binutils-gdb.git
7 years agoHandle symbol defined in IR and referenced in DSO
H.J. Lu [Thu, 20 Apr 2017 14:48:24 +0000 (07:48 -0700)]
Handle symbol defined in IR and referenced in DSO

We need to make an IR symbol visible if it is defined in an IR object
and referenced in a dynamic object.  When --as-needed is used, since
linker removes the IR symbol reference of the dynamic object if the
dynamic object isn't needed in the first pass, the IR definition isn't
visible to the dynamic object even if the dynamic object becomes needed
in the second pass.  Add dynamic_ref_after_ir_def to bfd_link_hash_entry
to track IR symbol which is defined in an IR object and later referenced
in a dynamic object.  dynamic_ref_after_ir_def is preserved when restoring
the symbol table for unneeded dynamic object.

bfd/

PR ld/21382
* elflink.c (elf_link_add_object_symbols): Preserve
dynamic_ref_after_ir_def when restoring the symbol table for
unneeded dynamic object.

include/

PR ld/21382
* bfdlink.h (bfd_link_hash_entry): Add dynamic_ref_after_ir_def.

ld/

PR ld/21382
* plugin.c (is_visible_from_outside): Symbol may be visible
from outside if dynamic_ref_after_ir_def is set.
(plugin_notice): Set dynamic_ref_after_ir_def if the symbol is
defined in an IR object and referenced in a dynamic object.
* testsuite/ld-plugin/lto.exp: Run PR ld/21382 tests.
* testsuite/ld-plugin/pr21382a.c: New file.
* testsuite/ld-plugin/pr21382b.c: Likewise.

7 years agoAutomatic date update in version.in
GDB Administrator [Thu, 20 Apr 2017 00:00:41 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agox86-64: Handle undefined IFUNC symbol with -z now
H.J. Lu [Wed, 19 Apr 2017 16:59:14 +0000 (09:59 -0700)]
x86-64: Handle undefined IFUNC symbol with -z now

Since undefined IFUNC symbol is treated as normal FUNC symbol, don't
abort on undefined IFUNC symbol in the second PLT.

bfd/

PR ld/21401
* elf64-x86-64.c (elf_x86_64_finish_dynamic_symbol): Don't abort
on on undefined IFUNC symbol in the second PLT.

ld/

PR ld/21401
* testsuite/ld-ifunc/ifunc.exp: Add a libtest-2-now.so test with
-z now.

7 years agogdbthread.h: Fix comment typo
Pedro Alves [Wed, 19 Apr 2017 12:16:05 +0000 (13:16 +0100)]
gdbthread.h: Fix comment typo

gdb/ChangeLog:
2017-04-19  Pedro Alves  <palves@redhat.com>

* gdbthread.h (thread): Add missing closing parenthesis in
comment.

7 years agoFix removing inferiors from within "thread apply" commands
Pedro Alves [Wed, 19 Apr 2017 12:12:23 +0000 (13:12 +0100)]
Fix removing inferiors from within "thread apply" commands

This patch fixes an internal error exposed by a test that does
something like:

  define kill-and-remove
    kill inferiors 2
    remove-inferiors 2
  end

  # Start one inferior.
  start

  # Start another inferior.
  add-inferior 2
  inferior 2
  start

  # Kill and remove inferior 1 while inferior 2 is selected.
  thread apply 1.1 kill-and-remove

The internal error looks like this:

 Thread 1.1 (Thread 0x7ffff7fc2700 (LWP 20677)):
 [Switching to inferior 1 [process 20677] (gdb/testsuite/outputs/gdb.threads/threadapply/threadapply)]
 [Switching to thread 1.1 (Thread 0x7ffff7fc2700 (LWP 20677))]
 #0  main () at src/gdb/testsuite/gdb.threads/threadapply.c:38
 38          for (i = 0; i < NUM; i++)
 src/gdb/inferior.c:66: internal-error: void set_current_inferior(inferior*): Assertion `inf != NULL' failed.
 A problem internal to GDB has been detected,
 further debugging may prove unreliable.
 Quit this debugging session? (y or n) FAIL: gdb.threads/threadapply.exp: kill_and_remove_inferior: try kill-and-remove: thread apply 1.1 kill-and-remove (GDB internal error)

There are several problems around this area of the code.  One is that
in do_restore_current_thread_cleanup, we do a look up of inferior by
ptid, which can find the wrong inferior if the previously selected
inferior exited and some other inferior was started with a reused pid
(rare, but still...).

The other problem is that the "remove-inferiors" command rejects
attempts to remove the current inferior, but when we get to
"remove-inferiors" in a "thread apply THR remove-inferiors 2" command,
the current inferior is the inferior of thread THR, not the previously
selected inferior, so if the previously selected inferior was inferior
2, that command still manages to wipe it, and then gdb restores the
old selected inferior, which is now a dangling pointer...

So the fix here is:

- Make make_cleanup_restore_current_thread store a pointer to the
  previously selected inferior directly, and use it directly instead
  of doing ptid look ups.

- Add a refcount to inferiors, very similar to thread_info's refcount,
  that is incremented/decremented by
  make_cleanup_restore_current_thread, and checked before deleting an
  inferior.  To avoid duplication, a new refcounted_object type is
  added, that both thread_info and inferior inherit from.

gdb/ChangeLog:
2017-04-19  Pedro Alves  <palves@redhat.com>

* common/refcounted-object.h: New file.
* gdbthread.h: Include "common/refcounted-object.h".
(thread_info): Inherit from refcounted_object and add comments.
(thread_info::incref, thread_info::decref)
(thread_info::m_refcount): Delete.
(thread_info::deletable): Use the refcounted_object::refcount()
method.
* inferior.c (current_inferior_): Add comment.
(set_current_inferior): Increment/decrement refcounts.
(prune_inferiors, remove_inferior_command): Skip inferiors marked
not-deletable instead of comparing with the current inferior.
(initialize_inferiors): Increment the initial inferior's refcount.
* inferior.h (struct inferior): Forward declare.
Include "common/refcounted-object.h".
(current_inferior, set_current_inferior): Move declaration to
before struct inferior's definition, and fix comment.
(inferior): Inherit from refcounted_object.  Add comments.
* thread.c (switch_to_thread_no_regs): Reference the thread's
inferior pointer directly instead of doing a ptid lookup.
(switch_to_no_thread): New function.
(switch_to_thread(thread_info *)): New function, factored out
from ...
(switch_to_thread(ptid_t)): ... this.
(restore_current_thread): Delete.
(current_thread_cleanup): Remove 'inf_id' and 'was_removable'
fields, and add 'inf' field.
(do_restore_current_thread_cleanup): Check whether old->inf is
alive instead of looking up an inferior by ptid.  Use
switch_to_thread and switch_to_no_thread.
(restore_current_thread_cleanup_dtor): Use old->inf directly
instead of lookup up an inferior by id.  Decref the inferior.
Don't restore 'removable'.
(make_cleanup_restore_current_thread): Same the inferior pointer
in old, instead of the inferior number.  Incref the inferior.
Don't save/clear 'removable'.

gdb/testsuite/ChangeLog:
2017-04-19  Pedro Alves  <palves@redhat.com>

* gdb.threads/threadapply.exp (kill_and_remove_inferior): New
procedure.
(top level): Call it.
* lib/gdb.exp (gdb_define_cmd): New procedure.

7 years agoMake inferior::detaching a bool, and introduce scoped_restore::release()
Pedro Alves [Wed, 19 Apr 2017 12:12:23 +0000 (13:12 +0100)]
Make inferior::detaching a bool, and introduce scoped_restore::release()

I left making inferior::detaching a bool to a separate patch, because
doing that makes a make_cleanup_restore_integer call in
infrun.c:prepare_for_detach no longer compile (passing a 'bool *' when
an 'int *' is expected).  Since we want to get rid of cleanups anyway,
I looked at converting that to a scoped_restore.  However,
prepare_for_detach wants to discard the cleanup on success, and
scoped_restore doesn't have an equivalent for that.  So I added one --
I called it "release()" because it seems like a natural fit in the way
standard components call similarly-spirited methods, and, it's also
what the proposal for a generic scope guard calls it too, AFAICS:

  http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2014/n4189.pdf

I've added some scoped_guard unit tests, while at it.

gdb/ChangeLog:
2017-04-19  Pedro Alves  <palves@redhat.com>

* Makefile.in (SUBDIR_UNITTESTS_SRCS): Add
unittests/scoped_restore-selftests.c.
(SUBDIR_UNITTESTS_OBS): Add scoped_restore-selftests.o.
* common/scoped_restore.h (scoped_restore_base): Make "class".
(scoped_restore_base::release): New public method.
(scoped_restore_base::scoped_restore_base): New protected ctor.
(scoped_restore_base::m_saved_var): New protected field.
(scoped_restore_tmpl::scoped_restore_tmpl(T*)): Initialize the
scoped_restore_base base class instead of m_saved_var directly.
(scoped_restore_tmpl::scoped_restore_tmpl(T*, T2)): Likewise.
(scoped_restore_tmpl::scoped_restore_tmpl(const
scoped_restore_tmpl<T>&)): Likewise.
(scoped_restore_tmpl::~scoped_restore_tmpl): Use the saved_var
method.
(scoped_restore_tmpl::saved_var): New method.
(scoped_restore_tmpl::m_saved_var): Delete.
* inferior.h (inferior::detaching): Now a bool.
* infrun.c (prepare_for_detach): Use a scoped_restore instead of a
cleanup.
* unittests/scoped_restore-selftests.c: New file.

7 years agogdb/Makefile.in: Re-sort SUBDIR_UNITTESTS_SRCS/SUBDIR_UNITTESTS_OBS
Pedro Alves [Wed, 19 Apr 2017 12:06:20 +0000 (13:06 +0100)]
gdb/Makefile.in: Re-sort SUBDIR_UNITTESTS_SRCS/SUBDIR_UNITTESTS_OBS

Note to self: 'o' before 'p'.

gdb/ChangeLog:
2017-04-19  Pedro Alves  <palves@redhat.com>

* Makefile.in (SUBDIR_UNITTESTS_SRCS, SUBDIR_UNITTESTS_OBS):
Re-sort in alphabetic order.

7 years agoCorrect PE reloc display
Wedson Almeida Filho [Wed, 19 Apr 2017 08:32:53 +0000 (18:02 +0930)]
Correct PE reloc display

* peXXigen.c (pe_print_reloc): Correct chunk_end.

7 years agoImplement -z dynamic-undefined-weak
Alan Modra [Tue, 18 Apr 2017 15:56:57 +0000 (01:26 +0930)]
Implement -z dynamic-undefined-weak

-z nodynamic-undefined-weak is only implemented for x86.  (The sparc
backend has some support code but doesn't enable the option by
including ld/emulparams/dynamic_undefined_weak.sh, and since the
support looks like it may be broken I haven't enabled it.)  This patch
adds the complementary -z dynamic-undefined-weak, extends both options
to affect building of shared libraries as well as executables, and
adds support for the option on powerpc.

include/
* bfdlink.h (struct bfd_link_info <dynamic_undefined_weak>):
Revise comment.
bfd/
* elflink.c (_bfd_elf_adjust_dynamic_symbol): Hide undefweak
or make dynamic for info->dynamic_undefined_weak 0 and 1.
* elf32-ppc.c:Formatting.
(ensure_undefweak_dynamic): Don't make dynamic when
info->dynamic_undefined_weak is zero.
(allocate_dynrelocs): Discard undefweak dyn_relocs for
info->dynamic_undefined_weak.  Discard undef dyn_relocs when
not default visibility.  Discard undef and undefweak
dyn_relocs earlier.
(ppc_elf_relocate_section): Adjust to suit.
* elf64-ppc.c: Formatting.
(ensure_undefweak_dynamic): Don't make dynamic when
info->dynamic_undefined_weak is zero.
(allocate_dynrelocs): Discard undefweak dyn_relocs for
info->dynamic_undefined_weak.  Discard them earlier.
ld/
* ld.texinfo (dynamic-undefined-weak): Document.
(nodynamic-undefined-weak): Document that this option now can
be used with shared libs.
* emulparams/dynamic_undefined_weak.sh: Support -z
dynamic-undefined-weak.
* emulparams/elf32ppccommon.sh: Include dynamic_undefined_weak.sh.
* testsuite/ld-undefined/weak-undef.exp (undef_weak_so),
(undef_weak_exe): New.  Use them.  Add -z dynamic-undefined-weak
and -z nodynamic-undefined-weak tests.
* Makefile.am: Update powerpc dependencies.
* Makefile.in: Regenerate.

