openwrt/staging/blocktrron.git
7 years agosunxi: sysupgrade: don't write partitions twice
Matthias Schiffer [Thu, 4 May 2017 05:29:58 +0000 (07:29 +0200)]
sunxi: sysupgrade: don't write partitions twice

When existing partitions are retained, the dd call writing the uboot image
in the space before the first partition was accidentally writing the whole
image, making the code for individual partitions redundant. Limit the copy
to 1016KiB (the first 8KiB are skipped, and the first partition starts at
1024KiB).

In addition, conv=notrunc is replaced with conv=fsync. It seems this was an
oversight, as notrunc doesn't make sense for block devices and all other dd
commands use conv=fsync.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agox86: sysupgrade: explicitly rescan disk after writing partition table
Matthias Schiffer [Wed, 3 May 2017 07:08:29 +0000 (09:08 +0200)]
x86: sysupgrade: explicitly rescan disk after writing partition table

This should ensure that the kernel partition can be mounted in
platform_copy_config when its size has changed.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agox86: sysupgrade: refactor platform_do_upgrade
Matthias Schiffer [Wed, 3 May 2017 07:05:25 +0000 (09:05 +0200)]
x86: sysupgrade: refactor platform_do_upgrade

By returning early when no upgrade device can be found and handling the
SAVE_PARTITIONS=0 case differently, we can get rid of two levels of if.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agox86: sysupgrade: move partition table change check to platform_check_image
Matthias Schiffer [Wed, 3 May 2017 06:57:29 +0000 (08:57 +0200)]
x86: sysupgrade: move partition table change check to platform_check_image

The staged sysupgrade will prevent us from using ask_bool in
platform_do_upgrade; therefore, the check is moved to platform_check_image.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agoramips: sysupgrade: move nand_do_upgrade call to platform_do_upgrade
Matthias Schiffer [Sat, 22 Apr 2017 19:27:04 +0000 (21:27 +0200)]
ramips: sysupgrade: move nand_do_upgrade call to platform_do_upgrade

All targets with NAND support should gradually move their nand_do_upgrade
calls from platform_pre_upgrade to platform_do_upgrade.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agobase-files: add support for staged sysupgrades from failsafe mode
Matthias Schiffer [Sun, 23 Apr 2017 23:31:04 +0000 (01:31 +0200)]
base-files: add support for staged sysupgrades from failsafe mode

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agobase-files: sysupgrade cleanup
Matthias Schiffer [Sun, 23 Apr 2017 21:33:14 +0000 (23:33 +0200)]
base-files: sysupgrade cleanup

Some functions only used by stage2 are moved there from common.sh.

One piece that could still use more cleanup is platform_pre_upgrade: many
targets reference files from there are aren't available in the ramfs, so
we need to evaluate it before the switch; conversely, flash writes happen
in that function on some targets. Targets that do the latter should be
fixed eventually to use platform_do_upgrade for that purpose.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agofstools: snapshot: handle jffs2 conversion using upgraded
Matthias Schiffer [Wed, 3 May 2017 06:29:55 +0000 (08:29 +0200)]
fstools: snapshot: handle jffs2 conversion using upgraded

We can reuse the kill_remaining and run_ramfs facilities of the stage2 run
by upgraded.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agofstools: clean up trailing whitespace in snapshot script
Matthias Schiffer [Wed, 3 May 2017 06:29:24 +0000 (08:29 +0200)]
fstools: clean up trailing whitespace in snapshot script

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agobase-files: always use staged sysupgrade
Matthias Schiffer [Fri, 21 Apr 2017 22:54:50 +0000 (00:54 +0200)]
base-files: always use staged sysupgrade

Support for the -d and -p options is dropped; it may be added again at some
point by adding these flags to the ubus sysupgrade call.

A downside of this is that we get a lot less information about the progress
of the upgrade: as soon as the actual upgrade starts, all shell sessions
are killed to allow unmounting the root filesystem.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agoprocd: remove procd-nand package
Matthias Schiffer [Fri, 21 Apr 2017 18:37:58 +0000 (20:37 +0200)]
procd: remove procd-nand package

We always want to support staged upgrades now, so it's better to include
upgraded into the main package. /lib/upgrade/nand.sh is moved to
base-files.

The procd-nand-firstboot package is removed for now, it may return later
as a separate package.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agoprocd: update to latest git HEAD
Matthias Schiffer [Mon, 29 May 2017 21:15:22 +0000 (23:15 +0200)]
procd: update to latest git HEAD

992b796 rcS: add missing fcntl.h include
63789e5 init: add support for sysupgrades triggered from preinit
5b1fb35 Remove code that has become unnecessary after sysupgrade changes
5918b6d upgraded: add support for passing a "command" argument to stage2
056d8dd upgraded: link dynamically, chroot during exec
7c6cf55 system: always support staged sysupgrade
d42b21e procd/rcS: Use /dev/null as stdin
e0098d4 service/instance: add an auto start option
1247db1 procd: Log initscript output prefixed with script name
8d720b2 procd: Don't use syslog before its initialization
2555474 procd: Add missing \n in debug message

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agoprocd: prepare NAND sysupgrade for making upgraded dynamically linked
Matthias Schiffer [Wed, 3 May 2017 06:28:05 +0000 (08:28 +0200)]
procd: prepare NAND sysupgrade for making upgraded dynamically linked

Use install_bin to copy upgraded with all dependencies. The old name
/tmp/upgraded is temporarily retained as a symlink to avoid breaking
things.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agoprocd: clean up trailing whitespace in nand.sh
Matthias Schiffer [Wed, 3 May 2017 06:27:40 +0000 (08:27 +0200)]
procd: clean up trailing whitespace in nand.sh

