This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
[intercu] Update DW_AT_specification handling
- From: Daniel Jacobowitz <drow at false dot org>
- To: gdb-patches at sources dot redhat dot com
- Date: Sun, 22 Feb 2004 01:16:01 -0500
- Subject: [intercu] Update DW_AT_specification handling
As promised earlier. By storing the full name on the partial DIE obstack, I
increase peak memory usage slightly for C++ during psymtab building (for
monotone peak goes up about 8MB out of 190MB). This could certainly be
ameliorated by a little creative timing, since the full names that I store
on the partial_die_obstack are immediately copied into the psymbol obstack;
I could save that copy in the partial_die_info. Will do that soon.
This patch also corrects comment formatting/wording (thanks for the
proofreading, Eli!); corrects a thinko about DW_TAG_enumeration_type
(we can't pre-handle them since we need to recurse to their children,
but we can pre-handle all their children); and removes a use-after-free bug
in the previous namespace handling caused by add_partial_structure being
cleverer than I expected it to be.
Committed to the intercu branch.
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer
2004-02-21 Daniel Jacobowitz <drow@mvista.com>
* dwarf2read.c (struct partial_die_info): Add full_name.
(scan_partial_symbols, add_partial_symbol, add_partial_structure)
(add_partial_enumeration): Change namespace argument to
have_namespace_info flag. Move DW_TAG_namespace handling to
add_partial_symbol. Update comments.
(pdi_needs_namespace): Remove namespace argument. Handle
DW_TAG_namespace.
(partial_determine_prefix, partial_determine_prefix_aux): Remove.
(dwarf2_build_psymtabs_hard): Update call to scan_partial_symbols.
(partial_die_full_name): New function.
(load_partial_die): Correct comment formatting. Don't handle
DW_TAG_enumeration_type.
(load_partial_dies): Correct comment formatting.
Index: dwarf2read.c
===================================================================
RCS file: /cvs/src/src/gdb/dwarf2read.c,v
retrieving revision 1.135.2.15
diff -u -p -r1.135.2.15 dwarf2read.c
--- dwarf2read.c 22 Feb 2004 01:34:13 -0000 1.135.2.15
+++ dwarf2read.c 22 Feb 2004 06:05:29 -0000
@@ -404,6 +404,7 @@ struct partial_die_info
unsigned int has_specification : 1;
unsigned int has_pc_info : 1;
char *name;
+ char *full_name;
struct dwarf_block *locdesc;
CORE_ADDR lowpc;
CORE_ADDR highpc;
@@ -647,26 +648,24 @@ static void dwarf2_build_psymtabs_hard (
static void scan_partial_symbols (struct partial_die_info *,
CORE_ADDR *, CORE_ADDR *,
- struct dwarf2_cu *,
- const char *namespace);
+ struct dwarf2_cu *, int);
static void add_partial_symbol (struct partial_die_info *, struct dwarf2_cu *,
- const char *namespace);
+ int have_namespace_info);
-static int pdi_needs_namespace (enum dwarf_tag tag, const char *namespace);
+static int pdi_needs_namespace (enum dwarf_tag tag);
static void add_partial_namespace (struct partial_die_info *pdi,
CORE_ADDR *lowpc, CORE_ADDR *highpc,
- struct dwarf2_cu *cu,
- const char *namespace);
+ struct dwarf2_cu *cu);
static void add_partial_structure (struct partial_die_info *struct_pdi,
struct dwarf2_cu *cu,
- const char *namespace);
+ int have_namespace_info);
static void add_partial_enumeration (struct partial_die_info *enum_pdi,
struct dwarf2_cu *cu,
- const char *namespace);
+ int have_namespace_info);
static char *locate_pdi_sibling (struct partial_die_info *orig_pdi,
char *info_ptr,
@@ -790,10 +789,6 @@ static char *determine_prefix (struct di
static char *determine_prefix_aux (struct die_info *die, struct dwarf2_cu *);
