GitLab

This patch addds multi console interface for ZynqMP
platform

Change-Id: I508a61412df2b71d04bca6a1139c8f32cbd7dccd
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Tested-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>

The new API becomes the default one.

Change-Id: Ic1d602da3dff4f4ebbcc158b885295c902a24fec
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>

The default implementations are defined in crash_console_helpers.S. The
platforms have to define plat_crash_console_*.

Implemented placeholders for platforms that were missing helpers.

Change-Id: Iea60b6f851956916e421dfd8c34a62d96eb9148e
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>

Allow EL2 (e.g. U-Boot) to load the configuration object at runtime
into the Xilinx ZynqMP PMU firmware. This allows booting with U-Boot
and U-Boot SPL with PMU FW without hard-coding the configuration
object.
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>

Many parts of the code were duplicating symbols that are defined in
include/common/bl_common.h. It is better to only use the definitions in
this header.

As all the symbols refer to virtual addresses, they have to be
uintptr_t, not unsigned long. This has also been fixed in bl_common.h.

Change-Id: I204081af78326ced03fb05f69846f229d324c711
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

PLAT_ARM_NS_IMAGE_OFFSET is in fact not an offset relative to some base
address, it is an absolute address. Rename it to avoid any confusion.

Change-Id: I1f7f5e8553cb267786afe7e5f3cd4d665b610d3f
Signed-off-by: Sandrine Bailleux <sandrine.bailleux@arm.com>

Use full include paths like it is done for common includes.

This cleanup was started in commit d40e0e08283a ("Sanitise includes
across codebase"), but it only cleaned common files and drivers. This
patch does the same to Arm platforms.

Change-Id: If982e6450bbe84dceb56d464e282bcf5d6d9ab9b
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

pm_service ipi functions can be used by other xilinx
platforms. So move it to common directory. Also change
node_id member type in pm_proc structure so it can be
used for versal where device IDs are used instead of
node IDs.

To accommodate this change header files are re-organized.
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

ZynqMP pm_service ipi file uses platform specific macros names.
pm_service ipi functions can be used by other Xilinx platforms
also. Make rename macros to use generic names so that it can be
used by common file.

pm_service ipi functions will be moved to common file in next patch.
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Use generic name for pm_api structure member, so that pm_api
structure can be used other Xilinx platforms.

Structure definition will be moved to common file in upcoming
patch.
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

ipi_mb functions can be used for other Xilinx platform.
Move it to common file from platform specific files.
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

ipi_mb function uses platform specific ipi configuration table.
These ipi_mb functions can be used for other Xilinx platform.
So, instead of using direct data structure, initialize IPI
configuration data by passing platform specific ipi table.
Macros are updated accordingly for this ipi table change.

This change is done so that ipi_mb functions can be moved to
common file without major changes. All common functions now would
be moved to common file in next patch.
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Move ipi configuration structure definition to common header file
and rename it to common name, so it can be used for Xilinx specific
other platforms in upcoming changes.
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Move zynqmp_ipi.h to platform specific include directory.
Rename it to plat_ipi.h instead of platform name. So, it can
be used to common source files which needs platform specific
data.
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Move zynqmp_def.h to platform specific include directory.
Also, update source file to include header file from updated
path of zynqmp_def.h
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Move zynqmp_private.h to platform specific include directory.
Also, rename it to plat_private.h instead of having platform
name. So, it can be used to common source files which needs
platform specific data.
Signed-off-by: Tejas Patel <tejas.patel@xilinx.com>
Reviewed-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

This clock does not drive any clock in LPD so there is no need for
Linux to try to initialize it.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Linux has no reason to use these system and debug clocks and therefore
shouldn't access them. These clocks are marked as invalid in order to
prevent Linux from registering and querying them.

Note that despite clocks being marked as invalid a security issue
still remains in place as there is nothing that prevents the
non-secure world from gating these clocks and that way causing
damage to the system.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

These clocks are marked as invalid in order to prevent Linux from
registering them.

Note that despite clocks being marked as invalid a security issue
still remains in place as there is nothing that prevents the
non-secure world from gating these clocks and that way halt
the whole APU subsystem.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

