For a long time, our shutdown procedure has basically been:
- Acquire big process lock.
- Switch framebuffer to Kernel debug console.
- Sync and lock all file systems so that disk caches are flushed and
files are in a good state.
- Use firmware and architecture-specific functionality to perform
hardware shutdown.
This naive and simple shutdown procedure has multiple issues:
- No processes are terminated properly, meaning they cannot perform more
complex cleanup work. If they were in the middle of I/O, for instance,
only the data that already reached the Kernel is written to disk, and
data corruption due to unfinished writes can therefore still occur.
- No file systems are unmounted, meaning that any important unmount work
will never happen. This is important for e.g. Ext2, which has
facilites for detecting improper unmounts (see superblock's s_state
variable) and therefore requires a proper unmount to be performed.
This was also the starting point for this PR, since I wanted to
introduce basic Ext2 file system checking and unmounting.
- No hardware is properly shut down beyond what the system firmware does
on its own.
- Shutdown is performed within the write() call that asked the Kernel to
change its power state. If the shutdown procedure takes longer (i.e.
when it's done properly), this blocks the process causing the shutdown
and prevents any potentially-useful interactions between Kernel and
userland during shutdown.
In essence, current shutdown is a glorified system crash with minimal
file system cleanliness guarantees.
Therefore, this commit is the first step in improving our shutdown
procedure. The new shutdown flow is now as follows:
- From the write() call to the power state SysFS node, a new task is
started, the Power State Switch Task. Its only purpose is to change
the operating system's power state. This task takes over shutdown and
reboot duties, although reboot is not modified in this commit.
- The Power State Switch Task assumes that userland has performed all
shutdown duties it can perform on its own. In particular, it assumes
that all kinds of clean process shutdown have been done, and remaining
processes can be hard-killed without consequence. This is an important
separation of concerns: While this commit does not modify userland, in
the future SystemServer will be responsible for performing proper
shutdown of user processes, including timeouts for stubborn processes
etc.
- As mentioned above, the task hard-kills remaining user processes.
- The task hard-kills all Kernel processes except itself and the
Finalizer Task. Since Kernel processes can delay their own shutdown
indefinitely if they want to, they have plenty opportunity to perform
proper shutdown if necessary. This may become a problem with
non-cooperative Kernel tasks, but as seen two commits earlier, for now
all tasks will cooperate within a few seconds.
- The task waits for the Finalizer Task to clean up all processes.
- The task hard-kills and finalizes the Finalizer Task itself, meaning
that it now is the only remaining process in the system.
- The task syncs and locks all file systems, and then unmounts them. Due
to an unknown refcount bug we currently cannot unmount the root file
system; therefore the task is able to abort the clean unmount if
necessary.
- The task performs platform-dependent hardware shutdown as before.
This commit has multiple remaining issues (or exposed existing ones)
which will need to be addressed in the future but are out of scope for
now:
- Unmounting the root filesystem is impossible due to remaining
references to the inodes /home and /home/anon. I investigated this
very heavily and could not find whoever is holding the last two
references.
- Userland cannot perform proper cleanup, since the Kernel's power state
variable is accessed directly by tools instead of a proper userland
shutdown procedure directed by SystemServer.
The recently introduced Firmware/PowerState procedures are removed
again, since all of the architecture-independent code can live in the
power state switch task. The architecture-specific code is kept,
however.
Since we never check a kernel process's state like a userland process,
it's possible for a kernel process to ignore the fact that someone is
trying to kill it, and continue running. This is not desireable if we
want to properly shutdown all processes, including Kernel ones.
This is correct since unmount doesn't treat bind mounts specially. If we
don't do this, unmounting bind mounts will call
prepare_for_last_unmount() on the guest FS much too early, which will
most likely fail due to a busy file system.
This is a preparation before we can create a usable mechanism to use
filesystem-specific mount flags.
To keep some compatibility with userland code, LibC and LibCore mount
functions are kept being usable, but now instead of doing an "atomic"
syscall, they do multiple syscalls to perform the complete procedure of
mounting a filesystem.
The FileBackedFileSystem IntrusiveList in the VFS code is now changed to
be protected by a Mutex, because when we mount a new filesystem, we need
to check if a filesystem is already created for a given source_fd so we
do a scan for that OpenFileDescription in that list. If we fail to find
an already-created filesystem we create a new one and register it in the
list if we successfully mounted it. We use a Mutex because we might need
to initiate disk access during the filesystem creation, which will take
other mutexes in other parts of the kernel, therefore making it not
possible to take a spinlock while doing this.
Instead of using ifdefs to use the correct platform-specific methods, we
can just use the same pattern we use for the microseconds_delay function
which has specific implementations for each Arch CPU subdirectory.
When linking a kernel image, the actual correct and platform-specific
power-state changing methods will be called in Firmware/PowerState.cpp
file.
All code that is related to PC BIOS should not be in the Kernel/Firmware
directory as this directory is for abstracted and platform-agnostic code
like ACPI (and device tree parsing in the future).
This fixes a problem with the aarch64 architecure, as these machines
don't have any PC-BIOS in them so actually trying to access these memory
locations (EBDA, BIOS ROM) does not make any sense, as they're specific
to x86 machines only.
