summaryrefslogtreecommitdiffhomepage
AgeCommit message (Collapse)Author
2019-05-21Refactor fdbased endpoint dispatcher code.Bhasker Hariharan
This is in preparation to support an fdbased endpoint that can read/dispatch packets from multiple underlying fds. Updates #231 PiperOrigin-RevId: 249337074 Change-Id: Id7d375186cffcf55ae5e38986e7d605a96916d35
2019-05-21Add basic plumbing for splice and stub implementation.Adin Scannell
This does not actually implement an efficient splice or sendfile. Rather, it adds a generic plumbing to the file internals so that this can be added. All file implementations use the stub fileutil.NoSplice implementation, which causes sendfile and splice to fall back to an internal copy. A basic splice system call interface is added, along with a test. PiperOrigin-RevId: 249335960 Change-Id: Ic5568be2af0a505c19e7aec66d5af2480ab0939b
2019-05-21Remove unused struct member.Neel Natu
Remove unused struct member. PiperOrigin-RevId: 249300446 Change-Id: Ifb16538f684bc3200342462c3da927eb564bf52d
2019-05-20Forward named pipe creation to the goferMichael Pratt
The backing 9p server must allow named pipe creation, which the runsc fsgofer currently does not. There are small changes to the overlay here. GetFile may block when opening a named pipe, which can cause a deadlock: 1. open(O_RDONLY) -> copyMu.Lock() -> GetFile() 2. open(O_WRONLY) -> copyMu.Lock() -> Deadlock A named pipe usable for writing must already be on the upper filesystem, but we are still taking copyMu for write when checking for upper. That can be changed to a read lock to fix the common case. However, a named pipe on the lower filesystem would still deadlock in open(O_WRONLY) when it tries to actually perform copy up (which would simply return EINVAL). Move the copy up type check before taking copyMu for write to avoid this. p9 must be modified, as it was incorrectly removing the file mode when sending messages on the wire. PiperOrigin-RevId: 249154033 Change-Id: Id6637130e567b03758130eb6c7cdbc976384b7d6
2019-05-20Fix incorrect tmpfs timestamp updatesMichael Pratt
* Creation of files, directories (and other fs objects) in a directory should always update ctime. * Same for removal. * atime should not be updated on lookup, only readdir. I've also renamed some misleading functions that update mtime and ctime. PiperOrigin-RevId: 249115063 Change-Id: I30fa275fa7db96d01aa759ed64628c18bb3a7dc7
2019-05-17Return EPERM for mknodMichael Pratt
This more directly matches what Linux does with unsupported nodes. PiperOrigin-RevId: 248780425 Change-Id: I17f3dd0b244f6dc4eb00e2e42344851b8367fbec
2019-05-17Fix gofer rename ctime and cleanup stat_times testMichael Pratt
There is a lot of redundancy that we can simplify in the stat_times test. This will make it easier to add new tests. However, the simplification reveals that cached uattrs on goferfs don't properly update ctime on rename. PiperOrigin-RevId: 248773425 Change-Id: I52662728e1e9920981555881f9a85f9ce04041cf
2019-05-17Bump rules_go to 0.18.5 and go tool chain to 1.12.5.Nicolas Lacasse
PiperOrigin-RevId: 248719758 Change-Id: I0a87d8befd56c5a791585818e9b84ac0c443b0df
2019-05-16tiny fix: avoid panicing when OpenSpec failedLiu Hua
Signed-off-by: Liu Hua <sdu.liu@huawei.com> Change-Id: I11a4620394a10a7d92036b0341e0c21ad50bd122 PiperOrigin-RevId: 248621810
