This patch fixes a number of bugs in ppc32 plt stub matching code.
1) The 4-insn stubs for shared libs and PIEs weren't matched.
2) The executable stub miscalculated PLT entry address (by oring a
sign-extended quantity rather than adding).
3) Comments were not accurate.
In addition, the insn arrays are made const.
* ppc-linux-tdep.c (powerpc32_plt_stub): Make const.
(powerpc32_plt_stub_so_1): Rename from powerpc32_plt_stub_so.
Remove nop. Make const. Comment.
(powerpc32_plt_stub_so_2): New.
(POWERPC32_PLT_CHECK_LEN): Rename from POWERPC32_PLT_STUB_LEN.
Correct count. Update uses.
(ppc_skip_trampoline_code): Match powerpc32_plt_stub_so_2 too.
Move common code reading PLT entry word. Correct
powerpc32_plt_stub PLT address calculation.
* ppc64-tdep.c (ppc64_standard_linkage1): Make const.
(ppc64_standard_linkage2, ppc64_standard_linkage3): Likewise.
(ppc64_standard_linkage4, ppc64_standard_linkage5): Likewise.
(ppc64_standard_linkage6, ppc64_standard_linkage7): Likewise.
(ppc64_standard_linkage8): Likewise.
* rs6000-tdep.c (ppc_insns_match_pattern): Make pattern const.
Correct insns description.
* ppc-tdep.h (ppc_insns_match_pattern): Update prototype.
Reviewed-By: Yao Qi <qiyaoltc@gmail.com>
This patch fixes a case where a user had a C-representable named
section in both the executable and shared libraries, and of course
wanted the size of the local section in the executable, not the
dynamic section. It does mean that __start and __stop symbols don't
behave exactly like PROVIDEd symbols, but I think that's a reasonable
difference particularly since this is the way they used to behave.
* elflink.c (bfd_elf_define_start_stop): Override symbols when
they are defined dynamically.
https://bugzilla.redhat.com/show_bug.cgi?id=1523457
I haven't analyzed this myself, I'm relying on Nick's excellent
analysis. What I believe is happening is that after some number of
stub sizing iterations, a long-branch stub needs to be converted to a
plt-branch, but either due to stub alignment or other stubs shrinking
in size, the stub group section size doesn't change.
That means we exit from ppc64_elf_size_stubs after sizing with an
incorrect layout, in fact the additional .branch_lt entry overlays
.got! Since .TOC. is normally set to .got + 0x8000 the stub sizing
code decides that entry is within +/-32k of the TOC pointer and so a
three insn stub is sufficient. When we come to build the stubs using
a correct non-overlaying layout, a four insn plt-branch stub is
generated and the stub group size doesn't match that calculated
earlier.
* elf64-ppc.c (ppc64_elf_size_stubs): Iterate sizing when
.branch_lt changes size.
There are actually 2 different bugs:
1. TLS transition is broken in PIE mode.
2. TLS is broken in PIC/PIE mode when the __tls_get_addr symbol
is versioned (as is the case on Linux and Solaris at least).
The 1st bug is fixed by reverting the problematic change for now
(note that the associated test doesn't pass on SPARC because of another
issue so there is no formal regression in the testsuite). The 2nd bug
is fixed by changing the call to _bfd_generic_link_add_one_symbol
on __tls_get_addr into a mere lookup in _bfd_sparc_elf_check_relocs.
bfd/
* elfxx-sparc.c (_bfd_sparc_elf_check_relocs) <R_SPARC_TLS_GD_CALL>:
Do a mere lookup of the __tls_get_addr symbol instead of adding it.
Revert
2017-10-19 H.J. Lu <hongjiu.lu@intel.com>
PR ld/22263
* elfxx-sparc.c (sparc_elf_tls_transition): Replace
bfd_link_pic with !bfd_link_executable, !bfd_link_pic with
bfd_link_executable for TLS check.
(_bfd_sparc_elf_check_relocs): Likewise.
(allocate_dynrelocs): Likewise.
(_bfd_sparc_elf_relocate_section): Likewise.
ld/
* testsuite/ld-sparc/sparc.exp (32-bit: Helper shared library):
Link with a version script.
(32-bit: TLS -fpie): New test.
(64-bit: Helper shared library): Link with a version script.
(64-bit: TLS -fpie): New test.
(64-bit: GOTDATA relocations): Pass -Av9 to the assembler.
* testsuite/ld-sparc/tlslib.ver: New file.
* testsuite/ld-sparc/tlspie32.dd: Likewise.
* testsuite/ld-sparc/tlspie32.s: Likewise.
