H.J. Lu [Fri, 27 May 2016 13:55:42 +0000 (06:55 -0700)]
Correct CpuMax in i386-opc.h
CpuMax should be CpuIntel64, not CpuNo64. i386-gen.c is updated to
verify that CpuMax is correct. X86 assembler is updated to properly
set cpuamd64 and cpuintel64.
gas/
PR gas/20154
* config/tc-i386.c (intel64): New.
(cpu_flags_match): Set cpuamd64 and cpuintel64.
(md_parse_option): Set intel64 instead of cpuamd64 and
cpuintel64.
opcodes/
PR gas/20154
* i386-gen.c (main): Fail if CpuMax is incorrect.
* i386-opc.h (CpuMax): Set to CpuIntel64.
* i386-tbl.h: Regenerated.
Nick Clifton [Fri, 27 May 2016 12:49:58 +0000 (13:49 +0100)]
Improve the MSP430 disassembler's handling of memory read errors.
PR target/20150
* msp430-dis.c (msp430dis_read_two_bytes): New function.
(msp430dis_opcode_unsigned): New function.
(msp430dis_opcode_signed): New function.
(msp430_singleoperand): Use the new opcode reading functions.
Only disassenmble bytes if they were successfully read.
(msp430_doubleoperand): Likewise.
(msp430_branchinstr): Likewise.
(msp430x_callx_instr): Likewise.
(print_insn_msp430): Check that it is safe to read bytes before
attempting disassembly. Use the new opcode reading functions.
Andrew Burgess [Tue, 24 May 2016 15:53:58 +0000 (16:53 +0100)]
gdb: Forward VALUE_LVAL when avoiding side effects for STRUCTOP_STRUCT
When evaluating an expression with EVAL_AVOID_SIDE_EFFECTS if the value
we return is forced to be of type not_lval then GDB will be unable to
take the address of the returned value.
Instead, we should properly initialise the LVAL of the returned value.
This commit builds on two previous commits 2520f728b710 (Forward
VALUE_LVAL when avoiding side effects for STRUCTOP_STRUCT) and ac775bf4d35b (gdb: Forward VALUE_LVAL when avoiding side effects for
STRUCTOP_PTR), which in turn build on ac1ca910d74d (Fixes for PR
exp/15364).
This commit is currently untested due to my lack of access to an OpenCL
compiler, however, if follows the same pattern as the first two commits
mentioned above and so I believe that it is correct.
gdb/ChangeLog:
* opencl-lang.c (evaluate_subexp_opencl): If
EVAL_AVOID_SIDE_EFFECTS mode, forward the VALUE_LVAL attribute to
the returned value in the STRUCTOP_STRUCT case.
Andrew Burgess [Tue, 24 May 2016 15:17:52 +0000 (16:17 +0100)]
gdb: Forward VALUE_LVAL when avoiding side effects for STRUCTOP_PTR
Assume that we have a C program like this:
struct foo_type
{
int var;
} foo;
struct foo_type *foo_ptr = &foo;
int
main ()
{
return foo_ptr->var;
}
Then GDB should be able to evaluate the following, however, it currently
does not:
(gdb) start
...
(gdb) whatis &(foo_ptr->var)
Attempt to take address of value not located in memory.
The problem is that in EVAL_AVOID_SIDE_EFFECTS mode,
eval.c:evaluate_subexp_standard always returns a not_lval value as the
result for a STRUCTOP_PTR operation. As a consequence, the rest of
the code believes that one cannot take the address of the returned
value.
This patch fixes STRUCTOP_PTR handling so that the VALUE_LVAL
attribute for the returned value is properly initialized. After this
change, the above session becomes:
(gdb) start
...
(gdb) whatis &(foo_ptr->var)
type = int *
This commit is largely the same as commit 2520f728b710 (Forward
VALUE_LVAL when avoiding side effects for STRUCTOP_STRUCT) but applied
to STRUCTOP_PTR rather than STRUCTOP_STRUCT. Both of these commits are
building on top of commit ac1ca910d74d (Fixes for PR exp/15364).
gdb/ChangeLog:
* eval.c (evaluate_subexp_standard): If EVAL_AVOID_SIDE_EFFECTS
mode, forward the VALUE_LVAL attribute to the returned value in
the STRUCTOP_PTR case.
gdb/testsuite/ChangeLog:
* gdb.base/whatis.c: Extend the test case.
* gdb.base/whatis.exp: Add additional tests.
Trevor Saunders [Mon, 23 May 2016 11:42:14 +0000 (07:42 -0400)]
metag: make an array's type unsigned char[]
It contains values between 128 and 256 which fit in an unsigned char, but not a
signed char, so we should explicitly use unsigned char to not rely on how these
values are converted to signed char.
MIPS/BFD: Don't stop processing on `bfd_reloc_outofrange'
Upon a `bfd_reloc_outofrange' error continue processing so that any
further issues are also reported, similarly to how `bfd_reloc_overflow'
is handled. Adjust message formatting accordingly, using `%X' to abort
processing at conclusion.
