While looking at the DAP spec, I noticed that the breakpointLocations
request is gated behind a capability. This patch changes gdb to
report this capability.
I've also added a comment to explain the fact that arguments to
breakpointLocations are not optional, even though the spec says they
are.
The canonical form to discard all sections not mentioned earlier in
the script is
/DISCARD/ : { *(*) }
not
/DISCARD/ : { *(.*) }
".*" happens to work with the usual section names starting with a dot,
but let's not promote something not quite right.
Two fixes in gdb.python/py-thread-exited.exp:
- fix the copyright notice validity range (PR testsuite/30687):
2022-202 -> 2022-2023, and
- add missing "require allow_python_tests".
Tested on x86_64-linux.
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30687
This little patch makes the BPF assembler accept '#' as an inline
comment character, which clang -S seems to use.
gas/
* config/tc-bpf.c (comment_chars): Add '#'.
* doc/c-bpf.texi (BPF Special Characters): Add note that '#' may
be used for inline comments.
When building gdb with -O2 -fsanitize-threads, I ran into
a Werror=stringop-truncation.
The problem is here in coff_getfilename in coffread.c:
...
strncpy (buffer, aux_entry->x_file.x_n.x_fname, FILNMLEN);
buffer[FILNMLEN] = '\0';
...
The constant FILNMLEN is expected to designate the size of
aux_entry->x_file.x_n.x_fname, but that's no longer the case since commit
60ebc25751 ("Fixes a buffer overflow when compiling assembler for the MinGW
targets.").
Fix this by using "sizeof (aux_entry->x_file.x_n.x_fname)" instead.
Likewise in xcoffread.c.
Tested on x86_64-linux.
Approved-By: Tom Tromey <tom@tromey.com>
PR build/30669
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30669
As reported in PR testsuite/30633, when running test-case
gdb.dwarf2/typeddwarf.exp with target board native-gdbserver on Ubuntu
22.04.2, we run into:
...
(gdb) continue^M
Continuing.^M
^M
Program received signal SIGSEGV, Segmentation fault.^M
0x0000000000000001 in ?? ()^M
(gdb) FAIL: gdb.dwarf2/typeddwarf.exp: runto: run to main
...
We run into the FAIL as follows:
- due to using gdbserver, we attach at the point of the first instruction, in
_start
- we then set a breakpoint at main
- the test-case is a .s file, that has main renamed to _start in the assembly,
but not in the debuginfo
- setting a breakpoint at main sets the breakpoint at the same instruction
we're currently stopped at
- continue doesn't hit the breakpoint, and we return out of _start, which
causes a sigsegv
Note that this is for the amd64 case (using gdb.dwarf2/typeddwarf-amd64.S).
For the i386 case (using gdb.dwarf2/typeddwarf.S), setting a breakpoint in
main sets it one insn after function entry, and consequently the problem does
not occur.
The FAIL is a regression since commit 90cce6c055 ("[gdb/testsuite] Add nopie
in a few test-cases").
Without nopie the executable is PIE, with nopie it's static instead.
In the PIE case, we attach at the point of _start in the dynamic linker, and
consequently we do not skip the breakpoint in main, and also don't run into
the FAIL.
Fix this by:
- removing the -nostdlib setting, and
- renaming _start to main in both .S files.
The change to use -nostdlib and rename main to _start was originally added
in commit 6edba76fe8 (submitted here:
https://sourceware.org/pipermail/gdb-patches/2011-May/082657.html ) , I assume
to fix the problem now fixed by using nopie.
Tested on x86_64-linux.
Reported-By: Simon Marchi <simon.marchi@efficios.com>
Tested-By: Simon Marchi <simon.marchi@efficios.com>
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30633
With CLI, a session defining a command looks like:
...
(gdb) define foo
Type commands for definition of "foo".
End with a line saying just "end".
>bar
>end
(gdb)
...
With TUI however, we get the same secondary prompts, and type the same, but
are left with:
...
(gdb) define foo
Type commands for definition of "foo".
End with a line saying just "end".
(gdb)
...
Fix this by calling tui_inject_newline_into_command_window in
gdb_readline_wrapper_line, as is done in tui_command_line_handler.
Tested on x86_64-linux.
