SOLANA·HUB
DE
Security

A svmgov bug would have made SGP-0002 fail

According to Asymmetric Research, svmgov dropped override votes when the validator never voted. The fix reached mainnet on 3 August 2026. SGP-0002 stayed at 67.00 percent.

SOLANA·HUB Editorial
Source: Asymmetric Research

On 14 October 2026, Asymmetric Research described a bug in svmgov, Solana’s governance program. For the program’s first six weeks on mainnet, an override was left out of the tally when the validator never voted. Kai Parker wrote the post.

The fix was on mainnet before the August votes. Asymmetric Research dates that deployment to 3 August 2026. The SGP-0002 result did not change.

Asymmetric Research pointed to the post on X on 19 October.

Override votes without a validator ballot

Validators vote with the stake delegated to them. Delegators can replace that choice with an override. If the validator has already voted, the program moves the stake in the tally at once.

If the validator has not voted, the program stores the override in the VoteOverrideCache. The tally picks up that override only when the validator votes. Finalizing a proposal does not read the cache.

When the validator never votes, the override stays out of the total. Delegators still hold an on-chain receipt. The August votes ran on the fixed program.

How the fix landed

Asymmetric Research reported the bug to the Solana Foundation on 10 July 2026. GitHub shows pull request 110 merged on 30 July 2026. The v0.4.0 release notes from 2 August 2026 name the same fix.

The note says override votes were not counted if the validator never voted. Asymmetric Research dates the mainnet deployment to 3 August 2026. According to the same timeline, voting opened on 24 August 2026.

The post thanks Tim Garcia of the Solana Foundation and Tiago Carvalho for the fix.

The SGP-0002 count

SGP-0002 needed 66.67 percent. The vote reached 67.00 percent. That is 0.33 percentage points above the threshold.

SOLANA·HUB recorded the same 67.00 percent from the governance dashboard on 28 August. Without the fix, Asymmetric Research puts the result at 66.65 percent. That is 0.02 percentage points below the threshold.

On that figure, SGP-0002 would have failed. The 66.65 percent is the firm’s calculation for the unfixed program. The counted result is the one after the fix.

Across 102 validators that cast no vote of their own, delegators filed 212 overrides. Together those were 2.81 million SOL. Of that stake, 99.5 percent was yes.

Across the three August proposals, 8.45 million SOL in delegator votes would have been dropped. Asymmetric Research says SGP-0001 and SGP-0003 would still have ended as they did. The vote itself is in the on-chain governance report.

What to watch

  • Whether the Solana Foundation describes the bug itself

Sources

#svmgov #governance #security #sgp

Related Articles