These NEWS entries had been missed when the instructions were added.
gas/ChangeLog:
2019-09-19 Matthew Malcomson <matthew.malcomson@arm.com>
* NEWS: Add SVE2 and TME entries.
The newly added objdump -S tests check for source line mapping of a static
variable. But the test doesn't dump any data sections so this should never
pass.
This changes the test to dump all sections so the source mapping can be found.
binutils/ChangeLog:
* testsuite/binutils-all/objdump.exp (objdump -S): Update testcases.
In commit 81dc3ab594 "[gdb/testsuite] Handle unreachable network in
server-connect.exp" a regression was introduced in gdb_target_cmd, causing
ERRORs like this:
...
ERROR: tcl error sourcing src/gdb/testsuite/gdb.server/abspath.exp.
ERROR: wrong # args: should be "gdb_target_cmd {$args}"
while executing
"gdb_target_cmd $gdbserver_protocol $gdbserver_gdbport"
...
Fix the argument passing in gdb_target_cmd.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-09-19 Tom de Vries <tdevries@suse.de>
* lib/gdbserver-support.exp (gdb_target_cmd): Fix argument passing.
When running gdb.server/server-connect.exp I run into:
...
FAIL: gdb.server/server-connect.exp: tcp6: connect to gdbserver using tcp6:::1
FAIL: gdb.server/server-connect.exp: tcp6-with-brackets: connect to gdbserver \
using tcp6:[::1]
FAIL: gdb.server/server-connect.exp: udp6: connect to gdbserver using udp6:::1
FAIL: gdb.server/server-connect.exp: udp6-with-brackets: connect to gdbserver \
using udp6:[::1]
...
The FAIL is caused by the fact that the ipv6 loopback address is not available:
...
PASS: gdb.server/server-connect.exp: tcp6: start gdbserver
target remote tcp6:::1:2347^M
A program is being debugged already. Kill it? (y or n) y^M
tcp6:::1:2347: Network is unreachable.^M
(gdb) FAIL: gdb.server/server-connect.exp: tcp6: connect to gdbserver using tcp6:::1
...
This should be marked UNSUPPORTED rather than FAIL.
Furthermore, the test-case takes about 4 minutes, because the 'Network is
unreachable' response is not explicitly handled in gdb_target_cmd, so instead
it runs into the timeout case.
Fix this by handling the 'Network is unreachable' response as UNSUPPORTED.
This reduces testing time from 4 minutes to about 2 seconds.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-09-19 Tom de Vries <tdevries@suse.de>
* lib/gdbserver-support.exp (gdb_target_cmd_ext): Return 2 (meaning
UNSUPPORTED) for 'Network is unreachable' message. Factor out of ...
(gdb_target_cmd): ... here.
* gdb.server/server-connect.exp: Use gdb_target_cmd_ext, handle return
value 2.
... to include PR breakpoints/24541 (for documentation purposes).
For the reader's convenience, the commit in question was the following:
commit 677052f2a5
Author: Sergio Durigan Junior <sergiodj@redhat.com>
Date: Thu May 16 16:23:24 2019 -0400
Subject: Make stap-probe.c:stap_parse_register_operand's "regname"
an std::string
Readline 8.0 has a feature that lets an application name a keymap.
This in turn makes it simpler for users to bind keys in keymaps in
their .inputrc.
This patch gives a name to the TUI SingleKey keymap, so that
additional bindings can be made there. For example:
$if gdb
set keymap SingleKey
"X": "echo hello\\n\n"
$endif
The call to rl_initialize, in tui_initialize_readline, had to be
removed so that .inputrc was not read too early. Note that Readline
explicitly documents that this call is not needed.
gdb/ChangeLog
2019-09-18 Tom Tromey <tom@tromey.com>
* NEWS: Add entry.
* tui/tui.c (tui_initialize_readline): Set name of keymap. Do not
call rl_initialize.
(tui_enable): Do not call rl_initialize.
gdb/doc/ChangeLog
2019-09-18 Tom Tromey <tom@tromey.com>
* gdb.texinfo (Editing): Document readline application name.
(TUI Single Key Mode): Document TUI SingleKey keymap name.
Fixes:
CXX alpha-linux-nat.o
alpha-linux-nat.c: In member function 'virtual CORE_ADDR alpha_linux_nat_target::register_u_offset(gdbarch*, int, int)':
alpha-linux-nat.c:93:16: error: 'gdbarch_pc_regnum' was not declared in this scope
if (regno == gdbarch_pc_regnum (gdbarch))
^~~~~~~~~~~~~~~~~
alpha-linux-nat.c:93:16: note: suggested alternative: 'gdbarch_info'
if (regno == gdbarch_pc_regnum (gdbarch))
^~~~~~~~~~~~~~~~~
gdbarch_info
alpha-linux-nat.c:97:15: error: 'gdbarch_fp0_regnum' was not declared in this scope
if (regno < gdbarch_fp0_regnum (gdbarch))
^~~~~~~~~~~~~~~~~~
alpha-linux-nat.c:97:15: note: suggested alternative: 'gdbarch_info'
if (regno < gdbarch_fp0_regnum (gdbarch))
^~~~~~~~~~~~~~~~~~
gdbarch_info
make[2]: *** [Makefile:1638: alpha-linux-nat.o] Error 1
make[2]: Leaving directory '/home/chris/tmp/binutils-gdb/gdb'
2019-09-12 Christian Groessler <chris@groessler.org>
* alpha-linux-nat.c: Include gdbarch.h.
