For binary ouput, we don't have an ELF bfd output so can't access
elf_elfheader. The elf64-ppc.c changes are really just a tidy,
triggered by looking at all places where the abiversion bits are
accessed.
bfd/
* elf64-ppc.c (ppc64_elf_before_check_relocs): Do .opd processing
even when output is not ppc64 ELF. Remove redundant tests on
type of input bfd.
ld/
PR 17488
* emultempl/ppc64elf.em (gld${EMULATION_NAME}_finish): Don't attempt
to access ELF header e_flags when not ppc64 ELF output.
* ldlex.l (INPUTLIST): New start condition.
(comment pattern, ",", "(", ")", "AS_NEEDED")
({FILENAMECHAR1}{FILENAMECHAR}*, "-l"{FILENAMECHAR}+)
(quoted string pattern, whitespace pattern): Add INPUTLIST to
valid start conditions.
(<INPUTLIST>"="{FILENAMECHAR1}{FILENAMECHAR}*): New NAME rule.
(ldlex_inputlist): New start-condition-setter function.
* ldgram.y (input_list1): Rename from input_list. All recursive
use changed.
(input_list): New wrapper rule for input_list1, setting
INPUTLIST lexer state for the duration of parsing input_list1.
All this to say INPUT(=/path/to/file) and not be forced to use
INPUT("=/path/to/file") whenever there's a need to force a sysroot-
prefix. Still, IMHO it seems better to make use of a previously
invalid syntax and not only change the meaning of quoted =-prefixed
paths (though arguably that's not very useful before this patchset).
This got a little bit hairier than I'd expected: I had to add a new
lexer state (aka. start condition) to avoid a first "=" being lexed as
the token "=", despite that not making sense in constructs expecting
file-names in the first place. (The grammar doesn't allow for
expressions in any part of those lists.) I guess I *could* have made
it work using that token anyway, but I didn't like the idea that you
would be able to separate the "=" from the rest of the file-name with
whitespace.
* ldlang.c (lang_add_input_file): If the first character in the
filename is '=', prepend the sysroot and force the context of that
input file to non-sysroot.
The "input_flags.sysrooted = 0" thing described in the comment is
covered by the testsuite part ("root-anchored =-prefixed script
inside"), but only observable for --with-sysroot configurations.
* ld.texinfo (input files in linker scripts): When mentioning
behavior of first character "/" on scripts within sysroot, also
mention that effect can be forced by prefixing with "=" and
refer to SEARCH_DIR.
The idea here is to drop .eh_frame FDEs corresponding to dropped
comdat group sections or linkonce sections, but not perform changes in
encoding.
bfd/
PR 17467
* elf-eh-frame.c (ENSURE_NO_RELOCS): Don't stop at first NONE reloc.
(_bfd_elf_parse_eh_frame): When relocatable output, don't set
flags enabling conversion of CIEs and FDEs to use relative encoding.
(find_merged_cie): Similarly.
(_bfd_elf_write_section_eh_frame): Don't edit FDEs when
relocatable, except for CIE pointer.
* elflink.c (bfd_elf_reloc_symbol_deleted_p): Return true for
relocs against symbols in dropped comdat group sections.
(bfd_elf_discard_info): Do some eh_frame optimisation when
relocatable.
ld/
* ldlang.c (lang_add_section): Set up map_head.s and map_tail.s when
relocatable.
Refactor each relaxation pattern to raise the maintainability.
In origin, all patterns is analysed in nds32_elf_relax_section,
so it is hard to debug and maintain. Therefore, we classify all
patterns into different functions in this patch.
Moreover, we adjust all optimizations into nds32_elf_relax_section
to take these optimizations in turn. This can promise all relaxation
being done after calling gld${EMULATION_NAME}_after_allocation.
bfd/
* aoutx.h (NAME (aout, machine_type)): Add mips32r6 and mips64r6.
* archures.c (bfd_architecture): Likewise.
* bfd-in2.h (bfd_architecture): Likewise.
(bfd_reloc_code_real): Add relocs BFD_RELOC_MIPS_21_PCREL_S2,
BFD_RELOC_MIPS_26_PCREL_S2, BFD_RELOC_MIPS_18_PCREL_S3 and
BFD_RELOC_MIPS_19_PCREL_S2.
* cpu-mips.c (arch_info_struct): Add mips32r6 and mips64r6.
* elf32-mips.c: Define relocs R_MIPS_PC21_S2, R_MIPS_PC26_S2
R_MIPS_PC18_S3, R_MIPS_PC19_S2, R_MIPS_PCHI16 and R_MIPS_PCLO16.
(mips_reloc_map): Add entries for BFD_RELOC_MIPS_21_PCREL_S2,
BFD_RELOC_MIPS_26_PCREL_S2, BFD_RELOC_MIPS_18_PCREL_S3,
BFD_RELOC_MIPS_19_PCREL_S2, BFD_RELOC_HI16_S_PCREL and
BFD_RELOC_LO16_PCREL.
* elf64-mips.c: Define REL, and RELA relocations R_MIPS_PC21_S2,
R_MIPS_PC26_S2, R_MIPS_PC18_S3, R_MIPS_PC19_S2, R_MIPS_PCHI16
and R_MIPS_PCLO16.
(mips_reloc_map): Add entries for BFD_RELOC_MIPS_21_PCREL_S2,
BFD_RELOC_MIPS_26_PCREL_S2, BFD_RELOC_MIPS_18_PCREL_S3,
BFD_RELOC_MIPS_19_PCREL_S2, BFD_RELOC_HI16_S_PCREL and
BFD_RELOC_LO16_PCREL.
* elfn32-mips.c: Likewise.
* elfxx-mips.c (MIPSR6_P): New define.
(mipsr6_exec_plt_entry): New array.
(hi16_reloc_p): Add support for R_MIPS_PCHI16.
(lo16_reloc_p): Add support for R_MIPS_PCLO16.
(aligned_pcrel_reloc_p): New function.
(mips_elf_relocation_needs_la25_stub): Add support for relocs:
R_MIPS_PC21_S2 and R_MIPS_PC26_S2.
(mips_elf_calculate_relocation): Add support for relocs:
R_MIPS_PC21_S2, R_MIPS_PC26_S2, R_MIPS_PC18_S3, R_MIPS_PC19_S2,
R_MIPS_PCHI16 and R_MIPS_PCLO16.
(_bfd_elf_mips_mach): Add support for mips32r6 and mips64r6.
(mips_elf_add_lo16_rel_addend): Add support for R_MIPS_PCHI16.
(_bfd_mips_elf_check_relocs): Add support for relocs:
R_MIPS_PC21_S2 and R_MIPS_PC26_S2.
(_bfd_mips_elf_relocate_section): Add a check for unaligned
pc relative relocs.
(_bfd_mips_elf_finish_dynamic_symbol): Add support for MIPS r6
plt entry.
(mips_set_isa_flags): Add support for mips32r6 and mips64r6.
(_bfd_mips_elf_print_private_bfd_data): Likewise.
(mips_32bit_flags_p): Add support for mips32r6.
* libbfd.h (bfd_reloc_code_real_names): Add entries for
BFD_RELOC_MIPS_21_PCREL_S2, BFD_RELOC_MIPS_26_PCREL_S2,
BFD_RELOC_MIPS_18_PCREL_S3 and BFD_RELOC_MIPS_19_PCREL_S2.
* reloc.c: Document relocs BFD_RELOC_MIPS_21_PCREL_S2,
BFD_RELOC_MIPS_26_PCREL_S2, BFD_RELOC_MIPS_18_PCREL_S3 and
BFD_RELOC_MIPS_19_PCREL_S2.
binutils/
* readelf.c (get_machine_flags): Add support for mips32r6 and
mips64r6.
elfcpp/
* mips.h (E_MIPS_ARCH_32R6, E_MIPS_ARCH_64R6): New enum constants.
gas/
* config/tc-mips.c (mips_nan2008): New static global.
