Nick Clifton [Mon, 15 Feb 2016 11:11:46 +0000 (11:11 +0000)]
Enhance GAS's .section directive so that it can take numeric values for the flags and type fields. (ELF only)
gas * doc/as.texinfo (.section): Document that numeric values can now
be used for the flags and type fields of the ELF target's .section
directive. Add notes about the restrictions on setting flags and
types.
* config/obj-elf.c (obj_elf_change_section): Allow known sections
to be given processor specific section types. Allow processor and
application specific flags of a section to be set after
definition.
(obj_elf_parse_section_letters): Handle parsing numeric values.
(obj_elf_section_type): Handle parsing numeric values.
(obj_elf_section): Allow numeric type values.
* config/obj-elf.h (obj_elf_change_section): Update prototype.
* testsuite/gas/elf/section10.d: New test.
* testsuite/gas/elf/section10.s: Source file for new test.
* testsuite/gas/elf/elf.exp: Run the new test.
* testsuite/gas/i386/ilp32/x86-64-unwind.d: Remove dependency upon
the description of the flags produced by readelf.
* testsuite/gas/tic6x/scomm-directive-4.d: Likewise.
* NEWS: Mention the new feature.
bfd * elf-bfd.h (struct bfd_elf_special_section): Use unsigned values
for length and type fields. Use a signed value for the
suffix_length field.
binutils* readelf.c (get_section_type_name): Add hex prefix to offsets
printed for LOPROC and LOOS values. Ensure that a result is
always returned for the V850 target, even when an unrecognised
processor specific value is encountered.
(process_section_headers): Display key values in the order in
which they appear to the user. Add the "C (compressed)" value to
the list.
* s390.cc (Target_s390::match_view_u): New helper method.
(Target_s390::do_is_call_to_non_split): New method.
(Target_s390::ss_code_st_r14): New const.
(Target_s390::ss_code_l_r14): New const.
(Target_s390::ss_code_bras_8): New const.
(Target_s390::ss_code_l_basr): New const.
(Target_s390::ss_code_a_basr): New const.
(Target_s390::ss_code_ear): New const.
(Target_s390::ss_code_c): New const.
(Target_s390::ss_code_larl): New const.
(Target_s390::ss_code_brasl): New const.
(Target_s390::ss_code_jg): New const.
(Target_s390::ss_code_jgl): New const.
(Target_s390::ss_match_mcount): New helper method.
(Target_s390::ss_match_l): New helper method.
(Target_s390::ss_match_ahi): New helper method.
(Target_s390::ss_match_alfi): New helper method.
(Target_s390::ss_match_cr): New helper method.
(Target_s390::do_calls_non_split): New method.
* testsuite/Makefile.am: Added new tests.
* testsuite/split_s390.sh: New test.
* testsuite/split_s390_1_a1.s: New test.
* testsuite/split_s390_1_a2.s: New test.
* testsuite/split_s390_1_n1.s: New test.
* testsuite/split_s390_1_n2.s: New test.
* testsuite/split_s390_1_z1.s: New test.
* testsuite/split_s390_1_z2.s: New test.
* testsuite/split_s390_1_z3.s: New test.
* testsuite/split_s390_1_z4.s: New test.
* testsuite/split_s390_2_ns.s: New test.
* testsuite/split_s390_2_s.s: New test.
* testsuite/split_s390x_1_a1.s: New test.
* testsuite/split_s390x_1_a2.s: New test.
* testsuite/split_s390x_1_n1.s: New test.
* testsuite/split_s390x_1_n2.s: New test.
* testsuite/split_s390x_1_z1.s: New test.
* testsuite/split_s390x_1_z2.s: New test.
* testsuite/split_s390x_1_z3.s: New test.
* testsuite/split_s390x_1_z4.s: New test.
* testsuite/split_s390x_2_ns.s: New test.
* testsuite/split_s390x_2_s.s: New test.
These issues get fixed (or removed if no longer applicable) by attached patch.
It is based on Googled:
http://www.cs.rpi.edu/~szymansk/OOF90/bugs.html#5
When a pointer is declared its status is undefined, and cannot be
safely queried with the associated intrinsic.
-> nullify(VARNAME)
+
https://software.intel.com/en-us/forums/intel-fortran-compiler-for-linux-and-mac-os-x/topic/268786
ALLOCATE is not supposed to initialize the array.
-> Remove checks like an initial print is: \\( *0, *0, *0...\\)
These regressions remain:
-PASS: gdb.fortran/library-module.exp: print var_i in lib
+FAIL: gdb.fortran/library-module.exp: print var_i in lib
-PASS: gdb.fortran/library-module.exp: print var_i in main
+FAIL: gdb.fortran/library-module.exp: print var_i in main
I believe it is more a GDB bug (in a code contributed by me), filed:
gdb.fortran/library-module.exp false regression on GCC upgrade
https://sourceware.org/bugzilla/show_bug.cgi?id=19635
gdb/testsuite/ChangeLog
2016-02-14 Jan Kratochvil <jan.kratochvil@redhat.com>
> +static int max_value_size = 65536; /* 64k bytes */
FAIL: gdb.fortran/vla-value-sub.exp: print array2 in foo after it was filled (passed fixed array)
FAIL: gdb.fortran/vla-value-sub.exp: print array2 in foo after it was mofified in debugger (passed fixed array)
FAIL: gdb.fortran/vla-value-sub-finish.exp: print array2 in foo after it was filled
FAIL: gdb.fortran/vla-value-sub-finish.exp: print array2 in foo after it was mofified in debugger
print array2
value requires 296352 bytes, which is more than max-value-size
(gdb) FAIL: gdb.fortran/vla-value-sub.exp: print array2 in foo after it was filled (passed fixed array)
gdb/testsuite/ChangeLog
2016-02-14 Jan Kratochvil <jan.kratochvil@redhat.com>
* gdb.fortran/vla-value-sub-finish.exp (set max-value-size 1024*1024):
New test.
* gdb.fortran/vla-value-sub.exp: Likewise.
