From a5425969edfdb895021a287e62f25350d86ca720 Mon Sep 17 00:00:00 2001
From: pptx704 <rafeed@omukk.dev>
Date: Sun, 17 May 2026 01:47:56 +0600
Subject: [PATCH] fix: assorted bug fixes for CH migration

Fix resource leaks, race conditions, and error handling across host
agent and control plane: proper sparse file cleanup on close error,
connect error wrapping for MakeDir, CoW file cleanup on pause failure,
per-sandbox VM directories, deferred map deletion to avoid race in VM
destroy, and goroutine launch for extension background workers.
---
 internal/devicemapper/devicemapper.go | 6 +++++-
 internal/hostagent/server.go          | 6 +++++-
 internal/sandbox/manager.go           | 1 +
 internal/vm/config.go                 | 2 +-
 internal/vm/manager.go                | 5 ++++-
 pkg/cpserver/run.go                   | 2 +-
 6 files changed, 17 insertions(+), 5 deletions(-)

diff --git a/internal/devicemapper/devicemapper.go b/internal/devicemapper/devicemapper.go
index f65ec6a..f53b109 100644
--- a/internal/devicemapper/devicemapper.go
+++ b/internal/devicemapper/devicemapper.go
@@ -379,5 +379,9 @@ func createSparseFile(path string, sizeBytes int64) error {
 		os.Remove(path)
 		return err
 	}
-	return f.Close()
+	if err := f.Close(); err != nil {
+		os.Remove(path)
+		return err
+	}
+	return nil
 }
diff --git a/internal/hostagent/server.go b/internal/hostagent/server.go
index 816a99f..faf6424 100644
--- a/internal/hostagent/server.go
+++ b/internal/hostagent/server.go
@@ -301,7 +301,7 @@ func (s *Server) MakeDir(
 
 	resp, err := client.MakeDir(ctx, msg.Path)
 	if err != nil {
-		return nil, fmt.Errorf("make dir: %w", err)
+		return nil, connect.NewError(connect.CodeInternal, fmt.Errorf("make dir: %w", err))
 	}
 
 	return connect.NewResponse(&pb.MakeDirResponse{
@@ -373,6 +373,8 @@ func (s *Server) ExecStream(
 					Error:    ev.Error,
 				},
 			}
+		default:
+			continue
 		}
 		if err := stream.Send(&resp); err != nil {
 			return err
@@ -889,6 +891,8 @@ func (s *Server) ConnectProcess(
 					Error:    ev.Error,
 				},
 			}
+		default:
+			continue
 		}
 		if err := stream.Send(&resp); err != nil {
 			return err
diff --git a/internal/sandbox/manager.go b/internal/sandbox/manager.go
index 8b3f85d..9aed83f 100644
--- a/internal/sandbox/manager.go
+++ b/internal/sandbox/manager.go
@@ -541,6 +541,7 @@ func (m *Manager) Pause(ctx context.Context, sandboxID string) error {
 			warnErr("snapshot dir cleanup error", sandboxID, os.RemoveAll(pauseDir))
 			warnErr("network cleanup error during pause", sandboxID, network.RemoveNetwork(sb.slot))
 			m.slots.Release(sb.SlotIndex)
+			os.Remove(sb.dmDevice.CowPath)
 			if sb.baseImagePath != "" {
 				m.loops.Release(sb.baseImagePath)
 			}
diff --git a/internal/vm/config.go b/internal/vm/config.go
index f3c2341..565bacb 100644
--- a/internal/vm/config.go
+++ b/internal/vm/config.go
@@ -77,7 +77,7 @@ func (c *VMConfig) applyDefaults() {
 		c.SocketPath = fmt.Sprintf("/tmp/ch-%s.sock", c.SandboxID)
 	}
 	if c.SandboxDir == "" {
-		c.SandboxDir = "/tmp/ch-vm"
+		c.SandboxDir = fmt.Sprintf("/tmp/ch-vm-%s", c.SandboxID)
 	}
 	if c.TapDevice == "" {
 		c.TapDevice = "tap0"
diff --git a/internal/vm/manager.go b/internal/vm/manager.go
index 5f8539e..3233a49 100644
--- a/internal/vm/manager.go
+++ b/internal/vm/manager.go
@@ -143,7 +143,6 @@ func (m *Manager) Destroy(ctx context.Context, sandboxID string) error {
 		m.mu.Unlock()
 		return fmt.Errorf("VM not found: %s", sandboxID)
 	}
-	delete(m.vms, sandboxID)
 	m.mu.Unlock()
 
 	slog.Info("destroying VM", "sandbox", sandboxID)
@@ -161,6 +160,10 @@ func (m *Manager) Destroy(ctx context.Context, sandboxID string) error {
 
 	os.Remove(vm.Config.SocketPath)
 
+	m.mu.Lock()
+	delete(m.vms, sandboxID)
+	m.mu.Unlock()
+
 	slog.Info("VM destroyed", "sandbox", sandboxID)
 	return nil
 }
diff --git a/pkg/cpserver/run.go b/pkg/cpserver/run.go
index 1fbf6eb..2b7a7a1 100644
--- a/pkg/cpserver/run.go
+++ b/pkg/cpserver/run.go
@@ -255,7 +255,7 @@ func Run(opts ...Option) {
 	// Start extension background workers.
 	for _, ext := range o.extensions {
 		for _, worker := range ext.BackgroundWorkers(sctx) {
-			worker(ctx)
+			go worker(ctx)
 		}
 	}