Spectre bug is back, this time to haunt JIT engines
Researchers disclose Branch Target Reuse, a Spectre v2 JIT attack that can leak Linux root password hashes.
Researchers from Vrije Universiteit and Scuola Superiore Sant'Anna describe Branch Target Reuse, an in-place Spectre v2 attack against JIT engines that reuses stale indirect branch-prediction entries after code-cache reuse. They demonstrated proofs of concept on an Intel Linux kernel that leak the root password hash despite cBPF constant binding, at about 5.7 KB/s on Raptor Cove and 5.4 KB/s on Lion Cove. Affected engines include Linux cBPF, Oracle GraalVM, and Mozilla SpiderMonkey. Linux and Oracle deployed mitigations and CVE-2026-64507 and CVE-2026-64508 were assigned; Mozilla is prioritizing site isolation instead of a direct fix. The paper was accepted to ACM CCS 2026.
- Branch Target Reuse reuses stale indirect-branch predictions in JIT code caches.
- Linux cBPF proofs of concept leaked root password hashes at about 5.5 KB/s.
- Engines named include Linux cBPF, Oracle GraalVM, and Mozilla SpiderMonkey.
- Linux and Oracle mitigated the issue; Mozilla is prioritizing site isolation.
- CVE-2026-64507 and CVE-2026-64508 were assigned; the paper is at CCS 2026.
- Organizations
- Vrije UniversiteitScuola Superiore Sant'Anna
- Countries
- NetherlandsItaly
Vulnerabilities mentioned
All| CVE | Vulnerability | CVSS | EPSS | Signals | Published |
|---|---|---|---|---|---|
CVE-2026-64507+1 related CVE | Linux kernel: x86/bugs: Enable IBPB flush on BPF JIT allocation Enable hardening against JIT spraying when Spectre-v2… | 5.5 | <1% | PoC |
From the original article
2 min read at theregister.comResearchers find a way to recover stale indirect branch prediction entries
By Thomas Claburn · theregister.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.