Tony Ambardar [Fri, 14 Aug 2020 05:11:22 +0000 (22:11 -0700)]
malta: update MIPS64 ISA to R2
Usage of current R1 ISA is inconsistent with the MIPS32 subtarget, little
used and has limited utility for testing.
Many distros target a minimum R2 ISA. Debian MIPS 32-bit/64-bit ports all
use MIPS R2 ISA since Stretch, for example. Fedora's MIPS arch also targets
the R2 ISA for 32-bit/64-bit.
Widely used MIPS64 platforms like Octeon are based on the MIPS R2 ISA or
later, and benefit from having a compatible test platform in OpenWRT.
While Linux does support MIPS64 R1 targets, its usefulness for development
and testing is limited. As an example, the modern Linux eBPF JIT requires
a MIPS R2 ISA or later.
dnsmasq: support tftp_unique_root in /etc/config/dhcp
The TFTP server provided by dnsmasq supports serving a select boot image
based on the client's MAC or IP address. This allows an administrator
to activate this feature in /etc/config/dhcp. Here is an example
/etc/config/dhcp that configures dnsmasq with --tftp-unique-root=mac:
With this configuration, dnsmasq will serve
/usr/libexec/tftpboot/00-11-22-33-44-55/openwrt-initramfs-kernel.bin to
the client with MAC address 00:11:22:33:44:55.
Hans Dedecker [Mon, 14 Sep 2020 19:55:01 +0000 (21:55 +0200)]
vxlan: add extra config options
Add config options:
srcportmin/srcportmax : range of port numbers to use as UDP source ports
to communicate to the remote VXLAN tunnel endpoint
ageing : lifetime in seconds of FDB entries learnt by the kernel
maxaddress : maximum number of FDB entries
learning : enable/disable entering unknown source link layer addresses
and IP addresses into the VXLAN device FDB.
rsc : enable/disable route short circuit
proxy : enable/disable ARP proxy
l2miss : enable/disable netlink LLADDR miss notifications
l3miss : enable/disable netlink IP ADDR miss notifications
gbp : enable/disable the Group Policy extension
This might fail due to the A/B nature of this device. When flashing, OEM
firmware writes over the non-booted partition. If booted from 'A',
flashing over 'B' won't work. To get around this, you should flash the
OEM image over itself. This will then boot the router from 'B' and
allow you to flash OpenWRT without problems.
Reverting to factory firmware:
Hard-reset the router three times to force it to boot from 'B.' This is
where the stock firmware resides. To remove any traces of OpenWRT from
your router simply flash the OEM image at this point.
Signed-off-by: J. Scott Heppler <shep971@centurylink.net>
David Bauer [Sun, 20 Sep 2020 21:49:35 +0000 (23:49 +0200)]
hostapd: ubus: make (B)SSID optional for neighbor report
Make the BSSID and SSID fields optional when configuring a neighbor
report into hostapd.
Both options can now be an empty string. For the BSSID, the first 6 byte
are copied from the neighbor report. For the SSID, the SSID for the
affected hostapd BSS is used.
David Bauer [Tue, 22 Sep 2020 13:25:57 +0000 (15:25 +0200)]
hostapd: ubus: send notification instead of event
Rafal Milecki pointed out that ubus events are meant for low-level ubus
events only (e.g. addition or removal of an object). Higher level
events should happen as notifications on the ubus object itself.
Dispatch BSS events on the main hostapd ubus object instead of
publishing them as ubus events.
bluetooth on mt7622 needs a firmware to start. It can't be built-in or
it tries to load firmware before rootfs is mounted, and then fails.
build it as a kernel module to fix that.
David Bauer [Thu, 17 Sep 2020 23:11:13 +0000 (01:11 +0200)]
hostapd: send ubus event on BSS update
hostapd will emit a ubus event with the eventname hostapd.<ifname>.<event>
when adding, removing or reloading a BSS.
This way, services which install state (for example the RMM neighbor
list) can on-demand reinstall this information for the BSS without
polling this state.
UCI defaults scripts are supposed to be numbered, but odhcpd's lacked numbering, which
turned out to mess up my custom scripts numbered 9[0-9]_*. The idea is to have high number
(custom) scripts executed last. Jow confirmed numbering is the default case, not the
exception (thanks).
