oweals/openwrt.git
5 years agoxfsprogs: Replace valloc with posix_memalign
Rosen Penev [Mon, 6 May 2019 21:57:18 +0000 (14:57 -0700)]
xfsprogs: Replace valloc with posix_memalign

Fixes compilation under uClibc-ng.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
5 years agogeneric-4.19: Backport spi-nand support for GigaDevice A/E
Jeff Kletsky [Fri, 10 May 2019 13:01:47 +0000 (06:01 -0700)]
generic-4.19: Backport spi-nand support for GigaDevice A/E

This patch backports verbatim the commits from Linux 5.0 and 5.1
that implemented support for GigaDevice SPI NAND A and E variants.

Supported only in Linux 4.19 and later as based on the upstream
drivers/mtd/nand/spi/ framework.

mtd-spinand-add-support-for-GigaDevice-GD5FxGQ4xA.patch
commit c93c613214ac (5.0)

mtd-spinand-Add-support-for-GigaDevice-GD5F1GQ4UExxG.patch
commit c40c7a990a46 (5.1)

Run-tested-on: GL.iNet AR750S

Signed-off-by: Jeff Kletsky <git-commits@allycomm.com>
5 years agokernel: Reorder generic configuration
Hauke Mehrtens [Fri, 10 May 2019 20:46:49 +0000 (22:46 +0200)]
kernel: Reorder generic configuration

This was done like this:
./scripts/kconfig.pl '+' target/linux/generic/config-4.14 /dev/null > target/linux/generic/config-4.14-new
mv target/linux/generic/config-4.14-new target/linux/generic/config-4.14

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agogemini: Make kernel text and rodata read-only
Hauke Mehrtens [Sun, 7 Apr 2019 14:53:11 +0000 (16:53 +0200)]
gemini: Make kernel text and rodata read-only

This is activate for all other targets except gemini, also activate it
there.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agocns3xxx: Activate CONFIG_CPU_SW_DOMAIN_PAN
Hauke Mehrtens [Sun, 7 Apr 2019 16:03:54 +0000 (18:03 +0200)]
cns3xxx: Activate CONFIG_CPU_SW_DOMAIN_PAN

This is already activated for all other ARM targets

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agox86: Activate CONFIG_X86_SMAP
Hauke Mehrtens [Sun, 7 Apr 2019 17:38:04 +0000 (19:38 +0200)]
x86: Activate CONFIG_X86_SMAP

This activates "Supervisor Mode Access Prevention". modern CPUs will
prevent the kernel code from accessing any data from the userspace
without the usage of copy_to_user() or copy_from_user()

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agokernel: Deactivate CONFIG_BINFMT_MISC
Hauke Mehrtens [Sun, 7 Apr 2019 17:08:18 +0000 (19:08 +0200)]
kernel: Deactivate CONFIG_BINFMT_MISC

CONFIG_BINFMT_MISC allows it to add support for new executable formats
to the kernel from user space, the kernel will then detect for example a
java binary and call the java execution program automatically. I am not
aware that this feature is used in OpenWrt and this could be used to
exploit something. Deactivate it for all targets for now.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agokernel: Do not set CONFIG_DEVMEM or CONFIG_DEVKMEM
Hauke Mehrtens [Sun, 7 Apr 2019 17:03:54 +0000 (19:03 +0200)]
kernel: Do not set CONFIG_DEVMEM or CONFIG_DEVKMEM

This is deactivated by default and should be manually activated in the
OpenWrt kernel configuration

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agokernel: Activate CONFIG_SYN_COOKIES for all targets
Hauke Mehrtens [Sun, 7 Apr 2019 17:00:45 +0000 (19:00 +0200)]
kernel: Activate CONFIG_SYN_COOKIES for all targets

Some targets deactivated CONFIG_SYN_COOKIES, for unknown reasons, use
the default setting from the generic configuration which activates
CONFIG_SYN_COOKIES.

This should prevent SYN flooding.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agokernel: Remove CONFIG_COMPAT
Hauke Mehrtens [Sun, 7 Apr 2019 15:14:24 +0000 (17:14 +0200)]
kernel: Remove CONFIG_COMPAT

This removes support for executing old 32 bit applications on 64 bit ARM
and MIPS kernels.
On OpenWrt we normally compile all the user space applications on our
own and do not support third party binary only modules especial not 32
bit applications on 64 bit CPUs.

This reduces the attack surface on such systems and should also save
some memory.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agokernel: Remove CONFIG_DEBUG_RODATA and CONFIG_DEBUG_SET_MODULE_RONX
Hauke Mehrtens [Sun, 7 Apr 2019 14:55:38 +0000 (16:55 +0200)]
kernel: Remove CONFIG_DEBUG_RODATA and CONFIG_DEBUG_SET_MODULE_RONX

These were renamed to CONFIG_STRICT_KERNEL_RWX and CONFIG_STRICT_MODULE_RWX and are
activated in kernel 4.14 and later by default.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agokernel: Activate CONFIG_ARM64_SW_TTBR0_PAN
Hauke Mehrtens [Sun, 7 Apr 2019 16:06:34 +0000 (18:06 +0200)]
kernel: Activate CONFIG_ARM64_SW_TTBR0_PAN

This activates "Emulate Privileged Access Never using TTBR0_EL1
switching" on ARM64.

This should prevent the kernel from reading code from user space in
kernel context.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoKernel: Activate CONFIG_HARDENED_USERCOPY
Hauke Mehrtens [Sun, 7 Apr 2019 14:38:44 +0000 (16:38 +0200)]
Kernel: Activate CONFIG_HARDENED_USERCOPY

This adds additional checks to the copy_from_user() and copy_to_user()
functions. The details are described in this article:
https://lwn.net/Articles/695991/

This should only have a very small performance impact on system calls
and should not affect routing performance.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoutils/spidev_test: build package directly from Linux
Christian Lamparter [Fri, 10 May 2019 21:31:01 +0000 (23:31 +0200)]
utils/spidev_test: build package directly from Linux

Jeff Kletsky noted in his patch titled:
"utils/spidev_test: Update to current source from upstream Linux"
that the spidev_test utility OpenWrt ships is severly out of date.

Instead of updating the spidev_test.c from the current kernel,
this patch replaces the package building code to utilize the
very file that gets shipped with the kernel we compiling for
anyway much like the "perf" package already does.

Reported-by: Jeff Kletsky <git-commits@allycomm.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
5 years agoath79: Add missing read-only properties
Adrian Schmutzler [Thu, 9 May 2019 13:31:05 +0000 (15:31 +0200)]
ath79: Add missing read-only properties

Add some read-only properties to protect partitions from
accidental changes.

Also fixed two whitespaces error on the way.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
5 years agolldpd: add option to edit hostname
Lucian Cristian [Thu, 9 May 2019 21:19:07 +0000 (00:19 +0300)]
lldpd: add option to edit hostname

also fixes the annoying repeating syslog
lldp[]: unable to get system name

Signed-off-by: Lucian Cristian <lucian.cristian@gmail.com>
5 years agolldpd: update to 1.0.3
Lucian Cristian [Tue, 30 Apr 2019 14:32:47 +0000 (17:32 +0300)]
lldpd: update to 1.0.3

Support for CDP PD PoE

Signed-off-by: Lucian Cristian <lucian.cristian@gmail.com>
5 years agoath10k-ct: Update to current version
Robert Marko [Thu, 9 May 2019 15:27:58 +0000 (17:27 +0200)]
ath10k-ct: Update to current version

This patch updates ath10k-ct to current version.
Changes are:
     ath10k-ct:  Fix printing PN in peer stats.

     Previous logic was incorrect.  Also add set-special API to enable
     returning PN.

Patches refreshed and tested on 8devices Jalapeno dev board(IPQ4019)

Signed-off-by: Robert Marko <robimarko@gmail.com>
5 years agoath10k-firmware: update Candela Tech firmware images
Robert Marko [Thu, 9 May 2019 15:27:57 +0000 (17:27 +0200)]
ath10k-firmware: update Candela Tech firmware images

Release notes since last time:

Release notes for wave-1:

2019-04-02: Support some get/set API for eeprom rate power tables.
    Mostly backported from 10.2

2019-04-02: Support adaptive-CCA, backported from 10.2

2019-04-02: Support adding eeprom configAddr pairs via the
            set-special API. These configAddrs can be used to change
            the default register settings for up to 12 registers.

2019-05-03: Fix tx-power settings for 2x2, 3x3 rates.
    Original logic I put in back in 2016 set 2x2 and 3x3 lower
    than the needed to be when using most NICs (very high
    powered NICs would not have been affected I think, not sure
    any of those exist though.)

    This improves throughput for 2x2 and 3x3 devices,
    especially when the signal is weaker.

Release notes for wave-2:

2019-04-08: When setting keys, if high bit of high value of
    key_rsc_counter is set to 0x1, then the lower 48 bits will
    be used as the PN value.  By default, PN is set to 1 each
    time the key is set.

2019-04-08: Pack PN into un-used 'excretries' aka
    'num_pkt_loss_excess_retry' high 16 bits.
    This lets us report peer PN, but *only* if driver has
    previously set a PN when setting key (or set-special cmd is
    used to enable PN reporting).

    This is done so that we know the driver is recent
            enough to deal with the PN stat reporting.

2019-04-16: Support specifying tx rate on a per-beacon packet.
    See ath10k_wmi_op_gen_beacon_dma and
    ath10k_convert_hw_rate_to_rate_info for API details.

     Driver needs additional work to actually enable this
     feature currently.

2019-04-30: Compile out tx-prefetch caching logic.
    It is full of tricky bugs that cause tx hangs.
    I fixed at least one, but more remain and I have wasted too
    much time on this already.

2019-05-08: Start rate-ctrl at mcs-3 instead of mcs-5.
    This significantly helps DHCP happen quickly, probably
    because the initial rate being too high would take a while
    to ramp down, especially since there are few packets sent
    by the time DHCP needs to start.

    This bug was triggered by me decreasing retries of 0x1e
    (upstream default) to 0x4.  But, I think it is better to
    start with lower initial MCS instead of always having a
    very high retry count.

Tested on 8devices Jalapeno dev board(IPQ4019)

Signed-off-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com> [neatify]
5 years agoramips: fix R6120 factory image
David Bauer [Thu, 9 May 2019 16:04:06 +0000 (18:04 +0200)]
ramips: fix R6120 factory image

The factory firmware omits the JFFS2 end-marker while flashing via
web-interface. Add a 64k padding after the marker fixes this problem.

When the end-marker is not present, OpenWRT won't save the overlayfs
after initial flash.

Reported-by: Andreas Ziegler <dev@andreas-ziegler.de>
Signed-off-by: David Bauer <mail@david-bauer.net>
5 years agomvebu: fix board_name condition in 79_move_config
Klaus Kudielka [Wed, 8 May 2019 18:40:51 +0000 (20:40 +0200)]
mvebu: fix board_name condition in 79_move_config

The correct board_name for the Turris Omnia is armada-385-turris-omnia.

Fixes: 4e8345ff68 ("mvebu: base-files: autodetect upgrade device")
Signed-off-by: Klaus Kudielka <klaus.kudielka@gmail.com>
5 years agobase-files: improve lib/upgrade/common.sh
Klaus Kudielka [Wed, 8 May 2019 18:40:50 +0000 (20:40 +0200)]
base-files: improve lib/upgrade/common.sh

Recently, upgrade device autodetection has been added to the mvebu target.
This exposes some shortcomings of the generic export_bootdevice function,
e.g. on the Turris Omnia: export_bootdevice silently reports the root
partition to be the boot device. This makes the sysupgrade process fail at
several places.

Fix this by clearly distinguishing between /proc/cmdline arguments which
specify the boot disk, and those which specify the root partition. Only in
the latter case, strip off the partition, and do it consistently.
root=PARTUUID=<pseudo PARTUUID for MBR> (any partition) and root=/dev/*
(any partition) are accepted.

The root of the problem is that the *existing* export_bootdevice in
/lib/upgrade/common.sh behaves differently, if the kernel is booted with
root=/dev/..., or if it is booted with root=PARTUUID=...

In the former case, it reports back major/minor of the root partition,
in the latter case it reports back major/minor of the complete boot disk.

Targets, which boot with root=/dev/... *and* use export_bootdevice /
export_partdevice, have added workarounds to this behaviour, by specifying
*negative* increments to the export_partdevice function.

Consequently, those targets have to be adapted to use positive increments,
otherwise they are broken by the change to export_bootdevice.

Fixes: 4e8345ff68 ("mvebu: base-files: autodetect upgrade device")
Signed-off-by: Klaus Kudielka <klaus.kudielka@gmail.com>
Tested-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
5 years agoipq40xx: add factory image for EnGenius ENS620EXT
Steve Glennon [Tue, 9 Apr 2019 20:46:32 +0000 (14:46 -0600)]
ipq40xx: add factory image for EnGenius ENS620EXT

Extended  mksenaofw to support new "capwap" header structure.
This supports flashing from factory 3.0.0, 3.0.1, 3.1.0 and 3.5.5
firmware.

Note that the factory image format changes for 3.1 and later firmware,
and that the 3.1.0 and 3.5.5 Engenius firmware will refuse the
factory_30.bin file. Similarly, the 3.0.0 and 3.0.1 Engenius firmware
will refuse the factory_35.bin file.

Flashing from the Engenius 3.1.0 firmware with the factory_35.bin
firmware has not been tested, as 3.1.0 firmware (Engenius "middleFW")
is only intended as part of the upgrade path to 3.5.5 firmware.

Modified ipq40xx image Makefile to appropriately invoke mksenaofw
with new parameters to configure the capwap header.

Note that there is currently no method to return to factory firmware,
so this is a one-way street.

Path from factory 3.0.0 and 3.0.1 (EnGenius) software to OpenWrt is
to navigate to 192.168.1.1 on the stock firmware and navigate to the
firmware menu. Then copy the URL you have for that page, something like
http://192.168.1.1/cgi-bin/luci/;stok=12345abcdef/admin/system/flashops
and replace the trailing /admin/system/flashops with just /easyflashops

You should then be presented with a simple "Firmware Upgrade" page.
On that page, BE SURE TO CLEAR the "Keep Settings:" checkbox.

Choose the openwrt-ipq40xx-engenius_ens620ext-squashfs-factory_30.bin,
click "Upgrade" and on the following page select "Proceed".

Path from factory 3.5.5 (EnGenius) software to OpenWrt is simply to
use the stock firmware update menu. Choose the
openwrt-ipq40xx-engenius_ens620ext-squashfs-factory_35.bin and click
"Upload" and "Proceed".

The device should then flash the OpenWrt firmware and reboot. Note
that this resets the device to a default configuration with Wi-Fi
disabled, LAN1/PoE acting as a WAN port (running DHCP client) and LAN2
acting as a LAN port with a DHCP server on 192.168.1.x (AP is at
192.168.1.1)

Signed-off-by: Steve Glennon <s.glennon@cablelabs.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
[sorry, for unfixing the 80-lines eyesores.]

5 years agoipq40xx: Fix reboot on EnGenius ENS620EXT
Steve Glennon [Tue, 23 Apr 2019 20:46:37 +0000 (14:46 -0600)]
ipq40xx: Fix reboot on EnGenius ENS620EXT

This patch works around an issue where reboot would cause the AP
to power down and not reboot.

The ipq4019 restart controller reboot causes the system
to power down and not recover. Fix is to disable the restart
controller in the device tree and the device reverts to
using the watchdog to perform the reset.

The real problem is due to the buggy bootloader that ships
with the device. Steve Glennon reported in the PR for this
patch: <https://github.com/openwrt/openwrt/pull/2009> that:

"the problem was due to a bad u-boot that ships with the device.

Using the u-boot that comes with 3.5.5.3 EnGenius factory
software now allows the old code (using the do_msm_reboot)
to reboot successfully.

On to the bad news:
Well 3.5.5.3 is a bad path. Finally managed to recover. You
CANNOT use prior EnGenius firmware to downgrade.

Findings:

* They now password protect the serial console with a new, unkown
  password.
* They changed the protection on their walled-garden. I have to
  use the ssh admin@ip /bin/sh --login to get out of their
  walled-garden.
* Attempts to flash the original 3.0.0 or 3.0.1 EnGenius firmware
  fail through the UI and sysupgrade. Their firmware update GUI now
  seem to detect regular openwrt images, but they fail to flash
  Attempts to flash a normal OpenWrt image with sysupgrade fail.
[..]

Attempts to sysupgrade with EnGenius firmware fail with the same
"mandatory section(s) missing" error, so you cannot downgrade to
3.0.0 or 3.0.1."

Signed-off-by: Steve Glennon <s.glennon@cablelabs.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com> [added valuable
findings from github discussion]

5 years agokirkwood: enable sata port multiplier
Alberto Bursi [Sat, 20 Apr 2019 15:25:29 +0000 (17:25 +0200)]
kirkwood: enable sata port multiplier

Marvell sata controllers in all kirkwood SoCs support
sata port multipliers, just like mvebu.
Enable this feature in the default kernel config
so it is available in normal builds.
tested and working on nsa310b

Signed-off-by: Alberto Bursi <alberto.bursi@outlook.it>
5 years agobuild: add a config option for enabling a testing version of the target kernel
Felix Fietkau [Sat, 27 Apr 2019 17:13:07 +0000 (19:13 +0200)]
build: add a config option for enabling a testing version of the target kernel

If the target supports a newer kernel version that is not used by default
yet, it can be enabled with this option

Signed-off-by: Felix Fietkau <nbd@nbd.name>
5 years agotoolchain/nasm: update to version 2.14
Daniel Golle [Fri, 10 May 2019 23:25:12 +0000 (01:25 +0200)]
toolchain/nasm: update to version 2.14

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
5 years agoramips: mt7620: fix dependencies
Deng Qingfang [Fri, 19 Apr 2019 10:04:56 +0000 (18:04 +0800)]
ramips: mt7620: fix dependencies

MT7620 integrated WMAC does not need RT2x00 PCI driver or firmware
Also corrected kmod-eeprom-93cx6 and kmod-lib-crc-itu-t dependencies
according to original Kconfig and lsmod output

This will remove some unnecessary packages from MT7620 target to
save some space

Signed-off-by: Deng Qingfang <dengqf6@mail2.sysu.edu.cn>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
[75 characters per line in the commit message]

5 years agoprocd: fix compile issue
Hans Dedecker [Thu, 9 May 2019 16:31:01 +0000 (18:31 +0200)]
procd: fix compile issue

1361b97 container: include stdbool.h

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agonetifd: update to latest git HEAD
Hans Dedecker [Wed, 8 May 2019 19:52:20 +0000 (21:52 +0200)]
netifd: update to latest git HEAD

f6fb700 interface-ip: fine tune IPv6 mtu warning
975a5c4 interface: tidy ipv6 mtu warning

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agoprocd: update to latest git HEAD
Hans Dedecker [Wed, 8 May 2019 18:50:09 +0000 (20:50 +0200)]
procd: update to latest git HEAD

9b35439 procd: detect lxc container and behave accordingly

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agokernel: bump 4.19 to 4.19.38
Koen Vandeputte [Thu, 2 May 2019 12:31:34 +0000 (14:31 +0200)]
kernel: bump 4.19 to 4.19.38

Refreshed all patches.

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

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.14 to 4.14.115
Koen Vandeputte [Thu, 2 May 2019 12:30:50 +0000 (14:30 +0200)]
kernel: bump 4.14 to 4.14.115

Refreshed all patches.

Compile-tested on: ar71xx, cns3xxx, imx6, x86_64
Runtime-tested on: ar71xx, cns3xxx, imx6

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.9 to 4.9.172
Koen Vandeputte [Thu, 2 May 2019 12:25:32 +0000 (14:25 +0200)]
kernel: bump 4.9 to 4.9.172

Refreshed all patches.

Compile-tested on: ar7
Runtime-tested on: none

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agoelfutils: Fix compile with uClibc-ng
Rosen Penev [Wed, 1 May 2019 17:05:20 +0000 (10:05 -0700)]
elfutils: Fix compile with uClibc-ng

Probably glibc too. argp_help takes a char *. not const char *.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
[updated with upstream version of the patch]

5 years agouboot-tegra: bump to 2019.04
Tomasz Maciej Nowak [Thu, 25 Apr 2019 13:38:16 +0000 (15:38 +0200)]
uboot-tegra: bump to 2019.04

This version has important change for tegra boards which is reserving
32MB memory for Linux kernel instead of current 16MB.

Signed-off-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
5 years agoath79: glinet_gl-ar750s: Use QCA9887 firmware
Jeff Kletsky [Fri, 3 May 2019 11:59:59 +0000 (04:59 -0700)]
ath79: glinet_gl-ar750s: Use QCA9887 firmware

The GL.iNet AR750S is based around the QCA9563
and requires the QCA9887 firmware for operation.

