cosmos-sdk/baseapp/abci.go

1363 lines
46 KiB
Go

package baseapp
import (
"context"
"errors"
"fmt"
"sort"
"strings"
"time"
abcitypes "github.com/cometbft/cometbft/abci/types"
abci "github.com/cometbft/cometbft/api/cometbft/abci/v1"
cmtproto "github.com/cometbft/cometbft/api/cometbft/types/v1"
"github.com/cosmos/gogoproto/proto"
"google.golang.org/grpc/codes"
grpcstatus "google.golang.org/grpc/status"
corecomet "cosmossdk.io/core/comet"
coreheader "cosmossdk.io/core/header"
errorsmod "cosmossdk.io/errors"
"cosmossdk.io/store/rootmulti"
snapshottypes "cosmossdk.io/store/snapshots/types"
storetypes "cosmossdk.io/store/types"
"github.com/cosmos/cosmos-sdk/codec"
"github.com/cosmos/cosmos-sdk/telemetry"
sdk "github.com/cosmos/cosmos-sdk/types"
sdkerrors "github.com/cosmos/cosmos-sdk/types/errors"
)
// Supported ABCI Query prefixes and paths
const (
QueryPathApp = "app"
QueryPathCustom = "custom"
QueryPathP2P = "p2p"
QueryPathStore = "store"
QueryPathBroadcastTx = "/cosmos.tx.v1beta1.Service/BroadcastTx"
)
func (app *BaseApp) InitChain(req *abci.InitChainRequest) (*abci.InitChainResponse, error) {
if req.ChainId != app.chainID {
return nil, fmt.Errorf("invalid chain-id on InitChain; expected: %s, got: %s", app.chainID, req.ChainId)
}
// On a new chain, we consider the init chain block height as 0, even though
// req.InitialHeight is 1 by default.
initHeader := cmtproto.Header{ChainID: req.ChainId, Time: req.Time}
app.logger.Info("InitChain", "initialHeight", req.InitialHeight, "chainID", req.ChainId)
// Set the initial height, which will be used to determine if we are proposing
// or processing the first block or not.
app.initialHeight = req.InitialHeight
if app.initialHeight == 0 { // If initial height is 0, set it to 1
app.initialHeight = 1
}
// if req.InitialHeight is > 1, then we set the initial version on all stores
if req.InitialHeight > 1 {
initHeader.Height = req.InitialHeight
if err := app.cms.SetInitialVersion(req.InitialHeight); err != nil {
return nil, err
}
}
// initialize states with a correct header
app.setState(execModeFinalize, initHeader)
app.setState(execModeCheck, initHeader)
// Store the consensus params in the BaseApp's param store. Note, this must be
// done after the finalizeBlockState and context have been set as it's persisted
// to state.
if req.ConsensusParams != nil {
err := app.StoreConsensusParams(app.finalizeBlockState.Context(), *req.ConsensusParams)
if err != nil {
return nil, err
}
}
defer func() {
// InitChain represents the state of the application BEFORE the first block,
// i.e. the genesis block. This means that when processing the app's InitChain
// handler, the block height is zero by default. However, after Commit is called
// the height needs to reflect the true block height.
initHeader.Height = req.InitialHeight
app.checkState.SetContext(app.checkState.Context().WithBlockHeader(initHeader).
WithHeaderInfo(coreheader.Info{
ChainID: req.ChainId,
Height: req.InitialHeight,
Time: req.Time,
}))
app.finalizeBlockState.SetContext(app.finalizeBlockState.Context().WithBlockHeader(initHeader).
WithHeaderInfo(coreheader.Info{
ChainID: req.ChainId,
Height: req.InitialHeight,
Time: req.Time,
}))
}()
if app.initChainer == nil {
return &abci.InitChainResponse{}, nil
}
// add block gas meter for any genesis transactions (allow infinite gas)
app.finalizeBlockState.SetContext(app.finalizeBlockState.Context().WithBlockGasMeter(storetypes.NewInfiniteGasMeter()))
res, err := app.initChainer(app.finalizeBlockState.Context(), req)
if err != nil {
return nil, err
}
if len(req.Validators) > 0 {
if len(req.Validators) != len(res.Validators) {
return nil, fmt.Errorf(
"len(RequestInitChain.Validators) != len(GenesisValidators) (%d != %d)",
len(req.Validators), len(res.Validators),
)
}
sort.Sort(abcitypes.ValidatorUpdates(req.Validators))
for i := range res.Validators {
if !proto.Equal(&res.Validators[i], &req.Validators[i]) {
return nil, fmt.Errorf("genesisValidators[%d] != req.Validators[%d] ", i, i)
}
}
}
// NOTE: We don't commit, but FinalizeBlock for block InitialHeight starts from
// this FinalizeBlockState.
return &abci.InitChainResponse{
ConsensusParams: res.ConsensusParams,
Validators: res.Validators,
AppHash: app.LastCommitID().Hash,
}, nil
}
func (app *BaseApp) Info(_ *abci.InfoRequest) (*abci.InfoResponse, error) {
lastCommitID := app.cms.LastCommitID()
appVersion := InitialAppVersion
if lastCommitID.Version > 0 {
ctx, err := app.CreateQueryContext(lastCommitID.Version, false)
if err != nil {
return nil, fmt.Errorf("failed creating query context: %w", err)
}
appVersion, err = app.AppVersion(ctx)
if err != nil {
return nil, fmt.Errorf("failed getting app version: %w", err)
}
}
return &abci.InfoResponse{
Data: app.name,
Version: app.version,
AppVersion: appVersion,
LastBlockHeight: lastCommitID.Version,
LastBlockAppHash: lastCommitID.Hash,
}, nil
}
// Query implements the ABCI interface. It delegates to CommitMultiStore if it
// implements Queryable.
