This patch adds AArch32 support for -march=armv9.[123]-a.
The behaviour of the new options can be expressed using a
combination of existing feature flags and tables.
The cpu_arch_ver entries for ARM_ARCH_V9_2A and ARM_ARCH_V9_3A
are technically redundant but it seemed less surprising to include
them anyway.
include/
* opcode/arm.h (ARM_ARCH_V9_1A, ARM_ARCH_V9_2A): New macros.
(ARM_ARCH_V9_3A): Likewise.
This patch adds AArch32 support for -march=armv8.[78]-a.
The behaviour of the new options can be expressed using a
combination of existing feature flags and tables.
The cpu_arch_ver entries are technically redundant but
it seemed less surprising to include them anyway.
include/
* opcode/arm.h (ARM_ARCH_V8_7A, ARM_ARCH_V8_8A): New macros.
gas/
* doc/c-arm.texi: Add armv8.7-a and armv8.8-a.
* config/tc-arm.c (armv87a_ext_table, armv88a_ext_table): New macros.
(arm_archs): Add armv8.7-a and armv8.8-a.
(cpu_arch_ver): Add ARM_ARCH_V8_7A and ARM_ARCH_V8_8A.
* NEWS: Mention the above.
* testsuite/gas/arm/attr-march-armv8_7-a.d: New test.
* testsuite/gas/arm/attr-march-armv8_8-a.d: Likewise.
* testsuite/gas/arm/bfloat16-armv8.7-a.d: Likewise.
* testsuite/gas/arm/bfloat16-armv8.8-a.d: Likewise.
* testsuite/gas/arm/i8mm-armv8.7-a.d: Likewise.
* testsuite/gas/arm/i8mm-armv8.8-a.d: Likewise.
This patch adds AArch64 support for -march=armv9.[123]-a.
The behaviour of the new options can be expressed using a
combination of existing feature flags, so we don't need to
eat into the vanishing number of spare AARCH64_FEATURE_* bits.
Hoewver, it was more convenient to separate out the |s of
feature flags so that Armv9.1-A could reuse the set for
Armv8.6-A, and so on.
include/
* opcode/aarch64.h (AARCH64_ARCH_V8_FEATURES): New macro,
split out from...
(AARCH64_ARCH_V8): ...here.
(AARCH64_ARCH_V8_1_FEATURES): New macro, split out from...
(AARCH64_ARCH_V8_1): ...here.
(AARCH64_ARCH_V8_2_FEATURES): New macro, split out from...
(AARCH64_ARCH_V8_2): ...here.
(AARCH64_ARCH_V8_3_FEATURES): New macro, split out from...
(AARCH64_ARCH_V8_3): ...here.
(AARCH64_ARCH_V8_4_FEATURES): New macro, split out from...
(AARCH64_ARCH_V8_4): ...here.
(AARCH64_ARCH_V8_5_FEATURES): New macro, split out from...
(AARCH64_ARCH_V8_5): ...here.
(AARCH64_ARCH_V8_6_FEATURES): New macro, split out from...
(AARCH64_ARCH_V8_6): ...here.
(AARCH64_ARCH_V8_7_FEATURES): New macro, split out from...
(AARCH64_ARCH_V8_7): ...here.
(AARCH64_ARCH_V8_8_FEATURES): New macro, split out from...
(AARCH64_ARCH_V8_8): ...here.
(AARCH64_ARCH_V9_FEATURES): New macro, split out from...
(AARCH64_ARCH_V9): ...here.
(AARCH64_ARCH_V9_1_FEATURES, AARCH64_ARCH_V9_1): New macros.
(AARCH64_ARCH_V9_2_FEATURES, AARCH64_ARCH_V9_2): New macros.
(AARCH64_ARCH_V9_3_FEATURES, AARCH64_ARCH_V9_3): New macros.
Nelson Chu [Thu, 22 Jul 2021 05:47:07 +0000 (13:47 +0800)]
RISC-V: Support svinval extension with frozen version 1.0.
According to the privileged spec, there are five new instructions for
svinval extension. Two of them (HINVAL.VVMA and HINVAL.GVMA) need to
enable the hypervisor extension. But there is no implementation of
hypervisor extension in mainline for now, so let's consider the related
issues later.
This patch is cherry-picked from the riscv integration branch since the
svinval extension is frozen for now. Besides, we fix the funct7 encodings
of hinval.vvma and hinval.gvma, from 0x0011011 and 0x0111011 to 0x0010011
and 0x0110011.
Mike Frysinger [Tue, 30 Nov 2021 07:30:46 +0000 (02:30 -0500)]
sim: mips/or1k: drop redundant arg to bitsize macro
These are just using the default behavior for the 3rd arg, so drop
it to make it more clear. This also makes them match all other
ports that only use the first 2 arguments.
Mike Frysinger [Thu, 16 Dec 2021 07:25:21 +0000 (02:25 -0500)]
sim: fix mingw builds with replacement gnulib open
The header shuffling in here broke the workaround for gnulib defining
"open". Move it back before the sim-specific includes to fix. This
is because the callback struct in the headers has an "open" member and
this file tries to call that.
Sandra Loosemore [Wed, 15 Dec 2021 23:05:41 +0000 (16:05 -0700)]
Adjust compare_link_order for unstable qsort
In a cross toolchain for nios2-elf target and x86_64-w64-mingw32 host
using binutils 2.37, we observed a failure that didn't show up on
x86_64-linux-gnu host: testcase pr25490-5.s was failing with
C:\path\to\nios2-elf-ld.exe: looping in map_segments
FAIL: __patchable_function_entries section 5
* ldelfgen.c (compare_link_order): Don't use section id in
sorting. Keep original ordering instead. Update comments.
Alan Modra [Thu, 16 Dec 2021 00:20:58 +0000 (10:50 +1030)]
Re: Fix an undefined behaviour in the BFD library's DWARF parser
Using an unsigned int cast (to 32 bits) on a pointer difference (of
possibly 64 bits) is wrong. Even though it will work on all real
object files, the fuzzers will eventually find this hole.
PR 28687
* dwarf1.c (parse_die): Cast pointer difference to size_t.
Catch another possible pointer overflow.
Alan Modra [Wed, 15 Dec 2021 04:36:26 +0000 (15:06 +1030)]
PR28691, validate dwarf attribute form
PR28691 is a fuzzing PR that triggers a non-problem of "output changes
per run" with PIEs and/or different compilers. I've closed similar
PRs before as wontfix, but I guess there will be no end of this type
of PR. The trigger is an attribute that usually takes one of the
offset/constant reference DW_FORMs being given an indexed string
DW_FORM. The bfd reader doesn't support indexed strings and returns
an error string instead. The address of the string varies with PIE
runs and/or compiler, and we allow that address to appear in output.
Fix this by validating integer attribute forms, as we do for string
form attributes.
PR 28691
* dwarf2.c (is_str_attr): Rename to..
(is_str_form): ..this. Change param type. Update calls.
(is_int_form): New function.
(read_attribute_value): Handle DW_FORM_addrx2.
(find_abstract_instance): Validate form when using attr.u.val.
(scan_unit_for_symbols, parse_comp_unit): Likewise.
Luis Machado [Fri, 26 Nov 2021 14:31:18 +0000 (11:31 -0300)]
New --enable-threading configure option to control use of threads in GDB/GDBserver
Add the --enable-threading configure option so multithreading can be disabled
at configure time. This is useful for statically-linked builds of
GDB/GDBserver, since the thread library doesn't play well with that setup.
If you try to run a statically-linked GDB built with threading, it will crash
when setting up the number of worker threads.
This new option is also convenient when debugging GDB in a system with lots of
threads, where the thread discovery code in GDB will emit too many messages,
like so:
Alan Modra [Tue, 14 Dec 2021 11:42:17 +0000 (22:12 +1030)]
loongarch32 build failure on 32-bit host
gas/config/tc-loongarch.c: In function ‘assember_macro_helper’:
gas/config/tc-loongarch.c:915:28: error: right shift count >= width of type [-Werror=shift-count-overflow]
915 | hi32 = insn->args[1] >> 32;
| ^~
One possible fix is to make offsetT a 64-bit type for loongarch32.
This also makes bfd/targmatch.h (generated from bfd/config.bfd)
consistent since the loongarch32 match is inside #ifdef BFD64.
