plat-stm32mp1: introduce PWR regulators

Introduce a voltage regulator driver for the voltage controllers
driven through PWR interface of stm32mp1 SoCs.

Signed-off-by: Etienne Carriere <etienne.carr

plat-stm32mp1: introduce PWR regulators

Introduce a voltage regulator driver for the voltage controllers
driven through PWR interface of stm32mp1 SoCs.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: ffa: Fix the FF-A version returned to Secure Partition

The current Firmware Framework(FF-A) specification defines the major
version of 0x1 and minor version of 0x0. Return these values when
re

core: ffa: Fix the FF-A version returned to Secure Partition

The current Firmware Framework(FF-A) specification defines the major
version of 0x1 and minor version of 0x0. Return these values when
requested through the FFA_VERSION call instead of returning the
FFA_VERSION function id

Signed-off-by: Sughosh Ganu <sughosh.ganu@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: refactor locking during PTA init

tee_ta_init_pseudo_ta_session() locks tee_ta_mutex only when needed.
This makes this function similar to sec_part_init_session() and
tee_ta_init_user_ta_sessio

core: refactor locking during PTA init

tee_ta_init_pseudo_ta_session() locks tee_ta_mutex only when needed.
This makes this function similar to sec_part_init_session() and
tee_ta_init_user_ta_session() in usage.

Reviewed-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: refactor locking during StMM load

sec_part_init_session() locks tee_ta_mutex only when needed. This avoids
a window where deadlock can occur in case pgt_alloc() goes to sleep.

Reviewed-by: Je

core: refactor locking during StMM load

sec_part_init_session() locks tee_ta_mutex only when needed. This avoids
a window where deadlock can occur in case pgt_alloc() goes to sleep.

Reviewed-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: refactor locking during user TA load

tee_ta_init_user_ta_session() unlocks tee_ta_mutex while loading ldelf
and the user TA. This avoids a window where deadlock can occur in case
pgt_alloc() g

core: refactor locking during user TA load

tee_ta_init_user_ta_session() unlocks tee_ta_mutex while loading ldelf
and the user TA. This avoids a window where deadlock can occur in case
pgt_alloc() goes to sleep.

A new condition variable, tee_ta_init_cv, is added to signal that a TA
has changed state while initializing.

tee_ta_init_session_with_context waits for a TA context be fully
initialized before assigning it to another session.

This also simplifies tee_ta_try_set_busy() since it now doesn't have to
handle the case where it has to wait for a TA context to be fully
initialized.

Reviewed-by: Jerome Forissier <jerome@forissier.org>
Suggested-by: Opensource <opensource@sequiturlabs.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: add CFG_CORE_MAX_SYSCALL_RECURSION

Adds CFG_CORE_MAX_SYSCALL_RECURSION to define the limit for the number
of levels TAs may call each other. If this number is too high we may run
over the thre

core: add CFG_CORE_MAX_SYSCALL_RECURSION

Adds CFG_CORE_MAX_SYSCALL_RECURSION to define the limit for the number
of levels TAs may call each other. If this number is too high we may run
over the thread stack in OP-TEE Core.

Reviewed-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: use vaddr_t instead of uint32_t for object IDs

Some function incorrectly use uint32_t for object identifiers:
tee_obj_get(), tee_svc_cryp_get_state() and tee_svc_storage_get_enum().
Those obje

core: use vaddr_t instead of uint32_t for object IDs

Some function incorrectly use uint32_t for object identifiers:
tee_obj_get(), tee_svc_cryp_get_state() and tee_svc_storage_get_enum().
Those object IDs are actually virtual addresses so they need to be of
type vaddr_t.

Link: https://github.com/OP-TEE/optee_os/issues/4035#issuecomment-680037072
Signed-off-by: Jerome Forissier <jerome@forissier.org>

show more ...

plat-stm32mp1: add inline descriptions in pmic drivers

Add a bit of inline description comments in STM32MP1 PMIC driver.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jero

plat-stm32mp1: add inline descriptions in pmic drivers

Add a bit of inline description comments in STM32MP1 PMIC driver.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

plat-stm32mp1: factorize FDT parsing in pmic driver

Factorize looping into the FDT regulator nodes for STM32MP1 PMIC driver.
Boot-on, low power and non-secure access configuration are done for
each

plat-stm32mp1: factorize FDT parsing in pmic driver

Factorize looping into the FDT regulator nodes for STM32MP1 PMIC driver.
Boot-on, low power and non-secure access configuration are done for
each regulator sub-node of the PMIC node.

This change merges save_power_configurations() sequence into
parse_regulator_fdt_nodes() implementation. The sequence panics on
error as the previous implementation did. Some error level traces
are changed to debug level.