Reduce the number of test cases by grouping relocations the handling of
which can now be verified together with a single source and dump.
bfd/
* elfxx-mips.c (_bfd_mips_elf_relocate_section)
<bfd_reloc_outofrange>: Use the `%X%H' rather than `%C' format
for message. Continue processing rather than returning failure.
MIPS/BFD: Report `bfd_reloc_outofrange' errors as such
A `bfd_reloc_outofrange' condition from `mips_elf_calculate_relocation'
currently triggers the warning callback, which in the case of LD prints
messages like:
foo.o: In function `foo':
(.text+0x0): warning: JALX to a non-word-aligned address
or:
foo.o: In function `foo':
(.text+0x0): warning: PC-relative load from unaligned address
and nothing else, which suggests this is a benign condition and link has
otherwise successfully run to completion. This is however not the case,
the link terminates right away with no further messages and no output
produced.
Use the general error or warning info callback then, preserving the
message format. Also set a BFD error condition so that a failure is
unambiguously reported. Complement the change with a set of suitable
test suite additions.
bfd/
* elfxx-mips.c (_bfd_mips_elf_relocate_section)
<bfd_reloc_outofrange>: Call `->einfo' rather than `->warning'.
Call `bfd_set_error'.
ld/
* testsuite/ld-mips-elf/unaligned-jalx-0.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-1.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-2.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-mips16-0.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-mips16-1.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-mips16-2.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-micromips-0.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-micromips-1.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-micromips-2.d: New test.
* testsuite/ld-mips-elf/unaligned-lwpc-0.d: New test.
* testsuite/ld-mips-elf/unaligned-lwpc-1.d: New test.
* testsuite/ld-mips-elf/unaligned-lwpc-2.d: New test.
* testsuite/ld-mips-elf/unaligned-lwpc-3.d: New test.
* testsuite/ld-mips-elf/unaligned-ldpc-0.d: New test.
* testsuite/ld-mips-elf/unaligned-ldpc-1.d: New test.
* testsuite/ld-mips-elf/unaligned-ldpc-2.d: New test.
* testsuite/ld-mips-elf/unaligned-ldpc-3.d: New test.
* testsuite/ld-mips-elf/unaligned-ldpc-4.d: New test.
* testsuite/ld-mips-elf/unaligned-jalx-0.s: New test source.
* testsuite/ld-mips-elf/unaligned-jalx-1.s: New test source.
* testsuite/ld-mips-elf/unaligned-jalx-2.s: New test source.
* testsuite/ld-mips-elf/unaligned-insn.s: New test source.
* testsuite/ld-mips-elf/unaligned-lwpc-0.s: New test source.
* testsuite/ld-mips-elf/unaligned-lwpc-1.s: New test source.
* testsuite/ld-mips-elf/unaligned-lwpc-2.s: New test source.
* testsuite/ld-mips-elf/unaligned-lwpc-3.s: New test source.
* testsuite/ld-mips-elf/unaligned-ldpc-0.s: New test source.
* testsuite/ld-mips-elf/unaligned-ldpc-1.s: New test source.
* testsuite/ld-mips-elf/unaligned-ldpc-2.s: New test source.
* testsuite/ld-mips-elf/unaligned-ldpc-3.s: New test source.
* testsuite/ld-mips-elf/unaligned-ldpc-4.s: New test source.
* testsuite/ld-mips-elf/unaligned-syms.s: New test source.
* testsuite/ld-mips-elf/mips-elf.exp: Run the new tests.
H.J. Lu [Wed, 25 May 2016 16:46:47 +0000 (09:46 -0700)]
Enable 64-bit archives in ar and ranlib
Since existing ld and gold support the 64-bit (MIPS) ELF archives, we
can use the 64-bit (MIPS) ELF archives as 64-bit archives. Since the
plugin target is used to create archive in plugin-enabled ar, we need
a way to enable 64-bit archives in the plugin target. This patch adds
--enable-64-bit-archive to bfd to force 64-bit archives in ar and
ranlib. Since both 64-bit MIPS and s390 ELF targets currently use
64-bit archives, 64-bit archives are enabled by default for them.
64-bit archive is generated automatically if the archive is too big.
Tested on Linux/x86 and Linux/x86-64 with existing ld and gold.
bfd/
PR binutils/14625
* archive.c (bfd_slurp_armap): Replace
bfd_elf64_archive_slurp_armap with
_bfd_archive_64_bit_slurp_armap.
(bsd_write_armap): Call _bfd_archive_64_bit_write_armap if
BFD64 is defined and the archive is too big.
(coff_write_armap): Likewise.
* archive64.c (bfd_elf64_archive_slurp_armap): Renamed to ...
(_bfd_archive_64_bit_slurp_armap): This.
(bfd_elf64_archive_write_armap): Renamed to ...
(_bfd_archive_64_bit_write_armap): This.
* configure.ac: Add --enable-64-bit-archive.
(want_64_bit_archive): New. Set to true by default for 64-bit
MIPS and s390 ELF targets.
(USE_64_BIT_ARCHIVE): New AC_DEFINE.
* config.in: Regenerated.
* configure: Likewise.
* elf64-mips.c (bfd_elf64_archive_functions): Removed.
