libutee: riscv: Fix relocation type of function call into __utee_panic()

In RISC-V, "j" instruction has R_RISCV_JAL relocation that can represent
an even signed 21-bit offset (+-1MiB). However, this

libutee: riscv: Fix relocation type of function call into __utee_panic()

In RISC-V, "j" instruction has R_RISCV_JAL relocation that can represent
an even signed 21-bit offset (+-1MiB). However, this range is not enough
to be position independent code, and the linker generates linking error.
Fix it by using "tail" instruction which has R_RISCV_CALL_PLT relocation
that the execution can jump to +-2GB location.

Note that we won't return from _utee_panic(), that's why we use "tail"
instead of "call" instruction so that we won't generate redundant return
instruction.

Signed-off-by: Alvin Chang <alvinga@andestech.com>
Acked-by: liushiwei <liushiwei@eswincomputing.com>

show more ...

drivers: pinctrl: fix stubbed pinctrl_free_state()

Fixes pinctrl_free_state() when CFG_DRIVERS_PINCTRL is disabled
as the API function has no return value.

The issue is reported by GCC with an erro

drivers: pinctrl: fix stubbed pinctrl_free_state()

Fixes pinctrl_free_state() when CFG_DRIVERS_PINCTRL is disabled
as the API function has no return value.

The issue is reported by GCC with an error trace like the below:

core/include/drivers/pinctrl.h: In function ‘pinctrl_free_state’:
lib/libutee/include/tee_api_defines.h:117:43: error: ‘return’ with a value, in function returning void [-Werror=return-type]
117 | #define TEE_ERROR_NOT_SUPPORTED 0xFFFF000A
| ^~~~~~~~~~
core/include/drivers/pinctrl.h:158:16: note: in expansion of macro ‘TEE_ERROR_NOT_SUPPORTED’
158 | return TEE_ERROR_NOT_SUPPORTED;
| ^~~~~~~~~~~~~~~~~~~~~~~
In file included from core/include/drivers/stm32_uart.h:10,
from core/arch/arm/plat-stm32mp1/main.c:14:
core/include/drivers/pinctrl.h:156:6: note: declared here
156 | void pinctrl_free_state(struct pinctrl_state *state __unused)
| ^~~~~~~~~~~~~~~~~~

Fixes: 9aec039ec0d7 ("drivers: pinctrl: add pinctrl support")
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

core: riscv: fix interrupt_main_handler() reference

Fixes itr_core_handler() reference in RiscV architecture that was renamed
interrupt_main_handler() in commit referred below.

Fixes: 358bf47c0612

core: riscv: fix interrupt_main_handler() reference

Fixes itr_core_handler() reference in RiscV architecture that was renamed
interrupt_main_handler() in commit referred below.

Fixes: 358bf47c0612 ("core: interrupt: rename itr_core_handler()")
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

core: ffa: initialize my_rxtx.size

Adds missing initialization of my_rxtx.size with CFG_CORE_SEL2_SPMC=y.

Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Balint Dobszay <balin

core: ffa: initialize my_rxtx.size

Adds missing initialization of my_rxtx.size with CFG_CORE_SEL2_SPMC=y.

Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Balint Dobszay <balint.dobszay@arm.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: ffa: rename nw_rxtx to my_rxtx

Renames nw_rxtx to my_rxtx to be more clear when we have an SPMC at
S-EL2 and the rxtx buffer is shared with the SPMC instead of normal
world.

Reviewed-by: Jero

core: ffa: rename nw_rxtx to my_rxtx

Renames nw_rxtx to my_rxtx to be more clear when we have an SPMC at
S-EL2 and the rxtx buffer is shared with the SPMC instead of normal
world.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Balint Dobszay <balint.dobszay@arm.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: ffa: support FFA_VERSION from S-EL0 SPs only

FFA_VERSION using ERET as conduit is not permitted in the FF-A
specification. So remove support for it in thread_spmc_msg_recv() but
keep it in spm

core: ffa: support FFA_VERSION from S-EL0 SPs only

FFA_VERSION using ERET as conduit is not permitted in the FF-A
specification. So remove support for it in thread_spmc_msg_recv() but
keep it in spmc_sp_msg_handler() for S-EL0 SPs where the conduit is SVC.

Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Balint Dobszay <balint.dobszay@arm.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: ffa: support FFA_SPM_ID_GET

Supports the FFA_SPM_ID_GET function introduced with FF-A v1.1.

Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Balint Dobszay <balint.dobsza

core: ffa: support FFA_SPM_ID_GET

Supports the FFA_SPM_ID_GET function introduced with FF-A v1.1.

Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Balint Dobszay <balint.dobszay@arm.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: ffa: rename spmc_get_id() to get_my_id()

Renames spmc_get_id() to get_my_id() in order to avoid confusion with
the function FFA_SPM_ID_GET introduced with FF-A v1.1.

