Before this patch the configuration took place only for CP0 and CP1, but
since new platforms can contains up to 3 CPs update is required.
Change-Id: Iebd50bbe7b9772063e2c4efb3a7ecbfd593e950d
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
The Marvell Armada 37xx SoCs-based platforms contain a bit
awkward directory structure because the currently only one
supported PLAT and PLAT_FAMILY are the same. Modify the latter
to 'a3k' in order to improve it and keep plat/marvell/armada
tree more consistent:
plat/marvell/
├── armada
│ ├── a3k
│ │ ├── a3700
[...]
│ ├── a8k
│ │ ├── a70x0
[...]
Change-Id: I693a6ef88e6ce49a326a3328875c90bbc186066a
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
Change a topology map from internal database
to SPD based for 32bit bus width mode
Change-Id: I803166893ddc2fd916fc8a1c27fffd34b6ec0c72
Signed-off-by: Alex Leibovich <alexl@marvell.com>
Remove pll powerdown from pll configuration sequence to improve
stability. Remove redundant cases, which no longer exist.
Also get rid of irrelevant definition of CPU_2200_DDR_1200_RCLK_1200,
which is not used by 806/807.
Change-Id: If911e7dee003dfb9a42fafd7ffe34662f026fd23
Signed-off-by: Alex Leibovich <alexl@marvell.com>
This commit introduces 32-bit DDR topology map initialization.
For that purpose a new DDR32 build flag is added, with
according documentation update.
Change-Id: I169ff358c2923afd984e27bc126dc551dcaefc01
Signed-off-by: Alex Leibovich <alexl@marvell.com>
This commit introduces two changes:
- remove hardcoded references to mci0 from the driver
- perform mci optimization for all mci interfaces
It fixes performance issues observed on cn9132 CP2.
Change-Id: I4e040cd54ff95c9134035ac89b87d8feb28e9eba
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
The mci_initialize function name was misleading. The function itself
doesn't initialize MCI in general but performs MCI link tuning for
performance improvement.
Change-Id: I13094ad2235182a14984035bbe58013ebde84a7e
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
Snoop filter needs to be enabled once per cluster.
Change-Id: I241e72f21982142ba290c7547df6f434e6a6a98d
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
By default all external CPs start with configuration address space set to
0xf200_0000. To overcome this issue, go in the loop and initialize the
CP one by one, using temporary window configuration which allows to access
each CP and update its configuration space according to decoding
windows scheme defined for each platform.
In case of cn9130 after this procedure bellow addresses will be used:
CP0 - f2000000
CP1 - f4000000
CP2 - f6000000
When the re-configuration is done there is need to restore previous
decoding window configuration(init_io_win).
Change-Id: I1a652bfbd0bf7106930a7a4e949094dc9078a981
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
As a preparation for upcoming support for CN9130 platform, which is
classified as OcteonTx2 product but inherits functionality from a8k,
allow to use a8k_common.mk and mss_common.mk from outside of
PLAT_FAMILY_BASE.
Above is done by introducing BOARD_DIR which needs to be set by each
platform, before including a8k_common.mk and mss_common.mk. This will
allow to use mentioned mk files not only for platforms located under
previously defined PLAT_FAMILY_BASE.
Change-Id: I22356c99bc0419a40ae11e42f37acd50943ea134
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
This commit is a preparation for upcoming support for OcteonTX and
OcteonTX2 product families. Armada platform related files (docs,
plat, include/plat) are moved to the new "armada" sub-folder.
Change-Id: Icf03356187078ad6a2e56c9870992be3ca4c9655
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
BL31_CACHE_DISABLE flag was introduced as a work-around
for the older SoC revisions. Since it is not relevant in the
newest versions, toggle it to be disabled by default.
One can still specify it by adding 'BL31_CACHE_DISABLE=1'
string to the build command.
Change-Id: I11b52dade3ff7f8ee643b8078c6e447c45946570
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
This patch moves all GICv3 driver files into new added
'gicv3.mk' makefile for the benefit of the generic driver
which can evolve in the future without affecting platforms.
The patch adds GICv3 driver configuration flags
'GICV3_IMPL', 'GICV3_IMPL_GIC600_MULTICHIP' and
'GICV3_OVERRIDE_DISTIF_PWR_OPS' described in
'GICv3 driver options' section of 'build-option.rst'
document.
NOTE: Platforms with GICv3 driver need to be modified to
include 'drivers/arm/gic/v3/gicv3.mk' in their makefiles.
Change-Id: If055f6770ff20f5dee5a3c99ae7ced7cdcac5c44
Signed-off-by: Alexei Fedorov <Alexei.Fedorov@arm.com>
This patch provides separation of GICD, GICR accessor
functions and adds new macros for GICv3 registers access
as a preparation for GICv3.1 and GICv4 support.
NOTE: Platforms need to modify to include both
'gicdv3_helpers.c' and 'gicrv3_helpers.c' instead of the
single helper file previously.