* testsuite/ld-sparc/tlspie64.dd: Likewise.
* testsuite/ld-sparc/tlspie64.s: Likewise.
* testsuite/ld-sparc/tlssunbin32.dd: Adjust for versioned symbol.
* testsuite/ld-sparc/tlssunbin32.rd: Likewise.
* testsuite/ld-sparc/tlssunbin32.sd: Likewise.
* testsuite/ld-sparc/tlssunbin64.dd: Likewise.
* testsuite/ld-sparc/tlssunbin64.rd: Likewise.
* testsuite/ld-sparc/tlssunbin64.sd: Likewise.
GCC PR83906 [1] is about a GCC/libstdc++ GDB/Python type printer
testcase failing randomly, as shown by running (in libstdc++'s
testsuite):
make check RUNTESTFLAGS=prettyprinters.exp=80276.cc
in a loop. Sometimes you get this:
FAIL: libstdc++-prettyprinters/80276.cc whatis p4
I.e., this:
type = std::unique_ptr<std::vector<std::unique_ptr<std::list<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >>[]>>[99]>
instead of this:
type = std::unique_ptr<std::vector<std::unique_ptr<std::list<std::string>[]>>[99]>
Jonathan Wakely tracked it on the printer side to this bit in
libstdc++'s type printer:
if self.type_obj == type_obj:
return strip_inline_namespaces(self.name)
This assumes the two types resolve to the same gdb.Type but some times
the comparison unexpectedly fails.
Running the testcase manually under Valgrind finds the problem in GDB:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
==6118== Conditional jump or move depends on uninitialised value(s)
==6118== at 0x4C35CB0: bcmp (vg_replace_strmem.c:1100)
==6118== by 0x6F773A: check_types_equal(type*, type*, VEC_type_equality_entry_d**) (gdbtypes.c:3515)
==6118== by 0x6F7B00: check_types_worklist(VEC_type_equality_entry_d**, bcache*) (gdbtypes.c:3618)
==6118== by 0x6F7C03: types_deeply_equal(type*, type*) (gdbtypes.c:3655)
==6118== by 0x4D5B06: typy_richcompare(_object*, _object*, int) (py-type.c:1007)
==6118== by 0x63D7E6C: PyObject_RichCompare (object.c:961)
==6118== by 0x646EAEC: PyEval_EvalFrameEx (ceval.c:4960)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
That "bcmp" call is really a memcmp call in check_types_equal. The
problem is that gdb is memcmp'ing two objects that are equal in value:
(top-gdb) p *TYPE_RANGE_DATA (type1)
$1 = {low = {kind = PROP_CONST, data = {const_val = 0, baton = 0x0}}, high = {kind = PROP_CONST, data = {const_val = 15, baton = 0xf}}, flag_upper_bound_is_count = 0,
flag_bound_evaluated = 0}
(top-gdb) p *TYPE_RANGE_DATA (type2)
$2 = {low = {kind = PROP_CONST, data = {const_val = 0, baton = 0x0}}, high = {kind = PROP_CONST, data = {const_val = 15, baton = 0xf}}, flag_upper_bound_is_count = 0,
flag_bound_evaluated = 0}
but differ in padding. Notice the 4-byte hole:
(top-gdb) ptype /o range_bounds
/* offset | size */ type = struct range_bounds {
/* 0 | 16 */ struct dynamic_prop {
/* 0 | 4 */ dynamic_prop_kind kind;
/* XXX 4-byte hole */
/* 8 | 8 */ union dynamic_prop_data {
/* 8 */ LONGEST const_val;
/* 8 */ void *baton;
/* total size (bytes): 8 */
} data;
which is filled with garbage:
(top-gdb) x /40bx TYPE_RANGE_DATA (type1)
0x2fa7ea0: 0x01 0x00 0x00 0x00 0x43 0x01 0x00 0x00
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
0x2fa7ea8: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0x2fa7eb0: 0x01 0x00 0x00 0x00 0xfe 0x7f 0x00 0x00
0x2fa7eb8: 0x0f 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0x2fa7ec0: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
(top-gdb) x /40bx TYPE_RANGE_DATA (type2)
0x20379b0: 0x01 0x00 0x00 0x00 0xfe 0x7f 0x00 0x00
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
0x20379b8: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0x20379c0: 0x01 0x00 0x00 0x00 0xfe 0x7f 0x00 0x00
0x20379c8: 0x0f 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0x20379d0: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
(top-gdb) p memcmp (TYPE_RANGE_DATA (type1), TYPE_RANGE_DATA (type2), sizeof (*TYPE_RANGE_DATA (type1)))
$3 = -187
In some cases objects of type range_bounds are memset when allocated,
but then their dynamic_prop low/high fields are copied over from some
template dynamic_prop object that wasn't memset. E.g.,
create_static_range_type's low/high locals are left with garbage in
the padding, and then that padding is copied over to the range_bounds
object's low/high fields.
