UART_IMAGE is not set when WTP isn't. The error rules will then provide
a recipe for $(BUILD_PLAT). When building with a lot of cores (64) this
rule might be called before the directory is made, causing a build
failure.
Hoist the definition so that the depended path is correct.
Change-Id: I167e7398e576e667d0c5c1fc0f07ab8c8ef939a8
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
To allow for generic handling of a wakeup, this hook is no longer
expected to call wfi itself. Update the name everywhere to reflect this
expectation so that future platform implementers don't get misled.
Change-Id: Ic33f0b6da74592ad6778fd802c2f0b85223af614
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
This commit streamlines directory creation by introducing a single
pattern rule to automatically make directories for which there is a
dependency.
We currently use several macros to generate rules to create directories
upon dependence, which is a significant amount of code and a lot of
redundancy. The rule introduced by this change represents a catch-all:
any rule dependency on a path ending in a forward slash is automatically
created.
Now, rules can rely on an unordered dependency (`|`) on `$$(@D)/` which,
when secondary expansion is enabled, expands to the directory of the
target being built, e.g.:
build/main.o: main.c | $$(@D)/ # automatically creates `build/`
Change-Id: I7e554efa2ac850e779bb302fd9c7fbb239886c9f
Signed-off-by: Chris Kay <chris.kay@arm.com>
This change introduces a few helper variables for dealing with verbose
and silent build modes: `silent`, `verbose`, `q` and `s`.
The `silent` and `verbose` variables are boolean values determining
whether the build system has been configured to run silently or
verbosely respectively (i.e. with `--silent` or `V=1`).
These two modes cannot be used together - if `silent` is truthy then
`verbose` is always falsy. As such:
make --silent V=1
... results in a silent build.
In addition to these boolean variables, we also introduce two new
variables - `s` and `q` - for use in rule recipes to conditionally
suppress the output of commands.
When building silently, `s` expands to a value which disables the
command that follows, and `q` expands to a value which supppresses
echoing of the command:
$(s)echo 'This command is neither echoed nor executed'
$(q)echo 'This command is executed but not echoed'
When building verbosely, `s` expands to a value which disables the
command that follows, and `q` expands to nothing:
$(s)echo 'This command is neither echoed nor executed'
$(q)echo 'This command is executed and echoed'
In all other cases, both `s` and `q` expand to a value which suppresses
echoing of the command that follows:
$(s)echo 'This command is executed but not echoed'
$(q)echo 'This command is executed but not echoed'
The `s` variable is predominantly useful for `echo` commands, where you
always want to suppress echoing of the command itself, whilst `q` is
more useful for all other commands.
Change-Id: I8d8ff6ed714d3cb401946c52955887ed7dca602b
Signed-off-by: Chris Kay <chris.kay@arm.com>
Refactor console_flush() and console_switch_state(CONSOLE_FLAG_RUNTIME)
to bl31_main(). This has been done per the recommendation in TF-A
mailing list. These calls need to be the last calls, after any runtime
initialization has been done, before BL31 exits.
All platforms that override the generic implementation of
bl31_plat_runtime_setup() have been refactored. The console_flush()
and console_switch_state() calls have been removed as they become
part of bl31_main() function.
Any platform that don't need to make any change to the generic (weak)
implementation of bl31_plat_runtime_setup() don't need to override it
in their platforms.
Change-Id: I6d04d6daa9353daeaa7e3df9e9adf6f322a917b8
Signed-off-by: Salman Nabi <salman.nabi@arm.com>
Flush the FIFO before switching to runtime. This is so that there are
no lingering chars in the FIFO when we move to the runtime console.
TF-A plans to refactor the console_Switch_state(CONSOLE_FLAG_RUNTIME)
and console_flush() calls and make them the last calls in bl31_main()
(before BL31 exits). Until then they are being left as the last calls
in bl31_plat_runtime_setup() for testing before refactoring.
This patch only affects the Armada SoC of Marvell's platform.
Change-Id: I7082fdb8c5507cd1ce5915d67e61e638605982e0
Signed-off-by: Salman Nabi <salman.nabi@arm.com>
Add code that acknowledges all SoC interrupts and resets the Generic
Interrupt Controller before resetting the SoC via the Cortex-M3 secure
coprocessor.
Recall that Turris MOX has a HW bug wherein a SoC reset initiated by
writing the magic value to the North Bridge Warm Reset register may
randomly freeze the board.
