* wasi: avoid buffer underflow with shared memory This change fixes an issue identified when using wasi-threads to perform file reads. In order to maintain Rust safety guarantees in the presence of WebAssembly shared memory, which can be modified concurrently by any of the running threads, the WASI implementations of `fd_read` and `fd_pread` were given special code paths when shared memory is detected: in these cases, the data is first read into a host-limited buffer and then subsequently copied into linear memory. The problem was that the rather-complex logic for doing this "buffer then copy" idea for multiple IO vectors could fail due to buffer underflow. If, e.g., a read was limited by the host to 64K (or even if the read returned less than the total buffer size) the `UnsafeGuestSlice::copy_from_slice` logic would fail, complaining that the sizes of both buffers were unequal. This change both simplifies and fixes the logic: - only the first IO vector is filled; this could represent a performance penalty for threaded programs, but the "buffer then copy" idea already imposes a non-trivial overhead. This simplifies the logic, allowing us to... - resize the shared memory buffer to the exact number of bytes read * review: early return when no IO vectors passed to shared memory * fix: add empty RoFlags on early exit
1162 lines
36 KiB
Rust
1162 lines
36 KiB
Rust
use crate::file::{FileCaps, FileEntryExt, TableFileExt};
|
|
use crate::sched::{
|
|
subscription::{RwEventFlags, SubscriptionResult},
|
|
Poll, Userdata,
|
|
};
|
|
use crate::snapshots::preview_1::types as snapshot1_types;
|
|
use crate::snapshots::preview_1::wasi_snapshot_preview1::WasiSnapshotPreview1 as Snapshot1;
|
|
use crate::snapshots::preview_1::MAX_SHARED_BUFFER_SIZE;
|
|
use crate::{ErrorExt, WasiCtx};
|
|
use cap_std::time::Duration;
|
|
use std::collections::HashSet;
|
|
use std::convert::{TryFrom, TryInto};
|
|
use std::io::{IoSlice, IoSliceMut};
|
|
use std::ops::Deref;
|
|
use wiggle::GuestPtr;
|
|
|
|
wiggle::from_witx!({
|
|
witx: ["$WASI_ROOT/phases/old/snapshot_0/witx/wasi_unstable.witx"],
|
|
errors: { errno => trappable Error },
|
|
async: *,
|
|
wasmtime: false,
|
|
});
|
|
|
|
use types::Error;
|
|
|
|
impl ErrorExt for Error {
|
|
fn not_found() -> Self {
|
|
types::Errno::Noent.into()
|
|
}
|
|
fn too_big() -> Self {
|
|
types::Errno::TooBig.into()
|
|
}
|
|
fn badf() -> Self {
|
|
types::Errno::Badf.into()
|
|
}
|
|
fn exist() -> Self {
|
|
types::Errno::Exist.into()
|
|
}
|
|
fn illegal_byte_sequence() -> Self {
|
|
types::Errno::Ilseq.into()
|
|
}
|
|
fn invalid_argument() -> Self {
|
|
types::Errno::Inval.into()
|
|
}
|
|
fn io() -> Self {
|
|
types::Errno::Io.into()
|
|
}
|
|
fn name_too_long() -> Self {
|
|
types::Errno::Nametoolong.into()
|
|
}
|
|
fn not_dir() -> Self {
|
|
types::Errno::Notdir.into()
|
|
}
|
|
fn not_supported() -> Self {
|
|
types::Errno::Notsup.into()
|
|
}
|
|
fn overflow() -> Self {
|
|
types::Errno::Overflow.into()
|
|
}
|
|
fn range() -> Self {
|
|
types::Errno::Range.into()
|
|
}
|
|
fn seek_pipe() -> Self {
|
|
types::Errno::Spipe.into()
|
|
}
|
|
fn perm() -> Self {
|
|
types::Errno::Perm.into()
|
|
}
|
|
}
|
|
|
|
impl wiggle::GuestErrorType for types::Errno {
|
|
fn success() -> Self {
|
|
Self::Success
|
|
}
|
|
}
|
|
|
|
impl From<wiggle::GuestError> for Error {
|
|
fn from(err: wiggle::GuestError) -> Error {
|
|
snapshot1_types::Error::from(err).into()
|
|
}
|
|
}
|
|
|
|
impl From<snapshot1_types::Error> for Error {
|
|
fn from(error: snapshot1_types::Error) -> Error {
|
|
match error.downcast() {
|
|
Ok(errno) => Error::from(types::Errno::from(errno)),
|
|
Err(trap) => Error::trap(trap),
|
|
}
|
|
}
|
|
}
|
|
|
|
impl From<std::num::TryFromIntError> for Error {
|
|
fn from(_err: std::num::TryFromIntError) -> Error {
|
|
types::Errno::Overflow.into()
|
|
}
|
|
}
|
|
|
|
// Type conversions
|
|
// The vast majority of the types defined in `types` and `snapshot1_types` are identical. However,
|
|
// since they are defined in separate places for mechanical (wiggle) reasons, we need to manually
|
|
// define conversion functions between them.
|
|
// Below we have defined these functions as they are needed.
|
|
|
|
/// Fd is a newtype wrapper around u32. Unwrap and wrap it.
|
|
impl From<types::Fd> for snapshot1_types::Fd {
|
|
fn from(fd: types::Fd) -> snapshot1_types::Fd {
|
|
u32::from(fd).into()
|
|
}
|
|
}
|
|
/// Fd is a newtype wrapper around u32. Unwrap and wrap it.
