These fails were introduced by git commit 2469b3c584 with the
inroduction of "input_from_string":
m68k-linux FAIL: MRI structured for
m68k-linux FAIL: MRI structured if
m68k-linux FAIL: MRI structured repeat
m68k-linux FAIL: MRI structured while
Since the m68k parser called expression() without setting
input_from_string, get_symbol_name rejected FAKE_LABEL_CHAR in names.
* config/m68k-parse.y (yylex): Use temp_ilp and restore_ilp.
This one looks to be a bug going back to 2009, git commit e054468f6c
"STT_GNU_IFUNC support for PowerPC". That bug was carried over with
git commit 49c09209d0 "Rearrange PLT reloc output on powerpc".
If the refcount for an ifunc local sym plt entry was zero,
ppc_elf_size_dynamic_sections would correctly set plt.offset to -1 but
leave glink_offset uninitialized. That leads to occasional segfaults
(which can be made solid with MALLOC_PERTURB_=1 when using glibc).
So, guard the write_glink_stub call with plt.offset != -1. Also,
remove the totally ineffective attempt at writing multiple-use glink
stubs only once.
PR 24355
* elf32-ppc.c (ppc_finish_symbols): Don't call write_glink_stub
for local iplt syms with ent->plt.offset == -1. Remove ineffective
attempt at writing glink stubs only once.
* testsuite/ld-elf/merge.d: Remove csky from xfails, add moxie.
* testsuite/ld-elf/pr21884.d: Remove csky from xfails.
* testsuite/ld-elf/shared.exp: Add csky to list not xfailing pr22374.
* testsuite/ld-unique/pr21529.d: Remove csky from xfails
The build failure was noticed by Helmut Jarausch in
https://bugs.gentoo.org/680232:
$ ./configure CXXFLAGS='-std=c++17 -Os'
...
CXXLD gdb
ld: init.o: in function `initialize_all_files()':
init.c:(.text+0x113): undefined reference to `_initialize_string_view_selftests()'
It happens because '_initialize_string_view_selftests()' is
conditionally defined based on C++ default.
The change defines '_initialize_string_view_selftests()'
unconditionally and leaves implementation a no-op on c++17
compilers.
gdb/ChangeLog:
2019-03-17 Sergei Trofimovich <siarheit@google.com>
* unittests/string_view-selftests.c: Define
_initialize_string_view_selftests unconditionally.
Since all AVX512 processors support AVX, we can encode 256-bit/512-bit
VEX/EVEX vector register clearing instructions with 128-bit VEX vector
register clearing instructions at -O1.
* config/tc-i386.c (optimize_encoding): Encode 256-bit/512-bit
VEX/EVEX vector register clearing instructions with 128-bit VEX
vector register clearing instructions at -O1.
* doc/c-i386.texi: Update -O1 and -O2 documentation.
* testsuite/gas/i386/i386.exp: Run optimize-1a and
x86-64-optimize-2a.
* testsuite/gas/i386/optimize-1a.d: New file.
* testsuite/gas/i386/x86-64-optimize-2a.d: Likewise.
There is unused variable text_vma in function windows_make_so. This
leads to build error on Windows using Cygwin.
gdb/ChangeLog:
2019-03-17 Vladimir Martyanov <vilgeforce@gmail.com>
PR gdb/24350
* windows-nat.c (windows_make_so): Remove unused text_vma variable.
There are a wrong format strings in function display_selector() in
file windows-nat.c. This leads to build error using Cygwin on Windows.
LDT_ENTRY.HighWord is a DWORD, which is unsigned long int, so the
format specification should be for long int, not simply int.
gdb/ChangeLog:
2019-03-17 Vladimir Martyanov <vilgeforce@gmail.com>
PR gdb/24351
* windows-nat.c (display_selector): Format specifications fixed
Without this change, when the current line is longer than the source
window width, redisplaying that line overwrites the window frame and
also portions of the next line.
gdb/ChangeLog:
2019-03-17 Eli Zaretskii <eliz@gnu.org>
* tui/tui-winsource.c (tui_set_is_exec_point_at): Call
tui_refill_source_window instead of tui_refresh_win, to update the
current execution line. This fixes redisplay of the current line
when stepping through very long lines with "next" or "step".
