Max Filippov [Mon, 23 Jul 2018 01:59:11 +0000 (18:59 -0700)]
xtensa: move dynamic relocations sections consistency check
The function elf_xtensa_finish_dynamic_sections checks that sizes of
sections .rela.dyn and .rela.plt match number of corresponding relocation
records, but the check is only done when .rela.plt is non-empty, so, e.g.
it is never run for the static PIE.
Rearrange the test so that .rela.dyn and .rela.plt are checked always.
bfd/
2018-07-23 Max Filippov <jcmvbkbc@gmail.com>
* elf32-xtensa.c (elf_xtensa_finish_dynamic_sections): Move
relocation sections consistency check to always check both
.rela.dyn and .rela.plt when they exist. Rearrange variable
definition and assignment places.
Max Filippov [Sun, 22 Jul 2018 20:52:28 +0000 (13:52 -0700)]
xtensa: fix relaxation of undefined weak references in shared objects
The change
c451bb34ae8b ("xtensa: don't emit dynamic relocation for weak
undefined symbol") didn't properly handle shrinking of relocation
sections due to coalescing of references to a dynamic undefined weak
symbol in a shared object, which resulted in the following assertion
failure in ld when linking uClibc-ng libthread_db for xtensa:
BFD (GNU Binutils) 2.31 internal error, aborting at elf32-xtensa.c:3269
in elf_xtensa_finish_dynamic_sections
Shrink dynamic relocations section for dynamic undefined weak symbols
when linking a shared object.
bfd/
2018-07-23 Max Filippov <jcmvbkbc@gmail.com>
* elf32-xtensa.c (shrink_dynamic_reloc_sections): Shrink dynamic
relocations section for dynamic undefined weak symbols when
linking a shared object.
Tom Tromey [Mon, 16 Jul 2018 00:07:40 +0000 (18:07 -0600)]
Avoid ubsan complaint in BFD
I built gdb with ubsan and ran the test suite.
One complaint was due to bfd_get_elf_phdrs passing NULL to memcpy.
This patch avoids the complaint.
bfd/ChangeLog
2018-07-23 Tom Tromey <tom@tromey.com>
* elf.c (bfd_get_elf_phdrs): Don't call memcpy with size 0.
H.J. Lu [Mon, 23 Jul 2018 13:32:20 +0000 (06:32 -0700)]
x86: Add a test for missing broadcast
For
.intel_syntax noprefix
vcvtps2qq xmm0, DWORD PTR [rax]
we should get
Error: broadcast is needed for operand of such type for `vcvtps2qq'
* testsuite/gas/i386/inval-avx512f.s: Add a test for missing
broadcast.
* testsuite/gas/i386/x86-64-inval-avx512f.s: Likewise.
* testsuite/gas/i386/inval-avx512f.l: Updated.
* testsuite/gas/i386/x86-64-inval-avx512f.l: Likewise.
H.J. Lu [Mon, 23 Jul 2018 12:43:43 +0000 (05:43 -0700)]
x86: Remove broadcast_not_on_src_operand
Remove broadcast_not_on_src_operand since it is unused.
* config/tc-i386.c (i386_error): Remove
broadcast_not_on_src_operand.
(match_template): Likewse.
Claudiu Zissulescu [Mon, 23 Jul 2018 09:09:43 +0000 (11:09 +0200)]
[ARC] Fix decoding of w6 signed short immediate.
gas/
Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/st.d: Fix test.
opcodes/
Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (extract_w6): Fix extending the sign.
Claudiu Zissulescu [Mon, 23 Jul 2018 09:09:43 +0000 (11:09 +0200)]
[ARC] Allow vewt instruction for ARC EM family.
opcode/
Claudiu Zissulescu <claziss@synopsys.com>
* arc-tbl.h (vewt): Allow it for ARC EM family.
claziss [Mon, 23 Jul 2018 09:09:43 +0000 (11:09 +0200)]
[ARC] Fix case-sensitivity for extension instructions.
In ARC assembler, we accept case insensitive mnemonics, but this was
not the case for extension instruction, fix it and add a test.
gas/
Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (tokenize_extinsn): Convert to lower case the
name of extension instructions.
* testsuite/gas/arc/textinsn_case.d: New file.
* testsuite/gas/arc/textinsn_case.s: Likewise.
Andreas Krebbel [Mon, 23 Jul 2018 08:29:37 +0000 (10:29 +0200)]
S/390: Suppress string diagnostics for pre-release GCC
Building for S/390 currently fails when using a GCC 8.1 pre-release
version like 8.0.1 (as it is used by F28). Fixed with the attached patch.
bfd/ChangeLog:
2018-07-23 Andreas Krebbel <krebbel@linux.ibm.com>
* elf32-s390.c (elf_s390_write_core_note): Make GCC version 8.0 to
be covered by the check as well.
* elf64-s390.c (elf_s390_write_core_note): Likewise.
Alan Modra [Mon, 23 Jul 2018 03:01:18 +0000 (12:31 +0930)]
PowerPC64 fails reloc-discard test
PowerPC64 fails here due to emitting a warning about discarded dynamic
relocations. That's a good thing, since accidentally discarding
dynamic relocations will likely result in a non-working binary.
* testsuite/ld-elf/reloc-discard-warn.d: New.
* testsuite/ld-elf/reloc-discard.d: Skip test on PowerPC64.
Alan Modra [Mon, 23 Jul 2018 03:29:23 +0000 (12:59 +0930)]
power9 mfupmc/mtupmc
PR 23419
* ppc-opc.c (powerpc_opcodes): Add mtupmc/mfupmc/mfpmc extended
opcode variants for mtspr/mfspr encodings.
Simon Marchi [Mon, 23 Jul 2018 03:23:11 +0000 (23:23 -0400)]
Remove some unused variables
Found using the newly-enabled -Wunused-variable.
gdb/ChangeLog:
* aarch64-linux-nat.c
(aarch64_linux_nat_target::stopped_data_address): Remove unused
variable.
* arm-linux-nat.c (fetch_regs): Likewise.
(store_regs): Likewise.
(fetch_vfp_regs): Likewise.
(store_vfp_regs): Likewise.
(arm_linux_nat_target::insert_hw_breakpoint): Likewise.
(arm_linux_nat_target::remove_hw_breakpoint): Likewise.
(arm_linux_nat_target::insert_watchpoint): Likewise.
(arm_linux_nat_target::remove_watchpoint): Likewise.
* mips-linux-nat.c (mips_linux_nat_target::insert_watchpoint):
Likewise.
* nat/aarch64-sve-linux-ptrace.c (aarch64_sve_get_sveregs):
Likewise.
* ppc-linux-nat.c (fetch_register): Likewise.
(fetch_all_gp_regs): Likewise.
(fetch_ppc_registers): Likewise.
(store_all_gp_regs): Likewise.
(store_ppc_registers): Likewise.
(hwdebug_insert_point): Likewise.
(can_use_watchpoint_cond_accel): Likewise.
* remote-sim.c (gdb_os_write_stdout): Likewise.
gdb/gdbserver/ChangeLog:
* linux-mips-low.c (mips_collect_ptrace_register): Remove unused
variable.
(mips_supply_ptrace_register): Likewise.
GDB Administrator [Mon, 23 Jul 2018 00:01:17 +0000 (00:01 +0000)]
Automatic date update in version.in
Tom Tromey [Mon, 9 Jul 2018 03:05:41 +0000 (21:05 -0600)]
Add -Wunused-variable to warnings.m4
This adds -Wunused-variable to the build. This required a special
check in configure in order to work around a bug in GCC 4.9. Simon
ound the correct test to use, so I've added him to the ChangeLog.
gdb/ChangeLog
2018-07-22 Simon Marchi <simon.marchi@polymtl.ca>
Tom Tromey <tom@tromey.com>
* warning.m4 (AM_GDB_WARNINGS): Add -Wunused-variable and special
test for it.
* configure: Rebuild.
gdb/gdbserver/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
Tom Tromey [Mon, 9 Jul 2018 02:31:58 +0000 (20:31 -0600)]
Conditionally define xmltarget_${name} variable in regdat.sh
This changes regdat.sh to emit the xmltarget_${name} variable inside
the #ifndef IN_PROCESS_AGENT block. This avoids a -Wunused-variable
warning for some builds. Thanks to Pedro for investigating this one.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* regformats/regdat.sh: Define xmltarget_${name} inside
#ifndef IN_PROCESS_AGENT.
Tom Tromey [Sun, 8 Jul 2018 18:09:22 +0000 (12:09 -0600)]
Remove unused variables from gdbserver
This removes a few unused variables from gdbserver.
The x86-tdesc.h change is a bit unusual for this series. This file
was not defining the multiple-include guard symbol, so I've added that
here. Also, it is hard to determine when i386_expedite_regs will be
needed, so this patch simply marks it ATTRIBUTE_UNUSED.
gdb/gdbserver/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* win32-low.c (win32_create_inferior): Remove unused variables.
* gdbreplay.c (remote_open): Remove unused variable.
* remote-utils.c (remote_prepare): Remove unused variable.
* x86-tdesc.h (X86_TDESC_H): Define.
(amd64_expedite_regs): Define conditionally.
(i386_expedite_regs): Mark ATTRIBUTE_UNUSED.
* linux-x86-tdesc.c (i386_tdescs): Move inside #if.
* remote-utils.c (readchar): Remove unused variable.
Tom Tromey [Sun, 8 Jul 2018 18:18:19 +0000 (12:18 -0600)]
Remove unused declaration from value.c
This removes an unused declaration from value_fetch_lazy_bitfield.
Because it is not completely clear if the check_typedef call is needed
somewhere beneath this function, this is broken out into a separate
patch.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* value.c (value_fetch_lazy_bitfield): Remove unused variable.
Tom Tromey [Sun, 8 Jul 2018 18:18:09 +0000 (12:18 -0600)]
Pass the correct argument to the observer in reread_symbols
This is actually a patch I found via another route. Joel had asked me
to write a test, but I still have not found the time to do this.
