Security readout for executives and security teams
Plain-English summary
A flaw in Linux’s Intel Xe graphics driver can trigger an invalid processor shift while clearing GPU address translations for an exceptionally large memory range. A local, low-privileged user may be able to reach the vulnerable path. The supplied CVSS rating is 7.8, but the sources do not demonstrate a working attack or confirmed real-world exploitation.
Executive priority
Treat this as a high-priority kernel maintenance issue on systems using Intel Xe graphics, especially shared compute hosts. It is less urgent for systems without the Xe driver or without untrusted local access. Patch through supported kernel channels; emergency incident response is not justified solely by the supplied evidence.
Technical view
The Xe driver passes an oversized TLB invalidation length to roundup_pow_of_two(), causing a shift exponent equal to the 64-bit word size. UBSAN observed this during xe_svm_invalidate() while releasing process memory. The kernel fix uses full TLB invalidation above a calculated range limit, avoiding the undefined shift.
Likely exposure
Exposure is limited to Linux systems using the Xe DRM driver and a vulnerable kernel build. The bundle lists version records including 6.8, 6.12.25, 6.14.4, and 6.15, but does not provide reliable complete range semantics. Distribution backports and the running kernel’s inclusion of the referenced commits must be checked.
Exploitation context
The CVSS vector describes a local, low-complexity attack requiring low privileges and no user interaction. The supplied record is not in KEV and contains no evidence of active exploitation. A UBSAN failure is documented, but practical privilege escalation, data compromise, and reliable denial of service are not demonstrated by these sources.
Researcher notes
The observed path involves SVM invalidation during memory teardown. The defect is an undefined 64-bit shift caused by power-of-two rounding of an extreme range length. The fix changes oversized requests to full TLB invalidation. The bundle does not establish attacker control sufficient for exploitation, complete affected-version boundaries, or a reproducible security impact beyond the UBSAN report.
Mitigation direction
Install a vendor-supported kernel containing the referenced stable fix, then reboot into it.
Check distribution guidance for corrected package versions and backports.
Prioritize shared systems where untrusted local users can access Intel Xe GPU functionality.
If updates are delayed, ask the vendor for supported temporary mitigations.
Validation and detection
Inventory running kernel versions and identify systems actively using the Xe driver.
Verify the installed kernel contains an applicable referenced stable commit or vendor backport.
Review kernel logs for shift-out-of-bounds reports involving log2.h or xe_gt_tlb_invalidation_range.
Confirm the updated kernel is active after reboot and exercise normal GPU workloads.
Assess whether untrusted local users or workloads can access affected GPU memory-management paths.
Generated from the cited source records. This long-tail analysis has not been individually reviewed by a named human.
Potential ATT&CK relevance
Conservative CVE-to-ATT&CK context
These mappings and lookup hints may be relevant to the vulnerability behavior, CWE, affected product, or exposure path. Glexia-inferred context is not an official MITRE, ATT&CK, CWE, or CVE Program mapping.
ATT&CK lookup starting points
Use these exact CWE pages and searches to review the Glexia ATT&CK library from this CVE's weakness and description context.
cve · low confidence lookup
CVE-2025-37761 mapping review
Open the CVE-to-ATT&CK bridge for reviewed, inferred, or future official mappings tied to this CVE.
These fields come from the CVE record and ADP containers, not from Glexia's Take. They preserve time-varying source decisions such as CISA SSVC, KEV status, CVSS metrics, and provider references.
1CVSS vectors
3Timeline events
0ADP providers
4Source links
CVSS vector scores
1 official score
We collect every scored CVSS vector available in the official CNA and ADP containers. When more than one version is present, the table keeps the source vectors side by side instead of collapsing them into the highest score.