All execution commands currently have this pattern:
/* If we must run in the background, but the target can't do it,
error out. */
if (async_exec && !target_can_async_p ())
error (_("Asynchronous execution not supported on this target."));
/* If we are not asked to run in the bg, then prepare to run in the
foreground, synchronously. */
if (!async_exec && target_can_async_p ())
{
/* Simulate synchronous execution. */
async_disable_stdin ();
}
This patch factors that into a shared function.
attach_command installs a cleanup to re-enable stdin, but that's not
necessary, as per the comment in prepare_execution_command. In any
case, if someday it turns out necessary, we have a single place to
install it now.
Tested on x86_64 Fedora 17, sync and async modes.
gdb/
2014-03-12 Pedro Alves <palves@redhat.com>
* infcmd.c (prepare_execution_command): New function, factored out
from several execution commands.
(run_command_1, continue_command, step_1, jump_command)
(signal_command, until_command, advance_command, finish_command)
(attach_command): Use prepare_execution_command.
This patch updates arm native support for hwbreak-/watchpoints to enable
support for hwbreak-/watchpoints across fork/vfork. This involves changes to
hwbreak-/watchpoint insertion mechanism to the modern way, by marking debug
registers as needing update, but only really updating them on resume, which is
necessary for supporting watchpoints in non-stop mode. This also updates a
previously maintained per thread hwbreak-/watchpoint cache to a per process
cache which allows target specific code to come in sync with gdb-linux calls to
threads create/destroy and process fork/exit hooks.
I noticed 'make check TESTS="..."' works when ran from gdb/testsuite/,
but TESTS is ignored when "make check" is ran from gdb/.
The issue is that TESTS isn't being passed to the testsuite subdir
make invocation.
gdb/
2014-03-12 Pedro Alves <palves@redhat.com>
* Makefile.in (TARGET_FLAGS_TO_PASS): Add TESTS.
A patch in the target cleanup series caused a regression when using
record with target-async. Version 4 of the patch is here:
https://sourceware.org/ml/gdb-patches/2014-03/msg00159.html
The immediate problem is that record supplies to_can_async_p and
to_is_async_p methods, but does not supply a to_async method. So,
when target-async is set, record claims to support async -- but if the
underlying target does not support async, then the to_async method
call will end up in that method's default implementation, namely
tcomplain.
This worked previously because the record target used to provide a
to_async method; one that (erroneously, only at push time) checked the
other members of the target stack, and then simply dropped to_async
calls in the "does not implement async" case.
My first thought was to simply drop tcomplain as the default for
to_async. This works, but Pedro pointed out that the only reason
record has to supply to_can_async_p and to_is_async_p is that these
default to using the find_default_run_target machinery -- and these
defaults are only needed by "run" and "attach".
So, a nicer solution presents itself: change run and attach to
explicitly call into the default run target when needed; and change
to_is_async_p and to_can_async_p to default to "return 0". This makes
the target stack simpler to use and lets us remove the method
implementations from record. This is also in harmony with other plans
for the target stack; namely trying to reduce the impact of
find_default_run_target. This approach makes it clear that
find_default_is_async_p is not needed -- it is asking whether a target
that may not even be pushed is actually async, which seems like a
nonsensical question.
While an improvement, this approach proved to introduce the same bug
when using the core target. Looking a bit deeper, the issue is that
code in "attach" and "run" may need to use either the current target
stack or the default run target -- but different calls into the target
API in those functions could wind up querying different targets.
This new patch makes the target to use more explicit in "run" and
"attach". Then these commands explicitly make the needed calls
against that target. This ensures that a single target is used for
all relevant operations. This lets us remove a couple find_default_*
functions from various targets, including the dummy target. I think
this is a decent understandability improvement.
One issue I see with this patch is that the new calls in "run" and
"attach" are not very much like the rest of the target API. I think
fundamentally this is due to bad factoring in the target API, which
may need to be fixed for multi-target. Tackling that seemed ambitious
for a regression fix.
While working on this I noticed that there don't seem to be any test
cases that involve both target-async and record, so this patch changes
break-precsave.exp to add some. It also changes corefile.exp to add
some target-async tests; these pass with current trunk and with this
patch applied, but fail with the v1 patch.
This patch differs from v4 in that it moves initialization of
to_can_async_p and to_supports_non_stop into inf-child, adds some
assertions to complete_target_initialization, and adds some comments
to target.h.
Built and regtested on x86-64 Fedora 20.
2014-03-12 Tom Tromey <tromey@redhat.com>
* inf-child.c (return_zero): New function.