-static char *partial_determine_prefix (struct partial_die_info *die, struct dwarf2_cu *);
-
-static char *partial_determine_prefix_aux (struct partial_die_info *die, struct dwarf2_cu *);
-
static char *typename_concat (const char *prefix, const char *suffix);
static char *class_name (struct die_info *die, struct dwarf2_cu *);
@@ -1370,7 +1365,7 @@ dwarf2_build_psymtabs_hard (struct objfi
first_die = load_partial_dies (abfd, info_ptr, &cu);
- scan_partial_symbols (first_die, &lowpc, &highpc, &cu, NULL);
+ scan_partial_symbols (first_die, &lowpc, &highpc, &cu, 0);
cu.partial_dies = NULL;
obstack_free (&cu.partial_die_obstack, NULL);
@@ -1420,7 +1415,7 @@ dwarf2_build_psymtabs_hard (struct objfi
static void
scan_partial_symbols (struct partial_die_info *first_die, CORE_ADDR *lowpc,
CORE_ADDR *highpc, struct dwarf2_cu *cu,
- const char *namespace)
+ int have_namespace_info)
{
struct objfile *objfile = cu->objfile;
bfd *abfd = objfile->obfd;
@@ -1458,7 +1453,7 @@ scan_partial_symbols (struct partial_die
}
if (!pdi_p->is_declaration)
{
- add_partial_symbol (pdi_p, cu, namespace);
+ add_partial_symbol (pdi_p, cu, have_namespace_info);
}
}
break;
@@ -1467,89 +1462,112 @@ scan_partial_symbols (struct partial_die
case DW_TAG_union_type:
if (!pdi_p->is_declaration)
{
- add_partial_symbol (pdi_p, cu, namespace);
+ add_partial_symbol (pdi_p, cu, have_namespace_info);
}
break;
case DW_TAG_class_type:
case DW_TAG_structure_type:
if (!pdi_p->is_declaration)
{
- add_partial_structure (pdi_p, cu, namespace);
+ add_partial_structure (pdi_p, cu, have_namespace_info);
}
break;
case DW_TAG_enumeration_type:
if (!pdi_p->is_declaration)
- add_partial_enumeration (pdi_p, cu, namespace);
+ add_partial_enumeration (pdi_p, cu, have_namespace_info);
break;
case DW_TAG_base_type:
case DW_TAG_subrange_type:
/* File scope base type definitions are added to the partial
symbol table. */
- add_partial_symbol (pdi_p, cu, namespace);
+ add_partial_symbol (pdi_p, cu, have_namespace_info);
break;
case DW_TAG_namespace:
/* We've hit a DW_TAG_namespace entry, so we know this
file has been compiled using a compiler that
generates them; update NAMESPACE to reflect that. */
- if (namespace == NULL)
- namespace = "";
- add_partial_namespace (pdi_p, lowpc, highpc, cu, namespace);
+ have_namespace_info = 1;
+ add_partial_namespace (pdi_p, lowpc, highpc, cu);
break;
default:
break;
}
}
- /* FIXME unnecessary now */
- if (pdi_p->tag == 0)
- break;
-
- /* If the die has a sibling, skip to the sibling, unless another
- function has already updated info_ptr for us. */
-
- /* NOTE: carlton/2003-06-16: This is a bit hackish, but whether
- or not we want to update this depends on enough stuff (not
- only pdi_p->tag but also whether or not pdi_p->name is NULL) that
- this seems like the easiest way to handle the issue. */
+ /* If the die has a sibling, skip to the sibling. */
pdi_p = pdi_p->die_sibling;
- // if (pdi_p)
- // fprintf_unfiltered (gdb_stderr, "scan: Advancing to DIE %x\n", pdi_p->offset);
}
}
+static char *
+partial_die_full_name (struct partial_die_info *pdi, int allocate,
+ struct dwarf2_cu *cu)
+{
+ char *full_name, *parent_name;
+ struct partial_die_info *real_parent;
+
+ /* Note: this code could be micro-optimized. We could avoid redoing
+ the hash table lookup, which we've probably already done once,
+ and we might be able to use real_pdi->full_name if there is a
+ specification. */
+
+ if (pdi->has_specification)
+ real_parent = find_partial_die (pdi->spec_offset, cu)->die_parent;
+ else
+ real_parent = pdi->die_parent;
+
+ if (real_parent == NULL)
+ return NULL;
+
+ parent_name = real_parent->full_name;
+ if (parent_name == NULL)
+ parent_name = real_parent->name;
+
+ if (allocate)
+ full_name = obconcat (&cu->partial_die_obstack,
+ parent_name, "::", pdi->name);
+ else
+ full_name = concat (parent_name, "::", pdi->name, NULL);
+
+ return full_name;
+}
+
static void
add_partial_symbol (struct partial_die_info *pdi,
- struct dwarf2_cu *cu, const char *namespace)
+ struct dwarf2_cu *cu, int have_namespace_info)
{
struct objfile *objfile = cu->objfile;
CORE_ADDR addr = 0;
- char *actual_name = pdi->name;
+ char *actual_name;
const char *my_prefix;
const struct partial_symbol *psym = NULL;
CORE_ADDR baseaddr;
+ int free_actual_name = 0;
baseaddr = ANOFFSET (objfile->section_offsets, SECT_OFF_TEXT (objfile));
- /* If we're not in the global namespace and if the namespace name
- isn't encoded in a mangled actual_name, add it. */
-
- if (pdi->spec_offset
- && pdi_needs_namespace (pdi->tag, "foo"))
- my_prefix = partial_determine_prefix (pdi, cu);
- else
- my_prefix = namespace;
-
- if (pdi_needs_namespace (pdi->tag, my_prefix))
+ if (pdi_needs_namespace (pdi->tag) && pdi->full_name == NULL)
{
- actual_name = alloca (strlen (pdi->name) + 2 + strlen (my_prefix) + 1);
- strcpy (actual_name, my_prefix);
- strcat (actual_name, "::");
- strcat (actual_name, pdi->name);
+ if (pdi->tag == DW_TAG_namespace
+ || pdi->tag == DW_TAG_class_type
+ || pdi->tag == DW_TAG_structure_type
+ || pdi->tag == DW_TAG_union_type)
+ {
+ pdi->full_name = partial_die_full_name (pdi, 1, cu);
+ actual_name = pdi->full_name;
+ }
+ else
+ {
+ actual_name = partial_die_full_name (pdi, 0, cu);
+ if (actual_name)
+ free_actual_name = 1;
+ }
+ if (actual_name == NULL)
+ actual_name = pdi->name;
}
-
- if (my_prefix)
- free ((char *) my_prefix);
+ else
+ actual_name = pdi->name;
switch (pdi->tag)
{
@@ -1623,6 +1641,12 @@ add_partial_symbol (struct partial_die_i
&objfile->static_psymbols,
0, (CORE_ADDR) 0, cu->language, objfile);
break;
+ case DW_TAG_namespace:
+ add_psymbol_to_list (actual_name, strlen (actual_name),
+ VAR_DOMAIN, LOC_TYPEDEF,
+ &objfile->global_psymbols,
+ 0, (CORE_ADDR) 0, cu->language, objfile);
+ break;
case DW_TAG_class_type:
case DW_TAG_structure_type:
case DW_TAG_union_type:
@@ -1668,25 +1692,26 @@ add_partial_symbol (struct partial_die_i
mangled name to begin with. */
if (cu->language == language_cplus
- && namespace == NULL
+ && have_namespace_info == 0
&& psym != NULL
&& SYMBOL_CPLUS_DEMANGLED_NAME (psym) != NULL)
cp_check_possible_namespace_symbols (SYMBOL_CPLUS_DEMANGLED_NAME (psym),
objfile);
+
+ if (free_actual_name)
+ xfree (actual_name);
}
-/* Determine whether a die of type TAG living in the C++ namespace
- NAMESPACE needs to have the name of the namespace prepended to the
+/* Determine whether a die of type TAG living in a C++ class or
+ namespace needs to have the name of the scope prepended to the
name listed in the die. */
static int
-pdi_needs_namespace (enum dwarf_tag tag, const char *namespace)
+pdi_needs_namespace (enum dwarf_tag tag)
{
- if (namespace == NULL || namespace[0] == '\0')
- return 0;
-
switch (tag)
{
+ case DW_TAG_namespace:
case DW_TAG_typedef:
case DW_TAG_class_type:
case DW_TAG_structure_type:
@@ -1706,37 +1731,23 @@ pdi_needs_namespace (enum dwarf_tag tag,
static void
add_partial_namespace (struct partial_die_info *pdi,
CORE_ADDR *lowpc, CORE_ADDR *highpc,
- struct dwarf2_cu *cu, const char *namespace)
+ struct dwarf2_cu *cu)
{
struct objfile *objfile = cu->objfile;
- const char *new_name = pdi->name;
char *full_name;
/* Calculate the full name of the namespace that we just entered. */
