This pipeline is run as part of IRGen and has access to the IRGenModule.
Passes that run as part of this pipeline can query for the IRGenModule.
We will use it for the AllocStackHoisting pass. It wants to know if a type is of
non-fixed size.
To break the cyclic dependency between IRGen -> SILOptimizer -> IRGen that would
arise from the SILPassManager having to know about the createIRGENPASS()
function IRGen passes instead of exposing this function dynamically have to add
themselves to the pass manager.
Following classes provide symbol mangling for specific purposes:
*) Mangler: the base mangler class, just providing some basic utilities
*) ASTMangler: for mangling AST declarations
*) SpecializationMangler: to be used in the optimizer for mangling specialized function names
*) IRGenMangler: mangling all kind of symbols in IRGen
All those classes are not used yet, so it’s basically a NFC.
Another change is that some demangler node types are added (either because they were missing or the new demangler needs them).
Those new nodes also need to be handled in the old demangler, but this should also be a NFC as those nodes are not created by the old demangler.
My plan is to keep the old and new mangling implementation in parallel for some time. After that we can remove the old mangler.
Currently the new implementation is scoped in the NewMangling namespace. This namespace should be renamed after the old mangler is removed.
Previously there was a mismatch between SIL's concept of a
struct (padding is implicit) and LLVM's (padding is explicit) in IRGen's
constant evaluator. The explicit padding fields need to be given a value
in the IR.
This patch also moves the (currently small) list of constant evaluation
functions into their own file.
Fixes SR-716.
As a first step to allowing the build script to build *only*
static library versions of the stdlib, change `add_swift_library`
such that callers must pass in `SHARED`, `STATIC`, or `OBJECT_LIBRARY`.
Ideally, only these flags would be used to determine whether to
build shared, static, or object libraries, but that is not currently
the case -- `add_swift_library` also checks whether the library
`IS_STDLIB` before performing certain additional actions. This will be
cleaned up in a future commit.
- Implement emission of type references for nominal type field
reflection, using a small custom encoder resulting in packed
structs, not strings. This will let us embed 7-bit encoded
32-bit relative offsets directly in the structure (not yet
hooked in).
- Use the AST Mangler for encoding type references
Archetypes and internal references were complicating this before, so we
can take the opportunity to reuse this machinery and avoid unique code
and new ABI.
Next up: Tests for reading the reflection sections and converting the
demangle tree into a tree of type references.
Todo: For concrete types, serialize the types for associated types of
their conformances to bootstrap the typeref substitution process.
rdar://problem/15617914
There are several interesting new features here.
The first is that, when emitting a SILFunction, we're now able to
cache type data according to the full dominance structure of the
original function. For example, if we ask for type metadata, and
we've already computed it in a dominating position, we're now able
to re-use that value; previously, we were limited to only doing this
if the value was from the entry block or the LLVM basic block
matched exactly. Since this tracks the SIL dominance relationship,
things in IRGen which add their own control flow must be careful
to suppress caching within blocks that may not dominate the
fallthrough; this mechanism is currently very crude, but could be
made to allow a limited amount of caching within the
conditionally-executed blocks.
This query is done using a proper dominator tree analysis, even at -O0.
I do not expect that we will frequently need to actually build the
tree, and I expect that the code-size benefits of doing a real
analysis will be significant, especially as we move towards making
more metadata lazily computed.
The second feature is that this adds support for "abstract"
cache entries, which indicate that we know how to derive the metadata
but haven't actually done so. This code isn't yet tested, but
it's going to be the basis of making a lot of things much lazier.
I believe this was not caught before since the users of IRGen included those.
swift-llvm-opt only includes swiftIRGen so missing symbols resulted while
linking.
Swift SVN r32805
- GenProto.cpp for protocols and protocol conformances
- GenExistential.cpp for existential type layout and operations
- GenArchetype.cpp for archetype type layout and operations
Swift SVN r32493
Using LLVM large integers to represent enum payloads has been causing compiler performance and code size problems with large types, and has also exposed a long tail of backend bugs. Replace them with an "EnumPayload" abstraction that manages breaking a large opaque binary value into chunks, along with masking, testing, and extracting typed data from the binary blob. For now, use a word-sized chunking schema always, though the architecture here is set up to eventually allow the use of an arbitrary explosion schema, which would benefit single-payload enums by allowing the payload to follow the explosion schema of the contained value.
