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.
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>
* 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>
* egraph-based midend: draw the rest of the owl.
* Rename `egg` submodule of cranelift-codegen to `egraph`.
* Apply some feedback from @jsharp during code walkthrough.
* Remove recursion from find_best_node by doing a single pass.
Rather than recursively computing the lowest-cost node for a given
eclass and memoizing the answer at each eclass node, we can do a single
forward pass; because every eclass node refers only to earlier nodes,
this is sufficient. The behavior may slightly differ from the earlier
behavior because we cannot short-circuit costs to zero once a node is
elaborated; but in practice this should not matter.
* Make elaboration non-recursive.
Use an explicit stack instead (with `ElabStackEntry` entries,
alongside a result stack).
* Make elaboration traversal of the domtree non-recursive/stack-safe.
* Work analysis logic in Cranelift-side egraph glue into a general analysis framework in cranelift-egraph.
* Apply static recursion limit to rule application.
* Fix aarch64 wrt dynamic-vector support -- broken rebase.
* Topo-sort cranelift-egraph before cranelift-codegen in publish script, like the comment instructs me to!
* Fix multi-result call testcase.
* Include `cranelift-egraph` in `PUBLISHED_CRATES`.
* Fix atomic_rmw: not really a load.
* Remove now-unnecessary PartialOrd/Ord derivations.
* Address some code-review comments.
* Review feedback.
* Review feedback.
* No overlap in mid-end rules, because we are defining a multi-constructor.
* rustfmt
* Review feedback.
* Review feedback.
* Review feedback.
* Review feedback.
* Remove redundant `mut`.
* Add comment noting what rules can do.
* Review feedback.
* Clarify comment wording.
* Update `has_memory_fence_semantics`.
* Apply @jameysharp's improved loop-level computation.
Co-authored-by: Jamey Sharp <jamey@minilop.net>
* Fix suggestion commit.
* Fix off-by-one in new loop-nest analysis.
* Review feedback.
* Review feedback.
* Review feedback.
* Use `Default`, not `std::default::Default`, as per @fitzgen
Co-authored-by: Nick Fitzgerald <fitzgen@gmail.com>
* Apply @fitzgen's comment elaboration to a doc-comment.
Co-authored-by: Nick Fitzgerald <fitzgen@gmail.com>
* Add stat for hitting the rewrite-depth limit.
* Some code motion in split prelude to make the diff a little clearer wrt `main`.
* Take @jameysharp's suggested `try_into()` usage for blockparam indices.
Co-authored-by: Jamey Sharp <jamey@minilop.net>
* Take @jameysharp's suggestion to avoid double-match on load op.
Co-authored-by: Jamey Sharp <jamey@minilop.net>
* Fix suggestion (add import).
* Review feedback.
* Fix stack_load handling.
* Remove redundant can_store case.
* Take @jameysharp's suggested improvement to FuncEGraph::build() logic
Co-authored-by: Jamey Sharp <jamey@minilop.net>
* Tweaks to FuncEGraph::build() on top of suggestion.
* Take @jameysharp's suggested clarified condition
Co-authored-by: Jamey Sharp <jamey@minilop.net>
* Clean up after suggestion (unused variable).
* Fix loop analysis.
* loop level asserts
* Revert constant-space loop analysis -- edge cases were incorrect, so let's go with the simple thing for now.
* Take @jameysharp's suggestion re: result_tys
Co-authored-by: Jamey Sharp <jamey@minilop.net>
* Fix up after suggestion
* Take @jameysharp's suggestion to use fold rather than reduce
Co-authored-by: Jamey Sharp <jamey@minilop.net>
* Fixup after suggestion
* Take @jameysharp's suggestion to remove elaborate_eclass_use's return value.
* Clarifying comment in terminator insts.
Co-authored-by: Jamey Sharp <jamey@minilop.net>
Co-authored-by: Nick Fitzgerald <fitzgen@gmail.com>
* Fix StructReturn handling: properly mark the clobber, and offset actual rets.
