summaryrefslogtreecommitdiffhomepage
path: root/pkg/sentry/vfs/vfs.go
AgeCommit message (Collapse)Author
2021-02-19Don't hold baseEndpoint.mu while calling EventUpdate().Nicolas Lacasse
This removes a three-lock deadlock between fdnotifier.notifier.mu, epoll.EventPoll.listsMu, and baseEndpoint.mu. A lock order comment was added to epoll/epoll.go. Also fix unsafe access of baseEndpoint.connected/receiver. PiperOrigin-RevId: 358515191
2021-02-11Internal change.gVisor bot
PiperOrigin-RevId: 357090170
2021-01-28Add O_PATH support in vfs2gVisor bot
PiperOrigin-RevId: 354367665
2020-11-18Port filesystem metrics to VFS2.Jamie Liu
PiperOrigin-RevId: 343196927
2020-10-23Support VFS2 save/restore.Jamie Liu
Inode number consistency checks are now skipped in save/restore tests for reasons described in greatest detail in StatTest.StateDoesntChangeAfterRename. They pass in VFS1 due to the bug described in new test case SimpleStatTest.DifferentFilesHaveDifferentDeviceInodeNumberPairs. Fixes #1663 PiperOrigin-RevId: 338776148
2020-10-13[vfs2] Add FilesystemType.Release to avoid reference leaks.Dean Deng
Singleton filesystem like devpts and devtmpfs have a single filesystem shared among all mounts, so they acquire a "self-reference" when initialized that must be released when the entire virtual filesystem is released at sandbox exit. PiperOrigin-RevId: 336828852
2020-10-12[vfs2] Don't leak disconnected mounts.Dean Deng
PiperOrigin-RevId: 336694658
2020-09-24Add basic stateify annotations.Adin Scannell
Updates #1663 PiperOrigin-RevId: 333539293
2020-09-09Unlock VFS.mountMu before FilesystemImpl calls for ↵Jamie Liu
/proc/[pid]/{mounts,mountinfo}. Also move VFS.MakeSyntheticMountpoint() (which is a utility wrapper around VFS.MkdirAllAt(), itself a utility wrapper around VFS.MkdirAt()) to not be in the middle of the implementation of these proc files. Fixes #3878 PiperOrigin-RevId: 330843106
2020-09-08[vfs] overlayfs: Fix socket tests.Ayush Ranjan
- BindSocketThenOpen test was expecting the incorrect error when opening a socket. Fixed that. - VirtualFilesystem.BindEndpointAt should not require pop.Path.Begin.Ok() because the filesystem implementations do not need to walk to the parent dentry. This check also exists for MknodAt, MkdirAt, RmdirAt, SymlinkAt and UnlinkAt but those filesystem implementations also need to walk to the parent denty. So that check is valid. Added some syscall tests to test this. PiperOrigin-RevId: 330625220
2020-09-08[vfs] Capitalize x in the {Get/Set/Remove/List}xattr functions.Ayush Ranjan
PiperOrigin-RevId: 330554450
2020-08-20[vfs] Create recursive dir creation util.Ayush Ranjan
Refactored the recursive dir creation util in runsc/boot/vfs.go to be more flexible. PiperOrigin-RevId: 327719100
2020-08-20Fix tabs in lock-ordering doc.Nicolas Lacasse
PiperOrigin-RevId: 327654207
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
2020-07-08Add shared mount hints to VFS2Fabricio Voznika
Container restart test is disabled for VFS2 for now. Updates #1487 PiperOrigin-RevId: 320296401
2020-06-23Support inotify in vfs2 gofer fs.Dean Deng
Because there is no inode structure stored in the sandbox, inotify watches must be held on the dentry. This would be an issue in the presence of hard links, where multiple dentries would need to share the same set of watches, but in VFS2, we do not support the internal creation of hard links on gofer fs. As a result, we make the assumption that every dentry corresponds to a unique inode. Furthermore, dentries can be cached and then evicted, even if the underlying file has not be deleted. We must prevent this from occurring if there are any watches that would be lost. Note that if the dentry was deleted or invalidated (d.vfsd.IsDead()), we should still destroy it along with its watches. Additionally, when a dentry’s last watch is removed, we cache it if it also has zero references. This way, the dentry can eventually be evicted from memory if it is no longer needed. This is accomplished with a new dentry method, OnZeroWatches(), which is called by Inotify.RmWatch and Inotify.Release. Note that it must be called after all inotify locks are released to avoid violating lock order. Stress tests are added to make sure that inotify operations don't deadlock with gofer.OnZeroWatches. Updates #1479. PiperOrigin-RevId: 317958034