Daniel Golle [Mon, 21 Sep 2020 17:12:24 +0000 (18:12 +0100)]
policycoreutils: break into smaller packages
Instead of vaguely describing dependencies in the package description
actually split-up into individual packages, each with their
dependencies expressed accurately.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Daniel Golle [Mon, 21 Sep 2020 12:47:57 +0000 (13:47 +0100)]
libselinux: split utility packages and add PKG_LICENSE
Split utility packages similar to coreutils in packages feed, adding
ALTERNATIVES for those which are also provided by busybox-selinux.
Also add missing license information.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Felix Fietkau [Mon, 21 Sep 2020 15:56:04 +0000 (17:56 +0200)]
mac80211: fix regression in station connection monitor optimization
When the nulldata frame was acked, the probe send count needs to be reset,
otherwise it will keep increasing until the connection is considered dead,
even though it fine.
Reported-by: Georgi Valkov <gvalkov@abv.bg> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Rui Salvaterra [Fri, 14 Aug 2020 12:21:22 +0000 (13:21 +0100)]
zram-swap: explicitly use mkswap/swapon/swapoff from /sbin
The required BusyBox applets are enabled by default, so we can rely on them
being present in the system. This way, we make sure there are no conflicts
with less featured variants of these same applets which might also be
present in the system.
mac80211: select the first available channel for 5GHz interfaces
Some 5GHz wifi interfaces, especially in Tri-band routers, can't use
channel 36. In these cases, the default configuration for 5GHz
interfaces, once enabled, doesn't work.
This patch selects the first non-disabled channel for 5GHz interfaces.
David Woodhouse [Tue, 8 Sep 2020 08:26:34 +0000 (09:26 +0100)]
mediatek: mt7623: use bash for generating bootable images
It turns out that 'echo -e' isn't portable; it doesn't work in the dash
builtin echo and Ubuntu users are complaining.
I can't even get octal (specified by POSIX) to work consistently because
those variants of 'echo' which *do* support -e don't seem to interpret
octalwithout it.
I could switch to /bin/echo but using -e with that isn't actually
portable *either* even though it works today.
For now just stick with bash, and use its builtin. We may end up using
something else entirely; perhaps perl.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
David Woodhouse [Tue, 8 Sep 2020 08:26:32 +0000 (09:26 +0100)]
mediatek: mt7623n-preloader: add preloader for Banana Pi R64
We want to be able to make full system images for this system too, just
as we now can for the MT7623 platforms.
The package directory (mt7623n) is now a bit misnamed as it's overly
specific, but the precise set of platforms which we support this way
is evolving and we'll fix it up when the dust settles and we know what
nomenclature makes most sense.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
This fixes several stylistic and functional errors of the recently
added Edgecore ECW5410:
- fix call in 11-ath10k-caldata
- use hex notation in 11-ath10k-caldata
- remove redundant definitions from DTS that are already in DTSI
- use proper sorting in image/Makefile
- use DEVICE_VENDOR/DEVICE_MODEL instead of DEVICE_TITLE
- use SOC instead of DEVICE_DTS
This bug applied where mtd partition end address,
or erase start address, was not cleanly divisible by parent mtd erasesize.
This error would cause the bits following the end of the partition
to the next erasesize block boundary to be erased,
and this partition-overflow data to be written to the partition erase
address (missing additional partition offset address)
of the mtd (top) parent device.
Fixes: FS#2428 Signed-off-by: John Thomson <git@johnthomson.fastmail.com.au> Tested-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
[shorten commit title, add Fixes, fix kernel 4.19 as well] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Andre Heider [Sat, 2 Nov 2019 12:14:08 +0000 (13:14 +0100)]
arm-trusted-firmware-mvebu: add support for espressobin
Use build variants to cover all 11 hardware options [0]:
espressobin-512mb
espressobin-v3-v5-1gb-1cs
espressobin-v3-v5-1gb-1cs-emmc
espressobin-v3-v5-1gb-2cs
espressobin-v3-v5-1gb-2cs-emmc
espressobin-v3-v5-2gb
espressobin-v3-v5-2gb-emmc
espressobin-v7-1gb
espressobin-v7-1gb-emmc
espressobin-v7-2gb
espressobin-v7-2gb-emmc
CLOCKSPRESET is set to CPU_800_DDR_800 for all builds, which is the only
stable configuration. That actually matches what Globalscale shipped as
CPU_1000_DDR_800 combined with kernel versions < v4.19.42. [1][2].
Andre Heider [Sun, 30 Aug 2020 04:52:59 +0000 (06:52 +0200)]
arm-trusted-firmware-mvebu: update a3700-utils
Update to current head of the branch A3700_utils-armada-18.12-fixed: 0967979 ddr: Add DDR3 2CS layout for EspressoBin v5 2GB board 486523e ddr: fix typo for ESPRESSObin 2GB layout 490b2b3 TBB: Fix building for Crypto++ 6.0 and later 0141dd1 TBB: Split INCDIR from LIBDIR
Tony Ambardar [Sun, 30 Aug 2020 22:54:51 +0000 (15:54 -0700)]
bpftools: support NLS, fix ppc build and update to 5.8.9
With global NLS support enabled (CONFIG_BUILD_NLS), the linked libelf.so
and libbfd.so libraries will depend on libintl.so. Import the nls.mk helper
to set library prefixes and flags accordingly, and also conditionally add
"-lintl" as link-time library.
Fix a build error on ppc due to a EDEADLOCK redefinition in errno.h.
Use upstream stable kernel 5.8.9, and fix overriding of feature detection
to only allow/hide detected features. Also refresh existing patches.
* New upstream microcode datafile 20200616
+ Downgraded microcodes (to a previously shipped revision):
sig 0x000406e3, pf_mask 0xc0, 2019-10-03, rev 0x00d6, size 101376
sig 0x000506e3, pf_mask 0x36, 2019-10-03, rev 0x00d6, size 101376
* Works around hangs on boot on Skylake-U/Y and Skylake Xeon E3,
* This update *removes* the SRBDS mitigations from the above processors
* Note that Debian had already downgraded 0x406e3 in release 3.20200609.2
* REGRESSION FIX: 0x406e3: rollback to rev 0xd6 and document regression
* Microcode rollbacks (closes: LP#1883002)
sig 0x000406e3, pf_mask 0xc0, 2019-10-03, rev 0x00d6, size 101376
* THIS REMOVES THE SECURITY FIXES FOR SKYLAKE-U/Y PROCESSORS
* Avoid hangs on boot on (some?) Skylake-U/Y processors,
* ucode-blacklist: blacklist models 0x8e and 0x9e from late-loading,
just in case. Note that Debian does not do late loading by itself.
Refer to LP#1883002 for the report, 0x806ec hangs upon late load.
* SECURITY UPDATE
* For most processors: SRBDS and/or VRDS, L1DCES mitigations depending
on the processor model
* For Skylake HEDT and Skylake Xeons with signature 0x50654: VRDS and
L1DCES mitigations, plus mitigations described in the changelog entry
for package release 3.20191112.1.
* Expect some performance impact, the mitigations are enabled by
default. A Linux kernel update will be issued that allows one to
selectively disable the mitigations.
