https://bugzilla.suse.com/show_bug.cgi?id=1176095 https://bugzilla.suse.com/show_bug.cgi?id=1176095#c4 --- Comment #4 from Tom de Vries <tdevries@suse.com> --- The assert happens as follows. We dereference tag here, and tag is 0: ... Program received signal SIGSEGV, Segmentation fault. 0x00007ffff7f65d92 in tag__is_pointer (tag=0x0) at /home/vries/dwarves/src/dwarves.h:446 446 return tag->tag == DW_TAG_pointer_type; ... because member_type is 0 here: ... (gdb) up #2 0x00007ffff7f69ad4 in type__natural_alignment (type=0x15b1e00, cu=0xd7ed50) at /home/vries/dwarves/src/dwarves.c:1469 1469 size_t member_natural_alignment = tag__natural_alignment(member_type, cu); (gdb) p member_type $1 = (struct tag *) 0x0 ... The member_type is set just before: ... 1468 struct tag *member_type = tag__strip_typedefs_and_modifiers(&member->tag, cu); 1469 size_t member_natural_alignment = tag__natural_alignment(member_type, cu); ... But it's set to 0 because member->tag.type is 0: ... (gdb) p *member $16 = {tag = {node = {next = 0x15b1ff0, prev = 0x15b1f30}, type = 0, tag = 13, visited = false, top_level = false, recursivity_level = 0, priv = 0x15b1440}, name = 55224, bit_offset = 128, bit_size = 0, byte_offset = 16, byte_size = 0, bitfield_offset = 0 '\000', bitfield_size = 0 '\000', bit_hole = 0 '\000', bitfield_end = 0 '\000', const_value = 0, alignment = 8, visited = 0 '\000', is_static = 0 '\000', accessibility = 0 '\000', virtuality = 0 '\000', hole = 0} ... Looking in a bit more detail, we can see that the member is at offset 0x426ba, and the type at 0x726e: ... (gdb) p /x *(struct dwarf_tag*)member.tag.priv $5 = {hash_node = {next = 0x15aad50, pprev = 0x15c4340}, type = {from_types = 0x0, off = 0x726e}, id = 0x426ba, {abstract_origin = { from_types = 0x0, off = 0x0}, containing_type = {from_types = 0x0, off = 0x0}}, tag = 0x15b1f90, small_id = 0x0, decl_file = 0x0, decl_line = 0x0} ... We have a matching complaint earlier: ... namespace__recode_dwarf_types: couldn't find 0x726e type for 0x426ba (member)! ... -- You are receiving this mail because: You are on the CC list for the bug.