binutils/
* testsuite/lib/binutils-common.exp (supports_gnu_unique): New
procedure.
* testsuite/binutils-all/objcopy.exp: Use `supports_gnu_unique'
with the `strip-10' test.
ld/
* testsuite/ld-unique/unique.exp: Use `is_elf_format' and
`supports_gnu_unique' to qualify testing.
Yao Qi [Tue, 17 May 2016 07:24:26 +0000 (08:24 +0100)]
Use unsuspend_all_lwps
This patch is to replace find_inferior (&all_threads, unsuspend_one_lwp, NULL)
with unsuspend_all_lwps (NULL), which is shorter. They are equivalent
to each other.
gdb/gdbserver:
2016-05-17 Yao Qi <yao.qi@linaro.org>
* linux-low.c (linux_stabilize_threads): Call unsuspend_all_lwps
instead of find_inferior.
Yao Qi [Mon, 16 May 2016 16:32:43 +0000 (17:32 +0100)]
Match shell_prompt # in batch-preserve-term-settings.exp
batch-preserve-term-settings.exp fails if the shell prompt isn't $. It
is # in our testing env. In fact, the shell prompt can be anything.
The perfect solution would be "set_board_info shell_prompt" in the
host board file, and use board_info shell_prompt in
batch-preserve-term-settings.exp. This is a little bit overkill to
me, and we still need to figure out the different prompts on different
shells. I also tried to start shell with the prompt preset, but there is
not unique way to set shell prompt in different shells, so I give up.
It is reasonably simple to match either $ or # for the shell prompt, and
we can easily extend it to match other char, like >.
V850/BFD: Call `_bfd_elf_copy_private_bfd_data' again
Correct a regression introduced with commit 685080f21003 ("Adds support
for generating notes in V850 binaries.") which replaced rather than
extending the call to `_bfd_elf_copy_private_bfd_data' with
`v850_elf_copy_private_bfd_data'. Consequently ELFOSABI_GNU marking is
not propagated to output by `objcopy' from objects containing
STB_GNU_UNIQUE symbols.
bfd/
* elf32-v850.c (v850_elf_copy_notes): New function, factored out
from...
(v850_elf_copy_private_bfd_data): ... here. Call the new
function and `_bfd_elf_copy_private_bfd_data'.
binutils/
* testsuite/binutils-all/objcopy.exp: Don't skip the `strip-10'
test for the V850.
Trevor Saunders [Sat, 14 May 2016 04:34:05 +0000 (00:34 -0400)]
stop defining linkrelax in multiple places
Defining linkrelax to have different values in as.c and tc-msp430.c /
tc-mn10300.c is at least rather tricky, and seems fragile, when we can just set
it in md_begin instead.
Trevor Saunders [Sat, 14 May 2016 19:27:26 +0000 (15:27 -0400)]
m68hc11: make some vars type bfd_reloc_code_real_type
These variables only hold values from the bfd_reloc_code_real_type enum, and
are passed to functions that expect the argument to be of type
bfd_reloc_code_real_type, so it seems to make sense that there type is
bfd_reloc_code_real_type rather than int.
* config/tc-m68hc11.c (fixup8): Change variables type from int to
bfd_reloc_code_real_type where appropriate.
(fixup16): Likewise.
(fixup8_xg): Likewise.
Alan Modra [Fri, 13 May 2016 02:18:56 +0000 (11:48 +0930)]
Set dynamic tag VMA and size from dynamic section when possible
Rather than searching the output for a specific named section, it's
better, where possible, to use a linker created dynamic section to set
a dynamic tag's value. That way ld doesn't depend on the output
section name, making it possibile to map dynamic sections differently.
bfd/
* elf-m10300.c (_bfd_mn10300_elf_finish_dynamic_sections): Use
linker dynamic sections in calculating size and address of
* dynamic tags rather than using output sections. Remove asserts.
* elf32-arm.c (elf32_arm_finish_dynamic_sections): Likewise.
* elf32-cr16.c (_bfd_cr16_elf_finish_dynamic_sections): Likewise.