Alan Modra [Tue, 14 Dec 2021 11:36:29 +0000 (22:06 +1030)]
loongarch64 build failure on 32-bit host
gas/config/tc-loongarch.c: In function ‘loongarch_args_parser_can_match_arg_helper’:
gas/config/tc-loongarch.c:661:13: error: cast from pointer to integer of different size [-Werror=pointer
-to-int-cast]
661 | imm = (offsetT) str_hash_find (r_htab, arg);
| ^
Cast it to the correct size int, relying on normal integer promotions
if offsetT is larger than a pointer.
* config/tc-loongarch.c (loongarch_args_parser_can_match_arg_helper):
Cast return from str_hash_find to intptr_t, not offsetT.
Alan Modra [Tue, 14 Dec 2021 10:43:21 +0000 (21:13 +1030)]
XCOFF C_STSYM test failure on 32-bit host
This test was failing here and on another similar symbol:
[ 4](sec 1)(fl 0x00)(ty 0)(scl 143) (nx 0) 0x05d1745d11745d21 .bs
where correct output is
[ 4](sec 1)(fl 0x00)(ty 0)(scl 143) (nx 0) 0x000000000000000a .bs
The problem is caused by a 32-bit host pointer being sign-extended
when stored into a 64-bit bfd_vma, and then that value not being
trimmed back to 32 bits when used. The following belt-and-braces
patch fixes both the store and subsequent reads.
* coffcode.h (coff_slurp_symbol_table): Do not sign extend
when storing a host pointer to syment.n_value.
* coffgen.c (coff_get_symbol_info): Cast syment.n_value to a
bfd_hostptr_t before using in arithmetic.
(coff_print_symbol): Likewise.
... trying to start an application with libinproctrace.so loaded
crashes:
$ LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libasan.so.6:./libinproctrace.so /bin/ls
/home/smarchi/src/binutils-gdb/gdbserver/../gdbsupport/common-utils.cc:69: A problem internal to GDBserver in-process agent has been detected.
xsnprintf: Assertion `ret < size' failed.
Looking at the rest of the socket initialization code, the intent seems
to be that if something goes wrong, we warn but let the program
execute. So crashing on this failed assertions seems against the intent.
Commit 6cebaf6e1ae4 ("use xsnprintf instead of snprintf.") changed this
code to use xsnprintf instead of snprintf, introducing this assertion.
Before that, snprintf would return a value bigger that UNIX_PATH_MAX and
the "if" after would catch it and emit a warning, which is exactly what
we want. That change was done because LynxOS didn't have snprintf.
Since LynxOS isn't supported anymore, we can simply revert to use
snprintf there.
With this patch, we get a warning (printed by the caller of
gdb_agent_socket_init), but the program keeps executing:
$ LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libasan.so.6:./libinproctrace.so /bin/ls
ipa: could not create sync socket
...
Simon Marchi [Tue, 14 Dec 2021 19:35:03 +0000 (14:35 -0500)]
gdbserver/tracepoint.cc: work around -Wstringop-truncation error
When building gdb with on AArch64 with g++ 11.1.0 (and some preceding
versions too), -O2 and -fsanitize=address, I get this error.
CXX tracepoint-ipa.o
cc1plus: warning: command-line option ‘-Wmissing-prototypes’ is valid for C/ObjC but not for C++
In file included from /usr/include/string.h:519,
from ../gnulib/import/string.h:41,
from /home/simark/src/binutils-gdb/gdbserver/../gdbsupport/common-defs.h:95,
from /home/simark/src/binutils-gdb/gdbserver/server.h:22,
from /home/simark/src/binutils-gdb/gdbserver/tracepoint.cc:19:
In function ‘char* strncpy(char*, const char*, size_t)’,
inlined from ‘int init_named_socket(const char*)’ at /home/simark/src/binutils-gdb/gdbserver/tracepoint.cc:6902:11,
inlined from ‘int gdb_agent_socket_init()’ at /home/simark/src/binutils-gdb/gdbserver/tracepoint.cc:6953:26,
inlined from ‘void* gdb_agent_helper_thread(void*)’ at /home/simark/src/binutils-gdb/gdbserver/tracepoint.cc:7204:41:
/usr/include/bits/string_fortified.h:95:34: error: ‘char* __builtin_strncpy(char*, const char*, long unsigned int)’ output may be truncated copying 107 bytes from a string of length 107 [-Werror=stringop-truncation]
95 | return __builtin___strncpy_chk (__dest, __src, __len,
| ~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
96 | __glibc_objsize (__dest));
| ~~~~~~~~~~~~~~~~~~~~~~~~~
Note that _FORTIFY_SOURCE changes the message a bit, but I get a similar
error if I use -D_FORTIFY_SOURCE=0.
I am pretty sure it's spurious and might be related to this GCC bug:
From what I can see, we are copying from a static 108 bytes long buffer
(the global array agent_socket_name) to a 108 bytes long array,
sockaddr_un::sun_path. I don't see anything wrong.
Still, it would make things easier if we didn't see this error. Change
the code to check that the source string length is smaller than the
destination buffer (including space for the NULL byte) and use strcpy.
For anybody who would like to try to reproduce, the full command line
is:
Simon Marchi [Fri, 3 Dec 2021 21:31:09 +0000 (16:31 -0500)]
bfd: fix -Wunused errors with clang 13+
Clang 13 and 14 produce some -Wunused-but-set-{variable,parameter} for
situations where gcc doesn't. In particular, when a variable is set and
then used in a way to update its own value. For example, if `i` is only
used in this way:
int i = 2;
i++;
i = i + 1;
gcc won't warn, but clang will.
Fix all such errors found in an --enable-targets=all build. It would be
important for somebody who knows what they're doing to just make sure
that these variables can indeed be deleted, and that there a no cases
where it's a bug, and the variable should actually be used.
The first instance of this error fix by this patch is:
CC elf32-score.lo
/home/simark/src/binutils-gdb/bfd/elf32-score.c:450:11: error: variable 'relocation' set but not used [-Werror,-Wunused-but-set-variable]
bfd_vma relocation;
^
Jan Vrany [Mon, 8 Nov 2021 14:00:13 +0000 (14:00 +0000)]
gdb/mi: rename mi_cmd to mi_command
Just give this class a new name, more inline with the name of the
sub-classes. I've also updated mi_cmd_up to mi_command_up in
mi-cmds.c inline with this new naming scheme.
There should be no user visible changes after this commit.
Jan Vrany [Tue, 23 Jun 2020 13:45:38 +0000 (14:45 +0100)]
gdb/mi: use separate classes for different types of MI command
This commit changes the infrastructure in mi-cmds.{c,h} to add new
sub-classes for the different types of MI command. Instances of these
sub-classes are then created and added into mi_cmd_table.
The existing mi_cmd class becomes the abstract base class, this has an
invoke method and takes care of the suppress notifications handling,
before calling a do_invoke virtual method which is implemented by all
of the sub-classes.
There's currently two different sub-classes, one of pure MI commands,
and a second for MI commands that delegate to CLI commands.
There should be no user visible changes after this commit.
Andrew Burgess [Mon, 8 Nov 2021 12:25:56 +0000 (12:25 +0000)]
gdb/mi: int to bool conversion in mi_execute_cli_command
Change an argument of mi_execute_cli_command from int to bool. Update
the callers to take this into account. Within mi_execute_cli_command,
update a comparison of a pointer to 0 with a comparison to nullptr,
and add an assert, if we are not using the argument string then the
string should be nullptr. Also removed a cryptic 'gdb_????' comment,
which isn't really helpful.
There should be no user visible changes after this commit.
Jan Vrany [Tue, 23 Jun 2020 13:45:38 +0000 (14:45 +0100)]
gdb/mi: use std::map for MI commands in mi-cmds.c
This changes the hashmap used in mi-cmds.c from a custom structure to
std::map. Not only is replacing a custom container with a standard
one an improvement, but using std::map will make it easier to
dynamically add commands; which is something that is planned for a
later series, where we will allow MI commands to be implemented in
Python.
There should be no user visible changes after this commit.
Nelson Chu [Tue, 14 Dec 2021 04:53:48 +0000 (12:53 +0800)]
RISC-V: Added ld testcases for the medlow and medany code models.
There are two linker scripts, code-model-01.ld and code-model-02.ld,
which are corresponding to the two different memory layouts,
* code-model-01.ld: the text section is in the 32-bit address range, but
the data section is far away from the text section, which means the data
section is over the 32-bit address range.
* code-model-02.ld: the text section is over the 32-bit address range, but
the data section is placed nearly zero address.
We use the two linker scripts, to test the current medlow and medany behaviors
of GNU ld, including the weak symbol references and the relaxations behaviors.
Besides, these testcases also show the limits of the current medlow and medany
code models, that is - we may get the truncated to fit errors when linking
with the above two linker scripts.
ld/
* testsuite/ld-riscv-elf/code-model-01.ld: New testcases to test the
behaviors of the current medlow and medany code models.
* testsuite/ld-riscv-elf/code-model-02.ld: Likewise.
* testsuite/ld-riscv-elf/code-model-medany-01.d: Likewise.
* testsuite/ld-riscv-elf/code-model-medany-02.d: Likewise.
* testsuite/ld-riscv-elf/code-model-medany-weakref-01.d: Likewise.
* testsuite/ld-riscv-elf/code-model-medany-weakref-02.d: Likewise.
* testsuite/ld-riscv-elf/code-model-medlow-01.d: Likewise.
* testsuite/ld-riscv-elf/code-model-medlow-02.d: Likewise.
* testsuite/ld-riscv-elf/code-model-medlow-weakref-01.d: Likewise.
* testsuite/ld-riscv-elf/code-model-medlow-weakref-02.d: Likewise.
* testsuite/ld-riscv-elf/code-model-relax-medany-01.d: Likewise.
* testsuite/ld-riscv-elf/code-model-relax-medany-02.d: Likewise.
* testsuite/ld-riscv-elf/code-model-relax-medany-weakref-01.d: Likewise.
* testsuite/ld-riscv-elf/code-model-relax-medany-weakref-02.d: Likewise.
* testsuite/ld-riscv-elf/code-model-relax-medlow-01.d: Likewise.
* testsuite/ld-riscv-elf/code-model-relax-medlow-02.d: Likewise.
* testsuite/ld-riscv-elf/code-model-relax-medlow-weakref-01.d: Likewise.
* testsuite/ld-riscv-elf/code-model-relax-medlow-weakref-02.d: Likewise.
* testsuite/ld-riscv-elf/code-model.s: Likewise.
* testsuite/ld-riscv-elf/ld-riscv-elf.exp: Updated.
Andrew Burgess [Fri, 15 Oct 2021 14:16:20 +0000 (15:16 +0100)]
gdb: improve reuse of value contents when fetching array elements
While working on a Python script, which was interacting with a remote
target, I noticed some weird slowness in GDB. In my program I had a
structure something like this:
struct foo_t
{
int array[5];
};
struct foo_t global_foo;
Then in the Python script I was fetching a complete copy of global
foo, like:
val = gdb.parse_and_eval('global_foo')
val.fetch_lazy()
Then I would work with items in foo_t.array, like:
print(val['array'][1])
I called the fetch_lazy method specifically because I knew I was going
to end up accessing almost all of the contents of val, and so I wanted
GDB to do a single remote protocol call to fetch all the contents in
one go, rather than trying to do lazy fetches for a couple of bytes at
a time.
What I observed was that, after the fetch_lazy call, GDB does,
correctly, fetch the entire contents of global_foo, including all of
the contents of array, however, when I access val.array[1], GDB still
goes and fetches the value of this element from the remote target.
What's going on is that in valarith.c, in value_subscript, for C like
languages, we always end up treating the array value as a pointer, and
then doing value_ptradd, and value_ind, the second of these calls
always returns a lazy value.
My guess is that this approach allows us to handle indexing off the
end of an array, when working with zero element arrays, or when
indexing a raw pointer as an array. And, I agree, that in these
cases, where, even when the original value is non-lazy, we still will
not have the content of the array loaded, we should be using the
value_ind approach.
However, for cases where we do have the array contents loaded, and we
do know the bounds of the array, I think we should be using
value_subscripted_rvalue, which is what we use for non C like
languages.
One problem I did run into, exposed by gdb.base/charset.exp, was that
value_subscripted_rvalue stripped typedefs from the element type of
the array, which means the value returned will not have the same type
as an element of the array, but would be the raw, non-typedefed,
type. In charset.exp we got back an 'int' instead of a
'wchar_t' (which is a typedef of 'int'), and this impacts how we print
the value. Removing typedefs from the resulting value just seems
wrong, so I got rid of that, and I don't see any test regressions.
With this change in place, my original Python script is now doing no
additional memory accesses, and its performance increases about 10x!
gdbserver: replace direct assignments to current_thread
Replace the direct assignments to current_thread with
switch_to_thread. Use scoped_restore_current_thread when appropriate.
There is one instance remaining in linux-low.cc's wait_for_sigstop.
This will be handled in a separate patch.
Regression-tested on X86-64 Linux using the native-gdbserver and
native-extended-gdbserver board files.
gdbserver: introduce scoped_restore_current_thread and switch_to_thread
Introduce a class for restoring the current thread and a function to
switch to the given thread. This is a preparation for a refactoring
that aims to remove direct assignments to 'current_thread'.
Andrew Burgess [Fri, 12 Nov 2021 11:18:58 +0000 (11:18 +0000)]
gdb: make post_startup_inferior a virtual method on inf_ptrace_target
While working on a later patch that required me to understand how GDB
starts up inferiors, I was confused by the
target_ops::post_startup_inferior method.
The post_startup_inferior target function is only called from
inf_ptrace_target::create_inferior.
Part of the target class hierarchy looks like this:
Every sub-class of inf_ptrace_target, except rs6000_nat_target,
implements ::post_startup_inferior. The rs6000_nat_target picks up
the implementation of ::post_startup_inferior not from
inf_ptrace_target, but from inf_child_target.
No descendent of inf_child_target, outside the inf_ptrace_target
sub-tree, implements ::post_startup_inferior, which isn't really
surprising, as they would never see the method called (remember, the
method is only called from inf_ptrace_target::create_inferior).
What I find confusing is the role inf_child_target plays in
implementing, what is really a helper function for just one of its
descendents.
In this commit I propose that we formally make ::post_startup_inferior
a helper function of inf_ptrace_target. To do this I will remove the
::post_startup_inferior from the target_ops API, and instead make this
a protected, pure virtual function on inf_ptrace_target.
I'll remove the empty implementation of ::post_startup_inferior from
the inf_child_target class, and add a new empty implementation to the
rs6000_nat_target class.
All the other descendents of inf_ptrace_target already provide an
implementation of this method and so don't need to change beyond
making the method protected within their class declarations.
To me, this makes much more sense now. The helper function, which is
only called from within the inf_ptrace_target class, is now a part of
the inf_ptrace_target class.
The only way in which this change is visible to a user is if the user
turns on 'set debug target 1'. With this debug flag on, prior to this
patch the user would see something like:
After this patch these lines are no longer present, as the
post_startup_inferior is no longer a top level target method. For me,
this is an acceptable change.
Andrew Burgess [Fri, 12 Nov 2021 12:04:08 +0000 (12:04 +0000)]
gdb: have mips_nbsd_nat_target inherit from nbsd_nat_target
While working on another patch I had reason to look at
mips-netbsd-nat.c, and noticed that the class mips_nbsd_nat_target
inherits directly from inf_ptrace_target.
This is a little strange as alpha_bsd_nat_target,
arm_netbsd_nat_target, hppa_nbsd_nat_target, m68k_bsd_nat_target,
ppc_nbsd_nat_target, sh_nbsd_nat_target, and vax_bsd_nat_target all
inherit from nbsd_nat_target.
When the target tree was converted to C++, all of the above classes
inherited from inf_ptrace_target except for hppa_nbsd_nat_target,
which was the only class that originally inherited from
nbsd_nat_target.
Later on all the remaining targets (except mips) were converted to
inherit from nbsd_nat_target, these are the commits:
I could only find mailing list threads for ppc and sh in the archive ,
and unfortunately, none of the commits has any real detail that might
explain why mips was missed out, the only extra context I could find
was this message:
Which says that "proper" OS support is going to be added to
nbsd_nat_target, hence the need to inherit from that class.
My guess is that leaving mips_nbsd_nat_target unchanged was an
oversight, so, in this commit, I propose changing mips_nbsd_nat_target
to inherit from nbsd_nat_target just like all the other nbsd targets.
My motivation for this patch relates to the post_startup_inferior
target method. In a future commit I want to change how this method is
handled. Currently the mips_nbsd_nat_target will pick up the empty
implementation of inf_child_target::post_startup_inferior rather than
the version in netbsd-nat.c. This feels like a bug to me, as surely,
enabling of proc events is something that would need to be done for
all netbsd targets, regardless of architecture.
