Using annotation attributes

Project Infrastructure LLVM Dev List Archives
1

Hi,

I'm trying to annotate certain functions in C code, and do something with
these functions in my LLVM pass. I annotate the C code like this:

int __attribute__((annotate("annot"))) function() {

This nicely gets added to the LLVM bitcode in an
@llvm.global.annotations global. Now I had hoped that it'd be easy to extract
a list of functions annotated with my annotation using

AnnotationManager::getID("annot")
and later on:
function->getAnnotation(AnnotID).

This does not seem to work, unfortunately. Is this supposed to work in this
way, or am I using the wrong functions? I could of course just try to
manually parse the @llvm.global.annotations constant, but that seems rather
tiresome :slight_smile:

Thanks,
Bart Coppens

2

I'm trying to annotate certain functions in C code, and do something with
these functions in my LLVM pass. I annotate the C code like this:

int __attribute__((annotate("annot"))) function() {

This nicely gets added to the LLVM bitcode in an
@llvm.global.annotations global. Now I had hoped that it'd be easy to extract
a list of functions annotated with my annotation using

AnnotationManager::getID("annot")
and later on:
function->getAnnotation(AnnotID).

This does not seem to work, unfortunately. Is this supposed to work in this
way, or am I using the wrong functions? I could of course just try to
manually parse the @llvm.global.annotations constant, but that seems rather
tiresome :slight_smile:

There is no way to get it from the Function*. You have to get the global annotations array and parse it yourself. You could easily write a pass to do this for you and store it in a map.

I think the main reason is that we don't want to bloat the IR by making space for annotations in the Function itself.

-Tanya

3

There is no way to get it from the Function*. You have to get the global
annotations array and parse it yourself. You could easily write a pass to
do this for you and store it in a map.

Ok, I will do that. In this context, is it possible that the code in
lib/Transforms/IPO/InlineSimple.cpp
has bitrotted somewhat? "Noinline" seems to reside in @llvm.global.annotations
as well, but the code doesn't look like it is designed to handle that.

I think the main reason is that we don't want to bloat the IR by making
space for annotations in the Function itself.

Right. I had just hoped that even though the link is not explicit in the IR,
it would have been preprocessed anyway (given the existence of
AnnotationManager).

Bart

4

Hi all,

I've also been developing an interest in using IR annotations for my compiler.
Some discussion with Bart turns out that he has implemented some code to parse
the llvm.globals.annotations array, but in no way integrated or reusable.
We've spent some thought about how this could be done properly, which I will
share here.

Firstly, however, I was wondering about the format of the
llvm.globals.annotations array. It does not seem to be defined in the LLVM
language reference, shouldn't it be? It's name suggests that it is a reserved
variable name with a fixed type (similar to intrinsic functions?).

Furthermore, it seems that the AnnotationManager that is currently implemented
is capable of keeping a list of Annotations for any Annotatable (currently
only Function). These annotations are kept in memory only and really have
nothing to do at all with the annotations in the IR.

Still, it seems that using the AnnotationManager to make the IR annotations
accessible seems like a decent approach.

The way I see this is having some pass, or probably the assembly reader or the
AnnotationManager itself, parsing the llvm.global.annotations variable and
adding annotations to the corresponding GlobalValues. This would just leave the
annotations in the IR as well, so that transformation passes would properly
preserve them (and, just like debug info, sometimes be prevented from
modifying some annotated global values unless they are taught how to preserve
the annotations).

By using a subclass of Annotation (say, GlobalAnnotation) we can distinguish
between annotations that are (or should be) in the IR and (the existing)
annotations that should be in memory only. This would also allow for newly
added annotations to be immediately be added to the IR, ensuring that the
AnnotationManager's view remains consistent with the IR.

In this way, any annotations added will be implicitely output when the IR is
outputted, without any special support on the output end.

A problem I could imagine using this approach would be name conflicts. Since
any annotation name could come from the IR, these could conflict by the other
names already in use (such as "CodeGen::MachineFunction" IIRC). This could be
solved by using a "GlobalAnnotation::" prefix for the name, or something
similar.

A completely alternative approach would be to just use a global map to
facilitate annotation lookups, not using AnnotationManager at all. I can see a
couple of drawbacks here. Firstly, it's a bit confusing that AnnotationManager
would not handle IR annotations. Some decent comments and/or renaming could
sovle that. Secondly, there is a speed gain from using AnnotationManager.
Because annotations are stored at the Annotatable in question, instead of in a
global map, annotations can be retrieved quickly. Also, annotations names are
mapped onto integers, allowing for even faster comparisons and lookups.

So, any thoughts?

Gr.

Matthijs

5

Hi all,

Howdy Matthijs,

I've also been developing an interest in using IR annotations for my compiler.
Some discussion with Bart turns out that he has implemented some code to parse
the llvm.globals.annotations array, but in no way integrated or reusable.
We've spent some thought about how this could be done properly, which I will
share here.

