* [PATCH] Introduce objfile::intern
@ 2020-02-24 2:58 Tom Tromey
2020-03-04 23:47 ` Tom Tromey
0 siblings, 1 reply; 3+ messages in thread
From: Tom Tromey @ 2020-02-24 2:58 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This introduces a string cache on the per-BFD object, replacing the
macro and filename caches. Both of these caches just store strings,
so this consolidation by itself saves a little memory (about the size
of a bcache per objfile).
Then this patch switches some allocations on the objfile obstack to
use this bcache instead. This saves more space; and turns out to be a
bit faster as well.
Here are the before and after "maint time" + "maint space" results of
"file ./gdb":
Command execution time: 4.664021 (cpu), 4.728518 (wall)
Space used: 39190528 (+29212672 for this command)
Command execution time: 4.216209 (cpu), 4.107023 (wall)
Space used: 36667392 (+26689536 for this command)
The main interface to the string cache is a new pair of overloaded
methods, objfile::intern.
gdb/ChangeLog
2020-02-23 Tom Tromey <tom@tromey.com>
* symmisc.c (print_symbol_bcache_statistics)
(print_objfile_statistics): Update.
* symfile.c (allocate_symtab): Use intern.
* psymtab.c (partial_symtab::partial_symtab): Use intern.
* objfiles.h (struct objfile_per_bfd_storage) <filename_cache,
macro_cache>: Remove.
<string_cache>: New member.
(struct objfile) <intern>: New methods.
* elfread.c (elf_symtab_read): Use intern.
* dwarf2/read.c (fixup_go_packaging): Intern package name.
(dwarf2_compute_name, dwarf2_physname)
(create_dwo_unit_in_dwp_v1, create_dwo_unit_in_dwp_v2): Intern
names.
(guess_partial_die_structure_name): Update.
(partial_die_info::fixup): Intern name.
(dwarf2_canonicalize_name): Change parameter to objfile. Intern
name.
(dwarf2_name): Intern name. Update.
* buildsym.c (buildsym_compunit::get_macro_table): Use
string_cache.
---
gdb/ChangeLog | 23 +++++++++++++++++++++++
gdb/buildsym.c | 2 +-
gdb/dwarf2/read.c | 43 ++++++++++++++++---------------------------
gdb/elfread.c | 6 +-----
gdb/objfiles.h | 24 ++++++++++++++++++------
gdb/psymtab.c | 4 +---
gdb/symfile.c | 4 +---
gdb/symmisc.c | 10 +++-------
8 files changed, 64 insertions(+), 52 deletions(-)
diff --git a/gdb/buildsym.c b/gdb/buildsym.c
index 4965b552b32..84cb44277a4 100644
--- a/gdb/buildsym.c
+++ b/gdb/buildsym.c
@@ -120,7 +120,7 @@ buildsym_compunit::get_macro_table ()
{
if (m_pending_macros == nullptr)
m_pending_macros = new_macro_table (&m_objfile->per_bfd->storage_obstack,
- &m_objfile->per_bfd->macro_cache,
+ &m_objfile->per_bfd->string_cache,
m_compunit_symtab);
return m_pending_macros;
}
diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
index 46d510eb274..fe54680eb1d 100644
--- a/gdb/dwarf2/read.c
+++ b/gdb/dwarf2/read.c
@@ -1449,7 +1449,7 @@ static const gdb_byte *read_full_die (const struct die_reader_specs *,
static void process_die (struct die_info *, struct dwarf2_cu *);
static const char *dwarf2_canonicalize_name (const char *, struct dwarf2_cu *,
- struct obstack *);
+ struct objfile *);
static const char *dwarf2_name (struct die_info *die, struct dwarf2_cu *);
@@ -9056,8 +9056,7 @@ fixup_go_packaging (struct dwarf2_cu *cu)
if (package_name != NULL)
{
struct objfile *objfile = cu->per_cu->dwarf2_per_objfile->objfile;
- const char *saved_package_name
- = obstack_strdup (&objfile->per_bfd->storage_obstack, package_name.get ());
+ const char *saved_package_name = objfile->intern (package_name.get ());
struct type *type = init_type (objfile, TYPE_CODE_MODULE, 0,
saved_package_name);
struct symbol *sym;
