Give the function a better name (drop "maybe_") and update the header
comment.
gdb/ChangeLog:
* disasm.c (maybe_add_dis_line_entry): Rename to...
(add_dis_line_entry): ...this, and update header comment.
(do_mixed_source_and_assembly): Now use add_dis_line_entry.
Currently, even when built with --enable-build-with-cxx, gdb uses
CFLAGS instead of CXXFLAGS. This commit fixes it.
CXXFLAGS set in the environment when configure was run is now honored
in the generated gdb/Makefile, and you can also override CXXFLAGS in
the command like at make time, with the usual 'make CXXFLAGS="..."'
Objects built with a C compiler (e.g., gnulib) still honor CFLAGS
instead.
gdb/ChangeLog:
2016-01-21 Pedro Alves <palves@redhat.com>
* Makefile.in (COMPILER_CFLAGS): New.
(CXXFLAGS): Get it from configure.
(INTERNAL_CFLAGS_BASE, INTERNAL_LDFLAGS): Use COMPILER_CFLAGS
instead of CFLAGS.
* build-with-cxx.m4 (GDB_AC_BUILD_WITH_CXX): Set and AC_SUBST
COMPILER_CFLAGS.
* configure: Regenerate.
gdb/gdbserver/ChangeLog:
2016-01-21 Pedro Alves <palves@redhat.com>
* Makefile.in (COMPILER_CFLAGS, CXXFLAGS): New.
(INTERNAL_CFLAGS_BASE): Use COMPILER_CFLAGS instead of CFLAGS.
* configure: Regenerate.
PR ld/19455
* elf32-arm.c (elf32_arm_create_dynamic_sections): Set the ELF
class of the linker stub bfd.
(elf32_arm_check_relocs): Skip check for pic format after
processing a vxWorks R_ARM_ABS12 reloc.
* elflink.c (bfd_elf_final_link): Check for ELFCLASSNONE when
reporting a class mismatch.
* testsuite/ld-arm/vxworks1-lib.dd: Update for current
disassmebler output.
* testsuite/ld-arm/vxworks1-lib.rd: Likewise.
* testsuite/ld-arm/vxworks1.dd: Likewise.
* testsuite/ld-arm/vxworks1.rd: Likewise.
* testsuite/ld-arm/vxworks1.ld: Set the output format.
A relatively recent patch support for explicit locations, and part
of that patch cleaned up the way we parse breakpoint locations.
Unfortunatly, a small regression crept in for "*<EXPR>" breakpoint
locations. In particular, on PIE programs, one can see the issue by
doing the following, with any program:
(gdb) b *main
Breakpoint 1 at 0x51a: file hello.c, line 3.
(gdb) run
Starting program: /[...]/hello
Error in re-setting breakpoint 1: Warning:
Cannot insert breakpoint 1.
Cannot access memory at address 0x51a
Warning:
Cannot insert breakpoint 1.
Cannot access memory at address 0x51a
Just for the record, this regression was introduced by:
commit a06efdd6ef
Date: Tue Aug 11 17:09:35 2015 -0700
Subject: Explicit locations: introduce address locations
What happens is that the patch makes the implicit assumption that
the address computed the first time is static, as if it was designed
to only support litteral expressions (Eg. "*0x1234"). This allows
the shortcut of not re-computing the breakpoint location's address
when re-setting breakpoints.
However, this does not work in general, as demonstrated in the example
above.
This patch plugs that hole simply by saving the original expression
used to compute the address as part of the address location, so as
to then re-evaluate that expression during breakpoint re-set.
gdb/ChangeLog:
* location.h (new_address_location): Add new parameters
"addr_string" and "addr_string_len".
(get_address_string_location): Add declaration.
* location.c (new_address_location): Add new parameters
"addr_string" and "addr_string_len". If not NULL, store
a copy of the addr_string in the new location as well.
(get_address_string_location): New function.
(string_to_event_location): Update call to new_address_location.
* linespec.c (event_location_to_sals) <ADDRESS_LOCATION>:
Save the event location in the parser's state before
passing it to convert_address_location_to_sals.
* breakpoint.c (create_thread_event_breakpoint): Update call
to new_address_location.
(init_breakpoint_sal): Get the event location's string, if any,
and use it to update call to new_address_location.
* python/py-finishbreakpoint.c (bpfinishpy_init):
Update call to new_address_location.
* spu-tdep.c (spu_catch_start): Likewise.
* config/djgpp/fnchange.lst: Add entries for
gdb/testsuite/gdb.base/break-fun-addr1.c and
gdb/testsuite/gdb.base/break-fun-addr2.c.
gdb/testsuite/ChangeLog:
* gdb.base/break-fun-addr.exp: New file.
* gdb.base/break-fun-addr1.c: New file.
* gdb.base/break-fun-addr2.c: New file.
As defined at AArch64 ELF Specification (4.6.7 Call and Jump
relocations), symbol with type of non STT_FUNC but in different input
section with relocation place should insert long branch veneer also.
Meanwhile the current long branch veneer infrastructure havn't considered
the situation where the branch destination is "sym_value + rela->addend".
This was OK because we only insert veneer for long call destination is
STT_FUNC symbol for which the addend is always zero. But as we relax the
support to other situations by this patch, we need to handle addend be
non-zero value. For example, for static function, relocation against
"local symbol" are turned into relocation against "section symbol + offset"
where there is a valid addend.
bfd/
* elfnn-aarch64.c (aarch64_type_of_stub): Allow insert long branch
veneer for sym_sec != input_sec.
(elfNN_aarch64_size_stub): Support STT_SECTION symbol.
