This was originally exposed by the OpenJDK Python-based unwinder.
If an unwinder attempts to call parse_and_eval from within its
sniffing method, GDB's unwinding machinery enters infinite recursion.
However, parse_and_eval is a pretty reasonable thing to call, because
Python/Scheme-based unwinders will often need to read globals out of
inferior memory. The recursion happens because:
- get_current_frame() is called soon after the target stops.
- current_frame is NULL, and so we unwind it from the sentinel frame
(which is special and has level == -1).
- We reach get_prev_frame_if_no_cycle, which does cycle detection
based on frame id, and thus tries to compute the frame id of the new
frame.
- Frame id computation requires an unwinder, so we go through all
unwinder sniffers trying to see if one accepts the new frame (the
current frame).
- the unwinder's sniffer calls parse_and_eval().
- parse_and_eval depends on the selected frame/block, and if not set
yet, the selected frame is set to the current frame.
- get_current_frame () is called again. current_frame is still NULL,
so ...
gdb doesn't recurse forever simply because the Python unwinder
contains code to detect and stop the recursion itself. However, GDB
goes downhill from here, e.g., by showing the sentinel frame as
current frame (note the -1):
Breakpoint 1, ccc (arg=<unavailable>) at py-recurse-unwind.c:23
23 }
(gdb) bt
#-1 ccc (arg=<unavailable>) at py-recurse-unwind.c:23
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
That "-1" frame level comes from this:
if (catch_exceptions (current_uiout, unwind_to_current_frame,
sentinel_frame, RETURN_MASK_ERROR) != 0)
{
/* Oops! Fake a current frame? Is this useful? It has a PC
of zero, for instance. */
current_frame = sentinel_frame;
}
which is bogus. It's never correct to set the current frame to the
sentinel frame. The only reason this has survived so long is that
getting here normally indicates something wrong has already happened
before and we fix that. And this case is no exception -- it doesn't
really matter how precisely we managed to get to that bogus code (it
has to do with the the stash), because anything after recursion
happens is going to be invalid.
So the fix is to avoid the recursion in the first place.
Observations:
#1 - The recursion happens because we try to do cycle detection from
within get_prev_frame_if_no_cycle. That requires computing the
frame id of the frame being unwound, and that itself requires
calling into the unwinders.
#2 - But, the first time we're unwinding from the sentinel frame,
when we reach get_prev_frame_if_no_cycle, there's no frame chain
at all yet:
- current_frame is NULL.
- the frame stash is empty.
Thus, there's really no need to do cycle detection the first time we
reach get_prev_frame_if_no_cycle, when building the current frame.
So we can break the recursion by making get_current_frame call a
simplified version of get_prev_frame_if_no_cycle that results in
setting the current_frame global _before_ computing the current
frame's id.
But, we can go a little bit further. As there's really no reason
anymore to compute the current frame's frame id immediately, we can
defer computing it to when some caller of get_current_frame might need
it. This was actually how the frame id was computed for all frames
before the stash-based cycle detection was added. So in a way, this
patch reintroduces the lazy frame id computation, but unlike before,
only for the case of the current frame, which turns out to be special.
This lazyness, however, requires adjusting
gdb.python/py-unwind-maint.exp, because that assumes unwinders are
immediately called as side effect of some commands. I didn't see a
need to preserve the behavior expected by that test (all it would take
is call get_frame_id inside get_current_frame), so I adjusted the
test.
gdb/ChangeLog:
2016-09-05 Pedro Alves <palves@redhat.com>
PR backtrace/19927
* frame.c (get_frame_id): Compute the frame id if not computed
yet.
(unwind_to_current_frame): Delete.
(get_current_frame): Use get_prev_frame_always_1 to get the
current frame and assert that that always succeeds.
(get_prev_frame_if_no_cycle): Skip cycle detection if returning
the current frame.
gdb/testsuite/ChangeLog:
2016-09-05 Pedro Alves <palves@redhat.com>
PR backtrace/19927
* gdb.python/py-unwind-maint.exp: Adjust tests to not expect that
unwinders are immediately called as side effect of "source" or
"disable unwinder" commands.
* gdb.python/py-recurse-unwind.exp: Remove setup_kfail calls.
Tom Tromey [Mon, 22 Aug 2016 22:56:52 +0000 (16:56 -0600)]
Handle DW_OP_form_tls_address
Currently gdb supports DW_OP_GNU_push_tls_address, but not
DW_OP_form_tls_address. I think it would be better if the toolchain
as a whole moved to using the standard opcode, and the prerequisite to
this is getting gdb to recognize it.
GCC can sometimes emit DW_OP_form_tls_address for emultls targets. As
far as I know, nobody has ever tried this with gdb (since it wouldn't
work at all).
I don't think there's a major drawback to using a single opcode for
all targets, because computing the location of a thread-local is
already target specific.
This is PR gdb/11616.
I don't know how to write a test case for this; though it's worth
noting that there aren't explicit tests for DW_OP_GNU_push_tls_address
either -- and if I change GCC, these paths will be tested to the same
extent they are now.
Yao Qi [Fri, 2 Sep 2016 14:49:57 +0000 (15:49 +0100)]
[GDBserver] Replace "reinsert_breakpoint" with "single_step_breakpoint"
reinsert_breakpoint is used for software single step, so it is more
clear to rename it to single_step_breakpoint. This was pointed out in
the review https://sourceware.org/ml/gdb-patches/2016-05/msg00429.html
I don't rename "other_breakpoint" in this patch.
gdb/gdbserver:
2016-09-02 Yao Qi <yao.qi@linaro.org>
* linux-low.c: Replace "reinsert_breakpoints" with
"single_step_breakpoints". Replace "reinsert breakpoints"
with "single-step breakpoints".
* mem-break.c: Likewise.
* mem-break.h: Likewise.
Yao Qi [Fri, 2 Sep 2016 08:22:13 +0000 (09:22 +0100)]
Detect broken ptrace in gdb_skip_float_test
We recently found a ARM kernel ptrace bug
http://lists.infradead.org/pipermail/linux-arm-kernel/2016-May/431962.html
Details can be found in the comment in gdb_skip_float_test. We can
skip floating point tests if the kernel bug is detected.
This patch adds more code in gdb_skip_float_test to detect the broken
ptrace on arm-linux. Such detection should be done at the beginning
of the test, because it starts a fresh GDB, so change the test cases
to invoke gdb_skip_float_test at the beginning of test, and use its
return value afterwards.
Since gdb_skip_float_test becomes a gdb_caching_proc, so it can't
have an argument, this patch also removes argument "msg", which isn't
useful.
gdb/testsuite:
2016-09-02 Yao Qi <yao.qi@linaro.org>
* gdb.arch/arm-neon.exp: Skip it if gdb_skip_float_test returns
true.
