svn commit: r300973 - head/usr.sbin/rpcbind

[Ngie Cooper]

> On May 29, 2016, at 14:52, Bjoern A. Zeeb <bzeeb-lists at lists.zabbadoz.net> wrote:
> 
> 
>> On 29 May 2016, at 20:28 , Garrett Cooper <ngie at FreeBSD.org> wrote:
>> 
>> Author: ngie
>> Date: Sun May 29 20:28:01 2016
>> New Revision: 300973
>> URL: https://svnweb.freebsd.org/changeset/base/300973
>> 
>> Log:
>> Follow up to r300932
>> 
>> In the event MK_INET6 != no in userspace, but is disabled in the
>> kernel, or if there aren't any IPv6 addresses configured in userspace
>> (for lo0 and all physical interfaces), rpcbind would terminate
>> immediately instead of silently failing on
>> 
>> Skip over the IPv6 block to its respective cleanup with freeifaddrs if
>> creating the socket failed instead of terminating rpcbind immediately
> 
> Stupid question (without me looking at the code):  what does it do in case I have no IPv4 support in the kernel anymore (as I do on some machines)?

The socket call only applied to IPv6. It intentionally sets the multicast address and leaks the socket, which causes coverity to note the leak as fd is local to the function.

I'll definite try out this case as well tonight.. init_transport(..) (another function in rpcbind that uses rpc(3) under the covers) was not working under certain edge cases with some changes I had put up for CR.

I feel like I wandered into another blackberry bramble patch with rpc(3). I have some ideas for refactoring rpcbind and friends to ease maintenance that I'll work on after 11.0-RELEASE hits.

Thanks!
-Ngie