linux-kernel - Re: [Question] refcount of DT node

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 14 Apr 2016 11:10:54 +0100
From:	Mark Rutland <mark.rutland@....com>
To:	Masahiro Yamada <yamada.masahiro@...ionext.com>
Cc:	devicetree@...r.kernel.org,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
	Rob Herring <robh+dt@...nel.org>,
	Arnd Bergmann <arnd@...db.de>,
	Frank Rowand <frowand.list@...il.com>,
	pantelis.antoniou@...sulko.com
Subject: Re: [Question] refcount of DT node

On Thu, Apr 14, 2016 at 04:47:57PM +0900, Masahiro Yamada wrote:
> Hi experts.
> 
> My understanding of refcount of DT node is poor.
> Please help me understand it correctly.
> 
> Sorry if I am asking stupid questions.
> 
> 
> [1] Does this reference count exist for Overlay?
>     Is a node freed when its refcount becomes zero?

I'm not familiar with the way that overlays are intended to work, but
generally this is true, and I believe the same applies.

Pantelis, please correct me if I am wrong on that front.

> [2] When of_node_put() should be called,
>      or should not be called?
> 
> 
> Shouldn't of_node_put() be called
> when we are still referencing to any of its properties?
> 
> For example,  cpu_read_enable_method()
> in arch/arm64/kernel/cpu_ops.c
> returns a pointer to the property value
> instead of creating a copy of it.
> 
> In this case, of_node_put() should not be called
> because we are still referencing the DT property
> (in other words, referencing to the DT node indirectly).
> 
> Am I right?

Yes, the node should not be freed while its data is referred to.

We are leaking a ref there, though, as we no longer refer to that data
after cpu_read_ops().

Fixing that will require some restructuring. We don't expect a CPU node
to need to disappear, so while it's currently not strictly correct the
code shouldn't lead to any adverse behaviour.

> [3] Is the following code correct?
> 
>    np = of_find_compatible_node(NULL, NULL,"foo-node");
>    of_node_put(np);
>    ret = of_address_to_resource(np, 0, &res);
>    if (ret) {
>             pr_err("failed to get resource\n");
>             return ret;
>    }
> 
> Actually I wrote the code above, and it was applied.
> 
> But, the node is still referenced while of_address_to_resource() is being run.
> 
> So the correct code should be as follows?
> 
>    np = of_find_compatible_node(NULL, NULL,"foo-node");
>    ret = of_address_to_resource(np, 0, &res);
>    of_node_put(np);
>    if (ret) {
>             pr_err("failed to get resource\n");
>             return ret;
>    }

It is correctly balanced, yes.

If you don't need to keep the node for future use, this is fine.

Mark.