Merge "fix(cm): deprecate use of NS_TIMER_SWITCH" into integration

feat(ti): de-assert AINACTS at boot

De-asserting AINACTS at startup to enable ACP interface. This is
required to enable limited system level IO coherency via ACP port
(eg.: DMA traffic on AM62L driv

feat(ti): de-assert AINACTS at boot

De-asserting AINACTS at startup to enable ACP interface. This is
required to enable limited system level IO coherency via ACP port
(eg.: DMA traffic on AM62L driving ASEL = 14/15).

Change-Id: Iceaf992dec5dc37eae6dc06895585ea712f23496
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>

show more ...

fix(cm): deprecate use of NS_TIMER_SWITCH

On AArch64, secure world has it's own EL3 physical timer registers
accessible to secure EL1 in absence of S-EL2. With S-EL2 there is
virtualized view availa

fix(cm): deprecate use of NS_TIMER_SWITCH

On AArch64, secure world has it's own EL3 physical timer registers
accessible to secure EL1 in absence of S-EL2. With S-EL2 there is
virtualized view available for EL1 timer registers. So it is
unreasonable for secure world to use non-secure EL1 physical timer
registers. Moreover, the non-secure operating system (Linux in our case)
relies heavily on these EL1 physical timer registers for scheduling
decisions. If NS_TIMER_SWITCH is enabled, it simply breaks the preemption
model of the non-secure world by disabling non-secure timer interrupts
leading to RCU stalls being observed on long running secure world tasks.

The only arch timer register which will benefit from context management
is cntkctl_el1: Counter-timer Kernel Control Register. This enables the
secure and non-secure worlds to independently control accesses to EL0
for counter-timer registers. This is something that OP-TEE uses to
enable ftrace feature for Trusted Applications and SPM_MM uses for EL0
access as well.

Lets enable context management of cntkctl_el1 by default and deprecate
conditional context management of non-secure EL1 physical timer
registers for whom there isn't any upstream user. With that deprecate
this NS_TIMER_SWITCH build option which just adds confusion for the
platform maintainers. It will be eventually dropped following
deprecation policy of TF-A.

Reported-by: Stauffer Thomas MTANA <thomas.stauffer@mt.com>
Reported-by: Andrew Davis <afd@ti.com>
Change-Id: Ifb3a919dc0bf8c05c38895352de5fe94b4f4387e
Signed-off-by: Sumit Garg <sumit.garg@oss.qualcomm.com>

show more ...

feat(ti): am62lx init: boot notif and version msg

Consume the boot notification on this platform before we start further
communication with the system-firmware.

Change-Id: Ia9e5d8d616d7d9cd50ee5de2

feat(ti): am62lx init: boot notif and version msg

Consume the boot notification on this platform before we start further
communication with the system-firmware.

Change-Id: Ia9e5d8d616d7d9cd50ee5de2e4c8abe06104dc05
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

feat(ti): add mmu regions for am62l soc

Update the k3low bl31 platform setup to map required device regions
(USART, GIC, GTC, MMR, and mailbox) in the MMU. This ensures that all
necessary hardware b

feat(ti): add mmu regions for am62l soc

Update the k3low bl31 platform setup to map required device regions
(USART, GIC, GTC, MMR, and mailbox) in the MMU. This ensures that all
necessary hardware blocks are accessible to the A53 cores on the
AM62L SoC.
Use 4K aligned address sizes wherever applicable, and update the
file header comment from "K3 SOC specific bl31_setup" to "k3low SoC
specific bl31_setup" to accurately represent the platform specific
nature of this file.
As part of the effort, rename WKUP_CTRL_MMR0_DEVICE_MANAGEMENT_BASE
to WKUP_CTRL_MMR0_BASE to make name shorter.

Change-Id: I58209bc9c780db3e452b09c2c939bb0c47a63ed1
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

feat(ti): build generic timer

Also, make sure we init the generic timer as part of the soc init
on am62lx as we use it later for delays

Change-Id: I921f4e4120ddaba588eb5b876231435fff3f7f3c
Signed-o

feat(ti): build generic timer

Also, make sure we init the generic timer as part of the soc init
on am62lx as we use it later for delays

Change-Id: I921f4e4120ddaba588eb5b876231435fff3f7f3c
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

fix(ti): specify allowable rcv_addr in mailbox

The ti_sci_transport_recv function had an overly restrictive check
where we enforced the rcv_addr to be always at the start of the
RX_START region in s

fix(ti): specify allowable rcv_addr in mailbox

The ti_sci_transport_recv function had an overly restrictive check
where we enforced the rcv_addr to be always at the start of the
RX_START region in shared memory. This started failing unnecessarily
when messages started being received at a location other than the
beginning.
However, the sender might send the messages at any location within
the share memory region if it does some sort of buffer management.
In such a case, TFA needs to just make sure the message being read
is well within the allocated shared memory region. Not necessarily
at the start.

Fixes: 9347ff4561e445d4 ("feat(ti): add support for TI mailbox driver")
Change-Id: I32cf0f5e4b9fedf49e40ace0bef06b7fc4c016f6
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

feat(ti): enable power state validation for K3 SoCs

Add power state validation for K3 SoCs to ensure proper PSCI power state
transitions. Defining validate_power_state enables the
PSCI_CPU_SUSPEND_A

feat(ti): enable power state validation for K3 SoCs

Add power state validation for K3 SoCs to ensure proper PSCI power state
transitions. Defining validate_power_state enables the
PSCI_CPU_SUSPEND_AARCH64 capability and, in turn, CPU_SUSPEND.

validate_power_state checks if the requested power level and power state
are in a valid configuration. If the power state requested is standby,
set the core power state to the maximum retention state.

Change-Id: Ied3ebdc583b2e5b7c86923705954d763997b324e
Signed-off-by: Kendall Willis <k-willis@ti.com>

show more ...

fix(ti): remove validate_power_state definition

validate_power_state should not be defined since CPU suspend is not yet
supported.

Defining the validate_power_state function without any meaningful

fix(ti): remove validate_power_state definition

validate_power_state should not be defined since CPU suspend is not yet
supported.

Defining the validate_power_state function without any meaningful logic
caused a bug where a crash would happen if a device tried to enter a power
state that was not yet supported by TFA. This is because
validate_power_state would return a pass when the state was not properly
checked.

Removing the validate_power_state definition disables the
PSCI_CPU_SUSPEND_AARCH64 capability which will prevent the HLOS from
trying to suspend the CPU all together. This will fix the aforementioned
bug from occurring.

Fixes: 2e9c9e829964 ("ti: k3: common: Add PSCI stubs")
Change-Id: I3bbbd228fbddea64e72f0cf50afc6b25fb6d317c
Signed-off-by: Kendall Willis <k-willis@ti.com>

show more ...

fix(ti): fix UNUSED_VALUE in AM62L PSCI Driver

New Defects reported by Coverity Scan found value "-6" to "ret" here,
but that stored value is overwritten before it can be used.
We'd rather just retu

fix(ti): fix UNUSED_VALUE in AM62L PSCI Driver

New Defects reported by Coverity Scan found value "-6" to "ret" here,
but that stored value is overwritten before it can be used.
We'd rather just return from those places to avoid this, similar
to how it was being done in k3_psci as well.

Change-Id: I7661586f191aad541387c06ab221c1d7cb58e461
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

feat(ti): introduce basic support for the AM62L

The AM62L is a lite, low power and performance optimized family of
application processors that are built for Linux application development.

Some high

feat(ti): introduce basic support for the AM62L

The AM62L is a lite, low power and performance optimized family of
application processors that are built for Linux application development.

Some highlights of AM62L SoC are:
- Single to Dual 64-bit Arm® Cortex®-A53 microprocessor subsystem
- 16-bit DDR Subsystem that supports LPDDR4, DDR4 memory types.
- Multiple low power modes support, ex: Deep sleep and RTC+DDR
- Mailbox transport layer for TI SCI

For more information check out our Technical Reference Manual (TRM)
which is loacted here:

https://www.ti.com/lit/pdf/sprujb4

Change-Id: I9d7c707b5b220c5ec13bd2de67f872b3da3c308a
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

feat(ti): introduce PSCI Driver for AM62L

The AM62L is a lite, low power and performance optimized family of
application processors that are built for Linux application development

There is enough

feat(ti): introduce PSCI Driver for AM62L

The AM62L is a lite, low power and performance optimized family of
application processors that are built for Linux application development

There is enough deviation from the previous K3 gen SoCs with regards to
how the PSCI functionality looks like on this device. For example, it
no longer does reset or LPSC turn ON/OFF operations using any external
Device Management entity like it did earlier.

