commit 3322c5d9a1 ("Remove unneeded explicit .o targets") broke the
build with clang, because -Wno-format-nonliteral was in fact needed.
This patch fixes the problem by introducing
DIAGNOSTIC_IGNORE_FORMAT_NONLITERAL and using it in printcmd.c. This
seems preferable to reverting the patch because now the warning
suppression is more targeted.
gdb/ChangeLog
2018-09-05 Simon Marchi <simon.marchi@ericsson.com>
* printcmd.c (printf_c_string): Use
DIAGNOSTIC_IGNORE_FORMAT_NONLITERAL.
(printf_wide_c_string, printf_pointer, ui_printf): Likewise.
include/ChangeLog
2018-09-05 Simon Marchi <simon.marchi@ericsson.com>
* diagnostics.h (DIAGNOSTIC_IGNORE_FORMAT_NONLITERAL): New macro.
I noticed a couple of unnecessary casts in cli-cmds.c. This patch
removes them.
Tested by rebuilding. I'm checking this in.
gdb/ChangeLog
2018-09-05 Tom Tromey <tom@tromey.com>
* cli/cli-cmds.c (shell_escape, edit_command): Remove cast.
Consider a vla variable 'a' in function f1:
...
<2><1a7>: Abbrev Number: 11 (DW_TAG_variable)
<1a8> DW_AT_description : a
<1aa> DW_AT_abstract_origin: <0x311>
...
with abstract origin 'a':
...
<2><311>: Abbrev Number: 3 (DW_TAG_variable)
<312> DW_AT_name : a
<317> DW_AT_type : <0x325>
...
and inherited abstract vla type:
...
<1><325>: Abbrev Number: 9 (DW_TAG_array_type)
<326> DW_AT_type : <0x33a>
<2><32e>: Abbrev Number: 10 (DW_TAG_subrange_type)
<32f> DW_AT_type : <0x2ea>
<333> DW_AT_upper_bound : 5 byte block: fd 1b 3 0 0
(DW_OP_GNU_variable_value: <0x31b>)
...
where the upper bound refers to this artificial variable D.1922 without location
attribute:
...
<2><31b>: Abbrev Number: 8 (DW_TAG_variable)
<31c> DW_AT_description : (indirect string, offset: 0x39a): D.1922
<320> DW_AT_type : <0x2ea>
<324> DW_AT_artificial : 1
...
Currently, when we execute "p sizeof (a)" in f1, the upper bound is calculated
by evaluating the DW_OP_GNU_variable_value expression referring to D.1922, but
since that die doesn't have a location attribute, we get:
...
value has been optimized out
...
However, there's also artificial variable D.4283 that is sibling of vla
variable 'a', has artificial variable D.1922 as abstract origin, and has a
location attribute:
...
<2><1ae>: Abbrev Number: 12 (DW_TAG_variable)
<1af> DW_AT_description : (indirect string, offset: 0x1f8): D.4283
<1b3> DW_AT_abstract_origin: <0x31b>
<1b7> DW_AT_location : 11 byte block: 75 1 8 20 24 8 20 26 31 1c 9f
(DW_OP_breg5 (rdi):1; DW_OP_const1u: 32;
DW_OP_shl; DW_OP_const1u: 32; DW_OP_shra;
DW_OP_lit1; DW_OP_minus; DW_OP_stack_value)
...
The intended behaviour for DW_OP_GNU_variable_value is to find a die that
refers to D.1922 as abstract origin, has a location attribute and is
'in scope', so the expected behaviour is:
...
$1 = 6
...
The 'in scope' concept can be thought of as variable D.1922 having name
attribute "D.1922", and variable D.4283 inheriting that attribute, resulting
in D.4283 being declared with name "D.1922" alongside vla a in f1, and when we
lookup "DW_OP_GNU_variable_value D.1922", it should work as if we try to find
the value of a variable named "D.1922" on the gdb command line using
"p D.1922", and we should return the value of D.4283.
This patch fixes the case described above, by:
- adding a field abstract_to_concrete to struct dwarf2_per_objfile,
- using that field to keep track of which concrete dies are instances of an
abstract die, and
- using that information when getting the value DW_OP_GNU_variable_value.
Build and reg-tested on x86_64-linux.
2018-09-05 Tom de Vries <tdevries@suse.de>
* dwarf2loc.c (sect_variable_value): Call indirect_synthetic_pointer
with resolve_abstract_p == true.
(indirect_synthetic_pointer): Add resolve_abstract_p parameter,
defaulting to false. Propagate resolve_abstract_p to
dwarf2_fetch_die_loc_sect_off.
