Linux Kernel CVE-2026-53321

The io_uring subsystem's NAPI (New API) integration exposes a busy_poll_to configuration option - a timeout (in microseconds) governing how long the kernel's NAPI polling loop will spin waiting for network events before yielding. Prior to this fix, no upper bound was enforced on this value. Critically, the polling loop contains no conditional rescheduling (cond_resched() or equivalent), meaning an arbitrarily large busy_poll_to with no arriving events causes the kernel thread to monopolize the CPU until the kernel's soft lockup detector fires (typically at 10-20 seconds). Although no CWE is formally assigned, the root cause maps to CWE-400 (Uncontrolled Resource Consumption) / CWE-835 (Loop with Unreachable Exit Condition). Affected CPE: cpe:2.3:a:linux:linux. Note: the intelligence tag 'Information Disclosure' appears to be a metadata error - the CVSS vector (C:N/I:N/A:H) and CVE description confirm exclusively an availability impact with no confidentiality or integrity implications.