- Remove un-needed files
- Disable printf and uart
- Modify include paths
- Guard macros from mbed_lib with ifndef
(cherry picked from commit 1f30b52488)
(cherry picked from commit 71cd34df32)
(cherry picked from commit 185d2865da)
Volatile makes no real difference when we're using assembler, or locked
functions, but leaving it off could be more efficient for the basic
loads and stores. So add non-volatile overloads in C++ for them.
Get rid of a volatile, and use atomics to synchronise with the interrupt
routine instead.
Useful as a non-RTOS basic compilation check for the atomics - the
fuller atomic test relies on the RTOS.
These are platform tests, but rely on the RTOS to run multiple threads
to exercise it.
(The atomics are still useful in non-RTOS, to protect against interrupt
handlers, but testing versus other threads is easier. The implementation
is the same either way, so doesn't seem worth testing non-RTOS
specifically).
Reimplement atomic code in inline assembly. This can improve
optimisation, and avoids potential architectural problems with using
LDREX/STREX intrinsics.
API further extended:
* Bitwise operations (fetch_and/fetch_or/fetch_xor)
* fetch_add and fetch_sub (like incr/decr, but returning old value -
aligning with C++11)
* compare_exchange_weak
* Explicit memory order specification
* Basic freestanding template overloads for C++
This gives our existing C implementation essentially all the functionality
needed by C++11.
An actual Atomic<T> template based upon these C functions could follow.
One gets this compiler warning from nvstore.cpp:
```
Compile [ 48.6%]: nvstore.cpp
[Warning] nvstore.cpp@814,9: variable 'os_ret' set but not used [-Wunused-but-set-variable]
```
Turns out it's caused by the fact that the variable is only used
with MBED_ASSERTs, which get optimized out or not, depending on your
build profile. In reality we do not need a separate variable for that
in my opinion though, so we can just use the ret-variable instead
and drop the os_ret variable completely and thus avoid this
compiler warning.
SPI peripherals' asynch transaction buffers are now wrapped by
SingletonPtr, which needs to take the singleton_lock Mutex when first
accessed. If it was first accessed by an asynch transaction started from
IRQ, that would not be possible.
Add a SingletonPtr::get() call to the SPI construction process so that
the peripheral's buffer is fully constructed before any SPI methods can
be called, meaning asynch methods won't fail from IRQ.
(Other pre-existing synchronisation issues with async remain, but this
avoids a new trap in Mbed OS 5.12).
Oren Cohen
JarekParal
Oren Cohen
Michael Schwarcz
Andrew Chong
Laurent Meunier
Anna Bridge
Kevin Bracey
Mirela Chirica
Kimmo Vaisanen
Steven Cooreman
Malavika Sajikumar
Janne Kiiskila