It was noticed that the the whole MAC can hang when transferring data from
one ar40xx port (WAN ports) to the CPU and from the CPU back to another
ar40xx port (LAN ports). The CPU was doing only NATing in that process.
Usually, the problem first starts with a simple data corruption:
$ wget https://cdimage.debian.org/debian-cd/current/amd64/iso-cd/debian-10.4.0-amd64-netinst.iso -O /dev/null
...
Connecting to saimei.ftp.acc.umu.se (saimei.ftp.acc.umu.se)|2001:6b0:19::138|:443... connected.
...
Read error at byte 48807936/352321536 (Decryption has failed.). Retrying.
But after a short while, the whole MAC will stop to react. No traffic can
be transported anymore from the CPU port from/to the AR40xx PHY/switch and
the MAC has to be resetted.
The whole problem can be avoided by disabling the TSO for IPv6 for this
ethernet MAC driver.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Acked-by: John Crispin <john@phrozen.org>
(backported from commit 6785695056,
with updated commit message)
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Fixes the following build error with GCC 10:
/usr/bin/ld: read_fs.o:(.bss+0x0): multiple definition of `swap'; mksquashfs.o:(.bss+0x1b2a88): first defined here
And a compile warning.
Fixes: FS#3104, FS#3119
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 1bbc1aa884)
There is a restriction in the number of parameters(10) that may be passed to
the SetupHostCommand macro so continually adding explicit gcc'n' version
checks ends up breaking the compiler check for the later versions and
oddballs like Darwin as was done in 835d1c68a0 which added gcc10.
Drop all the explicitly specified gcc version checks. If a suitable gcc
compiler is not found, it may be specified at the dependency checking
stage after which that version will be symlinked into the build staging
host directory.
eg. 'CC=gccfoo CXX=g++foo make prereq'
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Acked-by: Jo-Philipp Wich <jo@mein.io>
(cherry picked from commit 1fb3c003d6)
Lets add GCC 10 detection to the build system as distributions like Fedora 32 have started shipping with it.
Some tools like mtd-utils need work to compile under GCC10, but that will be next step.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
(cherry picked from commit 835d1c68a0)
The property "ralink,port-map" has been obsolete long before
this device was added, and the device is a one-port anyway.
Just remove it.
Fixes: 5ef79af4f8 ("ramips: add support for Ravpower WD03")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
(cherry picked from commit c00b2df6c8)
In file included from ./arch/mips/include/asm/io.h:34,
from ./arch/mips/include/asm/mmiowb.h:5,
from ./include/linux/spinlock.h:60,
from ./include/linux/irq.h:14,
from drivers/irqchip/irq-bcm6345-ext.c:10:
drivers/irqchip/irq-bcm6345-ext.c: In function 'bcm6345_ext_intc_of_init':
./arch/mips/include/asm/mach-bcm63xx/ioremap.h:48:9: warning: 'base' may be used uninitialized in this function [-Wmaybe-uninitialized]
return is_bcm63xx_internal_registers((unsigned long)addr);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/irqchip/irq-bcm6345-ext.c:255:16: note: 'base' was declared here
void __iomem *base;
^~~~
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 29c3bb5f41)
drivers/irqchip/irq-bcm6345-periph.c: In function 'bcm6345_periph_irq_handle':
drivers/irqchip/irq-bcm6345-periph.c:55:21: warning: 'block' may be used uninitialized in this function [-Wmaybe-uninitialized]
struct intc_block *block;
^~~~~
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit f2f2cf07a6)
Security fixes for:
* CVE-2020-10932
* a potentially remotely exploitable buffer overread in a DTLS client
* bug in DTLS handling of new associations with the same parameters
Full release announement:
https://tls.mbed.org/tech-updates/releases/mbedtls-2.16.6-and-2.7.15-released
Signed-off-by: Magnus Kroken <mkroken@gmail.com>
(cherry picked from commit 02fcbe2f3d)
This backports upstream fixes for the out of bounds write vulnerability in json-c.
It was reported and patches in this upstream PR: https://github.com/json-c/json-c/pull/592
Addresses CVE-2020-12762
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Signed-off-by: Luka Perkov <luka.perkov@sartura.hr>
[bump PKG_RELEASE, rebase patches on top of json-c 0.12]
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
(backported from commit bc0288b768)
* compat: timeconst.h is a generated artifact
Before we were trying to check for timeconst.h by looking in the kernel
source directory. This isn't quite correct on configurations in which
the object directory is separate from the kernel source directory, for
example when using O="elsewhere" as a make option when building the
kernel. The correct fix is to use $(CURDIR), which should point to
where we want.
* compat: use bash instead of bc for HZ-->USEC calculation
This should make packaging somewhat easier, as bash is generally already
available (at least for dkms), whereas bc isn't provided by distros by
default in their build meta packages.
* socket: remove errant restriction on looping to self
It's already possible to create two different interfaces and loop
packets between them. This has always been possible with tunnels in the
kernel, and isn't specific to wireguard. Therefore, the networking stack
already needs to deal with that. At the very least, the packet winds up
exceeding the MTU and is discarded at that point. So, since this is
already something that happens, there's no need to forbid the not very
exceptional case of routing a packet back to the same interface; this
loop is no different than others, and we shouldn't special case it, but
rather rely on generic handling of loops in general. This also makes it
easier to do interesting things with wireguard such as onion routing.
