Age | Commit message (Collapse) | Author |
|
Update RegistryImpl functions to return file descriptions, instead of
queues, and use Views in queue inodes.
Updates #136
|
|
View makes it easier to handle O_RDONLY, O_WRONLY, and ORDWR options in
mq_open(2).
Updates #136
|
|
Updates #136
|
|
Move root dentry and filesystem creation from GetFilesystem to
NewRegistryImpl, create IPCNamespace.InitPosixQueues to create a new
mqueue filesystem for each ipc namespace, and update GetFilesystem to
retreive fs and root dentry from IPCNamespace and return them.
Updates #136
|
|
CtxIPCNamespace is needed by mqfs package to be able to retreive an
IPCNamespace using ctx.Value. As ctx.Value compares keys as interfaces,
we need to use type kernel.contextID in package mqfs, which is not
possible due to circular depenedency, so move it to kernel/ipc instead.
Updates #136
|
|
Define a POSIX message queue Registry and RegistryImpl in mq package,
implement RegistryImpl in mqfs, and add a Registry object to
IPCNamespace initialized at filesystem creation.
Updates #136
|
|
Implement inode and file description representing a POSIX message queue,
and other utilities needed to implement file operations.
Updates #136
|
|
Create package mq to implement POSIX message queues, and define initial
struct definitions.
Updates #136
|
|
Updates #136
|
|
rootInode represents the root inode for mqueue filesystem (/dev/mqueue).
Updates #136
|
|
...and EAFNOSUPPORT for unexpected address family.
To comply with Linux.
Updates #6021, #6575.
PiperOrigin-RevId: 396893590
|
|
Previously, we weren't making a copy when a sysv message queue was
receiving a message with the MSG_COPY flag. This flag indicates the
message being received should be left in the queue and a copy of the
message should be returned to userspace. Without the copy, a racing
process can modify the original message while it's being marshalled to
user memory.
Reported-by: syzbot+cb15e644698b20ff4e17@syzkaller.appspotmail.com
PiperOrigin-RevId: 396712856
|
|
PiperOrigin-RevId: 396042572
|
|
Previously, gVisor did not represent loopback devices as an ethernet
device as Linux does. To maintain Linux API compatibility for packet
sockets, a workaround was used to add an ethernet header if a link
header was not already present in the packet buffer delivered to a
packet endpoint.
However, this workaround is a bug for non-ethernet based interfaces; not
all links use an ethernet header (e.g. pure L3/TUN interfaces).
As of 3b4bb947517d0d9010120aaa1c3989fd6abf278e, gVisor represents
loopback devices as an ethernet-based device so this workaround can
now be removed.
BUG: https://fxbug.dev/81592
Updates #6530, #6531.
PiperOrigin-RevId: 395819151
|
|
PiperOrigin-RevId: 395338926
|
|
This defines common infrastructure for dynamically-configured security checks,
including an example usage in the clone(2) path.
PiperOrigin-RevId: 394797270
|
|
If we want to take two endpoint locks, we need to be sure that we always
take them in the same order.
Accept() locks the listening endpoint to work with acceptedChan and then
it calls GetLocalAddress that locks an accepted endpoint. Actually, we
can release the listening endpoint lock before calling GetLocalAddress.
Reported-by: syzbot+f52bd603f51a4ae91054@syzkaller.appspotmail.com
PiperOrigin-RevId: 394553823
|
|
...through the loopback interface, only.
This change only supports sending on packet sockets through the loopback
interface as the loopback interface is the only interface used in packet
socket syscall tests - the other link endpoints are not excercised with
the existing test infrastructure.
Support for sending on packet sockets through the other interfaces will
be added as needed.
BUG: https://fxbug.dev/81592
PiperOrigin-RevId: 394368899
|
|
...from the UDP endpoint.
Datagram-based transport endpoints (e.g. UDP, RAW IP) can share a lot
of their write path due to the datagram-based nature of these endpoints.
Extract the common facilities from UDP so they can be shared with other
transport endpoints (in a later change).
Test: UDP syscall tests.
