https://carbon.compiler-explorer.com/z/88K9Kh5Wo shows the program
exiting with a garbage value copied from uninitialized memory.
This PR modifies `lower` to detect if the function lowered is the
entry point and doesn't specify a return type. If so, it emits
different LLVM IR to return int32 0, and modifies the lowered
function signature to match the int32 return type.
See
[here](https://docs.google.com/document/d/1rWcueFwIfZox6GKVGxiUG4cBzjrZ6djXiIDGyJDtrE4/edit?tab=t.0)
for the design doc.
This also removes the default value of the `result_type_inst_id`
parameter of `HandleAction`, moves it before the action in the parameter
list, and documents it. This solves two problems:
- The default made it easy to forget, leading to unnecessary
`TypeOfInst` instructions.
- When it was present, putting it after the fairly "bulky" action
argument tended to make the callsite harder to read.
- A function with a return declaration always has exactly one
`ReturnSlotPattern`, representing the whole return declaration (whereas
previously that was omitted for value and reference returns).
- The `ReturnSlotPattern` always has a subpattern with the same form.
`OutParamPattern` already plays that role for initializing forms, and
`TuplePattern` will play that role for tuple forms. This change
introduces `ValueReturnPattern` and `RefReturnPattern` to represent
value and reference return forms.
- As before, the `ReturnSlotPattern` has a corresponding `ReturnSlot`
that represents the output that is initialized by a `return` statement.
Its structure parallels the structure of the `ReturnSlotPattern`, so we
need `ValueReturn` and `RefReturn` insts that correspond to
`ValueReturnPattern` and `RefReturnPattern`.
This is a step toward supporting generic return forms, where the
`ReturnSlotPattern`'s subpattern may be an action: this change ensures
that evaluating the action for a specific form produces the same SemIR
as if the form were concrete to begin with. More speculatively, this
should simplify the implementation of `return` statements with compound
return forms.
---------
Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
When initializing `.base` in class initialization, use `partial Base` as
the destination type rather than `Base`. Treat `partial Base` as not
being abstract even when `Base` is.
Allow conversion from a `partial T` initializer to a `T` initializer.
Store the vptr while performing the conversion. Do not store the vptr
when performing a `partial T` initialization, only when performing a
non-partial `T` initialization.
This allows us to capture the location at which a type literal was used,
even in the cases where we don't otherwise need to create a new
instruction to represent the type such as for `char` or `str`.
The logic used to build the underlying type is now marked as desugaring.
For cases such as `iN`, this causes the call to `Core.Int` to no longer
be added as a dedicated IR instruction, and instead its constant value
is used directly as the value of the `type_literal`. This results in
this being on balance a reduction in the size of the IR.
This also fixes a crash in C++ interop when using a `char` literal as a
template argument. The crash was caused by the template argument not
having an associated location when mapping to a C++ location. See
changes to check/testdata/interop/cpp/template/type_param.carbon for an
example that used to crash before this change.
Update alias handling to allow an alias to point at any type literal,
reinstating support for aliases for type literals such as `bool` and
`i32` that had previously worked but stopped working when we
transitioned those types to being defined in the prelude. See changes to
toolchain/check/testdata/alias/builtins.carbon.
All the test changes other than the two mentioned above are mechanical
autoupdate changes switching to the new instruction.
This is needed to model things like the category of `x` in the body of
`fn Foo(F:! Core.Form, x:? F)`, where the category of `x` is determined
by the concrete value of `F` (see #5389 for the design of `:?`
bindings).
This will be used in a follow-up PR.
The primary change in this PR is to split the `Initializing` expression
category into separate `ReprInitializing` and `InPlaceInitializing`
categories, depending on whether initialization uses the types
initializing representation, or is guaranteed to be in place. It also
rationalizes and documents the SemIR-level semantics of those categories
(including where #5545's "ephemeral entire reference" category will
fit), and introduces two new inst kinds to close gaps exposed in the
process.
