| Age | Commit message (Collapse) | Author |
|---|
|
RFC 4861 section 4.4 comments the Target link-layer address option is sometimes
optional in a Neighbor Advertisement packet:
"When responding to a unicast Neighbor Solicitation this option SHOULD be
included."
Tests:
pkg/tcpip/stack:stack_test
- TestEntryStaleToReachableWhenSolicitedConfirmationWithoutAddress
- TestEntryDelayToReachableWhenSolicitedConfirmationWithoutAddress
- TestEntryProbeToReachableWhenSolicitedConfirmationWithoutAddress
pkg/tcpip/network/ipv6:ipv6_test
- TestCallsToNeighborCache
PiperOrigin-RevId: 337396493
|
|
This change also brings back the stack.Route.ResolveWith method so that
we can immediately resolve a route when sending an NA in response to a
a NS with a source link layer address option.
Test: ipv6_test.TestNeighorSolicitationResponse
PiperOrigin-RevId: 337185461
|
|
PiperOrigin-RevId: 336339194
|
|
PiperOrigin-RevId: 336304024
|
|
When a completed entry exists for a neighbor, there is no need to block
while reachability is (re)confirmed. The stack should continue to use
the neighbor's link address while NUD is performed.
Test: stack_test.TestNeighborCacheReplace
PiperOrigin-RevId: 336199043
|
|
When a response needs to be sent to an incoming packet, the stack should
consult its neighbour table to determine the remote address's link
address.
When an entry does not exist in the stack's neighbor table, the stack
should queue the packet while link resolution completes. See comments.
PiperOrigin-RevId: 336185457
|
|
We can get the network endpoint directly from the NIC.
This is a preparatory CL for when a Route needs to hold a dedicated NIC
as its output interface. This is because when forwarding is enabled,
packets may be sent from a NIC different from the NIC a route's local
address is associated with.
PiperOrigin-RevId: 335484500
|
|
Adds support for the IPv6-compatible redirect target. Redirection is a limited
form of DNAT, where the destination is always the localhost.
Updates #3549.
PiperOrigin-RevId: 334698344
|
|
Before this change, OutgoingPacketErrors was incremented in the
stack.Route methods. This was going to be a problem once
IPv4/IPv6 WritePackets support fragmentation because Route.WritePackets
might now know how many packets are left after an error occurs.
Test:
- pkg/tcpip/network/ipv4:ipv4_test
- pkg/tcpip/network/ipv6:ipv6_test
PiperOrigin-RevId: 334687983
|
|
Add code in IPv6 to send ICMP packets while processing extension headers.
Add some accounting in processing IPV6 Extension headers which
allows us to report meaningful information back in ICMP parameter
problem packets.
IPv4 also needs to send a message when an unsupported protocol
is requested.
Add some tests to generate both ipv4 and ipv6 packets with
various errors and check the responses.
Add some new checkers and cleanup some inconsistencies in
the messages in that file.
Add new error types for the ICMPv4/6 generators.
Fix a bug in the ICMPv4 generator that stopped it from generating
"Unknown protocol" messages.
Updates #2211
PiperOrigin-RevId: 334661716
|
|
PiperOrigin-RevId: 334652998
|
|
PiperOrigin-RevId: 334531794
|
|
Test: stack_test.TestGetMainNICAddressWhenNICDisabled
PiperOrigin-RevId: 334513286
|
|
As per relevant IP RFCS (see code comments), broadcast (for IPv4) and
multicast addresses are not allowed. Currently checks for these are
done at the transport layer, but since it is explicitly forbidden at
the IP layers, check for them there.
This change also removes the UDP.InvalidSourceAddress stat since there
is no longer a need for it.
Test: ip_test.TestSourceAddressValidation
PiperOrigin-RevId: 334490971
|
|
Like matchers, targets should use a module-like register/lookup system. This
replaces the brittle switch statements we had before.
