Roy Marples [Mon, 29 Jul 2024 15:17:08 +0000 (16:17 +0100)]
DHCP: No longer set interface mtu (#346)
We've been enforcing an interface MTU that is slightly larger
than the minimum for some time.
Instead, log an error than the MTU is smaller than the minimum
to send a BOOTP message.
The DHCP MTU is only used when adding routes as setting the
interface MTU can cause a PHY reset which is bad.
DHCP: Handle option 108 correctly when receiving 0.0.0.0 OFFER (#342)
* DHCP: Handle option 108 correctly when receiving 0.0.0.0 OFFER
According to RFC8925 section 3.3.1, when the server supports both option
108 (IPv6-Only Preferred) and option 116 (Auto-Configure), and the
client only sends IPv6-Only Preferred option, then the server SHOULD
return 0.0.0.0 as the offered address, and not setting the
Auto-Configure option.
However, in our current client code, the IPv6-Only Preferred option in a
0.0.0.0 OFFER is only handled correctly when the Auto-Configure option
is present. This patch fixes this issue.
Leo Ruan [Thu, 25 Jul 2024 08:37:11 +0000 (16:37 +0800)]
IPv4LL: Restart ARP probling on address conflict (#340)
* IPv4LL: Restart ARP probling on address conflict
When IPv4LL address conflict is detected, it is failed to restart IPv4LL
since IPv4LL is running. The commit fixes the problem by restarting ARP
probing instead of restarting IPv4LL.
Roy Marples [Fri, 14 Jun 2024 16:29:35 +0000 (17:29 +0100)]
DHCP6: Don't remove state in DISCOVER
As we might still have config that failed to CONFIRM/REBIND.
Try and remove timers more sensibly rather than blaket statements
so that expire timeout still works.
Roy Marples [Tue, 9 Jul 2024 12:39:07 +0000 (13:39 +0100)]
DHCP6: Persist configuration on confirm/rebind failure
only expire the lease once the last address has expired.
This is more in accordance with RFC8415 18.2.
Note, we still require an initial confirm/rebind on link state
change to ensure the configuration for link is valid.
Extending the lastlease in DHCP6 has been removed.
Roy Marples [Thu, 27 Jun 2024 05:27:22 +0000 (06:27 +0100)]
linux: allow roaming without IFF_DORMANT
IFF_DORMANT is apparently only set by supplicants.
For roaming without supplicants, we need to check that IFF_UP and
IFF_LOWER_UP are set, but IFF_RUNNING is not.
Thanks to Boris Krasnovskiy for the patch.
Fixes #335.
While here, remove our define if IFF_LOWER_UP as we can now include
linux/if.h
Roy Marples [Wed, 26 Jun 2024 08:36:35 +0000 (09:36 +0100)]
DHCP6: Don't remove delegated prefix addresses on start or fail
In the event where an interface is first delegated and then
started, we don't want the reading of a lease file to remove the
delegated prefixes or addresses.
Neither do we want any DHCP6 failure to remove the delegated
addresses.
As such, we also can't test the actual DHCP6 state when building
routes. Instead, just change the test for a delegated prefix
or not so we can still prefer non delegated routes.
While here, fix reject routes being restored for privsep.
Roy Marples [Tue, 25 Jun 2024 10:40:14 +0000 (11:40 +0100)]
DHCP6: Don't remove delegated prefix addresses on start
In the event where an interface is first delegated and then
started, we don't want the reading of a lease file to remove the
delegated prefixes or addresses.
As such, we also can't test the actual DHCP6 state when building
routes. Instead, just change the test for a delegated prefix
or not so we can still prefer non delegated routes.
Roy Marples [Fri, 21 Jun 2024 11:58:38 +0000 (12:58 +0100)]
IPv6: DUPLICATED could be announced by RTM_DELADDR
This at least is true on Linux when addresses are not permanent.
As such, update address flags on all paths and run DAD hooks
even for address deletions.
This helps to give reason to the messages that pid 0 has deleted
your newly added address.
Brett Holman [Sat, 20 Jan 2024 01:43:26 +0000 (18:43 -0700)]
Force dumplease to parse stdin
Checking if data exists in the stdin buffer via an ioctl is
unreliable. Allow a user to pass '-' to force stdin parsing.
This provides a fix that will allow "no interfaces" to still
mean "dump all interfaces".
Roy Marples [Fri, 24 May 2024 10:30:29 +0000 (10:30 +0000)]
DHCP: use request_time, fallback_time and ipv4ll_time
Rather than reboot time.
This allows reboot time of zero to skip the using old leases
while still allowing REQUESTs to gracefully fallback to DISCOVER.
request_time has a default of 180 seconds to mirror the DHCPv6
equivalent.
fallback_time and ipv4_ll time have a default of 5 seconds
to mirror the default reboot time.
Benjamin Drung [Tue, 7 May 2024 09:20:04 +0000 (11:20 +0200)]
hooks/30-hostname: Exit with 0 if setting hostname is not needed
If the hostname is already set before dhcpcd is started, `need_hostname`
will hit the "No old hostname" case and will call `false` to return 1.
`set_hostname` will return with the same return value. Then
`30-hostname` will exit with the same exit value:
```
dhcpcd-10.0.6 starting
dev: loaded udev
DUID 00:03:00:01:3c:97:0e:e9:32:3c
enp0s25: IAID 0e:e9:32:3c
enp0s25: soliciting a DHCP lease
enp0s25: offered 192.168.3.61 from 192.168.1.7
enp0s25: ignoring offer of 192.168.3.61 from 192.168.1.8
enp0s25: probing address 192.168.3.61/20
enp0s25: leased 192.168.3.61 for 3600 seconds
enp0s25: adding route to 192.168.0.0/20
enp0s25: adding default route via 192.168.1.1
script_status: /usr/lib/dhcpcd/dhcpcd-run-hooks: WEXITSTATUS 1
exiting due to oneshot
dhcpcd exited
```
Return with value 0 in case setting the hostname is not needed.
Bug-Ubuntu: https://launchpad.net/bugs/2064926 Signed-off-by: Benjamin Drung <benjamin.drung@canonical.com>
Roy Marples [Sat, 4 May 2024 22:33:11 +0000 (23:33 +0100)]
DHCPv6: Don't re-INFORM if the RA changes
RFC 8415 21.23
If the Reply to an Information-request message does not contain this
option, the client MUST behave as if the option with the value
IRT_DEFAULT was provided.
So we should not be influenced by only the RA changing the pl/vl
times of it's addresses.
Roy Marples [Sat, 4 May 2024 11:33:24 +0000 (12:33 +0100)]
privsep: Reduce fd use
On start close all FD's above stderr.
Close some fd's we don't need in processes spawned from priv.
Ensure we init some FD's to -1 to ensure we don't close stdin.
If DEBUG_FD is defined, we log FD's opened by pid.
Audit process FD usage and document it so I don't forget it.
Taoyu Li [Tue, 12 Mar 2024 02:38:28 +0000 (11:38 +0900)]
Respect IPV6_PREFERRED_ONLY flag on DHS_NONE state
Current IPv6_PREFERRED_ONLY (option 108) handling code is only effective
when current state is DHS_DISCOVER and DHS_REBOOT. However, when we
receive multiple ACKs upon our REQUEST, the first ACK will trigger the
use_v6only code path and dhcp_drop() us into DHS_NONE state, as a result
the option 108 on the second ACK won't be handled correctly and we'll
bind to the lease instead.
This patch fixes the issue by adding DHS_NONE as a state to respect
option 108 as well.
Roy Marples [Tue, 2 Apr 2024 08:55:18 +0000 (09:55 +0100)]
compat: stub out _rs_forkhandler for compat/arc4random.c
We need to keep the fd open at fork, but we retained the code to
handle a fork.
The original update to chacha avoided this by guarding the call
but left the code alive which produced an unused function warning
on the GitHub Ubuntu runner.
This update fixes that.
Daniel Goertzen [Sat, 9 Mar 2024 01:27:57 +0000 (19:27 -0600)]
add RFC4191 support (#297)
* add RFC4191 support
- handles route information options from RAs.