* New upstream microcode datafile 20200609
* Implements mitigation for CVE-2020-0543 Special Register Buffer Data
Sampling (SRBDS), INTEL-SA-00320, CROSSTalk
* Implements mitigation for CVE-2020-0548 Vector Register Data Sampling
(VRDS), INTEL-SA-00329
* Implements mitigation for CVE-2020-0549 L1D Cache Eviction Sampling
(L1DCES), INTEL-SA-00329
* Known to fix the regression introduced in release 2019-11-12 (sig
0x50564, rev. 0x2000065), which would cause several systems with
Skylake Xeon, Skylake HEDT processors to hang while rebooting
* Updated Microcodes:
sig 0x000306c3, pf_mask 0x32, 2019-11-12, rev 0x0028, size 23552
sig 0x000306d4, pf_mask 0xc0, 2019-11-12, rev 0x002f, size 19456
sig 0x00040651, pf_mask 0x72, 2019-11-12, rev 0x0026, size 22528
sig 0x00040661, pf_mask 0x32, 2019-11-12, rev 0x001c, size 25600
sig 0x00040671, pf_mask 0x22, 2019-11-12, rev 0x0022, size 14336
sig 0x000406e3, pf_mask 0xc0, 2020-04-27, rev 0x00dc, size 104448
sig 0x00050653, pf_mask 0x97, 2020-04-24, rev 0x1000157, size 32768
sig 0x00050654, pf_mask 0xb7, 2020-04-24, rev 0x2006906, size 34816
sig 0x00050656, pf_mask 0xbf, 2020-04-23, rev 0x4002f01, size 52224
sig 0x00050657, pf_mask 0xbf, 2020-04-23, rev 0x5002f01, size 52224
sig 0x000506e3, pf_mask 0x36, 2020-04-27, rev 0x00dc, size 104448
sig 0x000806e9, pf_mask 0x10, 2020-04-27, rev 0x00d6, size 103424
sig 0x000806e9, pf_mask 0xc0, 2020-04-27, rev 0x00d6, size 103424
sig 0x000806ea, pf_mask 0xc0, 2020-04-27, rev 0x00d6, size 103424
sig 0x000806eb, pf_mask 0xd0, 2020-04-27, rev 0x00d6, size 103424
sig 0x000806ec, pf_mask 0x94, 2020-04-23, rev 0x00d6, size 103424
sig 0x000906e9, pf_mask 0x2a, 2020-04-23, rev 0x00d6, size 103424
sig 0x000906ea, pf_mask 0x22, 2020-04-27, rev 0x00d6, size 102400
sig 0x000906eb, pf_mask 0x02, 2020-04-23, rev 0x00d6, size 103424
sig 0x000906ec, pf_mask 0x22, 2020-04-27, rev 0x00d6, size 102400
sig 0x000906ed, pf_mask 0x22, 2020-04-23, rev 0x00d6, size 103424
* Restores the microcode-level fixes that were reverted by release
3.20191115.2 for sig 0x50654 (Skylake Xeon, Skylake HEDT)
* New microcode update packages from AMD upstream:
+ Removed Microcode updates (known to cause issues):
sig 0x00830f10, patch id 0x08301025, 2019-07-11
* README: update for new release
* New microcode update packages from AMD upstream:
+ New Microcodes:
sig 0x00830f10, patch id 0x08301025, 2019-07-11
+ Updated Microcodes:
sig 0x00800f12, patch id 0x08001250, 2019-04-16
sig 0x00800f82, patch id 0x0800820d, 2019-04-16
kernel: drop kmod-bmp085, kmod-bmp085-i2c and kmod-bmp085-spi
CONFIG_BMP085* is replaced by CONFIG_BMP280 since 4.9[1] and this package is empty.
OpenWRT also has kmod-iio-bmp280* package and we can drop old packages.
The patch (001-dont-build-tests-man.patch) no longer works with the current version.
Follow the patch notes:
Adding the --enable-static and --disable-tests parameters should replace the patch.
