CVE-2026-22778 vLLM leaks a heap address when PIL throws an error
CriticalCVSS 9.8
EPSS 18.3%
Scanner template
- Published
- Updated
- Exploit status
- No public exploit known
- CISA SSVC
- No known exploitation · automatable: yes · technical impact: total
- Technology
- Red Hat
Description
vLLM is an inference and serving engine for large language models (LLMs). From 0.8.3 to before 0.14.1, when an invalid image is sent to vLLM's multimodal endpoint, PIL throws an error. vLLM returns this error to the client, leaking a heap address. With this leak, we reduce ASLR from 4 billion guesses to ~8 guesses. This vulnerability can be chained a heap overflow with JPEG2000 decoder in OpenCV/FFmpeg to achieve remote code execution. This vulnerability is fixed in 0.14.1.
Upgrade path
17 packages: 3 fixed, 1 partially fixed, 11 with no fix yet.
| Package | Affected | Fixed in | Status |
|---|---|---|---|
| redhat/rhaiis/vllm-cuda-rhel9:1772160593(Red Hat ai_inference_server 3.2 el9) | All versions | No fix listed | No fix yet |
| redhat/rhaiis/vllm-cuda-rhel9:1782352847(Red Hat ai_inference_server 3.3 el9) | All versions | No fix listed | No fix yet |
| redhat/rhaiis/vllm-rocm-rhel9:1772160625(Red Hat ai_inference_server 3.2 el9) | All versions | No fix listed | No fix yet |
| redhat/rhaiis/vllm-rocm-rhel9:1782353093(Red Hat ai_inference_server 3.3 el9) | All versions | No fix listed | No fix yet |
| redhat/rhaiis/vllm-spyre-rhel9:1782352919(Red Hat ai_inference_server 3.3 el9) | All versions | No fix listed | No fix yet |
| redhat/rhoai/odh-vllm-cpu-rhel9:1772093436(Red Hat openshift_ai 2.25 el9) | All versions | No fix listed | No fix yet |
| redhat/rhoai/odh-vllm-cpu-rhel9:1778264363(Red Hat openshift_ai 3.3 el9) | All versions | No fix listed | No fix yet |
| redhat/rhoai/odh-vllm-cuda-rhel9:1772093276(Red Hat openshift_ai 2.25 el9) | All versions | No fix listed | No fix yet |
| redhat/rhoai/odh-vllm-gaudi-rhel9:1770956034(Red Hat openshift_ai 3.3 el9) | All versions | No fix listed | No fix yet |
| redhat/rhoai/odh-vllm-gaudi-rhel9:1772093278(Red Hat openshift_ai 2.25 el9) | All versions | No fix listed | No fix yet |
| redhat/rhoai/odh-vllm-rocm-rhel9:1772093237(Red Hat openshift_ai 2.25 el9) | All versions | No fix listed | No fix yet |
| redhat openshift ai | < 1770956034; < 1772093237; < 1772093276; < 1772093278 +3 more | 1770956034, 1772093237, 1772093276, 1772093278, 1772093436, 1778264363 | Partially fixed |
| pypi vllm(PyPI) | = 0.8.3; = 0.8.4; = 0.8.5; = 0.8.5.post1 +15 more | 0.14.1 | Fixed |
| redhat ai inference server | < 1772160593; < 1772160625; < 1782352847; < 1782352919 +2 more | 1772160593, 1772160625, 1782352847, 1782352919, 1782353093 | Fixed |
| vllm | >= 0.8.3, < 0.14.1 | 0.14.1 | Fixed |
| redhat enterprise linux ai | = 3 | No fix listed | Versions listed |
| vllm-project vllm | = >= 0.8.3, < 0.14.1 | No fix listed | Versions listed |
Affected products
| Product | Affected versions | Fixed in |
|---|---|---|
| pypi vllm(PyPI) | >= 0.8.3, < 0.14.1 | 0.14.1 |
| pypi vllm(PyPI) | = 0.10.0 | No fix listed |
| pypi vllm(PyPI) | = 0.10.1 | No fix listed |
| pypi vllm(PyPI) | = 0.10.1.1 | No fix listed |
| pypi vllm(PyPI) | = 0.10.2 | No fix listed |
| pypi vllm(PyPI) | = 0.11.0 | No fix listed |
| pypi vllm(PyPI) | = 0.11.1 | No fix listed |
| pypi vllm(PyPI) | = 0.11.2 | No fix listed |
| pypi vllm(PyPI) | = 0.12.0 | No fix listed |
| pypi vllm(PyPI) | = 0.13.0 | No fix listed |
| pypi vllm(PyPI) | = 0.14.0 | No fix listed |
| pypi vllm(PyPI) | = 0.8.3 | No fix listed |
| pypi vllm(PyPI) | = 0.8.4 | No fix listed |
| pypi vllm(PyPI) | = 0.8.5 | No fix listed |
| pypi vllm(PyPI) | = 0.8.5.post1 | No fix listed |