* gdb.base/call-ar-st.exp: Invoke gdb_skip_float_test.
* gdb.base/call-rt-st.exp: Likewise.
* gdb.base/call-sc.exp: Invoke gdb_skip_float_test and use its
return value instead of gdb,skip_float_test.
* gdb.base/callfuncs.exp: Invoke gdb_skip_float_test.
(do_function_calls): Use its return value instead of
gdb,skip_float_test.
* gdb.base/finish.exp: Likewise.
* gdb.base/funcargs.exp: Likewise.
* gdb.base/return.exp: Likewise.
* gdb.base/return2.exp: Likewise.
* gdb.base/varargs.exp: Likewise.
* lib/gdb.exp (gdb_skip_float_test): Change it to
gdb_caching_proc. Detect the broken ptrace on arm-linux.
Share target_wait prototype between GDB and gdbserver
This commit moves the target_wait prototype from the GDB-specific
target.h header to the common target/target.h header. Then, it
creates a compatible implementation of target_wait on gdbserver using
the_target->wait, and adjusts the (only) caller (mywait function).
Pretty straightforward, no regressions introduced.
* target.c (target_wait): Mention that the function's prototype
can be found at target/target.h.
* target.h (target_wait): Move prototype from here...
* target/target.h (target_wait): ... to here.
Use target_continue{,_no_signal} instead of target_resume
This commit implements a new function, target_continue, on top of the
target_resume function. Then, it replaces all calls to target_resume
by calls to target_continue or to the already existing
target_continue_no_signal.
This is one of the (many) necessary steps needed to consolidate the
target interface between GDB and gdbserver. In particular, I am
interested in the impact this change will have on the unification of
the fork_inferior function (which I have been working on).
Tested on the BuildBot, no regressions introduced.
* server.c (start_inferior): New variable 'ptid'. Replace calls
to the_target->resume by target_continue{,_no_signal}, depending
on the case.
* target.c (target_stop_and_wait): Call target_continue_no_signal
instead of the_target->resume.
(target_continue): New function.
* fork-child.c (startup_inferior): Replace calls to target_resume
by target_continue{,_no_signal}, depending on the case.
* linux-nat.c (cleanup_target_stop): Call
target_continue_no_signal instead of target_resume.
* procfs.c (procfs_wait): Likewise.
* target.c (target_continue): New function.
* target/target.h (target_continue): New prototype.
2016-09-01 Thomas Preud'homme <thomas.preudhomme@arm.com>
bfd/
* elf32-arm.c (cmse_entry_fct_p): Store instruction encoding in an
array of bytes and use bfd_get_16 to interpret its encoding according
to endianness of target.
Alan Modra [Thu, 1 Sep 2016 05:26:52 +0000 (14:56 +0930)]
Don't treat .opd section specially when ELFv2
Fixes a gdb segfault if a section named .opd is found in ELFv2 binaries.
* elf64-ppc.c (synthetic_opd): New static var.
(compare_symbols): Don't treat symbols in .opd specially for ELFv2.
(ppc64_elf_get_synthetic_symtab): Likewise. Comment.
Antoine Tremblay [Wed, 31 Aug 2016 17:46:22 +0000 (13:46 -0400)]
Fix lwp_suspend/unsuspend imbalance in linux_wait_1
This patch fixes imbalanced lwp_suspend/unsuspend calls caused by the
premature choosing of another event for fairness.
select_event_lwp would switch the event before a call to
unsuspend_all_lwps, thus it would be called with the wrong event.
This caused an assertion failure: unsuspend LWP xx, suspended=-1 when
testing gdb.threads/non-stop-fair-events.exp with ARM range stepping in
GDBServer.
This patch moves the switch of event after the unsuspend/unstop calls.
No regressions, tested on ubuntu 14.04 ARMv7 and x86.
With gdbserver-native.
gdb/gdbserver/ChangeLog:
* linux-low.c (linux_wait_1): Move event switch after unsuspend_lwps.
Alan Modra [Wed, 31 Aug 2016 07:36:32 +0000 (17:06 +0930)]
[GOLD] Further tidy to powerpc can_add_to_stub_group
This patch adds a little more debug output, and replaces two variables
with one, tracking current max group size by group_size_ rather than
by has14_.
* powerpc.cc (class Stub_control): Delete stub14_group_size_
and has14_. Add group_size_.
(Stub_control::can_add_to_stub_group): Adjust to suit. Print
debug info when switching to adding sections before stubs.
Alan Modra [Tue, 30 Aug 2016 11:32:58 +0000 (21:02 +0930)]
PowerPC64, correct grouping of stubs for ld.bfd
Like 57f6d32d, this patch ensures that sections containing external
conditional branches limit the group size.
* elf64-ppc.c (group_sections): Delete stub14_group_size. Instead,
track max group size with a new "group_size" var that is reduced
by a factor of 1024 from the 24-bit branch size whenever a 14-bit
branch is seen.
Alan Modra [Wed, 31 Aug 2016 03:48:34 +0000 (13:18 +0930)]
PowerPC VLE sh_flags and p_flags
ELF section sh_flags SHF_PPC_VLE was being set based on arch/mach,
which meant all code sections in an object file has the flag or all
lacked it. We can do better than that. Only those code sections
where VLE is enabled ought to have the flag, allowing an object file
to contain both VLE and non-VLE code.
Also, ELF header p_flags PF_PPC_VLE wasn't being set, and segments
were being split unnecessarily.
bfd/
* elf32-ppc.c (ppc_elf_section_processing): Delete.
(elf_backend_section_processing): Don't define.
(ppc_elf_modify_segment_map): Set p_flags and mark valid. Don't
split on non-exec sections differing in SHF_PPC_VLE. When
splitting segments, mark size invalid.
gas/
* config/tc-ppc.c (md_assemble): Set sh_flags for VLE. Test
ppc_cpu rather than calling ppc_mach to determine VLE mode.
(ppc_frag_check, ppc_handle_align): Likewise use ppc_cpu.
Andreas Arnez [Tue, 30 Aug 2016 15:03:55 +0000 (17:03 +0200)]
Fix order of inferiors in "thread apply all"
This inserts missing parentheses in the calculation of the comparison
result between two different inferior numbers. The problem was found by
Philipp Rudo.
Nick Clifton [Tue, 30 Aug 2016 12:51:43 +0000 (13:51 +0100)]
Partially revert previous delta - move limit testing code to first scan over symbol file.
PR gprof/20499
* corefile.c (num_of_syms_in): Return an unsigned int.