Felix Fietkau [Fri, 18 Sep 2020 15:28:30 +0000 (17:28 +0200)]
mt76: update to the latest version
3b946a6dc588 mt76: dma: cache dma map address/len in struct mt76_queue_entry c4c8b6a20d3b mt76: mt7915: fix HE BSS info 15391c1c947f mt76: fix tx hang on non-AQL frame limit 72c8a81e64e8 mt76: mt7915: fix encap offload multicast traffic with 4-address mode 69b3f868d14b mt76: mt7915: use napi_consume_skb to bulk-free tx skbs 5f080033ec7d mt76: move txwi handling code to dma.c, since it is mmio specific b1f425686125 mt76: mt7915: fix VHT LDPC capability 8f48855f5d14 mt76: mt7915: simplify mt7915_lmac_mapping cfaf40858718 mt76: mt7915: fix queue/tid mapping for airtime reporting 115b62efac21 mt76: remove retry_q from struct mt76_txq and related code e22c65cdc585 mt76: mt7915: simplify checks for the 802.3 offload path bab866a01e4f mt76: mt7915: fix unexpected firmware mode 0fc3c5eb61d0 mt76: dma: queue more rx frames internally before passing them to the stack 130e5de09364 Revert "mt76: dma: queue more rx frames internally before passing them to the stack" e3af31409d41 update mt7915 firmware to the latest version e2b8a4ec9891 mt76: testmode: add a limit for queued tx_frames packets 146488631f7b mt76: mt7615: Remove set but unused variable 'index' 0b7d2b76288e mt76: mt7615: fix VHT LDPC capability 848f4a6334a8 mt76: mt7622: fix fw hang on mt7622 0a955d944bd0 mt76: mt7663s: do not use altx for ctl/mgmt traffic 13b96411513b mt76: mt7663s: split mt7663s_tx_update_sched in mt7663s_tx_{pick,update}_quota d62ba15b1bbf mt76: mt7663s: introduce __mt7663s_xmit_queue routine fdf14d1b6aec mt76: move pad estimation out of mt76_skb_adjust_pad d048f8e87ba0 mt76: mt7663s: fix possible quota leak in mt7663s_refill_sched_quota 979c0fdc5d27 mt76: mt7663s: introduce sdio tx aggregation 56e77a3a3ade mt76: mt7663: check isr read return value in mt7663s_rx_work f96cffa03e57 mt76: mt7615: unlock dfs bands 1ccd31bbe1f4 mt76: Use fallthrough pseudo-keyword 448cd2d36ee2 mt76: mt76x0: Move tables used only by init.c to their own header file 17ba3432f5af Revert "mt76: mt7615: unlock dfs bands" fee1f4a8e87f mt76: mt7915: fix possible memory leak in mt7915_mcu_add_beacon 5b78e5292777 mt76: Fix unsigned expressions compared with zero ec84891a4d23 mt76: mt7915: convert to use le16_add_cpu()
ath79/mikrotik: fix soft_config location for SXT 5n
The soft_config partition for these devices lays between 0xe000 and
0xf000 (as correctly detected by the RouterBoard platform driver),
before the bootloader2 partition which starts at 0x10000.
This commit correctly sorts the partitions, fixing the parsing error.
Martin Schiller [Tue, 15 Sep 2020 06:24:22 +0000 (08:24 +0200)]
lantiq: use uniform "u-boot-env" mtd label
This is the most popular choice in the linux kernel tree.
Within OpenWrt, this change will establish consistency with ath79
and ramips targets.
Signed-off-by: Martin Schiller <ms@dev.tdt.de>
[extend commit message, include netgear_dm200, update base-files] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Vladimir Vid [Mon, 14 Sep 2020 15:24:24 +0000 (17:24 +0200)]
mvebu: image: add check for fdt_add_r and kernel_addr_r variables
fdt_addr and kernel_addr variables are getting obsolete in the mainline
u-boot in favor of fdt_addr_r and kernel_addr_r.
By checking if the new variables exist, we can make sure that devices with newer
version of u-boot will work while not breaking support for the existing ones.
Signed-off-by: Vladimir Vid <vladimir.vid@sartura.hr> Acked-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
David Bauer [Tue, 15 Sep 2020 09:19:14 +0000 (11:19 +0200)]
ipq40xx: enable FRITZRepeater 3000 ports on switch
The ethernet ports on the AVM FRITZRepeater 3000 are not separated
between LAN and WAN in the stock firmware. OpenWrt currently abstracts
port 4 as eth0 and port 5 as eth1, bridging them in the kernel.
This patch adjusts the GMAC port bitmasks and default bitmask for ar40xx
to bridge them on the switch, avoiding traffic on both ports to pass
thru the CPU.
Follow the Ubiquiti TFTP recovery procedure for this device.
1. Hold down the reset button while connecting power for 10 seconds.
2. Transfer the factory image via TFTP to the AP (192.168.1.20)
3. Wait 2 minutes for the AP to write the firmware to flash. The device
will automatically reboot to OpenWrt.
zram-swap: default to lzo instead of lzo-rle compression
On devices with small amounts of RAM, zram-swap fails to initialise due to the
default compression algorithm (lzo-rle). Startup example on an AirGrid M2, with
32 MiB of RAM:
root@airgrid:/etc/config# /etc/init.d/zram start
zram_start: activating '/dev/zram0' for swapping (13 MegaBytes)
zram_reset: enforcing defaults via /sys/block/zram0/reset
sh: write error: Out of memory
mkswap: image is too small
swapon: /dev/zram0: Invalid argument
root@airgrid:/etc/config#
Fix this by defaulting to traditional lzo, which works fine and is always
available.
