Re: [PATCH 3/4] firmware: Add support for Qualcomm UEFI Secure Application

[Johan Hovold]

On Sun, Jul 24, 2022 at 12:49:48AM +0200, Maximilian Luz wrote:
> On platforms using the Qualcomm UEFI Secure Application (uefisecapp),
> EFI variables cannot be accessed via the standard interface in EFI
> runtime mode. The respective functions return EFI_UNSUPPORTED. On these
> platforms, we instead need to talk to uefisecapp. This commit provides
> support for this and registers the respective efivars operations to
> access EFI variables from the kernel.
> 
> Communication with uefisecapp follows the standard Qualcomm Trusted
> Environment (TEE or TrEE) / Secure OS conventions via the respective SCM
> call interface. This is also the reason why variable access works
> normally while boot services are active. During this time, said SCM
> interface is managed by the boot services. When calling
> ExitBootServices(), the ownership is transferred to the kernel.
> Therefore, UEFI must not use that interface itself (as multiple parties
> accessing this interface at the same time may lead to complications) and
> cannot access variables for us.
> 
> Signed-off-by: Maximilian Luz <luzmaximilian@xxxxxxxxx>
> ---

> +static struct platform_driver qcom_uefisecapp_driver = {
> +	.probe = qcom_uefisecapp_probe,
> +	.remove = qcom_uefisecapp_remove,
> +	.driver = {
> +		.name = "qcom_tee_uefisecapp",
> +		.of_match_table = qcom_uefisecapp_dt_match,
> +		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
> +	},
> +};
> +module_platform_driver(qcom_uefisecapp_driver);

I noticed that for efivarfs to work, you're currently relying on having
the firmware still claim that the variable services are supported in the
RT_PROP table so that efi core registers the default ops at subsys init
time (which are later overridden by this driver).

Otherwise efivarfs may fail to initialise when built in:

	static __init int efivarfs_init(void)
	{
		if (!efivars_kobject())
			return -ENODEV;

		return register_filesystem(&efivarfs_type);
	}

	module_init(efivarfs_init);

With recent X13s firmware the corresponding bit in the RT_PROP table has
been cleared so that efivarfs would fail to initialise. Similar problem
when booting with 'efi=noruntime'.

One way to handle this is to register also the qcom_uefisecapp_driver at
subsys init time and prevent it from being built as a module (e.g. as is
done for the SCM driver). I'm using the below patch for this currently.

I guess the Google GSMI implementation suffers from a similar problem.

Johan