* OpenBLT: linker file: use last 16 bytes of ram for shared area
This area is used for communication between FW and OpenBLT.
* OpenBLT: startup: invalidate only DFU breadcumb from startup code
Same location (but different seed) is used for shared param's
signature in OpenBLT.
* OpenBLT: enable shared params
* OpenBLT: enable hooks
If 0x01 is stored in shared param idx=0, then stay in OpenBLT forever.
* Jump to OpenBLT: store proper shared param when OpenBLT is requested
TODO: calculate and store CRC too!
* OpenBLT: move shared_params to start of RAM
* Bootloader: Allocate 16 bytes at start of RAM for bootloader shared area
* OpenBLT: shared_params: reuse same .c and .h file for main application
* OpenBLT: shared_params: enable CRC check
* OpenBLT: no need to manualy set breadcumb before reset
* Revert "OpenBLT: startup: invalidate only DFU breadcumb from startup code"
This reverts commit 6b064714b953560227ae53e38355332cf9f46f7f.
* OpenBLT: guard for builds without OpenBLT
* OpenBLT: reboot to OpenBLT on CAN request with ID=0x667 and DLC=2
* config: options to enable OpenBLT on CAN interfaces
* OpenBLT: comments
* adc mux
* cypress is a special ADC size case
Co-authored-by: Matthew Kennedy <matthewkennedy@outlook.com>
Co-authored-by: rusefillc <sdfsdfqsf2334234234>
* dead code in stm32
* rusefi.cpp can have some noreturn
* handle lua panic
* unused error codes
* simplifications of error_handling.cpp
* comment
* put that back
* guard
* use backup ram, hard fault handler improvement
* wire it up
* upload elf
* s
* I guess I'll just use strncpy
* s
* what is it even complaining about?
* neat, we don't need a cast
* cypress and kinetis
* cleanup
* ci
* guard for tests
* make things c++ friendly
* install various canaries in coal mines
* using is much clearer than typedef
* fixing builds
* lps25
* trigger includes were a mess
* c -> c++ cleanup
* os util
* why did so many include os_util.h?
* fwd declare
* cypress/kinetis
* s
* eg33 happy
* happy sim
* cypress usb
* pch
* cypress
* doesn't need to include all that
* h7
* I2C EEPROM support #3860
let's use contrib on all F4
* I2C EEPROM support #3860
let's use contrib on all F4, step 2
* I2C EEPROM support #3860
let's use contrib on all F4, step 3
* I2C EEPROM support #3860
let's use contrib on all F4, step 4
* I2C EEPROM support #3860
huh
* I2C EEPROM support #3860
all the way
Co-authored-by: rusefillc <sdfsdfqsf2334234234>
* ports: stm32f4: not all stm32f4xx are stm32f407, let board decide
* Do not include mcuconf_community.h in port's mcuconf.h usign relative path
as override file can be provided in board directory.
Some board (like stm32f429-disco) with SDRAM need clock to be
enabled in __early_init (before .bss and .data initialization).
So if STM32_HSECLK defined - skip clock detection.
* Stop & wakeup working on rising edge
* Standby & Sleep mode working Using PA0 only.
* Stop & standby works with wake from PA0.
Added alternative interrupt method to STOP
* Removed hand screwery on registers
* testing stop/standby
* comments
* guard
* s
* we shall figure out f7 later
* ok maybe we do worry about f7
* comments and tweaks
* f7
* f4 maybe probably
This way you can use the console/TunerStudio with the ST-Link
In the process, combine TS_PRIMARY_UART and TS_PRIMARY_SERIAL into TS_PRIMARY_PORT, to make UART vs
SERIAL selection more robus. Ditto for TS_SECONDARY_*. Also change use of TS_NO_PRIMARY to be #if
not #ifdef, so that it can be properly set as a compile flag and not be overwritten by various
header files.
* fast exti
* test code snuck in
* full interrupt disable :(
* do it the old fashioned way
* enable interrupt
* consume stored timestamp
* dead
* h7 maybe
* guard maybe
* non-stm32
* exti 16 wrong on f4/f7
* CORTEX_MAXIMUM_PRIORITY
* safer but uglier
* s
* chibios
* no const
* initializers
Co-authored-by: Matthew Kennedy <makenne@microsoft.com>
* static functions with (void)
* more
* Revert "more"
This reverts commit 246e53441f935451437df186ac92d7df26b62fb6.
* s
Co-authored-by: Matthew Kennedy <makenne@microsoft.com>
* Makefile: define BOOTLOADER=1 for linker in case of OpenBLT too
This will reserve first 32K of flash for bootloader.
* OpenBLT: include into build
* board: subaru eg33: add OpenBLT board code
* board: subaru eg33: use OpenBLT
* Board: Subaru EG33: OpenBLT: reuse HAL and CMSIS from OpenBLT submodule
* Board: Subaru EG33: OpenBLT: reuse linker file from OpenBLT too
* OpenBLT for MRE
* OpenBLT: MRE: adjust LD memory map for smallest variat of MCU on MRE
* OpenBLT: enable for MRE
* OpenBLT: disable CRC check of user application
This allows us to use DFU tool to flash main application too.
* hex2dfu: fresh binary for linux
* OpenBLT: extract common part of OpenBLT makefiles to openblt.mk
* OpenBLT: enable CRC check of user application for MRE and EG33
CRC (actually just a summ) of few first vectors is stored at 0x1c
offset. This is reserved vector for Cortex-M3, M4 and M7
* common_make: append OpenBLT CRC to dfu files
This should allow to use DFU to do main application update when
when OpenBLT is used as bootloader. hex2dfu will store same
styled CRC as OpenBLT expects to see in user app.
* OpenBLT reorganization
Move common stuff to hw_layer/ports/
* OpenBLT: proteus
* Proteus: enable OpenBLT for F4 and F7
Compilation tested only
* dead fast tps
* oooooh map avg on hh7
* adc v4 fast support
* new fast API
* hardware.cpp
* adc v2
* warning
* guard
* no check required
* stub cypress/kinetis
* kinetis and cypress stubs
* big pch energy
* put back ramdisk stub
* tests are happy
* h743 nucleo
* kinetis
* I love deleting code!
* make stepper happy
Co-authored-by: Matthew Kennedy <makenne@microsoft.com>
* STM32F7xx: flash write parallelism settings
Datasheed allows 32bit program operation only for 2.7..3.0
While RM defines wider range...
* Add parentheses to defines
* intFlash: show all errors to user
Currently only erase error was exported to user. Flash write error
was silent. Also define few additional error codes and show it to
user.
* detect hse
* implementation
* these boards don't need to set their own HSECLK
* assertions
* name
* tweaks
* how did this compile?
* s
* biiiig comment
* this script doesn't need to set 25mhz any more
* ....or PLLM
Co-authored-by: Matthew Kennedy <makenne@microsoft.com>
Andrey
Andrey G
rusefillc
rusefillc
Matthew Kennedy
NMSTEC
Scott Smith