The Server Base System Architecture document (ARM-DEN-0029)
specifies a generic UART device. The programmer's view of this
generic UART is a subset of the ARM PL011 UART. However, the
current PL011 driver in Trusted Firmware uses some features
that are outside the generic UART specification.
This patch modifies the PL011 driver to exclude features outside
the SBSA generic UART specification by setting the boolean build
option 'PL011_GENERIC_UART=1'. Default value is 0 (use full
PL011 features).
User guide updated.
FixesARM-software/tf-issues#216
Change-Id: I6e0eb86f9d69569bc3980fb57e70d6da5d91a737
Firmware Update (FWU) introduces a new set of images called
SCP_BL2U, BL2U and NS_BL2U, which can be packed in a FWU FIP file.
This patch introduces new UUIDs for the Firmware Update images
and extends the 'fip'create' tool so that these new images can be
packed in a FIP file.
Change-Id: I7c60211b4f3cc265411efb131e6d3c624768f522
This patch adds support for Firmware update in BL2U for ARM
platforms such that TZC initialization is performed on all
ARM platforms and (optionally) transfer of SCP_BL2U image on
ARM CSS platforms.
BL2U specific functions are added to handle early_platform and
plat_arch setup. The MMU is configured to map in the BL2U
code/data area and other required memory.
Change-Id: I57863295a608cc06e6cbf078b7ce34cbd9733e4f
The Firmware Update (FWU) feature needs support for an optional
secure world image, BL2U, to allow additional secure world
initialization required by FWU, for example DDR initialization.
This patch adds generic framework support to create BL2U.
NOTE: A platform makefile must supply additional `BL2U_SOURCES`
to build the bl2u target. A subsequent patch adds bl2u
support for ARM platforms.
Change-Id: If2ce036199bb40b39b7f91a9332106bcd4e25413
This patch adds Firmware Update support for ARM platforms.
New files arm_bl1_fwu.c and juno_bl1_setup.c were added to provide
platform specific Firmware update code.
BL1 now includes mmap entry for `ARM_MAP_NS_DRAM1` to map DRAM for
authenticating NS_BL2U image(For both FVP and JUNO platform).
Change-Id: Ie116cd83f5dc00aa53d904c2f1beb23d58926555
Firmware update(a.k.a FWU) feature is part of the TBB architecture.
BL1 is responsible for carrying out the FWU process if platform
specific code detects that it is needed.
This patch adds support for FWU feature support in BL1 which is
included by enabling `TRUSTED_BOARD_BOOT` compile time flag.
This patch adds bl1_fwu.c which contains all the core operations
of FWU, which are; SMC handler, image copy, authentication, execution
and resumption. It also adds bl1.h introducing #defines for all
BL1 SMCs.
Following platform porting functions are introduced:
int bl1_plat_mem_check(uintptr_t mem_base, unsigned int mem_size,
unsigned int flags);
This function can be used to add platform specific memory checks
for the provided base/size for the given security state.
The weak definition will invoke `assert()` and return -ENOMEM.
__dead2 void bl1_plat_fwu_done(void *cookie, void *reserved);
This function can be used to initiate platform specific procedure
to mark completion of the FWU process.
The weak definition waits forever calling `wfi()`.
plat_bl1_common.c contains weak definitions for above functions.
FWU process starts when platform detects it and return the image_id
other than BL2_IMAGE_ID by using `bl1_plat_get_next_image_id()` in
`bl1_main()`.
NOTE: User MUST provide platform specific real definition for
bl1_plat_mem_check() in order to use it for Firmware update.
Change-Id: Ice189a0885d9722d9e1dd03f76cac1aceb0e25ed
As of now BL1 loads and execute BL2 based on hard coded information
provided in BL1. But due to addition of support for upcoming Firmware
Update feature, BL1 now require more flexible approach to load and
run different images using information provided by the platform.
This patch adds new mechanism to load and execute images based on
platform provided image id's. BL1 now queries the platform to fetch
the image id of the next image to be loaded and executed. In order
to achieve this, a new struct image_desc_t was added which holds the
information about images, such as: ep_info and image_info.
This patch introduces following platform porting functions:
unsigned int bl1_plat_get_next_image_id(void);
This is used to identify the next image to be loaded
and executed by BL1.
struct image_desc *bl1_plat_get_image_desc(unsigned int image_id);
This is used to retrieve the image_desc for given image_id.
void bl1_plat_set_ep_info(unsigned int image_id,
struct entry_point_info *ep_info);
This function allows platforms to update ep_info for given
image_id.
The plat_bl1_common.c file provides default weak implementations of
all above functions, the `bl1_plat_get_image_desc()` always return
BL2 image descriptor, the `bl1_plat_get_next_image_id()` always return
BL2 image ID and `bl1_plat_set_ep_info()` is empty and just returns.
These functions gets compiled into all BL1 platforms by default.
Platform setup in BL1, using `bl1_platform_setup()`, is now done
_after_ the initialization of authentication module. This change
provides the opportunity to use authentication while doing the
platform setup in BL1.
In order to store secure/non-secure context, BL31 uses percpu_data[]
to store context pointer for each core. In case of BL1 only the
primary CPU will be active hence percpu_data[] is not required to
store the context pointer.
This patch introduce bl1_cpu_context[] and bl1_cpu_context_ptr[] to
store the context and context pointers respectively. It also also
re-defines cm_get_context() and cm_set_context() for BL1 in
bl1/bl1_context_mgmt.c.
BL1 now follows the BL31 pattern of using SP_EL0 for the C runtime
environment, to support resuming execution from a previously saved
context.
NOTE: THE `bl1_plat_set_bl2_ep_info()` PLATFORM PORTING FUNCTION IS
NO LONGER CALLED BY BL1 COMMON CODE. PLATFORMS THAT OVERRIDE
THIS FUNCTION MAY NEED TO IMPLEMENT `bl1_plat_set_ep_info()`
INSTEAD TO MAINTAIN EXISTING BEHAVIOUR.
Change-Id: Ieee4c124b951c2e9bc1c1013fa2073221195d881
The upcoming Firmware Update feature needs transitioning across
Secure/Normal worlds to complete the FWU process and hence requires
context management code to perform this task.
Currently context management code is part of BL31 stage only.
This patch moves the code from (include)/bl31 to (include)/common.
Some function declarations/definitions and macros have also moved
to different files to help code sharing.
Change-Id: I3858b08aecdb76d390765ab2b099f457873f7b0c
The primary usage of `RUN_IMAGE` SMC function id, used by BL2 is to
make a request to BL1 to execute BL31. But BL2 also uses it as
opcode to check if it is allowed to execute which is not the
intended usage of `RUN_IMAGE` SMC.
This patch removes the usage of `RUN_IMAGE` as opcode passed to
next EL to check if it is allowed to execute.
Change-Id: I6aebe0415ade3f43401a4c8a323457f032673657
This patch overrides the default weak definition of
`bl31_plat_runtime_setup()` for ARM Standard platforms to
specify a BL31 runtime console. ARM Standard platforms are
now expected to define `PLAT_ARM_BL31_RUN_UART_BASE` and
`PLAT_ARM_BL31_RUN_UART_CLK_IN_HZ` macros which is required
by `arm_bl31_plat_runtime_setup()` to initialize the runtime
console.
The system suspend resume helper `arm_system_pwr_domain_resume()`
is fixed to initialize the runtime console rather than the boot
console on resumption from system suspend.