(mips_flag_nan2008): Removed.
(LL_SC_FMT): New define.
(COP12_FMT): Updated.
(ISA_IS_R6): New define.
(ISA_HAS_64BIT_REGS): Add mips64r6.
(ISA_HAS_DROR): Likewise.
(ISA_HAS_64BIT_FPRS): Add mips32r6 and mips64r6.
(ISA_HAS_ROR): Likewise.
(ISA_HAS_ODD_SINGLE_FPR): Likewise.
(ISA_HAS_MXHC1): Likewise.
(hilo_interlocks): Likewise.
(md_longopts): Likewise.
(ISA_HAS_LEGACY_NAN): New define.
(options): Add OPTION_MIPS32R6 and OPTION_MIPS64R6.
(mips_ase): Add field rem_rev.
(mips_ases): Updated to add which ISA an ASE was removed in.
(mips_isa_rev): Add support for mips32r6 and mips64r6.
(mips_check_isa_supports_ase): Add support to check if an ASE
has been removed in the specified MIPS ISA revision.
(validate_mips_insn): Skip '-' character.
(macro_build): Likewise.
(mips_check_options): Prevent R6 working with fp32, mips16,
micromips, or branch relaxation.
(file_mips_check_options): Set R6 floating point registers to
64 bit. Also deal with the nan2008 option.
(limited_pcrel_reloc_p): Add relocs: BFD_RELOC_MIPS_21_PCREL_S2,
BFD_RELOC_MIPS_26_PCREL_S2, BFD_RELOC_MIPS_18_PCREL_S3,
BFD_RELOC_MIPS_19_PCREL_S2, BFD_RELOC_HI16_S_PCREL and
BFD_RELOC_LO16_PCREL.
(operand_reg_mask): Add support for OP_SAME_RS_RT, OP_CHECK_PREV
and OP_NON_ZERO_REG.
(match_check_prev_operand): New static function.
(match_same_rs_rt_operand): New static function.
(match_non_zero_reg_operand): New static function.
(match_operand): Added entries for: OP_SAME_RS_RT, OP_CHECK_PREV
and OP_NON_ZERO_REG.
(insns_between): Added case to deal with forbidden slots.
(append_insn): Added support for relocs: BFD_RELOC_MIPS_21_PCREL_S2
and BFD_RELOC_MIPS_26_PCREL_S2.
(match_insn): Add support for operands -A, -B, +' and +". Also
skip '-' character.
(mips_percent_op): Add entries for %pcrel_hi and %pcrel_lo.
(md_parse_option): Add support for mips32r6 and mips64r6. Also
update the nan option handling.
(md_pcrel_from): Add cases for relocs: BFD_RELOC_MIPS_21_PCREL_S2,
BFD_RELOC_MIPS_26_PCREL_S2.
(mips_force_relocation): Prevent forced relaxation for MIPS r6.
(md_apply_fix): Add support for relocs: BFD_RELOC_MIPS_21_PCREL_S2,
BFD_RELOC_MIPS_26_PCREL_S2, BFD_RELOC_MIPS_18_PCREL_S3,
BFD_RELOC_MIPS_19_PCREL_S2, BFD_RELOC_HI16_S_PCREL and
BFD_RELOC_LO16_PCREL.
(s_mipsset): Add support for mips32r6 and mips64r6.
(s_nan): Update to support the new nan2008 framework.
(tc_gen_reloc): Add relocs: BFD_RELOC_MIPS_21_PCREL_S2,
BFD_RELOC_MIPS_26_PCREL_S2, BFD_RELOC_MIPS_18_PCREL_S3,
BFD_RELOC_MIPS_19_PCREL_S2, BFD_RELOC_HI16_S_PCREL and
BFD_RELOC_LO16_PCREL.
(mips_elf_final_processing): Updated to use the mips_nan2008.
(mips_cpu_info_table): Add entries for mips32r6 and mips64r6.
(macro): Enable ldc2, sdc2, ll, lld, swc2, sc, scd, cache, pref
macros for R6.
(mips_fix_adjustable): Make PC relative R6 relocations relative
to the symbol and not the section.
* configure.ac: Add support for mips32r6 and mips64r6.
* configure: Regenerate.
* doc/c-mips.texi: Document the -mips32r6 and -mips64r6 command line
options.
* doc/as.texinfo: Likewise.
gas/testsuite/
* gas/mips/24k-triple-stores-1.s: If testing for r6 prevent
non-supported instructions from being tested.
* gas/mips/24k-triple-stores-2.s: Likewise.
* gas/mips/24k-triple-stores-3.s: Likewise.
* gas/mips/24k-triple-stores-6.s: Likewise.
* gas/mips/beq.s: Likewise.
* gas/mips/eva.s: Likewise.
* gas/mips/ld-zero-3.s: Likewise.
* gas/mips/mips32-cp2.s: Likewise.
* gas/mips/mips32.s: Likewise.
* gas/mips/mips4.s: Likewise.
* gas/mips/add.s: Don't test the add instructions if r6, and
add padding.
* gas/mips/add.d: Check for a triple dot not a nop at the end of the
disassembly output.
* gas/mips/micromips@add.d: Likewise.
* gas/mips/mipsr6@24k-branch-delay-1.d: New file.
* gas/mips/mipsr6@24k-triple-stores-1.d: New file.
* gas/mips/mipsr6@24k-triple-stores-2-llsc.d: New file.
* gas/mips/mipsr6@24k-triple-stores-2.d: New file.
* gas/mips/mipsr6@24k-triple-stores-3.d: New file.
* gas/mips/mipsr6@24k-triple-stores-6.d: New file.
* gas/mips/mipsr6@add.d: New file.
* gas/mips/mipsr6@attr-gnu-4-1-msingle-float.l: New file.
* gas/mips/mipsr6@attr-gnu-4-1-msingle-float.s: New file.
* gas/mips/mipsr6@attr-gnu-4-1-msoft-float.l: New file.
* gas/mips/mipsr6@attr-gnu-4-1-msoft-float.s: New file.
* gas/mips/mipsr6@attr-gnu-4-2-mdouble-float.l: New file.
* gas/mips/mipsr6@attr-gnu-4-2-mdouble-float.s: New file.
* gas/mips/mipsr6@beq.d: New file.
* gas/mips/mipsr6@bge.d: New file.
* gas/mips/mipsr6@bgeu.d: New file.
* gas/mips/mipsr6@blt.d: New file.
* gas/mips/mipsr6@bltu.d: New file.
* gas/mips/mipsr6@branch-misc-1.d: New file.
* gas/mips/mipsr6@branch-misc-2-64.d: New file.
* gas/mips/mipsr6@branch-misc-2pic-64.d: New file.
* gas/mips/mipsr6@branch-misc-4-64.d: New file.
* gas/mips/mipsr6@cache.d: New file.
* gas/mips/mipsr6@eva.d: New file.
* gas/mips/mipsr6@jal-svr4pic-noreorder.d: New file.
* gas/mips/mipsr6@jal-svr4pic.d: New file.
* gas/mips/mipsr6@ld-zero-2.d: New file.
* gas/mips/mipsr6@ld-zero-3.d: New file.
* gas/mips/mipsr6@loc-swap-dis.d: New file.
* gas/mips/mipsr6@mips32-cp2.d: New file.
* gas/mips/mipsr6@mips32-imm.d: New file.
* gas/mips/mipsr6@mips32.d: New file.
