Alan Modra [Fri, 24 Jan 2014 12:47:28 +0000 (23:17 +1030)]
TLS orphan section placement
Ensures TLS orphans are placed adjacent to existing TLS sections,
and fixes places where the output_section_statement flags (which might
not be set) were tested when bfd_section flags were available.
* ldlang.c (lang_output_section_find_by_flags): Be careful to
test look->bfd_section->flags if available rather than
look->flags. Separate SEC_THREAD_LOCAL handling from
SEC_READONLY loop, and rewrite.
Alan Modra [Fri, 24 Jan 2014 03:52:10 +0000 (14:22 +1030)]
Fixes powerpc64le ld segfaults when --emit-relocs is used.
ELFv2 needs fewer relocs to annotate plt call stubs. I correctly
allocated a smaller buffer and wrote the proper relocs, but stupidly
bumped the reloc count as for ELFv1.
* elf64-ppc.c (ppc_build_one_stub): Correct reloc count passed
to get_relocs for ELFv2.
Nick Clifton [Thu, 23 Jan 2014 17:08:24 +0000 (17:08 +0000)]
Remove the display of known MCU names from the MSP430 port of GAS.
New MSP430 MCU parts are being created by TI all the time and the
list is basically always out of date. Instead any name will be
accepted by the -mmcu= command line option. ISA selection is now
based upon the -mcpu= command line option, just as is done for GCC.
gas/ChangeLog
* config/tc-msp430.c (show_mcu_list): Delete.
(md_parse_option): Accept any MCU name. Accept several more
variants for the -mcpu option.
(md_show_usage): Do not call show_mcu_list.
H.J. Lu [Thu, 23 Jan 2014 16:15:17 +0000 (08:15 -0800)]
Issue a linker error if TLS sections are not adjacent
Bad linker script may lead to TLS sections separated by non-TLS sections
in output. This patch changes linker assert to a linker error to
provide better linker diagnosis.
PR ld/16498
* elf.c (_bfd_elf_map_sections_to_segments): Issue a linker error
if TLS sections are not adjacent.
Tom Tromey [Wed, 22 Jan 2014 15:52:15 +0000 (08:52 -0700)]
avoid python exception in FrameDecorator.py
This fixes a bug in FrameDecorator.py.
FrameVars seems to assume that Frame.block can return None if there is
no block. However, it actually throws an exception.
I saw this bug while developing a frame filter, but unfortunately I
don't know how to reproduce it. It seems to me that the SAL tests in
_is_limited_frame should exclude the bad cases; and in my attempts to
write a test they do.
Nevertheless I think the fix is reasonably obvious and ought to go in.
Tom Tromey [Wed, 22 Jan 2014 15:44:41 +0000 (08:44 -0700)]
fix erroneous error-handling in frame filter code
This fixes PR python/16487.
The bug here is that the function-name-handling code in py_print_frame
had a small logic error (really a misplaced closing brace). This
error could lead to a Py_DECREF(NULL), which crashes.
This patch fixes the bug in the obvious way.
Built and regtested on x86-64 Fedora 18. New test case included.
2014-01-23 Tom Tromey <tromey@redhat.com>
PR python/16487:
* python/py-framefilter.c (py_print_frame): Don't call Py_DECREF
on a NULL pointer. Move "goto error" to correct place.
2014-01-23 Tom Tromey <tromey@redhat.com>
PR python/16487:
* gdb.python/py-framefilter.exp: Add test using "Error" filter.
* gdb.python/py-framefilter.py (ErrorInName, ErrorFilter): New
classes.
Tom Tromey [Wed, 22 Jan 2014 15:10:01 +0000 (08:10 -0700)]
fix crash in frame filters
apply_frame_filter calls ensure_python_env before computing the
gdbarch to use. This means that python_gdbarch can be NULL while in
Python code, and if a frame filter depends on this somehow (easy to
do), gdb will crash.
The fix is to compute the gdbarch first.
Built and regtested on x86-64 Fedora 18.
New test case included.
2014-01-23 Tom Tromey <tromey@redhat.com>
PR python/16491:
* python/py-framefilter.c (apply_frame_filter): Call
ensure_python_env after computing gdbarch.
2014-01-23 Tom Tromey <tromey@redhat.com>
PR python/16491:
* gdb.python/py-framefilter.py (Reverse_Function.function): Read a
string from an inferior frame.
* gdb.python/py-framefilter-mi.exp: Update.
The purpose of this patch is to provide a way for one object file
to require the inclusion of another object, without having to
allocate space for a .word address reference.
Andreas Arnez [Wed, 22 Jan 2014 17:54:43 +0000 (18:54 +0100)]
gdb/ChangeLog:
* syscalls/s390x-linux.xml: New file.
* syscalls/s390-linux.xml: New file.
* s390-linux-tdep.c (XML_SYSCALL_FILENAME_S390): New macro.
(XML_SYSCALL_FILENAME_S390X): Likewise.
(op_svc): New enum value for SVC opcode.
(s390_sigtramp_frame_sniffer): Replace literal by 'op_svc'.
(s390_linux_get_syscall_number): New function.
(s390_gdbarch_init): Register '*get_syscall_number' and the
syscall xml file name.
* data-directory/Makefile.in (SYSCALLS_FILES): Add
"s390-linux.xml" and "s390x-linux.xml".
* NEWS: Announce new feature.
gdb/testsuite/ChangeLog:
* gdb.base/catch-syscall.exp: Activate test on s390*-linux.
Andreas Arnez [Wed, 22 Jan 2014 15:37:41 +0000 (15:37 +0000)]
Fix regression on s390x with entry-values.exp.
The trace-specific test case 'entry-values' concludes fairly late in
the process that this platform doesn't support trace. Before that,
there are some platform specifics that don't work on s390x. The fix
addresses two aspects:
(1) Removal of an excess space character in the regex for the
disassembly. This is needed when there is a function alignment
gap, because then the hex address is immediately followed by a
colon, like in the first 'nopr' line below:
(gdb) disassemble foo+50,+10
Dump of assembler code from 0x32 to 0x3c:
0x0000000000000032 <foo+50>: br %r4
0x0000000000000034: nopr %r7
0x0000000000000036: nopr %r7
0x0000000000000038 <bar+0>: stmg %r11,%r15,88(%r15)
End of assembler dump.
(2) Handling for the s390-specific call instruction.
gdb/testsuite/ChangeLog:
* gdb.trace/entry-values.exp: Remove excess space character from
regex patterns. Handle s390 call instruction.
This patch reverts to the original method, re-introducing '_start'
symbols. In addition, it adds sufficient alignment before the label,
such that the label never points into an alignment gap.
gdb/testsuite/ChangeLog:
* gdb.dwarf2/dw2-dir-file-name.c (FUNC): Insert alignment and
define "*_start" label. Make "name" static.
* gdb.dwarf2/dw2-dir-file-name.exp: Replace references to
${name} by references to ${name}_start.
Andreas Arnez [Wed, 22 Jan 2014 15:31:10 +0000 (15:31 +0000)]
Prevent appending "-g" after "-g3" to compile options in info-macros.exp.
When upstream gcc is given a command line with the "-g" option after
"-g3", it doesn't generate a ".debug_macro" section. This is because
the last option wins, thus downgrading the debug level again. Without
any macro debug information in the executable, info-macros.exp
obviously produces many failures.
Since the "-g" option is appended by DejaGnu's target_compile whenever
the "debug" option is set, the fix just removes that option.
gdb/testsuite/ChangeLog:
* gdb.base/info-macros.exp: Remove "debug" from the compile
options.
Joel Brobecker [Tue, 21 Jan 2014 10:16:57 +0000 (14:16 +0400)]
Add ARI (ok) marker for __func__ reference in common-utils.h
The ARI script flagged the use of the __func__ variable, which
is normally not allowed (not defined in C90). However, this particular
use is OK, as the reference is only made when __STDC_VERSION__ >=
199901L. So, add an "ARI:" comment to explicitly OK this use.
gdb/ChangeLog:
* common/common-utils.h: Add "ARI:" comment beside __func__
reference.
Joel Brobecker [Tue, 21 Jan 2014 10:20:50 +0000 (14:20 +0400)]
Expand documentation of common-utils.h::FUNCTION_NAME
While looking at this macro, I noticed that it wasn't always necessarily
defined. That prompted me to search the current sources to make sure
that all uses were adequately protected, which they were. But to help
prevent future uses to be made unprotected, this patch expands the
current macro documentation a bit.
gdb/ChangeLog:
* common/common-utils.h (FUNCTION_NAME): Expand the macro's
documentation a bit.
Alan Modra [Wed, 22 Jan 2014 01:13:03 +0000 (11:43 +1030)]
Make assignments to dot keep an empty output section.
An assignment to dot in an output section that allocates space of
course keeps the output section. Here, I'm changing the behaviour for
assignments that don't allocate space. The idea is not so much to
allow people to force output of an empty section with ". = .", but
to fix cases where an otherwise empty section has padding added by an
alignment expression that changes with relaxation or .eh_frame
editing. Such a section might have zero size before relaxation and so
be stripped incorrectly.
ld/
* ld.texinfo (Output Section Discarding): Mention assigning to dot
as a way of keeping otherwise empty sections.
