diff options
Diffstat (limited to 'vendor/github.com/vishvananda')
40 files changed, 2182 insertions, 235 deletions
diff --git a/vendor/github.com/vishvananda/netlink/.travis.yml b/vendor/github.com/vishvananda/netlink/.travis.yml index 7d14af4d6..80219c69d 100644 --- a/vendor/github.com/vishvananda/netlink/.travis.yml +++ b/vendor/github.com/vishvananda/netlink/.travis.yml @@ -1,8 +1,8 @@ language: go go: - - "1.10.x" - - "1.11.x" - "1.12.x" + - "1.13.x" + - "1.14.x" before_script: # make sure we keep path in tact when we sudo - sudo sed -i -e 's/^Defaults\tsecure_path.*$//' /etc/sudoers @@ -14,6 +14,7 @@ before_script: - sudo modprobe nf_conntrack_ipv4 - sudo modprobe nf_conntrack_ipv6 - sudo modprobe sch_hfsc + - sudo modprobe sch_sfq install: - - go get github.com/vishvananda/netns + - go get -v -t ./... go_import_path: github.com/vishvananda/netlink diff --git a/vendor/github.com/vishvananda/netlink/addr.go b/vendor/github.com/vishvananda/netlink/addr.go index f08c95696..653f540db 100644 --- a/vendor/github.com/vishvananda/netlink/addr.go +++ b/vendor/github.com/vishvananda/netlink/addr.go @@ -17,6 +17,7 @@ type Addr struct { Broadcast net.IP PreferedLft int ValidLft int + LinkIndex int } // String returns $ip/$netmask $label diff --git a/vendor/github.com/vishvananda/netlink/addr_linux.go b/vendor/github.com/vishvananda/netlink/addr_linux.go index 28746d5af..71da251ca 100644 --- a/vendor/github.com/vishvananda/netlink/addr_linux.go +++ b/vendor/github.com/vishvananda/netlink/addr_linux.go @@ -11,9 +11,6 @@ import ( "golang.org/x/sys/unix" ) -// IFA_FLAGS is a u32 attribute. -const IFA_FLAGS = 0x8 - // AddrAdd will add an IP address to a link device. // // Equivalent to: `ip addr add $addr dev $link` @@ -125,7 +122,7 @@ func (h *Handle) addrHandle(link Link, addr *Addr, req *nl.NetlinkRequest) error } else { b := make([]byte, 4) native.PutUint32(b, uint32(addr.Flags)) - flagsData := nl.NewRtAttr(IFA_FLAGS, b) + flagsData := nl.NewRtAttr(unix.IFA_FLAGS, b) req.AddData(flagsData) } } @@ -156,10 +153,10 @@ func (h *Handle) addrHandle(link Link, addr *Addr, req *nl.NetlinkRequest) error // value should be "forever". To compensate for that, only add the attributes if at least one of the values is // non-zero, which means the caller has explicitly set them if addr.ValidLft > 0 || addr.PreferedLft > 0 { - cachedata := nl.IfaCacheInfo{ - IfaValid: uint32(addr.ValidLft), - IfaPrefered: uint32(addr.PreferedLft), - } + cachedata := nl.IfaCacheInfo{unix.IfaCacheinfo{ + Valid: uint32(addr.ValidLft), + Prefered: uint32(addr.PreferedLft), + }} req.AddData(nl.NewRtAttr(unix.IFA_CACHEINFO, cachedata.Serialize())) } @@ -196,12 +193,12 @@ func (h *Handle) AddrList(link Link, family int) ([]Addr, error) { var res []Addr for _, m := range msgs { - addr, msgFamily, ifindex, err := parseAddr(m) + addr, msgFamily, err := parseAddr(m) if err != nil { return res, err } - if link != nil && ifindex != indexFilter { + if link != nil && addr.LinkIndex != indexFilter { // Ignore messages from other interfaces continue } @@ -216,11 +213,11 @@ func (h *Handle) AddrList(link Link, family int) ([]Addr, error) { return res, nil } -func parseAddr(m []byte) (addr Addr, family, index int, err error) { +func parseAddr(m []byte) (addr Addr, family int, err error) { msg := nl.DeserializeIfAddrmsg(m) family = -1 - index = -1 + addr.LinkIndex = -1 attrs, err1 := nl.ParseRouteAttr(m[msg.Len():]) if err1 != nil { @@ -229,7 +226,7 @@ func parseAddr(m []byte) (addr Addr, family, index int, err error) { } family = int(msg.Family) - index = int(msg.Index) + addr.LinkIndex = int(msg.Index) var local, dst *net.IPNet for _, attr := range attrs { @@ -254,12 +251,12 @@ func parseAddr(m []byte) (addr Addr, family, index int, err error) { addr.Broadcast = attr.Value case unix.IFA_LABEL: addr.Label = string(attr.Value[:len(attr.Value)-1]) - case IFA_FLAGS: + case unix.IFA_FLAGS: addr.Flags = int(native.Uint32(attr.Value[0:4])) - case nl.IFA_CACHEINFO: + case unix.IFA_CACHEINFO: ci := nl.DeserializeIfaCacheInfo(attr.Value) - addr.PreferedLft = int(ci.IfaPrefered) - addr.ValidLft = int(ci.IfaValid) + addr.PreferedLft = int(ci.Prefered) + addr.ValidLft = int(ci.Valid) } } @@ -394,7 +391,7 @@ func addrSubscribeAt(newNs, curNs netns.NsHandle, ch chan<- AddrUpdate, done <-c continue } - addr, _, ifindex, err := parseAddr(m.Data) + addr, _, err := parseAddr(m.Data) if err != nil { if cberr != nil { cberr(fmt.Errorf("could not parse address: %v", err)) @@ -403,7 +400,7 @@ func addrSubscribeAt(newNs, curNs netns.NsHandle, ch chan<- AddrUpdate, done <-c } ch <- AddrUpdate{LinkAddress: *addr.IPNet, - LinkIndex: ifindex, + LinkIndex: addr.LinkIndex, NewAddr: msgType == unix.RTM_NEWADDR, Flags: addr.Flags, Scope: addr.Scope, diff --git a/vendor/github.com/vishvananda/netlink/class.go b/vendor/github.com/vishvananda/netlink/class.go index dcc22d9e9..10ceffed8 100644 --- a/vendor/github.com/vishvananda/netlink/class.go +++ b/vendor/github.com/vishvananda/netlink/class.go @@ -132,7 +132,10 @@ func (class *GenericClass) Type() string { return class.ClassType } -// ServiceCurve is the way the HFSC curve are represented +// ServiceCurve is a nondecreasing function of some time unit, returning the amount of service +// (an allowed or allocated amount of bandwidth) at some specific point in time. The purpose of it +// should be subconsciously obvious: if a class was allowed to transfer not less than the amount +// specified by its service curve, then the service curve is not violated. type ServiceCurve struct { m1 uint32 d uint32 @@ -144,6 +147,21 @@ func (c *ServiceCurve) Attrs() (uint32, uint32, uint32) { return c.m1, c.d, c.m2 } +// Burst returns the burst rate (m1) of the curve +func (c *ServiceCurve) Burst() uint32 { + return c.m1 +} + +// Delay return the delay (d) of the curve +func (c *ServiceCurve) Delay() uint32 { + return c.d +} + +// Rate returns the rate (m2) of the curve +func (c *ServiceCurve) Rate() uint32 { + return c.m2 +} + // HfscClass is a representation of the HFSC class type HfscClass struct { ClassAttrs @@ -152,35 +170,44 @@ type HfscClass struct { Usc ServiceCurve } -// SetUsc sets the Usc curve +// SetUsc sets the USC curve. The bandwidth (m1 and m2) is specified in bits and the delay in +// seconds. func (hfsc *HfscClass) SetUsc(m1 uint32, d uint32, m2 uint32) { - hfsc.Usc = ServiceCurve{m1: m1 / 8, d: d, m2: m2 / 8} + hfsc.Usc = ServiceCurve{m1: m1, d: d, m2: m2} } -// SetFsc sets the Fsc curve +// SetFsc sets the Fsc curve. The bandwidth (m1 and m2) is specified in bits and the delay in +// seconds. func (hfsc *HfscClass) SetFsc(m1 uint32, d uint32, m2 uint32) { - hfsc.Fsc = ServiceCurve{m1: m1 / 8, d: d, m2: m2 / 8} + hfsc.Fsc = ServiceCurve{m1: m1, d: d, m2: m2} } -// SetRsc sets the Rsc curve +// SetRsc sets the Rsc curve. The bandwidth (m1 and m2) is specified in bits and the delay in +// seconds. func (hfsc *HfscClass) SetRsc(m1 uint32, d uint32, m2 uint32) { - hfsc.Rsc = ServiceCurve{m1: m1 / 8, d: d, m2: m2 / 8} + hfsc.Rsc = ServiceCurve{m1: m1, d: d, m2: m2} } -// SetSC implements the SC from the tc CLI +// SetSC implements the SC from the `tc` CLI. This function behaves the same as if one would set the +// USC through the `tc` command-line tool. This means bandwidth (m1 and m2) is specified in bits and +// the delay in ms. func (hfsc *HfscClass) SetSC(m1 uint32, d uint32, m2 uint32) { - hfsc.Rsc = ServiceCurve{m1: m1 / 8, d: d, m2: m2 / 8} - hfsc.Fsc = ServiceCurve{m1: m1 / 8, d: d, m2: m2 / 8} + hfsc.SetRsc(m1, d, m2) + hfsc.SetFsc(m1, d, m2) } -// SetUL implements the UL from the tc CLI +// SetUL implements the UL from the `tc` CLI. This function behaves the same as if one would set the +// USC through the `tc` command-line tool. This means bandwidth (m1 and m2) is specified in bits and +// the delay in ms. func (hfsc *HfscClass) SetUL(m1 uint32, d uint32, m2 uint32) { - hfsc.Usc = ServiceCurve{m1: m1 / 8, d: d, m2: m2 / 8} + hfsc.SetUsc(m1, d, m2) } -// SetLS implements the LS from the tc CLI +// SetLS implements the LS from the `tc` CLI. This function behaves the same as if one would set the +// USC through the `tc` command-line tool. This means bandwidth (m1 and m2) is specified in bits and +// the delay in ms. func (hfsc *HfscClass) SetLS(m1 uint32, d uint32, m2 uint32) { - hfsc.Fsc = ServiceCurve{m1: m1 / 8, d: d, m2: m2 / 8} + hfsc.SetFsc(m1, d, m2) } // NewHfscClass returns a new HFSC struct with the set parameters @@ -193,6 +220,7 @@ func NewHfscClass(attrs ClassAttrs) *HfscClass { } } +// String() returns a string that contains the information and attributes of the HFSC class func (hfsc *HfscClass) String() string { return fmt.Sprintf( "{%s -- {RSC: {m1=%d d=%d m2=%d}} {FSC: {m1=%d d=%d m2=%d}} {USC: {m1=%d d=%d m2=%d}}}", diff --git a/vendor/github.com/vishvananda/netlink/class_linux.go b/vendor/github.com/vishvananda/netlink/class_linux.go index 31091e501..e664ade7f 100644 --- a/vendor/github.com/vishvananda/netlink/class_linux.go +++ b/vendor/github.com/vishvananda/netlink/class_linux.go @@ -43,12 +43,12 @@ func NewHtbClass(attrs ClassAttrs, cattrs HtbClassAttrs) *HtbClass { if buffer == 0 { buffer = uint32(float64(rate)/Hz() + float64(mtu)) } - buffer = uint32(Xmittime(rate, buffer)) + buffer = Xmittime(rate, buffer) if cbuffer == 0 { cbuffer = uint32(float64(ceil)/Hz() + float64(mtu)) } - cbuffer = uint32(Xmittime(ceil, cbuffer)) + cbuffer = Xmittime(ceil, cbuffer) return &HtbClass{ ClassAttrs: attrs, @@ -56,9 +56,9 @@ func NewHtbClass(attrs ClassAttrs, cattrs HtbClassAttrs) *HtbClass { Ceil: ceil, Buffer: buffer, Cbuffer: cbuffer, - Quantum: 10, Level: 0, - Prio: 0, + Prio: cattrs.Prio, + Quantum: cattrs.Quantum, } } @@ -179,12 +179,15 @@ func classPayload(req *nl.NetlinkRequest, class Class) error { case "hfsc": hfsc := class.(*HfscClass) opt := nl.HfscCopt{} - opt.Rsc.Set(hfsc.Rsc.Attrs()) - opt.Fsc.Set(hfsc.Fsc.Attrs()) - opt.Usc.Set(hfsc.Usc.Attrs()) - options.AddRtAttr(nl.TCA_HFSC_RSC, nl.SerializeHfscCurve(&opt.Rsc)) - options.AddRtAttr(nl.TCA_HFSC_FSC, nl.SerializeHfscCurve(&opt.Fsc)) - options.AddRtAttr(nl.TCA_HFSC_USC, nl.SerializeHfscCurve(&opt.Usc)) + rm1, rd, rm2 := hfsc.Rsc.Attrs() + opt.Rsc.Set(rm1/8, rd, rm2/8) + fm1, fd, fm2 := hfsc.Fsc.Attrs() + opt.Fsc.Set(fm1/8, fd, fm2/8) + um1, ud, um2 := hfsc.Usc.Attrs() + opt.Usc.Set(um1/8, ud, um2/8) + nl.NewRtAttrChild(options, nl.TCA_HFSC_RSC, nl.SerializeHfscCurve(&opt.Rsc)) + nl.NewRtAttrChild(options, nl.TCA_HFSC_FSC, nl.SerializeHfscCurve(&opt.Fsc)) + nl.NewRtAttrChild(options, nl.TCA_HFSC_USC, nl.SerializeHfscCurve(&opt.Usc)) } req.AddData(options) return nil @@ -315,11 +318,11 @@ func parseHfscClassData(class Class, data []syscall.NetlinkRouteAttr) (bool, err m1, d, m2 := nl.DeserializeHfscCurve(datum.Value).Attrs() switch datum.Attr.Type { case nl.TCA_HFSC_RSC: - hfsc.Rsc = ServiceCurve{m1: m1, d: d, m2: m2} + hfsc.Rsc = ServiceCurve{m1: m1 * 8, d: d, m2: m2 * 8} case nl.TCA_HFSC_FSC: - hfsc.Fsc = ServiceCurve{m1: m1, d: d, m2: m2} + hfsc.Fsc = ServiceCurve{m1: m1 * 8, d: d, m2: m2 * 8} case nl.TCA_HFSC_USC: - hfsc.Usc = ServiceCurve{m1: m1, d: d, m2: m2} + hfsc.Usc = ServiceCurve{m1: m1 * 8, d: d, m2: m2 * 8} } } return detailed, nil diff --git a/vendor/github.com/vishvananda/netlink/conntrack_linux.go b/vendor/github.com/vishvananda/netlink/conntrack_linux.go index 4bff0dcba..ab91f4e55 100644 --- a/vendor/github.com/vishvananda/netlink/conntrack_linux.go +++ b/vendor/github.com/vishvananda/netlink/conntrack_linux.go @@ -318,18 +318,25 @@ func parseRawData(data []byte) *ConntrackFlow { // --mask-src ip Source mask address // --mask-dst ip Destination mask address +// Layer 4 Protocol common parameters and options: +// TCP, UDP, SCTP, UDPLite and DCCP +// --sport, --orig-port-src port Source port in original direction +// --dport, --orig-port-dst port Destination port in original direction + // Filter types type ConntrackFilterType uint8 const ( - ConntrackOrigSrcIP = iota // -orig-src ip Source address from original direction - ConntrackOrigDstIP // -orig-dst ip Destination address from original direction - ConntrackReplySrcIP // --reply-src ip Reply Source IP - ConntrackReplyDstIP // --reply-dst ip Reply Destination IP - ConntrackReplyAnyIP // Match source or destination reply IP - ConntrackNatSrcIP = ConntrackReplySrcIP // deprecated use instead ConntrackReplySrcIP - ConntrackNatDstIP = ConntrackReplyDstIP // deprecated use instead ConntrackReplyDstIP - ConntrackNatAnyIP = ConntrackReplyAnyIP // deprecated use instaed ConntrackReplyAnyIP + ConntrackOrigSrcIP = iota // -orig-src ip Source address from original direction + ConntrackOrigDstIP // -orig-dst ip Destination address from original direction + ConntrackReplySrcIP // --reply-src ip Reply Source IP + ConntrackReplyDstIP // --reply-dst ip Reply Destination IP + ConntrackReplyAnyIP // Match source or destination reply IP + ConntrackOrigSrcPort // --orig-port-src port Source port in original direction + ConntrackOrigDstPort // --orig-port-dst port Destination port in original direction + ConntrackNatSrcIP = ConntrackReplySrcIP // deprecated