The legalization of `StructReturn` was causing issues in the new
call-handling code: the `StructReturn` ret was included in the `SigData` as
if it were an actual CLIF-level return value, but it is not.
Prior to using regalloc constraints for return values, we
unconditionally included rax (or the architecture's usual return
register) as a def, so it would be properly handled as "clobbered" by
the regalloc. With the new scheme, we include defs on the call only for
CLIF-level outputs. Callees with `StructReturn` args were thus not known
to clobber the return-value register, and values might be corrupted.
This PR updates the code to include a `StructReturn` ret as a clobber
rather than a returned value in the relevant spots. I observed it
causing saves/restores of rax in some CLIF that @bjorn3 provided me, but
I was having difficulty minimizing this into a test-case that I would be
comfortable including as a precise-output case (including the whole
thing verbatim would lock down a bunch of other irrelevant details and
cause test-update noise later). If we can find a more minimized example
I'm happy to include it as a filetest.
Fixes#5018.
* Replace resize+copy_from_slice with extend_from_slice
Vec::resize initializes the new space, which is wasted effort if we're
just going to call `copy_from_slice` on it immediately afterward. Using
`extend_from_slice` is simpler, and very slightly faster.
If the new size were bigger than the buffer we're copying from, then it
would make sense to initialize the excess. But it isn't: it's always
exactly the same size.
* Move helpers from Context to CompiledCode
These methods only use information from Context::compiled_code, so they
should live on CompiledCode instead.
* Remove an unnecessary #[cfg_attr]
There are other uses of `#[allow(clippy::too_many_arguments)]` in this
file, so apparently it doesn't need to be guarded by the "cargo-clippy"
feature.
* Fix a few comments
Two of these were wrong/misleading:
- `FunctionBuilder::new` does not clear the provided func_ctx. It does
debug-assert that the context is already clear, but I don't think
that's worth a comment.
- `switch_to_block` does not "create values for the arguments." That's
done by the combination of `append_block_params_for_function_params`
and `declare_wasm_parameters`.
* wasmtime-cranelift: Misc cleanups
The main change is to use the `CompiledCode` reference we already had
instead of getting it out of `Context` repeatedly. This removes a bunch
of `unwrap()` calls.
* wasmtime-cranelift: Factor out uncached compile
Resolve overlap in the ISLE prelude and the x64 inst module by introducing new types that allow better sharing of extractor resuls, or falling back on priorities.
* Port branches to ISLE (AArch64)
Ported the existing implementations of the following opcodes for AArch64
to ISLE:
- `Brz`
- `Brnz`
- `Brif`
- `Brff`
- `BrIcmp`
- `Jump`
- `BrTable`
Copyright (c) 2022 Arm Limited
* Remove dead code
Copyright (c) 2022 Arm Limited
* Cranelift: use regalloc2 constraints on caller side of ABI code.
This PR updates the shared ABI code and backends to use register-operand
constraints rather than explicit pinned-vreg moves for register
arguments and return values.
The s390x backend was not updated, because it has its own implementation
of ABI code. Ideally we could converge back to the code shared by x64
and aarch64 (which didn't exist when s390x ported calls to ISLE, so the
current situation is underestandable, to be clear!). I'll leave this for
future work.
This PR exposed several places where regalloc2 needed to be a bit more
flexible with constraints; it requires regalloc2#74 to be merged and
pulled in.
* Update to regalloc2 0.3.3.
In addition to version bump, this required removing two asserts as
`SpillSlot`s no longer carry their class (so we can't assert that they
have the correct class).
* Review comments.
* Filetest updates.
* Add cargo-vet audit for regalloc2 0.3.2 -> 0.3.3 upgrade.
* Update to regalloc2 0.4.0.
* Port `icmp` to ISLE (AArch64)
Ported the existing implementation of `icmp` (and, by extension, the
`lower_icmp` function) to ISLE for AArch64.