At first, I considered making sure to always memset range_bounds
objects, thinking that maybe type objects are being put in some bcache
instance somewhere. But then I hacked bcache/bcache_full to poison
non-pod types, and made dynamic_prop a non-pod, and GDB still
compiled.
So given that, it seems safest to not assume padding will always be
memset, and instead treat them as regular value types, implementing
(in)equality operators and using those instead of memcmp.
This fixes the random FAILs in GCC's testcase.
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83906
gdb/ChangeLog:
2018-01-24 Pedro Alves <palves@redhat.com>
GCC PR libstdc++/83906
* gdbtypes.c (operator==(const dynamic_prop &,
const dynamic_prop &)): New.
(operator==(const range_bounds &, const range_bounds &)): New.
(check_types_equal): Use them instead of memcmp.
* gdbtypes.h (operator==(const dynamic_prop &,
const dynamic_prop &)): Declare.
(operator!=(const dynamic_prop &, const dynamic_prop &)): Declare.
(operator==(const range_bounds &, const range_bounds &)): Declare.
(operator!=(const range_bounds &, const range_bounds &)): Declare.
Correct an issue with the `mips64*-ps2-elf*' target introduced with
commit e407c74b5b ("Support for MIPS R5900 (Sony Playstation 2)"),
<https://sourceware.org/ml/binutils/2012-12/msg00240.html> and make
the n32 ABI the default for GAS, consistently with how BFD and LD
are configured for this target.
gas/
* configure.ac: Also set `mips_default_abi' to N32_ABI for
`mips64*-ps2-elf*'.
* configure: Regenerate.
Update a stale reference to `mips_elf32_section_processing', following a
merge of the function into `_bfd_mips_elf_section_processing' made with
commit 103186c62b ("PATCH for MIPS 64-bit ABI suport"),
<https://sourceware.org/ml/binutils/1999-q3/msg00072.html>.
bfd/
* elfxx-mips.c (_bfd_mips_elf_final_link): Update a stale
`mips_elf32_section_processing' comment reference.
Remove an issue with `as --help' always reporting `o32' as the default
ABI regardless of what the default actually is, originally caused by
commit cac012d6d3 ("check mips abi x linker emulation compatibility"),
<https://sourceware.org/ml/binutils/2003-05/msg00187.html> missing an
update here.
gas/
* config/tc-mips.c (md_show_usage): Correctly indicate the
configuration-specific default ABI.
Correct a commit 25499ac7ee ("MIPS16e2: Add MIPS16e2 ASE support") GAS
bug and add missing help text for the `-mmips16e2' and `-mno-mips16e2'
options added with said commit.
gas/
* config/tc-mips.c (md_show_usage): Report `-mmips16e2' and
`-mno-mips16e2' options.
After moving big parts of the code to the new s390-tdep.c file
s390-linux-tdep.c now contains many includes it doesn't need anymore.
Furthermore, there are some functions lacking a description.
Fix both and order the remaining includes alphabetically.
gdb/ChangeLog
* s390-linux-tdep.c: Remove unneeded includes and order them
alphabetically. Add comments to functions without description.
Record-replay is independent of the OS. So it can be moved to the common
s390 code without problem.
gdb/ChangeLog:
* s390-linux-tdep.c (s390_record_address_mask)
(s390_record_calc_disp_common, s390_record_calc_disp)
(s390_record_calc_disp_vsce, s390_record_calc_rl, s390_popcnt)
(s390_record_gpr_g, s390_record_gpr_h, s390_record_vr)
(s390_process_record): Move to s390-tdep.c.
(s390_linux_init_abi_any): Adjust.
* s390-tdep.c (s390_record_address_mask)
(s390_record_calc_disp_common, s390_record_calc_disp)
(s390_record_calc_disp_vsce, s390_record_calc_rl, s390_popcnt)
(s390_record_gpr_g, s390_record_gpr_h, s390_record_vr)
(s390_process_record): Moved from s390-linux-tdep.c
(s390_gdbarch_init): Adjust.
Currently all target dependent code for s390 is in one file,
s390-linux-tdep.c. This includes code general for the architecture as
well as code specific for uses in GNU/Linux (user space). Up until now
this was OK as GNU/Linux was the only supported OS. In preparation to
support the new Linux kernel 'OS' split up the existing s390 code into a
general s390-tdep and a GNU/Linux-specific s390-linux-tdep.
