Joel Brobecker [Tue, 1 Jan 2019 06:01:51 +0000 (10:01 +0400)]
Update copyright year range in all GDB files.
This commit applies all changes made after running the gdb/copyright.py
script.
Note that one file was flagged by the script, due to an invalid
copyright header
(gdb/unittests/basic_string_view/element_access/char/empty.cc).
As the file was copied from GCC's libstdc++-v3 testsuite, this commit
leaves this file untouched for the time being; a patch to fix the header
was sent to gcc-patches first.
Use gdb::unique_xmalloc_ptr<char> in command_line_input to fix a leak
Following the change of logic where the input_handler gets a
gdb::unique_xmalloc_ptr<char>, a call to readline directly
followed by a call to handle_line_of_input is missing a free,
and causes the below leak.
Use gdb::unique_xmalloc_ptr<char> to solve the leak.
==16291== VALGRIND_GDB_ERROR_BEGIN
==16291== 64 bytes in 1 blocks are definitely lost in loss record 1,815 of 4,111
==16291== at 0x4C2E2B3: realloc (vg_replace_malloc.c:836)
==16291== by 0x41EB1C: xrealloc (common-utils.c:62)
==16291== by 0x41DBD3: buffer_grow(buffer*, char const*, unsigned long) [clone .part.1] (buffer.c:40)
==16291== by 0x66E8FF: buffer_grow_char (buffer.h:40)
==16291== by 0x66E8FF: gdb_readline_no_editing (top.c:798)
==16291== by 0x66E8FF: command_line_input(char const*, char const*) (top.c:1249)
==16291== by 0x66EBD8: read_command_file(_IO_FILE*) (top.c:421)
==16291== by 0x412C0C: script_from_file(_IO_FILE*, char const*) (cli-script.c:1547)
==16291== by 0x40BE90: source_script_from_stream (cli-cmds.c:569)
==16291== by 0x40BE90: source_script_with_search(char const*, int, int) (cli-cmds.c:606)
==16291== by 0x54D567: catch_command_errors(void (*)(char const*, int), char const*, int) (main.c:379)
==16291== by 0x54EA84: captured_main_1 (main.c:994)
==16291== by 0x54EA84: captured_main (main.c:1167)
==16291== by 0x54EA84: gdb_main(captured_main_args*) (main.c:1193)
==16291== by 0x29DA27: main (gdb.c:32)
==16291==
==16291== VALGRIND_GDB_ERROR_END
gdb/ChangeLog
2018-12-31 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* top.c (command_line_input): Use unique_xmalloc_ptr to
manage memory allocated by readline.
Alan Modra [Mon, 31 Dec 2018 06:36:25 +0000 (17:06 +1030)]
PR24042, Global-buffer-overflow problem in output_rel_find
place_orphan handled ELF SHT_REL/SHT_RELA specially, output_rel_find
didn't. This mismatch was a bug and also meant it was possible to
craft an object where ld accessed section->name out of bounds.
PR 24042
* emultempl/elf32.em (output_rel_find): Drop "sec" param. Add
"rela".
(gld${EMULATION_NAME}_place_orphan): Use sh_type to calculate
"rela" param of output_rel_find when ELF. Tidy uses of elfinput.
Alan Modra [Mon, 31 Dec 2018 01:41:42 +0000 (12:11 +1030)]
[PowerPC64] Nop out ld 2,24(1) after old-style __tls_get_addr
When optimising inline plt calls to __tls_get_addr without tls marker
relocs, ld should zap any toc restore insn after the bctrl, to stop a
load-hit-store stall.
* elf64-ppc.c (ppc64_elf_relocate_section <tls_ldgd_opt>): When
editing an old-style __tls_get_addr call, replace a toc restore
insn with a nop.
The layout for 'help set address|variable' is strange, e.g.:
(gdb) help set style address
style address
List of show Address display styling
Configure address colors and display intensity subcommands:
show Address display styling
Configure address colors and display intensity background -- Set the background color for this property
show Address display styling
Configure address colors and display intensity foreground -- Set the foreground color for this property
show Address display styling
Configure address colors and display intensity intensity -- Set the display intensity color for this property
Type "help show Address display styling
Configure address colors and display intensity" followed by show Address display styling
Configure address colors and display intensity subcommand name for full documentation.
Type "apropos word" to search for commands related to "word".
Command name abbreviations are allowed if unambiguous.
(gdb)
The help for 'set style function|filename' gives help for 'Show':
(gdb) help set style filename
Filename display styling
Configure filename colors and display intensity.
List of show style filename subcommands:
show style filename background -- Set the background color for this property
show style filename foreground -- Set the foreground color for this property
show style filename intensity -- Set the display intensity color for this property
The help for 'show style function|filename' is equally strange, as it speaks
about commands, instead of sub commands:
(gdb) help show style filename
Filename display styling
Configure filename colors and display intensity.
List of commands:
background -- Show the background color for this property
foreground -- Show the foreground color for this property
intensity -- Show the display intensity color for this property
Type "help" followed by command name for full documentation.
Type "apropos word" to search for commands related to "word".
Command name abbreviations are allowed if unambiguous.
(gdb)
This patch fixes all this.
Note that the 'set style' and 'show style' have the same prefix_doc:
(gdb) help show style
Style-specific settings
Configure various style-related variables, such as colors
...
(gdb) help set style
Style-specific settings
Configure various style-related variables, such as colors
...
Other similar commands (such as set|show history) have typically
a more specific prefix:
(gdb) help show history
Generic command for showing command history parameters.
...
(gdb) help set history
Generic command for setting command history parameters.
...
This could be fixed by having set_prefix_doc and show_prefix_doc instead of
the single prefix_doc argument to cli_style_option::add_setshow_commands.
That could be improved if deemed better.
2018-12-29 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* cli/cli-style.c (cli_style_option::add_setshow_commands):
Initialize m_set_prefix with "set", instead of re-assigning
m_show_prefix. Use m_set_prefix for set_list and m_show_prefix
for show_list.
(_initialize_cli_style): Correct the order of arguments in
variable_name_style.add_setshow_commands and
address_style.add_setshow_commands calls.
Tom Tromey [Sat, 29 Dec 2018 04:07:24 +0000 (21:07 -0700)]
Fix the build when GNU Source Highlight is not available
The builder pointed out that, when GNU Source Highlight is not
available, get_language_name is not used. This patch makes it
conditional, fixing the build problem.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
Tom Tromey [Wed, 19 Dec 2018 18:44:28 +0000 (11:44 -0700)]
Fix a crash in jit.c
A user at Mozilla pointed out a crash in jit.c. In his situation, an
inferior using the JIT API exec'd an executable that did not use it.
This caused an assertion failure when jit.c:free_objfile_data called
delete_breakpoint with NULL.
This patch fixes the problem in the obvious way. New test case
included.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
* jit.c (free_objfile_data): Only delete breakpoint if non-null.
gdb/testsuite/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
Simon Marchi <simark@simark.ca>
* gdb.base/jit-exec.exp: New file.
* gdb.base/jit-exec.c: New file.
* gdb.base/jit-execd.c: New file.
Tom Tromey [Mon, 3 Sep 2018 18:42:59 +0000 (12:42 -0600)]
Make ANSI terminal escape sequences work in TUI
PR tui/14126 notes that ANSI terminal escape sequences don't affect
the colors shown in the TUI. A simple way to see this is to try the
extended-prompt example from the gdb manual.
Curses does not pass escape sequences through to the terminal.
Instead, it replaces non-printable characters with a visible
representation, for example "^[" for the ESC character.
This patch fixes the problem by adding a simple ANSI terminal sequence
parser to gdb. These sequences are decoded and those that are
recognized are turned into the appropriate curses calls.
The curses approach to color handling is unusual and so there are some
oddities in the implementation.
