fix: stmgr: make the tipset and height agree when estimating gas #10216

Stebalien · 2023-02-08T19:52:16Z

Related Issues

Should fix filecoin-project/ref-fvm#1648.

Proposed Changes

Specifically re-execute all messages in the current tipset, tacking the new message onto the end. That way, the epoch is the epoch of the current tipset.

We could try to "make" a fake block and use that, but that's unlikely to work well.

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, market, mempool, multisig, networking, paych, proving, sealing, wallet, deps
Tests exist for new functionality or change in behavior
CI is green

maciejwitowski · 2023-02-09T17:12:13Z

Tested #10216, it works now.

snissn · 2023-02-09T23:31:41Z

This test also confirms this PR works #10206

snissn · 2023-02-09T23:36:33Z

chain/stmgr/call.go

@@ -142,14 +136,14 @@ func (sm *StateManager) callInternal(ctx context.Context, msg *types.Message, pr
 	buffStore := blockstore.NewTieredBstore(sm.cs.StateBlockstore(), blockstore.NewMemorySync())
 	vmopt := &vm.VMOpts{
 		StateBase:      stateCid,
-		Epoch:          vmHeight,
+		Epoch:          ts.Height(),


we're changing the value of Epoch here from vmHeight (ts.Height()+1) to ts.Height() is that off by 1 change ok?

Yes, that's the point.

magik6k · 2023-02-10T11:51:53Z

chain/stmgr/call.go

+	// Unless executing on a specific state cid, apply all the messages from the current tipset
+	// first. Unfortunately, we can't just execute the tipset, because that will run cron. We
+	// don't want to apply miner messages after cron runs in a given epoch.
+	if stateCid == cid.Undef {
+		stateCid = ts.ParentState()
+		tsMsgs, err := sm.cs.MessagesForTipset(ctx, ts)
+		if err != nil {
+			return nil, xerrors.Errorf("failed to lookup messages for parent tipset: %w", err)
 		}
+		priorMsgs = append(tsMsgs, priorMsgs...)


Are we sure this isn't breaking StateCall / lotus state compute-state?

Exchanges rely on this to track transactions.

That uses

lotus/chain/stmgr/utils.go

Line 68 in 7aa6663

func ComputeState(ctx context.Context, sm *StateManager, height abi.ChainEpoch, msgs []*types.Message, ts *types.TipSet) (cid.Cid, []*api.InvocResult, error) {

Which is also incorrect? Well, it's correct if you don't pass any messages.

Hm. It does affect StateCall. I'm going to rework this a bit.

Done. I now only do this for CallWithGas.

Specifically re-execute all messages in the current tipset, tacking the new message onto the end. That way, the epoch is the epoch of the current tipset. We could try to "make" a fake block and use that, but that's unlikely to work well.

