The `wasmtime` crate currently lives in `crates/api` for historical
reasons, because we once called it `wasmtime-api` crate. This creates a
stumbling block for new contributors.
As discussed on Zulip, rename the directory to `crates/wasmtime`.
Several links were broken by line-breaks between the link caption and
the link itself. This commit fixes them by moving each on its own line.
Co-authored-by: k.bagrov <k.bagrov@g.nsu.ru>
If the scratch register is caller-saved, then it might appear in fixed
ranges because of call clobbers. Instead, use a register that's not
caller-saved and has no predefined use in the ABI.
SmallVec<[Value; 1]>, not as a Vec<Value>. This isn't a useful change for any
non-developer use of Cranelift, but it does significantly reduce the amount of
allocation "noise" seen when tuning the new backend pipeline as driven by
clif-util reading .clif files. In one case the number of malloc calls
declined by about 20% with this change.
In stark contrast with every reasonable architecture, X86-32 does not
pass any parameters in registers. Because of that we have to resort
to reading arguments from stack without being able to use the stack
slot machinery.
(This wouldn't have been avoidable even by pinning a register because
there is a trampoline in wasmtime with the C ABI that Cranelift needs
to be able to call.)
compilation.
This saves ~0.14% instruction count, ~0.18% allocated bytes, and ~1.5%
allocated blocks on a `clif-util wasm` compilation of `bz2.wasm` for
aarch64.
that only one value ever flows. Has been observed to improve generated code
run times by up to 8%. Compilation cost increases by about 0.6%, but up to 7%
total cost has been observed to be saved; iow it can be a significant win in
terms of compilation time, overall.
This is an incomplete version of a Cranelift IR interpreter: only a small subset of instructions are implemented and (known) missing parts are marked with TODO or FIXME.
* Introduce strongly-typed system primitives
This commit does a lot of reshuffling and even some more. It introduces
strongly-typed system primitives which are: `OsFile`, `OsDir`, `Stdio`,
and `OsOther`. Those primitives are separate structs now, each implementing
a subset of `Handle` methods, rather than all being an enumeration of some
supertype such as `OsHandle`. To summarise the structs:
* `OsFile` represents a regular file, and implements fd-ops
of `Handle` trait
* `OsDir` represents a directory, and primarily implements path-ops, plus
`readdir` and some common fd-ops such as `fdstat`, etc.
* `Stdio` represents a stdio handle, and implements a subset of fd-ops
such as `fdstat` _and_ `read_` and `write_vectored` calls
* `OsOther` currently represents anything else and implements a set similar
to that implemented by `Stdio`
This commit is effectively an experiment and an excercise into better
understanding what's going on for each OS resource/type under-the-hood.
It's meant to give us some intuition in order to move on with the idea
of having strongly-typed handles in WASI both in the syscall impl as well
as at the libc level.
Some more minor changes include making `OsHandle` represent an OS-specific
wrapper for a raw OS handle (Unix fd or Windows handle). Also, since `OsDir`
is tricky across OSes, we also have a supertype of `OsHandle` called
`OsDirHandle` which may store a `DIR*` stream pointer (mainly BSD). Last but not
least, the `Filetype` and `Rights` are now computed when the resource is created,
rather than every time we call `Handle::get_file_type` and `Handle::get_rights`.
Finally, in order to facilitate the latter, I've converted `EntryRights` into
`HandleRights` and pushed them into each `Handle` implementor.
* Do not adjust rights on Stdio
* Clean up testing for TTY and escaping writes
* Implement AsFile for dyn Handle
This cleans up a lot of repeating boilerplate code todo with
dynamic dispatch.
* Delegate definition of OsDir to OS-specific modules
Delegates defining `OsDir` struct to OS-specific modules (BSD, Linux,
Emscripten, Windows). This way, `OsDir` can safely re-use `OsHandle`
for raw OS handle storage, and can store some aux data such as an
initialized stream ptr in case of BSD. As a result, we can safely
get rid of `OsDirHandle` which IMHO was causing unnecessary noise and
overcomplicating the design. On the other hand, delegating definition
of `OsDir` to OS-specific modules isn't super clean in and of itself
either. Perhaps there's a better way of handling this?
* Check if filetype of OS handle matches WASI filetype when creating
It seems prudent to check if the passed in `File` instance is of
type matching that of the requested WASI filetype. In other words,
we'd like to avoid situations where `OsFile` is created from a
pipe.
* Make AsFile fallible
Return `EBADF` in `AsFile` in case a `Handle` cannot be made into
a `std::fs::File`.
