1 files changed, 398 insertions, 0 deletions
diff --git a/pkg/tcpip/network/ipv4/igmp.go b/pkg/tcpip/network/ipv4/igmp.go
new file mode 100644
index 000000000..e1de58f73
--- /dev/null
+++ b/pkg/tcpip/network/ipv4/igmp.go
@@ -0,0 +1,398 @@
+// 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 ipv4
+
+import (
+	"fmt"
+	"sync"
+	"time"
+
+	"gvisor.dev/gvisor/pkg/tcpip"
+	"gvisor.dev/gvisor/pkg/tcpip/buffer"
+	"gvisor.dev/gvisor/pkg/tcpip/header"
+	"gvisor.dev/gvisor/pkg/tcpip/stack"
+)
+
+const (
+	// igmpV1PresentDefault is the initial state for igmpV1Present in the
+	// igmpState. As per RFC 2236 Page 9 says "No IGMPv1 Router Present ... is
+	// the initial state."
+	igmpV1PresentDefault = false
+
+	// v1RouterPresentTimeout from RFC 2236 Section 8.11, Page 18
+	// See note on igmpState.igmpV1Present for more detail.
+	v1RouterPresentTimeout = 400 * time.Second
+
+	// v1MaxRespTimeTenthSec from RFC 2236 Section 4, Page 5. "The IGMPv1 router
+	// will send General Queries with the Max Response Time set to 0. This MUST
+	// be interpreted as a value of 100 (10 seconds)."
+	v1MaxRespTimeTenthSec = 100
+
+	// UnsolicitedReportIntervalMaxTenthSec from RFC 2236 Section 8.10, Page 19.
+	// As all IGMP delay timers are set to a random value between 0 and the
+	// interval, this is technically a maximum.
+	UnsolicitedReportIntervalMaxTenthSec = 100
+)
+
+// igmpState is the per-interface IGMP state.
+//
+// igmpState.init() MUST be called after creating an IGMP state.
+type igmpState struct {
+	// The IPv4 endpoint this igmpState is for.
+	ep *endpoint
+
+	mu struct {
+		sync.RWMutex
+
+		// memberships contains the map of host groups to their state, timer, and
+		// flag info.
+		memberships map[tcpip.Address]membershipInfo
+
+		// igmpV1Present is for maintaining compatibility with IGMPv1 Routers, from
+		// RFC 2236 Section 4 Page 6: "The IGMPv1 router expects Version 1
+		// Membership Reports in response to its Queries, and will not pay
+		// attention to Version 2 Membership Reports.  Therefore, a state variable
+		// MUST be kept for each interface, describing whether the multicast
+		// Querier on that interface is running IGMPv1 or IGMPv2.  This variable
+		// MUST be based upon whether or not an IGMPv1 query was heard in the last
+		// [Version 1 Router Present Timeout] seconds"
+		igmpV1Present bool
+
+		// igmpV1Job is scheduled when this interface receives an IGMPv1 style
+		// message, upon expiration the igmpV1Present flag is cleared.
+		// igmpV1Job may not be nil once igmpState is initialized.
+		igmpV1Job *tcpip.Job
+	}
+}
+
+// membershipInfo holds the IGMPv2 state for a particular multicast address.
+type membershipInfo struct {
+	// state contains the current IGMP state for this member.
+	state hostState
+
+	// lastToSendReport is true if this was "the last host to send a report from
+	// this group."
+	// RFC 2236, Section 6, Page 9. This is used to track whether or not there
+	// are other hosts on this subnet that belong to this group - RFC 2236
+	// Section 3, Page 5.
+	lastToSendReport bool
+
+	// delayedReportJob is used to delay sending responses to IGMP messages in
+	// order to reduce duplicate reports from multiple hosts on the interface.
+	// Must not be nil.
+	delayedReportJob *tcpip.Job
+}
+
+type hostState int
+
+// From RFC 2236, Section 6, Page 7.
+const (
+	// "'Non-Member' state, when the host does not belong to the group on
+	// the interface.  This is the initial state for all memberships on
+	// all network interfaces; it requires no storage in the host."
+	_ hostState = iota
+
+	// delayingMember is the "'Delaying Member' state, when the host belongs to
+	// the group on the interface and has a report delay timer running for that
+	// membership."
