Murilo Belluzzo
fe48901db0
AP_ADC_ADS1115: Remove resource leak.
...
Constructor allocate '_samples' but there was no destructor to
deallocate it. Also, initializes '_gain' to silence CID 9144 although
'_gain' wasn't being used uninitialized because of ::init().
CID 91424
2016-08-03 00:24:55 -03:00
Murilo Belluzzo
36bdd7f1f1
AP_InertialSensor: MPU6000: Add missing read() check
2016-08-03 00:18:20 -03:00
Lucas De Marchi
0d4caa3ccc
AP_Math: remove declaration of not implemented functions
2016-08-03 00:16:35 -03:00
Murilo Belluzzo
70942472d3
AP_Math: Matrix: Change deallocator to match allocator used
2016-08-03 00:09:11 -03:00
Murilo Belluzzo
e17fdb2aa9
DataFlash: Avoid null pointer dereference
...
Confirms that the loop successfully found a log structure before using it.
Should fix CID 126742
2016-08-03 00:02:19 -03:00
Randy Mackay
595aea236a
Copter: consolidate throw mode state into structure
2016-08-02 20:25:52 +09:00
Randy Mackay
381397c7bd
Copter: set throw mode stage on init
2016-08-02 20:25:52 +09:00
Randy Mackay
807e930251
Copter: throw uses motor spooling instead of interlock
2016-08-02 20:25:52 +09:00
Randy Mackay
f4f13bbe6b
Copter: add throw mode logging
2016-08-02 20:25:52 +09:00
Randy Mackay
8c5c8eec44
Copter: minor throw mode comment fix
...
No functional change
2016-08-02 20:25:52 +09:00
Randy Mackay
8e58ea9277
Copter: add throw mode defines for required speed
...
No functional change
2016-08-02 20:25:52 +09:00
chambana
25940c8e0f
Copter: add THROW_TYPE and allow dropping vehicle to trigger motors
2016-08-02 20:25:52 +09:00
Randy Mackay
f0f87a2f0c
Copter: add throw_nextmode
...
vehicle switches to mode specified in THROW_NEXTMODE parameter after the
throw is completed.
2016-08-02 20:25:52 +09:00
Randy Mackay
64ac18da6c
Copter: move throw state into structure
...
This makes it easier to add more state which is required for the follow throw_nextmode change
2016-08-02 20:25:52 +09:00
Randy Mackay
40db19549e
Copter: allow taking off in Guided mode's attitude control sub mode
2016-08-02 18:26:05 +09:00
Randy Mackay
4e92f08bf1
Copter: only accept attitude targets in Guided mode
2016-08-02 18:26:04 +09:00
Randy Mackay
1161417d7f
Copter: add GUIDED_NOGPS flight mode
...
This mode is a cut down version of Guided mode that only accepts attitude commands.
This mode does not require a GPS lock
2016-08-02 18:25:59 +09:00
Holger Steinhaus
cc4fafc3e4
AP_GPS_PX4: fixed GPS epoch calculations
2016-08-02 11:20:48 +10:00
Gustavo Jose de Sousa
5ccd3dfec8
waf: create ap_version taskgen in the corresponding function
2016-08-01 22:04:16 -03:00
Gustavo Jose de Sousa
2e1ac0de8d
waf: make get_legacy_defines() "public"
...
That function will be used by an upcoming separate tool.
2016-08-01 22:04:16 -03:00
Gustavo Jose de Sousa
f2dec971e7
waf: add "ap_" prefix to keywords "libraries" and "vehicle"
...
As an effort to keep things specific to ardupilot API in the build system
consistent and easy way to identify.
2016-08-01 22:04:16 -03:00
Tom Pittenger
282817eeb9
AP_ADSB: revert default enable = 0
2016-08-01 17:56:04 -07:00
Tom Pittenger
b5ec37b9d8
AP_ADSB: change adsb.enable param to be a flag
2016-08-01 17:51:24 -07:00
Tom Pittenger
f90957e7ed
Plane: remove all threat logic in adsb lib in favor of avoidance lib
2016-08-01 15:27:41 -07:00
Tom Pittenger
2be32e9daa
AP_ADSB: remove all threat logic in adsb lib in favor of avoidance lib
2016-08-01 15:27:40 -07:00
Tom Pittenger
b2b63b4d80
AP_ADSB: change param ADSB_ENABLE default from 0 to 1
2016-08-01 15:27:39 -07:00
Tom Pittenger
89271c29ab
AP_ADSB: move variable into struct like the rest
2016-08-01 15:27:38 -07:00
Tom Pittenger
f7c2df8e91
AP_ADSB: sanity check data via valid_flags before forwarding to avoidance lib
2016-08-01 12:47:18 -07:00
Tom Pittenger
5c0b5f5cd1
AP_ADSB: reject packets that are invalid or already stale
2016-08-01 12:44:20 -07:00
Tom Pittenger
b12edf6a71
AP_ADSB: fix bug where the update timeout wasn't being stored correctly
...