The only behavior change is supporing IPT_GET_REVISION_TARGET. This makes it
much easier to add IPv6 redirect in the next change.
Updates #3549.
PiperOrigin-RevId: 334469418
|
|
* Remove Capabilities and NICID methods from NetworkEndpoint.
* Remove linkEP and stack parameters from NetworkProtocol.NewEndpoint.
The LinkEndpoint can be fetched from the NetworkInterface. The stack
is passed to the NetworkProtocol when it is created so the
NetworkEndpoint can get it from its protocol.
* Remove stack parameter from TransportProtocol.NewEndpoint.
Like the NetworkProtocol/Endpoint, the stack is passed to the
TransportProtocol when it is created.
PiperOrigin-RevId: 334332721
|
|
* Add network address to network endpoints.
Hold network-specific state in the NetworkEndpoint instead of the stack.
This results in the stack no longer needing to "know" about the network
endpoints and special case certain work for various endpoints
(e.g. IPv6 DAD).
* Provide NetworkEndpoints with an NetworkInterface interface.
Instead of just passing the NIC ID of a NIC, pass an interface so the
network endpoint may query other information about the NIC such as
whether or not it is a loopback device.
* Move NDP code and state to the IPv6 package.
NDP is IPv6 specific so there is no need for it to live in the stack.
* Control forwarding through NetworkProtocols instead of Stack
Forwarding should be controlled on a per-network protocol basis so
forwarding configurations are now controlled through network protocols.
* Remove stack.referencedNetworkEndpoint.
Now that addresses are exposed via AddressEndpoint and only one
NetworkEndpoint is created per interface, there is no need for a
referenced NetworkEndpoint.
* Assume network teardown methods are infallible.
Fixes #3871, #3916
PiperOrigin-RevId: 334319433
|
|
Network or transport protocols may want to reach the stack. Support this
by letting the stack create the protocol instances so it can pass a
reference to itself at protocol creation time.
Note, protocols do not yet use the stack in this CL but later CLs will
make use of the stack from protocols.
PiperOrigin-RevId: 334260210
|
|
Generic ICMP errors were required because the transport dispatcher was
given the responsibility of sending ICMP errors in response to transport
packet delivery failures. Instead, the transport dispatcher should let
network layer know it failed to deliver a packet (and why) and let the
network layer make the decision as to what error to send (if any).
Fixes #4068
PiperOrigin-RevId: 333962333
|
|
segment_queue today has its own standalone limit of MaxUnprocessedSegments but
this can be a problem in UnlockUser() we do not release the lock till there are
segments to be processed. What can happen is as handleSegments dequeues packets
more keep getting queued and we will never release the lock. This can keep
happening even if the receive buffer is full because nothing can read() till we
release the lock.
Further having a separate limit for pending segments makes it harder to track
memory usage etc. Unifying the limits makes it easier to reason about memory in
use and makes the overall buffer behaviour more consistent.
PiperOrigin-RevId: 333508122
|
|
PiperOrigin-RevId: 333405169
|
|
Store transport protocol number on packet buffers for use in ICMP error
generation.
Updates #2211.
PiperOrigin-RevId: 333252762
|
|
PiperOrigin-RevId: 333138701
|
|
PiperOrigin-RevId: 332760843
|
|
PiperOrigin-RevId: 332486383
|
|
`ip6tables -t filter` is now usable. NAT support will come in a future CL.
#3549
PiperOrigin-RevId: 332381801
|
|
The lifetime of addreses in a loopback interface's associated subnets
should be bound to their respective permanent addresses.
This change also fixes a race when the stack attempts to get an IPv4
rereferencedNetworkEndpoint for an address in an associated subnet on
a loopback interface. Before this change, the stack would only check
if an IPv4 address is contained in an associated subnet while holding
a read lock but wouldn't do this same check after releasing the read
lock for a write lock to create a temporary address. This may cause
the stack to bind the lifetime of the address to a new (temporary)
endpoint instead of the associated subnet's permanent address.
