Add delayed response support when !BUILD_HAS_MULTITHREADING is not
defined. This will enable all build modes to utilize deferred
responses.

Change-Id: I2c3d6b020ae3a5b9af605c18d6244dea679a913b
Tested-by: Nicola Mazzucato <nicola.mazzucato@arm.com>
Signed-off-by: Souvik Chakravarty <souvik.chakravarty@arm.com>

We need to support delayed responses in non-MULTITHREADED mode. As a
first step, we split delayed response functionality from
fwk_multi_thread.c into a separate file which can be used by both
MULTITHREADED and non-MULTITHREADED modes of compilation. We also
include the appropriate headers to fix compilation issues in
affected files as a result of this change.

Change-Id: I3c48e130d3c8a61b06d2432bcf51760bfd19abd8
Tested-by: Nicola Mazzucato <nicola.mazzucato@arm.com>
Signed-off-by: Souvik Chakravarty <souvik.chakravarty@arm.com>

The value of status wasn't used after being assigned, propagate the
status to caller.

Change-Id: Ida0f1b4b99f41f643397c9363bac882746d54fd1
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

The file tools/cppcheck_suppress_list.txt contains the list of warnings
we want to suppress when running Cppcheck.

Change-Id: I5998e58fccfe7706d16d59dcb4232f14ea0f0dd0
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

The variable spurious was incremented but never read as well as local to
the function.

Change-Id: If5d4eaf48ed02e8623dca0fc256f7bb82b47c0c4
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

This script can be run locally via ./tools/check_api.py. When forbidden
functions are found, filename and line number for each match are printed on the
error output before the function name.

Change-Id: I1668e9e4f2a57d37b62ee984086edff9651132ec
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

In production, the assert are deactivated so the status needs to be
propagated to the caller whether or not it is successful.

Change-Id: I8889715907453db22cbd2cc4fa500bb87971d308
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

Assignment operator was used instead of equal operator. This patch
corrects occurrences in framework/test/.

Change-Id: I4c117b764497ced9a437abcdd854e2e9e11c3673
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

This error was picked up by static analyser. The condition here will make
sure we are always in the valid case and never dereference a NULL
pointer.

Change-Id: I60d95bbc977e02b9c7eb46154aeb927f67d6f4a6
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

The last ternary operator (for speed == 1333) was useless because it did
not differentiate between true or false. If more values are needed maybe
a switch-case would be more appropriate.

Change-Id: I43392fa4ac9c2487e523cbec5c1de37ef8c3f11b
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

Cppcheck detected that status variable was updated but not used
afterward. This patch fix this by propagating the status' value.

Change-Id: I9eb8e3a39e69d2d334548042f4e6d4f5d1ba3820
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

Cppcheck detected a possible NULL dereference. This patch corrects this
bug.

Change-Id: I42b75e0ac29eee173839e9727c8c6832400ea925
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

Cppcheck detected that the incorrect format letter was used within
printf's format string. This patch fixes this warning.

Change-Id: I26957134ded3e441d51d7e8b0318d5adfecae20f
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

Two undefined behaviour errors were detected by cppcheck due to bitshift
on signed integer. This patch corrects them.

Change-Id: Iccf51e91b199c2ec5199daa883079169fe24df24
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

While running cppcheck static analysis on the code base, two undefined
behaviour errors were detected due to bitshift. This patch corrects
them.

Change-Id: I5cba7e0293a97c56be38e9cc3def35c9dfa67bb2
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

While running cppcheck static analysis on code base, an
uninitialised variable error was detected. This patch corrects it.

Change-Id: I4853572e9261e76fd6ed324de1ed3ebd5f109f66
Signed-off-by: Raphael Gault <raphael.gault@arm.com>

This patch adds a mocker for sensors.
This module implements the sensor driver API and
defers the get_value requests. This module requires
the timer facilities in order to setup an alarm.
This would emulate a sensor that provides a value
through an ISR and a call to the driver_response API
of the sensor module.

This module can be used to test sensor call flow
in platforms that do not provide asynchronous sensor
drivers.

