Signed-off-by: Steev Klimaszewski <steev@kali.org>

Signed-off-by: Steev Klimaszewski <steev@kali.org>

On platforms using the Qualcomm UEFI Secure Application (uefisecapp),
EFI variables cannot be accessed via the standard interface in EFI
runtime mode. The respective functions return EFI_UNSUPPORTED. On these
platforms, we instead need to talk to uefisecapp. This commit provides
support for this and registers the respective efivars operations to
access EFI variables from the kernel.

Communication with uefisecapp follows the standard Qualcomm Trusted
Environment (TEE or TrEE) / Secure OS conventions via the respective SCM
call interface. This is also the reason why variable access works
nomrally while boot services are active. During this time, said SCM
interface is managed by the boot services. When calling
ExitBootServices(), the ownership is transferred to the kernel.
Therefore, UEFI must not use that interface itself (as multiple partys
accessing this interface at the same time may lead to complications) and
cannot access variables for us.

Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>

Add support for SCM calls to Secure OS and the Trusted Execution
Environment (TEE/TrEE) residing in the TrustZone (TZ). This allows
communication with Secure/TZ applications, for example 'uefisecapp'
managing access to UEFI variables.

While this patch introduces only a very basic interface without the more
advanced features (such as reentrant and blocking SCM calls and
listeners/callbacks), this is enough to talk to the aforemetioned
'uefisecapp'.

Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>

Add an eDP panel entry for IVO M133NW4J-R3.

Due to lack of documentation, use the delay_200_500_e50 timings for now.

Signed-off-by: Steev Klimaszewski <steev@kali.org>

Make qcom_scm_call and qcom_scm_call_atomic and associated types
accessible to other modules.

Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>

Add support for the alternate (second-source) touchpad found on some
X13s laptops.

The alternate touchpad interrupt line is active low and the ACPI table
declares that it should have its internal pull-up enabled.

Note that the pin configuration must currently be handled by the i2c-bus
node even though only one of these nodes will ever be successfully
probed on a single device.

FIXME: verify
FIXME: revisit "shared" pin config

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

When booting, we see the message

[    0.997647] cpu cpu0: failed to update OPP for freq=300000

So, lets add in 300MHz to make it happy

Signed-off-by: Steev Klimaszewski <steev@kali.org>

The removal of the AR9170_MAC_SNIFFER_ENABLE_PROMISC flag to fix an issue
many years ago caused the AR9170 to not be able to pass probe response
packets with different MAC addresses back up to the driver. In general
operation, this doesn't matter, but in the case of packet injection with
aireplay-ng it is important. aireplay-ng specifically injects packets with
spoofed MAC addresses on the probe requests and looks for probe responses
back to those addresses. No other combination of filter flags seem to fix
this issue and so AR9170_MAC_SNIFFER_ENABLE is required to get these packets.

This was originally caused by commit e0509d3b which
removed this flag in order to avoid spurious ack noise from the hardware.
In testing for this issue, keeping this flag but not restoring the
AR9170_MAC_RX_CTRL_ACK_IN_SNIFFER flag on the rc_ctrl seems to solve this
issue, at least with the most current firmware v1.9.9.

Signed-off-by: Steve deRosier <derosier@cal-sierra.com>

Add backlight to the mdss0_dp3/aux-bus/panel node so that the backlight
will actually turn off when it is supposed to.

Signed-off-by: Steev Klimaszewski <steev@kali.org>

Fix up the DP PHY node which had the wrong unit address.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Fix the usb_0 HS PHY reference clock which was mistakingly replaced with
the first usb_2 PHY clock.

Fixes: 152d1faf ("arm64: dts: qcom: add SC8280XP platform")
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Commit 6c846d02 ("gpio: Don't fiddle with irqchips marked as
immutable") added a warning to indicate if the gpiolib is altering the
internals of irqchips.

Following this change the following warning is now observed for the SPMI
PMIC pinctrl driver:
gpio gpiochip1: (200f000.spmi:pmic@0:gpio@c000): not an immutable chip, please consider fixing it!

Fix this by making the irqchip in the SPMI PMIC pinctrl driver immutable.

Signed-off-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>

Actual fix is in debian-installer...

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>

Copied from Johan Hovold's tree

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>

Add the misc-3p3 regulator as supply for the touchscreen, touchpad and
keyboard.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Use descriptive names for the input device nodes.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Fix the misc-3p3 regulator enable gpio which was off by one and
prevented the regulator from being probed.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Enable the PCIe2 interface accessible through an M.2 connector.

FIXME: add SSD regulator

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Enable the Qualcomm QCNFA765 Wireless Network Adapter connected to
PCIe4.

FIXME: power sequencing

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Force the ath11k driver to load a board file for a different machine
that's already in linux-firmware.

Not-Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Enable the SDX55 modem connected to PCIe3a.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

The SC8280XP reference design has an on-PCB SDX55 modem which uses
MBIM.

Reuse the Foxconn SDX55 configuration for now.

FIXME: revisit config

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Rename the backlight and misc regulators according to the net names.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Enable both PCIe2 controllers in x2 mode to be able to test the
corresponding nodes.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

The SA8295P-ADP has up to four PCIe interfaces implemented by three or
four controllers: PCIe2A, PCIe3A/PCIe3B and PCIe4.

PCIe2 is used in x4 mode, while PCIe3 can be used in either x2 or x4
mode. Enable both PCIe3A and PCI3B in x2 mode for now.

FIXME: Configure PCIe2 PHY for x4

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

The SC8280XP platform has seven PCIe controllers:

	PCIe0	USB4
	PCIe1	USB4
	PCIe2A	4-lane
	PCIe2B	2-lane
	PCIe3A	4-lane
	PCIe3B	2-lane
	PCIe4	1-lane

while SA8540P only has five (PCIe2-4).

Add devicetree nodes for the PCIe2-4 controllers and their PHYs.

FIXME: modem compatible
FIXME: 2b and 3b register addresses

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Use the new phy-mux clock implementation for the PCIe pipe clock muxes
so that the pipe clock source is set to the QMP PHY PLL when the
downstream pipe clock is enabled and restored to the always-on XO when
it is again disabled.

This is needed to prevent the corresponding GDSC from hanging when
enabling or disabling the PCIe power domain, something which requires a
ticking source.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Use the new optional devres helper to get the pipediv2 clock.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Add devres helper to retrieve an optional clock from a child node.

Note that of_clk_get_by_name() returns -EINVAL when the optional clock
is not found.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Set bit 3 of the QSERDES_V5_COM_BIAS_EN_CLKBUFLR_EN register to allow
four lanes to be used.

FIXME: use separate compatible

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

FIXME: address FIXMEs, drop comments
FIXME: modem compatible name
FIXME: revisit variant abstraction

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Some QMP PHYs have a second fixed-divider pipe clock that needs to be
enabled along with the pipe clock.

Add support for an optional "pipediv2" clock.

Note that there's no point in adding a "lane" suffix to the clock name
as there will only ever be one pipe clock per PHY (the same is true for
the "pipe0" clock which could be renamed "pipe").

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Add two v5 PCIe power-state config defines based on the corresponding v4
defines and vendor kernel USB3_5NM_UNI_PCS_PCIE defines that are needed
for SC8280XP.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Use lower case hex consistently for define values.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Clean up the QMP defines by removing some stray white space and making
sure values are aligned.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Clean up the QMP V4 and V5 defines by moving a few entries that were out
of order.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>

Sort the device-id table entries alphabetically by compatible string to
make it easier to find entries and add new ones.

Signed-off-by: Johan Hovold <johan+linaro@kernel.org>