Merge pull request #2130 from nybidari:iptables

PiperOrigin-RevId: 303208407

20 files changed, 278 insertions, 17 deletions

diff --git a/pkg/abi/linux/netfilter.go b/pkg/abi/linux/netfilter.go
index 80dc09aa9..a8d4f9d69 100644
--- a/pkg/abi/linux/netfilter.go
+++ b/pkg/abi/linux/netfilter.go
@@ -509,3 +509,44 @@ const (
 	// Enable all flags.
 	XT_UDP_INV_MASK = 0x03
 )
+
+// IPTOwnerInfo holds data for matching packets with owner. It corresponds
+// to struct ipt_owner_info in libxt_owner.c of iptables binary.
+type IPTOwnerInfo struct {
+	// UID is user id which created the packet.
+	UID uint32
+
+	// GID is group id which created the packet.
+	GID uint32
+
+	// PID is process id of the process which created the packet.
+	PID uint32
+
+	// SID is session id which created the packet.
+	SID uint32
+
+	// Comm is the command name which created the packet.
+	Comm [16]byte
+
+	// Match is used to match UID/GID of the socket. See the
+	// XT_OWNER_* flags below.
+	Match uint8
+
+	// Invert flips the meaning of Match field.
+	Invert uint8
+}
+
+// SizeOfIPTOwnerInfo is the size of an XTOwnerMatchInfo.
+const SizeOfIPTOwnerInfo = 34
+
+// Flags in IPTOwnerInfo.Match. Corresponding constants are in
+// include/uapi/linux/netfilter/xt_owner.h.
+const (
+	// Match the UID of the packet.
+	XT_OWNER_UID = 1 << 0
+	// Match the GID of the packet.
+	XT_OWNER_GID = 1 << 1
+	// Match if the socket exists for the packet. Forwarded
+	// packets do not have an associated socket.
+	XT_OWNER_SOCKET = 1 << 2
+)
diff --git a/pkg/abi/linux/netfilter_test.go b/pkg/abi/linux/netfilter_test.go
index 21e237f92..565dd550e 100644
--- a/pkg/abi/linux/netfilter_test.go
+++ b/pkg/abi/linux/netfilter_test.go
@@ -29,6 +29,7 @@ func TestSizes(t *testing.T) {
 		{IPTGetEntries{}, SizeOfIPTGetEntries},
 		{IPTGetinfo{}, SizeOfIPTGetinfo},
 		{IPTIP{}, SizeOfIPTIP},
+		{IPTOwnerInfo{}, SizeOfIPTOwnerInfo},
 		{IPTReplace{}, SizeOfIPTReplace},
 		{XTCounters{}, SizeOfXTCounters},
 		{XTEntryMatch{}, SizeOfXTEntryMatch},
diff --git a/pkg/sentry/kernel/task.go b/pkg/sentry/kernel/task.go
index 8452ddf5b..d6546735e 100644
--- a/pkg/sentry/kernel/task.go
+++ b/pkg/sentry/kernel/task.go
@@ -863,3 +863,15 @@ func (t *Task) SetOOMScoreAdj(adj int32) error {
 	atomic.StoreInt32(&t.tg.oomScoreAdj, adj)
 	return nil
 }
+
+// UID returns t's uid.
+// TODO(gvisor.dev/issue/170): This method is not namespaced yet.
+func (t *Task) UID() uint32 {
+	return uint32(t.Credentials().EffectiveKUID)
+}
+
+// GID returns t's gid.
+// TODO(gvisor.dev/issue/170): This method is not namespaced yet.
+func (t *Task) GID() uint32 {
+	return uint32(t.Credentials().EffectiveKGID)
+}
diff --git a/pkg/sentry/socket/netfilter/BUILD b/pkg/sentry/socket/netfilter/BUILD
index e801abeb8..721094bbf 100644
--- a/pkg/sentry/socket/netfilter/BUILD
+++ b/pkg/sentry/socket/netfilter/BUILD
@@ -7,6 +7,7 @@ go_library(
     srcs = [
         "extensions.go",
         "netfilter.go",
+        "owner_matcher.go",
         "targets.go",
         "tcp_matcher.go",
         "udp_matcher.go",
diff --git a/pkg/sentry/socket/netfilter/netfilter.go b/pkg/sentry/socket/netfilter/netfilter.go
index 55bcc3ace..878f81fd5 100644
--- a/pkg/sentry/socket/netfilter/netfilter.go
+++ b/pkg/sentry/socket/netfilter/netfilter.go
@@ -517,11 +517,10 @@ func SetEntries(stk *stack.Stack, optVal []byte) *syserr.Error {
 	}
 