(elfNN_aarch64_final_link_relocate): Take rela addend into account when
calculation destination.
ld/
* testsuite/ld-aarch64/farcall-section.d: Delete.
* testsuite/ld-aarch64/farcall-section.s: Delete.
* testsuite/ld-aarch64/farcall-b-section.d: New expectation file.
* testsuite/ld-aarch64/farcall-bl-section.d: Likewise.
* testsuite/ld-aarch64/farcall-b-section.s: New testcase.
* testsuite/ld-aarch64/farcall-bl-section.s: Likewise.
* testsuite/ld-aarch64/aarch64-elf.exp: Likewise.
Since it will sometime have more than just prpsinfo. Also,
elf32-ppc.c needlessly includes this header.
* elf-linux-core.h: Rename from elf-linux-psinfo.h.
* elf.c: Adjust #include.
* elf32-ppc.c: Don't #include elf-linux-psinfo.h
* Makefile.am (SOURCE_HFILES): Update.
* Makefile.in: Regenerate.
* po/SRC-PORFILES.in: Regenerate.
Corefile code should be moved after running config.bfd, because it
uses want64.
* configure.ac: Move corefile selection later in file. Move
tdefaults code immediately after other target vector code.
* configure: Regenerate.
This patch fixes the following regression introduced by commit d0e59a68
step^M
39 } /* handler */^M
1: x/i $pc^M
=> 0x8740 <handler+80>: sub sp, r11, #0^M
(gdb) step^M
^M
Program received signal SIGSEGV, Segmentation fault.^M
setitimer () at ../sysdeps/unix/syscall-template.S:81^M
81 ../sysdeps/unix/syscall-template.S: No such file or directory.^M
1: x/i $pc^M
=> 0xb6eff9c0 <setitimer>: push {r7}^M
(gdb) FAIL: gdb.base/sigstep.exp: continue to handler, si+advance in handler, step from handler: leave handler
in my test setting, program is compiled in arm mode, but the glibc
is built in thumb mode, so when we do 'step' to step over syscall
instruction svc for SIGRETURN, GDB should set breakpoint for arm mode
in the program, even though the current program in glibc is in thumb
mode. Current GDB doesn't consider the case that the mode of program
SIGRETURN goes to can be different from current program mode.
In fact, GDB has taken care of this arm/thumb mode changes already,
see
/* Copy the value of next pc of sigreturn and rt_sigrturn into PC,
return 1. In addition, set IS_THUMB depending on whether we
will return to ARM or Thumb code. Return 0 if it is not a
rt_sigreturn/sigreturn syscall. */
static int
arm_linux_sigreturn_return_addr (struct frame_info *frame,
unsigned long svc_number,
CORE_ADDR *pc, int *is_thumb)
but in the commit d0e59a68
> - arm_linux_sigreturn_return_addr (frame, svc_number, &return_addr, &is_thumb);
> + if (svc_number == ARM_SIGRETURN || svc_number == ARM_RT_SIGRETURN)
> + next_pc = arm_linux_sigreturn_next_pc (regcache, svc_number);
the IS_THUMB setting is lost, so it is a regression.
gdb:
2016-01-21 Yao Qi <yao.qi@linaro.org>
* arm-linux-tdep.c (arm_linux_sigreturn_next_pc): Add parameter
is_thumb and set it according to CPSR saved on the stack.
(arm_linux_get_next_pcs_syscall_next_pc): Pass is_thumb to
arm_linux_sigreturn_next_pc.
gdb/gdbserver:
2016-01-21 Yao Qi <yao.qi@linaro.org>
* linux-arm-low.c (arm_sigreturn_next_pc): Add parameter
is_thumb and set it according to CPSR saved on the stack.
(get_next_pcs_syscall_next_pc): Pass is_thumb to
arm_sigreturn_next_pc.
The lambda function used to sort the enumerator list does not work
properly. This list consists of tuples, (enum label, enum value). The
key function returns x.enumval. enumval not being defined for a tuple,
we see this exception in the test log:
Python Exception <class 'AttributeError'> 'tuple' object has no attribute 'enumval'
The function should return the second item of the tuple, which is the
enumval.
The pretty-printer still worked mostly correctly, except that the
enumeration values were not sorted. The test still passed because the
enumeration values are already sorted where they are defined. The test
also passed despite the exception being printed, because the right output
was printed after the exception:
print (enum flag_enum) (FLAG_1)
Python Exception <type 'exceptions.AttributeError'> 'tuple' objecthas no attribute 'enumval':M
$7 = 0x1 [FLAG_1]
(gdb) PASS: gdb.python/py-pp-maint.exp: print FLAG_1
New in v2:
- Improved test case, I stole Pedro's example directly. It verifies
that the sorting of enumerators by value works, by checking that
printing FOO_MASK appears as FOO_1 | FOO_2 | FOO_3.
I noticed that I could change the regexps to almost anything and the
tests would still pass. I think it was because of the | in there. I
made them more robust by using string_to_regexp. I used curly braces
{ } instead of quoting marks " " for strings, so that I could use
square brackets [ ] in them without having to escape them all. I also
removed the "message" part of the tests, since they are redundant with
the command, and it's just more maintenance to have to update them.
Tested with Python 2.7 and 3.5.
gdb/ChangeLog:
* python/lib/gdb/printing.py (FlagEnumerationPrinter.__call__):
Fix enumerators sort key function.
gdb/testsuite/ChangeLog:
* gdb.python/py-pp-maint.exp: Change/add enum flag tests.
* gdb.python/py-pp-maint.c (enum flag_enum): Use more complex
enum flag values.