Signed-off-by: Jeff Kletsky <git-commits@allycomm.com>
5 years agoar71xx: fix failsafe interface for TL-WR940nv6
Joachim Fünfer [Mon, 29 Apr 2019 10:53:08 +0000 (12:53 +0200)]
ar71xx: fix failsafe interface for TL-WR940nv6

Switches failsafe mode interface from WAN to LAN ports.

Tested on TL-WR940Nv6.0 and TL-WR940Nv6.1

Signed-off-by: Joachim Fünfer <joachim.fuenfer@stusta.net>
5 years agoath79: fix OCEDO Raccoon 10Mbit PLL value
David Bauer [Mon, 22 Apr 2019 19:29:42 +0000 (21:29 +0200)]
ath79: fix OCEDO Raccoon 10Mbit PLL value

This corrects the PLL value for 10 Mbit/s links on the OCEDO Raccoon.
Prior to this patch, 10 Mbit/s links would not transmit data.

It is worth mentioning that the vendor firmware used the same PLL
settings and 10Mbit/s was also not working there.

All other link-modes are working correctly without any packet loss.

Signed-off-by: David Bauer <mail@david-bauer.net>
5 years agobase-files: add service_stopped as a post stop hook
Arthur Skowronek [Wed, 30 Jan 2019 13:58:47 +0000 (14:58 +0100)]
base-files: add service_stopped as a post stop hook

Purpose of these changes is to introduce a hook for post service
shutdown in a similar fashion to the existing hook service_started. I
found it to be useful to specify a hook that is called once the service
has been stopped and not before the service is stopped like the
stop_service hook does.

The concrete use case I have for this is that I'm running a binary that
takes over the hardware watchdog timer. Said binary unfortunately can
not use ubus directly to tell procd to hand over the watchdog timer so
this has to be done in the service file for the binary in question. In
order to support a clean handover of the watchdog timer back to procd,
the service init script has to dispatch the ubus invocation once the
binary in question has been stopped.

Signed-off-by: Arthur Skowronek <ags@digineo.de>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
[added commit message, use the same form as other hooks]

5 years agoar71xx: mute script error in ar922x-led-fix script
Rosy Song [Tue, 16 Apr 2019 01:05:14 +0000 (09:05 +0800)]
ar71xx: mute script error in ar922x-led-fix script

This commit fixes following script error in syslog:

 cat: can't open '/sys/devices/platform/ehci-platform/usb1/1-1/1-1.2/1-1.2:1.4/ieee80211/phy*/name': No such file or directory
 sh: add: unknown operand
 sh: add: unknown operand

Signed-off-by: Rosy Song <rosysong@rosinson.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
[commit subject and message tweaks]

5 years agokernel: Rename UPROBE_EVENT to UPROBE_EVENTS
Petr Štetiar [Wed, 10 Apr 2019 15:07:34 +0000 (17:07 +0200)]
kernel: Rename UPROBE_EVENT to UPROBE_EVENTS

Upstream has renamed UPROBE_EVENT to UPROBE_EVENTS in the following
commit:

 commit 6b0b7551428e4caae1e2c023a529465a9a9ae2d4
 Author: Anton Blanchard <anton@samba.org>
 Date:   Thu Feb 16 17:00:50 2017 +1100

     perf/core: Rename CONFIG_[UK]PROBE_EVENT to CONFIG_[UK]PROBE_EVENTS

     We have uses of CONFIG_UPROBE_EVENT and CONFIG_KPROBE_EVENT as
     well as CONFIG_UPROBE_EVENTS and CONFIG_KPROBE_EVENTS.

     Consistently use the plurals.

So I'm changing it to this plural option in order to make kconfig happy
and stop asking about it if kernel is compiled with verbose logging:

 Enable uprobes-based dynamic events (UPROBE_EVENTS) [Y/n/?] (NEW)

Signed-off-by: Petr Štetiar <ynezz@true.cz>
5 years agoconfig: kernel: Add KPROBE_EVENTS config option
Petr Štetiar [Wed, 10 Apr 2019 15:07:34 +0000 (17:07 +0200)]
config: kernel: Add KPROBE_EVENTS config option

Upstream has renamed KPROBE_EVENT to KPROBE_EVENTS in the following
commit:

 commit 6b0b7551428e4caae1e2c023a529465a9a9ae2d4
 Author: Anton Blanchard <anton@samba.org>
 Date:   Thu Feb 16 17:00:50 2017 +1100

     perf/core: Rename CONFIG_[UK]PROBE_EVENT to CONFIG_[UK]PROBE_EVENTS

     We have uses of CONFIG_UPROBE_EVENT and CONFIG_KPROBE_EVENT as
     well as CONFIG_UPROBE_EVENTS and CONFIG_KPROBE_EVENTS.

     Consistently use the plurals.

So I'm adding this plural option in order to make kconfig happy and stop
asking about it if kernel is compiled with verbose logging:

  Enable kprobes-based dynamic events (KPROBE_EVENTS) [Y/n/?] (NEW)

Signed-off-by: Petr Štetiar <ynezz@true.cz>
5 years agokernel: Remove support for kernel 3.18
Hauke Mehrtens [Mon, 25 Mar 2019 14:29:06 +0000 (15:29 +0100)]
kernel: Remove support for kernel 3.18

No target is using kernel 3.18 anymore, remove all the generic
support for kernel 3.18.

The removed packages are depending on kernel 3.18 only and are not used on
any recent kernel.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoxburst: Remove unmaintained target
Hauke Mehrtens [Mon, 25 Mar 2019 14:15:57 +0000 (15:15 +0100)]
xburst: Remove unmaintained target

This target only supports kernel 3.18, which is not supported in OpenWrt
any more for multiple releases. It also looks like there is no active
maintainer for this target.
Remove the code and all the packages which are only used by this target.

To add this target to OpenWrt again port it to a recent and supported
kernel version.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoomap24xx: Remove unmaintained target
Hauke Mehrtens [Mon, 25 Mar 2019 14:20:59 +0000 (15:20 +0100)]
omap24xx: Remove unmaintained target

This target only supports kernel 4.1, which is not supported in OpenWrt
any more for multiple releases. It also looks like there is no active
maintainer for this target.
Remove the code and all the packages which are only used by this target.

To add this target to OpenWrt again port it to a recent and supported
kernel version.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoppc44x: Remove unmaintained target
Hauke Mehrtens [Mon, 25 Mar 2019 14:19:15 +0000 (15:19 +0100)]
ppc44x: Remove unmaintained target

This target only supports kernel 3.18, which is not supported in OpenWrt
any more for multiple releases. It also looks like there is no active
maintainer for this target.
Remove the code and all the packages which are only used by this target.

To add this target to OpenWrt again port it to a recent and supported
kernel version.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoppc40x: Remove unmaintained target
Hauke Mehrtens [Mon, 25 Mar 2019 14:16:59 +0000 (15:16 +0100)]
ppc40x: Remove unmaintained target

This target only supports kernel 3.18, which is not supported in OpenWrt
any more for multiple releases. It also looks like there is no active
maintainer for this target.
Remove the code and all the packages which are only used by this target.

