There are extensions to 386 psABI:
https://groups.google.com/forum/#!topic/ia32-abi/awsRSvJOJfs
to call tls_get_addr via GOT:
call *___tls_get_addr@GOT(%reg)
where EBX register isn't required as GOT base.
Since direct call is 4-byte long and indirect call, is 5-byte long, the
extra one byte must be handled properly.
For general dynamic model, 7-byte lea instruction before call
instruction is replaced by 6-byte one to make room for indirect call.
For local dynamic model, we simply use 5-byte indirect call.
TLS linker optimization is updated to recognize new instruction
patterns. For local dynamic model to local exec model transition,
we generate a 6-byte lea instruction as nop, instead of a 1-byte nop
plus a 4-byte lea instruction.
PR gold/20308
* i386.cc (Target_i386::Relocate::relocate): Allow
R_386_GOT32X relocation against ___tls_get_addr.
(Target_i386::Relocate::tls_gd_to_ie): Support indirect
call to __tls_get_addr.
(Target_i386::Relocate::tls_gd_to_le): Likewise.
(Target_i386::Relocate::tls_ld_to_le): Likewise.
* testsuite/Makefile.am (check_PROGRAMS): Add pr20308a_test,
pr20308b_test, pr20308c_test, pr20308d_test, pr20308e_test.
(pr20308a_test_SOURCES): New.
(pr20308a_test_DEPENDENCIES): Likewise.
(pr20308a_test_CFLAGS): Likewise.
(pr20308a_test_LDFLAGS): Likewise.
(pr20308a_test_LDADD): Likewise.
(pr20308b_test_SOURCES): Likewise.
(pr20308b_test_DEPENDENCIES): Likewise.
(pr20308b_test_CFLAGS): Likewise.
(pr20308b_test_LDFLAGS): Likewise.
(pr20308b_test_LDADD): Likewise.
(pr20308c_test_SOURCES): Likewise.
(pr20308c_test_DEPENDENCIES): Likewise.
(pr20308c_test_CFLAGS): Likewise.
(pr20308c_test_LDFLAGS): Likewise.
(pr20308c_test_LDADD): Likewise.
(pr20308d_test_SOURCES): Likewise.
(pr20308d_test_DEPENDENCIES): Likewise.
(pr20308d_test_CFLAGS): Likewise.
(pr20308d_test_LDFLAGS): Likewise.
(pr20308d_test_LDADD): Likewise.
(pr20308e_test_SOURCES): Likewise.
(pr20308e_test_DEPENDENCIES): Likewise.
(pr20308e_test_CFLAGS): Likewise.
(pr20308e_test_LDFLAGS): Likewise.
(pr20308e_test_LDADD): Likewise.
(pr20308a.so): Likewise.
(pr20308b.so): Likewise.
(pr20308_gd.o): Likewise.
(pr20308_ld.o): Likewise.
(MOSTLYCLEANFILES): Add pr20308a.so pr20308b.so.
* testsuite/Makefile.in: Regenerated.
* testsuite/pr20308_def.c: New file.
* testsuite/pr20308_gd.S: Likewise.
* testsuite/pr20308_ld.S: Likewise.
* testsuite/pr20308_main.c: Likewise.
There are extensions to x86-64 psABI:
https://groups.google.com/forum/#!topic/x86-64-abi/de5_KnLHxtI
to call tls_get_addr via GOT:
call *__tls_get_addr@GOTPCREL(%rip)
Since direct call is 4-byte long and indirect call, is 5-byte long, the
extra one byte must be handled properly.
For general dynamic model, one 0x66 prefix before call instruction is
removed to make room for indirect call. For local dynamic model, we
simply use 5-byte indirect call.
TLS linker optimization is updated to recognize new instruction
patterns. For local dynamic model to local exec model transition, we
generate 4 0x66 prefixes, instead of 3, before mov instruction in 64-bit
and generate a 5-byte nop, instead of 4-byte, before mov instruction in
32-bit.