use instead ConntrackReplySrcIP + ConntrackNatDstIP = ConntrackReplyDstIP // deprecated use instead ConntrackReplyDstIP + ConntrackNatAnyIP = ConntrackReplyAnyIP // deprecated use instead ConntrackReplyAnyIP ) type CustomConntrackFilter interface { @@ -339,7 +346,9 @@ type CustomConntrackFilter interface { } type ConntrackFilter struct { - ipFilter map[ConntrackFilterType]net.IP + ipFilter map[ConntrackFilterType]net.IP + portFilter map[ConntrackFilterType]uint16 + protoFilter uint8 } // AddIP adds an IP to the conntrack filter @@ -354,38 +363,89 @@ func (f *ConntrackFilter) AddIP(tp ConntrackFilterType, ip net.IP) error { return nil } +// AddPort adds a Port to the conntrack filter if the Layer 4 protocol allows it +func (f *ConntrackFilter) AddPort(tp ConntrackFilterType, port uint16) error { + switch f.protoFilter { + // TCP, UDP, DCCP, SCTP, UDPLite + case 6, 17, 33, 132, 136: + default: + return fmt.Errorf("Filter attribute not available without a valid Layer 4 protocol: %d", f.protoFilter) + } + + if f.portFilter == nil { + f.portFilter = make(map[ConntrackFilterType]uint16) + } + if _, ok := f.portFilter[tp]; ok { + return errors.New("Filter attribute already present") + } + f.portFilter[tp] = port + return nil +} + +// AddProtocol adds the Layer 4 protocol to the conntrack filter +func (f *ConntrackFilter) AddProtocol(proto uint8) error { + if f.protoFilter != 0 { + return errors.New("Filter attribute already present") + } + f.protoFilter = proto + return nil +} + // MatchConntrackFlow applies the filter to the flow and returns true if the flow matches the filter // false otherwise func (f *ConntrackFilter) MatchConntrackFlow(flow *ConntrackFlow) bool { - if len(f.ipFilter) == 0 { + if len(f.ipFilter) == 0 && len(f.portFilter) == 0 && f.protoFilter == 0 { // empty filter always not match return false } - match := true - // -orig-src ip Source address from original direction - if elem, found := f.ipFilter[ConntrackOrigSrcIP]; found { - match = match && elem.Equal(flow.Forward.SrcIP) + // -p, --protonum proto Layer 4 Protocol, eg. 'tcp' + if f.protoFilter != 0 && flow.Forward.Protocol != f.protoFilter { + // different Layer 4 protocol always not match + return false } - // -orig-dst ip Destination address from original direction - if elem, found := f.ipFilter[ConntrackOrigDstIP]; match && found { - match = match && elem.Equal(flow.Forward.DstIP) - } + match := true - // -src-nat ip Source NAT ip - if elem, found := f.ipFilter[ConntrackReplySrcIP]; match && found { - match = match && elem.Equal(flow.Reverse.SrcIP) - } + // IP conntrack filter + if len(f.ipFilter) > 0 { + // -orig-src ip Source address from original direction + if elem, found := f.ipFilter[ConntrackOrigSrcIP]; found { + match = match && elem.Equal(flow.Forward.SrcIP) + } + + // -orig-dst ip Destination address from original direction + if elem, found := f.ipFilter[ConntrackOrigDstIP]; match && found { + match = match && elem.Equal(flow.Forward.DstIP) + } - // -dst-nat ip Destination NAT ip - if elem, found := f.ipFilter[ConntrackReplyDstIP]; match && found { - match = match && elem.Equal(flow.Reverse.DstIP) + // -src-nat ip Source NAT ip + if elem, found := f.ipFilter[ConntrackReplySrcIP]; match && found { + match = match && elem.Equal(flow.Reverse.SrcIP) + } + + // -dst-nat ip Destination NAT ip + if elem, found := f.ipFilter[ConntrackReplyDstIP]; match && found { + match = match && elem.Equal(flow.Reverse.DstIP) + } + + // Match source or destination reply IP + if elem, found := f.ipFilter[ConntrackReplyAnyIP]; match && found { + match = match && (elem.Equal(flow.Reverse.SrcIP) || elem.Equal(flow.Reverse.DstIP)) + } } - // Match source or destination reply IP - if elem, found := f.ipFilter[ConntrackReplyAnyIP]; match && found { - match = match && (elem.Equal(flow.Reverse.SrcIP) || elem.Equal(flow.Reverse.DstIP)) + // Layer 4 Port filter + if len(f.portFilter) > 0 { + // -orig-port-src port Source port from original direction + if elem, found := f.portFilter[ConntrackOrigSrcPort]; match && found { + match = match && elem == flow.Forward.SrcPort + } + + // -orig-port-dst port Destination port from original direction + if elem, found := f.portFilter[ConntrackOrigDstPort]; match && found { + match = match && elem == flow.Forward.DstPort + } } return match diff --git a/vendor/github.com/vishvananda/netlink/filter.go b/vendor/github.com/vishvananda/netlink/filter.go index 88792eab0..2dc34b995 100644 --- a/vendor/github.com/vishvananda/netlink/filter.go +++ b/vendor/github.com/vishvananda/netlink/filter.go @@ -213,10 +213,11 @@ const ( type TunnelKeyAction struct { ActionAttrs - Action TunnelKeyAct - SrcAddr net.IP - DstAddr net.IP - KeyID uint32 + Action TunnelKeyAct + SrcAddr net.IP + DstAddr net.IP + KeyID uint32 + DestPort uint16 } func (action *TunnelKeyAction) Type() string { diff --git a/vendor/github.com/vishvananda/netlink/filter_linux.go b/vendor/github.com/vishvananda/netlink/filter_linux.go index c56f314cd..ef6fabe81 100644 --- a/vendor/github.com/vishvananda/netlink/filter_linux.go +++ b/vendor/github.com/vishvananda/netlink/filter_linux.go @@ -6,7 +6,6 @@ import ( "encoding/hex" "errors" "fmt" - "net" "syscall" "github.com/vishvananda/netlink/nl" @@ -89,7 +88,7 @@ func NewFw(attrs FilterAttrs, fattrs FilterFwAttrs) (*Fw, error) { if CalcRtable(&police.Rate, rtab[:], rcellLog, fattrs.Mtu, linklayer) < 0 { return nil, errors.New("TBF: failed to calculate rate table") } - police.Burst = uint32(Xmittime(uint64(police.Rate.Rate), uint32(buffer))) + police.Burst = Xmittime(uint64(police.Rate.Rate), uint32(buffer)) } police.Mtu = fattrs.Mtu if police.PeakRate.Rate != 0 { @@ -456,6 +455,9 @@ func EncodeActions(attr *nl.RtAttr, actions []Action) error { } else { return fmt.Errorf("invalid dst addr %s for tunnel_key action", action.DstAddr) } + if action.DestPort != 0 { + aopts.AddRtAttr(nl.TCA_TUNNEL_KEY_ENC_DST_PORT, htons(action.DestPort)) + } } case *SkbEditAction: table := attr.AddRtAttr(tabIndex, nil) @@ -566,12 +568,12 @@ func parseActions(tables []syscall.NetlinkRouteAttr) ([]Action, error) { action.(*TunnelKeyAction).Action = TunnelKeyAct(tun.Action) case nl.TCA_TUNNEL_KEY_ENC_KEY_ID: action.(*TunnelKeyAction).KeyID = networkOrder.Uint32(adatum.Value[0:4]) - case nl.TCA_TUNNEL_KEY_ENC_IPV6_SRC: - case nl.TCA_TUNNEL_KEY_ENC_IPV4_SRC: - action.(*TunnelKeyAction).SrcAddr = net.IP(adatum.Value[:]) - case nl.TCA_TUNNEL_KEY_ENC_IPV6_DST: - case nl.TCA_TUNNEL_KEY_ENC_IPV4_DST: - action.(*TunnelKeyAction).DstAddr = net.IP(adatum.Value[:]) + case nl.TCA_TUNNEL_KEY_ENC_IPV6_SRC, nl.TCA_TUNNEL_KEY_ENC_IPV4_SRC: + action.(*TunnelKeyAction).SrcAddr = adatum.Value[:] + case nl.TCA_TUNNEL_KEY_ENC_IPV6_DST, nl.TCA_TUNNEL_KEY_ENC_IPV4_DST: + action.(*TunnelKeyAction).DstAddr = adatum.Value[:] + case nl.TCA_TUNNEL_KEY_ENC_DST_PORT: + action.(*TunnelKeyAction).DestPort = ntohs(adatum.Value) } case "skbedit": switch adatum.Attr.Type { @@ -783,7 +785,7 @@ func CalcRtable(rate *nl.TcRateSpec, rtab []uint32, cellLog int, mtu uint32, lin } for i := 0; i < 256; i++ { sz = AdjustSize(uint((i+1)<<uint32(cellLog)), uint(mpu), linklayer) - rtab[i] = uint32(Xmittime(uint64(bps), uint32(sz))) + rtab[i] = Xmittime(uint64(bps), uint32(sz)) } rate.CellAlign = -1 rate.CellLog = uint8(cellLog) diff --git a/vendor/github.com/vishvananda/netlink/go.mod b/vendor/github.com/vishvananda/netlink/go.mod index 09ee60e77..9d8455ee7 100644 --- a/vendor/github.com/vishvananda/netlink/go.mod +++ b/vendor/github.com/vishvananda/netlink/go.mod @@ -3,6 +3,6 @@ module github.com/vishvananda/netlink go 1.12 require ( - github.com/vishvananda/netns v0.0.0-20191106174202-0a2b9b5464df - golang.org/x/sys v0.0.0-20190606203320-7fc4e5ec1444 + github.com/vishvananda/netns v0.0.0-20200728191858-db3c7e526aae + golang.org/x/sys v0.0.0-20200728102440-3e129f6d46b1 ) diff --git a/vendor/github.com/vishvananda/netlink/go.sum b/vendor/github.com/vishvananda/netlink/go.sum index 402d14ec5..ed5d30891 100644 --- a/vendor/github.com/vishvananda/netlink/go.sum +++ b/vendor/github.com/vishvananda/netlink/go.sum @@ -1,4 +1,6 @@ -github.com/vishvananda/netns v0.0.0-20191106174202-0a2b9b5464df h1:OviZH7qLw/7ZovXvuNyL3XQl8UFofeikI1NW1Gypu7k= -github.com/vishvananda/netns v0.0.0-20191106174202-0a2b9b5464df/go.mod h1:JP3t17pCcGlemwknint6hfoeCVQrEMVwxRLRjXpq+BU= -golang.org/x/sys v0.0.0-20190606203320-7fc4e5ec1444 h1:/d2cWp6PSamH4jDPFLyO150psQdqvtoNX8Zjg3AQ31g= -golang.org/x/sys v0.0.0-20190606203320-7fc4e5ec1444/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs= +github.com/vishvananda/netns v0.0.0-20200728191858-db3c7e526aae h1:4hwBBUfQCFe3Cym0ZtKyq7L16eZUtYKs+BaHDN6mAns= +github.com/vishvananda/netns v0.0.0-20200728191858-db3c7e526aae/go.mod h1:DD4vA1DwXk04H54A1oHXtwZmA0grkVMdPxx/VGLCah0= +golang.org/x/sys v0.0.0-20200217220822-9197077df867 h1:JoRuNIf+rpHl+VhScRQQvzbHed86tKkqwPMV34T8myw= +golang.org/x/sys v0.0.0-20200217220822-9197077df867/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs= +golang.org/x/sys v0.0.0-20200728102440-3e129f6d46b1 h1:sIky/MyNRSHTrdxfsiUSS4WIAMvInbeXljJz+jDjeYE= +golang.org/x/sys v0.0.0-20200728102440-3e129f6d46b1/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs= diff --git a/vendor/github.com/vishvananda/netlink/handle_unspecified.go b/vendor/github.com/vishvananda/netlink/handle_unspecified.go index ef914dcb8..df341f706 100644 --- a/vendor/github.com/vishvananda/netlink/handle_unspecified.go +++ b/vendor/github.com/vishvananda/netlink/handle_unspecified.go @@ -85,7 +85,7 @@ func (h *Handle) LinkSetVfRate(link Link, vf, minRate, maxRate int) error { return ErrNotImplemented } -func (h *Handle) LinkSetMaster(link Link, master *Bridge) error { +func (h *Handle) LinkSetMaster(link Link, master Link) error { return ErrNotImplemented } diff --git a/vendor/github.com/vishvananda/netlink/inet_diag.go b/vendor/github.com/vishvananda/netlink/inet_diag.go new file mode 100644 index 000000000..72c1fcb59 --- /dev/null +++ b/vendor/github.com/vishvananda/netlink/inet_diag.go @@ -0,0 +1,30 @@ +package netlink + +// INET_DIAG constatns +const ( + INET_DIAG_NONE = iota + INET_DIAG_MEMINFO + INET_DIAG_INFO + INET_DIAG_VEGASINFO + INET_DIAG_CONG + INET_DIAG_TOS + INET_DIAG_TCLASS + INET_DIAG_SKMEMINFO + INET_DIAG_SHUTDOWN + INET_DIAG_DCTCPINFO + INET_DIAG_PROTOCOL + INET_DIAG_SKV6ONLY + INET_DIAG_LOCALS + INET_DIAG_PEERS + INET_DIAG_PAD + INET_DIAG_MARK + INET_DIAG_BBRINFO + INET_DIAG_CLASS_ID + INET_DIAG_MD5SIG + INET_DIAG_MAX +) + +type InetDiagTCPInfoResp struct { + InetDiagMsg *Socket + TCPInfo *TCPInfo +} diff --git a/vendor/github.com/vishvananda/netlink/ipset_linux.go b/vendor/github.com/vishvananda/netlink/ipset_linux.go new file mode 100644 index 000000000..5487fc1cc --- /dev/null +++ b/vendor/github.com/vishvananda/netlink/ipset_linux.go @@ -0,0 +1,335 @@ +package netlink + +import ( + "log" + "net" + "syscall" + + "github.com/vishvananda/netlink/nl" + "golang.org/x/sys/unix" +) + +// IPSetEntry is used for adding, updating, retreiving and deleting entries +type IPSetEntry struct { + Comment string + MAC net.HardwareAddr + IP net.IP + Timeout *uint32 + Packets *uint64 + Bytes *uint64 + + Replace bool // replace existing entry +} + +// IPSetResult is the result of a dump request for a set +type IPSetResult struct { + Nfgenmsg *nl.Nfgenmsg + Protocol uint8 + Revision uint8 + Family uint8 + Flags uint8 + SetName string + TypeName string + + HashSize uint32 + NumEntries uint32 + MaxElements uint32 + References uint32 + SizeInMemory uint32 + CadtFlags uint32 + Timeout *uint32 + + Entries []IPSetEntry +} + +// IpsetCreateOptions is the options struct for creating a new ipset +type IpsetCreateOptions struct { + Replace bool // replace existing ipset + Timeout *uint32 + Counters bool + Comments bool + Skbinfo bool +} + +// IpsetProtocol returns the ipset protocol version from the kernel +func IpsetProtocol() (uint8, error) { + return pkgHandle.IpsetProtocol() +} + +// IpsetCreate creates a new ipset +func IpsetCreate(setname, typename string, options IpsetCreateOptions) error { + return pkgHandle.IpsetCreate(setname, typename, options) +} + +// IpsetDestroy destroys an existing ipset +func IpsetDestroy(setname string) error { + return pkgHandle.IpsetDestroy(setname) +} + +// IpsetFlush flushes an existing ipset +func IpsetFlush(setname string) error { + return pkgHandle.IpsetFlush(setname) +} + +// IpsetList dumps an specific ipset. +func IpsetList(setname string) (*IPSetResult, error) { + return pkgHandle.IpsetList(setname) +} + +// IpsetListAll dumps all ipsets. +func IpsetListAll() ([]IPSetResult, error) { + return pkgHandle.IpsetListAll() +} + +// IpsetAdd adds an entry to an existing ipset. +func IpsetAdd(setname string, entry *IPSetEntry) error { + return pkgHandle.ipsetAddDel(nl.IPSET_CMD_ADD, setname, entry) +} + +// IpsetDele deletes an entry from an existing ipset. +func IpsetDel(setname string, entry *IPSetEntry) error { + return pkgHandle.ipsetAddDel(nl.IPSET_CMD_DEL, setname, entry) +} + +func (h *Handle) IpsetProtocol() (uint8, error) { + req := h.newIpsetRequest(nl.IPSET_CMD_PROTOCOL) + msgs, err := req.Execute(unix.NETLINK_NETFILTER, 0) + + if err != nil { + return 0, err + } + + return ipsetUnserialize(msgs).Protocol, nil +} + +func (h *Handle) IpsetCreate(setname, typename string, options IpsetCreateOptions) error { + req := h.newIpsetRequest(nl.IPSET_CMD_CREATE) + + if !options.Replace { + req.Flags |= unix.NLM_F_EXCL + } + + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_SETNAME, nl.ZeroTerminated(setname))) + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_TYPENAME, nl.ZeroTerminated(typename))) + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_REVISION, nl.Uint8Attr(0))) + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_FAMILY, nl.Uint8Attr(0))) + + data := nl.NewRtAttr(nl.IPSET_ATTR_DATA|int(nl.NLA_F_NESTED), nil) + + if timeout := options.Timeout; timeout != nil { + data.AddChild(&nl.Uint32Attribute{Type: nl.IPSET_ATTR_TIMEOUT | nl.NLA_F_NET_BYTEORDER, Value: *timeout}) + } + + var cadtFlags uint32 + + if options.Comments { + cadtFlags |= nl.IPSET_FLAG_WITH_COMMENT + } + if options.Counters { + cadtFlags |= nl.IPSET_FLAG_WITH_COUNTERS + } + if options.Skbinfo { + cadtFlags |= nl.IPSET_FLAG_WITH_SKBINFO + } + + if cadtFlags != 0 { + data.AddChild(&nl.Uint32Attribute{Type: nl.IPSET_ATTR_CADT_FLAGS | nl.NLA_F_NET_BYTEORDER, Value: cadtFlags}) + } + + req.AddData(data) + _, err := ipsetExecute(req) + return err +} + +func (h *Handle) IpsetDestroy(setname string) error { + req := h.newIpsetRequest(nl.IPSET_CMD_DESTROY) + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_SETNAME, nl.ZeroTerminated(setname))) + _, err := ipsetExecute(req) + return err +} + +func (h *Handle) IpsetFlush(setname string) error { + req := h.newIpsetRequest(nl.IPSET_CMD_FLUSH) + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_SETNAME, nl.ZeroTerminated(setname))) + _, err := ipsetExecute(req) + return err +} + +func (h *Handle) IpsetList(name string) (*IPSetResult, error) { + req := h.newIpsetRequest(nl.IPSET_CMD_LIST) + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_SETNAME, nl.ZeroTerminated(name))) + + msgs, err := ipsetExecute(req) + if err != nil { + return nil, err + } + + result := ipsetUnserialize(msgs) + return &result, nil +} + +func (h *Handle) IpsetListAll() ([]IPSetResult, error) { + req := h.newIpsetRequest(nl.IPSET_CMD_LIST) + + msgs, err := ipsetExecute(req) + if err != nil { + return nil, err + } + + result := make([]IPSetResult, len(msgs)) + for i, msg := range msgs { + result[i].unserialize(msg) + } + + return result, nil +} + +func (h *Handle) ipsetAddDel(nlCmd int, setname string, entry *IPSetEntry) error { + req := h.newIpsetRequest(nlCmd) + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_SETNAME, nl.ZeroTerminated(setname))) + data := nl.NewRtAttr(nl.IPSET_ATTR_DATA|int(nl.NLA_F_NESTED), nil) + + if !entry.Replace { + req.Flags |= unix.NLM_F_EXCL + } + + if entry.Timeout != nil { + data.AddChild(&nl.Uint32Attribute{Type: nl.IPSET_ATTR_TIMEOUT | nl.NLA_F_NET_BYTEORDER, Value: *entry.Timeout}) + } + if entry.MAC != nil { + data.AddChild(nl.NewRtAttr(nl.IPSET_ATTR_ETHER, entry.MAC)) + } + + data.AddChild(&nl.Uint32Attribute{Type: nl.IPSET_ATTR_LINENO | nl.NLA_F_NET_BYTEORDER, Value: 0}) + req.AddData(data) + + _, err := ipsetExecute(req) + return err +} + +func (h *Handle) newIpsetRequest(cmd int) *nl.NetlinkRequest { + req := h.newNetlinkRequest(cmd|(unix.NFNL_SUBSYS_IPSET<<8), nl.GetIpsetFlags(cmd)) + + // Add the netfilter header + msg := &nl.Nfgenmsg{ + NfgenFamily: uint8(unix.AF_NETLINK), + Version: nl.NFNETLINK_V0, + ResId: 0, + } + req.AddData(msg) + req.AddData(nl.NewRtAttr(nl.IPSET_ATTR_PROTOCOL, nl.Uint8Attr(nl.IPSET_PROTOCOL))) + + return req +} + +func ipsetExecute(req *nl.NetlinkRequest) (msgs [][]byte, err error) { + msgs, err = req.Execute(unix.NETLINK_NETFILTER, 0) + + if err != nil { + if errno := int(err.(syscall.Errno)); errno >= nl.IPSET_ERR_PRIVATE { + err = nl.IPSetError(uintptr(errno)) + } + } + return +} + +func ipsetUnserialize(msgs [][]byte) (result IPSetResult) { + for _, msg := range msgs { + result.unserialize(msg) + } + return result +} + +func (result *IPSetResult) unserialize(msg []byte) { + result.Nfgenmsg = nl.DeserializeNfgenmsg(msg) + + for attr := range nl.ParseAttributes(msg[4:]) { + switch attr.Type { + case nl.IPSET_ATTR_PROTOCOL: + result.Protocol = attr.Value[0] + case nl.IPSET_ATTR_SETNAME: + result.SetName = nl.BytesToString(attr.Value) + case nl.IPSET_ATTR_TYPENAME: + result.TypeName = nl.BytesToString(attr.Value) + case nl.IPSET_ATTR_REVISION: + result.Revision = attr.Value[0] + case nl.IPSET_ATTR_FAMILY: + result.Family = attr.Value[0] + case nl.IPSET_ATTR_FLAGS: + result.Flags = attr.Value[0] + case nl.IPSET_ATTR_DATA | nl.NLA_F_NESTED: + result.parseAttrData(attr.Value) + case nl.IPSET_ATTR_ADT | nl.NLA_F_NESTED: + result.parseAttrADT(attr.Value) + default: + log.Printf("unknown ipset attribute from kernel: %+v %v", attr, attr.Type&nl.NLA_TYPE_MASK) + } + } +} + +func (result *IPSetResult) parseAttrData(data []byte) { + for attr := range nl.ParseAttributes(data) { + switch attr.Type { + case nl.IPSET_ATTR_HASHSIZE | nl.NLA_F_NET_BYTEORDER: + result.HashSize = attr.Uint32() + case nl.IPSET_ATTR_MAXELEM | nl.NLA_F_NET_BYTEORDER: + result.MaxElements = attr.Uint32() + case nl.IPSET_ATTR_TIMEOUT | nl.NLA_F_NET_BYTEORDER: + val := attr.Uint32() + result.Timeout = &val + case nl.IPSET_ATTR_ELEMENTS | nl.NLA_F_NET_BYTEORDER: + result.NumEntries = attr.Uint32() + case nl.IPSET_ATTR_REFERENCES | nl.NLA_F_NET_BYTEORDER: + result.References = attr.Uint32() + case nl.IPSET_ATTR_MEMSIZE | nl.NLA_F_NET_BYTEORDER: + result.SizeInMemory = attr.Uint32() + case nl.IPSET_ATTR_CADT_FLAGS | nl.NLA_F_NET_BYTEORDER: + result.CadtFlags = attr.Uint32() + default: + log.Printf("unknown ipset data attribute from kernel: %+v %v", attr, attr.Type&nl.NLA_TYPE_MASK) + } + } +} + +func (result *IPSetResult) parseAttrADT(data []byte) { + for attr := range nl.ParseAttributes(data) { + switch attr.Type { + case nl.IPSET_ATTR_DATA | nl.NLA_F_NESTED: + result.Entries = append(result.Entries, parseIPSetEntry(attr.Value)) + default: + log.Printf("unknown ADT attribute from kernel: %+v %v", attr, attr.Type&nl.NLA_TYPE_MASK) + } + } +} + +func parseIPSetEntry(data []byte) (entry IPSetEntry) { + for attr := range nl.ParseAttributes(data) { + switch attr.Type { + case nl.IPSET_ATTR_TIMEOUT | nl.NLA_F_NET_BYTEORDER: + val := attr.Uint32() + entry.Timeout = &val + case nl.IPSET_ATTR_BYTES | nl.NLA_F_NET_BYTEORDER: + val := attr.Uint64() + entry.Bytes = &val + case nl.IPSET_ATTR_PACKETS | nl.NLA_F_NET_BYTEORDER: + val := attr.Uint64() + entry.Packets = &val + case nl.IPSET_ATTR_ETHER: + entry.MAC = net.HardwareAddr(attr.Value) + case nl.IPSET_ATTR_COMMENT: + entry.Comment = nl.BytesToString(attr.Value) + case nl.IPSET_ATTR_IP | nl.NLA_F_NESTED: + for attr := range nl.ParseAttributes(attr.Value) { + switch attr.Type { + case nl.IPSET_ATTR_IP: + entry.IP = net.IP(attr.Value) + default: + log.Printf("unknown nested ADT attribute from kernel: %+v", attr) + } + } + default: + log.Printf("unknown ADT attribute from kernel: %+v", attr) + } + } + return +} diff --git a/vendor/github.com/vishvananda/netlink/link.go b/vendor/github.com/vishvananda/netlink/link.go index 886d88d1b..e2441bd71 100644 --- a/vendor/github.com/vishvananda/netlink/link.go +++ b/vendor/github.com/vishvananda/netlink/link.go @@ -65,6 +65,17 @@ type VfInfo struct { LinkState uint32 MaxTxRate uint32 // IFLA_VF_RATE Max TxRate MinTxRate uint32 // IFLA_VF_RATE Min TxRate + RxPackets uint64 + TxPackets uint64 + RxBytes uint64 + TxBytes uint64 + Multicast uint64 + Broadcast uint64 + RxDropped uint64 + TxDropped uint64 + + RssQuery uint32 + Trust uint32 } // LinkOperState represents the values of the IFLA_OPERSTATE link @@ -103,7 +114,8 @@ func (s LinkOperState) String() string { // NewLinkAttrs returns LinkAttrs structure filled with default values func NewLinkAttrs() LinkAttrs { return LinkAttrs{ - TxQLen: -1, + NetNsID: -1, + TxQLen: -1, } } @@ -196,10 +208,11 @@ type LinkStatistics64 struct { } type LinkXdp struct { - Fd int - Attached bool - Flags uint32 - ProgId uint32 + Fd int + Attached bool + AttachMode uint32 + Flags uint32 + ProgId uint32 } // Device links cannot be created via netlink. These links @@ -246,6 +259,7 @@ func (ifb *Ifb) Type() string { type Bridge struct { LinkAttrs MulticastSnooping *bool + AgeingTime *uint32 HelloTime *uint32 VlanFiltering *bool } @@ -338,6 +352,7 @@ type Veth struct { LinkAttrs PeerName string // veth on create only PeerHardwareAddr net.HardwareAddr + PeerNamespace interface{} } func (veth *Veth) Attrs() *LinkAttrs { @@ -348,6 +363,19 @@ func (veth *Veth) Type() string { return "veth" } +// Wireguard represent links of type "wireguard", see https://www.wireguard.com/ +type Wireguard struct { + LinkAttrs +} + +func (wg *Wireguard) Attrs() *LinkAttrs { + return &wg.LinkAttrs +} + +func (wg *Wireguard) Type() string { + return "wireguard" +} + // GenericLink links represent types that are not currently understood // by this netlink library. type GenericLink struct { @@ -878,10 +906,14 @@ type Ip6tnl struct { Remote net.IP Ttl uint8 Tos uint8 - EncapLimit uint8 Flags uint32 Proto uint8 FlowInfo uint32 + EncapLimit uint8 + EncapType uint16 + EncapFlags uint16 + EncapSport uint16 + EncapDport uint16 } func (ip6tnl *Ip6tnl) Attrs() *LinkAttrs { @@ -895,11 +927,13 @@ func (ip6tnl *Ip6tnl) Type() string { type Sittun struct { LinkAttrs Link uint32 - Local net.IP - Remote net.IP Ttl uint8 Tos uint8 PMtuDisc uint8 + Proto uint8 + Local net.IP + Remote net.IP + EncapLimit uint8 EncapType uint16 EncapFlags uint16 EncapSport uint16 diff --git a/vendor/github.com/vishvananda/netlink/link_linux.go b/vendor/github.com/vishvananda/netlink/link_linux.go index ec915a0b9..c02fa63b8 100644 --- a/vendor/github.com/vishvananda/netlink/link_linux.go +++ b/vendor/github.com/vishvananda/netlink/link_linux.go @@ -237,6 +237,37 @@ func (h *Handle) macvlanMACAddrChange(link Link, addrs []net.HardwareAddr, mode return err } +// LinkSetMacvlanMode sets the mode of a macvlan or macvtap link device. +// Note that passthrough mode cannot be set to and from and will fail. +// Equivalent to: `ip link set $link type (macvlan|macvtap) mode $mode +func LinkSetMacvlanMode(link Link, mode MacvlanMode) error { + return pkgHandle.LinkSetMacvlanMode(link, mode) +} + +// LinkSetMacvlanMode sets the mode of the macvlan or macvtap link device. +// Note that passthrough mode cannot be set to and from and will fail. +// Equivalent to: `ip link set $link type (macvlan|macvtap) mode $mode +func (h *Handle) LinkSetMacvlanMode(link Link, mode MacvlanMode) error { + base := link.Attrs() + h.ensureIndex(base) + req := h.newNetlinkRequest(unix.RTM_NEWLINK, unix.NLM_F_ACK) + + msg := nl.NewIfInfomsg(unix.AF_UNSPEC) + msg.Index = int32(base.Index) + req.AddData(msg) + + linkInfo := nl.NewRtAttr(unix.IFLA_LINKINFO, nil) + linkInfo.AddRtAttr(nl.IFLA_INFO_KIND, nl.NonZeroTerminated(link.Type())) + + data := linkInfo.AddRtAttr(nl.IFLA_INFO_DATA, nil) + data.AddRtAttr(nl.IFLA_MACVLAN_MODE, nl.Uint32Attr(macvlanModes[mode])) + + req.AddData(linkInfo) + + _, err := req.Execute(unix.NETLINK_ROUTE, 0) + return err +} + func BridgeSetMcastSnoop(link Link, on bool) error { return pkgHandle.BridgeSetMcastSnoop(link, on) } @@ -247,6 +278,16 @@ func (h *Handle) BridgeSetMcastSnoop(link Link, on bool) error { return h.linkModify(bridge, unix.NLM_F_ACK) } +func BridgeSetVlanFiltering(link Link, on bool) error { + return pkgHandle.BridgeSetVlanFiltering(link, on) +} + +func (h *Handle) BridgeSetVlanFiltering(link Link, on bool) error { + bridge := link.