The Ameba RTC driver converts time_t to
second/minute/hour/day/month/year in rtc_write and back to time_t in
rtc_read. Replace this with an implementation which uses time_t
directly.
Add MBED_APP_START and MBED_APP_SIZE to the linker scripts
so the start and size of an image can be specified. This allows the
ROM to be split into a bootloader region and an application region.
Fix vector table
The address of the vector table is hardcoded to the start of flash.
This patch updates make it properly handle updating the VTOR with
a bootloader.
When initializing the RTC on Kinetis devices, handle the case where
the time overflow interrupt is pending and the case where the time
alarm flag is pending. These flags persist across reset and if not
handled will cause a crash when powering up the low power ticker.
This problem manifested as a lp_ticker test failure on the K22F and
K64F on CI only when running a nightly. This problem has been present
but was made obvious by PR #4094 which configures all tickers to
interrupt at least every MBED_TICKER_INTERRUPT_TIMESTAMP_MAX_DELTA
(~31 minutes). This caused the RTC alarm to fire 31 minutes after the lp_ticker
or lp_timeout test and caused the next run of the lp_ticker test to
crash on boot.
Reworked the serial_format() function for STM32F0x
devices to take the format in the form:
data_bits - parity - stop_bits
E.g. 8 - N - 1
where data_bits exclude the parity bit.
Added a case for 7 bits data as at least the chips
STM32F0x1/STM32F0x2/STM32F0x8 support 7 bits data.
Consolidated serial_format() and uart_init()
functions into a general TARGET_STM serial_api.c
file since the functions are common to all STM targets.
Fixes#4189
For STM32 targets using a 32-bit timer for the microsecond ticker, the
driver did not properly handle timestamps that are in the past. It
would just blindly set the compare register to the requested timestamp,
resulting in the interrupt being serviced up to 4295 seconds late
(i.e. after the 32-bit timer counts all the way around to hit the
timestamp again).
This problem can easily be reproduced by creating a Timeout object
then calling the timeout's attach_us() member function to attach a
callback with a timeout of 0 us. The callback will not get called for
over 2147 seconds, and possibly up to 4295 seconds late if no other
microsecond ticker events are getting scheduled in the meantime.
Now, after the compare register has been set, the timestamp is checked
against the current time to see if the timestamp is in the past, and
if so, the compare event is manually set.
NOTE: By checking if the timestamp is in the past after configuring the
capture register, we ensure proper handling in the case where the timer
updates past the timestamp while setting the capture register.
TimMasterHandle.Instance initialization can be removed from here,
because it will either have been already done previously,
or it will be done in the us_ticker_init() call immediately below.
Following previous fixes on 16 tickers handling, the overflow flag
condition will not happen anymore, so the work-around in place is
not needed anymore
This commit simplifies ticker interrupt set function and handler.
There were issues around the 16 bits timer wrap-around timing as we were
aligning interrupts with wrap-around limits (0xFFFF) and then reading
TIM_MST->CNT again in timer_update_irq_handler which could lead
to crossing case with the wrap-around (TIM_FLAG_UPDATE) case.
Now we're using the 16 lower bits of the timestamp as the reference from
using in set_compare and never changing it. There is also no need to set
comparator again in timer_update_irq_handler. This is more robust and
also more efficient.
Move to a single more reliable implementation of us_ticker_read()
There were historically 2 versions of us_ticker_read() implementation.
The one removed here was not reliable because us_ticker_read() can be
called in interrupt context which means that TIM_MST->CNT would have
wrapped around while SlaveCounter is not yet updated. So there is a need
to check the TIM_FLAG_UPDATE inside this function, which was not done in
the implementation that is removed here.
As reported in issue #4214, there are seen issues seen first on
NUCLEO_F103RB in case of successive Reads of 1 byte at a time.
This issue is due to a wrong state management in the end of read sequence.
Also F1 i2c driver was not fully aligned to others, which is updated here.
There were still side effects, in particular on I2C master slave test,
when setting by default the Pin Speed for F1 family. So for F1 family,
let's do it only in case of Output which is the only case where this
actually applies on this family.
There is an easy default implementation of spi_master_block_write that
just calls spi_master_write in a loop, so the default implementation
of spi_master_block_write has been added to all targets.
Russ Butler
Laurent MEUNIER
Francisco J. Manno
Bradley Scott
arostm
Sam Grove
Leon Lindenfelser
Yuguo Zou