Copyright (c) 2022 Arm Limited
* Allow 'producer chains', eliminating `Nop0`s
Copyright (c) 2022 Arm Limited
* ABI: implement register arguments with constraints.
Currently, Cranelift's ABI code emits a sequence of moves from physical
registers into vregs at the top of the function body, one for every
register-carried argument.
For a number of reasons, we want to move to operand constraints instead,
and remove the use of explicitly-named "pinned vregs"; this allows for
better regalloc in theory, as it removes the need to "reverse-engineer"
the sequence of moves.
This PR alters the ABI code so that it generates a single "args"
pseudo-instruction as the first instruction in the function body. This
pseudo-inst defs all register arguments, and constrains them to the
appropriate registers at the def-point. Subsequently the regalloc can
move them wherever it needs to.
Some care was taken not to have this pseudo-inst show up in
post-regalloc disassemblies, but the change did cause a general regalloc
"shift" in many tests, so the precise-output updates are a bit noisy.
Sorry about that!
A subsequent PR will handle the other half of the ABI code, namely, the
callsite case, with a similar preg-to-constraint conversion.
* Update based on review feedback.
* Review feedback.
Previously, Cranelift panicked (via a a panic in regalloc2) when the
virtual-register limit of 2M (2^21) was reached. This resulted in a
perplexing and unhelpful failure when the user provided a too-large
input (such as the Wasm module in #4865).
This PR adds an explicit check when allocating vregs that fails with a
"code too large" error when the limit is hit, producing output such as
(on the minimized testcase from #4865):
```
Error: failed to compile wasm function 3785 at offset 0xa3f3
Caused by:
Compilation error: Code for function is too large
```
Fixes#4865.
* Initial forward-edge CFI implementation
Give the user the option to start all basic blocks that are targets
of indirect branches with the BTI instruction introduced by the
Branch Target Identification extension to the Arm instruction set
architecture.
Copyright (c) 2022, Arm Limited.
* Refactor `from_artifacts` to avoid second `make_executable` (#1)
This involves "parsing" twice but this is parsing just the header of an
ELF file so it's not a very intensive operation and should be ok to do
twice.
* Address the code review feedback
Copyright (c) 2022, Arm Limited.
Co-authored-by: Alex Crichton <alex@alexcrichton.com>
The previous implementation assumed that nothing had clobbered the
LR register since the current function had started executing, so
it would be incorrect for a non-leaf function, for example, that
contains the `get_return_address` operation right after a call.
The operation is valid only if the `preserve_frame_pointers` flag
is enabled, which implies that the presence of a frame record on
the stack is guaranteed.
Copyright (c) 2022, Arm Limited.
* cranelift-codegen: Remove all uses of DataValue
This type is only used by the interpreter, cranelift-fuzzgen, and
filetests. I haven't found another convenient crate for those to all
depend on where this type can live instead, but this small refactor at
least makes it obvious that code generation does not in any way depend
on the implementation of this type.
* Make DataValue, not Ieee32/64, respect IEEE754
This fixes#4857 by partially reverting #4849.
It turns out that Ieee32 and Ieee64 need bitwise equality semantics so
they can be used as hash-table keys.
Moving the IEEE754 semantics up a layer to DataValue makes sense in
conjunction with #4855, where we introduced a DataValue::bitwise_eq
alternative implementation of equality for those cases where users of
DataValue still want the bitwise equality semantics.
* cranelift-interpreter: Use eq/ord from DataValue
This fixes#4828, again, now that the comparison operators on DataValue
have the right IEEE754 semantics.
* Add regression test from issue #4857
Add a function_alignment function to the TargetIsa trait, and use it to align functions when generating objects. Additionally, collect the maximum alignment required for pc-relative constants in functions and pass that value out. Use the max of these two values when padding functions for alignment.
This fixes a bug on x86_64 where rip-relative loads to sse registers could cause a segfault, as functions weren't always guaranteed to be aligned to 16-byte addresses.
