Files
serenity/Kernel/FileSystem/OpenFileDescription.cpp
T
Liav A. 2a4a096e0f Kernel+runc: Make unshare syscalls more fd-oriented
Instead of creating a new resource that has its own ID number and work
with it directly, we can create a file that describes the unshared
resource, execute ioctl calls on it and only enter into it in the end,
essentially creating the resource only during the last call instead
of the previous method of creation of a resource when "attaching" to
that resource.

We can enter a resource for current program execution, after the exec
syscall, or both.
That change allows userspace to create a resource and attach to it only
in the new program, which makes it more comfortable to do cleanups or
track the new process, outside of the created container.

It should be noted that until this commit, we entered a resource without
detaching the old one, essentially leaking the attach counter of a
resource. While this bug didn't have severe effects, it was obvious that
a proper cleanup userspace code later on wouldn't work in that situation
anyway, so this commit changes the way we work, and the terminology of
entering a resource is actually to **replace** it.

These changes essentially open an opportunity to extend runc to be a
container manager rather being launcher of a containerized environment,
which makes it possible to do all sorts of nice cleanups and tracking of
containers' states.
2026-03-14 11:45:37 +01:00

577 lines
16 KiB
C++

/*
* Copyright (c) 2018-2022, Andreas Kling <[email protected]>
* Copyright (c) 2021, sin-ack <[email protected]>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#include "UnsharedResourceFile.h"
#include <AK/MemoryStream.h>
#include <Kernel/API/POSIX/errno.h>
#include <Kernel/Devices/BlockDevice.h>
#include <Kernel/Devices/TTY/MasterPTY.h>
#include <Kernel/Devices/TTY/TTY.h>
#include <Kernel/FileSystem/Custody.h>
#include <Kernel/FileSystem/FIFO.h>
#include <Kernel/FileSystem/InodeFile.h>
#include <Kernel/FileSystem/InodeWatcher.h>
#include <Kernel/FileSystem/MountFile.h>
#include <Kernel/FileSystem/OpenFileDescription.h>
#include <Kernel/FileSystem/VirtualFileSystem.h>
#include <Kernel/Memory/MemoryManager.h>
#include <Kernel/Net/Socket.h>
#include <Kernel/Tasks/Process.h>
#include <Kernel/UnixTypes.h>
namespace Kernel {
ErrorOr<NonnullRefPtr<OpenFileDescription>> OpenFileDescription::try_create(Custody& custody)
{
auto inode_file = TRY(InodeFile::create(custody.inode()));
auto description = TRY(adopt_nonnull_ref_or_enomem(new (nothrow) OpenFileDescription(move(inode_file))));
description->m_state.with([&](auto& state) { state.custody = custody; });
TRY(description->attach());
return description;
}
ErrorOr<NonnullRefPtr<OpenFileDescription>> OpenFileDescription::try_create(File& file)
{
auto description = TRY(adopt_nonnull_ref_or_enomem(new (nothrow) OpenFileDescription(file)));
TRY(description->attach());
return description;
}
OpenFileDescription::OpenFileDescription(File& file)
: m_file(file)
{
if (file.is_inode())
m_inode = static_cast<InodeFile&>(file).inode();
auto metadata = this->metadata();
m_state.with([&](auto& state) { state.is_directory = metadata.is_directory(); });
}
OpenFileDescription::~OpenFileDescription()
{
m_file->detach(*this);
// FIXME: Should this error path be observed somehow?
(void)m_file->close();
if (m_inode) {
m_inode->detach(*this);
m_inode->remove_flocks_for_description(*this);
}
}
ErrorOr<void> OpenFileDescription::attach()
{
if (m_inode)
TRY(m_inode->attach(*this));
return m_file->attach(*this);
}
void OpenFileDescription::set_original_custody(Custody& custody)
{
m_state.with([&](auto& state) { state.custody = custody; });
}
Thread::FileBlocker::BlockFlags OpenFileDescription::should_unblock(Thread::FileBlocker::BlockFlags block_flags) const
{
using BlockFlags = Thread::FileBlocker::BlockFlags;
BlockFlags unblock_flags = BlockFlags::None;
if (has_flag(block_flags, BlockFlags::Read) && can_read())
unblock_flags |= BlockFlags::Read;
if (has_flag(block_flags, BlockFlags::Write) && can_write())
unblock_flags |= BlockFlags::Write;
// TODO: Implement Thread::FileBlocker::BlockFlags::Exception
if (has_any_flag(block_flags, BlockFlags::SocketFlags)) {
auto const* sock = socket();
VERIFY(sock);
if (has_flag(block_flags, BlockFlags::Accept) && sock->can_accept())
unblock_flags |= BlockFlags::Accept;
if (has_flag(block_flags, BlockFlags::Connect) && sock->setup_state() == Socket::SetupState::Completed)
unblock_flags |= BlockFlags::Connect;
}
return unblock_flags;
}
ErrorOr<struct stat> OpenFileDescription::stat()
{
// FIXME: This is due to the Device class not overriding File::stat().
