In the drv_mac80211_setup function, mac80211_interface_cleanup
is called to ask the kernel to delete all existing interfaces
for the phy that is being configured via netlink.
Later in the first function, mac80211_prepare_vif is called to
set up the new interfaces as required.
But sometimes, when mac80211_prepare_vif (and so the relevant
`iw phy x interface add y` command) runs, the kernel might still
be cleaning up the old interface with the same ifname. It usually
takes very few time to do that; possibly a few milliseconds of
sleep in the script after detecting this error condition could be
enough, but the busybox sh does not support sub-second sleep
intervals.
When this happens, iw obviously fails to create the new interface;
and the following message is printed in the system log, followed by
subsequent failure messages from hostapd in case this would have been
an AP interface.
Tue Mar 14 04:21:57 2017 daemon.notice netifd: radio1 (2767): command failed: Too many open files in system (-23)
This was a long-standing issue existing since at least OpenWrt Backfire,
and today I finally managed to debug and (hopefully) solve it.
It was happening very few times on most devices; but it was happening
a lot more frequently on fast platforms with multiple radios, such as
the powerpc-based dual-ath9k-radio tl-wdr4900-v1.
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
On the ath10k-devel ML Michael Kazior stated:
"board-2 is a key-value store of actual board files.
Some devices, notably qca61x4 hw3+ and qca4019 need
distinct board files to be uploaded. Otherwise they
fail in various ways." [0].
Later on Rajkumar Manoharan explained:
"In QCA4019 platform, only radio specific calibration
(pre-cal-data) is stored in flash. Board specific contents
are read from board-2.bin. For each radio appropriate board
data should be loaded. To fetch correct board data from
board-2.bin bundle, pre-cal/radio specific caldata should
be loaded first to get proper board id.
|My understanding until now was that:
|
| * pre-cal data + board-2.bin info == actual calibration data
Correct." [1].
The standard board-2.bin from the ath10k-firmware-qca4019
barely works on the RT-AC58U. Especially 5GHz clients fail
to connect at all and if they do, they have very low
throughput even right next to the router.
Currently, the solution for this problem is to supply a
custom board-2.bin for every device.
To implement this feature, this method makes use of:
Rafał Miłecki's "base-files: add support for overlaying
rootfs content". This comes with a few limitations:
1. Since there can only be one board-2.bin at the right
location, there can only one board overwrite installed
at any time. (All packages CONFLICT with each other.
It's also not possible to "builtin" multiple package.)
2. updating ath10k-firmware-qca4019 will also replace
the board-2.bin. For this cases the user needs to
manually reinstall the wifi-board package once the
ath10k-firmware-qca4019 is updated.
To create the individual board-2.bin: Use the ath10k-bdencoder
utility from the qca-swiss-army-knife repository:
<https://github.com/qca/qca-swiss-army-knife>
The raw board.bin files have to be extracted from the
vendor's source GPL.tar archieves.
Signed-off-by: Alexis Green <agreen@cococorp.com>
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
This patch adds support for AVM FRITZ!Box 4040.
hardware highlights:
SOC: IPQ4018 / QCA Dakota
CPU: Quad-Core ARMv7 Processor rev 5 (v7l) Cortex-A7
DRAM: 256 MiB Nanya NT5CC128M16IP
FLASH: 32 MiB MXIC MX25L25635FMI
ETH: Qualcomm Atheros QCA8075 Gigabit Switch (4 x LAN, 1 x WAN)
USB: 1 x 3.0 (via Synopsys DesignWare DWC3 controller in the SoC)
1 x 2.0 (via Synopsys DesignWare DWC3 controller in the SoC)
WLAN1: Qualcomm Atheros QCA4018 2.4GHz 802.11bgn 2:2x2
WLAN2: Qualcomm Atheros QCA4018 5GHz 802.11a/n/ac 2:2x2
INPUT: one WLAN and one WPS button
LEDS: Power, WAN/Internet, WIFI, INFO (red and amber) and LAN.
Serial:
WARNING: The serial port needs a TTL/RS-232 v3.3 level converter!
The Serial setting is 115200-8-N-1. The SoC's serial port is right
next to the MXIC FLASH chip. The board has a unpopulated 1x4 0.1"
header for it. Use a multimeter to figure out the pinout!
This board currently needs an additional u-boot image in order to boot
properly. Booting with EVA isn't possible ATM.
Install Procedure:
0. It's highly recommended to connect to the serial port.
The serial settings are listed above.
1. install a u-boot image for AVM Fritz!Box 4040
(see <https://github.com/chunkeey/FritzBox-4040-UBOOT/releases> and
<https://github.com/chunkeey/FritzBox-4040-UBOOT/blob/master/upload-to-f4040.sh>)
2. upload the initramfs.itb image via tftp (u-boot listens to
192.168.1.1 - use binary transfer mode!)
3. connect to the FB4040 and use sysupgrade sysupgrade.bin
to install the image.
Works:
- Switch and Ethernet (99%)
- Buttons (WLAN, WPS)
- FLASH (1 x 32MiB NOR Chip)
- WLAN2G and WLAN5G
- CPUFREQ scaling
- PRNG
- serial
- Crypto Accelerator
- sysupgrade (Read the flash instructions to avoid bricking)
- full LEDE Install (Read the flash instructions to avoid bricking)
- LEDs (Power, WAN, Info (red and amber), LAN)
The LEDs are connected to the QCA8075 LED ports.
The AR40xx driver contains a gpio-controller to
handle these special "GPIOs".
- USB Both 3.0 and 2.0 ports
- many packages from other ARMv7 boards
(This does include the RaspberryPi Model 2!)
- ...