PR gold/20216
* configure.ac (DEFAULT_TARGET_X86_64_OR_X32): New
AM_CONDITIONAL.
* configure: Regenerated.
* x86_64.cc (Target_x86_64<size>::Relocate::relocate): Allow
R_X86_64_GOTPCRELX relocation against __tls_get_addr.
(Target_x86_64<size>::Relocate::tls_gd_to_ie): Support indirect
call to __tls_get_addr.
(Target_x86_64<size>::Relocate::tls_gd_to_le): Likewise.
(Target_x86_64<size>::Relocate::tls_ld_to_le): Likewise.
* testsuite/Makefile.am (check_PROGRAMS): Add pr20216a_test,
pr20216b_test, pr20216c_test, pr20216d_test, pr20216e_test.
(pr20216a_test_SOURCES): New.
(pr20216a_test_DEPENDENCIES): Likewise.
(pr20216a_test_CFLAGS): Likewise.
(pr20216a_test_LDFLAGS): Likewise.
(pr20216a_test_LDADD): Likewise.
(pr20216b_test_SOURCES): Likewise.
(pr20216b_test_DEPENDENCIES): Likewise.
(pr20216b_test_CFLAGS): Likewise.
(pr20216b_test_LDFLAGS): Likewise.
(pr20216b_test_LDADD): Likewise.
(pr20216c_test_SOURCES): Likewise.
(pr20216c_test_DEPENDENCIES): Likewise.
(pr20216c_test_CFLAGS): Likewise.
(pr20216c_test_LDFLAGS): Likewise.
(pr20216c_test_LDADD): Likewise.
(pr20216d_test_SOURCES): Likewise.
(pr20216d_test_DEPENDENCIES): Likewise.
(pr20216d_test_CFLAGS): Likewise.
(pr20216d_test_LDFLAGS): Likewise.
(pr20216d_test_LDADD): Likewise.
(pr20216e_test_SOURCES): Likewise.
(pr20216e_test_DEPENDENCIES): Likewise.
(pr20216e_test_CFLAGS): Likewise.
(pr20216e_test_LDFLAGS): Likewise.
(pr20216e_test_LDADD): Likewise.
(pr20216a.so): Likewise.
(pr20216b.so): Likewise.
(pr20216_gd.o): Likewise.
(pr20216_ld.o): Likewise.
(MOSTLYCLEANFILES): Add pr20216a.so pr20216b.so.
* testsuite/Makefile.in: Regenerated.
* testsuite/pr20216_def.c: New file.
* testsuite/pr20216_gd.S: Likewise.
* testsuite/pr20216_ld.S: Likewise.
* testsuite/pr20216_main.c: Likewise.
Building gdb with --enable-build-with-cxx=no trips on a warning:
../../binutils-gdb/gdb/rust-lang.c:173:15: error: saveptr may be used
uninitialized in this function [-Werror=maybe-uninitialized]
ret.name = concat (TYPE_NAME (type), "::", token, (char *) NULL);
The problem is that gcc doesn't understand that "tail" can never be
NULL in the call to strtok_r:
name = xstrdup (TYPE_FIELD_NAME (type, 0));
cleanup = make_cleanup (xfree, name);
tail = name + strlen (RUST_ENUM_PREFIX);
...
for (token = strtok_r (tail, "$", &saveptr);
Fix this by always initializing saveptr.
2016-06-29 Manish Goregaokar <manish@mozilla.com>
gdb/ChangeLog:
* rust-lang.c (rust_get_disr_info): Initialize saveptr to NULL.
Currently, we use 123456789 as unknown or illegal syscall number, and
expect program return ENOSYS. Although 123456789 is an illegal syscall
number on arm linux, kernel sends SIGILL rather than returns -ENOSYS.
