← All calls
ARBORIST#122

Zebra and ZcashD security patches in quick succession

Zebra 4.3 fixed two CVEs, and ZcashD also patched a Sprout proof vulnerability.

Agenda summary 10

  1. Zebra update

    Zcash Foundation · Marek

    • Last week the team released Zebra 4.3, which fixed two CVEs.
    • One could cause a chain split on vulnerable Zebra nodes, and the other could bring down a vulnerable node.
    • This release includes the ZIP 235 implementation and a fix so that blocks propagate correctly on regtest.
    • It also fixed the pre-Canopy founders reward calculation in the getblocksubsidy RPC and an issue where one CPU thread was pinned at 100% on testnet.
    • It confirmed that testnet sync is slow and work to improve it is underway.
    • This also applies to mainnet and should make it faster in the next release.
  2. ZODL update

    ZODL · Kris Nuttycombe

    • Str4d went on a three-month sabbatical starting yesterday.
    • For the Zaino chain index integration work Str4d had been doing, Kris is taking over a branch that applies the Zcash primitives 0.27 release to Zaino.
    • Zingo Labs' Zaino PR 853 also needs to be merged, and the team asked for community help.
    • The full block scanning change on top of Zallet is nearly done, and importing P2SH addresses, needed for zcashd wallet compatibility, is now supported.
    • The team plans to bring up past zcashd versions in Docker and create regtest wallet files containing various transactions from the Sprout, Overwinter and Sapling eras, to use in Zallet integration tests.
    • It finished a revision 2 draft of ZIP 316, the Unified Address standard.
    • It fixed Unified viewing keys to express P2SH descriptors and introduced the ZU prefix, which has no transparent element, and the TU prefix, which allows transparent elements and address expiry metadata.
    • It asked for feedback.
    • ZIP 248, the extensible transaction format to be included in NU 7, is also receiving comments on an open PR in the zips repository.
  3. zcashd vulnerability fix and quantum computing discussion

    ZODL, Zcash Foundation · Kris Nuttycombe, Daira Emma, Zooko, Arya

    • ZODL fixed a zcashd vulnerability that let a malicious miner put invalid Sprout proofs on the chain, and released zcashd 6.12.
    • Shielded Labs coordinated the disclosure, and the patch was distributed to major mining pools before disclosure, which mitigated it.
    • The patch was mainly written by Str4d and reviewed by Daira Emma and Kris.
    • Zooko said that Zebra was not vulnerable to that bug, which helped reason about the extent of exploitation, and that Zebra is safer than zcashd.
    • The AI-assisted security researcher who reported the vulnerability remarked that Zebra is well hardened, and Bootstrap, Shielded Labs, ZODL and ZF donated 50 ZEC as a reward.
    • Arya disclosed that Zebra also had a bug where the auth digest of v5 transactions was not checked, so the proof fields could be tampered with.
    • Zooko compared them, saying the Sprout bug could have stolen about 25,000 ZEC, while the Zebra bug had a much lower impact.
    • Daira Emma mentioned Google's post-quantum timeline paper and said that a fully post-quantum Zcash design and the deployment of quantum recoverability must be hastened.
    • They are writing a blog post saying the paper omitted an attack that needs only a single discrete log computation.
  4. Zaino update

    Zingo Labs · Hazel

    • Since the last call the team has released the library and zainod, and aims to ship the next release more regularly.
    • One method that Zallet needs is still on a feature branch on top of the release and needs to be merged.
    • It is upgrading to Zebra 4.3, which has improvements such as pre-Canopy indexing, and for a while will focus on paying down technical debt and improving pass-through documentation.
    • RPC work for block explorers also remains, and the next release is expected in one to two weeks.
  5. zcashd deprecation update

    Electric Coin Company, Zcash Foundation, Zingo Labs · Pacu, Zooko, Daira Emma

    • The team is coordinating in response to Str4d's sabbatical and is building integration tests in the light client working group that compare the gRPC API parity of lightwalletd and Zaino.
    • Zaino released 0.20, and the next, 0.21, will include Zebra 4.3 support and a fix to make the get block range gRPC API, which caused the nonlinear sync problem, match lightwalletd.
    • Daira Emma said that since this bug came from Sprout code written in C++ that had not been touched since 2020, it is important to remove attack surface from the protocol.
    • Pacu said the team is aware of the community's concerns about deprecating shielded pools and welcomes discussion and ZIP proposals.
    • Daira Emma said they would write up ideas on how to support, in the long term, a long-term storage pool that uses more conservative cryptography.
  6. ZSA update

