Merge #6239: feat: increase the number of block headers able to be do…

…wnloaded at once to 8000 in protocol version `70235`

48c7f98 doc: drop trailing whitespace (pasta)
697743d test: add missing import (UdjinM6)
cfe99fd docs: add release notes for 6239 (pasta)
a6bbaac fix: GetHeadersLimit is used for getheaders(2) and headers(2), refactor it to accept `compressed` instead of `msg_type` (UdjinM6)
b224f3f bump p2p_version in tests (PastaPastaPasta)
b423f42 refactor: sort imports (UdjinM6)
f6c68ba refactor: simplify _compute_requested_block_headers (UdjinM6)
07876b2 use `MAX_HEADERS_UNCOMPRESSED_RESULT` not `MAX_HEADERS_UNCOMPRESSED_RESULTS` ; use `MAX_HEADERS_UNCOMPRESSED_RESULT` in RPC to avoid breaking changes (pasta)
b137280 change to _COMPRESSED or _UNCOMPRESSED (pasta)
303bc7a fix: increase it for headers2 only (UdjinM6)
e23410f trivial: rename `MAX_HEADERS_RESULTS_NEW` to `MAX_HEADERS_RESULTS` (Kittywhiskers Van Gogh)
bcf0320 trivial: move the headers limit determination to `GetHeadersLimit()` (Kittywhiskers Van Gogh)
993c7c0 feat: increase the number of block headers able to be downloaded at once to 8000 in protocol version `70234` (pasta)

Pull request description:

  ## Issue being fixed or feature implemented
  We did some testing quite a while ago that found that sending 8000 headers at a time could speed stuff up. But we wanted to wait until compressed headers were implemented. Well, they've been implemented!

  ## What was done?
  Bump 2000 -> 8000 triggered by protocol version

  ## How Has This Been Tested?
  Hasn't, we should setup a few nodes running this and sync them from each other

  ## Breaking Changes
  New protocol version, not breaking but should add notes? I should probably add release notes

  ## Checklist:
    _Go over all the following points, and put an `x` in all the boxes that apply._
  - [ ] I have performed a self-review of my own code
  - [ ] I have commented my code, particularly in hard-to-understand areas
  - [ ] I have added or updated relevant unit/integration/functional/e2e tests
  - [ ] I have made corresponding changes to the documentation
  - [x] I have assigned this pull request to a milestone _(for repository code-owners and collaborators only)_

ACKs for top commit:
  UdjinM6:
    light ACK 48c7f98
  knst:
    utACK 48c7f98

Tree-SHA512: 54c68b9496131ab7f32504d44398d776a151df809d0120d093bbabb18904a783bd9b58796820209f5d75552df5476e30eaa09d68f7c5057882f94b5766a64f4c

doc/release-notes-6239.md

@@ -0,0 +1,5 @@
+P2P and Network Changes
+-----------------------
+
+The max number of compressed block headers which can be requested at once has been increased from 2000 to 8000. This
+change activates with the protocol version `70235` and only applies to compressed block headers.

src/net_processing.cpp

@@ -1041,6 +1041,15 @@ static bool IsLimitedPeer(const Peer& peer)
              (peer.m_their_services & NODE_NETWORK_LIMITED));
 }
 
+/** Get maximum number of headers that can be included in one batch */
+static uint16_t GetHeadersLimit(const CNode& pfrom, bool compressed)
+{
+    if (pfrom.GetCommonVersion() >= INCREASE_MAX_HEADERS2_VERSION && compressed) {
+        return MAX_HEADERS_COMPRESSED_RESULT;
+    }
+    return MAX_HEADERS_UNCOMPRESSED_RESULT;
+}
+
 static void PushInv(Peer& peer, const CInv& inv)
 {
     // Dash always initializes m_tx_relay
@@ -2963,16 +2972,17 @@ void PeerManagerImpl::ProcessHeadersMessage(CNode& pfrom, Peer& peer,
     }
 