* gas/mips/mipsr6@mips32r2.d: New file.
* gas/mips/mipsr6@mips4-fp.d: New file.
* gas/mips/mipsr6@mips4-fp.l: New file.
* gas/mips/mipsr6@mips4-fp.s: New file.
* gas/mips/mipsr6@mips4.d: New file.
* gas/mips/mipsr6@mips5-fp.d: New file.
* gas/mips/mipsr6@mips5-fp.l: New file.
* gas/mips/mipsr6@mips5-fp.s: New file.
* gas/mips/mipsr6@mips64.d: New file.
* gas/mips/mipsr6@msa-branch.d: New file.
* gas/mips/mipsr6@msa.d: New file.
* gas/mips/mipsr6@pref.d: New file.
* gas/mips/mipsr6@relax-swap3.d: New file.
* gas/mips/r6-64-n32.d: New file.
* gas/mips/r6-64-n64.d: New file.
* gas/mips/r6-64-removed.l: New file.
* gas/mips/r6-64-removed.s: New file.
* gas/mips/r6-64.s: New file.
* gas/mips/r6-attr-none-double.d: New file.
* gas/mips/r6-n32.d: New file.
* gas/mips/r6-n64.d: New file.
* gas/mips/r6-removed.l: New file.
* gas/mips/r6-removed.s: New file.
* gas/mips/r6.d: New file.
* gas/mips/r6.s: New file.
* gas/mips/mipsr6@mips32-dsp.d: New file.
* gas/mips/mipsr6@mips32-dspr2.d: New file.
* gas/mips/mipsr6@mips32r2-ill.l: New file.
* gas/mips/mipsr6@mips32r2-ill.s: New file.
* gas/mips/cache.s: Add r6 instruction varients.
* gas/mips/mips.exp: Add support for the mips32r6 and mips64r6
architectures. Also prevent non r6 supported tests from running.
Finally, add in support for running the new r6 tests.
(run_dump_test_arch): Add support for mipsr6 tests.
(run_list_test_arch): Add support for using files of the
form arch@testname.l .
include/elf/
* mips.h: Add relocs: R_MIPS_PC21_S2, R_MIPS_PC26_S2, R_MIPS_PC18_S3,
R_MIPS_PC19_S2, R_MIPS_PCHI16 and R_MIPS_PCLO16.
(E_MIPS_ARCH_32R6): New define.
(E_MIPS_ARCH_64R6): New define.
include/opcode/
* mips.h (mips_operand_type): Add new entries: OP_SAME_RS_RT,
OP_CHECK_PREV and OP_NON_ZERO_REG. Add descriptions for the MIPS R6
instruction arguments: -a, -b, -d, -s, -t, -u, -v, -w, -x, -y, -A, -B,
+I, +O, +R, +:, +\, +", +;
(mips_check_prev_operand): New struct.
(INSN2_FORBIDDEN_SLOT): New define.
(INSN_ISA32R6): New define.
(INSN_ISA64R6): New define.
(INSN_UPTO32R6): New define.
(INSN_UPTO64R6): New define.
(mips_isa_table): Add INSN_UPTO32R6 and INSN_UPTO64R6.
(ISA_MIPS32R6): New define.
(ISA_MIPS64R6): New define.
(CPU_MIPS32R6): New define.
(CPU_MIPS64R6): New define.
(cpu_is_member): Add cases for CPU_MIPS32R6, and CPU_MIPS64R6.
ld/
* ldmain.c (get_emulation): Add support for -mips32r6 and -mips64r6.
opcodes/
* mips-dis.c (mips_arch_choices): Add entries for mips32r6 and
mips64r6.
(parse_mips_dis_option): Allow MSA and virtualization support for
mips64r6.
(mips_print_arg_state): Add fields dest_regno and seen_dest.
(mips_seen_register): New function.
(print_insn_arg): Refactored code to use mips_seen_register
function. Add support for OP_SAME_RS_RT, OP_CHECK_PREV and
OP_NON_ZERO_REG. Changed OP_REPEAT_DEST_REG case to print out
the register rather than aborting.
(print_insn_args): Add length argument. Add code to correctly
calculate the instruction address for pc relative instructions.
(validate_insn_args): New static function.
(print_insn_mips): Prevent jalx disassembling for r6. Use
validate_insn_args.
(print_insn_micromips): Use validate_insn_args.
all the arguments are valid.
* mips-formats.h (PREV_CHECK): New define.
* mips-opc.c (decode_mips_operand): Add support for -a, -b, -d, -s,
-t, -u, -v, -w, -x, -y, -A, -B, +I, +O, +R, +:, +\, +", +;
(RD_pc): New define.
(FS): New define.
(I37): New define.
(I69): New define.
(mips_builtin_opcodes): Add MIPS R6 instructions. Exclude recoded
MIPS R6 instructions from MIPS R2 instructions.
The section name used to store the build-id on pe/coff is arbitrary, as it's
contents should be located using the pe/coff header's DataDirectory debug data
entry, not by using the section name.
But '.build-id' is not a good choice for that section name, as it is 9
characters long, and hence truncated to 8 characters when
--disable-long-section-names is used (which is the default, when producing an
executable with no dwarf debug sections, e.g. using ld --strip-all --build-id)
This truncation then breaks 'objcopy --only-keep-debug', which does use the
section name, due to concerns that keeping an arbitrary section which contains
the debug directory is not sensible.
binutils/ChangeLog
2014-09-01 Jon TURNEY <jon.turney@dronecode.org.uk>
* objcopy.c (is_nondebug_keep_contents_section): Change pe/coff
build-id section name from '.build-id' to '.buildid'.
ld/ChangeLog
2014-09-01 Jon TURNEY <jon.turney@dronecode.org.uk>
* emultempl/pe.em (write_build_id, setup_build_id): Change pe/coff
build-id section name from '.build-id' to '.buildid'.
* emultempl/pep.em (write_build_id, setup_build_id): Ditto.
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
I noticed recently that .eh_frame FDEs generated by the linker for
call stubs and .glink weren't being indexed in .eh_frame_hdr, due to
bfd_elf_discard_info being run before the linker generated .eh_frame
sections were available for parsing. This patch moves code around in
elf64-ppc.c and ppc64elf.em to avoid that problem.
Another problem fixed here is that --gc-sections parses .eh_frame
early, and the existing machinery allows only one go at parsing the
.eh_frame sections. That resulted in the linker generated .eh_frame
CIEs not being merged and no .eh_frame_hdr index entries for those
FDEs. It turns out that all the info from parsing .eh_frame is
attached to the section, so order of parsing isn't important, and
after parsing sec_info_type being set will prevent a section being
parsed again. At least, when parsing doesn't hit an error. So there
isn't really any need for "parsed_eh_frame". "merge_cies" is also
redundant, which means _bfd_elf_{begin,end}_eh_frame_parsing can also
disappear.
bfd/
* elf-bfd.h (struct eh_frame_hdr_info): Delete merge_cies and
parsed_eh_frames.
(_bfd_elf_begin_eh_frame_parsing): Delete.
(_bfd_elf_end_eh_frame_parsing): Delete.
* elf-eh-frame.c (_bfd_elf_begin_eh_frame_parsing): Delete.
(_bfd_elf_end_eh_frame_parsing): Delete.
(_bfd_elf_parse_eh_frame): Don't test parsed_eh_frame. Test
!info->relocatable in place of merge_cies.
* elflink.c (bfd_elf_gc_sections, bfd_elf_discard_info): Adjust.
* elf64-ppc.c (glink_eh_frame_cie): Pad to multiple of 8.
(ppc64_elf_size_stubs): Likewise pad stub FDE.