Back in 2021 we introduced the CM3_SYSTEM_RESET build option for the
Armada 3700 platform, which, when enabled, adds code to the PSCI reset
handler so that the SoC reset is done by requesting the firmware in the
Cortex-M3 secure coprocessor to do it, instead of writing the Warm Reset
register.
The secure coprocessor firmware tried various things to put the board
into a state where the SoC reset circuit would work correctly. This
managed to fix the issue for some boards, but not for all of them.
Another considered method to overcome this issue was to reset all the
SoC peripheral controllers one by one by writing to specific registers,
instead of triggering the SoC reset circuit via the Warm Reset register.
This method was not used because until now, there was one peripheral
that I could not find a way how to reset properly: the Generic Interrupt
Controller (GIC).
After 3 years I have finally found a way how to reset the GIC, and it
needs to be done by the main processor, before the secure coprocessor
resets the main processor.
Change-Id: Icc23251ef97738b6b48af514d5118440ec21cdd7
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Some of our specialized sections are not prefixed with the conventional
period. The compiler uses input section names to derive certain other
section names (e.g. `.rela.text`, `.relacpu_ops`), and these can be
difficult to select in linker scripts when there is a lack of a
delimiter.
This change introduces the period prefix to all specialized section
names.
BREAKING-CHANGE: All input and output linker section names have been
prefixed with the period character, e.g. `cpu_ops` -> `.cpu_ops`.
Change-Id: I51c13c5266d5975fbd944ef4961328e72f82fc1c
Signed-off-by: Chris Kay <chris.kay@arm.com>
The following build system variables have been renamed:
- `LINKERFILE` -> `DEFAULT_LINKER_SCRIPT`
- `BL_LINKERFILE` -> `DEFAULT_LINKER_SCRIPT_SOURCE`
- `<IMAGE>_LINKERFILE` -> `<IMAGE>_DEFAULT_LINKER_SCRIPT_SOURCE`
These new names better reflect how each variable is used:
1. the default linker script is passed via `-dT` instead of `-T`
2. linker script source files are first preprocessed
Additionally, linker scripts are now placed in the build directory
relative to where they exist in the source directory. For example,
the `bl32/sp_min/sp_min.ld.S` would now preprocess to
`sp_min/sp_min.ld` instead of just `bl32.ld`
BREAKING-CHANGE: The `LINKERFILE`, `BL_LINKERFILE` and
`<IMAGE_LINKERFILE>` build system variables have been renamed. See the
commit message for more information.
Change-Id: If8cef65dcb8820e8993736702c8741e97a66e6cc
Signed-off-by: Chris Kay <chris.kay@arm.com>
Current RAS framework in TF-A only supports handling errors originating
from NS world but the HANDLE_EA_EL3_FIRST flag configures it for all
lower Els. To make the current design of RAS explicit, rename this macro
to HANDLE_EA_EL3_FIRST_NS and set EA bit in scr_el3 only when
switching to NS world.
Note: I am unaware of any platform which traps errors originating in
Secure world to EL3, if there is any such platform then it need to
be explicitly implemented in TF-A
Signed-off-by: Manish Pandey <manish.pandey2@arm.com>
Change-Id: If58eb201d8fa792c16325c85c26056e9b409b750
get_el_str() was implemented under ENABLE_BACKTRACE macro but being
used at generic places too, this causes multiple definition of this
function.
Remove duplicate definition of this function and move it out of
backtrace scope. Also, this patch fixes a small bug where in default
case S-EL1 is returned which ideally should be EL1, as there is no
notion of security state in EL string.
Signed-off-by: Manish Pandey <manish.pandey2@arm.com>
Change-Id: Ib186ea03b776e2478eff556065449ebd478c3538
Currently Tf-A uses whatever openssl binary is on the system to sign
images. However if OPENSSL_DIR is specified in the build flags this can
lead to linking issues as the system binary can end up being linked
against shared libraries provided in OPENSSL_DIR/lib if both binaries
(the system's and the on in OPENSSL_DIR/bin) are the same version.
This patch ensures that the binary used is always the one given by
OPENSSL_DIR to avoid those link issues.
Signed-off-by: Salome Thirot <salome.thirot@arm.com>
Change-Id: Ib534e06ebc8482e4391e376d3791a87968de4a99
This allows TF-A's a3700_system_reset() function to try Warm reset
method when CM3 reset method is not implemented by WTMI firmware.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I7303197373e1a8ca5a44ba0b1e90b48855d6c0c3
A53 AP BootROM is just 16 kB long and is mapped to address range
0xFFFF0000-0xFFFF4000. RVBAR_EL3 register has value 0xFFFF0000.