Set optimize to INT_MAX, instead of -1, for -Os so that -Os will include
-O2 optimization.
PR gas/24353
* config/tc-i386.c (md_parse_option): Set optimize to INT_MAX
for -Os.
* testsuite/gas/i386/optimize-2.s: Add a test.
* testsuite/gas/i386/x86-64-optimize-3.s: Likewise.
* testsuite/gas/i386/optimize-2.d: Updated.
* testsuite/gas/i386/x86-64-optimize-3.d: Likewise.
We can optimize 512-bit EVEX to 128-bit EVEX encoding for upper 16
vector registers only when AVX512VL is enabled. We can't optimize
EVEX to 128-bit VEX encoding when AVX isn't enabled.
PR gas/24352
* config/tc-i386.c (optimize_encoding): Encode 512-bit EVEX
with 128-bit VEX encoding only when AVX is enabled and with
128-bit EVEX encoding only when AVX512VL is enabled.
* testsuite/gas/i386/i386.exp: Run PR gas/24352 tests.
* testsuite/gas/i386/optimize-6.s: New file.
* testsuite/gas/i386/optimize-6a.d: Likewise.
* testsuite/gas/i386/optimize-6b.d: Likewise.
* testsuite/gas/i386/optimize-6c.d: Likewise.
* testsuite/gas/i386/x86-64-optimize-7.s: Likewise.
* testsuite/gas/i386/x86-64-optimize-7a.d: Likewise.
* testsuite/gas/i386/x86-64-optimize-7b.d: Likewise.
* testsuite/gas/i386/x86-64-optimize-7c.d: Likewise.
* testsuite/gas/i386/x86-64-optimize-2.d: Updated.
gdb/ChangeLog:
2019-03-16 Eli Zaretskii <eliz@gnu.org>
* source-cache.c (source_cache::get_source_lines): Call
find_source_lines to initialize s->nlines. This fixes vertical
scrolling of TUI source window when the DOWN arrow is pressed.
gdb/ChangeLog:
2019-03-16 Eli Zaretskii <eliz@gnu.org>
* tui/tui-winsource.c (tui_show_source_line): Revert "Use
wclrtoeol in tui_show_source_line". This reverts changes made in
commit 4a3045920b.
Reverts commit 1ff31e135f, fixing the problem more generally. There
are likely other place that will segfault on a NULL section.
PR 24337
* elf.c (_bfd_elf_rela_local_sym): Revert last change.
(_bfd_elf_rel_local_sym): Likewise.
* elflink.c (elf_link_input_bfd): Use bfd_und_section for
section of symbols with unrecognized shndx.
This changes struct minimal_symbol to inherit from general_symbol_info
and updates various macros to cope.
Because MSYMBOL_SET_LANGUAGE and MSYMBOL_SET_NAMES were only used from
a single spot, this patch removes them in favor of simply inlining
their definitions. I consider this to be somewhat cleaner, not least
because the "phony polymorphism" provided by such macros is not useful
in practice.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* symtab.h (struct minimal_symbol): Derive from
general_symbol_info.
(MSYMBOL_VALUE, MSYMBOL_VALUE_RAW_ADDRESS)
(MSYMBOL_VALUE_ADDRESS, MSYMBOL_VALUE_BYTES)
(MSYMBOL_BLOCK_VALUE, MSYMBOL_VALUE_CHAIN, MSYMBOL_LANGUAGE)
(MSYMBOL_SECTION, MSYMBOL_OBJ_SECTION, MSYMBOL_NATURAL_NAME)
(MSYMBOL_LINKAGE_NAME, MSYMBOL_DEMANGLED_NAME)
(MSYMBOL_SEARCH_NAME): Update.
(MSYMBOL_SET_LANGUAGE, MSYMBOL_SET_NAMES): Remove.
