Like decl code generation, also unify the wasm backend and the wasm linker to call into
the general purpose `codegen.zig` to generate the code for a function.
When an union had a zero-sized payload type, we would lower the tag twice. This is fixed
by exiting early when `payload_size` is 0.
With regards to error unions, we were only accounting for padding for the payload field.
However, the errorset value can have a smaller alignment than the payload as well, i.e. error!usize.
We fix this by also accounting for padding/alignment of the error set tag of an error union.
Several issues with pointer types are fixed:
Prior to this commit, Zig would not canonicalize a pointer type with
an explicit alignment to alignment=0 if it matched the pointee ABI
alignment. In order to fix this, `Type.ptr` now takes a Target
parameter. I also moved the host_size canonicalization to `Type.ptr`
since target is now available. Similarly, is_allowzero in the case of
C pointers is now treated as a canonicalization done by the function
rather than a precondition.
in-memory coercion for pointers now properly checks ABI alignment
of pointee types instead of incorrectly treating the 0 value as an
alignment.
Type equality is completely reworked based on the tag() rather than the
zigTypeTag(). It's still semantically based on zigTypeTag() but that
knowledge is implied rather than dictating the control flow of the
logic. Importantly, this fixes cases for opaques, structs, tuples,
enums, and unions, where type equality was incorrectly returning based
on whether the tag() values were equal.
Additionally, pointer type equality now takes into account alignment.
Because we canonicalize non-zero alignment which equals pointee type ABI
alignment to alignment=0, this now can be a simple integer comparison.
Type hashing is implemented for pointers and floats. Array types now
additionally hash their sentinels.
This regressed some behavior tests that were passing but only because
of bugs regarding type equality.
The C backend has a noticeable problem with lowering differently-aligned
pointers (particularly slices) as the same type, causing C compilation
errors due to duplicate declarations.
In `getDeclVAddr`, it may happen that the target `Decl` has not
been allocated space in virtual memory. In this case, we store a
relocation in the linker-global table which we will iterate over
when flushing the module, and fill in any missing address in the
final binary. Note that for optimisation, if the address was resolved
at the time of a call to `getDeclVAddr`, we skip relocating this
atom.
This commit also adds the glue code for lowering const slices in
the ARM backend.
* pass more x64 behavior tests
* return with a TODO error when lowering a decl with no runtime bits
* insert some debug logs for tracing recursive descent down the
type-value tree when lowering types
* print `Decl`'s name when print debugging `decl_ref` value
* pad out (non-packed) struct fields when lowering to bytes to be
saved in the binary - prior to this change, fields would be
saved at non-aligned addresses leading to wrong accesses
* add a matching test case to `behavior/struct.zig` tests
* fix offset to field calculation in `struct_field_ptr` on `x86_64`
AstGen:
* rename the known_has_bits flag to known_non_opv to make it better
reflect what it actually means.
* add a known_comptime_only flag.
* make the flags take advantage of identifiers of primitives and the
fact that zig has no shadowing.
* correct the known_non_opv flag for function bodies.
Sema:
* Rename `hasCodeGenBits` to `hasRuntimeBits` to better reflect what it
does.
- This function got a bit more complicated in this commit because of
the duality of function bodies: on one hand they have runtime bits,
but on the other hand they require being comptime known.
* WipAnonDecl now takes a LazySrcDecl parameter and performs the type
resolutions that it needs during finish().
* Implement comptime `@ptrToInt`.
Codegen:
* Improved handling of lowering decl_ref; make it work for
comptime-known ptr-to-int values.
- This same change had to be made many different times; perhaps we
should look into merging the implementations of `genTypedValue`
across x86, arm, aarch64, and riscv.
It is the job of codegen backends to mark Decls that are referenced as
alive so that the frontend does not sweep them with the garbage. This
commit unifies the code between the backends with an added method on
Decl.
The implementation is more complete than before, switching on the Decl
val tag and recursing into sub-values.
As a result, two more array tests are passing.
