Tom Tromey [Fri, 23 Apr 2021 01:51:54 +0000 (19:51 -0600)]
Remove and modernize dependencies in sim
Some spots in the sim build used manual dependencies, and some spots
did a compilation by hand but did not use the automatic dependency
tracking code. This patch fixes these spots.
I didn't touch ppc, because it doesn't use the common Makefile code.
I also didn't touch objects that are for the build machine, because
automatic dependencies don't work for those.
sim/arm/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* Makefile.in (armemu26.o, armemu32.o): Use COMPILE and
POSTCOMPILE.
sim/bpf/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 23 Apr 2021 01:51:54 +0000 (19:51 -0600)]
Require GNU make
GDB has required GNU make for quite some time, and this patch applies
this approach to the sim as well. Requiring GNU make means that
automatic dependency tracking can be simple and reliable, and it also
makes other refactorings simpler.
sim/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* m4/sim_ac_common.m4 (SIM_AC_COMMON): Don't check for make.
sim/aarch64/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/arm/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/avr/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/bfin/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/common/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* Make-common.in: Require GNU make.
sim/cr16/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/cris/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/d10v/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/erc32/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/example-synacor/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/frv/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/ft32/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/h8300/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/iq2000/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/lm32/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/m32c/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/m32r/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/m68hc11/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/mcore/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/microblaze/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/mips/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/mn10300/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/moxie/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/msp430/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/or1k/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/pru/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/riscv/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/rl78/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/rx/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/sh/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
sim/v850/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 23 Apr 2021 01:51:54 +0000 (19:51 -0600)]
Introduce stamp file for hw-config.h
This adds a stamp file for hw-config.h, to avoid unnecessary rebuilds.
It also arranges to remove hw-config.h in "mostlyclean", because the
file is created by "make".
sim/common/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* Make-common.in (stamp-hw): New target.
(hw-config.h): Depend on stamp-hw.
(mostlyclean): Remove stamp-hw and hw-config.h.
Tom Tromey [Fri, 23 Apr 2021 01:51:54 +0000 (19:51 -0600)]
Make mostlyclean an alias for clean
I found out by accident that "mostlyclean" in a sim subdir removes all
the configure artifacts. The usual rule is:
* If the maintainer built it, maintainer-clean should remove it;
* If configure built it, distclean should remove it;
* If make built it, "clean" should remove it;
* If there is a handy subset of "clean" that is "easy" to rebuild,
"mostlyclean" should remove it; otherwise mostlyclean should be an
alias for clean
This patch makes mostlyclean an alias for clean.
sim/common/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* Make-common.in (mostlyclean): Now an alias for clean, not
distclean.
Tom Tromey [Thu, 22 Apr 2021 22:50:32 +0000 (16:50 -0600)]
Fix sim build failure
On x86-64 Fedora 32, the sim was failing to build.
sim_events_schedule was passing a 'dummy' argument to
sim_events_schedule_vtracef, which caused an error because the format
parameter was NULL. However, removing this dummy argument caused an
error because too few arguments were being passed -- catch 22.
This patch fixes the build problem by using sim_events_schedule_tracef
instead.
sim/common/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* sim-events.c (sim_events_schedule): Use
sim_events_schedule_tracef.
Tom Tromey [Fri, 23 Apr 2021 01:39:56 +0000 (19:39 -0600)]
Fix ptype/o bug with "<no data fields>"
I noticed that when using ptype/o, the "<no data fields>" text that
may be emitted is indented incorrectly. This patch fixes the bug and
adds a new test case.
I also removed a stray backslash from ptype-offsets.exp that I noticed
while writing the test. This seemed too trivial to warrant a separate
patch.
gdb/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* c-typeprint.c (c_type_print_base_struct_union): Use
print_spaces_filtered_with_print_options.
gdb/testsuite/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* gdb.base/ptype-offsets.cc (struct empty_member): New.
(main): Use empty_member.
* gdb.base/ptype-offsets.exp: Add new test.
Simon Marchi [Thu, 22 Apr 2021 19:01:28 +0000 (15:01 -0400)]
gdb: fix getting range of flexible array member in Python
As reported in bug 27757, we get an internal error when doing:
$ cat test.c
struct foo {
int len;
int items[];
};
struct foo *p;
int main() {
return 0;
}
$ gcc test.c -g -O0 -o test
$ ./gdb -q -nx --data-directory=data-directory ./test -ex 'python gdb.parse_and_eval("p").type.target()["items"].type.range()'
Reading symbols from ./test...
/home/simark/src/binutils-gdb/gdb/gdbtypes.h:435: internal-error: LONGEST dynamic_prop::const_val() const: Assertion `m_kind == PROP_CONST' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)
This is because the Python code (typy_range) blindly reads the high
bound of the type of `items` as a constant value. Since it is a
flexible array member, it has no high bound, the property is undefined.
Since commit 8c2e4e0689 ("gdb: add accessors to struct dynamic_prop"),
the getters check that you are not getting a property value of the wrong
kind, so this causes a failed assertion.
Fix it by checking if the property is indeed a constant value before
accessing it as such. Otherwise, use 0. This restores the previous GDB
behavior: because the structure was zero-initialized, this is what was
returned before. But now this behavior is explicit and not accidental.
Add a test, gdb.python/flexible-array-member.exp, that is derived from
gdb.base/flexible-array-member.exp. It tests the same things, but
through the Python API. It also specifically tests getting the range
from the various kinds of flexible array member types (AFAIK it wasn't
possible to do the equivalent through the CLI).
gdb/ChangeLog:
PR gdb/27757
* python/py-type.c (typy_range): Check that bounds are constant
before accessing them as such.