The gnu_vector test case yields a new FAIL on s390x:
FAIL: gdb.base/gnu_vector.exp: verify vector return value
It was introduced by commit 77ae9c1933 "gdb.base/gnu_vector.exp:
Don't test output from the inferior". That commit dropped the special
handling for GDB's inability (on some targets) to set the return value.
This change re-establishes the logic from before, converting the above
FAIL to a KFAIL (PRMS gdb/8549).
gdb/testsuite/ChangeLog:
* gdb.base/gnu_vector.exp: Re-establish handling for should_kfail
when GDB can not set the vector return value. Add more comments
for clarification.
On Ubuntu 14.04 the following failure would be seen when running the
tspeed.exp test on a target that supports fast tracepoints like x86_64:
Target returns error code '.In-process agent library not loaded in
process. Fast and static tracepoints unavailable.'.
(gdb) FAIL: gdb.trace/tspeed.exp: start trace experiment
This is because the default is to link with --as-needed and the
gdb_compile for the test is using the libs argument instead of shlib which
corrects this issue since 6ebea266fd by
adding -Wl,--no-as-needed.
This patch fixes the issue by passing the lib as the shlib argument to
gdb_compile.
Tested on Ubuntu 14.04 x86_64.
gdb/testsuite/ChangeLog:
* gdb.trace/tspeed.exp: Use shlib instead of libs in gdb_compile
command.
PR 19457
* testsuite/ld-scripts/script.exp (extract_symbol_test): Stop test
early for PE based targets.
* testsuite/ld-scripts/align.t: Use 0x1000 as VMA alignment.
* testsuite/ld-pe/tlssec32.d: Allow for relocatable output.
In the instruction to write to the ARMv8.2 PSTATE field UAO,
MSR UAO, #<imm>, the immediate should be either 0 or 1 but GAS accepts
any unsigned 4-bit integer.
This patch implements the constraint on the immediate, generating an
error if the immediate operand is invalid, and adds tests for the
illegal forms.
opcodes/
2016-01-20 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-opc.c (operand_general_constraint_met_p): Check validity
of MSR UAO immediate operand.
gas/
2016-01-20 Matthew Wahab <matthew.wahab@arm.com>
* testsuite/gas/aarch64/armv8_2-a-illegal.d: New.
* testsuite/gas/aarch64/armv8_2-a-illegal.l: New.
* testsuite/gas/aarch64/armv8_2-a-illegal.s: New.
Change-Id: Ibdec4967c00b1ef3be9dbc43d23b2c70d1a0b28c
bfd/ChangeLog:
* elf32-arm.c ((elf32_arm_special_sections): Remove catch of noread
section using '.text.noread' pattern.
gas/ChangeLog:
* config/obj-elf.c (obj_elf_change_section) : Allow arm section with
SHF_ARM_NOREAD section flag.
* config/tc-arm.h (md_elf_section_letter) : Implement this hook to
handle letter 'y'.
(arm_elf_section_letter) : Declare it.
* config/tc-arm.c (arm_elf_section_letter): Handle letter 'y' to set
SHF_ARM_NOREAD section flag.
* doc/c-arm.texi (ARM section attribute 'y'): Document it.
gas/testsuite/ChangeLog:
* gas/arm/section-execute-only.s: New test case.
* gas/arm/section-execute-only.d: Expected output.
ld/testsuite/ChangeLog:
* ld-arm/thumb1-noread-not-present-mixing-two-section.s: Add 'y'
attribute usage.
* ld-arm/thumb1-noread-present-one-section.s: Likewise.
* ld-arm/thumb1-noread-present-two-section.s: Likewise.
* ld-arm/thumb1-input-section-flag-match.s: Likewise.
binutils/ChangeLog:
* readelf.c (get_elf_section_flags): Display y letter for section
with SHF_ARM_NOREAD section flag in readelf section output.
(process_section_headers): Add y letter in readelf section output
key mapping for ARM architecture.
(GNU Coding Standard...)
gdb/ChangeLog:
* printcmd.c (print_scalar_formatted): move binary operator from
end of line to beginning of next line. Adjust formatting
accordingly.
gdb/ChangeLog:
* fbsd-tdep.c (find_stop_signal): Remove.
(struct fbsd_collect_regset_section_cb) <lwp>: New field.
<stop_signal>: New field.
<abort_iteration>: New field.
(fbsd_collect_regset_section_cb): Use new fields.
(fbsd_collect_thread_registers): New function.
(struct fbsd_corefile_thread_data): New structure.
(fbsd_corefile_thread): New function.
(fbsd_make_corefile_notes): Use new function to dump notes for each
non-exited thread in a process.
Older versions of FreeBSD supported userland threading via a pure
user-space threading library (N threads scheduled on 1 process) and
a N:M model (N threads scheduled on M LWPs). However, modern FreeBSD
versions only support a M:M threading model where each user thread is
backed by a dedicated LWP. This thread target only supports this
threading model. It also uses ptrace to query and alter LWP state
directly rather than using libthread_db to simplify the implementation.
FreeBSD recently gained support for reporting LWP events (birth and death
of LWPs). GDB will use LWP events when present. For older systems it
fetches the list of LWPs in the to_update_thread_list target op to update
the list of threads on each stop.
This target supports scheduler locking by using ptrace to suspend
individual LWPs as necessary before resuming a process.
gdb/ChangeLog:
* configure.ac: Check for support for LWP names on FreeBSD.
* fbsd-nat.c [PT_LWPINFO] New variable debug_fbsd_lwp.
[TDP_RFPPWAIT || HAVE_STRUCT_PTRACE_LWPINFO_PL_TDNAME]
(fbsd_fetch_kinfo_proc): Move function earlier.
[PT_LWPINFO] (fbsd_thread_alive): New function.