Meanwhile, -Wunused-variable also found this error.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* symfile.c (reread_symbols): Notify iter, not objfile.
Tom Tromey [Sun, 8 Jul 2018 18:17:51 +0000 (12:17 -0600)]
Fix ravenscar-thread.c to use arch_ops
The change to turn target ops into methods accidentally introduced a
bug in ravenscar-thread.c, changing some calls that were using
"arch_ops" to use the target beneath.
This patch changes ravenscar-thread.c to use these variables where
appropriate.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* ravenscar-thread.c (ravenscar_thread_target::store_registers):
Use arch_ops.
(ravenscar_thread_target::prepare_to_store): Likewise.
Tom Tromey [Sun, 8 Jul 2018 18:17:36 +0000 (12:17 -0600)]
Call value_fetch_lazy when needed in pretty-printers
This removes some unused variables, and replaces the calls to
value_contents_for_printing with a call to value_fetch_lazy, when
needed.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* python/py-prettyprint.c (gdbpy_apply_val_pretty_printer): Remove
unused variable. Call value_fetch_lazy when needed.
* guile/scm-pretty-print.c (gdbscm_apply_val_pretty_printer):
Remove unused variable. Call value_fetch_lazy when needed.
Tom Tromey [Sun, 8 Jul 2018 18:17:23 +0000 (12:17 -0600)]
Remove declarations from m32c-tdep.c
This removes some unused declarations from m32c-tdep.c. Initially I
had thought that this entire code block was dead, but Simon pointed
out that it is not; so this patch simply removes some declarations.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* m32c-tdep.c (mark_dma): Return void.
(make_regs): Remove unused declarations.
Tom Tromey [Sun, 8 Jul 2018 18:16:20 +0000 (12:16 -0600)]
Call some functions in guile/ for effect
This changes a couple of spots in guile/ to remove a variable
declaration but to still call a function for effect.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* guile/scm-cmd.c (gdbscm_dont_repeat): Call
cmdscm_get_valid_command_smob_arg_unsafe for effect.
* guile/scm-block.c (gdbscm_make_block_syms_iter): Call
bkscm_get_valid_block_smob_arg_unsafe for effect.
Tom Tromey [Sun, 8 Jul 2018 18:16:00 +0000 (12:16 -0600)]
Use a previously unused variable in bfin-tdep.c
This changes bfin_push_dummy_call to use the result of check_typedef.
Calling check_typedef for effect was probably ok as well, but this
seemed a little nicer.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* bfin-tdep.c (bfin_push_dummy_call): Use arg_type, not
value_type.
Tom Tromey [Sun, 8 Jul 2018 18:39:36 +0000 (12:39 -0600)]
Unused variable fixes related to conditional compilation
This patch fixes various unused variable warnings that are related to
conditional compilation. In these cases, either the variable is now
protected by the same #if as its uses, or the declaration is simply
lowered into the conditionally-compiled block.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* windows-nat.c (saved_context): Conditionally define.
* remote.c (remote_target::remote_btrace_maybe_reopen):
Conditionally declare "warned".
* inflow.c (sigquit_ours): Conditionally define.
(new_tty): Move "tty" declaration inside #if.
* guile/guile.c (guile_datadir): Conditionally define.
* charset.c (set_be_le_names): Move some declarations inside #if.
* btrace.c (parse_xml_btrace): Move "errcode" declaration inside
#if.
(parse_xml_btrace_conf): Likewise.
Tom Tromey [Mon, 16 Jul 2018 15:33:42 +0000 (09:33 -0600)]
Remove an unused variable from spu_get_overlay_table
This removes an unused variable from spu_get_overlay_table, replacing
it with an explanatory comment.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* spu-tdep.c (spu_get_overlay_table): Remove unused variable.
Tom Tromey [Sun, 8 Jul 2018 18:15:04 +0000 (12:15 -0600)]
Simple unused variable removals
This patch holds all the straightforward unused variable deletions.
gdb/ChangeLog
2018-07-22 Tom Tromey <tom@tromey.com>
* guile/scm-value.c (gdbscm_value_call): Remove unused variables.
* guile/scm-math.c (vlscm_unop_gdbthrow, vlscm_binop_gdbthrow)
(vlscm_convert_typed_value_from_scheme): Remove unused variable.
* buildsym-legacy.c (get_macro_table): Remove unused variable.
* stack.c (frame_apply_level_command): Remove unused variable.
* tic6x-tdep.c (tic6x_push_dummy_call): Remove unused variable.
* sparc64-tdep.c (adi_examine_command): Remove unused variable.
* rs6000-lynx178-tdep.c (rs6000_lynx178_push_dummy_call): Remove
unused variable.
* nios2-tdep.c (nios2_push_dummy_call): Remove unused variable.
* mep-tdep.c (mep_push_dummy_call): Remove unused variable.
* ada-lang.c (ada_lookup_symbol_list_worker): Remove unused
variable.
* amd64-tdep.c (amd64_supply_xsave): Remove unused variable.
* arm-tdep.c (arm_record_data_proc_misc_ld_str): Remove unused
variable.
* breakpoint.c (check_no_tracepoint_commands, update_watchpoint):
Remove unused variable.
* cli/cli-script.c (recurse_read_control_structure): Remove unused
variable.
* common/tdesc.c (print_xml_feature::visit): Remove unused
variable.
* compile/compile-object-load.c (store_regs): Remove unused
variables.
* complaints.c (clear_complaints): Remove unused variable.
* corelow.c (core_target_open): Remove unused variable.
* fbsd-tdep.c (fbsd_core_info_proc_status): Remove unused
variable.
* guile/scm-frame.c (gdbscm_frame_read_var): Remove unused
variable.
* guile/scm-symtab.c (stscm_print_sal_smob): Remove unused
variable.
* guile/scm-type.c (gdbscm_field_baseclass_p): Remove unused
variable.
* guile/scm-utils.c (gdbscm_parse_function_args): Remove unused
variable.
* hppa-tdep.c (hppa_stub_frame_unwind_cache): Remove unused
variable.
* ia64-tdep.c (examine_prologue): Remove unused variable.
* infcall.c (run_inferior_call): Remove unused variable.
* inferior.c (exit_inferior): Remove unused variable.
* infrun.c (infrun_thread_ptid_changed): Remove unused variable.
* linespec.c (decode_line_2): Remove unused variable.
* linux-nat.c (super_close): Remove.
* linux-tdep.c (linux_info_proc): Remove unused variable.
* mi/mi-main.c (mi_execute_command): Remove unused variable.
* microblaze-linux-tdep.c (microblaze_linux_sigtramp_cache):
Remove unused variable.
* parse.c (find_minsym_type_and_address): Remove unused variable.
* printcmd.c (info_symbol_command, printf_floating): Remove unused
variable.
* python/py-breakpoint.c (bppy_set_commands): Remove unused
variable.
* python/py-unwind.c (unwind_infopy_dealloc): Remove unused
variables.
* record-btrace.c (record_btrace_target::store_registers): Remove
unused variable.
(cmd_show_record_btrace_cpu): Remove unused variable.
* riscv-tdep.c (riscv_register_reggroup_p)
(riscv_push_dummy_call, riscv_return_value): Remove unused
variable.
* rust-exp.y (literal): Remove unused variable.
* rust-lang.c (rust_evaluate_subexp) <OP_RUST_ARARAY>: Remove
unused variable.
<STRUCTOP_ANONYMOUS>: Likewise.
* s390-linux-tdep.c (s390_linux_init_abi_31)
(s390_linux_init_abi_64): Remove unused variable.
* ser-ming2.c (ser_windows_read_prim, pipe_select_thread)
(file_select_thread, net_windows_open, _initialize_ser_windows):
Remove unused variables.
* symtab.c (find_pc_sect_line): Remove unused variable.
* target-memory.c (compute_garbled_blocks): Remove unused
variable.
(target_write_memory_blocks): Remove unused variable.
* target.c (target_stack::unpush): Remove unused variables.
* tracepoint.c (start_tracing, all_tracepoint_actions)
(merge_uploaded_trace_state_variables)
(print_one_static_tracepoint_marker): Remove unused variable.
* unittests/basic_string_view/element_access/char/1.cc (test01):
Remove unused variable.
* windows-nat.c (windows_continue, windows_add_all_dlls)
(do_initial_windows_stuff, windows_nat_target::create_inferior):
Remove unused variables.
H.J. Lu [Sun, 22 Jul 2018 19:00:39 +0000 (12:00 -0700)]
x86: Determine vector length from the last vector operand
Determine VEX/EVEXE vector length from the last multi-length vector
operand.
* config/tc-i386.c (build_vex_prefix): Determine vector
length from the last multi-length vector operand.
(build_evex_prefix): Likewise.
Simon Marchi [Sun, 22 Jul 2018 02:21:57 +0000 (22:21 -0400)]
Fix -Wunused-variable warnings related to conditional compilation
Some variables are not used when !HAVE_ELF, so avoid declaring them in
that case.
gdb/ChangeLog:
* arm-tdep.c (arm_gdbarch_init): Declare attr_arch and
attr_profile in HAVE_ELF.
* rs6000-tdep.c (bfd_uses_spe_extensions): Declare vector_abi in
HAVE_ELF.
GDB Administrator [Sun, 22 Jul 2018 00:01:12 +0000 (00:01 +0000)]
Automatic date update in version.in
H.J. Lu [Sat, 21 Jul 2018 17:57:54 +0000 (10:57 -0700)]
gas/config/tc-i386.c: Break long line
* config/tc-i386.c (match_simd_size): Break long line.
(match_mem_size): Likewise.
Eric Botcazou [Sat, 21 Jul 2018 13:00:40 +0000 (15:00 +0200)]
Initialize GOT slot for local symbol in non-PIC link.
On ARM/VxWorks 7, the R_ARM_TARGET2 relocation used for exception handling
is R_ARM_GOT_PREL; moreover in Ada you can define local exceptions. In this
case, you may end up with a GOT relocation against a local symbol in a
non-PIC link and the ARM linker leaves the GOT slot uninitialized, unlike
for example the i386 or the SPARC linkers in the same situation.
bfd/
* elf32-arm.c (elf32_arm_final_link_relocate) <R_ARM_GOT32>: Small
cleanup for the case of a global symbol that binds locally. Also
install a value in the GOT slot in the case of a local symbol in
a non-PIC link.
Simon Marchi [Sat, 21 Jul 2018 02:19:04 +0000 (22:19 -0400)]
Rename some frame unwind function parameters
I am currently working with these functions, and though this renaming
could help to reason about the code. Some functions take a frame and
will return the value associated to that frame, others will return the
value associated to the previous frame. Those usually conveniently
contain "unwind" in their name, but naming the variable next_frame
instead of frame helps remembering which frame we are dealing with.
I also included a little typo fix at the top of frame.h.
gdb/ChangeLog:
* frame.c (frame_register_unwind): Change parameter name.
(frame_unwind_register): Likewise.
(frame_unwind_register_value): Likewise.
(frame_unwind_register_signed): Likewise.
(frame_unwind_register_unsigned): Likewise.
* frame.h (frame_register_unwind): Likewise.
(frame_unwind_register): Likewise.
(frame_unwind_register_value): Likewise.
(frame_unwind_register_signed): Likewise.
(frame_unwind_register_unsigned): Likewise.
(frame_unwind_arch): Likewise.
GDB Administrator [Sat, 21 Jul 2018 00:00:33 +0000 (00:00 +0000)]
Automatic date update in version.in
Maciej W. Rozycki [Fri, 20 Jul 2018 23:14:01 +0000 (00:14 +0100)]
Update my e-mail address, limit maintenance to MIPS I-IV ISAs
binutils/
* MAINTAINERS: Update my e-mail address, downgrade to MIPS I-IV
ISA maintenance.
gdb/
* MAINTAINERS: Update my e-mail address, downgrade to MIPS I-IV
ISA maintenance.
sim/
* MAINTAINERS: Update my e-mail address, downgrade to MIPS I-IV
ISA maintenance.
H.J. Lu [Fri, 20 Jul 2018 19:47:25 +0000 (12:47 -0700)]
x86: Rename match_reg_size to match_operand_size
match_reg_size checks size for both memory and register operands. This
patch renamed match_reg_size to match_operand_size and updated comments
for
commit
3ac21baa8498d3aa9951f79e2c3336d532eeff7b
Author: Jan Beulich <jbeulich@novell.com>
Date: Mon Jul 16 08:19:21 2018 +0200
x86: fix operand size checking
which added one argument to match_reg_size, match_simd_size and
match_mem_size.
* config/tc-i386.c (match_reg_size): Renamed to ...
(match_operand_size): This. Update comments.
(match_simd_size): Update comments. Replace match_reg_size
with match_operand_size.
(match_mem_size): Likewise.
(operand_size_match): Replace match_reg_size with
match_operand_size.
H.J. Lu [Fri, 20 Jul 2018 16:52:35 +0000 (09:52 -0700)]
ld: Discard .rel.* and .rela.* sections
On many targets, like x86, .rel.dyn/.rela.dyn section is combined from
different .rel/.rela sections by -z combreloc. To discard dynamic
relocation section, we need to discard .rel.* and .rela.* sections
instead of .rel.dyn and .rela.dyn sections.
Note: This test fails for targets which issue a warning when dynamic
section is discarded.
* testsuite/ld-elf/reloc-discard.ld: Discard .rel.* and .rela.*
sections instead of .rel.dyn and .rela.dyn sections.
Maciej W. Rozycki [Fri, 20 Jul 2018 16:28:08 +0000 (17:28 +0100)]
MIPS/Linux/native: Convert to `get_ptrace_pid' for PID extraction
Use `get_ptrace_pid' to get the ptrace PID from `inferior_ptid' rather
than extracting it by hand.
gdb/
* mips-linux-nat.c (mips_linux_nat_target::read_description):
Call `get_ptrace_pid' rather than extracting the ptrace PID by
hand.
H.J. Lu [Fri, 20 Jul 2018 16:18:47 +0000 (09:18 -0700)]
x86: Add a GNU_PROPERTY_X86_ISA_1_USED note if needed
When -z separate-code, which is enabled by default for Linux/x86, is
used to create executable, ld won't place any data in the code-only
PT_LOAD segment. If there are no data sections placed before the
code-only PT_LOAD segment, the program headers won't be mapped into
any PT_LOAD segment. When the executable tries to access it (based
on the program header address passed in AT_PHDR), it will lead to
segfault. This patch inserts a GNU_PROPERTY_X86_ISA_1_USED note if
there may be no data sections before the text section so that the
first PT_LOAD segment won't be code-only and will contain the program
header.
Testcases are adjusted to either pass "-z noseparate-code" to ld or
discard the .note.gnu.property section. A Linux/x86 run-time test is
added.
bfd/
PR ld/23428
* elfxx-x86.c (_bfd_x86_elf_link_setup_gnu_properties): If the
separate code program header is needed, make sure that the first
read-only PT_LOAD segment has no code by adding a
GNU_PROPERTY_X86_ISA_1_USED note.
ld/
PR ld/23428
* testsuite/ld-elf/linux-x86.S: New file.
* testsuite/ld-elf/linux-x86.exp: Likewise.
* testsuite/ld-elf/pr23428.c: Likewise.
* testsuite/ld-elf/sec64k.exp: Pass "-z noseparate-code" to ld
for Linux/x86 targets.
* testsuite/ld-i386/abs-iamcu.d: Likewise.
* testsuite/ld-i386/abs.d: Likewise.
* testsuite/ld-i386/pr12718.d: Likewise.
* testsuite/ld-i386/pr12921.d: Likewise.
* testsuite/ld-x86-64/abs-k1om.d: Likewise.
* testsuite/ld-x86-64/abs-l1om.d: Likewise.
* testsuite/ld-x86-64/abs.d: Likewise.
* testsuite/ld-x86-64/pr12718.d: Likewise.
* testsuite/ld-x86-64/pr12921.d: Likewise.
* testsuite/ld-linkonce/zeroeh.ld: Discard .note.gnu.property
section.
* testsuite/ld-scripts/print-memory-usage.t: Likewise.
* testsuite/ld-scripts/size-2.t: Likewise.
* testsuite/lib/ld-lib.exp (run_ld_link_exec_tests): Use ld
to create executable if language is "asm".
Keith Seitz [Sun, 8 Jul 2018 16:19:10 +0000 (10:19 -0600)]
Rename remaining members of buildsym_compunit to start with "m_"
This renames all the remaining members of buildsym_compunit to start
with "m_" to follow the general naming convention.
gdb/ChangeLog
2018-07-20 Keith Seitz <keiths@redhat.com>
* buildsym.h (struct buildsym_compunit) <m_objfile, m_subfiles,
m_main_subfile, m_comp_dir, m_producer, m_debugformat,
m_compunit_symtab, m_language>: Add "m_" prefix.
Update all uses.
* buildsym.c: Update all uses.
Tom Tromey [Wed, 23 May 2018 04:47:15 +0000 (22:47 -0600)]
Remove record_line_ftype
The record_line_ftype typedef was only used in the DWARF reader, and
we removed those uses a few patches ago. So, remove the typedef.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym-legacy.h (record_line): Don't use record_line_ftype.
* buildsym.h (record_line_ftype): Remove typedef.
Tom Tromey [Wed, 23 May 2018 04:44:44 +0000 (22:44 -0600)]
Remove some unused buildsym functions
Now that the DWARF reader uses the builder-based API, we can remove a
few "legacy" functions that were only ever called by it.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym-legacy.h (augment_type_symtab): Don't declare.
(end_expandable_symtab): Likewise.
(end_symtab_get_static_block): Likewise.
(end_symtab_from_static_block): Likewise.
* buildsym-legacy.c (augment_type_symtab): Remove.
(end_expandable_symtab): Remove.
(end_symtab_get_static_block): Remove.
(end_symtab_from_static_block): Remove.
Tom Tromey [Tue, 22 May 2018 21:22:53 +0000 (15:22 -0600)]
Convert the DWARF reader to new-style buildysm
This converts the DWARF reader to use the new-style buildsym API. A
new buildsym_compunit is created for each CU and is used to construct
the symbol table. In some cases the CU must be passed to functions
which previously did not accept it. FWIW I tend to think that most
methods in the DWARF reader ought to be methods on the dwarf2_cu
object.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* dwarf2read.c: Include buildsym.h.
(struct dwarf2_cu) <builder>: New method.
(fixup_go_packaging): Update.
(process_full_comp_unit, process_full_type_unit): Update. Don't
use scoped_free_pendings.
(using_directives): Add "cu" parameter, remove "language".
(read_import_statement, setup_type_unit_groups, )
(read_func_scope, read_lexical_block_scope)
(dwarf2_record_block_ranges, read_namespace): Update.
(lnp_state_machine::lnp_state_machine): Add cu parameter.
(lnp_state_machine::handle_end_sequence): Update.
(class lnp_state_machine) <m_cu>: New member.
<m_record_line_callback>: Remove.
<m_currently_recording_lines>: New member.
(lnp_state_machine::handle_set_file): Update.
(noop_record_line): Remove.
(dwarf_record_line_p): Add cu parameter.
(dwarf_record_line_1, dwarf_finish_line): Likewise.
(lnp_state_machine::record_line)
(lnp_state_machine::lnp_state_machine)
(lnp_state_machine::check_line_address, dwarf_decode_lines_1)
(dwarf_decode_lines): Update.
(dwarf2_start_subfile): Add cu parameter.
(dwarf2_start_symtab, new_symbol): Update.
(macro_start_file, dwarf_decode_macro_bytes): Add cu parameter.
Remove dwarf2_per_objfile parameter.
(dwarf_decode_macros): Update.
