op-batcher: use sync.WaitGroup for fallback-auth watchers

authGroup was an errgroup.Group whose goroutines always returned nil
(failures are reported via receiptsCh, not the group error), so the
error branch in waitForAuthGroup was unreachable. Switch to a plain
sync.WaitGroup, drop the dead error handling, and correct the field
comment: watcher creation is back-pressured by, not hard-bounded by,
MaxPendingTransactions. Also document the receipts-loop-outlives-authGroup
invariant that keeps the final receiptsCh send from blocking.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: palango

op-batcher/batcher/driver.go

@@ -127,11 +127,12 @@ type BatchSubmitter struct {
 
 	// authGroup tracks the fallback batcher's receipt-watcher goroutines (one
 	// per auth+batch pair) so the publishing loop can drain them via
-	// waitForAuthGroup before closing receiptsCh. It is intentionally unbounded:
-	// pending-tx throttling is enforced by the txmgr Queue (queue.Send blocks at
-	// MaxPendingTransactions), and the live watcher count is derived from the
-	// queue's in-flight tx count, so it inherits the same bound.
-	authGroup errgroup.Group
+	// waitForAuthGroup before closing receiptsCh. New watchers are back-pressured
+	// (not hard-bounded) by the txmgr Queue: queue.Send blocks at
+	// MaxPendingTransactions, so watchers are created no faster than txs drain,
+	// though a slow receipts loop can briefly leave more than that parked on their
+	// final receiptsCh send.
+	authGroup sync.WaitGroup
 }
 
 // NewBatchSubmitter initializes the BatchSubmitter driver from a preconfigured DriverSetup

op-batcher/batcher/espresso_driver.go

@@ -1,22 +1,19 @@
 package batcher
 
 import (
-	"context"
-	"errors"
 	"fmt"
 
 	"github.com/ethereum-optimism/optimism/op-service/txmgr"
 )
 
-// waitForAuthGroup blocks until all in-flight fallback-auth submissions have
-// completed. Called from publishingLoop's tail; blocks until killCtx is
-// cancelled if any auth retries are still in flight.
+// waitForAuthGroup blocks until all in-flight fallback-auth watcher goroutines
+// have finished. publishingLoop calls it before closing receiptsCh: each watcher
+// is a sender on receiptsCh, so the receipts loop must still be draining
+// receiptsCh at this point or a watcher's final send would block forever. Each
+// watcher always terminates because the txmgr Queue emits exactly one receipt per
+// Send, even on context cancellation.
 func (l *BatchSubmitter) waitForAuthGroup() {
-	if err := l.authGroup.Wait(); err != nil {
-		if !errors.Is(err, context.Canceled) {
-			l.Log.Error("error waiting for fallback-auth transactions to complete", "err", err)
-		}
-	}
+	l.authGroup.Wait()
 }
 
 // dispatchAuthenticatedSendTx routes sendTx through the fallback-batcher

op-batcher/batcher/fallback_auth.go

@@ -97,10 +97,11 @@ func (l *BatchSubmitter) sendTxWithFallbackAuth(txdata txData, isCancel bool, ca
 	queue.Send(transactionReference, verifyCandidate, authReceiptCh)
 	queue.Send(transactionReference, *candidate, batchReceiptCh)
 
-	l.authGroup.Go(func() error {
+	l.authGroup.Add(1)
+	go func() {
+		defer l.authGroup.Done()
 		l.watchFallbackAuthReceipts(transactionReference, authReceiptCh, batchReceiptCh, receiptsCh)
-		return nil
-	})
+	}()
 }
 
 // watchFallbackAuthReceipts collects the auth and batch receipts for a fallback-auth pair,

op-batcher/batcher/fallback_auth_test.go

@@ -80,7 +80,7 @@ func TestFallbackAuth_OrderingAndSuccess(t *testing.T) {
 	receiptsCh := make(chan txmgr.TxReceipt[txRef], 1)
 
 	l.sendTxWithFallbackAuth(txdata, false, candidate, queue, receiptsCh)
-	require.NoError(t, l.authGroup.Wait())
+	l.authGroup.Wait()
 
 	require.Len(t, queue.sends, 2)
 	// First send must target the BatchAuthenticator (the auth tx), giving it the
@@ -110,7 +110,7 @@ func TestFallbackAuth_AuthFailureRetried(t *testing.T) {
 	receiptsCh := make(chan txmgr.TxReceipt[txRef], 1)
 
 	l.sendTxWithFallbackAuth(txdata, false, candidate, queue, receiptsCh)
-	require.NoError(t, l.authGroup.Wait())
+	l.authGroup.Wait()
 
 	got := <-receiptsCh
 	require.Error(t, got.Err)
@@ -131,7 +131,7 @@ func TestFallbackAuth_BatchFailureRetried(t *testing.T) {
 	receiptsCh := make(chan txmgr.TxReceipt[txRef], 1)
 
 	l.sendTxWithFallbackAuth(txdata, false, candidate, queue, receiptsCh)
-	require.NoError(t, l.authGroup.Wait())
+	l.authGroup.Wait()
 
 	got := <-receiptsCh
 	require.Error(t, got.Err)
@@ -155,7 +155,7 @@ func TestFallbackAuth_AuthRevertedRetried(t *testing.T) {
 	receiptsCh := make(chan txmgr.TxReceipt[txRef], 1)
 
 	l.sendTxWithFallbackAuth(txdata, false, candidate, queue, receiptsCh)
-	require.NoError(t, l.authGroup.Wait())
+	l.authGroup.Wait()
 
 	got := <-receiptsCh
 	require.Error(t, got.Err)
@@ -180,7 +180,7 @@ func TestFallbackAuth_WindowViolationRetried(t *testing.T) {
 	receiptsCh := make(chan txmgr.TxReceipt[txRef], 1)
 
 	l.sendTxWithFallbackAuth(txdata, false, candidate, queue, receiptsCh)
-	require.NoError(t, l.authGroup.Wait())
+	l.authGroup.Wait()
 
 	got := <-receiptsCh
 	require.Error(t, got.Err)
@@ -207,7 +207,7 @@ func TestFallbackAuth_WindowBoundaryAccepted(t *testing.T) {
 	receiptsCh := make(chan txmgr.TxReceipt[txRef], 1)
 
 	l.sendTxWithFallbackAuth(txdata, false, candidate, queue, receiptsCh)
-	require.NoError(t, l.authGroup.Wait())
+	l.authGroup.Wait()
 
 	got := <-receiptsCh
 	require.NoError(t, got.Err)