The actual power ON/OFF operations will be implemented in a later stage
once all the PM related drivers are integrated. Such places in this
driver have been marked as TODO and will be replaced with actual call
to power on/off the cores at a later point.

This warranted the addition of a new am62l_psci driver for the AM62L
family of devices. For further details of how this new device looks
like, refer to the TRM [1].

[1] https://www.ti.com/lit/pdf/sprujb4

Change-Id: Ic53096e7bbc25fa55386ac4b6bef364dd6d0cf3b
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

refactor(ti): move out k3/common to ti/common

This will allow us to support more platforms that share commonalities
like the k3_gicv3, console, helpers, etc.
With this new common folder location, we

refactor(ti): move out k3/common to ti/common

This will allow us to support more platforms that share commonalities
like the k3_gicv3, console, helpers, etc.
With this new common folder location, we can move the previously
created ti_bl31_setup file into the new location so it can be shared
across multiple TI SoCs when need comes.
With this, also update all copyright dates.

Change-Id: Ie4365e32cd3b4b5870fe2cd03843400506e46265
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

refactor(ti): introduce ti_bl31_setup

There maybe a lot of reuse between platforms for bl31_setup in future
which may require us to have a common bl31 setup driver for TI.
The ti_bl31_setup is expec

refactor(ti): introduce ti_bl31_setup

There maybe a lot of reuse between platforms for bl31_setup in future
which may require us to have a common bl31 setup driver for TI.
The ti_bl31_setup is expected to contain all the reusable bits
and the soc_bl31_setup can have the custom soc init functionality
like the mmap regions and any special sequences needed at
device boot
While at it, also fix the path of the reference to the kernel doc
for booting.

Change-Id: Ie574f08cf3ba75362c45f85306499061ef89c964
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

refactor(ti): add top level ti_sci transport layer

The TI SCI transport header will allow us to abstract all transport
related calls from the actual medium of transport which can vary
from device to

refactor(ti): add top level ti_sci transport layer

The TI SCI transport header will allow us to abstract all transport
related calls from the actual medium of transport which can vary
from device to device.
For eg. it is sec proxy for current TI K3 devices, but in future it
maybe mailbox like in the TI AM62L.
With this change, we no longer need to include anything from the IPC
folder, so drop it from plat_common.mk

Change-Id: Ic99209688cf69f20e694e31e553ce4ec74254669
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

refactor(ti): move TI SCI and sec proxy to drivers

Prepare to support more devices and abstract TI SCI and it's transport
layers. This refactor will help keep things clean when new ipc drivers
get a

refactor(ti): move TI SCI and sec proxy to drivers

Prepare to support more devices and abstract TI SCI and it's transport
layers. This refactor will help keep things clean when new ipc drivers
get added.

Change-Id: I05673f379b1398c0b6a2bd9e1b5392165d12f151
Signed-off-by: Dhruva Gole <d-gole@ti.com>

show more ...

feat(ti): implement DM_MANAGED suspend

DM_MANAGED is the new mode to support a model where DM chooses the
suspend mode. Previously suspend always chose deep sleep. Now DM may
also choose other modes

feat(ti): implement DM_MANAGED suspend

DM_MANAGED is the new mode to support a model where DM chooses the
suspend mode. Previously suspend always chose deep sleep. Now DM may
also choose other modes like IO+DDR or other modes in the future. DM has
more knowledge about constraints regarding wakeup latency or which parts
of the system need to work continuously, so it is better able to decide
what suspend mode should be used.

To support DM_MANAGED in TF-A the next system mode needs to be fetched
and passed to the enter sleep message. The new ti_sci command to fetch
the next system mode is added which is ti_sci_lpm_get_next_sys_mode().

DM_MANAGED is only enabled if the firmware capability is supported.

Change-Id: I7244e27793cc60e1673271990d7cd840294d6cc3
Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>

show more ...