Tom Tromey [Tue, 22 May 2018 20:44:09 +0000 (14:44 -0600)]
Parameterize cp_scan_for_anonymous_namespaces
This changes cp_scan_for_anonymous_namespaces to use the
buildsym_compunit API, rather than the function-based API.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* stabsread.c (define_symbol): Update.
* buildsym-legacy.h (get_buildsym_compunit): Declare.
* dwarf2read.c (new_symbol): Update.
* cp-support.h (cp_scan_for_anonymous_namespaces): Update.
* cp-namespace.c: Include buildsym.h.
(cp_scan_for_anonymous_namespaces): Add "compunit" parameter.
* buildsym-legacy.c (get_buildsym_compunit): New function.
Tom Tromey [Tue, 22 May 2018 20:39:24 +0000 (14:39 -0600)]
Introduce buildsym-legacy.h
This introduces a new header, buildsym-legacy.h, and changes all the
symbol readers to use it. The idea is to put the function-based
interface, that relies on the buildsym_compunit global, into a
separate header. Then when a symbol reader is updated to use the new
interface, it can simply not include buildsym-legacy.h, so it's easy
to be sure that the new API is used everywhere.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* xcoffread.c: Include buildsym-legacy.h.
* windows-nat.c: Include buildsym-legacy.h.
* stabsread.c: Include buildsym-legacy.h.
* mdebugread.c: Include buildsym-legacy.h.
* buildsym-legacy.h: New file.
* buildsym-legacy.c: New file, from buildsym.c.
* go32-nat.c: Include buildsym-legacy.h.
* dwarf2read.c: Include buildsym-legacy.h.
* dbxread.c: Include buildsym-legacy.h.
* cp-namespace.c: Include buildsym-legacy.h.
* coffread.c: Include buildsym-legacy.h.
* buildsym.h: Move some contents to buildsym-legacy.h.
* buildsym.c: Include buildsym-legacy.h. Move many functions to
buildsym-legacy.c.
* Makefile.in (HFILES_NO_SRCDIR): Add buildsym-legacy.h.
Tom Tromey [Tue, 22 May 2018 20:27:43 +0000 (14:27 -0600)]
Move struct buildsym_compunit to buildsym.h
This moves struct buildsym_compunit to buildsym.h. Now that the
members are private, and it no longer affects any global state in
buildsym.c, an instance can be used directly for symtab creation.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.h (struct buildsym_compunit): Move from buildsym.c.
* buildsym.c (struct buildsym_compunit): Move to buildsym.h.
(buildsym_compunit::buildsym_compunit)
(buildsym_compunit::~buildsym_compunit)
(buildsym_compunit::get_macro_table): Define.
Tom Tromey [Tue, 22 May 2018 20:23:32 +0000 (14:23 -0600)]
Remove reset_symtab_globals
This patch arranges for the remaining buildsym global --
buildsym_compunit -- to only be cleared by the wrapper functions, not
by methods on struct buildsym_compunit. In the process,
reset_symtab_globals is removed.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.c (reset_symtab_globals): Remove.
(buildsym_compunit::end_symtab_from_static_block): Update.
(buildsym_compunit::augment_type_symtab): Update.
(end_symtab_from_static_block): Call free_buildsym_compunit.
(augment_type_symtab, end_symtab, end_expandable_symtab):
Likewise.
Tom Tromey [Tue, 22 May 2018 20:13:46 +0000 (14:13 -0600)]
Do not use buildsym.h in some files
A few files no longer need to include buildsym.h.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* arch-utils.c: Do not include buildsym.h.
* mipsread.c: Do not include buildsym.h.
* machoread.c: Do not include buildsym.h.
* elfread.c: Do not include buildsym.h.
Tom Tromey [Mon, 21 May 2018 15:29:29 +0000 (09:29 -0600)]
Add many methods to buildsym_compunit
This adds many methods to buildsym_compunit and makes the data members
private. Essentially the entire buildsym API is now available as a
method on buildsym_compunit. However, standalone functions are still
provided, as this is what the sybmol readers actually use.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.c (buildsym_compunit::buildsym_compunit): Do more
initialization.
(buildsym_compunit): Add new constructor.
(struct buildsym_compunit) <get_last_source_file, finish_block,
record_block_range, start_subfile, patch_subfile_names,
push_subfile, pop_subfile, record_line, get_compunit_symtab,
set_last_source_start_addr, get_last_source_start_addr,
get_local_using_directives, set_local_using_directives,
get_global_using_directives, outermost_context_p,
get_current_context_stack, get_context_stack_depth,
get_current_subfile, get_local_symbols, get_file_symbols,
get_global_symbols, record_debugformat, record_producer,
push_context, pop_context, end_symtab_get_static_block,
end_symtab_from_static_block, end_symtab, end_expandable_symtab>:
New public methods.
<record_pending_block, finish_block_internal, make_blockvector,
watch_main_source_file_lossage, end_symtab_with_blockvector>: New
private methods.
Update all users.
Tom Tromey [Thu, 12 Jul 2018 04:55:17 +0000 (22:55 -0600)]
Remove parameter from record_pending_block
This removes a redundant parameter from record_pending_block. It also
moves record_pending_block earlier in the file, so that a forward
declaration is no longer needed.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.c (record_pending_block): Move earlier. Remove objfile
parameter.
(finish_block_internal): Update.
Tom Tromey [Mon, 21 May 2018 15:21:18 +0000 (09:21 -0600)]
Remove EXTERN from buildsym.h
Nothing in buildsym.h relies on the "EXTERN" method of
declaration/definition, so remove the traces.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.h (EXTERN): Don't define or undef.
* buildsym.c (EXTERN): Don't define.
Tom Tromey [Mon, 21 May 2018 15:20:31 +0000 (09:20 -0600)]
Remove a TODO
This removes a TODO comment -- the work has been done.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.c: Remove TODO comment.
Tom Tromey [Mon, 21 May 2018 15:16:16 +0000 (09:16 -0600)]
Remove buildsym_init
Now that buildsym_init does nothing, it can be removed.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* coffread.c (coff_symtab_read): Update.
* xcoffread.c (xcoff_psymtab_to_symtab_1): Update.
(xcoff_new_init): Update.
* mipsread.c (mipscoff_new_init): Update.
* mdebugread.c (mdebug_build_psymtabs): Update.
* elfread.c (elf_new_init): Update.
* dwarf2read.c (process_full_comp_unit, process_full_type_unit):
Update.
* dbxread.c (dbx_new_init, dbx_psymtab_to_symtab_1)
(coffstab_build_psymtabs, elfstab_build_psymtabs)
(stabsect_build_psymtabs): Update.
* buildsym.h (buildsym_init): Don't declare.
* buildsym.c: Update comment.
(prepare_for_building): Remove.
(start_symtab, restart_symtab): Update.
(reset_symtab_globals): Update comment.
(buildsym_init): Remove.
Tom Tromey [Mon, 21 May 2018 15:12:54 +0000 (09:12 -0600)]
Move the symbol lists to buildsym_compunit
This moves the global symbol lists into buildsym_compunit, adds
accessors, and updates all the users.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* xcoffread.c (read_xcoff_symtab, process_xcoff_symbol): Update.
* stabsread.c (patch_block_stabs, define_symbol, read_type)
(read_enum_type, common_block_start, common_block_end)
(cleanup_undefined_types_1, finish_global_stabs): Update.
* mdebugread.c (psymtab_to_symtab_1): Update.
* dwarf2read.c (fixup_go_packaging, read_func_scope)
(read_lexical_block_scope, new_symbol): Update.
* dbxread.c (process_one_symbol): Update.
* coffread.c (coff_symtab_read, process_coff_symbol)
(coff_read_enum_type): Update.
* buildsym.h (file_symbols, global_symbols, local_symbols): Don't
declare.
(get_local_symbols, get_file_symbols, get_global_symbols): New
functions.
* buildsym.c (~buildsym_compunit): Clean up m_file_symbols and
m_global_symbols.
<m_file_symbols, m_local_symbols, m_global_symbols>: New members.
(~scoped_free_pendings): Update.
(finish_block, prepare_for_building, reset_symtab_globals)
(end_symtab_get_static_block, end_symtab_with_blockvector)
(augment_type_symtab, push_context): Update.
(get_local_symbols, get_file_symbols, get_global_symbols): New
functions.
(buildsym_init): Update.
Tom Tromey [Tue, 22 May 2018 05:30:47 +0000 (23:30 -0600)]
Set list_in_scope later in DWARF reader
Currently the DWARF reader sets the list_in_scope member of the CU
when first starting to process symbols. Future changes will make this
assert -- code will not be able to refer to these lists until after
start_symtab has been called. This patch prepares for the problem by
arranging to initialize list_in_scope in dwarf2_start_symtab.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* dwarf2read.c (process_full_comp_unit): Do not set list_in_scope.
(process_full_type_unit): Likewise.
(dwarf2_start_symtab): Set list_in_scope.
Tom Tromey [Tue, 22 May 2018 05:23:02 +0000 (23:23 -0600)]
Do not look at file symbols when reading psymtabs
The DWARF reader was setting the list_in_scope member when reading
partial symbols. however, this member is only useful when reading
full symbols. Future patches will make this assert, so remove these
unneeded initializations.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* dwarf2read.c (process_psymtab_comp_unit_reader)
(build_type_psymtabs_reader): Do not set list_in_scope.
Tom Tromey [Mon, 21 May 2018 14:35:43 +0000 (08:35 -0600)]
Remove free_pendings
buildsym.c currently keeps a free list of "struct pending"s. However,
this didn't seem necessary to me, and so this patch removes the free
list.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.c (free_pendings): Remove.
(add_symbol_to_list, scoped_free_pendings)
(finish_block_internal, buildsym_init): Update.
Tom Tromey [Mon, 21 May 2018 06:41:21 +0000 (00:41 -0600)]
Remove the "listhead" argument from finish_block
finish_block is only ever called with "&local_symbols" as the
"listhead" argument. So, remove this argument.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* xcoffread.c (read_xcoff_symtab): Update.
