summaryrefslogtreecommitdiffhomepage
path: root/pkg
AgeCommit message (Collapse)Author
2021-06-21netstack: don't ACK SYNs in TIME-WAITKevin Krakauer
It was possible for a SYN to arrive after the endpoint sent an ACK as part of the transition to TIME-WAIT, but before returning from handleSegmentsLocked(). This caused the SYN to be dequeued and ACK'd despite the change in EndpointState. Deflakes TestTCPTimeWaitNewSyn. Tested with: blaze test --config=gotsan --runs_per_test 10000 \ //third_party/gvisor/pkg/tcpip/transport/tcp:tcp_x_test -j 2000 \ // --test_filter TestTCPTimeWaitNewSyn PiperOrigin-RevId: 380639808
2021-06-18Add endpoints to map only if registerEndpoint succeeds.Bhasker Hariharan
epsByNIC.registerEndpoint can add a multiportEndpoint to its map of nic->multiportEndpoint even if multiport.Endpoint.singleRegisterEndpoint failed. Same for transportDemuxer.singleRegisterEndpoint which ends up adding an entry to nic->epsByNIC even if epsByNIC.registerEndpoint fails. These breaks an invariant which the code assumes that a multiportEndpoint/endpointsByNIC always have at least one valid entry. PiperOrigin-RevId: 380310115
2021-06-18Include off-link route's preference in update eventsGhanan Gowripalan
RFC 4191 supports the notion of a preference value for default routers and more-specific routes, so update the OffLinkRouteUpdate event to include this preference value so integrators may prioritize routes based on a route's advertised preference value. Note, more-specific route discovery is not supported yet, but will be in a later change. Updates #6172. Test: ndp_test.TestRouterDiscovery PiperOrigin-RevId: 380243716
2021-06-17[syserror] Change p9 server to use linuxerr.Zach Koopmans
Change the p9 server to use *errors.Error defined in pkg linuxerr. Done separate from the client so that we ensure different p9 server/client versions work with each other. PiperOrigin-RevId: 380084491
2021-06-17remove outdated ip6tables TODOsKevin Krakauer
IPv6 SO_ORIGINAL_DST is supported, and the flag check as-written will detect when other flags are needed. Fixes #3549. PiperOrigin-RevId: 380059115
2021-06-17Move tcpip.Clock impl to TimekeeperTamir Duberstein
...and pass it explicitly. This reverts commit b63e61828d0652ad1769db342c17a3529d2d24ed. PiperOrigin-RevId: 380039167
2021-06-17raw sockets: don't overwrite destination addressKevin Krakauer
Also makes the behavior of raw sockets WRT fragmentation clearer, and makes the ICMPv4 header-length check explicit. Fixes #3160. PiperOrigin-RevId: 380033450
2021-06-16Fix broken hdrincl testKevin Krakauer
Fixes #3159. PiperOrigin-RevId: 379814096
2021-06-16[syserror] Refactor linuxerr and error package.Zach Koopmans
Move Error struct to pkg/errors package for use in multiple places. Move linuxerr static definitions under pkg/errors/linuxerr. Add a lookup list for quick lookup of *errors.Error by errno. This is useful when converting syserror errors and unix.Errno/syscall.Errrno values to *errors.Error. Update benchmarks routines to include conversions. The below benchmarks show *errors.Error usage to be comparable to using unix.Errno. BenchmarkAssignUnix BenchmarkAssignUnix-32 787875022 1.284 ns/op BenchmarkAssignLinuxerr BenchmarkAssignLinuxerr-32 1000000000 1.209 ns/op BenchmarkAssignSyserror BenchmarkAssignSyserror-32 759269229 1.429 ns/op BenchmarkCompareUnix BenchmarkCompareUnix-32 1000000000 1.310 ns/op BenchmarkCompareLinuxerr BenchmarkCompareLinuxerr-32 1000000000 1.241 ns/op BenchmarkCompareSyserror BenchmarkCompareSyserror-32 147196165 8.248 ns/op BenchmarkSwitchUnix BenchmarkSwitchUnix-32 373233556 3.664 ns/op BenchmarkSwitchLinuxerr BenchmarkSwitchLinuxerr-32 476323929 3.294 ns/op BenchmarkSwitchSyserror BenchmarkSwitchSyserror-32 39293408 29.62 ns/op BenchmarkReturnUnix BenchmarkReturnUnix-32 1000000000 0.5042 ns/op BenchmarkReturnLinuxerr BenchmarkReturnLinuxerr-32 1000000000 0.8152 ns/op BenchmarkConvertUnixLinuxerr BenchmarkConvertUnixLinuxerr-32 739948875 1.547 ns/op BenchmarkConvertUnixLinuxerrZero BenchmarkConvertUnixLinuxerrZero-32 977733974 1.489 ns/op PiperOrigin-RevId: 379806801
2021-06-16kvm: mark UpperHalf PTE-s as globalAndrei Vagin
