| Age | Commit message (Collapse) | Author |
|---|
|
@protocols is an array of protocol identifiers that are implemented,
excluding the Base protocol. Four protocol identifiers are packed into
each array element. The number of elements of @protocols is specified by
callee-side.
Signed-off-by: Ye Li <[email protected]>
Signed-off-by: Alice Guo <[email protected]>
|
|
This patch is used to add SCMI clock control permissions to sandbox for
testing.
Signed-off-by: Alice Guo <[email protected]>
|
|
This patch adds SCMI pin control protocol support to make the pin
controller driver based on SCMI, such as
drivers/pinctrl/nxp/pinctrl-imx-scmi.c, can be bound to the SCMI agent
device whose protocol id is 0x19.
Signed-off-by: Alice Guo <[email protected]>
|
|
If there is a SoC specific SCMI protocol driver, using
scmi_proto_driver_get() function can avoid to add SoC specific code to
scmi_agent-uclass.c.
Signed-off-by: Alice Guo <[email protected]>
|
|
linker-genetated array
U_BOOT_SCMI_PROTO_DRIVER macro is used to add a SCMI protocol driver to
scmi_proto_driver list. scmi_proto_driver_get() function can be used to
match a SCMI protocol id and its driver.
Signed-off-by: Alice Guo <[email protected]>
|
|
It is very surprising that such an uclass, specifically designed to
handle resources that may be shared by different devices, is not keeping
the count of the number of times a power domain has been
enabled/disabled to avoid shutting it down unexpectedly or disabling it
several times.
Doing this causes troubles on eg. i.MX8MP because disabling power
domains can be done in recursive loops were the same power domain
disabled up to 4 times in a row. PGCs seem to have tight FSM internal
timings to respect and it is easy to produce a race condition that puts
the power domains in an unstable state, leading to ADB400 errors and
later crashes in Linux.
Some drivers implement their own mechanism for that, but it is probably
best to add this feature in the uclass and share the common code across
drivers. In order to avoid breaking existing drivers, refcounting is
only enabled if the number of subdomains a device node supports is
explicitly set in the probe function. ->xlate() callbacks will return
the power domain ID which is then being used as the array index to reach
the correct refcounter.
As we do not want to break existing users while stile getting
interesting error codes, the implementation is split between:
- a low-level helper reporting error codes if the requested transition
could not be operated,
- a higher-level helper ignoring the "non error" codes, like EALREADY and
EBUSY.
CI tests using power domains are slightly updated to make sure the count
of on/off calls is even and the results match what we *now* expect. They
are also extended to test the low-level functions.
Signed-off-by: Miquel Raynal <[email protected]>
|
|
The current code attempts to bind scmi_voltage_domain to regulator subnode
of the SCMI protocol node, so scmi_voltage_domain can then bind regulators
directly to subnodes of its node. This kind of behavior should not be in
core code, move it into scmi_voltage_domain driver code. Let the driver
descend into regulator node and bind regulators to its subnodes.
Fixes: 1f213ee4dbf2 ("firmware: scmi: voltage regulator")
Signed-off-by: Marek Vasut <[email protected]>
[Alice Guo: Fix scmi_regulator_bind]
Signed-off-by: Peng Fan <[email protected]>
|
|
Unfortunately this change breaks boot on K3 platform.
U-Boot will hang after:
U-Boot SPL 2025.04-01050-ga40fc5afaec0 (Apr 14 2025 - 07:31:32 +0000)
SYSFW ABI: 3.1 (firmware rev 0x0009 '9.2.7--v09.02.07 (Kool Koala)')
This reverts commit 197376fbf300e92afa0a1583815d9c9eb52d613a as
suggested in [1].
[1] https://lists.denx.de/pipermail/u-boot/2025-April/587032.html
Signed-off-by: Wadim Egorov <[email protected]>
Acked-by: Miquel Raynal <[email protected]>
|
|
i.MX95 System Manager uses interrupt driven communication which requires
the caller to set Bit[0] of channel flags to 1. When transmission
completes and the previous general purpose interrupt has been processed
by the other core, i.MX95 System Manager will set General Purpose
Interrupt Control Register (GCR). U-Boot polls General-purpose Status
(GSR) to check if the operation is finished.
