Replace the hardcoded value calculation of the memory pool block size
with the RTX preprocessor macro (via a shim layer).
This is in preparation for the replacement of the direct access of RTX
functionalities in Mbed OS with an access via CMSIS.
In rtos/mbed_lib.json:
Enlarge idle-thread-stack-size-debug-extra for all Nuvoton targets.
In rtos/TARGET_CORTEX/mbed_rtx_conf.h:
MBED_DEBUG also must be defined for this configuration parameter to take effect.
As the timer code became more generic, coping with initialization on
demand, and variable width and speed us_ticker_api implementations,
wait_us has gradually gotten slower and slower.
Some platforms have reportedly seen overhead of wait_us() increase from
10µs to 30µs. These changes should fully reverse that drop, and even
make it better than ever.
Add fast paths for platforms that provide compile-time information about
us_ticker. Speed and code size is improved further if:
* Timer has >= 2^32 microsecond range, or better still is 32-bit 1MHz.
* Platform implements us_ticker_read() as a macro
* Timer is initialised at boot, rather than first use
The latter initialisation option is the default for STM, as this has
always been the case.
Building binaries with MicroLib and Arm compiler 6 has been failing
because of a bug in Arm compiler 6. The compiler introduces a
symbol, `__scanf_mbtowc`, which is not present in MicroLib C library.
The present commit adds a weak reference to the symbol to allow a
successful linkage. This is a temporary fix until the bug in the
compiler is fixed. It should be removed after the compiler bug is fixed.
Deprecate wait() in favour of acquire(), try_acquire(),
try_acquire_for() and try_acquire_until().
Brings Semaphore more into line with CMSIS-RTOS 2 (which uses "acquire"),
itself (as it has "release"), and other classes having "try", "try for"
and "try until".
Also steps away from vague "wait" term - the primary operation here is
to acquire the semaphore, and this will of course sleep.
Once a fatal error is in progress, it's not useful to trap RTX errors
or mutex problems, so short-circuit the checks.
This makes it more likely that we may be able to get the console
initialised if it is being written to for the first time by `mbed_error`
in a difficult context - such as an RTX error callback from inside an
SVCall.
For example, the one-line program
osMutexAcquire(NULL, 0);
will generate an RTX error trap, then `mbed_error` will try to call
`write(STDERR_FILENO)` to print the error, which will prompt mbed_retarget to
construct a singleton `UARTSerial`. This would trap in the mutex
for the singleton or the construction of the UARTSerial itself, if
we didn't allow this leniency. If we clear the mutex checks, then
`UARTSerial::write_unbuffered` will work.
CMSIS repo does not support pre-processor defines, hence multiple assembly
files are added for secure/non-secure and floating point tools.
Mbed OS tools support assembly file pre-processing, but the build system
does not support multiple assembly files for each target, hence updating
the assembly files.
1. Patch RTX so irq_cm4f.S files work with no FPU targets
2. Patch RTX so irq_armv8mml.S files to work with and without FPU
2. Patch RTX so irq_armv8mml.S and irq_armv8mbl.S files to work with secure and
non-secure builds
(cherry picked from commit 094662790a)
In case of ARM compiler, idle and timer thread stack though assigned
to `.bss.os` section since not zero initialized are part of `data` section.
In this commit, we are moving stacks of idle and timer thread to bss
section and thereby saving ROM space.
(cherry picked from commit ba92372b8d)
mbed OS used older RTX4 version and with osThreadDef accepting only 3
parameters, to preserve compatibility we hardcode the 'instances'
parameter to 1.
(cherry picked from commit 428acae1b2ac15c3ad523e8d40755a9301220822)
(cherry picked from commit 4360b7bbf8)
For single-threaded apps __iar_Initlocks() is not required and must
not be called because the proper IAR library is brought in only if
"--threaded_lib" is defined in the linker command.
Instead of user defined symbols in assembly files or C files,
use linker scripts to add heap and stack - this is inconsistent
with ARM std linker scripts
Assembler in rtx_core_ca.h is on the whole unified syntax such that it
should work fine if assembled as either ARM or Thumb-2.
The exception was __get_PSP, which uses STM^, which is only available
in ARM state.
Flagging for this function was incorrect, except for IAR - it switched
assembler state without telling the compiler, meaning that the resulting
assembler output could be incorrect, and that the function itself would
not be correctly marked as an ARM function - the CPU would enter in
Thumb state.
Alternative fix would be to switch to System mode, which would work
as either ARM or Thumb-2 assembler, like the rest of the file, but
this is the minimal change.
Fixes#526.
Optimize the tick computation in the following ways:
1. Use relative time rather than absolute time
2. Replace multiplication/division pair with just multiplication
or division:
"* 1000000ULL / OS_TICK_FREQ" -> "* US_IN_TICK"
"* OS_TICK_FREQ / 1000000" -> "/ US_IN_TICK"
The low power ticker wrapper layer adds a large amount of interrupt
latency. This can cause dropped bytes at a baud of 115200 on some
devices. To prevent this by default use the microsecond ticker for
tickless on devices which make use of the low power ticker wrapper.
Martin Kojtal
Laurent Meunier
Hugues Kamba
Chun-Chieh Li
Kevin Bracey
Lingkai Dong
George Beckstein
Oren Cohen
Martin Kojtal
Cruz Monrreal
jeromecoutant
Bartek Szatkowski
deepikabhavnani
RFulchiero
d-kato
Russ Butler