* config/tc-ppc.h (struct ppc_tc_sy): Complete comment on within.
(tc_new_dot_label): Define.
(ppc_new_dot_label): Declare.
* config/tc-ppc.c (ppc_frob_label): Set within target field.
(ppc_fix_adjustable): Use this field to adjust the reloc.
(ppc_new_dot_label): New function.
Alan Modra [Tue, 14 Jun 2011 02:45:12 +0000 (02:45 +0000)]
PR ld/12851
* elflink.c (_bfd_elf_gc_mark_extra_sections): New function.
(elf_gc_sweep): Don't treat debug and sections like .comment
specially here.
(bfd_elf_gc_sections): Treat note sections as gc roots only when
not part of a group. Always call gc_mark_extra_sections.
* elf-bfd.h (_bfd_elf_gc_mark_extra_sections): Declare.
* elfxx-target.h (elf_backend_gc_mark_extra_sections): Default to
_bfd_elf_gc_mark_extra_sections.
* elf32-arm.c (elf32_arm_gc_mark_extra_sections): Call
_bfd_elf_gc_mark_extra_sections.
* elf32-tic6x.c (elf32_tic6x_gc_mark_extra_sections): Likewise.
* common.h: Add EM_TILEGX.
* tilegx.h: New file.
* tilepro.h: New file.
* Makefile.am (ALL_EMULATION_SOURCES): Add eelf32tilegx.c and
eelf32tilepro.c.
(ALL_64_EMULATION_SOURCES): Add eelf64tilegx.c.
(eelf32tilegx.c): New target.
(eelf32tilepro.c): Likewise.
(eelf64tilegx.c): Likewise.
* Makefile.in: Regenerate.
* configure.tgt: Handle tilegx-*-* and tilepro-*-*.
* emulparams/elf32tilegx.sh: New file.
* emulparams/elf64tilegx.sh: New file.
* emulparams/elf32tilepro.sh: New file.
* ld-elf/eh5.d: Don't run on tile*.
* ld-srec/srec.exp: xfail on tile*.
* ld-tilegx/external.s: New file.
* ld-tilegx/reloc.d: New file.
* ld-tilegx/reloc.s: New file.
* ld-tilegx/tilegx.exp: New file.
* ld-tilepro/external.s: New file.
* ld-tilepro/reloc.d: New file.
* ld-tilepro/reloc.s: New file.
* ld-tilepro/tilepro.exp: New file.
* Makefile.am (TARGET_LIBOPCODES_CFILES): Add tilegx-dis.c,
tilegx-opc.c, tilepro-dis.c, and tilepro-opc.c.
* Makefile.in: Regenerate.
* configure.in: Handle bfd_tilegx_arch and bfd_tilepro_arch.
* configure: Regenerate.
* disassemble.c (disassembler): Add ARCH_tilegx and ARCH_tilepro.
* po/POTFILES.in: Regenerate.
* tilegx-dis.c: New file.
* tilegx-opc.c: New file.
* tilepro-dis.c: New file.
* tilepro-opc.c: New file.
Mark Kettenis [Sun, 12 Jun 2011 20:46:11 +0000 (20:46 +0000)]
* i386-tdep.c (i386_epilogue_frame_cache): Simplify code. Call
get_frame_func instead of get_frame_pc to determine the code
address used to construct the frame ID.
(i386_epilogue_frame_unwind_stop_reason): Fix coding style.
(i386_epilogue_frame_this_id): Likewise.
(i386_epilogue_frame_prev_register): New function.
(i386_epilogue_frame_unwind): Use i386_epilogue_frame_prev_register.
(i386_stack_tramp_frame_sniffer): Fix coding style.
(i386_stack_tramp_frame_unwind): Use i386_epilogue_frame_prev_register.
(i386_gdbarch_init): Fix comments.
Jan Kratochvil [Fri, 10 Jun 2011 21:48:04 +0000 (21:48 +0000)]
gdb/
Code cleanup.
* dwarf2read.c (dw2_map_symbol_filenames): Use symbol_filename_ftype
for fun.
* psymtab.c (map_symbol_filenames_psymtab)
(map_partial_symbol_filenames): Likewise.
* psymtab.h: Include symfile.h.
(map_partial_symbol_filenames): Use symbol_filename_ftype for fun.
* symfile.h (symbol_filename_ftype): New.
(struct quick_symbol_functions): Use symbol_filename_ftype for fun of
map_symbol_filenames, clarify more the naming in comment.
Tom Tromey [Fri, 10 Jun 2011 16:21:47 +0000 (16:21 +0000)]
* gdb.cp/temargs.exp: Let tests pass if compiler bug is fixed.
Add tests for pointer-to-member-function.
* gdb.cp/temargs.cc (S::somefunc): New function.
(K2): New class.