Change-Id: I1641bd6d217d6eb7d1228be3c4177b2d556da60a
Signed-off-by: Alexei Fedorov <Alexei.Fedorov@arm.com>
Now that different UARTs share the same console_t struct, we can
simplify the console selection for the Marvell platforms:
We share the same console_t pointers, just change the name of the
console register functions, depending on the selected platform.
Change-Id: I6fe3e49fd7f208a9b3372c5deef43236a12867bc
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Since now the generic console_t structure holds the UART base address as
well, let's use that generic location and drop the UART driver specific
data structure at all.
Change-Id: I89c3ab2ed85ab941d8b38ced48474feb4aaa8b7e
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Since now the generic console_t structure holds the UART base address as
well, let's use that generic location and drop the UART driver specific
data structure at all.
Change-Id: I5c2fe3b6a667acf80c808cfec4a64059a2c9c25f
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Extend possible images to 8, additionaly add another type which will be
used with platform containing up to 3 CPs.
Change-Id: Ib68092d11af9801e344d02de839f53127e056e46
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
In order to access MG SRAM, the amb bridge needs to be configured which is
done in bl2 platform init.
For MG CM3, the image is only loaded to its SRAM and the CM3 itself is
left in reset. It is because the next stage bootloader (e.g. u-boot)
will trigger action which will take it out of reset when needed. This
can happen e.g. when appropriate device-tree setup (which has enabled
802.3 auto-neg) will be chosen. In other cases the MG CM3 should not be
running.
Change-Id: I816ea14e3a7174eace068ec44e3cc09998d0337e
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
Implements a way to add platform specific power off code to a
Marvell Armada 8K platform.
Marvell Armada 8K boards can now add a board/system_power.c file
that contains a system_power_off() function.
This function can now send a command to a power management MCU or
other board periferals before shutting the board down.
Signed-off-by: Luka Kovacic <luka.kovacic@sartura.hr>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Change-Id: Iaba20bc2f603195679c54ad12c0c18962dd8e3db
---
I am working on a device that will be ported later, which has a
custom power management MCU that handles LEDs, board power and fans
and requires this separation.
Introduce timeout_init_us/timeout_elapsed() delay tracking with CNTPCT.
timeout_init_us(some_timeout_us); returns a reference to detect
timeout for the provided microsecond delay value from current time.
timeout_elapsed(reference) return true/false whether the reference
timeout is elapsed.
Cherry picked from OP-TEE implementation [1].
[1] commit 33d30a74502b ("core: timeout detection support")
Minor:
- Remove stm32mp platform duplicated implementation.
- Add new include in marvell ble.mk
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Lionel Debieve <lionel.debieve@st.com>
Change-Id: Iaef6d43c11a2e6992fb48efdc674a0552755ad9c
NOTE: __ASSEMBLY__ macro is now deprecated in favor of __ASSEMBLER__.
All common C compilers predefine a macro called __ASSEMBLER__ when
preprocessing a .S file. There is no reason for TF-A to define it's own
__ASSEMBLY__ macro for this purpose instead. To unify code with the
export headers (which use __ASSEMBLER__ to avoid one extra dependency),
let's deprecate __ASSEMBLY__ and switch the code base over to the
predefined standard.
Change-Id: Id7d0ec8cf330195da80499c68562b65cb5ab7417
Signed-off-by: Julius Werner <jwerner@chromium.org>
When the link goes down (e.g. during a retrain), accessing the device
configuration space can trigger an ARM64 SError interrupt. Such
conditions cannot be predicted, so to avoid a crash the SError is
ignored.
Signed-off-by: Remi Pommarel <repk@triplefau.lt>
Change-Id: I2b1fd3296cc1c88b9ca1fe21c0924cb324eed58d
This consists of ensuring that the left operand of each shift is
unsigned when the operation might overflow into the sign bit.
Change-Id: I78f386f5ac171d6e52383a3e42003e6fb3e96b57
Signed-off-by: Justin Chadwell <justin.chadwell@arm.com>
The argument passed via x1 is used as a base address for comphy related
routines. Nevertheless validation of this address wasn't good enough and
allowed some non comphy related addresses to slip over.
To overcome this issue make sure that the address passed via SMC points
to valid CP range and allow to proceed comphy initializations only with
correct comphy offset.
This could be fixed in a different way e.g. by passing CP id from the
caller, but since this API is already used with various Linux, U-Boot
and UEFI versions it can't be changed.
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
Change-Id: Ia74dbc36efcfbefc4a102d31191e6af5808c4a82
Also, update platform_def.h guidelines about includes in the porting
guide.
Change-Id: I1ae338c9dd3242b309f6d53687ba2cb755d488c3
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
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>
The definitions in bl1/bl1_private.h and bl2/bl2_private.h are useful for
platforms that may need to access them.
Change-Id: Ifd1880f855ddafcb3bfcaf1ed4a4e0f121eda174
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
The macro has n > 32.
It has to be n > 31 since GPIO 0-31 are on Register 0 and 32-63 on Register 1.
