summaryrefslogtreecommitdiffhomepage
path: root/pkg
AgeCommit message (Collapse)Author
2020-08-10Run GC before sandbox exit when leak checking is enabled.Dean Deng
Running garbage collection enqueues all finalizers, which are used by the refs/refs_vfs2 packages to detect reference leaks. Note that even with GC, there is no guarantee that all finalizers will be run before the program exits. This is a best effort attempt to activate leak checks as much as possible. Updates #3545. PiperOrigin-RevId: 325834438
2020-08-08Use unicast source for ICMP echo repliesGhanan Gowripalan
Packets MUST NOT use a non-unicast source address for ICMP Echo Replies. Test: integration_test.TestPingMulticastBroadcast PiperOrigin-RevId: 325634380
2020-08-07Add context.FullStateChanged()Andrei Vagin
It indicates that the Sentry has changed the state of the thread and next calls of PullFullState() has to do nothing. PiperOrigin-RevId: 325567415
2020-08-07[vfs2] Fix tmpfs mounting.Ayush Ranjan
Earlier we were using NLink to decide if /tmp is empty or not. However, NLink at best tells us about the number of subdirectories (via the ".." entries). NLink = n + 2 for n subdirectories. But it does not tell us if the directory is empty. There still might be non-directory files. We could also not rely on NLink because host overlayfs always returned 1. VFS1 uses Readdir to decide if the directory is empty. Used a similar approach. We now use IterDirents to decide if the "/tmp" directory is empty. Fixes #3369 PiperOrigin-RevId: 325554234
2020-08-07Don't hold gofer.filesystem.renameMu during dentry destruction.Jamie Liu
PiperOrigin-RevId: 325546629
2020-08-07Merge pull request #3069 from lubinszARM:pr_serr_injection2gVisor bot
PiperOrigin-RevId: 325546308
2020-08-07Mark dropped pages unevictable in fsimpl/gofer.dentry.destroyLocked.Jamie Liu
PiperOrigin-RevId: 325531657
2020-08-07Fix panic during Address Resolution of neighbor entry created by NSSam Balana
When a Neighbor Solicitation is received, a neighbor entry is created with the remote host's link layer address, but without a link layer address resolver. If the host decides to send a packet addressed to the IP address of that neighbor entry, Address Resolution starts with a nil pointer to the link layer address resolver. This causes the netstack to panic and crash. This change ensures that when a packet is sent in that situation, the link layer address resolver will be set before Address Resolution begins. Tests: pkg/tcpip/stack:stack_test + TestEntryUnknownToStaleToProbeToReachable - TestNeighborCacheEntryNoLinkAddress Updates #1889 Updates #1894 Updates #1895 Updates #1947 Updates #1948 Updates #1949 Updates #1950 PiperOrigin-RevId: 325516471
2020-08-07Port Ruby benchmark.Zach Koopmans
PiperOrigin-RevId: 325500772
2020-08-07tcp: change the limit of TCP_LINGER2Andrei Vagin
It was changed in the Linux kernel: commit f0628c524fd188c3f9418e12478dfdfadacba815 Date: Fri Apr 24 16:06:16 2020 +0800 net: Replace the limit of TCP_LINGER2 with TCP_FIN_TIMEOUT_MAX PiperOrigin-RevId: 325493859
2020-08-07Support separate read/write handles in fsimpl/gofer.dentry.Jamie Liu
PiperOrigin-RevId: 325490674
2020-08-07Try to update atime and mtime on VFS2 gofer files on dentry eviction.Jamie Liu
PiperOrigin-RevId: 325388385
2020-08-06Add LinkAt support to goferFabricio Voznika
Updates #1198 PiperOrigin-RevId: 325350818
2020-08-06Add reference counting utility to VFS2.Dean Deng
