This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
[RFA 4/4] Improve DWARF Type Unit performance
- From: dje at google dot com (Doug Evans)
- To: gdb-patches at sourceware dot org
- Date: Tue, 10 Jul 2012 02:01:03 -0700 (PDT)
- Subject: [RFA 4/4] Improve DWARF Type Unit performance
This patch is the main patch in the series and contains
the changes to dwarf2read.c.
2012-07-09 Doug Evans <dje@google.com>
* dwarf2read.c (dwarf2_per_objfile): New members n_type_unit_groups,
all_type_unit_groups, type_unit_groups, tu_stats.
(dwarf2_per_cu_data): Move "imported_symtabs" into new union "s".
All uses updated. Add type_unit_group to union "s".
(type_unit_group): New struct.
(IS_TYPE_UNIT_GROUP): New macro.
(abbrev_table): Delete unused member "section".
(dw2_do_instantiate_symtab): Early exit if type_unit_group.
(dw2_get_cu): Assert not used with type_unit_group.
(dw2_get_primary_cu): New function.
(dw2_build_type_unit_groups_reader): New function.
(dw2_build_type_unit_groups): New function.
(dw2_get_file_names): Assert not called on type units.
(dw2_map_symtabs_matching_filename): Call dw2_build_type_unit_groups.
Redo loop to iterate over type unit groups instead of type units.
(dw2_expand_symtabs_matching, dw2_map_symbol_filenames): Ditto.
(read_abbrev_offset): New function.
(init_cutu_and_read_dies): New arg "abbrev_table". All callers
updated.
(create_partial_symtab): New function.
(process_psymtab_comp_unit_reader): Assert not used with type units.
Call create_partial_symtab.
(process_psymtab_type_unit): Delete.
(hash_type_unit_group, eq_type_unit_group): New functions.
(allocate_type_unit_groups_table): New function.
(NO_STMT_LIST_TYPE_UNIT_PSYMTAB): New macro.
(NO_STMT_LIST_TYPE_UNIT_PSYMTAB_SIZE): New macro.
(create_type_unit_group, get_type_unit_group): New functions.
(tu_abbrev_offset): New struct.
(sort_tu_by_abbrev_offset): New function.
(add_type_unit_group_to_table): New function.
(build_type_unit_groups): New function.
(build_type_psymtabs_reader): New function.
(build_type_psymtab_dependencies): New function.
(build_type_psymtabs): Rewrite.
(scan_partial_symbols): Flag an error if a DW_TAG_imported_unit
is seen in a type unit.
(process_queue): Move symtab expansion debugging printfs here.
Call process_full_type_unit for type units.
(compute_symtab_includes): Assert not called for type units.
(process_cu_includes): Don't call compute_symtab_includes for
type units.
(process_full_type_unit): New function.
(process_imported_unit_die): Flag an error if called for type units.
(handle_DW_AT_stmt_list): Delete arg "want_line_info". All callers
updated. Assert not called for type units.
(read_file_scope): Call dwarf2_start_symtab.
(setup_type_unit_groups): New function.
(read_type_unit_scope): Rewrite.
(abbrev_table_read_table): Initialize abbrev_table->offset.
(abbrev_table_free_cleanup): New function.
(dwarf2_start_symtab): New function.
(load_full_type_unit): Assert not called for type unit groups.
Index: dwarf2read.c
===================================================================
RCS file: /cvs/src/src/gdb/dwarf2read.c,v
retrieving revision 1.683
diff -u -p -r1.683 dwarf2read.c
--- dwarf2read.c 10 Jul 2012 01:05:06 -0000 1.683
+++ dwarf2read.c 10 Jul 2012 06:11:18 -0000
@@ -214,14 +214,34 @@ struct dwarf2_per_objfile
/* The .debug_types-related CUs (TUs). */
struct signatured_type **all_type_units;
- /* A chain of compilation units that are currently read in, so that
- they can be freed later. */
- struct dwarf2_per_cu_data *read_in_chain;
+ /* The number of entries in all_type_unit_groups. */
+ int n_type_unit_groups;
+
+ /* Table of type unit groups.
+ This exists to make it easy to iterate over all CUs and TU groups. */
+ struct type_unit_group **all_type_unit_groups;
+
+ /* Table of struct type_unit_group objects.
+ The hash key is the DW_AT_stmt_list value. */
+ htab_t type_unit_groups;
/* A table mapping .debug_types signatures to its signatured_type entry.
This is NULL if the .debug_types section hasn't been read in yet. */
htab_t signatured_types;
+ /* Type unit statistics, to see how well the scaling improvements
+ are doing. */
+ struct tu_stats {
+ int nr_uniq_abbrev_tables;
+ int nr_symtabs;
+ int nr_symtab_sharers;
+ int nr_stmt_less_type_units;
+ } tu_stats;
+
+ /* A chain of compilation units that are currently read in, so that
+ they can be freed later. */
+ struct dwarf2_per_cu_data *read_in_chain;
+
/* A table mapping DW_AT_dwo_name values to struct dwo_file objects.
This is NULL if the table hasn't been allocated yet. */
htab_t dwo_files;
@@ -533,11 +553,19 @@ struct dwarf2_per_cu_data
struct dwarf2_per_cu_quick_data *quick;
} v;
- /* The CUs we import using DW_TAG_imported_unit. This is filled in
- while reading psymtabs, used to compute the psymtab dependencies,
- and then cleared. Then it is filled in again while reading full
- symbols, and only deleted when the objfile is destroyed. */
- VEC (dwarf2_per_cu_ptr) *imported_symtabs;
+ union
+ {
+ /* The CUs we import using DW_TAG_imported_unit. This is filled in
+ while reading psymtabs, used to compute the psymtab dependencies,
+ and then cleared. Then it is filled in again while reading full
+ symbols, and only deleted when the objfile is destroyed. */
+ VEC (dwarf2_per_cu_ptr) *imported_symtabs;
+
+ /* Type units are grouped by their DW_AT_stmt_list entry so that they
+ can share them. If this is a TU, this points to the containing
+ symtab. */
+ struct type_unit_group *type_unit_group;
+ } s;
};
/* Entry in the signatured_types hash table. */
@@ -564,6 +592,47 @@ struct signatured_type
sect_offset type_offset_in_section;
};
+/* Each element of dwarf2_per_objfile->type_unit_groups is a pointer to
+ an object of this type. */
+
+struct type_unit_group
+{
+ /* dwarf2read.c's main "handle" on the symtab.
+ To simplify things we create an artificial CU that "includes" all the
+ type units using this stmt_list so that the rest of the code still has
+ a "per_cu" handle on the symtab.
+ This PER_CU is recognized by having no section. */
+#define IS_TYPE_UNIT_GROUP(per_cu) ((per_cu)->info_or_types_section == NULL)
+ struct dwarf2_per_cu_data *per_cu;
+
+ /* The TUs that share this DW_AT_stmt_list entry.
+ This is added to while parsing type units to build partial symtabs,
+ and is deleted afterwards and not used again. */
+ VEC (dwarf2_per_cu_ptr) *tus;
+
+ /* The primary symtab.
+ Type units don't have DW_AT_name so we create an essentially
+ anonymous symtab as the primary symtab. */
+ struct symtab *primary_symtab;
+
+ /* Offset in .debug_line. This is the hash key. */
+ sect_offset line_offset;
+
+ /* The number of symtabs from the line header.
+ The value here must match line_header.num_file_names. */
+ unsigned int num_symtabs;
+
+ /* The symbol tables for this TU (obtained from the files listed in
+ DW_AT_stmt_list).
+ WARNING: The order of entries here must match the order of entries
+ in the line header. After the first TU using this type_unit_group, the
+ line header for the subsequent TUs is recreated from this. This is done
+ because we need to use the same symtabs for each TU using the same
+ DW_AT_stmt_list value. Also note that symtabs may be repeated here,
+ there's no guarantee the line header doesn't have duplicate entries. */
+ struct symtab **symtabs;
+};
+
/* These sections are what may appear in a "dwo" file. */
struct dwo_sections
@@ -804,8 +873,8 @@ struct attr_abbrev
struct abbrev_table
{
- /* Where the abbrev table came from. */
- struct dwarf2_section_info *section;
+ /* Where the abbrev table came from.