In my future patch I have a choice then, either (a) add a new, empty
implementation of post_startup_inferior to mips_nbsd_nat_target,
or (b) this commit, have mips_nbsd_nat_target inherit from
nbsd_nat_target. Option (b) seems like the right way to go, hence,
this commit.
I've done absolutely no testing for this change, not even building it,
as that would require at least an environment in which I can x-build
mips-netbsd applications, which I have no idea how to set up.
Andrew Burgess [Thu, 9 Dec 2021 17:51:10 +0000 (17:51 +0000)]
gdb: only include mips and riscv targets if building with 64-bit bfd
While testing another patch I was trying to build different
configurations of GDB, and, during one test build I ran into a
problem, I configured with `--enable-targets=all
--host=i686-w64-mingw32` and saw this error while linking GDB:
.../i686-w64-mingw32/bin/ld: mips-tdep.o: in function `mips_gdbarch_init':
.../src/gdb/mips-tdep.c:8730: undefined reference to `disassembler_options_mips'
.../i686-w64-mingw32/bin/ld: riscv-tdep.o: in function `riscv_gdbarch_init':
.../src/gdb/riscv-tdep.c:3851: undefined reference to `disassembler_options_riscv'
So the `disassembler_options_mips` and `disassembler_options_riscv`
symbols are missing.
This turns out to be because mips-dis.c and riscv-dis.c, in which
these symbols are defined, are in the TARGET64_LIBOPCODES_CFILES list
in opcodes/Makefile.am, these files are only built when we are
building with a 64-bit bfd.
If we look further, into bfd/Makefile.am, we can see that all the
files matching elf*-riscv.lo are in the BFD64_BACKENDS list, as are
the elf*-mips.lo files, and (I know because I tried), the two
disassemblers do, not surprisingly, depend on features supplied from
libbfd.
So, though we can build most of GDB just fine for riscv and mips with
a 32-bit bfd, if I understand correctly, the final GDB
executable (assuming we could get it to link) would not understand
these architectures at the bfd level, nor would there be any
disassembler available. This sounds like a pretty useless GDB to me.
So, in this commit, I move the riscv and mips targets into GDB's list
of targets that require a 64-bit bfd. After this I can build GDB just
fine with the configure options given above.
This was discussed on the mailing list in a couple of threads:
and it is agreed, that it is unfortunate that the 32-bit riscv and
32-bit mips targets require a 64-bit bfd. If in the future this
situation ever changes then it would be expected that some (or all) of
this patch would be reverted. Until then though, this patch allows
GDB to build when configured with --enable-targets=all, and when using
a 32-bit libbfd.
Tom Tromey [Tue, 26 Oct 2021 00:20:24 +0000 (18:20 -0600)]
C++-ify path substitution code
I found some uses of xfree in the path substitution code in source.c.
C++-ifying struct substitute_path_rule both simplifies the code and
removes manual memory management.
Jan Vrany [Fri, 10 Dec 2021 11:43:20 +0000 (11:43 +0000)]
gdb/testsuite: respect GDBSERVER variable in remote-stdio-gdbserver "board"
The comment on top of gdb/testsuite/boards/remote-stdio-gdbserver.exp says
that user can specify path to gdbserver on remote system by setting
GDBSERVER variable. However, this variable was ignored and /usr/bin/gdbserver
was used unconditionally.
This commit updates the code to respect GDBSERVER if set and fall back to
/usr/bin/gdbserver if not.
Alan Modra [Wed, 8 Dec 2021 09:34:25 +0000 (20:04 +1030)]
Set sh_link for reloc sections created as normal sections
binutils-all/strip-13 and binutils-all/strip-14 tests create
SHT_REL/SHT_RELA sections by hand. These don't have sh_link set to
the .symtab section as they should, leading to readelf warnings if you
happen to be looking at the object files.
* elf.c (assign_section_numbers): Formatting. Set sh_link for
reloc sections created as normal sections in relocatable
objects.
I suppose this code was copied from GDBserver and this ifndef was left
there. As far as I know, IN_PROCESS_AGENT will never be defined when
building this file, so we can remove this.
CXX microblaze-tdep.o
/home/simark/src/binutils-gdb/gdb/microblaze-tdep.c:207:7: error: variable 'flags' set but not used [-Werror,-Wunused-but-set-variable]
int flags = 0;
^
CXX csky-tdep.o
/home/simark/src/binutils-gdb/gdb/csky-tdep.c:332:7: error: variable 'need_dummy_stack' set but not used [-Werror,-Wunused-but-set-variable]
int need_dummy_stack = 0;
^
/home/simark/src/binutils-gdb/gdb/csky-tdep.c:805:12: error: variable 'offset' set but not used [-Werror,-Wunused-but-set-variable]
int offset = 0;
^
Simon Marchi [Wed, 8 Dec 2021 02:56:48 +0000 (21:56 -0500)]
gdb/testsuite: fix default behavior of runto
The documented behavior of proc runto is to not emit a PASS when
succeeding to to run to the specified location, but emit a FAIL when
failing to do so. I suppose the intent is that it won't pollute the
results normally passing tests (although I don't see why we would care),
but make visible any problems.
However, it seems like the implementation makes it default to never
print anything. "no-message" is prependend to "args", so if "message"
is not passed, we will always take the path that sets print_fail to 0,
which will silence any failure.
This unfortunately means that tests relying on runto_main won't emit a
FAIL if failing to run to main. And since commit 4dfef5be6812
("gdb/testsuite: make runto_main not pass no-message to runto"), tests
don't emit a FAIL themselves when failing to run to main. This means
that tests failing to run to main just fail silently, and that's bad.
This can be reproduced by hacking gdb.base/template.exp like so:
diff --git a/gdb/testsuite/gdb.base/template.c b/gdb/testsuite/gdb.base/template.c
index bcf39c377d92..052be5b79d73 100644
--- a/gdb/testsuite/gdb.base/template.c
+++ b/gdb/testsuite/gdb.base/template.c
@@ -15,6 +15,14 @@
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>. */
+#include <stdlib.h>
+
+__attribute__((constructor))
+static void c (void)
+{
+ exit (1);
+}
+
int
main (void)
{
Running the modified gdb.base/template.exp shows that it exits without
printing any result.
Remove the line that prepends no-message to args, that should make
runto's behavior match its documentation.
This patch will appear to add many failures, but in reality they already
existed, they were just silenced.
On ELFv1, the _start symbol must point to the *function descriptor* (in
the .opd section), not to the function code (in the .text section) like
with ELFv2 and other architectures.
Tom de Vries [Thu, 9 Dec 2021 10:15:34 +0000 (11:15 +0100)]
[gdb/testsuite] Fix gdb.base/maint.exp with -readnow
With test-case gdb.base/maint.exp and target board -readnow, I run into:
...
FAIL: gdb.base/maint.exp: maint info line-table w/o a file name
...
The problem is that this and other regexps anchored using '^':
...
-re "^$gdb_prompt $" {
...
don't trigger because other regexps don't consume the entire preceding line.
This is partly due to the addition of the IS-STMT column.
Fix this by making the regexps consume entire lines.
Tested on x86_64-linux with native and target board readnow, as well as
check-read1 and check-readmore.
Tom de Vries [Thu, 9 Dec 2021 10:15:34 +0000 (11:15 +0100)]
[gdb/testsuite] Fix gdb.base/include-main.exp with -readnow
With test-case gdb.base/include-main.exp and target board readnow, I run into:
...
FAIL: gdb.base/include-main.exp: maint info symtab
...
The corresponding check in gdb.base/include-main.exp:
...
gdb_test_no_output "maint info symtab"
...
checks that no CU was expanded, while -readnow ensures that all CUs are
expanded.
Fix this by skipping the check with -readnow.
Tested on x86_64-linux, with native and target board readnow.
* Don't allow to add or remove the base extensions in the .option arch
directive. Instead, users should reset the whole architecture string
while they want to change the base extension.
* The operator "+" won't update the version of extension, if the
extension is already in the subset list.
bfd/
* elfxx-riscv.c (riscv_add_subset): Don't update the version
if the extension is already in the subset list.
(riscv_update_subset): To be consistent with -march option,
removed the "=" operator when user want to reset the whole
architecture string. Besides, Don't allow to add or remove
the base extensions in the .option arch directive.
gas/
* testsuite/gas/riscv/option-arch-01.s: Updated since we cannot
add or remove the base extensions in the .option arch directive.
