Mocks are provided for all framework files providing
function prototypes. Other framework files are excluded
as their generated mocks are only stubs, and their inclusion
may mislead that they provide mocked functions when they
do not (macro functions in particular may mislead in this
way, as these can't currently be mocked automatically).

Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: Ie1e5605c4ce625d6eeb797b18771c92de53fe113

To enable the generation of framework mocks, stripable
macros have been added and modified. Regexes for macros with
parameters have been appended to remove the parameters as well
as the macro name.

Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: I83f1bf7a53a472a2f4206cff75a4764f0c3d73cb

The address of the cfg.yml file in unit_test/ is
provided as the default value for this flag. This
enables a more ergonomic use of gm.rb, reducing
boilerplate arguments, for example:

gm.rb -m mocks/ -f file_to_mock.h

Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: I1d772f741e40663ca31a04234437c4b46444e9d1

This enables the use of gm.rb in directories other
than its own, particularly in test/ directories under
modules.

Change-Id: Iced3e0146b9baf214be456ddc2c155524852fa3c
Signed-off-by: Brett Warren <brett.warren@arm.com>

When using the mocked version of a file,
the real source needs to be removed and replaced
with the mocked sources. For certain files such as
fwk_notification some build flags need to be enabled
for proper function. Macros are provided to bundle the
required steps together.

Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: If625b39baed2f03b07d6b0589844ef0df2bb2d95

New dependencies are added via contrib/cmock:

    - CMock, a mock/stub generator and runtime, designed for use with:
    - Unity Test, a simple unit test framework built for C.

These are integrated into the project via unit_test, which provides:
    - configuration for the framework, including project-wide mocks
    - tools for generating new mocks
    - cmake build system configuration allowing modules to be added to the
      test framework
    - doccumentation instructing how to configure and run the test suite and
      add new module test directories.

The new test framework can be built and run via the 'mod_test' target,
and it is also run in conjuction with framework tests under the existing
'test' target (to build framework tests independently, use 'fwk_test' target).

Signed-off-by: srevad01 <sreehari.vadlamudi@arm.com>
Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: I6174ca8bb78f6fd34ed71d300873ce5de5a5d27b

Upcoming unit testing patches autogenerate mock
files which shouldn't be subject to the usual code style
checks. Module-specific mocks will be held in "module/*/test/mocks",
so an expanded glob is used to exclude all of these
directories.

Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: I840ebce8d14344d625476cde9fd1fd71501af475

is_valid_directory was not recognising excluded directories
because of faulty logic, this is rectified here.

Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: I6f278f45f4a21122a3a9586fc1f1f47387533391

Allow python files outside of tools/ to be checked.

Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: I27b0c9f6bf51cc1608862f54af73bcd6a15e4076

Signed-off-by: Brett Warren <brett.warren@arm.com>
Change-Id: I4319394a1f83db3d6f755a3194e2b327542e23d0

Add support to publish a list of application CPUs that should not be
used on the platform. The MPID of such CPUs are put together into a
isolated CPU list and published via SDS. Typical reasons for isolating
certain CPUs include faulty CPUs being identified or restricting the
number of CPUs available for use on the platform.

The isolated CPUs list is supplied from the platform system module's
config data. This list is then published via the SDS to allow the next
stages of boot software to obtain the list of CPUs to isolate.

Signed-off-by: Nishant Sharma <nishant.sharma@arm.com>
Change-Id: I6f7d1045ea98ad4f33ae13bc6014dfd0768692c6

Update CPU info in SDS. The CPU info contains the logical ID of
the primary boot CPU on the platform and this is picked up from the
platform system module's context information.

Signed-off-by: Nishant Sharma <nishant.sharma@arm.com>
Change-Id: Ib9085e04c63509f3af36aff4dbdf8ae274c48648

Primary CPU number is hardcoded at multiple places and so this makes it
difficult to use a different CPU as the primary CPU. So add support to
specify the primary CPU's MPID number via platform config and calculate
logical number during runtime.