7 years agoAutomatic date update in version.in
GDB Administrator [Wed, 19 Apr 2017 00:00:40 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoEliminate obstack_printf
Pedro Alves [Tue, 18 Apr 2017 20:39:25 +0000 (21:39 +0100)]
Eliminate obstack_printf

Not used anywhere.

gdb/ChangeLog:
2017-04-18  Pedro Alves  <palves@redhat.com>

* xml-support.c (obstack_xml_printf): Delete.
* xml-support.h (obstack_xml_printf): Delete.

7 years agogdb_xml_parser: make data fields private and make more functions methods
Pedro Alves [Tue, 18 Apr 2017 20:39:25 +0000 (21:39 +0100)]
gdb_xml_parser: make data fields private and make more functions methods

This patch makes the data fields of gdb_xml_parser private, and makes
more functions be gdb_xml_parser methods.  This is mostly for better
encapsulation.

Some free functions have their parsing-related guts converted to
methods, while the free functions remain, as they're used as expat
callbacks.  Now their only job is to be small shims that restore back
the gdb_xml_parser type, defer work to the corresponding method, and
make sure C++ exceptions don't cross expat.

More C++-fycation of the XML parsers built on top of gdb_xml_parser
could follow, but this was my stopping point.

gdb/ChangeLog:
2017-04-18  Pedro Alves  <palves@redhat.com>

* xml-support.c (gdb_xml_parser) <use_dtd, dtd_name, parse,
vdebug, verror, body_text, start_element, end_element, name,
user_data, set_is_xinclude, set_error, expat_parser>: New methods.
<name, user_data, expat_parser, scopes, error, last_line, dtd_name,
is_xinclude>: Make private and add m_ prefix.
(gdb_xml_parser::body_text): New method, based on ...
(gdb_xml_body_text): ... this.  Adjust.
(gdb_xml_parser::vdebug): New method, based on ...
(gdb_xml_debug): ... this.  Adjust.
(gdb_xml_parser::verror): New method, based on ...
(gdb_xml_error): ... this.  Adjust.
(gdb_xml_parser::start_element): New method, based on ...
(gdb_xml_start_element): ... this.  Adjust.
(gdb_xml_start_element_wrapper): Defer to
gdb_xml_parser::start_element and gdb_xml_parser::set_error.
(gdb_xml_parser::end_element): New method, based on ...
(gdb_xml_end_element_wrapper): ... this.  Adjust.
(gdb_xml_parser::~gdb_xml_parser): Adjust.
(gdb_xml_parser::gdb_xml_parser): Adjust to field renames.
(gdb_xml_parser::use_dtd): New method, based on ...
(gdb_xml_use_dtd): ... this.  Adjust.
(gdb_xml_parser::parse): New method, based on ...
(gdb_xml_parse): ... this.  Adjust.
(gdb_xml_parse_quick): Adjust to call the parser's parse method.
(xinclude_start_include): Adjust to call the parser's name method.
(xml_xinclude_default, xml_xinclude_start_doctype)
(xml_xinclude_end_doctype): Adjust to call the parser's user_data
method.
(xml_process_xincludes): Adjust to call parser methods.
* xml-support.h (gdb_xml_use_dtd, gdb_xml_parse): Delete
declarations.

7 years agoxml-support.c: Use std::string for growing string buffer
Pedro Alves [Tue, 18 Apr 2017 20:39:25 +0000 (21:39 +0100)]
xml-support.c: Use std::string for growing string buffer

This main idea behind this patch is this change to xml-support.c:scope_level

  -  /* Body text accumulation.  This is an owning pointer.  */
  -  struct obstack *body;
  +  /* Body text accumulation.  */
  +  std::string body;

... which allows simplifying other parts of the code.

In target_fetch_description_xml, we want to distinguish between
returning "success + empty std::string" and "no success", and
gdb::optional is a natural fit for that.

gdb/ChangeLog:
2017-04-18  Pedro Alves  <palves@redhat.com>

* tracefile-tfile.c (tfile_write_tdesc): Adjust to use
gdb::optional<std::string>.
* xml-support.c: Include <string>.
(scope_level::scope_level(scope_level &&))
(scope_level::~scope_level): Delete.
(scope_level::body): Now a std::string.
(gdb_xml_body_text, gdb_xml_end_element): Adjust.
(xinclude_parsing_data::xinclude_parsing_data): Add 'output'
parameter.
(xinclude_parsing_data::~xinclude_parsing_data): Delete.
(xinclude_parsing_data::output): Now a std::string reference.
(xinclude_start_include): Adjust.
(xml_xinclude_default): Adjust.
(xml_process_xincludes): Add 'output' parameter, and return bool.
* xml-support.h (xml_process_xincludes): Add 'output' parameter,
and return bool.
* xml-tdesc.c: Include <unordered_map> and <string>.
(tdesc_xml_cache): Delete.
(tdesc_xml_cache_s): Delete.
(xml_cache): Now an std::unordered_map.
(tdesc_parse_xml): Adjust to use std::string and unordered_map.
(target_fetch_description_xml): Change return type to
gdb::optional<std::string>, and adjust.
* xml-tdesc.h: Include "common/gdb_optional.h" and <string>.
(target_fetch_description_xml): Change return type to
gdb::optional<std::string>.

7 years agogdb::optional unit tests
Pedro Alves [Tue, 18 Apr 2017 20:39:24 +0000 (21:39 +0100)]
gdb::optional unit tests

I thought I'd add some unit tests to make sure gdb::optional behaved
correctly, and started writing some, but then thought/realized that
libstdc++ already has extensive testing for C++17 std::optional, which
gdb::optional is a subset of, and thought why bother writing something
from scratch.  So I tried copying over a subset of libstdc++'s tests
(that ones that cover the subset supported by gdb::optional), and was
positively surprised that they mostly work OOTB.  This did help shake
out a few bugs from what I was implementing in the previous patch to
gdb::optional.  Still, it's a good chunk of code being copied over, so
if people dislike this copying/duplication, I can drop this patch.

gdb/ChangeLog:
2017-04-18  Pedro Alves  <palves@redhat.com>

* Makefile.in (SUBDIR_UNITTESTS_SRCS): Add
unittests/optional-selftests.c.
(SUBDIR_UNITTESTS_OBS): Add optional-selftests.o.
* unittests/optional-selftests.c: New file.
* unittests/optional/assignment/1.cc: New file.
* unittests/optional/assignment/2.cc: New file.
* unittests/optional/assignment/3.cc: New file.
* unittests/optional/assignment/4.cc: New file.
* unittests/optional/assignment/5.cc: New file.
* unittests/optional/assignment/6.cc: New file.
* unittests/optional/assignment/7.cc: New file.
* unittests/optional/cons/copy.cc: New file.
* unittests/optional/cons/default.cc: New file.
* unittests/optional/cons/move.cc: New file.
* unittests/optional/cons/value.cc: New file.
* unittests/optional/in_place.cc: New file.
* unittests/optional/observers/1.cc: New file.
* unittests/optional/observers/2.cc: New file.

7 years agoMore gdb::optional features
Pedro Alves [Tue, 18 Apr 2017 20:39:24 +0000 (21:39 +0100)]
More gdb::optional features

Currently we can't use gdb::optional<T> as function return type,
because gdb::optional's copy ctor is deleted.  For example, with:

  gdb::optional<int> function ()
  {
    gdb::optional<int> opt;
    ....
    return opt;

we get:

  src/gdb/foo.c: In function ‘gdb::optional<int> foo()’:
  src/gdb/foo.c:75:10: error: use of deleted function ‘gdb::optional<T>::optional(const gdb::optional<T>&) [with T = int]’
     return opt;
    ^
  In file included from src/gdb/foo.c:68:0:
  src/gdb/common/gdb_optional.h:53:3: note: declared here
     optional (const optional &other) = delete;
     ^

I started by fixing that, and then ran into another missing feature,
also fixed by this patch.

The next feature I'm missing most from gdb::optional<T> compared to
std::optional<T> is construction/move/assignment from a T, instead of
having to default construct an gdb::optional and then use
optional::emplace(....).

For example:
  gdb::optional<std::string> function ()
  {
    gdb::optional<std::string> opt;
    std::string str;
    ...
    opt.emplace (std::move (str));
    return opt;
vs
  gdb::optional<std::string> function ()
  {
    std::string str;
    ...
    return str;

The copy/move ctor/assign methods weren't initialy implemented because
std::optional supports construction from a type U if U is convertible
to T too, and has rules to decide whether the ctors are
explicit/implicit based on that, and rules for whether the ctor should
be trivial or not, etc., which leads to a much more complicated
implementation.

If we stick to supporting copy/move construction/assignment of/to an
optional<T> from exactly only optional<T> and T, then all that
conversion-related complication disappears, and we still gain
convenience in most use cases.

The patch also makes emplace return a reference to the constructor
object, per C++17 std::optional, and adds a reset method, againt
because std::optional has one and it's trivial to support it.  These
two changes are a requirement of the gdb::optional unit testing patch
that will follow.

gdb/ChangeLog:
2017-04-18  Pedro Alves  <palves@redhat.com>

* common/gdb_optional.h: Include common/traits.h.
(in_place_t): New type.
(in_place): New constexpr variable.
(optional::optional): Remove member initialization of
m_instantiated.
(optional::optional(in_place_t...)): New constructor.
(optional::~optional): Use reset.
(optional::optional(const optional&)): New.
(optional::optional(const optional&&)): New.
(optional::optional(T &)): New.
(optional::optional(T &&)): New.
(operator::operator=(const optional &)): New.
(operator::operator=(optional &&)): New.
(operator::operator= (const T &))
(operator::operator= (T &&))
(operator::emplace (Args &&... args)): Return a T&.  Use reset.
(operator::reset): New.
(operator::m_instantiated):: Add in-class initializer.
* common/traits.h: Include <type_traits>.
(struct And): New types.

7 years agoxml-support.c: Use std::vector
Pedro Alves [Tue, 18 Apr 2017 20:39:24 +0000 (21:39 +0100)]
xml-support.c: Use std::vector

scope_level::scope_level needed both a move ctor and a dtor explicitly
coded, but those will be eliminated in a following patch.

gdb/ChangeLog:
2017-04-18  Pedro Alves  <palves@redhat.com>

* xml-support.c: Include <vector>.
(scope_level::scope_level(const gdb_xml_element *))
(scope_level::scope_level(scope_level&&)): New.
(scope_level::~scope_level): New.
(scope_level_s): Delete.
(gdb_xml_parser::scopes): Now a std::vector.
(gdb_xml_body_text, gdb_xml_start_element, gdb_xml_end_element):
Use std::vector.
(gdb_xml_parser::~gdb_xml_parser): Remove now unnecessary
scope cleanup code.
(gdb_xml_parser::gdb_xml_parser): Remove explicit initialization
of the scopes member.  Use std::vector::emplace_back.

7 years agoC++-ify gdb/xml-support.c a bit to eliminate cleanups
Pedro Alves [Tue, 18 Apr 2017 20:39:24 +0000 (21:39 +0100)]
C++-ify gdb/xml-support.c a bit to eliminate cleanups

Basically convert cleanups to destructors in gdb_xml_parser and
xinclude_parsing_data, and then allocate objects of those types on the
stack.

