[llvm-dev] An issue with new PM's requirements on call graph changes

[Philip Reames via llvm-dev]

On 06/30/2017 01:02 AM, Chandler Carruth via llvm-dev wrote:
> I have hit a fairly isolated practical issue deploying the new PM, but 
> it does point to a latent theoretical issues as well. I see various 
> ways to address it, but want feedback from others before moving forward.
>
> The issue is that we can introduce out-of-thin-air calls to known 
> library functions (`SimplifyLibCalls`, etc). These can be introduced 
> in function passes (`InstCombine` in particular) and that seems highly 
> desirable.
>
> These all look like one of these cases:
> 1a) Introducing a new call to an LLVM intrinsic
> 1b) Replacing an existing call with a call to an LLVM intrinsic
> 2a) Introducing a new call to a declared library function (but not 
> defined)
> 2b) Replacing an existing call with a call to a declared library function
> 3a) Introducing a new call to a defined library function
> 3b) Replacing an existing call with a call to a defined library function
>
> Both #1 and #2 are easy to handle in reality. Intrinsics and declared 
> functions don't impact the PM's call graph because there is no need to 
> order the walk over them. But #3 is a real issue.
>
> The only case I have found that actually hits #3 at all hits #3b when 
> building FORTIFY code with the new pass manager because after inlining 
> we do a lot of (really nice) optimizations on library calls to remove 
> unnecessary FORTIFY checks. But this is in *theory* a problem when 
> LTO-ing with libc. More likely it could be a problem when LTO-ing with 
> a vector math library.
>
> So what do we do?
>
> My initial idea: find all *defined* library functions in the module, 
> and every time we create a ref edge to one of them, synthesize a ref 
> edge to all of them. This should completely solve #3b above. But it 
> doesn't really address #3a at all.
>
> Is that OK? It would be very convenient to say that if we want to 
> introduce truly novel and new calls to library functions, we should 
> have an LLVM intrinsic to model those routines.
>
> But we actually have an example (I think) of #3a, introducing a call 
> to a library function out of the blue: memset_pattern. =/
Out of curiosity, is this the only example we have?  In the context of 
https://reviews.llvm.org/D34885, I was thinking about whether it might 
make sense to have intrinsic form of memset_pattern anyways. If we did 
that, could we disallow such cases in practice?  On the other hand, it 
does seem less than desirable to prevent inlining of such cases when we 
do in fact have the implementation linked in.
>
> The only way I see to reasonably handle #3a is to have *every* 
> function implicitly contain a reference edge to every defined library 
> function in the module. This is, needless to say, amazingly wasteful. 
> Hence my email. How important is this?
>
> If we need to correctly handle this, I think I would probably 
> implement this by actually changing the *iteration* of reference edges 
> in the graph to just implicitly walk the list of defined library 
> functions so that we didn't burn any space on this. But it will make 
> iteration of reference edges slower and add a reasonable amount of 
> complexity. So I'd like to hear some other opinions before going down 
> either of these roads.
>
>
> Thanks,
> -Chandler
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170701/61892485/attachment.html>