Reviewed-by: Jerome Foris

core: ffa: rename spmc_get_id() to get_my_id()

Renames spmc_get_id() to get_my_id() in order to avoid confusion with
the function FFA_SPM_ID_GET introduced with FF-A v1.1.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Balint Dobszay <balint.dobszay@arm.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

ci: build CFG_CORE_SEL1_SPMC=y CFG_NS_VIRTUALIZATION=y

Add a build with both CFG_CORE_SEL1_SPMC=y and CFG_NS_VIRTUALIZATION=y.

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

ci: build CFG_CORE_SEL1_SPMC=y CFG_NS_VIRTUALIZATION=y

Add a build with both CFG_CORE_SEL1_SPMC=y and CFG_NS_VIRTUALIZATION=y.

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

show more ...

core: spmc: support virtualization with SPMC at S-EL1

Adds support for virtualization with OP-TEE as SPMC at S-EL1. This if
the FF-A counterpart of SMC based ABI with virtualization.

Reviewed-by: B

core: spmc: support virtualization with SPMC at S-EL1

Adds support for virtualization with OP-TEE as SPMC at S-EL1. This if
the FF-A counterpart of SMC based ABI with virtualization.

Reviewed-by: Balint Dobszay <balint.dobszay@arm.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: register spmc_init() with boot_final()

In case of virtualization registers spmc_init() with boot_final()
instead of service_init() to have my_endpoint_id initialized as part of
boot initializa

core: register spmc_init() with boot_final()

In case of virtualization registers spmc_init() with boot_final()
instead of service_init() to have my_endpoint_id initialized as part of
boot initialization instead of delayed initialization when the OP-TEE
partition is created.

This guarantees that my_endpoint_id holds the correct value when the
first FF-A request is received.

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

show more ...

core: add memory type MEM_AREA_NEX_NSEC_SHM

Adds the memory type MEM_AREA_NEX_NSEC_SHM used to map non-secure shared
memory in the nexus.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>

core: add memory type MEM_AREA_NEX_NSEC_SHM

Adds the memory type MEM_AREA_NEX_NSEC_SHM used to map non-secure shared
memory in the nexus.

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

show more ...

core: add virt_get_current_guest_id()

Adds the helper function virt_get_current_guest_id().

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

core: add virt_get_current_guest_id()

Adds the helper function virt_get_current_guest_id().

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

show more ...

core: spmc: support FF-A 1.1

Adds support for FF-A 1.1. Now OP-TEE will need to be able to work with
both version 1.0 and 1.1 depending on the other endpoint. The callee
supplies its implemented ver

core: spmc: support FF-A 1.1

Adds support for FF-A 1.1. Now OP-TEE will need to be able to work with
both version 1.0 and 1.1 depending on the other endpoint. The callee
supplies its implemented version and OP-TEE chooses the highest common
version and returns that. This is done per endpoint so some endpoint may
very well use version 1.0 while another uses version 1.1.

Two data structures, struct ffa_mem_transaction and struct
ffa_partition_info, are affected. Runtime conditionals are used to
select which version to use based on the negotiated FF-A version.

Reviewed-by: Balint Dobszay <balint.dobszay@arm.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: spmc: improve checks in handle_mem_share()

For the two FF-A functions FFA_MEM_SHARE_64 and FFA_MEM_SHARE_32 update
the checks in handle_mem_share() for:
- 32-bit vs 64-bit calling convention,

core: spmc: improve checks in handle_mem_share()

For the two FF-A functions FFA_MEM_SHARE_64 and FFA_MEM_SHARE_32 update
the checks in handle_mem_share() for:
- 32-bit vs 64-bit calling convention, that is, when to mask of the
upper 32 bits.
- that the reported fragment length does not exceed the total length of
the memory transaction descriptor.

Reviewed-by: Balint Dobszay <balint.dobszay@arm.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: arm64: use adr_l for __nex_bss_start and __nex_bss_end

Fixes the following linker errors when CFG_NS_VIRTUALIZATION is enabled:
.../entry_a64.o: in function `clear_bss':
.../entry_a64.

core: arm64: use adr_l for __nex_bss_start and __nex_bss_end

Fixes the following linker errors when CFG_NS_VIRTUALIZATION is enabled:
.../entry_a64.o: in function `clear_bss':
.../entry_a64.S:237:(.text._start+0x8c): relocation truncated to fit: R_AARCH64_ADR_PREL_LO21 against symbol `__nex_bss_start' defined in .bss.mempool_default section in all_objs.o
.../entry_a64.S:238:(.text._start+0x90): relocation truncated to fit: R_AARCH64_ADR_PREL_LO21 against symbol `__nex_bss_end' defined in .bss.mempool_default section in all_objs.o

Use the adr_l macro instead of adr to get the addresses for start and
end of .nex_bss.

