This patch set for the generic BFD a.out backend removes a dead #define and makes aoutx.h self-contained:  [PATCH 1/2]: bfd: remove unused NO_WRITE_HEADER_KLUDGE #define  [PATCH 2/2]: bfd: make aoutx.h self-contained

* aout64.c (BMAGIC, QMAGIC): Do not define.
* aoutx.h (N_IS_BMAGIC, N_SET_QMAGIC): New defines.
(NAME (aout, some_aout_object_p)): Use N_IS_QMAGIC and N_IS_BMAGIC
to check the file format.
(adjust_z_magic): Use N_SET_QMAGIC to set file format.
* i386aout.c (NO_WRITE_HEADER_KLUDGE): Delete define.
* libaout.h (NO_WRITE_HEADER_KLUDGE): Do not define.

ELF: Consolidate maybe_set_textrel

All maybe_set_textrel implementations are the same.  Consolidate them
to a single _bfd_elf_maybe_set_textrel.

* elf-bfd.h (_bfd_elf_maybe_set_textrel): New
* elf32-arm.c (maybe_set_textrel): Removed.
(elf32_arm_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-csky.c (maybe_set_textrel): Removed.
(csky_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-hppa.c (maybe_set_textrel): Removed.
(elf32_hppa_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-lm32.c (maybe_set_textrel): Removed.
(lm32_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-m32r.c (maybe_set_textrel): Removed.
(m32r_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-metag.c (maybe_set_textrel): Removed.
(elf_metag_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-nds32.c (maybe_set_textrel): Removed.
(nds32_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-or1k.c (maybe_set_textrel): Removed.
(or1k_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-ppc.c (maybe_set_textrel): Removed.
(ppc_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-s390.c (maybe_set_textrel): Removed.
(elf_s390_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-sh.c (maybe_set_textrel): Removed.
(sh_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-tic6x.c (maybe_set_textrel): Removed.
(elf32_tic6x_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf32-tilepro.c (maybe_set_textrel): Removed.
(tilepro_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf64-ppc.c (maybe_set_textrel): Removed.
(ppc64_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elf64-s390.c (maybe_set_textrel): Removed.
(elf_s390_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elfnn-aarch64.c (maybe_set_textrel): Removed.
(elfNN_aarch64_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elfnn-riscv.c (maybe_set_textrel): Removed.
(riscv_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elfxx-sparc.c (maybe_set_textrel): Removed.
(_bfd_sparc_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elfxx-tilegx.c (maybe_set_textrel): Removed.
(tilegx_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elfxx-x86.c (maybe_set_textrel): Removed.
(_bfd_x86_elf_size_dynamic_sections): Replace maybe_set_textrel
with _bfd_elf_maybe_set_textrel.
* elflink.c (_bfd_elf_maybe_set_textrel): New.

ELF: Copy dyn_relocs in _bfd_elf_link_hash_copy_indirect

Copy dyn_relocs in _bfd_elf_link_hash_copy_indirect instead of in each
target backend.

PR ld/26067
* elf32-arm.c (elf32_arm_copy_indirect_symbol): Don't copy
dyn_relocs.
* elf32-csky.c (csky_elf_copy_indirect_symbol): Likewise.
* elf32-hppa.c (elf32_hppa_copy_indirect_symbol): Likewise.
* elf32-metag.c (elf_metag_copy_indirect_symbol): Likewise.
* elf32-microblaze.c (microblaze_elf_copy_indirect_symbol):
Likewise.
* elf32-nds32.c (nds32_elf_copy_indirect_symbol): Likewise.
* elf32-nios2.c (nios2_elf32_copy_indirect_symbol): Likewise.
* elf32-or1k.c (or1k_elf_copy_indirect_symbol): Likewise.
* elf32-s390.c (elf_s390_copy_indirect_symbol): Likewise.
* elf32-sh.c (sh_elf_copy_indirect_symbol): Likewise.
* elf32-tilepro.c (tilepro_elf_copy_indirect_symbol): Likewise.
* elf64-s390.c (elf_s390_copy_indirect_symbol): Likewise.
* elfnn-aarch64.c (elfNN_aarch64_copy_indirect_symbol): Likewise.
* elfnn-riscv.c (riscv_elf_copy_indirect_symbol): Likewise.
* elfxx-sparc.c (_bfd_sparc_elf_copy_indirect_symbol): Likewise.
* elfxx-tilegx.c (tilegx_elf_copy_indirect_symbol): Likewise.
* elfxx-x86.c (_bfd_x86_elf_copy_indirect_symbol): Likewise.
* elf32-lm32.c (lm32_elf_copy_indirect_symbol): Removed.
(elf_backend_copy_indirect_symbol): Likewise.
* elf32-m32r.c (m32r_elf_copy_indirect_symbol): Removed.
(elf_backend_copy_indirect_symbol): Likewise.
* elflink.c (_bfd_elf_link_hash_copy_indirect): Copy dyn_relocs.

ELF: Consolidate readonly_dynrelocs

All readonly_dynrelocs implementations are the same.  Consolidate them
to a single _bfd_elf_readonly_dynrelocs.

PR ld/26067
* elf-bfd.h (_bfd_elf_readonly_dynrelocs): New.
* elf32-arm.c (readonly_dynrelocs): Removed.
(maybe_set_textrel): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
* elf32-csky.c (readonly_dynrelocs): Removed.
(maybe_set_textrel): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
*  elf32-hppa.c(readonly_dynrelocs): Removed.
(alias_readonly_dynrelocs): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elf32-lm32.c (readonly_dynrelocs): Removed.
(lm32_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elf32-m32r.c (readonly_dynrelocs): Removed.
(m32r_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elf32-metag.c (readonly_dynrelocs): Removed.
(elf_metag_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elf32-microblaze.c (readonly_dynrelocs): Removed.
(microblaze_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
* elf32-nds32.c (readonly_dynrelocs): Removed.
(nds32_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elf32-or1k.c (readonly_dynrelocs): Removed.
(or1k_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
* elf32-ppc.c (readonly_dynrelocs): Removed.
(alias_readonly_dynrelocs): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
(ppc_elf_adjust_dynamic_symbol): Likewise.
(maybe_set_textrel): Likewise.
* elf32-s390.c (readonly_dynrelocs): Removed.
(elf_s390_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elf32-sh.c (readonly_dynrelocs): Removed.
(sh_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elf32-tic6x.c (readonly_dynrelocs): Removed.
(maybe_set_textrel): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
* elf32-tilepro.c (readonly_dynrelocs): Removed.
(tilepro_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elf64-ppc.c (readonly_dynrelocs): Removed.
(alias_readonly_dynrelocs): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
(ppc64_elf_adjust_dynamic_symbol): Likewise.
(maybe_set_textrel): Likewise.
* elf64-s390.c (readonly_dynrelocs): Removed.
(elf_s390_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elflink.c (_bfd_elf_readonly_dynrelocs): New.
* elfnn-aarch64.c (readonly_dynrelocs): Removed.
(maybe_set_textrel): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
* elfnn-riscv.c (readonly_dynrelocs): Removed.
(riscv_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elfxx-sparc.c (readonly_dynrelocs): Removed.
(_bfd_sparc_elf_adjust_dynamic_symbol): Replace
readonly_dynrelocs with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elfxx-tilegx.c (readonly_dynrelocs): Removed.
(tilegx_elf_adjust_dynamic_symbol): Replace readonly_dynrelocs
with _bfd_elf_readonly_dynrelocs.
(maybe_set_textrel): Likewise.
* elfxx-x86.c (readonly_dynrelocs): Removed.
(maybe_set_textrel): Replace readonly_dynrelocs with
_bfd_elf_readonly_dynrelocs.
(_bfd_x86_elf_adjust_dynamic_symbol): Likewise.

ld: Pass -fno-sanitize=all to tests with linker

When binutils is compiled with -fsanitize=undefined, many tests with
linker, instead of $CC, fail with undefined symbol references to sanitize
library.  Define NOSANTIZE_CFLAGS to -fno-sanitize=all if target compiler
supports it and compile such tests with $NOSANTIZE_CFLAGS.

* testsuite/config/default.exp (NOSANTIZE_CFLAGS): New.
* testsuite/ld-elf/linux-x86.exp: Add $NOSANTIZE_CFLAGS to
tests with run_ld_link_exec_tests.
* testsuite/ld-elf/shared.exp: Add $NOSANTIZE_CFLAGS to tests
with run_ld_link_tests.
* testsuite/ld-elf/tls.exp: Likewise.
* testsuite/ld-elfweak/elfweak.exp: Add $NOSANTIZE_CFLAGS to
tests with ld_link.
* testsuite/ld-gc/gc.exp: Add $NOSANTIZE_CFLAGS to cflags.
* testsuite/ld-plugin/lto.exp: Add $NOSANTIZE_CFLAGS to tests
with run_ld_link_tests.a
* testsuite/ld-plugin/plugin.exp: Append $NOSANTIZE_CFLAGS to
CFLAGS.
* testsuite/ld-selective/selective.exp: Add $NOSANTIZE_CFLAGS
to cflags and cxxflags.
* testsuite/ld-srec/srec.exp: Append $NOSANTIZE_CFLAGS to CC
and CXX.
* testsuite/ld-x86-64/plt-main-ibt-x32.dd: Updated for
-fsanitize=undefined.
* testsuite/ld-x86-64/plt-main-ibt.dd: Likewise.
* testsuite/ld-x86-64/x86-64.exp: Add $NOSANTIZE_CFLAGS to
tests with run_cc_link_tests and run_ld_link_tests.

ld: Add --export-dynamic-symbol and --export-dynamic-symbol-list

--export-dynamic-symbol-list is like a dynamic list, but without
the symbolic property for unspecified symbols.

When creating an executable, --export-dynamic-symbol-list is treated
like --dynamic-list.

When creating a shared library, it is treated like --dynamic-list if
-Bsymbolic or --dynamic-list are used,  otherwise, it is ignored, so
that references to matched symbols will not be bound to the definitions
within the shared library.

PR ld/25910
* NEWS: Mention --export-dynamic-symbol[-list].
* ld.texi: Document --export-dynamic-symbol[-list].
* ldgram.y: Pass current_dynamic_list_p to
lang_append_dynamic_list.
* ldlang.c (current_dynamic_list_p): New.
(ang_append_dynamic_list): Updated to take a pointer to
struct bfd_elf_dynamic_list * argument instead of using
link_info.dynamic_list.
(lang_append_dynamic_list_cpp_typeinfo): Pass
&link_info.dynamic_list to ang_append_dynamic_list.
(lang_append_dynamic_list_cpp_new): Likewise.
* ldlang.h (current_dynamic_list_p): New.
(lang_append_dynamic_list): Add a pointer to
struct bfd_elf_dynamic_list * argument.
* ldlex.h (option_values): Add OPTION_EXPORT_DYNAMIC_SYMBOL and
OPTION_EXPORT_DYNAMIC_SYMBOL_LIST.
* lexsup.c (ld_options): Add entries for
OPTION_EXPORT_DYNAMIC_SYMBOL and
OPTION_EXPORT_DYNAMIC_SYMBOL_LIST.
(parse_args): Handle --export-dynamic-symbol and
--export-dynamic-symbol-list.
* testsuite/ld-dynamic/export-dynamic-symbol-1.d: New.
* testsuite/ld-dynamic/export-dynamic-symbol-2.d: New.
* testsuite/ld-dynamic/export-dynamic-symbol-glob.d: New.
* testsuite/ld-dynamic/export-dynamic-symbol-list-1.d: New.
* testsuite/ld-dynamic/export-dynamic-symbol-list-2.d: New.
* testsuite/ld-dynamic/export-dynamic-symbol-list-glob.d: New.
* testsuite/ld-dynamic/export-dynamic-symbol.exp: New.
* testsuite/ld-dynamic/export-dynamic-symbol.s: New.
* testsuite/ld-dynamic/foo-bar.list: New.
* testsuite/ld-dynamic/foo.list: New.
* testsuite/ld-dynamic/foo.s: New.
* testsuite/ld-dynamic/fstar.list: New.
* testsuite/ld-elf/dlempty.list: New.
* testsuite/ld-elf/shared.exp: Add tests for
--export-dynamic-symbol and --export-dynamic-symbol-list.

x86: Silence -fsanitize=undefined

Replace "&(EH)->elf" with "(struct elf_link_hash_entry *) (EH)" to
silence -fsanitize=undefined.

* elfxx-x86.h (GENERATE_DYNAMIC_RELOCATION_P): Replace
"&(EH)->elf" with "(struct elf_link_hash_entry *) (EH)".

ld: fix ld-elf/linux-x86.exp for r/o source tree

Since the copying ofthe individual files gets done more than once, make
sure the destinations can be copied to on subsequent copying attempts.

PR26069, strip/objcopy misaligned address accesses

PR 26069
PR 18758
* peicode.h (pe_ILF_make_a_section): Align data for compilers
other than gcc.
(pe_ILF_build_a_bfd): Likewise.

PR26069, strip/objcopy memory leaks

PR 26069
* objcopy.c (copy_relocations_in_section): Free relpp on error.
Don't accidentally free isection->orelocation.

PR26069, strip/objcopy memory leaks

PR 26029
* elf.c (_bfd_elf_close_and_cleanup): Free elf_shstrtab for
core files as well as objects.

RISC-V: Fix minor bugs in .insn docs.

This fixes some minor bugs in the docs for the .insn directive pointed out
by Frédéric Pétrot, and I added a few more cleanups since I was changing
the docs.

gas/
PR 26051
* doc/c-riscv.texi (RISC-V-Formats): Add missing I format using
simm12(rs1).  Correct S format to use simm12(rs1).  Drop SB and B
formats using simm12(rs1).  Correct SB and B to use rs1 and rs2.
Move B before SB.  Move J before UJ.

RISC-V: Fix the error when building RISC-V linux native gdbserver.

The original report is as follow,
https://sourceware.org/pipermail/binutils/2020-June/111383.html

Inlcude the bfd.h in the include/opcode/riscv.h may cause gdbserver fail
to build.  I just want to use the `bfd_boolean` in the opcodes/riscv-opc.c,
but I didn't realize this cause the build failed.  Fortunately, I can also
use the `int` as the function return types just like others in the
opcodes/riscv-opc.c.

include/
* opcode/riscv.h: Remove #include "bfd.h".  And change the return
types of riscv_get_isa_spec_class and riscv_get_priv_spec_class
from bfd_boolean to int.

opcodes/
* riscv-opc.c (riscv_get_isa_spec_class): Change bfd_boolean to int.
(riscv_get_priv_spec_class): Likewise.