(*Bridge) + bridge.VlanFiltering = &on + return h.linkModify(bridge, unix.NLM_F_ACK) +} + func SetPromiscOn(link Link) error { return pkgHandle.SetPromiscOn(link) } @@ -1048,6 +1089,10 @@ func (h *Handle) LinkAdd(link Link) error { return h.linkModify(link, unix.NLM_F_CREATE|unix.NLM_F_EXCL|unix.NLM_F_ACK) } +func (h *Handle) LinkModify(link Link) error { + return h.linkModify(link, unix.NLM_F_REQUEST|unix.NLM_F_ACK) +} + func (h *Handle) linkModify(link Link, flags int) error { // TODO: support extra data for macvlan base := link.Attrs() @@ -1060,8 +1105,6 @@ func (h *Handle) linkModify(link Link, flags int) error { } if isTuntap { - // TODO: support user - // TODO: support group if tuntap.Mode < unix.IFF_TUN || tuntap.Mode > unix.IFF_TAP { return fmt.Errorf("Tuntap.Mode %v unknown", tuntap.Mode) } @@ -1089,21 +1132,64 @@ func (h *Handle) linkModify(link Link, flags int) error { } req.Flags |= uint16(tuntap.Mode) - + const TUN = "/dev/net/tun" for i := 0; i < queues; i++ { localReq := req - file, err := os.OpenFile("/dev/net/tun", os.O_RDWR, 0) + fd, err := unix.Open(TUN, os.O_RDWR|syscall.O_CLOEXEC, 0) if err != nil { cleanupFds(fds) return err } - fds = append(fds, file) - _, _, errno := unix.Syscall(unix.SYS_IOCTL, file.Fd(), uintptr(unix.TUNSETIFF), uintptr(unsafe.Pointer(&localReq))) + _, _, errno := unix.Syscall(unix.SYS_IOCTL, uintptr(fd), uintptr(unix.TUNSETIFF), uintptr(unsafe.Pointer(&localReq))) if errno != 0 { + // close the new fd + unix.Close(fd) + // and the already opened ones cleanupFds(fds) return fmt.Errorf("Tuntap IOCTL TUNSETIFF failed [%d], errno %v", i, errno) } + + _, _, errno = syscall.Syscall(syscall.SYS_IOCTL, uintptr(fd), syscall.TUNSETOWNER, uintptr(tuntap.Owner)) + if errno != 0 { + cleanupFds(fds) + return fmt.Errorf("Tuntap IOCTL TUNSETOWNER failed [%d], errno %v", i, errno) + } + + _, _, errno = syscall.Syscall(syscall.SYS_IOCTL, uintptr(fd), syscall.TUNSETGROUP, uintptr(tuntap.Group)) + if errno != 0 { + cleanupFds(fds) + return fmt.Errorf("Tuntap IOCTL TUNSETGROUP failed [%d], errno %v", i, errno) + } + + // Set the tun device to non-blocking before use. The below comment + // taken from: + // + // https://github.com/mistsys/tuntap/commit/161418c25003bbee77d085a34af64d189df62bea + // + // Note there is a complication because in go, if a device node is + // opened, go sets it to use nonblocking I/O. However a /dev/net/tun + // doesn't work with epoll until after the TUNSETIFF ioctl has been + // done. So we open the unix fd directly, do the ioctl, then put the + // fd in nonblocking mode, an then finally wrap it in a os.File, + // which will see the nonblocking mode and add the fd to the + // pollable set, so later on when we Read() from it blocked the + // calling thread in the kernel. + // + // See + // https://github.com/golang/go/issues/30426 + // which got exposed in go 1.13 by the fix to + // https://github.com/golang/go/issues/30624 + err = unix.SetNonblock(fd, true) + if err != nil { + cleanupFds(fds) + return fmt.Errorf("Tuntap set to non-blocking failed [%d], err %v", i, err) + } + + // create the file from the file descriptor and store it + file := os.NewFile(uintptr(fd), TUN) + fds = append(fds, file) + // 1) we only care for the name of the first tap in the multi queue set // 2) if the original name was empty, the localReq has now the actual name // @@ -1114,6 +1200,7 @@ func (h *Handle) linkModify(link Link, flags int) error { if i == 0 { link.Attrs().Name = strings.Trim(string(localReq.Name[:]), "\x00") } + } // only persist interface if NonPersist is NOT set @@ -1193,6 +1280,11 @@ func (h *Handle) linkModify(link Link, flags int) error { nameData := nl.NewRtAttr(unix.IFLA_IFNAME, nl.ZeroTerminated(base.Name)) req.AddData(nameData) + if base.Alias != "" { + alias := nl.NewRtAttr(unix.IFLA_IFALIAS, []byte(base.Alias)) + req.AddData(alias) + } + if base.MTU > 0 { mtu := nl.NewRtAttr(unix.IFLA_MTU, nl.Uint32Attr(uint32(base.MTU))) req.AddData(mtu) @@ -1272,12 +1364,28 @@ func (h *Handle) linkModify(link Link, flags int) error { if base.TxQLen >= 0 { peer.AddRtAttr(unix.IFLA_TXQLEN, nl.Uint32Attr(uint32(base.TxQLen))) } + if base.NumTxQueues > 0 { + peer.AddRtAttr(unix.IFLA_NUM_TX_QUEUES, nl.Uint32Attr(uint32(base.NumTxQueues))) + } + if base.NumRxQueues > 0 { + peer.AddRtAttr(unix.IFLA_NUM_RX_QUEUES, nl.Uint32Attr(uint32(base.NumRxQueues))) + } if base.MTU > 0 { peer.AddRtAttr(unix.IFLA_MTU, nl.Uint32Attr(uint32(base.MTU))) } if link.PeerHardwareAddr != nil { peer.AddRtAttr(unix.IFLA_ADDRESS, []byte(link.PeerHardwareAddr)) } + if link.PeerNamespace != nil { + switch ns := link.PeerNamespace.(type) { + case NsPid: + val := nl.Uint32Attr(uint32(ns)) + peer.AddRtAttr(unix.IFLA_NET_NS_PID, val) + case NsFd: + val := nl.Uint32Attr(uint32(ns)) + peer.AddRtAttr(unix.IFLA_NET_NS_FD, val) + } + } case *Vxlan: addVxlanAttrs(link, linkInfo) case *Bond: @@ -1509,7 +1617,11 @@ func LinkDeserialize(hdr *unix.NlMsghdr, m []byte) (Link, error) { return nil, err } - base := LinkAttrs{Index: int(msg.Index), RawFlags: msg.Flags, Flags: linkFlags(msg.Flags), EncapType: msg.EncapType()} + base := NewLinkAttrs() + base.Index = int(msg.Index) + base.RawFlags = msg.Flags + base.Flags = linkFlags(msg.Flags) + base.EncapType = msg.EncapType() if msg.Flags&unix.IFF_PROMISC != 0 { base.Promisc = 1 } @@ -1543,6 +1655,8 @@ func LinkDeserialize(hdr *unix.NlMsghdr, m []byte) (Link, error) { link = &Vlan{} case "veth": link = &Veth{} + case "wireguard": + link = &Wireguard{} case "vxlan": link = &Vxlan{} case "bond": @@ -2080,6 +2194,13 @@ func parseVlanData(link Link, data []syscall.NetlinkRouteAttr) { func parseVxlanData(link Link, data []syscall.NetlinkRouteAttr) { vxlan := link.(*Vxlan) for _, datum := range data { + // NOTE(vish): Apparently some messages can be sent with no value. + // We special case GBP here to not change existing + // functionality. It appears that GBP sends a datum.Value + // of null. + if len(datum.Value) == 0 && datum.Attr.Type != nl.IFLA_VXLAN_GBP { + continue + } switch datum.Attr.Type { case nl.IFLA_VXLAN_ID: vxlan.VxlanId = int(native.Uint32(datum.Value[0:4])) @@ -2513,7 +2634,8 @@ func parseLinkXdp(data []byte) (*LinkXdp, error) { case nl.IFLA_XDP_FD: xdp.Fd = int(native.Uint32(attr.Value[0:4])) case nl.IFLA_XDP_ATTACHED: - xdp.Attached = attr.Value[0] != 0 + xdp.AttachMode = uint32(attr.Value[0]) + xdp.Attached = xdp.AttachMode != 0 case nl.IFLA_XDP_FLAGS: xdp.Flags = native.Uint32(attr.Value[0:4]) case nl.IFLA_XDP_PROG_ID: @@ -2577,7 +2699,7 @@ func parseIptunData(link Link, data []syscall.NetlinkRouteAttr) { case nl.IFLA_IPTUN_ENCAP_FLAGS: iptun.EncapFlags = native.Uint16(datum.Value[0:2]) case nl.IFLA_IPTUN_COLLECT_METADATA: - iptun.FlowBased = int8(datum.Value[0]) != 0 + iptun.FlowBased = true } } } @@ -2601,10 +2723,14 @@ func addIp6tnlAttrs(ip6tnl *Ip6tnl, linkInfo *nl.RtAttr) { data.AddRtAttr(nl.IFLA_IPTUN_TTL, nl.Uint8Attr(ip6tnl.Ttl)) data.AddRtAttr(nl.IFLA_IPTUN_TOS, nl.Uint8Attr(ip6tnl.Tos)) - data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_LIMIT, nl.Uint8Attr(ip6tnl.EncapLimit)) data.AddRtAttr(nl.IFLA_IPTUN_FLAGS, nl.Uint32Attr(ip6tnl.Flags)) data.AddRtAttr(nl.IFLA_IPTUN_PROTO, nl.Uint8Attr(ip6tnl.Proto)) data.AddRtAttr(nl.IFLA_IPTUN_FLOWINFO, nl.Uint32Attr(ip6tnl.FlowInfo)) + data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_LIMIT, nl.Uint8Attr(ip6tnl.EncapLimit)) + data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_TYPE, nl.Uint16Attr(ip6tnl.EncapType)) + data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_FLAGS, nl.Uint16Attr(ip6tnl.EncapFlags)) + data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_SPORT, htons(ip6tnl.EncapSport)) + data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_DPORT, htons(ip6tnl.EncapDport)) } func parseIp6tnlData(link Link, data []syscall.NetlinkRouteAttr) { @@ -2616,17 +2742,25 @@ func parseIp6tnlData(link Link, data []syscall.NetlinkRouteAttr) { case nl.IFLA_IPTUN_REMOTE: ip6tnl.Remote = net.IP(datum.Value[:16]) case nl.IFLA_IPTUN_TTL: - ip6tnl.Ttl = uint8(datum.Value[0]) + ip6tnl.Ttl = datum.Value[0] case nl.IFLA_IPTUN_TOS: - ip6tnl.Tos = uint8(datum.Value[0]) - case nl.IFLA_IPTUN_ENCAP_LIMIT: - ip6tnl.EncapLimit = uint8(datum.Value[0]) + ip6tnl.Tos = datum.Value[0] case nl.IFLA_IPTUN_FLAGS: ip6tnl.Flags = native.Uint32(datum.Value[:4]) case nl.IFLA_IPTUN_PROTO: - ip6tnl.Proto = uint8(datum.Value[0]) + ip6tnl.Proto = datum.Value[0] case nl.IFLA_IPTUN_FLOWINFO: ip6tnl.FlowInfo = native.Uint32(datum.Value[:4]) + case nl.IFLA_IPTUN_ENCAP_LIMIT: + ip6tnl.EncapLimit = datum.Value[0] + case nl.IFLA_IPTUN_ENCAP_TYPE: + ip6tnl.EncapType = native.Uint16(datum.Value[0:2]) + case nl.IFLA_IPTUN_ENCAP_FLAGS: + ip6tnl.EncapFlags = native.Uint16(datum.Value[0:2]) + case nl.IFLA_IPTUN_ENCAP_SPORT: + ip6tnl.EncapSport = ntohs(datum.Value[0:2]) + case nl.IFLA_IPTUN_ENCAP_DPORT: + ip6tnl.EncapDport = ntohs(datum.Value[0:2]) } } } @@ -2653,8 +2787,10 @@ func addSittunAttrs(sittun *Sittun, linkInfo *nl.RtAttr) { data.AddRtAttr(nl.IFLA_IPTUN_TTL, nl.Uint8Attr(sittun.Ttl)) } + data.AddRtAttr(nl.IFLA_IPTUN_PROTO, nl.Uint8Attr(sittun.Proto)) data.AddRtAttr(nl.IFLA_IPTUN_TOS, nl.Uint8Attr(sittun.Tos)) data.AddRtAttr(nl.IFLA_IPTUN_PMTUDISC, nl.Uint8Attr(sittun.PMtuDisc)) + data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_LIMIT, nl.Uint8Attr(sittun.EncapLimit)) data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_TYPE, nl.Uint16Attr(sittun.EncapType)) data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_FLAGS, nl.Uint16Attr(sittun.EncapFlags)) data.AddRtAttr(nl.IFLA_IPTUN_ENCAP_SPORT, htons(sittun.EncapSport)) @@ -2670,11 +2806,13 @@ func parseSittunData(link Link, data []syscall.NetlinkRouteAttr) { case nl.IFLA_IPTUN_REMOTE: sittun.Remote = net.IP(datum.Value[0:4]) case nl.IFLA_IPTUN_TTL: - sittun.Ttl = uint8(datum.Value[0]) + sittun.Ttl = datum.Value[0] case nl.IFLA_IPTUN_TOS: - sittun.Tos = uint8(datum.Value[0]) + sittun.Tos = datum.Value[0] case nl.IFLA_IPTUN_PMTUDISC: - sittun.PMtuDisc = uint8(datum.Value[0]) + sittun.PMtuDisc = datum.Value[0] + case nl.IFLA_IPTUN_PROTO: + sittun.Proto = datum.Value[0] case nl.IFLA_IPTUN_ENCAP_TYPE: sittun.EncapType = native.Uint16(datum.Value[0:2]) case nl.IFLA_IPTUN_ENCAP_FLAGS: @@ -2761,6 +2899,9 @@ func addBridgeAttrs(bridge *Bridge, linkInfo *nl.RtAttr) { if bridge.MulticastSnooping != nil { data.AddRtAttr(nl.IFLA_BR_MCAST_SNOOPING, boolToByte(*bridge.MulticastSnooping)) } + if bridge.AgeingTime != nil { + data.AddRtAttr(nl.IFLA_BR_AGEING_TIME, nl.Uint32Attr(*bridge.AgeingTime)) + } if bridge.HelloTime != nil { data.AddRtAttr(nl.IFLA_BR_HELLO_TIME, nl.Uint32Attr(*bridge.HelloTime)) } @@ -2773,6 +2914,9 @@ func parseBridgeData(bridge Link, data []syscall.NetlinkRouteAttr) { br := bridge.(*Bridge) for _, datum := range data { switch datum.Attr.Type { + case nl.IFLA_BR_AGEING_TIME: + ageingTime := native.Uint32(datum.Value[0:4]) + br.AgeingTime = &ageingTime case nl.IFLA_BR_HELLO_TIME: helloTime := native.Uint32(datum.Value[0:4]) br.HelloTime = &helloTime @@ -2852,6 +2996,24 @@ func parseVfInfo(data []syscall.NetlinkRouteAttr, id int) VfInfo { vfr := nl.DeserializeVfRate(element.Value[:]) vf.MaxTxRate = vfr.MaxTxRate vf.MinTxRate = vfr.MinTxRate + case nl.IFLA_VF_STATS: + vfstats := nl.DeserializeVfStats(element.Value[:]) + vf.RxPackets = vfstats.RxPackets + vf.TxPackets = vfstats.TxPackets + vf.RxBytes = vfstats.RxBytes + vf.TxBytes = vfstats.TxBytes + vf.Multicast = vfstats.Multicast + vf.Broadcast = vfstats.Broadcast + vf.RxDropped = vfstats.RxDropped + vf.TxDropped = vfstats.TxDropped + + case nl.IFLA_VF_RSS_QUERY_EN: + result := nl.DeserializeVfRssQueryEn(element.Value) + vf.RssQuery = result.Setting + + case nl.IFLA_VF_TRUST: + result := nl.DeserializeVfTrust(element.Value) + vf.Trust = result.Setting } } return vf diff --git a/vendor/github.com/vishvananda/netlink/neigh_linux.go b/vendor/github.com/vishvananda/netlink/neigh_linux.go index cb3b55d35..fb220d141 100644 --- a/vendor/github.com/vishvananda/netlink/neigh_linux.go +++ b/vendor/github.com/vishvananda/netlink/neigh_linux.go @@ -243,6 +243,18 @@ func (h *Handle) NeighListExecute(msg Ndmsg) ([]Neigh, error) { // Ignore messages from other interfaces continue } + if msg.Family != 0 && ndm.Family != msg.Family { + continue + } + if msg.State != 0 && ndm.State != msg.State { + continue + } + if msg.Type != 0 && ndm.Type != msg.Type { + continue + } + if msg.Flags != 0 && ndm.Flags != msg.Flags { + continue + } neigh, err := NeighDeserialize(m) if err != nil { diff --git a/vendor/github.com/vishvananda/netlink/netlink_unspecified.go b/vendor/github.com/vishvananda/netlink/netlink_unspecified.go index 42d3acf91..71436f25c 100644 --- a/vendor/github.com/vishvananda/netlink/netlink_unspecified.go +++ b/vendor/github.com/vishvananda/netlink/netlink_unspecified.go @@ -16,7 +16,7 @@ func LinkSetMTU(link Link, mtu int) error { return ErrNotImplemented } -func LinkSetMaster(link Link, master *Bridge) error { +func LinkSetMaster(link Link, master Link) error { return ErrNotImplemented } @@ -72,6 +72,10 @@ func LinkSetXdpFd(link Link, fd int) error { return ErrNotImplemented } +func LinkSetXdpFdWithFlags(link Link, fd, flags int) error { + return ErrNotImplemented +} + func LinkSetARPOff(link Link) error { return ErrNotImplemented } diff --git a/vendor/github.com/vishvananda/netlink/nl/addr_linux.go b/vendor/github.com/vishvananda/netlink/nl/addr_linux.go index 50db3b4cd..6bea4ed02 100644 --- a/vendor/github.com/vishvananda/netlink/nl/addr_linux.go +++ b/vendor/github.com/vishvananda/netlink/nl/addr_linux.go @@ -54,24 +54,18 @@ func (msg *IfAddrmsg) Len() int { // __u32 tstamp; /* updated timestamp, hundredths of seconds */ // }; -const IFA_CACHEINFO = 6 -const SizeofIfaCacheInfo = 0x10 - type IfaCacheInfo struct { - IfaPrefered uint32 - IfaValid uint32 - Cstamp uint32 - Tstamp uint32 + unix.IfaCacheinfo } func (msg *IfaCacheInfo) Len() int { - return SizeofIfaCacheInfo + return unix.SizeofIfaCacheinfo } func DeserializeIfaCacheInfo(b []byte) *IfaCacheInfo { - return (*IfaCacheInfo)(unsafe.Pointer(&b[0:SizeofIfaCacheInfo][0])) + return (*IfaCacheInfo)(unsafe.Pointer(&b[0:unix.SizeofIfaCacheinfo][0])) } func (msg *IfaCacheInfo) Serialize() []byte { - return (*(*[SizeofIfaCacheInfo]byte)(unsafe.Pointer(msg)))[:] + return (*(*[unix.SizeofIfaCacheinfo]byte)(unsafe.Pointer(msg)))[:] } diff --git a/vendor/github.com/vishvananda/netlink/nl/conntrack_linux.go b/vendor/github.com/vishvananda/netlink/nl/conntrack_linux.go index 