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
7 years agomt76: update to the latest version, improves mt7603/7628 stability
Felix Fietkau [Mon, 29 May 2017 12:42:16 +0000 (14:42 +0200)]
mt76: update to the latest version, improves mt7603/7628 stability

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: fix kmod package build on non-GNU systems
Felix Fietkau [Mon, 29 May 2017 12:26:36 +0000 (14:26 +0200)]
build: fix kmod package build on non-GNU systems

BSD paste requires a filename argument, and it accepts - to use stdin as
intended.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agotools/sparse: Update to snapshot 2017-03-31
Daniel Engberg [Wed, 10 May 2017 09:34:10 +0000 (11:34 +0200)]
tools/sparse: Update to snapshot 2017-03-31

Update sparse to snapshot 2017-03-31
Switch to HTTPS

Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
7 years agotoolchain/insight: Remove from tree
Daniel Engberg [Mon, 29 May 2017 07:41:15 +0000 (09:41 +0200)]
toolchain/insight: Remove from tree

Insight hasn't been touched since 2009 and I can't find any references at
least 6 years back of being used. Ubuntu and Debian removed insight years ago
from their repos so I think we can safely remove this too.

Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
7 years agolantiq: pinctrl: xway: fix copy/paste error in xrx200_grps
Martin Schiller [Mon, 29 May 2017 04:17:07 +0000 (06:17 +0200)]
lantiq: pinctrl: xway: fix copy/paste error in xrx200_grps

Signed-off-by: Martin Schiller <ms@dev.tdt.de>
7 years agoipq806x: add qpic nand and bam dma node's in ipq4019 dts tree
Ram Chandra Jangir [Wed, 24 May 2017 23:31:03 +0000 (05:01 +0530)]
ipq806x: add qpic nand and bam dma node's in ipq4019 dts tree

This change adds QPIC BAM dma and NAND driver node's in IPQ4019
device tree, also enable this for AP-DK04.1 based boards.

Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
7 years agoipq806x: Add support for custom data mapping in bam_dma dmaengine
Ram Chandra Jangir [Wed, 24 May 2017 23:31:02 +0000 (05:01 +0530)]
ipq806x: Add support for custom data mapping in bam_dma dmaengine

This change adds a new function to support for preparing DMA descriptor
for custom data.

Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
7 years agoipq806x: Added bam transaction and support additional CSRs
Ram Chandra Jangir [Wed, 24 May 2017 23:31:01 +0000 (05:01 +0530)]
ipq806x: Added bam transaction and support additional CSRs

This change adds support for below:
- Bam transaction which will be used for any NAND request.
- Reset function for NAND BAM transaction
- Add support for additional CSRs.

Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
7 years agoipq806x: Add bam_dma support in qcom_nand driver
Ram Chandra Jangir [Wed, 24 May 2017 23:31:00 +0000 (05:01 +0530)]
ipq806x: Add bam_dma support in qcom_nand driver

The existing qcom_nand driver supports ADM DMA which is mainly
required for ipq806x family based boards,
IPQ40xx based boards uses BAM DMA in NAND driver, so this patch
adds BAM DMA support with compatible string as qcom,ebi2-nandc-bam.

Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
7 years agoipq806x: Add support to configure ipq40xx GPIO_PULL bits
Ram Chandra Jangir [Wed, 24 May 2017 23:30:59 +0000 (05:00 +0530)]
ipq806x: Add support to configure ipq40xx GPIO_PULL bits

GPIO_PULL bits configurations in TLMM_GPIO_CFG register
differs for IPQ40xx from rest of the other qcom SoC's.
This change add support to configure the msm_gpio_pull
bits for ipq40xx, It is required to fix the proper
configurations of gpio-pull bits for nand pins mux.

Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
7 years agofirewall: fix stray continue statement
Jo-Philipp Wich [Sat, 27 May 2017 14:14:10 +0000 (16:14 +0200)]
firewall: fix stray continue statement

The previous commit introduced a faulty continue statement which might
lead to faulty rules not getting freed or reported.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agonetifd: fix 6rd regression (FS#812)
Hans Dedecker [Sat, 27 May 2017 11:18:33 +0000 (13:18 +0200)]
netifd: fix 6rd regression (FS#812)

08f1875 system-linux: fix 6rd regression

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
7 years agofirewall: extend ubus support, exception handling, parse fixes
Jo-Philipp Wich [Sun, 14 May 2017 19:00:26 +0000 (21:00 +0200)]
firewall: extend ubus support, exception handling, parse fixes

Update to latest Git HEAD in order to import a number of fixes and other
improvements:

3d2c18a options: improve handling of negations when parsing space separated values
0e5dd73 iptables: support -i, -o, -s and -d in option extra
4cb06c7 ubus: increase ubus network interface dump timeout
e5dfc82 iptables: add exception handling
f625954 firewall3: add check_snat() function
7d3d9dc firewall3: display the section type for UBUS rules
53ef9f1 firewall3: add UBUS support for include scripts
5cd4af4 firewall3: add UBUS support for ipset sections
02d6832 firewall3: add UBUS support for forwarding sections
0a7d36d firewall3: add UBUS support for redirect sections
d44f418 firewall3: add fw3_attr_parse_name_type() function
e264c8e firewall3: replace warn_rule() by warn_section()
6039c7f firewall3: check the return value of fw3_parse_options()

Fixes FS#548, FS#806, FS#811.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agobrcm47xx: remove target specific network preinit config
Jo-Philipp Wich [Wed, 17 May 2017 12:36:23 +0000 (14:36 +0200)]
brcm47xx: remove target specific network preinit config

The generic preinit code is now able to setup network and switch vlan settings
from the /etc/board.json file, therefor drop the target specific code.

