Alan Modra [Fri, 26 May 2017 00:32:29 +0000 (10:02 +0930)]
PPC64_OPT_LOCALENTRY
ELFv2 functions with localentry:0 are those with a single entry point,
ie. global entry == local entry, and that have no requirement on r2 or
r12, and guarantee r2 is unchanged on return. Such an external
function can be called via the PLT without saving r2 or restoring it
on return, avoiding a common load-hit-store for small functions. The
optimization is attractive. The TOC pointer load-hit-store is a major
reason why calls to small functions that need no register saves, or
with shrink-wrap, no register saves on a fast path, are slow on
powerpc64le.
To be safe, this optimization needs ld.so support to check that the
run-time matches link-time function implementation. If a function
in a shared library with st_other localentry non-zero is called
without saving and restoring r2, r2 will be trashed on return, leading
to segfaults. For that reason the optimization does not happen for
weak functions since a weak definition is a fairly solid hint that the
function will likely be overridden. I'm also not enabling the
optimization by default unless glibc-2.26 is detected, which should
have the ld.so checks implemented.
bfd/
* elf64-ppc.c (struct ppc_link_hash_table): Add has_plt_localentry0.
(ppc64_elf_merge_symbol_attribute): Merge localentry bits from
dynamic objects.
(is_elfv2_localentry0): New function.
(ppc64_elf_tls_setup): Default params->plt_localentry0.
(plt_stub_size): Adjust size for tls_get_addr_opt stub.
(build_tls_get_addr_stub): Use a simpler stub when r2 is not saved.
(ppc64_elf_size_stubs): Leave stub_type as ppc_stub_plt_call for
optimized localentry:0 stubs.
(ppc64_elf_build_stubs): Save r2 in ELFv2 __glink_PLTresolve.
(ppc64_elf_relocate_section): Leave nop unchanged for optimized
localentry:0 stubs.
(ppc64_elf_finish_dynamic_sections): Set PPC64_OPT_LOCALENTRY in
DT_PPC64_OPT.
* elf64-ppc.h (struct ppc64_elf_params): Add plt_localentry0.
include/
* elf/ppc64.h (PPC64_OPT_LOCALENTRY): Define.
ld/
* emultempl/ppc64elf.em (params): Init plt_localentry0 field.
(enum ppc64_opt): New, replacing OPTION_* defines. Add
OPTION_PLT_LOCALENTRY, and OPTION_NO_PLT_LOCALENTRY.
(PARSE_AND_LIST_*): Support --plt-localentry and --no-plt-localentry.
* testsuite/ld-powerpc/elfv2so.d: Update.
* testsuite/ld-powerpc/powerpc.exp (TLS opt 5): Use --no-plt-localentry.
* testsuite/ld-powerpc/tlsopt5.d: Update.
Andreas Krebbel [Thu, 1 Jun 2017 11:46:15 +0000 (13:46 +0200)]
S/390: idte/ipte fixes
Later CPU generations added optional operands to the ipte/idte
instructions. I've added these with:
https://sourceware.org/ml/binutils/2017-05/msg00316.html ... but
supported the optional operands only with the specific hardware
levels. However, it is more useful to have the optional operands
already in the first versions. Of course they need to be zero there.
Regression-tested with on s390 and s390x. Committed to mainline.
Bye,
-Andreas-
opcodes/ChangeLog:
2017-06-01 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
* s390-opc.txt: Support the optional parameters with the first
versions of ipte/idte.
gas/ChangeLog:
2017-06-01 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
Markus Metzger [Mon, 21 Nov 2016 13:27:57 +0000 (14:27 +0100)]
btrace: support decoder events
Newer versions of libipt support instruction flow decoder events instead of
indicating those events with flags in struct pt_insn. Add support for them in
GDB.
gdb/
* btrace.c (handle_pt_insn_events): New.
(ftrace_add_pt): Call handle_pt_insn_events. Rename ERRCODE into
STATUS. Split into this and ...
(handle_pt_insn_event_flags): ... this.
Markus Metzger [Mon, 21 Nov 2016 10:46:15 +0000 (11:46 +0100)]
config, btrace: check for pt_insn_event in libipt
Version 2 of libipt adds an event system to instruction flow decoders and
deprecates indicating events via flags in struct pt_insn. Add configuration
checks to determine which version we have.
gdb/
* configure.ac: Check for pt_insn_event, struct pt_insn.enabled,
and struct pt_insn.resynced.
* configure: Regenerated.
* config.in: Regenerated.
Anton Kolesov [Thu, 16 Mar 2017 12:21:31 +0000 (15:21 +0300)]
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
Anton Kolesov [Fri, 7 Apr 2017 14:22:39 +0000 (17:22 +0300)]
[ARC] Add arc-cpu.def with processor definitions
This patch extracts ARC CPU definitions from gas/config/tc-arc.c (cpu_types)
into a separate file arc-cpu.def. This will allow reuse of CPU type definition
in multiple places where it might be needed, for example in disassembler. This
will help ensure that gas and disassembker use same option values for CPUs.
arc-cpu.def file relies on preprocessor macroses which are defined somewhere
else. This for example multiple C files to include arc-cpu.def, but define
different macroses, therefore creating different structures.
include/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* elf/arc-cpu.def: New file.
gas/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* config/tc-arc.c (cpu_types): Include arc-cpu.def
Signed-off-by: Anton Kolesov <Anton.Kolesov@synopsys.com>
Anton Kolesov [Fri, 17 Mar 2017 15:37:42 +0000 (18:37 +0300)]
[ARC] Implement compatible function for ARC BFD architectures
The general rule for bfd_arch_info_type->compatible (A, B) is that if A and B
are compatible, then this function should return architecture that is more
"feature-rich", that is, can run both A and B. ARCv2, EM and HS all has same
mach number, so bfd_default_compatible assumes they are the same, and returns
an A. That causes issues with GDB, because GDB assumes that if machines are
compatible, then "compatible ()" always returns same machine regardless of
argument order. As a result GDB gets confused because, for example,
compatible(ARCv2, EM) returns ARCv2, but compatible(EM, ARCv2) returns EM,
hence GDB is not sure if they are compatible and prints a warning.
bfd/ChangeLog:
yyyy-mm-dd Anton Kolesov Anton.Kolesov@synopsys.com
H.J. Lu [Tue, 30 May 2017 13:34:05 +0000 (06:34 -0700)]
Add bfd_get_file_size to get archive element size
We can't use stat() to get archive element size. Add bfd_get_file_size
to get size for both normal files and archive elements.
bfd/
PR binutils/21519
* bfdio.c (bfd_get_file_size): New function.
* bfd-in2.h: Regenerated.
binutils/
PR binutils/21519
* objdump.c (dump_relocs_in_section): Replace get_file_size
with bfd_get_file_size to get archive element size.
* testsuite/binutils-all/objdump.exp (test_objdump_f): New
proc.
(test_objdump_h): Likewise.
(test_objdump_t): Likewise.
(test_objdump_r): Likewise.
(test_objdump_s): Likewise.
Add objdump tests on archive.
Tim Wiederhake [Tue, 30 May 2017 10:47:37 +0000 (12:47 +0200)]
btrace: Replace struct btrace_function::segment.
This used to hold a pair of pointers to the previous and next function segment
that belong to this function call. Replace with a pair of indices into the
vector of function segments.
Tim Wiederhake [Tue, 30 May 2017 10:47:37 +0000 (12:47 +0200)]
btrace: Remove struct btrace_function::flow.
This used to hold a pair of pointers to the previous and next function segment
in execution flow order. It is no longer necessary as the previous and next
function segments now are simply the previous and next elements in the vector
of function segments.
These are no longer needed and might hold invalid addresses once we change the
vector of function segment pointers into a vector of function segment objects
where a reallocation of the vector changes the address of its elements.
Tim Wiederhake [Tue, 30 May 2017 10:47:37 +0000 (12:47 +0200)]
btrace: Transfer ownership of pointers.
Directly insert new btrace_function pointers into the vector and have the
vector own these pointers. This allows us to later retrieve these objects by
their number directly after creation whereas at the moment we have to wait
until the vector is fully populated.
This requires to pull btrace_thread_info through different functions but
cleans up the code for freeing the trace.
Simon Marchi [Tue, 30 May 2017 09:24:52 +0000 (11:24 +0200)]
gdb.base/watch-cond-infcall.exp: Don't run if target doesn't support infcalls
This test requires calling a function in the inferior, and therefore it
doesn't make sense to run it if the target doesn't support calling
functions from GDB.
gdb/testsuite/ChangeLog:
* gdb.base/watch-cond-infcall.exp: Don't run if target doesn't
support function calls from GDB.
Andreas Krebbel [Mon, 29 May 2017 10:34:56 +0000 (12:34 +0200)]
S/390: Improve error checking for optional operands
So far we only had an instruction flag which made an arbitrary number
of operands optional. This limits error checking capabilities for
instructions marked that way. With this patch the optparm flag only
allows a single optional parameter and another one is added (optparm2)
allowing 2 optional arguments. Hopefully we won't need more than that
in the future. So far there will be only a single use of optparm2.
gas/ChangeLog:
2017-05-30 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
* config/tc-s390.c (md_gather_operands): Support new optparm2
instruction flag.
include/ChangeLog:
2017-05-30 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
* opcode/s390.h: Add new instruction flags optparm2.
opcodes/ChangeLog:
2017-05-30 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
* s390-dis.c (s390_print_insn_with_opcode): Support new optparm2
instruction flag.
* s390-mkopc.c (main): Recognize the new instruction flag when
parsing instruction list.
Yao Qi [Wed, 24 May 2017 21:15:23 +0000 (22:15 +0100)]
Add unit test to gdbarch methods register_to_value and value_to_register
This patch adds one unit test for gdbarch methods register_to_value and
value_to_register. The test pass different combinations of {regnu, type}
to gdbarch_register_to_value and gdbarch_value_to_register. In order
to do the test, add a new function create_new_frame to create a fake
frame. It can be improved after we converted frame_info to class.
In order to isolate regcache (from target_ops operations on writing
registers, like target_store_registers), the sub-class of regcache in the
test override raw_write. Also, in order to get the right regcache from
get_thread_arch_aspace_regcache, the sub-class of regcache inserts itself
to current_regcache.
Suppose I incorrectly modified the size of buffer as below,
@@ -1228,7 +1228,7 @@ ia64_register_to_value (struct frame_info *frame, int regnum,
int *optimizedp, int *unavailablep)
{
struct gdbarch *gdbarch = get_frame_arch (frame);
- gdb_byte in[MAX_REGISTER_SIZE];
+ gdb_byte in[1];
/* Convert to TYPE. */
if (!get_frame_register_bytes (frame, regnum, 0,
build GDB with "-fsanitize=address" and run unittest.exp, asan can detect
such error
==2302==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x7fff98193870 at pc 0xbd55ea bp 0x7fff981935a0 sp 0x7fff98193598
WRITE of size 16 at 0x7fff98193870 thread T0
#0 0xbd55e9 in frame_register_unwind(frame_info*, int, int*, int*, lval_type*, unsigned long*, int*, unsigned char*) /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:1119
#1 0xbd58c8 in frame_register(frame_info*, int, int*, int*, lval_type*, unsigned long*, int*, unsigned char*) /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:1147
#2 0xbd6e25 in get_frame_register_bytes(frame_info*, int, unsigned long, int, unsigned char*, int*, int*) /home/yao/SourceCode/gnu/gdb/git/gdb/frame.c:1427
#3 0x70080a in ia64_register_to_value /home/yao/SourceCode/gnu/gdb/git/gdb/ia64-tdep.c:1236
#4 0xbf570e in gdbarch_register_to_value(gdbarch*, frame_info*, int, type*, unsigned char*, int*, int*) /home/yao/SourceCode/gnu/gdb/git/gdb/gdbarch.c:2619
#5 0xc05975 in register_to_value_test /home/yao/SourceCode/gnu/gdb/git/gdb/gdbarch-selftests.c:131
Or, even if GDB is not built with asan, GDB just crashes.
Yao Qi [Wed, 24 May 2017 21:15:23 +0000 (22:15 +0100)]
Move current_regcache to regcache::current_regcache
This patches moves global variable current_regcache to a class regcache
static variable (protected) so that the unit test I add in the following
patch can access it (by means of extending class regcache in unit test).
gdb:
2017-05-24 Yao Qi <yao.qi@linaro.org>
* regcache.c (current_regcache): Change it to
regcache::current_regcache.
(regcache_observer_target_changed): Update.
(regcache_thread_ptid_changed): Make it a regcache static
method.
(regcache_thread_ptid_changed): Update.
(class regcache_access): New.
(current_regcache_test): Update.
(_initialize_regcache): Update.
* regcache.h: Include forward_list.
(regcache): Declare regcache_thread_ptid_changed and declare
registers_changed_ptid as friend.
Yao Qi [Wed, 24 May 2017 21:15:23 +0000 (22:15 +0100)]
Restrict alpha_convert_register_p
This patch restricts alpha_convert_register_p from
"TYPE_LENGTH (type) != 8" to "TYPE_LENGTH (type) == 4", because,
- we have check "TYPE_LENGTH (valtype) == 4" in alpha_register_to_value
and alpha_value_to_register,
- alpha lds and sts instruction access 4 bytes,
- comments "It might need to convert the [float] register into the
corresponding [integer] type (see Alpha)" and integer is 4-byte on
alpha,
I think it is the right restrict condition to "TYPE_LENGTH (valtype) == 4".
gdb:
2017-05-24 Yao Qi <yao.qi@linaro.org>
* alpha-tdep.c (alpha_convert_register_p): Return true if type
length is 4.
(alpha_register_to_value): Remove type length check.
(alpha_value_to_register): Likewise.
Yao Qi [Wed, 24 May 2017 21:15:23 +0000 (22:15 +0100)]
Restrict m68k_convert_register_p
We need to convert register if the type is float. Suppose we get a value
from float point register, but its type is integer, we don't have to convert.
This case may not exist in real code, but exist in my unit test case.
warning: Cannot convert floating-point register value to non-floating-point type.
Self test failed: arch m68k: self-test failed at gdb/git/gdb/findvar.c:1072
Yao Qi [Wed, 24 May 2017 16:23:52 +0000 (17:23 +0100)]
Move print_insn_XXX to an opcodes internal header
With the changes done in previous patches, print_insn_XXX functions
don't have to be external visible out of opcodes, because both gdb
and objdump select disassemblers through a single interface.
This patch moves these print_insn_XXX declarations from
include/dis-asm.h to opcodes/disassemble.h, which is a new header
added by this patch.
include:
2017-05-24 Yao Qi <yao.qi@linaro.org>
* dis-asm.h: Move some function declarations to
opcodes/disassemble.h.
Yao Qi [Wed, 24 May 2017 16:23:52 +0000 (17:23 +0100)]
Use disassemble.c:disassembler select rl78 disassembler
This patch changes rl78 to let disassble.c:disassembler select
disassembler. rl78_get_disassembler doesn't handle the case
that abfd is NULL, so this patch also fix it.
switch (info.bfd_arch_info->mach)
{
case bfd_mach_h8300:
...
set_gdbarch_print_insn (gdbarch, print_insn_h8300);
case bfd_mach_h8300h:
case bfd_mach_h8300hn:
...
set_gdbarch_print_insn (gdbarch, print_insn_h8300h);
case bfd_mach_h8300s:
case bfd_mach_h8300sn:
...
set_gdbarch_print_insn (gdbarch, print_insn_h8300s);
so we can leave disassble.c:disassembler doing the selection.
Yao Qi [Wed, 24 May 2017 16:23:52 +0000 (17:23 +0100)]
Delegate opcodes to select disassembler in GDB
This patch changes GDB to use disassembler selected by opcodes in
default, so that we don't have to duplicate the selection logic again
in GDB side. For example, gdb/score-tdep.c has
Yao Qi [Wed, 24 May 2017 16:23:52 +0000 (17:23 +0100)]
Refactor disassembler selection
Nowadays, opcodes/disassemble.c:disassembler selects the proper
disassembler according to ABFD only. However, it actually
selects disassemblers according to arch, mach, endianess, and
abfd. This patch adds them to the parameters of disassembler,
so that its caller can still select disassemblers in case that
abfd is NULL (a typical case in GDB).
There isn't any functionality change.
binutils:
2017-05-24 Yao Qi <yao.qi@linaro.org>
* objdump.c (disassemble_data): Caller update.
include:
2017-05-24 Yao Qi <yao.qi@linaro.org>
* dis-asm.h (disassembler): Update declaration.
opcodes:
2017-05-24 Yao Qi <yao.qi@linaro.org>
* disassemble.c (disassembler): Add arguments a, big and mach.
Use them.
John Baldwin [Tue, 18 Apr 2017 17:49:44 +0000 (10:49 -0700)]
Cleanups to FreeBSD/mips native register operations.
Compare against the "raw" PC register number instead of the cooked
register number when determining if a register was handled by
PT_GETREGS. Previously the register fetch/store operations only tried
PT_GETREGS to fetch any individual register. The result was that
fetching or storing an individual register not covered by PT_GETREGS
(such as floating point registers) did not work.
While here, remove an early exit to simplify the code flow from the
PT_GETREGS / PT_SETREGS case, and add a getfpregs_supplies similar to
getregs_supplies to describe the registers supplied by PT_GETFPREGS
and PT_SETFPREGS.
gdb/ChangeLog:
* mips-fbsd-nat.c (getregs_supplies): Fix upper bound comparison.
(getpfpregs_supplies): New function.
(mips_fbsd_fetch_inferior_registers): Remove early exit and use
getfpregs_supplies.
(mips_fbsd_store_inferior_registers): Likewise.
Andrew Burgess [Mon, 15 May 2017 09:11:57 +0000 (10:11 +0100)]
gdb: Document vMustReplyEmpty remote packet
Add mention of the vMustReplyEmpty to the remote serial protocol
documentation. It is important that this packet be treated in the same
fashion as any other unknown 'v' packet, and I have tried to reflect
this in the description of the packet, it is not simply the case that we
_must_ return the empty string for this packet.
As the intention is that we should treat this packet as unknown then an
argument could be made that we should not document it, however, for
someone implementing a gdbserver from scratch, seeing an undocumented
packet arrive from gdb is confusing, and will probably cause them to
have to read the code in order to check how this packet should be
handled, which is not ideal.
H.J. Lu [Mon, 22 May 2017 18:02:46 +0000 (11:02 -0700)]
x86: Add NOTRACK prefix support
For register indirect branches, NOTRACK prefix (0x3e), which is also
the DS segment register prefix, can be used to ignore the CET indirect
branch track.
gas/
* config/tc-i386.c (REX_PREFIX): Changed to 7.
(NOTRACK_PREFIX): New.
(MAX_PREFIXES): Changed to 8.
(_i386_insn): Add notrack_prefix.
(PREFIX_GROUP): Add PREFIX_DS.
(add_prefix): Return PREFIX_DS for DS_PREFIX_OPCODE.
(md_assemble): Check if NOTRACK prefix is supported.
(parse_insn): Set notrack_prefix and issue an error for
other prefixes after NOTRACK prefix.
* testsuite/gas/i386/i386.exp: Run tests for NOTRACK prefix.
* testsuite/gas/i386/notrack-intel.d: New file.
* testsuite/gas/i386/notrack.d: Likewise.
* testsuite/gas/i386/notrack.s: Likewise.
* testsuite/gas/i386/notrackbad.l: Likewise.
* testsuite/gas/i386/notrackbad.s: Likewise.
* testsuite/gas/i386/x86-64-notrack-intel.d: Likewise.
* testsuite/gas/i386/x86-64-notrack.d: Likewise.
* testsuite/gas/i386/x86-64-notrack.s: Likewise.
* testsuite/gas/i386/x86-64-notrackbad.l: Likewise.
* testsuite/gas/i386/x86-64-notrackbad.s: Likewise.
Jiong Wang [Mon, 22 May 2017 08:50:29 +0000 (09:50 +0100)]
[AArch64, ld] Support ILP32 triplet aarch64*-linux-gnu_ilp32
This patch allows AArch64 LD defaulting to ILP32 if it is configured with
aarch64*-linux-gnu_ilp32.
ld/
* configure.tgt: Set "targ_emul" to "aarch64linux32b" for
aarch64_be-*-linux-gnu_ilp32. Set "targ_emul" to "aarch64linux32" for
aarch64-*-linux-gnu_ilp32.
Jiong Wang [Mon, 22 May 2017 08:50:19 +0000 (09:50 +0100)]
[AArch64, gas] Support ILP32 triplet aarch64*-linux-gnu_ilp32
This patch allows AArch64 GAS defaulting to ILP32 if it is configured with
aarch64*-linux-gnu_ilp32.
"md_after_parse_args" is implemented to update ABI into ILP32 if DEFAULT_ARCH is
"aarch64:32".
gas/
* configure.tgt: Set "arch" to "aarch64" if ${cpu} equals "aarch64".
Recognize the new triplet name aarch64*-linux-gnu_ilp32.
* configure.ac: Output DEFAULT_ARCH macro for AArch64.
* configure: Regenerate.
* config/tc-aarch64.h (aarch64_after_parse_args): New declaration.
(md_after_parse_args): New define.
* config/tc-aarch64.c (aarch64_abi_type): New enumeration
AARCH64_ABI_NONE.
(DEFAULT_ARCH): New define.
(aarch64_abi): Set default value to AARCH64_ABI_NONE.
(aarch64_after_parse_args): New function.