FixesARM-software/tf-issues#220
Change-Id: I80eafe5b6adcfc7f1fdf8b99659aca1c64d96975
It is not ideal for BL31 to continue to use boot console at
runtime which could be potentially uninitialized. This patch
introduces a new optional platform porting API
`bl31_plat_runtime_setup()` which allows the platform to perform
any BL31 runtime setup just prior to BL31 exit during cold boot.
The default weak implementation of this function will invoke
`console_uninit()` which will suppress any BL31 runtime logs.
On the ARM Standard platforms, there is an anomaly that
the boot console will be reinitialized on resumption from
system suspend in `arm_system_pwr_domain_resume()`. This
will be resolved in the following patch.
NOTE: The default weak definition of `bl31_plat_runtime_setup()`
disables the BL31 console. To print the BL31 runtime
messages, platforms must override this API and initialize a
runtime console.
FixesARM-software/tf-issues#328
Change-Id: Ibaf8346fcceb447fe1a5674094c9f8eb4c09ac4a
Allowing console base address to be set to NULL conveniently
allows console driver to ignore further invocations to console_putc()
and console_getc(). This patch adds `console_uninit()` API to the
console driver which sets console base address as NULL. The BL images can
invoke this API to finish the use of console and ignore any further
invocations to print to the console.
Change-Id: I00a1762b3e0b7c55f2be2f9c4c9bee3967189dde
In the situation that EL1 is selected as the exception level for the
next image upon BL31 exit for a processor that supports EL2, the
context management code must configure all essential EL2 register
state to ensure correct execution of EL1.
VTTBR_EL2 should be part of this set of EL2 registers because:
- The ARMv8-A architecture does not define a reset value for this
register.
- Cache maintenance operations depend on VTTBR_EL2.VMID even when
non-secure EL1&0 stage 2 address translation are disabled.
This patch initializes the VTTBR_EL2 register to 0 when bypassing EL2
to address this issue. Note that this bug has not yet manifested
itself on FVP or Juno because VTTBR_EL2.VMID resets to 0 on the
Cortex-A53 and Cortex-A57.
Change-Id: I58ce2d16a71687126f437577a506d93cb5eecf33
This patch enables support for EL3 interrupts in the Interrupt Management
Framework (IMF) of ARM Trusted Firmware. Please note that although the
registration of the EL3 interrupt type is now supported, it has not been
tested on any of the ARM Standard platforms.
Change-Id: If4dcdc7584621522a2f3ea13ea9b1ad0a76bb8a1
Suport for ARM GIC v2.0 and v3.0 drivers has been reworked to create three
separate drivers instead of providing a single driver that can work on both
versions of the GIC architecture. These drivers correspond to the following
software use cases:
1. A GICv2 only driver that can run only on ARM GIC v2.0 implementations
e.g. GIC-400
2. A GICv3 only driver that can run only on ARM GIC v3.0 implementations
e.g. GIC-500 in a mode where all interrupt regimes use GICv3 features
3. A deprecated GICv3 driver that operates in legacy mode. This driver can
operate only in the GICv2 mode in the secure world. On a GICv3 system, this
driver allows normal world to run in either GICv3 mode (asymmetric mode)
or in the GICv2 mode. Both modes of operation are deprecated on GICv3
systems.
ARM platforms implement both versions of the GIC architecture. This patch adds a
layer of abstraction to help ARM platform ports chose the right GIC driver and
corresponding platform support. This is as described below:
1. A set of ARM common functions have been introduced to initialise the GIC and
the driver during cold and warm boot. These functions are prefixed as
"plat_arm_gic_". Weak definitions of these functions have been provided for
each type of driver.
2. Each platform includes the sources that implement the right functions
directly into the its makefile. The FVP can be instantiated with different
versions of the GIC architecture. It uses the FVP_USE_GIC_DRIVER build option
to specify which of the three drivers should be included in the build.
3. A list of secure interrupts has to be provided to initialise each of the
three GIC drivers. For GIC v3.0 the interrupt ids have to be further
categorised as Group 0 and Group 1 Secure interrupts. For GIC v2.0, the two
types are merged and treated as Group 0 interrupts.
The two lists of interrupts are exported from the platform_def.h. The lists
are constructed by adding a list of board specific interrupt ids to a list of
ids common to all ARM platforms and Compute sub-systems.
This patch also makes some fields of `arm_config` data structure in FVP redundant
and these unused fields are removed.
Change-Id: Ibc8c087be7a8a6b041b78c2c3bd0c648cd2035d8
This patch adds platform helpers for the new GICv2 and GICv3 drivers in
plat_gicv2.c and plat_gicv3.c. The platforms can include the appropriate
file in their build according to the GIC driver to be used. The existing
plat_gic.c is only meant for the legacy GIC driver.
In the case of ARM platforms, the major changes are as follows:
1. The crash reporting helper macro `arm_print_gic_regs` that prints the GIC CPU
interface register values has been modified to detect the type of CPU
interface being used (System register or memory mappped interface) before
using the right interface to print the registers.
2. The power management helper function that is called after a core is powered
up has been further refactored. This is to highlight that the per-cpu
distributor interface should be initialised only when the core was originally
powered down using the CPU_OFF PSCI API and not when the CPU_SUSPEND PSCI API
was used.
3. In the case of CSS platforms, the system power domain restore helper
`arm_system_pwr_domain_resume()` is now only invoked in the `suspend_finish`
handler as the system power domain is always expected to be initialized when
the `on_finish` handler is invoked.
Change-Id: I7fc27d61fc6c2a60cea2436b676c5737d0257df6
Add compile time `__warn_deprecated` flag to public api's in CCI-400
specific driver so that user is aware of the driver being deprecated.
Similarly, it also adds an error message when `ERROR_DEPRECATED` is set
to prevent succesful compilation if CCI-400 specific driver is used.
Change-Id: Id7e61a560262abc01cbbd432ca85b9bf448a194d
On a GICv2 system, interrupts that should be handled in the secure world are
typically signalled as FIQs. On a GICv3 system, these interrupts are signalled
as IRQs instead. The mechanism for handling both types of interrupts is the same
in both cases. This patch enables the TSP to run on a GICv3 system by:
1. adding support for handling IRQs in the exception handling code.
2. removing use of "fiq" in the names of data structures, macros and functions.
The build option TSPD_ROUTE_IRQ_TO_EL3 is deprecated and is replaced with a
new build flag TSP_NS_INTR_ASYNC_PREEMPT. For compatibility reasons, if the
former build flag is defined, it will be used to define the value for the
new build flag. The documentation is also updated accordingly.
Change-Id: I1807d371f41c3656322dd259340a57649833065e
The TSP is expected to pass control back to EL3 if it gets preempted due to
an interrupt while handling a Standard SMC in the following scenarios:
1. An FIQ preempts Standard SMC execution and that FIQ is not a TSP Secure
timer interrupt or is preempted by a higher priority interrupt by the time
the TSP acknowledges it. In this case, the TSP issues an SMC with the ID
as `TSP_EL3_FIQ`. Currently this case is never expected to happen as only
the TSP Secure Timer is expected to generate FIQ.
2. An IRQ preempts Standard SMC execution and in this case the TSP issues
an SMC with the ID as `TSP_PREEMPTED`.
In both the cases, the TSPD hands control back to the normal world and returns
returns an error code to the normal world to indicate that the standard SMC it
had issued has been preempted but not completed.
