binutils-gdb.git
7 years agoX86: Ignore REX_B bit for 32-bit XOP instructions
Amit Pawar [Mon, 28 Nov 2016 17:21:05 +0000 (09:21 -0800)]
X86: Ignore REX_B bit for 32-bit XOP instructions

While decoding 32-bit XOP instructions, 64 bit registers names are printed.
This patch fixes this by ignoring REX_B bit in 32-bit mode.

opcodes/

PR binutils/20637
* i386-dis.c (get_valid_dis386): Ignore REX_B for 32-bit XOP
instructions.

gas/

PR binutils/20637
* testsuite/gas/i386/xop32reg.d: New file.
* testsuite/gas/i386/xop32reg.s: New file.
* testsuite/gas/i386/i386.exp: Run new test.

7 years agoRestrict checking value.lval on using address
Yao Qi [Mon, 28 Nov 2016 17:09:26 +0000 (17:09 +0000)]
Restrict checking value.lval on using address

With the previous change, value.location.address is only valid for
lval_memory.  This patch restrict some checking on value.lval on
using address.  Since we have a check on VALUE_VAL in
set_value_address, we need to set VALUE_VAL properly before
set_value_address too.

gdb:

2016-11-25  Yao Qi  <yao.qi@linaro.org>

* ada-lang.c (ensure_lval): Call set_value_address after setting
VALUE_LVAL.
* elfread.c (elf_gnu_ifunc_resolve_addr): Set VALUE_LVAL to
lval_memory.
(elf_gnu_ifunc_resolver_return_stop): Likewise.
* value.c (value_fn_field): Likewise.
(value_from_contents_and_address_unresolved): Likewise.
(value_from_contents_and_address): Likewise.
(value_address): Check value->lval isn't
lval_memory.
(value_raw_address): Likewise.
(set_value_address): Assert value->lval is lval_memory.

7 years agoAdjust Value.location for lval_register
Yao Qi [Mon, 28 Nov 2016 17:09:26 +0000 (17:09 +0000)]
Adjust Value.location for lval_register

value.regnum and value.next_frame_id are only used for lval_register,
so this patch moves them to union value.location.  As a result, when
we copy value, only copy location, don't need to copy regnum and
next_frame_id.

This patch also changes regnum's type to int as there is no space
constraint, so update deprecated_value_regnum_hack return type too.

gdb:

2016-11-28  Yao Qi  <yao.qi@linaro.org>

* valops.c (value_slice): Don't set frame id of slice.
* value.c (struct value) <regnum, next_frame_id>: Move them to...
(struct value) <location>: ... here.  Update comments.
(allocate_value_lazy): Don't set frame id and regnum.
(deprecated_value_next_frame_id_hack): Adjust.
(deprecated_value_regnum_hack): Adjust.
(value_copy): Don't copy frame id and regnu.
(value_primitive_field): Likewise.
(value_from_component): Likewise.
(deprecated_value_regnum_hack): Return int *.
* value.h (deprecated_value_regnum_hack): Update declaration.

7 years agoMove computed value's frame id to piece_closure
Yao Qi [Mon, 28 Nov 2016 17:09:26 +0000 (17:09 +0000)]
Move computed value's frame id to piece_closure

Nowadays, we set computed value's frame id, which is a misuse to me.
The computed value itself doesn't care about frame id, but function
value_computed_funcs (val)->read (or read_pieced_value) cares about
which frame the register is relative to, so 'struct piece_closure' is
a better place to fit frame id.

This patch adds a frame id in 'struct piece_closure', and use it
instead of using computed value's frame id.

gdb:

2016-11-28  Yao Qi  <yao.qi@linaro.org>

* dwarf2loc.c (struct piece_closure) <frame_id>: New field.
(allocate_piece_closure): Add new parameter 'frame' and set
closure's frame_id field accordingly.
(read_pieced_value): Get frame from closure instead of value.
(dwarf2_evaluate_loc_desc_full): Remove code getting frame id.
Don't set value's frame id.

7 years agoProperly hide hidden versioned symbol in executable
H.J. Lu [Mon, 28 Nov 2016 16:03:46 +0000 (08:03 -0800)]
Properly hide hidden versioned symbol in executable

A hidden versioned symbol in executable should be forced local if it is
locally defined, not referenced by shared library and not exported.  We
must do it before _bfd_elf_link_renumber_dynsyms.

bfd/

* elflink.c (_bfd_elf_fix_symbol_flags): Hide hidden versioned
symbol in executable.
(elf_link_output_extsym): Don't change bind from global to
local when linking executable.

ld/

* testsuite/ld-elf/indirect.exp: Add a test for PR 18720.
* testsuite/ld-elf/pr18720.rd: New file.

7 years agoAutomatic date update in version.in
GDB Administrator [Mon, 28 Nov 2016 00:00:22 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoFix powerpc vle test for sorting of program headers
Alan Modra [Sun, 27 Nov 2016 09:37:08 +0000 (20:07 +1030)]
Fix powerpc vle test for sorting of program headers

1a9ccd70f changed the order of headers.

PR 20815
* testsuite/ld-powerpc/vle-multiseg-5.d: Update.

7 years agoFix spelling in comments in .y files (ld)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:54 +0000 (21:01 +0100)]
Fix spelling in comments in .y files (ld)

* deffilep.y: Fix spelling in comments.

7 years agoFix spelling in comments in .y files (binutils)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:53 +0000 (21:01 +0100)]
Fix spelling in comments in .y files (binutils)

* arparse.y: Fix spelling in comments.

7 years agoFix spelling in comments in .sc files (ld)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:52 +0000 (21:01 +0100)]
Fix spelling in comments in .sc files (ld)

* scripttempl/ia64vms.sc: Fix spelling in comments.
* scripttempl/ip2k.sc: Fix spelling in comments.
* scripttempl/v850.sc: Fix spelling in comments.
* scripttempl/v850_rh850.sc: Fix spelling in comments.

7 years agoFix spelling in comments in .l files (gas)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:51 +0000 (21:01 +0100)]
Fix spelling in comments in .l files (gas)

* config/bfin-lex.l: Fix spelling in comments.

7 years agoFix spelling in comments in .em files (ld)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:50 +0000 (21:01 +0100)]
Fix spelling in comments in .em files (ld)

* emultempl/avrelf.em: Fix spelling in comments.
* emultempl/elf32.em: Fix spelling in comments.
* emultempl/pe.em: Fix spelling in comments.
* emultempl/pep.em: Fix spelling in comments.
* emultempl/spuelf.em: Fix spelling in comments.

7 years agoFix spelling in comments in .def files (gold)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:49 +0000 (21:01 +0100)]
Fix spelling in comments in .def files (gold)

* aarch64-reloc.def: Fix spelling in comments.

7 years agoFix spelling in comments in Expect scripts (ld)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:48 +0000 (21:01 +0100)]
Fix spelling in comments in Expect scripts (ld)

* testsuite/ld-sh/arch/arch.exp: Fix spelling in comments.
* testsuite/ld-sh/rd-sh.exp: Fix spelling in comments.
* testsuite/ld-sh/sh64/rd-sh64.exp: Fix spelling in comments.
* testsuite/ld-undefined/undefined.exp: Fix spelling in comments.

7 years agoFix spelling in comments in Expect scripts (gas)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:47 +0000 (21:01 +0100)]
Fix spelling in comments in Expect scripts (gas)

* testsuite/gas/all/gas.exp: Fix spelling in comments.
* testsuite/gas/cris/cris.exp: Fix spelling in comments.
* testsuite/gas/hppa/basic/basic.exp: Fix spelling in comments.
* testsuite/gas/hppa/parse/parse.exp: Fix spelling in comments.
* testsuite/gas/hppa/reloc/reloc.exp: Fix spelling in comments.
* testsuite/gas/sh/arch/arch.exp: Fix spelling in comments.
* testsuite/gas/tic4x/tic4x.exp: Fix spelling in comments.

7 years agoFix spelling in comments in Assembler files (ld)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:46 +0000 (21:01 +0100)]
Fix spelling in comments in Assembler files (ld)

* testsuite/ld-arm/stm32l4xx-fix-all.s: Fix spelling in comments.
* testsuite/ld-arm/thumb2-b-interwork.s: Fix spelling in comments.
* testsuite/ld-arm/thumb2-bl.s: Fix spelling in comments.
* testsuite/ld-s390/tlspic1.s: Fix spelling in comments.
* testsuite/ld-s390/tlspic1_64.s: Fix spelling in comments.
* testsuite/ld-scripts/section-match-1.d: Fix spelling in comments.

7 years agoFix spelling in comments in Assembler files (gas)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:45 +0000 (21:01 +0100)]
Fix spelling in comments in Assembler files (gas)

* testsuite/gas/arm/local_function.d: Fix spelling in comments.
* testsuite/gas/arm/req.s: Fix spelling in comments.
* testsuite/gas/arm/vfp1.s: Fix spelling in comments.
* testsuite/gas/arm/vfp1_t2.s: Fix spelling in comments.
* testsuite/gas/arm/vfp1xD.s: Fix spelling in comments.
* testsuite/gas/arm/vfp1xD_t2.s: Fix spelling in comments.
* testsuite/gas/mcore/allinsn.s: Fix spelling in comments.
* testsuite/gas/mips/24k-triple-stores-5.s: Fix spelling in comments.
* testsuite/gas/mips/delay.d: Fix spelling in comments.
* testsuite/gas/mips/nodelay.d: Fix spelling in comments.
* testsuite/gas/mips/r5900-full.s: Fix spelling in comments.
* testsuite/gas/mips/r5900.s: Fix spelling in comments.

7 years agoFix spelling in comments in C source files (ld)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:44 +0000 (21:01 +0100)]
Fix spelling in comments in C source files (ld)

* deffile.h: Fix spelling in comments.
* ld.h: Fix spelling in comments.
* ldlang.c: Fix spelling in comments.
* ldmisc.c: Fix spelling in comments.
* pe-dll.c: Fix spelling in comments.

7 years agoFix spelling in comments in C source files (gprof)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:43 +0000 (21:01 +0100)]
Fix spelling in comments in C source files (gprof)

* basic_blocks.c: Fix spelling in comments.
* cg_arcs.c: Fix spelling in comments.
* cg_print.c: Fix spelling in comments.
* corefile.c: Fix spelling in comments.

7 years agoFix spelling in comments in C source files (gold)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:42 +0000 (21:01 +0100)]
Fix spelling in comments in C source files (gold)

* aarch64.cc: Fix spelling in comments.
* arm.cc: Fix spelling in comments.
* icf.cc: Fix spelling in comments.
* layout.cc: Fix spelling in comments.
* layout.h: Fix spelling in comments.
* mips.cc: Fix spelling in comments.
* output.h: Fix spelling in comments.
* plugin.h: Fix spelling in comments.
* script-sections.h: Fix spelling in comments.
* script.h: Fix spelling in comments.
* stringpool.h: Fix spelling in comments.
* tilegx.cc: Fix spelling in comments.

7 years agoFix spelling in comments in C source files (gas)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:41 +0000 (21:01 +0100)]
Fix spelling in comments in C source files (gas)

* as.h: Fix spelling in comments.
* config/obj-ecoff.c: Fix spelling in comments.
* config/obj-macho.c: Fix spelling in comments.
* config/tc-aarch64.c: Fix spelling in comments.
* config/tc-arc.c: Fix spelling in comments.
* config/tc-arm.c: Fix spelling in comments.
* config/tc-avr.c: Fix spelling in comments.
* config/tc-cr16.c: Fix spelling in comments.
* config/tc-epiphany.c: Fix spelling in comments.
* config/tc-frv.c: Fix spelling in comments.
* config/tc-hppa.c: Fix spelling in comments.
* config/tc-hppa.h: Fix spelling in comments.
* config/tc-i370.c: Fix spelling in comments.
* config/tc-m68hc11.c: Fix spelling in comments.
* config/tc-m68k.c: Fix spelling in comments.
* config/tc-mcore.c: Fix spelling in comments.
* config/tc-mep.c: Fix spelling in comments.
* config/tc-metag.c: Fix spelling in comments.
* config/tc-mips.c: Fix spelling in comments.
* config/tc-mn10200.c: Fix spelling in comments.
* config/tc-mn10300.c: Fix spelling in comments.
* config/tc-nds32.c: Fix spelling in comments.
* config/tc-nios2.c: Fix spelling in comments.
* config/tc-ns32k.c: Fix spelling in comments.
* config/tc-pdp11.c: Fix spelling in comments.
* config/tc-ppc.c: Fix spelling in comments.
* config/tc-riscv.c: Fix spelling in comments.
* config/tc-rx.c: Fix spelling in comments.
* config/tc-score.c: Fix spelling in comments.
* config/tc-score7.c: Fix spelling in comments.
* config/tc-sparc.c: Fix spelling in comments.
* config/tc-tic54x.c: Fix spelling in comments.
* config/tc-vax.c: Fix spelling in comments.
* config/tc-xgate.h: Fix spelling in comments.
* config/tc-xtensa.c: Fix spelling in comments.
* config/tc-z80.c: Fix spelling in comments.
* dwarf2dbg.c: Fix spelling in comments.
* input-file.h: Fix spelling in comments.
* itbl-ops.c: Fix spelling in comments.
* read.c: Fix spelling in comments.
* stabs.c: Fix spelling in comments.
* symbols.c: Fix spelling in comments.
* write.c: Fix spelling in comments.
* testsuite/gas/all/itbl-test.c: Fix spelling in comments.
* testsuite/gas/tic4x/opclasses.h: Fix spelling in comments.

7 years agoFix spelling in comments in C source files (binutils)
Ambrogino Modigliani [Fri, 25 Nov 2016 20:01:40 +0000 (21:01 +0100)]
Fix spelling in comments in C source files (binutils)

* dwarf.c: Fix spelling in comments.
* dwarf.h: Fix spelling in comments.
* objcopy.c: Fix spelling in comments.
* od-macho.c: Fix spelling in comments.
* rclex.c: Fix spelling in comments.
* readelf.c: Fix spelling in comments.
* stabs.c: Fix spelling in comments.

7 years agoRemove verbosity from ui_out_message and friends
Simon Marchi [Sun, 27 Nov 2016 03:06:07 +0000 (22:06 -0500)]
Remove verbosity from ui_out_message and friends

That concept is never actually used, so it's just a burden.  Removing it
facilitates the refactoring in upcoming patches.

gdb/ChangeLog:

* mi/mi-out.c (mi_message): Remove verbosity argument.
* ada-tasks.c (print_ada_task_info, info_task, task_command):
Update call.
* auto-load.c (auto_load_info_scripts): Likewise.
* breakpoint.c (breakpoint_1, watchpoints_info, tracepoints_info):
Likewise.
* cli-out.c (cli_message): Remove verbosity argument.
* inferior.c (print_inferior): Update call.
* linux-thread-db.c (info_auto_load_libthread_db): Likewise.
* probe.c (info_probes_for_ops): Likewise.
* skip.c (skip_info): Likewise.
* solib.c (info_sharedlibrary_command): Likewise.
* symfile.c (load_progress): Likewise.
* thread.c (print_thread_info_1): Likewise.
* ui-out.c (uo_message, ui_out_message): Remove verbosity argument.
(ui_out_get_verblvl): Remove.
* ui-out.h (ui_out_message): Remove verbosity argument.
(ui_out_get_verblvl): Remove.
(message_ftype): Remove verbosity argument.

7 years agoConstify wrap_here/wrap_hint code path
Simon Marchi [Sun, 27 Nov 2016 03:05:42 +0000 (22:05 -0500)]
Constify wrap_here/wrap_hint code path

Constify the data path between ui_out_wrap_hint and the wrap_indent
global, because we can.  It's clearer that the argument passed to
wrap_hint is not intended to be modified by the ui_out implementation.

gdb/ChangeLog:

* mi/mi-out.c (mi_wrap_hint): Constify argument.
* cli-out.c (cli_wrap_hint): Likewise.
* ui-out.c (ui_out_wrap_hint, uo_wrap_hint): Likewise.
* ui-out.h (ui_out_wrap_hint, wrap_hint_ftype): Likewise.
* utils.c (wrap_here): Likewise.
(wrap_indent): Constify.
* utils.h (wrap_here): Constify argument.

7 years agoFix return value of uo_redirect
Simon Marchi [Sun, 27 Nov 2016 03:05:25 +0000 (22:05 -0500)]
Fix return value of uo_redirect

The wrapper uo_redirect seems like it should return the return value
from of implementation function, since callers rely on it, but it
doesn't.

gdb/ChangeLog:

* ui-out.c (uo_redirect): Return the return value from the
implementation function.

7 years agoRemove ui_out_destroy
Simon Marchi [Sun, 27 Nov 2016 03:05:06 +0000 (22:05 -0500)]
Remove ui_out_destroy

It's not actually used, and removing it simplifies the upcoming patches
a bit.  After the whole series, destroying an ui_out object will be
simply "delete uiout", which will call the default destructor.

gdb/ChangeLog:

* ui-out.c (ui_out_destroy, uo_data_destroy): Remove.
* ui-out.h (ui_out_destroy): Remove.

7 years agoRename ui_out_data to mi_ui_out_data
Simon Marchi [Sun, 27 Nov 2016 03:04:49 +0000 (22:04 -0500)]
Rename ui_out_data to mi_ui_out_data

Just a little cleanup, so the name is more consistent with the naming of
the equivalent structures of cli and tui.  It goes away in subsequent
patches anyway, but it might help follow the changes in those patches...

gdb/ChangeLog:

* mi/mi-out.c (ui_out_data): Rename to ...
(mi_ui_out_data): ... this.

7 years agoRemove unused functions and declarations
Simon Marchi [Sun, 27 Nov 2016 03:02:56 +0000 (22:02 -0500)]
Remove unused functions and declarations

gdb/ChangeLog:

* ui-out.c (_initialize_ui_out): Remove.
(ui_out_set_flags): Remove.
(ui_out_clear_flags): Remove.
* ui-out.h (ui_out_begin_cleanup_end): Remove.
(ui_out_begin_cleanup_end): Remove.
(ui_out_set_flags): Remove.
(ui_out_clear_flags): Remove.
* mi/mi-out.c (_initialize_mi_out): Remove.
(mi_out_buffered): Remove.
* mi/mi-out.h (mi_out_buffered): Remove.

7 years agoAutomatic date update in version.in
GDB Administrator [Sun, 27 Nov 2016 00:00:30 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoRemove stale comments
Simon Marchi [Thu, 24 Nov 2016 15:23:07 +0000 (10:23 -0500)]
Remove stale comments

The default ui_out is gone since commit
23ff98d2fed4a1eaeb815e18cd4169e5aa7aaa60.

gdb/ChangeLog:

* ui-out.h (struct ui_out_impl): Remove comment.
* ui-out.c (struct ui_out): Remove comment.

7 years agoAutomatic date update in version.in
GDB Administrator [Sat, 26 Nov 2016 00:00:09 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoRemove check requiring void argument to functions with no parameters.
John Baldwin [Fri, 25 Nov 2016 18:29:50 +0000 (10:29 -0800)]
Remove check requiring void argument to functions with no parameters.

C++ treats an empty parameter list as no parameters unlike C.

gdb/ChangeLog:

* contrib/ari/gdb_ari.sh (no parameter function): Remove check.

7 years agoFix typo in Makefile
Simon Marchi [Fri, 25 Nov 2016 14:58:02 +0000 (09:58 -0500)]
Fix typo in Makefile

Fix a typo I made in my previous Makefile cleanup series.

Thanks to Patrick Monnerat for reporting.

gdb/ChangeLog:

* Makefile.in: Fix typo.

7 years agoFix typos in comment
Simon Marchi [Fri, 25 Nov 2016 14:41:58 +0000 (09:41 -0500)]
Fix typos in comment

gdb/ChangeLog:

* record-full.c (record_full_resume): Fix typos in comment.

7 years agoFix typo in comment
Simon Marchi [Fri, 25 Nov 2016 14:07:35 +0000 (09:07 -0500)]
Fix typo in comment

gdb/ChangeLog:

* infcmd.c (interrupt_command): Fix typo in comment.

7 years agogas: fix CBCOND diagnostics for invalid immediate operands.
Jose E. Marchesi [Fri, 25 Nov 2016 11:40:15 +0000 (03:40 -0800)]
gas: fix CBCOND diagnostics for invalid immediate operands.

This patch fixes two problems in the SPARC assembler:

- The diagnostic message

  Error: Illegal operands: Immediate value in cbcond is out of range.

  is incorrectly issued for non-CBCOND instructions that feature a
  simm5 immediate field, such as MPMUL, MONTMUL, etc.

- When an invalid immediate operand is used in a CBCOND
  instruction, two redundant error messages are issued to the
  user, the second due to a stale fixup (this happens since
  commit 85024cd8bcb93f4112470ecdbd6c10fc2aea724f).

Some diagnostic tests for the CBCOND instructions are also
included in the patch.

Tested in both sparc64-linux-gnu and sparcv9-linux-gnu targets.

gas/ChangeLog:

2016-11-25  Jose E. Marchesi  <jose.marchesi@oracle.com>

* config/tc-sparc.c (sparc_ip): Avoid emitting a cbcond error
messages for non-cbcond instructions.
* testsuite/gas/sparc/cbcond-diag.s: New file.
* testsuite/gas/sparc/cbcond-diag.l: Likewise.
* testsuite/gas/sparc/sparc.exp (gas_64_check): Run cbcond-diag tests.

7 years agoPrevent problems with section alignment by not shrinking the .rsrc section.
Jon Turney [Fri, 25 Nov 2016 09:47:31 +0000 (09:47 +0000)]
Prevent problems with section alignment by not shrinking the .rsrc section.

PR ld/20193
* peXXigen.c (rsrc_process_section): Do not shrink the merged
.rsrc section.

8 years agoAutomatic date update in version.in
GDB Administrator [Fri, 25 Nov 2016 00:00:09 +0000 (00:00 +0000)]
Automatic date update in version.in

8 years agoDo not use std::move when assigning an anonymous object to a unique_ptr.
John Baldwin [Thu, 24 Nov 2016 20:01:24 +0000 (12:01 -0800)]
Do not use std::move when assigning an anonymous object to a unique_ptr.

Using std::move forces an extra copy of the object.  These changes fix
-Wpessimizing-move warnings from clang.

gdb/ChangeLog:

* ada-lang.c (create_excep_cond_exprs): Do not use 'std::move'.
* ax-gdb.c (agent_eval_command_one): Likewise.
(agent_eval_command_one): Likewise.
* breakpoint.c (parse_cond_to_aexpr): Likewise.
(parse_cmd_to_aexpr): Likewise.
* dtrace-probe.c (dtrace_process_dof_probe): Likewise.
* parse.c (parse_expression_for_completion): Likewise.

8 years agoAdd noexcept to custom non-throwing new operators.
John Baldwin [Thu, 24 Nov 2016 20:01:24 +0000 (12:01 -0800)]
Add noexcept to custom non-throwing new operators.

Both libc++ and libstdc++ declare non-throwing new operators as
noexcept and overloads must also be noexcept.  This fixes a
-Wmissing-exception-spec warning with clang.

gdb/ChangeLog:

* common/new-op.c (operator new): Mark 'noexcept'.
(operator new[]): Likewise.

8 years agoOptimize byte-aligned copies in copy_bitwise()
Andreas Arnez [Thu, 24 Nov 2016 16:48:04 +0000 (17:48 +0100)]
Optimize byte-aligned copies in copy_bitwise()

The function copy_bitwise used for copying DWARF pieces can potentially
be invoked for large chunks of data.  For instance, consider a large
struct one of whose members is currently located in a register.  In this
case copy_bitwise would still copy the data bitwise in a loop, which is
much slower than necessary.

This change uses memcpy for the large part instead, if possible.

gdb/ChangeLog:

* dwarf2loc.c (copy_bitwise): Use memcpy for the middle part, if
it is byte-aligned.

8 years agoAdd unit test for copy_bitwise
Andreas Arnez [Thu, 24 Nov 2016 16:48:04 +0000 (17:48 +0100)]
Add unit test for copy_bitwise

This adds a unit test for the copy_bitwise function in dwarf2loc.c.
With the old (broken) version of copy_bitwise this test would generate
the following failure message:

(gdb) maintenance selftest
Self test failed: copy_bitwise 11000000 != 10000000 (7+2 -> 0)

gdb/ChangeLog:
2016-11-24  Andreas Arnez  <arnez@linux.vnet.ibm.com>
    Pedro Alves  <palves@redhat.com>

* dwarf2loc.c (bits_to_str, check_copy_bitwise)
(copy_bitwise_tests): New functions.
(_initialize_dwarf2loc): Register the new function
copy_bitwise_tests as a unit test.
* selftest.c (run_self_tests): Improve the failure message's
wording and formatting.

8 years agoFix copy_bitwise()
Andreas Arnez [Thu, 24 Nov 2016 16:48:03 +0000 (17:48 +0100)]
Fix copy_bitwise()

When the user writes or reads a variable whose location is described
with DWARF pieces (DW_OP_piece or DW_OP_bit_piece), GDB's helper
function copy_bitwise is invoked for each piece.  The implementation of
this function has a bug that may result in a corrupted copy, depending
on alignment and bit size.  (Full-byte copies are not affected.)

This rewrites copy_bitwise, replacing its algorithm by a fixed version,
and adding an appropriate test case.  Without the fix the new test case
fails, e.g.:

  print def_t
  $2 = {a = 0, b = 4177919}
  (gdb) FAIL: gdb.dwarf2/nonvar-access.exp: print def_t

Written in binary, the wrong result above looks like this:

  01111111011111111111111

Which means that two zero bits have sneaked into the copy of the
original all-one bit pattern.  The test uses this simple all-one value
in order to avoid another GDB bug that causes the DWARF piece of a
DW_OP_stack_value to be taken from the wrong end on big-endian
architectures.

gdb/ChangeLog:

* dwarf2loc.c (extract_bits_primitive): Remove.
(extract_bits): Remove.
(copy_bitwise): Rewrite.  Fixes a possible corruption that may
occur for non-byte-aligned copies.

gdb/testsuite/ChangeLog:

* gdb.dwarf2/nonvar-access.exp: Add a test for accessing
non-byte-aligned bit fields.

8 years agoFix PR12616 - gdb does not implement DW_AT_data_bit_offset
Andreas Arnez [Thu, 24 Nov 2016 16:48:03 +0000 (17:48 +0100)]
Fix PR12616 - gdb does not implement DW_AT_data_bit_offset

The DW_AT_data_bit_offset attribute was introduced by DWARF V4 and
allows specifying the offset of a data member within its containing
entity.  But although the new attribute was intended to replace
DW_AT_bit_offset for this purpose, GDB ignores it, and thus GCC still
emits DW_AT_bit_offset instead.  See also
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71669.

This change fixes GDB's lack of support for DW_AT_data_bit_offset and
adds an appropriate test case.

gdb/ChangeLog:

PR gdb/12616
* dwarf2read.c (dwarf2_add_field): Handle the DWARF V4 attribute
DW_AT_data_bit_offset.

gdb/testsuite/ChangeLog:

PR gdb/12616
* gdb.dwarf2/nonvar-access.exp: New testcase.  Check that GDB
respects the DW_AT_data_bit_offset attribute.

8 years ago[ARM] Bind defined symbol locally in PIE
Jiong Wang [Thu, 24 Nov 2016 14:01:53 +0000 (14:01 +0000)]
[ARM] Bind defined symbol locally in PIE

bfd/
PR target/20737
* elf32-arm.c (elf32_arm_final_link_relocate): Bind defined symbol
locally in PIE.

ld/
* testsuite/ld-arm/pie-bind-locally-a.s: New test source.
* testsuite/ld-arm/pie-bind-locally-b.s: Likewise.
* testsuite/ld-arm/pie-bind-locally.d: New testcase.
* testsuite/ld-arm/arm-elf.exp: Run new testcase.

8 years agoFix snafu parsing $ORIGIN.
Nick Clifton [Thu, 24 Nov 2016 10:00:20 +0000 (10:00 +0000)]
Fix snafu parsing $ORIGIN.

PR ld/20858
* emultempl/elf32.em (_search_needed): Allow for path separator
and terminating NUL byte when allocating space for new $ORIGIN
path.

8 years agoRISC-V/bfd: Fix bitsize of R_RISCV_ADD8.
Kuan-Lin Chen [Thu, 24 Nov 2016 02:22:15 +0000 (10:22 +0800)]
RISC-V/bfd: Fix bitsize of R_RISCV_ADD8.

bfd/ChangeLog:
* bfd/elfxx-riscv.c (howto_table): Fix bitsize of R_RISCV_ADD8.

8 years agoAutomatic date update in version.in
GDB Administrator [Thu, 24 Nov 2016 00:00:32 +0000 (00:00 +0000)]
Automatic date update in version.in

8 years agogdb: Use C++11 std::chrono
Pedro Alves [Wed, 23 Nov 2016 15:36:26 +0000 (15:36 +0000)]
gdb: Use C++11 std::chrono

This patch fixes a few problems with GDB's time handling.

#1 - It avoids problems with gnulib's C++ namespace support

On MinGW, the struct timeval that should be passed to gnulib's
gettimeofday replacement is incompatible with libiberty's
timeval_sub/timeval_add.  That's because gnulib also replaces "struct
timeval" with its own definition, while libiberty expects the
system's.

E.g., in code like this:

  gettimeofday (&prompt_ended, NULL);
  timeval_sub (&prompt_delta, &prompt_ended, &prompt_started);
  timeval_add (&prompt_for_continue_wait_time,
               &prompt_for_continue_wait_time, &prompt_delta);

That's currently handled in gdb by not using gnulib's gettimeofday at
all (see common/gdb_sys_time.h), but that #undef hack won't work with
if/when we enable gnulib's C++ namespace support, because that mode
adds compile time warnings for uses of ::gettimeofday, which are hard
errors with -Werror.

#2 - But there's an elephant in the room: gettimeofday is not monotonic...

We're using it to:

  a) check how long functions take, for performance analysis
  b) compute when in the future to fire events in the event-loop
  c) print debug timestamps

But that's exactly what gettimeofday is NOT meant for.  Straight from
the man page:

~~~
       The time returned by gettimeofday() is affected by
       discontinuous jumps in the system time (e.g., if the system
       administrator manually changes the system time).  If you need a
       monotonically increasing clock, see clock_gettime(2).
~~~

std::chrono (part of the C++11 standard library) has a monotonic clock
exactly for such purposes (std::chrono::steady_clock).  This commit
switches to use that instead of gettimeofday, fixing all the issues
mentioned above.

gdb/ChangeLog:
2016-11-23  Pedro Alves  <palves@redhat.com>

* Makefile.in (SFILES): Add common/run-time-clock.c.
(HFILES_NO_SRCDIR): Add common/run-time-clock.h.
(COMMON_OBS): Add run-time-clock.o.
* common/run-time-clock.c, common/run-time-clock.h: New files.
* defs.h (struct timeval, print_transfer_performance): Delete
declarations.
* event-loop.c (struct gdb_timer) <when>: Now a
std::chrono::steady_clock::time_point.
(create_timer): use std::chrono::steady_clock instead of
gettimeofday.  Use new instead of malloc.
(delete_timer): Use delete instead of xfree.
(duration_cast_timeval): New.
(update_wait_timeout): Use std::chrono::steady_clock instead of
gettimeofday.
* maint.c: Include <chrono> instead of "gdb_sys_time.h", <time.h>
and "timeval-utils.h".
(scoped_command_stats::~scoped_command_stats)
(scoped_command_stats::scoped_command_stats): Use
std::chrono::steady_clock instead of gettimeofday.  Use
user_cpu_time_clock instead of get_run_time.
* maint.h: Include "run-time-clock.h" and <chrono>.
(scoped_command_stats): <m_start_cpu_time>: Now a
user_cpu_time_clock::time_point.
<m_start_wall_time>: Now a std::chrono::steady_clock::time_point.
* mi/mi-main.c: Include "run-time-clock.h" and <chrono> instead of
"gdb_sys_time.h" and <sys/resource.h>.
(rusage): Delete.
(mi_execute_command): Use new instead of XNEW.
(mi_load_progress): Use std::chrono::steady_clock instead of
gettimeofday.
(timestamp): Rewrite in terms of std::chrono::steady_clock,
user_cpu_time_clock and system_cpu_time_clock.
(timeval_diff): Delete.
(print_diff): Adjust to use std::chrono::steady_clock,
user_cpu_time_clock and system_cpu_time_clock.
* mi/mi-parse.h: Include "run-time-clock.h" and <chrono> instead
of "gdb_sys_time.h".
(struct mi_timestamp): Change fields types to
std::chrono::steady_clock::time_point, user_cpu_time_clock::time
and system_cpu_time_clock::time_point, instead of struct timeval.
* symfile.c: Include <chrono> instead of <time.h> and
"gdb_sys_time.h".
(struct time_range): New.
(generic_load): Use std::chrono::steady_clock instead of
gettimeofday.
(print_transfer_performance): Replace timeval parameters with a
std::chrono::steady_clock::duration parameter.  Adjust.
* utils.c: Include <chrono> instead of "timeval-utils.h",
"gdb_sys_time.h", and <time.h>.
(prompt_for_continue_wait_time): Now a
std::chrono::steady_clock::duration.
(defaulted_query, prompt_for_continue): Use
std::chrono::steady_clock instead of
gettimeofday/timeval_sub/timeval_add.
(reset_prompt_for_continue_wait_time): Use
std::chrono::steady_clock::duration instead of struct timeval.
(get_prompt_for_continue_wait_time): Return a
std::chrono::steady_clock::duration instead of struct timeval.
(vfprintf_unfiltered): Use std::chrono::steady_clock instead of
gettimeofday.  Use std::string.  Use '.' instead of ':'.
* utils.h: Include <chrono>.
(get_prompt_for_continue_wait_time): Return a
std::chrono::steady_clock::duration instead of struct timeval.

gdb/gdbserver/ChangeLog:
2016-11-23  Pedro Alves  <palves@redhat.com>

* debug.c: Include <chrono> instead of "gdb_sys_time.h".
(debug_vprintf): Use std::chrono::steady_clock instead of
gettimeofday.  Use '.' instead of ':'.
* tracepoint.c: Include <chrono> instead of "gdb_sys_time.h".
(get_timestamp): Use std::chrono::steady_clock instead of
gettimeofday.

8 years agoAdjust linker test for arm-vxworks in wake of patch for PR 20815.
Nick Clifton [Wed, 23 Nov 2016 14:57:51 +0000 (14:57 +0000)]
Adjust linker test for arm-vxworks in wake of patch for PR 20815.

8 years agoMinor formatting fixups in Makefiles
Simon Marchi [Tue, 22 Nov 2016 21:14:25 +0000 (16:14 -0500)]
Minor formatting fixups in Makefiles

Mostly some whitespace changes to make things a bit more consistent.

gdb/ChangeLog:

* Makefile.in: Fix whitespace formatting.

gdb/gdbserver/ChangeLog:

* Makefile.in: Fix whitespace formatting.

8 years agoNormalize names of some source files
Simon Marchi [Tue, 22 Nov 2016 21:14:24 +0000 (16:14 -0500)]
Normalize names of some source files

Most tdep/nat files are named:

  <cpu>-<os>-tdep.c
  <cpu>-<os>-nat.c

A few files do not respect this scheme.  This patch renames them so that
they are consistent with the rest of the files.  It builds fine with
--enable-targets=all, but that doesn't test the nat files.  I can only
hope that my grep skill is good enough.

gdb/ChangeLog:

* Makefile.in (ALL_64_TARGET_OBS, ALL_TARGET_OBS,
HFILES_NO_SRCDIR, ALLDEPFILES): Rename files.
* alphabsd-nat.c: Rename to ...
* alpha-bsd-nat.c: ... this, adjust include.
* alphabsd-tdep.c: Rename to ...
* alpha-bsd-tdep.c: ... this, adjust include.
* alphabsd-tdep.h: Rename to ...
* alpha-bsd-tdep.h: ... this, adjust include barrier and comment.
* alphafbsd-tdep.c: Rename to ...
* alpha-fbsd-tdep.c: ... this.
* alphanbsd-tdep.c: Rename to ...
* alpha-nbsd-tdep.c: ... this, adjust include.
* alphaobsd-tdep.c: Rename to ...
* alpha-obsd-tdep.c: ... this, adjust include.
* amd64bsd-nat.c: Rename to ...
* amd64-bsd-nat.c: ... this, adjust include.
* amd64fbsd-nat.c: Rename to ...
* amd64-fbsd-nat.c: ... this, adjust include.
* amd64fbsd-tdep.c: Rename to ...
* amd64-fbsd-tdep.c: ... this, adjust include.
* amd64nbsd-nat.c: Rename to ...
* amd64-nbsd-nat.c: ... this.
* amd64nbsd-tdep.c: Rename to ...
* amd64-nbsd-tdep.c: ... this.
* amd64obsd-nat.c: Rename to ...
* amd64-obsd-nat.c: ... this.
* amd64obsd-tdep.c: Rename to ...
* amd64-obsd-tdep.c: ... this.
* amd64-tdep.h: Update comments.
* armbsd-tdep.c: Rename to ...
* arm-bsd-tdep.c: ... this.
* armnbsd-nat.c: Rename to ...
* arm-nbsd-nat.c: ... this.
* armnbsd-tdep.c: Rename to ...
* arm-nbsd-tdep.c: ... this.
* armobsd-tdep.c: Rename to ...
* arm-obsd-tdep.c: ... this.
* arm-tdep.h: Update comments.
* hppabsd-tdep.c: Rename to ...
* hppa-bsd-tdep.c: ... this, adjust include.
* hppabsd-tdep.h: Rename to ...
* hppa-bsd-tdep.h: ... this, adjust include barrier and comment.
* hppanbsd-nat.c: Rename to ...
* hppa-nbsd-nat.c: ... this.
* hppanbsd-tdep.c: Rename to ...
* hppa-nbsd-tdep.c: ... this, adjust include.
* hppaobsd-nat.c: Rename to ...
* hppa-obsd-nat.c: ... this.
* hppaobsd-tdep.c: Rename to ...
* hppa-obsd-tdep.c: ... this, adjust include.
* i386bsd-nat.c: Rename to ...
* i386-bsd-nat.c: ... this, adjust include.
* i386bsd-nat.h: Rename to ...
* i386-bsd-nat.h: ... this, adjust include barrier and comment.
* i386bsd-tdep.c: Rename to ...
* i386-bsd-tdep.c: ... this.
* i386fbsd-nat.c: Rename to ...
* i386-fbsd-nat.c: ... this, adjust include.
* i386fbsd-tdep.c: Rename to ...
* i386-fbsd-tdep.c: ... this, adjust include.
* i386fbsd-tdep.h: Rename to ...
* i386-fbsd-tdep.h: ... this, adjust include barrier and comment.
* i386gnu-nat.c: Rename to ...
* i386-gnu-nat.c: ... this.
* i386gnu-tdep.c: Rename to ...
* i386-gnu-tdep.c: ... this.
* i386nbsd-nat.c: Rename to ...
* i386-nbsd-nat.c: ... this, adjust include.
* i386nbsd-tdep.c: Rename to ...
* i386-nbsd-tdep.c: ... this.
* i386obsd-nat.c: Rename to ...
* i386-obsd-nat.c: ... this, adjust include.
* i386obsd-tdep.c: Rename to ...
* i386-obsd-tdep.c: ... this.
* i386v4-nat.c: Rename to ...
* i386-v4-nat.c: ... this.
* i386-tdep.h: Update comments.
* m68k-tdep.h: Update comments.
* m68kbsd-nat.c: Rename to ...
* m68k-bsd-nat.c: ... this.
* m68kbsd-tdep.c: Rename to ...
* m68k-bsd-tdep.c: ... this.
* m68klinux-nat.c: Rename to ...
* m68k-linux-nat.c: ... this.
* m68klinux-tdep.c: Rename to ...
* m68k-linux-tdep.c: ... this.
* m88kbsd-nat.c: Rename to ...
* m88k-bsd-nat.c: ... this.
* mipsnbsd-nat.c: Rename to ...
* mips-nbsd-nat.c: ... this, adjust include.
* mipsnbsd-tdep.c: Rename to ...
* mips-nbsd-tdep.c: ... this, adjust include.
* mipsnbsd-tdep.h: Rename to ...
* mips-nbsd-tdep.h: ... this, adjust include barrier and comment.
* mips64obsd-nat.c: Rename to ...
* mips64-obsd-nat.c: ... this.
* mips64obsd-tdep.c: Rename to ...
* mips64-obsd-tdep.c: ... this.
* ppcfbsd-nat.c: Rename to ...
* ppc-fbsd-nat.c: ... this, adjust include.
* ppcfbsd-tdep.c: Rename to ...
* ppc-fbsd-tdep.c: ... this, adjust include.
* ppcfbsd-tdep.h: Rename to ...
* ppc-fbsd-tdep.h: ... this, adjust include barrier and comment.
* ppcnbsd-nat.c: Rename to ...
* ppc-nbsd-nat.c: ... this, adjust include.
* ppcnbsd-tdep.c: Rename to ...
* ppc-nbsd-tdep.c: ... this, adjust include.
* ppcnbsd-tdep.h: Rename to ...
* ppc-nbsd-tdep.h: ... this, adjust include barrier and comment.
* ppcobsd-nat.c: Rename to ...
* ppc-obsd-nat.c: ... this, adjust include.
* ppcobsd-tdep.c: Rename to ...
* ppc-obsd-tdep.c: ... this, adjust include.
* ppcobsd-tdep.h: Rename to ...
* ppc-obsd-tdep.h: ... this, adjust include barrier and comment.
* shnbsd-nat.c: Rename to ...
* sh-nbsd-nat.c: ... this.
* shnbsd-tdep.c: Rename to ...
* sh-nbsd-tdep.c: ... this.
* sparcnbsd-nat.c: Rename to ...
* sparc-nbsd-nat.c: ... this.
* sparcnbsd-tdep.c: Rename to ...
* sparc-nbsd-tdep.c: ... this.
* sparcobsd-tdep.c: Rename to ...
* sparc-obsd-tdep.c: ... this.
* sparc64fbsd-nat.c: Rename to ...
* sparc64-fbsd-nat.c: ... this.
* sparc64fbsd-tdep.c: Rename to ...
* sparc64-fbsd-tdep.c: ... this.
* sparc64nbsd-nat.c: Rename to ...
* sparc64-nbsd-nat.c: ... this.
* sparc64nbsd-tdep.c: Rename to ...
* sparc64-nbsd-tdep.c: ... this.
* sparc64obsd-nat.c: Rename to ...
* sparc64-obsd-nat.c: ... this.
* sparc64obsd-tdep.c: Rename to ...
* sparc64-obsd-tdep.c: ... this.
* sparc64-tdep.h: Update comments.
* vaxbsd-nat.c: Rename to ...
* vax-bsd-nat.c: ... this.
* vaxnbsd-tdep.c: Rename to ...
* vax-nbsd-tdep.c: ... this.
* vaxobsd-tdep.c: Rename to ...
* vax-obsd-tdep.c: ... this.
* x86bsd-nat.h: Rename to ...
* x86-bsd-nat.h: ... this, adjust include barrier and comment.
* x86bsd-nat.c: Rename to ...
* x86-bsd-nat.c: ... this, adjust include.
* configure.tgt: Update renamed files.
* config/alpha/fbsd.mh: Update renamed files.
* config/alpha/nbsd.mh: Update renamed files.
* config/arm/nbsdelf.mh: Update renamed files.
* config/djgpp/fnchange.lst: Update renamed files.
* config/i386/fbsd.mh: Update renamed files.
* config/i386/fbsd64.mh: Update renamed files.
* config/i386/i386gnu.mh: Update renamed files.
* config/i386/i386sol2.mh: Update renamed files.
* config/i386/nbsd64.mh: Update renamed files.
* config/i386/nbsdelf.mh: Update renamed files.
* config/i386/obsd.mh: Update renamed files.
* config/i386/obsd64.mh: Update renamed files.
* config/i386/sol2-64.mh: Update renamed files.
* config/m68k/linux.mh: Update renamed files.
* config/m68k/nbsdelf.mh: Update renamed files.
* config/m68k/obsd.mh: Update renamed files.
* config/m88k/obsd.mh: Update renamed files.
* config/mips/nbsd.mh: Update renamed files.
* config/mips/obsd64.mh: Update renamed files.
* config/pa/nbsd.mh: Update renamed files.
* config/pa/obsd.mh: Update renamed files.
* config/powerpc/fbsd.mh: Update renamed files.
* config/powerpc/nbsd.mh: Update renamed files.
* config/powerpc/obsd.mh: Update renamed files.
* config/sh/nbsd.mh: Update renamed files.
* config/sparc/fbsd.mh: Update renamed files.
* config/sparc/nbsd64.mh: Update renamed files.
* config/sparc/nbsdelf.mh: Update renamed files.
* config/sparc/obsd64.mh: Update renamed files.
* config/vax/nbsdelf.mh: Update renamed files.
* config/vax/obsd.mh: Update renamed files.

8 years agoMakefiles: Flatten and sort file lists
Simon Marchi [Tue, 22 Nov 2016 21:14:22 +0000 (16:14 -0500)]
Makefiles: Flatten and sort file lists

I find the big file lists in the Makefiles a bit ugly and not very
practical.  Since there are multiple filenames on each line (as much as
fits in 80 columns), it's not easy to add, remove or change a name in
the middle.  As a result, we have a mix of long and short lines in no
particular order (ALL_TARGET_OBS is a good example).

I therefore suggest flattening the lists (one name per line) and keeping
them in alphabetical order.  The diffs will be much clearer and merge
conflicts will be easier to resolve.

A nice (IMO) side-effect I observed is that the files are compiled
alphabetically by make, so it gives a rough idea of the progress of the
build.

I added a comment in gdb/Makefile.in to mention to keep the file lists
ordered, and gave the general guidelines on what order to respect.  I
added a comment in other Makefiles which refers to gdb/Makefile.in, to
avoid duplication.

Running the patch through the buildbot found that gdb.base/default.exp
started to fail.  The languages in the error message shown when typing
"set language" have changed order.  We could probably improve gdb so
that it prints them in a stable order, regardless of the order of the
object list passed to the linked, but just fixing the test is easier for
now.

New in v2:

 - Change ordering style, directories go at the end.
 - Cleanup gdbserver's and data-directory's Makefile as well.
 - Add comments at top of Makefiles about the ordering.
 - Remove wrong trailing backslahes.
 - Fix test gdb.base/default.exp.

gdb/ChangeLog:

* Makefile.in: Add comment about file lists ordering.
(SUBDIR_CLI_OBS, SUBDIR_CLI_SRCS, SUBDIR_MI_OBS, SUBDIR_MI_SRCS,
SUBDIR_TUI_OBS, SUBDIR_TUI_SRCS, SUBDIR_GCC_COMPILE_OBS,
SUBDIR_GCC_COMPILE_SRCS, SUBDIR_GUILE_OBS, SUBDIR_GUILE_SRCS,
SUBDIR_PYTHON_OBS, SUBDIR_PYTHON_SRCS, SUBDIR_GDBTK_OBS,
SUBDIR_GDBTK_SRCS, XMLFILES, REMOTE_OBS, ALL_64_TARGET_OBS,
ALL_TARGET_OBS, SFILES, HFILES_NO_SRCDIR, HFILES_WITH_SRCDIR,
COMMON_OBS, YYFILES, YYOBJ, generated_files, ALLDEPFILES):
Flatten list and order alphabetically.
* data-directory/Makefile.in: Add comment about file lists
ordering.
(GEN_SYSCALLS_FILES, PYTHON_FILE_LIST): Flatten list and order
alphabetically.

gdb/gdbserver/ChangeLog:

* Makefile.in (SFILES, OBS): Flatten list and order
alphabetically.

gdb/testsuite/ChangeLog:

* gdb.base/default.exp: Fix output of "set language".

8 years agoFix the linker so that it will not silently generate ELF binaries with invalid progra...
Nick Clifton [Wed, 23 Nov 2016 11:10:39 +0000 (11:10 +0000)]
Fix the linker so that it will not silently generate ELF binaries with invalid program headers.  Fix readelf to report such invalid binaries.

PR ld/20815
bfd * elf.c (elf_modify_segment_map): Allow empty LOAD segments if
they contain the program headers.
(_bfd_elf_map_sections_to_segments): If the linker created the
PHDR segment then always attempt to include it in a LOAD segment.
(assign_file_positions_for_non_load_sections): Allow LOAD segments
to overlap PHDR segments.
(phdr_sorter): New function.  Sorts program headers.
(assign_file_positions_except_relocs): Sort the program headers
before writing them out.  Issue an error if the PHDR segment is
not covered by a LOAD segment, unless the backend allows it.
* elf-bfd.h (struct elf_backend_data): Add
elf_backend_allow_non_load_phdr.
* elfxx-target.h (elf_backend_allow_non_load_phdr): Provide
default definition that returns FALSE.
(elfNN_bed): Initialise the elf_backend_allow_non_load_phdr
field.
* elf64-hppa.c (elf64_hppa_allow_non_load_phdr): New function.
Returns TRUE.
(elf_backend_allow_non_load_phdr): Define.
* elf-m10300.c (_bfd_mn10300_elf_size_dynamic_sections): Do not
place the interpreter string into the .interp section if the
nointerp flag is set in the link info structure.
* elf32-arc.c (elf_arc_size_dynamic_sections): Likewise.
* elf32-score7.c (score_elf_final_link_relocate): Allow for the
_gp symbol not being part of the output.

binutils* readelf.c (process_program_headers): Check PT_LOAD and PT_PHDR
segments for validity.

ld * ld.texinfo: Note that PT_TLS can be used as a segment type.
* testsuite/ld-discard/discard.ld: Add space for program headers.
* testsuite/ld-elf/flags1.ld: Likewise.
* testsuite/ld-elf/maxpage3.t: Likewise.
* testsuite/ld-elf/noload-1.t: Likewise.
* testsuite/ld-elf/orphan.ld: Likewise.
* testsuite/ld-elf/overlay.t: Likewise.
* testsuite/ld-elf/pr14052.t: Likewise.
* testsuite/ld-elf/pr19539.t: Likewise.
* testsuite/ld-elf/provide-hidden-1.ld: Likewise.
* testsuite/ld-elf/provide-hidden-s.ld: Likewise.
* testsuite/ld-elf/weak-dyn-1.ld: Likewise.
* testsuite/ld-i386/pr19539.t: Likewise.
* testsuite/ld-scripts/defined.t: Likewise.
* testsuite/ld-scripts/defined6.t: Likewise.
* testsuite/ld-scripts/dynamic-sections.t: Likewise.
* testsuite/ld-scripts/empty-aligned.t: Likewise.
* testsuite/ld-scripts/provide-2.t: Likewise.
* testsuite/ld-scripts/provide-4.t: Likewise.
* testsuite/ld-vax-elf/plt-local.ld: Likewise.
* testsuite/ld-x86-64/pr19539.t: Likewise.
* testsuite/ld-elf/ehdr_start-missing.d: Do not initialise the
dynamic linker.
* testsuite/ld-elf/ehdr_start-weak.d: Likewise.
* testsuite/ld-elf/elf.exp (pr14170, pr17068): Likewise.
* testsuite/ld-elf/loadaddr1.d: Update expected readelf output.
* testsuite/ld-elf/noload-2.d: Likewise.
* testsuite/ld-powerpc/vxworks2.sd: Likewise.
* testsuite/ld-scripts/phdrs3a.d: Likewise.
* testsuite/ld-scripts/size-2.d: Likewise.
* testsuite/ld-elf/group.ld: Add program headers.
* testsuite/ld-elf/overlay.d: Skip for SPU.
* testsuite/ld-elf/flags1.d: Skip for RX.
* testsuite/ld-elf/pr19162.d: Skip for HPPA64.
* testsuite/ld-elf/pr19539.d: Skip for ALPHA.
* testsuite/ld-scripts/empty-orphan.t: Update program headers.
* testsuite/ld-scripts/size-2.t: Likewise.

8 years agogas: run the hwcaps-bump tests with 64-bit sparc objects only.
Jose E. Marchesi [Wed, 23 Nov 2016 11:04:17 +0000 (03:04 -0800)]
gas: run the hwcaps-bump tests with 64-bit sparc objects only.

gas/ChangeLog:

2016-11-23  Jose E. Marchesi  <jose.marchesi@oracle.com>

* testsuite/gas/sparc/sparc.exp (gas_64_check): Make sure the
hwcaps-bump test is run with 64-bit objects.

8 years agoRISCV/GAS Add missing break in md_apply_fix.
Kuan-Lin Chen [Wed, 23 Nov 2016 05:18:59 +0000 (13:18 +0800)]
RISCV/GAS Add missing break in md_apply_fix.

gdb/ChangeLog:
* config/tc-riscv.c: Add missing break.

8 years agoelf_backend_dtrel_excludes_plt
Alan Modra [Wed, 23 Nov 2016 04:37:17 +0000 (15:07 +1030)]
elf_backend_dtrel_excludes_plt

Now that all targets creating .rel.plt/.rela.plt use the ELF hash
table shortcut srelplt, the generic ELF code can set up DT_RELSZ/
DT_RELASZ and DT_REL/DT_RELA for targets that don't want PLT relocs
included in those tags.

* elf-bfd.h (struct elf_backend_data): Add dtrel_excludes_plt.
* elfxx-target.h (elf_backend_dtrel_excludes_plt): Define.
(elfNN_bed): Init new field.
* elflink.c (bfd_elf_final_link): Add and use htab variable.  Handle
dtrel_excludes_plt.
* elf-m10300.c (_bfd_mn10300_elf_finish_dynamic_sections): Delete
DT_RELASZ code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-arc.c (elf_arc_finish_dynamic_sections): Delete DT_RELASZ code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-arm.c (elf32_arm_finish_dynamic_sections): Delete code
subtracting off plt relocs from DT_RELSZ, DT_RELASZ.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-cr16.c (_bfd_cr16_elf_finish_dynamic_sections): Delete
DT_RELASZ code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-cris.c (elf_cris_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-hppa.c (elf32_hppa_finish_dynamic_sections): Delete DT_RELASZ
and DT_RELA code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-i386.c (elf_i386_finish_dynamic_sections): Delete DT_RELSZ
and DT_REL code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-lm32.c (lm32_elf_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-m32r.c (m32r_elf_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-m68k.c (elf_m68k_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-metag.c (elf_metag_finish_dynamic_sections): Delete DT_RELASZ
and DT_RELA code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-microblaze.c (microblaze_elf_finish_dynamic_sections): Delete
DT_RELASZ and DT_RELA code.  Use ELF htab shortcuts for other
dynamic sections.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-mips.c (elf_backend_dtrel_excludes_plt): Define.
* elf32-nds32.c (nds32_elf_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-nios2.c (nios2_elf32_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-or1k.c (or1k_elf_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-ppc.c (ppc_elf_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-sh.c (sh_elf_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-sparc.c (elf_backend_dtrel_excludes_plt): Define.
* elf32-vax.c (elf_vax_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf32-xtensa.c (elf_xtensa_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf64-alpha.c (elf64_alpha_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf64-ppc.c (ppc64_elf_finish_dynamic_sections): Delete DT_RELASZ
and DT_RELA code.
(elf_backend_dtrel_excludes_plt): Define.
* elf64-sh64.c (sh64_elf64_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elf64-x86-64.c (elf_x86_64_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elfnn-aarch64.c (elfNN_aarch64_finish_dynamic_sections): Delete
DT_RELASZ code.
(elf_backend_dtrel_excludes_plt): Define.
* elfnn-ia64.c (elfNN_ia64_finish_dynamic_sections): Delete DT_RELASZ
code.
(elf_backend_dtrel_excludes_plt): Define.
* elfxx-mips.c (_bfd_mips_elf_finish_dynamic_sections): Delete
DT_RELASZ code.
* elfxx-sparc.c (sparc_finish_dyn): Delete DT_RELASZ code.

8 years agoDelete duplicate target short-cuts to dynamic sections
Alan Modra [Wed, 23 Nov 2016 04:36:34 +0000 (15:06 +1030)]
Delete duplicate target short-cuts to dynamic sections

We'd like to have the elf_link_hash_table srelplt field rather than
some private target field used to save short-cuts to a PLT relocation
section.  This save a little space but mainly is so that the generic
ELF code can access the field.  Ditto for other dynamic sections.

* elf-m10300.c (mn10300_elf_check_relocs): Use elf htab shortcuts
to dynamic sections.
(mn10300_elf_final_link_relocate): Likewise.
(_bfd_mn10300_elf_adjust_dynamic_symbol): Likewise.
(_bfd_mn10300_elf_size_dynamic_sections): Likewise.
(_bfd_mn10300_elf_finish_dynamic_symbol): Likewise.
(_bfd_mn10300_elf_finish_dynamic_sections): Likewise.
* elf32-bfin.c (bfin_check_relocs): Likewise.
(bfin_relocate_section): Likewise.
(bfin_gc_sweep_hook): Likewise.
(struct bfinfdpic_elf_link_hash_table): Delete sgot, sgotrel, splt
and spltrel.
(bfinfdpic_got_section, bfinfdpic_gotrel_section,
bfinfdpic_plt_section, bfinfdpic_pltrel_section): Define using elf
shortcut sections.
(_bfin_create_got_section): Use elf htab shortcuts to dyn sections.
Delete dead code.
(bfin_finish_dynamic_symbol): Use elf htab shortcuts to dyn sections.
(bfin_size_dynamic_sections): Likewise.
* elf32-cr16.c (_bfd_cr16_elf_create_got_section): Likewise.
(cr16_elf_check_relocs): Likewise.
(cr16_elf_final_link_relocate): Likewise.
(_bfd_cr16_elf_create_dynamic_sections): Likewise.
(_bfd_cr16_elf_adjust_dynamic_symbol): Likewise.
(_bfd_cr16_elf_size_dynamic_sections): Likewise.
(_bfd_cr16_elf_finish_dynamic_symbol): Likewise.
(_bfd_cr16_elf_finish_dynamic_sections): Likewise.
* elf32-cris.c (cris_elf_relocate_section): Likewise.
(elf_cris_finish_dynamic_symbol): Likewise.
(elf_cris_finish_dynamic_sections): Likewise.
(cris_elf_gc_sweep_hook): Likewise.
(elf_cris_adjust_gotplt_to_got): Likewise.
(elf_cris_adjust_dynamic_symbol): Likewise.
(cris_elf_check_relocs): Likewise.  Delete dead code.
(elf_cris_size_dynamic_sections): Use elf htab shortcuts to dynamic
sections.
(elf_cris_discard_excess_program_dynamics): Likewise.
* elf32-frv.c (struct frvfdpic_elf_link_hash_table): Delete sgot,
sgotrel, splt and spltrel.
(frvfdpic_got_section, frvfdpic_gotrel_section,
frvfdpic_plt_section, frvfdpic_pltrel_section): Define using elf
shortcut sections.
(_frv_create_got_section): Likewise.
* elf32-hppa.c (struct elf32_hppa_link_hash_table): Delete sgot,
srelgot, splt and srelplt.
(hppa_build_one_stub): Use elf htab shortcuts to dynamic sections.
(elf32_hppa_create_dynamic_sections): Likewise.
(elf32_hppa_check_relocs): Likewise.
(allocate_plt_static): Likewise.
(allocate_dynrelocs): Likewise.
(elf32_hppa_size_dynamic_sections): Likewise.
(elf32_hppa_relocate_section): Likewise.
(elf32_hppa_finish_dynamic_symbol): Likewise.
(elf32_hppa_finish_dynamic_sections): Likewise.
* elf32-i370.c (i370_elf_finish_dynamic_sections): Likewise.
* elf32-lm32.c (struct elf_lm32_link_hash_table): Delete sgot,
sgotplt, srelgot, splt and srelplt.
(lm32fdpic_got_section, lm32fdpic_gotrel_section): Define using elf
shortcut sections.
(create_got_section): Delete.  Use _bfd_elf_create_got_section instead.
(lm32_elf_relocate_section): Use elf htab shortcuts to dyn sections.
(lm32_elf_check_relocs): Likewise.
(lm32_elf_finish_dynamic_sections): Likewise.
(lm32_elf_finish_dynamic_symbol): Likewise.
(allocate_dynrelocs): Likewise.
(lm32_elf_size_dynamic_sections): Likewise.
(lm32_elf_create_dynamic_sections): Likewise.
* elf32-m32c.c (m32c_elf_relocate_section): Likewise.
(m32c_elf_check_relocs): Likewise.
(m32c_elf_finish_dynamic_sections): Likewise.
(m32c_elf_always_size_sections): Likewise.
* elf32-m32r.c (struct elf_m32r_link_hash_table): Delete sgot,
sgotplt, srelgot, splt and srelplt.
(create_got_section): Delete.  Use _bfd_elf_create_got_section instead.
(m32r_elf_create_dynamic_sections): Use elf htab shortcuts to dynamic
sections.
(allocate_dynrelocs): Likewise.
(m32r_elf_size_dynamic_sections): Likewise.
(m32r_elf_relocate_section): Likewise.
(m32r_elf_finish_dynamic_symbol): Likewise.
(m32r_elf_finish_dynamic_sections): Likewise.
(m32r_elf_check_relocs): Likewise.
* elf32-m68k.c (elf_m68k_partition_multi_got): Likewise.
(elf_m68k_check_relocs): Likewise.
(elf_m68k_adjust_dynamic_symbol): Likewise.
(elf_m68k_size_dynamic_sections): Likewise.
(elf_m68k_relocate_section): Likewise.
(elf_m68k_finish_dynamic_symbol): Likewise.
(elf_m68k_finish_dynamic_sections): Likewise.
* elf32-metag.c (struct elf_metag_link_hash_table): Delete sgot,
sgotplt, srelgot, splt and srelplt.
(elf_metag_relocate_section): Use elf htab shortcuts to dynamic
sections.
(elf_metag_create_dynamic_sections): Likewise.  Allocate got header
here in .got.
(elf_metag_check_relocs): Use elf htab shortcuts to dynamic sections.
(allocate_dynrelocs): Likewise.
(elf_metag_size_dynamic_sections): Likewise.
(elf_metag_finish_dynamic_symbol): Likewise.
(elf_metag_finish_dynamic_sections): Likewise.
(elf_metag_size_stubs): Likewise.
(elf_backend_got_header_size): Don't define.
(elf_backend_want_got_plt): Define.
* elf32-microblaze.c (struct elf32_mb_link_hash_table): Delete sgot,
sgotplt, srelgot, splt and srelpl.
(microblaze_elf_relocate_section): Use elf htab shortcuts to dynamic
sections.
(create_got_section): Delete.  Use _bfd_elf_create_got_section instead.
(microblaze_elf_check_relocs): Use elf htab shortcuts to dyn sections.
(microblaze_elf_create_dynamic_sections): Likewise.
(allocate_dynrelocs): Likewise.
(microblaze_elf_size_dynamic_sections): Likewise.
(microblaze_elf_finish_dynamic_symbol): Likewise.
(microblaze_elf_finish_dynamic_sections): Likewise.
* elf32-nds32.c (nds32_elf_link_hash_table_create): Don't NULL
already zero fields.
(create_got_section): Delete.  Use _bfd_elf_create_got_section instead.
(nds32_elf_create_dynamic_sections): Use elf htab shortcuts to dynamic
sections.
(allocate_dynrelocs): Likewise.
(nds32_elf_size_dynamic_sections): Likewise.
(nds32_elf_relocate_section): Likewise.
(nds32_elf_finish_dynamic_symbol): Likewise.
(nds32_elf_finish_dynamic_sections): Likewise.
(nds32_elf_check_relocs): Likewise.
(calculate_plt_memory_address): Likewise.
(calculate_got_memory_address): Likewise.
* elf32-nds32.h (struct elf_nds32_link_hash_table): Delete sgot,
sgotplt, srelgot, splt and srelplt.
* elf32-or1k.c (struct elf_or1k_link_hash_table): Likewise.
(or1k_elf_relocate_section): Use elf htab shortcuts to dyn sections.
(create_got_section): Delete.  Use _bfd_elf_create_got_section instead.
(or1k_elf_check_relocs): Use elf htab shortcuts to dynamic sections.
(or1k_elf_finish_dynamic_sections): Likewise.
(or1k_elf_finish_dynamic_symbol): Likewise.
(allocate_dynrelocs): Likewise.
(or1k_elf_size_dynamic_sections): Likewise.
(or1k_elf_create_dynamic_sections): Likewise.
* elf32-ppc.c (struct ppc_elf_link_hash_table): Delete got, relgot,
plt, relplt, iplt, reliplt and sgotplt.
(ppc_elf_create_got): Use elf htab shortcuts to dynamic sections.
(ppc_elf_create_glink): Likewise.
(ppc_elf_create_dynamic_sections): Likewise.
(ppc_elf_check_relocs): Likewise.
(ppc_elf_select_plt_layout): Likewise.
(ppc_elf_tls_setup): Likewise.
(allocate_got): Likewise.
(allocate_dynrelocs): Likewise.
(ppc_elf_size_dynamic_sections): Likewise.
(ppc_elf_relax_section): Likewise.
(ppc_elf_relocate_section): Likewise.
(ppc_elf_finish_dynamic_symbol): Likewise.
(ppc_elf_reloc_type_class): Likewise.
(ppc_elf_finish_dynamic_sections): Likewise.
* elf32-rl78.c (rl78_elf_relocate_section): Likewise.
(rl78_elf_check_relocs): Likewise.
(rl78_elf_finish_dynamic_sections): Likewise.
(rl78_elf_always_size_sections): Likewise.
* elf32-s390.c  (create_got_section): Delete.
(elf_s390_create_dynamic_sections): Use _bfd_elf_create_got_section.
(elf_s390_check_relocs): Likewise.
* elf32-score.c (score_elf_create_got_section): Set elf shortcuts.
(s3_bfd_score_elf_finish_dynamic_sections): Use elf shortcuts.
* elf32-score7.c (score_elf_create_got_section): As above.
(s7_bfd_score_elf_finish_dynamic_sections): As above.
* elf32-sh.c (struct elf_sh_link_hash_table): Delete sgot,
sgotplt, srelgot, splt and srelplt.
(create_got_section): Don't set them.
(sh_elf_create_dynamic_sections): Use elf htab shortcuts to dynamic
sections.
(allocate_dynrelocs): Likewise.
(sh_elf_size_dynamic_sections): Likewise.
(sh_elf_add_rofixup): Likewise.
(sh_elf_relocate_section): Likewise.
(sh_elf_check_relocs): Likewise.
(sh_elf_finish_dynamic_symbol): Likewise.
(sh_elf_finish_dynamic_sections): Likewise.
* elf32-tic6x.c (elf32_tic6x_finish_dynamic_symbol): Likewise.
* elf32-tilepro.c (tilepro_elf_create_got_section): Likewise.
* elf32-vax.c (elf_vax_check_relocs): Likewise.
(elf_vax_adjust_dynamic_symbol): Likewise.
(elf_vax_always_size_sections): Likewise.
(elf_vax_instantiate_got_entries): Likewise.
(elf_vax_relocate_section): Likewise.
(elf_vax_finish_dynamic_symbol): Likewise.
(elf_vax_finish_dynamic_sections): Likewise.
* elf32-xstormy16.c (xstormy16_elf_check_relocs): Likewise.
(xstormy16_elf_always_size_sections): Likewise.
(xstormy16_elf_relocate_section): Likewise.
(xstormy16_elf_finish_dynamic_sections): Likewise.
* elf32-xtensa.c (struct elf_xtensa_link_hash_table): Delete sgot,
sgotplt, srelgot, splt and srelplt.
(elf_xtensa_create_dynamic_sections): Use elf htab shortcuts to
dynamic sections.
(elf_xtensa_allocate_dynrelocs): Likewise.
(elf_xtensa_allocate_local_got_size): Likewise.
(elf_xtensa_size_dynamic_sections): Likewise.
(elf_xtensa_relocate_section): Likewise.
(elf_xtensa_finish_dynamic_sections): Likewise.
(shrink_dynamic_reloc_sections): Likewise.
(elf_xtensa_get_plt_section): Likewise.
(elf_xtensa_get_gotplt_section): Likewise.
(xtensa_callback_required_dependence): Likewise.
* elf64-alpha.c (elf64_alpha_create_dynamic_sections): Set elf htab
shortcuts to dynamic sections.
(elf64_alpha_adjust_dynamic_symbol): Use elf htab shortcuts to
dynamic sections.
(elf64_alpha_size_plt_section): Likewise.
(elf64_alpha_size_rela_got_1): Likewise.
(elf64_alpha_size_rela_got_section): Likewise.
(elf64_alpha_relocate_section): Likewise.
(elf64_alpha_finish_dynamic_symbol): Likewise.
(elf64_alpha_finish_dynamic_sections): Likewise.
* elf64-hppa.c (elf64_hppa_size_dynamic_sections): Likewise.
* elf64-s390.c (create_got_section): Delete.
(elf_s390_create_dynamic_sections): Use _bfd_elf_create_got_section.
(elf_s390_check_relocs): Likewise.
* elf64-sh64.c (sh_elf64_relocate_section): Use elf htab shortcuts to
dynamic sections.
(sh_elf64_check_relocs): Likewise.
(sh64_elf64_adjust_dynamic_symbol): Likewise.
(sh64_elf64_size_dynamic_sections): Likewise.
(sh64_elf64_finish_dynamic_symbol): Likewise.
(sh64_elf64_finish_dynamic_sections): Likewise.
* elflink.c (_bfd_elf_create_got_section): Likewise.
* elfnn-aarch64.c (aarch64_elf_create_got_section): Likewise.
* elfnn-ia64.c (elfNN_ia64_size_dynamic_sections): Likewise.
(elfNN_ia64_finish_dynamic_sections): Likewise.
* elfnn-riscv.c (riscv_elf_create_got_section): Likewise.
* elfxx-mips.c (struct mips_elf_link_hash_table): Delete srellt,
sgotplt, splt and sgot.
(mips_elf_initialize_tls_slots): Use elf htab shortcuts to dynamic
sections.
(mips_elf_gotplt_index): Likewise.
(mips_elf_primary_global_got_index): Likewise.
(mips_elf_global_got_index): Likewise.
(mips_elf_got_offset_from_index): Likewise.
(mips_elf_create_local_got_entry): Likewise.
(mips_elf_create_got_section): Likewise.
(mips_elf_calculate_relocation): Likewise.
(_bfd_mips_elf_create_dynamic_sections): Likewise.
(_bfd_mips_elf_adjust_dynamic_symbol): Likewise.
(mips_elf_lay_out_got): Likewise.
(mips_elf_set_plt_sym_value): Likewise.
(_bfd_mips_elf_size_dynamic_sections): Likewise.
(_bfd_mips_elf_finish_dynamic_symbol): Likewise.
(_bfd_mips_vxworks_finish_dynamic_symbol): Likewise.
(mips_finish_exec_plt): Likewise.
(mips_vxworks_finish_exec_plt): Likewise.
(mips_vxworks_finish_shared_plt): Likewise.
(_bfd_mips_elf_finish_dynamic_sections): Likewise.
* elfxx-sparc.c (sparc_finish_dyn): Likewise.
* elfxx-tilegx.c (tilegx_elf_create_got_section): Likewise.

8 years agoRegen POTFILES.in
Alan Modra [Wed, 23 Nov 2016 04:34:24 +0000 (15:04 +1030)]
Regen POTFILES.in

bfd/
* po/BLD-POTFILES.in: Regenerate.
* po/SRC-POTFILES.in: Regenerate.
gas/
* po/POTFILES.in: Regenerate.

8 years agogdbserver: Use warning for warnings
Pedro Alves [Thu, 17 Nov 2016 23:15:34 +0000 (23:15 +0000)]
gdbserver: Use warning for warnings

gdb/gdbserver/ChangeLog:
2016-11-23  Pedro Alves  <palves@redhat.com>

* event-loop.c (handle_file_event): Use warning.
* linux-low.c (linux_resume_one_lwp_throw): Use warning.
* mem-break.c (add_breakpoint_condition, add_breakpoint_commands):
Use warning.

8 years agogdbserver: Use debug_printf for debug output
Pedro Alves [Fri, 18 Nov 2016 00:07:10 +0000 (00:07 +0000)]
gdbserver: Use debug_printf for debug output

gdb/gdbserver/ChangeLog:
2016-11-23  Pedro Alves  <palves@redhat.com>

* linux-low.c (check_zombie_leaders): Use debug_printf for debug
output.
* notif.c (handle_notif_ack, notif_event_enque): Likewise.
* remote-utils.c (putpkt_binary_1, readchar, getpkt): Use
debug_printf and debug_flush for debug output.
* server.c (handle_general_set): Likewise.
* thread-db.c (try_thread_db_load): Use debug_printf for debug
output.

8 years agoAutomatic date update in version.in
GDB Administrator [Wed, 23 Nov 2016 00:00:36 +0000 (00:00 +0000)]
Automatic date update in version.in

8 years agoPrint the default for all binary options; clean up --help output.
Cary Coutant [Tue, 22 Nov 2016 23:44:56 +0000 (15:44 -0800)]
Print the default for all binary options; clean up --help output.

gold/
PR gold/20346
* options.cc (One_option::print): Print "(default)" when appropriate.
* options.h: Clean up and re-sort options.
(One_option::is_default): New data member.
(One_option::One_option): Add is_default parameter; adjust all calls.
(DEFINE_var): Add is_default__ parameter; adjust all calls.
(DEFINE_bool): Set is_default based on default_value__.
(DEFINE_bool_ignore): New macro.
(--no-eh-frame-hdr): New option.
(--enable-new-dtags): Remove mention of DT_FLAGS.

8 years agoFix spelling mistakes in comments in shell scripts
Ambrogino Modigliani [Tue, 22 Nov 2016 16:05:00 +0000 (16:05 +0000)]
Fix spelling mistakes in comments in shell scripts

gdb/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * contrib/expect-read1.sh: Fix spelling in comments.
        * gdb_buildall.sh: Fix spelling in comments.
        * gdb_mbuild.sh: Fix spelling in comments.

8 years agoFix spelling mistakes in comments in configure scripts
Ambrogino Modigliani [Tue, 22 Nov 2016 15:43:03 +0000 (15:43 +0000)]
Fix spelling mistakes in comments in configure scripts

All changes are limited to comments, and no run-time behavior is
affected.

bfd/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * warning.m4: Fix spelling in comments.
        * configure.ac: Fix spelling in comments.
        * configure: Regenerate.

binutils/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * configure: Regenerate.

gdb/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * configure.ac: Fix spelling in comments.
        * configure: Regenerate.

gas/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * configure: Regenerate.

gold/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * configure: Regenerate.

gprof/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * configure: Regenerate.

ld/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * configure: Regenerate.

opcodes/ChangeLog:
2016-11-22  Ambrogino Modigliani  <ambrogino.modigliani@gmail.com>

        * configure: Regenerate.

8 years agoChange gdbarch software_single_step frame_info to regcache
Yao Qi [Tue, 22 Nov 2016 14:05:06 +0000 (14:05 +0000)]
Change gdbarch software_single_step frame_info to regcache

This patch changes gdbarch method software_single_step's parameter from
"struct frame_info *" to "struct regcache *, IOW, software_single_step
starts to use current regcache rather than current frame for software
single.

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* gdbarch.sh (software_single_step): Change parameter from frame_info
to regcache.
* gdbarch.c, gdbarch.h: Regenerated.
* aarch64-tdep.c (aarch64_software_single_step): Change parameter
from frame_info to regcache.  Don't call get_current_regcache.
* alpha-tdep.c (alpha_deal_with_atomic_sequence): Likewise.
(alpha_software_single_step): Likewise.
* alpha-tdep.h (alpha_software_single_step): Update declaration.
* arm-linux-tdep.c (arm_linux_software_single_step): Likewise.
* arm-tdep.c (arm_software_single_step): Likewise.
* arm-tdep.h (arm_software_single_step): Likewise.
* breakpoint.c (insert_single_step_breakpoint): Pass regcache to
gdbarch_software_single_step.
* cris-tdep.c (cris_software_single_step): Change parameter from
frame_info to regcache.  Don't call get_current_regcache.
* mips-tdep.c (mips_software_single_step): Likewise.
* mips-tdep.h (mips_software_single_step): Update declaration.
* moxie-tdep.c (moxie_software_single_step): Likewise.
* nios2-tdep.c (nios2_software_single_step): Likewise.
* ppc-tdep.h (ppc_deal_with_atomic_sequence): Update declaration.
* rs6000-aix-tdep.c (rs6000_software_single_step): Likewise.
* rs6000-tdep.c (ppc_deal_with_atomic_sequence): Likewise.
* s390-linux-tdep.c (s390_software_single_step): Likewise.
* sparc-tdep.c (sparc_software_single_step): Likewise.
* spu-tdep.c (spu_software_single_step): Likewise.
* tic6x-tdep.c (tic6x_software_single_step): Likewise.

8 years agogdbarch software_single_step frame_info to regcache: spu
Yao Qi [Tue, 22 Nov 2016 14:05:05 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: spu

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* spu-tdep.c (spu_software_single_step): Call get_regcache_arch
instead of get_frame_arch.  Call regcache_read_pc instead of
get_frame_pc.  Call regcache_raw_get_unsigned instead of
get_frame_register_unsigned.

8 years agogdbarch software_single_step frame_info to regcache: tic6x
Yao Qi [Tue, 22 Nov 2016 14:05:05 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: tic6x

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* tic6x-tdep.c (tic6x_condition_true): Replace frame with
regcache.  Call regcache_raw_get_signed instead of
get_frame_register_signed.
(tic6x_get_next_pc): Likewise.  Caller updated.

8 years agogdbarch software_single_step frame_info to regcache: rs6000
Yao Qi [Tue, 22 Nov 2016 14:05:05 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: rs6000

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* rs6000-aix-tdep.c (branch_dest): Replace parameter frame with
regcache.  Call get_regcache_arch instead of get_frame_arch.
Call regcache_raw_get_unsigned instead of
get_frame_register_unsigned.
(rs6000_software_single_step): Likewise.
* rs6000-tdep.c (ppc_deal_with_atomic_sequence): Call
get_regcache_arch instead of get_frame_arch.  Call
regcache_read_pc instead of get_frame_pc.

8 years agogdbarch software_single_step frame_info to regcache: s390
Yao Qi [Tue, 22 Nov 2016 14:05:05 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: s390

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* s390-linux-tdep.c (s390_software_single_step): Call
get_regcache_arch instead of get_frame_arch.  Call
regcache_read_pc instead of get_frame_pc.

8 years agogdbarch software_single_step frame_info to regcache: sparc
Yao Qi [Tue, 22 Nov 2016 14:05:05 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: sparc

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* sparc-tdep.c (sparc_analyze_control_transfer): Replace parameter
frame with regcache.  Call get_current_frame.
(sparc_software_single_step): Call get_regcache_arch instead of
get_frame_arch.  Call regcache_raw_get_unsigned instead of
get_frame_register_unsigned.

8 years agogdbarch software_single_step frame_info to regcache: nios2
Yao Qi [Tue, 22 Nov 2016 14:05:05 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: nios2

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* nios2-tdep.c (nios2_get_next_pc): Replace parameter frame
with regcache.  Call regcache_raw_get_signed instead of
get_frame_register_unsigned.
(nios2_software_single_step): Call get_regcache_arch
instead of get_frame_arch.

8 years agogdbarch software_single_step frame_info to regcache: moxie
Yao Qi [Tue, 22 Nov 2016 14:05:05 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: moxie

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* moxie-tdep.c (moxie_software_single_step): Call
get_regcache_arch instead of get_frame_arch.  Call
regcache_read_pc instead of get_frame_pc.

8 years agogdbarch software_single_step frame_info to regcache: mips
Yao Qi [Tue, 22 Nov 2016 14:05:05 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: mips

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* mips-tdep.c (mips32_bc1_pc): Replace parameter frame with
regcache.  Call regcache_raw_get_unsigned instead of
get_frame_register_unsigned.
(mips32_next_pc): Likewise.
(micromips_bc1_pc): Likewise.
(micromips_next_pc): Likewise.
(extended_mips16_next_pc): Likewise.
(mips16_next_pc): Likewise.
(mips_next_pc): Likewise.
(mips_software_single_step): Call get_regcache_arch instead
of get_frame_arch.

8 years agogdbarch software_single_step frame_info to regcache: cris
Yao Qi [Tue, 22 Nov 2016 14:05:04 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: cris

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* cris-tdep.c (find_step_target): Replace parameter frame
with regcache.  Call get_regcache_arch instead of
get_frame_arch.  Call regcache_raw_get_unsigned instead of
get_frame_register_unsigned.
(cris_software_single_step): Call get_regcache_arch instead
of get_frame_arch.

8 years agogdbarch software_single_step frame_info to regcache: alpha
Yao Qi [Tue, 22 Nov 2016 14:05:04 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: alpha

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* alpha-tdep.c (alpha_deal_with_atomic_sequence): Call
get_regcache_arch instead of get_frame_arch.  Call
regcache_read_pc instead of get_frame_pc.
(alpha_next_pc): Replace parameter frame with regcache.
Call regcache_raw_get_unsigned instead of
get_frame_register_unsigned.

8 years agogdbarch software_single_step frame_info to regcache: aarch64
Yao Qi [Tue, 22 Nov 2016 14:05:04 +0000 (14:05 +0000)]
gdbarch software_single_step frame_info to regcache: aarch64

Use regcache in software_single_step.

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* aarch64-tdep.c (aarch64_software_single_step): Call
get_regcache_arch instead of get_frame_arch.  Call
regcache_read_pc instead of get_frame_pc.

8 years agoNew regcache_raw_get_signed
Yao Qi [Tue, 22 Nov 2016 14:05:04 +0000 (14:05 +0000)]
New regcache_raw_get_signed

This patch adds a new regcache api regcache_raw_get_signed.

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* regcache.c (regcache_raw_get_signed): New function.
* regcache.h (regcache_raw_get_signed): Declare.

8 years agogas,opcodes: fix hardware capabilities bumping in the sparc assembler.
Jose E. Marchesi [Tue, 22 Nov 2016 12:40:37 +0000 (04:40 -0800)]
gas,opcodes: fix hardware capabilities bumping in the sparc assembler.

When the assembler finds an instruction which is part of a higher
opcode architecture it bumps the current opcode architecture.  For
example:

   $ echo "mwait" | as -bump
   {standard input}: Assembler messages:
   {standard input}:1: Warning: architecture bumped from "v6" to "v9m" on "mwait"

However, when two instructions pertaining to the same opcode
architecture but associated to different SPARC hardware capabilities
are found in the input stream, and no GAS architecture is specified in
the command line, the assembler bangs:

   $ echo "mwait; wr %g0,%g1,%mcdper" | as -bump
   {standard input}: Assembler messages:
   {standard input}:1: Warning: architecture bumped from "v6" to "v9m" on "mwait"
   {standard input}:1: Error: Hardware capability "sparc5" not enabled for "wr".

... and it should'nt, as WRMCDPER pertains to the same architecture
level than MWAIT.

This patch fixes this by extending the definition of sparc opcode
architectures to contain a set of hardware capabilities and making the
assembler to take these capabilities into account when updating the
set of allowed hwcaps when an architecture bump is triggered by some
instruction.

This way, hwcaps associated to architecture levels are maintained in
opcodes, while the assembler keeps the flexibiity of defining GAS
architectures including additional hwcaps (like -Asparcfmaf or the
v8plus* variants).

A test covering this failure case is included.

gas/ChangeLog:

2016-11-22  Jose E. Marchesi  <jose.marchesi@oracle.com>

        * config/tc-sparc.c: Move HWS_* and HWS2_* definitions to
        opcodes/sparc-opc.c.
        (sparc_arch): Clarify the new role of the hwcap_allowed and
        hwcap2_allowed fields.
        (sparc_arch_table): Remove HWS_* and HWS2_* instances from
        hwcap_allowed and hwcap2_allowed respectively.
        (md_parse_option): Include the opcode arch hwcaps when processing
        -A.
        (sparc_ip): Use the current opcode arch hwcaps to update
        hwcap_allowed, as well of the hwcaps of the instruction triggering
        the bump.
        * testsuite/gas/sparc/hwcaps-bump.s: New file.
        * testsuite/gas/sparc/hwcaps-bump.l: Likewise.
        * testsuite/gas/sparc/sparc.exp (gas_64_check): Run tests in
        hwcaps-bump.

include/ChangeLog:

2016-11-22  Jose E. Marchesi  <jose.marchesi@oracle.com>

        * opcode/sparc.h (sparc_opcode_arch): New fields hwcaps and
        hwcaps2.

opcodes/ChangeLog:

2016-11-22  Jose E. Marchesi  <jose.marchesi@oracle.com>

        * sparc-opc.c (HWS_V8): Definition moved from
        gas/config/tc-sparc.c.
        (HWS_V9): Likewise.
        (HWS_VA): Likewise.
        (HWS_VB): Likewise.
        (HWS_VC): Likewise.
        (HWS_VD): Likewise.
        (HWS_VE): Likewise.
        (HWS_VV): Likewise.
        (HWS_VM): Likewise.
        (HWS2_VM): Likewise.
        (sparc_opcode_archs): Initialize hwcaps and hwcaps2 fields of
        existing entries.

8 years ago[ARC] Fix printing 'b' mnemonics.
Claudiu Zissulescu [Wed, 9 Nov 2016 14:30:35 +0000 (15:30 +0100)]
[ARC] Fix printing 'b' mnemonics.

gas/
2016-11-22  Claudiu Zissulescu  <claziss@synopsys.com>

* testsuite/gas/arc/b.d: Update test result.

opcode/
2016-11-22  Claudiu Zissulescu  <claziss@synopsys.com>

* arc-tbl.h: Reorder conditional flags with delay flags for 'b'
instructions.

8 years agoPR20744, Incorrect PowerPC VLE relocs
Alan Modra [Tue, 22 Nov 2016 08:15:29 +0000 (18:45 +1030)]
PR20744, Incorrect PowerPC VLE relocs

VLE 16A and 16D relocs were functionally swapped.

