resolver: a reply marked recursion available from a root server's address stops the lookup (closes #300)

dnswatcher never asks for recursion and a root server never offers it,
so a reply marked recursion available from a root server's address
came from a resolver on the network answering in its place. The first
such reply, a refusal included, now stops the lookup at once with
"this network intercepts DNS queries", naming the address, whether or
not that resolver had the name cached. queryDNS returns a refused reply
along with ErrRefused so the mark can be seen; a refusal without the
mark is handled as before. On such a network dnswatcher and its tests
fail with that error, by design, as the owner ruled on the issue.

Model: opus-5-5

README.md

@@ -561,6 +561,18 @@ This approach ensures:
 - Ability to detect split-horizon or inconsistent responses across authoritative
   servers.
 
+A root server never offers recursion, so a reply marked "recursion available"
+from a root server's address came from a recursive resolver on the network
+answering in its place, as on a network where a router rule, a VPN or a DNS
+filter sends all DNS to one resolver. The first such reply from a root server's
+address, a refusal included, stops the lookup at once with the error "this
+network intercepts DNS queries", naming that address, and no other root server
+is asked. A lookup that every root server refuses without that mark fails with
+the same error. Every lookup starts at the root servers, so on such a network
+every lookup fails with that error, whether or not that resolver has the name
+cached, and dnswatcher's tests fail there too. That is by design: nothing asks
+again another way or otherwise works around the interception.
+
 A watched name's records are stored as its nameservers return them, CNAME
 included. When they return a CNAME and no address, the chain of every CNAME
 target they gave is followed (with a depth limit to prevent loops) to the A and

TODO.md

@@ -19,6 +19,14 @@ milestone: https://git.eeqj.de/sneak/dnswatcher/milestone/7
 
 # Completed Steps
 
+- 2026-10-16: the first reply marked "recursion available" from a root server's
+  address, a refusal included, stops the lookup with "this network intercepts
+  DNS queries", so on a network that sends DNS to a resolver dnswatcher and its
+  tests fail, by design (closes #300).
+- 2026-10-15: a query its caller's context ends before it is sent records no
+  reason about the nameserver, and a try the context's deadline cuts short is a
+  timeout, so a nameserver whose only failures were timeouts is `timeout`, not a
+  network error or `ok` (closes #298).
 - 2026-10-15: the resolver no longer says its query timeout grows on each try;
   every try waits the same 2 seconds, and the unused `timeoutMultiplier` is gone
   (closes #296).

internal/resolver/errors.go

@@ -53,9 +53,13 @@ var (
 		"reply truncated and its retry over TCP failed",
 	)
 
-	// ErrIntercepted is returned when every root server refused a
+	// ErrIntercepted is returned when a root server's address gave a
-	// query. Root servers refuse no query, so the refusals came from
+	// reply marked recursion available, a refusal included, which ends
-	// something on the network answering in their place.
+	// the lookup at once, or when every root server refused a query.
+	// Root servers do neither, so those replies came from something on
+	// the network answering in their place. On such a network
+	// dnswatcher fails with it, by design: nothing works around the
+	// interception.
 	ErrIntercepted = errors.New("this network intercepts DNS queries")
 
 	// ErrCNAMEDepthExceeded is returned when a CNAME chain

internal/resolver/iterative.go

@@ -62,7 +62,10 @@ func checkCtx(ctx context.Context) error {
 }
 
 // tryExchange sends msg to addr over UDP, up to maxRetries times, until
-// a reply comes. Every try waits the same queryTimeoutDuration.
+// a reply comes. Every try waits the same queryTimeoutDuration, or less
+// when ctx's deadline comes first. When ctx has ended before a further
+// try, it returns the error of the try already made, such as a timeout,
+// or ErrContextCanceled when none was made.
 func (r *Resolver) tryExchange(
 	ctx context.Context,
 	msg *dns.Msg,
@@ -74,7 +77,11 @@ func (r *Resolver) tryExchange(
 
 	for range maxRetries {
 		if checkCtx(ctx) != nil {
-			return nil, ErrContextCanceled
+			if err == nil {
+				return nil, ErrContextCanceled
+			}
+
+			return nil, err
 		}
 
 		resp, _, err = r.client.ExchangeContext(ctx, msg, addr)
@@ -108,7 +115,8 @@ func (r *Resolver) retryTCP(
 }
 
 // queryDNS sends a DNS query to a specific server IP, never asking it
-// for recursion. A reply of REFUSED is returned as ErrRefused.
+// for recursion. A reply of REFUSED is returned together with
+// ErrRefused, so that the caller can still read the reply's flags.
 func (r *Resolver) queryDNS(
 	ctx context.Context,
 	serverIP string,
@@ -132,7 +140,7 @@ func (r *Resolver) queryDNS(
 	}
 
 	if resp.Rcode == dns.RcodeRefused {
-		return nil, fmt.Errorf(
+		return resp, fmt.Errorf(
 			"query %s @%s: %w", name, serverIP, ErrRefused,
 		)
 	}
@@ -277,10 +285,15 @@ func shuffled(
 }
 
 // queryServers asks servers, the servers of zone, about name in a random
-// order until one gives a usable reply. A server that times out, refuses
+// order until one gives a usable reply. When they are the root servers,
-// or gives a reply that is not usable is passed over for the next. When
+// the first reply marked recursion available, a refusal included, ends
-// every server refused, the error says so, and when they are the root
+// the lookup with ErrIntercepted, naming the address it came from, and
-// servers it is ErrIntercepted.
+// no other root server is asked: root servers offer no recursion, so
+// that reply came from something on the network answering in their
+// place. Otherwise a server that times out, refuses or gives a reply
+// that is not usable is passed over for the next. When every server
+// refused, the error says so, and when they are the root servers it is
+// ErrIntercepted.
 func (r *Resolver) queryServers(
 	ctx context.Context,
 	servers []string,
@@ -297,7 +310,17 @@ func (r *Resolver) queryServers(
 			return nil, ErrContextCanceled
 		}
 
+		// queryDNS returns a refusal together with its reply, so a
+		// refusal's mark is checked here too.
 		resp, err := r.queryDNS(ctx, ip, name, qtype)
+		if resp != nil && zone == "." && resp.RecursionAvailable {
+			return nil, fmt.Errorf(
+				"query %s @%s: reply marked recursion available, "+
+					"so not from a root server: %w",
+				name, ip, ErrIntercepted,
+			)
+		}
+
 		if err == nil && !usableReply(resp, zone, name) {
 			err = fmt.Errorf(
 				"query %s @%s: %w", name, ip, ErrUnusableReply,
@@ -863,6 +886,10 @@ func (r *Resolver) querySingleType(
 			if err != nil {
 				err = fmt.Errorf("query %s @%s: %w", hostname, nsIP, err)
 			}
+		case errors.Is(err, ErrContextCanceled):
+			// The caller's context ended before the query was sent.
+			// That says nothing about the nameserver, so no reason is
+			// recorded.
 		case isTimeout(err):
 			state.gotTimeout = true
 		case errors.Is(err, ErrRefused):

internal/resolver/iterative_internal_test.go

@@ -1,6 +1,8 @@
 package resolver
 
 import (
+	"context"
+	"log/slog"
 	"strconv"
 	"syscall"
 	"testing"
@@ -132,3 +134,33 @@ func TestReadReply(t *testing.T) {
 		})
 	}
 }
+
+// TestQuerySingleType_ContextEnded asks a nameserver about one record
+// type with a context that has already ended, so no query is sent. A
+// query the context ended before it was sent says nothing about the
+// nameserver and records no reason: a nameserver whose earlier query
+// timed out is still reported as timed out, not as a network error.
+func TestQuerySingleType_ContextEnded(t *testing.T) {
+	t.Parallel()
+
+	r := NewFromLogger(slog.New(slog.DiscardHandler))
+
+	ctx, cancel := context.WithCancel(t.Context())
+	cancel()
+
+	resp := &NameserverResponse{Records: map[string][]string{}, Status: StatusOK}
+
+	// The query for an earlier record type timed out.
+	state := queryState{gotTimeout: true}
+
+	_, err := r.querySingleType(
+		ctx, []string{"192.0.2.1"}, "example.com.", dns.TypeA, resp, &state,
+	)
+	require.ErrorIs(t, err, ErrContextCanceled)
+	assert.Equal(t, queryState{gotTimeout: true}, state)
+
+	classifyResponse(resp, state)
+
+	assert.Equal(t, StatusTimeout, resp.Status)
+	assert.Equal(t, "all queries timed out", resp.Error)
+}

internal/resolver/resolver_test.go

@@ -850,6 +850,74 @@ func TestQueryServers_NotEveryRootServerRefused(t *testing.T) {
 	require.NotContains(t, err.Error(), "refused a query for")
 }
 
+// TestQueryServers_RecursiveResolversAsRootServers passes public
+// recursive resolvers to QueryServers as the servers of the root zone,
+// as a network that sends every DNS query to a resolver would. Their
+// replies are marked recursion available, which no root server's is:
+// SERVFAIL, REFUSED, or an answer from a resolver's cache, as 1.1.1.1
+// gives for com., which would be used if a root server gave it. The
+// first such reply ends the lookup with ErrIntercepted, naming the
+// resolver that gave it. In the first group the replies differ; every
+// resolver in the second refuses a query that does not ask for
+// recursion, as 9.9.9.9 does, so there a refusal ends the lookup
+// instead of being passed over. Each group has resolvers of four
+// operators, and a resolver that gives no reply is passed over for the
+// next, so one of them answering is enough.
+func TestQueryServers_RecursiveResolversAsRootServers(t *testing.T) {
+	t.Parallel()
+
+	r := newTestResolver(t)
+	groups := [][]string{
+		{"1.1.1.1", "8.8.8.8", "9.9.9.9", "208.67.222.222"},
+		{"64.6.64.6", "185.222.222.222", "4.2.2.1", "94.140.14.14"},
+	}
+
+	for _, resolvers := range groups {
+		for _, name := range []string{"com.", "desec.io."} {
+			var err error
+
+			livednstest.Retry(
+				t,
+				"QueryServers("+strings.Join(resolvers, ", ")+
+					" as root servers, "+name+")",
+				func(ctx context.Context) error {
+					_, err = r.QueryServers(
+						ctx, resolvers, ".", name, dns.TypeNS,
+					)
+
+					// When no resolver gave a reply marked recursion
+					// available, none may have replied at all, so the
+					// attempt is tried again.
+					if err != nil && !strings.Contains(
+						err.Error(), "marked recursion available",
+					) {
+						return fmt.Errorf(
+							"%w: %w", livednstest.ErrNoAnswer, err,
+						)
+					}
+
+					return nil
+				},
+			)
+
+			require.ErrorIs(t, err, resolver.ErrIntercepted)
+
+			// The error names the resolver whose reply ended the
+			// lookup.
+			messages := make([]string, 0, len(resolvers))
+			for _, ip := range resolvers {
+				messages = append(messages,
+					"query "+name+" @"+ip+": reply marked recursion "+
+						"available, so not from a root server: "+
+						"this network intercepts DNS queries",
+				)
+			}
+
+			require.Contains(t, messages, err.Error())
+		}
+	}
+}
+
 func TestQueryNameserver_RecordsSorted(t *testing.T) {
 	t.Parallel()
 
@@ -1380,10 +1448,10 @@ func TestQueryNameserverIP_Timeout(t *testing.T) {
 	r := newTestResolver(t)
 
 	// Nothing answers at 192.0.2.1, a documentation address. The
-	// resolver tries each query twice, and the first try gives up
+	// resolver tries each query twice, and each try gives up after two
-	// after two seconds. A deadline that ends during the first try
+	// seconds. This deadline ends during the second try of the first
-	// makes the status vary from run to run between error and
+	// record type; TestQueryNameserverIP_DeadlineDuringFirstTry covers
-	// timeout, so the deadline must outlast the first try.
+	// one that ends during the first.
 	ctx, cancel := context.WithTimeout(
 		context.Background(), 3*time.Second,
 	)
@@ -1399,6 +1467,62 @@ func TestQueryNameserverIP_Timeout(t *testing.T) {
 	assert.NotEmpty(t, resp.Error)
 }
 
+// TestQueryNameserverIP_DeadlineDuringFirstTry gives a query to
+// 192.0.2.1, where nothing answers, a deadline that ends during its
+// first try. That try ends at the deadline as a timeout, so the
+// nameserver is reported as timed out, not as answering.
+func TestQueryNameserverIP_DeadlineDuringFirstTry(t *testing.T) {
+	t.Parallel()
+
+	r := newTestResolver(t)
+
+	// Well under the two seconds one try waits.
+	ctx, cancel := context.WithTimeout(
+		context.Background(), 500*time.Millisecond,
+	)
+	t.Cleanup(cancel)
+
+	resp, err := r.QueryNameserverIP(
+		ctx, "unreachable.test.", "192.0.2.1",
+		"example.com",
+	)
+	require.NoError(t, err)
+
+	assert.Equal(t, resolver.StatusTimeout, resp.Status)
+	assert.NotEmpty(t, resp.Error)
+}
+
+// TestQueryNameserverIP_CancelDuringFirstTry cancels the context while
+// the first try of a query to 192.0.2.1, where nothing answers, waits
+// for a reply. That try's timeout is all that is learned about the
+// nameserver, so it is reported as timed out, not as answering.
+func TestQueryNameserverIP_CancelDuringFirstTry(t *testing.T) {
+	t.Parallel()
+
+	r := newTestResolver(t)
+
+	// The first try starts at once and waits two seconds for a reply;
+	// the cancel comes a second in. The DNS client cuts a try short at
+	// the context's deadline but not at a cancel, so the try still
+	// waits its full two seconds and ends as a timeout a second after
+	// the cancel. By then the context has certainly ended, so no second
+	// try is made. With a deadline instead, the try ends at the moment
+	// the context does, and whether a second try is made depends on
+	// which the resolver sees first.
+	ctx, cancel := context.WithCancel(context.Background())
+	t.Cleanup(cancel)
+	time.AfterFunc(time.Second, cancel)
+
+	resp, err := r.QueryNameserverIP(
+		ctx, "unreachable.test.", "192.0.2.1",
+		"example.com",
+	)
+	require.NoError(t, err)
+
+	assert.Equal(t, resolver.StatusTimeout, resp.Status)
+	assert.NotEmpty(t, resp.Error)
+}
+
 // TestQueryNameserverIP_CancelledLogsNothing cancels the context while
 // a query to 192.0.2.1, where nothing answers, is waiting for a reply,
 // as shutdown does. The query was cut short, not failed, so nothing is