Add BSP_VERBOSE_FATAL_EXTENSION to RTEMS_BSP_CLEANUP_OPTIONS to
optionally print the RTEMS version, the fatal source and the fatal code
in the shared bsp_fatal_extension().
Close#3248.
Remove BSP-specific defaults for RTEMS_BSP_CLEANUP_OPTIONS to simplify
the BSP configuration and documentation. Change defaults to:
BSP_PRESS_KEY_FOR_RESET=0
BSP_RESET_BOARD_AT_EXIT=1
BSP_PRINT_EXCEPTION_CONTEXT=1
Due to a new rtems_panic() implementation, it is possible to replace the
PowerPC-specific BSP_panic() with rtems_panic(). Remove BSP_panic()
implementations.
Close#3245.
Add device spin-lock around internal data structures. Since the driver
provides a low-level C API accessing the descriptors the application
still needs to implement part of the SMP synchonization needed between
Interrupt handler and tasks.
Close#2355.
Use the boot time to initialize the state. Use the state, the current
CPU counter and a very simple pseudo random number generator for
getentropy(). At least, this enables to pass the test "GETENTROPY 1" on
ERC32.
Update #3239.
If necessary, the BSP can now have it's clock and SDRAM initialization
in the SRAM instead of the SDRAM. This allows to change the clock
frequency during the startup of an SDRAM application.
SMP safe by spin-lock protection and semaphore. On spin-lock per SpW/AMBA
port to allow multiple parallel register operations per port. A common
semaphore for general SpW router configuration such as routing table.
Move to a C API instead of using the I/O Manager. The SpW router driver
does not perform any I/O only management of the router.
Update #2355.
* Use #ifdefs for 32/64 bit code
* Use unsigned long which is 32-bit on riscv32 and 64-bit on riscv64 (register size)
* Move the code to a new shared riscv folder to be shared between riscv32 and riscv64
* Rename RTEMS_CPU extracted from command line to shared riscv target s/riscv*/riscv
Update #3109
According to data sheet bit 31 in ASR22 must be written with a zero
to start the up-counter. GRMON starts the up-counter for us. This
patch is important when running in "flight" from ROM without GRMON.
It seems that the DWT CYCCNT does not advance when the CPU waits on a
WFI instruction.
That leads to the effect that for example on the atsamv BSP a sleep(1)
needs something in the range of a few minutes (depending on the
configured systick). A debugger might disables some deep sleep modes so
that the problem only appears if the application is executed without a
debugger.
Updates of the line control and baud divisor while transfers are in
progress may lead to unpredictable behaviour on some chips. Perform the
updates only if necessary.
Close#3198.
Initial idea has been that check for both, TMS570_SCI_FLR_TX_EMPTY
and TMS570_SCI_FLR_TXRDY is required before console driver parameters
update.
closes#2883.