Signed-off-by: Nishant Sharma <nishant.sharma@arm.com>
Change-Id: Ifb15f1525de44baabb9038050478b301c9d272b6

Validate the return value of the messaging_stack_ready function and
print error message on an error in updating the feature flags via SDS.

Signed-off-by: Nishant Sharma <nishant.sharma@arm.com>
Change-Id: Ie933aba4b30485d288321ac7df68ccee6393d3f1

Simplify the definition of SDS structure ID enum values by using a new
macro 'SDS_STRUCT_ID'. This macro appends the SDS ID value and the major
version of the SDS structure ID.

Signed-off-by: Nishant Sharma <nishant.sharma@arm.com>
Change-Id: I124b9df5377e43c219083a5d3e68fc7742af3cc4

Add support for RD-N2 MultiChip patform variant. This variant is a quad
chip platform where the chips are connected via a high speed CCIX link.
Each chip has 4 clusters. The "PLATFORM_VARIANT" value is set to 2.

In addition to this, increase the chip address space to 64 TiB to
account for the new changes in the increase in the address bus width
in the rdn2 variants.

Signed-off-by: Nishant Sharma <nishant.sharma@arm.com>
Signed-off-by: Vijayenthiran Subramaniam <vijayenthiran.subramaniam@arm.com>
Change-Id: I01256b9b254fa656676f08715c53350fac8b1bcc

Add FWU Multi Bank Update and boot BL2 from FIP image support for
SynQuacer platform.

When the DSW 3-4 is on, SynQuacer platform will decode FWU metadata
and platform metadata to choose the active bank of the FIP image
and boot BL2 from it.

This also enables the platform trial boot, which count up the boot
counter in the platform metadata, and when it reaches to the limit
(== 3) it switch back to the previous active bank. This counter will
stop when the BL33 (typically U-Boot) clears the counter.

Signed-off-by: Jassi Brar <jaswinder.singh@linaro.org>
Signed-off-by: Masami Hiramatsu <masami.hiramatsu@linaro.org>
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org>
Change-Id: I00b9c932fac5b4a9a231fcf2a050b766abefed61

The FastChannels functionality was introduced a while back to support
SCMI FastChannels as per SCMI spec v2.0. At that time only the
Performance protocol could advertise such support.
Now with the recent version 3.1[1] of the SCMI spec,also the Power
Capping protocol can support FastChannels.

It is now time to reflect which protocol the current FastChannel
support refers to, by adding _SCMI_PERF_ to its definition.
So:
- SCP_ENABLE_FAST_CHANNELS becomes SCP_ENABLE_SCMI_PERF_FAST_CHANNELS
- BUILD_HAS_FAST_CHANNELS becomes BUILD_HAS_SCMI_PERF_FAST_CHANNELS

This will allow precise inclusion of protocol-specific support for
SCMI FastChannels.

As a reminder, the CMake build option will become:
    make -f Makefile.cmake \
    MODE=<mode> \
    TOOLCHAIN=<toolchain> \
    PRODUCT=<product> \
    EXTRA_CONFIG_ARGS+="-DSCP_ENABLE_SCMI_PERF_FAST_CHANNELS=yes"

[1]https://developer.arm.com/documentation/den0056/d/?lang=en



Change-Id: I0ec4ffbc1ad5d797b9543aa979a98cceb267a798
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>

This patch change event type to lights events for temperature readings
in order to remove overhead and memory consumption.

Signed-off-by: Leandro Belli <leandro.belli@arm.com>
Change-Id: I65c31f3707956809959da052b742ab9e4ca056c6

This patch adds temperature protection functionality that
allows to configure two different alarms, warning and critical.
They can be configured independently but the critical level
threshold should be above the one for warning. Also when an alarm is
triggered it is reported to the driver through an API.

A common case of use is that the first threshold would be used
as first line of action against over temperature, while the
second could be used to initiate a sequence of hard power-off.

Signed-off-by: Leandro Belli <leandro.belli@arm.com>
Change-Id: I76304abd2feafcfc15836870ea9f5b4cfd5cc550

Currently there is support for one temperature sensor
and a number of actors are controlled by one single loop.
There are however examples of applications where it could be
necessary to have multiple independent thermal domains.
This patch adds the option to add multiple thermal management
controllers using different temperature sensors, in order to
have multiple loops working at the same time.