Standard curses has no notion of the default colors of the terminal.
So, if you set the foreground color, it is not possible to reset it --
you have to pick some other color. ncurses provides an extension to
handle this, so this patch updates configure and uses it when
available.
Second, in curses, colors always come in pairs: you cannot set just
the foreground. This patch handles this by tracking actually-used
pairs of colors and keeping a table of these for reuse.
Third, there are a limited number of such pairs available. In this
patch, if you try to use too many color combinations, gdb will just
ignore some color changes.
Finally, in addition to limiting the number of color pairs, curses
also limits the number of colors. This means that, when using
extended 8- or 24-bit color sequences, it may be possible to exhaust
the curses color table.
I am very sour on the curses design now.
I do not know how to write a test for this, so I did not.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
PR tui/14126:
* tui/tui.c (tui_enable): Call start_color and
use_default_colors.
* tui/tui-io.c (struct color_pair): New.
(color_pair_map, last_color_pair, last_style): New globals.
(tui_setup_io): Clean up color map when shutting down.
(curses_colors): New constant.
(get_color_pair, apply_ansi_escape): New functions.
(tui_write): Rewrite.
(tui_puts_internal): New function, from tui_puts. Add "height"
parameter.
(tui_puts): Use tui_puts_internal.
(tui_redisplay_readline): Use tui_puts_internal.
(_initialize_tui_io): New function.
(color_map): New globals.
(get_color): New function.
* configure.ac: Check for use_default_colors.
* config.in, configure: Rebuild.
Tom Tromey [Thu, 6 Sep 2018 05:32:16 +0000 (23:32 -0600)]
Style locations when setting a breakpoint
say_where does not use ui-out, so function and file names printed by
it were not styled. This patch changes say_where to use the low-level
style code directly.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
* breakpoint.c (say_where): Style file name.
gdb/testsuite/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
* gdb.base/style.exp: Add test for breakpoint setting.
Tom Tromey [Wed, 5 Sep 2018 18:12:19 +0000 (12:12 -0600)]
Style variable names
This adds style support for variable names. For the time being, this
is only done in backtraces, not in ptype or print; those places do not
use ui-out and so would need ad hoc changes.
This also adds styling to the names printed for local variables in
"backtrace full". This code does not use ui-out, so the styling is
done using the low-level API.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
Tom Tromey [Tue, 4 Sep 2018 04:56:33 +0000 (22:56 -0600)]
Add output styles to gdb
This adds some output styling to the CLI.
A style is currently a foreground color, a background color, and an
intensity (dim or bold). (This list could be expanded depending on
terminal capabilities.)
A style can be applied while printing. For ui-out, this is done by
passing the style constant as an argument. For low-level cases,
fprintf_styled and fputs_styled are provided.
Users can control the style via a number of new set/show commands. In
the interest of not typing many nearly-identical documentation
strings, I automated this. On the down side, this is not very
i18n-friendly.
I've chose some default colors to use. I think it would be good to
enable this by default, so that when users start the new gdb, they
will see the new feature.
Stylizing is done if TERM is set and is not "dumb". This could be
improved when the TUI is available by using the curses has_colors
call. That is, the lowest layer could call this without committing to
using curses everywhere; see my other patch for TUI colorizing.
I considered adding a new "set_style" method to ui_file. However,
because the implementation had to interact with the pager code, I
didn't take this approach. But, one idea might be to put the isatty
check there and then have it defer to the lower layers.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
Tom Tromey [Tue, 13 Nov 2018 18:59:03 +0000 (11:59 -0700)]
Change gdb test suite's TERM setting
This changes the gdb test suite to set TERM to "dumb" by default.
This setting disables terminal styling, so that the existing tests do
not need to be updated.
gdb/testsuite/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
* lib/gdb.exp (gdb_init): Set the TERM environment variable to
"dumb".
* gdb.base/readline.exp (operate_and_get_next): Save and restore
the TERM environment variable.
Tom Tromey [Sat, 10 Nov 2018 00:29:50 +0000 (17:29 -0700)]
Introduce ui_file_style
This introduces the new ui_file_style class and various helpers. This
class represents a terminal style and provides methods for parsing and
emitting the corresponding ANSI terminal escape sequences.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
* unittests/style-selftests.c: New file.
* ui-style.c: New file.
* ui-style.h: New file.
* ui-file.h: Include ui-style.h.
* Makefile.in (COMMON_SFILES): Add ui-style.c.
(HFILES_NO_SRCDIR): Add ui-style.h.
(SUBDIR_UNITTESTS_SRCS): Add style-selftests.c.
Tom Tromey [Thu, 6 Sep 2018 20:03:38 +0000 (14:03 -0600)]
Add a "context" argument to add_setshow_enum_cmd
This adds a "context" argument to add_setshow_enum_cmd. Now
add_setshow_enum_cmd will call set_cmd_context on both of the new
commands. This is used in a later patch.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
Tom Tromey [Wed, 5 Sep 2018 17:25:28 +0000 (11:25 -0600)]
Change wrap buffering to use a std::string
Currently wrap buffering is implemented by allocating a string that is
the same width as the window, and then writing characters into it.
However, if gdb emits terminal escapes, then these could possibly
overflow the buffer.
To prevent this, change the wrap buffer to be a std::string and update
the various uses.
This also changes utils.c to always emit characters to the wrap
buffer. This simplifies future patches which emit terminal escape
sequences, and also makes it possible for the "echo" and "printf"
commands to be used to emit terminal escapes and have these work in
the TUI.
gdb/ChangeLog
2018-12-28 Tom Tromey <tom@tromey.com>
* utils.c (filter_initialized): New global.
(wrap_buffer): Now a std::string.
(wrap_pointer): Remove.
(flush_wrap_buffer): New function.
(filtered_printing_initialized, set_width, wrap_here)
(fputs_maybe_filtered): Update.
Fix leak of set/show verbose doc, avoid xfree of static string
In the tests
py-pp-registration/gdb.log
default/gdb.log
foll-fork/gdb.log
setshow/gdb.log
break-interp/gdb.log
Valgrind detects a leak of the doc strings for the set and show verbose cmd.
Here is the stacktrace of the leaked set doc:
==25548== 15 bytes in 1 blocks are definitely lost in loss record 101 of 3,120
==25548== at 0x4C2BE6D: malloc (vg_replace_malloc.c:309)
==25548== by 0x409C27: xmalloc (common-utils.c:44)
==25548== by 0x778AF9: xstrdup (xstrdup.c:34)
==25548== by 0x3F860F: add_setshow_cmd_full(char const*, command_class, var_types, void*, char const*, char const*, char const*, void (*)(char const*, int, cmd_list_element*), void (*)(ui_file*, int, cmd_list_element*, char const*), cmd_list_element**, cmd_list_element**, cmd_list_element**, cmd_list_element**) [clone .constprop.10] (cli-decode.c:495)
==25548== by 0x3F8ADB: add_setshow_boolean_cmd(char const*, command_class, int*, char const*, char const*, char const*, void (*)(char const*, int, cmd_list_element*), void (*)(ui_file*, int, cmd_list_element*, char const*), cmd_list_element**, cmd_list_element**) (cli-decode.c:593)
==25548== by 0x3F7442: _initialize_cli_cmds() (cli-cmds.c:1768)
==25548== by 0x69EED3: initialize_all_files() (init.c:365)
==25548== by 0x658A84: gdb_init(char*) (top.c:2163)
==25548== by 0x5403E1: captured_main_1 (main.c:863)
==25548== by 0x5403E1: captured_main (main.c:1167)
==25548== by 0x5403E1: gdb_main(captured_main_args*) (main.c:1193)
==25548== by 0x289CA7: main (gdb.c:32)
The leak is created by top.c set_verbose 'elaborate joke':
the doc string is changed according to the verbosity:
(gdb) help set verbose
Set verbosity.
