As rk3399 reported the d8/octane scores drop 10% with cpu idle.
The root cause is thc cpu cluster enter the slow mode.
We don't need switch the clock to 24MHz if cpu cluster enter the
retention mode. In order to improve performance, it just needs for
cluster enter powering off mode.
Also, we shouldn't do anything for hlvl if the system is off.
Change-Id: I2a02962a01343abd0cba47ed63192c1cdf88b119
For the PMU design, we don't expect to get the interrupts before enter
the power mode. Since that will cause the confusion for the state
machine in the power mode.
Change-Id: Id8dee79ae617a66271b5caf92caf35f520f45099
If we don't enable the Schmitt trigger on the 32 kHz clock then systems
won't always resume from suspend properly. Presumably anything else in
the system that relies on the 32 kHz clock also will have problems
without the Schmitt trigger enabled.
Enable it always since having the 32 kHz clock on GPIO0_A0 isn't
exactly an optional feature, so all boards using rk3399 will need this.
Change-Id: Idc18c6cd1adc5be5f60efd9cb805d83d5cd40129
add auto_pd_dis_freq parameter, we can pass a frequency from kernel
to disable or enable ddr auto power down function.
Change-Id: Ie30914701336c59047c380381c6b75dd76a89562
add dram driver, and kernel can through sip function talk to bl31 to
do ddr frequency scaling. and ddr auto powerdown.
Change-Id: I0d0f2869aed95e336c6e23ba96a9310985c84840
In a previous change we mistakenly thought that PMU_24M_EN_CFG directly
controlled whether the PMU counts ran off the 32k vs. 24M clock.
Apparently that's not true. Real logic is now documented in code.
Also in the previous change we mistaknely though that PMU_24M_EN_CFG was
normally supposed to be 1 and we should "restore" it at resume time.
This is a terrible idea and made the system totally unreliable after
resume. Apparently PMU_24M_EN_CFG should always be 0 with all the
current code and settings.
Let's fix the above two problems. While we're changing all of this,
let's also:
1. Init at boot time. Many of these counts are used when the system is
running normally. We want the behavior at boot to match the behavior
after suspend/resume.
2. Init CPU counts to be 1 us. Although old code was trying to set this
to 1 ms (1000x slower) at suspend/resume time, we've been testing the
kernel with 1 us for a long time now. That's because the kernel (at
boot time) set these values to 24. Let's keep at 24 until we know
that's wrong.
3. Init GPU counts to be 1 us. Old code wasn't touching the GPU, but as
documented in comments it makes sense to init here. Do it.
4. Document the crap out of this code, since the SoC's behavior is
confusing and poorly documented in the TRM.
5. Increase some stabilization times to 30 ms (from 3 ms). It's unclear
that a full 30 ms is needed, but let's be safe for now.
This also inits the counts for the GPU.
(Thanks to Doug's patch that come from https://crosreview.com/372381)
Change-Id: Id1bc159a5a99916aeab043895e5c4585c4adab22
This patch adds a WFI instruction in the default implementations of
plat_error_handler() and plat_panic_handler(). This potentially reduces
power consumption by allowing the hardware to enter a low-power state.
The same change has been made to the FVP and Juno platform ports.
Change-Id: Ia4e6e1e5bf1ed42efbba7d0ebbad7be8d5f9f173
The pmusgrf reset-hold bits needs to be released, since the
pmusgrf reset-hold bits needs to be held.
Change-Id: Ia1eccc8fba18294f26b4cc07d47bc5e513dd9a1f
Sometimes this will cause the long delay for suspend/resume.
Since the 24M OCS will be turned off in power mode.
Also, remove the ERROR_DEPRECATED config define.
Change-Id: I78f21c35912c2250972e551695cdacc7bc4c020a
This patch updates some things for rk3399, as following:
1) Add the new file to handle the pwm. (e.g. the pwm regulator)
Make sure that good deal with the pwm related things.
Also, remove some pwm setting for pmu.c.
2) Set the plls slow mode and bypass in suspend, and restore them.
Change-Id: I112806700bf433c87763aac23d22fa7e6a7f5264
This patch implements the support for SP_MIN in FVP. The SP_MIN platform
APIs are implemented and the required makefile support is added for FVP.