- if (new_name == NULL)
- new_name = "(anonymous namespace)";
- full_name = alloca (strlen (namespace) + 2 + strlen (new_name) + 1);
- strcpy (full_name, namespace);
- if (*namespace != '\0')
- strcat (full_name, "::");
- strcat (full_name, new_name);
-
- /* FIXME: carlton/2003-10-07: We can't just replace this by a call
- to add_partial_symbol, because we don't have a way to pass in the
- full name to that function; that might be a flaw in
- add_partial_symbol's interface. */
-
- add_psymbol_to_list (full_name, strlen (full_name),
- VAR_DOMAIN, LOC_TYPEDEF,
- &objfile->global_psymbols,
- 0, 0, cu->language, objfile);
+ if (pdi->name == NULL)
+ pdi->name = "(anonymous namespace)";
+
+ add_partial_symbol (pdi, cu, 1);
/* Now scan partial symbols in that namespace. */
if (pdi->has_children)
scan_partial_symbols (pdi->die_child,
- lowpc, highpc, cu, full_name);
+ lowpc, highpc, cu, 1);
}
/* Read a partial die corresponding to a class or structure. */
@@ -1744,14 +1755,16 @@ add_partial_namespace (struct partial_di
static void
add_partial_structure (struct partial_die_info *struct_pdi,
struct dwarf2_cu *cu,
- const char *namespace)
+ int have_namespace_info)
{
bfd *abfd = cu->objfile->obfd;
- char *actual_class_name = NULL;
+ char *full_name;
+
+ if (struct_pdi->name == NULL)
+ struct_pdi->name = "(anonymous class)";
if (cu->language == language_cplus
- && namespace == NULL
- && struct_pdi->name != NULL
+ && have_namespace_info == 0
&& struct_pdi->has_children)
{
/* See if we can figure out if the class lives in a namespace
@@ -1771,9 +1784,16 @@ add_partial_structure (struct partial_di
{
if (child_pdi->tag == DW_TAG_subprogram)
{
- actual_class_name = class_name_from_physname (child_pdi->name);
+ char *actual_class_name
+ = class_name_from_physname (child_pdi->name);
if (actual_class_name != NULL)
- struct_pdi->name = actual_class_name;
+ {
+ struct_pdi->full_name
+ = obsavestring (actual_class_name,
+ strlen (actual_class_name),
+ &cu->partial_die_obstack);
+ xfree (actual_class_name);
+ }
break;
}
@@ -1781,22 +1801,21 @@ add_partial_structure (struct partial_di
}
}
- add_partial_symbol (struct_pdi, cu, namespace);
- xfree (actual_class_name);
+ add_partial_symbol (struct_pdi, cu, have_namespace_info);
}
/* Read a partial die corresponding to an enumeration type. */
static void
add_partial_enumeration (struct partial_die_info *enum_pdi,
- struct dwarf2_cu *cu, const char *namespace)
+ struct dwarf2_cu *cu, int have_namespace_info)
{
struct objfile *objfile = cu->objfile;
bfd *abfd = objfile->obfd;
struct partial_die_info *pdi_p;
if (enum_pdi->name != NULL)
- add_partial_symbol (enum_pdi, cu, namespace);
+ add_partial_symbol (enum_pdi, cu, have_namespace_info);
pdi_p = enum_pdi->die_child;
while (pdi_p)
@@ -1804,7 +1823,7 @@ add_partial_enumeration (struct partial_
if (pdi_p->tag != DW_TAG_enumerator || pdi_p->name == NULL)
complaint (&symfile_complaints, "malformed enumerator DIE ignored");
else
- add_partial_symbol (pdi_p, cu, namespace);
+ add_partial_symbol (pdi_p, cu, have_namespace_info);
pdi_p = pdi_p->die_sibling;
}
}
@@ -4548,18 +4567,18 @@ load_partial_dies (bfd *abfd, char *info
info_ptr = load_partial_die (part_die, abbrev, bytes_read,
abfd, info_ptr, cu);
- /* This two-pass algorithm for processing partial symbols has a high
- cost in cache pressure. Thus, handle some trivial cases here
- which cover the majority of C partial symbols. DIEs which
- neither have specification tags in them, nor could have specification
- tags elsewhere pointing at them, can simply be processed and
- discarded.