Yao Qi [Wed, 10 Feb 2016 14:21:38 +0000 (14:21 +0000)]
[ARM] Software single step cross kernel helpers
GDB step cross kernel helpers only works if the kernel helpers are tail
called, which is the case how it is used in glibc. See __aeabi_read_tp
in sysdeps/unix/sysv/linux/arm/aeabi_read_tp.S. In __aeabi_read_tp,
branch/jump to the kernel helper is the last instruction, and the next
instruction address is in LR, which is in caller function. GDB can
handle this correctly. For example, glibc function __GI___ctype_init
calls __aeabi_read_tp
once GDB or GDBserver single step instruction on 0xb6fef5d4, LR is
0xb6e19b38, which is right address of next instruction to set breakpoint
on.
However, if the kernel helpers are not tail-called, the LR is still the
address in the caller function of kernel helper's caller, which isn't
the right address of next instruction to set breakpoint on. For example,
we use kernel helper in main,
kernel helper is called on 0x0000862c and the expected next instruction
address is 0x00008630, but the LR now is the return address of main.
The problem here is LR may not have the right address because when we
single step the instruction, it isn't executed yet, so the LR isn't
updated. This patch fix this problem by decoding instruction, if the
instruction updates LR (BL and BLX), the next instruction address is
PC + INSN_SIZE, otherwise, get the address of next instruction from LR.
gdb:
2016-02-12 Yao Qi <yao.qi@linaro.org>
* arch/arm-linux.c (arm_linux_get_next_pcs_fixup): Calculate
nextpc according to instruction.
Yao Qi [Mon, 8 Feb 2016 16:05:31 +0000 (16:05 +0000)]
[ARM] Fixup PC in software single step
When I exercise GDBserver software single step, I see the following
error, which has been already handled by GDB properly.
In GDBserver log, we can see, GDBserver tries to single step instruction
on 0xb6e0a6e4, and destination address is 0xffff0fe0,
stop pc is 0xb6e0a6e4
Writing f001f0e7 to 0xffff0fe0 in process 7132
Failed to insert breakpoint at 0xffff0fe0 (Input/output error).
Failed to insert breakpoint at 0xffff0fe0 (-1).
(gdb) disassemble __aeabi_read_tp,+8
Dump of assembler code from 0xb6e0a6e0 to 0xb6e0a6e8:
0xb6e0a6e0 <__aeabi_read_tp+0>: mvn r0, #61440 ; 0xf000
0xb6e0a6e4 <__aeabi_read_tp+4>: sub pc, r0, #31
however, it fails inserting breakpoint there. This problem has already
fixed by GDB, see comments in arm-linux-tdep.c:arm_linux_software_single_step
/* The Linux kernel offers some user-mode helpers in a high page. We can
not read this page (as of 2.6.23), and even if we could then we
couldn't set breakpoints in it, and even if we could then the atomic
operations would fail when interrupted. They are all called as
functions and return to the address in LR, so step to there
instead. */
so we need to do the same thing in GDB side as well. This patch adds
a new field fixup in arm_get_next_pcs_ops, so that we can fix up PC
for arm-linux target. In this way, both GDB and GDBserver can single
step instructions going to kernel helpers.
gdb:
2016-02-12 Yao Qi <yao.qi@linaro.org>
* arch/arm-get-next-pcs.c (arm_get_next_pcs): Call
self->ops->fixup if it isn't NULL.
* arch/arm-get-next-pcs.h: Include gdb_vecs.h.
(struct arm_get_next_pcs_ops) <fixup>: New field.
* arch/arm-linux.c: Include common-regcache.h and
arch/arm-get-next-pcs.h.
(arm_linux_get_next_pcs_fixup): New function.
* arch/arm-linux.h (arm_linux_get_next_pcs_fixup): Declare.
* arm-linux-tdep.c (arm_linux_get_next_pcs_ops): Initialize
it with arm_linux_get_next_pcs_fixup.
(arm_linux_software_single_step): Move code to
arm_linux_get_next_pcs_fixup.
* arm-tdep.c (arm_get_next_pcs_ops): Initialize it.
gdb/gdbserver:
2016-02-12 Yao Qi <yao.qi@linaro.org>
* linux-arm-low.c (get_next_pcs_ops): Initialize it with
arm_linux_get_next_pcs_fixup.
Markus Metzger [Tue, 26 Jan 2016 13:58:44 +0000 (14:58 +0100)]
btrace, frame: fix crash in get_frame_type
In skip_artificial_frames we repeatedly call get_prev_frame_always until we get
a non-inline and non-tailcall frame assuming that there must be such a frame
eventually.
For record targets, however, we may have a frame chain that consists only of
artificial frames. This leads to a crash in get_frame_type when dereferencing a
NULL frame pointer.
Change skip_artificial_frames and skip_tailcall_frames to return NULL in such a
case and modify each caller to cope with a NULL return.
In frame_unwind_caller_pc and frame_unwind_caller_arch, we simply assert that
the returned value is not NULL. Their caller was supposed to check
frame_unwind_caller_id before calling those functions.
In other cases, we thrown an error.
In infcmd further move the skip_tailcall_frames call to the forward-stepping
case since we don't need a frame for reverse execution and we don't want to fail
because of that. Reverse-finish does make sense for a tailcall frame.
gdb/
* frame.h (skip_tailcall_frames): Update comment.
* frame.c (skip_artificial_frames, skip_tailcall_frames): Return NULL
if only artificial frames are found. Update comment.
(frame_unwind_caller_id): Handle NULL return.
(frame_unwind_caller_pc, frame_unwind_caller_arch): Assert that
skip_artificial_frames does not return NULL.
(frame_pop): Add an error if only tailcall frames are found.
* infcmd.c (finish_command): Move skip_tailcall_frames call into forward-
execution case. Add an error if only tailcall frames are found.