Note: The record-replay feature will be moved in a separate patch. This
is simply due to the fact that the combined patch would be too large for
the mailing list. This requires setting the process_record hook during
OSABI init to keep the code bisectable. The patch moving record-replay
cleans up this hack.
gdb/ChangeLog:
* s390-linux-nat.c (s390-tdep.h): New include.
* Makefile.in (ALL_TARGET_OBS): Add s390-tdep.o.
(HFILES_NO_SRCDIR): Add s390-tdep.h.
(ALLDEPFILES): Add s390-tdep.c.
* configure.tgt (s390*-*-linux*): Add s390-tdep.o.
* s390-linux-tdep.h (HWCAP_S390_*, S390_*_REGNUM): Move to...
* s390-tdep.h: ...this. New file.
* s390-linux-tdep.c (s390-tdep.h): New include.
(_initialize_s390_tdep): Rename to...
(_initialize_s390_linux_tdep): ...this and adjust.
(s390_abi_kind, s390_vector_abi_kind, gdbarch_tdep)
(enum named opcodes, S390_NUM_GPRS, S390_NUM_FPRS): Move to
s390-tdep.h.
(s390_break_insn, s390_breakpoint, s390_readinstruction, is_ri)
(is_ril, is_rr, is_rre, is_rs, is_rsy, is_rx, is_rxy)
(s390_is_partial_instruction, s390_software_single_step)
(is_non_branch_ril, s390_displaced_step_copy_insn)
(s390_displaced_step_fixup, s390_displaced_step_hw_singlestep)
(s390_prologue_data, s390_addr, s390_store, s390_load)
(s390_check_for_saved, s390_analyze_prologue, s390_skip_prologue)
(s390_register_call_saved, s390_guess_tracepoint_registers)
(s390_register_name, s390_dwarf_regmap, s390_dwarf_reg_to_regnum)
(regnum_is_gpr_full, regnum_is_vxr_full, s390_value_from_register)
(s390_pseudo_register_name, s390_pseudo_register_type)
(s390_pseudo_register_read, s390_pseudo_register_write)
(s390_pseudo_register_reggroup_p, s390_ax_pseudo_register_collect)
(s390_ax_pseudo_register_push_stack, s390_gen_return_address)
(s390_addr_bits_remove, s390_address_class_type_flags)
(s390_address_class_type_flags_to_name)
(s390_address_class_name_to_type_flags, s390_effective_inner_type)
(s390_function_arg_float, s390_function_arg_vector)
(is_power_of_two, s390_function_arg_integer, s390_arg_state)
(s390_handle_arg, s390_push_dummy_call, s390_dummy_id)
(s390_frame_align, s390_register_return_value, s390_return_value)
(s390_stack_frame_destroyed_p, s390_unwind_pc, s390_unwind_sp)
(s390_unwind_pseudo_register, s390_adjust_frame_regnum)
(s390_dwarf2_prev_register, s390_dwarf2_frame_init_reg)
(s390_trad_frame_prev_register, s390_unwind_cache)
(s390_prologue_frame_unwind_cache)
(s390_backchain_frame_unwind_cache, s390_frame_unwind_cache)
(s390_frame_this_id, s390_frame_prev_register, s390_frame_unwind)
(s390_stub_unwind_cache, s390_stub_frame_unwind_cache)
(s390_stub_frame_this_id, s390_stub_frame_prev_register)
(s390_stub_frame_sniffer, s390_stub_frame_unwind)
(s390_frame_base_address, s390_local_base_address)
(s390_frame_base, s390_gcc_target_options)
(s390_gnu_triplet_regexp, s390_stap_is_single_operand)
(s390_validate_reg_range, s390_tdesc_valid)
(s390_gdbarch_tdep_alloc, s390_gdbarch_init): Move to...
* s390-tdep.c: ...this. New file.
Most parts of s390_process_record are common for the architecture. Only
the system call handling differs between the OSes. In order to be able to
move s390_process_record to a common code file add a hook to record
syscalls to gdbarch_tdep. So every OS can implement their own handling.
gdb/ChangeLog:
* s390-linux-tdep.c (gdbarch_tdep.s390_syscall_record): New hook.
(s390_process_record, s390_gdbarch_tdep_alloc)
(s390_linux_init_abi_any): Use/set new hook.
Do what the title says and distinguish between 31- and 64-bit systems.
The goal is to init the OSABI as late as possible in gdbarch_init so the
OSABI has the chance to overwrite the defaults.