This patch unifies the handling of these two cases in the TSPD and ensures that
the TSP only uses TSP_PREEMPTED instead of separate SMC IDs. Also instead of 2
separate error codes, SMC_PREEMPTED and TSP_EL3_FIQ, only SMC_PREEMPTED is
returned as error code back to the normal world.
Background information: On a GICv3 system, when the secure world has affinity
routing enabled, in 2. an FIQ will preempt TSP execution instead of an IRQ. The
FIQ could be a result of a Group 0 or a Group 1 NS interrupt. In both case, the
TSPD passes control back to the normal world upon receipt of the TSP_PREEMPTED
SMC. A Group 0 interrupt will immediately preempt execution to EL3 where it
will be handled. This allows for unified interrupt handling in TSP for both
GICv3 and GICv2 systems.
Change-Id: I9895344db74b188021e3f6a694701ad272fb40d4
This patch renames the GICv3 interrupt group macros from
INT_TYPE_G0, INT_TYPE_G1S and INT_TYPE_G1NS to INTR_GROUP0,
INTR_GROUP1S and INTR_GROUP1NS respectively.
Change-Id: I40c66f589ce6234fa42205adcd91f7d6ad8f33d4
This patch fixes several issues with the SP804 delay timer on FVP:
* By default, the SP804 dual timer on FVP runs at 32 KHz. In order
to run the timer at 35 MHz (as specified in the FVP user manual)
the Overwrite bit in the SP810 control register must be set.
* The CLKMULT and CLKDIV definitions are mixed up:
delta(us) = delta(ticks) * T(us) = delta(ticks) / f(MHz)
From the delay function:
delta_us = (delta * ops->clk_mult) / ops->clk_div;
Matching both expressions:
1 / f(MHz) = ops->clk_mult / ops->clk_div
And consequently:
f(MHz) = ops->clk_div / ops->clk_mult
Which, for a 35 MHz timer, translates to:
ops->clk_div = 35
ops->clk_mult = 1
* The comment in the delay timer header file has been corrected:
The ratio of the multiplier and the divider is the clock period
in microseconds, not the frequency.
Change-Id: Iffd5ce0a5a28fa47c0720c0336d81b678ff8fdf1
This patch adds watchdog support on ARM platforms (FVP and Juno).
A secure instance of SP805 is used as Trusted Watchdog. It is
entirely managed in BL1, being enabled in the early platform setup
hook and disabled in the exit hook. By default, the watchdog is
enabled in every build (even when TBB is disabled).
A new ARM platform specific build option `ARM_DISABLE_TRUSTED_WDOG`
has been introduced to allow the user to disable the watchdog at
build time. This feature may be used for testing or debugging
purposes.
Specific error handlers for Juno and FVP are also provided in this
patch. These handlers will be called after an image load or
authentication error. On FVP, the Table of Contents (ToC) in the FIP
is erased. On Juno, the corresponding error code is stored in the
V2M Non-Volatile flags register. In both cases, the CPU spins until
a watchdog reset is generated after 256 seconds (as specified in
the TBBR document).
Change-Id: I9ca11dcb0fe15af5dbc5407ab3cf05add962f4b4
This patch adds ARM specific OIDs which will be used to extract
the extension data from the certificates. These OIDs are arranged
as a subtree whose root node has been specifically allocated for
ARM Ltd.
{ iso(1) identified-organization(3) dod(6) internet(1)
private(4) enterprise(1) 4128 }
Change-Id: Ice20b3c8a31ddefe9102f3bd42f7429986f3ac34
The TZC-400 driver implementation incorrectly uses the component
ID registers to detect the TZC-400 peripheral. As all ARM
peripherals share the same component ID, it doesn't allow to
uniquely identify the TZC-400 peripheral. This patch fixes the
TZC-400 driver by relying on the `part_number_0` and
`part_number_1` fields in the `PID` registers instead.
The `tzc_read_component_id` function has been replaced by
`tzc_read_peripheral_id`, which reads the 'part_number' values
and compares them with the TZC-400 peripheral ID.
Also, it adds a debug assertion to detect when the TZC driver
initialisation function is called multiple times.
Change-Id: I35949f6501a51c0a794144cd1c3a6db62440dce6
Based on SP805 Programmer's model (ARM DDI 0270B). This driver
provides three public APIs:
void sp805_start(uintptr_t base, unsigned long ticks);
void sp805_stop(uintptr_t base);
void sp805_refresh(uintptr_t base, unsigned long ticks);
Upon start, the watchdog starts counting down from the number of
ticks specified. When the count reaches 0 an interrupt is triggered.
The watchdog restarts counting down from the number of ticks
specified. If the count reaches 0 again, the system is reset. A
mechanism to handle the interrupt has not been implemented. Instead,
the API to refresh the watchdog should be used instead to prevent a
system reset.
Change-Id: I799d53f8d1213b10b341a4a67fde6486e89a3dab
FVP and Juno platforms include a NOR flash memory to store and
load the FIP, the kernel or a ramdisk. This NOR flash is arranged
as 2 x 16 bit flash devices and can be programmed using CFI
standard commands.
This patch provides a basic API to write single 32 bit words of
data into the NOR flash. Functions to lock/unlock blocks against
erase or write operations are also provided.
Change-Id: I1da7ad3105b1ea409c976adc863954787cbd90d2
This patch adds support for booting EL3 payloads on CSS platforms,
for example Juno. In this scenario, the Trusted Firmware follows
its normal boot flow up to the point where it would normally pass
control to the BL31 image. At this point, it jumps to the EL3
payload entry point address instead.
Before handing over to the EL3 payload, the data SCP writes for AP
at the beginning of the Trusted SRAM is restored, i.e. we zero the
first 128 bytes and restore the SCP Boot configuration. The latter
is saved before transferring the BL30 image to SCP and is restored
just after the transfer (in BL2). The goal is to make it appear that
the EL3 payload is the first piece of software to run on the target.
The BL31 entrypoint info structure is updated to make the primary
CPU jump to the EL3 payload instead of the BL31 image.
The mailbox is populated with the EL3 payload entrypoint address,
which releases the secondary CPUs out of their holding pen (if the
SCP has powered them on). The arm_program_trusted_mailbox() function
has been exported for this purpose.
The TZC-400 configuration in BL2 is simplified: it grants secure
access only to the whole DRAM. Other security initialization is
unchanged.
This alternative boot flow is disabled by default. A new build option
EL3_PAYLOAD_BASE has been introduced to enable it and provide the EL3
payload's entry point address. The build system has been modified
such that BL31 and BL33 are not compiled and/or not put in the FIP in
this case, as those images are not used in this boot flow.
Change-Id: Id2e26fa57988bbc32323a0effd022ab42f5b5077
This patch deprecates the legacy ARM GIC driver and related header files
(arm_gic.h, gic_v2.h, gic_v3.h). For GICv2 systems, platform ports should
use the GICv2 driver in include/drivers/arm/gicv2.h and for GICv3 systems,
platform ports should use the GICv3 driver in include/drivers/arm/gicv3.h
NOTE: The ARM Legacy GIC drivers have been deprecated with this patch.
Platform ports are encouraged to migrate to the new GIC drivers.