* elf32-cris.c (elf_cris_finish_dynamic_sections): Likewise.
* elf32-i370.c (i370_elf_finish_dynamic_sections): Likewise.
* elf32-lm32.c (lm32_elf_finish_dynamic_sections): Likewise.
* elf32-m32r.c (m32r_elf_finish_dynamic_sections): Likewise.
* elf32-m68k.c (elf_m68k_finish_dynamic_sections): Likewise.
* elf32-metag.c (elf_metag_finish_dynamic_sections): Likewise.
* elf32-microblaze.c (microblaze_elf_finish_dynamic_sections): Likewise.
* elf32-nds32.c (nds32_elf_finish_dynamic_sections): Likewise.
* elf32-nios2.c (nios2_elf32_finish_dynamic_sections): Likewise.
* elf32-or1k.c (or1k_elf_finish_dynamic_sections): Likewise.
* elf32-s390.c (elf_s390_finish_dynamic_sections): Likewise.
* elf32-score.c (s3_bfd_score_elf_finish_dynamic_sections): Likewise.
* elf32-score7.c (s7_bfd_score_elf_finish_dynamic_sections): Likewise.
* elf32-vax.c (elf_vax_finish_dynamic_sections): Likewise.
* elf32-xtensa.c (elf_xtensa_finish_dynamic_sections): Likewise.
* elf64-alpha.c (elf64_alpha_finish_dynamic_sections): Likewise.
* elf64-s390.c (elf_s390_finish_dynamic_sections): Likewise.
* elf64-sh64.c (sh64_elf64_finish_dynamic_sections): Likewise.
* elflink.c (bfd_elf_final_link): Likewise.
* elfxx-mips.c (_bfd_mips_elf_finish_dynamic_sections): Likewise.
* elfxx-sparc.c (sparc_finish_dyn): Likewise. Adjust error message.
* elf32-arc.c (GET_SYMBOL_OR_SECTION): Remove ASSERT arg and
don't set doit. Look up dynobj section.
(elf_arc_finish_dynamic_sections): Adjust GET_SYMBOL_OR_SECTION
invocation and dynamic tag vma calculation. Don't test
boolean var == TRUE.
* elfnn-aarch64.c (elfNN_aarch64_finish_dynamic_sections): Fix
DT_JMPREL calc.
ld/
* testsuite/ld-arm/arm-elf.exp: Adjust for arm-no-rel-plt now passing.
Use different output file name for static app without .rel.plt.
* testsuite/ld-arm/arm-no-rel-plt.ld: Align .rel.dyn and .rela.dyn.
* testsuite/ld-arm/arm-no-rel-plt.out: Delete.
* testsuite/ld-arm/arm-no-rel-plt.r: New.
* testsuite/ld-arm/arm-static-app.d: Don't check file name.
* testsuite/ld-arm/arm-static-app.r: Likewise.
Alan Modra [Fri, 13 May 2016 02:04:56 +0000 (11:34 +0930)]
ld -z combreloc elf_link_sort_relocs
A linker script may put .rela.plt in with other .rela.dyn relocs.
The change to elf_reloc_type_class puts any PLT relocs last. This
patch makes the input section layout better match the sorted relocs.
* elflink.c (elf_link_sort_relocs): Wrap overlong lines. Fix
octets_per_byte. Put dynamic .rela.plt last in link orders.
Assign output_offset for reloc sections rather than writing
sorted relocs from block corresponding to output_offset.
H.J. Lu [Thu, 12 May 2016 23:50:34 +0000 (16:50 -0700)]
Handle symbols defined/referenced only within IR
The plugin is called to claim symbols in an archive element from
plugin_object_p. But those symbols aren't needed to create output.
They are defined and referenced only within IR. get_symbols should
return resolution based on IR symbol kinds.
PR ld/20070
* Makefile.am (noinst_LTLIBRARIES): Add libldtestplug4.la.
(libldtestplug4_la_SOURCES): New.
(libldtestplug4_la_CFLAGS): Likewise.
