Tom Tromey [Mon, 13 Mar 2023 14:51:34 +0000 (08:51 -0600)]
Introduce type_allocator
This introduces a new type_allocator class. This class will be used
to abstract out the placement of new types, so that type-creation code
can be simplified and shared.
Reviewed-By: Simon Marchi <simon.marchi@efficios.com>
Tom Tromey [Thu, 16 Mar 2023 20:41:31 +0000 (14:41 -0600)]
Fix line table regression
Simon pointed out a line table regression, and after a couple of false
starts, I was able to reproduce it by hand using his instructions.
The bug is that most of the code in do_mixed_source_and_assembly uses
unrelocated addresses, but one spot does:
pc = low;
... after the text offset has been removed.
This patch fixes the problem by introducing a new type to represent
unrelocated addresses in the line table. This prevents this sort of
bug to some degree (it's still possible to manipulate a CORE_ADDR in a
bad way, this is unavoidable).
However, this did let the compiler flag a few spots in that function,
and now it's not possible to compare an unrelocated address from a
line table with an ordinary CORE_ADDR.
Regression tested on x86-64 Fedora 36, though note this setup never
reproduced the bug in the first place. I also tested it by hand on
the disasm-optim test program.
Simon Marchi [Fri, 25 Nov 2022 18:56:14 +0000 (13:56 -0500)]
gdb: introduce bp_loc_tracepoint
Since commit cb1e4e32c2d9 ("catch catch/throw/rethrow", breakpoint ->
catchpoint), this simple tracing scenario does not work:
$ gdb/gdb -nx -q --data-directory=gdb/data-directory ./test
Reading symbols from ./test...
(gdb) tar rem :1234
Remote debugging using :1234
Reading symbols from /lib64/ld-linux-x86-64.so.2...
(No debugging symbols found in /lib64/ld-linux-x86-64.so.2)
0x00007ffff7fe5730 in ?? () from /lib64/ld-linux-x86-64.so.2
(gdb) trace do_something
Tracepoint 1 at 0x555555555144: file test.c, line 5.
(gdb) tstart
(gdb) continue
Continuing.
Target returns error code '01'.
The root cause is that the bp_location::nserted flag does not transfer
anymore from an old bp_location to the new matching one. When a shared
library gets loaded, GDB computes new breakpoint locations for each
breakpoint in update_breakpoint_locations. The new locations are in the
breakpoint::loc chain, while the old locations are still in the
bp_locations global vector. Later, update_global_location_list is
called. It tries to map old locations to new locations, and if
necessary transfer some properties, like the inserted flag.
Since commit cb1e4e32c2d9, the inserted flag isn't transferred for
locations of tracepoints. This is because bl_address_is_meaningful used
to be implemented like this:
static int
breakpoint_address_is_meaningful (struct breakpoint *bpt)
{
enum bptype type = bpt->type;
return (type != bp_watchpoint && type != bp_catchpoint);
}
Because locations for tracepoints have the bp_loc_other type,
bl_address_is_meaningful started to return false for them, where it
returned true before. This made update_global_location_list skip the
part where it calls swap_insertion.
I think this can be solved by introduced a new bp_loc_tracepoint
bp_loc_type.
I don't know if it's accurate, but my understanding is that bp_loc_type
describes roughly "how do we ask the target to insert that location".
bp_loc_software_breakpoint are inserted using
target_ops::insert_breakpoint_location. bp_loc_hardware_breakpoint are
inserted using target_ops::insert_hw_breakpoint.
bp_loc_software_watchpoint and bp_loc_hardware_watchpoint are inserted
using target_ops::insert_watchpoint. For all these, the address is
meaningful, as we ask the target to insert the point at a specific
address. bp_loc_other is a catch-all for "the rest", in practice for
catchpoints that don't have a specific address (hence why
bl_address_is_meaningful returns false for them). For instance,
inserting a signal catchpoint is done by asking the target to report
that specific signal. GDB doesn't associate an address to that.
But tracepoints do have a meaningful address to thems, so they can't be
bp_loc_other, with that logic. They also can't be
bp_loc_software_breakpoint, because we don't want GDB to insert
breakpoints for them (even though they might be implemented using
software breakpoints by the remote side). So, the new bp_loc_tracepoint
type describes that the way to insert these locations is with
target_ops::download_tracepoint. It makes bl_address_is_meaningful
return true for them. And they'll be ignored by insert_bp_location and
GDB won't try to insert a memory breakpoint for them.
With this, I see a few instances of 'Target returns error code: 01'
disappearing from gdb.log, and the results of gdb.trace/*.exp improve a
little bit:
-# of expected passes 3765
+# of expected passes 3781
-# of unexpected failures 518
+# of unexpected failures 498
Carl Love [Thu, 9 Mar 2023 21:10:18 +0000 (16:10 -0500)]
PowerPC: fix for gdb.reverse/finish-precsave.exp and gdb.reverse/finish-reverse.exp
PPC64 multiple entry points, a normal entry point and an alternate entry
point. The alternate entry point is to setup the Table of Contents (TOC)
register before continuing at the normal entry point. When the TOC is
already valid, the normal entry point is used, this is typically the case.
