Security Advisory

CVE-2024-53104: Linux uvcvideo Out-of-Bounds Write

Last updated:
  • CVE-2024-53104

CVE-2024-53104 is an out-of-bounds write (CWE-787) in the Linux kernel's uvcvideo USB camera driver, triggered by frames of type UVC_VS_UNDEFINED in a device's descriptors. NVD scores it 7.8. CISA lists it as exploited, and Google reports limited, targeted exploitation on Android. Fixed in 6.1.117, 5.10.230 and the other stable releases.

Vulnerability summary

CVSS base score

7.8

High severity

EPSS score (FIRST)

3.4%

Probability of exploitation in the wild in the next 30 days

CISA KEV catalog

Yes

Listed as known exploited

Published

Vulnerability ID: CVE-2024-53104 · Source: NVD

Affected components

ComponentCPEVersion range
Linux kernel (uvcvideo, CONFIG_USB_VIDEO_CLASS)cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*2.6.26 to before 4.19.324, 4.20 to before 5.4.286, 5.5 to before 5.10.230, 5.11 to before 5.15.172, 5.16 to before 6.1.117, 6.2 to before 6.6.61, 6.7 to before 6.11.8, 6.12 to before 6.12.1 (NVD); unaffected from 6.13 (kernel CNA record)
Debian 11 (bullseye) linuxcpe:2.3:o:debian:debian_linux:11.0:*:*:*:*:*:*:*Fixed in 5.10.234-1 (DLA-4075-1); linux-6.1 fixed in 6.1.119-1~deb11u1 (DLA-4008-1); bookworm fixed in 6.1.119-1 (Debian security tracker)
Android (kernel, UVC subcomponent)—Fixed at security patch level 2025-02-05 or later (Android Security Bulletin February 2025, A-378455392)

What does this mean for affected devices?

CVE-2024-53104 is an out-of-bounds write (CWE-787) in uvcvideo, the Linux kernel's USB Video Class driver for webcams and other USB cameras. The kernel CNA's record is titled "media: uvcvideo: Skip parsing frames of type UVC_VS_UNDEFINED in uvc_parse_format" and explains that such frames "were not taken into account when calculating the size of the frames buffer in uvc_parse_streaming", which "can lead to out of bounds writes". In practice the input is the descriptor data a USB video device presents when it is plugged in. NVD's Primary score is 7.8, CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, with CWE-787. The Linux kernel CNA publishes no CVSS score of its own; CISA's ADP Vulnrichment container gives the same 7.8 vector (NVD, CVE record).

Which kernel versions are affected?

The bug dates from Linux 2.6.26. The CVE record lists the first unaffected release on each stable branch: 4.19.324, 5.4.286, 5.10.230, 5.15.172, 6.1.117, 6.6.61, 6.11.8, 6.12.1, and 6.13 onward, where mainline commit ecf2b43018da landed. NVD's CPE ranges match. The kernel CNA's dyad file gives the fix commit for each branch, which is the quickest way to check whether a vendor board-support kernel already carries the backport.

The only source file the fix touched is drivers/media/usb/uvc/uvc_driver.c. The uvc Makefile builds it into uvcvideo through the line obj-$(CONFIG_USB_VIDEO_CLASS) += uvcvideo.o, so a firmware image built without CONFIG_USB_VIDEO_CLASS does not contain the vulnerable code. Our Linux kernel CVE triage guide uses this CVE as its worked example of mapping a kernel CVE to a CONFIG_* option and recording "not affected" with the .config line as evidence.

Which devices are affected?

Any Linux or Android device that ships the uvcvideo driver on an unfixed kernel and has a USB port an attacker can reach: phones and tablets, camera-equipped kiosks, HMIs, video gateways and single-board computers. Google's Android Security Bulletin for February 2025, published 3 February 2025, lists CVE-2024-53104 in the Kernel section under the 2025-02-05 security patch level as an elevation of privilege rated High, subcomponent UVC (A-378455392); Android devices at patch level 2025-02-05 or later are fixed. Debian, per its security tracker, fixed the linux package in 5.10.234-1 for bullseye (DLA-4075-1), 6.1.119-1 for bookworm and 6.11.9-1 in unstable, and linux-6.1 for bullseye in 6.1.119-1~deb11u1 (DLA-4008-1). Distribution kernels carry backports, so compare package releases rather than the upstream version string.

How do you fix CVE-2024-53104?

Move the product kernel to at least the fixed release of its stable branch, for example 5.10.230, 5.15.172, 6.1.117 or 6.6.61, or to the latest release of that branch. Where the kernel cannot be rebased, backport the branch's fix commit from the dyad file. If the product does not use USB cameras, build without CONFIG_USB_VIDEO_CLASS or remove the uvcvideo module, and block or disable USB host ports that customers do not need.

Is CVE-2024-53104 exploited?