* ldexp.c (is_dot, is_value, is_sym_value, is_dot_ne_0,
is_dot_plus_0, is_align_conditional): New predicates.
(exp_fold_tree_1): Set SEC_KEEP when assigning to dot inside an
output section, except for some special cases.
* scripttempl/elfmicroblaze.sc: Use canonical form to align at
end of .heap and .stack.
ld/testsuite/
* ld-shared/elf-offset.ld: Align end of .bss with canonical form
of ALIGN that allows an empty .bss to be removed.
* ld-arm/arm-dyn.ld: Likewise.
* ld-arm/arm-lib.ld: Likewise.
* ld-elfvsb/elf-offset.ld: Likewise.
* ld-mips-elf/mips-dyn.ld: Likewise.
* ld-mips-elf/mips-lib.ld: Likewise.
* ld-arm/arm-no-rel-plt.ld: Remove duplicate ALIGN.
* ld-powerpc/vle-multiseg-1.ld: Remove ALIGN at start of section.
ALIGN address of section instead.
* ld-powerpc/vle-multiseg-2.ld: Likewise.
* ld-powerpc/vle-multiseg-3.ld: Likewise.
* ld-powerpc/vle-multiseg-4.ld: Likewise.
* ld-powerpc/vle-multiseg-6.ld: Likewise.
* ld-scripts/empty-aligned.d: Check section headers not program
headers. Remove xfail and notarget.
* ld-scripts/empty-aligned.t: Use canonical ALIGN for end of .text2.
H.J. Lu [Tue, 21 Jan 2014 23:42:43 +0000 (15:42 -0800)]
Check incompatible existing default symbol definition
After resolving a versioned reference, foo@VER1, to a default versioned
definition, foo@@VER1, from a shared object, we also merge it with
the existing regular default symbol definition, foo. When foo is IFUNC
and foo@@VER1 aren't, we will merge 2 incompatible definitions. This
patch avoids merging foo@@VER1 definition with foo definition if
one is IFUNC and the other isn't.
DJ Delorie [Tue, 21 Jan 2014 20:12:19 +0000 (15:12 -0500)]
Ensure that %func() expressions are outermost terms
This is to avoid expressions like: %hi(foo) + 1, which will
not do what you expect. The complex relocations can handle it,
but the internal fixups can't.
Roland McGrath [Tue, 21 Jan 2014 19:01:04 +0000 (11:01 -0800)]
gdb: Support install-strip target
gdb/
* configure.ac: Call AM_PROG_INSTALL_STRIP.
* configure: Regenerate.
* aclocal.m4: Regenerate.
* Makefile.in (install_sh, INSTALL_STRIP_PROGRAM, STRIP):
New substituted variables.
(install-strip): New target.
(INSTALL_SCRIPT): New substituted variable.
(FLAGS_TO_PASS): Add it.
(install-only): Use $(INSTALL_SCRIPT) rather than
$(INSTALL_PROGRAM) for gcore.
H.J. Lu [Tue, 21 Jan 2014 13:33:48 +0000 (05:33 -0800)]
Don't check shared/export_dynamic/ref_dynamic for type mismatch
There is nothing linker can do when a type mismatched default definition
are made dynamic by info->shared, info->export_dynamic or h->ref_dynamic.
But we do want to avoid exporting it when building PIE. Let's remove
those checks.
bfd/
PR ld/2404
* elflink.c (_bfd_elf_merge_symbol): Don't check info->shared,
info->export_dynamic, nor !h->ref_dynamic for type mismatch when
adding the default version.
ld/testsuite/
PR ld/2404
* ld-elf/shared.exp: Add a PIE test for PR ld/2404.
Tom Tromey [Fri, 27 Dec 2013 05:06:27 +0000 (22:06 -0700)]
better packing for command struct
This moves all the bitfields in struct cmd_list_element to be closer
together. This packs the structure somewhat better. On a 64 bit
machine, this simple rearrangement saves around 50k at startup.
2014-01-20 Tom Tromey <tromey@redhat.com>
* cli/cli-decode.h (struct cmd_list_element): Move all bitfields
together.
Tom Tromey [Fri, 27 Dec 2013 05:04:40 +0000 (22:04 -0700)]
convert flags to bitfields
This changes various flags struct cmd_list_element into bitfields. In
general I think bitfields are cleaner than flag words, at least in a
case like this where there is no need to pass the flags around
independently of the enclosing struct.
Baruch Siach [Mon, 20 Jan 2014 07:53:12 +0000 (09:53 +0200)]
gdb: xtensa: fix linux ptrace includes
Currently, xtensa code using the Linux ptrace interface only include
sys/ptrace.h. This file comes from the C library (glibc and uClibc,
at least), and includes a declaration of the ptrace() functions, along
with some cross architecture constants that are mostly copied from the
file located at include/uapi/linux/ptrace.h in recent Linux kernels.
For xtensa specific constants like PTRACE_GETXTREGS and
PTRACE_SETXTREGS the asm/ptrace.h include from the Linux kernel UAPI
is needed. The code in gdbserver xtensa specific part doesn't call
ptrace() directly, so we can remove the unneeded sys/ptrace.h include.
The gdb xtensa specific code needs both headers, since it calls
ptrace().
gdb/
* xtensa-linux-nat.c: Include asm/ptrace.h.
gdb/gdbserver/
* linux-xtensa-low.c: Include asm/ptrace.h instead of
sys/ptrace.h.
Alan Modra [Mon, 20 Jan 2014 10:58:42 +0000 (21:28 +1030)]
Allow self-assignment for absolute symbols defined in a linker script
Modifies ld machinery tracking linker script assignments to notice all
assignments, not just those symbols mentioned in DEFINED().
ld/
PR ld/14962
* ldlang.h (struct lang_definedness_hash_entry): Add by_object and
by_script. Make iteration a single bit field.
(lang_track_definedness, lang_symbol_definition_iteration): Delete.
(lang_symbol_defined): Declare.
* ldlang.c (lang_statement_iteration): Expand comment a little.
(lang_init <lang_definedness_table>): Make it bigger.
(lang_track_definedness, lang_symbol_definition): Delete.
(lang_definedness_newfunc): Update.
(lang_symbol_defined): New function.
(lang_update_definedness): Create entries here. Do track whether
script definition of symbol is valid, even when also defined in
an object file.
* ldexp.c (fold_name <DEFINED>): Update.
(fold_name <NAME>): Allow self-assignment for absolute symbols
defined in a linker script.
ld/testsuite/
* ld-scripts/pr14962-2.d,
* ld-scripts/pr14962-2.t: New test.
* ld-scripts/expr.exp: Run it.
Guy Martin [Mon, 20 Jan 2014 03:46:16 +0000 (14:16 +1030)]
Fix duplicate output section statement lookup
Tie output section statements to their associated output section via
output section userdata. This allows us to avoid hash lookups which
are slower and fail when multiple output sections have the same name.
* ldlang.h (lang_output_section_get): Define.
* ldlang.c (lang_output_section_get): Likewise.
(init_os): Set the output_section userdata to the output
section statement.
* emultempl/hppaelf.em: Use lang_output_section_get instead of
lang_output_section_find where applicable.
* emultempl/aarch64elf.em: Likewise.
* emultempl/aix.em: Likewise.
* emultempl/armelf.em: Likewise.
* emultempl/m68hc1xelf.em: Likewise.
* emultempl/metagelf.em: Likewise.
* emultempl/mipself.em: Likewise.
* emultempl/ppc64elf.em: Likewise.
* emultempl/spuelf.em: Likewise.
* d-lang.h (struct builtin_d_type): New data type.
(builtin_d_type): Add declaration.
* d-lang.c (d_language_arch_info, build_d_types)
(builtin_d_type): New functions.
(enum d_primitive_types): New data type.
(d_language_defn): Change c_language_arch_info to
d_language_arch_info.
(d_type_data): New static variable.
(_initialize_d_language): Initialize d_type_data.
Add "volatile" keyword to "struct gdb_exception" declaration
While doing something else, I found that those 2 places were incorrectly
declaring a "struct gdb_exception" without using the "volatile" keyword.
This commit fixes that.
Doug Evans [Fri, 17 Jan 2014 18:23:29 +0000 (10:23 -0800)]
Add delim_string_to_char_ptr_vec.
* common/gdb_vecs.c (delim_string_to_char_ptr_vec_append): New
function, contents of dirnames_to_char_ptr_vec_append moved here.
(delim_string_to_char_ptr_vec): New function.
(dirnames_to_char_ptr_vec_append): Rewrite.
* common/gdb_vecs.h (delim_string_to_char_ptr_vec): Declare.
Will Newton [Thu, 16 Jan 2014 10:08:45 +0000 (10:08 +0000)]
gas: ARM: Fix encoding of VCVTr.s32.f64 instructions
The direct rounding floating-point VCVT instructions introduced in
ARMv8 encode the s32.f64 variant incorrectly. The op bit should be
set to 1 for all signed conversions.
gas/ChangeLog:
2014-01-17 Will Newton <will.newton@linaro.org>
* config/tc-arm.c (do_vfp_nsyn_cvt_fpv8): Set OP to 1
for the s32.f64 flavours of VCVT.
gas/testsuite/ChangeLog:
2014-01-17 Will Newton <will.newton@linaro.org>
* gas/arm/armv8-a+fp.d: Correct encoding of vcvta.s32.f64.
Pedro Alves [Fri, 17 Jan 2014 13:33:30 +0000 (13:33 +0000)]
Fix PR PR16445 - gdbserver build failure on x86.
If gdb_proc_service.h ends up including linux/elf.h, we'll trip on
duplicate definitions:
In file included from ../../../gdb/gdbserver/linux-x86-low.c:29:0:
../../../gdb/gdbserver/../../include/elf/common.h:36:0: error: "ELFMAG0"
redefined [-Werror]
... etc ...
Handle this the same way linux-low.c and linux-arm-low.c handle this.
gdb/gdbserver/
2014-01-17 Pedro Alves <palves@redhat.com>
PR PR16445
* linux-x86-low.c (linux-x86-low.c): Don't include elf/common.h if
ELFMAG0 is defined after including gdb_proc_service.h.
Alan Modra [Fri, 17 Jan 2014 00:09:50 +0000 (10:39 +1030)]
Allow ld target customiser script to force on disk scripts.
Mucking with EMULATION_LIBPATH has unwanted side effects. Allow nds32
emulparams scripts to disable compiled-in scripts naturally, by
setting COMPILE_IN=no.
* genscripts.sh (COMPILE_IN): Don't set if already set.
* emulparams/nds32elf.sh: Don't clear EMULATION_LIBPATH, set
COMPILE_IN=no.
* emulparams/nds32elf16m.sh: Likewise.
* emulparams/nds32elf_linux.sh: Likewise.
* emultempl/aix.em: Test COMPILE_IN value is "yes".
* emultempl/armcoff.em: Likewise.
* emultempl/elf32.em: Likewise.
* emultempl/generic.em: Likewise.
* emultempl/gld960.em: Likewise.
* emultempl/gld960c.em: Likewise.
* emultempl/linux.em: Likewise.
* emultempl/lnk960.em: Likewise.
* emultempl/m68kcoff.em: Likewise.
* emultempl/sunos.em: Likewise.
* emultempl/ticoff.em: Likewise.
Tom Tromey [Sat, 28 Dec 2013 05:22:07 +0000 (22:22 -0700)]
rearrange struct value to save memory
This patch rearranges struct value a tiny bit, moving the "regnum"
field into a hole. This saves 8 bytes per value on a 64-bit machine,
and 4 bytes per value on a 32 bit machine. I think it does not
negatively affect readability or performance.
Tom Tromey [Wed, 18 Dec 2013 15:27:54 +0000 (08:27 -0700)]
remove extended_remote_create_inferior_1
I noticed that extended_remote_create_inferior_1 is called from a
single spot. This patch unifies the callee and caller. It's just a
simple cleanup that made the coming refactoring simpler.
2014-01-16 Tom Tromey <tromey@redhat.com>
* remote.c (extended_remote_create_inferior): Rename from
extended_remote_create_inferior_1. Add "ops" argument. Remove
old implementation.
tfile knows to infer the PC from the tracepoint's address if the PC
wasn't collected (tfile_fetch_registers) but, that only works on
targets whose PC register is a raw register, and on s390, the PC
register is a pseudo register.
But even if GDB doesn't know how to infer the value of PC, saying the
current frame is level -1 is a bug:
'-1' is the level of the sentinel frame, which should never be visible.
This is caused by the s390's heuristic unwinder accepting the frame
(the fallback heuristic unwinders _always_ accept the frame), but then
the unwind->this_id method throws that "PC not available\n" error.
IOW, the s390's heuristic unwinder was never adjusted to handle
unavailable register values gracefully, which can happen with e.g., a
trimmed core file too.
This is just the minimal necessary for
<unavailable> frames, which at least gets us:
(gdb) tfind
Found trace frame 0, tracepoint 1
#0 <unavailable> in ?? ()
That is, frame #0 instead of -1.
We could get better info out of "info frame" (this patch makes us show
"outermost"), but this change would still be necessary.
gdb/
2014-01-16 Pedro Alves <palves@redhat.com>
* s390-linux-tdep.c (s390_frame_unwind_cache): Swallow
NOT_AVAILABLE_ERROR errors while parsing the prologue or reading
the backchain.
Markus Metzger [Tue, 10 Sep 2013 10:27:14 +0000 (12:27 +0200)]
record-btrace: show trace from enable location
The btrace record target shows the branch trace from the location of the first
branch destination. This is the first BTS records.
After adding incremental updates, we can now add a dummy record for the current
PC when we enable tracing so we show the trace from the location where branch
tracing has been enabled.
2014-01-16 Markus Metzger <markus.t.metzger@intel.com>
Read branch trace data incrementally and extend the current trace rather than
discarding it and reading the entire trace buffer each time.
If the branch trace buffer overflowed, we can't extend the current trace so we
discard it and start anew by reading the entire branch trace buffer.
2014-01-16 Markus Metzger <markus.t.metzger@intel.com>
* common/linux-btrace.c (perf_event_read_bts, linux_read_btrace):
Support delta reads.
(linux_disable_btrace): Change return type.
* common/linux-btrace.h (linux_read_btrace): Change parameters
and return type to allow error reporting. Update users.
(linux_disable_btrace): Change return type. Update users.
* common/btrace-common.h (btrace_read_type) <BTRACE_READ_DELTA>:
New.
(btrace_error): New.
(btrace_block) <begin>: Comment on BEGIN == 0.
* btrace.c (btrace_compute_ftrace): Start from the end of
the current trace.
(btrace_stitch_trace, btrace_clear_history): New.
(btrace_fetch): Read delta trace, return if replaying.
(btrace_clear): Move clear history code to btrace_clear_history.
(parse_xml_btrace): Throw an error if parsing failed.
* target.h (struct target_ops) <to_read_btrace>: Change parameters
and return type to allow error reporting.
(target_read_btrace): Change parameters and return type to allow
error reporting.
* target.c (target_read_btrace): Update.
* remote.c (remote_read_btrace): Support delta reads. Pass
errors on.
* NEWS: Announce it.
gdbserver/
* target.h (target_ops) <read_btrace>: Change parameters and
return type to allow error reporting.
* server.c (handle_qxfer_btrace): Support delta reads. Pass
trace reading errors on.
* linux-low.c (linux_low_read_btrace): Pass trace reading
errors on.
(linux_low_disable_btrace): New.
Markus Metzger [Fri, 3 May 2013 08:51:13 +0000 (10:51 +0200)]
record-btrace: provide target_find_new_threads method
The "info threads" command tries to read memory, which is not possible during
replay. This results in an error message and aborts the command without showing
the existing threads.
Provide a to_find_new_threads target method to skip the search while replaying.
2014-01-16 Markus Metzger <markus.t.metzger@intel.com>
* record-btrace.c (record_btrace_find_new_threads)
(record_btrace_thread_alive): New.
(init_record_btrace_ops): Initialize to_find_new_threads and
to_thread_alive.
Markus Metzger [Wed, 24 Apr 2013 11:24:11 +0000 (13:24 +0200)]
record-btrace: provide xfer_partial target method
Provide the xfer_partial target method for the btrace record target.
Only allow memory read accesses to readonly memory while we're replaying,
except for inserting and removing breakpoints.
2014-01-16 Markus Metzger <markus.t.metzger@intel.com>
* record-btrace.c (record_btrace_xfer_partial)
(record_btrace_insert_breakpoint, record_btrace_remove_breakpoint)
(record_btrace_allow_memory_access): New.
(init_record_btrace_ops): Initialize new methods.
* target.c (raw_memory_xfer_partial): Bail out if target reports
that this memory is not available.
Markus Metzger [Mon, 18 Mar 2013 15:43:05 +0000 (16:43 +0100)]
frame, backtrace: allow targets to supply a frame unwinder
Allow targets to supply their own target-specific frame unwinders; one for
normal frames and one for tailcall frames. If a target-specific unwinder
is supplied, it will be chosen before any other unwinder.
The original patch has been split into this and the next two patches.
gdb/
2013-02-11 Jan Kratochvil <jan.kratochvil@redhat.com>
* frame-unwind.c: Include target.h.
(frame_unwind_try_unwinder): New function with code from ...
(frame_unwind_find_by_frame): ... here. New variable
unwinder_from_target, call also target_get_unwinder)
(target_get_tailcall_unwinder, and frame_unwind_try_unwinder for it.
* target.c (target_get_unwinder, target_get_tailcall_unwinder): New.
* target.h (struct target_ops): New fields to_get_unwinder and
to_get_tailcall_unwinder.
(target_get_unwinder, target_get_tailcall_unwinder): New declarations.
Markus Metzger [Wed, 17 Apr 2013 07:39:43 +0000 (09:39 +0200)]
record-btrace: make ranges include begin and end
The "record function-call-history" and "record instruction-history" commands
accept a range "begin, end". End is not included in both cases. Include it.
2014-01-16 Markus Metzger <markus.t.metzger@intel.com>
* record-btrace.c (record_btrace_insn_history_range): Include
end.
(record_btrace_insn_history_from): Adjust range.
(record_btrace_call_history_range): Include
end.
(record_btrace_call_history_from): Adjust range.
* NEWS: Announce changes.
Markus Metzger [Thu, 18 Apr 2013 08:58:05 +0000 (10:58 +0200)]
record-btrace: optionally indent function call history
Add a new modifier /c to the "record function-call-history" command to
indent the function name based on its depth in the call stack.
Also reorder the optional fields to have the indentation at the very beginning.
Prefix the insn range (/i modifier) with "inst ".
Prefix the source line (/l modifier) with "at ".
Change the range syntax from "begin-end" to "begin,end" to allow copy&paste to
the "record instruction-history" and "list" commands.
Adjust the respective tests and add new tests for the /c modifier.
2014-01-16 Markus Metzger <markus.t.metzger@intel.com>
testsuite/
* gdb.btrace/function_call_history.exp: Fix expected field
order for "record function-call-history".
Add new tests for "record function-call-history /c".
* gdb.btrace/exception.cc: New.
* gdb.btrace/exception.exp: New.
* gdb.btrace/tailcall.exp: New.
* gdb.btrace/x86-tailcall.S: New.
* gdb.btrace/x86-tailcall.c: New.
* gdb.btrace/unknown_functions.c: New.
* gdb.btrace/unknown_functions.exp: New.
* gdb.btrace/Makefile.in (EXECUTABLES): Add new.
doc/
* gdb.texinfo (Process Record and Replay): Document new /c
modifier accepted by "record function-call-history".
Add /i modifier to "record function-call-history" example.
Markus Metzger [Mon, 13 May 2013 12:57:42 +0000 (14:57 +0200)]
record-btrace: start counting at one
The record instruction-history and record-function-call-history commands start
counting instructions at zero. This is somewhat unintuitive when we start
navigating in the recorded instruction history. Start at one, instead.
2014-01-16 Markus Metzger <markus.t.metzger@intel.com>
* btrace.c (ftrace_new_function): Start counting at one.
* record-btrace.c (record_btrace_info): Adjust number of calls
and insns.
* NEWS: Announce it.
Markus Metzger [Fri, 22 Mar 2013 13:32:47 +0000 (14:32 +0100)]
btrace: change branch trace data structure
The branch trace is represented as 3 vectors:
- a block vector
- a instruction vector
- a function vector
Each vector (except for the first) is computed from the one above.
Change this into a graph where a node represents a sequence of instructions
belonging to the same function and where we have three types of edges to connect
the function segments:
- control flow
- same function (instance)
- call stack
This allows us to navigate in the branch trace. We will need this for "record
goto" and reverse execution.
This patch introduces the data structure and computes the control flow edges.
It also introduces iterator structs to simplify iterating over the branch trace
in control-flow order.
It also fixes PR gdb/15240 since now recursive calls are handled correctly.
Fix the test that got the number of expected fib instances and also the
function numbers wrong.
The current instruction had been part of the branch trace. This will look odd
once we start support for reverse execution. Remove it. We still keep it in
the trace itself to allow extending the branch trace more easily in the future.
2014-01-16 Markus Metzger <markus.t.metzger@intel.com>
* btrace.h (struct btrace_func_link): New.
(enum btrace_function_flag): New.
(struct btrace_inst): Rename to ...
(struct btrace_insn): ...this. Update all users.
(struct btrace_func) <ibegin, iend>: Remove.
(struct btrace_func_link): New.
(struct btrace_func): Rename to ...
(struct btrace_function): ...this. Update all users.
(struct btrace_function) <segment, flow, up, insn, insn_offset)
(number, level, flags>: New.
(struct btrace_insn_iterator): Rename to ...
(struct btrace_insn_history): ...this.
Update all users.
(struct btrace_insn_iterator, btrace_call_iterator): New.
(struct btrace_target_info) <btrace, itrace, ftrace>: Remove.
(struct btrace_target_info) <begin, end, level>
<insn_history, call_history>: New.