(ppc64_elf_build_stubs): Move code setting glink .eh_frame to..
(ppc64_elf_size_stubs): ..here and..
(ppc64_elf_finish_dynamic_sections): ..here.
ld/
* emultempl/ppc64elf.em (gld${EMULATION_NAME}_after_allocation): Call
bfd_elf_discard_info after generating glink .eh_frame. Delete
redundant test on ppc64_elf_setup_section_lists status.
At the moment it is possible to configure binutils for these triples
but the resulting linker defaults to little endian with huge numbers
of testsuite failures, which on the face of it does not appear to make
much sense.
This patch makes the behaviour similar to armeb-elf and the testsuite
is clean.
bfd/ChangeLog:
2014-08-20 Will Newton <will.newton@linaro.org>
* config.bfd: Default armeb-*-eabi* to big endian.
ld/ChangeLog:
2014-08-20 Will Newton <will.newton@linaro.org>
* configure.tgt: Default armeb-*-eabi* to big endian.
addresses outside of the 32-bit range before memory exhaustion. This
results in a higher entropy implementation of ASLR when used with the
DYNAMIC_BASE flag.
* include/coff/pe.h: Add HIGH_ENTROPY_VA flag
* ld/emultempl/pep.em: Add --high-entropy-va switch
* ld/ld.texinfo: Document the --high-entropy-va switch
Directories that don't use libtool need to add -ldl (on most *nix
hosts) to provide dlopen for libbfd.
config/
* plugins.m4 (AC_PLUGINS): If plugins are enabled, add -ldl to
LIBS via AC_SEARCH_LIBS.
gdb/
* acinclude.m4 (GDB_AC_CHECK_BFD): Don't add -ldl.
* config.in: Regenerate.
sim/ppc/
* configure.ac: Invoke AC_PLUGINS.
* config.in: Regenerate.
and regen lots of configure files.
PR 17287
bfd/
* elflink.c (on_needed_list): Only consider libraries that have
been loaded.
ld/
* ld.texinfo (--as-needed): Clarify that references from libraries
must be from needed libraries.
ld/testsuite/
* ld-plugin/needed3.c: New file.
* ld-elf/shared.exp: Add needed3 test.
Presents .eh_frame input sections to the optimisation machinery in
elf-eh-frame.c in the order they are given by the linker script.
PR 16563
bfd/
* elflink.c (bfd_elf_discard_info): Process .eh_frame and .stab
in the order they are mapped to output sections.
ld/
* ldlang.c (map_head_is_link_order): Rename from
stripped_excluded_sections.
(lang_clear_os_map): New function, extracted from..
(strip_excluded_output_sections): ..here.
* ldlang.h (lang_clear_os_map): Declare.
* ldwrite.c (ldwrite): Call lang_clear_os_map.
* emultempl/sh64elf.em (sh64_elf_${EMULATION_NAME}_after_allocation):
Likewise.
This also makes --enable-plugins default to on for hosts that can
support plugins, so we have consistent lto toolchain support. The
ACX_LARGEFILE moves aren't strictly necessary, but are harmless and
will be necessary if plugin support is extended to more hosts via
libtool's dlopen support. I started down that path then decided it
was more work than I was interested in doing. (ACX_LARGEFILE invokes
AC_PLUGINS.)
config/
* plugins.m4: Test for dlfcn.h or windows.h here to set default
for --enable-plugins. Report error if someone tries to enable
plugins on a host we don't support.
bfd/
* configure.ac: Delete redundant plugin related checks.
* configure: Regenerate.
binutils/
* configure.ac: Move ACX_LARGEFILE after LT_INIT.
* config.in: Regenerate.
* configure: Regenerate.
gas/
* configure.ac: Move ACX_LARGEFILE after LT_INIT.
* config.in: Regenerate.
* configure: Regenerate.
gprof/
* configure.ac: Move ACX_LARGEFILE after LT_INIT.
* configure: Regenerate.
* gconfig.in: Regenerate.
ld/
* configure.ac: Move AC_PROG_CC and other macros earlier. Delete
plugin checks now done in config/plugins.m4.
* config.in: Regenerate.
* configure: Regenerate.
The main aim of this change was to have non_ir_ref set correctly on
new indirect symbols. I could have added a "copy" param to the "notice"
function, so that indirect symbols could be created in plugin_notice,
but it seemed cleaner to create indirect syms earlier and pass them
rather than "string" to "notice".
include/
* bfdlink.h (struct bfd_link_callbacks <notice>): Remove "string"
param, add "inh".
bfd/
* coff-aux.c (coff_m68k_aux_link_add_one_symbol): Only call "notice"
here when not calling the generic add_symbol function. Formatting.
Correct handling of indirect symbols. Update notice call.
* elflink.c (_bfd_elf_notice_as_needed): Update notice call.
* linker.c (_bfd_generic_link_add_one_symbol): Create indirect
symbols early. Update notice call. Add comments regarding weak
symbols vs. indirect.
ld/
* ldmain.c (notice): Update args.
* plugin.c (plugin_notice): Likewise. Follow warning sym link.
Handle new indirect symbol.
Override the default value of 0x0000 defined in
TEXT_START_ADDR to avoid linux executables to be mapped
at zero page.
ld/
* emulparams/elf32or1k_linux.sh (TEXT_START_ADDR): Increase from
0x0 to first page boundary at 0x2000.
is_inconsistent_linkonce_section makes an assumption that section name
that starts with ".gnu.linkonce.prop." has one more dot in its suffix.
However gas generates such section name by insertion of "prop." right
after ".gnu.linkonce." part of the name of the original section. So, for
section named ".gnu.linkonce.this_module" corresponding property section
name does not satisfy the assumption. Such section names are common in
linux modules. This bug was exposed by the patch "a35d5e8 Fix alignment
for the first section frag on xtensa", that makes gas produce property
section for each section that has ".align" directive in it.
Use suffix that immediately follows ".gnu.linkonce.prop." when there are
no more dots following it.
2014-07-10 Max Filippov <jcmvbkbc@gmail.com>
ld/
* emultempl/xtensaelf.em (is_inconsistent_linkonce_section):
correctly handle missing dot in section name after
".gnu.linkonce.prop.".
Increase the maximum page size to 64kB and align the TEXT_START_ADDR
to a 64kB boundary. This brings AArch32 in line with AArch64 and
improves compatability under certain conditions.
bfd/ChangeLog:
2014-07-08 Will Newton <will.newton@linaro.org>
* elf32-arm.c (ELF_MAXPAGESIZE): Increase the default
value to 64kB and remove custom setting for NaCl.
ld/ChangeLog:
2014-07-08 Will Newton <will.newton@linaro.org>
* emulparams/armelf_linux.sh (TEXT_START_ADDR): Increase
alignment to 64kB boundary.
ld/testsuite/ChangeLog:
2014-07-08 Will Newton <will.newton@linaro.org>
* ld-arm/arm-lib.ld: Increase MAXPAGESIZE value to
match bfd.
* ld-arm/cortex-a8-fix-bl-rel-plt.d: Update offsets to
take into account increased segment alignment.
* ld-arm/ifunc-gdesc.r: Likewise.
* ld-arm/tls-lib.d: Likewise.
This removes usage of the obsolete AC_INIT and AM_INIT_AUTOMAKE in all
binutils configure.in files. The BFD version is now in bfd/version.m4
rather than bfd/configure.in, which allows automake to automatically
track this dependency.
bfd/
* version.m4: New file.
* configure.in: Include version.m4.
(AC_INIT): Update.
* Makefile.am (RELEASE): Delete.
(bfdver.h): Depend on development.sh, use instead of RELEASE.
* Makefile.in: Regenerate.