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

show more ...

riscv: plat-virt: Override default platform ISA extensions

RV64 virtual platform on QEMU supports C(compressed), Zicsr, and
Zifencei extensions. To specify the ISA extensions into RISC-V
toolchain s

riscv: plat-virt: Override default platform ISA extensions

RV64 virtual platform on QEMU supports C(compressed), Zicsr, and
Zifencei extensions. To specify the ISA extensions into RISC-V
toolchain so that toolchain can generate the code correctly, these
ISA extensions should be encoded into "-march" flag. This patch
overrides the default ISA extensions which is defined in riscv.mk
to specify the extension that the platform really supports.

Signed-off-by: Alvin Chang <alvinga@andestech.com>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: riscv: Add default variables for platform ISA, ABI and code model

In RISC-V, each platform may have different supported ISA extensions,
ABI, and code model. In this commit, we define the defau

core: riscv: Add default variables for platform ISA, ABI and code model

In RISC-V, each platform may have different supported ISA extensions,
ABI, and code model. In this commit, we define the default variables
of ISA extensions, ABI, and code model in RISC-V core Makefile. The
platform can further overrides the values in their plat-*/conf.mk.

Signed-off-by: Alvin Chang <alvinga@andestech.com>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

ta: pkcs11: invalidate handle of destroyed token objects

Marks as invalid any object handle that refers to the token object
that is being destroying so that any use of the related handle will
consid

ta: pkcs11: invalidate handle of destroyed token objects

Marks as invalid any object handle that refers to the token object
that is being destroying so that any use of the related handle will
consider the object handle as invalid.

Link: https://github.com/OP-TEE/optee_os/issues/6005
Reviewed-by: Valerii Chubar <valerii_chubar@epam.com>
Reviewed-by: Vesa Jääskeläinen <vesa.jaaskelainen@vaisala.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

ta: pkcs11: invalidating handle in a database

Introduces a specific pointer value for handles that are no more valid
and API function handle_invalidate() to mark that a handle is no
more valid.

Rev

ta: pkcs11: invalidating handle in a database

Introduces a specific pointer value for handles that are no more valid
and API function handle_invalidate() to mark that a handle is no
more valid.

Reviewed-by: Valerii Chubar <valerii_chubar@epam.com>
Reviewed-by: Vesa Jääskeläinen <vesa.jaaskelainen@vaisala.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

ta: pkcs11: factorize handle reference sanity test

Factorizes sanity validation of handle database and handle value
in pkcs11 TA handle functions

Reviewed-by: Valerii Chubar <valerii_chubar@epam.co

ta: pkcs11: factorize handle reference sanity test

Factorizes sanity validation of handle database and handle value
in pkcs11 TA handle functions

Reviewed-by: Valerii Chubar <valerii_chubar@epam.com>
Reviewed-by: Vesa Jääskeläinen <vesa.jaaskelainen@vaisala.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

core: dt_driver_test: move source file to core/pta/test/

Moves dt_driver_test.c from core/kernel/ to core/pta/test/ where most
embedded tests implementation are located. This is legitimate as the
te

core: dt_driver_test: move source file to core/pta/test/

Moves dt_driver_test.c from core/kernel/ to core/pta/test/ where most
embedded tests implementation are located. This is legitimate as the
test results are retrieved from the Invoke PTA interface, even if the
test is not effectively run from an invocation command of that PTA.

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

show more ...

core: dt_driver_test: fix test failure status

Before this change, failing clock/reset/gpios tests do print a
"not run" status message because related status dt_test_state.probe_xxx
remains to DEFAUL

core: dt_driver_test: fix test failure status

Before this change, failing clock/reset/gpios tests do print a
"not run" status message because related status dt_test_state.probe_xxx
remains to DEFAULT value and are never set to FAILED. Fix this and
move each test to a specific local function to ease error cases
management using 'goto' instructions.

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

show more ...

core: dt_driver_test: clarify "not run" status message

Changes trace message from "not passed" to "not run" when a test
is not run.

Suggested-by: Jerome Forissier <jerome.forissier@linaro.org>
Acke

core: dt_driver_test: clarify "not run" status message

Changes trace message from "not passed" to "not run" when a test
is not run.

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

show more ...

core: interrupt: rename itr_init()

Renames itr_init() to interrupt_main_init() as a later
change will modify interrupt chip API functions using interrupt_
as prefix.

Reviewed-by: Jens Wiklander <je

core: interrupt: rename itr_init()

Renames itr_init() to interrupt_main_init() as a later
change will modify interrupt chip API functions using interrupt_
as prefix.

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

show more ...