This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH v3.2 5/5] DWARF-5: .debug_names index consumer
- From: Pedro Alves <palves at redhat dot com>
- To: Jan Kratochvil <jan dot kratochvil at redhat dot com>, gdb-patches at sourceware dot org
- Cc: Victor Leschuk <vleschuk at accesssoftek dot com>
- Date: Fri, 8 Dec 2017 23:58:47 +0000
- Subject: Re: [PATCH v3.2 5/5] DWARF-5: .debug_names index consumer
- Authentication-results: sourceware.org; auth=none
- References: <149790572259.20186.14601775821404892582.stgit@host1.jankratochvil.net> <149790576669.20186.17091294504362630191.stgit@host1.jankratochvil.net> <20170628212114.GA23559@host1.jankratochvil.net> <20170702142957.GA22686@host1.jankratochvil.net>
On 07/02/2017 03:29 PM, Jan Kratochvil wrote:
>
> it is not regression-free against no-index but it is regression-free against
> .gdb_index. That is because .gdb_index is not regression-free against
> no-index.
I'm seeing a few regressions compared to .gdb_index, using
--target_board=dwarf4-gdb-index :
-PASS: gdb.base/enumval.exp: p ZERO
+FAIL: gdb.base/enumval.exp: p ZERO
-PASS: gdb.base/opaque.exp: ptype on opaque struct pointer (statically)
+FAIL: gdb.base/opaque.exp: ptype on opaque struct pointer (statically)
-PASS: gdb.base/type-opaque.exp: opaque struct type resolving
+FAIL: gdb.base/type-opaque.exp: opaque struct type resolving
-PASS: gdb.base/vla-stub.exp: p *local_struct.ptr
+FAIL: gdb.base/vla-stub.exp: p *local_struct.ptr
-PASS: gdb.cp/derivation.exp: before run: ptype A2::value_type
-PASS: gdb.cp/derivation.exp: before run: whatis A2::value_type
-PASS: gdb.cp/derivation.exp: before run: p (A2::value_type) 0
-PASS: gdb.cp/derivation.exp: before run: ptype D2::value_type
-PASS: gdb.cp/derivation.exp: before run: whatis D2::value_type
-PASS: gdb.cp/derivation.exp: before run: p (D2::value_type) 0
+FAIL: gdb.cp/derivation.exp: before run: ptype A2::value_type
+FAIL: gdb.cp/derivation.exp: before run: whatis A2::value_type
+FAIL: gdb.cp/derivation.exp: before run: p (A2::value_type) 0
+FAIL: gdb.cp/derivation.exp: before run: ptype D2::value_type
+FAIL: gdb.cp/derivation.exp: before run: whatis D2::value_type
+FAIL: gdb.cp/derivation.exp: before run: p (D2::value_type) 0
etc. There are a few more. That board has:
set_board_info debug_flags "-gdwarf-4 -fdebug-types-section"
and if I remove "-fdebug-types-section", then the series
is regression-free compared to .gdb_index. Have you seen this?
>
> Some testcases needed to be updated as they were missing .debug_aranges.
> While that does not matter for no-index (as GDB builds the mapping internally
> during dwarf2_build_psymtabs_hard) and neither for .gdb_index (as GDB uses that
> internally built mapping which it stores into .gdb_index) it does matter for
> .debug_names as that simply assumes existing .debug_aranges from GCC.
>
> I tried some performance checking but the index handling speed is negligible
> compared to the CU expansion associated with it. .debug_names looked even as
> a bit faster to me than .gdb_index which rather surprised me but I did not
> investigate it more.
>
Thanks! FWIW, I found this patch quite hard to read, until
I discovered "git diff --patience" (!). The default algorithm
makes it looks like this is changing much more code than
it is. With "git diff --patience", I was finally able to
see that this is mostly new code. I wish I'd known about
that switch before...
> +/* Read the address map data from DWARF-5 .debug_aranges, and use it to
> + populate the objfile's psymtabs_addrmap. */
> +
> +static void
> +create_addrmap_from_aranges (struct objfile *objfile,
> + struct dwarf2_section_info *section)
> +{
> + bfd *abfd = objfile->obfd;
> + struct gdbarch *gdbarch = get_objfile_arch (objfile);
> + const gdb_byte *iter, *end;
> + struct addrmap *mutable_map;
> + const CORE_ADDR baseaddr (ANOFFSET (objfile->section_offsets,
> + SECT_OFF_TEXT (objfile)));
> +
> + auto_obstack temp_obstack;
> +
> + mutable_map = addrmap_create_mutable (&temp_obstack);
> +
> + std::unordered_map<sect_offset,
> + dwarf2_per_cu_data *> debug_info_offset_to_per_cu;
This doesn't compile with 5.3.1 and earlier (at least):
In file included from /usr/include/c++/5.3.1/bits/hashtable.h:35:0,
from /usr/include/c++/5.3.1/unordered_set:47,
from src/gdb/dwarf2read.c:79:
/usr/include/c++/5.3.1/bits/hashtable_policy.h: In instantiation of ‘struct std::__detail::__is_noexcept_hash<sect_offset, std::hash<sect_offset> >’:
/usr/include/c++/5.3.1/type_traits:137:12: required from ‘struct std::__and_<std::__is_fast_hash<std::hash<sect_offset> >, std::__detail::__is_noexcept_hash<sect_offset, std::hash<sect_offset> > >’
/usr/include/c++/5.3.1/type_traits:148:38: required from ‘struct std::__not_<std::__and_<std::__is_fast_hash<std::hash<sect_offset> >, std::__detail::__is_noexcept_hash<sect_offset, std::hash<sect_offset> > > >’
/usr/include/c++/5.3.1/bits/unordered_map.h:100:66: required from ‘class std::unordered_map<sect_offset, dwarf2_per_cu_data*>’
src/gdb/dwarf2read.c:3260:30: required from here
/usr/include/c++/5.3.1/bits/hashtable_policy.h:85:34: error: no match for call to ‘(const std::hash<sect_offset>) (const sect_offset&)’
noexcept(declval<const _Hash&>()(declval<const _Key&>()))>
^
In file included from /usr/include/c++/5.3.1/bits/move.h:57:0,
from /usr/include/c++/5.3.1/bits/stl_pair.h:59,
from /usr/include/c++/5.3.1/bits/stl_algobase.h:64,
from /usr/include/c++/5.3.1/bits/char_traits.h:39,
from /usr/include/c++/5.3.1/string:40,
from /home/pedro/gdb/mygit/src/gdb/common/common-utils.h:23,
from /home/pedro/gdb/mygit/src/gdb/common/common-defs.h:78,
from /home/pedro/gdb/mygit/src/gdb/defs.h:28,
from /home/pedro/gdb/mygit/src/gdb/dwarf2read.c:31:
This is due to a C++11 language defect:
http://www.open-std.org/jtc1/sc22/wg21/docs/lwg-defects.html#2148
I've fixed this by adding a gdb::hash_enum type (in a separate
preparatory patch).
> +const gdb_byte *
> +dw2_debug_names_iterator::find_vec_in_debug_names
> + (const mapped_debug_names &map, const char *name)
> +{
> + struct cleanup *back_to = make_cleanup (null_cleanup, 0);
> + int (*cmp) (const char *, const char *);
> +
> + if (current_language->la_language == language_cplus
> + || current_language->la_language == language_fortran
> + || current_language->la_language == language_d)
> + {
> + /* NAME is already canonical. Drop any qualifiers as .gdb_index does
.gdb_index -> .debug_names.
> - if (cust->includes == NULL)
> +dwarf2_per_cu_data *
> +dw2_debug_names_iterator::next ()
> +{
> + if (addr == NULL)
> return NULL;
> + bfd *const abfd (dwarf2_per_objfile->objfile->obfd);
> + unsigned int bytes_read;
> + const ULONGEST abbrev (read_unsigned_leb128 (abfd, addr, &bytes_read));
> + addr += bytes_read;
> + if (abbrev == 0)
> + return NULL;
> + const auto indexval_it (map.abbrev_map.find (abbrev));
> + if (indexval_it == map.abbrev_map.cend ())
> + {
> + complaint (&symfile_complaints,
> + _("Wrong .debug_names undefined abbrev code %s "
> + "[in module %s]"),
> + pulongest (abbrev), objfile_name (dwarf2_per_objfile->objfile));
> + return NULL;
> + }
> + const mapped_debug_names::index_val &indexval (indexval_it->second);
> + bool have_is_static (false);
> + bool is_static;
> + dwarf2_per_cu_data *per_cu (NULL);
> + for (const mapped_debug_names::index_val::attr &attr : indexval.attr_vec)
> + {
> + ULONGEST ull;
> + switch (attr.form)
> + {
> + case DW_FORM_implicit_const:
> + ull = attr.implicit_const;
> + break;
> + case DW_FORM_flag_present:
> + ull = 1;
> + break;
> + case DW_FORM_udata:
> + ull = read_unsigned_leb128 (abfd, addr, &bytes_read);
> + addr += bytes_read;
> + break;
> + default:
> + complaint (&symfile_complaints,
> + _("Unsupported .debug_names form %s [in module %s]"),
> + dwarf_form_name (attr.form),
> + objfile_name (dwarf2_per_objfile->objfile));
> + return NULL;
> + }
> + switch (attr.dw_idx)
> + {
> + case DW_IDX_compile_unit:
> + /* Don't crash on bad data. */
> + if (ull >= (dwarf2_per_objfile->n_comp_units
> + + dwarf2_per_objfile->n_type_units))
> + {
> + complaint (&symfile_complaints,
> + _(".gdb_index entry has bad CU index %s"
.gdb_index -> .debug_names. I've fixed all the cases that I noticed.
