From 985280ec0b2828cbab248c359d3719ad0d5594ee Mon Sep 17 00:00:00 2001
From: behzad nouri <behzadnouri@gmail.com>
Date: Sun, 13 Jun 2021 14:08:08 +0000
Subject: [PATCH] excludes epoch-slots from nodes with unknown or different
 shred version (#17899)

Inspecting TDS gossip table shows that crds values of nodes with
different shred-versions are creeping in. Their epoch-slots are
accumulated in ClusterSlots causing bogus slots very far from current
root which are not purged and so cause ClusterSlots keep consuming more
memory:
https://github.com/solana-labs/solana/issues/17789
https://github.com/solana-labs/solana/issues/14366#issuecomment-769896036
https://github.com/solana-labs/solana/issues/14366#issuecomment-832754654

This commit updates ClusterInfo::get_epoch_slots, and discards entries
from nodes with unknown or different shred-version.

Follow up commits will patch gossip not to waste bandwidth and memory
over crds values of nodes with different shred-version.
---
 gossip/src/cluster_info.rs | 46 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 46 insertions(+)

diff --git a/gossip/src/cluster_info.rs b/gossip/src/cluster_info.rs
index 23026ce68a..3ad0510fb8 100644
--- a/gossip/src/cluster_info.rs
+++ b/gossip/src/cluster_info.rs
@@ -1170,10 +1170,19 @@ impl ClusterInfo {
             .map(map)
     }
 
+    /// Returns epoch-slots inserted since the given cursor.
+    /// Excludes entries from nodes with unkown or different shred version.
     pub fn get_epoch_slots(&self, cursor: &mut Cursor) -> Vec<EpochSlots> {
+        let self_shred_version = self.my_shred_version();
         let gossip = self.gossip.read().unwrap();
         let entries = gossip.crds.get_epoch_slots(cursor);
         entries
+            .filter(
+                |entry| match gossip.crds.get_contact_info(entry.value.pubkey()) {
+                    Some(node) => node.shred_version == self_shred_version,
+                    None => false,
+                },
+            )
             .map(|entry| match &entry.value.data {
                 CrdsData::EpochSlots(_, slots) => slots.clone(),
                 _ => panic!("this should not happen!"),
@@ -3839,6 +3848,43 @@ mod tests {
 
         let slots = cluster_info.get_epoch_slots(&mut cursor);
         assert!(slots.is_empty());
+
+        // Test with different shred versions.
+        let mut rng = rand::thread_rng();
+        let node_pubkey = Pubkey::new_unique();
+        let mut node = ContactInfo::new_rand(&mut rng, Some(node_pubkey));
+        node.shred_version = 42;
+        let epoch_slots = EpochSlots::new_rand(&mut rng, Some(node_pubkey));
+        let entries = vec![
+            CrdsValue::new_unsigned(CrdsData::ContactInfo(node)),
+            CrdsValue::new_unsigned(CrdsData::EpochSlots(0, epoch_slots)),
+        ];
+        {
+            let mut gossip = cluster_info.gossip.write().unwrap();
+            for entry in entries {
+                assert!(gossip.crds.insert(entry, /*now=*/ 0).is_ok());
+            }
+        }
+        // Should exclude other node's epoch-slot because of different
+        // shred-version.
+        let slots = cluster_info.get_epoch_slots(&mut Cursor::default());
+        assert_eq!(slots.len(), 1);
+        assert_eq!(slots[0].from, cluster_info.id);
+        // Match shred versions.
+        {
+            let mut node = cluster_info.my_contact_info.write().unwrap();
+            node.shred_version = 42;
+        }
+        cluster_info.push_self(
+            &HashMap::default(), // stakes
+            None,                // gossip validators
+        );
+        cluster_info.flush_push_queue();
+        // Should now include both epoch slots.
+        let slots = cluster_info.get_epoch_slots(&mut Cursor::default());
+        assert_eq!(slots.len(), 2);
+        assert_eq!(slots[0].from, cluster_info.id);
+        assert_eq!(slots[1].from, node_pubkey);
     }
 
     #[test]