Despite Westwood's theoretical advantages, in nearly
every benchmark we ran last year, TCP cubic won, whether it be
on correct RTT estimates, amount of buffering, responsiveness,
etc. on current hardware and software designs.
(both need timestamps on to work well, besides)
TCP cubic is better maintained and understood than westwood,
also.
While a scenario where westwood would win possibly exists,
there is too much buffering in the wifi stack in particular
at present, to see any improvement.
If you wish to exercise various TCPs under contention,
the current svn head of netperf (2.6) has options to switch
congestion control agorithms on the fly, as does iperf.
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32514 3c298f89-4303-0410-b956-a3cf2f4a3e73
Hardware filtering must always be enabled as long as there is an Ethernet
device registered, and use device tree for setting the link activity and
buffer shifting enable/disable bit.
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32486 3c298f89-4303-0410-b956-a3cf2f4a3e73
This caused stalls in the Ethernet DMA block, so until properly
written and sorted out, fallback to the assembly version instead.
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32470 3c298f89-4303-0410-b956-a3cf2f4a3e73
* use 'fsync' with 'dd' for image writing
* save config to boot partition
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32465 3c298f89-4303-0410-b956-a3cf2f4a3e73
This target currently only supports Moschip's MCS8140 SoC, but support
for other chips in the same family (MCS8142, MCS8144) will be easy to add.
Target support is entirely using Device Tree for probing peripherals.
Drivers support include:
- PCI
- USB 1 & 2
- watchdog
- random number generator
- UART
- timer
- internal Ethernet PHY
- Ethernet MAC core
Support for the following boards is included using Device Tree
- Devolo dLAN USB Extender
- Tigal RBT-832
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32462 3c298f89-4303-0410-b956-a3cf2f4a3e73
This patch helps properly detect a WNDRMAC device. Before this
patch the model is detected as "NETGEAR ?????????N".
Signed-off-by: Roman A. aka BasicXP <x12ozmouse@ya.ru>
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32453 3c298f89-4303-0410-b956-a3cf2f4a3e73
This adds kmod-leds-wndr3700-usb package to a default build for
WNDR3700, WNDR3800 and WNDRMAC in order to support the green USB LED.
Signed-off-by: Roman A. aka BasicXP <x12ozmouse@ya.ru>
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32452 3c298f89-4303-0410-b956-a3cf2f4a3e73
This patch adds support for the Korean made Petatel PSR-680W Wireless CDMA Router.
The platform is based on Ralink RT3052.
http://wiki.openwrt.org/toh/petatel/psr-680w
Signed-off-by: Dmitry Shmygov <shmygov at rambler.ru>
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32450 3c298f89-4303-0410-b956-a3cf2f4a3e73
This patch enables few extra kernel options for the kvm_guest subtarget:
- Tickless kernel to avoid timer ticks in idle guests, reduces CPU usage
- Enable paravirtualization steal time support
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32436 3c298f89-4303-0410-b956-a3cf2f4a3e73
BCM6338 and BCM6338 have their MSG_CONTROL register width of 8-bits instead of
16-bits. We were previously using a 16-bits write which corrupted the first
byte of the TX FIFO. Also the message type was always set to Full-duplex even
in the case of half-duplex messages.
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32409 3c298f89-4303-0410-b956-a3cf2f4a3e73
The dcache bug that it works around is a generic issue, not a brcm47xx cache quirk
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@32395 3c298f89-4303-0410-b956-a3cf2f4a3e73