Following the int -> bool conversion of boolean options (commit
491144b5e2 ("Change boolean options to bool instead of int")), I see
this ASAN error:
==357543==ERROR: AddressSanitizer: global-buffer-overflow on address 0x55555b25d440 at pc 0x5555583ce9e1 bp 0x7fffffffd390 sp 0x7fffffffd380
READ of size 4 at 0x55555b25d440 thread T0
#0 0x5555583ce9e0 in term_cli_styling /home/simark/src/binutils-gdb/gdb/ui-file.c:111
#1 0x5555583cf8b0 in stdio_file::can_emit_style_escape() /home/simark/src/binutils-gdb/gdb/ui-file.c:308
#2 0x5555584450d2 in set_output_style /home/simark/src/binutils-gdb/gdb/utils.c:1442
#3 0x5555584491af in fprintf_styled(ui_file*, ui_file_style const&, char const*, ...) /home/simark/src/binutils-gdb/gdb/utils.c:2143
#4 0x5555582fa13c in print_gdb_version(ui_file*, bool) /home/simark/src/binutils-gdb/gdb/top.c:1339
#5 0x555557b723ab in captured_main_1 /home/simark/src/binutils-gdb/gdb/main.c:981
#6 0x555557b7353b in captured_main /home/simark/src/binutils-gdb/gdb/main.c:1172
#7 0x555557b735d0 in gdb_main(captured_main_args*) /home/simark/src/binutils-gdb/gdb/main.c:1197
#8 0x55555700a53d in main /home/simark/src/binutils-gdb/gdb/gdb.c:32
#9 0x7ffff64c9ee2 in __libc_start_main (/usr/lib/libc.so.6+0x26ee2)
#10 0x55555700a30d in _start (/home/simark/build/binutils-gdb/gdb/gdb+0x1ab630d)
0x55555b25d441 is located 0 bytes to the right of global variable 'cli_styling' defined in '/home/simark/src/binutils-gdb/gdb/cli/cli-style.c:31:6' (0x55555b25d440) of size 1
The reason of this is that we have this declaration of cli_styling in cli/cli-style.h:
extern bool cli_styling;
but ui-file.c uses its own local declaration:
extern int cli_styling;
Because of that, the code in ui-file.c thinks the variable is 4 bytes
long, when it is actually 1 byte long, so the generated code reads past
the variable.
Fix it by removing the declaration and making ui-file.c include
cli/cli-style.h.
gdb/ChangeLog:
* ui-file.c: Include cli/cli-style.h.
(term_cli_styling): Remove cli_styling declaration.
Give a test an actual name in order to avoid having a directory name
in the .sum file.
gdb/testsuite/ChangeLog:
* gdb.base/source-dir.exp: Avoid having directory names in test
names.
There is a bunch of things left over after a `make clean` in the doc
directory. This patch changes the Makefile so that everything built by
make is cleaned:
- man pages (the man_MANS variable)
- cxxfilt.man
- binutils.info
I am not sure why the last one was in MAINTAINERCLEANFILES. It's
generated from binutils.texi using makeinfo, and it's not particularly
long to build, so I don't see why it shouldn't be cleaned like the other
files.
Note: with my automake 1,15.1 (built from source), I get a spurious diff
in Makefile.in, where the
runstatedir = @runstatedir@
line gets removed. I am not sure why, but I haven't included it.
binutils/ChangeLog:
* doc/Makefile.am (MOSTLYCLEANFILES): Clean all man pages,
cxxfilt.man and binutils.info.
(MAINTAINERCLEANFILES): Remove.
* doc/Makefile.in: Re-generate.
This variable doesn't exist (it evaluates to nothing).
It was introduced in
commit 42ecbf5eda
Author: Daniel Jacobowitz <drow@false.org>
Date: Sun May 15 18:19:45 2005 +0000
And removed in
commit c428fa83c2
Author: H.J. Lu <hjl.tools@gmail.com>
Date: Thu Mar 15 14:17:20 2007 +0000
Note: with my automake 1,15.1 (built from source), I get a spurious diff
in Makefile.in, where the
runstatedir = @runstatedir@
line gets removed. I am not sure why, but I haven't included it.
binutils/ChangeLog:
* doc/Makefile.am: Remove references to binutils_TEXINFOS.
* doc/Makefile.in: Re-generate.
The `install-data-local: install-info` part: the comment refers to when
the Makefile had the cygnus option. It no longer uses that, so the info
file should get installed without that rule. I verified that `make
install` still installed $(prefix)/share/info/binutils.info.
The `DISTCLEANFILES` part: the comment talks about Automake 1.9, we are
using 1.15.1, which doesn't appear to have this problem. binutils.info
is built fine without this snippet.
Note: with my automake 1,15.1 (built from source), I get a spurious diff
in Makefile.in, where the
runstatedir = @runstatedir@
line gets removed. I am not sure why, but I haven't included it.
binutils/ChangeLog:
* doc/Makefile.am (install-data-loca): Remove target.
(DISTCLEANFILES): Remove variable.
* doc/Makefile.in: Re-generate.
I get some spurious changes when running autoconf/automake for various
projects in the tree. This is likely because they were generated using
distro-patched tools last time.
I ran `autoreconf -f` in the various automake projects of the
binutils-gdb tree, and this is the result. The tools I am using have
been compiled from source, from the upstream release.
bfd/ChangeLog:
* Makefile.in: Re-generate.
* configure: Re-generate.
* doc/Makefile.in: Re-generate.
binutils/ChangeLog:
* Makefile.in: Re-generate.
* configure: Re-generate.
* doc/Makefile.in: Re-generate.
gas/ChangeLog:
* Makefile.in: Re-generate.
* configure: Re-generate.
* doc/Makefile.in: Re-generate.
gold/ChangeLog:
* testsuite/Makefile.in: Re-generate.
gprof/ChangeLog:
* Makefile.in: Re-generate.
