NAPI poll() function may be passed a budget value of zero, i.e. during
netpoll, which isn't NAPI context.
Therefore, napi_consume_skb() must be given budget value instead of
!force to truly discern netpoll-like scenarios.
https://lore.kernel.org/netdev/20220707141056.2644-1-liew.s.piaw@gmail.com/t/#m470f5c20225e76fb08c44d6cfa2f1b739ffaaea4
Signed-off-by: Sieng-Piaw Liew <liew.s.piaw@gmail.com>
[improve code format]
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Use existing rx processed count to track against budget, thereby making
budget decrement operation redundant.
rx_desc_count can be calculated outside the rx loop, making the loop a
bit smaller.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
napi_build_skb() reuses NAPI skbuff_head cache in order to save some
cycles on freeing/allocating skbuff_heads on every new rx or completed
tx.
Use napi_consume_skb() to feed the cache with skbuff_heads of completed
tx so it's never empty.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
Check if we're in NAPI context when refilling rx. Normally we're almost
always running in NAPI context. Dispatch to napi_alloc_frag() directly
instead of relying on netdev_alloc_frag() which does the same but with
the overhead of local_bh_disable/enable.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
[improve code format]
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
We can increase the efficiency of rx path by using buffers to receive
packets then build SKBs around them just before passing into the network
stack. In contrast, preallocating SKBs too early reduces CPU cache
efficiency.
Performance is slightly increased but the changes allow more
potential optimizations.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
[improve code format]
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
bmips is using Broadcom B53 DSA driver which means the ethernet driver
must compensate for 6 bytes tags from the internal switch.
This means using NET_IP_ALIGN actually misaligns the skb, lowering
performance significantly. Therefore napi_alloc_skb() which uses
NET_IP_ALIGN is changed to netdev_alloc_skb().
Performance in iperf3 is increased from ~47Mbps to 52Mbps.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
Add the missing definitions for the PoE passthrough functionality.
The relevant pin is already being exported, but it is missing from
the initial board configuration file. With this change, the user is
now able to toggle the PoE passthorough functionality via the uci cli
Signed-off-by: André Fonseca <mail@andrefonseca.pt>
They are enabled by selecting devices. Fixes build errors when enabling extra
devices without creating a new config from scratch.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Wrong pcie port number for WLAN causes missing 5g WLAN interface with 5.15
kernel on Arcadyan WE420223-99 (KPN Experia WiFi).
This changes port from pcie0 to pcie1.
[1.331556] mt7621-pci 1e140000.pcie: pcie0 no card, disable it (RST & CLK)
[1.345299] mt7621-pci 1e140000.pcie: pcie2 no card, disable it (RST & CLK)
[1.359116] mt7621-pci 1e140000.pcie: PCIE1 enabled
Signed-off-by: Mikhail Zhilkin <csharper2005@gmail.com>
Specifications:
* SoC: MT7621AT
* RAM: 256MB (NT5CC64M16GP-DI)
* Flash: 16MB NOR SPI flash (GD25Q127CSIG, using GD25Q128C driver)
* WiFi: MT7615DN (2.4GHz+5Ghz) with DBDC
* Ethernet: 4x1000M LAN, 1x 1000M WAN
* LEDs: Power Blue+Orange,Wan Blue+Orange,WPS Blue,"2.4G"Blue, "5G" Blue,
USB Blue
* Buttons: Reset,WPS, Wifi
* Serial interface: on board but not populated, pinout (from the DC jack
side to the WAN port side) is "3.3V Input Output Gnd". Baud rate is 57600,
settings are 8 data bits, no parity bit, one stop bit, and no flow control.
Stock flash layout:
```
GD25Q128C(c8 40180000) (16384 Kbytes)
mtd .name = raspi, .size = 0x01000000 (16M) .erasesize = 0x00010000 (64K)
.numeraseregions = 0
Creating 7 MTD partitions on "raspi":
0x000000000000-0x000001000000 : "ALL"
0x000000000000-0x000000030000 : "Bootloader"
0x000000030000-0x000000040000 : "Config"
0x000000040000-0x000000050000 : "Factory"
0x000000050000-0x000000060000 : "Config2"
0x000000060000-0x000000fb0000 : "Kernel"
0x000000fb0000-0x000001000000 : "Private"
```
The kernel partition will be replaced with the OpenWrt image, the other
partitions are left untouched.
