From: "Gerlicher, Klaus" <klaus.gerlicher@intel.com>
To: Guinevere Larsen <guinevere@redhat.com>,
"gdb-patches@sourceware.org" <gdb-patches@sourceware.org>
Cc: Guinevere Larsen <blarsen@redhat.com>,
Jan Kratochvil <jan@jankratochvil.net>
Subject: RE: [PATCH v5 5/5] gdb/testsuite: Test for a backtrace through object without debuginfo
Date: Thu, 3 Oct 2024 06:58:31 +0000 [thread overview]
Message-ID: <SN7PR11MB70914F7AF260BD8EC601673FE8712@SN7PR11MB7091.namprd11.prod.outlook.com> (raw)
In-Reply-To: <20241001184235.3710608-6-guinevere@redhat.com>
Hi Guienvere,
Just one suggestion inline.
Thanks
Klaus
> -----Original Message-----
> From: Guinevere Larsen <guinevere@redhat.com>
> Sent: Tuesday, October 1, 2024 8:43 PM
> To: gdb-patches@sourceware.org
> Cc: Guinevere Larsen <blarsen@redhat.com>; Jan Kratochvil
> <jan@jankratochvil.net>
> Subject: [PATCH v5 5/5] gdb/testsuite: Test for a backtrace through object
> without debuginfo
>
> From: Guinevere Larsen <blarsen@redhat.com>
>
> Fedora has been carrying this test since back in the Project Archer
> days. A change back then caused GDB to stop being able to backtrace when
> only some of the object files had debug information. Even though the
> changed code never seems to have made its way into the main GDB project,
> I think it makes sense to bring the test along to ensure something like
> this doesn't pass unnoticed.
>
> Co-Authored-By: Jan Kratochvil <jan@jankratochvil.net>
> ---
> .../backtrace-through-cu-nodebug-caller.c | 28 ++++++
> .../backtrace-through-cu-nodebug-main.c | 32 +++++++
> .../gdb.base/backtrace-through-cu-nodebug.exp | 90
> +++++++++++++++++++
> 3 files changed, 150 insertions(+)
> create mode 100644 gdb/testsuite/gdb.base/backtrace-through-cu-
> nodebug-caller.c
> create mode 100644 gdb/testsuite/gdb.base/backtrace-through-cu-
> nodebug-main.c
> create mode 100644 gdb/testsuite/gdb.base/backtrace-through-cu-
> nodebug.exp
>
> diff --git a/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug-caller.c
> b/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug-caller.c
> new file mode 100644
> index 00000000000..3a63d72a468
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug-caller.c
> @@ -0,0 +1,28 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> + Copyright 2005-2024 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/>. */
> +
> +typedef int (*callback_t) (void);
> +
> +int
> +caller (callback_t callback)
> +{
> + /* Ensure some frame content to push away the return address. */
> + volatile const long one = 1;
> +
> + /* Modify the return value to prevent any tail-call optimization. */
> + return (*callback) () - one;
> +}
> diff --git a/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug-main.c
> b/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug-main.c
> new file mode 100644
> index 00000000000..3e7ac57a166
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug-main.c
> @@ -0,0 +1,32 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> + Copyright 2005-2024 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/>. */
> +
> +typedef int (*callback_t) (void);
> +
> +extern int caller (callback_t callback);
> +
> +int
> +callback (void)
> +{
> + return 1;
> +}
> +
> +int
> +main (void)
> +{
> + return caller (callback);
> +}
> diff --git a/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug.exp
> b/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug.exp
> new file mode 100644
> index 00000000000..7951ab41141
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/backtrace-through-cu-nodebug.exp
> @@ -0,0 +1,90 @@
> +# Copyright 2010-2024 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/>.
> +
> +# Test that GDB can generate accurate backtraces even if some of the stack
> +# trace goes through a function with no debug information.
> +
> +standard_testfile -caller.c -main.c
> +set objmainfile ${testfile}-main.o
> +set objcallerfile ${testfile}-caller.o
> +
> +# recompile the inferior with or without CFI information, then run the
> +# inferior until the point where the important test starts
> +# returns TRUE on an ERROR.
> +proc prepare_test {has_cfi} {
> + global srcdir subdir srcfile srcfile2 objmainfile objcallerfile binfile
> + if {$has_cfi} {
> + set extension "cfi"
> + if {[gdb_compile "${srcdir}/${subdir}/${srcfile}" ${objcallerfile} \
> + object [list {additional_flags=-fomit-frame-pointer \
> + -funwind-tables -fasynchronous-unwind-tables}]] != "" } {
> + untested "couldn't compile with cfi"
> + return true
> + }
> + } else {
> + set extension "no-cfi"
> + if {[gdb_compile "${srcdir}/${subdir}/${srcfile}" ${objcallerfile} \
> + object [list {additional_flags=-fomit-frame-pointer \
> + -fno-unwind-tables \
> + -fno-asynchronous-unwind-tables}]] != "" } {
> + untested "couldn't compile without cfi"
> + return true
> + }
> + }
> + if {[gdb_compile "${objmainfile} ${objcallerfile}" \
> + "${binfile}-${extension}" binfile {}] != ""} {
> + untested "couldn't link object files"
> + return true
> + }
> +
> + clean_restart "$binfile-${extension}"
> +
> + with_test_prefix "${extension}" {
> +
> + if ![runto callback] then {
> + fail "has_cfi=$has_cfi: Can't run to callback"
> + return true
> + }
> + gdb_test_no_output "maint frame-unwinder disable ARCH"
> + return false
> + }
> +}
> +
> +if {[gdb_compile "${srcdir}/${subdir}/${srcfile2}" ${objmainfile} \
> + object {debug}] != "" } {
> + untested "couldn't compile main file"
> + return
> +}
> +
> +if { [prepare_test false] } {
> + untested ${testfile}.exp
> +} else {
> + gdb_test "bt" "frame_unwind_find_by_frame failed.*" \
> + "verify unwind fail without CFI"
> +}
> +
> +if { [prepare_test true] } {
> + untested ${testfile}.exp
> +} else {
> + if { [istarget "arm*-*-*"] } {
> + setup_kfail backtrace/31950 *-*-*
> + }
> + # #0 callback () at ...
> + # #1 0x00000000004004e9 in caller ()
> + # #2 0x00000000004004cd in main () at ...
> + gdb_test "bt" \
> + "#0 +callback \[^\r\n\]+\r\n#1 \[^\r\n\]+ in caller \[^\r\n\]+\r\n#2
> \[^\r\n\]+ in main \[^\r\n\]+" \
> + "verify unwinding works for CFI without DIEs"
> +}
You could do below to reduce the line to under 80 cols, I wonder why we don't have this defined somewhere
globally like "hex" or "decimal", though it's not really used that often.
set r {[^\r\n]+}
gdb_test "bt" "#0 +callback $r+\r\n#1 $r+ in caller $r+\r\n#2> $r+ in main $r+"
> --
> 2.46.2
>
Intel Deutschland GmbH
Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Sean Fennelly, Jeffrey Schneiderman, Tiffany Doon Silva
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928
next prev parent reply other threads:[~2024-10-03 6:59 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-01 18:42 [PATCH v5 0/5] Modernize frame unwinders and add disable feature Guinevere Larsen
2024-10-01 18:42 ` [PATCH v5 1/5] gdb: make gdbarch store a vector of frame unwinders Guinevere Larsen
2024-10-02 21:49 ` Thiago Jung Bauermann
2024-10-08 17:01 ` Guinevere Larsen
2024-10-03 18:33 ` Simon Marchi
2024-10-04 18:37 ` Tom Tromey
2024-10-12 1:34 ` Thiago Jung Bauermann
2024-10-14 18:18 ` Guinevere Larsen
2024-10-17 22:53 ` Tom Tromey
2024-10-18 17:40 ` Guinevere Larsen
2024-10-17 23:41 ` Tom Tromey
2024-10-01 18:42 ` [PATCH v5 2/5] gdb: add "unwinder class" to " Guinevere Larsen
2024-10-02 22:08 ` Thiago Jung Bauermann
2024-10-03 18:46 ` Simon Marchi
2024-10-08 18:22 ` Guinevere Larsen
2024-10-08 18:37 ` Simon Marchi
2024-10-01 18:42 ` [PATCH v5 3/5] gdb: Migrate frame unwinders to use C++ classes Guinevere Larsen
2024-10-03 0:23 ` Thiago Jung Bauermann
2024-10-09 18:16 ` Guinevere Larsen
2024-10-03 20:06 ` Simon Marchi
2024-10-04 5:21 ` Simon Marchi
2024-10-10 14:10 ` Guinevere Larsen
2024-10-10 16:28 ` Simon Marchi
2024-10-09 20:00 ` Guinevere Larsen
2024-10-01 18:42 ` [PATCH v5 4/5] gdb: introduce ability to disable frame unwinders Guinevere Larsen
2024-10-02 6:10 ` Eli Zaretskii
2024-10-04 17:57 ` Guinevere Larsen
2024-10-03 2:45 ` Thiago Jung Bauermann
2024-10-08 19:23 ` Guinevere Larsen
2024-10-06 2:51 ` Simon Marchi
2024-10-09 13:32 ` Guinevere Larsen
2024-10-09 15:38 ` Simon Marchi
2024-10-01 18:42 ` [PATCH v5 5/5] gdb/testsuite: Test for a backtrace through object without debuginfo Guinevere Larsen
2024-10-03 2:47 ` Thiago Jung Bauermann
2024-10-03 6:58 ` Gerlicher, Klaus [this message]
2024-10-09 14:56 ` Guinevere Larsen
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=SN7PR11MB70914F7AF260BD8EC601673FE8712@SN7PR11MB7091.namprd11.prod.outlook.com \
--to=klaus.gerlicher@intel.com \
--cc=blarsen@redhat.com \
--cc=gdb-patches@sourceware.org \
--cc=guinevere@redhat.com \
--cc=jan@jankratochvil.net \
/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