     // Consider fetching more headers.
-    if (nCount == MAX_HEADERS_RESULTS) {
-        std::string msg_type = UsesCompressedHeaders(peer) ? NetMsgType::GETHEADERS2 : NetMsgType::GETHEADERS;
+    const bool uses_compressed = UsesCompressedHeaders(peer);
+    const std::string msg_type = uses_compressed ? NetMsgType::GETHEADERS2 : NetMsgType::GETHEADERS;
+    if (nCount == GetHeadersLimit(pfrom, uses_compressed)) {
         // Headers message had its maximum size; the peer may have more headers.
         if (MaybeSendGetHeaders(pfrom, msg_type, m_chainman.ActiveChain().GetLocator(pindexLast), peer)) {
             LogPrint(BCLog::NET, "more %s (%d) to end to peer=%d (startheight:%d)\n",
                     msg_type, pindexLast->nHeight, pfrom.GetId(), peer.m_starting_height);
         }
     }
 
-    UpdatePeerStateForReceivedHeaders(pfrom, pindexLast, received_new_header, nCount == MAX_HEADERS_RESULTS);
+    UpdatePeerStateForReceivedHeaders(pfrom, pindexLast, received_new_header, nCount == GetHeadersLimit(pfrom, uses_compressed));
 
     // Consider immediately downloading blocks.
     HeadersDirectFetchBlocks(pfrom, peer, pindexLast);
@@ -4066,8 +4076,8 @@ void PeerManagerImpl::ProcessMessage(
                 pindex = m_chainman.ActiveChain().Next(pindex);
         }
 
-        const auto send_headers = [this /* for m_connman */, &hashStop, &pindex, &nodestate, &pfrom, &msgMaker](auto msg_type, auto& v_headers, auto callback) {
-            int nLimit = MAX_HEADERS_RESULTS;
+        const auto send_headers = [this /* for m_connman */, &hashStop, &pindex, &nodestate, &pfrom, &msgMaker](auto msg_type_internal, auto& v_headers, auto callback) {
+            int nLimit = GetHeadersLimit(pfrom, msg_type_internal == NetMsgType::HEADERS2);
             for (; pindex; pindex = m_chainman.ActiveChain().Next(pindex)) {
                 v_headers.push_back(callback(pindex));
 
@@ -4087,7 +4097,7 @@ void PeerManagerImpl::ProcessMessage(
             // will re-announce the new block via headers (or compact blocks again)
             // in the SendMessages logic.
             nodestate->pindexBestHeaderSent = pindex ? pindex : m_chainman.ActiveChain().Tip();
-            m_connman.PushMessage(&pfrom, msgMaker.Make(msg_type, v_headers));
+            m_connman.PushMessage(&pfrom, msgMaker.Make(msg_type_internal, v_headers));
         };
 
         LogPrint(BCLog::NET, "%s %d to %s from peer=%d\n", msg_type, (pindex ? pindex->nHeight : -1), hashStop.IsNull() ? "end" : hashStop.ToString(), pfrom.GetId());
@@ -4575,7 +4585,7 @@ void PeerManagerImpl::ProcessMessage(
 
         // Bypass the normal CBlock deserialization, as we don't want to risk deserializing 2000 full blocks.
         unsigned int nCount = ReadCompactSize(vRecv);
-        if (nCount > MAX_HEADERS_RESULTS) {
+        if (nCount > GetHeadersLimit(pfrom, msg_type == NetMsgType::HEADERS2)) {
             Misbehaving(pfrom.GetId(), 20, strprintf("headers message size = %u", nCount));
             return;
         }
@@ -6010,4 +6020,4 @@ bool PeerManagerImpl::SendMessages(CNode* pto)
         }
     } // release cs_main
     return true;
-}
+}

src/rpc/blockchain.cpp

@@ -997,7 +997,7 @@ static RPCHelpMan getblockheaders()
         "If verbose is true, each item is an Object with information about a single blockheader.\n",
         {
             {"blockhash", RPCArg::Type::STR_HEX, RPCArg::Optional::NO, "The block hash"},
-            {"count", RPCArg::Type::NUM, /* default */ strprintf("%s", MAX_HEADERS_RESULTS), ""},
+            {"count", RPCArg::Type::NUM, /* default */ strprintf("%s", MAX_HEADERS_UNCOMPRESSED_RESULT), ""},
             {"verbose", RPCArg::Type::BOOL, /* default */ "true", "true for a json object, false for the hex-encoded data"},
         },
         {
@@ -1055,11 +1055,11 @@ static RPCHelpMan getblockheaders()
         throw JSONRPCError(RPC_INVALID_ADDRESS_OR_KEY, "Block not found");
     }
 
