It looks like boot hangs on an AC-Mesh for unknown reasons. The last
message seen on the console is:
[ 0.000000] Inode-cache hash table entries: 8192 (order: 3, 32768 bytes, linear)
But interestingly, it seems like enabling AIO somehow works around this
problem. Changing any off the following options seem to have the same
effect at the moment for Linux 5.10.160+5.10.161
# CONFIG_KERNEL_AIO is not set
# CONFIG_KERNEL_CGROUPS is not set
# CONFIG_KERNEL_FANOTIFY is not set
# CONFIG_KERNEL_FHANDLE is not set
# CONFIG_KERNEL_IO_URING is not set
# CONFIG_KERNEL_IPV6_MROUTE is not set
# CONFIG_KERNEL_IPV6_SEG6_LWTUNNEL is not set
# CONFIG_KERNEL_IP_MROUTE is not set
CONFIG_KERNEL_PROC_STRIPPED=y
Just enable CONFIG_AIO until the actual problem was fixed.
Link: https://github.com/freifunk-gluon/gluon/issues/2784
We migrate to dnsmasq-full, while disabling most of its features.
Notably, dhcp and dnssec are compiled in, while other features of the
full variant are deselected.
The main difference between the non-EFI and EFI images generated by
OpenWrt is that the former uses an MS-DOS partition table, while the
latter uses GPT. The EFI images still have a BIOS-compatible MBR, so
they work fine on non-EFI systems.
Closes#2403
fixes OOM reboots due too limited ram with ath10k 5Ghz enabled
add some comments to describe the need for ath10k-ct replacement
tested stable on an TP-Link Archer C25v1
more details
694757a08f
AVM Fritz!Box 7520 and Fritz!Box 7530 use the same hardware platform and can
only be distinguished by using the urlader partition or the fritz-tffs tools
and read the ProductID (Fritz_Box_HW247).
Upstream added a standalone SPI kernel-loader which fixes the unbootable
image for the WDR4900. Thus, we can re-introduce this device to Gluon.
Signed-off-by: David Bauer <mail@david-bauer.net>
Specification:
SoC: MediaTek MT7628AN
RAM: 64MiB
Flash: 8MiB
Wifi:
- 2.4GHz: MT7628AN
- 5GHz: MT7612EN
LAN: 1x 10/100 Mbps
Flash instructions:
Flash factory image through stock firmware WEB UI.
Back to stock is possible by using TFTP and stripping down the Firmware
provided by TP-Link to a initramfs.
The flash space between 0x650000 and 0x7f0000
is blank in the stock firmware so I left it out as well.
Note: Buffalo has introduced hardware changes without bumping the
revision number. 19.07 did not support the rb-variant so there's no need
to implement a migration for the rb-variant.
Every g300nh supported by Gluon should either be the s-variant or
been flashed wrongly.
Gone due to
commit 45c84a117b ("ar71xx: drop target")