Change-Id: I37baa428c12d6aa1c3de2f6f534ab5ac158f2352
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>

This patch removes a redundant check from some of the
calls to the driver.

Change-Id: I90c9d3bd028c9fb571e8327c873ffd43bd654821
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>

This patch adds missing documentation details related
to the PENDING return status.

Change-Id: I1e75146fe6f2e96d66072c6a681aec3480e43387
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>

This patch adds enumeration of the indices for the different alarms
required for the Juno system.

Change-Id: I12577cb618532a43b346a0ef67f95a09a797493e
Signed-off-by: Elieva Pignat <Elieva.Pignat@arm.com>

Remove a redundant line of code from the SCMI clock module.

Change-Id: I1a1877427b2dc3c1688d4e686f2982297fd39841
Signed-off-by: Jim Quigley <jim.quigley@arm.com>

The SCMI clock _init() function allocates memory for the
clock_operations data incorrectly.

Change-Id: Iab812fbe0ef3cbc9a876ed74a3d5f0809f2bae53
Signed-off-by: Jim Quigley <jim.quigley@arm.com>

The SCMI clock protocol module uses incorrect permissions when checking
that the requested operation is allowed for the device.

Change-Id: Ia77c48ab030b061cbe699302360eafa0b009861b
Signed-off-by: Jim Quigley <jim.quigley@arm.com>

This patch adds support for returning FWK_PENDING from an interface
with the I2C HAL module. The function returns FWK_PENDING when the
transaction has been submitted but not completed. The completion will
be notified later by a response event.

At the same time, drivers and clients are updated to take into account
these changes.

Change-Id: I87d83e005ebf1b76f1707519131ec3d6eb5344b7
Co-authored-by: Nicola Mazzucato <nicola.mazzucato@arm.com>
Signed-off-by: Elieva Pignat <Elieva.Pignat@arm.com>
Signed-off-by: Jim Quigley <jim.quigley@arm.com>
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>

This patch:
-removes an unused event identifier
-optimize the parameters passed to internal helpers

Change-Id: Ib706f589fa0c7993f7bdf2dfb5c669c51d5c144a
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>

This patch fixes the status returned to scmi when
the result is provided through a response event. In the
case of a hardware failure, the agent is now correctly
informed of such state.

Change-Id: I96c09e3f63afa86bef64da81da346895ea738f22
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>

Change-Id: I8f7e372e2495f01ec2c44ca3d8b5e545469216da
Signed-off-by: Chris Kay <chris.kay@arm.com>

The original code co-opts the response event for its own response to
the original HAL event. This change creates a separate local event
which will be used as the HAL response.

Change-Id: Ic4aab04d63269a1a77be655b951cbb309ef09029
Signed-off-by: Chris Kay <chris.kay@arm.com>

This line currently incorrectly checks the element index instead of
the event index.

Change-Id: Id1cac5c150d52f79c0491f1b33d6d66b626e9a6e
Signed-off-by: Chris Kay <chris.kay@arm.com>

Prior this change was %e a valid format for printing error codes as
strings when identified. This change removes such support and replace
use of %e with use of standard format identifier and a error code
to string conversion function.

With this change, any standard printf-like format comply with
SCP-firmware traces implementation. fwk_status_str() helps traces to
get nice string identifiers for errors.

This change updates few impacted modules.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

This commit adds enumerations representing the indices of SGM-775's
`psu` and `mock_psu` elements, in order to avoid tying them to a
particular integer literal when we need to refer to an element.

Change-Id: I3f6aec07524b29ceb610dc05096eba8b5206ca25
Signed-off-by: Chris Kay <chris.kay@arm.com>

This commit adds an enumeration representing the indices of SGM-775's
`timer` elements, in order to avoid tying them to a particular integer
literal when we need to refer to an element.

Change-Id: Ic68cb4f1d9f1404f6b97dbc14de7f25fa2c2b0eb
Signed-off-by: Chris Kay <chris.kay@arm.com>

This commit allows elements of the `mock_psu` module to emulate
asynchronous operations.

Enabling asynchronous operations forces requests to the element to be
pended for later completion, and is useful for testing interfaces that
interact with `mock_psu` which need to support this functionality.

