Make sure to submit a change of option bytes and wait for the end of
the operation. It is done by checking the OPT_BUSY bit.
Signed-off-by: Dawid Niedzwiecki <dawidn@google.com>
The RDP level is set with the FLASH_OPTSR_PRG register and read with the
FLASH_OPTSR_CUR register, not FLASH_OPTCR.
Fix that to realy change and read RDP.
Signed-off-by: Dawid Niedzwiecki <dawidn@google.com>
Do not unlock CR at the end of initialization.
The lock mechanism is to protect unwanted registers modifications due to
software bugs. There is no point in leaving the registers unlock once
the driver is initialized.
Signed-off-by: Dawid Niedzwiecki <dawidn@google.com>
Adapt the stm32 flash driver for the stm32h7rs serie where some
bit of the FLASH registers are named differently from stm32h7 serie.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Implement the readout protection for the STM32H7 series.
Define the specific functions used by the flash_stm32h7.
Move the stm32h7 flash register manipulation in a write_optb()
to be called by the flash_stm32_set_rdp_level.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Place the flash_stm32_sem_take and flash_stm32_sem_give
function to header file for common to use in any flash driver
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Not all the stm32H7 have a clock enable for their flash;
only the dual-core.
For the st,stm32h7-flash-controller with "clocks" property, the driver
will enable the flash clock bit in the corresponding RCC register.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
When application code runs from M4 (resides on lower 896KiB of
the internal flash of bank2) and it does write to a block in
the upper 128KiB, write fails sometime (for example failed
1 out of 10 block write based on the tests) due to STRBERR.
As per technical reference manual section 4.7.4 of
STM32H745/755 SoC, application may ignore this error and
continue with write. So check for the flag and return okay
status for write(). Based on our tests, the application write
succeeds even when this error happens and match with document
description. So igoring this error flag is okay during write.
Signed-off-by: Murali Karicheri <murali.karicheri@sandc.com>
This variable is only used under DUAL_BANK condition and a warning
is generated in !DUAL_BANK case, so move its definition accordingly.
Signed-off-by: Erwan Gouriou <erwan.gouriou@st.com>
The offset is wrong when the second bank is used by M4 as the
offset is relative to the base which is 0. So add the least
significant 6 nibbles from the CONFIG_FLASH_BASE_ADDRESS to
find the sector.
Signed-off-by: Murali Karicheri <murali.karicheri@sandc.com>
Currently flash controller driver builds and runs only on M7.
This patch supports enablement of the driver on M4 CPUs. The
main issue in using the driver on M4 is that LL_GetFlashSize()
to read the flash size works only on M7 as the internal register
is not accessible from M4. So to use the driver on M4, add a dt
property, bank2-flash-size, to configure flash size of bank2.
this will allow gradual support of flash controller driver
on M4 of all supported STMH7 boards by defining the above
dt property and testing it. Currently this is verified only
on STM32H747i-disco board.
Signed-off-by: Murali Karicheri <murali.karicheri@sandc.com>
Add the support of the stm32h7rs serie to the flash stm32h7 driver
Remap Flash registers to the stm32h7rs serie.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
For the flash driver, the base address is the MCU internal flash
address (usually 0x8000000). This PR gets the that address
from the device tree node "st,stm32-nv-flash"
instead of relying on the CONFIG_FLASH_BASE_ADDRESS
which might differ when building for another flash memory.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Due to source data pointer having no alignment constraint,
extra care needs to be taken when reading source data
as dword
Signed-off-by: Tomislav Milkovic <tomislav.milkovic95@gmail.com>
Many device pointers are initialized at compile and never changed. This
means that the device pointer can be constified (immutable).
Automated using:
```
perl -i -pe 's/const struct device \*(?!const)(.*)= DEVICE/const struct
device *const $1= DEVICE/g' **/*.c
```
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
wait_write_queue() get passed a device and an offset to compute the
sector number and verify it is valid. However both operations are
already done by the caller, so let's pass directly the
flash_stm32_sector_t value instead.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The STM32H7 flash driver read-back a register after writing it to ensure
it is flushed. This is very fragile and might break if a new compiler
version slightly reorder the instructions. Instead use a __DSB()
barrier like done on other STM32 SoC, which ensures that the registers
and the data writes are flushed at the compiler level, but also at the
Cortex-M7 level.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
As the erase time for a sector differs by a few order of magnitude for the
various SoCs of the STM32H7 family, use the just added max-erase-time
property from the device tree instead of the hard-coded 4 seconds value.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The flash_stm32_write_range() function of the STM32H7x flash
driver partially uses a wrong flash program word size for certain
SOC types when calculating the flash write offset. If the used
SOC is not having a flash program word size of 256 bits / 32 bytes
the written data might get corrupted, as the flash write offset
value does not match the number of bytes that were actually
written.