Some additional secondary changes:
- Consistently format the storage arguments of initializers with `to`,
regardless of whether initialization is in-place, and document the `to`
notation.
- Rename some inst kinds and functions, and restructure some of the
code, for clarity and consistency with the new documentation.
- Resolve a TODO to handle more category conversions in
`CategoryConverter`, in order to make it easier to reason about category
conversions.
See #6588 and the review history of this PR for background.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
This resolves a TODO in `expr_info.cpp` by using the inst kind rather
than the bound value to track the binding's category.
Since we're churning all the `bind_name` insts in testdata anyway, I'm
also taking this opportunity to align the inst naming with the design's
terminology, by calling these insts "bindings" (this aspect of the PR is
dependent on #6231 resolving an ambiguity in that terminology). For
consistency we'll need to rename several other insts as well (see the
TODO on `RefBinding`); I'm deferring that to a separate PR to minimize
the review load, but I think those name changes are in-scope for this
review.
If a `BindSymbolicName` is converted to `type` and then to its exact
`FacetType`, we get a `FacetValue` wrapping the `BindSymbolicName` but
providing no different information: it has the same witnesses and
`FacetType` as the original `BindSymbolicName`. Yet it is a different
constant value, creating multiple canonical forms with the same meaning.
Now we make that `FacetValue` with the same `FacetType` as the
`BindSymbolicName` it wraps evaluate back to the `BindSymbolicName`,
making it the unique canonical form.
This makes the "shortcut" in convert for avoiding impl lookup when
converting from `FacetAccessType` to `FacetType` in this exact scenario
work the same as doing the full impl lookup.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Add `Dependent` value and initializing representations for types whose
representations are unknown because they are dependent. When generating
SemIR in such cases, use a worst-case initializing representation that
both provides a destination address and also propagates a potential
result value.
Use this to fix incorrect lowering and lowering crashes for specific
functions involving generic types that don't use a copy value
representation.
In lowering, be careful to distinguish between whether the initializing
representation for the generic return type uses a return slot (which
affects whether the SemIR declaration and call have one) and whether the
initializing representation for the specific return type uses a return
slot (which affects whether the LLVM IR declaration and call have one).
The goal was/is to reduce the overhead for vtables in generics - the
previous representation/prior to this patch caused a new vtable to be
created in every specific which isn't generally what we want for Carbon
generics (the whole specific/generic thing is meant to avoid creating
specific versions for things that can be a generic form parameterized by
a specific instead of manifest as a unique entity per specific)
So this moves vtables to a top level object (like functions, classes,
etc). Each dynamic class will have a vtable in this list.
Classes have a `vtable_ptr` instruction in them that points to the
vtable.
The actual generic support hasn't been implemented in this patch, as
I've been struggling with just getting this part of the migration going
& wanted to get it flushed out before adding the additional
complications.
It's possible more laziness when doing cross-file importing would be
suitable - for instance if we only need to reference the vtable from
another file, but don't need to know its individual contents, it may be
beneficial for the functions in the vtable to be import_refs (or to add
another layer of indirection - so it can be a single import_ref
all-or-nothing for the functions in the vtable).
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Co-authored-by: Carbon Infra Bot <carbon-external-infra@google.com>
Previously we created allocas for temporaries at whatever point in the
output LLVM function we'd reached. This would result in these being
dynamic allocas (performing a dynamic stack allocation), which is
inefficent and can lead to a stack overflow if it happens in a loop.
Switch to putting the allocas in the entry block instead, and instead
generate a lifetime start marker when we reach the point where the
temporary is introduced. We already did this for local variables; this
is just factoring out and reusing that code.
Update remaining parts of lowering, in particular the lowering of
aggregates, to handle lowering within a specific from a different file
than its generic. Look up information about a type in the current
specific and in its file rather than performing lookups for the type in
the generic and its file.
Remove or fix all remaining uses of raw `TypeId` in
lower/function_context and lower/handle*, so that the type from the
specific is consistently always used when lowering a specific function.
