Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
References: <20200819172134.11243-1-nramas@linux.microsoft.com>
 <20200819172134.11243-3-nramas@linux.microsoft.com>
User-agent: mu4e 1.4.10; emacs 27.1
From:   Thiago Jung Bauermann <bauerman@linux.ibm.com>
To:     Lakshmi Ramasubramanian <nramas@linux.microsoft.com>
Cc:     zohar@linux.ibm.com, robh@kernel.org, gregkh@linuxfoundation.org,
        james.morse@arm.com, catalin.marinas@arm.com, sashal@kernel.org,
        will@kernel.org, mpe@ellerman.id.au, benh@kernel.crashing.org,
        paulus@samba.org, robh+dt@kernel.org, frowand.list@gmail.com,
        vincenzo.frascino@arm.com, mark.rutland@arm.com,
        dmitry.kasatkin@gmail.com, jmorris@namei.org, serge@hallyn.com,
        pasha.tatashin@soleen.com, allison@lohutok.net,
        kstewart@linuxfoundation.org, takahiro.akashi@linaro.org,
        tglx@linutronix.de, masahiroy@kernel.org, bhsharma@redhat.com,
        mbrugger@suse.com, hsinyi@chromium.org, tao.li@vivo.com,
        christophe.leroy@c-s.fr, linux-integrity@vger.kernel.org,
        linux-kernel@vger.kernel.org, devicetree@vger.kernel.org,
        prsriva@linux.microsoft.com, balajib@linux.microsoft.com
Subject: Re: [PATCH v4 2/5] powerpc: Use libfdt functions to fetch IMA
 buffer properties
In-reply-to: <20200819172134.11243-3-nramas@linux.microsoft.com>
Date:   Thu, 27 Aug 2020 20:50:54 -0300
Message-ID: <8736478x7l.fsf@morokweng.localdomain>
MIME-Version: 1.0
Content-Type: text/plain
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


Lakshmi Ramasubramanian <nramas@linux.microsoft.com> writes:

> @@ -63,7 +29,22 @@ void remove_ima_buffer(void *fdt, int chosen_node)
>  	if (!prop)
>  		return;
>  
> -	ret = do_get_kexec_buffer(prop, len, &addr, &size);
> +	ret = fdt_address_cells(fdt, chosen_node);

This change was already present in the previous version of the patch but
it was just today that I noticed a problem: there's no #address-cells
property in /chosen. This code will still work though because if there's
no property this function returns 2 which is the correct value for
ppc64. But it's conceptually wrong. You need to pass the node offset for
/ so that it gets the #address-cells property from there.

> +	if (ret < 0)
> +		return;
> +	addr_cells = ret;
> +
> +	ret = fdt_size_cells(fdt, chosen_node);

Here we're not so lucky. The default value returned when no #size-cells
property is present is 1, which is wrong for ppc64 so this change
introduces a bug. You also need to pass the node offset for / here.

> +	if (ret < 0)
> +		return;
> +	size_cells = ret;
> +
> +	if (len < 4 * (addr_cells + size_cells))
> +		return;
> +
> +	addr = of_read_number(prop, addr_cells);
> +	size = of_read_number(prop + 4 * addr_cells, size_cells);
> +
>  	fdt_delprop(fdt, chosen_node, FDT_PROP_IMA_KEXEC_BUFFER);
>  	if (ret)
>  		return;
> @@ -129,9 +110,15 @@ int setup_ima_buffer(const struct kimage *image, void *fdt, int chosen_node)
>  	if (!image->arch.ima_buffer_size)
>  		return 0;
>  
> -	ret = get_addr_size_cells(&addr_cells, &size_cells);
> -	if (ret)
> +	ret = fdt_address_cells(fdt, chosen_node);
> +	if (ret < 0)
> +		return ret;
> +	addr_cells = ret;
> +
> +	ret = fdt_size_cells(fdt, chosen_node);
> +	if (ret < 0)
>  		return ret;
> +	size_cells = ret;
>  
>  	entry_size = 4 * (addr_cells + size_cells);

Ditto here.

-- 
Thiago Jung Bauermann
IBM Linux Technology Center