| pypi vllm(PyPI) | = 0.9.0 | No fix listed |
| pypi vllm(PyPI) | = 0.9.0.1 | No fix listed |
| pypi vllm(PyPI) | = 0.9.1 | No fix listed |
| pypi vllm(PyPI) | = 0.9.2 | No fix listed |
| vllm-project vllm | = >= 0.8.3, < 0.14.1 | No fix listed |
| redhat ai inference server | < 1772160593 | 1772160593 |
| redhat ai inference server | < 1772160625 | 1772160625 |
| redhat ai inference server | < 1782352919 | 1782352919 |
| redhat ai inference server | < 1782353093 | 1782353093 |
| redhat ai inference server | < 1782352847 | 1782352847 |
Show 25 more
| Product | Affected versions | Fixed in |
|---|---|---|
| redhat openshift ai | < 1772093436 | 1772093436 |
| redhat openshift ai | < 1772093276 | 1772093276 |
| redhat openshift ai | < 1772093278 | 1772093278 |
| redhat openshift ai | < 1772093237 | 1772093237 |
| redhat openshift ai | < 1778264363 | 1778264363 |
| redhat openshift ai | < 1770956034 | 1770956034 |
| redhat ai inference server | = 3 | No fix listed |
| redhat enterprise linux ai | = 3 | No fix listed |
| redhat openshift ai | All versions | No fix listed |
| redhat/rhaiis/vllm-cuda-rhel9:1772160593(Red Hat:ai_inference_server:3.2::el9) | All versions | No fix listed |
| redhat/rhaiis/vllm-rocm-rhel9:1772160625(Red Hat:ai_inference_server:3.2::el9) | All versions | No fix listed |
| redhat/rhaiis/vllm-spyre-rhel9:1782352919(Red Hat:ai_inference_server:3.3::el9) | All versions | No fix listed |
| redhat/rhaiis/vllm-rocm-rhel9:1782353093(Red Hat:ai_inference_server:3.3::el9) | All versions | No fix listed |
| redhat/rhaiis/vllm-cuda-rhel9:1782352847(Red Hat:ai_inference_server:3.3::el9) | All versions | No fix listed |
| redhat/rhoai/odh-vllm-cpu-rhel9:1772093436(Red Hat:openshift_ai:2.25::el9) | All versions | No fix listed |
| redhat/rhoai/odh-vllm-cuda-rhel9:1772093276(Red Hat:openshift_ai:2.25::el9) | All versions | No fix listed |
| redhat/rhoai/odh-vllm-gaudi-rhel9:1772093278(Red Hat:openshift_ai:2.25::el9) | All versions | No fix listed |
| redhat/rhoai/odh-vllm-rocm-rhel9:1772093237(Red Hat:openshift_ai:2.25::el9) | All versions | No fix listed |
| redhat/rhoai/odh-vllm-cpu-rhel9:1778264363(Red Hat:openshift_ai:3.3::el9) | All versions | No fix listed |
| redhat/rhoai/odh-vllm-gaudi-rhel9:1770956034(Red Hat:openshift_ai:3.3::el9) | All versions | No fix listed |
| redhat/rhelai3/bootc-aws-cuda-rhel9(Red Hat:enterprise_linux_ai:3) | All versions | No fix listed |
| redhat/rhelai3/bootc-azure-cuda-rhel9(Red Hat:enterprise_linux_ai:3) | All versions | No fix listed |
| redhat/rhelai3/bootc-cuda-rhel9(Red Hat:enterprise_linux_ai:3) | All versions | No fix listed |
| redhat/rhelai3/bootc-gcp-cuda-rhel9(Red Hat:enterprise_linux_ai:3) | All versions | No fix listed |
| vllm | >= 0.8.3, < 0.14.1 | 0.14.1 |
References
- https://github.com/vllm-project/vllm/security/advisories/GHSA-4r2x-xpjr-7cvv vendor-advisory
- nvd.nist.gov third-party-advisory vendor-advisory
- https://github.com/vllm-project/vllm/pull/32319 issue-tracking patch
- https://github.com/vllm-project/vllm/pull/31987 issue-tracking patch
- security.access.redhat.com
- pypi.org product
- https://github.com/vllm-project/vllm/releases/tag/v0.14.1 release-notes
- github.com product
- github.com
- github.com vendor-advisory
- RHBZ#2436113 issue-tracking
- Red Hat CVE page mitigation vdb-entry vendor-advisory
- RHSA-2026:3782 patch vendor-advisory
- RHSA-2026:3713 patch vendor-advisory
- RHSA-2026:3462 patch vendor-advisory
- RHSA-2026:3461 patch vendor-advisory
- RHSA-2026:30089 patch vendor-advisory
- RHSA-2026:30088 patch vendor-advisory
- RHSA-2026:30087 patch vendor-advisory
- RHSA-2026:19712 patch vendor-advisory
- www.cve.org