* stage2: put decls in different MachO sections
Use `getDeclVAddrWithReloc` when targeting MachO backend rather than
`getDeclVAddr` - this fn returns a zero vaddr and instead creates a
relocation on the linker side which will get automatically updated
whenever the target decl is moved in memory. This fn also records
a rebase of the target pointer so that its value is correctly slid
in presence of ASLR.
This commit enables `zig test` on x86_64-macos.
* stage2: fix output section selection for type,val pairs
const locals now detect if the value ends up being comptime known. In
such case, it replaces the runtime AIR instructions with a decl_ref
const.
In the backends, some more sophisticated logic for marking decls as
alive was needed to prevent Decls incorrectly being garbage collected
that were indirectly referenced in such manner.
* Introduce a mechanism into Sema for emitting a compile error when an
integer is too big and we need it to fit into a usize.
* Add `@intCast` where necessary
* link/MachO: fix an unnecessary allocation when all that was happening
was appending zeroes to an ArrayList.
* Add `error.Overflow` as a possible error to some codepaths, allowing
usage of `math.intCast`.
closes#9710
* incorporate Andrew's MIR draft as Mir.zig
* add skeleton for Emit.zig module - Emit will lower MIR into
machine code or textual ASM.
* implement push
* implement ret
* implement mov r/m, r
* implement sub r/m imm and sub r/m, r
* put encoding common ops together - some ops share impl such as
MOV and cmp so put them together and vary the actual opcode
with modRM ext only.
* implement pop
* implement movabs - movabs being a special-case of mov not
handled by general mov MIR instruction due to requirement to
handle 64bit immediates.
* store imm64 as a struct `Imm64{ msb: u32, lsb: u32 }` in extra data
for use with for instance movabs inst
* implement more mov variations
* implement adc
* implement add
* implement sub
* implement xor
* implement and
* implement or
* implement sbb
* implement cmp
* implement lea - lea doesn't follow the scheme as other inst above. Similarly, I
think bit shifts and rotates should be put in a separate basket too.
* implement adc_scale_src
* implement add_scale_src
* implement sub_scale_src
* implement xor_scale_src
* implement and_scale_src
* implement or_scale_src
* implement sbb_scale_src
* implement cmp_scale_src
* implement adc_scale_dst
* implement add_scale_dst
* implement sub_scale_dst
* implement xor_scale_dst
* implement and_scale_dst
* implement or_scale_dst
* implement sbb_scale_dst
* implement cmp_scale_dst
* implement mov_scale_src
* implement mov_scale_dst
* implement adc_scale_imm
* implement add_scale_imm
* implement sub_scale_imm
* implement xor_scale_imm
* implement and_scale_imm
* implement or_scale_imm
* implement sbb_scale_imm
* implement cmp_scale_imm
* port bin math to MIR
* backpatch stack size into prev MIR inst
* implement Function.gen() (minus dbg info)
* implement jmp/call [imm] - we can now call functions using indirect absolute
addressing, or via registers.
* port airRet to use MIR
* port airLoop to use MIR
* patch up performReloc to use inst indices
* implement conditional jumps (without relocs)
* implement set byte on condition
* implement basic lea r64, [rip + imm]
* implement calling externs
* implement callq in PIE
* implement lea RIP in PIE context
* remove all refs to Encoder from CodeGen
* implement basic imul ops
* pass all Linux tests!
* enable most of dbg info gen
* generate arg dbg info in Emit
The main problem that motivated these changes is that global constants
which are referenced by pointer would not be emitted into the binary.
This happened because `semaDecl` did not add `codegen_decl` tasks for
global constants, instead relying on the constant values being copied as
necessary. However when the global constants are referenced by pointer,
they need to be sent to the linker to be emitted.
After making global const arrays, structs, and unions get emitted, this
uncovered a latent issue: the anonymous decls that they referenced would
get garbage collected (via `deleteUnusedDecl`) even though they would
later be referenced by the global const.
In order to solve this problem, I introduced `anon_work_queue` which is
the same as `work_queue` except a lower priority. The `codegen_decl`
task for anon decls goes into the `anon_work_queue` ensuring that the
owner decl gets a chance to mark its anon decls as alive before they are
possibly deleted.
This caused a few regressions, which I made the judgement call to add
workarounds for. Two steps forward, one step back, is still progress.