if (m_inode)
return m_inode->metadata().stat();
return m_file->stat();
}
ErrorOr<off_t> OpenFileDescription::seek(off_t offset, int whence)
{
if (!m_file->is_seekable())
return ESPIPE;
auto metadata = this->metadata();
auto new_offset = TRY(m_state.with([&](auto& state) -> ErrorOr<off_t> {
off_t new_offset;
switch (whence) {
case SEEK_SET:
new_offset = offset;
break;
case SEEK_CUR:
if (Checked<off_t>::addition_would_overflow(state.current_offset, offset))
return EOVERFLOW;
new_offset = state.current_offset + offset;
break;
case SEEK_END:
if (!metadata.is_valid())
return EIO;
if (Checked<off_t>::addition_would_overflow(metadata.size, offset))
return EOVERFLOW;
new_offset = metadata.size + offset;
break;
default:
return EINVAL;
}
if (new_offset < 0)
return EINVAL;
state.current_offset = new_offset;
return new_offset;
}));
// FIXME: Return EINVAL if attempting to seek past the end of a seekable device.
m_file->did_seek(*this, new_offset);
if (m_inode)
m_inode->did_seek(*this, new_offset);
evaluate_block_conditions();
return new_offset;
}
ErrorOr<size_t> OpenFileDescription::read(UserOrKernelBuffer& buffer, u64 offset, size_t count)
{
if (Checked<u64>::addition_would_overflow(offset, count))
return EOVERFLOW;
return m_file->read(*this, offset, buffer, count);
}
ErrorOr<size_t> OpenFileDescription::write(u64 offset, UserOrKernelBuffer const& data, size_t data_size)
{
if (Checked<u64>::addition_would_overflow(offset, data_size))
return EOVERFLOW;
return m_file->write(*this, offset, data, data_size);
}
ErrorOr<size_t> OpenFileDescription::read(UserOrKernelBuffer& buffer, size_t count)
{
auto offset = TRY(m_state.with([&](auto& state) -> ErrorOr<off_t> {
if (Checked<off_t>::addition_would_overflow(state.current_offset, count))
return EOVERFLOW;
return state.current_offset;
}));
auto nread = TRY(m_file->read(*this, offset, buffer, count));
if (m_file->is_seekable())
m_state.with([&](auto& state) { state.current_offset = offset + nread; });
evaluate_block_conditions();
return nread;
}
ErrorOr<size_t> OpenFileDescription::write(UserOrKernelBuffer const& data, size_t size)
{
auto offset = TRY(m_state.with([&](auto& state) -> ErrorOr<off_t> {
if (Checked<off_t>::addition_would_overflow(state.current_offset, size))
return EOVERFLOW;
return state.current_offset;
}));
auto nwritten = TRY(m_file->write(*this, offset, data, size));
if (m_file->is_seekable())
m_state.with([&](auto& state) { state.current_offset = offset + nwritten; });
evaluate_block_conditions();
return nwritten;
}
bool OpenFileDescription::can_write() const
{
return m_file->can_write(*this, offset());
}
bool OpenFileDescription::can_read() const
{
return m_file->can_read(*this, offset());
}
ErrorOr<size_t> OpenFileDescription::get_dir_entries(UserOrKernelBuffer& output_buffer, size_t size)
{
if (!is_directory())
return ENOTDIR;
auto metadata = this->metadata();
if (!metadata.is_valid())
return EIO;
size_t remaining = size;
u8 stack_buffer[PAGE_SIZE];
Bytes temp_buffer(stack_buffer, sizeof(stack_buffer));
FixedMemoryStream stream { temp_buffer };
auto flush_stream_to_output_buffer = [&stream, &remaining, &temp_buffer, &output_buffer]() -> ErrorOr<void> {
auto buffered_size = TRY(stream.tell());
if (buffered_size == 0)
return {};
if (remaining < buffered_size)
return Error::from_errno(EINVAL);
TRY(output_buffer.write(temp_buffer.trim(buffered_size)));
output_buffer = output_buffer.offset(buffered_size);
remaining -= buffered_size;
TRY(stream.seek(0));
return {};
};
ErrorOr<void> result = m_inode->traverse_as_directory([&flush_stream_to_output_buffer, &stream, this](auto& entry) -> ErrorOr<void> {
// FIXME: Double check the calculation, at least the type for the name length mismatches.
