MLD support - #36
Open
zeeshanlakhani wants to merge 6 commits into
Open
Conversation
RFC 3376 §4.2.6 gives Aux Data Len in 32-bit words, but the `var_len` expression used it as bytes, so a record claiming two words consumed two octets and desynchronised every record after it. Conforming senders always set zero (§4.2.10), so this bites only on non-conforming input, where a snooper and the guest stack then disagree about the same packet. Also, we add the missing §4.2.12 record types `ALLOW_NEW_SOURCES` and `BLOCK_OLD_SOURCES`.
`Result::expect` takes its message by value, so the generated `.expect(&alloc::format!(...))` allocated a String on every owned emit, success included. The string has no interpolation, so no format is necessary. Also, we now document the caller-owns-the-length-field invariant on the `var_len` docs.
MLD constrains fields to multicast addresses. `MulticastIpv6Addr` wraps an `Ipv6Addr` proven multicast, in the `CRStr` shape. Refinement lives in constructor signatures only. The derive's Ref traits share the field type between owned and zero-copy forms, so a refined field would make a view assert an invariant its bytes were never checked against. A received address stays `Ipv6Addr` until refined through `new`. `NotMulticastError` keeps a generic message so a v4 counterpart can share it.
This adds MLDv1 (RFC 2710 §3) and the MLDv2 Query, Report and Multicast Address
Record (RFC 3810 §5.1, §5.2) types, modelled on the igmp module.
MLDv1 and MLDv2 Queries share ICMPv6 type 130 and are told apart by length
(RFC 3810 §8.1), so a v2 `TooSmall` means retry as v1, not a malformed
packet. `MldV1` builds through general_query, address_specific_query,
report and done rather than `Default`, since type zero is RESERVED. The
group-taking constructors require `MulticastIpv6Addr`, to be more
specific. Decoders cover the §5.1.3 and §5.1.9 floating-point Max Resp
Code and QQIC.
Also, we rename `MULTICAST_LISTENER_{REPORT,DONE}` to the `V1_`-prefixed forms,
matching `IgmpMessageType`. This is why we move to v3.
`RepeatedView` previously parsed until buffer exhaustion, ignoring the protocol's count field. That conflicts with RFC 3810 §5.2.11 and RFC 3376 §4.2.11, which allow additional data after the last record inside the IP payload. The count is the only reliable way to distinguish records from that additional data. For example, an empty IGMPv3 Report is 46 octets on the wire. Ethernet padding extends the frame to the 60-octet minimum, and exhaustion parsing interpreted that padding as a record and failed the parse. This adds `RepeatedView::parse_counted`, driven by the opt-in `#[ingot(subparse(count = "<expr>"))]` attribute. Headers without the attribute retain their existing generated behavior. A short count or a non-advancing element is treated as an error. `Unwanted` maps to `IllegalValue` so an enclosing choice cannot interpret an inner malformed count as "not my header". Zero-width elements are rejected explicitly. ```
`validate()` on the owned types checks what the RFCs constrain: the type matches the message, groups are multicast (or unspecified in a General Query, §5.1.5), sources are unicast (§5.1.11, §5.2.9) with `::` excluded per RFC 4291 §2.5.2, and a General Query carries no sources at all (§5.1.13). Loopback stays admissible, as rejecting it is forwarding policy.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR adds MLD support for RFC 2710 and RFC 3810, with related type and parsing fixes.
This includes:
alloc::format!from derivedEmit.