There are two pitfalls to be aware of:
First, the dwarf2 unwinder must be appended before the OSABI is
initialized. Otherwise the OS could add a default unwinder which always
takes control before the dwarf unwinder even gets a chance.
Second, tdesc_use_registers has to be handled with extra care. It sets
several gdbarch hooks, especially gdbarch_register_name, which has to be
overwritten again after the call. Furthermore it deletes the tdesc_data
without checking. Therefore there must not be a call to
tdesc_data_cleanup afterwards or GDB will crash with a double free.
gdb/ChangeLog:
* s390-linux-tdep.c (osabi.h): New include.
(s390_linux_init_abi_31, s390_linux_init_abi_64)
(s390_linux_init_abi_any): New functions.
(s390_gdbarch_init, _initialize_s390_tdep): Adjust.
Before doing the tdesc validation there is a check whether the tdesc has
registers or not. This check is not only unnecessary but wrong.
First the check is done after a default tdesc is assigned if the original
tdesc has no registers. These default tdescs always have registers so the
check alway returns true.
Second if the default tdesc would not have registers the check only skips
the tdesc validation instead of returning an error. This would trigger a
gdb_assert later on in tdesc_use_registers.
gdb/ChangeLog:
* s390-linux-tdep.c (s390_gdbarch_init): Use gdb_assert for
tdesc_has_registers check
Simplify s390_gdbarch_init by moving the target description validation to
a separate function.
gdb/ChangeLog:
* s390-linux-tdep.c (s390_tdesc_valid): New function.
(s390_validate_reg_range): New macro.
(s390_gdbarch_init): Adjust.
Add a field for the target description to gdbarch_tdep. This will later be
needed to pass the 'correct' target description from osabi_init to
gdbarch_init. Unfortunately this cannot be done using gdbarch_info as it
is only passed by copy, not reference.
gdb/ChangeLog:
* s390-linux-tdep.c (gdbarch_tdep) <tdesc>: New field.
(s390_gdbarch_tdep_alloc): Adjust.
(s390_gdbarch_init): Adjust.
Currently the gdbarch_tdep.have_* flags are a mix of int and bool. Clean
this up by making them all bool.
gdb/ChangeLog:
* s390-linux-tdep.c (gdbarch_tdep) <have_linux_v1, have_linux_v2>
<have_tdb>: Change type to bool.
(s390_gdbarch_tdep_alloc): Adjust.
(s390_gdbarch_init): Adjust.
Moving the allocation of gdbarch_tdep to the start of s390_gdbarch_init
allows us to use its fields for tracking the different features instead of
using separate variables. To make the code a little nicer move the actual
allocation and initialization to a separate function. Also move the
allocation of gdbarch to keep the two together.
gdb/ChangeLog:
* s390-linux-tdep (s390_abi_kind) <ABI_NONE>: New default field.
(gdbarch_tdep) <have_upper, have_vx>: New fields.
(s390_gdbarch_tdep_alloc): New function.
(s390_gdbarch_init): Allocate tdep at start and use its fields
instead of separate variables.
When initializing the gdbarch there is a check whether an appropriate
gdbarch already exists in the gdbarch_list. Failing of some of the checks
would lead to a different target description. However
gdbarch_list_lookup_by_info already checks for
if (info->target_desc != arches->gdbarch->target_desc)
continue;
Remove these duplicate checks.
gdb/ChangeLog:
* s390-linux-tdep.c (s390_gdbarch_init): Remove duplicate checks
when looking for cached gdbarch and add comment for remaining.
Compiling GDB with a recent GCC exposes a problem:
../../gdb/typeprint.c: In function 'void whatis_exp(const char*, int)':
../../gdb/typeprint.c:515:12: warning: 'val' may be used uninitialized in this function [-Wmaybe-uninitialized]
real_type = value_rtti_type (val, &full, &top, &using_enc);
~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The warning is correct. There are indeed code paths that use
uninitialized 'val', leading to crashes. Inside the
value_rtti_indirect_type/value_rtti_type calls here in whatis_exp:
if (opts.objectprint)
{
if (((TYPE_CODE (type) == TYPE_CODE_PTR) || TYPE_IS_REFERENCE (type))
&& (TYPE_CODE (TYPE_TARGET_TYPE (type)) == TYPE_CODE_STRUCT))
real_type = value_rtti_indirect_type (val, &full, &top, &using_enc);
else if (TYPE_CODE (type) == TYPE_CODE_STRUCT)
real_type = value_rtti_type (val, &full, &top, &using_enc);
}
We reach those calls above with "set print object on", and then with
any of:
(gdb) whatis struct some_structure_type
(gdb) whatis struct some_structure_type *
(gdb) whatis struct some_structure_type &
because "whatis" with a type argument enters this branch:
/* The behavior of "whatis" depends on whether the user
expression names a type directly, or a language expression
(including variable names). If the former, then "whatis"
strips one level of typedefs, only. If an expression,
"whatis" prints the type of the expression without stripping
any typedef level. "ptype" always strips all levels of
typedefs. */
if (show == -1 && expr->elts[0].opcode == OP_TYPE)
{
which does not initialize VAL. Trying the above triggers crashes like
this:
(gdb) set print object on
(gdb) whatis some_structure_type
Thread 1 "gdb" received signal SIGSEGV, Segmentation fault.