(inf_child_target): Set to_can_async_p, to_supports_non_stop.
* aix-thread.c (aix_thread_inferior_created): New function.
(aix_thread_attach): Remove.
(init_aix_thread_ops): Don't set to_attach.
(_initialize_aix_thread): Register inferior_created observer.
* corelow.c (init_core_ops): Don't set to_attach or
to_create_inferior.
* exec.c (init_exec_ops): Don't set to_attach or
to_create_inferior.
* infcmd.c (run_command_1): Use find_run_target. Make direct
target calls.
(attach_command): Use find_attach_target. Make direct target
calls.
* record-btrace.c (init_record_btrace_ops): Don't set
to_create_inferior.
* record-full.c (record_full_can_async_p, record_full_is_async_p):
Remove.
(init_record_full_ops, init_record_full_core_ops): Update. Don't
set to_create_inferior.
* target.c (complete_target_initialization): Add assertion.
(target_create_inferior): Remove.
(find_default_attach, find_default_create_inferior): Remove.
(find_attach_target, find_run_target): New functions.
(find_default_is_async_p, find_default_can_async_p)
(target_supports_non_stop, target_attach): Remove.
(init_dummy_target): Don't set to_create_inferior or
to_supports_non_stop.
* target.h (struct target_ops) <to_attach>: Add comment. Remove
TARGET_DEFAULT_FUNC.
<to_create_inferior>: Add comment.
<to_can_async_p, to_is_async_p, to_supports_non_stop>: Use
TARGET_DEFAULT_RETURN.
<to_can_async_p, to_supports_non_stop, to_can_run>: Add comments.
(find_attach_target, find_run_target): Declare.
(target_create_inferior): Remove.
(target_has_execution_1): Update comment.
(target_supports_non_stop): Remove.
* target-delegates.c: Rebuild.
2014-03-12 Tom Tromey <tromey@redhat.com>
* gdb.base/corefile.exp (corefile_test_run, corefile_test_attach):
New procs. Add target-async tests.
* gdb.reverse/break-precsave.exp (precsave_tests): New proc.
Add target-async tests.
that if multiple bignum values are encountered only the most recent is valid.
If such expressions are cached, eg to be emitted into a literal pool later on
in the assembly, then only one expression - the last - will be correct. This
patch fixes the problem for the AArch64 target by caching each bignum value
locally.
PR gas/16688
* config/tc-aarch64.c (literal_expression): New structure.
(literal_pool): Replace exp array with literal_expression array.
(add_to_lit_pool): When adding a bignum cache the big value.
(s_ltorg): When emitting a bignum initialise the global bignum
array from the cached value.
* gas/aarch64/litpool.s: New test case.
* gas/aarch64/litpool.d: Expected disassembly.
On PPC64, 'func' and 'main' are function descriptors and don't point
to the actual code. Thus the usage of these symbols in the DWARF
assembler source was broken. The patch introduces new labels
func_start and func_end for this purpose.
A "side effect" of the migration to Dwarf::assemble is that the DWARF
address size is now automatically adjusted to the target architecture.
The original assembler source hard-coded the DWARF address size to 4,
even on 64-bit architectures. This address size mismatch caused a
test case failure on s390x due to a wrong result from DW_OP_deref.
Now that prepare_for_testing etc. can cope with absolute path names,
this can be exploited for test cases with generated source files.
This is just to simplify the code and shouldn't cause any functional
change.
Test cases that produce source files in the build directory have not
been able to use prepare_for_testing and friends. This was because
build_executable_from_specs unconditionally prepended the source
directory path name to its arguments.
an ARM ELF binary into an AARCH64 ELF executable.
PR ld/16671
* elf32-arm.c (elf32_arm_add_symbol_hook): Check for ARM format
before testing for vxworks.
option coming *after* the command option not before it.
PR binutils/16652
* doc/binutils.texi (ar cmdline): Move --plugin command line
option to after the command option.
Like inf-child.c, this file is no longer used exclusively by Unix
targets anymore.
gdb/
2014-03-12 Pedro Alves <palves@redhat.com>
* inf-child.h: Update comment to not mention Unix.
This file is no longer used exclusively by Unix targets anymore.
gdb/
2014-03-12 Pedro Alves <palves@redhat.com>
* inf-child.c: Update top comment to not mention Unix. Add
generic comment describing how this target is meant to be used.
(inf_child_post_attach, inf_child_post_startup_inferior)
(inf_child_follow_fork, inf_child_pid_to_exec_file): Don't mention
Unix in comment.