* configure: Re-generate.
ld/ChangeLog:
* Makefile.in: Re-generate.
* configure: Re-generate.
opcodes/ChangeLog:
* Makefile.in: Re-generate.
* configure: Re-generate.
Fixes a few leaks in bfd and ld.
bfd/
* mach-o.c (bfd_mach_o_fat_member_init): Likewise. Replace
xstrdup and xmalloc with bfd_strdup and bfd_malloc. Return an
error status. Adjust calls.
* vms-lib.c (_bfd_vms_lib_get_module): Test mhd->id earlier.
Close bfd on failure. Replace xstrdup/bfd_alloc use with
bfd_malloc. Use bfd_set_filename.
gdb/
* solib-spu.c (spu_bfd_open): Use bfd_set_filename.
* spu-linux-nat.c (spu_bfd_open): Likewise.
ld/
* emultempl/pe.em (after_open): Use bfd_set_filename.
* emultempl/pep.em (after_open): Use bfd_set_filename.
Some versions of clang apparently generate non-PIC on powerpc64le to
access common symbols. Since a common symbol and a strong definition
with the same name should resolve to the strong definition we have the
possibility of non-PIC attempting to access shared library variables.
This is really a clanger since powerpc64le is supposed to be PIC by
default, but let's see if ld can cope by generating .dynbss copies.
* elf64-ppc.c (must_be_dyn_reloc): Return 0 for TOC16 relocs.
(ppc64_elf_check_relocs): Support dynamic/copy relocs for TOC16.
(ppc64_elf_adjust_dynamic_symbol): Don't keep dynamic reloc when
needs_copy even if all relocs are in rw sections.
(dec_dynrel_count): Handle TOC16 relocs.
(ppc64_elf_relocate_section): Support dynamic relocs for TOC16.
(ppc64_elf_finish_dynamic_symbol): Adjust to handle needs_copy
semantic change.
This was an oversight in 491144b5e21bbfd41969c175aebb663976f59058;
dwarf2loc.c has an extern definition for dwarf_always_disassemble
that I missed. This patch updates it from int to bool to match
the definition in dwarf2read.c.
gdb/ChangeLog:
2019-09-18 Christian Biesinger <cbiesinger@google.com>
* dwarf2loc.c: Change extern declaration of dwarf_always_disassemble
to bool to match definition in dwarf2read.c.
Generalize opcode arch dependencies so that we can support the
overlapping B extension Zb* subsets.
2019-09-17 Maxim Blinov <maxim.blinov@embecosm.com>
gas/
* config/tc-riscv.c (riscv_multi_subset_supports): Handle
insn_class enum rather than subset char string.
(riscv_ip): Update call to riscv_multi_subset_supports.
include/
* opcode/riscv.h (riscv_insn_class): New enum.
* opcode/riscv.h (struct riscv_opcode): Change
subset field to insn_class field.
opcodes/
* riscv-opc.c (riscv_opcodes): Change subset field
to insn_class field for all instructions.
(riscv_insn_types): Likewise.
The 'directory' command allows the user to provide a list of filesystem
directories in which to search for source code. The directories in this
search path are used as the base directory for the source filename from
the debug information (DW_AT_name). Thus the directory search path
provides alternatives to the existing compilation directory from the
debug information (DW_AT_comp_dir). Generally speaking, DW_AT_name
stores the filename argument passed to the compiler (including any
directory components), and DW_AT_comp_dir stores the current working
directory from which the compiler was executed. For example:
$ cd /path/to/project/subdir1
$ gcc -c a/test.c -g
The corresponding debug information will look like this:
DW_AT_name : a/test.c
DW_AT_comp_dir : /path/to/project/subdir1
When compiling with the -fdebug-prefix-map GCC option, the compilation
directory can be arbitrarily rewritten. In the above example, we may
rewrite the compilation directory as follows:
$ gcc -c a/test.c -g -fdebug-prefix-map=/path/to/project=
In this case, the corresponding debug information will look like:
DW_AT_name : a/test.c
DW_AT_comp_dir : /subdir1
This prevents GDB from finding the corresponding source code based on
the debug information alone. In some cases, a substitute-path command
can be used to re-map a consistent prefix in the rewritten compilation
directory to the real filesystem path. However, there may not be a
consistent prefix remaining in the debug symbols (for example in a
project that has source code in many subdirectories under the project's
root), thereby requiring multiple substitute-path rules. In this case,
it is easier to add the missing prefix to the directory search path via
the 'directory' command.
The function find_and_open_source currently searches in:
SEARCH_PATH/FILENAME
where SEARCH_PATH corresponds to each individual entry in the directory
search path (which is guaranteed to contain the compilation directory
from the debug information, as well as the current working directory).
FILENAME corresponds to the source filename (DW_AT_name), which may have
directory components in it. In addition, GDB searches in:
SEARCH_PATH/FILE_BASENAME
where FILE_BASENAME is the basename of the DW_AT_name entry.
This change modifies find_and_open_source to additionally search in:
SEARCH_PATH/COMP_DIR/FILENAME
where COMP_DIR is the compilation directory from the debug symbols. In
the example given earlier, running:
(gdb) directory /path/to/project
will now allow GDB to correctly locate the source code from the debug
information.
gdb/ChangeLog:
* source.c (prepare_path_for_appending): New function.
(openp): Make use of new function.
(find_and_open_source): Search for the compilation directory and
source file as a relative path beneath the directory search path.
gdb/doc/ChangeLog:
* gdb.texinfo (Source Path): Additional text to better describe
how the source path directory list is used when searching for
source files.
gdb/testsuite/ChangeLog:
* gdb.base/source-dir.exp: Add extra test for mapped compilation
directory.
