From: Hui Zhu <teawater@gmail.com>
To: Michael Snyder <msnyder@vmware.com>
Cc: gdb-patches@sourceware.org, marc.khouzam@ericsson.com
Subject: Re: [RFA] Reverse-step/next and shared libraries (with testsuite)
Date: Mon, 22 Jun 2009 07:45:00 -0000 [thread overview]
Message-ID: <daef60380906220045t7bac6addqaa102b54e3d465a4@mail.gmail.com> (raw)
In-Reply-To: <4A3EBDE9.9090002@vmware.com>
I try this patch in SLED 10. Everything is OK.
Thanks,
Hui
On Mon, Jun 22, 2009 at 07:10, Michael Snyder<msnyder@vmware.com> wrote:
> OK, this is a bit of a rewrite for the infrun code that
> handles reverse stepping into/over shared library functions
> and trampolines. I believe we now correctly handle various
> combinations of ...
> * reverse-step, reverse-next
> * trampolines, solib-dynsym-resolver
> * resolved and unresolved symbols
> * debuggable (user) and undebuggable (eg. libc) functions
>
> This also includes the first testsuite test case for reverse
> debugging. If this one's accepted, I've got more queued up.
>
> Hui, this patch will replace your patch
> http://sourceware.org/ml/gdb-patches/2009-03/msg00005.html,
> so I'll appreciate your review.
>
> Marc, this patch should fix all of your solib step/next bugs.
>
> No regressions on linux-i686.
>
>
> 2009-06-21 Michael Snyder <msnyder@vmware.com>
>
> * infrun.c (handle_inferior_event): Improve handling of
> reverse-stepping into and over shared library functions.
>
> 2009-06-21 Michael Snyder <msnyder@vmware.com>
>
> * gdb.base/solib-reverse.exp: New file.
> * gdb.base/solib-reverse.c: New file.
>
> Index: infrun.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/infrun.c,v
> retrieving revision 1.392
> diff -u -p -r1.392 infrun.c
> --- infrun.c 18 Jun 2009 23:55:57 -0000 1.392
> +++ infrun.c 21 Jun 2009 22:50:22 -0000
> @@ -3639,7 +3639,7 @@ infrun: not switching back to stepped th
> if (stop_pc >= ecs->event_thread->step_range_start
> && stop_pc < ecs->event_thread->step_range_end
> && (execution_direction != EXEC_REVERSE
> - || frame_id_eq (get_frame_id (get_current_frame ()),
> + || frame_id_eq (get_frame_id (frame),
> ecs->event_thread->step_frame_id)))
> {
> if (debug_infrun)
> @@ -3667,10 +3667,19 @@ infrun: not switching back to stepped th
> /* We stepped out of the stepping range. */
>
> /* If we are stepping at the source level and entered the runtime
> - loader dynamic symbol resolution code, we keep on single stepping
> - until we exit the run time loader code and reach the callee's
> - address. */
> - if (ecs->event_thread->step_over_calls == STEP_OVER_UNDEBUGGABLE
> + loader dynamic symbol resolution code...
> +
> + EXEC_FORWARD: we keep on single stepping until we exit the run
> + time loader code and reach the callee's address.
> +
> + EXEC_REVERSE: we've already executed the callee (backward), and
> + the runtime loader code is handled just like any other
> + undebuggable function call. Now we need only keep stepping
> + backward through the trampoline code, and that's handled further
> + down, so there is nothing for us to do here. */
> +
> + if (execution_direction != EXEC_REVERSE
> + && ecs->event_thread->step_over_calls == STEP_OVER_UNDEBUGGABLE
> && in_solib_dynsym_resolve_code (stop_pc))
> {
> CORE_ADDR pc_after_resolver =
> @@ -3740,12 +3749,29 @@ infrun: not switching back to stepped th
> /* Also, maybe we just did a "nexti" inside a prolog, so we
> thought it was a subroutine call but it was not. Stop as
> well. FENN */
> + /* And this works the same backward as frontward. MVS */
> ecs->event_thread->stop_step = 1;
> print_stop_reason (END_STEPPING_RANGE, 0);
> stop_stepping (ecs);
> return;
> }
>
> + /* Reverse stepping through solib trampolines. */
> +
> + if (execution_direction == EXEC_REVERSE
> + && (gdbarch_skip_trampoline_code (gdbarch, frame, stop_pc)
> + || (ecs->stop_func_start == 0
> + && in_solib_dynsym_resolve_code (stop_pc))))
> + {
> + /* Any solib trampoline code can be handled in reverse
> + by simply continuing to single-step. We have already
> + executed the solib function (backwards), and a few
> + steps will take us back through the trampoline to the
> + caller. */
> + keep_going (ecs);
> + return;
> + }
> +
> if (ecs->event_thread->step_over_calls == STEP_OVER_ALL)
> {
> /* We're doing a "next".