(bfd_elf64_archive_slurp_armap): Likewise.
(bfd_elf64_archive_write_armap): Likewise.
(bfd_elf64_archive_slurp_extended_name_table): Likewise.
(bfd_elf64_archive_construct_extended_name_table): Likewise.
(bfd_elf64_archive_truncate_arname): Likewise.
(bfd_elf64_archive_read_ar_hdr): Likewise.
(bfd_elf64_archive_write_ar_hdr): Likewise.
(bfd_elf64_archive_openr_next_archived_file): Likewise.
(bfd_elf64_archive_get_elt_at_index): Likewise.
(bfd_elf64_archive_generic_stat_arch_elt): Likewise.
(bfd_elf64_archive_update_armap_timestamp): Likewise.
* elf64-s390.c (bfd_elf64_archive_functions): Removed.
(bfd_elf64_archive_slurp_armap): Likewise.
(bfd_elf64_archive_write_armap): Likewise.
(bfd_elf64_archive_slurp_extended_name_table): Likewise.
(bfd_elf64_archive_construct_extended_name_table): Likewise.
(bfd_elf64_archive_truncate_arname): Likewise.
(bfd_elf64_archive_read_ar_hdr): Likewise.
(bfd_elf64_archive_write_ar_hdr): Likewise.
(bfd_elf64_archive_openr_next_archived_file): Likewise.
(bfd_elf64_archive_get_elt_at_index): Likewise.
(bfd_elf64_archive_generic_stat_arch_elt): Likewise.
(bfd_elf64_archive_update_armap_timestamp): Likewise.
* elfxx-target.h (TARGET_BIG_SYM): Use _bfd_archive_64_bit on
BFD_JUMP_TABLE_ARCHIVE if USE_64_BIT_ARCHIVE is defined and
bfd_elfNN_archive_functions isn't defined.
(TARGET_LITTLE_SYM): Likewise.
* libbfd-in.h (_bfd_archive_64_bit_slurp_armap): New prototype.
(_bfd_archive_64_bit_write_armap): Likewise.
(_bfd_archive_64_bit_slurp_extended_name_table): New macro.
(_bfd_archive_64_bit_construct_extended_name_table): Likewise.
(_bfd_archive_64_bit_truncate_arname): Likewise.
(_bfd_archive_64_bit_read_ar_hdr): Likewise.
(_bfd_archive_64_bit_write_ar_hdr): Likewise.
(_bfd_archive_64_bit_openr_next_archived_file): Likewise.
(_bfd_archive_64_bit_get_elt_at_index): Likewise.
(_bfd_archive_64_bit_generic_stat_arch_elt): Likewise.
(_bfd_archive_64_bit_update_armap_timestamp): Likewise.
* libbfd.h: Regenerated.
* plugin.c (plugin_vec): Use _bfd_archive_64_bit on
BFD_JUMP_TABLE_ARCHIVE if USE_64_BIT_ARCHIVE is defined.
H.J. Lu [Wed, 25 May 2016 15:40:52 +0000 (08:40 -0700)]
Skip an archive element if not added by linker
During archive rescan to resolve symbol references for files added by
LTO, linker add_archive_element callback is called to check if an
archive element should added. After all IR symbols have been claimed,
linker won't claim new IR symbols and shouldn't add the LTO archive
element. This patch updates linker add_archive_element callback to
return FALSE when seeing an LTO archive element during rescan and
changes ELF linker to skip such archive element.
PR ld/20103
* ldmain.c (add_archive_element): Don't claim new IR symbols
after all IR symbols have been claimed.
* plugin.c (plugin_call_claim_file): Remove no_more_claiming
check.
* testsuite/ld-plugin/lto.exp (pr20103): New proc.
Run PR ld/20103 tests.
* testsuite/ld-plugin/pr20103a.c: New file.
* testsuite/ld-plugin/pr20103b.c: Likewise.
* testsuite/ld-plugin/pr20103c.c: Likewise.
Only generate VMOV.I64 instructions for loading constant floating point values if this instruction is supported by the currently selected fpu.
PR target/2006764
* config/tc-arm.c (move_or_literal_pool): Only generate a VMOV.I64
instruction if supported by the currently selected fpu variant.
* testsuite/gas/arm/vfpv3-ldr_immediate.s: Add test of this PR.
* testsuite/gas/arm/vfpv3-ldr_immediate.d: Update expected disassembly.
Bernhard Heckel [Wed, 25 May 2016 06:47:18 +0000 (08:47 +0200)]
Fortran, typeprint: Forward level of details to be printed for pointers.
Variable "show" was hardcoded to zero for pointer and reference types.
This implementation didn't allow a correct "whatis" print
for those types and results in same output for "ptype" and "whatis".
Before:
(gdb) whatis t3p
type = PTR TO -> ( Type t3
integer(kind=4) :: t3_i
Type t2 :: t2_n
End Type t3 )
After:
(gdb) whatis t3p
type = PTR TO -> ( Type t3 )
Bernhard Heckel [Wed, 25 May 2016 06:47:17 +0000 (08:47 +0200)]
Fortran, typeprint: Decrease level of details when printing elements of a structure.