0x00000000005dda90 in check_typedef (type=0x6120736573756170) at src/gdb/gdbtypes.c:2388
2388 int instance_flags = TYPE_INSTANCE_FLAGS (type);
...
This is a regression caused by a recent-ish refactoring of the code on
'whatis_exp', introduced by:
commit c973d0aa4a
Date: Mon Aug 21 11:34:32 2017 +0100
Fix type casts losing typedefs and reimplement "whatis" typedef stripping
Fix this by setting VAL to NULL in the "whatis TYPE" case, and
skipping fetching the dynamic type if there's no value to fetch it
from.
New tests included.
gdb/ChangeLog:
2018-01-22 Pedro Alves <palves@redhat.com>
Sergio Durigan Junior <sergiodj@redhat.com>
* typeprint.c (whatis_exp): Initialize "val" in the "whatis type"
case.
gdb/testsuite/ChangeLog:
2018-01-22 Pedro Alves <palves@redhat.com>
Sergio Durigan Junior <sergiodj@redhat.com>
* gdb.base/whatis.exp: Add tests for 'set print object on' +
'whatis <struct>' 'whatis <struct> *' and 'whatis <struct> &'.
Following my recent transition from Imagination Technologies to the
reincarnated MIPS company update MAINTAINERS entries accordingly.
binutils/
* MAINTAINERS: Update my company e-mail address.
gdb/
* MAINTAINERS: Update my company e-mail address.
sim/
* MAINTAINERS: Update my company e-mail address.
Since my following patches will change how each gdbarch read and write
pseudo registers, it's better to write a unit test to
regcache::cooked_write, to make sure my following changes don't cause
any regressions. See the comments on cooked_write_test.
gdb:
2018-01-22 Yao Qi <yao.qi@linaro.org>
* regcache.c (cooked_write_test): New function.
(_initialize_regcache): Register the test.
The patch later in this series will move regcache's raw_read and
cooked_read methods to a new class regcache_read, and regcache is
dervied from it. Also pass regcache_read instead of regcache to gdbarch
methods pseudo_register_read and pseudo_register_read_value. In order
to prepare for this change, this patch changes regcache_raw_read to
regcache->raw_read. On the other hand, since we are in C++, I prefer
using class method (regcache->raw_read).
gdb:
2018-01-22 Yao Qi <yao.qi@linaro.org>
* aarch64-tdep.c (aarch64_pseudo_read_value): Call regcache
method raw_read instead of regcache_raw_read.
* amd64-tdep.c (amd64_pseudo_register_read_value): Likewise.
* arm-tdep.c (arm_neon_quad_read): Likewise.
* avr-tdep.c (avr_pseudo_register_read): Likewise.
* bfin-tdep.c (bfin_pseudo_register_read): Likewise.
* frv-tdep.c (frv_pseudo_register_read): Likewise.
* h8300-tdep.c (h8300_pseudo_register_read): Likewise.
* i386-tdep.c (i386_mmx_regnum_to_fp_regnum): Likewise.
(i386_pseudo_register_read_into_value): Likewise.
* mep-tdep.c (mep_pseudo_cr32_read): Likewise.
* msp430-tdep.c (msp430_pseudo_register_read): Likewise.
* nds32-tdep.c (nds32_pseudo_register_read): Likewise.
* rl78-tdep.c (rl78_pseudo_register_read): Likewise.
* s390-linux-tdep.c (s390_pseudo_register_read): Likewise.
* sparc-tdep.c (sparc32_pseudo_register_read): Likewise.
* sparc64-tdep.c (sparc64_pseudo_register_read): Likewise.
* spu-tdep.c (spu_pseudo_register_read_spu): Likewise.
* xtensa-tdep.c (xtensa_pseudo_register_read): Likewise.
This patch removes the MT port. The removal was annoucned
https://sourceware.org/ml/gdb-announce/2017/msg00006.html
I'll remove MT from the top-level configure later.
gdb:
2018-01-22 Yao Qi <yao.qi@linaro.org>
* Makefile.in (ALL_TARGET_OBS): Remove mt-tdep.o.
