Instead of bit banging SPI to talk to the GPIO chip, use the hardware
led controllers intended for controlling the LEDs.
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Fixes the following security vulnerabilities:
CVE-2017-8890
The inet_csk_clone_lock function in net/ipv4/inet_connection_sock.c in the
Linux kernel through 4.10.15 allows attackers to cause a denial of service
(double free) or possibly have unspecified other impact by leveraging use
of the accept system call.
CVE-2017-9074
The IPv6 fragmentation implementation in the Linux kernel through 4.11.1
does not consider that the nexthdr field may be associated with an invalid
option, which allows local users to cause a denial of service (out-of-bounds
read and BUG) or possibly have unspecified other impact via crafted socket
and send system calls.
CVE-2017-9075
The sctp_v6_create_accept_sk function in net/sctp/ipv6.c in the Linux kernel
through 4.11.1 mishandles inheritance, which allows local users to cause a
denial of service or possibly have unspecified other impact via crafted
system calls, a related issue to CVE-2017-8890.
CVE-2017-9076
The dccp_v6_request_recv_sock function in net/dccp/ipv6.c in the Linux
kernel through 4.11.1 mishandles inheritance, which allows local users to
cause a denial of service or possibly have unspecified other impact via
crafted system calls, a related issue to CVE-2017-8890.
CVE-2017-9077
The tcp_v6_syn_recv_sock function in net/ipv6/tcp_ipv6.c in the Linux kernel
through 4.11.1 mishandles inheritance, which allows local users to cause a
denial of service or possibly have unspecified other impact via crafted
system calls, a related issue to CVE-2017-8890.
CVE-2017-9242
The __ip6_append_data function in net/ipv6/ip6_output.c in the Linux kernel
through 4.11.3 is too late in checking whether an overwrite of an skb data
structure may occur, which allows local users to cause a denial of service
(system crash) via crafted system calls.
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-8890
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9074
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9075
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9076
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9077
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9242
Ref: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.31
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Fixes the following security vulnerabilities:
CVE-2017-8890
The inet_csk_clone_lock function in net/ipv4/inet_connection_sock.c in the
Linux kernel through 4.10.15 allows attackers to cause a denial of service
(double free) or possibly have unspecified other impact by leveraging use
of the accept system call.
CVE-2017-9074
The IPv6 fragmentation implementation in the Linux kernel through 4.11.1
does not consider that the nexthdr field may be associated with an invalid
option, which allows local users to cause a denial of service (out-of-bounds
read and BUG) or possibly have unspecified other impact via crafted socket
and send system calls.
CVE-2017-9075
The sctp_v6_create_accept_sk function in net/sctp/ipv6.c in the Linux kernel
through 4.11.1 mishandles inheritance, which allows local users to cause a
denial of service or possibly have unspecified other impact via crafted
system calls, a related issue to CVE-2017-8890.
CVE-2017-9076
The dccp_v6_request_recv_sock function in net/dccp/ipv6.c in the Linux
kernel through 4.11.1 mishandles inheritance, which allows local users to
cause a denial of service or possibly have unspecified other impact via
crafted system calls, a related issue to CVE-2017-8890.
CVE-2017-9077
The tcp_v6_syn_recv_sock function in net/ipv6/tcp_ipv6.c in the Linux kernel
through 4.11.1 mishandles inheritance, which allows local users to cause a
denial of service or possibly have unspecified other impact via crafted
system calls, a related issue to CVE-2017-8890.
CVE-2017-9242
The __ip6_append_data function in net/ipv6/ip6_output.c in the Linux kernel
through 4.11.3 is too late in checking whether an overwrite of an skb data
structure may occur, which allows local users to cause a denial of service
(system crash) via crafted system calls.
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-8890
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9074
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9075
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9076
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9077
Ref: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-9242
Ref: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.71
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
CONFIG_SG_POOL symbol is selected only by CONFIG_SCSI, since the last
one is disabled by default then disable CONFIG_SG_POOL by default too.