However, arm linux kernel returns -ENOSYS if syscall number is within
0xf0001..0xf07ff, so we can use 0xf07ff for unknown_syscall in test.
gdb/testsuite:
2016-06-29 Yao Qi <yao.qi@linaro.org>
* gdb.base/catch-syscall.c [__arm__]: Set unknown_syscall to
0x0f07ff.
bfd * elfnn-aarch64.c (is_aarch64_mapping_symbol): New function.
Returns TRUE for AArch64 mapping symbols.
(elfNN_aarch64_backend_symbol_processing): New function. Marks
mapping symbols as precious in object files so that they will not
be stripped.
(elf_backend_symbol_processing): Define.
* elf32-arm.c (is_arm_mapping_symbol): New function. Returns TRUE
for ARM mapping symbols.
(elf32_arm_backend_symbol_processing): Make use of the new function.
Remove an internal diagnostic regression introduced with the inclusion of
"libbfd.h" from write.c, added with:
commit e7ff5c732e
Author: Alan Modra <amodra@gmail.com>
Date: Fri Feb 16 03:40:17 2007 +0000
That change made "libbfd.h" override the `abort' definition provided by
"as.h" earlier on, making the message produced by any calls reached from
write.c, which is a part of the GAS proper, look like they came from
BFD, e.g.:
.../gas/testsuite/gas/elf/type.s: Assembler messages:
.../gas/testsuite/gas/elf/type.s:30: Error: symbol type "gnu_unique_object" is supported only by GNU targets
../as-new: BFD (GNU Binutils) 2.26.51.20160628 internal error, aborting at .../gas/write.c:608 in size_seg
../as-new: Please report this bug.
vs:
.../gas/testsuite/gas/elf/type.s: Assembler messages:
.../gas/testsuite/gas/elf/type.s:30: Error: symbol type "gnu_unique_object" is supported only by GNU targets
.../gas/testsuite/gas/elf/type.s: Internal error, aborting at .../gas/write.c:602 in size_seg
Please report this bug.
With the removal of "libbfd.h" restore the latter message format.
gas/
* write.c: Remove "libbfd.h" inclusion.
PowerPC gcc aligns char arrays to a word which results in the test
failing since it expects a char alignment. As the test already uses
gcc attributes to assign variables to sections, we may as well use an
attribute to align too.
* testsuite/plugin_layout_with_alignment.c: Explicitly align all
variables.
A target like powerpc64 that is PIC by default doesn't need copy relocs.
* testsuite/Makefile.am (copy_test_protected): Disable for powerpc.
* testsuite/Makefile.in: Regenerate.
According to "ELF for the ARM(r) 64-bit Architecture (AArch64)",
this relocation can be used "to prevent removal of sections that
might otherwise appear to be unused."
gold/
* aarch64-reloc.def (NONE): New relocation.
* aarch64.cc (Target_aarch64::Scan::local): Handle R_AARCH64_NONE.
(Target_aarch64::Scan::global): Likewise.
* testsuite/Makefile.am (aarch64_reloc_none): New test.
* testsuite/Makefile.in: Regenerate.
* testsuite/aarch64_reloc_none.s: New test source file.
* testsuite/aarch64_reloc_none.sh: New test script.
Please see patch discussion:
https://www.sourceware.org/ml/binutils/2016-05/msg00322.html
2016-06-28 Sriraman Tallam <tmsriram@google.com>
* x86_64.cc (Lazy_view): New class.
(can_convert_mov_to_lea): Templatize function. Make the function
check for appropriate relocation types and use the view parameter
to get section contents.
(can_convert_callq_to_direct): New function.
(Target_x86_64<size>::Scan::global): Refactor.
(Target_x86_64<size>::Relocate::relocate): Refactor. Change any indirect
call via GOT that can be converted.
* testsuite/Makefile.am (x86_64_indirect_call_to_direct.sh): New test.
* testsuite/Makefile.in: Regenerate.
* testsuite/x86_64_indirect_call_to_direct1.s: New file.