* doc/Makefile.in: Regenerate.
opcodes/
* configure.in: Include bfd/version.m4.
(AC_INIT, AM_INIT_AUTOMAKE): Use modern form.
(BFD_VERSION): Delete.
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Remove bfd/configure.in.
* configure: Regenerate.
* Makefile.in: Regenerate.
binutils/
* configure.in: Include bfd/version.m4.
(AC_INIT, AM_INIT_AUTOMAKE): Use modern form.
(BFD_VERSION): Delete.
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Remove bfd/configure.in.
* configure: Regenerate.
* Makefile.in: Regenerate.
* doc/Makefile.in: Regenerate.
gas/
* configure.in: Include bfd/version.m4.
(AC_INIT, AM_INIT_AUTOMAKE): Use modern form.
(BFD_VERSION): Delete.
* configure.com: Get bfd version from bfd/version.m4.
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Remove bfd/configure.in.
* configure: Regenerate.
* Makefile.in: Regenerate.
* doc/Makefile.in: Regenerate.
gprof/
* configure.in: Include bfd/version.m4.
(AC_INIT, AM_INIT_AUTOMAKE): Use modern form.
(BFD_VERSION): Delete.
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Remove bfd/configure.in.
* configure: Regenerate.
* Makefile.in: Regenerate.
ld/
* configure.in: Include bfd/version.m4.
(AC_INIT, AM_INIT_AUTOMAKE): Use modern form.
(BFD_VERSION): Delete.
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Remove bfd/configure.in.
* configure: Regenerate.
* Makefile.in: Regenerate.
When a shared library appears within --start-group/--end-group ld may
only discover a need for loading the library on the second or
subsequent pass over archive libraries, as more objects are extracted.
ld/
PR 17068
* ldlang.c (load_symbols): Always check flags.reload.
(open_input_bfds): Always reload --as-needed shared libraries,
not just when rescanning.
* ldlang.h (struct lang_input_statement_flags): Update reload comment.
* plugin.c (plugin_should_reload): Assume shared library arg.
* plugin.h (plugin_should_reload): Update comment.
ld/testsuite
* ld-elf/pr17068.s: New.
* ld-elf/pr17068a.s: New.
* ld-elf/pr17068b.s: New.
* ld-elf/pr17068c.s: New.
* ld-elf/pr17068d.s: New.
* ld-elf/pr17068e.s: New.
* ld-elf/pr17068ez.s: New.
* ld-elf/elf.exp: Run new test.
The undefined function symbols (with non-zero value) on global entry
stubs are discarded by objdump when disassembling, so give objdump
another symbol to mark the stubs.
Also fixes a couple of bugs:
- entry_section was set to .opd for ELFv2, which meant a hard error
rather than a warning when _start wasn't defined.
- global entry stubs were not built if they were the only type of
stub in an executable.
bfd/
* elf64-ppc.c (ppc_stub_type): Add ppc_stub_global_entry.
(struct ppc_link_hash_table): Increase size of stub_count array.
(build_global_entry_stubs): Emit symbol on global entry stub.
(ppc64_elf_build_stubs): NULL check htab->brlt. Add global entry
stub stats.
ld/
* emultempl/ppc64elf.em (stub_added): Delete.
(gld${EMULATION_NAME}_finish): Call ppc64_elf_build_stubs even when
none of the usual stubs have been added. Only change entry_section
for ELFv1.
sys/param.h on recent versions of powerpc glibc ends up including
asm/elf.h via asm/sigcontex.h. asm/elf.h defines R_PPC_* and R_PPC64_*
macros, which clash with our include/elf/ppc.h and include/elf/ppc64.h.
It turns out that no current source uses LD_PATHMAX, so there is no
need for limits.h or sys/param.h, except for one occurrence of UINT_MAX.
I don't have a quarrel with limits.h, but it seems unnecessary just
for UINT_MAX.
* sysdep.h: Don't include limits.h and sys/param.h. Don't
include unistd.h twice.
(LD_PATHMAX): Don't define.
* ldlang.c (lang_common): Don't use UINT_MAX.
* archures.c: add avrtiny architecture for avr target.
* bfd-in2.h: Regenerate.
* cpu-avr.c (arch_info_struct): add avrtiny arch info.
* elf32-avr.c (elf_avr_howto_table): new relocation R_AVR_LDS_STS_16
added for 16 bit LDS/STS instruction of avrtiny arch.
(avr_reloc_map): reloc R_AVR_LDS_STS_16 is mapped to
BFD_RELOC_AVR_LDS_STS_16.
(bfd_elf_avr_final_write_processing): select machine number avrtiny arch.
(elf32_avr_object_p): set machine number for avrtiny arch.
* libbfd.h: Regenerate.
* reloc.c: Add documentation for BFD_RELOC_AVR_LDS_STS_16 reloc.
* config/tc-avr.c (mcu_types): Add avrtiny arch.
Add avrtiny arch devices attiny4, attiny5, attiny9, attiny10, attiny20
and attiny40.
(md_show_usage): Add avrtiny arch in usage message.
(avr_operand): validate and issue error for invalid register for avrtiny.
add new reloc exp for 16 bit lds/sts instruction.
(md_apply_fix): check 16 bit lds/sts operand for out of range and encode.
(md_assemble): check ISA for arch and issue diagnostic.
* include/elf/avr.h (E_AVR_MACH_AVRTINY): define avrtiny machine number.
(R_AVR_LDS_STS_16): define 16 bit lds/sts reloc number.
* include/opcode/avr.h (AVR_ISA_TINY): define avrtiny specific ISA.
(AVR_ISA_2xxxa): define ISA without LPM.
(AVR_ISA_AVRTINY): define avrtiny arch ISA.
Add doc for contraint used in 16 bit lds/sts.
Adjust ISA group for icall, ijmp, pop and push.
Add 16 bit lds/sts encoding and update 32 bit lds/sts constraints.
* opcodes/avr-dis.c (avr_operand): Handle constraint j for 16 bit lds/sts.
(print_insn_avr): do not select opcode if insn ISA is avrtiny and machine
is not avrtiny.
* Makefile.am (ALL_EMULATION_SOURCES): add avrtiny emulation source.
(eavrtiny.c): add rules for avrtiny emulation source.
* Makefile.in: Regenerate.
* configure.tgt: Add avrtiny to avr target emulations.
* scripttempl/avrtiny.sc: New file.
linker script template for avrtiny arch.
* emulparams/avrtiny.sh: New file.
emulation parameters for avrtiny arch.
inserted. This is for compatibility with other, non-GNU tools. Deterministic
binaries can still be created by using the new --no-insert-timestamp command line
option.
* emultempl/pe.em: Initialise insert_timestamp to true.
Add a --no-insert-timestamp command line option.
* emultempl/pep.em: Likewise.
* ld.texinfo: Document that --insert-timestamp is enabled by
default and that it now has an inverse command line option.
* NEWS: Mention the new behaviour.
* Makefile.am (ALL_EMULATION_SOURCES): Move ei386pep.c from
here...
(ALL_64_EMULATION_SOURCES): ... to here.
(ALL_EMUL_EXTRA_OFILES): Move pep-dll.o from here...
(ALL_64_EMUL_EXTRA_OFILES): New. ... to here.
* configure.in (EMUL_EXTRA_OFILES): Include
ALL_64_EMUL_EXTRA_OFILES when making a 64-bit enabled build.
* Makefile.in: Regenerate.
* configure: Regenerate.
and fix more nds32 dependencies.
* scripttempl/elf.sc: Edit out __rela_iplt symbol assignments from
.rel sections, and __rel_iplt from .rela sections.
* scripttempl/nds32elf.sc: Likewise.