* guile/scm-type.c (gdbscm_type_range): Likewise.
gdb/testsuite/ChangeLog:
PR gdb/27757
* gdb.python/flexible-array-member.c: New test.
* gdb.python/flexible-array-member.exp: New test.
* gdb.guile/scm-type.exp (test_range): Add test for flexible
array member.
* gdb.guile/scm-type.c (struct flex_member): New.
(main): Use it.
Simon Marchi [Thu, 22 Apr 2021 16:07:14 +0000 (12:07 -0400)]
sim: sprinkle some ATTRIBUTE_PRINTF
Add some ATTRIBUTE_PRINTF attributes to functions that take a format
string, to fix a few -Wformat-nonliteral warnings. Use the
ATTRIBUTE_PRINTF macro like we use in GDB, instead of spelling out
__attribute__((format...)). Use ATTRIBUTE_NULL_PRINTF at one place,
because callers expect to be able to pass NULL.
sim/common/ChangeLog:
* callback.c (os_printf_filtered, os_vprintf_filtered,
os_evprintf_filtered, os_error): Use ATTRIBUTE_PRINTF.
* sim-engine.h (sim_engine_abort, sim_engine_vabort): Likewise.
* sim-events.h (sim_events_schedule_tracef,
sim_events_schedule_vtracef): Use ATTRIBUTE_NULL_PRINTF.
gdb/continuations: turn continuation functions into inferior methods
Turn continuations-related functions into methods of the inferior
class. This is a refactoring.
gdb/ChangeLog:
2021-04-22 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* Makefile.in (COMMON_SFILES): Remove continuations.c.
* inferior.c (inferior::add_continuation): New method, adapted
from 'add_inferior_continuation'.
(inferior::do_all_continuations): New method, adapted from
'do_all_inferior_continuations'.
(inferior::~inferior): Clear the list of continuations directly.
* inferior.h (class inferior) <continuations>: Rename into...
<m_continuations>: ...this and make private.
* continuations.c: Remove.
* continuations.h: Remove.
* event-top.c: Don't include "continuations.h".
gdb/continuations: use lambdas instead of function pointers
Use lambdas and std::list to track inferior continuations. This is a
refactoring.
gdb/ChangeLog:
2021-04-22 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* inferior.h (class inferior) <continuations>: Change the type
to be an std::list of std::function's.
Update the references and uses below.
* continuations.c (struct continuation): Delete.
(make_continuation): Delete.
(do_my_continuations_1): Delete.
(do_my_continuations): Delete.
(discard_my_continuations_1): Delete.
(discard_my_continuations): Delete.
(add_inferior_continuation): Update.
(do_all_inferior_continuations): Update.
(discard_all_inferior_continuations): Update.
* continuations.h (add_inferior_continuation): Update to take
an std::function as the parameter.
* infcmd.c (struct attach_command_continuation_args): Delete.
(attach_command_continuation): Delete.
(attach_command_continuation_free_args): Delete.
(attach_command): Update.
(notice_new_inferior): Update.
Inferior continuations are no longer used by the until and finish
command. It is used only by the attach command and the remote target
upon detecting new inferiors. Update the comment accordingly.
Also update another comment about non-existent thread continuations and
remove an unused #include.
gdb/ChangeLog:
2021-04-22 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* continuations.h: Update the general comment.
* inferior.h (class inferior) <continuations>: Update the comment.
* interps.c: Do not include "continuations.h".
gdb/continuations: remove the 'err' from 'do_all_inferior_continuations'
The 'err' parameter of 'do_all_inferior_continuations' is effectively
unused. There is only one place where the function is called, and
there the argument is a literal 0. So, remove the parameter.
gdb/ChangeLog:
2021-04-22 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
gdb/infcmd: remove the unused parameter 'args' in 'attach_post_wait'
The 'arg' parameter of 'attach_post_wait' is unused. Remove it.
gdb/ChangeLog:
2021-04-22 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* infcmd.c (attach_post_wait): Remove the unused parameter 'args'.
Update the references below.
(struct attach_command_continuation_args)
(attach_command_continuation)
(attach_command_continuation_free_args)
(attach_command)
(notice_new_inferior): Update to remove the reference to 'args'.
Tom de Vries [Thu, 22 Apr 2021 15:01:00 +0000 (17:01 +0200)]
[gdb] Fix assert in remote_async_get_pending_events_handler
Occassionally I run into the following assert:
...
(gdb) PASS: gdb.multi/multi-target-continue.exp: inferior 5
Remote debugging from host ::1, port 49990^M
Process multi-target-continue created; pid = 31241^M
src/gdb/remote-notif.c:113: internal-error: \
void remote_async_get_pending_events_handler(gdb_client_data): \
Assertion `target_is_non_stop_p ()' failed.^M
...
The assert checks target_is_non_stop_p, which is related to the current
target.
Fix this by changing the assert such that it checks non-stopness related to
the event it's handling.
Tested on x86_64-linux.
gdb/ChangeLog:
2021-04-22 Simon Marchi <simon.marchi@polymtl.ca>
Tom de Vries <tdevries@suse.de>
PR remote/27710
* remote.c (remote_target_is_non_stop_p): New function.
* remote.h (remote_target_is_non_stop_p): Declare.
* remote-notif.c (remote_async_get_pending_events_handler): Fix assert
to check non-stopness using notif_state->remote rather current target.
