github.com/tacshi/go-ethereum@v0.0.0-20230616113857-84a434e20921/docs/postmortems/2021-08-22-split-postmortem.md (about) 1 # Minority split 2021-08-27 post mortem 2 3 This is a post-mortem concerning the minority split that occurred on Ethereum mainnet on block [13107518](https://etherscan.io/block/13107518), at which a minority chain split occurred. 4 5 ## Timeline 6 7 - 2021-08-17: Guido Vranken submitted a bounty report. Investigation started, root cause identified, patch variations discussed. 8 - 2021-08-18: Made public announcement over twitter about upcoming security release upcoming Tuesday. Downstream projects were also notified about the upcoming patch-release. 9 - 2021-08-24: Released [v1.10.8](https://github.com/tacshi/go-ethereum/releases/tag/v1.10.8) containing the fix on Tuesday morning (CET). Erigon released [v2021.08.04](https://github.com/ledgerwatch/erigon/releases/tag/v2021.08.04). 10 - 2021-08-27: At 12:50:07 UTC, issue exploited. Analysis started roughly 30m later, 11 12 ## Bounty report 13 14 ### 2021-08-17 RETURNDATA corruption via datacopy 15 16 On 2021-08-17, Guido Vranken submitted a report to bounty@ethereum.org. This coincided with a geth-meetup in Berlin, so the geth team could fairly quickly analyse the issue. 17 18 He submitted a proof of concept which called the `dataCopy` precompile, where the input slice and output slice were overlapping but shifted. Doing a `copy` where the `src` and `dest` overlaps is not a problem in itself, however, the `returnData`slice was _also_ using the same memory as a backing-array. 19 20 #### Technical details 21 22 During CALL-variants, `geth` does not copy the input. This was changed at one point, to avoid a DoS attack reported by Hubert Ritzdorf, to avoid copying data a lot on repeated `CALL`s -- essentially combating a DoS via `malloc`. Further, the datacopy precompile also does not copy the data, but just returns the same slice. This is fine so far. 23 24 After the execution of `dataCopy`, we copy the `ret` into the designated memory area, and this is what causes a problem. Because we're copying a slice of memory over a slice of memory, and this operation modifies (shifts) the data in the source -- the `ret`. So this means we wind up with corrupted returndata. 25 26 ``` 27 1. Calling datacopy 28 29 memory: [0, 1, 2, 3, 4] 30 in (mem[0:4]) : [0,1,2,3] 31 out (mem[1:5]): [1,2,3,4] 32 33 2. dataCopy returns 34 35 returndata (==in, mem[0:4]): [0,1,2,3] 36 37 3. Copy in -> out 38 39 => memory: [0,0,1,2,3] 40 => returndata: [0,0,1,2] 41 ``` 42 43 #### Summary 44 45 A memory-corruption bug within the EVM can cause a consensus error, where vulnerable nodes obtain a different `stateRoot` when processing a maliciously crafted transaction. This, in turn, would lead to the chain being split: mainnet splitting in two forks. 46 47 #### Handling 48 49 On the evening of 17th, we discussed options on how to handle it. We made a state test to reproduce the issue, and verified that neither `openethereum`, `nethermind` nor `besu` were affected by the same vulnerability, and started a full-sync with a patched version of `geth`. 50 51 It was decided that in this specific instance, it would be possible to make a public announcement and a patch release: 52 53 - The fix can be made pretty 'generically', e.g. always copying data on input to precompiles. 54 - The flaw is pretty difficult to find, given a generic fix in the call. The attacker needs to figure out that it concerns the precompiles, specifically the datcopy, and that it concerns the `RETURNDATA` buffer rather than the regular memory, and lastly the special circumstances to trigger it (overlapping but shifted input/output). 