Signed-off-by: Leandro Belli <leandro.belli@arm.com>
Change-Id: I92182e9ddd21d8d35ec1e47a7286eea482288508

Configure a MHUv3 LCP->SCP connection via a doorbell channel.

Signed-off-by: Tomás Agustín González Orlando <tomasagustin.gonzalezorlando@arm.com>
Change-Id: Ibffae0185e81427903f163a0102bb35fdd8c7901

Define some possible MHUv3 connections and devices in rdfremont and
set the addresses of all the Mailbox (MBX) and PostBox (PBX) in the
LCP.
These addresses will be used by the MHUv3 driver to perform
MHUv3 operations like triggering a message transfer.

Change-Id: I106fcf6889877cc9e8d096ab34764d7b8b148e69
Signed-off-by: Tomás Agustín González Orlando <tomasagustin.gonzalezorlando@arm.com>

This change adds support for MHUv3 hardware.
This is intended to be used to communicate between:

SCP <---> AP
SCP <---> RSS
SCP <---> LCP
SCP <---> MCP
SCP <---> SCP
MCP <---> MCP
LCP <---> AP

At the moment only doorbell channels are supported.

Change-Id: I62d09503ab77612c01b4a99707867fa68c024f6a
Co-authored-by: Tomás Agustín González Orlando <tomasagustin.gonzalezorlando@arm.com>
Signed-off-by: Girish Pathak <girish.pathak@arm.com>
Signed-off-by: Tomás Agustín González Orlando <tomasagustin.gonzalezorlando@arm.com>

Add uart pl011 configuration to validate uart on LCP.

Signed-off-by: Ahmed Gadallah <ahmed.gadallah@arm.com>
Change-Id: I8707b97d5dde4b427b6a4faf539ab7dbf9a3ab70

Add timer and interrupt configuration to validate interrupt
handling in the LCP.

Signed-off-by: Ahmed Gadallah <ahmed.gadallah@arm.com>
Change-Id: I01bbf36890914fb17bab4c6eecb8b1d26b830996

Update readme.md so that the reference to SCMI points to v3.1 instead of v1.0.

Signed-off-by: Chuyue Luo <Chuyue.Luo@arm.com>
Change-Id: I85d04d53b1da3c45f58c15f1524c6e9e4ad07a03

Some compilers complain the return type error as below.

 | arch/none/host/include/arch_helpers.h: In function ‘arch_interrupts_disable’:
 | arch/none/host/include/arch_helpers.h:25:1: error: no return statement in
 |	function returning non-void [-Werror=return-type]

Fix the same.

Change-Id: Ia3343fbbea57ae783b272f0c4db05a7060256f85
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>

The changes in this patch are to introduce a typedef, which will
allow either signed or unsigned sensor values. This new typedef is
used to be able to conditionally support int64, allowing the
introduction of negative value sensors.

It will be automatically enabled with SCMI-sensor-v2, however it
can also be enabled on its own using with the cmake tag
SCP_ENABLE_SENSOR_SIGNED_VALUE.

Existing platforms that cannot handle signed values will return
FWK_E_SUPPORT when the above build options are set.

Signed-off-by: Katherine Vincent <katherine.vincent@arm.com>
Change-Id: I0f24e31d4bc3861b1724d814e866ef97f09c4b88

N1SDP SoC platform specific build option,
SCP_N1SDP_SENSOR_LIB_PATH has been added which
imports the n1sdp sensor library.

Also, added a readme.md file which describes this option.

Signed-off-by: sahil <sahil@arm.com>
Change-Id: Ie99d738a54d881bcb315e36efad5cb0b58d5a847

Replace Sensor Library with dummy Interface Implementation,
allowing relocation of closed source IP outside SCP codebase.

Signed-off-by: sahil <sahil@arm.com>
Change-Id: I6041f903e75dfaa804c2c299e033de6ae8b3ad65

Adding rdfremont along with an LCP target consisting of
armv8m MPU configuration example.

