This change adds support for immediate to memory moves in x64 which
are needed by Winch for zeroing local slots.
This change follows the guideline in `isa/x64/inst/emit` and uses
other instructions (immediate to register moves) as a base for the
test cases.
The instruction encoding expectation was derived by assembling each
instruction and inspecting the assembly with `objdump`.
Introduce a temporary for an intermediate value in the lowering of div in the x64 backend. Additionally, add a src argument to the shift_r smart constructor, which is why the diff got larger than just the div lowering.
Avoid naming %rcx as written by the CoffTlsGetAddr pseudo-instruction in the x64 backend, and instead emit a fixed-def constraint for a fresh VReg and %rcx.
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
Add a new pseudo-instruction, XmmUnaryRmRImm, to handle instructions like roundss that only use their first register argument for the instruction's result. This has the added benefit of allowing the isle wrappers for those instructions to take an XmmMem argument, allowing for more cases where loads may be merged.
* x64: clean up regalloc-related semantics on several instructions.
This PR removes all uses of "modify" operands on instructions in the x64
backend, and also removes all uses of "pinned vregs", or vregs that are
explicitly tied to particular physical registers. In place of both of
these mechanisms, which are legacies of the old regalloc design and
supported via compatibility code, the backend now uses operand
constraints. This is more flexible as it allows the regalloc to see the
liveranges and constraints without "reverse-engineering" move instructions.
Eventually, after removing all such uses (including in other backends
and by the ABI code), we can remove the compatibility code in regalloc2,
significantly simplifying its liverange-construction frontend and
thus allowing for higher confidence in correctness as well as possibly a
bit more compilation speed.
Curiously, there are a few extra move instructions now; they are likely
poor splitting decisions and I can try to chase these down later.
* Fix cranelift-codegen tests.
* Review feedback.
Lower extractlane, scalar_to_vector and splat in ISLE.
This PR also makes some changes to the SinkableLoad api
* change the return type of sink_load to RegMem as there are more functions available for dealing with RegMem
* add reg_mem_to_reg_mem_imm and register it as an automatic conversion
Lower `shuffle` and `swizzle` in ISLE.
This PR surfaced a bug with the lowering of `shuffle` when avx512vl and avx512vbmi are enabled: we use `vpermi2b` as the implementation, but panic if the immediate shuffle mask contains any out-of-bounds values. The behavior when the avx512 extensions are not present is that out-of-bounds values are turned into `0` in the result.
I've resolved this by detecting when the shuffle immediate has out-of-bounds indices in the avx512-enabled lowering, and generating an additional mask to zero out the lanes where those indices occur. This brings the avx512 case into line with the semantics of the `shuffle` op: 94bcbe8446/cranelift/codegen/meta/src/shared/instructions.rs (L1495-L1498)
Lower stack_addr, udiv, sdiv, urem, srem, umulhi, and smulhi in ISLE.
For udiv, sdiv, urem, and srem I opted to move the original lowering into an extern constructor, as the interactions with rax and rdx for the div instruction didn't seem meaningful to implement in ISLE. However, I'm happy to revisit this choice and move more of the embedding into ISLE.
This is the implementation of https://github.com/bytecodealliance/wasmtime/issues/4155, using the "inverted API" approach suggested by @cfallin (thanks!) in Cranelift, and trait object to provide a backend for an all-included experience in Wasmtime.
After the suggestion of Chris, `Function` has been split into mostly two parts:
- on the one hand, `FunctionStencil` contains all the fields required during compilation, and that act as a compilation cache key: if two function stencils are the same, then the result of their compilation (`CompiledCodeBase<Stencil>`) will be the same. This makes caching trivial, as the only thing to cache is the `FunctionStencil`.
- on the other hand, `FunctionParameters` contain the... function parameters that are required to finalize the result of compilation into a `CompiledCode` (aka `CompiledCodeBase<Final>`) with proper final relocations etc., by applying fixups and so on.
Most changes are here to accomodate those requirements, in particular that `FunctionStencil` should be `Hash`able to be used as a key in the cache:
- most source locations are now relative to a base source location in the function, and as such they're encoded as `RelSourceLoc` in the `FunctionStencil`. This required changes so that there's no need to explicitly mark a `SourceLoc` as the base source location, it's automatically detected instead the first time a non-default `SourceLoc` is set.
- user-defined external names in the `FunctionStencil` (aka before this patch `ExternalName::User { namespace, index }`) are now references into an external table of `UserExternalNameRef -> UserExternalName`, present in the `FunctionParameters`, and must be explicitly declared using `Function::declare_imported_user_function`.
