This adds reboot support to the rk808 pmic driver and enables it for
the rk809 and rk817 devices.
This only enables if the rockchip,system-power-controller flag is set.

Signed-off-by: Peter Geis <pgwipeout@gmail.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Link: https://lore.kernel.org/r/20220208194023.929720-1-pgwipeout@gmail.com

RK817 has a power-off bit in SYS_CFG3. Add support for powering
off the PMIC.

Signed-off-by: Ondrej Jirman <megous@megous.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

Add rk817 codec support cell to rk808 mfd driver.

Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
Tested-by: Maciej Matuszczyk <maccraft123mc@gmail.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

Constify a number of static struct resource. The only usage of the
structs are to assign their address to the resources field in the
mfd_cell struct. This allows the compiler to put them in read-only
memory. Done with the help of Coccinelle.

Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

RK805 has the same kind of dual-role sleep/shutdown pin as RK809/RK817,
so it makes little sense for the driver to have to have two completely
different mechanisms to handle essentially the same thing. Move RK805
over to the shutdown/suspend flow to clean things up.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

Rather than having 3 almost-identical functions plus the machinery to
keep track of them, it's far simpler to just dynamically select the
appropriate register field per variant.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

Setting the SLEEP pin to its shutdown function for appropriate PMICs
doesn't need to happen in single-CPU context, so there's really no point
involving the syscore machinery. Hook it up to the standard driver model
shutdown method instead. This also obviates the issue that the syscore
ops weren't being unregistered on probe failure or module removal.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

The RK809/RK817 suspend/resume hooks should not have to depend on
whether this driver owns the pm_power_off hook, and thus the global
rk808_i2c_client is set - indeed, the GPIO-based control is really
only relevant when PSCI firmware is in charge of power rather than
the kernel. As driver model callbacks, they have an appropriate
device argument to hand, so can just always use that.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

With the device tree property "rockchip,system-power-controller" we
explicitly request to use this PMIC to power off the system. So always
register our poweroff function, even if some other handler (probably
PSCI poweroff) was registered before.

This does tend to reveal a warning on shutdown due to the Rockchip I2C
driver not implementing an atomic transfer method, however since the
write to DEV_OFF takes effect immediately the I2C completion interrupt
is moot anyway, and as the very last thing written to the console it is
only visible to users going out of their way to capture serial output.

Signed-off-by: Soeren Moch <smoch@web.de>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
[ rm: note potential warning in commit message ]
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

Do not open code the definition, instead use the nice DEFINE_RES_IRQ
macro for it.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Lee Jones <lee.jones@linaro.org>