H.J. Lu [Thu, 11 Feb 2016 18:28:33 +0000 (10:28 -0800)]
Enable -Bsymbolic and -Bsymbolic-functions to PIE
Before binutils 2.26, -Bsymbolic and -Bsymbolic-functions were also
applied to PIE so that "ld -pie -Bsymbolic -E" can be used to export
symbols in PIE with local binding. This patch re-enables -Bsymbolic
and -Bsymbolic-functions for PIE.
PR ld/19615
* ld.texinfo: Document -Bsymbolic and -Bsymbolic-functions for
PIE.
* lexsup.c (parse_args): Enable -Bsymbolic and
-Bsymbolic-functions for PIE.
* testsuite/ld-i386/i386.exp: Run pr19175.
* testsuite/ld-i386/pr19615.d: New file.
* testsuite/ld-i386/pr19615.s: Likewise.
* testsuite/ld-x86-64/pr19615.d: Likewise.
* testsuite/ld-x86-64/pr19615.s: Likewise.
Wei-cheng Wang [Sat, 27 Jun 2015 16:21:38 +0000 (00:21 +0800)]
gdbserver: Remove tracepoint_action ops.
This patch removes 'ops' in tracepoint, and uses helper functions to
call action handler instead.
The object layout of tracepoint_action may differ in gdbserver and
inferior depend on the alignment rule of target ABI, so gdbserver cannot
simply copy the object from its memory to inferior memory.
Nick Clifton [Thu, 11 Feb 2016 15:30:55 +0000 (15:30 +0000)]
Allow the .cfi_sections directive to be reissued provided that CFI generation has not yet started.
PR gas/19614
* dw2gencfi.c (cfi_sections_set): Delay setting this variable
until it is actually used.
(cfi_set_sections): Set cfi_sections_set to true.
(dot_cfi_startproc): Likewise.
(dot_cfi_endproc): Likewise.
(dot_cfi_fde_data): Likewise.
(cfi_finish): Likewise.
(dot_cfi_sections): Do not set cfi_sections_set.
* doc/as.texinfo (.cfi_sections): Note that targets can provide
their own cfi section name. Also note that the directive can be
reissued provided that CFI generation has not started.
* testsuite/gas/mips/compact-eh-err2.s: Add .cfi_startproc and
.cfi_endproc directives so that the redefinition of .cfi_sections
will trigger the generation of the error message.
* testsuite/gas/mips/compact-eh-err2.l: Update expected line
number of error message.
This tests whether $ymm15 can be correctly collected and printed from
tfile. It covers:
- storing tdesc in tfile (without that, $ymm15 doesn't exist)
- ax_pseudo_register_collect for x86 (without that, $ymm15 cannot be
collected)
- register order in tfile_fetch_registers (without that, $ymm15h is
fetched from wrong position)
- off-by-one in tfile_fetch_registers (without that, $ymm15h is
incorrectly considered to be out of bounds)
- using proper tdesc in encoding tracepoint actions (without that,
internal error happens due to $ymm15h being considered unavailable)
gdb/testsuite/ChangeLog:
* gdb.trace/tfile-avx.c: New test.
* gdb.trace/tfile-avx.exp: New test.
Antoine Tremblay [Thu, 11 Feb 2016 13:14:12 +0000 (08:14 -0500)]
Use the target architecture when encoding tracepoint actions
This patch uses the target architecture rather then the objfile
architecture when encoding tracepoint actions.
The target architecture may contain additional registers. E.g. ARM VFP
registers. This information is needed to allow their collection. Since we
can never know whether the registers numbers in the target match the
binary's we have to use tdesc here.
One note about combined debuggers / multi-inferior from Pedro Alves:
In the combined debugger case taking Cell as the practical example that
gdb supports currently:
In that case, the main target_gdbarch() will be powerpc, but you may have set a
tracepoint on _spu_ code, which has a different gdbarch. so for that case,
target_gdbarch would be wrong. I think that in that case, we'd need to
find __the_ target/tdesc gdbarch that is (bfd) compatible with the
objfile's gdbarch.
I think cell/spu gdbserver doesn't support tracepoints, so we can ignore
this for now.
The multi-inferior/process case is somewhat related, but its simpler.
each inferior has its own gdbarch.
That is, target_gdbarch depends on the current inferior selected.
In fact, that just returns inferior->gdbarch nowaways.
No regressions, tested on ubuntu 14.04 ARMv7 and x86.
With gdbserver-{native,extended} / { -marm -mthumb }
gdb/ChangeLog:
* tracepoint.c (encode_actions_1): Use target_gdbarch () rather
than loc->gdbarch.
H.J. Lu [Thu, 11 Feb 2016 03:33:52 +0000 (19:33 -0800)]
Mask off the least significant bit in GOT offset
The least significant bit in GOT offset is to record whether we have
initialized the GOT entry in R_386_GOT32 processing. We need to mask
it off for R_386_GOT32X.
PR ld/19601
* elf32-i386.c (elf_i386_relocate_section): Mask off the least
significant bit in GOT offset for R_386_GOT32X.
Yao Qi [Wed, 10 Feb 2016 16:40:52 +0000 (16:40 +0000)]
Clear *VAL in regcache_raw_read_unsigned
We have function regcache_raw_read_unsigned defined in both GDB and
GDBserver, so that it is used in common like this,
ULONGEST value;
status = regcache_raw_read_unsigned (regcache, regnum, &value);
'value' is correctly set in GDB side, but may not be correctly set
in GDBserver, because &value is passed in regcache_raw_read_unsigned
but collect_register may only set part of the whole variable. In my
test, I see the top half of 'value' is garbage. This patch fixes this
problem by clearing *VAL before calling collect_register.
gdb.trace: Use g packet order in tfile_fetch_registers.
tfile_fetch_registers currently wrongly fetches registers using
gdb order instead of g packet order. On x86_64 with AVX, this causes
problems with ymm*h and orig_rax registers: gdb has ymm*h first, while
g packet has orig_rax first.
gdb/ChangeLog:
* tracefile-tfile.c (tfile_fetch_registers): Use g packet order
instead of gdb order.
gdb/doc/ChangeLog:
* gdb.texinfo (Trace File Format): Remove misleading information
about register block ordering.