|
|
impl From<snapshot1_types::Fd> for types::Fd {
|
|
fn from(fd: snapshot1_types::Fd) -> types::Fd {
|
|
u32::from(fd).into()
|
|
}
|
|
}
|
|
|
|
/// Trivial conversion between two c-style enums that have the exact same set of variants.
|
|
/// Could we do something unsafe and not list all these variants out? Probably, but doing
|
|
/// it this way doesn't bother me much. I copy-pasted the list of variants out of the
|
|
/// rendered rustdocs.
|
|
/// LLVM ought to compile these From impls into no-ops, inshallah
|
|
macro_rules! convert_enum {
|
|
($from:ty, $to:ty, $($var:ident),+) => {
|
|
impl From<$from> for $to {
|
|
fn from(e: $from) -> $to {
|
|
match e {
|
|
$( <$from>::$var => <$to>::$var, )+
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
convert_enum!(
|
|
snapshot1_types::Errno,
|
|
types::Errno,
|
|
Success,
|
|
TooBig,
|
|
Acces,
|
|
Addrinuse,
|
|
Addrnotavail,
|
|
Afnosupport,
|
|
Again,
|
|
Already,
|
|
Badf,
|
|
Badmsg,
|
|
Busy,
|
|
Canceled,
|
|
Child,
|
|
Connaborted,
|
|
Connrefused,
|
|
Connreset,
|
|
Deadlk,
|
|
Destaddrreq,
|
|
Dom,
|
|
Dquot,
|
|
Exist,
|
|
Fault,
|
|
Fbig,
|
|
Hostunreach,
|
|
Idrm,
|
|
Ilseq,
|
|
Inprogress,
|
|
Intr,
|
|
Inval,
|
|
Io,
|
|
Isconn,
|
|
Isdir,
|
|
Loop,
|
|
Mfile,
|
|
Mlink,
|
|
Msgsize,
|
|
Multihop,
|
|
Nametoolong,
|
|
Netdown,
|
|
Netreset,
|
|
Netunreach,
|
|
Nfile,
|
|
Nobufs,
|
|
Nodev,
|
|
Noent,
|
|
Noexec,
|
|
Nolck,
|
|
Nolink,
|
|
Nomem,
|
|
Nomsg,
|
|
Noprotoopt,
|
|
Nospc,
|
|
Nosys,
|
|
Notconn,
|
|
Notdir,
|
|
Notempty,
|
|
Notrecoverable,
|
|
Notsock,
|
|
Notsup,
|
|
Notty,
|
|
Nxio,
|
|
Overflow,
|
|
Ownerdead,
|
|
Perm,
|
|
Pipe,
|
|
Proto,
|
|
Protonosupport,
|
|
Prototype,
|
|
Range,
|
|
Rofs,
|
|
Spipe,
|
|
Srch,
|
|
Stale,
|
|
Timedout,
|
|
Txtbsy,
|
|
Xdev,
|
|
Notcapable
|
|
);
|
|
convert_enum!(
|
|
types::Clockid,
|
|
snapshot1_types::Clockid,
|
|
Realtime,
|
|
Monotonic,
|
|
ProcessCputimeId,
|
|
ThreadCputimeId
|
|
);
|
|
|
|
convert_enum!(
|
|
types::Advice,
|
|
snapshot1_types::Advice,
|
|
Normal,
|
|
Sequential,
|
|
Random,
|
|
Willneed,
|
|
Dontneed,
|
|
Noreuse
|
|
);
|
|
convert_enum!(
|
|
snapshot1_types::Filetype,
|
|
types::Filetype,
|
|
Directory,
|
|
BlockDevice,
|
|
CharacterDevice,
|
|
RegularFile,
|
|
SocketDgram,
|
|
SocketStream,
|
|
SymbolicLink,
|
|
Unknown
|
|
);
|
|
convert_enum!(types::Whence, snapshot1_types::Whence, Cur, End, Set);
|
|
|
|
/// Prestat isn't a c-style enum, its a union where the variant has a payload. Its the only one of
|
|
/// those we need to convert, so write it by hand.
|
|
impl From<snapshot1_types::Prestat> for types::Prestat {
|
|
fn from(p: snapshot1_types::Prestat) -> types::Prestat {
|
|
match p {
|
|
snapshot1_types::Prestat::Dir(d) => types::Prestat::Dir(d.into()),
|
|
}
|
|
}
|
|
}
|
|
|
|
/// Trivial conversion between two structs that have the exact same set of fields,
|
|
/// with recursive descent into the field types.
|
|
macro_rules! convert_struct {
|
|
($from:ty, $to:path, $($field:ident),+) => {
|
|
impl From<$from> for $to {
|
|
fn from(e: $from) -> $to {
|
|
$to {
|
|
$( $field: e.$field.into(), )+
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
convert_struct!(snapshot1_types::PrestatDir, types::PrestatDir, pr_name_len);
|
|
convert_struct!(
|
|
snapshot1_types::Fdstat,
|
|
types::Fdstat,
|
|
fs_filetype,
|
|
fs_rights_base,
|
|
fs_rights_inheriting,
|
|
fs_flags
|
|
);
|
|
|
|
/// Snapshot1 Filestat is incompatible with Snapshot0 Filestat - the nlink
|
|
/// field is u32 on this Filestat, and u64 on theirs. If you've got more than
|
|
/// 2^32 links I don't know what to tell you
|
|
impl From<snapshot1_types::Filestat> for types::Filestat {
|
|
fn from(f: snapshot1_types::Filestat) -> types::Filestat {
|
|
types::Filestat {
|
|
dev: f.dev.into(),
|
|
ino: f.ino.into(),
|
|
filetype: f.filetype.into(),
|
|
nlink: f.nlink.try_into().unwrap_or(u32::MAX),
|
|
size: f.size.into(),
|
|
atim: f.atim.into(),
|
|
mtim: f.mtim.into(),
|
|
ctim: f.ctim.into(),
|
|
}
|
|
}
|
|
}
|
|
|
|
/// Trivial conversion between two bitflags that have the exact same set of flags.
