On a system with multiple active SCMI agents, one agent(other than OSPM/
Linux or bootloader) would request to turn on a shared power domain
before the Linux boots/initialise the genpds. So when the Linux boots
and gets the power state as already ON, it just registers the genpd with
a default ON state.
However, when the driver that needs this shared power domain is probed
genpd sees that the power domain status is ON and never makes any SCMI
call to power it up which is correct. But, since Linux didn't make an
explicit request to turn on the shared power domain, the SCMI platform
firmware will not know if the OSPM agent is actively using it.
Suppose the other agent that requested the shared power domain to be
powered ON requests to power it OFF as it no longer needs it, the SCMI
platform firmware needs to turn it off if there are no active users of
it which in the above scenaro is the case.
As a result of SCMI platform firmware turning off the resource, OSPM/
Linux will crash the moment as it expects the shared power domain to be
powered ON.
Send an explicit request to set the current state when setting up the
genpd power domains so that OSPM registers its vote in the power domain
state with the SCMI platform firmware.
The other option is to not read the state and set the genpds as default
OFF, but it can't handle the scenario on certain platforms where SCMI
platform keeps all the power domains turned ON by default for faster boot
(or any other such variations) and expect the OSPM to turn off the unused
domains if power saving is required.
Link: https://lore.kernel.org/all/Z4aBkezSWOPCXcUh@bogus
Reported-by: Ranjani Vaidyanathan <ranjani.vaidyanathan@nxp.com>
Reported-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Reviewed-by: Cristian Marussi <cristian.marussi@arm.com>
Link: https://lore.kernel.org/r/20250115113931.1181309-1-sudeep.holla@arm.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Merge the pmdomain fixes for v6.13-rc[n] into the next branch, to allow them
to get tested together with the new changes that are targeted for v6.14.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add Airoha CPU PM Domain support to control frequency and power of CPU
present on Airoha EN7581 SoC.
Frequency and power can be controlled with the use of the SMC command by
passing the performance state. The driver also expose a read-only clock
that expose the current CPU frequency with SMC command.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Link: https://lore.kernel.org/r/20250109131313.32317-1-ansuelsmth@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Merge the pmdomain fixes for v6.13-rc[n] into the next branch, to allow them
to get tested together with the new changes that are targeted for v6.14.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The genpd device, which is really only used as a handle to lookup
OPP, but not even registered to the device core otherwise and thus
lifetime linked to the genpd struct it is contained in, is missing
a release function. After b8f7bbd1f4 ("pmdomain: core: Add
missing put_device()") the device will be cleaned up going through
the driver core device_release() function, which will warn when no
release callback is present for the device. Add a dummy release
function to shut up the warning.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Fixes: b8f7bbd1f4 ("pmdomain: core: Add missing put_device()")
Cc: stable@vger.kernel.org
Message-ID: <20241218184433.1930532-1-l.stach@pengutronix.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
imx_gpcv2_probe() leaks an OF node reference obtained by
of_get_child_by_name(). Fix it by declaring the device node with the
__free(device_node) cleanup construct.
This bug was found by an experimental static analysis tool that I am
developing.
Fixes: 03aa12629f ("soc: imx: Add GPCv2 power gating driver")
Signed-off-by: Joe Hattori <joe@pf.is.s.u-tokyo.ac.jp>
Cc: stable@vger.kernel.org
Message-ID: <20241215030159.1526624-1-joe@pf.is.s.u-tokyo.ac.jp>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
When a device supports IO daisy-chain wakeups, it uses a dedicated
wake IRQ. Devices with IO daisy-chain wakeups enabled should not set
wakeup constraints since these can happen even from deep power states,
so should not prevent the DM from picking deep power states.
Wake IRQs are set with dev_pm_set_wake_irq() or
dev_pm_set_dedicated_wake_irq(). The latter is used by the serial
driver used on K3 platforms (drivers/tty/serial/8250/8250_omap.c)
when the interrupts-extended property is used to describe the
dedicated wakeup interrupt.
Detect these wake IRQs in the suspend path, and if set, skip sending
constraint.
