It seems that RFC does not say anything about the timestamp of lease
we should use: time that the client sent a request or received a reply.
In DHCPv6 client and NDisc, we use a timestamp that we receive a packet,
rather than we sent something. So, let's consistently use the same
logic for DHCPv4 client.
By using the logic, we will hopefully not forget to set timestamp again,
which is fixed by 089362976c.
Then, stop client in the caller side.
This also makes
- ignore all errors except for resource errors like OOM when FORCERENEW
is received,
- trigger assertion when an message received even if the client is
stopped.
This should not change any functionality. Just refactoring.
If a server replies an ACK for the initial DISCOVER, previously
request_sent was not set, so networkd handle the lease timed out.
Follow-up for 808b65a087.
This implements the DHCPv4 equivalent of the DHCPv6 Rapid Commit option,
enabling a lease to be selected in an accelerated 2-message exchange
instead of the typical 4-message exchange.
We use it for more than just pipe() arrays. For example also for
socketpair(). Hence let's give it a generic name.
Also add EBADF_TRIPLET to mirror this for things like
stdin/stdout/stderr arrays, which we use a bunch of times.
This is convenient when the server supports IPv6 only mode.
Otherwise, we cannot request a new address during the client is waiting an
IPv6 connectivity. Note, the minimal timespan is 5min, and a server may
send a quite large value.
The condition was outdated, e.g. SELECTING state does not have a lease.
See client_handle_offer() and client_enter_requesting().
The condition based on the state may become much complex in the future.
Let's use simpler condition.
This is preparation for #28891, which adds a bunch more helpers around
"struct iovec", at which point this really deserves its own .c/.h file.
The idea is that we sooner or later can consider "struct iovec" as an
entirely generic mechanism to reference some binary blob, and is the
go-to type for this purpose whenever we need one.
Note, previously, we set 1 second if the time value is zero.
But the adjustment is dropped now, as for the lifetime we have explicit
check that the message has non-zero lifetime, and for T1 and T2 we have
better adjustment in client_set_lease_timeouts().
This splits sd_dhcp6_lease_get_address() into small pieces,
and introduce FOREACH_DHCP6_ADDRESS() macro.
Also, the lifetimes provided by _get_address_lifetime() are now in usec,
and _get_address_lifetime_timestamp() provides timestamp.
The same change is also applied for IA_PD.
Yu Watanabe
Nandakumar Raghavan
Ronan Pigott
Lennart Poettering
Frantisek Sumsal
Susant SahaniandYu Watanabe
Alexandre Peixoto Ferreira
Luca Boccassi
Zbigniew Jędrzejewski-Szmek