|
|
macro_rules! convert_flags {
|
|
($from:ty, $to:ty, $($flag:ident),+) => {
|
|
impl From<$from> for $to {
|
|
fn from(f: $from) -> $to {
|
|
let mut out = <$to>::empty();
|
|
$(
|
|
if f.contains(<$from>::$flag) {
|
|
out |= <$to>::$flag;
|
|
}
|
|
)+
|
|
out
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
/// Need to convert in both directions? This saves listing out the flags twice
|
|
macro_rules! convert_flags_bidirectional {
|
|
($from:ty, $to:ty, $($flag:tt)*) => {
|
|
convert_flags!($from, $to, $($flag)*);
|
|
convert_flags!($to, $from, $($flag)*);
|
|
}
|
|
}
|
|
|
|
convert_flags_bidirectional!(
|
|
snapshot1_types::Fdflags,
|
|
types::Fdflags,
|
|
APPEND,
|
|
DSYNC,
|
|
NONBLOCK,
|
|
RSYNC,
|
|
SYNC
|
|
);
|
|
convert_flags!(
|
|
types::Lookupflags,
|
|
snapshot1_types::Lookupflags,
|
|
SYMLINK_FOLLOW
|
|
);
|
|
convert_flags!(
|
|
types::Fstflags,
|
|
snapshot1_types::Fstflags,
|
|
ATIM,
|
|
ATIM_NOW,
|
|
MTIM,
|
|
MTIM_NOW
|
|
);
|
|
convert_flags!(
|
|
types::Oflags,
|
|
snapshot1_types::Oflags,
|
|
CREAT,
|
|
DIRECTORY,
|
|
EXCL,
|
|
TRUNC
|
|
);
|
|
convert_flags_bidirectional!(
|
|
types::Rights,
|
|
snapshot1_types::Rights,
|
|
FD_DATASYNC,
|
|
FD_READ,
|
|
FD_SEEK,
|
|
FD_FDSTAT_SET_FLAGS,
|
|
FD_SYNC,
|
|
FD_TELL,
|
|
FD_WRITE,
|
|
FD_ADVISE,
|
|
FD_ALLOCATE,
|
|
PATH_CREATE_DIRECTORY,
|
|
PATH_CREATE_FILE,
|
|
PATH_LINK_SOURCE,
|
|
PATH_LINK_TARGET,
|
|
PATH_OPEN,
|
|
FD_READDIR,
|
|
PATH_READLINK,
|
|
PATH_RENAME_SOURCE,
|
|
PATH_RENAME_TARGET,
|
|
PATH_FILESTAT_GET,
|
|
PATH_FILESTAT_SET_SIZE,
|
|
PATH_FILESTAT_SET_TIMES,
|
|
FD_FILESTAT_GET,
|
|
FD_FILESTAT_SET_SIZE,
|
|
FD_FILESTAT_SET_TIMES,
|
|
PATH_SYMLINK,
|
|
PATH_REMOVE_DIRECTORY,
|
|
PATH_UNLINK_FILE,
|
|
POLL_FD_READWRITE,
|
|
SOCK_SHUTDOWN
|
|
);
|
|
|
|
// This implementation, wherever possible, delegates directly to the Snapshot1 implementation,
|
|
// performing the no-op type conversions along the way.
|
|
#[wiggle::async_trait]
|
|
impl wasi_unstable::WasiUnstable for WasiCtx {
|
|
async fn args_get<'a>(
|
|
&mut self,
|
|
argv: &GuestPtr<'a, GuestPtr<'a, u8>>,
|
|
argv_buf: &GuestPtr<'a, u8>,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::args_get(self, argv, argv_buf).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn args_sizes_get(&mut self) -> Result<(types::Size, types::Size), Error> {
|
|
let s = Snapshot1::args_sizes_get(self).await?;
|
|
Ok(s)
|
|
}
|
|
|
|
async fn environ_get<'a>(
|
|
&mut self,
|
|
environ: &GuestPtr<'a, GuestPtr<'a, u8>>,
|
|
environ_buf: &GuestPtr<'a, u8>,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::environ_get(self, environ, environ_buf).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn environ_sizes_get(&mut self) -> Result<(types::Size, types::Size), Error> {
|
|
let s = Snapshot1::environ_sizes_get(self).await?;
|
|
Ok(s)
|
|
}
|
|
|
|
async fn clock_res_get(&mut self, id: types::Clockid) -> Result<types::Timestamp, Error> {
|
|
let t = Snapshot1::clock_res_get(self, id.into()).await?;
|
|
Ok(t)
|
|
}
|
|
|
|
async fn clock_time_get(
|
|
&mut self,
|
|
id: types::Clockid,
|
|
precision: types::Timestamp,
|
|
) -> Result<types::Timestamp, Error> {
|
|
let t = Snapshot1::clock_time_get(self, id.into(), precision).await?;
|
|
Ok(t)
|
|
}
|
|
|
|
async fn fd_advise(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
offset: types::Filesize,
|
|
len: types::Filesize,
|
|
advice: types::Advice,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::fd_advise(self, fd.into(), offset, len, advice.into()).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_allocate(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
offset: types::Filesize,
|
|
len: types::Filesize,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::fd_allocate(self, fd.into(), offset, len).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_close(&mut self, fd: types::Fd) -> Result<(), Error> {
|
|
Snapshot1::fd_close(self, fd.into()).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_datasync(&mut self, fd: types::Fd) -> Result<(), Error> {
|
|
Snapshot1::fd_datasync(self, fd.into()).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_fdstat_get(&mut self, fd: types::Fd) -> Result<types::Fdstat, Error> {
|
|
Ok(Snapshot1::fd_fdstat_get(self, fd.into()).await?.into())
|
|
}
|
|
|
|
async fn fd_fdstat_set_flags(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
