in IPv6 when dhcpcd is overriding the in-kernel RA support.
For Linux kernels (3.18+) which support IFA_F_MANAGETEMPADDR the bulk of
this changeset is compiled out and the kernel will manage the temporary
addresses entirely.
For BSD, this is a fully compliant implementation with the caveat
that when dhcpcd is restarted the last non deprecated temp address on the
interface will be treated as being created when it was last updated
rather when it was actually added. Thus this may voilate section 3.3.
As dhcpcd won't restart in normal operation, this isn't an issue.
For Linux (3.18+) which supports IFA_F_MANAGETEMPADDR, the bulk of this
changeset is compiled out as the kernel will manage the temporary addresses.
For older Linux this is a fully compliant implementation with the caveat that
when restarted new temporary addresses will be generated.
Fixes [2ddfcb190f]
The interface with the lowest metric gets the address.
When that interfaces loses the address, the next valid interface with the
lowest metric gets it.
natural string based on content except as noted:
* domain (RFC3397)/dname (string) is strict domain name allowance
(ie, [alnum] with _- (but not at the start or end))
* string is now printable ascii (1-127) until invalid
* ascii is all ascii (1-127) until invalid
* raw is all chars (1-255) until NUL
* binhex is a hex representation of the option including embedded NULs
* ssid is still escpaed octal because it's expected to be human readable
AND can technically be all NUL
* everything else has strict option -> value encoding
This removes all shell escaped encoding - dhcpcd will assume that IF the
--script option is a shell, it will quote variables correctly.
The stock dhcpcd-run-hooks does.
dhcpcd -V now prints how the variables will be decoded.
Changed some options in dhcpcd-definitions.conf to more sensible defaults.
If this is not found in the system headers, fall back to assuming WEXT.
This has been done because the WEXT module has been marked as deprecated
for a while now and I'm getting bored of people asking me why their
wireless wasn't working. Bloated dhcpcd by another 600 bytes as a result.
Only one call to write(2) should be performed for each POLLOUT check
via poll(2) so we should never see EAGAIN when writing to control
sockets ever again.
Each fd queue is limited to 100 entries so we don't OOM with badly
written control subscribers.
