In read_call_site_scope we have:
...
call_site_local.pc = pc;
slot = htab_find_slot (cu->call_site_htab, &call_site_local, INSERT);
...
The call passes a call_site pointer as element. OTOH, the hashtab is created
using hash_f == core_addr_hash and eq_f == core_addr_eq, so the element
will be accessed through a CORE_ADDR pointer.
This is not wrong (at least in C), given that pc is the first field in
call_site.
Nevertheless, as in call_site_for_pc, make the htab_find_slot call match the
used hash_f and eq_f by using &pc instead:
...
slot = htab_find_slot (cu->call_site_htab, &pc, INSERT);
...
Tested on x86_64-linux.
Co-Authored-By: Tom de Vries <tdevries@suse.de>
struct gdbarch *gdbarch = objfile->arch ();
CORE_ADDR pc, baseaddr;
struct attribute *attr;
- struct call_site *call_site, call_site_local;
+ struct call_site *call_site;
void **slot;
int nparams;
struct die_info *child_die;
cu->call_site_htab = htab_create_alloc_ex (16, core_addr_hash, core_addr_eq,
NULL, &objfile->objfile_obstack,
hashtab_obstack_allocate, NULL);
- call_site_local.pc = pc;
- slot = htab_find_slot (cu->call_site_htab, &call_site_local, INSERT);
+ slot = htab_find_slot (cu->call_site_htab, &pc, INSERT);
if (*slot != NULL)
{
complaint (_("Duplicate PC %s for DW_TAG_call_site "
struct call_site
{
- /* * Address of the first instruction after this call. It must be
- the first field as we overload core_addr_hash and core_addr_eq
- for it. */
+ /* Address of the first instruction after this call. */
CORE_ADDR pc;