Update header file appropriately and disable read for ownership
Note that the FIQ support implements a workaround that provides a performance
boost over the traditional upstream workaround which ensures cache lines
are exclusive on driver CPU using 'read for ownership'.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
target/linux/cns3xxx/config-3.3 | 2 +-
target/linux/cns3xxx/patches-3.3/460-cns3xxx_fiq_support.patch | 9 ++++-----
2 files changed, 5 insertions(+), 6 deletions(-)
SVN-Revision: 33827
- properly check for the link up condition in the phylib adjust callback
- do not assign skb->dev, eth_type_trans() already does it
- handle skb allocation failures in the refill path
SVN-Revision: 33780
Commit r33248 introduced a regression by passing the board name plus
'board=' in the third argument. The board name string has to be put
into the image as identifier of the image type.
[juhosg: fix OpenMesh template instead of use of string substitution]
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 33767
Handle the different vlynq device versions, and make the vlynq code more
in line with the original vlynq implementation.
Patch from Daniel Gimpelevich.
SVN-Revision: 33755
At least devices without a switch present on the MDIO bus still get a valid
link and work as expected (such as the Linksys WAG54Gv2).
SVN-Revision: 33751
The Laguna boards do not use all the same pins for SDHCI as the Cavium
reference board.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 33684
The ARM11MPCore Timer/Watchdog registers start at offset 0x600 which is where
all mpcore-wdt boards point the driver base too. I believe this is wrong
because 0x600 is aliased to the timer/watchdog of the 'current CPU' where
0x700 is CPU0's timer/watchdog, and 0x800 is CPU1's timer/watchdog. Thus
if your timer/watchdog application is switching between CPU's it can end up
writing to the wrong CPU's registers which results in random board resets
from watchdog timeouts etc.
This patch forces the timer/watchdog driver to use CPU0's registers always.
Its my opinion that other mpcore-wdt boards should be doing the same thing.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 33683
This allows sysupgrade for devices such as the Gateworks Avila/Cambria
product families based on the ixp4xx using the redboot bootloader with
combined FIS directory and RedBoot config partitions on larger FLASH
devices with larger eraseblocks.
This second iteration of this patch addresses previous issues:
- whitespace breakage fixed
- unlock in all scenarios
- simplification and fix logic bug
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 33681
Enabling OPTIMIZE_FOR_SIZE cuts the kernel binary down by 50K which is
critical for some small-footprint boards.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 33679
Several new features for newer boards:
- add additional UARTs present on some boards
- add additional LEDs present on some boards
- add HSS audio device codec present on some boards
- add support for GSC present on some boards
- add per model setup support for newer boards
- set FLASH window per-model for boards with larger FLASH
Some fixes:
- add IRQ mapping for additional PCI devices (USB Host)
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 33647
Several new features for newer boards:
- add irq mapping for additional devices
- add platform data for i2c bus to SFP modules
- add additional UARTs present on some boards
- increased R/W delay for expansion bus UARTs
- add additional LEDs present on some boards
- add GPIO exports and configuration
- add ENET switch config present on some boards
- add support for GSC present on some boards
- added per model setup support for newer boards
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 33644
Before it was tried to initialize the deactivated PCIe core in client
mode, but this causes the SoC to hang. Just do not initialize it at all
and ignore the core it is not working and nothing is connected to it
when the specific bit is set in the boardflags.
SVN-Revision: 33620
The new version of the ALL0256N now got 8MB SPI NOR flash instead of 4MB.
In order to expose the whole amount of flash, add another image which contains
the corresponding GENERIC_8M mtdparts.
[juhosg: change suffix from '8m' to '8M']
Patch by: Daniel Golle <dgolle@allnet.de>
SVN-Revision: 33614
Sometimes the PCIe card indicates that it has a sprom somewhere and we
are able to read the memory region, but it is empty and not valid. In
these cases we should try to use the fallback sprom as a last chance.
SVN-Revision: 33601
The chip common and the PCIe code are accessing the sprom struct which
is not filled when these cores are initialized. Fix this by adding an
early initialize and fill the sprom struct before accessing it in other
code.
SVN-Revision: 33600
The bcma based SoCs with a ieee80211 core on the SoC and an other
connected via PCIe or USB store the sprom for the SoC with a sb/1/
prefix. The SoC with just one wifi core do not use prefixes. The
BCM4706 do not use a prefix for the SoC part at all, because the prefix
is the path to the ieee80211 core and there is non on the BCM4706.
SVN-Revision: 33597
Resolves an issue where isochronouse USB would cause the driver to hang as
well as scheduling issues.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 33579
This feature was experimental in old kernels but
that flag has been removed in 2.6.36. Additionally,
the option is enabled by default since then.