 	// TODO(gvisor.dev/issue/170): Support other chains.
-	// Since we only support modifying the INPUT chain and redirect for
-	// PREROUTING chain right now, make sure all other chains point to
-	// ACCEPT rules.
+	// Since we only support modifying the INPUT, PREROUTING and OUTPUT chain right now,
+	// make sure all other chains point to ACCEPT rules.
 	for hook, ruleIdx := range table.BuiltinChains {
-		if hook != stack.Input && hook != stack.Prerouting {
+		if hook == stack.Forward || hook == stack.Postrouting {
 			if _, ok := table.Rules[ruleIdx].Target.(stack.AcceptTarget); !ok {
 				nflog("hook %d is unsupported.", hook)
 				return syserr.ErrInvalidArgument
diff --git a/pkg/sentry/socket/netfilter/owner_matcher.go b/pkg/sentry/socket/netfilter/owner_matcher.go
new file mode 100644
index 000000000..5949a7c29
--- /dev/null
+++ b/pkg/sentry/socket/netfilter/owner_matcher.go
@@ -0,0 +1,128 @@
+// Copyright 2020 The gVisor Authors.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package netfilter
+
+import (
+	"fmt"
+
+	"gvisor.dev/gvisor/pkg/abi/linux"
+	"gvisor.dev/gvisor/pkg/binary"
+	"gvisor.dev/gvisor/pkg/tcpip/stack"
+	"gvisor.dev/gvisor/pkg/usermem"
+)
+
+const matcherNameOwner = "owner"
+
+func init() {
+	registerMatchMaker(ownerMarshaler{})
+}
+
+// ownerMarshaler implements matchMaker for owner matching.
+type ownerMarshaler struct{}
+
+// name implements matchMaker.name.
+func (ownerMarshaler) name() string {
+	return matcherNameOwner
+}
+
+// marshal implements matchMaker.marshal.
+func (ownerMarshaler) marshal(mr stack.Matcher) []byte {
+	matcher := mr.(*OwnerMatcher)
+	iptOwnerInfo := linux.IPTOwnerInfo{
+		UID: matcher.uid,
+		GID: matcher.gid,
+	}
+
+	// Support for UID match.
+	// TODO(gvisor.dev/issue/170): Need to support gid match.
+	if matcher.matchUID {
+		iptOwnerInfo.Match = linux.XT_OWNER_UID
+	} else if matcher.matchGID {
+		panic("GID match is not supported.")
+	} else {
+		panic("UID match is not set.")
+	}
+
+	buf := make([]byte, 0, linux.SizeOfIPTOwnerInfo)
+	return marshalEntryMatch(matcherNameOwner, binary.Marshal(buf, usermem.ByteOrder, iptOwnerInfo))
+}
+
+// unmarshal implements matchMaker.unmarshal.
+func (ownerMarshaler) unmarshal(buf []byte, filter stack.IPHeaderFilter) (stack.Matcher, error) {
+	if len(buf) < linux.SizeOfIPTOwnerInfo {
+		return nil, fmt.Errorf("buf has insufficient size for owner match: %d", len(buf))
+	}
+
+	// For alignment reasons, the match's total size may
+	// exceed what's strictly necessary to hold matchData.
+	var matchData linux.IPTOwnerInfo
+	binary.Unmarshal(buf[:linux.SizeOfIPTOwnerInfo], usermem.ByteOrder, &matchData)