(libldtestplug4_la_LDFLAGS): Likewise.
* Makefile.in: Regenerated.
* plugin.c (get_symbols): Return resolution based on IR symbol
kinds for symbols defined/referenced only within IR.
* testplug4.c: New file.
* ld/testsuite/ld-plugin/pr20070.d: Likewise.
* ld/testsuite/ld-plugin/pr20070a.c: Likewise.
* ld/testsuite/ld-plugin/pr20070b.c: Likewise.
* testsuite/ld-plugin/plugin.exp (plugin4_name): New.
(plugin4_path): Likewise.
Add a test for ld/20070.
Doug Evans [Thu, 12 May 2016 16:20:57 +0000 (09:20 -0700)]
PR symtab/19999 gdb unable to resolve vars with fission+PIE
gdb/ChangeLog:
* dwarf2loc.c (dwarf2_find_location_expression): For DWO files still
add base_offset.
gdb/testsuite/ChangeLog:
* lib/dwarf.exp (build_executable_from_fission_assembler): Pass
$options when building executable.
* gdb.dwarf2/fission-loclists-pie.c: New file.
* gdb.dwarf2/fission-loclists-pie.exp: New file.
Alan Modra [Thu, 12 May 2016 14:29:26 +0000 (23:59 +0930)]
ld -z combreloc reloc sorting
PLT relocs don't appear in .rela.dyn, at least not when using
normal linker scripts. However, if they do, then they ought to be
placed last rather than in the middle of other relocs.
* elf-bfd.h (elf_reloc_type_class): Put reloc_class_plt last.
Matthew Fortune [Wed, 26 Nov 2014 11:15:01 +0000 (11:15 +0000)]
Add MIPS32 DSPr3 support.
bfd/
* elfxx-mips.c (print_mips_ases): Add DSPR3.
binutils/
* readelf.c (print_mips_ases): Add DSPR3.
gas/
* config/tc-mips.c (options): Add OPTION_DSPR3 and
OPTION_NO_DSPR3.
(md_longopts): Likewise.
(md_show_usage): Add help for -mdspr3 and -mno-dspr3.
(mips_ases): Define availability for DSPr3.
(mips_ase_groups): Add ASE_DSPR3 to the DSP group.
(mips_convert_ase_flags): Map ASE_DSPR3 to AFL_ASE_DSPR3.
* doc/as.texinfo: Document -mdspr3, -mno-dspr3. Fix -mdspr2
formatting.
* doc/c-mips.texi: Document -mdspr3, -mno-dspr3, .set dspr3 and
.set nodspr3. Fix -mdspr2 formatting.
* testsuite/gas/mips/mips32-dspr3.d: New file.
* testsuite/gas/mips/mips32-dspr3.s: Likewise.
* testsuite/gas/mips/mips.exp: Run mips32-dspr3 test.
include/
* elf/mips.h (AFL_ASE_DSPR3): New macro.
(AFL_ASE_MASK): Update to include AFL_ASE_DSPR3.
* opcode/mips.h (ASE_DSPR3): New macro.
opcodes/
* mips-dis.c (mips_arch_choices): Add ASE_DSPR3 to mips32r6 and
mips64r6.
* mips-opc.c (D34): New macro.
(mips_builtin_opcodes): Define bposge32c for DSPr3.
Nick Clifton [Wed, 11 May 2016 11:51:04 +0000 (12:51 +0100)]
Ensure that padding in the constant pool uses constant values.
PR target/20068
* config/tc-arm.c (add_to_lit_pool): Ensure that the padding added
to the pool uses O_constant.
* testsuite/gas/arm/pr20068.s: New test.
* testsuite/gas/arm/pr20068.d: Test driver.
2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com>
bfd/
* elf32-arm.c (enum elf32_arm_stub_type): New max_stub_type
enumerator.
(arm_stub_sym_claimed): New function.
(elf32_arm_create_stub): Use veneered symbol name and section if
veneer needs to claim its symbol, and keep logic unchanged otherwise.
(arm_stub_claim_sym): New function.
