Add dwarf2_per_objfile parameter to cutu_reader's constructors
authorSimon Marchi <simon.marchi@polymtl.ca>
Wed, 27 May 2020 15:13:55 +0000 (11:13 -0400)
committerSimon Marchi <simon.marchi@efficios.com>
Wed, 27 May 2020 15:15:55 +0000 (11:15 -0400)
The cutu_reader type is used for reading the CU represented by the
passed dwarf2_per_cu_data object.  This reading is done in the context
of a given obfile, which is currently the one associated to the passed
dwarf2_per_cu_data object.  Since the dwarf2_per_cu_data type will
become objfile-independent, we will need to pass the objfile separately.

This patch therefore adds a dwarf2_per_objfile parameter to the
cutu_reader constructors, as well as to their callers, up until the
point where we can get the dwarf2_per_objfile object from somewhere
else.  In the end, this allows removing the reference to
dwarf2_per_cu_data::dwarf2_per_objfile in cutu_reader::cutu_reader.

A few dwarf2_per_cu_data::dwarf2_per_objfile references are added (e.g.
in dwarf2_fetch_die_type_sect_off).  This is temporary, this will be
removed once these functions will get re-worked in subsequent patches.

gdb/ChangeLog:

* dwarf2/read.c (class cutu_reader) <cutu_reader>: Add
per_objfile parameter.
(load_full_type_unit): Add per_objfile parameter.
(read_signatured_type): Likewise.
(load_full_comp_unit): Likewise.
(load_cu): Likewise.
(dw2_do_instantiate_symtab): Likewise.
(dw2_get_file_names): Likewise.
(dw2_map_symtabs_matching_filename): Update.
(dw_expand_symtabs_matching_file_matcher): Update.
(dw2_map_symbol_filenames): Update.
(process_psymtab_comp_unit): Add per_objfile parameter.
(build_type_psymtabs_1): Update.
(process_skeletonless_type_unit): Update.
(dwarf2_build_psymtabs_hard): Update.
(load_partial_comp_unit): Add per_objfile parameter.
(scan_partial_symbols): Update.
(load_full_comp_unit): Add per_objfile parameter.
(process_imported_unit_die): Update.
(create_cus_hash_table): Update.
(find_partial_die): Update.
(dwarf2_read_addr_index): Update.
(follow_die_offset): Update.
(dwarf2_fetch_die_loc_sect_off): Update.
(dwarf2_fetch_constant_bytes): Update.
(dwarf2_fetch_die_type_sect_off): Update.
(follow_die_sig_1): Update.
(load_full_type_unit): Add per_objfile parameter.
(read_signatured_type): Likewise.

Change-Id: Ibd7bbc443df8b9b8b6f96ff18e93a60ee721b85f

gdb/ChangeLog
gdb/dwarf2/read.c

index 3ba5df428a80721590612d5123eec464c0d08ad4..d6c28a5a84ba4d355ba34683e4e02e1b4aa9e694 100644 (file)
@@ -1,3 +1,35 @@
+2020-05-27  Simon Marchi  <simon.marchi@polymtl.ca>
+
+       * dwarf2/read.c (class cutu_reader) <cutu_reader>: Add
+       per_objfile parameter.
+       (load_full_type_unit): Add per_objfile parameter.
+       (read_signatured_type): Likewise.
+       (load_full_comp_unit): Likewise.
+       (load_cu): Likewise.
+       (dw2_do_instantiate_symtab): Likewise.
+       (dw2_get_file_names): Likewise.
+       (dw2_map_symtabs_matching_filename): Update.
+       (dw_expand_symtabs_matching_file_matcher): Update.
+       (dw2_map_symbol_filenames): Update.
+       (process_psymtab_comp_unit): Add per_objfile parameter.
+       (build_type_psymtabs_1): Update.
+       (process_skeletonless_type_unit): Update.
+       (dwarf2_build_psymtabs_hard): Update.
+       (load_partial_comp_unit): Add per_objfile parameter.
+       (scan_partial_symbols): Update.
+       (load_full_comp_unit): Add per_objfile parameter.
+       (process_imported_unit_die): Update.
+       (create_cus_hash_table): Update.
+       (find_partial_die): Update.
+       (dwarf2_read_addr_index): Update.
+       (follow_die_offset): Update.
+       (dwarf2_fetch_die_loc_sect_off): Update.
+       (dwarf2_fetch_constant_bytes): Update.
+       (dwarf2_fetch_die_type_sect_off): Update.
+       (follow_die_sig_1): Update.
+       (load_full_type_unit): Add per_objfile parameter.
+       (read_signatured_type): Likewise.
+
 2020-05-27  Simon Marchi  <simon.marchi@efficios.com>
 