+ /* This two-pass algorithm for processing partial symbols has a
+ high cost in cache pressure. Thus, handle some simple cases
+ here which cover the majority of C partial symbols. DIEs
+ which neither have specification tags in them, nor could have
+ specification tags elsewhere pointing at them, can simply be
+ processed and discarded.
This segment is also optional; scan_partial_symbols and
add_partial_symbol will handle these DIEs if we simply chain
them in normally. When compilers which do not emit large
- quantities of duplicate debug information are more common,
- this code can probably be removed. */
+ quantities of duplicate debug information are more common,
+ this code can probably be removed. */
/* Any complete simple types at the top level (pretty much all
of them, for a language without namespaces), can be processed
@@ -4569,8 +4588,7 @@ load_partial_dies (bfd *abfd, char *info
&& part_die->is_declaration == 0
&& (part_die->tag == DW_TAG_typedef
|| part_die->tag == DW_TAG_base_type
- || part_die->tag == DW_TAG_subrange_type
- || part_die->tag == DW_TAG_enumeration_type))
+ || part_die->tag == DW_TAG_subrange_type))
{
if (part_die->name != NULL)
add_psymbol_to_list (part_die->name, strlen (part_die->name),
@@ -4620,11 +4638,10 @@ load_partial_dies (bfd *abfd, char *info
if (first_die == NULL)
first_die = part_die;
- /* Maybe add the DIE to the hash table. Not all DIEs need that
- we find interesting need to be in the hash table, because we
+ /* Maybe add the DIE to the hash table. Not all DIEs that we
+ find interesting need to be in the hash table, because we
also have the parent/sibling/child chains; only those that we
- might refer to by offset later during partial symbol
- reading.
+ might refer to by offset later during partial symbol reading.
For now this means things that might have be the target of a
DW_AT_specification, DW_AT_abstract_origin, or
@@ -6556,70 +6573,6 @@ determine_prefix_aux (struct die_info *d
else
retval = class_name (parent, cu);
}
- break;
- default:
- retval = parent_prefix;
- break;
- }
-
- if (retval != parent_prefix)
- xfree (parent_prefix);
- return retval;
- }
-}
-
-/* Likewise for partial DIEs. */
-
-static char *
-partial_determine_prefix (struct partial_die_info *die, struct dwarf2_cu *cu)
-{
- char *prefix = partial_determine_prefix_aux (die, cu);
-
- return prefix ? prefix : xstrdup ("");
-}
-
-/* Return the name of the namespace/class that DIE is defined
- within, or NULL if we can't tell. The caller should xfree the
- result. */
-
-static char *
-partial_determine_prefix_aux (struct partial_die_info *die, struct dwarf2_cu *cu)
-{
- struct partial_die_info *parent;
-
- if (cu->language != language_cplus)
- return NULL;
-
- if (die->spec_offset)
- die = find_partial_die (die->spec_offset, cu);
-
- parent = die->die_parent;
-
- if (parent == NULL)
- {
- return (processing_has_namespace_info ? xstrdup ("") : NULL);
- }
- else
- {
- char *parent_prefix = partial_determine_prefix_aux (parent, cu);
- char *retval;
-
- fixup_partial_die (parent, cu);
-
- switch (parent->tag) {
- case DW_TAG_namespace:
- if (parent->name != NULL)
- retval = typename_concat (parent_prefix, parent->name);
- else
- retval = typename_concat (parent_prefix, "(anonymous namespace)");
- break;
- case DW_TAG_class_type:
- case DW_TAG_structure_type:
- if (parent->name != NULL)
- retval = typename_concat (parent_prefix, parent->name);
- else
- retval = typename_concat (parent_prefix,
- "<<anonymous class>>");
break;
default:
retval = parent_prefix;