To add this target to OpenWrt again port it to a recent and supported
kernel version.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agomcs814x: Remove unmaintained target
Hauke Mehrtens [Mon, 25 Mar 2019 14:18:26 +0000 (15:18 +0100)]
mcs814x: Remove unmaintained target

This target only supports kernel 3.18, which is not supported in OpenWrt
any more for multiple releases. It also looks like there is no active
maintainer for this target.
Remove the code and all the packages which are only used by this target.

To add this target to OpenWrt again port it to a recent and supported
kernel version.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoau1000: Remove unmaintained target
Hauke Mehrtens [Mon, 25 Mar 2019 14:15:01 +0000 (15:15 +0100)]
au1000: Remove unmaintained target

This target only supports kernel 3.18, which is not supported in OpenWrt
any more for multiple releases. It also looks like there is no active
maintainer for this target.
Remove the code and all the packages which are only used by this target.

To add this target to OpenWrt again port it to a recent and supported
kernel version.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoadm8668: Remove unmaintained target
Hauke Mehrtens [Mon, 25 Mar 2019 14:17:43 +0000 (15:17 +0100)]
adm8668: Remove unmaintained target

This target only supports kernel 3.18, which is not supported in OpenWrt
any more for multiple releases. It also looks like there is no active
maintainer for this target.
Remove the code and all the packages which are only used by this target.

To add this target to OpenWrt again port it to a recent and supported
kernel version.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agoadm5120: Remove unmaintained target
Hauke Mehrtens [Mon, 25 Mar 2019 14:10:39 +0000 (15:10 +0100)]
adm5120: Remove unmaintained target

This target only supports kernel 3.18, which is not supported in OpenWrt
any more for multiple releases. It also looks like there is no active
maintainer for this target.
Remove the code and all the packages which are only used by this target.

To add this target to OpenWrt again port it to a recent and supported
kernel version.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
5 years agobrcm2708: fix renamed definition (FS#2265)
Koen Vandeputte [Fri, 3 May 2019 14:13:19 +0000 (16:13 +0200)]
brcm2708: fix renamed definition (FS#2265)

This target adds an alternative mmc driver using a patch.
Within this patch, a definition got renamed upstream.

Change it to fix following compile error:

  CC      drivers/mmc/host/bcm2835-mmc.o
In file included from ./include/linux/kernel.h:14:0,
                 from ./include/linux/delay.h:22,
                 from drivers/mmc/host/bcm2835-mmc.c:25:
drivers/mmc/host/bcm2835-mmc.c: In function 'bcm2835_mmc_dumpregs':
drivers/mmc/host/bcm2835-mmc.c:254:27: error: 'SDHCI_ACMD12_ERR' undeclared (first use in this function); did you mean 'SDHCI_CMD_CRC'?
   bcm2835_mmc_readw(host, SDHCI_ACMD12_ERR),
                           ^
./include/linux/printk.h:137:18: note: in definition of macro 'no_printk'
    printk(fmt, ##__VA_ARGS__); \
                  ^~~~~~~~~~~
drivers/mmc/host/bcm2835-mmc.c:253:2: note: in expansion of macro 'pr_debug'
  pr_debug(DRIVER_NAME ": AC12 err: 0x%08x | Slot int: 0x%08x\n",
  ^~~~~~~~
drivers/mmc/host/bcm2835-mmc.c:254:27: note: each undeclared identifier is reported only once for each function it appears in
   bcm2835_mmc_readw(host, SDHCI_ACMD12_ERR),
                           ^
./include/linux/printk.h:137:18: note: in definition of macro 'no_printk'
    printk(fmt, ##__VA_ARGS__); \
                  ^~~~~~~~~~~
drivers/mmc/host/bcm2835-mmc.c:253:2: note: in expansion of macro 'pr_debug'
  pr_debug(DRIVER_NAME ": AC12 err: 0x%08x | Slot int: 0x%08x\n",
  ^~~~~~~~
scripts/Makefile.build:326: recipe for target 'drivers/mmc/host/bcm2835-mmc.o' failed
make[8]: *** [drivers/mmc/host/bcm2835-mmc.o] Error 1

Upstream commit:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v4.14.114&id=45fd8679ea86bffb352132a1df4917c3d11375aa

Fixes: b765f4be407c ("kernel: bump 4.14 to 4.14.114")
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agoprocd: add procd_running() helper for checking running state
Rafał Miłecki [Wed, 1 May 2019 05:23:21 +0000 (07:23 +0200)]
procd: add procd_running() helper for checking running state

This should be helpful for implementing service_running() in procd init
scripts.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Acked-by: John Crispin <john@phrozen.org>
5 years agoprocd: update to latest git HEAD
Hans Dedecker [Thu, 2 May 2019 15:28:42 +0000 (17:28 +0200)]
procd: update to latest git HEAD

01f3dc8 instance: dump user and group as well

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agoprocd: allow passing optional group instance parameter
Michael Heimpold [Thu, 11 Apr 2019 19:01:09 +0000 (21:01 +0200)]
procd: allow passing optional group instance parameter

Sometimes is desirable to run a process with a specific group id
instead of the default one which is derived from passwd entry.
This can be achived now by using procd_set_param group $mygroup.

Signed-off-by: Michael Heimpold <mhei@heimpold.de>
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com> [PKG_SOURCE_VERSION update]
5 years agoprocd: allow passing optional syslog facility as instance parameter
Michael Heimpold [Thu, 11 Apr 2019 19:01:08 +0000 (21:01 +0200)]
procd: allow passing optional syslog facility as instance parameter

Optional syslog facility can be set by adding procd_set_param facility
$myfacility.
While at, also add stdout/stderr documentation.

Signed-off-by: Michael Heimpold <mhei@heimpold.de>
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com> [PKG_SOURCE_VERSION update]
5 years agokernel: bump 4.19 to 4.19.37
Koen Vandeputte [Mon, 29 Apr 2019 11:02:18 +0000 (13:02 +0200)]
kernel: bump 4.19 to 4.19.37

Refreshed all patches.

Altered patches:
- 150-bridge_allow_receiption_on_disabled_port.patch
- 201-extra_optimization.patch

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

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.14 to 4.14.114
Koen Vandeputte [Mon, 29 Apr 2019 11:01:53 +0000 (13:01 +0200)]
kernel: bump 4.14 to 4.14.114

Refreshed all patches.

Altered patches:
- 150-bridge_allow_receiption_on_disabled_port.patch
- 201-extra_optimization.patch

Remove upstreamed:
- 022-0006-crypto-crypto4xx-properly-set-IV-after-de-and-encryp.patch

Compile-tested on: ar71xx, cns3xxx, imx6, x86_64
Runtime-tested on: ar71xx, cns3xxx, imx6

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.9 to 4.9.171
Koen Vandeputte [Mon, 29 Apr 2019 10:05:49 +0000 (12:05 +0200)]
kernel: bump 4.9 to 4.9.171

Refreshed all patches.