size_t serialized_size = sizeof(ino_t) + sizeof(u8) + sizeof(size_t) + sizeof(char) * entry.name.length();
if (serialized_size > TRY(stream.size()) - TRY(stream.tell()))
TRY(flush_stream_to_output_buffer());
MUST(stream.write_value<u64>(entry.inode.index().value()));
MUST(stream.write_value(m_inode->fs().internal_file_type_to_directory_entry_type(entry)));
MUST(stream.write_value<u32>(entry.name.length()));
MUST(stream.write_until_depleted(entry.name.bytes()));
return {};
});
if (result.is_error()) {
// We should only return EFAULT when the userspace buffer is too small,
// so that userspace can reliably use it as a signal to increase its
// buffer size.
VERIFY(result.error().code() != EFAULT);
return result.release_error();
}
TRY(flush_stream_to_output_buffer());
return size - remaining;
}
bool OpenFileDescription::is_device() const
{
return m_file->is_device();
}
Device const* OpenFileDescription::device() const
{
if (!is_device())
return nullptr;
return static_cast<Device const*>(m_file.ptr());
}
Device* OpenFileDescription::device()
{
if (!is_device())
return nullptr;
return static_cast<Device*>(m_file.ptr());
}
bool OpenFileDescription::is_tty() const
{
return m_file->is_tty();
}
TTY const* OpenFileDescription::tty() const
{
if (!is_tty())
return nullptr;
return static_cast<TTY const*>(m_file.ptr());
}
TTY* OpenFileDescription::tty()
{
if (!is_tty())
return nullptr;
return static_cast<TTY*>(m_file.ptr());
}
bool OpenFileDescription::is_inode_watcher() const
{
return m_file->is_inode_watcher();
}
InodeWatcher const* OpenFileDescription::inode_watcher() const
{
if (!is_inode_watcher())
return nullptr;
return static_cast<InodeWatcher const*>(m_file.ptr());
}
InodeWatcher* OpenFileDescription::inode_watcher()
{
if (!is_inode_watcher())
return nullptr;
return static_cast<InodeWatcher*>(m_file.ptr());
}
bool OpenFileDescription::is_unshared_resource_file() const
{
return m_file->is_unshared_resource_file();
}
UnsharedResourceFile* OpenFileDescription::unshared_resource_file()
{
if (!is_unshared_resource_file())
return nullptr;
return static_cast<UnsharedResourceFile*>(m_file.ptr());
}
UnsharedResourceFile const* OpenFileDescription::unshared_resource_file() const
{
if (!is_unshared_resource_file())
return nullptr;
return static_cast<UnsharedResourceFile const*>(m_file.ptr());
}
bool OpenFileDescription::is_mount_file() const
{
return m_file->is_mount_file();
}
MountFile const* OpenFileDescription::mount_file() const
{
if (!is_mount_file())
return nullptr;
return static_cast<MountFile const*>(m_file.ptr());
}
MountFile* OpenFileDescription::mount_file()
{
if (!is_mount_file())
return nullptr;
return static_cast<MountFile*>(m_file.ptr());
}
bool OpenFileDescription::is_master_pty() const
{
return m_file->is_master_pty();
}
MasterPTY const* OpenFileDescription::master_pty() const
{
if (!is_master_pty())
return nullptr;
return static_cast<MasterPTY const*>(m_file.ptr());
}
MasterPTY* OpenFileDescription::master_pty()
{
if (!is_master_pty())
return nullptr;
return static_cast<MasterPTY*>(m_file.ptr());
}
ErrorOr<void> OpenFileDescription::close()
{
if (m_file->attach_count() > 0)
return {};
return m_file->close();
}
ErrorOr<NonnullOwnPtr<KString>> OpenFileDescription::original_absolute_path() const
{
if (auto custody = this->custody())
return custody->try_serialize_absolute_path();
return ENOENT;
}
ErrorOr<NonnullOwnPtr<KString>> OpenFileDescription::pseudo_path() const
{
if (auto custody = this->custody())
return custody->try_serialize_absolute_path();
return m_file->pseudo_path(*this);
}
InodeMetadata OpenFileDescription::metadata() const
{
if (m_inode)
return m_inode->metadata();
return {};
}
ErrorOr<File::VMObjectAndMemoryType> OpenFileDescription::vmobject_for_mmap(Process& process, Memory::VirtualRange const& range, u64& offset, bool shared)
{