---------
Co-authored-by: Geoff Romer <gromer@google.com>
When lowering a specific function whose generic was defined in a
different file, switch to that other file's `FileContext` and lower the
generic there. Also pass the `FileContext` corresponding to the specific
into the `FunctionContext`, and use that `FileContext` for resolving
requests for constants and types from the specific.
`IRBuilderBase::SetInsertPoint` weirdly replaces our debug location with
one copied from the new insertion point, so undo its damage after
calling it.
Also included: a couple of cleanups I made while tracking this down.
These constant instructions are all TypeInstId already in their type,
and this makes their names match.
Change the name of MakeSingletonInstId as well and update its comment.
We frequently want to operate on singletons. Per discussion, drop
`Singleton` to make the code shorter.
This started off as wanting to write `inst_id.is_error()`, but the
dependency relationship between ids.h and singleton_insts.h would
require some kind of delayed evaluation to allow the implementation to
remain in headers (which I suspect is helpful to have for inlining). I
could have added something like `IsErrorInst`, forward declared in ids.h
and defined in singleton_insts.h (which would always be included by
typed_insts.h), but the template approach felt like a decent balance
between (a) removing the boilerplate `::SingletonInstId`, (b)
understandability, (c) still visually mirroring if we immediately return
a singleton, and (d) flexibility for more than just `ErrorInst`. But TBH
I'd probably still have written `is_error()` if it didn't require
addressing the cross-header cycle.
Then I tried `SemIR::InstId::Is<SemIR::ErrorInst>`, which generally
worked with types but generated the complaint that it didn't shorten
*all* singleton uses. So pulling back on `::Is`, and instead just
dropping `Singleton`.
Get type in specific when lowering associated constants. This resolves
crash when getting type for type_id is not found, when the query is for
a specific.
Note: Currently lowering only supports mangling of impls that can viewed
as a single interface (single ImplConstraint).
Adds a mapping to keep track of vtable LLVM IR decls/defs for use.
Adds the vtable_id to the vtable_ptr initialize instruction for lookup.
Adds emission of vtable declarations for use outside the file that
defines the vtable. (this isn't done lazily, it's done for any imported
class - it could be done lazily & maybe eventually has to be lazy to
handle generics)
When performing a call through an impl witness, the callee that we
type-check against is the function in the interface, so we form a
specific for that callee. However, once the impl witness access
resolves, the eventual callee is a different function -- the function in
the impl -- so this would cause us to form a `SpecificFunction` where
the callee is one function but the specific refers to a different
function.
Address this by adding another instruction, `SpecificImplFunction`, that
takes a function in an impl and a specific for the corresponding
function in the interface, and computes and returns a `SpecificFunction`
referring to the corresponding specific function in the impl, or returns
a direct reference to the function in the `impl` if it's not a generic
function.
The InstId in this field is to an instruction that declares the
function, rather than the function itself, so that diagnostics can print
where the function is coming from. The type of the function (the
FunctionType instruction) is the type (the type_id) of the
function_decl_id. So we rename the field to help make this distinction
more clear.
Followup to #4739
Non-entry-block allocas will allocate new stack memory each time they're
reached, resulting in leaking stack memory over time for allocas in a
loop. Move all such allocas to the entry block instead, and use an LLVM
intrinsic to mark when the lifetime of the variable actually begins.
Such indexing operations are created by array initialization. Since we
switched integer literals to be of type IntLiteral we've been attempting
to index arrays with the (empty) representation of an IntLiteral rather
than with an actual integer value.
This fixes a crash in lowering, at least (though only initializes the
vptr to
null for now) - certainly open to naming feedback on the instruction, or
the
exact semantics (we could have a global vptr instruction that's
referenced from
the existing instructions for reading globals, for instance).
I guess we'll want one type parameter for the vptr_init instruction,
which is
the type that this is a vptr for? (can do that here or in a follow-on
patch)
Actually presenting two options in this one review - if you look at the
specific
commits in this PR, the first commit represents my first attempt - and
if you
look at the overall PR change for the second attempt.
