binutils_os216

Fork

Add a fake bfd to hold linker created dynamic sections

Currently, we hold linker created dynamic sections in an input shared
object, which has its own dynamic sections, when the first input file
from linker is a shared object. It may lead to conflicts between
linker created dynamic sections and shared object's dynamic sections.
We can use a a fake bfd to hold linker created dynamic sections.
Unfortunately, it doesn't work due to BFD_LINKER_CREATED. Dynamic
sections in bfd with BFD_LINKER_CREATED may be ignored.

Exclude linker created file from dynobj

Some ELF targets create a "linker stubs" fake bfd. Don't use it to
set dynobj.

* elflink.c (_bfd_elf_link_create_dynstrtab): Exclude linker
created file from dynobj.

Set dynobj to a normal input file if possible

When check_relocs is called after gc-sections has run,
_bfd_elf_link_create_dynstrtab may be called with an dynamic object
and hash_table->dynobj may be NULL. We may not set dynobj, an input
file holding linker created dynamic sections to the dynamic object,
which has its own dynamic sections. We need to find a normal input
file to hold linker created sections if possible. Otherwise ld will
crash during LTO input rescan when linker created dynamic section
overrides input dynamic section.

* elflink.c (_bfd_elf_link_create_dynstrtab): Set dynobj to a
normal input file if possible.

MIPS/GAS: Fix an ISA override not lifting ABI restrictions

Correct a regression introduced with commit 919731affbef ("Add MIPS
.module directive") causing code like:

.set mips3
dli $2, 0x9000000080000000

to fail assembly with the following error message produced:

Error: number (0x9000000080000000) larger than 32 bits

if built with `mips3' selected as the global ISA (e.g. `-march=mips3').
This is because a `.set' directive doing an ISA override does not lift
the ABI restriction on register sizes if the ISA remains unchanged.
Previously the directive always set register sizes from the ISA chosen,
which is what some code expects. Restore the old semantics then.

gas/
* config/tc-mips.c (code_option_type): New enum.
(parse_code_option): Return status indicating option type.
(s_mipsset): Update `parse_code_option' call site accordingly.
Always set register sizes from the ISA with ISA overrides.
(s_module): Update `parse_code_option' call site.
* testsuite/gas/mips/isa-override-1.d: New test.
* testsuite/gas/mips/micromips@isa-override-1.d: New test.
* testsuite/gas/mips/mips1@isa-override-1.d: New test.
* testsuite/gas/mips/mips2@isa-override-1.d: New test.
* testsuite/gas/mips/mips32@isa-override-1.d: New test.
* testsuite/gas/mips/mips32r2@isa-override-1.d: New test.
* testsuite/gas/mips/mips32r3@isa-override-1.d: New test.
* testsuite/gas/mips/mips32r5@isa-override-1.d: New test.
* testsuite/gas/mips/mips32r6@isa-override-1.d: New test.
* testsuite/gas/mips/mips64r2@isa-override-1.d: New test.
* testsuite/gas/mips/mips64r3@isa-override-1.d: New test.
* testsuite/gas/mips/mips64r5@isa-override-1.d: New test.
* testsuite/gas/mips/mips64r6@isa-override-1.d: New test.
* testsuite/gas/mips/r3000@isa-override-1.d: New test.
* testsuite/gas/mips/r3900@isa-override-1.d: New test.
* testsuite/gas/mips/r5900@isa-override-1.d: New test.
* testsuite/gas/mips/octeon@isa-override-1.d: New test.
* testsuite/gas/mips/octeon3@isa-override-1.d: New test.
* testsuite/gas/mips/isa-override-2.l: New list test.
* testsuite/gas/mips/mips1@isa-override-2.l: New list test.
* testsuite/gas/mips/mips2@isa-override-2.l: New list test.
* testsuite/gas/mips/mips32@isa-override-2.l: New list test.
* testsuite/gas/mips/mips32r2@isa-override-2.l: New list test.
* testsuite/gas/mips/mips32r3@isa-override-2.l: New list test.
* testsuite/gas/mips/mips32r5@isa-override-2.l: New list test.
* testsuite/gas/mips/mips32r6@isa-override-2.l: New list test.
* testsuite/gas/mips/r3000@isa-override-2.l: New list test.
* testsuite/gas/mips/r3900@isa-override-2.l: New list test.
* testsuite/gas/mips/octeon3@isa-override-2.l: New list test.
* testsuite/gas/mips/octeon3@isa-override-1.l: New stderr
output.
* testsuite/gas/mips/isa-override-1.s: New test source.
* testsuite/gas/mips/r5900@isa-override-1.s: New test source.
* testsuite/gas/mips/isa-override-2.s: New test source.
* testsuite/gas/mips/mips1@isa-override-2.s: New test source.
* testsuite/gas/mips/mips2@isa-override-2.s: New test source.
* testsuite/gas/mips/mips32@isa-override-2.s: New test source.
* testsuite/gas/mips/mips32r2@isa-override-2.s: New test source.
* testsuite/gas/mips/mips32r3@isa-override-2.s: New test source.
* testsuite/gas/mips/mips32r5@isa-override-2.s: New test source.
* testsuite/gas/mips/mips32r6@isa-override-2.s: New test source.
* testsuite/gas/mips/r3000@isa-override-2.s: New test source.
* testsuite/gas/mips/r3900@isa-override-2.s: New test source.
* testsuite/gas/mips/octeon3@isa-override-2.s: New test source.
* testsuite/gas/mips/mips.exp: Run the new tests.