Nick Clifton [Wed, 10 Feb 2016 11:25:59 +0000 (11:25 +0000)]
Correct assertion in NIOS2 linker to allow signed 16-buit immediate values.
PR 19405
* elf32-nios2.c (nios2_elf32_install_imm16): Allow for signed
immediate values.
* elf-eh-frame.c (_bfd_elf_discard_section_eh_frame): Limit the
number of messages about FDE encoding preventing .eh_frame_hdr
generation.
Keith Seitz [Mon, 8 Feb 2016 20:57:22 +0000 (12:57 -0800)]
breakpoints/19546: Fix crash after updating breakpoints
One of the last checks update_breakpoints_after_exec does while looping
over the list of breakpoints is check that the breakpoint has a valid
location spec. It uses event_location_empty_p to check if the location spec
is "empty", and if it is, the breakpoint is deleted.
momentary_breakpoint types rely on setting the breakpoint structure's
location spec to NULL, thereby causing an update to delete the breakpoint.
However, event_location_empty_p assumed that locations were never NULL.
As a result, GDB would crash dereferencing a NULL pointer whenever
update_breakpoints_after_exec would encounter a momentary_breakpoint.
This patch creates a new wrapper/helper function which tests that the given
breakpoint's location spec is non-NULL and if it is not "empty"
or "unspecified."
gdb/ChangeLog
PR breakpoints/19546
* breakpoint.c (breakpoint_event_location_empty_p): New function.
(update_breakpoints_after_exec, bkpt_re_set): Use this new function
instead of event_location_empty_p.
gdb/testsuite/ChangeLog
PR breakpoints/19546
* gdb.base/infcall-exec.c: New file.
* gdb.base/infcall-exec2.c: New file.
* gdb.base/infcall-exec.exp: New file.
Keith Seitz [Tue, 9 Feb 2016 18:02:54 +0000 (10:02 -0800)]
Enable/update legacy linespecs in MI.
MI is currently using string_to_event_location to enable the use of legacy
linespecs, but using this function (until this patchset) had the (as yet
unnoticed) side effect of allowing both MI and CLI representation for
explicit locations.
This patch simply changes MI to use the same legacy linespec functions
that the python and guile interpreters use. This eliminates the CLI syntax
for explicit locations (in MI).
gdb/ChangeLog
* mi/mi-cmd-break.c (mi_cmd_break_insert_1): Use
string_to_event_location_basic instead of string_to_event_location.
Keith Seitz [Tue, 9 Feb 2016 18:02:53 +0000 (10:02 -0800)]
Use string_to_event_location_basic in guile.
This patch, analogous to the previous python patch, implements proper
legacy linespec support in guile code using the newly introduced
string_to_event_location_basic.
gdb/ChangeLog
* guile/scm-breakpoint.c (gdbscm_register_breakpoint_x): Skip
leading whitespace and use string_to_event_location_basic instead
of new_linespec_location.
gdb/testsuite/ChangeLog
* gdb.guile/scm-breakpoint.exp (test_bkpt_address): New procedure.
(toplevel): Call test_bkpt_address.
Now that "legacy" linespecs benefit from consolidated support in
string_to_event_location_basic, python's Breakpoint command should use this
function to turn strings into event locations.
As a result, this patch fixes python/19506. Before:
(gdb) python gdb.Breakpoint("*main")
Traceback (most recent call last):
File "<string>", line 1, in <module>
RuntimeError: Function "*main" not defined.
Error while executing Python code.
After:
(gdb) python gdb.Breakpoint("*main")
Breakpoint 1 at 0x4005fb: file ../../../src/gdb/testsuite/gdb.python/py-breakpoint.c, line 32.
gdb/ChangeLog
PR python/19506
* python/py-breakpoint.c (bppy_init): Use
string_to_event_location_basic instead of new_linespec_location.
gdb/testsuite/ChangeLog
PR python/19506
* gdb.python/py-breakpoint.exp (test_bkpt_address): New procedure.
(toplevel): Call test_bkpt_address.
Keith Seitz [Tue, 9 Feb 2016 18:02:53 +0000 (10:02 -0800)]
Refactor string_to_event_location for legacy linespec support.
This patch refactors string_to_event_location, breaking it into two
separate functions:
1) string_to_event_location_basic
A "basic" string parser that implements support for "legacy" linespecs
(linespec, address, and probe locations). This function is intended to
be used by any UI wishing/needing to support this legacy behavior.
2) string_to_event_location
This is now intended as a CLI-only function which adds explicit location
parsing in a CLI-appropriate manner (in the form of traditional option/value
pairs).
Together these patches serve to simplify string-to-event location parsing
for all existing non-CLI interfaces (MI, guile, and python).
gdb/ChangeLog
* location.c (string_to_explicit_location): Note that "-p" is
reserved for probe locations and return NULL for any input
that starts with that.
(string_to_event_location): Move "legacy" linespec code to ...
(string_to_event_location_basic): ... here.
* location.h (string_to_event_location): Update comment.
(string_to_event_location_basic): New function.
Simon Marchi [Tue, 9 Feb 2016 14:01:58 +0000 (09:01 -0500)]
Modernize configure.ac's
Using AC_OUTPUT with arguments has been deprecated for some time in
autoconf, even in version 2.64, which we are using. This change should
not affect functionality.
I also removed the "exit 0"'s, they shouldn't be necessary.
gdb/ChangeLog:
* configure.ac: Use AC_CONFIG_FILES instead of passing arguments
to AC_OUTPUT. Remove "exit 0" at the end.
* configure: Regenerate.
gdb/testsuite/ChangeLog:
* configure.ac: Use AC_CONFIG_FILES instead of passing arguments
to AC_OUTPUT.
* configure: Regenerate.
gdb/gdbserver/ChangeLog:
* configure.ac: Use AC_CONFIG_FILES instead of passing arguments
to AC_OUTPUT.
* configure: Regenerate.
Pedro Alves [Tue, 9 Feb 2016 12:12:17 +0000 (12:12 +0000)]
Fix PR19548: Breakpoint re-set inserts breakpoints when it shouldn't
PR19548 shows that we still have problems related to 13fd3ff34329:
[PR17431: following execs with "breakpoint always-inserted on"]
https://sourceware.org/ml/gdb-patches/2014-09/msg00733.html
The problem this time is that we currently update the global location
list and try to insert breakpoint locations after re-setting _each_
breakpoint in turn.
Say:
- We have _more_ than one breakpoint set. Let's assume 2.
- There's a breakpoint with a pre-exec address that ends up being an
unmapped address after the exec.
- That breakpoint is NOT the first in the breakpoint list.
Then when handling an exec, and we re-set the first breakpoint in the
breakpoint list, we mistakently try to install the old pre-exec /
un-re-set locations of the other breakpoint, which fails:
(gdb) continue
Continuing.
process 28295 is executing new program: (...)/execl-update-breakpoints2
Error in re-setting breakpoint 1: Warning:
Cannot insert breakpoint 2.
Cannot access memory at address 0x1000764
Breakpoint 1, main (argc=1, argv=0x7fffffffd368) at /home/pedro/gdb/mygit/src/gdb/testsuite/gdb.base/execl-update-breakpoints.c:34
34 len = strlen (argv[0]);
(gdb)
Fix this by deferring the global location list update till after all
breakpoints are re-set.
Tested on x86_64 Fedora 20, native and gdbserver.
gdb/ChangeLog:
2016-02-09 Pedro Alves <palves@redhat.com>
PR breakpoints/19548
* breakpoint.c (create_overlay_event_breakpoint): Don't update
global location list here.
(create_longjmp_master_breakpoint)
(create_std_terminate_master_breakpoint)
(create_exception_master_breakpoint, create_jit_event_breakpoint)
(update_breakpoint_locations):
(breakpoint_re_set): Update global location list after all
breakpoints are re-set.
gdb/testsuite/ChangeLog:
2016-02-09 Pedro Alves <palves@redhat.com>
PR breakpoints/19548
* gdb.base/execl-update-breakpoints.c (some_function): New
function.
(main): Call it.
* gdb.base/execl-update-breakpoints.exp: Add a second breakpoint.
Tighten expected GDB output.
Nick Clifton [Tue, 9 Feb 2016 09:56:21 +0000 (09:56 +0000)]
Add a more helpful warning message to explain why some AArch64 relocations can overflow.
bfd * elfnn-aarch64.c (elfNN_aarch64_relocate_section): Add a more
helpful warning message to explain why certain AArch64 relocs
might overflow.
ld * testsuite/ld-aarch64/reloc-overflow-bad.d: New test.
* testsuite/ld-aarch64/reloc-overflow-1.s: New source file.
* testsuite/ld-aarch64/reloc-overflow-2.s: New source file.
* testsuite/ld-aarch64/aarch64-elf.exp: Run the new test.
Nick Clifton [Tue, 9 Feb 2016 09:56:21 +0000 (09:56 +0000)]
Add a more helpful warning message to explain why some AArch64 relocations can overflow.
bfd * elfnn-aarch64.c (elfNN_aarch64_relocate_section): Add a more
helpful warning message to explain why certain AArch64 relocs
might overflow.
ld * testsuite/ld-aarch64/reloc-overflow-bad.d: New test.
* testsuite/ld-aarch64/reloc-overflow-1.s: New source file.
* testsuite/ld-aarch64/reloc-overflow-2.s: New source file.
* testsuite/ld-aarch64/aarch64-elf.exp: Run the new test.
Simon Marchi [Mon, 8 Feb 2016 19:02:36 +0000 (14:02 -0500)]
Always organize test artifacts in a directory hierarchy
When running tests in parallel, each test puts its generated files in a
different directory, under "outputs". I think it would be nice if it
was always the case, as it would isolate the test cases a bit more. An
artifact created by a test wouldn't get overwritten by another test.
Also, it makes it easier to clean up. A lot of executables are left all
over the place because their names do not appear in gdb.*/Makefile. If
everything is in "outputs", then we just have to delete that directory
(which we already do).
At the same time it makes the gdb.foo directories and their Makefiles
useless in the build directory, since they are pretty much only used for
cleaning.
Simon Marchi [Mon, 8 Feb 2016 19:00:49 +0000 (14:00 -0500)]
Fix in-tree, parallel running of Ada tests
While testing the following patch,
[PATCH] Always organize test artifacts in a directory hierarchy
https://sourceware.org/ml/gdb-patches/2016-01/msg00133.html
I noticed that it broke Ada testing. This lead me to think that
parallel testing when building in-tree didn't work previously in Ada.
It is confirmed by this test:
This patch fixes in-tree parallel testing for Ada, and consequently
serial and parallel testing when the aforementioned patch is applied.
The problem originates from the fact that Ada support code cd's to the
builddir before compiling. In itself it's not a problem, it allows to
place intermediate auto-generated files in that directory. The Ada
compilation refers to the source file, which is in another directory,
only by its base name (e.g. foo.adb). In serial mode, that worked
because builddir was the same as the source directory (e.g.
gdb.ada/fun_addr/). In an out-of-tree build, it works because the
source directory is added as an include directory (note: this is not the
same $srcdir as autoconf's):
set srcdir [file dirname $source]
additional_flags=-I$srcdir
However, when building in-tree, srcdir is relative: ./gdb.ada/fun_addr.
When using parallel or always-in-outputs-directory mode, we are cd'ed in
the outputs directory. So -I$srcdir is relative to the current
directory, which is wrong.
To fix it, I made the TCL variable srcdir (set in site.exp, from which
everything else is derived) always absolute. It is done by assigning
autoconf's abs_srcdir instead of autoconf's srcdir. This way -I$srcdir
will always be good, regardless of where we cd'ed to. A small apparent
change is that when running tests, DejaGnu will say:
Cary Coutant [Sun, 7 Feb 2016 06:42:16 +0000 (22:42 -0800)]
Fix incorrect x32 overflow checking for refs to weak undef symbols.