The regressions were:
* Two behavior tests having to do with unions. These tests were
intentionally exercising the LLVM constant value lowering, however,
due to the bug with garbage collection that was fixed in this commit,
the LLVM code was not getting exercised, and union types/values were
not implemented correctly, due to me forgetting that LLVM does not
allow bitcasting aggregate values.
- This is worked around by allowing those 2 test cases to regress,
moving them to the "passing for stage1 only" section.
* The test-stage2 test cases (in test/cases/*) for non-LLVM backends
previously did not have any calls to lower struct values, but now
they do. The code that was there was just `@panic("TODO")`. I
replaced that code with a stub that generates the wrong value. This
is an intentional miscompilation that will obviously need to get
fixed before any struct behavior tests pass. None of the current
tests we have exercise loading any values from these global const
structs, so there is not a problem until we try to improve these
backends.
* Fix backend using wrong union field of the slice instruction.
* LLVM backend properly sets alignment on global variables.
* Sema: add coercion for *T to *[1]T
* Sema: pointers to Decls with explicit alignment now have alignment
metadata in them.
AIR:
* div is renamed to div_trunc.
* Add div_float, div_floor, div_exact.
- Implemented in Sema and LLVM codegen. C backend has a stub.
Improvements to std.math.big.Int:
* Add `eqZero` function to `Mutable`.
* Fix incorrect results for `divFloor`.
Compiler-rt:
* Add muloti4 to the stage2 section.
* Restructure elemPtr a bit
* New AIR instruction: slice_elem_ptr, which returns a pointer to an element of a slice
* Value: adapt elemPtr to work on slices
* New AIR instruction: slice, which constructs a slice out of a pointer
and a length.
* AstGen: use `coerced_ty` for start and end expressions, use `none`
for the sentinel, and don't try to load the result of the slice
operation because it returns a by-value result.
* Sema: pointer arithmetic is extracted into analyzePointerArithmetic
and it is used by the implementation of slice.
- Also I implemented comptime pointer addition.
* Sema: extract logic into analyzeSlicePtr, analyzeSliceLen and use them
inside the slice semantic analysis.
- The approach in stage2 is much cleaner than stage1 because it uses
more granular analysis calls for obtaining the slice pointer, doing
arithmetic on it, and checking if the length is comptime-known.
* Sema: use the slice Value Tag for slices when doing coercion from
pointer-to-array.
* LLVM backend: detect when emitting a GEP instruction into a
pointer-to-array and add the extra index that is required.
* Type: ptrAlignment for c_void returns 0.
* Implement Value.hash and Value.eql for slices.
* Remove accidentally duplicated behavior test.
* std.os: take advantage of `@minimum`. It's probably time to
deprecate `std.min` and `std.max`.
* New AIR instructions: min and max
* Introduce SIMD vector support to stage2
* Add `@Type` support for vectors
* Sema: add `checkSimdBinOp` which can be re-used for other arithmatic
operators that want to support vectors.
* Implement coercion from vectors to arrays.
- In backends this is handled with bitcast for vector to array,
however maybe we want to reduce the amount of branching by
introducing an explicit AIR instruction for it in the future.
* LLVM backend: implement lowering vector types
* Sema: Implement `slice.ptr` at comptime
* Value: improve `numberMin` and `numberMax` to support floats in
addition to integers, and make them behave properly in the presence
of NaN.
* apply late symbol resolution for globals - instead of resolving
the exact location of a symbol in locals, globals or undefs,
we postpone the exact resolution until we have a full picture
for relocation resolution.
* fixup stubs to defined symbols - this is currently a hack rather
than a final solution. I'll need to work out the details to make
it more approachable. Currently, we preemptively create a stub
for a lazy bound global and fix up stub offsets in stub helper
routine if the global turns out to be undefined only. This is quite
wasteful in terms of space as we create stub, stub helper and lazy ptr
atoms but don't use them for defined globals.
* change log scope to .link for macho.
* remove redundant code paths from Object and Atom.
* drastically simplify the contents of Relocation struct (i.e., it is
now a simple superset of macho.relocation_info), clean up relocation
parsing and resolution logic.
