Feedback on the problem of force quitting the smp system when running algorithm

[Alan-19950616]

In the riscv_run_algorithm function, the single_hart parameter of riscv_resume is true. then at old_or_new_riscv_poll since both the halted and running variables are not equal to 0, then a call to riscv_ halt is then called to terminate the newly running algorithm.

static int riscv_run_algorithm(struct target *target, int num_mem_params,
		struct mem_param *mem_params, int num_reg_params,
		struct reg_param *reg_params, target_addr_t entry_point,
		target_addr_t exit_point, int timeout_ms, void *arch_info)
{
...
        /* Run algorithm */
	LOG_TARGET_DEBUG(target, "Resume at 0x%" TARGET_PRIxADDR, entry_point);
	if (riscv_resume(target, 0, entry_point, 0, 1, true) != ERROR_OK)
		return ERROR_FAIL;
...
}

int riscv_openocd_poll(struct target *target)
{
...
	if (should_remain_halted) {
		LOG_TARGET_DEBUG(target, "halt all; should_remain_halted=%d",
			should_remain_halted);
		riscv_halt(target);
	} else if (should_resume) {
		LOG_TARGET_DEBUG(target, "resume all");
		riscv_resume(target, true, 0, 0, 0, false);
	} else if (halted && running) {
		LOG_TARGET_DEBUG(target, "halt all; halted=%d",
			halted);
		riscv_halt(target);
	} else {
		/* For targets that were discovered to be halted, call the
		 * appropriate callback. */
		foreach_smp_target(entry, targets)
		{
			struct target *t = entry->target;
			struct riscv_info *info = riscv_info(t);
			if (info->halted_needs_event_callback) {
				target_call_event_callbacks(t, info->halted_callback_event);
				info->halted_needs_event_callback = false;
			}
		}
	}
...
}

[Alan-19950616]

This issue is nullified.