The simplifier was performing an optimization to replace bitselect
with vselect if the all bytes of the condition mask could be shown
to be all ones or all zeros.
This optimization only ever made any difference in codegen on the
x64 target. Therefore, move this optimization to the x64 back-end
and perform it in ISLE instead. Resulting codegen should be
unchanged, with slightly improved compile time.
This also eliminates a few endian-dependent bitcast operations.
* fuzzgen: Request only one variable for bswap
This was included by accident. Bswap only has one input, instead of two.
* cranelift: Add `bswap.i128` support
Adds support only for x86, AArch64, S390X.
RISCV does not yet have bswap.
This branch removes the trapif and trapff instructions, in favor of using an explicit comparison and trapnz. This moves us closer to removing iflags and fflags, but introduces the need to implement instructions like iadd_cout in the x64 and aarch64 backends.
This makes some rather tricky analysis available to other users besides
the current IR. It shouldn't change current behavior, except if a rule
attempts to bind its root term to a name. There's no Rust value for a
root term, so the existing code silently ignored such bindings and would
panic saying "Variable should already be bound" if a rule attempted to
use such bindings. With this commit, the initial attempt to bind the
name reports the error instead.
- Allow bitcast for vectors with differing lane widths
- Remove raw_bitcast IR instruction
- Change all users of raw_bitcast to bitcast
- Implement support for no-op bitcast cases across backends
This implements the second step of the plan outlined here:
https://github.com/bytecodealliance/wasmtime/issues/4566#issuecomment-1234819394
* Return `anyhow::Error` from host functions instead of `Trap`
This commit refactors how errors are modeled when returned from host
functions and additionally refactors how custom errors work with `Trap`.
At a high level functions in Wasmtime that previously worked with
`Result<T, Trap>` now work with `Result<T>` instead where the error is
`anyhow::Error`. This includes functions such as:
* Host-defined functions in a `Linker<T>`
* `TypedFunc::call`
* Host-related callbacks like call hooks
Errors are now modeled primarily as `anyhow::Error` throughout Wasmtime.
This subsequently removes the need for `Trap` to have the ability to
represent all host-defined errors as it previously did. Consequently the
`From` implementations for any error into a `Trap` have been removed
here and the only embedder-defined way to create a `Trap` is to use
`Trap::new` with a custom string.
After this commit the distinction between a `Trap` and a host error is
the wasm backtrace that it contains. Previously all errors in host
functions would flow through a `Trap` and get a wasm backtrace attached
to them, but now this only happens if a `Trap` itself is created meaning
that arbitrary host-defined errors flowing from a host import to the
other side won't get backtraces attached. Some internals of Wasmtime
itself were updated or preserved to use `Trap::new` to capture a
backtrace where it seemed useful, such as when fuel runs out.
The main motivation for this commit is that it now enables hosts to
thread a concrete error type from a host function all the way through to
where a wasm function was invoked. Previously this could not be done
since the host error was wrapped in a `Trap` that didn't provide the
ability to get at the internals.
A consequence of this commit is that when a host error is returned that
isn't a `Trap` we'll capture a backtrace and then won't have a `Trap` to
attach it to. To avoid losing the contextual information this commit
uses the `Error::context` method to attach the backtrace as contextual
information to ensure that the backtrace is itself not lost.
This is a breaking change for likely all users of Wasmtime, but it's
hoped to be a relatively minor change to workaround. Most use cases can
likely change `-> Result<T, Trap>` to `-> Result<T>` and otherwise
explicit creation of a `Trap` is largely no longer necessary.
* Fix some doc links
* add some tests and make a backtrace type public (#55)
* Trap: avoid a trailing newline in the Display impl
which in turn ends up with three newlines between the end of the
backtrace and the `Caused by` in the anyhow Debug impl
* make BacktraceContext pub, and add tests showing downcasting behavior of anyhow::Error to traps or backtraces
* Remove now-unnecesary `Trap` downcasts in `Linker::module`
* Fix test output expectations
* Remove `Trap::i32_exit`
This commit removes special-handling in the `wasmtime::Trap` type for
the i32 exit code required by WASI. This is now instead modeled as a
specific `I32Exit` error type in the `wasmtime-wasi` crate which is
returned by the `proc_exit` hostcall. Embedders which previously tested
for i32 exits now downcast to the `I32Exit` value.
