sock: fix sg page frag coalescing in sk_alloc_sg

Current sg coalescing logic in sk_alloc_sg() (latter is used by tls and
sockmap) is not quite correct in that we do fetch the previous sg entry,
however the subsequent check whether the refilled page frag from the
socket is still the same as from the last entry with prior offset and
length matching the start of the current buffer is comparing always the
first sg list entry instead of the prior one.

Fixes: 3c4d755 ("tls: kernel TLS support")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Dave Watson <davejwatson@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/core/sock.c

@@ -2277,9 +2277,9 @@ int sk_alloc_sg(struct sock *sk, int len, struct scatterlist *sg,
 		pfrag->offset += use;
 
 		sge = sg + sg_curr - 1;
-		if (sg_curr > first_coalesce && sg_page(sg) == pfrag->page &&
-		    sg->offset + sg->length == orig_offset) {
-			sg->length += use;
+		if (sg_curr > first_coalesce && sg_page(sge) == pfrag->page &&
+		    sge->offset + sge->length == orig_offset) {
+			sge->length += use;
 		} else {
 			sge = sg + sg_curr;
 			sg_unmark_end(sge);