flags: types::Fdflags,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::fd_fdstat_set_flags(self, fd.into(), flags.into()).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_fdstat_set_rights(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
fs_rights_base: types::Rights,
|
|
fs_rights_inheriting: types::Rights,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::fd_fdstat_set_rights(
|
|
self,
|
|
fd.into(),
|
|
fs_rights_base.into(),
|
|
fs_rights_inheriting.into(),
|
|
)
|
|
.await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_filestat_get(&mut self, fd: types::Fd) -> Result<types::Filestat, Error> {
|
|
Ok(Snapshot1::fd_filestat_get(self, fd.into()).await?.into())
|
|
}
|
|
|
|
async fn fd_filestat_set_size(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
size: types::Filesize,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::fd_filestat_set_size(self, fd.into(), size).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_filestat_set_times(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
atim: types::Timestamp,
|
|
mtim: types::Timestamp,
|
|
fst_flags: types::Fstflags,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::fd_filestat_set_times(self, fd.into(), atim, mtim, fst_flags.into()).await?;
|
|
Ok(())
|
|
}
|
|
|
|
// NOTE on fd_read, fd_pread, fd_write, fd_pwrite implementations:
|
|
// Because the arguments to these function sit behind GuestPtrs, they are not values we
|
|
// can convert and pass to the corresponding function in Snapshot1.
|
|
// Instead, we have copied the implementation of these functions from the Snapshot1 code.
|
|
// The implementations are identical, but the `types::` in scope locally is different.
|
|
// The bodies of these functions is mostly about converting the GuestPtr and types::-based
|
|
// representation to a std::io::IoSlice(Mut) representation.
|
|
|
|
async fn fd_read<'a>(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
iovs: &types::IovecArray<'a>,
|
|
) -> Result<types::Size, Error> {
|
|
let f = self
|
|
.table()
|
|
.get_file_mut(u32::from(fd))?
|
|
.get_cap_mut(FileCaps::READ)?;
|
|
|
|
let iovs: Vec<wiggle::GuestPtr<[u8]>> = iovs
|
|
.iter()
|
|
.map(|iov_ptr| {
|
|
let iov_ptr = iov_ptr?;
|
|
let iov: types::Iovec = iov_ptr.read()?;
|
|
Ok(iov.buf.as_array(iov.buf_len))
|
|
})
|
|
.collect::<Result<_, Error>>()?;
|
|
|
|
// If the first iov structure is from shared memory we can safely assume
|
|
// all the rest will be. We then read into memory based on the memory's
|
|
// shared-ness:
|
|
// - if not shared, we copy directly into the Wasm memory
|
|
// - if shared, we use an intermediate buffer; this avoids Rust unsafety
|
|
// due to holding on to a `&mut [u8]` of Wasm memory when we cannot
|
|
// guarantee the `&mut` exclusivity--other threads could be modifying
|
|
// the data as this functions writes to it. Though likely there is no
|
|
// issue with OS writing to io structs in multi-threaded scenarios,
|
|
// since we do not know here if `&dyn WasiFile` does anything else
|
|
// (e.g., read), we cautiously incur some performance overhead by
|
|
// copying twice.
|
|
let is_shared_memory = iovs
|
|
.iter()
|
|
.next()
|
|
.and_then(|s| Some(s.is_shared_memory()))
|
|
.unwrap_or(false);
|
|
let bytes_read: u64 = if is_shared_memory {
|
|
// For shared memory, read into an intermediate buffer. Only the
|
|
// first iov will be filled and even then the read is capped by the
|
|
// `MAX_SHARED_BUFFER_SIZE`, so users are expected to re-call.
|
|
let iov = iovs.into_iter().next();
|
|
if let Some(iov) = iov {
|
|
let mut buffer = vec![0; (iov.len() as usize).min(MAX_SHARED_BUFFER_SIZE)];
|
|
let bytes_read = f.read_vectored(&mut [IoSliceMut::new(&mut buffer)]).await?;
|
|
iov.get_range(0..bytes_read.try_into()?)
|
|
.expect("it should always be possible to slice the iov smaller")
|
|
.copy_from_slice(&buffer[0..bytes_read.try_into()?])?;
|
|
bytes_read
|
|
} else {
|
|
return Ok(0);
|
|
}
|
|
} else {
|
|
// Convert all of the unsafe guest slices to safe ones--this uses
|
|
// Wiggle's internal borrow checker to ensure no overlaps. We assume
|
|
// here that, because the memory is not shared, there are no other
|
|
// threads to access it while it is written to.
|
|
let mut guest_slices: Vec<wiggle::GuestSliceMut<u8>> = iovs
|
|
.into_iter()
|
|
.map(|iov| Ok(iov.as_slice_mut()?.unwrap()))
|
|
.collect::<Result<_, Error>>()?;
|
|
|
|
// Read directly into the Wasm memory.
|
|
let mut ioslices: Vec<IoSliceMut> = guest_slices
|
|
.iter_mut()
|
|
.map(|s| IoSliceMut::new(&mut *s))
|
|
.collect();
|
|
f.read_vectored(&mut ioslices).await?