Change-Id: Ic0460ef0427b54a6aac476279a7f29b81943e942
This patch adds a driver for ARM GICv2 systems, example GIC-400. Unlike
the existing GIC driver in `include/drivers/arm/arm_gic.h`, this driver
is optimised for GICv2 and does not support GICv3 systems in GICv2
compatibility mode. The driver interface has been implemented in
`drivers/arm/gic/v2/gicv2_main.c`. The corresponding header is in
`include/drivers/arm/gicv2.h`. Helper functions are implemented in
`drivers/arm/gic/v2/gicv2_helpers.c` and are accessible through the
`drivers/arm/gic/v2/gicv2_private.h` header.
Change-Id: I09fffa4e621fb99ba3c01204839894816cd89a2a
This patch adds a driver for ARM GICv3 systems that need to run software
stacks where affinity routing is enabled across all privileged exception
levels for both security states. This driver is a partial implementation
of the ARM Generic Interrupt Controller Architecture Specification, GIC
architecture version 3.0 and version 4.0 (ARM IHI 0069A). The driver does
not cater for legacy support of interrupts and asymmetric configurations.
The existing GIC driver has been preserved unchanged. The common code for
GICv2 and GICv3 systems has been refactored into a new file,
`drivers/arm/gic/common/gic_common.c`. The corresponding header is in
`include/drivers/arm/gic_common.h`.
The driver interface is implemented in `drivers/arm/gic/v3/gicv3_main.c`.
The corresponding header is in `include/drivers/arm/gicv3.h`. Helper
functions are implemented in `drivers/arm/gic/v3/arm_gicv3_helpers.c`
and are accessible through the `drivers/arm/gic/v3/gicv3_private.h`
header.
Change-Id: I8c3c834a1d049d05b776b4dcb76b18ccb927444a
This patch changes the build time behaviour when using deprecated API within
Trusted Firmware. Previously the use of deprecated APIs would only trigger a
build warning (which was always treated as a build error), when
WARN_DEPRECATED = 1. Now, the use of deprecated C declarations will always
trigger a build time warning. Whether this warning is treated as error or not
is determined by the build flag ERROR_DEPRECATED which is disabled by default.
When the build flag ERROR_DEPRECATED=1, the invocation of deprecated API or
inclusion of deprecated headers will result in a build error.
Also the deprecated context management helpers in context_mgmt.c are now
conditionally compiled depending on the value of ERROR_DEPRECATED flag
so that the APIs themselves do not result in a build error when the
ERROR_DEPRECATED flag is set.
NOTE: Build systems that use the macro WARN_DEPRECATED must migrate to
using ERROR_DEPRECATED, otherwise deprecated API usage will no longer
trigger a build error.
Change-Id: I843bceef6bde979af7e9b51dddf861035ec7965a
The default reset values for the L2 Data & Tag RAM latencies on the
Cortex-A72 on Juno R2 are not suitable. This patch modifies
the Juno platform reset handler to configure the right settings
on Juno R2.
Change-Id: I20953de7ba0619324a389e0b7bbf951b64057db8
As per Section D7.2.81 in the ARMv8-A Reference Manual (DDI0487A Issue A.h),
bits[29:28], bits[23:22], bit[20] and bit[11] in the SCTLR_EL1 are RES1. This
patch adds the missing bit[20] to the SCTLR_EL1_RES1 macro.
Change-Id: I827982fa2856d04def6b22d8200a79fe6922a28e
Patch 7e26fe1f deprecates IO specific return definitions in favour
of standard errno codes. This patch removes those definitions
and its usage from the IO framework, IO drivers and IO platform
layer. Following this patch, standard errno codes must be used
when checking the return value of an IO function.
Change-Id: Id6e0e9d0a7daf15a81ec598cf74de83d5768650f
This patch introduces a new function called 'print_entry_point_info'
that prints an entry_point_t structure for debugging purposes.
As such, it can be used to display the entry point address, SPSR and
arguments passed from a firmware image to the next one.
This function is now called in the following images transitions:
- BL1 to BL2
- BL1 to BL31
- BL31 to the next image (typically BL32 or BL33)
The following changes have been introduced:
- Fix the output format of the SPSR value : SPSR is a 32-bit value,
not a 64-bit one.
- Print all arguments values.
The entry_point_info_t structure allows to pass up to 8 arguments.
In most cases, only the first 2 arguments were printed.
print_entry_point_info() now prints all of them as 'VERBOSE'
traces.
Change-Id: Ieb384bffaa7849e6cb95a01a47c0b7fc2308653a
This patch fixes a compilation issue for platforms that are aligned to ARM
Standard platforms and include the `plat_arm.h` header in their platform port.
The compilation would fail for such a platform because `xlat_tables.h` which
has the definition for `mmap_region_t` is not included in `plat_arm.h`. This
patch fixes this by including `xlat_tables.h` in `plat_arm.h` header.
FixesARM-Software/tf-issues#318
Change-Id: I75f990cfb4078b3996fc353c8cd37c9de61d555e
This patch adds the capability to power down at system power domain level
on Juno via the PSCI SYSTEM SUSPEND API. The CSS power management helpers
are modified to add support for power management operations at system
power domain level. A new helper for populating `get_sys_suspend_power_state`
handler in plat_psci_ops is defined. On entering the system suspend state,
the SCP powers down the SYSTOP power domain on the SoC and puts the memory
into retention mode. On wakeup from the power down, the system components
on the CSS will be reinitialized by the platform layer and the PSCI client
is responsible for restoring the context of these system components.
According to PSCI Specification, interrupts targeted to cores in PSCI CPU
SUSPEND should be able to resume it. On Juno, when the system power domain
is suspended, the GIC is also powered down. The SCP resumes the final core
to be suspend when an external wake-up event is received. But the other
cores cannot be woken up by a targeted interrupt, because GIC doesn't
forward these interrupts to the SCP. Due to this hardware limitation,
we down-grade PSCI CPU SUSPEND requests targeted to the system power domain
level to cluster power domain level in `juno_validate_power_state()`
and the CSS default `plat_arm_psci_ops` is overridden in juno_pm.c.
A system power domain resume helper `arm_system_pwr_domain_resume()` is
defined for ARM standard platforms which resumes/re-initializes the
system components on wakeup from system suspend. The security setup also
needs to be done on resume from system suspend, which means
`plat_arm_security_setup()` must now be included in the BL3-1 image in
addition to previous BL images if system suspend need to be supported.
Change-Id: Ie293f75f09bad24223af47ab6c6e1268f77bcc47
This patch implements the necessary topology changes for supporting
system power domain on CSS platforms. The definition of PLAT_MAX_PWR_LVL and
PLAT_NUM_PWR_DOMAINS macros are removed from arm_def.h and are made platform
specific. In addition, the `arm_power_domain_tree_desc[]` and
`arm_pm_idle_states[]` are modified to support the system power domain
at level 2. With this patch, even though the power management operations
involving the system power domain will not return any error, the platform
layer will silently ignore any operations to the power domain. The actual
power management support for the system power domain will be added later.
Change-Id: I791867eded5156754fe898f9cdc6bba361e5a379
This patch adds an optional API to the platform port:
void plat_error_handler(int err) __dead2;
The platform error handler is called when there is a specific error
condition after which Trusted Firmware cannot continue. While panic()
simply prints the crash report (if enabled) and spins, the platform
error handler can be used to hand control over to the platform port
so it can perform specific bookeeping or post-error actions (for
example, reset the system). This function must not return.