So that all native targets inherit a single "superclass".
Target methods that are set to or do the same as inf-child.c's are
removed.
Not tested.
gdb/
2014-03-12 Pedro Alves <palves@redhat.com>
* nto-procfs.c: Include inf-child.h.
(procfs_ops): Delete global.
(procfs_can_run): Delete method.
(procfs_detach, procfs_mourn_inferior): Unpush the passed in
target pointer instead of referencing procfs_ops.
(procfs_prepare_to_store): Delete.
(init_procfs_ops): Delete function.
(procfs_target): New function, based on init_procfs_ops, but
inherit inf_child_target.
(_initialize_procfs): Use procfs_target.
So that all native targets inherit a single "superclass".
Target methods that are set to or do the same as inf-child.c's are
removed.
Tested by cross building on Fedora 17, and then confirming that
./gdb.exe ./gdb.exe -ex "set pagination off" -ex "start"
under Wine still works.
Also, Joel tested this with Adacore's internal testsuite.
gdb/
2014-03-12 Pedro Alves <palves@redhat.com>
* windows-nat.c: Include inf-child.h.
(windows_ops): Delete global.
(windows_open, windows_prepare_to_store, windows_can_run): Delete
methods.
(init_windows_ops): Delete function.
(windows_target): New function, based on init_windows_ops, but
inherit inf_child_target.
(_initialize_windows_nat): Use windows_target. Install x86
specific target methods here.
copy_elf_program_header has logic to reject non-alloc sections when
calculating p_vaddr offset for padding, but blithely assumed the
first section in a segment was allocated.
PR 16690
* elf.c (copy_elf_program_header): Ignore first section lma if
non-alloc.
coffcode.h uses an intptr_t cast inside an #ifdef RS6000COFF_C, so
ensure that intptr_t is defined. We don't see this when
cross-compiling from linux due to intptr_t being provided by
unistd.h.
PR 16686
* coff-rs6000.c: Include stdint.h.
* coff64-rs6000.c: Likewise.
This fix is simple:
msp430-sim.c: In function 'maybe_perform_syscall':
msp430-sim.c:898:10: warning: format '%d' expects argument of type 'int',
but argument 5 has type 'long int' [-Wformat]
This one we change to use casts like everyone else does in the code base:
msp430-sim.c: In function 'msp430_step_once':
msp430-sim.c:985:7: warning: passing argument 3 of 'init_disassemble_info'
from incompatible pointer type [enabled by default]
include/dis-asm.h:368:13: note: expected 'fprintf_ftype' but argument is
of type 'int (*)(struct FILE * __restrict__, const char * __restrict__)'
It is rare for people to want to modify the cmd arg. In general, they
really shouldn't be, but a few still do. For those who misbehave, dupe
the string locally so they can bang on it.
2014-03-10 Sasa Stankovic <Sasa.Stankovic@imgtec.com>
gold/
* symtab.cc (Symbol_table::sized_write_globals): Allow a target to
adjust dynamic symbol value.
* target.h (Target::adjust_dyn_symbol): New function.
(Target::do_adjust_dyn_symbol): New function.
2014-03-10 Sasa Stankovic <Sasa.Stankovic@imgtec.com>
gold/
* output.cc (Output_data_dynamic::Dynamic_entry::write):
Get the value of DYNAMIC_CUSTOM dynamic entry.
* output.h (Output_data_dynamic::add_custom): New function.
(Dynamic_entry::Dynamic_entry): New constructor for DYNAMIC_CUSTOM
dynamic entry.
(enum Dynamic_entry::Classification): Add DYNAMIC_CUSTOM.
* target.h (Target::dynamic_tag_custom_value): New function.
(Target::do_dynamic_tag_custom_value): New function.
2014-03-10 Sasa Stankovic <Sasa.Stankovic@imgtec.com>
gold/
* symtab.cc (Symbol_table::set_dynsym_indexes): Allow a target to set
dynsym indexes.
* target.h (Target::has_custom_set_dynsym_indexes): New function.
(Target::do_has_custom_set_dynsym_indexes): New function.
(Target::set_dynsym_indexes): New function.
(Target::do_set_dynsym_indexes): New function.
When evaluating an expression, if it is of a tagged type, GDB reads
the tag in memory and deduces the full view. At parsing time, however,
this operation is done only in the case of OP_VAR_VALUE. ptype does
not go through a full evaluation of expressions so it may return some
odd results:
(gdb) print c.menu_name
$1 = 0x0
(gdb) ptype $
type = system.strings.string_access
(gdb) ptype c.menu_name
type = <void>
This change removes this peculiarity by extending the tag resolution
to UNOP_IND and STRUCTOP_STRUCT. As in the case of OP_VAR_VALUE, this
implies switching from EVAL_AVOID_SIDE_EFFECTS to EVAL_NORMAL when a
tagged type is dereferenced.
gdb/
* ada-lang.c (ada_evaluate_subexp): Resolve tagged types to
full view in the case of UNOP_IND and STRUCTOP_STRUCT.
gdb/testsuite/
* gdb.ada/tagged_access: New testcase.
This function is for simple breakpoint. So I post a patch to remove "hardware".
Thanks,
Hui
2014-03-10 Hui Zhu <hui@codesourcery.com>
* target.h (target_insert_breakpoint): Remove "hardware" from its
comments.
Similar to the powerpc64 patch, this improves overflow checking in
elf32-ppc.c. Many reloc "howto" entries needed fixes, some just
cosmetic.
The patch also fixes the R_PPC_VLE_SDA21 reloc application code, which
was horribly broken. In fact, it may still be broken since Power ISA
2.07 says e_li behaves as
RT <- EXTS(li20 1:4 || li20 5:8 || li20 0 || li20 9:19)
where li20 is a field taken from bits 17..20, 11..15, 21..31 of the
instruction. Freescale VLEPEM says differently, and I assume
correctly, that
RT <- EXTS(li20 0:3 || li20 4:8 || li20 9:19)
The VLE_SDA21 relocation description matches this too.
Now the VLE_SDA21 relocation specifies in the case where e_addi16 is
converted to e_li for symbols in .PPC.EMB.sdata0 or .PPC.EMB.sbss0
(no base register), that the field is restricted to 16 bits, with the
sign bit being propagated to the top 4 bits. I don't see the sense in
restricting the value like this, so have allowed the full 20 bit
signed value. This of course is compatible with the reloc description
in that values in the 16 bit signed range will result in exactly the
same insn field as when the reloc description is followed to the
letter.
* elf32-ppc.c (ppc_elf_howto_raw): Correct overflow check for
many relocations. Correct bitsize and rightshift too for a number
of VLE relocs. Describe R_PPC_VLE_SDA21 and R_PPC_VLE_SDA21_LO.
Correct dst_mask on R_PPC_VLE_SDA21_LO.
(ppc_elf_vle_split16): Tidy, delete unnecessary prototype.
(ppc_elf_relocate_section): Modify overflow test for 16-bit
fields in instructions to signed/unsigned according to whether
the field takes a signed or unsigned value. Tidy vle split16 code.
Correct R_PPC_VLE_SDA21 and R_PPC_VLE_SDA21_LO handling.
R_PPC64_ADDR16 is used in three contexts:
- .short data relocation
- 16-bit signed insn fields, eg. addi
- 16-bit unsigned insn fields, eg. ori
In the first case we want to allow both signed and unsigned 16-bit
values, the latter two ought to error if the field exceeds the range
of values allowed for 16-bit signed and unsigned integers
respectively. These conflicting requirements meant that ld had to
choose the least restrictive overflow checks, and thus it is possible
to construct testcases where an addi field overflows but is not
reported by ld. Many relocations dealing with 16-bit insn fields have
this problem. What's more, some relocations that are only ever used
for signed fields of instructions woodenly copied the lax overflow
checking of R_PPC64_ADDR16.
bfd/
* elf64-ppc.c (ppc64_elf_howto_raw): Use complain_overflow_signed
for R_PPC64_ADDR14, R_PPC64_ADDR14_BRTAKEN, R_PPC64_ADDR14_BRNTAKEN,
R_PPC64_SECTOFF, R_PPC64_ADDR16_DS, R_PPC64_SECTOFF_DS,
R_PPC64_REL16 entries. Use complain_overflow_dont for R_PPC64_TOC.
(ppc64_elf_relocate_section): Modify overflow test for 16-bit
fields in instructions to signed/unsigned according to whether
the field takes a signed or unsigned value.
gold/
* powerpc.cc (Powerpc_relocate_functions::Overflow_check): Add
CHECK_UNSIGNED, CHECK_LOW_INSN, CHECK_HIGH_INSN.
(Powerpc_relocate_functions::has_overflow_unsigned): New function.
(Powerpc_relocate_functions::has_overflow_bitfield,
overflowed): Use the above.
(Target_powerpc::Relocate::relocate): Correct overflow checking
for a number of relocations. Modify overflow test for 16-bit
fields in instructions to signed/unsigned according to whether
the field takes a signed or unsigned value.