(main): Instantiate K2; call method.
Nick Clifton [Fri, 10 Jun 2011 14:04:25 +0000 (14:04 +0000)]
* elflink.c (_bfd_elf_link_create_dynamic_sections): If the
backend does not provide a function for creating dynamic sections
then fail.
(bfd_elf_final_link): Issue a warning message if a dynamic section
has the SHT_NOTE type.
(bfd_elf_final_link): Do not look for dynamic strings in a section
that does not have the SHT_STRTAB type or the name .dynstr.
* elf32-arm.c (elf32_arm_finish_dynamic_sections): Fail if the got
section is not in the output binary.
* elf32-hppa.c (elf32_hppa_finish_dynamic_sections): Likewise.
* ld-elf/elf.exp: Add test for linking a shared library with a
broken linker script that marks dynamic sections as being notes.
* ld-elf/note-3.s: New test source file.
* ld-elf/note-3.t: New test linker script.
* ld-elf/note-3.l: Expected output from the linker.
* lib/ld-lib.exp (run_ld_link_tests): Improve description.
Cary Coutant [Thu, 9 Jun 2011 18:18:44 +0000 (18:18 +0000)]
PR gold/12804
* gold/gold.cc (queue_initial_tasks): Warn if --incremental is
used with --compress-debug-sections.
* gold/object.cc (Sized_relobj_file::do_layout): Report
uncompressed size of compressed input sections.
Nick Clifton [Thu, 9 Jun 2011 15:43:52 +0000 (15:43 +0000)]
PR gas/12861
* config/tc-cr16.c (tc_gen_reloc): Remove unused local variable
code.
(check_cinv_options): Remove unused local variables. Make
function void.
(md_assemble): Remove unused local variable.
Nick Clifton [Thu, 9 Jun 2011 10:36:20 +0000 (10:36 +0000)]
PR ld/12845
* elf.c (_bfd_elf_init_private_section_data): Add an assertion
that the output section has an allocated bfd_elf_section_data
structure.
* elfxx-mips.c (mips_elf_check_symbols): Do not create a stub for
symbols in sections that have been removed by garbage collection.
* emultempl/mipself.em (mips_add_stub_section): Do not add stubs
for sections that have been removed by garbage collection.
Joel Brobecker [Wed, 8 Jun 2011 16:56:11 +0000 (16:56 +0000)]
py-inferior.exp: Make sure local var is allocated on the stack.
The testcase, at some point, is trying to change the contents
of a string that was defined as follow:
char *str = "hello, testsuite";
The problem is that the string is constant, and str is never used
to change the contents of the string in the program, so the compiler
is free to allocate it in a read-only section. This is what happens
on x86-windows, for instance.
As a result, trying to change the contents of the string during
the `python gdb.inferiors()[0].write_memory (addr, str)' results
in the following error:
(gdb) python gdb.inferiors()[0].write_memory (addr, str)
gdb: write target memory, 5 bytes at 0x00403064
Traceback (most recent call last):
File "<string>", line 1, in <module>
gdb.MemoryError: Cannot access memory at address 0x403064
Error while executing Python code.
This patch prevents this from happening by declaring str as an
array rather than a pointer.
gdb/testsuite/ChangeLog:
* gdb.python/py-inferior.c (f2): Make str an array rather
than a pointer.
* gdb.python/py-inferior.exp: Adjust testcase accordingly.
Joel Brobecker [Wed, 8 Jun 2011 16:44:28 +0000 (16:44 +0000)]
ppc sim: Improve invalid option error message
This patch improves the error message when an invalid option is
detected, by also printing the option that actually caused the error.
For instance, from GDB:
We also added the usage after an invalid long-name option (Eg: --hello)
to be in line with what's being done for all other invalid options
being detected.
sim/ppc/ChangeLog:
* psim.c (psim_options): Add option that cause the error
in invalid-option error messages. Print the usage when
detecting an invalid long-name option.
Joel Brobecker [Wed, 8 Jun 2011 16:44:20 +0000 (16:44 +0000)]
ppc sim: Allow --sysroot command-line option
There was a recent change that cuased the "target sim" command
to add a --sysroot option to the argument vector passed down to
the simulator. This caused a failure in the powerpc simulator,
as it did not recognize it. This patch fixes the problem by adding
support for the --sysroot option (it ignores it).
sim/ppc/ChangeLog:
* psim.c (psim_options): Accept and ignore `--sysroot=...'.
Nick Clifton [Wed, 8 Jun 2011 15:59:07 +0000 (15:59 +0000)]
PR binutils/12855
* readelf.c (process_version_sections): Handle binaries containing
corrupt version information.
(process_symbol_table): Stop processing a symbol's version
information if it could not be read in.
(get_data): Add comment describing the function.
(process_section_headers): Set dynamic_strings_length to 0 if the
dynamic strings could not be read in.