The parameter indicates the type of error using standard codes from
errno.h. Possible errors reported by the generic code are:
-EAUTH : a certificate or image could not be authenticated
(when Trusted Board Boot is enabled)
-ENOENT : the requested image or certificate could not be found
or an IO error was detected
-ENOMEM : resources exhausted. Trusted Firmware does not use
dynamic memory, so this error is usually an indication
of an incorrect array size
A default weak implementation of this function has been provided.
It simply implements an infinite loop.
Change-Id: Iffaf9eee82d037da6caa43b3aed51df555e597a3
This patch replaces custom definitions used as return values for
the load_auth_image() function with standard error codes defined
in errno.h. The custom definitions have been removed.
It also replaces the usage of IO framework error custom definitions,
which have been deprecated. Standard errno definitions are used
instead.
Change-Id: I1228477346d3876151c05b470d9669c37fd231be
This patch redefines the values of IO_FAIL, IO_NOT_SUPPORTED and
IO_RESOURCES_EXHAUSTED to match the corresponding definitions in
errno.h:
#define IO_FAIL (-ENOENT)
#define IO_NOT_SUPPORTED (-ENODEV)
#define IO_RESOURCES_EXHAUSTED (-ENOMEM)
NOTE: please note that the IO_FAIL, IO_NOT_SUPPORTED and
IO_RESOURCES_EXHAUSTED definitions are considered deprecated
and their usage should be avoided. Callers should rely on errno.h
definitions when checking the return values of IO functions.
Change-Id: Ic8491aa43384b6ee44951ebfc053a3ded16a80be
This patch does the following reorganization to psci power management (PM)
handler setup for ARM standard platform ports :
1. The mailbox programming required during `plat_setup_psci_ops()` is identical
for all ARM platforms. Hence the implementation of this API is now moved
to the common `arm_pm.c` file. Each ARM platform now must define the
PLAT_ARM_TRUSTED_MAILBOX_BASE macro, which in current platforms is the same
as ARM_SHARED_RAM_BASE.
2. The PSCI PM handler callback structure, `plat_psci_ops`, must now be
exported via `plat_arm_psci_pm_ops`. This allows the common implementation
of `plat_setup_psci_ops()` to return a platform specific `plat_psci_ops`.
In the case of CSS platforms, a default weak implementation of the same is
provided in `css_pm.c` which can be overridden by each CSS platform.
3. For CSS platforms, the PSCI PM handlers defined in `css_pm.c` are now
made library functions and a new header file `css_pm.h` is added to export
these generic PM handlers. This allows the platform to reuse the
adequate CSS PM handlers and redefine others which need to be customized
when overriding the default `plat_arm_psci_pm_ops` in `css_pm.c`.
Change-Id: I277910f609e023ee5d5ff0129a80ecfce4356ede
The CASSERT() macro introduces a typedef for the sole purpose of
triggering a compilation error if the condition to check is false.
This typedef is not used afterwards. As a consequence, when the
CASSERT() macro is called from withing a function block, the compiler
complains and outputs the following error message:
error: typedef 'msg' locally defined but not used [-Werror=unused-local-typedefs]
This patch adds the "unused" attribute for the aforementioned
typedef. This silences the compiler warning and thus makes the
CASSERT() macro callable from within function blocks as well.
Change-Id: Ie36b58fcddae01a21584c48bb6ef43ec85590479
This patch adds PM handlers to TLKD for the system suspend/resume and
system poweroff/reset cases. TLK expects all SMCs through a single
handler, which then fork out into multiple handlers depending on the
SMC. We tap into the same single entrypoint by restoring the S-EL1
context before passing the PM event via register 'x0'. On completion
of the PM event, TLK sends a completion SMC and TLKD then moves on
with the PM process.
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
The generic delay timer driver expects a pointer to a timer_ops_t
structure containing the specific timer driver information. It
doesn't make a copy of the structure, instead it just keeps the
pointer. Therefore, this pointer must remain valid over time.
The SP804 driver doesn't satisfy this requirement. The
sp804_timer_init() macro creates a temporary instanciation of the
timer_ops_t structure on the fly and passes it to the generic
delay timer. When this temporary instanciation gets deallocated,
the generic delay timer is left with a pointer to invalid data.
This patch fixes this bug by statically allocating the SP804
timer_ops_t structure.
Change-Id: I8fbf75907583aef06701e3fd9fabe0b2c9bc95bf
This patch adds a device driver which can be used to program the following
aspects of ARM CCN IP:
1. Specify the mapping between ACE/ACELite/ACELite+DVM/CHI master interfaces and
Request nodes.
2. Add and remove master interfaces from the snoop and dvm
domains.
3. Place the L3 cache in a given power state.
4. Configuring system adress map and enabling 3 SN striping mode of memory
controller operation.
Change-Id: I0f665c6a306938e5b66f6a92f8549b529aa8f325
On the ARMv8 architecture, cache maintenance operations by set/way on the last
level of integrated cache do not affect the system cache. This means that such a
flush or clean operation could result in the data being pushed out to the system
cache rather than main memory. Another CPU could access this data before it
enables its data cache or MMU. Such accesses could be serviced from the main
memory instead of the system cache. If the data in the sysem cache has not yet
been flushed or evicted to main memory then there could be a loss of
coherency. The only mechanism to guarantee that the main memory will be updated
is to use cache maintenance operations to the PoC by MVA(See section D3.4.11
(System level caches) of ARMv8-A Reference Manual (Issue A.g/ARM DDI0487A.G).
This patch removes the reliance of Trusted Firmware on the flush by set/way
operation to ensure visibility of data in the main memory. Cache maintenance
operations by MVA are now used instead. The following are the broad category of
changes:
1. The RW areas of BL2/BL31/BL32 are invalidated by MVA before the C runtime is
initialised. This ensures that any stale cache lines at any level of cache
are removed.
2. Updates to global data in runtime firmware (BL31) by the primary CPU are made
visible to secondary CPUs using a cache clean operation by MVA.
3. Cache maintenance by set/way operations are only used prior to power down.
NOTE: NON-UPSTREAM TRUSTED FIRMWARE CODE SHOULD MAKE EQUIVALENT CHANGES IN
ORDER TO FUNCTION CORRECTLY ON PLATFORMS WITH SUPPORT FOR SYSTEM CACHES.
FixesARM-software/tf-issues#205
Change-Id: I64f1b398de0432813a0e0881d70f8337681f6e9a
This patch updates ARM platform ports to use the new unified bakery locks
API. The caller does not have to use a different bakery lock API depending upon
the value of the USE_COHERENT_MEM build option.
NOTE: THIS PATCH CAN BE USED AS A REFERENCE TO UPDATE OTHER PLATFORM PORTS.
Change-Id: I1b26afc7c9a9808a6040eb22f603d30192251da7
This patch unifies the bakery lock api's across coherent and normal
memory implementation of locks by using same data type `bakery_lock_t`
and similar arguments to functions.
A separate section `bakery_lock` has been created and used to allocate
memory for bakery locks using `DEFINE_BAKERY_LOCK`. When locks are
allocated in normal memory, each lock for a core has to spread
across multiple cache lines. By using the total size allocated in a
separate cache line for a single core at compile time, the memory for
other core locks is allocated at link time by multiplying the single
core locks size with (PLATFORM_CORE_COUNT - 1). The normal memory lock
algorithm now uses lock address instead of the `id` in the per_cpu_data.
For locks allocated in coherent memory, it moves locks from
tzfw_coherent_memory to bakery_lock section.