And explicitly enable it only for platforms that use CONFIG_SCSI.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Add HAVE_IRQ_EXIT_ON_IRQ_STACK kernel configuration symbol that was
missed during backporting separate IRQ stack for MIPS from upstream.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
PHY core treats any positive return value as the auto-negotiation done
indication. Since we do not actually check any device register in this
callback then update it to return positive value with a neutral meaning
instead of the register flag to avoid confusing for future readers.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
The Marvel 88E6060 switch has an MDIO interface, but does not emulate
regular PHY behavior for the host. The network core can not detect using
the generic code, whether the connection via the attached PHY can be
used or not. The PHY's state machine is stuck in a state of
auto-negotiation and does not go any further so the Ethernet interface
of the router stay forever in the not-runing state.
Fix this issue by implementing the aneg_done callback to be able to
inform the network core that the Ethernet interface link to which the
switch is connected can be marked as RUNNING.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Drop the own PHY polling function and switch to using the kernel PHY
state machine. This change allows driver to work correctly with devices
that do not support PHY behaviour but whose driver could emulate
autonegotiation completion (e.g. MV88E6060 and IP17xx switches).
NB: earlier this driver rely on flaws in PHY core code and could use PHY
device without really starting it. But now (at least in kernel 4.9)
this trick no more work and network interface could stuck in not-running
state.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Few tiny fixes for issues caused by changes in the upstream:
- do not touch PHY IRQ array (core code initializes it itself now)
- add missed SET_NETDEV_DEV() invocation (causes segfault during phy
connection)
- use phy API inside the MDIO probe function instead of direct field
access (consider phy structure changes in upstream and prevent
similar issues in the future)
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
This patch adds support for the Ubiquiti EdgeRouter X-SFP and
improves support for the EdgeRouter X (PoE-passthrough).
Specification:
- SoC: MediaTek MT7621AT
- Flash: 256 MiB
- RAM: 265 MiB
- Ethernet: 5 x LAN (1000 Mbps)
- UART: 1 x UART on PCB (3.3V, RX, TX, GND) - 57600 8N1
- EdgeRouter X:
- 1 x PoE-Passtrough (Eth4)
- powered by Wallwart or passive PoE
- EdgeRouter X-SFP:
- 5 x PoE-Out (24V, passive)
- 1 x SFP (unknown status)
- powered by Wallwart (24V)
Doesn't work:
* SoC has crypto engine but no open driver.
* SoC has nat acceleration, but no open driver.
* This router has 2MB spi flash soldered in but MT
nand/spi drivers do not support pin sharing,
so it is not accessable and disabled. Stock
firmware could read it and it was empty.
Installation
via vendor firmware:
- build an Initrd-image (> 3MiB) and upload the factory-image
- initrd can have luci-mod-failsafe
- flash final firmware via LuCI / sysupgrade on rebooted system
via TFTP:
- stop uboot into tftp-load into option "1"
- upload factory.bin image
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
This change add IPQ40xx AP-DK04.1-C1 board image support,
enables ubi image for IPQ40xx AP-DK04.1-C1 board and also
add sysupgrage support for AP-DK04.1-C1 and generates a
sysupgrade.tar image.
Testing:
*Tested on IPQ40xx AP-DK04.1-C1:
a. NAND boot
b. ubi sysupgrade
Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
x86 board.d only contains a case for the APU2, not the APU1. This
causes, for example, network configuration not to be created correctly.
Even though the APU1 seems to reaching EOL, there a still a lot of them
out there.
The APU1 and APU2 is configured in the same way and this patch should
also be considered for stable, as the error also exists there.
Signed-off-by: Kristian Evensen <kristian.evensen@gmail.com>
Now that we always provide these partitions through DT, we don't need to
pass their data through parser data from board files anymore.
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Allow the parser to be invoked from DT without fixed cfe/linux/nvram
partitions. This allows flash to be probed from DT also for multi
flash-size images.
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
The patch 0008 removes the vanilla kernel gptu.c of lantiq.
But the replacement file timer.c does not work Amazon-SE and
is also protected therefore with an CONFIG_SOC_AMAZON_SE :-)
This commit will re-activate the default vanilla kernel code
in gptu.c for Amazon-SE.
Signed-off-by: Tino Reichardt <milky-lede@mcmilk.de>
[refresh patches and add same change for kernel 4.9]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This patch adds support for the Allnet ADSL2+ Modem ALL0333CJ.
