Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Tom de Vries <tdevries@suse.de>
To: Guinevere Larsen <guinevere@redhat.com>, gdb-patches@sourceware.org
Subject: Re: [PATCH] gdb/testsuite: add test for memory requirements of gcore
Date: Mon, 24 Feb 2025 16:24:58 +0100	[thread overview]
Message-ID: <5f6b3d35-bb54-45d3-9256-4c8714b9fc22@suse.de> (raw)
In-Reply-To: <20250224140407.252987-1-guinevere@redhat.com>

On 2/24/25 15:04, Guinevere Larsen wrote:
> For a long time, Fedora has been carrying an out-of-tree patch with a
> similar test to the one proposed in this patch, that ensures that the
> memory requirements don't grow with the inferior's memory. It's been
> so long that the context for why this test exists has been lost, but
> it looked like it could be interesting for upstream.
> 
> The test runs twice, once with the inferior allocating 4Mb of memory,
> and the other allocating 64Mb. My plan was to find the rate at which
> things increase based on inferior size, and have that tested to ensure
> we're not growing that requirement accidentally, but my testing
> actually showed memory requirements going down as the inferior increases,
> so instead I just hardcoded that we need less than 2Mb for the command,
> and it can be tweaked later if necessary.

I tried out the test-case, and it passed for me, but after doing:
...
$ echo "1" | sudo tee /proc/sys/kernel/yama/ptrace_scope
...
I get:
...
(gdb) attach 29408^M
Attaching to program: 
/data/vries/gdb/leap-15-6/build/gdb/testsuite/outputs/gdb.base/gcore-memory-usage/gcore-memory-usage, 
process 29408^M
ptrace: Operation not permitted.^M
(gdb) PASS: gdb.base/gcore-memory-usage.exp: 4 Mb: attach 29408
PASS: gdb.base/gcore-memory-usage.exp: 4 Mb: before: Managed to read the 
memory usage
gcore 
/data/vries/gdb/leap-15-6/build/gdb/testsuite/outputs/gdb.base/gcore-memory-usage/gcore-memory-usage.core^M
You can't do that without a process to debug.^M
(gdb) FAIL: gdb.base/gcore-memory-usage.exp: 4 Mb: create the corefile
...

So, this is missing "require can_spawn_for_attach".