Tom Tromey [Sun, 21 May 2017 23:00:10 +0000 (17:00 -0600)]
Print Rust unsized array types a bit more nicely
It's a bit difficult to create an unsized array type in Rust, but if
you do, right now ptype will show something like "[u8; ]". It really
should print "[u8]", though, which is what this patch implements.
This is part of PR 21466.
Built and regtested on x86-64 Fedora 25. I'm checking this in.
ChangeLog
2017-05-21 Tom Tromey <tom@tromey.com>
PR rust/21466:
* rust-lang.c (rust_print_type) <TYPE_CODE_ARRAY>: Print unsized
arrays as "[T]", not "[T; ]".
testsuite/ChangeLog
2017-05-21 Tom Tromey <tom@tromey.com>
PR rust/21466:
* gdb.rust/unsized.exp: New file.
* gdb.rust/unsized.rs: New file.
Jose E. Marchesi [Sat, 20 May 2017 06:45:29 +0000 (23:45 -0700)]
binutils: remove sparc64/mips64 workaround in objcopy build notes merge code
This patch removes a workaround recently installed in objcopy that
avoided removing duplicated notes in targets for which the number of
internal relocations may be bigger than the number of external
relocations. With the recent fixes in sparc64 and mips64, this
workaround is no longer necessary.
2017-05-19 Jose E. Marchesi <jose.marchesi@oracle.com>
* objcopy.c (merge_gnu_build_notes): Remove workaround that
prevented deleting relocations in duplicated notes in mips64 and
sparc.
Tom Tromey [Sun, 14 May 2017 17:12:14 +0000 (11:12 -0600)]
Use watchpoint's language when re-parsing expression
PR rust/21484 notes that watch -location does not work with Rust:
(gdb) watch -location a
syntax error in expression, near `) 0x00007fffffffe0f4'.
update_watchpoint tries to tell gdb that the new expression it creates
has C syntax:
/* The above expression is in C. */
b->language = language_c;
However, update_watchpoint doesn't actually use this language when
re-parsing the expression.
Originally I was going to fix this by saving and restoring the
language in update_watchpoint, but this regressed
gdb.dlang/watch-loc.exp, because the constructed expression actually
has D syntax (specifically the name is not parseable by C).
Next I looked at directly constructing an expression, and not relying
on the parser at all; but it seemed to me that upon a re-set, we'd
want to reparse the type, and there is no existing API to do this
correctly.
So, in the end I made a hook to let each language choose what
expression to use. I made all the languages other than Rust use the C
expression, because that is the status quo ante. However, this is
probably not truly correct. After this patch, at least, it is easy to
correct by someone who knows the language(s) in question.
MIPS/BFD: For n64 hold the number of internal relocs in `->reloc_count'
Revert parts of commit fee24f1c5bfe ("objdump improvements for mips
elf64"), <https://sourceware.org/ml/binutils/2003-03/msg00108.html>, and
make the `->reloc_count' member of `struct bfd_section' hold the actual
number of internal relocations stored in its `->relocation' vector. To
do so adjust `mips_elf64_slurp_one_reloc_table' to set `->reloc_count'
to the actual number of internal relocations retrieved and discard
`mips_elf64_canonicalize_reloc', `mips_elf64_canonicalize_dynamic_reloc'
and their corresponding target macros. Contrary to the description of
`mips_elf64_slurp_one_reloc_table', adjusted appropriately, this makes
generic relocation processing code happy and satisfies the "merge notes
section" binutils test case.
Add extra binutils test cases to expand the coverage of the generic
"merge notes section" test case, now passing with the n64 ABI, across
the MIPS o32, n32 and n64 ABIs regardless of the default ABI selected in
target configuration, and also to verify correctness of the relocations
produced. Conversely, do not provide any additional test cases for the
original issue addressed with the commit referred:
- objdump would display only 1/3 of the total number of relocations,
because it used the external relocation count, but each external
relocation is brought in as 3 internal relocations.
as n64 ABI relocation processing with `objdump -r' and `objdump -R' is
already widely covered across the GAS and LD test suites.
bfd/
* elf64-mips.c (mips_elf64_canonicalize_reloc): Remove prototype
and function.
(mips_elf64_canonicalize_dynamic_reloc): Likewise.
(mips_elf64_slurp_one_reloc_table): Set `reloc_count' to the
actual number of internal relocations retrieved. Adjust
function description.
(bfd_elf64_canonicalize_reloc): Remove macro.
(bfd_elf64_canonicalize_dynamic_reloc): Likewise.
binutils/
* testsuite/binutils-all/mips/mips-note-2.d: New test.