The alternate entry point is typically referred to as the global entry
point (GEP) in IBM. The normal entry point is typically referred to as
the local entry point (LEP).
When GDB is executing the finish command in reverse, the function
finish_backward currently sets the break point at the alternate entry point.
This issue is if the function, when executing in the forward direction,
entered the function via the normal entry point, execution in the reverse
direction will never sees the break point at the alternate entry point. In
this case, the reverse execution continues until the next break point is
encountered thus stopping at the wrong place.
This patch adds a new address to struct execution_control_state to hold the
address of the alternate entry point (GEP). The finish_backwards function
is updated, if the stopping point is between the normal entry point (LEP)
and the end of the function, a breakpoint is set at the normal entry point.
If the stopping point is between the entry points, a breakpoint is set at
the alternate entry point. This ensures that GDB will always stop at the
normal entry point. If the function did enter via the alternate entry
point, GDB will detect that and continue to execute backwards in the
function until the alternate entry point is reached.
The patch fixes the behavior of the reverse-finish command on PowerPC to
match the behavior of the command on other platforms, specifically X86.
The patch does not change the behavior of the command on X86.
A new test is added to verify the reverse-finish command on PowerPC
correctly stops at the instruction where the function call is made.
The patch fixes 11 regression errors in test gdb.reverse/finish-precsave.exp
and 11 regression errors in test gdb.reverse/finish-reverse.exp.
The patch has been tested on Power 10 and X86 processor with no new
regression failures.
Carl Love [Wed, 1 Mar 2023 16:45:43 +0000 (11:45 -0500)]
Move step_until procedure
Procedure step_until from test gdb.reverse/step-indirect-call-thunk.exp
is moved to lib/gdb.exp and renamed repeat_cmd_until. The existing procedure
gdb_step_until in lib/gdb.exp is simpler variant of the new repeat_cmd_until
procedure. The existing procedure gdb_step_until is changed to just call
the new repeat_cmd_until procedure with the command set to "step" and an
optional CURRENT string. The default CURRENT string is set to "\}" to work
with the existing uses of procedure gdb_step_until.
Tom de Vries [Fri, 17 Mar 2023 18:25:18 +0000 (19:25 +0100)]
[gdb/testsuite] Fix regexp in gdb.arch/ftrace-insn-reloc.exp
With test-case gdb.arch/ftrace-insn-reloc.exp and host board
local-remote-host-notty and target board native-gdbserver I run into:
...
(gdb) info sharedlibrary^M
From To Syms Read Shared Object Library^M
$hex $hex Yes /lib64/ld-linux-x86-64.so.2^M
$hex $hex Yes /home/remote-host/libinproctrace.so^M
$hex $hex Yes /lib64/libm.so.6^M
$hex $hex Yes /lib64/libc.so.6^M
$hex $hex Yes /lib64/libdl.so.2^M
$hex $hex Yes (*) /usr/lib64/libstdc++.so.6^M
$hex $hex Yes (*) /lib64/libgcc_s.so.1^M
$hex $hex Yes /lib64/libpthread.so.0^M
(*): Shared library is missing debugging information.^M
(gdb) FAIL: gdb.arch/ftrace-insn-reloc.exp: IPA loaded
...
due to trying to match libinproctrace.so using the target path, while the
command lists it using the host path.
Tom de Vries [Fri, 17 Mar 2023 18:25:18 +0000 (19:25 +0100)]
[gdb/testsuite] Handle remote host in gdb_load_shlib
With test-case gdb.arch/ftrace-insn-reloc.exp and host board
local-remote-host-notty and target board native-gdbserver I run into:
...
(gdb) tstart^M
Target returns error code '.In-process agent library not loaded in process. \
Fast and static trace points unavailable.'.^M
(gdb) FAIL: gdb.arch/ftrace-insn-reloc.exp: start trace experiment
...
Fix this by:
- handling remote host in gdb_load_shlib, and
- moving the gdb_load_shlib to after the clean_restart, such that the
set solib-search-path can take effect.
Tom de Vries [Fri, 17 Mar 2023 18:25:18 +0000 (19:25 +0100)]
[gdb/testsuite] Handle REMOTE_HOST_USERNAME in local-remote-host
Handle REMOTE_HOST_USERNAME in local-remote-host, similar to how that's done for
REMOTE_TARGET_USERNAME in remote-gdbserver-on-localhost.
This helps to keep the home dir clean.
Since the setup makes $build/gdb/testsuite on build unreadable for the remote
host, we run into permission problems for GDB and the data-directory, so fix
this (as was done for gdbserver in gdbserver-base.exp) using file normalize.
Tom de Vries [Fri, 17 Mar 2023 15:06:39 +0000 (16:06 +0100)]
[gdb/testsuite] Set remotedir by default in some boards
When doing a gdb_simple_compile, and downloading the resulting exec $obj
to target the result $target_obj may be a relative file path, which may give
problems when trying to do:
...
remote_exec target $target_obj
...
