summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAge
* DHCP: Fix reading RENEW messages from inet socketdhcpcd-8Roy Marples2020-12-20
| | | | It helps to use the correct buffer to store the control message ....
* DHCP6: Apply delegations to interface on carrier upRoy Marples2020-06-10
| | | | | Even with DHCP6 turned off for the interface. As long as it was activated by another interface this is fine.
* NetBSD: Gear up for kernel RA removal.Roy Marples2020-06-10
| | | | | | linkmtu will vanish from nd_ifinfo along with other things. Also, there is not need to pre-fetch ifinfo because flags are no longer updated when setting ifinfo back again.
* Linux: fix compile on old onesRoy Marples2020-06-10
|
* Bump date for priorRoy Marples2020-05-12
|
* Fix neighbor discovery option parsing (#16)Andrew Childs2020-05-12
| | | | | | | | | | | | * make_option_mask: use `nd_` prefix for nd options The prefix for neighbor discovery options was partially changed to `nd6_` in d5700a48edf7006369f766d225e70bbca3d8055c. However set_option_space continued to use `nd_`. This prevented either `nd6_prefix_information` or `nd_prefix_information` from being parsed since the former was in the wrong option space, and the latter did not have the prefix removed. * manual: fix nd_ and dhcp6_ option prefixes
* Release dhcpcd-8.1.9dhcpcd-8.1.9Roy Marples2020-04-21
|
* align CMSG bufferRoy Marples2020-04-21
|
* DHCP: Avoid mis-aligned BOOTP structureRoy Marples2020-04-20
| | | | This was caused by the recent change to pass back the frame header.
* cast is not neededRoy Marples2020-04-19
|
* Fix small builds.Roy Marples2020-04-12
|
* Release dhcpcd-8.1.8dhcpcd-8.1.8Roy Marples2020-04-12
|
* Fix build on old BSD.Roy Marples2020-04-12
|
* Release dhcpcd-8.1.7dhcpcd-8.1.7Roy Marples2020-04-02
|
* RA: Remove timecheck on sorted router listRoy Marples2020-03-11
| | | | Newer is always at the end, mirrors master branch.
* RA: Prefer older routers at the preferenceRoy Marples2020-03-11
|
* logerr: Use macros to call log functionsSergey Nikiforov2020-02-20
| | | | | | This allows downstream implementions to inject their own data into the log message and/or supply new logging targets whilst preserving macros such as __FILE__ and __LINE__.
* if: Decode vlid from the interface nameRoy Marples2020-02-15
| | | | This just clarifies the former fix
* if: . is part of the interface name, not an aliasRoy Marples2020-02-15
|
* dhcp: Cast away a compile warningfox2020-02-07
|
* Linux: mask the NETLINK_GET_STRICT_CHK: Protocol not available errorRoy Marples2020-02-07
| | | | | Just to appease Arch LTS users (packages are built with newer headers than what kernel supports).
* options: Improve AF_INET6 address parsing and loggingRoy Marples2020-02-07
|
* inet6: Calculate the prefix in the canonical formRoy Marples2020-02-07
| | | | Rather than being clever and getting it wrong.
* Add an extra fallthrough to priorRoy Marples2020-02-06
|
* inet6: Swap between DISCOVER and INFORM betterRoy Marples2020-02-06
| | | | | Most of the time this will be dictated by the RA, but let our config decide as well.
* DHCP6: Clean up old lease when we fail to confirm/rebind, etcRoy Marples2020-02-06
| | | | Also removed the TIMEOUT states which makes things easier to read.
* inet6: Ensure expired routers are cleared after a carrier lossRoy Marples2020-02-06
| | | | | | | | | | When we lose carrier, mark all RAs as willexpire and add the timeout to expire to it's own queue. On receipt of a matching RA, clear the willexpire flag. When the above timeout occours, set doexpire on all RA's with withexpire and then call the general expirera function. This is needed because expirera can be called at any point.
