Add `linear_function` and `linear_function_extract` instructions.
`linear_function` creates a `@differentiable(linear)` function-typed value from
an original function operand and a transpose function operand (optional).
`linear_function_extract` extracts either the original or transpose function
value from a `@differentiable(linear)` function.
Resolves TF-1142 and TF-1143.
Add `differentiable_function` and `differentiable_function_extract`
instructions.
`differentiable_function` creates a `@differentiable` function-typed
value from an original function operand and derivative function operands
(optional).
`differentiable_function_extract` extracts either the original or
derivative function value from a `@differentiable` function.
The differentiation transform canonicalizes `differentiable_function`
instructions, filling in derivative function operands if missing.
Resolves TF-1139 and TF-1140.
I enabled this recently by piggybacking it in the mega-commit:
commit badc5658bb
Author: Andrew Trick <atrick@apple.com>
Date: Mon Mar 2 16:23:53 2020
Fix SIL MemBehavior queries with access markers.
It miscompiles SwiftNIO's NIOPerformanceTester benchmark causing it to
segfault immediately.
We may consider debugging this to reenable it in the future...
<rdar://60068448> Teach IRGen to emit !invariant metadata for 'let' variables
Fixes <rdar://60038603> Swift CI: SwiftNIO_macOS NIOHTTP2_macOS fails
This is in prepration for other bug fixes.
Clarify the SIL utilities that return canonical address values for
formal access given the address used by some memory operation:
- stripAccessMarkers
- getAddressAccess
- getAccessedAddress
These are closely related to the code in MemAccessUtils.
Make sure passes use these utilities consistently so that
optimizations aren't defeated by normal variations in SIL patterns.
Create an isLetAddress() utility alongside these basic utilities to
make sure it is used consistently with the address corresponding to
formal access. When this query is used inconsistently, it defeats
optimization. It can also cause correctness bugs because some
optimizations assume that 'let' initialization is only performed on a
unique address value.
Functional changes to Memory Behavior:
- An instruction with side effects now conservatively still has side
effects even when the queried value is a 'let'. Let values are
certainly sensitive to side effects, such as the parent object being
deallocated.
- Return the correct MemBehavior for begin/end_access markers.
The `differentiability_witness_function` instruction looks up a
differentiability witness function (JVP, VJP, or transpose) for a referenced
function via SIL differentiability witnesses.
Add round-trip parsing/serialization and IRGen tests.
Notes:
- Differentiability witnesses for linear functions require more support.
`differentiability_witness_function [transpose]` instructions do not yet
have IRGen.
- Nothing currently generates `differentiability_witness_function` instructions.
The differentiation transform does, but it hasn't been upstreamed yet.
Resolves TF-1141.
Before comparing the potential sugared type for equality is needs to be mapped
into the context to resolve generic type parameters to primary archetypes.
<rdar://problem/59238327>
The 'fake use' asm gadget does not always keep variables alive. E.g., in
rdar://57754659, llvm was unable to preserve two local variables despite
the use of these gadgets. These variables were backed by a LoadInst and
an ExtractValueInst respectively. Instead of emitting shadow copies for
just those kinds of instructions, use shadow copies exclusively. This
may cause more variables to appear in the debugger window before they
are initialized, but should result in fewer variables being dropped.
rdar://57754659
In 1b2842bf902a8b52acbef2425120533b63be5ae3 this API was changed.
It seems the right thing to do is turn this into the new alignment
type by doing MaybeAlign(alignment) which is what this patch does.
SIL type lowering erases DynamicSelfType, so we generate
incorrect code when casting to DynamicSelfType. Fixing this
requires a fair amount of plumbing, but most of the
changes are mechanical.
Note that the textual SIL syntax for casts has changed
slightly; the target type is now a formal type without a '$',
not a SIL type.
Also, the unconditional_checked_cast_value and
checked_cast_value_br instructions now take the _source_
formal type as well, just like the *_addr forms they are
intended to replace.
If the size of a global variable is not known at compile time, we emit
a fixed size buffer together with initialization code when the global
variable is initialized.
Make sure the SIL optimizer does not convert this into a static
initialization, even if the size of the type is known inside the
module where the global is declared, because we don't have a way to
statically initialize something that looks like a fixed-size buffer
to client code.
