As Grygorii Strashko pointed out, the runtime PM use count of the
children can be whatever at suspend and we should not use it. So
let's just suspend ti-sysc at noirq level and get rid of some code.

Let's also remove the PM_SLEEP ifdef and use __maybe_unused as the
PM code already deals with the ifdefs.
Suggested-by: NGrygorii Strashko <grygorii.strashko@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

We're currently warning about busy children on suspend in
sysc_child_suspend_noirq() but the legacy code omap_device does
not do that. Let's just make it dev_dbg() instead of dev_warn().
Signed-off-by: NTony Lindgren <tony@atomide.com>

We want to see the names of detected devices when DEBUG is enabled.
Signed-off-by: NTony Lindgren <tony@atomide.com>

We need to detect timer and gpio on dra7 because of the
SYSC_QUIRK_LEGACY_IDLE flag for suspend and resume.
Signed-off-by: NTony Lindgren <tony@atomide.com>

When debug is enabled, we want to see what devices we're detecting
to make things a bit easier for us.
Signed-off-by: NTony Lindgren <tony@atomide.com>

When debug is enabled, we want to see what devices we're detecting
to make things a bit easier for us. Many of these devices will also
be available on am335x and dra7, and some just need updating the
revision register mask.
Signed-off-by: NTony Lindgren <tony@atomide.com>

We need to detect few new devices to tag for noirq_suspend and
pm_runtime_irq_safe to avoid causing regressions compared to
legacy platform data booting.

Let's update i2c, gpio, uart and wdt revision masks to detect
them on am437x. Note that we can remove the second wdt entry
with the updated mask. Note that we also have some uarts with
a different revision register.
Signed-off-by: NTony Lindgren <tony@atomide.com>

We don't care when we suspend but some our children do. In order to
avoid tagging various modules with SYSC_QUIRK_RESOURCE_PROVIDER, let's
do it automatically by tagging modules that are busy on suspend for
noirq suspend. This way we can just do module detection on define DEBUG.

Note that we still need to keep SYSC_QUIRK_LEGACY_IDLE flag around so
the our legacy single-child devices that set pm_runtime_irq_safe() can
manage the interconnect target module themselves.
Signed-off-by: NTony Lindgren <tony@atomide.com>

If no_console_suspend is set, we should keep console enabled during suspend.
Lets fix this by only producing a warning if we can't idle hardware during
suspend.

Fixes: ef55f821 ("bus: ti-sysc: Improve suspend and resume handling")
Signed-off-by: NTony Lindgren <tony@atomide.com>

We can have the interconnect target module control registers pretty
much anywhere within the module range. The current code attempts an
incomplete optimization of the ioremap size but does it wrong and
it only works for registers at the beginning of the module.

Let's just use the largest control register to calculate the ioremap
size. The ioremapped range is for most part cached anyways so there
is no need for size optimization. Let's also update the comments
accordingly.

Fixes: 0eecc636 ("bus: ti-sysc: Add minimal TI sysc interconnect
target driver")
Signed-off-by: NTony Lindgren <tony@atomide.com>

Add support for the software reset of a target interconnect
module using its sysconfig and sysstatus registers.
Signed-off-by: NFaiz Abbas <faiz_abbas@ti.com>
[tony@atomide.com: updated to check if sysconfig exists]
Signed-off-by: NTony Lindgren <tony@atomide.com>

The dra76x MCAN generic interconnect module has a its own
format for the bits in the control registers.

Therefore add a new module type, new regbits and new capabilities
specific to the MCAN module.
Acked-by: NRob Herring <robh@kernel.org>
CC: Tony Lindgren <tony@atomide.com>
Signed-off-by: NFaiz Abbas <faiz_abbas@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

This adjusts the allocator calls to use 2-factor argument call style, as
done treewide already for improved defense against allocation overflows.
Signed-off-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NTony Lindgren <tony@atomide.com>

We should be checking ddata->clocks[i] instead of clock_names[i]
for the optional clocks. Currently this just happens to work for
the typical case of one fck and one optional clock.