+	nflog("parseMatchers: parsed IPTOwnerInfo: %+v", matchData)
+
+	if matchData.Invert != 0 {
+		return nil, fmt.Errorf("invert flag is not supported for owner match")
+	}
+
+	// Support for UID match.
+	// TODO(gvisor.dev/issue/170): Need to support gid match.
+	if matchData.Match&linux.XT_OWNER_UID != linux.XT_OWNER_UID {
+		return nil, fmt.Errorf("owner match is only supported for uid")
+	}
+
+	// Check Flags.
+	var owner OwnerMatcher
+	owner.uid = matchData.UID
+	owner.gid = matchData.GID
+	owner.matchUID = true
+
+	return &owner, nil
+}
+
+type OwnerMatcher struct {
+	uid      uint32
+	gid      uint32
+	matchUID bool
+	matchGID bool
+	invert   uint8
+}
+
+// Name implements Matcher.Name.
+func (*OwnerMatcher) Name() string {
+	return matcherNameOwner
+}
+
+// Match implements Matcher.Match.
+func (om *OwnerMatcher) Match(hook stack.Hook, pkt stack.PacketBuffer, interfaceName string) (bool, bool) {
+	// Support only for OUTPUT chain.
+	// TODO(gvisor.dev/issue/170): Need to support for POSTROUTING chain also.
+	if hook != stack.Output {
+		return false, true
+	}
+
+	// If the packet owner is not set, drop the packet.
+	// Support for uid match.
+	// TODO(gvisor.dev/issue/170): Need to support gid match.
+	if pkt.Owner == nil || !om.matchUID {
+		return false, true
+	}
+
+	// TODO(gvisor.dev/issue/170): Need to add tests to verify
+	// drop rule when packet UID does not match owner matcher UID.
+	if pkt.Owner.UID() != om.uid {
+		return false, false
+	}
+
+	return true, false
+}
diff --git a/pkg/sentry/socket/netstack/provider.go b/pkg/sentry/socket/netstack/provider.go
index 5f181f017..eb090e79b 100644
--- a/pkg/sentry/socket/netstack/provider.go
+++ b/pkg/sentry/socket/netstack/provider.go
@@ -126,6 +126,12 @@ func (p *provider) Socket(t *kernel.Task, stype linux.SockType, protocol int) (*
 		ep, e = eps.Stack.NewRawEndpoint(transProto, p.netProto, wq, associated)
 	} else {
 		ep, e = eps.Stack.NewEndpoint(transProto, p.netProto, wq)
+
+		// Assign task to PacketOwner interface to get the UID and GID for
+		// iptables owner matching.
+		if e == nil {
+			ep.SetOwner(t)
+		}
 	}
 	if e != nil {
 		return nil, syserr.TranslateNetstackError(e)
diff --git a/pkg/tcpip/network/ipv4/ipv4.go b/pkg/tcpip/network/ipv4/ipv4.go
index b3ee6000e..a7d9a8b25 100644
--- a/pkg/tcpip/network/ipv4/ipv4.go
+++ b/pkg/tcpip/network/ipv4/ipv4.go
@@ -244,6 +244,14 @@ func (e *endpoint) WritePacket(r *stack.Route, gso *stack.GSO, params stack.Netw
 	ip := e.addIPHeader(r, &pkt.Header, pkt.Data.Size(), params)
 	pkt.NetworkHeader = buffer.View(ip)
 