Fixes FS#790.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agosamba: bump PKG_RELEASE
Jo-Philipp Wich [Sat, 27 May 2017 10:15:06 +0000 (12:15 +0200)]
samba: bump PKG_RELEASE

The previous CVE bugfix commit did not adjust PKG_RELEASE, therefor the
fixed samba package does not appear as opkg update.

Bump the PKG_RELEASE to signify upgrades to downstream users.

Ref: https://forum.lede-project.org/t/sambacry-are-lede-devices-affected/3972/4

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agox86: fix build of geode target
Hauke Mehrtens [Sat, 27 May 2017 09:29:55 +0000 (11:29 +0200)]
x86: fix build of geode target

The build bot complained that the OLPC option was not set.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agokernel: add missing config options
Hauke Mehrtens [Sat, 27 May 2017 09:29:28 +0000 (11:29 +0200)]
kernel: add missing config options

This broke the build for the x86 generic target.
This was found by the build bot.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agoMove enablemodem from ramips to new package adb-enablemodem and make it used also...
Filip Moc [Thu, 25 May 2017 19:55:32 +0000 (21:55 +0200)]
Move enablemodem from ramips to new package adb-enablemodem and make it used also by TL-MR6400

Signed-off-by: Filip Moc <lede@moc6.cz>
7 years agoar71xx: Add support for TP-Link MR6400
Filip Moc [Thu, 25 May 2017 19:51:37 +0000 (21:51 +0200)]
ar71xx: Add support for TP-Link MR6400

You can flash via tftp recovery (serve factory image as /mr6400_tp_recovery.bin
on 192.168.0.66/24, connect to any ethernet port and power on device while
holding the reset button). Flashing via OEM web interface does not work.

Hardware Specification (v1.0 EU):
 - SoC: QCA9531
 - Flash: Winbond W25Q64FV (8MiB)
 - RAM: EtronTech EM6AB160TSE-5G (64MiB)
 - Wireless: SoC platform only (2.4GHz b/g/n, 2x internal antenna)
 - Ethernet: 2NIC (3x100M + 1x100M)
 - WWAN: TP-LINK LTE MODULE (2x external detachable antenna)
 - Power: DC 12V 1A

Signed-off-by: Filip Moc <lede@moc6.cz>
7 years agoramips: remove fictional LEDs from AsiaRF AWM00x device tree
Russell Senior [Wed, 24 May 2017 17:22:05 +0000 (10:22 -0700)]
ramips: remove fictional LEDs from AsiaRF AWM00x device tree

Neither the AsiaRF AWM002 or AWM003 actually has an LED on the module
board.  The ld1 and ld2 do not represent actual LEDs.  These pins might
connect to LEDS on an eval board or other carrier board, but that is
outside the scope of this device tree file.

Signed-off-by: Russell Senior <russell@personaltelco.net>
7 years agousbmode: update usb-modeswitch-data to 20170205
Julian Labus [Wed, 24 May 2017 14:32:17 +0000 (16:32 +0200)]
usbmode: update usb-modeswitch-data to 20170205

add support for new hardware

Signed-off-by: Julian Labus <julian@labus-online.de>
7 years agousbmode: update to latest version
Julian Labus [Wed, 24 May 2017 14:32:16 +0000 (16:32 +0200)]
usbmode: update to latest version

453da8e convert-modeswitch.pl: fix message indices

Signed-off-by: Julian Labus <julian@labus-online.de>
7 years agobuild: fix possible issue with kmod package having multiple AutoLoad's
Yousong Zhou [Sat, 27 May 2017 02:22:02 +0000 (10:22 +0800)]
build: fix possible issue with kmod package having multiple AutoLoad's

This commit contains the following changes

 - Use local shell var where appliable
 - The $(sort $$$$$$$$mods) call will have no expected effect
 - Avoid EEXIST when creating symlinks in /etc/modules-boot.d/
 - Avoid duplicate arguments for insert_modules() in postinst-pkg

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
7 years agokernel: update kernel 4.4 to 4.4.70
Hauke Mehrtens [Fri, 26 May 2017 21:38:19 +0000 (23:38 +0200)]
kernel: update kernel 4.4 to 4.4.70

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agokernel: update kernel 4.9 to 4.9.30
Hauke Mehrtens [Fri, 26 May 2017 20:56:32 +0000 (22:56 +0200)]
kernel: update kernel 4.9 to 4.9.30

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agouml: Fix sample command line
Florian Fainelli [Tue, 23 May 2017 03:21:55 +0000 (20:21 -0700)]
uml: Fix sample command line

Provide paths that match where LEDE is staging images.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agouml: Check for glibc static libraries
Florian Fainelli [Tue, 23 May 2017 02:35:00 +0000 (19:35 -0700)]
uml: Check for glibc static libraries

UML statically links against libutil, librt and libpthread. Some hosts do not
necessarily have these libraries installed and we should find out sooner than
later (during the final vmlinux linking stage) about that.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agokernel: Make KERNEL_PERF_EVENTS selectable
Florian Fainelli [Mon, 22 May 2017 23:46:46 +0000 (16:46 -0700)]
kernel: Make KERNEL_PERF_EVENTS selectable

The kernel itself allows enabling/disabling CONFIG_PERF_EVENTS, so allow
doing the same thing.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agokernel: Hide kernel options behind a menu
Florian Fainelli [Mon, 22 May 2017 23:45:52 +0000 (16:45 -0700)]
kernel: Hide kernel options behind a menu