"Config2" seems to be the config storage used by the stock firmware.
"Private" is a 320kB empty JFFS2 partition that comes with the stock
firmware. One can get a larger space for OpenWrt by merging it with
"Kernel".
OpenWrt flash layout:
```
0x000000000000-0x000000030000 : "u-boot"
0x000000030000-0x000000040000 : "u-boot-env"
0x000000040000-0x000000050000 : "factory"
0x000000050000-0x000000060000 : "config2_stock"
0x000000060000-0x000000fb0000 : "firmware"
0x000000fb0000-0x000001000000 : "private_stock"
```
The OpenWrt image must have 96 bytes of padding in the header.
MAC addresses on OEM firmware:
| | location on the flash | notes |
|------ |----------------------- |---------- |
| lan (eth2) | factory + 0xe000 | on label |
| wan (eth3) | factory + 0xe006 | |
| 2.4g (rax0) | not on flash | lan + 1 |
| 5g (ra0) | not on flash | lan + 2 |
Mac addresses of the 2.4g and 5g interface are stored as ASCII strings in
the u-boot-env partition, but they are not used. OpenWrt calculates
Wifi Mac addresses based on the LAN Mac.
Flash and test instructions:
Flash the encrypted image (available in the OpenWrt forum) through the
stock D-Dink web interface.
1. Open the case, and solder the 4-pin header near the WAN port.
2. Connect it to a USB-UART TTL (3.3V) adapter, no need to connect VCC.
3. Open a terminal emulator (e.g. `screen /dev/ttyUSB0` on linux) with
the settings mentioned above.
4. Setup a TFTP server on your PC that can serve
`xxx-ramips-mt7621-dlink_dir-853-a1-initramfs-kernel.bin`.
5. Connect any LAN port to your PC and set a static IPv4 address to
192.168.0.101 (netmask 255.255.255.0).
6. Power on the device and keeps pressing 1 until you see the prompt.
7. Use default IP addresses and enter the file name accordingly, then hit
enter.
8. Wait until it boots to OpenWrt, the default IP address is 192.168.1.1,
you need to change your PC network adapter to use DHCP in order to access
LUCI.
9. So far, the OpenWrt runs in RAM and the flash contents are not touched.
You can try OpenWrt without having to overwrite the stock firmware, a
reboot clears all changes.
10. Optionally, backup the stock firmware (the "firmware" partition) in
Luci.
11. To permantly install OpenWrt to the device , click
on "System -> Backup/Flash Firmware" in Luci and flash
`xxx-ramips-mt7621-dlink_dir-853-a1-squashfs-sysupgrade.bin`
Known problems:
* WLAN0 defaults to 5G after a fresh installation, to enable 2.4G network,
you need to config it manually in LUCI.
* If you see jffs2 related warnings/errors after updating from the stock
web interface, you need to do a reset in LUCI. The error will be gone after
a cold reboot.
Signed-off-by: Hang Zhou <929513338qq@gmail.com>
On WXR-5950AX12, squashfs-factory.ubi is unnecessary for OpenWrt
installation and other purposes, so drop it from IMAGES and don't
generate to prevent confusion of users.
Reviewed-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
This fixes spurious boot-errors with some ath79 MIPS 74Kc boards such
as the AC Lite as well as Archer C7 v2.
The missing barrier leads to the icache flush being executed before the
dcache writeback, which results in the CPU executing the dummy infinite
loop in tlbmiss_handler_setup_pgd.
Applying this patch from upstream ensures the dcache is written back
before flushing the icache.