2019-05-16Add test for duplicate proc entries.Ian Gudger
The issue with duplicate /proc/sys entries seems to have been fixed in: PiperOrigin-RevId 229305982 Git hash dc8450b5676d4c4ac9bcfa23cabd862e0060527d Fixes google/gvisor#125 PiperOrigin-RevId: 248571903 Change-Id: I76ff3b525c93dafb92da6e5cf56e440187f14579
2019-05-15gofer: don't call hostfile.Close if hostFile is nilAndrei Vagin
PiperOrigin-RevId: 248437159 Change-Id: Ife71f6ca032fca59ec97a82961000ed0af257101
2019-05-15Automated rollback of changelist 247964961Andrei Vagin
PiperOrigin-RevId: 248411456 Change-Id: I21c3767b0b7e5948536d4c0b78be46ba35cf76cb
2019-05-15Cleanup around urpc file payload handlingFabricio Voznika
urpc always closes all files once the RPC function returns. PiperOrigin-RevId: 248406857 Change-Id: I400a8562452ec75c8e4bddc2154948567d572950
2019-05-15gvisor/runsc: use a veth link address instead of generating a new oneAndrei Vagin
PiperOrigin-RevId: 248367340 Change-Id: Id792afcfff9c9d2cfd62cae21048316267b4a924
2019-05-14Update out of date commentMichael Pratt
PiperOrigin-RevId: 248265524 Change-Id: Ib9082f08d24ba10535079cf89c714fb22a4fdf10
2019-05-14Start of support for /proc/pid/cgroup file.Nicolas Lacasse
PiperOrigin-RevId: 248263378 Change-Id: Ic057d2bb0b6212110f43ac4df3f0ac9bf931ab98
2019-05-14Remove false commentMichael Pratt
PiperOrigin-RevId: 248249285 Change-Id: I9b6d267baa666798b22def590ff20c9a118efd47
2019-05-14kokoro: run tests with a default docker container runtimeAndrei Vagin
We want to know that our environment set up properly and docker tests pass with a native runtime. PiperOrigin-RevId: 248229294 Change-Id: I06c221e5eeed6e01bdd1aa935333c57e8eadc498
2019-05-13gvisor/hostnet: restart epoll_wait after epoll_ctlAndrei Vagin
Otherwise changes of epoll_ctl will not have affect. PiperOrigin-RevId: 247964961 Change-Id: I9fbb35c44766421af45d9ed53760e0c324d80d99
2019-05-13gvisor/runsc/tests: set timeout for http.Get()Andrei Vagin
WaitForHTTP tries GET requests on a port until the call succeeds or timeout. But we want to be sure that one of our attempts will not stuck for the whole timeout. All timeouts are increased to 30 seconds, because test cases with smaller timeouts fail sometimes even for the native container runtime (runc). PiperOrigin-RevId: 247888467 Change-Id: I03cfd3275286bc686a78fd26da43231d20667851
2019-05-10Modernize mknod testMichael Pratt
PiperOrigin-RevId: 247704588 Change-Id: I1e63e2b310145695fbe38429b91e44d72473fcd6
2019-05-10Add pgalloc.DelayedEvictionManual.Jamie Liu
PiperOrigin-RevId: 247667272 Change-Id: I16b04e11bb93f50b7e05e888992303f730e4a877
2019-05-09Implement fallocate(2)Fabricio Voznika
Closes #225 PiperOrigin-RevId: 247508791 Change-Id: I04f47cf2770b30043e5a272aba4ba6e11d0476cc
2019-05-09Remove dhcp clientTamir Duberstein
This was upstreamed from Fuchsia, but it is pretty buggy and doesn't rely on any private APIs. Thus it can be checked into the Fuchsia source tree without forking netstack, where we can more easily iterate on (and eventually remove) it. PiperOrigin-RevId: 247506582 Change-Id: Ifb1b60c6c4941c374a59c5570a6a9cacf2468981
2019-05-08Fix types that are subtly incorrect.Googler
PiperOrigin-RevId: 247294093 Change-Id: Iac8c76e50bbc15c240ae7da7f5786f9968e7057c
2019-05-08Set the FilesytemType in MountSource from the Filesystem.Nicolas Lacasse
