feat(versal2): support alternate core as primary (non-cpu0)

Primary core was hardcoded to CPU0, causing a panic when another
core booted first.
Update logic to allow any booting core to become the p

feat(versal2): support alternate core as primary (non-cpu0)

Primary core was hardcoded to CPU0, causing a panic when another
core booted first.
Update logic to allow any booting core to become the primary and
gate secondary core startup inline to existing implementation for
secondary cores.

Change-Id: I6a5d76f23d4d4c4139d95bbaf55edf1244f2dbfe
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>

show more ...

fix(versal2): align comment about invalid console selection

Error message should be aligned actual symbol used for console section
which has been changed by commit 2333ab4cd214 ("fix(versal2): renam

fix(versal2): align comment about invalid console selection

Error message should be aligned actual symbol used for console section
which has been changed by commit 2333ab4cd214 ("fix(versal2): rename
console build arg to generic").

Change-Id: I230892875a6343ca8ffc55e0fac251f6586cf3f4
Signed-off-by: Michal Simek <michal.simek@amd.com>

show more ...

fix(versal2): remove PM_ABORT_SUSPEND API implementation

The API is not getting called by Linux. Removing it to reduce dead
code and improve maintainability.

Change-Id: I88025fa0213e40c510a2f1edb65

fix(versal2): remove PM_ABORT_SUSPEND API implementation

The API is not getting called by Linux. Removing it to reduce dead
code and improve maintainability.

Change-Id: I88025fa0213e40c510a2f1edb6566fc849a8dbb6
Signed-off-by: Madhav Bhatt <madhav.bhatt@amd.com>

show more ...

fix(versal2): enable graceful system shutdown

Power down the APU cores before sending SYSTEM_SHUTDOWN call to
PLM firmware. This ensures graceful system shutdown.

Change-Id: I89a132e72c9a2530620ffb

fix(versal2): enable graceful system shutdown

Power down the APU cores before sending SYSTEM_SHUTDOWN call to
PLM firmware. This ensures graceful system shutdown.

Change-Id: I89a132e72c9a2530620ffb58aaec752796d01d3e
Signed-off-by: Naman Trivedi <naman.trivedimanojbhai@amd.com>

show more ...

Merge changes from topic "xlnx_enhance_tfa_feature_check" into integration

* changes:
chore(xilinx): add deprecation warning to pm_feature_check
refactor(xilinx): rename eemi_feature_check to tf

Merge changes from topic "xlnx_enhance_tfa_feature_check" into integration

* changes:
chore(xilinx): add deprecation warning to pm_feature_check
refactor(xilinx): rename eemi_feature_check to tfa_api_feature_check

show more ...

fix(versal2): handle debugfs specific APIs before EEMI handler

In Versal Gen 2 SoC, all PM APIs use the extended SMC format including
the debugfs-specific APIs. So, call eemi_psci_debugfs_handler be

fix(versal2): handle debugfs specific APIs before EEMI handler

In Versal Gen 2 SoC, all PM APIs use the extended SMC format including
the debugfs-specific APIs. So, call eemi_psci_debugfs_handler before
eemi_api_handler. This ensures that debugfs-specific PM APIs are handled
correctly by TF-A and are not forwarded to the PLM firmware.

Change-Id: Ibab08c851c853a8f4272783b210040ddf7291d76
Signed-off-by: Naman Trivedi <naman.trivedimanojbhai@amd.com>

show more ...

fix(versal2): use PM_STATE_CPU_OFF for core power down in SELF_SUSPEND

Currently TF-A provides PM_STATE_CPU_IDLE state during SELF_SUSPEND
to power down the core. PM_STATE_CPU_IDLE is intended for C

fix(versal2): use PM_STATE_CPU_OFF for core power down in SELF_SUSPEND

Currently TF-A provides PM_STATE_CPU_IDLE state during SELF_SUSPEND
to power down the core. PM_STATE_CPU_IDLE is intended for CPU-idle
suspend paths (when Linux CPU idle is enabled) and is not the correct
state for a full core power-off.