More C++-ification is possible / will follow, but this removes a few
make_cleanup calls already.

gdb/ChangeLog:
2017-04-18  Pedro Alves  <palves@redhat.com>

* xml-support.c (gdb_xml_parser): Add ctor/dtor.  Make is_xinclude
a bool.
(gdb_xml_end_element): Change type of first parameter.
(gdb_xml_cleanup): Rename to ...
(gdb_xml_parser::~gdb_xml_parser): ... this.
(gdb_xml_create_parser_and_cleanup): Delete with ...
(gdb_xml_parser::gdb_xml_parser): ... creation parts factored out
to this new ctor.
(gdb_xml_parse_quick): Create a local gdb_xml_parser instead of
using gdb_xml_create_parser_and_cleanup.
(xinclude_parsing_data): Add ctor/dtor.
(xml_xinclude_cleanup): Delete.
(xml_process_xincludes): Create a local xinclude_parsing_data
instead of heap-allocating one.  Create a local gdb_xml_parser
instead of heap-allocating one with
gdb_xml_create_parser_and_cleanup.

7 years agoPR threads/20743: Don't attempt to suspend or resume exited threads.
John Baldwin [Tue, 18 Apr 2017 16:44:32 +0000 (09:44 -0700)]
PR threads/20743: Don't attempt to suspend or resume exited threads.

When resuming a native FreeBSD process, ignore exited threads when
suspending/resuming individual threads prior to continuing the process.

gdb/ChangeLog:

PR threads/20743
* fbsd-nat.c (resume_one_thread_cb): Remove.
(resume_all_threads_cb): Remove.
(fbsd_resume): Use ALL_NON_EXITED_THREADS instead of
iterate_over_threads.

7 years agoAutomatic date update in version.in
GDB Administrator [Tue, 18 Apr 2017 00:00:36 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoPut soname in the version definition section
H.J. Lu [Mon, 17 Apr 2017 14:47:17 +0000 (07:47 -0700)]
Put soname in the version definition section

commit 902e9fc76a0ec9f642cefa71ef88cca1c675ad54
Author: Maciej W. Rozycki <macro@imgtec.com>
Date:   Tue Feb 21 01:46:42 2017 +0000

    PR ld/20828: Move symbol version processing ahead of GC symbol sweep

breaks version definition with --version-script --soname.  This patch
fixes it by getting soname index before generating the version definition
section.

bfd/

PR ld/21389
* elflink.c (bfd_elf_size_dynamic_sections): Get soname index
before generating the version definition section.

ld/

PR ld/21389
* testsuite/ld-elf/pr21389.map: New file.
* testsuite/ld-elf/pr21389.s: Likewise.
* testsuite/ld-elf/pr21389a.d: Likewise.
* testsuite/ld-elf/pr21389b.d: Likewise.
* testsuite/ld-elf/pr21389c.d: Likewise.

7 years agoUpdate NEWS post GDB 8.0 branch creation.
Joel Brobecker [Mon, 17 Apr 2017 14:30:29 +0000 (07:30 -0700)]
Update NEWS post GDB 8.0 branch creation.

gdb/ChangeLog:

* NEWS: Create a new section for the next release branch.
Rename the section of the current branch, now that it has
been cut.

7 years agoBump version to 8.0.50.DATE-git
Joel Brobecker [Mon, 17 Apr 2017 14:20:49 +0000 (07:20 -0700)]
Bump version to 8.0.50.DATE-git

Now that the GDB 8.0 branch has been created, we should bump
the GDB version accordingly.

gdb/ChangeLog:

GDB 8.0 branch created (725bf5cf125783c2a7ca4ab63d3768e220bab2db):
* version.in: Bump version to 7.99.90.DATE-git.

7 years agoUndo dynamic symbol state after regular object sym type mismatch
Alan Modra [Mon, 17 Apr 2017 09:03:52 +0000 (18:33 +0930)]
Undo dynamic symbol state after regular object sym type mismatch

We already handle the case of an object file first defining a symbol
that a later shared library also defines where the symbol types are
incompatible.  This patch handles the reverse, when a later object
file defines an incompatible symbol defined by an earlier shared
library.

bfd/
* elflink.c (_bfd_elf_merge_symbol): Undo dynamic linking
state when a regular object file defines a symbol with
incompatible type to that defined by an earlier shared lib.
ld/
* testsuite/ld-elf/indirect5a.c,
* testsuite/ld-elf/indirect5b.c,
* testsuite/ld-elf/indirect5.map,
* testsuite/ld-elf/indirect5.out: New test.
* testsuite/ld-elf/indirect6a.c: Likewise.
* testsuite/ld-elf/indirect.exp (check_dynamic_syms): New proc.
Run new tests and check dynsyms.

7 years agoAutomatic date update in version.in
GDB Administrator [Mon, 17 Apr 2017 00:00:33 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoAutomatic date update in version.in
GDB Administrator [Sun, 16 Apr 2017 00:00:35 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoAutomatic date update in version.in
GDB Administrator [Sat, 15 Apr 2017 00:00:36 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoFix build breakage on Cygwin (PR gdb/21385)
Sergio Durigan Junior [Fri, 14 Apr 2017 01:01:53 +0000 (21:01 -0400)]
Fix build breakage on Cygwin (PR gdb/21385)

On gdb/windows-nat.c:windows_create_inferior, ALLARGS needs to be
declared independently of the host that we're building for.  This
fixes a build breakage on Cygwin.

2017-04-13  Sergio Durigan Junior  <sergiodj@redhat.com>

PR gdb/21385
* windows-nat.c (windows_create_inferior): Declare 'allargs'
independently of the host, and fix build breakage on Cygwin.

7 years agoAutomatic date update in version.in
GDB Administrator [Fri, 14 Apr 2017 00:00:35 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoMake inferior a class with cdtors, and use new/delete
Pedro Alves [Thu, 13 Apr 2017 15:15:35 +0000 (16:15 +0100)]
Make inferior a class with cdtors, and use new/delete

struct inferior became a non-POD when enum_flags was made a non-POD,
so we should be allocating/destroying inferiors with new/delete, etc.
That's what this commit does.

Note: this commit makes all boolean fields of inferior be "bool",
except the "detaching" field.  That'll require more work, so I split
it to a separate patch.

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

* inferior.c (free_inferior): Convert to ...
(inferior::~inferior): ... this dtor.
(inferior::inferior): New ctor, factored out from ...
(add_inferior_silent): ... here.  Allocate the inferior with a new
expression.
(delete_inferior): Call delete instead of free_inferior.
* inferior.h (gdb_environ, continuation): Forward declare.
(inferior): Now a class.  Add in-class initialization to all
members.  Make boolean fields bool, except 'detaching'.
(inferior::inferior): New explicit ctor.
(inferior::~inferior): New.

7 years agoGC inferior.c:init_inferior_list
Pedro Alves [Thu, 13 Apr 2017 15:15:34 +0000 (16:15 +0100)]
GC inferior.c:init_inferior_list

Not used anywhere.  This was actually never used.  It came in because
I originally created inferior.c by copying thread.c, and doing
s/thread/inferior/g, and missed that nothing needs this.  :-)

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

* inferior.c (init_inferior_list): Delete.
* inferior.h (init_inferior_list): Delete.

7 years agoImprove coverage of the PR threads/13217 regression test
Pedro Alves [Thu, 13 Apr 2017 15:15:34 +0000 (16:15 +0100)]
Improve coverage of the PR threads/13217 regression test

- Make sure we end up with no thread selected after the detach.

- Test both "thread apply all" and "thread apply $some_threads", for
  completeness.

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

PR threads/13217
* gdb.threads/threadapply.exp (thr_apply_detach): New procedure.
(top level): Call it twice, with different thread sets.

7 years agoC++fy thread_apply_all_command
Pedro Alves [Thu, 13 Apr 2017 15:15:34 +0000 (16:15 +0100)]
C++fy thread_apply_all_command

This eliminates a couple cleanups.

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

* thread.c: Include <algorithm>.
(thread_array_cleanup): Delete.
(scoped_inc_dec_ref): New class.
(live_threads_count): New function.
(set_thread_refcount): Delete.
(tp_array_compar_ascending): Now a bool.
(tp_array_compar): Convert to a std::sort comparison function.
(thread_apply_all_command): Use std::vector and scoped_inc_dec_ref
and live_threads_count.

7 years agoFix follow-fork latent bug
Pedro Alves [Thu, 13 Apr 2017 15:15:34 +0000 (16:15 +0100)]
Fix follow-fork latent bug

A later patch in the series adds an assertion to switch_to_thread that
the resulting inferior_ptid always matches the "current_inferior()"
inferior.  This exposed a latent bug in the follow-fork code, where
we're building the fork child inferior.  We're switching
inferior_ptid, but not the current inferior object...

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

* infrun.c (follow_fork_inferior): Also switch the current
inferior.

7 years agowatch_command_1: Fix dangling frame access
Pedro Alves [Thu, 13 Apr 2017 15:15:34 +0000 (16:15 +0100)]
watch_command_1: Fix dangling frame access

While working on some changes to switch_to_thread, I inadvertently
make switch_to_thread call reinit_frame_cache more frequently, even
when the thread didn't change.  This exposed a latent bug in
watch_command_1, where we're referencing a frame after
creating/inserting breakpoints, which potentially calls
reinit_frame_cache if it needs to install breakpoints with a different
thread selected.

Handle this similarly to how it's already handled in other similar
cases.  I.e., save any frame-related information we might need before
creating a breakpoint.

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

* breakpoint.c (watch_command_1): Save watchpoint-frame info
before calling create_internal_breakpoint.

7 years agoreadelf: fix out of range subtraction, seg fault from a NULL pointer and memory exhau...
Nick Clifton [Thu, 13 Apr 2017 15:06:30 +0000 (16:06 +0100)]
readelf: fix out of range subtraction, seg fault from a NULL pointer and memory exhaustion, all from parsing corrupt binaries.

PR binutils/21379
* readelf.c (process_dynamic_section): Detect over large section
offsets in the DT_SYMTAB entry.

PR binutils/21345
* readelf.c (process_mips_specific): Catch an unfeasible memory
allocation before it happens and print a suitable error message.

7 years agoAdd note merging to strip and add code to merge stack size notes.
Nick Clifton [Thu, 13 Apr 2017 13:50:56 +0000 (14:50 +0100)]
Add note merging to strip and add code to merge stack size notes.

* objcopy.c: Add --no-merge-notes option to disable note merging.
Add --[no-]merge-notes option to strip, and enable it by default.
(num_bytes): New function.
(merge_gnu_build_notes): Add code to merge stack size notes.
* binutils.texi: Update strip and objcopy documentation.
* readelf.c (print_gnu_build_attribute_name): Use defined
constants for note types.

7 years agoRegen cgen files
Alan Modra [Thu, 13 Apr 2017 01:03:53 +0000 (10:33 +0930)]
Regen cgen files

* epiphany-desc.c: Regenerate.
* fr30-desc.c: Regenerate.
* frv-desc.c: Regenerate.
* ip2k-desc.c: Regenerate.
* iq2000-desc.c: Regenerate.
* lm32-desc.c: Regenerate.
* m32c-desc.c: Regenerate.
* m32r-desc.c: Regenerate.
* mep-desc.c: Regenerate.
* mt-desc.c: Regenerate.
* or1k-desc.c: Regenerate.
* xc16x-desc.c: Regenerate.
* xstormy16-desc.c: Regenerate.

7 years agofork-child.c: Avoid unnecessary heap-allocation / string copying
Pedro Alves [Thu, 13 Apr 2017 10:46:07 +0000 (11:46 +0100)]
fork-child.c: Avoid unnecessary heap-allocation / string copying

The previous change to fork-child.c converted the argv building from
an alloca-allocated array of non-owning arg pointers, to a std::vector
of owning pointers, which results in N string dups, with N being the
number of arguments in the vector, and then requires manually
releasing the pointers owned by the vector.

This patch makes the vector hold non-owning pointers, and avoids the
string dups, by doing one single string copy of the arguments upfront,
and replacing separators with NULL terminators in place, like we used
to.  All the logic to do that is encapsulated in a new class.

With this, there's no need to remember to manually release the argv
elements with free_vector_argv either.

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

* fork-child.c (execv_argv): New class.
(breakup_args): Refactored as ...
(execv_argv::init_for_no_shell): .. this method of execv_argv.
Copy arguments to storage and replace separators with NULL
terminators in place.
(escape_bang_in_quoted_argument): Adjust to return bool.
(execv_argv::execv_argv): New ctor.
(execv_argv::init_for_shell): New method, factored out from
fork_inferior.  Don't strdup strings into the vector.
(fork_inferior): Eliminate "shell" local and use execv_argv.  Use
Remove free_vector_argv call.

7 years ago * config.sub: Sync with master version in config project.
Andrew Jenner [Thu, 13 Apr 2017 09:34:19 +0000 (02:34 -0700)]
* config.sub: Sync with master version in config project.

7 years agoAdd ChangeLog entries
Iain Buclaw [Thu, 13 Apr 2017 08:39:13 +0000 (10:39 +0200)]
Add ChangeLog entries

ChangeLog entries were left unstaged in my previous commit on March 30th.

7 years agoWrap long lines
Alan Modra [Thu, 13 Apr 2017 03:54:18 +0000 (13:24 +0930)]
Wrap long lines

Not a comprehensive change, just some split out from fixes made for
the %A and %B changes.

* coffcode.h: Wrap some overly long _bfd_error_handler args.
* elf.c: Likewise.
* elf32-arm.c: Likewise.
* elf32-i386.c: Likewise.
* elf32-mep.c: Likewise.
* elf64-ia64-vms.c: Likewise.
* elf64-x86-64.c: Likewise.
* elflink.c: Likewise.
* elfnn-ia64.c: Likewise.
* elfxx-mips.c: Likewise.

7 years agoUse %A and %B in more error messages
Alan Modra [Thu, 13 Apr 2017 03:50:15 +0000 (13:20 +0930)]
Use %A and %B in more error messages

* aoutx.h: Use %B and %A in error messages throughout file.
* aout-cris.c: Likewise.
* archive.c: Likewise.
* binary.c: Likewise.
* coff-rs6000.c: Likewise.
* coff-tic4x.c: Likewise.
* coffcode.h: Likewise.
* coffgen.c: Likewise.
* cofflink.c: Likewise.
* coffswap.h: Likewise.
* cpu-arm.c: Likewise.
* elf-eh-frame.c: Likewise.
* elf-m10300.c: Likewise.
* elf.c: Likewise.
* elf32-arc.c: Likewise.
* elf32-arm.c: Likewise.
* elf32-bfin.c: Likewise.
* elf32-frv.c: Likewise.
* elf32-iq2000.c: Likewise.
* elf32-m32c.c: Likewise.
* elf32-microblaze.c: Likewise.
* elf32-nds32.c: Likewise.
* elf32-rl78.c: Likewise.
* elf32-rx.c: Likewise.
* elf32-score.c: Likewise.
* elf32-score7.c: Likewise.
* elf32-sh64.c: Likewise.
* elf32-v850.c: Likewise.
* elf32-vax.c: Likewise.
* elf32-visium.c: Likewise.
* elf64-ia64-vms.c: Likewise.
* elf64-mmix.c: Likewise.
* elf64-sh64.c: Likewise.
* elfcode.h: Likewise.
* elfnn-aarch64.c: Likewise.
* elfnn-ia64.c: Likewise.
* elfxx-mips.c: Likewise.
* hpux-core.c: Likewise.
* ieee.c: Likewise.
* ihex.c: Likewise.
* linker.c: Likewise.
* merge.c: Likewise.
* mmo.c: Likewise.
* oasys.c: Likewise.
* pdp11.c: Likewise.
* peXXigen.c: Likewise.
* rs6000-core.c: Likewise.
* vms-alpha.c: Likewise.
* xcofflink.c: Likewise.

7 years agoRewrite bfd error handler
Alan Modra [Thu, 13 Apr 2017 01:28:40 +0000 (10:58 +0930)]
Rewrite bfd error handler

This steals _doprnt from libiberty, extended to handle %A and %B.
Which lets us do away with the current horrible %A and %B handling
that requires all %A and %B arguments to be passed first, rather than
in the natural order.

* bfd.c (PRINT_TYPE): Define.
(_doprnt): New function.
(error_handler_internal): Use _doprnt.
* coff-arm.c: Put %A and %B arguments to _bfd_error_handler
calls in their natural order, throughout file.
* coff-mcore.c: Likewise.
* coff-ppc.c: Likewise.
* coff-tic80.c: Likewise.
* cofflink.c: Likewise.
* elf-s390-common.c: Likewise.
* elf.c: Likewise.
* elf32-arm.c: Likewise.
* elf32-i386.c: Likewise.
* elf32-m32r.c: Likewise.
* elf32-msp430.c: Likewise.
* elf32-spu.c: Likewise.
* elf64-ia64-vms.c: Likewise.
* elf64-sparc.c: Likewise.
* elf64-x86-64.c: Likewise.
* elflink.c: Likewise.
* elfnn-aarch64.c: Likewise.
* elfnn-ia64.c: Likewise.
* elfxx-mips.c: Likewise.

7 years agoMissing _bfd_error_handler args
Alan Modra [Thu, 13 Apr 2017 01:14:20 +0000 (10:44 +0930)]
Missing _bfd_error_handler args

* elf32-arm.c (arm_type_of_stub): Supply missing args to "long
branch veneers" error.  Fix double space and format message.
* elf32-avr.c (avr_add_stub): Do not pass NULL as %B arg.
* elf64-ppc.c (tocsave_find): Supply missing %B arg.

7 years agoRegen bfd-in2.h
Alan Modra [Thu, 13 Apr 2017 01:03:11 +0000 (10:33 +0930)]
Regen bfd-in2.h

* bfd-in2.h: Regenerate.

7 years agoFix a typo in rx_fpsw_type
Yao Qi [Thu, 13 Apr 2017 07:29:44 +0000 (08:29 +0100)]
Fix a typo in rx_fpsw_type

gdb:

2017-04-13  Yao Qi  <yao.qi@linaro.org>

* rx-tdep.c (rx_fpsw_type): Check tdep->rx_fpsw_type instead of
tdep->rx_psw_type.

7 years agoXCNEW gdbarch_tdep in rl78 and rx
Yao Qi [Thu, 13 Apr 2017 07:29:43 +0000 (08:29 +0100)]
XCNEW gdbarch_tdep in rl78 and rx

"struct gdbarch_tdep" is XNEW'ed in rl78 and rx, so the memory is not
cleared.  As the result, tdep->rl78_psw_type is never initialized
properly.

  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  if (tdep->rl78_psw_type == NULL)
    {
      tdep->rl78_psw_type = arch_flags_type (gdbarch,
     "builtin_type_rl78_psw", 1);

The bug is found by my unit test in the following patch.

gdb:

2017-04-13  Yao Qi  <yao.qi@linaro.org>

* rl78-tdep.c (rl78_gdbarch_init): Use XCNEW instead of XNEW.
* rx-tdep.c (rx_gdbarch_init): Likewise.

7 years agostruct breakpoint: Fix indentation
Pedro Alves [Thu, 13 Apr 2017 02:07:21 +0000 (03:07 +0100)]
struct breakpoint: Fix indentation

I'm going to need to touch all these fields to add in-class
initialization anyway, might as well take the opportunity to finally
fix this...

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

* breakpoint.h (struct breakpoint): Reindent.

7 years agobreakpoint.c: bp_location (the array) shadows bp_location (the type)
Pedro Alves [Thu, 13 Apr 2017 01:12:53 +0000 (02:12 +0100)]
breakpoint.c: bp_location (the array) shadows bp_location (the type)

The bp_location array has the same name as the "struct bp_location",
type preventing refering to the structure without the "struct" inside
breakpoint.c.  I.e., we must write:

 "new struct bp_location;"

instead of:

 "new bp_location"

Rename the array and the associated variables/functions to avoid the
shadowing.

gdb/ChangeLog:
2017-04-13  Pedro Alves  <palves@redhat.com>

* breakpoint.c (bp_location): Rename to ...
(bp_locations): ... this.  All references updated.
(bp_location_count): Rename to ...
(bp_locations_count): ... this.  All references updated.
(bp_location_placed_address_before_address_max): Rename to ...
(bp_locations_placed_address_before_address_max): ... this.  All
references updated.
(bp_location_shadow_len_after_address_max): Rename to ...
(bp_locations_shadow_len_after_address_max): ... this.  All
references updated.
(bp_location_compare_addrs): Rename to ...
(bp_locations_compare_addrs): ... this.  All references updated.
(bp_location_compare):Rename to ...
(bp_locations_compare): ... this.  All references updated.
(bp_location_target_extensions_update): Rename to ...
(bp_locations_target_extensions_update): ... this.  All references
updated.

7 years agoAutomatic date update in version.in
GDB Administrator [Thu, 13 Apr 2017 00:00:38 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoCreate gdb_termios.h (and cleanup gdb/{,gdbserver/}terminal.h)
Sergio Durigan Junior [Wed, 12 Apr 2017 01:05:12 +0000 (21:05 -0400)]
Create gdb_termios.h (and cleanup gdb/{,gdbserver/}terminal.h)

As requested, I'm sending this as a separate patch because it is ready
to be included as-is.

The idea here is that both gdb/terminal.h and gdb/gdbserver/terminal.h
share the same code, which is responsible for setting a bunch of
defines on based on the presence of termios.h and a few other headers.
This simple patch just moves this common code to common/gdb_termios.h
and makes the necessary adjustments on both GDB and gdbserver so that
they can use this new header.  It also implements the some header
checks on common/common.m4.

As a bonus, gdb/gdbserver/terminal.h can be removed because it's now
empty.

Built on x86_64, no regressions found.

gdb/ChangeLog:
2017-04-12  Sergio Durigan Junior  <sergiodj@redhat.com>

* Makefile.in (HFILES_NO_SRCDIR): Add "common/gdb_termios.h".
* common/common.m4: Check headers 'termios.h', 'termio.h' and
'sgtty.h'.
* common/gdb_termios.h: New file, with parts of "terminal.h".
* inflow.c: Include "gdb_termios.h".
* ser-unix.c: Include "gdb_termios.h".
* terminal.h: Move terminal-related defines to
"common/gdb_termios.h".

gdb/gdbserver/ChangeLog:
2017-04-12  Sergio Durigan Junior  <sergiodj@redhat.com>

* remote-utils.c: Include "gdb_termios.h" instead of
"terminal.h".
* terminal.h: Delete file.

7 years agoChange linespec_result::location to be an event_location_up
Tom Tromey [Mon, 10 Apr 2017 21:53:22 +0000 (15:53 -0600)]
Change linespec_result::location to be an event_location_up

This is a follow-up to another patch.  It changes
linespec_result::location to be an event_location_up.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* probe.c (parse_probes): Update.
* location.h (delete_event_location): Don't declare.
(event_location_deleter::operator()): Update.
* location.c (event_location_deleter::operator()): Rename from
delete_event_location.
* linespec.h (linespec_result) <location>: Change type to
event_location_up.
* linespec.c (canonicalize_linespec, event_location_to_sals)
(decode_objc): Update.
(linespec_result): Don't call delete_event_location.
* breakpoint.c (create_breakpoints_sal)
(bkpt_probe_create_sals_from_location)
(strace_marker_create_sals_from_location): Update.

7 years agoAdd a constructor and destructor to linespec_result
Tom Tromey [Mon, 10 Apr 2017 21:47:21 +0000 (15:47 -0600)]
Add a constructor and destructor to linespec_result

linespec_result is only ever allocated on the stack, so it's
relatively easy to convert to having a constructor and a destructor.
This patch makes this change.  This removes some cleanups.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* linespec.h (struct linespec_result): Add constructor and
destructor.
(init_linespec_result, destroy_linespec_result)
(make_cleanup_destroy_linespec_result): Don't declare.
* linespec.c (init_linespec_result): Remove.
(linespec_result::~linespec_result): Rename from
destroy_linespec_result.  Update.
(cleanup_linespec_result, make_cleanup_destroy_linespec_result):
Remove.
* breakpoint.c (create_breakpoint, break_range_command)
(decode_location_default): Update.
* ax-gdb.c (agent_command_1): Update.

7 years agoChange breakpoint event locations to event_location_up
Tom Tromey [Mon, 10 Apr 2017 21:41:25 +0000 (15:41 -0600)]
Change breakpoint event locations to event_location_up

This is a follow-up to an earlier patch.  It changes breakpoint's
location and location_range_end members to be of type
event_location_up, then fixes up the users.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* remote.c (remote_download_tracepoint): Update.
* python/py-breakpoint.c (bppy_get_location): Update.
* guile/scm-breakpoint.c (bpscm_print_breakpoint_smob)
(gdbscm_breakpoint_location): Update.
* elfread.c (elf_gnu_ifunc_resolver_return_stop): Update.
* breakpoint.h (struct breakpoint) <location, location_range_end>:
Change type to event_location_up.
* breakpoint.c (create_overlay_event_breakpoint)
(create_longjmp_master_breakpoint)
(create_std_terminate_master_breakpoint)
(create_exception_master_breakpoint)
(breakpoint_event_location_empty_p, print_breakpoint_location)
(print_one_breakpoint_location, create_thread_event_breakpoint)
(init_breakpoint_sal, create_breakpoint)
(print_recreate_ranged_breakpoint, break_range_command)
(init_ada_exception_breakpoint, say_where): Update.
(base_breakpoint_dtor): Don't call delete_event_location.
(bkpt_print_recreate, tracepoint_print_recreate)
(dprintf_print_recreate, update_static_tracepoint)
(breakpoint_re_set_default): Update.

7 years agoUse std::vector in compile-loc2c.c
Tom Tromey [Fri, 7 Apr 2017 22:05:42 +0000 (16:05 -0600)]
Use std::vector in compile-loc2c.c

This changes compile-loc2c.c to use std::vector in place of a VEC,
allowing the removal of a cleanup.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* compile/compile-loc2c.c (compute_stack_depth_worker): Change
type of "to_do".  Update.
(compute_stack_depth): Use std::vector.

7 years agoUse std::vector in find_instruction_backward
Tom Tromey [Fri, 7 Apr 2017 22:01:10 +0000 (16:01 -0600)]
Use std::vector in find_instruction_backward

This changes find_instruction_backward to use std::vector, removing a
cleanup.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* printcmd.c (find_instruction_backward): Use std::vector.

7 years agoUse std::vector in reread_symbols
Tom Tromey [Fri, 7 Apr 2017 21:58:25 +0000 (15:58 -0600)]
Use std::vector in reread_symbols

This changes reread_symbols to use std::vector, removing a cleanup.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* symfile.c (objfilep): Remove typedef.
(reread_symbols): Use a std::vector.

7 years agoUse scoped_restore in more places
Tom Tromey [Fri, 7 Apr 2017 21:37:25 +0000 (15:37 -0600)]
Use scoped_restore in more places

This changes a few more places to use scoped_restore, allowing some
cleanup removals.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* mi/mi-main.c (exec_direction_forward): Remove.
(exec_reverse_continue, mi_execute_command): Use scoped_restore.
* guile/scm-ports.c (ioscm_with_output_to_port_worker): Use
scoped_restore.
* guile/guile.c (guile_repl_command, guile_command)
(gdbscm_execute_gdb_command): Use scoped_restore.
* go-exp.y (go_parse): Use scoped_restore.
* d-exp.y (d_parse): Use scoped_restore.
* cli/cli-decode.c (cmd_func): Use scoped_restore.
* c-exp.y (c_parse): Use scoped_restore.

7 years agoC++ify mi_parse
Tom Tromey [Fri, 7 Apr 2017 21:34:02 +0000 (15:34 -0600)]
C++ify mi_parse

This changes mi_parse to return a unique_ptr, and to use "new"; then
fixes up the users.  This allows removing one cleanup.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* mi/mi-parse.h (struct mi_parse): Add constructor, destructor.
(mi_parse): Update return type.
(mi_parse_free): Remove.
* mi/mi-parse.c (mi_parse::mi_parse): New constructor.
(mi_parse::~mi_parse): Rename from mi_parse_free.
(mi_parse_cleanup): Remove.
(mi_parse): Return a unique_ptr.  Use new.
* mi/mi-main.c (mi_execute_command): Update.

7 years agoRemove some cleanups from location.c
Tom Tromey [Fri, 7 Apr 2017 05:26:14 +0000 (23:26 -0600)]
Remove some cleanups from location.c

This removes some more cleanups from location.c by using
unique_xmalloc_ptr.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* location.c (explicit_location_lex_one): Return a
unique_xmalloc_ptr.
(string_to_explicit_location): Update.  Remove cleanups.

7 years agoRemove some cleanups from gnu-v3-abi.c
Tom Tromey [Fri, 7 Apr 2017 04:48:27 +0000 (22:48 -0600)]
Remove some cleanups from gnu-v3-abi.c

This removes some cleanups from gnu-v3-abi.c, by using std::vector
rather than VEC.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* gnu-v3-abi.c (value_and_voffset_p): Remove typedef.
(compare_value_and_voffset): Change type.  Update.
(compute_vtable_size): Change type of "offset_vec".
(gnuv3_print_vtable): Use std::vector.  Remove cleanups.
(gnuv3_get_typeid): Remove extraneous declaration.

7 years agoFix up wchar_iterator comment
Tom Tromey [Fri, 7 Apr 2017 03:07:00 +0000 (21:07 -0600)]
Fix up wchar_iterator comment

This fixes up a comment in charset.h that has been obsolete for a
while.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* charset.h (wchar_iterator): Fix comment.

7 years agoRemove cleanup_iconv
Tom Tromey [Fri, 7 Apr 2017 02:50:38 +0000 (20:50 -0600)]
Remove cleanup_iconv

This introduces a new "iconv_wrapper" class, to be used in
convert_between_encodings.  This allows the removal of cleanup_iconv.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* charset.c (iconv_wrapper): New class.
(cleanup_iconv): Remove.
(convert_between_encodings): Use it.

7 years agoChange increment_reading_symtab to return a scoped_restore
Tom Tromey [Fri, 7 Apr 2017 02:27:10 +0000 (20:27 -0600)]
Change increment_reading_symtab to return a scoped_restore

This changes increment_reading_symtab to return a scoped_restore, then
fixes up the users.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* symfile.h (increment_reading_symtab): Update type.
* symfile.c (decrement_reading_symtab): Remove.
(increment_reading_symtab): Return a scoped_restore_tmpl<int>.
* psymtab.c (psymtab_to_symtab): Update.
* dwarf2read.c (dw2_instantiate_symtab): Update.

7 years agoIntroduce gdb_dlhandle_up
Tom Tromey [Fri, 7 Apr 2017 02:09:42 +0000 (20:09 -0600)]
Introduce gdb_dlhandle_up

This introduces gdb_dlhandle_up, a unique_ptr that can close a
dlopen'd library.  All the functions working with dlopen handles are
updated to use this new type.

I did not try to build this on Windows.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* jit.c (struct jit_reader): Declare separately.  Add constructor
and destructor.  Change type of "handle".
(loaded_jit_reader): Define separately.
(jit_reader_load): Update.  New "new".
(jit_reader_unload_command): Use "delete".
* gdb-dlfcn.h (struct dlclose_deleter): New.
(gdb_dlhandle_up): New typedef.
(gdb_dlopen, gdb_dlsym): Update types.
(gdb_dlclose): Remove.
* gdb-dlfcn.c (gdb_dlopen): Return a gdb_dlhandle_up.
(gdb_dlsym): Change type of "handle".
(make_cleanup_dlclose): Remove.
(dlclose_deleter::operator()): Rename from gdb_dlclose.
* compile/compile-c-support.c (load_libcc): Update.

7 years agoChange find_pcs_for_symtab_line to return a std::vector
Tom Tromey [Thu, 6 Apr 2017 22:38:56 +0000 (16:38 -0600)]
Change find_pcs_for_symtab_line to return a std::vector

This changes find_pcs_for_symtab_line to return a std::vector.  This
allows the removal of some cleanups.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* symtab.h (find_pcs_for_symtab_line): Change return type.
* symtab.c (find_pcs_for_symtab_line): Change return type.
* python/py-linetable.c (build_line_table_tuple_from_pcs): Change
type of "vec".  Update.
(ltpy_get_pcs_for_line): Update.
* linespec.c (decode_digits_ordinary): Update.

7 years agoIntroduce command_line_up
Tom Tromey [Thu, 6 Apr 2017 03:14:09 +0000 (21:14 -0600)]
Introduce command_line_up

This introduces command_line_up, a unique_ptr for command_line
objects, and changes many places to use it.  This removes a number of
cleanups.

Command lines are funny in that sometimes they are reference counted.
Once there is more C++-ification of some of the users, perhaps all of
these can be changed to use shared_ptr instead.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* tracepoint.c (actions_command): Update.
* python/python.c (python_command, python_interactive_command):
Update.
* mi/mi-cmd-break.c (mi_cmd_break_commands): Update.
* guile/guile.c (guile_command): Update.
* defs.h (read_command_lines, read_command_lines_1): Return
command_line_up.
(command_lines_deleter): New struct.
(command_line_up): New typedef.
* compile/compile.c (compile_code_command)
(compile_print_command): Update.
* cli/cli-script.h (get_command_line, copy_command_lines): Return
command_line_up.
(make_cleanup_free_command_lines): Remove.
* cli/cli-script.c (get_command_line, read_command_lines_1)
(copy_command_lines): Return command_line_up.
(while_command, if_command, read_command_lines, define_command)
(document_command): Update.
(do_free_command_lines_cleanup, make_cleanup_free_command_lines):
Remove.
* breakpoint.h (breakpoint_set_commands): Change type of
"commands".
* breakpoint.c (breakpoint_set_commands): Change type of
"commands".  Update.
(do_map_commands_command, update_dprintf_command_list)
(create_tracepoint_from_upload): Update.

7 years agoIntroduce event_location_up
Tom Tromey [Thu, 6 Apr 2017 02:44:01 +0000 (20:44 -0600)]
Introduce event_location_up

This removes make_cleanup_delete_event_location and instead changes
the various location functions to return an event_location_up, a new
unique_ptr typedef.

This is largely straightforward, but be sure to examine the
init_breakpoint_sal change.  I believe the code I deleted there is
dead, because "location != NULL" can never be true in that branch; but
you should double-check.

gdb/ChangeLog
2017-04-12  Tom Tromey  <tom@tromey.com>

* tracepoint.c (scope_info): Update.
* spu-tdep.c (spu_catch_start): Update.
* python/python.c (gdbpy_decode_line): Update.
* python/py-finishbreakpoint.c (bpfinishpy_init): Update.
* python/py-breakpoint.c (bppy_init): Update.
* probe.c (parse_probes): Update.
* mi/mi-cmd-break.c (mi_cmd_break_insert_1): Update.
* location.h (event_location_deleter): New struct.
(event_location_up): New typedef.
(new_linespec_location, new_address_location, new_probe_location)
(new_explicit_location, copy_event_location)
(string_to_event_location, string_to_event_location_basic)
(string_to_explicit_location): Update return type.
(make_cleanup_delete_event_location): Remove.
* location.c (new_linespec_location, new_address_location)
(new_probe_location, new_explicit_location, copy_event_location):
Return event_location_up.
(delete_event_location_cleanup)
(make_cleanup_delete_event_location): Remove.
(string_to_explicit_location, string_to_event_location_basic)
(string_to_event_location): Return event_location_up.
* linespec.c (canonicalize_linespec, event_location_to_sals)
(decode_line_with_current_source)
(decode_line_with_last_displayed, decode_objc): Update.
* guile/scm-breakpoint.c (gdbscm_register_breakpoint_x): Update.
* completer.c (location_completer): Update.
* cli/cli-cmds.c (edit_command, list_command): Update.
* breakpoint.c (create_overlay_event_breakpoint)
(create_longjmp_master_breakpoint)
(create_std_terminate_master_breakpoint)
(create_exception_master_breakpoint)
(create_thread_event_breakpoint): Update.
(init_breakpoint_sal): Update.  Remove some dead code.
(create_breakpoint_sal): Change type of "location".  Update.
(create_breakpoints_sal, create_breakpoint, break_command_1)
(dprintf_command, break_range_command, until_break_command)
(init_ada_exception_breakpoint)
(strace_marker_create_sals_from_location)
(update_static_tracepoint, trace_command, ftrace_command)
(strace_command, create_tracepoint_from_upload): Update.
* break-catch-throw.c (re_set_exception_catchpoint): Update.
* ax-gdb.c (agent_command_1): Update.

7 years agogdb: Move DJGPP/go32 bits to their own tdep file
Pedro Alves [Wed, 12 Apr 2017 15:00:04 +0000 (16:00 +0100)]
gdb: Move DJGPP/go32 bits to their own tdep file

I posit that this makes them easier to find.

The other day while working on the wchar_t patch, I had a bit of
trouble finding the DJGPP/go32 tdep bits.  My initial reaction was
looking for a go32-specific tdep file, but there's none.

Confirmed that a --host=i586-pc-msdosdjgpp GDB still builds
successfully and includes the  i386-go32-tdep.o object.

Confirmed that an --enable-targets=all build of GDB on x86-64
GNU/Linux includes the DJGPP/go32 bits too.

gdb/ChangeLog:
2017-04-12  Pedro Alves  <palves@redhat.com>

* Makefile.in (ALL_TARGET_OBS): Add i386-go32-tdep.o.
* configure.tgt: Handle i[34567]86-*-go32* and
i[34567]86-*-msdosdjgpp*.
* i386-tdep.c (i386_svr4_reg_to_regnum):
Make extern.
(i386_go32_init_abi, i386_coff_osabi_sniffer): Moved to
i386-go32-tdep.c.
(_initialize_i386_tdep): DJGPP bits moved to i386-go32-tdep.c.
* i386-go32-tdep.c: New file.
* i386-tdep.h (tdesc_i386_mmx, i386_svr4_reg_to_regnum): New
declarations.

7 years agoFix build error in aix-thread.c
Simon Marchi [Wed, 12 Apr 2017 14:46:13 +0000 (10:46 -0400)]
Fix build error in aix-thread.c

Obvious fix for:

aix-thread.c: In function 'char* pd_status2str(int)':
aix-thread.c:163:33: error: deprecated conversion from string constant to 'char*' [-Werror=write-strings]
     case PTHDB_SUCCESS:  return "SUCCESS";
                                 ^
gdb/ChangeLog:

* aix-thread.c (pd_status2str): Change return type to const char *.

7 years agoRemove unnecessary set_gdbarch_gnu_triplet_regexp calls
Pedro Alves [Fri, 31 Mar 2017 12:41:45 +0000 (13:41 +0100)]
Remove unnecessary set_gdbarch_gnu_triplet_regexp calls

i386_gdbarch_init already does this unconditionally for all x86 ports.

Tested on x86-64 Fedora 23.

gdb/ChangeLog:
2017-04-12  Pedro Alves  <palves@redhat.com>

* i386-tdep.c (i386_elf_init_abi, i386_go32_init_abi): Remove
calls to set_gdbarch_gnu_triplet_regexp.

7 years agoTeach GDB that wchar_t is a built-in type in C++ mode
Pedro Alves [Wed, 12 Apr 2017 13:00:49 +0000 (14:00 +0100)]
Teach GDB that wchar_t is a built-in type in C++ mode

GDB is currently not aware that wchar_t is a built-in type in C++
mode.  This is usually not a problem because the debug info describes
the type, so when you have a program loaded, you don't notice this.
However, if you try expressions involving wchar_t before a program is
loaded, gdb errors out:

 (gdb) p (wchar_t)-1
 No symbol table is loaded.  Use the "file" command.
 (gdb) p L"hello"
 No type named wchar_t.
 (gdb) ptype L"hello"
 No type named wchar_t.

This commit teaches gdb about the type.  After:

 (gdb) p (wchar_t)-1
 $1 = -1 L'\xffffffff'
 (gdb) p L"hello"
 $2 = L"hello"
 (gdb) ptype L"hello"
 type = wchar_t [6]

Unlike char16_t/char32_t, unfortunately, the underlying type of
wchar_t is implementation dependent, both size and signness.  So this
requires adding a couple new gdbarch hooks.

I grepped the GCC code base for WCHAR_TYPE and WCHAR_TYPE_SIZE, and it
seems to me that the majority of the ABIs have a 4-byte signed
wchar_t, so that's what I made the default for GDB too.  And then I
looked for which ports have a 16-bit and/or unsigned wchar_t, and made
GDB follow suit.

gdb/ChangeLog:
2017-04-12  Pedro Alves  <palves@redhat.com>

PR gdb/21323
* c-lang.c (cplus_primitive_types) <cplus_primitive_type_wchar_t>:
New enum value.
(cplus_language_arch_info): Register cplus_primitive_type_wchar_t.
* gdbtypes.h (struct builtin_type) <builtin_wchar>: New field.
* gdbtypes.c (gdbtypes_post_init): Create the "wchar_t" type.
* gdbarch.sh (wchar_bit, wchar_signed): New per-arch values.
* gdbarch.h, gdbarch.c: Regenerate.
* aarch64-tdep.c (aarch64_gdbarch_init): Override
gdbarch_wchar_bit and gdbarch_wchar_signed.
* alpha-tdep.c (alpha_gdbarch_init): Likewise.
* arm-tdep.c (arm_gdbarch_init): Likewise.
* avr-tdep.c (avr_gdbarch_init): Likewise.
* h8300-tdep.c (h8300_gdbarch_init): Likewise.
* i386-nto-tdep.c (i386nto_init_abi): Likewise.
* i386-tdep.c (i386_go32_init_abi): Likewise.
* m32r-tdep.c (m32r_gdbarch_init): Likewise.
* moxie-tdep.c (moxie_gdbarch_init): Likewise.
* nds32-tdep.c (nds32_gdbarch_init): Likewise.
* rs6000-aix-tdep.c (rs6000_aix_init_osabi): Likewise.
* sh-tdep.c (sh_gdbarch_init): Likewise.
* sparc-tdep.c (sparc32_gdbarch_init): Likewise.
* sparc64-tdep.c (sparc64_init_abi): Likewise.
* windows-tdep.c (windows_init_abi): Likewise.
* xstormy16-tdep.c (xstormy16_gdbarch_init): Likewise.

gdb/testsuite/ChangeLog:
2017-04-12  Pedro Alves  <palves@redhat.com>

PR gdb/21323
* gdb.cp/wide_char_types.c: Include <wchar.h>.
(wchar): New global.
* gdb.cp/wide_char_types.exp (wide_char_types_program)
(do_test_wide_char, wide_char_types_no_program, top level): Add
wchar_t testing.

7 years agoFix PR c++/21323: GDB thinks char16_t and char32_t are signed in C++
Pedro Alves [Wed, 12 Apr 2017 13:00:49 +0000 (14:00 +0100)]
Fix PR c++/21323: GDB thinks char16_t and char32_t are signed in C++

While the C++ standard says that char16_t and char32_t are unsigned types:

 Types char16_t and char32_t denote distinct types with the same size,
 signedness, and alignment as uint_least16_t and uint_least32_t,
 respectively, in <cstdint>, called the underlying types.

... gdb treats them as signed currently:

 (gdb) p (char16_t)-1
 $1 = -1 u'\xffff'

There are actually two places in gdb that hardcode these types:

- gdbtypes.c:gdbtypes_post_init, when creating the built-in types,
  seemingly used by the "x /s" command (judging from commit 9a22f0d0).

- dwarf2read.c, when reading base types with DW_ATE_UTF encoding
  (which is what is used for these types, when compiling for C++11 and
  up).  Despite the comment, the type created does end up used.

Both places need fixing.  But since I couldn't tell why dwarf2read.c
needs to create a new type, I've made it use the per-arch built-in
types instead, so that the types are only created once per arch
instead of once per objfile.  That seems to work fine.

While writting the test, I noticed that the C++ language parser isn't
actually aware of these built-in types, so if you try to use them
without a program that uses them, you get:

 (gdb) set language c++
 (gdb) ptype char16_t
 No symbol table is loaded.  Use the "file" command.
 (gdb) ptype u"hello"
 No type named char16_t.
 (gdb) p u"hello"
 No type named char16_t.

That's fixed by simply adding a couple entries to C++'s built-in types
array in c-lang.c.  With that, we get the expected:

 (gdb) ptype char16_t
 type = char16_t
 (gdb) ptype u"hello"
 type = char16_t [6]
 (gdb) p u"hello"
 $1 = u"hello"

gdb/ChangeLog:
2017-04-12  Pedro Alves  <palves@redhat.com>

PR c++/21323
* c-lang.c (cplus_primitive_types) <cplus_primitive_type_char16_t,
cplus_primitive_type_char32_t>: New enum values.
(cplus_language_arch_info): Register cplus_primitive_type_char16_t
and cplus_primitive_type_char32_t.
* dwarf2read.c (read_base_type) <DW_ATE_UTF>: If bit size is 16 or
32, use the archtecture's built-in type for char16_t and char32_t,
respectively.  Otherwise, fallback to init_integer_type as before,
but make the type unsigned, and issue a complaint.
* gdbtypes.c (gdbtypes_post_init): Make char16_t and char32_t unsigned.

gdb/testsuite/ChangeLog:
2017-04-12  Pedro Alves  <palves@redhat.com>

PR c++/21323
* gdb.cp/wide_char_types.c: New file.
* gdb.cp/wide_char_types.exp: New file.

7 years agoFix Changelog for ab0538b875c054468aa205d8d36550d1223a3bfd
Alan Hayward [Wed, 12 Apr 2017 08:51:04 +0000 (09:51 +0100)]
Fix Changelog for ab0538b875c054468aa205d8d36550d1223a3bfd

7 years agoAdd M32R_ARG_REGISTER_SIZE
Alan Hayward [Wed, 12 Apr 2017 08:19:55 +0000 (09:19 +0100)]
Add M32R_ARG_REGISTER_SIZE

gdb/
* m32r-tdep.c M32R_ARG_REGISTER_SIZE: Added.
(m32r_push_dummy_call): Use M32R_ARG_REGISTER_SIZE.

7 years agoFix build breakage from last commit (window-nat.c:windows_create_inferior)
Sergio Durigan Junior [Wed, 12 Apr 2017 05:16:50 +0000 (01:16 -0400)]
Fix build breakage from last commit (window-nat.c:windows_create_inferior)

Forgot to declare the variable 'toexec' (from
window-nat.c:windows_create_inferior) as 'const char *', which caused
a build breakage.

gdb/ChangeLog:
2017-04-12  Sergio Durigan Junior  <sergiodj@redhat.com>

* windows-nat.c (windows_create_inferior): Declare 'toexec' as
'const char *'.

7 years agoC++-fy and prepare for sharing fork_inferior
Sergio Durigan Junior [Thu, 23 Mar 2017 01:54:49 +0000 (21:54 -0400)]
C++-fy and prepare for sharing fork_inferior

As a preparation for the next patch, which will move fork_inferior
from GDB to common/ (and therefore share it with gdbserver), it is
interesting to convert a few functions to C++.

This patch touches functions related to parsing command-line arguments
to the inferior (see gdb/fork-child.c:breakup_args), the way the
arguments are stored on fork_inferior (using std::vector instead of
char **), and the code responsible for dealing with argv also on
gdbserver.

I've taken this opportunity and decided to constify a few arguments to
fork_inferior/create_inferior as well, in order to make the code
cleaner.  And now, on gdbserver, we're using xstrdup everywhere and
aren't checking for memory allocation failures anymore, as requested
by Pedro:

  <https://sourceware.org/ml/gdb-patches/2017-03/msg00191.html>
  Message-Id: <025ebdb9-90d9-d54a-c055-57ed2406b812@redhat.com>

  Pedro Alves wrote:

  > On the "== NULL" check: IIUC, the old NULL check was there to
  > handle strdup returning NULL due to out-of-memory.
  > See NULL checks and comments further above in this function.
  > Now that you're using a std::vector, that doesn't work or make
  > sense any longer, since if push_back fails to allocate space for
  > its internal buffer (with operator new), our operator new replacement
  > (common/new-op.c) calls malloc_failure, which aborts gdbserver.
  >
  > Not sure it makes sense to handle out-of-memory specially in
  > the gdb/rsp-facing functions nowadays (maybe git blame/log/patch
  > submission for that code shows some guidelines).  Maybe (or, probably)
  > it's OK to stop caring about it, but then we should consistently remove
  > left over code, by using xstrdup instead and remove the NULL checks.

IMO this refactoring was very good to increase the readability of the
code as well, because some parts of the argument handling were
unnecessarily confusing before.

gdb/ChangeLog:
2017-04-12  Sergio Durigan Junior  <sergiodj@redhat.com>

* common/common-utils.c (free_vector_argv): New function.
* common/common-utils.h: Include <vector>.
(free_vector_argv): New prototype.
* darwin-nat.c (darwin_create_inferior): Rewrite function
prototype in order to constify "exec_file" and accept a
"std::string" for "allargs".
* fork-child.c: Include <vector>.
(breakup_args): Rewrite function, using C++.
(fork_inferior): Rewrite function header, constify "exec_file_arg"
and accept "std::string" for "allargs".  Update the code to
calculate "argv" based on "allargs".  Update calls to "exec_fun"
and "execvp".
* gnu-nat.c (gnu_create_inferior): Rewrite function prototype in
order to constify "exec_file" and accept a "std::string" for
"allargs".
* go32-nat.c (go32_create_inferior): Likewise.
* inf-ptrace.c (inf_ptrace_create_inferior): Likewise.
* infcmd.c (run_command_1): Constify "exec_file".  Use
"std::string" for inferior arguments.
* inferior.h (fork_inferior): Update prototype.
* linux-nat.c (linux_nat_create_inferior): Rewrite function
prototype in order to constify "exec_file" and accept a
"std::string" for "allargs".
* nto-procfs.c (procfs_create_inferior): Likewise.
* procfs.c (procfs_create_inferior): Likewise.
* remote-sim.c (gdbsim_create_inferior): Likewise.
* remote.c (extended_remote_run): Update code to accept
"std::string" as argument.
(extended_remote_create_inferior): Rewrite function prototype in
order to constify "exec_file" and accept a "std::string" for
"allargs".
* rs6000-nat.c (super_create_inferior): Likewise.
(rs6000_create_inferior): Likewise.
* target.h (struct target_ops) <to_create_inferior>: Likewise.
* windows-nat.c (windows_create_inferior): Likewise.

gdb/gdbserver/ChangeLog:
2017-04-12  Sergio Durigan Junior  <sergiodj@redhat.com>

* server.c: Include <vector>.
<program_argv, wrapper_argv>: Convert to std::vector.
(start_inferior): Rewrite function to use C++.
(handle_v_run): Likewise.  Update code that calculates the argv
based on the vRun packet; use C++.
(captured_main): Likewise.

7 years agoAutomatic date update in version.in
GDB Administrator [Wed, 12 Apr 2017 00:00:43 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoRemove the extra `\n' in warning/error messages
H.J. Lu [Tue, 11 Apr 2017 22:41:00 +0000 (15:41 -0700)]
Remove the extra `\n' in warning/error messages

* elf-properties.c (_bfd_elf_parse_gnu_properties): Remove the
extra `\n' in warning/error messages.
* elf32-i386.c (elf_i386_parse_gnu_properties): Likewise.
* elf64-x86-64.c (elf_x86_64_parse_gnu_properties): Likewise.

7 years agoIgnore processor-specific GNU program properties
H.J. Lu [Tue, 11 Apr 2017 22:07:41 +0000 (15:07 -0700)]
Ignore processor-specific GNU program properties

Skip processor-specific GNU program properties with generic ELF target
vector.  They should be handled by the matching ELF target vector.

* elf-properties.c (_bfd_elf_parse_gnu_properties): Ignore
processor-specific properties with generic ELF target vector.

7 years agogdb/thread.c: Fix whitespace throughout
Pedro Alves [Tue, 11 Apr 2017 12:57:52 +0000 (13:57 +0100)]
gdb/thread.c: Fix whitespace throughout

gdb/ChangeLog:
2017-04-11  Pedro Alves  <palves@redhat.com>

* thread.c: Fix whitespace throughout.

7 years agoFix read after xfree in linux-nat.c:linux_nat_detach
Philipp Rudo [Tue, 11 Apr 2017 13:28:51 +0000 (14:28 +0100)]
Fix read after xfree in linux-nat.c:linux_nat_detach

At the end of linux_nat_detach the main_lwp is deleted (delete_lwp).
This is problematic as during detach (detach_one_lwp and
linux_fork_detach) main_lwp already gets freed.  Thus calling
delete_lwp causes a read after free.  Fix it by removing the
unnecessary delete_lwp.

gdb/ChangeLog:
2017-04-11  Philipp Rudo  <prudo@linux.vnet.ibm.com>

* linux-nat.c (linux_nat_detach): Remove delete_lwp call.

7 years agoRemove MAX_REGISTER_SIZE from arm-tdep.c
Alan Hayward [Tue, 11 Apr 2017 12:51:58 +0000 (13:51 +0100)]
Remove MAX_REGISTER_SIZE from arm-tdep.c

gdb/
* arm-tdep.c (arm_store_return_value): Use FP_REGISTER_SIZE

7 years agoPR 21274, ld segfaults linking PE DLL
Alan Modra [Tue, 11 Apr 2017 10:07:51 +0000 (19:37 +0930)]
PR 21274, ld segfaults linking PE DLL

Don't use fixed size buffers for symbol names.

PR 21274
PR 18466
* emultempl/pe.em (pe_find_data_imports): Don't use fixed size
symbol buffer.  Instead, xmalloc max size needed with space for
prefix.  Wrap overlong lines.  Formatting.  Pass symbol buffer
copy of name to pe_walk_relocs_of_symbol.
(make_inport_fixup): Add "name" param, pass to pe_create_import_fixup.
* emultempl/pe.em (pep_find_data_imports): As for pe_find_data_imports.
(make_import_fixup): Add "name" param, pass to pep_create_import_fixup.
Use bfd_get_signed_* and remove unnecessary casts.  Formatting.
* pe-dll.c (pe_walk_relocs_of_symbol): Add "name" param.  Pass to
callback.
(make_import_fixup_mark): Add "name" param.  Make use of prefix
space rather than xmalloc here.
(pe_create_import_fixup): Likewise.
* pe-dll.h (pe_walk_relocs_of_symbol): Update prototype.
(pe_create_import_fixup): Likewise.
* pep-dll.h (pep_walk_relocs_of_symbol): Likewise.
(pep_create_import_fixup): Likewise.

7 years agoAutomatic date update in version.in
GDB Administrator [Tue, 11 Apr 2017 00:00:28 +0000 (00:00 +0000)]
Automatic date update in version.in

7 years agoReorder PPC_OPCODE_* and set PPC_OPCODE_TMR for e6500
Alan Modra [Mon, 10 Apr 2017 22:13:21 +0000 (07:43 +0930)]
Reorder PPC_OPCODE_* and set PPC_OPCODE_TMR for e6500

PPC_OPCODE_* renumbered to fill the gaps left by previous patches,
and reordered chronologically just because.  I kept PPC_OPCODE_TMR
because presumably it might be used in future APUinfo for e6500.

include/
* opcode/ppc.h (PPC_OPCODE_*): Renumber and order chronologically.
(PPC_OPCODE_SPE): Comment on this and other bits used for APUinfo.
opcodes/
* ppc-dis.c (ppc_opts): Formatting.  Set PPC_OPCODE_TMR for e6500.
* ppc-opc.c (powerpc_opcodes <mftmr, mttmr>): Remove now
unnecessary E6500.

7 years agoBye bye PPC_OPCODE_HTM and -mhtm
Alan Modra [Mon, 10 Apr 2017 22:10:24 +0000 (07:40 +0930)]
Bye bye PPC_OPCODE_HTM and -mhtm

The -mhtm option is fairly useless too.

include/
* opcode/ppc.h (PPC_OPCODE_HTM): Delete.
gas/
* config/tc-ppc.c (md_show_usage): Delete mention of -mhtm.
* testsuite/gas/ppc/htm.d: Pass -mpower8 and -Mpower8.
opcodes/
* ppc-dis.c (ppc_opts): Remove PPC_OPCODE_HTM and "htm".
* ppc-opc.c (PPCHTM): Define as PPC_OPCODE_POWER8.

7 years agoBye Bye PPC_OPCODE_VSX3
Alan Modra [Mon, 10 Apr 2017 22:06:43 +0000 (07:36 +0930)]
Bye Bye PPC_OPCODE_VSX3

This bit is also useless as it can be replaced with PPC_OPCODE_POWER9.
Defining the VSX2 and VSX3 selection based on cpu bits also lets the
assembler/disassembler distinguish between the power7 VSX opcodes and
the power8 ones.  Note that this change means -mvsx now reverts back
to just adding the power7 VSX insns.

include/
* opcode/ppc.h (PPC_OPCODE_VSX3): Delete.
opcodes/
* ppc-dis.c (ppc_opts): Remove PPC_OPCODE_VSX3.
* ppc-opc.c (PPCVSX2): Define as PPC_OPCODE_POWER8.
(PPCVSX3): Define as PPC_OPCODE_POWER9.

7 years agoBye bye PPC_OPCODE_ALTIVEC2
Alan Modra [Mon, 10 Apr 2017 22:03:50 +0000 (07:33 +0930)]
Bye bye PPC_OPCODE_ALTIVEC2

This bit is worse than useless.  Using it prevents the assembler and
disassembler distinguishing between opcodes added for power8 and those
added for power9.

include/
* opcode/ppc.h (PPC_OPCODE_ALTIVEC2): Delete.
opcodes/
* ppc-dis.c (ppc_opts): Remove PPC_OPCODE_ALTIVEC2.
* ppc-opc.c (PPCVEC2): Define as PPC_OPCODE_POWER8|PPC_OPCODE_E6500.
(PPCVEC3): Define as PPC_OPCODE_POWER9.

7 years agoFix PR gdb/21364: Dead code due to an unreachable condition in osdata.c
Sergio Durigan Junior [Mon, 10 Apr 2017 16:43:44 +0000 (12:43 -0400)]
Fix PR gdb/21364: Dead code due to an unreachable condition in osdata.c

Pedro's recent commits enabling -Wwrite-strings has changed a bit the
logic of info_osdata.  Now, 'type' is always non-NULL, so we have to
check if it's an empty string instead of NULL.  One of the checks was
fixed, but there is another that was left behind.  This commit fixes
it.

gdb/ChangeLog:
2017-04-10  Sergio Durigan Junior  <sergiodj@redhat.com>

PR gdb/21364
* osdata.c (info_osdata): Check if 'type' is an empty string
instead of NULL.