* testsuite/gas/riscv/option-arch-02.s: Likewise.
* testsuite/gas/riscv/option-arch-fail.l: Likewise.
* testsuite/gas/riscv/option-arch-fail.s: Likewise.
* testsuite/gas/riscv/option-arch-01a.d: Set -misa-spec=2.2.
* testsuite/gas/riscv/option-arch-01b.d: Likewise.
* testsuite/gas/riscv/option-arch-02.d: Updated since the .option
arch, + won't change the version of extension, if the extension is
already in the subset list.
* testsuite/gas/riscv/option-arch-03.s: Removed the "=" operator
when resetting the whole architecture string.
Mike Frysinger [Thu, 9 Dec 2021 06:40:28 +0000 (01:40 -0500)]
sim: use ## for automake comments
The ## marker tells automake to not include the comment in its
generated output, so use that in most places where the comment
only makes sense in the inputs.
Simon Marchi [Wed, 1 Dec 2021 14:40:03 +0000 (09:40 -0500)]
gdb, gdbserver: detach fork child when detaching from fork parent
While working with pending fork events, I wondered what would happen if
the user detached an inferior while a thread of that inferior had a
pending fork event. What happens with the fork child, which is
ptrace-attached by the GDB process (or by GDBserver), but not known to
the core? Sure enough, neither the core of GDB or the target detach the
child process, so GDB (or GDBserver) just stays ptrace-attached to the
process. The result is that the fork child process is stuck, while you
would expect it to be detached and run.
Make GDBserver detach of fork children it knows about. That is done in
the generic handle_detach function. Since a process_info already exists
for the child, we can simply call detach_inferior on it.
GDB-side, make the linux-nat and remote targets detach of fork children
known because of pending fork events. These pending fork events can be
stored in:
- thread_info::pending_waitstatus, if the core has consumed the event
but then saved it for later (for example, because it got the event
while stopping all threads, to present an all-stop stop on top of a
non-stop target)
- thread_info::pending_follow: if we ran to a "catch fork" and we
detach at that moment
Additionally, pending fork events can be in target-specific fields:
- For linux-nat, they can be in lwp_info::status and
lwp_info::waitstatus.
- For the remote target, they could be stored as pending stop replies,
saved in `remote_state::notif_state::pending_event`, if not
acknowledged yet, or in `remote_state::stop_reply_queue`, if
acknowledged. I followed the model of remove_new_fork_children for
this: call remote_notif_get_pending_events to process /
acknowledge any unacknowledged notification, then look through
stop_reply_queue.
Update the gdb.threads/pending-fork-event.exp test (and rename it to
gdb.threads/pending-fork-event-detach.exp) to try to detach the process
while it is stopped with a pending fork event. In order to verify that
the fork child process is correctly detached and resumes execution
outside of GDB's control, make that process create a file in the test
output directory, and make the test wait $timeout seconds for that file
to appear (it happens instantly if everything goes well).
This test catches a bug in linux-nat.c, also reported as PR 28512
("waitstatus.h:300: internal-error: gdb_signal target_waitstatus::sig()
const: Assertion `m_kind == TARGET_WAITKIND_STOPPED || m_kind ==
TARGET_WAITKIND_SIGNALLED' failed.). When detaching a thread with a
pending event, get_detach_signal unconditionally fetches the signal
stored in the waitstatus (`tp->pending_waitstatus ().sig ()`). However,
that is only valid if the pending event is of type
TARGET_WAITKIND_STOPPED, and this is now enforced using assertions (iit
would also be valid for TARGET_WAITKIND_SIGNALLED, but that would mean
the thread does not exist anymore, so we wouldn't be detaching it). Add
a condition in get_detach_signal to access the signal number only if the
wait status is of kind TARGET_WAITKIND_STOPPED, and use GDB_SIGNAL_0
instead (since the thread was not stopped with a signal to begin with).
Add another test, gdb.threads/pending-fork-event-ns.exp, specifically to
verify that we consider events in pending stop replies in the remote
target. This test has many threads constantly forking, and we detach
from the program while the program is executing. That gives us some
chance that we detach while a fork stop reply is stored in the remote
target. To verify that we correctly detach all fork children, we ask
the parent to exit by sending it a SIGUSR1 signal and have it write a
file to the filesystem before exiting. Because the parent's main thread
joins the forking threads, and the forking threads wait for their fork
children to exit, if some fork child is not detach by GDB, the parent
will not write the file, and the test will time out. If I remove the
new remote_detach_pid calls in remote.c, the test fails eventually if I
run it in a loop.
There is a known limitation: we don't remove breakpoints from the
children before detaching it. So the children, could hit a trap
instruction after being detached and crash. I know this is wrong, and
it should be fixed, but I would like to handle that later. The current
patch doesn't fix everything, but it's a step in the right direction.
Simon Marchi [Wed, 1 Dec 2021 14:40:03 +0000 (09:40 -0500)]
gdb: move clearing of tp->pending_follow to follow_fork_inferior
A following patch will change targets so that when they detach an
inferior, they also detach any pending fork children this inferior may
have. While doing this, I hit a case where we couldn't differentiate
two cases, where in one we should detach the fork detach but not in the
other.
Suppose we continue past a fork with "follow-fork-mode == child" &&
"detach-on-fork on". follow_fork_inferior calls target_detach to detach
the parent. In that case the target should not detach the fork
child, as we'll continue debugging the child. As of now, the
tp->pending_follow field of the thread who called fork still contains
the details about the fork.
Then, suppose we run to a fork catchpoint and the user types "detach".
In that case, the target should detach the fork child in addition to the
parent. In that case as well, the tp->pending_follow field contains
the details about the fork.
To allow targets to differentiate the two cases, clear
tp->pending_follow a bit earlier, when following a fork. Targets will
then see that tp->pending_follow contains TARGET_WAITKIND_SPURIOUS, and
won't detach the fork child.
As of this patch, no behavior changes are expected.
Simon Marchi [Wed, 1 Dec 2021 14:40:02 +0000 (09:40 -0500)]
gdb/remote.c: refactor pending fork status functions
In preparation for a following patch, refactor a few things that I did
find a bit awkward, and to make them a bit more reusable.
- Pass an inferior to kill_new_fork_children instead of a pid. That
allows iterating on only this inferior's threads and avoid further
filtering on the thread's pid.
- Change thread_pending_fork_status to return a non-nullptr value only
if the thread does have a pending fork status.
- Remove is_pending_fork_parent_thread, as one can just use
thread_pending_fork_status and check for nullptr.
- Replace is_pending_fork_parent with is_fork_status, which just
returns if the given target_waitkind if a fork or a vfork. Push
filtering on the pid to the callers, when it is necessary.
Simon Marchi [Wed, 1 Dec 2021 14:40:02 +0000 (09:40 -0500)]
gdb/remote.c: move some things up
Move the stop_reply and a few functions up. Some code above them in the
file will need to use them in a following patch. No behavior changes
expected here.
Simon Marchi [Wed, 1 Dec 2021 14:40:02 +0000 (09:40 -0500)]
gdb/linux-nat: factor ptrace-detach code to new detach_one_pid function
The following patch will add some code paths that need to ptrace-detach
a given PID. Factor out the code that does this and put it in its own
function, so that it can be re-used.
Simon Marchi [Wed, 1 Dec 2021 14:40:02 +0000 (09:40 -0500)]
gdbserver: hide fork child threads from GDB
This patch aims at fixing a bug where an inferior is unexpectedly
created when a fork happens at the same time as another event, and that
other event is reported to GDB first (and the fork event stays pending
in GDBserver). This happens for example when we step a thread and
another thread forks at the same time. The bug looks like (if I
reproduce the included test by hand):
(gdb) show detach-on-fork
Whether gdb will detach the child of a fork is on.
(gdb) show follow-fork-mode
Debugger response to a program call of fork or vfork is "parent".
(gdb) si
[New inferior 2]
Reading /home/simark/build/binutils-gdb/gdb/testsuite/outputs/gdb.threads/step-while-fork-in-other-thread/step-while-fork-in-other-thread from remote target...
Reading /home/simark/build/binutils-gdb/gdb/testsuite/outputs/gdb.threads/step-while-fork-in-other-thread/step-while-fork-in-other-thread from remote target...
Reading symbols from target:/home/simark/build/binutils-gdb/gdb/testsuite/outputs/gdb.threads/step-while-fork-in-other-thread/step-while-fork-in-other-thread...