Signed-off-by: Sven Auhagen <sven.auhagen@voleatech.de>
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: e0ea0928d5 ("Fix gpio includes of mt8173 platform
to avoid collision."). More recently, this patch has had similar
problems: 46f9b2c3a2 ("drivers: add tzc380 support").
This problem was introduced in commit 4ecca33988 ("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>
This patch makes the build system link the console framework code by
default, like it already does with other common libraries (e.g. cache
helpers). This should not make a difference in practice since TF is
linked with --gc-sections, so the linker will garbage collect all
functions and data that are not referenced by any other code. Thus, if a
platform doesn't want to include console code for size reasons and
doesn't make any references to console functions, the code will not be
included in the final binary.
To avoid compatibility issues with older platform ports, only make this
change for the MULTI_CONSOLE_API.
Change-Id: I153a9dbe680d57aadb860d1c829759ba701130d3
Signed-off-by: Julius Werner <jwerner@chromium.org>
Very rarely, during cpuidle operations the following error
is seen: "PM MSG Trigger Timeout".
This is caused by slow handling of message interrutps
in the PM FW running on CM3 (under heavy PM operation load).
This is not a real issue, so we extend the timeout to
avoid the error prints.
Change-Id: I92fd6f2ff1ddf208b216c123880ded28a00b6e0e
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/59670
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Add static configuration for SFI+ 10Gbps interface on SERDES
Lane 4.
This is just a copy of Lane 2 static values, not optimized.
Board-to-board iperf test shows up to 6Gbps transfer speed.
Change-Id: I024d2ac132f7fa6c342a64367f3dca2123a27e97
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-by: Igal Liberman <igall@marvell.com>
- Number of open power switches for CPUs should be three
and now two.
- This patch updates the value of open power switches from
0xfd (two power-switches) to 0xfc (three power-switches).
Change-Id: I2783ab7f04bbbb6da78eeedcabe4636f9a774512
Signed-off-by: Christine Gharzuzi <chrisg@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Add support for SVC test builds for tuning AVS values.
Update the SVC procedure and add EEPROM access.
Add support for AP807 AVS values (10 bits wide).
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
- Update CPU frequency on AP807 to 2GHz for SAR 0x0.
- Increase AVS to 0.88V for 2GHz clock
Change-Id: Ic945b682ab2f8543e34294bfc56c3eae2c5e0c8e
Signed-off-by: Christine Gharzuzi <chrisg@marvell.com>
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
In Armada 8K DB boards, PCIe initialization can be executed only once
because PCIe reset performed during chip power on and it cannot be
executed via GPIO later.
This means that power on can be executed only once, when it's called
from the bootloader.
Power on:
Read bit 21 of the mode, it marks if the caller is
the bootloader or the Linux Kernel.
Power off:
Check if the comphy was already configured to PCIe, if yes,
check if the caller is bootloader, if both conditions are true
(PCIe mode and called by Linux) - skip the power-off.
In addition, fix incorrect documentation describing mode fields -
PCIe width is 3 bits, not 2.
NOTE: with this patch, please use LK4.14.76 (LK4.4.120 may not work
with it).
Change-Id: I4b929011f97a0a1869a51ba378687e78b3eca4ff
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-by: Grzegorz Jaszczyk <jaz@semihalf.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Do not power off the CPU1 since there is no way to wake it up
(wake-up is causing CPU0 reset as well duo to HW bug). Quote from errata
Ref #13 [In power saving mode, both cores must be powered off]:
"When Core 0 is on and Core 1 is in power-off state, a Core 1
wake-up resets Core 0 as well and puts Core 0 back to ROM".
To overcome described HW bug instead of powering the CPU off, let it
reach WFI instruction, which is invoked by generic psci_do_cpu_off
function after platform handler finishes. This will put the core in low
power state and give a chance to wake it up.
Before this change, after running secondary kernel via kexec, only one
core was up, now both cores are up.
Change-Id: I87f144867550728055d9b8a2edb84a14539acab7
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Migrate Marvell platforms from legacy console API to
multi-console API.
Change-Id: I647f5f49148b463a257a747af05b5f0c967f267c
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
According to "openssl" manual:
-K key
The actual key to use: this must be represented as a string
comprised only of hex digits. If only the key is specified,
the IV must additionally specified using the -iv option.
When both a key and a password are specified, the key given
with the -K option will be used and the IV generated from the
password will be taken.
It does not make much sense to specify both key and password.
This patch removes "-k 0" parameter from the encryption command
since we are already using "-K" and "-iv" for the key and IV.
Change-Id: Ia333cedaa3207e643c95d2ec7c229f50eeab96db
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/60745
Reviewed-by: Igal Liberman <igall@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Sharon Habet <sharonh@marvell.com>
Grzegorz Jaszczyk
Marcin Wojtas
Alex Leibovich
Alexei Fedorov
Andre Przywara
Deepika Bhavnani
Luka Kovacic
Lionel Debieve
Julius Werner
Soby Mathew
Remi Pommarel
Justin Chadwell
Ambroise Vincent
Antonio Nino Diaz
Sven Auhagen
Konstantin Porotchkin
Igal Liberman
Christine Gharzuzi