MIPS: Go back with the default Linux # of registers to 90

Set the number of registers for non-XML-described Linux targets to 90,
reverting a change made here with the addition of DSP register support:

commit 1faeff088bbbd037d7769d214378b4faf805fa2e
Author: Maciej W. Rozycki <macro@linux-mips.org>
Date: Thu Mar 1 22:19:48 2012 +0000

and fixing a regression introduced for legacy `gdbserver' targets
causing a "Remote 'g' packet reply is too long" error message where the
amount of register data received with a `g' packet (90) exceeds the
maximum number of registers expected (79).

Update the setting for XML-described targets, reflecting the actual
number of registers which have been assigned numbers, matching the:

gdb_assert (gdbarch_num_regs (gdbarch) <= MIPS_RESTART_REGNUM);

requirement in `mips_linux_init_abi'.

gdb/
* mips-tdep.c (mips_gdbarch_init): For GDB_OSABI_LINUX set
`num_regs' to 90 rather than 79. Where a target description is
present adjust the setting appropriately.

Automatic date update in version.in

Switch gdb's TRY/CATCH to sjlj again

We don't currently handle the case of gdb's readline callback throwing
gdb C++ exceptions across a readline that wasn't built with
-fexceptions. The end result is:

(gdb) whatever-command-that-causes-an-error
terminate called after throwing an instance of 'gdb_exception_RETURN_MASK_ERROR'
Aborted
$

Until that is fixed, revert back to sjlj-based exceptions again.

gdb/ChangeLog:
2016-04-21 Pedro Alves <palves@redhat.com>

* common/common-exceptions.h (GDB_XCPT_TRY): Add comment.
(GDB_XCPT): Always define as GDB_XCPT_SJMP.

Always run LTO tests on Linux with GCC 4.9 or newer

Check for LTO availability will hide LTO bugs in ld. Since GCC 4.9 adds
-ffat-lto-objects, we always run LTO tests on Linux with GCC 4.9 or newer.

* testsuite/lib/ld-lib.exp (check_lto_available): Return 1 on
Linux with GCC 4.9 or newer.
(check_lto_fat_available): Likewise.
(check_lto_shared_available): Likewise.

Add support for non-ELF targets to check their relocs.

bfd * aout-adobe.c: Use _bfd_generic_link_check_relocs.
* aout-target.h: Likewise.
* aout-tic30.c: Likewise.
* binary.c: Likewise.
* bout.c: Likewise.
* coff-alpha.c: Likewise.
* coff-rs6000.c: Likewise.
* coff64-rs6000.c: Likewise.
* coffcode.h: Likewise.
* i386msdos.c: Likewise.
* i386os9k.c: Likewise.
* ieee.c: Likewise.
* ihex.c: Likewise.
* libbfd-in.h: Likewise.
* libecoff.h: Likewise.
* mach-o-target.c: Likewise.
* mmo.c: Likewise.
* nlm-target.h: Likewise.
* oasys.c: Likewise.
* pef.c: Likewise.
* plugin.c: Likewise.
* ppcboot.c: Likewise.
* som.c: Likewise.
* srec.c: Likewise.
* tekhex.c: Likewise.
* versados.c: Likewise.
* vms-alpha.c: Likewise.
* xsym.c: Likewise.
* elfxx-target.h: Use _bfd_elf_link_check_relocs.
* linker.c (bfd_link_check_relocs): New function.
(_bfd_generic_link_check_relocs): New function.
* targets.c (BFD_JUMP_TABLE_LINK): Add initialization of
_bfd_link_check_relocs field.
(struct bfd_target)L Add _bfd_link_check_relocs field.
* bfd-in2.h: Regenerate.
* libbfd.h: Regenerate.

ld * ldlang.c (lang_check_relocs): Use bfd_link_check_relocs in
prefernce to _bfd_elf_link_check_relocs. Drop test for ELF
targets. Do not stop the checks when problems are encountered.

include * bfdlink.h: Add prototype for bfd_link_check_relocs.

Fix AIX gdb build with C++ compiler

We currently get:

../../src/gdb/aix-thread.c: In function 'int pdc_read_data(pthdb_user_t, void*, pthdb_addr_t, size_t)':
../../src/gdb/aix-thread.c:465:46: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
status = target_read_memory (addr, buf, len);
^


../../src/gdb/aix-thread.c: In function 'void aix_thread_resume(target_ops*, ptid_t, int, gdb_signal)':
../../src/gdb/aix-thread.c:1010:46: error: invalid conversion from 'void*' to 'int*' [-fpermissive]
gdb_signal_to_host (sig), (void *) tid);
^
../../src/gdb/aix-thread.c:243:1: error: initializing argument 5 of 'int ptrace64aix(int, int, long long int, int, int*)' [-fpermissive]
ptrace64aix (int req, int id, long long addr, int data, int *buf)


../../src/gdb/rs6000-nat.c: In function 'gdb_byte* rs6000_ptrace_ldinfo(ptid_t)':
../../src/gdb/rs6000-nat.c:596:36: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
gdb_byte *ldi = xmalloc (ldi_size);
^
../../src/gdb/rs6000-nat.c:615:36: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
ldi = xrealloc (ldi, ldi_size);
^

(and more instances of the same).

gdb/ChangeLog:
2016-04-21 Pedro Alves <palves@redhat.com>

* aix-thread.c (pdc_read_data, pdc_write_data): Add cast.
(aix_thread_resume): Use PTRACE_TYPE_ARG5.
* rs6000-nat.c (rs6000_ptrace64): Use PTRACE_TYPE_ARG5.
(rs6000_ptrace_ldinfo): Change type of 'ldi' local to void
pointer, and cast return to gdb_byte pointer.

Fix s390 GNU/Linux gdb and gdbserver builds

Now that gdb/gdbserver compile as C++ programs by default, the s390
GNU/Linux build started failing with:

In file included from ../../src/gdb/common/common-defs.h:64:0,
from ../../src/gdb/defs.h:28,
from ../../src/gdb/s390-linux-nat.c:22:
../../src/gdb/s390-linux-nat.c: In function ‘void fetch_regset(regcache*, int, int, int, const regset*)’:
../../src/gdb/../include/libiberty.h:711:38: error: invalid conversion from ‘void*’ to ‘gdb_byte* {aka unsigned char*}’ [-fpermissive]
# define alloca(x) __builtin_alloca(x)
^
../../src/gdb/s390-linux-nat.c:297:19: note: in expansion of macro ‘alloca’
gdb_byte *buf = alloca (regsize);
^

etc.

gdb/ChangeLog:
2016-04-21 Pedro Alves <palves@redhat.com>

* s390-linux-nat.c (fetch_regset, store_regset, check_regset): Use
void * instead of gdb_byte *.

gdb/gdbserver/ChangeLog:
2016-04-21 Pedro Alves <palves@redhat.com>

* linux-s390-low.c (s390_collect_ptrace_register)
(s390_supply_ptrace_register, s390_get_hwcap): Use gdb_byte * and
add casts.
(s390_check_regset): Use void * instead of gdb_byte *.

Add missing sentinel 'char *' casts in concat/reconcat calls

The wildebeest-debian-wheezy-i686 buildslave's build is broken due to:

../../binutils-gdb/gdb/python/python.c: In function void _initialize_python():
../../binutils-gdb/gdb/python/python.c:1709:36: error: missing sentinel in function call [-Werror=format]

Reproduced on Fedora 23 by sticking a few:

#undef NULL
#define 0

in build/gdb/build-gnulib/{stddef|signal|stdio}.h. Hopefully this
caught all instances.

gdb/ChangeLog:
2016-04-21 Pedro Alves <palves@redhat.com>

* dwarf2read.c (try_open_dwop_file, open_dwo_file)
(file_file_name, file_full_name): Add char * cast to sentinel in
concat/reconcat calls.
* event-top.c (top_level_prompt): Likewise.
* guile/guile.c (initialize_scheme_side): Likewise.
* linux-tdep.c (linux_fill_prpsinfo): Likewise.
* macrotab.c (macro_source_fullname): Likewise.
* main.c (get_init_files, captured_main): Likewise.
* psymtab.c (psymtab_to_fullname): Likewise.
* python/python.c (_initialize_python)
(gdbpy_finish_initialization): Likewise.
* source.c (symtab_to_fullname): Likewise.

New NOCROSSREFS_TO tests

Fixes failures on hppa-linux and alpha-linux due to not merging
.data.* and .sdata into .data. cross3.t modified too since it is the
template for the NOCROSSREFS_TO scripts.

* testsuite/ld-scripts/cross3.t: Add commonly used data
and text section names to output section statements.
* testsuite/ld-scripts/cross4.t: Likewise.
* testsuite/ld-scripts/cross5.t: Likewise.
* testsuite/ld-scripts/cross6.t: Likewise.
* testsuite/ld-scripts/cross7.t: Likewise.

Remove x86 gc_sweep_hook

Since x86 backends never see the removed sections, there is no need
for gc_sweep_hook.

* elf32-i386.c (elf_i386_gc_sweep_hook): Removed.
(elf_backend_gc_sweep_hook): Likewise.
* elf64-x86-64.c (elf_x86_64_gc_sweep_hook): Likewise.
(elf_backend_gc_sweep_hook): Likewise.

Don't check relocations in excluded sections

When checking relocations after gc-sections has run, the unused sections
have been removed. Don't check relocations in excluded sections.

* elflink.c (_bfd_elf_link_check_relocs): Don't check relocations
in excluded sections

Move ELF relocation check after lang_gc_sections

Move ELF relocation check after lang_gc_sections so that all the
reference counting code for plt and got relocs can be removed. This
only affects ELF targets which check relocations after opening all
input file.

* ldlang.c (lang_check_relocs): New function.
(lang_process): Call lang_check_relocs after lang_gc_sections.
* emultempl/elf32.em (gld${EMULATION_NAME}_before_parse): Don't
call _bfd_elf_link_check_relocs here.

Automatic date update in version.in

Build GDB as a C++ program by default

This makes --enable-build-with-cxx be "yes" by default.

One must now configure with --enable-build-with-cxx=no in order to
build with a C compiler.

gdb/ChangeLog:
2016-04-20 Pedro Alves <palves@redhat.com>

* build-with-cxx.m4 (GDB_AC_BUILD_WITH_CXX): Default to yes.
* configure: Renegerate.

gdb/gdbserver/ChangeLog:
2016-04-20 Pedro Alves <palves@redhat.com>

* configure: Renegerate.

Fix host signal vs gdb signal mixup in gdb/darwin-nat.c

Building in C++ mode caught a bug here:

.../src/gdb/darwin-nat.c: In function 'ptid_t darwin_decode_message(mach_msg_header_t*, darwin_thread_t**, inferior**, target_waitstatus*)':
.../src/gdb/darwin-nat.c:1016:25: error: invalid conversion from 'int' to 'gdb_signal' [-fpermissive]
status->value.sig = WTERMSIG (wstatus);
^

gdb/ChangeLog:
2016-04-20 Pedro Alves <palves@redhat.com>

* darwin-nat.c (darwin_decode_message): Use gdb_signal_from_host.

Fix "incompatible pointer type" warning in gdb/aarch64-tdep.c

Fixes, with x86_64-apple-darwin15-gcc (gcc 5.3.0):

.../src/gdb/aarch64-tdep.c: In function 'aarch64_record_load_store':
.../src/gdb/aarch64-tdep.c:3479:67: error: passing argument 3 of 'regcache_raw_read_unsigned' from incompatible pointer type [-Werror=incompatible-pointer-types]
bits (aarch64_insn_r->aarch64_insn, 16, 20), &reg_rm_val);
^
In file included from .../src/gdb/regcache.h:23:0,
from .../src/gdb/gdbarch.h:69,
from .../src/gdb/defs.h:620,
from .../src/gdb/aarch64-tdep.c:21:
.../src/gdb/common/common-regcache.h:60:29: note: expected 'ULONGEST * {aka long unsigned int *}' but argument is of type 'uint64_t * {aka long long unsigned int *}'
extern enum register_status regcache_raw_read_unsigned
^

gdb/ChangeLog:
2016-04-20 Pedro Alves <palves@redhat.com>

* aarch64-tdep.c (aarch64_record_load_store): Change type of
'reg_rm_val' local to ULONGEST.

gdb/darwin-nat.c: Fix "cast to pointer from integer of different size" warning

Fixes, with gcc 5.3.0:

.../src/gdb/darwin-nat.c: In function 'void darwin_resume_thread(inferior*, darwin_thread_t*, int, int)':
.../src/gdb/darwin-nat.c:731:21: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
(caddr_t)thread->gdb_port, nsignal);
^
.../src/gdb/darwin-nat.c:84:35: note: in definition of macro 'PTRACE'
darwin_ptrace(#CMD, CMD, (PID), (ADDR), (SIG))
^

thread->gdb_port is an unsigned int, caddr_t is a void pointer.

gdb/ChangeLog:
2016-04-20 Pedro Alves <palves@redhat.com>

* darwin-nat.c (darwin_resume_thread): Add uintptr_t cast.

Check R_386_NONE/R_X86_64_NONE in eh6.d

There is no need for run-time relocation when converting pointers
in .eh_frame section to DW_EH_PE_pcrel encoding. R_386_NONE and
R_X86_64_NONE are expected since the space for run-time relocation
has been allocated. This is an optimization.

PR ld/19972
* testsuite/ld-elf/eh6.d: Pass -rW to readelf and check for
R_386_NONE or R_X86_64_NONE.

Add ld-x86-64/pic1 test

* testsuite/ld-x86-64/pic1.d: New file.
* testsuite/ld-x86-64/pic1.s: Likewise.
* testsuite/ld-x86-64/x86-64.exp: Run pic1.

Add ld-x86-64/pie2 test

* testsuite/ld-x86-64/pie2.d: New file.
* testsuite/ld-x86-64/pie2.s: Likewise.
* testsuite/ld-x86-64/x86-64.exp: Run pie2.

