Jo-Philipp Wich [Tue, 28 Jun 2016 20:43:22 +0000 (22:43 +0200)]
mac80211: fix skb size calculation in 4addr mode (FS#24)
The PDU length of incoming LLC frames is set to the total skb payload size
in __ieee80211_data_to_8023() of net/wireless/util.c which incorrectly
includes the length of the IEEE 802.11 header.
The resulting LLC frame header has a too large PDU length, causing the
llc_fixup_skb() function of net/llc/llc_input.c to reject the incoming
skb, effectively breaking STP.
Solve the problem by properly substracting the IEEE 802.11 frame header size
from the PDU length, allowing the LLC processor to pick up the incoming
control messages.
Special thanks to Gerry Rozema for tracking down the regression and proposing
a suitable patch.
Daniel Dickinson [Wed, 18 May 2016 07:05:42 +0000 (03:05 -0400)]
lldpd: Use /etc/os-release instead of /etc/openwrt_*
With the addition of /etc/os-release patching lldpd to use
/etc/openwrt_release and to have the initscript use
/etc/openwrt_release and/or /etc/openwrt_version becomes
unnecessary.
Signed-off-by: Daniel Dickinson <lede@daniel.thecshore.com>
kernel: remove full cache flush in fuse_copy_do() for MIPS
This patch was introduced in commit r16412 for the brcm47xx target only
and then moved to generic in commit r32395. It was initially added
because of ticket #5186 and should fix some problems with fuse file
systems and MIPS caches. The commit comment in r32395 says that this a
generic problem in MIPS CPUs, but does not name any specifics about
that. There was a fix added to kernel 2.6.21 in commit commit 7575a49f20 "[MIPS] Implement flush_anon_page()." that should fix this
problem, but that was already available before both commits were done
to OpenWrt.
I just tested fuse with ntfs.3g without this patch on a BCM4704
(BMIPS3300 V0.6) SoC and haven't seen any problems. Someone reported
that removing this patch improves some fuse operations by 5 times on
some modern MIPS cores.
My test was only a simple "dd if=/dev/zero of=/mnt/zero bs=5000" to an
USB stick.
This patch removes the patch to OpenWrt, because I assume that it is
not needed any more and Felix, the orginal author, also thinks so.
Alin Năstac [Fri, 24 Jun 2016 12:41:10 +0000 (14:41 +0200)]
libnetfilter_queue: fix checksum computation
There are 2 issues fixed by this patch:
- UDP checksum is computed incorrectly, the used pseudo IP header
contains transport protocol 6 iso 17
- on big endian arches the UDP/TCP checksum is incorrectly
computed when payload length is odd
Signed-off-by: Alin Nastac <alin.nastac@gmail.com> Signed-off-by: Felix Fietkau <nbd@nbd.name> [refresh patch]
Sergey Sergeev [Thu, 16 Jun 2016 19:04:51 +0000 (22:04 +0300)]
build: Adds the ability to disable personal initramfs build for target device
If KERNEL_INITRAMFS := in the target/linux/*/image/Makefile->Device/%NAME% section is set to ''
then personal initramfs file for this target device will not be created.
This var is similar to the Device/Build/kernel KERNEL_INSTALL :=
Felix Fietkau [Tue, 14 Jun 2016 09:27:20 +0000 (11:27 +0200)]
ar71xx: enable flow control for ethernet MACs with built-in switch
Should fix LAN speed issues on some devices. This is an updated version
of the previously reverted commit with the same name.
It improves the check for MACs connected to a built-in switch
Hauke Mehrtens [Sat, 25 Jun 2016 16:55:50 +0000 (18:55 +0200)]
linux-firmware: fix md5sum
The copy on the mirror has a different md5sum as specified in this
package Makefile. The content of the file on the mirror is the same as
in the checkout so just update our md5sum.
Jo-Philipp Wich [Fri, 24 Jun 2016 13:04:27 +0000 (15:04 +0200)]
ebtables: fix segmentation fault due to uninitialized extension data
The ebtables code relies on the `-nostartfiles` linker argument to execute the
extension modules' `_init()` functions automatically which is not working
reliably across all supported targets and gcc versions.
Running an ebtables executable linked this way just crashes with a segmentation
fault at runtime on program startup, e.g. on ARM architectures.
In order to fix the issue ...
- remove the use of the -nostartfiles linker flag
- rename the init procedures to a generic name without implicit semantics
- explicitely annotate those init procedures as constructors
The patch has been taken from the Alpine Linux distribution at
http://git.alpinelinux.org/cgit/aports/tree/main/ebtables/fix-extension-init.patch
base-files: sysfixtime no longer exclude dnsmasq.time
dnsmasq's dnssec time checking method now uses a ntp hotplug mechanism,
therefore dnsmasq.time is redudant and no longer needs to be explicitly
excluded from sysfixtime.