The bakery locks are allocated as part of bss or in coherent memory
depending on usage of coherent memory. Both these regions are
initialised to zero as part of run_time_init before locks are used.
Hence, bakery_lock_init() is made an empty function as the lock memory
is already initialised to zero.
The above design lead to the removal of psci bakery locks from
non_cpu_power_pd_node to psci_locks.
NOTE: THE BAKERY LOCK API WHEN USE_COHERENT_MEM IS NOT SET HAS CHANGED.
THIS IS A BREAKING CHANGE FOR ALL PLATFORM PORTS THAT ALLOCATE BAKERY
LOCKS IN NORMAL MEMORY.
Change-Id: Ic3751c0066b8032dcbf9d88f1d4dc73d15f61d8b
Currently, on ARM platforms(ex. Juno) non-secure access to specific
peripheral regions, config registers which are inside and outside CSS
is done in the soc_css_security_setup(). This patch separates the CSS
security setup from the SOC security setup in the css_security_setup().
The CSS security setup involves programming of the internal NIC to
provide access to regions inside the CSS. This is needed only in
Juno, hence Juno implements it in its board files as css_init_nic400().
Change-Id: I95a1fb9f13f9b18fa8e915eb4ae2f15264f1b060
On Juno and FVP platforms, the Non-Secure System timer corresponds
to frame 1. However, this is a platform-specific decision and it
shouldn't be hard-coded. Hence, this patch introduces
PLAT_ARM_NSTIMER_FRAME_ID which should be used by all ARM platforms
to specify the correct non-secure timer frame.
Change-Id: I6c3a905d7d89200a2f58c20ce5d1e1d166832bba
This patch replaces the `ARM_TZC_BASE` constant with `PLAT_ARM_TZC_BASE` to
support different TrustZone Controller base addresses across ARM platforms.
Change-Id: Ie4e1c7600fd7a5875323c7cc35e067de0c6ef6dd
In certain Trusted OS implementations it is a requirement to pass them the
highest power level which will enter a power down state during a PSCI
CPU_SUSPEND or SYSTEM_SUSPEND API invocation. This patch passes this power level
to the SPD in the "max_off_pwrlvl" parameter of the svc_suspend() hook.
Currently, the highest power level which was requested to be placed in a low
power state (retention or power down) is passed to the SPD svc_suspend_finish()
hook. This hook is called after emerging from the low power state. It is more
useful to pass the highest power level which was powered down instead. This
patch does this by changing the semantics of the parameter passed to an SPD's
svc_suspend_finish() hook. The name of the parameter has been changed from
"suspend_level" to "max_off_pwrlvl" as well. Same changes have been made to the
parameter passed to the tsp_cpu_resume_main() function.
NOTE: THIS PATCH CHANGES THE SEMANTICS OF THE EXISTING "svc_suspend_finish()"
API BETWEEN THE PSCI AND SPD/SP IMPLEMENTATIONS. THE LATTER MIGHT NEED
UPDATES TO ENSURE CORRECT BEHAVIOUR.
Change-Id: If3a9d39b13119bbb6281f508a91f78a2f46a8b90
BL2 loads secure runtime code(BL3-1, BL3-2) and hence it has to
run in secure world otherwise BL3-1/BL3-2 have to execute from
non-secure memory. Hence, This patch removes the change_security_state()
call in bl1_run_bl2() and replaces it with an assert to confirm
the BL2 as secure.
FixesARM-software/tf-issues#314
Change-Id: I611b83f5c4090e58a76a2e950b0d797b46df3c29
ARM TF configures all interrupts as non-secure except those which
are present in irq_sec_array. This patch updates the irq_sec_array
with the missing secure interrupts for ARM platforms.
It also updates the documentation to be inline with the latest
implementation.
FixesARM-software/tf-issues#312
Change-Id: I39956c56a319086e3929d1fa89030b4ec4b01fcc
This patch adds macros suitable for programming the Advanced
SIMD/Floating-point (only Cortex-A53), CPU and L2 dynamic
retention control policy in the CPUECTLR_EL1 and L2ECTLR
registers.
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
BL3-2 image (Secure Payload) is optional. If the image cannot be
loaded a warning message is printed and the boot process continues.
According to the TBBR document, this behaviour should not apply in
case of an authentication error, where the boot process should be
aborted.
This patch modifies the load_auth_image() function to distinguish
between a load error and an authentication error. The caller uses
the return value to abort the boot process or continue.
In case of authentication error, the memory region used to store
the image is wiped clean.
Change-Id: I534391d526d514b2a85981c3dda00de67e0e7992
This patch reworks the PSCI generic implementation to conform to ARM
Trusted Firmware coding guidelines as described here:
https://github.com/ARM-software/arm-trusted-firmware/wiki
This patch also reviews the use of signed data types within PSCI
Generic code and replaces them with their unsigned counterparts wherever
they are not appropriate. The PSCI_INVALID_DATA macro which was defined
to -1 is now replaced with PSCI_INVALID_PWR_LVL macro which is defined
to PLAT_MAX_PWR_LVL + 1.
Change-Id: Iaea422d0e46fc314e0b173c2b4c16e0d56b2515a
This patch adds the necessary documentation updates to porting_guide.md
for the changes in the platform interface mandated as a result of the new
PSCI Topology and power state management frameworks. It also adds a
new document `platform-migration-guide.md` to aid the migration of existing
platform ports to the new API.
The patch fixes the implementation and callers of
plat_is_my_cpu_primary() to use w0 as the return parameter as implied by
the function signature rather than x0 which was used previously.
Change-Id: Ic11e73019188c8ba2bd64c47e1729ff5acdcdd5b
This patch implements the platform power managment handler to verify
non secure entrypoint for ARM platforms. The handler ensures that the
entry point specified by the normal world during CPU_SUSPEND, CPU_ON
or SYSTEM_SUSPEND PSCI API is a valid address within the non secure
DRAM.
Change-Id: I4795452df99f67a24682b22f0e0967175c1de429
As per PSCI1.0 specification, the error code to be returned when an invalid
non secure entrypoint address is specified by the PSCI client for CPU_SUSPEND,
CPU_ON or SYSTEM_SUSPEND must be PSCI_E_INVALID_ADDRESS. The current PSCI
implementation returned PSCI_E_INVAL_PARAMS. This patch rectifies this error
and also implements a common helper function to validate the entrypoint
information to be used across these PSCI API implementations.
Change-Id: I52d697d236c8bf0cd3297da4008c8e8c2399b170
Since there is a unique warm reset entry point, the FVP and Juno
port can use a single mailbox instead of maintaining one per core.
The mailbox gets programmed only once when plat_setup_psci_ops()
is invoked during PSCI initialization. This means mailbox is not
zeroed out during wakeup.
Change-Id: Ieba032a90b43650f970f197340ebb0ce5548d432
This patch adds support to the Juno and FVP ports for composite power states
with both the original and extended state-id power-state formats. Both the
platform ports use the recommended state-id encoding as specified in
Section 6.5 of the PSCI specification (ARM DEN 0022C). The platform build flag
ARM_RECOM_STATE_ID_ENC is used to include this support.
By default, to maintain backwards compatibility, the original power state
parameter format is used and the state-id field is expected to be zero.