* Remove unnecessary as_file conversion
* Remove unnecessary check for TTY for Stdio handle type
* Fix incorrect stdio ctors on Unix
* Split Stdio into three separate types: Stdin, Stdout, Stderr
* Rename PendingEntry::File to PendingEntry::OsHandle to avoid confusion
* Rename OsHandle to RawOsHandle
Also, since `RawOsHandle` on *nix doesn't need interior mutability
wrt the inner raw file descriptor, we can safely swap the `RawFd`
for `File` instance.
* Add docs explaining what OsOther is
* Allow for stdio to be non-character-device (e.g., piped)
* Return error on bad preopen rather than panic
This PR changes the aarch64 ABI implementation to use positive offsets
from SP, rather than negative offsets from FP, to refer to spill slots
and stack-local storage. This allows for better addressing-mode options,
and hence slightly better code: e.g., the unsigned scaled 12-bit offset
mode can be used to reach anywhere in a 32KB frame without extra
address-construction instructions, whereas negative offsets are limited
to a signed 9-bit unscaled mode (-256 bytes).
To enable this, the PR introduces a notion of "nominal SP offsets" as a
virtual addressing mode, lowered during the emission pass. The offsets
are relative to "SP after adjusting downward to allocate stack/spill
slots", but before pushing clobbers. This allows the addressing-mode
expressions to be generated before register allocation (or during it,
for spill/reload sequences).
To convert these offsets into *true* offsets from SP, we need to track
how much further SP is moved downward, and compensate for this. We do so
with "virtual SP offset adjustment" pseudo-instructions: these are seen
by the emission pass, and result in no instruction (0 byte output), but
update state that is now threaded through each instruction emission in
turn. In this way, we can push e.g. stack args for a call and adjust
the virtual SP offset, allowing reloads from nominal-SP-relative
spillslots while we do the argument setup with "real SP offsets" at the
same time.
Previously, the logic was wrong on two counts:
- It used the bits of the entire vector (e.g. i32x4 -> 128) instead of just the lane bits (e.g. i32x4 -> 32).
- It used the type of the first operand before it was bitcast to its correct type. Remember that, by default, vectors are handed around as i8x16 and we must bitcast them to their correct type for Cranelift's verifier; see https://github.com/bytecodealliance/wasmtime/issues/1147 for discussion on this. This fix simply uses the type of the instruction itself, which is equivalent and hopefully less fragile to any changes.
* Update `Table::grow`'s return to be the previous size
This brings it in line with `Memory::grow` and the `table.grow`
instruction which return the size of the table previously, not the size
of the table currently.
* Comment successful return
Previously, every call was lowered on AArch64 to a `call` instruction, which
takes a signed 26-bit PC-relative offset. Including the 2-bit left shift, this
gives a range of +/- 128 MB. Longer-distance offsets would cause an impossible
relocation record to be emitted (or rather, a record that a more sophisticated
linker would fix up by inserting a shim/veneer).
This commit adds a notion of "relocation distance" in the MachInst backends,
and provides this information for every call target and symbol reference. The
intent is that backends on architectures like AArch64, where there are different
offset sizes / addressing strategies to choose from, can either emit a regular
call or a load-64-bit-constant / call-indirect sequence, as necessary. This
avoids the need to implement complex linking behavior.
The MachInst driver code provides this information based on the "colocated" bit
in the CLIF symbol references, which appears to have been designed for this
purpose, or at least a similar one. Combined with the `use_colocated_libcalls`
setting, this allows client code to ensure that library calls can link to
library code at any location in the address space.
Separately, the `simplejit` example did not handle `Arm64Call`; rather than doing
so, it appears all that is necessary to get its tests to pass is to set the
`use_colocated_libcalls` flag to false, to make use of the above change. This
fixes the `libcall_function` unit-test in this crate.
* Improve output display of RunCommand
The previous use of Debug for displaying `print` and `run` results was less than clear.
* Avoid checking the types of vectors during trampoline construction
Because DataValue only understands `V128` vectors, we avoid type-checking vector values when constructing the trampoline arguments.
* Improve the documentation of the filetest `run` command
Adds an up-to-date example of how to use the `run` and `print` directives and includes an actual use of the new directives in a SIMD arithmetic filetest.
Previously we initialized trap handling (signals/etc) once-per-instance
but that's a bit too granular since we only need to do this as
one-time per-program initialization. This moves the initialization to
`Store` instead which means that we'll call this at least once per
thread, which some platforms may need (none currently do, they all only
need per-program initialization, but Fuchsia will need per-thread
initialization).
Looks like everything is in general passing now so it's probably time to
close#1521 and all other remaining tests that are failing are
classified under new more focused issues.
Closes#1521