Altered patches:
- 150-bridge_allow_receiption_on_disabled_port.patch
- 201-extra_optimization.patch

Compile-tested on: ar7
Runtime-tested on: none

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: iio: Fix BMP280 Auto probing
Robert Marko [Tue, 30 Apr 2019 15:58:42 +0000 (17:58 +0200)]
kernel: iio: Fix BMP280 Auto probing

Currently Auto probing for BMP/BME280 does not work because kernel
module name in the call is not correct.
Package name was used instead of kernel module name.

Signed-off-by: Robert Marko <robimarko@gmail.com>
5 years agolibroxml: bump to the 3.0.1 version
Rafał Miłecki [Tue, 23 Apr 2019 07:39:14 +0000 (09:39 +0200)]
libroxml: bump to the 3.0.1 version

Some of changes:
* Support for local-name()
* General refactoring
* Better parsing performance
* Fix possible buffer overflow & memleak
* Validation checks
* More commit functions (file, buffer, fd)

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
5 years agoprocd: update to latest git HEAD
Hans Dedecker [Mon, 29 Apr 2019 19:32:12 +0000 (21:32 +0200)]
procd: update to latest git HEAD

cfaed56 procd: add SIGPWR as signal
a30a8fd procd: copy the respawn property of new instance

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agomac80211: rt2x00: replace patches with upstream version
Daniel Golle [Mon, 29 Apr 2019 16:21:34 +0000 (18:21 +0200)]
mac80211: rt2x00: replace patches with upstream version

Support for RT3883/RT3663 was merged upstream [1]. Use that patch
instead of our original series. The resulting source tree is
exactly identical, this commit is merely reorganizing the patches.

[1]: https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next.git/commit/?id=d0e61a0f7cca51ce340a5a73595189972122ff25

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
5 years agoglibc: update to latest 2.27 commit (BZ #18035)
Hans Dedecker [Mon, 29 Apr 2019 12:56:04 +0000 (14:56 +0200)]
glibc: update to latest 2.27 commit (BZ #18035)

1961e5c7296 elf: Fix pldd (BZ#18035)

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agokernel: add missing symbol in 4.19 config
Koen Vandeputte [Mon, 29 Apr 2019 13:28:03 +0000 (15:28 +0200)]
kernel: add missing symbol in 4.19 config

Intel AtomISP2 dummy / power-management driver (INTEL_ATOMISP2_PM) [N/m/y/?] (NEW)

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agotools/ccache: update to 3.7
Deng Qingfang [Mon, 29 Apr 2019 05:26:36 +0000 (13:26 +0800)]
tools/ccache: update to 3.7

Update ccache to 3.7

Release notes:
https://ccache.dev/releasenotes.html#_ccache_3_7

Signed-off-by: Deng Qingfang <dengqf6@mail2.sysu.edu.cn>
5 years agokernel: fix regression on 4.19 with 613-netfilter_optional_tcp_window_check.patch...
Felix Fietkau [Sat, 27 Apr 2019 16:56:30 +0000 (18:56 +0200)]
kernel: fix regression on 4.19 with 613-netfilter_optional_tcp_window_check.patch (FS#2253)

Since ct->proto.tcp.last_win isn't updated when nf_ct_tcp_no_window_check is
enabled, the retransmission timeout check needs to be bypassed.

Based on patch by Rob Mosher

Signed-off-by: Felix Fietkau <nbd@nbd.name>
5 years agoopenssl: build kmods only if engines are selected
Eneas U de Queiroz [Thu, 25 Apr 2019 19:23:40 +0000 (19:23 +0000)]
openssl: build kmods only if engines are selected

Add a conditional to the individual package's for the kmods in DEPENDS.
This avoids the need to compile the kernel modules when the crypto
engine packages are not selected.  The final binares are not affected by
this.

Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Tested-by: Rosen Penev <rosenp@gmail.com>
5 years agoscripts/qemustart: add notes on adding new options
Yousong Zhou [Fri, 26 Apr 2019 04:17:16 +0000 (04:17 +0000)]
scripts/qemustart: add notes on adding new options

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
5 years agoscripts/qemustart: fix network device emulation
Yousong Zhou [Thu, 25 Apr 2019 14:05:24 +0000 (14:05 +0000)]
scripts/qemustart: fix network device emulation

We only enabled pcnet driver for malta machine, and x86/legacy target
was supposed to only support e1000 model

Cc: Petr Štetiar <ynezz@true.cz>
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
5 years agoscripts/qemustart: no network by default and enable with -n
Yousong Zhou [Thu, 25 Apr 2019 16:07:18 +0000 (16:07 +0000)]
scripts/qemustart: no network by default and enable with -n

This should make it easier to set up and possibly provide a "work out of
the box" experience for most test usage.  Typical wan&lan networking
setup is only two characters argument away.

Cc: Petr Štetiar <ynezz@true.cz>
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
Acked-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agosdk: fix building external modules when CONFIG_STACK_VALIDATION=y
Yousong Zhou [Thu, 25 Apr 2019 05:26:23 +0000 (05:26 +0000)]
sdk: fix building external modules when CONFIG_STACK_VALIDATION=y

The option is at the moment enabled for x86/64.  The selection chain
should be

config RETPOLINE
  select STACK_VALIDATION if HAVE_STACK_VALIDATION
config x86
  select HAVE_STACK_VALIDATION            if X86_64

cryptdev-linux:

make[4]: Entering directory '/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/cryptodev-linux-cryptodev-linux-1.10'
make -C /store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/linux-4.14.113 M=/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/cryptodev-linux-cryptodev-linux-1.10 ARCH=x86 CROSS_COMPILE=x86_64-openwrt-linux-musl- modules
make[5]: Entering directory '/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/linux-4.14.113'
make[6]: *** No rule to make target '/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/cryptodev-linux-cryptodev-linux-1.10/ioctl.o', needed by '/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/cryptodev-linux-cryptodev-linux-1.10/cryptodev.o'.  Stop.
Makefile:1533: recipe for target '_module_/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/cryptodev-linux-cryptodev-linux-1.10' failed

Open vSwitch intree kmods

make[7]: Entering directory '/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/openvswitch-2.11.0/datapath/linux'
make -C /store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/linux-4.14.111 ARCH=x86 M=/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/openvswitch-2.11.0/datapath/linux modules
make[8]: Entering directory '/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/linux-4.14.111'
make[9]: *** No rule to make target 'tools/objtool/objtool', needed by '/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/openvswitch-2.11.0/datapath/linux/actions.o'.  Stop.
Makefile:1533: recipe for target '_module_/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/openvswitch-2.11.0/datapath/linux' failed
make[8]: *** [_module_/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/openvswitch-2.11.0/datapath/linux] Error 2
make[8]: Leaving directory '/store/buildbot/slave/x86_64/build/sdk/build_dir/target-x86_64_musl/linux-x86_64/linux-4.14.111'

Reference: builddeb: include objtool binary in headers package, https://patchwork.kernel.org/patch/9014071/
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
5 years agoelfutils: bump to 0.176
Jose Olivera [Wed, 24 Apr 2019 04:17:29 +0000 (12:17 +0800)]
elfutils: bump to 0.176

*Fixes:
  -CVE-2019-7150
  -CVE-2019-7149
  -CVE-2019-7146
  -CVE-2019-7665
  -CVE-2019-7664
  -CVE-2019-7148

*Refresh 003-libintl-compatibility.patch

*Also reset PKG_RELEASE.

