* clif-util: Make the `-t` flag work with the `wasm` subcommand
The presence of the flag was checked in the code, so it was essentially already
supported, but it was not properly configured when constructing
the CLI arguments parser.
* clif-util: also enable the `-c` flag for the `wasm` subcommand
Similar to the parent commit, this functionality is already supported, just a
mix up of the CLI parser construction made it not show up.
This commit introduces initial support for multi-value Wasm. Wasm blocks and
calls can now take and return an arbitrary number of values.
The encoding for multi-value blocks means that we need to keep the contents of
the "Types" section around when translating function bodies. To do this, we
introduce a `WasmTypesMap` type that maps the type indices to their parameters
and returns, construct it when parsing the "Types" section, and shepherd it
through a bunch of functions and methods when translating function bodies.
ReplaceBuilder is available in the public API through
`DataFlowGraph::replace`, however it's documentation is not available
through rustdoc as the type isn't publicly importable.
Only i16x8 and i32x4 are encoded in this commit mainly because i8x16 and i64x2 do not have simple encodings in x86. i64x2 is not required by the SIMD spec and there is discussion (https://github.com/WebAssembly/simd/pull/98#issuecomment-530092217) about removing i8x16.
The x86 ISA has (at least) two encodings for PEXTRW:
1. in the SSE2 opcode (66 0f c5) the XMM operand uses r/m and the GPR operand uses reg
2. in the SSE4.1 opcode (66 0f 3a 15) the XMM operand uses reg and the GPR operand uses r/m
This changes the 16-bit x86_pextr encoding from 1 to 2 to match the other PEXTR* implementations (all #2 style).
Add documentation to the icmp instruction text for both signed and
unsigned overflow, making it very clear why unsigned overflow is
complicated and where to find it.
`VisibleTranslationState` was a wrapper around a `TranslationState` that was
meant to public API consumers outside of this crate. However, the internal
`TranslationState` and all its methods were still publicly exposed! This commit
simplifies and remedies the situation by combining them into a single
`TranslationState` type. Most of its methods are only `pub(crate)` now, not
visible to the entire world. The only methods that are `pub` are the ones that
`VisibleTranslationState` exposed.
This move allows the `IntCC`/`FloatCC` enums to be used in both meta (for predicate matching) and in codegen. To avoid breaking any code dependent on the previous location of condcodes.rs (`cranelift-codegen/src/condcodes.rs`), the module is re-exported under `cranelift_codegen::ir`.
Instead of using MOVUPS to expensively load bits from memory, this change uses a predicate to optimize vconst without a memory access:
- when the 128-bit immediate is all zeroes in all bits, use PXOR to zero out an XMM register
- when the 128-bit immediate is all ones in all bits, use PCMPEQB to set an XMM register to all ones
This leaves the constant data in the constant pool, which may increase code size (TODO)