(arm_map_one_stub): Call arm_stub_claim_sym if veneer needs to claim
veneered symbol, otherwise create local symbol as before.
2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com>
bfd/
* elf32-arm.c (elf32_arm_size_stubs): Use new macros
ARM_GET_SYM_BRANCH_TYPE and ARM_SET_SYM_BRANCH_TYPE to respectively get
and set branch type of a symbol.
(bfd_elf32_arm_process_before_allocation): Likewise.
(elf32_arm_relocate_section): Likewise and fix identation along the
way.
(allocate_dynrelocs_for_symbol): Likewise.
(elf32_arm_finish_dynamic_symbol): Likewise.
(elf32_arm_swap_symbol_in): Likewise.
(elf32_arm_swap_symbol_out): Likewise.
gas/
* config/tc-arm.c (arm_adjust_symtab): Use ARM_SET_SYM_BRANCH_TYPE to
set branch type of a symbol.
gdb/
* arm-tdep.c (arm_elf_make_msymbol_special): Use
ARM_GET_SYM_BRANCH_TYPE to get branch type of a symbol.
include/
* arm.h (enum arm_st_branch_type): Add new ST_BRANCH_ENUM_SIZE
enumerator.
(NUM_ENUM_ARM_ST_BRANCH_TYPE_BITS): New macro.
(ENUM_ARM_ST_BRANCH_TYPE_BITMASK): Likewise.
(ARM_SYM_BRANCH_TYPE): Replace by ...
(ARM_GET_SYM_BRANCH_TYPE): This and ...
(ARM_SET_SYM_BRANCH_TYPE): This in two versions depending on whether
BFD_ASSERT is defined or not.
ld/
* emultempl/armelf.em (gld${EMULATION_NAME}_finish): Use
ARM_GET_SYM_BRANCH_TYPE to get branch type of a symbol.
opcodes/
* arm-dis.c (get_sym_code_type): Use ARM_GET_SYM_BRANCH_TYPE to get
branch type of a symbol.
(print_insn): Likewise.
Allow stubs without associated input section in ARM backend
2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com>
bfd/
* bfd-in.h (elf32_arm_size_stubs): Add an output section parameter.
* bfd-in2.h: Regenerated.
* elf32-arm.c (struct elf32_arm_link_hash_table): Add an output section
parameter to add_stub_section callback.
(elf32_arm_create_or_find_stub_sec): Get output section from link_sec
and pass it down to add_stub_section.
(elf32_arm_add_stub): Set section to stub_sec if NULL before using it
for error message.
(elf32_arm_size_stubs): Add output section parameter to
add_stub_section function pointer parameter.
ld/
* emultempl/armelf.em (elf32_arm_add_stub_section): Add output_section
parameter and rename input_section parameter to after_input_section.
Append input stub section to the output section if after_input_section
is NULL.
Refactor Cortex-A8 erratum workaround in preparation
2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com>
bfd/
* elf32-arm.c (enum elf32_arm_stub_type): Delete
arm_stub_a8_veneer_lwm enumerator.
(arm_stub_a8_veneer_lwm): New unsigned constant to replace
aforementioned enumerator.
(struct elf32_arm_stub_hash_entry): Delete target_addend
field and add source_value.
(struct a8_erratum_fix): Delete addend field and add target_offset.
(stub_hash_newfunc): Initialize source_value field amd remove
initialization for target_addend.
(arm_build_one_stub): Stop special casing Thumb relocations: promote
the else to being always executed, moving the
arm_stub_a8_veneer_b_cond specific code in it. Remove
stub_entry->target_addend from points_to computation.
(cortex_a8_erratum_scan): Store in a8_erratum_fix structure the offset
to target symbol from start of section rather than the offset from the
stub address.
(elf32_arm_size_stubs): Set stub_entry's source_value and target_value
fields from struct a8_erratum_fix's offset and target_offset
respectively.
(make_branch_to_a8_stub): Rename target variable to loc. Compute
veneered_insn_loc and loc using stub_entry's source_value.
gas/
* NEWS: Document ARMv8-M and ARMv8-M Security and DSP Extensions.