* solib.c (gdb_bfd_lookup_symbol_from_symtab): Don't use memset.
* minsyms.c (minimal_symbol_reader::record_full): Update.
minimal_symbol_reader::install copies minsyms from the msym_bunch
objects into the allocated memory. It seemed better to me to do this
via memcpy, as that is frequently optimized in libc.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* minsyms.c (minimal_symbol_reader::install): Use memcpy.
Currently, minimal symbols are allocated on the per-BFD obstack.
However, it is also possible for multiple symbol readers to create
minimal symbols for a given objfile. In this case, the minimal
symbols will be reallocated on the obstack, leading to some waste of
storage.
This is a memory leak, but I think it won't be caught by tools like
valgrind, because valgrind doesn't know about obstacks.
This patch fixes the problem by using malloc to allocate the storage
for minimal symbols.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* objfiles.h (struct objfile_per_bfd_storage) <msymbols>: Now a
unique_xmalloc_ptr.
(objfile::msymbols_range::begin, objfile::msymbols_range::end):
Update.
* minsyms.c (lookup_minimal_symbol_by_pc_section)
(build_minimal_symbol_hash_tables)
(minimal_symbol_reader::install): Update.
This changes objfile_per_bfd_storage::demangled_names_hash to be an
htab_up. This lets us remove some manual management code from the
objfile_per_bfd_storage destructor.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* symtab.c (create_demangled_names_hash): Update.
(symbol_set_names): Update.
* objfiles.h (struct objfile_per_bfd_storage)
<demangled_names_hash>: Now an htab_up.
* objfiles.c (objfile_per_bfd_storage): Simplify.
There's no reason that the objfile_per_bfd_storage must be allocated
via bfd_alloc. This patch changes objfile_per_bfd_storage to be
managed more simply, via ordinary new and delete; and moves some code
into its (new) destructor.
While doing this I also noticed an extra initialization of
language_of_main, and removed it.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* objfiles.h (struct objfile_per_bfd_storage): Declare
destructor.
* objfiles.c (objfile_per_bfd_storage::~objfile_per_bfd_storage):
New.
(get_objfile_bfd_data): Use new. Don't initialize
language_of_main.
(free_objfile_per_bfd_storage): Remove.
(objfile_bfd_data_free, objfile::~objfile): Use delete.
I was curious what used the terminating "null" minimal symbol; and
after looking I could not find anything. This patch removes
terminate_minimal_symbol_table and the extra minimal symbol that is
allocated for it.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* symfile.c (reread_symbols): Update.
* objfiles.c (objfile::objfile): Update.
* minsyms.h (terminate_minimal_symbol_table): Don't declare.
* minsyms.c (lookup_minimal_symbol_by_pc_section): Update
comment.
(minimal_symbol_reader::install): Update.
(terminate_minimal_symbol_table): Remove.
* jit.c (jit_object_close_impl): Update.
minimal_symbol_reader::record_full does not need to initialize any
minsym fields to 0, because that was already done implicitly via the
use of XCNEW when allocating the msym_bunch.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* minsyms.c (minimal_symbol_reader::record_full): Remove some
initializations.
I noticed that objfile_per_bfd_storage::demangled_hash_languages is a
std::vector, which seemed quite large for something that,
fundamentally, can be represented as a bitset. This patch
reimplements it as a std::bitset.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* objfiles.h (struct objfile_per_bfd_storage)
<demangled_hash_languages>: Now a bitset.
* minsyms.c (add_minsym_to_demangled_hash_table): Update.
(lookup_minimal_symbol): Update.
Only one caller of minimal_symbol_reader::record_with_info was using
the return value, so this patch simplifies this code by having it
return void and changing that caller to use record_full instead.
gdb/ChangeLog
2019-03-15 Tom Tromey <tom@tromey.com>
* minsyms.h (class minimal_symbol_reader) <record_with_info>:
Don't return the symbol.
* coffread.c (record_minimal_symbol): Use record_full.
For LTO, a symbol may defined in discarded section. We should mark it
as undefined so that LTO plugin will make IR definition available.