> + " [in module %s]"),
> + pulongest (ull),
> + objfile_name (dwarf2_per_objfile->objfile));
> + continue;
> + }
> + per_cu = dw2_get_cutu (ull);
> + break;
> + case DW_IDX_GNU_static:
> + if (!map.augmentation_is_gdb)
> + break;
> + have_is_static = true;
> + is_static = true;
> + break;
> + case DW_IDX_GNU_external:
> + if (!map.augmentation_is_gdb)
> + break;
> + have_is_static = true;
> + is_static = false;
> + break;
> + }
> + }
>
> - for (i = 0; cust->includes[i]; ++i)
> + /* Skip if already read in. */
> + if (per_cu->v.quick->compunit_symtab)
> + return next ();
These next() calls are recursing, which carries an implicit
assumption that this will tail call. That doesn't seem like a
good idea to me, because we wouldn't want to risk blowing the
stack if we're testing at -O0 and end up skipping a lot
of symtabs. I applied a minimal tweak using "goto again;" to
jump to the top again.
>
> @@ -22903,6 +24019,8 @@ dwarf2_free_objfile (struct objfile *objfile)
> htab_delete (dwarf2_per_objfile->line_header_hash);
>
> /* Everything else should be on the objfile obstack. */
> +
> + delete dwarf2_per_objfile->debug_names_table;
This should have been above the "everything else" comment.
But I've actually make dwarf2_per_objfile->debug_names_table
be a std::unique_ptr instead, so the delete is no longer
necessary.
> index 8e51d41..847358f 100644
> --- a/gdb/symfile.h
> +++ b/gdb/symfile.h
> @@ -606,6 +606,8 @@ struct dwarf2_debug_sections {
> struct dwarf2_section_names frame;
> struct dwarf2_section_names eh_frame;
> struct dwarf2_section_names gdb_index;
> + struct dwarf2_section_names debug_names;
> + struct dwarf2_section_names debug_aranges;
> /* This field has no meaning, but exists solely to catch changes to
> this structure which are not reflected in some instance. */
> int sentinel;
> diff --git a/gdb/testsuite/gdb.base/maint.exp b/gdb/testsuite/gdb.base/maint.exp
> index 782a21c..8e79451 100644
> --- a/gdb/testsuite/gdb.base/maint.exp
> +++ b/gdb/testsuite/gdb.base/maint.exp
> @@ -91,8 +91,11 @@ if ![runto_main] then {
>
> # If we're using .gdb_index there will be no psymtabs.
I've tweaked several places in the tests that I noticed
were still referencing .gdb_index only, both in comments
and in test messages.
I applied the same many editorial fixups that I
applied to the producer patch throughout, and the necessary
tweaks to adjust to current master, but otherwise the
algorithms etc. are unmodified from your original patch.
Thanks much, and below's what I pushed, generated with
"git format-patch --patience".
Let me know if I messed something up too badly.
I'll follow up with a few patches that adjust this
to handle the new symbol_name_match_type::WILD, since
currently as is, this only supports fully-qualified matching.
>From 927aa2e778dce440f4de5de8fc37ead1683a804e Mon Sep 17 00:00:00 2001
From: Jan Kratochvil <jan.kratochvil@redhat.com>
Date: Fri, 8 Dec 2017 22:44:12 +0000
Subject: [PATCH 6/9] DWARF-5: .debug_names index consumer
Some testcases needed to be updated as they were missing
.debug_aranges. While that does not matter for no-index (as GDB
builds the mapping internally during dwarf2_build_psymtabs_hard) and
neither for .gdb_index (as GDB uses that internally built mapping
which it stores into .gdb_index) it does matter for .debug_names as
that simply assumes existing .debug_aranges from GCC.
gdb/ChangeLog:
2017-12-08 Jan Kratochvil <jan.kratochvil@redhat.com>
Pedro Alves <palves@redhat.com>
* defs.h (elf_sym_fns_debug_names): New declaration.
* dwarf2read.c: Include "hash_enum.h".
(mapped_debug_names): New.
(struct dwarf2_per_objfile): Add debug_names, debug_aranges and
debug_names_table.
(dwarf2_elf_names): Add ".debug_names" and ".debug_aranges".
(struct dwz_file): Add debug_names.
(dwarf2_per_objfile::locate_sections): Handle debug_names and
debug_aranges.
(locate_dwz_sections): Handle debug_names.
(create_signatured_type_table_from_debug_names)
(create_addrmap_from_aranges): New.
(dwarf2_read_index): Update function comment.
(dwarf5_augmentation): Moved up.
(read_debug_names_from_section, create_cus_from_debug_names_list)
(create_cus_from_debug_names, dwarf2_read_debug_names): New.
(dwarf5_djb_hash): Moved up.
(dw2_debug_names_iterator): New.
(read_indirect_string_at_offset): New declaration.
(mapped_debug_names::namei_to_name)
(dw2_debug_names_iterator::find_vec_in_debug_names)
(dw2_debug_names_iterator::next, dw2_debug_names_lookup_symbol)
(dw2_debug_names_dump, dw2_debug_names_expand_symtabs_for_function)
(dw2_debug_names_expand_symtabs_matching, dwarf2_debug_names_functions):
New.
(dwarf2_initialize_objfile): Return also elf_sym_fns_debug_names.
(debug_names::djb_hash): Rename it to dwarf5_djb_hash.
(debug_names::build): Update djb_hash caller.
(write_debug_names): Move out and rename augmentation to
dwarf5_augmentation.
* elfread.c (elf_sym_fns_debug_names): New.
* psymtab.h (dwarf2_debug_names_functions): New declaration.
* symfile.h (struct dwarf2_debug_sections): Add debug_names and
debug_aranges.
* xcoffread.c (dwarf2_xcoff_names): Add debug_names and debug_aranges.
gdb/testsuite/ChangeLog:
2017-12-08 Jan Kratochvil <jan.kratochvil@redhat.com>
Pedro Alves <palves@redhat.com>
* gdb.base/maint.exp (check for .gdb_index): Check also for
.debug_names.
* gdb.dlang/watch-loc.c (.debug_aranges): New.
* gdb.dwarf2/dw2-case-insensitive-debug.S: Likewise.
* gdb.dwarf2/gdb-index.exp (check if index present, .gdb_index used)
(.gdb_index used after symbol reloading): Support also .debug_names.
* gdb.mi/dw2-ref-missing-frame-func.c (.debug_aranges): New.
---
gdb/ChangeLog | 39 +
gdb/testsuite/ChangeLog | 11 +
gdb/defs.h | 1 +
gdb/dwarf2read.c | 1126 +++++++++++++++++++-
gdb/elfread.c | 17 +
gdb/psymtab.h | 1 +
gdb/symfile.h | 2 +
gdb/testsuite/gdb.base/maint.exp | 9 +-
gdb/testsuite/gdb.dlang/watch-loc.c | 20 +
.../gdb.dwarf2/dw2-case-insensitive-debug.S | 18 +
gdb/testsuite/gdb.dwarf2/gdb-index.exp | 19 +-
gdb/testsuite/gdb.mi/dw2-ref-missing-frame-func.c | 21 +
gdb/xcoffread.c | 2 +
13 files changed, 1257 insertions(+), 29 deletions(-)
diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index 09cafd0..ee2f1c4 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,42 @@
+2017-12-08 Jan Kratochvil <jan.kratochvil@redhat.com>
+ Pedro Alves <palves@redhat.com>
+
+ * defs.h (elf_sym_fns_debug_names): New declaration.
+ * dwarf2read.c: Include "hash_enum.h".
+ (mapped_debug_names): New.
+ (struct dwarf2_per_objfile): Add debug_names, debug_aranges and
+ debug_names_table.
+ (dwarf2_elf_names): Add ".debug_names" and ".debug_aranges".
+ (struct dwz_file): Add debug_names.
+ (dwarf2_per_objfile::locate_sections): Handle debug_names and
+ debug_aranges.
+ (locate_dwz_sections): Handle debug_names.
+ (create_signatured_type_table_from_debug_names)
+ (create_addrmap_from_aranges): New.
+ (dwarf2_read_index): Update function comment.
+ (dwarf5_augmentation): Moved up.
+ (read_debug_names_from_section, create_cus_from_debug_names_list)
+ (create_cus_from_debug_names, dwarf2_read_debug_names): New.
+ (dwarf5_djb_hash): Moved up.
+ (dw2_debug_names_iterator): New.
+ (read_indirect_string_at_offset): New declaration.
+ (mapped_debug_names::namei_to_name)
+ (dw2_debug_names_iterator::find_vec_in_debug_names)
+ (dw2_debug_names_iterator::next, dw2_debug_names_lookup_symbol)
+ (dw2_debug_names_dump, dw2_debug_names_expand_symtabs_for_function)
+ (dw2_debug_names_expand_symtabs_matching, dwarf2_debug_names_functions):
+ New.
+ (dwarf2_initialize_objfile): Return also elf_sym_fns_debug_names.
+ (debug_names::djb_hash): Rename it to dwarf5_djb_hash.
+ (debug_names::build): Update djb_hash caller.