(gdb) set verbose on
(gdb) help set verbose
Set verbose printing of informational messages.
(gdb)
set_verbose creates the leak as it replaces the string allocated in
the above stacktrace by a static (non translated) string:
...
if (info_verbose)
{
c->doc = "Set verbose printing of informational messages.";
...
Also, this can possibly trigger a call to 'free' of a static string,
as c->doc_allocated is kept true, while the string is not allocated anymore.
This patch:
* fixes the leak by freeing the previous docs if doc_allocated.
* internationalize the messages.
* properly sets doc_allocated to 0 once doc strings are static.
gdb/ChangeLog
2018-12-28 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* top.c (set_verbose): Free previous docs if doc_allocated.
Internationalize messages. Set doc_allocated to 0.
Eli Zaretskii [Fri, 28 Dec 2018 07:02:04 +0000 (09:02 +0200)]
Avoid internal errors when stepping outside 'main' on MinGW
When one steps with "next" past the 'main's 'return' statement
in MinGW programs built by mingw.org's tools, PC lands in a
function whose symbol is not in any symtab. GDB then looks
up the nearest symbol, and should find none, because all those
with addresses below PC are not real functions. Having
unresolved symbols, whose address is zero, in minsyms tricked
GDB into using these bogus symbols, which then caused
assertion violation and internal_error. See the discussion at
https://sourceware.org/ml/gdb-patches/2018-12/msg00176.html
for more details.
gdb/ChangeLog
2018-12-28 Eli Zaretskii <eliz@gnu.org>
* coffread.c (coff_symtab_read): Don't record in minsyms symbols
that are unresolved. This avoids triggering an internal error
when stepping outside of 'main' in MinGW programs.
Alan Modra [Thu, 27 Dec 2018 23:11:44 +0000 (09:41 +1030)]
PR24015, glibc-2.28 on little-endian mips32 broken
Commit 2bf2bf23da exposed a bug on targets that create common sections
other than the standard ELF SHN_COMMON. If these are output by ld -r,
then their type becomes SHT_PROGBITS unless the target handles them
specially (eg. by elf_backend_special_sections), and if they are
merged into .bss/.sbss by ld -r then that section becomes SHT_PROGBITS.
Worse, if they are output by ld -r, then their size is increased by
bfd_generic_define_common_symbol during final link, which leads to
bogus file contents being copied to output.
For mips, it seems to me that the .scommon section should not be
output for ld -r, but I haven't made that change in this patch.
PR 24015
* elf.c (bfd_elf_get_default_section_type): Make common sections
SHT_NOBITS.
* linker.c (bfd_generic_define_common_symbol): Clear
SEC_HAS_CONTENTS.
Alan Modra [Thu, 27 Dec 2018 23:04:28 +0000 (09:34 +1030)]
PR23966, mingw failure due to 32-bit long
PR 23966
* libbfd.c (SSIZE_MAX): Define.
(bfd_malloc, bfd_realloc): Don't cast size to long to check for
"negative" values, compare against SSIZE_MAX instead.
Alan Modra [Thu, 27 Dec 2018 23:01:42 +0000 (09:31 +1030)]
PR24028, PPC_INT_FMT
PPC_INT_FMT is redundant now that bfd.h pulls in inttypes.h if
available. Apparently MacOS Mojave defines int64_t as long long even
though long is also 64 bits, which confuses the logic selecting
PPC_INT_FMT (and BFD_PRI64 too). Hopefully inttypes.h is available on
Mojave.
Tom Tromey [Tue, 25 Dec 2018 19:38:01 +0000 (12:38 -0700)]
Translate PyExc_KeyboardInterrupt to gdb "quit"
A while back I typed "info pretty-printers" with a large number of
printers installed, and I typed "q" to stop the pagination. I noticed
that gdb printed a Python exception in this case.
It seems to me that, instead, quitting pagination (or control-c'ing a
Python command generally) should be handled the same way that gdb
normally handles a quit.
This patch implements this idea by changing gdbpy_handle_exception to
treat PyExc_KeyboardInterrupt specially.
gdb/ChangeLog
2018-12-27 Tom Tromey <tom@tromey.com>
* python/py-utils.c (gdbpy_handle_exception): Translate
PyExc_KeyboardInterrupt to quit.
gdb/testsuite/ChangeLog
2018-12-27 Tom Tromey <tom@tromey.com>
Tom Tromey [Tue, 25 Dec 2018 18:44:58 +0000 (11:44 -0700)]
Consolidate some Python exception-printing functions
A few places in the Python code would either call gdbpy_print_stack,
or throw a gdb "quit", depending on the pending exception. This patch
consolidates these into a helper function.
gdb/ChangeLog
2018-12-27 Tom Tromey <tom@tromey.com>
* python/python-internal.h (gdbpy_print_stack_or_quit): Declare.
* python/py-unwind.c (pyuw_sniffer): Use
gdbpy_print_stack_or_quit.
* python/py-framefilter.c (throw_quit_or_print_exception):
Remove.
(gdbpy_apply_frame_filter): Use gdbpy_print_stack_or_quit.
* python/python.c (gdbpy_print_stack_or_quit): New function.
Tom Tromey [Tue, 25 Dec 2018 18:00:21 +0000 (11:00 -0700)]
Use gdbpy_convert_exception in a few more spots
I noticed a few places were converting a gdb exception to a Python
exception "by hand". It's better to use the existing
gdbpy_convert_exception helper function, as this handles memory errors
correctly, and in the future may be enhanced in other ways.
gdb/ChangeLog
2018-12-27 Tom Tromey <tom@tromey.com>
* python/py-value.c (convert_value_from_python): Use
gdbpy_convert_exception.
* python/py-param.c (parmpy_init): Use gdbpy_convert_exception.
* python/py-cmd.c (cmdpy_init): Use gdbpy_convert_exception.
* python/py-breakpoint.c (bppy_init): Use
gdbpy_convert_exception.
Tom Tromey [Mon, 24 Dec 2018 19:01:34 +0000 (12:01 -0700)]
Build gdb "nat" files in subdirectory
This moves the various "nat" object files into the nat/ subdirectory.
This allows for the removal of a pattern rule from the gdb Makefile,
which is a small cleanup.
I made the configure.nat change in a (semi-) automated way, hopefully
meaning that it is more likely to be correct than had I done it by
hand.
Eventually I would like for the various configure scripts to only
mention source files, and let the Makefile compute the object file
names.
gdb/ChangeLog
2018-12-27 Tom Tromey <tom@tromey.com>
* configure.nat (NATDEPFILES): Use nat/ prefix.
* Makefile.in (CONFIG_SRC_SUBDIR): Add nat.
(%.o: ${srcdir}/nat/%.c): Remove rule.
(INIT_FILES): Do not filter out NATDEPFILES.
Tom Tromey [Mon, 24 Dec 2018 18:44:10 +0000 (11:44 -0700)]
Make init.c depend on source files
I noticed that init.c depends on the object files that go into gdb.
Because init.c actually only requires the contents of the
corresponding source files, this unnecessarily serializes the step
that builds init.c.
This patch changes gdb's Makefile to make init.c depend on the source
files. This also simplifies the rule to build init.c.
gdb/ChangeLog
2018-12-27 Tom Tromey <tom@tromey.com>
* Makefile.in (INIT_FILES): Redefine.
(stamp-init): Remove sed, tr invocations. Use for loop. Don't
set LANG or LC_ALL.
Tom Tromey [Mon, 24 Dec 2018 17:07:51 +0000 (10:07 -0700)]
Remove gdbtypes special case from init.c rule
The rule to make init.c has a special case for gdbtypes, with a long
explanatory comment. All of this is obsolete, as the globals referred
to by the comment no longer exist. This patch simplifies the rule.
gdb/ChangeLog
2018-12-27 Tom Tromey <tom@tromey.com>
* Makefile.in (stamp-init): Remove gdbtypes special case.