No functional changes.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

plat-stm32mp1: store PMIC regulators secure assignment

Implement stm32mp_nsec_can_access_regu() for non-secure service to
get whether a voltage regulator driven by PMIC can be accessed.
The status i

plat-stm32mp1: store PMIC regulators secure assignment

Implement stm32mp_nsec_can_access_regu() for non-secure service to
get whether a voltage regulator driven by PMIC can be accessed.
The status is set according to embedded FDT, based status value.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jerome Forissier <jerome@forissier.org>

show more ...

core: non-lpae: allocate one more translation table for ASLR

Commit 1579bdf39b3e ("core: lpae: allocate one more translation table
for ASLR") has fixed an issue with LPAE (QEMUv8), but a similar one

core: non-lpae: allocate one more translation table for ASLR

Commit 1579bdf39b3e ("core: lpae: allocate one more translation table
for ASLR") has fixed an issue with LPAE (QEMUv8), but a similar one
occurs with non-LPAE. More specifically, running xtest 1013 on QEMU
with CFG_ULIBS_SHARED=y:

E/TC:0 0 Panic 'Failed to spread pgdir on small tables' at core/arch/arm/mm/core_mmu.c:1737 <core_mmu_map_pages>
E/TC:0 0 TEE load address @ 0x5a9b5000
E/TC:0 0 Call stack:
E/TC:0 0 0x5a9bcba1 print_kernel_stack at optee_os/core/arch/arm/kernel/unwind_arm32.c:109
E/TC:0 0 0x5a9c8293 __do_panic at optee_os/core/kernel/panic.c:31
E/TC:0 0 0x5a9bf357 core_mmu_map_pages at optee_os/core/arch/arm/mm/core_mmu.c:1737
E/TC:0 0 0x5a9c2a2f mobj_reg_shm_inc_map at optee_os/core/arch/arm/mm/mobj_dyn_shm.c:200
E/TC:0 0 0x5a9e9067 mobj_inc_map at optee_os/core/arch/arm/include/mm/mobj.h:93
E/TC:0 0 0x5a9e92b1 mobj_mapped_shm_alloc at optee_os/core/arch/arm/mm/mobj_dyn_shm.c:412
E/TC:0 0 0x5a9ee9d9 msg_param_mobj_from_noncontig at optee_os/core/kernel/msg_param.c:141
E/TC:0 0 0x5a9b9d43 get_rpc_alloc_res at optee_os/core/arch/arm/kernel/thread_optee_smc.c:541
E/TC:0 0 0x5a9e479b thread_rpc_alloc at optee_os/core/arch/arm/kernel/thread_optee_smc.c:580
E/TC:0 0 0x5a9e47bf thread_rpc_alloc_payload at optee_os/core/arch/arm/kernel/thread_optee_smc.c:585
E/TC:0 0 0x5a9b67db rpc_load at optee_os/core/arch/arm/kernel/ree_fs_ta.c:99
E/TC:0 0 0x5a9e28e9 ree_fs_ta_open at optee_os/core/arch/arm/kernel/ree_fs_ta.c:146
E/TC:0 0 0x5a9c9f77 system_open_ta_binary at optee_os/core/pta/system.c:259
E/TC:0 0 0x5a9ca873 invoke_command at optee_os/core/pta/system.c:890
E/TC:0 0 0x5a9e3553 pseudo_ta_enter_invoke_cmd at optee_os/core/arch/arm/kernel/pseudo_ta.c:198
E/TC:0 0 0x5a9ef059 tee_ta_invoke_command at optee_os/core/kernel/tee_ta_manager.c:767
E/TC:0 0 0x5a9f345f syscall_invoke_ta_command at optee_os/core/tee/tee_svc.c:887
E/TC:0 0 0x5a9c3128 tee_svc_do_call at optee_os/core/arch/arm/tee/arch_svc_a32.S:54

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: non-lpae: increase PGT_CACHE_SIZE when CFG_NUM_THREADS=2

xtest on QEMU shows allocation errors when CFG_ULIBS_SHARED=y:

# xtest 1010
...
o regression_1010.10 Invalid memory access 5 with 1

core: non-lpae: increase PGT_CACHE_SIZE when CFG_NUM_THREADS=2

xtest on QEMU shows allocation errors when CFG_ULIBS_SHARED=y:

# xtest 1010
...
o regression_1010.10 Invalid memory access 5 with 1024 bytes memref
regression_1000.c:500: [...] 0xffff000c = TEEC_ERROR_OUT_OF_MEMORY [...]
regression_1000.c:505: [...] 0xffff000c = TEEC_ERROR_OUT_OF_MEMORY [...]
regression_1010.10 FAILED
...
E/TC:? 0 alloc_pgt:147 5 page tables not available
E/TC:? 0 alloc_pgt:147 5 page tables not available

This configuration needs at least 5 page tables. Use 8 to avoid wasting
space.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

drivers: scmi-msg: add SCMI Voltage Domain protocol

SCMI Voltage Domain protocol in defined in the SCMI specification
since its version 3 [1]. This protocol allows a SCMI server to expose
voltage re

drivers: scmi-msg: add SCMI Voltage Domain protocol

SCMI Voltage Domain protocol in defined in the SCMI specification
since its version 3 [1]. This protocol allows a SCMI server to expose
voltage regulator control services. The current specification defines
services to discover the exposed regulators, to enable/disable them
and to set/get the regulator voltage level.

The protocol driver is embedded upon configuration switch
CFG_SCMI_MSG_VOLTAGE_DOMAIN.