+	// iptables filtering. All packets that reach here are locally
+	// generated.
+	ipt := e.stack.IPTables()
+	if ok := ipt.Check(stack.Output, pkt); !ok {
+		// iptables is telling us to drop the packet.
+		return nil
+	}
+
 	if r.Loop&stack.PacketLoop != 0 {
 		// The inbound path expects the network header to still be in
 		// the PacketBuffer's Data field.
@@ -280,7 +288,14 @@ func (e *endpoint) WritePackets(r *stack.Route, gso *stack.GSO, pkts []stack.Pac
 		return len(pkts), nil
 	}
 
+	// iptables filtering. All packets that reach here are locally
+	// generated.
+	ipt := e.stack.IPTables()
 	for i := range pkts {
+		if ok := ipt.Check(stack.Output, pkts[i]); !ok {
+			// iptables is telling us to drop the packet.
+			continue
+		}
 		ip := e.addIPHeader(r, &pkts[i].Header, pkts[i].DataSize, params)
 		pkts[i].NetworkHeader = buffer.View(ip)
 	}
diff --git a/pkg/tcpip/stack/packet_buffer.go b/pkg/tcpip/stack/packet_buffer.go
index 9505a4e92..9367de180 100644
--- a/pkg/tcpip/stack/packet_buffer.go
+++ b/pkg/tcpip/stack/packet_buffer.go
@@ -13,7 +13,10 @@
 
 package stack
 
-import "gvisor.dev/gvisor/pkg/tcpip/buffer"
+import (
+	"gvisor.dev/gvisor/pkg/tcpip"
+	"gvisor.dev/gvisor/pkg/tcpip/buffer"
+)
 
 // A PacketBuffer contains all the data of a network packet.
 //
@@ -59,6 +62,10 @@ type PacketBuffer struct {
 	// Hash is the transport layer hash of this packet. A value of zero
 	// indicates no valid hash has been set.
 	Hash uint32
+
+	// Owner is implemented by task to get the uid and gid.
+	// Only set for locally generated packets.
+	Owner tcpip.PacketOwner
 }
 
 // Clone makes a copy of pk. It clones the Data field, which creates a new
diff --git a/pkg/tcpip/stack/transport_test.go b/pkg/tcpip/stack/transport_test.go
index 8ca9ac3cf..3084e6593 100644
--- a/pkg/tcpip/stack/transport_test.go
+++ b/pkg/tcpip/stack/transport_test.go
@@ -56,6 +56,8 @@ func (f *fakeTransportEndpoint) Stats() tcpip.EndpointStats {
 	return nil
 }
 
+func (f *fakeTransportEndpoint) SetOwner(owner tcpip.PacketOwner) {}
+
 func newFakeTransportEndpoint(s *stack.Stack, proto *fakeTransportProtocol, netProto tcpip.NetworkProtocolNumber, uniqueID uint64) tcpip.Endpoint {
 	return &fakeTransportEndpoint{stack: s, TransportEndpointInfo: stack.TransportEndpointInfo{NetProto: netProto}, proto: proto, uniqueID: uniqueID}
 }
diff --git a/pkg/tcpip/tcpip.go b/pkg/tcpip/tcpip.go
index 3dc5d87d6..2ef3271f1 100644
--- a/pkg/tcpip/tcpip.go
+++ b/pkg/tcpip/tcpip.go
@@ -336,6 +336,15 @@ type ControlMessages struct {
 	PacketInfo IPPacketInfo
 }
 
+// PacketOwner is used to get UID and GID of the packet.
+type PacketOwner interface {
+	// UID returns UID of the packet.
+	UID() uint32
+
+	// GID returns GID of the packet.
+	GID() uint32
+}
+
 // Endpoint is the interface implemented by transport protocols (e.g., tcp, udp)
 // that exposes functionality like read, write, connect, etc. to users of the
 // networking stack.
@@ -470,6 +479,9 @@ type Endpoint interface {
 
 	// Stats returns a reference to the endpoint stats.
 	Stats() EndpointStats
+
+	// SetOwner sets the task owner to the endpoint owner.
+	SetOwner(owner PacketOwner)
 }
 