This time, adjust the assertion in emitCompare not to perform a check before we've established that the payload is empty, since APInt doesn't have a 0-bit state and the default-constructed form is nondeterminisitic. (We should probably use a more-tailored representation for enum payload bit patterns than APInt or ClusteredBitVector.)
Swift SVN r28985
Using LLVM large integers to represent enum payloads has been causing compiler performance and code size problems with large types, and has also exposed a long tail of backend bugs. Replace them with an "EnumPayload" abstraction that manages breaking a large opaque binary value into chunks, along with masking, testing, and extracting typed data from the binary blob. For now, use a word-sized chunking schema always, though the architecture here is set up to eventually allow the use of an arbitrary explosion schema, which would benefit single-payload enums by allowing the payload to follow the explosion schema of the contained value.
Swift SVN r28982
There must be a better way ... to get us moving forward i am committing this
until somebody tells me how to do this right (tm).
rdar://20265255
Swift SVN r26455
This adds the -profile-generate flag, which enables LLVM's
instrumentation based profiling. It implements the instrumentation
for basic control flow, such as if statements, loops, and closures.
Swift SVN r25155
We've had a rash of bugs due to inconsistencies between how IRGen and the runtime think types are laid out. Add a '-verify-type-layout' mode to the frontend that causes IRGen to emit a bunch of code that compares its static assumptions against what the runtime value witness does.
Swift SVN r24918
Teach IRGen and the runtime about the extra inhabitants
of function pointers, and take advantage of that in
thin and thick function types.
Also add runtime entrypoints for thin function type
metadata.
Swift SVN r24346
OptimizeARC does not only contain an optimize arc pass: the library also
includes aa. What this really is a repository of the extra passes and
infrastructure that we inject into LLVM. Thus LLVMPasses is a more descriptive
name. It also matches SILPasses.
I also taught lit how to use the new llvm-opt driver for running swift llvm
passes through opt without having to remember how to setup the dynamic swift
llvm pass dylib. You can use this in lit tests by using the substitution
%llvm-opt.
Swift SVN r21654
This handles things like NSSwapHostLongLongToBig and MKMapRectMake that
are static inline functions that themselves call other static inline
functions.
<rdar://problem/17227237>
Swift SVN r21080
This allows IRGen to complain about types it doesn't know how to lower yet, while still recovering well enough not to take the compiler down with it. This reduces the common "unimplemented enum layout" error to be a mere error instead of a compiler crash.
Swift SVN r20773
Finishes the removal of the old "wrapped" module section that was made
unnecessary in r12922/3. Now that we no longer use the old compiler, we
don't need this at all. That also removes the need for SwiftTargetMachine.
No functionality change; this was all dead code.
Swift SVN r14758
Start out with pointer spare bits, heap object alignment, null page size, and ObjC reserved bits, and provide static values for X86-64 and for a worst-case generic platform.
Swift SVN r7833
proposal.
When compiling with debug info, build a swiftmodule that contains all the
type decls referenced by DWARF and emit it into a special __apple_swiftast
section in the .o file.
Swift SVN r7398
We haven't fully updated references to union cases, and enums still are not
their own thing yet, but "oneof" is gone. Long live "union"!
Swift SVN r6783
their types.
- DebugTypeInfo holds all type info we need to emit debug information.
- Type info is limited to name, location, and storage size.
- As a side-effect: verbose LLVM IR allocas in debug builds!
Swift SVN r5980
and lexical scopes, which can be enabled through the new -g option.
When -g is enabled, line tables and scopes compile all the way
down to DWARF.
Changes to SIL:
- In addition to a SILLocation, every instruction now also has a pointer
to a SILDebugScope (its containing lexical scope).
- Added LexicalScope, which is to be used for all Scopes we want to show
up in the debug info.
Swift SVN r5772
Sever the last load-bearing link between SILFunction and SILConstant by naming SILFunctions with their mangled symbol names. Move the core of the mangler up to SIL, and teach SILGen how to use it to mangle a SILConstant.
Swift SVN r4964