|
|
};
|
|
|
|
Ok(types::Size::try_from(bytes_read)?)
|
|
}
|
|
|
|
async fn fd_pread<'a>(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
iovs: &types::IovecArray<'a>,
|
|
offset: types::Filesize,
|
|
) -> Result<types::Size, Error> {
|
|
let f = self
|
|
.table()
|
|
.get_file_mut(u32::from(fd))?
|
|
.get_cap_mut(FileCaps::READ | FileCaps::SEEK)?;
|
|
|
|
let iovs: Vec<wiggle::GuestPtr<[u8]>> = iovs
|
|
.iter()
|
|
.map(|iov_ptr| {
|
|
let iov_ptr = iov_ptr?;
|
|
let iov: types::Iovec = iov_ptr.read()?;
|
|
Ok(iov.buf.as_array(iov.buf_len))
|
|
})
|
|
.collect::<Result<_, Error>>()?;
|
|
|
|
// If the first iov structure is from shared memory we can safely assume
|
|
// all the rest will be. We then read into memory based on the memory's
|
|
// shared-ness:
|
|
// - if not shared, we copy directly into the Wasm memory
|
|
// - if shared, we use an intermediate buffer; this avoids Rust unsafety
|
|
// due to holding on to a `&mut [u8]` of Wasm memory when we cannot
|
|
// guarantee the `&mut` exclusivity--other threads could be modifying
|
|
// the data as this functions writes to it. Though likely there is no
|
|
// issue with OS writing to io structs in multi-threaded scenarios,
|
|
// since we do not know here if `&dyn WasiFile` does anything else
|
|
// (e.g., read), we cautiously incur some performance overhead by
|
|
// copying twice.
|
|
let is_shared_memory = iovs
|
|
.iter()
|
|
.next()
|
|
.and_then(|s| Some(s.is_shared_memory()))
|
|
.unwrap_or(false);
|
|
let bytes_read: u64 = if is_shared_memory {
|
|
// For shared memory, read into an intermediate buffer. Only the
|
|
// first iov will be filled and even then the read is capped by the
|
|
// `MAX_SHARED_BUFFER_SIZE`, so users are expected to re-call.
|
|
let iov = iovs.into_iter().next();
|
|
if let Some(iov) = iov {
|
|
let mut buffer = vec![0; (iov.len() as usize).min(MAX_SHARED_BUFFER_SIZE)];
|
|
let bytes_read = f
|
|
.read_vectored_at(&mut [IoSliceMut::new(&mut buffer)], offset)
|
|
.await?;
|
|
iov.get_range(0..bytes_read.try_into()?)
|
|
.expect("it should always be possible to slice the iov smaller")
|
|
.copy_from_slice(&buffer[0..bytes_read.try_into()?])?;
|
|
bytes_read
|
|
} else {
|
|
return Ok(0);
|
|
}
|
|
} else {
|
|
// Convert all of the unsafe guest slices to safe ones--this uses
|
|
// Wiggle's internal borrow checker to ensure no overlaps. We assume
|
|
// here that, because the memory is not shared, there are no other
|
|
// threads to access it while it is written to.
|
|
let mut guest_slices: Vec<wiggle::GuestSliceMut<u8>> = iovs
|
|
.into_iter()
|
|
.map(|iov| Ok(iov.as_slice_mut()?.unwrap()))
|
|
.collect::<Result<_, Error>>()?;
|
|
|
|
// Read directly into the Wasm memory.
|
|
let mut ioslices: Vec<IoSliceMut> = guest_slices
|
|
.iter_mut()
|
|
.map(|s| IoSliceMut::new(&mut *s))
|
|
.collect();
|
|
f.read_vectored_at(&mut ioslices, offset).await?
|
|
};
|
|
|
|
Ok(types::Size::try_from(bytes_read)?)
|
|
}
|
|
|
|
async fn fd_write<'a>(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
ciovs: &types::CiovecArray<'a>,
|
|
) -> Result<types::Size, Error> {
|
|
let f = self
|
|
.table()
|
|
.get_file_mut(u32::from(fd))?
|
|
.get_cap_mut(FileCaps::WRITE)?;
|
|
|
|
let guest_slices: Vec<wiggle::GuestCow<u8>> = ciovs
|
|
.iter()
|
|
.map(|iov_ptr| {
|
|
let iov_ptr = iov_ptr?;
|
|
let iov: types::Ciovec = iov_ptr.read()?;
|
|
Ok(iov.buf.as_array(iov.buf_len).as_cow()?)
|
|
})
|
|
.collect::<Result<_, Error>>()?;
|
|
|
|
let ioslices: Vec<IoSlice> = guest_slices
|
|
.iter()
|
|
.map(|s| IoSlice::new(s.deref()))
|
|
.collect();
|
|
let bytes_written = f.write_vectored(&ioslices).await?;
|
|
|
|
Ok(types::Size::try_from(bytes_written)?)
|
|
}
|
|
|
|
async fn fd_pwrite<'a>(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
ciovs: &types::CiovecArray<'a>,
|
|
offset: types::Filesize,
|
|
) -> Result<types::Size, Error> {
|
|
let f = self
|
|
.table()
|
|
.get_file_mut(u32::from(fd))?
|
|
.get_cap_mut(FileCaps::WRITE | FileCaps::SEEK)?;
|
|
|
|
let guest_slices: Vec<wiggle::GuestCow<u8>> = ciovs
|
|
.iter()
|
|
.map(|iov_ptr| {
|
|
let iov_ptr = iov_ptr?;
|
|
let iov: types::Ciovec = iov_ptr.read()?;
|
|
Ok(iov.buf.as_array(iov.buf_len).as_cow()?)
