CVE-2026-31720: USB audio module trusts host length

You should be aware that camera firmware version 1.0.0 contains a stack overflow in its USB audio gadget module. Version 1.0.1 validates control-request lengths, but the original modules remain independently compiled and cannot be assumed to include the upstream fix.

CVE-2026-31720: USB audio module trusts host length

Key points

  • The vulnerable f_audio_complete function copies the host-controlled request length into a four-byte stack variable, allowing an oversized payload to overwrite the return address.
  • The researcher demonstrated the flaw by executing the extracted instructions in QEMU user mode, where 32 bytes of 0x41 overwrote the return address with 0x41414140.
  • Version 1.0.1 accepts only one-byte Mute and two-byte Volume data and stalls invalid requests; upstream addressed the issue in commit 6e0e34d85cd46ceb37d16054e97a373a32770f6c.

Context

For developers maintaining embedded USB audio gadgets, this case shows why vendor-specific kernel modules require their own review even after a defect is fixed upstream. The full-device exploit path was not reached because enumeration failed with error -32 at a vendor-specific sound-card structure offset, so the demonstrated overflow came from isolated module code.

What this means for you

Directly applicable

If you maintain embedded USB audio modules or device firmware, check whether your compiled artifacts contain the length-validation fix rather than assuming that an upstream kernel patch also applies to vendor code.

Who benefits: Developers and technical decision-makers responsible for legacy embedded firmware can use this case to prioritize module-level review, reproducible validation, and patch verification in their own products.

Share article:

Does a report affect your system?

Send the link to the report and two sentences about your system. You get an assessment of whether action is needed and what the fix looks like.