See: http://www.spinics.net/lists/linux-usb/msg31993.html
SVN-Revision: 33574
For cns3xxx SCU_CONFIGURATION always shows multipe cores but SCU_CPU_STATUS
shows which ones are active.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
SVN-Revision: 33566
Sometimes the port 0 was the WAN port or there could be some device
where the port 0 is not wired to an actual Ethernet port. Now telnet in
failsafe will listen on all ports of the switch so user do not have to
search for the correct port any more. This change will make the switch
leak traffic between the WAN and the LAN port, but you should not use
failsafe while connected to a untrusted network.
SVN-Revision: 33544
With these images, it is possible to use the 'unused'
partition of the flash. The 'fat' images can be installed
with the sysupgrade command. When a 'fat' image is
installed from a regular one, the platform specific
sysupgrade script copies the calibration data to the
end of the flash. Likewise, when a regular image is
installed from the 'fat' version the sysupgrade script
copies the calibration data back to the original location.
SVN-Revision: 33540
When not removing support for the wait instruction, when ruining on the
BCM4706 SoC the system stops after msleep() without an exception. I
haven't seen this problem on the other Broadcom SoCs with a 74K CPU,
but in the Broadcom SDK it is deactivated for all CPUs of this type.
SVN-Revision: 33496
It should be implemmented in a cleaner way.
This reverts the following commits:
[ar71xx] dir-825-b1: Fix wholeflash images
[ar71xx] dir825b1: Add image for DIR-825-B1 that uses the whole flash"
SVN-Revision: 33459
This is a patch to connect eth0.1 (lan) to the only
ethernet connector available.
D-Link DAP-1350 goes bricked with AA 12.09-beta.
This is because the current esw configuration for the
board connects eth0.2 (wan) to the *ONE and ONLY*
ethernet connector available, preventing initial access
to the board through 192.168.1.1 on eth0.1 after flashing,
effectively bricking the board.
There are things that should be done to make this board
really useful, but for AA 12.09, this one liner is
indispensable.
Signed of by Yoichi Shinoda <shinoda@jaist.ac.jp>
SVN-Revision: 33411
Somehow detecting the RAM size in common/setup.c doesn't
work here, it always detects 64M and then crashes on devices
with less RAM.
Probably using MEMC_REG_SDRAM_CFG1 to know the RAM size is how
it could be, for now I use the mem=32M kernel parameter to get
stuff working.
Signed-off-by: Daniel Golle <dgolle@allnet.de>
SVN-Revision: 33381
This patch adds preliminary support for the Mikrotik RB2011L.
The NAND flash is not yet supported, so only the initramfs
kernel can be used for now.
[juhosg: remove unnecessary PCI/USB related stuff, the board
does not use that, update commit message]
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
SVN-Revision: 33344
Add missing andmask to ramips_esw register read for recv_good value.
Without the mask, recv_bad leaks into the recv_good packet count.
Didn't notice the bug before since you don't usually get bad
packets, so I only saw it when I was playing with overlength packets
earlier...
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33322
Use doubletagging to disable ramips_esw vlan by default, it seems
more reliable.
Daniel Golle found an issue where sometimes (possibly only for
RT3352) the default vlan disable method (clearing en_vlan, untag,
doubletag and putting all ports into vlan 0) doesn't work and the
packets get sent out vlan-tagged with vlan 0.
Instead switch to using the doubletagging method (allow doubletagged
packets, put all ports into vlan 0 with untag enabled) by default.
Unless someone figures out a way to really globally disable vlan for
this switch, this seems like the best (most reliable) option.
I did some tests regarding maximum packet size and did not see any
difference between the two methods, both allow for slightly bigger packets
than the ramips_main.c ethernet driver (ping stops going through
above "ping -s 1472" (1514 bytes), on the switch packets are recv_good until
"ping -s 1490", or about 1532 bytes).
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33321
Power down phy on disabled switch ports.
Haven't measured this myself yet, but according to this
http://www.8devices.com/community/viewtopic.php?f=6&t=156
it can save about 300mW of power.
[juhosg: fix checkpatch warning]
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33304
Add support for 8devices Carambola dev board.
The Carambola is a small RT3050-based development board with two
ethernet ports, on-board chip antenna, usb and plenty of accessible
gpio ports, sold by 8devices.
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33303
Rename POC registers.
The current code uses POC1-POC3.
The datasheet uses:
POC1: Port Control 0
POC1: Port Control 1
POC2: Port Control 2
So the first POC1 is a typo that should have been POC0, rename the
registers to POC0-POC2 accordingly.
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33302
Stop handling VLAN setup in the kernel.
Removes the obsolete RT305X_ESW_VLAN_CONFIG_BYPASS option I added for
WL-351 and add some extra comments.
Also removes the en_vlan per-port flag that isn't very useful really, it now
is only controlled by the global enable_vlan flag.
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33301
Add switch setup to uci network defaults.