Fail if the count exceeds the maximum possible allocatable size.
(core_create_syms_from): Exit early if num_of_syms_in returns a
failure code.
Nick Clifton [Tue, 30 Aug 2016 12:38:54 +0000 (13:38 +0100)]
Fix more potential seg-faults in gprof.
PR gprof/20499
* corefile.c (num_of_syms_in): Return an unsigned int.
(core_create_syms_from): Catch a possible integer overflow
computing the argument to xmalloc. Also allow for the possibility
that an integer overflow in num_of_syms_in means that less space
has been allocated than expected.
Alan Modra [Tue, 30 Aug 2016 02:28:05 +0000 (11:58 +0930)]
[GOLD] correct grouping of stubs
This patch rewrites the rather obscure can_add_to_stub_group, fixing
a problem with the handling of sections containing conditional
external branches. When a section group contains any such section,
the group size needs to be limited to a much smaller size than groups
with only non-conditional external branches.
PR 20523
* powerpc.cc (class Stub_control): Add has14_. Comment owner_.
(Stub_control::can_add_to_stub_group): Correct grouping of
sections containing 14-bit external branches. When returning
false, set state_ to reflect the fact that we have one section
for the next group. Rewrite most of function for clarity.
Add and expand comments.
(Target_powerpc::do_relax): Print stub group size retry in hex.
H.J. Lu [Mon, 29 Aug 2016 15:49:20 +0000 (08:49 -0700)]
Check the external compression header size
Since the internal compression header size can be bigger than the
external compression header size, we should check the external
compression header size.
* readelf.c (load_specific_debug_section): Check the external
compression header size.
H.J. Lu [Mon, 29 Aug 2016 15:12:59 +0000 (08:12 -0700)]
i386: Issue an error on non-PIC call to IFUNC in PIC object
On i386, IFUNC function must be called via PLT. Since PLT in PIC
object uses EBX register, R_386_PLT32 relocation must be used to
call IFUNC function even when IFUNC function is defined locally.
Linker should issue an error when R_386_PC32 relocation is used
to call IFUNC function.
Since PR ld/19784 tests doesn't use PLT relocation to local IFUNC
function, they are moved to the x86-64 test directory.
bfd/
PR ld/14961
PR ld/20515
* elf32-i386.c (elf_i386_check_relocs): Issue an error when
R_386_PC32 relocation is used to call IFUNC function in PIC
object.
ld/
PR ld/14961
PR ld/20515
* testsuite/ld-i386/i386.exp: Run pr20515.
* testsuite/ld-i386/pr20515.d: New file.
* testsuite/ld-i386/pr20515.s: Likewise.
* testsuite/ld-ifunc/ifunc-14a.s: Use R_386_PLT32 to call IFUNC
function.
* testsuite/ld-ifunc/ifunc-14c.s: Likewise.
* testsuite/ld-ifunc/ifunc-2-i386.s: Likewise.
* testsuite/ld-ifunc/ifunc-2-local-i386.s: Likewise.
* testsuite/ld-ifunc/ifunc.exp: Move PR ld/19784 tests to ...
* testsuite/ld-x86-64/x86-64.exp: Here.
* testsuite/ld-ifunc/pr19784a.c: Moved to ...
* testsuite/ld-x86-64/pr19784a.c: Here.
* testsuite/ld-ifunc/pr19784b.c: Moved to ...
* testsuite/ld-x86-64/pr19784b.c: Here.
* testsuite/ld-ifunc/pr19784c.c: Moved to ...
* testsuite/ld-x86-64/pr19784c.c: Here.
Han Shen [Fri, 26 Aug 2016 22:50:15 +0000 (15:50 -0700)]
PR gold/20529 - relaxing loop never ends.
gold/ChangeLog
2016-08-26 Han Shen <shenhan@google.com>
* powerpc.cc (Stub_table::min_size_threshold_): New member to
limit size.
(Stub_table::set_min_size_threshold): New member function.
(Stub_table::set_address_and_size): Add code to only allow size
increase.
(Target_powerpc::do_relax): Add code to record last size.
Commit 980aa3e6 was supposed to cure dyn_reloc counting problems, but
did the opposite. For PIC we count two types of dyn_reloc, those on
pc-relative relocs, and the total. If a sym needs pc-relative dyn
relocs then all the relocs are dynamic. If not, then only those that
are must_be_dyn_reloc are dynamic.
PR 20519
* elf64-ppc.c (pc_dynrelocs): New function.
(ppc64_elf_relocate_section): Use it and must_be_dyn_reloc to
handle pic dynamic relocs.
Reduce parameter list in bfd_elf32_arm_target_relocs
2016-08-26 Thomas Preud'homme <thomas.preudhomme@arm.com>
bfd/
* bfd-in.h (struct elf32_arm_params): Define.
(bfd_elf32_arm_set_target_relocs): Rename into ...
(bfd_elf32_arm_set_target_params): This. Use a struct
elf32_arm_params to pass all parameters but the bfd and bfd_link_info.
* bfd-in2.h: Regenerate.
* elf32-arm.c (bfd_elf32_arm_set_target_relocs): Rename into ...
(bfd_elf32_arm_set_target_params): This. Pass all values via a struct
elf32_arm_params rather than as individual parameters.
ld/
* emultempl/armelf.em (params): New static variable.
(thumb_entry_symbol, byteswap_code, target1_is_rel, target2_type,
fix_v4bx, use_blx, vfp11_denorm_fix, stm32l4xx_fix, fix_cortex_a8,
no_enum_size_warning, no_wchar_size_warning, pic_veneer,
merge_exidx_entries, fix_arm1176, cmse_implib): move as part of the
above new structure.
(arm_elf_before_allocation): Access static variable from the params
structure.
(gld${EMULATION_NAME}_finish): Likewise.
(arm_elf_create_output_section_statements): Likewise and pass the
address of that structure to bfd_elf32_arm_set_target_relocs instead
of the static variables.
(PARSE_AND_LIST_ARGS_CASES): Access static variable from the params
structure.
2016-08-26 Thomas Preud'homme <thomas.preudhomme@arm.com>
bfd/
* elf32-arm.c (elf32_arm_get_stub_entry): Assert that we don't access
passed the end of htab->stub_group array.
(elf32_arm_create_or_find_stub_sec): Likewise.
(elf32_arm_create_stub): Likewise.
2016-08-26 Thomas Preud'homme <thomas.preudhomme@arm.com>
gas/
* config/tc-arm.c (v7m_psrs): Add MSPLIM, PSPLIM, MSPLIM_NS,
PSPLIM_NS, PRIMASK_NS, BASEPRI_NS, FAULTMASK_NS, CONTROL_NS, SP_NS and
their lowecase counterpart special registers. Write register
identifier in hex.
