[LLVMdev] Potential Google Summer of Code Applicant

Mon Apr 16 21:18:21 PDT 2012

You are probably correct about consuming the debug information at run-time,
I think where my idea might be useful is only having one set of details
about where stuff ends up after optimization passes get run. In this
respect the debug information should have enough information to tell you
where a particular value ended up at each point in the program and I
believe it could accurately tell you its type too which could help with the
propagation issues you mention. From their, a compiler specific post
processor could build the required run-time tables. If an approach like
this where feasible, it would allow the GC code to share bug fixes and test
coverage with debug information which I imagine gets a good amount of
attention.

On Mon, Apr 16, 2012 at 8:29 PM, Talin <viridia at gmail.com> wrote:

> Actually, I'm pretty happy with the way that LLVM handles this aspect of
> garbage collection now. LLVM does not itself generate any data related to
> garbage collection - all it does is supply a plugin interface that lets
> your code know where on the stack the roots are. Your code is responsible
> for generating any static data structures that would be read by your
> garbage collector. So if you wanted to generate something that is
> DWARF-like, you could certainly do so. Note that you probably want
> something that is optimized for speed, and decoding DWARF DIEs is fairly
> involved. In general I've found that it's not all that difficult to
> generate this data, and the size of the data isn't all that large,
> especially if you take advantage of the fact that a lot of functions can
> share the same static data, since they have identical stack layouts.
>
> Note that tracking stack roots is the only thing that LLVM does for you in
> the way of garbage collection - your compiler's runtime is responsible for
> managing the heap, implementing tracing, and all of the other tasks related
> to garbage collection. And this is fine, for several reasons: First,
> because different languages are going to have different strategies for
> garbage collection, and second, because the tracking of stack roots is the
> only place where the compiler really needs to be involved (well, that and
> read/write barriers if you are doing an advanced collector.) Everything
> else is just library code.
>
> So I have no complaints about the *output* of LLVM's GC support. It's the
> *input* side where I think serious effort is needed. The method for telling
> LLVM which variables are roots and which are not is clunky and has a tacked
> on feel.
>
> The way things work now is that there is an intrinsic function called
> llvm.gcroot which is a "do-nothing" function - it takes one argument (which
> must be a pointer that was created by an Alloca instruction) and simply
> returns that value unchanged. However, the presence of this function call
> is used as a "marker" by the various code generators and other backend
> passes, letting them know that this variable should be treated as a root.
> The advantage of doing it this way is that it avoids having to store a bit
> in the value itself (because bits are precious). The downside, however, is
> that you can't easily propagate the root-ness attribute from one value to
> another - so if you load that value into a register / SSA value, the
> root-ness of the value is lost. In order to get around *that* problem, the
> frontend has to do some fairly convoluted logic - which could all be
> avoided if we had a better way to track roots.
>
> In virtually all modern languages, garbage collection is driven by type -
> so if you have a pointer to an Object, you know it's garbage collectable,
> and you know how to trace it. And because types travel along with their
> values, so that copying from one variable to another preserves the type
> information, it's easy for the garbage collection information to go along
> for the ride. Furthermore, it doesn't matter whether the value is in memory
> or is in a register, the type tells you all you need to know.
>
> So it seems obvious that instead of marking specific *values* as garbage
> collectible, we should instead be marking *types*.
>
> (Of course, there are occasionally instances where a given type may be
> *conditionally* garbage collectible (such as Java's "permgen" objects which
> live in permanent memory). But that's an edge case that can easily be
> handled by a number of techniques. Telling the compiler that a given type
> is garbage collectable in no way compels the runtime library to actually
> trace it.)
>
> On Sun, Apr 15, 2012 at 8:00 PM, Nathan Jeffords <blunted2night at gmail.com>wrote:
>
>> Just a thought, but it would it make sense for garbage collection to be
>> some sort of minimal debug information for potentially optimized code.
>> Store just enough debug information to reconstruct call stacks and know
>> where gc-roots are. Perhaps an approach like this could minimize the work
>> required as it is shared between gc-support and debug information support.
>> From what I understand, DWARF exception handling is similar in that it
>> makes use of similar information to understand where things are during
>> unwinding.