[PT_LWPINFO] (fbsd_pid_to_str): New function.
[HAVE_STRUCT_PTRACE_LWPINFO_PL_TDNAME] (fbsd_thread_name): New function.
[PT_LWP_EVENTS] (fbsd_enable_lwp_events): New function.
[PT_LWPINFO] (fbsd_add_threads): New function.
[PT_LWPINFO] (fbsd_update_thread_list): New function.
[PT_LWPINFO] New variable super_resume.
[PT_LWPINFO] (resume_one_thread_cb): New function.
[PT_LWPINFO] (resume_all_threads_cb): New function.
[PT_LWPINFO] (fbsd_resume): New function.
(fbsd_remember_child): Save full ptid instead of plain pid.
(fbsd_is_child_pending): Return ptid of saved child process.
(fbsd_wait): Include lwp in returned ptid and switch to LWP ptid on
first stop.
[PT_LWP_EVENTS] Handle LWP events.
[TDP_RFPPWAIT] Include LWP in child ptid.
(fbsd_post_startup_inferior) [PT_LWP_EVENTS]: Enable LWP events.
(fbsd_post_attach) [PT_LWP_EVENTS]: Enable LWP events.
Add threads for existing processes.
(fbsd_nat_add_target) [PT_LWPINFO]: Set "to_thread_alive" to
"fbsd_thread_alive".
Set "to_pid_to_str" to "fbsd_pid_to_str".
[HAVE_STRUCT_PTRACE_LWPINFO_PL_TDNAME]: Set "to_thread_name" to
"fbsd_thread_name".
[PT_LWPINFO]: Set "to_update_thread_list" to "fbsd_update_thread_list".
Set "to_has_thread_control" to "tc_schedlock".
Set "to_resume" to "fbsd_resume".
(_initialize_fbsd_nat): New function.
* configure: Regenerate.
* config.in: Regenerate.
gdb/doc/ChangeLog:
* gdb.texinfo (Debugging Output): Document "set/show debug fbsd-lwp".
This allows gdb to fetch per-thread registers for multi-threaded FreeBSD
processes.
Export get_ptrace_pid() from inf-ptrace.c and use it to determine the PID
to pass to ptrace in pan-BSD native targets. NetBSD and OpenBSD also accept
LWP IDs for ptrace requests to fetch per-thread state.
gdb/ChangeLog:
* amd64bsd-nat.c (amd64bsd_fetch_inferior_registers): Use
get_ptrace_pid.
(amd64bsd_store_inferior_registers): Use get_ptrace_pid.
(amd64bsd_dr_get): Use get_ptrace_pid.
(amd64bsd_dr_set): Use get_ptrace_pid.
* i386bsd-nat.c (i386bsd_fetch_inferior_registers): Use get_ptrace_pid.
(i386bsd_store_inferior_registers): Use get_ptrace_pid.
(i386bsd_dr_get): Use get_ptrace_pid.
(i386bsd_dr_set): Use get_ptrace_pid.
* inf-ptrace.c (get_ptrace_pid): Export.
* inf-ptrace.h (get_ptrace_pid): Declare.
* ppcfbsd-nat.c (ppcfbsd_fetch_inferior_registers): Use lwp id.
(ppcfbsd_store_inferior_registers): Use lwp id.
Display the LWP ID of each thread in a FreeBSD core. Extract thread
names from the per-thread THRMISC note.
gdb/ChangeLog:
* fbsd_tdep.c (fbsd_core_pid_to_str): New function.
(fbsd_core_thread_name): New function.
(fbsd_init_abi): Add "core_pid_to_str" gdbarch method.
Add "core_thread_name" gdbarch method.
Add a new gdbarch method to extract a thread name from a core for a
given thread. Use this new method in core_thread_name to implement the
to_thread_name target op.
gdb/ChangeLog:
* corelow.c (core_thread_name): New function.
(init_core_ops): Use "core_thread_name" for the "to_thread_name"
target op.
* gdbarch.sh (core_thread_name): New gdbarch callback.
* gdbarch.h: Re-generate.
* gdbarch.c: Re-generate.
Using Python 3.5 (I assume it's the same with 3.4 and lower, but I didn't
test), I see this:
print (enum flag_enum) (FLAG_1)^M
Python Exception <class 'TypeError'> %x format: an integer is required, not gdb.Value: ^M
$7 = ^M
(gdb) FAIL: gdb.python/py-pp-maint.exp: print FLAG_1
Apparently, this idiom, where v is a gdb.Value, was possible with Python 2,
but not with Python 3:
'%x' % v
In Python 2, it would automatically get converted to an integer. To solve
it, I simply added wrapped v in a call to int().
'%x' % int(v)
In Python 2, the int type is implemented with a "long" in C, so on x86-32 it's
32-bits. I was worried that doing int(v) would truncate the value and give
wrong results for enum values > 32-bits. However, the int type != the int
function. The int function does the right thing, selecting the right integer
type for the given value. I tested with large enum values on x86-32 and
Python 2, and everything works as expected.
gdb/ChangeLog:
* python/lib/gdb/printing.py (_EnumInstance.to_string): Explicitly
convert gdb.Value to integer type using int().
By default, if a test driver (a test .exp) ends with an uncaught
error/exception, the runtest command will still have a return code of 0
(success). However, if a test (or the environment) is broken and does
not work properly, it should be considered as failed so that we can
notice it and fix it.
Passing the --status flag to runtest will make it return an error if one
of the test it runs ends up with an uncaught error.
gdb/testsuite/ChangeLog:
* Makefile.in (check-single): Pass --status to runtest.
(check/%.exp): Likewise.