Specifications:
- SoC: Amazon-SE
- RAM: 16MB
- Storage: 4MB NOR FLash
- LEDs: Power, Ethernet, DSL (Sync)
- Buttons: 1x Reset Button (not currently supported, will be done)
- 1x Ethernet-Port: 1x RJ45 10/100BaseTX
- 1x WAN-Port: AnnexB & J, G.992.1(ADSL), G.992.3(ADSL2), G.992.5(ADSL2+)
Installation:
- can be done via telnet+tftp or serial console
- default passwords:
- via telnet (root:admin)
- via webinterface http://172.16.1.254:8235/ (admin:coolwhite)
Installation via telnet / tftp:
host: # your own host must have an tftpd daemon, then do this:
host: ifconfig eth0:172 172.16.1.1 up
host: # telnet to allnet modem, root:admin
host: telnet 172.16.1.254
modem: # kill this daemon, it spams your console
modem: killall dsl_cpe_control
modem: # change to some place in ram:
modem: cd /ramdisk/tftp_upload
modem: # load lede-lantiq-ase-ALL0333CJ-squashfs-sysupgrade.bin alias uImage
modem: tftp -g -r uImage 172.16.1.1
modem: # load these extra tools from your tftpd via:
modem: # source: https://github.com/mcmilk/uboot-utils/releases/download/0.1/mips32.tar.gz
modem: tftp -g -r flashwrite 172.16.1.1
modem: tftp -g -r fw_setenv 172.16.1.1
modem: tftp -g -r fw.conf 172.16.1.1
modem: ln -s fw_setenv fw_printenv
modem: chmod +x *
modem: ./fw_setenv disable_recovery y
modem: ./fw_setenv kernel_addr 0xb0010000
modem: ./flashwrite /dev/mtd/1 uImage 0
Installation via serial line at uboot:
uboot: # erase everything exept bootloader
uboot: protect on b0000000 +10000; protect on b03f0000 +10000
uboot: erase all
uboot: # get new firmware via tftp:
uboot: tftpboot 0x80100000 uImage; setenv kernel_addr 0xb0010000
uboot: # copy to flash:
uboot: cp.b 0x80100000 $(kernel_addr) $(filesize)
uboot: # disable proprietary image checking:
uboot: setenv disable_recovery=y; saveenv
uboot: # reboot with LEDE ;)
uboot: reset
Ethernet works as expected, DSL syncronization does not work properly
currently, I am working on this issue.
Signed-off-by: Tino Reichardt <milky-lede@mcmilk.de>
Enable the building of Netgear DGN1000B images. I have an
example device here, but it was not tested currently.
Also WLAN and USB is not ready in the moment.
This is just the proper way, so that developing on this
device is known to be done in the near future.
Signed-off-by: Tino Reichardt <milky-lede@mcmilk.de>
Add block device sd to kernel config otherwise AHCI/eSATA devices won't get enumerated in /dev
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Explicitely disable X2APIC support on legacy targets since the targeted
processor types do not support it anyway there.
Fixes FS#285.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
mvebu was modifying RAMFS_COPY_BIN and RAMFS_COPY_DATA from a
sysupgrade_pre_upgrade hook. As the ramfs is created from stage2, this
did not have an effect anymore after the staged sysupgrade changes.
As it doesn't really hurt to copy fw_printenv and fw_setenv
unconditionally, simply add them in /lib/upgrade/platform.sh, so stage2
will see them.
Config copying is moved to a function called by platform_copy_config, where
it belongs.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Fixes: FS#821
Fixes: 30f61a34b4 "base-files: always use staged sysupgrade"
If nand chip has no NAND_NO_SUBPAGE_WRITE flag on its options
ubifs can't use it mtd devices and the kernel crashes with error:
__nand_correct_data: uncorrectable ECC error
Signed-off-by: Sergey Sergeev <adron@yapic.net>
Fixes these prints on boot:
[ 13.785600] imx-uart 2020000.serial: DMA transaction error.
[ 13.793134] imx-uart 2020000.serial: DMA transaction error.
[ 13.798721] imx-uart 2020000.serial: DMA transaction error.