* Makefile.am (ends32*.c): Depend on nds32elf.sc.
* Makefile.in: Regenerate.
The goal of this patch is to provide an easy way to make
--disable-werror the default when building binutils, or the parts
of binutils that need to get built when building GDB. In development
mode, we want to continue making -Werror the default with GCC.
But, when making releases, I think we want to make it as easy as
possible for regular users to successfully build from sources.
GDB already has this kind of feature to turn -Werror as well as
the use of the libmcheck library. As GDB Release Manager, I take
advantage of it to turn those off after having cut the branch.
I'd like to be able to do the same for the binutils bits. And
perhaps Tristan will want to do the same for his releases too
(not sure, binutils builders might be a little savvier than GDB
builders).
This patch introduces a new file, called development.sh, which
just sets a variable called $development. In our development branches
(Eg. "master"), it's set to true. But setting it to false would allow
us to change the default behavior of various development-related
features to be turned off; in this case, it turns off the use of
-Werror by default (use --enable-werror to turn it back on).
bfd/ChangeLog:
* development.sh: New file.
* warning.m4 (AM_BINUTILS_WARNINGS): Source bfd/development.sh.
Make -Werror the default with GCC only if DEVELOPMENT is true.
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Add
$(srcdir)/development.sh.
* Makefile.in, configure: Regenerate.
binutils/ChangeLog:
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Add dependency on
bfd's development.sh.
* Makefile.in, configure: Regenerate.
gas/ChangeLog:
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Add dependency on
bfd's development.sh.
* Makefile.in, configure: Regenerate.
gold/ChangeLog:
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): New.
* Makefile.in, configure: Regenerate.
gprof/ChangeLog:
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Add dependency on
bfd's development.sh.
* Makefile.in, configure: Regenerate.
ld/ChangeLog:
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Add dependency on
bfd's development.sh.
* Makefile.in, configure: Regenerate.
opcodes/ChangeLog:
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Add dependency on
bfd's development.sh.
* Makefile.in, configure: Regenerate.
gdb/ChangeLog:
* development.sh: Delete.
* Makefile.in (config.status): Adjust dependency on development.sh.
* configure.ac: Adjust development.sh source call.
* configure: Regenerate.
gdb/gdbserver/ChangeLog:
* configure.ac: Adjust development.sh source call.
* Makefile.in (config.status): Adjust dependency on development.sh.
* configure: Regenerate.
Tested on x86_64-linux by building two ways: One with DEVELOPMENT
set to true, and one with DEVELOPMENT set to false. In the first
case, I could see the use of -Werror, while it disappeared in
the second case.
bfin, msp, and score all used an extra parameter to genscripts.sh
to select a "customizer_script" different from the standard one
named from the emulation. This patch renames the scripts to avoid
the need, tidying them in the process.
* emulparams/elf32bfin.sh: Rename from bfin.sh.
* emulparams/elf32bfinfd.sh: Update to suit.
* emulparams/: Delete.
* emulparams/msp430.sh: Rename from msp430all.sh. Remove
MSP430_NAME and msp430X vars.
* emulparams/msp430X.sh: New.
* emulparams/: Delete.
* emulparams/score3_elf.sh: Rename from scoreelf.sh. Remove
SCORE_NAME and score7_elf ARCH setting.
* emulparams/score7_elf.sh: New.
* Makefile.am (eelf32bfin.c, eelf32bfinfd.c): Update dependencies.
(emsp430.c, emsp430X.c, escore3_elf.c, escore7_elf.c): Likewise.
* Makefile.in: Regenerate.
* genscripts.sh: Delete customizer_script param.
* Makefile.am: Change all rules with ${GENSCRIPTS}
invocations to be just dependencies.
($(ALL_EMULATION_SOURCES) $(ALL_64_EMULATION_SOURCES))
(run-genscripts): New rules.
* Makefile.in: Regenerate.
1) _SDA_BASE_ and _SDA2_BASE_ and defined automatically, in a similar
manner to the way _GLOBAL_OFFSET_TABLE_ is handled. It's a little
more complicated to remove the symbols because _SDA_BASE_ needs to
be there if either .sdata or .sbss is present, and similarly for
_SDA2_BASE.
2) The linker created .sdata and .sdata2 sections used for
R_PPC_EMB_SDAI16 and R_PPC_EMB_SDA2I16 pointers are created early.
Nowadays we strip unneeded sections from the output, so it isn't
necessary to delay creating the sections.
3) The output section for targets of various SDA relocs is now checked
as per the ABI(s). We previously allowed .sdata.foo and similar,
most likely because at some stage we were checking input sections.
Also, the patch fixes a long-standing bug in size_input_sections
that affects the values of symbols defined in stripped input
sections.
PR 16952
bfd/
* elf32-ppc.c (ppc_elf_create_linker_section): Move earlier.
Remove redundant setting of htab->elf.dynobj. Don't align.
Define .sdata symbols using _bfd_elf_define_linkage_sym.
(ppc_elf_create_glink): Call ppc_elf_create_linker_section.
(create_sdata_sym): Delete.
(elf_allocate_pointer_linker_section): Rename from
elf_create_pointer_linker_section. Align section.
(ppc_elf_check_relocs): Don't call ppc_elf_creat_linker_section
directly here, or create_sdata_sym. Set ref_regular on _SDA_BASE_
and _SDA2_BASE_.
(ppc_elf_size_dynamic_sections): Remove ATTRIBUTE_UNUSED on param.
Remove unnecessary tests on _SDA_BASE_ sym.
(maybe_strip_sdasym, ppc_elf_maybe_strip_sdata_syms): New functions.
(ppc_elf_relocate_section): Tighten SDA reloc symbol section checks.
* elf32-ppc.h (ppc_elf_set_sdata_syms): Delete.
(ppc_elf_maybe_strip_sdata_syms): Declare.
ld/
* emulparams/elf32ppccommon.sh (_SDA_BASE_, _SDA2_BASE_): Delete.
* emultempl/ppc32elf.em (ppc_before_allocation): Call
ppc_elf_maybe_strip_sdata_syms.
* ldlang.c (size_input_section): Correct output_offset value
for excluded input sections.
ld:
* ldlang.c (lang_finish): Don't call bfd_link_hash_table_free here.
(output_bfd_hash_table_free_fn): New variable.
(open_output): Save the _bfd_link_hash_table_free function for the
output_bfd into output_bfd_hash_table_free_fn.
* ldmain.c (ld_cleanup): If set, call output_bfd_hash_table_free_fn
on link_info.hash.
* ldlang.h (output_bfd_hash_table_free_fn): Declare.
ld/testsuite:
* ld-mmix/wrap1.d, ld-mmix/wrap1a.s, ld-mmix/wrap1b.s,
ld-mmix/wrap1c.s, ld-mmix/wrap2.d, ld-mmix/wrap3.d,
ld-mmix/wrap3a.s, ld-mmix/wrap3b.s, ld-mmix/wrap4.d: New
tests.
ld/
* emultempl/alphaelf.em (alpha_after_parse): Enable 2 relax passes.
bfd/
* elf64-alpha.c (elf64_alpha_size_got_sections): New may_merge
parameter; honor it and disable got merging when false.
(elf64_alpha_relax_got_load): Do not relax to GPREL relocs during
the first pass of relaxation.
(elf64_alpha_relax_with_lituse): Likewise. Move relaxed relocs to
the end of the LITERAL+LITUSE chain.
(elf64_alpha_relax_section): Only process LITERAL relocs during the
second pass of relaxation.
I got the ppc476 workaround wrong. bctr (and bctrl) as the last
instruction in a page can hit the icache bug if the preceding mtctr
insn is close by, and the destination is in the first few instructions
on the next page. This scenario can occur with code generated by gcc
to implement switch statements, or in code generated to call by
function pointer.