* testsuite/gas/arm/archv8m-cmse-msr.s: Reorganize tests per
operation, special register and then case. Use different register for
each operation. Add tests for new special registers.
* testsuite/gas/arm/archv8m-cmse-msr-base.d: Adapt expected result
accordingly.
* testsuite/gas/arm/archv8m-cmse-msr-main.d: Likewise.
* testsuite/gas/arm/archv8m-main-dsp-4.d: Likewise.
opcodes/
* arm-dis.c (psr_name): Use hex as case labels. Add detection for
MSPLIM, PSPLIM, MSPLIM_NS, PSPLIM_NS, PRIMASK_NS, BASEPRI_NS,
FAULTMASK_NS, CONTROL_NS and SP_NS special registers.
Forgot to set should_relocate to TRUE in case of GOT and TLS relocations of
undefined symbols for shared libraries.
In dynamic libraries if symbol is not known the instruction relocation would
not be resolved to point to the respective .got entry.
A test was created to detect similar future mistakes.
bfd/ChangeLog:
Cupertino Miranda <cmiranda@synopsys.com>
* elf32-arc.c (elf_arc_relocate_section): Changed. Set should_relocate
to TRUE for GOT and TLS relocs.
ld/ChangeLog:
Cupertino Miranda <cmiranda@synopsys.com>
* ld/testsuite/ld-arc/tls_gd-01.s: Added a testcase for this patch.
* ld/testsuite/ld-arc/tls_gd-01.d: Likewise.
Fixed conditions related to dynamic relocs relative offset patching.
Added arc_link_hash_table to be able to always generate and track
.rela.bss section.
bfd/ChangeLog:
Cupertino Miranda <cmiranda@synopsys.com>
* elf-bfd.h: Added ARC_ELF_DATA to enum elf_target_id.
* elf32-arc.c (struct elf_arc_link_hash_entry): Added.
(struct elf_arc_link_hash_table): Likewise.
(elf_arc_link_hash_newfunc): Likewise.
(elf_arc_link_hash_table_free): Likewise.
(arc_elf_link_hash_table_create): Likewise.
(elf_arc_relocate_section): Fixed conditions related to dynamic
(elf_arc_check_relocs): Likewise.
(arc_elf_create_dynamic_sections): Added
(elf_arc_adjust_dynamic_symbol): Changed access to .rela.bss to be done
through the hash table.
Added support for ARC_SDA_12 reloc.
Fixed ARC_N32_ME.
Added ME (middle-endian) to ARC_SDA_12 reloc.
bfd/ChangeLog:
Cupertino Miranda <cmiranda@synopsys.com>
* reloc.c: Fixed type in ARC_SECTOFF relocations. Added ARC_SDA_12
relocation.
* bfd-in2.h: Regenerated from the previous changes.
* libbfd.h: Regenerated from the previous changes.
include/ChangeLog:
Cupertino Miranda <cmiranda@synopsys.com>
* elf/arc-reloc.def: Fixed relocation formula for N*, SDA, SDA_12,
SDA_16_LD*, S13_PCREL, N32_ME, SECTOFF_* relocations.
* opcode/arc-func.h (replace_disp12s): Added. Used for SDA_12 relocation.
Add support for stable secure gateway veneers addresses
2016-08-26 Thomas Preud'homme <thomas.preudhomme@arm.com>
bfd/
* bfd-in.h (bfd_elf32_arm_set_target_relocs): Add a new parameter for
the input import library bfd.
* bfd-in2.h: Regenerate.
* elf32-arm.c (struct elf32_arm_link_hash_table): New in_implib_bfd
and new_cmse_stub_offset fields.
(stub_hash_newfunc): Initialize stub_offset and stub_template_size to
-1.
(elf32_arm_add_stub): Likewise for stub_offset.
(arm_new_stubs_start_offset_ptr): New function.
(arm_build_one_stub): Only allocate a stub_offset if it is -1. Allow
empty SG veneers to have zero relocations.
(arm_size_one_stub): Only initialize stub size and template
information for non empty veneers. Do not update veneer section size
if veneer already has an offset.
(elf32_arm_create_stub): Return the stub entry pointer or NULL instead
of a boolean indicating success or failure.
(cmse_scan): Change stub_changed parameter into an integer pointer
parameter cmse_stub_created to count the number of stub created and
adapt to change of return value in elf32_arm_create_stub.
(cmse_entry_fct_p): New function.
(arm_list_new_cmse_stub): Likewise.
(set_cmse_veneer_addr_from_implib): Likewise.
(elf32_arm_size_stubs): Define cmse_stub_created, pass its address to
cmse_scan instead of that of cmse_stub_changed to compute the number
of stub created and use it to initialize stub_changed. Call
set_cmse_veneer_addr_from_implib after all cmse_scan. Adapt to change
of return value in elf32_arm_create_stub. Use
arm_stub_section_start_offset () if not NULL to initialize size of
secure gateway veneers section. Initialize stub_offset of Cortex-A8
erratum fix to -1. Use ret to hold return value.
(elf32_arm_build_stubs): Use arm_stub_section_start_offset () if not
NULL to initialize size of secure gateway veneers section. Adapt
comment to stress the importance of zeroing veneer section content.
(bfd_elf32_arm_set_target_relocs): Add new in_implib_bfd parameter to
initialize eponymous field in struct elf32_arm_link_hash_table.
ld/
* emultempl/armelf.em (in_implib_filename): Declare and initialize new
variable.
(arm_elf_create_output_section_statements): Open import input library
file for writing and pass resulting in_implib_bfd to
bfd_elf32_arm_set_target_relocs.
(PARSE_AND_LIST_PROLOGUE): Define OPTION_IN_IMPLIB option.
(PARSE_AND_LIST_LONGOPTS): Define --in-implib option.
(PARSE_AND_LIST_OPTIONS): Add help message for --in-implib option.
(PARSE_AND_LIST_ARGS_CASES): Handle new OPTION_IN_IMPLIB case.
* ld.texinfo (--cmse-implib): Update to mention --in-implib.
(--in-implib): Document new option.
* NEWS: Likewise.
* testsuite/ld-arm/arm-elf.exp
(Secure gateway import library generation): add --defsym VER=1 to gas
CLI.
(Secure gateway import library generation: errors): Likewise.
(Input secure gateway import library): New test.
(Input secure gateway import library: no output import library):
Likewise.
(Input secure gateway import library: not an SG input import library):
Likewise.
(Input secure gateway import library: earlier stub section base):
Likewise.
(Input secure gateway import library: later stub section base):
Likewise.