func (app *BaseApp) Query(_ context.Context, req *abci.QueryRequest) (resp *abci.QueryResponse, err error) {
// add panic recovery for all queries
//
// Ref: https://github.com/cosmos/cosmos-sdk/pull/8039
defer func() {
if r := recover(); r != nil {
resp = queryResult(errorsmod.Wrapf(sdkerrors.ErrPanic, "%v", r), app.trace)
}
}()
// when a client did not provide a query height, manually inject the latest
if req.Height == 0 {
req.Height = app.LastBlockHeight()
}
telemetry.IncrCounter(1, "query", "count")
telemetry.IncrCounter(1, "query", req.Path)
defer telemetry.MeasureSince(telemetry.Now(), req.Path)
if req.Path == QueryPathBroadcastTx {
return queryResult(errorsmod.Wrap(sdkerrors.ErrInvalidRequest, "can't route a broadcast tx message"), app.trace), nil
}
// handle gRPC routes first rather than calling splitPath because '/' characters
// are used as part of gRPC paths
if grpcHandler := app.grpcQueryRouter.Route(req.Path); grpcHandler != nil {
return app.handleQueryGRPC(grpcHandler, req), nil
}
path := SplitABCIQueryPath(req.Path)
if len(path) == 0 {
return queryResult(errorsmod.Wrap(sdkerrors.ErrUnknownRequest, "no query path provided"), app.trace), nil
}
switch path[0] {
case QueryPathApp:
// "/app" prefix for special application queries
resp = handleQueryApp(app, path, req)
case QueryPathStore:
resp = handleQueryStore(app, path, *req)
case QueryPathP2P:
resp = handleQueryP2P(app, path)
default:
resp = queryResult(errorsmod.Wrap(sdkerrors.ErrUnknownRequest, "unknown query path"), app.trace)
}
return resp, nil
}
// ListSnapshots implements the ABCI interface. It delegates to app.snapshotManager if set.
func (app *BaseApp) ListSnapshots(req *abci.ListSnapshotsRequest) (*abci.ListSnapshotsResponse, error) {
resp := &abci.ListSnapshotsResponse{Snapshots: []*abci.Snapshot{}}
if app.snapshotManager == nil {
return resp, nil
}
snapshots, err := app.snapshotManager.List()
if err != nil {
app.logger.Error("failed to list snapshots", "err", err)
return nil, err
}
for _, snapshot := range snapshots {
abciSnapshot, err := snapshot.ToABCI()
if err != nil {
app.logger.Error("failed to convert ABCI snapshots", "err", err)
return nil, err
}
resp.Snapshots = append(resp.Snapshots, &abciSnapshot)
}
return resp, nil
}
// LoadSnapshotChunk implements the ABCI interface. It delegates to app.snapshotManager if set.
func (app *BaseApp) LoadSnapshotChunk(req *abci.LoadSnapshotChunkRequest) (*abci.LoadSnapshotChunkResponse, error) {
if app.snapshotManager == nil {
return &abci.LoadSnapshotChunkResponse{}, nil
}
chunk, err := app.snapshotManager.LoadChunk(req.Height, req.Format, req.Chunk)
if err != nil {
app.logger.Error(
"failed to load snapshot chunk",
"height", req.Height,
"format", req.Format,
"chunk", req.Chunk,
"err", err,
)
return nil, err
}
return &abci.LoadSnapshotChunkResponse{Chunk: chunk}, nil
}
// OfferSnapshot implements the ABCI interface. It delegates to app.snapshotManager if set.
func (app *BaseApp) OfferSnapshot(req *abci.OfferSnapshotRequest) (*abci.OfferSnapshotResponse, error) {
if app.snapshotManager == nil {
app.logger.Error("snapshot manager not configured")
return &abci.OfferSnapshotResponse{Result: abci.OFFER_SNAPSHOT_RESULT_ABORT}, nil
}
if req.Snapshot == nil {
app.logger.Error("received nil snapshot")
return &abci.OfferSnapshotResponse{Result: abci.OFFER_SNAPSHOT_RESULT_REJECT}, nil
}
snapshot, err := snapshottypes.SnapshotFromABCI(req.Snapshot)
if err != nil {
app.logger.Error("failed to decode snapshot metadata", "err", err)
return &abci.OfferSnapshotResponse{Result: abci.OFFER_SNAPSHOT_RESULT_REJECT}, nil
}
err = app.snapshotManager.Restore(snapshot)
switch {
case err == nil:
return &abci.OfferSnapshotResponse{Result: abci.OFFER_SNAPSHOT_RESULT_ACCEPT}, nil
case errors.Is(err, snapshottypes.ErrUnknownFormat):
return &abci.OfferSnapshotResponse{Result: abci.OFFER_SNAPSHOT_RESULT_REJECT_FORMAT}, nil
case errors.Is(err, snapshottypes.ErrInvalidMetadata):
app.logger.Error(
"rejecting invalid snapshot",
"height", req.Snapshot.Height,
"format", req.Snapshot.Format,
"err", err,
)
return &abci.OfferSnapshotResponse{Result: abci.OFFER_SNAPSHOT_RESULT_REJECT}, nil
default:
// CometBFT errors are defined here: https://github.com/cometbft/cometbft/blob/main/statesync/syncer.go
// It may happen that in case of a CometBFT error, such as a timeout (which occurs after two minutes),
// the process is aborted. This is done intentionally because deleting the database programmatically
// can lead to more complicated situations.
app.logger.Error(
"failed to restore snapshot",
"height", req.Snapshot.Height,
"format", req.Snapshot.Format,
"err", err,
)
// We currently don't support resetting the IAVL stores and retrying a
// different snapshot, so we ask CometBFT to abort all snapshot restoration.
return &abci.OfferSnapshotResponse{Result: abci.OFFER_SNAPSHOT_RESULT_ABORT}, nil
}
}
// ApplySnapshotChunk implements the ABCI interface. It delegates to app.snapshotManager if set.
