.

Such a mix&match would obviously only work if both OpenWrt versions/
builds are identical, and especially fail after sysupgrade upgraded
the system disk on SATA 0:1.

The fallback to SATA 1:1 needs to be kept for MBL Single (only has
SATA 1:1) and MBL Duo with one disk inserted on SATA 1:1. To speed
up booting in those cases, the unneccesarily doubled "sata init"
will only be called once. (In theory it could be omitted completely
since the on-flash boot script already initializes SATA to load the
on-disk boot script.)

Tested on MBL Duo (all possible combination of disks) and MBL Single

Acked-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Freddy Leitner <hello@square.wf>
6 years agoramips: drop pointless default led definitions
Mathias Kresin [Sat, 18 Aug 2018 10:07:08 +0000 (12:07 +0200)]
ramips: drop pointless default led definitions

The LEDs should be triggered/lit by any kind of state change instead of
turned on/off unconditional.

If LEDs really need to be turned off by default, it should be done via
the default-state devicetree led property.

The handling of the wndr3700v5 and wt3020 power led is at least
strange. Something is for sure wrong with them. Either the leds are
misnamed, the default off trigger is a typo or the polarity of the
gpios is wrong. Drop the power led from userspace and wait for someone
with access to the hardware to fix it properly.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: add missing zbt-cpe102 diag led
Mathias Kresin [Sat, 18 Aug 2018 09:42:50 +0000 (11:42 +0200)]
ramips: add missing zbt-cpe102 diag led

Based on the userspace led configuration it's quite obvious that the
4g-0 led should be used for boot status indication.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: express diag led handling via devicetree
Mathias Kresin [Sat, 18 Aug 2018 09:38:13 +0000 (11:38 +0200)]
ramips: express diag led handling via devicetree

Use the default-state property to express the desired led handling in
the devicetree source file instead of the userspace.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: remove default on userspace trigger for diag leds
Mathias Kresin [Sat, 18 Aug 2018 09:09:29 +0000 (11:09 +0200)]
ramips: remove default on userspace trigger for diag leds

All the LEDs are turned on by diag.sh at the end of the boot process.
No need to do the same via userspace configuration again.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: fix multi colour led handling
Mathias Kresin [Sat, 18 Aug 2018 10:22:25 +0000 (12:22 +0200)]
ramips: fix multi colour led handling

All boards either have a multi colour led or a single lightpipe. It
makes it impossible to handle the LEDs individual. Change the LED
config for these boards to take it into account.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agoramips: rt-n12p: use the boardname helper variable
Mathias Kresin [Wed, 29 Aug 2018 06:59:46 +0000 (08:59 +0200)]
ramips: rt-n12p: use the boardname helper variable

Use the helper variable for the Asus RT-N12P as it is done by all
boards.

Signed-off-by: Mathias Kresin <dev@kresin.me>
6 years agocomgt: increase timeout on runcommands
Giuseppe Lippolis [Sun, 26 Aug 2018 08:52:27 +0000 (10:52 +0200)]
comgt: increase timeout on runcommands

Some combination of modem/wireless operator requires more time to
execute the commands.
Tested on DWR-512 embedded wwan modem and italian operator iliad (new
virtual operator).

Signed-off-by: Giuseppe Lippolis <giu.lippolis@gmail.com>
6 years agobase-files: provide more tolerant xterm detection
Paul Wassi [Thu, 23 Aug 2018 07:24:43 +0000 (09:24 +0200)]
base-files: provide more tolerant xterm detection

Set the window title not only in "xterm", but also in
e.g. "xterm-256color", "xterm-color", etc.
The case statement is taken from Debian / Ubuntu.

Signed-off-by: Paul Wassi <p.wassi@gmx.at>