Fix this by providing PM_STATE_CPU_OFF state to power down the core.

Change-Id: I25585b32fe90372b0404a1ad89544f1aaa2f34a2
Signed-off-by: Naman Trivedi <naman.trivedimanojbhai@amd.com>

show more ...

refactor(xilinx): rename eemi_feature_check to tfa_api_feature_check

Rename eemi_feature_check() to tfa_api_feature_check() for better
clarity. The new name clearly indicates its purpose of handling

refactor(xilinx): rename eemi_feature_check to tfa_api_feature_check

Rename eemi_feature_check() to tfa_api_feature_check() for better
clarity. The new name clearly indicates its purpose of handling TF-A
specific feature checks and improves code maintainability.

Change-Id: Ia74b12933427ccadbc8ede5ddc2a7a4822766264
Signed-off-by: Devanshi Chauhan <devanshi.chauhan@amd.com>

show more ...

feat(xilinx): use common SECURE/NON_SECURE macro

Remove platform-specific macro definitions such as SECURE_FLAG and
NON_SECURE_FLAG, and replace them with the common macros SECURE and
NON_SECURE acr

feat(xilinx): use common SECURE/NON_SECURE macro

Remove platform-specific macro definitions such as SECURE_FLAG and
NON_SECURE_FLAG, and replace them with the common macros SECURE and
NON_SECURE across all AMD-Xilinx platforms.

Change-Id: I95465e29ac8a9370da135c2113203c3206ecfec0
Signed-off-by: Ronak Jain <ronak.jain@amd.com>

show more ...

fix(xilinx): incorrect usage of SECURE_FLAG for psci

As per the PSCI specification, the PSCI SMC call always expects from
the NON_SECURE world. However, in the platform specific file SECURE
flag was

fix(xilinx): incorrect usage of SECURE_FLAG for psci

As per the PSCI specification, the PSCI SMC call always expects from
the NON_SECURE world. However, in the platform specific file SECURE
flag was passed to the firmware which is incorrect. Pass NON_SECURE
flag from the platform specific file to the firmware in order to
align with the PSCI specification.

Change-Id: Iabe2cb45467cf63fe36626d323513ff05548eb3b
Signed-off-by: Ronak Jain <ronak.jain@amd.com>

show more ...

feat(xilinx): deprecate PM_REQ_SUSPEND EEMI API

Deprecate the use of the PM_REQ_SUSPEND EEMI API from the Versal,
Versal-Net and Versal Gen 2 platforms. This is because the API is
intended for suspe

feat(xilinx): deprecate PM_REQ_SUSPEND EEMI API

Deprecate the use of the PM_REQ_SUSPEND EEMI API from the Versal,
Versal-Net and Versal Gen 2 platforms. This is because the API is
intended for suspending cross-subsystems, and the same functionality
can now be achieved using the ForcePowerdown API. Therefore,
continuing to use PM_REQ_SUSPEND API may no longer be necessary.
Hence deprecating the same.

Change-Id: I967d7803da4cf433fabfe8d87c32305954f69884
Signed-off-by: Ronak Jain <ronak.jain@amd.com>

show more ...

Merge "chore(versal2): rename versal2 to Versal Gen 2" into integration

Merge "feat(versal2): remove handoff entry from tl" into integration

Merge changes I5fcf6578,Ic7792603 into integration

* changes:
fix(xilinx): fix missing security flag in suspend path
feat(zynqmp): mark IPI calls secure/non-secure

chore(versal2): rename versal2 to Versal Gen 2

The term Versal2 should be updated to Versal Gen 2.

Change-Id: Iac94ef32604f88ec030bf95ec35484b72a0f7ffa
Signed-off-by: Prasad Kummari <prasad.kummari

chore(versal2): rename versal2 to Versal Gen 2

The term Versal2 should be updated to Versal Gen 2.