Signed-off-by: Viorel Suman <[email protected]>
Signed-off-by: Alice Guo <[email protected]>
Reviewed-by: Ye Li <[email protected]>
Reviewed-by: Marek Vasut <[email protected]>
|
|
It is very surprising that such an uclass, specifically designed to
handle resources that may be shared by different devices, is not keeping
the count of the number of times a power domain has been
enabled/disabled to avoid shutting it down unexpectedly or disabling it
several times.
Doing this causes troubles on eg. i.MX8MP because disabling power
domains can be done in recursive loops were the same power domain
disabled up to 4 times in a row. PGCs seem to have tight FSM internal
timings to respect and it is easy to produce a race condition that puts
the power domains in an unstable state, leading to ADB400 errors and
later crashes in Linux.
CI tests using power domains are slightly updated to make sure the count
of on/off calls is even and the results match what we *now* expect.
As we do not want to break existing users while stile getting
interesting error codes, the implementation is split between:
- a low-level helper reporting error codes if the requested transition
could not be operated,
- a higher-level helper ignoring the "non error" codes, like EALREADY and
EBUSY.
Signed-off-by: Miquel Raynal <[email protected]>
|
|
As part of bringing the master branch back in to next, we need to allow
for all of these changes to exist here.
Reported-by: Jonas Karlman <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
When bringing in the series 'arm: dts: am62-beagleplay: Fix Beagleplay
Ethernet"' I failed to notice that b4 noticed it was based on next and
so took that as the base commit and merged that part of next to master.
This reverts commit c8ffd1356d42223cbb8c86280a083cc3c93e6426, reversing
changes made to 2ee6f3a5f7550de3599faef9704e166e5dcace35.
Reported-by: Jonas Karlman <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
Remove <common.h> from this driver directory and when needed
add missing include files directly.
Reviewed-by: Ilias Apalodimas <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
This change will be useful when we manually test SCMI on sandbox
by enabling/disabling a specific SCMI protocol.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
A power domain id on sandbox should be in the range from zero to
ARRAY_SIZE(scmi_pwdom) - 1. Correct the validity check logic.
Addresses-Coverity-ID: 467401 ("Out-of-bounds write")
Addresses-Coverity-ID: 467405 ("Out-of-bounds read")
Signed-off-by: AKASHI Takahiro <[email protected]>
|
|
SCMI power domain management protocol is supported on sandbox
for test purpose. Add fake agent interfaces and associated
power domain devices.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Add power domain driver based on SCMI power domain management protocol.
Signed-off-by: AKASHI Takahiro <[email protected]>
|
|
In this patch, added are helper functions to directly manipulate
SCMI power domain management protocol. DM compliant power domain
driver will be implemented on top of those interfaces in a succeeding
patch.
Signed-off-by: AKASHI Takahiro <[email protected]>
|
|
The sess variable in open_channel was not entirely
cleared to zero at the start of this function.
This commit ensures that the entire struct is cleared.
Signed-off-by: Francois Berder <[email protected]>
|
|
Now that we have Base protocol support, we will be able to check if a given
protocol is really supported by the SCMI server (firmware).
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
SCMI base protocol is mandatory, and once SCMI node is found in a device
tree, the protocol handle (udevice) is unconditionally installed to
the agent. Then basic information will be retrieved from SCMI server via
the protocol and saved into the agent instance's local storage.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
Adding SCMI base protocol makes it inconvenient to hold the agent instance
(udevice) locally since the agent device will be re-created per each test.
Just remove it and simplify the test flows.
The test scenario is not changed at all.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
This is a simple implementation of SCMI base protocol for sandbox.
The main use is in SCMI unit test.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
In SCMI base protocol version 2 (0x20000), new interfaces,
BASE_SET_DEVICE_PERMISSIONS/BASE_SET_PROTOCOL_PERMISSIONS/
BASE_RESET_AGENT_CONFIGURATION, were added. Moreover, the api of
BASE_DISCOVER_AGENT was changed to support self-agent discovery.
So the driver expects SCMI firmware support version 2 of base protocol.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
SCMI base protocol is mandatory according to the SCMI specification.
With this patch, SCMI base protocol can be accessed via SCMI transport
layers. All the commands, except SCMI_BASE_NOTIFY_ERRORS, are supported.
