The arc700 nps variant uses some specific named sections to
facilitate its CMEM based instructions, which operate on memory
regions fixed at specific addresses.
This commit adds the arclinux_nps emulation, which is based on the
arclinux emulation. The new emulation uses the OTHER_SECTIONS
variable in the emulation parameters script to provide symbols
mapping to various locations in the CMEM space on nps.
The accompanying test ensures that all symbols are provided at the
correct address in the linker script.
ld/ChangeLog:
* Makefile.am: Add earclinux_nps.c target and add to
ALL_EMULATION_SOURCES.
* Makefile.in: Likewise, regenerated.
* configure.tgt: Add arclinux_nps as an extra emulation for
arc*-*-elf* and arc*-*-linux*.
* emulparams/arc-nps.sh: New file.
* emulparams/arclinux_nps.sh: New file.
* testsuite/ld-arc/arclinux-nps.d,
* testsuite/ld-arc/arclinux-nps.s: New test.
gas/
* read.c (emit_leb128_expr): Extended unsigned big number for
sleb128.
* testsuite/gas/all/gas.exp (test_cond): Add sleb128-8 test.
* testsuite/gas/all/sleb128.d: New test.
* testsuite/gas/all/sleb128.s: New test source.
PowerPC64 defines .TOC. rather than _GLOBAL_OFFSET_TABLE_, and
what's more, doesn't define it at all unless referenced. For ELFv1
ABI the symbol isn't normally referenced, so modify the test to accept
.TOC. as a variant of _GLOBAL_OFFSET_TABLE_ and 0 or 1 occurrences.
copy_test_relro as written doesn't need copy relocs on PowerPC64.
PowerPC64 is always PIC. So, modify copy_test_relro to test that the
existing vars are in fact read-only directly by deliberately causing a
sigsegv, and add another couple of vars that do cause copy relocs even
when PIC.
* testsuite/ver_test_8.sh: Accept .TOC. in lieu of
_GLOBAL_OFFSET_TABLE_. Allow zero count.
* testsuite/copy_test_relro_1.cc (c, q): New vars.
* testsuite/copy_test_relro.cc: Rewrite to test read-only
status of variables directly. Reference new vars in
read-only data.
Added just to accept, and ignore. gcc since 2015-10-21, when
configured with --enable-secureplt passes this option to the linker.
As powerpc gold cannot link --bss-plt code successfully, gold needs to
accept the option or the gcc specs file needs to be changed.
The patch also make gold detect --bss-plt code and error out rather
than producing a binary that crashes.
* options.h: Add --secure-plt option.
* powerpc.cc (Target_powerpc::Scan::local): Detect and error
on -fPIC -mbss-plt code.
(Target_powerpc::Scan::global): Likewise.
Before this patch, some functions would read the inferior memory with
(*the_target)->read_memory, which returns the raw memory, rather than the
shadowed memory.
This is wrong since these functions do not expect to read a breakpoint
instruction and can lead to invalid behavior.
Use of raw memory in get_next_pcs_read_memory_unsigned_integer for example
could lead to get_next_pc returning an invalid pc.
Here's how this would happen:
In non-stop:
the user issues:
thread 1
step&
thread 2
step&
thread 3
step&
In a similar way as non-stop-fair-events.exp (threads are looping).
GDBServer:
linux_resume is called
GDBServer has pending events,
threads are not resumed and single-step breakpoint for thread 1 not installed.
linux_wait_1 is called with a pending event on thread 2 at pc A
GDBServer handles the event and calls proceed_all_lwps
This calls proceed_one_lwp and installs single-step breakpoints on all
the threads that need one.
Now since thread 1 needs to install a single-step breakpoint and is at pc B
(different than thread 2), a step-over is not initiated and get_next_pc
is called to figure out the next instruction from pc B.
However it may just be that thread 3 as a single step breakpoint at pc
B. And thus get_next_pc fails.
This situation is tested with non-stop-fair-events.exp.
In other words, single-step breakpoints are installed in proceed_one_lwp
for each thread. GDBserver proceeds two threads for resume_step, as
requested by GDB, and the thread proceeded later may see the single-step
breakpoints installed for the thread proceeded just now.
Tested on gdbserver-native/-m{thumb,arm} no regressions.
gdb/gdbserver/ChangeLog:
* linux-aarch32-low.c (arm_breakpoint_kind_from_pc): Use
target_read_memory.
* linux-arm-low.c (get_next_pcs_read_memory_unsigned_integer): Likewise.
(get_next_pcs_syscall_next_pc): Likewise.
Previously, some pseudoinstructions like "call" only accepted
symbolic addresses and rejected constant addresses with an
esoteric internal error. This patch enables them by deferring
application of constant relocations to md_apply_fix, rather than
eagerly applying them during instruction assembly.
gas/ChangeLog
2017-01-09 Andrew Waterman <andrew@sifive.com>
* config/tc-riscv.c (append_insn): Don't eagerly apply relocations
against constants.
(md_apply_fix): Mark relocations against constants as "done."
TLS relocs against constants previously segfaulted, and illegal
symbol subtractions were silently ignored.
The previous behavior was to segfault.
gas/ChangeLog
2017-01-09 Andrew Waterman <andrew@sifive.com>
* config/tc-riscv.c (md_apply_fix): Report TLS relocations against
constants. Report disallowed symbol subtractions.
This was added so compressed loads could have smaller offsets for
accessing the data section, but the result was that writable sections
ended up in INITIAL_READONLY_SECTIONS. This is a bad idea. The fix is
to just remove this micro-optimization.
Thanks to Alan Morda for finding the problem!
ld/ChangeLog
2017-01-09 Palmer Dabbelt <palmer@dabbelt.com>
Kito Cheng <kito.cheng@gmail.com>
* emulparams/elf32lriscv-defs.sh (INITIAL_READONLY_SECTIONS):
Removed.
(SDATA_START_SYMBOLS): Likewise.
bfd * dwarf2.c (lookup_address_in_function_table): Return early if
there are no functions in the given comp unit, or if the high
address of the last function in the comp unit is less than the
desired address.
binutils * objdump.c (display_file): Add new parameter 'last_file'. If
last_file is true, do not call bfd_close at the end of the
function.
(main): Set the value of the last_file parameter when calling
display_file.
While casting works as expected with expression debugging turned off,
this seems to be an indication that the D language parser function is
doing something wrong in the building of the expression.
Without changing the grammar, using UNOP_CAST_TYPE is the right thing to
do here, as the TypeExp handler has already wrapped the type around a
pair of OP_TYPE opcodes.
gdb/ChangeLog:
* d-exp.y (CastExpression): Emit UNOP_CAST_TYPE.
gdb/testsuite/ChangeLog:
* gdb.dlang/debug-expr.exp: New file.
Building PIE executable from non-PIC code results in broken binaries.
With this patch the problem is detected at link-time.
bfd/ChangeLog:
2017-01-07 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
* elf64-s390.c (elf_s390_relocate_section): Issue error for
non-PLT relocs of shared libary symbol in exectuable.
Plus some paranoia in symval_for_branch. We shouldn't get there with
dynamic symbols, but if we ever did the static_cast to Powerpc_relobj
would be wrong.
* powerpc.cc: Use shorter equivalent elfcpp typedef for
Reltype and reloc_size throughout.
(Target_powerpc::symval_for_branch): Exclude dynamic symbols.
(Target_powerpc::Scan::local): Use local var r_sym.
(Target_powerpc::Scan::global: Likewise.
(Target_powerpc::Relocate::relocate): Delete shadowing r_sym.
$ make check-headers CHECK_HEADERS="x86-linux-nat.h"
...
../../binutils-gdb/gdb/x86-linux-nat.h:29:8: error: 'ps_err_e' does not name a type
extern ps_err_e x86_linux_get_thread_area (pid_t pid, void *addr,
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* x86-linux-nat.h: Include gdb_proc_service.h.
$ make check-headers CHECK_HEADERS="ser-base.h"
...
../../binutils-gdb/gdb/ser-base.h:33:8: error: 'serial_ttystate' does not name a type
extern serial_ttystate ser_base_get_tty_state (struct serial *scb);
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* ser-base.h: Include serial.h.
$ make check-headers CHECK_HEADERS="ppc-linux-tdep.h"
...
