Commit Graph

58 Commits

Author SHA1 Message Date
Jakub Konka
5c51940100 Refactor use of Refs and RefMuts in wasi-common (#1412)
* Refactor use of Refs and RefMuts in wasi-common

This commit refactors the use of `Ref`s and `RefMut`s in `wasi-common`.
Now, `Entry` is stored behind an `Rc` inside the `EntryTable`. The `Entry`
itself on the other hand now stores rights behind a `RefCell` and the
descriptor as `Rc<RefCell<..>>` combo to enable easy reference tracking
and interior mutability which is required down the line in a couple of
syscalls. In essence, this implies that we no longer have need for
mutable accessor to `Entry` from `WasiCtx`, and so all related methods
go away (`get_entry_mut`, etc.).

While here, I've also simplified handling and aggregating of rights on
the `Entry` object. Instead of storing base and inheriting rights as
separate fields, they are now aggregated into one struct `EntryRights`
which features convenient constructors for each possible combination; i.e.,
when only base rights are set, or both base and inheriting are set, or
both are left as empty. Since we do need to be able to mutate those
rights down the line in `fdstat_set_rights` syscall, this object
is kept behind a `RefCell` (note no `Rc` since we don't need to pass it
around anywhere).

The descriptor field in `Entry` is now kept behind `Rc<RefCell<..>>` combo
since we not only need to mutate it down the line, but we also need to
be able to pass it around (as part of the machinery making `poll_oneoff`
work).

I've also removed `as_file` and `try_clone` methods on `Descriptor` struct
since they were adding more noise than necessary, and making them work
with `Rc` was unnecessarily complicated.

Finally, I've converted the `get_dir_from_os_handle` function into a
method attached to the `OsHandle` itself, called `dir_stream`. IMHO,
it makes more sense to have it there directly as a method than as a separate
function.

* Use Cell for types that are Copy
2020-03-27 09:34:52 +01:00
Alex Crichton
a628dc315e Shuffle around the wiggle crates (#1414)
* Shuffle around the wiggle crates

This commit reorganizes the wiggle crates slightly by performing the
following transforms:

* The `crates/wiggle` crate, previously named `wiggle`, was moved to
  `crates/wiggle/crates/macro` and is renamed to `wiggle-macro`.

* The `crates/wiggle/crates/runtime` crate, previously named
  `wiggle-runtime`, was moved to `crates/wiggle` and is renamed to
  `wiggle`.

* The new `wiggle` crate depends on `wiggle-macro` and reexports the macro.

The goal here is that consumers only deal with the `wiggle` crate
itself. No more crates depend on `wiggle-runtime` and all dependencies
are entirely on just the `wiggle` crate.

* Remove the `crates/wiggle/crates` directory

Move everything into `crates/wiggle` directly, like `wasi-common`

* Add wiggle-macro to test-all script

* Fixup a test
2020-03-26 18:34:50 -05:00
Pat Hickey
bc1a11435e wiggle: emit a metadata module containing witx document (#1387)
* wiggle: emit a metadata module containing witx document

* wiggle: put metadata module behind a wiggle_metadata feature

* wasi-common: add wiggle_metadata feature and optional witx dep

* refactor according to alex's advice

* wasi-common: make snapshots pub

* wasi-common: i do need a wiggle_metadata feature to be available

* Tweak features and such

* wiggle: fix tests by passing metadata flag to wiggle-runtime

* wiggle: need to move wiggle-runtime to a non-dev dependency

so that the feature resolves for external users of the crates

Co-authored-by: Alex Crichton <alex@alexcrichton.com>
2020-03-25 14:57:44 -05:00
Jakub Konka
1f6890e070 [wasi-common] Clean up the use of mutable Entry (#1395)
* Clean up the use of mutable Entry

Until now, several syscalls including `fd_pwrite` etc. were relying on
mutating `&mut Entry` by mutating its inner file handle. This is
unnecessary in almost all cases since all methods mutating `std::fs::File`
in Rust's libstd are also implemented for `&std::fs::File`.