>>
>>
>> On Fri, Apr 13, 2012 at 3:05 PM, Talin <viridia at gmail.com> wrote:
>>
>>> I realize that this was written in a hurry, and may not have been
>>> entirely clear. If there are any questions, critiques, etc., I'd be happy
>>> to respond to them. I'd really like it if LLVM's garbage collection support
>>> didn't continue to languish...
>>>
>>>
>>> On Fri, Apr 6, 2012 at 12:43 PM, Talin <viridia at gmail.com> wrote:
>>>
>>>> Sorting through all of the discussions would be difficult, as the ideas
>>>> have morphed over the years. Also, some of the discussion took place
>>>> offline at various LLVM dev conferences.
>>>>
>>>> I can summarize the main points here:
>>>>
>>>> The biggest improvement in GC would be to allow SSA values to be
>>>> declared as GC roots - currently only alloca values, that is, values in
>>>> memory, can be GC roots. This means that frontends need to copy traceable
>>>> values to memory before calling any function that might trigger a
>>>> collection. This makes the resulting IR complex and inefficient.
>>>>
>>>> Chris's proposal is that we add metadata to the Pointer type (perhaps
>>>> using the address space field) to indicate that a pointer is a GC root.
>>>> Note that this would not be a replacement for the existing llvm.gcroot
>>>> intrinsic, but rather an addition to it - that is, you could choose to use
>>>> either the Pointer attribute or the llvm.gcroot intrinsic. The reason for
>>>> this is that for some languages, there are non-pointer GC roots (my
>>>> language has discriminated unions that sometimes contains pointers and
>>>> sometimes not.) These frontends would need to continue to use the existing
>>>> intrinsics, but fortunately such cases are fairly rare and this is not a
>>>> terrible burden. For most languages, the Pointer attribute would be a much
>>>> easier way to deal with roots, since the "root-ness" of a value is simply a
>>>> property of it's type, and so gets carried along with the value
>>>> automatically.
>>>>
>>>> Thus, in this new scheme, Pointer types which were marked as roots
>>>> could be traced regardless of whether they lived within a memory location
>>>> or a register. These could be used to generate either stack maps or
>>>> register maps. Note that these maps are language-specific and are
>>>> implemented via a language-specific plugin, so you would need to update the
>>>> current API to deal with items in registers. (An alternative plan is to
>>>> allow the frontend to request via a flag that roots be copied to local
>>>> variables at each safe point, so that register maps would not be needed.
>>>> This alternate plan might be a good first step, and then move on to the
>>>> more difficult problem of register maps if the first step is successful.)
>>>>
>>>> The reason why this is difficult is that the presence of garbage
>>>> collection roots has a major impact on the optimizer and backend code
>>>> generators - certain optimizations can cause the stack map to be incorrect,
>>>> so these optimizations must be prevented or compensated for. However, an
>>>> advantage is that variables which are optimized away no longer need to be
>>>> included in stack maps - something that is not possible with the current
>>>> approach.
>>>>
>>>> One other limitation of the Pointer approach over the existing
>>>> llvm.gcroot system, is that the latter allows complex metadata to be
>>>> associated with each root. This is useful in languages that don't use
>>>> tagged objects, that is, the type of every object it known at compile time.
>>>> However, creating a metadata pointer for every Pointer type would be
>>>> expensive, so the Pointer roots would only be used for languages which use
>>>> tagged objects - which is fortunately most languages that use GC.
>>>>
>>>> An even more ambitious plan would be to allow structs in SSA values to
>>>> be declared as roots, which would be useful for languages like mine. We
>>>> wouldn't use register maps for these, since a struct might get "sliced"
>>>> into multiple registers. Instead, the code generator would automatically
>>>> spill the struct value to memory during a safe point, and otherwise treat
>>>> the struct like the existing llvm.gcroot() intrinsic does. Note that this
>>>> is a much harder problem, and would not be needed for languages like Java
>>>> or Python where objects are always passed by reference. I wouldn't expect
>>>> an initial implementation to attempt to tackle this harder problem.
>>>>
>>>> That would be the biggest improvement that I can think of. There are a
>>>> few other minor improvements that I would also like to see:
>>>>
>>>> One would be a set of intrinsics that would allow efficient iteration
>>>> of stack frames in an efficient manner. The existing LLVM stack frame
>>>> intrinsics are inefficient and cannot be relied upon in many cases.