…in-project#71) * Test and fix eth_FeeHistory * Add test for eth_getStorageAt * Add test for eth_call * Change EthBlock.Extradata type to EthBytes * fix: EthAPI: Drop hack in GetStorageAt * Eth JSON-RPC: implement web3_clientVersion * fix lint * add genesis for ipc gateway * Remove debug logging * chore: cli Standarize cli/code functions similar to: filecoin-project#9317 - cctx.NArg() instead of cctx.Args().xxx - Add check for args and print help on functions that did not have it * fix: rpcenc: Don't hang when source dies * make docsgen-cli make docsgen-cli * fix: EthAPI: Handle EthCall return correctly * feat: vm: Assert empty object CID when dumping state * Update codeql to v2 * Enable code ql for release branches * less strict ArgsCheck Less strict ArgsCheck * feat: wdpost: Emit more detailed errors * Fix panic in EthGetCode * Update codeql to v2 * Enable code ql for release branches * Add actors to circ-supply - Consider funds to EAM as locked - Add evm, placeholder and ethaccount to circ supply * devnets starts with nv17 * Use new kit.DisableEthRPC method in test * make gen * init git submodules in build-docker CI * Remove old Snapcraft and AppDir folders * Always include checked in files in docker context The .dockerignore file is symlinked to the .gitignore file, and checked in files should not be removed from our docker context otherwise they result in dirty git state when we build our images. * Check git state in Dockerfile to catch .dockerignore bugs * make jen - Add builtin.EthereumAddressManagerActorAddr to builtin.go.template and make jen - Rename to EthereumAddressManagerActorAddr to match pattern of other actors (CronActorAddr/etc) * chore: all: bump go-libipfs to replace go-block-format Includes changes from: - ipfs/go-block-format#37 - ipfs/boxo#58 * init git submodules in build-docker CI * Remove old Snapcraft and AppDir folders * Always include checked in files in docker context The .dockerignore file is symlinked to the .gitignore file, and checked in files should not be removed from our docker context otherwise they result in dirty git state when we build our images. * Check git state in Dockerfile to catch .dockerignore bugs * Fix lint errors * Skip some rpc conformance tests * make gen * Changed skip message * fix: don't replace the method in EthSendRawTransaction (filecoin-project#10129) This will just cause signature validation to fail. * Remove stm: #integration comment * Fix comment * fix: devnets: default to starting from nv17 * Refactor to ensure conformance test can run in circleci * go fmt * Standardize path variable * Update to go-state-types v0.10.0-alpha-11 * chore: node: migrate go-bitswap to go-libipfs/bitswap This was migrated in ipfs/boxo#63. * simplify transaction hash db queries, prevent fd leaks * Refactor: Unify EthTx to FilecoinMessage methods v2 (filecoin-project#10095) * Refactor: Unify EthTx to FilecoinMessage methods * Filecoin messages can again be converted to Eth Txs * All BLS messages should calculated tx hash with unsigned message * Refactor newEthTxReceipt * fill in from and to for non-eth transactions * Hoist nil check out of newEthTxFromMessageLookup --------- Co-authored-by: Aayush <arajasek94@gmail.com> Co-authored-by: Raúl Kripalani <raul@protocol.ai> * Populate transaction hash database if the database has not been set up before * rpc: Switch eth_subscribe to reverse calls * gateway: eth_subscribe support * itests: Fix TestEthSubscribeLogs * ethtypes: Serialize EthFilterID/EthSubscriptionID correctly * eth rpc: Params are optional in eth_subscribe * fix: cli: add ArgsUsage fix: cli: add ArgsUsage * make libp2p user agent overridable * move UserAgentOption upwards * Eth JSON-RPC: from in eth_getTransactionByHash is not correctly populated filecoin-project#1614 * Move assigning `from` address to the API level * Improve the error message * Add test for EthGetTransactionByHash * fix: should not serve non v0 api in v1 * Fix: typo Fix: typo * Fix: typos Fix: typos * Fix: typo Fix: typo * Fix: typo Fix: typo * Fix: typo Fix: typo * Fix: typos Fix: typos * Fix: typos Fix: typos * test: fevm: add in tests for deploying, destroying contracts, recursive calls, sending value (filecoin-project#10082) adds the following tests to itests/fevm_test.go: - recursive tests - delegate call tests - delegate call recursive tests - revert tests - destruct tests - contract deploy address tests - send value to contracts - gas limit on value transfer tests - sending value to destroyed contracts adds the test to itests/fevm_address_test.go: - deploy contract and confirm address is different second deploy * merge * tests that use create2 and destroy to validate evm state * wip: ipc_gateway in genesis * deps: Update go-jsonrpc to v0.2.1 * Test: assert all fields returned from EthGetTransactionByHash have expected values Related to filecoin-project#10151 (comment) * Fix merge conflicts * Allow f4 address to send to all address types if ID address exists on chain * fix: extend LOTUS_CHAIN_BADGERSTORE_DISABLE_FSYNC to the markset Without doing this walking a badger markset on a non-nvme knocks the node hopelessly out of sync during a compaction. * review fixes * gateway: Support all EthModule methods * node builder: Use gateway eth module in lite mode * feat: ethrpc: Support filtering by address in subscribe * itests: Fix TestEthFilterAPIDisabledViaConfig * Check decoding params for new methods * Review fixes * feat: evm: align events implementation with FIP-0049 and FIP-0054. (filecoin-project#10152) - Event keys are now t1, t2, t3, t4 for topics; and d for data. - ref-fvm no longer stores events in the blockstore for us. It just returns events to the client, who is now responsible for handling them as it wishes / according to its configuration. - Add a flag to VMOpts to have the events AMT be written in the blockstore. - Add a flag to the ChainStore to advertise to the rest of the system if the ChainStore is storing events. - Enable that flag if the EthRPC is enabled (can also add an explicit configuration flag if wanted). * cli: fix extend cmd to get the right sector number * todo: rebase master to include release/v1.20.0 * feat: compute a better gas limit for recursive external contract calls * make gen * fix: ethtypes: Correct 'no uncles' hash in