* config/tc-arm.c (arm_ext_dsp): New feature for Thumb DSP
instructions.
(arm_extensions): Add dsp extension for ARMv8-M Mainline.
(aeabi_set_public_attributes): Memorize the feature bits of the
architecture selected for Tag_CPU_arch. Use it to set
Tag_DSP_extension to 1 for ARMv8-M Mainline with DSP extension.
(arm_convert_symbolic_attribute): Define Tag_DSP_extension.
* testsuite/gas/arm/arch7em-bad.d: Rename to ...
* testsuite/gas/arm/arch7em-bad-1.d: This.
* testsuite/gas/arm/arch7em-bad-2.d: New file.
* testsuite/gas/arm/arch7em-bad-3.d: Likewise.
* testsuite/gas/arm/archv8m-main-dsp-1.d: Likewise.
* testsuite/gas/arm/archv8m-main-dsp-2.d: Likewise.
* testsuite/gas/arm/archv8m-main-dsp-3.d: Likewise.
* testsuite/gas/arm/archv8m-main-dsp-4.d: Likewise.
* testsuite/gas/arm/archv8m-main-dsp-5.d: Likewise.
* testsuite/gas/arm/attr-march-armv8m.main.dsp.d: Likewise.
include/
* elf/arm.h (Tag_DSP_extension): Define.
ld/
* testsuite/ld-arm/arm-elf.exp (EABI attribute merging 10 (DSP)): New
test.
* testsuite/ld-arm/attr-merge-10b-dsp.s: New file.
* testsuite/ld-arm/attr-merge-10-dsp.attr: Likewise.
Allow extension availability to depend on several architecture bits
2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com>
gas/
* config/tc-arm.c (struct arm_option_extension_value_table): Make
allowed_archs an array with 2 entries.
(ARM_EXT_OPT): Adapt to only fill the first entry of allowed_archs.
(ARM_EXT_OPT2): New macro filling the two entries of allowed_archs.
(arm_extensions): Use separate entries in allowed_archs when several
archs are allowed to use an extension and change ARCH_ANY in
ARM_ARCH_NONE in allowed_archs.
(arm_parse_extension): Check that, for each allowed_archs entry, all
bits are set in the current architecture, ignoring ARM_ANY entries.
(s_arm_arch_extension): Likewise.
Add support for ARMv8-M security extensions instructions
2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com>
gas/
* config/tc-arm.c (arm_ext_m): Add feature bit ARM_EXT2_V8M_MAIN.
(arm_ext_v8m_main): New feature set for bit ARM_EXT2_V8M_MAIN.
(arm_ext_v8m_m_only): New feature set for instructions in ARMv8-M not
shared with a non M profile architecture.
(do_rn): New function.
(known_t32_only_insn): Check opcode against arm_ext_v8m_m_only rather
than arm_ext_v8m.
(v7m_psrs): Add ARMv8-M security extensions new special registers.
(insns): Add ARMv8-M Security Extensions instructions.
(aeabi_set_public_attributes): Use arm_ext_v8m_m_only instead of
arm_ext_v8m_m to decide the profile and the Thumb ISA.
* testsuite/gas/arm/archv8m-cmse.s: New file.
* testsuite/gas/arm/archv8m-cmse-main.s: Likewise..
* testsuite/gas/arm/archv8m-cmse-msr.s: Likewise.
* testsuite/gas/arm/any-cmse.d: Likewise.
* testsuite/gas/arm/any-cmse-main.d: Likewise.
* testsuite/gas/arm/archv8m-cmse-base.d: Likewise.
* testsuite/gas/arm/archv8m-cmse-msr-base.d: Likewise.
* testsuite/gas/arm/archv8m-cmse-main-1.d: Likewise.
* testsuite/gas/arm/archv8m-cmse-main-2.d: Likewise.
* testsuite/gas/arm/archv8m-cmse-msr-main.d: Likewise.
include/
* opcode/arm.h (ARM_EXT2_V8M_MAIN): new feature bit.