        * dwarf2/read.c (lookup_dwo_unit): Use bfd_get_filename instead
index 2ee98cc1b8cebd341ed1dae34e925f0975db3d33..ec32804bd1d654f446769da602382a2334ab9d58 100644 (file)
@@ -912,12 +912,14 @@ class cutu_reader : public die_reader_specs
 {
 public:
 
-  cutu_reader (struct dwarf2_per_cu_data *this_cu,
+  cutu_reader (dwarf2_per_cu_data *this_cu,
+              dwarf2_per_objfile *per_objfile,
               struct abbrev_table *abbrev_table,
               int use_existing_cu,
               bool skip_partial);
 
   explicit cutu_reader (struct dwarf2_per_cu_data *this_cu,
+                       dwarf2_per_objfile *per_objfile,
                        struct dwarf2_cu *parent_cu = nullptr,
                        struct dwo_file *dwo_file = nullptr);
 
@@ -1510,9 +1512,11 @@ static struct type *get_DW_AT_signature_type (struct die_info *,
                                              const struct attribute *,
                                              struct dwarf2_cu *);
 
-static void load_full_type_unit (struct dwarf2_per_cu_data *per_cu);
+static void load_full_type_unit (dwarf2_per_cu_data *per_cu,
+                                dwarf2_per_objfile *per_objfile);
 
-static void read_signatured_type (struct signatured_type *);
+static void read_signatured_type (signatured_type *sig_type,
+                                 dwarf2_per_objfile *per_objfile);
 
 static int attr_to_dynamic_prop (const struct attribute *attr,
                                 struct die_info *die, struct dwarf2_cu *cu,
@@ -1562,8 +1566,10 @@ static void create_all_comp_units (struct dwarf2_per_objfile *dwarf2_per_objfile
 
 static int create_all_type_units (struct dwarf2_per_objfile *dwarf2_per_objfile);
 
-static void load_full_comp_unit (struct dwarf2_per_cu_data *, bool,
-                                enum language);
+static void load_full_comp_unit (dwarf2_per_cu_data *per_cu,
+                                dwarf2_per_objfile *per_objfile,
+                                bool skip_partial,
+                                enum language pretend_language);
 
 static void process_full_comp_unit (struct dwarf2_per_cu_data *,
                                    enum language);
@@ -2323,17 +2329,18 @@ create_quick_file_names_table (unsigned int nr_initial_entries)
                                     delete_file_name_entry, xcalloc, xfree));
 }
 
-/* Read in PER_CU->CU.  This function is unrelated to symtabs, symtab would
-   have to be created afterwards.  You should call age_cached_comp_units after
-   processing PER_CU->CU.  dw2_setup must have been already called.  */
+/* Read in CU (dwarf2_cu object) for PER_CU in the context of PER_OBJFILE.  This
+   function is unrelated to symtabs, symtab would have to be created afterwards.
+   You should call age_cached_comp_units after processing the CU.  */
 
 static void
-load_cu (struct dwarf2_per_cu_data *per_cu, bool skip_partial)
+load_cu (dwarf2_per_cu_data *per_cu, dwarf2_per_objfile *per_objfile,
+        bool skip_partial)
 {
   if (per_cu->is_debug_types)
-    load_full_type_unit (per_cu);
+    load_full_type_unit (per_cu, per_objfile);
   else
-    load_full_comp_unit (per_cu, skip_partial, language_minimal);
+    load_full_comp_unit (per_cu, per_objfile, skip_partial, language_minimal);
 
   if (per_cu->cu == NULL)
     return;  /* Dummy CU.  */
@@ -2361,7 +2368,7 @@ dw2_do_instantiate_symtab (dwarf2_per_cu_data *per_cu,
   if (!dwarf2_per_objfile->symtab_set_p (per_cu))
     {
       queue_comp_unit (per_cu, language_minimal);
-      load_cu (per_cu, skip_partial);
+      load_cu (per_cu, dwarf2_per_objfile, skip_partial);
 
