From: Mark Kettenis <mark.kettenis@xs4all.nl>
To: jan.kratochvil@redhat.com
Cc: drow@false.org, gdb-patches@sourceware.org
Subject: Re: [patch] Fix `return' of long/long-long results with no debuginfo
Date: Sat, 21 Feb 2009 13:44:00 -0000 [thread overview]
Message-ID: <200902211054.n1LAsjYv013808@brahms.sibelius.xs4all.nl> (raw)
In-Reply-To: <20090221000409.GA8771@host0.dyn.jankratochvil.net> (message from Jan Kratochvil on Sat, 21 Feb 2009 01:04:09 +0100)
> Date: Sat, 21 Feb 2009 01:04:09 +0100
> From: Jan Kratochvil <jan.kratochvil@redhat.com>
>
> On Wed, 11 Feb 2009 23:44:21 +0100, Mark Kettenis wrote:
> > Thinking a bit more of this now, things all depend on the calling
> > convention. I'm not convinced casting to `long long' is safe in all
> > cases, especially on 32-bit big-endian machines. It really might do
> > the wrong thing there, exposing garbage or the wrong 32 bits of the
> > 64-bit value.
>
> OK to check-in this patch?
Hell no! What is all that extra code about?
> gdb/
> 2009-02-20 Jan Kratochvil <jan.kratochvil@redhat.com>
>
> * stack.c (return_extended_value_cast, return_extended_value): New.
> (return_command): Returned type for functions with no debug info has
> now at least the width of the type of user specified expression while
> it tries to use the largest compatible width.
>
> gdb/testsuite/
> 2009-02-20 Jan Kratochvil <jan.kratochvil@redhat.com>
>
> * gdb.base/return-nodebug.exp, gdb.base/return-nodebug.S: New.
>
> --- gdb/stack.c 11 Feb 2009 16:07:28 -0000 1.185
> +++ gdb/stack.c 20 Feb 2009 21:47:34 -0000
> @@ -1777,7 +1777,75 @@ down_command (char *count_exp, int from_
> down_silently_base (count_exp);
> print_stack_frame (get_selected_frame (NULL), 1, SRC_AND_LOC);
> }
> -\f
> +
> +/* Verify whether if RETURN_VALUE width gets extended to EXT_TYPE it will still
> + be the same value after reading it back using the RETURN_VALUE type. */
> +
> +static int
> +return_extended_value_cast (struct value *return_value, struct type *ext_type)
> +{
> + struct regcache *current_regcache = get_current_regcache ();
> + struct gdbarch *gdbarch = get_regcache_arch (current_regcache);
> + struct type *return_type = value_type (return_value);
> + struct value *ext_value, *compare_value;
> +
> + if (gdbarch_return_value (gdbarch, NULL, ext_type, NULL, NULL, NULL)
> + != RETURN_VALUE_REGISTER_CONVENTION)
> + return 0;
> +
> + ext_value = value_cast (ext_type, return_value);
> + gdbarch_return_value (gdbarch, NULL, ext_type,
> + current_regcache, NULL /*read*/,
> + value_contents (ext_value) /*write*/);
> +
> + compare_value = allocate_value (return_type);
> + gdbarch_return_value (gdbarch, NULL, return_type, current_regcache,
> + value_contents_raw (compare_value) /*read*/,
> + NULL /*write*/);
> +
> + return value_equal (return_value, compare_value);
> +}
> +
> +/* Set RETURN_VALUE extended to the largest type width which will still be the
> + same value after reading it back using the RETURN_VALUE type. */
> +
> +static void
> +return_extended_value (struct value *return_value)
> +{
> + struct type *return_type = value_type (return_value);
> + struct regcache *current_regcache = get_current_regcache ();
> + struct gdbarch *gdbarch = get_regcache_arch (current_regcache);
> + const struct builtin_type *builtins = builtin_type (gdbarch);
> + struct type **extp, *ext_tab[] =
> + {
> + builtins->builtin_long_long,
> + builtins->builtin_long,
> + return_type
> + };
> + unsigned width_tried = 0;
> +
> + for (extp = ext_tab; extp < ext_tab + ARRAY_SIZE (ext_tab); extp++)
> + {
> + struct type *ext_type = *extp;
> +
> + /* Do not retry extension to the integer of an already tried width. */
> + if (ext_type != return_type && width_tried == TYPE_LENGTH (ext_type))
> + continue;
> +
> + /* Do not extend to non-original smaller (or the same size) type. */
> + if (ext_type != return_type
> + && TYPE_LENGTH (ext_type) <= TYPE_LENGTH (return_type))
> + continue;
> +
> + gdb_assert (TYPE_LENGTH (return_type) <= TYPE_LENGTH (ext_type));
> + width_tried = TYPE_LENGTH (ext_type);
> + if (return_extended_value_cast (return_value, ext_type))
> + break;
> + }
> +
> + /* Ensure at least the attempt with original RETURN_TYPE was successful. */
> + gdb_assert (extp < ext_tab + ARRAY_SIZE (ext_tab));
> +}
>
> void
> return_command (char *retval_exp, int from_tty)
> @@ -1806,6 +1874,8 @@ return_command (char *retval_exp, int fr
> the cast fail, this call throws an error. */
> if (thisfun != NULL)
> return_type = TYPE_TARGET_TYPE (SYMBOL_TYPE (thisfun));
> + else
> + return_type = value_type (return_value);
> if (return_type == NULL)
> return_type = builtin_type (get_frame_arch (thisframe))->builtin_int;
> CHECK_TYPEDEF (return_type);
> @@ -1862,9 +1932,13 @@ If you continue, the return value that y
> gdb_assert (gdbarch_return_value (gdbarch, func_type, return_type, NULL,
> NULL, NULL)
> == RETURN_VALUE_REGISTER_CONVENTION);
> - gdbarch_return_value (gdbarch, func_type, return_type,
> - get_current_regcache (), NULL /*read*/,
> - value_contents (return_value) /*write*/);
> +
> + if (thisfun != NULL)
> + gdbarch_return_value (gdbarch, func_type, return_type,
> + get_current_regcache (), NULL /*read*/,
> + value_contents (return_value) /*write*/);
> + else
> + return_extended_value (return_value);
> }
>
> /* If we are at the end of a call dummy now, pop the dummy frame
> --- /dev/null 1 Jan 1970 00:00:00 -0000
> +++ gdb/testsuite/gdb.base/return-nodebug.c 20 Feb 2009 21:47:34 -0000
> @@ -0,0 +1,49 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> + 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/>. */
> +
> +#include <stdio.h>
> +
> +static TYPE
> +init (void)
> +{
> + return 0;
> +}
> +
> +static TYPE
> +func (void)
> +{
> + return 31;
> +}
> +
> +static void
> +marker (void)
> +{
> +}
> +
> +int
> +main (void)
> +{
> + /* Preinitialize registers to 0 to avoid false PASS by leftover garbage. */
> + init ();
> +
> + printf ("result=" FORMAT "\n", CAST func ());
> +
> + /* Cannot `next' with no debug info. */
> + marker ();
> +
> + return 0;
> +}
> --- /dev/null 1 Jan 1970 00:00:00 -0000
> +++ gdb/testsuite/gdb.base/return-nodebug.exp 20 Feb 2009 21:47:34 -0000
> @@ -0,0 +1,54 @@
> +# Copyright (C) 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/>.
> +
> +proc do_test {type} {
> + set typenospace [string map {{ } -} $type]
> +
> + global pf_prefix
> + set old_prefix $pf_prefix
> + lappend pf_prefix "$typenospace:"
> +
> + if {[runto "func"]} {
> + # Test return from a function with no debug info (symbol; still it may
> + # have a minimal-symbol) if it does not crash.
> + gdb_test "return -1" "#0 .* main \\(.*" \
> + "return from function with no debug info" \
> + "Make selected stack frame return now\\? \\(y or n\\) " "y"
> +
> + # And if it returned the full width of the result.
> + gdb_test "adv marker" "\r\nresult=-1\r\n.* in marker \\(.*" \
> + "full width of the returned result"
> + }
> +
> + set pf_prefix $old_prefix
> +}
> +
> +foreach case {{{signed char} %d (int)} \
> + {{short} %d (int)} \
> + {{int} %d} \
> + {{long} %ld} \
> + {{long long} %lld}} {
> + set type [lindex $case 0]
> + set format [lindex $case 1]
> + set cast [lindex $case 2]
> +
> + set typeesc [string map {{ } {\ }} $type]
> + set typenospace [string map {{ } -} $type]
> +
> + if {[prepare_for_testing return-nodebug.exp "return-nodebug-$typenospace" "return-nodebug.c" \
> + [list "additional_flags=-DFORMAT=\"$format\" -DTYPE=$typeesc -DCAST=$cast"]] == 0} {
> + do_test $type
> + }
> +}
>
next prev parent reply other threads:[~2009-02-21 10:55 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-02-11 19:45 Jan Kratochvil
2009-02-11 20:40 ` Mark Kettenis
2009-02-11 20:50 ` Daniel Jacobowitz
2009-02-11 21:23 ` Mark Kettenis
2009-02-11 21:47 ` Jan Kratochvil
2009-02-11 21:58 ` Mark Kettenis
2009-02-11 22:08 ` Jan Kratochvil
2009-02-11 22:38 ` Mark Kettenis
2009-02-11 22:50 ` Jan Kratochvil
2009-03-03 18:10 ` Tom Tromey
2009-03-04 21:29 ` Mark Kettenis
2009-03-05 0:15 ` Tom Tromey
2009-03-09 1:55 ` Jan Kratochvil
2009-03-09 22:38 ` Mark Kettenis
2009-03-11 16:49 ` Joel Brobecker
2009-03-11 20:23 ` Tom Tromey
2009-03-11 21:48 ` Joel Brobecker
2009-03-13 19:50 ` Jan Kratochvil
2009-03-13 23:00 ` Joel Brobecker
2009-03-14 10:45 ` Eli Zaretskii
2009-03-14 22:09 ` Jan Kratochvil
2009-03-14 22:18 ` Eli Zaretskii
2009-03-15 9:22 ` Joel Brobecker
2009-03-15 18:15 ` Jan Kratochvil
2009-03-18 4:36 ` Pedro Alves
2009-03-18 14:44 ` Joel Brobecker
2009-03-18 15:39 ` Jan Kratochvil
2009-03-18 15:46 ` Pedro Alves
2009-02-11 22:44 ` Mark Kettenis
2009-02-12 9:41 ` Jan Kratochvil
2009-02-12 14:36 ` Pierre Muller
2009-02-12 14:44 ` Jan Kratochvil
2009-02-14 21:59 ` Mark Kettenis
2009-02-21 12:47 ` Jan Kratochvil
2009-02-21 13:44 ` Mark Kettenis [this message]
2009-02-21 15:58 ` Jan Kratochvil
2009-02-21 16:21 ` Mark Kettenis
2009-02-21 16:32 ` Jan Kratochvil
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=200902211054.n1LAsjYv013808@brahms.sibelius.xs4all.nl \
--to=mark.kettenis@xs4all.nl \
--cc=drow@false.org \
--cc=gdb-patches@sourceware.org \
--cc=jan.kratochvil@redhat.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