Signed-off-by: Jose Olivera <oliverajeo@gmail.com>
5 years agoscripts/qemustart: fix usage with networking support
Hans Dedecker [Wed, 24 Apr 2019 17:54:44 +0000 (19:54 +0200)]
scripts/qemustart: fix usage with networking support

Commit 8bf500eb79 added support for qemu usage without networking
support but broke networking support as -n and -z do not work with
unquoted argument; fix this by quoting the arguments.

Fixes 8bf500eb79

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agomac80211: fix regression in skb resizing optimization in monitor mode (FS#2254)
Felix Fietkau [Wed, 24 Apr 2019 07:32:22 +0000 (09:32 +0200)]
mac80211: fix regression in skb resizing optimization in monitor mode (FS#2254)

struct ieee80211_local needs to be passed in separately instead of
dereferencing the (potentially NULL) sdata

Signed-off-by: Felix Fietkau <nbd@nbd.name>
5 years agotoolchain/musl: update to version 1.1.22
Koen Vandeputte [Fri, 12 Apr 2019 14:25:49 +0000 (16:25 +0200)]
toolchain/musl: update to version 1.1.22

new features:
- priority-inheritance mutexes
- membarrier syscall, pre-registration to use it, fallback emulation
- header-level support for new linux features in 4.19, 4.20, 5.0

major internal changes:
- complete, async-safe view of all existent threads as global list
- robust __synccall based on new thread list
- new dynamic TLS is installed synchronously at dlopen
- TLSDESC resolver functions no longer make bad ABI assumptions to call C
- resolved shared library dependencies are now recorded

compatibility & conformance:
- dependency-order shared library constructor execution
- sigaltstack no longer rejects SS_AUTODISARM, future flags
- FILE is now a complete (dummy) type in pre-C11 feature profiles
- setvbuf reports failure on invalid arguments
- TSVTX is exposed unconditionally in tar.h
- multithreaded set*id() no longer depends on /proc
- key slot reuse after pthread_key_delete no longer depends on /proc

bugs fixed:
- failures in multithreaded set*id() with concurrent thread creation/exit
- interposed free was called from invalid/inconsistent contexts
- freeaddrinfo performed invalid free of some partial results lists
- dlsym dependency order search had false negatives and false positives
- dn_skipname gave wrong results for labels with 8-bit content
- dcngettext clobbered errno, often breaking printing of error messages
- sscanf read past end of buffer under certain conditions (1.1.21 regression)
- pthread_key_create spuriously failed under race condition (1.1.21 regression)
- fdopendir wrongly succeeded with O_PATH file descriptors
- gets behaved incorrectly in presence of null bytes
- namespace violations in c11 tsd and mutex function dependencies
- incorrect prototype for makecontext (unimplemented)

arch-specfic bugs fixed:
- s390x had wrong values for POSIX_FADV_DONTNEED/_NOREUSE

Extensively tested on dozens of devices, covering most popular architectures.

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Signed-off-by: Dainis Jonitis <dainis.jonitis@ubnt.com>
Signed-off-by: Roman Yeryomin <roman@advem.lv>
5 years agoar71xx: add support for Mikrotik RouterBOARD 2011UiAS-2HnD r2
Koen Vandeputte [Fri, 12 Apr 2019 09:08:43 +0000 (11:08 +0200)]
ar71xx: add support for Mikrotik RouterBOARD 2011UiAS-2HnD r2

This board got a new hardware revision, without any change in functionality.

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agoath10k-ct: Update to 2019-04-08
Koen Vandeputte [Tue, 16 Apr 2019 08:54:52 +0000 (10:54 +0200)]
ath10k-ct: Update to 2019-04-08

9cd701a4f028 ath10k-ct:  Add PN get/set API for wave-2 firmware.
5c8a4668323b ath10k-ct:  Support over-riding the power ctl table in eeprom
75e2705f31bb ath10k-ct:  CCA, eeprom, other changes.
a696e602a0fc ath10k-ct:  Attempt to fix-out-of-tree compile for 4.16
a2aec62262df ath10k:  Improve beacon tx status for 4.20 kernel.
be5c21a82b15 ath10k-ct:  Fix out-of-tree compile for 4.20, pull in stable changes for 4.19

Fixes compile errors when using the 4.20 flavour.
Also the amount of beacon errors seems to have dropped.

Tested on a Mikrotik RB912UAGS-5HPacD

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.19 to 4.19.36
Koen Vandeputte [Mon, 22 Apr 2019 19:52:42 +0000 (21:52 +0200)]
kernel: bump 4.19 to 4.19.36

Refreshed all patches.

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

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.14 to 4.14.113
Koen Vandeputte [Mon, 22 Apr 2019 19:51:43 +0000 (21:51 +0200)]
kernel: bump 4.14 to 4.14.113

Refreshed all patches.

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

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.9 to 4.9.170
Koen Vandeputte [Mon, 22 Apr 2019 19:18:27 +0000 (21:18 +0200)]
kernel: bump 4.9 to 4.9.170

Refreshed all patches.

Compile-tested on: ar7
Runtime-tested on: none

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agoRevert "uhttpd: disable concurrent requests by default"
Jo-Philipp Wich [Tue, 23 Apr 2019 06:12:49 +0000 (08:12 +0200)]
Revert "uhttpd: disable concurrent requests by default"

This reverts commit c6aa9ff38870a30dbe6da17e4edad6039fe10ddf.

Further testing has revealed that we will need to allow concurrent
requests after all, especially for situations where CGI processes
initiate further HTTP requests to the local host.

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
5 years agoopenssl: add Eneas U de Queiroz as maintainer
Eneas U de Queiroz [Mon, 22 Apr 2019 19:28:12 +0000 (19:28 +0000)]
openssl: add Eneas U de Queiroz as maintainer

Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
5 years agokernel: bump 4.19 to 4.19.35
Koen Vandeputte [Thu, 18 Apr 2019 13:05:04 +0000 (15:05 +0200)]
kernel: bump 4.19 to 4.19.35

Refreshed all patches.

Fixes:
- CVE-2019-3887

New symbol:
- CONFIG_LDISC_AUTOLOAD

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

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.14 to 4.14.112
Koen Vandeputte [Thu, 18 Apr 2019 13:04:42 +0000 (15:04 +0200)]
kernel: bump 4.14 to 4.14.112

Refreshed all patches.

New symbol:
- CONFIG_LDISC_AUTOLOAD

Compile-tested on: ar71xx, cns3xxx, imx6, x86_64
Runtime-tested on: ar71xx, cns3xxx, imx6

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agokernel: bump 4.9 to 4.9.169
Koen Vandeputte [Thu, 18 Apr 2019 08:04:35 +0000 (10:04 +0200)]
kernel: bump 4.9 to 4.9.169

Refreshed all patches.