55 56 Since we had merged the removal of `ETH65`, if the entire network were to upgrade, then nodes which have not yet implemented `ETH66` would be cut off from the network. After further discussions, we decided to: 57 58 - Announce an upcoming security release on Tuesday (August 24th), via Twitter and official channels, plus reach out to downstream projects. 59 - Temporarily revert the `ETH65`-removal. 60 - Place the fix into the PR optimizing the jumpdest analysis [233381](https://github.com/tacshi/go-ethereum/pull/23381). 61 - After 4-8 weeks, release details about the vulnerability. 62 63 ## Exploit 64 65 At block [13107518](https://etherscan.io/block/13107518), mined at Aug-27-2021 12:50:07 PM +UTC, a minority chain split occurred. The discord user @AlexSSD7 notified the allcoredevs-channel on the Eth R&D discord, on Aug 27 13:09 UTC. 66 67 At 14:09 UTC, it was confirmed that the transaction `0x1cb6fb36633d270edefc04d048145b4298e67b8aa82a9e5ec4aa1435dd770ce4` had triggered the bug, leading to a minority-split of the chain. The term 'minority split' means that the majority of miners continued to mine on the correct chain. 68 69 At 14:17 UTC, @mhswende tweeted out about the issue [2]. 70 71 The attack was sent from an account funded from Tornado cash. 72 73 It was also found that the same attack had been carried out on the BSC chain at roughly the same time -- at a block mined [12 minutes earlier](https://bscscan.com/tx/0xf667f820631f6adbd04a4c92274374034a3e41fa9057dc42cb4e787535136dce), at Aug-27-2021 12:38:30 PM +UTC. 74 75 The blocks on the 'bad' chain were investigated, and Tim Beiko reached out to those mining operators on the minority chain who could be identified via block extradata. 76 77 ## Lessons learned 78 79 ### Disclosure decision 80 81 The geth-team have an official policy regarding [vulnerability disclosure](https://geth.ethereum.org/docs/vulnerabilities/vulnerabilities). 82 83 > The primary goal for the Geth team is the health of the Ethereum network as a whole, and the decision whether or not to publish details about a serious vulnerability boils down to minimizing the risk and/or impact of discovery and exploitation. 84 85 In this case, it was decided that public pre-announce + patch would likely lead to sufficient update-window for a critical mass of nodes/miners to upgrade in time before it could be exploited. In hindsight, this was a dangerous decision, and it's unlikely that the same decision would be reached were a similar incident to happen again. 86 87 ### Disclosure path 88 89 Several subprojects were informed about the upcoming security patch: 90 91 - Polygon/Matic 92 - MEV 93 - Avalanche 94 - Erigon 95 - BSC 96 - EWF 97 - Quorum 98 - ETC 99 - xDAI 100 101 However, some were 'lost', and only notified later 102 103 - Optimism 104 - Summa 105 - Harmony 106 107 Action point: create a low-volume geth-announce@ethereum.org email list where dependent projects/operators can receive public announcements. 108 109 - This has been done. If you wish to receive release- and security announcements, sign up [here](https://groups.google.com/a/ethereum.org/g/geth-announce/about) 110 111 ### Fork monitoring 112 113 The fork monitor behaved 'ok' during the incident, but had to be restarted during the evening. 114 115 Action point: improve the resiliency of the forkmon, which is currently not performing great when many nodes are connected. 116 117 Action point: enable push-based alerts to be sent from the forkmon, to speed up the fork detection. 