Fix gdb.ada/fun_renaming.exp by using more unique names.
The test fails due to conflict between var 'next' and s-pooloc.adb next:
(gdb) print next(1)
Multiple matches for next
[0] cancel
[1] pack.next (integer) return integer at /bd/home/philippe/gdb/git/binutils-gdb/gdb/testsuite/gdb.ada/fun_renaming/pack.adb:19
[2] system.pool_local.next (system.address) return system.pool_local.acc_address at s-pooloc.adb:151
> FAIL: gdb.ada/fun_renaming.exp: print next(1) (timeout)
Fix by making the names and renamings more unique.
gdb/testsuite/ChangeLog
2018-12-26 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* gdb.ada/fun_renaming/pack.ads (Next): Rename to Fun_Rename_Test_Next.
(Renamed_Next): Rename to Renamed_Fun_Rename_Test_Next.
gdb.ada/fun_renaming/pack.adb (Next): Rename to Fun_Rename_Test_Next.
gdb.ada/fun_renaming/fun_renaming.adb (N): Rename to Fun_Rename_Test_N.
gdb.ada/fun_renaming.exp: Update accordingly.
Improve gdb.ada/rename_subscript_param.exp by using more unique names.
With old compilers, the test fails because no debug info is generated
for 'B' and GDB finds some 'b' in atnat.h:
(gdb) print b
Multiple matches for b
[0] cancel
[1] b at ../sysdeps/ieee754/dbl-64/atnat.h:106
[2] b at ../sysdeps/ieee754/dbl-64/atnat.h:106
[3] b at ../sysdeps/ieee754/dbl-64/atnat.h:106
> FAIL: gdb.ada/rename_subscript_param.exp: print b before changing its value (timeout)
Avoid the timeout by renaming 'b' to rename_subscript_param_b.
Also, change 'before' to 'after' in the gdb_test message that prints
the value after changing it.
The test still fails with old compilers that do not properly
generate debug info for this renaming:
(gdb) print rename_subscript_param_b
No definition of "rename_subscript_param_b" in current context.
(gdb) FAIL: gdb.ada/rename_subscript_param.exp: print rename_subscript_param_b before changing its value
Note: if the compiler would generate the correct debug info, the test should
succeed with the name B. However, waiting for this fix, changing the name
ensures that the test fails directly, instead of causing a timeout.
2018-12-26 Philippe Waroquiers <philippe.waroquiers@skynet.be>
PR ada/23381
* gdb.ada/rename_subscript_param/pkg.adb (B): Rename to
Rename_Subscript_Param_B. All users updated.
gdb.ada/rename_subscript_param.exp: Test names made unique.
Note that PR ada/23381 is only fully fixed when using a recent
compiler.
Fix gdb.ada/packed_array_assign.exp by using more unique names.
The test gdb.ada/packed_array_assign fails due to conflict between component 'w'
and system.dim.mks.w:
(gdb) print pra := ((x => 2, y => 0, w => 17), pr, (x => 7, y => 1, w => 23))
Unknown component name: system.dim.mks.w.
(gdb) FAIL: gdb.ada/packed_array_assign.exp: print pra := ((x => 2, y => 0, w => 17), pr, (x => 7, y => 1, w => 23))
Also, depending on the compiler version, the component w might be reordered
and placed before components x and y.
So, change the component order in the source, so that both an old
compiler (GNATMAKE 6.3.0, gcc (Debian 6.3.0-18+deb9u1) 6.3.0 20170516)
and a new compiler (GNATMAKE Pro 20.0w (20181210-82), based on gcc 8.2.1)
produce the same component order (checked by using -gnatR3s).
So, update to test the new (more unique) names in the source order.
2018-12-26 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* gdb.ada/packed_array_assign/aggregates.ads (Packed_Rec):
Rename components to Packed_Array_Assign_[X|Y|W]. Place
component Packed_Array_Assign_W as first component, to ensure
old and new compilers have the same representation.
All users updated.
Simon Marchi [Thu, 27 Dec 2018 01:14:08 +0000 (20:14 -0500)]
target.c: Remove struct keyword in range-based for
I get this when compiling with a gcc 6.3.0-based cross-compiler:
CXX target.o
/home/simark/src/binutils-gdb/gdb/target.c: In static member function 'static void target_terminal::restore_inferior()':
/home/simark/src/binutils-gdb/gdb/target.c:396:10: error: types may not be defined in a for-range-declaration [-Werror]
for (struct inferior *inf : all_inferiors ())
^~~~~~
Accomodate it by dropping the unnecessary struct keyword. Actually, I used
"::inferior", otherwise it resolves to the inferior method of the
target_terminal class.
Simon Marchi [Wed, 26 Dec 2018 16:49:51 +0000 (11:49 -0500)]
Improve "set debug separate-debug-file"
"set debug separate-debug-file" shows which candidates are considered,
when trying to find separate debug info. But it's not clear if GDB used
a certain candidate, and if not, why not. This patch adds some
precision:
Before:
Looking for separate debug info (debug link) for /lib/x86_64-linux-gnu/libc.so.6
Trying /lib/x86_64-linux-gnu/libc-2.23.so
Trying /lib/x86_64-linux-gnu/.debug/libc-2.23.so
Trying /usr/lib/debug//lib/x86_64-linux-gnu/libc-2.23.so
After:
Looking for separate debug info (debug link) for /lib/x86_64-linux-gnu/libc.so.6
Trying /lib/x86_64-linux-gnu/libc-2.23.so... no, same file as the objfile.
Trying /lib/x86_64-linux-gnu/.debug/libc-2.23.so... no, unable to open.
Trying /usr/lib/debug//lib/x86_64-linux-gnu/libc-2.23.so... yes!
Andrew Burgess [Thu, 13 Dec 2018 18:25:25 +0000 (18:25 +0000)]
gdb: Allow struct fields named double
The 64-bit RISC-V target currently models the floating point registers
as having the following type:
union riscv_double
{
builtin_type_ieee_single float;
builtin_type_ieee_double double;
}
Notice the choice of names for the fields of this struct, possibly not
ideal choices, as these are not valid field names in C. However, this
type is only ever defined within GDB (or in the target description),
and no restriction seems to exist on the field names in that case.
The problem though is that currently:
(gdb) info registers $ft0
ft0 {float = 0, double = 0} (raw 0x0000000000000000)
(gdb) p $ft0.float
$1 = 0
(gdb) p $ft0.double
A syntax error in expression, near `double'.
We can access the 'float' field, but not the 'double' field. This is
because the string 'double' is handled differently to the string
'float' in c-exp.y.
In both cases the string '$ft0' is parsed as a VARIABLE expression.
In the 'float' case, the string 'float' becomes a generic NAME token
in 'lex_one_token', which then allows the rule "exp '.' name" to match
and the field name lookup to occur.
The 'double' case is different. In order to allow parsing of the type
string 'long double', the 'double' string becomes the token
DOUBLE_KEYWORD. At this point there's no rule to match "exp '.'
DOUBLE_KEYWORD", so we can never lookup the field named 'double'.
We could rename the fields for RISC-V, and maybe that would be the
best solution. However, its not hard to allow for fields named
'double', which is what this patch does.
A new case is added to the 'field_name' rule to match the
DOUBLE_KEYWORD, and create a suitable 'struct stoken'. With this done
the "exp '.' field_name" pattern can now match, and we can lookup the
double field.