According to the typeprint's description, the level of details is
decreased by one for the typeprint of elements of a structure.
Before:
(gdb) ptype t3v
type = Type t3
integer(kind=4) :: t3_i
Type t2
integer(kind=4) :: t2_i
Type t1
integer(kind=4) :: t1_i
real(kind=4) :: t1_r
End Type t1 :: t1_n
End Type t2 :: t2_n
End Type t3
After:
(gdb) ptype t3v
type = Type t3
integer(kind=4) :: t3_i
Type t2 :: t2_n
End Type t3
Bernhard Heckel [Wed, 25 May 2016 06:47:17 +0000 (08:47 +0200)]
Fortran, typeprint: Take level of details into account when printing elements of a structure.
According to the typeprint's description, elements of a structure
should not be printed when show is < 1.
This variable is also used to distinguish the level of details
between "ptype" and "whatis" expressions.
Before:
(gdb) whatis t1v
type = Type t1
integer(kind=4) :: t1_i
real(kind=4) :: t1_r
End Type t1
Tom Tromey [Sun, 14 Sep 2014 04:24:50 +0000 (22:24 -0600)]
Fix PR python/17386 - add __index__ method to gdb.Value
This patch fixes PR python/17386.
The bug is that gdb.Value does not implement the Python __index__
method. This method is needed to convert a Python object to an index
and is used by various operations in Python, such as indexing an
array.
The fix is to implement the nb_index method for gdb.Value.
nb_index was added in Python 2.5. I don't have a good way to test
Python 2.4, but I made an attempt to accomodate it.
I chose to use valpy_long in all cases because this simplifies porting
to Python 3, and because there didn't seem to be any harm.
Tom Tromey [Mon, 23 May 2016 16:58:51 +0000 (10:58 -0600)]
add nb_inplace_divide for python 2
Python 2's PyNumberMethods has nb_inplace_divide, but Python 3 does
not. This patch adds it for Python 2.
This buglet didn't cause much fallout because the only non-NULL entry
in value_object_as_number after this is for valpy_divide; and the
missing slot caused it to slide up to nb_floor_divide (where
nb_true_divide was intended).
Tom Tromey [Thu, 19 May 2016 03:41:28 +0000 (21:41 -0600)]
Fix PR python/17981
PR python/17981 notes that gdb.breakpoints() returns None when there
are no breakpoints; whereas an empty list or tuple would be more in
keeping with Python and the documentation.
This patch fixes the bug by changing the no-breakpoint return to make
an empty tuple.
Pedro Alves [Tue, 24 May 2016 13:47:57 +0000 (14:47 +0100)]
Fix PR gdb/19828: gdb -p <process from a container>: internal error
When GDB attaches to a process, it looks at the /proc/PID/task/ dir
for all clone threads of that process, and attaches to each of them.
Usually, if there is more than one clone thread, it means the program
is multi threaded and linked with pthreads. Thus when GDB soon after
attaching finds and loads a libthread_db matching the process, it'll
add a thread to the thread list for each of the initially found
lower-level LWPs.
If, however, GDB fails to find/load a matching libthread_db, nothing
is adding the LWPs to the thread list. And because of that, "detach"
hits an internal error:
(gdb) PASS: gdb.threads/clone-attach-detach.exp: fg attach 1: attach
info threads
Id Target Id Frame
* 1 LWP 6891 "clone-attach-de" 0x00007f87e5fd0790 in __nanosleep_nocancel () at ../sysdeps/unix/syscall-template.S:84
(gdb) FAIL: gdb.threads/clone-attach-detach.exp: fg attach 1: info threads shows two LWPs
detach
.../src/gdb/thread.c:1010: internal-error: is_executing: Assertion `tp' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)
FAIL: gdb.threads/clone-attach-detach.exp: fg attach 1: detach (GDB internal error)
From here:
...
#8 0x00000000007ba7cc in internal_error (file=0x98ea68 ".../src/gdb/thread.c", line=1010, fmt=0x98ea30 "%s: Assertion `%s' failed.")
at .../src/gdb/common/errors.c:55
#9 0x000000000064bb83 in is_executing (ptid=...) at .../src/gdb/thread.c:1010
#10 0x00000000004c23bb in get_pending_status (lp=0x12c5cc0, status=0x7fffffffdc0c) at .../src/gdb/linux-nat.c:1235
#11 0x00000000004c2738 in detach_callback (lp=0x12c5cc0, data=0x0) at .../src/gdb/linux-nat.c:1317
#12 0x00000000004c1a2a in iterate_over_lwps (filter=..., callback=0x4c2599 <detach_callback>, data=0x0) at .../src/gdb/linux-nat.c:899
#13 0x00000000004c295c in linux_nat_detach (ops=0xe7bd30, args=0x0, from_tty=1) at .../src/gdb/linux-nat.c:1358
#14 0x000000000068284d in delegate_detach (self=0xe7bd30, arg1=0x0, arg2=1) at .../src/gdb/target-delegates.c:34
#15 0x0000000000694141 in target_detach (args=0x0, from_tty=1) at .../src/gdb/target.c:2241
#16 0x0000000000630582 in detach_command (args=0x0, from_tty=1) at .../src/gdb/infcmd.c:2975
...