UpperHalf is shared with all address spaces. PiperOrigin-RevId: 379790539
2021-06-16Merge pull request #5991 from zhlhahaha:2165gVisor bot
PiperOrigin-RevId: 379766106
2021-06-14Support parsing Prf field in RAsGhanan Gowripalan
This change prepares for a later change which actually handles the Prf field in RAs to discover default routers with preference values, as per RFC 4191. Updates #6172. Test: header_test.TestNDPRouterAdvert PiperOrigin-RevId: 379421710
2021-06-14Rename DefaultRouter event to OffLinkRoute eventGhanan Gowripalan
This change prepares for a later change which supports the NDP Route Information option to discover more-specific routes, as per RFC 4191. Updates #6172. PiperOrigin-RevId: 379361330
2021-06-14Fix typoMichael Pratt
PiperOrigin-RevId: 379337677
2021-06-14Cleanup iptables bug TODOsKevin Krakauer
There are many references to unimplemented iptables features that link to #170, but that bug is about Istio support specifically. Istio is supported, so the references should change. Some TODOs are addressed, some removed because they are not features requested by users, and some are left as implementation notes. Fixes #170. PiperOrigin-RevId: 379328488
2021-06-14Always accept discovered configurations from NDPGhanan Gowripalan
Before this change, the NDPDispatcher was allowed to "cancel" the discovery of default routers/prefixes and auto-generate addresses. No use case exists for this today so we drop this for now. If a use case comes up in the future, we should instead invalidate the discovered configuration through the stack instead of during discovery. PiperOrigin-RevId: 379327009
2021-06-13Remove usermem dependency from marshalIan Lewis
Both marshal and usermem are depended on by many packages and a dependency on marshal can often create circular dependencies. marshal should consider adding internal dependencies carefully moving forward. Fixes #6160 PiperOrigin-RevId: 379199882
2021-06-11Fix //test/syscalls:exec_test_nativeZach Koopmans
Later kernels add empty arguments to argv, throwing off return values for the exec_basic_workload.cc binary. This is result of a bug introduced by ccbb18b67323b "exec/binfmt_script: Don't modify bprm->buf and then return - ENOEXEC". Before this change, an empty interpreter string was reported if the first non-space/non-tab character after "#!" was '\0' (end of file, previously- overwritten trailing space or tab, or previously-overwritten first newline). After this change, an empty interpreter string is reported if all characters after "#!" are spaces or tabs, or the first non-space non-tab character is at i_end, which is the position of the first newline after "#!". However, if there is no newline after "#!" (as in ExecTest.InterpreterScriptNoPath), then i_end = buf_end (= bprm->buf + sizeof(bprm->buf) - 1, the last possible byte in the buffer) and neither condition holds. Change white space for script inputs to take into account the above bug. Co-authored-by: Andrei Vagin <avagin@gmail.com> PiperOrigin-RevId: 378997171
2021-06-11Rework the workaround of the XCR0 issueAndrei Vagin
XCR0 has to be synchronized with the host. We can call xsave from the host context and then call xrstor from the guest context and vise versa. This means we need to support the same set of FPU features in both contexts. PiperOrigin-RevId: 378988281
2021-06-10Minor VFS2 xattr changes.Jamie Liu
- Allow the gofer client to use most xattr namespaces. As documented by the updated comment, this is consistent with e.g. Linux's FUSE client, and allows gofers to provide extended attributes from FUSE filesystems. - Make tmpfs' listxattr omit xattrs in the "trusted" namespace for non-privileged users. PiperOrigin-RevId: 378778854
2021-06-10Fix lock ordering issue when enumerating cgroup tasks.Rahat Mahmood
The control files enumerating tasks and threads residing in cgroupfs incorrectly locks cgroupfs.filesystem.tasksMu before kernel.TaskSet.mu. The contents of these control files are inherently racy anyways, so use a snapshot of the tasks in the cgroup and drop tasksMu before resolving pids/tids (which acquires TaskSet.mu). PiperOrigin-RevId: 378767060