+ This is used as a sanity check when the table is used. */
sect_offset offset;
/* Storage for the abbrev table. */
@@ -1105,6 +1174,8 @@ static struct abbrev_table *abbrev_table
static void abbrev_table_free (struct abbrev_table *);
+static void abbrev_table_free_cleanup (void *);
+
static void dwarf2_read_abbrevs (struct dwarf2_cu *,
struct dwarf2_section_info *);
@@ -1155,6 +1226,9 @@ static LONGEST read_offset (bfd *, gdb_b
static LONGEST read_offset_1 (bfd *, gdb_byte *, unsigned int);
+static sect_offset read_abbrev_offset (struct dwarf2_section_info *,
+ sect_offset);
+
static gdb_byte *read_n_bytes (bfd *, gdb_byte *, unsigned int);
static char *read_direct_string (bfd *, gdb_byte *, unsigned int *);
@@ -1203,6 +1277,9 @@ static void dwarf_decode_lines (struct l
static void dwarf2_start_subfile (char *, const char *, const char *);
+static void dwarf2_start_symtab (struct dwarf2_cu *,
+ char *, char *, CORE_ADDR);
+
static struct symbol *new_symbol (struct die_info *, struct type *,
struct dwarf2_cu *);
@@ -1387,6 +1464,11 @@ static void load_full_type_unit (struct
static void read_signatured_type (struct signatured_type *);
+static struct type_unit_group *get_type_unit_group
+ (struct dwarf2_per_cu_data *, struct attribute *);
+
+static void build_type_unit_groups (die_reader_func_ftype *, void *);
+
/* memory allocation interface */
static struct dwarf_block *dwarf_alloc_block (struct dwarf2_cu *);
@@ -1451,6 +1533,9 @@ static void load_full_comp_unit (struct
static void process_full_comp_unit (struct dwarf2_per_cu_data *,
enum language);
+static void process_full_type_unit (struct dwarf2_per_cu_data *,
+ enum language);
+
static void dwarf2_add_dependence (struct dwarf2_cu *,
struct dwarf2_per_cu_data *);
@@ -1482,7 +1567,8 @@ static char *file_full_name (int file, s
const char *comp_dir);
static void init_cutu_and_read_dies
- (struct dwarf2_per_cu_data *this_cu, int use_existing_cu, int keep,
+ (struct dwarf2_per_cu_data *this_cu, struct abbrev_table *abbrev_table,
+ int use_existing_cu, int keep,
die_reader_func_ftype *die_reader_func, void *data);
static void init_cutu_and_read_dies_simple
@@ -1491,8 +1577,6 @@ static void init_cutu_and_read_dies_simp
static htab_t allocate_signatured_type_table (struct objfile *objfile);
-static void process_psymtab_comp_unit (struct dwarf2_per_cu_data *, int);
-
static htab_t allocate_dwo_unit_table (struct objfile *objfile);
static struct dwo_unit *lookup_dwo_comp_unit
@@ -2026,6 +2110,11 @@ dw2_do_instantiate_symtab (struct dwarf2
{
struct cleanup *back_to;
+ /* Skip type_unit_groups, reading the type units they contain
+ is handled elsewhere. */
+ if (IS_TYPE_UNIT_GROUP (per_cu))
+ return;
+
back_to = make_cleanup (dwarf2_release_queue, NULL);
if (dwarf2_per_objfile->using_index
@@ -2064,16 +2153,63 @@ dw2_instantiate_symtab (struct dwarf2_pe
return per_cu->v.quick->symtab;
}
-/* Return the CU given its index. */
+/* Return the CU given its index.
+
+ This is intended for loops like:
+
+ for (i = 0; i < (dwarf2_per_objfile->n_comp_units
+ + dwarf2_per_objfile->n_type_units); ++i)
+ {
+ struct dwarf2_per_cu_data *per_cu = dw2_get_cu (i);
+
+ ...;
+ }
+*/
static struct dwarf2_per_cu_data *
dw2_get_cu (int index)
{
if (index >= dwarf2_per_objfile->n_comp_units)
{
+ struct dwarf2_per_cu_data *per_cu;
+
index -= dwarf2_per_objfile->n_comp_units;
- return &dwarf2_per_objfile->all_type_units[index]->per_cu;
+ per_cu = &dwarf2_per_objfile->all_type_units[index]->per_cu;
+ gdb_assert (! IS_TYPE_UNIT_GROUP (per_cu));
+ return per_cu;
}
+
+ return dwarf2_per_objfile->all_comp_units[index];
+}
+
+/* Return the primary CU given its index.
+ The difference between this function and dw2_get_cu is in the handling
+ of type units (TUs). Here we return the type_unit_group object.
+
+ This is intended for loops like:
+
+ for (i = 0; i < (dwarf2_per_objfile->n_comp_units
+ + dwarf2_per_objfile->n_type_unit_groups); ++i)
+ {
+ struct dwarf2_per_cu_data *per_cu = dw2_get_primary_cu (i);
+
+ ...;
+ }
+*/
+
+static struct dwarf2_per_cu_data *
+dw2_get_primary_cu (int index)
+{
+ if (index >= dwarf2_per_objfile->n_comp_units)
+ {
+ struct dwarf2_per_cu_data *per_cu;
+
+ index -= dwarf2_per_objfile->n_comp_units;
+ per_cu = dwarf2_per_objfile->all_type_unit_groups[index]->per_cu;
+ gdb_assert (IS_TYPE_UNIT_GROUP (per_cu));
+ return per_cu;
+ }
+
return dwarf2_per_objfile->all_comp_units[index];
}
@@ -2474,6 +2610,41 @@ dw2_setup (struct objfile *objfile)
gdb_assert (dwarf2_per_objfile);
}
+/* Reader function for dw2_build_type_unit_groups. */
+
+static void
+dw2_build_type_unit_groups_reader (const struct die_reader_specs *reader,
+ gdb_byte *info_ptr,
+ struct die_info *type_unit_die,
+ int has_children,
+ void *data)
+{
+ struct dwarf2_cu *cu = reader->cu;
+ struct dwarf2_per_cu_data *per_cu = cu->per_cu;
+ struct attribute *attr;
+ struct type_unit_group *tu_group;
+
+ gdb_assert (data == NULL);
+
+ if (! has_children)
+ return;
+
+ attr = dwarf2_attr_no_follow (type_unit_die, DW_AT_stmt_list);
+ /* Call this for its side-effect of creating the associated
+ struct type_unit_group if it doesn't already exist. */
+ tu_group = get_type_unit_group (per_cu, attr);
+}
+
+/* Build dwarf2_per_objfile->type_unit_groups.
