GitLab

To quote jwerner in T925:
"The __sramdata in the declaration is a mistake, the correct target
section for that global needs to be .pmusram.data. This used to be
in .sram.data once upon a time but then the suspend.c stuff got added
and required it to be moved to PMUSRAM. I guess they forgot to update
that part in the declaration and since the old GCC seemed to silently
prefer the attribute in the definition, nobody noticed."

This fixes building with gcc 11.

fix #T925

Change-Id: I2b91542277c95cf487eaa1344927294d5d1b8f2b
Signed-off-by: Patrick Georgi <pgeorgi@google.com>

Compiling BL31 for the Rockchip platform now produces a message about
the deprecation of gic_common.c.
Follow the advice and use include gicv2.mk instead.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Change-Id: I396b977d57975dba27cfed801ad5264bbbde2b5e

It uses the system timer as "entropy" source in the same
way as QEMU, layerscape and others.

Change-Id: Icda17b78e85255bea96109ca2ee0e091187d62ac
Signed-off-by: Christoph Müllner <christophm30@gmail.com>

When we parse the param from BL2, we try to parse it as a FDT and then,
if that fails, as aux params. However, we don't sufficiently distinguish
between failure modes in the first step: specifically, if we are given
an FDT with good magic that we can't parse for some other reason (e.g.
not enough space in our buffer), we still attempt to parse it as aux
params even though that's guaranteed to fatal. Instead, we should either
fail with a more descriptive message or continue to boot without parsing
the FDT.

This patch takes the latter approach, since all we currently get from
the FDT is non-critical UART params.
Signed-off-by: Thomas Hebb <tommyhebb@gmail.com>
Change-Id: I1e98f1fcda4f78e6b45e86956288bafe58b113e4

The RK3368 has two clusters of 4 cores and it's cluster id starts at
bit 8 of the MPIDR.  To convert from the cluster id (0 or 1) to the
lowest CPU-ID in the respective cluster, we thus need to shift by 6
(i.e. shift by 8 to extract the cluster-id and multiply by 4).

This change is required to ensure the PSCI support can index the
per-cpu entry-address array correctly.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Change-Id: I64a76038f090a85a47067f09f750e96e3946e756

The size of buffer currently used to store the FDT passed from U-Boot as
a platform parameter is not large enough to store some RK3399 device
trees. The largest RK3399 device tree currently in U-Boot (for the
Pinebook Pro) is about 70KB in size when passed to TF-A, so increase the
buffer size to 128K which gives some headroom for possibly larger FDTs
in future.
Signed-off-by: Hugh Cole-Baker <sigmaris@gmail.com>
Change-Id: I414caf20683cd47c02ee470dfa988544f3809919

Current value is 16, count the MAP_REGION calls gets us at least 17,
so increase the max value to 20 to have a bit of a margin.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Change-Id: I93d0324f3d483758366e758f8f663545d365e03f

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>

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>

In extreme cases, the number of secure regions is one more than
non-secure regions. So array "s_base" and "s_top"s size
in struct param_ddr_usage need to be adjust to "DDR_REGION_NR_MAX + 1".
Signed-off-by: XiaoDong Huang <derrick.huang@rock-chips.com>
Change-Id: Ifc09da2c8f8afa1aebcc78f8fbc21ac95abdece2

The Secure Configuration Register is 64-bits in AArch64 and 32-bits in
AArch32. Use u_register_t instead of unsigned int to reflect this.

Change-Id: I51b69467baba36bf0cfaec2595dc8837b1566934
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>

PLATFORM_CORE_COUNT - Unsigned int
PLATFORM_CLUSTER_COUNT - Unsigned int
PLATFORM_MAX_CPUS_PER_CLUSTER - Unsigned int
PLATFORM_CORE_COUNT_PER_CLUSTER - Unsigned int
Signed-off-by: Deepika Bhavnani <deepika.bhavnani@arm.com>
Change-Id: I624c15d569db477506a74964bc828e1a932181d4

In further patches, we wish to enable -wredundant-decls check as
part of warning flags by default.

