GitLab

497a59f8a52a22f63cb1af415484b.

Reverting README.md

- fix of cleaning wrong library name
Signed-off-by: mz-fuzzy <mzfuzzy800@gmail.com>

in c++ build command line. Fix for ubunutu compilation.
Signed-off-by: mz-fuzzy <mzfuzzy800@gmail.com>

Signed-off-by: mz-fuzzy <mzfuzzy800@gmail.com>

Signed-off-by: mz-fuzzy <mzfuzzy800@gmail.com>

Mistake: Only the underlying variable has changed, the printout should
still say CONFIG as the radio register is CONFIG.

It seems mutex locking is still required when using SPIDEV. Errors only
become apparent with very heavy radio traffic (HARDWARE FAIL: RADIO NOT
RESPONDING )

Was finally able to successfully recreate conditions to consistently
cause 'HARDWARE FAILURE RADIO NOT RESPONDING' messages on RPI master
node:
1: Using RPi B2.0 with Raspbian 4.1.20+ armv61
2. Using 2 matching nrf24l01 non-PA modules w/ext antenna
3. Using RF24Gateway + RF24Ethernet
4. Using Arduino Due as RF24Ethernet node

Test scenario: Mesh working normally for very long periods of time.
Re-introduced an old node using Arduino Due. Immediately started
noticing error msgs on master node.

Solution: Increasing the wait time from 85 to 95ms seems to prevent
false indications of hardware failures. It appears, the code was just
not waiting long enough.

First attempt to open   /sys/class/gpio/gpio<GPIO NUMBER>/direction
fails when running as non-root user. Added loop w/delay to retry for up
to 10 seconds by default.

Was getting error -  ln: failed to create symbolic link
‘/usr/local/lib/librf24-bcm.so’: Permission denied  when calling make
upload for remote build

#231

Signed-off-by: mz-fuzzy <mzfuzzy800@gmail.com>

Signed-off-by: mz-fuzzy <mzfuzzy800@gmail.com>

interrupt module added to SPIDEV
changed to Ofast optimization
littlewire example fix
Signed-off-by: mz-fuzzy <mzfuzzy800@gmail.com>

Linux library name and symlinks adaptations
Linux library version taken from library.properties (including python lib installation)
ldconfig bugfixing
Linux makefile cleanup
Minor configure/makefile fixes

Mistake: Only the underlying variable has changed, the printout should
still say CONFIG as the radio register is CONFIG.

It seems mutex locking is still required when using SPIDEV. Errors only
become apparent with very heavy radio traffic (HARDWARE FAIL: RADIO NOT
RESPONDING )

makefile: upload actions announce error when remote machine not set
makefile: report destination directories when installing/uploading
synchronizing makefiles for linux examples
RF24_config.h - changed detection of makefile build
minor configure/makefile improvements

Was finally able to successfully recreate conditions to consistently
cause 'HARDWARE FAILURE RADIO NOT RESPONDING' messages on RPI master
node:
1: Using RPi B2.0 with Raspbian 4.1.20+ armv61
2. Using 2 matching nrf24l01 non-PA modules w/ext antenna
3. Using RF24Gateway + RF24Ethernet
4. Using Arduino Due as RF24Ethernet node

Test scenario: Mesh working normally for very long periods of time.
Re-introduced an old node using Arduino Due. Immediately started
noticing error msgs on master node.

Solution: Increasing the wait time from 85 to 95ms seems to prevent
false indications of hardware failures. It appears, the code was just
not waiting long enough.

First attempt to open   /sys/class/gpio/gpio<GPIO NUMBER>/direction
fails when running as non-root user. Added loop w/delay to retry for up
to 10 seconds by default.

bugfixes
RPi driver set as preferred

cpu optimalization improvements

- BBB renamed to SPIDEV
- SPIDEV - support for spidev device specification by RF24 constructor
- LittleWire support unified with other supported linux devices
- configure script introduced for auto detection/configuration
- support for cross compilation including library upload/installation to a remote host
- examples_RPi renamed to examples_linux
- Python wrapper - preparation for cross compiling support

Interrupt functionality is exactly the same as with BCM lib