- some refactorings have been made for function names:
- `ExternalName` was used as the type for a `Function`'s name; while it thus allowed `ExternalName::Libcall` in this place, this would have been quite confusing to use it there. Instead, a new enum `UserFuncName` is introduced for this name, that's either a user-defined function name (the above `UserExternalName`) or a test case name.
- The future of `ExternalName` is likely to become a full reference into the `FunctionParameters`'s mapping, instead of being "either a handle for user-defined external names, or the thing itself for other variants". I'm running out of time to do this, and this is not trivial as it implies touching ISLE which I'm less familiar with.
The cache computes a sha256 hash of the `FunctionStencil`, and uses this as the cache key. No equality check (using `PartialEq`) is performed in addition to the hash being the same, as we hope that this is sufficient data to avoid collisions.
A basic fuzz target has been introduced that tries to do the bare minimum:
- check that a function successfully compiled and cached will be also successfully reloaded from the cache, and returns the exact same function.
- check that a trivial modification in the external mapping of `UserExternalNameRef -> UserExternalName` hits the cache, and that other modifications don't hit the cache.
- This last check is less efficient and less likely to happen, so probably should be rethought a bit.
Thanks to both @alexcrichton and @cfallin for your very useful feedback on Zulip.
Some numbers show that for a large wasm module we're using internally, this is a 20% compile-time speedup, because so many `FunctionStencil`s are the same, even within a single module. For a group of modules that have a lot of code in common, we get hit rates up to 70% when they're used together. When a single function changes in a wasm module, every other function is reloaded; that's still slower than I expect (between 10% and 50% of the overall compile time), so there's likely room for improvement.
Fixes#4155.
* Add a test for the existing behavior of fcvt_from_unit
* Migrate the I8, I16, I32 cases of fcvt_from_uint
* Implement the I64 case of fcvt_from_uint
* Add a test for the existing behavior of fcvt_from_uint.f64x2
* Migrate fcvt_from_uint.f64x2 to ISLE
* Lower the last case of `fcvt_from_uint`
* Add a test for `fcvt_from_uint`
* Finish lowering fcmp_from_uint
* Format
* x64: Add VEX Instruction Encoder
This uses a similar builder pattern to the EVEX Encoder.
Does not yet support memory accesses.
* x64: Add FMA Flag
* x64: Implement SIMD `fma`
* x64: Use 4 register Vex Inst
* x64: Reorder VEX pretty print args
* x64: port `atomic_rmw` to ISLE
This change ports `atomic_rmw` to ISLE for the x64 backend. It does not
change the lowering in any way, though it seems possible that the fixed
regs need not be as fixed and that there are opportunities for single
instruction lowerings. It does rename `inst_common::AtomicRmwOp` to
`MachAtomicRmwOp` to disambiguate with the IR enum with the same name.
* x64: remove remaining hardcoded register constraints for `atomic_rmw`
* x64: use `SyntheticAmode` in `AtomicRmwSeq`
* review: add missing reg collector for amode
* review: collect memory registers in the 'late' phase
- Handle call instructions' clobbers with the clobbers API, using RA2's
clobbers bitmask (bytecodealliance/regalloc2#58) rather than clobbers
list;
- Pull in changes from bytecodealliance/regalloc2#59 for much more sane
edge-case behavior w.r.t. liverange splitting.
`idiv` on x86-64 only reads `rdx`/`edx`/`dx`/`dl` for divides with width
greater than 8 bits; for an 8-bit divide, it reads the whole 16-bit
divisor from `ax`, as our CISC ancestors intended. This PR fixes the
metadata to avoid a regalloc panic (due to undefined `rdx`) in this
case. Does not affect Wasmtime or other Wasm-frontend embedders.
x64 backend: add lowerings with load-op-store fusion.
These lowerings use the `OP [mem], reg` forms (or in AT&T syntax, `OP
%reg, (mem)`) -- i.e., x86 instructions that load from memory, perform
an ALU operation, and store the result, all in one instruction. Using
these instruction forms, we can merge three CLIF ops together: a load,
an arithmetic operation, and a store.
This PR switches Cranelift over to the new register allocator, regalloc2.
See [this document](https://gist.github.com/cfallin/08553421a91f150254fe878f67301801)
for a summary of the design changes. This switchover has implications for
core VCode/MachInst types and the lowering pass.