The utility has several differences from the VFS1 equivalent: - There are no weak references, which have a significant overhead - In order to print useful debug messages with the type of the reference- counted object, we use a generic Refs object with the owner type as a template parameter. In vfs1, this was accomplished by storing a type name and caller stack directly in the ref count (as in vfs1), which increases the struct size by 6x. (Note that the caller stack was needed because fs types like Dirent were shared by all fs implementations; in vfs2, each impl has its own data structures, so this is no longer necessary.) Updates #1486. PiperOrigin-RevId: 325271469
2020-08-06Only register /dev/net/tun if supported.Dean Deng
PiperOrigin-RevId: 325266487
2020-08-06Join IPv4 all-systems group on NIC enableGhanan Gowripalan
Test: - stack_test.TestJoinLeaveMulticastOnNICEnableDisable - integration_test.TestIncomingMulticastAndBroadcast PiperOrigin-RevId: 325185259
2020-08-05Add loss recovery option for TCP.Nayana Bidari
/proc/sys/net/ipv4/tcp_recovery is used to enable RACK loss recovery in TCP. PiperOrigin-RevId: 325157807
2020-08-05Correctly decrement link counts in tmpfs rename operations.Dean Deng
When a directory is replaced by a rename operation, its link count should reach zero. We were missing the link from `dir/.` PiperOrigin-RevId: 325141730
2020-08-05Support receiving broadcast IPv4 packetsGhanan Gowripalan
Test: integration_test.TestIncomingSubnetBroadcast PiperOrigin-RevId: 325135617
2020-08-05Release extra memfd reference.Dean Deng
PiperOrigin-RevId: 325122849
2020-08-05Prefer RLock over Lock in functions that don't need Lock().Bhasker Hariharan
Updates #231 PiperOrigin-RevId: 325097683
2020-08-05Stop profiling when the sentry exitsFabricio Voznika
Also removes `--profile-goroutine` because it's equivalent to `debug --stacks`. PiperOrigin-RevId: 325061502
2020-08-05Add missing case in tmpfs.inode.direntType.Dean Deng
This was discovered by syzkaller. PiperOrigin-RevId: 325025193
2020-08-04Update variables for implementation of RACK in TCPNayana Bidari
RACK (Recent Acknowledgement) is a new loss detection algorithm in TCP. These are the fields which should be stored on connections to implement RACK algorithm. PiperOrigin-RevId: 324948703
2020-08-04Handle EOF in vfs2 sendfile.Dean Deng
Discovered by syzkaller. PiperOrigin-RevId: 324938438
2020-08-04Inline gofer.regularFileFD.pwriteLockedFabricio Voznika
Go compiler barely inlines anything, so inline by hand pwriteLocked since it's called from a single place. PiperOrigin-RevId: 324937734
2020-08-04Automated rollback of changelist 324906582Dean Deng
PiperOrigin-RevId: 324931854
2020-08-04Use 1 fragmentation component per IP stackGhanan Gowripalan
This will help manage memory consumption by IP reassembly when receiving IP fragments on multiple network endpoints. Previously, each endpoint would cap memory consumption at 4MB, but with this change, each IP stack will cap memory consumption at 4MB. No behaviour changes. PiperOrigin-RevId: 324913904
2020-08-04Add reference counting utility to VFS2.Dean Deng
The utility has several differences from the VFS1 equivalent: - There are no weak references, which have a significant overhead - In order to print useful debug messages with the type of the reference- counted object, we use a generic Refs object with the owner type as a template parameter. In vfs1, this was accomplished by storing a type name and caller stack directly in the ref count (as in vfs1), which increases the struct size by 6x. (Note that the caller stack was needed because fs types like Dirent were shared by all fs implementations; in vfs2, each impl has its own data structures, so this is no longer necessary.) As an example, the utility is added to tmpfs.inode. Updates #1486. PiperOrigin-RevId: 324906582
2020-08-04Internal change.gVisor bot
PiperOrigin-RevId: 324826968
2020-08-04Internal change.gVisor bot
PiperOrigin-RevId: 324819246
2020-08-03Add callbacks to support lazy loading/restoring thread statesAndrei Vagin
