Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 23 Oct 2019 16:15:28 -0500
From:   Josh Poimboeuf <jpoimboe@redhat.com>
To:     Petr Mladek <pmladek@suse.com>
Cc:     Jiri Kosina <jikos@kernel.org>, Miroslav Benes <mbenes@suse.cz>,
        Joe Lawrence <joe.lawrence@redhat.com>,
        Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>,
        Nicolai Stange <nstange@suse.de>,
        live-patching@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3 3/5] livepatch: Allow to distinguish different version
 of system state changes
Message-ID: <20191023211528.nfstzbuzzxsyffqh@treble>
References: <20191003090137.6874-1-pmladek@suse.com>
 <20191003090137.6874-4-pmladek@suse.com>
MIME-Version: 1.0
In-Reply-To: <20191003090137.6874-4-pmladek@suse.com>
User-Agent: NeoMutt/20180716
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Petr,

Sorry for taking so long...

On Thu, Oct 03, 2019 at 11:01:35AM +0200, Petr Mladek wrote:
> diff --git a/include/linux/livepatch.h b/include/linux/livepatch.h
> index 726947338fd5..42907c4a0ce8 100644
> --- a/include/linux/livepatch.h
> +++ b/include/linux/livepatch.h
> @@ -133,10 +133,12 @@ struct klp_object {
>  /**
>   * struct klp_state - state of the system modified by the livepatch
>   * @id:=09=09system state identifier (non-zero)
> + * @version:=09version of the change (non-zero)

Is it necessary to assume that 'version' is non-zero?  It would be easy
for a user to not realize that and start with version 0.  Then the patch
state would be silently ignored.

I have the same concern about 'id', but I guess at least one of them has
to be non-zero to differentiate valid entries from the array terminator.

> +/* Check if the patch is able to deal with the given system state. */
> +static bool klp_is_state_compatible(struct klp_patch *patch,
> +=09=09=09=09    struct klp_state *state)
> +{
> +=09struct klp_state *new_state;
> +
> +=09new_state =3D klp_get_state(patch, state->id);
> +
> +=09if (new_state)
> +=09=09return new_state->version >=3D state->version;
> +
> +=09/* Cumulative livepatch must handle all already modified states. */
> +=09return !patch->replace;
> +}

From my perspective I view '!new_state' as an error condition.  I'd find
it easier to read if the ordering were changed to check for the error
first:

=09if (!new_state) {
=09=09/*
=09=09 * A cumulative livepatch must handle all already
=09=09 * modified states.
=09=09 */
=09=09return !patch->replace;
=09}

=09return new_state->version >=3D state->version;


> +
> +/*
> + * Check that the new livepatch will not break the existing system state=
s.
> + * Cumulative patches must handle all already modified states.
> + * Non-cumulative patches can touch already modified states.
> + */
> +bool klp_is_patch_compatible(struct klp_patch *patch)
> +{
> +=09struct klp_patch *old_patch;
> +=09struct klp_state *state;
> +
> +
> +=09klp_for_each_patch(old_patch) {

Extra newline above.

> +=09=09klp_for_each_state(old_patch, state) {
> +=09=09=09if (!klp_is_state_compatible(patch, state))
> +=09=09=09=09return false;
> +=09=09}
> +=09}

I think renaming 'state' to 'old_state' would make the intention a
little clearer, and would be consistent with 'old_patch'.

--=20
Josh