summaryrefslogtreecommitdiffhomepage
AgeCommit message (Collapse)Author
2021-08-19Merge release-20210806.0-45-g3b4bb9475 (automated)gVisor bot
2021-08-19Add loopback interface as an ethernet-based deviceGhanan Gowripalan
...to match Linux behaviour. We can see evidence of Linux representing loopback as an ethernet-based device below: ``` # EUI-48 based MAC addresses. $ ip link show lo 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 # tcpdump showing ethernet frames when sniffing loopback and logging the # link-type as EN10MB (Ethernet). $ sudo tcpdump -i lo -e -c 2 -n tcpdump: verbose output suppressed, use -v[v]... for full protocol decode listening on lo, link-type EN10MB (Ethernet), snapshot length 262144 bytes 03:09:05.002034 00:00:00:00:00:00 > 00:00:00:00:00:00, ethertype IPv4 (0x0800), length 66: 127.0.0.1.9557 > 127.0.0.1.36828: Flags [.], ack 3562800815, win 15342, options [nop,nop,TS val 843174495 ecr 843159493], length 0 03:09:05.002094 00:00:00:00:00:00 > 00:00:00:00:00:00, ethertype IPv4 (0x0800), length 66: 127.0.0.1.36828 > 127.0.0.1.9557: Flags [.], ack 1, win 6160, options [nop,nop,TS val 843174496 ecr 843159493], length 0 2 packets captured 116 packets received by filter 0 packets dropped by kernel ``` Wireshark shows a similar result as the tcpdump example above. Linux's loopback setup: https://github.com/torvalds/linux/blob/5bfc75d92efd494db37f5c4c173d3639d4772966/drivers/net/loopback.c#L162 PiperOrigin-RevId: 391836719
2021-08-19Merge release-20210806.0-44-g50ed6b2e0 (automated)gVisor bot
2021-08-19Use a hash function to generate tcp timestamp offsetZeling Feng
Also fix an option parsing error in checker.TCPTimestampChecker while I am here. PiperOrigin-RevId: 391828329
2021-08-19Merge release-20210806.0-43-ga4ae5fed3 (automated)gVisor bot
2021-08-18Split TCP secrets from Stack to tcp.protocolZeling Feng
Use different secrets for different purposes (port picking, ISN generation, tsOffset generation) and moved the secrets from stack.Stack to tcp.protocol. PiperOrigin-RevId: 391641238
2021-08-18Merge release-20210806.0-42-g75b5a4f45 (automated)gVisor bot
2021-08-18Add control configsChong Cai
Also plumber the controls through runsc PiperOrigin-RevId: 391594318
2021-08-18Merge release-20210806.0-41-ge0bf52250 (automated)gVisor bot
2021-08-18Declare default outputs from nogo_stdlibMichael Pratt
nogo_stdlib propogate facts and findings to downstream nogo_aspects via NogoStdlibInfo. This all works fine except one case: directly building a nogo_stdlib. e.g., bazel build //tools/nogo:stdlib. In this case, nothing is requesting the NogoStdlibInfo, and thus the target has nothing to do. This can be rather confusing when trying to debug failures in :stdlib, as building :stdlib does nothing. Fix this by declaring the facts and findings as default outputs from nogo_stdlib. Now direct bazel build will request these outputs and actually trigger the analysis. Standard aspect builds are unaffected. PiperOrigin-RevId: 391580126
2021-08-18Merge release-20210806.0-40-g216b74066 (automated)gVisor bot
2021-08-17[op] Deflake SNMP Metric proc_net tests.Ayush Ranjan
Earlier the tests were checking for equality of system-wide metrics before and after some network related operations. That is inherently racy for native tests because depending on the testing infrastructure, multiple tests might run parallely hence trampling over each other's metrics. Tests should only compare metrics that are increasing in nature. The comparison should not be a hard comparison, instead a less-than/greater-than relation test. I have changed the checks and also removed tests for tcpCurrEstab metric which has "SYNTAX Gauge" and hence can not be tested reliably. PiperOrigin-RevId: 391460081
2021-08-18Merge release-20210806.0-39-gb495ae599 (automated)gVisor bot
2021-08-17Merge pull request #6262 from sudo-sturbia:msgqueue/syscalls3gVisor bot
PiperOrigin-RevId: 391416650
2021-08-17Merge release-20210806.0-34-g8f6c54c8c (automated)gVisor bot
2021-08-17Deflake test/perf:randread_benchmarkAndrei Vagin
The test expects that pread reads the full buffer, it means that the pread offset has to be equal or less than file_size - buffer_size. PiperOrigin-RevId: 391356863
2021-08-17Implement stub for msgctl(2).Zyad A. Ali
Add support for msgctl and enable tests. Fixes #135
2021-08-17Implement control operations on msgqueue.Zyad A. Ali
For IPCInfo, update value of MSGSEG constant in abi to avoid overflow in MsgInfo.MsgSeg. MSGSEG was originaly simplified in abi, and is unused (by us and within the kernel), so updating it is okay. Updates #135
2021-08-17Implement ipc.Object.Set and use it in ipc mechanisms.Zyad A. Ali
Set provides functionality of {sem,shm,msg}ctl(IPC_SET).