@@ -10183,14 +10182,13 @@ dwarf2_compute_name (const char *name,
if (cu->language == language_cplus)
canonical_name
= dwarf2_canonicalize_name (intermediate_name.c_str (), cu,
- &objfile->per_bfd->storage_obstack);
+ objfile);
/* If we only computed INTERMEDIATE_NAME, or if
INTERMEDIATE_NAME is already canonical, then we need to
- copy it to the appropriate obstack. */
+ intern it. */
if (canonical_name == NULL || canonical_name == intermediate_name.c_str ())
- name = obstack_strdup (&objfile->per_bfd->storage_obstack,
- intermediate_name);
+ name = objfile->intern (intermediate_name);
else
name = canonical_name;
}
@@ -10310,7 +10308,7 @@ dwarf2_physname (const char *name, struct die_info *die, struct dwarf2_cu *cu)
retval = canon;
if (need_copy)
- retval = obstack_strdup (&objfile->per_bfd->storage_obstack, retval);
+ retval = objfile->intern (retval);
return retval;
}
@@ -11636,8 +11634,7 @@ create_dwo_unit_in_dwp_v1 (struct dwarf2_per_objfile *dwarf2_per_objfile,
virtual_dwo_name.c_str ());
}
dwo_file = new struct dwo_file;
- dwo_file->dwo_name = obstack_strdup (&objfile->objfile_obstack,
- virtual_dwo_name);
+ dwo_file->dwo_name = objfile->intern (virtual_dwo_name);
dwo_file->comp_dir = comp_dir;
dwo_file->sections.abbrev = sections.abbrev;
dwo_file->sections.line = sections.line;
@@ -11832,8 +11829,7 @@ create_dwo_unit_in_dwp_v2 (struct dwarf2_per_objfile *dwarf2_per_objfile,
virtual_dwo_name.c_str ());
}
dwo_file = new struct dwo_file;
- dwo_file->dwo_name = obstack_strdup (&objfile->objfile_obstack,
- virtual_dwo_name);
+ dwo_file->dwo_name = objfile->intern (virtual_dwo_name);
dwo_file->comp_dir = comp_dir;
dwo_file->sections.abbrev =
create_dwp_v2_section (dwarf2_per_objfile, &dwp_file->sections.abbrev,
@@ -17953,8 +17949,7 @@ partial_die_info::read (const struct die_reader_specs *reader,
struct objfile *objfile = dwarf2_per_objfile->objfile;
name
- = dwarf2_canonicalize_name (DW_STRING (&attr), cu,
- &objfile->per_bfd->storage_obstack);
+ = dwarf2_canonicalize_name (DW_STRING (&attr), cu, objfile);
}
break;
}
@@ -18279,9 +18274,7 @@ guess_partial_die_structure_name (struct partial_die_info *struct_pdi,
if (actual_class_name != NULL)
{
struct objfile *objfile = cu->per_cu->dwarf2_per_objfile->objfile;
- struct_pdi->name
- = obstack_strdup (&objfile->per_bfd->storage_obstack,
- actual_class_name.get ());
+ struct_pdi->name = objfile->intern (actual_class_name.get ());
}
break;
}
@@ -18361,7 +18354,7 @@ partial_die_info::fixup (struct dwarf2_cu *cu)
base = demangled.get ();
struct objfile *objfile = cu->per_cu->dwarf2_per_objfile->objfile;
- name = obstack_strdup (&objfile->per_bfd->storage_obstack, base);
+ name = objfile->intern (base);
}
}
@@ -21678,7 +21671,7 @@ sibling_die (struct die_info *die)
static const char *
dwarf2_canonicalize_name (const char *name, struct dwarf2_cu *cu,
- struct obstack *obstack)
+ struct objfile *objfile)
{
if (name && cu->language == language_cplus)
{
@@ -21687,7 +21680,7 @@ dwarf2_canonicalize_name (const char *name, struct dwarf2_cu *cu,
if (!canon_name.empty ())
{
if (canon_name != name)
- name = obstack_strdup (obstack, canon_name);
+ name = objfile->intern (canon_name);
}
}
@@ -21761,10 +21754,7 @@ dwarf2_name (struct die_info *die, struct dwarf2_cu *cu)
const char *base;
- /* FIXME: we already did this for the partial symbol... */
- DW_STRING (attr)
- = obstack_strdup (&objfile->per_bfd->storage_obstack,
- demangled.get ());
+ DW_STRING (attr) = objfile->intern (demangled.get ());
DW_STRING_IS_CANONICAL (attr) = 1;
/* Strip any leading namespaces/classes, keep only the base name.
@@ -21784,9 +21774,8 @@ dwarf2_name (struct die_info *die, struct dwarf2_cu *cu)
if (!DW_STRING_IS_CANONICAL (attr))
{
- DW_STRING (attr)
- = dwarf2_canonicalize_name (DW_STRING (attr), cu,
- &objfile->per_bfd->storage_obstack);
+ DW_STRING (attr) = dwarf2_canonicalize_name (DW_STRING (attr), cu,
+ objfile);
DW_STRING_IS_CANONICAL (attr) = 1;
}
return DW_STRING (attr);
diff --git a/gdb/elfread.c b/gdb/elfread.c
index 453bca527e9..710ecbf6e17 100644
--- a/gdb/elfread.c
+++ b/gdb/elfread.c
@@ -353,11 +353,7 @@ elf_symtab_read (minimal_symbol_reader &reader,
if (type == ST_DYNAMIC && !stripped)
continue;
if (sym->flags & BSF_FILE)
- {
- filesymname
- = ((const char *) objfile->per_bfd->filename_cache.insert
- (sym->name, strlen (sym->name) + 1));
- }
+ filesymname = objfile->intern (sym->name);
else if (sym->flags & BSF_SECTION_SYM)
continue;
else if (sym->flags & (BSF_GLOBAL | BSF_LOCAL | BSF_WEAK
diff --git a/gdb/objfiles.h b/gdb/objfiles.h
index b71a8a9edb8..a568fa4bcda 100644
--- a/gdb/objfiles.h
+++ b/gdb/objfiles.h
@@ -275,13 +275,9 @@ struct objfile_per_bfd_storage
auto_obstack storage_obstack;
- /* Byte cache for file names. */
+ /* String cache. */
- gdb::bcache filename_cache;
-
- /* Byte cache for macros. */
-
- gdb::bcache macro_cache;
+ gdb::bcache string_cache;
/* The gdbarch associated with the BFD. Note that this gdbarch is
determined solely from BFD information, without looking at target
@@ -533,6 +529,22 @@ public:
return section_offsets[SECT_OFF_DATA (this)];
}
+ /* Intern STRING and return the unique copy. The copy has the same
+ lifetime as the per-BFD object. */
+ const char *intern (const char *str)
+ {
+ return (const char *) per_bfd->string_cache.insert (str, strlen (str) + 1);
+ }
+
+ /* Intern STRING and return the unique copy. The copy has the same
+ lifetime as the per-BFD object. */
+ const char *intern (const std::string &str)
+ {
+ return (const char *) per_bfd->string_cache.insert (str.c_str (),
+ str.size () + 1);
+ }
+
+
/* The object file's original name as specified by the user,
made absolute, and tilde-expanded. However, it is not canonicalized
(i.e., it has not been passed through gdb_realpath).
diff --git a/gdb/psymtab.c b/gdb/psymtab.c
index fd7fc8feff2..69176dbee47 100644
--- a/gdb/psymtab.c
+++ b/gdb/psymtab.c
@@ -1656,9 +1656,7 @@ partial_symtab::partial_symtab (const char *filename_, struct objfile *objfile)
{
objfile->partial_symtabs->install_psymtab (this);
- filename
- = ((const char *) objfile->per_bfd->filename_cache.insert
- (filename_, strlen (filename_) + 1));
+ filename = objfile->intern (filename_);
if (symtab_create_debug)
{
diff --git a/gdb/symfile.c b/gdb/symfile.c
index f7bada75f35..dc423f0344a 100644
--- a/gdb/symfile.c
+++ b/gdb/symfile.c
@@ -2778,9 +2778,7 @@ allocate_symtab (struct compunit_symtab *cust, const char *filename)
struct symtab *symtab
= OBSTACK_ZALLOC (&objfile->objfile_obstack, struct symtab);
- symtab->filename
- = ((const char *) objfile->per_bfd->filename_cache.insert
- (filename, strlen (filename) + 1));
+ symtab->filename = objfile->intern (filename);
symtab->fullname = NULL;
symtab->language = deduce_language_from_filename (filename);
diff --git a/gdb/symmisc.c b/gdb/symmisc.c
index a6a7e728c4a..1d7c3816670 100644
--- a/gdb/symmisc.c
+++ b/gdb/symmisc.c
@@ -71,9 +71,7 @@ print_symbol_bcache_statistics (void)
objfile_name (objfile));
objfile->partial_symtabs->psymbol_cache.print_statistics
("partial symbol cache");
- objfile->per_bfd->macro_cache.print_statistics
- ("preprocessor macro cache");
- objfile->per_bfd->filename_cache.print_statistics ("file name cache");
+ objfile->per_bfd->string_cache.print_statistics ("string cache");
}
}
@@ -135,10 +133,8 @@ print_objfile_statistics (void)
printf_filtered
(_(" Total memory used for psymbol cache: %d\n"),
objfile->partial_symtabs->psymbol_cache.memory_used ());
- printf_filtered (_(" Total memory used for macro cache: %d\n"),
- objfile->per_bfd->macro_cache.memory_used ());
- printf_filtered (_(" Total memory used for file name cache: %d\n"),
- objfile->per_bfd->filename_cache.memory_used ());
+ printf_filtered (_(" Total memory used for string cache: %d\n"),
+ objfile->per_bfd->string_cache.memory_used ());
}
}
--
2.17.2
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] Introduce objfile::intern
2020-02-24 2:58 [PATCH] Introduce objfile::intern Tom Tromey
@ 2020-03-04 23:47 ` Tom Tromey
2020-03-05 9:10 ` [committed][gdb/testsuite] Update maint.exp for string cache Tom de Vries
0 siblings, 1 reply; 3+ messages in thread
From: Tom Tromey @ 2020-03-04 23:47 UTC (permalink / raw)
To: Tom Tromey; +Cc: gdb-patches
>>>>> "Tom" == Tom Tromey <tom@tromey.com> writes:
Tom> This introduces a string cache on the per-BFD object, replacing the
Tom> macro and filename caches. Both of these caches just store strings,
Tom> so this consolidation by itself saves a little memory (about the size
Tom> of a bcache per objfile).
I'm checking this in now.