This is because U-Boot doesn't support interrupts and the current transport
layers are not able to handle asynchronous messages properly.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
This framework allows SCMI protocols to be installed and bound to the agent
so that the agent can manage and utilize them later.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
Move the location of scmi_bind_protocols() backward for changes
in later patches.
There is no change in functionality.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
In sandbox scmi agent, channels are not used at all. But in this patch,
dummy channels are supported in order to test protocol-specific channels.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
SCMI specification allows any protocol to have its own channel for
the transport. While the current SCMI driver may assign its channel
from a device tree, the core function, devm_scmi_process_msg(), doesn't
use a protocol's channel, but always use an agent's channel.
With this commit, devm_scmi_process_msg() tries to find and use
a protocol's channel. If it doesn't exist, use an agent's.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
The commit 85dc58289238 ("firmware: scmi: prepare uclass to pass channel
reference") added an explicit parameter, channel, but it seems to make
the code complex.
Hiding this parameter will allow for adding a generic (protocol-agnostic)
helper function, i.e. for PROTOCOL_VERSION, in a later patch.
Signed-off-by: AKASHI Takahiro <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Etienne Carriere <[email protected]>
|
|
scmi_to_linux_errno() is set to return an appropriate errno
which corresponds to a given SCMI status code.
But the current implementation always returns the same value.
Signed-off-by: AKASHI Takahiro <[email protected]>
|
|
Building U-Boot SPL with CLK_SCMI and SCMI_FIRMWARE Kconfig options
enabled and SPL_FIRMWARE disabled result in the following error.
drivers/clk/clk_scmi.o: in function `scmi_clk_gate':
drivers/clk/clk_scmi.c:84: undefined reference to `devm_scmi_process_msg'
drivers/clk/clk_scmi.c:88: undefined reference to `scmi_to_linux_errno'
drivers/clk/clk_scmi.o: in function `scmi_clk_get_rate':
drivers/clk/clk_scmi.c:113: undefined reference to `devm_scmi_process_msg'
drivers/clk/clk_scmi.c:117: undefined reference to `scmi_to_linux_errno'
drivers/clk/clk_scmi.o: in function `scmi_clk_set_rate':
drivers/clk/clk_scmi.c:139: undefined reference to `devm_scmi_process_msg'
drivers/clk/clk_scmi.c:143: undefined reference to `scmi_to_linux_errno'
drivers/clk/clk_scmi.o: in function `scmi_clk_probe':
drivers/clk/clk_scmi.c:157: undefined reference to `devm_scmi_of_get_channel'
make[1]: *** [scripts/Makefile.spl:527: spl/u-boot-spl] Error 1
make: *** [Makefile:2043: spl/u-boot-spl] Error 2
Add Kconfig option so that CLK_SCMI can be disabled in SPL to fix this.
Signed-off-by: Jonas Karlman <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
In scmi firmware driver, it is better to bind the scmi protocol driver
"scmi_voltage_domain" with the node name of the protocol 17 and not
the sub-node named "regulator", because is a fixed string which doesn't
provide information and because it is not aligned with the other scmi
protocol nodes.
For example on stm32mp135f-dk board with device tree in stm32mp131.dtsi
scmi: scmi {
compatible = "linaro,scmi-optee";
#address-cells = <1>;
#size-cells = <0>;
linaro,optee-channel-id = <0>;
shmem = <&scmi_shm>;
scmi_clk: protocol@14 {
reg = <0x14>;
#clock-cells = <1>;
};
scmi_reset: protocol@16 {
reg = <0x16>;
#reset-cells = <1>;
};
scmi_voltd: protocol@17 {
reg = <0x17>;
scmi_regu: regulators {
#address-cells = <1>;
#size-cells = <0>;
scmi_reg11: voltd-reg11 {
reg = <VOLTD_SCMI_REG11>;
regulator-name = "reg11";
};
scmi_reg18: voltd-reg18 {
reg = <VOLTD_SCMI_REG18>;
regulator-name = "reg18";
};
scmi_usb33: voltd-usb33 {
reg = <VOLTD_SCMI_USB33>;
regulator-name = "usb33";
};
};
};
};
Before the patch:
> dm tree
scmi_agent 0 [ + ] scmi-over-optee |-- scmi
clk 1 [ + ] scmi_clk | |-- protocol@14
...
