Core time: Configuration Changes

[eskimor]

Reconsider whether configurations should even differ (likely not) and then either rename configurations to apply for bulk and on-demand or add new ones for bulk. In particular:

pub max_availability_timeouts: u32,
pub ttl: BlockNumber,

If we settle on having them unified (even with legacy), remove [get_provider_config](https://github.com/paritytech/polkadot-sdk/blob/fc358fdf18c8aff62598cd3e3e072b74e32ba792/polkadot/runtime/parachains/src/assigner.rs#L176) from AssignmentProvider` interface and read configuration directly.

Configurations to change/migrate:

• Above options as described
• Rename on_demand_cores configuration to coretime_cores and change meaning to specify how many cores are controlled by coretime. (Will be all eventually.) This configuration will be set by the core time chain via request_core_count.

[bkchr]

• Rename on_demand_cores configuration to coretime_cores and change meaning to specify how many cores are controlled by coretime. (Will be all eventually.)

Why do we need this config at all? With coretime all the cores will be managed by it. Thus, I don't see any reason for adding a config option for this.

[eskimor]

It needs to be properly session buffered. An easy way to achieve this, is to simply have it as part of the configuration (which is already session buffered). This is not set in stone though, but it kinda makes sense. It is a configuration after all, just set by the coretime chain.

[eskimor]

@tdimitrov any updates on this one?

[tdimitrov]

I still haven't started working on this one. I'll do it by tomorrow.