Fixes#4812
* Cranelift: Deduplicate ABI signatures during lowering
This commit creates the `SigSet` type which interns and deduplicates the ABI
signatures that we create from `ir::Signature`s. The ABI signatures are now
referred to indirectly via a `Sig` (which is a `cranelift_entity` ID), and we
pass around a `SigSet` to anything that needs to access the actual underlying
`SigData` (which is what `ABISig` used to be).
I had to change a couple methods to return a `SmallInstVec` instead of emitting
directly to work around what would otherwise be shared and exclusive borrows of
the lowering context overlapping. I don't expect any of these to heap allocate
in practice.
This does not remove the often-unnecessary allocations caused by
`ensure_struct_return_ptr_is_returned`. That is left for follow up work.
This also opens the door for further shuffling of signature data into more
efficient representations in the future, now that we have `SigSet` to store it
all in one place and it is threaded through all the code. We could potentially
move each signature's parameter and return vectors into one big vector shared
between all signatures, for example, which could cut down on allocations and
shrink the size of `SigData` since those `SmallVec`s have pretty large inline
capacity.
Overall, this refactoring gives a 1-7% speedup for compilation on
`pulldown-cmark`:
```
compilation :: cycles :: benchmarks/pulldown-cmark/benchmark.wasm
Δ = 8754213.66 ± 7526266.23 (confidence = 99%)
dedupe.so is 1.01x to 1.07x faster than main.so!
[191003295 234620642.20 280597986] dedupe.so
[197626699 243374855.86 321816763] main.so
compilation :: cycles :: benchmarks/bz2/benchmark.wasm
No difference in performance.
[170406200 194299792.68 253001201] dedupe.so
[172071888 193230743.11 223608329] main.so
compilation :: cycles :: benchmarks/spidermonkey/benchmark.wasm
No difference in performance.
[3870997347 4437735062.59 5216007266] dedupe.so
[4019924063 4424595349.24 4965088931] main.so
```
* Use full path instead of import to avoid warnings in some build configurations
Warnings will then cause CI to fail.
* Move `SigSet` into `VCode`
Ported the existing implementation of `fcmp` for AArch64 to ISLE.
This also ports the `lower_vector_comparison` method to ISLE.
Copyright (c) 2022 Arm Limited
This retains `lower_amode` in the handwritten code (@akirilov-arm
reports that there is an upcoming patch to port this), but tweaks it
slightly to take a `Value` rather than an `Inst`.
Ensure that constants generated for the memory case of XmmMem values are always 16 bytes, ensuring that we don't accidantally perform an unaligned load.
Fixes#4761
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)
* Port `vconst` to ISLE (AArch64)
Ported the existing implementation of `vconst` to ISLE for AArch64, and
added support for 64-bit vector constants.
Also introduced 64-bit `vconst` support to the interpreter.
Copyright (c) 2022 Arm Limited
* Replace if-chains with match statements
Copyright (c) 2022 Arm Limited
Fixes#4736
Fix lowerings that were using values as both a Reg and a RegMem, making it look like a load could be sunk while its value in a register was still being used. Also add an assert that checks that loads that are sunk are never used.
* Cranelift: Remove the `ABICaller` trait
It has only one implementation: the `ABICallerImpl` struct. We can just use that
directly rather than having extra, unnecessary layers of generics and abstractions.
* Cranelift: Rename `ABICallerImpl` to `Caller`
* Cranelift: Remove `ABICallee` trait
It has only one implementation: the `ABICalleeImpl` struct. By using that
directly we can avoid unnecessary layers of generics and abstractions as well as
a couple `Box`es that were previously putting the single implementation into a
`Box<dyn>`.
* Cranelift: Rename `ABICalleeImpl` to `AbiCallee`
* Fix comments as per review
* Rename `AbiCallee` to `Callee`
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.
The trait had only one implementation: the `Lower` struct. It is easier to just
use that directly, and not introduce unnecessary layers of generics and
abstractions.