We are starting to add more and more kernel configurable options, to the
point where the Global build options menu is not really usable anymore,
hide all kernel-related configuration options behind a menu.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agoelfutils: Pass -Wno-unused-result to silence warnings as errors
Florian Fainelli [Mon, 22 May 2017 23:08:57 +0000 (16:08 -0700)]
elfutils: Pass -Wno-unused-result to silence warnings as errors

elfutils turns on -Werror by default, and patch 100-musl-compat.patch
changes how strerror_r is used and we no longer use the function's
return value. This causes the following build error/warning to occur
with glibc-based toolchains:

dwfl_error.c: In function 'dwfl_errmsg':
dwfl_error.c:158:18: error: ignoring return value of 'strerror_r',
declared with attribute warn_unused_result [-Werror=unused-result]
       strerror_r (error & 0xffff, s, sizeof(s));
                  ^
cc1: all warnings being treated as errors

Fixing this would be tricky as there are two possible signatures for
strerror_r (XSI and GNU), just turn off unused-result warnings instead.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agodnsmasq: add dhcp-script hook for other packages
Nick Brassel [Sat, 6 May 2017 10:56:26 +0000 (20:56 +1000)]
dnsmasq: add dhcp-script hook for other packages

Adds a script which acts as a hook for when dnsmasq creates/destroys a
lease, or completes a TFTP file transfer. The hook loops through scripts
in appropriate directories inside '/etc/hotplug.d', executing each one with
the same arguments supplied by dnsmasq.

In case dnsmasq is jailed by ujail the dhcp-script hook will not work as
expected as ujail does not yet support executing a script within a jail.

Signed-off-by: Nick Brassel <nick@tzarc.org>
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
7 years agotreewide: fix device tree path in scripts
Mathias Kresin [Fri, 26 May 2017 13:13:42 +0000 (15:13 +0200)]
treewide: fix device tree path in scripts

The device tree is at /proc/device-tree/ without a base subdir.

Fixes: da472e5b30f6 ("treewide: access device tree from userspace via /proc/")
Signed-off-by: Mathias Kresin <dev@kresin.me>
7 years agoRevert "sysupgrade: run only one instance at a time."
Jo-Philipp Wich [Fri, 26 May 2017 12:48:05 +0000 (14:48 +0200)]
Revert "sysupgrade: run only one instance at a time."

This reverts commit e96a9a9af82c00dcce606a84a7bb87a00411385d.

The change breaks sysupgrade through LuCI and two-stage sysupgrade on
NAND targets. There is also a mismatch of file paths in lock and unlock
operations.

This commit was apparently neither properly tested, nor reviewed, so
drop it for now.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
7 years agoramips: add support for GL-inet GL-MT300N-V2
Kyson Lok [Wed, 17 May 2017 10:18:45 +0000 (18:18 +0800)]
ramips: add support for GL-inet GL-MT300N-V2

This patch adds supports for the GL-inet GL-MT300N-V2.

Specification:
- SoC: MediaTek MT7628AN
- Flash: 16 MiB (W25Q128FVSG)
- RAM: 128 MiB DDR
- Ethernet: 1 x WAN (100 Mbps) and 1 x LAN (100 Mbps)
- USB: 1 x USB 2.0 port
- Button: 1 x switch button, 1 x reset button
- LED: 3 x LEDS (system power led is not GPIO controller)
- UART: 1 x UART on PCB (JP1: 3.3V, RX, TX, GND)

Installation through Luci:
- The original firmware is LEDE, so both LuCI or sysupgrade can be used.
- Do not keep settings, for sysupgrade please use the -n option.

Installation through bootloader webserver:
- Plug power and hold reset button until red LED blink to bright.
- Install sysupgrade image using web interface on 192.168.1.1.

Signed-off-by: Kyson Lok <kysonlok@gmail.com>
[match maximum image size with firmware partition]
Signed-off-by: Mathias Kresin <dev@kresin.me>
7 years agotreewide: access device tree from userspace via /proc/
Mathias Kresin [Wed, 10 May 2017 08:44:18 +0000 (10:44 +0200)]
treewide: access device tree from userspace via /proc/

Access the device tree via /proc/device-tree/ is the documented way to
access the properties. Everything else might not work in future.

Signed-off-by: Mathias Kresin <dev@kresin.me>
7 years agoflex: update to 2.6.4 (FS#809)
Eric Luehrsen [Fri, 26 May 2017 04:02:47 +0000 (00:02 -0400)]
flex: update to 2.6.4 (FS#809)

flex 2.6.3 has a bug which fails code generation and compile of some packages

Signed-off-by: Eric Luehrsen <ericluehrsen@hotmail.com>
7 years agox86/legacy: disable PAE again
Daniel Golle [Thu, 25 May 2017 21:44:23 +0000 (23:44 +0200)]
x86/legacy: disable PAE again

commit 961c0eacea ('x86: fix lifting kernel CPU requirements and always
enable PAE') broke some older geode boards such as Soekris net4826.
Hence disable PAE on x86/legacy again in order to still support those
very old non-PAE capable CPUs.