Tom Tromey [Thu, 22 Apr 2021 13:16:36 +0000 (07:16 -0600)]
Improve code coverage of Rust testing
I enabled code coverage and ran the gdb test suite, and noticed that
the new Rust parser was missing testing on a few lines that were easy
to cover. This patch mostly adds tests for certain syntax errors; but
this process also uncovered a couple of real bugs: I must have
cut-and-pasted the 'sizeof' parsing code from some other code, because
it is checking for KW_MUT (the old bison parser did not do this), and
the array length check is actually impossible because a negative
number like '-1' is parsed as two tokens.
gdb/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* rust-parse.c (rust_parser::parse_sizeof): Remove KW_MUT code.
(struct typed_val_int) <val>: Now ULONGEST.
(rust_parser::parse_array_type): Remove negative check.
(rust_lex_int_test): Change 'value' to ULONGEST.
gdb/testsuite/ChangeLog
2021-04-22 Tom Tromey <tom@tromey.com>
* gdb.rust/modules.exp: Add checks for syntax errors.
* gdb.rust/expr.exp: Add checks for syntax errors.
* gdb.rust/simple.exp: Add checks for syntax errors.
Harmonize and improve auxiliary entries support for XCOFF
bfd/ChangeLog:
* coff-rs6000.c (_bfd_xcoff_swap_aux_in): Add errors for
unsupported storage class or auxialiry entries.
Improve and adapt to new aux structures.
Add C_DWARF support.
(_bfd_xcoff_swap_aux_out): Likewise.
* coff64-rs6000.c (_bfd_xcoff64_swap_aux_in): Likewise.
(_bfd_xcoff64_swap_aux_out): Likewise.
binutils/ChangeLog:
* od-xcoff.c (dump_xcoff32_symbols): Adapt to new
aux structures.
include/ChangeLog:
* coff/internal.h (union internal_auxent):
Add x_sect structure.
* coff/rs6000.h (union external_auxent): Rework to
match official documentation.
* coff/rs6k64.h (union external_auxent): Likewise.
(_AUX_SECT): New define.
Jan Beulich [Thu, 22 Apr 2021 08:55:07 +0000 (10:55 +0200)]
x86-64/PE: adjust PR ld/26659 testcase for Cygwin
While the testcase put in place by 74edb473c9ec ("PE/Windows x86_64: Fix
weak undef symbols after image base change") is fine for MingW, it fails
for Cygwin. This is because the default image base is different there
(for whatever reason).
Andrew Burgess [Wed, 21 Apr 2021 09:59:38 +0000 (10:59 +0100)]
gdb: allow default_addressable_memory_unit_size to handle more cases
Currently default_addressable_memory_unit_size always returns 1,
indicating 1 byte is 1 octet. If a target has something other than
this (common) setup then the target should override the
default_addressable_memory_unit_size.
However, the bfd library already knows about each targets octets per
byte, so it seems redundant making targets override this method to
tell GDB something it already knows (through bfd).
In this commit I propose to make default_addressable_memory_unit_size
return a value based on bfd's bits per byte. I checked, and for every
target that GDB currently supports the bits per byte in bfd is 8, so
the current behaviour will not change.
In fact, the only targets in bfd that have bits per byte set to
something other than 8 can be found in cpu-tic4x.c and cpu-tic54x.c, I
don't believe these are supported by GDB right now.
I don't propose to remove the ability to override
default_addressable_memory_unit_size, this allows targets additional
flexibility for how to handle weird combinations of byte sizes.
This change was motivated by an out of tree target I was working on,
but it seemed like it was a good change that others might benefit
from.
There should be no user visible changes after this commit.
gdb/ChangeLog:
* arch-utils.c (default_addressable_memory_unit_size): Return a
value based on bfd's bits per byte.
This is because clang only warns by default when encountering an unknown
warning option, and the probe for supported warning flags is done
without -Werror. All flags are therefore accepted by configure, but
then it breaks when actually compiling a source file with -Werror.
Tom Tromey [Wed, 21 Apr 2021 22:10:44 +0000 (16:10 -0600)]
Shrink size of dwarf2_per_cu_data
I noticed some holes in struct dwarf2_per_cu_data. This patch
rearranges the type slightly, and shrinks the size of some fields.
This reduces it from 136 bytes to 112 bytes (on x86-64).
I also reduced the size of the DWARF "version" fields in a couple of
spots. It seemed needless to use a short to hold a value that ranges
from 2 to 5, and this also helped the goal of shrinking
dwarf2_per_cu_data.
Tom de Vries [Wed, 21 Apr 2021 19:54:03 +0000 (21:54 +0200)]
[gdb/build] Hardcode --with-included-regex
Currently gdb has a configure option:
...
$ ./src/gdb/configure --help
...
--without-included-regex
don't use included regex; this is the default on
systems with version 2 of the GNU C library (use
with caution on other system)
...
The configure option controls config.h macro USE_INCLUDED_REGEX, which is
used in gdb/gdb_regex.h to choose between:
- using regex from libiberty (which is included in the binutils-gdb.git repo,
hence the 'included' in USE_INCLUDED_REGEX), or
- using regex.h.
In the former case, the symbol regcomp is remapped to a symbol xregcomp, which
is then provided by libiberty.
In the latter case, the symbol regcomp is resolved at runtime, usually binding
to libc. However, there is no mechanism in place to enforce this.