       /* If we just loaded a CU from a DWO, and we're working with an index
         that may badly handle TUs, load all the TUs in that DWO as well.
@@ -3212,7 +3219,8 @@ dw2_get_file_names_reader (const struct die_reader_specs *reader,
    table for THIS_CU.  */
 
 static struct quick_file_names *
-dw2_get_file_names (struct dwarf2_per_cu_data *this_cu)
+dw2_get_file_names (dwarf2_per_cu_data *this_cu,
+                   dwarf2_per_objfile *per_objfile)
 {
   /* This should never be called for TUs.  */
   gdb_assert (! this_cu->is_debug_types);
@@ -3225,7 +3233,7 @@ dw2_get_file_names (struct dwarf2_per_cu_data *this_cu)
   if (this_cu->v.quick->no_file_data)
     return NULL;
 
-  cutu_reader reader (this_cu);
+  cutu_reader reader (this_cu, per_objfile);
   if (!reader.dummy_p)
     dw2_get_file_names_reader (&reader, reader.info_ptr, reader.comp_unit_die);
 
@@ -3341,7 +3349,8 @@ dw2_map_symtabs_matching_filename
       if (dwarf2_per_objfile->symtab_set_p (per_cu))
        continue;
 
-      quick_file_names *file_data = dw2_get_file_names (per_cu);
+      quick_file_names *file_data
+       = dw2_get_file_names (per_cu, dwarf2_per_objfile);
       if (file_data == NULL)
        continue;
 
@@ -3682,7 +3691,8 @@ dw2_expand_symtabs_with_fullname (struct objfile *objfile,
       if (dwarf2_per_objfile->symtab_set_p (per_cu))
        continue;
 
-      quick_file_names *file_data = dw2_get_file_names (per_cu);
+      quick_file_names *file_data
+       = dw2_get_file_names (per_cu, dwarf2_per_objfile);
       if (file_data == NULL)
        continue;
 
@@ -4671,7 +4681,8 @@ dw_expand_symtabs_matching_file_matcher
       if (dwarf2_per_objfile->symtab_set_p (per_cu))
        continue;
 
-      quick_file_names *file_data = dw2_get_file_names (per_cu);
+      quick_file_names *file_data
+       = dw2_get_file_names (per_cu, dwarf2_per_objfile);
       if (file_data == NULL)
        continue;
 
@@ -4855,7 +4866,8 @@ dw2_map_symbol_filenames (struct objfile *objfile, symbol_filename_ftype *fun,
          if (dwarf2_per_objfile->symtab_set_p (per_cu))
            continue;
 
-         quick_file_names *file_data = dw2_get_file_names (per_cu);
+         quick_file_names *file_data
+           = dw2_get_file_names (per_cu, dwarf2_per_objfile);
          if (file_data == NULL)
            continue;
 
@@ -6952,14 +6964,14 @@ cutu_reader::init_tu_and_read_dwo_dies (dwarf2_per_cu_data *this_cu,
    If USE_EXISTING_CU is non-zero, and THIS_CU->cu is non-NULL, then use it.
    Otherwise, a new CU is allocated with xmalloc.  */
 
-cutu_reader::cutu_reader (struct dwarf2_per_cu_data *this_cu,
+cutu_reader::cutu_reader (dwarf2_per_cu_data *this_cu,
+                         dwarf2_per_objfile *dwarf2_per_objfile,
                          struct abbrev_table *abbrev_table,
                          int use_existing_cu,
                          bool skip_partial)
   : die_reader_specs {},
     m_this_cu (this_cu)
 {
-  struct dwarf2_per_objfile *dwarf2_per_objfile = this_cu->dwarf2_per_objfile;
   struct objfile *objfile = dwarf2_per_objfile->objfile;
   struct dwarf2_section_info *section = this_cu->section;
   bfd *abfd = section->get_bfd_owner ();
@@ -7176,13 +7188,13 @@ cutu_reader::keep ()
    When parent_cu is passed, it is used to provide a default value for
    str_offsets_base and addr_base from the parent.  */
 