Tested-by: Dhruva Gole <d-gole@ti.com>
Signed-off-by: Kevin Hilman <khilman@baylibre.com>
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Message-ID: <20241206-lpm-v6-10-constraints-pmdomain-v6-3-833980158c68@baylibre.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
For each device in a TI SCI PM domain, check whether the device has
any resume latency constraints set via per-device PM QoS. If
constraints are set, send them to DM via the new SCI constraints API.
Checking for constraints happen for each device before system-wide
suspend (via ->suspend() hook.)
An important detail here is that the PM domain driver inserts itself
into the path of both the ->suspend() and ->resume() hook path
of *all* devices in the PM domain. This allows generic PM domain code
to handle the constraint management and communication with TI SCI.
Further, this allows device drivers to use existing PM QoS APIs to
add/update constraints.
DM firmware clears constraints during its resume, so Linux has
to check/update/send constraints each time system suspends.
Also note that the PM QoS framework uses usecs as the units for
latency whereas the TI SCI firmware uses msecs, so a conversion is
needed before passing to TI SCI.
Co-developed-by: Vibhore Vardhan <vibhore@ti.com>
Signed-off-by: Vibhore Vardhan <vibhore@ti.com>
Reviewed-by: Markus Schneider-Pargmann <msp@baylibre.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Tested-by: Dhruva Gole <d-gole@ti.com>
Signed-off-by: Kevin Hilman <khilman@baylibre.com>
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Message-ID: <20241206-lpm-v6-10-constraints-pmdomain-v6-1-833980158c68@baylibre.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The gpcv2 drivers on imx8m series are registered as platform
devices and this opens the possibility of reloading the driver
at runtime.
But this doesn't actually work. There are some hardware sequence
dependecy between blk ctrl and gpc, also power domains are used
by other peripherals, so fix this by explicitly suppressing bind
attrs.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Message-ID: <20241206112731.98244-2-peng.fan@oss.nxp.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The blk ctrl drivers on imx8m series are registered as platform
devices and this opens the possibility of reloading the driver
at runtime.
But this doesn't actually work. There are some hardware sequence
dependecy between blk ctrl and gpc, also power domains are used
by other peripherals, so fix this by explicitly suppressing bind
attrs.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Message-ID: <20241206112731.98244-1-peng.fan@oss.nxp.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Pull pmdomain fixes from Ulf Hansson:
"Core:
- Fix a couple of memory-leaks during genpd init/remove
Providers:
- imx: Adjust delay for gpcv2 to fix power up handshake
- mediatek: Fix DT bindings by adding another nested power-domain
layer"
* tag 'pmdomain-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/linux-pm:
pmdomain: imx: gpcv2: Adjust delay after power up handshake
pmdomain: core: Fix error path in pm_genpd_init() when ida alloc fails
pmdomain: core: Add missing put_device()
dt-bindings: power: mediatek: Add another nested power-domain layer
When the ida allocation fails we need to free up the previously allocated
memory before returning the error code. Let's fix this and while at it,
let's also move the ida allocation to genpd_alloc_data() and the freeing to
genpd_free_data(), as it better belongs there.
Fixes: 899f44531f ("pmdomain: core: Add GENPD_FLAG_DEV_NAME_FW flag")
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Message-ID: <20241122134207.157283-3-ulf.hansson@linaro.org>
The continual trickle of small conversion patches is grating on me, and
is really not helping. Just get rid of the 'remove_new' member
function, which is just an alias for the plain 'remove', and had a
comment to that effect:
/*
* .remove_new() is a relic from a prototype conversion of .remove().
* New drivers are supposed to implement .remove(). Once all drivers are
* converted to not use .remove_new any more, it will be dropped.
*/
This was just a tree-wide 'sed' script that replaced '.remove_new' with
'.remove', with some care taken to turn a subsequent tab into two tabs
to make things line up.
I did do some minimal manual whitespace adjustment for places that used
spaces to line things up.
Then I just removed the old (sic) .remove_new member function, and this
is the end result. No more unnecessary conversion noise.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge the pmdomain fixes for v6.12-rc[n] into the next branch, to allow them
to get tested together with the new changes that are targeted for v6.13.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The check condition should be 'i < bc->onecell_data.num_domains', not
'bc->onecell_data.num_domains' which will make the look never finish
and cause kernel panic.