Fixes <https://bugs.swift.org/browse/SR-11709>.
This option is useful to debug the compiler itself, to simulate debug info as it
would be generated when producing optimized code, but without the unpredictable
output of an optimizing debugger.
ProtocolConformanceRef already has an invalid state. Drop all of the
uses of Optional<ProtocolConformanceRef> and just use
ProtocolConformanceRef::forInvalid() to represent it. Mechanically
translate all of the callers and callsites to use this new
representation.
https://forums.swift.org/t/improving-the-representation-of-polymorphic-interfaces-in-sil-with-substituted-function-types/29711
This prepares SIL to be able to more accurately preserve the calling convention of
polymorphic generic interfaces by letting the type system represent "substituted function types".
We add a couple of fields to SILFunctionType to support this:
- A substitution map, accessed by `getSubstitutions()`, which maps the generic signature
of the function to its concrete implementation. This will allow, for instance, a protocol
witness for a requirement of type `<Self: P> (Self, ...) -> ...` for a concrete conforming
type `Foo` to express its type as `<Self: P> (Self, ...) -> ... for <Foo>`, preserving the relation
to the protocol interface without relying on the pile of hacks that is the `witness_method`
protocol.
- A bool for whether the generic signature of the function is "implied" by the substitutions.
If true, the generic signature isn't really part of the calling convention of the function.
This will allow closure types to distinguish a closure being passed to a generic function, like
`<T, U> in (*T, *U) -> T for <Int, String>`, from the concrete type `(*Int, *String) -> Int`,
which will make it easier for us to differentiate the representation of those as types, for
instance by giving them different pointer authentication discriminators to harden arm64e
code.
This patch is currently NFC, it just introduces the new APIs and takes a first pass at updating
code to use them. Much more work will need to be done once we start exercising these new
fields.
This does bifurcate some existing APIs:
- SILFunctionType now has two accessors to get its generic signature.
`getSubstGenericSignature` gets the generic signature that is used to apply its
substitution map, if any. `getInvocationGenericSignature` gets the generic signature
used to invoke the function at apply sites. These differ if the generic signature is
implied.
- SILParameterInfo and SILResultInfo values carry the unsubstituted types of the parameters
and results of the function. They now have two APIs to get that type. `getInterfaceType`
returns the unsubstituted type of the generic interface, and
`getArgumentType`/`getReturnValueType` produce the substituted type that is used at
apply sites.
Mark the condfail condition as expected-false, and arrange the generated code so that the trap
basic block is emitted out-of-line, and the happy path falls through.
This reverts commit b3ac66aeeb.
This is problematic when [thunk] functions get inlined into and the
function from which we inline expects a framepointer.
Specifically, _swift_os_log_return_address() reads the framepointer.
rdar://55852225
This removes it from the AST and largely replaces it with AnyObject
at the SIL and IRGen layers. Some notes:
- Reflection still uses the notion of "unknown object" to mean an
object with unknown refcounting. There's no real reason to make
this different from AnyObject (an existential containing a
single object with unknown refcounting), but this way nothing
changes for clients of Reflection, and it's consistent with how
native objects are represented.
- The value witness table and reflection descriptor for AnyObject
use the mangling "BO" instead of "yXl".
- The demangler and remangler continue to support "BO" because it's
still in use as a type encoding, even if it's not an AST-level
Type anymore.
- Type-based alias analysis for Builtin.UnknownObject was incorrect,
so it's a good thing we weren't using it.
- Same with enum layout. (This one assumed UnknownObject never
referred to an Objective-C tagged pointer. That certainly wasn't how
we were using it!)
While tightening the requirements of the debug info generator in
IRGenSIL I noticed that SILCloner didn't correctly transfer variable
debug info on alloc_box and alloc_stack instructions. In order to make
these mistakes easier to find I added an assertion to SILBuilder and
fixed all issues uncovered by that assertion, too.
The result is a moderate increase in debug info coverage in optimized code.
On stdlib/public/core/OSX/x86_64/Swift.o "variables with location"
increases from 60134 to 60299.
This mostly requires changing various entry points to pass around a
TypeConverter instead of a SILModule. I've left behind entry points
that take a SILModule for a few methods like SILType::subst() to
avoid creating even more churn.