79d2b6b89..14924027e 100644 --- a/vendor/github.com/vishvananda/netlink/nl/conntrack_linux.go +++ b/vendor/github.com/vishvananda/netlink/nl/conntrack_linux.go @@ -40,9 +40,10 @@ const ( NFNETLINK_V0 = 0 ) -// #define NLA_F_NESTED (1 << 15) const ( - NLA_F_NESTED = (1 << 15) + NLA_F_NESTED uint16 = (1 << 15) // #define NLA_F_NESTED (1 << 15) + NLA_F_NET_BYTEORDER uint16 = (1 << 14) // #define NLA_F_NESTED (1 << 14) + NLA_TYPE_MASK = ^(NLA_F_NESTED | NLA_F_NET_BYTEORDER) ) // enum ctattr_type { diff --git a/vendor/github.com/vishvananda/netlink/nl/ipset_linux.go b/vendor/github.com/vishvananda/netlink/nl/ipset_linux.go new file mode 100644 index 000000000..a60b4b09d --- /dev/null +++ b/vendor/github.com/vishvananda/netlink/nl/ipset_linux.go @@ -0,0 +1,222 @@ +package nl + +import ( + "strconv" + + "golang.org/x/sys/unix" +) + +const ( + /* The protocol version */ + IPSET_PROTOCOL = 6 + + /* The max length of strings including NUL: set and type identifiers */ + IPSET_MAXNAMELEN = 32 + + /* The maximum permissible comment length we will accept over netlink */ + IPSET_MAX_COMMENT_SIZE = 255 +) + +const ( + _ = iota + IPSET_CMD_PROTOCOL /* 1: Return protocol version */ + IPSET_CMD_CREATE /* 2: Create a new (empty) set */ + IPSET_CMD_DESTROY /* 3: Destroy a (empty) set */ + IPSET_CMD_FLUSH /* 4: Remove all elements from a set */ + IPSET_CMD_RENAME /* 5: Rename a set */ + IPSET_CMD_SWAP /* 6: Swap two sets */ + IPSET_CMD_LIST /* 7: List sets */ + IPSET_CMD_SAVE /* 8: Save sets */ + IPSET_CMD_ADD /* 9: Add an element to a set */ + IPSET_CMD_DEL /* 10: Delete an element from a set */ + IPSET_CMD_TEST /* 11: Test an element in a set */ + IPSET_CMD_HEADER /* 12: Get set header data only */ + IPSET_CMD_TYPE /* 13: Get set type */ +) + +/* Attributes at command level */ +const ( + _ = iota + IPSET_ATTR_PROTOCOL /* 1: Protocol version */ + IPSET_ATTR_SETNAME /* 2: Name of the set */ + IPSET_ATTR_TYPENAME /* 3: Typename */ + IPSET_ATTR_REVISION /* 4: Settype revision */ + IPSET_ATTR_FAMILY /* 5: Settype family */ + IPSET_ATTR_FLAGS /* 6: Flags at command level */ + IPSET_ATTR_DATA /* 7: Nested attributes */ + IPSET_ATTR_ADT /* 8: Multiple data containers */ + IPSET_ATTR_LINENO /* 9: Restore lineno */ + IPSET_ATTR_PROTOCOL_MIN /* 10: Minimal supported version number */ + + IPSET_ATTR_SETNAME2 = IPSET_ATTR_TYPENAME /* Setname at rename/swap */ + IPSET_ATTR_REVISION_MIN = IPSET_ATTR_PROTOCOL_MIN /* type rev min */ +) + +/* CADT specific attributes */ +const ( + IPSET_ATTR_IP = 1 + IPSET_ATTR_IP_FROM = 1 + IPSET_ATTR_IP_TO = 2 + IPSET_ATTR_CIDR = 3 + IPSET_ATTR_PORT = 4 + IPSET_ATTR_PORT_FROM = 4 + IPSET_ATTR_PORT_TO = 5 + IPSET_ATTR_TIMEOUT = 6 + IPSET_ATTR_PROTO = 7 + IPSET_ATTR_CADT_FLAGS = 8 + IPSET_ATTR_CADT_LINENO = IPSET_ATTR_LINENO /* 9 */ + IPSET_ATTR_MARK = 10 + IPSET_ATTR_MARKMASK = 11 + + /* Reserve empty slots */ + IPSET_ATTR_CADT_MAX = 16 + + /* Create-only specific attributes */ + IPSET_ATTR_GC = 3 + iota + IPSET_ATTR_HASHSIZE + IPSET_ATTR_MAXELEM + IPSET_ATTR_NETMASK + IPSET_ATTR_PROBES + IPSET_ATTR_RESIZE + IPSET_ATTR_SIZE + + /* Kernel-only */ + IPSET_ATTR_ELEMENTS + IPSET_ATTR_REFERENCES + IPSET_ATTR_MEMSIZE + + SET_ATTR_CREATE_MAX +) + +/* ADT specific attributes */ +const ( + IPSET_ATTR_ETHER = IPSET_ATTR_CADT_MAX + iota + 1 + IPSET_ATTR_NAME + IPSET_ATTR_NAMEREF + IPSET_ATTR_IP2 + IPSET_ATTR_CIDR2 + IPSET_ATTR_IP2_TO + IPSET_ATTR_IFACE + IPSET_ATTR_BYTES + IPSET_ATTR_PACKETS + IPSET_ATTR_COMMENT + IPSET_ATTR_SKBMARK + IPSET_ATTR_SKBPRIO + IPSET_ATTR_SKBQUEUE +) + +/* Flags at CADT attribute level, upper half of cmdattrs */ +const ( + IPSET_FLAG_BIT_BEFORE = 0 + IPSET_FLAG_BEFORE = (1 << IPSET_FLAG_BIT_BEFORE) + IPSET_FLAG_BIT_PHYSDEV = 1 + IPSET_FLAG_PHYSDEV = (1 << IPSET_FLAG_BIT_PHYSDEV) + IPSET_FLAG_BIT_NOMATCH = 2 + IPSET_FLAG_NOMATCH = (1 << IPSET_FLAG_BIT_NOMATCH) + IPSET_FLAG_BIT_WITH_COUNTERS = 3 + IPSET_FLAG_WITH_COUNTERS = (1 << IPSET_FLAG_BIT_WITH_COUNTERS) + IPSET_FLAG_BIT_WITH_COMMENT = 4 + IPSET_FLAG_WITH_COMMENT = (1 << IPSET_FLAG_BIT_WITH_COMMENT) + IPSET_FLAG_BIT_WITH_FORCEADD = 5 + IPSET_FLAG_WITH_FORCEADD = (1 << IPSET_FLAG_BIT_WITH_FORCEADD) + IPSET_FLAG_BIT_WITH_SKBINFO = 6 + IPSET_FLAG_WITH_SKBINFO = (1 << IPSET_FLAG_BIT_WITH_SKBINFO) + IPSET_FLAG_CADT_MAX = 15 +) + +const ( + IPSET_ERR_PRIVATE = 4096 + iota + IPSET_ERR_PROTOCOL + IPSET_ERR_FIND_TYPE + IPSET_ERR_MAX_SETS + IPSET_ERR_BUSY + IPSET_ERR_EXIST_SETNAME2 + IPSET_ERR_TYPE_MISMATCH + IPSET_ERR_EXIST + IPSET_ERR_INVALID_CIDR + IPSET_ERR_INVALID_NETMASK + IPSET_ERR_INVALID_FAMILY + IPSET_ERR_TIMEOUT + IPSET_ERR_REFERENCED + IPSET_ERR_IPADDR_IPV4 + IPSET_ERR_IPADDR_IPV6 + IPSET_ERR_COUNTER + IPSET_ERR_COMMENT + IPSET_ERR_INVALID_MARKMASK + IPSET_ERR_SKBINFO + + /* Type specific error codes */ + IPSET_ERR_TYPE_SPECIFIC = 4352 +) + +type IPSetError uintptr + +func (e IPSetError) Error() string { + switch int(e) { + case IPSET_ERR_PRIVATE: + return "private" + case IPSET_ERR_PROTOCOL: + return "invalid protocol" + case IPSET_ERR_FIND_TYPE: + return "invalid type" + case IPSET_ERR_MAX_SETS: + return "max sets reached" + case IPSET_ERR_BUSY: + return "busy" + case IPSET_ERR_EXIST_SETNAME2: + return "exist_setname2" + case IPSET_ERR_TYPE_MISMATCH: + return "type mismatch" + case IPSET_ERR_EXIST: + return "exist" + case IPSET_ERR_INVALID_CIDR: + return "invalid cidr" + case IPSET_ERR_INVALID_NETMASK: + return "invalid netmask" + case IPSET_ERR_INVALID_FAMILY: + return "invalid family" + case IPSET_ERR_TIMEOUT: + return "timeout" + case IPSET_ERR_REFERENCED: + return "referenced" + case IPSET_ERR_IPADDR_IPV4: + return "invalid ipv4 address" + case IPSET_ERR_IPADDR_IPV6: + return "invalid ipv6 address" + case IPSET_ERR_COUNTER: + return "invalid counter" + case IPSET_ERR_COMMENT: + return "invalid comment" + case IPSET_ERR_INVALID_MARKMASK: + return "invalid markmask" + case IPSET_ERR_SKBINFO: + return "skbinfo" + default: + return "errno " + strconv.Itoa(int(e)) + } +} + +func GetIpsetFlags(cmd int) int { + switch cmd { + case IPSET_CMD_CREATE: + return unix.NLM_F_REQUEST | unix.NLM_F_ACK | unix.NLM_F_CREATE + case IPSET_CMD_DESTROY, + IPSET_CMD_FLUSH, + IPSET_CMD_RENAME, + IPSET_CMD_SWAP, + IPSET_CMD_TEST: + return unix.NLM_F_REQUEST | unix.NLM_F_ACK + case IPSET_CMD_LIST, + IPSET_CMD_SAVE: + return unix.NLM_F_REQUEST | unix.NLM_F_ACK | unix.NLM_F_ROOT | unix.NLM_F_MATCH | unix.NLM_F_DUMP + case IPSET_CMD_ADD, + IPSET_CMD_DEL: + return unix.NLM_F_REQUEST | unix.NLM_F_ACK + case IPSET_CMD_HEADER, + IPSET_CMD_TYPE, + IPSET_CMD_PROTOCOL: + return unix.NLM_F_REQUEST + default: + return 0 + } +} diff --git a/vendor/github.com/vishvananda/netlink/nl/link_linux.go b/vendor/github.com/vishvananda/netlink/nl/link_linux.go index afb16a9c1..faee2fa03 100644 --- a/vendor/github.com/vishvananda/netlink/nl/link_linux.go +++ b/vendor/github.com/vishvananda/netlink/nl/link_linux.go @@ -1,6 +1,8 @@ package nl import ( + "bytes" + "encoding/binary" "unsafe" ) @@ -243,7 +245,9 @@ const ( IFLA_VF_STATS_TX_BYTES IFLA_VF_STATS_BROADCAST IFLA_VF_STATS_MULTICAST - IFLA_VF_STATS_MAX = IFLA_VF_STATS_MULTICAST + IFLA_VF_STATS_RX_DROPPED + IFLA_VF_STATS_TX_DROPPED + IFLA_VF_STATS_MAX = IFLA_VF_STATS_TX_DROPPED ) const ( @@ -326,6 +330,59 @@ func (msg *VfTxRate) Serialize() []byte { return (*(*[SizeofVfTxRate]byte)(unsafe.Pointer(msg)))[:] } +//struct ifla_vf_stats { +// __u64 rx_packets; +// __u64 tx_packets; +// __u64 rx_bytes; +// __u64 tx_bytes; +// __u64 broadcast; +// __u64 multicast; +//}; + +type VfStats struct { + RxPackets uint64 + TxPackets uint64 + RxBytes uint64 + TxBytes uint64 + Multicast uint64 + Broadcast uint64 + RxDropped uint64 + TxDropped uint64 +} + +func DeserializeVfStats(b []byte) VfStats { + var vfstat VfStats + stats, err := ParseRouteAttr(b) + if err != nil { + return vfstat + } + var valueVar uint64 + for _, stat := range stats { + if err := binary.Read(bytes.NewBuffer(stat.Value), NativeEndian(), &valueVar); err != nil { + break + } + switch stat.Attr.Type { + case IFLA_VF_STATS_RX_PACKETS: + vfstat.RxPackets = valueVar + case IFLA_VF_STATS_TX_PACKETS: + vfstat.TxPackets = valueVar + case IFLA_VF_STATS_RX_BYTES: + vfstat.RxBytes = valueVar + case IFLA_VF_STATS_TX_BYTES: + vfstat.TxBytes = valueVar + case IFLA_VF_STATS_MULTICAST: + vfstat.Multicast = valueVar + case IFLA_VF_STATS_BROADCAST: + vfstat.Broadcast = valueVar + case IFLA_VF_STATS_RX_DROPPED: + vfstat.RxDropped = valueVar + case IFLA_VF_STATS_TX_DROPPED: + vfstat.TxDropped = valueVar + } + } + return vfstat +} + // struct ifla_vf_rate { // __u32 vf; // __u32 min_tx_rate; /* Min Bandwidth in Mbps */ @@ -478,6 +535,14 @@ const ( IFLA_XDP_MAX = IFLA_XDP_PROG_ID ) +// XDP program attach mode (used as dump value for IFLA_XDP_ATTACHED) +const ( + XDP_ATTACHED_NONE = iota + XDP_ATTACHED_DRV + XDP_ATTACHED_SKB + XDP_ATTACHED_HW +) + const ( IFLA_IPTUN_UNSPEC = iota IFLA_IPTUN_LINK diff --git a/vendor/github.com/vishvananda/netlink/nl/nl_linux.go b/vendor/github.com/vishvananda/netlink/nl/nl_linux.go index aaf56c671..cef64b82e 100644 --- a/vendor/github.com/vishvananda/netlink/nl/nl_linux.go +++ b/vendor/github.com/vishvananda/netlink/nl/nl_linux.go @@ -259,6 +259,29 @@ func NewIfInfomsgChild(parent *RtAttr, family int) *IfInfomsg { return msg } +type Uint32Attribute struct { + Type uint16 + Value uint32 +} + +func (a *Uint32Attribute) Serialize() []byte { + native := NativeEndian() + buf := make([]byte, rtaAlignOf(8)) + native.PutUint16(buf[0:2], 8) + native.PutUint16(buf[2:4], a.Type) + + if a.Type&NLA_F_NET_BYTEORDER != 0 { + binary.BigEndian.PutUint32(buf[4:], a.Value) + } else { + native.PutUint32(buf[4:], a.Value) + } + return buf +} + +func (a *Uint32Attribute) Len() int { + return 8 +} + // Extend RtAttr to handle data and children type RtAttr struct { unix.RtAttr @@ -439,10 +462,7 @@ done: if m.Header.Pid != pid { continue } - if m.Header.Type == unix.NLMSG_DONE { - break done - } - if m.Header.Type == unix.NLMSG_ERROR { + if m.Header.Type == unix.NLMSG_DONE || m.Header.Type == unix.NLMSG_ERROR { native := NativeEndian() error := int32(native.Uint32(m.Data[0:4])) if error == 0 { diff --git a/vendor/github.com/vishvananda/netlink/nl/parse_attr.go b/vendor/github.com/vishvananda/netlink/nl/parse_attr.go new file mode 100644 index 000000000..19eb8f28e --- /dev/null +++ b/vendor/github.com/vishvananda/netlink/nl/parse_attr.go @@ -0,0 +1,67 @@ +package nl + +import ( + "encoding/binary" + "fmt" +) + +type Attribute struct { + Type uint16 + Value []byte +} + +func ParseAttributes(data []byte) <-chan Attribute { + native := NativeEndian() + result := make(chan Attribute) + + go func() { + i := 0 + for i+4 < len(data) { + length := int(native.Uint16(data[i : i+2])) + + result <- Attribute{ + Type: native.Uint16(data[i+2 : i+4]), + Value: data[i+4 : i+length], + } + i += rtaAlignOf(length) + } + close(result) + }() + + return result +} + +func PrintAttributes(data []byte) { + printAttributes(data, 0) +} + +func printAttributes(data []byte, level int) { + for attr := range ParseAttributes(data) { + for i := 0; i < level; i++ { + print("> ") + } + nested := attr.Type&NLA_F_NESTED != 0 + fmt.Printf("type=%d nested=%v len=%v %v\n", attr.Type&NLA_TYPE_MASK, nested, len(attr.Value), attr.Value) + if nested { + printAttributes(attr.Value, level+1) + } + } +} + +// Uint32 returns the uint32 value respecting the NET_BYTEORDER flag +func (attr *Attribute) Uint32() uint32 { + if attr.Type&NLA_F_NET_BYTEORDER != 0 { + return binary.BigEndian.Uint32(attr.Value) + } else { + return NativeEndian().Uint32(attr.Value) + } +} + +// Uint64 returns the uint64 value respecting the NET_BYTEORDER flag +func (attr *Attribute) Uint64() uint64 { + if attr.Type&NLA_F_NET_BYTEORDER != 0 { + return binary.BigEndian.Uint64(attr.Value) + } else { + return NativeEndian().Uint64(attr.Value) + } +} diff --git a/vendor/github.com/vishvananda/netlink/nl/syscall.go b/vendor/github.com/vishvananda/netlink/nl/syscall.go index f7f7f92e6..4a01e6e59 100644 --- a/vendor/github.com/vishvananda/netlink/nl/syscall.go +++ b/vendor/github.com/vishvananda/netlink/nl/syscall.go @@ -21,6 +21,13 @@ const ( FRA_TABLE /* Extended table id */ FRA_FWMASK /* mask for netfilter mark */ FRA_OIFNAME + FRA_PAD + FRA_L3MDEV /* iif or oif is l3mdev goto its table */ + FRA_UID_RANGE /* UID range */ + FRA_PROTOCOL /* Originator of the rule */ + FRA_IP_PROTO /* ip proto */ + FRA_SPORT_RANGE /* sport */ + FRA_DPORT_RANGE /* dport */ ) // ip rule netlink request types diff --git a/vendor/github.com/vishvananda/netlink/nl/tc_linux.go b/vendor/github.com/vishvananda/netlink/nl/tc_linux.go index 501f554b2..c24d53eb7 100644 --- a/vendor/github.com/vishvananda/netlink/nl/tc_linux.go +++ b/vendor/github.com/vishvananda/netlink/nl/tc_linux.go @@ -94,6 +94,9 @@ const ( SizeofTcTunnelKey = SizeofTcGen + 0x04 SizeofTcSkbEdit = SizeofTcGen SizeofTcPolice = 2*SizeofTcRateSpec + 0x20 + SizeofTcSfqQopt = 0x0b + SizeofTcSfqRedStats = 0x18 + SizeofTcSfqQoptV1 = SizeofTcSfqQopt + SizeofTcSfqRedStats + 0x1c ) // struct tcmsg { @@ -735,7 +738,13 @@ const ( TCA_TUNNEL_KEY_ENC_IPV6_SRC TCA_TUNNEL_KEY_ENC_IPV6_DST