(Input secure gateway import library: veneer comeback): Likewise.
(Input secure gateway import library: entry function change):
Likewise.
* testsuite/ld-arm/cmse-implib.s: Add input import library testing.
* testsuite/ld-arm/cmse-implib.rd: Update accordingly.
* testsuite/ld-arm/cmse-new-implib.out: New file.
* testsuite/ld-arm/cmse-new-implib.rd: Likewise.
* testsuite/ld-arm/cmse-new-implib-no-output.out: Likewise.
* testsuite/ld-arm/cmse-new-implib-not-sg-in-implib.out: Likewise.
* testsuite/ld-arm/cmse-new-earlier-later-implib.out: Likewise.
* testsuite/ld-arm/cmse-new-comeback-implib.rd: Likewise.
* testsuite/ld-arm/cmse-new-wrong-implib.out: Likewise.
Andreas Arnez [Thu, 25 Aug 2016 17:13:57 +0000 (19:13 +0200)]
S390: Add support for core dump NOTE sections
This enhances the 32-bit and 64-bit s390 ELF backends with support for
reading and writing the core dump note sections NT_PRSTATUS and
NT_PRPSINFO. Byte swapping is done as appropriate, such that core files
can now be processed correctly on non-s390 platforms.
bfd/ChangeLog:
* elf32-s390.c (stdarg.h): New include.
(elf_s390_grok_psinfo): New function.
(elf_s390_write_core_note): New function.
(elf_backend_grok_psinfo): Declare backend hook.
(elf_backend_write_core_note): Likewise.
* elf64-s390.c (stdarg.h): New include.
(elf_s390_grok_prstatus): New function.
(elf_s390_grok_psinfo): New function.
(elf_s390_write_core_note): New function.
(elf_backend_grok_prstatus): Declare backend hook.
(elf_backend_grok_psinfo): Likewise.
(elf_backend_write_core_note): Likewise.
Andreas Arnez [Thu, 25 Aug 2016 17:13:57 +0000 (19:13 +0200)]
xtensa: Avoid designated inits, for C++ compliance
C++ does not officially support designators in initializer lists. Thus
some compilers may issue errors when encountering them. Modern versions
of GCC seem to allow them by default, as a GCC extension, even though
the GCC documentation explicitly states otherwise: "[...] This extension
is not implemented in GNU C++." But some older GCC versions (like
4.4.7) did indeed emit an error instead, like this:
.../gdb/xtensa-config.c:219: error: expected primary-expression before
‘.’ token
This patch removes the only such instance I've seen when building with
'--enable-targets=all'.
gdb/ChangeLog:
* xtensa-tdep.h (XTENSA_GDBARCH_TDEP_INSTANTIATE): Replace
designated initializer list by plain initializer list, for C++
compliance.
Alan Modra [Thu, 25 Aug 2016 06:41:40 +0000 (16:11 +0930)]
powerpc ld configure
This fixes a few problems in the powerpc ld configury.
1) Using powerpc*le-* to test for powerpcle and powerpc64le would
incorrectly match a target triple like powerpc-apple-elf.
2) In the same vein, *64* could match 64 in the user supplied MANUF-OS
part of the target triple.
3) tooldir vars were missing, and some target aliases would result in
ridiculous values for those tdir_* vars given.
4) Since 2016-04-25, BE support was added automatically when asking
for an LE target. If that is a good idea, then so is adding LE
support when asking for BE.
* configure.tgt (powerpc*-*-linux* et al): Rewrite, adding LE
support for BE. First output all target endian configury
values, then opposite endian. Handle more tooldirs. Fix
bogus matches with strings in MANUF-OS part of target triple.
GLIBC BZ#20311 [1] proc_service.h install patch also remove 'const'
attributes from ps_get_thread_area and comment #15 discuss why to remove
the const attribute (basically since it a callback with the struct
ps_prochandle owned by the client it should be able to modify it if
it the case).
On default build this is not the issue and current g++ does not trigger
any issue with this mismatch declaration. However, on some bootstrap
build configuration where gdbserver is build with gcc instead this
triggers:
error: conflicting types for 'ps_get_thread_area'
This patch fixes it by syncing the declaration with GLIBC.
Breakpoint 1, ccc (arg=<unavailable>) at py-recurse-unwind.c:23
23 }
(gdb) bt
#-1 ccc (arg=<unavailable>) at py-recurse-unwind.c:23
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
[I've shortened pathnames for easier reading.]
The desired / expected behavior looks like this:
(gdb) source py-recurse-unwind.py
Python script imported
(gdb) b ccc
Breakpoint 1 at 0x4004bd: file py-recurse-unwind.c, line 23.
(gdb) run
Starting program: py-recurse-unwind
Breakpoint 1, ccc (arg=789) at py-recurse-unwind.c:23
23 }
(gdb) bt
#0 ccc (arg=789) at py-recurse-unwind.c:23
#1 0x00000000004004d5 in bbb (arg=456) at py-recurse-unwind.c:28
#2 0x00000000004004ed in aaa (arg=123) at py-recurse-unwind.c:34
#3 0x00000000004004fe in main () at py-recurse-unwind.c:40
Note that GDB's problems go well beyond the fact that it invokes the
unwinder recursively. In the process it messes up some internal state
(the frame stash) leading to display of (only) the sentinel frame in
the backtrace.
gdb/testsuite/ChangeLog:
* gdb.python/py-recurse-unwind.c: New file.
* gdb.python/py-recurse-unwind.py: New file.
* gdb.python/py-recurse-unwind.exp: New file.
Simon Marchi [Wed, 24 Aug 2016 17:10:08 +0000 (13:10 -0400)]
Allow resetting an empty inferior-tty
This patch allows the user to set the inferior-tty to "empty", in order
to come back to the default behaviour of using the same tty as gdb is
using.
This is already supported in MI (and tested in gdb.mi/mi-basics.exp).
I added a new test, set-inferior-tty.exp, where I test only the setting
and unsetting of the parameter. It would be nice to actually test that
the inferior output properly goes to the separate tty, but that will be
for another day.
gdb/ChangeLog:
* infcmd.c (set_inferior_io_terminal): Set inferior terminal to
NULL if terminal_name is an empty string.
(_initialize_infcmd): Make the argument of "set inferior-tty"
optional, mention it in the help doc.
gdb/doc/ChangeLog:
* gdb.texinfo (Input/Output): Mention possibility to unset
inferior-tty.
gdb/testsuite/ChangeLog:
* gdb.base/set-inferior-tty.exp: New file.
* gdb.base/set-inferior-tty.c: New file.
