Andrew Tridgell
0b25ff0a16
HAL_PX4: added uartD
2013-11-23 19:29:22 +11:00
Andrew Tridgell
8dc6b758f3
HAL_PX4: switch to delay_microseconds_semaphore() for UART timer
...
this may prevent some timing jitter on the GPS UARTs
2013-11-05 14:41:24 +11:00
Andrew Tridgell
5800c2a2c8
HAL_PX4: prevent excessive writes on startup from blocking
...
this could cause copter on PX4 to hang on startup
2013-10-28 16:12:44 +11:00
Andrew Tridgell
622f0dcc1d
HAL_PX4: prevent threads running before subsystems are initialised
...
this fixes a bug where the timer thread would hang waiting for the
console on startup. This caused the "hit enter" behaviour with recent
PX4 bugs, as AnalogIn tried to read from fd==0, which was a console
read
2013-10-26 18:12:35 +11:00
Andrew Tridgell
1c091cbca5
HAL_PX4: fixed a scheduler bug that affects UART IO
...
we need to detect if we are in a timer task using a PID, not a
flag. This is the same type of bug I fixed last week, but now fixed
properly.
The bug could well be the "USB bug" that Craig has been chasing
2013-10-19 16:50:38 +11:00
Andrew Tridgell
d973730b88
HAL_PX4: use hrt timer and semaphores to improve timing
...
this gives much more accurate microsecond delays, while also ensuring
we yield the CPU when possible
2013-10-14 08:29:30 +11:00
Andrew Tridgell
63bee0b2d8
HAL_PX4: disable checks for in_timerprocess
...
these are not valid in this form for the way the timer thread works on
PX4.
2013-10-11 16:06:35 +11:00
Andrew Tridgell
67db1cedaa
HAL_PX4: fixed delay_microseconds() on PX4
...
we could underflow and delay a very long time!
2013-10-08 17:31:15 +11:00
Andrew Tridgell
942ae70651
HAL_PX4: updates for AP_HAL::MemberProc
2013-09-30 21:06:42 +10:00
Andrew Tridgell
7ec242146d
HAL_PX4: update for new scheduler API
2013-09-28 21:24:03 +10:00
Andrew Tridgell
fb56feb6b2
HAL_PX4: support all analog input pins, with scaling
...
this allows voltage/current/Vcc sensing on FMUv2
2013-09-12 13:26:00 +10:00
Andrew Tridgell
fa107bdb58
PX4: moved UART handling to its own thread
...
this prevents slow writes to the SD card from interfering with MAVLink
or GPS serial communication
2013-09-04 13:15:27 +10:00
Andrew Tridgell
7082e0f8aa
AP_HAL: added hold_in_bootloader parameter to scheduler->reboot()
2013-09-04 11:58:13 +10:00
Andrew Tridgell
e5f19e314d
PX4: systemreset() now takes a boolean argument
...
this chooses whether to stay in the bootloader
2013-08-30 13:01:33 +10:00
Julian Oes
56e94d70e9
PX4: Adaptations to latest build changes
2013-08-30 13:01:32 +10:00
Andrew Tridgell
af7381e9e0
AP_HAL: added register_io_process()
...
this is used to register a low priority IO task. Used for file IO in
DataFlash_File.cpp
2013-04-17 21:33:50 +10:00
Andrew Tridgell
8e2a20bea8
HAL_PX4: enable ttyACM0 as main console, ttyS1 as telemetry port
2013-02-18 13:55:33 +11:00
Andrew Tridgell
02b4ecc273
HAL_PX4: get RC input from PX4IO board
...
this allows us to support DSM and SBUS receivers
2013-01-27 12:52:14 +11:00
Andrew Tridgell
8f2f4b1bc5
HAL_PX4: use write() to /dev/pwm_output for servo output
...
this lowest the cost of PWM output a lot, but relies on the new I2C
based IO firmware
2013-01-25 20:44:36 +11:00
Andrew Tridgell
70227d5d93
HAL_PX4: added scheduler performance counters
2013-01-24 15:04:57 +11:00
Andrew Tridgell
966b9b9b9c
HAL_PX4: format panic messages with a newline
2013-01-23 16:54:44 +11:00
Andrew Tridgell
952df2fced
HAL_PX4: new buffered storeage driver for microsd cards
...
this does all IO in the timer thread, avoids writes that don't change
data, and does all writes in multiples of 128 byte chunks. This should
be about as friendly to SD cards as we can get.
2013-01-23 16:37:40 +11:00
Andrew Tridgell
2a10727902
HAL_PX4: run timer thread as SCHED_FIFO
...
this ensures it never yields to the main thread, which would break our
locking assumptions
2013-01-23 09:21:50 +11:00
Andrew Tridgell
6ffa18fa61
HAL_PX4: rewrite UART driver to be much more efficient
...
this does the IO in the timer thread, and uses buffers in the main
task to avoid system call costs in the flight code.
The cost of PX4 read and write system calls is quite high - about 10
to 15 usec. We can't afford to pay that per byte
2013-01-22 21:22:01 +11:00
Andrew Tridgell
39e28d48c2
HAL_PX4: switched scheduler to use a pthread
...
this allows the timer tasks to access file descriptors in the main APM
task, which makes writing PX4 device drivers much easier
2013-01-21 13:54:09 +11:00
Andrew Tridgell
7c7a215934
HAL_PX4: fixes for new Scheduler API
2013-01-11 12:25:26 +11:00
Andrew Tridgell
5829d44a22
HAL_PX4: disable failsafe handler
...
this won't work until we can set RC output values in timers
2013-01-04 22:26:16 +11:00
Andrew Tridgell
9f423a24ad
HAL_PX4: yield CPU in delay()
...
this allows other apps to run
2013-01-04 12:34:35 +11:00
Andrew Tridgell
1e69b88261
HAL_PX4: use write() in panic()
...
this allows panic from the UARTDriver
2013-01-03 21:11:37 +11:00
Andrew Tridgell
b30fa6535b
HAL_PX4: support reboot
2013-01-03 19:34:36 +11:00
Andrew Tridgell
c6305b5876
HAL_PX4: rework Scheduler using hrt calls
...
thanks to Julian Oes for the suggestion
2013-01-03 13:15:57 +11:00
Andrew Tridgell
5053fb8acc
HAL_PX4: implement atomic operations
2013-01-02 21:45:17 +11:00
Andrew Tridgell
5a70f3becf
HAL_PX4: added a scheduler implementation
2013-01-02 21:39:26 +11:00