prevent the creation of a fake NUMA node
saving a few CPU cycles each DRAM access:
[ 0.000000] NUMA: No NUMA configuration found
[ 0.000000] NUMA: Faking a node at ...
[ 0.000000] NUMA: NODE_DATA [mem 0xff9c2100-0xff9c3fff]
if you boot on QEMU make sure to review vCPU topology:
set socket to 1
This is needed to get ethernet working on most 64-bit sunxi boards.
Basically just synced with armv7 config a bit.
[ci:skip-build] already built successfully in CI
This patch is obsolete now and even breaks the detection
of headphones. Headsets are properly detected at all times.
Dropping this, allows ALSA to see the 'Headphone Jack' event
when headphones are plugged in and removed. This way, ALSA UCM
can act on these jack events as well.
[ci:skip-build] Already built on CI in a MR
Next release of ModemManager needs Qualcomm RMNET support
for QMAP mobile data support.
armv7 config doesn't have ethernet enabled at all,
so this is only enabled for the aarch64 config.
I rebased all of the patches (only 2 needed some changes, the rest I
just did for good measure).
These two patches were dropped since they are already merged in the new
kernel:
0008-leds-gpio-Set-max-brightness-to-1.patch
0009-dts-pinephone-Add-pine64-pinephone-to-compat-list.patch was dropped
The kconfig changes are just after running kconfig edit, and making no
changes so that it would generate a config with any new options in 5.12,
to make maintenance easier moving forward.
The i2c bus for the pogo pins was locked up because the level shifter is
powered by the regulator used for the cameras. This enables that
specific regular for the gpios the i2c bus is on so the pull-ups work
when reading and writing from the pogo pins.
[ci:skip-build] Already built in CI
Wireguard's userspace tool 'wg-quick' requires certain network and
netfilter options in the kernel config to be enabled, since it uses
policy routing and some 'iptables' modules. An overview of the options
and why they are required can be found in pmaports#909.
NOTE: after rebasing this patch on the one that enables anbox related
kconfig options, only CONFIG_IP_MULTIPLE_TABLES=y was left.
wireguard needs the ipv6 multitable option to make a seperate routing
table for the inside and outside vpn part of the network.
[ci:skip-build]: already built successfully in CI
Update ov5640 autofocus related patch. We had it working with Martijn's
patch in 5.9.1_git20201019-r2 (MR 1654), but then dropped the patch in
favor of the patch that was added to megi's tree in 5.9.3_git20201101
(MR 1678).
It turns out that the patch added to megi's tree was different than the
patch we were carrying, it was a WIP version where loading the autofocus
firmware did not work anymore.
Megi decided to drop the autofocus patch from his tree for now, as the
firmware uploader takes more time in Martijn's proper patch. According
to Martijn it takes about a second, which seems reasonable for the
postmarketOS use case. I've added the previous version again as
out-of-tree patch.
Martijn found that for some reason the continuous autofocus still
doesn't want to start with this kernel and current patch version, as for
some reason the firmware load fails. But he extended the patch to make
autofocus work again with "tap to focus" at least.
0005-dts-pinephone-Disable-flash-led-in-OV5640-node.patch is not needed
anymore, the same change is in megi's tree.
Co-Authored-By: Martijn Braam <martijn@brixit.nl>
[ci:skip-build]: already built successfully in CI
This patches from the megi solves the issue of the MIPI-DSI panels
framerate being at 2/3rd of actual or expected rate.
[ci:skip-build]: already built successfully in CI
Switch from the 5.8.x kernel with allwinner specific patches to megi's
5.9 based tree. He drives the PinePhone related kernel development
forward like no other, his tree and changes are well documented on
xnux.eu, and he frequently publishes tested "build this one" commits with
the latest improvements. It makes a faster and less time consuming
workflow to simply package these versions (after a smoke test) into
postmarketOS edge, and after some time to catch possible regressions
ship the same version in postmarketOS stable.
This kernel should make HDMI work for most people, and improve call
quality (thanks to smaeul's patches, which are integrated into megi's
tree).
Kernel config changes:
* Enable MODEM_POWER in the kernel config, among other improvements this
allows us to get rid of the 30s delay on power off.
* Enable CONFIG_DRM_PANEL_ILITEK_ILI9881C (new PineTab panel)
* Enable CONFIG_ZRAM
* Enable CONFIG_HID_MULTITOUCH
Related: https://xnux.eu/devices/feature/modem-pp.html#toc-modem-power-driver
Changelog: https://megous.com/git/linux/tag/?h=orange-pi-5.9-20201019-1553
Co-Authored-By: Martijn Braam <martijn@brixit.nl>
Co-Authored-By: Clayton Craft <clayton@craftyguy.net>
Enable the freezer cgroup. This allows a user to start a large application like
firefox, freeze it to save some power, and unfreeze it to continue using it,
without having to restart the application completely. Stopping and restarting
the application completely results in more battery usage as well as lost state.
Signed-off-by: Serge Hallyn <serge@hallyn.com>
Sloppy patch that makes the displayport alt-mode output signal work on
the PinePhone convergence edition. Proper fix is in the 5.8 tree
[ci:skip-build]: Martijn made sure that it builds and works. This is
time critical, so skip the CI build.
Enable CONFIG_CRYPTO_XTS for each kernel, so we can switch to using
aes-xts-plain64 as default cipher for cryptsetup (override with
"pmbootstrap --cipher"), instead of aes-cbc-plain64 (pmbootstrap#1940).
I have executed "pmbootstrap kconfig edit" on each kernel, and manually
toggled the option. The diff is not always clean, because for some
kernels it is apparently the first time, that menuconfig was executed on
the configs like that. In a few instances, it turned out that
CONFIG_ANDROID_PARANOID_NETWORK needed to be disabled too (this is
already a requirement, but as the config was incomplete, it was not
visible that this option was enabled). Very few times, I had to enable
CONFIG_EXPERIMENTAL in order to see and enable CONFIG_CRYPTO_XTS.
It would be great if we could automate such mass kconfig edits in the
future, see pmbootstrap#1942.
[skip ci]: I have verified, that every single one of these kernels builds.
CI will likely run out of time while downloading source tarballs.