A53 AP BootROM itself is in the BootROM window which is 1 MB long and
mapped to address range 0xFFF00000-0xFFFFFFFF.
CM3 BootROM is not accessible from A53 core.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I5d4a4c7b1e7550c4738c67a872d341f945d48bbc
These registers make it is possible to do external resets of A3700
peripherals. Most peripherals are reset by clearing a particular bit,
but some need setting the bit. Reflect this via "_N" suffix in macro
names.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: Iacef5e671746b831b5beea9e4fdcc59d8de84edc
Use long instead of long long on aarch64 for 64_t stdint types.
Introduce inttypes.h to properly support printf format specifiers for
fixed width types for such change.
Change-Id: I0bca594687a996fde0a9702d7a383055b99f10a1
Signed-off-by: Scott Branden <scott.branden@broadcom.com>
'#' needs to be before TAB, otherwise comment is printed on stdout during build.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I502374ef35d91e194dc35b78d31d6884a466fab2
This change adds just comments why some checks are required. They check
that ENV variables and external repos are correctly set for TF-A builds.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I2f8af5061411c0c92d3875917f4d97b60dc2cf10
Add support for Globalscale MOCHAbin board.
Its based on Armada 7040 SoC and ships in multiple DRAM options:
* 2GB DDR4 (1CS)
* 4GB DDR4 (1CS)
* 8GB DDR4 (2CS)
Since it ships in multiple DRAM configurations, an
Armada 3k style DDR_TOPOLOGY variable is added.
Currently, this only has effect on the MOCHAbin, but
I expect more boards with multiple DRAM sizes to be
supported.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Change-Id: I8a1ec9268fed34f6a81c5cbf1e891f638d461305
The current Makefile assumes all TF-A binaries
have BL prefixes (BL1, BL2, etc). Now that we
have other binary names with FEAT_RME feature, remove
this assumption. With this change, we need to pass
the full name of a binary when using build macros.
Signed-off-by: Zelalem Aweke <zelalem.aweke@arm.com>
Change-Id: I44e094b2366aa526f807d92dffa709390d14d145
Erratum 1530924 affects Armada 37xx CPU, since it affects all Cortex-A53
revisions from r0p0 to r0p4.
Enable the workaround for this erratum.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Change-Id: I753225040e49e956788d5617cd7ce76d5e6ea8e8
This patch adds the necessary files to support
the SolidRun CN913X CEx7 Evaluation Board.
Because the DRAM connectivity and SerDes settings
is shared with the CN913X DB - reuse relevant
board-specific files.
Change-Id: I75a4554a4373953ca3fdf3b04c4a29c2c4f8ea80
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
The common makefile used by every a8k/cn913x platform
(a8k_common.mk) assumed default paths in PLAT_INCLUDES,
BLE/BL31_PORTING_SOURCES. Allow overriding those
variables, in order to avoid code duplication.
It can be helpful in case using multiple board variants
or sharing common settings between different platforms.
Change-Id: Idce603e44ed04d99fb1e3e11a2bb395d552e2bf7
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
It was enabled in commit 3c7dcdac5c ("marvell/a3700: Prevent SError
accessing PCIe link while it is down") with a workaround for a bug found
in U-Boot and Linux kernel driver pci-aardvark.c (PCIe controller driver
for Armada 37xx SoC) which results in SError interrupt caused by AXI
SLVERR on external access (syndrome 0xbf000002) and immediate kernel
panic.
Now when proper patches are in both U-Boot and Linux kernel projects,
this workaround in TF-A should not have to be enabled by default
anymore as it has unwanted side effects like propagating all external
aborts, including non-fatal/correctable into EL3 and making them as
fatal which cause immediate abort.
Add documentation for HANDLE_EA_EL3_FIRST build option into Marvell
Armada build section.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: Ic92b65bf9923505ab682830afb66c2f6cec70491
A3700 plat_ea_handler was introduced into TF-A codebase just because of
bugs in U-Boot and Linux kernel PCIe controller driver pci-aardvark.c.