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Change dnsmasq's dnssec time check handling to use time validity
indicated by ntpd rather than maintaining a cross boot/upgrade
/etc/dnsmasq.time timestamp file. This saves flash device wear.
If ntpd client is configured in uci and you're using dnssec, then
dnsmasq will not check dnssec timestamp validity until ntpd hotplug
indicates sync via a stratum change. The ntpd hotplug leaves a status
flag file to indicate to dnsmasq.init that time is valid and that it
should now start in 'check dnssec timestamp valid' mode.
If ntpd client is not configured and you're using dnssec, then it is
presumed you're using an alternate time sync mechanism and that time is
correct, thus dnsmasq checks dnssec timestamps are valid from 1st start.
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
V2 - stratum & step ntp changes indicate time is valid
V3 - on initial flag file step signal dnsmasq with SIGHUP if running
V4 - only accept step ntp changes. Accepting both stratum & step could
result in unpleasant script race conditions
V5 - Actually only accepting stratum is the correct thing to do after
further testing
V6 - improve handling of non busybox ntpd
if sysntpd not executable
dnsmasq checks dnssec timestamps
else
sysntp script disabled - look for timestamp file - allows external mechanism to use hotplug flag file
sysntp script enabled & uci ntp enabled - look for timestamp file
sysntp script enabled & uci ntp disabled - dnsmasq checks dnssec
timestamps
fi
Daniel Dickinson [Wed, 18 May 2016 13:07:30 +0000 (09:07 -0400)]
base-files: Add standard os-release file
/etc/os-release is the standard distribution release information
file, therefore add it (and image configuration options for
fields not previously present in LEDE). Once it is deemed
reasonable the non-standard openwrt_release, openwrt_version,
and device_info files could be removed (that is with this patch
we consider them deprecated in favour of the standard file).
Signed-off-by: Daniel Dickinson <lede@daniel.thecshore.com>
Jo-Philipp Wich [Fri, 24 Jun 2016 11:32:22 +0000 (13:32 +0200)]
ar71xx: fix legacy image building
After the conversion from legacy device profiles to the newer profile
information embedded in the image building code, the legacy recipes got
triggered twice with different eval depths, leading to shell syntax errors
when processing certain images.
The double processing was caused by the remaining Image/Build macro in
legacy.mk which serves as main entry point for the new style image build code
in conjunction with the newly introduced LegacyDevice/* macros which caused
the legacy image build fallback code to kick in.
In order to fix the issue, rework all legacy macros to work under the legacy
image build wrapper and remove the Image/Build macro of legacy.mk to prevent
legacy profiles getting executed in the context of the new build code.
The ZBT APE522II is a dual-radio outdoor CPE based on the MT7620a SoC. It has
64 MB RAM, 8 MB flash, 2 Fast Ethernet ports via internal switch (one with
802.3af 48V PoE support), a 802.11b/g/n SoC 2.4 GHz radio and an 802.11a/n/ac
MT7612E-based 5 GHz radio.
Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
Adrian Panella [Tue, 21 Jun 2016 16:12:26 +0000 (11:12 -0500)]
ipq806x: enable ieee80211 phy hotplug and patch macaddress
Calibration data for QCA99x0 in this device has bogus macaddress.
The data cannot be modified directly, as it breaks checksum control.
Instead change the macaddress from phy add hotplug event.
Signed-off-by: Adrian Panella <ianchi74@outlook.com>
Hauke Mehrtens [Tue, 21 Jun 2016 20:05:44 +0000 (22:05 +0200)]
squashfs4: use upstream xz compression header format
In the upstream kernel and the upstream squashfs4 tools the xz
compression header looks the following:
struct disk_comp_opts {
__le32 dictionary_size;
__le32 flags;
};
We added some other members and also moved some existing members. Place
the members which are already in upstream header at the same position
as in that kernel and add our own at the end. The kernel should not
have a problem when there are some additional members and just ignore
them.
Alexey Brodkin [Wed, 22 Jun 2016 12:47:20 +0000 (15:47 +0300)]
linux: arc: disable kernel unwinding to fix modules loading
With update of binutils for ARC (this is now based on upstream 2.26)
we noticed issues with loadable kernel modules.