Check run-time R_X86_64_32 relocation overflow

Since elf_x86_64_check_relocs is called after opening all input files,
we can detect dynamic R_X86_64_32 relocation overflow there.

bfd/

PR ld/19969
* elf64-x86-64.c (check_relocs_failed): New.
(elf_x86_64_need_pic): Moved before elf_x86_64_check_relocs.
Support relocation agaist local symbol. Set check_relocs_failed.
(elf_x86_64_check_relocs): Use elf_x86_64_need_pic. Check
R_X86_64_32 relocation overflow.
(elf_x86_64_relocate_section): Skip if check_relocs failed.
Update one elf_x86_64_need_pic and remove one elf_x86_64_need_pic.

ld/

PR ld/19969
* testsuite/ld-x86-64/pr19969.d: New file.
* testsuite/ld-x86-64/pr19969a.S: Likewise.
* testsuite/ld-x86-64/pr19969b.S: Likewise.
* testsuite/ld-x86-64/x86-64.exp: Run pr19969 tests.

symmisc.c (dump_symtab_1): Print owning compunit for identical blockvectors.

* symmisc.c (dump_symtab_1): Print owning compunit for identical
blockvectors.

Include arch/arm-linux.h in aarch32-linux-nat.c

Fix the compilation failure by including arch/arm-linux.h in
aarch32-linux-nat.c.

gdb:

2016-04-20 Yao Qi <yao.qi@linaro.org>

* aarch32-linux-nat.c: Include "arch/arm-linux.h".

Call _bfd_elf_create_ifunc_sections only for ifunc

Since x86 check_relocs is called after opening all input files, we
need to call _bfd_elf_create_ifunc_sections only for STT_GNU_IFUNC
symbols.

* elf32-i386.c (elf_i386_check_relocs): Call
_bfd_elf_create_ifunc_sections only for STT_GNU_IFUNC symbol.
* elf64-x86-64.c (elf_x86_64_check_relocs): Likewise.

Check ELF relocs after opening all input files

Delaying checking ELF relocations until opening all input files so
that symbol information is final when relocations are checked. This
is only enabled for x86 targets.

bfd/

* elf-bfd.h (_bfd_elf_link_check_relocs): New.
* elflink.c (_bfd_elf_link_check_relocs): New function.
(elf_link_add_object_symbols): Call _bfd_elf_link_check_relocs
if check_relocs_after_open_input is FALSE.

include/

* bfdlink.h (bfd_link_info): Add check_relocs_after_open_input.

ld/

* emulparams/elf32_x86_64.sh (CHECK_RELOCS_AFTER_OPEN_INPUT):
New.
* emulparams/elf_i386.sh (CHECK_RELOCS_AFTER_OPEN_INPUT):
Likewise.
* emulparams/elf_i386_be.sh (CHECK_RELOCS_AFTER_OPEN_INPUT):
Likewise.
* emulparams/elf_i386_chaos.sh (CHECK_RELOCS_AFTER_OPEN_INPUT):
Likewise.
* emulparams/elf_i386_ldso.sh (CHECK_RELOCS_AFTER_OPEN_INPUT):
Likewise.
* emulparams/elf_i386_vxworks.sh (CHECK_RELOCS_AFTER_OPEN_INPUT):
Likewise.
* emulparams/elf_x86_64.sh (CHECK_RELOCS_AFTER_OPEN_INPUT):
Likewise.
* emulparams/i386nto.sh (CHECK_RELOCS_AFTER_OPEN_INPUT):
Likewise.
* emultempl/elf32.em (gld${EMULATION_NAME}_before_parse):
Set check_relocs_after_open_input to TRUE if
CHECK_RELOCS_AFTER_OPEN_INPUT is yes.
(gld${EMULATION_NAME}_after_open): Call
_bfd_elf_link_check_relocs on all inputs if
check_relocs_after_open_input is TRUE.

Move ARM_CPSR_GREGNUM to arch/arm-linux.h

This patch moves macro ARM_CPSR_GREGNUM to arch/arm-linux.h so that it
can be used in GDBserver side.

gdb:

2016-04-20 Yao Qi <yao.qi@linaro.org>

* arm-linux-tdep.h (ARM_CPSR_GREGNUM): Move it to ...
* arch/arm-linux.h: ... here.

gdb/gdbserver:

2016-04-20 Yao Qi <yao.qi@linaro.org>

* linux-aarch32-low.c: Include "arch/arm-linux.h".
(arm_fill_gregset): Use ARM_CPSR_GREGNUM rather than magic
number 16.
(arm_store_gregset): Likewise.