At the same time, we add a selftest for this, ensuring that both onion
routing works and infinite routing loops do not crash the kernel. We
also add a test case for wireguard interfaces nesting packets and
sending traffic between each other, as well as the loop in this case
too. We make sure to send some throughput-heavy traffic for this use
case, to stress out any possible recursion issues with the locks around
workqueues.
* send: cond_resched() when processing tx ringbuffers
Users with pathological hardware reported CPU stalls on CONFIG_
PREEMPT_VOLUNTARY=y, because the ringbuffers would stay full, meaning
these workers would never terminate. That turned out not to be okay on
systems without forced preemption. This commit adds a cond_resched() to
the bottom of each loop iteration, so that these workers don't hog the
core. We don't do this on encryption/decryption because the compat
module here uses simd_relax, which already includes a call to schedule
in preempt_enable.
* selftests: initalize ipv6 members to NULL to squelch clang warning
This fixes a worthless warning from clang.
* send/receive: use explicit unlikely branch instead of implicit coalescing
Some code readibility cleanups.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
(cherry picked from commit 4f6343ffe7)
As announced on the mailing list, WireGuard will be in Linux 5.6. As a
result, the wg(8) tool, used by OpenWRT in the same manner as ip(8), is
moving to its own wireguard-tools repo. Meanwhile, the out-of-tree
kernel module for kernels 3.10 - 5.5 moved to its own wireguard-linux-
compat repo. Yesterday, releases were cut out of these repos, so this
commit bumps packages to match. Since wg(8) and the compat kernel module
are versioned and released separately, we create a wireguard-tools
Makefile to contain the source for the new tools repo. Later, when
OpenWRT moves permanently to Linux 5.6, we'll drop the original module
package, leaving only the tools. So this commit shuffles the build
definition around a bit but is basically the same idea as before.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
(cherry picked from commit ea980fb9c6)
f4d759b dhcp.c: further improve validation
Further improve input validation for CVE-2020-11752
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit 9e7d11f3e2)
cdac046 dns.c: fix input validation fix
Due to a slight foobar typo, failing to de-reference a pointer, previous
fix not quite as complete as it should have been.
Improve CVE-2020-11750 fix
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit 9f7c8ed078)
ab7a39a umdns: fix unused error
45c4953 dns: explicitly endian-convert all fields in header and question
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit 22ae8bd50e)
gcc 8 & 9 appear to be more picky with regards access alignment to
packed structures, leading to this warning in dns.c:
dns.c:261:2: error: converting a packed ‘struct dns_question’ pointer
(alignment 1) to a ‘uint16_t’ {aka ‘short unsigned int’} pointer
(alignment 2) may result in an unaligned pointer value
[-Werror=address-of-packed-member]
261 | uint16_t *swap = (uint16_t *) q;
Work around what I think is a false positive by turning the warning off.
Not ideal, but not quite as not ideal as build failure.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit 02640f0147)
kmod-usb-dwc2 and kmod-usb-ledtrig-usbport are not target default packages, and
Belkin F7C027 does not have a USB port anyway. Just drop it.
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
(cherry picked from commit 1dedad2a00)
This service file has been misplaced from the very beginning.
Fixes: dcc34574ef ("oxnas: bring in new oxnas target")
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
(cherry picked from commit 01961f163d)
ZyXEL Keenetic has 8MB flash, but OpenWrt uses only 4MB.
This commit fixes the problem.
WikiDevi page [1] says that ZyXEL Keenetic has FLA1: 8 MiB, there is
an article with specs [2] (in Russian).
[1] https://wikidevi.wi-cat.ru/ZyXEL_Keenetic
[2] https://3dnews.ru/608774/page-2.html
Fixes: FS#2487
Fixes: a7cbf59e0e ("ramips: add new device ZyXEL Keenetic as kn")
Signed-off-by: Alexey Dobrovolsky <dobrovolskiy.alexey@gmail.com>
(cherry picked from commit fea232ae8f)
Use power led for device status.
The status led behavior has already been fixed in af28d8a539
("ath79: add support for GL.iNet GL-AR750S") when porting the
device to ath79. This fixes it for ar71xx as well.
Signed-off-by: Jan Alexander <jan@nalx.net>
[minor commit title/message adjustments]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
(cherry picked from commit d394c354ee)
The labels on the LAN ports of the TP-Link Archer C60 v1/v2 are
actually inverted compared to the ports of the internal switch.
Add this information to 02_network.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
(cherry picked from commit 14a07fa1f0)
The adjustment of the MAC address for Archer C60 v2 in 10_fix_wifi_mac
is broken since a "mac" partition is not set up for this device on
ar71xx. Instead, the MAC address is already patched correctly in
11-ath10k-caldata.
Remove the useless adjustment.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
(cherry picked from commit cbdc919024)
The MAC addresses for lan/wan are swapped compared to the vendor
firmware. This adjusts to vendor configuration, which is:
lan *:7b label
wan *:7c label+1
2.4g *:7b label
5g *:7a label-1
Only one address is stored in <&mac 0x8>, corresponding to the label.
This has been checked on revisions v1, v2 and v3.
Since ar71xx calculates the ath10k MAC address based on the ethernet
addresses, the number there is adjusted, too.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
(cherry picked from commit 14eb54938b)
Doing up & down on non-Ethernet devices (e.g. monitor mode interface)
was consuming memory.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
(cherry picked from commit ec8e8e2ef0)
This reverts commit cc78f934a9 since it
didn't contain a reference to the CVE it addresses. The next commit
will re-add the commit including a CVE reference in its commit message.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
