> > The debug info on your testcase looks as expected (though, > we really don't emit the DW_OP_GNU_parameter_ref for aggregates anyway > - not sure if we could do something about those at least in the easy cases > where the argument used to be stored into the struct shortly before the > call.
Cool, I was basically curious if we can chain the changes w/o suprises. > > Anyway, here is an updated patch where I'm not adding anything for > skipped parameters that have tree_map replacement, and fixes some issues > found during bootstrap/regtest (x86_64-linux and i686-linux). > > Ok for trunk? > > 2015-12-18 Jakub Jelinek <ja...@redhat.com> > > PR debug/68860 > * ipa-split.c (split_function): Only perform caller side > modifications for decl_debug_args here. > * cgraph.c: Include gimplify.h. > (cgraph_edge::redirect_call_stmt_to_callee): Add caller side > debug stmts for decl_debug_args. Spelling fix in a comment. > * tree-inline.c (tree_function_versioning): Populate decl_debug_args > for args_to_skip arguments and add callee side debug stmts. > Formatting fixes. Avoid shadowing i variable. > > * gcc.dg/guality/pr68860-1.c: New test. > * gcc.dg/guality/pr68860-2.c: New test. Looks good to me. Thanks again for working on this! Honza