7 years agoPrevent a bigus warning from readelf about a gdb-index table being too big.
John Delsignor [Mon, 10 Apr 2017 15:27:05 +0000 (16:27 +0100)]
Prevent a bigus warning from readelf about a gdb-index table being too big.

PR binutils/21319
* dwarf.c (display_gdb_index): Correct test for a corrupt address
table size.

7 years agoDocument undocumented linker command line options.
Nick Clifton [Mon, 10 Apr 2017 14:56:55 +0000 (15:56 +0100)]
Document undocumented linker command line options.

ld  * ld.texinfo (--strip-discarded): Document.
(--embedded-relocs): Document.
(--spare-dynamic-tags): Document.
(--task-link): Document.

7 years agothread.c: ptid_equal -> operator==
Pedro Alves [Mon, 10 Apr 2017 14:54:57 +0000 (15:54 +0100)]
thread.c: ptid_equal -> operator==

gdb/ChangeLog:
2017-04-10  Pedro Alves  <palves@redhat.com>

* thread.c (add_thread_silent, delete_thread_1, find_thread_ptid)
(ptid_to_global_thread_id, in_thread_list)
(do_captured_list_thread_ids, set_resumed, set_running)
(set_executing, set_stop_requested, finish_thread_state)
(validate_registers_access, can_access_registers_ptid)
(print_thread_info_1, switch_to_thread)
(do_restore_current_thread_cleanup)
(make_cleanup_restore_current_thread, thread_command)
(thread_name_command): Use operator== instead of ptid_equal.

7 years agoGC gdb/thread.c:current_thread_cleanup_chain
Pedro Alves [Mon, 10 Apr 2017 14:18:49 +0000 (15:18 +0100)]
GC gdb/thread.c:current_thread_cleanup_chain

Commit 803bdfe43083475c7df3db38dc96f4e20d05457d ("Don't delete
thread_info if refcount isn't zero") eliminated
restore_current_thread_ptid_changed, so current_thread_cleanup_chain
is no longer necessary either.

gdb/ChangeLog:
2017-04-10  Pedro Alves  <palves@redhat.com>

* thread.c (struct current_thread_cleanup) <next>: Delete field.
(current_thread_cleanup_chain): Delete.
(restore_current_thread_cleanup_dtor)
(make_cleanup_restore_current_thread): Remove references to
current_thread_cleanup_chain.

7 years agomsp430: Don't use REG_UNKNOWN
Alan Hayward [Mon, 10 Apr 2017 14:01:53 +0000 (15:01 +0100)]
msp430: Don't use REG_UNKNOWN

gdb/
* msp430-tdep.c (msp430_pseudo_register_read): Never return
REG_UNKNOWN.

7 years agoDon't delete thread_info if refcount isn't zero
Yao Qi [Mon, 10 Apr 2017 13:39:41 +0000 (14:39 +0100)]
Don't delete thread_info if refcount isn't zero

I build GDB with asan, and run test case hook-stop.exp, and threadapply.exp,
I got the following asan error,

=================================================================^M
^[[1m^[[31m==2291==ERROR: AddressSanitizer: heap-use-after-free on address 0x6160000999c4 at pc 0x000000826022 bp 0x7ffd28a8ff70 sp 0x7ffd28a8ff60^M
^[[1m^[[0m^[[1m^[[34mREAD of size 4 at 0x6160000999c4 thread T0^[[1m^[[0m^M
    #0 0x826021 in release_stop_context_cleanup ../../binutils-gdb/gdb/infrun.c:8203^M
    #1 0x72798a in do_my_cleanups ../../binutils-gdb/gdb/common/cleanups.c:154^M
    #2 0x727a32 in do_cleanups(cleanup*) ../../binutils-gdb/gdb/common/cleanups.c:176^M
    #3 0x826895 in normal_stop() ../../binutils-gdb/gdb/infrun.c:8381^M
    #4 0x815208 in fetch_inferior_event(void*) ../../binutils-gdb/gdb/infrun.c:4011^M
    #5 0x868aca in inferior_event_handler(inferior_event_type, void*) ../../binutils-gdb/gdb/inf-loop.c:44^M
....
^[[1m^[[32m0x6160000999c4 is located 68 bytes inside of 568-byte region [0x616000099980,0x616000099bb8)^M
^[[1m^[[0m^[[1m^[[35mfreed by thread T0 here:^[[1m^[[0m^M
    #0 0x7fb0bc1312ca in __interceptor_free (/usr/lib/x86_64-linux-gnu/libasan.so.2+0x982ca)^M
    #1 0xb8c62f in xfree(void*) ../../binutils-gdb/gdb/common/common-utils.c:100^M
    #2 0x83df67 in free_thread ../../binutils-gdb/gdb/thread.c:207^M
    #3 0x83dfd2 in init_thread_list() ../../binutils-gdb/gdb/thread.c:223^M
    #4 0x805494 in kill_command ../../binutils-gdb/gdb/infcmd.c:2595^M
....

Detaching from program: /home/yao.qi/SourceCode/gnu/build-with-asan/gdb/testsuite/outputs/gdb.threads/threadapply/threadapply, process 2399^M
=================================================================^M
^[[1m^[[31m==2387==ERROR: AddressSanitizer: heap-use-after-free on address 0x6160000a98c0 at pc 0x00000083fd28 bp 0x7ffd401c3110 sp 0x7ffd401c3100^M
^[[1m^[[0m^[[1m^[[34mREAD of size 4 at 0x6160000a98c0 thread T0^[[1m^[[0m^M
    #0 0x83fd27 in thread_alive ../../binutils-gdb/gdb/thread.c:741^M
    #1 0x844277 in thread_apply_all_command ../../binutils-gdb/gdb/thread.c:1804^M
....
^M
^[[1m^[[32m0x6160000a98c0 is located 64 bytes inside of 568-byte region [0x6160000a9880,0x6160000a9ab8)^M
^[[1m^[[0m^[[1m^[[35mfreed by thread T0 here:^[[1m^[[0m^M
    #0 0x7f59a7e322ca in __interceptor_free (/usr/lib/x86_64-linux-gnu/libasan.so.2+0x982ca)^M
    #1 0xb8c62f in xfree(void*) ../../binutils-gdb/gdb/common/common-utils.c:100^M
    #2 0x83df67 in free_thread ../../binutils-gdb/gdb/thread.c:207^M
    #3 0x83dfd2 in init_thread_list() ../../binutils-gdb/gdb/thread.c:223^M

This patch fixes the issue by deleting thread_info object if it is
deletable, otherwise, mark it as exited (by set_thread_exited).
Function set_thread_exited is shared from delete_thread_1.  This patch
also moves field "refcount" to private and methods incref and
decref.  Additionally, we stop using "ptid_t" in
"struct current_thread_cleanup" to reference threads, instead we use
"thread_info" directly.  Due to this change, we don't need
restore_current_thread_ptid_changed anymore.

gdb:

2017-04-10  Yao Qi  <yao.qi@linaro.org>

PR gdb/19942
* gdbthread.h (thread_info::deletable): New method.
(thread_info::incref): New method.
(thread_info::decref): New method.
(thread_info::refcount): Move it to private.
* infrun.c (save_stop_context): Call inc_refcount.
(release_stop_context_cleanup): Likewise.
* thread.c (set_thread_exited): New function.
(init_thread_list): Delete "tp" only it is deletable, otherwise
call set_thread_exited.
(delete_thread_1): Call set_thread_exited.
(current_thread_cleanup) <inferior_pid>: Remove.
<thread>: New field.
(restore_current_thread_ptid_changed): Removed.
(do_restore_current_thread_cleanup): Adjust.
(restore_current_thread_cleanup_dtor): Don't call
find_thread_ptid.
(set_thread_refcount): Use dec_refcount.
(make_cleanup_restore_current_thread): Adjust.
(thread_apply_all_command): Call inc_refcount.
(_initialize_thread): Don't call
observer_attach_thread_ptid_changed.

7 years agoHoist code on marking thread as exited
Yao Qi [Mon, 10 Apr 2017 13:39:41 +0000 (14:39 +0100)]
Hoist code on marking thread as exited

This patch hoists code on marking thread as exited, so more code is shared
for two different paths (thread_info is deleted or is not deleted).

gdb:

2017-04-10  Yao Qi  <yao.qi@linaro.org>

* thread.c (delete_thread_1): Hoist code on marking thread as
exited.

7 years agogas: xtensa: fix incorrect code generated with auto litpools
Max Filippov [Mon, 10 Apr 2017 12:12:52 +0000 (13:12 +0100)]
gas: xtensa: fix incorrect code generated with auto litpools

* config/tc-xtensa.c (xtensa_maybe_create_literal_pool_frag):
Initialize lps->frag_count with auto_litpool_limit.
(xg_promote_candidate_litpool): New function.
(xtensa_move_literals): Extract candidate litpool promotion code
into separate function. Call it for all possible found
candidates.
(xtensa_switch_to_literal_fragment): Drop 'recursive' flag and
call to xtensa_mark_literal_pool_location that it guards.
Replace it with call to xtensa_maybe_create_literal_pool_frag.
Initialize pool_location with created literal pool candidate.
* testsuite/gas/xtensa/all.exp: Add new tests.
* testsuite/gas/xtensa/auto-litpools-first1.d: New test results.
* testsuite/gas/xtensa/auto-litpools-first1.s: New test.
* testsuite/gas/xtensa/auto-litpools-first2.d: New test results.
* testsuite/gas/xtensa/auto-litpools-first2.s: New test.
* testsuite/gas/xtensa/auto-litpools.d: Fix offsets changed due
to additional jump instruction.

7 years agoPort the bug fix for PR 19704 (Missing dynamic relocation against undefined weak...
Qing Zhao [Mon, 10 Apr 2017 11:46:30 +0000 (12:46 +0100)]
Port the bug fix for PR 19704 (Missing dynamic relocation against undefined weak symbol) to the SPARC architecture.

       * elf32-sparc.c (elf_backend_fixup_symbol): New.
        * elf64-sparc.c (elf_backend_fixup_symbol): New.
        * elfxx-sparc.c (UNDEFINED_WEAK_RESOLVED_TO_ZERO): New.
        (_bfd_sparc_elf_link_hash_entry): Add has_got_reloc and
        has_non_got_reloc.
        (link_hash_newfunc): Initialize has_got_reloc and
has_non_got_reloc.
        (_bfd_sparc_elf_size_dynamic_sections): Set interp to .interp
        section.
        (_bfd_sparc_elf_copy_indirect_symbol): Copy has_got_reloc and
        has_non_got_reloc.
        (_bfd_sparc_elf_check_relocs): Set has_got_reloc and
        has_non_got_reloc.
        (_bfd_sparc_elf_fixup_symbol): New function.
        (allocate_dynrelocs): Don't allocate space for dynamic
        relocations and discard relocations against resolved undefined
        weak symbols in executable.  Don't make resolved undefined weak
        symbols in executable dynamic.  Keep dynamic non-GOT/non-PLT
        relocation against undefined weak symbols in PIE.
        (_bfd_sparc_elf_relocate_section): Don't generate dynamic
        relocations against resolved undefined weak symbols in PIE
        (_bfd_sparc_elf_finish_dynamic_symbol): Keep PLT/GOT entries
        without ynamic PLT/GOT relocations for resolved undefined weak
        symbols.
        Don't generate dynamic relocation against resolved undefined
        weak symbol in executable.
        (pie_finish_undefweak_symbol): New function.
        (_bfd_sparc_elf_finish_dynamic_sections): Call
        pie_finish_undefweak_symbol on all symbols in PIE.
        * elfxx-sparc.h (_bfd_sparc_elf_link_hash_table): Add interp.
        (_bfd_sparc_elf_fixup_symbol): New function.

7 years agoRemove the ns32k target from the obsolete list.
Nick Clifton [Mon, 10 Apr 2017 10:38:21 +0000 (11:38 +0100)]
Remove the ns32k target from the obsolete list.

* config.bfd: Remove ns32k from obsolete list.