mirror/binutils-gdb
2
0
Fork 0
mirror of https://sourceware.org/git/binutils-gdb.git synced 2025-02-11 13:02:10 +08:00
Code Issues Packages Projects Releases Wiki Activity
105,081 Commits 392 Branches 138 Tags 632 MiB
bfece7562d
Commit Graph

7 Commits

Author SHA1 Message Date
Simon Marchi
2189c31265 gdb: add remote_debug_printf 
This is the next in the new-style debug macro series.

For this one, I decided to omit the function name from the "Sending packet" /
"Packet received" kind of prints, just because it's not very useful in that
context and hinders readability more than anything else.  This is completely
arbitrary.

This is with:

  [remote] putpkt_binary: Sending packet: $qTStatus#49...
  [remote] getpkt_or_notif_sane_1: Packet received: T0;tnotrun:0;tframes:0;tcreated:0;tfree:500000;tsize:500000;circular:0;disconn:0;starttime:0;stoptime:0;username:;notes::

and without:

  [remote] Sending packet: $qTStatus#49...
  [remote] Packet received: T0;tnotrun:0;tframes:0;tcreated:0;tfree:500000;tsize:500000;circular:0;disconn:0;starttime:0;stoptime:0;username:;notes::

A difference is that previously, the query packet and its reply would be
printed on the same line, like this:

  Sending packet: $qTStatus#49...Packet received: T0;tnotrun:0;tframes:0;tcreated:0;tfree:500000;tsize:500000;circular:0;disconn:0;starttime:0;stoptime:0;username:;notes::

Now, they are printed on two lines, since each remote_debug_printf{,_nofunc}
prints its own complete message including an end of line.  It's probably
a matter of taste, but I prefer the two-line version, it's easier to
follow, especially when the query packet is long.

As a result, lib/range-stepping-support.exp needs to be updated, as it
currently expects the vCont packet and the reply to be on the same line.
I think it's sufficient in that context to just expect the vCont packet
and not the reply, since the goal is just to count how many vCont;r GDB
sends.

gdb/ChangeLog:

	* remote.h (remote_debug_printf): New.
	(remote_debug_printf_nofunc): New.
	(REMOTE_SCOPED_DEBUG_ENTER_EXIT): New.
	* remote.c: Use above macros throughout file.

gdbsupport/ChangeLog:

	* common-debug.h (debug_prefixed_printf_cond_nofunc): New.
	* common-debug.c (debug_prefixed_vprintf): Handle a nullptr
	func.

gdb/testsuite/ChangeLog:

	* lib/range-stepping-support.exp (exec_cmd_expect_vCont_count):
	Adjust to "set debug remote" changes.

Change-Id: Ica6dead50d3f82e855c7d763f707cef74bed9fee
 2021-01-22 12:43:27 -05:00
Simon Marchi
3ec3145c5d gdb: introduce scoped debug prints 
I spent a lot of time reading infrun debug logs recently, and I think
they could be made much more readable by being indented, to clearly see
what operation is done as part of what other operation.  In the current
format, there are no visual cues to tell where things start and end,
it's just a big flat list.  It's also difficult to understand what
caused a given operation (e.g. a call to resume_1) to be done.

To help with this, I propose to add the new scoped_debug_start_end
structure, along with a bunch of macros to make it convenient to use.

The idea of scoped_debug_start_end is simply to print a start and end
message at construction and destruction.  It also increments/decrements
a depth counter in order to make debug statements printed during this
range use some indentation.  Some care is taken to handle the fact that
debug can be turned on or off in the middle of such a range.  For
example, a "set debug foo 1" command in a breakpoint command, or a
superior GDB manually changing the debug_foo variable.

Two macros are added in gdbsupport/common-debug.h, which are helpers to
define module-specific macros:

  - scoped_debug_start_end: takes a message that is printed both at
    construction / destruction, with "start: " and "end: " prefixes.
  - scoped_debug_enter_exit: prints hard-coded "enter" and "exit"
    messages, to denote the entry and exit of a function.

I added some examples in the infrun module to give an idea of how it can
be used and what the result looks like.  The macros are in capital
letters (INFRUN_SCOPED_DEBUG_START_END and
INFRUN_SCOPED_DEBUG_ENTER_EXIT) to mimic the existing SCOPE_EXIT, but
that can be changed if you prefer something else.

Here's an excerpt of the debug
statements printed when doing "continue", where a displaced step is
started:

    [infrun] proceed: enter
      [infrun] proceed: addr=0xffffffffffffffff, signal=GDB_SIGNAL_DEFAULT
      [infrun] global_thread_step_over_chain_enqueue: enqueueing thread Thread 0x7ffff75a5640 (LWP 2289301) in global step over chain
      [infrun] start_step_over: enter
        [infrun] start_step_over: stealing global queue of threads to step, length = 1
        [infrun] start_step_over: resuming [Thread 0x7ffff75a5640 (LWP 2289301)] for step-over
        [infrun] resume_1: step=1, signal=GDB_SIGNAL_0, trap_expected=1, current thread [Thread 0x7ffff75a5640 (LWP 2289301)] at 0x5555555551bd
        [displaced] displaced_step_prepare_throw: displaced-stepping Thread 0x7ffff75a5640 (LWP 2289301) now
        [displaced] prepare: selected buffer at 0x5555555550c2
        [displaced] prepare: saved 0x5555555550c2: 1e fa 31 ed 49 89 d1 5e 48 89 e2 48 83 e4 f0 50
        [displaced] amd64_displaced_step_copy_insn: copy 0x5555555551bd->0x5555555550c2: c7 45 fc 00 00 00 00 eb 13 8b 05 d4 2e 00 00 83
        [displaced] displaced_step_prepare_throw: prepared successfully thread=Thread 0x7ffff75a5640 (LWP 2289301), original_pc=0x5555555551bd, displaced_pc=0x5555555550c2
        [displaced] resume_1: run 0x5555555550c2: c7 45 fc 00
        [infrun] infrun_async: enable=1
        [infrun] prepare_to_wait: prepare_to_wait
        [infrun] start_step_over: [Thread 0x7ffff75a5640 (LWP 2289301)] was resumed.
        [infrun] operator(): step-over queue now empty
      [infrun] start_step_over: exit
      [infrun] proceed: start: resuming threads, all-stop-on-top-of-non-stop
        [infrun] proceed: resuming Thread 0x7ffff7da7740 (LWP 2289296)
        [infrun] resume_1: step=0, signal=GDB_SIGNAL_0, trap_expected=0, current thread [Thread 0x7ffff7da7740 (LWP 2289296)] at 0x7ffff7f7d9b7
        [infrun] prepare_to_wait: prepare_to_wait
        [infrun] proceed: resuming Thread 0x7ffff7da6640 (LWP 2289300)
        [infrun] resume_1: thread Thread 0x7ffff7da6640 (LWP 2289300) has pending wait status status->kind = stopped, signal = GDB_SIGNAL_TRAP (currently_stepping=0).
        [infrun] prepare_to_wait: prepare_to_wait
        [infrun] proceed: [Thread 0x7ffff75a5640 (LWP 2289301)] resumed
        [infrun] proceed: resuming Thread 0x7ffff6da4640 (LWP 2289302)
        [infrun] resume_1: thread Thread 0x7ffff6da4640 (LWP 2289302) has pending wait status status->kind = stopped, signal = GDB_SIGNAL_TRAP (currently_stepping=0).
        [infrun] prepare_to_wait: prepare_to_wait
      [infrun] proceed: end: resuming threads, all-stop-on-top-of-non-stop
    [infrun] proceed: exit

We can easily see where the call to `proceed` starts and end.  We can
also see why there are a bunch of resume_1 calls, it's because we are
resuming threads, emulating all-stop on top of a non-stop target.

We also see that debug statements nest well with other modules that have
been migrated to use the "new" debug statement helpers (because they all
use debug_prefixed_vprintf in the end.  I think this is desirable, for
example we could see the debug statements about reading the DWARF info
of a library nested under the debug statements about loading that
library.

Of course, modules that haven't been migrated to use the "new" helpers
will still print without indentations.  This will be one good reason to
migrate them.

I think the runtime cost (when debug statements are disabled) of this is
reasonable, given the improvement in readability.  There is the cost of
the conditionals (like standard debug statements), one more condition
(if (m_must_decrement_print_depth)) and the cost of constructing a stack
object, which means copying a fews pointers.

Adding the print in fetch_inferior_event breaks some tests that use "set
debug infrun", because it prints a debug statement after the prompt.  I
adapted these tests to cope with it, by using the "-prompt" switch of
gdb_test_multiple to as if this debug statement is part of the expected
prompt.  It's unfortunate that we have to do this, but I think the debug
print is useful, and I don't want a few tests to get in the way of
adding good debug output.

gdbsupport/ChangeLog:

	* common-debug.h (debug_print_depth): New.
	(struct scoped_debug_start_end): New.
	(scoped_debug_start_end): New.
	(scoped_debug_enter_exit): New.
	* common-debug.cc (debug_prefixed_vprintf): Print indentation.

gdb/ChangeLog:

	* debug.c (debug_print_depth): New.
	* infrun.h (INFRUN_SCOPED_DEBUG_START_END): New.
	(INFRUN_SCOPED_DEBUG_ENTER_EXIT): New.
	* infrun.c (start_step_over): Use
	INFRUN_SCOPED_DEBUG_ENTER_EXIT.
	(proceed): Use INFRUN_SCOPED_DEBUG_ENTER_EXIT and
	INFRUN_SCOPED_DEBUG_START_END.
	(fetch_inferior_event): Use INFRUN_SCOPED_DEBUG_ENTER_EXIT.

gdbserver/ChangeLog:

	* debug.cc (debug_print_depth): New.

gdb/testsuite/ChangeLog:

        * gdb.base/ui-redirect.exp: Expect infrun debug print after
	prompt.
        * gdb.threads/ia64-sigill.exp: Likewise.
        * gdb.threads/watchthreads-reorder.exp: Likewise.

Change-Id: I7c3805e6487807aa63a1bae318876a0c69dce949
 2021-01-04 12:00:54 -05:00
Joel Brobecker
3666a04883 Update copyright year range in all GDB files 
This commits the result of running gdb/copyright.py as per our Start
of New Year procedure...

gdb/ChangeLog

        Update copyright year range in copyright header of all GDB files.
 2021-01-01 12:12:21 +04:00
Simon Marchi
74b773fcd6 gdb: factor out debug_prefixed_printf_cond 
The same pattern happens often to define a "debug_printf" macro:

    #define displaced_debug_printf(fmt, ...) \
      do \
        { \
          if (debug_displaced) \
    	debug_prefixed_printf ("displaced", __func__, fmt, ##__VA_ARGS__); \
        } \
      while (0)

Move this pattern behind a helper macro, debug_prefixed_printf_cond and
update the existing macros to use it.

gdb/ChangeLog:

	* displaced-stepping.h (displaced_debug_printf): Use
	debug_prefixed_printf_cond.
	* dwarf2/read.c (dwarf_read_debug_printf): Likewise.
	(dwarf_read_debug_printf_v): Likewise.
	* infrun.h (infrun_debug_printf): Likewise.
	* linux-nat.c (linux_nat_debug_printf): Likewise.

gdbsupport/ChangeLog:

	* common-debug.h (debug_prefixed_printf_cond): New.
	* event-loop.h (event_loop_debug_printf): Use
	debug_prefixed_printf_cond.

Change-Id: I1ff48b98b8d1cc405d1c7e8da8ceadf4e3a17f99
 2020-12-11 14:01:12 -05:00
Simon Marchi
17417fb0ec gdb, gdbsupport: add debug_prefixed_printf, remove boilerplate functions 
The *_debug_print_1 functions are all very similar, the only difference
being the subsystem name.  Remove them all and make the logging macros
use a new debug_prefixed_printf function directly.

gdb/ChangeLog:

	* infrun.c (infrun_debug_printf_1): Remove.
	(displaced_debug_printf_1): Remove.
	(stop_all_threads): Use debug_prefixed_printf.
	* infrun.h (infrun_debug_printf_1): Remove.
	(infrun_debug_printf): Use debug_prefixed_printf.
	(displaced_debug_printf_1): Remove.
	(displaced_debug_printf): Use debug_prefixed_printf.
	* linux-nat.c (linux_nat_debug_printf_1): Remove.
	(linux_nat_debug_printf): Use debug_prefixed_printf.

gdbsupport/ChangeLog:

	* common-debug.cc (debug_prefixed_printf): New.
	* common-debug.h (debug_prefixed_printf): New declaration.
	* event-loop.cc (event_loop_debug_printf_1): Remove.
	* event-loop.h (event_loop_debug_printf_1): Remove.
	(event_loop_debug_printf): Use debug_prefixed_printf.
	(event_loop_ui_debug_printf): Use debug_prefixed_printf.

Change-Id: Ib323087c7257f0060121d302055c41eb64aa60c6
 2020-10-31 09:15:13 -04:00
Simon Marchi
ba98841943 gdb: move debug_prefixed_vprintf here 
The following patch needs to output debug prints from gdbsupport code.
Move debug_prefixed_vprintf so that it is possible to use it from
gdbsupport.

gdb/ChangeLog:

	* debug.c (debug_prefixed_vprintf): Move to gdbsupport.
	* debug.h: Remove.
	* infrun.c: Include gdbsupport/common-debug.h.
	* linux-nat.c: Likewise.

gdbsupport/ChangeLog:

	* common-debug.cc (debug_prefixed_vprintf): Move here.
	* common-debug.h (debug_prefixed_vprintf): Move here.

Change-Id: I5170065fc10a7a49c0f1bba67c691decb2cf3bcb
 2020-10-02 14:47:26 -04:00
Tom Tromey
01027315f5 Move gdbsupport to the top level 
This patch moves the gdbsupport directory to the top level.  This is
the next step in the ongoing project to move gdbserver to the top
level.

The bulk of this patch was created by "git mv gdb/gdbsupport gdbsupport".

This patch then adds a build system to gdbsupport and wires it into
the top level.  Then it changes gdb to use the top-level build.

gdbserver, on the other hand, is not yet changed.  It still does its
own build of gdbsupport.

ChangeLog
2020-01-14  Tom Tromey  <tom@tromey.com>

	* src-release.sh (GDB_SUPPORT_DIRS): Add gdbsupport.
	* MAINTAINERS: Add gdbsupport.
	* configure: Rebuild.
	* configure.ac (configdirs): Add gdbsupport.
	* gdbsupport: New directory, move from gdb/gdbsupport.
	* Makefile.def (host_modules, dependencies): Add gnulib.
	* Makefile.in: Rebuild.

gdb/ChangeLog
2020-01-14  Tom Tromey  <tom@tromey.com>

	* nat/x86-linux-dregs.c: Include configh.h.
	* nat/linux-ptrace.c: Include configh.h.
	* nat/linux-btrace.c: Include configh.h.
	* defs.h: Include config.h, bfd.h.
	* configure.ac: Don't source common.host.
	(CONFIG_OBS, CONFIG_SRCS): Remove gdbsupport files.
	* configure: Rebuild.
	* acinclude.m4: Update path.
	* Makefile.in (SUPPORT, LIBSUPPORT, INCSUPPORT): New variables.
	(CONFIG_SRC_SUBDIR): Remove gdbsupport.
	(INTERNAL_CFLAGS_BASE): Add INCSUPPORT.
	(CLIBS): Add LIBSUPPORT.
	(CDEPS): Likewise.
	(COMMON_SFILES): Remove gdbsupport files.
	(HFILES_NO_SRCDIR): Likewise.
	(stamp-version): Update path to create-version.sh.
	(ALLDEPFILES): Remove gdbsupport files.

gdb/gdbserver/ChangeLog
2020-01-14  Tom Tromey  <tom@tromey.com>

	* server.h: Include config.h.
	* gdbreplay.c: Include config.h.
	* configure: Rebuild.
	* configure.ac: Don't source common.host.
	* acinclude.m4: Update path.
	* Makefile.in (INCSUPPORT): New variable.
	(INCLUDE_CFLAGS): Add INCSUPPORT.
	(SFILES): Update paths.
	(version-generated.c): Update path to create-version.sh.
	(gdbsupport/%-ipa.o, gdbsupport/%.o): Update paths.

gdbsupport/ChangeLog
2020-01-14  Tom Tromey  <tom@tromey.com>

	* common-defs.h: Add GDBSERVER case.  Update includes.
	* acinclude.m4, aclocal.m4, config.in, configure, configure.ac,
	Makefile.am, Makefile.in, README: New files.
	* Moved from ../gdb/gdbsupport/

Change-Id: I07632e7798635c1bab389bf885971e584fb4bb78
 2020-01-14 16:25:02 -07:00