func (app *BaseApp) ApplySnapshotChunk(req *abci.ApplySnapshotChunkRequest) (*abci.ApplySnapshotChunkResponse, error) {
if app.snapshotManager == nil {
app.logger.Error("snapshot manager not configured")
return &abci.ApplySnapshotChunkResponse{Result: abci.APPLY_SNAPSHOT_CHUNK_RESULT_ABORT}, nil
}
_, err := app.snapshotManager.RestoreChunk(req.Chunk)
switch {
case err == nil:
return &abci.ApplySnapshotChunkResponse{Result: abci.APPLY_SNAPSHOT_CHUNK_RESULT_ACCEPT}, nil
case errors.Is(err, snapshottypes.ErrChunkHashMismatch):
app.logger.Error(
"chunk checksum mismatch; rejecting sender and requesting refetch",
"chunk", req.Index,
"sender", req.Sender,
"err", err,
)
return &abci.ApplySnapshotChunkResponse{
Result: abci.APPLY_SNAPSHOT_CHUNK_RESULT_RETRY,
RefetchChunks: []uint32{req.Index},
RejectSenders: []string{req.Sender},
}, nil
default:
app.logger.Error("failed to restore snapshot", "err", err)
return &abci.ApplySnapshotChunkResponse{Result: abci.APPLY_SNAPSHOT_CHUNK_RESULT_ABORT}, nil
}
}
// CheckTx implements the ABCI interface and executes a tx in CheckTx mode. In
// CheckTx mode, messages are not executed. This means messages are only validated
// and only the AnteHandler is executed. State is persisted to the BaseApp's
// internal CheckTx state if the AnteHandler passes. Otherwise, the ResponseCheckTx
// will contain relevant error information. Regardless of tx execution outcome,
// the ResponseCheckTx will contain relevant gas execution context.
func (app *BaseApp) CheckTx(req *abci.CheckTxRequest) (*abci.CheckTxResponse, error) {
var mode execMode
switch {
case req.Type == abci.CHECK_TX_TYPE_CHECK:
mode = execModeCheck
case req.Type == abci.CHECK_TX_TYPE_RECHECK:
mode = execModeReCheck
default:
return nil, fmt.Errorf("unknown RequestCheckTx type: %s", req.Type)
}
gInfo, result, anteEvents, err := app.runTx(mode, req.Tx)
if err != nil {
return responseCheckTxWithEvents(err, gInfo.GasWanted, gInfo.GasUsed, anteEvents, app.trace), nil
}
return &abci.CheckTxResponse{
GasWanted: int64(gInfo.GasWanted), // TODO: Should type accept unsigned ints?
GasUsed: int64(gInfo.GasUsed), // TODO: Should type accept unsigned ints?
Log: result.Log,
Data: result.Data,
Events: sdk.MarkEventsToIndex(result.Events, app.indexEvents),
}, nil
}
// PrepareProposal implements the PrepareProposal ABCI method and returns a
// ResponsePrepareProposal object to the client. The PrepareProposal method is
// responsible for allowing the block proposer to perform application-dependent
// work in a block before proposing it.
//
// Transactions can be modified, removed, or added by the application. Since the
// application maintains its own local mempool, it will ignore the transactions
// provided to it in RequestPrepareProposal. Instead, it will determine which
// transactions to return based on the mempool's semantics and the MaxTxBytes
// provided by the client's request.
//
// Ref: https://github.com/cosmos/cosmos-sdk/blob/main/docs/architecture/adr-060-abci-1.0.md
// Ref: https://github.com/cometbft/cometbft/blob/main/spec/abci/abci%2B%2B_basic_concepts.md
func (app *BaseApp) PrepareProposal(req *abci.PrepareProposalRequest) (resp *abci.PrepareProposalResponse, err error) {
if app.prepareProposal == nil {
return nil, errors.New("PrepareProposal handler not set")
}
// Always reset state given that PrepareProposal can timeout and be called
// again in a subsequent round.
header := cmtproto.Header{
ChainID: app.chainID,
Height: req.Height,
Time: req.Time,
ProposerAddress: req.ProposerAddress,
NextValidatorsHash: req.NextValidatorsHash,
AppHash: app.LastCommitID().Hash,
}
app.setState(execModePrepareProposal, header)
// CometBFT must never call PrepareProposal with a height of 0.
//
// Ref: https://github.com/cometbft/cometbft/blob/059798a4f5b0c9f52aa8655fa619054a0154088c/spec/core/state.md?plain=1#L37-L38
if req.Height < 1 {
return nil, errors.New("PrepareProposal called with invalid height")
}
app.prepareProposalState.SetContext(app.getContextForProposal(app.prepareProposalState.Context(), req.Height).
WithVoteInfos(toVoteInfo(req.LocalLastCommit.Votes)). // this is a set of votes that are not finalized yet, wait for commit
WithBlockHeight(req.Height).
WithProposer(req.ProposerAddress).
WithExecMode(sdk.ExecModePrepareProposal).
WithCometInfo(corecomet.Info{
Evidence: sdk.ToSDKEvidence(req.Misbehavior),
ValidatorsHash: req.NextValidatorsHash,
ProposerAddress: req.ProposerAddress,
LastCommit: sdk.ToSDKExtendedCommitInfo(req.LocalLastCommit),
}).
WithHeaderInfo(coreheader.Info{
ChainID: app.chainID,
Height: req.Height,
Time: req.Time,
}))
app.prepareProposalState.SetContext(app.prepareProposalState.Context().
WithConsensusParams(app.GetConsensusParams(app.prepareProposalState.Context())).
WithBlockGasMeter(app.getBlockGasMeter(app.prepareProposalState.Context())))
defer func() {
if err := recover(); err != nil {
app.logger.Error(
"panic recovered in PrepareProposal",
"height", req.Height,
"time", req.Time,
"panic", err,
)
resp = &abci.PrepareProposalResponse{Txs: req.Txs}
}
}()
resp, err = app.prepareProposal(app.prepareProposalState.Context(), req)
if err != nil {
app.logger.Error("failed to prepare proposal", "height", req.Height, "time", req.Time, "err", err)
return &abci.PrepareProposalResponse{Txs: req.Txs}, nil
}
return resp, nil
}
// ProcessProposal implements the ProcessProposal ABCI method and returns a
// ResponseProcessProposal object to the client. The ProcessProposal method is
// responsible for allowing execution of application-dependent work in a proposed
// block. Note, the application defines the exact implementation details of
// ProcessProposal. In general, the application must at the very least ensure
// that all transactions are valid. If all transactions are valid, then we inform
// CometBFT that the Status is ACCEPT. However, the application is also able
// to implement optimizations such as executing the entire proposed block
// immediately.
//
// If a panic is detected during execution of an application's ProcessProposal
// handler, it will be recovered and we will reject the proposal.