* dwarf2loc.h (dwarf2_fetch_die_loc_sect_off): Add resolve_abstract_p
parameter, defaulting to false.
* dwarf2read.c (read_variable): Add variable to abstract_to_concrete.
(dwarf2_fetch_die_loc_sect_off): Add and handle resolve_abstract_p
parameter.
* dwarf2read.h (struct die_info): Forward-declare.
(die_info_ptr): New typedef.
(struct dwarf2_per_objfile): Add abstract_to_concrete field.
* gdb.dwarf2/varval.exp: Add test.
This patch avoids a duplicated error message when an invalid
relocation number is read from an object file in sparc-* ELF targets:
$ strip -g test.o
strip: test.o: unsupported relocation type 0xd7
strip: test.o: unsupported relocation type 0xd7
strip: test.o: bad value
Tested in x86_64-linux-gnu, sparc64-linux-gnu and sparc-linux-gnu
targets.
bfd/ChangeLog:
2018-09-04 Jose E. Marchesi <jose.marchesi@oracle.com>
* elfxx-sparc.c (_bfd_sparc_elf_info_to_howto): Do not issue an
error when an invalid relocation is passed; this is already done
by `_bfd_sparc_elf_info_to_howto_ptr'.
The function `elf64_sparc_slurp_one_reloc_table' in elf64-sparc.c
currently checks that the symbol indexes read in the r_sym fields of
relocations are in range. This is done for both dynamic and
non-dynamic symbols. This avoids subsequent invalid memory accesses.
However, no error is issued to the user.
This patch makes BFD to issue an error when the read symbol index is
out of range, following the same behavior implemented in both the
generic ELF routines and other ELF backends (such as mips64).
Tested in x86_64-linux-gnu, sparc64-linux-gnu, and
--enable-targets=all.
2018-09-04 Jose E. Marchesi <jose.marchesi@oracle.com>
* elf64-sparc.c (elf64_sparc_slurp_one_reloc_table): Issue an
error when an invalid symbol index is retrieved in ELF64_R_SYM of
a relocation seen in an input file.
When we update gnulib using our "update-gnulib.sh" tool, it doesn't
automatically update the list of M4 files present at
gnulib/Makefile.in:aclocal_m4_deps. This patch extends the tool to do
that. It also puts "aclocal_m4_deps" in its own file (a Makefile
fragment), so that it's easier to update it programatically.
Tested by generating the file and diff'ing the results against the
current version of "aclocal_m4_deps".
gdb/ChangeLog:
2018-09-04 Sergio Durigan Junior <sergiodj@redhat.com>
Pedro Alves <palves@redhat.com>
* gnulib/Makefile.in (aclocal_m4_deps): Move to
"aclocal-m4-deps.mk". Include file here.
$(srcdir)/aclocal.m4: Add "configure.ac".
* gnulib/aclocal-m4-deps.mk: New file.
* gnulib/update-gnulib.sh: Automatically update
"aclocal-m4-deps.mk".
gdb's configure script accepts --enable-multi-ice, but the code this
refers to is long gone. This patch removes the option entirely.
gdb/ChangeLog
2018-09-04 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
* configure.ac: Remove multi-ice code.
The ada-exp.o rule no longer needs to pass -Wno-old-style-definition
to the compiler, as this option has no meaning in C++. So, This patch
simplifies the explicit ada-exp.o rule in the Makefile. The rule is
still needed because, according to the comment, ada-exp.c may appear
in the srcdir.
gdb/ChangeLog
2018-09-04 Tom Tromey <tom@tromey.com>
* Makefile.in (GDB_WARN_CFLAGS_NO_DEFS): Remove.
(ada-exp.o): Update.
Makefile.in had special cases to compile printcmd.o and target-float.o
with a different set of warnings. However, this is no longer
required, so this patch removes those rules.
gdb/ChangeLog
2018-09-04 Tom Tromey <tom@tromey.com>
* Makefile.in (printcmd.o, target-float.o): Remove.
(GDB_WARN_CFLAGS_NO_FORMAT): Remove.
This removes an obsolete comment from Makefile.in. This was copied
into gnulib/Makefile.in, so this removes that comment as well.
gdb/ChangeLog
2018-09-04 Tom Tromey <tom@tromey.com>
* gnulib/Makefile.in: Remove obsolete comment.
* Makefile.in: Remove obsolete comment.