+	delayingMember
+
+	// idleMember is the "Idle Member" state, when the host belongs to the group
+	// on the interface and does not have a report delay timer running for that
+	// membership.
+	idleMember
+)
+
+// init sets up an igmpState struct, and is required to be called before using
+// a new igmpState.
+func (igmp *igmpState) init(ep *endpoint) {
+	igmp.mu.Lock()
+	defer igmp.mu.Unlock()
+	igmp.ep = ep
+	igmp.mu.memberships = make(map[tcpip.Address]membershipInfo)
+	igmp.mu.igmpV1Present = igmpV1PresentDefault
+	igmp.mu.igmpV1Job = igmp.ep.protocol.stack.NewJob(&igmp.mu, func() {
+		igmp.mu.igmpV1Present = false
+	})
+}
+
+func (igmp *igmpState) handleIGMP(pkt *stack.PacketBuffer) {
+	stats := igmp.ep.protocol.stack.Stats()
+	received := stats.IGMP.PacketsReceived
+	headerView, ok := pkt.Data.PullUp(header.IGMPMinimumSize)
+	if !ok {
+		received.Invalid.Increment()
+		return
+	}
+	h := header.IGMP(headerView)
+
+	// Temporarily reset the checksum field to 0 in order to calculate the proper
+	// checksum.
+	wantChecksum := h.Checksum()
+	h.SetChecksum(0)
+	gotChecksum := ^header.ChecksumVV(pkt.Data, 0 /* initial */)
+	h.SetChecksum(wantChecksum)
+
+	if gotChecksum != wantChecksum {
+		received.ChecksumErrors.Increment()
+		return
+	}
+
+	switch h.Type() {
+	case header.IGMPMembershipQuery:
+		received.MembershipQuery.Increment()
+		if len(headerView) < header.IGMPQueryMinimumSize {
+			received.Invalid.Increment()
+			return
+		}
+		igmp.handleMembershipQuery(h.GroupAddress(), h.MaxRespTime())
+	case header.IGMPv1MembershipReport:
+		received.V1MembershipReport.Increment()
+		if len(headerView) < header.IGMPReportMinimumSize {
+			received.Invalid.Increment()
+			return
+		}
+		igmp.handleMembershipReport(h.GroupAddress())
+	case header.IGMPv2MembershipReport:
+		received.V2MembershipReport.Increment()
+		if len(headerView) < header.IGMPReportMinimumSize {
+			received.Invalid.Increment()
+			return
+		}
+		igmp.handleMembershipReport(h.GroupAddress())
+	case header.IGMPLeaveGroup:
+		received.LeaveGroup.Increment()
+		// As per RFC 2236 Section 6, Page 7: "IGMP messages other than Query or
+		// Report, are ignored in all states"
+
+	default:
+		// As per RFC 2236 Section 2.1 Page 3: "Unrecognized message types should
+		// be silently ignored. New message types may be used by newer versions of
+		// IGMP, by multicast routing protocols, or other uses."
+		received.Unrecognized.Increment()
+	}
+}
+
+func (igmp *igmpState) handleMembershipQuery(groupAddress tcpip.Address, maxRespTime byte) {
+	igmp.mu.Lock()
+	defer igmp.mu.Unlock()
+
+	// As per RFC 2236 Section 6, Page 10: If the maximum response time is zero
+	// then change the state to note that an IGMPv1 router is present and
+	// schedule the query received Job.
+	if maxRespTime == 0 {
+		igmp.mu.igmpV1Job.Cancel()
+		igmp.mu.igmpV1Job.Schedule(v1RouterPresentTimeout)
+		igmp.mu.igmpV1Present = true
+		maxRespTime = v1MaxRespTimeTenthSec
+	}
+
+	// IPv4Any is the General Query Address.
+	if groupAddress == header.IPv4Any {
+		for membershipAddress, info := range igmp.mu.memberships {
+			igmp.setDelayTimerForAddressRLocked(membershipAddress, &info, maxRespTime)
+			igmp.mu.memberships[membershipAddress] = info
+		}
+	} else if info, ok := igmp.mu.memberships[groupAddress]; ok {
+		igmp.setDelayTimerForAddressRLocked(groupAddress, &info, maxRespTime)
+		igmp.mu.memberships[groupAddress] = info
+	}
+}
+
+// setDelayTimerForAddressRLocked modifies the passed info only and does not
+// modify IGMP state directly.