-    int nCount = MAX_HEADERS_RESULTS;
+    int nCount = MAX_HEADERS_UNCOMPRESSED_RESULT;
     if (!request.params[1].isNull())
         nCount = request.params[1].get_int();
 
-    if (nCount <= 0 || nCount > (int)MAX_HEADERS_RESULTS)
+    if (nCount <= 0 || nCount > (int)MAX_HEADERS_UNCOMPRESSED_RESULT)
         throw JSONRPCError(RPC_INVALID_PARAMETER, "Count is out of range");
 
     bool fVerbose = true;
@@ -1135,7 +1135,7 @@ static RPCHelpMan getmerkleblocks()
         {
             {"filter", RPCArg::Type::STR_HEX, RPCArg::Optional::NO, "The hex-encoded bloom filter"},
             {"blockhash", RPCArg::Type::STR_HEX, RPCArg::Optional::NO, "The block hash"},
-            {"count", RPCArg::Type::NUM, /* default */ strprintf("%s", MAX_HEADERS_RESULTS), ""},
+            {"count", RPCArg::Type::NUM, /* default */ strprintf("%s", MAX_HEADERS_UNCOMPRESSED_RESULT), ""},
         },
         RPCResult{
                 RPCResult::Type::ARR, "", "",
@@ -1164,11 +1164,11 @@ static RPCHelpMan getmerkleblocks()
         throw JSONRPCError(RPC_INVALID_ADDRESS_OR_KEY, "Block not found");
     }
 
-    int nCount = MAX_HEADERS_RESULTS;
+    int nCount = MAX_HEADERS_UNCOMPRESSED_RESULT;
     if (!request.params[2].isNull())
         nCount = request.params[2].get_int();
 
-    if (nCount <= 0 || nCount > (int)MAX_HEADERS_RESULTS) {
+    if (nCount <= 0 || nCount > (int)MAX_HEADERS_UNCOMPRESSED_RESULT) {
         throw JSONRPCError(RPC_INVALID_PARAMETER, "Count is out of range");
     }
 

src/validation.h

@@ -85,7 +85,8 @@ static const int MAX_SCRIPTCHECK_THREADS = 15;
 static const int DEFAULT_SCRIPTCHECK_THREADS = 0;
 /** Number of headers sent in one getheaders result. We rely on the assumption that if a peer sends
  *  less than this number, we reached its tip. Changing this value is a protocol upgrade. */
-static const unsigned int MAX_HEADERS_RESULTS = 2000;
+static const unsigned int MAX_HEADERS_UNCOMPRESSED_RESULT = 2000;
+static const unsigned int MAX_HEADERS_COMPRESSED_RESULT = 8000;
 
 static const int64_t DEFAULT_MAX_TIP_AGE = 6 * 60 * 60; // ~144 blocks behind -> 2 x fork detection time, was 24 * 60 * 60 in bitcoin
 

src/version.h

@@ -11,7 +11,7 @@
  */
 
 
-static const int PROTOCOL_VERSION = 70234;
+static const int PROTOCOL_VERSION = 70235;
 
 //! initial proto version, to be increased after version/verack negotiation
 static const int INIT_PROTO_VERSION = 209;
@@ -58,6 +58,9 @@ static const int NO_LEGACY_ISLOCK_PROTO_VERSION = 70231;
 //! Inventory type for DSQ messages added
 static const int DSQ_INV_VERSION = 70234;
 
+//! Maximum header count for HEADRES2 message was increased from 2000 to 8000 in this version
+static const int INCREASE_MAX_HEADERS2_VERSION = 70235;
+
 // Make sure that none of the values above collide with `ADDRV2_FORMAT`.
 
 #endif // BITCOIN_VERSION_H

test/functional/p2p_invalid_messages.py

@@ -10,11 +10,13 @@
     CInv,
     msg_ping,
     ser_string,
-    MAX_HEADERS_RESULTS,
+    MAX_HEADERS_COMPRESSED_RESULT,
+    MAX_HEADERS_UNCOMPRESSED_RESULT,
     MAX_INV_SIZE,
     MAX_PROTOCOL_MESSAGE_LENGTH,
     msg_getdata,
     msg_headers,
+    msg_headers2,
     msg_inv,
     MSG_TX,
     msg_version,
@@ -153,9 +155,13 @@ def test_oversized_getdata_msg(self):
         self.test_oversized_msg(msg_getdata([CInv(MSG_TX, 1)] * size), size)
 
     def test_oversized_headers_msg(self):
-        size = MAX_HEADERS_RESULTS + 1
+        size = MAX_HEADERS_UNCOMPRESSED_RESULT + 1
         self.test_oversized_msg(msg_headers([CBlockHeader()] * size), size)
 