With this patch in place I now see this behaviour:
(gdb) info registers $ft0
ft0 {float = 0, double = 0} (raw 0x0000000000000000)
(gdb) p $ft0.float
$1 = 0
(gdb) p $ft0.double
$2 = 0
I've gone ahead and handled INT_KEYWORD, LONG, SHORT, SIGNED_KEYWORD,
and UNSIGNED as well within field_name.
I've added a new test for this functionality.
This change was tested on x86-64 GNU/Linux with no regressions.
gdb/ChangeLog:
* c-exp.y (field_name): Allow DOUBLE_KEYWORD, INT_KEYWORD, LONG,
SHORT, SIGNED_KEYWORD, and UNSIGNED tokens to act as a field
names.
(typename_stoken): New function.
gdb/testsuite/ChangeLog:
* gdb.dwarf2/dw2-unusual-field-names.c: New file.
* gdb.dwarf2/dw2-unusual-field-names.exp: New file.
Andrew Burgess [Mon, 17 Dec 2018 11:21:08 +0000 (11:21 +0000)]
gdb: Extend the comments in c-exp.y
In an attempt to fix PR gdb/13368 this commit adds some comments to
c-exp.y which hopefully makes the type parsing code a little clearer.
There are no code changes here, so there should be no user visible
changes after this commit.
gdb/ChangeLog:
PR gdb/13368
* c-exp.y (typebase): Extend the comment.
(ident_tokens): Likewise.
Tom Tromey [Mon, 24 Dec 2018 16:55:10 +0000 (09:55 -0700)]
Simplify dwarf2_find_containing_comp_unit
In an earlier patch discussion we noticed that
dwarf2_find_containing_comp_unit takes the address of sect_off, but
doesn't actually need to. This is a leftover from before
C++-ification. This patch simplifies the function.
Tested using gdb.dwarf2 on x86-64 Fedora 28.
gdb/ChangeLog
2018-12-18 Tom Tromey <tom@tromey.com>
* dwarf2read.c (dwarf2_find_containing_comp_unit): Don't take
address of sect_off.
Fix gdb.ada bp_fun_addr failure due to conflict between fun 'a' and s-dimmks.ads 'A'.
The test fails (timeout) due to:
(gdb) PASS: gdb.ada/bp_fun_addr.exp: break *a'address
run
Starting program: /bd/home/philippe/gdb/git/build_info_t/gdb/testsuite/outputs/gdb.ada/bp_fun_addr/a
Multiple matches for a
[0] cancel
[1] a at /bd/home/philippe/gdb/git/info_t/gdb/testsuite/gdb.ada/bp_fun_addr/a.adb:18
[2] system.dim.mks.a at s-dimmks.ads:115
> FAIL: gdb.ada/bp_fun_addr.exp: run until breakpoint at a'address (timeout)
testcase /home/philippe/gdb/git/build_info_t/gdb/testsuite/../../../info_t/gdb/testsuite/gdb.ada/bp_fun_addr.exp completed in 10 seconds
Fix this by using a fun name that has more chances to be unique.
2018-12-24 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* gdb.ada/bp_fun_addr/a.adb (a): Rename to bp_fun_addr.
Filename a.adb changed to bp_fun_addr.adb.
gdb.ada/bp_fun_addr.exp: Update test accordingly.
Andrew Burgess [Tue, 27 Nov 2018 13:41:44 +0000 (13:41 +0000)]
gdb/riscv: Prevent buffer overflow in riscv_return_value
The existing code for reading and writing the return value can
overflow the passed in buffers in a couple of situations. This commit
aims to resolve these issues.
The problems were detected using valgrind, here are two examples,
first from gdb.base/structs.exp:
(gdb) p/x fun9()
==31353== Invalid write of size 8
==31353== at 0x4C34153: memmove (vg_replace_strmem.c:1270)
==31353== by 0x632EBB: memcpy (string_fortified.h:34)
==31353== by 0x632EBB: readable_regcache::raw_read(int, unsigned char*) (regcache.c:538)
==31353== by 0x659D3F: riscv_return_value(gdbarch*, value*, type*, regcache*, unsigned char*, unsigned char const*) (riscv-tdep.c:2593)
==31353== by 0x583641: get_call_return_value (infcall.c:448)
==31353== by 0x583641: call_thread_fsm_should_stop(thread_fsm*, thread_info*) (infcall.c:546)
==31353== by 0x59BBEC: fetch_inferior_event(void*) (infrun.c:3883)
==31353== by 0x53890B: check_async_event_handlers (event-loop.c:1064)
==31353== by 0x53890B: gdb_do_one_event() [clone .part.4] (event-loop.c:326)
==31353== by 0x6CA34B: wait_sync_command_done() (top.c:503)
==31353== by 0x584653: run_inferior_call (infcall.c:621)
...
And from gdb.base/call-sc.exp:
(gdb) advance fun
fun () at /gdb/gdb/testsuite/gdb.base/call-sc.c:41
41 return foo;
(gdb) finish
==1968== Invalid write of size 8
==1968== at 0x4C34153: memmove (vg_replace_strmem.c:1270)
==1968== by 0x632EBB: memcpy (string_fortified.h:34)
==1968== by 0x632EBB: readable_regcache::raw_read(int, unsigned char*) (regcache.c:538)
==1968== by 0x659D01: riscv_return_value(gdbarch*, value*, type*, regcache*, unsigned char*, unsigned char const*) (riscv-tdep.c:2576)
==1968== by 0x5891E4: get_return_value(value*, type*) (infcmd.c:1640)
==1968== by 0x5892C4: finish_command_fsm_should_stop(thread_fsm*, thread_info*) (infcmd.c:1808)
==1968== by 0x59BBEC: fetch_inferior_event(void*) (infrun.c:3883)
==1968== by 0x53890B: check_async_event_handlers (event-loop.c:1064)
==1968== by 0x53890B: gdb_do_one_event() [clone .part.4] (event-loop.c:326)
==1968== by 0x6CA34B: wait_sync_command_done() (top.c:503)
...
There are a couple of problems with the existing code, that are all
related.
In riscv_call_arg_struct we incorrectly rounded up the size of a
structure argument. This is unnecessary, and caused GDB to read too
much data into the output buffer when extracting a struct return
value.
In fixing this it became clear that we were incorrectly assuming that
any value being placed in a register (or read from a register) would
always access the entire register. This is not true, for example a
9-byte struct on a 64-bit target places 8-bytes in one registers and
1-byte in a second register (assuming available registers). To handle
this I switch from using cooked_read to cooked_read_part.
Finally, when processing basic integer return value types these are
extended to xlen sized types and then passed in registers. We
currently don't handle this type expansion in riscv_return_value, but
we do in riscv_push_dummy_call. The result is that small integer
types (like char) result in a full xlen sized register being written
into the output buffer, which results in buffer overflow. To address
this issue we now create a value of the expanded type and use this
values contents buffer to hold the return value before casting the
value down to the smaller expected type.
This patch resolves all of the valgrind issues I have found so far,
and causes no regressions. Tested against RV32/64 with and without
floating point support.
gdb/ChangeLog:
* riscv-tdep.c (riscv_call_arg_struct): Don't adjust size before
assigning locations.
(riscv_return_value): Take more care not to read/write outside of
argument buffer. Cast return value between the declared type and
the abi type.
Andrew Burgess [Thu, 13 Dec 2018 19:06:23 +0000 (19:06 +0000)]
gdb/riscv: Add float status registers to save and restore reggroups
We should save and restore the floating point status registers. This
became an issue when testing 32-bit float on a target with 64-bit with
the gdb.base/callfuncs.exp test.
gdb/ChangeLog:
* riscv-tdep.c (riscv_register_reggroup_p): Save and restore fcsr,
fflags, and frm registers.