New symbols:
- CONFIG_LDISC_AUTOLOAD
- CONFIG_PPC_BARRIER_NOSPEC

Compile-tested on: ar7
Runtime-tested on: none

Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
5 years agoopenssl: fix OPENSSL_config bug affecting wget
Eneas U de Queiroz [Mon, 22 Apr 2019 16:39:10 +0000 (16:39 +0000)]
openssl: fix OPENSSL_config bug affecting wget

This applies an upstream patch that fixes a OPENSSL_config() bug that
causes SSL initialization to fail when the openssl.cnf file is not
found.  The config file is not installed by default.

Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
5 years agonghttp2: bump to 1.38.0
Hans Dedecker [Mon, 22 Apr 2019 11:39:35 +0000 (13:39 +0200)]
nghttp2: bump to 1.38.0

4a9d2005 Update manual pages
acf6a922 Bump up version number to 1.38.0, LT revision to 31:3:17
4ff45821 Update AUTHORS
42dce01e Merge branch 'nghttpx-fix-backend-selection-on-retry'
a35059e3 nghttpx: Fix bug that altered authority and path affect backend selection
5a30fafd Merge branch 'nghttpx-fix-chunked-request-stall'
dce91ad3 Merge branch 'nghttpx-dont-log-authorization'
2cff8b43 nghttpx: Fix bug that chunked request stalls
be96654d nghttpx: Don't log authorization request header field value with -LINFO
ce962c3f Merge branch 'update-http-parser'
f931504e Update http-parser to v2.9.1
d978f351 Fix bug that on_header callback is still called after stream is closed
ec519f22 Merge pull request #1270 from baitisj/master
e8b213e3 Bump up version number to 1.38.0-DEV

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agotoolchain/binutils: refresh patches
Kevin Darbyshire-Bryant [Sun, 21 Apr 2019 13:57:44 +0000 (14:57 +0100)]
toolchain/binutils: refresh patches

Fix hunk offset warnings.

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
5 years agoodhcpd: update to latest git HEAD (FS#2243, FS#2244)
Hans Dedecker [Fri, 19 Apr 2019 17:21:27 +0000 (19:21 +0200)]
odhcpd: update to latest git HEAD (FS#2243, FS#2244)

6633efe router: fix dns search list option

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agodropbear: allow build without dbclient
Rosy Song [Thu, 18 Apr 2019 02:20:19 +0000 (10:20 +0800)]
dropbear: allow build without dbclient

This can save ~16KBytes size for the ipk

Signed-off-by: Rosy Song <rosysong@rosinson.com>
5 years agoarmvirt64: enable qemu drm drivers
Lucian Cristian [Thu, 18 Apr 2019 04:26:23 +0000 (07:26 +0300)]
armvirt64: enable qemu drm drivers

adds the qemu video drivers and also fixes the build failure
introduced with drm-kms-helper

Signed-off-by: Lucian Cristian <lucian.cristian@gmail.com>
5 years agokernel: add missing drm symbol
Lucian Cristian [Thu, 18 Apr 2019 04:41:10 +0000 (07:41 +0300)]
kernel: add missing drm symbol

dependency introduced with module drm-kms-helper

Signed-off-by: Lucian Cristian <lucian.cristian@gmail.com>
5 years agoinclude: unpack.mk: remove CRLF_WORKAROUND
Alexandru Ardelean [Wed, 17 Apr 2019 06:11:14 +0000 (09:11 +0300)]
include: unpack.mk: remove CRLF_WORKAROUND

This patch doesn't seem to have any more users.
The only one used to be target in
`target/linux/etrax/image/e100boot/Makefile`.

That target has since been removed via commit
1080f68b2b422f1e72e402765a45d0cc3c550825 .

There doesn't seem to be any reason left for this workaround.

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
5 years agomac80211: brcm: backport brcmfmac 5.2 patches
Rafał Miłecki [Thu, 18 Apr 2019 08:13:13 +0000 (10:13 +0200)]
mac80211: brcm: backport brcmfmac 5.2 patches

This includes some USB fixes and early work on FullMAC firmware crash
recovery.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
5 years agolantiq: tdw90x0: Fix WLAN LED on TP-Link W8970 v1.2 (FS#2232)
Petr Štetiar [Wed, 10 Apr 2019 08:10:33 +0000 (10:10 +0200)]
lantiq: tdw90x0: Fix WLAN LED on TP-Link W8970 v1.2 (FS#2232)

This patch fixes disfunctional WLAN LED on TP-Link W8970. The LED was
reported working in the CC release[1], but doesn't work anymore in 18.06.2.

1. https://github.com/openwrt/chaos_calmer/commit/420cb24d41fcd650795f09d2b26e0e1d2cf21165

Tested-by: Damian Janarek <dzanar18@o2.pl>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
5 years agoodhcpd: update to latest git HEAD (FS#2206)
Hans Dedecker [Wed, 17 Apr 2019 12:39:57 +0000 (14:39 +0200)]
odhcpd: update to latest git HEAD (FS#2206)

38bc630 router: use ra_lifetime as lifetime for RA options (FS#2206)
0523bdd router: improve code readibility
0a3b279 Revert "router:"
207f8e0 treewide: align syslog loglevels
f1d7da9 router:
0e048ac treewide: fix compiler warnings
83698f6 CMakeList.txt: enable extra compiler checks

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
5 years agoopenssl: change defaults: ENGINE:on, NPN:off, misc
Eneas U de Queiroz [Tue, 16 Apr 2019 20:12:47 +0000 (20:12 +0000)]
openssl: change defaults: ENGINE:on, NPN:off, misc

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.
Enable engine support by default.  Right now, some packages require
this, so it is always enabled by the bots.  Many packages will compile
differently when engine support is detected, needing engine symbols from
the libraries.

However, being off by default, a user compiling its own image will fail
to run some popular packages from the official repo.
Note that disabling engines did not work in 1.0.2, so this problem never
showed up before.

NPN support has been removed in major browsers & servers, and has become
a small bloat, so it does not make sense to leave it on by default.

Remove deprecated CONFIG_ENGINE_CRYPTO symbol that is no longer needed.

Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
5 years agokernel: Fix kmod-drm-amdgpu and kmod-drm-radeon dependencies
Lucian Cristian [Sat, 13 Apr 2019 17:26:07 +0000 (20:26 +0300)]
kernel: Fix kmod-drm-amdgpu and kmod-drm-radeon dependencies

Currently the Geode builds fails on following kernel module missing
dependencies:

 Package kmod-drm-amdgpu is missing dependencies for the following libraries:
 backlight.ko
 drm_kms_helper.ko
 fb.ko
 ttm.ko

So this patch tries to fix the kmod-drm-amdgpu module dependecies.

Fixes: 2f239c0 ("x86: video: add amdgpu DRM kernel package")
Fixes: 2f6918e ("x86: video: add radeon DRM module support")
Tested-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
Signed-off-by: Lucian Cristian <lucian.cristian@gmail.com>