Tom
^ permalink raw reply [flat|nested] 3+ messages in thread
* [committed][gdb/testsuite] Update maint.exp for string cache
2020-03-04 23:47 ` Tom Tromey
@ 2020-03-05 9:10 ` Tom de Vries
0 siblings, 0 replies; 3+ messages in thread
From: Tom de Vries @ 2020-03-05 9:10 UTC (permalink / raw)
To: Tom Tromey; +Cc: gdb-patches
[-- Attachment #1: Type: text/plain, Size: 529 bytes --]
[ was: Re: [PATCH] Introduce objfile::intern ]
On 05-03-2020 00:47, Tom Tromey wrote:
>>>>>> "Tom" == Tom Tromey <tom@tromey.com> writes:
>
> Tom> This introduces a string cache on the per-BFD object, replacing the
> Tom> macro and filename caches. Both of these caches just store strings,
> Tom> so this consolidation by itself saves a little memory (about the size
> Tom> of a bcache per objfile).
>
> I'm checking this in now.
This caused a regression in gdb.base/maint.exp, fixed by committed patch
below.
Thanks,
- Tom
[-- Attachment #2: 0001-gdb-testsuite-Update-maint.exp-for-string-cache.patch --]
[-- Type: text/x-patch, Size: 2482 bytes --]
[gdb/testsuite] Update maint.exp for string cache
When running gdb.base/maint.exp, I see:
...
FAIL: gdb.base/maint.exp: maint print statistics
...
This is due to commit be1e3d3eab "Introduce objfile::intern", which replaces
the macro and filename caches with a string cache.
Update maint.exp accordingly.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2020-03-05 Tom de Vries <tdevries@suse.de>
* gdb.base/maint.exp: Update "main print statistics" expected output.
---
gdb/testsuite/gdb.base/maint.exp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gdb/testsuite/gdb.base/maint.exp b/gdb/testsuite/gdb.base/maint.exp
index fe25e0fc61..f6eeb98a20 100644
--- a/gdb/testsuite/gdb.base/maint.exp
+++ b/gdb/testsuite/gdb.base/maint.exp
@@ -200,7 +200,7 @@ if [istarget "*-*-cygwin*"] {
send_gdb "maint print statistics\n"
gdb_expect {
- -re "Statistics for\[^\n\r\]*maint\[^\n\r\]*:\r\n Number of \"minimal\" symbols read: $decimal\r\n( Number of \"partial\" symbols read: $decimal\r\n)? Number of \"full\" symbols read: $decimal\r\n Number of \"types\" defined: $decimal\r\n( Number of psym tables \\(not yet expanded\\): $decimal\r\n)?( Number of read CUs: $decimal\r\n Number of unread CUs: $decimal\r\n)? Number of symbol tables: $decimal\r\n Number of symbol tables with line tables: $decimal\r\n Number of symbol tables with blockvectors: $decimal\r\n Total memory used for objfile obstack: $decimal\r\n Total memory used for BFD obstack: $decimal\r\n Total memory used for psymbol cache: $decimal\r\n Total memory used for macro cache: $decimal\r\n Total memory used for file name cache: $decimal\r\n" {
+ -re "Statistics for\[^\n\r\]*maint\[^\n\r\]*:\r\n Number of \"minimal\" symbols read: $decimal\r\n( Number of \"partial\" symbols read: $decimal\r\n)? Number of \"full\" symbols read: $decimal\r\n Number of \"types\" defined: $decimal\r\n( Number of psym tables \\(not yet expanded\\): $decimal\r\n)?( Number of read CUs: $decimal\r\n Number of unread CUs: $decimal\r\n)? Number of symbol tables: $decimal\r\n Number of symbol tables with line tables: $decimal\r\n Number of symbol tables with blockvectors: $decimal\r\n Total memory used for objfile obstack: $decimal\r\n Total memory used for BFD obstack: $decimal\r\n Total memory used for psymbol cache: $decimal\r\n Total memory used for string cache: $decimal\r\n" {
gdb_expect {
-re "$gdb_prompt $" {
pass "maint print statistics"
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-03-05 9:10 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-02-24 2:58 [PATCH] Introduce objfile::intern Tom Tromey
2020-03-04 23:47 ` Tom Tromey
2020-03-05 9:10 ` [committed][gdb/testsuite] Update maint.exp for string cache Tom de Vries
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).