The 'lan' attribute is used to decide into which vlan to put the
port.
Currently 'disable' is never set, but the intention is to use this
for devices like the 8devices Carambola, which only has two ports
hooked up to the SoC.
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33300
Add swconfig support to ramips_esw.c
This patch adds swconfig support for ramips_esw:
Tested on both D-LINK DIR-300 B1 and Sitecom WL-351 (external
rtl8366rb on internal port 5).
I've made sure that in the enable_vlan=0 case it behaves like a dumb
switch, so external switches should work fine with vlans and
verified this on the WL-351.
The current state shown by swconfig is always read directly from HW
registers, new settings only show after 'swconfig dev rt305x set apply'.
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33299
Minor indentation cleanup.
Prepare for the main swconfig patch by cleaning up indentation a bit.
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 33298
An improved LED support for Buffalo WZR-HP-AG300H.
There will be two new packages to go with this patch
(ar922x-led-fix-hotplug and ar922x-led-fix-init).
These packages essentially perform the same task of
disabling "JTAG over GPIO pins" function which is
currently required for LEDs on AR922x emac GPIO to work,
and are both considered a temporary fix that circumvent
this problem.
[juhosg: add the hotplug script from the proposed
ar922x-led-fix-hotplug package to ar71xx/base-files]
Signed-off-by: Yoichi Shinoda <shinoda@jaist.ac.jp>
SVN-Revision: 33294
Commit 48788063266ad99bdab03b53cbe7e1a564c2d9b3 (r33251) introduced a
regression with board= kernel command line parameter incorrectly
initialised.
The faulty builds propogated to the snapshots download area. This makes
the device unbootable and TP-Link users have to attach the serial to
recover.
Here is the q&d way i used to identify MIPS boards that didn't have a
corresponding entry in the image/Makefile:
for i in `find ../files/arch/mips/ath79/ -type f -exec cat {} \; \
| sed -n '/MIPS_MACHINE/ { s/^[^"]*"\([^"]*\)".*$/\1/; p; }'`; do
grep -q ,$i, Makefile || echo $i;
done > missing-boards.new
Reported-by: Casper on IRC
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
SVN-Revision: 33274
OpenWRT runs in a eSATA sheevaplug like a charm. This patch just enables
CONFIG_MACH_ESATA_SHEEVAPLUG so the proper MACHINE_START definition
is compiled in.
Signed-off-by: Unai Uribarri <unaiur@gmail.com>
SVN-Revision: 33266
For the pure convenience of having a correctly named
image and system name in /proc/cpuinfo , until we can
do that by having system names in DTS...
Signed-off-by: Daniel Golle <dgolle@allnet.de>
SVN-Revision: 33224
The TL-WDR3600 is identical to the TL-WDR4300 with the exception that is has
only two antennas.
[juhosg: remove the custom machine type, change the board name instead]
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 33219
Generate image for the ALL0239-3G which can be flashed through the
chipset-vendor SDK based firmware's web-interface and bootloader.
The bootloader seems to ignore uImage checksum errors, but does complain about
them once the 0xDEADC0DE was replaced by an actual JFFS2 page.
I'm working on implementing fixtrx for uImage in the mtd package to solve this.
Signed-off-by: Daniel Golle <dgolle@allnet.de>
SVN-Revision: 33206
rt2x00 still needs some patching as the radio doesn't come to life.
Installation via webflash.
[juhosg: fix whitespace issues, remove rt305x_register_usb
from machine setup because the board has no USB port]
Signed-off-by: Mikko Hissa <mikko.hissa@uta.fi>
SVN-Revision: 33205
* Creation of uImage for WNR854T only done once (before 2x for jffs2 build and 1x for squashfs build)
* Got rid of unneccessary padding of rootfs partition
* ARM zImages always need a machine id, therefore do not copy generic (=no id) uImage to BIN_DIR, instead copy zImage
* Generalized functions for easier re-using and enhancing (e.g. D-Link DNS 323 implementation would be only a couple lines)
* Copy rootfs partitions to BIN_DIR, just like it is done for D-Link DNS 323
* Use variables to allows easily changing for custom builds, e.g. kernel mtd size for symbols
* Size check of kernel files to avoid builds that break devices
* Use for "-sysupgrade" and "-factory" in image names (like ar71xx, brcm63xx, etc.) to avoid questions about which image to use
Signed-off by: Matthias Buecher <mail@maddes.net>
SVN-Revision: 33174
In order to get OHCI/EHCI working on the Rt3352, the platform device must be
named so rt3883-?hci will recognize it.
Signed-off-by: Daniel Golle <dgolle@allnet.de>
SVN-Revision: 33145
As the userspace has no means to determine the maximum possible timeout, use
that as the default and let the userspace lower it when necessary.