PR27681 is an example of where that causes problems. On openSUSE Tumbleweed,
the ncurses package got the --with-pcre2 configure switch enabled, and solved
the resulting dependencies using:
...
$ cat /usr/lib64/libncursesw.so
/* GNU ld script */
-INPUT(/lib64/libncursesw.so.6 AS_NEEDED(-ltinfo -ldl))
+INPUT(/lib64/libncursesw.so.6 AS_NEEDED(-ltinfo -ldl -lpcre2-posix -lpcre2-8))
...
This lead to regcomp being bound to libpcre2-posix instead of libc.
This causes problems in several ways:
- by compiling using regex.h, we've already chosen a specific regex_t
implementation, and the one from pcre2-posix is not the same.
- in gdb_regex.c we use GNU regex function re_search, which pcre2-posix
doesn't provide, so while regcomp binds to pcre2-posix, re_search binds to
libc.
A note on the latter: it's actually a bug to compile a regex using regcomp and
then pass it to re_search. The GNU regex interface requires one to use
re_compile_pattern or re_compile_fastmap. But as long we're using one of the
GNU regex incarnations in gnulib, glibc or libiberty, we get away with this.
The PR could be fixed by adding -lc in a specific position in the link line,
to force regcomp to be bound to glibc. But this solution was considered
in the discussion in the PR as being brittle, and possibly causing problems
elsewhere.
Another solution offered was to restrict regex usage to posix, and no longer
use the GNU regex API. This however could mean having to reproduce some of
that functionality locally, which would mean maintaining the same
functionality in more than one place.
The solution chosen here, is to hardcode --with-included-regex, that is, using
libiberty.
The option of using glibc for regex was introduced because glibc became the
authorative source for GNU regex, so it offered the possibility to link
against a more up-to-date regex version.
In that aspect, this patch is a step back. But we have the option of using a
more up-to-date regex version as a follow-up step: by using the regex from
gnulib.
Eli Zaretskii [Wed, 21 Apr 2021 16:38:52 +0000 (17:38 +0100)]
Extend the description of PE header flags.
PR 27760
include * coff/pe.h (IMAGE_DLLCHARACTERISTICS_APPCONTAINER): Define.
(IMAGE_DLLCHARACTERISTICS_GUARD_CF): Define.
bfd * peXXigen.c (_bfd_XX_print_private_bfd_data_common): Add display
of IMAGE_FILE_REMOVABLE_RUN_FROM_SWAP,
IMAGE_FILE_NET_RUN_FROM_SWAP and IMAGE_FILE_UP_SYSTEM_ONLY flags.
Decode the contents of the DllCharacteristics field.
gdb/breakpoint: add a 'force_condition' parameter to 'create_breakpoint'
The 'create_breakpoint' function takes a 'parse_extra' argument that
determines whether the condition, thread, and force-condition
specifiers should be parsed from the extra string or be used from the
function arguments. However, for the case when 'parse_extra' is
false, there is no way to pass the force-condition specifier. This
patch adds it as a new argument.
Also, in the case when parse_extra is false, the current behavior is
as if the condition is being forced. This is a bug. The default
behavior should reject the breakpoint. See below for a demo of this
incorrect behavior. (The MI command '-break-insert' uses the
'create_breakpoint' function with parse_extra=0.)
$ gdb -q --interpreter=mi3 /tmp/simple
=thread-group-added,id="i1"
=cmd-param-changed,param="history save",value="on"
=cmd-param-changed,param="auto-load safe-path",value="/"
~"Reading symbols from /tmp/simple...\n"
(gdb)
-break-insert -c junk -f main
&"warning: failed to validate condition at location 1, disabling:\n "
&"No symbol \"junk\" in current context.\n"
^done,bkpt={number="1",type="breakpoint",disp="keep",enabled="y",addr="<MULTIPLE>",cond="junk",times="0",original-location="main",locations=[{number="1.1",enabled="N",addr="0x000000000000114e",func="main",file="/tmp/simple.c",fullname="/tmp/simple.c",line="2",thread-groups=["i1"]}]}
(gdb)
break main if junk
&"break main if junk\n"
&"No symbol \"junk\" in current context.\n"
^error,msg="No symbol \"junk\" in current context."
(gdb)
break main -force-condition if junk
&"break main -force-condition if junk\n"
~"Note: breakpoint 1 also set at pc 0x114e.\n"
&"warning: failed to validate condition at location 1, disabling:\n "
&"No symbol \"junk\" in current context.\n"
~"Breakpoint 2 at 0x114e: file /tmp/simple.c, line 2.\n"
=breakpoint-created,bkpt={number="2",type="breakpoint",disp="keep",enabled="y",addr="<MULTIPLE>",cond="junk",times="0",original-location="main",locations=[{number="2.1",enabled="N",addr="0x000000000000114e",func="main",file="/tmp/simple.c",fullname="/tmp/simple.c",line="2",thread-groups=["i1"]}]}
^done
(gdb)
After applying this patch, we get the behavior below:
(gdb)
-break-insert -c junk -f main
^error,msg="No symbol \"junk\" in current context."
This restores the behavior that is present in the existing releases.
gdb/ChangeLog:
2021-04-21 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* breakpoint.h (create_breakpoint): Add a new parameter,
'force_condition'.
* breakpoint.c (create_breakpoint): Use the 'force_condition'
argument when 'parse_extra' is false to check if the condition
is invalid at all of the breakpoint locations.
Update the users below.
(break_command_1)
(dprintf_command)
(trace_command)
(ftrace_command)
(strace_command)
(create_tracepoint_from_upload): Update.