Fixes: 09dfe581 ("bus: ti-sysc: Add handling for clkctrl opt clocks")
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Let's show module info if DEBUG is enabled to make it easier to follow
what happens on the suspend and resume path.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Looks like these two device drivers don't yet behave properly for suspend
unless configured with the legacy option.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Starting with omap4, UARTs have different revision register that we need to
detect to enable SYSC_QUIRK_LEGACY_IDLE. Otherwise UARTs won't idle properly.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Starting with omap4 some timers have different sysc registers (type2) compared
to the omap2 timers (type1). We need to detect these to enable the quirk for
SYSC_QUIRK_LEGACY_IDLE, otherwise these won't be idling properly.
Siganed-off-by: NTony Lindgren <tony@atomide.com>

Some modules need to use external resets in the rstctrl bits. Typically
only one of the rstctrl bits is for the interconnect target module while
the others are for various child devices.

For ti-sysc driver, we just need the module rstctrl bit mapped. The rest
of the rstctrl bits can be directly mapped to the child devices.

Cc: Suman Anna <s-anna@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Based on testing with more devices I noticed that some devices
don't suspend or resume properly. We need to PM runtime suspend
and resume devices if we have ddata->needs_resume set.

Let's also improve the error handling and add few debug statements
to make it easier to notice suspend and resume related issues if
DEBUG is set.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Modules that provide resources for other modules need to be suspended
and resumed in the noirq calls. Tag the resource providing modules.
Signed-off-by: NTony Lindgren <tony@atomide.com>

There can be up to eight optional device functional gate gate clocks for
each clkctrl instance in clkctrl register bits 8 to 15. Some of them are
only needed for module level reset while others may always be needed
during use. Let's add support for those and update the binding doc
accordingly.

Note that the optional clkctrl mux and divider clocks starting at bit 20
can be directly mapped to the child devices, and ti-sysc does not need to
manage those.

And as GPIOs need the optional clocks for reset, we can now add it with
SYSC_QUIRK_OPT_CLKS_IN_RESET.

Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Tero Kristo <t-kristo@ti.com>
Cc: devicetree@vger.kernel.org
Reviewed-by: NRob Herring <robh@kernel.org>
Signed-off-by: NTony Lindgren <tony@atomide.com>

In order to prepare supporting clkctrl optional clocks, we need to
make the current child clock handling more generic so we can use the
clock role names for the optional clocks in the following patch.

Cc: Tero Kristo <t-kristo@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Otherwise child devices that some interconnect target module devices
have won't probe using simple-bus.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Now that ti-sysc can manage child devices, we must also be backwards
compatible with the current omap_device code. With omap_device, we
assume that the child device manages the interconnect target module
directly.

The drivers needing special handling are the ones that still set
pm_runtime_irq_safe(). In the long run we want to update those drivers
as otherwise they will cause problems with genpd as a permanent PM
runtime usage count is set on the parent device.

We can handle omap_device these devices by improving the ti-sysc quirk
handling to detect the devices needing special handling based on
register map and revision register if usable. We also need to implement
dev_pm_domain for these child devices just like omap_device does.
Signed-off-by: NTony Lindgren <tony@atomide.com>

We want to pass the device tree configuration for interconnect target
modules from ti-sysc driver to the existing platform hwmod code.

This allows us to first validate the dts data against the existing
platform data before we start dropping the platform data in favor of
device tree data.

To do this, let's add platform data callbacks for PM runtime functions
to call for the interconnect target modules if platform data is
available.

Note that as ti-sysc driver can rebind, omap_auxdata_lookup and related
functions can no longer be __init.
Signed-off-by: NTony Lindgren <tony@atomide.com>

We already show the status for the interconnect target module
when debugging is enabled, there's no need to be more verbose
about that. So let's just cut down the noise and remove the
extra debug statements.
Signed-off-by: NTony Lindgren <tony@atomide.coM>

At least earlycon needs a delayed idle before the 8250 driver probes
to avoid glitches in the console output. Let's handle the delayed idle
for devices tagged with ti,no-reset-on-init and ti,no-idle-on-init with
delayed_work. Others don't need it, and there should be no need to use
runtime PM autosuspend for the interconnect target driver as it's the
child device drivers that should configure it.
Signed-off-by: NTony Lindgren <tony@atomide.com>

If we have stdout-path specified for earlycon, we must prevent
the debug console from idling until runtime PM kicks in.
Signed-off-by: NTony Lindgren <tony@atomide.com>

This allows us to idle the module on suspend after the children
are suspended.
Signed-off-by: NTony Lindgren <tony@atomide.com>

The functional clock is used by several child device drivers to query
the rate for the child device internal configuration. The functional
clock is really for the whole interconnect target module, and not just
for the child device, and can also be shared across multiple children.
At least the timers, i2c and mmc driver query the fck for rate.