+ (write_debug_names): Move out and rename augmentation to
+ dwarf5_augmentation.
+ * elfread.c (elf_sym_fns_debug_names): New.
+ * psymtab.h (dwarf2_debug_names_functions): New declaration.
+ * symfile.h (struct dwarf2_debug_sections): Add debug_names and
+ debug_aranges.
+ * xcoffread.c (dwarf2_xcoff_names): Add debug_names and debug_aranges.
+
2017-12-08 Pedro Alves <palves@redhat.com>
* common/hash_enum.h: New file.
diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index 53a2ca4..48bad29 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,14 @@
+2017-12-08 Jan Kratochvil <jan.kratochvil@redhat.com>
+ Pedro Alves <palves@redhat.com>
+
+ * gdb.base/maint.exp (check for .gdb_index): Check also for
+ .debug_names.
+ * gdb.dlang/watch-loc.c (.debug_aranges): New.
+ * gdb.dwarf2/dw2-case-insensitive-debug.S: Likewise.
+ * gdb.dwarf2/gdb-index.exp (check if index present, .gdb_index used)
+ (.gdb_index used after symbol reloading): Support also .debug_names.
+ * gdb.mi/dw2-ref-missing-frame-func.c (.debug_aranges): New.
+
2017-12-08 Yao Qi <yao.qi@linaro.org>
* gdb.arch/aarch64-tagged-pointer.c (main): Update.
diff --git a/gdb/defs.h b/gdb/defs.h
index a873a55..c2e145e 100644
--- a/gdb/defs.h
+++ b/gdb/defs.h
@@ -683,6 +683,7 @@ extern void initialize_inferiors (void);
extern const struct sym_fns elf_sym_fns_lazy_psyms;
extern const struct sym_fns elf_sym_fns_gdb_index;
+extern const struct sym_fns elf_sym_fns_debug_names;
/* * Special block numbers */
diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 80d4857..ca2b04d 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -74,6 +74,7 @@
#include "common/gdb_optional.h"
#include "common/underlying.h"
#include "common/byte-vector.h"
+#include "common/hash_enum.h"
#include "filename-seen-cache.h"
#include "producer.h"
#include <fcntl.h>
@@ -282,6 +283,44 @@ struct mapped_index
find_name_components_bounds (const lookup_name_info &ln_no_params) const;
};
+/* A description of the mapped .debug_names.
+ Uninitialized map has CU_COUNT 0. */
+struct mapped_debug_names
+{
+ bfd_endian dwarf5_byte_order;
+ bool dwarf5_is_dwarf64;
+ bool augmentation_is_gdb;
+ uint8_t offset_size;
+ uint32_t cu_count = 0;
+ uint32_t tu_count, bucket_count, name_count;
+ const gdb_byte *cu_table_reordered, *tu_table_reordered;
+ const uint32_t *bucket_table_reordered, *hash_table_reordered;
+ const gdb_byte *name_table_string_offs_reordered;
+ const gdb_byte *name_table_entry_offs_reordered;
+ const gdb_byte *entry_pool;
+
+ struct index_val
+ {
+ ULONGEST dwarf_tag;
+ struct attr
+ {
+ /* Attribute name DW_IDX_*. */
+ ULONGEST dw_idx;
+
+ /* Attribute form DW_FORM_*. */
+ ULONGEST form;
+
+ /* Value if FORM is DW_FORM_implicit_const. */
+ LONGEST implicit_const;
+ };
+ std::vector<attr> attr_vec;
+ };
+
+ std::unordered_map<ULONGEST, index_val> abbrev_map;
+
+ const char *namei_to_name (uint32_t namei) const;
+};
+
typedef struct dwarf2_per_cu_data *dwarf2_per_cu_ptr;
DEF_VEC_P (dwarf2_per_cu_ptr);
@@ -334,6 +373,8 @@ public:
dwarf2_section_info frame {};
dwarf2_section_info eh_frame {};
dwarf2_section_info gdb_index {};
+ dwarf2_section_info debug_names {};
+ dwarf2_section_info debug_aranges {};
VEC (dwarf2_section_info_def) *types = NULL;
@@ -399,6 +440,9 @@ public:
/* The mapped index, or NULL if .gdb_index is missing or not being used. */
mapped_index *index_table = NULL;
+ /* The mapped index, or NULL if .debug_names is missing or not being used. */
+ std::unique_ptr<mapped_debug_names> debug_names_table;
+
/* When using index_table, this keeps track of all quick_file_names entries.
TUs typically share line table entries with a CU, so we maintain a
separate table of all line table entries to support the sharing.
@@ -453,6 +497,8 @@ static const struct dwarf2_debug_sections dwarf2_elf_names =
{ ".debug_frame", ".zdebug_frame" },
{ ".eh_frame", NULL },
{ ".gdb_index", ".zgdb_index" },
+ { ".debug_names", ".zdebug_names" },
+ { ".debug_aranges", ".zdebug_aranges" },
23
};
@@ -1119,6 +1165,7 @@ struct dwz_file
struct dwarf2_section_info line;
struct dwarf2_section_info macro;
struct dwarf2_section_info gdb_index;
+ struct dwarf2_section_info debug_names;
/* The dwz's BFD. */
bfd *dwz_bfd;
@@ -1726,6 +1773,9 @@ static const char *read_indirect_line_string (bfd *, const gdb_byte *,
const struct comp_unit_head *,
unsigned int *);
+static const char *read_indirect_string_at_offset (bfd *abfd,
+ LONGEST str_offset);
+
static const char *read_indirect_string_from_dwz (struct dwz_file *, LONGEST);
static LONGEST read_signed_leb128 (bfd *, const gdb_byte *, unsigned int *);
@@ -2547,6 +2597,16 @@ dwarf2_per_objfile::locate_sections (bfd *abfd, asection *sectp,
this->gdb_index.s.section = sectp;
this->gdb_index.size = bfd_get_section_size (sectp);
}
+ else if (section_is_p (sectp->name, &names.debug_names))
+ {
+ this->debug_names.s.section = sectp;
+ this->debug_names.size = bfd_get_section_size (sectp);
+ }
+ else if (section_is_p (sectp->name, &names.debug_aranges))
+ {
+ this->debug_aranges.s.section = sectp;
+ this->debug_aranges.size = bfd_get_section_size (sectp);
+ }
if ((bfd_get_section_flags (abfd, sectp) & (SEC_LOAD | SEC_ALLOC))
&& bfd_section_vma (abfd, sectp) == 0)
@@ -2743,6 +2803,11 @@ locate_dwz_sections (bfd *abfd, asection *sectp, void *arg)
dwz_file->gdb_index.s.section = sectp;
dwz_file->gdb_index.size = bfd_get_section_size (sectp);
}
+ else if (section_is_p (sectp->name, &dwarf2_elf_names.debug_names))
+ {
+ dwz_file->debug_names.s.section = sectp;
+ dwz_file->debug_names.size = bfd_get_section_size (sectp);
+ }
}
/* Open the separate '.dwz' debug file, if needed. Return NULL if
@@ -3192,6 +3257,65 @@ create_signatured_type_table_from_index (struct objfile *objfile,
dwarf2_per_objfile->signatured_types = sig_types_hash;
}
+/* Create the signatured type hash table from .debug_names. */
+
+static void
+create_signatured_type_table_from_debug_names
+ (struct objfile *objfile,
+ const mapped_debug_names &map,
+ struct dwarf2_section_info *section,
+ struct dwarf2_section_info *abbrev_section)
+{
+ dwarf2_read_section (objfile, section);
+ dwarf2_read_section (objfile, abbrev_section);
+
+ dwarf2_per_objfile->n_type_units
+ = dwarf2_per_objfile->n_allocated_type_units
+ = map.tu_count;
+ dwarf2_per_objfile->all_type_units
+ = XNEWVEC (struct signatured_type *, dwarf2_per_objfile->n_type_units);
+
+ htab_t sig_types_hash = allocate_signatured_type_table (objfile);
+
+ for (uint32_t i = 0; i < map.tu_count; ++i)
+ {
+ struct signatured_type *sig_type;
+ ULONGEST signature;
+ void **slot;
+ cu_offset type_offset_in_tu;
+
+ sect_offset sect_off
+ = (sect_offset) (extract_unsigned_integer
+ (map.tu_table_reordered + i * map.offset_size,
+ map.offset_size,
+ map.dwarf5_byte_order));
+
+ comp_unit_head cu_header;
+ read_and_check_comp_unit_head (&cu_header, section, abbrev_section,
+ section->buffer + to_underlying (sect_off),
