When StandaloneMm used with SPM_MM, TF-A should create
PHIT Hob to boot it.
This patch supports Hob creation for StandaloneMm in synquacer platform.
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
Change-Id: Ifa3ae1f0aa37f389aabb14f48be307502ae6fc2c
The commit
1922875233 ("fix(spm-mm): carve out NS buffer TZC400 region")
removes overlaps of ns shared buffer in secure memory region.
Unfortunately, this separation increases 1 region and over maximum
number of TZC programmable regions when they include
extended memory map regions (DRAM3 to DRAM6).
This causes boot failure of StandaloneMm with spmc_el3 && sp_el0 with
ASSERT: drivers/arm/tzc/tzc400.c:256.
To fix this, like SPM_MM, exclude setting extended memory map regions when
it uses SPMC_AT_EL3 && SPC_AT_EL3_SEL0_SP.
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
Change-Id: I2d40bea066ca030050dfe951218cd17171010676
Support StandaloneMm running with FF-A as S-EL0 SP
when TF-A is built with EL3 SPMC partition manager.
For this
1. add manifest file describing StandaloneMm partition.
2. add number of page mapping area.
3. StandaloneMm should use SRAM with 512K.
while enabling, StandaloneMm, BL1 image requires more size:
aarch64-none-elf/bin/ld: BL31 image has exceeded its limit.
aarch64-none-elf/bin/ld: region `RAM' overflowed by 16384 bytes
So, when using SRAM size with 512K configuration,
increase size limit of BL1 binary.
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
Change-Id: Idaa1db510340ebb812cfd13588610b2eea941918
This is a small change adding accessor functions for the Debug Power
Control register (DBGPRCR_EL1) to the common architectural helpers.
Change-Id: I72261fbf0395d900347b46af320093ed946aa73d
Signed-off-by: Chris Kay <chris.kay@arm.com>
This variable had a misleading name, as it is the length
of the header only when the ROTPK is a hash.
Also rename arm_rotpk_header to match the new pattern.
Change-Id: I36c29998eebf50c356a6ca959ec9223c8837b540
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
No longer hard code SHA-256 hashed rsa dev keys,
now the keys can use pair of key alg: rsa, p256, p384
and hash alg: sha256, sha384, sha512.
All public keys are now generated at build-time from the dev
keys.
Change-Id: I669438b7d1cd319962c4a135bb0e204e44d7447e
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
No longer hard code SHA-256 hashed dev rotpks, instead
use the algorithm given by HASH_ALG. This means that
we no longer need the plat_arm_configs (once the protpk and
swd_rotpk are also updated to use HASH_ALG).
The rot public key is now generated at build time, as is
the header for the key.
Also support some default 3k and 4k RSA keys.
Change-Id: I33538124aeb4fa7d67918d878d17f2a84d3a6756
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
Commit a6485b2 ("refactor(delay-timer): add timer callback
functions") introduced a requirement for timer-related APIs
to have a timer object initialized before use. This caused
assertion failures in SMMU routines on Neoverse platforms,
as they relied on timer APIs.
Resolve this issue by initializing the timer early during
platform boot to set up the timer_ops object properly.
Change-Id: I3d9ababdb7897185f23e9ccf982b9aab6c666b8c
Signed-off-by: Sammit Joshi <sammit.joshi@arm.com>
This corrects the MISRA violation C2012-10.3:
The value of an expression shall not be assigned to an object with a
narrower essential type or of a different essential type category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I37ec9f8d716347df9acea5eb084f5a423a32a058
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.3:
The value of an expression shall not be assigned to an object with a
narrower essential type or of a different essential type category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: Ie82297e7eb5faa5d45b1a613c59516052e0c5ecb
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.3:
The value of an expression shall not be assigned to an object with a
narrower essential type or of a different essential type category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: Ie2d32d5554d251cde8a9c8b7c7a85666ea505a15
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.3:
The value of an expression shall not be assigned to an object with a
narrower essential type or of a different essential type category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I1606422aadfd64b283fd9948b6dadcddecdf61e0
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.3:
The value of an expression shall not be assigned to an object with a
narrower essential type or of a different essential type category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I141fbc554265173df0ca90c2ddc7f28137c6b0f1
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.1:
Operands shall not be of an inappropriate essential type.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: If0a6ffa84c4d1ce5ae08337a4eb20c9a221d7795
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.1:
Operands shall not be of an inappropriate essential type.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I41b08349fc6023458ffc6e126f58293a9ef37422
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.1:
Operands shall not be of an inappropriate essential type.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I675f1b2ac408b70a9ca307fb5161ebb8e597897c
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.1:
Operands shall not be of an inappropriate essential type.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I016f9df3811d80cd230257b5533d4d15a15fe14f
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.4:
Both operands of an operator in which the usual arithmetic
conversions are performed shall have the same essential type
category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: Ia352e3cf261b52777c1c431701e1e6c3be9cd279
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.4:
Both operands of an operator in which the usual arithmetic
conversions are performed shall have the same essential type
category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I5add78285ff0e48aa6c0fb639e7e2924f5bf9000
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.4:
Both operands of an operator in which the usual arithmetic
conversions are performed shall have the same essential type
category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: Ide520aa8ec900d0e23e80753d7082e34b6897e8f
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.4:
Both operands of an operator in which the usual arithmetic conversions
are performed shall have the same essential type category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I9110ea86f5ee49af0b21be78fd0890742ef95ddf
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.4:
Both operands of an operator in which the usual arithmetic conversions
are performed shall have the same essential type category.