Change-Id: Iac94ef32604f88ec030bf95ec35484b72a0f7ffa
Signed-off-by: Prasad Kummari <prasad.kummari@amd.com>

show more ...

feat(versal2): add SPMD support for SPMC at S-EL1

Added support for SPMD when SPMC is running at S-EL1 on Versal Gen 2
platform. Added DTB with manifest addresses to BL32 for proper
initialization.

feat(versal2): add SPMD support for SPMC at S-EL1

Added support for SPMD when SPMC is running at S-EL1 on Versal Gen 2
platform. Added DTB with manifest addresses to BL32 for proper
initialization. Added `plat_spmd_handle_group0_interrupt` to handle
Group0 interrupts in SPMD. Added a new manifest source file compliant
with FFA 1.0 specification in which load_address and entrypoint
points to BL32 base address.

Change-Id: I518e2e799d3b86fcd67f9fee0af42503ca705488
Signed-off-by: Pranav Tilak <pranav.vinaytilak@amd.com>

show more ...

fix(xilinx): fix missing security flag in suspend path

Suspend flow was always programming wakeup sources with a fixed
secure flag, regardless of whether the caller was secure or
non-secure. This ma

fix(xilinx): fix missing security flag in suspend path

Suspend flow was always programming wakeup sources with a fixed
secure flag, regardless of whether the caller was secure or
non-secure. This may cause incorrect behavior for non-secure
suspend requests.

Fix this by passing the caller's security state (flag) through
pm_client_suspend() and pm_client_set_wakeup_sources() to ensure
that wakeup sources are set with the correct context.

Fixes: <4697164a3fa8> ("plat: xilinx: versal: Mark IPI calls secure/non-secure")

Change-Id: I5fcf65788a54010b4759b0d08e4f54c6e5037e47
Signed-off-by: Ronak Jain <ronak.jain@amd.com>

show more ...

fix(versal2): move plat_core_pos_by_mpidr to asm

In the current implementation, plat_core_pos_by_mpidr() is defined in
C.When BL31 is compiled with Armclang, a call to plat_core_pos_by_mpidr()
from

fix(versal2): move plat_core_pos_by_mpidr to asm

In the current implementation, plat_core_pos_by_mpidr() is defined in
C.When BL31 is compiled with Armclang, a call to plat_core_pos_by_mpidr()
from plat_my_core_pos() results in the return address stored in register
x30 becoming invalid and register x9 (used later) ends up with the value
0x0. Consequently, the CPU branches to address 0x0, triggering a
synchronous exception. TF-A then invokes the BHB flush code before
resuming execution. However, since the stack is not properly initialized
at this stage, the system eventually enters plat_panic_handler().
In the updated implementation, the platform_get_core_pos() function
is redefined in assembly to provide tighter control during early boot
stages. The MPIDR_EL1 register contains three affinity levels: Aff0
(bits [0:7]), Aff1 (bits [8:15]), and Aff2 (bits [16:23]). In this
assembly function, the core ID is extracted from Aff1
(MPIDR_AFF1_SHIFT), and the cluster ID from Aff2 (MPIDR_AFF2_SHIFT).
cluster/core ID calculation. The macro PLATFORM_MPIDR_AFFINITY_MASK
introduced to mask MPIDR_EL1 register.

Change-Id: Id532bbcd68f18e87ceba01c8f961d8c15962a1a3
Signed-off-by: Prasad Kummari <prasad.kummari@amd.com>

show more ...

feat(versal2): remove handoff entry from tl

In Versal Gen 2, handoff passing mechanism is via transfer list.
Transfer list is packaged with handoff addresses of BL32, BL33 along
with DT blob as part

feat(versal2): remove handoff entry from tl

In Versal Gen 2, handoff passing mechanism is via transfer list.
Transfer list is packaged with handoff addresses of BL32, BL33 along
with DT blob as part of it.
Once TF-A process the hand off details, rest of the components
(primarily U-Boot) should not parse these details at non-secure world.

Post retrieval of handoff information, remove entry point structures
catering to OP-TEE and U-Boot.