+ This function may be called multiple times. */
+
+static void
+dw2_build_type_unit_groups (void)
+{
+ if (dwarf2_per_objfile->type_unit_groups == NULL)
+ build_type_unit_groups (dw2_build_type_unit_groups_reader, NULL);
+}
+
/* die_reader_func for dw2_get_file_names. */
static void
@@ -2558,6 +2729,10 @@ static struct quick_file_names *
dw2_get_file_names (struct objfile *objfile,
struct dwarf2_per_cu_data *this_cu)
{
+ /* For TUs this should only be called on the parent group. */
+ if (this_cu->is_debug_types)
+ gdb_assert (IS_TYPE_UNIT_GROUP (this_cu));
+
if (this_cu->v.quick->file_names != NULL)
return this_cu->v.quick->file_names;
/* If we know there is no line data, no point in looking again. */
@@ -2569,7 +2744,8 @@ dw2_get_file_names (struct objfile *objf
However, that's not the case for TUs where DW_AT_stmt_list lives in the
DWO file. */
if (this_cu->is_debug_types)
- init_cutu_and_read_dies (this_cu, 0, 0, dw2_get_file_names_reader, NULL);
+ init_cutu_and_read_dies (this_cu, NULL, 0, 0,
+ dw2_get_file_names_reader, NULL);
else
init_cutu_and_read_dies_simple (this_cu, dw2_get_file_names_reader, NULL);
@@ -2675,11 +2851,13 @@ dw2_map_symtabs_matching_filename (struc
dw2_setup (objfile);
+ dw2_build_type_unit_groups ();
+
for (i = 0; i < (dwarf2_per_objfile->n_comp_units
- + dwarf2_per_objfile->n_type_units); ++i)
+ + dwarf2_per_objfile->n_type_unit_groups); ++i)
{
int j;
- struct dwarf2_per_cu_data *per_cu = dw2_get_cu (i);
+ struct dwarf2_per_cu_data *per_cu = dw2_get_primary_cu (i);
struct quick_file_names *file_data;
/* We only need to look at symtabs not already expanded. */
@@ -2997,8 +3175,8 @@ dw2_find_symbol_file (struct objfile *ob
if (per_cu->v.quick->symtab != NULL)
return per_cu->v.quick->symtab->filename;
- init_cutu_and_read_dies (per_cu, 0, 0, dw2_get_primary_filename_reader,
- &filename);
+ init_cutu_and_read_dies (per_cu, NULL, 0, 0,
+ dw2_get_primary_filename_reader, &filename);
return filename;
}
@@ -3040,6 +3218,8 @@ dw2_expand_symtabs_matching
struct cleanup *cleanup;
htab_t visited_found, visited_not_found;
+ dw2_build_type_unit_groups ();
+
visited_found = htab_create_alloc (10,
htab_hash_pointer, htab_eq_pointer,
NULL, xcalloc, xfree);
@@ -3050,10 +3230,10 @@ dw2_expand_symtabs_matching
make_cleanup_htab_delete (visited_not_found);
for (i = 0; i < (dwarf2_per_objfile->n_comp_units
- + dwarf2_per_objfile->n_type_units); ++i)
+ + dwarf2_per_objfile->n_type_unit_groups); ++i)
{
int j;
- struct dwarf2_per_cu_data *per_cu = dw2_get_cu (i);
+ struct dwarf2_per_cu_data *per_cu = dw2_get_primary_cu (i);
struct quick_file_names *file_data;
void **slot;
@@ -3223,6 +3403,8 @@ dw2_map_symbol_filenames (struct objfile
cleanup = make_cleanup_htab_delete (visited);
dw2_setup (objfile);
+ dw2_build_type_unit_groups ();
+
/* We can ignore file names coming from already-expanded CUs. */
for (i = 0; i < (dwarf2_per_objfile->n_comp_units
+ dwarf2_per_objfile->n_type_units); ++i)
@@ -3239,10 +3421,10 @@ dw2_map_symbol_filenames (struct objfile
}
for (i = 0; i < (dwarf2_per_objfile->n_comp_units
- + dwarf2_per_objfile->n_type_units); ++i)
+ + dwarf2_per_objfile->n_type_unit_groups); ++i)
{
int j;
- struct dwarf2_per_cu_data *per_cu = dw2_get_cu (i);
+ struct dwarf2_per_cu_data *per_cu = dw2_get_primary_cu (i);
struct quick_file_names *file_data;
void **slot;
@@ -3541,6 +3723,26 @@ read_and_check_type_unit_head (struct co
return info_ptr;
}
+/* Fetch the abbreviation table offset from a comp or type unit header. */
+
+static sect_offset
+read_abbrev_offset (struct dwarf2_section_info *section,
+ sect_offset offset)
+{
+ bfd *abfd = section->asection->owner;
+ gdb_byte *info_ptr;
+ unsigned int length, initial_length_size, offset_size;
+ sect_offset abbrev_offset;
+
+ dwarf2_read_section (dwarf2_per_objfile->objfile, section);
+ info_ptr = section->buffer + offset.sect_off;
+ length = read_initial_length (abfd, info_ptr, &initial_length_size);
+ offset_size = initial_length_size == 4 ? 4 : 8;
+ info_ptr += initial_length_size + 2 /*version*/;
+ abbrev_offset.sect_off = read_offset_1 (abfd, info_ptr, offset_size);
+ return abbrev_offset;
+}
+
/* Allocate a new partial symtab for file named NAME and mark this new
partial symtab as being an include of PST. */
@@ -3876,6 +4078,10 @@ init_cu_die_reader (struct die_reader_sp
/* Initialize a CU (or TU) and read its DIEs.
If the CU defers to a DWO file, read the DWO file as well.
+ ABBREV_TABLE, if non-NULL, is the abbreviation table to use.
+ Otherwise the table specified in the comp unit header is read in and used.
+ This is an optimization for when we already have the abbrev table.
+
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.
@@ -3887,6 +4093,7 @@ init_cu_die_reader (struct die_reader_sp
static void
init_cutu_and_read_dies (struct dwarf2_per_cu_data *this_cu,
+ struct abbrev_table *abbrev_table,
int use_existing_cu, int keep,
die_reader_func_ftype *die_reader_func,
void *data)
@@ -4006,7 +4213,14 @@ init_cutu_and_read_dies (struct dwarf2_p
done. Note that it's important that if the CU had an abbrev table
on entry we don't free it when we're done: Somewhere up the call stack
it may be in use. */
- if (cu->abbrev_table == NULL)
+ if (abbrev_table != NULL)
+ {
+ gdb_assert (cu->abbrev_table == NULL);
+ gdb_assert (cu->header.abbrev_offset.sect_off
+ == abbrev_table->offset.sect_off);
+ cu->abbrev_table = abbrev_table;
+ }
+ else if (cu->abbrev_table == NULL)
{
dwarf2_read_abbrevs (cu, abbrev_section);
make_cleanup (dwarf2_free_abbrev_table, cu);
@@ -4138,8 +4352,16 @@ init_cutu_and_read_dies (struct dwarf2_p
}
/* Discard the original CU's abbrev table, and read the DWO's. */
- dwarf2_free_abbrev_table (cu);
- dwarf2_read_abbrevs (cu, dwo_abbrev_section);
+ if (abbrev_table == NULL)
+ {
+ dwarf2_free_abbrev_table (cu);
+ dwarf2_read_abbrevs (cu, dwo_abbrev_section);
+ }
+ else
+ {
+ dwarf2_read_abbrevs (cu, dwo_abbrev_section);
+ make_cleanup (dwarf2_free_abbrev_table, cu);
+ }
/* Read in the die, but leave space to copy over the attributes
from the stub. This has the benefit of simplifying the rest of
@@ -4294,6 +4516,32 @@ init_cutu_and_read_dies_simple (struct d
die_reader_func, data);
}
+/* Create a psymtab named NAME and assign it to PER_CU.
+
+ The caller must fill in the following details:
+ dirname, textlow, texthigh. */
+
+static struct partial_symtab *
+create_partial_symtab (struct dwarf2_per_cu_data *per_cu, const char *name)
+{
+ struct objfile *objfile = per_cu->objfile;
+ struct partial_symtab *pst;
+
+ pst = start_psymtab_common (objfile, objfile->section_offsets,
+ name, 0,
+ objfile->global_psymbols.next,
+ objfile->static_psymbols.next);
+
+ pst->psymtabs_addrmap_supported = 1;
+
+ /* This is the glue that links PST into GDB's symbol API. */
+ pst->read_symtab_private = per_cu;
+ pst->read_symtab = dwarf2_psymtab_to_symtab;
+ per_cu->v.psymtab = pst;
+
+ return pst;
+}
+
/* die_reader_func for process_psymtab_comp_unit. */
static void
@@ -4319,6 +4567,8 @@ process_psymtab_comp_unit_reader (const
|| !*want_partial_unit_ptr))
return;
+ gdb_assert (! per_cu->is_debug_types);
+
prepare_one_comp_unit (cu, comp_unit_die, language_minimal);
cu->list_in_scope = &file_symbols;
@@ -4329,27 +4579,16 @@ process_psymtab_comp_unit_reader (const
filename = "";
else
filename = DW_STRING (attr);
- pst = start_psymtab_common (objfile, objfile->section_offsets,
- filename,
- /* TEXTLOW and TEXTHIGH are set below. */
- 0,
- objfile->global_psymbols.next,
- objfile->static_psymbols.next);
- pst->psymtabs_addrmap_supported = 1;
+ pst = create_partial_symtab (per_cu, filename);
+
+ /* This must be done before calling dwarf2_build_include_psymtabs. */
attr = dwarf2_attr (comp_unit_die, DW_AT_comp_dir, cu);
if (attr != NULL)
pst->dirname = DW_STRING (attr);
- pst->read_symtab_private = per_cu;
-
baseaddr = ANOFFSET (objfile->section_offsets, SECT_OFF_TEXT (objfile));
- /* Store the function that reads in the rest of the symbol table. */
- pst->read_symtab = dwarf2_psymtab_to_symtab;
-
- per_cu->v.psymtab = pst;
-
dwarf2_find_base_address (comp_unit_die, cu);
/* Possibly set the default values of LOWPC and HIGHPC from
@@ -4401,10 +4640,10 @@ process_psymtab_comp_unit_reader (const
(objfile->static_psymbols.list + pst->statics_offset);
sort_pst_symbols (pst);
- if (!VEC_empty (dwarf2_per_cu_ptr, cu->per_cu->imported_symtabs))
+ if (!VEC_empty (dwarf2_per_cu_ptr, cu->per_cu->s.imported_symtabs))
{
int i;
- int len = VEC_length (dwarf2_per_cu_ptr, cu->per_cu->imported_symtabs);
+ int len = VEC_length (dwarf2_per_cu_ptr, cu->per_cu->s.imported_symtabs);
struct dwarf2_per_cu_data *iter;
/* Fill in 'dependencies' here; we fill in 'users' in a
@@ -4413,25 +4652,17 @@ process_psymtab_comp_unit_reader (const
pst->dependencies = obstack_alloc (&objfile->objfile_obstack,
len * sizeof (struct symtab *));
for (i = 0;
- VEC_iterate (dwarf2_per_cu_ptr, cu->per_cu->imported_symtabs,
+ VEC_iterate (dwarf2_per_cu_ptr, cu->per_cu->s.imported_symtabs,
i, iter);
++i)
pst->dependencies[i] = iter->v.psymtab;
- VEC_free (dwarf2_per_cu_ptr, cu->per_cu->imported_symtabs);
+ VEC_free (dwarf2_per_cu_ptr, cu->per_cu->s.imported_symtabs);
}
- if (per_cu->is_debug_types)
- {
- /* It's not clear we want to do anything with stmt lists here.