John Crispin [Mon, 7 Sep 2020 11:08:41 +0000 (13:08 +0200)]
ipq40xx: add Edgecore OAP-100 support
flashing the unit
* first update to latest edcore FW as per the PDF instructions
* boot the initramfs
- tftpboot 0x88000000 openwrt-ipq40xx-generic-edgecore_oap100-initramfs-fit-uImage.itb; bootm
* inside the initramfs call the following commiands
- ubiattach -p /dev/mtd0
- ubirmvol /dev/ubi0 -n0
- ubirmvol /dev/ubi0 -n1
- ubirmvol /dev/ubi0 -n2
* scp the sysupgrade image to the board and call
- sysupgrade -n openwrt-ipq40xx-generic-edgecore_oap100-squashfs-nand-sysupgrade.bin
Robert Marko [Mon, 7 Sep 2020 10:43:37 +0000 (12:43 +0200)]
ipq806x: add Edgecore ECW5410 support
This patch adds support for the Edgecore ECW5410 indoor AP.
Specification:
- SoC: Qualcomm Atheros IPQ8068 ARMv7 2x Cortex A-15
- RAM: 256MB(225 usable) DDR3
- NOR Flash: 16MB SPI NOR
- NAND Flash: 128MB S34MS01G2 Parallel NAND
- Ethernet: 2 x 1G via 2x AR8033 PHY-s connected directly to GMAC2 and GMAC3 via SGMII (802.3af POE IN on eth0)
- USB: 1 x USB 3.0 SuperSpeed
- WLAN: 2x QCA9994 AC Wawe 2 (1x 2GHz bgn, 1x 5GHz acn)
- CC2540 BLE
- UART console on RJ45 next to ethernet ports exposed.
Its Cisco pin compatible, 115200 8n1 baud.
Installation instructions:
Through stock firmware or initramfs.
1.Connect to console
2. Login with root account, if password is unknown then interrupt the boot with f and reset it in failsafe.
3. Transfer factory image
4. Flash the image with ubiformat /dev/mtd1 -y -f <your factory image path>
This will replace the rootfs2 with OpenWrt, if you are currently running from rootfs2 then simply change /dev/mtd1 to /dev/mtd0
Note
Initramfs:
1. Connect to console
2. Transfer the image from TFTP server with tftpboot,
or by using DHCP advertised image with dhcp command.
3. bootm
4. Run ubiformat /dev/mtd1
You need to interrupt the bootloader after rebooting and run:
run altbootcmd
This will switch your active rootfs partition to one you wrote to and boot from it.
So if rootfs1 is active, then it will change it to rootfs2.
This will format the rootfs2 partition, if your active partition is 2 then simply change /dev/mtd1 with /dev/mtd0
If you dont format the partition you will be writing too, then sysupgrade will find existing UBI rootfs and kernel volumes and update those.
This will result in wrong ordering and OpenWrt will panic on boot.
5. Transfer sysupgrade image
6. Flash with sysupgrade -n.
Note that sysupgrade will write the image to rootfs partition that is not currently in use.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Robert Marko [Mon, 18 May 2020 10:35:15 +0000 (12:35 +0200)]
ipq806x: import bootargs-append from IPQ40xx
This imports the patch that adds bootargs-append support from IPQ40xx.
This way we can append additional boot arguments from DTS instead of only being able to overwrite them.
This way dual firmware devices can use the rootfs number that bootloader passes to decide from what to boot.
But we still need to append console info and ubi root info.
This is used by Edgecore ECW5410.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Robert Marko [Mon, 18 May 2020 10:34:06 +0000 (12:34 +0200)]
firmware: ipq-wifi: enable use on IPQ806x
This enables the ipq-wifi package to be used on IPQ806x target.
Its needed for boards using a different BDF than one shipped in the upstream board-2.bin.
