Age | Commit message (Collapse) | Author |
|
Signed-off-by: Emanuele Fia <name29@meta.com>
|
|
|
|
|
|
|
|
`NewInform()` exposes a consumer-facing interface that accepts modifiers, but
this functionality is never tested unlike with for example `NewRequest()` and
`NewRenew()`. To ensure that modifiers are correctly respected and any future
modifications won't break this support, add a modifier test for `NewInform()`.
Signed-off-by: Dennis Marttinen <twelho@welho.tech>
|
|
This is just a minor readability improvement that makes `NewInform()`
consistent with the neighboring functions like `NewRequest()`, `NewReply()`,
`NewRelease()` etc. Having the prepended modifiers on the same line helps
with separating the modifiers that provided by the function from the modifiers
passed in by the consumer of the function.
Signed-off-by: Dennis Marttinen <twelho@welho.tech>
|
|
Any parameter passed to `require.NotNil()` after the first two will not be
tested but instead appended to the message that is printed if the test fails.
This means that `lease.ACK` was never tested to not be nil.
Signed-off-by: Dennis Marttinen <twelho@welho.tech>
|
|
Signed-off-by: Dennis Marttinen <twelho@welho.tech>
|
|
|
|
|
|
|
|
Add `SO_REUSEPORT` to aid in debugging and testing.
Signed-off-by: Jacob Weinstock <jakobweinstock@gmail.com>
|
|
|
|
Signed-off-by: Emanuele Fia <name29@fb.com>
|
|
|
|
Signed-off-by: Dennis Marttinen <twelho@welho.tech>
|
|
Signed-off-by: Dennis Marttinen <twelho@welho.tech>
|
|
Signed-off-by: Dennis Marttinen <twelho@welho.tech>
|
|
Signed-off-by: kayos@tcp.direct <kayos@tcp.direct>
|
|
mdlayher/raw supports them, so why don't we support them, too.
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
This should fix the panic we've seen with the malformed packets on the
wire. Remaining buffer size should be checked before trying to access
the data, otherwise Go might panic on out of bounds slice operation.
Signed-off-by: Andrey Smirnov <andrey.smirnov@talos-systems.com>
|
|
RFC 2131 Section 4.4.1 specifies that REQUEST messages should set
'siaddr' to 0.
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
|
|
|
|
Fix payload buffer to end at actual payload end. UDP header length includes the 8 bytes of the header so we do not want the +8 at the end of the buffer.
Signed-off-by: noiz <noiz@users.noreply.github.com>
Co-authored-by: Pablo Mazzini <pmazzini@gmail.com>
|
|
* Adding support for Ciena DHCPv4
Signed-off-by: Emanuele Fia <name29@fb.com>
|
|
Signed-off-by: noiz <noiz@users.noreply.github.com>
|
|
empty (#444)
* Adding check for empty serial number in opt 61
|
|
As described in RFC8357 this is a 0-length suboption used to indicate
that the relay supports the arbitrary UDP port behavior from RFC8357
Since this is 0-length the constant is enough to handle the full
behavior
Signed-off-by: Anatole Denis <natolumin@unverle.fr>
|
|
* Adding support to parse vendor data for cisco firepower
|
|
|
|
u-root -> dhcp -> u-root is the circular dependency. Which is fine,
except we messed up modules in u-root big time. We need to break the
circular dep for now.
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
|
|
* Add dhcp Opt 124(vivc) parsing to ztp
Signed-off-by: Hariharakumar Narasimhakumar <hhkumar@fb.com>
|
|
Signed-off-by: Hariharakumar Narasimhakumar <hhkumar@fb.com>
Co-authored-by: Hariharakumar Narasimhakumar <hhkumar@fb.com>
|
|
* Adding xe interface format for Juniper qfx to list of circuit regexes
Signed-off-by: Hariharakumar Narasimhakumar <hhkumar@fb.com>
* Consolidating regexes for et and xe into one entry
Signed-off-by: Hariharakumar Narasimhakumar <hhkumar@fb.com>
Co-authored-by: Hariharakumar Narasimhakumar <hhkumar@fb.com>
|
|
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
Ensure server identifier matches the offer server identifier when
looking for ACKs or NAKs.
This protects against rogue DHCP servers NAKing requests the shouldn't.
Signed-off-by: Moritz Fischer <moritzf@google.com>
Co-authored-by: Moritz Fischer <moritzf@google.com>
|
|
Signed-off-by: Eugene Burkov <e.burkov@adguard.com>
|
|
Signed-off-by: Marcel Moolenaar <mmoolena@amazon.com>
|
|
RFC 2131 gives two examples of padding in section 4.1. For both the
examples, the 'end' option preceeds the padding with the 'pad' option.
These examples relate to overloading pre-defined fields for options,
not padding the packet for a minimal size.
Older versions of the U-Boot firmware support DHCP, but do not expect
the 'pad' option. The firmware logs the following warning for each
occurence:
*** Unhandled DHCP Option in OFFER/ACK: 0
This warning disappears when the 'end' option preceeds the 'pad'
option.
Signed-off-by: Marcel Moolenaar <mmoolena@amazon.com>
|
|
Signed-off-by: yaocw2020 <yaocanwu@gmail.com>
|
|
to raiSubOptionCode
- a few minor changes to address latest feedback
Signed-off-by: Hu Jun <hujun.work@gmail.com>
|
|
|
|
For any users to write their own Discover or Request methods, they need
access to the ifaceHWAddr and serverAddr.
Discovered while trying to move pinterest/bender to nclients.
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
Signed-off-by: Chris Koch <chrisko@google.com>
|
|
- add raiValue in option_relay_agent_information.go to implement fmt.Stringer
- change test cases accordingly
Signed-off-by: Hu Jun <hujun.work@gmail.com>
|