Date:   Tue, 15 Dec 2020 12:16:01 -0600
From:   Rob Herring <robh@...nel.org>
To:     Lakshmi Ramasubramanian <nramas@...ux.microsoft.com>
Cc:     zohar@...ux.ibm.com, bauerman@...ux.ibm.com,
        takahiro.akashi@...aro.org, gregkh@...uxfoundation.org,
        will@...nel.org, catalin.marinas@....com, mpe@...erman.id.au,
        james.morse@....com, sashal@...nel.org, benh@...nel.crashing.org,
        paulus@...ba.org, frowand.list@...il.com,
        vincenzo.frascino@....com, mark.rutland@....com,
        dmitry.kasatkin@...il.com, jmorris@...ei.org, serge@...lyn.com,
        pasha.tatashin@...een.com, allison@...utok.net,
        masahiroy@...nel.org, bhsharma@...hat.com, mbrugger@...e.com,
        hsinyi@...omium.org, tao.li@...o.com, christophe.leroy@....fr,
        prsriva@...ux.microsoft.com, balajib@...ux.microsoft.com,
        linux-integrity@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org, devicetree@...r.kernel.org
Subject: Re: [PATCH v11 2/8] powerpc: Move ima buffer functions to
 drivers/of/kexec.c

On Mon, Dec 14, 2020 at 11:18:48AM -0800, Lakshmi Ramasubramanian wrote:
> The functions do_get_kexec_buffer() and get_addr_size_cells(),
> defined in arch/powerpc/kexec/ima.c, retrieve the address and size
> of the given property from the device tree blob. These functions do
> not have architecture specific code, but are currently limited to
> powerpc. do_get_kexec_buffer() correctly handles a device tree property
> that is a child node of the root node, but not anything other than
> the immediate root child nodes.
> 
> Move architecture independent functions get_ima_kexec_buffer() and
> get_root_addr_size_cells() to "drivers/of/kexec.c". These functions
> retrieve the chosen node "linux,ima-kexec-buffer" from the device tree,
> and return the address and size of the buffer used for carrying forward
> the IMA measurement log across kexec system call.
> 
> Co-developed-by: Prakhar Srivastava <prsriva@...ux.microsoft.com>
> Signed-off-by: Prakhar Srivastava <prsriva@...ux.microsoft.com>
> Signed-off-by: Lakshmi Ramasubramanian <nramas@...ux.microsoft.com>
> ---
>  drivers/of/kexec.c | 68 ++++++++++++++++++++++++++++++++++++++++++++++
>  include/linux/of.h |  3 ++
>  2 files changed, 71 insertions(+)
> 
> diff --git a/drivers/of/kexec.c b/drivers/of/kexec.c
> index 66787be081fe..9af5371340b1 100644
> --- a/drivers/of/kexec.c
> +++ b/drivers/of/kexec.c
> @@ -30,6 +30,10 @@
>  /**
>   * fdt_find_and_del_mem_rsv - delete memory reservation with given address and size
>   *
> + * @fdt: Flattened Device Tree to update
> + * @start: Starting address of the reservation to delete
> + * @size: Size of the reservation to delete
> + *

This belongs in the patch adding fdt_find_and_del_mem_rsv.

>   * Return: 0 on success, or negative errno on error.
>   */
>  static int fdt_find_and_del_mem_rsv(void *fdt, unsigned long start, unsigned long size)
> @@ -226,3 +230,67 @@ int of_kexec_setup_new_fdt(const struct kimage *image, void *fdt,
>  
>  	return 0;
>  }
> +
> +/**
> + * get_root_addr_size_cells - Get address and size of root node
> + *
> + * @addr_cells: Return address of the root node
> + * @size_cells: Return size of the root node
> + *
> + * Return: 0 on success, or negative errno on error.
> + */
> +int get_root_addr_size_cells(int *addr_cells, int *size_cells)
> +{
> +	struct device_node *root;
> +
> +	root = of_find_node_by_path("/");
> +	if (!root)
> +		return -EINVAL;
> +
> +	*addr_cells = of_n_addr_cells(root);
> +	*size_cells = of_n_size_cells(root);
> +
> +	of_node_put(root);
> +
> +	return 0;
> +}
> +
> +/**
> + * get_ima_kexec_buffer - Get address and size of IMA kexec buffer
> + *
> + * @fdt: Flattened Device Tree
> + * @chosen_node: Offset of chosen node in the FDT
> + * @addr: Return address of the node
> + * @size: Return size of the node
> + *
> + * Return: 0 on success, or negative errno on error.
> + */
> +int get_ima_kexec_buffer(void *fdt, int chosen_node,
> +			 unsigned long *addr, size_t *size)
> +{
> +	const void *prop;
> +	int addr_cells, size_cells, prop_len;
> +	int rc;
> +
> +	rc = get_root_addr_size_cells(&addr_cells, &size_cells);
> +	if (rc)
> +		return rc;
> +
> +	if (fdt)
> +		prop = fdt_getprop(fdt, chosen_node,
> +				   "linux,ima-kexec-buffer", &prop_len);
> +	else
> +		prop = of_get_property(of_chosen,
> +				       "linux,ima-kexec-buffer", &prop_len);

This is an odd structure. The DT APIs are generally of_foo() for 
unflattened tree and of_flat_foo() for FDT.

> +
> +	if (!prop)
> +		return -ENOENT;
> +
> +	if (prop_len < 4 * (addr_cells + size_cells))
> +		return -EINVAL;
> +
> +	*addr = of_read_number(prop, addr_cells);
> +	*size = of_read_number(prop + 4 * addr_cells, size_cells);
> +
> +	return 0;
> +}
> diff --git a/include/linux/of.h b/include/linux/of.h
> index 3375f5295875..fb2ef274135d 100644
> --- a/include/linux/of.h
> +++ b/include/linux/of.h
> @@ -562,6 +562,9 @@ struct kimage;
>  int of_kexec_setup_new_fdt(const struct kimage *image, void *fdt,
>  			   unsigned long initrd_load_addr, unsigned long initrd_len,
>  			   const char *cmdline);
> +int get_root_addr_size_cells(int *addr_cells, int *size_cells);
> +int get_ima_kexec_buffer(void *fdt, int chosen_node,
> +			 unsigned long *addr, size_t *size);

The whole point of moving code out of arch code was to avoid all these 
cross tree calls.

I was expecting the next step to be moving setup_ima_buffer() call into 
of_kexec_setup_new_fdt with all the code in powerpc/kexec/ima.c moved 
here. Then the end result to enable on arm64 is just selecting 
HAVE_IMA_KEXEC and adding fields to kimage_arch.

Rob

Powered by blists - more mailing lists