|
|
})
|
|
.collect::<Result<_, Error>>()?;
|
|
|
|
let ioslices: Vec<IoSlice> = guest_slices
|
|
.iter()
|
|
.map(|s| IoSlice::new(s.deref()))
|
|
.collect();
|
|
let bytes_written = f.write_vectored_at(&ioslices, offset).await?;
|
|
|
|
Ok(types::Size::try_from(bytes_written)?)
|
|
}
|
|
|
|
async fn fd_prestat_get(&mut self, fd: types::Fd) -> Result<types::Prestat, Error> {
|
|
Ok(Snapshot1::fd_prestat_get(self, fd.into()).await?.into())
|
|
}
|
|
|
|
async fn fd_prestat_dir_name<'a>(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
path: &GuestPtr<'a, u8>,
|
|
path_max_len: types::Size,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::fd_prestat_dir_name(self, fd.into(), path, path_max_len).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_renumber(&mut self, from: types::Fd, to: types::Fd) -> Result<(), Error> {
|
|
Snapshot1::fd_renumber(self, from.into(), to.into()).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_seek(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
offset: types::Filedelta,
|
|
whence: types::Whence,
|
|
) -> Result<types::Filesize, Error> {
|
|
Ok(Snapshot1::fd_seek(self, fd.into(), offset, whence.into()).await?)
|
|
}
|
|
|
|
async fn fd_sync(&mut self, fd: types::Fd) -> Result<(), Error> {
|
|
Snapshot1::fd_sync(self, fd.into()).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn fd_tell(&mut self, fd: types::Fd) -> Result<types::Filesize, Error> {
|
|
Ok(Snapshot1::fd_tell(self, fd.into()).await?)
|
|
}
|
|
|
|
async fn fd_readdir<'a>(
|
|
&mut self,
|
|
fd: types::Fd,
|
|
buf: &GuestPtr<'a, u8>,
|
|
buf_len: types::Size,
|
|
cookie: types::Dircookie,
|
|
) -> Result<types::Size, Error> {
|
|
Ok(Snapshot1::fd_readdir(self, fd.into(), buf, buf_len, cookie).await?)
|
|
}
|
|
|
|
async fn path_create_directory<'a>(
|
|
&mut self,
|
|
dirfd: types::Fd,
|
|
path: &GuestPtr<'a, str>,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::path_create_directory(self, dirfd.into(), path).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn path_filestat_get<'a>(
|
|
&mut self,
|
|
dirfd: types::Fd,
|
|
flags: types::Lookupflags,
|
|
path: &GuestPtr<'a, str>,
|
|
) -> Result<types::Filestat, Error> {
|
|
Ok(
|
|
Snapshot1::path_filestat_get(self, dirfd.into(), flags.into(), path)
|
|
.await?
|
|
.into(),
|
|
)
|
|
}
|
|
|
|
async fn path_filestat_set_times<'a>(
|
|
&mut self,
|
|
dirfd: types::Fd,
|
|
flags: types::Lookupflags,
|
|
path: &GuestPtr<'a, str>,
|
|
atim: types::Timestamp,
|
|
mtim: types::Timestamp,
|
|
fst_flags: types::Fstflags,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::path_filestat_set_times(
|
|
self,
|
|
dirfd.into(),
|
|
flags.into(),
|
|
path,
|
|
atim,
|
|
mtim,
|
|
fst_flags.into(),
|
|
)
|
|
.await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn path_link<'a>(
|
|
&mut self,
|
|
src_fd: types::Fd,
|
|
src_flags: types::Lookupflags,
|
|
src_path: &GuestPtr<'a, str>,
|
|
target_fd: types::Fd,
|
|
target_path: &GuestPtr<'a, str>,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::path_link(
|
|
self,
|
|
src_fd.into(),
|
|
src_flags.into(),
|
|
src_path,
|
|
target_fd.into(),
|
|
target_path,
|
|
)
|
|
.await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn path_open<'a>(
|
|
&mut self,
|
|
dirfd: types::Fd,
|
|
dirflags: types::Lookupflags,
|
|
path: &GuestPtr<'a, str>,
|
|
oflags: types::Oflags,
|
|
fs_rights_base: types::Rights,
|
|
fs_rights_inheriting: types::Rights,
|
|
fdflags: types::Fdflags,
|
|
) -> Result<types::Fd, Error> {
|
|
Ok(Snapshot1::path_open(
|
|
self,
|
|
dirfd.into(),
|
|
dirflags.into(),
|
|
path,
|
|
oflags.into(),
|
|
fs_rights_base.into(),
|
|
fs_rights_inheriting.into(),
|
|
fdflags.into(),
|
|
)
|
|
.await?
|
|
.into())
|
|
}
|
|
|
|
async fn path_readlink<'a>(
|
|
&mut self,
|
|
dirfd: types::Fd,
|
|
path: &GuestPtr<'a, str>,
|
|
buf: &GuestPtr<'a, u8>,
|
|
buf_len: types::Size,
|
|
) -> Result<types::Size, Error> {
|
|
Ok(Snapshot1::path_readlink(self, dirfd.into(), path, buf, buf_len).await?)