 // EndpointInfo is the interface implemented by each endpoint info struct.
diff --git a/pkg/tcpip/transport/icmp/endpoint.go b/pkg/tcpip/transport/icmp/endpoint.go
index 613b12ead..b007302fb 100644
--- a/pkg/tcpip/transport/icmp/endpoint.go
+++ b/pkg/tcpip/transport/icmp/endpoint.go
@@ -73,6 +73,9 @@ type endpoint struct {
 	route         stack.Route `state:"manual"`
 	ttl           uint8
 	stats         tcpip.TransportEndpointStats `state:"nosave"`
+
+	// owner is used to get uid and gid of the packet.
+	owner tcpip.PacketOwner
 }
 
 func newEndpoint(s *stack.Stack, netProto tcpip.NetworkProtocolNumber, transProto tcpip.TransportProtocolNumber, waiterQueue *waiter.Queue) (tcpip.Endpoint, *tcpip.Error) {
@@ -133,6 +136,10 @@ func (e *endpoint) Close() {
 // ModerateRecvBuf implements tcpip.Endpoint.ModerateRecvBuf.
 func (e *endpoint) ModerateRecvBuf(copied int) {}
 
+func (e *endpoint) SetOwner(owner tcpip.PacketOwner) {
+	e.owner = owner
+}
+
 // IPTables implements tcpip.Endpoint.IPTables.
 func (e *endpoint) IPTables() (stack.IPTables, error) {
 	return e.stack.IPTables(), nil
@@ -321,7 +328,7 @@ func (e *endpoint) write(p tcpip.Payloader, opts tcpip.WriteOptions) (int64, <-c
 
 	switch e.NetProto {
 	case header.IPv4ProtocolNumber:
-		err = send4(route, e.ID.LocalPort, v, e.ttl)
+		err = send4(route, e.ID.LocalPort, v, e.ttl, e.owner)
 
 	case header.IPv6ProtocolNumber:
 		err = send6(route, e.ID.LocalPort, v, e.ttl)
@@ -415,7 +422,7 @@ func (e *endpoint) GetSockOpt(opt interface{}) *tcpip.Error {
 	}
 }
 
-func send4(r *stack.Route, ident uint16, data buffer.View, ttl uint8) *tcpip.Error {
+func send4(r *stack.Route, ident uint16, data buffer.View, ttl uint8, owner tcpip.PacketOwner) *tcpip.Error {
 	if len(data) < header.ICMPv4MinimumSize {
 		return tcpip.ErrInvalidEndpointState
 	}
@@ -444,6 +451,7 @@ func send4(r *stack.Route, ident uint16, data buffer.View, ttl uint8) *tcpip.Err
 		Header:          hdr,
 		Data:            data.ToVectorisedView(),
 		TransportHeader: buffer.View(icmpv4),
+		Owner:           owner,
 	})
 }
 
diff --git a/pkg/tcpip/transport/packet/endpoint.go b/pkg/tcpip/transport/packet/endpoint.go
index df49d0995..23158173d 100644
--- a/pkg/tcpip/transport/packet/endpoint.go
+++ b/pkg/tcpip/transport/packet/endpoint.go
@@ -392,3 +392,5 @@ func (ep *endpoint) Info() tcpip.EndpointInfo {
 func (ep *endpoint) Stats() tcpip.EndpointStats {
 	return &ep.stats
 }
+
+func (ep *endpoint) SetOwner(owner tcpip.PacketOwner) {}
diff --git a/pkg/tcpip/transport/raw/endpoint.go b/pkg/tcpip/transport/raw/endpoint.go
index 536dafd1e..337bc1c71 100644
--- a/pkg/tcpip/transport/raw/endpoint.go
+++ b/pkg/tcpip/transport/raw/endpoint.go
@@ -80,6 +80,9 @@ type endpoint struct {
 	// Connect(), and is valid only when conneted is true.
 	route stack.Route                  `state:"manual"`
 	stats tcpip.TransportEndpointStats `state:"nosave"`
+
+	// owner is used to get uid and gid of the packet.
+	owner tcpip.PacketOwner
 }
 
 // NewEndpoint returns a raw  endpoint for the given protocols.
@@ -159,6 +162,10 @@ func (e *endpoint) Close() {
 // ModerateRecvBuf implements tcpip.Endpoint.ModerateRecvBuf.
 func (e *endpoint) ModerateRecvBuf(copied int) {}
 