Test: integration_test.TestLoopbackSubnetLifetimeBoundToAddr
PiperOrigin-RevId: 332094719
|
|
When a broadcast packet is received by the stack, the packet should be
delivered to each endpoint that may be interested in the packet. This
includes all any address and specified broadcast address listeners.
Test: integration_test.TestReuseAddrAndBroadcast
PiperOrigin-RevId: 332060652
|
|
The routing table (in its current) form should not be used to make
decisions about whether a remote address is a broadcast address or
not (for IPv4).
Note, a destination subnet does not always map to a network.
E.g. RouterA may have a route to 192.168.0.0/22 through RouterB,
but RouterB may be configured with 4x /24 subnets on 4 different
interfaces.
See https://github.com/google/gvisor/issues/3938.
PiperOrigin-RevId: 331819868
|
|
Fixes pkg/tcpip/stack:stack_test flake experienced while running
TestCacheResolution with gotsan. This occurs when the test-runner takes longer
than the resolution timeout to call linkAddrCache.get.
In this test we don't care about the resolution timeout, so set it to the
maximum and rely on test-runner timeouts to avoid deadlocks.
PiperOrigin-RevId: 330566250
|
|
The existing implementation for TransportProtocol.{Set}Option take
arguments of an empty interface type which all types (implicitly)
implement; any type may be passed to the functions.
This change introduces marker interfaces for transport protocol options
that may be set or queried which transport protocol option types
implement to ensure that invalid types are caught at compile time.
Different interfaces are used to allow the compiler to enforce read-only
or set-only socket options.
RELNOTES: n/a
PiperOrigin-RevId: 330559811
|
|
stack.cleanupEndpoints is protected by the stack.mu but that can cause
contention as the stack mutex is already acquired in a lot of hot paths during
new endpoint creation /cleanup etc. Moving this to a fine grained mutex should
reduce contention on the stack.mu.
PiperOrigin-RevId: 330026151
|
|
b/166980357#comment56 shows:
- 837 goroutines blocked in:
gvisor/pkg/sync/sync.(*RWMutex).Lock
gvisor/pkg/tcpip/stack/stack.(*Stack).StartTransportEndpointCleanup
gvisor/pkg/tcpip/transport/tcp/tcp.(*endpoint).cleanupLocked
gvisor/pkg/tcpip/transport/tcp/tcp.(*endpoint).completeWorkerLocked
gvisor/pkg/tcpip/transport/tcp/tcp.(*endpoint).protocolMainLoop.func1
gvisor/pkg/tcpip/transport/tcp/tcp.(*endpoint).protocolMainLoop
- 695 goroutines blocked in:
gvisor/pkg/sync/sync.(*RWMutex).Lock
gvisor/pkg/tcpip/stack/stack.(*Stack).CompleteTransportEndpointCleanup
gvisor/pkg/tcpip/transport/tcp/tcp.(*endpoint).cleanupLocked
gvisor/pkg/tcpip/transport/tcp/tcp.(*endpoint).completeWorkerLocked
gvisor/pkg/tcpip/transport/tcp/tcp.(*endpoint).protocolMainLoop.func1
gvisor/pkg/tcpip/transport/tcp/tcp.(*endpoint).protocolMainLoop
- 3882 goroutines blocked in:
gvisor/pkg/sync/sync.(*RWMutex).Lock
gvisor/pkg/tcpip/stack/stack.(*Stack).GetTCPProbe
gvisor/pkg/tcpip/transport/tcp/tcp.newEndpoint
gvisor/pkg/tcpip/transport/tcp/tcp.(*protocol).NewEndpoint
gvisor/pkg/tcpip/stack/stack.(*Stack).NewEndpoint
All of these are contending on Stack.mu. Stack.StartTransportEndpointCleanup()
and Stack.CompleteTransportEndpointCleanup() insert/delete TransportEndpoints
in a map (Stack.cleanupEndpoints), and the former also does endpoint
unregistration while holding Stack.mu, so it's not immediately clear how
feasible it is to replace the map with a mutex-less implementation or how much
doing so would help. However, Stack.GetTCPProbe() just reads a function object
(Stack.tcpProbeFunc) that is almost always nil (as far as I can tell,
Stack.AddTCPProbe() is only called in tests), and it's called for every new TCP
endpoint. So converting it to an atomic.Value should significantly reduce
contention on Stack.mu, improving TCP endpoint creation latency and allowing
TCP endpoint cleanup to proceed.
