maintenance mode and shard retirement

[brenzi]

closes #1598
closes #1259
closes #1115

1. make use of L1 maintenance mode flag to filter incoming state transitions. This will allow Integritee Network governance to put a shard into maintenance mode, effectively pausing state transitions
2. retire shard (force-unshield all user funds) if maintenance mode lasts for a "long" time. If maintenance mode isn't lifted within 30d on production systems, we assume the problem can't be solved and therefore recover user funds by unshielding all balances to L1 (to equal accounts). This does leak some sensitive information (account balances) but prevents loss of funds

effects of maintenance mode = true:

• Getters still get executed.
• Sidechain blocks still get produced
• A new TrustedCall filter applies. Only shielding calls get through as we can't prevent them
• After a hardcoded max duration of maintenance mode, the shard gets retired automatically

shard retirement:

• for all accounts except some technical accounts (like enclave signer, fee collection aso.):
  • unshield_all assets and native tokens
• do this for every block as long as we're running because shielding may happen (which should be immediately unshielded)

other changes:

• make stf aware of its own shard & ShardConfig for additional checks
  • StorageVersion can be checked before attempting to decoded L1 mirrored storage
• allow opaque L1 state mirroring to L2 and accessing it from stf (with examples for ShardConfig, ShardStatus, FinalizedSidechainBlock)
• introduce new sgx-runtime pallet "shard-management" which shall serve to track fee payouts and maintenance mode stuff
• introduce ShardMode with sticky Retired variant
• new cli getter get-shard-info to query config, status and mode

followup tasks:

• [CI] test maintenance mode and shard retirement #1695
• shard retirement sanity checks and edge case mitigation #1696

[clangenb]

LGTM

[clangenb]

LGTM

[brenzi]

tested successfully on Incognitee Paseo staging environment