Something like that was happening:
--------------------->8-------------------
mbcache: unknown relocation: 49
insmod: can't insert './mbcache.ko': invalid module format
--------------------->8-------------------
More details could be found in that discussion in binutils mailing list:
http://thread.gmane.org/gmane.comp.gnu.binutils/74662
As of now the simplest work-around is to disable in-kernel unwinder
for now. That will at least allow us to use modules again.
Hauke Mehrtens [Tue, 21 Jun 2016 21:52:55 +0000 (23:52 +0200)]
packages: prefer http over git for git protocol
In company networks everything except the http and https protocol is
often causes problems, because the network administrators try to block
everything else. To make it easier to use LEDE in company networks use
the https/http protocol for git access when possible.
lantiq: add support for ARV7506PW11 (Alice/O2 IAD 4421)
Ethernet, ADSL2+ and LEDs are fully functional.
Supporting the two TAE ports and SIP gateway was not attempted.
The WiFi is unreliable, due to experimental support for rt35xx family
devices by the rt2800pci driver.
Signed-off-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
[rebase to LEDE HEAD]
[switch to normal image instead of brnboot image]
[remove not required pinmux child nodes keys, leds, ebu, exin, pci_in and pci_out]
[remove switch_rst pinmux child node (no support for hw reset in driver/setting a default GPIO value in DT]
[enable usage of the wireless LED]
[fixup mac address configuration] Sgned-off-by: Mathias Kresin <dev@kresin.me>
The bootloader uses 30 MHz as the SPI frequency for flash on the Germany and
North America models, and 50 MHz for it on the worldwide model, but the Lantiq
SPI driver in OpenWrt and LEDE may access the flash differently such that
writes are capped at 20 MHz, leading to read errors reported on the worldwide
model at 30 MHz.
Signed-off-by: Daniel Gimpelevich <daniel@gimpelevich.san-francisco.ca.us> Acked-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Hauke Mehrtens [Sat, 18 Jun 2016 22:17:39 +0000 (00:17 +0200)]
lantiq: make EASY80920 work with both chip versions
The EASY80920 is available with the A1X and the A2X chip version
depending on the board version. Add both firmware versions to device
tree and make the driver load the correct version depending on the chip
version.
Hauke Mehrtens [Sat, 11 Jun 2016 14:14:52 +0000 (16:14 +0200)]
lantiq: build fullimage.img for EASY80920NAND
This image format is used by Lantiq's / Intel's UGW version 6.1 to 7.1.
These images can be flashed onto a board with the SoC vendor boot
loader as a replacement for the vendor firmware.
Hauke Mehrtens [Tue, 7 Jun 2016 21:10:45 +0000 (23:10 +0200)]
lantiq: remove patch adding support for building without reset controller
The lantiq kernel arch code selects CONFIG_RESET_CONTROLLER always, so
it is always selected when the lantiq target is build. we do not need
support for unselected CONFIG_RESET_CONTROLLER option.
Added optional command line option for patch-image tool
Default 16KB size is still maintained as this is an optional argument.
if one wants to specify or increase size they can provide this option.
sample usage: patch-dtb <file> <dtb> [dtb max size]
brcm63xx: simplify block size and image offset options
Instead of having CFE_EXTRAS for every device that need specific block sizes
and image offset, let's define a couple of vars which can be customized for
each device.
Some devices require padded images in order to prevent CFE from flashing them
to the wrong offset.
For small flashes (4/8) this is no big deal, but for larger flashes (16/32) this
implies 8 and 16 MB images, which is way too large and some devices aren't
capable of flashing them through sysupgrade.
Daniel Dickinson [Tue, 10 May 2016 22:26:42 +0000 (18:26 -0400)]
sdk: Fix keeping unset as unset
The configuration preservation mechanism fails to keep items was are unset
as unset. For busybox this causes issues when building with custom
settings in the SDK. Therefore preserve busybox unset settings.
In addition we preserve non-package selection unset settings because
they may be important for making sure we compile and identical
package for packages that are recompiled in the SDK.
Now that SDK not longer ships .config this becomes relevant for
any use of the SDK.
Signed-off-by: Daniel Dickinson <lede@daniel.thecshore.com>
Daniel Gonzalez [Mon, 11 Apr 2016 11:26:15 +0000 (13:26 +0200)]
brcm63xx: fix external IRQs
External IRQs are currently broken, fix the BCM6345_EXT_IRQ driver.
Since the adoption of the new driver irq-bcm6345-ext, in Chaos Calmer,
external IRQs don't work. It seems there were some minor errors at the initial
development of the driver.
