-
Notifications
You must be signed in to change notification settings - Fork 156
[CIR][CodeGen] Introduce CIR CXXSpecialMember attribute #1711
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for adding more of the building blocks here. In the future it'd be nice to have a representation for struct/class more closer to C++ and it will probably be easier to extract this type of information, but to prevent over engineering early, it feels right to incrementally add pieces that can enable us to better analyze C++ code and make transformations easier to write.
Can you look into changing LifetimeChecker.cpp to use this attribute instead of the current AST approach?
(@andykaylor @erichkeane @dkolsen-pgi in case you have any extra thoughts here) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do we also want to set this attribute on cir.func definitions/declarations?
68ff6ef
to
8c7235c
Compare
✅ With the latest revision this PR passed the C/C++ code formatter. |
A few updates:
|
77c8851
to
d4c51f3
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
1 comment, else seems reasonable, I'm happy when the others are.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the update, one more round of suggestions
ab65415
to
c1f8a62
Compare
c1f8a62
to
f6aaa4d
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the updates, one more round!
@@ -2798,6 +2813,15 @@ void cir::FuncOp::print(OpAsmPrinter &p) { | |||
p.printAttribute(annotations); | |||
} | |||
|
|||
if (getCxxSpecialMember()) { | |||
p << " cxx_special_member<"; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why do we need to wrap it in cxx_special_member
? It is apparent from cir.cxx_ctor/dtor
that this is the special member function.
a5f2b05
to
c40a840
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for your patience, more comments!
@@ -2618,6 +2620,37 @@ ParseResult cir::FuncOp::parse(OpAsmParser &parser, OperationState &state) { | |||
state.addAttribute(annotationsNameAttr, annotations); | |||
} | |||
|
|||
// Parse CXXSpecialMember attribute | |||
if (mlir::succeeded(parser.parseOptionalKeyword("cxx_ctor"))) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Still sounds simple enough that we shouldn't need these to be hand written, what specific problems are you hitting?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I can update this once the syntax of the attribute is approved. I am also happy with member_function_of<type, ctor<copy>>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! I'll defer this to @xlauko cause I'm going to be out until next week. One important part here is not to have a custom parser/printer because this is simple enough that it doesn't need one (see more examples in the LLVM dialect for inspiration).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bcardosolopes I tried working around these custom printers/parsers, and I ran into some issues.
Using the CXXCtorAttr
attribute as an example, the format we want to use is member_function_of<type, ctor<Kind>>
. The issue with using parseOptionalAttribute
following a printAttribute
, for example, to avoid manual parsing is that it tries to parse the full attribute, starting with its attribute mnemonic #cir.cxx_ctor
!!
Both printAttribute
and parseAttribute
/parseOptionalAttribute
start with the attribute mnemonic!
No matter how much you change the assembly format the mnemonic starts anyway, because CXXCtorAttr
inherits from CIR_Attr
. This is why I arrived at this version using the "neater parseOptionalAttribute"
previously.
I think the easiest solution is having custom printers/parsers for these attributes and for CXXCtorAttr
(as an example) we can print/parse the type
and the ctorKind
as we want, using whatever format.
I played around with these printers/parers a bit and I wasn't able to find alternatives, unfortunately.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for your patience and many different attempts here. Let's try to expedite this so you can move on to more interesting work. What I'd like to see:
- No custom printing/parsing. You can drop the
special_member_of
and just have a#cir.cxx_ctor
if that's what it takes. If you introduce an alias for it you can reduce it further, but I would leave that to another PR to land this faster. - Follow the TBAA route of attribute inheritance and drop the AnyAttrOf for the SpecialMemberOf attribute.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the feedback.
I have updated the PR and now there is no custom printing/parsing. The constructor attribute, for example, looks like special_member<#cir.cxx_ctor<Type, Kind>>
.
For AnyAttrOf
, the idea was actually gotten following the TBAA route. I do not think we are able to use direct attribute inheritance. I tried and got some crashes related to attribute constraints.
// CIR-NEXT: cir.func private @_ZN1EntEv(!cir.ptr<!rec_E>) -> !rec_E | ||
// CIR-NEXT: cir.func private @_ZN1ED1Ev(!cir.ptr<!rec_E>) extra(#fn_attr) | ||
// CIR-NEXT: cir.func private @_ZN1ED1Ev(!cir.ptr<!rec_E>) cxx_dtor<!rec_E> extra(#fn_attr) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sounds like we could also drop "cxx": dtor<!rec_E>
is probably good enough? No need to change this right now, let's converge on the syntax first (see other comment).
def CIR_CXXCtorAttr | ||
: CIR_Attr<"CXXCtor", "cxx_ctor"> { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
def CIR_CXXCtorAttr | |
: CIR_Attr<"CXXCtor", "cxx_ctor"> { | |
def CIR_CXXCtorAttr : CIR_Attr<"CXXCtor", "cxx_ctor"> { |
def CIR_CXXDtorAttr | ||
: CIR_Attr<"CXXDtor", "cxx_dtor"> { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
def CIR_CXXDtorAttr | |
: CIR_Attr<"CXXDtor", "cxx_dtor"> { | |
def CIR_CXXDtorAttr : CIR_Attr<"CXXDtor", "cxx_dtor"> { |
let assemblyFormat = [{ | ||
`<` $type `>` | ||
}]; | ||
// Printing and parsing also available in CIRDialect.cpp |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
// Printing and parsing also available in CIRDialect.cpp |
let assemblyFormat = [{ | ||
`<` $type `,` $ctorKind `>` | ||
}]; | ||
// Printing and parsing also available in CIRDialect.cpp |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
// Printing and parsing also available in CIRDialect.cpp |
Attribute should either use default printer/parser or custom, only in latter case it is available in CIRDialect.cpp
which obvious from the context.
let builders = | ||
[AttrBuilderWithInferredContext<(ins "mlir::Type":$type), [{ | ||
return $_get(type.getContext(), type); | ||
}]>]; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
let builders = | |
[AttrBuilderWithInferredContext<(ins "mlir::Type":$type), [{ | |
return $_get(type.getContext(), type); | |
}]>]; | |
let builders = [ | |
AttrBuilderWithInferredContext<(ins "mlir::Type":$type), [{ | |
return $_get(type.getContext(), type); | |
}]> | |
]; |
let builders = | ||
[AttrBuilderWithInferredContext<(ins "mlir::Type":$type, | ||
CArg<"CtorKind", "cir::CtorKind::Custom">:$ctorKind), [{ | ||
return $_get(type.getContext(), type, ctorKind); | ||
}]>]; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
let builders = | |
[AttrBuilderWithInferredContext<(ins "mlir::Type":$type, | |
CArg<"CtorKind", "cir::CtorKind::Custom">:$ctorKind), [{ | |
return $_get(type.getContext(), type, ctorKind); | |
}]>]; | |
let builders = [ | |
AttrBuilderWithInferredContext<(ins "mlir::Type":$type, | |
CArg<"CtorKind", "cir::CtorKind::Custom">:$ctorKind), [{ | |
return $_get(type.getContext(), type, ctorKind); | |
}]> | |
]; |
I think this one is self-explanatory, so I will not write much 🙂
Adding this attribute helps in optimizations like #1653, and using the attribute it's easy to create operations like
cir.std.vector.ctor
/cir.std.vector.dtor
by just modifyingIdiomRecognizer
a bit. I believe it will also be useful for future optimizations. Finally, I updated quite a number of tests so they now reflect this attribute.Please, let me know if you see any issues.