Fix/workaround this on some target boards by setting remotedir by default, and
add a corresponding test in gdb.testsuite/board-sanity.exp.
This doesn't work for host/target board local-remote-host-native, so xfail this.
Tom de Vries [Fri, 17 Mar 2023 15:06:39 +0000 (16:06 +0100)]
[gdb/testsuite] Fix have_avx for remote target
In proc have_avx we compile some source into an exec, resulting in a file $obj
on build, and then attempt to execute it on target:
...
set result [remote_exec target $obj]
...
Fix this by using gdb_remote_download target.
Likewise in a few other procs that use "remote_exec target".
Tom de Vries [Fri, 17 Mar 2023 15:06:39 +0000 (16:06 +0100)]
[gdb/testsuite] Handle precise-aligned-alloc.c for remote host
With test-case gdb.arch/i386-sse.exp (and likewise gdb.arch/i386-avx.exp) and
host board local-remote-host-notty and target board native-gdbserver I run
into:
...
gdb compile failed, i386-sse.c:68:10: fatal error: \
../lib/precise-aligned-alloc.c: No such file or directory
#include "../lib/precise-aligned-alloc.c"
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
...
Fix this using '#include "precise-aligned-alloc.c"' and making that work with
non-remote and remote host.
Tom de Vries [Fri, 17 Mar 2023 12:29:13 +0000 (13:29 +0100)]
[gdb/testsuite] Add escape_for_host
In gdb_compile we have:
...
lappend new_options "ldflags=-Wl,-rpath,\\\$ORIGIN"
...
and we could improve readability by using {} rather than "":
...
lappend new_options {ldflags=-Wl,-rpath,\$ORIGIN}
...
But rather than manually adding escapes in a string, add a new proc
escape_for_host that care of this for us, allowing us to write:
...
lappend new_options [escape_for_host {ldflags=-Wl,-rpath,$ORIGIN}]
...
Jan Beulich [Fri, 17 Mar 2023 09:07:05 +0000 (10:07 +0100)]
gas: apply md_register_arithmetic also to unary '+'
Even a unary '+' has to be considered arithmetic; at least on x86 in
Intel Syntax mode otherwise bogus insn operands may be accepted.
Convert this specific case to binary + (i.e. 0 + <register>). (An
implication is that md_operator(,1,) would need to deal with arch-
specific equivalents of unary '+' is a similar way, if such an arch-
specific variant would be specified in the first place.)
To avoid duplicating what make_expr_symbol() does to construct a
constant-zero expression, simply make its previously local variable a
file-scope static one. This way there's also no need to invoke
clean_up_expression().
Jan Beulich [Fri, 17 Mar 2023 09:06:18 +0000 (10:06 +0100)]
gas: expose flag_macro_alternate globally
Yet again with the removal of gasp about 20 years ago this extra level
of indirection isn't necessary anymore either. Drop macro.c's local
variable and make as.c's global.
While doing the conversion, switch the variable to "bool".
Jan Beulich [Fri, 17 Mar 2023 09:05:57 +0000 (10:05 +0100)]
gas: use flag_mri directly in macro processing
Again with the removal of gasp about 20 years ago the extra level of
indirection isn't necessary anymore. Drop macro.c's local variable and
use the global flag directly.
Jan Beulich [Fri, 17 Mar 2023 09:05:32 +0000 (10:05 +0100)]
gas: isolate macro_strip_at to macro.c
This removes a leftover from i960 support; with that nothing is left
which would set macro_strip_at to non-zero, so the variable is converted
to a #define (retaining the logic in case a new user would appear) and
macro_init()'s respective parameter is dropped.
Jan Beulich [Fri, 17 Mar 2023 09:05:05 +0000 (10:05 +0100)]
gas: drop function pointer parameter from macro_init()
With the removal of gasp (about 20 years ago) the need for this kind-
of-hook has disappeared. Go a step beyond merely moving the to be called
function: Inline its contents right at the sole call site.
Tom de Vries [Fri, 17 Mar 2023 09:01:07 +0000 (10:01 +0100)]
[gdb/testsuite] Fix filename in gdb.debuginfod/crc_mismatch.exp
After running test-case gdb.debuginfod/crc_mismatch.exp, I find a dir called '$':
...
$ ls $build/gdb/testsuite/
$ config.log gdb.log lib outputs site.exp
cache config.status gdb.sum Makefile site.bak temp
...
Fix this by removing the stray '$' here:
...
set debugfile "$[standard_output_file ${testfile}.debug]"
...
Andrew Burgess [Mon, 16 Jan 2023 17:38:10 +0000 (17:38 +0000)]
gdb/doc: extended documentation for inferior function calls
I noticed that the documentation for inferior function calls doesn't
say much about what happens if/when an inferior function call is
interrupted, i.e. it doesn't describe what the dummy frame looks like
on the stack, or how GDB behaves when the inferior is continued and
reaches the dummy frame.
