mod file_attr;
mod inode_tracker;
mod inodes;
#[cfg(test)]
mod tests;
use crate::{
blobservice::{BlobReader, BlobService},
directoryservice::DirectoryService,
fuse::{
file_attr::gen_file_attr,
inodes::{DirectoryInodeData, InodeData},
},
pathinfoservice::PathInfoService,
proto::{node::Node, NamedNode},
B3Digest, Error,
};
use fuser::{FileAttr, ReplyAttr, Request};
use nix_compat::store_path::StorePath;
use std::io::{self, Read, Seek};
use std::os::unix::ffi::OsStrExt;
use std::str::FromStr;
use std::sync::Arc;
use std::{collections::HashMap, time::Duration};
use tracing::{debug, info_span, warn};
use self::inode_tracker::InodeTracker;
/// This implements a read-only FUSE filesystem for a tvix-store
/// with the passed [BlobService], [DirectoryService] and [PathInfoService].
///
/// We don't allow listing on the root mountpoint (inode 0).
/// In the future, this might be made configurable once a listing method is
/// added to [self.path_info_service], and then show all store paths in that
/// store.
///
/// Linux uses inodes in filesystems. When implementing FUSE, most calls are
/// *for* a given inode.
///
/// This means, we need to have a stable mapping of inode numbers to the
/// corresponding store nodes.
///
/// We internally delegate all inode allocation and state keeping to a
/// [InodeTracker], and store the currently "explored" store paths together with
/// root inode of the root.
///
/// There's some places where inodes are allocated / data inserted into
/// [self.inode_tracker], if not allocated before already:
/// - Processing a `lookup` request, either in the mount root, or somewhere
/// deeper
/// - Processing a `readdir` request
///
/// Things pointing to the same contents get the same inodes, irrespective of
/// their own location.
/// This means:
/// - Symlinks with the same target will get the same inode.
/// - Regular/executable files with the same contents will get the same inode
/// - Directories with the same contents will get the same inode.
///
/// Due to the above being valid across the whole store, and considering the
/// merkle structure is a DAG, not a tree, this also means we can't do "bucketed
/// allocation", aka reserve Directory.size inodes for each PathInfo.
pub struct FUSE {
blob_service: Arc<dyn BlobService>,
directory_service: Arc<dyn DirectoryService>,
path_info_service: Arc<dyn PathInfoService>,
/// This maps a given StorePath to the inode we allocated for the root inode.
store_paths: HashMap<StorePath, u64>,
/// This keeps track of inodes and data alongside them.
inode_tracker: InodeTracker,
/// This holds all open file handles
file_handles: HashMap<u64, Box<dyn BlobReader>>,
next_file_handle: u64,
}
impl FUSE {
pub fn new(
blob_service: Arc<dyn BlobService>,
directory_service: Arc<dyn DirectoryService>,
path_info_service: Arc<dyn PathInfoService>,
) -> Self {
Self {
blob_service,
directory_service,
path_info_service,
store_paths: HashMap::default(),
inode_tracker: Default::default(),
file_handles: Default::default(),
next_file_handle: 1,
}
}
/// This will turn a lookup request for [std::ffi::OsStr] in the root to
/// a ino and [InodeData].
/// It will peek in [self.store_paths], and then either look it up from
/// [self.inode_tracker],
/// or otherwise fetch from [self.path_info_service], and then insert into
/// [self.inode_tracker].
fn name_in_root_to_ino_and_data(
&mut self,
name: &std::ffi::OsStr,
) -> Result<Option<(u64, Arc<InodeData>)>, Error> {
// parse the name into a [StorePath].
let store_path = if let Some(name) = name.to_str() {
match StorePath::from_str(name) {
Ok(store_path) => store_path,
Err(e) => {
debug!(e=?e, "unable to parse as store path");
// This is not an error, but a "ENOENT", as someone can stat
// a file inside the root that's no valid store path
return Ok(None);
}
}
} else {
debug!("{name:?} is no string");
// same here.
return Ok(None);
};
if let Some(ino) = self.store_paths.get(&store_path) {
// If we already have that store path, lookup the inode from
// self.store_paths and then get the data from [self.inode_tracker],
// which in the case of a [InodeData::Directory] will be fully
// populated.
