Initially, since c6c37250e98f113755e0d787f7070e2ac80ce77e (in 1999),
in order to fix linking against Microsoft import libraries, ld did
internally rename members of such libraries. At that point, the
criteria for being considered a Microsoft import library was that
every archive member had the same name (no regard for exactly what
that name was).
This was later amended in 44dbf3639f127af46d569ad96b6242dfbc4c0a89
(in 2003) to allow for Microsoft import libraries with intermixed
static object files. At this point, the criteria were extended, so
that all members following the first member named *.dll either had
the exact same member name, or be named *.obj. (Curiously, this would
allow members with any name if it precedes the first one named *.dll.)
In practice, Microsoft style import libraries can contain
members for linking against more than one DLL (built by merging
multiple regular import libraries into one).
Instead of trying to do validation of the whole archive before
considering it a Microsoft style import library, relax the criteria
for doing the member renaming: If an archive member is named *.dll
and it contains .idata sections, assume that that member is a
Microsoft import file, and apply the renaming scheme.
This works for imports for any number of DLLs in the same library,
intermixed with other static object files (regardless of their
names), and vastly simplifies the code.
LLVM generates Microsoft style import libraries, and Rust builds
seem to bundle up multiple import libraries together with some
Rust specific static objects. This fixes linking directly against
them with ld.bfd.
When building with Clang 14 (using gcc 12 libstdc++ headers), I get:
CXX dwarf2/read.o
In file included from /home/simark/src/binutils-gdb/gdb/dwarf2/read.c:94:
/home/simark/src/binutils-gdb/gdb/../gdbsupport/parallel-for.h:142:21: error: 'result_of<(lambda at /home/simark/src/binutils-gdb/gdb/dwarf2/read.c:7124:5) (__gnu_debug::_Safe_iterator<__gnu_cxx::__normal_iterator<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter> *, std::__cxx1998::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>>, std::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>, std::random_access_iterator_tag>, __gnu_debug::_Safe_iterator<__gnu_cxx::__normal_iterator<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter> *, std::__cxx1998::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>>, std::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>, std::random_access_iterator_tag>)>' is deprecated: use 'std::invoke_result' instead [-Werror,-Wdeprecated-declarations]
= typename std::result_of<RangeFunction (RandomIt, RandomIt)>::type;
^
/home/simark/src/binutils-gdb/gdb/dwarf2/read.c:7122:14: note: in instantiation of function template specialization 'gdb::parallel_for_each<__gnu_debug::_Safe_iterator<__gnu_cxx::__normal_iterator<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter> *, std::__cxx1998::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>>, std::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>, std::random_access_iterator_tag>, (lambda at /home/simark/src/binutils-gdb/gdb/dwarf2/read.c:7124:5)>' requested here
= gdb::parallel_for_each (1, per_bfd->all_comp_units.begin (),
^
/usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/12.1.1/../../../../include/c++/12.1.1/type_traits:2597:9: note: 'result_of<(lambda at /home/simark/src/binutils-gdb/gdb/dwarf2/read.c:7124:5) (__gnu_debug::_Safe_iterator<__gnu_cxx::__normal_iterator<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter> *, std::__cxx1998::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>>, std::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>, std::random_access_iterator_tag>, __gnu_debug::_Safe_iterator<__gnu_cxx::__normal_iterator<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter> *, std::__cxx1998::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>>, std::vector<std::unique_ptr<dwarf2_per_cu_data, dwarf2_per_cu_data_deleter>>, std::random_access_iterator_tag>)>' has been explicitly marked deprecated here
{ } _GLIBCXX17_DEPRECATED_SUGGEST("std::invoke_result");
^
/usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/12.1.1/../../../../include/c++/12.1.1/x86_64-pc-linux-gnu/bits/c++config.h:120:45: note: expanded from macro '_GLIBCXX17_DEPRECATED_SUGGEST'
# define _GLIBCXX17_DEPRECATED_SUGGEST(ALT) _GLIBCXX_DEPRECATED_SUGGEST(ALT)
^
/usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/12.1.1/../../../../include/c++/12.1.1/x86_64-pc-linux-gnu/bits/c++config.h:96:19: note: expanded from macro '_GLIBCXX_DEPRECATED_SUGGEST'
__attribute__ ((__deprecated__ ("use '" ALT "' instead")))
^
It complains about the use of std::result_of, which is deprecated in
C++17 and removed in C++20:
https://en.cppreference.com/w/cpp/types/result_of
Given we'll have to transition to std::invoke_result eventually, make a
GDB wrapper to mimimc std::invoke_result, which uses std::invoke_result
for C++ >= 17 and std::result_of otherwise. This way, it will be easy
to remove the wrapper in the future, just replace gdb:: with std::.
Tested by building with gcc 12 in -std=c++11 and -std=c++17 mode, and
clang in -std=c++17 mode (I did not test fully with clang in -std=c++11
mode because there are other unrelated issues).
Change-Id: I50debde0a3307a7bc67fcf8fceefda51860efc1d
GDB overwrites Python's sys.stdout and sys.stderr, but does not
properly implement the 'flush' method -- it only ever will flush
stdout. This patch fixes the bug. I couldn't find a straightforward
way to write a test for this.
PR 29529
* dwarf2.c (struct line_info_table): Add new field:
use_dir_and_file_0.
(concat_filename): Use new field to help select the correct table
slot.
(read_formatted_entries): Do not skip entry 0.
(decode_line_info): Set new field depending upon the version of
DWARF being parsed. Initialise filename based upon the setting of
the new field.
The print_one_detail_ranged_breakpoint has been renamed to
ranged_breakpoint::print_one_detail in this commit:
commit ec45bb676c9c69c30783bcf35ffdac8280f3b8bc
Date: Sat Jan 15 16:34:51 2022 -0700
Convert ranged breakpoints to vtable ops
So their comments should be updated as well.
Running configure and make in binutils-gdb.
$ ./configure
$ make
In file included from ./as.h:37,
from ./config/loongarch-lex.l:21,
from config/loongarch-lex-wrapper.c:20:
./config.h:206: error: “PACKAGE” redefined [-Werror]
#define PACKAGE "gas"
...
gas/config
* loongarch-lex-wrapper.c
Three-part patch set from Tsukasa OI to support zmmul in assembler.
The 'Zmmul' is a RISC-V extension consisting of only multiply instructions
(a subset of 'M' which has multiply and divide instructions).
bfd/
* elfxx-riscv.c (riscv_implicit_subsets): Add 'Zmmul' implied by 'M'.
(riscv_supported_std_z_ext): Add 'Zmmul' extension.
(riscv_multi_subset_supports): Add handling for new instruction class.
gas/
* testsuite/gas/riscv/attribute-09.d: Updated implicit 'Zmmul' by 'M'.
* testsuite/gas/riscv/option-arch-02.d: Likewise.
* testsuite/gas/riscv/m-ext.s: New test.
* testsuite/gas/riscv/m-ext-32.d: New test (RV32).
* testsuite/gas/riscv/m-ext-64.d: New test (RV64).
* testsuite/gas/riscv/zmmul-32.d: New expected output.
* testsuite/gas/riscv/zmmul-64.d: Likewise.
* testsuite/gas/riscv/m-ext-fail-xlen-32.d: New test (failure
by using RV64-only instructions in RV32).
* testsuite/gas/riscv/m-ext-fail-xlen-32.l: Likewise.
* testsuite/gas/riscv/m-ext-fail-zmmul-32.d: New failure test
(RV32 + Zmmul but with no M).
* testsuite/gas/riscv/m-ext-fail-zmmul-32.l: Likewise.
* testsuite/gas/riscv/m-ext-fail-zmmul-64.d: New failure test
(RV64 + Zmmul but with no M).
* testsuite/gas/riscv/m-ext-fail-zmmul-64.l: Likewise.
* testsuite/gas/riscv/m-ext-fail-noarch-64.d: New failure test
(no Zmmul or M).
* testsuite/gas/riscv/m-ext-fail-noarch-64.l: Likewise.
include/
* opcode/riscv.h (enum riscv_insn_class): Added INSN_CLASS_ZMMUL.
ld/
* testsuite/ld-riscv-elf/attr-merge-arch-01.d: We don't care zmmul in
these testcases, so just replaced m by a.
* testsuite/ld-riscv-elf/attr-merge-arch-01a.s: Likewise.
* testsuite/ld-riscv-elf/attr-merge-arch-01b.s: Likewise.
* testsuite/ld-riscv-elf/attr-merge-arch-02.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-arch-02a.s: Likewise.
* testsuite/ld-riscv-elf/attr-merge-arch-03.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-arch-03a.s: Likewise.
* testsuite/ld-riscv-elf/attr-merge-user-ext-01.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-user-ext-rv32i2p1_a2p0.s: Renamed.
* testsuite/ld-riscv-elf/attr-merge-user-ext-rv32i2p1_a2p1.s: Renamed.
opcodes/
* riscv-opc.c (riscv_opcodes): Updated multiply instructions to zmmul.
When running the included test-case, we run into:
...
(gdb) break _start^M
read.h:309: internal-error: set_length: \
Assertion `m_length == length' failed.^M
...
The problem is that while there are two CUs:
...
$ readelf -wi debug-names-missing-cu | grep @
Compilation Unit @ offset 0x0:
Compilation Unit @ offset 0x2d:
...
the CU table in the .debug_names section only contains the first one:
...
CU table:
[ 0] 0x0
...
The incomplete CU table makes create_cus_from_debug_names_list set the size of
the CU at 0x0 to the actual size of both CUs combined.
This eventually leads to the assert, when we read the actual size from the CU
header.
While having an incomplete CU table in a .debug_names section is incorrect,
we need a better failure mode than asserting.
The easiest way to fix this is to set the length to 0 (meaning: unkown) in
create_cus_from_debug_names_list.
This makes the failure mode to accept the incomplete CU table, but to ignore
the missing CU.
It would be nice to instead reject the .debug_names index, and build a
complete CU list, but the point where we find this out is well after
dwarf2_initialize_objfile, so it looks rather intrusive to restart at that
point.
Tested on x86_64-linux.
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=29453
I tried out the script gdb/gdb_mbuild.sh, and ran into:
...
score-elf ...
... configure --target=score-elf
... make score-elf
... run score-elf
score-elf: gdb dumped core
Terminated
...
Gdb runs into this internal error in initialize_current_architecture:
...
if (! gdbarch_update_p (info))
internal_error (__FILE__, __LINE__,
_("initialize_current_architecture: Selection of "
"initial architecture failed"));
...
The call to gdbarch_update_p fails because commit 575b4c298a6 ("gdb: Remove
support for S+core") removed support for the architecture.
Fix this by adding score-*-* to the list of obsolete targets in
gdb/configure.tgt, such that we're no longer able to build the configuration:
...
*** Configuration score-unknown-elf is obsolete.
*** Support has been REMOVED.
make: *** [Makefile:12806: configure-gdb] Error 1
...
Also remove the related line from the "Target Instruction Set Architectures"
list in gdb/MAINTAINERS, such that gdb/gdb_mbuild.sh no longer tries to build
it.
out_inc_line_addr and relax_inc_line_addr are passed INT_MAX as
line_delta to flag end of section. This filters its way down to
size_inc_line_addr and emit_inc_line_addr. Pass line_delta on to
scale_addr_delta where it can be used to omit an unaligned opcode
error.
PR 29494
* dwarf2dbg.c (scale_addr_delta): Delete unnecessary forward decl.
Add line_delta param. Don't print error at end of section, just
round the address down.
(size_inc_line_addr, emit_inc_line_addr): Adjust calls.
Similar to 911438f9f4 ("gdbsupport: fix array-view compilation with
c++11 && _GLIBCXX_DEBUG"), but for gdb::optional.
I get this error when building with Clang 14 and -std=c++11:
CXX agent.o
In file included from /home/simark/src/binutils-gdb/gdbsupport/agent.cc:20:
In file included from /home/simark/src/binutils-gdb/gdbsupport/common-defs.h:210:
In file included from /home/simark/src/binutils-gdb/gdbsupport/common-debug.h:23:
/home/simark/src/binutils-gdb/gdbsupport/../gdbsupport/gdb_optional.h:213:5: error: use of this statement in a constexpr function is a C++14 extension [-Werror,-Wc++14-extensions]
gdb_assert (this->has_value ());
^
/home/simark/src/binutils-gdb/gdbsupport/gdb_assert.h:35:3: note: expanded from macro 'gdb_assert'
((void) ((expr) ? 0 : \
^
Change-Id: If0cf55607fc9dbd1925ccb97cd9abbf8993ff264
Change from int to target_waitkind, which is really what is is. While
at it, remove some outdated doc. The return value is described by a
relatively self-describing enum, not a numerical value like the doc
says.
Change-Id: Id899c853a857c7891c45e5b1639024067d5b59cd
Factor out the code that checks that a value is yes/no or yes/no/auto.
Add two macros to gdbsupport/common.m4 and use them in gdb/configure.ac
I inspected the changes to configure. Other than whitespace changes, we
have some benign changes to the error messages (one of them had an error
actually). There are changes to the --enable-source-highlight and
--enable-libbacktrace handling, but setting enable_source_highlight /
enable_libbacktrace was not really useful anyway, they already had the
right value.
Change-Id: I92587aec36874309e1605e2d60244649f09a757a
The fail was due to -Werror and headers included by dlfcn.h and
elf-bfd.h disagreeing about AT_DCACHEBSIZE and other AT_*. Not a
serious problem obviously, since release versions of binutils don't
enable -Werror and the defines are not used. Anyway, reduce the
number of files that might hit this problem by only including dlfcn.h
where it is needed.
PR 12265
* sysdep.h: Don't include dlfcn.h here.
* plugin.c: Include it here.
Compared to the previous version, this version fixes the comments reported by
Tom Tromey and ensures that the 'help some-user-documented-alias'
shows the alias definition to ensure the user understands this is an
alias even if specifically documented.
When using 'help ALIASNAME', GDB shows the help of the aliased command.
This is a good default behaviour.
However, GDB alias command allows to define aliases with arguments
possibly changing or tuning significantly the behaviour of
the aliased command. In such a case, showing the help of the aliased
command might not be ideal.
This is particularly true when defining an alias as a set of
nested 'with' followed by a last command to launch, such as:
(gdb) alias pp10 = with print pretty -- with print elements 10 -- print
Asking 'help pp10' shows the help of the 'with' command, which is
not particularly useful:
(gdb) help pp10
with, pp10, w
alias pp10 = with print pretty -- with print elements 10 -- print
Temporarily set SETTING to VALUE, run COMMAND, and restore SETTING.
Usage: with SETTING [VALUE] [-- COMMAND]
....
Such an alias can now be documented by the user:
(gdb) document pp10
>Pretty printing an expressiong, printing 10 elements.
>Usage: pp10 [PRINT-COMMAND-OPTIONS] EXP
>See 'help print' for more information.
>end
(gdb) help pp10
alias pp10 = with print pretty -- with print elements 10 -- print
Pretty printing an expressiong, printing 10 elements.
Usage: pp10 [PRINT-COMMAND-OPTIONS] EXP
See 'help print' for more information.
(gdb)
When a user-defined alias is documented specifically, help and apropos
use the provided alias documentation instead of the documentation of
the aliased command.
Such a documented alias is also not shown anymore in the help of the
aliased command, and the alias is not listed anymore in the help
of the aliased command. In particular for cases such as pp10 example above,
indicating that pp10 is an alias of the 'with' command is confusing.
On x86, the PLT entry in executable may be used as function address for
functions in shared libraries. If functions are protected, the function
address used in executable can be different from the function address
used in shared library. This will lead to incorrect run-time behavior
if function pointer equality is needed. By default, x86 linker issues
an error in this case.
On Solaris, linker issued an error for
struct tm *tb = (kind == CPP_time_kind::FIXED ? gmtime : localtime) (&tt);
where gmtime is a protected function in libc.so. Use gmtime's PLT entry
in executable as function address is safe since function pointer equality
isn't needed. Ignore protected visibility in shared libraries on Solaris
to disable linker error. If function pointer equality is needed, linker
will silently generate executable with incorrect run-time behavior on
Solaris.
PR ld/29512
* elf32-i386.c (elf_i386_scan_relocs): Ignore protected
visibility in shared libraries on Solaris.
* elf64-x86-64.c (elf_x86_64_scan_relocs): Likewise.
PR 29517
* dwarf2dbg.c (GAS_ABBREV_COMP_UNIT): New defined constant.
(GAS_ABBREV_SUBPROG): New defined constant.
(GAS_ABBREV_NO_TYPE): New defined constant.
(out_debug_abbrev): Use the new defined constants when emitting
abbreviation numbers. Generate an abbreviation for an unspecified
type.
(out_debug_info): Use the new defined constants when referring to
abbreviations. Generate a use of the no_type abbreviation.
Reference the use when generating DIEs for functions.
* testsuite/gas/elf/dwarf-3-func.d: Update to allow for newly
extended output from the assembler.
* testsuite/gas/elf/dwarf-5-func-global.d: Likewise.
* testsuite/gas/elf/dwarf-5-func-local.d: Likewise.
* testsuite/gas/elf/dwarf-5-func.d: Likewise.
PR 29519
* config/tc-aarch64.c (s_unreq): Use find_end_of_line().
(s_aarch64_cpu): Likewise.
(s_aarch64_arch): Likewise.
(s_aarch64_arch_extension): Likewise.
* testsuite/gas/aarch64/pr29519.d: New test driver file.
* testsuite/gas/aarch64/pr29519.s: New test source file.
Don't overload bfd_reloc_outofrange with what is really a domain error
(target at odd address), or an overflow.
PR 11290
* reloc.c (bfd_reloc_other): Correct comment.
* elf32-avr.c (avr_final_link_relocate): Return bfd_reloc_other
for unaligned reloc target values. Return bfd_reloc_overflow
when stubs are too far away and when R_AVR_LDS_STS_16,
R_AVR_PORT6, or R_AVR_PORT5 overflow.
(elf32_avr_relocate_section): Report more descriptive relocation
errors.
* bfd-in2.h: Regenerate.
On Windows, filename_cmp is case insensitive, but when cross compiling
with libraries that may contain members with uppercase file names, we
should keep those comparisons case insensitive when running the build
tools on other OSes too.
Also make the check for .def consistent with the other ones, fixing
out of bounds reads if file names are shorter than 4 characters.
There's been a long-standing bug in the arm backend where
target-specific directives did not correctly handle lines with
multiple statements. This patch fixes the issue for all the cases
I've been able to find.
It does result in a slight change in behaviour when errors are
encountered: where, previously,
.cpu arm6 bar
would result in the error "junk at end of line, first unrecognized
character is `b'", we now get "unknown cpu `arm6 bar'", which I think
is slightly more helpful anyway. Similar errors are generated for
other directives.
When debugging a certain class of GDB bug, I often end up wanting to
know what GDB thinks the frame-id is in a particular frame. It's
not too hard to pull this from some debug output, but I thought it
might be nice if there was a maintenance command that could tell us.
This commit adds 'maint print frame-id' which prints the frame-id of
the currently selected frame. You can also pass a frame level number
to find the frame-id for a specific frame.
There's a new test too.
Fix the bug that can not generate func@plt
when linking a undefined function with cmodel=medium.
Add testcase.
bfd/
* elfnn-loongarch.c
ld/testsuite/ld-loongarch-elf/
* cmodel-libjirl.dd
* cmodel.exp
* libjirl.s
I don't think it makes any sense for a SHT_RELR section to specify a
symbol table with sh_link. SHT_RELR relocations don't use symbols.
There is no real need to specify sh_info either, SHT_RELR is not for
relocatable object files. Anyway, fuzzers of course don't restrict
themselves to even half-sensible objects. So they found a hole in
objcopy using a non-alloc SHT_RELR in an ET_EXEC. In that case BFD
set up the SHT_RELR section as if it were a SHT_REL against the
sh_info target section. When it came to reading in the target section
relocs, the count was horribly wrong which caused a buffer overflow.
* elf.c (bfd_section_from_shdr <SHT_RELR>): Always just make a
normal section, don't treat it as a reloc section.
PR 29489
* dlltool.c (deterministic): New variable.
(gen_lib_file): If deterministic is true set the
BFD_DETERMINISTIC_OUTPUT flag.
(usage): Mention --deterministic-libraries and
--non-deterministic-libraries.
(long_options): Add new options.
(main): Parse new options.
* doc/binutils.texi: Document the new options.
* NEWS: Mention the new feature.
This implements target async for Windows. The basic idea is to have
the worker thread block in WaitForDebugEvent, then notify the event
loop when an event is seen. In a few situations, this blocking
behavior is undesirable, so the functions passed to do_synchronously
are changed to return a boolean indicating which behavior is needed.
On Windows, certain debugging APIs can only be called from the thread
that started (or attached) to the inferior. Also, there is no way on
Windows to wait for a debug event in addition to other events.
Therefore, in order to implement target async for Windows, gdb will
have to call some functions in a worker thread.
This patch implements the worker thread and moves the necessary
operations there. Target async isn't yet implemented, so this patch
does not cause any visible changes.