* dwarf2read.c (read_func_scope, read_lexical_block_scope):
Update.
* dbxread.c (process_one_symbol): Update.
* coffread.c (coff_symtab_read): Update.
* buildsym.h (finish_block): Update.
* buildsym.c (finish_block): Remove "listhead" argument.
(end_symtab_get_static_block): Update.
Tom Tromey [Mon, 21 May 2018 06:34:26 +0000 (00:34 -0600)]
Move pending_blocks and pending_block_obstack to buildsym_compunit
This moves the pending_blocks and pending_block_obstack into
buildsym_compunit.
The obstack could perhaps be merged with the addrmap obstack, but I
did not do that in this series.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.h (class scoped_free_pendings): Remove constructor.
* buildsym.c (struct buildsym_compunit) <free_pending_blocks>: New
method.
<m_pending_block_obstack, m_pending_blocks>: New members.
(pending_block_obstack, pending_blocks): Remove.
(scoped_free_pendings::scoped_free_pendings): Default.
(~scoped_free_pendings): Update.
(free_pending_blocks): Remove.
(finish_block_internal, record_pending_block, make_blockvector)
(end_symtab_get_static_block, augment_type_symtab, push_context)
(buildsym_init): Update.
Tom Tromey [Mon, 21 May 2018 06:28:56 +0000 (00:28 -0600)]
Move pending addrmap globals to buildsym_compunit
This moves the pending addrmap-related globals into buildsym_compunit.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* buildsym.c (struct buildsym_compunit) <m_pending_addrmap,
m_pending_addrmap_obstack, m_pending_addrmap_interesting>: New
members.
(pending_addrmap, pending_addrmap_obstack)
(pending_addrmap_interesting): Remove.
(scoped_free_pendings, record_block_range, make_blockvector)
(prepare_for_building, reset_symtab_globals, buildsym_init):
Update.
Tom Tromey [Mon, 21 May 2018 06:15:12 +0000 (00:15 -0600)]
Move current_subfile to buildsym_compunit
This moves the global current_subfile into buildsym_compunit.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* xcoffread.c (process_linenos): Update.
* stabsread.c (define_symbol, read_type, read_enum_type): Update.
* mdebugread.c (psymtab_to_symtab_1): Update.
* dwarf2read.c (setup_type_unit_groups)
(lnp_state_machine::handle_set_file, dwarf_record_line_p)
(lnp_state_machine::record_line, dwarf_decode_lines): Update.
* dbxread.c (process_one_symbol): Update.
* coffread.c (coff_symtab_read, enter_linenos)
(process_coff_symbol): Update.
* buildsym.h (current_subfile): Don't declare.
(get_current_subfile): Declare.
* buildsym.c (struct buildsym_compunit) <m_current_subfile>: New
member.
(start_subfile, free_buildsym_compunit, push_subfile)
(prepare_for_building, start_symtab): Update.
(get_current_subfile): New function.
Tom Tromey [Mon, 21 May 2018 05:58:35 +0000 (23:58 -0600)]
Move the context stack to buildsym_compunit
This moves the context stack globals to be members of
buildsym_compunit, changing the type to a std::vector in the process.
Because the callers expect the context stack object to be valid after
being popped, at Simon's suggestion I've changed pop_context to return
the object rather than the pointer.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* coffread.c (coff_symtab_read): Update.
* xcoffread.c (read_xcoff_symtab): Update.
* dwarf2read.c (new_symbol): Update.
(read_func_scope, read_lexical_block_scope): Update.
* dbxread.c (process_one_symbol): Update.
* buildsym.h (context_stack, context_stack_depth): Don't declare.
(outermost_context_p): Remove macro.
(outermost_context_p, get_current_context_stack)
(get_context_stack_depth): Declare.
(pop_context): Return struct context_stack.
* buildsym.c (struct buildsym_compunit) <m_context_stack: New
member.
(context_stack_size): Remove.
(INITIAL_CONTEXT_STACK_SIZE): Remove.
(prepare_for_building, end_symtab_get_static_block)
(augment_type_symtab, push_context): Update.
(pop_context): Return struct context_stack.
(outermost_context_p, get_current_context_stack)
(get_context_stack_depth): New functions.
(buildsym_init): Update.
Tom Tromey [Sat, 2 Jun 2018 04:20:23 +0000 (22:20 -0600)]
Make the Rust parser pure
This makes the Rust parser a pure parser and removes all the
(non-constant) globals from rust-exp.y. This seemed like a nice
simplification to me and I think it should probably be applied to all
the parsers. Perhaps it would be good to go even one step farther and
have all parsers derive from parser_state.
Tested on x86-64 Fedora 26.
gdb/ChangeLog
2018-07-20 Tom Tromey <tom@tromey.com>
* rust-exp.y: Now a pure parser. Update all rules.
(%union): Move earlier.
(current_parser, work_obstack): Remove globals.
(rust_parser, ~rust_parser): Update.
(class rust_parser) <copy_name, concat3, crate_name, super_name,
lex_character, lex_number, lex_string, lex_identifier,
rust_lookup_type, convert_params_to_types, convert_ast_to_type,
convert_name, convert_params_to_expression,
convert_ast_to_expression, ast_basic_type, ast_operation,
ast_compound_assignment, rust_op, ast_literal, ast_dliteral,
ast_structop, ast_structop_anonymous, ast_unary, ast_cast,
ast_call_ish, ast_path, ast_string, ast_struct, ast_range,
ast_array_type, ast_slice_type, ast_reference_type,
ast_pointer_type, ast_function_type, ast_tuple_type>: New methods.
(rust_parse): Update.
(rustyyerror, rustyylex): Add parser parameter.
(rust_lex_test_one, rust_lex_int_test, rust_lex_exception_test)
(rust_lex_stringish_test, rust_lex_test_sequence)
(rust_lex_test_trailing_dot, rust_lex_test_completion)
(rust_lex_test_push_back, rust_lex_tests): Update.
Andre Simoes Dias Vieira [Thu, 19 Jul 2018 15:18:28 +0000 (16:18 +0100)]
[PATCH, LD, AArch64] Fix ifunc testisms
This patch fixes some ifunc testisms after H.J. Lu's patch to enable the
use of IFUNC pointers in position dependent code for binutils. See PR
LD/23169 in binutils bugzilla.
The aarch64 ifunc error message test was changed to no longer expect
this error message as this is now an accepted combination. This patch
also disables the executable tests added by H.J. Lu for aarch64, just as
Alan Modra did with his patch, as these tests only seem to work on some
architectures.
ld/ChangeLog:
2018-07-19 Andre Vieira <andre.simoesdiasvieira@arm.com>
* testsuite/ld-aarch64/ifunc-9.d: Remove no longer expected error.
* testsuite/ld-ifunc/ifunc.exp: Disable tests for aarch64.
Tom de Vries [Fri, 20 Jul 2018 14:40:10 +0000 (16:40 +0200)]
[gdb/testsuite] Make noclone conditional in vla-optimized-out.c
Make the noclone attribute on f1 in vla-optimized-out.c conditional. This
makes the test-case by default identical to
gcc/testsuite/gcc.dg/guality/vla-1.c.
Tested on x86_64-linux.
2018-07-20 Tom de Vries <tdevries@suse.de>
* gdb.base/vla-optimized-out.c: Make noclone attribute conditional on
NOCLONE macro.
* gdb.base/vla-optimized-out.exp: Use additional_flags -DNOCLONE.
Nick Clifton [Fri, 20 Jul 2018 14:05:34 +0000 (15:05 +0100)]
Close memory and resource leaks detected by coverity in the binutils directory.
* objcopy.c (add_specific_symbols): Free buffer on exit.
(add_redefine_syms_file): Close file handle on exit.
(copy_object): Close file handle on early exit.
Free buffer on early exit.
Free gaps buffers once they are no longer needed.
* dwarf.c (display_debug_frames): Free allocated memory on exit.
(load_separate_debug_info): Free allocate memory on early exit.
Chenghua Xu [Fri, 20 Jul 2018 12:21:33 +0000 (13:21 +0100)]
MIPS/GAS: Split Loongson MMI Instructions from loongson2f/3a
The MMI instruction set has been implemented in many Loongson
processors. There is a lot of software optimized for MMI. This patch
splits MMI from loongson2f/3a, and adds GAS and disassembler options for
MMI instructions.
2018-07-20 Chenghua Xu <paul.hua.gm@gmail.com>
Maciej W. Rozycki <macro@mips.com>
bfd/
* elfxx-mips.c (print_mips_ases): Add MMI extension.
binutils/
* readelf.c (print_mips_ases): Add MMI extension.
gas/
* NEWS: Mention MultiMedia extensions Instructions (MMI)
support.
* config/tc-mips.c (options): Add OPTION_LOONGSON_MMI and
OPTION_NO_LOONGSON_MMI.
(md_longopts): Likewise.
(mips_ases): Define availability for MMI.
(mips_convert_ase_flags): Map ASE_LOONGSON_MMI to
AFL_ASE_LOONGSON_MMI.
(mips_cpu_info_table): Add ASE_LOONGSON_MMI for loongson2f/3a.
(md_show_usage): Add help for -mloongson-mmi and
-mno-loongson-mmi.
* doc/as.texi: Document -mloongson-mmi, -mno-loongson-mmi.
* doc/c-mips.texi: Document -mloongson-mmi, -mno-loongson-mmi,
.set loongson-mmi and .set noloongson-mmi.
* testsuite/gas/mips/loongson-2f.d: Move mmi test to ...
* testsuite/gas/mips/loongson-2f-mmi.d: Here. Add ISA/ASE
flag verification.
* testsuite/gas/mips/loongson-2f.s: Move mmi test to ...
* testsuite/gas/mips/loongson-2f-mmi.s: Here.
* testsuite/gas/mips/loongson-3a.d: Move mmi test to ...
