OpenWRT changed the default fq_codel sch->limit from 10240 to 1024,
without also adjusting q->flows_cnt. Eric Dumazet explains below that
you must also adjust the buckets (q->flows_cnt) for this not to break.
Eric explains: Limit of 1024 packets and 1024 flows is not wise I think.
(If all buckets are in use, each bucket has a virtual queue of 1 packet,
which is almost the same than having no queue at all)
I suggest to have at least 8 packets per bucket, to let Codel have a
chance to trigger. So you could either reduce number of buckets to 128
(if memory is tight), or increase limit to 8192.
flows_cnt is now set to 1024/8=128
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Add node aliases to dtsi files.
Reword dts files so they're more in-line with upstream.
Fix some more warnings and errors reported by dtc
Signed-off-by: Stanislav Galabov <sgalabov@gmail.com>
Building for octeon fails with
'arch/mips/vdso/vdso-n32.so.dbg' already contains a '.MIPS.abiflags'
section
if the file already exists from a prior build.
Use the same workaround as the one for vdso.so.dbg committed in
9eb155353a.
Commit 91f205acaf extended the workaround
to cover vdso-o32.so.dbg but missed the vdso-n32.so.dbg which is added
now by this change.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
This patch introduces serial0 aliases in the ramips DTS files, which can
then be used to denote the active console instead of relying on bootargs.
Signed-off-by: Stanislav Galabov <sgalabov@gmail.com>
This router only has one ethernet port, so a VLAN is useless here, now that the rt3050 TCP bug that happened without VLANs has been fixed for a very long time.
Add this router to the VLAN-less config that is used by other single-port routers.
Also fix MAC address detection code since this router has no WAN port.
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
Some routers only have one port, so eth0 is used without VLANs for them.
Revision r47720 introduced some changes, but wrongly confused "enable" with "reset".
VLANs need to be disabled for those routers, and the switch may be reset.
Fix this, by explicitly disabling VLANs instead of resetting the switch for these routers.
Also merge duplicate configuration for the "m2m".
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
The new rt3050 switch driver doesn't have problems with TCP when not
using VLANs.
This piece of code also broke failsafe for all routers where the LAN
port is not wired to port 0 of the internal switch.
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49293
Port 0 is the only ethernet port on this router, so disable all other PHYs
in order to save power.
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49292
Port 4 is the only ethernet port on this router, so disable all other PHYs
in order to save power.
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49291
This patch allows configuring ports to be disabled in the device tree; this
saves power, since disabling ports here actually disables power to ethernet
PHYs.
Line 444 enables all ethernet ports, so line 487 is getting zero ports to be
disabled, except for port 5 in SoCs where this is not implemented as it will
be sticky disabled in register POC0. Because of this, the code will still read
the switch configuration and OR it to the device tree setting.
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49290
Line 444 is actually enabling all switch ports by setting the disable bits
to 0. This needs to be done because the bootloader sets all ports to disabled
by default (which is the case for at least one router based on RT5350).
So, this patch fixes the comment in line 443.
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49289
The FCT2 esw register should be set to 0x2500C to have "unknown IPv6
multicast" packets broadcasted to every port, instead of dropped.
The previous value only let those packets go through ports 1 and 3.
"Unknown IPv6 multicast" packets include packets needed by ICMPv6 echo
requests addressed to well-known addresses, such as ff02::1 (MAC address
is 33:33:00:00:00:01 in this case).
Please note that by default ICMPv6 echo requests to ff02::1 are not replied
to by the router because of ip6tables considering those packets to be invalid.
But this is another bug/patch. ;)
Signed-off-by: Vittorio Gambaletta <openwrt@vittgam.net>
SVN-Revision: 49287
VGV7510KW2 with VRX288 v1.2 has brnboot 1.8 installed. Starting with
this brnboot version, the "GPHY Clock Source" isn't set anymore by
brnboot, with the result that xrx200-net fails to probe/initialize the
phys.
Use the phy clock source device tree binding to specify the clock source.
Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49284
brnboot based devices can have two Image partitions. When flashing
images via the brnboot recovery web interface, the Image partitions are
written alternating.
The current active Image partition is stored in the first byte of the
Primary_Setting partition by using 0x00 for Code_Image_0 and 0x01 for
Code_Image_1.
By using the information about the active "Code Image", it is possible
to ensure that the rootfs belongs to the current booted Image/Kernel.
Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49281
Starting with kernel 4.4, the use of partitions as direct subnodes of the
mtd device is discouraged and only supported for backward compatiblity
reasons.
Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49280
Use the same name for TP-Link images as it was with the old image build
code.
Move the BOARD_ID export to the TP-Link image build recipe, to indicate
that the variable is only related in this context.
Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49279
Based on the vg3503j_gphy_led.sh script published in the VG3503J wiki
article, the OEM Firmware uses the following PHY led functionality:
gphy led 0: LINK/ACTIVITY
gphy led 1: LINK
gphy led 2: ACTIVITY
Signed-off-by: Mathias Kresin <openwrt@kresin.me>
SVN-Revision: 49278