//
// Ref: https://github.com/cosmos/cosmos-sdk/blob/main/docs/architecture/adr-060-abci-1.0.md
// Ref: https://github.com/cometbft/cometbft/blob/main/spec/abci/abci%2B%2B_basic_concepts.md
func (app *BaseApp) ProcessProposal(req *abci.ProcessProposalRequest) (resp *abci.ProcessProposalResponse, err error) {
if app.processProposal == nil {
return nil, errors.New("ProcessProposal handler not set")
}
// CometBFT must never call ProcessProposal with a height of 0.
// Ref: https://github.com/cometbft/cometbft/blob/059798a4f5b0c9f52aa8655fa619054a0154088c/spec/core/state.md?plain=1#L37-L38
if req.Height < 1 {
return nil, errors.New("ProcessProposal called with invalid height")
}
// Always reset state given that ProcessProposal can timeout and be called
// again in a subsequent round.
header := cmtproto.Header{
ChainID: app.chainID,
Height: req.Height,
Time: req.Time,
ProposerAddress: req.ProposerAddress,
NextValidatorsHash: req.NextValidatorsHash,
AppHash: app.LastCommitID().Hash,
}
app.setState(execModeProcessProposal, header)
// Since the application can get access to FinalizeBlock state and write to it,
// we must be sure to reset it in case ProcessProposal timeouts and is called
// again in a subsequent round. However, we only want to do this after we've
// processed the first block, as we want to avoid overwriting the finalizeState
// after state changes during InitChain.
if req.Height > app.initialHeight {
// abort any running OE
app.optimisticExec.Abort()
app.setState(execModeFinalize, header)
}
app.processProposalState.SetContext(app.getContextForProposal(app.processProposalState.Context(), req.Height).
WithVoteInfos(req.ProposedLastCommit.Votes). // this is a set of votes that are not finalized yet, wait for commit
WithBlockHeight(req.Height).
WithHeaderHash(req.Hash).
WithProposer(req.ProposerAddress).
WithCometInfo(corecomet.Info{
ProposerAddress: req.ProposerAddress,
ValidatorsHash: req.NextValidatorsHash,
Evidence: sdk.ToSDKEvidence(req.Misbehavior),
LastCommit: sdk.ToSDKCommitInfo(req.ProposedLastCommit),
},
).
WithExecMode(sdk.ExecModeProcessProposal).
WithHeaderInfo(coreheader.Info{
ChainID: app.chainID,
Height: req.Height,
Time: req.Time,
}))
app.processProposalState.SetContext(app.processProposalState.Context().
WithConsensusParams(app.GetConsensusParams(app.processProposalState.Context())).
WithBlockGasMeter(app.getBlockGasMeter(app.processProposalState.Context())))
defer func() {
if err := recover(); err != nil {
app.logger.Error(
"panic recovered in ProcessProposal",
"height", req.Height,
"time", req.Time,
"hash", fmt.Sprintf("%X", req.Hash),
"panic", err,
)
resp = &abci.ProcessProposalResponse{Status: abci.PROCESS_PROPOSAL_STATUS_REJECT}
}
}()
resp, err = app.processProposal(app.processProposalState.Context(), req)
if err != nil {
app.logger.Error("failed to process proposal", "height", req.Height, "time", req.Time, "hash", fmt.Sprintf("%X", req.Hash), "err", err)
return &abci.ProcessProposalResponse{Status: abci.PROCESS_PROPOSAL_STATUS_REJECT}, nil
}
// Only execute optimistic execution if the proposal is accepted, OE is
// enabled and the block height is greater than the initial height. During
// the first block we'll be carrying state from InitChain, so it would be
// impossible for us to easily revert.
// After the first block has been processed, the next blocks will get executed
// optimistically, so that when the ABCI client calls `FinalizeBlock` the app
// can have a response ready.
if resp.Status == abci.PROCESS_PROPOSAL_STATUS_ACCEPT &&
app.optimisticExec.Enabled() &&
req.Height > app.initialHeight {
app.optimisticExec.Execute(req)
}
return resp, nil
}
// ExtendVote implements the ExtendVote ABCI method and returns a ResponseExtendVote.
// It calls the application's ExtendVote handler which is responsible for performing
// application-specific business logic when sending a pre-commit for the NEXT
// block height. The extensions response may be non-deterministic but must always
// be returned, even if empty.
//
// Agreed upon vote extensions are made available to the proposer of the next
// height and are committed in the subsequent height, i.e. H+2. An error is
// returned if vote extensions are not enabled or if extendVote fails or panics.
func (app *BaseApp) ExtendVote(_ context.Context, req *abci.ExtendVoteRequest) (resp *abci.ExtendVoteResponse, err error) {
// Always reset state given that ExtendVote and VerifyVoteExtension can timeout
// and be called again in a subsequent round.
var ctx sdk.Context
// If we're extending the vote for the initial height, we need to use the
// finalizeBlockState context, otherwise we don't get the uncommitted data
// from InitChain.
if req.Height == app.initialHeight {
ctx, _ = app.finalizeBlockState.Context().CacheContext()
} else {
ms := app.cms.CacheMultiStore()
ctx = sdk.NewContext(ms, false, app.logger).WithStreamingManager(app.streamingManager).WithChainID(app.chainID).WithBlockHeight(req.Height)
}
if app.extendVote == nil {
return nil, errors.New("application ExtendVote handler not set")
}
// If vote extensions are not enabled, as a safety precaution, we return an
// error.
cp := app.GetConsensusParams(ctx)
// Note: In this case, we do want to extend vote if the height is equal or
// greater than VoteExtensionsEnableHeight. This defers from the check done
// in ValidateVoteExtensions and PrepareProposal in which we'll check for
// vote extensions on VoteExtensionsEnableHeight+1.
extsEnabled := cp.Feature != nil && req.Height >= cp.Feature.VoteExtensionsEnableHeight.Value && cp.Feature.VoteExtensionsEnableHeight.Value != 0
if !extsEnabled {
// check abci params
extsEnabled = cp.Abci != nil && req.Height >= cp.Abci.VoteExtensionsEnableHeight && cp.Abci.VoteExtensionsEnableHeight != 0
if !extsEnabled {
return nil, fmt.Errorf("vote extensions are not enabled; unexpected call to ExtendVote at height %d", req.Height)
}
}
ctx = ctx.
WithConsensusParams(cp).
WithBlockGasMeter(storetypes.NewInfiniteGasMeter()).
WithBlockHeight(req.Height).
WithHeaderHash(req.Hash).