PiperOrigin-RevId: 324748508
2020-08-03[vfs2] Implement /sys/devices/system/cpu/cpuX.Ayush Ranjan
Fixes #3364 PiperOrigin-RevId: 324724614
2020-08-03Internal change.gVisor bot
PiperOrigin-RevId: 324695672
2020-08-03Remove old TODO.Dean Deng
Fixes #2920. PiperOrigin-RevId: 324695118
2020-08-03Plumbing context.Context to DecRef() and Release().Nayana Bidari
context is passed to DecRef() and Release() which is needed for SO_LINGER implementation. PiperOrigin-RevId: 324672584
2020-08-03Merge pull request #3460 from zhlhahaha:1927gVisor bot
PiperOrigin-RevId: 324658881
2020-08-03Add inotify events for fallocate and tests for fallocate/sendfile.Dean Deng
Updates #1479, #2923. PiperOrigin-RevId: 324658826
2020-08-03AARCH64:fix variable name collision with register nameHoward Zhang
The variable name is g which is collision with the reserved name for R28. This leads to bazel build failure on ARM with following information: (register+register) not supported on this architecture rename it from g to ptr (referenced from golang source code) Signed-off-by: Howard Zhang <howard.zhang@arm.com>
2020-07-31Merge pull request #3300 from lubinszARM:pr_fpsimd_usrgVisor bot
PiperOrigin-RevId: 324309862
2020-07-31Support fragments from different sourcesGhanan Gowripalan
Prevent fragments with different source-destination pairs from conflicting with each other. Test: - ipv6_test.TestReceiveIPv6Fragments - ipv4_test.TestReceiveIPv6Fragments PiperOrigin-RevId: 324283246
2020-07-31Merge pull request #3348 from kevinGC:so-orig-dstgVisor bot
PiperOrigin-RevId: 324279280
2020-07-31Internal change.gVisor bot
PiperOrigin-RevId: 324259991
2020-07-31iptables: support SO_ORIGINAL_DSTKevin Krakauer
Envoy (#170) uses this to get the original destination of redirected packets.
2020-07-31Clean up vfs2 fallocate.Dean Deng
Move to setstat.go and add a FileDescription wrapper method. PiperOrigin-RevId: 324165277
2020-07-30Fix TCP CurrentConnected counter updates.Mithun Iyer
CurrentConnected counter is incorrectly decremented on close of an endpoint which is still not connected. Fixes #3443 PiperOrigin-RevId: 324155171
2020-07-30Merge pull request #3448 from lubinszARM:pr_tls_testsgVisor bot
PiperOrigin-RevId: 324127810
2020-07-30Merge pull request #3028 from lubinszARM:pr_kvm_hello1gVisor bot
PiperOrigin-RevId: 324125938
2020-07-30Merge pull request #3179 from jinmouil:fuse_initgVisor bot
PiperOrigin-RevId: 324100220
2020-07-30Revert change to default buffer size.Bhasker Hariharan
In https://github.com/google/gvisor/commit/ca6bded95dbce07f9683904b4b768dfc2d4a09b2 we reduced the default buffer size to 32KB. This mostly works fine except at high throughput where we hit zero window very quickly and the TCP receive buffer moderation is not able to grow the window. This can be seen in the benchmarks where with a 32KB buffer and 100 connections downloading a 10MB file we get about 30 requests/s vs the 1MB buffer gives us about 53 requests/s. A proper fix requires a few changes to when we send a zero window as well as when we decide to send a zero window update. Today we consider available space below 1MSS as zero and send an update when it crosses 1MSS of available space. This is way too low and results in the window staying very small once we hit a zero window condition as we keep sending updates with size barely over 1MSS. Linux and BSD are smarter about this and use different thresholds. We should separately update our logic to match linux or BSD so that we don't send window updates that are really tiny or wait until we drop below 1MSS to advertise a zero window. PiperOrigin-RevId: 324087019
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-18 20:19:01 +0000