Currently needed for Edgecore ECW5410.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
ath79: fix button types for WZR-HP-AG300H and WZR-600DHP
Router and Movie "keys" are actually switches for both devices
according to the manual. This has been properly implemented in ar71xx,
but overlooked when porting to ath79.
Fixes: 480bf2827392 ("ath79: add support for Buffalo WZR-HP-AG300H") Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Installation of OpenWRT from vendor firmware:
- Connect to the Web-interface at http://192.168.11.1
- Go to “Administration” → “Firmware Upgrade”
- Upload the OpenWrt factory image
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
ath79: consistently use "info" label for default-mac partition
The tp-link safeloader devices typically contain a partition
"default-mac" which stores the MAC addresses. It is followed by other
partitions containing device info, like
In DTS, we typically assign a 0x10000 sized partition for these,
which is mostly labelled "mac" or "info". In rarer cases, the
partitions have been enclosed in a larger "tplink" or "config"
partition.
However, when comparing different devices, the implementation appears
relatively arbitrary at the moment.
Thus, this PR aims at harmonizing these partitions by always using
the name "info" for the DTS partition containing "default-mac".
"info" is preferred over "mac" as we never just have "default-mac"
alone, but always some other device-info partitions as well.
While at it, this also establishes a similar partitioning for the
few devices where the "info" partitions are part of a bigger
unspecific "config" partition or similar.
Besides the harmonization itself, this also allows to merge a few
cases in 11-ath10k-caldata.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
vconfig has been disabled by default since 2015 [1] and there are
no remaining uses in entire OpenWrt trunk. However, we still set up
a specific name_type for it during boot.
While this setup is properly implemented to be only triggered when
vconfig is present, it still seems anachronistic and unnecessary
to set up a standard for a tool that is not used anymore.
Therefore, this removes the set_name_type initialization and leaves
it for those people actually using the tool to configure it as needed.
The reverted shell code is a very poor reimplementation of the existing
package-metadata.pl usergroup subcommand and the resulting file is not
used anymore, so drop this code.
Use the prepared .packageusergroup file to lookup user and group names
when processing the passed file mode.
Also replace the various subshell/cut invocations with a sequence of
standard variable interpolations which fixes paths with embedded colons
as a side-effect.
Paul Spooren [Mon, 14 Sep 2020 02:02:02 +0000 (16:02 -1000)]
build: add user/group ID resolve function
With the introduction of `./tmp/userids` the `ipkg-build` script can now
resolve values of "PKG_FILE_MODES", allowing users to set names rather
than numeric values.
Paul Spooren [Mon, 14 Sep 2020 01:55:52 +0000 (15:55 -1000)]
build: create tmp/userids file
Multiple packages contain a USERID variable defining required user and
group for the package to run. With the recent addition of
"PKG_FILE_MODES" it is possible to define user and group of specific
files, replacing (possibly insecure) post-inst scripts. These modes are
set during build time and put directly into the packages.
To allow user and group names rather than the numeric values, a mapping
like `/etc/passwd` is required by the `ipkg-build` script, mapping names
defined in "PKG_FILE_MODES" to a numeric value, as the build system does
not create any users during build.
This commit adds a single line to the `prepare-tmpinfo` target, so that
everytime the feeds are updated the *passwd like* content of
`./tmp/userids` is updated.
This adds support for the RTL838x Architecture.
SoCs of this type are used in managed and un-managed Switches and Routers
with 8-28 ports. Drivers are provided for SoC initialization, GPIOs, Flash,
Ethernet including a DSA switch driver and internal and external PHYs used
with these switches.
Supported SoCs:
RTL8380M
RTL8381M
RTL8382M
The kernel will also boot on the following RTL839x SoCs, however driver
support apart from spi-nor is missing:
RTL8390
RTL8391
RTL8393
The following PHYs are supported:
RTL8214FC (Quad QSGMII multiplexing GMAC and SFP port)
RTL8218B internal: internal PHY of the RTL838x chips
RTL8318b external (QSGMII 8-port GMAC phy)
RTL8382M SerDes for 2 SFP ports
Initialization sequences for the PHYs are provided in the form of
firmware files.
Flash driver supports 3 / 4 byte access
DSA switch driver supports VLANs, port isolation, STP and port mirroring.
