This provides the following benefits:
- We can now use pkg/fd package which does not take ownership
of the file descriptor. So it does not close the fd when garbage collected.
This reduces scope of errors from unexpected garbage collection of io.File.
- It enforces the offset parameter in every read call.
It does not affect the fd offset nor is it affected by it. Hence reducing
scope of error of using stale offsets when reading.
- We do not need to serialize the usage of any global file descriptor anymore.
So this drops the mutual exclusion req hence reducing complexity and
congestion.
PiperOrigin-RevId: 260635174
Allocate a larger memory buffer and combine multiple copies into one copy,
to reduce the number of copies from kernel memory to user memory.
Signed-off-by: Hang Su <darcy.sh@antfin.com>
This introduces two new types of Emitters:
1. MultiEmitter, which will forward events to other registered Emitters, and
2. RateLimitedEmitter, which will forward events to a wrapped Emitter, subject
to given rate limits.
The methods in the eventchannel package itself act like a multiEmitter, but is
not actually an Emitter. Now we have a DefaultEmitter, and the methods in
eventchannel simply forward calls to the DefaultEmitter.
The unimplemented syscall handler now uses a RateLimetedEmitter that wraps the
DefaultEmitter.
PiperOrigin-RevId: 260612770
This allows the user code to add a network address with a subnet prefix length.
The prefix length value is stored in the network endpoint and provided back to
the user in the ProtocolAddress type.
PiperOrigin-RevId: 259807693
The different containers in a sandbox used only one pid
namespace before. This results in that a container can see
the processes in another container in the same sandbox.
This patch use different pid namespace for different containers.
Signed-off-by: chris.zn <chris.zn@antfin.com>
This keeps all container filesystem completely separate from eachother
(including from the root container filesystem), and allows us to get rid of the
"__runsc_containers__" directory.
It also simplifies container startup/teardown as we don't have to muck around
in the root container's filesystem.
PiperOrigin-RevId: 259613346
m.available.Signal() has to be called under m.mu.RLock, otherwise it can
race with machine.Get:
m.Get | m.Put
-------------------------------------
m.mu.Lock() |
Seatching available vcpu|
| m.available.Signal()
m.available.Wait |
PiperOrigin-RevId: 259394051
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
This proc file reports the stats of interfaces. We could use ifconfig
command to check the result.
Signed-off-by: Jianfeng Tan <henry.tjf@antfin.com>
Change-Id: Ia7c1e637f5c76c30791ffda68ee61e861b6ef827
COPYBARA_INTEGRATE_REVIEW=https://gvisor-review.googlesource.com/c/gvisor/+/18282/
PiperOrigin-RevId: 258303936
Now we call FUTEX_WAKE with ^uintptr(0) of waiters, but in this case only one
waiter will be waked up. If we want to wake up all of them, the number of
waiters has to be set to math.MaxInt32.
PiperOrigin-RevId: 258285286
iptables also relies on IPPROTO_RAW in a way. It opens such a socket to
manipulate the kernel's tables, but it doesn't actually use any of the
functionality. Blegh.
PiperOrigin-RevId: 257903078
Adds support to set/get the TCP_MAXSEG value but does not
really change the segment sizes emitted by netstack or
alter the MSS advertised by the endpoint. This is currently
being added only to unblock iperf3 on gVisor. Plumbing
this correctly requires a bit more work which will come
in separate CLs.
PiperOrigin-RevId: 257859112
The image is of size 64Kb which supports 64 1k blocks
and 16 inodes. This is the smallest size mkfs.ext4 works with.
Added README.md documenting how this was created and included
all files on the device under assets.
PiperOrigin-RevId: 257712672
Renamed ext4 to ext since we are targeting ext(2/3/4).
Removed fs.go since we are targeting VFS2.
Added ext.go with filesystem struct.
PiperOrigin-RevId: 257689775
A userspace process (CPL=3) can access an i/o port if the bit corresponding to
the port is set to 0 in the I/O permission bitmap.
Configure the I/O permission bitmap address beyond the last valid byte in the
TSS so access to all i/o ports is blocked.
Signed-off-by: Liu Hua <sdu.liu@huawei.com>
Change-Id: I3df76980c3735491db768f7210e71703f86bb989
PiperOrigin-RevId: 257336518
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
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
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
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
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
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
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
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
Readdir of /proc/x/task/ will get direntry entries
from tasks of specified taskgroup. Now the tasks
slice is unsorted, use sort.SearchInts search entry
from the slice may cause infinity loops.
The fix is sort the slice before search.
This issue could be easily reproduced via following
steps, revise Readdir in pkg/sentry/fs/proc/task.go,
force set taskInts into test slice
[]int{1, 11, 7, 5, 10, 6, 8, 3, 9, 2, 4},
then run docker image and run ls /proc/1/task, the
command will cause infinity loops.
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
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
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
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
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
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
Makes CLOCK_BOOTTIME available with
* clock_gettime
* timerfd_create
* clock_gettime vDSO
CLOCK_BOOTTIME is implemented as an alias to CLOCK_MONOTONIC.
CLOCK_MONOTONIC already keeps track of time across save
and restore. This is the closest possible behavior to Linux
CLOCK_BOOTIME, as there is no concept of suspend/resume.
Updates google/gvisor#218
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
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
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
There will be a deadloop when we use getdents to read /proc/{pid}/task
of an exited process
Like this:
Process A is running
Process B: open /proc/{pid of A}/task
Process A exits
Process B: getdents /proc/{pid of A}/task
Then, process B will fall into deadloop, and return "." and ".."
in loops and never ends.
This patch returns ENOENT when use getdents to read /proc/{pid}/task
if the process is just exited.
Signed-off-by: chris.zn <chris.zn@antfin.com>
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
Otherwise every call to, say, fs.ContextCanAccessFile() in a benchmark
using contexttest allocates new auth.Credentials, a new
auth.UserNamespace, ...
PiperOrigin-RevId: 254261051
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
This allows tasks to have distinct mount namespace, instead of all sharing the
kernel's root mount namespace.
Currently, the only way for a task to get a different mount namespace than the
kernel's root is by explicitly setting a different MountNamespace in
CreateProcessArgs, and nothing does this (yet).
In a follow-up CL, we will set CreateProcessArgs.MountNamespace when creating a
new container inside runsc.
Note that "MountNamespace" is a poor term for this thing. It's more like a
distinct VFS tree. When we get around to adding real mount namespaces, this
will need a better naem.
PiperOrigin-RevId: 254009310
Test fails because it's reading 4KB instead of the
expected 64KB. Changed the test to read pipe buffer
size instead of hardcode and added some logging in
case the reason for failure was not pipe buffer size.
PiperOrigin-RevId: 253916040