This commit adds calls to remote_close and clear_gdb_spawn_id to
gdb.base/batch-exit-status.exp, fixing failures reported by buildbot
on Fedora 28 where gdb_spawn_id not being reset by the previous test
caused default_gdb_spawn to return without spawning.
This commit also changes the test to use detect GDB's exit using
gdb_test_multiple expecting 'eof', rather than using 'wait -i' alone.
This means the testcase won't hang forever on failure as fixed in
gdb.base/quit.exp by commit 15763a09d4 ("Fix 'gdb.base/quit.exp
hangs forever' if the test fails").
gdb/testsuite/ChangeLog:
* gdb.base/batch-exit-status.exp: Use gdb_test_multiple and expect
'eof' before 'wait -i'. Use remote_close and clear_gdb_spawn_id.
Sparc V8 does not allow fpop2 instructions (floating point
comparisons) immediately before floating point branches. From the
SPARC Architecture Manual Version 8, section B.22 "Branch on
Floating-point Condition Codes Instructions":
"If the instruction executed immediately before an FBfcc is an FPop2
instruction, the result of the FBfcc is undefined. Therefore, at
least one non FPop2 instruction should be executed between the FPop2
instruction and the FBfcc instruction."
The existing check in GAS, however, does not allow any kind of
floating point instruction before the branch. This patch adds an
extra condition to only disallow fpop2 instructions.
gas/ChangeLog:
2018-09-04 Daniel Cederman <cederman@gaisler.com>
* config/tc-sparc.c (md_assemble): Allow non-fpop2 instructions
before floating point branches for Sparc V8 and earlier.
* testsuite/gas/sparc/sparc.exp: Execute the new test.
* testsuite/gas/sparc/v8branch.d: New test.
* testsuite/gas/sparc/v8branch.s: New test.
This patch fixes an ARI violation in riscv-tdep.c (line ends with
'+').
gdb/ChangeLog:
* riscv-tdep.c (riscv_frame_cache): Fix ARI warning, don't end a
line with '+'.
PR gas/23570
* config/tc-avr.c (md_pseudo_table): Add entry for "secction".
(avr_set_section): New function. Ensures that the .noinit section
gets the NOBITS ELF section type.
Collects information during the prologue scan and uses this to unwind
registers when no DWARF information is available.
This patch has been tested by disabling the DWARF stack unwinders, and
running the complete GDB testsuite against a range of RISC-V targets.
The results are comparable to running with the DWARF unwinders in
place.
gdb/ChangeLog:
* riscv-tdep.c: Add 'prologue-value.h' include.
(struct riscv_unwind_cache): New struct.
(riscv_debug_unwinder): New global.
(riscv_scan_prologue): Update arguments, capture register details
from prologue scan.
(riscv_skip_prologue): Reformat arguments line, move end of
prologue calculation into riscv_scan_prologue.
(riscv_frame_cache): Update return type, create
riscv_unwind_cache, scan the prologue, and fill in remaining cache
details.
(riscv_frame_this_id): Use frame id computed in riscv_frame_cache.
(riscv_frame_prev_register): Use the trad_frame within the
riscv_unwind_cache.
(_initialize_riscv_tdep): Add 'set/show debug riscv unwinder'
flag.
Adds two new functions to the trad-frame API and update the internals
of trad-frame to use the new functions. These functions will be used
in later commits.
gdb/ChangeLog:
* trad-frame.h (trad_frame_set_realreg): Declare.
(trad_frame_set_addr): Declare.
* trad-frame.c (trad_frame_set_realreg): Define new function.
(trad_frame_set_addr): Define new function.
(trad_frame_set_reg_realreg): Use new function.
(trad_frame_set_reg_addr): Use new function.
VMA of the first section in the segment containing the ELF file header
(and possibly section headers too) can't be used to reliably find the
size of the headers plus padding. What's really needed is sh_offset
of the first section assuming it has contents (vma does have a
relationship to sh_offset, but is only guaranteed in demand paged
executables).
If the first section is SHT_NOBITS and it hasn't been converted to
have file contents by the existence of a following SHT_PROGBITS
section in the same segment, the sh_offset value also isn't reliable.
PR 23595
elf.c (copy_elf_program_header): When first segment contains
only the headers and SHT_NOBITS sections, use segment p_filesz
to calculate header and padding size. Use filepos of the first
section otherwise.
Fixes pr23591 test failures on hppa64-hpux and score-elf, and xfails
frv-linux and lm32-linux.
PR ld/23591
* testsuite/ld-elf/pr23591a.s,
* testsuite/ld-elf/pr23591b.s,
* testsuite/ld-elf/pr23591c.s: Don't start directives in first column.