[New Thread 965190.965190]
[Switching to Thread 965190.965190]
Remote 'g' packet reply is too long (expected 560 bytes, got 816 bytes): ... <long series of bytes>
The sequence of events leading to the problem is:
- We are using the all-stop user-visible mode as well as the
synchronous / all-stop variant of the remote protocol
- We have two threads, thread A that we single-step and thread B that
calls fork at the same time
- GDBserver's linux_process_target::wait pulls the "single step
complete SIGTRAP" and the "fork" events from the kernel. It
arbitrarily choses one event to report, it happens to be the
single-step SIGTRAP. The fork stays pending in the thread_info.
- GDBserver send that SIGTRAP as a stop reply to GDB
- While in stop_all_threads, GDB calls update_thread_list, which ends
up querying the remote thread list using qXfer:threads:read.
- In the reply, GDBserver includes the fork child created as a result
of thread B's fork.
- GDB-side, the remote target sees the new PID, calls
remote_notice_new_inferior, which ends up unexpectedly creating a new
inferior, and things go downhill from there.
The problem here is that as long as GDB did not process the fork event,
it should pretend the fork child does not exist. Ultimately, this event
will be reported, we'll go through follow_fork, and that process will be
detached.
The remote target (GDB-side), has some code to remove from the reported
thread list the threads that are the result of forks not processed by
GDB yet. But that only works for fork events that have made their way
to the remote target (GDB-side), but haven't been consumed by the core
yet, so are still lingering as pending stop replies in the remote target
(see remove_new_fork_children in remote.c). But in our case, the fork
event hasn't made its way to the GDB-side remote target. We need to
implement the same kind of logic GDBserver-side: if there exists a
thread / inferior that is the result of a fork event GDBserver hasn't
reported yet, it should exclude that thread / inferior from the reported
thread list.
This was actually discussed a while ago, but not implemented AFAIK:
Implementation details-wise, the fix for this is all in GDBserver. The
Linux layer of GDBserver already tracks unreported fork parent / child
relationships using the lwp_info::fork_relative, in order to avoid
wildcard actions resuming fork childs unknown to GDB. This information
needs to be made available to the handle_qxfer_threads_worker function,
so it can filter the reported threads. Add a new thread_pending_parent
target function that allows the Linux target to return the parent of an
eventual fork child.
Testing-wise, the test replicates pretty-much the sequence of events
shown above. The setup of the test makes it such that the main thread
is about to fork. We stepi the other thread, so that the step completes
very quickly, in a single event. Meanwhile, the main thread is resumed,
so very likely has time to call fork. This means that the bug may not
reproduce every time (if the main thread does not have time to call
fork), but it will reproduce more often than not. The test fails
without the fix applied on the native-gdbserver and
native-extended-gdbserver boards.
At some point I suspected that which thread called fork and which thread
did the step influenced the order in which the events were reported, and
therefore the reproducibility of the bug. So I made the test try both
combinations: main thread forks while other thread steps, and vice
versa. I'm not sure this is still necessary, but I left it there
anyway. It doesn't hurt to test a few more combinations.
Tom Tromey [Fri, 3 Dec 2021 21:45:37 +0000 (14:45 -0700)]
Use for-each more in gdb
There are some loops in gdb that use ARRAY_SIZE (or a wordier
equivalent) to loop over a static array. This patch changes some of
these to use foreach instead.
Tom Tromey [Wed, 8 Dec 2021 19:52:34 +0000 (12:52 -0700)]
Fix error in file_and_directory patch
In my earlier C++-ization patch for file_and_directory, I introduced
an error:
- if (strcmp (fnd.name, "<unknown>") != 0)
+ if (fnd.is_unknown ())
This change inverted the sense of the test, which causes failures with
.debug_names.
This patch fixes the bug. Regression tested on x86-64 Fedora 34. I
also tested it using the AdaCore internal test suite, with
.debug_names -- this was failing before, and now it works.
Andrew Burgess [Wed, 17 Nov 2021 11:55:37 +0000 (11:55 +0000)]
gdb/python: Use tp_init instead of tp_new to setup gdb.Value
The documentation suggests that we implement gdb.Value.__init__,
however, this is not currently true, we really implement
gdb.Value.__new__. This will cause confusion if a user tries to
sub-class gdb.Value. They might write:
class MyVal (gdb.Value):
def __init__ (self, val):
gdb.Value.__init__(self, val)
obj = MyVal(123)
print ("Got: %s" % obj)
But, when they source this code they'll see:
(gdb) source ~/tmp/value-test.py
Traceback (most recent call last):
File "/home/andrew/tmp/value-test.py", line 7, in <module>
obj = MyVal(123)
File "/home/andrew/tmp/value-test.py", line 5, in __init__
gdb.Value.__init__(self, val)
TypeError: object.__init__() takes exactly one argument (the instance to initialize)
(gdb)
The reason for this is that, as we don't implement __init__ for
gdb.Value, Python ends up calling object.__init__ instead, which
doesn't expect any arguments.
The Python docs suggest that the reason why we might take this
approach is because we want gdb.Value to be immutable:
But I don't see any reason why we should require gdb.Value to be
immutable when other types defined in GDB are not. This current
immutability can be seen in this code:
In this commit I propose that we switch to using __init__ to
initialize gdb.Value objects.
This does introduce some additional complexity, during the __init__
call a gdb.Value might already be associated with a gdb value object,
in which case we need to cleanly break that association before
installing the new gdb value object. However, the cost of doing this
is not great, and the benefit - being able to easily sub-class
gdb.Value seems worth it.
After this commit the first example above works without error, while
the second example now prints:
Got: 1234
Got: 5678
In order to make it easier to override the gdb.Value.__init__ method,
I have tweaked the definition of gdb.Value.__init__. The second,
optional argument to __init__ is a gdb.Type, if this argument is not
present then GDB figures out a suitable type.
However, if we want to override the __init__ method in a sub-class,
and still support the default argument, it is easier to write:
Currently, passing None for the Type will result in an error:
TypeError: type argument must be a gdb.Type.
After this commit I now allow the type argument to be None, in which
case GDB figures out a suitable type just as if the type had not been
passed at all.
Unless a user is trying to reinitialize a value, or create sub-classes
of gdb.Value, there should be no user visible changes after this
commit.
This is interesting because it shows a case where the libopcodes
disassembler is returning -1 without first calling the
memory_error_func callback. Indeed, the return from libopcodes
happens from this code snippet in i386-dis.c in the print_insn
function:
if (address_mode == mode_64bit && sizeof (bfd_vma) < 8)
{
(*info->fprintf_func) (info->stream,
_("64-bit address is disabled"));
return -1;
}
Notice how, prior to the return the disassembler tries to print a
helpful message out, but GDB doesn't print this message.
The reason this message goes missing is the call stack, it looks like
this:
When i386-dis.c:print_insn returns -1 this is handled in
gdb_disassembler::print_insn, where an exception is thrown. However,
the actual printing of the disassembler output is done in
gdb_pretty_print_disassembler::pretty_print_insn, and is only done if
an exception is not thrown.
In this commit I change this. The pretty_print_insn now uses
try/catch around the call to gdb_disassembler::print_insn, if we catch
an error then we first print any pending output in the instruction
buffer, before rethrowing the exception. As a result, even if an
exception is thrown we still print any pending disassembler output to
the screen; in the above case the helpful message will now be shown.
Before my patch we might expect to see this output:
(gdb) disassemble 0x0,+4
Dump of assembler code from 0x0 to 0x4:
0x0000000000000000: unknown disassembler error (error = -1)
(gdb)
But now we see this:
(gdb) disassemble 0x0,+4
Dump of assembler code from 0x0 to 0x4:
0x0000000000000000: 64-bit address is disabled
unknown disassembler error (error = -1)
If the disassembler returns -1 without printing a helpful message then
we would still expect a change in output, something like:
(gdb) disassemble 0x0,+4
Dump of assembler code from 0x0 to 0x4:
0x0000000000000000:
unknown disassembler error (error = -1)
Which I think is still acceptable, though at this point I think a
strong case can be made that this is a disassembler bug (not printing
anything, but still returning -1).
Notice however, that the error message is always printed on a new line
now. This is also true for the memory error case, where before we
might see this:
(gdb) disassemble 0x0,+4
Dump of assembler code from 0x0 to 0x4:
0x00000000: Cannot access memory at address 0x0
We now get this:
(gdb) disassemble 0x0,+4
Dump of assembler code from 0x0 to 0x4:
0x00000000:
Cannot access memory at address 0x0
For me, I'm happy to accept this change, having the error on a line by
itself, rather than just appended to the end of the previous line,
seems like an improvement, but I'm aware others might feel
differently, so I'd appreciate any feedback.