* testsuite/x86_64_indirect_jump_to_direct1.s: New file.
In AArch64 displaced stepping and fast tracepoint, GDB/GDBserver needs
to check whether the offset can fit in the range. We are using int32_t
for offset, it is sufficient to get an offset from an instruction, but
it is not enough to get an offset from two addresses. For example,
we have a BL in shared lib which is at 0x0000002000040774, and the
scratch pad for displaced stepping is at 0x400698. The offset can't
fit in 28 bit imm. However, since we are using int32_t for offset, GDB
thinks the offset can fit it, and generate the B instruction with wrong
offset.
It fixes the following fail,
-FAIL: gdb.base/dso2dso.exp: next over call to sub2
gdb:
2016-06-28 Yao Qi <yao.qi@linaro.org>
* aarch64-tdep.c (aarch64_displaced_step_b): Use int64_t for
variable new_offset.
gdb/gdbserver:
2016-06-28 Yao Qi <yao.qi@linaro.org>
* linux-aarch64-low.c (aarch64_ftrace_insn_reloc_b): Use int64_t
for variable new_offset.
(aarch64_ftrace_insn_reloc_b_cond): Likewise.
(aarch64_ftrace_insn_reloc_cb): Likewise.
(aarch64_ftrace_insn_reloc_tb): Likewise.
(aarch64_install_fast_tracepoint_jump_pad): Likewise. Use
PRIx64 instead of PRIx32.
No need to check version if symbol is unreferenced and undefined.
bfd/
PR ld/20306
* elflink.c (elf_link_check_versioned_symbol): Return false
for unreferenced undefined symbol.
ld/testsuite/
* testsuite/ld-gc/gc.exp: Run pr20306 test.
* ld-gc/pr20306.c: New file.
* ld-gc/pr20306.d: Likewise.
bfin * elf32-bfin.c (bfin_adjust_dynamic_symbol): Fail if a COPY reloc
is needed.
ld * testsuite/ld-elf/comm-data.exp: Expect comm-data2 test to fail
for bfin.
* testsuite/ld-elf/elf.exp: Expect pr14170 and symbolic function
tests to fail for bfin.
* testsuite/ld-elf/endsym.d: Expect to fail with cr16, crx, dlx,
nds32 and visium.
* testsuite/ld-elf/var1.d: Expect to fail with d30v, dlx, ft32 and
microblaze.
* testsuite/ld-pe/pe.exp: Expect foreign symbol test to fail for
mcore-pe.
Complement commit a43942db49 ("LD/ELF: Unify STB_GNU_UNIQUE handling")
and use `supports_gnu_unique' with the `unique_symbol' and `type' tests,
fixing failures like:
.../binutils/testsuite/binutils-all/unique.s: Assembler messages:
.../binutils/testsuite/binutils-all/unique.s:2: Error: symbol type "gnu_unique_object" is supported only by GNU targets
ERROR: .../binutils/testsuite/binutils-all/unique.s: assembly failed
UNRESOLVED: ar unique symbol in archive
.../binutils/ar -s -r -c tmpdir/artest.a tmpdir/unique.o
Executing on host: .../binutils/ar -s -r -c tmpdir/artest.a tmpdir/unique.o (timeout = 300)
.../binutils/ar: tmpdir/unique.o: No such file or directory
FAIL: ar unique symbol in archive
and:
.../gas/testsuite/gas/elf/type.s: Assembler messages:
.../gas/testsuite/gas/elf/type.s:30: Error: symbol type "gnu_unique_object" is supported only by GNU targets
../as-new: BFD (GNU Binutils) 2.26.51.20160628 internal error, aborting at .../gas/write.c:608 in size_seg
../as-new: Please report this bug.