Tested on x86-64 Fedora 23. Also confirmed the test passes against
gdbserver with "maint set target-non-stop".
Pedro Alves [Tue, 24 May 2016 13:47:57 +0000 (14:47 +0100)]
Make gdb/linux-nat.c consider a waitstatus pending on the infrun side
Working on the fix for gdb/19828, I saw
gdb.threads/attach-many-short-lived-threads.exp fail once in an
unusual way. Unfortunately I didn't keep debug logs, but it's an
issue similar to what's been fixed in remote.c a while ago --
linux-nat.c was not fetching the pending status from the right place.
Pedro Alves [Tue, 24 May 2016 13:47:57 +0000 (14:47 +0100)]
[Linux] Optimize PID -> struct lwp_info lookup
Hacking the gdb.threads/attach-many-short-lived-threads.exp test to
spawn thousands of threads instead of dozens, and running gdb under
perf, I saw that GDB was spending most of the time in find_lwp_pid:
attach_proc_task_lwp_callback is called once for each LWP that we
attach to, found by listing the /proc/PID/task/ directory. In turn,
attach_proc_task_lwp_callback calls find_lwp_pid to check whether the
LWP we're about to try to attach to is already known. Since
find_lwp_pid does a linear walk over the whole LWP list, this becomes
quadratic. We do the /proc/PID/task/ listing until we get two
iterations in a row where we found no new threads. So the second and
following times we walk the /proc/PID/task/ dir, we're going to take
an even worse find_lwp_pid hit.
Fix this by adding a hash table keyed by LWP PID, for fast lookup.
The linked list embedded in the LWP structure itself is kept, and made
a double-linked list, so that removals from that list are O(1). An
earlier version of this patch got rid of this list altogether, but
that revealed hidden dependencies / assumptions on how the list is
sorted. For example, killing a process and then waiting for all the
LWPs status using iterate_over_lwps only works as is because the
leader LWP is always last in the list. So I thought it better to take
an incremental approach and make this patch concern itself _only_ with
the PID lookup optimization.
Pedro Alves [Tue, 24 May 2016 13:47:57 +0000 (14:47 +0100)]
[Linux] Avoid refetching core-of-thread if thread hasn't run
Hacking the gdb.threads/attach-many-short-lived-threads.exp test to
spawn thousands of threads instead of dozens, I saw GDB having trouble
keeping up with threads being spawned too fast, when it tried to stop
them all. This was because while gdb is doing that, it updates the
thread list to make sure no new thread has sneaked in that might need
to be paused. It does this a few times until it sees no-new-threads
twice in a row. The thread listing update itself is not that
expensive, however, in the Linux backend, updating the threads list
calls linux_common_core_of_thread for each LWP to record on which core
each LWP was last seen running, which opens/reads/closes a /proc file
for each LWP which becomes expensive when you need to do it for
thousands of LWPs.
perf shows gdb in linux_common_core_of_thread 44% of the time, in the
stop_all_threads -> update_thread_list path in this use case.
This patch simply makes linux_common_core_of_thread avoid updating the
core the thread is bound to if the thread hasn't run since the last
time we updated that info. This makes linux_common_core_of_thread
disappear into the noise in the perf report.
Pedro Alves [Tue, 24 May 2016 13:47:56 +0000 (14:47 +0100)]
[Linux] Read vDSO range from /proc/PID/task/PID/maps instead of /proc/PID/maps
... as it's _much_ faster.
Hacking the gdb.threads/attach-many-short-lived-threads.exp test to
spawn thousands of threads instead of dozens to stress and debug
timeout problems with gdb.threads/attach-many-short-lived-threads.exp,
I saw that GDB would spend several seconds just reading the
/proc/PID/smaps file, to determine the vDSO mapping range. GDB opens
and reads the whole file just once, and caches the result, but even
that is too slow. For example, with almost 8000 threads:
$ ls /proc/3518/task/ | wc -l
7906
reading the /proc/PID/smaps file grepping for "vdso" takes over 15
seconds :
The problem is that /proc/PID/smaps wants to show the mappings as
being thread stack, and that has the kernel iterating over all threads
in the thread group, for each mapping.
The fix is to use the "map" file under /proc/PID/task/PID/ instead of
the /proc/PID/ one, as the former doesn't mark thread stacks for all
threads.
That alone drops the timing to the millisecond range on my machine:
Pedro Alves [Tue, 24 May 2016 13:47:56 +0000 (14:47 +0100)]
Linux native thread create/exit events support
A following patch (fix for gdb/19828) makes linux-nat.c add threads to
GDB's thread list earlier in the "attach" sequence, and that causes a
surprising regression on
gdb.threads/attach-many-short-lived-threads.exp on my machine. The
extra "thread x exited" handling and traffic slows down that test
enough that GDB core has trouble keeping up with new threads that are
spawned while trying to stop existing ones.
I saw the exact same issue with remote/gdbserver a while ago and fixed
it in 65706a29bac5 (Remote thread create/exit events) so part of the
fix here is the exact same -- add support for thread created events to
gdb/linux-nat.c. infrun.c:stop_all_threads enables those events when
it tries to stop threads, which ensures that new threads never get a
chance to themselves start new threads, thus fixing the race.
PR gdb/19828
* linux-nat.c (report_thread_events): New global.
(linux_handle_extended_wait): Report
TARGET_WAITKIND_THREAD_CREATED if thread event reporting is
enabled.
(wait_lwp, linux_nat_filter_event): Report all thread exits if
thread event reporting is enabled. Remove comment.
(filter_exit_event): New function.
(linux_nat_wait_1): Use it.
(linux_nat_thread_events): New function.
(linux_nat_add_target): Install it as target_thread_events method.
MIPS/GAS: Treat local jump relocs the same no matter if REL or RELA
Do not convert jump relocs against local MIPS16 or microMIPS symbols to
refer to a section symbol instead even on RELA targets, as it makes it
impossible for the linker to make a JAL to JALX conversion based on ISA
symbol annotation, breaking regular and compressed MIPS interlinking.
gas/
* config/tc-mips.c (mips_fix_adjustable): Also return 0 for
jump relocations against MIPS16 or microMIPS symbols on RELA
targets.
* testsuite/gas/mips/jalx-local.d: New test.
* testsuite/gas/mips/jalx-local-n32.d: New test.
* testsuite/gas/mips/jalx-local-n64.d: New test.
* testsuite/gas/mips/jalx-local.s: New test source.
* testsuite/gas/mips/mips.exp: Run the new tests.
ld/
* testsuite/ld-mips-elf/jalx-local.d: New test.
* testsuite/ld-mips-elf/jalx-local-n32.d: New test.
* testsuite/ld-mips-elf/jalx-local-n64.d: New test.
* testsuite/ld-mips-elf/mips-elf.exp: Run the new tests.
MIPS/GAS: Cut TLS reloc dead code path in `md_apply_fix'
With code refactoring made in commit b886a2ab0d52 and the addition of
`calculate_reloc' and a separate test for TLS relocs against constants
made there the preexisting fall-through from the TLS reloc switch case
has effectively become a dead execution path. This is because the call
to `calculate_reloc' present there is only made if `fixP->fx_done' is
true, which can only be the case if `fixP->fx_addsy' is NULL, which in
turn has already triggered the TLS reloc test and made execution break
out of the switch statement.
Remove the fall-through then and reshape code accordingly.
* config/tc-xtensa.c (struct suffix_reloc_map): Change type of field
operator to operatorT.
(map_suffix_reloc_to_operator): Change return type to operatorT.
Trevor Saunders [Sat, 21 May 2016 08:21:32 +0000 (04:21 -0400)]
ft32: fixup TARGET_FORMAT
Nothing ever assigns to ft32_target_format, so its always null, which means the
bfd target arch is the default one. It looks like ft32 only has one target
format, so we can just define TARGET_FORMAT to be that literal string.
Yao Qi [Mon, 23 May 2016 16:32:56 +0000 (17:32 +0100)]
Skip unwritable frames in command "finish"
Nowadays, GDB can't insert breakpoint on the return address of the
exception handler on ARM M-profile, because the address is a magic
one 0xfffffff9,
(gdb) bt
#0 CT32B1_IRQHandler () at ../src/timer.c:67
#1 <signal handler called>
#2 main () at ../src/timer.c:127
(gdb) info frame
Stack level 0, frame at 0x200ffa8:
pc = 0x4ec in CT32B1_IRQHandler (../src/timer.c:67); saved pc = 0xfffffff9
called by frame at 0x200ffc8
source language c.
Arglist at 0x200ffa0, args:
Locals at 0x200ffa0, Previous frame's sp is 0x200ffa8
Saved registers:
r7 at 0x200ffa0, lr at 0x200ffa4
(gdb) x/x 0xfffffff9
0xfffffff9: Cannot access memory at address 0xfffffff9
(gdb) finish
Run till exit from #0 CT32B1_IRQHandler () at ../src/timer.c:67
Ed:15: Target error from Set break/watch: Et:96: Pseudo-address (0xFFFFFFxx) for EXC_RETURN is invalid (GDB error?)
Warning:
Cannot insert hardware breakpoint 0.
Could not insert hardware breakpoints:
You may have requested too many hardware breakpoints/watchpoints.
Command aborted.
even some debug probe can't set hardware breakpoint on the magic
address too,
(gdb) hbreak *0xfffffff9
Hardware assisted breakpoint 2 at 0xfffffff9
(gdb) c
Continuing.
Ed:15: Target error from Set break/watch: Et:96: Pseudo-address (0xFFFFFFxx) for EXC_RETURN is invalid (GDB error?)
Warning:
Cannot insert hardware breakpoint 2.
Could not insert hardware breakpoints:
You may have requested too many hardware breakpoints/watchpoints.
Command aborted.
The problem described above is quite similar to PR 8841, in which GDB
can't set breakpoint on signal trampoline, which is mapped to a read-only
page by kernel. The rationale of this patch is to skip "unwritable"
frames when looking for caller frames in command "finish", and a new
gdbarch method code_of_frame_writable is added. This patch fixes
the problem on ARM cortex-m target, but it can be used to fix
PR 8841 too.