PR ld/24267
* coffgen.c (_bfd_coff_section_already_linked): Skip discarded
section.
* cofflink.c (coff_link_add_symbols): Check for symbols defined
in discarded section.
Commit b4be1b0648 ("Fix MI output for multi-location breakpoints")
broke the build of gdb.texinfo. The problem is simply the use of "@end
@table", which should be "@end table".
The error was:
/home/smarchi/src/binutils-gdb/gdb/doc/gdb.texinfo:27870: warning: @table should not appear in @end
/home/smarchi/src/binutils-gdb/gdb/doc/gdb.texinfo:27870: table requires an argument: the formatter for @item
/home/smarchi/src/binutils-gdb/gdb/doc/gdb.texinfo:27870: no matching `@end table'
/home/smarchi/src/binutils-gdb/gdb/doc/gdb.texinfo:27870: bad argument to @end: @table
/home/smarchi/src/binutils-gdb/gdb/doc/gdb.texinfo:27867: warning: @table has text but no @item
/home/smarchi/src/binutils-gdb/gdb/doc/gdb.texinfo:27879: @node seen before @end table
gdb/doc/ChangeLog:
* gdb.texinfo (GDB/MI Development and Front Ends): Fix closing
of table, "@end @table" -> "@end table".
PR 24334
* dwarf2.c (struct dwarf2_debug): Add sec_vma_count field.
(save_section_vma): Initialise field to the number of entries in
the sec_vma table.
(section_vma_same): Check that the number of entries in the
sec_vma table matches the number of sections in the bfd.
The MS-Windows port of ncurses fails to switch to a color pair if
one or both of the colors are the implicit default colors. This
change records the default colors when TUI is initialized, and
then specifies them explicitly when a color pair uses the default
colors. This allows color styling in TUI mode on MS-Windows.
gdb/ChangeLog:
2019-03-14 Eli Zaretskii <eliz@gnu.org>
* tui/tui-io.c [__MINGW32__]: Include windows.h. Declare
ncurses_norm_attr.
(tui_initialize_io) [__MINGW32__]: Record the default terminal
colors in ncurses_norm_attr.
(apply_ansi_escape) [__MINGW32__]: If a color in a color pair is
"none", replace it with the default color recorded in
ncurses_norm_attr.
PR 24332
* elflink.c (elf_link_add_object_symbols): Add new local variable
extversym_end. Initialise it to point to the end of the version
symbol table, if present. Check it when initialising and updating
the ever pointer.
If the first requested line is larger than the number of lines in the
source buffer, source_cache::extract_lines could crash, because it
would try to pass string::npos" to string::substr.
This patch avoids the crash by checking for this case.
This version of the patch changes get_source_lines to return
std::string.
gdb/ChangeLog
2019-03-14 Tom Tromey <tromey@adacore.com>
* source-cache.h (class source_cache) <get_source_lines>: Return
std::string.
* source-cache.c (source_cache::extract_lines): Handle case where
first_pos==npos. Return std::string.
(source_cache::get_source_lines): Update.
When the user toggles "set style enabled", the TUI should react by
redrawing the source window, if necessary. This patch implements this
behavior.
No test because the TUI is generally not tested.
This version of the patch incorporates Pedro's patch to provide a
clean way to force the TUI to update the source window's contents.
gdb/ChangeLog
2019-03-14 Pedro Alves <palves@redhat.com>
Tom Tromey <tromey@adacore.com>
* tui/tui-winsource.h (tui_refill_source_window): Declare.
* tui/tui-winsource.c (tui_refill_source_window): New function,
from...
(tui_horizontal_source_scroll): ... here. Move some logic.
* cli/cli-style.c (set_style_enabled): Notify new observable.
* tui/tui-hooks.c (tui_redisplay_source): New function.
(tui_attach_detach_observers): Attach or detach
tui_redisplay_source.
* observable.h (source_styling_changed): New observable.
* observable.c: Define source_styling_changed observable.