Previously, reads would only be successful for offset 0. For this
reason, the maximum size that could be correctly read from the PCI
expansion ROM SysFS node was limited to the block size, and
subsequent blocks would fail. This commit fixes the computation of
the number of bytes to read.
Like the HID, Audio and Storage subsystem, the Graphics subsystem (which
handles GPUs technically) exposes unix device files (typically in /dev).
To ensure consistency across the repository, move all related files to a
new directory under Kernel/Devices called "GPU".
Also remove the redundant "GPU" word from the VirtIO driver directory,
and the word "Graphics" from GraphicsManagement.{h,cpp} filenames.
This has KString, KBuffer, DoubleBuffer, KBufferBuilder, IOWindow,
UserOrKernelBuffer and ScopedCritical classes being moved to the
Kernel/Library subdirectory.
Also, move the panic and assertions handling code to that directory.
When deleting a directory, the rmdir syscall should fail if the path was
unveiled without the 'c' permission. This matches the same behavior that
OpenBSD enforces when doing this kind of operation.
When deleting a file, the unlink syscall should fail if the path was
unveiled without the 'w' permission, to ensure that userspace is aware
of the possibility of removing a file only when the path was unveiled as
writable.
When using the userdel utility, we now unveil that directory path with
the unveil 'c' permission so removal of an account home directory is
done properly.
The Storage subsystem, like the Audio and HID subsystems, exposes Unix
device files (for example, in the /dev directory). To ensure consistency
across the repository, we should make the Storage subsystem to reside in
the Kernel/Devices directory like the two other mentioned subsystems.
The contents of the directory inode could change if we are not taking so
we must take the m_inode_lock to prevent corruption when reading the
directory contents.
This is not needed, because when we are doing this traversing, functions
that are called from this function are using proper and more "atomic"
locking.
"Wherever applicable" = most places, actually :^), especially for
networking and filesystem timestamps.
This includes changes to unzip, which uses DOSPackedTime, since that is
changed for the FAT file systems.
That's what this class really is; in fact that's what the first line of
the comment says it is.
This commit does not rename the main files, since those will contain
other time-related classes in a little bit.
The Raspberry Pi hardware doesn't support a proper software-initiated
shutdown, so this instead uses the watchdog to reboot to a special
partition which the firmware interprets as an immediate halt on
shutdown. When running under Qemu, this causes the emulator to exit.
These functions would have caused a `-Woverloaded-virtual` warning with
GCC 13, as they shadow `File::{attach,detach}(OpenFileDescription&)`.
Both of these functions had a single call site. This commit inlines
`attach` into its only caller, `FIFO::open_direction`.
Instead of explicitly checking `is_fifo()` in `~OpenFileDescription`
before running the `detach(Direction)` overload, let's just override the
regular `detach(OpenFileDescription&)` for `FIFO` to perform this action
instead.
Whenever an entry is added to the cache, the last element is removed to
make space for the new entry(if the cache is full). To make this an LRU
cache, the entry needs to be moved to the front of the list when there
is a cache hit so that the least recently used entry moves to the end
to be evicted first.
This was the last change that was needed to be able boot with the flag
of LOCK_IN_CRITICAL_DEBUG. That flag is not always enabled because there
are still other issues in which we hold a spinlock and still try to lock
a mutex.
Instead of using one global mutex we can protect internal structures of
the InodeWatcher class with SpinlockProtected wrappers. This in turn
allows the InodeWatcher code to be called from other parts in the kernel
while holding a prior spinlock properly.
`process.fds()` is protected by a Mutex, which causes issues when we try
to acquire it while holding a Spinlock. Since nothing seems to use this
value, let's just remove it entirely for now.
The existing `read_entire` is quite slow due to allocating and copying
multiple times, but it is simultaneously quite hard to get rid of in a
single step. As a replacement, add a new function that reads as much as
possible directly into a user-provided buffer.
To do this we also need to get rid of LockRefPtrs in the USB code as
well.
Most of the SysFS nodes are statically generated during boot and are not
mutated afterwards.
The same goes for general device code - once we generate the appropriate
SysFS nodes, we almost never mutate the node pointers afterwards, making
locking unnecessary.
We have a problem with the original utimensat syscall because when we
do call LibC futimens function, internally we provide an empty path,
and the Kernel get_syscall_path_argument method will detect this as an
invalid path.
This happens to spit an error for example in the touch utility, so if a
user is running "touch non_existing_file", it will create that file, but
the user will still see an error coming from LibC futimens function.
This new syscall gets an open file description and it provides the same
functionality as utimensat, on the specified open file description.
The new syscall will be used later by LibC to properly implement LibC
futimens function so the situation described with relation to the
"touch" utility could be fixed.
These were easy to pick-up as these pointers are assigned during the
construction point and are never changed afterwards.
This small change to these pointers will ensure that our code will not
accidentally assign these pointers with a new object which is always a
kind of bug we will want to prevent.
These were stored in a bunch of places. The main one that's a bit iffy
is the Mutex::m_holder one, which I'm going to simplify in a subsequent
commit.
In Plan9FS and WorkQueue, we can't make the NNRPs const due to
initialization order problems. That's probably doable with further
cleanup, but left as an exercise for our future selves.
Before starting this, I expected the thread blockers to be a problem,
but as it turns out they were super straightforward (for once!) as they
don't mutate the thread after initiating a block, so they can just use
simple const-ified NNRPs.