* dwarf2read.c (read_cutu_die_from_dwo): Fix function comment.
Remove unused local comp_dir_attr. Assert exactly one of
stub_comp_unit_die, stub_comp_dir is non-NULL.
Expand a bit the comments to answer some questions I had when looking
at why a target of mine would not have some default methods set.
gdb/ChangeLog:
* target.h (complete_target_initialization, add_target):
Add comment.
So that all native targets inherit a single "superclass".
Target methods that are set to or do the same as inf-child.c's are
removed.
Tested by cross building on Fedora 17.
gdb/
2014-03-07 Pedro Alves <palves@redhat.com>
* go32-nat.c: Include inf-child.h.
(go32_ops): Delete global.
(go32_close, go32_detach, go32_prepare_to_store, go32_can_run):
Delete methods.
(go32_create_inferior): Push the passed in target pointer instead
of referencing go32_ops.
(init_go32_ops): Delete function. Moved parts to _initialize_go32_nat.
(go32_target): New function, based on init_go32_ops, but inherit
inf_child_target.
(_initialize_go32_nat): Use go32_target. Move parts of
init_go32_ops here.
I noticed GDB was failing to enable threading support for 32-bit AIX
cores. I traced it to failure to read variables from libpthreads.a.
The issue is that data for that library is loaded at a high address,
and bfd is sign extending the section addresses:
(gdb) info files
Symbols from "/home/palves/crash".
Local core dump file:
`/home/palves/core', file type aixcoff-rs6000.
0x2ff22000 - 0x2ff23000 is .stack
0x20000000 - 0x200316e0 is .data
0x20000e90 - 0x200016c0 is .data
0xfffffffff0254000 - 0xfffffffff0297920 is .data
0xfffffffff07b46a8 - 0xfffffffff07b47c8 is .data
0xfffffffff0298000 - 0xfffffffff029bfcc is .data
0xfffffffff06dafe0 - 0xfffffffff07b3838 is .data
Local exec file:
`/home/palves/crash', file type aixcoff-rs6000.
Entry point: 0x20001394
0x10000150 - 0x10000e90 is .text
0x20000e90 - 0x2000149c is .data
0x2000149c - 0x200016c0 is .bss
0xd053b124 - 0xd053e15f is .text in /usr/lib/libpthreads.a(shr_comm.o)
0xf0254000 - 0xf0297920 is .data in /usr/lib/libpthreads.a(shr_comm.o)
0xf0254450 - 0xf0297920 is .bss in /usr/lib/libpthreads.a(shr_comm.o)
0xd053a280 - 0xd053aabe is .text in /usr/lib/libcrypt.a(shr.o)
0xf07b46a8 - 0xf07b47c8 is .data in /usr/lib/libcrypt.a(shr.o)
0xf07b47c8 - 0xf07b47c8 is .bss in /usr/lib/libcrypt.a(shr.o)
0xd04fb180 - 0xd053917e is .text in /usr/lib/libpthreads.a(shr_xpg5.o)
0xf0298000 - 0xf029bfcc is .data in /usr/lib/libpthreads.a(shr_xpg5.o)
0xf029bf64 - 0xf029bfcc is .bss in /usr/lib/libpthreads.a(shr_xpg5.o)
0xd0100900 - 0xd04fa39c is .text in /usr/lib/libc.a(shr.o)
0xf06dafe0 - 0xf07b3838 is .data in /usr/lib/libc.a(shr.o)
0xf0751e94 - 0xf07b3838 is .bss in /usr/lib/libc.a(shr.o)
Notice:
...
0xfffffffff0298000 - 0xfffffffff029bfcc is .data
...
Those are the bfd section start/end addresses. It't not visible here:
...
0xf0298000 - 0xf029bfcc is .data in /usr/lib/libpthreads.a(shr_xpg5.o)
...
... just because GDB trims that number to 32-bit when printing.
GDB then fails to find the memory for libpthreads.a variables in the
core, and falls back to reading it directly from the executable (which
yields the values as originally initialized in the code).
E.g.:
(gdb) p &__n_pthreads
$2 = (<data variable, no debug info> *) 0xf074fda8 <__n_pthreads>
(gdb) p __n_pthreads
$1 = -1
That should have returned 2 instead of -1.
bfd/
2014-03-07 Pedro Alves <palves@redhat.com>
* rs6000-core.c (rs6000coff_core_p): Cast pointers to bfd_vma
through ptr_to_uint instead of through long.