(process_dynamic_section): Likewise.
(process_section_groups): Stop processing the group information if
the data could not be read in.
(hppa_processs_unwind): Assert that there is only one string table
in the file.
(arm_process_unwind): Likewise.
(ia64_process_unwind): Likewise.
Set the size of the unwind auxillary information to 0 if the data
could not be read.
(load_specific_debug_section): Handle a failure to read in the
section.
(process_mips_specific): Stop display of the PLT GOT section if it
could not be read in.
Cary Coutant [Wed, 8 Jun 2011 04:34:22 +0000 (04:34 +0000)]
* layout.cc (Layout::set_segment_offsets): Don't adjust layout
for incremental links.
* output.cc (Output_segment::set_section_list_addresses): Remove
FIXME and test for TLS or BSS.
Cary Coutant [Wed, 8 Jun 2011 03:50:12 +0000 (03:50 +0000)]
* copy-relocs.cc (Copy_relocs::copy_reloc): Call make_copy_reloc
instead of emit_copy_reloc.
(Copy_relocs::emit_copy_reloc): Refactor.
(Copy_relocs::make_copy_reloc): New function.
(Copy_relocs::add_copy_reloc): Remove.
* copy-relocs.h (Copy_relocs::emit_copy_reloc): Move to public
section.
(Copy_relocs::make_copy_reloc): New function.
(Copy_relocs::add_copy_reloc): Remove.
* gold.cc (queue_middle_tasks): Emit old COPY relocations from
unchanged input files.
* incremental-dump.cc (dump_incremental_inputs): Print "COPY" flag.
* incremental.cc (Sized_incremental_binary::do_reserve_layout):
Reserve BSS space for COPY relocations.
(Sized_incremental_binary::do_emit_copy_relocs): New function.
(Output_section_incremental_inputs::write_info_blocks): Record
whether a symbol is copied from a shared object.
(Sized_incr_dynobj::do_add_symbols): Record COPY relocations.
* incremental.h (enum Incremental_shlib_symbol_flags): New type.
(INCREMENTAL_SHLIB_SYM_FLAGS_SHIFT): New constant.
(Incremental_input_entry_reader::get_output_symbol_index): Add
is_copy parameter.
(Incremental_binary::emit_copy_relocs): New function.
(Incremental_binary::do_emit_copy_relocs): New function.
(Sized_incremental_binary::Sized_incremental_binary): Initialize
new data member.
(Sized_incremental_binary::add_copy_reloc): New function.
(Sized_incremental_binary::do_emit_copy_relocs): New function.
(Sized_incremental_binary::Copy_reloc): New struct.
(Sized_incremental_binary::Copy_relocs): New typedef.
(Sized_incremental_binary::copy_relocs_): New data member.
* symtab.cc (Symbol_table::add_from_incrobj): Change return type.
* symtab.h (Symbol_table::add_from_incrobj): Change return type.
* target.h (Sized_target::emit_copy_reloc): New function.
* x86_64.cc (Target_x86_64::emit_copy_reloc): New function.
2011-05-30 Chris Metcalf <cmetcalf@tilera.com>
Ben Elliston <bje@gnu.org>
* config.guess (tile*:Linux:*:*): Use vendor "unknown", not
"tilera", for consistency with other architectures.
* config.sub (tile*-*, tilegx-*): Use a more general pattern for
"tile" to allow matching tilepro, tilegx32 and other variants.
Mike Frysinger [Sat, 4 Jun 2011 17:44:22 +0000 (17:44 +0000)]
sim: bfin: import testsuite
Now that the common sim testsuite code supports .S and .c files, we
can import the Blackfin testsuite. There are about ~800 tests here,
so I'm only attaching a compressed patch of them. Other than adding
files to sim/testsuite/sim/bfin/, the sim/configure.tgt file was
updated to mark Blackfin as having a testsuite, and sim/configure
regenerated.
Mike Frysinger [Sat, 4 Jun 2011 17:18:04 +0000 (17:18 +0000)]
sim: bfin: add support for glued SIC interrupt lines
The BF537 family glues a bunch of peripherals into single interrupt lines
that run into the SIC. To model this same behavior in the sim, we need to
use the glue-or device, and in order to use that, we need to tweak things
a bit in the mach code to allow declaring of these new devices.
Mike Frysinger [Sat, 4 Jun 2011 17:11:19 +0000 (17:11 +0000)]
sim: bfin: push SIC mappings to device tree
The machs.c file is the best place for holding cpu-specific details, so
restructure the way the SIC manages its ports to do just that. Now the
SIC's have a standard set of input pins and the different line routing
from peripherals is kept in the device tree only. This better models
the hardware where the SIC doesn't care about the exact peripheral that
is sending it stuff, just which input pin it gets it on.