- Waiting to see what gcc ultimately does. */
- }
- else
- {
- /* Get the list of files included in the current compilation unit,
- and build a psymtab for each of them. */
- dwarf2_build_include_psymtabs (cu, comp_unit_die, pst);
- }
+ /* Get the list of files included in the current compilation unit,
+ and build a psymtab for each of them. */
+ dwarf2_build_include_psymtabs (cu, comp_unit_die, pst);
if (dwarf2_read_debug)
{
@@ -4464,38 +4695,399 @@ process_psymtab_comp_unit (struct dwarf2
free_one_cached_comp_unit (this_cu);
gdb_assert (! this_cu->is_debug_types);
- init_cutu_and_read_dies (this_cu, 0, 0, process_psymtab_comp_unit_reader,
+ init_cutu_and_read_dies (this_cu, NULL, 0, 0,
+ process_psymtab_comp_unit_reader,
&want_partial_unit);
/* Age out any secondary CUs. */
age_cached_comp_units ();
}
-/* Traversal function for htab_traverse_noresize.
- Process one .debug_types comp-unit. */
+static hashval_t
+hash_type_unit_group (const void *item)
+{
+ const struct type_unit_group *symtab = item;
+
+ return symtab->line_offset.sect_off;
+}
static int
-process_psymtab_type_unit (void **slot, void *info)
+eq_type_unit_group (const void *item_lhs, const void *item_rhs)
{
- struct signatured_type *sig_type = (struct signatured_type *) *slot;
- struct dwarf2_per_cu_data *per_cu = &sig_type->per_cu;
+ const struct type_unit_group *lhs = item_lhs;
+ const struct type_unit_group *rhs = item_rhs;
- gdb_assert (per_cu->is_debug_types);
- gdb_assert (info == NULL);
+ return lhs->line_offset.sect_off == rhs->line_offset.sect_off;
+}
- /* If this compilation unit was already read in, free the
- cached copy in order to read it in again. This is
- necessary because we skipped some symbols when we first
- read in the compilation unit (see load_partial_dies).
- This problem could be avoided, but the benefit is unclear. */
- if (per_cu->cu != NULL)
- free_one_cached_comp_unit (per_cu);
+/* Allocate a hash table for type unit groups. */
+
+static htab_t
+allocate_type_unit_groups_table (void)
+{
+ return htab_create_alloc_ex (3,
+ hash_type_unit_group,
+ eq_type_unit_group,
+ NULL,
+ &dwarf2_per_objfile->objfile->objfile_obstack,
+ hashtab_obstack_allocate,
+ dummy_obstack_deallocate);
+}
+
+/* Type units that don't have DW_AT_stmt_list are grouped into their own
+ partial symtabs. We combine several TUs per psymtab to not let the size
+ of any one psymtab grow too big. */
+#define NO_STMT_LIST_TYPE_UNIT_PSYMTAB (1 << 31)
+#define NO_STMT_LIST_TYPE_UNIT_PSYMTAB_SIZE 10
+
+/* Helper routine for build_type_psymtabs_reader.
+ Create the type_unit_group object used to hold one or more TUs. */
+
+static struct type_unit_group *
+create_type_unit_group (struct dwarf2_per_cu_data *per_cu,
+ sect_offset line_offset_struct)
+{
+ struct objfile *objfile = dwarf2_per_objfile->objfile;
+ struct type_unit_group *tu_group;
+ struct partial_symtab *pst;
+ unsigned int line_offset;
+ char *name;
- init_cutu_and_read_dies (per_cu, 0, 0, process_psymtab_comp_unit_reader,
- NULL);
+ line_offset = line_offset_struct.sect_off;
- /* Age out any secondary CUs. */
- age_cached_comp_units ();
+ /* Give the symtab a useful name for debug purposes. */
+ if ((line_offset & NO_STMT_LIST_TYPE_UNIT_PSYMTAB) != 0)
+ name = xstrprintf ("<type_units_%d>",
+ (line_offset & ~NO_STMT_LIST_TYPE_UNIT_PSYMTAB));
+ else
+ name = xstrprintf ("<type_units_at_0x%x>", line_offset);
+
+ tu_group = OBSTACK_ZALLOC (&objfile->objfile_obstack,
+ struct type_unit_group);
+
+ per_cu = OBSTACK_ZALLOC (&objfile->objfile_obstack,
+ struct dwarf2_per_cu_data);
+ per_cu->objfile = objfile;
+ per_cu->is_debug_types = 1;
+ per_cu->s.type_unit_group = tu_group;
+
+ pst = create_partial_symtab (per_cu, name);
+ pst->anonymous = 1;
+
+ xfree (name);
+
+ tu_group->per_cu = per_cu;
+ tu_group->line_offset.sect_off = line_offset;
+
+ return tu_group;
+}
+
+/* Look up the type_unit_group for PER_CU, and create it if necessary.
+ STMT_LIST is an DW_AT_stmt_list attribute. */
+
+static struct type_unit_group *
+get_type_unit_group (struct dwarf2_per_cu_data *per_cu,
+ struct attribute *stmt_list)
+{
+ struct tu_stats *tu_stats = &dwarf2_per_objfile->tu_stats;
+ struct type_unit_group *tu_group;
+ void **slot;
+ unsigned int line_offset;
+ struct type_unit_group type_unit_group_for_lookup;
+
+ if (dwarf2_per_objfile->type_unit_groups == NULL)
+ {
+ dwarf2_per_objfile->type_unit_groups =
+ allocate_type_unit_groups_table ();
+ }
+
+ /* Do we need to create a new group, or can we use an existing one? */
+
+ if (stmt_list)
+ {
+ line_offset = DW_UNSND (stmt_list);
+ ++tu_stats->nr_symtab_sharers;
+ }
+ else
+ {
+ /* Ugh, no stmt_list. Rare, but we have to handle it.
+ We can do various things here like create one group per TU or
+ spread them over multiple groups to split up the expansion work.
