Historically, the VERSION message contains an "addrMe" and an "addrYou". As
these are sent before version negotiation is complete, the protocol version is
INIT_PROTO_VERSION (209), and in that protocol, CAddress is serialized without
nTime.
This is in fact the only situation left where a CAddress is (de)serialized
without nTime. As it's such a simple structure (CService for ip/port + uint64_t
for nServices), just inline that structure in the few places where it occurs,
and remove the logic for dealing with missing nTime from CAddress.
Before this commit, CAddress disk serialization was messy. It stored
CLIENT_VERSION in the first 4 bytes, optionally OR'ed with ADDRV2_FORMAT.
- All bits except ADDRV2_FORMAT were ignored, making it hard to use for actual
future format changes.
- ADDRV2_FORMAT determines whether or not nServices is serialized in LE64
format or in CompactSize format.
- Whether or not the embedded CService is serialized in V1 or V2 format is
determined by the stream's version having ADDRV2_FORMAT (as opposed to the
nServices encoding, which is determined by the disk version).
To improve the situation, this commit introduces the following disk
serialization format, compatible with earlier versions, but better defined for
future changes:
- The first 4 bytes store a format version number. Its low 19 bits are ignored
(as it historically stored the CLIENT_VERSION), but its high 13 bits specify
the serialization exactly:
- 0x00000000: LE64 encoding for nServices, V1 encoding for CService
- 0x20000000: CompactSize encoding for nServices, V2 encoding for CService
- Any other value triggers an unsupported format error on deserialization,
and can be used for future format changes.
- The ADDRV2_FORMAT flag in the stream's version does not impact the actual
serialization format; it only determines whether V2 encoding is permitted;
whether it's actually enabled depends solely on the disk version number.
Operationally the changes to the deserializer are:
- Failure when the stored format version number is unexpected.
- The embedded CService's format is determined by the stored format version
number rather than the stream's version number.
These do no introduce incompatibilities, as no code versions exist that write
any value other than 0 or 0x20000000 in the top 13 bits, and no code paths
where the stream's version differs from the stored version.
Introduce a new message `sendaddrv2` to signal support for ADDRv2.
Send the new message immediately after sending the `VERACK` message.
Add support for receiving and parsing ADDRv2 messages.
Send ADDRv2 messages (instead of ADDR) to a peer if he has
advertised support for it.
Co-authored-by: Carl Dong <contact@carldong.me>
Change the serialization of `CAddrMan` to serialize its addresses
in ADDRv2/BIP155 format by default. Introduce a new `CAddrMan` format
version (3).
Add support for ADDRv2 format in `CAddress` (un)serialization.
Co-authored-by: Carl Dong <contact@carldong.me>
This moves header size and netmagic checking out of net_processing and
into net. This check now runs in ReadHeader, so that net can exit early
out of receiving bytes from the peer. IsValid is now slimmed down, so
it no longer needs a MessageStartChars& parameter.
Additionally this removes the rest of the m_valid_* members from
CNetMessage.
This commit removes the single-parameter contructor of CMessageHeader
and replaces it with a default constructor.
The single parameter contructor isn't used anywhere except for tests.
There is no reason to initialize a CMessageHeader with a particular
messagestart. This messagestart should always be replaced when
deserializing an actual message header so that we can run checks on it.
The default constructor initializes it to zero, just like the command
and checksum.
This also removes a parameter of a V1TransportDeserializer constructor,
as it was only used for this purpose.
fixes issue #19678 UBSan implicit-integer-sign-change
Credit to Eugene (Crypt-iQ) for finding and reporting the issue
and to Vasil Dimov (vasild) for the original suggestion
