Add 5.4 kernel version as a new testing kernel option.
Ref: https://github.com/openwrt/openwrt/pull/2793
Tested-by: Hannu Nyman <hannu.nyman@iki.fi> [ipq8065, R7800]
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [ipq8065, NBG6817]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
[added Tested-by tags]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Due to changes in syscon driver, the phy dwc3 driver
needs to use device_node_to_regmap since it has to skip
the new introduced clk check. This fix broken usb3 on this
target.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
As mdio0 is used in every dts move it to general ipq8064
dts and use label to set device specific definition.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Currently the watchdog timer is broken as it tries to
get an interrupt to setup pretimeout. Since our platform
have a different type of interrupt disable it and use
legacy watchdog probe.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
This was created by Chunkeey some time ago. Since mdio driver
works or doesn't work and since this was tested by me for 1 year,
include it to remove the use of the generic bitbang gpio driver for
switch driver.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Rework tsens driver.
Since in the new kernel 5.4 init common do more than it
should, inizialize the kernel memory directly in the driver and
drop use of this function. Rework all the patch with the new
variable names.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Rework l2 scaling patch to fix some compile warning
and to imporve the caling timings by removing call to unnecessary
function.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Since this dtsi now have wrong definition in the upstream version,
include it to overwrite and remove any problem.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Refactor l2 freq scaling patch to support voltage
scaling and add support to base cache scaling on
cpu freq scaling. Update the dtsi files with the new
definition used in the new code.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
It's wrong set the mux to bias-disable. The best way to
do this is by creating a separate group and disable the
specific pins.
By documentation, any subgroup with no bias definition
is ignored so the mux definition is useless.
Rework the definition by sremoving the mux subgroup and
set the remaining subgroup with the mux function and
drive-strength
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
The opp_notifier introduced with patch 0054-Handle-OPP-voltage adjust
for some reason missed the actual registration in the opp struct,
resulting in never being registred.
This was present in kernel 4.9 patchset but dropped in the transition
to 4.14. Reintroduce this and fix patch 0055 about L2 cache scaling.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
This patch uses the SOC variable to calculate DTS names automatically
based on the SOC and the device definition node name.
This reduces redundancy and (by having to choose DTS name
appropriately) will unify the naming of a device in different places
(image/Makefile, DTS name, compatible, image name). This is supposed
to make life easier for developers and reviewers.
Since the kernel uses a "soc-device.dts" scheme for this target, we
use this for the derivation of DEVICE_DTS, too, leaving the DTS names
unchanged for this target.
Note that for some devices the kernel itself uses inconsistent names
(DTS naming scheme vs. compatible), leaving us with a manual overwrite
of DEVICE_DTS for those cases.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
We have only 11 sensors on ipq806x. Fix the reg property
to load the right amount of data instead of the entire
space.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Tsense driver for ipq806x have various problem.
- Emit wrong error. On probing of this driver, nvmem driver can be
not ready and this cause a EDEFER error. Actually this is not an
error as the kernel will retry to probe the driver after the
nvmem driver is loaded.
- Use uninitialized value on trigger of critical temp
- Doesn't free allocated memory
Because of this, rework the driver and improve it by removing extra
load of data.
Change the logic of loading data. Use the backup calib data only
when the calib data is not present. As the calibration is only
needed to set the temp offset, we don't really need to read
both calib data and set the offset based only on the backup one.
Also change how the notifier function work. At times when we
output the trigger message, we already have read the temp so
remove the extra read and the wrong uninitialized data that
probably caused a kernel panic for null pointer exception.
(Think we never experience this bug because the router
never reached that temp ever... So just lucky)
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
The config name for cpufreq driver has changed.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
The new driver use opp table to register frequency.
Drop psv bindings as they are not used anymore.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
The new driver use opp table to register frequency.
Drop psv bindings as they are not used anymore.
Adds speedbin definition for nvmem driver
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
This patch has been proposed but never actually merged to
mainline. It was accepted but never re proposed by the
creator.
Rework it, fix kernel panic cause by double kfree.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Backport patch applied to qcom-cpufreq-kryo
driver as krait cpu will base on this driver.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Drop old cpufreq as now we have new driver that
can use normal kernel opp definition
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Rework 0052-PM-OPP-Update-the-voltage-tolerance-when-adjusting-t
to reflect changes upstream.
- Skip unnecessary allocation of buffer to set u_volt
- Change opp u_volt directly
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Update 0049-PM-OPP-Support-adjusting-OPP-voltages-at-runtime with
upstream version.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
It has been notice a buf in L2 cache scaling where the scaling is not
done proprely if the frequency is set to the initial state before
the new frequency.
From: https://patchwork.kernel.org/patch/10565443/
* The clocks are set to aux clock rate first to make sure the
* secondary mux is not sourcing off of QSB. The rate is then set to
* two different rates to force a HFPLL reinit under all
* circumstances.
In the initial stage of boot to force a new frequency to apply, is
needed to first set the frequency back to the lowest one (aux_rate)
and then to the target one. This force and make sure the controller
actually switch the frequency to the right one. Apply the same
mechanism to L2 frequency scaling. Before scaling to the target
frequency, first set the frequency to the aux_rate to force the
transition, then scale it to the target frequency. Doing the wrong way
can produce unexpected results and could lock the scaling mechanism
until a full reboot is done (Causing a full reset by the krait-cc driver)
From: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=77612720a2362230af726baa4149c40ec7a7fb05
When the Hfplls are reprogrammed during the rate change,
the primary muxes which are sourced from the same hfpll
for higher frequencies, needs to be switched to the 'safe
secondary mux' as the parent for that small window. This
is done by registering a clk notifier for the muxes and
switching to the safe parent in the PRE_RATE_CHANGE notifier
and back to the original parent in the POST_RATE_CHANGE notifier.
This should apply also to L2 scaling... as we can't relly use
the notifier, we manually do this on L2 scaling.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [nbg6817/ipq8065]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Deactivate CONFIG_SFP for kernel 4.19 in the generic configuration.
The CONFIG_SFP configuration option was not set to anything in the
ath79 build for me, set it to deactivated by default.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
It has been used by several people for some time already and feedback
has been mostly positive.
Ref: https://github.com/openwrt/openwrt/pull/2472
Tested-by: Hannu Nyman <hannu.nyman@iki.fi> [ipq8065, R7800]
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [ipq8065, NBG6817]
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
[separate commit, commit subject and description facelift]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
- Use new dwc3-qcom usb driver.
- Drop dwc3-of-simple as we have a dedicated driver now.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
[split into separate commit, commit subject facelift]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Voltage tolerance is accounted per core, not per cpu, so add
missing DT entry.
Signed-off-by: Pavel Kubelun <be.dissent@gmail.com>
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Added patch:
- 063-3 Fix tsense shared memory problem
Recent changes in ioremap_resource function are causing fails for the
memory areas which are already mapped. This changes are causing tsense
driver failures during initialization:
qcom-tsens 900000.thermal-sensor: tsens init failed
So this patch uses simple ioremap in order to use this shared memory
space.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
[commit subject and desciption facelitf]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