* eloop: define eloop queue numbers in common.hRoy Marples2020-02-06
| | | | Allows for easier maintainance.
* compat: Fix a typo in pidfile when O_CLOEXEC is not definedPetr Gotthard2020-02-05
| | | | While here, include fcntl.h in if.c for older environments.
* ARP: Only copy out frame header if we have it.Roy Marples2020-02-05
| | | | Frameless does exist.
* BPF: Don't use magic in buffer length assignmentsRoy Marples2020-02-05
| | | | | Define frame header lengths and frame size maximums. Use this to calculate required buffer sizes.
* DHCP: Support jumbo framesRoy Marples2020-02-05
| | | | | Also important as we now need to store the frame header in the initial buffer.
* BPF: Return the frame header with the dataRoy Marples2020-02-05
| | | | | | | | | | | | | For DHCP, we then just skip over the frame header. For ARP, we extract the frame source and destination addresses so we can log the source in the event of a conflict. This is important as a user has found a router which sets the ARP source and destination hardware addresses to all zeros but unicasts the ARP straight to our hardware address. https://serverfault.com/questions/297425/ip-address-conflict-with-mac-address-000000000000 https://discussions.flightaware.com/t/piaware-wont-reconnect-to-wifi-network-if-it-drops-off/59789
* BSD: ignore XEN DOM0 virtual interface xvifRoy Marples2020-02-04
|
* if: decode XEN vif1.2 and xvif1i2 as vif1:2Roy Marples2020-02-04
|
* Release dhcpcd-8.1.6dhcpcd-8.1.6Roy Marples2020-01-27
|
* IPv4LL: A state might not always exist when running the script.Christos Zoulas2020-01-27
| | | | | For example, we deleted existing IPv4LL addresses that dhcpcd doesn't directly control when we obtained a DHCP lease.
* BSD: Fix INET flags on RTM_NEWADDR messages.Roy Marples2020-01-22
|
* DHCP: Take interface reference on Address Defend failureRoy Marples2020-01-22
| | | | As the astate could be freed on dropping the DHCP state.
* BSD: Improve RTM_NEWADDR/RTM_DELADDR validaton some moreRoy Marples2020-01-22
|
* BSD: Validate RTM_NEWADDR as wellRoy Marples2020-01-17
| | | | | This is important because the address could be added and quickly deleted and RTM_NEWADDR arrvives after deletion.
* inet6: Temporary addresses only work if prefix + ifidlen == 128Roy Marples2020-01-17
| | | | Cater for various prefix lengths when using temporay addresses.
* inet6: Support a /128 autoconf prefix from RARoy Marples2020-01-17
| | | | This is apparently a thing. Unsure why, but heh ho.
* Release dhcpcd-8.1.5dhcpcd-8.1.5Roy Marples2020-01-03
|
* Move some option finialisation so it's available globally.Roy Marples2020-01-02
|
* inet: Allow forcing a host route from an interface without a leaseRoy Marples2020-01-02
|
* Linux: fix RA time unit confusionAnders Kaseorg2020-01-01
| | | | | | | | | | The RA times are provided in milliseconds, but commit 569051c8aa8fc297eb8edb7bd228e0fd353d30c1 (dhcpcd-8.1.3~18) “Linux: prefer ms RA times” incorrectly scaled them by an extra factor of 1000 before writing them to retrans_time_ms and base_reachable_time_ms. Instead, write the values in milliseconds directly to the *_ms files. Signed-off-by: Anders Kaseorg <andersk@mit.edu>
* Release dhcpcd-8.1.4dhcpcd-8.1.4Roy Marples2019-12-20
|
* options: Fix allocating the script optionRoy Marples2019-12-20
| | | | | | When passing PARSE_STRING_NULL we expect to store the string NULL terminated. As such, allocate space for it an ensure we have space for it.
* Release dhcpcd-8.1.3dhcpcd-8.1.3Roy Marples2019-12-20
|
generated by cgit v1.2.3 (git 2.25.1) at 2021-03-13 22:53:19 +0000