The ALLNET ALL-SG8208M is supported as Proof of Concept:
RTL8382M SoC
1 MIPS 4KEc core @ 500MHz
8 Internal PHYs (RTL8218B)
128MB DRAM (Nanya NT5TU128MB)
16MB NOR Flash (MXIC 25L128)
8 GBEthernet ports with one green status LED each (SoC controlled)
1 Power LED (not configurable)
1 SYS LED (configurable)
1 On-Off switch (not configurable)
1 Reset button at the right behind right air-vent (not configurable)
1 Reset button on front panel (configurable)
12V 1A barrel connector
1 serial header with populated standard pin connector and with markings
GND TX RX Vcc(3.3V), connection properties: 115200 8N1
To install, upload the sysupgrade image to the OEM webpage.
Klaus Kudielka [Fri, 27 Dec 2019 18:15:31 +0000 (19:15 +0100)]
uboot-envtools: mvebu: update uci defaults for Turris Omnia
On the Turris Omnia 2019, u-boot environment is located at 0xF0000, instead
of 0xC0000. The switch happened with u-boot-omnia package version 2019-04-2
(May 10, 2019).
Check the installed u-boot release, and set the default accordingly.
Signed-off-by: Klaus Kudielka <klaus.kudielka@gmail.com>
[bump PKG_RELEASE, use lower case for hex offset] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
ramips: mt7621: pbr-m1: add pcie reset for asm1061
this board has a pcie to sata bridge connected to pcie2 with a
separated pcie reset on gpio7.
add reset-gpios and corresponding pinctrl nodes into dts.
Felix Fietkau [Sun, 13 Sep 2020 07:34:37 +0000 (09:34 +0200)]
ramips/mediatek: improve GRO performance, fix PPE packet parsing
Backport upstream changes to initialize GDM settings and reset PPE
Allow GMAC to recognize the special tag to fix PPE packet parsing
Improve GRO performance by passing PPE L4 hash as skb hash
TP-Link EAP225-Wall v2 is an AC1200 (802.11ac Wave-2) wall plate access
point. UART access and debricking require fine soldering.
The device was kindly provided for porting by Stijn Segers.
Device specifications:
* SoC: QCA9561 @ 775MHz
* RAM: 128MiB DDR2
* Flash: 16MiB SPI-NOR (GD25Q127CSIG)
* Wireless 2.4GHz (SoC): b/g/n, 2x2
* Wireless 5Ghz (QCA9886): a/n/ac, 2x2 MU-MIMO
* Ethernet (SoC): 4× 100Mbps
* Eth0 (back): 802.3af/at PoE in
* Eth1, Eth2 (bottom)
* Eth3 (bottom): PoE out (can be toggled by GPIO)
* One status LED
* Two buttons (both work as failsafe)
* LED button, implemented as KEY_BRIGHTNESS_TOGGLE
* Reset button
Flashing instructions, requires recent firmware (tested on 1.20.0):
* ssh into target device and run `cliclientd stopcs`
* Upgrade with factory image via web interface
Debricking:
* Serial port can be soldered on PCB J4 (1: TXD, 2: RXD, 3: GND, 4: VCC)
* Bridge unpopulated resistors R162 (TXD) and R165 (RXD)
Do NOT bridge R164
* Use 3.3V, 115200 baud, 8n1
* Interrupt bootloader by holding CTRL+B during boot
* tftp initramfs to flash via sysupgrade or LuCI web interface
MAC addresses:
MAC address (as on device label) is stored in device info partition at
an offset of 8 bytes. ath9k device has same address as ethernet, ath10k
uses address incremented by 1.
From OEM ifconfig:
br0 Link encap:Ethernet HWaddr 50:...:04
eth0 Link encap:Ethernet HWaddr 50:...:04
wifi0 Link encap:UNSPEC HWaddr 50-...-04-...
wifi1 Link encap:UNSPEC HWaddr 50-...-05-...
Signed-off-by: Sander Vanheule <sander@svanheule.net>
[fix IMAGE_SIZE] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Signed-off-by: John Audia <graysky@archlinux.us>
[add community build/run tests to commit message] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
projects / thirdparty / openwrt.git / log