Fixes FS#773 - PAE broke Soekris net4826

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
7 years agobuild: fix QUILT related overrides
Felix Fietkau [Thu, 25 May 2017 18:38:10 +0000 (20:38 +0200)]
build: fix QUILT related overrides

They need to be defined before including quilt.mk

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoiptables: fix typos in 600-shared-libext.patch (FS#711)
Felix Fietkau [Thu, 25 May 2017 17:30:06 +0000 (19:30 +0200)]
iptables: fix typos in 600-shared-libext.patch (FS#711)

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoDWR-512: adding wwan support for the dwr-512 3G modem
Giuseppe Lippolis [Thu, 22 Dec 2016 21:12:44 +0000 (22:12 +0100)]
DWR-512: adding wwan support for the dwr-512 3G modem

This PR allow the 3G modem embedded in the DWR-512 to be managed
by the wwan-ncm scripts. The modem will use the usb-option and
usb-cdc-ether drivers.
The DWR-512 DT is updated accordingly.

Signed-off-by: Giuseppe Lippolis <giu.lippolis@gmail.com>
7 years agogcc: fix documentation entries added by 910-mbsd_multi.patch
Felix Fietkau [Wed, 10 May 2017 12:01:54 +0000 (14:01 +0200)]
gcc: fix documentation entries added by 910-mbsd_multi.patch

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoscripts/download.pl: print the command used to download files
Felix Fietkau [Tue, 9 May 2017 13:20:42 +0000 (15:20 +0200)]
scripts/download.pl: print the command used to download files

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agoxfsprogs: update to 4.11.0
Felix Fietkau [Tue, 9 May 2017 13:20:23 +0000 (15:20 +0200)]
xfsprogs: update to 4.11.0

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agotools/libressl: update to 2.5.4
Hannu Nyman [Sun, 7 May 2017 17:52:32 +0000 (20:52 +0300)]
tools/libressl: update to 2.5.4

Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
7 years agotools/flex: update to 2.6.3
Hannu Nyman [Fri, 28 Apr 2017 13:53:42 +0000 (16:53 +0300)]
tools/flex: update to 2.6.3

* update flex to 2.6.3
* download .tar.gz (as upstream has dropped .xz)
* refresh patches

Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
7 years agotools/cmake: update to 3.8.1
Hannu Nyman [Mon, 8 May 2017 19:18:25 +0000 (22:18 +0300)]
tools/cmake: update to 3.8.1

* update cmake to 3.8.1
* refresh patches

Release notes:
https://cmake.org/cmake/help/v3.8/release/3.8.html

Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
7 years agox86_64: add drivers for Xeon controller hub based GPIO
Philip Prindeville [Tue, 28 Mar 2017 00:25:24 +0000 (18:25 -0600)]
x86_64: add drivers for Xeon controller hub based GPIO

These drivers are in many reference-design Xeon, iCore, or
Atom64 based server boards.

Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
7 years agojson-c: disable implicit fallthrough warning (gcc 7)
Felix Fietkau [Thu, 4 May 2017 13:45:35 +0000 (15:45 +0200)]
json-c: disable implicit fallthrough warning (gcc 7)

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agofirewall: update to the latest version, fixes a gcc7 build error
Felix Fietkau [Thu, 4 May 2017 14:11:27 +0000 (16:11 +0200)]
firewall: update to the latest version, fixes a gcc7 build error

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agotoolchain: add gcc 7.1.0 support
Felix Fietkau [Thu, 4 May 2017 13:18:13 +0000 (15:18 +0200)]
toolchain: add gcc 7.1.0 support

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agotools/isl: update to 0.18
Syrone Wong [Sun, 7 May 2017 11:39:12 +0000 (19:39 +0800)]
tools/isl: update to 0.18

Signed-off-by: Syrone Wong <wong.syrone@gmail.com>
7 years agoocteon: remove linux 4.4 support
Felix Fietkau [Thu, 4 May 2017 10:57:47 +0000 (12:57 +0200)]
octeon: remove linux 4.4 support

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: fix quilt for mixed package/host builds
Felix Fietkau [Wed, 3 May 2017 21:33:59 +0000 (23:33 +0200)]
build: fix quilt for mixed package/host builds

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: set QUILT=1 automatically when calling package host build refresh
Felix Fietkau [Wed, 3 May 2017 21:33:10 +0000 (23:33 +0200)]
build: set QUILT=1 automatically when calling package host build refresh

Makes behavor consistent with package builds and regular host builds

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agobuild: stop overriding STAGING_DIR_HOST for toolchain build
Felix Fietkau [Wed, 3 May 2017 13:01:03 +0000 (15:01 +0200)]
build: stop overriding STAGING_DIR_HOST for toolchain build

This causes various issues in other places that assume that host
binaries are staged in STAGING_DIR_HOST.
Since all the right places use HOST_BUILD_PREFIX, override that instead.
This fixes some issues with quilt on toolchain dirs

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agokernel: fix autoloading arch-specific modules
Yousong Zhou [Thu, 25 May 2017 06:41:34 +0000 (14:41 +0800)]
kernel: fix autoloading arch-specific modules

Fixes FS#745

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
7 years agobacklight-pwm: fix module description
Yousong Zhou [Thu, 25 May 2017 06:40:36 +0000 (14:40 +0800)]
backlight-pwm: fix module description

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
7 years agosysupgrade: run only one instance at a time.
Kenneth Johansson [Wed, 24 May 2017 11:45:37 +0000 (13:45 +0200)]
sysupgrade: run only one instance at a time.

Things do not work well if running multiple instances of
upgrade at the same time.

Signed-off-by: Kenneth Johansson <kenneth.johansson@inteno.se>
7 years agoinclude: Determine MODULES_DIR correctly for external/git kernels
Florian Fainelli [Thu, 11 May 2017 21:33:43 +0000 (14:33 -0700)]
include: Determine MODULES_DIR correctly for external/git kernels

When using external or git cloned kernels, any kind of modifications
will alter KERNELRELEASE. LEDE still tries to stage modules in
lib/modules/$(LINUX_UNAME_VERSION) and LINUX_UNAME_VERSION is based on
KERNEL_PATCHVER (indirectly) so this does not work, and we lose all
kinds of automatic modules loading.