2020-06-05Implement mount(2) and umount2(2) for VFS2.Rahat Mahmood
This is mostly syscall plumbing, VFS2 already implements the internals of mounts. In addition to the syscall defintions, the following mount-related mechanisms are updated: - Implement MS_NOATIME for VFS2, but only for tmpfs and goferfs. The other VFS2 filesystems don't implement node-level timestamps yet. - Implement the 'mode', 'uid' and 'gid' mount options for VFS2's tmpfs. - Plumb mount namespace ownership, which is necessary for checking appropriate capabilities during mount(2). Updates #1035 PiperOrigin-RevId: 315035352
2020-05-29Implement IN_EXCL_UNLINK inotify option in vfs2.Dean Deng
Limited to tmpfs. Inotify support in other filesystem implementations to follow. Updates #1479 PiperOrigin-RevId: 313828648
2020-05-29Port inotify to vfs2, with support in tmpfs.Dean Deng
Support in other filesystem impls is still needed. Unlike in Linux and vfs1, we need to plumb inotify down to each filesystem implementation in order to keep track of links/inode structures properly. IN_EXCL_UNLINK still needs to be implemented, as well as a few inotify hooks that are not present in either vfs1 or vfs2. Those will be addressed in subsequent changes. Updates #1479. PiperOrigin-RevId: 313781995
2020-05-28Split VFS.MountAt into VFS.MountDisconnected and VFS.ConnectMountAt.Jamie Liu
This makes it straightforward to create bind mounts internally in VFS2: Given a bind mount root represented by vfs.VirtualDentry vd: - Create a new mount with VFS.NewDisconnectedMount(vd.Mount().Filesystem(), vd.Dentry()). - Connect the resulting mount in the appropriate namespace with VFS.ConnectMountAt(). Note that the resulting bind mount is non-recursive; recursive bind mounting requires explicitly duplicating all children of the original mount, which is best handled internally by VFS. Updates #179 PiperOrigin-RevId: 313703963
2020-04-28Support pipes and sockets in VFS2 gofer fs.Dean Deng
Named pipes and sockets can be represented in two ways in gofer fs: 1. As a file on the remote filesystem. In this case, all file operations are passed through 9p. 2. As a synthetic file that is internal to the sandbox. In this case, the dentry stores an endpoint or VFSPipe for sockets and pipes respectively, which replaces interactions with the remote fs through the gofer. In gofer.filesystem.MknodAt, we attempt to call mknod(2) through 9p, and if it fails, fall back to the synthetic version. Updates #1200. PiperOrigin-RevId: 308828161
2020-04-24Fixes for procfsFabricio Voznika
- Return ENOENT for /proc/[pid]/task if task is zoombied or terminated - Allow directory to be Seek() to the end - Construct synthetic files for /proc/[pid]/ns/* - Changed GenericDirectoryFD.Init to not register with FileDescription, otherwise other implementation cannot change behavior. Updates #1195,1193 PiperOrigin-RevId: 308294649
2020-04-16Implement pipe(2) and pipe2(2) for VFS2.Jamie Liu
Updates #1035 PiperOrigin-RevId: 306968644
2020-04-14Pass O_LARGEFILE in syscalls/linux/vfs2.openat.Jamie Liu
Needed for PipeTest_Flags: files opened by open() and openat() get O_LARGEFILE (on architectures with 64-bit off_t), but not FDs created by other syscalls such as pipe(). Updates #1035 PiperOrigin-RevId: 306504788
2020-04-10Port extended attributes to VFS2.Dean Deng
As in VFS1, we only support the user.* namespace. Plumbing is added to tmpfs and goferfs. Note that because of the slightly different order of checks between VFS2 and Linux, one of the xattr tests needs to be relaxed slightly. Fixes #2363. PiperOrigin-RevId: 305985121
2020-03-27Add FilesystemType.Name method, and FilesystemType field to Filesystem struct.Nicolas Lacasse
Both have analogues in Linux: * struct file_system_type has a char *name field. * struct super_block keeps a pointer to the file_system_type. These fields are necessary to support the `filesystem type` field in /proc/[pid]/mountinfo. PiperOrigin-RevId: 303434063
