This is trying to make it clearer when vectors are being indexed with
`CheckIRId`.
The only one that I still kind of want to change is the
`SmallVector<std::unique_ptr<CompilationUnit>>`, but because it's a
`unique_ptr` that's a little more complex. I may not bother.
Note, some of the changes around nuanced `SmallVector` interactions were
based on trying to copy the way `SmallVector` itself takes arguments,
like with range passing.
Suggested by zygoloid while looking at #5678
```
CHECK failure at toolchain/lower/context.cpp:62: !llvm::verifyModule(*llvm_module_, &errs): Verifier errors: Instruction does not dominate all uses!
%.loc17_46.1.temp = alloca { i1, i32, i32 }, align 8, !dbg !13
%tuple.elem0.loc17_46.2.tuple.elem = getelementptr inbounds nuw { i1, i32, i32 }, ptr %.loc17_46.1.temp, i32 0, i32 0, !dbg !13
Instruction does not dominate all uses!
%.loc17_46.1.temp = alloca { i1, i32, i32 }, align 8, !dbg !13
%tuple.elem1.loc17_46.2.tuple.elem = getelementptr inbounds nuw { i1, i32, i32 }, ptr %.loc17_46.1.temp, i32 0, i32 1, !dbg !13
Instruction does not dominate all uses!
%.loc17_46.1.temp = alloca { i1, i32, i32 }, align 8, !dbg !13
%tuple.elem2.loc17_46.2.tuple.elem = getelementptr inbounds nuw { i1, i32, i32 }, ptr %.loc17_46.1.temp, i32 0, i32 2, !dbg !13
```
Adds a `--llvm-verifier` flag to be able to turn this off easily,
particularly for debugging the LLVM IR.
The call workaround is due to a verifier requirement `inlinable function
call in a function with debug info must have a !dbg location`. It
specifically comes up for the `++x` case, with `%1 = call i32
@"_CConvert.8b3d5d6a6c17be04:ImplicitAs.Core.b88d1103f417c6d4"(i32
%other)`. I think #5397 is in the direction of a fix for that, but #5397
was set aside because it puts the debug info in too many places.
Instead, address this by adding a stub location for calls that don't
have a good location. I'm deliberately putting this next to the TODO so
that it's easier to understand the association.
---------
Co-authored-by: Dana Jansens <danakj@orodu.net>
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>
Factor out the logic for mapping from a `LocId` into a diagnostic
location from check into sem_ir so it can be reused by lowering. Include
the function and instruction being lowered in the pretty stack trace.
Example stack trace:
```carbon
2. filename: examples/sieve.carbon
3. core/prelude/types/int.carbon:213:3: lowering function Core.Op(Core.IntLiteral as Core.ImplicitAs(i32))
fn Op[addr self: Self*](other: Self) = "int.sadd_assign";
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4. core/prelude/operators/arithmetic.carbon:22:27: lowering call
fn Op[addr self: Self*](other: Other);
^~~~~~~~~~~~
```
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.
Previously we walked the global variables defined by the current file
and emitted an LLVM global variable definition for each of them. Now
instead, when emitting a constant reference to a global variable, we
emit an LLVM global variable declaration, and we then subsequently walk
the global variables defined by the current file and convert each of
them from a declaration to a definition.
In order to make import of names of global variables work, add support
for import of `var`, as well as support for importing `tuple_access` and
`tuple_pattern` in the case where the `var` has a tuple pattern in its
declaration. Also treat `bind_name`s that are reference bindings to
`var`s as having the same constant reference value as their `var` so
that we can properly import and lower them.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
Co-authored-by: Carbon Infra Bot <carbon-external-infra@google.com>
If a function contains instructions whose locations are in another file,
skip providing debug locations for those instructions rather than
CHECK-failing.
This happens when emitting a thunk where the signature is declared in
one file and the call target is in another file: some parts of the thunk
use the original signature as their locations, whereas other parts of it
use the location of the call target.
`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.
Provide builtins for compound assignments instead of defining them in
the prelude as a use of a binary operator and an assignment. This allows
us to lower compound assignment directly to LLVM operations instead of
producing a function call. In the short term this also allows us to
define a type-generic compound assignment in the prelude.