This commit aims to add some of this missing information.
Tom Tromey [Thu, 16 Mar 2023 13:48:22 +0000 (07:48 -0600)]
Fix build breakage in rs6000-aix-tdep.c
A recent change to rs6000-aix-tdep.c broke the build. This patch
fixes it by declaring a few target descriptions in ppc-tdep.h and then
not including the various features .c files in rs6000-aix-tdep.c.
Hui Li [Sat, 4 Mar 2023 00:45:28 +0000 (08:45 +0800)]
gdb/testsuite: Add support for LoongArch in gdb.base/float.exp
The test results on LoongArch as follows:
Without this patch:
```
$ make check-gdb TESTS="gdb.base/float.exp"
=== gdb Summary ===
# of expected passes 2
# of unexpected failures 1
```
With this patch:
```
$ make check-gdb TESTS="gdb.base/float.exp"
=== gdb Summary ===
# of expected passes 3
```
Signed-off-by: Hui Li <lihui@loongson.cn> Reviewed-By: Tom Tromey <tom@tromey.com> Approved-By: Simon Marchi <simon.marchi@efficios.com> Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Andrew Burgess [Wed, 8 Mar 2023 15:26:32 +0000 (15:26 +0000)]
gdb/doc: spring clean the Python unwinders documentation
The documentation for the Python Unwinders API could do with some
improvement. The 'Unwinder Skeleton Code' has an error: it says
'unwinders' when it should say 'unwinder' in one case.
Additionally, by placing the 'Unwinder Skeleton Code' before the
section 'Registering an Unwinder' we have skipping including the
registration line in the skeleton code. But this is confusion for
users (I think) as the skeleton code is almost complete, except for
one missing line which the user has to figure out for themselves. By
reordering the sections, it is now obvious that the registration
should be included in the skeleton code, and the example is therefore
almost complete.
Additionally, in the example skeleton code the way in which the
frame-id was being built (using the current stack point and program
counter is (a) not correct, and (b) counter to what is laid out in the
'Unwinder Input' section when describing building a frame-id.
I've removed the incorrect code and replaced it with more generic
comments indicating what needs to be done. As the actual actions that
need to be performed are both architecture specific, and dependent on
the function being unwound, it's almost impossible to include more
exact code here, but I think what I'm proposing is less misleading
than what we had before.
aarch64nto32 emulation isn't supported. The tests will then fall back
on aarch64elf32. It does work but some extra warnings are being
generated because the "-z relro" being added aarch64nto but ignored by
aarch64elf32 emulation.
Skip the tests to avoid any problems.
ld: add support of QNX stack arguments for aarch64nto
QNX is handling the stack argument using a .note section. Generate it
according to ELF argument -zexecstack, -zstack-size and a new NTO
argument --lazy-stack. Another NTO argument --stack mimicking
-zstack-size is added in order to ensure compatibility with previously
made NTO linkers.
This requires a new emultempl nto.em which is applied above the default
${ARCH}elf.em.
ld/ChangeLog:
* emulparams/aarch64nto.sh: Move to nto.em.
* emultempl/nto.em: New file.
* testsuite/ld-aarch64/aarch64-nto.exp: New test.
* testsuite/ld-aarch64/nto-stack-note-1.d: New test.
* testsuite/ld-aarch64/nto-stack-note-2.d: New test.
* testsuite/ld-aarch64/start.s: New test.
* readelf.c (get_qnx_elfcore_note_type): New function.
(print_qnx_note): New function.
(process_note): Add support for QNX support.
include/ChangeLog:
* elf/common.h (QNT_DEBUG_FULLPATH): New define.
(QNT_DEBUG_RELOC): New define.
(QNT_STACK): New define.
(QNT_GENERATOR): New define.
(QNT_DEFAULT_LIB): New define.
(QNT_CORE_SYSINFO): New define.
(QNT_CORE_INFO): New define.
(QNT_CORE_STATUS): New define.
(QNT_CORE_GREG): New define.
(QNT_CORE_FPREG): New define.
(QNT_LINK_MAP): New define.
Alan Modra [Thu, 16 Mar 2023 06:40:07 +0000 (17:10 +1030)]
Re: Add --enable-linker-verssion
Output sections without any input sections to initialise their flags
have their flags initialised by data statements to LOAD, ALLOC,
HAS_CONTENTS by default. This is wrong for .comment. Fix that by
making the script initialise the section type to INFO, one of the
noalloc section types. That also allows the address of .comment to be
set to zero, as is usual for non-alloc sections.
Also, use source_sh for all of the sourced scripts to set up make
dependencies.
Alan Modra [Wed, 15 Mar 2023 07:37:44 +0000 (18:07 +1030)]
PR30217, dynamic relocations using local dynamic symbols
glibc's ld.so ignores local dynamic symbols. It's been that way
forever. We therefore can't use them on dynamic relocations. Fixing
that problem uncovered another problem in sorting of dynamic relocs,
caused no doubt by copying make_iplt_section (where we don't want
reloc sorting by the generic gold function, we want iplt relocs last)
to make_lplt_section (where we do want sorting).
PR 30217
* powerpc.cc (branch_needs_plt_entry): New function.
(Target_powerpc::plt_off): Use it here..
(Target_powerpc::Scan::global): ..and here to correct PLT16 reloc
handling for forced-local global symbols.
(Output_data_plt_powerpc::add_entry): Rename "stash"
parameter "is_local". Emit relative relocs for globals that
are forced local, and don't set_needs_dynsym_entry.
(Target_powerpc::make_lplt_section): Don't create a separate
reloc section, use rela_dyn.
(Target_powerpc::make_brlt_section): Likewise.
Tom de Vries [Wed, 15 Mar 2023 15:38:03 +0000 (16:38 +0100)]
[gdb/testsuite] Fix gdb.arch/amd64-stap-special-operands.exp for remote host
With test-case gdb.arch/amd64-stap-special-operands.exp and host board
local-remote-host-notty and target board native-gdbserver I run into:
...
(gdb) break -pstap three_arg^M
No probe matching objfile=`<any>', provider=`<any>', name=`three_arg'^M
Make breakpoint pending on future shared library load? (y or [n]) n^M
(gdb) FAIL: gdb.arch/amd64-stap-special-operands.exp: probe: three_arg: \
gdb_breakpoint: set breakpoint at -pstap three_arg
...
due to compiling two executables with the same name, and when uploading the
second one from host to build, we run into:
...
Upload from 127.0.0.1 failed, \
$outputs/gdb.arch/amd64-stap-special-operands/amd64-stap-special-operands: \
Text file busy.
...
Tom de Vries [Wed, 15 Mar 2023 15:38:03 +0000 (16:38 +0100)]
[gdb/testsuite] Unset DEBUGINFOD_URLS on remote host
When running test-case gdb.arch/i386-pkru.exp with host board
local-remote-host-notty and target board native-gdbserver on openSUSE
Tumbleweed (with DEBUGINFOD_URLS set), I run into:
...
This GDB supports auto-downloading debuginfo from the following URLs:^M
<https://debuginfod.opensuse.org/>^M
Enable debuginfod for this session? (y or [n]) ^CQuit^M
(gdb) FAIL: gdb.arch/i386-pkru.exp: runto: run to main
...
The problem is that the unsetenv for DEBUGINFOD_URLS in default_gdb_init:
...
# If DEBUGINFOD_URLS is set, gdb will try to download sources and
# debug info for f.i. system libraries. Prevent this.
unset -nocomplain ::env(DEBUGINFOD_URLS)
...
doesn't work on remote host.
Fix this by using "set debuginfod enabled off" for remote host.
Tom de Vries [Wed, 15 Mar 2023 15:38:03 +0000 (16:38 +0100)]
[gdb/testsuite] Fix gdb.arch/amd64*.exp with local-remote-host-native.exp
There's a number of gdb.arch/amd64*.exp test-cases that fail with host+target
board local-remote-host-native.exp because of using a .S file, generated from
a .c file.
If a test-case compiles the .S file when executing on remote host,
the .S file is already copied from build to host, such that it's available for
the compiler.
But that's not the case for the .c file, which is needed by gdb to show a
source line:
...
(gdb) continue^M
Continuing.^M
^M
Breakpoint 2, fn2 (y=y@entry=25, x=x@entry=6) at amd64-entry-value-inline.c:32^M
32 in gdb.arch/amd64-entry-value-inline.c^M
(gdb) FAIL: gdb.arch/amd64-entry-value-inline.exp: continue to breakpoint: \
break-here
...
Fix this by using "gdb_remote_download host <.c file>".
Tested on x86_64-linux, with host+target board local-remote-host-native.
Tom Tromey [Mon, 13 Feb 2023 16:56:58 +0000 (09:56 -0700)]
Implement DAP variables, scopes, and evaluate requests
The DAP code already claimed to implement "scopes" and "evaluate", but
this wasn't done completely correctly. This patch implements these
and also implements the "variables" request.
After this patch, variables and scopes correctly report their
sub-structure. This also interfaces with the gdb pretty-printer API,
so the output of pretty-printers is available.
Tom Tromey [Thu, 23 Feb 2023 17:45:47 +0000 (10:45 -0700)]
Hide the implementation of gdb_mpf
This renames the data member of gdb_mpf and makes it private. It also
adds a single new method to aid in this change. Unlike the earlier
changes here, I did this one all together because gdb_mpf has very few
uses.
Tom Tromey [Thu, 23 Feb 2023 14:13:40 +0000 (07:13 -0700)]
Clean up gmp-utils.h includes
gmp-utils.h includes "defs.h", but normally the rule in gdb is that
the .c files include this first. This patch changes this code to
match the rest of gdb.
Add a test-case that tests the sanity of target/host boards.
It contains a number of tests related to remote file manipulation, exercising:
- remote_upload
- remote_download
- remote_file exists
- remote_file delete
which check that these work together as expected.
Tested on x86_64-linux, with all relevant gdb/testsuite/boards/*.exp boards.
For target board remote-stdio-gdbserver.exp, this revealed a trivial problem
with the return value of proc ${board}_file for delete, so fix this.
The test-case shows that the proc ${board}_download in
local-remote-host-native.exp is broken, so remove it.
Likewise for board local-remote-host.exp, so remove proc ${board}_download and
associated ${board}_file.
Alan Modra [Fri, 10 Mar 2023 09:58:22 +0000 (20:28 +1030)]
gas/read.c: init more statics
* read.c (current_name, current_label, dwarf_file, dwarf_line): Move
to file scope.
(pobegin): Tidy pop_override_ok.
(read_a_source_file): Make last_eol an auto var.
(s_reloc): Constify bfd_relocs.
(read_begin): Init more variables.
Alan Modra [Sun, 12 Mar 2023 02:57:38 +0000 (13:27 +1030)]
gas .include and .incbin
This fixes a bug in .include and .incbin where given an absolute path
the -I dirs would be searched for the path.
* read.c (include_dir_count, include_dir_maxlen): Make them size_t.
(search_and_open): New function.
(s_incbin, s_include): Use search_and_open.
(init_include_dir): New function.
(add_include_dir): Don't set initial "." dir here.
* read.h (include_dir_count, include_dir_maxlen): Update.
(init_include_dir, search_and_open): Declare.
* as.c (gas_early_init): Call init_include_dir.
* config/tc-rx.c (rx_include): Avoid warning by using size_t.
* config/tc-tic54x.c (tic54x_set_default_include): Simplify and
use notes for include path.
(tic54x_mlib): Use search_and_open.
Alan Modra [Fri, 10 Mar 2023 09:56:51 +0000 (20:26 +1030)]
gas/ecoff.c: don't use zero struct copies to init
It might have made sense once upon a time, but doesn't nowadays when
compilers expand memset inline.
* ecoff.c (add_aux_sym_tir, allocate_scope, allocate_vlinks),
(allocate_shash, allocate_thash, allocate_tag, allocate_forward),
(allocate_thead, allocate_lineno_list): Use memset rather than
copying zero struct.
Andrew Burgess [Thu, 23 Feb 2023 11:45:11 +0000 (11:45 +0000)]
gdb: add gdbarch::displaced_step_buffer_length
The gdbarch::max_insn_length field is used mostly to support displaced
stepping; it controls the size of the buffers allocated for the
displaced-step instruction, and is also used when first copying the
instruction, and later, when fixing up the instruction, in order to
read in and parse the instruction being stepped.
However, it has started to be used in other places in GDB, for
example, it's used in the Python disassembler API, and it is used on
amd64 as part of branch-tracing instruction classification.
The problem is that the value assigned to max_insn_length is not
always the maximum instruction length, but sometimes is a multiple of
that length, as required to support displaced stepping, see rs600,
ARM, and AArch64 for examples of this.
It seems to me that we are overloading the meaning of the
max_insn_length field, and I think that could potentially lead to
confusion.
I propose that we add a new gdbarch field,
gdbarch::displaced_step_buffer_length, this new field will do
exactly what it says on the tin; represent the required displaced step
buffer size. The max_insn_length field can then do exactly what it
claims to do; represent the maximum length of a single instruction.
As some architectures (e.g. i386, and amd64) only require their
displaced step buffers to be a single instruction in size, I propose
that the default for displaced_step_buffer_length will be the
value of max_insn_length. Architectures than need more buffer space
can then override this default as needed.
I've updated all architectures to setup the new field if appropriate,
and I've audited all calls to gdbarch_max_insn_length and switched to
gdbarch_displaced_step_buffer_length where appropriate.
There should be no user visible changes after this commit.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Andrew Burgess [Fri, 10 Mar 2023 15:35:52 +0000 (15:35 +0000)]
gdbarch: make invalid=True the default for all Components
This commit switches the default value for the 'invalid' field from
False to True. All components that previous set the invalid field to
True explicitly have had the field removed.
I think that True is a good choice for the default, this means that we
now get the validity checks by default, and if anyone adds a new
Component they need to make a choice to add an 'invalid=False' line
and disable the validation.
The flip side of this is that 'invalid=False' seems to be far more
common than 'invalid=True'. But I don't see a huge problem with this,
we shouldn't be aiming to reduce our typing, rather we should choose
based on which is least likely to introduce bugs. I think assuming
that we should do a validity check will achieve that.
Some additional components need to have an 'invalid=False' line added
to their definition, these are components that have a predefault
value, which is sufficient; the tdep code doesn't need to replace this
value if it doesn't want to.
Without adding the 'invalid=False' these components would be
considered to be invalid if they have not changed from their
predefault value -- but the predefault is fine.