Carl E. Love [Wed, 24 Aug 2016 15:08:58 +0000 (08:08 -0700)]
Fix for gdb.base/pc-fp.exp.
It is my understanding that GDB used to require each architecture to
define a Frame Pointer (fp). However, this functionality was deprecated
some time ago so the call to setup the fp_reg was changed to deprecated
(set_gdbarch_deprecated_fp_regnum). It should have been removed from the
Power code.
That said, the code "set_gdbarch_deprecated_fp_regnum
(gdbarch, PPC_R0_REGNUM + 1);" sets up register r1 as the frame pointer.
Register r1 is no longer used to hold the frame pointer on Power. By
removing the fp definition for Power in GDB, it causes GDB to fall back
to the call get_frame_base_address (frame) which returns the correct value
depending on the specific senario but most of the time is the DWARF
canonical frame address.
gdb/ChangeLog
2016-08-24 Carl Love <cel@us.ibm.com>
* rs6000-tdep.c (rs6000_gdbarch_init): Remove call
set_gdbarch_deprecated_fp_regnum() from initialization function.
Pedro Alves [Tue, 23 Aug 2016 15:03:28 +0000 (16:03 +0100)]
Fix PR20494 - User input stops being echoed in CLI
This patch fixes a problem that problem triggers if you start an
inferior, e.g., with the "start" command, in a UI created with the
new-ui command, and then run a foreground execution command in the
main UI. Once the program stops for the latter command, typing in the
main UI no longer echoes back to the user.
The problem revolves around this:
- gdb_has_a_terminal computes its result lazily, on first call.
that is what saves gdb's initial main UI terminal state (the UI
associated with stdin):
This is the state that target_terminal_ours() restores.
- In this scenario, the gdb_has_a_terminal function happens to be
first ever called from within the target_terminal_init call in
startup_inferior:
(top-gdb) bt
#0 gdb_has_a_terminal () at src/gdb/inflow.c:157
#1 0x000000000079db22 in child_terminal_init_with_pgrp () at src/gdb/inflow.c:217
[...]
#4 0x000000000065bacb in target_terminal_init () at src/gdb/target.c:456
#5 0x00000000004676d2 in startup_inferior () at src/gdb/fork-child.c:531
[...]
#7 0x000000000046b168 in linux_nat_create_inferior () at src/gdb/linux-nat.c:1112
[...]
#9 0x00000000005f20c9 in start_command (args=0x0, from_tty=1) at src/gdb/infcmd.c:657
If the command to start the inferior is issued on the main UI, then
readline will have deprepped the terminal when we reach the above, and
the problem doesn't appear.
If however the command is issued on a non-main UI, then when we reach
that gdb_has_a_terminal call, the main UI's terminal state is still
set to whatever readline has sets it to in rl_prep_terminal, which
happens to have echo disabled. Later, when the following synchronous
execution command finishes, we'll call target_terminal_ours to restore
gdb's the main UI's terminal settings, and that restores the terminal
state with echo disabled...
Conceptually, the fix is to move the gdb_has_a_terminal call earlier,
to someplace during GDB initialization, before readline/ncurses have
had a chance to change terminal settings. Turns out that
"set_initial_gdb_ttystate" is exactly such a place.
I say conceptually, because the fix actually inlines the
gdb_has_a_terminal part that saves the terminal state in
set_initial_gdb_ttystate and then simplifies gdb_has_a_terminal, since
there's no point in making gdb_has_a_terminal do lazy computation.
gdb/ChangeLog:
2016-08-23 Pedro Alves <palves@redhat.com>
PR gdb/20494
* inflow.c (our_terminal_info, initial_gdb_ttystate): Update
comments.
(enum gdb_has_a_terminal_flag_enum, gdb_has_a_terminal_flag):
Delete.
(set_initial_gdb_ttystate): Record our_terminal_info here too,
instead of ...
(gdb_has_a_terminal): ... here. Reimplement in terms of
initial_gdb_ttystate. Make static.
* terminal.h (gdb_has_a_terminal): Delete declaration.
(set_initial_gdb_ttystate): Add comment.
* top.c (show_interactive_mode): Use input_interactive_p instead
of gdb_has_a_terminal.
gdb/testsuite/ChangeLog:
2016-08-23 Pedro Alves <palves@redhat.com>
PR gdb/20494
* gdb.base/new-ui-echo.c: New file.
* gdb.base/new-ui-echo.exp: New file.
Nick Clifton [Tue, 23 Aug 2016 14:41:01 +0000 (15:41 +0100)]
Better fix for PR 20499, including preventing strlen from being called on an uninitialised name field.
PR gprof/20499
* corefile.c (BUFSIZE): Define.
(STR_BUFSIZE): Define.
(read_function_mappings): Use BUFSIZE and STR)BUFSIZE.
(num_of_syms_in): Move buf, address and name arrays out of
function and declare as static BUFSIZE arrays.
Use STR_BUFSIZE when scanning for name and address.
(core_create_syms_from): Revert previous delta. Instead
short circuit the parsing of a symbol if all three fields
could not be found.
as we can see, there are two instances of connect-stopped-target or
connect-stopped-target in the command line spawning gdbserver, but
none of these gets parameters from command line. In these two
tests, gdbserver is spawned via "gdbserver_spawn ${binfile}". However,
the argument of gdbserver_spawn is the argument passed the child
inferior, not the program itself.
# Start a gdbserver process running SERVER_EXEC, and connect GDB
# to it. CHILD_ARGS are passed to the inferior.
#
# Returns the target protocol and socket to connect to.
proc gdbserver_spawn { child_args } {
set target_exec [gdbserver_download_current_prog]
GDBserver gets the program via last_loaded_file, which is set by
gdb_file_cmd. In each test, we don't need to pass ${binfile}.
Yao Qi [Tue, 23 Aug 2016 13:15:50 +0000 (14:15 +0100)]
Fix signals-state-child.exp in remote testing
Remote testing isn't considered in signals-state-child.exp, so the it
fails like
shell diff -s /scratch/yao/gdb/build-git/aarch64-linux-gnu/gdb/testsuite/outputs/gdb.base/signals-state-child/standalone.txt /scratch/yao/gdb/build-git/aarch64-linux-gnu/gdb/testsuite/outputs/gdb.base/signals-state-child/gdb.txt^M
diff: /scratch/yao/gdb/build-git/aarch64-linux-gnu/gdb/testsuite/outputs/gdb.base/signals-state-child/standalone.txt: No such file or directory^M
(gdb) FAIL: gdb.base/signals-state-child.exp: signals states are identical
This patch is to fix it.
gdb/testsuite:
2016-08-23 Yao Qi <yao.qi@linaro.org>
* gdb.base/signals-state-child.exp: Set variables gdb_txt and
standalone_txt. Delete gdb_txt and standalone_txt on host
and target. Spawn the binary on target. Copy files from
target to host.
