core: spmc: Fix flags and tag in retrieve request

Inside check_retrieve_request() the flags and tag where swapped.
Store them in the correct variable.

Fixes: de66193d9849 ("core: FF-A: ADD FFA_MEM_

core: spmc: Fix flags and tag in retrieve request

Inside check_retrieve_request() the flags and tag where swapped.
Store them in the correct variable.

Fixes: de66193d9849 ("core: FF-A: ADD FFA_MEM_RETRIEVE for SPs")
Signed-off-by: Jelle Sels <jelle.sels@arm.com>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

plat-stm32mp1: allow non-secure to read DDR registers

Changes stm32mp1 ETZPC configuration to allow non-secure read access
to DDRCTRL and DDRPHYC interface register. This change will be needed
by ne

plat-stm32mp1: allow non-secure to read DDR registers

Changes stm32mp1 ETZPC configuration to allow non-secure read access
to DDRCTRL and DDRPHYC interface register. This change will be needed
by next U-Boot release v2022.04 since merge of [1] in order to
dynamically compute the DDR size.

Link: [1] https://source.denx.de/u-boot/u-boot/-/commit/d72e7bbe7c2841f161848d57b723495a731d0121
Acked-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

core: imx: add support for i.MX8DXL

Add the i.MX 8DXL SoC support.

Signed-off-by: Silvano di Ninno <silvano.dininno@nxp.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

core: imx: crypto_conf: change CAAM default job ring for mscale platforms

Change the default CAAM job ring index from 0 to 2 for the following
platforms:
* imx8mmevk
* imx8mnevk
* imx8mpevk
* im

core: imx: crypto_conf: change CAAM default job ring for mscale platforms

Change the default CAAM job ring index from 0 to 2 for the following
platforms:
* imx8mmevk
* imx8mnevk
* imx8mpevk
* imx8mqevk
It leaves JR0 available for the HAB to authenticate and decrypt boot
images.

Fixes: 2d7a8964 ("driver: implement CAAM driver")
Signed-off-by: Silvano di Ninno <silvano.dininno@nxp.com>
Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: imx_wdog: fix compilation warning on watchdog driver

$ make PLATFORM=imx-mx7ulpevk CFG_DT=n
core/drivers/imx_wdog.c:42:13: warning: ext_reset_output defined but not used [-Wunused-variable]

drivers: imx_wdog: fix compilation warning on watchdog driver

$ make PLATFORM=imx-mx7ulpevk CFG_DT=n
core/drivers/imx_wdog.c:42:13: warning: ext_reset_output defined but not used [-Wunused-variable]
42 | static bool ext_reset_output;
| ^~~~~~~~~~~~~~~~

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: imx_wdog: fix compilation watchdog driver on mx7ulp

Define WDOG_BASE and WDOG_SIZE value for mx7ulp platform.

$ make PLATFORM=imx-mx7ulpevk CFG_DT=n
core/drivers/imx_wdog.c:151:42: error:

drivers: imx_wdog: fix compilation watchdog driver on mx7ulp

Define WDOG_BASE and WDOG_SIZE value for mx7ulp platform.

$ make PLATFORM=imx-mx7ulpevk CFG_DT=n
core/drivers/imx_wdog.c:151:42: error: WDOG_BASE undeclared here (not in a function); did you mean WDOG_CS?
151 | register_phys_mem_pgdir(MEM_AREA_IO_SEC, WDOG_BASE, CORE_MMU_PGDIR_SIZE);
|

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

core: imx: enable the CAAM on imx8q platforms

Enabled the CAAM on the following platforms:
* imx8qm
* imx8qxp

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Jerome Forissier <jero

core: imx: enable the CAAM on imx8q platforms

Enabled the CAAM on the following platforms:
* imx8qm
* imx8qxp

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: caam: fix aligned buffer allocation for DMA

For aligned memory buffer and DMA CAAM access, the allocated buffer size
must be rounded up to a certain value depending of the DMA behaviour on

drivers: caam: fix aligned buffer allocation for DMA

For aligned memory buffer and DMA CAAM access, the allocated buffer size
must be rounded up to a certain value depending of the DMA behaviour on
the platform.
For the imx8qm/qxp, the allocated aligned buffer size must be rounded up
to 4 bytes.

Signed-off-by: Remi Koman <remi.koman@nxp.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

core: imx: set CAAM configuration for imx8q platforms

Define the JR block size, JR index and the JR interruption number for
the following platforms:
* imx8qm
* imx8qxp

Signed-off-by: Clement Faur

core: imx: set CAAM configuration for imx8q platforms

Define the JR block size, JR index and the JR interruption number for
the following platforms:
* imx8qm
* imx8qxp

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Jerome Forissier <jerome@forissier.org>

show more ...

drivers: clk: move stm32mp1 clock driver to clock drivers directory

Moves stm32mp15 clock driver to core/drivers/clk and adds configuration
switch CFG_STM32MP15_CLK to embed or not the driver. Platf

drivers: clk: move stm32mp1 clock driver to clock drivers directory

Moves stm32mp15 clock driver to core/drivers/clk and adds configuration
switch CFG_STM32MP15_CLK to embed or not the driver. Platform stm32mp1
mandates CFG_STM32MP15_CLK=y.