Change-Id: Id50bd6093eccbd5e38894e3fd2b20d5baeac5452
This patch adds AArch32 support for FVP and implements common platform APIs
like `plat_get_my_stack`, `plat_set_my_stack`, `plat_my_core_cos` for AArch32.
Only Multi Processor(MP) implementations of these functions are considered in
this patch. The ARM Standard platform layer helpers are implemented for
AArch32 and the common makefiles are modified to cater for both AArch64 and
AArch32 builds. Compatibility with the deprecated platform API is not
supported for AArch32.
Change-Id: Iad228400613eec91abf731b49e21a15bcf2833ea
This patch migrates the upstream platform makefiles to include the
console drivers from the new location in ARM Trusted Firmware code
base.
Change-Id: I866d6c4951e475de1f836ce8a8c1d5e6da9577e3
1. Remove the AP_PWROFF in ATF, should configure it in kernel.
2. Save and restore the PWMs pin/regs for suspend/resume.
3. The pmusgrf reset-hold bits needs to be released. since the
pmusgrf reset-hold bits needs to be held.
4. Configure the PMU power up/down cycles about delay 3ms.
5. With the MMIO register block as one big mapping.
6. Fix the build error with psci_entrypoint since PSCI lib updated.
Fixes the commit
9ec78bd ("rockchip: support the suspend/resume for rk3399").
Change-Id: I112806700bf433c87763aac23d22fa7e6a7f5264
Compile option `ARM_BOARD_OPTIMISE_MMAP` has been renamed to
`ARM_BOARD_OPTIMISE_MEM` because it now applies not only to defines
related to the translation tables but to the image size as well.
The defines `PLAT_ARM_MAX_BL1_RW_SIZE`, `PLAT_ARM_MAX_BL2_SIZE` and
`PLAT_ARM_MAX_BL31_SIZE` have been moved to the file board_arm_def.h.
This way, ARM platforms no longer have to set their own values if
`ARM_BOARD_OPTIMISE_MEM=0` and they can specify optimized values
otherwise. The common sizes have been set to the highest values used
for any of the current build configurations.
This is needed because in some build configurations some images are
running out of space. This way there is a common set of values known
to work for all of them and it can be optimized for each particular
platform if needed.
The space reserved for BL2 when `TRUSTED_BOARD_BOOT=0` has been
increased. This is needed because when memory optimisations are
disabled the values for Juno of `PLAT_ARM_MMAP_ENTRIES` and
`MAX_XLAT_TABLES` are higher. If in this situation the code is
compiled in debug mode and with "-O0", the code won't fit.
Change-Id: I70a3d8d3a0b0cad1d6b602c01a7ea334776e718e
Previously the definition of `plat_get_syscnt_freq2()` in `arm_common.c` was
conditionally defined based on the ERROR_DEPRECATED flag. This patch makes
this function available irrespective of the flag and removes the deprecated
`plat_get_syscnt_freq()` definition.
Change-Id: I250ca787ca1b5e867096c6ba8f2bb444db44c97b
The PRELOADED_BL33_BASE build option allows to preload a BL33 and bypass its
loading by BL2. In ARM standard platforms, the conditional behaviour of
PRELOADED_BL33_BASE is moved within the implementation of
`plat_get_ns_image_entrypoint()` so that all callers may benefit from this
feature.
Change-Id: Iea060e204ec72f8081087837854535c4e320da4e
This patch moves the `arm_common.c` file from `plat/arm/common/aarch64/`
to the parent directory since the functions implemented in the file are
not AArch64 specific. The platform makefiles are also modified for this
change.
Change-Id: I776d2e4958f59041476cf2f53a9adb5b2d304ee0
The `plat_psci_common.c` was moved to the new location `plat/common`
and a stub file was retained at previous location for compatibility. This
patch modifies the platform makefiles to include the file from the new
location.
Change-Id: Iabddeeb824e9a5d72d176d7c644735966c8c0699
This patch moves the PSCI services and BL31 frameworks like context
management and per-cpu data into new library components `PSCI` and
`el3_runtime` respectively. This enables PSCI to be built independently from
BL31. A new `psci_lib.mk` makefile is introduced which adds the relevant
PSCI library sources and gets included by `bl31.mk`. Other changes which
are done as part of this patch are:
* The runtime services framework is now moved to the `common/` folder to
enable reuse.