For consistency with the previous two patches, this one
adds a macro for the two ARMv8.2 table entries. Both table
entries need a non-null aarch64_op field.
I haven't added macros for the RAS and STAT_PROFILE entries
since there's only one of each. The series isn't getting
rid of braced entries altogether, so I've only looked at
replacing things that occur more than once.
opcodes/
* aarch64-tbl.h (V8_2_INSN): New macro.
(aarch64_opcode_table): Use it.
[AArch64] Add OP parameter to aarch64-tbl.h macros
Nick recently wrapped most of aarch64-tbl.h entries in macros
like CORE_INSN. These new macros assumed that the aarch64_op
"op" field of aarch64_opcode is 0 and that the new "verifier"
field is NULL.
However, there are a lot of CORE, SIMD and FP insns whose table
entries need a nonzero aarch64_op field, so these entries
continued to use a braced list instead of a macro. This makes
the table entries less consistent and means that there are still
quite a few braced entries that need to be updated when making
further changes to the aarch64_opcode structure.
I think the number of entries that need a nonzero aarch64_op
field is high enough to justify having an explicit aarch64_op
entry for all CORE, SIMD and FP entries. This patch adds
one and updates all existing uses of the macros. A following
patch makes more use of the macros.
I've followed existing practice by using 0 instead of OP_NIL
for empty aarch64_op fields. Empty fields are still the norm
and you need to know what the fields are when reading the table
anyway, so it was hard to justify an additional patch to replace
all 0 op fields with OP_NIL.
Pedro Alves [Mon, 22 Aug 2016 19:05:09 +0000 (20:05 +0100)]
Fix PR gdb/20505 - Make vDSO detection work with core files
Loading a core dump that was either generated on a system running
pristine glibc master, or on a Fedora/RHEL system with LD_DEBUG=unused
set in the environment, solib-svr4.c:svr4_current_sos fails to filter
out the vDSO, resulting in:
(gdb) core-file corefile.core^M
[New LWP 2362]^M
warning: Could not load shared library symbols for linux-vdso.so.1.^M
Do you need "set solib-search-path" or "set sysroot"?^M
Core was generated by `build-gdb/gdb/testsuite/outputs/gdb.base/corefile/'.^M
...
The problem is that gdbarch_vsyscall_range does not support core
inferiors at all.
When live debugging, we're finding the vDSO's start address with
auxv/AT_SYSINFO_EHDR, and then we find the vDSO's size by look for the
corresponding mapping, by parsing /proc/PID/maps. When debugging a
core dump, we can also determine the starting address from
auxv/AT_SYSINFO_EHDR. However, we obviously can't read the core
mappings out of the host's /proc. But we can instead look for a
corresponding load segment in the core's bfd.
gdb/ChangeLog:
2016-08-22 Pedro Alves <palves@redhat.com>
PR gdb/20505
* linux-tdep.c (linux_vsyscall_range_raw): For core inferiors,
find the vDSO's start address with AT_SYSINFO_EHDR too, and
determine the vDSO's size by finding the PT_LOAD segment that
matches AT_SYSINFO_EHDR.
gdb/testsuite/ChangeLog:
2016-08-22 Pedro Alves <palves@redhat.com>
PR gdb/20505
* gdb.base/vdso-warning.exp: Test core dumps too. Use
with_test_prefix. Factor out bits to ...
(test_no_vdso): ... this new procedure.
Alan Modra [Mon, 22 Aug 2016 01:12:26 +0000 (10:42 +0930)]
Error on unsupported PowerPC ifuncs
The pr19784 tests fail on ppc32 due to a gcc bug. The failure should
be noticed when building both libpr19784a.so and libpr19784b.so,
rather than ld building a buggy libpr19784a.so that fails at run time.
This patch fixes that by moving the @local ifunc check out of
check_relocs, where a call destination may not yet be known to be
ifunc. The patch also adds a related error for -mbss-plt code.
* elf32-ppc.c (ppc_elf_check_relocs): Move error for @local ifunc..
(ppc_elf_relocate_section): ..to here. Comment. Error on
detecting -mbss-plt -fPIC local ifuncs too.
(ppc_elf_size_dynamic_sections): Comment on unnecessary glink
branch table entries.
Carl E. Love [Fri, 19 Aug 2016 18:06:38 +0000 (11:06 -0700)]
Fix missing files for ld when test suite not compiled in the source directory
This patch fixes an issues with six test suite expect files that do not
run correctly when the test suite is not built in the source directory. The
issue is these tests are not using the current "standard_testfile" call
but rather using the older set command to initialize the "testfile",
"srcfile" and "binprefix" variables or are missing the set for the
"binprefix" variable.
-----------------------------------------------
gdb/testsuite/ChangeLog
2016-08-19 Carl Love <cel@us.ibm.com>
* gdb.arch/altivec-regs.exp: Use standard_testfile instead of
maintaining separate logic for constructing the output path.
* gdb.arch/powerpc-d128-regs.exp: Likewise.
* gdb.arch/ppc-dfp.exp: Likewise.
* gdb.arch/ppc-fp.exp: Likewise.
* gdb.arch/vsx-regs.exp: Likewise.
* gdb.arch/altivec-abi.exp: Likewise, plus added local variable
binprefix for generating the additional binary files.
Yao Qi [Fri, 19 Aug 2016 13:49:31 +0000 (14:49 +0100)]
[AArch64] Match instruction "STP with base register" in prologue
Nowadays, we only match pre-indexed STP in prologue. Due to the change
in gcc, https://gcc.gnu.org/ml/gcc-patches/2016-07/msg01933.html, it
may generate "STP with base register" in prologue, which GDB doesn't
handle. That is to say, previously GCC generates prologue like this,
Yao Qi [Thu, 11 Aug 2016 16:04:11 +0000 (17:04 +0100)]
null-terminate string in linespec_location_completer
If I build gdb with -fsanitize=address and run tests, I get error,
malformed linespec error: unexpected colon^M
(gdb) PASS: gdb.linespec/ls-errs.exp: lang=C: break :
break :=================================================================^M
==3266==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x602000051451 at pc 0x2b5797a972a8 bp 0x7fffd8e0f3c0 sp 0x7fffd8e0f398^M
READ of size 2 at 0x602000051451 thread T0
#0 0x2b5797a972a7 in __interceptor_strlen (/usr/lib/x86_64-linux-gnu/libasan.so.1+0x322a7)^M
#1 0x7bd004 in compare_filenames_for_search(char const*, char const*) /home/yao/SourceCode/gnu/gdb/git/gdb/symtab.c:316^M
#2 0x7bd310 in iterate_over_some_symtabs(char const*, char const*, int (*)(symtab*, void*), void*, compunit_symtab*, compunit_symtab*) /home/yao/SourceCode/gnu/gdb/git/gdb/symtab.c:411^M
#3 0x7bd775 in iterate_over_symtabs(char const*, int (*)(symtab*, void*), void*) /home/yao/SourceCode/gnu/gdb/git/gdb/symtab.c:481^M
#4 0x7bda15 in lookup_symtab(char const*) /home/yao/SourceCode/gnu/gdb/git/gdb/symtab.c:527^M
#5 0x7d5e2a in make_file_symbol_completion_list_1 /home/yao/SourceCode/gnu/gdb/git/gdb/symtab.c:5635^M
#6 0x7d61e1 in make_file_symbol_completion_list(char const*, char const*, char const*) /home/yao/SourceCode/gnu/gdb/git/gdb/symtab.c:5684^M
#7 0x88dc06 in linespec_location_completer /home/yao/SourceCode/gnu/gdb/git/gdb/completer.c:288
....