While here, I've also modified `OsHandle` in BSD to include `RefCell<Option<Dir>>`
rather than `Option<Mutex<Dir>>` as was until now. While `RefCell`
could easily be replaced with `RefCell`, since going multithreading
will require a lot of (probably even) conceptual changes to `wasi-common`,
I thought it'd be best not to mix single- with multithreading contexts
and swap all places at once when it comes to it.

I've also had to make some modifications to virtual FS which mainly
swapped mutability for interior mutability in places.

* Use one-liners wherever convenient
2020-03-25 14:00:52 +01:00
Dan Gohman
2fdc7f1a8e [wiggle] Don't make generated structs and unions Copy.
Some structs and unions are large enough that making them `Copy` isn't
ideal. wasi-common only needed `Copy` in a few places that were easy to
fix. `SubscriptionClock` is 32 bytes, so it's not a bad a idea to pass
it by reference anyway.
2020-03-23 18:01:25 +01:00
Dan Gohman
ff18c5b05c [wasi-common] Fix a warning about unreachable code. (#1374)
`exit` doesn't return, so code after it is unreachable.
2020-03-20 16:28:27 -07:00
Pat Hickey
73fe49cd65 wasi-common: update trait methods to take &GuestPtr args. 2020-03-20 14:14:47 -07:00
Jakub Konka
32595faba5 It's wiggle time! (#1202)
* Use wiggle in place of wig in wasi-common

This is a rather massive commit that introduces `wiggle` into the
picture. We still use `wig`'s macro in `old` snapshot and to generate
`wasmtime-wasi` glue, but everything else is now autogenerated by `wiggle`.
In summary, thanks to `wiggle`, we no longer need to worry about
serialising and deserialising to and from the guest memory, and
all guest (WASI) types are now proper idiomatic Rust types.

While we're here, in preparation for the ephemeral snapshot, I went
ahead and reorganised the internal structure of the crate. Instead of
modules like `hostcalls_impl` or `hostcalls_impl::fs`, the structure
now resembles that in ephemeral with modules like `path`, `fd`, etc.
Now, I'm not requiring we leave it like this, but I reckon it looks
cleaner this way after all.

* Fix wig to use new first-class access to caller's mem

* Ignore warning in proc_exit for the moment

* Group unsafes together in args and environ calls

* Simplify pwrite; more unsafe blocks

* Simplify fd_read

* Bundle up unsafes in fd_readdir

* Simplify fd_write

* Add comment to path_readlink re zero-len buffers

* Simplify unsafes in random_get

* Hide GuestPtr<str> to &str in path::get

* Rewrite pread and pwrite using SeekFrom and read/write_vectored

I've left the implementation of VirtualFs pretty much untouched
as I don't feel that comfortable in changing the API too much.
Having said that, I reckon `pread` and `pwrite` could be refactored
out, and `preadv` and `pwritev` could be entirely rewritten using
`seek` and `read_vectored` and `write_vectored`.

* Add comment about VirtFs unsafety

* Fix all mentions of FdEntry to Entry

* Fix warnings on Win

* Add aux struct EntryTable responsible for Fds and Entries

This commit adds aux struct `EntryTable` which is private to `WasiCtx`
and is basically responsible for `Fd` alloc/dealloc as well as storing
matching `Entry`s. This struct is entirely private to `WasiCtx` and
as such as should remain transparent to `WasiCtx` users.

* Remove redundant check for empty buffer in path_readlink

* Preserve and rewind file cursor in pread/pwrite

* Use GuestPtr<[u8]>::copy_from_slice wherever copying bytes directly

* Use GuestPtr<[u8]>::copy_from_slice in fd_readdir

* Clean up unsafes around WasiCtx accessors

* Fix bugs in args_get and environ_get

* Fix conflicts after rebase
2020-03-20 21:54:44 +01:00