PR tui/30636
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30636
With a gdb build with -O2 -flto=auto and gcc 7.5.0 and test-case
gdb.gdb/python-helper.exp I run into:
...
(outer-gdb) continue^M
Continuing.^M
print 1^M
^M
Thread 1 "xgdb" hit Breakpoint 2, \
_Z11value_printP5valueP7ui_filePK19value_print_options (val=0x22e2590, \
stream=0x1f65480, options=0x7fffffffcdc0) at gdb/valprint.c:1193^M
1193 {^M
(outer-gdb) FAIL: gdb.gdb/python-helper.exp: hit breakpoint in outer gdb
...
This is the "value_print" variant of the problem with "c_print_type" I fixed
in commit 0d332f1112 ("[gdb/testsuite] Fix gdb.gdb/python-helper.exp with -O2
-flto=auto and gcc 7.5.0").
Fix this likewise.
Tested on x86_64-linux.
In gdb/tui/tui-layout.c, we have:
...
static window_types_map known_window_types;
...
and in gdb/python/py-tui.c:
...
/* A global list of all gdbpy_tui_window_maker objects. */
static intrusive_list<gdbpy_tui_window_maker> m_window_maker_list;
};
/* See comment in class declaration above. */
intrusive_list<gdbpy_tui_window_maker>
gdbpy_tui_window_maker::m_window_maker_list;
...
With a gdb build with -O0 or -O2, the static destructor calling order seems to be:
- first gdb/tui/tui-layout.c,
- then gdb/python/py-tui.c.
So when running test-case gdb.python/tui-window-factory.exp, we see the
following order of events:
- the destructor for known_window_types is called, which triggers calling the
destructor for the only element E of m_window_maker_list. The destructor
destroys E, and also removes E from m_window_maker_list, leaving it empty.
- the destructor for m_window_maker_list is called. It's empty, so it's a nop.
However, when building gdb with -O2 -flto=auto, the static destructor calling
order seems to be reversed.
Instead, we have these events:
- the destructor for m_window_maker_list is called. This doesn't destroy it's
only element E, but it does make m_window_maker_list empty.
- the destructor for known_window_types is called, which triggers calling the
destructor for E. An attempt is done to remove E from m_window_maker_list,
but we run into an assertion failure, because the list is empty.
Fix this by checking is_linked () before attempting to remove from
m_window_maker_list, similar to how things were addressed in commit 995a34b177
("Guard against frame.c destructors running before frame-info.c's").
Tested on x86_64-linux.
PR tui/30646
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30646
When running test-case gdb.base/step-over-syscall.exp without glibc debuginfo
installed, I get:
...
(gdb) continue^M
Continuing.^M
^M
Breakpoint 2, 0x00007ffff7d4405e in vfork () from /lib64/libc.so.6^M
(gdb) PASS: gdb.base/step-over-syscall.exp: vfork: displaced=off: \
continue to vfork (1st time)
...
but with glibc debuginfo installed I get instead:
...
(gdb) continue^M
Continuing.^M
^M
Breakpoint 2, 0x00007ffff7d4405e in __libc_vfork () at \
../sysdeps/unix/sysv/linux/x86_64/vfork.S:44^M
44 ENTRY (__vfork)^M
(gdb) FAIL: gdb.base/step-over-syscall.exp: vfork: displaced=off: \
continue to vfork (1st time)
...
The FAIL is due to a mismatch with regexp:
...
"Breakpoint \[0-9\]+, (.* in |__libc_|)$syscall \\(\\).*"
...
because it cannot match both ".* in " and the __libc_ prefix.
Fix this by using instead the regexp:
...
"Breakpoint \[0-9\]+, (.* in )?(__libc_)?$syscall \\(\\).*"
...
Tested on x86_64-linux.
This patch fixes the semantics of the neg and neg32 BPF instructions
in the simulator, and also updates the corresponding tests
accordingly.
Tested in target bpf-unknown-none.
This patch fixes the BPF_INSN_NEGR and BPF_INSN_NEG32R BPF
instructions to not use their source registers.