+ rcuh_kind::TYPE);
+
+ sig_type = OBSTACK_ZALLOC (&objfile->objfile_obstack,
+ struct signatured_type);
+ sig_type->signature = cu_header.signature;
+ sig_type->type_offset_in_tu = cu_header.type_cu_offset_in_tu;
+ sig_type->per_cu.is_debug_types = 1;
+ sig_type->per_cu.section = section;
+ sig_type->per_cu.sect_off = sect_off;
+ sig_type->per_cu.objfile = objfile;
+ sig_type->per_cu.v.quick
+ = OBSTACK_ZALLOC (&objfile->objfile_obstack,
+ struct dwarf2_per_cu_quick_data);
+
+ slot = htab_find_slot (sig_types_hash, sig_type, INSERT);
+ *slot = sig_type;
+
+ dwarf2_per_objfile->all_type_units[i] = sig_type;
+ }
+
+ dwarf2_per_objfile->signatured_types = sig_types_hash;
+}
+
/* Read the address map data from the mapped index, and use it to
populate the objfile's psymtabs_addrmap. */
@@ -3247,6 +3371,165 @@ create_addrmap_from_index (struct objfile *objfile, struct mapped_index *index)
&objfile->objfile_obstack);
}
+/* Read the address map data from DWARF-5 .debug_aranges, and use it to
+ populate the objfile's psymtabs_addrmap. */
+
+static void
+create_addrmap_from_aranges (struct objfile *objfile,
+ struct dwarf2_section_info *section)
+{
+ bfd *abfd = objfile->obfd;
+ struct gdbarch *gdbarch = get_objfile_arch (objfile);
+ const CORE_ADDR baseaddr = ANOFFSET (objfile->section_offsets,
+ SECT_OFF_TEXT (objfile));
+
+ auto_obstack temp_obstack;
+ addrmap *mutable_map = addrmap_create_mutable (&temp_obstack);
+
+ std::unordered_map<sect_offset,
+ dwarf2_per_cu_data *,
+ gdb::hash_enum<sect_offset>>
+ debug_info_offset_to_per_cu;
+ for (int cui = 0; cui < dwarf2_per_objfile->n_comp_units; ++cui)
+ {
+ dwarf2_per_cu_data *per_cu = dw2_get_cutu (cui);
+ const auto insertpair
+ = debug_info_offset_to_per_cu.emplace (per_cu->sect_off, per_cu);
+ if (!insertpair.second)
+ {
+ warning (_("Section .debug_aranges in %s has duplicate "
+ "debug_info_offset %u, ignoring .debug_aranges."),
+ objfile_name (objfile), to_underlying (per_cu->sect_off));
+ return;
+ }
+ }
+
+ dwarf2_read_section (objfile, section);
+
+ const bfd_endian dwarf5_byte_order = gdbarch_byte_order (gdbarch);
+
+ const gdb_byte *addr = section->buffer;
+
+ while (addr < section->buffer + section->size)
+ {
+ const gdb_byte *const entry_addr = addr;
+ unsigned int bytes_read;
+
+ const LONGEST entry_length = read_initial_length (abfd, addr,
+ &bytes_read);
+ addr += bytes_read;
+
+ const gdb_byte *const entry_end = addr + entry_length;
+ const bool dwarf5_is_dwarf64 = bytes_read != 4;
+ const uint8_t offset_size = dwarf5_is_dwarf64 ? 8 : 4;
+ if (addr + entry_length > section->buffer + section->size)
+ {
+ warning (_("Section .debug_aranges in %s entry at offset %zu "
+ "length %s exceeds section length %s, "
+ "ignoring .debug_aranges."),
+ objfile_name (objfile), entry_addr - section->buffer,
+ plongest (bytes_read + entry_length),
+ pulongest (section->size));
+ return;
+ }
+
+ /* The version number. */
+ const uint16_t version = read_2_bytes (abfd, addr);
+ addr += 2;
+ if (version != 2)
+ {
+ warning (_("Section .debug_aranges in %s entry at offset %zu "
+ "has unsupported version %d, ignoring .debug_aranges."),
+ objfile_name (objfile), entry_addr - section->buffer,
+ version);
+ return;
+ }
+
+ const uint64_t debug_info_offset
+ = extract_unsigned_integer (addr, offset_size, dwarf5_byte_order);
+ addr += offset_size;
+ const auto per_cu_it
+ = debug_info_offset_to_per_cu.find (sect_offset (debug_info_offset));
+ if (per_cu_it == debug_info_offset_to_per_cu.cend ())
+ {
+ warning (_("Section .debug_aranges in %s entry at offset %zu "
+ "debug_info_offset %s does not exists, "
+ "ignoring .debug_aranges."),
+ objfile_name (objfile), entry_addr - section->buffer,
+ pulongest (debug_info_offset));
+ return;
+ }
+ dwarf2_per_cu_data *const per_cu = per_cu_it->second;
+
+ const uint8_t address_size = *addr++;
+ if (address_size < 1 || address_size > 8)
+ {
+ warning (_("Section .debug_aranges in %s entry at offset %zu "
+ "address_size %u is invalid, ignoring .debug_aranges."),
+ objfile_name (objfile), entry_addr - section->buffer,
+ address_size);
+ return;
+ }
+
+ const uint8_t segment_selector_size = *addr++;
+ if (segment_selector_size != 0)
+ {
+ warning (_("Section .debug_aranges in %s entry at offset %zu "
+ "segment_selector_size %u is not supported, "
+ "ignoring .debug_aranges."),
+ objfile_name (objfile), entry_addr - section->buffer,
+ segment_selector_size);
+ return;
+ }
+
+ /* Must pad to an alignment boundary that is twice the address
+ size. It is undocumented by the DWARF standard but GCC does
+ use it. */
+ for (size_t padding = ((-(addr - section->buffer))
+ & (2 * address_size - 1));
+ padding > 0; padding--)
+ if (*addr++ != 0)
+ {
+ warning (_("Section .debug_aranges in %s entry at offset %zu "
+ "padding is not zero, ignoring .debug_aranges."),
+ objfile_name (objfile), entry_addr - section->buffer);
+ return;
+ }
+
+ for (;;)
+ {
+ if (addr + 2 * address_size > entry_end)
+ {
+ warning (_("Section .debug_aranges in %s entry at offset %zu "
+ "address list is not properly terminated, "
+ "ignoring .debug_aranges."),
+ objfile_name (objfile), entry_addr - section->buffer);
+ return;
+ }
+ ULONGEST start = extract_unsigned_integer (addr, address_size,
+ dwarf5_byte_order);
+ addr += address_size;
+ ULONGEST length = extract_unsigned_integer (addr, address_size,
+ dwarf5_byte_order);
+ addr += address_size;
+ if (start == 0 && length == 0)
+ break;
+ if (start == 0 && !dwarf2_per_objfile->has_section_at_zero)
+ {
+ /* Symbol was eliminated due to a COMDAT group. */
+ continue;
+ }
+ ULONGEST end = start + length;
+ start = gdbarch_adjust_dwarf2_addr (gdbarch, start + baseaddr);
+ end = gdbarch_adjust_dwarf2_addr (gdbarch, end + baseaddr);
+ addrmap_set_empty (mutable_map, start, end - 1, per_cu);
+ }
+ }
+
+ objfile->psymtabs_addrmap = addrmap_create_fixed (mutable_map,
+ &objfile->objfile_obstack);
+}
+
/* The hash function for strings in the mapped index. This is the same as
SYMBOL_HASH_NEXT, but we keep a separate copy to maintain control over the
implementation. This is necessary because the hash function is tied to the
@@ -3457,8 +3740,7 @@ to use the section anyway."),
return 1;
}
-
-/* Read the index file. If everything went ok, initialize the "quick"
+/* Read .gdb_index. If everything went ok, initialize the "quick"
elements of all the CUs and return 1. Otherwise, return 0. */
static int
@@ -5330,6 +5612,827 @@ const struct quick_symbol_functions dwarf2_gdb_index_functions =
dw2_map_symbol_filenames
};
+/* DWARF-5 debug_names reader. */
+
+/* DWARF-5 augmentation string for GDB's DW_IDX_GNU_* extension. */
+static const gdb_byte dwarf5_augmentation[] = { 'G', 'D', 'B', 0 };
+
+/* A helper function that reads the .debug_names section in SECTION
+ and fills in MAP. FILENAME is the name of the file containing the
+ section; it is used for error reporting.