* testsuite/binutils-all/mips/mips-note-2r.d: New test.
* testsuite/binutils-all/mips/mips-note-2-n32.d: New test.
* testsuite/binutils-all/mips/mips-note-2-n64.d: New test.
* testsuite/binutils-all/mips/mips-note-2r-n32.d: New test.
* testsuite/binutils-all/mips/mips-note-2r-n64.d: New test.
* testsuite/binutils-all/mips/mips.exp: Define `has_newabi'.
Run the new tests.
binutils/testsuite: Permit the reuse of dump patterns
Complement commit 89210bdc8fd2 ("GAS: Permit the reuse of dump
patterns"), <https://sourceware.org/ml/binutils/2010-07/msg00269.html>,
and commit ef2b5578f693 ("MIPS: Enable NewABI tests for SDE targets"),
<https://sourceware.org/ml/binutils/2012-08/msg00017.html>, and like the
GAS and LD versions make the binutils version of `run_dump_test' also
support the `dump' keyword, for reusing dump patterns between tests.
binutils/
* testsuite/lib/utils-lib.exp (run_dump_test): Handle the `dump'
option.
eorg-Johann Lay [Fri, 19 May 2017 14:06:33 +0000 (15:06 +0100)]
Update avrxmega3 linker emulation to support avrxmega2 devices with flash memory visible in the SRAM address range.
PR ld/21472
ld * emulparams/avrxmega3.sh (RODATA_PM_OFFSET): Set to 0x8000.
* scripttempl/avr.sc
(__RODATA_PM_OFFSET__) [RODATA_PM_OFFSET]: Use RODATA_PM_OFFSET
as default if not already defined.
(.data) [!RODATA_PM_OFFSET]: Don't include .rodata and friends.
(.rodata) [RODATA_PM_OFFSET]: Put at an offset of
__RODATA_PM_OFFSET__.
Rainer Orth [Fri, 19 May 2017 13:08:45 +0000 (15:08 +0200)]
Fix tui compilation with Solaris libcurses: clear define (PR tui/21482)
On both mainline and the 8.0 branch, gdb compilation fails on Solaris 10
with the native libcurses like this:
In file included from /vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/gdb_curses.h:42:
0,
from /vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-data.h:2
6,
from /vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-disasm.c
:31:
/vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-disasm.c: In function `CORE_A
DDR tui_disassemble(gdbarch*, tui_asm_line*, CORE_ADDR, int)':
/vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-disasm.c:71:19: error: `class
string_file' has no member named `wclear'; did you mean `clear'?
gdb_dis_out.clear ();
^
/vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-disasm.c:78:19: error: `class
string_file' has no member named `wclear'; did you mean `clear'?
gdb_dis_out.clear ();
^
make[2]: *** [Makefile:1927: tui-disasm.o] Error 1
It turned out this happens because <curses.h> has
#define clear() wclear(stdscr)
This can be avoided by defining NOMACROS, which the patch below does.
ncurses potentially has a similar problem, which can be avoided by defining
NCURSES_NOMACROS.
Rainer Orth [Fri, 19 May 2017 12:16:55 +0000 (14:16 +0200)]
Fix tui compilation with Solaris libcurses: non-const last arg to mvwaddstr (PR tui/21482)
On both mainline and the 8.0 branch, gdb compilation fails on Solaris 10
with the native libcurses in gdb/tui for several instances of the same problem:
/vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-winsource.c: In function `void tui_erase_source_content(tui_win_info*, int)':
/vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-winsource.c:257:18: error: invalid conversion from `const char*' to `char*' [-fpermissive]
no_src_str);
^
In file included from /vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/gdb_curses.h:42:0,
from /vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-data.h:26,
from /vol/src/gnu/gdb/gdb-8.0-branch/local/gdb/tui/tui-winsource.c:33:
/vol/gcc-7/lib/gcc/sparc-sun-solaris2.10/7.1.0/include-fixed/curses.h:699:12: note: initializing argument 4 of `int mvwaddstr(WINDOW*, int, int, char*)'
extern int mvwaddstr(WINDOW *, int, int, char *);
^~~~~~~~~
make[2]: *** [Makefile:1927: tui-winsource.o] Error 1
Unlike ncurses, <curses.h> declares
extern int mvwaddstr(WINDOW *, int, int, char *);
i.e. the last arg is char *, not const char *.