- refactor `sa_fromprefix()` to expose lower level functionality
- refactor `ipv6nd_rtprefix()` to be usable outside of `struct ra` context
* changes as requested by RM
- mostly minor/cosmetic changes
- functional change: "no longer a default router" warning moved to capture changes from routeinfo options
Lloyd Parkes [Sun, 25 Feb 2024 12:21:12 +0000 (01:21 +1300)]
Define the Azure Endpoint and other site-specific options (#299)
Added the azureendpoint site-specific option as an ipaddress
definition to make it easier for Azure VMs using dhcpcd to get their
WireServer endpoint address.
Added binhex definitions for all otherwise undefined site-specific
options so that site-specific hooks can use them.
Petr MenÅ¡Ãk [Fri, 16 Feb 2024 16:15:35 +0000 (17:15 +0100)]
Move dhcp(v4) packet size check earlier (#295)
dhcp_handlebootp handled zero sized packets correctly, but
dhcp_redirect_dhcp did not have such protection. Move size check before
both of them. Size when called from dhcp_packet is checked by
is_packet_udp_bootp call. Only dhcp_recvmsg needs earlier checking to be
added.
Roy Marples [Sun, 24 Dec 2023 12:07:25 +0000 (12:07 +0000)]
DHCP: DECLINE address on ARP defend failure
Also, drop the lease.
This should get us a new address from the DHCP server when we
re-enter DISCOVER to avoid looping on the same address and fail
again.
dhcpcd: Add support for arp persist defence (#273)
RFC 5227 recommends 3 ways to deal with address conflict detection.
a) Stop everything.
b) Defend and then stop on fail - this is what dhcpcd currently does.
c) Notify and carry on.
The current change implements the option c. A new option arp_persistdefence
has been added and when this is enabled, the a defence is attempted upon a
conflict and when that fails, an error is logged on every other conflict
within the DEFEND_INTERVAL and the current IP address is retained.
Roy Marples [Mon, 18 Dec 2023 12:22:13 +0000 (12:22 +0000)]
DHCP6: For Prefix Delegation, the - interface means no assignment
- is an invalid interface name.
So we take this to mean don't assign the Delegated Prefix to
any interfaces.
The reject route for the Delegated Prefix is still installed.
Roy Marples [Mon, 18 Dec 2023 12:15:10 +0000 (12:15 +0000)]
DHCP6: Improve logging when changing IA type
Changing from PD to IA or IA to PD can result in a diagnostic
when there is no address to confirm AND we haven't loaded
a lease.
This improves the check and no more Success errors should
be reported.
Roy Marples [Mon, 13 Nov 2023 10:29:58 +0000 (10:29 +0000)]
dhcpcd: Remove stdio callback and detach on daemonise
For some reason, the stdio callback is extremely flaky on
*some* Linux based distributions making it very hard to debug some
things.
Removing it is fine because we now enforce that we have file descriptors
for stdin, stdout and stdrr on launch and dup them to /dev/null on daemonise.
It's also interesting to see behavioural differences between
some socketpair implementations that emit a HANGUP and some don't.
As such, we now close the fork socket on daemonise once more AND
in the fork_cb depending on if we hangup or read zero first.
Roy Marples [Sun, 12 Nov 2023 12:16:53 +0000 (12:16 +0000)]
dhcpcd: Detach from launcher before stopping root process
This fixes non privsep builds where the launcher reports dhcpcd
hungup. Unsure why this happens, but it should not be a problem.
While here, shutdown has no effect on non STREAM sockets and
remove the silly error logging in fork_cb that we read an
error. We already printed the error so this makes no sense.
Roy Marples [Wed, 8 Nov 2023 11:09:52 +0000 (11:09 +0000)]
dhcpcd: Close fork_fd on hangup
Closing it early results in zero length reads in some situations.
Logging that we forked via the launcher process also make more sense
and allows us to use log* functions.
While here, handle error condtions better by forcing a return
rather than handling an invalid state.