To remedy that, just cat $(LINUX_DIR)/include/config/kernel.release
which is late enough the kernel has prepared this file, and is correctly
tracking changes done throughout the kernel.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agoinclude: Do not alter KERNELRELEASE for external/git kernels
Florian Fainelli [Thu, 11 May 2017 21:33:42 +0000 (14:33 -0700)]
include: Do not alter KERNELRELEASE for external/git kernels

In case we use external and/or git cloned kernels, let the kernel
determine the appropriate KERNELRELEASE. We cannot used
LINUX_UNAME_VERSION because that one gets determined at a later time,
when the kernel is already built proper.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agoRevert "kernel: prevent addition of scm marker to localversion"
Florian Fainelli [Thu, 11 May 2017 21:33:41 +0000 (14:33 -0700)]
Revert "kernel: prevent addition of scm marker to localversion"

This reverts commit 0df2c6563a3537ed21b28a9fb6874bf2718afd05 since it
gets in the way of identifying properly which kernel we are running.
This is particularly important if LEDE is using external kernels/git
cloned kernels. We want to make sure we only load modules from that
specific kernel.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agokernel: add kmod-usb-net-pl package
Roman Spychała [Tue, 25 Apr 2017 11:07:37 +0000 (13:07 +0200)]
kernel: add kmod-usb-net-pl package

Kernel support for Prolific PL-2301/2302/25A1 based cables

Signed-off-by: Roman Spychała <roed@onet.eu>
7 years agoMake GBit switch work on RB2011
René Mayrhofer [Thu, 27 Apr 2017 08:08:39 +0000 (10:08 +0200)]
Make GBit switch work on RB2011

This change is required to make the GBit switch work on my Mikrotik Routerboard RB2011UiAS-RM, and I assume that the other RB2011 variants are exactly the same in terms of the switch. I have tested the board without and with the patch and confirm that the GBit ports are not supported at all (i.e. no communication works) with the current version in trunk and that everything works with the patch applied. The test box has been running for a few days with the patch applied, and does not show any performance problems in a test setting. I have not used it with LEDE in production so far, but with a previous turnk version of OpenWRT for many years - with the same patch applied. I therefore have good indication that it is stable.

For the record, the switch chip on my test box is identified as
switch0: Atheros AR8327 rev. 4 switch registered on ag71xx-mdio.0

The value 0x6f000000 has been taken from the table at https://wiki.openwrt.org/toh/mikrotik/rb2011uias with the previous discussion thread still online at https://lists.openwrt.org/pipermail/openwrt-devel/2014-December/029949.html.
One definite improvement from the older OpenWRT trunk version I have been running in production and current LEDE trunk is that the SFP interface can be kept in the default configuration without excessive kernel messages about it constantly going up and down. I have not yet tested an actual SFP module, though.

Performance seems to be reasonable. Routing between two GBit ports on that switch separated by different VLANs with the default firewall ruleset (and one additional rule two allow traffic between the VLANs), but without NAT, iperf3 results are:
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec   508 MBytes   426 Mbits/sec  102             sender
[  4]   0.00-10.00  sec   506 MBytes   425 Mbits/sec                  receiver
With a connection going through NAT (also 2 ports on the same GBit switch, same ruleset, but NAT active), routing performance drops to around 250 MBit/s.
(Note that RouterOS achieves beyond 900 MBit/s on the same hardware with the default rule set and the FastTrack rule active even for NAT, see https://wiki.mikrotik.com/index.php?title=Manual:IP/Fasttrack and http://www.mikrotik.com/download/share/FastTrack.pdf).

Summarizing, I strongly recommend to apply this patch in trunk, so that the GBit switch chip rev. 4 can be supported upstream in the next LEDE release (hopefully soon).

Signed-off-by: René Mayrhofer <rene@mayrhofer.eu.org>
7 years agoar71xx: change image version for ubiquiti devices
Matthias Fritzsche [Wed, 17 May 2017 03:44:49 +0000 (05:44 +0200)]
ar71xx: change image version for ubiquiti devices

changes the image version from hardcoded OpenWrt to
$VERSION_DIST. AirOS shows a notification with the image version
during a firmware upgrade.

fixes #582

Signed-off-by: Matthias Fritzsche <txt.file@txtfile.eu>
7 years agolldpd: bump to 0.9.7
Stijn Tintel [Wed, 24 May 2017 12:56:22 +0000 (14:56 +0200)]
lldpd: bump to 0.9.7

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
7 years agosamba: fix CVE-2017-7494
Stijn Tintel [Wed, 24 May 2017 12:44:03 +0000 (14:44 +0200)]
samba: fix CVE-2017-7494

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
7 years agobrcm2708: enable cpufreq
Stijn Tintel [Sun, 14 May 2017 01:40:11 +0000 (03:40 +0200)]
brcm2708: enable cpufreq

With cpufreq disabled, the CPU stays locked at the frequency set by the
bootloader. This severely degrades performance as the bootloader sets
the CPU at the lowest frequency by default.

Enable cpufreq for all subtargets and use the ondemand governor.

Tested bcm2708 on RPi0W. Tested bcm2709 and bcm2710 on RPi3.

