On Sun, Oct 18, 2026, at 3:22 PM, [email protected] wrote: > - [Medium] tls_sw_recvmsg() (net/tls/tls_sw.c) still has no bound on a > run of zero-length application_data records for any caller that has no…
The series bounds tls_sw_read_sock() because its consumer is never entered during a run of empty records and so cannot act. A recvmsg() caller runs in its own context and holds only its own socket's lock. The cover letter draws that line, and v6 will keep that behavior. A bound for an in-kernel sock_recvmsg() caller belongs in that caller. v6 will also drop the cap that v5 added to tls_sw_splice_read(), so the two system call paths behave the same way. -- Chuck Lever (Come to NFS bake-a-thon! https://nfsv4bat.org)