Automatic date update in version.in

gdb: Convert language skip_trampoline field to a method

This commit changes the language_data::skip_trampoline function
pointer member variable into a member function of language_defn.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_language_data): Delete skip_trampoline
initializer.
* c-lang.c (c_language_data): Likewise.
(cplus_language_data): Likewise.
(cplus_language::skip_trampoline): New member function.
(asm_language_data): Delete skip_trampoline initializer.
(minimal_language_data): Likewise.
* d-lang.c (d_language_data): Likewise.
* f-lang.c (f_language_data): Likewise.
* go-lang.c (go_language_data): Likewise.
* language.c (unk_lang_trampoline): Delete function.
(skip_language_trampoline): Update.
(unknown_language_data): Delete skip_trampoline initializer.
(auto_language_data): Likewise.
* language.h (language_data): Delete skip_trampoline field.
(language_defn::skip_trampoline): New function.
* m2-lang.c (m2_language_data): Delete skip_trampoline
initializer.
* objc-lang.c (objc_skip_trampoline): Delete function, move
implementation to objc_language::skip_trampoline.
(objc_language_data): Delete skip_trampoline initializer.
(objc_language::skip_trampoline): New member function with
implementation from objc_skip_trampoline.
* opencl-lang.c (opencl_language_data): Delete skip_trampoline
initializer.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.

gdb: Convert language la_demangle field to a method

This commit changes the language_data::la_demangle function pointer
member variable into a member function of language_defn.

The only slightly "weird" change in this commit is in f-lang.c, where
I have given the Fortran language a demangle method that is identical
to the default language_defn::demangle.  The only reason for this is
to give me somewhere to copy the comment that was previously embedded
within the f_language_data structure.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_language_data): Delete la_demangle initializer.
(ada_language::demangle): New member function.
* c-lang.c (c_language_data): Delete la_demangle initializer.
(cplus_language_data): Delete la_demangle initializer.
(cplus_language::demangle): New member function.
(asm_language_data): Delete la_demangle initializer.
(minimal_language_data): Delete la_demangle initializer.
* d-lang.c (d_language_data): Delete la_demangle initializer.
(d_language::demangle): New member function.
* f-lang.c (f_language_data): Delete la_demangle initializer.
(f_language::demangle): New member function.
* go-lang.c (go_language_data): Delete la_demangle initializer.
(go_language::demangle): New member function.
* language.c (language_demangle): Update.
(unk_lang_demangle): Delete.
(unknown_language_data): Delete la_demangle initializer.
(unknown_language::demangle): New member function.
(auto_language_data): Delete la_demangle initializer.
(auto_language::demangle): New member function.
* language.h (language_data): Delete la_demangle field.
(language_defn::demangle): New function.
* m2-lang.c (m2_language_data): Delete la_demangle initializer.
* objc-lang.c (objc_language_data): Delete la_demangle
initializer.
(objc_language::demangle): New member function.
* opencl-lang.c (opencl_language_data): Delete la_demangle
initializer.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.
(rust_language::demangle): New member functi

gdb: Convert language la_print_type field to a method

This commit changes the language_data::la_print_type function pointer
member variable into a member function of language_defn.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_language_data): Delete la_print_type
initializer.
(ada_language::print_type): New member function.
* c-lang.c (c_language_data): Delete la_print_type initializer.
(c_language::print_type): New member function.
(cplus_language_data): Delete la_print_type initializer.
(cplus_language::print_type): New member function.
(asm_language_data): Delete la_print_type initializer.
(asm_language::print_type): New member function.
(minimal_language_data): Delete la_print_type initializer.
(minimal_language::print_type): New member function.
* d-lang.c (d_language_data): Delete la_print_type initializer.
(d_language::print_type): New member function.
* f-lang.c (f_language_data): Delete la_print_type initializer.
(f_language::print_type): New member function.
* go-lang.c (go_language_data): Delete la_print_type initializer.
(go_language::print_type): New member function.
* language.c (unk_lang_print_type): Delete.
(unknown_language_data): Delete la_print_type initializer.
(unknown_language::print_type): New member function.
(auto_language_data): Delete la_print_type initializer.
(auto_language::print_type): New member function.
* language.h (language_data): Delete la_print_type field.
(language_defn::print_type): New function.
(LA_PRINT_TYPE): Update.
* m2-lang.c (m2_language_data): Delete la_print_type initializer.
(m2_language::print_type): New member function.
* objc-lang.c (objc_language_data): Delete la_print_type
initializer.
(objc_language::print_type): New member function.
* opencl-lang.c (opencl_print_type): Delete, implementation moved
to opencl_language::print_type.
(opencl_language_data): Delete la_print_type initializer.
(opencl_language::print_type): New member function, implementation
from opencl_print_type.
* p-lang.c (pascal_language_data): Delete la_print_type
initializer.
(pascal_language::print_type): New member function.
* rust-lang.c (rust_print_type): Delete, implementation moved to
rust_language::print_type.
(rust_language_data): Delete la_print_type initializer.
(rust_language::print_type): New member function, implementation
from rust_print_type.

gdb: Convert language la_sniff_from_mangled_name field to a method

This commit changes the language_data::la_sniff_from_mangled_name
function pointer member variable into a member function of
language_defn.

Previously the la_sniff_from_mangled_name pointer was NULL for some
languages, however, all uses of this function pointer were through the
function language_sniff_from_mangled_name which provided a default
implementation.

This default implementation now becomes the implementation in the base
class language_defn, which is then overridden as required in various
language sub-classes.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_sniff_from_mangled_name): Delete function,
implementation moves to...
(ada_language::sniff_from_mangled_name): ...here.  Update return
type.
(ada_language_data): Delete la_sniff_from_mangled_name
initializer.
* c-lang.c (c_language_data): Likewise.
(cplus_language_data): Likewise.
(cplus_language::sniff_from_mangled_name): New member function,
implementation taken from gdb_sniff_from_mangled_name.
(asm_language_data): Delete la_sniff_from_mangled_name
initializer.
(minimal_language_data): Likewise.
* cp-support.c (gdb_sniff_from_mangled_name): Delete,
implementation moves to cplus_language::sniff_from_mangled_name.
* cp-support.h (gdb_sniff_from_mangled_name): Delete declaration.
* d-lang.c (d_sniff_from_mangled_name): Delete, implementation
moves to...
(d_language::sniff_from_mangled_name): ...here.
(d_language_data): Delete la_sniff_from_mangled_name initializer.
* f-lang.c (f_language_data): Likewise.
* go-lang.c (go_sniff_from_mangled_name): Delete, implementation
moves to...
(go_language::sniff_from_mangled_name): ...here.
(go_language_data): Delete la_sniff_from_mangled_name initializer.
* language.c (language_sniff_from_mangled_name): Delete.
(unknown_language_data): Delete la_sniff_from_mangled_name
initializer.
(auto_language_data): Likewise.
* language.h (language_data): Delete la_sniff_from_mangled_name
field.
(language_defn::sniff_from_mangled_name): New function.
(language_sniff_from_mangled_name): Delete declaration.
* m2-lang.c (m2_language_data): Delete la_sniff_from_mangled_name
field.
* objc-lang.c (objc_sniff_from_mangled_name): Delete,
implementation moves to...
(objc_language::sniff_from_mangled_name): ...here.
(objc_language_data): Delete la_sniff_from_mangled_name initializer.
* opencl-lang.c (opencl_language_data): Likewise.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_sniff_from_mangled_name): Delete,
implementation moves to...
(rust_language::sniff_from_mangled_name): ...here.
(rust_language_data): Delete la_sniff_from_mangled_name
initializer.
* symtab.c (symbol_find_demangled_name): Call
sniff_from_mangled_name member function.

gdb: Convert language la_search_name_hash field to a method

This commit changes the language_data::la_search_name_hash
function pointer member variable into a member function of
language_defn.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_language_data): Delete la_search_name_hash
initializer.
* c-lang.c (c_language_data): Likewise.
(cplus_language_data): Likewise.
(cplus_language::search_name_hash): New member function.
(asm_language_data): Delete la_search_name_hash initializer.
(minimal_language_data): Likewise.
* d-lang.c (d_language_data): Likewise.
* dictionary.c (default_search_name_hash): Rename to...
(language_defn::search_name_hash): ...this.
* f-lang.c (f_language_data): Likewise.
(f_language::search_name_hash): New member function.
* go-lang.c (go_language_data): Delete la_search_name_hash
initializer.
* language.c (unknown_language_data): Likewise.
(auto_language_data): Likewise.
* language.h (struct language_data): Delete la_search_name_hash
field.
(language_defn::search_name_hash): Declare new member function.
(default_search_name_hash): Delete declaration.
* m2-lang.c (m2_language_data): Delete la_search_name_hash
initializer.
* objc-lang.c (objc_language_data): Likewise.
* opencl-lang.c (opencl_language_data): Likewise.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.
* symtab.c (search_name_hash): Update call.

gdb: Convert language la_get_compile_instance field to a method

This commit changes the language_data::la_get_compile_instance
function pointer member variable into a member function of
language_defn.  Unlike previous commits converting fields of
language_data to member function in language_defn, this field is NULL
for some languages.  As a result I had to change the API slightly so
that the base language_defn class provides an implementation.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_language_data): Delete la_get_compile_instance
initializer.
* c-lang.c (class compile_instance): Declare.
(c_language_data): Delete la_get_compile_instance initializer.
(c_language::get_compile_instance): New member function.
(cplus_language_data): Delete la_get_compile_instance initializer.
(cplus_language::get_compile_instance): New member function.
(asm_language_data): Delete la_get_compile_instance initializer.
(minimal_language_data): Likewise.
* c-lang.h (c_get_compile_context): Update comment.
(cplus_get_compile_context): Update comment.
* compile/compile.c (compile_to_object): Update calls, don't rely
on function pointer being NULL.
* d-lang.c (d_language_data): Delete la_get_compile_instance
initializer.
* f-lang.c (f_language_data): Likewise.
* go-lang.c (go_language_data): Likewise.
* language.c (unknown_language_data): Likewise.
(auto_language_data): Likewise.
* language.h (language_data): Delete la_get_compile_instance field.
(language_defn::get_compile_instance): New member function.
* m2-lang.c (m2_language_data): Delete la_get_compile_instance
initializer.
* objc-lang.c (objc_language_data): Likewise.
* opencl-lang.c (opencl_language_data): Likewise.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.

gdb: Convert language la_iterate_over_symbols field to a method

This commit changes the language_data::la_iterate_over_symbols
function pointer member variable into a member function of
language_defn.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_add_all_symbols): Update comment.
(ada_iterate_over_symbols): Delete, move implementation to...
(ada_language::iterate_over_symbols): ...here, a new member
function, rewrite to use range based for loop.
(ada_language_data): Delete la_iterate_over_symbols initializer.
* c-lang.c (c_language_data): Likewise.
(cplus_language_data): Likewise.
(asm_language_data): Likewise.
(minimal_language_data): Likewise.
* d-lang.c (d_language_data): Likewise.
* f-lang.c (f_language_data): Likewise.
* go-lang.c (go_language_data): Likewise.
* language.c (unknown_language_data): Likewise.
(auto_language_data): Likewise.
* language.h (language_data): Delete la_iterate_over_symbols field.
(language_defn::iterate_over_symbols): New member function.
(LA_ITERATE_OVER_SYMBOLS): Update.
* linespec.c (iterate_over_all_matching_symtabs): Update.
* m2-lang.c (m2_language_data): Delete la_iterate_over_symbols
initializer.
* objc-lang.c (objc_language_data): Likewise.
* opencl-lang.c (opencl_language_data): Likewise.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.

gdb: Convert language la_lookup_transparent_type field to a method

This commit changes the language_data::la_lookup_transparent_type
function pointer member variable into a member function of
language_defn.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_language_data): Delete
la_lookup_transparent_type initializer.
* c-lang.c (c_language_data): Likewise.
(cplus_language_data): Likewise.
(cplus_language::lookup_transparent_type): New member function.
(asm_language_data): Delete la_lookup_transparent_type
initializer.
(minimal_language_data): Likewise.
* d-lang.c (d_language_data): Likewise.
* f-lang.c (f_language_data): Likewise.
* go-lang.c (go_language_data): Likewise.
* language.c (unknown_language_data): Likewise.
(auto_language_data): Likewise.
* language.h (struct language_data): Delete
la_lookup_transparent_type field.
(language_defn::lookup_transparent_type): New member function.
* m2-lang.c (m2_language_data): Delete la_lookup_transparent_type
initializer.
* objc-lang.c (objc_language_data): Likewise.
* opencl-lang.c (opencl_language_data): Likewise.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.
* symtab.c (symbol_matches_domain): Update call.

gdb: Convert language la_language_arch_info field to a method

This commit changes the language_data::la_language_arch_info function
pointer member variable into a member function of language_defn.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_language_arch_info): Delete function, move
implementation to...
(ada_language::language_arch_info): ...here, a new member
function.
(ada_language_data): Delete la_language_arch_info.
* c-lang.c (c_language_data): Likewise.
(c_language::language_arch_info): New member function.
(cplus_language_arch_info): Delete function, move
implementation to...
(cplus_language::language_arch_info): ...here, a new member
function.
(cplus_language_data): Delete la_language_arch_info.
(asm_language_data): Likewise.
(asm_language::language_arch_info): New member function.
(minimal_language_data): Delete la_language_arch_info.
(minimal_language::language_arch_info): New member function.
* d-lang.c (d_language_arch_info): Delete function, move
implementation to...
(d_language::language_arch_info): ...here, a new member
function.
(d_language_data): Delete la_language_arch_info.
* f-lang.c (f_language_arch_info): Delete function, move
implementation to...
(f_language::language_arch_info): ...here, a new member
function.
(f_language_data): Delete la_language_arch_info.
* go-lang.c (go_language_arch_info): Delete function, move
implementation to...
(go_language::language_arch_info): ...here, a new member
function.
(go_language_data): Delete la_language_arch_info.
* language.c (unknown_language_data): Likewise.
(unknown_language::language_arch_info): New member function.
(auto_language_data): Delete la_language_arch_info.
(auto_language::language_arch_info): New member function.
(language_gdbarch_post_init): Update call to
la_language_arch_info.
* language.h (language_data): Delete la_language_arch_info
function pointer.
(language_defn::language_arch_info): New function.
* m2-lang.c (m2_language_arch_info): Delete function, move
implementation to...
(m2_language::language_arch_info): ...here, a new member
function.
(m2_language_data): Delete la_language_arch_info.
* objc-lang.c (objc_language_arch_info): Delete function, move
implementation to...
(objc_language::language_arch_info): ...here, a new member
function.
(objc_language_data): Delete la_language_arch_info.
* opencl-lang.c (opencl_language_arch_info): Delete function, move
implementation to...
(opencl_language::language_arch_info): ...here, a new member
function.
(opencl_language_data): Delete la_language_arch_info.
* p-lang.c (pascal_language_arch_info): Delete function, move
implementation to...
(pascal_language::language_arch_info): ...here, a new member
function.
(pascal_language_data): Delete la_language_arch_info.
* rust-lang.c (rust_language_arch_info): Delete function, move
implementation to...
(rust_language::language_arch_info): ...here, a new member
function.
(rust_language_data): Delete la_language_arch_info.

gdb: Convert language la_pass_by_reference field to a method

This commit changes the language_data::la_pass_by_reference function
pointer member variable into a member function of language_defn.