Link: [1] https://developer.arm.com/documentation/den0056/c
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

plat-stm32mp1: use SCMI reset to manage MCU hold boot

Adding the MCU hold boot management through a SCMI dedicated
reset domain. MCU hold boot controls the MCU reboot sequence together
with MCU rese

plat-stm32mp1: use SCMI reset to manage MCU hold boot

Adding the MCU hold boot management through a SCMI dedicated
reset domain. MCU hold boot controls the MCU reboot sequence together
with MCU reset controller already exposed to SCMI agent 0.

Signed-off-by: Lionel Debieve <lionel.debieve@st.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: stm32_bsec: add local helper otp_upper_base()

Use local helper function otp_upper_base() to factorize where upper OTP
base ID starts.

Signed-off-by: Etienne Carriere <etienne.carriere@lina

drivers: stm32_bsec: add local helper otp_upper_base()

Use local helper function otp_upper_base() to factorize where upper OTP
base ID starts.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: stm32_gpio: fix incorrect alternate mask

Alternate configuration is using 4 bits, current mask
was resetting partially the other alternate configuration.

Signed-off-by: Lionel Debieve <lio

drivers: stm32_gpio: fix incorrect alternate mask

Alternate configuration is using 4 bits, current mask
was resetting partially the other alternate configuration.

Signed-off-by: Lionel Debieve <lionel.debieve@st.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

plat-stm32mp1: prevent panicking when secure clock has no parent

Remove the Core panic instruction when looking for parents of a
clock to be secure. If the clock has no parent, there is no parent
cl

plat-stm32mp1: prevent panicking when secure clock has no parent

Remove the Core panic instruction when looking for parents of a
clock to be secure. If the clock has no parent, there is no parent
clock to secure, no need to panic.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

plat-stm32mp1: disable RTCAPB clock when not needed

Disable RTCAPB clock when not used. The clock is currently needed
when enabling or disabling the secondary code since this one reads
secure entry

plat-stm32mp1: disable RTCAPB clock when not needed

Disable RTCAPB clock when not used. The clock is currently needed
when enabling or disabling the secondary code since this one reads
secure entry point address from a SoC interface relying on this clock.

Prior this change was RTCAPB clock enabled at boot time and never
disabled. This change disables the clock when secondary core is brought
to life and enables it back before secondary core enters its
power down sequence.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: stpmic1: add stpmic1_regulator_is_valid()

Add driver helper API function stpmic1_regulator_is_valid() to
check if a regulator name identifier references a STPMIC1 regulator.

Signed-off-by:

drivers: stpmic1: add stpmic1_regulator_is_valid()

Add driver helper API function stpmic1_regulator_is_valid() to
check if a regulator name identifier references a STPMIC1 regulator.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: stpmic1: API functions to query regulators levels

Add API function in STPMIC1 driver to allow Core to query the
regulators supported voltage levels. This change is needed by
coming SCMI Vol

drivers: stpmic1: API functions to query regulators levels

Add API function in STPMIC1 driver to allow Core to query the
regulators supported voltage levels. This change is needed by
coming SCMI Voltage Domain for regulator resource discovery
services.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: stpmic1: don't force panic on invalid regulator ID

Returns an explicit invalid reference rather than panicking straight
in local helper function get_regulator_data(). This change lets calle

drivers: stpmic1: don't force panic on invalid regulator ID

Returns an explicit invalid reference rather than panicking straight
in local helper function get_regulator_data(). This change lets caller
function decide whether to panic, propagate an error status or take
the applicable behavior when it occurs.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: scmi-msg: remove #ifdef in switch/case

Remove #ifdef from scmi_process_message() switch/case block and define
stubs for when related SCMI protocol is not embedded.

Signed-off-by: Etienne C

drivers: scmi-msg: remove #ifdef in switch/case

Remove #ifdef from scmi_process_message() switch/case block and define
stubs for when related SCMI protocol is not embedded.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: scmi-msg: fix clock rates query

Fix implementation of the clock rate query platform functions to
consider rates start index which was missing.

Fixes: a7a9e3ba71dd ("drivers/scmi-msg: suppo

drivers: scmi-msg: fix clock rates query

Fix implementation of the clock rate query platform functions to
consider rates start index which was missing.

Fixes: a7a9e3ba71dd ("drivers/scmi-msg: support for clock protocol")
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

core: Fix RPMB fat entry cache buffer overflow

Ensure that fat_entry_dir_update can only update entries less than
the current cache size and not just the maximum size limit of the cache.

Signed-off

core: Fix RPMB fat entry cache buffer overflow

Ensure that fat_entry_dir_update can only update entries less than
the current cache size and not just the maximum size limit of the cache.

Signed-off-by: Neil Shipp <neilsh@microsoft.com>
Reviewed-by: Joakim Bech <joakim.bech@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

Use fallthrough pseudo-keyword in switch/case statements

Use fallthrough; in switch/case statements. Imported libraries
(libtomcrypt, libmbedtls) are not modified to minimize differences with
upstre

Use fallthrough pseudo-keyword in switch/case statements

Use fallthrough; in switch/case statements. Imported libraries
(libtomcrypt, libmbedtls) are not modified to minimize differences with
upstream.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...