And stop storing the Filesystem in the MountSource. This allows us to decouple the MountSource filesystem type from the name of the filesystem. PiperOrigin-RevId: 247292982 Change-Id: I49cbcce3c17883b7aa918ba76203dfd6d1b03cc8
2019-05-07Check GSO for nil in WritePacketGoogler
Testing: Unit tests added PiperOrigin-RevId: 247096269 Change-Id: I849c010eadcb53caf45896a15ef38162d66a9568
2019-05-07Add gonet.DialContextTCP.Ian Gudger
Allows cancellation and timeouts. PiperOrigin-RevId: 247090428 Change-Id: I91907f12e218677dcd0e0b6d72819deedbd9f20c
2019-05-07Remove defers from gofer.contextFileFabricio Voznika
Most are single line methods in hot paths. PiperOrigin-RevId: 247050267 Change-Id: I428d78723fe00b57483185899dc8fa9e1f01e2ea
2019-05-06Ensure all uses of MM.brk occur under MM.mappingMu in MM.Brk().Jamie Liu
PiperOrigin-RevId: 246921386 Change-Id: I71d8908858f45a9a33a0483470d0240eaf0fd012
2019-05-05Fix raw socket behavior and tests.Kevin Krakauer
Some behavior was broken due to the difficulty of running automated raw socket tests. Change-Id: I152ca53916bb24a0208f2dc1c4f5bc87f4724ff6 PiperOrigin-RevId: 246747067
2019-05-03Add arm64 support to pkg/seccompBin Lu
Signed-off-by: Bin Lu <bin.lu@arm.com> PiperOrigin-RevId: 246622505 Change-Id: I803639a0c5b0f75959c64fee5385314214834d10
2019-05-03Fix runsc restore to be compatible with docker start --checkpoint ...Andrei Vagin
Change-Id: I02b30de13f1393df66edf8829fedbf32405d18f8 PiperOrigin-RevId: 246621192
2019-05-03Update tcpip Clock description.Ian Gudger
The tcpip.Clock comment stated that times provided by it should not be used for netstack internal timekeeping. This comment was from before the interface supported monotonic times. The monotonic times that it provides are now be the preferred time source for netstack internal timekeeping. PiperOrigin-RevId: 246618772 Change-Id: I853b720e3d719b03fabd6156d2431da05d354bda
2019-05-03gvisor: run bazel in a docker containerAndrei Vagin
bazel has a lot of dependencies and users don't want to install them just to build gvisor. These changes allows to run bazel in a docker container. A bazel cache is on the local file system (~/.cache/bazel), so incremental builds should be fast event after recreating a bazel container. Here is an example how to build runsc: make BAZEL_OPTIONS="build runsc:runsc" bazel Change-Id: I8c0a6d0c30e835892377fb6dd5f4af7a0052d12a PiperOrigin-RevId: 246570877
2019-05-03gofer: don't leak file descriptorsAndrei Vagin
Fixes #219 PiperOrigin-RevId: 246568639 Change-Id: Ic7afd15dde922638d77f6429c508d1cbe2e4288a
2019-05-03Support IPv4 fragmentation in netstackGoogler
Testing: Unit tests and also large ping in Fuchsia OS PiperOrigin-RevId: 246563592 Change-Id: Ia12ab619f64f4be2c8d346ce81341a91724aef95
2019-05-03Add netfilter ABI for iptables support.Kevin Krakauer
Change-Id: Ifbd2abf63ea8062a89b83e948d3e9735480d8216 PiperOrigin-RevId: 246559904
2019-05-03Fix transport/raw copybara exportTamir Duberstein
- include packet_list.go - exclude state.go (by renaming to include an underscore) Also rename raw.go to endpoint.go for consistency. PiperOrigin-RevId: 246547912 Change-Id: I19c8331c794ba683a940cc96a8be6497b53ff24d
2019-05-03runsc: add a bazel target to build a debian packageAndrei Vagin