Ok(Some((
*ino,
self.inode_tracker.get(*ino).expect("must exist"),
)))
} else {
// If we don't have it, look it up in PathInfoService.
match self.path_info_service.get(store_path.digest)? {
// the pathinfo doesn't exist, so the file doesn't exist.
None => Ok(None),
Some(path_info) => {
// The pathinfo does exist, so there must be a root node
let root_node = path_info.node.unwrap().node.unwrap();
// The name must match what's passed in the lookup, otherwise we return nothing.
if root_node.get_name() != store_path.to_string().as_bytes() {
return Ok(None);
}
// insert the (sparse) inode data and register in
// self.store_paths.
// FUTUREWORK: change put to return the data after
// inserting, so we don't need to lookup a second
// time?
let ino = self.inode_tracker.put((&root_node).into());
self.store_paths.insert(store_path, ino);
Ok(Some((ino, self.inode_tracker.get(ino).unwrap())))
}
}
}
}
/// This will lookup a directory by digest, and will turn it into a
/// [InodeData::Directory(DirectoryInodeData::Populated(..))].
/// This is both used to initially insert the root node of a store path,
/// as well as when looking up an intermediate DirectoryNode.
fn fetch_directory_inode_data(&self, directory_digest: &B3Digest) -> Result<InodeData, Error> {
match self.directory_service.get(directory_digest) {
Err(e) => {
warn!(e = e.to_string(), directory.digest=%directory_digest, "failed to get directory");
Err(e)
}
// If the Directory can't be found, this is a hole, bail out.
Ok(None) => {
tracing::error!(directory.digest=%directory_digest, "directory not found in directory service");
Err(Error::StorageError(format!(
"directory {} not found",
directory_digest
)))
}
Ok(Some(directory)) => Ok(directory.into()),
}
}
}
impl fuser::Filesystem for FUSE {
#[tracing::instrument(skip_all, fields(rq.inode = ino))]
fn getattr(&mut self, _req: &Request, ino: u64, reply: ReplyAttr) {
debug!("getattr");
if ino == fuser::FUSE_ROOT_ID {
reply.attr(&Duration::MAX, &file_attr::ROOT_FILE_ATTR);
return;
}
match self.inode_tracker.get(ino) {
None => reply.error(libc::ENOENT),
Some(node) => {
debug!(node = ?node, "found node");
reply.attr(&Duration::MAX, &file_attr::gen_file_attr(&node, ino));
}
}
}
#[tracing::instrument(skip_all, fields(rq.parent_inode = parent_ino, rq.name = ?name))]
fn lookup(
&mut self,
_req: &Request,
parent_ino: u64,
name: &std::ffi::OsStr,
reply: fuser::ReplyEntry,
) {
debug!("lookup");
// This goes from a parent inode to a node.
// - If the parent is [fuser::FUSE_ROOT_ID], we need to check
// [self.store_paths] (fetching from PathInfoService if needed)
// - Otherwise, lookup the parent in [self.inode_tracker] (which must be
// a [InodeData::Directory]), and find the child with that name.
if parent_ino == fuser::FUSE_ROOT_ID {
match self.name_in_root_to_ino_and_data(name) {
Err(e) => {
warn!("{}", e);
reply.error(libc::EIO);
}
Ok(None) => {
reply.error(libc::ENOENT);
}
Ok(Some((ino, inode_data))) => {
debug!(inode_data=?&inode_data, ino=ino, "Some");
reply_with_entry(reply, &gen_file_attr(&inode_data, ino));
}
}
} else {
// This is the "lookup for "a" inside inode 42.
// We already know that inode 42 must be a directory.
// It might not be populated yet, so if it isn't, we do (by
// fetching from [self.directory_service]), and save the result in
// [self.inode_tracker].
