GHSA-F3W5-V9XX-RP8P

Vulnerability from github – Published: 2021-12-20 18:17 – Updated: 2021-05-20 20:10
VLAI?
Summary
Signature verification failure in Tendermint
Details

The root cause of this security vulnerability is in the Tendermint specification, and this advisory is a duplicate of https://github.com/tendermint/spec/security/advisories/GHSA-jqfc-687g-59pw.

Impact

Tendermint light clients running versions 0.34.0 to 0.34.8 are unable to detect and punish a new kind of attack. We’re calling this a “forward lunatic attack,” or FLA. The severity of this vulnerability is moderate.

Note that an FLA cannot be successfully executed unless there are already ⅓+ Byzantine validators, and therefore outside of Tendermint’s security model; however, it is important to be able to detect and punish these kinds of attacks in order to incentivize correct behavior.

In an FLA, an attacking validator (with ⅓+ voting power) signs commit messages for arbitrary application state associated with a block height that hasn’t been seen yet, hence the name “forward lunatic attacks.” A malicious validator effectively executes a lunatic attack, but signs messages for a target block that is higher than the current block. This can be dangerous: Typically, misbehavior evidence is only created when there are conflicting blocks at the same height, but by targeting a block height that is far “ahead” of the current chain height, it’s possible that the chain will not produce a (conflicting) block at the target height in time to create evidence.

Prior to Tendermint v0.34.9, the light client could accept a bad header from its primary witness, and would not be able to form evidence of this deception, even if all the secondary witnesses were correct. Because the light client is responsible for verifying cross-chain state for IBC, a successful FLA could result in loss of funds. However, it is important to note that FLAs are only possible outside the Tendermint security model.

All FLAs, attempted and successful, leave traces of provable misbehavior on-chain. A faulty header contains signatures from the faulty validator, and even in unpatched versions of Tendermint Core, networks could use social consensus (off-chain action) to recover the network. The patches introduced in Tendermint Core v0.34.9 handle all evidence automatically and on-chain.

Note that this fix also allows for successful automatic reporting of FLAs, even after a chain halt. By adding a time to FetchBlock, light clients effectively have a backup way to determine if a halted chain should have continued, and it will be able to submit evidence as soon as the chain resumes.

Patches

This problem has been patched in Tendermint Core v0.34.9.

Workarounds

There are no workarounds. All users are recommended to upgrade to Tendermint Core v0.34.9 at their earliest possible convenience.

Credits

Thank you to @MaximilianDiez for originally surfacing this issue, and to @cmwaters, @josef-widder, and @milosevic for creating fixes at both the implementation and specification level.

For more information

If you have any questions or comments about this advisory: * Open an issue in tendermint/tendermint * Email us at security@tendermint.com

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "Go",
        "name": "github.com/tendermint/tendermint"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0.34.0"
            },
            {
              "fixed": "0.34.9"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [],
  "database_specific": {
    "cwe_ids": [
      "CWE-347"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2021-05-20T20:10:47Z",
    "nvd_published_at": null,
    "severity": "MODERATE"
  },
  "details": "_The root cause of this security vulnerability is in the Tendermint specification, and this advisory is a duplicate of https://github.com/tendermint/spec/security/advisories/GHSA-jqfc-687g-59pw._\n\n\n### Impact\nTendermint light clients running versions 0.34.0 to 0.34.8 are unable to detect and punish a new kind of attack. We\u2019re calling this a \u201cforward lunatic attack,\u201d or FLA. The severity of this vulnerability is _moderate_. \n\nNote that an FLA cannot be successfully executed unless there are already \u2153+ Byzantine validators, and therefore outside of Tendermint\u2019s security model; however, it is important to be able to detect and punish these kinds of attacks in order to incentivize correct behavior.\n\nIn an FLA, an attacking validator (with \u2153+ voting power) signs commit messages for arbitrary application state associated with a block height that hasn\u2019t been seen yet, hence the name \u201cforward lunatic attacks.\u201d A malicious validator effectively executes a [lunatic attack](https://docs.tendermint.com/master/spec/light-client/accountability/#the-misbehavior-of-faulty-validators), but signs messages for a target block that is higher than the current block. This can be dangerous: Typically, misbehavior evidence is only created when there are conflicting blocks at the same height, but by targeting a block height that is far \u201cahead\u201d of the current chain height, it\u2019s possible that the chain will not produce a (conflicting) block at the target height in time to create evidence. \n\nPrior to Tendermint v0.34.9, the light client could accept a bad header from its primary witness, and would not be able to form evidence of this deception, even if all the secondary witnesses were correct. Because the light client is responsible for verifying cross-chain state for IBC, a successful FLA could result in loss of funds. However, it is important to note that FLAs are only possible outside the Tendermint security model. \n\nAll FLAs, attempted and successful, leave traces of provable misbehavior on-chain. A faulty header contains signatures from the faulty validator, and even in unpatched versions of Tendermint Core, networks could use social consensus (off-chain action) to recover the network. The patches introduced in Tendermint Core v0.34.9 handle all evidence automatically and on-chain. \n\nNote that this fix also allows for successful automatic reporting of FLAs, even after a chain halt. By adding a time to FetchBlock, light clients effectively have a backup way to determine if a halted chain should have continued, and it will be able to submit evidence as soon as the chain resumes. \n\n### Patches\nThis problem has been patched in Tendermint Core v0.34.9. \n\n### Workarounds\nThere are no workarounds. All users are recommended to upgrade to Tendermint Core v0.34.9 at their earliest possible convenience. \n\n### Credits\n\nThank you to @MaximilianDiez for originally surfacing this issue, and to @cmwaters, @josef-widder, and @milosevic for creating fixes at both the implementation and specification level.\n\n### For more information\n\nIf you have any questions or comments about this advisory:\n* Open an issue in [tendermint/tendermint](https://github.com/tendermint/tendermint)\n* Email us at [security@tendermint.com](mailto:security@tendermint.com)",
  "id": "GHSA-f3w5-v9xx-rp8p",
  "modified": "2021-05-20T20:10:47Z",
  "published": "2021-12-20T18:17:41Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/tendermint/tendermint/security/advisories/GHSA-f3w5-v9xx-rp8p"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [],
  "summary": "Signature verification failure in Tendermint"
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Sightings

Author Source Type Date

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.


Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…