So let's just create a clock alias for the child fck if it does not
yet exits. We can do this with the BUS_NOTIFY_ADD_DEVICE before the
child is probed.

Note that we need to now also remove the legacy mode check for getting
the dts clocks in ti-sysc driver.
Signed-off-by: NTony Lindgren <tony@atomide.com>

We are currently only checking for the first entry in the table while
we should check them all. Usual no-idle-on-init is together with
no-reset-on-init, so this has gone unnoticed.

Fixes: 566a9b05 ("bus: ti-sysc: Handle module quirks based dts
configuration")
Signed-off-by: NTony Lindgren <tony@atomide.com>

The enawakeup bit is in a different location for smartreflex compared
to the "ti,sysc-omap2" compatible.

Fixes: 70a65240 ("bus: ti-sysc: Add register bits for interconnect
target modules")
Signed-off-by: NTony Lindgren <tony@atomide.com>

We need to configure the interconnect target module based on the
device three configuration.

Let's also add a new quirk for SYSC_QUIRK_RESET_STATUS to indicate
that the SYSCONFIG reset bit changes after the reset is done.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Let's configure few module quirks via device tree using the
properties for "ti,no-idle-on-init", "ti,no-reset-on-init"
and "ti,sysc-delay-us".

Let's also reorder the probe a bit so we have pdata available
earlier, and move the PM runtime calls to sysc_init_module()
from sysc_read_revision().
Signed-off-by: NTony Lindgren <tony@atomide.com>

We can easily detect i2c based on it's non-standard module registers that
consist of two 32-bit registers accessed in 16-bit mode.

So far we don't have other 16-bit modules, so there's currently no need
to add a custom property for 16-bit register access.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Let's add data for the known interconnect target module types by mapping
their register bits.

Note that we can handle many quirks for the older omap2 type1 modules
directly in the driver without a need for adding custom properties.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Looks like we're missing remove() that's needed if a driver instance
rebound. Otherwise we will get "Unbalanced pm_runtime_enable!".
Signed-off-by: NTony Lindgren <tony@atomide.com>

The new bus driver causes a harmless compile-time warning when
CONFIG_PM is disabled:

drivers/bus/ti-sysc.c:440:12: error: 'sysc_runtime_resume' defined but not used [-Werror=unused-function]
 static int sysc_runtime_resume(struct device *dev)
            ^~~~~~~~~~~~~~~~~~~
drivers/bus/ti-sysc.c:421:12: error: 'sysc_runtime_suspend' defined but not used [-Werror=unused-function]
 static int sysc_runtime_suspend(struct device *dev)
            ^~~~~~~~~~~~~~~~~~~~

This marks the two unused functions as __maybe_unused to shut up
that warning.

Fixes: 0eecc636 ("bus: ti-sysc: Add minimal TI sysc interconnect target driver")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NTony Lindgren <tony@atomide.com>

We can handle the sysc interconnect target module in a generic way
for many TI SoCs. Initially let's just enable runtime PM with
autosuspend, and probe the children. This can already be used for
idling interconnect target modules that don't have any device driver
available for the child devices.

For now, the "ti,hwmods" custom binding is still required. That will
be eventually deprecated in later patches. And more features will be
added, such as parsing for sysc capabilities so we can continue
removing the legacy platform data.

Cc: Benoît Cousson <bcousson@baylibre.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Nishanth Menon <nm@ti.com>
Cc: Matthijs van Duin <matthijsvanduin@gmail.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Peter Ujfalusi <peter.ujfalusi@ti.com>
Cc: Sakari Ailus <sakari.ailus@iki.fi>
Cc: Tero Kristo <t-kristo@ti.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NTony Lindgren <tony@atomide.com>