In the existing model for ACPU clock the mux, divider, and gate were
represented as one clock and ACPU_HALF was modelled as child of
ACPU clock. This is not correct. ACPU clock model contains only
mux and the divider, and it has 2 children: ACPU_FULL and ACPU_HALF
clocks which have only gates. The models of ACPU and ACPU_HALF clocks
are fixed and ACPU_FULL clock is added.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Clock get parent EEMI API is reimplemented to use system-level clock
and pll EEMI APIs rather than direct MMIO read/write accesses to clock
and pll control registers.
Since linux still uses clock set parent API to get pre_src, post_src, div2
and bypasss, in the implementation of pm_clock_get_parent() we need to
workaround this by distinguishing two cases:
1) if the given clock ID corresponds to a PLL-related clock ID (*_PRE_SRC,
   *_POST_SRC, *_INT_MUX or *_PLL clock IDs); or
2) given clock ID is truly an on-chip clock.
For case 1) we'll map the call onto PLL-specific EEMI API with the
respective parameter ID. For case 2) the call is passed to the PMU.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Clock set parent EEMI API is reimplemented to use system-level clock
and pll EEMI APIs rather than direct MMIO read/write accesses to clock
and pll control registers.
Since linux still uses clock set parent API to set pre_src, post_src, div2
and bypass, in the implementation of pm_clock_set_parent() we need to
workaround this by distinguishing two cases:
1) if the given clock ID corresponds to a PLL-related clock ID (*_PRE_SRC,
*_POST_SRC, *_INT_MUX or *PLL clock IDs); or 2) given clock ID is truly
an on-chip clock.
For case 1) we'll map the call onto PLL set parameter EEMI API with the
respective parameter ID. Since clock set parent interface to EL1/2 receives
parent index (mux select value), the value is just passed to PMU.
Functions that appear to be unused after this change is made are removed.

Setting the parent of *PLL clocks, that actually model bypass, is not
possible. This is already ensured by the existing clock model having the
CLK_SET_RATE_NO_REPARENT flag. The API also doesn't allow changing the
bypass parent. Bypass is controlled only by the PMU firmware.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Clock set/get rate are not implemented and will likely never be.
Remove empty function stubs.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Clock get divider EEMI API is reimplemented to use system-level clock
get divider EEMI API rather than direct MMIO read/write accesses to clock
control registers.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Clock set divider EEMI API is reimplemented to use system-level clock
set divider EEMI API rather than direct MMIO read/write accesses to clock
control registers.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Clock get state EEMI API is reimplemented to use system-level clock
and pll EEMI APIs rather than direct MMIO read/write accesses to clock
and pll control registers.
Since linux is_enabled method for PLLs still uses clock get state API
get the PLL state, in the implementation of pm_clock_getstate() we need
to workaround this by distinguishing two cases: 1) if the given clock ID
corresponds to a PLL output clock ID; or 2) given clock ID is truly an
on-chip clock whose state of the gate should be returned.
For case 1) we'll call pm_api_clock_pll_getstate() implemented in
pm_api_clock.h/c. This function will query the PLL state from PMU using
the system-level PLL get mode EEMI API.
For case 2) we'll call the PMU to query the clock gate state using
system-level clock get status EEMI API.
Functions that appear to be unused after this change is made are removed.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Clock disable EEMI API is reimplemented to use system-level clock
and pll EEMI APIs rather than direct MMIO read/write accesses to clock
and pll control registers.
Since linux still uses clock disable API to reset the PLL in the
implementation of pm_clock_disable() we need to workaround this by
distinguishing two cases: 1) if the given clock ID corresponds to a PLL
output clock ID; or 2) given clock ID is truly an on-chip clock that can
be gated.
For case 1) we'll call pm_api_clock_pll_disable() implemented in
pm_api_clock.h/c. This function will reset the PLL using the system-level
PLL set mode EEMI API with the reset mode argument.
For case 2) we'll call the PMU to configure the clock gate. This is done
using system-level clock disable EEMI API.
Functions that appear to be unused after this change is made are removed.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Clock enable EEMI API is reimplemented to use system-level clock
and pll EEMI APIs rather than direct MMIO read/write accesses to clock
and pll control registers.
Since linux still uses clock enable API to trigger locking of the PLLs
in the pm_clock_enable() implementation we need to workaround this by
distinguishing two cases: 1) if the given clock ID corresponds to a PLL
output clock ID; or 2) given clock ID is truly an on-chip clock that can
be gated.
For case 1) we'll call pm_api_clock_pll_enable() implemented in
pm_api_clock.h/c. This function checks what is the buffered PLL mode and
calls the system-level PLL set mode EEMI API with the buffered mode value
specified as argument. Long term, if linux driver get fixed to use PLL
EEMI API to control PLLs, this case could be removed from ATF.
For case 2) we'll call the PMU to configure the clock gate. This is done
using system-level clock enable EEMI API.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

When linux calls pm_ioctl_get_pll_frac_mode() it doesn't expect the actual
mode to be read from hardware, but the value that it is intending to
program. Therefore, we return the buffered value to linux.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