+ To avoid worst case scenarios (too many groups or too large groups)
+ we, umm, group them in bunches. */
+ line_offset = (NO_STMT_LIST_TYPE_UNIT_PSYMTAB
+ | (tu_stats->nr_stmt_less_type_units
+ / NO_STMT_LIST_TYPE_UNIT_PSYMTAB_SIZE));
+ ++tu_stats->nr_stmt_less_type_units;
+ }
+
+ type_unit_group_for_lookup.line_offset.sect_off = line_offset;
+ slot = htab_find_slot (dwarf2_per_objfile->type_unit_groups,
+ &type_unit_group_for_lookup, INSERT);
+ if (*slot != NULL)
+ {
+ tu_group = *slot;
+ gdb_assert (tu_group != NULL);
+ }
+ else
+ {
+ sect_offset line_offset_struct;
+
+ line_offset_struct.sect_off = line_offset;
+ tu_group = create_type_unit_group (per_cu, line_offset_struct);
+ *slot = tu_group;
+ ++tu_stats->nr_symtabs;
+ }
+
+ return tu_group;
+}
+
+/* Struct used to sort TUs by their abbreviation table offset. */
+
+struct tu_abbrev_offset
+{
+ struct signatured_type *sig_type;
+ sect_offset abbrev_offset;
+};
+
+/* Helper routine for build_type_unit_groups, passed to qsort. */
+
+static int
+sort_tu_by_abbrev_offset (const void *ap, const void *bp)
+{
+ const struct tu_abbrev_offset * const *a = ap;
+ const struct tu_abbrev_offset * const *b = bp;
+ unsigned int aoff = (*a)->abbrev_offset.sect_off;
+ unsigned int boff = (*b)->abbrev_offset.sect_off;
+
+ return (aoff > boff) - (aoff < boff);
+}
+
+/* A helper function to add a type_unit_group to a table. */
+
+static int
+add_type_unit_group_to_table (void **slot, void *datum)
+{
+ struct type_unit_group *tu_group = *slot;
+ struct type_unit_group ***datap = datum;
+
+ **datap = tu_group;
+ ++*datap;
+
+ return 1;
+}
+
+/* Efficiently read all the type units, calling init_cutu_and_read_dies on
+ each one passing FUNC,DATA.
+
+ The efficiency is because we sort TUs by the abbrev table they use and
+ only read each abbrev table once. In one program there are 200K TUs
+ sharing 8K abbrev tables.
+
+ The main purpose of this function is to support building the
+ dwarf2_per_objfile->type_unit_groups table.
+ TUs typically share the DW_AT_stmt_list of the CU they came from, so we
+ can collapse the search space by grouping them by stmt_list.
+ The savings can be significant, in the same program from above the 200K TUs
+ share 8K stmt_list tables.
+
+ FUNC is expected to call get_type_unit_group, which will create the
+ struct type_unit_group if necessary and add it to
+ dwarf2_per_objfile->type_unit_groups. */
+
+static void
+build_type_unit_groups (die_reader_func_ftype *func, void *data)
+{
+ struct objfile *objfile = dwarf2_per_objfile->objfile;
+ struct tu_stats *tu_stats = &dwarf2_per_objfile->tu_stats;
+ struct cleanup *cleanups;
+ struct abbrev_table *abbrev_table;
+ sect_offset abbrev_offset;
+ struct tu_abbrev_offset *sorted_by_abbrev;
+ struct type_unit_group **iter;
+ int i;
+
+ /* It's up to the caller to not call us multiple times. */
+ gdb_assert (dwarf2_per_objfile->type_unit_groups == NULL);
+
+ if (dwarf2_per_objfile->n_type_units == 0)
+ return;
+
+ /* TUs typically share abbrev tables, and there can be way more TUs than
+ abbrev tables. Sort by abbrev table to reduce the number of times we
+ read each abbrev table in.
+ Alternatives are to punt or to maintain a cache of abbrev tables.
+ This is simpler and efficient enough for now.
+
+ Later we group TUs by their DW_AT_stmt_list value (as this defines the
+ symtab to use). Typically TUs with the same abbrev offset have the same
+ stmt_list value too so in practice this should work well.
+
+ The basic algorithm here is:
+
+ sort TUs by abbrev table
+ for each TU with same abbrev table:
+ read abbrev table if first user
+ read TU top level DIE
+ [IWBN if DWO skeletons had DW_AT_stmt_list]
+ call FUNC */
+
+ if (dwarf2_read_debug)
+ fprintf_unfiltered (gdb_stdlog, "Building type unit groups ...\n");
+
+ /* Sort in a separate table to maintain the order of all_type_units
+ for .gdb_index: TU indices directly index all_type_units. */
+ sorted_by_abbrev = XNEWVEC (struct tu_abbrev_offset,
+ dwarf2_per_objfile->n_type_units);
+ for (i = 0; i < dwarf2_per_objfile->n_type_units; ++i)
+ {
+ struct signatured_type *sig_type = dwarf2_per_objfile->all_type_units[i];
+
+ sorted_by_abbrev[i].sig_type = sig_type;
+ sorted_by_abbrev[i].abbrev_offset =
+ read_abbrev_offset (sig_type->per_cu.info_or_types_section,
+ sig_type->per_cu.offset);
+ }
+ cleanups = make_cleanup (xfree, sorted_by_abbrev);
+ qsort (sorted_by_abbrev, dwarf2_per_objfile->n_type_units,
+ sizeof (struct tu_abbrev_offset), sort_tu_by_abbrev_offset);
+
+ memset (tu_stats, 0, sizeof (*tu_stats));
+ abbrev_offset.sect_off = ~(unsigned) 0;
+ abbrev_table = NULL;
+ make_cleanup (abbrev_table_free_cleanup, &abbrev_table);
+
+ for (i = 0; i < dwarf2_per_objfile->n_type_units; ++i)
+ {
+ const struct tu_abbrev_offset *tu = &sorted_by_abbrev[i];
+
+ /* Switch to the next abbrev table if necessary. */
+ if (abbrev_table == NULL
+ || tu->abbrev_offset.sect_off != abbrev_offset.sect_off)
+ {
+ if (abbrev_table != NULL)
+ {
+ abbrev_table_free (abbrev_table);
+ /* Reset to NULL in case abbrev_table_read_table throws
+ an error: abbrev_table_free_cleanup will get called. */
+ abbrev_table = NULL;
+ }
+ abbrev_offset = tu->abbrev_offset;
+ abbrev_table =
+ abbrev_table_read_table (&dwarf2_per_objfile->abbrev,
+ abbrev_offset);
+ ++tu_stats->nr_uniq_abbrev_tables;
+ }
+
+ init_cutu_and_read_dies (&tu->sig_type->per_cu, abbrev_table, 0, 0,
+ func, data);
+ }
+
+ /* Create a vector of pointers to primary type units to make it easy to
+ iterate over them and CUs. See dw2_get_primary_cu. */
+ dwarf2_per_objfile->n_type_unit_groups =
+ htab_elements (dwarf2_per_objfile->type_unit_groups);
+ dwarf2_per_objfile->all_type_unit_groups =
+ obstack_alloc (&objfile->objfile_obstack,
+ dwarf2_per_objfile->n_type_unit_groups
+ * sizeof (struct type_unit_group *));
+ iter = &dwarf2_per_objfile->all_type_unit_groups[0];
+ htab_traverse_noresize (dwarf2_per_objfile->type_unit_groups,
+ add_type_unit_group_to_table, &iter);
+ gdb_assert (iter - &dwarf2_per_objfile->all_type_unit_groups[0]
+ == dwarf2_per_objfile->n_type_unit_groups);
+
+ do_cleanups (cleanups);
+
+ if (dwarf2_read_debug)
+ {
+ fprintf_unfiltered (gdb_stdlog, "Done building type unit groups:\n");
+ fprintf_unfiltered (gdb_stdlog, " %d TUs\n",
+ dwarf2_per_objfile->n_type_units);
+ fprintf_unfiltered (gdb_stdlog, " %d uniq abbrev tables\n",
+ tu_stats->nr_uniq_abbrev_tables);
+ fprintf_unfiltered (gdb_stdlog, " %d symtabs from stmt_list entries\n",
+ tu_stats->nr_symtabs);
+ fprintf_unfiltered (gdb_stdlog, " %d symtab sharers\n",
+ tu_stats->nr_symtab_sharers);
+ fprintf_unfiltered (gdb_stdlog, " %d type units without a stmt_list\n",
+ tu_stats->nr_stmt_less_type_units);
+ }
+}
+
+/* Reader function for build_type_psymtabs. */
+
+static void
+build_type_psymtabs_reader (const struct die_reader_specs *reader,
+ gdb_byte *info_ptr,
+ struct die_info *type_unit_die,
+ int has_children,
+ void *data)
+{
+ struct objfile *objfile = dwarf2_per_objfile->objfile;
+ struct dwarf2_cu *cu = reader->cu;
+ struct dwarf2_per_cu_data *per_cu = cu->per_cu;
+ struct type_unit_group *tu_group;
+ struct attribute *attr;
+ struct partial_die_info *first_die;
+ CORE_ADDR lowpc, highpc;
+ struct partial_symtab *pst;
+
+ gdb_assert (data == NULL);
+
+ if (! has_children)
+ return;
+
+ attr = dwarf2_attr_no_follow (type_unit_die, DW_AT_stmt_list);