The condition is explicitly checked against 0U, appending 'U' and
typecasting for unsigned comparison.
Change-Id: I847af07f5e4f139384c1ed50bee765b892a6e9cd
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.7:
If a composite expression is used as one operand of an operator in
which the usual arithmetic conversions are performed then the other
operand shall not have wider essential type.
Explicitly type casted to match the data type of both the operands.
Change-Id: I670304682cc4945b8575f125ac750d0dc69079a7
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.6:
The value of a composite expression shall not be assigned to
an object with wider essential type.
Explicitly type casted to match the data type of composite
expression.
Change-Id: I6497453f9f7455ae2f1ad8a18760ff0ef41d7c40
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
This corrects the MISRA violation C2012-10.6:
The value of a composite expression shall not be assigned to an
object with wider essential type.
Explicitly type casted to match the data type of composite
expression.
Change-Id: I0fd845496b4d6ac702027eb2075a23b15849f7d6
Signed-off-by: Nithin G <nithing@amd.com>
Signed-off-by: Maheedhar Bollapalli <maheedharsai.bollapalli@amd.com>
* changes:
feat(tc): add MCN PMU nodes in dts for TC4
feat(tc): add 'kaslr-seed' node in device tree for TC3
feat(tc): enable MCN non-secure access to pmu counters on TC4
feat(tc): define MCN related macros for TC4
MCN PMU counters are by default non-accesible from non-secure world,
so enable the non-secure access to those PMU counters so that linux
perf driver can read them.
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen Zeyada <Icen.Zeyada2@arm.com>
Change-Id: I1cf1f88f97e9062592fd5603a78fd36f15a15f89
Define MCN related macros for TC4 to add TC4 specific MCN PMU
nodes in dts and to enable MCN PMU NS access in further commits.
Signed-off-by: Jagdish Gediya <jagdish.gediya@arm.com>
Signed-off-by: Icen Zeyada <Icen.Zeyada2@arm.com>
Change-Id: Ifc02fcd833888a9953fac404585468316aa0168c
Add APU SMMU hardware semaphore operations to make APU SMMU
able to sync the power status.
Change-Id: I1926cab990fba54a2ea694ac6d9e87135dfb19cf
Signed-off-by: Karl Li <karl.li@mediatek.com>
1. Add smpu protection for APU secure memory.
2. Move emi mpu protection for mt8188 to platform folder
Since the smpu driver has not upstream, we currently leave the interface
and do nothing until smpu driver is ready.
Change-Id: Id70162e90a7deb64befe90f09a841a0903535482
Signed-off-by: Karl Li <karl.li@mediatek.com>
APU RCX is a sub-domain in apusys, connecting several APU components.
The APU RCX DevAPC control lives in APU and can only be set after
APU is powered on.
The APUSYS kernel driver will trigger RCX DevAPC init by smc call.
Change-Id: I3a9b014ea1be7ee80fd6861ad088f1dec5410872
Signed-off-by: Karl Li <karl.li@mediatek.com>
Add APU kernel control operations to provide the bootup init functions.
1. Add software workaround for certain operations on mt8196.
2. Add APU logger operations.
3. Add function to clear mbox spare register, which is used in APU
booting process.
4. Add function to setup CE binary to make sure the CE binary version
is align with the APU firmware.