reset 1 [ ] scmi_reset_domain | |-- protocol@16
nop 2 [ + ] scmi_voltage_domain | `-- regulators
regulator 0 [ + ] scmi_regulator | |-- voltd-reg11
regulator 1 [ + ] scmi_regulator | |-- voltd-reg18
regulator 2 [ + ] scmi_regulator | |-- voltd-usb33
...
after the patch:
> dm tree
scmi_agent 0 [ + ] scmi-over-optee |-- scmi
clk 1 [ + ] scmi_clk | |-- protocol@14
...
reset 1 [ ] scmi_reset_domain | |-- protocol@16
nop 2 [ + ] scmi_voltage_domain | `-- protocol@17
regulator 0 [ + ] scmi_regulator | |-- voltd-reg11
regulator 1 [ + ] scmi_regulator | |-- voltd-reg18
regulator 2 [ + ] scmi_regulator | |-- voltd-usb33
...
Signed-off-by: Patrick Delaunay <[email protected]>
|
|
When the CCF is activated, the dev->parent is not necessary
the reference to SCMI transport and the function devm_scmi_of_get_channel
failed for the registered SCMI clock, child for protocol@14,
the channel is null and the SCMI clock driver crash for any operations.
This patch changes the first parameter of the ops of_get_channel(),
aligned with other process_msg() to pass directly the good reference,
i.e. parent result of find_scmi_transport_device(dev)
which return the reference of the scmi transport device.
Fixes: 8e96801aa6a ("firmware: scmi: add multi-channel support")
Signed-off-by: Patrick Delaunay <[email protected]>
|
|
This function is also available as ofnode_is_enabled(), so use that
instead.
Signed-off-by: Simon Glass <[email protected]>
|
|
Updates .process_msg operators of the SCMI transport drivers that
supports multi-channel to use it now that drivers do provide
the reference through channel argument. These are the mailbox
agent, the optee agent and the smccc agent.
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Implements multi SCMI channel support in OP-TEE SCMI transport. An
SCMI protocol may use a dedicated channel, specified by the DT.
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Updates SCMI SMCCC transport driver to get SCMI channel reference
at initialization and use when posting SCMI messages.
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Updates SCMI mailbox transport driver to get SCMI channel reference
at initialization and use when posting SCMI messages.
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Adds resources for SCMI protocols to possibly use a dedicated SCMI
channel instead of the default channel allocated by the SCMI agent
during initialization. As per DT binding documentation, some SCMI
transports can define a specific SCMI communication channel for
given SCMI protocols. It allows SCMI protocols to pass messages
concurrently each other.
This change introduces new scmi agent uclass API function
devm_scmi_of_get_channel() for SCMI drivers probe sequences to get
a reference to the SCMI channel assigned to its related SCMI protocol.
The function queries the channel reference to its SCMI transport driver
through new scmi agent uclass operator .of_get_channel that uses Device
Tree information from related SCMI agent node.
Operator .of_get_channel returns a reference to the SCMI channel
assigned to SCMI protocol used by the caller device. SCMI transport
drivers that do not support multi-channel are not mandated to register
this operator. When so, API function devm_scmi_of_get_channel() returns
NULL and SCMI transport driver are expected to retrieve by their own
means the reference to the unique SCMI channel, for example using
platform data as these drivers currently do in U-Boot source tree.
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Defines local helper function find_scmi_transport_device() with the
instructions to find the SCMI transport device from a SCMI protocol
device.
Cc: Patrick Delaunay <[email protected]>
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Changes SCMI transport operator ::process_msg to pass the SCMI channel
reference provided by caller SCMI protocol device.
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Changes SCMI driver API function devm_scmi_process_msg() to add
an SCMI channel reference argument for when SCMI agent supports
SCMI protocol specific channels. First argument of devm_scmi_process_msg()
is also change to point to the caller SCMI protocol device rather
than its parent device (the SCMI agent device).
The argument is a pointer to opaque struct scmi_channel known from
the SCMI transport drivers. It is currently unused and caller a pass
NULL value. A later change will enable such support once SCMI protocol
drivers have means to get the channel reference during initialization.
Cc: Lukasz Majewski <[email protected]>
Cc: Sean Anderson <[email protected]>
Cc: Jaehoon Chung <[email protected]>
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Removes inaccurate inline description of OP-TEE SCMI PTA command
PTA_SCMI_CMD_GET_CHANNEL.