Also disable runtime to address
"imx93-blk-ctrl 4ac10000.system-controller: Unbalanced pm_runtime_enable!"
Fixes: e9aa77d413 ("soc: imx: add i.MX93 media blk ctrl driver")
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Stefan Wahren <wahrenst@gmx.net>
Cc: stable@vger.kernel.org
Message-ID: <20241101101252.1448466-1-peng.fan@oss.nxp.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Merge the pmdomain fixes for v6.12-rc[n] into the next branch, to allow them
to get tested together with the new changes that are targeted for v6.13.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
In the single PM domain case there is no need for platform code to specify
the index of the corresponding required OPP in DT, as the index must be
zero. This allows us to assign a required dev for the required OPP from
genpd, while attaching a device to its PM domain.
In this way, we can remove some of the genpd specific code in the OPP core
for the single PM domain case. Although, this cleanup is made from a
subsequent change.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://lore.kernel.org/r/20241002122232.194245-7-ulf.hansson@linaro.org
Merge the pmdomain fixes for v6.12-rc[n] into the next branch, to allow them
to get tested together with the new changes that are targeted for v6.13.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The MediaTek power domain driver contains a hack that assigns the device
node of the power domain to the struct device of the power domain
controller in order to use the devres regulator API.
Now that there is a proper OF-specific regulator API, and even a devres
version, replace the hack with proper code.
This change is incompatible with incomplete device trees. Instead of
assigning the dummy regulator in cases where the power domain requires
a supply but the device tree does not provide one, the driver will just
error out. This will be seen on the MT8390 EVK, which is missing
supplies for the IMG_VCORE and CAM_VCORE domains. And likely all the
MediaTek EVBs, which have no power domain supplies specified. This is
however the correct behavior. If the power domain's supply is missing,
then it should not work. Relying on other parts of the system to keep
the unattached regulator enabled is likely to break in ways less easier
to understand.
Signed-off-by: Chen-Yu Tsai <wenst@chromium.org>
Link: https://lore.kernel.org/r/20240930044525.2043884-4-wenst@chromium.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Commit 9094e53ff5 ("pmdomain: core: Use dev_name() instead of
kobject_get_path() in debugfs") severely shortened the names of devices
in a PM Domain. Now the most common format[1] consists of a 32-bit
unit-address (8 characters), followed by a dot and a node name (20
characters for "air-pollution-sensor" and "interrupt-controller", which
are the longest generic node names documented in the Devicetree
Specification), for a typical maximum of 29 characters.
This offers a good opportunity to reduce the table width of the debug
summary:
- Reduce the device name field width from 50 to 30 characters, which
matches the PM Domain name width,
- Reduce the large inter-column space between the "performance" and
"managed by" columns.
Visual impact:
- The "performance" column now starts at a position that is a
multiple of 16, just like the "status" and "children" columns,
- All of the "/device", "runtime status", and "managed by" columns are
now indented 4 characters more than the columns right above them,
- Everything fits in (one less than) 80 characters again ;-)
[1] Note that some device names (e.g. TI AM335x interconnect target
modules) do not follow this convention, and may be much longer, but
these didn't fit in the old 50-character column width either.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/f8e1821364b6d5d11350447c128f6d2b470f33fe.1725459707.git.geert+renesas@glider.be
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The removed code was added to handle the case where the power domain is
already on during the driver's probing. In this use case, the "is_off"
parameter is passed as false to pm_genpd_init() to inform it not to call
the power_on() callback, as it's unnecessary to perform the hardware
power-on procedure since the power domain is already on. Therefore, with
the call to clk_bulk_prepare_enable() by probe(), the system is in the
same operational state as when "is_off" is passed as true after the
power_on() callback execution:
probe() -> is_off == true -> clk_bulk_prepare_enable() called by power_on()
probe() -> is_off == false -> clk_bulk_prepare_enable() called by probe()
Reaching the same logical and operational state, it follows that upon
driver removal, there is no need to perform different actions depending
on the power domain's on/off state during probing.
Signed-off-by: Dario Binacchi <dario.binacchi@amarulasolutions.com>
Link: https://lore.kernel.org/r/20240825143428.556439-2-dario.binacchi@amarulasolutions.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