+
+ Returns true if all went well, false otherwise. */
+
+static bool
+read_debug_names_from_section (struct objfile *objfile,
+ const char *filename,
+ struct dwarf2_section_info *section,
+ mapped_debug_names &map)
+{
+ if (dwarf2_section_empty_p (section))
+ return false;
+
+ /* Older elfutils strip versions could keep the section in the main
+ executable while splitting it for the separate debug info file. */
+ if ((get_section_flags (section) & SEC_HAS_CONTENTS) == 0)
+ return false;
+
+ dwarf2_read_section (objfile, section);
+
+ map.dwarf5_byte_order = gdbarch_byte_order (get_objfile_arch (objfile));
+
+ const gdb_byte *addr = section->buffer;
+
+ bfd *const abfd = get_section_bfd_owner (section);
+
+ unsigned int bytes_read;
+ LONGEST length = read_initial_length (abfd, addr, &bytes_read);
+ addr += bytes_read;
+
+ map.dwarf5_is_dwarf64 = bytes_read != 4;
+ map.offset_size = map.dwarf5_is_dwarf64 ? 8 : 4;
+ if (bytes_read + length != section->size)
+ {
+ /* There may be multiple per-CU indices. */
+ warning (_("Section .debug_names in %s length %s does not match "
+ "section length %s, ignoring .debug_names."),
+ filename, plongest (bytes_read + length),
+ pulongest (section->size));
+ return false;
+ }
+
+ /* The version number. */
+ uint16_t version = read_2_bytes (abfd, addr);
+ addr += 2;
+ if (version != 5)
+ {
+ warning (_("Section .debug_names in %s has unsupported version %d, "
+ "ignoring .debug_names."),
+ filename, version);
+ return false;
+ }
+
+ /* Padding. */
+ uint16_t padding = read_2_bytes (abfd, addr);
+ addr += 2;
+ if (padding != 0)
+ {
+ warning (_("Section .debug_names in %s has unsupported padding %d, "
+ "ignoring .debug_names."),
+ filename, padding);
+ return false;
+ }
+
+ /* comp_unit_count - The number of CUs in the CU list. */
+ map.cu_count = read_4_bytes (abfd, addr);
+ addr += 4;
+
+ /* local_type_unit_count - The number of TUs in the local TU
+ list. */
+ map.tu_count = read_4_bytes (abfd, addr);
+ addr += 4;
+
+ /* foreign_type_unit_count - The number of TUs in the foreign TU
+ list. */
+ uint32_t foreign_tu_count = read_4_bytes (abfd, addr);
+ addr += 4;
+ if (foreign_tu_count != 0)
+ {
+ warning (_("Section .debug_names in %s has unsupported %lu foreign TUs, "
+ "ignoring .debug_names."),
+ filename, static_cast<unsigned long> (foreign_tu_count));
+ return false;
+ }
+
+ /* bucket_count - The number of hash buckets in the hash lookup
+ table. */
+ map.bucket_count = read_4_bytes (abfd, addr);
+ addr += 4;
+
+ /* name_count - The number of unique names in the index. */
+ map.name_count = read_4_bytes (abfd, addr);
+ addr += 4;
+
+ /* abbrev_table_size - The size in bytes of the abbreviations
+ table. */
+ uint32_t abbrev_table_size = read_4_bytes (abfd, addr);
+ addr += 4;
+
+ /* augmentation_string_size - The size in bytes of the augmentation
+ string. This value is rounded up to a multiple of 4. */
+ uint32_t augmentation_string_size = read_4_bytes (abfd, addr);
+ addr += 4;
+ map.augmentation_is_gdb = ((augmentation_string_size
+ == sizeof (dwarf5_augmentation))
+ && memcmp (addr, dwarf5_augmentation,
+ sizeof (dwarf5_augmentation)) == 0);
+ augmentation_string_size += (-augmentation_string_size) & 3;
+ addr += augmentation_string_size;
+
+ /* List of CUs */
+ map.cu_table_reordered = addr;
+ addr += map.cu_count * map.offset_size;
+
+ /* List of Local TUs */
+ map.tu_table_reordered = addr;
+ addr += map.tu_count * map.offset_size;
+
+ /* Hash Lookup Table */
+ map.bucket_table_reordered = reinterpret_cast<const uint32_t *> (addr);
+ addr += map.bucket_count * 4;
+ map.hash_table_reordered = reinterpret_cast<const uint32_t *> (addr);
+ addr += map.name_count * 4;
+
+ /* Name Table */
+ map.name_table_string_offs_reordered = addr;
+ addr += map.name_count * map.offset_size;
+ map.name_table_entry_offs_reordered = addr;
+ addr += map.name_count * map.offset_size;
+
+ const gdb_byte *abbrev_table_start = addr;
+ for (;;)
+ {
+ unsigned int bytes_read;
+ const ULONGEST index_num = read_unsigned_leb128 (abfd, addr, &bytes_read);
+ addr += bytes_read;
+ if (index_num == 0)
+ break;
+
+ const auto insertpair
+ = map.abbrev_map.emplace (index_num, mapped_debug_names::index_val ());
+ if (!insertpair.second)
+ {
+ warning (_("Section .debug_names in %s has duplicate index %s, "
+ "ignoring .debug_names."),
+ filename, pulongest (index_num));
+ return false;
+ }
+ mapped_debug_names::index_val &indexval = insertpair.first->second;
+ indexval.dwarf_tag = read_unsigned_leb128 (abfd, addr, &bytes_read);
+ addr += bytes_read;
+
+ for (;;)
+ {
+ mapped_debug_names::index_val::attr attr;
+ attr.dw_idx = read_unsigned_leb128 (abfd, addr, &bytes_read);
+ addr += bytes_read;
+ attr.form = read_unsigned_leb128 (abfd, addr, &bytes_read);
+ addr += bytes_read;
+ if (attr.form == DW_FORM_implicit_const)
+ {
+ attr.implicit_const = read_signed_leb128 (abfd, addr,
+ &bytes_read);
+ addr += bytes_read;
+ }
+ if (attr.dw_idx == 0 && attr.form == 0)
+ break;
+ indexval.attr_vec.push_back (std::move (attr));
+ }
+ }
+ if (addr != abbrev_table_start + abbrev_table_size)
+ {
+ warning (_("Section .debug_names in %s has abbreviation_table "
+ "of size %zu vs. written as %u, ignoring .debug_names."),
+ filename, addr - abbrev_table_start, abbrev_table_size);
+ return false;
+ }
+ map.entry_pool = addr;
+
+ return true;
+}
+
+/* A helper for create_cus_from_debug_names that handles the MAP's CU
+ list. */
+
+static void
+create_cus_from_debug_names_list (struct objfile *objfile,
+ const mapped_debug_names &map,
+ dwarf2_section_info §ion,
+ bool is_dwz, int base_offset)
+{
+ sect_offset sect_off_prev;
+ for (uint32_t i = 0; i <= map.cu_count; ++i)
+ {
+ sect_offset sect_off_next;
+ if (i < map.cu_count)
+ {
+ sect_off_next
+ = (sect_offset) (extract_unsigned_integer
+ (map.cu_table_reordered + i * map.offset_size,
+ map.offset_size,
+ map.dwarf5_byte_order));
+ }
+ else
+ sect_off_next = (sect_offset) section.size;
+ if (i >= 1)
+ {
+ const ULONGEST length = sect_off_next - sect_off_prev;
+ dwarf2_per_objfile->all_comp_units[base_offset + (i - 1)]
+ = create_cu_from_index_list (objfile, §ion, is_dwz,
+ sect_off_prev, length);
+ }
+ sect_off_prev = sect_off_next;
+ }
+}
+
+/* Read the CU list from the mapped index, and use it to create all
+ the CU objects for this objfile. */
+
+static void
+create_cus_from_debug_names (struct objfile *objfile,
+ const mapped_debug_names &map,
+ const mapped_debug_names &dwz_map)
+{
+
+ dwarf2_per_objfile->n_comp_units = map.cu_count + dwz_map.cu_count;
+ dwarf2_per_objfile->all_comp_units
+ = XOBNEWVEC (&objfile->objfile_obstack, struct dwarf2_per_cu_data *,
+ dwarf2_per_objfile->n_comp_units);
+
+ create_cus_from_debug_names_list (objfile, map, dwarf2_per_objfile->info,
+ false /* is_dwz */,
+ 0 /* base_offset */);
+
+ if (dwz_map.cu_count == 0)
+ return;
+
+ dwz_file *dwz = dwarf2_get_dwz_file ();
+ create_cus_from_debug_names_list (objfile, dwz_map, dwz->info,
+ true /* is_dwz */,
+ map.cu_count /* base_offset */);
+}
+
+/* Read .debug_names. If everything went ok, initialize the "quick"
+ elements of all the CUs and return true. Otherwise, return false. */
+
+static bool
+dwarf2_read_debug_names (struct objfile *objfile)
+{
+ mapped_debug_names local_map, dwz_map;
+
+ if (!read_debug_names_from_section (objfile, objfile_name (objfile),
+ &dwarf2_per_objfile->debug_names,
+ local_map))
+ return false;
+
+ /* Don't use the index if it's empty. */
+ if (local_map.name_count == 0)
+ return false;
+
+ /* If there is a .dwz file, read it so we can get its CU list as
+ well. */
+ dwz_file *dwz = dwarf2_get_dwz_file ();
+ if (dwz != NULL)
+ {
+ if (!read_debug_names_from_section (objfile,
+ bfd_get_filename (dwz->dwz_bfd),
+ &dwz->debug_names, dwz_map))
+ {
+ warning (_("could not read '.debug_names' section from %s; skipping"),
+ bfd_get_filename (dwz->dwz_bfd));
+ return false;
+ }
+ }
+
+ create_cus_from_debug_names (objfile, local_map, dwz_map);
+
+ if (local_map.tu_count != 0)
+ {
+ /* We can only handle a single .debug_types when we have an
+ index. */
+ if (VEC_length (dwarf2_section_info_def, dwarf2_per_objfile->types) != 1)
+ return false;
+
+ dwarf2_section_info *section = VEC_index (dwarf2_section_info_def,
+ dwarf2_per_objfile->types, 0);
+
+ create_signatured_type_table_from_debug_names
+ (objfile, local_map, section, &dwarf2_per_objfile->abbrev);
+ }
+
+ create_addrmap_from_aranges (objfile, &dwarf2_per_objfile->debug_aranges);
+
+ dwarf2_per_objfile->debug_names_table.reset (new mapped_debug_names);
+ *dwarf2_per_objfile->debug_names_table = std::move (local_map);
+ dwarf2_per_objfile->using_index = 1;
+ dwarf2_per_objfile->quick_file_names_table =
+ create_quick_file_names_table (dwarf2_per_objfile->n_comp_units);
+
+ return true;
+}
+
+/* Symbol name hashing function as specified by DWARF-5. */
+
+static uint32_t
+dwarf5_djb_hash (const char *str_)
+{
+ const unsigned char *str = (const unsigned char *) str_;
+
+ /* Note: tolower here ignores UTF-8, which isn't fully compliant.
