On 04.08.23 12:15, EYSSARTIER Kevin wrote:
I agree that my solution is not good but moving
_RISCV_Start_multitasking out of the ‘#ifdef RTEMS_SMP’ barrier seems to
bring many side effects.
You have to provide this function for all configurations, however, with
different implementations.
--
Classified as: {OPEN}
I don't know how to run the tests.
I observe that _Thread_Do_dispatch fails with
INTERNAL_ERROR_BAD_THREAD_DISPATCH_ENVIRONMENT because
RTEMS_SCORE_ROBUST_THREAD_DISPATCH is activated and the _ISR_Is_enabled returns
false.
Indeed, the _CPU_Start_multitasking should be
On 03.08.23 15:29, Kevin Eyssartier wrote:
The ISR was not activated when running without RTEMS_SMP.
Is there a test which fails due to this bug?
Please use the already existing CPU port interface
_CPU_Start_multitasking() to fix this bug.
--
embedded brains GmbH
Herr Sebastian HUBER
The ISR was not activated when running without RTEMS_SMP.
---
cpukit/score/cpu/riscv/include/rtems/score/cpu.h | 3 +++
cpukit/score/cpu/riscv/riscv-context-switch.S| 5 +
cpukit/score/src/threadstartmultitasking.c | 2 +-
3 files changed, 9 insertions(+), 1 deletion(-)
diff --git