WithExecMode(sdk.ExecModeVoteExtension).
WithHeaderInfo(coreheader.Info{
ChainID: app.chainID,
Height: req.Height,
Hash: req.Hash,
})
// add a deferred recover handler in case extendVote panics
defer func() {
if r := recover(); r != nil {
app.logger.Error(
"panic recovered in ExtendVote",
"height", req.Height,
"hash", fmt.Sprintf("%X", req.Hash),
"panic", err,
)
err = fmt.Errorf("recovered application panic in ExtendVote: %v", r)
}
}()
resp, err = app.extendVote(ctx, req)
if err != nil {
app.logger.Error("failed to extend vote", "height", req.Height, "hash", fmt.Sprintf("%X", req.Hash), "err", err)
return &abci.ExtendVoteResponse{VoteExtension: []byte{}}, nil
}
return resp, err
}
// VerifyVoteExtension implements the VerifyVoteExtension ABCI method and returns
// a ResponseVerifyVoteExtension. It calls the applications' VerifyVoteExtension
// handler which is responsible for performing application-specific business
// logic in verifying a vote extension from another validator during the pre-commit
// phase. The response MUST be deterministic. An error is returned if vote
// extensions are not enabled or if verifyVoteExt fails or panics.
// We highly recommend a size validation due to performance degradation,
// see more here https://docs.cometbft.com/v1.0/references/qa/cometbft-qa-38#vote-extensions-testbed
func (app *BaseApp) VerifyVoteExtension(req *abci.VerifyVoteExtensionRequest) (resp *abci.VerifyVoteExtensionResponse, err error) {
if app.verifyVoteExt == nil {
return nil, errors.New("application VerifyVoteExtension handler not set")
}
var ctx sdk.Context
// If we're verifying the vote for the initial height, we need to use the
// finalizeBlockState context, otherwise we don't get the uncommitted data
// from InitChain.
if req.Height == app.initialHeight {
ctx, _ = app.finalizeBlockState.Context().CacheContext()
} else {
ms := app.cms.CacheMultiStore()
ctx = sdk.NewContext(ms, false, app.logger).WithStreamingManager(app.streamingManager).WithChainID(app.chainID).WithBlockHeight(req.Height)
}
// If vote extensions are not enabled, as a safety precaution, we return an
// error.
cp := app.GetConsensusParams(ctx)
// Note: we verify votes extensions on VoteExtensionsEnableHeight+1. Check
// comment in ExtendVote and ValidateVoteExtensions for more details.
extsEnabled := cp.Feature.VoteExtensionsEnableHeight != nil && req.Height >= cp.Feature.VoteExtensionsEnableHeight.Value && cp.Feature.VoteExtensionsEnableHeight.Value != 0
if !extsEnabled {
// check abci params
extsEnabled = cp.Abci != nil && req.Height >= cp.Abci.VoteExtensionsEnableHeight && cp.Abci.VoteExtensionsEnableHeight != 0
if !extsEnabled {
return nil, fmt.Errorf("vote extensions are not enabled; unexpected call to VerifyVoteExtension at height %d", req.Height)
}
}
// add a deferred recover handler in case verifyVoteExt panics
defer func() {
if r := recover(); r != nil {
app.logger.Error(
"panic recovered in VerifyVoteExtension",
"height", req.Height,
"hash", fmt.Sprintf("%X", req.Hash),
"validator", fmt.Sprintf("%X", req.ValidatorAddress),
"panic", r,
)
err = fmt.Errorf("recovered application panic in VerifyVoteExtension: %v", r)
}
}()
ctx = ctx.
WithConsensusParams(cp).
WithBlockGasMeter(storetypes.NewInfiniteGasMeter()).
WithBlockHeight(req.Height).
WithHeaderHash(req.Hash).
WithExecMode(sdk.ExecModeVerifyVoteExtension).
WithHeaderInfo(coreheader.Info{
ChainID: app.chainID,
Height: req.Height,
Hash: req.Hash,
})
resp, err = app.verifyVoteExt(ctx, req)
if err != nil {
app.logger.Error("failed to verify vote extension", "height", req.Height, "err", err)
return &abci.VerifyVoteExtensionResponse{Status: abci.VERIFY_VOTE_EXTENSION_STATUS_REJECT}, nil
}
return resp, err
}
// internalFinalizeBlock executes the block, called by the Optimistic
// Execution flow or by the FinalizeBlock ABCI method. The context received is
// only used to handle early cancellation, for anything related to state app.finalizeBlockState.Context()
// must be used.
func (app *BaseApp) internalFinalizeBlock(ctx context.Context, req *abci.FinalizeBlockRequest) (*abci.FinalizeBlockResponse, error) {
var events []abci.Event
if err := app.checkHalt(req.Height, req.Time); err != nil {
return nil, err
}
if err := app.validateFinalizeBlockHeight(req); err != nil {
return nil, err
}
if app.cms.TracingEnabled() {
app.cms.SetTracingContext(storetypes.TraceContext(
map[string]any{"blockHeight": req.Height},
))
}
header := cmtproto.Header{
ChainID: app.chainID,
Height: req.Height,
Time: req.Time,
ProposerAddress: req.ProposerAddress,
NextValidatorsHash: req.NextValidatorsHash,
AppHash: app.LastCommitID().Hash,
}
// finalizeBlockState should be set on InitChain or ProcessProposal. If it is
// nil, it means we are replaying this block and we need to set the state here
// given that during block replay ProcessProposal is not executed by CometBFT.
if app.finalizeBlockState == nil {
app.setState(execModeFinalize, header)
}
// Context is now updated with Header information.
app.finalizeBlockState.SetContext(app.finalizeBlockState.Context().
WithBlockHeader(header).
WithHeaderHash(req.Hash).
WithHeaderInfo(coreheader.Info{
ChainID: app.chainID,
Height: req.Height,
Time: req.Time,
Hash: req.Hash,
AppHash: app.LastCommitID().Hash,
}).
WithConsensusParams(app.GetConsensusParams(app.finalizeBlockState.Context())).
WithVoteInfos(req.DecidedLastCommit.Votes).
WithExecMode(sdk.ExecModeFinalize).