* testsuite/gas/mips/loongson-3a-mmi.d: Here. Add ISA/ASE
flag verification.
* testsuite/gas/mips/loongson-3a.s: Move mmi test to ...
* testsuite/gas/mips/loongson-3a-mmi.s: Here.
* testsuite/gas/mips/mips.exp: Run loongson-2f-mmi and
loongson-3a-mmi tests.
include/
* elf/mips.h (AFL_ASE_MMI): New macro.
(AFL_ASE_MASK): Update to include AFL_ASE_LOONGSON_MMI.
* opcode/mips.h (ASE_LOONGSON_MMI): New macro.
opcodes/
* mips-dis.c (mips_arch_choices): Add MMI to loongson2f and
loongson3a descriptors.
(parse_mips_ase_option): Handle -M loongson-mmi option.
(print_mips_disassembler_options): Document -M loongson-mmi.
* mips-opc.c (LMMI): New macro.
(mips_opcodes): Replace IL2F|IL3A marking with LMMI for MMI
instructions.
Maciej W. Rozycki [Fri, 20 Jul 2018 12:21:33 +0000 (13:21 +0100)]
MIPS/LD: Fix crashing with a discarded dynamic relocation section
Fix a crash that occurs in `_bfd_mips_elf_finish_dynamic_sections' if a
dynamic relocation section has been created, but marked to be discarded
by an assignment to the /DISCARD/ output section in a linker script.
In that case the output section is the absolute section, which has no
ELF section data attached, so trying to set its `sh_size' parameter
causes a null pointer dereference.
This is only done as the value for the DT_RELSZ dynamic entry is being
set, so fix the problem by not creating DT_REL, DT_RELSZ or DT_RELENT
dynamic entries in the first place if the dynamic relocation section
will not be output, as with no dynamic relocation data present these
would not serve their purpose anyway.
Add a generic ELF test case to verify that no dynamic relocation data is
reported in the dynamic segment.
bfd/
* elfxx-mips.c (_bfd_mips_elf_size_dynamic_sections): Do not
create DT_REL, DT_RELSZ or DT_RELENT dynamic entries if the
dynamic relocation section will be discarded from output.
(_bfd_mips_elf_finish_dynamic_sections) <DT_RELSZ>: Assert that
the dynamic relocation section will be retained in output.
ld/
* testsuite/ld-elf/reloc-discard.d: New test.
* testsuite/ld-elf/reloc-discard.ld: New test linker script.
* testsuite/ld-elf/reloc-discard.s: New test source.
Rainer Orth [Fri, 20 Jul 2018 09:15:57 +0000 (11:15 +0200)]
Specify architecture for SPARC gas tests
A couple of SPARC gas tests FAIL on Solaris/SPARC (both
sparc-sun-solaris2.11 and sparcv9-sun-solaris2.11):
FAIL: sparc PAUSE
FAIL: sparc CBCOND
FAIL: sparc CFR
FAIL: sparc CRYPTO
FAIL: sparc HPC+VIS3
FAIL: sparc IMA
FAIL: sparc OSA2015 %mwait asr and MWAIT instruction
FAIL: sparc OSA2015 %mcdper asr
FAIL: sparc SPARC5 and VIS4.0
FAIL: OSA2015 crypto instructions
FAIL: SPARC6
FAIL: FPCMPSHL OSA2017 instructions
FAIL: OSA2017 ONADD/ONSUB/ONMUL/ONDIV instructions.
FAIL: OSA2017 RLE instructions
FAIL: sparc64 rdasr
FAIL: sparc64 rdpr
FAIL: sparc64 rdhpr
FAIL: sparc64 wrasr
FAIL: sparc64 wrpr
It turns out there's a common pattern here: failures happen for all
tests that use SPARC ISA extensions beyond sparcv9, e.g. for the sparc
PAUSE test:
regexp_diff match failure
regexp "^ 0: b7 80 40 02 wr %g1, %g2, %pause$"
line " 0: b7 80 40 02 wr %g1, %g2, %asr27"
[...]
regexp_diff match failure
regexp "^ 8: b7 80 20 08 pause 8$"
line " 8: b7 80 20 08 wr 8, %asr27"
[...]
The fix is easy, actually: just as the tests specify the ISA extension
to use as a gas flag, the same needs to be done for objdump. For the
test above, which has -Av9v, this means passing -msparc:v9v to objdump.
Doing so makes all but two (unrelated; to be reported separately)
failures go away.
This doesn't happen on Linux/SPARC, where gas emits GNU object attributes
matching the hardcare capabilities used. Since gas doesn't yet implement
Solaris-style object capabilites, the explicit -march is needed, but only
passed on Solaris.
Tested on both sparc-sun-solaris2.11 and sparcv9-sun-solaris2.11.
* testsuite/gas/sparc/sparc.exp (set_tests_arch): New proc.
Prefix v9c, v9d, v9v, v9m, v9m8 tests with corresponding
set_tests_arch.
DJ Delorie [Fri, 20 Jul 2018 01:34:37 +0000 (21:34 -0400)]
Remove myself from target-specific MAINTAINERS
[binutils]
* MAINTAINERS (RL78, RX): Remove myself as maintainer.
[sim]
* MAINTAINERS (rl78, m32c, rx, v850): Remove myself as maintainer.
GDB Administrator [Fri, 20 Jul 2018 00:00:47 +0000 (00:00 +0000)]
Automatic date update in version.in
Pedro Alves [Thu, 19 Jul 2018 15:34:55 +0000 (16:34 +0100)]
Eliminate remaining gdb/guile cleanups
The remaining gdb/guile cleanups all handle the memory returned by
gdbscm_scm_to_c_string.
This commit makes gdbscm_scm_to_c_string return a
gdb::unique_xmalloc_ptr instead of a naked pointer, and eliminates the
remaining cleanups.
gdb/ChangeLog:
2018-07-19 Pedro Alves <palves@redhat.com>
* guile/guile-internal.h (gdbscm_scm_to_c_string): Now returns a
gdb::unique_xmalloc_ptr.
* guile/scm-breakpoint.c (gdbscm_set_breakpoint_condition_x):
Adjust to use dbscm_wrap and gdb::unique_xmalloc_ptr.
* guile/scm-exception.c (gdbscm_exception_message_to_string): Use
copy-initialization.
* guile/scm-pretty-print.c (ppscm_print_children): Use
gdb::unique_xmalloc_ptr instead of cleanups.
(gdbscm_apply_val_pretty_printer): Remove cleanups.
* guile/scm-string.c (gdbscm_scm_to_c_string): Now returns a
gdb::unique_xmalloc_ptr.
* guile/scm-type.c (gdbscm_type_field, gdbscm_type_has_field_p):
Adjust to use gdb::unique_xmalloc_ptr.
* guile/scm-utils.c (extract_arg): Adjust.
* guile/scm-value.c (gdbscm_value_field): Adjust to use
gdb::unique_xmalloc_ptr instead of a cleanup.
Tom Tromey [Thu, 19 Jul 2018 14:49:58 +0000 (08:49 -0600)]
Remove make_cleanup_value_free_to_mark
Pedro's patch to introduce gdbscm_wrap removed the last caller of
make_cleanup_value_free_to_mark. This patch removes this function.
I'm checking this in as obvious. Tested by rebuilding, and by
grepping.
gdb/ChangeLog
2018-07-19 Tom Tromey <tom@tromey.com>
* utils.c (do_value_free_to_mark)
(make_cleanup_value_free_to_mark): Remove.
* utils.h (make_cleanup_value_free_to_mark): Remove.
Tom de Vries [Thu, 19 Jul 2018 14:24:16 +0000 (16:24 +0200)]
[gdb/testsuite] Fix regexp in list-thread-groups-available.exp
I ran into a gdb.mi/list-thread-groups-available.exp failure:
...
Running gdb/testsuite/gdb.mi/list-thread-groups-available.exp ...
FAIL: gdb.mi/list-thread-groups-available.exp:
list available thread groups (unexpected output)
PASS: gdb.mi/list-thread-groups-available.exp:
list available thread groups with filter
...
When doing an experiment of running it 100 times in a row, the failure
reproduced 3 times.
Analyzing the original failure led to insufficient quoting of square brackets
in a regexp. This patch fixes the regexp, which resulted in 0 failures in a
100-in-a-row run.
Tested on x86_64-linux.
2018-07-19 Tom de Vries <tdevries@suse.de>
* gdb.mi/list-thread-groups-available.exp (cores_re): Fix quoting in
regular expression.
Eli Zaretskii [Thu, 19 Jul 2018 13:54:06 +0000 (16:54 +0300)]
Fix GDB build using mingw.org's MinGW
libiberty/
2018-07-18 Eli Zaretskii <eliz@gnu.org>
PR gdb/23434
* libiberty/simple-object-elf.c (ENOTSUP): If not defined by
errno.h, redirect ENOTSUP to ENOSYS.
Nick Clifton [Thu, 19 Jul 2018 13:23:31 +0000 (14:23 +0100)]
Stop potential memory leaks in dwarf CIE reader.
* dwarf.c (read_cie): Free allocated memory before returning with
a failure result.
Pedro Alves [Thu, 19 Jul 2018 09:56:47 +0000 (10:56 +0100)]
gdbscm_wrap, really forward args
Commit
557e56be2648 ("Eliminate most remaining cleanups under
gdb/guile/") missed adding the && to Args to really forward the
arguments properly. Noticed by inspection.
gdb/ChangeLog:
2018-07-19 Pedro Alves <palves@redhat.com>
* guile/guile-internal.h (gdbscm_wrap): Really make 'args' a
forwarding reference.
Andreas Krebbel [Thu, 19 Jul 2018 08:00:51 +0000 (10:00 +0200)]
S/390: Set the htm flag on PPA
The PPA instruction will be emitted by GCC transactional execution
builtins so it needs to be accepted with just -mhtm and without
-march=zEC12.
opcodes/ChangeLog:
2018-07-19 Andreas Krebbel <krebbel@linux.ibm.com>
* s390-opc.txt (PPA): Add the htm flag.