0x602000051451 is located 0 bytes to the right of 1-byte region [0x602000051450,0x602000051451)^M
mallocated by thread T0 here:
#0 0x2b5797ab97ef in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.1+0x547ef)^M
#1 0xbbfb8d in xmalloc /home/yao/SourceCode/gnu/gdb/git/gdb/common/common-utils.c:43^M
#2 0x88dabd in linespec_location_completer /home/yao/SourceCode/gnu/gdb/git/gdb/completer.c:273^M
#3 0x88e5ef in location_completer(cmd_list_element*, char const*, char const*) /home/yao/SourceCode/gnu/gdb/git/gdb/completer.c:531^M
#4 0x8902e7 in complete_line_internal /home/yao/SourceCode/gnu/gdb/git/gdb/completer.c:964^
The code in question is here
file_to_match = (char *) xmalloc (colon - text + 1);
strncpy (file_to_match, text, colon - text + 1);
it is likely that file_to_match is not null-terminated. The patch is
to strncpy 'colon - text' bytes and explicitly set '\0'.
gdb:
2016-08-19 Yao Qi <yao.qi@linaro.org>
* completer.c (linespec_location_completer): Make file_to_match
null-terminated.
This test only collects the PC, and thus expects to only see one
register in the output of -trace-frame-collected. However, while on
the 64-bit ABI gdb only exposes 64-bit $pc/$rip (register 16 above),
on x32, GDB exposes 32-bit $eip as well, as a pseudo-register
(register 204 above). Thus, collecting $pc/$rip automatically always
collects $eip as well.
gdb/testsuite/ChangeLog:
2016-08-19 Pedro Alves <palves@redhat.com>
* gdb.trace/mi-trace-frame-collected.exp
(test_trace_frame_collected): On x32, expect two registers.
Pedro Alves [Fri, 19 Aug 2016 11:07:46 +0000 (12:07 +0100)]
x32: gdb: Fix 'call' insn relocation with qRelocInsn
Running the fast tracepoints tests against x32 gdbserver exposes a
latent bug. E.g.,:
(gdb)
continue
Continuing.
Reading /media/sf_host-pedro/gdb/mygit/build-ubuntu-x32/gdb/testsuite/outputs/gdb.trace/change-loc/change-loc-2.sl from remote target...
Thread 1 "change-loc" received signal SIGSEGV, Segmentation fault.
func4 () at /home/pedro/gdb/src/gdb/testsuite/gdb.trace/change-loc.h:24
24 }
(gdb) FAIL: gdb.trace/change-loc.exp: 1 ftrace: continue to marker 2
The test sets a fast tracepoint on a shared library. On x32, shared
libraries end up loaded somewhere in the upper 2GB of the 4GB address
space x32 has access to. When gdbserver needs to copy an instruction
to execute it in the jump pad, it asks gdb to relocate/adjust it, with
the qRelocInsn packet. gdb converts "call" instructions into a "push
$<2GB-4GB addr> + jmp" sequence, however, the "pushq" instruction sign
extends its operand, so later when the called function returns, it
returns to an incorrectly sign-extended address. E.g.,
0xfffffffffabc0000 instead of 0xfabc0000, resulting in the
segmentation fault.
Fix this by converting calls at such addresses to "sub + mov + jmp"
sequences instead.
gdb/ChangeLog:
2016-08-19 Pedro Alves <palves@redhat.com>
* amd64-tdep.c (amd64_relocate_instruction) <callq>: Handle return
addresses over 0x7fffffff.
Running fast tracepoint tests on x32 exposes a latent bug in the agent
bytecode jitting. There's a code path that forgets to emit the call
opcode... Whoops. Fixes a bunch of gdb.trace/trace-condition.exp
FAILs, like:
(gdb)
continue
Continuing.
Thread 1 "trace-condition" received signal SIGSEGV, Segmentation fault.
0x7ffec016 in ?? ()
(gdb) FAIL: gdb.trace/trace-condition.exp: ftrace: $rip == *set_point: advance through tracing
gdb/gdbserver/ChangeLog:
2016-08-19 Pedro Alves <palves@redhat.com>
Pedro Alves [Fri, 19 Aug 2016 11:07:45 +0000 (12:07 +0100)]
x32: Avoid unsigned long when installing fast tracepoint jump pads
We're casting through unsigned long to write a 64-bit immediate
operand of movabs (the comment said movl, but that was incorrect).
The problem is that unsigned long is 32-bit on x32, so we were writing
fewer bytes than necessary.
Fix this by using an 8 byte memcpy like in other similar places in the
function.
gdb/gdbserver/ChangeLog:
2016-08-19 Pedro Alves <palves@redhat.com>
* linux-x86-low.c (amd64_install_fast_tracepoint_jump_pad): Fix
comment. Use memcpy instead of casting through unsigned long.
Pedro Alves [Fri, 19 Aug 2016 11:07:45 +0000 (12:07 +0100)]
x32 Fast tracepoints: Customize jump pad address
MAP_32BIT is ignored on x32, meaning the jump pad can end up somewhere
between 2GB and 4GB, too far away from the executable for 5-byte
relative jumps (JMP rel32). So on x32, try explicitly placing the
jump pad near the middle of the available address space.
gdb/gdbserver/ChangeLog:
2016-08-19 Pedro Alves <palves@redhat.com>
* linux-amd64-ipa.c (alloc_jump_pad_buffer) [__ILP32__]: Try
allocating around 0x80000000.
projects / thirdparty / binutils-gdb.git / log
