Date:   Sat, 4 Feb 2023 09:37:30 -0800
From:   Josh Poimboeuf <jpoimboe@kernel.org>
To:     Petr Mladek <pmladek@suse.com>
Cc:     Marcos Paulo de Souza <mpdesouza@suse.com>,
        linux-kernel@vger.kernel.org, live-patching@vger.kernel.org,
        jpoimboe@redhat.com, joe.lawrence@redhat.com
Subject: Re: [PATCH v2 4/4] livepatch/shadow: Add garbage collection of
 shadow variables
Message-ID: <20230204173730.d2w6rwjlok33ah7h@treble>
References: <20221026194122.11761-1-mpdesouza@suse.com>
 <20221026194122.11761-5-mpdesouza@suse.com>
 <20230201001817.ej7n3ehtgaxwr3pq@treble>
 <Y9uM94UOJWcNxnkI@alley>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <Y9uM94UOJWcNxnkI@alley>
Precedence: bulk

On Thu, Feb 02, 2023 at 11:14:15AM +0100, Petr Mladek wrote:
> On Tue 2023-01-31 16:18:17, Josh Poimboeuf wrote:
> > On Wed, Oct 26, 2022 at 04:41:22PM -0300, Marcos Paulo de Souza wrote:
> > > The shadow variable type might then be added into a new @shadow_types
> > > member of struct klp_object. They will get then automatically registered
> > > and unregistered when the object is being livepatched. The registration
> > > increments the reference count. Unregistration decreases the reference
> > > count. All shadow variables of the given type are freed when the reference
> > > count reaches zero.
> > 
> > How does the automatic unregistration work for replaced patches?
> > 
> > I see klp_unpatch_replaced_patches() is called, but I don't see where it
> > ends up calling klp_shadow_unregister() for the replaced patch(es).
> 
> Great catch! I forgot that replaced patches are handled separately.
> We should do the following (on top of this patch):
> 
> --- a/kernel/livepatch/transition.c
> +++ b/kernel/livepatch/transition.c
> @@ -123,8 +123,11 @@ static void klp_complete_transition(void)
>  			continue;
>  		if (klp_target_state == KLP_PATCHED)
>  			klp_post_patch_callback(obj);
> -		else if (klp_target_state == KLP_UNPATCHED) {
> +		else if (klp_target_state == KLP_UNPATCHED)
>  			klp_post_unpatch_callback(obj);
> +
> +		if (klp_target_state == KLP_UNPATCHED ||
> +		    klp_transition_patch->replace) {
>  			klp_unregister_shadow_types(obj);
>  		}
>  	}

That calls klp_unregister_shadow_types() on the objects for
klp_transition_patch, which is the replacement patch, not the replaced
patch(es).

-- 
Josh