Andrew Burgess [Thu, 13 Dec 2018 17:57:14 +0000 (17:57 +0000)]
gdb/riscv: Add gdb to dwarf register number mapping
Provide a mapping between GDB's register numbers and DWARF's register
numbers. This resolves some failures that I was seeing on
gdb.base/store.exp when running on an rv64imfdc target.
gdb/ChangeLog:
* riscv-tdep.c (riscv_dwarf_reg_to_regnum): New function.
(riscv_gdbarch_init): Register new function with gdbarch.
* riscv-tdep.h: New enum to define RISC-V DWARF register numbers.
I added a printf gated by "set debug symtab-create" to be able to
quickly see all minimal symbols recorded by GDB. I thought it would be
useful to have it built-in, for the future. Here's how the output
looks:
Jan Vrany [Thu, 13 Dec 2018 15:20:49 +0000 (15:20 +0000)]
Fix various tests to use -no-pie linker flag when needed
Various tests use test code written in i386 / x86_64 assembly that cannot
be used to create PIE executables. Therefore compilation of test programs
failed on systems where the compiler default is to create PIE executable.
The solution is to use -no-pie linker flag, however, such flag may not
(is not) supported by all compilers GDB needs to support (e.g. gcc 4.8).
To handle this, introduce a new flag to gdb_compile - nopie - which
inserts -no-pie linker flag where supported and is no-op where it is
not. By default, -no-pie flag is inserted since most modern compiler do
support it.
John Baldwin [Fri, 21 Dec 2018 18:18:11 +0000 (10:18 -0800)]
Workaround a FreeBSD kernel bug resulting in spurious SIGTRAP events.
The ptrace command PT_LWPINFO to request detailed information about a
stopped thread can return stale signal information from an earlier
stop. Events which are reporting an intercepted signal will always
report the correct information, but signal stops for some other events
such as system call enter/exit events might include stale siginfo from
an earlier signal. In particular, if a thread reports a system call
entry or exit event after previously reporting a single-step or
breakpoint event via SIGTRAP, fbsd_handle_debug_trap believed the
system call event was the previous event and claimed it resulting in a
spurious SIGTRAP event.
True breakpoint and single-step events will never report another event
in the pl_flags member of struct ptrace_lwpinfo. Use this to detect
stale siginfo by requiring pl_flags to have only the PL_FLAG_SI flag
and no other flags before treating a SIGTRAP as a single-step or
breakpoint trap.
gdb/ChangeLog:
* fbsd-nat.c (fbsd_handle_debug_trap): Require pl.pl_flags to
equal PL_FLAG_SI.
(fbsd_nat_target::stopped_by_sw_breakpoint): Likewise.
Paul Marechal [Fri, 21 Dec 2018 17:02:33 +0000 (12:02 -0500)]
gdb: Fix "info os <unknown>" command
Running `info os someUnknownOsType` is crashing when gdb is built with
-D_GLIBCXX_DEBUG:
/usr/include/c++/5/debug/vector:439:error: attempt to
access an element in an empty container.
In target_read_stralloc from target.c, the call to
target_read_alloc_1 can return an empty vector, we then call vector::back on
this vector, which is invalid.
This commit adds a check for emptiness before trying to call
vector::back on it. It also adds test to check for `info os <unknown>`
to return the proper error message.
This is a regression in gdb 8.2 and this patch restores the behavior of
previous versions.
gdb/ChangeLog:
PR gdb/23974
* target.c (target_read_stralloc): Check for empty vector.
Дилян Палаузов [Fri, 21 Dec 2018 15:09:40 +0000 (19:09 +0400)]
when printing the GDB config, explicitly say if configured without python
When using the --configuration command line switch, or using
the "show configuration" command with a version of GDB which
was configured without Python supoprt, this patch changes
the resulting output to include...
--without-python
... instead of not printing anything about Python support.
gdb/ChangeLog:
* top.c (print_gdb_configuration): Print "--without-python"
if GDB was configured without Python.
Tested on x86_64-linux by rebuilding GDB with and without Python,
and checking the output of "gdb --configuration" in both cases.
Dave Murphy [Fri, 21 Dec 2018 16:14:28 +0000 (11:14 -0500)]
Fix compile error with clang 3.8
When compiling with clang 3.8 (default clang version on Debian
Stretch, the current stable), we get errors like this:
CXX dtrace-probe.o
../../binutils-gdb/gdb/dtrace-probe.c:103:31: error: default initialization of an object of const type 'const dtrace_static_probe_ops' without a user-provided default constructor
const dtrace_static_probe_ops dtrace_static_probe_ops;
^
Silence them by value-initializing those objects. It's not necessary
with other compilers (later clang versions, gcc), but it shouldn't
hurt either.
H.J. Lu [Thu, 20 Dec 2018 21:25:41 +0000 (13:25 -0800)]
x86: Call rtype_to_howto to get reloc_howto_type pointer
* elf32-i386.c (elf_i386_relocate_section): Call
elf_i386_rtype_to_howto to get reloc_howto_type pointer.
* elf64-x86-64.c (elf_x86_64_relocate_section): Call
elf_x86_64_rtype_to_howto to get reloc_howto_type pointer.
Ensure deterministic result order in gdb.ada/info_auto_lang.exp
standard_ada_testfile, standard_test_file and the explicit
csrcfile assignment in info_auto_lang.exp all gives similar pathnames
prefix for a source, such as
/home/philippe/gdb/git/build_binutils-gdb/gdb/testsuite/../../../binutils-gdb/gdb/testsuite/gdb.<something>.
Note that the above pathnames contain ../ which appears when a relative
pathname is used to call configure.
In any case, the gnat compiler normalizes Ada sources path when compiling.
So, the 'Ada' .o object are referencing a pathname such as
/home/philippe/gdb/git/binutils-gdb/gdb/testsuite/gdb.ada/info_auto_lang/proc_in_ada.adb,
while the 'C' .o object still references the not normalized pathname.
As the results of 'info functions | ...' are sorted by pathname first,
the order of the results depends on the comparison between different directories,
leading to results that can change depending on these directories.
=> Ensure the result order is always the same, by normalising the C source file,
which makes the results independent of the way configure is launched.
Tested by running the testcase in 2 different builds, that without normalize
were giving different results.
Note: such 'set csrcfile' is used in 4 other tests mixing Ada and C.
After discussion, it was deemed sufficient to just normalize the pathname
for this test.
gdb/testsuite/ChangeLog
2018-12-20 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* gdb.ada/info_auto_lang.exp: Normalize some_c source file.
Update order of results accordingly.
Andrew Burgess [Fri, 7 Sep 2018 19:04:44 +0000 (20:04 +0100)]
gdb: Add default frame methods to gdbarch
Supply default gdbarch methods for gdbarch_dummy_id,
gdbarch_unwind_pc, and gdbarch_unwind_sp. This patch doesn't actually
convert any targets to use these methods, and so, there will be no
user visible changes after this commit.
The implementations for default_dummy_id and default_unwind_sp are
fairly straight forward, these just take on the pattern used by most
targets. Once these default methods are in place then most targets
will be able to switch over.
The implementation for default_unwind_pc is also fairly straight
forward, but maybe needs some explanation.
This patch has gone through a number of iterations:
This is basically a clone of default_unwind_sp, but using $pc. It was
pointed out that we could potentially do better, and in version 2 the
implementation became:
CORE_ADDR
default_unwind_pc (struct gdbarch *gdbarch, struct frame_info *next_frame)
{
struct type *type;
int pc_regnum;
CORE_ADDR addr;
struct value *value;
The idea was to try split out some of the steps of unwinding the $pc,
steps that are on some (or many) targets no-ops, and so allow targets
that do override these methods, to make use of default_unwind_pc.
This implementation remained in place for version 2, 3, and 4.
However, I realised that I'd made a mistake, most targets simply use
frame_unwind_register_unsigned to unwind the $pc, and this throws an
error if the register value is optimized out or unavailable. My new
proposed implementation doesn't do this, I was going to end up
breaking many targets.
