The state argument of the pm_self_suspend API encodes the state to
which the APU intends to suspend. The state can be:
- PM_APU_STATE_CPU_IDLE - processor power down, all memories remain
on
- PM_APU_STATE_SUSPEND_TO_RAM - all processors powered down, L2$
powered down, all OCM banks in retention and DDR in
self-refresh.
The calls for setting requirements for L2$ and OCM banks are now
redundant and removed.
Signed-off-by: Filip Drazic <filip.drazic@aggios.com>
[ sb
- remove redundant #defines
]
Signed-off-by: Sören Brinkmann <soren.brinkmann@xilinx.com>
Implementation is based on arm_validate_power_state().
This function is called during CPU_SUSPEND PSCI call to validate
power_state parameter. If state is valid this function populate it
in req_state array as power domain level specific local state.
ATF platform migration guide chapter 2.2 defines this function as
mandatory for PSCIv1.0 CPU_SUSPEND support.
Signed-off-by: Stefan Krsmanovic <stefan.krsmanovic@aggios.com>
When moving the ATF into the DRAM address space an additional
translation table is required.
Reported-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
The OCM space was reorganized to use the space more efficiently. Adjust
the default ATF location to be aligned with other ZynqMP software
components.
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Arm provided error injection support. To enable this error injection,
we need to set L2DEIEN in L2ACTLR_EL1 register and L1DEIEN in
CPUACTLR_EL1 register.
This is needed for our cortexa53 edac linux driver testing.
These registers need write access from non secure EL1 i.e linux
at the time of setting the above bits.
Signed-off-by: Naga Sureshkumar Relli <nagasure@xilinx.com>
ZynqMP only supports builds with RESET_TO_BL31=1. Set this option
through the platform makefile on default.
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Currently the GICv3 driver mandates that platform populate
both G1S and G0 interrupts. However, it is possible that a
given platform is not interested in both the groups and
just needs to specify either one of them.
This patch modifies the `gicv3_rdistif_init()` & `gicv3_distif_init()`
functions to allow either G1S or G0 interrupts to be configured.
FixesARM-software/tf-issues#400
Change-Id: I43572b0e08ae30bed5af9334f25d35bf439b0d2b
This feature allows one to quickly verify that the expected
image is contained in the FIP without extracting the image and
running sha256sum(1) on it.
The sha256 digest is only shown when the verbose flag is used.
This change requires libssl-dev to be installed in order to build
Trusted Firmware. Previously, libssl-dev was optionally needed only
to support Trusted Board Boot configurations.
FixesARM-Software/tf-issues#124
Change-Id: Ifb1408d17f483d482bb270a589ee74add25ec5a6
As the checkpatch reports the warning or error.
plat/rockchip/common/plat_pm.c:96:
ERROR: do not set execute permissions for source files
plat/rockchip/rk3399/drivers/pmu/pmu.c:294:
ERROR: do not set execute permissions for source files
plat/rockchip/common/plat_pm.c:286: WARNING: line over 80 characters
plat/rockchip/common/plat_pm.c:287: WARNING: line over 80 characters
Change-Id: Ib347da21c56551c31df3f90f03777b13c75d5c26
for compatible 32bit and 64bit, we use 0x82xxxxxx as function ID,
we modify SIP call function return value to 32 bit.
Change-Id: Ib99b03a9ea423853aaa296dcc634ee82c622a552
For save power cosumption, if gpio power supply shut down, we need to
set gpio2 ~ gpio4 to input and HiZ status when suspend, and recovery
they status when rusume. we do it base on apio pass from loader.
Change-Id: I59fd2395e5e37e63425472a39f519822c9197e4c
some specific board need to disable/enable specific gpio when
suspend/resume, so we add this function, bootloader can pass the
specific gpio, and we can handle these gpios in bl31 suspend/resuem
function.
Change-Id: I373b03ef9202ee4a05a2b9caacdfa01b47ee2177
We may need gpio pull mode later, so add this function.
Besides fix a set pull mode bug, and save gpio clock gate,
when operate the gpio, we will enable gpio clock, when
finish gpio operate, restore gpio clock gate status.
Change-Id: Ia1d602804f571a17f5ddc499908663b968b02974
The `psci_plat_pm_ops` global pointer is initialized during cold boot by the
primary CPU and will be accessed by the secondary CPUs before enabling data
cache during warm boot. This patch adds a missing data cache flush of
`psci_plat_psci_ops` after initialization during psci_setup() so that
secondaries can see the updated `psci_plat_psci_ops` pointer.
FixesARM-software/tf-issues#424
Change-Id: Id4554800b5646302b944115a33be69507d53cedb
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
This patch fixes a bug in context management library when writing
SCTLR register during context initialization. The write happened
prior to initialization of the register context pointer. This
resulted in the compiler optimizing the write sequence from the
final binary and hence SCTLR remains uninitialized when
entering normal world. The bug is fixed by doing the
initialization of the register context pointer earlier in the
sequence.
Change-Id: Ic7465593a74534046b79f40446ffa1165c52ed76
Given the service name and timestamp id, this assembler macro
calculates the offset into a memory region where the per-cpu timestamp
value is located.
Change-Id: I47f6dfa2a17be182675e2ca0489d6eed42433209
More headers will be needed soon so better to move these to their own
directory to avoid cluttering include/lib.
Change-Id: I6a72dc5b602d6f51954cf60aadd1beb52a268670
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
Instead of hardcoding a level 1 table as the base translation level
table, let the code decide which level is the most appropriate given
the virtual address space size.
As the table granularity is 4 KB, this allows the code to select
level 0, 1 or 2 as base level for AArch64. This way, instead of
limiting the virtual address space width to 39-31 bits, widths of
48-25 bit can be used.
For AArch32, this change allows the code to select level 1 or 2
as the base translation level table and use virtual address space
width of 32-25 bits.
Also removed some unused definitions related to translation tables.
FixesARM-software/tf-issues#362
Change-Id: Ie3bb5d6d1a4730a26700b09827c79f37ca3cdb65
This patch removes the tight loop that calls `plat_report_exception`
in unhandled exceptions in AArch64 state.
The new behaviour is to call the `plat_report_exception` only
once followed by call to `plat_panic_handler`.
This allows platforms to take platform-specific action when
there is an unhandled exception, instead of always spinning
in a tight loop.
Note: This is a subtle break in behaviour for platforms that
expect `plat_report_exception` to be continuously executed
when there is an unhandled exception.
Change-Id: Ie2453804b9b7caf9b010ee73e1a90eeb8384e4e8
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
At the end of successful image authentication in load_auth_image(),
the data cache for the virtual address range corresponding to the
image is invalidated (by a call to inv_dcache_range()). The intent
seems to be to ensure the data caches do not contain any sensitive
data used during authentication, which subsequent code can read.
However, this same address range is already flushed (cleaned and
invalidated by a call to flush_dcache_range()) at the end of
load_image(), and the subsequent invalidate has no functional
effect.
This patch removes the redundant call to inv_dcache_range(). It
also moves the flush_dcache_range() call from the end of load_image()
to the end of load_auth_image(), so the image data will remain in
the caches during authentication, improving performance.
This also improves the comments that explain the rationale for
calling flush_dcache_range() after image loading/authentication.
Change-Id: I14f17ad2935075ef6f3d1327361c5088bfb2d284