fix: lotus-miner: remove provecommit1 method

[LexLuthr]

Related Issues

Proposed Changes

This PR drops the support for ProveCommit1.

Additional Info

Checklist

Before you mark the PR ready for review, please make sure that:

• Commits have a clear commit message.
• PR title is in the form of of <PR type>: <area>: <change being made>
  • example: fix: mempool: Introduce a cache for valid signatures
  • PR type: fix, feat, build, chore, ci, docs, perf, refactor, revert, style, test
  • area, e.g. api, chain, state, mempool, multisig, networking, paych, proving, sealing, wallet, deps
• Update CHANGELOG.md or signal that this change does not need it.
  • If the PR affects users (e.g., new feature, bug fix, system requirements change), update the CHANGELOG.md and add details to the UNRELEASED section.
  • If the change does not require a CHANGELOG.md entry, do one of the following:
    • Add [skip changelog] to the PR title
    • Add the label skip/changelog to the PR
• New features have usage guidelines and / or documentation updates in
  • Lotus Documentation
  • Discussion Tutorials
• Tests exist for new functionality or change in behavior
• CI is green

[magik6k]

Pretty sure we need to call processPieces with the last arg set to true to convert dealid sectors to ddo-style -

lotus/storage/pipeline/states_sealing.go

Line 854 in 2ea1009

pams, dealIDs, err := m.processPieces(ctx.Context(), sector, false)

Also pretty sure we need to convert all precommits to ddo-style, otherwise deals in f05 won't get activated

[magik6k]

Don't see anything wrong here, looks good!

[rvagg]

Continuing the discussion above (#12251 (comment)) about how we invoke this removed code path post nv22. (Permalinks below point to v1.28.0)

In

lotus/storage/pipeline/commit_batch.go

Lines 240 to 245 in 28a578f

	if !individual && !cfg.AggregateAboveBaseFee.Equals(big.Zero()) {
	if ts.MinTicketBlock().ParentBaseFee.LessThan(cfg.AggregateAboveBaseFee) {
	individual = true
	}
	}

we have the basefee check, where if the basefee is below the batch balancer level we individual = true.

Since nv22, it always goes into this block, regardless of what individual is, and any sectors that return from processBatchV2() having been processed are removed from the todo list:

lotus/storage/pipeline/commit_batch.go

Lines 246 to 278 in 28a578f

	if nv >= MinDDONetworkVersion {
	// After nv21, we have a new ProveCommitSectors2 method, which supports
	// batching without aggregation, but it doesn't support onboarding
	// sectors which were precommitted with DealIDs in the precommit message.
	// We prefer it for all other sectors, so first we use the new processBatchV2
	var sectors []abi.SectorNumber
	for sn := range b.todo {
	sectors = append(sectors, sn)
	}
	res, err = b.processBatchV2(cfg, sectors, nv, !individual)
	if err != nil {
	err = xerrors.Errorf("processBatchV2: %w", err)
	}
	// Mark sectors as done
	for _, r := range res {
	if err != nil {
	r.Error = err.Error()
	}
	for _, sn := range r.Sectors {
	for _, ch := range b.waiting[sn] {
	ch <- r // buffered
	}
	delete(b.waiting, sn)
	delete(b.todo, sn)
	delete(b.cutoffs, sn)
	}
	}
	}

When we get to processIndividually(), we end up without anything in todo so don't process anything:

lotus/storage/pipeline/commit_batch.go

Lines 287 to 290 in 28a578f

	if individual {
	resV1, err = b.processIndividually(cfg)
	} else {
	var sectors []abi.SectorNumber

But the catch is in processBatchV2(), where we have a check for deal IDs and we don't process those sectors:

lotus/storage/pipeline/commit_batch.go

Lines 359 to 363 in 28a578f

	if b.todo[sector].DealIDPrecommit {
	// can't process sectors precommitted with deal IDs with ProveCommitSectors2
	continue
	}

So the effect of this is that CC sectors can go in via PCS3, but deal sectors go in via PCS.

There's additional changes in here that relate to deal IDs as they pertain to inspecting pieces. In particular this code inspects pieces, looking for non-filler pieces and always opts out of DDO (false). It's only true when it's used for ReplicaUpdate. Either way we get a list of deal IDs back for non-CC sectors, which forces us into the PCS onboarding state as long as the basefee condition holds:

lotus/storage/pipeline/states_sealing.go

Lines 854 to 875 in 28a578f

	pams, dealIDs, err := m.processPieces(ctx.Context(), sector, false)
	if err != nil {
	return err
	}
	res, err := m.commiter.AddCommit(ctx.Context(), sector, AggregateInput{
	Info: proof.AggregateSealVerifyInfo{
	Number: sector.SectorNumber,
	Randomness: sector.TicketValue,
	InteractiveRandomness: sector.SeedValue,
	SealedCID: *sector.CommR,
	UnsealedCID: *sector.CommD,
	},
	Proof: sector.Proof,
	Spt: sector.SectorType,
	ActivationManifest: miner2.SectorActivationManifest{
	SectorNumber: sector.SectorNumber,
	Pieces: pams,
	},
	DealIDPrecommit: len(dealIDs) > 0,
	})

[jennijuju]

Thanks! would've been great if its in PR description in the first place

[TippyFlitsUK]

Started sealing a non-DDO deal on mainnet 1.27.2 to WaitSeed stage. Upgraded to 1.28.1 and the sector fails to finalize as expected.

12251 Test.txt