On x32, a pc-relative reference to an undef weak symbol (value 0)
with a negative addend (typically -4) generates a spurious overflow
error because Symbol_value::value() returns a 32-bit negative number
as an unsigned number, which gets zero-extended before subtracting
the PC value. This patch fixes the problem by special-casing the
negative addend, and adding it to the value after widening it to
64 bits. Symbol_value::value() does not need the addend if it's
negative, since it is only important when processing section
symbols for merge sections, where a positive addend provides the
input section offset of the merged constant.
gold/
* x86_64.cc (X86_64_relocate_functions::pcrela32_check): Fix x32
overflow checking when symbol value + addend < 0.
Simon Marchi [Sun, 7 Feb 2016 14:45:02 +0000 (09:45 -0500)]
varobj: Cleanup dead code
This patch removes some dead code.
I noticed that varobj_delete was always called with dellist == NULL, so
I started removing that parameter. That allows removing a good chunk of
the code in varobj_delete, making it almost trivial. We can also remove
the resultp parameters in that whole trail. In turn, this shows that
struct cpstack, cppush and cppop were only used fo that mechanism, so
they can be removed as well.
I also moved the function comment to the header file to comply with
today's guideline, even though the rest of the file does not respect it
(yet).
Cary Coutant [Sun, 7 Feb 2016 02:18:51 +0000 (18:18 -0800)]
Fix compile errors about shift counts too large.
In order to get around the optimizer and newer compiler warnings
about shift counts, the overflow checking code had resorted to
some messy shifting, and with the never-before-seen instantiations
of the template functions, we were still running afoul of the
compiler checks.
This patch replaces those messy shift sequences with a simple
class template that provides the min and max limits for any
bit size up to 64, with a specialization for 64 that prevents
the compiler from complaining.
gold/
PR gold/19577
* reloc.h (Limits): New class.
(Bits::has_overflow32): Use min/max values from Limits.
(Bits::has_unsigned_overflow32): Likewise.
(Bits::has_signed_unsigned_overflow32): Likewise.
(Bits::has_overflow): Likewise.
(Bits::has_unsigned_overflow): Likewise.
(Bits::has_signed_unsigned_overflow64): Likewise.
Cary Coutant [Sat, 6 Feb 2016 22:47:05 +0000 (14:47 -0800)]
Fix overflow checking for 32-bit pc-relative relocations on x32.
The problem here is that x32 is really using 64-bit addressing,
while pretending to be 32-bit. Even though the object file format
is 32-bit, we need to do the overflow checking with 64-bit
arithmetic (because that's what the hardware will be using).
This patch overrides the pcrela32_check functions in reloc.h
with target-specific versions that do 64-bit checking.
I've also updated the test case to use -Tdata instead of adding
a huge .space directive, to reduce the size of the .o files.
gold/
PR gold/19567
* reloc.h (Relocate_functions::Overflow_check): Add comments.
* x86_64.cc (X86_64_relocate_functions): New class.
(Target_x86_64::Relocate::relocate): Use the new class.
* testsuite/Makefile.am (x86_64_overflow_pc32): Add -Tdata option.
(x32_overflow_pc32): New test case.
* testsuite/Makefile.in: Regenerate.
* testsuite/x32_overflow_pc32.sh: New script.
* testsuite/x86_64_overflow_pc32.s: Remove .space directive.
Sriraman Tallam [Fri, 5 Feb 2016 23:07:45 +0000 (15:07 -0800)]
2016-02-05 Sriraman Tallam <tmsriram@google.com>
* icf.cc (get_rel_addend): New function.
(get_section_contents): Move merge section addend computation to a
new function. Ignore negative values for SHT_REL and SHT_RELA addends.
Fix bug to not read past the length of the section.
Fix bug related to addend computation for MERGE sections.
Cary Coutant [Fri, 5 Feb 2016 17:19:47 +0000 (09:19 -0800)]
Add some relocation overflow checks for x86_64.
2016-02-05 Cary Coutant <ccoutant@gmail.com>
Andrew Senkevich <andrew.senkevich@intel.com>
gold/
PR gold/18695
* x86_64.cc (Target_x86_64::Relocate::relocate): Add overflow
checking for R_X86_64_32, R_X86_64_32S, R_X86_64_PC32, and
R_X86_64_PLT32.
* testsuite/Makefile.am (x86_64_overflow_pc32): New test.
* testsuite/x86_64_overflow_pc32.sh: New test script.
* testsuite/x86_64_overflow_pc32.s: New source file.
Change the default architecture value for ARC bfds so that they have the lowest possible value, and hence can be merged with other ARC binaries wihtout changing their architecture value.
bfd * cpu-arc.c: Change default archure from bfd_mach_arc_arcv2
to bfd_mach_arc_arc600.
binutils * testsuite/binutils-all/objdump.exp: Update expected default
architecture value for ARC binaries.
Nick Clifton [Thu, 4 Feb 2016 16:27:06 +0000 (16:27 +0000)]
Prevent possible undefined behaviour computing the size of the scache by usingunsigned integers instead of signed integers.
* cgen-scache.c (scache_option_handler): Prevent possible
undefined behaviour computing the size of the scache by using
unsigned integers instead of signed integers.
Yao Qi [Thu, 4 Feb 2016 15:46:37 +0000 (15:46 +0000)]
[testsuite] Remove BASEDIR
BASEDIR was added by https://sourceware.org/ml/gdb-patches/2013-10/msg00587.html
in order to handle the different directory layout in serial testing
and parallel testing. BASEDIR is "gdb.base" in serial testing and is
"outputs/gdb.base/TESTNAME" in parallel testing. However, it doesn't
work if the GDBserver is in remote target, like this,
$ make check RUNTESTFLAGS='--target_board=remote-gdbserver-on-localhost foll-vfork.exp foll-exec.exp'
FAIL: gdb.base/foll-exec.exp: continue to first exec catchpoint (the program exited)
FAIL: gdb.base/foll-vfork.exp: exec: vfork and exec child follow, to main bp: continue to bp (the program exited)
FAIL: gdb.base/foll-vfork.exp: exec: vfork child follow, finish after tcatch vfork: finish (the program exited)
FAIL: gdb.base/foll-vfork.exp: exec: vfork relations in info inferiors: continue to bp (the program exited)
these tests fail because the executable can't be found. With target
board native-gdbserver, the program is spawned this way,
so BASEDIR is correct. However, with target board
remote-gdbserver-on-localhost, the program is spawned
spawn /usr/bin/ssh -l yao localhost /scratch/yao/gdb/build-git/x86_64/gdb/testsuite/../gdbserver/gdbserver --once :2346 /scratch/yao/gdb/build-git/x86_64/gdb/testsuite/gdb.base/foll-vfork
so BASEDIR (either "gdb.base" or "outputs/gdb.base/TESTNAME") makes no
sense.