Overall, this change brings improvements to both compile time and speed of
generated code (runtime), as reported in #3942:
```
Benchmark Compilation (wallclock) Execution (wallclock)
blake3-scalar 25% faster 28% faster
blake3-simd no diff no diff
meshoptimizer 19% faster 17% faster
pulldown-cmark 17% faster no diff
bz2 15% faster no diff
SpiderMonkey, 21% faster 2% faster
fib(30)
clang.wasm 42% faster N/A
```
This primary motivation of this large commit (apologies for its size!) is to
introduce `Gpr` and `Xmm` newtypes over `Reg`. This should help catch
difficult-to-diagnose register class mixup bugs in x64 lowerings.
But having a newtype for `Gpr` and `Xmm` themselves isn't enough to catch all of
our operand-with-wrong-register-class bugs, because about 50% of operands on x64
aren't just a register, but a register or memory address or even an
immediate! So we have `{Gpr,Xmm}Mem[Imm]` newtypes as well.
Unfortunately, `GprMem` et al can't be `enum`s and are therefore a little bit
noisier to work with from ISLE. They need to maintain the invariant that their
registers really are of the claimed register class, so they need to encapsulate
the inner data. If they exposed the underlying `enum` variants, then anyone
could just change register classes or construct a `GprMem` that holds an XMM
register, defeating the whole point of these newtypes. So when working with
these newtypes from ISLE, we rely on external constructors like `(gpr_to_gpr_mem
my_gpr)` instead of `(GprMem.Gpr my_gpr)`.
A bit of extra lines of code are included to add support for register mapping
for all of these newtypes as well. Ultimately this is all a bit wordier than I'd
hoped it would be when I first started authoring this commit, but I think it is
all worth it nonetheless!
In the process of adding these newtypes, I didn't want to have to update both
the ISLE `extern` type definition of `MInst` and the Rust definition, so I move
the definition fully into ISLE, similar as aarch64.
Finally, this process isn't complete. I've introduced the newtypes here, and
I've made most XMM-using instructions switch from `Reg` to `Xmm`, as well as
register class-converting instructions, but I haven't moved all of the GPR-using
instructions over to the newtypes yet. I figured this commit was big enough as
it was, and I can continue the adoption of these newtypes in follow up commits.
Part of #3685.
Peepmatic was an early attempt at a DSL for peephole optimizations, with the
idea that maybe sometime in the future we could user it for instruction
selection as well. It didn't really pan out, however:
* Peepmatic wasn't quite flexible enough, and adding new operators or snippets
of code implemented externally in Rust was a bit of a pain.
* The performance was never competitive with the hand-written peephole
optimizers. It was *very* size efficient, but that came at the cost of
run-time efficiency. Everything was table-based and interpreted, rather than
generating any Rust code.
Ultimately, because of these reasons, we never turned Peepmatic on by default.
These days, we just landed the ISLE domain-specific language, and it is better
suited than Peepmatic for all the things that Peepmatic was originally designed
to do. It is more flexible and easy to integrate with external Rust code. It is
has better time efficiency, meeting or even beating hand-written code. I think a
small part of the reason why ISLE excels in these things is because its design
was informed by Peepmatic's failures. I still plan on continuing Peepmatic's
mission to make Cranelift's peephole optimizer passes generated from DSL rewrite
rules, but using ISLE instead of Peepmatic.
Thank you Peepmatic, rest in peace!
On the build side, this commit introduces two things:
1. The automatic generation of various ISLE definitions for working with
CLIF. Specifically, it generates extern type definitions for clif opcodes and
the clif instruction data `enum`, as well as extractors for matching each clif
instructions. This happens inside the `cranelift-codegen-meta` crate.
2. The compilation of ISLE DSL sources to Rust code, that can be included in the
main `cranelift-codegen` compilation.
Next, this commit introduces the integration glue code required to get
ISLE-generated Rust code hooked up in clif-to-x64 lowering. When lowering a clif
instruction, we first try to use the ISLE code path. If it succeeds, then we are
done lowering this instruction. If it fails, then we proceed along the existing
hand-written code path for lowering.
Finally, this commit ports many lowering rules over from hand-written,
open-coded Rust to ISLE.
In the process of supporting ISLE, this commit also makes the x64 `Inst` capable
of expressing SSA by supporting 3-operand forms for all of the existing
instructions that only have a 2-operand form encoding:
dst = src1 op src2
Rather than only the typical x86-64 2-operand form:
dst = dst op src
This allows `MachInst` to be in SSA form, since `dst` and `src1` are
disentangled.
("3-operand" and "2-operand" are a little bit of a misnomer since not all
operations are binary operations, but we do the same thing for, e.g., unary
operations by disentangling the sole operand from the result.)
