GPU VulnDB

Database/Kernel, userspace & hypervisor

Linux kernel scsi/ses: out-of-bounds desc_ptr accesses while parsing enclosure data

CVSS 7.1CVE-2023-53675Kernel, userspace & hypervisorcurated

Impact

ses_enclosure_data_process() could walk desc_ptr past the end of the buffer returned by a SCSI enclosure, giving out-of-bounds reads and possible memory corruption in kernel context. This touches the enclosure-services path that storage shelves and JBOD backplanes in dense GPU and HPC racks expose, so a machine with no SES-capable enclosure never runs the code. Where it does run, the data being parsed comes from the enclosure firmware rather than from a tenant, so the practical exposure is a local user or a compromised enclosure provoking a kernel crash; the record gives no evidence of a remote or cross-tenant path.

Who can reach it

Local, per the CVSS vector (AV:L, PR:L) - a user able to drive SES queries on a host with an attached SCSI enclosure. Not reachable from a GPU pod on a node without enclosure hardware.

What to do

Pick up a stable kernel containing the fix (commits listed in the record, e.g. 414418abc19f) from your distribution and reboot each affected storage or head node. The fix is kernel-resident, so there is no hot-patch path short of a vendor livepatch; drain and reboot in your normal kernel maintenance window rather than scheduling an emergency one, given the local-only exposure.

References

Related entries

All Kernel, userspace & hypervisor entries

This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.