Change-Id: I43410d6dbf40361a503c16d94ccf0f4cf29615b7
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>

Enable the workaround for Cortex-A53 erratum 855873 for the Rockchip
RK3328, silencing a warning at startup.

Change-Id: I5aa29d674d23c096c599abcb5e7dac970f9607d8
Signed-off-by: Simon South <simon@simonsouth.net>

Transfering the regions of ddr memory to additionally protect is very much
specific to some rockchip internal first stage bootloader and doesn't get
used in either mainline uboot or even Rockchip's published vendor uboot
sources.

This results in a big error
    ERROR:   over or zero region, nr=0, max=10
getting emitted on every boot for most users and such a message coming
from early firmware might actually confuse developers working with the
system.

As this mechanism seems to be only be used by Rockchip's internal miniloader
hide it behind a build conditional, so it doesn't confuse people too much.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Change-Id: I52c02decc60fd431ea78c7486cad5bac82bdbfbe

So far the px30-related ddr security was loading data for regions to secure
from a pre-specified memory location and also setting region0 to secure
the first megabyte of memory in hard-coded setting (top=0, end=0, meaning
1MB).

To make things more explicit and easier to read add a function doing
the settings for specified memory areas, like other socs have and also
add an assert to make sure any descriptor read from memory does not
overlap the TZRAM security in region0 and TEE security in region1.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Change-Id: I78441875112bf66a62fde5f1789f4e52a78ef95f

Similar to others like rk3399 and rk3288 move the secure init to a
separate file to unclutter the soc init a bit.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Change-Id: Iebb38e24f1c7fe5353f139c896fb8ca769bf9691

The calls to secure ddr regions on rk3288 and rk3399 use parameters of
base and size - as it custom for specifying memory regions, but the
functions themself expect start and endpoints of the area.

This only works by chance for the TZRAM, as it starts a 0x0 and therefore
its end location is the same as its size.

To not fall into a trap later on adapt the functions to really take
base+size parameters.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Change-Id: Idb9fab38aa081f3335a4eca971e7b7f6757fbbab

The agreed upon division of early boot locations is 0x40000 for bl31
to leave enough room for u-boot-spl and 0x100000 for bl33 (u-boot).

rk3288 and rk3399 already correctly secure the ddr up to the 1MB boundary
so pull the other platforms along to also give the Rockchip TF-A enough
room to comfortably live in.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Change-Id: Ie9e0c927d3074a418b6fd23b599d2ed7c15c8c6f

Instead of stringizing the paths to binary files, add them as string
defines on the command line (e.g. -DFOO=\"BAR\" instead of -DFOO=BAR).
This prevents macros from being expanded inside the string value itself.
For example, -DFOO=/path/with-linux-in-it would have been expanded to
"/path/with-1-in-it" because `linux=1` is one of the standard GCC
defines.

Change-Id: I7b65df3c9930faed4f1aff75ad726982ae3671e6
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>

HDCP is using a binary driver, add macro PLAT_RK_DP_HDCP to make it as
an option.

Change-Id: I54ef1a3635a28e8ae56654bd1e91dfe011520a7f
Signed-off-by: Ziyuan Xu <xzy.xu@rock-chips.com>
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>

This patch fixes hangs that happen after soft resetting of rk3399.
Signed-off-by: Piotr Szczepanik <piter75@gmail.com>
Change-Id: If41b12ba1dfcb2ba937361b58eafd50bf5c483d4

Add the UART3 base definition for serial output, which is used on some
PX30 SoM boards.
Signed-off-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Change-Id: I8490b15c9f129a33c01cb78bd78675014bc7b015

Variable shadowing is, according to the C standard, permitted and valid
behaviour. However, allowing a local variable to take the same name as a
global one can cause confusion and can make refactoring and bug hunting
more difficult.