There are two motivations for this change:
1. To allow ISLE lowering code to have value-equivalence semantics. We want ISLE
lowering to translate a CLIF expression that evaluates to some value into a
`MachInst` expression that evaluates to the same value. We want both the
lowering itself and the resulting `MachInst` to be pure and referentially
transparent. This is both a nice paradigm for compiler writers that are
authoring and maintaining lowering rules and is a prerequisite to any sort of
formal verification of our lowering rules in the future.
2. Better align `MachInst` with `regalloc2`'s API, which requires that the input
be in SSA form.
This commit removes the Lightbeam backend from Wasmtime as per [RFC 14].
This backend hasn't received maintenance in quite some time, and as [RFC
14] indicates this doesn't meet the threshold for keeping the code
in-tree, so this commit removes it.
A fast "baseline" compiler may still be added in the future. The
addition of such a backend should be in line with [RFC 14], though, with
the principles we now have for stable releases of Wasmtime. I'll close
out Lightbeam-related issues once this is merged.
[RFC 14]: https://github.com/bytecodealliance/rfcs/pull/14
* Add support for x64 packed promote low
* Add support for x64 packed floating point demote
* Update vector promote low and demote by adding constraints
Also does some renaming and minor refactoring
Previously, the multiple flags for certain AVX512 instructions were
checked using `OR`: e.g., if the CPU has AVX512VL `OR` AVX512DQ,
emit `VPMULLQ`. This is incorrect--the logic should be `AND`. The Intel
Software Developer Manual, vol. 1, sec. 15.4, has more information on
this (notable there is the suggestion to check with `XGETBV` that the OS
is allowing the use of the XMM registers--but that is a separate issue).
This change switches to `AND` logic in the new backend.
When shuffling values from two different registers, the x64 lowering for
`i8x16.shuffle` must first shuffle each register separately and then OR
the results with SSE instructions. With `VPERMI2B`, available in
AVX512VL + AVX512VBMI, this can be done in a single instruction after
the shuffle mask has been moved into the destination register. This
change uses `VPERMI2B` for that case when the CPU supports it.
When AVX512VL or AVX512BITALG are available, Wasm SIMD's `popcnt`
instruction can be lowered to a single x64 instruction, `VPOPCNTB`,
instead of 8+ instructions.
When AVX512VL and AVX512F are available, use a single instruction
(`VCVTUDQ2PS`) instead of a length 9-instruction sequence. This
optimization is a port from the legacy x86 backend.
This change implements `vselect` using SSE4.1's `BLENDVPS`, `BLENDVPD`,
and `PBLENDVB`. `vselect` is a lane-selecting instruction that is used
by
[simple_preopt.rs](fa1faf5d22/cranelift/codegen/src/simple_preopt.rs (L947-L999))
to lower `bitselect` to a single x86 instruction when the condition mask
is known to be boolean (all 1s or 0s, e.g., from a conversion). This is
better than `bitselect` in general, which lowers to 4-5 instructions.
The old backend had the `vselect` lowering; this simply introduces it to
the new backend.
This adds the machinery to encode the VPMULLQ instruction which is
available in AVX512VL and AVX512DQ. When these feature sets are
available, we use this instruction instead of a lengthy 12-instruction
sequence.
* x64: add EVEX encoding mechanism
Also, includes an empty stub module for the VEX encoding.
* x64: lower abs.i64x2 to VPABSQ when available
* x64: refactor EVEX encodings to use `EvexInstruction`
This change replaces the `encode_evex` function with a builder-style struct, `EvexInstruction`. This approach clarifies the code, adds documentation, and results in slight speedups when benchmarked.
* x64: rename encoding CodeSink to ByteSink
This instruction has a single instruction lowering in AVX512F/VL and a three instruction lowering in AVX but neither is currently supported in the x64 backend. To implement this, we instead subtract the vector from 0 and use a blending instruction to pick the lanes containing the absolute value.
This fixes#2672 and #2679, and also fixes an incorrect instruction
emission (`test` with small immediate) that we had missed earlier.
The shift-related fixes have to do with (i) shifts by 0 bits, as a
special case that must be handled; and (ii) shifts by a 128-bit amount,
which we can handle by just dropping the upper half (we only use 3--7
bits of shift amount).
This adjusts the lowerings appropriately, and also adds run-tests to
ensure that the lowerings actually execute correctly (previously we only
had compile-tests with golden lowerings; I'd like to correct this for
more ops eventually, adding run-tests beyond what the Wasm spec and
frontend covers).
