psci_suspend_to_standby_finisher and psci_cpu_suspend_finish do mostly
the same stuff, besides the system management associated with their
respective wakeup paths. So bring the wake from standby path in line
with the wake from reset path - caller acquires locks and manages
context. This way both behave in vaguely the same way. We can also bring
their names in line so it's more apparent how they are different.
This is in preparation for cores waking from sleep, coming in another
patch. No functional change is expected.
Change-Id: I0e569d12f65d231606080faa0149d22efddc386d
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
The target_pwrlvl field in the psci cpu data struct only stores the
highest power domain that a CPU_SUSPEND call affected, and is used to
resume those same domains on warm reset. If the cpu is otherwise OFF
(never turned on or CPU_OFF), then this needs to be the highest power
level because we don't know the highest level that will be off.
So skip the invalidation and always keep the field to the maximum value.
During suspend the field will be lowered to the appropriate value and
then put back after wakeup.
Also, do that in the suspend to standby path as well as it will have
been written before the sleep and it might end up incorrect.
Change-Id: I614272ec387e1d83023c94700780a0f538a9a6b6
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
The comment was written when cache maintenance had to be considered when
calling this function. But that argument was dropped a while back and
this comment no longer makes any sense.
Change-Id: Ib68293f23cc3edca3010164dfe8866956b8e1a63
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
Patch fc7dca72ba656e5f147487b20f9f0fb6eb39e116 changed the owning
security states of the TRBE and SPE buffers to NS. The thinking was that
this would assist SMCCC feature availability to more easily determine
if the feature is enabled or disabled. However, that only changed bit 0
while the SMCCC feature only looks at bit 1 so this change is redundant.
It was also meant to tighten security but that was done by
73d98e3759 instead.
Annoyingly, FEAT_TRBE has TRBIDR_EL1 which reports that programming is
allowed when the current security state owns the buffer even when the
MDCR_EL3 setting disallows this in practice.
So revert the functional aspect of the patch as it causes linux panics
with ERRATA_A520_2938996. Keep the defines as they are used elsewhere.
Change-Id: I39463d585df89aee44d1996137616da85d678f41
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
* changes:
feat(el3_spmc): ffa error handling in direct msg
feat(ff-a): support FFA_MSG_SEND_DIRECT_REQ2/RESP2
feat(ff-a): add FFA_MEM_PERM_GET/SET_SMC64
feat(el3-spmc): support Hob list to boot S-EL0 SP
feat(synquacer): add support Hob creation
fix(fvp): exclude extend memory map TZC regions
feat(fvp): add StandaloneMm manifest in fvp
feat(spm): use xfer list with Hob list in SPM_MM
When an FFA_ERROR happens while handling a direct message
from normal world, return to normal world with
FFA_ERROR. Otherwise, the system would re-enter the secure partition
with FFA_ERROR.
Change-Id: I3d9a68a41b4815c1a8e10354cfcf68fec9f4b800
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
StandaloneMm which is S-EL0 partition uses
FFA_MSG_SEND_DIRECT_REQ2/RESP2 to handle multiple services.
For this, add support for FFA_MSG_SEND_DIRECT_REQ2/RESP2 in el3_spmc
restrictly up to use 8 registers.
although FF-A v1.2 defines FFA_MSG_SEND_DIRECT_REQ2/RESP2
with ability to pass/return up to 18 registers.
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
Change-Id: I8ab1c332d269d9d131330bb2debd10d75bdba1ee
Agilex5 platform:
Boot scratch COLD6 register is meant for Customer use only.
So, use Intel specific COLD3 register with [5:2]bits to
determine the warm reset and SMP boot requests.
Also handle the unaligned DEVICE/IO memory store and load
in the assembly entrypoint startup code.
Agilex, Stratix10, N5X platforms:
Use only the LSB 4bits [3:0] of the boot scratch COLD6 register
to detect the warm reset request.
Change-Id: I4fd6e63fe0bd42ddcb4a3f81c7a7295bdc8ca65f
Signed-off-by: Girisha Dengi <girisha.dengi@intel.com>
Signed-off-by: Jit Loon Lim <jit.loon.lim@altera.com>
If MTKLIB_PATH is provided, the build will use the library provided by
MTKLIB_PATH. Otherwise, it will use stub implementation.
Change-Id: I218e724231c8bbc6cc851a240c6bbc4f6f49f154
Signed-off-by: Gavin Liu <gavin.liu@mediatek.corp-partner.google.com>
Switch the state back to non-secure after a secure interrupt
using spmd_smc_switch_state with FFA_NORMAL_WORLD_RESUME
to reduce the number of control flow paths for world switches.
Fixes an issue where FP registers were not correctly restored
after secure interrupts.
Upstreamed from https://r.android.com/3345999, tested on Trusty.