Replace the large and growing `TryEvalInstInContext` function with one
function per kind. While we still have special-case handling for a small
number of instruction kinds, most instructions are now handled either
fully automatically or use a common codepath that evaluates the
instruction operands and then performs an eval-context-independent
evaluation of the instruction.
To support this, `InstConstantKind` is expanded to describe more
fine-grained details about how each kind of instruction interacts with
constant evaluation. Also, the operand kinds of instructions become
slightly more fine-grained: we now distinguish between operands that
describe the destination of an initializing expression (`DestInstId`)
from other `InstId` operands, because `DestInstId` operands need
different treatment during constant evaluation. In particular, an
initializing expression can have a constant value even if its
destination is non-constant or has not yet been set, because evaluation
of an initializing expression doesn't include the store to the
destination.
Some minor test changes:
- We now more consistently propagate errors into the results of constant
evaluation, so more instructions that depend on errors have a constant
value of `<error>`.
- Diagnostic location for invalid array types now point at the whole
array type rather than the array index expression, because
`EvalConstantinst` doesn't have access to the original expression.
- Diagnostic for failed `RequireCompleteType` doesn't print the original
type any more because `EvalConstantInst` doesn't have access to the
original expression.
As a follow-up, some of this -- in particular, the `EvalConstantInst`
overloads -- will be moved to a separate file, in an effort to split the
overall constant evaluation machinery apart from the logic to evaluate
each individual kind of instruction.
Noted CopyNameFromImportIR while glancing around (this one's interesting
because it's NameId, not void nor auto), did a scan just for a few other
cases. Not an exhaustive fix, and TBH assuming we'd prefer `auto ... ->
auto` since equivalent Carbon syntax would probably be `fn ... -> auto`
High level, replacing `Id::Invalid` with `Id::None` and `Id::is_valid`
with `Id::has_value` for clarity, as discussed
[here](https://discord.com/channels/655572317891461132/655578254970716160/1331664574545395794).
The `IntId` refactoring is needed together with `AnyIdBase` because it's
also used with `ValueStore`.
Note, trying to be careful not to rewrite `EnumBase::InvalidIndex`, or
`is_valid` in general (e.g., `IdKind::is_valid`).
I've tried to sequence commits here:
1. Automatic replacements:
- `((?:Id|Index)(?: |::|\(|Base(?:\(|::)))Invalid((?:Index)?\W)` ->
`$1None$2`
- `<invalid>` -> `<none>`
- `InvalidNodeId` -> `NoneNodeId`
- `/\*invalid\*/` -> `/*none*/`
- `id((?:_|\(\))(?:\.|->))is_valid` -> `id$1has_value`
2. Manual edits:
- In `int.h` and `int_test.cpp`
- `IntT` has `is_value`, which I'm renaming to `is_embedded_value`.
- Manual edits to comments in this file.
- `AnyIdBase` and `IdBase`
- Declaration of `is_valid` -> `has_value`, `InvalidIndex` ->
`NoneIndex`.
- In `ids.h` and `ids.cpp`
- `is_valid` -> `has_value`
- `// An explicitly invalid ID.` -> `// An ID with no value.`; similar
for index
- Various math on `InvalidIndex` -> `NoneIndex`
- Various mentions of "valid" in comments
- In `value_store.h`, for `IdT::Invalid`, plus one comment
- In `impl.h` and `tokenized_buffer.h`, we had different initialization
of `::None` values (versus `ids.h` syntax) that I fixed manually.
- Spot checks to compile
- Particularly where `is_valid` replacements didn't catch spots due to
different naming.
3. Autoupdate tests
4. verbose.carbon (NOAUTOUPDATE)
5. Comment spot checks
Note there are probably other mentions of "Invalid" that should be swept
up, but I'd like to argue for merging and separating out remaining
cleanup since this is so sweeping (and likely to hit merge conflicts
from churn). We'll probably have lingering mentions of "invalid" for a
bit regardless, just because there are uses of "invalid" in non-Id APIs.
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>
Mainly, changes the default from -1 to 0 in DiagnosticLoc, still trying
to keep reusing that. Nothing except for the lowered output is affected,
so I think this is fine.
Also, have lowering consistently call GetDiagnosticLoc.
Pulls in one of the CHECKs suggested from #4251
Co-authored-by: David Blaikie <dblaikie@gmail.com>
The source line debug info generation assumed that the function would
have debug info. Check for a non-null di_subprogram_ to ensure we are
emitting debug info for the function.