* testsuite/ld-elf/pr23591.d: xfail frv-linux and lm32-linux.
Allow __start___sancov_cntrs as a local symbol.
This patch fixes two violations of the ARI (use of ATTRIBUTE_UNUSED and
"%ll").
gdb/ChangeLog
* compile/compile-cplus-types.c (compile_cplus_debug_output_1): Use
pulongest instead of "%lld".
* compile/compile-cplus-symbols.c (gcc_cplus_convert_symbol): Remove
ATTRIBUTE_UNUSED.
An IR object may have an unknown architecture. But it is compatible
with other architecture.
PR ld/23600
* archures.c (bfd_arch_get_compatible): Allow an IR object with
unknown architecture.
gdb represents a DW_TAG_variant_part as a union. While normally DWARF
would not set the size of a DW_TAG_variant_part, gdb's representation
requires the TYPE_LENGTH to be set.
This patch arranges to set the TYPE_LENGTH of a variant part if it has
not already been set.
This fixes some Rust regressions when testing against a version of
rustc that emits DW_TAG_variant_part.
gdb/ChangeLog
2018-08-31 Tom Tromey <tom@tromey.com>
* dwarf2read.c (dwarf2_add_field): Set the TYPE_LENGTH of the
variant part type.
I noticed that gdb.rust/simple.rs had two local variables named "v".
This didn't previous cause problems, but with a newer rust compiler
this resulted in a test failure. (It should have failed all along, so
I suppose earlier passes were due to a compiler bug.)
This patch renames the second variable.
gdb/testsuite/ChangeLog
2018-08-31 Tom Tromey <tom@tromey.com>
* gdb.rust/simple.rs: Rename second variable "v".
The previous commit included a stale gdbarch.h from an earlier version
of that patch by mistake.
gdb/ChangeLog:
2018-08-31 Pedro Alves <palves@redhat.com>
* gdbarch.h: Regenerate.
target_have_continuable_watchpoint isn't used anywhere so remove it.
The property isn't necessary because checking for "continuable" is the
same as checking for "!steppable && !non-steppable".
gdb/ChangeLog:
2018-08-31 Pedro Alves <palves@redhat.com>
* nto-procfs.c (nto_procfs_target::have_continuable_watchpoint):
Delete.
* s390-linux-nat.c
(s390_linux_nat_target::have_continuable_watchpoint): Delete.
* target.h (target_ops::have_continuable_watchpoint): Delete.
(target_have_continuable_watchpoint): Delete.
* x86-nat.h (x86_nat_target::have_continuable_watchpoint): Delete.
* target-delegates.c: Regenerate.
We will get an XPASS when h8300 port peculiarities are fixed, and will
fix the testsuite too.
* testsuite/gas/elf/section14.d: Change skip to xfail.
* testsuite/lib/gas-defs.exp (run_dump_test): Add xfail support.
When there are more than one input sections with the same section name,
SECNAME, linker picks the first one to define __start_SECNAME and
__stop_SECNAME symbols. When the first input section is removed by
comdat group, we need to check if there is still an output section
with section name SECNAME.
PR ld/23591
* ldlang.c (undef_start_stop): Lookup section in output with
the same name.
* testsuite/ld-elf/pr23591.d: New file.
* testsuite/ld-elf/pr23591a.s: Likewise.
* testsuite/ld-elf/pr23591b.s: Likewise.
* testsuite/ld-elf/pr23591c.s: Likewise.
It was pointed by Pedro that gnulib/Makefile.in should be updated
accordingly after our local gnulib is also updated. The specific part
that needs to be refreshed is the "aclocal_m4_deps" variable, which
lists the .m4 files present under the "gnulib/import/m4/" directory.
This patch does that.
No regressions introduced.
gdb/ChangeLog:
2018-08-31 Sergio Durigan Junior <sergiodj@redhat.com>
* gnulib/Makefile.in (aclocal_m4_deps): Update according to
the files present in "gnulib/import/m4/".
This patch uses the newly defined high-part REL16 relocs to emit
relocations on the notoc stubs as we already do for other stubs.
* elf64-ppc.c (num_relocs_for_offset): New function.
(emit_relocs_for_offset): New function.
(use_global_in_relocs): New function, split out from..
(ppc_build_one_stub): ..here. Output relocations for notoc stubs.
(ppc_size_one_stub): Calculate reloc count for notoc stubs.
(ppc64_elf_size_stubs): Don't count undefined syms in stub_globals.