Based on following upstream patch by Tim Harvey (Gateworks):
80a01b6582
TX complete DMA messages are getting missed.
This is also currently an issue in mainline.
For now we will disable DMA in serial/imx.c.
This resolves an issue encountered with RS485 transmit.
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
The "reserved" partition should probably be read-only, just in case. Even
not knowing it's content, other devices have marked it as such, so it
seems a good idea to do so also for this device.
Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
CC: Mathias Kresin <dev@kresin.me>
CC: Hanqing Wong <hquu@outlook.com>
This fixes wrong GPIO numbers for LEDs and button in Wallys DR344 board
and sets color of all LEDs to green as the mass production boards have
only green one.
Actually, DR344 has 6 GPIO-connected LEDs and one button:
- GPIO11: status
- GPIO12: sig1
- GPIO13: sig2
- GPIO14: sig3
- GPIO15: sig4
- GPIO16: reset button
- GPIO17: lan
WAN LED is connected directly with AR8035 PHY.
Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
This aligns default network interfaces configuration with vendor
firmware: GE (eth0) -> wan, FE (eth1) -> lan.
Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
GMAC0 interface of AR9344 SOC in Wallys DR344 board is connected with
AR8035, not with AR8327. Without this fix, GE interface doesn't work at
all or shows high packet loss ratio.
Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
EnGenius ENS202EXT is an outdoor wireless access point with
2-port 10/100 switch, detachable antennas and proprietery PoE.
The device is based on Qualcomm/Atheros AR9341 v1.
Specifications:
- 535/400/200 MHz (CPU/DDR/AHB)
- 64 MB of RAM
- 16 MB of FLASH
- UART (J1) header on PCB (unpopulated)
- 2x 10/100 Mbps Ethernet
- 2.4 GHz, up to 26dBm
- 2x external, detachable antennas
- 7x LED, 1x button
Flash instructions:
You have three options:
- Use the vendor firmware upgrade page on the web interface and give
it the factory.img. This is the easiest way to go about it.
- If you have serial access during u-boot, interrupt the normal boot
(any key before timeout) and run 'run failsafe_boot'; this will bring
you to a minimal openwrt luci image on ip 192.168.1.1 useful if you've
bricked the normal firmware.
- Use the vendor's management cli, which can be accessed via telnet
with the same credentials as the web login (default admin:admin), then
issue the following commands:
*** Hi admin, welcome to use cli(V-1.6.7) ***
---========= Commands Help =========---
stat -- Status
sys -- System
wless2 -- 2.4G-Wireless
mgmt -- Management
tree -- Tree
help -- Help
reboot -- Reboot
ens202ext>mgmt
Management
---========= Commands Help =========---
admin -- Administration
mvlan -- Management VLAN settings
snmp -- SNMP settings
backup -- Backup/Restore settings
autorb -- Auto reboot settings
fwgrade -- Firmware upgrade
time -- Time settings
wifisch -- Wifi schedule
log -- Log
diag -- Diagnostics
disc -- Device Discovery
logout -- Logout
help -- Help
exit -- Exit
ens202ext/mgmt>fwgrade
Management --> Firmware upgrade
---========= Commands Help =========---
fwup -- Firmware upgrade
help -- Help
exit -- Exit
ens202ext/mgmt/fwgrade>fwup http://web.server/lede-ar71xx-generic-ens202ext-squashfs-factory.bin
Signed-off-by: Marty Plummer <ntzrmtthihu777@gmail.com>
The single SFP port is meant for direct media access to WAN, such as
VDSL2, GPON). While it could also be used for in-home fiber, it is much
more likely that the LAN is standard ethernet, especially considering
that SFP is the onmly port that can go beyond 1Gbps.
Signed-off-by: Josua Mayer <josua.mayer97@gmail.com>
Take explicit note of what physical ports eth{0,1,2} refer to.
Also repair port assignment:
At some point between 4.9.20 and 4.9.29 the numbering changed.
Keep previous port assignment that was:
LAN = SFP+Switch
WAN = standalone ethernet
Also use the same assignment for Clearfog-Base to avoid confusion.
Signed-off-by: Josua Mayer <josua.mayer97@gmail.com>
Installing all armada-388-clearfog-* DTBs in the same sdcard image,
it now becomes much easier to swap sdcards between different device variants.
