rpcbind fails to start in a jail with only one ipv4 address assigned on amd64 (no other platform available for testing) Fix: As a workaround I disabled udp6 and tcp6 in /etc/netconfig, rpcbind works fine now, the last part of a "ktrace rpcbind" is shown below 13860 rpcbind GIO fd 7 wrote 70 bytes "<27>Feb 18 03:55:12 rpcbind: cannot bind ::1 on udp6: Invalid argument" 13860 rpcbind RET sendto 70/0x46 13860 rpcbind CALL socket(PF_INET6,SOCK_DGRAM,IPPROTO_UDP) 13860 rpcbind RET socket 8 13860 rpcbind PSIG SIGSEGV SIG_DFL 13860 rpcbind NAMI "rpcbind.core" How-To-Repeat: Install 7.1-STABLE to a version after the multiple ip jail changes Start rpcbind, it fails with "Segmentation fault (core dumped)"
Responsible Changed From-To: freebsd-bugs->freebsd-jail Over to maintainer(s).
Hi, this jail behaviour will soon change again that opening an IPv6 socket will be possible w/o the IP though this currently matches the behavior of the base system. The problem here seems to be that (without looking at the code) rpcbind is not checking return codes properly. /bz -- Bjoern A. Zeeb The greatest risk is not taking one.
From the jail side this should have been fixed with the latest MFC from jamie@ . But this really is an rpcbind bug. -- Bjoern A. Zeeb The greatest risk is not taking one.
> From the jail side this should have been fixed with the latest MFC > from jamie@ . > > But this really is an rpcbind bug. > > -- > Bjoern A. Zeeb The greatest risk is not taking one. > Hi Bjoern, I can confirm it now works again without editing the /etc/netconfig file, should we leave the p.r. open for reference to the rpcbind problem? Thanks, Rob Evers -- A: Because it reverses the logical flow of conversation. Q: Why is top posting frowned upon?
Responsible Changed From-To: freebsd-jail->freebsd-bugs
For bugs matching the following criteria: Status: In Progress Changed: (is less than) 2014-06-01 Reset to default assignee and clear in-progress tags. Mail being skipped