WithCometInfo(corecomet.Info{
Evidence: sdk.ToSDKEvidence(req.Misbehavior),
ValidatorsHash: req.NextValidatorsHash,
ProposerAddress: req.ProposerAddress,
LastCommit: sdk.ToSDKCommitInfo(req.DecidedLastCommit),
}))
// GasMeter must be set after we get a context with updated consensus params.
gasMeter := app.getBlockGasMeter(app.finalizeBlockState.Context())
app.finalizeBlockState.SetContext(app.finalizeBlockState.Context().WithBlockGasMeter(gasMeter))
if app.checkState != nil {
app.checkState.SetContext(app.checkState.Context().
WithBlockGasMeter(gasMeter).
WithHeaderHash(req.Hash))
}
if err := app.preBlock(req); err != nil {
return nil, err
}
beginBlock, err := app.beginBlock(req)
if err != nil {
return nil, err
}
// First check for an abort signal after beginBlock, as it's the first place
// we spend any significant amount of time.
select {
case <-ctx.Done():
return nil, ctx.Err()
default:
// continue
}
events = append(events, beginBlock.Events...)
// Reset the gas meter so that the AnteHandlers aren't required to
gasMeter = app.getBlockGasMeter(app.finalizeBlockState.Context())
app.finalizeBlockState.SetContext(app.finalizeBlockState.Context().WithBlockGasMeter(gasMeter))
// Iterate over all raw transactions in the proposal and attempt to execute
// them, gathering the execution results.
//
// NOTE: Not all raw transactions may adhere to the sdk.Tx interface, e.g.
// vote extensions, so skip those.
txResults := make([]*abci.ExecTxResult, 0, len(req.Txs))
for _, rawTx := range req.Txs {
response := app.deliverTx(rawTx)
// check after every tx if we should abort
select {
case <-ctx.Done():
return nil, ctx.Err()
default:
// continue
}
txResults = append(txResults, response)
}
if app.finalizeBlockState.ms.TracingEnabled() {
app.finalizeBlockState.ms = app.finalizeBlockState.ms.SetTracingContext(nil).(storetypes.CacheMultiStore)
}
endBlock, err := app.endBlock(app.finalizeBlockState.Context())
if err != nil {
return nil, err
}
// check after endBlock if we should abort, to avoid propagating the result
select {
case <-ctx.Done():
return nil, ctx.Err()
default:
// continue
}
events = append(events, endBlock.Events...)
cp := app.GetConsensusParams(app.finalizeBlockState.Context())
return &abci.FinalizeBlockResponse{
Events: events,
TxResults: txResults,
ValidatorUpdates: endBlock.ValidatorUpdates,
ConsensusParamUpdates: &cp,
}, nil
}
// FinalizeBlock will execute the block proposal provided by RequestFinalizeBlock.
// Specifically, it will execute an application's BeginBlock (if defined), followed
// by the transactions in the proposal, finally followed by the application's
// EndBlock (if defined).
//
// For each raw transaction, i.e. a byte slice, BaseApp will only execute it if
// it adheres to the sdk.Tx interface. Otherwise, the raw transaction will be
// skipped. This is to support compatibility with proposers injecting vote
// extensions into the proposal, which should not themselves be executed in cases
// where they adhere to the sdk.Tx interface.
func (app *BaseApp) FinalizeBlock(req *abci.FinalizeBlockRequest) (res *abci.FinalizeBlockResponse, err error) {
defer func() {
// call the streaming service hooks with the FinalizeBlock messages
for _, streamingListener := range app.streamingManager.ABCIListeners {
if err := streamingListener.ListenFinalizeBlock(app.finalizeBlockState.Context(), *req, *res); err != nil {
app.logger.Error("ListenFinalizeBlock listening hook failed", "height", req.Height, "err", err)
}
}
}()
if app.optimisticExec.Initialized() {
// check if the hash we got is the same as the one we are executing
aborted := app.optimisticExec.AbortIfNeeded(req.Hash)
// Wait for the OE to finish, regardless of whether it was aborted or not
res, err = app.optimisticExec.WaitResult()
// only return if we are not aborting
if !aborted {
if res != nil {
res.AppHash = app.workingHash()
}
return res, err
}
// if it was aborted, we need to reset the state
app.finalizeBlockState = nil
app.optimisticExec.Reset()
}
// if no OE is running, just run the block (this is either a block replay or a OE that got aborted)
res, err = app.internalFinalizeBlock(context.Background(), req)
if res != nil {
res.AppHash = app.workingHash()
}
return res, err
}
// checkHalt checks if height or time exceeds halt-height or halt-time respectively.
func (app *BaseApp) checkHalt(height int64, time time.Time) error {
var halt bool
switch {
case app.haltHeight > 0 && uint64(height) > app.haltHeight:
halt = true
case app.haltTime > 0 && time.Unix() > int64(app.haltTime):
halt = true
}
if halt {
return fmt.Errorf("halt per configuration height %d time %d", app.haltHeight, app.haltTime)
}
return nil
}
// Commit implements the ABCI interface. It will commit all state that exists in
// the deliver state's multi-store and includes the resulting commit ID in the
// returned abci.ResponseCommit. Commit will set the check state based on the
// latest header and reset the deliver state. Also, if a non-zero halt height is
// defined in config, Commit will execute a deferred function call to check
// against that height and gracefully halt if it matches the latest committed
// height.
func (app *BaseApp) Commit() (*abci.CommitResponse, error) {
header := app.finalizeBlockState.Context().BlockHeader()
retainHeight := app.GetBlockRetentionHeight(header.Height)
if app.precommiter != nil {
app.precommiter(app.finalizeBlockState.Context())
}
rms, ok := app.cms.(*rootmulti.Store)
if ok {
rms.SetCommitHeader(header)
}
app.cms.Commit()
resp := &abci.CommitResponse{
RetainHeight: retainHeight,
}
abciListeners := app.streamingManager.ABCIListeners
if len(abciListeners) > 0 {
ctx := app.finalizeBlockState.Context()
blockHeight := ctx.BlockHeight()
changeSet := app.cms.PopStateCache()
for _, abciListener := range abciListeners {
if err := abciListener.ListenCommit(ctx, *resp, changeSet); err != nil {
app.logger.Error("Commit listening hook failed", "height", blockHeight, "err", err)
}
}
}
// Reset the CheckTx state to the latest committed.
