Discussion:
gdb failure: "core file format not supported"
(too old to reply)
Finn Thain
2023-03-10 09:40:01 UTC
Permalink
I've been attempting to debug a dash crash (reported in a different
thread). I was able to capture some core dumps but gdb doesn't like them.

***@panmac:~# file core.?
core.0: ELF 32-bit MSB core file, Motorola m68k, 68020, version 1 (SYSV),
SVR4-style, from '/bin/sh /etc/init.d/mountkernfs.sh reload', real uid: 0,
effective uid: 0, real gid: 0, effective gid: 0, execfn:
'/etc/init.d/mountkernfs.sh'
core.1: ELF 32-bit MSB core file, Motorola m68k, 68020, version 1 (SYSV),
SVR4-style, from '/bin/sh /etc/rcS.d/S08mountall.sh start', real uid: 0,
effective uid: 0, real gid: 0, effective gid: 0, execfn:
'/etc/rcS.d/S08mountall.sh'
core.2: ELF 32-bit MSB core file, Motorola m68k, 68020, version 1 (SYSV),
SVR4-style, from '/bin/sh /etc/rcS.d/S08mountall.sh start', real uid: 0,
effective uid: 0, real gid: 0, effective gid: 0, execfn:
'/etc/rcS.d/S08mountall.sh'
***@panmac:~# gdb --core core.0
GNU gdb (Debian 13.1-2) 13.1
Copyright (C) 2023 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "m68k-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word".
"/root/core.0": Core file format not supported
(gdb) quit
***@panmac:~#
***@panmac:~# gdb /bin/dash core.0
GNU gdb (Debian 13.1-2) 13.1
Copyright (C) 2023 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "m68k-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /bin/dash...
Reading symbols from
/usr/lib/debug/.build-id/aa/4160f84f3eeee809c554cb9f3e1ef0686b8dcc.debug...
"/root/core.0": Core file format not supported
(gdb) quit
***@panmac:~#


I tried a different system (Gentoo/m68k, gdb 10.2) which I installed a
couple of years ago and it works there.

Has anyone else encountered this?
Finn Thain
2023-03-22 02:00:01 UTC
Permalink
[Cc: linux-m68k]
Post by Finn Thain
GNU gdb (Debian 13.1-2) 13.1
Copyright (C) 2023 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "m68k-linux-gnu".
Type "show configuration" for configuration details.
<https://www.gnu.org/software/gdb/bugs/>.
<http://www.gnu.org/software/gdb/documentation/>.
For help, type "help".
Type "apropos word" to search for commands related to "word".
"/root/core.0": Core file format not supported
(gdb) quit
I was able to run gdb under gdb to get a few hints. Here's the error
message:


if (!m_core_gdbarch
|| !gdbarch_iterate_over_regset_sections_p (m_core_gdbarch))
error (_("\"%s\": Core file format not supported"),
bfd_get_filename (core_bfd));


And the reason for the message:


bool
gdbarch_iterate_over_regset_sections_p (struct gdbarch *gdbarch)
{
gdb_assert (gdbarch != NULL);
return gdbarch->iterate_over_regset_sections != NULL;
}


So gdbarch->iterate_over_regset_sections == NULL somehow:


(gdb) print *m_core_gdbarch
$3 = {initialized_p = true, obstack = {<obstack> = {chunk_size = 4064,
chunk = 0x8092bf80, object_base = 0x8092bfb8 "",
next_free = 0x8092bfb8 "", chunk_limit = 0x8092cf60 "", temp = {
i = 853591194, p = 0x32e0c49a}, alignment_mask = 1, chunkfun = {
plain = 0x800791d8 <xmalloc(size_t)>,
extra = 0x800791d8 <xmalloc(size_t)>}, freefun = {
plain = 0x8007b5d2 <xfree<void>(void*)>,
extra = 0x8007b5d2 <xfree<void>(void*)>}, extra_arg = 0x3300c49a,
use_extra_arg = 0, maybe_empty_object = 0,
alloc_failed = 0}, <No data fields>}, registry_fields = {
m_fields = std::vector of length 23, capacity 23 = {0x0, 0x0, 0x0, 0x0,
0x8090b608, 0x8090b5f8, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}},
bfd_arch_info = 0x80795aae <bfd_m68k_arch>, byte_order = BFD_ENDIAN_BIG,
byte_order_for_code = BFD_ENDIAN_BIG, osabi = GDB_OSABI_SVR4,
target_desc = 0x0, tdep = 0x808e7fb0,
dump_tdep = 0x8028c242 <m68k_dump_tdep(gdbarch*, ui_file*)>, nr_data = 0,
data = 0x0, short_bit = 16, int_bit = 32, long_bit = 32, long_long_bit = 64,
bfloat16_bit = 16, bfloat16_format = 0x807b545a <floatformats_bfloat16>,
half_bit = 16, half_format = 0x807b54ba <floatformats_ieee_half>,
float_bit = 32, float_format = 0x807b54b2 <floatformats_ieee_single>,
double_bit = 64, double_format = 0x807b54aa <floatformats_ieee_double>,
long_double_bit = 96,
long_double_format = 0x807b548a <floatformats_m68881_ext>, wchar_bit = 32,
wchar_signed = 1,
floatformat_for_type = 0x8007d812 <default_floatformat_for_type(gdbarch*, char const*, int)>, ptr_bit = 32, addr_bit = 32, dwarf2_addr_size = 4,
char_signed = 1, read_pc = 0x0, write_pc = 0x0,
virtual_frame_pointer = 0x8007f5c2 <legacy_virtual_frame_pointer(gdbarch*, unsigned long long, int*, long long*)>, pseudo_register_read = 0x0,
pseudo_register_read_value = 0x0, pseudo_register_write = 0x0,
num_regs = 29, num_pseudo_regs = 0, ax_pseudo_register_collect = 0x0,
ax_pseudo_register_push_stack = 0x0, report_signal_info = 0x0,
sp_regnum = 15, pc_regnum = 17, ps_regnum = 16, fp0_regnum = 18,
stab_reg_to_regnum = 0x8007b4ba <no_op_reg_to_regnum(gdbarch*, int)>,
ecoff_reg_to_regnum = 0x8007b4ba <no_op_reg_to_regnum(gdbarch*, int)>,
sdb_reg_to_regnum = 0x8007b4ba <no_op_reg_to_regnum(gdbarch*, int)>,
dwarf2_reg_to_regnum = 0x8028c298 <m68k_dwarf_reg_to_regnum(gdbarch*, int)>,
register_name = 0x8028c248 <m68k_register_name(gdbarch*, int)>,
register_type = 0x8028ca62 <m68k_register_type(gdbarch*, int)>,
dummy_id = 0x8028e526 <m68k_dummy_id(gdbarch*, frame_info_ptr)>,
deprecated_fp_regnum = -1,
push_dummy_call = 0x8028c838 <m68k_push_dummy_call(gdbarch*, value*, regcache*, CORE_ADDR, int, value**, CORE_ADDR, function_call_return_method, CORE_ADDR)>, call_dummy_location = AT_ENTRY_POINT, push_dummy_code = 0x0,
code_of_frame_writable = 0x8007b4a2 <default_code_of_frame_writable(gdbarch*, frame_info_ptr)>,
print_registers_info = 0x8022edfc <default_print_registers_info(gdbarch*, ui_file*, frame_info_ptr, int, int)>,
print_float_info = 0x8022f840 <default_print_float_info(gdbarch*, ui_file*, frame_info_ptr, char const*)>, print_vector_info = 0x0,
register_sim_regno = 0x8007f228 <legacy_register_sim_regno(gdbarch*, int)>,
cannot_fetch_register = 0x8007b4dc <cannot_register_not(gdbarch*, int)>,
cannot_store_register = 0x8007b4dc <cannot_register_not(gdbarch*, int)>,
get_longjmp_target = 0x0, believe_pcc_promotion = 1,
convert_register_p = 0x8028cd58 <m68k_convert_register_p(gdbarch*, int, type*)>,
register_to_value = 0x8028e926 <m68k_register_to_value(frame_info_ptr, int, type*, gdb_byte*, int*, int*)>,
value_to_register = 0x8028e700 <m68k_value_to_register(frame_info_ptr, int, type*, gdb_byte const*)>,
value_from_register = 0x801f5370 <default_value_from_register(gdbarch*, type*, int, frame_id)>,
pointer_to_address = 0x801f5092 <unsigned_pointer_to_address(gdbarch*, type*, unsigned char const*)>,
address_to_pointer = 0x801f50ee <unsigned_address_to_pointer(gdbarch*, type*, unsigned char*, unsigned long long)>, integer_to_address = 0x0,
return_value = 0x8028c04a <m68k_svr4_return_value(gdbarch*, value*, type*, regcache*, gdb_byte*, gdb_byte const*)>,
get_return_buf_addr = 0x8007b50e <default_get_return_buf_addr(type*, frame_info_ptr)>,
return_in_first_hidden_param_p = 0x8028bc3c <m68k_return_in_first_hidden_param_p(gdbarch*, type*)>,
skip_prologue = 0x8028c7e2 <m68k_skip_prologue(gdbarch*, CORE_ADDR)>,
skip_main_prologue = 0x0, skip_entrypoint = 0x0,
inner_than = 0x8007b984 <core_addr_lessthan(unsigned long long, unsigned long long)>,
breakpoint_from_pc = 0x80081c1c <default_breakpoint_from_pc(gdbarch*, unsigned long long*, int*)>,
breakpoint_kind_from_pc = 0x8028bc56 <bp_manipulation<2, &m68k_break_insn>::kind_from_pc(gdbarch*, CORE_ADDR*)>,
sw_breakpoint_from_kind = 0x8028bc5a <bp_manipulation<2, &m68k_break_insn>::bp_from_kind(gdbarch*, int, int*)>,
breakpoint_kind_from_current_state = 0x80081b28 <default_breakpoint_kind_from_current_state(gdbarch*, regcache*, unsigned long long*)>,
adjust_breakpoint_address = 0x0,
memory_insert_breakpoint = 0x802a5ba4 <default_memory_insert_breakpoint(gdbarch*, bp_target_info*)>,
memory_remove_breakpoint = 0x802a5c34 <default_memory_remove_breakpoint(gdbarch*, bp_target_info*)>, decr_pc_after_break = 0,
deprecated_function_start_offset = 0,
remote_register_number = 0x8007b770 <default_remote_register_number(gdbarch*, int)>, fetch_tls_load_module_address = 0x0, get_thread_local_address = 0x0,
frame_args_skip = 8,
unwind_pc = 0x8028e442 <m68k_unwind_pc(gdbarch*, frame_info_ptr)>,
unwind_sp = 0x801fa23c <default_unwind_sp(gdbarch*, frame_info_ptr)>,
frame_num_args = 0x0,
frame_align = 0x8028bc30 <m68k_frame_align(gdbarch*, CORE_ADDR)>,
stabs_argument_has_addr = 0x8007b4e4 <default_stabs_argument_has_addr(gdbarch*, type*)>, frame_red_zone_size = 0,
convert_from_func_ptr_addr = 0x8007b4b0 <convert_from_func_ptr_addr_identity(gdbarch*, unsigned long long, target_ops*)>,
addr_bits_remove = 0x8007b4a6 <core_addr_identity(gdbarch*, unsigned long long)>,
remove_non_address_bits = 0x8007b756 <default_remove_non_address_bits(gdbarch*, unsigned long long)>,
memtag_to_string = 0x8007b514 <default_memtag_to_string[abi:cxx11](gdbarch*, value*)>,
tagged_address_p = 0x8007b47c <default_tagged_address_p(gdbarch*, value*)>,
memtag_matches_p = 0x8007b480 <default_memtag_matches_p(gdbarch*, value*)>,
set_memtags = 0x8007b484 <default_set_memtags(gdbarch*, value*, unsigned int, std::vector<unsigned char, gdb::default_init_allocator<unsigned char, std::allocator<unsigned char> > > const&, memtag_type)>,
get_memtag = 0x8007b488 <default_get_memtag(gdbarch*, value*, memtag_type)>,
memtag_granule_size = 0, software_single_step = 0x0,
single_step_through_delay = 0x0,
print_insn = 0x8007b696 <default_print_insn(unsigned long, disassemble_info*)>,
skip_trampoline_code = 0x8007b48e <generic_skip_trampoline_code(frame_info_ptr, unsigned long long)>, so_ops = 0x80791610 <solib_target_so_ops>,
skip_solib_resolver = 0x8007b494 <generic_skip_solib_resolver(gdbarch*, unsigned long long)>,
in_solib_return_trampoline = 0x8007b49a <generic_in_solib_return_trampoline(gdbarch*, unsigned long long, char const*)>,
in_indirect_branch_thunk = 0x8007b502 <default_in_indirect_branch_thunk(gdbarch*, unsigned long long)>,
stack_frame_destroyed_p = 0x8007b49e <generic_stack_frame_destroyed_p(gdbarch*, unsigned long long)>, elf_make_msymbol_special = 0x0,
coff_make_msymbol_special = 0x8007b4c0 <default_coff_make_msymbol_special(int, minimal_symbol*)>,
make_symbol_special = 0x8007b4c2 <default_make_symbol_special(symbol*, objfile*)>,
adjust_dwarf2_addr = 0x8007b4c4 <default_adjust_dwarf2_addr(unsigned long long)>,
adjust_dwarf2_line = 0x8007b4ce <default_adjust_dwarf2_line(unsigned long long, int)>, cannot_step_breakpoint = 0, have_nonsteppable_watchpoint = 0,
address_class_type_flags = 0x0, address_class_type_flags_to_name = 0x0,
execute_dwarf_cfa_vendor_op = 0x8007b4d8 <default_execute_dwarf_cfa_vendor_op(gdbarch*, unsigned char, dwarf2_frame_state*)>,
address_class_name_to_type_flags = 0x0,
register_reggroup_p = 0x80351864 <default_register_reggroup_p(gdbarch*, int, reggroup const*)>, fetch_pointer_argument = 0x0,
iterate_over_regset_sections = 0x0, make_corefile_notes = 0x0,
find_memory_regions = 0x0, create_memtag_section = 0x0,
fill_memtag_section = 0x0, decode_memtag_section = 0x0,
core_xfer_shared_libraries = 0x0, core_xfer_shared_libraries_aix = 0x0,
core_pid_to_str = 0x0, core_thread_name = 0x0, core_xfer_siginfo = 0x0,
gcore_bfd_target = 0x0, vtable_function_descriptors = 0, vbit_in_delta = 0,
skip_permanent_breakpoint = 0x80081932 <default_skip_permanent_breakpoint(regcache*)>, max_insn_length = 0, displaced_step_copy_insn = 0x0,
displaced_step_hw_singlestep = 0x80083572 <default_displaced_step_hw_singlestep(gdbarch*)>, displaced_step_fixup = 0x0, displaced_step_prepare = 0x0,
displaced_step_finish = 0x0, displaced_step_copy_insn_closure_by_addr = 0x0,
displaced_step_restore_all_in_ptid = 0x0, relocate_instruction = 0x0,
overlay_update = 0x0, core_read_description = 0x0,
sofun_address_maybe_missing = 0, process_record = 0x0,
process_record_signal = 0x0, gdb_signal_from_target = 0x0,
gdb_signal_to_target = 0x0, get_siginfo_type = 0x0,
record_special_symbol = 0x0, get_syscall_number = 0x0,
xml_syscall_file = 0x0, syscalls_info = 0x0, stap_integer_prefixes = 0x0,
stap_integer_suffixes = 0x0, stap_register_prefixes = 0x0,
stap_register_suffixes = 0x0, stap_register_indirection_prefixes = 0x0,
stap_register_indirection_suffixes = 0x0, stap_gdb_register_prefix = 0x0,
stap_gdb_register_suffix = 0x0, stap_is_single_operand = 0x0,
stap_parse_special_token = 0x0, stap_adjust_register = 0x0,
dtrace_parse_probe_argument = 0x0, dtrace_probe_is_enabled = 0x0,
dtrace_enable_probe = 0x0, dtrace_disable_probe = 0x0,
has_global_solist = 0, has_global_breakpoints = 0,
has_shared_address_space = 0x8007b4e8 <default_has_shared_address_space(gdbarch*)>,
fast_tracepoint_valid_at = 0x8007b4ec <default_fast_tracepoint_valid_at(gdbarch*, unsigned long long, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >*)>,
guess_tracepoint_registers = 0x8007eb92 <default_guess_tracepoint_registers(gdbarch*, regcache*, unsigned long long)>,
auto_charset = 0x800f3e5a <default_auto_charset()>,
auto_wide_charset = 0x800f3e74 <default_auto_wide_charset()>,
solib_symbols_extension = 0x0, has_dos_based_file_system = 0,
gen_return_address = 0x8007b534 <default_gen_return_address(gdbarch*, agent_expr*, axs_value*, unsigned long long)>, info_proc = 0x0, core_info_proc = 0x0,
iterate_over_objfiles_in_search_order = 0x802d77dc <default_iterate_over_objfiles_in_search_order(gdbarch*, gdb::function_view<bool (objfile*)>, objfile*)>,
ravenscar_ops = 0x0,
insn_is_call = 0x8007b760 <default_insn_is_call(gdbarch*, unsigned long long)>,
insn_is_ret = 0x8007b764 <default_insn_is_ret(gdbarch*, unsigned long long)>,
insn_is_jump = 0x8007b768 <default_insn_is_jump(gdbarch*, unsigned long long)>,
program_breakpoint_here_p = 0x800819a2 <default_program_breakpoint_here_p(gdbarch*, unsigned long long)>, auxv_parse = 0x0,
print_auxv_entry = 0x80096fda <default_print_auxv_entry(gdbarch*, ui_file*, unsigned long long, unsigned long long)>,
vsyscall_range = 0x8007b76c <default_vsyscall_range(gdbarch*, mem_range*)>,
infcall_mmap = 0x8007b554 <default_infcall_mmap(unsigned long long, unsigned int)>,
infcall_munmap = 0x8007b500 <default_infcall_munmap(unsigned long long, unsigned long long)>,
gcc_target_options = 0x8007dbbe <default_gcc_target_options[abi:cxx11](gdbarch*)>, gnu_triplet_regexp = 0x8007ce2e <default_gnu_triplet_regexp(gdbarch*)>,
addressable_memory_unit_size = 0x8007ce42 <default_addressable_memory_unit_size(gdbarch*)>, disassembler_options_implicit = 0x0, disassembler_options = 0x0,
valid_disassembler_options = 0x0,
type_align = 0x8007b506 <default_type_align(gdbarch*, type*)>,
get_pc_address_flags = 0x8007b776 <default_get_pc_address_flags[abi:cxx11](frame_info_ptr, unsigned long long)>,
read_core_file_mappings = 0x8007b50c <default_read_core_file_mappings(gdbarch*, bfd*, gdb::function_view<void (unsigned long long)>, gdb::function_view<void (int, unsigned long long, unsigned long long, unsigned long long, char const*, bfd_build_id const*)>)>}
(gdb)
(gdb)


But it should get initialized in m68k_linux_init_abi()...

https://sources.debian.org/src/gdb/13.1-2/gdb/m68k-linux-tdep.c/?hl=385#L385


set_gdbarch_skip_trampoline_code (gdbarch, find_solib_trampoline_target);

/* Core file support. */
set_gdbarch_iterate_over_regset_sections
(gdbarch, m68k_linux_iterate_over_regset_sections);


Yet somehow iterate_over_regset_sections ends up NULL.

I could not see why that happened so I built the current upstream snapshot
(at the time), gdb-14.0.50.20230316. That took a few attempts and a couple
of days of building, but it has the same problem:

# /opt/gdb-14.0.50.20230316/bin/gdb
GNU gdb (GDB) 14.0.50.20230316-git
Copyright (C) 2023 Free Software Foundation, Inc.
...
(gdb) core ~/core.0
"/root/core.0": Core file format not supported
(gdb)

I guess the next step will be a formal bug report.
Finn Thain
2023-03-27 01:10:01 UTC
Permalink
I wasn't able to add the linux-m68k list to the bug report Cc list.

I'm forwarding this as I don't know the answer to Tom's question...
https://sourceware.org/bugzilla/show_bug.cgi?id=30273
What |Removed |Added
----------------------------------------------------------------------------
CC| |tromey at sourceware dot org
--- Comment #1 from Tom Tromey <tromey at sourceware dot org> ---
commit 3eba3a011a89c75c10bd1860eee4589e65697165
Date: Mon Sep 14 08:30:10 2020 -0600
Various m68k fixes for gdb
When I debug gdb I see it decides this core file is
GDB_OSABI_SVR4, not GDB_OSABI_LINUX. This happens
if (elfosabi == ELFOSABI_NONE)
return GDB_OSABI_SVR4;
#define ELFOSABI_NONE 0 /* UNIX System V ABI */
#define ELFOSABI_HPUX 1 /* HP-UX operating system */
#define ELFOSABI_NETBSD 2 /* NetBSD */
#define ELFOSABI_GNU 3 /* GNU */
#define ELFOSABI_LINUX 3 /* Alias for ELFOSABI_GNU */
So I guess one question is why this core file isn't marked
as ELFOSABI_LINUX.
Geert Uytterhoeven
2023-03-27 07:00:01 UTC
Permalink
Hi Finn,
Post by Finn Thain
I wasn't able to add the linux-m68k list to the bug report Cc list.
I'm forwarding this as I don't know the answer to Tom's question...
https://sourceware.org/bugzilla/show_bug.cgi?id=30273
What |Removed |Added
----------------------------------------------------------------------------
CC| |tromey at sourceware dot org
--- Comment #1 from Tom Tromey <tromey at sourceware dot org> ---
commit 3eba3a011a89c75c10bd1860eee4589e65697165
Date: Mon Sep 14 08:30:10 2020 -0600
Various m68k fixes for gdb
When I debug gdb I see it decides this core file is
GDB_OSABI_SVR4, not GDB_OSABI_LINUX. This happens
if (elfosabi == ELFOSABI_NONE)
return GDB_OSABI_SVR4;
#define ELFOSABI_NONE 0 /* UNIX System V ABI */
#define ELFOSABI_HPUX 1 /* HP-UX operating system */
#define ELFOSABI_NETBSD 2 /* NetBSD */
#define ELFOSABI_GNU 3 /* GNU */
#define ELFOSABI_LINUX 3 /* Alias for ELFOSABI_GNU */
So I guess one question is why this core file isn't marked
as ELFOSABI_LINUX.
Only PA-RISC and the remoteproc core dumper seem to use
ELFOSABI_LINUX?

arch/parisc/include/asm/elf.h:#define ELF_OSABI ELFOSABI_LINUX
arch/parisc/kernel/ptrace.c: .name = "parisc", .e_machine =
ELF_ARCH, .ei_osabi = ELFOSABI_LINUX,
arch/parisc/kernel/ptrace.c: .name = "parisc", .e_machine =
EM_PARISC, .ei_osabi = ELFOSABI_LINUX,

include/uapi/linux/elf.h:#ifndef ELF_OSABI
include/uapi/linux/elf.h:#define ELF_OSABI ELFOSABI_NONE
include/uapi/linux/elf.h-#endif

fs/proc/kcore.c: [EI_OSABI] = ELF_OSABI,

drivers/remoteproc/remoteproc_elf_helpers.h:
hdr->e_ident[EI_OSABI] = ELFOSABI_NONE;

Or perhaps it's set originally somewhere else, but overwritten on
m68k (and a few others)?

arch/arm/kernel/ptrace.c: .name = "arm", .e_machine =
ELF_ARCH, .ei_osabi = ELF_OSABI,
arch/hexagon/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/m68k/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/mips/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/mips/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/mips/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/nios2/kernel/ptrace.c: .ei_osabi = ELF_OSABI,

Does this work on other architectures?

Gr{oetje,eeting}s,

Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ***@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
Finn Thain
2023-03-27 08:50:02 UTC
Permalink
Post by Geert Uytterhoeven
Post by Finn Thain
I wasn't able to add the linux-m68k list to the bug report Cc list.
I'm forwarding this as I don't know the answer to Tom's question...
https://sourceware.org/bugzilla/show_bug.cgi?id=30273
What |Removed |Added
----------------------------------------------------------------------------
CC| |tromey at sourceware dot org
--- Comment #1 from Tom Tromey <tromey at sourceware dot org> ---
commit 3eba3a011a89c75c10bd1860eee4589e65697165
Date: Mon Sep 14 08:30:10 2020 -0600
Various m68k fixes for gdb
When I debug gdb I see it decides this core file is
GDB_OSABI_SVR4, not GDB_OSABI_LINUX. This happens
if (elfosabi == ELFOSABI_NONE)
return GDB_OSABI_SVR4;
#define ELFOSABI_NONE 0 /* UNIX System V ABI */
#define ELFOSABI_HPUX 1 /* HP-UX operating system */
#define ELFOSABI_NETBSD 2 /* NetBSD */
#define ELFOSABI_GNU 3 /* GNU */
#define ELFOSABI_LINUX 3 /* Alias for ELFOSABI_GNU */
So I guess one question is why this core file isn't marked
as ELFOSABI_LINUX.
Only PA-RISC and the remoteproc core dumper seem to use
ELFOSABI_LINUX?
arch/parisc/include/asm/elf.h:#define ELF_OSABI ELFOSABI_LINUX
arch/parisc/kernel/ptrace.c: .name = "parisc", .e_machine =
ELF_ARCH, .ei_osabi = ELFOSABI_LINUX,
arch/parisc/kernel/ptrace.c: .name = "parisc", .e_machine =
EM_PARISC, .ei_osabi = ELFOSABI_LINUX,
include/uapi/linux/elf.h:#ifndef ELF_OSABI
include/uapi/linux/elf.h:#define ELF_OSABI ELFOSABI_NONE
include/uapi/linux/elf.h-#endif
fs/proc/kcore.c: [EI_OSABI] = ELF_OSABI,
hdr->e_ident[EI_OSABI] = ELFOSABI_NONE;
Or perhaps it's set originally somewhere else, but overwritten on
m68k (and a few others)?
arch/arm/kernel/ptrace.c: .name = "arm", .e_machine =
ELF_ARCH, .ei_osabi = ELF_OSABI,
arch/hexagon/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/m68k/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/mips/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/mips/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/mips/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
arch/nios2/kernel/ptrace.c: .ei_osabi = ELF_OSABI,
Does this work on other architectures?
I'm not sure that it matters. If Tom is right that the regression was
caused by his commit 3eba3a011a89 then the issue would be confined to m68k
and it's ABIs, and how to tell them apart given ELFOSABI_NONE:
https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=3eba3a011a89

BTW, I suspect that the best place to discuss this would be bugzilla
otherwise we'll need to Cc Tom.

Loading...