    Qedit · Vivek, Daira Emma

    • The team answered all the reviews on Orchard ZSA PR 471 and merged upstream into the branch.
    • It improved issuance handling and added the visibility needed in Zebra, and hopes that it is nearly done and the review continues.
    • It opened a PR updating sapling-crypto for the zcash_note_encryption interface, which changed to encrypt notes of multiple sizes.
    • Daira Emma said to merge it to unblock work, but that it falls short for the memo bundle and needs further refactoring.
    • A librustzcash upstream PR 1931 is prepared and is next in line for review once the Orchard review is finished.
    • Zebra has nearly finished issued-asset state management and is internally reviewing sync as of 4.2.0.
    • Daira Emma said ZIP 2005 already takes ZSA into account, and that its interaction with ZIP 248 (the extensible format) and the memo bundle ZIP 231 needs to be checked.
    • The ZSA ZIP itself has not changed.
  7. Network Sustainability Mechanism (NSM) update

    Shielded Labs · Judah

    • ZIP 235 was merged and released behind a configuration flag (a compile-time flag).
    • What remains are ZIP 234 and the refactoring PR.
    • The 234 PR is blocked by the refactoring PR, so a review was requested.
  8. Crosslink update (Trailing Finality Layer)

    Shielded Labs, ZODL · Mark, Daira Emma, Zooko

    • The first incentivized testnet starts with the April 15 workshop.
    • The team is polishing stability, especially the new networking component for sync, collaborating with ZF, and also receiving feedback from the Tachyon team.
    • The ZIPs for Crosslink started at the stub level, and proper stubs will be submitted around the workshop.
    • When Daira Emma asked how many ZIPs there are, Mark answered at least three, and that it will not be one monolithic ZIP.
    • Daira Emma said that since the phased deployment path is clear, the polling questions should be organized by deployment phase, such as phase 1 and phase 2, rather than by ZIP boundaries.
  9. Dynamic fees update

    Shielded Labs · Mark, Daira Emma, Arya

    • The team withdrew the preliminary ZIP it had submitted for NU 7 polling and restarted the design and exploration of dynamic fees with a UX focus.
    • Since it was not adopted into NU 7 in the coinholder vote, there is time to explore.
    • It started wallet outreach with the Unstoppable team and plans to also hear about the fee experience from ZODL, Zingo, Edge and others.
    • Under the principle that mainnet data matters most for fees, it is looking for ways to obtain mainnet fee data.
    • Mark asked about lowering the marginal fee from 5000 to 1000, considering recent price changes and stability.
    • Daira Emma said it is too large a change, and that it would be a problem without a quick way to roll back if sandblasting recurs.
    • Arya said one should look at fees relative to the chain's total value, not only fiat-based fees.
    • Daira Emma said both extremes can be attack vectors: fees getting too high, and fees getting so low that they become dust.
  10. Open discussion: DAG BFT and Google's quantum computing paper

    Arya, Daira Emma, Mark, Zooko

    • When Arya asked whether DAG BFT such as Avalanche had been considered for Crosslink, Daira Emma said they envision keeping modularity so that an arbitrary BFT can be plugged into a gadget that gives better security properties.
    • Mark said Avalanche does not have the finality of something like Tendermint.
    • Daira Emma pointed out the two key points of Google's quantum computer paper (a reduction in the gate count of the discrete log quantum circuit and an analysis of quantum attacks on each cryptocurrency).
    • The paper analyzed Zcash in depth, but they rebutted its claim that Halo 2 is better because it has no trusted setup.
    • They said that computing just one discrete log breaks the proof system, the value commitments and the note commitments alike, so whether the reference string is structured or unstructured does not matter.
    • They will submit a correction to the paper and explain it in a blog post.

This summary is based on the original meeting notes; check them for exact details.