David Powell wrote: > Tony Nguyen wrote: >>> Since the string pointed to by 'p' is 'l' bytes long, after the >>> strncpy 'port_str' won't neccesarily be NUL terminated, which >>> could >>> cause atol to fail. >> >> I thought the prior bzero call would prevent that problem, especially >> if we make sure 'p' doesn't advance past 'addr'. Or would you prefer >> should be replacing it with "strlcpy(port_str, p, l + 1)" ? > > No, you're right. The bzero should handle it. > I replaced these str* and atol calls with strtol to simplify and make it clearer.
>>> svc_getrpcinfo: >>> >>> I'm still wondering where the "12" in "buf[12]" comes from. It >>> should either be documented, or a constant (one hopefully derived >>> from whatever determines the length of the buffer). >> >> Yes, a constant would be more clear but I'm not sure what the optimal >> size should be. The buffer holds either port number (2^16 value) or a >> netid (currently limited to tcp, udp, tcp6, udp6) and an extra space. >> Does 12 seems reasonable? > > 12 sounds fine, I would just like for it to be a #define with a > comment explaining where it comes from. > will do. Thanks a bunch, tony