These bugs were finally fixed in both U-Boot and Linux kernel drivers:
eccbd4ad8ehttps://git.kernel.org/stable/c/f18139966d072dab8e4398c95ce955a9742e04f7
Add all these information into comments, including printing error
message into a3k plat_ea_handler. Also check that abort is really
asynchronous and comes from lower level than EL3.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I46318d221b39773d5e25b3a0221d7738736ffdf1
Put default ea handler implementation into function plat_default_ea_handler()
which just print verbose information and panic, so it can be called also
from overwritten / weak function plat_ea_handler() implementation.
Replace every custom implementation of printing verbose error message of
external aborts in custom plat_ea_handler() functions by a common
implementation from plat_default_ea_handler() function.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I15897f61b62b4c3c29351e693f51d4df381f3b98
* changes:
fix(plat/marvell/a3k): Fix building uart-images.tgz.bin archive
refactor(plat/marvell/a3k): Rename *_CFG and *_SIG variables
refactor(plat/marvell/a3k): Rename DOIMAGETOOL to TBB
refactor(plat/marvell/a3k): Remove useless DOIMAGEPATH variable
fix(plat/marvell/a3k): Fix check for external dependences
fix(plat/marvell/a8k): Add missing build dependency for BLE target
fix(plat/marvell/a8k): Correctly set include directories for individual targets
fix(plat/marvell/a8k): Require that MV_DDR_PATH is correctly set
For UART secure boot it is required also TIMN image, so pack it into
uart-images.tgz.bin archive which is created by mrvl_uart target.
$(TIMN_IMAGE) and $(TIM_IMAGE) variables are used only for UART images
so their content needs to be initialized from $(TIMN_UART_CFG) and
$(TIM_UART_CFG) config files. And not from $(TIMN_CFG) and $(TIM_CFG) as
it is now because they are not generated during mrvl_uart target. Fix it
to allow building mrvl_uart target before mrvl_flash target.
To match usage of these variables, rename them to $(TIMN_UART_IMAGE) and
$(TIM_UART_IMAGE).
To not complicate rule for building uart-images.tgz.bin archive, set
list of image files into a new $(UART_IMAGES) variable.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I83b980abb4047a3afb3ce3026842e1d873c490bf
For TIM config file use TIM name instead of DOIMAGE and use underscores
to make variable names more readable.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I1282ce11f1431c15458a143ae7bfcee85eed2432
Armada 3700 uses external TBB tool for creating images and does not use
internal TF-A doimage tool from tools/marvell/doimage/
Therefore set correct name of variable.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I38a94dca78d483de4c79da597c032e1e5d06d92d
Old Marvell a3700_utils and mv-ddr tarballs do not have to work with
latest TF-A code base. Marvell do not provide these old tarballs on
Extranet anymore. Public version on github repository contains all
patches and is working fine, so for public TF-A builds use only public
external dependencies from git.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: Iee5ac6daa9a1826a5b80a8d54968bdbb8fe72f61
BLE source files depend on external Marvell mv-ddr-marvell tree
(specified in $(MV_DDR_PATH) variable) and its header files. Add
dependency on $(MV_DDR_LIB) target which checks that variable
$(MV_DDR_PATH) is correctly set and ensures that make completes
compilation of mv-ddr-marvell tree.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I73968b24c45d9af1e3500b8db7a24bb4eb2bfa47
Do not set all include directories, including those for external targets
in one PLAT_INCLUDES variable.
Instead split them into variables:
* $(PLAT_INCLUDES) for all TF-A BL images
* BLE target specific $(PLAT_INCLUDES) only for Marvell BLE image
* $(MV_DDR_INCLUDES) for targets in external Marvell mv-ddr-marvell tree
Include directory $(CURDIR)/drivers/marvell is required by TF-A BL
images, so move it from ble.mk to a8k_common.mk.
Include directory $(MV_DDR_PATH) is needed only by Marvell BLE image, so
move it into BLE target specific $(PLAT_INCLUDES) variable.
And remaining include directories specified in ble.mk are needed only
for building external dependences from Marvell mv-ddr tree, so move them
into $(MV_DDR_INCLUDES) variable and correctly use it in $(MV_DDR_LIB)
target.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I331f7de675dca2bc70733d56b768f00d56ae4a67
Target mrvl_flash depends on external mv_ddr source code which is not
part of TF-A project. Do not expect that it is pre-downloaded at some
specific location and require user to specify correct path to mv_ddr
source code via MV_DDR_PATH build option.
TF-A code for Armada 37x0 platform also depends on mv_ddr source code
and already requires passing correct MV_DDR_PATH build option.