This patch fixes the ticket https://dev.openwrt.org/ticket/21613
Signed-off-by: Daniel Gonzalez <dgcbueu@gmail.com> Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Michael Lee [Thu, 7 Jan 2016 13:50:24 +0000 (21:50 +0800)]
ramips: update i2c drivers
* add rt_i2c structure to store driver data
* rewrite read/write check function and add i2c error status check.
so we don't need to wait until time out.
* add 10 bits address support. according to the data sheet i think
it is possible. but i haven't verify it.
* the most important is start transfer only need once. otherwise
it cause I2C_STARTERR status.
* add set i2c clock speed register by dts options "clock-frequency".
not just hard code it.
* add mt7621 i2c driver. i just copy i2c-ralink.c and change register
names. and the hardware don't support error status. so i remove it.
but the logic is the same.
Mathias Kresin [Sun, 19 Jun 2016 14:33:44 +0000 (16:33 +0200)]
lantiq: diag - switch off the power led on failsafe
This patch is a follow up to commit 4cf3fd4 "add support for indicating
the boot state using three leds".
At the time of writing the patch, I wasn't aware that it's possible to
switch info failsafe after boot (factory reset).
Enabling the failsafe led without disabling the running led causes an
unexpected led colour on devices using a single multicolour led to
indicate the boot state.
Rafał Miłecki [Tue, 14 Jun 2016 10:55:41 +0000 (12:55 +0200)]
kernel: rename B53 symbols to avoid upstream kernel conflict
In kernel 4.7 there is upstreamed b53 driver using (mostly?) the same
symbols as our b53 does. Change our symbols so both drivers can coexist
in kernel tree.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com> Acked-by: Jonas Gorski <jonas.gorski@gmail.com>
Rafał Miłecki [Mon, 20 Jun 2016 05:58:29 +0000 (07:58 +0200)]
kernel: mtdsplit: calculate kernel partition precisely for Seama
So far "kernel" partition didn't contain just a kernel. It also included
Seama header and meta data. This was making kernel update complex and it
wasn't trivial to read kernel size.
Fix it by making "kernel" parition contain just a kernel image.
Daniel Golle [Sat, 18 Jun 2016 16:34:22 +0000 (18:34 +0200)]
kernel: backport cosmetic ubifs patch
Despite the MS_SILENT flag being set when probing for ubifs rootfs a
logline indicating an error is generated during boot:
UBIFS error (pid: 1): cannot open "ubi0:rootfs", error -19
This leads to confusion and there shouldn't be lines containing
the word 'error' twice in a bootlog if actually everything is fine
(just the rootfs happens to be something else than ubifs)
The patch added has been submitted and was accepted upstream, see:
http://lists.infradead.org/pipermail/linux-mtd/2016-June/068056.html
http://patchwork.ozlabs.org/patch/637491
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Jo-Philipp Wich [Fri, 17 Jun 2016 15:39:54 +0000 (17:39 +0200)]
octeon: fix image build
The generic sysupgrade image attempted to use the wrong filesystem type due
to premature variable interpolation leading to the following error on the
buildbot system:
cp: cannot stat `.../root.squashfs': No such file or directory
make[4]: *** [.../tmp/lede-octeon-generic-ext4-sysupgrade.tar] Error 1
Reordered the VLANs so the LAN ports are set to VLAN 1 and the WAN port is set to VLAN 2, as in the other routers in the config file. Moreover, this model had this VLAN mapping in OpenWRT Chaos Calmer. It seems that the VLAN were switched when fixing a bug in the port mapping ( OpenWRT changeset 47799 )
Signed-off-by: David Pinilla Caparrós <dpinitux@gmail.com>
Adrian Panella [Sat, 11 Jun 2016 00:10:15 +0000 (19:10 -0500)]
ipq806x: activate ATAG DTB mangle and EA8500 rootblock in dts
The command-line arguments provided by the boot loader will be
appended to a new device tree property: bootloader-args.
If there is a property "append-rootblock" in DT under /chosen
and a root= option in bootloaders command line it will be parsed
and added to DT bootargs with the form: <append-rootblock>XX.
Only command line ATAG will be processed, the rest of the ATAGs
sent by bootloader will be ignored.
This is usefull in dual boot systems, to get the current root partition
without afecting the rest of the system.
Signed-off-by: Adrian Panella <ianchi74@outlook.com>
Koen Vandeputte [Thu, 16 Jun 2016 11:48:39 +0000 (13:48 +0200)]
cns3xxx: fix RX softIRQ loop
Already reschedule when 1 or more frames came in.
Checking for a full queue could produce a re-schedule loop as
the checked RX ring location could contain undefined values
depending on activity in previous loops.
projects / thirdparty / openwrt.git / log