2021-06-10Set RLimits during `runsc exec`Fabricio Voznika
PiperOrigin-RevId: 378726430
2021-06-10Add /proc/sys/vm/max_map_countFabricio Voznika
Set it to int32 max because gVisor doesn't have a limit. Fixes #2337 PiperOrigin-RevId: 378722230
2021-06-10Parse mmap protection and flags in straceFabricio Voznika
PiperOrigin-RevId: 378712518
2021-06-10Report task exit in /proc/[pid]/{stat,status} before task goroutine exit.Jamie Liu
Between when runExitNotify.execute() returns nil (indicating that the task goroutine should exit) and when Task.run() advances Task.gosched.State to TaskGoroutineNonexistent (indicating that the task goroutine is exiting), there is a race window in which the Task is waitable (since TaskSet.mu is unlocked and Task.exitParentNotified is true) but will be reported by /proc/[pid]/status as running. Close the window by checking Task.exitState before task goroutine exit. PiperOrigin-RevId: 378711484
2021-06-10[op] Move SignalInfo to abi/linux package.Ayush Ranjan
Fixes #214 PiperOrigin-RevId: 378680466
2021-06-10Merge pull request #6103 from sudo-sturbia:semaphore-errgVisor bot
PiperOrigin-RevId: 378607458
2021-06-10[op] Move SignalStack to abi/linux package.Ayush Ranjan
Updates #214 PiperOrigin-RevId: 378594929
2021-06-09[op] Move SignalAct to abi/linux package.Ayush Ranjan
There were also other duplicate definitions of the same struct that I have now removed. Updates #214 PiperOrigin-RevId: 378579954
2021-06-09Change TODO bug to a more specific issueKevin Krakauer
This lets us close a tracking bug that's too widely-scoped to be reasonably finished. PiperOrigin-RevId: 378563203
2021-06-09Decommit huge-page-aligned regions during reclaim under manual zeroing.Jamie Liu
PiperOrigin-RevId: 378546551
2021-06-09Merge pull request #6128 from kevinGC:fanout-pidgVisor bot
PiperOrigin-RevId: 378506076
2021-06-09Change TODO to NOTE.Nicolas Lacasse
It's in VFS1 code, so we probably will not do it. PiperOrigin-RevId: 378474174
2021-06-09Avoid fanout group collisions with best effortKevin Krakauer
Running multiple instances of netstack in the same network namespace can cause collisions when enabling packet fanout for fdbased endpoints. The only bulletproof fix is to run in different network namespaces, but by using `getpid()` instead of 0 as the fanout ID starting point we can avoid collisions in the common case, particularly when testing/experimenting. Addresses #6124
2021-06-07Exclusively lock IPv6 EP when modifying addressesGhanan Gowripalan
...as address add/removal updates multicast group memberships and NDP state. This partially reverts the change made to the IPv6 endpoint in https://github.com/google/gvisor/commit/ebebb3059f7c5dbe42af85715f1c51c. PiperOrigin-RevId: 378061726
2021-06-07Remove unsupported syscall event for setsockopt(*, SOL_SOCKET, SO_OOBINLINE).Nicolas Lacasse
Netstack behaves as if SO_OOBINLINE is always set, and was logging an unsupported syscall event if the app tries to disable it. We don't have a real use case for TCP urgent mechanisms (and RFC6093 says apps SHOULD NOT use it). This CL keeps the current behavior, but removes the unsupported syscall event. Fixes #6123 PiperOrigin-RevId: 378026059
2021-06-07cgroupfs: don't add a task in the root cgroup if it is already there.Andrei Vagin
PiperOrigin-RevId: 377975013
2021-06-07Implement RENAME_NOREPLACE for all VFS2 filesystem implementations.Jamie Liu
PiperOrigin-RevId: 377966969
2021-06-05Use the NIC packets arrived at when filteringGhanan Gowripalan
As per https://linux.die.net/man/8/iptables, ``` Parameters -i, --in-interface [!] name Name of an interface via which a packet was received (only for packets entering the INPUT, FORWARD and PREROUTING chains). ``` Before this change, iptables would use the NIC that a packet was delivered to after forwarding a packet locally (when forwarding is enabled) instead of the NIC the packet arrived at. Updates #170, #3549. Test: iptables_test.TestInputHookWithLocalForwarding PiperOrigin-RevId: 377714971
2021-06-04Honor data and FIN from the ACK completing handshakeMithun Iyer