2021-08-17Add tests for msgctl(2).Zyad A. Ali
Updates #135
2021-08-17Merge release-20210806.0-33-gebf76b30c (automated)gVisor bot
2021-08-17Internal change.gVisor bot
PiperOrigin-RevId: 391331401
2021-08-17Merge release-20210806.0-32-gfa32136ac (automated)gVisor bot
2021-08-16Internal change.gVisor bot
PiperOrigin-RevId: 391217339
2021-08-17Merge release-20210806.0-31-g6294a7a6e (automated)gVisor bot
2021-08-16test/syscalls/proc_net: /proc/net/snmp can contain system-wide statisticsAndrei Vagin
This is a new kernel feature that are controlled by the net.core.mibs_allocation sysctl. PiperOrigin-RevId: 391215784
2021-08-16Merge release-20210806.0-30-gbb13d015a (automated)gVisor bot
2021-08-16imges/syzkaller: add --allow-releaseinfo-change to apt updateAndrei Vagin
Otherwise, it fails with this error: Get:3 http://security.debian.org/debian-security buster/updates InRelease Reading package lists... E: Repository 'http://deb.debian.org/debian buster InRelease' changed its 'Suite' value from 'stable' to 'oldstable' PiperOrigin-RevId: 391155532
2021-08-14Merge release-20210806.0-29-gce58d71fd (automated)gVisor bot
2021-08-13[syserror] Remove pkg syserror.Zach Koopmans
Removes package syserror and moves still relevant code to either linuxerr or to syserr (to be later removed). Internal errors are converted from random types to *errors.Error types used in linuxerr. Internal errors are in linuxerr/internal.go. PiperOrigin-RevId: 390724202
2021-08-13Merge release-20210806.0-28-g868ed0e80 (automated)gVisor bot
2021-08-13[benchmarks] Update BenchmarkStartEmpty benchmark.Zach Koopmans
Update the start benchmark on empty to only "Start" a container, not wait for its completion. TL:DR only measure the actual start call for the empty container. Previously, we were measuring the completion of /bin/true in container alpine AND the cleanup. This was fine until profiling started failing all the time on ptrace. This is a cost that runc is not paying. These changes will reduce the over all timing of the benchmark, but it will give more sane results. Instead, use "Spawn" which is similar to `docker run --detach alpine /bin/sleep 100`. Call sleep so containers stick around long enough for the profiler to read profile data from them. PiperOrigin-RevId: 390705431
2021-08-13Merge release-20210806.0-27-g6eb8596f7 (automated)gVisor bot
2021-08-13Add Event controlsChong Cai
Add Event controls and implement "stream" commands. PiperOrigin-RevId: 390691702
2021-08-13Merge release-20210806.0-26-ga7b59445d (automated)gVisor bot
2021-08-13Fix minor typoMichael Pratt
PiperOrigin-RevId: 390659097
2021-08-13Merge release-20210806.0-25-ged8960216 (automated)gVisor bot
2021-08-13Disable SA1019 (deprecation check)Michael Pratt
On Go tip (pre-1.18), http://golang.org/issue/44195 is making SA1019 mistake uses of reflect.Value.Len for reflect.Value.InterfaceData, which is deprecated. It is thus mistakenly raising deprecation errors on uses of reflect.Value.Len. Suppress these errors by disabling SA1019 entirely. This is a bit overkill, but it is unclear to me if we want hard errors on deprecation anyways. That can be reevaluated when http://golang.org/issue/44195 is fixed. The other staticcheck analyzers are moved to alphabetical order. Updates golang/go#44195 PiperOrigin-RevId: 390655918
2021-08-13Merge release-20210806.0-24-g8f2b11a87 (automated)gVisor bot
2021-08-13Update `core` allowed dependenciesMichael Pratt
This list has gotten a little out-of-date. Note that `clockwork` used to be used but was removed in gvisor.dev/pr/5384. PiperOrigin-RevId: 390644841
2021-08-13Merge release-20210806.0-23-geb0f24c6c (automated)gVisor bot
2021-08-13Free multicastMemberships on UDP endpoint closeGhanan Gowripalan
tcpip.Endpoint.Close is documented to free all resources associated with an endpoint so we don't need to create an empty map to clear the multicast memberships. PiperOrigin-RevId: 390609826
2021-08-13Merge release-20210806.0-22-gddcf884e9 (automated)gVisor bot
2021-08-12Add Usage controlsChong Cai
Add Usage controls and implement "usage/usagefd" commands. PiperOrigin-RevId: 390507423
2021-08-12Merge release-20210806.0-21-g02370bbd3 (automated)gVisor bot
2021-08-12[syserror] Convert remaining syserror definitions to linuxerr.Zach Koopmans
Convert remaining public errors (e.g. EINTR) from syserror to linuxerr. PiperOrigin-RevId: 390471763
2021-08-12Merge release-20210806.0-20-g5f132ae1f (automated)gVisor bot
2021-08-12Clear Merkle files before measuring verity fsChong Cai
PiperOrigin-RevId: 390467957
2021-08-12Merge release-20210806.0-19-g345eb4a66 (automated)gVisor bot
2021-08-12fix typoKevin Krakauer
PiperOrigin-RevId: 390463819
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-28 05:43:53 +0000