Reviewed-by: Lionel Debieve <lionel.debieve@foss.st.com>
Acked-by: Gabriel Fernandez <gabriel.fernandez@foss.st.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

plat-stm32mp1: mandate generic clock driver support

Changes stm32mp1 platform to mandate use of the generic clock driver
(CFG_DRVIERS_CLK=y). The platform clock driver implementation is
updated to r

plat-stm32mp1: mandate generic clock driver support

Changes stm32mp1 platform to mandate use of the generic clock driver
(CFG_DRVIERS_CLK=y). The platform clock driver implementation is
updated to remove all implementation related to when CFG_DRVIERS_CLK
is disabled.

CFG_DRIVERS_CLK_DT must be disabled if there is no embedded DTB for
that platform.

Reviewed-by: Lionel Debieve <lionel.debieve@foss.st.com>
Acked-by: Gabriel Fernandez <gabriel.fernandez@foss.st.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

plat-stm32mp1: register to clock framework

Changes stm32mp1 clock driver to register clocks in the clk framework
upon CFG_DRIVERS_CLK=y. CFG_DRIVERS_CLK=y mandates CFG_EMBED_DTB=y
for that platform.

plat-stm32mp1: register to clock framework

Changes stm32mp1 clock driver to register clocks in the clk framework
upon CFG_DRIVERS_CLK=y. CFG_DRIVERS_CLK=y mandates CFG_EMBED_DTB=y
for that platform.

When CFG_DRIVERS_CLK=y, static array stm32mp1_clk[] holds all registered
clock instances, relating to either a clock gate referred in array
stm32mp1_clk_gate[] and an always on clock from array stm32mp1_clk_on[].

Defines local helper functions clock_id_to_gate_index() and
clock_id_to_always_on_index() to convert generic clock references
into a platform local clock identifier that is the index of the target
clock in its relative clock references array.


When CFG_DRIVERS_CLK is disabled, stm32mp1 clock legacy functions
stm32_clock_*() call local clock driver. When CFG_DRIVERS_CLK=y, they
call the generic clock API functions clk_*(). These platform clock legacy
functions are preserved since used in platform specific functions
implementation.

To optimize unpaged memory footprint, only few clock names are
embedded and only upon debug trace level and only required clk_ops
operators are linked in an unpaged memory section.

Acked-by: Lionel Debieve <lionel.debieve@foss.st.com>
Acked-by: Gabriel Fernandez <gabriel.fernandez@foss.st.com>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

core: link: platform can define pager dummy resources

Allows platform to define paged and init dummy resources to allows
platform specific resources to be linked in unpaged memory sections
without p

core: link: platform can define pager dummy resources

Allows platform to define paged and init dummy resources to allows
platform specific resources to be linked in unpaged memory sections
without propagating their unpaged constraint to the resources they
depend on.

Platform should implement source files link_dummies_paged.c and/or
link_dummies_init.c from their platform directory when needed.

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

show more ...

drivers: add reset controller framework

Adds a common reset controller framework rstctrl for interfacing
reset controllers exposed by a platform.

Reset controller consumers can act on relate reset

drivers: add reset controller framework

Adds a common reset controller framework rstctrl for interfacing
reset controllers exposed by a platform.

Reset controller consumers can act on relate reset level with
rstctrl_assert(), rstctrl_deassert() and friends.

Reset controller consumers can claim exclusive access to the reset
level woth rstctrl_get_exclusive(), rstctrl_put_exclusive().

Reset controller provider drivers call rstctrl_register_provider()
to allow other drivers to get a reset control reference from
a devicetree reference. Reset controller driver are identified with
type DT_DRIVER_RSTCTRL.

A reset controller provider exposes struct rstctrl instances made of
an opaque private reference (a private data pointer or an unsigned
integer identifier), an reset controller operators reference and
the exclusive claim state.

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

show more ...

io.h: add WRITE_ONCE macro

Implements WRITE_ONCE() macro that ensures compiler will write memory
only once. It is simple wrapper over __compiler_atomic_store() but
its name emphasizes its purpose.

io.h: add WRITE_ONCE macro

Implements WRITE_ONCE() macro that ensures compiler will write memory
only once. It is simple wrapper over __compiler_atomic_store() but
its name emphasizes its purpose.

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

show more ...

core: remove __rodata_dtdrv_start and __rodata_dtdrv_end

Commit 61bdedea9452 ("core: define DT drivers using scattered arrays")
omitted to remove the declarations and "dummy" definitions for symbols

core: remove __rodata_dtdrv_start and __rodata_dtdrv_end

Commit 61bdedea9452 ("core: define DT drivers using scattered arrays")
omitted to remove the declarations and "dummy" definitions for symbols
__rodata_dtdrv_start and __rodata_dtdrv_end, which are not used anymore.
Remove them.

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

show more ...

core: define DT drivers using scattered arrays

Replace the specific mechanism used to define and enumerate DT drivers
with scattered arrays. Doing so simplifies the TEE linker file a bit.