This provides a singular instruction for convert an unmanaged value to a ref,
then strong_retain it. I expanded the definition of UNCHECKED_REF_STORAGE to
include these copy like instructions. This instruction is valid in all SIL.
The reason why I am adding this instruction is that currently when we emit an
access to an unowned (unsafe) ivar, we use an unmanaged_to_ref and a strong
retain. This can look to the optimizer like a strong retain that can potentially
be optimized. By combining the two together into a new instruction, we can avoid
this potential problem since the pattern matching will break.
In too many places, we were calling into `emitDynamicTypeOfOpaqueHeapObject` even when we had
more specific type information about the heap object we were querying. Replace all calls with
`emitDynamicTypeOfHeapObject`, which uses the best available access path and completely avoids
runtime calls for pure Swift classes and heap objects. When targeting non-ObjC-interop platforms,
we also know we never need to call `swift_getObjectType`, so avoid doing so altogether.
This indicates that the "self" argument to the current function is always dynamically of the exact
static base class type, allowing metadata accesses in IRGen to use the local self metadata to answer
metadata requests for the class type. Set this attribute on allocating entry points of designated
inits, which is one of the most common places where we emit redundant metadata accesses.
Class methods always have a "self" argument that can be used to get the metadata of the dynamic
Self type, which in final classes is always the same as the statically-known base class. Use this
to avoid reconstructing the static base class type.
If we're allowed to know at IRGen time what the underlying type of an opaque type is, we can
satisfy references to the opaque type's metadata or protocol witness tables by directly referencing
the underlying type instead.
To display a failure message in the debugger, create a function in the debug info which has the name of the failure message.
The debug location of the trap/cond_fail is then wrapped into this function and the function is declared as "inlined".
In case the debugger stops at the trap instruction, it displays the inline function, which looks like the failure message.
For example:
* thread #1, queue = 'com.apple.main-thread', stop reason = EXC_BAD_INSTRUCTION (code=EXC_I386_INVOP, subcode=0x0)
frame #0: 0x0000000100000cbf a.out`testit3(_:) [inlined] Unexpectedly found nil while unwrapping an Optional value at test.swift:14:11 [opt]
11
12 @inline(never)
13 func testit(_ a: Int?) -> Int {
-> 14 return a!
15 }
16
This change is currently not enabled by default, but can be enabled with the option "-Xllvm -enable-trap-debug-info".
Enabling this feature needs some changes in lldb. When the lldb part is done, this option can be removed and the feature enabled by default.
To display a failure message in the debugger, create a function in the debug info which has the name of the failure message.
The debug location of the trap/cond_fail is then wrapped into this function and the function is declared as "inlined".
In case the debugger stops at the trap instruction, it displays the inline function, which looks like the failure message.
For example:
* thread #1, queue = 'com.apple.main-thread', stop reason = EXC_BAD_INSTRUCTION (code=EXC_I386_INVOP, subcode=0x0)
frame #0: 0x0000000100000cbf a.out`testit3(_:) [inlined] Unexpectedly found nil while unwrapping an Optional value at test.swift:14:11 [opt]
11
12 @inline(never)
13 func testit(_ a: Int?) -> Int {
-> 14 return a!
15 }
16
This change is currently not enabled by default, but can be enabled with the option "-Xllvm -enable-trap-debug-info".
Enabling this feature needs some changes in lldb. When the lldb part is done, this option can be removed and the feature enabled by default.
With the option -Xllvm -basic-dynamic-replacement the runtime functions are not called (so it works with an old swift library).
But calling the original of a replaced function is not supported in this case.
Instead of a thunk insert the dispatch into the original function.
If the original function should be executed the prolog just jumps to the "real" code in the function. Otherwise the replacement function is called.
There is one little complication here: when the replacement function calls the original function, the original function should not dispatch to the replacement again.
To pass this information, we use a flag in thread local storage.
The setting and reading of the flag is done in two new runtime functions.
rdar://problem/51043781
With the advent of dynamic_function_ref the actual callee of such a ref
my vary. Optimizations should not assume to know the content of a
function referenced by dynamic_function_ref. Introduce
getReferencedFunctionOrNull which will return null for such function
refs. And getInitialReferencedFunction to return the referenced
function.
Use as appropriate.
rdar://50959798