Reported-by: Bryan Mayland <bmayland@capnbry.net>
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
7 years agodnsmasq: bump to 2.77rc5
Hans Dedecker [Mon, 22 May 2017 20:56:20 +0000 (22:56 +0200)]
dnsmasq: bump to 2.77rc5

Some small tweaks and improvements :

9828ab1 Fix compiler warning.
f77700a Fix compiler warning.
0fbd980 Fix compiler warning.
43cdf1c Remove automatic IDN support when building i18n.
ff19b1a Fix &/&& confusion.
2aaea18 Add .gitattributes to substitute VERSION on export.

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
7 years ago6rd: add 6rd specific settings as nested json object
Hans Dedecker [Thu, 18 May 2017 09:34:44 +0000 (11:34 +0200)]
6rd: add 6rd specific settings as nested json object

Add 6rd specific settings prefix, relay-prefix as a nested data json object

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
7 years agonetifd: update to git HEAD version
Hans Dedecker [Mon, 22 May 2017 19:35:21 +0000 (21:35 +0200)]
netifd: update to git HEAD version

7573880 system-linux: parse 6rd specific settings as nested json data object
a063705 system-linux: remove redundant check for strtoul() return value
e6ebe0b build: disable unknown warning option error in clang
08d8f47 interface: add new "ifup-failed" hotplug event
20a1bac bridge: reset primary only after marking the member not present
6b9c267 build: suppress format truncation warnings to avoid errors with gcc7

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
7 years agoumdns: update to the version 2017-05-22
Rafał Miłecki [Mon, 22 May 2017 10:04:01 +0000 (12:04 +0200)]
umdns: update to the version 2017-05-22

This includes following changes:
0e8b948 Support specifying instance name in JSON file
49fdb9f Support PTR queries for a specific service
26ce7dc Allow filtering with instance name in service_reply
920c62a Store instance name in the struct service
ff09d9a Rename service_name function to the service_instance_name
64f78f1 Rename mdns_hostname variable to the umdns_host_label

Previous package update pulled commit 70c66fbbcde86 ("Fix sending
replies to PTR questions") which introduced a regression which this
update fixes.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
7 years agofstools: update to the latest version
Felix Fietkau [Thu, 4 May 2017 14:04:21 +0000 (16:04 +0200)]
fstools: update to the latest version

88d48d5 libfstools: silence mkfs.{ext4,f2fs}
a19f2b3 build: disable the format-truncation warning error to fix gcc 7 build errors
633a8d0 libfstools: fix multiple volume_identify usages with the same volume
c43ae11 fstools: use -Wno-format-truncation instead of -Wno-error=format-truncation

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 years agolibunwind: update to 1.2
Yousong Zhou [Mon, 22 May 2017 02:35:10 +0000 (10:35 +0800)]
libunwind: update to 1.2

Addresses CVE-2015-3239: Off-by-one error in the dwarf_to_unw_regnum
function in include/dwarf_i.h in libunwind 1.1 allows local users to
have unspecified impact via invalid dwarf opcodes.

Upstream stable-v1.2 fixed the missing unwind_i.h issue but no new
tarball is released yet

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
7 years agodropbear: bump to 2017.75
Kevin Darbyshire-Bryant [Sat, 20 May 2017 11:54:11 +0000 (12:54 +0100)]
dropbear: bump to 2017.75

- Security: Fix double-free in server TCP listener cleanup A double-free
in the server could be triggered by an authenticated user if dropbear is
running with -a (Allow connections to forwarded ports from any host)
This could potentially allow arbitrary code execution as root by an
authenticated user.  Affects versions 2013.56 to 2016.74. Thanks to Mark
Shepard for reporting the crash.
CVE-2017-9078 https://secure.ucc.asn.au/hg/dropbear/rev/c8114a48837c

- Security: Fix information disclosure with ~/.ssh/authorized_keys
symlink.  Dropbear parsed authorized_keys as root, even if it were a
symlink.  The fix is to switch to user permissions when opening
authorized_keys

A user could symlink their ~/.ssh/authorized_keys to a root-owned file
they couldn't normally read. If they managed to get that file to contain
valid authorized_keys with command= options it might be possible to read
other contents of that file.
This information disclosure is to an already authenticated user.
Thanks to Jann Horn of Google Project Zero for reporting this.
CVE-2017-9079 https://secure.ucc.asn.au/hg/dropbear/rev/0d889b068123

Refresh patches, rework 100-pubkey_path.patch to work with new
authorized_keys validation.

Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
7 years agomac80211: add support for rtl8821ae pcie adapter
Hans Ulli Kroll [Fri, 19 May 2017 18:17:08 +0000 (20:17 +0200)]
mac80211: add support for rtl8821ae pcie adapter

Add support for Realtek RTL8821AE/RTL8812AE PCIe adapter.
This device supports 802.11ac and bluetooth

testet on PC Engines APU with AP and STA mode

Signed-off-by: Hans Ulli Kroll <ulli.kroll@googlemail.com>
7 years agolinux-firmware: add firmware for rtl8821ae support
Hans Ulli Kroll [Fri, 19 May 2017 18:17:07 +0000 (20:17 +0200)]
linux-firmware: add firmware for rtl8821ae support

Add needed firmware to support rtl8821ae pcie adapter

Signed-off-by: Hans Ulli Kroll <ulli.kroll@googlemail.com>
7 years agolantiq: spi: double time out tolerance
Hauke Mehrtens [Sun, 21 May 2017 19:20:44 +0000 (21:20 +0200)]
lantiq: spi: double time out tolerance

The generic SPI code calculates how long the issued transfer would take
and adds 100ms in addition to the timeout as tolerance. On my 500 MHz
Lantiq Mips SoC I am getting timeouts from the SPI like this when the
system boots up:

m25p80 spi32766.4: SPI transfer timed out
blk_update_request: I/O error, dev mtdblock3, sector 2
SQUASHFS error: squashfs_read_data failed to read block 0x6e