* guile/scm-breakpoint.c (gdbscm_register_breakpoint_x): Update.
* mi/mi-cmd-break.c (mi_cmd_break_insert_1): Update.
* python/py-breakpoint.c (bppy_init): Update.
* python/py-finishbreakpoint.c (bpfinishpy_init): Update.
gdb/testsuite/ChangeLog:
2021-04-21 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdb.mi/mi-break.exp: Extend with checks for invalid breakpoint
conditions.
gdb/breakpoint: display "N" on MI for disabled-by-condition locations
For breakpoint locations that are disabled because of an invalid
condition, CLI displays "N*" in the 'enabled' field, where '*' refers
to the footnote below the table:
(*): Breakpoint condition is invalid at this location.
This is not necessary for MI, where we shall simply print "N" without
the footnote.
Update the document to mention the "N" value for the MI. Also remove
the line about the 'enable' field, because there is no such field for
locations.
gdb/ChangeLog:
2021-04-21 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* breakpoint.c (print_one_breakpoint_location): Display "N" for
disabled-by-condition locations on MI-like output.
(breakpoint_1): Do not display the disabled-by-condition footnote
if the output is MI-like.
gdb/doc/ChangeLog:
2021-04-21 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdb.texinfo (GDB/MI Breakpoint Information): Update the
description for the 'enabled' field of breakpoint locations.
Nick Lott [Wed, 21 Apr 2021 14:33:13 +0000 (15:33 +0100)]
Add ability to select numeric base when displaying symbol values in readelf.
PR 27672
* readelf.c (sym_base): New variable.
(enum print_mode): Add more modes.
(print_vma): Add suport for new modes.
(options): Add sym-base.
(usage): Add sym-base.
(parse_args): Add support for --sym-base.
(print_dynamic_symbol_size): New function.
(print_dynamic_symbol): Use new function.
* doc/binutils.texi: Document the new feature.
* NEWS: Mention the new feature.
Nick Clifton [Wed, 21 Apr 2021 09:34:32 +0000 (10:34 +0100)]
Adjust readelf's output so that section symbols without a name as shown with their section name.
binutils* readelf.c (print_dynamic_symbol): Print the section name for
section symbols without a name of their own.
* testsuite/binutils-all/mips/global-local-symtab-sort-n64t.d:
Adjust expected output to allow for named section symbols.
* testsuite/binutils-all/mips/global-local-symtab-sort-o32t.d:
Likewise.
* testsuite/binutils-all/readelf.s-64: Likewise.
* testsuite/binutils-all/readelf.ss-64-unused: Likewise.
* testsuite/binutils-all/readelf.ss-tmips: Likewise.
* testsuite/binutils-all/readelf.ss-unused: Likewise.
Rework the R_NEG support on both gas and ld for the PowerPC AIX targets, in order to manage C++ exceptions built with GCC.
bfd PR binutils/21700
* reloc.c (BFD_RELOC_PPC_NEG): New relocation.
* bfd-in2.h: Regenerate.
* libbfd.h: Regenerate.
* coff-rs6000.c (_bfd_xcoff_reloc_type_lookup): Add
BFD_RELOC_PPC_NEG handler.
(xcoff_reloc_type_neg): Correctly substract addend.
* coff64-rs6000.c (xcoff64_howto_table): Add R_NEG_32
howto.
(xcoff64_rtype2howto): Add handler for R_NEG_32.
(xcoff64_reloc_type_lookup): Add BFD_RELOC_PPC_NEG handler.
* xcofflink.c (xcoff_need_ldrel_p): Check output section
for R_POS-like relocations. New argument added.
(xcoff_mark): Adapt to new xcoff_need_ldrel_p argument.
(xcoff_link_input_bfd): Likewise.
gas * config/tc-ppc.c (ppc_get_csect_to_adjust): New function.
(ppc_fix_adjustable): Manage fx_subsy part.
(tc_gen_reloc): Create second relocation when both
fx_addsy and fx_subsy are provided.
* config/tc-ppc.h (RELOC_EXPANSION_POSSIBLE): New define.
(MAX_RELOC_EXPANSION): Likewise.
(TC_FORCE_RELOCATION_SUB_SAME): Likewise
(UNDEFINED_DIFFERENCE_OK): Likewise
* testsuite/gas/all/gas.exp: Skip difference between two
undefined symbols test.
ld * testsuite/ld-powerpc/aix52.exp: Add new test.
* testsuite/ld-powerpc/aix-neg-reloc-32.d: New test.
* testsuite/ld-powerpc/aix-neg-reloc-64.d: New test.
* testsuite/ld-powerpc/aix-neg-reloc.ex: New test.
* testsuite/ld-powerpc/aix-neg-reloc.s: New test.
Mike Frysinger [Tue, 20 Apr 2021 13:21:28 +0000 (09:21 -0400)]
sim: rl78/rx: drop unnecessary getopt.h probing
Since getopt.h is provided by libiberty, there's no need to probe for
a system version of it. Plus we already assume it exists in other
parts of the sim.
Felix Willgerodt [Tue, 20 Apr 2021 06:44:17 +0000 (08:44 +0200)]
gdb: Fix reduce/reduce conflicts for qualifier_seq_noopt in the C parser.
This fixes a problem with GDB's address space qualifier parsing. GDB uses
'@' as a way to express an address space in expression evaluation. This can
currently lead to a crash for "Add support for the __flash qualifier on AVR"
(487d975399dfcb2bb2f0998a7d12bd62acdd9fa1), the only user I am aware of.
int
main (void)
{
const __flash char *pointer_to_flash = &data_in_flash;
}
~~~
Before:
~~~
(gdb) p data_in_flash
$1 = -85 '\253'
(gdb) p *(const char * @flash) pointer_to_flash
$2 = -85 '\253'
(gdb) p *(@flash const char *) pointer_to_flash
type-stack.c:201: internal-error: type* type_stack::follow_types(type*): unrecognized tp_ value in follow_types
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)
~~~
Note that how the binding of this qualifier is interpreted and resolved for an
address/pointer is target specific. Hence only the prepended qualifier works
for AVR, even if it seems syntactically incorrect. I won't change this for
AVR, as I am not familiar with that target.
Felix Willgerodt [Tue, 20 Apr 2021 06:43:08 +0000 (08:43 +0200)]
gdb: Allow address space qualifier parsing in C++.
The goal of this patch is to allow target dependent address space qualifiers
in the C++ expression parser. This can be useful for memory examination on
targets that actually use different address spaces in hardware without
having to deep-dive into implementation details of the whole solution.
GDB uses the @ symbol to parse address space qualifiers. The only current
user that I am aware of is the __flash support for avr, which was added in
"Add support for the __flash qualifier on AVR"
(487d975399dfcb2bb2f0998a7d12bd62acdd9fa1)
and only works for C.
One use-case of the AVR patch is:
~~~
const __flash char data_in_flash = 0xab;
int
main (void)
{
const __flash char *pointer_to_flash = &data_in_flash;
}
~~~
I want to enable a similar usage of e.g. @local in C++.
Before this patch (using "set debug parser on"):
~~~
(gdb) p *(int* @local) 0x1234
(...)
Reading a token: Next token is token '@' ()
Shifting token '@' ()
Entering state 46
Reading a token: Next token is token UNKNOWN_CPP_NAME (ssym<name=local, sym=(null), field_of_this=0>)
A syntax error in expression, near `local) &x'.
~~~
After:
~~~
(gdb) p *(int* @local) 0x1234
(...)
Reading a token: Next token is token '@' ()
Shifting token '@' ()
Entering state 46
Reading a token: Next token is token UNKNOWN_CPP_NAME (ssym<name=local, sym=(null), field_of_this=0>)
Shifting token UNKNOWN_CPP_NAME (ssym<name=local, sym=(null), field_of_this=0>)
Entering state 121
Reducing stack by rule 278 (line 1773):
$1 = token UNKNOWN_CPP_NAME (ssym<name=local, sym=(null), field_of_this=0>)
-> $$ = nterm name ()
Stack now 0 49 52 76 222 337 46
Entering state 167
Reducing stack by rule 131 (line 1225):
$1 = token '@' ()
$2 = nterm name ()
Unknown address space specifier: "local"
~~~
The "Unknown address space qualifier" is the right behaviour, as I ran this
on a target that doesn't have multiple address spaces and therefore obviously
no support for such qualifiers.
gdb/ChangeLog:
2021-04-20 Felix Willgerodt <felix.willgerodt@intel.com>
aarch64: New instructions for maintenance of GPT entries cached in a TLB
This patch adds support to four new system registers (RPAOS, RPALOS, PAALLOS,
PAALL) in conjunction with TLBI instruction. This change is part of RME (Realm
Management Extension).
aarch64: Add new data cache maintenance operations
This patch adds support to two new system registers (CIPAPA, CIGDPAPA) in
conjunction with DC instruction. This change is part of RME (Realm Management
Extension).
Tom Tromey [Mon, 19 Apr 2021 13:16:34 +0000 (07:16 -0600)]
Fix kfail patterns in inline-locals.exp
PR gdb/27742 points out that my recent change to
print_variable_and_value caused a regression in inline-locals.exp. I
can't reproduce this, but I came up with this patch based on the
output shown in the bug.
gdb/testsuite/ChangeLog
2021-04-19 Tom Tromey <tromey@adacore.com>
Jan Beulich [Mon, 19 Apr 2021 11:28:35 +0000 (13:28 +0200)]
gas: drop sprint_value()
Its (documented) behavior is unhelpful in particular in 64-bit build
environments: While printing large 32-bit numbers in decimal already
isn't very meaningful to most people, this even more so goes for yet
larger 64-bit numbers. bfd_sprintf_vma() still tries to limit the number
of digits printed (without depending on a build system property), but
uniformly produces hex output.
Mike Frysinger [Mon, 19 Apr 2021 02:48:31 +0000 (22:48 -0400)]
sim: moxie: switch syscalls to common nltvals
Rather than hand duplicate the syscall constants, switch to the
common nltvals framework. I made sure the constants have the
same values before & after too :).
Mike Frysinger [Mon, 19 Apr 2021 02:43:06 +0000 (22:43 -0400)]
sim: sh: switch syscalls to common nltvals
Rather than hand duplicate the syscall table, switch to the common
nltvals framework. We have to tweak the constant names, but we get
everything else for free. I made sure the constants have the same
values before & after too :).
Mike Frysinger [Mon, 19 Apr 2021 02:35:19 +0000 (22:35 -0400)]
sim: rx: switch syscalls to common nltvals
Rather than hand duplicate the syscall table, switch to the common
nltvals framework. We have to tweak the constant names, but we get
everything else for free. I made sure the constants have the same
values before & after too :).