So for A8K implement same checks for validity of MV_DDR_PATH option as
are already used by TF-A code for Armada 37x0 platform.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I792f2bfeab0cec89b1b64e88d7b2c456e22de43a
The UART code for the A3K platform assumes that UART parent clock rate
is always 25 MHz. This is incorrect, because the xtal clock can also run
at 40 MHz (this is board specific).
The frequency of the xtal clock is determined by a value on a strapping
pin during SOC reset. The code to determine this frequency is already in
A3K's comphy driver.
Move the get_ref_clk() function from the comphy driver to a separate
file and use it for UART parent clock rate determination.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I8bb18a2d020ef18fe65aa06ffa4ab205c71be92e
Macros PLAT_MARVELL_BOOT_UART* and PLAT_MARVELL_CRASH_UART* are defined
to same values. De-duplicate them into PLAT_MARVELL_UART* macros.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: Iae5daf7cad6a971e6f3dbe561df3d0174106ca7f
Macros PLAT_MARVELL_BL31_RUN_UART* are not used since commit
d7c4420cb8 ("plat/marvell: Migrate to multi-console API").
Remove them.
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I5ec959ef4de87dcfb332c017ad2599bf8af6ffc3
UART parent clock is by default the platform's xtal clock, which is
25 MHz.
The value defined in the driver, though, is 25.8048 MHz. This is a hack
for the suboptimal divisor calculation
Divisor = UART clock / (16 * baudrate)
which does not use rounding division, resulting in a suboptimal value
for divisor if the correct parent clock rate was used.
Change the code for divisor calculation to
Divisor = Round(UART clock / (16 * baudrate))
and change the parent clock rate value to 25 MHz.
The final UART divisor for default baudrate 115200 is not affected by
this change.
(Note that the parent clock rate should not be defined via a macro,
since the xtal clock can also be 40 MHz. This is outside of the scope of
this fix, though.)
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: Iaa401173df87aec94f2dd1b38a90fb6ed0bf0ec6
This new compile option is only for Armada 3720 Development Board. When
it is set to 1 then TF-A will setup PM wake up src configuration.
By default this new option is disabled as it is board specific and no
other A37xx board has PM wake up src configuration.
Currently neither upstream U-Boot nor upstream Linux kernel has wakeup
support for A37xx platforms, so having it disabled does not cause any
issue.
Prior this commit PM wake up src configuration specific for Armada 3720
Development Board was enabled for every A37xx board. After this change it
is enabled only when compiling with build flag A3720_DB_PM_WAKEUP_SRC=1
Signed-off-by: Pali Rohár <pali@kernel.org>
Change-Id: I09fea1172c532df639acb3bb009cfde32d3c5766
Subversion is not reflecting the Marvell sources variant anymore.
This patch removes version.mk from Marvell plafroms.
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Change-Id: I8f3afbe3fab3a38da68876f77455f449f5fe0179
Move efuse definitions to a separate header file for later
usage with other FW modules.
Change-Id: I2e9465f760d0388c8e5863bc64a4cdc57de2417f
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: https://sj1git1.cavium.com/c/IP/SW/boot/atf/+/47313
Tested-by: sa_ip-sw-jenkins <sa_ip-sw-jenkins@marvell.com>
Reviewed-by: Yi Guo <yi.guo@cavium.com>
Use single 64b register for the return value instead of two 32b.
Report an error if caller requested larger than than 64b random
number in a single SMC call.
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Change-Id: Ib8756cd3c0808b78c359f90c6f6913f7d16ac360
Reviewed-on: https://sj1git1.cavium.com/c/IP/SW/boot/atf/+/33280
Tested-by: sa_ip-sw-jenkins <sa_ip-sw-jenkins@marvell.com>
Reviewed-by: Nadav Haklai <nadavh@marvell.com>
This patch forces rx training on 10G ports
as part of comphy_smc call from Linux.
Signed-off-by: Alex Evraev <alexev@marvell.com>
Change-Id: Iebe6ea7c8b21cbdce5c466c8a69b92e9d7c8a8ca
Reviewed-on: https://sj1git1.cavium.com/c/IP/SW/boot/atf/+/30763
Tested-by: sa_ip-sw-jenkins <sa_ip-sw-jenkins@marvell.com>
Reviewed-by: Stefan Chulski <stefanc@marvell.com>
Reviewed-by: Nadav Haklai <nadavh@marvell.com>