After increasing the tolerance for the timeout to 200ms I haven't seen
these SPI transfer time outs any more.
The Lantiq SPI driver in use here has an extra work queue in between,
which gets triggered when the controller send the last word and the
hardware FIFOs used for reading and writing are only 8 words long.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agokernel: update kernel 4.9 to 4.9.29
Koen Vandeputte [Mon, 15 May 2017 11:11:05 +0000 (13:11 +0200)]
kernel: update kernel 4.9 to 4.9.29

- Refresh all patches
- Removed upstreamed
- Adapted 1

Compile tested on: bcm53xx, cns3xxx, imx6, lantiq
Run tested on: cns3xxx & imx6

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
[update from 4.9.28 to 4.9.29]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agokernel: update kernel 4.4 to version 4.4.69
Kevin Darbyshire-Bryant [Mon, 15 May 2017 14:03:47 +0000 (15:03 +0100)]
kernel: update kernel 4.4 to version 4.4.69

Refresh patches. A number of patches have landed upstream & hence are no
longer required locally:

062-[1-6]-MIPS-* series
042-0004-mtd-bcm47xxpart-fix-parsing-first-block

Reintroduced lantiq/patches-4.4/0050-MIPS-Lantiq-Fix-cascaded-IRQ-setup
as it was incorrectly included upstream thus dropped from LEDE.
As it has now been reverted upstream it needs to be included again for
LEDE.

Run tested ar71xx Archer C7 v2 and lantiq.

Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
[update from 4.4.68 to 4.4.69]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
7 years agobcm53xx: add support for TP-LINK Archer C5 V2
Rafał Miłecki [Sun, 21 May 2017 15:41:41 +0000 (17:41 +0200)]
bcm53xx: add support for TP-LINK Archer C5 V2

This model also contains few partitions non-discoverable partitions we
need to "protect". Othen than that it uses non-deprecated serial entry
in DTS that doesn't work with LEDE so we need to workaround it as well.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
7 years agobase-files: fix default procd reload
Alexandru Ardelean [Fri, 19 May 2017 13:19:20 +0000 (16:19 +0300)]
base-files: fix default procd reload

Bug introduced with 6713694.

I did not count on procd handling reload as mentioned
in this doc:
https://wiki.openwrt.org/inbox/procd-init-scripts

```
procd_set_param file /var/etc/your_service.conf # /etc/init.d/your_service reload will restart the daemon if these files have changed
procd_set_param netdev dev # likewise, except if dev's ifindex changes.
procd_set_param data name=value ... # likewise, except if this data changes.
```

The service would be restarted regardless of any of those params.

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
7 years agokernel: enable CRASH_DUMP on supported platforms
Daniel Golle [Fri, 19 May 2017 22:20:35 +0000 (00:20 +0200)]
kernel: enable CRASH_DUMP on supported platforms

While we have CRASHLOG on MIPS it makes sense to support 'classic'
kexec-based CRASH_DUMP on x86 and arm platforms.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
7 years agoRevert "ar71xx: Add support for Teltonika RUT900"
John Crispin [Thu, 18 May 2017 13:31:26 +0000 (15:31 +0200)]
Revert "ar71xx: Add support for Teltonika RUT900"

This reverts commit 224e5f5efa22b8a0132522367afd3b22f05d53e1.

pepe2k pointed out that this was not ready to merge

Signed-off-by: John Crispin <john@phrozen.org>
7 years agoramips: support jumbo frame on mt7621 up to 2k
Michael Lee [Thu, 27 Apr 2017 01:04:31 +0000 (09:04 +0800)]
ramips: support jumbo frame on mt7621 up to 2k

Signed-off-by: Michael Lee <igvtee@gmail.com>
7 years agoar71xx: Add support for Teltonika RUT900
Steffen Weinreich [Thu, 27 Apr 2017 21:59:22 +0000 (23:59 +0200)]
ar71xx: Add support for Teltonika RUT900

Teltonika RUT900 is a Router with LTE dual SIM, WiFi, 4x Ethernet
ports, I/O, RS232, RS485, GPS.

The device ist based on a Atheros AR9344 rev 3,

Specifications:
- 560/450/225 MHz (CPU/DDR/AHB)
- 128 MB of RAM
- 16 MB of FLASH
- Serial Console header on a Card Board edge connector
- 4x 10/100 Mbps Ethernet (3x LAN, 1x WAN)
- 2.4 GHz Wifi
- 2x external, detachable Wifi antennas
- LTE Modem Huawei ME909u-521 (Also other Modem seen)
- 2x LTE antennas
- 1x GPS antenna
- 7x LED, 1x button
- 1x USB Connector
- 1x Serial RS232
- 1x Serial RS485
- 1x MicroSD Card

The GPL sources of the device are available at www.teltonika.lt/gpl/
and are based on OpenWRT Barrier Breaker (14.07)

Running from tftp:

The Router starts into the uboot Webupdater if the Button ist pressed
more than 3 seconds, if no Network cable is attached it starts the
uboot serial console, from there the router loads the firmware image
via tftpboot from 192.168.1.2:firmware.bin (the router has the
192.168.1.1). With bootm the loaded image will be booted.

Signed-off-by: Steffen Weinreich <steve@weinreich.org>
7 years agof2fs-tools: Switch to gz tarball
Daniel Engberg [Wed, 10 May 2017 09:04:26 +0000 (11:04 +0200)]
f2fs-tools: Switch to gz tarball

At some point kernel.org decided to drop xz generated tarballs, switch to gz which they still provide.

Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>