The zerocopy path ultimately calls iov_iter_get_pages, which defines the step function for ITER_KVECs as simply, return -EFAULT. Taking the non-zerocopy path for ITER_KVECs avoids the unnecessary fallback.
See https://lore.kernel.org/lkml/20150401023311.gl29...@zeniv.linux.org.uk/T/#u for a discussion of why zerocopy for vmalloc data is not a good idea. Discovered while testing NBD traffic encrypted with ktls. Fixes: c46234ebb4d1 ("tls: RX path for ktls") Signed-off-by: Doron Roberts-Kedes <doro...@fb.com> --- net/tls/tls_sw.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index 4618f1c31137..ef15e35232dd 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -426,7 +426,8 @@ int tls_sw_sendmsg(struct sock *sk, struct msghdr *msg, size_t size) full_record = true; } - if (full_record || eor) { + bool iskvec = msg->msg_iter.type & ITER_KVEC; + if (!iskvec && (full_record || eor)) { ret = zerocopy_from_iter(sk, &msg->msg_iter, try_to_copy, &ctx->sg_plaintext_num_elem, &ctx->sg_plaintext_size, @@ -804,7 +805,8 @@ int tls_sw_recvmsg(struct sock *sk, page_count = iov_iter_npages(&msg->msg_iter, MAX_SKB_FRAGS); to_copy = rxm->full_len - tls_ctx->rx.overhead_size; - if (to_copy <= len && page_count < MAX_SKB_FRAGS && + bool iskvec = msg->msg_iter.type & ITER_KVEC; + if (!iskvec && to_copy <= len && page_count < MAX_SKB_FRAGS && likely(!(flags & MSG_PEEK))) { struct scatterlist sgin[MAX_SKB_FRAGS + 1]; int pages = 0; -- 2.17.1