Jan Vrany [Wed, 8 Dec 2021 10:46:49 +0000 (10:46 +0000)]
ppc: recognize all program traps
Permanent program breakpoints (ones inserted into the code) other than
the one GDB uses for POWER (0x7fe00008) did not result in stop but
caused GDB to loop infinitely.
This was because GDB did not recognize trap instructions other than
"trap". For example, "tw 12, 4, 4" was not be recognized, causing GDB
to loop forever.
This commit fixes this by providing POWER specific hook
(gdbarch_program_breakpoint_here_p) recognizing all tw, twi, td and tdi
instructions.
Tested on Linux on PowerPC e500 and on QEMU PPC64le.
Jan Vrany [Wed, 8 Dec 2021 10:46:49 +0000 (10:46 +0000)]
ppc: use "trap" ("tw, 31, 0, 0") as breakpoint instruction
Power ISA 3.0 B spec [1], sections 3.3.11 "Fixed-Point Trap Instructions"
and section C.6 "Trap Mnemonics" specify "tw, 31, 0, 0" (encoded as
0x7fe00008) as canonical unconditional trap instruction.
This commit changes the breakpoint instruction used by GDB from
"tw 12, r2, r2" to unconditional "trap".
Simon Marchi [Fri, 3 Dec 2021 21:35:38 +0000 (16:35 -0500)]
gdb: make struct linespect contain vectors, not pointers to vectors
struct linespec contains pointers to vectors, instead of containing
vectors directly. This is probably historical, when linespec_parser
(which contains a struct linespec field) was not C++-ified yet. But it
seems easy to change the pointers to vectors to just vectors today.
This simplifies the code, we don't need to manually allocate and delete
the vectors and there's no pointer that can be NULL.
As far as I understand, there was not meaningful distinction between a
NULL pointer to vector and an empty vector. So all NULL checks are
changed for !empty checks.
Simon Marchi [Fri, 3 Dec 2021 21:35:37 +0000 (16:35 -0500)]
gdb/linespec.c: simplify condition
We can remove the empty check: if the vector has size 1, it is obviously
not empty. This code ended up like this because the empty check used to
be a NULL check.
Simon Marchi [Tue, 7 Dec 2021 20:56:25 +0000 (15:56 -0500)]
gdb: rename "maint agent" functions
Functions agent_eval_command and agent_command are used to implement
maintenance commands, rename them accordingly (with the maint_ prefix),
as well as the agent_command_1 helper function.
John Baldwin [Tue, 7 Dec 2021 18:29:01 +0000 (10:29 -0800)]
Support AT_FXRNG and AT_KPRELOAD on FreeBSD.
FreeBSD's kernel has recently added two new ELF auxiliary vector
entries. AT_FXRNG points to a root seed version for the kernel's
PRNG. Userland can use this to reseed a userland PRNG after the
kernel's PRNG has reseeded. AT_KPRELOAD is the base address of a
kernel-provided vDSO.
This change displays the proper name and description of these entries
in 'info auxv'.
Tom Tromey [Mon, 6 Dec 2021 19:20:28 +0000 (12:20 -0700)]
Avoid extra work in global_symbol_searcher::expand_symtabs
I noticed that global_symbol_searcher::expand_symtabs always passes a
file matcher to expand_symtabs_matching. However, if 'filenames' is
empty, then this always returns true. It's slightly more efficient to
pass a null file matcher in this case, because that lets the "quick"
symbol implementations skip any filename checks.
Tom de Vries [Tue, 7 Dec 2021 08:44:36 +0000 (09:44 +0100)]
[gdb/testsuite] Fix options arg handling in compile_jit_elf_main_as_so
In commit 80ad340c902 ("[gdb/testsuite] use -Ttext-segment for jit-elf tests")
the following change was made:
...
proc compile_jit_elf_main_as_so {main_solib_srcfile main_solib_binfile options} {
- set options [concat $options debug]
+ global jit_load_address jit_load_increment
+
+ set options [list \
+ additional_flags="-DMAIN=jit_dl_main" \
+ additional_flags=-DLOAD_ADDRESS=$jit_load_address \
+ additional_flags=-DLOAD_INCREMENT=$jit_load_increment \
+ debug]
...
Before the change, the options argument was used, but after the change not
anymore.
Fix this by reverting back to using "set options [concat $options ...]".
Fixing this gets us twice the -DMAIN=jit_dl_main bit, once from a caller, and
once from compile_jit_elf_main_as_so. Fix this by removing the bit from
compile_jit_elf_main_as_so, which makes the code similar to compile_jit_main.
Tom de Vries [Tue, 7 Dec 2021 07:16:42 +0000 (08:16 +0100)]
[gdb/testsuite] Fix FAIL in gdb.tui/basic.exp
On openSUSE Leap 15.2 aarch64 I ran into:
...
FAIL: gdb.tui/basic.exp: check main is where we expect on the screen
...
while this is passing on x86_64.
On x86_64-linux we have at the initial screen dump for "list -q main":
...
0 +-/home/vries/gdb_versions/devel/src/gdb/testsuite/gdb.tui/tui-layout.c--+
1 | 15 You should have received a copy of the GNU General Public |
2 | 16 along with this program. If not, see <http://www.gnu.org/|
3 | 17 |
4 | 18 int |
5 | 19 main () |
6 | 20 { |
7 | 21 return 0; |
8 | 22 } |
9 | 23 |
...
but on aarch64:
...
0 +-/home/tdevries/gdb/src/gdb/testsuite/gdb.tui/tui-layout.c--------------+
1 | 16 along with this program. If not, see <http://www.gnu.org/|
2 | 17 |
3 | 18 int |
4 | 19 main () |
5 | 20 { |
6 | 21 return 0; |
7 | 22 } |
8 | 23 |
9 | 24 |
...
The cause of the diffferent placement is that we have as line number for main
on x86_64:
...
$ gdb -q -batch outputs/gdb.tui/basic/basic -ex "info line main"
Line 20 of "tui-layout.c" starts at address 0x4004a7 <main> \
and ends at 0x4004ab <main+4>.
...
and on aarch64 instead:
...
$ gdb -q -batch outputs/gdb.tui/basic/basic -ex "info line main"
Line 21 of "tui-layout.c" starts at address 0x4005f4 <main> \
and ends at 0x4005f8 <main+4>.
...
Fix this by using a new source file main-one-line.c, that implements the
entire main function on a single line, in order to force the compiler to use
that line number.
Tom de Vries [Tue, 7 Dec 2021 07:07:18 +0000 (08:07 +0100)]
[gdb/tdep] Fix inferior plt calls in PIE for i386
Consider test-case test.c:
...
int main (void) {
void *p = malloc (10);
return 0;
}
...
When compiled to a non-PIE exec:
...
$ gcc -m32 test.c
...
the call sequence looks like:
... 8048447: 83 ec 0c sub $0xc,%esp 804844a: 6a 0a push $0xa 804844c: e8 bf fe ff ff call 8048310 <malloc@plt>
...
which calls to:
... 08048310 <malloc@plt>: 8048310: ff 25 0c a0 04 08 jmp *0x804a00c 8048316: 68 00 00 00 00 push $0x0 804831b: e9 e0 ff ff ff jmp 8048300 <.plt>
...
where the first insn at 0x8048310 initially jumps to the following address
0x8048316, read from the .got.plt @ 0x804a00c:
... 804a0000c9f0408000000000000000016830408 ................ 804a01026830408 &...
...
Likewise, when compiled as a PIE:
...
$ gcc -m32 -fPIE -pie test.c
...
we have this call sequence (with %ebx setup to point to the .got.plt):
... 0000055d <main>:
579: 83 ec 0c sub $0xc,%esp
57c: 6a 0a push $0xa
57e: 89 c3 mov %eax,%ebx
580: e8 6b fe ff ff call 3f0 <malloc@plt>
...
which calls to:
... 000003f0 <malloc@plt>:
3f0: ff a3 0c 00 00 00 jmp *0xc(%ebx)
3f6: 68 00 00 00 00 push $0x0
3fb: e9 e0 ff ff ff jmp 3e0 <.plt>
...
where the insn at 0x3f0 initially jumps to following address 0x3f6, read from
the .got.plt at offset 0xc:
...