* Remove the `Trap::new` constructor
This commit removes the ability to create a trap with an arbitrary error
message. The purpose of this commit is to continue the prior trend of
leaning into the `anyhow::Error` type instead of trying to recreate it
with `Trap`. A subsequent simplification to `Trap` after this commit is
that `Trap` will simply be an `enum` of trap codes with no extra
information. This commit is doubly-motivated by the desire to always use
the new `BacktraceContext` type instead of sometimes using that and
sometimes using `Trap`.
Most of the changes here were around updating `Trap::new` calls to
`bail!` calls instead. Tests which assert particular error messages
additionally often needed to use the `:?` formatter instead of the `{}`
formatter because the prior formats the whole `anyhow::Error` and the
latter only formats the top-most error, which now contains the
backtrace.
* Merge `Trap` and `TrapCode`
With prior refactorings there's no more need for `Trap` to be opaque or
otherwise contain a backtrace. This commit parse down `Trap` to simply
an `enum` which was the old `TrapCode`. All various tests and such were
updated to handle this.
The main consequence of this commit is that all errors have a
`BacktraceContext` context attached to them. This unfortunately means
that the backtrace is printed first before the error message or trap
code, but given all the prior simplifications that seems worth it at
this time.
* Rename `BacktraceContext` to `WasmBacktrace`
This feels like a better name given how this has turned out, and
additionally this commit removes having both `WasmBacktrace` and
`BacktraceContext`.
* Soup up documentation for errors and traps
* Fix build of the C API
Co-authored-by: Pat Hickey <pat@moreproductive.org>
* Pull `Module` out of `ModuleTextBuilder`
This commit is the first in what will likely be a number towards
preparing for serializing a compiled component to bytes, a precompiled
artifact. To that end my rough plan is to merge all of the compiled
artifacts for a component into one large object file instead of having
lots of separate object files and lots of separate mmaps to manage. To
that end I plan on eventually using `ModuleTextBuilder` to build one
large text section for all core wasm modules and trampolines, meaning
that `ModuleTextBuilder` is no longer specific to one module. I've
extracted out functionality such as function name calculation as well as
relocation resolving (now a closure passed in) in preparation for this.
For now this just keeps tests passing, and the trajectory for this
should become more clear over the following commits.
* Remove component-specific object emission
This commit removes the `ComponentCompiler::emit_obj` function in favor
of `Compiler::emit_obj`, now renamed `append_code`. This involved
significantly refactoring code emission to take a flat list of functions
into `append_code` and the caller is responsible for weaving together
various "families" of functions and un-weaving them afterwards.
* Consolidate ELF parsing in `CodeMemory`
This commit moves the ELF file parsing and section iteration from
`CompiledModule` into `CodeMemory` so one location keeps track of
section ranges and such. This is in preparation for sharing much of this
code with components which needs all the same sections to get tracked
but won't be using `CompiledModule`. A small side benefit from this is
that the section parsing done in `CodeMemory` and `CompiledModule` is no
longer duplicated.
* Remove separately tracked traps in components
Previously components would generate an "always trapping" function
and the metadata around which pc was allowed to trap was handled
manually for components. With recent refactorings the Wasmtime-standard
trap section in object files is now being generated for components as
well which means that can be reused instead of custom-tracking this
metadata. This commit removes the manual tracking for the `always_trap`
functions and plumbs the necessary bits around to make components look
more like modules.
* Remove a now-unnecessary `Arc` in `Module`
Not expected to have any measurable impact on performance, but
complexity-wise this should make it a bit easier to understand the
internals since there's no longer any need to store this somewhere else
than its owner's location.
* Merge compilation artifacts of components
This commit is a large refactoring of the component compilation process
to produce a single artifact instead of multiple binary artifacts. The
core wasm compilation process is refactored as well to share as much
code as necessary with the component compilation process.
This method of representing a compiled component necessitated a few
medium-sized changes internally within Wasmtime:
* A new data structure was created, `CodeObject`, which represents
metadata about a single compiled artifact. This is then stored as an
`Arc` within a component and a module. For `Module` this is always
uniquely owned and represents a shuffling around of data from one
owner to another. For a `Component`, however, this is shared amongst
all loaded modules and the top-level component.
* The "module registry" which is used for symbolicating backtraces and
for trap information has been updated to account for a single region
of loaded code holding possibly multiple modules. This involved adding
a second-level `BTreeMap` for now. This will likely slow down
instantiation slightly but if it poses an issue in the future this
should be able to be represented with a more clever data structure.
This commit additionally solves a number of longstanding issues with
components such as compiling only one host-to-wasm trampoline per
signature instead of possibly once-per-module. Additionally the
`SignatureCollection` registration now happens once-per-component
instead of once-per-module-within-a-component.
* Fix compile errors from prior commits
* Support AOT-compiling components
This commit adds support for AOT-compiled components in the same manner
as `Module`, specifically adding:
* `Engine::precompile_component`
* `Component::serialize`
* `Component::deserialize`
* `Component::deserialize_file`
Internally the support for components looks quite similar to `Module`.
All the prior commits to this made adding the support here
(unsurprisingly) easy. Components are represented as a single object
file as are modules, and the functions for each module are all piled
into the same object file next to each other (as are areas such as data
sections). Support was also added here to quickly differentiate compiled
components vs compiled modules via the `e_flags` field in the ELF
header.
* Prevent serializing exported modules on components
The current representation of a module within a component means that the
implementation of `Module::serialize` will not work if the module is
exported from a component. The reason for this is that `serialize`
doesn't actually do anything and simply returns the underlying mmap as a
list of bytes. The mmap, however, has `.wasmtime.info` describing
component metadata as opposed to this module's metadata. While rewriting
this section could be implemented it's not so easy to do so and is
otherwise seen as not super important of a feature right now anyway.
* Fix windows build
* Fix an unused function warning
* Update crates/environ/src/compilation.rs
Co-authored-by: Nick Fitzgerald <fitzgen@gmail.com>
Co-authored-by: Nick Fitzgerald <fitzgen@gmail.com>
One big change here is to stop using `Term::extractor_sig`, which was
the only call that used a `TypeEnv`. However that function only uses
type information to construct the fully-qualified name of the extractor,
which is not used when building the IR. So removing it and removing the
now-unused `typeenv` parameters removes all uses of `TypeEnv` from the
`ir` and `trie` modules.
In addition, this completes the changes started in "More consistent use
of `add_inst`" (e63771f2d9), by always
using `add_inst` to get an `InstId`.
I also removed a number of unnecessary intermediate allocations.
* Cranelift: Use a single, shared vector allocation for all `ABIArg`s
Instead of two `SmallVec`s per `SigData`.
* Remove `Deref` and `DerefMut` impls for `ArgsAccumulator`
Adds Bswap to the Cranelift IR. Implements the Bswap instruction
in the x64 and aarch64 codegen backends. Cranelift users can now:
```
builder.ins().bswap(value)
```
to get a native byteswap instruction.
* x64: implements the 32- and 64-bit bswap instruction, following
the pattern set by similar unary instrutions (Neg and Not) - it
only operates on a dst register, but is parameterized with both
a src and dst which are expected to be the same register.
As x64 bswap instruction is only for 32- or 64-bit registers,
the 16-bit swap is implemented as a rotate left by 8.
Updated x64 RexFlags type to support emitting for single-operand
instructions like bswap
* aarch64: Bswap gets emitted as aarch64 rev16, rev32,
or rev64 instruction as appropriate.
* s390x: Bswap was already supported in backend, just had to add
a bit of plumbing
* For completeness, added bswap to the interpreter as well.