When using the check-parallel target, the return code of make is always 0,
regardless of test results. This patch makes it return the same code as
the "make do-check-parallel" sub-command. So if there is a FAIL somewhere,
non-zero will be returned by make.
For the sake of example, I introduced a failure in gdb.base/break.exp.
$ make check-single TESTS="gdb.base/break.exp gdb.python/py-value.exp" && echo 'Success :D' || echo 'Fail :('
...
FAIL: gdb.base/break.exp: allo
...
Fail :(
I think the parallel run should do the same. Currently:
$ make check-parallel TESTS="gdb.base/break.exp gdb.python/py-value.exp" && echo 'Success :D' || echo 'Fail :('
...
FAIL: gdb.base/break.exp: allo
...
Success :D
And with the patch (no big surprises there):
$ make check-parallel TESTS="gdb.base/break.exp gdb.python/py-value.exp" && echo 'Success :D' || echo 'Fail :('
...
FAIL: gdb.base/break.exp: allo
...
Fail :(
What do you think?
gdb/testsuite/ChangeLog:
* Makefile.in (check-parallel): Propagate return code from make
do-check-parallel.
Include <sys/types.h> as a prerequisite for <machine/reg.h> when checking
for the r_fs and r_gs members in struct reg. Note that the previous test
for <machine/reg.h> already includes <sys/types.h> as a prerequisite.
gdb/ChangeLog:
* configure.ac: Include <sys/types.h when checking for "r_fs" in
"struct reg".
* configure: Regenerate.
bfd/ChangeLog:
* arc-plt.def: New file.
* arc-plt.h: Likewise.
* elf32-arc.c (elf_arc_abs_plt0_entry, elf_arc_abs_pltn_entry,
elf_arcV2_abs_plt0_entry, elf_arcV2_abs_pltn_entry,
elf_arc_pic_plt0_entry, elf_arc_pic_pltn_entry,
elf_arcV2_pic_plt0_entry, elf_arcV2_pic_pltn_entry): Remove.
(name_for_global_symbol): Added.
(ADD_RELA): Helper to create dynamic relocs.
(new_got_entry_to_list): Create a new got entry in linked list.
(symbol_has_entry_of_type): Search for specific type of entry in
list.
(is_reloc_for_GOT): return FALSE for any TLS related relocs.
(is_reloc_for_TLS, arc_elf_set_private_flags)
(arc_elf_print_private_bfd_data, arc_elf_copy_private_bfd_data)
(arc_elf_merge_private_bfd_data): New functions.
(debug_arc_reloc): Cleaned debug info printing.
(PDATA reloc): Changed not to perform address alignment.
(reverse_me): Added. Fix for ARC_32 relocs.
(arc_do_relocation): Return bfd_reloc_of when no relocation should
occur.
(arc_get_local_got_ents): Renamed from arc_get_local_got_offsets.
Changed function to access an array of list of GOT entries instead
of just an array of offsets.
(elf_arc_relocate_section): Added support for PIC and TLS related relocations.
(elf_arc_check_relocs): Likewise.
(elf_arc_adjust_dynamic_symbol, elf_arc_finish_dynamic_symbol,
(elf_arc_finish_dynamic_sections): Likewise
(arc_create_dynamic_sections): Modified conditions to create
dynamic sections.
(ADD_SYMBOL_REF_SEC_AND_RELOC): New macro.
(plt_do_relocs_for_symbol, relocate_plt_for_symbol)
(relocate_plt_for_entry): Changed to support new way to define PLT
related code.
(add_symbol_to_plt): Likewise.
(arc_elf_link_hash_table_create): New function.
include/ChangeLog:
* elf/arc-reloc.def (ARC_32, ARC_GOTPC, ARC_TLS_GD_GOT)
(ARC_TLS_IE_GOT, ARC_TLS_DTPOFF, ARC_TLS_DTPOFF_S9, ARC_TLS_LE_S9)
(ARC_TLS_LE_32): Fixed formula.
(ARC_TLS_GD_LD): Use new special function.
* opcode/arc-func.h: Changed all the replacement
functions to clear the patching bits before doing an or it with the value
argument.
Currently, we always re-set all locations of all breakpoints. This
commit makes us re-set only locations of the current program space.
If we loaded symbols to a program space (e.g., "file" command or some
shared library was loaded), GDB must run through all breakpoints and
determine if any new locations need to be added to the breakpoint.
However, there's no reason to recreate locations for _other_ program
spaces, as those haven't changed.
Similarly, when we create a new inferior, through e.g., a fork, GDB
must run through all breakpoints and determine if any new locations
need to be added to the breakpoint. There's no reason to destroy the
locations of the parent inferior and other inferiors. We know those
won't change.
In addition to being inneficient, resetting breakpoints of inferiors
that are currently running is problematic, because:
- some targets can't read memory while the inferior is running.
- the inferior might exit while we're re-setting its breakpoints,
which may confuse prologue skipping.
I went through all the places where we call breakpoint_re_set, and it
seems to me that all can be changed to only re-set locations of the
current program space.
The patch that reversed threads order in "info threads" etc. happened
to make gdb.threads/fork-plus-thread.exp expose this problem when
testing on x86/-m32. The problem was latent and masked out by chance
by the code-cache:
https://sourceware.org/ml/gdb-patches/2016-01/msg00213.html
Tested on x86-64 F20, native (-m64/-m32) and extended-remote
gdbserver.