Rather than checking the di_builder_ - this way if we implement
`nodebug` function attributes, it'll fall out naturally (by creating a
null di_subprogram_) rather than having to come back and change this
from "is debug info enabled" to "is debug info enabled for this
function" later on.
Seems to work with lldb ( https://pastebin.com/igKkNECm ), though gdb
has /some/ trouble with the paths (they aren't complete - just using the
filename directly, not providing the working directory - might be some
quick hacks that can help there).
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.
This works to leverage the capabilities of the hashtable as much as
possible, for example using the key context in the value stores.
However, there may still be opportunities to refactor more deeply and
use the functionality even better. Hopefully this is at least
a reasonable start and gets us a clean baseline.
On an Arm M1, this is a 15% improvement on my large lexing stress test,
but ends up a wash on my x86-64 server. This is a smaller benefit than
I expected, and it's because we're using a set-of-IDs and looking up
values with a key context for things like identifiers. This pattern has
a surprising tradeoff. The new hashtable uses significantly less memory,
a 10% peak RSS reduction just from the hashtable change. But indirecting
through the vector of values makes growing the hashtable dramatically
less cache-friendly: it causes growth to randomly access every key when
rehashing. On x86, everything gained by the faster hashtable is lost in
even slower growth. And even on Arm, this eats into the benefits.
But I have a plan to tweak how identifiers specifically work to avoid
most of the growth, and so I suspect this is the right tradeoff on the
whole. It gives us significant working set size reduction and we can
likely avoid the regressed operation (growth with rehash) in most cases
by clever reserving and if necessary by adding a hash caching layer to
the table infrastructure.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
Make constant emission non-recursive, and stop building a bogus
FunctionContext to emit constants.
To support this, move `InstConstantKind` from the typed instruction
definition into the `.def` file, and add more macros to allow us to
generate case labels based on whether an instruction is a constant.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
Co-authored-by: Carbon Infra Bot <carbon-external-infra@google.com>
First steps towards using constant values in lowering.
For now, we reuse the regular instruction lowering to lower constants.
This mostly works, because we don't actually need an `llvm::Function` or
a current basic block when lowering a constant most of the time.
However, a special case is needed for lowering aggregate value constants
because they would otherwise create a stack alloca to store the
constant. Separate constant lowering code will be added in a future
change to clean this up.
When lowering a constant initializing expression, the result is a value
of the destination type, rather than code to initialize the destination,
so a separate copy step is required when finishing initialization from a
constant for a type that uses in-place initialization. Handling this
required extending `ReturnExpr` to track its destination location.
We currently often create non-constant `*_access` SemIR instructions
that are only used by constant `*_init` instructions. These cause
lowering to leave behind `getelementptr` instructions in the lowered IR
that are now unused. It should be possible to detect this case and avoid
producing these instructions, or to produce them lazily, but for now
we're just leaving them around for LLVM to clean up.
Factor out `SemIR::InstNamer` and also use it when lowering to LLVM IR.
Automatically name all instructions created with our `IRBuilder` based
on the name computed by the `InstNamer`, and likewise name basic blocks
using the label generated by the `InstNamer`.
Move some of the existing naming logic out from lower into `InstNamer`
so that it's also used in SemIR. In particular, we now name call
instructions after their callee, or after the builtin name for calls to
builtins.
Computing and adding these names isn't completely free. This instruction
naming is designed to be optional, so that we can turn it off for builds
where the LLVM IR will only be converted to assembly and won't be seen
by a human, but so far it's enabled unconditionally. We can tune that
later as needed.
Similar to #3705, we actually have a mix of `Make` and `Create` in
factory functions too, so this PR is normalizing on `Make`. It's
intended to be consistent with the naming choice for Carbon factory
functions.
Note, MakeSyntheticBlock is the only one I feel a little weird about
because llvm's own APIs use Create, and this is essentially wrapping
LLVM calls. But the flipside is it also feels like a vague line to draw,
when we also differ from LLVM coding style in other ways.
Recent runs of `clang-tidy` for me started showing more errors, and this
is a collection of changes to address them.