118 119 ## Links 120 121 - [1] https://twitter.com/go_ethereum/status/1428051458763763721 122 - [2] https://twitter.com/mhswende/status/1431259601530458112 123 124 ## Appendix 125 126 ### Subprojects 127 128 The projects were sent variations of the following text: 129 130 ``` 131 We have identified a security issue with go-ethereum, and will issue a 132 new release (v1.10.8) on Tuesday next week. 133 134 At this point, we will not disclose details about the issue, but 135 recommend downstream/dependent projects to be ready to take actions to 136 upgrade to the latest go-ethereum codebase. More information about the 137 issue will be disclosed at a later date. 138 139 https://twitter.com/go_ethereum/status/1428051458763763721 140 141 ``` 142 143 ### Patch 144 145 ```diff 146 diff --git a/core/vm/instructions.go b/core/vm/instructions.go 147 index f7ef2f900e..6c8c6e6e6f 100644 148 --- a/core/vm/instructions.go 149 +++ b/core/vm/instructions.go 150 @@ -669,6 +669,7 @@ func opCall(pc *uint64, interpreter *EVMInterpreter, scope *ScopeContext) ([]byt 151 } 152 stack.push(&temp) 153 if err == nil || err == ErrExecutionReverted { 154 + ret = common.CopyBytes(ret) 155 scope.Memory.Set(retOffset.Uint64(), retSize.Uint64(), ret) 156 } 157 scope.Contract.Gas += returnGas 158 @@ -703,6 +704,7 @@ func opCallCode(pc *uint64, interpreter *EVMInterpreter, scope *ScopeContext) ([ 159 } 160 stack.push(&temp) 161 if err == nil || err == ErrExecutionReverted { 162 + ret = common.CopyBytes(ret) 163 scope.Memory.Set(retOffset.Uint64(), retSize.Uint64(), ret) 164 } 165 scope.Contract.Gas += returnGas 166 @@ -730,6 +732,7 @@ func opDelegateCall(pc *uint64, interpreter *EVMInterpreter, scope *ScopeContext 167 } 168 stack.push(&temp) 169 if err == nil || err == ErrExecutionReverted { 170 + ret = common.CopyBytes(ret) 171 scope.Memory.Set(retOffset.Uint64(), retSize.Uint64(), ret) 172 } 173 scope.Contract.Gas += returnGas 174 @@ -757,6 +760,7 @@ func opStaticCall(pc *uint64, interpreter *EVMInterpreter, scope *ScopeContext) 175 } 176 stack.push(&temp) 177 if err == nil || err == ErrExecutionReverted { 178 + ret = common.CopyBytes(ret) 179 scope.Memory.Set(retOffset.Uint64(), retSize.Uint64(), ret) 180 } 181 scope.Contract.Gas += returnGas 182 diff --git a/core/vm/interpreter.go b/core/vm/interpreter.go 183 index 9cf0c4e2c1..9fb83799c9 100644 184 --- a/core/vm/interpreter.go 185 +++ b/core/vm/interpreter.go 186 @@ -262,7 +262,7 @@ func (in *EVMInterpreter) Run(contract *Contract, input []byte, readOnly bool) ( 187 // if the operation clears the return data (e.g. it has returning data) 188 // set the last return to the result of the operation. 189 if operation.returns { 190 - in.returnData = common.CopyBytes(res) 191 + in.returnData = res 192 } 193 194 switch { 195 ``` 196 197 ### Statetest to test for the issue 198 199 ```json 200 { 201 "trigger-issue": { 202 "env": { 203 "currentCoinbase": "b94f5374fce5edbc8e2a8697c15331677e6ebf0b", 204 "currentDifficulty": "0x20000", 205 "currentGasLimit": "0x26e1f476fe1e22", 206 "currentNumber": "0x1", 207 "currentTimestamp": "0x3e8", 208 "previousHash": "0x0000000000000000000000000000000000000000000000000000000000000000" 209 }, 210 "pre": { 211 "0x00000000000000000000000000000000000000bb": { 212 "code": "0x6001600053600260015360036002536004600353600560045360066005536006600260066000600060047f7ef0367e633852132a0ebbf70eb714015dd44bc82e1e55a96ef1389c999c1bcaf13d600060003e596000208055", 213 "storage": {}, 214 "balance": "0x5", 215 "nonce": "0x0" 216 }, 217 "0xa94f5374fce5edbc8e2a8697c15331677e6ebf0b": { 218 "code": "0x", 219 "storage": {}, 220 "balance": "0xffffffff", 221 "nonce": "0x0" 222 } 223 }, 224 "transaction": { 225 "gasPrice": "0x1", 226 "nonce": "0x0", 227 "to": "0x00000000000000000000000000000000000000bb", 228 "data": ["0x"], 229 "gasLimit": ["0x7a1200"], 230 "value": ["0x01"], 231 "secretKey": "0x45a915e4d060149eb4365960e6a7a45f334393093061116b197e3240065ff2d8" 232 }, 233 "out": "0x", 234 "post": { 235 "Berlin": [ 236 { 237 "hash": "2a38a040bab1e1fa499253d98b2fd363e5756ecc52db47dd59af7116c068368c", 238 "logs": "1dcc4de8dec75d7aab85b567b6ccd41ad312451b948a7413f0a142fd40d49347", 239 "indexes": { 240 "data": 0, 241 "gas": 0, 242 "value": 0 243 } 244 } 245 ] 246 } 247 } 248 } 249 ```