New in v2:
- Addressed comments about doc, updated the MI version table
- New doc for the Breakpoint information format
- New -fix-multi-location-breakpoint-output command, with associated
doc, test and NEWS updated accordingly
- Fixed the output, the locations list is now actually in the tuple
representing the breakpoint.
Various MI commands or events related to breakpoints output invalid MI
records when printing information about a multi-location breakpoint.
For example:
-break-insert allo
^done,bkpt={...,addr="<MULTIPLE>",...},{number="1.1",...},{number="1.2",...}
The problem is that according to the syntax [1], the top-level elements
are of type "result" and should be of the form "variable=value".
This patch changes the output to wrap the locations in a list:
^done,bkpt={...,addr="<MULTIPLE>",locations=[{number="1.1",...},{number="1.2",...}]}
The events =breakpoint-created, =breakpoint-modified, as well as the
-break-info command also suffer from this (and maybe others I didn't
find).
Since this is a breaking change for MI, we have to deal somehow with
backwards compatibility. The approach taken by this patch is to bump
the MI version, use the new syntax in MI3 while retaining the old syntax
in MI2. Frontends are expected to use a precise MI version (-i=mi2), so
if they do that they should be unaffected.
The patch also adds the command -fix-multi-location-breakpoint-output,
which front ends can use to enable this behavior with MI <= 2.
[1] https://sourceware.org/gdb/onlinedocs/gdb/GDB_002fMI-Output-Syntax.html#GDB_002fMI-Output-Syntax
gdb/ChangeLog:
* NEWS: Mention that the new default MI version is 3. Mention
changes to the output of commands and events that deal with
multi-location breakpoints.
* breakpoint.c: Include "mi/mi-out.h".
(print_one_breakpoint): Change output syntax if using MI version
>= 3.
* mi/mi-main.h (mi_cmd_fix_multi_location_breakpoint_output):
New.
(mi_multi_location_breakpoint_output_fixed): New.
* mi/mi-main.c (fix_multi_location_breakpoint_output): New.
(mi_cmd_fix_multi_location_breakpoint_output): New.
(mi_multi_location_breakpoint_output_fixed): New.
* mi/mi-cmds.c (mi_cmds): Register command
-fix-multi-location-breakpoint-output.
* mi/mi-out.c (mi_out_new): Instantiate version 3 when using
interpreter "mi".
gdb/testsuite/ChangeLog:
* mi-breakpoint-location-ena-dis.exp: Rename to ...
* mi-breakpoint-multiple-locations.exp: ... this.
(make_breakpoints_pattern): New proc.
(do_test): Add mi_version parameter, test -break-insert,
-break-info and =breakpoint-created.
gdb/doc/ChangeLog:
* gdb.texinfo (Mode Options): Mention mi3.
(Interpreters): Likewise.
(GDB/MI Development and Front Ends): Add entry for MI 3 in
version table. Document -fix-multi-location-breakpoint-output.
(GDB/MI Breakpoint Information): Document format of breakpoint
location output.
When re-reviewing this [1] I noticed that there were two spots encoding
the logic of instantiating an mi_ui_out object based on the interpreter
name ("mi", "mi1", "mi2" or "mi3"):
- mi_interp::init
- mi_load_progress
Both encode the logic to choose what the default version is when the
interpreter name is "mi". I had forgotten the one in mi_load_progress.
Therefore, I propose extracting that logic to a single function. I
started to add a new overload of mi_out_new, then realized the current
mi_out_new wasn't very useful, being just a thing wrapper around "new
mi_ui_out". So I ended up with just an mi_out_new function taking the
interp name as parameter.
I ran the gdb.mi tests, and verified manually the behavior (including
the load command).
[1] https://sourceware.org/ml/gdb-patches/2019-01/msg00427.html
gdb/ChangeLog:
* mi/mi-out.h (mi_out_new): Change parameter to const char *.
* mi/mi-out.c (mi_out_new): Change parameter to const char *,
instantiate mi_ui_out based on interpreter name.
* mi/mi-interp.c (mi_interp::init): Use the new mi_out_new.
* mi/mi-main.c (mi_load_progress): Likewise.