return m_file->vmobject_and_memory_type_for_mmap(process, range, offset, shared);
}
ErrorOr<void> OpenFileDescription::truncate(u64 length)
{
return m_file->truncate(length);
}
ErrorOr<void> OpenFileDescription::sync()
{
return m_file->sync();
}
bool OpenFileDescription::is_fifo() const
{
return m_file->is_fifo();
}
FIFO* OpenFileDescription::fifo()
{
if (!is_fifo())
return nullptr;
return static_cast<FIFO*>(m_file.ptr());
}
bool OpenFileDescription::is_socket() const
{
return m_file->is_socket();
}
Socket* OpenFileDescription::socket()
{
if (!is_socket())
return nullptr;
return static_cast<Socket*>(m_file.ptr());
}
Socket const* OpenFileDescription::socket() const
{
if (!is_socket())
return nullptr;
return static_cast<Socket const*>(m_file.ptr());
}
void OpenFileDescription::set_file_flags(u32 flags)
{
m_state.with([&](auto& state) {
state.is_blocking = !(flags & O_NONBLOCK);
state.should_append = flags & O_APPEND;
state.direct = flags & O_DIRECT;
state.file_flags = flags;
});
}
ErrorOr<void> OpenFileDescription::chmod(Credentials const& credentials, mode_t mode)
{
return m_file->chmod(credentials, *this, mode);
}
ErrorOr<void> OpenFileDescription::chown(Credentials const& credentials, UserID uid, GroupID gid)
{
return m_file->chown(credentials, *this, uid, gid);
}
FileBlockerSet& OpenFileDescription::blocker_set()
{
return m_file->blocker_set();
}
ErrorOr<void> OpenFileDescription::apply_flock(Process const& process, Userspace<flock const*> lock, ShouldBlock should_block)
{
if (!m_inode)
return EBADF;
return m_inode->apply_flock(process, *this, lock, should_block);
}
ErrorOr<void> OpenFileDescription::get_flock(Userspace<flock*> lock) const
{
if (!m_inode)
return EBADF;
return m_inode->get_flock(*this, lock);
}
bool OpenFileDescription::is_readable() const
{
return m_state.with([](auto& state) { return state.readable; });
}
bool OpenFileDescription::is_writable() const
{
return m_state.with([](auto& state) { return state.writable; });
}
void OpenFileDescription::set_readable(bool b)
{
m_state.with([&](auto& state) { state.readable = b; });
}
void OpenFileDescription::set_writable(bool b)
{
m_state.with([&](auto& state) { state.writable = b; });
}
void OpenFileDescription::set_rw_mode(int options)
{
m_state.with([&](auto& state) {
state.readable = (options & O_RDONLY) == O_RDONLY;
state.writable = (options & O_WRONLY) == O_WRONLY;
});
}
bool OpenFileDescription::is_direct() const
{
return m_state.with([](auto& state) { return state.direct; });
}
bool OpenFileDescription::is_directory() const
{
return m_state.with([](auto& state) { return state.is_directory; });
}
bool OpenFileDescription::is_blocking() const
{
return m_state.with([](auto& state) { return state.is_blocking; });
}
void OpenFileDescription::set_blocking(bool b)
{
m_state.with([&](auto& state) { state.is_blocking = b; });
}
bool OpenFileDescription::should_append() const
{
return m_state.with([](auto& state) { return state.should_append; });
}
u32 OpenFileDescription::file_flags() const
{
return m_state.with([](auto& state) { return state.file_flags; });
}
FIFO::Direction OpenFileDescription::fifo_direction() const
{
return m_state.with([](auto& state) { return state.fifo_direction; });
}
void OpenFileDescription::set_fifo_direction(Badge<FIFO>, FIFO::Direction direction)
{
m_state.with([&](auto& state) { state.fifo_direction = direction; });
}
OwnPtr<OpenFileDescriptionData>& OpenFileDescription::data()
{
return m_state.with([](auto& state) -> OwnPtr<OpenFileDescriptionData>& { return state.data; });
}
off_t OpenFileDescription::offset() const
{
return m_state.with([](auto& state) { return state.current_offset; });
}
RefPtr<Custody const> OpenFileDescription::custody() const
{
return m_state.with([](auto& state) { return state.custody; });
}
RefPtr<Custody> OpenFileDescription::custody()
{
return m_state.with([](auto& state) { return state.custody; });
}
}