Commit Graph

636 Commits

Author SHA1 Message Date
Adin Scannell dea3cb92f2 build: add nogo for static validation
PiperOrigin-RevId: 257297820
2019-07-09 16:44:06 -07:00
Adin Scannell cceef9d2cf Cleanup straggling syscall dependencies.
PiperOrigin-RevId: 257293198
2019-07-09 16:18:02 -07:00
Nicolas Lacasse 6db3f8d54c Don't mask errors in createAt loop.
The error set in the loop in createAt was being masked
by other errors declared with ":=". This allowed an
ErrResolveViaReadlink error to escape, which can cause
a sentry panic.

Added test case which repros without the fix.

PiperOrigin-RevId: 257061767
2019-07-08 14:57:15 -07:00
Nicolas Lacasse 659bebab8e Don't try to execute a file that is not regular.
PiperOrigin-RevId: 257037608
2019-07-08 12:56:48 -07:00
Ayush Ranjan 8f9b1ca8e7 ext4: disklayout: inode impl.
PiperOrigin-RevId: 257010414
2019-07-08 10:44:11 -07:00
Andrei Vagin 67f2cefce0 Avoid importing platforms from many source files
PiperOrigin-RevId: 256494243
2019-07-03 22:51:26 -07:00
Ian Lewis da57fb9d25 Fix syscall doc for getresgid
PiperOrigin-RevId: 256481284
2019-07-03 20:13:19 -07:00
Neel Natu 9f2f9f0cab futex: compare keys for equality when doing a FUTEX_UNLOCK_PI.
PiperOrigin-RevId: 256453827
2019-07-03 16:01:38 -07:00
Andrei Vagin 116cac053e netstack/udp: connect with the AF_UNSPEC address family means disconnect
PiperOrigin-RevId: 256433283
2019-07-03 14:19:02 -07:00
gVisor bot f10862696c Merge pull request #493 from ahmetb:reticulating-splines
PiperOrigin-RevId: 256319059
2019-07-03 01:10:34 -07:00
Yong He 85b27a9f8f Solve BounceToKernel may hang issue
BounceToKernel will make vCPU quit from guest ring3 to guest ring0, but
vCPUWaiter is not cleared when we unlock the vCPU, when next time this vCPU
enter guest mode ring3, vCPU may enter guest mode with vCPUWaiter bit setted,
this will cause the following BounceToKernel to this vCPU hangs at
waitUntilNot.

Halt may workaroud this issue, because halt process will reset vCPU status into
vCPUUser, and notify all waiter for vCPU state change, but if there is no
exception or syscall in this period, BounceToKernel will hang at waitUntilNot.

PiperOrigin-RevId: 256299660
2019-07-02 22:03:28 -07:00
Adin Scannell 753da9604e Remove map from fd_map, change to fd_table.
This renames FDMap to FDTable and drops the kernel.FD type, which had an entire
package to itself and didn't serve much use (it was freely cast between types,
and served as more of an annoyance than providing any protection.)

Based on BenchmarkFDLookupAndDecRef-12, we can expect 5-10 ns per lookup
operation, and 10-15 ns per concurrent lookup operation of savings.

This also fixes two tangential usage issues with the FDMap. Namely, non-atomic
use of NewFDFrom and associated calls to Remove (that are both racy and fail to
drop the reference on the underlying file.)

PiperOrigin-RevId: 256285890
2019-07-02 19:28:59 -07:00
Ian Lewis 3f14caeb99 Add documentation for remaining syscalls (fixes #197, #186)
Adds support level documentation for all syscalls. Removes the Undocumented
utility function to discourage usage while leaving SupportUndocumented as the
default support level for Syscall structs.

PiperOrigin-RevId: 256281927
2019-07-02 18:45:16 -07:00
Ayush Ranjan d8ec2fb671 Ext4: DiskLayout: Inode interface.
PiperOrigin-RevId: 256234390
2019-07-02 14:04:31 -07:00
Nicolas Lacasse 4f2f44320f Simplify (and fix) refcounts in createAt.
fileOpAt holds references on the Dirents passed as arguments to the callback,
and drops refs when finished, so we don't need to DecRef those Dirents
ourselves

However, all Dirents that we get from FindInode/FindLink must be DecRef'd.

This CL cleans up the ref-counting logic, and fixes some refcount issues in the
process.

PiperOrigin-RevId: 256220882
2019-07-02 12:58:58 -07:00
Ahmet Alp Balkan 4cd28c6e27
sentry/kernel: add syslog message
It feels like "reticulating splines" is missing from the list of meaningless
syslog messages.

Signed-off-by: Ahmet Alp Balkan <ahmetb@google.com>
2019-07-02 12:05:41 -07:00
Ian Gudger 0aa9418a77 Fix unix/transport.queue reference leaks.
Fix two leaks for connectionless Unix sockets:
* Double connect: Subsequent connects would leak a reference on the previously
  connected endpoint.
* Close unconnected: Sockets which were not connected at the time of closure
  would leak a reference on their receiver.