Change-Id: Ie721b961957eaecaca5bf417a30952fe0627ef10
This patch migrates ARM reference platforms, Juno and FVP, to the new platform
API mandated by the new PSCI power domain topology and composite power state
frameworks. The platform specific makefiles now exports the build flag
ENABLE_PLAT_COMPAT=0 to disable the platform compatibility layer.
Change-Id: I3040ed7cce446fc66facaee9c67cb54a8cd7ca29
This patch migrates the rest of Trusted Firmware excluding Secure Payload and
the dispatchers to the new platform and context management API. The per-cpu
data framework APIs which took MPIDRs as their arguments are deleted and only
the ones which take core index as parameter are retained.
Change-Id: I839d05ad995df34d2163a1cfed6baa768a5a595d
This patch defines deprecated platform APIs to enable Trusted
Firmware components like Secure Payload and their dispatchers(SPD)
to continue to build and run when platform compatibility is disabled.
This decouples the migration of platform ports to the new platform API
from SPD and enables them to be migrated independently. The deprecated
platform APIs defined in this patch are : platform_get_core_pos(),
platform_get_stack() and platform_set_stack().
The patch also deprecates MPIDR based context management helpers like
cm_get_context_by_mpidr(), cm_set_context_by_mpidr() and cm_init_context().
A mechanism to deprecate APIs and identify callers of these APIs during
build is introduced, which is controlled by the build flag WARN_DEPRECATED.
If WARN_DEPRECATED is defined to 1, the users of the deprecated APIs will be
flagged either as a link error for assembly files or compile time warning
for C files during build.
Change-Id: Ib72c7d5dc956e1a74d2294a939205b200f055613
This commit does the switch to the new PSCI framework implementation replacing
the existing files in PSCI folder with the ones in PSCI1.0 folder. The
corresponding makefiles are modified as required for the new implementation.
The platform.h header file is also is switched to the new one
as required by the new frameworks. The build flag ENABLE_PLAT_COMPAT defaults
to 1 to enable compatibility layer which let the existing platform ports to
continue to build and run with minimal changes.
The default weak implementation of platform_get_core_pos() is now removed from
platform_helpers.S and is provided by the compatibility layer.
Note: The Secure Payloads and their dispatchers still use the old platform
and framework APIs and hence it is expected that the ENABLE_PLAT_COMPAT build
flag will remain enabled in subsequent patch. The compatibility for SPDs using
the older APIs on platforms migrated to the new APIs will be added in the
following patch.
Change-Id: I18c51b3a085b564aa05fdd98d11c9f3335712719
The new PSCI topology framework and PSCI extended State framework introduces
a breaking change in the platform port APIs. To ease the migration of the
platform ports to the new porting interface, a compatibility layer is
introduced which essentially defines the new platform API in terms of the
old API. The old PSCI helpers to retrieve the power-state, its associated
fields and the highest coordinated physical OFF affinity level of a core
are also implemented for compatibility. This allows the existing
platform ports to work with the new PSCI framework without significant
rework. This layer will be enabled by default once the switch to the new
PSCI framework is done and is controlled by the build flag ENABLE_PLAT_COMPAT.
Change-Id: I4b17cac3a4f3375910a36dba6b03d8f1700d07e3
There used to be 2 warm reset entry points:
- the "on finisher", for when the core has been turned on using a
PSCI CPU_ON call;
- the "suspend finisher", entered upon resumption from a previous
PSCI CPU_SUSPEND call.
The appropriate warm reset entry point used to be programmed into the
mailboxes by the power management hooks.
However, it is not required to provide this information to the PSCI
entry point code, as it can figure it out by itself. By querying affinity
info state, a core is able to determine on which execution path it is.
If the state is ON_PENDING then it means it's been turned on else
it is resuming from suspend.
This patch unifies the 2 warm reset entry points into a single one:
psci_entrypoint(). The patch also implements the necessary logic
to distinguish between the 2 types of warm resets in the power up
finisher.
The plat_setup_psci_ops() API now takes the
secure entry point as an additional parameter to enable the platforms
to configure their mailbox. The platform hooks `pwr_domain_on`
and `pwr_domain_suspend` no longer take secure entry point as
a parameter.
Change-Id: I7d1c93787b54213aefdbc046b8cd66a555dfbfd9
The state-id field in the power-state parameter of a CPU_SUSPEND call can be
used to describe composite power states specific to a platform. The current PSCI
implementation does not interpret the state-id field. It relies on the target
power level and the state type fields in the power-state parameter to perform
state coordination and power management operations. The framework introduced
in this patch allows the PSCI implementation to intepret generic global states
like RUN, RETENTION or OFF from the State-ID to make global state coordination
decisions and reduce the complexity of platform ports. It adds support to
involve the platform in state coordination which facilitates the use of
composite power states and improves the support for entering standby states
at multiple power domains.
The patch also includes support for extended state-id format for the power
state parameter as specified by PSCIv1.0.
The PSCI implementation now defines a generic representation of the power-state
parameter. It depends on the platform port to convert the power-state parameter
(possibly encoding a composite power state) passed in a CPU_SUSPEND call to this
representation via the `validate_power_state()` plat_psci_ops handler. It is an
array where each index corresponds to a power level. Each entry contains the
local power state the power domain at that power level could enter.
The meaning of the local power state values is platform defined, and may vary
between levels in a single platform. The PSCI implementation constrains the
values only so that it can classify the state as RUN, RETENTION or OFF as
required by the specification:
* zero means RUN
* all OFF state values at all levels must be higher than all RETENTION
state values at all levels
* the platform provides PLAT_MAX_RET_STATE and PLAT_MAX_OFF_STATE values
to the framework
The platform also must define the macros PLAT_MAX_RET_STATE and
PLAT_MAX_OFF_STATE which lets the PSCI implementation find out which power
domains have been requested to enter a retention or power down state. The PSCI
implementation does not interpret the local power states defined by the
platform. The only constraint is that the PLAT_MAX_RET_STATE <
PLAT_MAX_OFF_STATE.
For a power domain tree, the generic implementation maintains an array of local
power states. These are the states requested for each power domain by all the
cores contained within the domain. During a request to place multiple power
domains in a low power state, the platform is passed an array of requested
power-states for each power domain through the plat_get_target_pwr_state()
API. It coordinates amongst these states to determine a target local power
state for the power domain. A default weak implementation of this API is
provided in the platform layer which returns the minimum of the requested
power-states back to the PSCI state coordination.
Finally, the plat_psci_ops power management handlers are passed the target
local power states for each affected power domain using the generic
representation described above. The platform executes operations specific to
these target states.
The platform power management handler for placing a power domain in a standby
state (plat_pm_ops_t.pwr_domain_standby()) is now only used as a fast path for
placing a core power domain into a standby or retention state should now be
used to only place the core power domain in a standby or retention state.
The extended state-id power state format can be enabled by setting the
build flag PSCI_EXTENDED_STATE_ID=1 and it is disabled by default.
Change-Id: I9d4123d97e179529802c1f589baaa4101759d80c
This patch removes the assumption in the current PSCI implementation that MPIDR
based affinity levels map directly to levels in a power domain tree. This
enables PSCI generic code to support complex power domain topologies as
envisaged by PSCIv1.0 specification. The platform interface for querying
the power domain topology has been changed such that:
1. The generic PSCI code does not generate MPIDRs and use them to query the
platform about the number of power domains at a particular power level. The
platform now provides a description of the power domain tree on the SoC
through a data structure. The existing platform APIs to provide the same
information have been removed.