As the result the usual OpenWrt install (with busybox's watchdog trying to set
the timeout to 60s on start) is using a 33s timeout on an RT3052 clocked at
384MHz instead of the current 20s default.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
SVN-Revision: 33144
D-Link's DIR-620 allows to flash a uImage directly from its web-interface but
for that the image name should be set to DIR_620.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
SVN-Revision: 33143
Full functionality runtime tested, installation via OEM web-interface requires
a follow-up patch.
Thanks go to Sergey Vasilyugin for his patch that pointed me at the missing
bits (including the antiparallel wps led).
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
SVN-Revision: 33142
Fixes this section mismatch warning:
WARNING: vmlinux.o(.meminit.text+0xb14): Section mismatch in reference
from the function memblock_insert_region() to the function
.init.text:crashlog_init_memblock()
The function __meminit memblock_insert_region() references
a function __init crashlog_init_memblock().
If crashlog_init_memblock is only used by memblock_insert_region then
annotate crashlog_init_memblock with a matching annotation.
Also, remove the __init annotations from 'include/linux/crashlog.h'
SVN-Revision: 33137
The following patch corrects a hard reset that occurs on alix and
soekris net5501 hardware. Specifically, this is an issue with the via
rhine NIC driver. Under periods of extreme load, the via rhine driver
can cause a reset of the entire system. When this happens no output is
seen on the console, the device simply reboots. This was reported in the
following ticket:
Consistent crash on Soekris NET5501
https://dev.openwrt.org/ticket/11882
I was able to reproduce the problem locally with numerous alix 2d13
platforms. Under heavy 64 byte frame size load from a Smartbits traffic
generator, the alix hardware will hard reset in under 5 minutes. The
same hardware, under the same load, on FreeBSD 9.0 does not reset.
The patch below backports most of the via rhine changes from kernel 3.5.
With OpenWrt trunk being on kernel 3.3.8, it seemed prudent to bring
these changes in to avoid this serious issue.
I've tested this heavily in my test environment and was able to keep the
device stable under load for extended periods of time without any resets.
Also, a nice side effect of this change is that it significantly raises
the performance the platform. Without this patch the alix 2d13 can move
approximately 29,000 packets per second at 64 byte frame sizes. After
this patch the alix can move 52,000 packets per second at 64 byte frame
sizes.
I put this patch under the x86 patches as those are the only platforms
with via rhine hardware that I'm aware of. If it needs to go some place
else please let me know.
Thanks,
Adam
Signed-off-by: Adam Gensler <openwrt@kristenandadam.net>
SVN-Revision: 33072
This patch fixes the GPIO ALTSEL settings for some of the GPIOs used by the PCI subsystem in Lantiq Danube. These changes are required for more than one PCI device to work. Tested with an ARV7510PW having two PCI-devices; a VIA USB controller and a Ralink WLAN mini-PCI card.
Signed-off-by: Matti Laakso <malaakso@elisanet.fi>
SVN-Revision: 33054
The ARV4525 board init is passing the wrong offset to
arv_register_ath5k() resulting in a bogus mac.
Signed-off-by: Tobias Diedrich <ranma+openwrt@tdiedrich.de>
SVN-Revision: 32993
One condition before codel_Newton_step() was not good if
we never left the dropping state for a flow. As a result
rec_inv_sqrt was 0, instead of the ~0 initial value.
codel control law was then set to a very aggressive mode, dropping
many packets before reaching 'target' and recovering from this problem.
Brought over from 3.5-stable
Signed-off-by: Dave Taht <dave.taht@bufferbloat.net>
SVN-Revision: 32950
Here is the uci-defaults version of the led support for
the Buffalo WLAE-AG300N.
Signed-off-by: Yoichi Shinoda <shinoda@jaist.ac.jp>
SVN-Revision: 32932
The patch set for Buffalo WLAE-AG300N initial support.
There is another patch for wireless led support that
is posted separately.
Note on serial console:
This unit has buffalo standard 4 pin console, but the unit may not power
on if some console apparatus is connected. This is probably due to some
electronic interaction between the unit's electronic power switch circuit
and the serial console apparatus. If this happens, it is required to power
on the unit without the console, then quickly (re-)attach it.
[juhosg: fix coding style]
Signed-off-by: Yoichi Shinoda <shinoda@jaist.ac.jp>
SVN-Revision: 32931
Signed-off-by: Miguel GAIO <miguel.gaio@efixo.com>
Signed-off-by: Jonas Gorski <jonas.gorski+openwrt@gmail.com>
[jonas.gorski@gmail.com: add network config for nb6, make switch fix a
platform patch]
SVN-Revision: 32924
Gabor Juhos
Felix Fietkau
John Crispin
Florian Fainelli
Imre Kaloz
Jo-Philipp Wich
Jonas Gorski
Hauke Mehrtens
Daniel Dickinson
Luka Perkov