Mike Frysinger [Mon, 19 Apr 2021 02:31:19 +0000 (22:31 -0400)]
sim: m32c: switch syscalls to common nltvals
Rather than hand duplicate the syscall table, switch to the common
nltvals framework. We have to tweak the constant names, but we get
everything else for free. I made sure the constants have the same
values before & after too :).
Mike Frysinger [Mon, 19 Apr 2021 02:24:37 +0000 (22:24 -0400)]
sim: iq2000: switch syscalls to common nltvals
Rather than hand duplicate the syscall table, switch to the common
nltvals framework. We have to tweak the constant names, but we get
everything else for free. I made sure the constants have the same
values before & after too :).
Mike Frysinger [Mon, 19 Apr 2021 01:01:00 +0000 (21:01 -0400)]
sim: cr16: fix build warnings
The printf fix is obvious enough, but the hash one is a real bug:
cr16/interp.c: In function 'sim_open':
cr16/interp.c:560:17: error: 'h' may be used uninitialized in this function [-Werror=maybe-uninitialized]
560 | h = h->next;
| ~~^~~~~~~~~
It happens to not cause a problem currently because the first entry in
the generated table that this loop operates matches a codepath where h
is initialized. Then when later entries don't match, the previous value
is pointing at the end of a valid hash table already, and the rest of
the code does nothing.
With this tidied up, we can delete the SIM_AC_OPTION_WARNINGS(no) call
to get the default common behavior where -Werror is enabled.
Tom Tromey [Sat, 17 Apr 2021 19:56:36 +0000 (13:56 -0600)]
Handle unaligned mapping of .gdb_index
The .gdb_index was designed such that all data would be aligned.
Unfortunately, we neglected to require this alignment in the objcopy
instructions in the manual. As a result, in many cases, a .gdb_index
in the wild will not be properly aligned by mmap. This yields
undefined behavior, which is PR gdb/23743.
This patch fixes the bug by always assuming that the mapping is
unaligned, and using extract_unsigned_integer when needed. A new
helper class is introduced to make this less painful.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
quick_symbol_functions::map_matching_symbols is only used by the Ada
code. Currently, it both expands certain psymtabs and then walks over
the full symtabs -- including any already-expanded ones -- calling a
callback.
It appears to work lazily as well, in that if the callback returns
false, iteration stops. However, only the psymtab implementation does
this; the DWARF index implementations are not lazy. It turns out,
though, that the only callback that is ever passed here never returns
false.
This patch simplifies this method by removing the callback. The
method is also renamed. In the new scheme, the caller is responsible
for walking the full symtabs, which removes some redundancy as well.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
* psymtab.c (psymbol_functions::expand_matching_symbols): Rename
from map_matching_symbols. Change parameters.
* psympriv.h (struct psymbol_functions) <expand_matching_symbols>:
Rename from map_matching_symbols. Change parameters.
* dwarf2/read.c (struct dwarf2_gdb_index)
<expand_matching_symbols>: Rename from map_matching_symbols.
Change parameters.
(struct dwarf2_debug_names_index) <expand_matching_symbols>:
Rename from map_matching_symbols. Change parameters.
(dwarf2_gdb_index::expand_matching_symbols): Rename from
dw2_map_matching_symbols. Change parameters.
(dwarf2_gdb_index::expand_matching_symbols): Remove old
implementation.
(dwarf2_debug_names_index::expand_matching_symbols): Rename from
map_matching_symbols. Change parameters.
* objfiles.h (struct objfile) <expand_matching_symbols>: Rename
from map_matching_symbols. Change parameters.
* symfile-debug.c (objfile::expand_matching_symbols): Rename from
map_matching_symbols. Change parameters.
* ada-lang.c (map_matching_symbols): New function.
(add_nonlocal_symbols): Update.
This removes quick_symbol_functions::expand_symtabs_with_fullname,
replacing it with a call to expand_symtabs_matching. As with the
previous patches, the implementation is consolidated in the objfile
method.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
This removes quick_symbol_functions::expand_symtabs_for_function,
replacing it with a call to expand_symtabs_matching. As with the
previous patches, the implementation is consolidated in the objfile
method.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
This replaces quick_symbol_functions::map_symtabs_matching_filename
with a call to expand_symtabs_matching. As with the previous patch,
rather than update all callers, the implementation is consolidated in
objfile::map_symtabs_matching_filename.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 17 Apr 2021 15:35:04 +0000 (09:35 -0600)]
Remove quick_symbol_functions::lookup_symbol
This removes quick_symbol_functions, replacing it with calls to
expand_symtabs_matching. Because the replacement is somewhat verbose,
objfile::lookup_symbol is not removed. This consolidates some
duplicated code into this one spot.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 17 Apr 2021 15:35:04 +0000 (09:35 -0600)]
Add 'domain' parameter to expand_symtabs_matching
Currently, expand_symtabs_matching only accepts a search_domain
parameter. However, lookup_symbol uses a domain_enum instead, and the
two, confusingly, do quite different things -- one cannot emulate the
other. So, this patch adds a domain_enum parameter to
expand_symtabs_matching, with UNDEF_DOMAIN used as a wildcard.
This is another step toward replacing lookup_symbol with
expand_symtabs_matching.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 17 Apr 2021 15:35:04 +0000 (09:35 -0600)]
Add search_flags to expand_symtabs_matching
This adds a block search flags parameter to expand_symtabs_matching.
All callers are updated to search both the static and global blocks,
as that was the implied behavior before this patch.
