ardupilot

Commit Graph

Author	SHA1	Message	Date
Peter Barker	723c37bcf7	DataFlash: backend/frontend split	10 years ago
Daniel Frenzel	74460cd526	DataFlash: Delete unused BufferRead function Signed-off-by: Daniel Frenzel <dgdanielf@gmail.com>	11 years ago
Andrew Tridgell	e6bafa2d8f	DataFlash: added structures to Init() of dataflash this will allow the get_log_data() call to add FMT headers if a log has wrapped	11 years ago
Andrew Tridgell	987cdaf3e1	DataFlash: make it safe to do DataFlash logging from timers this uses a non-blocking semaphore take if we are in a timer process. Useful for logging failsafe events	12 years ago
Andrew Tridgell	c52ef80f06	DataFlash: added Block layer in classes this will allow the addition of a DataFlash_File implementation of the DataFlash API which will store logs in a traditional filesystem. That will align better with the PX4 design, and be more useful for fast transfer of logs to a host computer	12 years ago
Andrew Tridgell	28a0ba6c4a	DataFlash: added block based dataflash writes these are about 10x faster than the old writes	12 years ago
Pat Hickey	ef1db1a3fc	DataFlash: don't take a delay_cb	12 years ago
Pat Hickey	a81de007d8	DataFlash: headers switch to use AP_HAL SPI device driver types	12 years ago
Pat Hickey	cc699a2a4e	DataFlash: conform to ordinary stdint types * s/unsigned char/uint8_t/g * changed all type 'byte' to uint8_t	12 years ago
rmackay9	785d9ca6a7	DataFlash_APM2: renamed semaphore to spi3_semaphore to make it more obvious which semaphore is required.	12 years ago
Pat Hickey	605f89fa69	DataFlash_APM2: private AP_Semaphore* rather than use extern AP_Semaphore_spi3 The AP_Semaphore* argument to the constructor can be null (and is by default for compatibility). Semaphore is only used when non-null.	13 years ago
rmackay9	b9963f5b5c	DataFlash: modified to use AP_Semaphore for SPI3 bus when required. Also fixed test sketch which seemed broken.	13 years ago
uncrustify	c68a341633	uncrustify libraries/DataFlash/DataFlash_APM2.h	13 years ago
rmackay9	2b24e422de	DataFlash: revert delay callbacks to use "unsigned long"	13 years ago
rmackay9	20cb2e56a0	DataFlash: change "unsigned long" to "uint32_t" in callback functions	13 years ago
Andrew Tridgell	dd24330645	DataFlash: fallback to BlockErase if ChipErase fails The errata on the APM2 dataflash chip says that ChipErase may not work on some chips	13 years ago
Andrew Tridgell	fc73fd6531	DataFlash: use ChipErase() instead of PageErase() on all pages This makes the DataFlash erase much faster (about 6 seconds instead of about 60 seconds). We need to test and ensure the behaviour is equivalent apart from the speed	13 years ago
Andrew Tridgell	5e4f66e525	DataFlash: moved high level logging logic to library this moves out all the high level logic from ArduPlane/ArduCopter to the library. It also adds a "config page", as the last page in the flash. This is used to check if the flash needs erasing. We only erase now if the DF_LOGGING_FORMAT has changed. This patch also adds a public CardInserted() method, which is used to disable logging on APM2 if a dataflash card is not inserted	13 years ago
Andrew Tridgell	977326e991	dataflash: reworked classes to move most code to parent class this moves all the non-hardware specific code to the parent DataFlash_Class class, keeping the hw specific code much smaller and simpler. This should prevent discrepancies creeping in between APM1 and APM2 support again	13 years ago
Andrew Tridgell	13b2ddcbe1	dataflash: fixed APM2 flash logging we had log wrap disabled on APM2, and also had the wrong log size. The mainline code assumed 4096, but it is 8192 pages long.	13 years ago
Andrew Tridgell	02f25ac36d	libraries: rename purple to APM2	13 years ago
Andrew Tridgell	2b7e0c8494	DataFlash_Purple: added new logging methods needed for ArduPlane builds, but not tested yet	13 years ago
Andrew Tridgell	bb7af8c15a	dataflash: fixed datatypes to be compatible with desktop build	13 years ago
Pat Hickey	82e7551c90	purple: rework DataFlash to separate purple and APM1 hardware support the purple support is still a work in progress, but the APM1 support should be fine	13 years ago

21 Commits (f18802bc462792ad8e320419c4a084dace2f14b5)