I had a fix that pass absolute directory to BASEDIR, but it assumes
that directory structure is the same on build and target, and it
doesn't work in remote host case. The current fix in this patch is
to get the directory from argv[0]. In any case, the program to be
exec'ed is at the same directory with the main program.
Note that these tests do "next N" to let program stop at the desired
line, but it is fragile, because GDB for different targets may skip
function prologue slightly differently, so I replace some of them by
"tbreak on LINE NUMBER and continue".
gdb/testsuite:
2016-02-04 Yao Qi <yao.qi@linaro.org>
* gdb.base/foll-exec-mode.c: Include limits.h.
(main): Add parameters argc and argv. Get directory from
argv[0].
* gdb.base/foll-exec-mode.exp: Don't pass -DBASEDIR in
compilation.
* gdb.base/foll-exec.c: Include limits.h.
(main): Add parameters argc and argv.
Get directory from argv[0].
* gdb.base/foll-exec.exp: Don't pass -DBASEDIR in compilation.
Adjust tests on the number of lines as source code changed.
* gdb.base/foll-vfork-exit.c: Include limits.h.
(main): Add one line of statement before vfork.
* gdb.base/foll-vfork.c: Include limits.h and string.h.
(main): Add parameters argc and argv. Get directory from
argv[0].
* gdb.base/foll-vfork.exp: Don't pass -DBASEDIR in compilation.
(setup_gdb): Set tbreak to skip some source lines.
* gdb.multi/bkpt-multi-exec.c: Include limits.h.
(main): Add parameters argc and argv. Get directory from
argv[0].
* gdb.multi/bkpt-multi-exec.exp: Don't pass -DBASEDIR in
compilation.
* gdb.multi/multi-arch-exec.c: Include limits.h and string.h.
(main): Add parameters argc and argv. Get directory from
argv[0].
* gdb.multi/multi-arch-exec.exp: Don't pass -DBASEDIR in
compilation.
Yao Qi [Thu, 4 Feb 2016 15:09:09 +0000 (15:09 +0000)]
waiting_for_stop_reply around remote_fileio_request
Hi,
I see this error when GDB connects with qemu,
(gdb) n
....
Sending packet: $vCont;c#a8...Ack
Packet received: Ffstat,00000001,f6fff038
Cannot execute this command while the target is running.
Use the "interrupt" command to stop the target
and then try again.
looks we don't set rs->waiting_for_stop_reply to zero
before handle fileio request,
#10 0x00000000005edb64 in target_write (len=64, offset=4143968312, buf=0x7fffffffd570 "\375\377\377\377", annex=0x0, object=TARGET_OBJECT_MEMORY,
ops=<optimised out>) at /home/yao/SourceCode/gnu/gdb/git/gdb/target.c:1922
#11 target_write_memory (memaddr=memaddr@entry=4143968312, myaddr=myaddr@entry=0x7fffffffd6a0 "", len=len@entry=64)
at /home/yao/SourceCode/gnu/gdb/git/gdb/target.c:1500
#12 0x00000000004b2b41 in remote_fileio_func_fstat (buf=0x127b258 "") at /home/yao/SourceCode/gnu/gdb/git/gdb/remote-fileio.c:1037
#13 0x00000000004b1878 in do_remote_fileio_request (uiout=<optimised out>, buf_arg=buf_arg@entry=0x127b240)
at /home/yao/SourceCode/gnu/gdb/git/gdb/remote-fileio.c:1204
#14 0x00000000005b8c7c in catch_exceptions_with_msg (func_uiout=<optimised out>, func=func@entry=0x4b1800 <do_remote_fileio_request>,
func_args=func_args@entry=0x127b240, gdberrmsg=gdberrmsg@entry=0x0, mask=mask@entry=RETURN_MASK_ALL)
at /home/yao/SourceCode/gnu/gdb/git/gdb/exceptions.c:187
#15 0x00000000005b8dea in catch_exceptions (uiout=<optimised out>, func=func@entry=0x4b1800 <do_remote_fileio_request>, func_args=func_args@entry=0x127b240,
mask=mask@entry=RETURN_MASK_ALL) at /home/yao/SourceCode/gnu/gdb/git/gdb/exceptions.c:167
#16 0x00000000004b2fff in remote_fileio_request (buf=0x127b240 "Xf6fff038,0:", ctrlc_pending_p=0) at /home/yao/SourceCode/gnu/gdb/git/gdb/remote-fileio.c:1255
#17 0x0000000000496f12 in remote_wait_as (ptid=..., status=0x7fffffffdb20, options=1) at /home/yao/SourceCode/gnu/gdb/git/gdb/remote.c:6997
however, we did set rs->waiting_for_stop_reply to zero before Luis's
patch https://sourceware.org/ml/gdb-patches/2015-10/msg00336.html
In fact, Luis's patch v1
https://sourceware.org/ml/gdb-patches/2015-08/msg00809.html is about
setting rs->waiting_for_stop_reply back to one after
remote_fileio_request, which is correct. However during the review, the
patch is changed and ends up with "not setting rs->waiting_for_stop_reply
to zero".
I manually test GDB, but I don't have a way to run regression tests.
gdb:
2016-02-04 Yao Qi <yao.qi@linaro.org>
* remote.c (remote_wait_as): Set rs->waiting_for_stop_reply to
0 before handling 'F' and set it back afterwards.