If the ACK completing the handshake has FIN or data, requeue the segment for further processing by the newly established endpoint. Otherwise, the segments would have to be retransmitted by the peer to be processed by the established endpoint. Doing this, keeps the behavior in parity with Linux. This also addresses a test flake with TCPNonBlockingConnectClose where the ACK (completing the handshake) and multiple retransmitted FINACKs from the peer could be dropped by the listener, when using syncookies and the accept queue is full. The handshake could eventually get completed with a retransmitted FINACK, without actual processing of FIN. This can cause the poll with POLLRDHUP on the accepted socket to sometimes time out before the next FINACK retransmission. PiperOrigin-RevId: 377651695
2021-06-03Add additional mmap seccomp ruleFabricio Voznika
HostFileMapper.RegenerateMappings calls mmap with MAP_SHARED|MAP_FIXED and these were not allowed. Closes #6116 PiperOrigin-RevId: 377428463
2021-06-03Implement stringer for ExitStatusTamir Duberstein
PiperOrigin-RevId: 377370807
2021-06-03Initialize metrics at initTamir Duberstein
Avoids a race condition at kernel initialization. Updates #6057. PiperOrigin-RevId: 377357723
2021-06-01Ensure full shutdown of endpoint on notifyCloseMithun Iyer
Address a race with non-blocking connect and socket close, causing the FIN (because of socket close) to not be sent out, even after completing the handshake. The race occurs with this sequence: (1) endpoint Connect starts handshake, sending out SYN (2) handshake complete() releases endpoint lock, waiting on sleeper.Fetch() (3) endpoint Close acquires endpoint lock, does not enqueue FIN (as the endpoint is not yet connected) and asserts notifyClose (4) SYNACK from peer gets enqueued asserting newSegmentWaker (5) handshake complete() re-aqcuires lock, first processes newSegmentWaker event, transitions to ESTABLISHED and proceeds to protocolMainLoop() (6) protocolMainLoop() exits while processing notifyClose When the execution follows the above sequence, no FIN is sent to the peer. This causes the listener side to have a half-open connection sitting in the accept queue. Fix this by ensuring that the protocolMainLoop() performs clean shutdown when the endpoint state is still ESTABLISHED. This would not be a bug, if during handshake complete(), sleeper.Fetch() prioritized notificationWaker over newSegmentWaker. In that case, the handshake would not have completed in (5) above. Fixes #6067 PiperOrigin-RevId: 376994395
2021-06-01Move sync generics to their own packagesTamir Duberstein
The presence of multiple packages in a single directory sometimes confuses `go mod`, producing output like: go: downloading gvisor.dev/gvisor v0.0.0-20210601174640-77dc0f5bc94d $GOMODCACHE/gvisor.dev/gvisor@v0.0.0-20210601174640-77dc0f5bc94d/pkg/linewriter/linewriter.go:21:2: found packages sync (aliases.go) and seqatomic (generic_atomicptr_unsafe.go) in $GOMODCACHE/gvisor.dev/gvisor@v0.0.0-20210601174640-77dc0f5bc94d/pkg/sync imports.go:67:2: found packages tcp (accept.go) and rcv (rcv_test.go) in $GOMODCACHE/gvisor.dev/gvisor@v0.0.0-20210601174640-77dc0f5bc94d/pkg/tcpip/transport/tcp PiperOrigin-RevId: 376956213
2021-06-01vfs: Don't allow to mount anything on top of detached mountsAndrei Vagin
PiperOrigin-RevId: 376932659
2021-06-01Ignore RST received for a TCP listenerMithun Iyer
The current implementation has a bug where TCP listener does not ignore RSTs from the peer. While handling RST+ACK from the peer, this bug can complete handshakes that use syncookies. This results in half-open connection delivered to the accept queue. Fixes #6076 PiperOrigin-RevId: 376868749
2021-05-31Update comments on ambient caps to point to bugIan Lewis
PiperOrigin-RevId: 376747671
2021-05-31Use syserror.ENOSPC for system-wide semaphore limits.Zyad A. Ali
semget(2) man page specifies that ENOSPC should be used if "the system limit for the maximum number of semaphore sets (SEMMNI), or the system wide maximum number of semaphores (SEMMNS), would be exceeded."
2021-05-28Clean up warningsTamir Duberstein
- Typos - Unused arguments - Useless conversions PiperOrigin-RevId: 376362730
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-26 06:11:49 +0000