Signed-off-by: Josua Mayer <josua.mayer97@gmail.com>
U-Boot provides standard variables for load addresses, and
filesystem-agnostic load-commands. Furthermore thanks to distro-boot,
the device and partition from which the system boots is known.
The new boot-script makes use of all this information.
Tested on the only board that uses this boot-script: Clearfog Pro
Signed-off-by: Josua Mayer <josua.mayer97@gmail.com>
sunxi sysupgrade was based on the x86 implementation; sync fixes and other
changes from the current x86 version:
x86: fix sysupgrades on disks with 4k block size
x86: sysupgrade: move partition table change check to platform_check_image
x86: sysupgrade: refactor platform_do_upgrade
x86: sysupgrade: explicitly rescan disk after writing partition table
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
When existing partitions are retained, the dd call writing the uboot image
in the space before the first partition was accidentally writing the whole
image, making the code for individual partitions redundant. Limit the copy
to 1016KiB (the first 8KiB are skipped, and the first partition starts at
1024KiB).
In addition, conv=notrunc is replaced with conv=fsync. It seems this was an
oversight, as notrunc doesn't make sense for block devices and all other dd
commands use conv=fsync.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
This should ensure that the kernel partition can be mounted in
platform_copy_config when its size has changed.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
By returning early when no upgrade device can be found and handling the
SAVE_PARTITIONS=0 case differently, we can get rid of two levels of if.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
The staged sysupgrade will prevent us from using ask_bool in
platform_do_upgrade; therefore, the check is moved to platform_check_image.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
All targets with NAND support should gradually move their nand_do_upgrade
calls from platform_pre_upgrade to platform_do_upgrade.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
This change adds QPIC BAM dma and NAND driver node's in IPQ4019
device tree, also enable this for AP-DK04.1 based boards.
Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
This change adds support for below:
- Bam transaction which will be used for any NAND request.
- Reset function for NAND BAM transaction
- Add support for additional CSRs.
Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
The existing qcom_nand driver supports ADM DMA which is mainly
required for ipq806x family based boards,
IPQ40xx based boards uses BAM DMA in NAND driver, so this patch
adds BAM DMA support with compatible string as qcom,ebi2-nandc-bam.
Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
GPIO_PULL bits configurations in TLMM_GPIO_CFG register
differs for IPQ40xx from rest of the other qcom SoC's.
This change add support to configure the msm_gpio_pull
bits for ipq40xx, It is required to fix the proper
configurations of gpio-pull bits for nand pins mux.
Signed-off-by: Ram Chandra Jangir <rjangir@codeaurora.org>
The generic preinit code is now able to setup network and switch vlan settings
from the /etc/board.json file, therefor drop the target specific code.
Fixes FS#790.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
You can flash via tftp recovery (serve factory image as /mr6400_tp_recovery.bin
on 192.168.0.66/24, connect to any ethernet port and power on device while
holding the reset button). Flashing via OEM web interface does not work.
Hardware Specification (v1.0 EU):
- SoC: QCA9531
- Flash: Winbond W25Q64FV (8MiB)
- RAM: EtronTech EM6AB160TSE-5G (64MiB)
- Wireless: SoC platform only (2.4GHz b/g/n, 2x internal antenna)
- Ethernet: 2NIC (3x100M + 1x100M)
- WWAN: TP-LINK LTE MODULE (2x external detachable antenna)
- Power: DC 12V 1A
Signed-off-by: Filip Moc <lede@moc6.cz>
Neither the AsiaRF AWM002 or AWM003 actually has an LED on the module
board. The ld1 and ld2 do not represent actual LEDs. These pins might
connect to LEDS on an eval board or other carrier board, but that is
outside the scope of this device tree file.
Signed-off-by: Russell Senior <russell@personaltelco.net>
UML statically links against libutil, librt and libpthread. Some hosts do not
necessarily have these libraries installed and we should find out sooner than
later (during the final vmlinux linking stage) about that.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
The device tree is at /proc/device-tree/ without a base subdir.
Fixes: da472e5b30 ("treewide: access device tree from userspace via /proc/")
Signed-off-by: Mathias Kresin <dev@kresin.me>
This patch adds supports for the GL-inet GL-MT300N-V2.