Change-Id: I3ce33f7657c13b999969ebb8957d5d4b6c3aa634
Signed-off-by: Andrei Homescu <ahomescu@google.com>
Rename TC_FPGA_ANDROID_IMG_IN_RAM to TC_FPGA_FS_IMG_IN_RAM
to use it for debian loading to ram as well.
Change-Id: I70b68b06501d17dcebbe78bee8fec0a701106c92
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
* changes:
feat(qemu): hand off TPM event log via TL
feat(handoff): common API for TPM event log handoff
feat(handoff): transfer entry ID for TPM event log
fix(qemu): fix register convention in BL31 for qemu
fix(handoff): fix register convention in opteed
Commit 71c4443886ff ("fix(lib/rmm_el3_ifc): add console name to checksum
calculation") on TF-RMM updated the checksum calcualtion of the RMM
manifest to include the console names.
Include console names in the QEMU manifest to remain compatible with
RMM, just like commit aa99881d30 ("fix(rme): add console name to
checksum calculation") did for FVP.
Checksum calculation is done by adding together 64-bit values. Add a
helper that does this.
Change-Id: Ica6cab628160593830270bef1acdeb475d1c0c36
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
* changes:
feat(smccc): implement SMCCC_ARCH_FEATURE_AVAILABILITY
refactor(cm): clean up per-world context
refactor(cm): change owning security state when a feature is disabled
Modify ethernet base addr and irq numbers for TC4 FPGA in dts to match
with its RoS configuration.
Change-Id: I7b180c3eb90d7557d0011a25a742106f703cd264
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
Modify gpio controller base addr for TC4 FPGA in dts to match
with its RoS configuration.
Change-Id: Id4ad925d23937d302adfb3e0d4b1573e5ec717c1
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
TC4 FPGA DPU base addr and irq doesn't match with TC3 FPGA
so refactor the code to manage it accordingly.
Change-Id: Ie31933e0bcbd489945935829940a5c5434e6b1d7
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
Modify mmc base addr and irq numbers for TC4 FPGA in dts to match
with its RoS configuration.
Change-Id: Ie8fe1f1d3aef1c020ac85db7c3b81dfad3722e2f
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
TC4 FPGA have a UART clock of 4000000 so modify the value
of TC_UARTCLK for TC4.
Change-Id: I8de84d58bce8b7277bf356136a5185c008ab4c28
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
Implement stub functions for the EMI driver to ensure that the build
can pass when a prebuilt library is not available.
Change-Id: I296945a3df6766a3a133cd385a1e5038ca979403
Signed-off-by: Gavin Liu <gavin.liu@mediatek.corp-partner.google.com>
The arguments struct needs to be changed to remove a non-standard entry
(`scp_image_info[]`) and also makes use of a built-in arguments parser.
Since the `scp_image_info[]` entry is removed in U-Boot SPL-compatible builds,
SCP_BL2 image info is hardcoded.
Change-Id: Id3cc887c61c3b940c8a21d9da7f2b6845da51af8
Signed-off-by: Ferass El Hafidi <funderscore@postmarketos.org>
As per GPU team, this change should be helpful to improve
the performance.
Change-Id: I10a3fc1d0ddf1ba0a17da6dc4f2a80f5fe567db6
Signed-off-by: Vishnu Satheesh <vishnu.satheesh@arm.com>
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
Currently used stream id 0x200 gives below fault,
[ 9.547393][ C0] mali 2d000000.gpu: Unexpected Page fault in firmware address space at VA 0x0000000000000000
[ 9.547393][ C0] raw fault status: 0x400D02C0
[ 9.547393][ C0] exception type 0xC0: TRANSLATION_FAULT at level 0
[ 9.547393][ C0] access type 0x2: READ
As per the GPU team, GPU stream id is 0 on TC4-FPGA so change it.
Change-Id: I3aed62289c5b96fb850f0022ea7f5172c606eb95
Signed-off-by: Vishnu Satheesh <vishnu.satheesh@arm.com>
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
In the SMMU root port programming model, changing both
SMMU_ROOT_CR0.GPCEN and ACCESSEN bits in the same MMIO write operation
is permitted by the architecture but left to the SMMU IP implementation
to determine the order of completing one or the other operation.
Enforce more determinism by setting CR0.GPCEN, wait for CR0ACK.GPCEN
completion, then setting CR0.ACCESSEN and wait for CR0ACK.ACCESSEN
completion.
Signed-off-by: Olivier Deprez <olivier.deprez@arm.com>
Change-Id: I36ba5fbc13d06c6243226008d18a2d57477b0d28
Move platform.mk inclusion in top level Makefile to permit a platform
specifying BRANCH_PROTECTION option.
Signed-off-by: Olivier Deprez <olivier.deprez@arm.com>
Change-Id: I1f662f82cd949eedfdbb61b9f66de15c46fb3106
Similarly to BL1 and BL31, use EL3_PAS macro from xlat_tables header
(depends on ENABLE_RME) in BL2 to define MAP_BL2_TOTAL.
