This commit is the result of running the gdb/copyright.py script,
which automated the update of the copyright year range for all
source files managed by the GDB project to be updated to include
year 2023.
Pedro pointed out that gdb-add-index is much slower with the new DWARF
indexer. He also noticed that, in some cases, the generated
.gdb_index would have the wrong fully-qualified name for a method.
I tracked this down to a bug in the indexer. If a type could have
methods but was marked as a declaration, the indexer was ignoring it.
However, this meant that the internal map to find the qualified name
was not updated for this container.
The new DIE scanner works more or less along the lines indicated by
the text for the .debug_names section, disregarding the bugs in the
specification.
While working on this, I noticed that whether a DIE is interesting is
a static property of the DIE's abbrev. It also turns out that many
abbrevs imply a static size for the DIE data, and additionally that
for many abbrevs, the sibling offset is stored at a constant offset
from the start of the DIE.
This patch changes the abbrev reader to analyze each abbrev and stash
the results on the abbrev. These combine to speed up the new indexer.
If the "interesting" flag is false, GDB knows to skip the DIE
immediately. If the sibling offset is statically known, skipping can
be done without reading any attributes; and in some other cases, the
DIE can be skipped using simple arithmetic.
The replacement for the DWARF psymbol reader works in a somewhat
different way. The current reader reads and stores all the DIEs that
might be interesting. Then, if it is missing a DIE, it re-scans the
CU and reads them all. This approach is used for both intra- and
inter-CU references.
I instrumented the partial DIE hash to see how frequently it was used:
[ 0] -> 1538165
[ 1] -> 4912
[ 2] -> 96102
[ 3] -> 175
[ 4] -> 244
That is, most DIEs are never used, and some are looked up twice -- but
this is just an artifact of the implementation of
partial_die_info::fixup, which may do two lookups.
Based on this, the new implementation doesn't try to store any DIEs,
but instead just re-scans them on demand. In order to do this,
though, it is convenient to have a cache of DWARF abbrevs. This way,
if a second CU is needed to resolve an inter-CU reference, the abbrevs
for that CU need only be computed a single time.
This commit brings all the changes made by running gdb/copyright.py
as per GDB's Start of New Year Procedure.
For the avoidance of doubt, all changes in this commits were
performed by the script.
abbrev_table::read has a workaround for Irix 6. The last release of
Irix was in 2006, and (according to Wikipedia) hardware produced after
2007 cannot run Irix. I think this workaround can safely be retired.
gdb/ChangeLog
2021-03-13 Tom Tromey <tom@tromey.com>
* dwarf2/abbrev.c (abbrev_table::read): Remove Irix 6 workaround.
Currently, and abbrev_info points to a separately allocated array of
attr_abbrev objects. This array is constructed in a temporary vector,
then copied to the abbrev table's obstack.
This patch changes abbrev_info to use the struct hack to store the
objects directly, and changes abbrev_table::read to avoid an extra
copy when allocating, using the "growing objects" capability of
obstacks.
This saves a bit of space, and also perhaps a little time.
2021-03-06 Tom Tromey <tom@tromey.com>
* dwarf2/read.c (read_attribute): Make 'abbrev' const.
* dwarf2/abbrev.c (abbrev_table::alloc_abbrev): Remove.
(abbrev_table::read): Update.
* dwarf2/abbrev.h (struct attr_abbrev): Move earlier.
(struct abbrev_info): Reformat.
<attrs>: Now an array.
(struct abbrev_table) <alloc_abbrev>: Remove.
This commits the result of running gdb/copyright.py as per our Start
of New Year procedure...
gdb/ChangeLog
Update copyright year range in copyright header of all GDB files.
In a longer series that I am working on, I needed to remove the
objfile parameter from abbrev_table::read. It seemed to me that this
was a simple and relatively harmless patch, so I'm sending it now.
gdb/ChangeLog
2020-11-05 Tom Tromey <tom@tromey.com>
* dwarf2/read.c (read_cutu_die_from_dwo)
(cutu_reader::cutu_reader, cutu_reader::cutu_reader)
(build_type_psymtabs_1): Update.
* dwarf2/abbrev.h (struct abbrev_table): Remove objfile
parameter.
* dwarf2/abbrev.c (abbrev_table::read): Remove objfile parameter.
Don't read section. Add assert.
Profiling showed that calls to abbrev_table::lookup_abbrev were "too
visible". As these are just forwarding calls to the hash table, this
patch inlines the lookup. Also, htab_find_with_hash is used, avoiding
another call.
The run previous to this had times of (see the first patch in the
series for an explanation):
gdb 1.69
libxul 2.02
Ada 2.52
This patch improves the times to:
gdb 1.64
libxul 1.99
Ada 2.47
gdb/ChangeLog
2020-05-27 Tom Tromey <tromey@adacore.com>
* dwarf2/abbrev.h (struct abbrev_table) <lookup_abbrev>: Inline.
Use htab_find_with_hash.
<add_abbrev>: Remove "abbrev_number" parameter.
* dwarf2/abbrev.c (abbrev_table::add_abbrev): Remove
"abbrev_number" parameter. Use htab_find_slot_with_hash.
(hash_abbrev): Add comment.
(abbrev_table::lookup_abbrev): Move to header file.
(abbrev_table::read): Update.
I compiled gdb with -fsanitize=undefined and ran the test suite.
A couple of reports came from passing NULL to memcpy, e.g.:
[...]btrace-common.cc:176:13: runtime error: null pointer passed as argument 2, which is declared to never be null
While it would be better to fix this in the standard, in the meantime
it seems easy to avoid this error.
gdb/ChangeLog
2020-03-31 Tom Tromey <tromey@adacore.com>
* dwarf2/abbrev.c (abbrev_table::read): Conditionally call
memcpy.
gdbsupport/ChangeLog
2020-03-31 Tom Tromey <tromey@adacore.com>
* btrace-common.cc (btrace_data_append): Conditionally call
memcpy.
abbrev_table::read increments cur_abbrev->num_attrs in the inner loop,
but there's no need to do this, as the information is already stored
in the temporary vector.
2020-02-08 Tom Tromey <tom@tromey.com>
* dwarf2/abbrev.c (abbrev_table::read): Simplify.
Change-Id: I765f12850ffa1c6066e884bb22c94468d1abdba4
This changes abbrev_table to use an htab_up rather than an ad hoc,
bucket-based hash table.
2020-02-08 Tom Tromey <tom@tromey.com>
* dwarf2/abbrev.c (abbrev_table): Move constructor from header.
Rewrite.
(abbrev_table::add_abbrev, abbrev_table::lookup_abbrev): Rewrite.
* dwarf2/abbrev.h (struct abbrev_info) <next>: Remove.
(abbrev_table::abbrev_table): No longer inline.
(ABBREV_HASH_SIZE): Remove.
(abbrev_table::m_abbrevs): Now an htab_up.
Change-Id: Icbaa8e49501f9c43218d6a81a7e8c4d3a77d65dc
This cleans up the DWARF abbrev_table API a bit, primarily by making
various methods and members private.
2020-02-08 Tom Tromey <tom@tromey.com>
* dwarf2/read.c (read_cutu_die_from_dwo): Update.
(cutu_reader): Update.
(build_type_psymtabs_1): Update.
* dwarf2/abbrev.c (abbrev_table::read): Rename.
(abbrev_table::alloc_abbrev): Update.
* dwarf2/abbrev.h (abbrev_table_up): Move earlier.
(abbrev_table::read): New static method, renamed from
abbrev_table_read_table.
(abbrev_table::alloc_abbrev)
(abbrev_table::add_abbrev): Now private.
(abbrev_table::abbrev_table): Now private.
(abbrev_table::m_abbrev_obstack): Now private. Rename.
Change-Id: I320dca83b799f672909ae66f73b7aca266adbaf9
This moves the abbrev table code out of dwarf2read.c and into new
files dwarf2/abbrev.[ch].
gdb/ChangeLog
2020-02-08 Tom Tromey <tom@tromey.com>
* dwarf2read.c (abbrev_table_up, struct abbrev_info)
(struct attr_abbrev, ABBREV_HASH_SIZE, struct abbrev_table):
Move.
(read_cutu_die_from_dwo, build_type_psymtabs_1): Update.
(abbrev_table::alloc_abbrev, abbrev_table::add_abbrev)
(abbrev_table::lookup_abbrev, abbrev_table_read_table): Move to
abbrev.c.
* dwarf2/abbrev.h: New file.
* dwarf2/abbrev.c: New file, from dwarf2read.c.
* Makefile.in (COMMON_SFILES): Add dwarf2/abbrev.c.
Change-Id: I87911bc5297de4407587ca849fef8e8d19136c30