I considered duplicating the code from frame_unwind_register_unsigned
that throws the errors into my new default_unwind_pc, however, this
felt really overly complex. So, what I instead went with was to
simply revert back to using frame_unwind_register_unsigned. Almost
all existing targets already use this. Some of the ones that don't can
be converted to, which means almost all targets could end up using the
default.
One addition I have made over the version 1 implementation is to add a
call to gdbarch_addr_bits_remove. For most targets this is a no-op,
but for a handful, having this call in place will mean that they can
use the default method. After all this, the new default_unwind_pc now
looks like this:
CORE_ADDR
default_unwind_pc (struct gdbarch *gdbarch, struct frame_info *next_frame)
{
int pc_regnum = gdbarch_pc_regnum (gdbarch);
CORE_ADDR pc = frame_unwind_register_unsigned (next_frame, pc_regnum);
pc = gdbarch_addr_bits_remove (gdbarch, pc);
return pc;
}
gdb/ChangeLog:
* gdb/dummy-frame.c (default_dummy_id): Defined new function.
* gdb/dummy-frame.h (default_dummy_id): Declare new function.
* gdb/frame-unwind.c (default_unwind_pc): Define new function.
(default_unwind_sp): Define new function.
* gdb/frame-unwind.h (default_unwind_pc): Declare new function.
(default_unwind_sp): Declare new function.
* gdb/frame.c (frame_unwind_pc): Assume gdbarch_unwind_pc is
available.
(get_frame_sp): Assume that gdbarch_unwind_sp is available.
* gdb/gdbarch.c: Regenerate.
* gdb/gdbarch.h: Regenerate.
* gdb/gdbarch.sh: Update definition of dummy_id, unwind_pc, and
unwind_sp. Add additional header files to be included in
generated file.
H.J. Lu [Wed, 19 Dec 2018 20:21:56 +0000 (12:21 -0800)]
x86: Properly handle PLT expression in directive
For PLT expressions, we should subtract the PLT relocation size only for
jump instructions. Since PLT relocations are PC relative, we only allow
"symbol@PLT" in PLT expression.
gas/
PR gas/23997
* config/tc-i386.c (x86_cons): Check for invalid PLT expression.
(md_apply_fix): Subtract the PLT relocation size only for jump
instructions.
* testsuite/gas/i386/reloc32.s: Add test for invalid PLT
expression.
* testsuite/gas/i386/reloc64.s: Likewise.
* testsuite/gas/i386/ilp32/reloc64.s: Likewise.
* testsuite/gas/i386/reloc32.l: Updated.
* testsuite/gas/i386/reloc64.l: Likewise.
* testsuite/gas/i386/ilp32/reloc64.l: Likewise.
Dimitar Dimitrov [Mon, 17 Dec 2018 19:30:52 +0000 (21:30 +0200)]
Fix build with latest GCC 9.0 tree
A recent patch [1] to fix a GCC PR [2] actually broke the GDB build.
To fix, remove the stack pointer clobber. GCC will ignore the clobber
marker, and will not save or restore the stack pointer.
I ran "make check-gdb" on x86_64 to ensure there are no regressions.
Alan Modra [Tue, 18 Dec 2018 08:33:51 +0000 (19:03 +1030)]
Include bfd_stdint.h in bfd.h
This patch adds bfd_stdint.h to bfd.h, so that BFD can use size_t
where appropriate in function parameters and return values. I also
tidy a few other cases where headers are included twice.
bfd/
* Makefile.am (bfdinclude_HEADERS): Add bfd_stdint.h.
(BFD_H_DEPS): Add include/diagnostics.h.
(LOCAL_H_DEPS): Add bfd_stdint.h.
* bfd-in.h: Include bfd_stdint.h.
* arc-plt.h: Don't include stdint.h.
* coff-rs6000.c: Likewise.
* coff64-rs6000.c: Likewise.
* elfxx-riscv.c: Likewise.
* cache.c: Don't include bfd_stdint.h.
* elf32-arm.c: Likewise.
* elf32-avr.c: Likewise.
* elf32-nds32.c: Likewise.
* elf32-rl78.c: Likewise.
* elf32-rx.c: Likewise.
* elf32-wasm32.c: Likewise.
* elf64-nfp.c: Likewise.
* elflink.c: Likewise.
* elfnn-aarch64.c: Likewise.
* elfnn-ia64.c: Likewise.
* elfxx-ia64.c: Likewise.
* elfxx-x86.h: Likewise.
* wasm-module.c: Likewise, and don't include sysdep.h twice.
* elf-nacl.h: Don't include bfd.h.
* mach-o.h: Likewise.
* elfxx-aarch64.c: Include bfd.h and elf-bfd.h.
* elfxx-aarch64.h: Don't include bfd.h, elf-bfd.h or stdint.h.
* mach-o-aarch64.c: Include mach-o.h later.
* mach-o-arm.c: Likewise.
* mach-o-i386.c: Likewise.
* mach-o-x86-64.c: Likewise.
* mach-o.c: Likewise.
* sysdep.h: Don't include ansidecl.h or sys/stat.h.
* Makefile.in: Regenerate.
* bfd-in2.h: Regenerate.
opcodes/
* arm-dis.c: Include bfd.h.
* aarch64-opc.c: Include bfd_stdint.h rather than stdint.h.
* csky-dis.c: Likewise.
* nds32-asm.c: Likewise.
* riscv-dis.c: Likewise.
* s12z-dis.c: Likewise.
* wasm32-dis.c: Likewise.
Alan Modra [Tue, 18 Dec 2018 00:17:44 +0000 (10:47 +1030)]
[GOLD] Tweak keep_text_section_prefix test for PowerPC64 ELFv1
This test checks code layout by function symbol ordering, but that
doesn't work on powerpc64 ELFv1 where the function symbol is on a
descriptor. A simple work-around is to have nm emit synthetic symbols
marking the code entry point of functions. Since the text segment is
laid out before the data segment, the synthetic symbols will have
lower addresses than function descriptor symbols and be seen first in
nm -n output.
On other targets, nm --synthetic typically emits symbols on plt
entries. Since the testcase doesn't call any of the functions of
interest there shouldn't be plt entries for those functions, so there
should be no potentially confusing extra symbols.
* testsuite/Makefile.am (keep_text_section_prefix_nm.stdout):
Pass --synthetic to nm.
* testsuite/Makefile.in: Regenerate.
Alan Modra [Mon, 17 Dec 2018 22:50:06 +0000 (09:20 +1030)]
PR23980, assertion fail
All of the backend relocate_section functions that interpret reloc
numbers assuming the input file is of the expected type (ie. same as
output or very similar) really ought to be checking input file type.
Not many do, and those that do currently just assert. This patch
replaces the assertion with a more graceful exit.
PR 23980
* elf32-i386.c (elf_i386_relocate_section): Exit with wrong format
error rather than asserting input file is as expected.
* elf32-s390.c (elf_s390_relocate_section): Likewise.
* elf32-sh.c (sh_elf_relocate_section): Likewise.
* elf32-xtensa.c (elf_xtensa_relocate_section): Likewise.
* elf64-ppc.c (ppc64_elf_relocate_section): Likewise.
* elf64-s390.c (elf_s390_relocate_section): Likewise.
* elf64-x86-64.c (elf_x86_64_relocate_section): Likewise.
* elf32-ppc.c (ppc_elf_relocate_section): Exit with wrong format
error if input file is not ppc32 ELF.
Andrew Burgess [Mon, 5 Feb 2018 11:15:38 +0000 (11:15 +0000)]
sim: Don't overwrite stored errno in sim_syscall_multi
The host syscall callback mechanism should take care of updating the
errcode within the CB_SYSCALL struct, and we should not be adjusting
the error code once the syscall has completed. We especially, should
not be rewriting the syscall errcode based on the value of errno some
time after running the host syscall, as there is no guarantee that
errno has not be overwritten.