Signed-off-by: David Bauer <mail@david-bauer.net>
This backports the third version [1], which is awaiting upstream merge. It
adds support for watchdog max6370, which is connected via GPIO. It
is useful primarily for P2020 RDB and Turris 1.x routers, which are
not yet supported.
[1] https://www.spinics.net/lists/linux-watchdog/msg23299.html
Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>
* Fix incorrect error message in case input file opening fails
* Don't close files in case the pointers are invalid
Signed-off-by: David Bauer <mail@david-bauer.net>
This helps bring down rx CPU usage by avoiding calls to the rx handlers in
the slow path. Supports forwarding and local rx, including A-MSDU.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The padding intended to avoid corrupted non-zero padding payload was
accidentally adding too many padding bytes, tripping up some setups.
Fix this by using eth_skb_pad instead.
Fixes#11942.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
71d84bfb343e wifi: mt76: mt76x0u: report firmware version through ethtool
99d13130b517 wifi: mt76: support ww power config in dts node
09c614734880 Revert "wifi: mt76: mt7996: rely on mt76_connac2_mac_decode_he_radiotap"
e1c9c1cb50a8 mt76: mt7921: Let PCI core handle power state and use pm_sleep_ptr()
34064dbcd72a wifi: mt76: mt7921e: add pci .shutdown() support
18ccfa73a9e2 wifi: mt76: remove redundent MCU_UNI_CMD_* definitions
282845ce7f3d wifi: mt76: mt7921: fix wrong command to set STA channel
546934dacfd4 wifi: mt76: mt7921: fix PCI DMA hang after reboot
fc2ed0dfc5b0 wifi: mt76: mt7996: Remove unneeded semicolon
1b602db9c235 wifi: mt76: mt7915: unlock on error in mt7915_thermal_temp_store()
ce2438aa616a wifi: mt76: mt7996: fix radiotap bitfield
17ec2146b8f0 wifi: mt76: dynamic channel bandwidth changes in AP mode
ab2d3650a456 wifi: mt76: mt7915: expose device tree match table
90d78253498e wifi: mt76: mt7915: add dev->hif2 support for mt7916 WED device
a69c34a60451 wifi: mt76: mt7915: rework init flow in mt7915_thermal_init()
39079b5e44a7 wifi: mt76: drop the incorrect scatter and gather frame
f9ca70d6367a wifi: mt76: mt7915: add back 160MHz channel width support for MT7915
eff7666e1aa4 wifi: mt76: handle failure of vzalloc in mt7615_coredump_work
920bc6e1fc8e wifi: mt76: do not run mt76_unregister_device() on unregistered hw
b0721b96927b wifi: mt76: connac: refresh tx session timer for WED device
c32d6d849c43 wifi: mt76: usb: fix use-after-free in mt76u_free_rx_queue
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This was build tested with all core packages on all targets
successfully.
This was run tested on the following systems:
* lantiq/xrx200 musl
* x86/64 musl
* x86/64 glibc
* malta/be
The size of some of the initramfs images from lantiq/xrx increased by
240 bytes.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add the patches with real changes from the binutils 2.40 stable branch.
I am not aware that we ran into any of these problems, but I think it is
better to take the existing stable patches.
They were exported like this:
git format-patch binutils-2_40...origin/binutils-2_40-branch
I removed the patches changing the version numbers and updating the
translations only.
I removed the following patches:
*Automatic-date-updat*
001-Re-enable-development.-Update-version-to-2.40.0.patch
004-Updated-translations-for-the-gas-and-binutils-sub-di.patch
015-Updated-Swedish-translation-for-the-binutils-sub-dir.patch
027-Updated-Swedish-translation-for-the-binutils-sub-dir.patch
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Fix hwmon-gsc driver by replacing out-of-tree hwmon-gsc driver with in-tree
driver that was merged in Linux v5.8:
- remove the old out-of-tree module
- add configuration for the in-tree modules
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Several devices depend on fw_printenv during sysupgrade. Make sure
it always is present in all images, including initramfs images built
by the buildbots.