PiperOrigin-RevId: 256070451
2019-07-01 17:46:24 -07:00
Nicolas Lacasse 06537129a6 Check remaining traversal limit when creating a file through a symlink.
This fixes the case when an app tries to create a file that already exists, and
is a symlink to itself. A test was added.

PiperOrigin-RevId: 256044811
2019-07-01 15:25:22 -07:00
Ian Gudger 45566fa4e4 Add finalizer on AtomicRefCount to check for leaks.
PiperOrigin-RevId: 255711454
2019-06-28 20:07:52 -07:00
Adin Scannell 7dae043fec Drop ashmem and binder.
These are unfortunately unused and unmaintained. They can be brought back in
the future if need requires it.

PiperOrigin-RevId: 255697132
2019-06-28 17:20:25 -07:00
Nicolas Lacasse d3f97aec49 Remove events from name_to_handle_at and open_by_handle_at.
These syscalls require filesystem support that gVisor does not provide, and is
not planning to implement. Their absense should not trigger an event.

PiperOrigin-RevId: 255692871
2019-06-28 16:50:24 -07:00
Ayush Ranjan c4da599e22 ext4: disklayout: SuperBlock interface implementations.
PiperOrigin-RevId: 255687771
2019-06-28 16:18:29 -07:00
Nicolas Lacasse 295078fa7a Automated rollback of changelist 255263686
PiperOrigin-RevId: 255679453
2019-06-28 15:28:41 -07:00
Andrei Vagin e21d49c2d8 platform/ptrace: return more detailed errors
Right now, if we can't create a stub process, we will see this error:
panic: unable to activate mm: resource temporarily unavailable

It would be better to know the root cause of this "resource temporarily
unavailable".

PiperOrigin-RevId: 255656831
2019-06-28 13:23:36 -07:00
Ayush Ranjan 7c13789818 Superblock interface in the disk layout package for ext4.
PiperOrigin-RevId: 255644277
2019-06-28 12:07:28 -07:00
Fabricio Voznika b2907595e5 Complete pipe support on overlayfs
Get/Set pipe size and ioctl support were missing from
overlayfs. It required moving the pipe.Sizer interface
to fs so that overlay could get access.

Fixes #318

PiperOrigin-RevId: 255511125
2019-06-27 17:22:53 -07:00
Michael Pratt 5b41ba5d0e Fix various spelling issues in the documentation
Addresses obvious typos, in the documentation only.

COPYBARA_INTEGRATE_REVIEW=https://github.com/google/gvisor/pull/443 from Pixep:fix/documentation-spelling 4d0688164eafaf0b3010e5f4824b35d1e7176d65
PiperOrigin-RevId: 255477779
2019-06-27 14:25:50 -07:00
Michael Pratt 085a907565 Cache directory entries in the overlay
Currently, the overlay dirCache is only used for a single logical use of
getdents. i.e., it is discard when the FD is closed or seeked back to
the beginning.

But the initial work of getting the directory contents can be quite
expensive (particularly sorting large directories), so we should keep it
as long as possible.

This is very similar to the readdirCache in fs/gofer.

Since the upper filesystem does not have to allow caching readdir
entries, the new CacheReaddir MountSourceOperations method controls this
behavior.

This caching should be trivially movable to all Inodes if desired,
though that adds an additional copy step for non-overlay Inodes.
(Overlay Inodes already do the extra copy).

PiperOrigin-RevId: 255477592
2019-06-27 14:24:03 -07:00
Andrei Vagin e276083903 gvisor/ptrace: grub initial thread registers only once
PiperOrigin-RevId: 255465635
2019-06-27 13:59:57 -07:00
Fabricio Voznika 42e212f6b7 Preserve permissions when checking lower
The code was wrongly assuming that only read access was
required from the lower overlay when checking for permissions.
This allowed non-writable files to be writable in the overlay.

Fixes #316

PiperOrigin-RevId: 255263686
2019-06-26 14:24:44 -07:00
Nicolas Lacasse 857e5c47e9 Follow symlinks when creating a file, and create the target.
If we have a symlink whose target does not exist, creating the symlink (either
via 'creat' or 'open' with O_CREAT flag) should create the target of the
symlink. Previously, gVisor would error with EEXIST in this case

PiperOrigin-RevId: 255232944
2019-06-26 11:49:20 -07:00
Michael Pratt e98ce4a2c6 Add TODO reminder to remove tmpfs caching options
Updates #179

PiperOrigin-RevId: 255081565
2019-06-25 17:12:34 -07:00
Andrei Vagin 03ae91c662 gvisor: lockless read access for task credentials
Credentials are immutable and even before these changes we could read them
without locks, but we needed to take a task lock to get a credential object
from a task object.

