https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64694
--- Comment #7 from H.J. Lu <hjl.tools at gmail dot com> --- (In reply to Jan Hubicka from comment #6) > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64694 > > > > --- Comment #5 from Jan Hubicka <hubicka at ucw dot cz> --- > > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64694 > > > > > > --- Comment #4 from H.J. Lu <hjl.tools at gmail dot com> --- > > > r219076 introduced: > > > > > > if (current_badness != badness) > > > { > > > if (edge_heap.min () && badness > edge_heap.min_key ()) > > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > > > Does it look right? > > > > This is a thinko, indeed. I am testing a fix, but I doubdt this can affect > > correctness This patch --- diff --git a/gcc/ipa-inline.c b/gcc/ipa-inline.c index da1afc6..aa93928 100644 --- a/gcc/ipa-inline.c +++ b/gcc/ipa-inline.c @@ -1794,7 +1794,7 @@ inline_small_functions (void) #endif if (current_badness != badness) { - if (edge_heap.min () && badness > edge_heap.min_key ()) + if (edge_heap.min () && current_badness > edge_heap.min_key ()) { edge->aux = edge_heap.insert (current_badness, edge); continue; --- works for me. > > of a testcase. I see it does inlining, but why it depends on inline > > decisions? > I meant unwinding. > That is a very interesting question.