Not planned:
- WAN<->LAN short-cut
- Qualcomm Secure Execution Environment
- ...
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John Crispin <john@phrozen.org>
With this patch, the nand.sh script no longer overwrites
CI_KERNELPART and CI_UBIPART if they have been set before.
This allows platforms specify alternative values for these
variables.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This patch enables the ATH10K_AHB support for the QCA4019
devices on the AHB bus.
This patch also removes 936-ath10k_skip_otp_check.patch
because it breaks the AHB device identification.
"Patch is wrong. I find it frustrating OpenWRT/LEDE doesn't
try to work with upstream on ixing these things right."
[1] <https://www.mail-archive.com/ath10k@lists.infradead.org/msg05896.html>
It also limits ath10k memory hunger (This is a problem with 128MiB RAM)
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This change protects the openvpn instances to be marked as "in a crash
loop" and thereby the connection retries will run infinitely.
When the remote site of an openvpn connection goes down for some time
(network failure etc.) the openvpn instance in an openwrt/lede device
should not stop retrying to establish the connection.
With the current limit of 5 retries, there is a user interaction
required, which isn't really what you want when the device should
simply do everything to keep the vpn connection up.
Signed-off-by: Martin Schiller <ms@dev.tdt.de>
Add actual mirror and use main site as last resport
Source: http://www.tcpdump.org/mirrors.html
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Add mirror and use main site as last resort.
Source: http://www.tcpdump.org/mirrors.html
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Update util-linux to 2.29.2
Remote 0001-fix-uClibc-ng-scanf-check.patch as it's been merged upstream.
Refresh patches
Change ncurses to ncursesw to fix compilation and avoid confusion
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Refresh mirror list, some doesn't offer OpenSSL and add main site as last resort.
Source: https://www.openssl.org/source/mirror.html
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
This includes following changes:
480d7bc Fix sending unicast questions on cache expire
a0403cd Keep source sockaddr for every cached DNS record
1478293 Fix code freeing cached non-A(AAA) records too early
9f1cc22 Fix replying to "QU" questions received on unicast interface
943bedb Fix reading port of incoming packets
c725494 Use MCAST_PORT define for port 5353
ce7e9e9 Use one define for DNS-Based Service Discovery service name
e1bacef Drop entries cached for interface we're going to delete
496aeba Fix comment typo in cache_gc_timer
f89986b Fix refreshing cached A(AAA) records that expire
Previous updates made umdns work as expected on startup but there were
still many bugs. They were mostly related to runtime - cache management
and requests + responses. E.g. umdns was never able to send question on
DNS record expire. It was also ignoring all incoming unicast questions.
Since these issues are quite serious it makes sense to backport this
update to the stable branch.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
a032166 interface-ip: set prefix indicator flag when IPv6 prefix lifetime changes
b4f8984 system-linux: parse vti specific settings as nested json data object
7e3b89a system-linux: parse gre specific settings as nested json data object
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
This adds support for describing GPIO chips placed on PCIe cards. Thanks
to this we get working 2.4 GHz LED on Tenda AC9.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This is the standard way we handle this. Please note (it seems) I could
drop few symbols as they are hidden under (disabled) DRM_LEGACY now.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
For targets with i2c not built-in this fixes following error:
Package kmod-drm is missing dependencies for the following libraries:
i2c-core.ko
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This can be used to indicate that a target does not support the optional mips16
extension even when it is a mips32r2 or later CPU.
This will generate a separate toolchain and a separate package folder,
e.g. mips_24kc_nomips16
Signed-off-by: Hauke Mehrtens <hauke.mehrtens@intel.com>
Without this change the code checked if the string was contained in the
feature option and not if the string matches the complete word. This only
removes the nand option from the omap24xx target, the other changes are
only removing options which were added twice.
Signed-off-by: Hauke Mehrtens <hauke.mehrtens@intel.com>
Remove upx from repo, it hasn't been used for a very long time and serves little to no purpose.
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Update to 4.1
Remove 0001-mkfs-Default-to-64-32-heads-sectors-for-targets-smal.patch
as it's included in upstream repo.
Refresh patch
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Add back support for the independent_clocks definition that has been
removed between kernel 4.4 and 4.9 by upstream commits
eb96924acddc709db58221c210ca05cd9effb1df and
e86eee6bc2aaa6b3637f6497b26beee09a91bde9
* extend the new cpufreq_dt_platform_data definition in cpufreq-dt.h
* revert the removal of its usage in cpufreq-dt.c
* use new cpufreq-dt.h in qcom-cpufreq.c
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
For IPQ806x targets, TZ protects the registers that are used to
configure the routing of interrupts to a target processor.
To resolve this, this patch uses scm call to route GPIO interrupts
to application processor. Also the scm call interface is changed.
Signed-off-by: Pavel Kubelun <be.dissent@gmail.com>
At the moment as a workaround definition for scm firmware in DT is used as if it is
apq8064 board. This leads to incomplete scm firmware initialization and as a result
cpuidle driver fails to configure.
By design unlike other qcom boards ipq do not use clocks to connect to scm.
Considering this we're removing from DT and scm driver clocks for ipq boards.
As a result cpuidle does not produce errors about failed configuration anymore.
Signed-off-by: Pavel Kubelun <be.dissent@gmail.com>
Felix Fietkau
Vittorio Gambaletta
John Crispin
Christian Lamparter
Christian Lamparter
Martin Schiller
Daniel Engberg
Rafał Miłecki
Hans Dedecker
Hauke Mehrtens
Paul Wassi
Furong Xu
Vaclav Svoboda
Hannu Nyman
Pavel Kubelun
Yousong Zhou