Jan Beulich [Thu, 19 Jul 2018 06:36:19 +0000 (08:36 +0200)]
x86: fold narrowing VCVT* templates
When multiple (here: two) forms of an insn take different width inputs
but produce identical size outputs (here: RegXMM), the templates can be
combined.
Also drop IgnoreSize (and the now redundant size specifiers) wherever
applicable.
Jan Beulich [Thu, 19 Jul 2018 06:35:38 +0000 (08:35 +0200)]
x86: fold VFPCLASSP{D,S} templates
These are special because they may not have a register operand to derive
the vector length from, which requires to also deal with the braodcast
case when determining vector length in build_evex_prefix().
Also drop IgnoreSize (and the now redundant size specifiers) from their
suffixed counterparts.
Jan Beulich [Thu, 19 Jul 2018 06:34:45 +0000 (08:34 +0200)]
x86: fold various AVX512* templates
Jan Beulich [Thu, 19 Jul 2018 06:34:01 +0000 (08:34 +0200)]
x86: fold various AVX512DQ templates
Jan Beulich [Thu, 19 Jul 2018 06:32:17 +0000 (08:32 +0200)]
x86: fold various AVX512BW templates
Jan Beulich [Thu, 19 Jul 2018 06:31:24 +0000 (08:31 +0200)]
x86: fold various AVX512CD templates
Jan Beulich [Thu, 19 Jul 2018 06:29:35 +0000 (08:29 +0200)]
x86: fold various AVX512VL templates into their AVX512F counterparts
Jan Beulich [Thu, 19 Jul 2018 06:28:29 +0000 (08:28 +0200)]
x86: pre-process opcodes table before parsing
Instead of expanding macro-like constructs in i386-gen, have the C pre-
processor do this for us. Besides being a prerequisite for the next
template folding steps, this also paves the way for removing various
hidden dependencies between #define-s in i386-opc.h and plain literal
numbers used in i386-opc.tbl.
The #undef of None is solely to leave the generated i386-tbl.h entirely
unchanged.
GDB Administrator [Thu, 19 Jul 2018 00:00:39 +0000 (00:00 +0000)]
Automatic date update in version.in
Pedro Alves [Wed, 18 Jul 2018 22:41:06 +0000 (23:41 +0100)]
Eliminate cleanup in gdbscm_execute_gdb_command
Note: the "may be modified" comment is no longer true nowadays.
gdb/ChangeLog:
2018-07-18 Pedro Alves <palves@redhat.com>
* guile/guile.c (gdbscm_execute_gdb_command): Adjust to use
gdbscm_wrap. Use gdb::unique_xmalloc_ptr<char> instead of a
cleanup.
Pedro Alves [Wed, 18 Jul 2018 21:55:59 +0000 (22:55 +0100)]
Eliminate most remaining cleanups under gdb/guile/
The main complication with the Guile code is that we have two types of
exceptions to consider. GDB/C++ exceptions, and Guile/SJLJ
exceptions. Code that is facing the Guile interpreter must not throw
GDB exceptions, instead Scheme exceptions must be thrown. Also,
because Guile exceptions are SJLJ based, Guile-facing code must not
use local objects with dtors, unless wrapped in a scope with a
TRY/CATCH, because the dtors won't otherwise be run when a Guile
exceptions is thrown.
This commit adds a new gdbscm_wrap wrapper function than encapsulates
a pattern I noticed in many of the functions using
GDBSCM_HANDLE_GDB_EXCEPTION_WITH_CLEANUPS. The wrapper is written
such that you can pass either a lambda to it, or a function plus a
variable number of forwarded args. I used a lambda when its body
would be reasonably short, and a separate function in the larger
cases.
This also convers a few functions that were using
GDBSCM_HANDLE_GDB_EXCEPTION to use gdbscm_wrap too because they
followed a similar pattern.
A few cases of make_cleanup calls are replaced with explicit xfree
calls. The make_cleanup/do_cleanups calls in those cases are
pointless, because do_cleanups won't be called when a Scheme exception
is thrown.
We also have a couple cases of Guile-facing code using RAII-type
objects to manage memory, but those are incorrect, exactly because
their dtor won't be called if a Guile exception is thrown.
gdb/ChangeLog:
2018-07-18 Pedro Alves <palves@redhat.com>
* guile/guile-internal.h: Add comment about mixing GDB and Scheme
exceptions.
(GDBSCM_HANDLE_GDB_EXCEPTION_WITH_CLEANUPS): Delete.
(gdbscm_wrap): New.
* guile/scm-frame.c (gdbscm_frame_read_register): Use xfree
directly instead of a cleanup.
* guile/scm-math.c (vlscm_unop_gdbthrow): New, factored out from ...
(vlscm_unop): ... this. Reimplement using gdbscm_wrap.
(vlscm_binop_gdbthrow): New, factored out from ...
(vlscm_binop): ... this. Reimplement using gdbscm_wrap.
(vlscm_rich_compare): Use gdbscm_wrap.
* guile/scm-symbol.c (gdbscm_lookup_symbol): Use xfree directly
instead of a cleanup.
(gdbscm_lookup_global_symbol): Use xfree directly instead of a
cleanup.
* guile/scm-type.c (gdbscm_type_field, gdbscm_type_has_field_p):
Use xfree directly instead of a cleanup.
* guile/scm-value.c (gdbscm_make_value, gdbscm_make_lazy_value):
Adjust to use gdbscm_wrap and scoped_value_mark.
(gdbscm_value_optimized_out_p): Adjust to use gdbscm_wrap.
(gdbscm_value_address, gdbscm_value_dereference)
(gdbscm_value_referenced_value): Adjust to use gdbscm_wrap and
scoped_value_mark.
(gdbscm_value_dynamic_type): Use scoped_value_mark.
(vlscm_do_cast, gdbscm_value_field): Adjust to use gdbscm_wrap and
scoped_value_mark.
(gdbscm_value_subscript, gdbscm_value_call): Adjust to use
gdbscm_wrap and scoped_value_mark.
(gdbscm_value_to_string): Use xfree directly instead of a
cleanup. Move 'buffer' unique_ptr to TRY scope.
(gdbscm_value_to_lazy_string): Use xfree directly instead of a
cleanup. Move 'buffer' unique_ptr to TRY scope. Use
scoped_value_mark.
(gdbscm_value_fetch_lazy_x): Use gdbscm_wrap.
(gdbscm_parse_and_eval): Adjust to use gdbscm_wrap and
scoped_value_mark.
(gdbscm_history_ref, gdbscm_history_append_x): Adjust to use
gdbscm_wrap.
Tom de Vries [Tue, 3 Jul 2018 09:06:45 +0000 (11:06 +0200)]
[gdb/exp] Fix printing of type of optimized out vla
Consider this snippet from gcc/testsuite/gcc.dg/guality/vla-1.c:
...
int __attribute__((noinline))
f1 (int i)
{
char a[i + 1];
a[0] = 5; /* { dg-final { gdb-test .+1 "i" "5" } } */
return a[0]; /* { dg-final { gdb-test . "sizeof (a)" "6" } } */
}
...
When we compile the test-case with -O1 -g, and query the size of optimized
out vla 'a', we get:
...
$ ./gdb -batch -ex "b f1" -ex "r" -ex "p sizeof (a)" vla-1.exe
Breakpoint 1 at 0x4004a8: file vla-1.c, line 17.
Breakpoint 1, f1 (i=i@entry=5) at vla-1.c:17
17 return a[0];
$1 = 0
...
while we expect a size of '6'.
The problem is that default_read_var_value does not resolve the dynamic type
of a variable if the variable is optimized out.
This patch fixes that, and consequently gdb prints '6', as expected.
Tested on x86_64-linux.
2018-07-18 Tom de Vries <tdevries@suse.de>
* findvar.c (default_read_var_value): Also resolve dynamic type for
LOC_OPTIMIZED_OUT vars.
* gdb.base/vla-optimized-out.c: New test.
* gdb.base/vla-optimized-out.exp: New file.
Maciej W. Rozycki [Wed, 18 Jul 2018 18:17:39 +0000 (19:17 +0100)]
microMIPS: Add SYSCALL instruction decoding
Fix a bug with commit
4cc0665f24bb ("microMIPS support"),
<https://sourceware.org/ml/gdb-patches/2012-05/msg00724.html>, and add
missing microMIPS SYSCALL instruction decoding needed to determine the
location to put a breakpoint at when single-stepping though a syscall.
gdb/
* mips-tdep.c (micromips_next_pc): Add SYSCALL instruction
decoding.
Maciej W. Rozycki [Wed, 18 Jul 2018 16:51:54 +0000 (17:51 +0100)]
BFD/XCOFF: Fix storage class setting for weak defined symbols
Fix an issue with commit
8602d4fea60d ("Add AIX weak support"),
<https://sourceware.org/ml/binutils/2009-03/msg00189.html>, and use the
correct condition to set the storage class for weak defined symbols.
The context here is as follows:
else if ((h->root.type == bfd_link_hash_defined
|| h->root.type == bfd_link_hash_defweak)
&& h->smclas == XMC_XO)
{
BFD_ASSERT (bfd_is_abs_section (h->root.u.def.section));
isym.n_value = h->root.u.def.value;
isym.n_scnum = N_UNDEF;
if (h->root.type == bfd_link_hash_undefweak
&& C_WEAKEXT == C_AIX_WEAKEXT)
isym.n_sclass = C_WEAKEXT;
else
isym.n_sclass = C_EXT;
aux.x_csect.x_smtyp = XTY_ER;
}
so clearly the inner condition can never be true. Correct the condition
then to check for the `bfd_link_hash_defweak' symbol type instead here,
and in a similar place a little further down in the same function.
bfd/
* xcofflink.c (xcoff_write_global_symbol): Fix symbol type
checks for defined weak symbols.