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Changes implementation when using TEE dynamically allocated shared
memory to synchronize with the Linux implementation where the legacy
SMT protocol cannot be used with such memory since it is expected from
device mapped memory whereas OP-TEE shared memory is cached and
hence should not be accessed using memcpy_toio()/memcpy_fromio().
This change implements the MSG shared memory protocol introduced
in Linux [1]. The protocol uses a simplified SMT header of 32bit
named MSG_SMT to carry SCMI protocol information and uses side channel
means to carry exchanged buffer size information, as TEE invocation API
parameters when used in the SCMI OP-TEE transport.
Link: [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=f301bba0ca7392d16a6ea4f1d264a91f1fadea1a
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Implements SCMI APIs to retrieve the number exposed SCMI clocks using
SCMI_PROTOCOL_ATTRIBUTES messages and the names of the clocks using
SCMI_CLOCK_ATTRIBUTES messages.
This change updates sandbox SCMI clock test driver to manage these
2 new message IDs.
Cc: Lukasz Majewski <[email protected]>
Cc: Sean Anderson <[email protected]>
Cc: Clement Leger <[email protected]>
Cc: Patrick Delaunay <[email protected]>
Reviewed-by: Patrick Delaunay <[email protected]>
Signed-off-by: Gabriel Fernandez <[email protected]>
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Updates sandbox SCMI clock driver and tests since enabling CCF will
mandate clock discovery that is all exposed SCMI clocks shall be
discovered at initialization. For this reason, sandbox SCMI clock
driver must emulate all clocks exposed by SCMI server, not only those
effectively consumed by some other U-Boot devices.
Therefore the sandbox SCMI test driver exposes 3 clocks (IDs 0, 1 and 2)
and sandbox SCMI clock consumer driver gets 2 of them.
Cc: Simon Glass <[email protected]>
Reviewed-by: Patrick Delaunay <[email protected]>
Signed-off-by: Etienne Carriere <[email protected]>
|
|
Changes devm_scmi_process_msg() first argument from target parent device
to current SCMI device and lookup the SCMI agent device among SCMI device
parents for find the SCMI agent operator needed for communication with
the firmware.
This change is needed in order to support CCF in clk_scmi driver unless
what CCF will fail to find the right udevice related to exposed SCMI
clocks.
This patch allows to simplify the caller sequence, using SCMI device
reference as parameter instead of knowing SCMI uclass topology. This
change also adds some protection in case devm_scmi_process_msg() API
function is called for an invalid device type.
Cc: Lukasz Majewski <[email protected]>
Cc: Sean Anderson <[email protected]>
Cc: Jaehoon Chung <[email protected]>
Cc: Patrick Delaunay <[email protected]>
Reviewed-by: Patrick Delaunay <[email protected]>
Signed-off-by: Etienne Carriere <[email protected]>
|
|
As per DT bindings since Linux kernel v5.14, the device tree can define
only 1 SCMI agent node that is named scmi [1]. As a consequence, change
implementation of the SCMI driver test through sandbox architecture to
reflect that.
This change updates sandbox test DT and sandbox SCMI driver accordingly
since all these are impacted.
Cc: Simon Glass <[email protected]>
Reviewed-by: Patrick Delaunay <[email protected]>
Signed-off-by: Etienne Carriere <[email protected]>
|
|
This change implements an SCMI transport for agent interfacing the
OP-TEE SCMI service. OP-TEE provides an SCMI PTA (Pseudo-TA) for
non-secure world to send SCMI messages over an identified channel.
The driver implemented here uses a SMT shared memory for passing
messages between client and server.
The implementation opens and releases channel resources for each
passed SCMI message so that resources allocated (sessions) or
registered (shared memory areas) in OP-TEE firmware are released for
example before relocation as the driver will likely allocate/register
them back when probed after relocation.
The integration of the driver using dedicated config switch
CONFIG_SCMI_AGENT_OPTEE is designed on the model posted to the
U-Boot ML by Patrick Delaunay [1].
Link: [1] https://lore.kernel.org/all/20211028191222.v3.4.Ib2e58ee67f4d023823d8b5404332dc4d7e847277@changeid/
Cc: Patrick Delaunay <[email protected]>
Cc: Wolfgang Denk <[email protected]>
Signed-off-by: Etienne Carriere <[email protected]>
Reviewed-by: Patrick Delaunay <[email protected]>
|