- causes immediate timeout so all samples were dropped
- now the timeout is current and respects the time-since-last-comms/packet so we can ignore stale packets
2016-08-01 12:43:47 -07:00
Tom Pittenger
e0bb77726e
AP_ADSB: do not service transceiver if it is not detected
...
- also allow forced detected (setting chan in header) for testing by not having it time out in that special case (timer = 0)
2016-08-01 09:06:05 -07:00
Pierre Kancir
07eb170b39
Autotest: correct mavproxy unloaded map error
2016-08-01 10:37:15 -03:00
Gustavo Jose de Sousa
bd8526fd68
waf: update to version 1.9.2
...
- That contains our previous cherry-picked commits.
- That fixes the issue of builds not storing data in
`build/<board>/compile_commands.json` files (reported by #4580 ).
2016-08-01 10:33:30 -03:00
Peter Barker
2844190e5b
AP_RangeFinder: correct DisplayName comments
2016-08-01 09:55:25 -03:00
Peter Barker
1bb73383ec
AP_RangeFinder: group _ADDR with other first-rangefinder params
2016-08-01 09:55:24 -03:00
Peter Barker
d5ffc8a863
AP_RangeFinder: define 3_ADDR when we have three rangefinders
...
Without this patch this field would only be defined when 4 rangefinders
were specified
2016-08-01 09:55:24 -03:00
Randy Mackay
d5c18dc987
AC_AttControl: remove unused 100hz definition
2016-08-01 17:21:04 +09:00
Randy Mackay
2dc111ff39
Copter: reduce WPNAV_LOIT_MINA parameter description range
2016-08-01 17:09:20 +09:00
Randy Mackay
7d0cac15fb
AP_ICEngine: add casting to avoid compiler warnings
2016-08-01 09:41:32 +09:00
Lucas De Marchi
e217faacab
AP_Baro: MS56XX: cosmetic changes to members
...
- reorder and document members.
- remove tentative of vertical alignement
- like was done for accumulated values, move the calibration
values to a struct
2016-07-30 00:55:28 -03:00
Lucas De Marchi
f7b453359d
AP_Baro: MS56XX: convert to threaded bus
...
This converts MS56XX to use the thread started by SPI/I2C instead of
using the timer thread. This also fixes a possible starvation of the
main thread:
1) INS driver registers itself to be sampled on timer thread
2) MS56XX registers itself to be sampled on timer thread
3) Main thread waits for a sample from INS with
ins.wait_for_sample()
4) timer thread is waiting on update from MS56XX and consequently
the main thread is waiting on an I2C/SPI transfer
Besides this starvation there's another one due to reuse of the timer
lock in order to pump values from the timer thread to the main thread. A
call to the update() method when we have a sample available would need
to wait on any other driver holding the timer lock.
Now there's a lock just to pass the new values from the bus thread to
the main thread with a very tiny critical region, not waiting on any
bus transfers and/or syscalls.
2016-07-30 00:55:28 -03:00
Lucas De Marchi
052f30bc70
AP_HAL: remove macros to use MS5611 in the main thread
2016-07-30 00:55:28 -03:00
Lucas De Marchi
929348ff75
AP_Baro: MS5611: always use timer thread
...
Remove support to run MS5611 on main thread since we shouldn't be doing
I2C transactions there and we are moving to "thread per bus"
nonetheless.
2016-07-30 00:55:28 -03:00
Lucas De Marchi
7b7e73680f
AP_Math: add USEC_PER_MSEC
2016-07-30 00:55:28 -03:00
Lucas De Marchi
8541d5339d
AP_HAL_Linux: Implement threaded SPI
2016-07-30 00:55:27 -03:00
Lucas De Marchi
ea0595c7ac
AP_HAL_Linux: SPIDevice: move implementation to outside
...
As we add methods, it's becoming too complex to be implemented together
with the declaration.
2016-07-30 00:55:27 -03:00
Lucas De Marchi
590539f674
AP_HAL_Linux: implement threaded I2C
2016-07-30 00:55:27 -03:00
Lucas De Marchi
11ddbb8fb8
AP_HAL_Linux: add PollerThread
...
This is the infrastructure needed by both SPIDevice and I2CDevice to
start a poller thread per bus.
2016-07-30 00:55:27 -03:00
Lucas De Marchi
7d451ac876
AP_HAL_Linux: allow to check if thread is running
2016-07-30 00:55:27 -03:00
Lucas De Marchi
eca6f45b68
AP_HAL_Linux: remove unused part of Poller infra
...
We aren't going to use all the poller infra for now and we need it's
behavior a little bit different for what we are going to use:
- Do not use any "fair" time for each ready fd since we don't want
to set a timeout
- Allow to set the fd on Pollable after constructing it since we are
likely to embed Pollable inside other structs and just later be
able to open an fd.
- Let caller use the epoll flags directly - this is not in AP_HAL,
so there's no need to abstract them
2016-07-30 00:55:27 -03:00