.../gas/testsuite/../../binutils/readelf -s dump.o | grep "1 *\[FIONTCU\]" > dump.out
Executing on host: sh -c {.../gas/testsuite/../../binutils/readelf -s dump.o >readelf.out 2>gas.stderr} /dev/null (timeout = 300)
readelf: Error: dump.o: Failed to read file's magic number
FAIL: elf type list
on MIPS/FreeBSD targets:
mips-freebsd -FAIL: ar unique symbol in archive
mips-freebsd -FAIL: elf type list
mips64-freebsd -FAIL: ar unique symbol in archive
mips64-freebsd -FAIL: elf type list
mips64el-freebsd -FAIL: ar unique symbol in archive
mips64el-freebsd -FAIL: elf type list
mipsel-freebsd -FAIL: ar unique symbol in archive
mipsel-freebsd -FAIL: elf type list
binutils/
* testsuite/binutils-all/ar.exp: Use `supports_gnu_unique' with
the `unique_symbol' test.
gas/
* testsuite/gas/elf/elf.exp: Use `supports_gnu_unique' with the
`type' test.
* elf32-arm.c (elf32_arm_backend_symbol_processing): New
function. Marks mapping symbols in object files as precious, so
that strip will not remove them.
(elf_backend_symbol_processing): Define.
ld * testsuite/ld-elf/merge.d: Add m68hc11 to list of targets that
expect to fail this test.
* testsuite/ld-scripts/overlay-size.d: Skip the entire test for
RX.
* testsuite/ld-scripts/rgn-at10.d: No longer expect this test to
fail for the RX.
* testsuite/ld-scripts/rgn-at11.d: Likewise.
* testsuite/ld-scripts/rgn-at2.d: Likewise.
* testsuite/ld-scripts/rgn-at6.d: Likewise.
* testsuite/ld-scripts/rgn-at7.d: Likewise.
* testsuite/ld-scripts/rgn-at8.d: Likewise.
When I implement linux_target_ops.get_syscall_trapinfo for aarch64 and arm,
I find the second parameter sysret isn't used at all. In RSP, we don't
need syscall return value either, because GDB can figure out the return
value from registers content got by 'g' packet.
This patch is to remove them.
gdb/gdbserver:
2016-06-28 Yao Qi <yao.qi@linaro.org>
* linux-low.c (get_syscall_trapinfo): Remove parameter sysret.
Callers updated.
* linux-low.h (struct linux_target_ops) <get_syscall_trapinfo>:
Remove parameter sysno.
* linux-x86-low.c (x86_get_syscall_trapinfo): Remove parameter
sysret.
In 82075af2c1 (Implement 'catch syscall'
for gdbserver), only x86 is supported, but the test can still be run
on other linux targets, like aarch64 and ppc, with native-gdbserver.
This causes many new fails.
This patch removes the check on isnative and on target triplets.
Instead, we can insert catch point, and resume the program to see whether
catch syscall is supported or not.
gdb/testsuite:
2016-06-28 Yao Qi <yao.qi@linaro.org>
* gdb.base/catch-syscall.exp: Remove check on isnative and target
triplets. Start gdb, execute catch syscall, and continue. Check
gdb's output to determine catch syscall is supported.
bfd * elfxx-sparc.c (_bfd_sparc_elf_relocate_section): Don't convert
R_SPARC_32 to R_SPARC_RELATIVE if class is ELFCLASS64.
gold * sparc.cc (Target_sparc::Scan::local): Don't convert R_SPARC_32
to R_SPARC_RELATIVE if class is ELFCLASS64.
(Target_sparc::Scan::global): Likewise.
ld * testsuite/ld-elf/symbolic-func.r: Allow non-zero offsets from
.text.
aarch64_opnd_info used bitfields to hold vector element indices,
but values were stored into those bitfields before their ranges had
been checked. This meant large invalid indices could be silently
truncated to smaller valid indices.
The two obvious fixes were to do the range checking earlier or use
a full 64-bit field for the index. I went for the latter for two
reasons:
- Doing the range checking in operand_general_constraint_met_p
seems structurally cleaner than doing it while parsing.
