Description:
systemctl is unable to disable any unit because it does not retry socket receives that fail with EAGAIN.


Additional info:
* package version(s)
204-1
x86_64 ArchLinux up-to-date as of today (2013-05-22).
* config and/or log files etc.


Steps to reproduce:

Try to disable any unit with systemctl, inactive or active, eg. the active 'proc-xen.mount' :
$ systemctl | grep proc-xen
proc-xen.mount loaded active mounted /proc/xen
$ systemctl disable proc-xen.mount
Failed to issue method call: No such file or directory
or the inactive 'display-manager.service' :
$ systemctl -a | grep display
display-manager.service error inactive dead display-manager.service
$ systemctl disable display-manager.service

An strace shows why these commands fail:
$ strace systemctl disable proc-xen.mount
...
sendmsg(3, {msg_name(0)=NULL, msg_iov(2)=[{"l\1\0\1\34\0\0\0\1\0\0\0\251\0\0\0\1\1o\0\31\0\0\0/org/fre"..., 192}, {"\23\0\0\0\16\0\0\0proc-xen.mount\0\0\0\0\0\0", 28}], msg_controllen=0,
msg_flags=0}, MSG_NOSIGNAL) = 220
clock_gettime(CLOCK_MONOTONIC, {73496, 166603036}) = 0
poll([{fd=3, events=POLLIN}], 1, 25000) = 1 ([{fd=3, revents=POLLIN}])
recvmsg(3, {msg_name(0)=NULL, msg_iov(1)=[{"l\4\1\1\0\0\0\0\1\0\0\0q\0\0\0\1\1o\0\31\0\0\0/org/fre"..., 2048}], msg_controllen=0, msg_flags=MSG_CMSG_CLOEXEC}, MSG_CMSG_CLOEXEC) = 382
recvmsg(3, 0x7fffdeb0fff0, MSG_CMSG_CLOEXEC) = -1 EAGAIN (Resource temporarily unavailable)
writev(2, [{"Failed to issue method call: No "..., 54}, {"\n", 1}], 2Failed to issue method call: No such file or directory) = 55
close(3) = 0
exit_group(1) = ?

This happens for every 'disable' command sent.

I strongly suggest that the systemd programmers read the 'recvmsg(2)', recv(2), 'read(2)' and 'socket(7)' manual pages ,
where EAGAIN is documented as being returned by read() / recvmsg() when there is currently no data to be received
and the socket is marked as non-blocking / O_ASYNC . If they are expecting a reply to be sent, they should retry
the operation if the receive fails with errno set to EAGAIN.

This problem effectively makes my whole arch linux system unusable unless I'm running Xen, since
systemd always fails to enter the default "runlevel", and no units can be disabled.

I've done an 'egrep -RI proc-xen /' with no results found - it must be in a binary database somewhere -
any ideas where this database might be / how to reinitialize it so that NO units are enabled ?
( I want to disable systemd completely, and replace it with SysVinit, since it is not of sufficient
quality to be useful yet).