>>>> Basically you'd want 3 intrinsics, which would work on all supported
>>>> platforms: The first would return an opaque reference to the current stack
>>>> frame; The second would take a stack frame as it's argument and return a
>>>> pointer to it's parent stack frame; And the third would take a stack frame
>>>> argument and return a base address for the local variables of that frame.
>>>> The lanuguage specific runtime would then use this base address, along with
>>>> the generated stack maps, to access all of the stack roots for that frame.
>>>>
>>>> An example of how this stack walking is done can be seen here:
>>>> http://code.google.com/p/tart/source/browse/trunk/runtime/lib/gc_common.cpp#155However, this code only works on x86 - the intrinsics that I envision would
>>>> work on a much wider set of backend targets.
>>>>
>>>> Note that these items are just a tiny part of a complete collector,
>>>> however, the design of LLVM is that each language is supposed to implement
>>>> its own collector, and LLVM only supplies the parts that need to be
>>>> integrated into the code generator.
>>>>
>>>> I can also suggest ways to test the new features without having to
>>>> build a complete garbage collector. For example, one can create a trivial
>>>> stack walker that merely counts the number of non-null root pointers, and
>>>> write various unit tests that verify that the results are as expected.
>>>>
>>>> On Fri, Apr 6, 2012 at 6:33 AM, Michael Thorpe <mthorpe at netcraft.com>wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I'm currently working for the next 6 months, but I would be very
>>>>> interested in looking into this. Are there any discussions in this mailing
>>>>> list that would be useful in finding out more information?
>>>>>
>>>>> Regards
>>>>>
>>>>> Michael Thorpe
>>>>> Internet Services Developer
>>>>> Netcraft Ltd
>>>>>
>>>>>
>>>>> -----Original Message-----
>>>>> From: llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu]
>>>>> On Behalf Of Yiannis Tsiouris
>>>>> Sent: 06 April 2012 09:25
>>>>> To: llvmdev at cs.uiuc.edu
>>>>> Subject: Re: [LLVMdev] Potential Google Summer of Code Applicant
>>>>>
>>>>> On 4/6/12 2:21 AM, Talin wrote:
>>>>> > I would really like to see someone work on LLVM's garbage collection
>>>>> > support - it hasn't been updated in 4 years, and while there's been a
>>>>> > lot of talk about ways that it could be improved, there's been no
>>>>> action.
>>>>> That is *sooo* true! :-) I'm one of the authors of an LLVM backend for
>>>>> Erlang (ErLLVM [1]); we have tested and measured our backend  and noticed
>>>>> that with the current GC infrastructure we see 20-40% performance
>>>>> degradation (because of the loads/stores on the stack for all gcroots). It
>>>>> is clear to me and the rest of the team that with this infrastructure the
>>>>> LLVM might not be suitable for languages with explicit garbage collection,
>>>>> like Erlang. I've also studied the way the Vmkit project handles GC and
>>>>> they seem to face the same deficiency too.
>>>>>
>>>>> offtopic: I am working on an email (more like an RFC) with all the
>>>>> details and patches to the LLVM project in order to support our Erlang
>>>>> backend. I hope I will be able to send it by next week. Note, that we have
>>>>> already talked with the Ericsson/OTP team about integrating our work in a
>>>>> future release of Erlang/OTP (as a new HiPE backend).
>>>>>
>>>>> Yiannis
>>>>>
>>>>> [1]: http://erllvm.softlab.ntua.gr
>>>>>
>>>>> --
>>>>> Yiannis Tsiouris
>>>>> Ph.D. student,
>>>>> Software Engineering Laboratory,
>>>>> National Technical University of Athens
>>>>> WWW: http://www.softlab.ntua.gr/~gtsiour
>>>>>
>>>>> _______________________________________________
>>>>> LLVM Developers mailing list
>>>>> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
>>>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>>>>>
>>>>> _______________________________________________
>>>>> LLVM Developers mailing list
>>>>> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
>>>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> -- Talin
>>>>
>>>
>>>
>>>
>>> --
>>> -- Talin
>>>
>>> _______________________________________________
>>> LLVM Developers mailing list
>>> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>>>
>>>
>>
>
>
> --
> -- Talin
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20120416/51fac506/attachment.html>