* added filetests and runtests for each ISA
* added bswap to fuzzgen, thanks to afonso360 for the code there
* 128-bit swaps are not yet implemented, that can be done later
* aarch64: Fix incorrect masking for small types on bmask
`bmask` was accidentally relying on the uppermost bits of the register
for small types.
This was found by fuzzgen, when it generated a shift left followed by
a bmask, the shift left shifted the bits out of the range of the input
type (i8), however these are not automatically cleared since they
remained inside the 32 bits of the register.
That caused issues when the bmask tried to compare the whole register
instead of just the bottom bits. The solution here is to mask the upper
bits for small types.
* aarch64: Emit 32bit cmp on bmask
This fixes an issue where bmask was accidentally comparing the
upper bits of the register by always using a 64bit cmp.
* riscv: Mask high bits in bmask
* riscv: Add compile tests for br{z,nz}
* riscv: Use shifts to mask 32bit values
This produces less code than the AND since that version needs to
load an immediate constant from memory.
* cranelift: Update test input to hexadecimal values
This makes it a bit more clear what is being tested.
* riscv: Use addiw for masking 32 bit values
Co-authored-by: Trevor Elliott <telliott@fastly.com>
* aarch64: Update bmask rule priority
Co-authored-by: Trevor Elliott <telliott@fastly.com>
Add a new instruction uadd_overflow_trap, which is a fused version of iadd_ifcout and trapif. Adding this instruction removes a dependency on the iflags type, and would allow us to move closer to removing it entirely.
The instruction is defined for the i32 and i64 types only, and is currently only used in the legalization of heap_addr.
Now that we aren't trying to do overlap checking in parallel, we can
fuse the loop that generates a list of rule pairs with the loop that
checks those pairs.
Removing the intermediate vector of pairs should save a little time and
memory. But it also means we're no longer borrowing from the `by_term`
HashMap, so we can use `into_iter` instead of `values` to move ownership
out of the map. That in turn means that we can use `into_iter` on each
vector of rules as well, which turns out to offer a slightly nicer idiom
for looping over all pairs, and also means we drop allocations as soon
as possible.
I also pushed grouping by priority earlier, so the O(n^2) all-pairs loop
runs over smaller lists. If we later find we want to know about overlaps
across different priorities, the definition of the map key is an easy
place to make that change.
* Cranelift: disable egraphs in fuzzing for now.
As per [this comment], with a few recent discussions it's become clear
that we want to refactor egraphs in a way that will subsume, or make
irrelevant, some of the recent fuzzbugs that have arisen (and likely
lead to others, which we'll want to fix!). Rather than chase these down
then refactor later, it probably makes sense not to spend the human time
or fuzzing time doing so. This PR turns off egraphs support in fuzzing
configurations for now, to be re-enabled later.
[this comment]: https://github.com/bytecodealliance/wasmtime/issues/5126#issuecomment-1291222515
* Disable in cranelift-fuzzgen as well.
In particular, this was found to happen in #5099 because a `Result`
projection node was not deduplicating across two separate `isplit`s that
created it. (This is a separate issue we should also fix; `needs_dedup`
is I think overly conservative because `Result` can project out a single
value from a pure or impure node, but the projection itself should be
treated like any other pure operator.)
In any case, if we have a value `v0` and two separate `Result { value:
v0, result: N, ty }` nodes, each of these will fill in the type `ty` for
the `N`th output of `v0`, and the second will idempotently overwrite the
first; we should loosen the assert so that it allows this case.
Fixes#5099. Fixes#5100.
* Cranelift: pass iterators to `ABIMachineSpec::compute_arg_locs`
Instead of slices. This gives us more flexibility to pass custom sequences
without needing to allocate a `Vec` to hold them and pass in as a slice.
* Cranelift: Avoid cloning `ir::Signature`s in `SigData::from_func_sig`
This avoids two heap allocations per signature that are unnecessary 99% of the
time.
* fix typo
* Simplify condition in `missing_struct_return`
As discussed in the 2022/10/19 meeting, this PR removes many of the branch and select instructions that used iflags, in favor if using brz/brnz and select in their place. Additionally, it reworks selectif_spectre_guard to take an i8 input instead of an iflags input.
