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, 21 Nov 2018 14:59:29 +0100 (CET)
From:   Miroslav Benes <mbenes@suse.cz>
To:     Petr Mladek <pmladek@suse.com>
cc:     Jiri Kosina <jikos@kernel.org>,
        Josh Poimboeuf <jpoimboe@redhat.com>,
        Jason Baron <jbaron@akamai.com>,
        Joe Lawrence <joe.lawrence@redhat.com>,
        Evgenii Shatokhin <eshatokhin@virtuozzo.com>,
        live-patching@vger.kernel.org, linux-kernel@vger.kernel.org,
        Jessica Yu <jeyu@kernel.org>
Subject: Re: [PATCH v13 04/12] livepatch: Consolidate klp_free functions
In-Reply-To: <20181015123713.25868-5-pmladek@suse.com>
Message-ID: <alpine.LSU.2.21.1811211450280.7578@pobox.suse.cz>
References: <20181015123713.25868-1-pmladek@suse.com> <20181015123713.25868-5-pmladek@suse.com>
User-Agent: Alpine 2.21 (LSU 202 2017-01-01)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

> -/*
> - * Free all functions' kobjects in the array up to some limit. When limit is
> - * NULL, all kobjects are freed.
> - */
> -static void klp_free_funcs_limited(struct klp_object *obj,
> -				   struct klp_func *limit)
> +static void klp_free_funcs(struct klp_object *obj)
>  {
>  	struct klp_func *func;
>  
> -	for (func = obj->funcs; func->old_name && func != limit; func++)
> -		kobject_put(&func->kobj);
> +	klp_for_each_func(obj, func) {
> +		/* Might be called from klp_init_patch() error path. */
> +		if (func->kobj.state_initialized)
> +			kobject_put(&func->kobj);
> +	}
>  }

I have not noticed till today, but state_initialized is probably not the 
best idea. kobject_init_and_add() sets it to 1 in kobject_init() part but 
then _add() is called which could result in error. So we would end up with 
state_initialized equal to 1 and kobject reference equal to 0. Later call 
to kobject_put() in klp_free_funcs() (or elsewhere) would not call 
->release method, because refcount would be 0 by then.

I think that all would end up well, but that does not mean we should not 
fix it.

We could use state_in_sysfs, but I do not think it guarantees anything. 
Both are internal states and maybe we should not rely on them.

So kref_read() and check the reference?

Miroslav