commit r46299 broke the mvebu target because one patch does not apply
any more. This commit closes#20070.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46303
This updates valgrind to version 3.10.1 and make configure work with
kernel version > 4.0. This patch was taken from upstream valgrind.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46302
The musl build "fix" introduced in r45108 removed all netinet/ether.h
includes, which made the prototypes of ether_aton and ether_ntoa
unavailable. As a result, the compiler assumed they return int instead
of a pointer. This currupted the pointer on 64bit targets, causing ebtables
to segfault in commands containing MAC addresses.
Since r46161 made it possible to include both the kernel and the libc
if_ether.h as long as the libc version is included first, this patch
changes the fix to remove the linux/if_ether.h from the ebtables source
(so the fixed version from the kernel is used) and ensures netinet/ether.h
is included early.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 46292
During certificate verification, OpenSSL (starting from version 1.0.1n and
1.0.2b) will attempt to find an alternative certificate chain if the first
attempt to build such a chain fails. An error in the implementation of this
logic can mean that an attacker could cause certain checks on untrusted
certificates to be bypassed, such as the CA flag, enabling them to use a valid
leaf certificate to act as a CA and "issue" an invalid certificate.
This issue will impact any application that verifies certificates including
SSL/TLS/DTLS clients and SSL/TLS/DTLS servers using client authentication.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 46285
In fb6f62e97733312053ab593fcf68eea47a21169e several settings
are set on the ethernet device, but they are not working.
Fix Ethernet by setting the right values.
Signed-off-by: Christian Mehlis <christian@m3hlis.de>
SVN-Revision: 46281
no-op since 2.6.35
removed in Kernel 4.1
see https://lwn.net/Articles/380931/
Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 46280
This patch adds the RFKill GPIO control switch and enables another GPIO to
control power supply to USB Ports by emulating an LED GPIO for WDR4900v1.
Signed-off-by: Guo Wei Lim <alphasparc@gmail.com>
SVN-Revision: 46279
Make wpan depend on not kernel 3.18 so it will get activated on all
other kernel versions like 4.0 and 4.1.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46276
Internal GPIO pins are used for PoE passthrough setups in multi-port
routers. This patch implemnets control over this hardware feature for
Ubiquiti Nanostations and TP-Link CPE510.
Signed-off-by: Lars Kruse <lists@sumpfralle.de>
SVN-Revision: 46271
This reverts commit r45579.
With the latest change r46251 ("nvram: fix "Segmentation fault" caused
by setting memory out of buffer") nvram utility shouldn't crash anymore.
It was tested on 3 brcm47xx devices:
1) Unknown with 0x10000 NVRAM size (0x8000 offset)
2) Linksys E1000 V2.1 with 0x10000 (0x8000 offset)
3) Linksys WRT300N V1 with 0x10000 (0x8000 offset)
And 3 bcm53xx devices:
1) Buffalo WZR-600DHP2 with 0x160000 NVRAM size
2) Buffalo WZR-1750DHP with 0x10000 NVRAM size
3) Netgear R6250 V1 with 0x180000 NVRAM size
(all using 0 offset)
This is an important change as it allows reading whole NVRAM. This may
critical when reading some basic configuration (e.g. switch ports).
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 46253
Some MTD partitions with NVRAM have content starting in the middle. In
such case offset is set and nvram_header returns pointer to the middle.
It means we have to respect offset when calculating remaining space.
By the way use real MTD partition size (nvram_part_size variable) as we
may want to bump NVRAM_SPACE in the (very near) future.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 46251
This patch adds support for the Cisco WAP4410N, an access point that uses the
AR9132 SoC. Web upgrades from stock are not yet possible, UART access required
for the initial flash.
Signed-off-by: Ryan A Young <rayoung@utexas.edu>
SVN-Revision: 46250
The mips reloc patch introduced new allocations which were done before
add_unformed_module but never freed them in case of an error. A new hook in
Linux 3.19 called module_arch_freeing_init can be used for freeing memory
which were allocated during this init phase.
The problem can be seen when trying to load a module (via busybox insmod)
when it was already loaded.
free -m
for i in `seq 1 100`; do
/sbin/insmod /lib/modules/*/ath9k.ko >& /dev/null
done
free -m
This simple loop would leak ~3.2 MB.
Signed-off-by: Sven Eckelmann <sven@open-mesh.com>
SVN-Revision: 46247
Ath10k has now a proper method of providing calibration data via
the kernel firmware API. This patch can be dropped as all boards
now use the proper method.
Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 46245
Currently one board uses the "proper" way of providing firmware
with the name ath10k/cal-pci-<pci address>.bin, whereas other
boards patch board.bin template in preinit and rely on a custom
ath10k patch to use calibration data from the template.
This patch makes all boards use the first method.
Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 46244