The interesting thing in this commit is that I have removed the
default_pass_by_reference function entirely.  This function only ever
returned a language_pass_by_ref_info struct in its default state, so
all uses of this function can be replaced by just default
initialisation of a language_pass_by_ref_info variable.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_language_data): Delete la_pass_by_reference
initializer.
* c-lang.c (c_language_data): Likewise.
(cplus_language_data): Likewise.
(cplus_language::pass_by_reference_info): New method.
(asm_language_data): Delete la_pass_by_reference initializer.
(minimal_language_data): Likewise.
* cp-abi.c (cp_pass_by_reference): Remove use of
default_pass_by_reference.
* d-lang.c (d_language_data): Likewise.
* f-lang.c (f_language_data): Likewise.
* gnu-v3-abi.c (gnuv3_pass_by_reference): Remove use of
default_pass_by_reference.
* go-lang.c (go_language_data): Likewise.
* language.c (language_pass_by_reference): Update.
(default_pass_by_reference): Delete.
(unknown_language_data): Delete la_pass_by_reference
initializer.
(auto_language_data): Likewise.
* language.h (struct language_data): Delete la_pass_by_reference
field.
(language_defn::pass_by_reference_info): New member function.
(default_pass_by_reference): Delete declaration.
* m2-lang.c (m2_language_data): Delete la_pass_by_reference
initializer.
* objc-lang.c (objc_language_data): Likewise.
* opencl-lang.c (opencl_language_data): Likewise.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.

gdb: Convert language la_read_var_value field to a method

This commit changes the language_data::la_read_var_value function
pointer member variable into a member function of language_defn.

An interesting aspect of this change is that the implementation of
language_defn::read_var_value is actually in findvar.c.  This is
partly historical, the new language_defn::read_var_value is a rename
of default_read_var_value, which was already in that file, but also,
that is the file that contains the helper functions needed by the
read_var_value method, so it makes sens that the method implementation
should continue to live there (I think).

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_read_var_value): Delete function, move
implementation to...
(ada_language::read_var_value): ...here.
(ada_language_data): Delete la_read_var_value initializer.
* c-lang.c (c_language_data): Likewise.
(cplus_language_data): Likewise.
(minimal_language_data): Likewise.
* d-lang.c (d_language_data): Likewise.
* f-lang.c (f_language_data): Likewise.
* findvar.c (default_read_var_value): Rename to...
(language_defn::read_var_value): ...this.
* findvar.c (read_var_value): Update header comment, and change to
call member function instead of function pointer.
* go-lang.c (go_language_data): Likewise.
* language.c (unknown_language_data): Delete la_read_var_value
initializer.
(auto_language_data): Likewise.
* language.h (struct language_data): Delete la_read_var_value
field.
(language_defn::read_var_value): New member function.
(default_read_var_value): Delete declaration.
* m2-lang.c (m2_language_data): Delete la_read_var_value
initializer.
* objc-lang.c (objc_language_data): Likewise.
* opencl-lang.c (opencl_language_data): Likewise.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.
* value.h (default_read_var_value): Delete declaration.

gdb: Convert language la_print_array_index field to a method

This commit changes the language_data::la_print_array_index function
pointer member variable into a member function of language_defn.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* ada-lang.c (ada_print_array_index): Delete function, move
implementation to...
(ada_language::print_array_index): ...here.
(ada_language_data): Delete la_print_array_index initializer.
* c-lang.c (c_language_data): Likewise.
(cplus_language_data): Likewise.
(minimal_language_data): Likewise.
* d-lang.c (d_language_data): Likewise.
* f-lang.c (f_language_data): Likewise.
* go-lang.c (go_language_data): Likewise.
* language.c (default_print_array_index): Delete function, move
implementation to...
(language_defn::print_array_index): ...here.
(unknown_language_data): Delete la_print_array_index initializer.
(auto_language_data): Likewise.
* language.h (struct language_data): Delete la_print_array_index
field.
(language_defn::print_array_index): New member function.
(LA_PRINT_ARRAY_INDEX): Update.
(default_print_array_index): Delete declaration.
* m2-lang.c (m2_language_data): Delete la_print_array_index
initializer.
* objc-lang.c (objc_language_data): Likewise.
* opencl-lang.c (opencl_language_data): Likewise.
* p-lang.c (pascal_language_data): Likewise.
* rust-lang.c (rust_language_data): Likewise.

gdb: Represent all languages as sub-classes of language_defn

This commit converts all languages to sub-classes of a language_defn
base class.

The motivation for this change is to make it easier to add new methods
onto languages without having to update all of the individual language
structures.  In the future it might be possible to move more things,
like expression parsing, into the language class(es) for better
encapsulation, however I have no plans to tackle this in the short
term.

This commit sets up a strategy for transitioning from the current
language system, where each language is an instance of the
language_defn structure, to the class hierarchy system.

The plan is to rename the existing language_defn into language_data,
and make this a base class for the new language_defn class, something
like this:

  struct language_data
  {
    ... old language_defn fields here ...
  };

  struct language_defn : public language_data
  {
    language_defn (const language_data d)
      : language_data (d)
    { .... }
  };

Then each existing language, for example ada_language_defn can be
converted into an instance of language_data, and passed into the
constructor of a new language class, something like this:

  language_data ada_language_data =
  {
    ... old ada_language_defn values here ...
  };

  struct ada_language : public language_defn
  {
    ada_language (ada_language_data)
    { .... }
  };

What this means is that immediately after the conversion nothing much
changes.  Every language is now its own class, but all the old
language fields still exist and can be accessed in the same way.

In later commits I will convert function pointers from the old
language_defn structure into real class methods on language_defn, with
overrides on sub-classes where needed.

At this point I imagine that those fields of the old language_defn
structure that contained only data will probably remain as data fields
within the new language_data base structure, it is only the methods
that I plan to change initially.

I tweaked how we manage the list of languages a bit, each language is
now registered as it is created, and this resulted in a small number
of changes in language.c.

Most of the changes in the *-lang.c files are identical.

There should be no user visible changes after this commit.

gdb/ChangeLog:

* gdb/ada-lang.c (ada_language_defn): Convert to...
(ada_language_data): ...this.
(class ada_language): New class.
(ada_language_defn): New static global.
* gdb/c-lang.c (c_language_defn): Convert to...
(c_language_data): ...this.
(class c_language): New class.
(c_language_defn): New static global.
(cplus_language_defn): Convert to...
(cplus_language_data): ...this.
(class cplus_language): New class.
(cplus_language_defn): New static global.
(asm_language_defn): Convert to...
(asm_language_data): ...this.
(class asm_language): New class.
(asm_language_defn): New static global.
(minimal_language_defn): Convert to...
(minimal_language_data): ...this.
(class minimal_language): New class.
(minimal_language_defn): New static global.
* gdb/d-lang.c (d_language_defn): Convert to...
(d_language_data): ...this.
(class d_language): New class.
(d_language_defn): New static global.
* gdb/f-lang.c (f_language_defn): Convert to...
(f_language_data): ...this.
(class f_language): New class.
(f_language_defn): New static global.
* gdb/go-lang.c (go_language_defn): Convert to...
(go_language_data): ...this.
(class go_language): New class.
(go_language_defn): New static global.
* gdb/language.c (unknown_language_defn): Remove declaration.
(current_language): Initialize to nullptr, real initialization is
moved to _initialize_language.
(languages): Delete global.
(language_defn::languages): Define.
(set_language_command): Use language_defn::languages.
(set_language): Likewise.
(range_error): Likewise.
(language_enum): Likewise.
(language_def): Likewise.
(add_set_language_command): Use language_def::languages for the
language list, and language_def to lookup language pointers.
(skip_language_trampoline): Use language_defn::languages.
(unknown_language_defn): Convert to...
(unknown_language_data): ...this.
(class unknown_language): New class.
(unknown_language_defn): New static global.
(auto_language_defn): Convert to...
(auto_language_data): ...this.
(class auto_language): New class.
(auto_language_defn): New static global.
(language_gdbarch_post_init): Use language_defn::languages.
(_initialize_language): Initialize current_language.
* gdb/language.h (struct language_defn): Rename to...
(struct language_data): ...this.
(struct language_defn): New.
(auto_language_defn): Delete.
(unknown_language_defn): Delete.
(minimal_language_defn): Delete.
(ada_language_defn): Delete.
(asm_language_defn): Delete.
(c_language_defn): Delete.
(cplus_language_defn): Delete.
(d_language_defn): Delete.
(f_language_defn): Delete.
(go_language_defn): Delete.
(m2_language_defn): Delete.
(objc_language_defn): Delete.
(opencl_language_defn): Delete.
(pascal_language_defn): Delete.
(rust_language_defn): Delete.
* gdb/m2-lang.c (m2_language_defn): Convert to...
(m2_language_data): ...this.
(class m2_language): New class.
(m2_language_defn): New static global.
* gdb/objc-lang.c (objc_language_defn): Convert to...
(objc_language_data): ...this.
(class objc_language): New class.
(objc_language_defn): New static global.
* gdb/opencl-lang.c (opencl_language_defn): Convert to...
(opencl_language_data): ...this.
(class opencl_language): New class.
(opencl_language_defn): New static global.
* gdb/p-lang.c (pascal_language_defn): Convert to...
(pascal_language_data): ...this.
(class pascal_language): New class.
(pascal_language_defn): New static global.
* gdb/rust-exp.y (rust_lex_tests): Use language_def to find
language pointer, update comment format.
* gdb/rust-lang.c (rust_language_defn): Convert to...
(rust_language_data): ...this.
(class rust_language): New class.
(rust_language_defn): New static global.

[gdb/testsuite] Fix scrolling in gdb.dwarf2/multidictionary.exp

Consider a gdb_load patch to call the gdb_file_cmd twice:
...
 proc gdb_load { arg } {
     if { $arg != "" } {
+       set res [gdb_file_cmd $arg]
+       if { $res != 0 } {
+           return $res
+       }
        return [gdb_file_cmd $arg]
     }
     return 0
 }
...

With this patch, I run into:
...
(gdb) kill^M
The program is not being run.^M
(gdb) ^M</outputs/gdb.dwarf2/multidictionary/multidictionary^M
<.dwarf2/multidictionary/multidictionary"? (y or n)
ERROR: Couldn't load outputs/gdb.dwarf2/multidictionary/multidictionary \
  into gdb (timeout).
p 1^M
Please answer y or n.^M
<.dwarf2/multidictionary/multidictionary"? (y or n) n^M
Not confirmed.^M
(gdb) UNRESOLVED: gdb.dwarf2/multidictionary.exp: GDB is alive \
  (got interactive prompt)
...

The problem is that the second file command results in a prompt, which is
normally handled by gdb_file_cmd, but not recognized because the initial part
of the prompt is scrolled out.

This in turn is caused by using gdb_spawn_with_cmdline_opts without a
subsequent "set width 0".

Fix this by avoiding gdb_spawn_with_cmdline_opts, and forcing -readline by
temporarily modifying GDBFLAGS instead.

Tested on x86_64-linux.

gdb/testsuite/ChangeLog:

2020-06-02  Tom de Vries  <tdevries@suse.de>

* gdb.dwarf2/multidictionary.exp: Don't use
gdb_spawn_with_cmdline_opts.

binutils archive tests

There are a number of targets that don't support thin archives (*),
and vms doesn't even support file name extensions other than .obj for
archives containing object files.  This patch adjusts the testsuite
to cater for the vms restriction, and reenables testing for non-ELF
alpha targets.  That adds a few alpha-dec-vms fails and one
alpha-linuxecoff fail but testsuite fails on those targets are nothing
new.

(*) It might seem like they do if binutils is built with
--enable-plugins but the plugin archive support is broken, causing the
wrong type of archives to be created by ar for those targets.

* testsuite/binutils-all/ar.exp (obj): Set up object file name
extension.  Use throughout.  Don't completely exclude non-ELF
alpha targets.  Run long_filenames test for tic30.  Exclude thin
archive tests for aix, ecoff and vms.
* estsuite/binutils-all/objdump.exp (obj): Set up object file name
extension.  Use throughout.  Don't exclude non-ELF alpha targets
from "bintest.a".

ELF: Move dyn_relocs to struct elf_link_hash_entry

All ELF backends with shared library support have

  /* Track dynamic relocs copied for this symbol.  */
  struct elf_dyn_relocs *dyn_relocs;

in symbol hash entry.  Move dyn_relocs to struct elf_link_hash_entry
to reduce code duplication.

PR ld/26067
* elf-bfd.h (elf_link_hash_entry): Add dyn_relocs after size.
* elf-s390-common.c (s390_elf_allocate_ifunc_dyn_relocs):
Updated.
* elf32-arc.c (elf_arc_link_hash_entry): Remove dyn_relocs.
(elf_arc_link_hash_newfunc): Updated.
* elf32-arm.c (elf32_arm_link_hash_entry): Remove dyn_relocs.
(elf32_arm_link_hash_newfunc): Updated.
(elf32_arm_copy_indirect_symbol): Likewise.
(elf32_arm_check_relocs): Likewise.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs_for_symbol): Likewise.
* elf32-csky.c (csky_elf_link_hash_entry): Remove dyn_relocs.
(csky_elf_link_hash_newfunc): Updated.
(csky_allocate_dynrelocs): Likewise.
(readonly_dynrelocs): Likewise.
(csky_elf_copy_indirect_symbol): Likewise.
* elf32-hppa.c (elf32_hppa_link_hash_entry): Remove dyn_relocs.
(hppa_link_hash_newfunc): Updated.
(elf32_hppa_copy_indirect_symbol): Likewise.
(elf32_hppa_hide_symbol): Likewise.
(elf32_hppa_adjust_dynamic_symbol): Likewise.
(allocate_dynrelocs): Likewise.
(elf32_hppa_relocate_section): Likewise.
* elf32-i386.c (elf_i386_check_relocs): Likewise.
* elf32-lm32.c (elf_lm32_link_hash_entry): Removed.
(lm32_elf_link_hash_newfunc): Likewise.
(lm32_elf_link_hash_table_create): Updated.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
(lm32_elf_copy_indirect_symbol): Likewise.
* elf32-m32r.c (elf_m32r_link_hash_entry): Removed.
(m32r_elf_link_hash_newfunc): Likewise.
(m32r_elf_link_hash_table_create): Updated.
(m32r_elf_copy_indirect_symbol): Likewise.
(allocate_dynrelocs): Likewise.
* elf32-metag.c (elf_metag_link_hash_entry): Remove dyn_relocs.
(metag_link_hash_newfunc): Updated.
(elf_metag_copy_indirect_symbol): Likewise.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
* elf32-microblaze.c (elf32_mb_link_hash_entry): Remove
dyn_relocs.
(link_hash_newfunc): Updated.
(microblaze_elf_check_relocs): Likewise.
(microblaze_elf_copy_indirect_symbol): Likewise.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
* elf32-nds32.c (elf_nds32_link_hash_entry): Remove dyn_relocs.
(nds32_elf_link_hash_newfunc): Updated.
(nds32_elf_copy_indirect_symbol): Likewise.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
(nds32_elf_check_relocs): Likewise.
* elf32-nios2.c (elf32_nios2_link_hash_entry): Remove dyn_relocs.
(link_hash_newfunc): Updated.
(nios2_elf32_copy_indirect_symbol): Likewise.
(nios2_elf32_check_relocs): Likewise.
(allocate_dynrelocs): Likewise.
* elf32-or1k.c (elf_or1k_link_hash_entry): Remove dyn_relocs.
(or1k_elf_link_hash_newfunc): Updated.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
(or1k_elf_copy_indirect_symbol): Likewise.
* elf32-ppc.c (ppc_elf_link_hash_entry): Remove dyn_relocs.
(ppc_elf_link_hash_newfunc): Updated.
(ppc_elf_copy_indirect_symbol): Likewise.
(ppc_elf_check_relocs): Likewise.
(readonly_dynrelocs): Likewise.
(ppc_elf_adjust_dynamic_symbol): Likewise.
(allocate_dynrelocs): Likewise.
(ppc_elf_relocate_section): Likewise.
* elf32-s390.c (elf_s390_link_hash_entry): Remove dyn_relocs.
(link_hash_newfunc): Updated.
(elf_s390_copy_indirect_symbol): Likewise.
(readonly_dynrelocs): Likewise.
(elf_s390_adjust_dynamic_symbol): Likewise.
(allocate_dynrelocs): Likewise.
* elf32-sh.c (elf_sh_link_hash_entry): Remove dyn_relocs.
(sh_elf_link_hash_newfunc): Updated.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
(sh_elf_copy_indirect_symbol): Likewise.
(sh_elf_check_relocs): Likewise.
* elf32-tic6x.c (elf32_tic6x_link_hash_entry): Removed.
(elf32_tic6x_link_hash_newfunc): Likewise.
(elf32_tic6x_link_hash_table_create): Updated.
(readonly_dynrelocs): Likewise.
(elf32_tic6x_check_relocs): Likewise.
(elf32_tic6x_allocate_dynrelocs): Likewise.
* elf32-tilepro.c (tilepro_elf_link_hash_entry): Remove
dyn_relocs.
(link_hash_newfunc): Updated.
(tilepro_elf_copy_indirect_symbol): Likewise.
(tilepro_elf_check_relocs): Likewise.
(allocate_dynrelocs): Likewise.
* elf64-ppc.c (ppc_link_hash_entry): Remove dyn_relocs.
(ppc64_elf_copy_indirect_symbol): Updated.
(ppc64_elf_check_relocs): Likewise.
(readonly_dynrelocs): Likewise.
(ppc64_elf_adjust_dynamic_symbol): Likewise.
(dec_dynrel_count): Likewise.
(allocate_dynrelocs): Likewise.
(ppc64_elf_relocate_section): Likewise.
* elf64-s390.c (elf_s390_link_hash_entry): Remove dyn_relocs.
(link_hash_newfunc): Updated.
(elf_s390_copy_indirect_symbol): Likewise.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
* elf64-x86-64.c (elf_x86_64_check_relocs): Likewise.
* elfnn-aarch64.c (elf_aarch64_link_hash_entry): Remove
dyn_relocs.
(elfNN_aarch64_link_hash_newfunc): Updated.
(elfNN_aarch64_copy_indirect_symbol): Likewise.
(readonly_dynrelocs): Likewise.
(need_copy_relocation_p): Likewise.
(elfNN_aarch64_allocate_dynrelocs): Likewise.
(elfNN_aarch64_allocate_ifunc_dynrelocs): Likewise.
* elfnn-riscv.c (riscv_elf_link_hash_entry): Remove dyn_relocs.
(link_hash_newfunc): Updated.
(riscv_elf_copy_indirect_symbol): Likewise.
(riscv_elf_check_relocs): Likewise.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
* elfxx-sparc.c (_bfd_sparc_elf_link_hash_entry): Remove
dyn_relocs.
(link_hash_newfunc): Updated.
(_bfd_sparc_elf_copy_indirect_symbol): Likewise.
(_bfd_sparc_elf_check_relocs): Likewise.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
* elfxx-tilegx.c (tilegx_elf_link_hash_entry): Remove dyn_relocs.
(link_hash_newfunc): Updated.
(tilegx_elf_copy_indirect_symbol): Likewise.
(tilegx_elf_check_relocs): Likewise.
(readonly_dynrelocs): Likewise.
(allocate_dynrelocs): Likewise.
* elfxx-x86.c (elf_x86_allocate_dynrelocs): Likewise.
(readonly_dynrelocs): Likewise.
(_bfd_x86_elf_copy_indirect_symbol): Likewise.
* elfxx-x86.h (elf_x86_link_hash_entry): Remove dyn_relocs.

Automatic date update in version.in

gas: Fix checking for backwards .org with negative offset

This patch fixes internal errors in (at least) arm and aarch64 GAS
when assembling code that attempts a negative .org.  The bug appears
to be a regression introduced in binutils-2.29 by commit 9875b36538d.

* write.c (relax_segment): Fix handling of negative offset when
relaxing an rs_org frag.
* testsuite/gas/aarch64/org-neg.d: New test.
* testsuite/gas/aarch64/org-neg.l: Error output for test.
* testsuite/gas/aarch64/org-neg.s: Input for test.
* testsuite/gas/arm/org-neg.d: New test.
* testsuite/gas/arm/org-neg.l: Error output for test.
* testsuite/gas/arm/org-neg.s: Input for test.

alpha-vms: ETIR checks

Better validity checks, and remove a fuzzer vulnerability of sorts that
targeted the store-immediate-repeat command with a zero length but
very large repeat counts to chew cpu.

* vms-alpha.c (_bfd_vms_slurp_etir): Check bound for the current
command against cmd_length, not the end of record.  For
ETIR__C_STO_IMMR check size against cmd_length, mask repeat count
to 32-bits and break out on zero size.  Add ETIR__C_STC_LP_PSB
cmd_length test.

Regen opcodes/bpf-desc.c

The last regen used an old version of cgen.

* bpf-desc.c: Regenerate.

gdb: Preserve is-stmt lines when switch between files

After the is-stmt support commit:

  commit 8c95582da858ac981f689a6f599acacb8c5c490f
  Date:   Mon Dec 30 21:04:51 2019 +0000

      gdb: Add support for tracking the DWARF line table is-stmt field

A regression was observed where a breakpoint could no longer be placed
in some cases.

Consider a line table like this:

  File 1: test.c
  File 2: test.h

  | Addr | File | Line | Stmt |
  |------|------|------|------|
  | 1    | 1    | 16   | Y    |
  | 2    | 1    | 17   | Y    |
  | 3    | 2    | 21   | Y    |
  | 4    | 2    | 22   | Y    |
  | 4    | 1    | 18   | N    |
  | 5    | 2    | 23   | N    |
  | 6    | 1    | 24   | Y    |
  | 7    | 1    | END  | Y    |
  |------|------|------|------|

Before the is-stmt patch GDB would ignore any non-stmt lines, so GDB
built two line table structures:

  File 1                 File 2
  ------                 ------

  | Addr | Line |        | Addr | Line |
  |------|------|        |------|------|
  | 1    | 16   |        | 3    | 21   |
  | 2    | 17   |        | 4    | 22   |
  | 3    | END  |        | 6    | END  |
  | 6    | 24   |        |------|------|
  | 7    | END  |
  |------|------|

After the is-stmt patch GDB now records non-stmt lines, so the
generated line table structures look like this:

  File 1                   File 2
  ------                   ------

  | Addr | Line | Stmt |  | Addr | Line | Stmt |
  |------|------|------|  |------|------|------|
  | 1    | 16   | Y    |  | 3    | 21   | Y    |
  | 2    | 17   | Y    |  | 4    | 22   | Y    |
  | 3    | END  | Y    |  | 4    | END  | Y    |
  | 4    | 18   | N    |  | 5    | 23   | N    |
  | 5    | END  | Y    |  | 6    | END  | Y    |
  | 6    | 24   | Y    |  |------|------|------|
  | 7    | END  | Y    |
  |------|------|------|

The problem is that in 'File 2', end END marker at address 4 causes
the previous line table entry to be discarded, so we actually end up
with this:

  File 2
  ------

  | Addr | Line | Stmt |
  |------|------|------|
  | 3    | 21   | Y    |
  | 4    | END  | Y    |
  | 5    | 23   | N    |
  | 6    | END  | Y    |
  |------|------|------|

When a user tries to place a breakpoint in file 2 at line 22, this is
no longer possible.

The solution I propose here is that we ignore line table entries that
would trigger a change of file if:

  1. The new line being added is at the same address as the previous
  line, and

  2. We have previously seen an is-stmt line at the current address.

The result of this is that GDB switches file, and knows that some line
entry (or entries) are going to be discarded, prefer to keep is-stmt
lines and discard non-stmt lines.

After this commit the lines tables are now:

  File 1                   File 2
  ------                   ------

  | Addr | Line | Stmt |  | Addr | Line | Stmt |
  |------|------|------|  |------|------|------|
  | 1    | 16   | Y    |  | 3    | 21   | Y    |
  | 2    | 17   | Y    |  | 4    | 22   | Y    |
  | 3    | END  | Y    |  | 5    | 23   | N    |
  | 5    | END  | Y    |  | 6    | END  | Y    |
  | 6    | 24   | Y    |  |------|------|------|
  | 7    | END  | Y    |
  |------|------|------|

We've lost the non-stmt entry for file 1, line 18, but retained the
is-stmt entry for file 2, line 22.  The user can now place a
breakpoint at that location.

One problem that came from this commit was the test
gdb.cp/step-and-next-inline.exp, which broke in several places.  After
looking at this test again I think that in some cases this test was
only ever passing by pure luck.  The debug GCC is producing for this
test is pretty broken.  I raised this GCC bug:

  https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94474

for this and disabled one entire half of the test.  There are still
some cases in here that do pass, and if/when GCC is fixed it would be
great to enable this test again.

gdb/ChangeLog:

* dwarf2/read.c (class lnp_state_machine) <m_last_address>: New
member variable.
<m_stmt_at_address>: New member variable.
(lnp_state_machine::record_line): Don't record some lines, update
tracking of is_stmt at the same address.
(lnp_state_machine::lnp_state_machine): Initialise new member
variables.

gdb/testsuite/ChangeLog:

* gdb.cp/step-and-next-inline.exp (do_test): Skip all tests in the
use_header case.
* gdb.dwarf2/dw2-inline-header-1.exp: New file.
* gdb.dwarf2/dw2-inline-header-2.exp: New file.
* gdb.dwarf2/dw2-inline-header-3.exp: New file.
* gdb.dwarf2/dw2-inline-header-lbls.c: New file.
* gdb.dwarf2/dw2-inline-header.c: New file.
* gdb.dwarf2/dw2-inline-header.h: New file.

hurd: Add shared mig declarations

We are using -Werror=missing-declarations, and the _S.h files generated
by mig do not currently include a declaration for the server routine.
gnu-nat.c used to have its own external declarations, but better just
share them between gnu-nat.c and the _S.c files.

Fixes

exc_request_S.c:177:24: error: no previous declaration for ‘exc_server’ [-Werror=missing-declarations]
  177 | mig_external boolean_t exc_server

gdb/ChangeLog:

* config/i386/i386gnu.mn [%_S.o %_U.o] (COMPILE.post): Add
"-include gnu-nat-mig.h".
* gnu-nat-mig.h: New file.
* gnu-nat.c: Include "gnu-nat-mig.h".
(exc_server, msg_reply_server, notify_server,
process_reply_server): Remove declarations.

Automatic date update in version.in

gnu-nat: Move local functions inside gnu_nat_target class

This allows to have the process_stratum_target object at hand for future use in
the gdb API, and only use gnu_target from external calls.

gdb/Changelog:

* gnu-nat.h (inf_validate_procs, inf_suspend, inf_set_traced,
steal_exc_port, proc_get_state, inf_clear_wait, inf_cleanup,
inf_startup, inf_update_suspends, inf_set_pid, inf_steal_exc_ports,
inf_validate_procinfo, inf_validate_task_sc, inf_restore_exc_ports,
inf_set_threads_resume_sc, inf_set_threads_resume_sc_for_signal_thread,
inf_resume, inf_set_step_thread, inf_detach, inf_attach, inf_signal,
inf_continue, make_proc, proc_abort, _proc_free, proc_update_sc,
proc_get_exception_port, proc_set_exception_port, _proc_get_exc_port,
proc_steal_exc_port, proc_restore_exc_port, proc_trace): Move functions
to gnu_nat_target class.
* gnu-nat.c: Likewise.
(inf_update_procs, S_proc_wait_reply, set_task_pause_cmd,
set_task_exc_port_cmd, set_signals_cmd, set_thread_pause_cmd,
set_thread_exc_port_cmd): Call inf_validate_procs through gnu_target
object.
(gnu_nat_target::create_inferior, gnu_nat_target::detach): Pass `this'
instead of `gnu_target'.

Automatic date update in version.in

hurd: unwinding support over signal trampolines

This allows to get full backtrace from signal handlers, otherwise the
backtrace stops at the trampoline that calls the handler.

This needs special knowledge how the trampoline records register context
for the sigreturn call after signal handling.

gdb/ChangeLog:

* i386-gnu-tdep.c: Include "gdbcore.h"
(gnu_sigtramp_code, i386_gnu_sc_reg_offset): New arrays.
        (GNU_SIGTRAMP_LEN, GNU_SIGTRAMP_TAIL,
        I386_GNU_SIGCONTEXT_THREAD_STATE_OFFSET): New macros
        (i386_gnu_sigtramp_start, i386_gnu_sigtramp_p,
        i386_gnu_sigcontext_addr): New functions
        (i386gnu_init_abi): Register i386_gnu_sigtramp_p,
        i386_gnu_sigcontext_addr, and i386_gnu_sc_reg_offset in the gdbarch
        tdep.

hurd: fix pushing target on inferior creation

This fixes creating inferiors, which was broken since 5b6d1e4fa
('Multi-target support')

gdb/ChangeLog:

* gnu-nat.c (gnu_nat_target::create_inferior): Move push_target call
before fork_inferior call. Avoid calling it if target_is_pushed returns
true.

hurd: add gnu_target pointer to fix thread API calls

Fixes

../../gdb/gnu-nat.c:1110:28: error: cannot convert ‘ptid_t’ to ‘process_stratum_target*’
 1110 |        thread_change_ptid (inferior_ptid, ptid);

and others related to 5b6d1e4fa ("Multi-target support")

gdb/ChangeLog:

* gnu-nat.h (gnu_target): New variable declaration.
* i386-gnu-nat.c (_initialize_i386gnu_nat): Initialize
gnu_target.
* gnu-nat.c (gnu_target): New variable.
(inf_validate_procs): Pass gnu_target to thread_change_ptid,
add_thread_silent, and add_thread calls.
(gnu_nat_target::create_inferior): Pass gnu_target to
add_thread_silent, thread_change_ptid call.
(gnu_nat_target::detach): Pass gnu_target to detach_inferior
call.

hurd: remove unused variables

Fixes

../../gdb/gnu-nat.c:2554:7: error: unused variable ‘res’ [-Werror=unused-variable]
 2554 |   int res;
../../gdb/gnu-nat.c:2644:20: error: unused variable ‘old_address’ [-Werror=unused-variable]
 2644 |       vm_address_t old_address = region_address;

gdb/ChangeLog:

* gnu-nat.c (gnu_xfer_auxv): Remove unused `res' variable.
(gnu_nat_target::find_memory_regions): Remove unused
`old_address' variable.

hurd: add missing include

Fixes

../../gdb/gnu-nat.c:2522:14: error: ‘target_gdbarch’ was not declared in this scope; did you mean ‘target_detach’?
 2522 |    paddress (target_gdbarch (), memaddr), pulongest (len),

gdb/Changelog:

* gnu-nat.c: Include "gdbarch.h".

hurd: make function cast stronger

Fixes

process_reply_S.c:104:23: error: function called through a non-compatible type [-Werror]
  104 |      OutP->RetCode = (*(kern_return_t (*)(mach_port_t, kern_return_t)) S_proc_setmsgport_reply) (In0P->Head.msgh_request_port, In0P-

As the existing comment says, it is in general not safe to drop some
parameters like this, but this is the error handling case, where the
called function does not actually read them, and mig is currently planned
to be used on i386 and x86_64 only, where this is not a problem. As the
existing comment says, fixing it properly would be far from trivial:
we can't just pass 0 for them, as they might not be scalar.

gdb/ChangeLog:

* reply_mig_hack.awk (Error return): Cast function through
void *, to bypass compiler function call check.

hurd: add missing awk script dependency

To make sure the *_reply_S.[ch] files get regenerated whenever we change
the awk script.

gdb/ChangeLog:

* config/i386/i386gnu.mn (%_reply_S.c): Add dependency on
$(srcdir)/reply_mig_hack.awk.

hurd: fix gnu_debug_flag type

Fixes

../../gdb/gnu-nat.c:96:6: error: conflicting declaration ‘bool gnu_debug_flag’
   96 | bool gnu_debug_flag = false;
../../gdb/gnu-nat.c: In function ‘void _initialize_gnu_nat()’:
../../gdb/gnu-nat.c:3511:7: error: cannot

gdb/ChangeLog:

* gnu-nat.h (gnu_debug_flag): Set type to bool.

gdb: change bug URL to https

gdb/ChangeLog:

* configure.ac (ACX_BUGURL): change bug URL to https.

Signed-off-by: Jonny Grant <jg@jguk.org>
Change-Id: If8d939e50c830e3e452c3e8f7a7aee06d9c96645

replace_typedefs: handle templates in namespaces

GDB currently crashes with infinite recursion, if you set a breakpoint
on a function inside a namespace that includes a template on its fully
qualified name, and, the template's name is also used as typedef in
the global scope that expands to a name that includes the template
name in its qualified name.  For example, from the testcase added by
this commit:

 namespace NS1 { namespace NS2 {

 template<typename T> struct Templ1
 {
   T x;

   Templ1 (object_p) {}
 }} // namespace NS1::NS2

 using Templ1 = NS1::NS2::Templ1<unsigned>;

Setting a breakpoint like so:

(gdb) break NS1::NS2::Templ1<int>::Templ1(NS1::NS2::object*)

Results in infinite recursion, with this cycle (started by
cp_canonicalize_string_full) repeating until the stack is exhausted:

 ...
 #1709 0x000000000055533c in inspect_type (info=0x38ff720, ret_comp=0xd83be10, finder=0x0, data=0x0) at /home/pedro/gdb/mygit/src/gdb/cp-support.c:267
 #1710 0x0000000000555a6f in replace_typedefs (info=0x38ff720, ret_comp=0xd83be10, finder=0x0, data=0x0) at /home/pedro/gdb/mygit/src/gdb/cp-support.c:475
 #1711 0x0000000000555a36 in replace_typedefs (info=0x38ff720, ret_comp=0xd83be70, finder=0x0, data=0x0) at /home/pedro/gdb/mygit/src/gdb/cp-support.c:470
 #1712 0x0000000000555800 in replace_typedefs_qualified_name (info=0x38ff720, ret_comp=0xd839470, finder=0x0, data=0x0) at /home/pedro/gdb/mygit/src/gdb/cp-support.c:389
 #1713 0x0000000000555a8c in replace_typedefs (info=0x38ff720, ret_comp=0xd839470, finder=0x0, data=0x0) at /home/pedro/gdb/mygit/src/gdb/cp-support.c:479
 ...

The demangle component tree for that symbol name looks like this:

d_dump tree for NS1::NS2::Templ1<int>::Templ1(NS1::NS2::object*):
typed name
  qualified name
    name 'NS1'
    qualified name
      name 'NS2'
      qualified name
        template                  <<<<<<<<<<
          name 'Templ1'
          template argument list
            builtin type int
        name 'Templ1'
  function type
    argument list
      pointer
        qualified name
          name 'NS1'
          qualified name
            name 'NS2'
            name 'object'

The recursion starts at replace_typedefs_qualified_name, which doesn't
handle the "template" node, and thus doesn't realize that the template
name actually has the fully qualified name NS1::NS2::Templ1.
replace_typedefs_qualified_name calls into replace_typedefs on the
template node, and that ends up in inspect_type looking up for a
symbol named "Templ1", which finds the global namespace typedef, which
itself expands to NS1::NS2::Templ1.  GDB then tries replacing typedefs
in that newly expanded name, which ends up again in
replace_typedefs_qualified_name, trying to expand a fully qualified
name with "NS::NS2::Templ1<unsigned>" in its name, which results in
recursion whenever the template node is reached.

Fix this by teaching replace_typedefs_qualified_name how to handle
template nodes.  It needs handling in two places: the first spot
handles the symbol above; the second spot handles a symbol like this,
from the new test:

(gdb) b NS1::NS2::grab_it(NS1::NS2::Templ1<int>*)
d_dump tree for NS1::NS2::grab_it(NS1::NS2::Templ1<int>*):
typed name
  qualified name
    name 'NS1'
    qualified name
      name 'NS2'
      name 'grab_it'
  function type
    argument list
      pointer
        qualified name
          name 'NS1'
          qualified name
            name 'NS2'
            template             <<<<<<<<
              name 'Templ1'
              template argument list
                builtin type int

What's different in this case is that the template node appears on the
right child node of a qualified name, instead of on the left child.

The testcase includes a test that checks whether template aliases are
correctly replaced by GDB too.  That fails with GCC due to GCC PR
95437, which makes GDB not know about a typedef for
"NS1::NS2::AliasTempl<int>".  GCC emits a typedef named
"NS1::NS2::AliasTempl" instead, with no template parameter info.  The
test passes with Clang (5.0.2 at least).  See more details here:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95437

gdb/ChangeLog:
2020-05-30  Pedro Alves  <palves@redhat.com>

* cp-support.c (replace_typedefs_template): New.
(replace_typedefs_qualified_name): Handle
DEMANGLE_COMPONENT_TEMPLATE.

gdb/testsuite/ChangeLog:
2020-05-30  Pedro Alves  <palves@redhat.com>

* gdb.linespec/cp-replace-typedefs-ns-template.cc: New.
* gdb.linespec/cp-replace-typedefs-ns-template.exp: New.

Automatic date update in version.in

gdb: rename dwarf2_per_objfile variables/fields to per_objfile

While doing the psymtab-sharing patchset, I avoided renaming variables
unnecessarily to avoid adding noise to patches, but I'd like to do it
now.  Basically, we have these dwarf2 per-something structures:

  - dwarf2_per_objfile
  - dwarf2_per_bfd
  - dwarf2_per_cu_data

I named the instances of dwarf2_per_bfd `per_bfd` and most of instances
of dwarf2_per_cu_data are called `per_cu`.  Most pre-existing instances
of dwarf2_per_objfile are named `dwarf2_per_objfile`.  For consistency
with the other type, I'd like to rename them to just `per_objfile`.  The
`dwarf2_` prefix is superfluous, since it's already clear we are in
dwarf2 code.  It also helps reducing the line wrapping by saving 7
precious columns.

gdb/ChangeLog:

* dwarf2/comp-unit.c, dwarf2/comp-unit.h, dwarf2/index-cache.c,
dwarf2/index-cache.h, dwarf2/index-write.c,
dwarf2/index-write.h, dwarf2/line-header.c,
dwarf2/line-header.h, dwarf2/macro.c, dwarf2/macro.h,
dwarf2/read.c, dwarf2/read.h: Rename struct dwarf2_per_objfile
variables and fields from `dwarf2_per_objfile` to just
`per_objfile` throughout.

Change-Id: I3c45cdcc561265e90df82cbd36b4b4ef2fa73aef

Fix build errors in with clang in gdb.compile/compile-cplus.c

Clang fails to compile the file, with the following error:
  fatal error: 'iostream' file not found

This prevents the following testcase from executing:
  gdb.compile/compile-cplus.exp

The testcase sets additional_flags when building with GCC, which
this commit causes to also be set when building with clang.  This
makes the testcase fail to build with a different error:
  warning: treating 'c' input as 'c++' when in C++ mode, this behavior
  is deprecated [-Wdeprecated]
so this commit adds -Wno-deprecated in two places to sidestep this.
Note that, while allowing the testcase to build, this commit reveals
failures when the testsuite is built using clang.

gdb/testsuite/ChangeLog:

* gdb.compile/compile-cplus.exp (additional_flags): Also
set when building with clang.
(additional_flags, srcfilesoptions): Pass -Wno-deprecated
when building with clang.

Fix file-not-found error with clang in gdb.arch/i386-{avx,sse}.c

Clang fails to compile two testcases with the following error:
  fatal error: 'nat/x86-cpuid.h' file not found

This prevents the following testcases from executing:
  gdb.arch/i386-avx.exp
  gdb.arch/i386-sse.exp

Both testcases set additional_flags when building with GCC.
This commit causes the additional_flags to also be used when
building with clang.  Note that, while fixing the build, this
commit reveals several new failures when using clang to build
the testsuite.

gdb/testsuite/ChangeLog:

* gdb.arch/i386-avx.exp (additional_flags): Also set when
building with clang.
* gdb.arch/i386-sse.exp (additional_flags): Likewise.

Build two gdb.cp testcases with -Wno-unused-comparison

Clang fails to compile two testcases with the following error:
  warning: equality comparison result unused [-Wunused-comparison]

This prevents the following testcases from executing:
  gdb.cp/koenig.exp
  gdb.cp/operator.exp

This commit builds those testcases with -Wno-unused-comparison, to
avoid the failure.  Note that this commit reveals a new failure,
"FAIL: gdb.cp/koenig.exp: p foo (p_union)" when the testsuite is
compiled using clang.

gdb/testsuite/ChangeLog:

* gdb.cp/koenig.exp (prepare_for_testing): Add
additional_flags=-Wno-unused-comparison.
* gdb.cp/operator.exp (prepare_for_testing): Likewise.

binutils: Add myself as Xtensa maintainer

2020-05-28  Max Filippov  <jcmvbkbc@gmail.com>
binutils/
* MAINTAINERS (Xtensa): Add myself as maintainer.

Automatic date update in version.in

bfd: fix handling of R_BPF_INSN_{32,64} relocations.

2020-05-28  David Faust  <david.faust@oracle.com>

* elf64-bpf.c (bpf_elf_relocate_section): Fix handling of
R_BPF_INSN_{32,64} relocations.

cpu,opcodes: add instruction semantics to bpf.cpu and minor fixes

This patch adds semantic RTL descriptions to the eBPF instructions
defined in cpu/bpf.cpu.  It also contains a couple of minor
improvements.

Tested in bpf-unknown-none targets.
No regressions.

cpu/ChangeLog:

2020-05-28  Jose E. Marchesi  <jose.marchesi@oracle.com>
    David Faust <david.faust@oracle.com>

* bpf.cpu (define-alu-insn-un): Add definitions of semantics.
(define-alu-insn-mov): Likewise.
(daib): Likewise.
(define-alu-instructions): Likewise.
(define-endian-insn): Likewise.
(define-lddw): Likewise.
(dlabs): Likewise.
(dlind): Likewise.
(dxli): Likewise.
(dxsi): Likewise.
(dsti): Likewise.
(define-ldstx-insns): Likewise.
(define-st-insns): Likewise.
(define-cond-jump-insn): Likewise.
(dcji): Likewise.
(define-condjump-insns): Likewise.
(define-call-insn): Likewise.
(ja): Likewise.
("exit"): Likewise.
(define-atomic-insns): Likewise.
(sem-exchange-and-add): New macro.
* bpf.cpu ("brkpt"): New instruction.
(bpfbf): Set word-bitsize to 32 and insn-endian big.
(h-gpr): Prefer r0 to `a' and r6 to `ctx'.
(h-pc): Expand definition.
* bpf.opc (bpf_print_insn): Set endian_code to BIG.

opcodes/ChangeLog:

2020-05-28  Jose E. Marchesi  <jose.marchesi@oracle.com>
    David Faust <david.faust@oracle.com>

* bpf-desc.c: Regenerate.
* bpf-opc.h: Likewise.
* bpf-opc.c: Likewise.
* bpf-dis.c: Likewise.

gdb: add comment in dwarf_evaluate_loc_desc::push_dwarf_reg_entry_value

Add a comment to clarify why we temporarily override some of the
context's fields, and especially the per_objfile field.  A longer
explanation can be found in this previous commit

    44486dcf19b ("gdb: use caller objfile in dwarf_evaluate_loc_desc::push_dwarf_reg_entry_value")

gdb/ChangeLog:

* dwarf2/loc.c (class dwarf_evaluate_loc_desc)
<push_dwarf_reg_entry_value>: Add comment.

Change-Id: I60c6e1062799f729b30a9db78bcb6448783324b4

Fix Python3.9 related runtime problems

Python3.9b1 is now available on Rawhide.  GDB w/ Python 3.9 support
can be built using the configure switch -with-python=/usr/bin/python3.9.

Attempting to run gdb/Python3.9 segfaults on startup:

    #0  0x00007ffff7b0582c in PyEval_ReleaseLock () from /lib64/libpython3.9.so.1.0
    #1  0x000000000069ccbf in do_start_initialization ()
at worktree-test1/gdb/python/python.c:1789
    #2  _initialize_python ()
at worktree-test1/gdb/python/python.c:1877
    #3  0x00000000007afb0a in initialize_all_files () at init.c:237
    ...

Consulting the the documentation...

https://docs.python.org/3/c-api/init.html

...we find that PyEval_ReleaseLock() has been deprecated since version
3.2.  It recommends using PyEval_SaveThread or PyEval_ReleaseThread()
instead.  In do_start_initialization, in gdb/python/python.c, we
can replace the calls to PyThreadState_Swap() and PyEval_ReleaseLock()
with a single call to PyEval_SaveThread.   (Thanks to Keith Seitz
for working this out.)

With that in place, GDB gets a little bit further.  It still dies
on startup, but the backtrace is different:

    #0  0x00007ffff7b04306 in PyOS_InterruptOccurred ()
       from /lib64/libpython3.9.so.1.0
    #1  0x0000000000576e86 in check_quit_flag ()
at worktree-test1/gdb/extension.c:776
    #2  0x0000000000576f8a in set_active_ext_lang (now_active=now_active@entry=0x983c00 <extension_language_python>)
at worktree-test1/gdb/extension.c:705
    #3  0x000000000069d399 in gdbpy_enter::gdbpy_enter (this=0x7fffffffd2d0,
gdbarch=0x0, language=0x0)
at worktree-test1/gdb/python/python.c:211
    #4  0x0000000000686e00 in python_new_inferior (inf=0xddeb10)
at worktree-test1/gdb/python/py-inferior.c:251
    #5  0x00000000005d9fb9 in std::function<void (inferior*)>::operator()(inferior*) const (__args#0=<optimized out>, this=0xccad20)
at /usr/include/c++/10/bits/std_function.h:617
    #6  gdb::observers::observable<inferior*>::notify (args#0=0xddeb10,
this=<optimized out>)
at worktree-test1/gdb/../gdbsupport/observable.h:106
    #7  add_inferior_silent (pid=0)
at worktree-test1/gdb/inferior.c:113
    #8  0x00000000005dbcb8 in initialize_inferiors ()
at worktree-test1/gdb/inferior.c:947
    ...

We checked with some Python Developers and were told that we should
acquire the GIL prior to calling any Python C API function.  We
definitely don't have the GIL for calls of PyOS_InterruptOccurred().

I moved class_gdbpy_gil earlier in the file and use it in
gdbpy_check_quit_flag() to acquire (and automatically release) the
GIL.

With those changes in place, I was able to run to a GDB prompt.  But,
when trying to quit, it segfaulted again due to due to some other
problems with gdbpy_check_quit_flag():

    Thread 1 "gdb" received signal SIGSEGV, Segmentation fault.
    0x00007ffff7bbab0c in new_threadstate () from /lib64/libpython3.9.so.1.0
    (top-gdb) bt 8
    #0  0x00007ffff7bbab0c in new_threadstate () from /lib64/libpython3.9.so.1.0
    #1  0x00007ffff7afa5ea in PyGILState_Ensure.cold ()
       from /lib64/libpython3.9.so.1.0
    #2  0x000000000069b58c in gdbpy_gil::gdbpy_gil (this=<synthetic pointer>)
at worktree-test1/gdb/python/python.c:278
    #3  gdbpy_check_quit_flag (extlang=<optimized out>)
at worktree-test1/gdb/python/python.c:278
    #4  0x0000000000576e96 in check_quit_flag ()
at worktree-test1/gdb/extension.c:776
    #5  0x000000000057700c in restore_active_ext_lang (previous=0xe9c050)
at worktree-test1/gdb/extension.c:729
    #6  0x000000000088913a in do_my_cleanups (
pmy_chain=0xc31870 <final_cleanup_chain>,
old_chain=0xae5720 <sentinel_cleanup>)
at worktree-test1/gdbsupport/cleanups.cc:131
    #7  do_final_cleanups ()
at worktree-test1/gdbsupport/cleanups.cc:143

In this case, we're trying to call a Python C API function after
Py_Finalize() has been called from finalize_python().  I made
finalize_python set gdb_python_initialized to false and then cause
check_quit_flag() to return early when it's false.

With these changes in place, GDB seems to be working again with
Python3.9b1.  I think it likely that there are other problems lurking.
I wouldn't be surprised to find that there are other calls into Python
where we don't first make sure that we have the GIL.  Further changes
may well be needed.

I see no regressions testing on Rawhide using a GDB built with the
default Python version (3.8.3) versus one built using Python 3.9b1.

I've also tested on Fedora 28, 29, 30, 31, and 32 (all x86_64) using
the default (though updated) system installed versions of Python on
those OSes.  This means that I've tested against Python versions
2.7.15, 2.7.17, 2.7.18, 3.7.7, 3.8.2, and 3.8.3.  In each case GDB
still builds without problem and shows no regressions after applying
this patch.

gdb/ChangeLog:

2020-MM-DD  Kevin Buettner  <kevinb@redhat.com>
    Keith Seitz  <keiths@redhat.com>

* python/python.c (do_start_initialization): For Python 3.9 and
later, call PyEval_SaveThread instead of PyEval_ReleaseLock.
(class gdbpy_gil): Move to earlier in file.
(finalize_python): Set gdb_python_initialized.
(gdbpy_check_quit_flag): Acquire GIL via gdbpy_gil.  Return early
when not initialized.

Fix all unexpected failures in gas testsuite for pdp11-aout.

These failures were caused by the PDP11's mix of little-endian octets
in shorts but shorts in big endian order for long or quad so regexps
did not match.  Also tests used addresses as values in .long which
required BRD_RELOC_32 that was not implemented.

* gas/config/tc-pdp11.c (md_number_to_chars): Implement .quad
* gas/testsuite/gas/all/gas.exp: Select alternate test scripts for
pdp11, skip octa test completely.
* gas/testsuite/gas/all/eqv-dot-pdp11.s: Identical to eqv-dot.s
* gas/testsuite/gas/all/eqv-dot-pdp11.d: Match different octet order.
* gas/testsuite/gas/all/cond-pdp11.l: Match different octet order.

* bfd/pdp11.c: Implement BRD_RELOC_32 to relocate the low 16 bits of
addreses in .long (used in testsuites) and .stab values.

Fix "enumeration values not handled in switch" error in testsuite

When running the testsuite with clang, gdb.base/sigaltstack.c
fails to compile with the following error:
  warning: enumeration values 'LEAF' and 'NR_LEVELS' not handled
    in switch [-Wswitch]

This prevents the gdb.base/sigaltstack.exp from executing.
This commit fixes.

gdb/testsuite/ChangeLog:

* gdb.base/sigaltstack.c (catcher): Add default case to switch
statement.

Have the linker fail if any attempt to link in an executable is made.

PR 26047
* ldelf.c (ldelf_after_open): Fail if attempting to link one
executable into another.  Ensure that the test is made for all
forms of linking.

Stop the linker from accepting executable ELF files as inputs to other links.

PR 26047
* ldelf.c (ldelf_after_open): Fail if attempting to link one
executable into another.

gdb: use caller objfile in dwarf_evaluate_loc_desc::push_dwarf_reg_entry_value

In commit

    89b07335fe ("Add dwarf2_per_objfile to dwarf_expr_context and dwarf2_frame_cache")

I replaced the offset property of dwarf_expr_context by a per_objfile
property (since we can get the text offset from the objfile).  The
previous code in dwarf_evaluate_loc_desc::push_dwarf_reg_entry_value
(dwarf_evaluate_loc_desc derives from dwarf_expr_context) did
temporarily override the offset property while evaluating a DWARF
sub-expression.  I speculated that this sub-expression always came from
the same objfile as the outer expression, so I didn't see the need to
temporarily override the per_objfile property in the new code.  A later
commit:

    9f47c70716 ("Remove dwarf2_per_cu_data::objfile ()")

added the following assertion to verify this:

    gdb_assert (this->per_objfile == caller_per_objfile);

It turns out that this is not true.  Call sites can refer to function in
another objfile, and therefore the caller's objfile can be different
from the callee's objfile.  This can happen when the call site DIE in the
DWARF represents a function call done through a function pointer.  The
DIE can't describe statically which function is being called, since it's
variable and not known at compile time.  Instead, it provides an
expression that evaluates to the address of the function being called.
In this case, the called function can very well be in a separate
objfile.

Fix this by overriding the per_objfile property while evaluating the
sub-expression.

This was exposed by the gdb.base/catch-load.exp test failing on openSUSE
Tumbleweed with the glibc debug info installed.  It was also reported to
fail on Fedora.

When I investigated the problem, the particular call site on which we
did hit the assert was coming from this DIE, in
/usr/lib/debug/lib64/libc-2.31.so-2.31-5.1.x86_64.debug on openSUSE
Tumbleweed:

    0x0091aa10:     DW_TAG_GNU_call_site
                      DW_AT_low_pc [DW_FORM_addr]   (0x00000000001398e0)
                      DW_AT_GNU_call_site_target [DW_FORM_exprloc]  (DW_OP_fbreg -272, DW_OP_deref)
                      DW_AT_sibling [DW_FORM_ref4]  (0x0091aa2b)

And for you curious out there, this call site is found in this function:

    0x0091a91d:   DW_TAG_subprogram
                    DW_AT_external [DW_FORM_flag_present]   (true)
                    DW_AT_name [DW_FORM_strp]       ("_dl_catch_exception")
                    DW_AT_decl_file [DW_FORM_data1] ("/usr/src/debug/glibc-2.31-5.1.x86_64/elf/dl-error-skeleton.c")
                    ...

Which is a function that indeed uses a function pointer.

gdb/ChangeLog:

* dwarf2/loc.c (class dwarf_evaluate_loc_desc)
<push_dwarf_reg_entry_value>: Remove assert.  Override
per_objfile with caller_per_objfile.

Change-Id: Ib227d767ce525c10607ab6621a373aaae982c67a

Pass -Wno-deprecated-register for gdb.cp that use "register"

Clang fails to compile three testcases with the following error:
  warning: 'register' storage class specifier is deprecated and
    incompatible with C++17 [-Wdeprecated-register]

This prevents the following testcases from executing:
  gdb.cp/classes.exp
  gdb.cp/inherit.exp
  gdb.cp/misc.exp

This commit builds those testcases with -Wno-deprecated-register, to
avoid the failure.  Note that this commit reveals five "wrong access
specifier for typedef" failures in gdb.cp/classes.exp when compiling
the testsuite with clang.

gdb/testsuite/ChangeLog:

* gdb.cp/classes.exp (prepare_for_testing): Add
additional_flags=-Wno-deprecated-register.
* gdb.cp/inherit.exp (prepare_for_testing): Likewise.
* gdb.cp/misc.exp: Likewise.

[gdb/symtab] Make gold index workaround more precise

There's a PR gold/15646 - "gold-generated .gdb_index has duplicated
symbols that gdb-generated index doesn't", that causes gold to generate
duplicate symbols in the index.

F.i., a namespace N1 declared in a header file can be listed for two CUs that
include the header file:
...
[759] N1:
        2 [global type]
        3 [global type]
...

This causes a gdb performance problem: f.i. when attempting to set a
breakpoint on a non-existing function N1::misspelled, the symtab for both CUs
will be expanded.

Gdb contains a workaround for this, added in commit 8943b87476 "Work around
gold/15646", that skips duplicate global symbols in the index.

However, the workaround does not check for the symbol kind ("type" in the
example above).

Make the workaround more precise by limiting it to symbol kind "type".

Tested on x86_64-linux, with target boards cc-with-gdb-index and
gold-gdb-index.

gdb/ChangeLog:

2020-05-28  Tom de Vries  <tdevries@suse.de>

* dwarf2/read.c (dw2_symtab_iter_next, dw2_expand_marked_cus): Limit
PR gold/15646 workaround to symbol kind "type".

[PATCH] gas: Fix comment on definition of frag_grow()

* frags.c (frag_grow): Fix comment.

Fix "'operator new' should not return NULL" errors in testsuite

When running the testsuite with clang, gdb.linespec/cpls-ops.cc
fails to compile with the following errors:
  warning: 'operator new' should not return a null pointer unless
    it is declared 'throw()' or 'noexcept' [-Wnew-returns-null]
  warning: 'operator new[]' should not return a null pointer unless
    it is declared 'throw()' or 'noexcept' [-Wnew-returns-null]

This prevents the gdb.linespec/cpls-ops.exp testcase from executing.
This commit fixes.

gdb/testsuite/ChangeLog:

* gdb.linespec/cpls-ops.cc (dummy): New static global.
(test_op_new::operator new): Add return statement.
(test_op_new_array::operator new[]): Likewise.

ubsan: nios2: undefined shift

* nios2-dis.c (nios2_print_insn_arg): Avoid shift left of negative
values.

PR26044, Some targets can't be compiled with GCC 10 (tilepro)

Since this value is used in fields of type tilepro_pipeline (as
NO_PIPELINE, see tc-tilepro.c) it is appropriate to put it in
the tilepro_pipelen enum.  This avoids a warning about converting from
one enum type to another.

PR 26044
* opcode/tilepro.h (TILEPRO_NUM_PIPELINE_ENCODINGS): Move to
tilepro_pipeline enum.

asan: ns32k: use of uninitialized value

* ns32k-dis.c (print_insn_arg): Handle d value of 'f' for
immediates.
(print_insn_ns32k): Revert last change.

ld: Mention --enable-textrel-check=yes is default for Linux/x86 targets

* NEWS: Mention --enable-textrel-check=yes is default for
Linux/x86 targets.

ld: Enable --warn-textrel by default for Linux/x86 targets

* configure.tgt (ac_default_ld_textrel_check): Set to if unset
for Linux/x86 targets.

ld: Add --enable-textrel-check=[no|yes|warning|error]

Add a configure option, --enable-textrel-check=[no|yes|warning|error],
to decide what ELF linker should do by default with DT_TEXTREL in an
executable or shared library.

PR ld/20824
* NEWS: Mention --enable-textrel-check=[no|yes|warning|error].
* configure.ac: Add --enable-textrel-check=[no|yes|warning|error].
(DEFAULT_LD_TEXTREL_CHECK): New AC_DEFINE_UNQUOTED.
(DEFAULT_LD_TEXTREL_CHECK_WARNING): Likewise.
* ldmain.c (main): Initialize link_info.textrel_check to
DEFAULT_LD_TEXTREL_CHECK.
* lexsup.c (ld_options): Check DEFAULT_LD_TEXTREL_CHECK_WARNING.
* config.in: Regenerated.
* configure: Likewise.

Fix a potential use of an uninitialised value in the ns32k disassembler.

* ns32k-dis.c (print_insn_ns32k): Change the arg_bufs array to
static.

Linker: Remove support for -Map= with an empty argument.

* lexsup.c (parse_args): Generate an error if a name is not
provided to the -Map option.
(ld_options): Mention that the -Map option supports a directory
name as an argument.
* NEWS: Remove mention of support for an empty string as an
argument to -Map.
* ld.texi: Likewise.

Fix PR gas/26001 (pdp11-*-*)

PR gas/26001
* gas/config/tc-pdp11.c (parse_reg): Distinguish register names from
symbols that begin with a register name.
* gas/testsuite/gas/pdp11/pdp11.exp: Add test of such symbols.
* gas/testsuite/gas/pdp11/pr26001.s: Likewise.
* gas/testsuite/gas/pdp11/pr26001.d: Likewise.

Automatic date update in version.in

RISC-V: Fix missing initialization of riscv_csr_extra structs

The next pointer of struct riscv_csr_extra was not always initilized
to NULL or a valid pointer, causing the assembler to attempt to read
through an uninitialized pointer on startup.

gas/ChangeLog:

        * config/tc-riscv.c (riscv_init_csr_hash): NULL initilize next
        pointer when creating struct riscv_csr_extra.

PR ld/22909 amendment; don't xfail ld-elf/pr19539.d for cris*-*-*.

Noticed by my autotester covering cris-axis-linux-gnu:
...
Running <x>/src/ld/testsuite/ld-elf/elf.exp ...
XPASS: ld-elf/pr19539
...

This was apparently as intended, so I'm just adjusting the xfail.

PR ld/22909
* testsuite/ld-elf/pr19539.d: Don't xfail for cris*-*-*.

cp-completion-aliases.exp: Use test_gdb_complete_{unique,multiple}

gdb.linespec/cp-completion-aliases.exp is calling
test_gdb_complete_{tab,cmd}_unique and
test_gdb_complete_{tab,cmd}_multiple separately for each use case.
I.e., testing once for TAB completion and once for the "complete"
command.  There's no need to do that explicitly and separately, we
have wrapper procedures to do that for us.

gdb/testsuite/ChangeLog:
2020-05-27  Pedro Alves  <palves@redhat.com>

* gdb.linespec/cp-completion-aliases.exp: Remove readline_is_used
check.  Use test_gdb_complete_unique instead of
test_gdb_complete_tab_unique + test_gdb_complete_cmd_unique.  Use
test_gdb_complete_multiple instead of
test_gdb_complete_tab_multiple + test_gdb_complete_cmd_multiple.

Use add_partial_symbol in load_partial_dies

An earlier patch added the add_partial_symbol helper function to
dwarf2/read.c.  However, a couple of calls to add_psymbol_to_list were
left in place.  It turns out that these calls slow down partial symbol
reading, because they still go via the path that tries to needlessly
demangle already-demangled names.

This patch improves the performance of partial symbol reading by
changing this code to use add_partial_symbol instead.

The run previous to this had times of (see the first patch in the
series for an explanation):

gdb    1.64
libxul 1.99
Ada    2.47

This patch improves the times to:

gdb    1.47
libxul 1.89
Ada    2.39

gdb/ChangeLog
2020-05-27  Tom Tromey  <tromey@adacore.com>

* dwarf2/read.c (load_partial_dies): Use add_partial_symbol.

Inline abbrev lookup

Profiling showed that calls to abbrev_table::lookup_abbrev were "too
visible".  As these are just forwarding calls to the hash table, this
patch inlines the lookup.  Also, htab_find_with_hash is used, avoiding
another call.

The run previous to this had times of (see the first patch in the
series for an explanation):

gdb    1.69
libxul 2.02
Ada    2.52

This patch improves the times to:

gdb    1.64
libxul 1.99
Ada    2.47

gdb/ChangeLog
2020-05-27  Tom Tromey  <tromey@adacore.com>

* dwarf2/abbrev.h (struct abbrev_table) <lookup_abbrev>: Inline.
Use htab_find_with_hash.
<add_abbrev>: Remove "abbrev_number" parameter.
* dwarf2/abbrev.c (abbrev_table::add_abbrev): Remove
"abbrev_number" parameter.  Use htab_find_slot_with_hash.
(hash_abbrev): Add comment.
(abbrev_table::lookup_abbrev): Move to header file.
(abbrev_table::read): Update.

Lazily compute partial DIE name

Currently the name of a partial DIE is computed eagerly.  However, the
name is not always needed.  This patch changes partial DIEs to compute
their name lazily, improving performance by avoiding unnecessary name
computations.

The run previous to this had times of (see the first patch in the
series for an explanation):

gdb    1.88
libxul 2.11
Ada    2.60

This patch improves the times to:

gdb    1.69
libxul 2.02
Ada    2.52

gdb/ChangeLog
2020-05-27  Tom Tromey  <tromey@adacore.com>

* dwarf2/read.c (struct partial_die_info) <name>: Declare new
method.
<canonical_name>: New member.
<raw_name>: Rename from "name".
(partial_die_info): Initialize canonical_name.
(scan_partial_symbols): Check raw_name.
(partial_die_parent_scope, partial_die_full_name)
(add_partial_symbol, add_partial_subprogram)
(add_partial_enumeration, load_partial_dies): Use "name" method.
(partial_die_info::name): New method.
(partial_die_info::read, guess_partial_die_structure_name)
(partial_die_info::fixup): Update.

Attribute method inlining

This inlines a couple of methods on struct attribute, improving the
performance of DWARF partial symbol reading.  These methods were
discovered as hot spots using callgrind.

For this patch, and for all the patches in this series, I tested gdb's
performance on three programs:

1. gdb itself -- I built gdb and copied it to /tmp, ensuring that the
   same version was used in all tests.

2. The system libxul.so, the main library of Firefox.  I installed the
   separate debuginfo and ensured that gdb read it.

3. A large-ish Ada program that I happen to have.

I ran gdb 10 times like:

  /bin/time -f %e \
    ./gdb/gdb --data-directory ./gdb/data-directory -nx \
    -iex 'set debug-file-directory /usr/lib/debug' \
    -batch $X

... where $X was the test executable.  Then I computed the mean time.
This was all done with a standard (-g -O2) build of gdb.

The baseline times were

gdb    1.90
libxul 2.12
Ada    2.61

This patch brings the numbers down to

gdb    1.88
libxul 2.11
Ada    2.60

Not a huge change, but still visible in the results.

gdb/ChangeLog
2020-05-27  Tom Tromey  <tromey@adacore.com>

* dwarf2/attribute.h (struct attribute) <form_is_ref>: Inline.
<get_ref_die_offset>: Inline.
<get_ref_die_offset_complaint>: New method.
* dwarf2/attribute.c (attribute::form_is_ref): Move to header.
(attribute::get_ref_die_offset_complaint): Rename from
get_ref_die_offset.  Just issue complaint.

Move exit_status_set_internal_vars out of GLOBAL_CURDIR

Fixes these testsuite fails on Windows:
FAIL: gdb.base/shell.exp: shell success exitcode
FAIL: gdb.base/shell.exp: shell fail exitcode

The convenience variables $_shell_exitcode and $_shell_exitsignal don't
depend on the GLOBAL_CURDIR define.

gdb/ChangeLog:

2020-05-27  Hannes Domani  <ssbssa@yahoo.de>

* cli/cli-cmds.c (shell_escape): Move exit_status_set_internal_vars.

Use errno value of first openp failure

Fixes this testsuite fail on Windows:
FAIL: gdb.base/bad-file.exp: directory

If both tries to open the file fail (without and with ".exe"), use the
errno value of the first try.

gdb/ChangeLog:

2020-05-27  Hannes Domani  <ssbssa@yahoo.de>

* exec.c (exec_file_attach): Use errno value of first openp failure.

Don't close process handle provided by WaitForDebugEvent

Only the process handle returned by OpenProcess or CreateProcess needs to
be closed, the one provided by WaitForDebugEvent is closed automatically.

gdbserver/ChangeLog:

2020-05-27  Hannes Domani  <ssbssa@yahoo.de>

* win32-low.cc (do_initial_child_stuff): Set open_process_used.
(win32_clear_inferiors): Use open_process_used.
(get_child_debug_event): Likewise.

Don't close thread handles provided by WaitForDebugEvent

I sometimes encountered a weird breakpoint failure when using start:

(gdb) start
Temporary breakpoint 2 at 0x40162d: file gdb-25911.c, line 4.
Starting program: C:\src\tests\gdb-25911.exe
Warning:
Cannot insert breakpoint 2.
Cannot access memory at address 0x401628

After trying a lot of combinations, I found a way to reproduce it:

(gdb) file gdb-25987.exe
Reading symbols from gdb-25987.exe...
(gdb) start
Temporary breakpoint 1 at 0x401638: file gdb-25987.cpp, line 13.
Starting program: C:\src\tests\gdb-25987.exe

Temporary breakpoint 1, main () at gdb-25987.cpp:13
13      int main() {
(gdb) c
Continuing.

Program received signal SIGSEGV, Segmentation fault.
MyClass::call (this=0x3d20d0) at gdb-25987.cpp:8
8           *(char*)(nullptr) = 1;
(gdb) kill
Kill the program being debugged? (y or n) y
[Inferior 1 (process 1140) killed]
(gdb) file gdb-25911.exe
Load new symbol table from "gdb-25911.exe"? (y or n) y
Reading symbols from gdb-25911.exe...
(gdb) start
Temporary breakpoint 2 at 0x40162d: file gdb-25911.c, line 4.
Starting program: C:\src\tests\gdb-25911.exe
Warning:
Cannot insert breakpoint 2.
Cannot access memory at address 0x401628

Command aborted.

The actual failure was that ReadProcessMemory used a process handle that
was no longer valid.
And the underlying reason was that the windows_thread_info destructor
closes a thread handle that was provided earlier by WaitForDebugEvent.

But since this is not allowed (and it was actually already closed at this
point, and the handle value reused), this closed another still-needed handle.

gdb/ChangeLog:

2020-05-27  Hannes Domani  <ssbssa@yahoo.de>

* nat/windows-nat.c (windows_thread_info::~windows_thread_info):
Don't close thread handle.

 [PATCH] allow empty string as argument to -Map

* lexsup.c (parse_args): If the map filename is defined but empty
create a name based upon the output file name.  If the name is
defined but refers to a directory create a file inside the
directory based on the output file name.
* ld.texi: Document the new feature.
* testsuite/ld-script/map-address.exp: Add test of new feature.
* NEWS: Mention the new feature.

Share DWARF partial symtabs

This changes the DWARF reader to share partial symtabs (or indices if
they are available) across objfiles.  This has a few parts.

* If multiple objfiles backed by the same BFD can share partial symtabs
  (see below), a single dwarf2_per_bfd is created.  It is stored in the
  per-bfd `dwarf2_per_bfd_bfd_data_key` registry.  Multiple
  dwarf2_per_objfile objects will point to the same instance.  The
  lifetime of these dwarf2_per_bfd objects is naturally handled.  When
  all the objfiles using the BFD are destroyed, the BFD's refount drops
  to 0, which triggers the removal of the corresponding dwarf2_per_bfd
  object from the registry and its destruction.

* If multiple objfiles backed by the same BFD can't share partial
  symtabs (see below), one dwarf2_per_bfd object is created for each
  objfile.  Each dwarf2_per_objfile will point to their own instance of
  dwarf2_per_bfd.  These instances of dwarf2_per_bfd are kept in a
  per-objfile registry, meaning that when the objfile gets destroyed,
  the dwarf2_per_bfd instance gets destroyed as well.

* objfile::partial_symtabs is changed to be a shared_ptr again.  This
  lets us stash a second reference in dwarf2_per_bfd; if the DWARF
  data is being shared, we can simply copy this value to the new
  objfile.