(ARM_AEXT2_V8M_MAIN): New architecture extension feature set.
(ARM_ARCH_V8M_MAIN): Use ARM_AEXT2_V8M_MAIN instead of ARM_AEXT2_V8M
for the high core bits.
opcodes/
* arm-dis.c (coprocessor_opcodes): Add entries for VFP ARMv8-M
Mainline Security Extensions instructions.
(thumb_opcodes): Add entries for narrow ARMv8-M Security
Extensions instructions.
(thumb32_opcodes): Add entries for wide ARMv8-M Security Extensions
instructions.
(psr_name): Add new MSP_NS and PSP_NS ARMv8-M Security Extensions
special registers.
Andrew Burgess [Fri, 29 Apr 2016 12:22:51 +0000 (13:22 +0100)]
binutils/doc: Update documentation for nm --size-sort
The documentation for 'nm --size-sort' is out of step with the code, the
special handling for ELF is not mentioned in the documentation, and
could cause confusion to a user.
binutils/ChangeLog:
* doc/binutils.texi (nm): Update description of --size-sort.
PR ld/20059
* elfxx-target.h (bfd_elfNN_bfd_copy_link_hash_symbol_type):
Define as _bfd_generic_copy_link_hash_symbol_type when using
generic hash table.
Alan Modra [Mon, 9 May 2016 07:37:33 +0000 (17:07 +0930)]
Correct script input section pattern
We don't want this to match .rela.text or similar.
* testsuite/ld-scripts/pr14962-2.t: Match .text, not *.text.
* testsuite/ld-scripts/rgn-at5.t: Similarly, .sec not *.sec.
* testsuite/ld-scripts/section-match-1.t: Likewise.
H.J. Lu [Fri, 6 May 2016 16:46:34 +0000 (09:46 -0700)]
Strip global symbol defined in discarded section
When a global symbol is defined in COMDAT group, we shouldn't leave an
undefined symbol in symbol table when the symbol section is discarded
unless there is a reference to the symbol outside of COMDAT group.
bfd/
PR ld/17550
* elf-bfd.h (elf_link_hash_entry): Update comments for indx,
documenting that indx == -3 if symbol is defined in a discarded
section.
* elflink.c (elf_link_add_object_symbols): Set indx to -3 if
symbol is defined in a discarded section.
(elf_link_output_extsym): Strip a global symbol defined in a
discarded section.
H.J. Lu [Thu, 5 May 2016 11:13:47 +0000 (04:13 -0700)]
Extract convert_load_reloc from x86 convert_load
* elf32-i386.c (elf_i386_convert_load): Extract the GOT load
conversion to ...
(elf_i386_convert_load_reloc): This. New function.
* elf64-x86-64.c (elf_x86_64_convert_load): Extract the GOT load
conversion to ...
(elf_x86_64_convert_load_reloc): This. New function.
Ulrich Weigand [Wed, 4 May 2016 23:42:09 +0000 (19:42 -0400)]
[spu] Fix C++ build problems
ChangeLog:
* spu-linux-nat.c (spu_bfd_iovec_pread): Add pointer cast for C++.
(spu_bfd_open): Likewise.
gdbserver/ChangeLog:
* spu-low.c (fetch_ppc_register): Cast PowerPC-Linux-specific value
used as first ptrace argument to PTRACE_TYPE_ARG1 for C++.
(fetch_ppc_memory_1, store_ppc_memory_1): Likewise.
LD/testsuite: Add STB_GNU_UNIQUE cross-linker test
Some targets are only really, or at least regularly, regression-tested
in a crossed configuration. Currently we only have native compiled test
cases for the STB_GNU_UNIQUE feature in the linker test suite. This is
nice, covering run-time semantics even, but quite often not run at all.
Consequently a regression may remain unnoticed for long.
Add a simple test case then to provide basic linker coverage with no
need for a compiler or a native toolchain.
ld/
* testsuite/ld-unique/unique.d: New test.
* testsuite/ld-unique/unique.exp: Run the new test. Adjust
messages for compiled tests.