+ tu_group = get_type_unit_group (per_cu, attr);
+
+ VEC_safe_push (dwarf2_per_cu_ptr, tu_group->tus, per_cu);
+
+ prepare_one_comp_unit (cu, type_unit_die, language_minimal);
+ cu->list_in_scope = &file_symbols;
+ pst = create_partial_symtab (per_cu, "");
+ pst->anonymous = 1;
+
+ first_die = load_partial_dies (reader, info_ptr, 1);
+
+ lowpc = (CORE_ADDR) -1;
+ highpc = (CORE_ADDR) 0;
+ scan_partial_symbols (first_die, &lowpc, &highpc, 0, cu);
+
+ pst->n_global_syms = objfile->global_psymbols.next -
+ (objfile->global_psymbols.list + pst->globals_offset);
+ pst->n_static_syms = objfile->static_psymbols.next -
+ (objfile->static_psymbols.list + pst->statics_offset);
+ sort_pst_symbols (pst);
+}
+
+/* Traversal function for build_type_psymtabs. */
+
+static int
+build_type_psymtab_dependencies (void **slot, void *info)
+{
+ struct objfile *objfile = dwarf2_per_objfile->objfile;
+ struct type_unit_group *tu_group = (struct type_unit_group *) *slot;
+ struct dwarf2_per_cu_data *per_cu = tu_group->per_cu;
+ struct partial_symtab *pst = per_cu->v.psymtab;
+ int len = VEC_length (dwarf2_per_cu_ptr, tu_group->tus);
+ struct dwarf2_per_cu_data *iter;
+ int i;
+
+ gdb_assert (len > 0);
+
+ pst->number_of_dependencies = len;
+ pst->dependencies = obstack_alloc (&objfile->objfile_obstack,
+ len * sizeof (struct psymtab *));
+ for (i = 0;
+ VEC_iterate (dwarf2_per_cu_ptr, tu_group->tus, i, iter);
+ ++i)
+ {
+ pst->dependencies[i] = iter->v.psymtab;
+ iter->s.type_unit_group = tu_group;
+ }
+
+ VEC_free (dwarf2_per_cu_ptr, tu_group->tus);
return 1;
}
@@ -4509,8 +5101,11 @@ build_type_psymtabs (struct objfile *obj
if (! create_all_type_units (objfile))
return;
- htab_traverse_noresize (dwarf2_per_objfile->signatured_types,
- process_psymtab_type_unit, NULL);
+ build_type_unit_groups (build_type_psymtabs_reader, NULL);
+
+ /* Now that all TUs have been processed we can fill in the dependencies. */
+ htab_traverse_noresize (dwarf2_per_objfile->type_unit_groups,
+ build_type_psymtab_dependencies, NULL);
}
/* A cleanup function that clears objfile's psymtabs_addrmap field. */
@@ -4626,7 +5221,8 @@ load_partial_comp_unit_reader (const str
static void
load_partial_comp_unit (struct dwarf2_per_cu_data *this_cu)
{
- init_cutu_and_read_dies (this_cu, 1, 1, load_partial_comp_unit_reader, NULL);
+ init_cutu_and_read_dies (this_cu, NULL, 1, 1,
+ load_partial_comp_unit_reader, NULL);
}
/* Create a list of all compilation units in OBJFILE.
@@ -4765,6 +5361,14 @@ scan_partial_symbols (struct partial_die
{
struct dwarf2_per_cu_data *per_cu;
+ /* For now we don't handle imported units in type units. */
+ if (cu->per_cu->is_debug_types)
+ {
+ error (_("Dwarf Error: DW_TAG_imported_unit is not"
+ " supported in type units [in module %s]"),
+ cu->objfile->name);
+ }
+
per_cu = dwarf2_find_containing_comp_unit (pdi->d.offset,
cu->objfile);
@@ -4772,8 +5376,8 @@ scan_partial_symbols (struct partial_die
if (per_cu->v.psymtab == NULL)
process_psymtab_comp_unit (per_cu, 1);
- VEC_safe_push (dwarf2_per_cu_ptr, cu->per_cu->imported_symtabs,
- per_cu);
+ VEC_safe_push (dwarf2_per_cu_ptr,
+ cu->per_cu->s.imported_symtabs, per_cu);
}
break;
default:
@@ -5558,7 +6162,30 @@ process_queue (void)
if (dwarf2_per_objfile->using_index
? !item->per_cu->v.quick->symtab
: (item->per_cu->v.psymtab && !item->per_cu->v.psymtab->readin))
- process_full_comp_unit (item->per_cu, item->pretend_language);
+ {
+ struct dwarf2_per_cu_data *per_cu = item->per_cu;
+
+ if (dwarf2_read_debug)
+ {
+ fprintf_unfiltered (gdb_stdlog,
+ "Expanding symtab of %s at offset 0x%x\n",
+ per_cu->is_debug_types ? "TU" : "CU",
+ per_cu->offset.sect_off);
+ }
+
+ if (per_cu->is_debug_types)
+ process_full_type_unit (per_cu, item->pretend_language);
+ else
+ process_full_comp_unit (per_cu, item->pretend_language);
+
+ if (dwarf2_read_debug)
+ {
+ fprintf_unfiltered (gdb_stdlog,
+ "Done expanding %s at offset 0x%x\n",
+ per_cu->is_debug_types ? "TU" : "CU",
+ per_cu->offset.sect_off);
+ }
+ }
item->per_cu->queued = 0;
next_item = item->next;
@@ -5716,8 +6343,8 @@ load_full_comp_unit (struct dwarf2_per_c
{
gdb_assert (! this_cu->is_debug_types);
- init_cutu_and_read_dies (this_cu, 1, 1, load_full_comp_unit_reader,
- &pretend_language);
+ init_cutu_and_read_dies (this_cu, NULL, 1, 1,
+ load_full_comp_unit_reader, &pretend_language);
}
/* Add a DIE to the delayed physname list. */
@@ -5880,7 +6507,7 @@ recursively_compute_inclusions (VEC (dwa
VEC_safe_push (dwarf2_per_cu_ptr, *result, per_cu);
for (ix = 0;
- VEC_iterate (dwarf2_per_cu_ptr, per_cu->imported_symtabs, ix, iter);
+ VEC_iterate (dwarf2_per_cu_ptr, per_cu->s.imported_symtabs, ix, iter);
++ix)
recursively_compute_inclusions (result, all_children, iter);
}
@@ -5891,7 +6518,9 @@ recursively_compute_inclusions (VEC (dwa
static void
compute_symtab_includes (struct dwarf2_per_cu_data *per_cu)
{
- if (!VEC_empty (dwarf2_per_cu_ptr, per_cu->imported_symtabs))
+ gdb_assert (! per_cu->is_debug_types);
+
+ if (!VEC_empty (dwarf2_per_cu_ptr, per_cu->s.imported_symtabs))
{
int ix, len;
struct dwarf2_per_cu_data *iter;
@@ -5907,7 +6536,7 @@ compute_symtab_includes (struct dwarf2_p
NULL, xcalloc, xfree);
for (ix = 0;
- VEC_iterate (dwarf2_per_cu_ptr, per_cu->imported_symtabs,
+ VEC_iterate (dwarf2_per_cu_ptr, per_cu->s.imported_symtabs,
ix, iter);
++ix)
recursively_compute_inclusions (&result_children, all_children, iter);
@@ -5942,7 +6571,10 @@ process_cu_includes (void)
VEC_iterate (dwarf2_per_cu_ptr, dwarf2_per_objfile->just_read_cus,
ix, iter);
++ix)
- compute_symtab_includes (iter);
+ {
+ if (! iter->is_debug_types)
+ compute_symtab_includes (iter);
+ }
VEC_free (dwarf2_per_cu_ptr, dwarf2_per_objfile->just_read_cus);
}
@@ -5961,14 +6593,6 @@ process_full_comp_unit (struct dwarf2_pe
struct cleanup *back_to, *delayed_list_cleanup;
CORE_ADDR baseaddr;
- if (dwarf2_read_debug)
- {
- fprintf_unfiltered (gdb_stdlog,
- "Expanding symtab of %s at offset 0x%x\n",
- per_cu->is_debug_types ? "TU" : "CU",
- per_cu->offset.sect_off);
- }
-
baseaddr = ANOFFSET (objfile->section_offsets, SECT_OFF_TEXT (objfile));
buildsym_init ();
@@ -6044,14 +6668,78 @@ process_full_comp_unit (struct dwarf2_pe
VEC_safe_push (dwarf2_per_cu_ptr, dwarf2_per_objfile->just_read_cus, per_cu);
do_cleanups (back_to);
+}
- if (dwarf2_read_debug)
+/* Generate full symbol information for type unit PER_CU, whose DIEs have
+ already been loaded into memory. */
+
+static void
+process_full_type_unit (struct dwarf2_per_cu_data *per_cu,
+ enum language pretend_language)
+{
+ struct dwarf2_cu *cu = per_cu->cu;
+ struct objfile *objfile = per_cu->objfile;
+ struct symtab *symtab;
+ struct cleanup *back_to, *delayed_list_cleanup;
+
+ buildsym_init ();
+ back_to = make_cleanup (really_free_pendings, NULL);
+ delayed_list_cleanup = make_cleanup (free_delayed_list, cu);
+
+ cu->list_in_scope = &file_symbols;
+
+ cu->language = pretend_language;
+ cu->language_defn = language_def (cu->language);
+
+ /* The symbol tables are set up in read_type_unit_scope. */
+ process_die (cu->dies, cu);
+
+ /* For now fudge the Go package. */
+ if (cu->language == language_go)
+ fixup_go_packaging (cu);
+
+ /* Now that we have processed all the DIEs in the CU, all the types
+ should be complete, and it should now be safe to compute all of the
+ physnames. */
+ compute_delayed_physnames (cu);
+ do_cleanups (delayed_list_cleanup);
+
+ /* TUs share symbol tables.