Fixes#45568
Signed-off-by: Christoph Heller <chris@metanetics.de>
In order to bring consistency in-tree, migrate all drivers to the new
prefix <zephyr/...>. Note that the conversion has been scripted, refer
to #45388 for more details.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Add a check to avoid invalidating the cache when the latter is disabled.
Indeed, doing so can lead to a bus fault.
Signed-off-by: Thomas Altenbach <taltenbach@witekio.com>
Refactors all of the on-chip flash drivers to use a shared driver class
initialization priority configuration, CONFIG_FLASH_INIT_PRIORITY, to
allow configuring flash drivers separately from other devices. This is
similar to other driver classes like I2C and SPI.
The default is set to CONFIG_KERNEL_INIT_PRIORITY_DEVICE to preserve the
existing default initialization priority for most drivers.
Driver-specific options for SPI-based flash drivers are left intact
because they need to be initialized at a different priority than on-chip
flash drivers.
Signed-off-by: Maureen Helm <maureen.helm@intel.com>
The STM32H7x flash has an integrated ECC that can correct single
errors and detect double errors. When a double ECC error is detected,
the DBECCERR1/2 flag is raised and there is a bus fault.
We now mask this bus fault and check the error flags. ECC errors are
logged with the offset of the data. Single ECC errors cause a warning
to be logged and double ECC errors return -EIO.
Fixes#33140.
Signed-off-by: Göran Weinholt <goran.weinholt@endian.se>
This commit replaces driver hard coded clk bus and enr definitions with
definitions from device tree.
Signed-off-by: Thomas Stranger <thomas.stranger@outlook.com>
flash_write_protection_set() API was deprecated so driver
implementation interface is left behind as well.
This patch removes all implementation pointed by the interface
'struct flash_driver_api.write_protection'.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
Flash write protection services were integrated into erase and write
procedures. This is step required for fixing following issue:
Multi-threading flash access is not supported by
flash_write_protection_set().
flash_write_protection_set() will be deprecated
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
This fixes the logic in erase_sector() for clearing the previous
sector number. The logic was reversed, clearing bits it should not
clear and keeping previous bits of the sector number. In practice this
does not seem to have had any effect on the current functionality, but
it will start to matter later if e.g. flash interrupts are enabled.
Signed-off-by: Göran Weinholt <goran.weinholt@endian.se>
Convert from device_get_binding to DEVICE_DT_GET. In doing this we
no longer need the label in the devicetree node so we remove that.
Removed all __ASSERT_NO_MSG(clk) since we'll get a build error if
DEVICE_DT_GET cant be satisfied, and the clock control api's will
handle reporting if the device_is_ready.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Some format strings were causing warnings.
As elsewhere in this file, offset is type-casted (or not,
depending on compiler) to long int, and then %ld is used in
format string.
Signed-off-by: Emil Lindqvist <emil@lindq.gr>
When flash is Dualbank and flash size is lower than 512K,
then there is a discontinuity between bank1 and bank2.
Also take into account bank swap capability.
Signed-off-by: Alexandre Bourdiol <alexandre.bourdiol@st.com>
Flash controller support is not yet ready on M4 core.
Remove the chosen declaration to make it clear.
Additionally, generate a build error if this driver is compiled
on M4 core.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Fixes#29831: Implements flash driver for stm32h7 devices.
The driver is independant from the other stm32 families (flash_stm32.c),
only the header interface is (mainly) common.
Signed-off-by: Nicolas VINCENT <nicolas.vincent@vossloh.com>
Fixes#29831: Implements flash driver for stm32h7 devices.
The driver is independant from the other stm32 families (flash_stm32.c),
only the header interface is (mainly) common.
Signed-off-by: Nicolas VINCENT <nicolas.vincent@vossloh.com>