Yao Qi [Wed, 4 May 2016 14:04:01 +0000 (15:04 +0100)]
Throw NOT_AVAILABLE_ERROR in read_stack and read_code
Nowadays, read_memory may throw NOT_AVAILABLE_ERROR (it is done by
patch http://sourceware.org/ml/gdb-patches/2013-08/msg00625.html)
however, read_stack and read_code still throws MEMORY_ERROR only. This
causes PR 19947, that is prologue unwinder is unable unwind because
code memory isn't available, but MEMORY_ERROR is thrown, while unwinder
catches NOT_AVAILABLE_ERROR.
#0 memory_error (err=err@entry=TARGET_XFER_E_IO, memaddr=memaddr@entry=140737349781158) at /home/yao/SourceCode/gnu/gdb/git/gdb/corefile.c:217
#1 0x000000000065f5ba in read_code (memaddr=memaddr@entry=140737349781158, myaddr=myaddr@entry=0x7fffffffd7b0 "\340\023<\001", len=len@entry=1)
at /home/yao/SourceCode/gnu/gdb/git/gdb/corefile.c:288
#2 0x000000000065f7b5 in read_code_unsigned_integer (memaddr=memaddr@entry=140737349781158, len=len@entry=1, byte_order=byte_order@entry=BFD_ENDIAN_LITTLE)
at /home/yao/SourceCode/gnu/gdb/git/gdb/corefile.c:363
#3 0x00000000004717e0 in amd64_analyze_prologue (gdbarch=gdbarch@entry=0x13c13e0, pc=140737349781158, current_pc=140737349781165, cache=cache@entry=0xda0cb0)
at /home/yao/SourceCode/gnu/gdb/git/gdb/amd64-tdep.c:2267
#4 0x0000000000471f6d in amd64_frame_cache_1 (cache=0xda0cb0, this_frame=0xda0bf0) at /home/yao/SourceCode/gnu/gdb/git/gdb/amd64-tdep.c:2437
#5 amd64_frame_cache (this_frame=0xda0bf0, this_cache=<optimised out>) at /home/yao/SourceCode/gnu/gdb/git/gdb/amd64-tdep.c:2508
#6 0x000000000047214d in amd64_frame_this_id (this_frame=<optimised out>, this_cache=<optimised out>, this_id=0xda0c50)
at /home/yao/SourceCode/gnu/gdb/git/gdb/amd64-tdep.c:2541
#7 0x00000000006b94c4 in compute_frame_id (fi=0xda0bf0) at /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:481
#8 get_prev_frame_if_no_cycle (this_frame=this_frame@entry=0xda0b20) at /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:1809
#9 0x00000000006bb6c9 in get_prev_frame_always_1 (this_frame=0xda0b20) at /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:1983
#10 get_prev_frame_always (this_frame=this_frame@entry=0xda0b20) at /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:1999
#11 0x00000000006bbe11 in get_prev_frame (this_frame=this_frame@entry=0xda0b20) at /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:2241
#12 0x00000000006bc13c in unwind_to_current_frame (ui_out=<optimised out>, args=args@entry=0xda0b20) at /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:1485
The fix is to let read_stack and read_code throw NOT_AVAILABLE_ERROR too,
in order to align with read_memory.
gdb:
2016-05-04 Yao Qi <yao.qi@linaro.org>
PR gdb/19947
* corefile.c (read_memory): Rename it to ...
(read_memory_object): ... it. Add parameter object.
(read_memory): Call read_memory_object.
(read_stack): Likewise.
(read_code): Likewise.
Simon Marchi [Wed, 4 May 2016 13:27:34 +0000 (09:27 -0400)]
Fix solib-display.exp remote check
This test currently uses [is_remote target] to check if the test is
supported. This is not quite correct, as the limitation is actually
that it requires support for "running", ruling out stub-like targets.
Therefore, it should check for use_gdb_stub.