//
// NOTE: This is safe because CometBFT holds a lock on the mempool for
// Commit. Use the header from this latest block.
app.setState(execModeCheck, header)
app.finalizeBlockState = nil
if app.prepareCheckStater != nil {
app.prepareCheckStater(app.checkState.Context())
}
// The SnapshotIfApplicable method will create the snapshot by starting the goroutine
app.snapshotManager.SnapshotIfApplicable(header.Height)
return resp, nil
}
// workingHash gets the apphash that will be finalized in commit.
// These writes will be persisted to the root multi-store (app.cms) and flushed to
// disk in the Commit phase. This means when the ABCI client requests Commit(), the application
// state transitions will be flushed to disk and as a result, but we already have
// an application Merkle root.
func (app *BaseApp) workingHash() []byte {
// Write the FinalizeBlock state into branched storage and commit the MultiStore.
// The write to the FinalizeBlock state writes all state transitions to the root
// MultiStore (app.cms) so when Commit() is called it persists those values.
app.finalizeBlockState.ms.Write()
// Get the hash of all writes in order to return the apphash to the comet in finalizeBlock.
commitHash := app.cms.WorkingHash()
app.logger.Debug("hash of all writes", "workingHash", fmt.Sprintf("%X", commitHash))
return commitHash
}
func handleQueryApp(app *BaseApp, path []string, req *abci.QueryRequest) *abci.QueryResponse {
if len(path) >= 2 {
switch path[1] {
case "simulate":
txBytes := req.Data
gInfo, res, err := app.Simulate(txBytes)
if err != nil {
return queryResult(errorsmod.Wrap(err, "failed to simulate tx"), app.trace)
}
simRes := &sdk.SimulationResponse{
GasInfo: gInfo,
Result: res,
}
bz, err := codec.ProtoMarshalJSON(simRes, app.interfaceRegistry)
if err != nil {
return queryResult(errorsmod.Wrap(err, "failed to JSON encode simulation response"), app.trace)
}
return &abci.QueryResponse{
Codespace: sdkerrors.RootCodespace,
Height: req.Height,
Value: bz,
}
case "version":
return &abci.QueryResponse{
Codespace: sdkerrors.RootCodespace,
Height: req.Height,
Value: []byte(app.version),
}
default:
return queryResult(errorsmod.Wrapf(sdkerrors.ErrUnknownRequest, "unknown query: %s", path), app.trace)
}
}
return queryResult(
errorsmod.Wrap(
sdkerrors.ErrUnknownRequest,
"expected second parameter to be either 'simulate' or 'version', neither was present",
), app.trace)
}
func handleQueryStore(app *BaseApp, path []string, req abci.QueryRequest) *abci.QueryResponse {
// "/store" prefix for store queries
queryable, ok := app.cms.(storetypes.Queryable)
if !ok {
return queryResult(errorsmod.Wrap(sdkerrors.ErrUnknownRequest, "multi-store does not support queries"), app.trace)
}
req.Path = "/" + strings.Join(path[1:], "/")
if req.Height <= 1 && req.Prove {
return queryResult(
errorsmod.Wrap(
sdkerrors.ErrInvalidRequest,
"cannot query with proof when height <= 1; please provide a valid height",
), app.trace)
}
sdkReq := storetypes.RequestQuery(req)
resp, err := queryable.Query(&sdkReq)
if err != nil {
return queryResult(err, app.trace)
}
resp.Height = req.Height
abciResp := abci.QueryResponse(*resp)
return &abciResp
}
func handleQueryP2P(app *BaseApp, path []string) *abci.QueryResponse {
// "/p2p" prefix for p2p queries
if len(path) < 4 {
return queryResult(errorsmod.Wrap(sdkerrors.ErrUnknownRequest, "path should be p2p filter <addr|id> <parameter>"), app.trace)
}
var resp *abci.QueryResponse
cmd, typ, arg := path[1], path[2], path[3]
switch cmd {
case "filter":
switch typ {
case "addr":
resp = app.FilterPeerByAddrPort(arg)
case "id":
resp = app.FilterPeerByID(arg)
}
default:
resp = queryResult(errorsmod.Wrap(sdkerrors.ErrUnknownRequest, "expected second parameter to be 'filter'"), app.trace)
}
return resp
}
// SplitABCIQueryPath splits a string path using the delimiter '/'.
//
// e.g. "this/is/funny" becomes []string{"this", "is", "funny"}
func SplitABCIQueryPath(requestPath string) (path []string) {
path = strings.Split(requestPath, "/")
// first element is empty string
if len(path) > 0 && path[0] == "" {
path = path[1:]
}
return path
}
// FilterPeerByAddrPort filters peers by address/port.
func (app *BaseApp) FilterPeerByAddrPort(info string) *abci.QueryResponse {
if app.addrPeerFilter != nil {
return app.addrPeerFilter(info)
}
return &abci.QueryResponse{}
}
// FilterPeerByID filters peers by node ID.
func (app *BaseApp) FilterPeerByID(info string) *abci.QueryResponse {
if app.idPeerFilter != nil {
return app.idPeerFilter(info)
}
return &abci.QueryResponse{}
}
// getContextForProposal returns the correct Context for PrepareProposal and
// ProcessProposal. We use finalizeBlockState on the first block to be able to
// access any state changes made in InitChain.