This is a step toward replacing lookup_symbol with
expand_symtabs_matching.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 17 Apr 2021 15:35:04 +0000 (09:35 -0600)]
Let expand_symtabs_matching short-circuit
This changes expand_symtabs_exp_notify_ftype to return bool, and
updates all the uses. Now, if the notification function returns
false, the call is short-circuited and stops examining symtabs. This
is a step toward replacing map_symtabs_matching_filename with
expand_symtabs_matching.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 17 Apr 2021 15:35:04 +0000 (09:35 -0600)]
Add block_search_flags
This adds block_search_flags, a flag enum. This will be used to by
certain search functions so that the caller can control which blocks
are searched more precisely.
gdb/ChangeLog
2021-04-17 Tom Tromey <tom@tromey.com>
* quick-symbol.h (enum block_search_flag_values): New.
(block_search_flags): New enum flags type.
Tom Tromey [Fri, 16 Apr 2021 22:34:07 +0000 (16:34 -0600)]
Rewrite the Rust expression parser
The Rust expression parser was written to construct its own AST, then
lower this to GDB expressions. I did this primarily because the old
expressions were difficult to work with; after rewriting those, I
realized I could remove the AST from the Rust parser.
After looking at this, I realized it might be simpler to rewrite the
parser. This patch reimplements it as a recursive-descent parser. I
kept a fair amount of the existing code -- the lexer is pulled in
nearly unchanged.
There are several benefits to this approach:
* The parser is shorter now (from 2882 LOC to 2351).
* The parser is just ordinary C++ code that can be debugged in the
usual way.
* Memory management in the parser is now straightforward, as
parsing methods simply return a unique pointer or vector.
This required a couple of minor changes to the test suite, as some
errors have changed.
While this passes the tests, it's possible there are lurking bugs,
particularly around error handling.
gdb/ChangeLog
2021-04-16 Tom Tromey <tom@tromey.com>
Simon Marchi [Fri, 16 Apr 2021 15:01:26 +0000 (11:01 -0400)]
gdb/testsuite: use -gdwarf-4 in simavr board
By default, when using the -g switch, avr-gcc generates stabs debug
information. I think it would make more sense to test GDB against DWARF
information, because stabs is obsolete by now, and nobody work on it.
So change the simavr board to pass -gdwarf-4 as the debug flag.
The downside is that users are probably more likely to use just -g, so
we don't test GDB the same way as users are likely to use it. But in
this case, if somebody comes and asks for help with GDB for AVR, I
suggest we encourage them to use -gdwarf-4.
I can't give stats about how that changes test results, because the
testsuite is too broken. There is an internal error that happens quite
frequently that needs to be investigated:
* testsuite/gas/aarch64/rme-invalid.d: New test.
* testsuite/gas/aarch64/rme-invalid.l: New test.
* testsuite/gas/aarch64/rme-invalid.s: New test.
* testsuite/gas/aarch64/rme.d: New test.
* testsuite/gas/aarch64/rme.s: New test.
Luis Machado [Wed, 14 Apr 2021 14:20:18 +0000 (11:20 -0300)]
Print bfloat16 DWARF types correctly
Even if the DWARF information contains a bfloat16 base type (__bf16), a
variable of such type will still be printed using the IEEE half float format,
which is wrong.
This patch teaches GDB how to pick the bfloat16 format for __bf16 types in
DWARF (based on the base type name) and uses IEEE half float for all the other
16-bit float formats.
Alan Modra [Fri, 16 Apr 2021 12:59:05 +0000 (22:29 +0930)]
PR27567, Linking PE files adds alignment section flags to executables
So don't set those flags for an executable. In the patch I also test
DYNAMIC even though the PE bfd code doesn't appear to set it for dlls.
I figure it doesn't hurt to include that flag too.
PR 27567
bfd/
* coffcode.h (styp_to_sec_flags): Use an unsigned long styp_flags.
(coff_write_object_contents): Pass bfd to COFF_ENCODE_ALIGNMENT,
ignore alignment checks when return is false. Formatting.
include/
* coff/internal.h (struct internal_scnhdr): Make s_flags unsigned long.
* coff/pe.h (COFF_ENCODE_ALIGNMENT): Don't set align flags for an
executable and return false. Do so for a relocatable object and
evaluate to true.
* coff/ti.h (COFF_ENCODE_ALIGNMENT): Add bfd arg and evaluate to true.
(COFF_DECODE_ALIGNMENT): Formatting.
* coff/z80.h (COFF_ENCODE_ALIGNMENT): Similarly.
(COFF_DECODE_ALIGNMENT): Similarly.
John Baldwin [Thu, 15 Apr 2021 23:03:15 +0000 (16:03 -0700)]
sim mn10300: Fix igen generation.
Only pass the top-level instruction decode table (mn10300.igen) to
igen via -i. The additional files passed previously caused igen to
exit its getopt loop in main and exit silently without generating any
output. In addition, when am33-2.igen was added, it was not included
from mn10300.igen, so was never used.
sim/mn10300/ChangeLog:
* Makefile.in: (tmp-igen) Only pass mn10300.igen to igen.
* mn10300.igen: Include am33-2.igen.
John Baldwin [Thu, 15 Apr 2021 23:03:15 +0000 (16:03 -0700)]
sim: Add SIM_EXTRA_CFLAGS after CSEARCH.
The bfin sim adds include paths for the SDL libraries. These include
paths might include headers for different version of binutils. Move
SIM_EXTRA_CFLAGS after CSEARCH to ensure local includes are always
preferred to external includes.
projects / thirdparty / binutils-gdb.git / log