To prevent the bctr problem it is also necessary to remove other
instructions that otherwise would be safe.
bfd/
* elf32-ppc.c (ppc_elf_relocate_section): Remove bctr from list
of safe ppc476 insns at end of page. Also remove non-branch insns.
Expand comments.
ld/
* emultempl/ppc32elf.em (no_zero_padding, ppc_finish): New functions.
(LDEMUL_FINISH): Define.
the Cygwin and MinGW targets. The manifest is now going to be handled by gcc.
* scripttempl/pe.sc (R_RSRC): Remove default manifest.
* scripttempl/pep.sc (R_RSRC): Remove default manifest.
The linker script documentation does not mention the optional comma
that may follow an output section command or an overlay command.
In some cases, where a fill expression is used, and the next
output section command begins with an operator (e.g., "/DISCARD/"),
the comma may be required to separate the two commands.
Currently, GNU ld doesn't require the comma, but gold does.
ld/
PR gold/16804
* ld.texinfo: Document optional comma following output section
command and overlay command.
When aligning input sections, we are supposed to take the fill pattern
from a FILL statement, if there is one in the output section statement.
ld/
* ldlang.c (lang_size_sections_1 <lang_input_section_enum>): Use
current "fill", not "output_section_statement->fill".
ld/testsuite/
* ld-scripts/fill.d, * ld-scripts/fill.t, * ld-scripts/fill_0.s,
* ld-scripts/fill_1.s, * ld-scripts/fill_2.s: New test.
* ld-scripts/data.exp: Run it.
would not honour the [no]execstack. Fully linked binaries use a special segment to indicate the
state of the stack, but relocatable object files only have sections, and .note.GNU-stack was not
being created.
* emultempl/elf32.em (_after_open): Create a .note.GNU-stack
section when performing a relocatable link with -z [no]execstack
specified.
This patch scans all input files for symbol reference warning if the
symbol reference doesn't exist in the current input file.
ld/
PR ld/16756
* ldmain.c (symbol_warning): New function.
(warning_callback): Use it. Scan all input files for a reference
to SYMBOL.
ld/testsuite/
PR ld/16756
* ld-plugin/lto.exp: Expect filename and line number for PR
ld/12760 test.
should work for all types of input .rsrc section.
* peXXigen.c (rsrc_process_section): Add code to scan input
sections and record their lengths. Use these lengths to find the
start of each merged .rsrc section.
* scripttempl/pe.sc (R_RSRC): Fix default-manifest exclusion.
(.rsrc): Add SUBALIGN(4). Remove SORT.
* scripttempl/pep.sc: Likewise.
For libraries without a soname, -l:libfoo.so set DT_NEEDED to the search
dir plus filename, while gold and -lfoo just use the filename. This
patch fixes the inconsistency.
* ldlang.h (full_name_provided): New input flag.
* ldlang.c (new_afile): Don't use lang_input_file_is_search_file_enum
for -l:namespec. Instead use lang_input_file_is_l_enum with
full_name_provided flag.
* ldlfile.c (ldfile_open_file_search): Don't complete lib name if
full_name_provided flag is set.
* emultempl/elf32.em (gld${EMULATION_NAME}_open_dynamic_archive):
Handle full_name_provided libraries. Tidy EXTRA_SHLIB_EXTENSION
support. Set DT_NEEDED for -l:namespec as namespec.
* emultempl/aix.em (ppc_after_open_output): Handle full_name_provided.
* emultempl/linux.em (gld${EMULATION_NAME}_open_dynamic_archive):
Don't handle full_name_provided libraries.
* emultempl/pe.em (gld${EMULATION_NAME}_open_dynamic_archive): Ditto.
* emultempl/pep.em (gld${EMULATION_NAME}_open_dynamic_archive): Ditto.
* emultempl/vms.em (gld${EMULATION_NAME}_open_dynamic_archive): Ditto.
built when constructing a toolchain configued from all targets.
* Makefile.am (ALL_EMUL_EXTRA_BINARIES): Remove default-manifest.
* Makefile.in: Regenerate.
Enabled via the use of a new linker command line option: --long-plt.
* bfd-in.h: Add export of bfd_elf32_arm_use_long_plt.
* bfd-in2.h: Regenerate.
* elf32-arm.c (elf32_arm_plt_entry_long): New array.
(elf32_arm_link_hash_table_create): Set plt_entry_size to 16 if
using long PLT entries.
(bfd_elf32_arm_use_long_plt): New function.
(elf32_arm_populate_plt_entry): Add support for long PLT entries.
* emultempl/armelf.em (OPTION_LONG_PLT): Define.
(PARSE_AND_LIST_LONGOPTS): Add long-plt.
(PARSE_AND_LIST_OPTIONS): Likewise.
(PARSE_AND_LIST_ARGS_CASES): Handle long-plt.
* ld.texinfo: Document --long-plt.
* ld-arm/long-plt-format.s: New test case.
* ld-arm/long-plt-format.d: Expected disassembly.
* ld-arm/arm-elf.exp: Run the new test.
The manifest is necessary in order for the linked binaries to be executed in a Windows 8 environment.
The manifest is added using a linker script so that this feature will be compiler-neutral. The resource
merging code in the linker means that if an application provides its own manifest then the default
manifest will be ignored.
* configure.in (all_emul_extra_binaries): New variable. Populated
by invoking configure.tgt.
(EMUL_EXTRA_BINARIES): New substitution.
* configure: Regenerate.
* configure.tgt (target_extra_binaries): New variable. Set to
default-manifest.o for Cygwin and MinGW targets.
* Makefile.am (EMUL_EXTRA_BINARIES): New variable. Initialised
by the configure script.
(ALL_EMUL_EXTRA_BINARIES): New variable.
(default-manifest.o): New rule to build the default manifest.
(ld_new_DEPENDENCIES): Add EMUL_EXTRA_BINARIES.
(install-data-local): Add EMUL_EXTRA_BINARIES.
* Makefile.in: Regenerate.
* ld.texinfo: Document default manifest support.
* emulparams/i386pe.sh (DEFAULT_MANIFEST): Define.
* emulparams/i386pep.sh (DEFAULT_MANIFEST): Define.
* emultempl/default-manifest.rc: New file.
* scripttempl/pe.sc (R_RSRC): Include DEFAULT_MANIFEST, if defined.
* scripttempl/pep.sc (R_RSRC): Likewise.
* ld-pe/longsecn-1.d: Allow for extra sections.
* ld-pe/longsecn-2.d: Likewise.
* ld-pe/longsecn.d: Likewise.
* ld-pe/secrel.d: Likewise.
5446cbdf82 broke powerpc-lynxos,
powerpc-netware, powerpc-windiss and powerpc-vxworks.
bfd/
* elf32-ppc.c (ppc_elf_link_hash_table_create): Provide default
params for targets that don't use ppc32elf.em.
ld/
* emulparams/elf32ppcvxworks.sh: Source plt_unwind.sh and
use ppc32elf.em.
* emultempl/ppc32elf.em (ppc_after_open): Don't compile for
vxworks.
(LDEMUL_AFTER_OPEN): Don't set for vxworks.
(PARSE_AND_LIST_LONGOPTS, PARSE_AND_LIST_OPTIONS): Exclude
-secure-plt, -bss-plt and -sdata-got when vxworks.
Intel MPX introduces 4 bound registers, which will be used for parameter
passing in x86-64. Bound registers are cleared by branch instructions.
Branch instructions with BND prefix will keep bound register contents.
This leads to 2 requirements to 64-bit MPX run-time:
1. Dynamic linker (ld.so) should save and restore bound registers during
symbol lookup.