Roy Marples [Mon, 23 Oct 2023 14:26:16 +0000 (15:26 +0100)]
compat: test for memset_explicit, explicit_bzero and memset_s
These won't be optimised away by the compiler and our arc4random
compat function should use them *if* available.
If none are then a warning will be emitted to say it's potentially insecure.
Hopefully only uclibc users will see this message.
Roy Marples [Thu, 19 Oct 2023 10:11:05 +0000 (11:11 +0100)]
privsep: Notify processes when dhcpcd has daemonised
This allows us to dup stdout and stderr onto stdin which is
guaranteed to be dupped to /dev/null.
This in turn avoids SIGPIPE when the privileged proccess launches
the script and it wants to write to stdout/stderr or stupidly
read from stdin.
Fabrice Fontaine [Mon, 16 Oct 2023 17:03:41 +0000 (19:03 +0200)]
compat/arc4random.c: use memset instead of explicit_bzero (#252)
Use memset instead of explicit_bzero to avoid the following build
failure with uclibc-ng since version 10.0.3 and
https://github.com/NetworkConfiguration/dhcpcd/commit/837d09e34c487edaa92aa2ae71a630d84c927f8e:
/home/fabrice/buildroot/output/host/lib/gcc/arm-buildroot-linux-uclibcgnueabi/12.3.0/../../../../arm-buildroot-linux-uclibcgnueabi/bin/ld: ../compat/arc4random.o: in function `_rs_stir_if_needed':
arc4random.c:(.text+0x8cc): undefined reference to `explicit_bzero'
musl libc doesn't have __NR_clock_gettime definition,
but has __NR_clock_gettime32. clock_gettime implementation
fallbacks to 32-bit version if 64-bit is not supported by the kernel.
Roy Marples [Wed, 4 Oct 2023 12:03:21 +0000 (13:03 +0100)]
options: introduce the uri option as opposed to a string
Currently we don't attempt to validate a uri given, aside from
not allowing any space characters within.
If the option is `array uri` then the first two bytes are the
length of the uri in network order and the rest is a uri element.
The uri's are space separated for the variable because space is not
allowed within the uri.
This allows us to implement RFC 8572, Secure Zero Touch Provisioning.
Send correct amount of used buffer for prefix exclude option (#250)
The payload of the prefix exclude option was correctly created
but the amount of bytes to send in the DHCPv6 request was
always set to 0 which resulted in an invalid prefix exclude option
This patch fixes this behavior by calculating the correct amount
of bytes to send.
Roy Marples [Mon, 25 Sep 2023 09:34:12 +0000 (10:34 +0100)]
Guard against handling many SIGTERM/SIGINT.
Privsep has a mini-eloop for reading data from other processes.
This mini-eloop processes signals as well so we can reap children.
During teardown we don't want to process SIGTERM or SIGINT again,
as that could trigger memory issues.
Detect libcrypto in configure script. Only fall back
to using libcrypto when /usr libs are allowed and no
other compatible implementation is available or when
--with-openssl is passed explicitly.
Make sure libcrypto and libmd are never linked at the
same time.
Add OpenSSL based SHA256 and HMAC compat shims in
compat/crypt_openssl. Depeding on version and build flags,
libcrypto ships with a compatible SHA256 API in
"openssl/sha.h". OpenSSL 3 has deprecated the SHA API,
so if it is not detected we fall back to an EVP_DIGEST
based version.
Because the API might still be in use in OpenSSL internally,
the compatibility wrappers have a dhcpcd_ prefix to avoid
symbol conflicts.
* Add sha256 tests based on the existing hmac-md5 tests.
When persistent is not specified, dhcpcd de-configures the interface at exit.
However, the default dhcpcd.conf example provided enables the option.
See https://github.com/NetworkConfiguration/dhcpcd/discussions/140
Tobias Heider [Tue, 15 Aug 2023 16:06:48 +0000 (18:06 +0200)]
privsep: fix strlcpy overflow in psp_ifname (#239)
When running our Ubuntu tests with libc6 and strlcpy overflow checks
enabled we found that the wrong size is passed to strlcpy resulting
in a crash because of an overflow.