Comment 6 for bug 1215911

Thank you.

By the way, is there any possibility that "the message of a block device which
the wait-for-root is waiting for" cannot be received after ENOBUFS?

I think that any messages of block devices which are generated while the socket
buffer is full cannot be received using recvmsg(). I worry that "the message of
a block device which the wait-for-root is waiting for" is by chance generated
as one of such messages.

If such case occurs, waiting at udev_monitor_receive_device() will not return
"the message of a block device which the wait-for-root is waiting for", and
wait-for-root will uselessly waits until interrupted by SIGALRM. (The boot
would succeed anyway because the device file would have been created by the
time interrupted by SIGALRM. But really useless delay...)