+ If this is the first TU to use this symtab, complete the construction
+ of it with end_symtab. Otherwise, complete the addition of this TU's
+ symbols to the existing symtab. */
+ if (per_cu->s.type_unit_group->primary_symtab == NULL)
+ {
+ symtab = end_expandable_symtab (0, objfile, SECT_OFF_TEXT (objfile));
+ per_cu->s.type_unit_group->primary_symtab = symtab;
+
+ if (symtab != NULL)
+ {
+ /* Set symtab language to language from DW_AT_language. If the
+ compilation is from a C file generated by language preprocessors,
+ do not set the language if it was already deduced by
+ start_subfile. */
+ if (!(cu->language == language_c && symtab->language != language_c))
+ symtab->language = cu->language;
+ }
+ }
+ else
{
- fprintf_unfiltered (gdb_stdlog,
- "Done expanding symtab of %s at offset 0x%x\n",
- per_cu->is_debug_types ? "TU" : "CU",
- per_cu->offset.sect_off);
+ augment_type_symtab (objfile,
+ per_cu->s.type_unit_group->primary_symtab);
+ symtab = per_cu->s.type_unit_group->primary_symtab;
}
+
+ if (dwarf2_per_objfile->using_index)
+ per_cu->v.quick->symtab = symtab;
+ else
+ {
+ struct partial_symtab *pst = per_cu->v.psymtab;
+ pst->symtab = symtab;
+ pst->readin = 1;
+ }
+
+ do_cleanups (back_to);
}
/* Process an imported unit DIE. */
@@ -6061,6 +6749,14 @@ process_imported_unit_die (struct die_in
{
struct attribute *attr;
+ /* For now we don't handle imported units in type units. */
+ if (cu->per_cu->is_debug_types)
+ {
+ error (_("Dwarf Error: DW_TAG_imported_unit is not"
+ " supported in type units [in module %s]"),
+ cu->objfile->name);
+ }
+
attr = dwarf2_attr (die, DW_AT_import, cu);
if (attr != NULL)
{
@@ -6075,7 +6771,7 @@ process_imported_unit_die (struct die_in
if (maybe_queue_comp_unit (cu, per_cu, cu->language))
load_full_comp_unit (per_cu, cu->language);
- VEC_safe_push (dwarf2_per_cu_ptr, cu->per_cu->imported_symtabs,
+ VEC_safe_push (dwarf2_per_cu_ptr, cu->per_cu->s.imported_symtabs,
per_cu);
}
}
@@ -6760,7 +7456,7 @@ read_import_statement (struct die_info *
do_cleanups (cleanups);
}
-/* Cleanup function for read_file_scope. */
+/* Cleanup function for handle_DW_AT_stmt_list. */
static void
free_cu_line_header (void *arg)
@@ -6811,17 +7507,19 @@ find_file_and_directory (struct die_info
*name = "<unknown>";
}
-/* Handle DW_AT_stmt_list for a compilation unit or type unit.
- DIE is the DW_TAG_compile_unit or DW_TAG_type_unit die for CU.
+/* Handle DW_AT_stmt_list for a compilation unit.
+ DIE is the DW_TAG_compile_unit die for CU.
COMP_DIR is the compilation directory.
WANT_LINE_INFO is non-zero if the pc/line-number mapping is needed. */
static void
handle_DW_AT_stmt_list (struct die_info *die, struct dwarf2_cu *cu,
- const char *comp_dir, int want_line_info)
+ const char *comp_dir)
{
struct attribute *attr;
+ gdb_assert (! cu->per_cu->is_debug_types);
+
attr = dwarf2_attr (die, DW_AT_stmt_list, cu);
if (attr)
{
@@ -6833,7 +7531,7 @@ handle_DW_AT_stmt_list (struct die_info
{
cu->line_header = line_header;
make_cleanup (free_cu_line_header, cu);
- dwarf_decode_lines (line_header, comp_dir, cu, NULL, want_line_info);
+ dwarf_decode_lines (line_header, comp_dir, cu, NULL, 1);
}
}
}
@@ -6879,19 +7577,12 @@ read_file_scope (struct die_info *die, s
if (cu->producer && strstr (cu->producer, "GNU Go ") != NULL)
set_cu_language (DW_LANG_Go, cu);
- /* We assume that we're processing GCC output. */
- processing_gcc_compilation = 2;
-
- processing_has_namespace_info = 0;
-
- start_symtab (name, comp_dir, lowpc);
- record_debugformat ("DWARF 2");
- record_producer (cu->producer);
+ dwarf2_start_symtab (cu, name, comp_dir, lowpc);
/* Decode line number information if present. We do this before
processing child DIEs, so that the line header table is available
for DW_AT_decl_file. */
- handle_DW_AT_stmt_list (die, cu, comp_dir, 1);
+ handle_DW_AT_stmt_list (die, cu, comp_dir);
/* Process all dies in compilation unit. */
if (die->child != NULL)
@@ -6931,81 +7622,141 @@ read_file_scope (struct die_info *die, s
do_cleanups (back_to);
}
-/* Process DW_TAG_type_unit.
- For TUs we want to skip the first top level sibling if it's not the
- actual type being defined by this TU. In this case the first top
- level sibling is there to provide context only. */
+/* TU version of handle_DW_AT_stmt_list for read_type_unit_scope.
+ Create the set of symtabs used by this TU, or if this TU is sharing
+ symtabs with another TU and the symtabs have already been created
+ then restore those symtabs in the line header.
+ We don't need the pc/line-number mapping for type units. */
static void
-read_type_unit_scope (struct die_info *die, struct dwarf2_cu *cu)
+setup_type_unit_groups (struct die_info *die, struct dwarf2_cu *cu)
{
- struct objfile *objfile = cu->objfile;
- struct cleanup *back_to = make_cleanup (null_cleanup, 0);
- CORE_ADDR lowpc;
+ struct objfile *objfile = dwarf2_per_objfile->objfile;
+ struct dwarf2_per_cu_data *per_cu = cu->per_cu;
+ struct type_unit_group *tu_group;
+ int first_time;
+ struct line_header *lh;
struct attribute *attr;
- char *name = NULL;
- char *comp_dir = NULL;
- struct die_info *child_die;
- bfd *abfd = objfile->obfd;
+ unsigned int i, line_offset;
- /* start_symtab needs a low pc, but we don't really have one.
