We were unnecessarily conservative here; generic metadata patterns support indirectable references
to value witness tables exported by the standard library, so if we have a fixed-layout generic
type that matches a known value witness table layout, use that instead of generating a new
value witness table.
This is essentially a long-belated follow-up to Arnold's #12606.
The key observation here is that the enum-tag-single-payload witnesses
are strictly more powerful than the XI witnesses: you can simulate
the XI witnesses by using an extra case count that's <= the XI count.
Of course the result is less efficient than the XI witnesses, but
that's less important than overall code size, and we can work on
fast-paths for that.
The extra inhabitant count is stored in a 32-bit field (always present)
following the ValueWitnessFlags, which now occupy a fixed 32 bits.
This inflates non-XI VWTs on 32-bit targets by a word, but the net effect
on XI VWTs is to shrink them by two words, which is likely to be the
more important change. Also, being able to access the XI count directly
should be a nice win.
@objc enums don't support reflection and aren't given extra inhabitants, so we never in practice use their enum value witnesses (and the ones we generated were wrong!), and we can share the prefab value witnesses for the builtin integer types for C enums. All uninhabited types like Never can also share an arbitrary value witness table; the value witnesses will never be invoked.
This allows us to layout-optimize Optional<T> when T is a struct with an
extra-inhabitant-bearing field anywhere in its definition, not only at
the beginning. rdar://problem/43019427
And update the existential container's initializeWithTake implementation
in the runtime. After only allowing bitwise takable values in the
inline buffer we can use memcpy to move existential container values.
rdar://31414907
SR-343
We are seeing some crashes due to this optimization where a type
contains a single retainable pointer. Temporarily disable it while
we get to the root cause.
rdar://problem/39629937
Fixes a regression in the source compatibility suite which I had a
lot of trouble extracting into a separate test case.
Most of this patch is just moving the outlining code into a separate
file and organizing it into a helper class instead of copy/pasting
so much code. The main functional change is implicit in the difference
between collecting formal metadata and collecting it for layout, which
then is exploited in bindMetadataParameters.
As a secondary change, stop collecting metadata for class-bounded
archetypes; we don't actually need it to do value operations.
Most of the work of this patch is just propagating metadata states
throughout the system, especially local-type-data caching and
metadata-path resolution. It took a few design revisions to get both
DynamicMetadataRequest and MetadataResponse to a shape that felt
right and seemed to make everything easier.
The design is laid out pretty clearly (I hope) in the comments on
DynamicMetadataRequest and MetadataResponse, so I'm not going to
belabor it again here. Instead, I'll list out the work that's still
outstanding:
- I'm sure there are places we're asking for complete metadata where
we could be asking for something weaker.
- I need to actually test the runtime behavior to verify that it's
breaking the cycles it's supposed to, instead of just not regressing
anything else.
- I need to add something to the runtime to actually force all the
generic arguments of a generic type to be complete before reporting
completion. I think we can get away with this for now because all
existing types construct themselves completely on the first request,
but there might be a race condition there if another asks for the
type argument, gets an abstract metadata, and constructs a type with
it without ever needing it to be completed.
- Non-generic resilient types need to be switched over to an IRGen
pattern that supports initialization suspension.
- We should probably space out the MetadataStates so that there's some
space between Abstract and Complete.
- The runtime just calmly sits there, never making progress and
permanently blocking any waiting threads, if you actually form an
unresolvable metadata dependency cycle. It is possible to set up such
a thing in a way that Sema can't diagnose, and we should detect it at
runtime. I've set up some infrastructure so that it should be
straightforward to diagnose this, but I haven't actually implemented
the diagnostic yet.
- It's not clear to me that swift_checkMetadataState is really cheap
enough that it doesn't make sense to use a cache for type-fulfilled
metadata in associated type access functions. Fortunately this is not
ABI-affecting, so we can evaluate it anytime.
- Type layout really seems like a lot of code now that we sometimes
need to call swift_checkMetadataState for generic arguments. Maybe
we can have the runtime do this by marking low bits or something, so
that a TypeLayoutRef is actually either (1) a TypeLayout, (2) a known
layout-complete metadata, or (3) a metadata of unknown state. We could
do that later with a flag, but we'll need to at least future-proof by
allowing the runtime functions to return a MetadataDependency.
If a type has the same layout as one of the basic integer types, or has a single refcounted pointer representation, we can use prefab value witness tables from the runtime instead of instantiating new ones. This saves quite a bit of code size, particularly in the Apple SDK overlays, where there are lots of swift_newtype wrappers and option set structs.
- Create the value witness table as a separate global object instead
of concatenating it to the metadata pattern.
- Always pass the metadata to the runtime and let the runtime handle
instantiating or modifying the value witness table.
- Pass the right layout algorithm version to the runtime; currently
this is always "Swift 5".
- Create a runtime function to instantiate single-case enums.
Among other things, this makes the copying of the VWT, and any
modifications of it, explicit and in the runtime, which is more
future-proof.
So far single payload enums were implemented in terms of runtime functions which
internally emitted several calls to value witnesses.
This commit adds value witnesses to get and store the enum tag side stepping the
need for witness calls as this information is statically available in many cases
/// int (*getEnumTagSinglePayload)(const T* enum, UINT_TYPE emptyCases)
/// Given an instance of valid single payload enum with a payload of this
/// witness table's type (e.g Optional<ThisType>) , get the tag of the enum.
/// void (*storeEnumTagSinglePayload)(T* enum, INT_TYPE whichCase,
/// UINT_TYPE emptyCases)
/// Given uninitialized memory for an instance of a single payload enum with a
/// payload of this witness table's type (e.g Optional<ThisType>), store the
/// tag.
A simple 'for element in array' loop in generic code operating on a
ContigousArray of Int is ~25% faster on arm64.
rdar://31408033
Replace `NameOfType foo = dyn_cast<NameOfType>(bar)` with DRY version `auto foo = dyn_cast<NameOfType>(bar)`.
The DRY auto version is by far the dominant form already used in the repo, so this PR merely brings the exceptional cases (redundant repetition form) in line with the dominant form (auto form).
See the [C++ Core Guidelines](https://github.com/isocpp/CppCoreGuidelines/blob/master/CppCoreGuidelines.md#es11-use-auto-to-avoid-redundant-repetition-of-type-names) for a general discussion on why to use `auto` to avoid redundant repetition of type names.
* IRGen: Change c-o-w existential implementation functions
* initialzeBufferWith(Copy|Take)OfBuffer value witness implementation for cow existentials
Implement and use initialzeBufferWith(Copy|Take)OfBuffer value witnesses for
copy-on-write existentials.
Before we used a free standing function but the overhead of doing so was
noticable (~20-30%) on micro benchmarks.
* IRGen: Use common getCopyOutOfLineBoxPointerFunction
* Add a runtime function to conditionally make a box unique
* Fix compilation of HeapObject.cpp on i386
* Fix IRGen test case
* Fix test case for i386
This is NFC in intent, but I had to restructure the code to emit more
of the lists "inline", which means I inevitably altered some IRGen
emission patterns in ways that are visible to tests:
- GenClass emits property/ivar/whatever descriptors in a somewhat
different order.
- An ext method type list is now emitted as just an array, not a struct
containing only that array.
- Protocol descriptors are no longer emitted as packed structs.
I was sorely tempted to stop using packed structs for all the metadata
emission, but didn't really want to update that many tests in one go.