But I'm totally open to completely different approaches/ideas - these
were just
my rough guesses.
The new `FacetValue` instruction represents `C as I` for some type `C`
and facet type `I`. It is named `FacetValue` instead of just `Facet` to
parallel the `FacetType` instruction.
This PR uses this instruction represent the facet value `Self` in an
`impl` declaration. This instruction will be used in the future to also
support things like:
* `C as I` where `C` is a class; and
* forming a specific for a generic with a `T:! I` parameter where `T` is
being given a concrete value.
(Here `I` is an interface or other non-`type` facet type.)
Also do some renaming and add some comments to make things a bit more
clear.
* `FacetTypeAccess` -> `FacetAccessType` to clarify this is not access
of a facet type, but access of the type of a facet
* `.facet_id` -> `.facet_value_inst_id` to parallel the `FacetValue`
instruction
`FacetAccessWitness` will be in a future PR.
---------
Co-authored-by: Josh L <josh11b@users.noreply.github.com>
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
In this case, the callee may be non-constant because it includes a
reference to `self`, so we need to be able to lower a non-constant
`specific_function`.
Also propagate the pattern IR along with the pattern-match IR, and use
it where appropriate.
Strictly speaking, some parts of the pattern-match IR are allocated
eagerly, while traversing the pattern's parse tree, but they still
aren't actually emitted until we traverse the associated pattern insts.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
This switches `DCHECK` and `FATAL` as well.
The goal is to reduce the code size impact of these assertions so that
we can keep more of them enabled. Currently, the largest cost I see from
`CHECK` is not the actual check or the cold code itself, but actually
the failure to inline trivial functions due to the presence of the cold
code. This means that our goal isn't to reduce apparent code size in the
final binary but the LLVM IR cost assessed for these routines in the
inliner, which closely correlates with code size but is a bit different.
As discussed in #4283, experimentation shows that a single function call
with a minimal number of arguments is the lowest cost model for these.
This is easily achieved with a format-string API that internally uses
`llvm::formatv`. This PR is essentially the `CHECK` version of #4283.
However, the check macros are substantially harder to make work with
both format strings and streaming because they also take a condition.
Also, unexpectedly, I was very successful at devising a regular
expression based automated rewrite from the streaming to the format
string form with only low 10s of manual fixes. This includes compacting
strings broken up across lines, etc. Given how well that went, I've
prepared this PR which just directly switches to the format string API
and migrate everything to use it.
One nice side-effect is that the format string approach ends up greatly
simplifying the implementation here as well.
This is ... *shockingly* effective. Parsing speeds up by more than 3%
with just this change. And checking speeds up by **8%** with this change
alone:
```
BM_CompileAPIFileDenseDecls<Phase::Parse>/256 86.3µs ± 1% 82.9µs ± 1% -3.94% (p=0.000 n=17+19)
BM_CompileAPIFileDenseDecls<Phase::Parse>/1024 431µs ± 1% 415µs ± 1% -3.76% (p=0.000 n=18+19)
BM_CompileAPIFileDenseDecls<Phase::Parse>/4096 1.77ms ± 1% 1.71ms ± 1% -3.18% (p=0.000 n=18+19)
BM_CompileAPIFileDenseDecls<Phase::Parse>/16384 7.44ms ± 1% 7.17ms ± 2% -3.56% (p=0.000 n=18+20)
BM_CompileAPIFileDenseDecls<Phase::Parse>/65536 30.7ms ± 1% 29.7ms ± 1% -3.15% (p=0.000 n=18+20)
BM_CompileAPIFileDenseDecls<Phase::Parse>/262144 131ms ± 1% 127ms ± 1% -2.81% (p=0.000 n=18+18)
BM_CompileAPIFileDenseDecls<Phase::Check>/256 878µs ± 2% 800µs ± 1% -8.91% (p=0.000 n=19+20)
BM_CompileAPIFileDenseDecls<Phase::Check>/1024 1.88ms ± 2% 1.72ms ± 1% -8.56% (p=0.000 n=19+20)
BM_CompileAPIFileDenseDecls<Phase::Check>/4096 5.78ms ± 2% 5.28ms ± 1% -8.70% (p=0.000 n=20+18)
BM_CompileAPIFileDenseDecls<Phase::Check>/16384 21.9ms ± 1% 20.1ms ± 1% -8.02% (p=0.000 n=18+20)
BM_CompileAPIFileDenseDecls<Phase::Check>/65536 90.4ms ± 2% 83.1ms ± 1% -8.04% (p=0.000 n=19+20)
BM_CompileAPIFileDenseDecls<Phase::Check>/262144 381ms ± 2% 352ms ± 1% -7.79% (p=0.000 n=19+19)
```
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Co-authored-by: josh11b <15258583+josh11b@users.noreply.github.com>
This adds just the debug info metadata for Compilation Units (the top
level container of debug info) - but without anything in them, LLVM
won't emit them at all, so while this is testable at the IR level, it
isn't observable at the object level until more debug info is added.