Once upon a time, there was hope that we would have other implementations of the
`LowerCtx` trait, that did things like lower CLIF to SMTLIB for
verification. However, this is not practical these days given the way that the
trait has evolved over time, and our verification efforts are focused on ISLE
now anyways, and we're actually making some progress on that front (much more
than anyone ever did on a second `LowerCtx` trait implementation!)
Implement the tls_value for s390 in the ELF general-dynamic mode.
Notable differences to the x86_64 implementation are:
- We use a __tls_get_offset libcall instead of __tls_get_addr.
- The current thread pointer (stored in a pair of access registers)
needs to be added to the result of __tls_get_offset.
- __tls_get_offset has a variant ABI that requires the address of
the GOT (global offset table) is passed in %r12.
This means we need a new libcall entries for __tls_get_offset.
In addition, we also need a way to access _GLOBAL_OFFSET_TABLE_.
The latter is a "magic" symbol with a well-known name defined
by the ABI and recognized by the linker. This patch introduces
a new ExternalName::KnownSymbol variant to support such names
(originally due to @afonso360).
We also need to emit a relocation on a symbol placed in a
constant pool, as well as an extra relocation on the call
to __tls_get_offset required for TLS linker optimization.
Needed by the cg_clif frontend.
This is a nonsensical constraint: the entry block must come first in the
compiled code's layout, so it cannot also be sunk to the end of the
function.
This PR modifies the CLIF verifier to disallow this situation entirely.
It also adds an assert during final block-order computation to catch the
problem (and avoid a silent miscompile) even if the verifier is
disabled.
Fixes#4656.
The `MachBuffer` applies a set of peephole-optimization rules to do
branch threading, leverage fallthrough paths, eliminate empty blocks,
and flip conditional branches where needed to make branches more
efficient starting from naive always-branch-at-end-of-BB code.
This works by applying the rules at every label-bind, which is
equivalent to applying them at the end of every basic block, where
branches are usually inserted.
However, this misses one case: the end of the buffer! Currently we
don't optimize any redundant or foldable branches at the very end of
the machine code.
This usually doesn't matter when the function ends in an epilogue with
`ret` as the last instruction. However, when cold blocks exist, it can
actually matter.
Thanks to @mchesser for pointing out this issue in #4636.
To determine whether we need to insert a "veneer island" of
branch-range extension veneers, we need to know ahead of emitting a
basic block the worst-case size of that block. This is because veneers
only go between blocks (we could plop one in the middle of a block but
that would require another jump around it and would probably pessimize
some code significantly), and we can't back up once we emit a block.
To compute this worst-case size, we take the number of instructions
and multiply by the largest possible size of one pseudoinst (e.g., on
aarch64, this is 44 bytes; it explicitly excludes the `EmitIsland`
pseudo-op which is used before large jumptable inline offset tables
are emitted). This is conservative, but it always works, and veneers
are somewhat rare in practice (function body >1MiB on aarch64 for
example).
Unfortunately this logic didn't account for the spill/reload/move
instructions inserted by the register allocator, and in one example in
issue #4629, a block had only one instruction but 482
edge-moves (!). This came at just the wrong time as we were
approaching the 1MiB limit on aarch64.
This PR fixes that issue, and fixes the logic to actually look at the
correct next block (next in `final_order` rather than numerically
next), as a bonus correctness fix.
Fixes#4629.
* Convert `fma`, `valltrue` & `vanytrue` to ISLE (AArch64)
Ported the existing implementations of the following opcodes to ISLE on
AArch64:
- `fma`
- Introduced missing support for `fma` on vector values, as per the
docs.
- `valltrue`
- `vanytrue`
Also fixed `fcmp` on scalar values in the interpreter, and enabled
interpreter tests in `simd-fma.clif`.
This introduces the `FMLA` machine instruction.
Copyright (c) 2022 Arm Limited
* Add comments for `Fmla` and `Bsl`
Copyright (c) 2022 Arm Limited
And the `ABICallerImpl::ir_sig` field that was used to implement that
method. This removes 56 bytes from the size of `ABICallerImpl` and gives us
speed ups to compilation of about 7% on all benchmarks.