2. Change the current 16-byte PLT0:
ff 35 08 00 00 00 pushq GOT+8(%rip)
ff 25 00 10 00 jmpq *GOT+16(%rip)
0f 1f 40 00 nopl 0x0(%rax)
and 16-byte PLT1:
ff 25 00 00 00 00 jmpq *name@GOTPCREL(%rip)
68 00 00 00 00 pushq $index
e9 00 00 00 00 jmpq PLT0
which clear bound registers, to preserve bound registers.
We use 2 new relocations:
to mark branch instructions with BND prefix.
When linker sees any R_X86_64_PC32_BND or R_X86_64_PLT32_BND relocations,
it switches to a different PLT0:
ff 35 08 00 00 00 pushq GOT+8(%rip)
f2 ff 25 00 10 00 bnd jmpq *GOT+16(%rip)
0f 1f 00 nopl (%rax)
to preserve bound registers for symbol lookup and it also creates an
external PLT section, .pl.bnd. Linker will create a BND PLT1 entry
in .plt:
68 00 00 00 00 pushq $index
f2 e9 00 00 00 00 bnd jmpq PLT0
0f 1f 44 00 00 nopl 0(%rax,%rax,1)
and a 8-byte BND PLT entry in .plt.bnd:
f2 ff 25 00 00 00 00 bnd jmpq *name@GOTPCREL(%rip)
90 nop
Otherwise, linker will create a legacy PLT1 entry in .plt:
68 00 00 00 00 pushq $index
e9 00 00 00 00 jmpq PLT0
66 0f 1f 44 00 00 nopw 0(%rax,%rax,1)
and a 8-byte legacy PLT in .plt.bnd:
ff 25 00 00 00 00 jmpq *name@GOTPCREL(%rip)
66 90 xchg %ax,%ax
The initial value of the GOT entry for "name" will be set to the the
"pushq" instruction in the corresponding entry in .plt. Linker will
resolve reference of symbol "name" to the entry in the second PLT,
.plt.bnd.
Prelink stores the offset of pushq of PLT1 (plt_base + 0x10) in GOT[1]
and GOT[1] is stored in GOT[3]. We can undo prelink in GOT by computing
the corresponding the pushq offset with
GOT[1] + (GOT offset - &GOT[3]) * 2
Since for each entry in .plt except for PLT0 we create a 8-byte entry in
.plt.bnd, there is extra 8-byte per PLT symbol.
We also investigated the 16-byte entry for .plt.bnd. We compared the
8-byte entry vs the the 16-byte entry for .plt.bnd on Sandy Bridge.
There are no performance differences in SPEC CPU 2000/2006 as well as
micro benchmarks.
Pros:
No change to undo prelink in dynamic linker.
Only 8-byte memory overhead for each PLT symbol.
Cons:
Extra .plt.bnd section is needed.
Extra 8 byte for legacy branches to PLT.
GDB is unware of the new layout of .plt and .plt.bnd.
bfd/
* elf64-x86-64.c (elf_x86_64_bnd_plt0_entry): New.
(elf_x86_64_legacy_plt_entry): Likewise.
(elf_x86_64_bnd_plt_entry): Likewise.
(elf_x86_64_legacy_plt2_entry): Likewise.
(elf_x86_64_bnd_plt2_entry): Likewise.
(elf_x86_64_bnd_arch_bed): Likewise.
(elf_x86_64_link_hash_entry): Add has_bnd_reloc and plt_bnd.
(elf_x86_64_link_hash_table): Add plt_bnd.
(elf_x86_64_link_hash_newfunc): Initialize has_bnd_reloc and
plt_bnd.
(elf_x86_64_copy_indirect_symbol): Also copy has_bnd_reloc.
(elf_x86_64_check_relocs): Create the second PLT for Intel MPX
in 64-bit mode.
(elf_x86_64_allocate_dynrelocs): Handle the second PLT for IFUNC
symbols. Resolve call to the second PLT if it is created.
(elf_x86_64_size_dynamic_sections): Keep the second PLT section.
(elf_x86_64_relocate_section): Resolve PLT references to the
second PLT if it is created.
(elf_x86_64_finish_dynamic_symbol): Use BND PLT0 and fill the
second PLT entry for BND relocation.
(elf_x86_64_finish_dynamic_sections): Use MPX backend data if
the second PLT is created.
(elf_x86_64_get_synthetic_symtab): New.
(bfd_elf64_get_synthetic_symtab): Likewise. Undefine for NaCl.
ld/
* emulparams/elf_x86_64.sh (TINY_READONLY_SECTION): New.
ld/testsuite/
* ld-x86-64/mpx.exp: Run bnd-ifunc-1 and bnd-plt-1.
* ld-x86-64/bnd-ifunc-1.d: New file.
* ld-x86-64/bnd-ifunc-1.s: Likewise.
* ld-x86-64/bnd-plt-1.d: Likewise.
This patch allows the user to override powerpc64-ld's default for
providing linker generated register save and restore functions as used
by gcc -Os code. Normally these are not provided by ld -r, so Linux
kernel modules have needed to include their own copies.
bfd/
* elf64-ppc.h (struct ppc64_elf_params): Add save_restore_funcs.
* elf64-ppc.c (ppc64_elf_func_desc_adjust): Use it to control
provision of out-of-line register save/restore routines.
ld/
* emultempl/ppc64elf.em (params): Init new field.
(ppc_create_output_section_statements): Set params.save_restore_funcs
default.
(PARSE_AND_LIST_*): Add support for --save-restore-funcs and
--no-save-restore-funcs.
Moves assorted variables used to communicate between ld and bfd into
a struct, hooks it into the bfd link_hash_table early, and removes
all other places where such variables were passed piecemeal.
bfd/
* elf64-ppc.h (struct ppc64_elf_params): Define.
(ppc64_elf_init_stub_bfd, ppc64_elf_edit_opd, ppc64_elf_tls_setup,
ppc64_elf_setup_section_lists, ppc64_elf_size_stubs,
ppc64_elf_build_stubs): Update prototype.
* elf64-ppp.c (struct ppc_link_hash_table): Add params, delete other
fields now in params. Adjust code throughout file.
(ppc64_elf_init_stub_bfd): Delete "abfd" parameter, add "params".
Save params pointer in htab.
(ppc64_elf_edit_opd, ppc64_elf_tls_setup,
ppc64_elf_setup_section_lists, ppc64_elf_size_stubs,
ppc64_elf_build_stubs): Remove parameters now in "params".
ld/
* emultemps/ppc64elf.em (params): New static struct replacing
various other static vars. Adjust code throughout file.
This fixes the glaring error that the ppc476 workaround wasn't
actually enabled for ld -r, and adjusts relocations to match moved
code.
bfd/
* elf32-ppc.c (ppc_elf_relocate_section): Move relocs on insns
patched for ppc476 workaround. Reapply branch taken/not taken
relocs.
ld/
* emultempl/ppc32elf.em (ppc_after_open_output): Really enable
ppc476 workaround for ld -r.
The Linux kernel builds modules using ld -r. These might need the
ppc476 workaround, so enable it for ld -r if sections have sufficient
alignment to tell location within a page.
bfd/
* elf32-ppc.c (ppc_elf_relax_section): Enable ppc476 workaround
for ld -r, when code sections are sufficiently aligned.
* elf32-ppc.h (struct ppc_elf_params): Delete pagesize. Add
pagesize_p2.
ld/
* emultempl/ppc32elf.em (pagesize): New static var.
(ppc_after_open_output): Set params.pagesize_p2 from pagesize.
(PARSE_AND_LIST_ARGS_CASES): Adjust to use pagesize.