- Do what read_file_scope would do in the absence of such info. */
- lowpc = ANOFFSET (objfile->section_offsets, SECT_OFF_TEXT (objfile));
+ gdb_assert (per_cu->is_debug_types);
- /* Find the filename. Do not use dwarf2_name here, since the filename
- is not a source language identifier. */
- attr = dwarf2_attr (die, DW_AT_name, cu);
- if (attr)
- name = DW_STRING (attr);
+ attr = dwarf2_attr (die, DW_AT_stmt_list, cu);
- attr = dwarf2_attr (die, DW_AT_comp_dir, cu);
- if (attr)
- comp_dir = DW_STRING (attr);
- else if (name != NULL && IS_ABSOLUTE_PATH (name))
+ /* If we're using .gdb_index (includes -readnow) then
+ per_cu->s.type_unit_group may not have been set up yet. */
+ if (per_cu->s.type_unit_group == NULL)
+ per_cu->s.type_unit_group = get_type_unit_group (per_cu, attr);
+ tu_group = per_cu->s.type_unit_group;
+
+ /* If we've already processed this stmt_list there's no real need to
+ do it again, we could fake it and just recreate the part we need
+ (file name,index -> symtab mapping). If data shows this optimization
+ is useful we can do it then. */
+ first_time = tu_group->primary_symtab == NULL;
+
+ /* We have to handle the case of both a missing DW_AT_stmt_list or bad
+ debug info. */
+ lh = NULL;
+ if (attr != NULL)
{
- comp_dir = ldirname (name);
- if (comp_dir != NULL)
- make_cleanup (xfree, comp_dir);
+ line_offset = DW_UNSND (attr);
+ lh = dwarf_decode_line_header (line_offset, cu);
+ }
+ if (lh == NULL)
+ {
+ if (first_time)
+ dwarf2_start_symtab (cu, "", NULL, 0);
+ else
+ {
+ gdb_assert (tu_group->symtabs == NULL);
+ restart_symtab (0);
+ }
+ /* Note: The primary symtab will get allocated at the end. */
+ return;
}
- if (name == NULL)
- name = "<unknown>";
+ cu->line_header = lh;
+ make_cleanup (free_cu_line_header, cu);
- prepare_one_comp_unit (cu, die, language_minimal);
+ if (first_time)
+ {
+ dwarf2_start_symtab (cu, "", NULL, 0);
- /* We assume that we're processing GCC output. */
- processing_gcc_compilation = 2;
+ tu_group->num_symtabs = lh->num_file_names;
+ tu_group->symtabs = XNEWVEC (struct symtab *, lh->num_file_names);
- processing_has_namespace_info = 0;
+ for (i = 0; i < lh->num_file_names; ++i)
+ {
+ char *dir = NULL;
+ struct file_entry *fe = &lh->file_names[i];
- start_symtab (name, comp_dir, lowpc);
- record_debugformat ("DWARF 2");
- record_producer (cu->producer);
+ if (fe->dir_index)
+ dir = lh->include_dirs[fe->dir_index - 1];
+ dwarf2_start_subfile (fe->name, dir, NULL);
- /* Decode line number information if present. We do this before
- processing child DIEs, so that the line header table is available
- for DW_AT_decl_file.
- We don't need the pc/line-number mapping for type units. */
- handle_DW_AT_stmt_list (die, cu, comp_dir, 0);
+ /* Note: We don't have to watch for the main subfile here, type units
+ don't have DW_AT_name. */
- /* Process the dies in the type unit. */
- if (die->child == NULL)
+ if (current_subfile->symtab == NULL)
+ {
+ /* NOTE: start_subfile will recognize when it's been passed
+ a file it has already seen. So we can't assume there's a
+ simple mapping from lh->file_names to subfiles,
+ lh->file_names may contain dups. */
+ current_subfile->symtab = allocate_symtab (current_subfile->name,
+ objfile);
+ }
+
+ fe->symtab = current_subfile->symtab;
+ tu_group->symtabs[i] = fe->symtab;
+ }
+ }
+ else
{
- dump_die_for_error (die);
- error (_("Dwarf Error: Missing children for type unit [in module %s]"),
- bfd_get_filename (abfd));
+ restart_symtab (0);
+
+ for (i = 0; i < lh->num_file_names; ++i)
+ {
+ struct file_entry *fe = &lh->file_names[i];
+
+ fe->symtab = tu_group->symtabs[i];
+ }
}
- child_die = die->child;
+ /* The main symtab is allocated last. Type units don't have DW_AT_name
+ so they don't have a "real" (so to speak) symtab anyway.
+ There is later code that will assign the main symtab to all symbols
+ that don't have one. We need to handle the case of a symbol with a
+ missing symtab (DW_AT_decl_file) anyway. */
+}
- while (child_die && child_die->tag)
- {
- process_die (child_die, cu);
+/* Process DW_TAG_type_unit.
+ For TUs we want to skip the first top level sibling if it's not the
+ actual type being defined by this TU. In this case the first top
+ level sibling is there to provide context only. */
- child_die = sibling_die (child_die);
- }
+static void
+read_type_unit_scope (struct die_info *die, struct dwarf2_cu *cu)
+{
+ struct die_info *child_die;
- do_cleanups (back_to);
+ prepare_one_comp_unit (cu, die, language_minimal);
+
+ /* Initialize (or reinitialize) the machinery for building symtabs.
+ We do this before processing child DIEs, so that the line header table
+ is available for DW_AT_decl_file. */
+ setup_type_unit_groups (die, cu);
+
+ if (die->child != NULL)
+ {
+ child_die = die->child;
+ while (child_die && child_die->tag)
+ {
+ process_die (child_die, cu);
+ child_die = sibling_die (child_die);
+ }
+ }
}
/* DWO files. */
@@ -11155,6 +11906,7 @@ abbrev_table_read_table (struct dwarf2_s
unsigned int allocated_attrs;
abbrev_table = XMALLOC (struct abbrev_table);
+ abbrev_table->offset = offset;
obstack_init (&abbrev_table->abbrev_obstack);
abbrev_table->abbrevs = obstack_alloc (&abbrev_table->abbrev_obstack,
(ABBREV_HASH_SIZE
@@ -11241,6 +11993,21 @@ abbrev_table_free (struct abbrev_table *
xfree (abbrev_table);
}
+/* Same as abbrev_table_free but as a cleanup.
+ We pass in a pointer to the pointer to the table so that we can
+ set the pointer to NULL when we're done. It also simplifies
+ build_type_unit_groups. */
+
+static void
+abbrev_table_free_cleanup (void *table_ptr)
+{
+ struct abbrev_table **abbrev_table_ptr = table_ptr;
+
+ if (*abbrev_table_ptr != NULL)
+ abbrev_table_free (*abbrev_table_ptr);
+ *abbrev_table_ptr = NULL;
+}
+
/* Read the abbrev table for CU from ABBREV_SECTION. */
static void
@@ -13600,6 +14367,23 @@ dwarf2_start_subfile (char *filename, co
xfree (fullname);
}
+/* Start a symtab for DWARF.
+ NAME, COMP_DIR, LOW_PC are passed to start_symtab. */
+
+static void
+dwarf2_start_symtab (struct dwarf2_cu *cu,
+ char *name, char *comp_dir, CORE_ADDR low_pc)
+{
+ start_symtab (name, comp_dir, low_pc);
+ record_debugformat ("DWARF 2");
+ record_producer (cu->producer);
+
+ /* We assume that we're processing GCC output. */
+ processing_gcc_compilation = 2;
+
+ processing_has_namespace_info = 0;
+}
+
static void
var_decode_location (struct attribute *attr, struct symbol *sym,
struct dwarf2_cu *cu)
@@ -15515,6 +16299,9 @@ load_full_type_unit (struct dwarf2_per_c
{
struct signatured_type *sig_type;
+ /* Caller is responsible for ensuring type_unit_groups don't get here. */
+ gdb_assert (! IS_TYPE_UNIT_GROUP (per_cu));
+
/* We have the per_cu, but we need the signatured_type.
Fortunately this is an easy translation. */
gdb_assert (per_cu->is_debug_types);
@@ -15577,7 +16364,8 @@ read_signatured_type (struct signatured_
gdb_assert (per_cu->is_debug_types);
gdb_assert (per_cu->cu == NULL);
- init_cutu_and_read_dies (per_cu, 0, 1, read_signatured_type_reader, NULL);
+ init_cutu_and_read_dies (per_cu, NULL, 0, 1,
+ read_signatured_type_reader, NULL);
}
/* Decode simple location descriptions.
@@ -17510,7 +18298,7 @@ dwarf2_per_objfile_free (struct objfile
for (ix = 0; ix < dwarf2_per_objfile->n_comp_units; ++ix)
VEC_free (dwarf2_per_cu_ptr,
- dwarf2_per_objfile->all_comp_units[ix]->imported_symtabs);
+ dwarf2_per_objfile->all_comp_units[ix]->s.imported_symtabs);
VEC_free (dwarf2_section_info_def, data->types);