* The `asm_macros.S` and `assert_macros.S` helpers are moved to architecture
specific folder.
* The `plat_psci_common.c` is moved from the `plat/common/aarch64/` folder
to `plat/common` folder. The original file location now has a stub which
just includes the file from new location to maintain platform compatibility.
Most of the changes wouldn't affect platform builds as they just involve
changes to the generic bl1.mk and bl31.mk makefiles.
NOTE: THE `plat_psci_common.c` FILE HAS MOVED LOCATION AND THE STUB FILE AT
THE ORIGINAL LOCATION IS NOW DEPRECATED. PLATFORMS SHOULD MODIFY THEIR
MAKEFILES TO INCLUDE THE FILE FROM THE NEW LOCATION.
Change-Id: I6bd87d5b59424995c6a65ef8076d4fda91ad5e86
This patch reworks type usage in generic code, drivers and ARM platform files
to make it more portable. The major changes done with respect to
type usage are as listed below:
* Use uintptr_t for storing address instead of uint64_t or unsigned long.
* Review usage of unsigned long as it can no longer be assumed to be 64 bit.
* Use u_register_t for register values whose width varies depending on
whether AArch64 or AArch32.
* Use generic C types where-ever possible.
In addition to the above changes, this patch also modifies format specifiers
in print invocations so that they are AArch64/AArch32 agnostic. Only files
related to upcoming feature development have been reworked.
Change-Id: I9f8c78347c5a52ba7027ff389791f1dad63ee5f8
1.Fixes the suspend/resume some bugs.
2.Add the power domain for saving power consumption.
3.Add cpu clusters suspend for rk3399 SoCs
Change-Id: Id602779016b41d6281f4ba40a20229d909b28e46
Software executing in the normal world and in the trusted world at
exception levels lower than EL3 will request runtime services using the
SMC instruction.
See the documentation here:
https://github.com/ARM-software/arm-trusted-firmware/blob/master/docs/
rt-svc-writers-guide.md
This to be implemented as an EL3 Runtime Service in rockchip BL31
platform port, using the "SiP Service Call" range as specified in the
SMC Calling Convention.
This doesn't support any SMC yet, we will support it in later.
Change-Id: I0a638dd0b653c28b08f79d89f77ed7c69864017d
On ARM CSS platforms, the whole flash used to be mapped as executable.
This is not required, given that the flash is used to store the BL1
and FIP images and:
- The FIP is not executed in place, its images are copied to RAM
and executed from there.
- BL1 is executed in place from flash but only its code needs to be
mapped as executable and platform code takes care of re-mapping
BL1's read-only section as executable.
Therefore, this patch now maps the flash as non-executable by default
on these platforms. This increases security by restricting the
executable region to what is strictly needed.
This patch also adds some comments to clarify the memory mapping
attributes on these platforms.
Change-Id: I4db3c145508bea1f43fbe0f6dcd551e1aec1ecd3
This patch adds some verbose traces in the arm_setup_page_tables()
function to print the extents of the different memory regions it maps.
Change-Id: Ia3ae1053e7ebf3579601ff9238b0e3791eb1e9e4
The arm_setup_page_tables() function used to expect a single set of
addresses defining the extents of the whole read-only section, code
and read-only data mixed up, which was mapped as executable.
This patch changes this behaviour. arm_setup_page_tables() now
expects 2 separate sets of addresses:
- the extents of the code section;
- the extents of the read-only data section.
The code is mapped as executable, whereas the data is mapped as
execute-never. New #defines have been introduced to identify the
extents of the code and the read-only data section. Given that
all BL images except BL1 share the same memory layout and linker
script structure, these #defines are common across these images.
The slight memory layout differences in BL1 have been handled by
providing values specific to BL1.
Note that this patch also affects the Xilinx platform port, which
uses the arm_setup_page_tables() function. It has been updated
accordingly, such that the memory mappings on this platform are
unchanged. This is achieved by passing null values as the extents
of the read-only data section so that it is ignored. As a result,
the whole read-only section is still mapped as executable.
