Skip to content
NewA calmer, faster ZeroHour

New Spectre v2 Variant Exposes Intel, AMD, Arm CPUs to Data Leaks

SecurityWeekby Eduard Kovacs ()4 min read1
Part of a story covered by 3 sourcesBranch Target Reuse: New Spectre-v2 Variant Leaks Linux Root Password Hashes From Fully Patched Intel, AMD, and Arm SystemsMerged summary & timeline
AI summary · glm-5.3Summaries by Vani AI

New Spectre-v2 variant BTR lets attackers leak arbitrary memory, including root password hashes, from Intel, AMD, and Arm systems via stale JIT branch predictions.

VUSec and Scuola Superiore Sant'Anna researchers disclosed Branch Target Reuse (BTR), a Spectre-v2 variant exploiting stale branch predictor entries left by self-modifying JIT code. End-to-end exploits against the Linux kernel's unprivileged cBPF JIT leak arbitrary memory at 8 bytes per second, recovering the root password hash on fully updated Intel systems with default mitigations enabled. Linux shipped an IBPB-based x86 mitigation, Oracle partially mitigated GraalVM via code-cache randomization, Mozilla is prioritizing site isolation, and AMD says existing Spectre-v2 guidance applies.

Key facts
  • Affects Intel, AMD, and Arm CPUs; no current hardware keeps branch predictor and code in sync.
  • cBPF kernel exploits leak memory at 8 bytes/second, defeating default mitigations.
  • Firefox exposed until site isolation completes; other tabs may share attacker address space.
  • Linux mitigation issues IBPB when cBPF reuses previously used memory regions.
Analyze this story with Vani AIImpact, who is affected and what defenders should do, in a new chat

From the original article

4 min read at securityweek.com

Branch Target Reuse (BTR) is a new Spectre v2 attack targeting JIT compilers in web browsers, language runtimes, and the operating system kernel The post New Spectre v2 Variant Exposes Intel, AMD, Arm CPUs to Data Leaks appeared first on SecurityWeek.

SecurityWeek

By Eduard Kovacs · securityweek.com

Read the full article at SecurityWeek

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.