> @@ -3763,35 +3789,10 @@ infrun: not switching back to stepped th
> {
> struct symtab_and_line sr_sal;
>
> - if (ecs->stop_func_start == 0
> - && in_solib_dynsym_resolve_code (stop_pc))
> - {
> - /* Stepped into runtime loader dynamic symbol
> - resolution code. Since we're in reverse,
> - we have already backed up through the runtime
> - loader and the dynamic function. This is just
> - the trampoline (jump table).
> -
> - Just keep stepping, we'll soon be home.
> - */
> - keep_going (ecs);
> - return;
> - }
> - if (gdbarch_skip_trampoline_code(current_gdbarch,
> - get_current_frame (),
> - stop_pc))
> - {
> - /* We are in a function call trampoline.
> - Keep stepping backward to get to the caller. */
> - ecs->event_thread->stepping_over_breakpoint = 1;
> - }
> - else
> - {
> - /* Normal function call return (static or dynamic). */
> - init_sal (&sr_sal);
> - sr_sal.pc = ecs->stop_func_start;
> - insert_step_resume_breakpoint_at_sal (sr_sal,
> null_frame_id);
> - }
> + /* Normal function call return (static or dynamic). */
> + init_sal (&sr_sal);
> + sr_sal.pc = ecs->stop_func_start;
> + insert_step_resume_breakpoint_at_sal (sr_sal, null_frame_id);
> }
> else
> insert_step_resume_breakpoint_at_caller (frame);
> Index: testsuite/gdb.base/solib-reverse.c
> ===================================================================
> RCS file: testsuite/gdb.base/solib-reverse.c
> diff -N testsuite/gdb.base/solib-reverse.c
> --- /dev/null 1 Jan 1970 00:00:00 -0000
> +++ testsuite/gdb.base/solib-reverse.c 21 Jun 2009 22:54:01 -0000
> @@ -0,0 +1,26 @@
> +/* Test reverse debugging of shared libraries. */
> +
> +#include <stdio.h>
> +
> +/* Shared library function */
> +extern int shr2(int);
> +
> +int main ()
> +{
> + char* cptr = "String 1";
> + int b[2] = {5,8};
> +
> + b[0] = shr2(12); /* begin part two */
> + b[1] = shr2(17); /* middle part two */
> +
> + b[0] = 6; b[1] = 9; /* generic statement, end part two
> */
> + printf ("message 1\n"); /* printf one */
> + printf ("message 2\n"); /* printf two */
> + printf ("message 3\n"); /* printf three */
> + sleep (0); /* sleep one */
> + sleep (0); /* sleep two */
> + sleep (0); /* sleep three */
> +
> + return 0; /* end part one */
> +}
> +
> Index: testsuite/gdb.base/solib-reverse.exp
> ===================================================================
> RCS file: testsuite/gdb.base/solib-reverse.exp
> diff -N testsuite/gdb.base/solib-reverse.exp
> --- /dev/null 1 Jan 1970 00:00:00 -0000
> +++ testsuite/gdb.base/solib-reverse.exp 21 Jun 2009 22:54:01 -0000
> @@ -0,0 +1,127 @@
> +# Copyright 2009
> +# Free Software Foundation, Inc.
> +
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program. If not, see <http://www.gnu.org/licenses/>.
> +
> +# This file was written by Michael Snyder (msnyder@vmware.com).
> +
> +#
> +# Test reverse debugging with shared libraries.
> +#
> +
> +if ![target_info exists gdb,can_reverse] {
> + return
> +}
> +
> +set testfile "solib-reverse"
> +set srcfile ${testfile}.c
> +set libfile "shr2"
> +set libsrc ${libfile}.c
> +set library ${objdir}/${subdir}/${libfile}.sl
> +set binfile ${objdir}/${subdir}/${testfile}
> +
> +if [get_compiler_info ${binfile}] {
> + return -1
> +}
> +
> +if { [gdb_compile_shlib ${srcdir}/${subdir}/${libsrc} ${library} "debug"]
> != "" } {
> + untested "Could not compile shared library."
> + return -1
> +}
> +
> +set exec_opts [list debug shlib=${library}]
> +
> +if { [gdb_compile ${srcdir}/${subdir}/${srcfile} ${binfile} executable
> $exec_opts] != "" } {
> + untested "Could not compile $binfile."
> + return -1
> +}
> +
> +# Start with a fresh gdb.
> +
> +gdb_exit
> +gdb_start
> +gdb_reinitialize_dir $srcdir/$subdir
> +gdb_load ${binfile}
> +
> +
> +runto main
> +
> +if [target_info exists gdb,use_precord] {
> + # Activate process record/replay
> + gdb_test "record" "" "Turn on process record"
> + # FIXME: command ought to acknowledge, so we can test if it succeeded.
> +}
> +
> +#
> +# Test reverse-step over undebuggable solib functions.
> +#
> +
> +# Run forward past some solib function calls.
> +
> +set end_part_one [gdb_get_line_number " end part one" "$srcfile"]
> +set end_part_two [gdb_get_line_number " end part two" "$srcfile"]
> +gdb_test "until $end_part_one" " end part one.*" "run until end part one"
> +
> +gdb_test "reverse-step" " sleep three .*" "reverse-step third sleep"
> +gdb_test "reverse-step" " sleep two .*" "reverse-step second sleep"
> +gdb_test "reverse-step" " sleep one .*" \
> + "reverse-step first sleep, dynsym resolve"
> +
> +gdb_test "reverse-step" " printf three .*" "reverse-step third printf"
> +gdb_test "reverse-step" " printf two .*" "reverse-step second printf"
> +gdb_test "reverse-step" " printf one .*" \
> + "reverse-step first printf, dynsym resolve"
> +gdb_test "reverse-step" " generic statement.*" "reverse-step generic"
> +
> +
> +#
> +# Test reverse-next over undebuggable solib functions.
> +#
> +
> +# Run forward again...
> +
> +gdb_test "until $end_part_one" " end part one.*" "forward to end part one"
> +
> +gdb_test "reverse-next" " sleep three .*" "reverse-next third sleep"
> +gdb_test "reverse-next" " sleep two .*" "reverse-next second sleep"
> +gdb_test "reverse-next" " sleep one .*" \
> + "reverse-next first sleep, dynsym resolve"
> +
> +gdb_test "reverse-next" " printf three .*" "reverse-next third printf"
> +gdb_test "reverse-next" " printf two .*" "reverse-next second printf"
> +gdb_test "reverse-next" " printf one .*" \
> + "reverse-next first printf, dynsym resolve"
> +gdb_test "reverse-next" " generic statement.*" "reverse-next generic"
> +
> +
> +#
> +# Test reverse-step into debuggable solib function
> +#
> +
> +gdb_test "reverse-step" "${libsrc}.*" "reverse-step into solib function
> one"
> +gdb_test "reverse-step" "return 2.x.*" "reverse-step within solib function
> one"
> +gdb_test "reverse-step" " middle part two.*" "reverse-step back to main
> one"
> +
> +gdb_test "reverse-step" "${libsrc}.*" "reverse-step into solib function
> two"
> +gdb_test "reverse-step" "return 2.x.*" "reverse-step within solib function
> two"
> +gdb_test "reverse-step" " begin part two.*" "reverse-step back to main two"
> +
> +#
> +# Test reverse-next over debuggable solib function
> +#
> +
> +gdb_test "until $end_part_two" " end part two.*" "run until end part two"
> +
> +gdb_test "reverse-next" " middle part two.*" "reverse-next over solib
> function one"
> +gdb_test "reverse-next" " begin part two.*" "reverse-next over solib
> function two"
>
>
next prev parent reply other threads:[~2009-06-22 7:45 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-06-21 23:11 Michael Snyder
2009-06-22 7:45 ` Hui Zhu [this message]
2009-06-27 18:35 ` Michael Snyder
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=daef60380906220045t7bac6addqaa102b54e3d465a4@mail.gmail.com \
--to=teawater@gmail.com \
--cc=gdb-patches@sourceware.org \
--cc=marc.khouzam@ericsson.com \
--cc=msnyder@vmware.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox