Alan Modra [Sun, 24 Jul 2022 23:55:49 +0000 (09:25 +0930)]
Re: PowerPC64 .branch_lt address
On seeing PR29369 my suspicion was naturally on a recent powerpc64
change, commit 0ab80031430e. Without a reproducer, I spent time
wondering what could have gone wrong, and while I doubt this patch
would have fixed the PR, there are some improvements that can be made
to cater for user silliness.
I also noticed that when -z relro -z now sections are created out of
order, with .got before .plt in the section headers but .got is laid
out at a higher address. That's due to the address expression for
.branch_lt referencing SIZEOF(.got) and so calling init_os (which
creates a bfd section) for .got before the .plt section is created.
Fix that by ignoring SIZEOF in exp_init_os. Unlike ADDR and LOADADDR
which need to reference section vma and lma respectively, SIZEOF can
and does cope with a missing bfd section by returning zero for its
size, which of course is correct.
PR 29369
* ldlang.c (exp_init_os): Don't create a bfd section for SIZEOF.
* emulparams/elf64ppc.sh (OTHER_RELRO_SECTIONS_2): Revise
.branch_lt address to take into account possible user sections
with alignment larger than 8 bytes.
Peter Bergner [Wed, 20 Jul 2022 23:16:05 +0000 (18:16 -0500)]
PowerPC: Create new MMA instruction masks and use them
The MMA instructions use XX3_MASK|3<<21 as an instruction mask, but that
misses the RC bit/bit 31, so if we disassemble a .long that represents an
MMA instruction except that it also has bit 31 set, we will erroneously
disassemble it to that MMA instruction. We create new masks defines that
contain bit 31 so that doesn't happen anymore.
Jan Beulich [Thu, 21 Jul 2022 10:37:07 +0000 (12:37 +0200)]
x86: replace wrong attributes on VCVTDQ2PH{X,Y}
A standalone (without SAE) StaticRounding attribute is meaningless, and
indeed all other similar insns have ATTSyntax there instead. I can only
assume this was some strange copy-and-paste mistake.
Jan Beulich [Thu, 21 Jul 2022 10:36:44 +0000 (12:36 +0200)]
x86/Intel: correct AVX512F scatter insn element sizes
I clearly screwed up in 6ff00b5e12e7 ("x86/Intel: correct permitted
operand sizes for AVX512 scatter/gather") giving all AVX512F scatter
insns Dword element size. Update testcases (also their gather parts),
utilizing that there previously were two identical lines each (for no
apparent reason).
Alan Modra [Thu, 21 Jul 2022 04:01:51 +0000 (13:31 +0930)]
PowerPC64: fix build error on 32-bit hosts
elf64-ppc.c:11673:33: error: format ‘%lx’ expects argument of type ‘long unsigned int’, but argument 3 has type ‘bfd_vma’ {aka ‘long long unsigned int’} [-Werror=format=]
11673 | fprintf (stderr, "offset = %#lx:", stub_entry->stub_offset);
| ~~~^ ~~~~~~~~~~~~~~~~~~~~~~~
| | |
| | bfd_vma {aka long long unsigned int}
| long unsigned int
| %#llx
H.J. Lu [Mon, 18 Jul 2022 18:44:32 +0000 (11:44 -0700)]
x86: Properly check invalid relocation against protected symbol
Only check invalid relocation against protected symbol defined in shared
object.
bfd/
PR ld/29377
* elf32-i386.c (elf_i386_scan_relocs): Only check invalid
relocation against protected symbol defined in shared object.
* elf64-x86-64.c (elf_x86_64_scan_relocs): Likewise.
ld/
PR ld/29377
* testsuite/ld-elf/linux-x86.exp: Run PR ld/29377 tests.
* testsuite/ld-elf/pr29377a.c: New file.
* testsuite/ld-elf/pr29377b.c: Likewise.
Jan Beulich [Mon, 18 Jul 2022 15:31:57 +0000 (17:31 +0200)]
x86: correct VMOVSH attributes
Both forms were missing VexW0 (thus allowing Evex.W=1 to be encoded by
suitable means, which would cause #UD). The memory operand form further
was using the wrong Masking value, thus allowing zeroing-masking to be
encoded for the store form (which would again cause #UD).
Alan Modra [Tue, 12 Jul 2022 01:21:52 +0000 (10:51 +0930)]
PR29355, ld segfaults with -r/-q and custom-named section .rela*
The bug testcase uses an output section named .rel or .rela which has
input .data sections mapped to it. The input .data section has
relocations. When counting output relocations SHT_REL and SHT_RELA
section reloc_count is ignored, with the justification that reloc
sections themselves can't have relocations and some backends use
reloc_count in reloc sections. However, the test wrongly used the
output section type (which normally would match input section type).
Fix that. Note that it is arguably wrong for ld to leave the output
.rel/.rela section type as SHT_REL/SHT_RELA when non-empty non-reloc
sections are written to it, but I'm not going to change that since it
might be useful to hand-craft relocs in a data section that is then
written to a SHT_REL/SHT_RELA output section.
PR 29355
* elflink.c (bfd_elf_final_link): Use input section type rather
than output section type to determine whether to exclude using
reloc_count from that section.
Alan Modra [Thu, 7 Jul 2022 12:33:15 +0000 (22:03 +0930)]
ppc gas: don't leak ppc_hash memory
* config/tc-ppc.c (insn_obstack): New.
(insn_calloc): New function.
(ppc_setup_opcodes): Use insn_obstack for ppc_hash.
(cherry picked from commit a887be69963c40ced36e319e5fb14b3de4b6658b)
Without ppc_md_end since the infrastructure to call that late isn't
available on the branch, and without the multiply overflow check.
Andrew Burgess [Tue, 5 Jul 2022 13:17:14 +0000 (14:17 +0100)]
libopcodes/s390: add support for disassembler styling
This commit adds disassembler style to the libopcodes s390
disassembler. This conversion was pretty straight forward, I just
converted the fprintf_func calls to fprintf_styled_func calls and
added an appropriate style.
For testing the new styling I just assembled then disassembled the
source files in gas/testsuite/gas/s390 and manually checked that the
styling looked reasonable.
If the user does not request styled output from objdump, then there
should be no change in the disassembler output after this commit.
At this point, waitpid has returned an "exited" status for some pid, so
pid is non-zero. Since inferior_ptid is set to null_ptid on entry, the
pid returned by wait is not equal to `inferior_ptid.pid ()`, so we reset
pid to -1 and go to waiting again. Since there are not more children to
wait for, waitpid then returns -1 so we get here:
if (pid == -1)
{
gdb_printf (gdb_stderr,
_("Child process unexpectedly missing: %s.\n"),
safe_strerror (save_errno));
/* Claim it exited with unknown signal. */
ourstatus->set_signalled (GDB_SIGNAL_UNKNOWN);
return inferior_ptid;
}
We therefore return a "signalled" status with a null_ptid (again,
inferior_ptid is null_ptid). This confuses infrun, because if the
target returns a "signalled" status, it should be coupled with a ptid
for an inferior that exists.
So, the first step is to fix the snippets above to not use
inferior_ptid. In the first snippet, use find_inferior_pid to see if
we know the event process. If there is no inferior with that pid, we
assume it's a detached child process to we ignore the event. That
should be enough to fix the problem, because it should make it so we
won't go into the second snippet. But still, fix the second snippet to
return an "ignore" status. This is copied from inf_ptrace_target::wait,
which is where rs6000_nat_target::wait appears to be copied from in the
first place.
These changes, are not sufficient, as the aix_thread_target, which sits
on top of rs6000_nat_target, also relies on inferior_ptid.
aix_thread_target::wait, by calling pd_update, assumes that
rs6000_nat_target has set inferior_ptid to the appropriate value (the
ptid of the event thread), but that's not the case. pd_update
returns inferior_ptid - null_ptid - and therefore
aix_thread_target::wait returns null_ptid too, and we still hit the
assert shown above.
Fix this by changing pd_activate, pd_update, sync_threadlists and
get_signaled_thread to all avoid using inferior_ptid. Instead, they
accept as a parameter the pid of the process we are working on.
With this patch, I am able to run the program to completion:
Pedro Alves [Thu, 7 Jul 2022 12:05:50 +0000 (13:05 +0100)]
Fix pedantically invalid DWARF in gdb.trace/unavailable-dwarf-piece.exp
The DWARF spec says:
Any debugging information entry representing the declaration of an object,
module, subprogram or type may have DW_AT_decl_file, DW_AT_decl_line and
DW_AT_decl_column attributes, each of whose value is an unsigned integer
^^^^^^^^
constant.
Grepping around the DWARF-assembler-based testcases, I noticed that
gdb.trace/unavailable-dwarf-piece.exp emits decl_line with
DW_FORM_sdata, a signed integer form. This commit tweaks it to use
DW_FORM_udata instead.
Unsurprisingly, this:
$ make check \
TESTS="gdb.trace/unavailable-dwarf-piece.exp" \
RUNTESTFLAGS="--target_board=native-gdbserver"
... still passes cleanly for me after this change.
I've noticed this because current llvm-dwarfdump crashed on an
ROCm-internal DWARF-assembler-based testcase that incorrectly used
signed forms for DW_AT_decl_file/DW_AT_decl_line.
The older llvm-dwarfdump found on Ubuntu 20.04 (LLVM 10) reads the
line numbers with signed forms as "0" instead of crashing. Here's the
before/after fix for gdb.trace/unavailable-dwarf-piece.exp with that
llvm-dwarfdump version:
GDB/testsuite: Add coverage for `print -elements' command
We currently have no coverage for the `print -elements ...' command (or
`p -elements ...' in the shortened form), so add a couple of test cases
mimicking ones using corresponding `set print elements ...' values.
Tiezhu Yang [Thu, 7 Jul 2022 06:33:19 +0000 (14:33 +0800)]
gdb: LoongArch: Implement the push_dummy_call gdbarch method
According to "Procedure Calling Convention" in "LoongArch ELF ABI
specification" [1], implement the push_dummy_call gdbarch method
as clear as possible.
Tsukasa OI [Mon, 27 Jun 2022 02:03:43 +0000 (11:03 +0900)]
RISC-V: Added Zfhmin and Zhinxmin.
This commit adds Zfhmin and Zhinxmin extensions (subsets of Zfh and
Zhinx extensions, respectively). In the process supporting Zfhmin and
Zhinxmin extension, this commit also changes how instructions are
categorized considering Zfhmin, Zhinx and Zhinxmin extensions.
Detailed changes,
* From INSN_CLASS_ZFH to INSN_CLASS_ZFHMIN:
flh, fsh, fmv.x.h and fmv.h.x.
* From INSN_CLASS_ZFH to INSN_CLASS_ZFH_OR_ZHINX:
fmv.h.
* From INSN_CLASS_ZFH_OR_ZHINX to INSN_CLASS_ZFH_OR_ZHINX:
* From INSN_CLASS_ZFH_OR_ZHINX to INSN_CLASS_ZFHMIN_OR_ZHINXMIN:
fcvt.s.h and fcvt.h.s.
* From INSN_CLASS_D_AND_ZFH_INX to INSN_CLASS_ZFHMIN_AND_D:
fcvt.d.h and fcvt.h.d.
* From INSN_CLASS_Q_AND_ZFH_INX to INSN_CLASS_ZFHMIN_AND_Q:
fcvt.q.h and fcvt.h.q.
bfd/ChangeLog:
* elfxx-riscv.c (riscv_implicit_subsets): Change implicit
subsets. Zfh->Zicsr is not needed and Zfh->F is replaced with
Zfh->Zfhmin and Zfhmin->F. Zhinx->Zicsr is not needed and
Zhinx->Zfinx is replaced with Zhinx->Zhinxmin and
Zhinxmin->Zfinx.
(riscv_supported_std_z_ext): Added zfhmin and zhinxmin.
(riscv_multi_subset_supports): Rewrite handling for new
instruction classes.
(riscv_multi_subset_supports_ext): Updated.
(riscv_parse_check_conflicts): Change error message to include
zfh and zfhmin extensions.
Tsukasa OI [Mon, 27 Jun 2022 02:03:44 +0000 (11:03 +0900)]
RISC-V: Fix disassembling Zfinx with -M numeric
This commit fixes floating point operand register names from ABI ones
to dynamically set ones.
gas/ChangeLog:
* testsuite/gas/riscv/zfinx-dis-numeric.s: Test new behavior of
Zfinx extension and -M numeric disassembler option.
* testsuite/gas/riscv/zfinx-dis-numeric.d: Likewise.
opcodes/ChangeLog:
* riscv-dis.c (riscv_disassemble_insn): Use dynamically set GPR
names to disassemble Zfinx instructions.
Tsukasa OI [Fri, 24 Jun 2022 02:59:04 +0000 (11:59 +0900)]
RISC-V: Fix requirement handling on Zhinx+{D,Q}
This commit fixes how instructions are masked on Zhinx+Z{d,q}inx.
fcvt.h.d and fcvt.d.h require ((D&&Zfh)||(Zdinx&&Zhinx)) and
fcvt.h.q and fcvt.q.h require ((Q&&Zfh)||(Zqinx&&Zhinx)).
bfd/ChangeLog:
* elfxx-riscv.c (riscv_multi_subset_supports): Fix feature gate
on INSN_CLASS_{D,Q}_AND_ZFH_INX.
(riscv_multi_subset_supports_ext): Fix feature gate diagnostics
on INSN_CLASS_{D,Q}_AND_ZFH_INX.
gas/ChangeLog:
* testsuite/gas/riscv/fp-zhinx-insns.d: Add Zqinx to -march
for proper testing.
Ruud van der Pas [Tue, 28 Jun 2022 17:37:19 +0000 (10:37 -0700)]
gprofng: implement a functional gp-display-html
This patch enables the first support for the "gprofng display html" command.
This command works for C/C++ applications on x86_64. Using one or more gprofng
experiment directories as input, a new directory with html files is created.
Through the index.html file in this directory, the performance results may be
viewed in a browser.
gprofng/Changelog:
2022-06-28 Ruud van der Pas <ruud.vanderpas@oracle.com>
* gp-display-html/gp-display-html.in: implement first support for x86_64 and C/C++
elf: Set p_align to the minimum page size if possible
may ignore p_align of PT_GNU_STACK when copying ELF program header if
the maximum page size is larger than p_align of PT_LOAD segments. Copy
p_align of PT_GNU_STACK since p_align of PT_GNU_STACK describes stack
alignment, not page size,
PR binutils/29319
* elf.c (copy_elf_program_header): Copy p_align of PT_GNU_STACK
for stack alignment.
Jan Beulich [Wed, 6 Jul 2022 13:40:04 +0000 (15:40 +0200)]
x86: make D attribute usable for XOP and FMA4 insns
This once again allows to reduce redundancy in (and size of) the opcode
table.
Don't go as far as also making D work on the two 5-operand XOP insns:
This would significantly complicate the code, as there the first
(immediate) operand would need special treatment in several places.
Note that the .s suffix isn't being enabled to have any effect, for
being deprecated. Whereas neither {load} nor {store} pseudo prefixes
make sense here, as the respective operands are inputs (loads) only
anyway, regardless of order. Hence there is (as before) no way for the
programmer to request the alternative encoding to be used for register-
only insns.
Note further that it is always the first original template which is
retained (and altered), to make sure the same encoding as before is
used for register-only insns. This has the slightly odd (but pre-
existing) effect of XOP register-only insns having XOP.W clear, but FMA4
ones having VEX.W set.
Jan Beulich [Wed, 6 Jul 2022 13:39:03 +0000 (15:39 +0200)]
x86: fix 3-operand insn reverse-matching
The middle operand would have gone entirely unchecked, allowing e.g.
vmovss %xmm0, %esp, %xmm2
to assemble successfully, or e.g.
vmovss %xmm0, $4, %xmm2
causing an internal error. Alongside dealing with this also drop a
related comment, which hasn't been applicable anymore since the
introduction of 3-operand patterns with D set (and which perhaps never
had been logical to be there, as reverse-matched insns don't make it
there in the first place).
Jan Beulich [Wed, 6 Jul 2022 07:22:47 +0000 (09:22 +0200)]
x86: introduce a state stack for .arch
When using just slightly non-trivial combinations of .arch, it can be
quite useful to be able to go back to prior state without needing to
re-invoke perhaps many earlier directives and without needing to invoke
perhaps many "negative" ones. Like some other architectures allow
saving (pushing) and restoring (popping) present/prior state.
For now require the same .code<N> to be in effect for ".arch pop" that
was in effect for the corresponding ".arch push".
Also change the global "no_cond_jump_promotion" to be bool, to match the
new struct field.
Jan Beulich [Wed, 6 Jul 2022 07:22:11 +0000 (09:22 +0200)]
x86: generalize disabling of sub-architectures
I never really understood upon what basis ".arch .no*" options were made
available. Let's not have any "criteria" at all, and simply allow
disabling of all of them. Then we also have all data for a sub-arch in
a single place, as we now only need a single table.
Jan Beulich [Wed, 6 Jul 2022 07:21:40 +0000 (09:21 +0200)]
x86: permit "default" with .arch
So far there was no way to reset the architecture to that assembly would
start with in the absence of any overrides (command line or directives).
Note that for Intel MCU "default" is merely an alias of "iamcu".
While there also zap a stray @item from the doc section, as noticed
when inspecting the generated output (which still has some quirks, but
those aren't easy to address without re-flowing almost the entire
section).
While it may not be necessary in i386_target_format() (but then setting
the variable to NULL also wouldn't be necessary), at least in the other
cases strings may already have accumulated.
Tom de Vries [Tue, 5 Jul 2022 20:41:25 +0000 (22:41 +0200)]
[gdb/exp] Fix internal error when printing C++ pointer-to-member
When running the test-case included with this patch, we run into:
...
(gdb) print ptm^M
$1 = gdb/gdbtypes.h:695: internal-error: loc_bitpos: \
Assertion `m_loc_kind == FIELD_LOC_KIND_BITPOS' failed.^M
...
while printing a c++ pointer-to-member.
Fix this by skipping static fields in cp_find_class_member, such that we have:
...
(gdb) print ptm^M
$1 = &A::i^M
...
Tom Tromey [Mon, 20 Jun 2022 18:32:52 +0000 (12:32 -0600)]
Add gdb.Objfile.is_file attribute
Sometimes an objfile comes from memory and not from a file. It can be
useful to be able to check this from Python, so this patch adds a new
"is_file" attribute.
Tom Tromey [Fri, 3 Jun 2022 13:59:49 +0000 (07:59 -0600)]
Make 'import gdb.events' work
Pierre-Marie noticed that, while gdb.events is a Python module, it
can't be imported. This patch changes how this module is created, so
that it can be imported, while also ensuring that the module is always
visible, just as it was in the past.
This new approach required one non-obvious change -- when running
gdb.base/warning.exp, where --data-directory is intentionally not
found, the event registries can now be nullptr. Consequently, this
patch probably also requires
Xi Ruoyao [Tue, 5 Jul 2022 11:30:12 +0000 (19:30 +0800)]
gdb: LoongArch: add orig_a0 into register set
The basic support for LoongArch has been merged into the upstream Linux
kernel since 5.19-rc1 on June 5, 2022. This commit adds orig_a0 which
is added into struct user_pt_regs [1] to match the upstream kernel, and
then the upstream GDB will work with the upstream kernel.
Note that orig_a0 was added into struct user_pt_regs in the development
cycle for merging LoongArch port into the upstream Linux kernel, so
earlier kernels (notably, the product kernel with version 4.19 used in
distros like UOS and Loongnix) don't have it. Inspect
arch/loongarch/include/uapi/asm/ptrace.h in the kernel tree to make sure.
To build upstream GDB for a kernel lacking orig_a0, it's necessary to
revert this commit locally.
Support for location and range lists for split-dwarf and dwarf-5.
Adding support for location and range lists for split-dwarf and dwarf-5.
Following issues are taken care.
1. Display of the index values for DW_FORM_loclistx and DW_FORM_rnglistx.
2. Display of .debug_loclists.dwo and .debug_rnglists.dwo sections.
* dwarf.c(read_and_display_attr_value): Handle DW_FORM_loclistx
and DW_FORM_rnglistx for .dwo files.
(process_debug_info): Load .debug_loclists.dwo and
.debug_rnglists.dwo if exists.
(load_separate_debug_files): Load .debug_loclists and
.debug_rnglists if exists.
Include 2 entries in debug_displays table.
* dwarf.h (enum dwarf_section_display_enum): Include 2 entries.
Jan Beulich [Tue, 5 Jul 2022 06:40:09 +0000 (08:40 +0200)]
x86: macro-ize cpu_arch[] entries
Putting individual elements behind macros, besides (imo) improving
readability, will make subsequent (and likely also future) changes less
intrusive.
Utilize this right away to pack the table a little more tightly, by
converting "skip" to bool and putting it earlier in a group of bitfields
together with "len".
Alan Modra [Mon, 4 Jul 2022 03:15:47 +0000 (12:45 +0930)]
alloc gas seginfo on notes obstack
Lots of memory used in gas should go on this obstack. The patch also
frees all the gas obstacks on exit, which isn't a completely trivial
task.
* subsegs.c (alloc_seginfo): New function.
(subseg_change, subseg_get): Use it.
(subsegs_end): New function.
* as.h (subsegs_end): Declare.
* output-file.c: Include subsegs.h
(stash_frchain_obs): New function.
(output_file_close): Save obstacks attached to output bfd before
closing. Call subsegs_end with the array of obstacks.
Alan Modra [Mon, 4 Jul 2022 01:54:22 +0000 (11:24 +0930)]
objcopy: bfd_alloc orelocation
This fixes an inconsequential objcopy memory leak. I'd normally
ignore reports of leaks like this one, that are merely one block or
fewer per section processed, since objcopy soon exits and frees all
memory. However I thought it worth providing support for allocating
memory on a bfd objalloc in objcopy and other utils.
PR 29233
* bucomm.c (bfd_xalloc): New function.
* bucomm.h (bfd_xalloc): Declare.
* objcopy.c (copy_relocations_in_section): Use it to allocate
array of reloc pointers. Rewrite code stripping relocs to do
without extra memory allocation.
* dwarf.c(process_debug_info): Include DW_TAG_skeleton_unit.
(display_debug_str_offsets): While dumping .debug_str_offsets.dwo,
pass proper str_offsets_base to fetch_indexed_string().
(load_separate_debug_files): Skip DWO ID dump for dwarf-5.
Marcus Nilsson [Mon, 4 Jul 2022 10:25:42 +0000 (11:25 +0100)]
opcodes/avr: Implement style support in the disassembler
* disassemble.c: (disassemble_init_for_target): Set
created_styled_output for AVR based targets.
* avr-dis.c: (print_insn_avr): Use fprintf_styled_ftype
instead of fprintf_ftype throughout.
(avr_operand): Pass in and fill disassembler_style when
parsing operands.
Tom de Vries [Mon, 4 Jul 2022 08:28:42 +0000 (10:28 +0200)]
[gdb/symtab] Add get/set functions for per_cu->lang/unit_type
The dwarf2_per_cu_data fields lang and unit_type both have a dont-know
initial value (respectively language_unknown and (dwarf_unit_type)0), which
allows us to add certain checks, f.i. checking that that a field is not read
before written.
Add get/set member functions for the two fields as a convenient location to
add such checks, make the fields private to enforce using the member
functions, and add the m_ prefix.
Jan Beulich [Mon, 4 Jul 2022 06:32:50 +0000 (08:32 +0200)]
x86: fold Disp32S and Disp32
The only case where 64-bit code uses non-sign-extended (can also be
considered zero-extended) displacements is when an address size override
is in place for a memory operand (i.e. particularly excluding
displacements of direct branches, which - if at all - are controlled by
operand size, and then are still sign-extended, just from 16 bits).
Hence the distinction in templates is unnecessary, allowing code to be
simplified in a number of places. The only place where logic becomes
more complicated is when signed-ness of relocations is determined in
output_disp().
The other caveat is that Disp64 cannot be specified anymore in an insn
template at the same time as Disp32. Unlike for non-64-bit mode,
templates don't specify displacements for both possible addressing
modes; the necessary adjustment to the expected ones has already been
done in match_template() anyway (but of course the logic there needs
tweaking now). Hence the single template so far doing so is split.
Jan Beulich [Mon, 4 Jul 2022 06:32:20 +0000 (08:32 +0200)]
x86: restore masking of displacement kinds
Commit 7d5e4556a375 rendered the check near the end of what is now
i386_finalize_displacement() entirely dead for AT&T mode, since for
operands involving a displacement .unspecified will always be set. But
the logic there is bogus anyway - Intel syntax operand size specifiers
are of no interest there either. The only thing which matters in the
"displacement only" determination is .baseindex.
Of course when masking displacement kinds we should not at the same time
also mask off other attributes.
Furthermore the type mask returned by lex_got() also needs to be
adjusted: The only case where we want Disp32 (rather than Disp32S) is
when dealing with 32-bit addressing mode in 64-bit code.
Jan Beulich [Mon, 4 Jul 2022 06:31:21 +0000 (08:31 +0200)]
x86-64: improve handling of branches to absolute addresses
There are two related problems here: The use of "addr32" on a direct
branch would, besides causing a warning, result in operands to be
permitted which mistakenly are refused without "addr32". Plus at some
point not too long ago I'm afraid it may have been me who regressed the
relocation addends emitted for such branches. Correct both problems,
adding a testcase to guard against regressing this again.
Tom de Vries [Sat, 2 Jul 2022 11:03:34 +0000 (13:03 +0200)]
[gdb/symtab] Fix data race on per_cu->dwarf_version
When building gdb with -fsanitize=thread and gcc 12, and running test-case
gdb.dwarf2/dwz.exp, we run into a data race between thread T2 and the main
thread in the same write:
...
Write of size 1 at 0x7b200000300c:^M
#0 cutu_reader::cutu_reader(dwarf2_per_cu_data*, dwarf2_per_objfile*, \
abbrev_table*, dwarf2_cu*, bool, abbrev_cache*) gdb/dwarf2/read.c:6252 \
(gdb+0x82f3b3)^M
...
which is here:
...
this_cu->dwarf_version = cu->header.version;
...
Both writes are called from the parallel for in dwarf2_build_psymtabs_hard,
this one directly:
...
#1 process_psymtab_comp_unit gdb/dwarf2/read.c:6774 (gdb+0x8304d7)^M
#2 operator() gdb/dwarf2/read.c:7098 (gdb+0x8317be)^M
#3 operator() gdbsupport/parallel-for.h:163 (gdb+0x872380)^M
...
and this via the PU import:
...
#1 cooked_indexer::ensure_cu_exists(cutu_reader*, dwarf2_per_objfile*, \
sect_offset, bool, bool) gdb/dwarf2/read.c:17964 (gdb+0x85c43b)^M
#2 cooked_indexer::index_imported_unit(cutu_reader*, unsigned char const*, \
abbrev_info const*) gdb/dwarf2/read.c:18248 (gdb+0x85d8ff)^M
#3 cooked_indexer::index_dies(cutu_reader*, unsigned char const*, \
cooked_index_entry const*, bool) gdb/dwarf2/read.c:18302 (gdb+0x85dcdb)^M
#4 cooked_indexer::make_index(cutu_reader*) gdb/dwarf2/read.c:18443 \
(gdb+0x85e68a)^M
#5 process_psymtab_comp_unit gdb/dwarf2/read.c:6812 (gdb+0x830879)^M
#6 operator() gdb/dwarf2/read.c:7098 (gdb+0x8317be)^M
#7 operator() gdbsupport/parallel-for.h:171 (gdb+0x8723e2)^M
...
Fix this by setting the field earlier, in read_comp_units_from_section.
The write in cutu_reader::cutu_reader() is still needed, in case
read_comp_units_from_section is not used (run the test-case with say, target
board cc-with-gdb-index).
Make the write conditional, such that it doesn't trigger if the field is
already set by read_comp_units_from_section. Instead, verify that the
field already has the value that we're trying to set it to.
Move this logic into into a member function set_version (in analogy to the
already present member function version) to make sure it's used consistenly,
and make the field private in order to enforce access through the member
functions, and rename it to m_dwarf_version.
While we're at it, make sure that the version is set before read, to avoid
say returning true for "per_cu.version () < 5" if "per_cu.version () == 0".
Tom de Vries [Sat, 2 Jul 2022 09:50:03 +0000 (11:50 +0200)]
[gdb/testsuite] Fix gdb.base/early-init-file.exp with -fsanitize=thread
When building gdb with -fsanitize=thread, I run into:
...
FAIL: gdb.base/early-init-file.exp: check startup version string has style \
version
...
due to this:
...
warning: Found custom handler for signal 7 (Bus error) preinstalled.^M
warning: Found custom handler for signal 8 (Floating point exception) \
preinstalled.^M
warning: Found custom handler for signal 11 (Segmentation fault) \
preinstalled.^M
Some signal dispositions inherited from the environment (SIG_DFL/SIG_IGN)^M
won't be propagated to spawned programs.^M
...
appearing before the "GNU gdb (GDB) $version" line.
This is similar to the problem fixed by commit f0bbba7886f
("gdb.debuginfod/fetch_src_and_symbols.exp: fix when GDB is built with
AddressSanitizer").
In that commit, the problem was fixed by starting gdb with -quiet, but using
that would mean the "GNU gdb (GDB) $version" line that we're trying to check
would disappear.
Fix this instead by updating the regexp to allow the message.
GDB/doc: Remove extraneous spaces from completion examples
Completion results are usually different when the operation is applied
to a word that is or is not followed by a space. In some cases they are
equivalent, however a space would not be produced if completion was used
earlier on in the word processed.
However in the manual we have completion examples given using a space
that actually prevents the example from working. E.g.:
(gdb) info bre <TAB>
(nothing) and:
(gdb) info bre <TAB><TAB>
Display all 200 possibilities? (y or n)
as it now goes on to propose the entire symbol table, while:
(gdb) info bre<TAB>
(gdb) info breakpoints
does the right thing, but is not what is shown in the manual.
In other cases an extraneous space is used that does not correspond to
the actual completion pattern shown, which gives an impression of
sloppiness.
Remove extraneous spaces then from completion examples as appropriate.
(gdb) complete set print elements
set print elements unlimited
(gdb)
(there is a space entered at the end of both commands, not shown here)
which also means if you strike <Tab> with `set print elements ' input,
it will, annoyingly, complete to `set print elements unlimited' right
away rather than showing a choice between `NUMBER' and `unlimited'.
Add `NUMBER' then as an available completion for such `set' commands:
(gdb) complete set print elements
set print elements NUMBER
set print elements unlimited
(gdb)
Adjust the testsuite accordingly. Also document the feature in the
Completion section of the manual in addition to the Command Options
section already there.
Bruno Larsen [Tue, 7 Jun 2022 19:07:07 +0000 (16:07 -0300)]
gdb/testsuite: Expand gdb.cp/mb-ctor.exp to test dynamic allocation
When testing GDB's ability to stop in constructors, gdb.cp/mb-ctor.exp
only tested objects allocated on the stack. This commit adds a couple of
dynamic allocations and tests if GDB can stop in it as well.
Nick Clifton [Thu, 30 Jun 2022 13:53:02 +0000 (14:53 +0100)]
Fix implementation of readelf's -wE and -wN options,
* dwarf.c (dwarf_select_sections_by_name): If the entry's value is
zero then clear the corresponding variable.
(dwarf_select_sections_by_letters): Likewise.
* testsuite/binutils-all/debuginfo.exp: Expect -WE and -wE
debuginfod tests to fail.
Tom de Vries [Thu, 30 Jun 2022 11:31:06 +0000 (13:31 +0200)]
[gdb] Block SIGTERM in worker threads
With gdb build with gcc-12 and -fsanitize=thread, and test-case
gdb.base/gdb-sigterm.exp, I run into:
...
WARNING: ThreadSanitizer: data race (pid=9722)^M
Write of size 4 at 0x00000325bc68 by thread T1:^M
#0 handle_sigterm(int) src/gdb/event-top.c:1211 (gdb+0x8ec01f)^M
...
Previous read of size 4 at 0x00000325bc68 by main thread:^M
[failed to restore the stack]^M
^M
Location is global 'sync_quit_force_run' of size 4 at \
0x00000325bc68 (gdb+0x325bc68)^M
...
SUMMARY: ThreadSanitizer: data race gdb/event-top.c:1211 in \
handle_sigterm(int)^M
...
and 3 more data races involving handle_sigterm and locations:
- active_ext_lang
- quit_flag
- heap block of size 40
(XNEW (async_signal_handler) in create_async_signal_handler)
This was reported in PR29297.
The testcase executes a "kill -TERM $gdb_pid", which generates a
process-directed signal.
A process-directed signal can be delivered to any thread, and what we see
here is the fallout of the signal being delivered to a worker thread rather
than the main thread.
Fix this by blocking SIGTERM in the worker threads.
[ I have not been able to reproduce this after it occurred for the first time,
so unfortunately I cannot confirm that the patch fixes the problem. ]
Tested on x86_64-linux, with and without -fsanitize=thread.
Andrew Burgess [Thu, 23 Jun 2022 12:55:02 +0000 (13:55 +0100)]
gdb/doc: fix column widths in MI compatibility table
In passing I noticed that the column headings for the table of MI
compatibility and breaking changes, were overlapping, at least when
the PDF is generated on my machine.
I propose giving slightly more space to the two version number
columns, this prevents the headers overlapping for me.
Pedro Alves [Wed, 29 Jun 2022 15:38:43 +0000 (16:38 +0100)]
Fix GDBserver regression due to change to avoid reading shell registers
Simon reported that the recent change to make GDB and GDBserver avoid
reading shell registers caused a GDBserver regression, caught with
ASan while running gdb.server/non-existing-program.exp:
$ /home/smarchi/build/binutils-gdb/gdb/testsuite/../../gdb/../gdbserver/gdbserver stdio non-existing-program
=================================================================
==127719==ERROR: AddressSanitizer: heap-use-after-free on address 0x60f0000000e9 at pc 0x55bcbfa301f4 bp 0x7ffd238a7320 sp 0x7ffd238a7310
WRITE of size 1 at 0x60f0000000e9 thread T0
#0 0x55bcbfa301f3 in scoped_restore_tmpl<bool>::~scoped_restore_tmpl() /home/smarchi/src/binutils-gdb/gdbserver/../gdbsupport/scoped_restore.h:86
#1 0x55bcbfa2ffe9 in post_fork_inferior(int, char const*) /home/smarchi/src/binutils-gdb/gdbserver/fork-child.cc:120
#2 0x55bcbf9c9199 in linux_process_target::create_inferior(char const*, std::__debug::vector<char*, std::allocator<char*> > const&) /home/smarchi/src/binutils-gdb/gdbserver/linux-low.cc:991
#3 0x55bcbf954549 in captured_main /home/smarchi/src/binutils-gdb/gdbserver/server.cc:3941
#4 0x55bcbf9552f0 in main /home/smarchi/src/binutils-gdb/gdbserver/server.cc:4084
#5 0x7ff9d663b0b2 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x240b2)
#6 0x55bcbf8ef2bd in _start (/home/smarchi/build/binutils-gdb/gdbserver/gdbserver+0x1352bd)
0x60f0000000e9 is located 169 bytes inside of 176-byte region [0x60f000000040,0x60f0000000f0)
freed by thread T0 here:
#0 0x7ff9d6c6f0c7 in operator delete(void*) ../../../../src/libsanitizer/asan/asan_new_delete.cpp:160
#1 0x55bcbf910d00 in remove_process(process_info*) /home/smarchi/src/binutils-gdb/gdbserver/inferiors.cc:164
#2 0x55bcbf9c4ac7 in linux_process_target::remove_linux_process(process_info*) /home/smarchi/src/binutils-gdb/gdbserver/linux-low.cc:454
#3 0x55bcbf9cdaa6 in linux_process_target::mourn(process_info*) /home/smarchi/src/binutils-gdb/gdbserver/linux-low.cc:1599
#4 0x55bcbf988dc4 in target_mourn_inferior(ptid_t) /home/smarchi/src/binutils-gdb/gdbserver/target.cc:205
#5 0x55bcbfa32020 in startup_inferior(process_stratum_target*, int, int, target_waitstatus*, ptid_t*) /home/smarchi/src/binutils-gdb/gdbserver/../gdb/nat/fork-inferior.c:515