A couple of starting points in this patch:
* A flag (`--debug-info`, seems to match the naming/style of other flags
in the carbon driver, though this is different from the naming
conventions of clang/gcc) that enables debug info when lowering. Open to
other names/approaches (on by default? historically debug info's been to
large/expensive to do this, so sticking with that precedent for now).
* Enabling that flag by default in the lowering tests - I do find the
churn on golden tests a bit rough, and adding more features to all the
tests means more churn, but it seems consistent with the approach so far
- keep an eye on this and perhaps revisit this if the churn gets too
annoying
---------
Co-authored-by: Carbon Infra Bot <carbon-external-infra@google.com>
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
Rename `ReturnInfo` to `ReturnTypeInfo`. Move it and `InitRepr` into
`type_info.h` alongside `ValueRepr`. Replace `ReturnSlot` with
`InitRepr`, and extend `InitRepr` to be able to represent the
incomplete-type case instead of CHECK-failing. Remove `has_return_slot`
from `InitRepr` and instead only provide that as part of
`ReturnTypeInfo`.
Renames Lower::Handle* to Lower::LowerFunctionInst. This allows writing
a templated handler for instructions, moving code out of the macro
expansion and removing some of the redundancy in things like
`HandleAddrOf(..., SemIR::AddrOf inst)`
This adds `DefinitionInfo` for `Define`-based configuration so that
parameters are optional. It also makes it easier to provide the
equivalent functions on both `Definition` and `Define`.
A common pattern used here is to change from a `switch` with in-line
`case`s to instead have `case`s that call an overloaded function. What's
happening here is that the instruction type is used to select an
overload, and if an overload is not defined, a compiler error would
result. Meanwhile, clusters of overloads are being defined using
`requires`-based templating, so that equivalent implementations are not
copied. This addresses a limitation of a vanilla `switch` approach where
it's hard to have redundant cases using conditional logic, while also
getting compiler errors when adding new `InstKind` entries, which had
been a significant part of why we used macros previously.
This starts hitting some odd clang-format edge cases causing
`CARBON_KIND_SWITCH(inst){` (missing space), which I haven't seen
before. Adding `CARBON_KIND_SWITCH` to .clang-format works around it.
This is to remove all the FatalIfEncountered handlers in handle.cpp.
They just feel like noise when reading the file. Plus it's one less bit
of boilerplate to add for instructions that don't lower.
Note that I left HandleParam/HandleAddrPattern. I'd be happy to change
those to just set lowered=false too, but was hesitant to given the
separate logic.
Also, I'm separately considering migrating the macro logic into similar
constexpr things. If I do, I might switch Define to take in a struct.
But for how this particular parameter works, the overload felt
reasonable, particularly since is_lowered is not used in combination
with TerminatorKind.
Lucile Rose Nihlen
Geoff Romer
Richard Smith
Dana JansensandRichard Smith
Alina Sbirlea