-cutu_reader::cutu_reader (struct dwarf2_per_cu_data *this_cu,
+cutu_reader::cutu_reader (dwarf2_per_cu_data *this_cu,
+                         dwarf2_per_objfile *dwarf2_per_objfile,
                          struct dwarf2_cu *parent_cu,
                          struct dwo_file *dwo_file)
   : die_reader_specs {},
     m_this_cu (this_cu)
 {
-  struct dwarf2_per_objfile *dwarf2_per_objfile = this_cu->dwarf2_per_objfile;
   struct objfile *objfile = dwarf2_per_objfile->objfile;
   struct dwarf2_section_info *section = this_cu->section;
   bfd *abfd = section->get_bfd_owner ();
@@ -7544,7 +7556,8 @@ process_psymtab_comp_unit_reader (const struct die_reader_specs *reader,
    Process compilation unit THIS_CU for a psymtab.  */
 
 static void
-process_psymtab_comp_unit (struct dwarf2_per_cu_data *this_cu,
+process_psymtab_comp_unit (dwarf2_per_cu_data *this_cu,
+                          dwarf2_per_objfile *per_objfile,
                           bool want_partial_unit,
                           enum language pretend_language)
 {
@@ -7556,7 +7569,7 @@ process_psymtab_comp_unit (struct dwarf2_per_cu_data *this_cu,
   if (this_cu->cu != NULL)
     free_one_cached_comp_unit (this_cu);
 
-  cutu_reader reader (this_cu, NULL, 0, false);
+  cutu_reader reader (this_cu, per_objfile, NULL, 0, false);
 
   switch (reader.comp_unit_die->tag)
     {
@@ -7737,8 +7750,8 @@ build_type_psymtabs_1 (struct dwarf2_per_objfile *dwarf2_per_objfile)
          ++tu_stats->nr_uniq_abbrev_tables;
        }
 
-      cutu_reader reader (&tu.sig_type->per_cu, abbrev_table.get (),
-                         0, false);
+      cutu_reader reader (&tu.sig_type->per_cu, dwarf2_per_objfile,
+                         abbrev_table.get (), 0, false);
       if (!reader.dummy_p)
        build_type_psymtabs_reader (&reader, reader.info_ptr,
                                    reader.comp_unit_die);
@@ -7843,7 +7856,7 @@ process_skeletonless_type_unit (void **slot, void *info)
   *slot = entry;
 
   /* This does the job that build_type_psymtabs_1 would have done.  */
-  cutu_reader reader (&entry->per_cu, NULL, 0, false);
+  cutu_reader reader (&entry->per_cu, dwarf2_per_objfile, NULL, 0, false);
   if (!reader.dummy_p)
     build_type_psymtabs_reader (&reader, reader.info_ptr,
                                reader.comp_unit_die);
@@ -7944,7 +7957,8 @@ dwarf2_build_psymtabs_hard (struct dwarf2_per_objfile *dwarf2_per_objfile)
       if (per_cu->v.psymtab != NULL)
        /* In case a forward DW_TAG_imported_unit has read the CU already.  */
        continue;
-      process_psymtab_comp_unit (per_cu, false, language_minimal);
+      process_psymtab_comp_unit (per_cu, dwarf2_per_objfile, false,
+                                language_minimal);
     }
 
   /* This has to wait until we read the CUs, we need the list of DWOs.  */
@@ -7977,9 +7991,10 @@ dwarf2_build_psymtabs_hard (struct dwarf2_per_objfile *dwarf2_per_objfile)
    This is also used when rereading a primary CU with load_all_dies.  */
 
 static void
-load_partial_comp_unit (struct dwarf2_per_cu_data *this_cu)
+load_partial_comp_unit (dwarf2_per_cu_data *this_cu,
+                       dwarf2_per_objfile *per_objfile)
 {
-  cutu_reader reader (this_cu, NULL, 1, false);
+  cutu_reader reader (this_cu, per_objfile, NULL, 1, false);
 
   if (!reader.dummy_p)
     {
@@ -8158,7 +8173,8 @@ scan_partial_symbols (struct partial_die_info *first_die, CORE_ADDR *lowpc,
 
                /* Go read the partial unit, if needed.  */
                if (per_cu->v.psymtab == NULL)
-                 process_psymtab_comp_unit (per_cu, true, cu->language);
+                 process_psymtab_comp_unit (per_cu, cu->per_objfile, true,
+                                            cu->language);
 
                cu->per_cu->imported_symtabs_push (per_cu);
              }
@@ -9089,13 +9105,14 @@ die_eq (const void *item_lhs, const void *item_rhs)
 /* Load the DIEs associated with PER_CU into memory.  */
 
 static void
-load_full_comp_unit (struct dwarf2_per_cu_data *this_cu,
+load_full_comp_unit (dwarf2_per_cu_data *this_cu,
+                    dwarf2_per_objfile *per_objfile,
                     bool skip_partial,
                     enum language pretend_language)
 {
   gdb_assert (! this_cu->is_debug_types);
 