2. The linear indices returned by plat_core_pos_by_mpidr() and
plat_my_core_pos() are used to retrieve core power domain nodes from the
power domain tree. Power domains above the core level are accessed using a
'parent' field in the tree node descriptors.
The platform describes the power domain tree in an array of 'unsigned
char's. The first entry in the array specifies the number of power domains at
the highest power level implemented in the system. Each susbsequent entry
corresponds to a power domain and contains the number of power domains that are
its direct children. This array is exported to the generic PSCI implementation
via the new `plat_get_power_domain_tree_desc()` platform API.
The PSCI generic code uses this array to populate its internal power domain tree
using the Breadth First Search like algorithm. The tree is split into two
arrays:
1. An array that contains all the core power domain nodes
2. An array that contains all the other power domain nodes
A separate array for core nodes allows certain core specific optimisations to
be implemented e.g. remove the bakery lock, re-use per-cpu data framework for
storing some information.
Entries in the core power domain array are allocated such that the
array index of the domain is equal to the linear index returned by
plat_core_pos_by_mpidr() and plat_my_core_pos() for the MPIDR
corresponding to that domain. This relationship is key to be able to use
an MPIDR to find the corresponding core power domain node, traverse to higher
power domain nodes and index into arrays that contain core specific
information.
An introductory document has been added to briefly describe the new interface.
Change-Id: I4b444719e8e927ba391cae48a23558308447da13
This patch introduces new platform APIs and context management helper APIs
to support the new topology framework based on linear core position. This
framework will be introduced in the follwoing patch and it removes the
assumption that the MPIDR based affinity levels map directly to levels
in a power domain tree. The new platforms APIs and context management
helpers based on core position are as described below:
* plat_my_core_pos() and plat_core_pos_by_mpidr()
These 2 new mandatory platform APIs are meant to replace the existing
'platform_get_core_pos()' API. The 'plat_my_core_pos()' API returns the
linear index of the calling core and 'plat_core_pos_by_mpidr()' returns
the linear index of a core specified by its MPIDR. The latter API will also
validate the MPIDR passed as an argument and will return an error code (-1)
if an invalid MPIDR is passed as the argument. This enables the caller to
safely convert an MPIDR of another core to its linear index without querying
the PSCI topology tree e.g. during a call to PSCI CPU_ON.
Since the 'plat_core_pos_by_mpidr()' API verifies an MPIDR, which is always
platform specific, it is no longer possible to maintain a default implementation
of this API. Also it might not be possible for a platform port to verify an
MPIDR before the C runtime has been setup or the topology has been initialized.
This would prevent 'plat_core_pos_by_mpidr()' from being callable prior to
topology setup. As a result, the generic Trusted Firmware code does not call
this API before the topology setup has been done.
The 'plat_my_core_pos' API should be able to run without a C runtime.
Since this API needs to return a core position which is equal to the one
returned by 'plat_core_pos_by_mpidr()' API for the corresponding MPIDR,
this too cannot have default implementation and is a mandatory API for
platform ports. These APIs will be implemented by the ARM reference platform
ports later in the patch stack.
* plat_get_my_stack() and plat_set_my_stack()
These APIs are the stack management APIs which set/return stack addresses
appropriate for the calling core. These replace the 'platform_get_stack()' and
'platform_set_stack()' APIs. A default weak MP version and a global UP version
of these APIs are provided for the platforms.
* Context management helpers based on linear core position
A set of new context management(CM) helpers viz cm_get_context_by_index(),
cm_set_context_by_index(), cm_init_my_context() and cm_init_context_by_index()
are defined which are meant to replace the old helpers which took MPIDR
as argument. The old CM helpers are implemented based on the new helpers to
allow for code consolidation and will be deprecated once the switch to the new
framework is done.
Change-Id: I89758632b370c2812973a4b2efdd9b81a41f9b69
As per Section 4.2.2. in the PSCI specification, the term "affinity"
is used in the context of describing the hierarchical arrangement
of cores. This often, but not always, maps directly to the processor
power domain topology of the system. The current PSCI implementation
assumes that this is always the case i.e. MPIDR based levels of
affinity always map to levels in a power domain topology tree.
This patch is the first in a series of patches which remove this
assumption. It removes all occurences of the terms "affinity
instances and levels" when used to describe the power domain
topology. Only the terminology is changed in this patch. Subsequent
patches will implement functional changes to remove the above
mentioned assumption.
Change-Id: Iee162f051b228828310610c5a320ff9d31009b4e
This patch optimizes the invocation of the platform power management hooks for
ON, OFF and SUSPEND such that they are called only for the highest affinity
level which will be powered off/on. Earlier, the hooks were being invoked for
all the intermediate levels as well.
This patch requires that the platforms migrate to the new semantics of the PM
hooks. It also removes the `state` parameter from the pm hooks as the `afflvl`
parameter now indicates the highest affinity level for which power management
operations are required.
Change-Id: I57c87931d8a2723aeade14acc710e5b78ac41732
This patch creates a copy of the existing PSCI files and related psci.h and
platform.h header files in a new `PSCI1.0` directory. The changes for the
new PSCI power domain topology and extended state-ID frameworks will be
added incrementally to these files. This incremental approach will
aid in review and in understanding the changes better. Once all the
changes have been introduced, these files will replace the existing PSCI
files.
Change-Id: Ibb8a52e265daa4204e34829ed050bddd7e3316ff
If Trusted Firmware is built with optimizations disabled (-O0), the
linker throws the following error:
undefined reference to 'xxx'
Where 'xxx' is a raw inline function defined in a header file. The
reason is that, with optimizations disabled, GCC may decide to skip
the inlining. If that is the case, an external definition to the
compilation unit must be provided. Because no external definition
is present, the linker throws the error.
This patch fixes the problem by declaring the following inline
functions static, so the internal definition is used:
inline void soc_css_security_setup(void)
inline const arm_config_t *get_arm_config(void)
Change-Id: Id650d6be1b1396bdb48af1ac8a4c7900d212e95f
Denver is NVIDIA's own custom-designed, 64-bit, dual-core CPU which is
fully ARMv8 architecture compatible. Each of the two Denver cores
implements a 7-way superscalar microarchitecture (up to 7 concurrent
micro-ops can be executed per clock), and includes a 128KB 4-way L1
instruction cache, a 64KB 4-way L1 data cache, and a 2MB 16-way L2
cache, which services both cores.
Denver implements an innovative process called Dynamic Code Optimization,
which optimizes frequently used software routines at runtime into dense,
highly tuned microcode-equivalent routines. These are stored in a
dedicated, 128MB main-memory-based optimization cache. After being read
into the instruction cache, the optimized micro-ops are executed,
re-fetched and executed from the instruction cache as long as needed and
capacity allows.
Effectively, this reduces the need to re-optimize the software routines.
Instead of using hardware to extract the instruction-level parallelism
(ILP) inherent in the code, Denver extracts the ILP once via software
techniques, and then executes those routines repeatedly, thus amortizing
the cost of ILP extraction over the many execution instances.
Denver also features new low latency power-state transitions, in addition
to extensive power-gating and dynamic voltage and clock scaling based on
workloads.
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
danh-arm
Juan Castillo
Yatharth Kochar
Soby Mathew
Sandrine Bailleux
Achin Gupta
Vikram Kanigiri
Juan Castillo
Varun Wadekar
Andrew Thoelke
Jimmy Huang