+//
+// Precondition: igmp.mu MUST be read locked.
+func (igmp *igmpState) setDelayTimerForAddressRLocked(groupAddress tcpip.Address, info *membershipInfo, maxRespTime byte) {
+	if info.state == delayingMember {
+		// As per RFC 2236 Section 3, page 3: "If a timer for the group is already
+		// running, it is reset to the random value only if the requested Max
+		// Response Time is less than the remaining value of the running timer.
+		// TODO: Reset the timer if time remaining is greater than maxRespTime.
+		return
+	}
+	info.state = delayingMember
+	info.delayedReportJob.Cancel()
+	info.delayedReportJob.Schedule(igmp.calculateDelayTimerDuration(maxRespTime))
+}
+
+func (igmp *igmpState) handleMembershipReport(groupAddress tcpip.Address) {
+	igmp.mu.Lock()
+	defer igmp.mu.Unlock()
+
+	// As per RFC 2236 Section 3, pages 3-4: "If the host receives another host's
+	// Report (version 1 or 2) while it has a timer running, it stops its timer
+	// for the specified group and does not send a Report"
+	if info, ok := igmp.mu.memberships[groupAddress]; ok {
+		info.delayedReportJob.Cancel()
+		info.lastToSendReport = false
+		igmp.mu.memberships[groupAddress] = info
+	}
+}
+
+// writePacket assembles and sends an IGMP packet with the provided fields,
+// incrementing the provided stat counter on success.
+func (igmp *igmpState) writePacket(destAddress tcpip.Address, groupAddress tcpip.Address, igmpType header.IGMPType) {
+	igmpData := header.IGMP(buffer.NewView(header.IGMPReportMinimumSize))
+	igmpData.SetType(igmpType)
+	igmpData.SetGroupAddress(groupAddress)
+	igmpData.SetChecksum(header.IGMPCalculateChecksum(igmpData))
+
+	pkt := stack.NewPacketBuffer(stack.PacketBufferOptions{
+		ReserveHeaderBytes: int(igmp.ep.MaxHeaderLength()),
+		Data:               buffer.View(igmpData).ToVectorisedView(),
+	})
+
+	// TODO(gvisor.dev/issue/4888): We should not use the unspecified address,
+	// rather we should select an appropriate local address.
+	r := stack.Route{
+		LocalAddress:  header.IPv4Any,
+		RemoteAddress: destAddress,
+	}
+	igmp.ep.addIPHeader(&r, pkt, stack.NetworkHeaderParams{
+		Protocol: header.IGMPProtocolNumber,
+		TTL:      header.IGMPTTL,
+		TOS:      stack.DefaultTOS,
+	})
+
+	// TODO(b/162198658): set the ROUTER_ALERT option when sending Host
+	// Membership Reports.
+	sent := igmp.ep.protocol.stack.Stats().IGMP.PacketsSent
+	if err := igmp.ep.nic.WritePacketToRemote(header.EthernetAddressFromMulticastIPv4Address(destAddress), nil /* gso */, header.IPv4ProtocolNumber, pkt); err != nil {
+		sent.Dropped.Increment()
+	} else {
+		switch igmpType {
+		case header.IGMPv1MembershipReport:
+			sent.V1MembershipReport.Increment()
+		case header.IGMPv2MembershipReport:
+			sent.V2MembershipReport.Increment()
+		case header.IGMPLeaveGroup:
+			sent.LeaveGroup.Increment()
+		default:
+			panic(fmt.Sprintf("unrecognized igmp type = %d", igmpType))
+		}
+	}
+}
+
+// sendReport sends a Host Membership Report in response to a query or after
+// this host joins a new group on this interface.
+//
+// Precondition: igmp.mu MUST be locked.
+func (igmp *igmpState) sendReportLocked(groupAddress tcpip.Address) {
+	igmpType := header.IGMPv2MembershipReport
+	if igmp.mu.igmpV1Present {
+		igmpType = header.IGMPv1MembershipReport
+	}
+	igmp.writePacket(groupAddress, groupAddress, igmpType)
+
+	// Update the state of the membership for this group. If the group no longer
+	// exists, do nothing since this report must have been a race with a remove
+	// or is in the process of being added.