The source_cache::get_line_charpos function can currently throw an
exception if the source file is missing, which doesn't match the
expected behaviour documented in the functions header file. The
documented behaviour is to return false on failure, and this is how
the function appears to be used throughout GDB.
I spotted this in the 'info source' command, currently for a missing
source file you'll see something like this:
(gdb) info source
Current source file is /path/to/src/file.c
Compilation directory is /path/to/build/
/path/to/src/file.c: No such file or directory.
(gdb)
After this patch we see this:
(gdb) info source
Current source file is /path/to/src/file.c
Compilation directory is /path/to/build/
Source language is c.
Producer is COMPILER VERSION AND FLAGS.
Compiled with DWARF 2 debugging format.
Does not include preprocessor macro info.
We don't currently indicate that the source file can't be found, and
maybe that would be something worth adding in the future.
gdb/ChangeLog:
* source-cache.c (source_cache::get_line_charpos): Catch
exceptions and return false, this matches the behaviour documented
in the header file.
gdb/testsuite/ChangeLog:
* gdb.base/list-missing-source.exp: New file.
A recent change enhanced a couple of notifications about Ada tasks to
include the task's name in addition to the task's number. That change
also modified the output of the "info tasks" when printing the detailed
information of one given task. I believe this change was unintentionally
left over from the initial version after it was decided that quoting
the task's name should not be done. This patch therefore undoes this
part of the patch.
gdb/ChangeLog:
* ada-tasks.c (info_task): Remove quoting of the task's name.
Tested on x86_64-linux, with both the official testsuite as well as
AdaCore's testsuite.
This just replaces the comment in the .c file with
"See symfile.h.".
gdb/ChangeLog:
2019-09-16 Christian Biesinger <cbiesinger@google.com>
* symfile.c (auto_solib_add): Replace comment with a reference
to the header file.
This was fixed in commit 272044897e.
2019-09-14 Christian Biesinger <cbiesinger@google.com>
* NEWS: Mention that gdb can now be compiled with Python 3
on Windows.
When running gdb.ada/rename_subscript_param.exp with gnatmake 7.4.1, we get:
...
FAIL: gdb.ada/rename_subscript_param.exp: print rename_subscript_param_b \
before changing its value
FAIL: gdb.ada/rename_subscript_param.exp: print rename_subscript_param_b \
after changing its value
...
The commit last touching the test-case (afcfda091e) states:
...
The test still fails with old compilers that do not properly
generate debug info for this renaming:
...
Fix this by requiring at least gnatmake 8 for the test-case.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-09-14 Tom de Vries <tdevries@suse.de>
PR teststuite/24599
* gdb.ada/rename_subscript_param.exp: Require gnatmake 8.
* lib/ada.exp (gnatmake_version_at_least): New proc.
I think this is safer than leaving an input_statement added during
open_input_bfds off the list. There are a number of places that
fiddle with various lists and might be confused by an off-list
statement, eg. orphan handling.
* ldlang.c (new_afile): Remove add_to_list parameter.
(lang_add_input_file): Update new_afile calls.
(lookup_name): Splice input_statement added by new_afile into
statement_list after current input_file_chain entry.
(lang_process): Update comment.
This commit:
commit aa17805fb9
Date: Sat Aug 31 23:44:40 2019 +0100
gdb: Have 'maint info sections' print all sections again
introduced a use of log10 that took an int as a parameter.
Unfortunately this was causing a compilation error on Solaris, see:
https://sourceware.org/ml/gdb-patches/2019-09/msg00230.htmlhttps://sourceware.org/ml/gdb-patches/2019-09/msg00231.html
because there was only a float, double, or long double version of
log10, and the compiler doesn't know which to choose.
This commit should resolve this issue by casting the argument to
float.
gdb/ChangeLog:
* maint.c (maint_print_section_data::maint_print_section_data):
Force use of 'float log10 (float)' by casting the argument to
float.
Make use of exec_has_index_section library function rather than
manually checking in the 'maintenance info sections' output. Should
make no difference to the test results, just makes the code easier to
read.
gdb/testsuite/ChangeLog:
* gdb.base/maint.exp: Use exec_has_index_section.
In this commit:
commit 6eac171f06
Date: Fri Aug 16 00:25:14 2019 +0200
[gdb] Make maint info sections print relocated addresses
A couple of things broke with the 'maintenance info sections' command,
here is some before output:
(gdb) maintenance info sections
Exec file:
`/path/to/gdb/build/gdb/testsuite/outputs/gdb.base/maint/maint',
file type elf64-x86-64.
[0] 0x00400238->0x00400254 at 0x00000238: .interp ALLOC LOAD READONLY DATA HAS_CONTENTS
[1] 0x00400254->0x00400274 at 0x00000254: .note.ABI-tag ALLOC LOAD READONLY DATA HAS_CONTENTS
[2] 0x00400274->0x00400298 at 0x00000274: .note.gnu.build-id ALLOC LOAD READONLY DATA HAS_CONTENTS
[3] 0x00400298->0x004002bc at 0x00000298: .gnu.hash ALLOC LOAD READONLY DATA HAS_CONTENTS
[4] 0x004002c0->0x00400380 at 0x000002c0: .dynsym ALLOC LOAD READONLY DATA HAS_CONTENTS
[5] 0x00400380->0x004003e3 at 0x00000380: .dynstr ALLOC LOAD READONLY DATA HAS_CONTENTS
[6] 0x004003e4->0x004003f4 at 0x000003e4: .gnu.version ALLOC LOAD READONLY DATA HAS_CONTENTS
[7] 0x004003f8->0x00400418 at 0x000003f8: .gnu.version_r ALLOC LOAD READONLY DATA HAS_CONTENTS
[8] 0x00400418->0x00400460 at 0x00000418: .rela.dyn ALLOC LOAD READONLY DATA HAS_CONTENTS
[9] 0x00400460->0x004004c0 at 0x00000460: .rela.plt ALLOC LOAD READONLY DATA HAS_CONTENTS
[10] 0x004004c0->0x004004d7 at 0x000004c0: .init ALLOC LOAD READONLY CODE HAS_CONTENTS
[11] 0x004004e0->0x00400530 at 0x000004e0: .plt ALLOC LOAD READONLY CODE HAS_CONTENTS
[12] 0x00400530->0x00400802 at 0x00000530: .text ALLOC LOAD READONLY CODE HAS_CONTENTS
[13] 0x00400804->0x0040080d at 0x00000804: .fini ALLOC LOAD READONLY CODE HAS_CONTENTS
[14] 0x00400810->0x0040084e at 0x00000810: .rodata ALLOC LOAD READONLY DATA HAS_CONTENTS
[15] 0x00400850->0x004008c4 at 0x00000850: .eh_frame_hdr ALLOC LOAD READONLY DATA HAS_CONTENTS
[16] 0x004008c8->0x00400ab8 at 0x000008c8: .eh_frame ALLOC LOAD READONLY DATA HAS_CONTENTS
[17] 0x00600e00->0x00600e08 at 0x00000e00: .init_array ALLOC LOAD DATA HAS_CONTENTS
[18] 0x00600e08->0x00600e10 at 0x00000e08: .fini_array ALLOC LOAD DATA HAS_CONTENTS
[19] 0x00600e10->0x00600ff0 at 0x00000e10: .dynamic ALLOC LOAD DATA HAS_CONTENTS
[20] 0x00600ff0->0x00601000 at 0x00000ff0: .got ALLOC LOAD DATA HAS_CONTENTS
[21] 0x00601000->0x00601038 at 0x00001000: .got.plt ALLOC LOAD DATA HAS_CONTENTS
[22] 0x00601038->0x0060103c at 0x00001038: .data ALLOC LOAD DATA HAS_CONTENTS
[23] 0x00601040->0x006012c8 at 0x0000103c: .bss ALLOC
[24] 0x00000000->0x0000002c at 0x0000103c: .comment READONLY HAS_CONTENTS
[25] 0x00000000->0x00000060 at 0x00001068: .debug_aranges READONLY HAS_CONTENTS
[26] 0x00000000->0x0000061b at 0x000010c8: .debug_info READONLY HAS_CONTENTS
[27] 0x00000000->0x00000264 at 0x000016e3: .debug_abbrev READONLY HAS_CONTENTS
[28] 0x00000000->0x000001e6 at 0x00001947: .debug_line READONLY HAS_CONTENTS
[29] 0x00000000->0x00000487 at 0x00001b2d: .debug_str READONLY HAS_CONTENTS
(gdb)
And here is the output after the above commit:
(gdb) maintenance info sections
+maintenance info sections
Exec file:
`/path/to/gdb/build/gdb/testsuite/outputs/gdb.base/maint/maint',
file type elf64-x86-64.
0x00400238->0x00400254 at 0x00000238: .interp ALLOC LOAD READONLY DATA HAS_CONTENTS
0x00400254->0x00400274 at 0x00000254: .note.ABI-tag ALLOC LOAD READONLY DATA HAS_CONTENTS
0x00400274->0x00400298 at 0x00000274: .note.gnu.build-id ALLOC LOAD READONLY DATA HAS_CONTENTS
0x00400298->0x004002bc at 0x00000298: .gnu.hash ALLOC LOAD READONLY DATA HAS_CONTENTS
0x004002c0->0x00400380 at 0x000002c0: .dynsym ALLOC LOAD READONLY DATA HAS_CONTENTS
0x00400380->0x004003e3 at 0x00000380: .dynstr ALLOC LOAD READONLY DATA HAS_CONTENTS
0x004003e4->0x004003f4 at 0x000003e4: .gnu.version ALLOC LOAD READONLY DATA HAS_CONTENTS
0x004003f8->0x00400418 at 0x000003f8: .gnu.version_r ALLOC LOAD READONLY DATA HAS_CONTENTS
0x00400418->0x00400460 at 0x00000418: .rela.dyn ALLOC LOAD READONLY DATA HAS_CONTENTS
0x00400460->0x004004c0 at 0x00000460: .rela.plt ALLOC LOAD READONLY DATA HAS_CONTENTS
0x004004c0->0x004004d7 at 0x000004c0: .init ALLOC LOAD READONLY CODE HAS_CONTENTS
0x004004e0->0x00400530 at 0x000004e0: .plt ALLOC LOAD READONLY CODE HAS_CONTENTS
0x00400530->0x00400802 at 0x00000530: .text ALLOC LOAD READONLY CODE HAS_CONTENTS
0x00400804->0x0040080d at 0x00000804: .fini ALLOC LOAD READONLY CODE HAS_CONTENTS
0x00400810->0x0040084e at 0x00000810: .rodata ALLOC LOAD READONLY DATA HAS_CONTENTS
0x00400850->0x004008c4 at 0x00000850: .eh_frame_hdr ALLOC LOAD READONLY DATA HAS_CONTENTS
0x004008c8->0x00400ab8 at 0x000008c8: .eh_frame ALLOC LOAD READONLY DATA HAS_CONTENTS
0x00600e00->0x00600e08 at 0x00000e00: .init_array ALLOC LOAD DATA HAS_CONTENTS
0x00600e08->0x00600e10 at 0x00000e08: .fini_array ALLOC LOAD DATA HAS_CONTENTS
0x00600e10->0x00600ff0 at 0x00000e10: .dynamic ALLOC LOAD DATA HAS_CONTENTS
0x00600ff0->0x00601000 at 0x00000ff0: .got ALLOC LOAD DATA HAS_CONTENTS
0x00601000->0x00601038 at 0x00001000: .got.plt ALLOC LOAD DATA HAS_CONTENTS
0x00601038->0x0060103c at 0x00001038: .data ALLOC LOAD DATA HAS_CONTENTS
0x00601040->0x006012c8 at 0x0000103c: .bss ALLOC
0x00000000->0x00000000 at 0x00000000: *COM* IS_COMMON
0x00000000->0x00000000 at 0x00000000: *UND*
0x00000000->0x00000000 at 0x00000000: *ABS*
0x00000000->0x00000000 at 0x00000000: *IND*
(gdb)
We lost the section index numbers, but more importantly, we lost the
information about the .debug* sections. We also gained entries for
the "fake" sections *COM*, *UND*, *ABS*, and *IND*.
I noticed this when running:
make check-gdb RUNTESTFLAGS="--target_board=cc-with-gdb-index gdb.base/maint.exp"
As this test relies on looking in the 'maint info sections' output to
see if we have a .debug_names or .gdb_index section, and these are
debug sections so they no longer show up in the 'main info sections'
output, the gdb.base/maint.exp test fails.
This commit restores the old behaviour while keeping the important
change that the above commit introduced, the addresses printed for
sections are the relocated addresses where appropriate. The above
commit mentions using this test:
make check-gdb RUNTESTFLAGS="CFLAGS_FOR_TARGET='-pie' gdb.base/compare-sections.exp"
And this still passes after this commit.
The output for 'maint info sections' now looks like this:
(gdb) maintenance info sections
Exec file:
`/home/andrew/projects/binutils-gdb/build/gdb/testsuite/outputs/gdb.base/maint/maint',
file type elf64-x86-64.
[0] 0x00400238->0x00400254 at 0x00000238: .interp ALLOC LOAD READONLY DATA HAS_CONTENTS
[1] 0x00400254->0x00400274 at 0x00000254: .note.ABI-tag ALLOC LOAD READONLY DATA HAS_CONTENTS
[2] 0x00400274->0x00400298 at 0x00000274: .note.gnu.build-id ALLOC LOAD READONLY DATA HAS_CONTENTS
[3] 0x00400298->0x004002bc at 0x00000298: .gnu.hash ALLOC LOAD READONLY DATA HAS_CONTENTS
[4] 0x004002c0->0x00400380 at 0x000002c0: .dynsym ALLOC LOAD READONLY DATA HAS_CONTENTS
[5] 0x00400380->0x004003e3 at 0x00000380: .dynstr ALLOC LOAD READONLY DATA HAS_CONTENTS
[6] 0x004003e4->0x004003f4 at 0x000003e4: .gnu.version ALLOC LOAD READONLY DATA HAS_CONTENTS
[7] 0x004003f8->0x00400418 at 0x000003f8: .gnu.version_r ALLOC LOAD READONLY DATA HAS_CONTENTS
[8] 0x00400418->0x00400460 at 0x00000418: .rela.dyn ALLOC LOAD READONLY DATA HAS_CONTENTS
[9] 0x00400460->0x004004c0 at 0x00000460: .rela.plt ALLOC LOAD READONLY DATA HAS_CONTENTS
[10] 0x004004c0->0x004004d7 at 0x000004c0: .init ALLOC LOAD READONLY CODE HAS_CONTENTS
[11] 0x004004e0->0x00400530 at 0x000004e0: .plt ALLOC LOAD READONLY CODE HAS_CONTENTS
[12] 0x00400530->0x00400802 at 0x00000530: .text ALLOC LOAD READONLY CODE HAS_CONTENTS
[13] 0x00400804->0x0040080d at 0x00000804: .fini ALLOC LOAD READONLY CODE HAS_CONTENTS
[14] 0x00400810->0x0040084e at 0x00000810: .rodata ALLOC LOAD READONLY DATA HAS_CONTENTS
[15] 0x00400850->0x004008c4 at 0x00000850: .eh_frame_hdr ALLOC LOAD READONLY DATA HAS_CONTENTS
[16] 0x004008c8->0x00400ab8 at 0x000008c8: .eh_frame ALLOC LOAD READONLY DATA HAS_CONTENTS
[17] 0x00600e00->0x00600e08 at 0x00000e00: .init_array ALLOC LOAD DATA HAS_CONTENTS
[18] 0x00600e08->0x00600e10 at 0x00000e08: .fini_array ALLOC LOAD DATA HAS_CONTENTS
[19] 0x00600e10->0x00600ff0 at 0x00000e10: .dynamic ALLOC LOAD DATA HAS_CONTENTS
[20] 0x00600ff0->0x00601000 at 0x00000ff0: .got ALLOC LOAD DATA HAS_CONTENTS
[21] 0x00601000->0x00601038 at 0x00001000: .got.plt ALLOC LOAD DATA HAS_CONTENTS
[22] 0x00601038->0x0060103c at 0x00001038: .data ALLOC LOAD DATA HAS_CONTENTS
[23] 0x00601040->0x006012c8 at 0x0000103c: .bss ALLOC
[24] 0x00000000->0x0000002c at 0x0000103c: .comment READONLY HAS_CONTENTS
[25] 0x00000000->0x00000060 at 0x00001068: .debug_aranges READONLY HAS_CONTENTS
[26] 0x00000000->0x0000061b at 0x000010c8: .debug_info READONLY HAS_CONTENTS
[27] 0x00000000->0x00000264 at 0x000016e3: .debug_abbrev READONLY HAS_CONTENTS
[28] 0x00000000->0x000001e6 at 0x00001947: .debug_line READONLY HAS_CONTENTS
[29] 0x00000000->0x00000487 at 0x00001b2d: .debug_str READONLY HAS_CONTENTS
(gdb)
This is basically as it was before, except that the index numbers are
now padded so the section information all lines up.
When GDB has relocated a section then the relocated addresses will be
printed, otherwise the non-relocated addresses from the bfd will be
printed.
I've added a test to gdb.base/maint.exp to do some basic validation of
the output format.
gdb/ChangeLog:
* maint.c: Add 'cmath' include.
(struct maint_print_section_data): New structure.
(print_section_index): New function.
(print_bfd_section_info): Add header comment, small whitespace
cleanup, and update to call new print_section_index function.
(print_objfile_section_info): Likewise.
(maint_obj_section_from_bfd_section): New function.
(print_bfd_section_info_maybe_relocated): New function.
(maintenance_info_sections): Add header comment, always use
bfd_map_over_sections instead of ALL_OBJFILE_OSECTIONS.
gdb/testsuite/ChangeLog:
* gdb.base/maint.exp: Add test for 'maint info sections'.
I found that these tests:
make check-gdb RUNTESTFLAGS="--target_board=cc-with-gdb-index gdb.base/maint.exp"
make check-gdb RUNTESTFLAGS="--target_board=cc-with-debug-names gdb.base/maint.exp"
were causing GDB to segfault. It turns out that this test runs this
command:
maint print psymbols -pc main /path/to/some/file
which tries to lookup the partial_symtab for 'main'. The problem is
that there is no partial_symtab for 'main' as we are using the
.gdb_index or .debug_names instead of partial_symtabs.
What happens is that maintenance_print_symbols calls
find_pc_sect_psymtab, which looks for the partial_symtab in the
objfile's objfile->partial_symtabs->psymtabs_addrmap.
This is a problem because when we are using the indexes
psymtabs_addrmap is reused to hold things other than partial_symtabs,
this can be seen in dwarf2read.c in create_addrmap_from_index and
create_addrmap_from_aranges. If we then lookup in psymtabs_addrmap we
end up returning a pointer to something that isn't really a
partial_symtab, after which everything goes wrong.
Initially I simply added a check at the start of find_pc_sect_psymtab
that the objfile had some partial_symtabs, like:
if (objfile->partial_symtabs->psymtabs == NULL)
return NULL;
Figuring that if there were no partial_symtabs at all then this
function should always return NULL, however, this caused a failure in
the test gdb.python/py-event.exp which I didn't dig into too deeply,
but seems to be that in this tests there are initially no psymtabs,
but the second part of find_pc_sect_psymtab does manage to read some
in from somewhere, with the check I added the test fails as we
returned NULL here and this caused GDB to load in the full symtabs
earlier than was expected.
Instead I chose to guard only the access to psymtabs_addrmap with a
check that the function has some psymtabs. This allows my original
tests to pass, and the py-event.exp test to pass too.
Now, a good argument can be made that we simply should never call
find_pc_sect_psymtab on an objfile that is using indexes instead of
partial_symtabs. I did consider this approach, we could easily add an
assert into find_pc_sect_psymtab that if we find a partial_symtab in
psymtabs_addrmap then the psymtabs pointer must be non-null. The
responsibility would then be on the user of find_pc_sect_psymtab to
ensure that the objfile being checked is suitable. In the end I
didn't take this approach as the check in find_pc_sect_psymtab is
cheap and this ensures that any future miss-uses of the function will
not cause problems.
I also extended the comment on psymtabs_addrmap to indicate that it
holds more than just partial_symtabs as this was not at all clear from
the original comment, and caused me some confusion when I was
initially debugging this problem.
gdb/ChangeLog:
* psymtab.c (find_pc_sect_psymtab): Move baseaddr local into more
inner scope, add check that the objfile has psymtabs before
checking psymtabs_addrmap.
* psymtab.h (psymtab_storage) <psymtabs_addrmap>: Extend comment.
The test-case gdb.base/store.exp fails with gcc 7.4.0:
...
nr of unexpected failures 27
...
The first FAIL:
...
110 l = add_float (l, r);
(gdb) PASS: gdb.base/store.exp: continue to wack_float
print l
$21 = <optimized out>
FAIL: gdb.base/store.exp: var float l; print old l, expecting -1
...
relates to this bit in the test-case (compiled at -O0):
...
106 float
107 wack_float (register float u, register float v)
108 {
109 register float l = u, r = v;
110 l = add_float (l, r);
111 return l + r;
112 }
...
and it expects to be able to read and modify variable l before executing line
110, but it already fails to read the value, because l has no DW_AT_location
attribute in the debug info.
Variable l is declared with the register keyword, and GCC implements the
register keyword at -O0 like so:
...
the compiler allocates distinct stack memory for all variables that do not
have the register storage-class specifier; if register is specified, the
variable may have a shorter lifespan than the code would indicate and may
never be placed in memory.
...
The fact that l has no DW_AT_location attribute, matches with the documented
"variable may have a shorter lifespan that code would indicate", (though it
is the most extreme case of it) so the gcc behaviour is valid. We can of
course improve gcc to generate better debuginfo (filed gcc PR91611), but
this not a wrong-debug problem.
[ The test-case passes with gcc 4.2.1, but for the failing test discussed
above, it passes simply because it doesn't store l in a register. ]
With the debug info missing for l, reading and setting l is unsupported, so
fix the FAIL by marking the test UNSUPPORTED instead.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-09-12 Tom de Vries <tdevries@suse.de>
* gdb.base/store.exp: Allow register variables to be optimized out at
-O0.
With this patch, we e.g. get:
[Switching to task 2 "task_list(1)"]
[Current task is 2 "task_list(1)"]
instead of
[Switching to task 2]
[Current task is 2]
The logic to produce the taskno optionally followed
by the task name has been factorized in the task_to_str function.
Task names are output between double quotes in the new messages,
similarly to what GDB does for thread names.
However, no quotes are put around task names in 'info tasks' Name column.
This was discussed with Tom, that preferred no quotes there, while I
was more in favour of visual consistency.
I discussed with a few more users, which led to (exactly) 50% preferring
quotes and 50% preferring no quotes :).
To arrive to the decision to remove the quotes, the following "killing args"
were used:
* To have quotes or to not have quotes, that is the question; yes
but not *THE* question :).
* If there is not a clear majority that prefers quotes, better to
not disturb the existing user basis for a (somewhat) irrelevant
aspect.
* The opinion of the reviewer has more weight.
So, compared to the previous version, this version remotes the quotes
in 'info tasks'.
It improves the alignement of 'info tasks' output.
With this patch, we get:
(gdb) info task
ID TID P-ID Pri State Name
* 1 555555759030 48 Runnable main_task
2 555555759e30 1 48 Selective Wait mit
(gdb)
instead of
(gdb) info task
ID TID P-ID Pri State Name
* 1 555555759030 48 Runnable main_task
2 555555759e30 1 48 Selective Wait mit
(gdb)
(e.g. the first one properly shows parent and priority under the
correct header).
This is version 4 of the 'task name' patch.
Compared to version 3, the changes are:
output task names between quotes but not in 'info tasks'
gdb/ChangeLog
2019-09-12 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* NEWS: Announce that Ada task names are now shown at more places,
and between quotes (except in info task output).
* gdb/ada-tasks.c (task_to_str): New function.
(display_current_task_id): Call task_to_str.
(task_command_1): Likewise.
(print_ada_task_info): In non-mi mode, Properly align headers and data
when task-id length is > 9 (9 is the default for a 32 bits CORE_ADDR).
gdb/doc/ChangeLog
2019-09-12 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* gdb.texinfo (Ada Tasks): Tell the task name is printed, update
examples.
gdb/testsuite/ChangeLog
2019-09-12 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* gdb.ada/rdv_wait.exp: Update to new task names.
* gdb.base/task_switch_in_core.exp: Likewise.
* gdb.base/info_sources_base.c: Likewise.
For some reason, commit 906e58cab5 2008-05-15 excluded input
statements with BFD_LINKER_CREATED BFDs from being printed to map
files. This isn't ideal since it loses claimed plugin BFDs.
* ldlang.c (print_input_statement): Do not exclude linker created
BFDs.
This fixes a problem with commit 128bf1fe60, a patch I made
2019-08-06. Apparently it is possible to trigger the assertion I
added during an LTO bootstrap, something I haven't reproduced.
However, I did find a case triggered by an odd linker script feature
that allows a file to be loaded from the script without specifying
that file on the command line. Regarding input sections:
"When you use a file name which is not an archive:file specifier
and does not contain any wild card characters, the linker will
first see if you also specified the file name on the linker command
line or in an INPUT command. If you did not, the linker will
attempt to open the file as an input file, as though it appeared on
the command line."
So putting
.foo : { foo.a(*) }
into a script supposedly extracts foo.a into .foo. Except it doesn't,
since this feature is meant for object files only. Well anyway,
assuming --whole-archive was given on the command line, foo.a contains
a -flto object and no other objects involved were -flto then we'll hit
the assert due to files added like foo.a here *not* having their input
statement put on the general statement list. Why these are not put on
the statement list isn't obvious but it has been that way since commit
193c5f93a1 in 1994.
PR 24981
* ldlang.c (lang_process): Remove assertion. Comment.
It's been a long-standing nuisance that gdb reported unaligned accesses
on Solaris/SPARC as SIGSEGV, contrary to the shells and truss which
correctly report SIGBUS instead.
I could trace this down to the fault handling code in procfs.c
(procfs_target::wait): when pr_why is set to PR_FAULTED, the current
code sets the signal based on the fault number. For one, the code gets
this wrong for FLTACCESS (the unaligned access case) where it uses
SIGSEGV. What's worse, it's completely unnecessary to make up the
signal number inside gdb. Instead, it should just take what procfs
reports to avoid mismatches, which is what this patch does. I've
completely removed the explicit handling of the various fault codes: for
one, the list has already been incomplete, lacking FLTCPCOVF which
existed since at least Solaris 8. Besides, there's no reason to error
out on unknown fault codes: either the fault causes a signal which can
then be reported from procfs, or it doesn't (as for FLTPAGE) and no
reporting is necessary.
Tested on sparcv9-sun-solaris2.11 and x86_64-pc-solaris2.11. Also
spot-checked manually for a couple of cases (unaligned access, division
by 0, NULL pointer dereference).
* procfs.c (procfs_target::wait) <PR_FAULTED>: Get signal from
prstatus.pr_lwp.pr_info instead of making it up.
Matthew Malcomson
Tamar Christina
Alan Modra
GDB Administrator
Tom de Vries
Joel Brobecker
Tom Tromey
Christian Groessler
Simon Marchi
Andrew Burgess
Christian Biesinger
Jim Wilson
Mike Gulick
Phil Blundell
Philippe Waroquiers
Rainer Orth