This patch rearranges ppc_size_one_stub to make it a little easier to
compare against ppc_build_one_stub, and makes a few other random
changes that might help for future maintenance. There should be no
functional changes here.
The patch also fixes code examples in comments. A couple of "ori"
instructions lacked the source register operand, and "@high" is the
correct reloc modifier to use in a sequence building a 64-bit value.
(@hi reports overflow of a 32-bit signed value.)
* elf64-ppc.c: Correct _notoc stub comments.
(ppc_build_one_stub): Simplify output of branch for notoc
long branch stub. Don't include label offset of 8 bytes in
"off" calculation for notoc plt stub. Don't emit insns to get pc.
(build_offset): Emit insns to get pc here instead.
(size_offset): Add 4 extra insns.
(plt_stub_size): Adjust for "off" and size_offset changes.
(ppc_size_one_stub): Rearrange code into a switch, duplicating
some to better match ppc_build_one_stub.
Pass -z separate-code to ld for property-x86-4a tests to generate
the expected .note.gnu.property section. Also run i386 tests for
x86_64-*-elf* targets.
* testsuite/ld-i386/i386.exp: Also run for x86_64-*-elf* targets.
* testsuite/ld-i386/property-x86-4a.d: Pass -z separate-code to
ld.
* testsuite/ld-x86-64/property-x86-4a-x32.d: Likewise.
* testsuite/ld-x86-64/property-x86-4a.d: Likewise.
Since x86 assembler may generate .note.gnu.property section and S-records
can't handle .note.gnu.property sections, add -mx86-used-note=no to
S-records tests on ELF/x86 targets to avoid .note.gnu.property section.
* testsuite/ld-srec/srec.exp: Append -Wa,-mx86-used-note=no to
CC and CXX on ELF/x86 targets.
Since x86 assembler may generate .note.gnu.property section, pass
-mx86-used-note=no to assembler for section2 test on ELF/x86 targets
to avoid .note.gnu.property section.
* testsuite/gas/elf/elf.exp: Pass -mx86-used-note=no to
assembler for section2 test on ELF/x86 targets.
Since cfi/cfi-label.d is ELF/x86 specific, we can pass -mx86-used-note=no
to assembler.
* testsuite/gas/cfi/cfi-label.d: Pass -mx86-used-note=no to
assembler.
Since x86 assembler may generate .note.gnu.property section, adjust
binutils tests to handle it.
* testsuite/binutils-all/readelf.s-64: Adjusted for
.note.gnu.property section.
* testsuite/binutils-all/strip-3.d: Also strip
.note.gnu.property section.
Add -mx86-used-note=[yes|no] option to generate (or not) GNU property
notes with GNU_PROPERTY_X86_FEATURE_2_USED and GNU_PROPERTY_X86_ISA_1_USED
properties. If the assembly input contains no instructions, set the
GNU_PROPERTY_X86_UINT32_VALID bit in GNU_PROPERTY_X86_FEATURE_2_USED
property. Add a --enable-x86-used-note configure time option to set the
default behavior. Set the default if the configure option is not used
to "no".
* NEWS: Mention -mx86-used-note=[no|yes].
* configure.ac: Add --enable-x86-used-note. Define
DEFAULT_X86_USED_NOTE.
* config.in: Regenerated.
* configure: Likewise.
* config/tc-i386.c (x86_isa_1_used): New.
(x86_feature_2_used): Likewise.
(x86_used_note): Likewise.
(_i386_insn): Add has_regmmx, has_regxmm, has_regymm and
has_regzmm.
(build_modrm_byte): Set i.has_regmmx, i.has_regzmm.
i.has_regymm and i.has_regxmm.
(x86_cleanup): New function.
(output_insn): Update x86_isa_1_used and x86_feature_2_used.
(OPTION_X86_USED_NOTE): New.
(md_longopts): Add -mx86-used-note=.
(md_parse_option): Handle OPTION_X86_USED_NOTE.
(md_show_usage): Display -mx86-used-note=.
* config/tc-i386.h (x86_cleanup): New prototype.
(md_cleanup): New.
* doc/c-i386.texi: Document -mx86-used-note=.
Simon Marchi
GDB Administrator
Tom Tromey
Tom de Vries
Joel Brobecker
Jose E. Marchesi
Sergio Durigan Junior
Gary Benson
Daniel Cederman
Andrew Burgess
Nick Clifton
Jozef Lawrynowicz
Alan Modra
Keith Seitz
H.J. Lu
Jim Wilson
Pedro Alves