Move defines and board_info structs back to their proper locations so
(de-)selecting SOC support won't break the build.
Reported-by: Álvaro Fernández Rojas <noltari@gmail.com>
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 41550
They should be unnecessary with fq_codel, and simplifying rules helps
with performance
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 41549
A reboot is issued by setting the watchdog timer to a low value, when
the value is decreased from 1 to 0 by the hardware the system will
reboot. Setting this value to 1 does not work every time, it is a
little bit more reliable when setting it to 3. This timer runs with
some MHz or KHz so nobody will notice the increased waiting.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41548
broadcom-wl now builds in the mips74k profile, so remove these chips
from the generic profile.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41546
When the original mac addresses are not valid assign the first one to
the internal wifi core, this matches the manufacture mac address in
most cases. In addition refresh the patches.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41545
The previous attempt (r40717) to fix squashfx made the system bootable but
overlayfs remained broken.
This enables block2mtd again using the right partition number and naming
and adds MTD_ROOTFS_SPLIT option.
However onboard rootfs partition name had to be renamed to place overlay on
CF instead of onboard NAND.
Signed-off-by: Roman Yeryomin <roman@advem.lv>
SVN-Revision: 41541
No config option should be set to m in the default config, because
there is also some script needed which picks the module up into the
image and that is missing this way. Always create a kmod when building
a module.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41539
Most ar7 devices have only 16 MiB RAM, so select low_mem to give them
a chance to work.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 41532
When the Ethernet controller is powered down and someone wants to
access the mdio bus like the witch driver (b53) the system crashed if
PCI_D3hot was set before. This patch deactivates this power sawing mode
when a switch driver is in use.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41527
According to the thread https://forum.openwrt.org/viewtopic.php?id=48281
b53 uses GPIO 7:
[ 4.470000] b53_common: [DBG] b53_switch_reset_gpio using 7
and causes device to self-reboot. GPIO 8 was found in CFE boot log:
"Reset switch via GPIO 8 ..."
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 41526
This updates linux-firmware to version
7f388b4885cf64d6b7833612052d20d4197af96f from 2014-06-04.
In addition it adds new firmware for Intel, Broadcom and Realtek devices.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41523
Some devices have an nvram setting, which make broadcom-wl turn of the
led all the time. When the driver is switched on and we find such a
setting we replace it with a better value.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41518
Now the mac address is increased by two for the first fallback mac
address. It could be that the address increased should be used for the
wan port.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41515
Some series of devices are using the same mac address, which causes
conflicts for these devices we generate a new mac address based on the
base address. This patch adds this functionality to braodcom-wl.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41514
The address prefix 00:90:4C is used by Broadcom in their initial
configuration. When a mac address with the prefix 00:90:4C is used
all devices from the same series are sharing the same mac address.
To prevent mac address collisions we replace them with a mac address
based on the base address.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41513
Without this patch the switch config for the adm switch did not worked,
because hey are not names eth0.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41512
This patch deactivates the statistics, adds a missing lock
initialization and fixes a waring.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 41511
Old OpenWrt releases were using network configs that are not valid
anymore. They were specifying ports tagging in a different way (or were
not tagging CPU at all) and were using VID 0 which is a reserved value.
Modifying network configuration to apply all needed changes would be
pretty tricky. Script /etc/init.d/netconfig that generates new config is
quite complex itself.
So instead let's save the most important settings, regenerate config
from the scratch and restore values. This should work for 99% of users.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 41500