2000 f41e00000000000000000000f6030000 ................
2010 06040000 ....
...
When instead doing an inferior call to malloc (with nosharedlib to force
malloc to resolve to malloc@plt rather than the functions in ld.so or libc.so)
with the non-PIE exec, we have the expected:
...
$ gdb -q -batch a.out -ex start -ex nosharedlib -ex "p /x (void *)malloc (10)"
Temporary breakpoint 1 at 0x8048444
Temporary breakpoint 1, 0x08048444 in main ()
$1 = 0x804b160
...
But with the PIE exec, we run into:
...
$ gdb -q -batch a.out -ex start -ex nosharedlib -ex "p /x (void *)malloc (10)"
Temporary breakpoint 1 at 0x56c
Temporary breakpoint 1, 0x5655556c in main ()
Program received signal SIGSEGV, Segmentation fault.
0x565553f0 in malloc@plt ()
...
The segfault happens because:
- the inferior call mechanism doesn't setup %ebx
- %ebx instead is 0
- the jump to "*0xc(%ebx)" reads from memory at 0xc
Fix this by setting up %ebx properly in i386_thiscall_push_dummy_call.
Fixes this failure with target board unix/-m32/-pie/-fPIE reported in
PR28467:
...
FAIL: gdb.base/nodebug.exp: p/c (int) array_index("abcdef",2)
...
Tested on x86_64-linux, with target board unix/-m32 and unix/-m32/-fPIE/-pie.
Tom de Vries [Tue, 7 Dec 2021 06:51:25 +0000 (07:51 +0100)]
[gdb/symtab] Support -readnow during reread
When running test-case gdb.base/cached-source-file.exp with target board
readnow, we run into:
...
FAIL: gdb.base/cached-source-file.exp: rerun program (the program exited)
...
The problem is that when rereading, the readnow is ignored.
Fix this by copying the readnow handling code from symbol_file_add_with_addrs
to reread_symbols.
Tom de Vries [Tue, 7 Dec 2021 06:35:10 +0000 (07:35 +0100)]
[gdb/ada] Fix assert in ada_is_unconstrained_packed_array_type
On openSUSE Leap 42.3, with system compiler gcc 4.8.5 I run into:
...
(gdb) print u_one_two_three^M
src/gdb/gdbtypes.h:1050: internal-error: field: \
Assertion `idx >= 0 && idx < num_fields ()' failed.^M
...
We run into trouble while doing this in
ada_is_unconstrained_packed_array_type:
...
1953 return TYPE_FIELD_BITSIZE (type, 0) > 0;
...
which tries to get field 0 from a type without fields:
...
(gdb) p type->num_fields ()
$6 = 0
...
which is the case because the type is a typedef:
...
(gdb) p type->code ()
$7 = TYPE_CODE_TYPEDEF
...
Fix this by using the type referenced by the typedef instead.
Alan Modra [Tue, 7 Dec 2021 02:06:31 +0000 (12:36 +1030)]
Re: Add support for AArch64 EFI (efi-*-aarch64)
Commit b69c9d41e8 was broken in multiple ways regarding the realloc
of the target string, most notably in that "-little" wasn't actually
appended to the input_target or output_target. This caused asan
errors and "FAIL: Check if efi app format is recognized". I also
noticed that the input_target string wasn't being copied but rather
the output_target when dealing with the input target. Fix that too.
PR 26206
* objcopy.c (convert_efi_target): Rewrite. Allocate modified
target strings here..
(copy_main): ..rather than here. Do handle input_target,
not output_target for input.
Carl Love [Fri, 19 Nov 2021 18:33:51 +0000 (18:33 +0000)]
gdb: Add PowerPC support to gdb.dwarf2/frame-inlined-in-outer-frame
This patch adds an #elif defined for PowerPC to setup the exit_0 macro.
This patch addes the needed macro definitionald logic to handle both elfV1
and elfV2.
The patch has been successfully tested on both PowerPC BE, Powerpc LE and
X86_64 with no regressions.
Tom de Vries [Mon, 6 Dec 2021 15:01:47 +0000 (16:01 +0100)]
[gdb/testsuite] Use precise align in gdb.arch/i386-{avx,sse}.exp
Test-cases gdb.arch/i386-{avx,sse}.exp use assembly instructions that require
the memory operands to be aligned to a certain boundary, and the test-cases
use C11's _Alignas to make that happen.
The draw-back of using _Alignas is that while it does enforce a minimum
alignment, the actual alignment may be bigger, which makes the following
scenario possible:
- copy say, gdb.arch/i386-avx.c as basis for a new test-case
- run the test-case and observe a PASS
- commit the new test-case in the supposition that the test-case is correct
and well-tested
- run later into a failure on a different test setup (which may be a setup
where reproduction and investigation is more difficult and time-consuming),
and find out that the specified alignment was incorrect and should have been
updated to say, 64 bytes. The initial PASS occurred only because the actual
alignment happened to be greater than required.
The idea of having precise alignment as a means of having more predictable
execution which allows flushing out bugs earlier, has been filed as PR
gcc/103095.
Add a new file lib/precise-aligned-alloc.c with functions
precise_aligned_alloc and precise_aligned_dup, to support precise alignment.
Use precise_aligned_dup in aforementioned test-cases to:
- verify that the specified alignment is indeed sufficient, rather
than too little but accidentally over-aligned.
- prevent the same type of problems in any new test-cases based on these
Tom de Vries [Mon, 6 Dec 2021 15:01:47 +0000 (16:01 +0100)]
[gdb/testsuite] Fix data alignment in gdb.arch/i386-{avx,sse}.exp
When running test-case gdb.arch/i386-avx.exp with clang I ran into:
...
(gdb) PASS: gdb.arch/i386-avx.exp: set first breakpoint in main
continue^M
Continuing.^M
^M
Program received signal SIGSEGV, Segmentation fault.^M
0x000000000040052b in main (argc=1, argv=0x7fffffffd3c8) at i386-avx.c:54^M
54 asm ("vmovaps 0(%0), %%ymm0\n\t"^M
(gdb) FAIL: gdb.arch/i386-avx.exp: continue to breakpoint: \
continue to first breakpoint in main
...
The problem is that the vmovaps insn requires an 256-bit (or 32-byte) aligned
address, and it's only 16-byte aligned:
...
(gdb) p /x $rax
$1 = 0x601030
...
Fix this by using a sufficiently aligned address, using _Alignas.
Alan Modra [Wed, 24 Nov 2021 13:30:50 +0000 (00:00 +1030)]
[GOLD] PowerPC64 inline plt sequences
The fixes gold failures to handle inline PLT sequences properly.
PowerPC gold was always turning these back into direct calls due to
gsym->use_plt_offset() returning false. This is fixed for dynamic
linking by correcting get_reference_flags, and for static linking by
overriding use_plt_offset() in relocate(). The rest of the patch
revolves around needing to create PLT entries for inline PLT calls
when statically linking (for gcc -mlongcall). The lplt section
handled that for local symbols, now it does globals too.
* powerpc.cc (Target_powerpc::plt_off): Return proper section
for static link.
(Target_powerpc::symval_for_branch): Make public.
(Target_powerpc::make_lplt_section): Add Symbol_table* param.
Adjust all calls.
(Target_powerpc::make_local_plt_entry): Likewise.
(Target_powerpc::make_local_plt_entry): New variant for global syms.
(Powerpc_relobj::do_relocate_sections): Don't write lplt contents.
(Output_data_plt_powerpc::do_write): Write lplt contents here.
(Output_data_plt_powerpc::Output_data_plt_powerpc): Save
symbol table pointer. Adjust all uses.
(Output_data_plt_powerpc::add_entry): Add stash parameter. Don't
do dynamic reloc handling when no reloc section. Save symbol
for local plt entries.
(Output_data_plt_powerpc::add_local_entry): Save symbol.
(Output_data_plt_powerpc::Local_plt_ent): New class.
(Output_data_plt_powerpc::sym_ents_): New vector.
(Target_powerpc::Scan::get_reference_flags): Return
FUNCTION_CALL|RELATIVE_REF for inline plt relocs.
(Target_powerpc::Scan::global): Make entries in lplt for inline
plt call relocation symbols.
(Target_powerpc::Relocate::relocate): Rename has_plt_offset to
use_plt_offset. Set use_plt_offset for inline plt relocs.
projects / thirdparty / binutils-gdb.git / log