|
|
}
|
|
|
|
async fn path_remove_directory<'a>(
|
|
&mut self,
|
|
dirfd: types::Fd,
|
|
path: &GuestPtr<'a, str>,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::path_remove_directory(self, dirfd.into(), path).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn path_rename<'a>(
|
|
&mut self,
|
|
src_fd: types::Fd,
|
|
src_path: &GuestPtr<'a, str>,
|
|
dest_fd: types::Fd,
|
|
dest_path: &GuestPtr<'a, str>,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::path_rename(self, src_fd.into(), src_path, dest_fd.into(), dest_path).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn path_symlink<'a>(
|
|
&mut self,
|
|
src_path: &GuestPtr<'a, str>,
|
|
dirfd: types::Fd,
|
|
dest_path: &GuestPtr<'a, str>,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::path_symlink(self, src_path, dirfd.into(), dest_path).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn path_unlink_file<'a>(
|
|
&mut self,
|
|
dirfd: types::Fd,
|
|
path: &GuestPtr<'a, str>,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::path_unlink_file(self, dirfd.into(), path).await?;
|
|
Ok(())
|
|
}
|
|
|
|
// NOTE on poll_oneoff implementation:
|
|
// Like fd_write and friends, the arguments and return values are behind GuestPtrs,
|
|
// so they are not values we can convert and pass to the poll_oneoff in Snapshot1.
|
|
// Instead, we have copied the implementation of these functions from the Snapshot1 code.
|
|
// The implementations are identical, but the `types::` in scope locally is different.
|
|
// The bodies of these functions is mostly about converting the GuestPtr and types::-based
|
|
// representation to use the Poll abstraction.
|
|
async fn poll_oneoff<'a>(
|
|
&mut self,
|
|
subs: &GuestPtr<'a, types::Subscription>,
|
|
events: &GuestPtr<'a, types::Event>,
|
|
nsubscriptions: types::Size,
|
|
) -> Result<types::Size, Error> {
|
|
if nsubscriptions == 0 {
|
|
return Err(Error::invalid_argument().context("nsubscriptions must be nonzero"));
|
|
}
|
|
|
|
// Special-case a `poll_oneoff` which is just sleeping on a single
|
|
// relative timer event, such as what WASI libc uses to implement sleep
|
|
// functions. This supports all clock IDs, because POSIX says that
|
|
// `clock_settime` doesn't effect relative sleeps.
|
|
if nsubscriptions == 1 {
|
|
let sub = subs.read()?;
|
|
if let types::SubscriptionU::Clock(clocksub) = sub.u {
|
|
if !clocksub
|
|
.flags
|
|
.contains(types::Subclockflags::SUBSCRIPTION_CLOCK_ABSTIME)
|
|
{
|
|
self.sched
|
|
.sleep(Duration::from_nanos(clocksub.timeout))
|
|
.await?;
|
|
events.write(types::Event {
|
|
userdata: sub.userdata,
|
|
error: types::Errno::Success,
|
|
type_: types::Eventtype::Clock,
|
|
fd_readwrite: fd_readwrite_empty(),
|
|
})?;
|
|
return Ok(1);
|
|
}
|
|
}
|
|
}
|
|
|
|
let table = &mut self.table;
|
|
let mut sub_fds: HashSet<types::Fd> = HashSet::new();
|
|
// We need these refmuts to outlive Poll, which will hold the &mut dyn WasiFile inside
|
|
let mut reads: Vec<(u32, Userdata)> = Vec::new();
|
|
let mut writes: Vec<(u32, Userdata)> = Vec::new();
|
|
let mut poll = Poll::new();
|
|
|
|
let subs = subs.as_array(nsubscriptions);
|
|
for sub_elem in subs.iter() {
|
|
let sub_ptr = sub_elem?;
|
|
let sub = sub_ptr.read()?;
|
|
match sub.u {
|
|
types::SubscriptionU::Clock(clocksub) => match clocksub.id {
|
|
types::Clockid::Monotonic => {
|
|
let clock = self.clocks.monotonic.deref();
|
|
let precision = Duration::from_nanos(clocksub.precision);
|
|
let duration = Duration::from_nanos(clocksub.timeout);
|
|
let deadline = if clocksub
|
|
.flags
|
|
.contains(types::Subclockflags::SUBSCRIPTION_CLOCK_ABSTIME)
|
|
{
|
|
self.clocks
|
|
.creation_time
|
|
.checked_add(duration)
|
|
.ok_or_else(|| Error::overflow().context("deadline"))?
|
|
} else {
|
|
clock
|
|
.now(precision)
|
|
.checked_add(duration)
|
|
.ok_or_else(|| Error::overflow().context("deadline"))?
|
|
};
|
|
poll.subscribe_monotonic_clock(
|
|
clock,
|
|
deadline,
|
|
precision,
|
|
sub.userdata.into(),
|
|
)
|
|
}
|
|
_ => Err(Error::invalid_argument()
|
|
.context("timer subscriptions only support monotonic timer"))?,
|
|
},
|
|
types::SubscriptionU::FdRead(readsub) => {
|
|
let fd = readsub.file_descriptor;
|
|
if sub_fds.contains(&fd) {
|
|
return Err(Error::invalid_argument()
|
|
.context("Fd can be subscribed to at most once per poll"));
|
|
} else {
|
|
sub_fds.insert(fd);
|
|
}
|
|
table
|
|
.get_file_mut(u32::from(fd))?
|
|
.get_cap_mut(FileCaps::POLL_READWRITE)?;
|
|
reads.push((u32::from(fd), sub.userdata.into()));
|
|
}
|
|
types::SubscriptionU::FdWrite(writesub) => {
|
|
let fd = writesub.file_descriptor;
|
|
if sub_fds.contains(&fd) {
|
|
return Err(Error::invalid_argument()
|
|
.context("Fd can be subscribed to at most once per poll"));
|
|
} else {
|
|
sub_fds.insert(fd);
|
|
}
|
|
table
|
|
.get_file_mut(u32::from(fd))?
|
|
.get_cap_mut(FileCaps::POLL_READWRITE)?;
|
|
writes.push((u32::from(fd), sub.userdata.into()));
|
|
}
|
|
}
|
|
}
|
|
|
|
self.sched.poll_oneoff(&mut poll).await?;
|
|
|
|
let results = poll.results();
|
|
let num_results = results.len();
|
|
assert!(
|
|
num_results <= nsubscriptions as usize,
|
|
"results exceeds subscriptions"
|
|
);
|
|
let events = events.as_array(
|
|
num_results
|
|
.try_into()
|
|
.expect("not greater than nsubscriptions"),
|
|
);
|
|
for ((result, userdata), event_elem) in results.into_iter().zip(events.iter()) {
|
|
let event_ptr = event_elem?;
|
|
let userdata: types::Userdata = userdata.into();
|
|
event_ptr.write(match result {
|
|
SubscriptionResult::Read(r) => {
|
|
let type_ = types::Eventtype::FdRead;
|
|
match r {
|
|
Ok((nbytes, flags)) => types::Event {
|
|
userdata,
|
|
error: types::Errno::Success,
|
|
type_,
|
|
fd_readwrite: types::EventFdReadwrite {
|
|
nbytes,
|
|
flags: types::Eventrwflags::from(&flags),
|
|
},
|
|
},
|
|
Err(e) => types::Event {
|
|
userdata,
|
|
error: types::Errno::from(e.downcast().map_err(Error::trap)?),
|
|
type_,
|
|
fd_readwrite: fd_readwrite_empty(),
|
|
},
|
|
}
|
|
}
|
|
SubscriptionResult::Write(r) => {
|
|
let type_ = types::Eventtype::FdWrite;
|
|
match r {
|
|
Ok((nbytes, flags)) => types::Event {
|
|
userdata,
|
|
error: types::Errno::Success,
|
|
type_,
|
|
fd_readwrite: types::EventFdReadwrite {
|
|
nbytes,
|
|
flags: types::Eventrwflags::from(&flags),
|
|
},
|
|
},
|
|
Err(e) => types::Event {
|
|
userdata,
|
|
error: types::Errno::from(e.downcast().map_err(Error::trap)?),
|
|
type_,
|
|
fd_readwrite: fd_readwrite_empty(),
|
|
},
|
|
}
|
|
}
|
|
SubscriptionResult::MonotonicClock(r) => {
|
|
let type_ = types::Eventtype::Clock;
|
|
types::Event {
|
|
userdata,
|
|
error: match r {
|
|
Ok(()) => types::Errno::Success,
|
|
Err(e) => types::Errno::from(e.downcast().map_err(Error::trap)?),
|
|
},
|
|
type_,
|
|
fd_readwrite: fd_readwrite_empty(),
|
|
}
|
|
}
|
|
})?;
|
|
}
|
|
|
|
Ok(num_results.try_into().expect("results fit into memory"))
|
|
}
|
|
|
|
async fn proc_exit(&mut self, status: types::Exitcode) -> anyhow::Error {
|
|
Snapshot1::proc_exit(self, status).await
|
|
}
|
|
|
|
async fn proc_raise(&mut self, _sig: types::Signal) -> Result<(), Error> {
|
|
Err(Error::trap(anyhow::Error::msg("proc_raise unsupported")))
|
|
}
|
|
|
|
async fn sched_yield(&mut self) -> Result<(), Error> {
|
|
Snapshot1::sched_yield(self).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn random_get<'a>(
|
|
&mut self,
|
|
buf: &GuestPtr<'a, u8>,
|
|
buf_len: types::Size,
|
|
) -> Result<(), Error> {
|
|
Snapshot1::random_get(self, buf, buf_len).await?;
|
|
Ok(())
|
|
}
|
|
|
|
async fn sock_recv<'a>(
|
|
&mut self,
|
|
_fd: types::Fd,
|
|
_ri_data: &types::IovecArray<'a>,
|
|
_ri_flags: types::Riflags,
|
|
) -> Result<(types::Size, types::Roflags), Error> {
|
|
Err(Error::trap(anyhow::Error::msg("sock_recv unsupported")))
|
|
}
|
|
|
|
async fn sock_send<'a>(
|
|
&mut self,
|
|
_fd: types::Fd,
|
|
_si_data: &types::CiovecArray<'a>,
|
|
_si_flags: types::Siflags,
|
|
) -> Result<types::Size, Error> {
|
|
Err(Error::trap(anyhow::Error::msg("sock_send unsupported")))
|
|
}
|
|
|
|
async fn sock_shutdown(&mut self, _fd: types::Fd, _how: types::Sdflags) -> Result<(), Error> {
|
|
Err(Error::trap(anyhow::Error::msg("sock_shutdown unsupported")))
|
|
}
|
|
}
|
|
|
|
impl From<&RwEventFlags> for types::Eventrwflags {
|
|
fn from(flags: &RwEventFlags) -> types::Eventrwflags {
|
|
let mut out = types::Eventrwflags::empty();
|
|
if flags.contains(RwEventFlags::HANGUP) {
|
|
out = out | types::Eventrwflags::FD_READWRITE_HANGUP;
|
|
}
|
|
out
|
|
}
|
|
}
|
|
|
|
fn fd_readwrite_empty() -> types::EventFdReadwrite {
|
|
types::EventFdReadwrite {
|
|
nbytes: 0,
|
|
flags: types::Eventrwflags::empty(),
|
|
}
|
|
}
|