Change-Id: Ia5ace44de68721dc73f29a07b1e79a9c97e4122a
Signed-off-by: Saivardhan Thatikonda <saivardhan.thatikonda@amd.com>

show more ...

fix(versal2): runtime console in debug mode

Whenever DEBUG is set to 1, the Makefile forces CONSOLE_RUNTIME to
pl011, regardless of the user-specified CONSOLE value. This causes a
build requested wi

fix(versal2): runtime console in debug mode

Whenever DEBUG is set to 1, the Makefile forces CONSOLE_RUNTIME to
pl011, regardless of the user-specified CONSOLE value. This causes a
build requested with CONSOLE=pl011_1 to register both pl011_1 and
pl011 as boot and runtime consoles. If the hardware is connected only
to UART1, this causes TF-A to hang when UART0 is selected as the
runtime console, since it waits indefinitely on the transmit FIFO.
The fix ensures that, in a DEBUG build, CONSOLE_RUNTIME defaults to
the same value as CONSOLE.

Change-Id: I2e29d5b77c43aa65f58224d226683f4a8d94271a
Signed-off-by: Prasad Kummari <prasad.kummari@amd.com>

show more ...

Merge "fix(versal2): remove extraneous parentheses" into integration

Merge "feat(versal2): add support of MMI_GEM as wakeup source" into integration

fix(versal2): remove extraneous parentheses

While compiling TF-A with Armclang, the following error was observed.
To fix it, the extraneous parentheses around the comparison were
removed to silence

fix(versal2): remove extraneous parentheses

While compiling TF-A with Armclang, the following error was observed.
To fix it, the extraneous parentheses around the comparison were
removed to silence the warning.

plat/amd/versal2/scmi.c:322:13: error: equality comparison with
extraneous parentheses [-Werror,-Wparentheses-equality]
if ((clock == NULL)) {

Change-Id: I51f955dd15d5020598c9e4ff2f7f92b55a909f9b
Signed-off-by: Prasad Kummari <prasad.kummari@amd.com>

show more ...

feat(versal2): add support of MMI_GEM as wakeup source

Add MMI_GEM device node ID and add support of MMI_GEM as wakeup
source.

Change-Id: I449e23df63e887978be1ff6d151542a27c44466f
Signed-off-by: Na

feat(versal2): add support of MMI_GEM as wakeup source

Add MMI_GEM device node ID and add support of MMI_GEM as wakeup
source.

Change-Id: I449e23df63e887978be1ff6d151542a27c44466f
Signed-off-by: Naman Trivedi <naman.trivedimanojbhai@amd.com>

show more ...

feat(versal2): add hooks for mmap and early setup

Add early setup hooks through custom_early_setup() and provide a
mechanism to support custom memory mapping, including the extension
of the memory m

feat(versal2): add hooks for mmap and early setup

Add early setup hooks through custom_early_setup() and provide a
mechanism to support custom memory mapping, including the extension
of the memory map via custom_mmap_add(). This change may also
require alignment of the MAX_XLAT_TABLE and MAX_XLAT_TABLES macros.
These can be defined within the custom_pkg.mk makefile as follows:

MAX_MMAP_REGIONS := XY
$(eval $(call add_define,MAX_MMAP_REGIONS))
MAX_XLAT_TABLES := XZ
$(eval $(call add_define,MAX_XLAT_TABLES))

If PLATFORM_STACK_SIZE is not already defined, a default value
should be used. This allows for configurability of the stack size
across different interfaces, such as custom packages. The
custom_early_setup() function enables early low-level operations
to bring the system into a correct state. Support for a custom
SiP service is also added. A basic implementation of
custom_smc_handler() is provided by the platform, while the actual
definition is expected to be supplied by the custom package. This
feature is designed for use by external libraries, such as those
that require status checking. This code introduces a generic
framework for integrating custom logic via the
$(CUSTOM_PKG_PATH)/custom_pkg.mk makefile, including
optional support for custom SMC functionality, which is determined
by the custom package.

Change-Id: I40281acf2dc48be43471b8642e2ab1a93b1cf8f6
Signed-off-by: Prasad Kummari <prasad.kummari@amd.com>

show more ...