(btrace_insn_get, btrace_insn_number, btrace_insn_begin)
(btrace_insn_end, btrace_insn_prev, btrace_insn_next)
(btrace_insn_cmp, btrace_find_insn_by_number, btrace_call_get)
(btrace_call_number, btrace_call_begin, btrace_call_end)
(btrace_call_prev, btrace_call_next, btrace_call_cmp)
(btrace_find_function_by_number, btrace_set_insn_history)
(btrace_set_call_history): New.
* btrace.c (btrace_init_insn_iterator)
(btrace_init_func_iterator, compute_itrace): Remove.
(ftrace_print_function_name, ftrace_print_filename)
(ftrace_skip_file): Change
parameter to const.
(ftrace_init_func): Remove.
(ftrace_debug): Use new btrace_function fields.
(ftrace_function_switched): Also consider gaining and
losing symbol information).
(ftrace_print_insn_addr, ftrace_new_call, ftrace_new_return)
(ftrace_new_switch, ftrace_find_caller, ftrace_new_function)
(ftrace_update_caller, ftrace_fixup_caller, ftrace_new_tailcall):
New.
(ftrace_new_function): Move. Remove debug print.
(ftrace_update_lines, ftrace_update_insns): New.
(ftrace_update_function): Check for call, ret, and jump.
(compute_ftrace): Renamed to ...
(btrace_compute_ftrace): ...this. Rewritten to compute call
stack.
(btrace_fetch, btrace_clear): Updated.
(btrace_insn_get, btrace_insn_number, btrace_insn_begin)
(btrace_insn_end, btrace_insn_prev, btrace_insn_next)
(btrace_insn_cmp, btrace_find_insn_by_number, btrace_call_get)
(btrace_call_number, btrace_call_begin, btrace_call_end)
(btrace_call_prev, btrace_call_next, btrace_call_cmp)
(btrace_find_function_by_number, btrace_set_insn_history)
(btrace_set_call_history): New.
* record-btrace.c (require_btrace): Use new btrace thread
info fields.
(record_btrace_info, btrace_insn_history)
(record_btrace_insn_history, record_btrace_insn_history_range):
Use new btrace thread info fields and new iterator.
(btrace_func_history_src_line): Rename to ...
(btrace_call_history_src_line): ...this. Use new btrace
thread info fields.
(btrace_func_history): Rename to ...
(btrace_call_history): ...this. Use new btrace thread info
fields and new iterator.
(record_btrace_call_history, record_btrace_call_history_range):
Use new btrace thread info fields and new iterator.
The test doesn't execute the compiled object's code, so GDB will try
to read memory from the binary's sections. Instructions on ARM are
4-byte wide, and thus ARM's prologue scanner reads in 4-byte chunks.
As the section 'func' is put at is only 1 byte long, and no other
section is allocated contiguously:
...
Sections:
Idx Name Size VMA LMA File off Algn
0 .text 00000001000000000000000000000034 2**0
CONTENTS, ALLOC, LOAD, READONLY, CODE
...
H.J. Lu [Wed, 15 Jan 2014 20:53:04 +0000 (12:53 -0800)]
Skip directories with LIBPATH_SUFFIX_SKIP suffix
On Linux/x86-64, when binutils is configured with --libdir=/usr/lib64,
genscripts.sh treats /usr/lib64 as the default search directory. It
puts /usr/lib64 in linker scripts for all emulations, like
/usr/lib6432 is odd and /usr/lib64 is wrong. This patch changes
genscripts.sh to check LIBPATH_SUFFIX_SKIP if it is defined. It
skips directories with LIBPATH_SUFFIX_SKIP suffix.
PR ld/16456
* genscripts.sh: Don't search directory with LIBPATH_SUFFIX_SKIP
suffix.
* emulparams/elf32_x86_64.sh (LIBPATH_SUFFIX_SKIP): Set to 64
for elf32_x86_64 emulation.
* emulparams/elf_i386.sh (LIBPATH_SUFFIX_SKIP): Set to 64
for elf_i386 emulation.
Alan Modra [Thu, 16 Jan 2014 01:17:25 +0000 (11:47 +1030)]
Tidy ld use of bfd_section userdata.
A long time ago ld made use of userdata to tie an output section to
its lang_input_statement_struct object file. Some time later Joern
made map file printing of symbols at lot faster, using userdata on
input sections. That complicated allocation of userdata, and when the
output section use disappeared a year later, the code wasn't properly
cleaned up. This patch does that cleanup, and also tidies the symbol
printing code to not allocate userdata where it won't be needed. We
don't print symbols defined in the absolute section or in output
sections.
projects / deliverable / binutils-gdb.git / log
