CVE-2025-39682

In the Linux kernel, the following vulnerability has been resolved:

tls: fix handling of zero-length records on the rx_list

Each recvmsg() call must process either

• only contiguous DATA records (any number of them)
• one non-DATA record

If the next record has different type than what has already been processed we break out of the main processing loop. If the record has already been decrypted (which may be the case for TLS 1.3 where we dont know type until decryption) we queue the pending record to 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 dont have an skb to queue (darg.skb points to the ciphertext skb for access to metadata like length).

Only data records are allowed zero-copy, and we break the processing loop after each non-data record. So we should never zero-copy and then find out that the record type has changed. The corner case we missed is when the initial record comes from rx_list, and its zero length.

Affected Software

References

• https://git.kernel.org/stable/c/2902c3ebcca52ca845c03182000e8d71d3a5196f
• https://git.kernel.org/stable/c/29c0ce3c8cdb6dc5d61139c937f34cb888a6f42e
• https://git.kernel.org/stable/c/3439c15ae91a517cf3c650ea15a8987699416ad9
• https://git.kernel.org/stable/c/62708b9452f8eb77513115b17c4f8d1a22ebf843
• https://git.kernel.org/stable/c/c09dd3773b5950e9cfb6c9b9a5f6e36d06c62677