PiperOrigin-RevId: 330004140
|
|
Accept on gVisor will return an error if a socket in the accept queue was closed
before Accept() was called. Linux will return the new fd even if the returned
socket is already closed by the peer say due to a RST being sent by the peer.
This seems to be intentional in linux more details on the github issue.
Fixes #3780
PiperOrigin-RevId: 329828404
|
|
An earlier change considered the loopback bound to all addresses in an
assigned subnet. This should have only be done for IPv4 to maintain
compatability with Linux:
```
$ ip addr show dev lo
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group ...
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
$ ping 2001:db8::1
PING 2001:db8::1(2001:db8::1) 56 data bytes
^C
--- 2001:db8::1 ping statistics ---
4 packets transmitted, 0 received, 100% packet loss, time 3062ms
$ ping 2001:db8::2
PING 2001:db8::2(2001:db8::2) 56 data bytes
^C
--- 2001:db8::2 ping statistics ---
3 packets transmitted, 0 received, 100% packet loss, time 2030ms
$ sudo ip addr add 2001:db8::1/64 dev lo
$ ping 2001:db8::1
PING 2001:db8::1(2001:db8::1) 56 data bytes
64 bytes from 2001:db8::1: icmp_seq=1 ttl=64 time=0.055 ms
64 bytes from 2001:db8::1: icmp_seq=2 ttl=64 time=0.074 ms
64 bytes from 2001:db8::1: icmp_seq=3 ttl=64 time=0.073 ms
64 bytes from 2001:db8::1: icmp_seq=4 ttl=64 time=0.071 ms
^C
--- 2001:db8::1 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3075ms
rtt min/avg/max/mdev = 0.055/0.068/0.074/0.007 ms
$ ping 2001:db8::2
PING 2001:db8::2(2001:db8::2) 56 data bytes
From 2001:db8::1 icmp_seq=1 Destination unreachable: No route
From 2001:db8::1 icmp_seq=2 Destination unreachable: No route
From 2001:db8::1 icmp_seq=3 Destination unreachable: No route
From 2001:db8::1 icmp_seq=4 Destination unreachable: No route
^C
--- 2001:db8::2 ping statistics ---
4 packets transmitted, 0 received, +4 errors, 100% packet loss, time 3070ms
```
Test: integration_test.TestLoopbackAcceptAllInSubnet
PiperOrigin-RevId: 329011566
|
|
The existing implementation for NetworkProtocol.{Set}Option take
arguments of an empty interface type which all types (implicitly)
implement; any type may be passed to the functions.
This change introduces marker interfaces for network protocol options
that may be set or queried which network protocol option types implement
to ensure that invalid types are caught at compile time. Different
interfaces are used to allow the compiler to enforce read-only or
set-only socket options.
PiperOrigin-RevId: 328980359
|
|
Reported-by: syzbot+074ec22c42305725b79f@syzkaller.appspotmail.com
PiperOrigin-RevId: 328963899
|
|
This change was already done as of
https://github.com/google/gvisor/commit/1736b2208f but
https://github.com/google/gvisor/commit/a174aa7597 conflicted with that
change and it was missed in reviews.
This change fixes the conflict.
PiperOrigin-RevId: 328920372
|
|
The existing implementation for {G,S}etSockOpt take arguments of an
empty interface type which all types (implicitly) implement; any
type may be passed to the functions.