func (app *BaseApp) getContextForProposal(ctx sdk.Context, height int64) sdk.Context {
if height == app.initialHeight {
ctx, _ = app.finalizeBlockState.Context().CacheContext()
// clear all context data set during InitChain to avoid inconsistent behavior
ctx = ctx.WithHeaderInfo(coreheader.Info{}).WithBlockHeader(cmtproto.Header{})
return ctx
}
return ctx
}
func (app *BaseApp) handleQueryGRPC(handler GRPCQueryHandler, req *abci.QueryRequest) *abci.QueryResponse {
ctx, err := app.CreateQueryContext(req.Height, req.Prove)
if err != nil {
return queryResult(err, app.trace)
}
resp, err := handler(ctx, req)
if err != nil {
resp = queryResult(gRPCErrorToSDKError(err), app.trace)
resp.Height = req.Height
return resp
}
return resp
}
func gRPCErrorToSDKError(err error) error {
status, ok := grpcstatus.FromError(err)
if !ok {
return errorsmod.Wrap(sdkerrors.ErrInvalidRequest, err.Error())
}
switch status.Code() {
case codes.NotFound:
return errorsmod.Wrap(sdkerrors.ErrKeyNotFound, err.Error())
case codes.InvalidArgument:
return errorsmod.Wrap(sdkerrors.ErrInvalidRequest, err.Error())
case codes.FailedPrecondition:
return errorsmod.Wrap(sdkerrors.ErrInvalidRequest, err.Error())
case codes.Unauthenticated:
return errorsmod.Wrap(sdkerrors.ErrUnauthorized, err.Error())
default:
return errorsmod.Wrap(sdkerrors.ErrUnknownRequest, err.Error())
}
}
func checkNegativeHeight(height int64) error {
if height < 0 {
return errorsmod.Wrap(sdkerrors.ErrInvalidRequest, "cannot query with height < 0; please provide a valid height")
}
return nil
}
// CreateQueryContext creates a new sdk.Context for a query, taking as args
// the block height and whether the query needs a proof or not.
func (app *BaseApp) CreateQueryContext(height int64, prove bool) (sdk.Context, error) {
if err := checkNegativeHeight(height); err != nil {
return sdk.Context{}, err
}
// use custom query multi-store if provided
qms := app.qms
if qms == nil {
qms = app.cms.(storetypes.MultiStore)
}
lastBlockHeight := qms.LatestVersion()
if lastBlockHeight == 0 {
return sdk.Context{}, errorsmod.Wrapf(sdkerrors.ErrInvalidHeight, "%s is not ready; please wait for first block", app.Name())
}
if height > lastBlockHeight {
return sdk.Context{},
errorsmod.Wrap(
sdkerrors.ErrInvalidHeight,
"cannot query with height in the future; please provide a valid height",
)
}
// when a client did not provide a query height, manually inject the latest
if height == 0 {
height = lastBlockHeight
}
if height <= 1 && prove {
return sdk.Context{},
errorsmod.Wrap(
sdkerrors.ErrInvalidRequest,
"cannot query with proof when height <= 1; please provide a valid height",
)
}
cacheMS, err := qms.CacheMultiStoreWithVersion(height)
if err != nil {
return sdk.Context{},
errorsmod.Wrapf(
sdkerrors.ErrNotFound,
"failed to load state at height %d; %s (latest height: %d)", height, err, lastBlockHeight,
)
}
// branch the commit multi-store for safety
ctx := sdk.NewContext(cacheMS, true, app.logger).
WithMinGasPrices(app.minGasPrices).
WithGasMeter(storetypes.NewGasMeter(app.queryGasLimit)).
WithHeaderInfo(coreheader.Info{
ChainID: app.chainID,
Height: height,
}).
WithBlockHeader(app.checkState.Context().BlockHeader()).
WithBlockHeight(height)
if height != lastBlockHeight {
rms, ok := app.cms.(*rootmulti.Store)
if ok {
cInfo, err := rms.GetCommitInfo(height)
if cInfo != nil && err == nil {
ctx = ctx.WithHeaderInfo(coreheader.Info{Height: height, Time: cInfo.Timestamp})
}
}
}
return ctx, nil
}
// GetBlockRetentionHeight returns the height for which all blocks below this height
// are pruned from CometBFT. Given a commitment height and a non-zero local
// minRetainBlocks configuration, the retentionHeight is the smallest height that
// satisfies:
//
// - Unbonding (safety threshold) time: The block interval in which validators
// can be economically punished for misbehavior. Blocks in this interval must be
// auditable e.g. by the light client.
//
// - Logical store snapshot interval: The block interval at which the underlying
// logical store database is persisted to disk, e.g. every 10000 heights. Blocks
// since the last IAVL snapshot must be available for replay on application restart.
//
// - State sync snapshots: Blocks since the oldest available snapshot must be
// available for state sync nodes to catch up (oldest because a node may be
// restoring an old snapshot while a new snapshot was taken).
//
// - Local (minRetainBlocks) config: Archive nodes may want to retain more or
// all blocks, e.g. via a local config option min-retain-blocks. There may also
// be a need to vary retention for other nodes, e.g. sentry nodes which do not
// need historical blocks.
func (app *BaseApp) GetBlockRetentionHeight(commitHeight int64) int64 {
// pruning is disabled if minRetainBlocks is zero
if app.minRetainBlocks == 0 {
return 0
}
minNonZero := func(x, y int64) int64 {
switch {
case x == 0:
return y
case y == 0:
return x
case x < y:
return x
default:
return y
}
}
// Define retentionHeight as the minimum value that satisfies all non-zero
// constraints. All blocks below (commitHeight-retentionHeight) are pruned
// from CometBFT.
var retentionHeight int64
// Define the number of blocks needed to protect against misbehaving validators
// which allows light clients to operate safely. Note, we piggy back of the
// evidence parameters instead of computing an estimated number of blocks based
// on the unbonding period and block commitment time as the two should be
// equivalent.
cp := app.GetConsensusParams(app.finalizeBlockState.Context())
if cp.Evidence != nil && cp.Evidence.MaxAgeNumBlocks > 0 {
retentionHeight = commitHeight - cp.Evidence.MaxAgeNumBlocks
}
if app.snapshotManager != nil {
snapshotRetentionHeights := app.snapshotManager.GetSnapshotBlockRetentionHeights()
if snapshotRetentionHeights > 0 {
retentionHeight = minNonZero(retentionHeight, commitHeight-snapshotRetentionHeights)
}
}
v := commitHeight - int64(app.minRetainBlocks)
retentionHeight = minNonZero(retentionHeight, v)
if retentionHeight <= 0 {
// prune nothing in the case of a non-positive height
return 0
}
return retentionHeight
}
// toVoteInfo converts the new ExtendedVoteInfo to VoteInfo.
func toVoteInfo(votes []abci.ExtendedVoteInfo) []abci.VoteInfo {
legacyVotes := make([]abci.VoteInfo, len(votes))
for i, vote := range votes {
legacyVotes[i] = abci.VoteInfo{
Validator: abci.Validator{
Address: vote.Validator.Address,
Power: vote.Validator.Power,
},
BlockIdFlag: vote.BlockIdFlag,
}
}
return legacyVotes
}