+	info, ok := igmp.mu.memberships[groupAddress]
+	if !ok {
+		return
+	}
+	info.state = idleMember
+	info.lastToSendReport = true
+	igmp.mu.memberships[groupAddress] = info
+}
+
+// sendLeave sends a Leave Group report to the IPv4 All Routers Group.
+//
+// Precondition: igmp.mu MUST be read locked.
+func (igmp *igmpState) sendLeaveRLocked(groupAddress tcpip.Address) {
+	// As per RFC 2236 Section 6, Page 8: "If the interface state says the
+	// Querier is running IGMPv1, this action SHOULD be skipped. If the flag
+	// saying we were the last host to report is cleared, this action MAY be
+	// skipped."
+	if igmp.mu.igmpV1Present || !igmp.mu.memberships[groupAddress].lastToSendReport {
+		return
+	}
+
+	igmp.writePacket(header.IPv4AllRoutersGroup, groupAddress, header.IGMPLeaveGroup)
+}
+
+// joinGroup handles adding a new group to the membership map, setting up the
+// IGMP state for the group, and sending and scheduling the required
+// messages.
+//
+// If the group already exists in the membership map, returns
+// tcpip.ErrDuplicateAddress.
+func (igmp *igmpState) joinGroup(groupAddress tcpip.Address) *tcpip.Error {
+	igmp.mu.Lock()
+	defer igmp.mu.Unlock()
+	if _, ok := igmp.mu.memberships[groupAddress]; ok {
+		// The group already exists in the membership map.
+		return tcpip.ErrDuplicateAddress
+	}
+
+	info := membershipInfo{
+		// There isn't a Job scheduled currently, so it's just idle.
+		state: idleMember,
+		// Joining a group immediately sends a report.
+		lastToSendReport: true,
+		delayedReportJob: igmp.ep.protocol.stack.NewJob(&igmp.mu, func() {
+			igmp.sendReportLocked(groupAddress)
+		}),
+	}
+
+	// As per RFC 2236 Section 3, Page 5: "When a host joins a multicast group,
+	// it should immediately transmit an unsolicited Version 2 Membership Report
+	// for that group" ... "it is recommended that it be repeated"
+	igmp.sendReportLocked(groupAddress)
+	igmp.setDelayTimerForAddressRLocked(groupAddress, &info, UnsolicitedReportIntervalMaxTenthSec)
+	igmp.mu.memberships[groupAddress] = info
+
+	return nil
+}
+
+// leaveGroup handles removing the group from the membership map, cancels any
+// delay timers associated with that group, and sends the Leave Group message
+// if required.
+//
+// If the group does not exist in the membership map, this function will
+// silently return.
+func (igmp *igmpState) leaveGroup(groupAddress tcpip.Address) {
+	igmp.mu.Lock()
+	defer igmp.mu.Unlock()
+	info, ok := igmp.mu.memberships[groupAddress]
+	if !ok {
+		return
+	}
+
+	// Clean up the state of the group before sending the leave message and
+	// removing it from the map.
+	info.delayedReportJob.Cancel()
+	info.state = idleMember
+	igmp.mu.memberships[groupAddress] = info
+
+	igmp.sendLeaveRLocked(groupAddress)
+	delete(igmp.mu.memberships, groupAddress)
+}
+
+// RFC 2236 Section 3, Page 3: The response time is set to a "random value...
+// selected from the range (0, Max Response Time]" where Max Resp Time is given
+// in units of 1/10 of a second.
+func (igmp *igmpState) calculateDelayTimerDuration(maxRespTime byte) time.Duration {
+	maxRespTimeDuration := DecisecondToSecond(maxRespTime)
+	return time.Duration(igmp.ep.protocol.stack.Rand().Int63n(int64(maxRespTimeDuration)))
+}
+
+// DecisecondToSecond converts a byte representing deci-seconds to a Duration
+// type. This helper function exists because the IGMP stack sends and receives
+// Max Response Times in deci-seconds.
+func DecisecondToSecond(ds byte) time.Duration {
+	return time.Duration(ds) * time.Second / 10
+}