* configure.tgt: Remove target mt.
* mt-tdep.c: Remove.
* regcache.c (cooked_read_test): Remove the check for mt.
gdbarch_pseudo_register_read_value is not implemented in every gdbarch, so
the predicate gdbarch_pseudo_register_read_value_p is needed before
calling it. However, there is no such guard in jit_frame_prev_register, I
am wondering how does jit work on the arch without having gdbarch method
pseudo_register_read_value.
The proper way to get register value is to call cooked_read, and then
create the value object from the buffer.
gdb:
2018-01-22 Yao Qi <yao.qi@linaro.org>
* jit.c (jit_frame_prev_register): Call regcache::cooked_read
instead of gdbarch_pseudo_register_read_value.
GCC was enhanced in 2011 to generate this attribute, so I think we can
now assume that it is available when using that compiler. Doing so
allows us to speed up what we call "parallel type" lookups when
processing certain types encoded using the GNAT encoding.
This patch changes need_gnat_info to always expect those attributes
to be generated when the language is Ada. This is an assumption
that on the surfcace looks like it might be a bit on the edge; but
in practice, it should be OK because this is only useful in the
context of handling GNAT-specific encodings. Other Ada compilers
would presumably produce debugging information using pure DWARF
constructs, so would not be impacted by this.
gdb/ChangeLog:
* dwarf2read.c (need_gnat_info): Return nonzero if the cu's
language is Ada.
Tested on x86_64-linux.
Consider the following situation, where we have one file containing...
$ cat -n body.inc
1 i = i + 1;
... we include that file from some code, like so:
$ cat -n cat -n small.c
[...]
17 int
18 next (int i)
19 {
20 #include "body.inc"
21 return i;
22 }
When trying to insert a breakpoint on line 18, for instance:
(gdb) b small.c:18
Breakpoint 1 at 0x40049f: file body.inc, line 18.
^^
||
Here, the issue is that GDB reports the breakpoint to be in file
body.inc, which is true, but with the line number that corresponding
to the user-requested location, which is not correct.
Although the simple reproducer may look slightly artificial,
the above is simply one way to reproduce the same issue observed
when trying to insert a breakpoint on a function provided in
a .h files and then subsequently inlined in a C file.
What happens is the following:
1. We resolve the small.c:18 linespec into a symtab_and_line which
has "small.c" and 18 as the symtab and line number.
2. Next, we call skip_prologue_sal, which calculates the PC
past the prologue, and updates the symtab_and_line: PC,
but also symtab (now body.inc) and the new line (now 1).
3. However, right after that, we do:
/* Make sure the line matches the request, not what was
found. */
intermediate_results.sals[i].line = val.line;
We should either restore both symtab and line, or leave the actual
line to match the actual symtab. This patch chose the latter.
This introduces a few changes in a few tests, which required some
updates, but looking at those change, I believe them to be expected.
gdb/ChangeLog:
* linespec.c (create_sals_line_offset): Remove code that preserved
the symtab_and_line's line number.
gdb/testsuite/ChangeLog:
* gdb.base/break-include.c, gdb.base/break-include.inc,
gdb.base/break-include.exp: New files.
* gdb.base/ending-run.exp: Minor adaptations due to the breakpoint's
line number now being the actual line number where the breakpoint
was inserted.
* gdb.mi/mi-break.exp: Likewise.
* gdb.mi/mi-reverse.exp: Likewise.
* gdb.mi/mi-simplerun.exp: Ditto.
Tested on x86_64-linux.
When creating a varobj with -var-create a user can create either fixed
varobj, or floating varobj.
A fixed varobj will always be evaluated within the thread/frame/block in
which the varobj was created, if that thread/frame/block is no longer
available then the varobj is considered out of scope.
A floating varobj will always be evaluated within the current
thread/frame/block.
Despite never using them GDB was storing the thread/frame/block into a
floating varobj, and the thread-id would then be displayed when GDB
reported on the state of the varobj, this could confuse a user into
thinking that the thread-id was relevant.
This commit prevents GDB storing the thread/frame/block onto floating
varobj, and updates the few tests where this impacts the results.
gdb/ChangeLog:
* varobj.c (varobj_create): Don't set valid_block when creating a
floating varobj.
gdb/testsuite/ChangeLog:
* gdb.python/py-mi.exp: Don't expect a thread-id for floating
varobj.
* gdb.mi/mi-var-create-rtti.exp: Likewise.