2020-03-26Add BoundEndpointAt filesystem operation.Dean Deng
BoundEndpointAt() is needed to support Unix sockets bound at a file path, corresponding to BoundEndpoint() in VFS1. Updates #1476. PiperOrigin-RevId: 303258251
2020-03-25Remove TODO to push down exec permission checkFabricio Voznika
Pushing it down requires all implementation to check for exec individualy which is not maintanable. Making it part of GenericCheckPermissions add extra cost to everyone that calls it. So it's better to keep is in VirtualFilesystem.OpenAt. Updates #1193 PiperOrigin-RevId: 302982993
2020-03-13Implement access/faccessat for VFS2.Dean Deng
Note that the raw faccessat system call does not actually take a flags argument; according to faccessat(2), the glibc wrapper implements the flags by using fstatat(2). Remove the flag argument that we try to extract from vfs1, which would just be a garbage value. Updates #1965 Fixes #2101 PiperOrigin-RevId: 300796067
2020-03-13Add support for mount flagsFabricio Voznika
Plumbs MS_NOEXEC and MS_RDONLY. Others are TODO. Updates #1623 #1193 PiperOrigin-RevId: 300764669
2020-02-25Add VFS.NewDisconnectedMount().Jamie Liu
Analogous to Linux's kern_mount(). PiperOrigin-RevId: 297259580
2020-02-25Port most syscalls to VFS2.Jamie Liu
pipe and pipe2 aren't ported, pending a slight rework of pipe FDs for VFS2. mount and umount2 aren't ported out of temporary laziness. access and faccessat need additional FSImpl methods to implement properly, but are stubbed to prevent googletest from CHECK-failing. Other syscalls require additional plumbing. Updates #1623 PiperOrigin-RevId: 297188448
2020-02-14Inline vfs.VirtualFilesystem in Kernel structgVisor bot
This saves one pointer dereference per VFS access. Updates #1623 PiperOrigin-RevId: 295216176
2020-02-14Plumb VFS2 inside the SentrygVisor bot
- Added fsbridge package with interface that can be used to open and read from VFS1 and VFS2 files. - Converted ELF loader to use fsbridge - Added VFS2 types to FSContext - Added vfs.MountNamespace to ThreadGroup Updates #1623 PiperOrigin-RevId: 295183950
2020-02-13Add FileExec flag to OpenOptionsgVisor bot
This allow callers to say whether the file is being opened to be executed, so that the proper checks can be done from FilesystemImpl.OpenAt() Updates #1623 PiperOrigin-RevId: 295042595
2020-01-28Add VFS2 support for epoll.Jamie Liu
PiperOrigin-RevId: 291997879
2020-01-28Implement an anon_inode equivalent for VFS2.Jamie Liu
PiperOrigin-RevId: 291986033
2020-01-27Update package locations.Adin Scannell
Because the abi will depend on the core types for marshalling (usermem, context, safemem, safecopy), these need to be flattened from the sentry directory. These packages contain no sentry-specific details. PiperOrigin-RevId: 291811289
2020-01-24Add anonymous device number allocation to VFS2.Jamie Liu
Note that in VFS2, filesystem device numbers are per-vfs.FilesystemImpl rather than global, avoiding the need for a "registry" type to handle save/restore. (This is more consistent with Linux anyway: compare e.g. mm/shmem.c:shmem_mount() => fs/super.c:mount_nodev() => (indirectly) set_anon_super().) PiperOrigin-RevId: 291425193
2020-01-09New sync package.Ian Gudger
* Rename syncutil to sync. * Add aliases to sync types. * Replace existing usage of standard library sync package. This will make it easier to swap out synchronization primitives. For example, this will allow us to use primitives from github.com/sasha-s/go-deadlock to check for lock ordering violations. Updates #1472 PiperOrigin-RevId: 289033387
2019-12-30Add VFS2 support for device special files.Jamie Liu
- Add FileDescriptionOptions.UseDentryMetadata, which reduces the amount of boilerplate needed for device FDs and the like between filesystems. - Switch back to having FileDescription.Init() take references on the Mount and Dentry; otherwise managing refcounts around failed calls to OpenDeviceSpecialFile() / Device.Open() is tricky. PiperOrigin-RevId: 287575574
2019-12-27Add VFS2 support for /proc/filesystems.Jamie Liu
Updates #1195 PiperOrigin-RevId: 287269106
2019-12-23Clean up vfs.FilesystemImpl methods that operate on parent directories.Jamie Liu