First, I've systematically applied the disabling tag to all C++ rules
under //explorer/... with `buildozer` so we don't spend time analyzing
this code or reporting errors from it. Not sure this was strictly
necessary, but it seemed like a nice consistency improvement.
Next, I disabled a buggy check for missing `default` cases in
`switch`es. It seems to get confused by the fancy conversions in our
`enum_base.h`. We don't miss much with this as the Clang compiler
warnings for `switch` catch most of our actual bugs. I also removed the
local disabling of this now that it is turned off centrally.
Lastly, I added error checking to two file descriptor manipulating calls
in the `file_test` infrastructure.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
Instead of ad-hoc conversion tracking on some kinds of nodes that
conversion creates, consolidate tracking into a single node kind. This
frees up an operand on `Init` instructions that can be used to store the
destination.
Adds a `BoundMethod` SemIR node to represent an `x.F` bound method, with
a new builtin type `BoundMethodType`. Reorganized conversion of call
expression arguments to also check and convert a `self` parameter in the
implicit parameters list.
In passing, improved diagnostics and error recovery for bad call
expressions. We now build a `call` node with the appropriate type and
value category, but with invalid arguments, if the argument conversion
failed, and diagnose calls to non-callable expressions.
`addr self` methods don't work properly yet; the `addr` is ignored for
now.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
1. In general, `semantics_ir` -> `sem_ir`, to match the directory name.
2. For the list of `ValueStore`-related accessors on `SemIR::File`, add
them to `check`'s `Context` object, shortening access.
Finishing what #3316 started, add more bespoke ValueStore-like
structures to File. With this, the things which previously had somewhat
boilerplate Add/Get functions are now all on side classes, giving a
uniform style of API for calling.
Note, I was on the fence about making things public on ValueStore. If
it's preferred that I make some things there protected I certainly can,
there's just a trade-off that may mean more distinct child/wrapper
types.
Using the computed value representation, fix lowering of struct and
tuple values to use the value representation rather than the object
representation. Fixes an issue found in the review of #3257.
This currently causes us to compute value representations of all types
as they are created, which generates substantially more SemIR to
represent types. We can get some of that back by deferring computation
of the value representation until the type is required to be complete,
but some of the additional cost here will persist with this approach.
I also considered making the computation of the value representation
type be something that lives entirely within the lowering phase, but I
think that's not the right approach in the longer term, because the
value representation will be semantically visible and relevant once we
start allowing it to be customized.
We should consider moving the nodes that exist to compute canonical
non-local types, including value representations, out into a separate
global block. That will clean up the SemIR representation substantially,
and make the SemIR produced for a function not depend on which types we
happen to have encountered beforehand. But that's not being done in this
PR.
---------
Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
This is shorter, more closely connected to code using the typed node
types, and avoids using the ambiguous word `Node` in places referring to
typed `SemIR` nodes.
Replace `SemIR::Node::GetAsFoo` and `SemIR::Node::Foo::Make` with
`SemIR::Foo` class that represents a particular kind of node, with named
fields.
Rename `SemIR::IntegerLiteral` and `SemIR::RealLiteral` to
`IntegerValue` / `RealValue` to better reflect their purpose and avoid a
name collision with the corresponding `SemIR` node kinds.
Remove `NodeKind::Invalid` and the `SemIR::Node` default constructor
entirely, as they were not used for anything.
Trust semantics to have put them in the right places.
Many parts of lowering still need to be updated to use the value
representation chosen at the semantics layer, but this is an incremental
step towards that.
Fix a bug where we would perform the computation of the return location
in SemIR after we have already used it in some cases, leading to
assertion failures during lowering. Instead, accumulate a sequence of
instructions to compute the return location in a temporary block, and
overwrite the return slot with those instructions when we perform
initialization.
StubReference is replaced by a more general SpliceBlock node, that takes
a code block and a result value, executes the instructions in the block,
and produces the result. This is used in the uncommon case where more
than one instruction is required to compute the return slot, which can
happen if we need to first emit a temporary and then index into it, or
if we need to perform multiple levels of indexing before we reach an
entity to initialize.
Continuing along with #3070. Note this is just a file rename, with BUILD
edits; every file previously in semantics/ should show as moved (except
maybe BUILDs, which split).
Jon Ross-Perkins
Richard Smith
Alina Sbirlea
Danial Klimkin
josh11b