Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
›

Xen flaws let a guest VM read and write other VMs' memory

The Xen Project disclosed two hypervisor flaws on November 3 that let a malicious guest virtual machine read and write the memory of other VMs on the same host, including dom0. Anthropic's Claude model found both.

TechnologyExplainerRachel NwosuPublished: 4 November 20263 min readSources 28
Xen flaws let a guest VM read and write other VMs' memory

The Xen Project published advisory XSA-523 on November 3, disclosing two hypervisor flaws that let a malicious guest virtual machine read and write the memory of other VMs on the same host, including dom0. The flaws are tracked as CVE-2026-102703 and CVE-2026-102704. Patches are out, and there is no known mitigation, the project said.

All Xen versions from 4.6 onward are vulnerable. Versions 4.5 and earlier are not. Dom0 is the VM that by default has full permissions and manages the others.

How the flaws work

The flaws sit in grant copy, the operation Xen uses to copy data between VMs. Each VM tells Xen which of its memory pages other VMs may use, and Xen checks those permissions. To save time on a batch of copies within the same memory page, Xen reuses checks it has already made. In some cases it did not check again when it should have. That mistake lets a malicious guest read and write memory it does not own.

XCP-ng, an open-source virtualization platform built around Xen, said in its update notes that "a single compromised guest can break out of its isolation and take over the whole host." Vates, which backs XCP-ng, rates the impact as Critical.

Claude found the bugs

The Xen Project credited AI company Anthropic with finding the flaws using its Claude model. Stephan Lipp of Ada Logics validated the finding. The CodeAnt AI Security Research Team later independently reported the same issue.

Fixes and restarts by platform

The advisory has patches for xen-unstable to Xen 4.19.x, and for Xen 4.18.x to 4.17.x. It carries no patch for Xen 4.6 to 4.16. Xen 4.16's security support ended on December 2, 2024.

XCP-ng 8.3 has a fix: xen-4.17.7-4.2.xcpng8.3, released November 3. Hosts must be rebooted.

Qubes OS, a security-focused desktop operating system that keeps software in separate Xen-based compartments called qubes, said every supported release and configuration is affected. Its fix is for Qubes 4.3, the only release it still supports. The Xen packages are 4.19.6-3 in dom0. Dom0 must be restarted, and Anti Evil Maid users must reseal their secret passphrase because the new Xen binaries change PCR18 and PCR19.

Because every Xen version from 4.6 on is vulnerable, other products built on those versions are affected until their makers add the XSA-523 patches. As of November 4, Debian's security tracker listed the xen package in its bookworm, trixie, and forky releases as vulnerable, and in sid as fixed.

What happens next

Qubes told users on November 3 to keep installing updates as normal. The fixed packages were then in its security-testing repository, due to move to the stable one over the next two weeks after community testing. The Xen advisory does not mention any attacks that exploit the flaws or specify what access within the VM an attacker needs. XSA-523 is separate from XSA-500, an earlier grant copy flaw that Xen disclosed on July 28, and it has its own patches.

Comments 0

Sources

28
  1. 01Xen Hypervisor Flaws Let One Malicious VM Read and Write Any Other VM's MemoryEN
  2. 02Cyber security flaws may put about 70% of websites at riskEN
  3. 0328 Zero-Day Findings in Less Than Two Days: What AI-Speed Discovery Means for RoboticsEN
  4. 04Gemini for Home now plugs into AT&T’s smart security systemEN
  5. 05Barracuda: Root-Sicherheitslücke bedroht unter anderem Email Security GatewayEN
  6. 06SonicWall Won't Patch EoL GMS Flaws That Could Expose Files, Alter Database RecordsEN
  7. 07SonicWall warns of unpatched flaws in end-of-life GMS softwareEN
  8. 08Adding Network Security to AI, HPC and Data CentersEN
  9. 09AI helps Stanford scientists discover a “natural Ozempic” with a surprising advantageEN
  10. 10All 5 genetic “letters” of DNA and RNA discovered in a 4.5-billion-year-old asteroidEN
  11. 11Trace Gas Orbiter discovers potential evidence of volcanic activity on Mars via atmospheric observationsEN
  12. 12Debian and Ubuntu APT Flaw Lets MitM Attackers Swap Verified Packages and Run Code as RootEN
  13. 13Security updates for WednesdayEN
  14. 14Critical Rails Flaw Lets Attackers Write Server Files via Crafted ImageMagick UploadsEN
  15. 15CISA flags WordPress RCE flaw as actively exploited in attacksEN
  16. 16Radio With No Radar: Why Agent Runtime Security Fails Without the Layers Beneath ItEN
  17. 17Docker Urges Isolation and Governance for AI Agent SecurityEN
  18. 18Google Issues Patches for Critical Android Flaw That Could Allow RCE Without User InteractionEN
  19. 19CISA Confirms Hackers Exploiting WordPress Core Flaw During Software UpgradesEN
  20. 20Scientists discover new Alzheimer’s drug that repairs DNA damage and reduces brain inflammationEN
  21. 21OpenAI hires former U.S. Army Secretary Dan Driscoll for national security roleEN
  22. 22 Nvidia Brev security breach may have exposed SSH keys used to access customer VMs EN
  23. 23Scientists discover that some “unhealthy” diets can boost cancer immunotherapy in miceEN
  24. 24NASA's Hubble and Webb discover ancient worlds that remember their pastEN
  25. 25Security updates for TuesdayEN
  26. 26Red Hat Patches 9.8-Rated Linux TCP RCE Flaw While Ubuntu Releases Fixes for TestingEN
  27. 27Attackers Exploit Critical Siemens IoT2050 Flaw to Plant Remote Command BackdoorsEN
  28. 28Attackers Target Unpatched CryptoPayment Gateway Flaw That Can Expose Wallet CredentialsEN

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Rachel Nwosu

Rachel Nwosu

AI, models and technology

Rachel Nwosu covers AI, models and technology for FLASH24, working from public model documentation, benchmark releases and repository histories rather than press summaries, and she skips announcements that arrive without reproducible numbers. She checks training-data claims against dataset cards and reruns reported metrics where code is available. She spends much of her week interviewing researchers and engineers, tracking model launch calendars, and comparing vendor benchmarks with independent evaluations. Outside the desk she runs 3D printers, restores old computers, and tests how models learn from internet junk. She does not publish benchmark figures she cannot trace to a source.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.