This fixes PR python/19438 and PR python/18393. Both bugs are about
invoking dir() on some Python object implemented by gdb, and getting a
crash.
The crash happens because the dictionary field of these objects was
not initialized. Apparently what happens is that this field can be
lazily initialized by Python when assigning to an attribute; and it
can also be handled ok when using dir() but without __dict__ defined;
but gdb defines __dict__ because this isn't supplied automatically by
Python.
The docs on this seem rather sparse, but this patch works ok.
An alternative might be to lazily create the dictionary in
gdb_py_generic_dict, but I went with this approach because it seemed
more straightforward.
Yao Qi [Mon, 23 May 2016 14:45:12 +0000 (15:45 +0100)]
Use standard_testfile in gdb.arch/thumb-prologue.exp and gdb.arch/thumb2-it.exp
This patch fixes the errors below:
Running /home/yao/SourceCode/gnu/gdb/git/gdb/testsuite/gdb.arch/thumb-prologue.exp ...
gdb compile failed, arm-linux-gnueabihf/bin/ld: cannot open output file /scratch/yao/gdb/build-git/arm-linux-gnueabihf/gdb/testsuite/gdb.arch/thumb-prologue: No such file or directory
collect2: error: ld returned 1 exit status
Running /home/yao/SourceCode/gnu/gdb/git/gdb/testsuite/gdb.arch/thumb2-it.exp ...
gdb compile failed, arm-linux-gnueabihf/bin/ld: cannot open output file /scratch/yao/gdb/build-git/arm-linux-gnueabihf/gdb/testsuite/gdb.arch/thumb2-it: No such file or directory
bfd/
* elf32-arm.c (arm_dedicated_stub_section_padding): New function.
(elf32_arm_size_stubs): Declare stub_type in a more outer scope and
account for padding for stub section requiring one.
(elf32_arm_build_stubs): Add comment to stress the importance of
zeroing veneer section content.
bfd/
* bfd-in.h (bfd_elf32_arm_keep_private_stub_output_sections): Declare
bfd hook.
* bfd-in2.h: Regenerate.
* elf32-arm.c (arm_dedicated_stub_output_section_required): New
function.
(arm_dedicated_stub_output_section_required_alignment): Likewise.
(arm_dedicated_stub_output_section_name): Likewise.
(arm_dedicated_stub_input_section_ptr): Likewise.
(elf32_arm_create_or_find_stub_sec): Add stub type parameter and
function description comment. Add support for dedicated output stub
section to given stub types.
(elf32_arm_add_stub): Add a stub type parameter and pass it down to
elf32_arm_create_or_find_stub_sec.
(elf32_arm_create_stub): Pass stub type down to elf32_arm_add_stub.
(elf32_arm_size_stubs): Pass stub type when calling
elf32_arm_create_or_find_stub_sec for Cortex-A8 erratum veneers.
(bfd_elf32_arm_keep_private_stub_output_sections): New function.
ld/
* emultempl/armelf.em (arm_elf_before_allocation): Call
bfd_elf32_arm_keep_private_stub_output_sections before generic
before_allocation function.
Jon Boden [Mon, 23 May 2016 07:46:33 +0000 (08:46 +0100)]
Search for libutil-freebsd as alternative to libutil
GDB needs kinfo_getvmmap() on GNU/kFreeBSD systems same as on
pure FreeBSD. However on these systems the FreeBSD version of libutil
is renamed to libutil-freebsd.
Trevor Saunders [Thu, 19 May 2016 02:39:18 +0000 (22:39 -0400)]
spu: make some constants unsigned
The field in spu_opcode is unsigned, and for some values of opcode we can end
up shifting into the high bit. So avoid possibly creating a negative number
and then assigning it to a unsigned field by shifting an unsigned constant.
Trevor Saunders [Thu, 19 May 2016 03:48:48 +0000 (23:48 -0400)]
tic54x: rename typedef of struct symbol_
generic gas code has a struct symbol, and tic54x typedefs a struct to symbol.
This seems at least rather confusing, and it seems like target specific headers
shouldn't put such generic names in the global namespace preventing other
generic code from using them.
MIPS: Fix the encoding of immediates with microMIPS JALX
The microMIPS JALX instruction shares the R_MICROMIPS_26_S1 relocation
with microMIPS J/JAL/JALS instructions, however unlike the latters its
encoded immediate argument is unusually shifted left by 2 rather than 1
in calculating the value used for the operation requested.
We already handle this exception in `mips_elf_calculate_relocation' in
LD, in a scenario where JALX is produced as a result of relaxing JAL for
the purpose of making a cross-mode jump. We also get it right in the
disassembler in `decode_micromips_operand'.
What we don't correctly do however is processing microMIPS JALX produced
by GAS from an assembly source, where a non-zero constant argument or a
symbol reference with a non-zero in-place addend has been used. In this
case the same calculation is made as for microMIPS J/JAL/JALS, causing
the wrong encoding to be produced by GAS on making an object file, and
then again by LD in the final link. The latter in particular causes the
calculation, where the addend fits in the relocatable field, to produce
different final addresses for the same source code depending on whether
REL or RELA relocations are used.