Fixes the regression discussed in the url above with --target_board=unix/-m32:
-FAIL: gdb.threads/fork-plus-threads.exp: detach-on-fork=off: inferior 1 exited
+PASS: gdb.threads/fork-plus-threads.exp: detach-on-fork=off: inferior 1 exited
-FAIL: gdb.threads/fork-plus-threads.exp: detach-on-fork=off: no threads left (timeout)
-FAIL: gdb.threads/fork-plus-threads.exp: detach-on-fork=off: only inferior 1 left (the program exited)
+PASS: gdb.threads/fork-plus-threads.exp: detach-on-fork=off: no threads left
+PASS: gdb.threads/fork-plus-threads.exp: detach-on-fork=off: only inferior 1 left
gdb/ChangeLog:
2016-01-19 Pedro Alves <palves@redhat.com>
* ax-gdb.c (agent_command_1): Adjust call to decode_line_full.
* break-catch-throw.c (re_set_exception_catchpoint): Pass the
current program space down to linespec decoding and breakpoint
location updating.
* breakpoint.c (parse_breakpoint_sals): Adjust calls to
decode_line_full.
(until_break_command): Adjust calls to decode_line_1.
(base_breakpoint_decode_location, bkpt_decode_location): Add
'search_pspace' parameter. Pass it along.
(bkpt_probe_create_sals_from_location): Adjust calls to
parse_probes.
(tracepoint_decode_location, tracepoint_probe_decode_location)
(strace_marker_decode_location): Add 'search_pspace' parameter.
Pass it along.
(all_locations_are_pending): Rewrite to take a breakpoint and
program space as arguments instead.
(hoist_existing_locations): New function.
(update_breakpoint_locations): Add 'filter_pspace' parameter. Use
hoist_existing_locations instead of always removing all locations,
and adjust to all_locations_are_pending change.
(location_to_sals): Add 'search_pspace' parameter. Pass it along.
Don't disable the breakpoint if there are other locations in
another program space.
(breakpoint_re_set_default): Adjust to pass down the current
program space as filter program space.
(decode_location_default): Add 'search_pspace' parameter and pass
it along.
(prepare_re_set_context): Don't switch program space here.
(breakpoint_re_set): Use save_current_space_and_thread instead of
save_current_program_space.
* breakpoint.h (struct breakpoint_ops) <decode_location>: Add
'search_pspace' parameter.
(update_breakpoint_locations): Add 'filter_pspace' parameter.
* cli/cli-cmds.c (edit_command, list_command): Adjust calls to
decode_line_1.
* elfread.c (elf_gnu_ifunc_resolver_return_stop): Pass the current
program space as filter program space.
* linespec.c (struct linespec_state) <search_pspace>: New field.
(create_sals_line_offset, convert_explicit_location_to_sals)
(parse_linespec): Pass the search program space down.
(linespec_state_constructor): Add 'search_pspace' parameter.
Store it.
(linespec_parser_new): Add 'search_pspace' parameter and pass it
along.
(linespec_lex_to_end): Adjust.
(decode_line_full, decode_line_1): Add 'search_pspace' parameter
and pass it along.
(decode_line_with_last_displayed): Adjust.
(collect_symtabs_from_filename, symtabs_from_filename): New
'search_pspace' parameter. Use it.
(find_function_symbols): Pass the search program space down.
* linespec.h (decode_line_1, decode_line_full): Add
'search_pspace' parameter.
* probe.c (parse_probes_in_pspace): New function, factored out
from ...
(parse_probes): ... this. Add 'search_pspace' parameter and use
it.
* probe.h (parse_probes): Add pspace' parameter.
* python/python.c (gdbpy_decode_line): Adjust.
* tracepoint.c (scope_info): Adjust.
This is fallout from f303dbd60d.
The testcases themselves are single-threaded, but they load the IPA library,
which injects a thread in the inferior - making them multithreaded.
This results in printing the thread numbers in breakpoint messages.
gdb/testsuite/ChangeLog:
* gdb.trace/ftrace.exp: Fix expected message on continue.
* gdb.trace/pending.exp: Fix expected message on continue.
* gdb.trace/trace-break.exp: Fix expected message on continue.
The POOL48A major opcode was defined in early revisions of the 64-bit
microMIPS ISA, has never been implemented, and was removed before the
64-bit microMIPS ISA specification[1] has been finalized.
This complements commit a6c7053929 ("MIPS/opcodes: Remove microMIPS
48-bit LI instruction").
References:
[1] "MIPS Architecture for Programmers, Volume II-B: The microMIPS64
Instruction Set", MIPS Technologies, Inc., Document Number: MD00594,
Revision 3.06, October 17, 2012, Table 6.2 "microMIPS64 Encoding of
Major Opcode Field", p. 578
gas/
* config/tc-mips.c (micromips_insn_length): Remove the mention
of 48-bit microMIPS instructions.
gdb/
* mips-tdep.c (mips_insn_size): Remove 48-bit microMIPS
instruction support.
(micromips_next_pc): Likewise.
(micromips_scan_prologue): Likewise.
(micromips_deal_with_atomic_sequence): Likewise.
(micromips_stack_frame_destroyed_p): Likewise.
(mips_breakpoint_from_pc): Likewise.
opcodes/
* mips-dis.c (print_insn_micromips): Remove 48-bit microMIPS
instruction support.
This fixes PR sim/19441. In the MIPS simulator the microMIPS
functions in micromips.igen were not predicated on the microMIPS
models. This was causing build issues for some target triples.
This patch sets all the microMIPS specific functions to only be built if
the micromips32, micromips64 or micromipsdsp models are used.
PR sim/19441
* micromips.igen (delayslot_micromips): Enable for `micromips32',
`micromips64' and `micromipsdsp' only.
(process_isa_mode): Enable for `micromips32' and `micromips64' only.
(do_micromips_jalr, do_micromips_jal): Likewise.