Maciej W. Rozycki [Wed, 18 Jul 2018 13:08:37 +0000 (14:08 +0100)]
MIPS/GAS/testsuite: Correct whitespace issues with Loongson tests
Remove CR characters, trailing whitespace and space characters appearing
immediately before a tab character, and replace spaces with tabs, all
across Loongson GAS tests.
gas/
* testsuite/gas/mips/loongson-2e.d: Correct whitespace issues.
* testsuite/gas/mips/loongson-2f.d: Likewise.
* testsuite/gas/mips/loongson-2f-2.d: Likewise.
* testsuite/gas/mips/loongson-2f-3.d: Likewise.
* testsuite/gas/mips/loongson-3a.d: Likewise.
* testsuite/gas/mips/loongson-3a-2.d: Likewise.
* testsuite/gas/mips/loongson-2e.s: Likewise.
* testsuite/gas/mips/loongson-2f.s: Likewise.
* testsuite/gas/mips/loongson-2f-3.s: Likewise.
* testsuite/gas/mips/loongson-3a.s: Likewise.
* testsuite/gas/mips/loongson-3a-2.s: Likewise.
H.J. Lu [Wed, 18 Jul 2018 12:33:36 +0000 (05:33 -0700)]
x86: Split vcvtps2{,u}qq and vcvttps2{,u}qq
After
commit
1b54b8d7e4fc8055f9220a5287e8a94d8a65a88d
Author: Jan Beulich <jbeulich@novell.com>
Date: Mon Dec 18 09:36:14 2017 +0100
x86: fold RegXMM/RegYMM/RegZMM into RegSIMD
... qualified by their respective sizes, allowing to drop FirstXmm0 at
the same time.
folded RegXMM, RegYMM and RegZMM into RegSIMD, it's no longer impossible
to distinguish if Xmmword can represent a memory reference when operand
specification contains SIMD register. For example, template operands
specification like these
RegXMM|...|Xmmword|...
and
RegXMM|...
The Xmmword bitfield is always set by RegXMM which is represented by
"RegSIMD|Xmmword". This patch splits each of vcvtps2qq, vcvtps2uqq,
vcvttps2qq and vcvttps2uqq into 2 templates: one template only has
RegXMM source operand and the other only has mempry source operand.
gas/
PR gas/23418
* testsuite/gas/i386/xmmword.s: Add tests for vcvtps2qq,
vcvtps2uqq, vcvttps2qq and vcvttps2uqq.
* testsuite/gas/i386/xmmword.l: Updated.
opcodes/
PR gas/23418
* i386-opc.h (Byte): Update comments.
(Word): Likewise.
(Dword): Likewise.
(Fword): Likewise.
(Qword): Likewise.
(Tbyte): Likewise.
(Xmmword): Likewise.
(Ymmword): Likewise.
(Zmmword): Likewise.
* i386-opc.tbl: Split vcvtps2qq, vcvtps2uqq, vcvttps2qq and
vcvttps2uqq.
* i386-tbl.h: Regenerated.
Andreas Krebbel [Thu, 21 Dec 2017 12:12:03 +0000 (13:12 +0100)]
S/390: Improve partial relro support for 64 bit
Currently on S/390 the .got.plt always comes first which prevents the
GNU_RELRO segment from being extended across the non-plt GOT entries.
Just swapping both unfortunately is not that simple since our ABI
requires the _GLOBAL_OFFSET_TABLE_ symbol to point to the very
beginning of the entire GOT. Of the 3 magic GOT entries the first is
accessed via got pointer while second and third are being accessed via
DT_PLTGOT. In order to keep them together we make DT_PLTGOT to point
to the .got instead of .got.plt. However, this violates an assumption
in the dynamic linker prelink undo code about the GOTPLT entries
starting at DT_PLTGOT + 3. We got rid of this requirement with a
Glibc patch already in version 2.24:
https://sourceware.org/ml/libc-alpha/2016-06/msg01302.html
So the S/390 relro GOT layout will look like this with this patch:
+----------------------------------+
|got[0]: DYNAMIC | <--- _GLOBAL_OFFSET_TABLE_ == DT_PLTGOT .got
|got[1]: link_map parm |
|got[2]: &_dl_runtime_resolve |
+----------------------------------+
| | non-plt GOT entries
| |
| |
+----------------------------------+
| | <--- .gotplt, PLT GOT entries
| |
| |
| |
+----------------------------------+
The patch detects the current layout in size_dynamic_section in order
to deal also with linker scripts not generated by this ld version.
With partial relro enabled we pick a linker script where .got and
.got.plt are swapped which then triggers the rest of the logic.
ld/ChangeLog:
2018-07-18 Andreas Krebbel <krebbel@linux.ibm.com>
* emulparams/elf64_s390.sh: Define GENERATE_RELRO_SCRIPT and
SEPARATE_GOTPLT.
* testsuite/ld-s390/gotreloc_64-relro-1.dd: New test.
* testsuite/ld-s390/gotreloc_64-norelro-1.dd: Renamed from ...
* testsuite/ld-s390/gotreloc_64-1.dd: ... this.
* testsuite/ld-s390/s390.exp: Split the GOT testcase into two.
bfd/ChangeLog:
2018-07-18 Andreas Krebbel <krebbel@linux.ibm.com>
* elf-s390-common.c (s390_gotplt_after_got_p): New function.
(s390_got_pointer): New function.
(s390_got_offset): New function.
(s390_gotplt_offset): New function.
* elf64-s390.c (allocate_dynrelocs): Adjust comment.
(elf_s390_size_dynamic_sections): Move space for magic GOT entries
from .got.plt to .got if necessary and pick the right location for
_GLOBAL_OFFSET_TABLE_.
(elf_s390_relocate_section): Use the wrapper functions from
elf-s390-common.c to deal with both possible layouts (either .got
or .got.plt first).
(elf_s390_finish_dynamic_sections): Likewise.
(elf_s390_finish_dynamic_symbol): Make the location of the GOT
magic entries conditional.
Andreas Krebbel [Mon, 11 Jun 2018 11:23:00 +0000 (13:23 +0200)]
ld: Enable using separate linker script for -z relro
With this patch dedicated linker scripts can be generated for partial
relro triggered by defining GENERATE_RELRO_SCRIPT in the target
specific scripts.
This is necessary for e.g. S/390 where usually the .got.plt comes
first and prevents the relro segment from being extended across the
non-plt GOT entries.
The patch started with the work from Marcin taken from the mwk user
branches. However, the patch needed substantial changes due to the
'separate code' feature which got committed in the meantime.
ld/ChangeLog:
2018-07-18 Andreas Krebbel <krebbel@linux.ibm.com>
Marcin KoĆcielnicki <koriakin@0x04.net>
* emultempl/elf32.em: Add code to pick dedicated linker scripts
for partial relro.
* genscripts.sh: Generate dedicated linker scripts for partial relro.
Nick Clifton [Wed, 18 Jul 2018 07:37:38 +0000 (08:37 +0100)]
Updated French translation for gold and Spanish translation for ld.
ld * po/es.po: Updated Spanish translation.
gold * po/fr.po: Updated French translation.
GDB Administrator [Wed, 18 Jul 2018 00:00:35 +0000 (00:00 +0000)]
Automatic date update in version.in
Tom Tromey [Sun, 27 May 2018 05:34:02 +0000 (23:34 -0600)]
Return unique_xmalloc_ptr from gdbscm_scm_to_string
This changes gdbscm_scm_to_string to return a unique_xmalloc_ptr and
then fixes all the callers. This allows for the removal of some
cleanups.
gdb/ChangeLog
2018-07-17 Tom Tromey <tom@tromey.com>
* guile/scm-param.c (pascm_set_func, pascm_show_func)
(compute_enum_list, pascm_set_param_value_x)
(gdbscm_parameter_value): Update.
* guile/guile-internal.h (gdbscm_scm_to_string): Update.
(gdbscm_scm_to_host_string): Update.
* guile/scm-math.c (vlscm_convert_typed_value_from_scheme):
Update.
* guile/scm-cmd.c (cmdscm_add_completion): Update.
* guile/scm-pretty-print.c (ppscm_print_string_repr): Update.
* guile/scm-string.c (gdbscm_scm_to_string): Return
unique_xmalloc_ptr.
(gdbscm_scm_to_host_string): Likewise.
Tom Tromey [Sun, 27 May 2018 05:26:39 +0000 (23:26 -0600)]
Return unique_xmalloc_ptr from gdbscm_safe_eval_string
This changes gdbscm_safe_eval_string to return a unique_xmalloc_ptr.
This allows for the removal of some cleanups. It also fixes a
potential latent memory leak in gdbscm_set_backtrace.
gdb/ChangeLog
2018-07-17 Tom Tromey <tom@tromey.com>
* guile/guile.c (gdbscm_eval_from_control_command): Update.
* guile/guile-internal.h (gdbscm_safe_eval_string): Update.
* guile/scm-objfile.c (gdbscm_execute_objfile_script): Update.
* guile/scm-safe-call.c (gdbscm_safe_eval_string): Return
unique_xmalloc_ptr.
Tom Tromey [Sun, 27 May 2018 05:21:23 +0000 (23:21 -0600)]
Change gdbscm_exception_message_to_string to return a unique_xmalloc_ptr
This changes gdbscm_exception_message_to_string to return a
unique_xmalloc_ptr, allowing for the removal of some cleanups.
unique_xmalloc_ptr was chosen because at the root of the call chains
is a function from Guile that returns a malloc'd string.
gdb/ChangeLog
2018-07-17 Tom Tromey <tom@tromey.com>
* guile/scm-param.c (pascm_signal_setshow_error): Update.
* guile/guile-internal.h (gdbscm_exception_message_to_string):
Update.
* guile/scm-cmd.c (cmdscm_function): Update.
* guile/scm-pretty-print.c
(ppscm_print_exception_unless_memory_error): Update.
* guile/scm-exception.c (gdbscm_exception_message_to_string):
Return unique_xmalloc_ptr.