```
compilation :: nanoseconds :: benchmarks/pulldown-cmark/benchmark.wasm
Δ = 8205119.48 ± 4069474.25 (confidence = 99%)
main.so is 0.91x to 0.97x faster than feature.so!
feature.so is 1.03x to 1.10x faster than main.so!
[117729152 132258110.36 167484097] main.so
[107486500 124052990.88 138008797] feature.so
compilation :: nanoseconds :: benchmarks/bz2/benchmark.wasm
Δ = 4645258.32 ± 1981104.59 (confidence = 99%)
main.so is 0.92x to 0.97x faster than feature.so!
feature.so is 1.03x to 1.08x faster than main.so!
[76562171 85504479.28 93116863] main.so
[75180650 80859220.96 90591978] feature.so
compilation :: nanoseconds :: benchmarks/spidermonkey/benchmark.wasm
Δ = 150575617.54 ± 65021102.57 (confidence = 99%)
main.so is 0.92x to 0.97x faster than feature.so!
feature.so is 1.03x to 1.08x faster than main.so!
[2573089039 2843117485.10 3175982602] main.so
[2559784932 2692541867.56 3143529008] feature.so
```
This adds full i128 support to the s390x target, including new filetests
and enabling the existing i128 runtest on s390x.
The ABI requires that i128 is passed and returned via implicit pointer,
but the front end still generates direct i128 types in call. This means
we have to implement ABI support to implicitly convert i128 types to
pointers when passing arguments.
To do so, we add a new variant ABIArg::ImplicitArg. This acts like
StructArg, except that the value type is the actual target type,
not a pointer type. The required conversions have to be inserted
in the prologue and at function call sites.
Note that when dereferencing the implicit pointer in the prologue,
we may require a temp register: the pointer may be passed on the
stack so it needs to be loaded first, but the value register may
be in the wrong class for pointer values. In this case, we use
the "stack limit" register, which should be available at this
point in the prologue.
For return values, we use a mechanism similar to the one used for
supporting multiple return values in the Wasmtime ABI. The only
difference is that the hidden pointer to the return buffer must
be the *first*, not last, argument in this case.
(This implements the second half of issue #4565.)
This adds support for StructArgument on s390x. The ABI for this
platform requires that the address of the buffer holding the copy
of the struct argument is passed from caller to callee as hidden
pointer, using a register or overflow stack slot.
To implement this, I've added an optional "pointer" filed to
ABIArg::StructArg, and code to handle the pointer both in common
abi_impl code and the s390x back-end.
One notable change necessary to make this work involved the
"copy_to_arg_order" mechanism. Currently, for struct args
we only need to copy the data (and that need to happen before
setting up any other args), while for non-struct args we only
need to set up the appropriate registers or stack slots.
This order is ensured by sorting the arguments appropriately
into a "copy_to_arg_order" list.
However, for struct args with explicit pointers we need to *both*
copy the data (again, before everything else), *and* set up a
register or stack slot. Since we now need to touch the argument
twice, we cannot solve the ordering problem by a simple sort.
Instead, the abi_impl common code now provided *two* callbacks,
emit_copy_regs_to_buffer and emit_copy_regs_to_arg, and expects
the back end to first call copy..to_buffer for all args, and
then call copy.._to_arg for all args. This required updates
to all back ends.
In the s390x back end, in addition to the new ABI code, I'm now
adding code to actually copy the struct data, using the MVC
instruction (for small buffers) or a memcpy libcall (for larger
buffers). This also requires a bit of new infrastructure:
- MVC is the first memory-to-memory instruction we use, which
needed a bit of memory argument tweaking
- We also need to set up the infrastructure to emit libcalls.
(This implements the first half of issue #4565.)
Give the user the option to sign and to authenticate function
return addresses with the operations introduced by the Pointer
Authentication extension to the Arm instruction set architecture.
Copyright (c) 2021, Arm Limited.