Specification:
- SoC: MediaTek MT7628AN
- Flash: 16 MiB (W25Q128FVSG)
- RAM: 128 MiB DDR
- Ethernet: 1 x WAN (100 Mbps) and 1 x LAN (100 Mbps)
- USB: 1 x USB 2.0 port
- Button: 1 x switch button, 1 x reset button
- LED: 3 x LEDS (system power led is not GPIO controller)
- UART: 1 x UART on PCB (JP1: 3.3V, RX, TX, GND)
Installation through Luci:
- The original firmware is LEDE, so both LuCI or sysupgrade can be used.
- Do not keep settings, for sysupgrade please use the -n option.
Installation through bootloader webserver:
- Plug power and hold reset button until red LED blink to bright.
- Install sysupgrade image using web interface on 192.168.1.1.
Signed-off-by: Kyson Lok <kysonlok@gmail.com>
[match maximum image size with firmware partition]
Signed-off-by: Mathias Kresin <dev@kresin.me>
Access the device tree via /proc/device-tree/ is the documented way to
access the properties. Everything else might not work in future.
Signed-off-by: Mathias Kresin <dev@kresin.me>
commit 961c0eacea ('x86: fix lifting kernel CPU requirements and always
enable PAE') broke some older geode boards such as Soekris net4826.
Hence disable PAE on x86/legacy again in order to still support those
very old non-PAE capable CPUs.
Fixes FS#773 - PAE broke Soekris net4826
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This PR allow the 3G modem embedded in the DWR-512 to be managed
by the wwan-ncm scripts. The modem will use the usb-option and
usb-cdc-ether drivers.
The DWR-512 DT is updated accordingly.
Signed-off-by: Giuseppe Lippolis <giu.lippolis@gmail.com>
These drivers are in many reference-design Xeon, iCore, or
Atom64 based server boards.
Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
This change is required to make the GBit switch work on my Mikrotik Routerboard RB2011UiAS-RM, and I assume that the other RB2011 variants are exactly the same in terms of the switch. I have tested the board without and with the patch and confirm that the GBit ports are not supported at all (i.e. no communication works) with the current version in trunk and that everything works with the patch applied. The test box has been running for a few days with the patch applied, and does not show any performance problems in a test setting. I have not used it with LEDE in production so far, but with a previous turnk version of OpenWRT for many years - with the same patch applied. I therefore have good indication that it is stable.
For the record, the switch chip on my test box is identified as
switch0: Atheros AR8327 rev. 4 switch registered on ag71xx-mdio.0
The value 0x6f000000 has been taken from the table at https://wiki.openwrt.org/toh/mikrotik/rb2011uias with the previous discussion thread still online at https://lists.openwrt.org/pipermail/openwrt-devel/2014-December/029949.html.
One definite improvement from the older OpenWRT trunk version I have been running in production and current LEDE trunk is that the SFP interface can be kept in the default configuration without excessive kernel messages about it constantly going up and down. I have not yet tested an actual SFP module, though.
Performance seems to be reasonable. Routing between two GBit ports on that switch separated by different VLANs with the default firewall ruleset (and one additional rule two allow traffic between the VLANs), but without NAT, iperf3 results are:
[ ID] Interval Transfer Bandwidth Retr
[ 4] 0.00-10.00 sec 508 MBytes 426 Mbits/sec 102 sender
[ 4] 0.00-10.00 sec 506 MBytes 425 Mbits/sec receiver
With a connection going through NAT (also 2 ports on the same GBit switch, same ruleset, but NAT active), routing performance drops to around 250 MBit/s.
(Note that RouterOS achieves beyond 900 MBit/s on the same hardware with the default rule set and the FastTrack rule active even for NAT, see https://wiki.mikrotik.com/index.php?title=Manual:IP/Fasttrack and http://www.mikrotik.com/download/share/FastTrack.pdf).
Summarizing, I strongly recommend to apply this patch in trunk, so that the GBit switch chip rev. 4 can be supported upstream in the next LEDE release (hopefully soon).
Signed-off-by: René Mayrhofer <rene@mayrhofer.eu.org>