Change-Id: Ic52221c510de8da60501a163b3f7dcacef6b0a3b
Signed-off-by: Chris Kay <chris.kay@arm.com>

This commit adds the extra handling required to support drivers that may
pend certain requests, as opposed to acting on them immediately.

This extra support is intended for drivers that might otherwise busy-
wait for extended periods, such as when they do not immediately know the
result of an operation they have begun.

Change-Id: I39aa9e4129d833044bd0e87b324c701d59a6908f
Signed-off-by: Chris Kay <chris.kay@arm.com>

Add missing declaration of exported functions fwk_interrupt_init()
and fwk_interrupt_set_isr_fault().

fwk_interrupt_init() is an effective exported API hence declared from
fwk_interrupt.h.

fwk_interrupt_set_isr_fault() is a test only API hence declared from
internal/fwk_interrupt.h.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

This patch increases the PCIe RP inbound region size to 48-bits
as there are cases where endpoint can generate a 48-bit VA
which gets translated by SMMU.

Change-Id: Ieca1c59b99f5da50e141237cd79f12ed17fc6ed7
Signed-off-by: Manoj Kumar <manoj.kumar3@arm.com>

This patch fixes the power domain tree view for master and slave
chips differently.

Master chip sees the power domains of both master and slave chips
whereas slave sees the power domain of only slave chip.

Master chip's SCP PD tree view will look like:

                    -----SYSTOP (LOGICAL)-------
                   /                            \
                  /                              \
            ---SYSTOP0--                   ---SYSTOP1--
           /      |     \                 /      |     \
          /       |      \               /       |      \
         /        |       \             /        |       \
    CLUS0       CLUS1    DBGTOP0   CLUS2       CLUS3     DBGTOP1
   /    \      /    \             /    \      /    \
 CPU0--CPU1--CPU2--CPU3---------CPU4--CPU5--CPU6--CPU7

 Here the SYSTOP LOGICAL is at PD level 3 but currently there is an
 issue  in handling level 3 in PD driver. So until it is fixed, the
 master chip's SCP uses the following PD tree view:

            --------------SYSTOP0--------------
           /        /        |         \       \
          /        /         |          \       \
         /        /          |           \       \
    CLUS0       CLUS1      CLUS2       CLUS3  DBGTOP0
   /    \      /    \      /    \      /    \
 CPU0--CPU1--CPU2--CPU3--CPU4--CPU5--CPU6--CPU7

 The slave chip's SCP however looks a single chip view of PD tree having
 its own local power domains and looks like below:

            ---SYSTOP0--
           /      |     \
          /       |      \
         /        |       \
    CLUS0       CLUS1    DBGTOP0
   /    \      /    \
 CPU0--CPU1--CPU2--CPU3

Change-Id: I83b339261d6ddda9660d5d8c6b575fadf48e70bd
Signed-off-by: Manoj Kumar <manoj.kumar3@arm.com>

Change-Id: I079a4966ef6674dad5886314883f1b20fce61342
Signed-off-by: Manoj Kumar <manoj.kumar3@arm.com>

This patch adds remote power domain driver to handle the slave
power domains from master.

In multichip scenario, the power domain tree view of master and slave
are different. Master sees the power domain of both local domains
and slave chips' domains. However slave sees only it own local
power domains.

In master chip, the power domain handling of local domains is managed
through ppu_v1 driver which accesses the PPU registers to manage the
domains. For slave chip's power domain, this driver is pointed such
that this driver binds to the C2C module and the request is sent to
slave chip over I2C and slave performs the action accordingly.

Change-Id: I8146f67f9c37ca5d749fb4860f5ef58c9566383a
Signed-off-by: Manoj Kumar <manoj.kumar3@arm.com>

This patch cleans the power domain API in C2C module and respective
handling in master and slave chips.

It is not possible for slave chip to report power state transition
to master hence on slave chip, sync calls are used to power ON & OFF
the power domains. For this purpose, C2C is treated as separate
thread on slave chips.

Change-Id: Ie356d19d5d4eeba3a62e53e9907f58c944d391b4
Signed-off-by: Manoj Kumar <manoj.kumar3@arm.com>