This patch fixes a problem with using the MI -var-update command
to access the values of registers in frames other than the current
frame. The patch includes a test that demonstrates the problem:
* run so there are several frames on the stack
* create a fixed varobj for $pc in each frame, #'s 1 and above
* step one instruction, to modify the value of $pc
* call -var-update for each of the previously created varobjs
to verify that they are not reported as having changed.
Without the patch, the -var-update command reported that $pc for all
frames 1 and above had changed to the value of $pc in frame 0.
A varobj is created as either fixed, the expression is evaluated within
the context of a specific frame, or floating, the expression is
evaluated within the current frame, whatever that may be.
When a varobj is created by -var-create we set two fields of the varobj
to track the context in which the varobj was created, these two fields
are varobj->root->frame and var->root->valid_block.
If a varobj is of type fixed, then, when we subsequently try to
reevaluate the expression associated with the varobj we must determine
if the original frame (and block) is still available, if it is not then
the varobj can no longer be evaluated.
The problem is that for register expressions varobj->root->valid_block
is not set correctly. This block tracking is done using the global
'innermost_block' which is set in the various parser files (for example
c-exp.y). However, this is not set for register expressions.
The fix then seems like it should be to just update the innermost block
when parsing register expressions, however, that solution causes several
test regressions.
The problem is that in some cases we rely on the expression parsing
code not updating the innermost block for registers, one example is
when we parse the expression for a 'display' command. The display
commands treats registers like floating varobjs, but symbols are
treated like fixed varobjs. So 'display $reg_name' will always show
the value of '$reg_name' even as the user moves from frame to frame,
while 'display my_variable' will only show 'my_variable' while it is
in the current frame and/or block, when the user moves to a new frame
and/or block (even one with a different 'my_variable' in) then the
display of 'my_variable' stops. For the case of 'display', without
the option to force fixed or floating expressions, the current
behaviour is probably the best choice. For the varobj system though,
we can choose between floating and fixed, and we should try to make
this work for registers.
There's only one existing test case that needs to be updated, in that
test a fixed varobj is created using a register, the MI output now
include the thread-id in which the varobj should be evaluated, which I
believe is correct behaviour. I also added a new floating test case
into the same test script, however, right now this also includes the
thread-id in the expected output, which I believe is an existing gdb
bug, which I plan to fix next.
Tested on x86_64 Linux native and native-gdbserver, no regressions.
gdb/ChangeLog:
PR mi/20395
* ada-exp.y (write_var_from_sym): Pass extra parameter when
updating innermost block.
* parse.c (innermost_block_tracker::update): Take extra type
parameter, and check types match before updating innermost block.
(write_dollar_variable): Update innermost block for registers.
* parser-defs.h (enum innermost_block_tracker_type): New enum.
(innermost_block_tracker::innermost_block_tracker): Initialise
m_types member.
(innermost_block_tracker::reset): Take type parameter.
(innermost_block_tracker::update): Take type parameter, and pass
type through as needed.
(innermost_block_tracker::m_types): New member.
* varobj.c (varobj_create): Pass type when reseting innermost
block.
gdb/testsuite/ChangeLog:
* gdb.mi/basics.c: Add new global.
* gdb.mi/mi-frame-regs.exp: New file.
* gdb.mi/mi-var-create-rtti.exp: Update expected results, add new
case.
This commit is preparation for a later change, at this point there
should be no user visible change.
We currently maintain a global innermost_block which tracks the most
inner block encountered when parsing an expression.
This commit wraps the innermost_block into a new class, and switches all
direct accesses to the variable to use the class API.
gdb/ChangeLog:
* ada-exp.y (write_var_from_sym): Switch to innermost_block API.
* ada-lang.c (resolve_subexp): Likewise.
* breakpoint.c (set_breakpoint_condition) Likewise.
(watch_command_1) Likewise.
* c-exp.y (variable): Likewise.
* d-exp.y (PrimaryExpression): Likewise.
* f-exp.y (variable): Likewise.
* go-exp.y (variable): Likewise.
* m2-exp.y (variable): Likewise.
* objfiles.c (objfile::~objfile): Likewise.
* p-exp.y (variable): Likewise.
* parse.c (innermost_block): Change type.
* parser-defs.h (class innermost_block_tracker): New.
(innermost_block): Change to innermost_block_tracker.
* printcmd.c (display_command): Switch to innermost_block API.
(do_one_display): Likewise.
* rust-exp.y (do_one_display): Likewise.
* symfile.c (clear_symtab_users): Likewise.
* varobj.c (varobj_create): Switch to innermost_block API, replace
use of innermost_block with block stored on varobj object.
Alan Hayward