GitLab

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>

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>

px30 is a Quad-core soc and Cortex-a53 inside.
This patch supports the following functions:
1. basic platform setup
2. power up/off cpus
3. suspend/resume cpus
4. suspend/resume system
5. reset system
6. power off system

Change-Id: I73d55aa978096c078242be921abe0ddca9e8f67e
Signed-off-by: XiaoDong Huang <derrick.huang@rock-chips.com>

The new API becomes the default one.

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

In the rockchip bl31 setup the __RO_START__ and __RO_END__ symbols are
currently imported into special BL31_RO_* constants while the general
code also imports them as BL_CODE_BASE and BL_CODE_END.

So we can just use the general symbols and can drop the duplication.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: Ibf1b48ad80bed897247a1690a32711030479262d

All supported Rockchip SoCs (RK3288, RK3328, RK3368 and RK3399)
have non-continuous memory areas in the linker script with a huge
gap between them. This results in extremely padded binary images
with a size of about 4 GiB.

E.g. on the RK3399 we have the following memory areas (and base addresses):
RAM (0x1000), SRAM (0xFF8C0000), and PMUSRAM (0xFF3B0000).

Consumers of the TF-A project (e.g. coreboot or U-Boot) therefore
use the ELF image instead, which has a size of a few hundred kBs.

In order to prevent the generation of a huge and useless file,
this patch disables the binary generation for all affected Rockchip
SoCs.
Signed-off-by: Christoph Müllner <christophm30@gmail.com>
Change-Id: I4ac65bdf1e598c3e1a59507897d183aee9a36916

Currently the compile-time constant PLAT_RK_UART_BASE defines
which UART is used as console device. E.g. on RK3399 it is set
to UART2. That means, that a single bl31 image can not be used
for two boards, which just differ on the UART console.

This patch addresses this limitation by parsing the "stdout-path"
property from the "chosen" node in the DTB. The expected property
string is expected to have the form "serialN:XXX", with
N being either 0, 1, 2, 3 or 4. When the property is found, it will
be used to override PLAT_RK_UART_BASE.

Tested on RK3399-Q7, with a stdout-path of "serial0:115200n8".
Signed-off-by: Christoph Müllner <christophm30@gmail.com>
Change-Id: Iafe1320e77ab006c121f8d52745d54cef68a48c7

params_setup.c provides the function params_early_setup, which
takes care of parsing ATF parameters (bl31_plat_param array,
fdt or coreboot table). As params_early_setup is defined as weak
symbol in bl31_plat_setup.c, providing a platform-specific
bl31_plat_setup implementation is optional.

This patch adds the rockchip-common params_setup.c to the sources
for RK3328. This streamlines the parameter handling for all supported
rockchip SoCs.
Signed-off-by: Christoph Müllner <christophm30@gmail.com>
Change-Id: I071c03106114364ad2fc408e49cc791fe5b35925

In order to set the UART base during bootup in common code of
plat/rockchip, we need to streamline the way the UART base addresses
are defined and add the missing definitions and mappings.

This patch does so by following the pattern UARTn_BASE, which is
already in use on RK3399 and RK3328. The numbering itself is derived
from the upstream Linux DTS files of the individual SoCs.
Signed-off-by: Christoph Müllner <christophm30@gmail.com>
Change-Id: I341a1996f4ceed5f82a2f6687d4dead9d7cc5c1f

While mainline u-boot always expects to submit the devicetree
as platform param, coreboot always uses the existing parameter
structure. As libfdt is somewhat big, it makes sense to limit
its inclusion to where necessary and thus only to non-coreboot
builds.

libfdt itself will get build in all cases, but only the non-
coreboot build will actually reference and thus include it.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: I4c5bc28405a14e6070917e48a526bfe77bab2fb7

The rk3288 is a 4-core Cortex-A12 SoC and shares a lot of features
with later SoCs.

Working features are general non-secure mode (the gic needs special
love for that), psci-based smp bringing cpu cores online and also
taking them offline again, psci-based suspend (the simpler variant
also included in the linux kernel, deeper suspend following later)
and I was also already able to test HYP-mode and was able to boot
a virtual kernel using kvm.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: Ibaaa583b2e78197591a91d254339706fe732476a

There are a number or ARMv7 Rockchip SoCs that are very similar in their
bringup routines to the existing arm64 SoCs, so there is quite a high
commonality possible here.

Things like virtualization also need psci and hyp-mode and instead of
trying to cram this into bootloaders like u-boot, barebox or coreboot
(all used in the field), re-use the existing infrastructure in TF-A
for this (both Rockchip plat support and armv7 support in general).

So add core support for aarch32 Rockchip SoCs, with actual soc support
following in a separate patch.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: I298453985b5d8434934fc0c742fda719e994ba0b

The cpuson_entry_point and cpuson_flags are already declared in
plat_private.h so there is no need to have it again declared in
the local pmu.h, especially as it may cause conflicts when the
other type changes.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: I80ae0e23d22f67109ed96f8ac059973b6de2ce87

Some older socs like the rk3288 do not have the necessary registers
to check the wfi/wfe state of the cpu cores. Allow this case an "just"
do an additional delay similar to how the Linux kernel handles smp
right now.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: I0f67af388b06b8bfb4a9bac411b4900ac266a77a

The current code doing power-management from sram is highly
arm64-specific so should live in a corresponding subdirectory
and not in the common area.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Change-Id: I3b79ac26f70fd189d4d930faa6251439a644c5d9

GCC complains for quite some versions, when compiling the M0 firmware
for Rockchip's rk3399 platform, about an invalid type of function 'main':

  warning: return type of 'main' is not 'int' [-Wmain]

This patch addresses this, by renaming the function to 'm0_main'.
Signed-off-by: Christoph Müllner <christophm30@gmail.com>
Change-Id: I10887f2bda6bdb48c5017044c264139004f7c785

The Rockchip ATF platform can be entered from both Coreboot and U-Boot.
While Coreboot does submit the list of linked parameter structs as
platform param, upstream u-boot actually always provides a pointer
to a devicetree as parameter.
This results in current ATF not running at all when started from U-Boot.

To fix this, add a stub that checks if the parameter is a fdt so we
can at least boot and not get stuck. Later on we can extend this with
actual parsing of information from the devicetree.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Cc: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

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>

After the removal of deprecated interfaces in TF 2.0 the migration to
the new GIC driver interfaces was done incorrectly in rk3328 and rk3368:
2d6f1f01

 ("rockchip: Migrate to new interfaces").

In the GICv2 driver it is mandated that all interrupts are Group 0
interrupts. This patch simply moves all Group 1 interrupts to Group 0.

Change-Id: I224c0135603eb5b81bd512976361500c0d129a91
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>