* Two dwarf2_per_objfile objects backed by the same BFD may share a
  dwarf2_per_bfd instance if:

  * No other symbol reader has found symbols, and
  * No BFD section rqeuires relocation

gdb/ChangeLog:

YYYY-MM-DD  Tom Tromey  <tom@tromey.com>
YYYY-MM-DD  Simon Marchi  <simon.marchi@efficios.com>

* objfiles.h (struct objfile) <partial_symtabs>: Now a
shared_ptr.
* dwarf2/read.h (struct dwarf2_per_objfile) <partial_symtabs>: New
member.
* dwarf2/read.c (dwarf2_per_bfd_bfd_data_key,
dwarf2_per_bfd_objfile_data_key>: New globals.
(dwarf2_has_info): Use shared dwarf2_per_bfd if possible.
(dwarf2_get_section_info): Use get_dwarf2_per_objfile.
(dwarf2_initialize_objfile): Consider cases where per_bfd can be
shared.
(dwarf2_build_psymtabs): Set objfile::partial_symtabs and
short-circuit when sharing.
(dwarf2_build_psymtabs): Set dwarf2_per_objfile::partial_symtabs.
(dwarf2_psymtab::expand_psymtab): Use free_cached_comp_units.

Change-Id: I868c64448589102ab8cbb8f06c31a8de50a14004

Move line_header_hash to dwarf2_per_objfile

The `line_header_hash` field of `struct dwarf2_per_bfd` contains some
`struct line_header` objects.  A `struct line_header` objects contains
some `file_entry` objects.  A `file_entry` object contains a pointer to
the `symtab` object created from it.  The `line_header_hash` is
therefore ultimately objfile-dependent and can't be shared as-is between
objfiles.

Move it from `dwarf2_per_bfd` to `dwarf2_per_objfile`.

gdb/ChangeLog:

* dwarf2/read.h (struct dwarf2_per_bfd) <line_header_hash>: Move
to...
(struct dwarf2_per_objfile) <line_header_hash>: ... here.
* dwarf2/read.c (handle_DW_AT_stmt_list): Update.

Change-Id: I8d2ee04df4f4847c2db99061fc976c35af98ac71

Make mapped_debug_names independent of objfile

mapped_debug_names currently has a dwarf2_per_objfile field.  Since we
want it to become objfile-independent, this field must be removed.

This patch removes it, and then arranges for all methods that needed it
to accept a dwarf2_per_objfile parameter.  This trickles down at various
places, like the dw2_debug_names_iterator type.

Ultimately, the objfile only seems to be needed because we might need to
read a string from the string section.  For that, we might need to read
in the section, and if it's a relocatable section, the objfile is needed
in order to do the relocation.  This pattern happens often (that we to
pass an objfile only because a section might be read).  I think it's a
bit ugly, but I don't have a good alternative right now.

gdb/ChangeLog:

* dwarf2/read.c (struct mapped_index_base) <symbol_name_at,
build_name_components, find_name_components_bounds>:
Add per_objfile parameter.
(struct mapped_index) <symbol_name_at>: Likewise.
(struct mapped_debug_names): Remove constructor.
<dwarf2_per_objfile>: Remove field.
<namei_to_name, symbol_name_at>: Add per_objfile parameter.
(mapped_index_base::find_name_components_bounds,
mapped_index_base::build_name_components,
dw2_expand_symtabs_matching_symbol): Likewise.
(class mock_mapped_index) <symbol_name_at>: Likewise.
(check_match): Likewise.
(check_find_bounds_finds): Likewise.
(test_mapped_index_find_name_component_bounds): Update.
(CHECK_MATCH): Update.
(dw2_expand_symtabs_matching): Update.
(class dw2_debug_names_iterator) <dw2_debug_names_iterator>: Add
per_objfile parameter.
<find_vec_in_debug_names>: Likewise.
<m_per_objfile>: New field.
(mapped_debug_names::namei_to_name): Add dwarf2_per_objfile
parameter.
(dw2_debug_names_iterator::find_vec_in_debug_names): Likewise.
(dw2_debug_names_iterator::next): Update.
(dw2_debug_names_lookup_symbol): Update.
(dw2_debug_names_expand_symtabs_for_function): Update.
(dw2_debug_names_map_matching_symbols): Update.
(dw2_debug_names_expand_symtabs_matching): Update.
(dwarf2_read_debug_names): Update.

Change-Id: I00ee0d939390d353442675c7d400a261307c57a1

Replace dwarf2_per_cu_data::cu backlink with per-objfile map

The dwarf2_per_cu_data type is going to become objfile-independent,
while the dwarf2_cu type will stay object-dependent.  This patch removes
the backlink from dwarf2_per_cu_data to dwarf2_cu, in favor of the
dwarf2_per_objfile::m_dwarf2_cus map.  It maps dwarf2_per_cu_data
objects to the corresponding dwarf2_cu objects for this objfile.  If a
CU has been read in in the context of this objfile, then an entry will
be present in the map.

The dwarf2_cu objects that are read in are currently kept in a linked
list rooted in the dwarf2_per_bfd.  Except that the dwarf2_cu objects
are not simply linked together, they are interleaved with their
corresponding dwarf2_per_cu_data objects.  So if we have CUs A and B
read in, the dwarf2_per_bfd::read_in_chain will point to a chain like
this (DPCD == dwarf2_per_cu_data, DC == dwarf2_cu):

 DPCD A -> DC A -> DPCD B -> DC B

Obviously, this can't stay as is, since a same CU can be read in for an
objfile but not read in for another objfile sharing the same BFD, and
the dwarf2_per_cu_data::cu link is removed.   This is all replaced by
the dwarf2_per_objfile::m_dwarf2_cus map.

gdb/ChangeLog:

* dwarf2/read.h (struct dwarf2_cu): Forward-declare.
(struct dwarf2_per_bfd) <free_cached_comp_units>: Remove,
move to dwarf2_per_objfile.
<read_in_chain>: Remove.
(struct dwarf2_per_objfile) <get_cu, set_cu, remove_cu,
remove_all_cus, age_comp_units>: New methods.
<m_dwarf2_cus>: New member.
(struct dwarf2_per_cu_data) <cu>: Remove.
* dwarf2/read.c (struct dwarf2_cu) <read_in_chain>: Remove.
(age_cached_comp_units, free_one_cached_comp_unit): Remove,
moved to methods of dwarf2_per_objfile.
(dwarf2_clear_marks): Remove.
(dwarf2_queue_item::~dwarf2_queue_item): Update.
(dwarf2_per_bfd::~dwarf2_per_bfd): Don't free dwarf2_cus.
(dwarf2_per_bfd::free_cached_comp_units): Remove.
(dwarf2_per_objfile::remove_all_cus): New.
(class free_cached_comp_units) <~free_cached_comp_units>:
Update.
(load_cu): Update.
(dw2_do_instantiate_symtab): Adjust.
(fill_in_sig_entry_from_dwo_entry): Adjust.
(cutu_reader::init_tu_and_read_dwo_dies): Update.
(cutu_reader::cutu_reader): Likewise.
(cutu_reader::keep): Use dwarf2_per_objfile::set_cu.
(cutu_reader::cutu_reader): Use dwarf2_per_objfile::get_cu.
(process_psymtab_comp_unit): Use dwarf2_per_objfile::remove_cu
and dwarf2_per_objfile::age_comp_units.
(load_partial_comp_unit): Update.
(maybe_queue_comp_unit): Use dwarf2_per_objfile::get_cu.
(process_queue): Likewise.
(find_partial_die): Use dwarf2_per_objfile::get_cu instead of cu
backlink.
(dwarf2_read_addr_index): Likewise.
(follow_die_offset): Likewise.
(dwarf2_fetch_die_loc_sect_off): Likewise.
(dwarf2_fetch_constant_bytes): Likewise.
(dwarf2_fetch_die_type_sect_off): Likewise.
(follow_die_sig_1): Likewise.
(load_full_type_unit): Likewise.
(read_signatured_type): Likewise.
(dwarf2_cu::dwarf2_cu): Don't set cu field.
(dwarf2_cu::~dwarf2_cu): Remove.
(dwarf2_per_objfile::get_cu): New.
(dwarf2_per_objfile::set_cu): New.
(age_cached_comp_units): Rename to...
(dwarf2_per_objfile::age_comp_units): ... this.  Adjust
to std::unordered_map.
(free_one_cached_comp_unit): Rename to...
(dwarf2_per_objfile::remove_cu): ... this.  Adjust
to std::unordered_map.
(dwarf2_per_objfile::~dwarf2_per_objfile): New.
(dwarf2_mark_helper): Use dwarf2_per_objfile::get_cu, expect
a dwarf2_per_objfile in data.
(dwarf2_mark): Pass dwarf2_per_objfile in data to htab_traverse.
(dwarf2_clear_marks): Remove.

Change-Id: Ia33ac71c79b2de4710569008e22a6563a1505cde

Pass existing_cu object to cutu_reader

It is possible, seemingly for a special case described in
find_partial_die, for cutu_reader to re-use an existing dwarf2_cu
instead of creating a new one.  This happens when running this test, for
example:

    make check TESTS="gdb.dwarf2/fission-reread.exp"

Right now the, `use_existing_cu` flag tells cutu_reader to use the
dwarf2_cu object at dwarf2_per_cu_data::cu.  However, we'll remove that
field, so we need to find another solution.

This situation arises when some caller up the stack has already created
the dwarf2_cu to read a dwarf2_per_cu_data, but needs to re-read it with
some other parameters.  Therefore, it's possible to just have that
caller pass down the dwarf2_cu object to use as a `existing_cu`
parameter.  If `existing_cu` is NULL, it tells cutu_reader that it needs
to instantiate a new one.

gdb/ChangeLog:

* dwarf2/read.c (class cutu_reader) <cutu_reader>: Replace
`int use_existing_cu` parameter with `dwarf2_cu *existing_cu`.
(init_tu_and_read_dwo_dies): Likewise.
(cutu_reader::init_tu_and_read_dwo_dies): Likewise.
(cutu_reader::cutu_reader): Likewise.
(load_partial_comp_unit): Likewise.
(process_psymtab_comp_unit): Update.
(build_type_psymtabs_1): Update.
(process_skeletonless_type_unit): Update.
(load_full_comp_unit): Update.
(find_partial_die): Update.
(dwarf2_read_addr_index): Update.
(read_signatured_type): Update.

Change-Id: Id03e3bc3de3cf99d9e4b4080ad83b029c93bf434

Add comp_unit_head to dwarf2_per_cu_data

The per_cu_header_read_in function allows obtaining a filled
comp_unit_head object for a given dwarf2_per_cu_data object.  If a
dwarf2_cu object exists for this dwarf2_per_cu_data, then it just
returns a pointer to the comp_unit_head from that dwarf2_cu.  Otherwise,
it reads the header into a temporary buffer provided by the caller, and
returns a pointer to that.

Since the dwarf2_per_cu_data::cu link is going to be removed
(dwarf2_per_cu_data will become objfile-independent while dwarf2_cu
stays objfile-dependent), we cannot rely anymore on returning the header
from the dwarf2_cu object.

The not too complex solution implemented by this patch is to keep a copy
of the header in the dwarf2_per_cu_data object, independent from the
copy in dwarf2_cu.  The new copy is only used in the addr_size,
offset_size and ref_addr_size methods of dwarf2_per_cu_data.

There's nothing intrinsic to the comp_unit_head object that prevents it
to be shared between two dwarf2_cu objects (belonging to different
objfiles) representing the same CU.  In other words, I think we could
eventually get rid of the copy in dwarf2_cu to only keep the one in
dwarf2_per_cu_data.  It is not trivial, however, so I have decided not
to do it for the moment.

gdb/ChangeLog:

* dwarf2/read.h (struct dwarf2_per_cu_data) <m_header,
m_header_read_in>: New fields.
<get_header>: New method.
* dwarf2/read.c (per_cu_header_read_in): Remove.
(dwarf2_per_cu_data::get_header): New.
(dwarf2_per_cu_data::addr_size): Update.
(dwarf2_per_cu_data::offset_size): Update.
(dwarf2_per_cu_data::ref_addr_size): Update.

Change-Id: Id7541fca7562843eba110ece21c4df38d45fca23

Make load_cu return the loaded dwarf2_cu

In a subsequent patch, the dwarf2_per_cu_data::cu link will be removed.
dwarf2_cu objects will instead need to be looked up from a per-objfile
map, using the dwarf2_per_cu_data object as the key.

To make it easier for some callers, this patch makes load_cu return the
dwarf2_cu it creates.  If the caller needs to use the created dwarf2_cu,
it will have it available right away, rather than having to do a map
lookup.

At the same time, this allows changing queue_and_load_all_dwo_tus to
take a dwarf2_cu instead of a dwarf2_per_cu_data.

gdb/ChangeLog:

* dwarf2/read.c (load_cu): Return dwarf2_cu.
(dw2_do_instantiate_symtab): Update.
(queue_and_load_all_dwo_tus): Change parameter from
dwarf2_per_cu_data to dwarf2_cu.
(dwarf2_fetch_die_loc_sect_off): Update.
(dwarf2_fetch_constant_bytes): Update.
(dwarf2_fetch_die_type_sect_off): Update.

Change-Id: I8a04c5d1b8cc661b8203f97999258ba8e04e1765

Pass dwarf2_cu to process_full_{comp,type}_unit

These two functions work on a dwarf2_cu.  It is currently obtained from
the per_cu->cu link, which we want to remove.  Make them accept the
dwarf2_cu directly as a parameter.  This moves the per_cu->cu references
one level up, but that one will be removed too in a subsequent patch.

gdb/ChangeLog:

* dwarf2/read.c (process_full_comp_unit,
process_full_type_unit): Remove per_cu, per_objfile paramters.
Add dwarf2_cu parameter.
(process_queue): Update.

Change-Id: I1027d36986073ac991e198e06f9d51341dc19c6e

Pass dwarf2_per_bfd instead of dwarf2_per_objfile to some index-related functions

All these functions actually only need to receive a dwarf2_per_bfd, pass
that instead of dwarf2_per_objfile.

gdb/ChangeLog:

* dwarf2/read.c (create_cu_from_index_list): Replace
dwarf2_per_objfile parameter with dwarf2_per_bfd.
(create_cus_from_index_list): Likewise.
(create_cus_from_index): Likewise.
(create_signatured_type_table_from_index): Likewise.
(create_cus_from_debug_names_list): Likewise.
(create_cus_from_debug_names): Likewise.
(dwarf2_read_gdb_index): Update.
(dwarf2_read_debug_names): Update.

Change-Id: I8cd7dc04bf815723a48745e7e9b283663dccc1ac