+func (e *endpoint) SetOwner(owner tcpip.PacketOwner) {
+	e.owner = owner
+}
+
 // IPTables implements tcpip.Endpoint.IPTables.
 func (e *endpoint) IPTables() (stack.IPTables, error) {
 	return e.stack.IPTables(), nil
@@ -348,10 +355,12 @@ func (e *endpoint) finishWrite(payloadBytes []byte, route *stack.Route) (int64,
 			}
 			break
 		}
+
 		hdr := buffer.NewPrependable(len(payloadBytes) + int(route.MaxHeaderLength()))
 		if err := route.WritePacket(nil /* gso */, stack.NetworkHeaderParams{Protocol: e.TransProto, TTL: route.DefaultTTL(), TOS: stack.DefaultTOS}, stack.PacketBuffer{
 			Header: hdr,
 			Data:   buffer.View(payloadBytes).ToVectorisedView(),
+			Owner:  e.owner,
 		}); err != nil {
 			return 0, nil, err
 		}
diff --git a/pkg/tcpip/transport/tcp/accept.go b/pkg/tcpip/transport/tcp/accept.go
index 375ca21f6..7a9dea4ac 100644
--- a/pkg/tcpip/transport/tcp/accept.go
+++ b/pkg/tcpip/transport/tcp/accept.go
@@ -276,7 +276,7 @@ func (l *listenContext) createConnectingEndpoint(s *segment, iss seqnum.Value, i
 // and then performs the TCP 3-way handshake.
 //
 // The new endpoint is returned with e.mu held.
-func (l *listenContext) createEndpointAndPerformHandshake(s *segment, opts *header.TCPSynOptions, queue *waiter.Queue) (*endpoint, *tcpip.Error) {
+func (l *listenContext) createEndpointAndPerformHandshake(s *segment, opts *header.TCPSynOptions, queue *waiter.Queue, owner tcpip.PacketOwner) (*endpoint, *tcpip.Error) {
 	// Create new endpoint.
 	irs := s.sequenceNumber
 	isn := generateSecureISN(s.id, l.stack.Seed())
@@ -284,6 +284,7 @@ func (l *listenContext) createEndpointAndPerformHandshake(s *segment, opts *head
 	if err != nil {
 		return nil, err
 	}
+	ep.owner = owner
 
 	// listenEP is nil when listenContext is used by tcp.Forwarder.
 	deferAccept := time.Duration(0)
@@ -414,7 +415,7 @@ func (e *endpoint) handleSynSegment(ctx *listenContext, s *segment, opts *header
 	}()
 	defer s.decRef()
 
-	n, err := ctx.createEndpointAndPerformHandshake(s, opts, &waiter.Queue{})
+	n, err := ctx.createEndpointAndPerformHandshake(s, opts, &waiter.Queue{}, e.owner)
 	if err != nil {
 		e.stack.Stats().TCP.FailedConnectionAttempts.Increment()
 		e.stats.FailedConnectionAttempts.Increment()
diff --git a/pkg/tcpip/transport/tcp/connect.go b/pkg/tcpip/transport/tcp/connect.go
index 1d245c2c6..3239a5911 100644
--- a/pkg/tcpip/transport/tcp/connect.go
+++ b/pkg/tcpip/transport/tcp/connect.go
@@ -745,7 +745,7 @@ func (e *endpoint) sendSynTCP(r *stack.Route, tf tcpFields, opts header.TCPSynOp
 
 func (e *endpoint) sendTCP(r *stack.Route, tf tcpFields, data buffer.VectorisedView, gso *stack.GSO) *tcpip.Error {
 	tf.txHash = e.txHash
-	if err := sendTCP(r, tf, data, gso); err != nil {
+	if err := sendTCP(r, tf, data, gso, e.owner); err != nil {
 		e.stats.SendErrors.SegmentSendToNetworkFailed.Increment()
 		return err
 	}
@@ -787,7 +787,7 @@ func buildTCPHdr(r *stack.Route, tf tcpFields, pkt *stack.PacketBuffer, gso *sta
 	}
 }
 