- Make FilesystemImpl methods that operate on parent directories require !rp.Done() (i.e. there is at least one path component to resolve) as precondition and postcondition (in cases where they do not finish path resolution due to mount boundary / absolute symlink), and require that they do not need to follow the last path component (the file being created / deleted) as a symlink. Check for these in VFS. - Add FilesystemImpl.GetParentDentryAt(), which is required to obtain the old parent directory for VFS.RenameAt(). (Passing the Dentry to be renamed instead has the wrong semantics if the file named by the old path is a mount point since the Dentry will be on the wrong Mount.) - Update memfs to implement these methods correctly (?), including RenameAt. - Change fspath.Parse() to allow empty paths (to simplify implementation of AT_EMPTY_PATH). - Change vfs.PathOperation to take a fspath.Path instead of a raw pathname; non-test callers will need to fspath.Parse() pathnames themselves anyway in order to detect absolute paths and select PathOperation.Start accordingly. PiperOrigin-RevId: 286934941
2019-12-18Add VFS2 plumbing for extended attributes.Jamie Liu
PiperOrigin-RevId: 286281274
2019-12-10Add most VFS methods for syscalls.Jamie Liu
PiperOrigin-RevId: 284892289
2019-11-25Minor VFS2 interface changes.Jamie Liu
- Remove the Filesystem argument from DentryImpl.*Ref(); in general DentryImpls that need the Filesystem for reference counting will probably also need it for other interface methods that don't plumb Filesystem, so it's easier to just store a pointer to the filesystem in the DentryImpl. - Add a pointer to the VirtualFilesystem to Filesystem, which is needed by the gofer client to disown dentries for cache eviction triggered by dentry reference count changes. - Rename FilesystemType.NewFilesystem to GetFilesystem; in some cases (e.g. sysfs, cgroupfs) it's much cleaner for there to be only one Filesystem that is used by all mounts, and in at least one case (devtmpfs) it's visibly incorrect not to do so, so NewFilesystem doesn't always actually create and return a *new* Filesystem. - Require callers of FileDescription.Init() to increment Mount/Dentry references. This is because the gofer client may, in the OpenAt() path, take a reference on a dentry with 0 references, which is safe due to synchronization that is outside the scope of this CL, and it would be safer to still have its implementation of DentryImpl.IncRef() check for an increment for 0 references in other cases. - Add FileDescription.TryIncRef. This is used by the gofer client to take references on "special file descriptions" (FDs for files such as pipes, sockets, and devices), which use per-FD handles (fids) instead of dentry-shared handles, for sync() and syncfs(). PiperOrigin-RevId: 282473364
2019-11-25Implement basic umounting for vfs2.Jamie Liu
This is required to test filesystems with a non-trivial implementation of FilesystemImpl.Release(). Propagation isn't handled yet, and umount isn't yet plumbed out to VirtualFilesystem.UmountAt(), but otherwise the implementation of umount is believed to be correct. - Move entering mountTable.seq writer critical sections to callers of mountTable.{insert,remove}Seqed. This is required since umount(2) must ensure that no new references are taken on the candidate mount after checking that it isn't busy, which is only possible by entering a vfs.mountTable.seq writer critical section before the check and remaining in it until after VFS.umountRecursiveLocked() is complete. (Linux does the same thing: fs/namespace.c:do_umount() => lock_mount_hash(), fs/pnode.c:propagate_mount_busy(), umount_tree(), unlock_mount_hash().) - It's not possible for dentry deletion to umount while only holding VFS.mountMu for reading, but it's also very unappealing to hold VFS.mountMu exclusively around e.g. gofer unlink RPCs. Introduce dentry.mu to avoid these problems. This means that VFS.mountMu is never acquired for reading, so change it to a sync.Mutex. PiperOrigin-RevId: 282444343
2019-07-18Sentry virtual filesystem, v2Jamie Liu
Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-25 02:39:58 +0000