// Now it for sure is populated, so we search for that name in the
// list of children and return the FileAttrs.
let parent_data = self.inode_tracker.get(parent_ino).unwrap();
let parent_data = match *parent_data {
InodeData::Regular(..) | InodeData::Symlink(_) => {
// if the parent inode was not a directory, this doesn't make sense
reply.error(libc::ENOTDIR);
return;
}
InodeData::Directory(DirectoryInodeData::Sparse(ref parent_digest, _)) => {
match self.fetch_directory_inode_data(parent_digest) {
Ok(new_data) => {
// update data in [self.inode_tracker] with populated variant.
// FUTUREWORK: change put to return the data after
// inserting, so we don't need to lookup a second
// time?
let ino = self.inode_tracker.put(new_data);
self.inode_tracker.get(ino).unwrap()
}
Err(_e) => {
reply.error(libc::EIO);
return;
}
}
}
InodeData::Directory(DirectoryInodeData::Populated(..)) => parent_data,
};
// now parent_data can only be a [InodeData::Directory(DirectoryInodeData::Populated(..))].
let (parent_digest, children) = if let InodeData::Directory(
DirectoryInodeData::Populated(ref parent_digest, ref children),
) = *parent_data
{
(parent_digest, children)
} else {
panic!("unexpected type")
};
let span = info_span!("lookup", directory.digest = %parent_digest);
let _enter = span.enter();
// in the children, find the one with the desired name.
if let Some((child_ino, _)) =
children.iter().find(|e| e.1.get_name() == name.as_bytes())
{
// lookup the child [InodeData] in [self.inode_tracker].
// We know the inodes for children have already been allocated.
let child_inode_data = self.inode_tracker.get(*child_ino).unwrap();
// Reply with the file attributes for the child.
// For child directories, we still have all data we need to reply.
reply_with_entry(reply, &gen_file_attr(&child_inode_data, *child_ino));
} else {
// Child not found, return ENOENT.
reply.error(libc::ENOENT);
}
}
}
// TODO: readdirplus?
#[tracing::instrument(skip_all, fields(rq.inode = ino, rq.offset = offset))]
fn readdir(
&mut self,
_req: &Request<'_>,
ino: u64,
_fh: u64,
offset: i64,
mut reply: fuser::ReplyDirectory,
) {
debug!("readdir");
if ino == fuser::FUSE_ROOT_ID {
reply.error(libc::EPERM); // same error code as ipfs/kubo
return;
}
// lookup the inode data.
let dir_inode_data = self.inode_tracker.get(ino).unwrap();
let dir_inode_data = match *dir_inode_data {
InodeData::Regular(..) | InodeData::Symlink(..) => {
warn!("Not a directory");
reply.error(libc::ENOTDIR);
return;
}
InodeData::Directory(DirectoryInodeData::Sparse(ref directory_digest, _)) => {
match self.fetch_directory_inode_data(directory_digest) {
Ok(new_data) => {
// update data in [self.inode_tracker] with populated variant.
// FUTUREWORK: change put to return the data after
// inserting, so we don't need to lookup a second
// time?
let ino = self.inode_tracker.put(new_data);
self.inode_tracker.get(ino).unwrap()
}
Err(_e) => {
reply.error(libc::EIO);
return;
}
}
}
InodeData::Directory(DirectoryInodeData::Populated(..)) => dir_inode_data,
};
// now parent_data can only be InodeData::Directory(DirectoryInodeData::Populated(..))
if let InodeData::Directory(DirectoryInodeData::Populated(ref _digest, ref children)) =
*dir_inode_data
{
for (i, (ino, child_node)) in children.iter().skip(offset as usize).enumerate() {
// the second parameter will become the "offset" parameter on the next call.