PiperOrigin-RevId: 394347774
|
|
reflect.ValueOf takes an interface{}, so when passed a slice the compiler emits
a call to runtime.convTslice to heap-allocate a copy of the slice header.
PiperOrigin-RevId: 394310052
|
|
PiperOrigin-RevId: 394300607
|
|
PiperOrigin-RevId: 394296687
|
|
Reported-by: syzbot+1aab6800bd14829609b8@syzkaller.appspotmail.com
PiperOrigin-RevId: 394279838
|
|
PiperOrigin-RevId: 393841270
|
|
PiperOrigin-RevId: 393831108
|
|
Document this ordering in mm/mm.go.
PiperOrigin-RevId: 393413203
|
|
PiperOrigin-RevId: 392982220
|
|
PiperOrigin-RevId: 392774712
|
|
PiperOrigin-RevId: 392554743
|
|
Right now, the first slot starts with an address of a memory region and its size is faultBlockSize,
but the second slot starts with (physicalStart + faultBlockSize) & faultBlockMask.
It means they will overlap if a start address of a memory region are not aligned to faultBlockSize.
The kernel doesn't allow to add overlapped regions, but we ignore the EEXIST error.
Signed-off-by: Andrei Vagin <avagin@google.com>
|
|
PiperOrigin-RevId: 392102898
|
|
We cannot hold mm.aioManager.mu while calling MUnmap, because MUnmap attempts
to aquire mm.mappingMu. This violates the lock order as documented in mm/mm.go.
PiperOrigin-RevId: 392102472
|
|
PiperOrigin-RevId: 392078690
|
|
Add an LRU cache to cache verity dentries when ref count drop to 0. This
way we don't need to hash and verify the previous opened files or
directories each time.
PiperOrigin-RevId: 391880157
|
|
The rationale given for using buffered copies is still valid, but it's unclear
whether holding MM locks or allocating buffers is better in practice, and the
former is at least consistent with gofer.regularFileFD (and VFS1), making
performance easier to reason about.
PiperOrigin-RevId: 391877913
|
|
Also plumber the controls through runsc
PiperOrigin-RevId: 391594318
|
|
PiperOrigin-RevId: 391416650
|
|
|
|
Add support for msgctl and enable tests.
Fixes #135
|
|
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
|
|
Set provides functionality of {sem,shm,msg}ctl(IPC_SET).
|
|
Changed sendSignal to sendSignalLocked because tg.pidns.owner.mu and tg.signalHandlers.mu
are already locked in SetForegroundProcess
Added a control to verify whether the calling process is ignoring SIGTTOU
before sending the signal
|
|
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
|
|
Add Event controls and implement "stream" commands.
PiperOrigin-RevId: 390691702
|
|
Add Usage controls and implement "usage/usagefd" commands.
PiperOrigin-RevId: 390507423
|
|
Convert remaining public errors (e.g. EINTR) from syserror to linuxerr.
PiperOrigin-RevId: 390471763
|
|
Send buffer size in TCP indicates the amount of bytes available for the sender
to transmit. This change will allow TCP to update the send buffer size when
- TCP enters established state.
- ACK is received.
The auto tuning is disabled when the send buffer size is set with the
SO_SNDBUF option.
PiperOrigin-RevId: 390312274
|
|
The dentry for each file/directory can be created/destroyed multiple
times during sandbox lifetime. We should not clear the Merkle file each
time a dentry is created.
PiperOrigin-RevId: 390277107
|
|
We were relying on children adding its name to parent's dentry to
populate parent's children list. However, this may not work since the
parent dentry could be destroyed if its reference count drops to zero.
In that case, a new dentry will be created when enabling the parent and
it does not contain the children names info. Therefore we need to
populate the child names list again to avoid missing children in the
directory.
PiperOrigin-RevId: 390270227
|
|
Allow creation and management of subcontainers through cgroupfs
directory syscalls. Also add a mechanism to specify a default root
container to start new jobs in.
This implements the filesystem support for subcontainers, but doesn't
implement hierarchical resource accounting or task migration.
PiperOrigin-RevId: 390254870
|