(compute_movep_src_reg): Likewise.
(compute_andi16_imm): Likewise.
(convert_fmt_micromips): Likewise.
(convert_fmt_micromips_cvt_d): Likewise.
(convert_fmt_micromips_cvt_s): Likewise.
(FMT_MICROMIPS): Likewise.
(FMT_MICROMIPS_CVT_D): Likewise.
(FMT_MICROMIPS_CVT_S): Likewise.
Fix a bug in `micromips_insn_at_pc_has_delay_slot' in instruction size
determination via `mips_insn_size'. In the microMIPS case the latter
function expects a lone 16-bit instruction word containing the major
opcode regardless of whether the opcode requires another 16-bit word to
follow, to form a complete 32-bit instruction. Code however passes the
16-bit word previously retrieved shifted left by 16 bits. Consequently
`mips_insn_size', which examines the low 16-bit only, always sees 0.
By pure coincidence a major opcode of 0 denotes a 32-bit instruction in
the microMIPS instruction set, so the size of 4 is always returned here,
and the following 16-bit word is then merged in the low 16 bits of the
instruction previously shifted by 16 bits. The resulting 32-bit value
is then passed to `micromips_instruction_has_delay_slot' for delay slot
presence determination. This function in turn first examines the high
16 bits of the instruction word received and ignores the low 16 bits for
16-bit instructions.
Consequently the only effect of this bug is an extraneous memory read
issued to retrieve a subsequent 16-bit word where a 16-bit instruction
is being examined. Which in turn may fail if the instruction is located
right at the end of a readable memory area, in which case the lack of a
delay slot will be reported to the caller, which may be incorrect.
This code is used in breakpoint maintenance, for delay slot avoidance,
so the bug would only trigger for the unlikely case of someone placing
a breakpoint in a delay slot of an instruction which is at the end of
readable memory. Which explains why the bug remained unnoticed so long.
gdb/
* mips-tdep.c (micromips_insn_at_pc_has_delay_slot): Pass
unshifted 16-bit microMIPS instruction word to `mips_insn_size'.
Those are unused since gdb_test_multiple was added, factoring out most
of the content of gdb_test.
gdb/testsuite/ChangeLog:
* lib/gdb.exp (gdb_test): Remove unused global references.
PR ld/19440
inc * coff/internal.h (internal_syment): Use int to hold section
number.
(N_UNDEF): Cast to int not short.
(N_ABS): Likewise.
(N_DEBUG): Likewise.
(N_TV): Likewise.
(P_TV): Likewise.
bfd PR ld/19440
* coff-rs6000.c (_bfd_xcoff_swap_sym_in): Sign extend external
section number into internal section number.
* coff64-rs6000.c (_bfd_xcoff64_swap_sym_in): Likewise.
* coffswap.h (coff_swap_sym_in): Likewise.
* peXXigen.c (_bfd_XXi_swap_sym_in): Likewise.
* coffcode.h (_coff_bigobj_swap_sym_in): Make sure that internal
section number field is big enough to hold the external value.
This commit changes GDB like this:
- Program received signal SIGINT, Interrupt.
+ Thread 1 "main" received signal SIGINT, Interrupt.
- Breakpoint 1 at 0x40087a: file threads.c, line 87.
+ Thread 3 "bar" hit Breakpoint 1 at 0x40087a: file threads.c, line 87.
... once the program goes multi-threaded. Until GDB sees a second
thread spawn, the output is still the same as before, per the
discussion back in 2012:
https://www.sourceware.org/ml/gdb/2012-11/msg00010.html
This helps non-stop mode, where you can't easily tell which thread hit
a breakpoint or received a signal:
(gdb) info threads
Id Target Id Frame
* 1 Thread 0x7ffff7fc1740 (LWP 19362) "main" (running)
2 Thread 0x7ffff7fc0700 (LWP 19366) "foo" (running)
3 Thread 0x7ffff77bf700 (LWP 19367) "bar" (running)
(gdb)
Program received signal SIGUSR1, User defined signal 1.
0x0000003616a09237 in pthread_join (threadid=140737353877248, thread_return=0x7fffffffd5b8) at pthread_join.c:92
92 lll_wait_tid (pd->tid);
(gdb) b threads.c:87
Breakpoint 1 at 0x40087a: file threads.c, line 87.
(gdb)
Breakpoint 1, thread_function1 (arg=0x1) at threads.c:87
87 usleep (1); /* Loop increment. */
The best the user can do is run "info threads" and try to figure
things out.
It actually also affects all-stop mode, in case of "handle SIG print
nostop":
...
Program received signal SIGUSR1, User defined signal 1.
Program received signal SIGUSR1, User defined signal 1.
Program received signal SIGUSR1, User defined signal 1.
Program received signal SIGUSR1, User defined signal 1.
...
The above doesn't give any clue that these were different threads
getting the SIGUSR1 signal.
I initially thought of lowercasing "breakpoint" in
"Thread 3 hit Breakpoint 1"
but then after trying it I realized that leaving "Breakpoint"
uppercase helps the eye quickly find the relevant information. It's
also easier to implement not showing anything about threads until the
program goes multi-threaded this way.
Here's a larger example session in non-stop mode:
(gdb) c -a&
Continuing.
(gdb) interrupt -a
(gdb)
Thread 1 "main" stopped.
0x0000003616a09237 in pthread_join (threadid=140737353877248, thread_return=0x7fffffffd5b8) at pthread_join.c:92
92 lll_wait_tid (pd->tid);
Thread 2 "foo" stopped.
0x0000003615ebc6ed in nanosleep () at ../sysdeps/unix/syscall-template.S:81
81 T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)
Thread 3 "bar" stopped.