let full = reply.add(
*ino,
offset + i as i64 + 1_i64,
match child_node {
Node::Directory(_) => fuser::FileType::Directory,
Node::File(_) => fuser::FileType::RegularFile,
Node::Symlink(_) => fuser::FileType::Symlink,
},
std::ffi::OsStr::from_bytes(child_node.get_name()),
);
if full {
break;
}
}
reply.ok();
} else {
panic!("unexpected type")
}
}
#[tracing::instrument(skip_all, fields(rq.inode = ino))]
fn open(&mut self, _req: &Request<'_>, ino: u64, _flags: i32, reply: fuser::ReplyOpen) {
// get a new file handle
let fh = self.next_file_handle;
if ino == fuser::FUSE_ROOT_ID {
reply.error(libc::ENOSYS);
return;
}
// lookup the inode
match *self.inode_tracker.get(ino).unwrap() {
// read is invalid on non-files.
InodeData::Directory(..) | InodeData::Symlink(_) => {
warn!("is directory");
reply.error(libc::EISDIR);
}
InodeData::Regular(ref blob_digest, _blob_size, _) => {
let span = info_span!("read", blob.digest = %blob_digest);
let _enter = span.enter();
match self.blob_service.open_read(blob_digest) {
Ok(None) => {
warn!("blob not found");
reply.error(libc::EIO);
}
Err(e) => {
warn!(e=?e, "error opening blob");
reply.error(libc::EIO);
}
Ok(Some(blob_reader)) => {
self.file_handles.insert(fh, blob_reader);
reply.opened(fh, 0);
// TODO: this will overflow after 2**64 operations,
// which is fine for now.
// See https://cl.tvl.fyi/c/depot/+/8834/comment/a6684ce0_d72469d1
// for the discussion on alternatives.
self.next_file_handle += 1;
}
}
}
}
}
#[tracing::instrument(skip_all, fields(rq.inode = ino, fh = fh))]
fn release(
&mut self,
_req: &Request<'_>,
ino: u64,
fh: u64,
_flags: i32,
_lock_owner: Option<u64>,
_flush: bool,
reply: fuser::ReplyEmpty,
) {
// remove and get ownership on the blob reader
let blob_reader = self.file_handles.remove(&fh).unwrap();
// drop it, which will close it.
drop(blob_reader);
reply.ok();
}
#[tracing::instrument(skip_all, fields(rq.inode = ino, rq.offset = offset, rq.size = size))]
fn read(
&mut self,
_req: &Request<'_>,
ino: u64,
fh: u64,
offset: i64,
size: u32,
_flags: i32,
_lock_owner: Option<u64>,
reply: fuser::ReplyData,
) {
debug!("read");
let blob_reader = self.file_handles.get_mut(&fh).unwrap();
// seek to the offset specified, which is relative to the start of the file.
let resp = blob_reader.seek(io::SeekFrom::Start(offset as u64));
match resp {
Ok(pos) => {
debug_assert_eq!(offset as u64, pos);
}
Err(e) => {
warn!("failed to seek to offset {}: {}", offset, e);
reply.error(libc::EIO);
return;
}
}
// now with the blobreader seeked to this location, read size of data
let data: std::io::Result<Vec<u8>> =
blob_reader.bytes().take(size.try_into().unwrap()).collect();
match data {
// respond with the requested data
Ok(data) => reply.data(&data),
Err(e) => reply.error(e.raw_os_error().unwrap()),
}
}
#[tracing::instrument(skip_all, fields(rq.inode = ino))]
fn readlink(&mut self, _req: &Request<'_>, ino: u64, reply: fuser::ReplyData) {
if ino == fuser::FUSE_ROOT_ID {
reply.error(libc::ENOSYS);
return;
}
// lookup the inode
match *self.inode_tracker.get(ino).unwrap() {
InodeData::Directory(..) | InodeData::Regular(..) => {
reply.error(libc::EINVAL);
}
InodeData::Symlink(ref target) => reply.data(target),
}
}
}
fn reply_with_entry(reply: fuser::ReplyEntry, file_attr: &FileAttr) {
reply.entry(&Duration::MAX, file_attr, 1 /* TODO: generation */);
}