This patch moves -Wshadow from WARNING2 into the general warning group
so it is always used. It also fixes all warnings that this introduces
by simply renaming the local variable to a new name

Change-Id: I6b71bdce6580c6e58b5e0b41e4704ab0aa38576e
Signed-off-by: Justin Chadwell <justin.chadwell@arm.com>

And return NULL if we didn't get them in bl aux params otherwise reset and poweroff
will be broken on platforms that do not have reset and poweroff GPIOs.

Fixes: c1185ffd

 ("plat/rockchip: Switch to use new common BL aux parameter library")
Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Change-Id: Ic6cf6383d8f05d745e2c5d5e1b1df38514ea8429

Rockchip platform is using the first 1MB of DRAM as secure ram space,
and there is a vendor loader who loads and runs the BL31/BL32/BL33,
this loader is usually load by SoC BootRom to the start addres of DRAM,
we need to reserve enough space for this loader so that it doesn't need
to do the relocate when loading the BL31. eg.
We use U-Boot SPL to load ATF BL31 and U-Boot proper as BL33, the SPL
TEXT BASE is offset 0 of DRAM which is decide by Bootrom; if we update
the BL31_BASE to offset 0x40000(256KB), then the 0~0x40000 should be
enough for SPL and no need to do the relocate while the space size
0x10000(64KB) may not enough for SPL.
After this update, the BL31 can use the rest 768KB of the first 1MB,
which is also enough, and the loader who is using BL31 elf file can
support this update without any change.

Change-Id: I66dc685594d77f10f9a49c3be015fd6729250ece
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>

The 'txet' should be 'text'.

Change-Id: I2217a1adf50c3b86f3087b83c77d9291b280627c
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>

The rk3399 suspend code saves and restores the debug uart settings, but
right now always does this for the default uart. Right now this works
only by chance for the majority of rk3399 boards, which do not deviate
from that default.

But both Coreboot as well as U-Boot-based platforms can actually use
different uarts for their output, which can be configured from either
devicetree or Coreboot-variables.

To fix this, just use the stored uart-base information instead of the
default constant.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: I1ea059d59a1126f6f8702315df7e620e632b686e

Rockchip platforms can be booted from either u-boot or coreboot.

So far the Coreboot-console was initizalized from a coreboot data struct
in the early_param2 callbacks and dt-based consoles with data from the
rockchip_get_uart_* functions.

But later code may also need this console information for example for
special suspend handling. To make this easy follow a suggestion from
Julius Werner and move the coreboot<->dt distinction into the
rockchip_get_uart_* functions, thus making correct data about the used
uart available to all Rockchip platform code at all times.

This includes a new rockchip_get_uart_clock as well, because while the
dt-platforms right now always just default the rate defined in a constant
Coreboot provides its own field for the clock rate and we don't want to
loose that information for the console init. Similarly the rk_uart_*
variables should move into the non-Coreboot code, to prevent them from
being marked as unused, which also requires the rk_get_uart_* functions
to move below the actual dt-parsing.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: I278d595d2aa6c6864187fc8979a9fbff9814feac

A previous patch already allowed to configure the uart output from the
devicetree, but on Rockchip platforms we also have the issue of different
vendors using different baudrates for their uarts.

For example, rk3399 has a default baudrate of 115200 which is true for
ChromeOS-devices and boards from Theobroma-Systems, while all the boards
using the vendor boot chain actually use a baudrate of 1500000.

Similarly the newly added px30 has a default of said 1500000 but some
boards may want to use the more widely used 115200.

The devicetree stdout-path node already contains the desired baudrate,
so add simple code to parse it from there and override the default,
which stays unchanged.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: I7412139c3df3073a1996eb508ec08642ec6af90d

The px30 mini-evb can use either uart2 (muxed with the sd-card pins) or
uart5 via its pin header for serial output. Uart5 is especially useful
when needing to boot from the sd-card, where uart2 obviously is not
useable.

