live-patching.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Petr Mladek <pmladek@suse.com>
To: Michael Vetter <mvetter@suse.com>
Cc: linux-kselftest@vger.kernel.org, live-patching@vger.kernel.org
Subject: Re: [PATCH v2] selftests: livepatch: test livepatching a kprobed function
Date: Thu, 19 Sep 2024 09:40:22 +0200	[thread overview]
Message-ID: <ZuvVZneTT7SY76wP@pathway.suse.cz> (raw)
In-Reply-To: <20240917135059.73346-1-mvetter@suse.com>

On Tue 2024-09-17 15:50:53, Michael Vetter wrote:
> The test proves that a function that is being kprobed and uses a
> post_handler cannot be livepatched.
> 
> Only one ftrace_ops with FTRACE_OPS_FL_IPMODIFY set may be registered
> to any given function at a time.
> 
> --- /dev/null
> +++ b/tools/testing/selftests/livepatch/test-kprobe.sh
> @@ -0,0 +1,65 @@
> +#!/bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (C) 2024 SUSE
> +# Author: Michael Vetter <mvetter@suse.com>
> +
> +. $(dirname $0)/functions.sh
> +
> +MOD_LIVEPATCH=test_klp_livepatch
> +MOD_KPROBE=test_klp_kprobe
> +
> +setup_config
> +
> +# Kprobe a function and verify that we can't livepatch that same function
> +# when it uses a post_handler since only one IPMODIFY maybe be registered
> +# to any given function at a time.
> +
> +start_test "livepatch interaction with kprobed function with post_handler"
> +
> +STATE=$(cat /sys/kernel/debug/kprobes/enabled)
> +echo 1 > /sys/kernel/debug/kprobes/enabled
> +
> +load_mod $MOD_KPROBE has_post_handler=true
> +load_failing_mod $MOD_LIVEPATCH
> +unload_mod $MOD_KPROBE
[...]
> +echo $STATE > /sys/kernel/debug/kprobes/enabled

The original state won't get restored when the test fails.

A better solution would be to integrate this into setup_config()
and cleanup() in /tools/testing/selftests/livepatch/functions.sh

Also it would be great to avoid hard-coding /sys/kernel/debug.
See KLP_SYSFS_DIR="/sys/kernel/livepatch" in
/tools/testing/selftests/livepatch/functions.sh

I would replace it with:

KERNEL_SYSFS_DIR="/sys/kernel"
KLP_SYSFS_DIR="$KERNEL_SYSFS_DIR/livepatch"
DEBUG_SYSFS_DIR="$KERNEL_SYSFS_DIR/debug"
KPROBES_SYSFS_DIR="$DEBUG_SYSFS_DIR/kprobes"

It would actually look better with switched ordering:

SYSFS_KERNEL_DIR="/sys/kernel"
SYSFS_KLP_DIR="$SYSFS_KERNEL_DIR/livepatch"
SYSFS_DEBUG_DIR="$SYSFS_KERNEL_DIR/debug"
SYSFS_KPROBES_DIR="$SYSFS_DEBUG_DIR/kprobes"

But this would require renaming KLP_SYSFS_DIR everywhere in
a separate (prerequisite) patch. It would be nice but
I do not want to force you to do it ;-)

> --- a/tools/testing/selftests/livepatch/test_modules/Makefile
> +++ b/tools/testing/selftests/livepatch/test_modules/Makefile
> @@ -11,7 +11,8 @@ obj-m += test_klp_atomic_replace.o \
>  	test_klp_state2.o \
>  	test_klp_state3.o \
>  	test_klp_shadow_vars.o \
> -	test_klp_syscall.o
> +	test_klp_syscall.o \
> +	test_klp_kprobe.o

Nit: Please, put it in alphabetical order. It is not a big deal.
     The ordering helps searching. The list already is almost sorted.
     And the patch needs to be updated anyway.

>  # Ensure that KDIR exists, otherwise skip the compilation
>  modules:
> diff --git a/tools/testing/selftests/livepatch/test_modules/test_klp_kprobe.c b/tools/testing/selftests/livepatch/test_modules/test_klp_kprobe.c
> new file mode 100644
> index 000000000000..49b579ea1054
> --- /dev/null
> +++ b/tools/testing/selftests/livepatch/test_modules/test_klp_kprobe.c
> @@ -0,0 +1,38 @@
[...]
> +module_init(kprobe_init)
> +module_exit(kprobe_exit)
> +MODULE_LICENSE("GPL");
> +MODULE_AUTHOR("Michael Vetter <mvetter@suse.com>");
> +MODULE_DESCRIPTION("Livepatch test: livepatch kprobed function");

Nit: The module adds a kprobe. I would invert the logic:

MODULE_DESCRIPTION("Livepatch test: kprobe livepatched function");

Best Regards,
Petr

      reply	other threads:[~2024-09-19  7:40 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-09-17 13:50 [PATCH v2] selftests: livepatch: test livepatching a kprobed function Michael Vetter
2024-09-19  7:40 ` Petr Mladek [this message]

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=ZuvVZneTT7SY76wP@pathway.suse.cz \
    --to=pmladek@suse.com \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=live-patching@vger.kernel.org \
    --cc=mvetter@suse.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;
as well as URLs for NNTP newsgroup(s).