Thanks,
- Tom

> ---
>   gdb/testsuite/gdb.base/gcore-memory-usage.c   | 37 +++++++
>   gdb/testsuite/gdb.base/gcore-memory-usage.exp | 97 +++++++++++++++++++
>   2 files changed, 134 insertions(+)
>   create mode 100644 gdb/testsuite/gdb.base/gcore-memory-usage.c
>   create mode 100644 gdb/testsuite/gdb.base/gcore-memory-usage.exp
> 
> diff --git a/gdb/testsuite/gdb.base/gcore-memory-usage.c b/gdb/testsuite/gdb.base/gcore-memory-usage.c
> new file mode 100644
> index 00000000000..514fdb905cf
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/gcore-memory-usage.c
> @@ -0,0 +1,37 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> +   Copyright 2025 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>
> +#include <stdlib.h>
> +#include <unistd.h>
> +
> +int main(int argc, char **argv)
> +{
> +  /* Use argv to calculate how many megabytes to allocate.
> +     Do this to see how much gcore memory usage increases
> +     based on inferior dynamic memory.  */
> +  int megs = atoi (argv[1]);
> +  char *p = malloc (megs * 1024 * 1024);
> +
> +  /* Wait a long time so GDB can attach to this.  */
> +  sleep (10);
> +
> +  /* Let's be nice citizens :).  */
> +  free (p);
> +  return 0;
> +}
> diff --git a/gdb/testsuite/gdb.base/gcore-memory-usage.exp b/gdb/testsuite/gdb.base/gcore-memory-usage.exp
> new file mode 100644
> index 00000000000..cdb6d73c11b
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/gcore-memory-usage.exp
> @@ -0,0 +1,97 @@
> +# Copyright (C) 2025 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 gcore doesn't end up using excessive memory.
> +
> +require {istarget "*-*-linux*"}
> +
> +standard_testfile
> +
> +if {[build_executable "failed to prepare" $testfile $srcfile {debug}] == -1} {
> +    return -1
> +}
> +
> +# Read the proc_pid_status page, to find how much memory the given
> +# PID is using.  This is meant to be used to find the
> +# memory usage for the GDB in this test.
> +# Returns memory usage in Kb, or -1 if it couldn't be found.
> +proc get_mem_usage {pid prefix} {
> +    set fd [open "/proc/$pid/status"]
> +    set memory -1
> +    while {[gets $fd line] != -1} {
> +	if {[regexp {VmSize:\s*([0-9]+) kB} $line full mem]} {
> +	    set memory $mem
> +	    break
> +	}
> +    }
> +    close $fd
> +
> +    gdb_assert {$memory != -1} "$prefix: Managed to read the memory usage"
> +
> +    return $memory
> +}
> +
> +# This proc restarts GDB, runs the inferior with the desired
> +# amount of memory, then checks how much memory is necessary
> +# to run the gcore command.  It will return -1 if the gcore
> +# command fails, 0 otherwise.
> +proc run_test {megs} {
> +    with_test_prefix "$megs Mb" {
> +	clean_restart $::testfile
> +
> +	set bin [standard_output_file $::testfile]
> +	set corefile [standard_output_file "${::testfile}.core"]
> +	set inferior_pid [eval exec $bin $megs &]
> +	set gdb_pid [exp_pid -i [board_info host fileid]]
> +
> +	# wait for memory allocation to finish.
> +	sleep 1
> +
> +	# Get the important info.
> +	gdb_test "attach $inferior_pid" "Attaching to.*"
> +	set mem_before [get_mem_usage $gdb_pid before]
> +	if {![gdb_gcore_cmd $corefile "create the corefile"]} {
> +	    return -1
> +	}
> +	set mem_after [get_mem_usage $gdb_pid after]
> +
> +	# Do the main part of the test: How much is the memory
> +	# usage of GDB going to grow after using the gcore command.
> +	set diff_k [expr $mem_after - $mem_before]
> +	set diff [expr $diff_k/1024]
> +	verbose -log "The gcore command used $diff Mb ($diff_k Kb)"
> +	# The original plan was to compare to a multiple of MEGS
> +	# but since the requirements don't seem to go up as the
> +	# inferior allocated more memory, we instead just hardcode
> +	# 2 megs, since sometimes 1 is used.
> +	gdb_assert {$diff < 2} \
> +	    "gdb did not use as much memory as the inferior"
> +
> +	# Kill the inferior so we don't have to wait until the process
> +	# finishes.
> +	gdb_test "signal SIGKILL" ".*The program no longer exists."
> +    }
> +    return 0
> +}
> +
> +# If we couldn't create the first corefile, there's no point
> +# in running the second part of the test.
> +if {[run_test 4] != 0} {
> +    return
> +}
> +# Surprisingly enough, the larger inferior doesn't seem to use
> +# any extra memory, it usually uses less memory.  Which is good,
> +# it means our memory requirements aren't growing with the inferior.
> +run_test 64


  reply	other threads:[~2025-02-24 15:25 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-24 14:04 Guinevere Larsen
2025-02-24 15:24 ` Tom de Vries [this message]
2025-02-24 19:47   ` Guinevere Larsen
2025-02-24 18:21 ` Andrew Burgess
2025-02-24 20:00   ` Guinevere Larsen
2025-02-24 20:18     ` Guinevere Larsen
2025-02-25 13:45 ` [PATCH v2] " Guinevere Larsen
2025-02-26 17:50   ` [PATCH v3] " Guinevere Larsen
2025-03-04 18:19     ` Tom Tromey
2025-03-05 20:03       ` Guinevere Larsen
2025-03-05 20:21         ` Tom Tromey
2025-03-05 20:39           ` Guinevere Larsen
2025-03-11 16:14     ` Simon Marchi
2025-03-11 16:43       ` 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=5f6b3d35-bb54-45d3-9256-4c8714b9fc22@suse.de \
    --to=tdevries@suse.de \
    --cc=gdb-patches@sourceware.org \
    --cc=guinevere@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