0x0000003615ebc6ed in nanosleep () at ../sysdeps/unix/syscall-template.S:81
81 T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)
b threads.c:87
Breakpoint 4 at 0x40087a: file threads.c, line 87.
(gdb) b threads.c:67
Breakpoint 5 at 0x400811: file threads.c, line 67.
(gdb) c -a&
Continuing.
(gdb)
Thread 3 "bar" hit Breakpoint 4, thread_function1 (arg=0x1) at threads.c:87
87 usleep (1); /* Loop increment. */
Thread 2 "foo" hit Breakpoint 5, thread_function0 (arg=0x0) at threads.c:68
68 (*myp) ++;
info threads
Id Target Id Frame
* 1 Thread 0x7ffff7fc1740 (LWP 31957) "main" (running)
2 Thread 0x7ffff7fc0700 (LWP 31961) "foo" thread_function0 (arg=0x0) at threads.c:68
3 Thread 0x7ffff77bf700 (LWP 31962) "bar" thread_function1 (arg=0x1) at threads.c:87
(gdb) shell kill -SIGINT 31957
(gdb)
Thread 1 "main" received signal SIGINT, Interrupt.
0x0000003616a09237 in pthread_join (threadid=140737353877248, thread_return=0x7fffffffd5b8) at pthread_join.c:92
92 lll_wait_tid (pd->tid);
info threads
Id Target Id Frame
* 1 Thread 0x7ffff7fc1740 (LWP 31957) "main" 0x0000003616a09237 in pthread_join (threadid=140737353877248, thread_return=0x7fffffffd5b8) at pthread_join.c:92
2 Thread 0x7ffff7fc0700 (LWP 31961) "foo" thread_function0 (arg=0x0) at threads.c:68
3 Thread 0x7ffff77bf700 (LWP 31962) "bar" thread_function1 (arg=0x1) at threads.c:87
(gdb) t 2
[Switching to thread 2, Thread 0x7ffff7fc0700 (LWP 31961)]
#0 thread_function0 (arg=0x0) at threads.c:68
68 (*myp) ++;
(gdb) catch syscall
Catchpoint 6 (any syscall)
(gdb) c&
Continuing.
(gdb)
Thread 2 "foo" hit Catchpoint 6 (call to syscall nanosleep), 0x0000003615ebc6ed in nanosleep () at ../sysdeps/unix/syscall-template.S:81
81 T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)
I'll work on documentation next if this looks agreeable.
This patch applies on top of the star wildcards thread IDs series:
https://sourceware.org/ml/gdb-patches/2016-01/msg00291.html
For convenience, I've pushed this to the
users/palves/show-which-thread-caused-stop branch.
gdb/doc/ChangeLog:
2016-01-18 Pedro Alves <palves@redhat.com>
* gdb.texinfo (Threads): Mention that GDB displays the ID and name
of the thread that hit a breakpoint or received a signal.
gdb/ChangeLog:
2016-01-18 Pedro Alves <palves@redhat.com>
* NEWS: Mention that GDB now displays the ID and name of the
thread that hit a breakpoint or received a signal.
* break-catch-sig.c (signal_catchpoint_print_it): Use
maybe_print_thread_hit_breakpoint.
* break-catch-syscall.c (print_it_catch_syscall): Likewise.
* break-catch-throw.c (print_it_exception_catchpoint): Likewise.
* breakpoint.c (maybe_print_thread_hit_breakpoint): New function.
(print_it_catch_fork, print_it_catch_vfork, print_it_catch_solib)
(print_it_catch_exec, print_it_ranged_breakpoint)
(print_it_watchpoint, print_it_masked_watchpoint, bkpt_print_it):
Use maybe_print_thread_hit_breakpoint.
* breakpoint.h (maybe_print_thread_hit_breakpoint): Declare.
* gdbthread.h (show_thread_that_caused_stop): Declare.
* infrun.c (print_signal_received_reason): Print which thread
received signal.
* thread.c (show_thread_that_caused_stop): New function.
gdb/testsuite/ChangeLog:
2016-01-18 Pedro Alves <palves@redhat.com>
* gdb.base/async-shell.exp: Adjust expected output.
* gdb.base/dprintf-non-stop.exp: Adjust expected output.
* gdb.base/siginfo-thread.exp: Adjust expected output.
* gdb.base/watchpoint-hw-hit-once.exp: Adjust expected output.
* gdb.java/jnpe.exp: Adjust expected output.
* gdb.threads/clone-new-thread-event.exp: Adjust expected output.
* gdb.threads/continue-pending-status.exp: Adjust expected output.
* gdb.threads/leader-exit.exp: Adjust expected output.
* gdb.threads/manythreads.exp: Adjust expected output.
* gdb.threads/pthreads.exp: Adjust expected output.
* gdb.threads/schedlock.exp: Adjust expected output.
* gdb.threads/siginfo-threads.exp: Adjust expected output.
* gdb.threads/signal-command-multiple-signals-pending.exp: Adjust
expected output.
* gdb.threads/signal-delivered-right-thread.exp: Adjust expected
output.
* gdb.threads/sigthread.exp: Adjust expected output.
* gdb.threads/watchpoint-fork.exp: Adjust expected output.
Andrew Burgess
Pedro Alves
Nick Clifton
Joel Brobecker
Jiong Wang
Alan Modra
Yao Qi
GDB Administrator
Simon Marchi
Andreas Arnez
Antoine Tremblay
Matthew Wahab
Mickael Guene
John Baldwin
Simon Marchi
Miranda Cupertino
Marcin Kościelnicki
Maciej W. Rozycki
Andrew Bennett