There's no change in the generated code after this commit, so there
will be no user visible changes after this commit.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Andrew Burgess [Fri, 10 Mar 2023 15:31:36 +0000 (15:31 +0000)]
gdbarch: remove some unneeded predefault="0" from gdbarch_components.py
I noticed that there are a bunch of 'predefault="0"' lines in
gdbarch_components.py, and that some (just some, not all) of these are
not needed.
The gdbarch is already zero initialized, but these lines seem to
exists so that we can know when to compare against "0" and when to
compare against "NULL". At least, this seems to be useful in some
places in the generated code.
Specifically, if we remove the predefault="0" line from the
max_insn_length component then we end up generating a line like:
gdb_assert (gdbarch->max_insn_length != NULL);
which doesn't compile as we compare a ULONGEST to NULL.
In this commit I remove all the predefault="0" lines that I claim are
obviously not needed. These are lines for components that are not
Values (i.e. the component holds a function pointer anyway), or for
Value components that hold a pointer type, in which case using NULL is
fine.
The only changes after this commit are some fields that have nullptr
as their initial value, and gcore_bfd_target now compares to NULL not
0 in gdbarch_gcore_bfd_target_p, which, given the field is of type
'const char *', seems like an improvement.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Andrew Burgess [Fri, 10 Mar 2023 15:13:55 +0000 (15:13 +0000)]
gdbarch: improve generation of validation in gdbarch getters
We currently generate some validation code within the gdbarch getter
methods.
This commit adjusts the algorithm used to generate this validation
slightly to make the gdbarch.py code (I think) clearer; there's no
significant changes to what is generated.
The validation algorithm for gdbarch values is now:
- If the Value has an 'invalid' field that is a string, use that for
validation,
- If the Value has its 'predicate' field set to true, then check the
predicate returns true, this ensures the predicate has been
called,
- If the Value has its 'invalid' field set to True, or the Value has
'postdefault' field, then check the fields has changed from its
initial value,
- Otherwise no validation is performed.
The only changes after this commit are:
- Some comments change slightly, and
- For 'gcore_bfd_target' and 'max_insn_length' we now validate by
calling the predicate rather than checking the field value
directly, the underlying check being performed is unchanged
though.
There should be no user visible changes after this commit.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
All use a postdefault but no predefault to set the default value for
the component.
As the postdefault values for these components are all constant
pointers that don't depend on other fields within the gdbarch, then I
don't see any reason why we couldn't use a predefault instead.
So lets do that.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Andrew Burgess [Mon, 6 Mar 2023 14:39:28 +0000 (14:39 +0000)]
gdb/gdbarch: remove the 'invalid=None' state from gdbarch_components.py
This commit ensures that the 'invalid' property of all components is
either True, False, or a string.
Additionally, this commit allows a component to have both a predicate
and for the 'invalid' property to be a string.
Removing the option for 'invalid' to be None allows us to simplify the
algorithms in gdbarch.py a little.
Allowing a component to have both a predicate and an 'invalid' string
means that we can validate the value that a tdep sets into a field,
but also allow a predicate to ensure that the field has changed from
the default.
This functionality isn't going to be used in this series, but I have
tested it locally and believe that it would work, and this might make
it easier for others to add new components in the future.
In gdbarch_types.py, I've updated the type annotations to show that
the 'invalid' field should not be None, and I've changed the default
for this field from None to False.
The change to using False as the default is temporary. Later in this
series I'm going to change the default to True, but we need more fixes
before that can be done.
Additionally, in gdbarch_types.py I've removed an assert from
Component.get_predicate. This assert ensured that we didn't have the
predicate field set to True and the invalid field set to a string.
However, no component currently uses this configuration, and after
this commit, that combination is now supported, so the assert can be
removed.
As a consequence of the gdbarch_types.py changes we see some
additional comments generated in gdbarch.c about verification being
skipped due to the invalid field being False. This comment is inline
with plenty of other getters that also have a similar comment. Plenty
of the getters do have validation, so I think it is reasonable to have
a comment noting that the validation has been skipped for a specific
reason, rather than due to some bug.
In gdbarch_components.py I've had to add 'invalid=True' for two
components: gcore_bfd_target and max_insn_length, without this the
validation in the gdbarch getter would disappear.
And in gdbarch.py I've reworked the logic for generating the
verify_gdbarch function, and for generating the getter functions.
The logic for generating the getter functions is still not ideal, I
ended up having to add this additional logic block:
elif c.postdefault is not None and c.predefault is not None:
print(" /* Check variable changed from pre-default. */", file=f)
print(f" gdb_assert (gdbarch->{c.name} != {c.predefault});", file=f)
which was needed to ensure we continued to generate the same code as
before, without this the fact that invalid is now False when it would
previously have been None, meant that we dropped the gdb_assert in
favour of a comment like:
Andrew Burgess [Thu, 23 Feb 2023 18:18:45 +0000 (18:18 +0000)]
gdb/gdbarch: split postdefault setup from invalid check in gdbarch.py
Restructure how gdbarch.py generates the verify_gdbarch function.
Previously the postdefault handling was bundled together with the
validation. This means that a field can't have both a postdefault,
and set its invalid attribute to a string.
This doesn't seem reasonable to me, I see no reason why a field can't
have both a postdefault (used when the tdep doesn't set the field),
and an invalid expression, which can be used to validate the value
that a tdep might set.
In this commit I restructure the verify_gdbarch generation code to
allow the above, there is no change in the actual generated code in
this commit, that will come in later commit.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Andrew Burgess [Mon, 6 Mar 2023 11:45:12 +0000 (11:45 +0000)]
gdb/gdbarch: remove yet more 'invalid=True' from gdbarch_components.py
Following on from the previous commit, this commit removes yet more
'invalid=True' lines from gdbarch_components.py where the invalid
setting has no effect.
Due to the algorithm used in gdbarch.py for generated verify_gdbarch,
if a component has a postdefault value then no invalid check will ever
be generated for the component, as such setting 'invalid=True' on the
component is pointless. This commit removes the setting of invalid.
There is no change in the generated code after this commit.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Andrew Burgess [Mon, 6 Mar 2023 11:40:51 +0000 (11:40 +0000)]
gdb/gdbarch: remove unused 'invalid=True' from gdbarch_components.py
Due to the algorithm used to generate verify_gdbarch in gdbarch.py, if
a component has a predicate, then a validation check will never be
generated.
There are a bunch of components that are declared with both a
predicate AND have 'invalid=True' set. The 'invalid=True' has no
effect.
In this commit I clean things up by removing all these additional
'invalid=True' lines. There's no change in any of the generated files
after this commit.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Tom Tromey [Fri, 24 Feb 2023 17:40:16 +0000 (10:40 -0700)]
Fix crash in inside_main_func
gdb 13.1 crashes while running the rust compiler's debugger tests.
The crash has a number of causes.
First, the rust compiler still uses the C++-like _Z mangling, but with
its own twist -- some hex digits added to the end of a symbol. So,
while gdb finds the correct name of "main":
(top-gdb) p name
$13 = 0x292e0c0 "rustc_gdb_1031745::main"
It isn't found in the minsyms, because C++ demangling yields:
[99] t 0x90c0 _ZN17rustc_gdb_10317454main17h5b5be7fe16a97225E section .text rustc_gdb_1031745::main::h5b5be7fe16a97225 zko06yobckx336v
This could perhaps be fixed. I also filed a new PR to suggest
preferring the linkage name of the main program.
Next, the rust compiler emits both a DW_TAG_subprogram and a
DW_TAG_namespace for "main". This happens because the file is named
"main.rs" -- i.e., the bug is specific to the source file name. The
crash also seems to require the nested function inside of 'main', at
least for me. The namespace always is generated, but perhaps this
changes the ordering in the DWARF.
When inside_main_func looks up the main symbol, it finds the namespace
symbol rather than the function. (I filed a bug about fixing gdb's
symbol tables -- long overdue.)
Meanwhile, as I think it's important to fix this crash sooner rather
than later, this patch changes inside_main_func to check that the
symbol that is found is LOC_BLOCK. This perhaps should have been done
in the first place, anyway.
Tom de Vries [Mon, 13 Mar 2023 16:20:09 +0000 (17:20 +0100)]
[gdb/testsuite] Fix gdb.python/tui-window-factory.exp for remote host
When running gdb.python/tui-window.exp with host board
local-remote-host-notty and target board native-gdbserver, I get:
...
FAIL: gdb.python/tui-window-factory.exp: msg_2: \
check test_window box (box check: ul corner is l, not +)
...
The problem is that the result of Term::prepare_for_tui is not checked.
Tom de Vries [Mon, 13 Mar 2023 16:20:09 +0000 (17:20 +0100)]
[gdb/testsuite] Fix gdb.python/tui-window.exp for remote host
When running gdb.python/tui-window.exp with host board
local-remote-host-notty and target board native-gdbserver, I get:
...
UNSUPPORTED: gdb.python/tui-window.exp: TUI not supported
FAIL: gdb.python/tui-window.exp: test title
...
Fix this by adding the missing return after the unsupported.
Tom de Vries [Mon, 13 Mar 2023 16:20:09 +0000 (17:20 +0100)]
[gdb/testsuite] Require ![is_remote host] for TUI
When running test-case gdb.tui/corefile-run.exp with both host and target board
local-remote-host-native.exp, we run into:
...
FAIL: gdb.tui/corefile-run.exp: load corefile
...
while this passes with USE_TUI=0.
The problem is that the TUI setup code uses "setenv TERM ansi", which has no
effect on remote host.
I can confirm this analysis by working around this problem in
local-remote-host-native.exp like this:
...
- spawn $RSH -t -l $username $remote $cmd
+ spawn $RSH -t -l $username $remote "export TERM=ansi; $cmd"
...
For now, simply make TUI unsupported for remote host, by returning 0 in
prepare_for_tui.
projects / thirdparty / binutils-gdb.git / log