- The bitfields didn't really buy us anything. The imm field
of the union is already 128 bits, so we can use a full int64_t
index without growing the structure.
The patch also adds missing range checks for the elements in a register
list index.
include/
* opcode/aarch64.h (aarch64_opnd_info): Change index fields to int64_t.
opcodes/
* aarch64-opc.c (operand_general_constraint_met_p): Check the
range of ldst_elemlist operands.
(print_register_list): Use PRIi64 to print the index.
(aarch64_print_operand): Likewise.
gas/
* testsuite/gas/aarch64/diagnostic.s,
testsuite/gas/aarch64/diagnostic.l: Add tests for out-of-range indices.
Rust prefers to not specify the return type of a function when it is unit
(`()`). The type is also referred to as "void" in debuginfo but not in actual
usage, so we should never be printing "void" when the language is Rust.
2016-06-27 Manish Goregaokar <manish@mozilla.com>
gdb/ChangeLog:
* rust-lang.c (rust_print_type): Print unit types as "()"
* rust-lang.c (rust_print_type): Omit return type for functions
returning unit
gdb/testsuite/ChangeLog:
* gdb.rust/simple.rs: Add test for returning unit in a function
* gdb.rust/simple.exp: Add expectation for functions returning unit
* config//tc-arc.c (tc_arc_frame_initial_instructions): Use
cfi_add_CFA_def_cfa to generate default CFA with offset
* testsuite/gas/cfi/cfi-arc-1.d: Update expected output.
PR ld/20302
* lexsup.c (set_segment_start): If resetting the start address of
a section, remember to generate a new script element as well.
* testsuite/ld-scripts/pr20302.d: New test.
* testsuite/ld-scripts/scripts.exp: Run the new test.
PR gas/20247
* as.h (do_not_pad_sections_to_alignment): New global variable.
* as.c (show_usage): Add --no-pad-sections.
(parse_args): Likewise.
* write.c (size_seg): Skip padding the end of the section if
requested from the command line.
(SUB_SEGMENT_ALIGN): Likewise.
* doc/as.texinfo: Document the new option.
* NEWS: Mention the new feature.
* testsuite/gas/elf/section11.s: New test.
* testsuite/gas/elf/section11.d: New test driver.
* testsuite/gas/elf/elf.exp: Run the new test.
b399102 fixed the testcase in this PR but it may be possible to
trigger the problem in other ways.
PR ld/19264
* elf64-ppc.c (STUB_SHRINK_ITER): Define.
(ppc64_elf_size_stubs): Exit stub sizing loop past STUB_SHRINK_ITER
if shrinking stubs.
(ppc64_elf_size_stubs): Adjust to suit.
When a Python script tries to create a breakpoint but fails to do so,
gdb.Breakpoint.__init__ raises an exception and the breakpoint does not
exist anymore in the Python interpreter. However, GDB still keeps a
reference to the Python object to be used for a later hook, which is
wrong.
This commit adds the necessary cleanup code so that there is no stale
reference to this Python object. It also adds a new testcase to
reproduce the bug and check the fix.
2016-06-25 Pierre-Marie de Rodat <derodat@adacore.com>
gdb/
* python/py-breakpoint.c (bppy_init): Clear bppy_pending_object
when there is an error during the breakpoint creation.
gdb/testsuite
* gdb.python/py-breakpoint-create-fail.c,
gdb.python/py-breakpoint-create-fail.exp,
gdb.python/py-breakpoint-create-fail.py: New testcase.
H.J. Lu
Manish Goregaokar
Yao Qi
Trevor Saunders
Nick Clifton
Maciej W. Rozycki
GDB Administrator
Alan Modra
Cary Coutant
Igor Kudrin
Sriraman Tallam
Walter Lee
James Clarke
Richard Sandiford
Vineet Gupta
Pierre-Marie de Rodat