https://sourceware.org/bugzilla/show_bug.cgi?id=30060
Alan Modra <amodra at gmail dot com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC|amodra at gmail dot com | Assignee|unassigned at sourceware dot org |amodra at gmail dot com Status|NEW |ASSIGNED --- Comment #1 from Alan Modra <amodra at gmail dot com> --- Things go wrong here at ar.c:1302 tmpfd = dup (tmpfd); if (!bfd_close (obfd)) bfd_fatal (old_name); The bfd_close returns an error, and when bfd_fatal calls exit the atexit handlers include bucomm.c:remove_output. remove_output calls bfd_cache_close, which accesses the output bfd. I reckon that remove_output should not be calling bfd_cache_close, but it might also be the case that code that does this sort of thing is common enough that we can't reclaim bfd memory in bfd_close or bfd_close_all_done after an error. -- You are receiving this mail because: You are on the CC list for the bug.