This change introduces marker interfaces for socket options that may be
set or queried which socket option types implement to ensure that invalid
types are caught at compile time. Different interfaces are used to allow
the compiler to enforce read-only or set-only socket options.
Fixes #3714.
RELNOTES: n/a
PiperOrigin-RevId: 328832161
|
|
In an upcoming CL, socket option types are made to implement a marker
interface with pointer receivers. Since this results in calling methods
of an interface with a pointer, we incur an allocation when attempting
to get an Endpoint's last error with the current implementation.
When calling the method of an interface, the compiler is unable to
determine what the interface implementation does with the pointer
(since calling a method on an interface uses virtual dispatch at runtime
so the compiler does not know what the interface method will do) so it
allocates on the heap to be safe incase an implementation continues to
hold the pointer after the functioon returns (the reference escapes the
scope of the object).
In the example below, the compiler does not know what b.foo does with
the reference to a it allocates a on the heap as the reference to a may
escape the scope of a.
```
var a int
var b someInterface
b.foo(&a)
```
This change removes the opportunity for that allocation.
RELNOTES: n/a
PiperOrigin-RevId: 328796559
|
|
More implementation+testing to follow.
#3549.
PiperOrigin-RevId: 328770160
|
|
This change adds an option to replace the current implementation of ARP through
linkAddrCache, with an implementation of NUD through neighborCache. Switching
to using NUD for both ARP and NDP is beneficial for the reasons described by
RFC 4861 Section 3.1:
"[Using NUD] significantly improves the robustness of packet delivery in the
presence of failing routers, partially failing or partitioned links, or nodes
that change their link-layer addresses. For instance, mobile nodes can move
off-link without losing any connectivity due to stale ARP caches."
"Unlike ARP, Neighbor Unreachability Detection detects half-link failures and
avoids sending traffic to neighbors with which two-way connectivity is
absent."
Along with these changes exposes the API for querying and operating the
neighbor cache. Operations include:
- Create a static entry
- List all entries
- Delete all entries
- Remove an entry by address
This also exposes the API to change the NUD protocol constants on a per-NIC
basis to allow Neighbor Discovery to operate over links with widely varying
performance characteristics. See [RFC 4861 Section 10][1] for the list of
constants.
Finally, an API for subscribing to NUD state changes is exposed through
NUDDispatcher. See [RFC 4861 Appendix C][3] for the list of edges.
Tests:
pkg/tcpip/network/arp:arp_test
+ TestDirectRequest
pkg/tcpip/network/ipv6:ipv6_test
+ TestLinkResolution
+ TestNDPValidation
+ TestNeighorAdvertisementWithTargetLinkLayerOption
+ TestNeighorSolicitationResponse
+ TestNeighorSolicitationWithSourceLinkLayerOption
+ TestRouterAdvertValidation
pkg/tcpip/stack:stack_test
+ TestCacheWaker
+ TestForwardingWithFakeResolver
+ TestForwardingWithFakeResolverManyPackets
+ TestForwardingWithFakeResolverManyResolutions
+ TestForwardingWithFakeResolverPartialTimeout
+ TestForwardingWithFakeResolverTwoPackets
+ TestIPv6SourceAddressSelectionScopeAndSameAddress
[1]: https://tools.ietf.org/html/rfc4861#section-10
[2]: https://tools.ietf.org/html/rfc4861#appendix-C
Fixes #1889
Fixes #1894
Fixes #1895
Fixes #1947
Fixes #1948
Fixes #1949
Fixes #1950
PiperOrigin-RevId: 328365034
|
|
When a loopback interface is configurd with an address and associated
subnet, the loopback should treat all addresses in that subnet as an
address it owns.
This is mimicking linux behaviour as seen below:
```
$ ip addr show dev lo
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group ...
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
$ ping 192.0.2.1
PING 192.0.2.1 (192.0.2.1) 56(84) bytes of data.
^C
--- 192.0.2.1 ping statistics ---
2 packets transmitted, 0 received, 100% packet loss, time 1018ms
$ ping 192.0.2.2
PING 192.0.2.2 (192.0.2.2) 56(84) bytes of data.
^C
--- 192.0.2.2 ping statistics ---
3 packets transmitted, 0 received, 100% packet loss, time 2039ms
$ sudo ip addr add 192.0.2.1/24 dev lo
$ ip addr show dev lo
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group ...
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet 192.0.2.1/24 scope global lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
$ ping 192.0.2.1
PING 192.0.2.1 (192.0.2.1) 56(84) bytes of data.
64 bytes from 192.0.2.1: icmp_seq=1 ttl=64 time=0.131 ms
64 bytes from 192.0.2.1: icmp_seq=2 ttl=64 time=0.046 ms
64 bytes from 192.0.2.1: icmp_seq=3 ttl=64 time=0.048 ms
^C
--- 192.0.2.1 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2042ms
rtt min/avg/max/mdev = 0.046/0.075/0.131/0.039 ms
$ ping 192.0.2.2
PING 192.0.2.2 (192.0.2.2) 56(84) bytes of data.
64 bytes from 192.0.2.2: icmp_seq=1 ttl=64 time=0.131 ms
64 bytes from 192.0.2.2: icmp_seq=2 ttl=64 time=0.069 ms
64 bytes from 192.0.2.2: icmp_seq=3 ttl=64 time=0.049 ms
64 bytes from 192.0.2.2: icmp_seq=4 ttl=64 time=0.035 ms
^C
--- 192.0.2.2 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3049ms
rtt min/avg/max/mdev = 0.035/0.071/0.131/0.036 ms
```
Test: integration_test.TestLoopbackAcceptAllInSubnet
PiperOrigin-RevId: 328188546
|
|
Our "Preconditions:" blocks are very useful to determine the input invariants,
but they are bit inconsistent throughout the codebase, which makes them harder
to read (particularly cases with 5+ conditions in a single paragraph).
I've reformatted all of the cases to fit in simple rules:
1. Cases with a single condition are placed on a single line.
2. Cases with multiple conditions are placed in a bulleted list.
This format has been added to the style guide.
I've also mentioned "Postconditions:", though those are much less frequently
used, and all uses already match this style.
PiperOrigin-RevId: 327687465
|
|
- Merges aleksej-paschenko's with HEAD
- Adds vfs2 support for ip_forward
|
|
Should have been removed in cl/326791119
https://github.com/google/gvisor/commit/9a7b5830aa063895f67ca0fdf653a46906374613
PiperOrigin-RevId: 327074156
|
|
Previously the netstack supported assignment of a range of addresses.
This feature is not used so remove it.
PiperOrigin-RevId: 326791119
|
|
The NetworkEndpoint does not need to be created for each address.
Most of the work the NetworkEndpoint does is address agnostic.
PiperOrigin-RevId: 326759605
|
|
Formerly, when a packet is constructed or parsed, all headers are set by the
client code. This almost always involved prepending to pk.Header buffer or
trimming pk.Data portion. This is known to prone to bugs, due to the complexity
and number of the invariants assumed across netstack to maintain.
In the new PacketHeader API, client will call Push()/Consume() method to
construct/parse an outgoing/incoming packet. All invariants, such as slicing
and trimming, are maintained by the API itself.
NewPacketBuffer() is introduced to create new PacketBuffer. Zero value is no
longer valid.
PacketBuffer now assumes the packet is a concatenation of following portions:
* LinkHeader
* NetworkHeader
* TransportHeader
* Data
Any of them could be empty, or zero-length.
PiperOrigin-RevId: 326507688
|
|
NetworkEndpoints set the number on outgoing packets in Write() and
NetworkProtocols set them on incoming packets in Parse().
Needed for #3549.
PiperOrigin-RevId: 325938745
|
