[llvm-dev] Interpreting DSCallGraph results

[Christian Convey via llvm-dev]

Any suggestions for how to interpret DSCallGraph's output for the following?

I'm trying to use DSCallGraph to get a conservative estimate of a
whole-program SCC call graph.  I wanted to see how it handles real
call-graph cycles involving functions both internal and external to the
module.  So I made a test program with the following actual call graph,
using the standard library's "qsort" function:

   main          --> DoSorting
   DoSorting     --> qsort(..., QsortCallback)
   qsort         --> QsortCallback (via callback)
   QsortCallback --> DoSorting


There were two things in DSCallGraph's results which surprised me:

   - "qsort" was placed into its own SCC (according to
   DSCallGraph::scc_begin/end).
   - "qsort" 's SCC did not have any callees (according to
   DSCallGraph::flat_callee_begin/end).

I'm not sure that DSCallGraph's output is *wrong*, because it does report
these caveats:

   - DSCallGraph::called_from_incomplete_site( @QsortCallback ) returns
   true.
   - For all three callsites in the Module,
   DSCallGraph::callee_is_complete(...) returns false.

Must I assume that the entire Module could be a single SCC whenever
"called_from_incomplete_site" returns true "callee_is_complete" returns
false?

And if I *do* need to be that pessimistic, anyone know why DSCallGraph
doesn't apply that pessimism itself and return a single-SCC call graph in
such cases?

Thanks,
Christian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20151228/0fa94950/attachment.html>