Fix harmless bug in rioConnRead (#7557)

this code is in use only if the master is disk-based, and the replica is
diskless. In this case we use a buffered reader, but we must avoid reading
past the rdb file, into the command stream. which Luckly rdb.c doesn't
really attempt to do (it knows how much it should read).

When rioConnRead detects that the extra buffering attempt reaches beyond
the read limit it should read less, but if the caller actually requested
more, then it should return with an error rather than a short read. the
bug would have resulted in short read.

in order to fix it, the code must consider the real requested size, and
not the extra buffering size.

src/rio.c

@@ -190,15 +190,18 @@ static size_t rioConnRead(rio *r, void *buf, size_t len) {
     /* If we don't already have all the data in the sds, read more */
     while (len > sdslen(r->io.conn.buf) - r->io.conn.pos) {
         size_t buffered = sdslen(r->io.conn.buf) - r->io.conn.pos;
-        size_t toread = len - buffered;
+        size_t needs = len - buffered;
         /* Read either what's missing, or PROTO_IOBUF_LEN, the bigger of
          * the two. */
-        if (toread < PROTO_IOBUF_LEN) toread = PROTO_IOBUF_LEN;
+        size_t toread = needs < PROTO_IOBUF_LEN ? PROTO_IOBUF_LEN: needs;
         if (toread > sdsavail(r->io.conn.buf)) toread = sdsavail(r->io.conn.buf);
         if (r->io.conn.read_limit != 0 &&
             r->io.conn.read_so_far + buffered + toread > r->io.conn.read_limit)
         {
-            if (r->io.conn.read_limit >= r->io.conn.read_so_far - buffered)
+            /* Make sure the caller didn't request to read past the limit.
+             * If they didn't we'll buffer till the limit, if they did, we'll
+             * return an error. */
+            if (r->io.conn.read_limit >= r->io.conn.read_so_far + needs)
                 toread = r->io.conn.read_limit - r->io.conn.read_so_far - buffered;
             else {
                 errno = EOVERFLOW;