+    def test_oversized_headers2_msg(self):
+        size = MAX_HEADERS_COMPRESSED_RESULT + 1
+        self.test_oversized_msg(msg_headers2([CBlockHeader()] * size), size)
+
     def test_resource_exhaustion(self):
         self.log.info("Test node stays up despite many large junk messages")
         conn = self.nodes[0].add_p2p_connection(P2PDataStore())

test/functional/test_framework/messages.py

@@ -42,7 +42,8 @@
 BIP125_SEQUENCE_NUMBER = 0xfffffffd  # Sequence number that is BIP 125 opt-in and BIP 68-opt-out
 
 MAX_PROTOCOL_MESSAGE_LENGTH = 3 * 1024 * 1024  # Maximum length of incoming protocol messages
-MAX_HEADERS_RESULTS = 2000  # Number of headers sent in one getheaders result
+MAX_HEADERS_UNCOMPRESSED_RESULT = 2000  # Number of headers sent in one getheaders result
+MAX_HEADERS_COMPRESSED_RESULT = 8000  # Number of headers2 sent in one getheaders2 result
 MAX_INV_SIZE = 50000  # Maximum number of entries in an 'inv' protocol message
 
 NODE_NETWORK = (1 << 0)

test/functional/test_framework/p2p.py

@@ -31,7 +31,8 @@
 from test_framework.messages import (
     CBlockHeader,
     CompressibleBlockHeader,
-    MAX_HEADERS_RESULTS,
+    MAX_HEADERS_COMPRESSED_RESULT,
+    MAX_HEADERS_UNCOMPRESSED_RESULT,
     NODE_HEADERS_COMPRESSED,
     msg_addr,
     msg_addrv2,
@@ -91,8 +92,8 @@
 # The minimum P2P version that this test framework supports
 MIN_P2P_VERSION_SUPPORTED = 60001
 # The P2P version that this test framework implements and sends in its `version` message
-# Version 70231 drops supports for legacy InstantSend locks
-P2P_VERSION = 70231
+# Version 70235 increased max header count for HEADERS2 message from 2000 to 8000
+P2P_VERSION = 70235
 # The services that this test framework offers in its `version` message
 P2P_SERVICES = NODE_NETWORK | NODE_HEADERS_COMPRESSED
 # The P2P user agent string that this test framework sends in its `version` message
@@ -712,7 +713,7 @@ def on_getdata(self, message):
             else:
                 logger.debug('getdata message type {} received.'.format(hex(inv.type)))
 
-    def _compute_requested_block_headers(self, locator, hash_stop):
+    def _compute_requested_block_headers(self, locator, hash_stop, max_results):
         # Assume that the most recent block added is the tip
         if not self.block_store:
             return
@@ -733,14 +734,14 @@ def _compute_requested_block_headers(self, locator, hash_stop):
                 break
 
         # Truncate the list if there are too many headers
-        headers_list = headers_list[:-MAX_HEADERS_RESULTS - 1:-1]
+        headers_list = headers_list[:-max_results - 1:-1]
 
         return headers_list
 
     def on_getheaders2(self, message):
         """Search back through our block store for the locator, and reply with a compressed headers message if found."""
 
-        headers_list = self._compute_requested_block_headers(message.locator, message.hashstop)
+        headers_list = self._compute_requested_block_headers(message.locator, message.hashstop, MAX_HEADERS_COMPRESSED_RESULT)
         compressible_headers_list = [CompressibleBlockHeader(h) for h in headers_list] if headers_list else None
         response = msg_headers2(compressible_headers_list)
 
@@ -750,7 +751,7 @@ def on_getheaders2(self, message):
     def on_getheaders(self, message):
         """Search back through our block store for the locator, and reply with a headers message if found."""
 
-        headers_list = self._compute_requested_block_headers(message.locator, message.hashstop)
+        headers_list = self._compute_requested_block_headers(message.locator, message.hashstop, MAX_HEADERS_UNCOMPRESSED_RESULT)
         response = msg_headers(headers_list)
 
         if response is not None: