Jose E. Marchesi [Fri, 19 Jul 2019 13:35:43 +0000 (15:35 +0200)]
cpu,opcodes,gas: use %r0 and %r6 instead of %a and %ctf in eBPF disassembler
This patch changes the eBPF CPU description to prefer the register
names %r0 and %r6 instead of %a and %ctx when disassembling. This
matches better with the current practice, vs. cBPF.
It also updates the GAS tests in order to reflect this change.
Tested in a x86_64 host.
cpu/ChangeLog:
2019-07-19 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf.cpu (h-gpr): when disassembling, use %r0 and %r6 instead of
%a and %ctx.
opcodes/ChangeLog:
2019-07-19 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf-desc.c: Regenerated.
gas/ChangeLog:
2019-07-19 Jose E. Marchesi <jose.marchesi@oracle.com>
* testsuite/gas/bpf/alu.d: Use %r6 instead of %ctx.
* testsuite/gas/bpf/lddw-be.d: Likewise.
* testsuite/gas/bpf/lddw.d: Likewise.
* testsuite/gas/bpf/alu-be.d: Likewise.
* testsuite/gas/bpf/alu32.d: Likewise.
After some discussion, we've decided to rename the +bitperm feature
flag to +sve2-bitperm, so that it's consistent with the other SVE2
feature flags. The associated internal macros already used
"SVE2_BITPERM", so only the feature flag itself needs to change.
2019-07-19 Richard Sandiford <richard.sandiford@arm.com>
Alan Modra [Fri, 19 Jul 2019 06:06:58 +0000 (15:36 +0930)]
[PowerPC64] pc-relative TLS relocations
This patch supports using pcrel instructions in TLS code sequences. A
number of new relocations are needed, gas operand modifiers to
generate those relocations, and new TLS optimisation. For
optimisation it turns out that the new pcrel GD and LD sequences can
be distinguished from the non-pcrel GD and LD sequences by there being
different relocations on the new sequence. The final "add ra,rb,13"
on IE sequences similarly needs a new relocation, or as I chose, a
modification of R_PPC64_TLS. On pcrel IE code, the R_PPC64_TLS points
one byte into the "add" instruction rather than being on the
instruction boundary.
GD:
pla 3,z@got@tlsgd@pcrel # R_PPC64_GOT_TLSGD34
bl __tls_get_addr@notoc(z@tlsgd) # R_PPC64_TLSGD and R_PPC64_REL24_NOTOC
edited to IE
pld 3,z@got@tprel@pcrel
add 3,3,13
edited to LE
paddi 3,13,z@tprel
nop
LD:
pla 3,z@got@tlsld@pcrel # R_PPC64_GOT_TLSLD34
bl __tls_get_addr@notoc(z@tlsld) # R_PPC64_TLSLD and R_PPC64_REL24_NOTOC
..
paddi 9,3,z2@dtprel
pld 10,z3@got@dtprel@pcrel
add 10,10,3
edited to LE
paddi 3,13,0x1000
nop
IE:
pld 9,z@got@tprel@pcrel # R_PPC64_GOT_TPREL34
add 3,9,z@tls@pcrel # R_PPC64_TLS at insn+1
ldx 4,9,z@tls@pcrel
lwax 5,9,z@tls@pcrel
stdx 5,9,z@tls@pcrel
edited to LE
paddi 9,13,z@tprel
nop
ld 4,0(9)
lwa 5,0(9)
std 5,0(9)
The first two of these allow you to get function type info and args out
of the types section give a type ID: astonishingly, this was missing
from libctf before now: so even though types of kind CTF_K_FUNCTION were
supported, you couldn't find out anything about them. (The existing
ctf_func_info and ctf_func_args only allow you to get info about
functions in the function section, i.e. given symbol table indexes, not
type IDs.)
The second of these allows you to get the raw undecorated name out of
the CTF section (strdupped for safety) without traversing subtypes to
build a full C identifier out of it. It's useful for things that are
already tracking the type kind etc and just need an unadorned name.
On Windows, passing a named pipe as terminal argument to the new-ui
command does not work.
The problem is that the new_ui_command function in top.c opens the
same tty three times, for stdin, stdout and stderr. With Windows
named pipes, the second and third calls to open fail.
Opening the file only once and passing the same stream for stdin,
stdout and stderr makes it work.
Pedro says:
I tried it on GNU/Linux and things still work.
I ran all the MI tests with forced new-ui, with:
$ make check TESTS="gdb.mi/*.exp" RUNTESTFLAGS="FORCE_MI_SEPARATE_UI=1"
Tom Tromey [Thu, 18 Jul 2019 15:29:15 +0000 (09:29 -0600)]
Constify main_name
This patch constifies the return type of main_name. There is a
comment indicating that this wasn't possible at some point in the
past, but whatever the barrier was, it is gone now.
Tested by rebuilding.
gdb/ChangeLog
2019-07-18 Tom Tromey <tromey@adacore.com>
H.J. Lu [Thu, 18 Jul 2019 15:01:45 +0000 (08:01 -0700)]
ld -r: Don't merge with member of output section group
When doing a relocatable link, members of input section group are
placed in their own output sections. We need to make sure that no
input sections are merged with member of output section group.
PR ld/24819
* emultempl/elf32.em (elf_orphan_compatible): Return FALSE for
member of output section group when doing a relocatable link.
* testsuite/ld-elf/pr24819.d: New file.
* testsuite/ld-elf/pr24819.s: Likewise.
Alan Modra [Thu, 18 Jul 2019 01:42:40 +0000 (11:12 +0930)]
[PowerPC64] Use STN_UNDEF internally for edited relocs
It's not correct to use non-STT_TLS symbols with TLS relocation, not
that it matters much when editing relocs, but this edited reloc can be
output by --emit-relocs. So don't use a symbol on the reloc.
* elf64-ppc.c (ppc64_elf_relocate_section): Don't bother selecting
a TLS section symbol for edited relocs. Tighten TLS symbol/reloc
match test.
Alan Modra [Thu, 18 Jul 2019 01:34:06 +0000 (11:04 +0930)]
[PowerPC64] Don't store TLS_EXPLICIT in tls_mask
This saves a bit in tls_mask, and fixes a bug that could be triggered
in the unlikely case that both @got (usual ELF style) and @toc
(PowerOpen style) code was used to set up args for __tls_get_addr.
* elf64-ppc.c (TLS_EXPLICIT): Define as 256.
(ppc64_elf_check_relocs): Don't store TLS_EXPLICIT even if char
is more than 8 bits.
(ppc64_elf_tls_optimize): Likewise. Make tls_set, tls_clear, and
tls_type vars unsigned int.
(ppc64_elf_relocate_section): Use r_type rather than TLS_EXPLICIT
to select r_type edit.
Alan Modra [Thu, 18 Jul 2019 01:28:01 +0000 (10:58 +0930)]
[PowerPC] Rename TLS_TPRELGD to TLS_GDIE
Choose a better name, that reflects why the flag is set (GD to IE
optimisation) rather than what the flag produces (TPREL64 reloc on
a single GOT entry replacing a tls_index pair).
* elf32-ppc.c (TLS_GDIE): Rename from TLS_TPRELGD throughout file.
Correct comment.
* elf64-ppc.c (TLS_GDIE): Likewise.
Jose E. Marchesi [Wed, 17 Jul 2019 20:57:23 +0000 (22:57 +0200)]
gas: support .half, .word and .dword directives in eBPF
This little patch adds support to the eBPF port of GAS for a few data
directives. The names for the directives have been chosen to be
coherent with the suffixes used in eBPF instructions: b, h, w and dw
for 8, 16, 32 and 64-bit values respectively.
Documentation and tests included.
Tested in a x86_64 host.
gas/ChangeLog:
2019-07-17 Jose E. Marchesi <jose.marchesi@oracle.com>
* config/tc-bpf.c (md_pseudo_table): .half, .word and .dword.
* testsuite/gas/bpf/data.s: New file.
* testsuite/gas/bpf/data.d: Likewise.
* testsuite/gas/bpf/data-be.d: Likewise.
* testsuite/gas/bpf/bpf.exp: Run data and data-be.
* doc/c-bpf.texi (BPF Directives): New section.
Tom Tromey [Tue, 2 Jul 2019 23:07:02 +0000 (17:07 -0600)]
Move source window common to code to tui-winsource.[ch]
Like the previous rearranging patches, this moves the source and
disassembly window base class code to tui-winsource.[ch]. The
execution info window is also moved, because it is associated with
this base class.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui-winsource.h (struct tui_exec_info_window)
(struct tui_source_window_base): Move from tui-data.h.
* tui/tui-winsource.c: Move many method definitions from
elsewhere. Remove "structuring" comments.
* tui/tui-wingeneral.c (tui_source_window_base::make_visible)
(tui_source_window_base::refresh_window): Move to
tui-winsource.c.
* tui/tui-win.c (tui_source_window_base::refresh_all)
(tui_source_window_base::update_tab_width)
(tui_source_window_base::set_new_height)
(tui_source_window_base::do_make_visible_with_new_height): Move to
tui-winsource.c.
* tui/tui-source.h: Update.
* tui/tui-source.c (tui_source_window_base::reset): Move to
tui-winsource.c.
* tui/tui-disasm.h: Update.
* tui/tui-data.h (struct tui_exec_info_window): Move to
tui-winsource.h.
(struct tui_source_window_base): Likewise.
* tui/tui-data.c (tui_source_window_base::clear_detail)
(tui_source_window_base, ~tui_source_window_base): Move to
tui-winsource.c.
Tom Tromey [Tue, 2 Jul 2019 21:48:59 +0000 (15:48 -0600)]
Move TUI command window code
Like the earlier change to the data window, this moves the TUI command
window code to tui-command.[ch], and removes the old "structuring"
comments from tui-command.c.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui.c: Update.
* tui/tui-win.c (tui_cmd_window::do_make_visible_with_new_height)
(tui_cmd_window::max_height): Move to tui-command.c.
* tui/tui-layout.c: Update.
* tui/tui-data.h (struct tui_cmd_window): Move to tui-command.h.
* tui/tui-data.c (tui_cmd_window::clear_detail): Move to
tui-command.c.
* tui/tui-command.h (struct tui_cmd_window): Move from
tui-data.h.
* tui/tui-command.c: Remove "structuring" comments.
(tui_cmd_window::clear_detail)
(tui_cmd_window::do_make_visible_with_new_height)
(tui_cmd_window::max_height): Move from elsewhere.
Tom Tromey [Tue, 2 Jul 2019 21:41:08 +0000 (15:41 -0600)]
Rearrange TUI data window code
An earlier patch caused tui-windata.h to be essentially empty. And,
other earlier patches implemented TUI data window methods in any spot
that happened to be convenient at the time.
This patch rearranges all the data window code to be somewhat more
organized. It moves tui_data_window to tui-regs.h, and moves the
implementation of all methods to tui-regs.c. It then removes
tui-windata.h and tui-windata.c.
It also removes the "structuring" comments from tui-regs.c; these are
not the usual gdb style, and were out of date anyhow. Finally, it
moves _initialize_tui_regs to the end of the file, per the usual gdb
convention.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui.c: Update.
* tui/tui-wingeneral.c (tui_data_window::refresh_window): Move to
tui-regs.c.
* tui/tui-windata.h: Remove file.
* tui/tui-windata.c: Remove file.
* tui/tui-win.c (tui_data_window::set_new_height)
(tui_data_window::do_make_visible_with_new_height): Move to
tui-regs.c.
* tui/tui-regs.h (struct tui_data_window): Move from tui-data.h.
* tui/tui-regs.c: Remove "structuring" comments.
(tui_data_window::first_data_item_displayed)
(tui_data_window::delete_data_content_windows)
(tui_data_window::erase_data_content)
(tui_data_window::display_all_data)
(tui_data_window::refresh_all)
(tui_data_window::do_scroll_vertical)
(tui_data_window::clear_detail, tui_data_window::set_new_height)
(tui_data_window::do_make_visible_with_new_height)
(tui_data_window::refresh_window): Move from elsewhere.
(_initialize_tui_regs): Move to end of file.
* tui/tui-layout.c: Update.
* tui/tui-hooks.c: Update.
* tui/tui-data.h (struct tui_data_window): Move to tui-regs.h.
* tui/tui-data.c (tui_data_window::clear_detail): Move to
tui-regs.c.
* Makefile.in (SUBDIR_TUI_SRCS): Remove tui-windata.c.
Tom Tromey [Tue, 2 Jul 2019 21:18:27 +0000 (15:18 -0600)]
Fix flushing bug in tui_puts_internal
A while back I changed gdb not to flush in some places. It turned out
that this broke the TUI a little. An easy way to see it is to run
"gdb -tui -nx", then "file gdb" at the gdb prompt. gdb will print the
usual "Reading symbols..." message -- but it won't appear on-screen
until the reading is complete.
This patch changes the TUI to do the equivalent of line buffering in
tui_puts_internal.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui-io.c (tui_puts_internal): Call wrefresh if newline is
seen.
Tom Tromey [Mon, 1 Jul 2019 20:44:29 +0000 (14:44 -0600)]
Remove has_locator method
Earlier changes made it obvious that the has_locator method can only
be called for source/disassembly windows. Because the only reference
to this now occurs in methods on this object, we can remove the
has_locator method entirely, in favor of using the member directly.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui-win.c (tui_source_window_base::set_new_height)
(tui_source_window_base::do_make_visible_with_new_height): Use
m_has_locator field directly.
* tui/tui-data.h (struct tui_win_info) <has_locator>: Remove
method.
(struct tui_source_window_base) <has_locator>: Likewise.
Tom Tromey [Mon, 1 Jul 2019 20:41:28 +0000 (14:41 -0600)]
Remove tui_make_visible and tui_make_invisible
tui_make_visible and tui_make_invisible are just wrappers for a method
call, so remove them and have the callers simply make the method call
themselves.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 1 Jul 2019 20:37:38 +0000 (14:37 -0600)]
Remove make_data_window
As with the previous patches, unifying the creation and
re-initialization cases for the data window lets us remove
make_data_window in favor of simply using "new".
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui-layout.c (make_data_window): Remove.
(show_data): Unify creation and re-initialization cases.
Tom Tromey [Mon, 1 Jul 2019 20:37:04 +0000 (14:37 -0600)]
Remove make_source_window and make_disasm_window
This unifies the remaining creation and re-initialization cases for
the source and disassembly windows. Once this is done, it's clear
that make_source_window and make_disasm_window aren't needed any more,
so remove them.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 1 Jul 2019 20:26:42 +0000 (14:26 -0600)]
Remove make_command_window
This unifies the creation and re-initialization cases for the command
window. When this is done, it becomes clear that make_command_window
isn't needed -- it can be replaced with a simple "new", so this is
removed as well.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 1 Jul 2019 02:46:18 +0000 (20:46 -0600)]
Simplify show_source_disasm_command
This is the first of a few patches to further simplify window
(re-)initialization in tui-layout.c. When changing the layout, a
window may be created or, if it already exists, simply resized. These
two cases normally are identical, but this was obscured by the way the
code was written. This patch changes show_source_disasm_command to
unify the creation and re-initialization cases.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 1 Jul 2019 02:43:32 +0000 (20:43 -0600)]
Clean up tui_layout_command
tui_layout_command is a simple wrapper for tui_set_layout_by_name.
This removes the extra layer and cleans up the resulting function a
bit -- changing it to call error rather than return a result. This
necessitated a small change to tui-regs.c, to avoid calling the
function that is being removed.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui.h (tui_set_layout_by_name): Don't declare.
* tui/tui-regs.c (tui_reg_layout): New function.
(tui_show_registers, tui_reg_command): Use it.
* tui/tui-layout.c (LAYOUT_USAGE): Remove.
(tui_layout_command): Rename from tui_set_layout_by_name. Change
parameters.
(tui_layout_command): Remove.
Tom Tromey [Mon, 1 Jul 2019 02:34:57 +0000 (20:34 -0600)]
Change tui_set_layout to return void
tui_set_layout can't meaningfully be called with UNDEFINED_LAYOUT; and
instead of trying to handle this case, simply assert and have the
function return void. No caller was checking the return value anyway.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 1 Jul 2019 02:32:35 +0000 (20:32 -0600)]
Remove reset_locator
reset_locator was introduced just a few patches ago, but it's already
time to remove it. It consists of a call to the locator's "reset"
method, plus a call to tui_make_window; but the latter is redundant at
all the places that call reset_locator.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 29 Jun 2019 13:32:47 +0000 (07:32 -0600)]
Remove the win_type parameter from tui_gen_win_info::reset
tui_gen_win_info::reset has a window type parameter that is only used
for an assertion. This made sense as a defensive measure when window
creation was more dynamic -- it ensured that one did not make
mistakes. However, there's no need for it any more, so this removes
it.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 29 Jun 2019 05:56:25 +0000 (23:56 -0600)]
Introduce reset_locator function in tui-layout.c
init_and_make_win in tui-layout.c is now only called for the locator
-- earlier changes have made most of the cases here obsolete. This
patch removes init_and_make_win and introduces a reset_locator
function. Window creation is now much simpler to follow, because it
is no longer quite so dynamic. (Though it will become even simpler in
coming patches.)
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui-layout.c (show_source_disasm_command): Use
reset_locator.
(reset_locator): New function.
(init_and_make_win): Remove.
(show_source_or_disasm_and_command): Use reset_locator.
Tom Tromey [Sat, 29 Jun 2019 05:54:25 +0000 (23:54 -0600)]
Always create an execution info window for a source window
A source or disassembly window will always have an "execution info"
window (the window along the side that displays breakpoint info), but
this isn't immediately clear from the source. As a result, some code
has checks to see whether the execution_info is NULL.
This changes the source window base class to always instantiate an
execution_info window, then updates the rest of the code. It also
simplifies window creation in tui-layout.c.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 29 Jun 2019 05:32:05 +0000 (23:32 -0600)]
Remove TUI data window special case
The TUI has a couple of special cases for updating the data window:
one in tui_rl_other_windowand one in tui_set_focus_command. As part
of the project to remove references to globals, I wanted to remove
these calls; but when I did, some simple operations (like "C-x o")
would cause the register window to blank.
This fixes the underlying problem by arranging for the data window's
refresh_window method to call the superclass method first, and then to
refresh the child windows. Then the special cases can be removed.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 29 Jun 2019 05:17:24 +0000 (23:17 -0600)]
Merge refresh and refresh_window methods
Earlier refactorings introduced the refresh and refresh_window
methods, following the previous TUI code. However, these methods are
essentially the same, so this patch merges them. It also removes some
redundant refresh_window calls, because the execution window is
updated automatically by the corresponding source (or disassembly)
window.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 29 Jun 2019 04:57:37 +0000 (22:57 -0600)]
Remove tui_source_window::content_in_use
Now that source window clearing has been simplified, we don't need a
special flag to say whether the source window is in use -- we can
simply check whether the contents are set. This patch implements this
idea, removing the content_in_use field.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 29 Jun 2019 04:55:45 +0000 (22:55 -0600)]
Simplify source window clearing
When a TUI source window is empty, it displays a "No Source Available"
message. The function tui_set_source_content_nil also made sure to
put this message into the window's "content" field.
However, I believe this isn't really necessary. Instead, it's simpler
to just empty the contents and let curses handle the refreshing.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 29 Jun 2019 04:48:33 +0000 (22:48 -0600)]
Remove unused parameter from two TUI functions
The "display_prompt" parameter of tui_erase_source_content and
tui_clear_source_content was never passed the NO_EMPTY_SOURCE_PROMPT
value, so remove the parameter. Once this is done, the
EMPTY_SOURCE_PROMPT and NO_EMPTY_SOURCE_PROMPT defines are unused, so
remove those as well.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 28 Jun 2019 22:44:23 +0000 (16:44 -0600)]
Move content_in_use to tui_source_window class
From scanning the source now, it's clear that the content_in_use field
is only used for the source window. This patch moves the field there,
and changes it to be a bool at the same time. (A future patch will
clean this up further, removing the field entirely.)
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 28 Jun 2019 20:48:45 +0000 (14:48 -0600)]
Make source windows be self-updating
This changes the TUI source window to register itself on the
source_styling_changed observable, and removes a bit of code from
tui-hooks.c. This reduces the number of uses of the TUI_SRC_WIN
global.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
* tui/tui-source.c (tui_source_window): New constructor. Add
observer.
(~tui_source_window): New destructor.
(tui_source_window::style_changed): New method.
* tui/tui-hooks.c (tui_redisplay_source): Remove.
(tui_attach_detach_observers): Update.
* tui/tui-data.h (struct tui_source_window): Make constructor not
inline. Add destructor.
(struct tui_source_window) <style_changed>: New method.
<m_observable>: New member.
Tom Tromey [Wed, 26 Jun 2019 22:13:39 +0000 (16:13 -0600)]
Introduce two more tui_data_window methods
This changes tui_display_registers_from and
tui_display_registers_from_line to be methods on tui_data_window,
allowing for the removal of more uses of the TUI_DATA_WIN global.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Wed, 26 Jun 2019 21:56:56 +0000 (15:56 -0600)]
Remove tui_display_data_from_line
tui_display_data_from_line is just a wrapper for
tui_display_registers_from_line, so remove it. Also, nothing passed 0
as the "force_display" parameter to tui_display_registers_from_line,
so remove that parameter as well.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
This changes tui_first_reg_element_no_inline to be a method on
tui_data_window. This again moves uses of the TUI_DATA_WIN global
from this function into other functions that are already using the
global.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
This changes tui_line_from_reg_element_no to be a method on
tui_data_window, allowing for the removal of some uses of the
TUI_DATA_WIN global. (Actually, in this particular patch, the number
of uses is not decreased, but rather the uses are moved to spots that
are already using the global, i.e., increasing the clustering.)
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 24 Jun 2019 03:25:46 +0000 (21:25 -0600)]
Remove deleted breakpoint from TUI display
PR tui/24722 points out that deleting a breakpoint does not cause the
"b" to be removed from the breakpoint display. The issue here was
that the breakpoint still exists at the moment the breakpoint-deleted
observer is notified. This fixes the problem by specially handling
the case where a breakpoint is being removed.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 24 Jun 2019 03:00:01 +0000 (21:00 -0600)]
Consolidate "if"s in tui_show_frame_info
tui_show_frame_info has two parallel "if"s with the same condition and
no intervening code. The second "if" perhaps wasn't obviously
identical, but earlier refactorings made it clearer, by pointing out
that only source and disassembly windows can be handled here. This
merges the two statements.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 24 Jun 2019 02:48:59 +0000 (20:48 -0600)]
Check can_highlight in tui_check_and_display_highlight_if_needed
tui_check_and_display_highlight_if_needed currently checks the
window's type; but this can be replaced with a check of
"can_highlight", making it more polymorphically-correct.
gdb/ChangeLog
2019-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Wed, 17 Jul 2019 14:34:23 +0000 (08:34 -0600)]
Rename field_int to field_signed
This renames ui_out::field_int to field_signed, and field_fmt_int to
field_fmt_signed; and changes the type of the "value" parameter from
int to LONGEST.
Tested by the buildbot.
gdb/ChangeLog
2019-07-17 Tom Tromey <tromey@adacore.com>
* ui-out.h (class ui_out) <field_signed, field_fmt_signed,
do_field_signed>: Rename. Change type of "value".
* ui-out.c (ui_out::field_signed): Rename from field_int.
Change type of "value".
(ui_out::field_fmt_signed): Rename from field_fmt_int. Change
type of "value".
* tui/tui-out.h (class tui_ui_out) <do_field_signed>: Rename from
do_field_int. Change type of "value".
* tui/tui-out.c (tui_ui_out::do_field_signed): Rename from
do_field_int. Change type of "value".
* tracepoint.c (trace_status_mi, tfind_1)
(print_one_static_tracepoint_marker): Update.
* thread.c (print_thread_info_1, print_selected_thread_frame):
Update.
* stack.c (print_frame, print_frame_info): Update.
* spu-tdep.c (info_spu_signal_command, info_spu_dma_cmdlist):
Update.
* source.c (print_source_lines_base): Update.
* skip.c (info_skip_command): Update.
* record-btrace.c (btrace_ui_out_decode_error)
(btrace_call_history_src_line): Update.
* python/py-framefilter.c (py_print_single_arg, py_print_frame):
Update.
* progspace.c (print_program_space): Update.
* mi/mi-symbol-cmds.c (mi_cmd_symbol_list_lines): Update.
* mi/mi-out.h (class mi_ui_out) <do_field_signed>: Rename from
do_field_int. Change type of "value".
* mi/mi-out.c (mi_ui_out::do_table_begin)
(mi_ui_out::do_table_header): Update.
(mi_ui_out::do_field_signed): Rename from do_field_int. Change
type of "value".
* mi/mi-main.c (mi_cmd_thread_list_ids, print_one_inferior)
(mi_cmd_data_list_changed_registers, output_register)
(mi_cmd_data_read_memory, mi_load_progress)
(mi_cmd_trace_frame_collected): Update.
* mi/mi-interp.c (mi_on_normal_stop_1, mi_output_solib_attribs):
Update.
* mi/mi-cmd-var.c (print_varobj, mi_cmd_var_create)
(mi_cmd_var_delete, mi_cmd_var_info_num_children)
(mi_cmd_var_list_children, varobj_update_one): Update.
* mi/mi-cmd-stack.c (mi_cmd_stack_info_depth)
(mi_cmd_stack_list_args, list_arg_or_local): Update.
* mi/mi-cmd-file.c (mi_cmd_file_list_exec_source_file): Update.
* inferior.c (print_inferior): Update.
* gdb_bfd.c (print_one_bfd): Update.
* disasm.c (gdb_pretty_print_disassembler::pretty_print_insn):
Update.
* darwin-nat-info.c (darwin_debug_regions_recurse): Update.
* cli-out.h (class cli_ui_out) <do_field_signed>: Rename from
do_field_int. Change type of "value".
* cli-out.c (cli_ui_out::do_field_signed): Rename from
do_field_int. Change type of "value".
* breakpoint.c (watchpoint_check, print_breakpoint_location)
(print_one_breakpoint_location, print_it_catch_fork)
(print_one_catch_fork, print_it_catch_vfork)
(print_one_catch_vfork, print_it_catch_solib)
(print_it_catch_exec, print_it_ranged_breakpoint)
(print_mention_watchpoint, print_mention_masked_watchpoint)
(bkpt_print_it, update_static_tracepoint): Update.
* break-catch-throw.c (print_it_exception_catchpoint): Update.
* break-catch-syscall.c (print_it_catch_syscall): Update.
* ada-tasks.c (print_ada_task_info): Update.
* ada-lang.c (print_it_exception, print_mention_exception):
Update.
This commit updates the gdbarch method gdbarch_stap_adjust_register,
and the one implementation i386_stap_adjust_register to avoid using a
non-const reference parameter.
I've also removed the kfail from the testsuite for bug 24541, as this
issue is now resolved.
gdb/ChangeLog:
PR breakpoints/24541
* gdbarch.c: Regenerate.
* gdbarch.h: Regenerate.
* gdbarch.sh: Adjust return type and parameter types for
'stap_adjust_register'.
(i386_stap_adjust_register): Adjust signature and return new
register name.
* stap-probe.c (stap_parse_register_operand): Adjust use of
'gdbarch_stap_adjust_register'.
gdb/testsuite/ChangeLog:
PR breakpoints/24541
* gdb.mi/mi-catch-cpp-exceptions.exp: Remove kfail due to 24541.
Jan Beulich [Wed, 17 Jul 2019 07:16:31 +0000 (09:16 +0200)]
x86: replace "anymem" checks where possible
Once operand parsing has completed, the simpler check of Operand_Mem can
be used in places where i.types[] got passed to operand_type_check().
Note that this has shown a couple of omissions of adjusting i.flags[]
when playing with i.op[] / i.types[] / i.tm.operand_types[]. Not all of
them get added here, just all of the ones needed in process_operands().
Jan Beulich [Wed, 17 Jul 2019 07:15:49 +0000 (09:15 +0200)]
x86: drop stale Mem enumerator
This was supposed to also be removed by c48dadc9a8 ('x86: drop "mem"
operand type attribute'). It's odd enough that this hasn't caused
build issues, considering the careful use of OTunused (apparently to
avoid "missing initializer" warnings).
To avoid such happening again introduce compile time consistency
checks.
Jan Beulich [Tue, 16 Jul 2019 07:31:36 +0000 (09:31 +0200)]
x86: make RegMem an opcode modifier
... instead of an operand type bit: It's an insn property, not an
operand one. There's just one actual change to be made to the
templates: Most are now required to have the (unswapped) destination go
into ModR/M.rm, so VMOVD template needs its opcode adjusted accordingly
and its operands swapped. {,V}MOVS{S,D}, otoh, are left alone in this
regard, as otherwise generated code would differ from what we've been
producing so far (which I don't think is wanted).
Take the opportunity and add a missing IgnoreSize to pextrb (leading to
an error in 16-bit mode), and take the liberty to once again drop stray
IgnoreSize attributes from lines changed and neighboring related ones.
Jan Beulich [Tue, 16 Jul 2019 07:30:29 +0000 (09:30 +0200)]
x86: fold SReg{2,3}
They're the only exception to there generally being no mix of register
kinds possible in an insn operand template, and there being two bits per
operand for their representation is also quite wasteful, considering the
low number of uses. Fold both bits and deal with the little bit of
fallout.
Also take the liberty and drop dead code trying to set REX_B: No segment
register has RegRex set on it.
Additionally I was quite surprised that PUSH/POP with the permitted
segment registers is not covered by the test cases. Add the missing
pieces.
John Baldwin [Tue, 16 Jul 2019 00:21:36 +0000 (17:21 -0700)]
Fix build for aarch64, arm, and riscv FreeBSD native targets.
Remove unused gdbarch argument to helper functions originally
copied from mips-fbsd-nat.c. Include regcache.h previously
included from defs.h via gdbarch.h.
Tom Tromey [Mon, 1 Jul 2019 18:14:48 +0000 (12:14 -0600)]
Introduce field_unsigned
This adds field_unsigned and changes various places using field_fmt
with "%u" to use this instead. This also replaces an existing
equivalent helper function in record-btrace.c.
2019-07-15 Tom Tromey <tromey@adacore.com>
* mi/mi-out.h (class mi_ui_out) <do_field_unsigned>: Declare.
* mi/mi-out.c (mi_ui_out::do_field_unsigned): New method.
* cli-out.h (class cli_ui_out) <do_field_unsigned>: Declare.
* cli-out.c (cli_ui_out::do_field_int): New method.
* ui-out.c (ui_out::field_unsigned): New method.
* symfile.c (generic_load): Use field_unsigned.
(print_transfer_performance): Likewise.
* record-btrace.c (ui_out_field_uint): Remove.
(btrace_call_history_insn_range, btrace_call_history): Use
field_unsigned.
* disasm.c (gdb_pretty_print_disassembler::pretty_print_insn): Use
field_unsigned.
* ui-out.h (class ui_out) <field_unsigned>: New method.
<do_field_unsigned>: Likewise.
Tom Tromey [Mon, 1 Jul 2019 18:08:15 +0000 (12:08 -0600)]
Use field_string in more places
This replaces uses of field_fmt with a "%s" format string to use
field_string instead. Also, one use of "%9lx" is replaced with a call
to phex_nz; the '9' is dropped as it is implicit in the field width.
2019-07-15 Tom Tromey <tromey@adacore.com>
* mi/mi-main.c (list_available_thread_groups): Use field_string.
* mi/mi-interp.c (mi_memory_changed): Use field_string.
* target.c (flash_erase_command): Use field_string.
* infrun.c (print_signal_received_reason): Use field_string.
* i386-tdep.c (i386_mpx_print_bounds): Use field_string.
* breakpoint.c (maybe_print_thread_hit_breakpoint): Use
field_string.
* ada-tasks.c (print_ada_task_info): Use field_string.
Tom Tromey [Mon, 1 Jul 2019 18:02:37 +0000 (12:02 -0600)]
Use field_core_addr in more places
This changes a few spots that use field_fmt to use field_core_addr
instead.
gdb/ChangeLog
2019-07-15 Tom Tromey <tromey@adacore.com>
* target.c (flash_erase_command): Use field_core_addr.
* symfile.c (generic_load): Use field_core_addr.
* sparc64-linux-tdep.c (sparc64_linux_handle_segmentation_fault):
Use field_core_addr.
* i386-linux-tdep.c (i386_linux_handle_segmentation_fault): Use
field_core_addr.
Jose E. Marchesi [Mon, 15 Jul 2019 14:00:28 +0000 (16:00 +0200)]
cpu,opcodes,gas: fix explicit arguments to eBPF ldabs instructions
This patch fixes the eBPF CPU description in order to reflect the
right explicit arguments passed to the ldabs{b,h,w,dw} instructions,
updates the corresponding GAS tests, and updates the BPF section of
the GAS manual.
cpu/ChangeLog:
2019-07-15 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf.cpu (dlabs): New pmacro.
(dlind): Likewise.
opcodes/ChangeLog:
2019-07-15 Jose E. Marchesi <jose.marchesi@oracle.com>
2019-07-15 Jose E. Marchesi <jose.marchesi@oracle.com>
* testsuite/gas/bpf/mem.s: ldabs instructions do not take a `src'
register as an argument.
* testsuite/gas/bpf/mem.d: Updated accordingly.
* testsuite/gas/bpf/mem-be.d: Likewise.
* doc/c-bpf.texi (BPF Opcodes): Update to reflect the correct
explicit arguments to ldabs and ldind instructions.
Alan Modra [Fri, 12 Jul 2019 05:58:19 +0000 (15:28 +0930)]
Fix __bss_start assertion failure in _bfd_elf_fix_symbol_flags
> Building LLVM 6.0 on FreeBSD/powerpc (devel/llvm60 port) the assertion
> in the subject trips (displays twice) when linking libLTO.so.1. The
> issue has been filed in FreeBSD's bugzilla, at
> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=237068 . It appears
> the 'llvm::hashing::detail::get_execution_seed()::seed@@JL_LLVM_6.0'
> symbol is being weakly aliased to an indirect symbol
> __bss_start@@JL_LLVM_6.0. Since __bss_start@@JL_LLVM_6.0 is an
> indirect symbol, it fails the assertion.
I haven't looked under a debugger at your testcase but I think I know
what is going on here. You have a shared library with a weakly
defined llvm::hashing::detail::get_execution_seed()::seed which
happens to be at the same location as __bss_start in that library. At
the time the linker loads symbols for that library, it sees they are
both versioned and thus introduces non-versioned indirect symbols for
them. The linker considers the symbols as possibly being aliases,
setting up h->u.alias and h->is_weakalias such that
__bss_start@@JL_LLVM_6.0 is the definition. No real problem so far,
the definition is bfd_link_hash_defined, except that the zero size, no
type __bss_start symbol possibly should not be considered an alias in
the first place.
Later, __bss_start as defined by the linker script is entered into the
linker symbol table. This is similar to __bss_start being defined by
a regular object file in that ELF symbol resolution rules say that the
value of __bss_start in the library is overridden by __bss_start in
the executable/library being produced. So to accomplish the override,
ld flips __bss_start from being an indirect symbol pointing at
__bss_start@@JL_LLVM_6.0 to __bss_start@@JL_LLVM_6.0 being an indirect
symbol pointing at __bss_start. That's how we get an unexpected
indirect symbol and hit the assert.
What should happen I think, is for the def->def_regular code above the
assert to run in this case. The symbols are no longer aliases.
* elflink.c (_bfd_elf_fix_symbol_flags): If the def for an
alias is no longer bfd_link_hash_defined, clear the alias.
Jose E. Marchesi [Sun, 14 Jul 2019 12:45:31 +0000 (14:45 +0200)]
cpu,opcodes,gas: fix arguments to ldabs and ldind eBPF instructions
The eBPF non-generic load instructions ldind{b,h,w,dw} and
ldabs{b,h,w,dw} do not take an explicit destination register as an
argument. Instead, they put the loaded value in %r0, implicitly.
This patch fixes the CPU BPF description to not expect a 'dst'
argument in these arguments, regenerates the corresponding files in
opcodes, and updates the impacted GAS tests.
Tested in a x86-64 host.
cpu/ChangeLog:
2019-07-14 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf.cpu (dlsi): ldabs and ldind instructions do not take an
explicit 'dst' argument.
opcodes/ChangeLog:
2019-07-14 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf-desc.c: Regenerate.
* bpf-opc.c: Likewise.
gas/ChangeLog:
2019-07-14 Jose E. Marchesi <jose.marchesi@oracle.com>
* testsuite/gas/bpf/mem.s: Do not use explicit arguments for
ldabs and ldind instructions.
* testsuite/gas/bpf/mem.d: Updated accordingly.
* testsuite/gas/bpf/mem-be.d: Likewise.
Joel Brobecker [Sun, 14 Jul 2019 01:00:32 +0000 (18:00 -0700)]
adjust src-release following the renaming of gdb/common/ to gdb/gdbsupport/
A recent change renamed the common/ directory into gdbsupport/ in gdb.
This causes problems in the getver function in the src-release script
which doesn't find the create-version.sh script anymore. As a result,
it falls back on using the version.in file verbatim, meaning that
the "DATE" placeholder doesn't get replaced with the snapshot date,
and the "-git" suffix doesn't get stripped. More precisely, we get
snapshots called "gdb-8.3.50-DATE-git.tar" instead of (e.g.)
"gdb-8.3-20190712.tar".
This commit fixes the issue by adding support for this situation.
I left the support for $tool/common/create-version.sh, because
the sim still uses that directory structure.
ChangeLog:
* src-release (getver): If $tool/gdbsupport/create-version.sh
exists, use that to determine the version number.
projects / thirdparty / binutils-gdb.git / log