+ See http://dwarfstd.org/ShowIssue.php?issue=161027.1. */
+
+ uint32_t hash = 5381;
+ while (int c = *str++)
+ hash = hash * 33 + tolower (c);
+ return hash;
+}
+
+/* Type used to manage iterating over all CUs looking for a symbol for
+ .debug_names. */
+
+class dw2_debug_names_iterator
+{
+public:
+ /* If WANT_SPECIFIC_BLOCK is true, only look for symbols in block
+ BLOCK_INDEX. Otherwise BLOCK_INDEX is ignored. */
+ dw2_debug_names_iterator (const mapped_debug_names &map,
+ bool want_specific_block,
+ block_enum block_index, domain_enum domain,
+ const char *name)
+ : m_map (map), m_want_specific_block (want_specific_block),
+ m_block_index (block_index), m_domain (domain),
+ m_addr (find_vec_in_debug_names (map, name))
+ {}
+
+ dw2_debug_names_iterator (const mapped_debug_names &map,
+ search_domain search, uint32_t namei)
+ : m_map (map),
+ m_search (search),
+ m_addr (find_vec_in_debug_names (map, namei))
+ {}
+
+ /* Return the next matching CU or NULL if there are no more. */
+ dwarf2_per_cu_data *next ();
+
+private:
+ static const gdb_byte *find_vec_in_debug_names (const mapped_debug_names &map,
+ const char *name);
+ static const gdb_byte *find_vec_in_debug_names (const mapped_debug_names &map,
+ uint32_t namei);
+
+ /* The internalized form of .debug_names. */
+ const mapped_debug_names &m_map;
+
+ /* If true, only look for symbols that match BLOCK_INDEX. */
+ const bool m_want_specific_block = false;
+
+ /* One of GLOBAL_BLOCK or STATIC_BLOCK.
+ Unused if !WANT_SPECIFIC_BLOCK - FIRST_LOCAL_BLOCK is an invalid
+ value. */
+ const block_enum m_block_index = FIRST_LOCAL_BLOCK;
+
+ /* The kind of symbol we're looking for. */
+ const domain_enum m_domain = UNDEF_DOMAIN;
+ const search_domain m_search = ALL_DOMAIN;
+
+ /* The list of CUs from the index entry of the symbol, or NULL if
+ not found. */
+ const gdb_byte *m_addr;
+};
+
+const char *
+mapped_debug_names::namei_to_name (uint32_t namei) const
+{
+ const ULONGEST namei_string_offs
+ = extract_unsigned_integer ((name_table_string_offs_reordered
+ + namei * offset_size),
+ offset_size,
+ dwarf5_byte_order);
+ return read_indirect_string_at_offset
+ (dwarf2_per_objfile->objfile->obfd, namei_string_offs);
+}
+
+/* Find a slot in .debug_names for the object named NAME. If NAME is
+ found, return pointer to its pool data. If NAME cannot be found,
+ return NULL. */
+
+const gdb_byte *
+dw2_debug_names_iterator::find_vec_in_debug_names
+ (const mapped_debug_names &map, const char *name)
+{
+ int (*cmp) (const char *, const char *);
+
+ if (current_language->la_language == language_cplus
+ || current_language->la_language == language_fortran
+ || current_language->la_language == language_d)
+ {
+ /* NAME is already canonical. Drop any qualifiers as
+ .debug_names does not contain any. */
+
+ if (strchr (name, '(') != NULL)
+ {
+ gdb::unique_xmalloc_ptr<char> without_params
+ = cp_remove_params (name);
+
+ if (without_params != NULL)
+ {
+ name = without_params.get();
+ }
+ }
+ }
+
+ cmp = (case_sensitivity == case_sensitive_on ? strcmp : strcasecmp);
+
+ const uint32_t full_hash = dwarf5_djb_hash (name);
+ uint32_t namei
+ = extract_unsigned_integer (reinterpret_cast<const gdb_byte *>
+ (map.bucket_table_reordered
+ + (full_hash % map.bucket_count)), 4,
+ map.dwarf5_byte_order);
+ if (namei == 0)
+ return NULL;
+ --namei;
+ if (namei >= map.name_count)
+ {
+ complaint (&symfile_complaints,
+ _("Wrong .debug_names with name index %u but name_count=%u "
+ "[in module %s]"),
+ namei, map.name_count,
+ objfile_name (dwarf2_per_objfile->objfile));
+ return NULL;
+ }
+
+ for (;;)
+ {
+ const uint32_t namei_full_hash
+ = extract_unsigned_integer (reinterpret_cast<const gdb_byte *>
+ (map.hash_table_reordered + namei), 4,
+ map.dwarf5_byte_order);
+ if (full_hash % map.bucket_count != namei_full_hash % map.bucket_count)
+ return NULL;
+
+ if (full_hash == namei_full_hash)
+ {
+ const char *const namei_string = map.namei_to_name (namei);
+
+#if 0 /* An expensive sanity check. */
+ if (namei_full_hash != dwarf5_djb_hash (namei_string))
+ {
+ complaint (&symfile_complaints,
+ _("Wrong .debug_names hash for string at index %u "
+ "[in module %s]"),
+ namei, objfile_name (dwarf2_per_objfile->objfile));
+ return NULL;
+ }
+#endif
+
+ if (cmp (namei_string, name) == 0)
+ {
+ const ULONGEST namei_entry_offs
+ = extract_unsigned_integer ((map.name_table_entry_offs_reordered
+ + namei * map.offset_size),
+ map.offset_size, map.dwarf5_byte_order);
+ return map.entry_pool + namei_entry_offs;
+ }
+ }
+
+ ++namei;
+ if (namei >= map.name_count)
+ return NULL;
+ }
+}
+
+const gdb_byte *
+dw2_debug_names_iterator::find_vec_in_debug_names
+ (const mapped_debug_names &map, uint32_t namei)
+{
+ if (namei >= map.name_count)
+ {
+ complaint (&symfile_complaints,
+ _("Wrong .debug_names with name index %u but name_count=%u "
+ "[in module %s]"),
+ namei, map.name_count,
+ objfile_name (dwarf2_per_objfile->objfile));
+ return NULL;
+ }
+
+ const ULONGEST namei_entry_offs
+ = extract_unsigned_integer ((map.name_table_entry_offs_reordered
+ + namei * map.offset_size),
+ map.offset_size, map.dwarf5_byte_order);
+ return map.entry_pool + namei_entry_offs;
+}
+
+/* See dw2_debug_names_iterator. */
+
+dwarf2_per_cu_data *
+dw2_debug_names_iterator::next ()
+{
+ if (m_addr == NULL)
+ return NULL;
+
+ bfd *const abfd = dwarf2_per_objfile->objfile->obfd;
+
+ again:
+
+ unsigned int bytes_read;
+ const ULONGEST abbrev = read_unsigned_leb128 (abfd, m_addr, &bytes_read);
+ m_addr += bytes_read;
+ if (abbrev == 0)
+ return NULL;
+
+ const auto indexval_it = m_map.abbrev_map.find (abbrev);
+ if (indexval_it == m_map.abbrev_map.cend ())
+ {
+ complaint (&symfile_complaints,
+ _("Wrong .debug_names undefined abbrev code %s "
+ "[in module %s]"),
+ pulongest (abbrev), objfile_name (dwarf2_per_objfile->objfile));
+ return NULL;
+ }
+ const mapped_debug_names::index_val &indexval = indexval_it->second;
+ bool have_is_static = false;
+ bool is_static;
+ dwarf2_per_cu_data *per_cu = NULL;
+ for (const mapped_debug_names::index_val::attr &attr : indexval.attr_vec)
+ {
+ ULONGEST ull;
+ switch (attr.form)
+ {
+ case DW_FORM_implicit_const:
+ ull = attr.implicit_const;
+ break;
+ case DW_FORM_flag_present:
+ ull = 1;
+ break;
+ case DW_FORM_udata:
+ ull = read_unsigned_leb128 (abfd, m_addr, &bytes_read);
+ m_addr += bytes_read;
+ break;
+ default:
+ complaint (&symfile_complaints,
+ _("Unsupported .debug_names form %s [in module %s]"),
+ dwarf_form_name (attr.form),
+ objfile_name (dwarf2_per_objfile->objfile));
+ return NULL;
+ }
+ switch (attr.dw_idx)
+ {
+ case DW_IDX_compile_unit:
+ /* Don't crash on bad data. */
+ if (ull >= (dwarf2_per_objfile->n_comp_units
+ + dwarf2_per_objfile->n_type_units))
+ {
+ complaint (&symfile_complaints,
+ _(".debug_names entry has bad CU index %s"
+ " [in module %s]"),
+ pulongest (ull),
+ objfile_name (dwarf2_per_objfile->objfile));
+ continue;
+ }
+ per_cu = dw2_get_cutu (ull);
+ break;
+ case DW_IDX_GNU_internal:
+ if (!m_map.augmentation_is_gdb)
+ break;
+ have_is_static = true;
+ is_static = true;
+ break;
+ case DW_IDX_GNU_external:
+ if (!m_map.augmentation_is_gdb)
+ break;
+ have_is_static = true;
+ is_static = false;
+ break;
+ }
+ }
+
+ /* Skip if already read in. */
+ if (per_cu->v.quick->compunit_symtab)
+ goto again;
+
+ /* Check static vs global. */
+ if (have_is_static)
+ {
+ const bool want_static = m_block_index != GLOBAL_BLOCK;
+ if (m_want_specific_block && want_static != is_static)
+ goto again;
+ }
+
+ /* Match dw2_symtab_iter_next, symbol_kind
+ and debug_names::psymbol_tag. */
+ switch (m_domain)
+ {
+ case VAR_DOMAIN:
+ switch (indexval.dwarf_tag)
+ {
+ case DW_TAG_variable:
+ case DW_TAG_subprogram:
+ /* Some types are also in VAR_DOMAIN. */
+ case DW_TAG_typedef:
+ case DW_TAG_structure_type:
+ break;
+ default:
+ goto again;
+ }
+ break;
+ case STRUCT_DOMAIN:
+ switch (indexval.dwarf_tag)
+ {
+ case DW_TAG_typedef:
+ case DW_TAG_structure_type:
+ break;
+ default:
+ goto again;
+ }
+ break;
+ case LABEL_DOMAIN:
+ switch (indexval.dwarf_tag)
+ {
+ case 0:
+ case DW_TAG_variable:
+ break;
+ default:
+ goto again;
+ }
+ break;
+ default:
+ break;
+ }
+
+ /* Match dw2_expand_symtabs_matching, symbol_kind and
+ debug_names::psymbol_tag. */
+ switch (m_search)
+ {
+ case VARIABLES_DOMAIN:
+ switch (indexval.dwarf_tag)
+ {
+ case DW_TAG_variable:
+ break;
+ default:
+ goto again;
+ }
+ break;
+ case FUNCTIONS_DOMAIN:
+ switch (indexval.dwarf_tag)
+ {
+ case DW_TAG_subprogram:
+ break;
+ default:
+ goto again;
+ }
+ break;
+ case TYPES_DOMAIN:
+ switch (indexval.dwarf_tag)
+ {
+ case DW_TAG_typedef:
+ case DW_TAG_structure_type:
+ break;
+ default:
+ goto again;
+ }
+ break;
+ default:
+ break;
+ }
+
+ return per_cu;
+}
+
+static struct compunit_symtab *
+dw2_debug_names_lookup_symbol (struct objfile *objfile, int block_index_int,
+ const char *name, domain_enum domain)
+{
+ const block_enum block_index = static_cast<block_enum> (block_index_int);
+ dw2_setup (objfile);
+
+ const auto &mapp = dwarf2_per_objfile->debug_names_table;
+ if (!mapp)
+ {
+ /* index is NULL if OBJF_READNOW. */
+ return NULL;
+ }
+ const auto &map = *mapp;
+
+ dw2_debug_names_iterator iter (map, true /* want_specific_block */,
+ block_index, domain, name);
+
+ struct compunit_symtab *stab_best = NULL;
+ struct dwarf2_per_cu_data *per_cu;
+ while ((per_cu = iter.next ()) != NULL)
+ {
+ struct symbol *sym, *with_opaque = NULL;
+ struct compunit_symtab *stab = dw2_instantiate_symtab (per_cu);
+ const struct blockvector *bv = COMPUNIT_BLOCKVECTOR (stab);
+ struct block *block = BLOCKVECTOR_BLOCK (bv, block_index);
+
+ sym = block_find_symbol (block, name, domain,
+ block_find_non_opaque_type_preferred,
+ &with_opaque);
+
+ /* Some caution must be observed with overloaded functions and
+ methods, since the index will not contain any overload
+ information (but NAME might contain it). */
+
+ if (sym != NULL
+ && strcmp_iw (SYMBOL_SEARCH_NAME (sym), name) == 0)
+ return stab;
+ if (with_opaque != NULL
+ && strcmp_iw (SYMBOL_SEARCH_NAME (with_opaque), name) == 0)
+ stab_best = stab;
+
+ /* Keep looking through other CUs. */
+ }
+
+ return stab_best;
+}
+
+/* This dumps minimal information about .debug_names. It is called
+ via "mt print objfiles". The gdb.dwarf2/gdb-index.exp testcase
+ uses this to verify that .debug_names has been loaded. */
+
+static void
+dw2_debug_names_dump (struct objfile *objfile)
+{
+ dw2_setup (objfile);
+ gdb_assert (dwarf2_per_objfile->using_index);
+ printf_filtered (".debug_names:");
+ if (dwarf2_per_objfile->debug_names_table)
+ printf_filtered (" exists\n");
+ else
+ printf_filtered (" faked for \"readnow\"\n");
+ printf_filtered ("\n");
+}
+
+static void
+dw2_debug_names_expand_symtabs_for_function (struct objfile *objfile,
+ const char *func_name)
+{
+ dw2_setup (objfile);
+
+ /* dwarf2_per_objfile->debug_names_table is NULL if OBJF_READNOW. */
+ if (dwarf2_per_objfile->debug_names_table)
+ {
+ const mapped_debug_names &map = *dwarf2_per_objfile->debug_names_table;
+
+ /* Note: It doesn't matter what we pass for block_index here. */
+ dw2_debug_names_iterator iter (map, false /* want_specific_block */,
+ GLOBAL_BLOCK, VAR_DOMAIN, func_name);
+
+ struct dwarf2_per_cu_data *per_cu;
+ while ((per_cu = iter.next ()) != NULL)
+ dw2_instantiate_symtab (per_cu);
+ }
+}
+
+static void
+dw2_debug_names_expand_symtabs_matching
+ (struct objfile *objfile,
+ gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
+ const lookup_name_info &lookup_name,
+ gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
+ gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
+ enum search_domain kind)
+{
+ dw2_setup (objfile);
+
+ /* debug_names_table is NULL if OBJF_READNOW. */
+ if (!dwarf2_per_objfile->debug_names_table)
+ return;
+
+ dw_expand_symtabs_matching_file_matcher (file_matcher);
+
+ const mapped_debug_names &map = *dwarf2_per_objfile->debug_names_table;
+
+ for (uint32_t namei = 0; namei < map.name_count; ++namei)
+ {
+ QUIT;
+
+ const char *const namei_string = map.namei_to_name (namei);
+ if (symbol_matcher != NULL && !symbol_matcher (namei_string))
+ continue;
+
+ /* The name was matched, now expand corresponding CUs that were
+ marked. */
+ dw2_debug_names_iterator iter (map, kind, namei);
+
+ struct dwarf2_per_cu_data *per_cu;
+ while ((per_cu = iter.next ()) != NULL)
+ dw2_expand_symtabs_matching_one (per_cu, file_matcher,
+ expansion_notify);
+ }
+}
+
+const struct quick_symbol_functions dwarf2_debug_names_functions =
+{
+ dw2_has_symbols,
+ dw2_find_last_source_symtab,
+ dw2_forget_cached_source_info,
+ dw2_map_symtabs_matching_filename,
+ dw2_debug_names_lookup_symbol,
+ dw2_print_stats,
+ dw2_debug_names_dump,
+ dw2_relocate,
+ dw2_debug_names_expand_symtabs_for_function,
+ dw2_expand_all_symtabs,
+ dw2_expand_symtabs_with_fullname,
+ dw2_map_matching_symbols,
+ dw2_debug_names_expand_symtabs_matching,
+ dw2_find_pc_sect_compunit_symtab,
+ NULL,
+ dw2_map_symbol_filenames
+};
+
/* Initialize for reading DWARF for this objfile. Return 0 if this
file will use psymtabs, or 1 if using the GNU index. */
@@ -5365,6 +6468,9 @@ dwarf2_initialize_objfile (struct objfile *objfile)
return elf_sym_fns_gdb_index;
}
+ if (dwarf2_read_debug_names (objfile))
+ return elf_sym_fns_debug_names;
+
if (dwarf2_read_index (objfile))
return elf_sym_fns_gdb_index;
@@ -24867,22 +25973,6 @@ recursively_write_psymbols (struct objfile *objfile,
1);
}
-/* Symbol name hashing function as specified by DWARF-5. */
-
-static uint32_t
-dwarf5_djb_hash (const char *str_)
-{
- const unsigned char *str = (const unsigned char *) str_;
-
- /* Note: tolower here ignores UTF-8, which isn't fully compliant.