Signed-off-by: Ahmed Gadallah <ahmed.gadallah@arm.com>
Change-Id: Ibe77e1a62ded74caca6b1bfa04cee51e565127b7

The ARMv8-M architecture comes with a different MPU
than the previous ARMv7-M architecture.

Hence a new MPU module is added in this change to
support it.

Signed-off-by: Ahmed Gadallah <ahmed.gadallah@arm.com>
Change-Id: Ie8d4ff62c05186ec141114ac2508799f0f81a867

ARMv7-M support arch directory is modified to support both ARMv7-M
and ARMv8-M. The modified directory is renamed from armv7-m to
arm-m. This in turn affects the SCP_ARCHITECTURE cmake flag which
was previuosly set to armv7-m for armv7m targets and now would be
set by arm-m.

The ARMv8-M support does not include the security extension.

This ARMv8-M support includes support for ARMv8-M Mainline and
ARMv8.1-M Mainline.

ARMv8.1-M Mainline is supported by GCC ver 10.1 onwards, the
latest version can be found here:
https://developer.arm.com/tools-and-software/open-source-software


/developer-tools/gnu-toolchain/gnu-rm/downloads
in addition to LLVM ver 13.0.0 onwards and ArmClang ver 6.14 onwards.

Signed-off-by: Ahmed Gadallah <ahmed.gadallah@arm.com>
Change-Id: I7bdac6c80efbb56a0874daad794111376866a46c

Update to CMSIS pack 5.8.0 is done as it supports cortex-m55
where the current version 5.2.0 doesn't.

Additionally __VTOR_PRESENT macro is defined for all platforms
that use cortex-m3/7 as it is required by the new cmsis pack.

Signed-off-by: Ahmed Gadallah <ahmed.gadallah@arm.com>
Change-Id: Id7f4119c04b04bbedc0c5e4b5a5b9fa8832bc357

When the scp is running as an application or sub-system, there is a need to
terminate the process gracefully. Let introduce the deinit and stop
sequence to allow the modules and their elements to cleanup and release
resources before terminating the process

Add a new fwk_arch_deinit() to loop on modules and stop them.

Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Tarek El-Sherbiny <tarek.el-sherbiny@arm.com>
Change-Id: I9da1486b8096043f94845c06980d3e7ae373b9e0

Update the interface of fwk_interrupt_global_{enable|disable} functions to
save and restore the current interrupt state.

Add a new framework interface to get current interrupt context state:
- fwk_is_interrupt_context()

No functional change expected to current arch.

Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Tarek El-Sherbiny <tarek.el-sherbiny@arm.com>
Change-Id: I61569d1ff21bbc7d7bcc82eee32fc6d7d28b8469

Some architectures don't want to loop indefinitly in the same context of
the initialization. Let architecture decides if it wants to stay in the
initialization context or finish processing the events raised during
init and return.

No functional changes expected for current arch.

fwk_process_event_queue() can then be called to process the pending events.

Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Tarek El-Sherbiny <tarek.el-sherbiny@arm.com>
Change-Id: I626c0159e3bde38c4ed4d2931ed993cc5a5821dc

The inclusion of the notification_id member is now dependent on
whether or not BUILD_HAS_SCMI_NOTIFICATIONS build flag is set.
But in the reset_domain implementation, the functionality associated
with that identifier is conditional to the availability of framework
notifications BUILD_HAS_NOTIFICATION.

This creates a misalignment of build option for the correct use of
the reset domain features.

Replace BUILD_HAS_SCMI_NOTIFICATIONS with BUILD_HAS_NOTIFICATION
for the inclusion of the notification_id.

Change-Id: If596761faae827a9a7b3fe4101b36039b0fb2720
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>

The HAL API type that should be returned by whoever binds to
the reset module should be `mod_reset_domain_api` and not
`mod_reset_domain_drv_api`.

Fix the type.

Change-Id: Ice6690c575db43af42e2c2ca014968aeb3c7bd96
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>