To perform a syscall we call cb_syscall (in syscall.c). To return
from cb_syscall control passes through one of two exit paths these are
labeled FinishSyscall and ErrorFinish and are reached using goto
statements scattered throughout the cb_syscall function.
In FinishSyscall we store the syscall result in 'sc->result', and the
error code is transated to target encoding, and stored in
'sc->errcode'.
In ErrorFinish, we again store the syscall result in 'sc->result', and
fill in 'sc->errcode' by fetching the actual errno from the host with
the 'cb->get_errno' callback.
In both cases 'sc->errcode' will have been filled in with an
appropriate value.
Further, if we look at a specific syscall example, CB_SYS_open, in
this case the first thing we do is fetch the path to open from the
target with 'get_path', if this fails then the errcode is returned,
and we jump to FinishSyscall. Notice that in this case, no host
syscall may have been performed, for example a failure to read the
path to open out of simulated memory can return EINVAL without
performing any host syscall. Given that no host syscall has been
performed, reading the host errno makes absolutely no sense.
This commit removes from sim_syscall_multi the rewriting of
sc->errcode based on the value of errno, and instead relies on the
value stored in the cb_syscall.
sim/common/ChangeLog:
* sim-syscall.c (sim_syscall_multi): Don't update sc->errcode at
this point, it should have already been set in cb_syscall.
Andrew Burgess [Wed, 31 Oct 2018 13:39:58 +0000 (13:39 +0000)]
gdb/dwarf: Convert some predicates from int to bool
In the dwarf reader we have a set of predicates, these include the
different producer predicates and also some control predicates. The
older ones are declared as integers, while newer ones (added since the
C++ conversion) are bool.
This commit makes them all bool for consistency. There should be no
user visible change after this commit.
gdb/ChangeLog:
* dwarf2read.c (struct dwarf2_cu): Convert the fields 'mark',
'has_loclist', 'checked_producer', 'producer_is_gxx_lt_4_6',
'producer_is_gcc_lt_4_3', 'producer_is_icc_lt_14',
'processing_has_namespace_info' from unsigned int to bool. Update
comments.
(producer_is_icc_lt_14): Update return type.
(producer_is_gcc_lt_4_3): Likewise.
(producer_is_gxx_lt_4_6): Likewise.
(process_die): Write true instead of 1 into predicate fields.
(dwarf2_start_symtab): Likewise.
(var_decode_location): Likewise.
(dwarf2_mark_helper): Likewise.
(dwarf2_mark): Likewise.
(dwarf2_clear_marks): Write false instead of 0 into predicate
field.
(dwarf2_cu::dwarf2_cu): Initialise predicate fields to false, not
0.
Szabolcs Nagy [Thu, 13 Dec 2018 17:47:17 +0000 (17:47 +0000)]
AArch64: Fix the gdb build with musl libc
Including asm/sigcontext.h together with libc headers is not valid. In
general linux headers may not work with libc headers, so mixing them
should be avoided, especially when the linux header defines types that
are also exposed in libc headers.
In case of asm/sigcontext.h glibc happens to work because glibc signal.h
directly includes it, but e.g. in musl libc signal.h replicates the
sigcontext.h definitions in an abi compatible way which are in conflict
with the linux definitions when both headers are included.
Since old linux headers or old libc headers may not have the necessary
definitions, gdb has to replicate the definitions it relies on anyway.
Which is fine since all definitions must be ABI stable. For linux apis
that are not available via libc headers, replicating the definitions in
gdb is the most reliable way to use them.
Note: asm/ptrace.h includes asm/sigcontext.h in some versions of linux
headers, which is just as problematic and should be fixed in linux.
gdb/ChangeLog:
* nat/aarch64-sve-linux-ptrace.h: Include signal.h instead of
asm/sigcontext.h.
Alan Modra [Sun, 16 Dec 2018 12:32:50 +0000 (23:02 +1030)]
PR23994, libbfd integer overflow
PR 23994
* aoutx.h: Include limits.h.
(get_reloc_upper_bound): Detect long overflow and return a file
too big error if it occurs.
* elf.c: Include limits.h.
(_bfd_elf_get_symtab_upper_bound): Detect long overflow and return
a file too big error if it occurs.
(_bfd_elf_get_dynamic_symtab_upper_bound): Likewise.
(_bfd_elf_get_dynamic_reloc_upper_bound): Likewise.
Factorize killing the children in linux-ptrace.c, and fix a 'process leak'.
Running the gdb testsuite under Valgrind started to fail after 100+ tests,
due to out of memory caused by lingering processes.
The lingering processes are caused by the combination
of a limitation in Valgrind signal handling when using PTRACE_TRACEME
and a (minor) bug in GDB.
The Valgrind limitation is : when a process is ptraced and raises
a signal, Valgrind will replace the raised signal by SIGSTOP as other
signals are masked by Valgrind when executing a system call.
Removing this limitation seems far to be trivial, valgrind signal
handling is very complex.
Due to this valgrind limitation, GDB linux_ptrace_test_ret_to_nx gets
a SIGSTOP signal instead of the expected SIGTRAP or SIGSEGV.
In such a case, linux_ptrace_test_ret_to_nx does an early return, but
does not kill the child (running under valgrind), child stays in a STOP-ped
state.
These lingering processes then eat the available system memory,
till launching a new process starts to fail.
This patch fixes the GDB minor bug by killing the child in case
linux_ptrace_test_ret_to_nx does an early return.
nat/linux-ptrace.c has 3 different logics to kill a child process.
So, this patch factorizes killing a child in the function kill_child.
The 3 different logics are:
* linux_ptrace_test_ret_to_nx is calling both kill (child, SIGKILL)
and ptrace (PTRACE_KILL, child, ...), and then is calling once
waitpid.
* linux_check_ptrace_features is calling ptrace (PTRACE_KILL, child, ...)
+ my_waitpid in a loop, as long as the waitpid status was WIFSTOPPED.
* linux_test_for_tracefork is calling once ptrace (PTRACE_KILL, child, ...)
+ my_waitpid.
The linux ptrace documentation indicates that PTRACE_KILL is deprecated,
and tells to not use it, as it might return success but not kill the tracee.
The documentation indicates to send SIGKILL directly.
I suspect that linux_ptrace_test_ret_to_nx calls both kill and ptrace just
to be sure ...
I suspect that linux_check_ptrace_features calls ptrace in a loop
to bypass the PTRACE_KILL limitation.
And it looks like linux_test_for_tracefork does not handle the PTRACE_KILL
limitation.
Also, 2 of the 3 logics are calling my_waitpid, which seems better,
as this is protecting the waitpid syscall against EINTR.
So, the logic in kill_child is just using kill (child, SIGKILL)
+ my_waitpid, and then does a few verifications to see everything worked
accordingly to the plan.
Tested on Debian/x86_64.
2018-12-16 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* nat/linux-ptrace.c (kill_child): New function.
(linux_ptrace_test_ret_to_nx): Use kill_child instead of local code.
Add a call to kill_child in case of early return after fork.
(linux_check_ptrace_features): Use kill_child instead of local code.
(linux_test_for_tracefork): Likewise.
Tom Tromey [Sat, 15 Dec 2018 01:58:32 +0000 (18:58 -0700)]
Minor gdb/Makefile.in cleanups
This removes an IMO not very useful comment in gdb/Makefile.in about
"alloca". It also removes INFOFILES, which I think probably has not
been useful since whenever the manual was moved into a subdirectory.
gdb/ChangeLog
2018-12-14 Tom Tromey <tom@tromey.com>
projects / thirdparty / binutils-gdb.git / log