-func sendTCPBatch(r *stack.Route, tf tcpFields, data buffer.VectorisedView, gso *stack.GSO) *tcpip.Error {
+func sendTCPBatch(r *stack.Route, tf tcpFields, data buffer.VectorisedView, gso *stack.GSO, owner tcpip.PacketOwner) *tcpip.Error {
 	optLen := len(tf.opts)
 	if tf.rcvWnd > 0xffff {
 		tf.rcvWnd = 0xffff
@@ -816,6 +816,7 @@ func sendTCPBatch(r *stack.Route, tf tcpFields, data buffer.VectorisedView, gso
 		pkts[i].DataSize = packetSize
 		pkts[i].Data = data
 		pkts[i].Hash = tf.txHash
+		pkts[i].Owner = owner
 		buildTCPHdr(r, tf, &pkts[i], gso)
 		off += packetSize
 		tf.seq = tf.seq.Add(seqnum.Size(packetSize))
@@ -833,14 +834,14 @@ func sendTCPBatch(r *stack.Route, tf tcpFields, data buffer.VectorisedView, gso
 
 // sendTCP sends a TCP segment with the provided options via the provided
 // network endpoint and under the provided identity.
-func sendTCP(r *stack.Route, tf tcpFields, data buffer.VectorisedView, gso *stack.GSO) *tcpip.Error {
+func sendTCP(r *stack.Route, tf tcpFields, data buffer.VectorisedView, gso *stack.GSO, owner tcpip.PacketOwner) *tcpip.Error {
 	optLen := len(tf.opts)
 	if tf.rcvWnd > 0xffff {
 		tf.rcvWnd = 0xffff
 	}
 
 	if r.Loop&stack.PacketLoop == 0 && gso != nil && gso.Type == stack.GSOSW && int(gso.MSS) < data.Size() {
-		return sendTCPBatch(r, tf, data, gso)
+		return sendTCPBatch(r, tf, data, gso, owner)
 	}
 
 	pkt := stack.PacketBuffer{
@@ -849,6 +850,7 @@ func sendTCP(r *stack.Route, tf tcpFields, data buffer.VectorisedView, gso *stac
 		DataSize:   data.Size(),
 		Data:       data,
 		Hash:       tf.txHash,
+		Owner:      owner,
 	}
 	buildTCPHdr(r, tf, &pkt, gso)
 
diff --git a/pkg/tcpip/transport/tcp/endpoint.go b/pkg/tcpip/transport/tcp/endpoint.go
index 1ebee0cfe..9b123e968 100644
--- a/pkg/tcpip/transport/tcp/endpoint.go
+++ b/pkg/tcpip/transport/tcp/endpoint.go
@@ -603,6 +603,9 @@ type endpoint struct {
 	// txHash is the transport layer hash to be set on outbound packets
 	// emitted by this endpoint.
 	txHash uint32
+
+	// owner is used to get uid and gid of the packet.
+	owner tcpip.PacketOwner
 }
 
 // UniqueID implements stack.TransportEndpoint.UniqueID.
@@ -1132,6 +1135,10 @@ func (e *endpoint) ModerateRecvBuf(copied int) {
 	e.rcvListMu.Unlock()
 }
 