For reference, the removed instructions are: br_icmp, brif, brff, trueif, trueff, and selectif.
* cranelift: Remove iconst.i128
* bugpoint: Report Changed when only one instruction is mutated
* cranelift: Fix egraph bxor rule
* cranelift: Remove some simple_preopt opts for i128
Eliminate a few remaining instances of non-SSA code.
Remove infrastructure previously used for non-SSA code emission.
Related cleanup around flags handling.
Using rayon adds a lot of dependencies to Cranelift. The total
unparallelized time the code that uses rayon takes is less than half a
second and it runs at compile time, so there is pretty much no benefit
to parallelizing it.
This can help rustc/llvm avoid bounds checks, but more importantly I will have
future changes here that remove indexing of params, and instead hand them out as
an iterator.
The `SigData::from_func_sig` constructor will already ensure that the struct
return pointer is returned, so this is a purely unnecessary call.
Note that this is not a performance speed up, since
`ensure_struct_return_ptr_is_returned` doesn't do any significant work if the
signature is already legalized.
This fixes#5086 by addressing two separate issues:
- The `ValueDataPacked::set_type()` helper had an embarrassing bitfield-manipulation bug that would mangle the rest of a `ValueDef` when setting its type. This is not normally used, only when the egraph elaboration fills in types after-the-fact on a multi-value node.
- The lowering rules for `isplit` on aarch64 and s390x were dispatching on the first output type, rather than the input type. When only the second output is used (as in the example in #5086), the first output type actually remains `INVALID` (and this is fine because it's never used).
Remove uses of reg_mod from the s390x backend. This required moving away from using r0/r1 as the result registers from a few different pseudo instructions, standardizing instead on r2/r3. That change was necessary as regalloc2 will not correctly allocate registers that aren't listed in the allocatable set, which r0/r1 are not.
Co-authored-by: Ulrich Weigand <ulrich.weigand@de.ibm.com>
Co-authored-by: Chris Fallin <chris@cfallin.org>
* fuzzgen: Test compiler flags
* cranelift: Generate `all()` function for all enum flags
This allows a user to iterate all flags that exist.
* fuzzgen: Minimize regalloc_checker compiles
* fuzzgen: Limit the amount of test case inputs
* fuzzgen: Add egraphs flag
It's finally here! 🥳
* cranelift: Add fuzzing comment to settings
* fuzzgen: Add riscv64
* fuzzgen: Unconditionally enable some flags
* egraphs: a few miscellaneous compile-time optimizations.
These optimizations together are worth about a 2% compile-time
reduction, as measured on one core with spidermonkey.wasm as an input,
using `hyperfine` on `wasmtime compile`.
The changes included are:
- Some better pre-allocation (blockparams and side-effects concatenated
list vecs);
- Avoiding the indirection of storing list-of-types for every Pure and
Inst node, when almost all nodes produce only a single result;
instead, store arity and single type if it exists, and allow result
projection nodes to fill in types otherwise;
- Pack the `MemoryState` enum into one `u32` (this together with the
above removal of the type slice allows `Node` to
shrink from 48 bytes to 32 bytes);
- always-inline an accessor (`entry` on `CtxHash`) that wasn't
(`always(inline)` appears to be load-bearing, rather than just
`inline`);
- Split the update-analysis path into two hotpaths, one for the union
case and one for the new-node case (and the former can avoid
recomputing for the contained node when replacing a node with
node-and-child eclass entry).
* Review feedback.
* Fix test build.
* Fix to lowering when unused output with invalid type is present.
Remove the boolean types from cranelift, and the associated instructions breduce, bextend, bconst, and bint. Standardize on using 1/0 for the return value from instructions that produce scalar boolean results, and -1/0 for boolean vector elements.
Fixes#3205
Co-authored-by: Afonso Bordado <afonso360@users.noreply.github.com>
Co-authored-by: Ulrich Weigand <ulrich.weigand@de.ibm.com>
Co-authored-by: Chris Fallin <chris@cfallin.org>