-  cutu_reader reader (this_cu, NULL, 1, skip_partial);
+  cutu_reader reader (this_cu, per_objfile, NULL, 1, skip_partial);
   if (reader.dummy_p)
     return;
 
@@ -9899,8 +9916,9 @@ process_imported_unit_die (struct die_info *die, struct dwarf2_cu *cu)
     {
       sect_offset sect_off = attr->get_ref_die_offset ();
       bool is_dwz = (attr->form == DW_FORM_GNU_ref_alt || cu->per_cu->is_dwz);
+      dwarf2_per_objfile *per_objfile = cu->per_objfile;
       dwarf2_per_cu_data *per_cu
-       = dwarf2_find_containing_comp_unit (sect_off, is_dwz, cu->per_objfile);
+       = dwarf2_find_containing_comp_unit (sect_off, is_dwz, per_objfile);
 
       /* We're importing a C++ compilation unit with tag DW_TAG_compile_unit
         into another compilation unit, at root level.  Regard this as a hint,
@@ -9912,7 +9930,7 @@ process_imported_unit_die (struct die_info *die, struct dwarf2_cu *cu)
 
       /* If necessary, add it to the queue and load its DIEs.  */
       if (maybe_queue_comp_unit (cu, per_cu, cu->language))
-       load_full_comp_unit (per_cu, false, cu->language);
+       load_full_comp_unit (per_cu, per_objfile, false, cu->language);
 
       cu->per_cu->imported_symtabs_push (per_cu);
     }
@@ -11313,7 +11331,7 @@ create_cus_hash_table (struct dwarf2_per_objfile *dwarf2_per_objfile,
       per_cu.sect_off = sect_offset (info_ptr - section.buffer);
       per_cu.section = &section;
 
-      cutu_reader reader (&per_cu, cu, &dwo_file);
+      cutu_reader reader (&per_cu, dwarf2_per_objfile, cu, &dwo_file);
       if (!reader.dummy_p)
        create_dwo_cu_reader (&reader, reader.info_ptr, reader.comp_unit_die,
                              &dwo_file, &read_unit);
@@ -12825,7 +12843,7 @@ queue_and_load_dwo_tu (void **slot, void *info)
         a real dependency of PER_CU on SIG_TYPE.  That is detected later
         while processing PER_CU.  */
       if (maybe_queue_comp_unit (NULL, sig_cu, per_cu->cu->language))
-       load_full_type_unit (sig_cu);
+       load_full_type_unit (sig_cu, per_cu->cu->per_objfile);
       per_cu->imported_symtabs_push (sig_cu);
     }
 
@@ -18665,7 +18683,7 @@ find_partial_die (sect_offset sect_off, int offset_in_dwz, struct dwarf2_cu *cu)
                                                 dwarf2_per_objfile);
 
       if (per_cu->cu == NULL || per_cu->cu->partial_dies == NULL)
-       load_partial_comp_unit (per_cu);
+       load_partial_comp_unit (per_cu, cu->per_objfile);
 
       per_cu->cu->last_used = 0;
       pd = per_cu->cu->find_partial_die (sect_off);
@@ -18684,7 +18702,7 @@ find_partial_die (sect_offset sect_off, int offset_in_dwz, struct dwarf2_cu *cu)
         DIEs alone (which can still be in use, e.g. in scan_partial_symbols),
         and clobber THIS_CU->cu->partial_dies with the hash table for the new
         set.  */
-      load_partial_comp_unit (per_cu);
+      load_partial_comp_unit (per_cu, cu->per_objfile);
 
       pd = per_cu->cu->find_partial_die (sect_off);
     }
@@ -19400,7 +19418,7 @@ dwarf2_read_addr_index (dwarf2_per_cu_data *per_cu, unsigned int addr_index)
     }
   else
     {
-      cutu_reader reader (per_cu, NULL, 0, false);
+      cutu_reader reader (per_cu, dwarf2_per_objfile, NULL, 0, false);
       addr_base = reader.cu->addr_base;
       addr_size = reader.cu->header.addr_size;
     }
@@ -22211,7 +22229,7 @@ follow_die_offset (sect_offset sect_off, int offset_in_dwz,
 