When linux calls pm_ioctl_set_pll_frac_mode() it doesn't expect the
fractional mode to be changed in hardware. Furthermore, even before this
patch setting the mode which is done by writing into register takes
no effect until the PLL reset is deasserted, i.e. until linux "enables"
the PLL. To adjust the code to system-level PLL EEMI API and avoid
unnecessary IPIs that would otherwise be issued, we buffer the mode
value set via IOCTL until the PLL mode really needs to be set.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Fractional data should be set using PLL set parameter EEMI API. This
stands for system-level communication (APU to PMU). Since linux
already uses a specific IOCTL function to do this and we need to
keep it that way, the pll clock ID given by linux has to be mapped
to the pll node ID that is communicated at the system-level (argument
of PLL set parameter API).
With this modification the function pm_api_clk_set_pll_frac_data is
removed from pm_api_clock.c/h because it became unused.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Fractional data should be get using PLL get parameter EEMI API. This
stands for system-level communication (APU to PMU). Since linux
already uses a specific IOCTL function to do this and we need to
keep it that way, the pll clock ID given by linux has to be mapped
to the pll node ID that is communicated at the system-level (argument
of PLL get parameter API).
With this modification the function pm_api_clk_get_pll_frac_data is
removed from pm_api_clock.c/h because it became unused.
The clock enum is defined as 'enum clock_id'.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

This API will be used to get the currently configured PLL mode:
reset (bypassed and unlocked), integer or fractional (locked).
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

This API will be used to set the PLL mode: reset (unlocked), integer
or fractional (locked). If reset mode is set the PM controller will
bypass the target PLL prior to asserting the reset. If integer or
fractional mode is set the PM controller will program and trigger
locking of the PLL. If success status is returned the PLL is locked
and its bypass is deasserted. If fractional mode is set the fractional
divider (data parameter) has to have a non-zero value prior to issuing
pll set fractional mode. The caller need to ensure that the data
parameter is properly set using pll get/set parameter EEMI API.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

This API will be used to get a parameter for the PLL. Parameter values
represent the values as defined in the Zynq MPSoC register reference
manual ug1087.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

This API will be used to set a parameter for the PLL. The parameter
value that is set will have effect once the PLL mode is set to integer
or fractional mode. Parameter values represent the values as defined
in the Zynq MPSoC register reference manual ug1087.
Signed-off-by: Mirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: Will Wong <WILLW@xilinx.com>
Signed-off-by: Jolly Shah <jollys@xilinx.com>

Enforce full include path for includes. Deprecate old paths.

The following folders inside include/lib have been left unchanged:

- include/lib/cpus/${ARCH}
- include/lib/el3_runtime/${ARCH}

The reason for this change is that having a global namespace for
includes isn't a good idea. It defeats one of the advantages of having
folders and it introduces problems that are sometimes subtle (because
you may not know the header you are actually including if there are two
of them).

For example, this patch had to be created because two headers were
called the same way: e0ea0928 ("Fix gpio includes of mt8173 platform
to avoid collision."). More recently, this patch has had similar
problems: 46f9b2c3 ("drivers: add tzc380 support").

This problem was introduced in commit 4ecca339

 ("Move include and
source files to logical locations"). At that time, there weren't too
many headers so it wasn't a real issue. However, time has shown that
this creates problems.

Platforms that want to preserve the way they include headers may add the
removed paths to PLAT_INCLUDES, but this is discouraged.

Change-Id: I39dc53ed98f9e297a5966e723d1936d6ccf2fc8f
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

Xilinx is introducing Versal, an adaptive compute acceleration platform
(ACAP), built on 7nm FinFET process technology. Versal ACAPs combine Scalar
Processing Engines, Adaptable Hardware Engines, and Intelligent Engines with
leading-edge memory and interfacing technologies to deliver powerful
heterogeneous acceleration for any application. The Versal AI Core series has
five devices, offering 128 to 400 AI Engines. The series includes dual-core Arm
Cortex-A72 application processors, dual-core Arm Cortex-R5 real-time
processors, 256KB of on-chip memory with ECC, more than 1,900 DSP engines
optimized for high-precision floating point with low latency.

This patch adds Virtual QEMU platform support for
this SoC "versal_virt".
Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

All identifiers, regardless of use, that start with two underscores are
reserved. This means they can't be used in header guards.

The style that this project is now to use the full name of the file in
capital letters followed by 'H'. For example, for a file called
"uart_example.h", the header guard is UART_EXAMPLE_H.

The exceptions are files that are imported from other projects:

- CryptoCell driver
- dt-bindings folders
- zlib headers

Change-Id: I50561bf6c88b491ec440d0c8385c74650f3c106e
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

This function is not related to Arm platforms and can be reused by other
platforms if needed.

Change-Id: Ia9c328ce57ce7e917b825a9e09a42b0abb1a53e8
Co-authored-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>