Correct these issues by special-casing microMIPS JALX in the places that
have been previously missed.
bfd/
* elfxx-mips.c (mips_elf_read_rel_addend): Adjust the addend for
microMIPS JALX.
gas/
* config/tc-mips.c (append_insn): Correct the encoding of a
constant argument for microMIPS JALX.
(tc_gen_reloc): Correct the encoding of an in-place addend for
microMIPS JALX.
* testsuite/gas/mips/jalx-addend.d: New test.
* testsuite/gas/mips/jalx-addend-n32.d: New test.
* testsuite/gas/mips/jalx-addend-n64.d: New test.
* testsuite/gas/mips/jalx-imm.d: New test.
* testsuite/gas/mips/jalx-imm-n32.d: New test.
* testsuite/gas/mips/jalx-imm-n64.d: New test.
* testsuite/gas/mips/jalx-addend.s: New test source.
* testsuite/gas/mips/jalx-imm.s: New test source.
* testsuite/gas/mips/mips.exp: Run the new tests.
ld/
* testsuite/ld-mips-elf/jalx-addend.d: New test.
* testsuite/ld-mips-elf/jalx-addend-n32.d: New test.
* testsuite/ld-mips-elf/jalx-addend-n64.d: New test.
* testsuite/ld-mips-elf/mips-elf.exp: Run the new tests.
Fix non-deterministic behavior when generating MIPS GOT.
* mips.cc (Mips_got_entry::Mips_got_entry): Remove object argument
for global got symbols, and set addend to 0.
(Mips_got_entry::hash): Change hash algorithm.
(Mips_got_entry::equals): Refactor.
(Mips_got_entry::object): Return input object for local got symbols
from union d.
(Mips_got_entry::addend): Change return of the relocation addend.
(Mips_got_entry::addend_): Move from union d.
(Mips_got_entry::object_): Move into union d.
(class Mips_symbol_hash): New class.
(Mips_got_info::Global_got_entry_set): New type.
(Mips_got_info::global_got_symbols): Change return type to
Global_got_entry_set.
(Mips_got_info::global_got_symbols_): Change type to
Global_got_entry_set.
(Mips_symbol::hash): New method.
(Mips_output_data_la25_stub::symbols_): Change type to std::vector.
(Mips_output_data_mips_stubs::Mips_stubs_entry_set): New type.
(Mips_output_data_mips_stubs::symbols_): Change type to
Mips_stubs_entry_set.
(Mips_got_info::record_global_got_symbol): Don't pass object
argument when creating global got symbol.
(Mips_got_info::record_got_entry): Remove find before inserting
got entries.
(Mips_got_info::add_reloc_only_entries): Change type of iterator
to Global_got_entry_set.
(Mips_got_info::count_got_symbols): Likewise.
(Mips_output_data_la25_stub::create_la25_stub): Use push_back
for adding entries to symbols_.
(Mips_output_data_la25_stub::do_write): Change type of iterator
to std::vector.
(Mips_output_data_mips_stubs::set_lazy_stub_offsets): Change type
of iterator to Mips_stubs_entry_set.
(Mips_output_data_mips_stubs::set_needs_dynsym_value): Likewise.
(Mips_output_data_mips_stubs::do_write): Likewise.
* emulparams/arcelf.sh: Changed.
* emulparams/arclinux.sh: Likewise.
* scripttempl/arclinux.sc: Moved to a more standard implementation
similar to elf.sc.
The `am34-*-linux*' target cannot be configured for, `am34' is not a CPU
name recognized by `config.sub'. It has never been, required code has
not been contributed to GNU config, neither before nor since the
addition of the target triplet to our configury with commit bfff16424942
("Add MN10300 linker relaxation support for symbol differences") back in
2007. Also there is no difference in actual tool configuration between
the `am34-*-linux*' and `am33_2.0-*-linux*' targets, except from a
different executable prefix and tooldir name.
Given the above remove the target triplet from our configuration.
Alan Modra [Wed, 18 May 2016 14:39:03 +0000 (00:09 +0930)]
ld-elf/shared.exp mix_pic_and_non_pic
Tweaks to make it easier to re-run these testcases by hand.
* testsuite/ld-elf/shared.exp (mix_pic_and_non_pic): Pass in
exe name rather than constructing testname. Fix typo in
sub-test name. Log copying. Use -rpath rather than -R.
Andrew Burgess [Fri, 6 May 2016 12:59:03 +0000 (13:59 +0100)]
gas/arc: Make member of arc_flags const
By making the flgp field of struct arc_flags constant we can remove a
place where we cast away the const-ness of a variable. Also, given that
the value assigned to this field almost always comes from compile-time
constant data, having the field non-constant is probably a bad thing.
gas/ChangeLog:
* config/tc-arc.c (find_opcode_match): Remove casting away of
const.
* config/tc-arc.h (struct arc_flags): Make flgp field const.
projects / binutils.git / log