Ok, cool. Annotations are tricky to do right :slight_smile:

Firstly, however, I was wondering about the format of the
llvm.globals.annotations array. It does not seem to be defined in the LLVM
language reference, shouldn't it be? It's name suggests that it is a reserved
variable name with a fixed type (similar to intrinsic functions?).

Yes, we should document it. It is a convention established by the __builtin_annotate function in the c compilers. We should standardize it and document it.

Furthermore, it seems that the AnnotationManager that is currently implemented
is capable of keeping a list of Annotations for any Annotatable (currently
only Function). These annotations are kept in memory only and really have
nothing to do at all with the annotations in the IR.

Yes, this is a really old mechanism that we should rip out. MachineFunction should be moved to be an analysis that is preserved as an actual part of the passmanager, instead of being a thing we tack onto the Function object. We have killed all uses of this old annotation mechanism except MachineFunction.

Still, it seems that using the AnnotationManager to make the IR annotations
accessible seems like a decent approach.

I agree that *having* an annotationmanager makes sense, but the existing one should die and be replaced. :slight_smile:

The way I see this is having some pass, or probably the assembly reader or the
AnnotationManager itself, parsing the llvm.global.annotations variable and
adding annotations to the corresponding GlobalValues. This would just leave the
annotations in the IR as well, so that transformation passes would properly
preserve them (and, just like debug info, sometimes be prevented from
modifying some annotated global values unless they are taught how to preserve
the annotations).

Makes sense. This is similar to how the MachineDebugInfo stuff deserializes debug info out of the LLVM IR and presents it for easy consumption of the code generator.

By using a subclass of Annotation (say, GlobalAnnotation) we can distinguish
between annotations that are (or should be) in the IR and (the existing)
annotations that should be in memory only. This would also allow for newly
added annotations to be immediately be added to the IR, ensuring that the
AnnotationManager's view remains consistent with the IR.

I think we need to distinguish between two forms of annotation:

1. there are some "annotations" like "readonly", "nounwind", etc that are baked into the LLVM IR and are/should be documented in LangRef.

2. There are annotations that are really "cheap extensions" of the LLVM IR that are either experimental, very domain specific, or that are just metadata about the code.

For #1, the current "parameter attributes" we have work reasonable well, and Devang is actually cooking up a proposal to extend them a bit (to fix some issues with LTO). #2 is something that llvm.annotate handles reasonable well, but I agree it would be great to have a nice interface to update/read them.

The advantage of #1 is that the compiler as a whole knows about the attributes, but this means that adding one is "hard". The advantage of #2 is that they are easy to add, but they have limitations and can impact codegen (e.g. they disable IPO in some cases).

A problem I could imagine using this approach would be name conflicts. Since
any annotation name could come from the IR, these could conflict by the other
names already in use (such as "CodeGen::MachineFunction" IIRC). This could be
solved by using a "GlobalAnnotation::" prefix for the name, or something
similar.

It could also be served by making them completely string based, and just provide a simple string interface? That way you don't need classes for each attribute.

-Chris

6

Hi all,

We are interested in implementing a loop-aware (in the sense memory dependences are tracked across back edges especially) memory profiling tool in LLVM. We need to annotate profile information on individual instructions. Basically our earlier profiling system used to work as follows:

(1) The first pass in the compiler instruments (individual load/store instructions) the program with calls to the profiling library which keeps track of the src/dst addresses of load/store instructions and how often they are executed.
(2) The instrumented program is run, with some input and the profiling library generates its output into a separate file.
(3) The second pass in the compiler reads in this profile information and annotates individual load/store instructions with the set of other load/store instructions (along with profile weights) that it is dependent on for the current input set. This information is then used by the subsequent passes and eventually dropped before code generation.

What would be a good way to get started on implementing this sort of system in LLVM ? Specifically: As far as we know, there is no way to give individual instructions ids/tags so that pass (3) can uniquely identify them using these tags. Is this correct or is there some other way to uniquely identify instructions (by just reading the .ll/.bc file) ?

7

Hi all

I have a project where LLVM annotations could be very useful. My current understanding of LLVM in general is still limited, hence a nice interface to annotations or sample code that uses annotations would help me to get started.

I was wondering what happened to the proposal for a better interface to handling of annotations, which has been discussed in this thread back in July. Has this proposal lead to any results yet?

Best regards,
  Christian

8

Hi Christian,

I was wondering what happened to the proposal for a better interface
to handling of annotations, which has been discussed in this thread
back in July. Has this proposal lead to any results yet?

So far, I've been using some code hacked up by Bart Coppens, which is very
specific and not generically useful, but enough for what we need. I'd look at
making something more generic, but this currently has no priority at all. Feel
free to code up something nice :slight_smile:

Gr.

Matthijs