> On Tue, Mar 5, 2024 at 1:45 PM H.J. Lu <hjl.to...@gmail.com> wrote: > > > > We can't instrument an IFUNC resolver nor its callees as it may require > > TLS which hasn't been set up yet when the dynamic linker is resolving > > IFUNC symbols. > > > > Add an IFUNC resolver caller marker to cgraph_node and set it if the > > function is called by an IFUNC resolver. Update tree_profiling to skip > > functions called by IFUNC resolver. > > > > Tested with profiledbootstrap on Fedora 39/x86-64. > > > > gcc/ChangeLog: > > > > PR tree-optimization/114115 > > * cgraph.h (symtab_node): Add check_ifunc_callee_symtab_nodes. > > (cgraph_node): Add called_by_ifunc_resolver. > > * cgraphunit.cc (symbol_table::compile): Call > > symtab_node::check_ifunc_callee_symtab_nodes. > > * symtab.cc (check_ifunc_resolver): New. > > (ifunc_ref_map): Likewise. > > (is_caller_ifunc_resolver): Likewise. > > (symtab_node::check_ifunc_callee_symtab_nodes): Likewise. > > * tree-profile.cc (tree_profiling): Do not instrument an IFUNC > > resolver nor its callees. > > > > gcc/testsuite/ChangeLog: > > > > PR tree-optimization/114115 > > * gcc.dg/pr114115.c: New test. > > PING.
I am bit worried about commonly used functions getting "infected" by being called once from ifunc resolver. I think we only use thread local storage for indirect call profiling, so we may just disable indirect call profiling for these functions. Also the patch will be noop with -flto -flto-partition=max, so probably we need to compute this flag at WPA time and stream to partitions. Honza