- See http://dwarfstd.org/ShowIssue.php?issue=161027.1. */
-
- uint32_t hash = 5381;
- while (int c = *str++)
- hash = hash * 33 + tolower (c);
- return hash;
-}
-
/* DWARF-5 .debug_names builder. */
class debug_names
{
diff --git a/gdb/elfread.c b/gdb/elfread.c
index 904696f..31288a9 100644
--- a/gdb/elfread.c
+++ b/gdb/elfread.c
@@ -1403,6 +1403,23 @@ const struct sym_fns elf_sym_fns_gdb_index =
&dwarf2_gdb_index_functions
};
+/* The same as elf_sym_fns, but not registered and uses the
+ DWARF-specific .debug_names index rather than psymtab. */
+const struct sym_fns elf_sym_fns_debug_names =
+{
+ elf_new_init, /* init anything gbl to entire symab */
+ elf_symfile_init, /* read initial info, setup for sym_red() */
+ elf_symfile_read, /* read a symbol file into symtab */
+ NULL, /* sym_read_psymbols */
+ elf_symfile_finish, /* finished with file, cleanup */
+ default_symfile_offsets, /* Translate ext. to int. relocatin */
+ elf_symfile_segments, /* Get segment information from a file. */
+ NULL,
+ default_symfile_relocate, /* Relocate a debug section. */
+ &elf_probe_fns, /* sym_probe_fns */
+ &dwarf2_debug_names_functions
+};
+
/* STT_GNU_IFUNC resolver vector to be installed to gnu_ifunc_fns_p. */
static const struct gnu_ifunc_fns elf_gnu_ifunc_fns =
diff --git a/gdb/psymtab.h b/gdb/psymtab.h
index f0c9ae7..17ceb22 100644
--- a/gdb/psymtab.h
+++ b/gdb/psymtab.h
@@ -33,6 +33,7 @@ extern struct bcache *psymbol_bcache_get_bcache (struct psymbol_bcache *);
extern const struct quick_symbol_functions psym_functions;
extern const struct quick_symbol_functions dwarf2_gdb_index_functions;
+extern const struct quick_symbol_functions dwarf2_debug_names_functions;
/* Ensure that the partial symbols for OBJFILE have been loaded. If
VERBOSE is non-zero, then this will print a message when symbols
diff --git a/gdb/symfile.h b/gdb/symfile.h
index 194dccf..e903c60 100644
--- a/gdb/symfile.h
+++ b/gdb/symfile.h
@@ -582,6 +582,8 @@ struct dwarf2_debug_sections {
struct dwarf2_section_names frame;
struct dwarf2_section_names eh_frame;
struct dwarf2_section_names gdb_index;
+ struct dwarf2_section_names debug_names;
+ struct dwarf2_section_names debug_aranges;
/* This field has no meaning, but exists solely to catch changes to
this structure which are not reflected in some instance. */
int sentinel;
diff --git a/gdb/testsuite/gdb.base/maint.exp b/gdb/testsuite/gdb.base/maint.exp
index 782a21c..8d1a4d0 100644
--- a/gdb/testsuite/gdb.base/maint.exp
+++ b/gdb/testsuite/gdb.base/maint.exp
@@ -89,10 +89,13 @@ if ![runto_main] then {
perror "tests suppressed"
}
-# If we're using .gdb_index there will be no psymtabs.
+# If we're using .gdb_index or .debug_names there will be no psymtabs.
set have_gdb_index 0
-gdb_test_multiple "maint info sections .gdb_index" "check for .gdb_index" {
- -re ": .gdb_index.*$gdb_prompt $" {
+gdb_test_multiple "maint info sections .gdb_index .debug_names" "check for .gdb_index" {
+ -re ": \\.gdb_index .*\r\n$gdb_prompt $" {
+ set have_gdb_index 1
+ }
+ -re ": \\.debug_names .*\r\n$gdb_prompt $" {
set have_gdb_index 1
}
-re ".*$gdb_prompt $" {
diff --git a/gdb/testsuite/gdb.dlang/watch-loc.c b/gdb/testsuite/gdb.dlang/watch-loc.c
index 0ffc377..3b993db 100644
--- a/gdb/testsuite/gdb.dlang/watch-loc.c
+++ b/gdb/testsuite/gdb.dlang/watch-loc.c
@@ -34,3 +34,23 @@ main (void)
return _Dmain ();
}
+/* The .debug_names-based index support depends on .debug_aranges
+ generated by GCC. (.gdb_index includes a gdb-generated map
+ instead.) */
+asm (
+" .pushsection .debug_aranges,\"\",@progbits \n"
+" .4byte .Laranges_end - .Laranges_start \n" // Length of Address Ranges Info
+".Laranges_start: \n"
+" .2byte 0x2 \n" // DWARF Version
+" .4byte 0 \n" // .Ldebug_info0 - Offset of Compilation Unit Info
+" .byte 4 \n" // Size of Address
+" .byte 0 \n" // Size of Segment Descriptor
+" .2byte 0 \n" // Pad to 16 byte boundary
+" .2byte 0 \n"
+" .4byte _Dmain \n" // Address
+" .4byte 0x1000 \n" // Length
+" .4byte 0 \n"
+" .4byte 0 \n"
+".Laranges_end: \n"
+" .popsection \n"
+);
diff --git a/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive-debug.S b/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive-debug.S
index 3bbd725..ad19f81 100644
--- a/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive-debug.S
+++ b/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive-debug.S
@@ -57,6 +57,24 @@
.byte 0 /* End of children of CU */
.Lcu1_end:
+ /* The .debug_names-based index support depends on
+ .debug_aranges generated by GCC. (.gdb_index includes a
+ gdb-generated map instead.) */
+ .section .debug_aranges,"",@progbits
+ .4byte .Laranges_end - .Laranges_start // Length of Address Ranges Info
+.Laranges_start:
+ .2byte 0x2 // DWARF Version
+ .4byte 0 // .Ldebug_info0 - Offset of Compilation Unit Info
+ .byte PTRBITS / 8 // Size of Address
+ .byte 0 // Size of Segment Descriptor
+ .2byte 0 // Pad to 16 byte boundary
+ .2byte 0
+ PTRBYTE cu_text_start // Address
+ PTRBYTE 0x1000 // cu_text_end - cu_text_start // Length
+ PTRBYTE 0
+ PTRBYTE 0
+.Laranges_end:
+
/* Abbrev table */
.section .debug_abbrev
.Labbrev1_begin:
diff --git a/gdb/testsuite/gdb.dwarf2/gdb-index.exp b/gdb/testsuite/gdb.dwarf2/gdb-index.exp
index c925b1e..c7ddbbf 100644
--- a/gdb/testsuite/gdb.dwarf2/gdb-index.exp
+++ b/gdb/testsuite/gdb.dwarf2/gdb-index.exp
@@ -58,7 +58,7 @@ proc add_gdb_index { program } {
return ${program_with_index}
}
-# Build a copy of the program with .gdb_index.
+# Build a copy of the program with an index (.gdb_index/.debug_names).
# But only if the toolchain didn't already create one: gdb doesn't support
# building an index from a program already using one.
@@ -67,6 +67,9 @@ gdb_test_multiple "mt print objfiles ${testfile}" $test {
-re "gdb_index.*${gdb_prompt} $" {
set binfile_with_index $binfile
}
+ -re "debug_names.*${gdb_prompt} $" {
+ set binfile_with_index $binfile
+ }
-re "Psymtabs.*${gdb_prompt} $" {
set binfile_with_index [add_gdb_index $binfile]
if { ${binfile_with_index} == "" } {
@@ -75,16 +78,16 @@ gdb_test_multiple "mt print objfiles ${testfile}" $test {
}
}
-# Ok, we have a copy of $binfile with .gdb_index.
+# Ok, we have a copy of $binfile with an index.
# Restart gdb and verify the index was used.
clean_restart ${binfile_with_index}
gdb_test "mt print objfiles ${testfile}" \
- "gdb_index.*" \
- ".gdb_index used"
+ "(gdb_index|debug_names).*" \
+ "index used"
-# Make gdb re-read symbols and see if .gdb_index still gets used.
-# symtab/15885
+# Make gdb re-read symbols and see if .gdb_index/.debug_names still
+# gets used. symtab/15885
# There is gdb_touch_execfile, but it doesn't handle remote hosts.
# Is touch portable enough?
@@ -98,5 +101,5 @@ if ![runto_main] {
return -1
}
gdb_test "mt print objfiles ${testfile}" \
- "gdb_index.*" \
- ".gdb_index used after symbol reloading"
+ "(gdb_index|debug_names).*" \
+ "index used after symbol reloading"
diff --git a/gdb/testsuite/gdb.mi/dw2-ref-missing-frame-func.c b/gdb/testsuite/gdb.mi/dw2-ref-missing-frame-func.c
index 0c2a153..785cf43 100644
--- a/gdb/testsuite/gdb.mi/dw2-ref-missing-frame-func.c
+++ b/gdb/testsuite/gdb.mi/dw2-ref-missing-frame-func.c
@@ -52,3 +52,24 @@ asm ("func_loopfb_end:");
asm (".globl cu_text_end");
asm ("cu_text_end:");
+
+/* The .debug_names-based index support depends on .debug_aranges
+ generated by GCC. (.gdb_index includes a gdb-generated map
+ instead.) */
+asm (
+" .pushsection .debug_aranges,\"\",@progbits \n"
+" .4byte .Laranges_end - .Laranges_start \n" // Length of Address Ranges Info
+".Laranges_start: \n"
+" .2byte 0x2 \n" // DWARF Version
+" .4byte 0 \n" // .Ldebug_info0 - Offset of Compilation Unit Info
+" .byte 4 \n" // Size of Address
+" .byte 0 \n" // Size of Segment Descriptor
+" .2byte 0 \n" // Pad to 16 byte boundary
+" .2byte 0 \n"
+" .4byte cu_text_start \n" // Address
+" .4byte cu_text_end - cu_text_start \n" // Length
+" .4byte 0 \n"
+" .4byte 0 \n"
+".Laranges_end: \n"
+" .popsection \n"
+);
diff --git a/gdb/xcoffread.c b/gdb/xcoffread.c
index b1c634c..142845b 100644
--- a/gdb/xcoffread.c
+++ b/gdb/xcoffread.c
@@ -174,6 +174,8 @@ static const struct dwarf2_debug_sections dwarf2_xcoff_names = {
{ ".dwframe", NULL },
{ NULL, NULL }, /* eh_frame */
{ NULL, NULL }, /* gdb_index */
+ { NULL, NULL }, /* debug_names */
+ { NULL, NULL }, /* debug_aranges */
23
};
--
2.5.5