Signed-of

core: define DT drivers using scattered arrays

Replace the specific mechanism used to define and enumerate DT drivers
with scattered arrays. Doing so simplifies the TEE linker file a bit.

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

show more ...

core: arm: mmu: fix find_map_by_pa() on areas end addresses

Fix find_map_by_pa() to test the inclusive end address of an area
to prevent issues when end address overlaps size field byte size.

Revie

core: arm: mmu: fix find_map_by_pa() on areas end addresses

Fix find_map_by_pa() to test the inclusive end address of an area
to prevent issues when end address overlaps size field byte size.

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

show more ...

core: the FF-A ABI is now a stable ABI

The OP-TEE FF-A driver in the Linux kernel has been merged, so the
changes in the ABI towards the Linux kernel from now on have to be
backwards compatible.

Ac

core: the FF-A ABI is now a stable ABI

The OP-TEE FF-A driver in the Linux kernel has been merged, so the
changes in the ABI towards the Linux kernel from now on have to be
backwards compatible.

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

show more ...

crypto: drivers: se050: ecc shared secret

Allow clients to inject their own keypairs to derive the secret - the
previous implementation only allowed for secure element NVM based
keypairs to be used.

crypto: drivers: se050: ecc shared secret

Allow clients to inject their own keypairs to derive the secret - the
previous implementation only allowed for secure element NVM based
keypairs to be used.

By default, the secure element does not store all the possible EC
curves in its internal memory; however attempting to inject a keypair
when the curve is not in the secure element would cause the injection
to fail.

This commit addresses that situation by generating those curves in the
SE whenever they are not available.

Tested with TEE_ALG_ECDH_P192, TEE_ALG_ECDH_P224, TEE_ALG_ECDH_P256
and TEE_ALG_ECDH_P384 and TEE_ALG_ECDH_P521 (xtest 4009 passing)

Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: fix race in ffa_inc_map()

Fixes a race in ffa_inc_map() when mapcount is 0. The problem goes like:

Thread 1 and 2 calls ffa_inc_map() at the same time and mapcount is 0.
Thread 1 takes the lo

core: fix race in ffa_inc_map()

Fixes a race in ffa_inc_map() when mapcount is 0. The problem goes like:

Thread 1 and 2 calls ffa_inc_map() at the same time and mapcount is 0.
Thread 1 takes the lock first and initializes mapcount to 1 and map the
mobj etc.

When thread 2 has the lock it discovers that mapcount has been
initialize while it was waiting for the lock.

Prior to this patch we where exiting the function doing nothing more
since the mobj was mapped, but by doing so we'll miss to increase
mapcount.

Fix this by restarting the call to refcount_inc() using a loop.

Fixes: 73e1d3f398b0 ("core: add mobj_ffa")
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: fix race in mobj_reg_shm_inc_map()

Fixes a race in mobj_reg_shm_inc_map() when mapcount is 0. The problem
goes like:

Thread 1 and 2 calls mobj_reg_shm_inc_map() at the same time and
mapcount

core: fix race in mobj_reg_shm_inc_map()

Fixes a race in mobj_reg_shm_inc_map() when mapcount is 0. The problem
goes like:

Thread 1 and 2 calls mobj_reg_shm_inc_map() at the same time and
mapcount is 0. Thread 1 takes the lock first and initializes mapcount to
1 and map the mobj etc.

When thread 2 has the lock it discovers that mapcount has been
initialize while it was waiting for the lock.

Prior to this patch we where exiting the function doing nothing more
since the mobj was mapped, but by doing so we'll miss to increase
mapcount.

Fix this by restarting the call to refcount_inc() using a loop.

Fixes: 37a6b717787b ("core: introduce CFG_CORE_DYN_SHM")
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: dt_driver: fix result argument description get_of_device_func

Fixes inline description of type get_of_device_func that falsely mentions
TEE_ERROR_BUSY instead of TEE_ERROR_DEFER_DRIVER_INIT wh

core: dt_driver: fix result argument description get_of_device_func

Fixes inline description of type get_of_device_func that falsely mentions
TEE_ERROR_BUSY instead of TEE_ERROR_DEFER_DRIVER_INIT when expected
resource requests deferral of the driver probing.

Fixes: d8b14b46af9d ("core: dt_driver: get return code when querying a device")
Acked-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

drivers: crypto: rsa: handle not implemented sign/verify operations

Route the unimplemented RSA sign/verify optional cases to their
software implementations.

Signed-off-by: Jorge Ramirez-Ortiz <jor

drivers: crypto: rsa: handle not implemented sign/verify operations

Route the unimplemented RSA sign/verify optional cases to their
software implementations.

Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: make mobj_get_va() more secure

Adds a length parameter to allow mobj_get_va() to check that the entire
va range requested is available.

Reviewed-by: Etienne Carriere <etienne.carriere@linaro.

core: make mobj_get_va() more secure

Adds a length parameter to allow mobj_get_va() to check that the entire
va range requested is available.

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

show more ...