Yes. CISA added it to the KEV catalog on 2025-02-05 with a due date of 2025-02-26, describing an out-of-bounds write in the UVC driver "that could allow for physical escalation of privilege"; known ransomware campaign use is listed as Unknown, as of 2026-11-19 (catalog version 2026.11.18). Google's February 2025 bulletin states "There are indications that CVE-2024-53104 may be under limited, targeted exploitation." FIRST EPSS gives it 0.03395, the 89.1st percentile, in the 2026-11-18 dataset (CISA KEV, FIRST EPSS). Related kernel CVEs from the same triage example that are also KEV-listed include CVE-2025-39964 in af_alg and CVE-2025-39682 in kernel TLS.

How to fix this class of issueThis advisory's weakness class is Memory corruption. Our practical guide covers how to detect, fix, verify and disclose it: How to fix memory corruption in embedded firmware.

Evidence

  • NVD: status Analyzed; [email protected] Primary CVSS 3.1 7.8 AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H; CISA-ADP Secondary 7.8 same vector; CWE-787; CPE linux_kernel 2.6.26 to before 4.19.324, 4.20-5.4.286, 5.5-5.10.230, 5.11-5.15.172, 5.16-6.1.117, 6.2-6.6.61, 6.7-6.11.8, 6.12-6.12.1, plus debian_linux 11.0; published 2024-12-02; read from cached API response 2026-11-19. — NVD
  • CVE record (CNA Linux, PUBLISHED 2024-12-02): title 'media: uvcvideo: Skip parsing frames of type UVC_VS_UNDEFINED in uvc_parse_format'; programFiles drivers/media/usb/uvc/uvc_driver.c; affected from 2.6.26; unaffected 4.19.324, 5.4.286, 5.10.230, 5.15.172, 6.1.117, 6.6.61, 6.11.8, 6.12.1, 6.13; no CNA CVSS score; CISA ADP Vulnrichment 7.8, CWE-787. — CVE.org
  • Kernel CNA dyad file: one vulnerable:fixed pair per branch from 2.6.26 (c0efd232929c), e.g. 6.1.117 beced2cb09b5, 5.10.230 faff5bbb2762, mainline 6.13 ecf2b43018da. — Linux kernel vulns.git
  • drivers/media/usb/uvc/Makefile: uvcvideo-objs includes uvc_driver.o; obj-$(CONFIG_USB_VIDEO_CLASS) += uvcvideo.o. — Linux kernel source
  • Android Security Bulletin February 2025 (published 2025-02-03): 'There are indications that CVE-2024-53104 may be under limited, targeted exploitation'; listed under Kernel, 2025-02-05 patch level, A-378455392, EoP, High, subcomponent UVC; patch levels 2025-02-05 or later address all issues. — Android Security Bulletin
  • Debian: linux fixed bullseye 5.10.234-1 (DLA-4075-1), bookworm 6.1.119-1, unstable 6.11.9-1; linux-6.1 bullseye 6.1.119-1~deb11u1 (DLA-4008-1). — Debian security tracker
  • CISA KEV catalog version 2026.11.18: Linux Kernel Out-of-Bounds Write Vulnerability, dateAdded 2025-02-05, dueDate 2025-02-26, knownRansomwareCampaignUse Unknown, 'could allow for physical escalation of privilege'; checked 2026-11-19. — CISA KEV
  • EPSS 0.03395, percentile 0.89071, dataset 2026-11-18, read 2026-11-19. — FIRST EPSS

Related advisories

Sources

  1. NVD
  2. CVE.org record CVE-2024-53104 (CNA: Linux)
  3. Android Security Bulletin: February 2025
  4. Debian security tracker: CVE-2024-53104
  5. Linux kernel vulns.git: CVE-2024-53104.dyad (vulnerable and fixed commits per branch)
  6. Linux kernel source: drivers/media/usb/uvc/Makefile
  7. Linux kernel mainline fix commit ecf2b43018da
  8. CISA KEV
  9. FIRST EPSS

Frequently asked questions

Which kernel versions fix CVE-2024-53104?

4.19.324, 5.4.286, 5.10.230, 5.15.172, 6.1.117, 6.6.61, 6.11.8, 6.12.1 and 6.13 onward; the bug dates from 2.6.26. Each stable branch has its own fixed release, so a 6.1.120 kernel is fixed (per the kernel CNA's CVE record and dyad file).

Is my firmware affected if it has no USB camera support?

Not if the kernel is built without CONFIG_USB_VIDEO_CLASS. The only file the fix touched, uvc_driver.c, is built into the uvcvideo driver only when that option is set (per the CVE record's programFiles and the drivers/media/usb/uvc/Makefile).

Which Android security patch level fixes CVE-2024-53104?

Security patch level 2025-02-05 or later. The bulletin lists it as a High-severity elevation of privilege in the kernel's UVC subcomponent and notes indications of limited, targeted exploitation (per the Android Security Bulletin for February 2025).

Is CVE-2024-53104 exploited in the wild?

Yes. CISA added it to the KEV catalog on 2025-02-05, with ransomware use listed as Unknown, as of 2026-11-19 (catalog version 2026.11.18). FIRST EPSS gives it 0.03395, the 89.1st percentile, in the 2026-11-18 dataset (per CISA KEV and FIRST EPSS).