       /* If necessary, add it to the queue and load its DIEs.  */
       if (maybe_queue_comp_unit (cu, per_cu, cu->language))
-       load_full_comp_unit (per_cu, false, cu->language);
+       load_full_comp_unit (per_cu, dwarf2_per_objfile, false, cu->language);
 
       target_cu = per_cu->cu;
     }
@@ -22219,7 +22237,8 @@ follow_die_offset (sect_offset sect_off, int offset_in_dwz,
     {
       /* We're loading full DIEs during partial symbol reading.  */
       gdb_assert (dwarf2_per_objfile->per_bfd->reading_partial_symbols);
-      load_full_comp_unit (cu->per_cu, false, language_minimal);
+      load_full_comp_unit (cu->per_cu, dwarf2_per_objfile, false,
+                          language_minimal);
     }
 
   *ref_cu = target_cu;
@@ -22274,7 +22293,7 @@ dwarf2_fetch_die_loc_sect_off (sect_offset sect_off,
   struct objfile *objfile = dwarf2_per_objfile->objfile;
 
   if (per_cu->cu == NULL)
-    load_cu (per_cu, false);
+    load_cu (per_cu, dwarf2_per_objfile, false);
   cu = per_cu->cu;
   if (cu == NULL)
     {
@@ -22412,7 +22431,7 @@ dwarf2_fetch_constant_bytes (sect_offset sect_off,
   struct objfile *objfile = per_cu->dwarf2_per_objfile->objfile;
 
   if (per_cu->cu == NULL)
-    load_cu (per_cu, false);
+    load_cu (per_cu, per_cu->dwarf2_per_objfile, false);
   cu = per_cu->cu;
   if (cu == NULL)
     {
@@ -22534,7 +22553,7 @@ dwarf2_fetch_die_type_sect_off (sect_offset sect_off,
   struct die_info *die;
 
   if (per_cu->cu == NULL)
-    load_cu (per_cu, false);
+    load_cu (per_cu, per_cu->dwarf2_per_objfile, false);
   cu = per_cu->cu;
   if (!cu)
     return NULL;
@@ -22576,7 +22595,7 @@ follow_die_sig_1 (struct die_info *src_die, struct signatured_type *sig_type,
   /* If necessary, add it to the queue and load its DIEs.  */
 
   if (maybe_queue_comp_unit (*ref_cu, &sig_type->per_cu, language_minimal))
-    read_signatured_type (sig_type);
+    read_signatured_type (sig_type, (*ref_cu)->per_objfile);
 
   sig_cu = sig_type->per_cu.cu;
   gdb_assert (sig_cu != NULL);
@@ -22739,7 +22758,8 @@ get_DW_AT_signature_type (struct die_info *die, const struct attribute *attr,
 /* Load the DIEs associated with type unit PER_CU into memory.  */
 
 static void
-load_full_type_unit (struct dwarf2_per_cu_data *per_cu)
+load_full_type_unit (dwarf2_per_cu_data *per_cu,
+                    dwarf2_per_objfile *per_objfile)
 {
   struct signatured_type *sig_type;
 
@@ -22753,7 +22773,7 @@ load_full_type_unit (struct dwarf2_per_cu_data *per_cu)
 
   gdb_assert (per_cu->cu == NULL);
 
-  read_signatured_type (sig_type);
+  read_signatured_type (sig_type, per_objfile);
 
   gdb_assert (per_cu->cu != NULL);
 }
@@ -22763,14 +22783,15 @@ load_full_type_unit (struct dwarf2_per_cu_data *per_cu)
    read in the real type from the DWO file as well.  */
 
 static void
-read_signatured_type (struct signatured_type *sig_type)
+read_signatured_type (signatured_type *sig_type,
+                     dwarf2_per_objfile *per_objfile)
 {
   struct dwarf2_per_cu_data *per_cu = &sig_type->per_cu;
 
   gdb_assert (per_cu->is_debug_types);
   gdb_assert (per_cu->cu == NULL);
 
-  cutu_reader reader (per_cu, NULL, 0, false);
+  cutu_reader reader (per_cu, per_objfile, NULL, 0, false);
 
   if (!reader.dummy_p)
     {