+func (e *endpoint) SetOwner(owner tcpip.PacketOwner) {
+	e.owner = owner
+}
+
 // IPTables implements tcpip.Endpoint.IPTables.
 func (e *endpoint) IPTables() (stack.IPTables, error) {
 	return e.stack.IPTables(), nil
diff --git a/pkg/tcpip/transport/tcp/forwarder.go b/pkg/tcpip/transport/tcp/forwarder.go
index a094471b8..808410c92 100644
--- a/pkg/tcpip/transport/tcp/forwarder.go
+++ b/pkg/tcpip/transport/tcp/forwarder.go
@@ -157,7 +157,7 @@ func (r *ForwarderRequest) CreateEndpoint(queue *waiter.Queue) (tcpip.Endpoint,
 		TSVal:         r.synOptions.TSVal,
 		TSEcr:         r.synOptions.TSEcr,
 		SACKPermitted: r.synOptions.SACKPermitted,
-	}, queue)
+	}, queue, nil)
 	if err != nil {
 		return nil, err
 	}
diff --git a/pkg/tcpip/transport/tcp/protocol.go b/pkg/tcpip/transport/tcp/protocol.go
index 1377107ca..dce9a1652 100644
--- a/pkg/tcpip/transport/tcp/protocol.go
+++ b/pkg/tcpip/transport/tcp/protocol.go
@@ -199,7 +199,7 @@ func replyWithReset(s *segment) {
 		seq:    seq,
 		ack:    ack,
 		rcvWnd: 0,
-	}, buffer.VectorisedView{}, nil /* gso */)
+	}, buffer.VectorisedView{}, nil /* gso */, nil /* PacketOwner */)
 }
 
 // SetOption implements stack.TransportProtocol.SetOption.
diff --git a/pkg/tcpip/transport/udp/endpoint.go b/pkg/tcpip/transport/udp/endpoint.go
index a3372ac58..120d3baa3 100644
--- a/pkg/tcpip/transport/udp/endpoint.go
+++ b/pkg/tcpip/transport/udp/endpoint.go
@@ -143,6 +143,9 @@ type endpoint struct {
 
 	// TODO(b/142022063): Add ability to save and restore per endpoint stats.
 	stats tcpip.TransportEndpointStats `state:"nosave"`
+
+	// owner is used to get uid and gid of the packet.
+	owner tcpip.PacketOwner
 }
 
 // +stateify savable
@@ -484,7 +487,7 @@ func (e *endpoint) write(p tcpip.Payloader, opts tcpip.WriteOptions) (int64, <-c
 		useDefaultTTL = false
 	}
 
-	if err := sendUDP(route, buffer.View(v).ToVectorisedView(), e.ID.LocalPort, dstPort, ttl, useDefaultTTL, e.sendTOS); err != nil {
+	if err := sendUDP(route, buffer.View(v).ToVectorisedView(), e.ID.LocalPort, dstPort, ttl, useDefaultTTL, e.sendTOS, e.owner); err != nil {
 		return 0, nil, err
 	}
 	return int64(len(v)), nil, nil
@@ -886,7 +889,7 @@ func (e *endpoint) GetSockOpt(opt interface{}) *tcpip.Error {
 
 // sendUDP sends a UDP segment via the provided network endpoint and under the
 // provided identity.
-func sendUDP(r *stack.Route, data buffer.VectorisedView, localPort, remotePort uint16, ttl uint8, useDefaultTTL bool, tos uint8) *tcpip.Error {
+func sendUDP(r *stack.Route, data buffer.VectorisedView, localPort, remotePort uint16, ttl uint8, useDefaultTTL bool, tos uint8, owner tcpip.PacketOwner) *tcpip.Error {
 	// Allocate a buffer for the UDP header.
 	hdr := buffer.NewPrependable(header.UDPMinimumSize + int(r.MaxHeaderLength()))
 
@@ -916,6 +919,7 @@ func sendUDP(r *stack.Route, data buffer.VectorisedView, localPort, remotePort u
 		Header:          hdr,
 		Data:            data,
 		TransportHeader: buffer.View(udp),
+		Owner:           owner,
 	}); err != nil {
 		r.Stats().UDP.PacketSendErrors.Increment()
 		return err
@@ -1356,3 +1360,7 @@ func (*endpoint) Wait() {}
 func isBroadcastOrMulticast(a tcpip.Address) bool {
 	return a == header.IPv4Broadcast || header.IsV4MulticastAddress(a) || header.IsV6MulticastAddress(a)
 }
+
+func (e *endpoint) SetOwner(owner tcpip.PacketOwner) {
+	e.owner = owner
+}