> > why disallow caller->indirect_calls? See testcase in comment #9 > > > + return false; > > + for (cgraph_edge *e2 = callee->callees; e2; e2 = e2->next_callee) > > I don't think this flys - it looks quadratic. Can we compute this > in the inline summary once instead?
I guess I can place a cache there. I think this check will become more global over time so it more fits IMO here. > > As for indirect calls, can we maybe mark initial direct GIMPLE call > stmts as "always-inline" and only look at that marking, thus an > indirect call will never become "always-inline"? Iff cgraph edges > prevail during all early inlining we could mark call edges for > this purpose? I also think we need call site specific info. Tagging gimple call statements and copying the info to gimple edges will probably be needed here. We want to keep the info from early inlining to late inlining since we output errors late. We already have plenty of GF_CALL_ flags, so adding one should be easy? Honza