Tested in bpf-unknown-none.
opcodes/ChangeLog:
2023-07-26 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf-opc.c (bpf_opcodes): Fix BPF_INSN_NEGR to not use a src
register.
gas/ChangeLog:
2023-07-26 Jose E. Marchesi <jose.marchesi@oracle.com>
* testsuite/gas/bpf/alu.s: The register neg instruction gets only
one argument.
* testsuite/gas/bpf/alu32-be-pseudoc.d: Likewise.
* testsuite/gas/bpf/alu32-pseudoc.d: Likewise.
* testsuite/gas/bpf/alu32-pseudoc.s: Likewise.
* testsuite/gas/bpf/alu-pseudoc.d: Likewise.
* testsuite/gas/bpf/alu-be-pseudoc.d: Likewise.
* testsuite/gas/bpf/alu-pseudoc.s: Likewise.
* testsuite/gas/bpf/alu-be.d: Likewise.
* testsuite/gas/bpf/alu.d: Likewise.
* testsuite/gas/bpf/alu32-be.d: Likewise.
* testsuite/gas/bpf/alu32.d: Likewise.
* testsuite/gas/bpf/alu32.s: Likewise.
* doc/c-bpf.texi (BPF Instructions): Update accordingly.
get_symbol_name currently returns "" for the usual STT_SECTION symbols
generated by gas. That's not very helpful, return the section name.
Demangle local symbols too, fixing an inconsistency in
issue_discarded_error where global symbols are demangled.
* object.cc (Sized_relobj_file::get_symbol_name): Return a
std::string. Return section name for STT_SECTION symbols with
zero st_name. Sanity check st_name, and don't run off the end
of an improperly terminated .strtab. Demangle sym names.
* object.h (Sized_relobj_file::get_symbol_name): Update decl.
* target-reloc.h (issue_discarded_error): Adjust.
* powerpc.cc (Target_powerpc::Relocate::relocate): Report reloc
type and symbol for relocation overflows.
Since the 'Zca' extension is the new base of the compressed instructions,
this commit enables RVC *also* when the 'Zca' extension is enabled
via ".option arch" directive.
gas/ChangeLog:
* config/tc-riscv.c (s_riscv_option): Enable RVC also when the
'Zca' extension is enabled after an ".option arch" directive.
The version 1.0.4-1 of the code size reduction specification clarifies
that 'Zcf' implies 'F' and 'Zcd' implies 'D'.
cf:
<https://github.com/riscv/riscv-code-size-reduction/releases/tag/v1.0.4-1>
This commit adds those implications.
bfd/ChangeLog:
* elfxx-riscv.c (riscv_implicit_subsets): Add two implications,
'Zcf' -> 'F' and 'Zcd' -> 'D'.
gas/ChangeLog:
* testsuite/gas/riscv/march-imply-zcd.d: New test.
* testsuite/gas/riscv/march-imply-zcf.d: New test.
As per:
<https://github.com/riscv/riscv-code-size-reduction/issues/221>,
the 'Zcf' extension does not exist on RV64. This is reflected on the
version 1.0.4-1 of the code size reduction specification:
<https://github.com/riscv/riscv-code-size-reduction/releases/tag/v1.0.4-1>.
This commit prohibits the combination: RV64 (or any ISA with XLEN > 32)
and the 'Zcf' extension.
bfd/ChangeLog:
* elfxx-riscv.c (riscv_parse_check_conflicts): Prohibit
combination of RV64 and 'Zcf'.
gas/ChangeLog:
* testsuite/gas/riscv/march-fail-rv64i_zcf.d: New test.
* testsuite/gas/riscv/march-fail-rv64i_zcf.l: Likewise.
bfd
* peXXigen.c (_bfd_XXi_only_swap_filehdr_out): If inserting a timestamp, use the value held in the SOURCE_DATE_EPOCH environment variable, if it is defined.
binutils
* doc/binutils.texi (objcopy): Document change in behaviour of objcopy's --preserve-dates command line option.
ld
* pe-dll.c (fill_edata): If inserting a timestamp, use the value held in the SOURCE_DATE_EPOCH environment variable, if it is defined.
* ld.texi (--insert-timestamp): Document change in behaviour.
In LoongArch, the R_LARCH_IRELATIVE relocations for local ifunc symbols are
in .rela.dyn. Before, this is done by loongarch_elf_finish_dynamic_sections.
But this function is called after elf_link_sort_relocs, it need to find a
null slot to insert IRELATIVE relocation.
Now, it is processed by elf_loongarch_output_arch_local_syms before
elf_link_sort_relocs, just need to call loongarch_elf_append_rela to
insert IRELATIVE relocation.
bfd/ChangeLog:
* elfnn-loongarch.c (elfNN_allocate_local_ifunc_dynrelocs): Return
type change to int.
(loongarch_elf_size_dynamic_sections): Delete (void *).
(loongarch_elf_finish_dynamic_symbol): Use loongarch_elf_append_rela
insert IRELATIVE relocation to .rela.dyn.
(elfNN_loongarch_finish_local_dynamic_symbol): Return type change to
int.
(loongarch_elf_finish_dynamic_sections): Delete process of local
ifunc symbols.
(elf_backend_output_arch_local_syms): New.
ld/ChangeLog:
* testsuite/ld-loongarch-elf/local-ifunc-reloc.d: Regenerated.
For B16/B21/B26/PCREL20_S2 relocations, if immediate overflow check after
rightshift, and the mask need to include sign bit.
Now, the immediate overflow check before rightshift for easier understand.
bfd/ChangeLog:
* elfxx-loongarch.c (reloc_bits_pcrel20_s2): Delete.
(reloc_bits_b16): Delete.
(reloc_bits_b21): Delete.
(reloc_bits_b26): Delete.
(reloc_sign_bits): New.
For extreme code mode, the instruction sequences is
pcalau12i $t0, hi20
addi.d $t1, $zero, lo12
lu32i.d $t1, lo20
lu52i.d $t1, hi12
add.d $t1, $t0, $t1
If lo12 > 0x7ff, hi20 need to add 0x1, lo20 need to sub 0x1.
If hi20 > 0x7ffff, lo20 need to add 0x1.
bfd/ChangeLog:
* elfnn-loongarch.c (RELOCATE_CALC_PC32_HI20): Redefined.
(RELOCATE_CALC_PC64_HI32): Redefined.
This patch adds support for the BPF V4 ISA byte swap instructions to
opcodes, assembler and disassembler.
Tested in bpf-unknown-none.
include/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* opcode/bpf.h (BPF_IMM32_BSWAP16): Define.
(BPF_IMM32_BSWAP32): Likewise.
(BPF_IMM32_BSWAP64): Likewise.
(enum bpf_insn_id): New entries BPF_INSN_BSWAP{16,32,64}.
opcodes/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf-opc.c (bpf_opcodes): Add entries for the BSWAP*
instructions.
gas/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* doc/c-bpf.texi (BPF Instructions): Document BSWAP* instructions.
* testsuite/gas/bpf/alu.s: Test BSWAP{16,32,64} instructions.
* testsuite/gas/bpf/alu.d: Likewise.
* testsuite/gas/bpf/alu-be.d: Likewise.
* testsuite/gas/bpf/alu-pseudoc.s: Likewise.
* testsuite/gas/bpf/alu-pseudoc.d: Likewise.
* testsuite/gas/bpf/alu-be-pseudoc.d: Likewise.
gas/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* doc/c-bpf.texi (BPF Instructions): The pseudoc syntax for MOVS*
doesn't use `s=' but `='.
This patch fixes the pseudoc syntax of the V4 instructions MOVS* and
LDXS* in order to reflect https://reviews.llvm.org/D144829.
opcodes/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf-opc.c (bpf_opcodes): Fix pseudo-c syntax for MOVS* and LDXS*
instructions.
gas/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* doc/c-bpf.texi (BPF Instructions): Fix pseudoc syntax for MOVS*
and LDXS* instructions.
* testsuite/gas/bpf/mem-pseudoc.d: Likewise.
* testsuite/gas/bpf/mem-be-pseudoc.d: Likewise.
* testsuite/gas/bpf/mem-pseudoc.s: Likewise.
* testsuite/gas/bpf/alu-pseudoc.s: Likewise.
* testsuite/gas/bpf/alu-pseudoc.d: Likewise.
* testsuite/gas/bpf/alu-be-pseudoc.d: Likewise.
* testsuite/gas/bpf/alu32-pseudoc.s: Likewise.
* testsuite/gas/bpf/alu32-pseudoc.d: Likewise.
* testsuite/gas/bpf/alu32-be-pseudoc.d: Likewise.
This patch adds support for the V4 BPF instruction jal/gotol, which is
like ja/goto but it supports a signed 32-bit PC-relative (in number of
64-bit words minus one) target operand instead of the 16-bit signed
operand of the other instruction. This greatly increases the jump
range in BPF programs.
Tested in bpf-unkown-none.
bfd/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* reloc.c: New reloc BFD_RELOC_BPF_DISPCALL32.
* elf64-bpf.c (bpf_reloc_type_lookup): Handle the new reloc.
* libbfd.h (bfd_reloc_code_real_names): Regenerate.
gas/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* config/tc-bpf.c (struct bpf_insn): New field `id'.
(md_assemble): Save the ids of successfully parsed instructions
and use the new BFD_RELOC_BPF_DISPCALL32 whenever appropriate.
(md_apply_fix): Adapt to the new BFD reloc.
* testsuite/gas/bpf/jump.s: Test JAL.
* testsuite/gas/bpf/jump.d: Likewise.
* testsuite/gas/bpf/jump-pseudoc.d: Likewise.
* testsuite/gas/bpf/jump-be.d: Likewise.
* testsuite/gas/bpf/jump-be-pseudoc.d: Likewise.
* doc/c-bpf.texi (BPF Instructions): Document new instruction
jal/gotol.
Document new operand type disp32.
include/ChangeLog:
2023-07-24 Jose E. Marchesi <jose.marchesi@oracle.com>
* opcode/bpf.h (enum bpf_insn_id): Add entry BPF_INSN_JAL.
(enum bpf_insn_id): Remove spurious entry BPF_INSN_CALLI.
opcodes/ChangeLog:
2023-07-23 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf-opc.c (bpf_opcodes): Add entry for jal.
The DAP start_thread helper function has a 'name' parameter that is
unused. Apparently I forgot to hook it up to the thread constructor.
This patch fixes the oversight.
While working on an experiment, I realized that I needed the DAP
block_signals function. I figured other developers may need it as
well, so this patch moves it from DAP to the gdb module and exports
it.
I also added a new subclass of threading.Thread that ensures that
signals are blocked in the new thread.
Finally, this patch slightly rearranges the documentation so that
gdb-side threading issues and functions are all discussed in a single
node.
This commit adjusts some of the debug output in linux-nat.c, but makes
no other functional changes to GDB.
In resume_lwp I've added the word "sibling" to one of the debug
messages. All the other debug messages in this function talk about
operating on the sibling thread, so I think it makes sense, for
consistency, if the message I've updated also talks about the sibling
thread.
In resume_stopped_resumed_lwps I've reordered the condition checks so
that the vfork-parent check now happens after the checks for whether
the thread is already resumed or not. This makes no functional
difference to GDB, but does, I think, mean we see more helpful debug
messages first.
Consider the situation where a vfork-parent thread is already resumed,
and resume_stopped_resumed_lwps is called. Previously the message
saying that the thread was not being resumed due to being a
vfork-parent, was printed. This might give the impression that the
thread is left in a not resumed state, which is misleading.
After this change we now get a message saying that the thread is not
being resumed due to it not being stopped (i.e. is already resumed).
With this message the already resumed nature of the thread is much
clearer.
I found this change helpful when debugging some vfork related issues.
For *reasons* I was hacking on gdb.base/foll-vfork.exp and wanted to
change the name of the binary that was created. Should be easy, I
adjusted the global $binfile variable .... but that didn't work.
In one place the script uses $testfile instead of $binfile.
Fixed this to use $binfile, now I can easily change the name of the
generated binary, and the test still works.
There's no change in what is tested after this commit.
I noticed in test-case gdb.arch/arm-pthread_cond_timedwait-bt.exp that
prepare_for_testing is used, followed by a clean_restart.
This calls clean_restart twice in a row.
Fix this by using build_executable instead.
Also, I noticed that the test-case requires an SVC instruction, so add a
require to limit the test-case to supported architectures.
While we're at it, run M-x indent-region in emacs to fix indentation.
Tested on x86_64-linux.