NewEthBlock * LOTUS_FEVM_ENABLEETHRPC: Fix env variable name in error * fix: ethtypes: Correct 'no transactions' hash in NewEthBlock * fix: null rounds: pass correct timestamp to the FVM. * itest: fix FEVM tests for upstream changes * itest: fix: test comment * itest: fix remaining fevm failures * improve evm error handling in itests (filecoin-project#10161) * use WithValue language for test * clean up test for recursive delegate call count. improved readability (filecoin-project#10195) * chain: explicitly check that gasLimit is above zero * add bundle git tag from pack.sh into builtin_actors_gen * fix: worker: add all task type flag Add all flag for the `lotus-worker tasks enable/disable` cmd * Update cmd/lotus-worker/tasks.go Co-authored-by: Łukasz Magiera <magik6k@users.noreply.github.com> * Update cmd/lotus-worker/tasks.go Co-authored-by: Łukasz Magiera <magik6k@users.noreply.github.com> * Update cmd/lotus-worker/tasks.go Co-authored-by: Łukasz Magiera <magik6k@users.noreply.github.com> * make docsgen-cli make docsgen-cli * retry make docsgen-cli retry make docsgen-cli with the new usage * remove bundle-gen from make gen - not actually useful today anyway * fvm: chore: update FVM This: 1. Updates the builtin actors bundle (for actors v10). 2. Updates the event entry type to include the codec. 3. Removes the cbor encoding and zero trimming from event data. I've chose to: 1. _Not_ add codec handling to the event filtering system for now. 2. _Skip_ events with unexpected codecs. We don't actually _allow_ these events in the FVM right now, and it simplifies the implementation. However, I _am_ recording the codecs in the database so we don't have to migrate it later. * Option to set-seal-delay seconds Add the option to set-seal-delay in seconds * make docsgen-cli make docsgen-cli * Error if backup file exists Error out if backup file already exists * fix: ethrpc: Don't send sub notifs in array * Update eth_subscribe itests to work with correct responses * test: build: ignore git tags in embedded metadata * fix: build: use actual v9 bundles for butterfly * fix: ethrpc: emit correct bloom filters (filecoin-project#10194) * skip problematic test vectors. * feat: actors: Add bytecode getter * itests: Test EVM bytecode getters, check bytecode hash * ethrpc: Add missing fields to EthTx * Update cli/backup.go Co-authored-by: Łukasz Magiera <magik6k@users.noreply.github.com> * feat: eth cli: Strip out empty spaces around contract bytes * fix: chain: make sure the head is empty, not the code * bump test vectors. * remove test vectors bleeding edge job. We no longer need it because specs-actors is deprecated. v7 vectors have been merged to master. * fix: Don't call WalletExport in msg signing flows * updates butterflynet reset artifacts * bootstrap node multiaddr * new genesis file these files are necessary for others to build lotus and join the new butterflynet, reset on February 08, 2023 * fix: gas: update ffi & correct the message inclusion cost in nv18 (filecoin-project#10228) Co-authored-by: Raúl Kripalani <raul@protocol.ai> * refactor: use EthHash for event topics This ensures they're always 32 bytes and padded, as required. * fix: eth: strict event parsing We now enforce the following rules: 1. No duplicate topics or data. 2. Topics must have 32 byte keys. 3. Topics may not be skipped. (e.g., no t1 & t3 without a t2). 4. Raw codecs. We _don't_ require that topics/data be emitted in any specific order. We _skip_ events with unknown keys. We _drop_ events that violate the above rules. * Apply suggestions from code review Co-authored-by: raulk <raul@protocol.ai> * fix: eth: log on unexpected events We can remove these later as we add more event types, but this will aid in debugging. * fix: make gen * fix config.yml * implement itest and handle optional params * fix typo * fix: stmgr: make the tipset and height agree when estimating gas (filecoin-project#10216) * fix: stmgr: make the tipset and height agree when estimating gas Specifically re-execute all messages in the current tipset, tacking the new message onto the end. That way, the epoch is the epoch of the current tipset. We could try to "make" a fake block and use that, but that's unlikely to work well. * fix: stmgr: only apply tipset messages for CallWithGas * fix: itest: window post dispute * eth_feeHistory: parse block param correctly. * add todo * Update to FFI v1.20.0-rc1 * fix bad test. * improve TODO. * Update to go-state-types v0.10.0-rc2 * disable adding git tags to bundle metadata * Update actors to v10.0.0-rc.1 * eth: FIP-0055: implement final version of transitory delegated signature. (filecoin-project#10239) * make lint happy and re-generate devgen.car * setup ipc gateway in genesis and assign initial balance * override with ipc-actors type to load ipc bundle automatically * set balance of reward actor according to network name and minor fix * linter fixes * make linter happy :) * deploy ipc subnet actor * fix config.yml --------- Co-authored-by: Ian Davis <nospam@iandavis.com> Co-authored-by: Aayush <arajasek94@gmail.com> Co-authored-by: ychiao <ychiaoli18@users.noreply.github.com> Co-authored-by: cryptoAtwill <willes.lau@protocol.ai> Co-authored-by: Phi <orjan.roren@gmail.com> Co-authored-by: Łukasz Magiera <magik6k@gmail.com> Co-authored-by: Geoff Stuart <geoff.vball@gmail.com> Co-authored-by: Jiaying Wang <42981373+jennijuju@users.noreply.github.com> Co-authored-by: Geoff Stuart <geoffrey.stuart@protocol.ai> Co-authored-by: Łukasz Magiera <magik6k@users.noreply.github.com> Co-authored-by: Jennifer Wang <jiayingw703@gmail.com> Co-authored-by: Ian Davis <jungziege@gmail.com> Co-authored-by: Jorropo <jorropo.pgm@gmail.com> Co-authored-by: Steven Allen <steven@stebalien.com> Co-authored-by: Raúl Kripalani <raul@protocol.ai> Co-authored-by: Richard Guan <richard.guan@protocol.ai> Co-authored-by: Maciej Witowski <maciej.witowski@protocol.ai> Co-authored-by: omahs <73983677+omahs@users.noreply.github.com> Co-authored-by: snissn <snissn@gmail.com> Co-authored-by: Peter Rabbitson <ribasushi@protocol.ai> Co-authored-by: 0x5459 <0x5459@protonmail.com> Co-authored-by: Travis Person <travisperson@users.noreply.github.com> Co-authored-by: Anton Evangelatov <anton.evangelatov@gmail.com> Co-authored-by: ognots <ognots@protonmail.com>