#6 0x55bcbfa2fdeb in post_fork_inferior(int, char const*) /home/smarchi/src/binutils-gdb/gdbserver/fork-child.cc:111
#7 0x55bcbf9c9199 in linux_process_target::create_inferior(char const*, std::__debug::vector<char*, std::allocator<char*> > const&) /home/smarchi/src/binutils-gdb/gdbserver/linux-low.cc:991
#8 0x55bcbf954549 in captured_main /home/smarchi/src/binutils-gdb/gdbserver/server.cc:3941
#9 0x55bcbf9552f0 in main /home/smarchi/src/binutils-gdb/gdbserver/server.cc:4084
#10 0x7ff9d663b0b2 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x240b2)
previously allocated by thread T0 here:
#0 0x7ff9d6c6e5a7 in operator new(unsigned long) ../../../../src/libsanitizer/asan/asan_new_delete.cpp:99
#1 0x55bcbf910ad0 in add_process(int, int) /home/smarchi/src/binutils-gdb/gdbserver/inferiors.cc:144
#2 0x55bcbf9c477d in linux_process_target::add_linux_process_no_mem_file(int, int) /home/smarchi/src/binutils-gdb/gdbserver/linux-low.cc:425
#3 0x55bcbf9c8f4c in linux_process_target::create_inferior(char const*, std::__debug::vector<char*, std::allocator<char*> > const&) /home/smarchi/src/binutils-gdb/gdbserver/linux-low.cc:985
#4 0x55bcbf954549 in captured_main /home/smarchi/src/binutils-gdb/gdbserver/server.cc:3941
#5 0x55bcbf9552f0 in main /home/smarchi/src/binutils-gdb/gdbserver/server.cc:4084
#6 0x7ff9d663b0b2 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x240b2)
Above we see that in the non-existing-program case, the process gets
deleted before the starting_up flag gets restored to false.
This happens because startup_inferior calls target_mourn_inferior
before throwing an error, and in GDBserver, unlike in GDB, mourning
deletes the process.
Fix this by not using a scoped_restore to manage the starting_up flag,
since we should only clear it when startup_inferior doesn't throw.
Match the whole error message expected to be given rather than omitting
the part about the "unlimited" keyword. There's no point in omitting
the missing part first, and second with an upcoming change the part in
parentheses will no longer be a fixed string, so doing a full match will
ensure the algorithm correctly builds the message expected here. Also
avoid any wildcard matches.
Andrew Burgess [Wed, 22 Jun 2022 10:39:00 +0000 (11:39 +0100)]
gdb/doc: improve description of --data-disassemble opcodes output
Extend the description of the MI command --data-disassemble.
Specifically, expand the description of the 'opcodes' field to explain
how the bytes are formatted.
Yvan Roux [Wed, 29 Jun 2022 12:01:45 +0000 (14:01 +0200)]
gdb/arm: Only stack S16..S31 when FPU registers are secure
The FPCCR.TS bit is used to identify if FPU registers are considered
non-secure or secure. If they are secure, then callee saved registers
(S16 to S31) are stacked on exception entry or otherwise skipped.
Andrew Burgess [Thu, 16 Jun 2022 12:46:41 +0000 (13:46 +0100)]
opcodes/aarch64: split off creation of comment text in disassembler
The function aarch64_print_operand (aarch64-opc.c) is responsible for
converting an instruction operand into the textual representation of
that operand.
In some cases, a comment is included in the operand representation,
though this (currently) only happens for the last operand of the
instruction.
In a future commit I would like to enable the new libopcodes styling
for AArch64, this will allow objdump and GDB[1] to syntax highlight
the disassembler output, however, having operands and comments
combined in a single string like this makes such styling harder.
In this commit, I propose to extend aarch64_print_operand to take a
second buffer. Any comments for the instruction are written into this
extra buffer. The two callers of aarch64_print_operand are then
updated to pass an extra buffer, and print any resulting comment.
In this commit no styling is added, that will come later. However, I
have adjusted the output slightly. Before this commit some comments
would be separated from the instruction operands with a tab character,
while in other cases the comment was separated with two single spaces.
After this commit I use a single tab character in all cases. This
means a few test cases needed updated. If people would prefer me to
move everyone to use the two spaces, then just let me know. Or maybe
there was a good reason why we used a mix of styles, I could probably
figure out a way to maintain the old output exactly if that is
critical.
Other than that, there should be no user visible changes after this
commit.
[1] GDB patches have not been merged yet, but have been posted to the
GDB mailing list:
https://sourceware.org/pipermail/gdb-patches/2022-June/190142.html
projects / thirdparty / binutils-gdb.git / log