FixesARM-software/tf-issues#85
Change-Id: I1f95865c53ce6e253a01286ff56e0aa1161abac5
This patch changes the base address of the "total" Trusted SRAM region
seen by the BL2U image. It used to start just after BL2U's read-only
section (i.e. at address BL2U_RO_LIMIT), it now starts from the base
address of the BL2U image (i.e. at address BL2U_BASE). In other words,
the "total" memory region now includes BL2U's own read-only section.
This does not change BL2U's resulting memory mappings because the
read-only section was already mapped in BL2U, it just wasn't part of
this total memory region.
Change-Id: I2da16ac842469023b41904eaa8d13ed678d65671
At the moment, on ARM platforms, BL1 maps everything from BL1_RO_BASE
to BL1_RO_LIMIT. BL1_RO_LIMIT, as defined in the porting guide, is
the maximum address in Trusted ROM that BL1's actual content _can_
occupy. The actual portion of ROM occupied by BL1 can be less than
that, which means that BL1 might map more Trusted ROM than it actually
needs to.
This patch changes BL1's memory mappings on ARM platforms to restrict
the region of Trusted ROM it maps. It uses the symbols exported by
the linker to figure out the actual extents of BL1's ROM footprint.
This change increases the number of page tables used on FVP by 1.
On FVP, we used to map the whole Trusted ROM. As it is 64MB large,
we used to map it as blocks of 2MB using level-2 translation table
entries. We now need a finer-grained mapping, which requires an
additional level-3 translation table.
On ARM CSS platforms, the number of translation tables is unchanged.
The BL1 image resides in flash at address 0x0BEC0000. This address is
not aligned on a 2MB-boundary so a level-3 translation table was
already required to map this memory.
Change-Id: I317a93fd99c40e70d0f13cc3d7a570f05c6c61eb
This patch introduces a new header file: include/lib/utils.h.
Its purpose is to provide generic macros and helper functions that
are independent of any BL image, architecture, platform and even
not specific to Trusted Firmware.
For now, it contains only 2 macros: ARRAY_SIZE() and
IS_POWER_OF_TWO(). These were previously defined in bl_common.h and
xlat_tables.c respectively.
bl_common.h includes utils.h to retain compatibility for platforms
that relied on bl_common.h for the ARRAY_SIZE() macro. Upstream
platform ports that use this macro have been updated to include
utils.h.
Change-Id: I960450f54134f25d1710bfbdc4184f12c049a9a9
This patch introduces the arm_setup_page_tables() function to
set up page tables on ARM platforms. It replaces the
arm_configure_mmu_elx() functions and does the same thing except
that it doesn't enable the MMU at the end. The idea is to reduce
the amount of per-EL code that is generated by the C preprocessor
by splitting the memory regions definitions and page tables creation
(which is generic) from the MMU enablement (which is the only per-EL
configuration).
As a consequence, the call to the enable_mmu_elx() function has been
moved up into the plat_arch_setup() hook. Any other ARM standard
platforms that use the functions `arm_configure_mmu_elx()` must be
updated.
Change-Id: I6f12a20ce4e5187b3849a8574aac841a136de83d
The per-cpu stacks should be aligned to the cache-line size and
the `declare_stack` helper in asm_macros.S macro assumed a
cache-line size of 64 bytes. The platform defines the cache-line
size via CACHE_WRITEBACK_GRANULE macro. This patch modifies
`declare_stack` helper macro to derive stack alignment from the
platform defined macro.
Change-Id: I1e1b00fc8806ecc88190ed169f4c8d3dd25fe95b
This patch enables optional PSCI functions `PSCI_STAT_COUNT` and
`PSCI_STAT_RESIDENCY` for ARM standard platforms. The optional platform
API 'translate_power_state_by_mpidr()' is implemented for the Juno
platform. 'validate_power_state()' on Juno downgrades PSCI CPU_SUSPEND
requests for the system power level to the cluster power level.
Hence, it is not suitable for validating the 'power_state' parameter
passed in a PSCI_STAT_COUNT/RESIDENCY call.
Change-Id: I9548322676fa468d22912392f2325c2a9f96e4d2
Add build time option 'cadence1' for ZYNQMP_CONSOLE to select the 2nd
UART available in the SoC.
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Acked-by: Michal Simek <michal.simek@xilinx.com>
Tony Xie
davidcunado-arm
Caesar Wang
danh-arm
Sandrine Bailleux
Soby Mathew
Antonio Nino Diaz
Soren Brinkmann