$ dpkg -s runsc Package: runsc Status: install ok installed Priority: optional Section: contrib/devel Maintainer: The gVisor Authors <gvisor-dev@googlegroups.com> Architecture: amd64 Version: 20190304.1-123-g861434f612ce-dirty Description: gVisor is a user-space kernel, written in Go, that implements a substantial portion of the Linux system surface. It includes an Open Container Initiative (OCI) runtime called runsc that provides an isolation boundary between the application and the host kernel. The runsc runtime integrates with Docker and Kubernetes, making it simple to run sandboxed containers. Homepage: https://gvisor.dev/ Built-Using: Bazel Change-Id: I6f161de8fba649f12272a87b99529ccfd22e499a PiperOrigin-RevId: 246546294
2019-05-03gvisor/kokoro: save runsc logsAndrei Vagin
PiperOrigin-RevId: 246542315 Change-Id: Ia9ba2bc104e0af3277d3b6102122c13d320ea802
2019-05-03Implement support for SACK based recovery(RFC 6675).Bhasker Hariharan
PiperOrigin-RevId: 246536003 Change-Id: I118b745f45040be9c70cb6a1028acdb06c78d8c9
2019-05-03Increase timeout to wait for port to become availableFabricio Voznika
TestHttpd fails sporadically waiting for the port on slow machines. PiperOrigin-RevId: 246525277 Change-Id: Ie0ea71e3c4664d24f580eabd8f7461e47079f734
2019-05-03Skip flaky ClockGettime.CputimeId take 2Fabricio Voznika
The test also times out when GCE machine has 2 CPUs. I cannot repro it locally with a 2 CPU cgroup though. Let's skip the test when there are 2 CPUs to stop the flakiness and retest it once the fix is available. PiperOrigin-RevId: 246523363 Change-Id: I9d9d922a5be3aa7bc91dff5a1807ca99f3f4a4f9
2019-05-03Proper cleanup of sockets that used REUSEPORTChris Kuiper
Fixed a small logic error that broke proper accounting of MultiPortEndpoints. PiperOrigin-RevId: 246502126 Change-Id: I1a7d6ea134f811612e545676212899a3707bc2c2
2019-05-02Support reception of multicast data on more than one socketChris Kuiper
This requires two changes: 1) Support for more than one socket to join a given multicast group. 2) Duplicate delivery of incoming multicast packets to all sockets listening for it. In addition, I tweaked the code (and added a test) to disallow duplicates IP_ADD_MEMBERSHIP calls for the same group and NIC. This is how Linux does it. PiperOrigin-RevId: 246437315 Change-Id: Icad8300b4a8c3f501d9b4cd283bd3beabef88b72
2019-05-02runsc: don't create an empty network namespace if NetworkHost is setAndrei Vagin
With this change, we will be able to run runsc do in a host network namespace. PiperOrigin-RevId: 246436660 Change-Id: I8ea18b1053c88fe2feed74239b915fe7a151ce34
2019-05-02runsc: move test_app in a separate directoryAndrei Vagin
Opensource tools (e. g. https://github.com/fatih/vim-go) can't hanlde more than one golang package in one directory. PiperOrigin-RevId: 246435962 Change-Id: I67487915e3838762424b2d168efc54ae34fb801f
2019-05-02Replace dynamic macros with constants in memfd test.Kevin Krakauer
PiperOrigin-RevId: 246433167 Change-Id: Idb9b6c20ee1da193176288dfd2f9d85ec0e69c54
2019-05-02Add [simple] network support to 'runsc do'Fabricio Voznika
Sandbox always runsc with IP 192.168.10.2 and the peer network adds 1 to the address (192.168.10.3). Sandbox IP can be changed using --ip flag. Here a few examples: sudo runsc do curl www.google.com sudo runsc do --ip=10.10.10.2 bash -c "echo 123 | netcat -l -p 8080" PiperOrigin-RevId: 246421277 Change-Id: I7b3dce4af46a57300350dab41cb27e04e4b6e9da
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-05 03:04:04 +0000