PR 20744
include/
* opcode/ppc.h: Define VLE insns using 16A and 16D relocs.
bfd/
* elf32-ppc.h (struct ppc_elf_params): Add vle_reloc_fixup field.
* elf32-ppc.c: Include opcode/ppc.h.
(ppc_elf_howto_raw): Correct dst_mask for R_PPC_VLE_LO16A,
R_PPC_VLE_LO16D, R_PPC_VLE_HI16A, R_PPC_VLE_HI16D, R_PPC_VLE_HA16A,
R_PPC_VLE_HA16D, R_PPC_VLE_SDAREL_LO16A, R_PPC_VLE_SDAREL_LO16D,
R_PPC_VLE_SDAREL_HI16A, R_PPC_VLE_SDAREL_HI16D,
R_PPC_VLE_SDAREL_HA16A, and R_PPC_VLE_SDAREL_HA16D relocs.
(ppc_elf_link_hash_table_create): Update default_params init.
(ppc_elf_vle_split16): Correct shift and mask.  Add params.
Report or fix insn/reloc mismatches.
(ppc_elf_relocate_section): Pass input_section, offset and fixup
to ppc_elf_vle_split16.
binutils/
* NEWS: Mention PowerPC VLE relocation error.
gas/
* config/tc-ppc.c: Delete VLE insn defines.
(md_assemble): Swap use_a_reloc and use_d_reloc.
* testsuite/gas/ppc/vle-reloc.d: Update.
ld/
* emultempl/ppc32elf.em (params): Update initializer.  Handle
--vle-reloc-fixup command line arg.

8 years agoUse input_bfd in relocate_section
Alan Modra [Tue, 22 Nov 2016 05:57:12 +0000 (16:27 +1030)]
Use input_bfd in relocate_section

It makes just a little more sense to use input_bfd when retrieving
insns for relocation, since the relocations match the endianness of
the input bfd.

* elf32-ppc.c (ppc64_elf_relocate_section): Calculate d_offset for
input_bfd.  Replace occurrences of output_bfd as bfd_get_32 and
bfd_put_32 param with input_bfd.
* elf32-ppc.c (ppc_elf_relocate_section): Likewise.  Also
ppc_elf_vle_split16 param.
(ppc_elf_vle_split16): Rename output_bfd param to input_bfd.

8 years agoUse VALUE_NEXT_FRAME_ID in value_from_component
Yao Qi [Tue, 22 Nov 2016 08:53:34 +0000 (08:53 +0000)]
Use VALUE_NEXT_FRAME_ID in value_from_component

We renamed VALUE_FRAME_ID to VALUE_NEXT_FRAME_ID recently,
https://sourceware.org/ml/gdb-patches/2016-11/msg00018.html
and we should use VALUE_NEXT_FRAME_ID in value_from_component
too.

gdb:

2016-11-22  Yao Qi  <yao.qi@linaro.org>

* value.c (value_from_component): Use VALUE_NEXT_FRAME_ID
instead of VALUE_FROM_ID.

8 years agoAutomatic date update in version.in
GDB Administrator [Tue, 22 Nov 2016 00:00:15 +0000 (00:00 +0000)]
Automatic date update in version.in

8 years agoFix bug where -u option with empty archive results in internal error.
Cary Coutant [Mon, 21 Nov 2016 22:05:40 +0000 (14:05 -0800)]
Fix bug where -u option with empty archive results in internal error.

gold/
PR gold/20693
* gold.cc (queue_middle_tasks): Force valid target earlier.

8 years agoAdd missing POSTCOMPILE step to mi/ file generation rules
Simon Marchi [Mon, 21 Nov 2016 21:05:57 +0000 (16:05 -0500)]
Add missing POSTCOMPILE step to mi/ file generation rules

A little oversight from my part, it caused the Makefile not to track
the dependencies from mi/*.c files.

gdb/ChangeLog:

* Makefile.in (%o: $(srcdir)/mi/%.c): Add missing POSTCOMPILE
step.

8 years agoAdd --build-id=uuid support for MinGW32.
Igor Kudrin [Mon, 21 Nov 2016 17:59:37 +0000 (09:59 -0800)]
Add --build-id=uuid support for MinGW32.

2016-11-21  Igor Kudrin  <ikudrin@accesssoftek.com>

gold/
* layout.cc: Include windows.h and rpcdce.h (for MinGW32).
(Layout::create_build_id): Generate uuid using UuidCreate().

8 years agoBFD/DWARF2: Correct an `index' global shadowing error
Maciej W. Rozycki [Mon, 21 Nov 2016 15:59:42 +0000 (15:59 +0000)]
BFD/DWARF2: Correct an `index' global shadowing error

Fix a commit 089e3718bd8d ("Greatly improve the speed if looking up
DWARF line number information.") build regression:

cc1: warnings being treated as errors
.../bfd/dwarf2.c: In function 'build_line_info_table':
.../bfd/dwarf2.c:1614: warning: declaration of 'index' shadows a global declaration
/usr/include/string.h:304: warning: shadowed declaration is here
.../bfd/dwarf2.c: In function 'build_lookup_funcinfo_table':
.../bfd/dwarf2.c:2262: warning: declaration of 'index' shadows a global declaration
/usr/include/string.h:304: warning: shadowed declaration is here
make[4]: *** [dwarf2.lo] Error 1

in a way following commit 91d6fa6a035c ("Add -Wshadow to the gcc command
line options used when compiling the binutils.").

bfd/
* dwarf2.c (build_line_info_table): Rename `index' local
variable to `line_index'.
(build_lookup_funcinfo_table): Rename `index' local variable to
`func_index'.

8 years agoCreate subobject value in pretty printer
Yao Qi [Mon, 21 Nov 2016 14:15:06 +0000 (14:15 +0000)]
Create subobject value in pretty printer

Nowadays, we create a value of subobject in pretty printer with 'address'
being used,

  value = value_from_contents_and_address (type, valaddr + embedded_offset,
   address + embedded_offset);

  set_value_component_location (value, val);
  /* set_value_component_location resets the address, so we may
     need to set it again.  */
  if (VALUE_LVAL (value) != lval_internalvar
      && VALUE_LVAL (value) != lval_internalvar_component
      && VALUE_LVAL (value) != lval_computed)
    set_value_address (value, address + embedded_offset);

value_from_contents_and_address creates a value from memory, but the
value we are pretty-printing may not from memory at all.

Instead of using value_from_contents_and_address, we create a value
of subobject with the same location as object's but different offset.
We avoid using address in this way.  As a result, parameter 'address'
in apply_val_pretty_printer is no longer needed, we can remove it in
next step.

We've already had the location of the 'whole' value, so it is safe
to assume we can create a value of 'component' or 'suboject' value
at the same location but with different offset.

gdb:

2016-11-21  Yao Qi  <yao.qi@linaro.org>

* guile/scm-pretty-print.c (gdbscm_apply_val_pretty_printer):
Don't call value_from_contents_and_address and
set_value_address.  Call value_from_component.
* python/py-prettyprint.c (gdbpy_apply_val_pretty_printer):
Likewise.
* value.c (value_from_component): New function.
* value.h (value_from_component): Likewise.
* valarith.c (value_subscripted_rvalue): Call
value_from_component.

8 years ago[GAS][ARM][PR20827]Fix gas error for two register form instruction (pre-UAL syntax).
Renlin Li [Mon, 21 Nov 2016 12:06:04 +0000 (12:06 +0000)]
[GAS][ARM][PR20827]Fix gas error for two register form instruction (pre-UAL syntax).

gas/

2016-11-21  Renlin Li  <renlin.li@arm.com>

PR gas/20827
* config/tc-arm.c (encode_arm_shift): Don't assert for operands not
presented.
* testsuite/gas/arm/add-shift-two.d: New.
* testsuite/gas/arm/add-shift-two.s: New.

8 years agoUse ACX_PROG_CMP_IGNORE_INITIAL in gas
Alan Modra [Mon, 21 Nov 2016 09:48:41 +0000 (20:18 +1030)]
Use ACX_PROG_CMP_IGNORE_INITIAL in gas

* configure.ac: Invoke ACX_PROG_CMP_IGNORE_INITIAL.
* Makefile.am (comparison): Rewrite using do_compare.
* configure: Regenerate.
* Makefile.in: Regenerate.
* doc/Makefile.in: Regenerate.

8 years agoAutomatic date update in version.in
GDB Administrator [Mon, 21 Nov 2016 00:00:33 +0000 (00:00 +0000)]
Automatic date update in version.in

8 years agoAutomatic date update in version.in
GDB Administrator [Sun, 20 Nov 2016 00:00:28 +0000 (00:00 +0000)]
Automatic date update in version.in

8 years agoARI: Add detection of printf_vma and sprintf_vma
Joel Brobecker [Sat, 19 Nov 2016 18:40:17 +0000 (10:40 -0800)]
ARI: Add detection of printf_vma and sprintf_vma

We shouldn't be using these, since their output goes straight to
stdout, which doesn't allow redirection. So this patch updates
the ARI to detect any such use.

gdb/ChangeLog:

        * contrib/ari/gdb_ari.sh: Add detection of printf_vma and
        sprintf_vma.

8 years agoRevert "bfd: allow negative offsets to _GLOBAL_OFFSET_TABLE_ in elf64 SPARC"
Jose E. Marchesi [Sat, 19 Nov 2016 12:39:09 +0000 (04:39 -0800)]
Revert "bfd: allow negative offsets to _GLOBAL_OFFSET_TABLE_ in elf64 SPARC"

This reverts commit b19753ce31da347605dfa903c6fd2158e2444f0d.

As it turns out, GCC (and the assembler) needs additional work in
order to support negative GOT offsets in 64-bit sparc.  This is
breaking TLS Local Dynamic in position-independent code.

8 years agoMakefile: fix typo
Simon Marchi [Sat, 19 Nov 2016 02:16:52 +0000 (21:16 -0500)]
Makefile: fix typo

Thanks to Patrick Monnerat for reporting this typo.

gdb/ChangeLog:

* Makefile.in (%.o: $(srcdir)/gdbtk/generic/%.c): Fix typo.

8 years agoAutomatic date update in version.in
GDB Administrator [Sat, 19 Nov 2016 00:00:32 +0000 (00:00 +0000)]
Automatic date update in version.in

8 years agogdb/doc: Add missing comma after xref
Andreas Arnez [Fri, 18 Nov 2016 15:59:00 +0000 (16:59 +0100)]
gdb/doc: Add missing comma after xref

Get rid of a warning for missing punctuation after xref.

gdb/doc/ChangeLog:

* gdb.texinfo (GDB/MI Async Records): Add missing comma after
xref.

8 years ago[ARC] Fix and extend features of .cpu directive.
Claudiu Zissulescu [Thu, 17 Nov 2016 12:26:54 +0000 (13:26 +0100)]
[ARC] Fix and extend features of .cpu directive.

gas/
2016-11-18  Claudiu Zissulescu  <claziss@synopsys.com>

* testsuite/gas/arc/cl-warn.s: New file.
* testsuite/gas/arc/cpu-pseudop-1.d: Likewise.
* testsuite/gas/arc/cpu-pseudop-1.s: Likewise.
* testsuite/gas/arc/cpu-pseudop-2.d: Likewise.
* testsuite/gas/arc/cpu-pseudop-2.s: Likewise.
* testsuite/gas/arc/cpu-warn2.s: Likewise.
* config/tc-arc.c (selected_cpu): Initialize.
(feature_type): New struct.
(feature_list): New variable.
(arc_check_feature): New function.
(arc_select_cpu): Check for .cpu duplicates. Don't overwrite the
current cpu features. Check if a feature is available for a given
cpu.
(md_parse_option): Test if features are available for a given cpu.

8 years agobfd: fix negative GOT offsets for non-local references on sparc64
James Clarke [Fri, 18 Nov 2016 11:51:40 +0000 (12:51 +0100)]
bfd: fix negative GOT offsets for non-local references on sparc64

bfd/ChangeLog:

2016-11-18  James Clarke  <jrtc27@jrtc27.com>

* elfxx-sparc.c (_bfd_sparc_elf_relocate_section): Don't convert
R_SPARC_GOTDATA_OP_HIX22 and R_SPARC_GOTDATA_OP_LOX10 to
R_SPARC_GOT* for non-local references. Instead, treat them like
R_SPARC_GOTDATA_HIX22/R_SPARC_GOTDATA_LOX10 when filling in the
immediate with the calculated relocation.