status-im
/
eth2.0-specs
mirror of https://github.com/status-im/eth2.0-specs.git
2
0
Fork 0
Code Issues Projects Releases Wiki Activity

cache more than just state root at start of state transition

This commit is contained in:
Danny Ryan 2019-03-08 09:48:46 -07:00
parent 0bcc350b7b
commit f081f23c31
No known key found for this signature in database
GPG Key ID: 2765A792E42CE07A
1 changed files with 19 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
specs/core/0_beacon-chain.md
View File

@ -110,7 +110,7 @@
- [Beacon chain processing](#beacon-chain-processing) - [Beacon chain processing](#beacon-chain-processing)
- [Beacon chain fork choice rule](#beacon-chain-fork-choice-rule) - [Beacon chain fork choice rule](#beacon-chain-fork-choice-rule)
- [Beacon chain state transition function](#beacon-chain-state-transition-function) - [Beacon chain state transition function](#beacon-chain-state-transition-function)
- [State-root caching](#state-root-caching) - [State caching](#state-caching)
- [Per-epoch processing](#per-epoch-processing) - [Per-epoch processing](#per-epoch-processing)
- [Helper functions](#helper-functions-1) - [Helper functions](#helper-functions-1)
- [Justification](#justification) - [Justification](#justification)
@ -1662,28 +1662,38 @@ def lmd_ghost(store: Store, start_state: BeaconState, start_block: BeaconBlock)
We now define the state transition function. At a high level the state transition is made up of four parts: We now define the state transition function. At a high level the state transition is made up of four parts:
1. State-root caching, which happens at the start of every slot. 1. State caching, which happens at the start of every slot.
2. The per-epoch transitions, which happens at the start of the first slot of every epoch. 2. The per-epoch transitions, which happens at the start of the first slot of every epoch.
3. The per-slot transitions, which happens at every slot. 3. The per-slot transitions, which happens at every slot.
4. The per-block transitions, which happens at every block. 4. The per-block transitions, which happens at every block.
Transition section notes: Transition section notes:
* The state-root caching, caches the state root of the previous slot; * The state caching, caches the state root of the previous slot.
* The per-epoch transitions focus on the [validator](#dfn-validator) registry, including adjusting balances and activating and exiting [validators](#dfn-validator), as well as processing crosslinks and managing block justification/finalization; * The per-epoch transitions focus on the [validator](#dfn-validator) registry, including adjusting balances and activating and exiting [validators](#dfn-validator), as well as processing crosslinks and managing block justification/finalization.
* The per-slot transitions focus on the slot counter and block roots records updates; * The per-slot transitions focus on the slot counter and block roots records updates.
* The per-block transitions generally focus on verifying aggregate signatures and saving temporary records relating to the per-block activity in the `BeaconState`. * The per-block transitions generally focus on verifying aggregate signatures and saving temporary records relating to the per-block activity in the `BeaconState`.
Beacon blocks that trigger unhandled Python exceptions (e.g. out-of-range list accesses) and failed `assert`s during the state transition are considered invalid. Beacon blocks that trigger unhandled Python exceptions (e.g. out-of-range list accesses) and failed `assert`s during the state transition are considered invalid.
_Note_: If there are skipped slots between a block and its parent block, run the steps in the [state-root](#state-root-caching), [per-epoch](#per-epoch-processing), and [per-slot](#per-slot-processing) sections once for each skipped slot and then once for the slot containing the new block. _Note_: If there are skipped slots between a block and its parent block, run the steps in the [state-root](#state-caching), [per-epoch](#per-epoch-processing), and [per-slot](#per-slot-processing) sections once for each skipped slot and then once for the slot containing the new block.
### State-root caching ### State caching
At every `slot > GENESIS_SLOT` run the following function: At every `slot > GENESIS_SLOT` run the following function:
```python ```python
def cache_state_root(state: BeaconState) -> None: def cache_state(state: BeaconState) -> None:
state.latest_state_roots[state.slot % SLOTS_PER_HISTORICAL_ROOT] = hash_tree_root(state) previous_slot_state_root = hash_tree_root(state)
# store the previous slot's post state transition root
state.latest_state_roots[state.slot % SLOTS_PER_HISTORICAL_ROOT] = previous_slot_state_root
# cache state root in stored latest_block_header if empty
if state.latest_block_header.state_root == ZERO_HASH:
state.latest_block_header.state_root = previous_slot_state_root
# store latest known block for previous slot
state.latest_block_roots[state.slot % SLOTS_PER_HISTORICAL_ROOT] = hash_tree_root(state.latest_block_header)
``` ```
### Per-epoch processing ### Per-epoch processing
@ -2218,9 +2228,6 @@ At every `slot > GENESIS_SLOT` run the following function:
```python ```python
def advance_slot(state: BeaconState) -> None: def advance_slot(state: BeaconState) -> None:
if state.latest_block_header.state_root == ZERO_HASH:
state.latest_block_header.state_root = state.latest_block_roots[state.slot % SLOTS_PER_HISTORICAL_ROOT]
state.latest_block_roots[state.slot % SLOTS_PER_HISTORICAL_ROOT] = hash_tree_root(state.latest_block_header)
state.slot += 1 state.slot += 1
``` ```