It is possible to avoid this lock, if we will guarantee that a credential
object will not be changed after setting it on a task.

PiperOrigin-RevId: 254989492
2019-06-25 09:52:49 -07:00
Andrei Vagin e9ea7230f7 fs: synchronize concurrent writes into files with O_APPEND
For files with O_APPEND, a file write operation gets a file size and uses it as
offset to call an inode write operation. This means that all other operations
which can change a file size should be blocked while the write operation doesn't
complete.

PiperOrigin-RevId: 254873771
2019-06-24 17:45:02 -07:00
Adin Scannell 7f5d0afe52 Add O_EXITKILL to ptrace options.
This prevents a race before PDEATH_SIG can take effect during
a sentry crash.

Discovered and solution by avagin@.

PiperOrigin-RevId: 254871534
2019-06-24 17:30:01 -07:00
Rahat Mahmood 94a6bfab5d Implement /proc/net/tcp.
PiperOrigin-RevId: 254854346
2019-06-24 15:56:36 -07:00
Andrei Vagin c5486f5122 platform/ptrace: specify PTRACE_O_TRACEEXIT for stub-processes
The tracee is stopped early  during  process  exit,  when registers are still
available, allowing the tracer to see where the exit occurred, whereas the
normal exit  notifi? cation  is  done  after  the process is finished exiting.

Without this option, dumpAndPanic fails to get registers.

PiperOrigin-RevId: 254852917
2019-06-24 15:48:58 -07:00
Nicolas Lacasse 87df9aab24 Use correct statx syscall number for amd64.
The previous number was for the arm architecture.

Also change the statx tests to force them to run on gVisor, which would have
caught this issue.

PiperOrigin-RevId: 254846831
2019-06-24 15:19:36 -07:00
Fabricio Voznika b21b1db700 Allow to change logging options using 'runsc debug'
New options are:
  runsc debug --strace=off|all|function1,function2
  runsc debug --log-level=warning|info|debug
  runsc debug --log-packets=true|false

Updates #407

PiperOrigin-RevId: 254843128
2019-06-24 15:03:02 -07:00
Nicolas Lacasse 35719d52c7 Implement statx.
We don't have the plumbing for btime yet, so that field is left off. The
returned mask indicates that btime is absent.

Fixes #343

PiperOrigin-RevId: 254575752
2019-06-22 13:29:26 -07:00
Andrei Vagin ab6774cebf gvisor/fs: getdents returns 0 if offset is equal to FileMaxOffset
FileMaxOffset is a special case when lseek(d, 0, SEEK_END) has been called.

PiperOrigin-RevId: 254498777
2019-06-21 17:25:17 -07:00
Ayush Ranjan 727375321f ext4 block group descriptor implementation in disk layout package.
PiperOrigin-RevId: 254482180
2019-06-21 15:42:46 -07:00
Fabricio Voznika 5ba16d51a9 Add list of stuck tasks to panic message
PiperOrigin-RevId: 254450309
2019-06-21 12:46:53 -07:00
Andrei Vagin f94653b3de kernel: call t.mu.Unlock() explicitly in WithMuLocked
defer here doesn't improve readability, but we know it slower that
the explicit call.

PiperOrigin-RevId: 254441473
2019-06-21 11:55:42 -07:00
Fabricio Voznika 054b5632ef Update comment
PiperOrigin-RevId: 254428866
2019-06-21 10:56:42 -07:00
Jamie Liu 7db8685100 Preallocate auth.NewAnonymousCredentials() in contexttest.TestContext.
Otherwise every call to, say, fs.ContextCanAccessFile() in a benchmark
using contexttest allocates new auth.Credentials, a new
auth.UserNamespace, ...

PiperOrigin-RevId: 254261051
2019-06-20 13:36:14 -07:00
Michael Pratt 292f70cbf7 Add package docs to seqfile and ramfs
These are the only packages missing docs:
https://godoc.org/gvisor.dev/gvisor

PiperOrigin-RevId: 254261022
2019-06-20 13:34:33 -07:00
Neel Natu 0b2135072d Implement madvise(MADV_DONTFORK)
PiperOrigin-RevId: 254253777
2019-06-20 12:56:00 -07:00
Ian Gudger 7e49515696 Deflake SendFileTest_Shutdown.
The sendfile syscall's backing doSplice contained a race with regard to
blocking. If the first attempt failed with syserror.ErrWouldBlock and then
the blocking file became ready before registering a waiter, we would just
return the ErrWouldBlock (even if we were supposed to block).

PiperOrigin-RevId: 254114432
2019-06-19 18:40:54 -07:00
Nicolas Lacasse 29f9e4fa87 fileOp{On,At} should pass the remaning symlink traversal count.
And methods that do more traversals should use the remaining count rather than
resetting.

PiperOrigin-RevId: 254041720
2019-06-19 11:56:34 -07:00