Nick Clifton [Thu, 4 Feb 2016 09:55:10 +0000 (09:55 +0000)]
Fix the encoding of the MSP430's RRUX instruction.
PR target/19561
opcdoe * msp430-dis.c (print_insn_msp430): Add a special case for
decoding an RRC instruction with the ZC bit set in the extension
word.
include * opcode/msp430.h (IGNORE_CARRY_BIT): New define.
(RRUX): Synthesise using case 2 rather than 7.
gas * config/tc-msp430.c (msp430_operands): Remove case 7. Use case 2
to handle encoding of RRUX instruction.
* testsuite/gas/msp430/msp430x.s: Add more tests of the extended
shift instructions.
* testsuite/gas/msp430/msp430x.d: Update expected disassembly.
Max Filippov [Tue, 2 Feb 2016 14:11:38 +0000 (17:11 +0300)]
xtensa: fix signedness of gas relocations
Change 1058c7532d0b "Use signed data type for R_XTENSA_DIFF* relocation
offsets." changed signedness of BFD_RELOC_XTENSA_DIFF* relocations
substituted for BFD_RELOC_*. This made it impossible to encode arbitrary
8-, 16- and 32-bit values, which broke e.g. debug info encoding by .loc
directive. Revert this part and add test.
gas/
2016-02-03 Max Filippov <jcmvbkbc@gmail.com>
* config/tc-xtensa.c (md_apply_fix): Mark BFD_RELOC_XTENSA_DIFF*
substitutions for BFD_RELOC_* as unsigned.
* gas/testsuite/gas/xtensa/all.exp: Add loc to list of xtensa
tests.
* gas/testsuite/gas/xtensa/loc.d: New file: loc test result
patterns.
* gas/testsuite/gas/xtensa/loc.s: New file: loc test.
H.J. Lu [Wed, 3 Feb 2016 16:25:15 +0000 (08:25 -0800)]
Add -mrelax-relocations= to x86 assembler
The x86 relax relocations introduced in binutils 2.26 aren't supported
by linker on Solaris older than Solaris 12. To use x86 assembler with
older Solaris linker, this patch adds
1. A command line option -mrelax-relocations= to x86 assembler to
control whether to generate relax relocations.
2. A configure option --enable-x86-relax-relocations to decide whether
x86 assembler should generate relax relocations by default. It is
defaulted to yes, except for x86 Solaris targets older than Solaris 12.
gas/
PR gas/19520
* NEWS: Mention new command line option -mrelax-relocations and
new configure option --enable-x86-relax-relocations for x86
target.
* config.in: Regenerated.
* configure.ac: Add --enable-x86-relax-relocations.
(ac_default_x86_relax_relocations): New. Default to 1 except
for x86 Solaris targets older than Solaris 12.
(DEFAULT_GENERATE_X86_RELAX_RELOCATIONS): Define.
* configure: Likewise.
* config/tc-i386.c (generate_relax_relocations): New.
(OPTION_MRELAX_RELOCATIONS): Likewise.
(output_disp): Don't generate relax relocations if
generate_relax_relocations is 0.
(md_longopts): Add -mrelax-relocations.
(md_show_usage): Likewise.
(md_parse_option): Handle OPTION_MRELAX_RELOCATIONS.
* doc/c-i386.texi: Document -mrelax-relocations=.
* testsuite/gas/i386/got-no-relax.d: New file.
* testsuite/gas/i386/x86-64-gotpcrel-no-relax.d: Likewise.
* testsuite/gas/i386/got.d: Pass -mrelax-relocations=yes to as.
* testsuite/gas/i386/localpic.d: Likewise.
* testsuite/gas/i386/mixed-mode-reloc32.d: Likewise.
* testsuite/gas/i386/reloc32.d: Likewise.
* testsuite/gas/i386/x86-64-gotpcrel.d: Likewise.
* testsuite/gas/i386/x86-64-localpic.d: Likewise.
* testsuite/gas/i386/ilp32/x86-64-gotpcrel.d: Likewise.
* testsuite/gas/i386/ilp32/x86-64-localpic.d: Likewise.
* testsuite/gas/i386/i386.exp: Run got-no-relax and
x86-64-gotpcrel-no-relax.
Kevin Buettner [Wed, 27 Jan 2016 19:44:38 +0000 (12:44 -0700)]
msp430: Set DWARF2_ADDR_SIZE to 4.
This change makes gas's notion of the msp430 dwarf2 address size match
that of gcc and gdb. This is needed so that the format of addresses
generated for DW_LNE_set_address in .debug_line will match the address
size for the compilation unit.
In gcc/config/msp430/msp430.h, it's set to 4:
#define DWARF2_ADDR_SIZE 4
Likewise in gdb/msp430-tdep.c:
set_gdbarch_dwarf2_addr_size (gdbarch, 4);
(As far as I can tell, however, GDB doesn't use this value when decoding
.debug_line. Instead, GDB uses the Pointer Size from the compilation
unit.)
readelf is able to seamlessly handle mismatches between these various
sizes by using the size of the DW_LNE_set_address instruction to
determine the address size. Another way to fix this problem is to
make GDB behave in a similar manner. In my opinion, GDB should detect
and inform the user about these mismatches; it's not clear to me if
it's correct for GDB to go ahead and read the address anyway when a
size mismatch is detected.
Without this change, addresses in .debug_line are encoded in two bytes
for some multilibs. When GDB reads the address for
DW_LNE_set_address, it uses the pointer size provided by the CU. When
these values don't match, GDB reads the wrong number of bytes. In the
cases that I've looked at, GDB is reading 4 bytes from a 2 byte
container, which results in a garbage address. GDB discards lines
which have a bogus address; the end result is that GDB records no line
number information for CUs which have a mismatch between the address
size (from the CU) and the format of the address used by
DW_LNE_set_address.
gas/ChangeLog:
* config/tc-msp430.h (DWARF2_ADDR_SIZE): Set to 4.