The patch fixes this by casting the last arg to mvwaddstr to char *,
as was recently done on mainline in a newterm() call (the only
difference between 8.0 and mainline gdb/tui).
gdb: fix TYPE_CODE_ARRAY handling in sparc targets
gdb has a special type (TYPE_CODE_ARRAY) to support the gcc extension
(https://gcc.gnu.org/onlinedocs/gcc/Vector-Extensions.html).
TYPE_CODE_ARRAY is handled incorrectly for both (32- and 64-bit) modes
on Sparc machines.
Tested on sparc64-linux-gnu and sparc-solaris (32- and 64-bit mode).
6 tests ( from gdb/testsuite/gdb.base/gnu_vector.exp) failed on
sparc64-Linux and on sparc-Solaris in 32- and 64-bit mode. Now all
these tests passed. gdb/testsuite/gdb.base/gnu_vector.exp has 117
different cases for small (and not small) arrays and structures.
No regressions.
gdb/ChangeLog:
2017-05-19 Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
* sparc-tdep.c (sparc_structure_return_p)
(sparc_arg_on_registers_p): New functions.
(sparc32_store_arguments): Use them.
* sparc64-tdep.c (sparc64_16_byte_align_p)
(sparc64_store_floating_fields, sparc64_extract_floating_fields):
Handle TYPE_CODE_ARRAY.
Alan Modra [Fri, 19 May 2017 03:10:10 +0000 (12:40 +0930)]
garbage collect debug sections when no alloc sections are kept
The pr20882 testcase fails on a number of targets that add attribute
or note sections to object files, and the default linker script says
those sections should be kept. This patch changes --gc-sections to
drop debug and special sections like .comment when no SEC_ALLOC
section in an object file is kept. The assumption is that debug
sections are describing code and data that will become part of the
final image in memory.
* elflink.c (_bfd_elf_gc_mark_extra_sections): Don't keep
debug and special sections when no non-note alloc sections in an
object are kept.
MIPS/binutils/testsuite: Bail out right away if !ELF
We have but ELF binutils tests in the MIPS subset, and non-ELF MIPS/GAS
ports are gone, making a future addition of any non-ELF tests unlikely.
Bail out right away then if non-ELF, consuming one level of indentation
across the actual tests run.
binutils/
* testsuite/binutils-all/mips/mips.exp: Bail out right away if
non-ELF.
Nick Clifton [Thu, 18 May 2017 14:07:59 +0000 (15:07 +0100)]
Treat a prefix of "$SYSROOT" in the same way as "=" when parsing linker search paths.
PR ld/21251
* ldfile.c (ldfile_add_library_path): If the path starts with
$SYSROOT then use the sysroot as the real prefix.
* ldlang.c (lang_add_input_file): Treat $SYSROOT in the same
way as =.
* ldlex.l: Add $SYSROOT as allow prefix for a filename.
* ld.texinfo (-L): Document that $SYSROOT acts like = when
prefixing a library search path.
(INPUT): Likewise.
* testsuite/ld-scripts/sysroot-prefix.exp: Add $SYSROOT prefix
tests.
Pedro Alves [Thu, 18 May 2017 10:40:08 +0000 (11:40 +0100)]
gdb.base/fileio.c: Fix several -Wmaybe-uninitialized warnings
src/gdb/testsuite/gdb.base/fileio.c: In function ‘test_write’:
src/gdb/testsuite/gdb.base/fileio.c:158:5: warning: ‘ret’ may be used uninitialized in this function [-Wmaybe-uninitialized]
printf ("write 1: ret = %d, errno = %d\n", ret, errno);
^
gdb/ChangeLog:
2017-05-18 Pedro Alves <palves@redhat.com>
* gdb.base/fileio.c (test_write, test_read, test_close)
(test_fstat): Don't print 'ret' in the fail path.
Pedro Alves [Thu, 18 May 2017 10:37:55 +0000 (11:37 +0100)]
gdb.base/fileio.c: Fix several -Wreturn-type warnings
All the "test_" functions warn like:
src/gdb/testsuite/gdb.base/fileio.c: In function ‘test_close’:
src/gdb/testsuite/gdb.base/fileio.c:280:1: warning: control reaches end of non-void function [-Wreturn-type]
}
^
Nothing looks at the return of these functions, so just make them
return void. While at it, "()" is not the same as "(void)" in C - fix
that too.
gdb/ChangeLog:
2017-05-18 Pedro Alves <palves@redhat.com>
projects / thirdparty / binutils-gdb.git / log
