1.7 KiB
CVE-2025-39682
Description
In the Linux kernel, the following vulnerability has been resolved:tls: fix handling of zero-length records on the rx_listEach recvmsg() call must process either - only contiguous DATA records (any number of them) - one non-DATA recordIf the next record has different type than what has already beenprocessed we break out of the main processing loop. If the recordhas already been decrypted (which may be the case for TLS 1.3 wherewe don't know type until decryption) we queue the pending recordto the rx_list. Next recvmsg() will pick it up from there.Queuing the skb to rx_list after zero-copy decrypt is not possible,since in that case we decrypted directly to the user space buffer,and we don't have an skb to queue (darg.skb points to the ciphertextskb for access to metadata like length).Only data records are allowed zero-copy, and we break the processingloop after each non-data record. So we should never zero-copy andthen find out that the record type has changed. The corner casewe missed is when the initial record comes from rx_list, and it'szero length.
POC
Reference
No PoCs from references.