Signed-off-by: Olivier Deprez <olivier.deprez@arm.com>
Change-Id: I59a3b297efd2eacd082a297de6b579b7c9052883
If TRANSFER_LIST is enabled, hand off TPM event log via TL instead
of DT; otherwise fallback to legacy way if TRANSFER_LIST is off or
errors observed.
Moreover, for updating the TL from secure to non-secure
memory before existing EL3, replace memcpy with function
transfer_list_relocate() for more accuracy.
Change-Id: I1d6bcf573f91efe99380bc89195198a8583b1def
Signed-off-by: Raymond Mao <raymond.mao@linaro.org>
print ni-tower discovery tree to understand ni-tower hierarchy which
might be useful during debugging.
Change-Id: Ib49fef9c63f7740e04b4d8371c1083bd040f6e09
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
Set console baurate to 38400 for fvp as well for code
simplicity.
Change-Id: I58ba6b7043541f6eb67e32257307da4eba0bb28a
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
remove redundant macro UARTCLK_FREQ and replace it with TC_UARTCLK
in dts.
Change-Id: Id463a9ddc1588278e552ffca3dfb738676229ce7
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
SMCCC_ARCH_FEATURE_AVAILABILITY [1] is a call to query firmware about
the features it is aware of and enables. This is useful when a feature
is not enabled at EL3, eg due to an older FW image, but it is present in
hardware. In those cases, the EL1 ID registers do not reflect the usable
feature set and this call should provide the necessary information to
remedy that.
The call itself is very lightweight - effectively a sanitised read of
the relevant system register. Bits that are not relevant to feature
enablement are masked out and active low bits are converted to active
high.
The implementation is also very simple. All relevant, irrelevant, and
inverted bits combined into bitmasks at build time. Then at runtime the
masks are unconditionally applied to produce the right result. This
assumes that context managers will make sure that disabled features
do not have their bits set and the registers are context switched if
any fields in them make enablement ambiguous.
Features that are not yet supported in TF-A have not been added. On
debug builds, calling this function will fail an assert if any bits that
are not expected are set. In combination with CI this should allow for
this feature to to stay up to date as new architectural features are
added.
If a call for MPAM3_EL3 is made when MPAM is not enabled, the call
will return INVALID_PARAM, while if it is FEAT_STATE_CHECK, it will
return zero. This should be fairly consistent with feature detection.
The bitmask is meant to be interpreted as the logical AND of the
relevant ID registers. It would be permissible for this to return 1
while the ID returns 0. Despite this, this implementation takes steps
not to. In the general case, the two should match exactly.
Finally, it is not entirely clear whether this call replies to SMC32
requests. However, it will not, as the return values are all 64 bits.
[1]: https://developer.arm.com/documentation/den0028/galp1/?lang=en
Co-developed-by: Charlie Bareham <charlie.bareham@arm.com>
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
Change-Id: I1a74e7d0b3459b1396961b8fa27f84e3f0ad6a6f
In preparation for SMCCC_ARCH_FEATURE_AVAILABILITY, it is useful for
context to be directly related to the underlying system. Currently,
certain bits like SCR_EL3.APK are always set with the understanding that
they will only take effect if the feature is present.
However, that is problematic for SMCCC_ARCH_FEATURE_AVAILABILITY (an
SMCCC call to report which features firmware enables), as simply reading
the enable bit may contradict the ID register, like the APK bit above
for a system with no Pauth present.
This patch is to clean up these cases. Add a check for PAuth's presence
so that the APK bit remains unset if not present. Also move SPE and TRBE
enablement to only the NS context. They already only enable the features
for NS only and disable them for Secure and Realm worlds. This change
only makes these worlds' context read 0 for easy bitmasking.
There's only a single snag on SPE and TRBE. Currently, their fields have
the same values and any world asymmetry is handled by hardware. Since we
don't want to do that, the buffers' ownership will change if we just set
the fields to 0 for non-NS worlds. Doing that, however, exposes Secure
state to a potential denial of service attack - a malicious NS can
enable profiling and call an SMC. Then, the owning security state will
change and since no SPE/TRBE registers are contexted, Secure state will
start generating records. Always have NS world own the buffers to
prevent this.
Finally, get rid of manage_extensions_common() as it's just a level of
indirection to enable a single feature.
Change-Id: I487bd4c70ac3e2105583917a0e5499e0ee248ed9
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
New slew rate applied.
SDMMC:
- for SD card and eMMC:
- clk at 2.
- cmd and data at 1.
- for Wifi
- clk at 1.
- cmd and data at 0.
SDMMC1:
- for dk board:
- clk at 2.
- cmd and data at 1.
- for eval board:
- clk at 3.
- cmd and data at 2.
Change-Id: I2dfa62aca08a613e0532746050246fc8dc476ff8
Signed-off-by: Christophe Kerello <christophe.kerello@foss.st.com>