This has no visible effect right now, but it will once we make the
native-gdbserver board non-dejagnu-remote.
gdb/testsuite/ChangeLog:
* gdb.base/solib-display.exp: Check for [use_gdb_stub] instead
of [is_remote target],
Simon Marchi [Wed, 4 May 2016 13:27:19 +0000 (09:27 -0400)]
Introduce procedure use_gdb_stub
This patch introduces the use_gdb_stub procedure, which allows getting
the right value of the use_gdb_stub variable/property in any all
situations.
When calling it before the $use_gdb_stub global variable has been set,
it will return the value of the use_gdb_stub property from the board
file. This happens when tests want to bail out early (even before gdb
has been started) when the current test setup is a stub.
Otherwise, it returns the value of the $use_gdb_stub global.
It's possible for these two to differ when a test file overrides the
value of the global.
ld * testsuite/ld-elf/pr18735.d: Allow other symbols.
* testsuite/ld-elf/sec64k.exp: Skip 64ksec for avr.
* testsuite/ld-gc/pr14265.d: Allow other symbols.
* testsuite/ld-plugin/plugin.exp: Add PR ld/17973 to
plugin_tests only if check_shared_lib_support is true.
* testsuite/ld-selective/selective.exp: Add --section-start
flag for avr.
Pedro Alves [Tue, 3 May 2016 11:16:56 +0000 (12:16 +0100)]
Fix gdb/python/python.c use-after-free
Valgrind shows:
==26964== Invalid read of size 1
==26964== at 0x6E14100: __GI_strcmp (strcmp.S:180)
==26964== by 0x6DB55AA: setlocale (setlocale.c:238)
==26964== by 0x4E0455: _initialize_python() (python.c:1731)
==26964== by 0x786731: initialize_all_files() (init.c:319)
==26964== by 0x72EF0A: gdb_init(char*) (top.c:1929)
==26964== by 0x60BCAC: captured_main(void*) (main.c:863)
==26964== by 0x606AD5: catch_errors(int (*)(void*), void*, char*, return_mask) (exceptions.c:234)
==26964== by 0x60C608: gdb_main(captured_main_args*) (main.c:1165)
==26964== by 0x40CAEC: main (gdb.c:32)
==26964== Address 0x81d30a0 is 0 bytes inside a block of size 181 free'd
==26964== at 0x4C29CF0: free (vg_replace_malloc.c:530)
==26964== by 0x6DB5B65: setname (setlocale.c:201)
==26964== by 0x6DB5B65: setlocale (setlocale.c:388)
==26964== by 0x4E037F: _initialize_python() (python.c:1712)
==26964== by 0x786731: initialize_all_files() (init.c:319)
==26964== by 0x72EF0A: gdb_init(char*) (top.c:1929)
==26964== by 0x60BCAC: captured_main(void*) (main.c:863)
==26964== by 0x606AD5: catch_errors(int (*)(void*), void*, char*, return_mask) (exceptions.c:234)
==26964== by 0x60C608: gdb_main(captured_main_args*) (main.c:1165)
==26964== by 0x40CAEC: main (gdb.c:32)
Jiong Wang [Tue, 3 May 2016 10:59:37 +0000 (11:59 +0100)]
[AArch64] Also puts value in place for R_AARCH64_RELATIVE
When handling absolute relocations for global symbols bind within the
shared object, AArch64 will generate one dynamic RELATIVE relocation,
but won't apply the value for this absolution relocations at static
linking stage. This is different from AArch64 gold linker and x86-64.
This is not a bug as AArch64 is RELA, there is only guarantee that
relocation addend is placed in the relocation entry. But some
system softwares originally writen for x86-64 might assume AArch64
bfd linker gets the same behavior as x86-64, then they could take
advantage of this buy skipping those RELATIVE dynamic relocations
if the load address is the same as the static linking address.
This patch makes AArch64 BFD linker applies absolution relocations at
static linking stage for scenario described above. Meanwhile old AArch64
android loader has a bug (PR19163) which relies on current linker behavior
as a workaround, so the same option --no-apply-dynamic-relocs added.
projects / thirdparty / binutils-gdb.git / log