So add the uart5 constants and it as uart option for the serial-param
handler.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: Ib88df7a55d761ee104d312c9953a13de3beba1c4

NOTE: AARCH32/AARCH64 macros are now deprecated in favor of __aarch64__.

All common C compilers pre-define the same macros to signal which
architecture the code is being compiled for: __arm__ for AArch32 (or
earlier versions) and __aarch64__ for AArch64. There's no need for TF-A
to define its own custom macros for this. In order to unify code with
the export headers (which use __aarch64__ to avoid another dependency),
let's deprecate the AARCH32 and AARCH64 macros and switch the code base
over to the pre-defined standard macro. (Since it is somewhat
unintuitive that __arm__ only means AArch32, let's standardize on only
using __aarch64__.)

Change-Id: Ic77de4b052297d77f38fc95f95f65a8ee70cf200
Signed-off-by: Julius Werner <jwerner@chromium.org>

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>

This change is needed for the platform to compile following the changes
made in commits cbdc72b5 and 3e02c743

.

Change-Id: I3468dd27f3b4f3095fb82f445d51cd8714311eb7
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>

"result of '1 << 31' requires 33 bits to represent, but 'int' only has
32 bits [-Werror=shift-overflow=]"

This is treated as an error since commit 93c690eb

 ("Enable
-Wshift-overflow=2 to check for undefined shift behavior")

Only the actual errors are being tackled by this patch. It is up to the
platform to choose whether there needs to be further modifications to
the code.

Change-Id: I70860ae5f2a34d7c684bd491b76da50aa04f778e
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>

The Rockchip platform is a prime candidate for switching to the new
bl31_params_parse_helper(), so switch it over. This will allow BL2
implementations on this platform to transparently switch over to the
version 2 parameter structure.

Change-Id: I540741d2425c93f66c8697ce749a351eb2b3a7e8
Signed-off-by: Julius Werner <jwerner@chromium.org>

This patch adds a new include/export/ directory meant for inclusion in
third-party code. This is useful for cases where third-party code needs
to interact with TF-A interfaces and data structures (such as a custom
BL2-implementation like coreboot handing off to BL31). Directly
including headers from the TF-A repository avoids having to duplicate
all these definitions (and risk them going stale), but with the current
header structure this is not possible because handoff API definitions
are too deeply intertwined with other TF code/headers and chain-include
other headers that will not be available in the other environment.

The new approach aims to solve this by separating only the parts that
are really needed into these special headers that are self-contained and
will not chain-include other (non-export) headers. TF-A code should
never include them directly but should instead always include the
respective wrapper header, which will include the required prerequisites
(like <stdint.h>) before including the export header. Third-party code
can include the export headers via its own wrappers that make sure the
necessary definitions are available in whatever way that environment can
provide them.

Change-Id: Ifd769320ba51371439a8e5dd5b79c2516c3b43ab
Signed-off-by: Julius Werner <jwerner@chromium.org>

This patch changes all Rockchip platforms to use the new common BL aux
parameter helpers. Since the parameter space is now cleanly split in
generic and vendor-specific parameters and the COREBOOT_TABLE
parameter is now generic, the parameter type number for that parameter
has to change. Since it only affects coreboot which always builds TF as
a submodule and includes its headers directly to get these constants,
this should not cause any issues. In general, after this point, we
should avoid changing already assigned parameter type numbers whenever
possible.

Change-Id: Ic99ddd1e91ff5e5fe212fa30c793a0b8394c9dad
Signed-off-by: Julius Werner <jwerner@chromium.org>

This consists of ensuring that the left operand of each shift is
unsigned when the operation might overflow into the sign bit.

Change-Id: Ib7fc54e4141cc4f1952a18241bc18671b36e2168
Signed-off-by: Justin Chadwell <justin.chadwell@arm.com>