Change-Id: Ic99adba1409c020c72179ea135e0d4291fc3f384
Signed-off-by: Karl Li <karl.li@mediatek.com>
1. Add APU power on/off functions
2. Refine the APU power on/off interface for mt8188
3. Add dcm setup function to support mt8196
Change-Id: Ie1caca40f89de71caac037fabe7e7455ff2a1872
Signed-off-by: Karl Li <karl.li@mediatek.com>
APUMMU is the MMU in APU, which is responsible for inner address
mapping. The APU kernel driver will setup the APUMMU by SMC call.
Change-Id: Iad7532883e42c288aeb0d23ab419f4dc6d8630f2
Signed-off-by: Karl Li <karl.li@mediatek.com>
Enable mt8196 apusys mailbox mpu protection and
move the mt8188 setting to platform folder
Change-Id: I76b68318bb88e56b12cdacd9e2b998699ca6b48e
Signed-off-by: Karl Li <karl.li@mediatek.com>
Remap the request from domain 5, 7, 14 to domain 6 and setup security
sideband
Change-Id: I06d377f4bcc542bf22e0a04ffb45cf52b7528a75
Signed-off-by: Karl Li <karl.li@mediatek.com>
Apusys AO DevAPC is a set of control registers inside APU, controlling
the access permission of APU AO (Always On) domain.
This patch add the mt8196 APU AO DevAPC setting to setup the protection.
Change-Id: I975a92795031cd1813442302890e29b671ef16f1
Signed-off-by: Karl Li <karl.li@mediatek.com>
Add the APU (AI processing unit) power init flow to prepare the hardware
setting before using APU power functions.
Change-Id: I595b1d5100a4f083263de6527f920e5168700b7a
Signed-off-by: Karl Li <karl.li@mediatek.com>
Mbed-TLS requires platforms to allocate it a heap for it's own internal
usage. This heap is typically between shared by BL1 and BL2 to conserve
memory.The base address and size of the heap are conveyed from BL1 to
BL2 through the config TB_FW_CONFIG.
This slightly awkward approach necessitates declaring a placeholder node
in the DTS. At runtime, this node is populated with the actual values of
the heap information. Instead, since this is dynamic information, and
simple to represent through C structures, transmit it to later stages
using the firmware handoff framework.
With this migration, remove references to TB_FW_CONFIG when firmware
handoff is enabled, as it is no longer needed. The setup code now relies
solely on TL structures to configure the TB firmware
Change-Id: Iff00dc742924a055b8bd304f15eec03ce3c6d1ef
Signed-off-by: Harrison Mutai <harrison.mutai@arm.com>
The initialization logic for the secure transfer list is currently
scattered and duplicated across platform setup code. This not only leads
to inefficiency but also complicates access to transfer lists from other
parts of the code without invoking setup functions. For instance,
arm_bl2_setup_next_ep_info acts as a thin wrapper in arm_bl2_setup.c to
provide access to the secure transfer list.
To streamline the interface, all setup code has been consolidated into a
central location.
Change-Id: I99d2a567ff39df88baa57e7e08607fccb8af189c
Signed-off-by: Harrison Mutai <harrison.mutai@arm.com>
Fix dangling comments around define guards, addressing leftovers from
fe94a21a6 ("fix(arm): move HW_CONFIG relocation into BL31") which
implicitly removed constraints on using HW_CONFIG with RESET_TO_BL2.
Change-Id: I19d61812fed6fa4b668875e5bf4eafd1a8a660f6
Signed-off-by: Harrison Mutai <harrison.mutai@arm.com>
The failure was caused by missing a variable definition, `status`
in the RSE initialisation patch.
Change-Id: I937a39e20fae39f3a6d14fe66af578c166545301
Signed-off-by: Icen.Zeyada <Icen.Zeyada2@arm.com>
When GICR_WAKER.ProcessorSleep == 1 (i.e. after gicv3_cpuif_disable())
the GIC will assert the WakeRequest signal to try and wake the core up
instead of delivering an interrupt. This is useful when a core is in
some kind of suspend state.
However, when the core is properly off (CPU_OFF), it shouldn't get woken
up in any way other than a CPU_ON call. In the general case interrupts
would be routed away so this doesn't matter. But in case they aren't, we
want the core to stay off.
So turn the redistributor off on CPU_OFF calls. This will prevent the
WakeRequest from being sent.
Change-Id: I7f20591d1c83a4a9639281ef86caa79d6669b536
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