Stebalien requested a review from a team as a code owner February 8, 2023 19:52

Stebalien mentioned this pull request Feb 8, 2023

Illegal argument error when calling block.difficulty filecoin-project/ref-fvm#1648

Closed

Stebalien requested a review from snissn February 8, 2023 19:54

Stebalien force-pushed the steb/fix-tipset-selection branch from f2373c5 to c0f83f2 Compare February 8, 2023 20:13

Stebalien linked an issue Feb 8, 2023 that may be closed by this pull request

Illegal argument error when calling block.difficulty filecoin-project/ref-fvm#1648

Closed

snissn mentioned this pull request Feb 9, 2023

test: itest: fevm block.difficulty #10206

Closed

7 tasks

snissn reviewed Feb 9, 2023

View reviewed changes

magik6k reviewed Feb 10, 2023

View reviewed changes

Stebalien added 2 commits February 10, 2023 09:19

fix: stmgr: only apply tipset messages for CallWithGas

72946f5

Stebalien force-pushed the steb/fix-tipset-selection branch from c0f83f2 to 72946f5 Compare February 10, 2023 17:23

fix: itest: window post dispute

1011ed4

arajasek approved these changes Feb 10, 2023

View reviewed changes

Stebalien enabled auto-merge (squash) February 10, 2023 20:47

Stebalien merged commit 01e616f into release/v1.20.0 Feb 10, 2023

Stebalien deleted the steb/fix-tipset-selection branch February 10, 2023 20:49

Fatman13 mentioned this pull request Feb 13, 2023

[venus] merge nv18 consensus / NV18 代码合并 filecoin-project/venus#5550

Closed

Stebalien mentioned this pull request Apr 19, 2024

fix: ETH RPC API: ETH Call should use the parent state root of the subsequent tipset #11905

Merged

8 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: stmgr: make the tipset and height agree when estimating gas #10216

fix: stmgr: make the tipset and height agree when estimating gas #10216

Stebalien commented Feb 8, 2023 •

edited by snissn

maciejwitowski commented Feb 9, 2023 •

edited

snissn commented Feb 9, 2023

snissn Feb 9, 2023

Stebalien Feb 10, 2023

magik6k Feb 10, 2023

Stebalien Feb 10, 2023

Stebalien Feb 10, 2023

Stebalien Feb 10, 2023

Stebalien Feb 10, 2023

fix: stmgr: make the tipset and height agree when estimating gas #10216

fix: stmgr: make the tipset and height agree when estimating gas #10216

Conversation

Stebalien commented Feb 8, 2023 • edited by snissn

Related Issues

Proposed Changes

Checklist

maciejwitowski commented Feb 9, 2023 • edited

snissn commented Feb 9, 2023

snissn Feb 9, 2023

Choose a reason for hiding this comment

Stebalien Feb 10, 2023

Choose a reason for hiding this comment

magik6k Feb 10, 2023

Choose a reason for hiding this comment

Stebalien Feb 10, 2023

Choose a reason for hiding this comment

Stebalien Feb 10, 2023

Choose a reason for hiding this comment

Stebalien Feb 10, 2023

Choose a reason for hiding this comment

Stebalien Feb 10, 2023

Choose a reason for hiding this comment

Stebalien commented Feb 8, 2023 •

edited by snissn

maciejwitowski commented Feb 9, 2023 •

edited