TCA_TUNNEL_KEY_ENC_KEY_ID - TCA_TUNNEL_KEY_MAX = TCA_TUNNEL_KEY_ENC_KEY_ID + TCA_TUNNEL_KEY_PAD + TCA_TUNNEL_KEY_ENC_DST_PORT + TCA_TUNNEL_KEY_NO_CSUM + TCA_TUNNEL_KEY_ENC_OPTS + TCA_TUNNEL_KEY_ENC_TOS + TCA_TUNNEL_KEY_ENC_TTL + TCA_TUNNEL_KEY_MAX ) type TcTunnelKey struct { @@ -872,3 +881,103 @@ const ( TCA_HFSC_FSC TCA_HFSC_USC ) + +// struct tc_sfq_qopt { +// unsigned quantum; /* Bytes per round allocated to flow */ +// int perturb_period; /* Period of hash perturbation */ +// __u32 limit; /* Maximal packets in queue */ +// unsigned divisor; /* Hash divisor */ +// unsigned flows; /* Maximal number of flows */ +// }; + +type TcSfqQopt struct { + Quantum uint8 + Perturb int32 + Limit uint32 + Divisor uint8 + Flows uint8 +} + +func (x *TcSfqQopt) Len() int { + return SizeofTcSfqQopt +} + +func DeserializeTcSfqQopt(b []byte) *TcSfqQopt { + return (*TcSfqQopt)(unsafe.Pointer(&b[0:SizeofTcSfqQopt][0])) +} + +func (x *TcSfqQopt) Serialize() []byte { + return (*(*[SizeofTcSfqQopt]byte)(unsafe.Pointer(x)))[:] +} + +// struct tc_sfqred_stats { +// __u32 prob_drop; /* Early drops, below max threshold */ +// __u32 forced_drop; /* Early drops, after max threshold */ +// __u32 prob_mark; /* Marked packets, below max threshold */ +// __u32 forced_mark; /* Marked packets, after max threshold */ +// __u32 prob_mark_head; /* Marked packets, below max threshold */ +// __u32 forced_mark_head;/* Marked packets, after max threshold */ +// }; +type TcSfqRedStats struct { + ProbDrop uint32 + ForcedDrop uint32 + ProbMark uint32 + ForcedMark uint32 + ProbMarkHead uint32 + ForcedMarkHead uint32 +} + +func (x *TcSfqRedStats) Len() int { + return SizeofTcSfqRedStats +} + +func DeserializeTcSfqRedStats(b []byte) *TcSfqRedStats { + return (*TcSfqRedStats)(unsafe.Pointer(&b[0:SizeofTcSfqRedStats][0])) +} + +func (x *TcSfqRedStats) Serialize() []byte { + return (*(*[SizeofTcSfqRedStats]byte)(unsafe.Pointer(x)))[:] +} + +// struct tc_sfq_qopt_v1 { +// struct tc_sfq_qopt v0; +// unsigned int depth; /* max number of packets per flow */ +// unsigned int headdrop; +// /* SFQRED parameters */ +// __u32 limit; /* HARD maximal flow queue length (bytes) */ +// __u32 qth_min; /* Min average length threshold (bytes) */ +// __u32 qth_max; /* Max average length threshold (bytes) */ +// unsigned char Wlog; /* log(W) */ +// unsigned char Plog; /* log(P_max/(qth_max-qth_min)) */ +// unsigned char Scell_log; /* cell size for idle damping */ +// unsigned char flags; +// __u32 max_P; /* probability, high resolution */ +// /* SFQRED stats */ +// struct tc_sfqred_stats stats; +// }; +type TcSfqQoptV1 struct { + TcSfqQopt + Depth uint32 + HeadDrop uint32 + Limit uint32 + QthMin uint32 + QthMax uint32 + Wlog byte + Plog byte + ScellLog byte + Flags byte + MaxP uint32 + TcSfqRedStats +} + +func (x *TcSfqQoptV1) Len() int { + return SizeofTcSfqQoptV1 +} + +func DeserializeTcSfqQoptV1(b []byte) *TcSfqQoptV1 { + return (*TcSfqQoptV1)(unsafe.Pointer(&b[0:SizeofTcSfqQoptV1][0])) +} + +func (x *TcSfqQoptV1) Serialize() []byte { + return (*(*[SizeofTcSfqQoptV1]byte)(unsafe.Pointer(x)))[:] +} diff --git a/vendor/github.com/vishvananda/netlink/nl/xfrm_state_linux.go b/vendor/github.com/vishvananda/netlink/nl/xfrm_state_linux.go index b6290fd54..43a947f22 100644 --- a/vendor/github.com/vishvananda/netlink/nl/xfrm_state_linux.go +++ b/vendor/github.com/vishvananda/netlink/nl/xfrm_state_linux.go @@ -13,7 +13,7 @@ const ( SizeofXfrmAlgoAuth = 0x48 SizeofXfrmAlgoAEAD = 0x48 SizeofXfrmEncapTmpl = 0x18 - SizeofXfrmUsersaFlush = 0x8 + SizeofXfrmUsersaFlush = 0x1 SizeofXfrmReplayStateEsn = 0x18 ) diff --git a/vendor/github.com/vishvananda/netlink/qdisc.go b/vendor/github.com/vishvananda/netlink/qdisc.go index af78305ac..8418569ee 100644 --- a/vendor/github.com/vishvananda/netlink/qdisc.go +++ b/vendor/github.com/vishvananda/netlink/qdisc.go @@ -338,3 +338,27 @@ func (qdisc *FqCodel) Attrs() *QdiscAttrs { func (qdisc *FqCodel) Type() string { return "fq_codel" } + +type Sfq struct { + QdiscAttrs + // TODO: Only the simplified options for SFQ are handled here. Support for the extended one can be added later. + Quantum uint8 + Perturb uint8 + Limit uint32 + Divisor uint8 +} + +func (sfq *Sfq) String() string { + return fmt.Sprintf( + "{%v -- Quantum: %v, Perturb: %v, Limit: %v, Divisor: %v}", + sfq.Attrs(), sfq.Quantum, sfq.Perturb, sfq.Limit, sfq.Divisor, + ) +} + +func (qdisc *Sfq) Attrs() *QdiscAttrs { + return &qdisc.QdiscAttrs +} + +func (qdisc *Sfq) Type() string { + return "sfq" +} diff --git a/vendor/github.com/vishvananda/netlink/qdisc_linux.go b/vendor/github.com/vishvananda/netlink/qdisc_linux.go index e9eee5908..d0e1ca194 100644 --- a/vendor/github.com/vishvananda/netlink/qdisc_linux.go +++ b/vendor/github.com/vishvananda/netlink/qdisc_linux.go @@ -278,6 +278,14 @@ func qdiscPayload(req *nl.NetlinkRequest, qdisc Qdisc) error { if qdisc.FlowDefaultRate > 0 { options.AddRtAttr(nl.TCA_FQ_FLOW_DEFAULT_RATE, nl.Uint32Attr((uint32(qdisc.FlowDefaultRate)))) } + case *Sfq: + opt := nl.TcSfqQoptV1{} + opt.TcSfqQopt.Quantum = qdisc.Quantum + opt.TcSfqQopt.Perturb = int32(qdisc.Perturb) + opt.TcSfqQopt.Limit = qdisc.Limit + opt.TcSfqQopt.Divisor = qdisc.Divisor + + options = nl.NewRtAttr(nl.TCA_OPTIONS, opt.Serialize()) default: options = nil } @@ -362,6 +370,8 @@ func (h *Handle) QdiscList(link Link) ([]Qdisc, error) { qdisc = &FqCodel{} case "netem": qdisc = &Netem{} + case "sfq": + qdisc = &Sfq{} default: qdisc = &GenericQdisc{QdiscType: qdiscType} } @@ -417,6 +427,10 @@ func (h *Handle) QdiscList(link Link) ([]Qdisc, error) { if err := parseNetemData(qdisc, attr.Value); err != nil { return nil, err } + case "sfq": + if err := parseSfqData(qdisc, attr.Value); err != nil { + return nil, err + } // no options for ingress } @@ -582,6 +596,17 @@ func parseTbfData(qdisc Qdisc, data []syscall.NetlinkRouteAttr) error { return nil } +func parseSfqData(qdisc Qdisc, value []byte) error { + sfq := qdisc.(*Sfq) + opt := nl.DeserializeTcSfqQoptV1(value) + sfq.Quantum = opt.TcSfqQopt.Quantum + sfq.Perturb = uint8(opt.TcSfqQopt.Perturb) + sfq.Limit = opt.TcSfqQopt.Limit + sfq.Divisor = opt.TcSfqQopt.Divisor + + return nil +} + const ( TIME_UNITS_PER_SEC = 1000000 ) @@ -598,10 +623,10 @@ func initClock() { return } parts := strings.Split(strings.TrimSpace(string(data)), " ") - if len(parts) < 3 { + if len(parts) < 4 { return } - var vals [3]uint64 + var vals [4]uint64 for i := range vals { val, err := strconv.ParseUint(parts[i], 16, 32) if err != nil { @@ -615,7 +640,12 @@ func initClock() { } clockFactor = float64(vals[2]) / TIME_UNITS_PER_SEC tickInUsec = float64(vals[0]) / float64(vals[1]) * clockFactor - hz = float64(vals[0]) + if vals[2] == 1000000 { + // ref https://git.kernel.org/pub/scm/network/iproute2/iproute2.git/tree/lib/utils.c#n963 + hz = float64(vals[3]) + } else { + hz = 100 + } } func TickInUsec() float64 { @@ -663,6 +693,7 @@ func latency(rate uint64, limit, buffer uint32) float64 { return TIME_UNITS_PER_SEC*(float64(limit)/float64(rate)) - float64(tick2Time(buffer)) } -func Xmittime(rate uint64, size uint32) float64 { - return TickInUsec() * TIME_UNITS_PER_SEC * (float64(size) / float64(rate)) +func Xmittime(rate uint64, size uint32) uint32 { + // https://git.kernel.org/pub/scm/network/iproute2/iproute2.git/tree/tc/tc_core.c#n62 + return time2Tick(uint32(TIME_UNITS_PER_SEC * (float64(size) / float64(rate)))) } diff --git a/vendor/github.com/vishvananda/netlink/rdma_link_linux.go b/vendor/github.com/vishvananda/netlink/rdma_link_linux.go index 2d0bdc8c3..ff014ca4c 100644 --- a/vendor/github.com/vishvananda/netlink/rdma_link_linux.go +++ b/vendor/github.com/vishvananda/netlink/rdma_link_linux.go @@ -77,28 +77,39 @@ func executeOneGetRdmaLink(data []byte) (*RdmaLink, error) { return &link, nil } -func execRdmaGetLink(req *nl.NetlinkRequest, name string) (*RdmaLink, error) { +func execRdmaSetLink(req *nl.NetlinkRequest) error { + + _, err := req.Execute(unix.NETLINK_RDMA, 0) + return err +} + +// RdmaLinkList gets a list of RDMA link devices. +// Equivalent to: `rdma dev show` +func RdmaLinkList() ([]*RdmaLink, error) { + return pkgHandle.RdmaLinkList() +} + +// RdmaLinkList gets a list of RDMA link devices. +// Equivalent to: `rdma dev show` +func (h *Handle) RdmaLinkList() ([]*RdmaLink, error) { + proto := getProtoField(nl.RDMA_NL_NLDEV, nl.RDMA_NLDEV_CMD_GET) + req := h.newNetlinkRequest(proto, unix.NLM_F_ACK|unix.NLM_F_DUMP) msgs, err := req.Execute(unix.NETLINK_RDMA, 0) if err != nil { return nil, err } + + var res []*RdmaLink for _, m := range msgs { link, err := executeOneGetRdmaLink(m) if err != nil { return nil, err } - if link.Attrs.Name == name { - return link, nil - } + res = append(res, link) } - return nil, fmt.Errorf("Rdma device %v not found", name) -} -func execRdmaSetLink(req *nl.NetlinkRequest) error { - - _, err := req.Execute(unix.NETLINK_RDMA, 0) - return err + return res, nil } // RdmaLinkByName finds a link by name and returns a pointer to the object if @@ -110,11 +121,16 @@ func RdmaLinkByName(name string) (*RdmaLink, error) { // RdmaLinkByName finds a link by name and returns a pointer to the object if // found and nil error, otherwise returns error code. func (h *Handle) RdmaLinkByName(name string) (*RdmaLink, error) { - - proto := getProtoField(nl.RDMA_NL_NLDEV, nl.RDMA_NLDEV_CMD_GET) - req := h.newNetlinkRequest(proto, unix.NLM_F_ACK|unix.NLM_F_DUMP) - - return execRdmaGetLink(req, name) + links, err := h.RdmaLinkList() + if err != nil { + return nil, err + } + for _, link := range links { + if link.Attrs.Name == name { + return link, nil + } + } + return nil, fmt.Errorf("Rdma device %v not found", name) } // RdmaLinkSetName sets the name of the rdma link device. Return nil on success diff --git a/vendor/github.com/vishvananda/netlink/route.go b/vendor/github.com/vishvananda/netlink/route.go index 58ff1af60..b16254174 100644 --- a/vendor/github.com/vishvananda/netlink/route.go +++ b/vendor/github.com/vishvananda/netlink/route.go @@ -29,25 +29,38 @@ type Encap interface { // Route represents a netlink route. type Route struct { - LinkIndex int - ILinkIndex int - Scope Scope - Dst *net.IPNet - Src net.IP - Gw net.IP - MultiPath []*NexthopInfo - Protocol int - Priority int - Table int - Type int - Tos int - Flags int - MPLSDst *int - NewDst Destination - Encap Encap - MTU int - AdvMSS int - Hoplimit int + LinkIndex int + ILinkIndex int + Scope Scope + Dst *net.IPNet + Src net.IP + Gw net.IP + MultiPath []*NexthopInfo + Protocol int + Priority int + Table int + Type int + Tos int + Flags int + MPLSDst *int + NewDst Destination + Encap Encap + MTU int + Window int + Rtt int + RttVar int + Ssthresh int + Cwnd int + AdvMSS int + Reordering int + Hoplimit int + InitCwnd int + Features int + RtoMin int + InitRwnd int + QuickACK int + Congctl string + FastOpenNoCookie int } func (r Route) String() string { diff --git a/vendor/github.com/vishvananda/netlink/route_linux.go b/vendor/github.com/vishvananda/netlink/route_linux.go index c69c595ed..4e778a417 100644 --- a/vendor/github.com/vishvananda/netlink/route_linux.go +++ b/vendor/github.com/vishvananda/netlink/route_linux.go @@ -33,6 +33,9 @@ const ( RT_FILTER_GW RT_FILTER_TABLE RT_FILTER_HOPLIMIT + RT_FILTER_PRIORITY + RT_FILTER_MARK + RT_FILTER_MASK ) const ( @@ -639,19 +642,70 @@ func (h *Handle) routeHandle(route *Route, req *nl.NetlinkRequest, msg *nl.RtMsg } var metrics []*nl.RtAttr - // TODO: support other rta_metric values if route.MTU > 0 { b := nl.Uint32Attr(uint32(route.MTU)) metrics = append(metrics, nl.NewRtAttr(unix.RTAX_MTU, b)) } + if route.Window > 0 { + b := nl.Uint32Attr(uint32(route.Window)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_WINDOW, b)) + } + if route.Rtt > 0 { + b := nl.Uint32Attr(uint32(route.Rtt)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_RTT, b)) + } + if route.RttVar > 0 { + b := nl.Uint32Attr(uint32(route.RttVar)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_RTTVAR, b)) + } + if route.Ssthresh > 0 { + b := nl.Uint32Attr(uint32(route.Ssthresh)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_SSTHRESH, b)) + } + if route.Cwnd > 0 { + b := nl.Uint32Attr(uint32(route.Cwnd)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_CWND, b)) + } if route.AdvMSS > 0 { b := nl.Uint32Attr(uint32(route.AdvMSS)) metrics = append(metrics, nl.NewRtAttr(unix.RTAX_ADVMSS, b)) } + if route.Reordering > 0 { + b := nl.Uint32Attr(uint32(route.Reordering)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_REORDERING, b)) + } if route.Hoplimit > 0 { b := nl.Uint32Attr(uint32(route.Hoplimit)) metrics = append(metrics, nl.NewRtAttr(unix.RTAX_HOPLIMIT, b)) } + if route.InitCwnd > 0 { + b := nl.Uint32Attr(uint32(route.InitCwnd)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_INITCWND, b)) + } + if route.Features > 0 { + b := nl.Uint32Attr(uint32(route.Features)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_FEATURES, b)) + } + if route.RtoMin > 0 { + b := nl.Uint32Attr(uint32(route.RtoMin)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_RTO_MIN, b)) + } + if route.InitRwnd > 0 { + b := nl.Uint32Attr(uint32(route.InitRwnd)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_INITRWND, b)) + } + if route.QuickACK > 0 { + b := nl.Uint32Attr(uint32(route.QuickACK)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_QUICKACK, b)) + } + if route.Congctl != "" { + b := nl.ZeroTerminated(route.Congctl) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_CC_ALGO, b)) + } + if route.FastOpenNoCookie > 0 { + b := nl.Uint32Attr(uint32(route.FastOpenNoCookie)) + metrics = append(metrics, nl.NewRtAttr(unix.RTAX_FASTOPEN_NO_COOKIE, b)) + } if metrics != nil { attr := nl.NewRtAttr(unix.RTA_METRICS, nil) @@ -903,10 +957,36 @@ func deserializeRoute(m []byte) (Route, error) { switch metric.Attr.Type { case unix.RTAX_MTU: route.MTU = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_WINDOW: + route.Window = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_RTT: + route.Rtt = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_RTTVAR: + route.RttVar = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_SSTHRESH: + route.Ssthresh = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_CWND: + route.Cwnd = int(native.Uint32(metric.Value[0:4])) case unix.RTAX_ADVMSS: route.AdvMSS = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_REORDERING: + route.Reordering = int(native.Uint32(metric.Value[0:4])) case unix.RTAX_HOPLIMIT: route.Hoplimit = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_INITCWND: + route.InitCwnd = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_FEATURES: + route.Features = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_RTO_MIN: + route.RtoMin = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_INITRWND: + route.InitRwnd = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_QUICKACK: + route.QuickACK = int(native.Uint32(metric.Value[0:4])) + case unix.RTAX_CC_ALGO: + route.Congctl = nl.BytesToString(metric.Value) + case unix.RTAX_FASTOPEN_NO_COOKIE: + route.FastOpenNoCookie = int(native.Uint32(metric.Value[0:4])) } } } @@ -938,15 +1018,27 @@ func deserializeRoute(m []byte) (Route, error) { return route, nil } +// RouteGetOptions contains a set of options to use with +// RouteGetWithOptions +type RouteGetOptions struct { + VrfName string +} + +// RouteGetWithOptions gets a route to a specific destination from the host system. +// Equivalent to: 'ip route get <> vrf <VrfName>'. +func RouteGetWithOptions(destination net.IP, options *RouteGetOptions) ([]Route, error) { + return pkgHandle.RouteGetWithOptions(destination, options) +} + // RouteGet gets a route to a specific destination from the host system. // Equivalent to: 'ip route get'. func RouteGet(destination net.IP) ([]Route, error) { return pkgHandle.RouteGet(destination) } -// RouteGet gets a route to a specific destination from the host system. -// Equivalent to: 'ip route get'. -func (h *Handle) RouteGet(destination net.IP) ([]Route, error) { +// RouteGetWithOptions gets a route to a specific destination from the host system. +// Equivalent to: 'ip route get <> vrf <VrfName>'. +func (h *Handle) RouteGetWithOptions(destination net.IP, options *RouteGetOptions) ([]Route, error) { req := h.newNetlinkRequest(unix.RTM_GETROUTE, unix.NLM_F_REQUEST) family := nl.GetIPFamily(destination) var destinationData []byte @@ -966,6 +1058,20 @@ func (h *Handle) RouteGet(destination net.IP) ([]Route, error) { rtaDst := nl.NewRtAttr(unix.RTA_DST, destinationData) req.AddData(rtaDst) + if options != nil { + link, err := LinkByName(options.VrfName) + if err != nil { + return nil, err + } + var ( + b = make([]byte, 4) + native = nl.NativeEndian() + ) + native.PutUint32(b, uint32(link.Attrs().Index)) + + req.AddData(nl.NewRtAttr(unix.RTA_OIF, b)) + } + msgs, err := req.Execute(unix.NETLINK_ROUTE, unix.RTM_NEWROUTE) if err != nil { return nil, err @@ -980,7 +1086,12 @@ func (h *Handle) RouteGet(destination net.IP) ([]Route, error) { res = append(res, route) } return res, nil +} +// RouteGet gets a route to a specific destination from the host system. +// Equivalent to: 'ip route get'. +func (h *Handle) RouteGet(destination net.IP) ([]Route, error) { + return h.RouteGetWithOptions(destination, nil) } // RouteSubscribe takes a chan down which notifications will be sent diff --git a/vendor/github.com/vishvananda/netlink/rule.go b/vendor/github.com/vishvananda/netlink/rule.go index 7fc8ae5df..95f2facfb 100644 --- a/vendor/github.com/vishvananda/netlink/rule.go +++ b/vendor/github.com/vishvananda/netlink/rule.go @@ -12,6 +12,7 @@ type Rule struct { Table int Mark int Mask int + Tos uint TunID uint Goto int Src *net.IPNet @@ -22,6 +23,8 @@ type Rule struct { SuppressIfgroup int SuppressPrefixlen int Invert bool + Dport *RulePortRange + Sport *RulePortRange } func (r Rule) String() string { @@ -40,3 +43,14 @@ func NewRule() *Rule { Flow: -1, } } + +// NewRulePortRange creates rule sport/dport range. +func NewRulePortRange(start, end uint16) *RulePortRange { + return &RulePortRange{Start: start, End: end} +} + +// RulePortRange represents rule sport/dport range. +type RulePortRange struct { + Start uint16 + End uint16 +} diff --git a/vendor/github.com/vishvananda/netlink/rule_linux.go b/vendor/github.com/vishvananda/netlink/rule_linux.go index e12569fe4..40474f30e 100644 --- a/vendor/github.com/vishvananda/netlink/rule_linux.go +++ b/vendor/github.com/vishvananda/netlink/rule_linux.go @@ -1,6 +1,7 @@ package netlink import ( + "bytes" "fmt" "net" @@ -55,6 +56,9 @@ func ruleHandle(rule *Rule, req *nl.NetlinkRequest) error { if rule.Table >= 0 && rule.Table < 256 { msg.Table = uint8(rule.Table) } + if rule.Tos != 0 { + msg.Tos = uint8(rule.Tos) + } var dstFamily uint8 var rtAttrs []*nl.RtAttr @@ -138,10 +142,10 @@ func ruleHandle(rule *Rule, req *nl.NetlinkRequest) error { } } if rule.IifName != "" { - req.AddData(nl.NewRtAttr(nl.FRA_IIFNAME, []byte(rule.IifName))) + req.AddData(nl.NewRtAttr(nl.FRA_IIFNAME, []byte(rule.IifName+"\x00"))) } if rule.OifName != "" { - req.AddData(nl.NewRtAttr(nl.FRA_OIFNAME, []byte(rule.OifName))) + req.AddData(nl.NewRtAttr(nl.FRA_OIFNAME, []byte(rule.OifName+"\x00"))) } if rule.Goto >= 0 { msg.Type = nl.FR_ACT_GOTO @@ -150,6 +154,16 @@ func ruleHandle(rule *Rule, req *nl.NetlinkRequest) error { req.AddData(nl.NewRtAttr(nl.FRA_GOTO, b)) } + if rule.Dport != nil { + b := rule.Dport.toRtAttrData() + req.AddData(nl.NewRtAttr(nl.FRA_DPORT_RANGE, b)) + } + + if rule.Sport != nil { + b := rule.Sport.toRtAttrData() + req.AddData(nl.NewRtAttr(nl.FRA_SPORT_RANGE, b)) + } + _, err := req.Execute(unix.NETLINK_ROUTE, 0) return err } @@ -163,6 +177,19 @@ func RuleList(family int) ([]Rule, error) { // RuleList lists rules in the system. // Equivalent to: ip rule list func (h *Handle) RuleList(family int) ([]Rule, error) { + return h.RuleListFiltered(family, nil, 0) +} + +// RuleListFiltered gets a list of rules in the system filtered by the +// specified rule template `filter`. +// Equivalent to: ip rule list +func RuleListFiltered(family int, filter *Rule, filterMask uint64) ([]Rule, error) { + return pkgHandle.RuleListFiltered(family, filter, filterMask) +} + +// RuleListFiltered lists rules in the system. +// Equivalent to: ip rule list +func (h *Handle) RuleListFiltered(family int, filter *Rule, filterMask uint64) ([]Rule, error) { req := h.newNetlinkRequest(unix.RTM_GETRULE, unix.NLM_F_DUMP|unix.NLM_F_REQUEST) msg := nl.NewIfInfomsg(family) req.AddData(msg) @@ -184,6 +211,7 @@ func (h *Handle) RuleList(family int) ([]Rule, error) { rule := NewRule() rule.Invert = msg.Flags&FibRuleInvert > 0 + rule.Tos = uint(msg.Tos) for j := range attrs { switch attrs[j].Attr.Type { @@ -225,10 +253,44 @@ func (h *Handle) RuleList(family int) ([]Rule, error) { rule.Goto = int(native.Uint32(attrs[j].Value[0:4])) case nl.FRA_PRIORITY: rule.Priority = int(native.Uint32(attrs[j].Value[0:4])) + case nl.FRA_DPORT_RANGE: + rule.Dport = NewRulePortRange(native.Uint16(attrs[j].Value[0:2]), native.Uint16(attrs[j].Value[2:4])) + case nl.FRA_SPORT_RANGE: + rule.Sport = NewRulePortRange(native.Uint16(attrs[j].Value[0:2]), native.Uint16(attrs[j].Value[2:4])) + } + } + + if filter != nil { + switch { + case filterMask&RT_FILTER_SRC != 0 && + (rule.Src == nil || rule.Src.String() != filter.Src.String()): + continue + case filterMask&RT_FILTER_DST != 0 && + (rule.Dst == nil || rule.Dst.String() != filter.Dst.String()): + continue + case filterMask&RT_FILTER_TABLE != 0 && + filter.Table != unix.RT_TABLE_UNSPEC && rule.Table != filter.Table: + continue + case filterMask&RT_FILTER_TOS != 0 && rule.Tos != filter.Tos: + continue + case filterMask&RT_FILTER_PRIORITY != 0 && rule.Priority != filter.Priority: + continue + case filterMask&RT_FILTER_MARK != 0 && rule.Mark != filter.Mark: + continue + case filterMask&RT_FILTER_MASK != 0 && rule.Mask != filter.Mask: + continue } } + res = append(res, *rule) } return res, nil } + +func (pr *RulePortRange) toRtAttrData() []byte { + b := [][]byte{make([]byte, 2), make([]byte, 2)} + native.PutUint16(b[0], pr.Start) + native.PutUint16(b[1], pr.End) + return bytes.Join(b, []byte{}) +} diff --git a/vendor/github.com/vishvananda/netlink/socket_linux.go b/vendor/github.com/vishvananda/netlink/socket_linux.go index c4d89c17e..e4e7f7ac3 100644 --- a/vendor/github.com/vishvananda/netlink/socket_linux.go +++ b/vendor/github.com/vishvananda/netlink/socket_linux.go @@ -4,6 +4,7 @@ import ( "errors" "fmt" "net" + "syscall" "github.com/vishvananda/netlink/nl" "golang.org/x/sys/unix" @@ -49,10 +50,15 @@ func (r *socketRequest) Serialize() []byte { native.PutUint32(b.Next(4), r.States) networkOrder.PutUint16(b.Next(2), r.ID.SourcePort) networkOrder.PutUint16(b.Next(2), r.ID.DestinationPort) - copy(b.Next(4), r.ID.Source.To4()) - b.Next(12) - copy(b.Next(4), r.ID.Destination.To4()) - b.Next(12) + if r.Family == unix.AF_INET6 { + copy(b.Next(16), r.ID.Source) + copy(b.Next(16), r.ID.Destination) + } else { + copy(b.Next(4), r.ID.Source.To4()) + b.Next(12) + copy(b.Next(4), r.ID.Destination.To4()) + b.Next(12) + } native.PutUint32(b.Next(4), r.ID.Interface) native.PutUint32(b.Next(4), r.ID.Cookie[0]) native.PutUint32(b.Next(4), r.ID.Cookie[1]) @@ -89,10 +95,15 @@ func (s *Socket) deserialize(b []byte) error { s.Retrans = rb.Read() s.ID.SourcePort = networkOrder.Uint16(rb.Next(2)) s.ID.DestinationPort = networkOrder.Uint16(rb.Next(2)) - s.ID.Source = net.IPv4(rb.Read(), rb.Read(), rb.Read(), rb.Read()) - rb.Next(12) - s.ID.Destination = net.IPv4(rb.Read(), rb.Read(), rb.Read(), rb.Read()) - rb.Next(12) + if s.Family == unix.AF_INET6 { + s.ID.Source = net.IP(rb.Next(16)) + s.ID.Destination = net.IP(rb.Next(16)) + } else { + s.ID.Source = net.IPv4(rb.Read(), rb.Read(), rb.Read(), rb.Read()) + rb.Next(12) + s.ID.Destination = net.IPv4(rb.Read(), rb.Read(), rb.Read(), rb.Read()) + rb.Next(12) + } s.ID.Interface = native.Uint32(rb.Next(4)) s.ID.Cookie[0] = native.Uint32(rb.Next(4)) s.ID.Cookie[1] = native.Uint32(rb.Next(4)) @@ -160,3 +171,68 @@ func SocketGet(local, remote net.Addr) (*Socket, error) { } return sock, nil } + +// SocketDiagTCPInfo requests INET_DIAG_INFO for TCP protocol for specified family type. +func SocketDiagTCPInfo(family uint8) ([]*InetDiagTCPInfoResp, error) { + s, err := nl.Subscribe(unix.NETLINK_INET_DIAG) + if err != nil { + return nil, err + } + defer s.Close() + + req := nl.NewNetlinkRequest(nl.SOCK_DIAG_BY_FAMILY, unix.NLM_F_DUMP) + req.AddData(&socketRequest{ + Family: family, + Protocol: unix.IPPROTO_TCP, + Ext: INET_DIAG_INFO, + States: uint32(0xfff), // All TCP states + }) + s.Send(req) + + var result []*InetDiagTCPInfoResp +loop: + for { + msgs, from, err := s.Receive() + if err != nil { + return nil, err + } + if from.Pid != nl.PidKernel { + return nil, fmt.Errorf("Wrong sender portid %d, expected %d", from.Pid, nl.PidKernel) + } + if len(msgs) == 0 { + return nil, errors.New("no message nor error from netlink") + } + + for _, m := range msgs { + switch m.Header.Type { + case unix.NLMSG_DONE: + break loop + case unix.NLMSG_ERROR: + native := nl.NativeEndian() + error := int32(native.Uint32(m.Data[0:4])) + return nil, syscall.Errno(-error) + } + sockInfo := &Socket{} + if err := sockInfo.deserialize(m.Data); err != nil { + return nil, err + } + attrs, err := nl.ParseRouteAttr(m.Data[sizeofSocket:]) + if err != nil { + return nil, err + } + var tcpInfo *TCPInfo + for _, a := range attrs { + if a.Attr.Type == INET_DIAG_INFO { + tcpInfo = &TCPInfo{} + if err := tcpInfo.deserialize(a.Value); err != nil { + return nil, err + } + break + } + } + r := &InetDiagTCPInfoResp{InetDiagMsg: sockInfo, TCPInfo: tcpInfo} + result = append(result, r) + } + } + return result, nil +} diff --git a/vendor/github.com/vishvananda/netlink/tcp.go b/vendor/github.com/vishvananda/netlink/tcp.go new file mode 100644 index 000000000..4a42ee5a6 --- /dev/null +++ b/vendor/github.com/vishvananda/netlink/tcp.go @@ -0,0 +1,18 @@ +package netlink + +// TCP States +const ( + TCP_ESTABLISHED = iota + 0x01 + TCP_SYN_SENT + TCP_SYN_RECV + TCP_FIN_WAIT1 + TCP_FIN_WAIT2 + TCP_TIME_WAIT + TCP_CLOSE + TCP_CLOSE_WAIT + TCP_LAST_ACK + TCP_LISTEN + TCP_CLOSING + TCP_NEW_SYN_REC + TCP_MAX_STATES +) diff --git a/vendor/github.com/vishvananda/netlink/tcp_linux.go b/vendor/github.com/vishvananda/netlink/tcp_linux.go new file mode 100644 index 000000000..741ea1655 --- /dev/null +++ b/vendor/github.com/vishvananda/netlink/tcp_linux.go @@ -0,0 +1,393 @@ +package netlink + +import ( + "bytes" + "io" +) + +type TCPInfo struct { + State uint8 + Ca_state uint8 + Retransmits uint8 + Probes uint8 + Backoff uint8 + Options uint8 + Snd_wscale uint8 // no uint4 + Rcv_wscale uint8 + Delivery_rate_app_limited uint8 + Fastopen_client_fail uint8 + Rto uint32 + Ato uint32 + Snd_mss uint32 + Rcv_mss uint32 + Unacked uint32 + Sacked uint32 + Lost uint32 + Retrans uint32 + Fackets uint32 + Last_data_sent uint32 + Last_ack_sent uint32 + Last_data_recv uint32 + Last_ack_recv uint32 + Pmtu uint32 + Rcv_ssthresh uint32 + Rtt uint32 + Rttvar uint32 + Snd_ssthresh uint32 + Snd_cwnd uint32 + Advmss uint32 + Reordering uint32 + Rcv_rtt uint32 + Rcv_space uint32 + Total_retrans uint32 + Pacing_rate uint64 + Max_pacing_rate uint64 + Bytes_acked uint64 /* RFC4898 tcpEStatsAppHCThruOctetsAcked */ + Bytes_received uint64 /* RFC4898 tcpEStatsAppHCThruOctetsReceived */ + Segs_out uint32 /* RFC4898 tcpEStatsPerfSegsOut */ + Segs_in uint32 /* RFC4898 tcpEStatsPerfSegsIn */ + Notsent_bytes uint32 + Min_rtt uint32 + Data_segs_in uint32 /* RFC4898 tcpEStatsDataSegsIn */ + Data_segs_out uint32 /* RFC4898 tcpEStatsDataSegsOut */ + Delivery_rate uint64 + Busy_time uint64 /* Time (usec) busy sending data */ + Rwnd_limited uint64 /* Time (usec) limited by receive window */ + Sndbuf_limited uint64 /* Time (usec) limited by send buffer */ + Delivered uint32 + Delivered_ce uint32 + Bytes_sent uint64 /* RFC4898 tcpEStatsPerfHCDataOctetsOut */ + Bytes_retrans uint64 /* RFC4898 tcpEStatsPerfOctetsRetrans */ + Dsack_dups uint32 /* RFC4898 tcpEStatsStackDSACKDups */ + Reord_seen uint32 /* reordering events seen */ + Rcv_ooopack uint32 /* Out-of-order packets received */ + Snd_wnd uint32 /* peer's advertised receive window after * scaling (bytes) */ +} + +func checkDeserErr(err error) error { + if err == io.EOF { + return nil + } + return err +} + +func (t *TCPInfo) deserialize(b []byte) error { + var err error + rb := bytes.NewBuffer(b) + + t.State, err = rb.ReadByte() + if err != nil { + return checkDeserErr(err) + } + + t.Ca_state, err = rb.ReadByte() + if err != nil { + return checkDeserErr(err) + } + + t.Retransmits, err = rb.ReadByte() + if err != nil { + return checkDeserErr(err) + } + + t.Probes, err = rb.ReadByte() + if err != nil { + return checkDeserErr(err) + } + + t.Backoff, err = rb.ReadByte() + if err != nil { + return checkDeserErr(err) + } + t.Options, err = rb.ReadByte() + if err != nil { + return checkDeserErr(err) + } + + scales, err := rb.ReadByte() + if err != nil { + return checkDeserErr(err) + } + t.Snd_wscale = scales >> 4 // first 4 bits + t.Rcv_wscale = scales & 0xf // last 4 bits + + rateLimAndFastOpen, err := rb.ReadByte() + if err != nil { + return checkDeserErr(err) + } + t.Delivery_rate_app_limited = rateLimAndFastOpen >> 7 // get first bit + t.Fastopen_client_fail = rateLimAndFastOpen >> 5 & 3 // get next two bits + + next := rb.Next(4) + if len(next) == 0 { + return nil + } + t.Rto = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Ato = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Snd_mss = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Rcv_mss = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Unacked = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Sacked = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Lost = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Retrans = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Fackets = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Last_data_sent = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Last_ack_sent = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Last_data_recv = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Last_ack_recv = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Pmtu = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Rcv_ssthresh = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Rtt = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Rttvar = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Snd_ssthresh = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Snd_cwnd = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Advmss = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Reordering = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Rcv_rtt = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Rcv_space = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Total_retrans = native.Uint32(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Pacing_rate = native.Uint64(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Max_pacing_rate = native.Uint64(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Bytes_acked = native.Uint64(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Bytes_received = native.Uint64(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Segs_out = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Segs_in = native.Uint32(next) + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Notsent_bytes = native.Uint32(next) + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Min_rtt = native.Uint32(next) + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Data_segs_in = native.Uint32(next) + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Data_segs_out = native.Uint32(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Delivery_rate = native.Uint64(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Busy_time = native.Uint64(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Rwnd_limited = native.Uint64(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Sndbuf_limited = native.Uint64(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Delivered = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Delivered_ce = native.Uint32(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Bytes_sent = native.Uint64(next) + + next = rb.Next(8) + if len(next) == 0 { + return nil + } + t.Bytes_retrans = native.Uint64(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Dsack_dups = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Reord_seen = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Rcv_ooopack = native.Uint32(next) + + next = rb.Next(4) + if len(next) == 0 { + return nil + } + t.Snd_wnd = native.Uint32(next) + return nil +} diff --git a/vendor/github.com/vishvananda/netns/go.mod b/vendor/github.com/vishvananda/netns/go.mod index 8221f782c..9cdf577ad 100644 --- a/vendor/github.com/vishvananda/netns/go.mod +++ b/vendor/github.com/vishvananda/netns/go.mod @@ -1,3 +1,5 @@ module github.com/vishvananda/netns go 1.12 + +require golang.org/x/sys v0.0.0-20200217220822-9197077df867 diff --git a/vendor/github.com/vishvananda/netns/go.sum b/vendor/github.com/vishvananda/netns/go.sum index e69de29bb..1d4adf451 100644 --- a/vendor/github.com/vishvananda/netns/go.sum +++ b/vendor/github.com/vishvananda/netns/go.sum @@ -0,0 +1,2 @@ +golang.org/x/sys v0.0.0-20200217220822-9197077df867 h1:JoRuNIf+rpHl+VhScRQQvzbHed86tKkqwPMV34T8myw= +golang.org/x/sys v0.0.0-20200217220822-9197077df867/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs= diff --git a/vendor/github.com/vishvananda/netns/netns.go b/vendor/github.com/vishvananda/netns/netns.go index dd2f21570..116befd54 100644 --- a/vendor/github.com/vishvananda/netns/netns.go +++ b/vendor/github.com/vishvananda/netns/netns.go @@ -10,7 +10,8 @@ package netns import ( "fmt" - "syscall" + + "golang.org/x/sys/unix" ) // NsHandle is a handle to a network namespace. It can be cast directly @@ -24,11 +25,11 @@ func (ns NsHandle) Equal(other NsHandle) bool { if ns == other { return true } - var s1, s2 syscall.Stat_t - if err := syscall.Fstat(int(ns), &s1); err != nil { + var s1, s2 unix.Stat_t + if err := unix.Fstat(int(ns), &s1); err != nil { return false } - if err := syscall.Fstat(int(other), &s2); err != nil { + if err := unix.Fstat(int(other), &s2); err != nil { return false } return (s1.Dev == s2.Dev) && (s1.Ino == s2.Ino) @@ -36,11 +37,11 @@ func (ns NsHandle) Equal(other NsHandle) bool { // String shows the file descriptor number and its dev and inode. func (ns NsHandle) String() string { - var s syscall.Stat_t if ns == -1 { return "NS(None)" } - if err := syscall.Fstat(int(ns), &s); err != nil { + var s unix.Stat_t + if err := unix.Fstat(int(ns), &s); err != nil { return fmt.Sprintf("NS(%d: unknown)", ns) } return fmt.Sprintf("NS(%d: %d, %d)", ns, s.Dev, s.Ino) @@ -49,11 +50,11 @@ func (ns NsHandle) String() string { // UniqueId returns a string which uniquely identifies the namespace // associated with the network handle. func (ns NsHandle) UniqueId() string { - var s syscall.Stat_t if ns == -1 { return "NS(none)" } - if err := syscall.Fstat(int(ns), &s); err != nil { + var s unix.Stat_t + if err := unix.Fstat(int(ns), &s); err != nil { return "NS(unknown)" } return fmt.Sprintf("NS(%d:%d)", s.Dev, s.Ino) @@ -67,7 +68,7 @@ func (ns NsHandle) IsOpen() bool { // Close closes the NsHandle and resets its file descriptor to -1. // It is not safe to use an NsHandle after Close() is called. func (ns *NsHandle) Close() error { - if err := syscall.Close(int(*ns)); err != nil { + if err := unix.Close(int(*ns)); err != nil { return err } (*ns) = -1 diff --git a/vendor/github.com/vishvananda/netns/netns_linux.go b/vendor/github.com/vishvananda/netns/netns_linux.go index b1e3b07c0..c76acd087 100644 --- a/vendor/github.com/vishvananda/netns/netns_linux.go +++ b/vendor/github.com/vishvananda/netns/netns_linux.go @@ -6,46 +6,30 @@ import ( "fmt" "io/ioutil" "os" + "path" "path/filepath" - "runtime" "strconv" "strings" "syscall" -) -// SYS_SETNS syscall allows changing the namespace of the current process. -var SYS_SETNS = map[string]uintptr{ - "386": 346, - "amd64": 308, - "arm64": 268, - "arm": 375, - "mips": 4344, - "mipsle": 4344, - "mips64le": 4344, - "ppc64": 350, - "ppc64le": 350, - "riscv64": 268, - "s390x": 339, -}[runtime.GOARCH] + "golang.org/x/sys/unix" +) // Deprecated: use syscall pkg instead (go >= 1.5 needed). const ( - CLONE_NEWUTS = 0x04000000 /* New utsname group? */ - CLONE_NEWIPC = 0x08000000 /* New ipcs */ - CLONE_NEWUSER = 0x10000000 /* New user namespace */ - CLONE_NEWPID = 0x20000000 /* New pid namespace */ - CLONE_NEWNET = 0x40000000 /* New network namespace */ - CLONE_IO = 0x80000000 /* Get io context */ + CLONE_NEWUTS = 0x04000000 /* New utsname group? */ + CLONE_NEWIPC = 0x08000000 /* New ipcs */ + CLONE_NEWUSER = 0x10000000 /* New user namespace */ + CLONE_NEWPID = 0x20000000 /* New pid namespace */ + CLONE_NEWNET = 0x40000000 /* New network namespace */ + CLONE_IO = 0x80000000 /* Get io context */ + bindMountPath = "/run/netns" /* Bind mount path for named netns */ ) // Setns sets namespace using syscall. Note that this should be a method // in syscall but it has not been added. func Setns(ns NsHandle, nstype int) (err error) { - _, _, e1 := syscall.Syscall(SYS_SETNS, uintptr(ns), uintptr(nstype), 0) - if e1 != 0 { - err = e1 - } - return + return unix.Setns(int(ns), nstype) } // Set sets the current network namespace to the namespace represented @@ -57,21 +41,64 @@ func Set(ns NsHandle) (err error) { // New creates a new network namespace, sets it as current and returns // a handle to it. func New() (ns NsHandle, err error) { - if err := syscall.Unshare(CLONE_NEWNET); err != nil { + if err := unix.Unshare(CLONE_NEWNET); err != nil { return -1, err } return Get() } +// NewNamed creates a new named network namespace and returns a handle to it +func NewNamed(name string) (NsHandle, error) { + if _, err := os.Stat(bindMountPath); os.IsNotExist(err) { + err = os.MkdirAll(bindMountPath, 0755) + if err != nil { + return None(), err + } + } + + newNs, err := New() + if err != nil { + return None(), err + } + + namedPath := path.Join(bindMountPath, name) + + f, err := os.OpenFile(namedPath, os.O_CREATE|os.O_EXCL, 0444) + if err != nil { + return None(), err + } + f.Close() + + nsPath := fmt.Sprintf("/proc/%d/task/%d/ns/net", os.Getpid(), syscall.Gettid()) + err = syscall.Mount(nsPath, namedPath, "bind", syscall.MS_BIND, "") + if err != nil { + return None(), err + } + + return newNs, nil +} + +// DeleteNamed deletes a named network namespace +func DeleteNamed(name string) error { + namedPath := path.Join(bindMountPath, name) + + err := syscall.Unmount(namedPath, syscall.MNT_DETACH) + if err != nil { + return err + } + + return os.Remove(namedPath) +} + // Get gets a handle to the current threads network namespace. func Get() (NsHandle, error) { - return GetFromThread(os.Getpid(), syscall.Gettid()) + return GetFromThread(os.Getpid(), unix.Gettid()) } // GetFromPath gets a handle to a network namespace // identified by the path func GetFromPath(path string) (NsHandle, error) { - fd, err := syscall.Open(path, syscall.O_RDONLY, 0) + fd, err := unix.Open(path, unix.O_RDONLY|unix.O_CLOEXEC, 0) if err != nil { return -1, err } |