Fixes: 2449a63208 ("ramips: mt7621: Add support for ZyXEL NR7101")
Signed-off-by: Bjørn Mork <bjorn@mork.no>
If a packet has reached its intended destination, it was bumped to the code
that accepts it, without first checking if a mesh_path needs to be created
based on the discovered source.
Fix this by moving the destination address check further down
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Add both ext4 and f2fs support for overlayfs. The fstools mount_root
application will choose f2fs if the overlay volume space available
exceeds 100MB, otherwise ext4 is used.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
The Gateworks Newport boards supported by the octeontx target have
the following on-board devices:
- Gateworks System Controller
- GPIO buttons
- GPIO leds
- GPS PPS
- Accelerometer
- MCP251X CAN controller
Add kernel drivers for these devices in DEFAULT_PACKAGES
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Forward-port from ar71xx target the board introduced in commit
eb9e3651dd (" ar71xx: add support for the MikroTik RB911-2Hn/5Hn
boards"). Citing:
The patch adds support for the MikroTik RB911-2Hn (911 Lite2)
and the RB911-5Hn (911 Lite5) boards:
https://mikrotik.com/product/RB911-2Hnhttps://mikrotik.com/product/RB911-5Hn
The two boards are using the same hardware design, the only difference
between the two is the supported wireless band.
Specifications:
* SoC: Atheros AR9344 (600MHz)
* RAM: 64MiB
* Storage: 16 MiB SPI NOR flash
* Ethernet: 1x100M (Passive PoE in)
* Wireless: AR9344 built-in wireless MAC, single chain
802.11b/g/n (911-2Hn) or 802.11a/g/n (911-5Hn)
Notes:
* Older versions of these boards might be equipped with a NAND
flash chip instead of the SPI NOR device. Those boards are not
supported (yet).[1]
* The MikroTik RB911-5HnD (911 Lite5 Dual) board also uses the
same hardware. Support for that can be added later with little
effort probably.[2]
End of citation.
Follow intallation instruction from that commit message, using
openwrt-ath79-mikrotik-mikrotik_routerboard-911-lite-initramfs-kernel.bin
and
openwrt-ath79-mikrotik-mikrotik_routerboard-911-lite-squashfs-sysupgrade.bin
images found in ath79/mikrotik directory. Be advised that the board
accepts 10-30 V on PoE input.
Known issues
Compared to ar71xx target image, there is still small leak of current to
user LED, which makes it lit, although weaker, even if brightness is set
to 0. The cause of that is still unknown.
1. https://github.com/openwrt/openwrt/pull/3652
2. RB911-5HnD should work with this commit or with [1], depending on
what flash topology was used.
Signed-off-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Most of boards from MikroTik with AR9344 SoC (supported and
un-supported) replicate the same schematic, so stack common device nodes
to a single dtsi.
ar9344_mikrotik_routerboard-16m-nor.dtsi:
- remove include paragraph and wmac node, make it single nor flash node
for others dts to include
ar9344_mikrotik_routerboard-lhg-5nd.dts:
- move all of the nodes to new file ar9344_mikrotik_routerboard.dtsi
and leave only power, user and lan LEDs which differ from sxt-5nd-r2
and other yet unsupported devices
ar9344_mikrotik_routerboard-sxt-5n.dtsi:
- remove, it made no sense to keep it, as only
ar9344_mikrotik_routerboard-sxt-5nd-r2.dts included this file and
added only compatible and model
ar9344_mikrotik_routerboard-sxt-5nd-r2.dts:
- include ar9344_mikrotik_routerboard.dtsi
- add nand gpio activating node, beeper, additional LEDs and flash chips
which previously have been in ar9344_mikrotik_routerboard-sxt-5n.dtsi
ar9344_mikrotik_routerboard.dtsi:
- inherited most of the content from ar9344_mikrotik_routerboard-lhg-5nd.dts
except three LEDs
- add wmac node, removed from ar9344_mikrotik_routerboard-16m-nor.dtsi
Signed-off-by: Tomasz Maciej Nowak <tmn505@gmail.com>