../../binutils-gdb/gdb/ppc-linux-tdep.h:34:24: error: 'PPC_NUM_REGS' was not declared in this scope
PPC_ORIG_R3_REGNUM = PPC_NUM_REGS,
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* ppc-linux-tdep.h: Include ppc-tdep.h.
$ make check-headers CHECK_HEADERS="nat/amd64-linux-siginfo.h"
....
../../binutils-gdb/gdb/nat/amd64-linux-siginfo.h:52:39: error: 'siginfo_t' was not declared in this scope
int amd64_linux_siginfo_fixup_common (siginfo_t *native, gdb_byte *inf,
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* nat/amd64-linux-siginfo.h: Include signal.h.
$ make check-headers CHECK_HEADERS="nat/aarch64-linux-hw-point.h"
...
../../binutils-gdb/gdb/nat/aarch64-linux-hw-point.h:169:37: error: use of enum 'target_hw_bp_type' without previous declaration
int aarch64_handle_breakpoint (enum target_hw_bp_type type, CORE_ADDR addr,
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* nat/aarch64-linux-hw-point.h: Include break-common.h.
$ make check-headers CHECK_HEADERS="mi/mi-parse.h"
...
../../binutils-gdb/gdb/mi/mi-parse.h:77:6: error: use of enum 'print_values' without previous declaration
enum print_values mi_parse_print_values (const char *name);
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* mi/mi-parse.h: Include mi-cmds.h.
$ make check-headers CHECK_HEADERS="target.h inf-loop.h"
...
../../binutils-gdb/gdb/inf-loop.h:23:42: error: use of enum 'inferior_event_type' without previous declaration
extern void inferior_event_handler (enum inferior_event_type event_type,
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* inf-loop.c: Don't include "target.h".
* inf-loop.h: Include it here.
$ make check-headers CHECK_HEADERS="dfp.h"
...
../../binutils-gdb/gdb/dfp.h:39:8: error: 'DOUBLEST' does not name a type
extern DOUBLEST decimal_to_doublest (const gdb_byte *from, int len,
^
../../binutils-gdb/gdb/dfp.h:41:33: error: use of enum 'exp_opcode' without previous declaration
extern void decimal_binop (enum exp_opcode,
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* dfp.h: Include "dboulest.h" and "expression.h".
$ make check-headers CHECK_HEADERS="ax-gdb.h"
...
../../binutils-gdb/gdb/ax-gdb.h:104:8: error: 'agent_expr_up' does not name a type
extern agent_expr_up gen_trace_for_expr (CORE_ADDR, struct expression *,
^
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* ax-gdb.h: Include "ax.h"
Commit e379037 (Move gdb_ptrace.h to nat/), so we should update
file name in HFILES_NO_SRCDIR too. Otherwise, 'make tags' complains,
$ make tags
make: *** No rule to make target `gdb_ptrace.h', needed by `TAGS'. Stop.
gdb:
2017-01-06 Yao Qi <yao.qi@linaro.org>
* Makefile.in (HFILES_NO_SRCDIR): Replace gdb_ptrace.h
with nat/gdb_ptrace.h.
This patch splits the expression before the && operator instead of
after it.
gdb:
2017-01-05 Yao Qi <yao.qi@linaro.org>
* mips-fbsd-tdep.c (mips_fbsd_sigframe_init): Move && to
new line.
(mips64_fbsd_sigframe_init): Likewise.
PR 20946
* frv-desc.c (lookup_mach_via_bfd_name): Return NULL if the name
could not be matched.
(frv_cgen_cpu_open): Allow for lookup_mach_via_bfd_name returning
NULL.
sim/aarch64/
* cpustate.c: Include math.h.
(aarch64_set_FP_float): Use signbit to check for signed zero.
(aarch64_set_FP_double): Likewise.
* simulator.c (do_vec_MOV_immediate, case 0x8): Add missing break.
(do_vec_mul): In all DO_VEC_WIDENING_MUL calls, make second and fourth
args same size as third arg.
(fmaxnm): Use isnan instead of fpclassify.
(fminnm, dmaxnm, dminnm): Likewise.
(do_vec_MLS): Reverse order of subtraction operands.
(dexSimpleFPCondSelect): Call aarch64_get_FP_double or
aarch64_get_FP_float to get source register contents.
(UINT_MIN, ULONG_MIN, FLOAT_UINT_MAX, FLOAT_UINT_MIN,
DOUBLE_UINT_MAX, DOUBLE_UINT_MIN, FLOAT_ULONG_MAX, FLOAT_ULONG_MIN,
DOUBLE_ULONG_MAX, DOUBLE_ULONG_MIN): New.
(do_fcvtzu): Use ULONG instead of LONG, and UINT instead of INT in
raise_exception calls.
sim/testsuite/sim/aarch64/
* fcsel.s: New.
* fcvtz.s: New.
* fminnm.s: New.
* mls.s: New.
* mul.s: New.
This has been tested for the n64 and o32 ABIs. Signal frame unwinders for
both ABIs are provided. FreeBSD/mips requires custom linkmap offsets since
it contains an additional l_off member in 'struct link_map' that other
FreeBSD platforms do not have. Support for collecting and supplying
general purpose and floating point register sets are provided. Common
routines for working with native format register sets are exported for
use by the native target.
gdb/ChangeLog:
* Makefile.in (ALL_TARGET_OBS): Add mips-fbsd-tdep.o.
(ALLDEPFILES): Add mips-fbsd-tdep.c.
* NEWS: Mention new FreeBSD/mips target.
* configure.tgt: Add mips*-*-freebsd*.
* mips-fbsd-tdep.c: New file.
* mips-fbsd-tdep.h: New file.
gdb/doc/ChangeLog:
* gdb.texinfo (Contributors): Add SRI International and University
of Cambridge for FreeBSD/mips.
This is a modified form of gcc's contrib/update-copyright.py.
* update-copyright.py: New file.
* add-log.el: Update copyright year range.
* texi2pod.pl: Likewise.
The weaker release consistency support of ARMv8.3-A is allowed as an optional
extension for ARMv8.2-A, so separate command line option and feature flag is
added: -march=armv8.2-a+rcpc turns LDAPR, LDAPRB, LDAPRH instructions on.
opcodes/
* aarch64-tbl.h (RCPC, RCPC_INSN): Define.
(aarch64_opcode_table): Use RCPC_INSN.
include/
* opcode/aarch64.h (AARCH64_FEATURE_RCPC): Define.
(AARCH64_ARCH_V8_3): Update.
gas/
* config/tc-aarch64.c (aarch64_features): Add rcpc.
* doc/c-aarch64.texi (AArch64 Extensions): Document rcpc.
* testsuite/gas/aarch64/ldst-exclusive-armv8_3.d: Rename to ...
* testsuite/gas/aarch64/ldst-rcpc.d: This.
* testsuite/gas/aarch64/ldst-exclusive-armv8_3.s: Rename to ...
* testsuite/gas/aarch64/ldst-rcpc.s: This.
* testsuite/gas/aarch64/ldst-rcpc-armv8_2.d: New test.
In ee40d8d (Move computed value's frame id to piece_closure), I only
updated read_pieced_value to use frame_id from piece_closure, but
forgot to update write_pieced_value, so it causes the following
internal error on arm-linux,
set variable l = 4^M
gdb/git/gdb/value.c:1579: internal-error: frame_id* deprecated_value_next_frame_id_hack(value*): Assertion `value->lval == lval_register' failed.^M
A problem internal to GDB has been detected,^M
further debugging may prove unreliable.^M
Quit this debugging session? (y or n) FAIL: gdb.base/store.exp: var longest l; setting l to 4 (GDB internal error)
This patch fixes the internal error.
gdb:
2017-01-04 Yao Qi <yao.qi@linaro.org>
* dwarf2loc.c (write_pieced_value): Don't use VALUE_FRAME_ID (to),
use c->frame_id when the piece location is DWARF_VALUE_REGISTER.
Nick Clifton
Graham Markall
Tristan Gingold
Alan Modra
GDB Administrator
Jim Wilson
Antoine Tremblay
Andrew Waterman
Palmer Dabbelt
Iain Buclaw
Andreas Krebbel
Yao Qi
John Baldwin
Dilan Palauzov
Jiong Wang
Szabolcs Nagy
Norm Jacobs