Re: [PATCH v5 2/5] tsm-mr: Add tsm-mr sample code

[Dan Williams <dan.j.williams@intel.com>]

Cedric Xing wrote:
> This sample kernel module demonstrates how to make MRs accessible to user
> mode through the tsm-mr library.
> 
> Once loaded, this module registers a `miscdevice` that host a set of
> emulated measurement registers as shown in the directory tree below.
> 
> /sys/class/misc/tsm_mr_sample
> └── mr
>     ├── config_mr
>     ├── report_digest:sha512
>     ├── rtmr0:sha256
>     ├── rtmr1:sha384
>     ├── rtmr_crypto_agile:sha256
>     ├── rtmr_crypto_agile:sha384
>     └── static_mr:sha384
> 
> Among the MRs in this example:
> 
> - `config_mr` demonstrates a hashless MR, like MRCONFIGID in Intel TDX or
>   HOSTDATA in AMD SEV.
> - `static_mr` demonstrates a static MR. The suffix `:sha384` indicates its
>   value is a sha384 digest.
> - `rtmr0` is an RTMR with `TSM_MR_F_WRITABLE` **cleared**, preventing
>   direct extensions; as a result, the attribute `rtmr0:sha256` is
>   read-only.
> - `rtmr1` is an RTMR with `TSM_MR_F_WRITABLE` **set**, permitting direct
>   extensions; thus, the attribute `rtmr1:sha384` is writable.
> - `rtmr_crypto_agile` demonstrates a "single" MR that supports multiple
>   hash algorithms. Each supported algorithm has a corresponding digest,
>   usually referred to as a "bank" in TCG terminology. In this specific
>   sample, the 2 banks are aliased to `rtmr0` and `rtmr1`, respectively.
> - `report_digest` contains the digest of the internal report structure
>   living in this sample module's memory. It is to demonstrate the use of
>   the `TSM_MR_F_LIVE` flag. Its value changes each time an RTMR is
>   extended.
> 
> Signed-off-by: Cedric Xing <cedric.xing@intel.com>
> ---
>  MAINTAINERS                    |   1 +
>  samples/Kconfig                |  11 ++++
>  samples/Makefile               |   1 +
>  samples/tsm-mr/Makefile        |   2 +
>  samples/tsm-mr/tsm_mr_sample.c | 137 +++++++++++++++++++++++++++++++++++++++++
>  5 files changed, 152 insertions(+)
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 1d1426af3a68..fd5bbf78ec8b 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -24618,6 +24618,7 @@ S:	Maintained
>  F:	Documentation/ABI/testing/configfs-tsm
>  F:	drivers/virt/coco/tsm*.c
>  F:	include/linux/tsm*.h
> +F:	samples/tsm/*
>  
>  TRUSTED SERVICES TEE DRIVER
>  M:	Balint Dobszay <balint.dobszay@arm.com>
> diff --git a/samples/Kconfig b/samples/Kconfig
> index 09011be2391a..6ade17cb16b4 100644
> --- a/samples/Kconfig
> +++ b/samples/Kconfig
> @@ -184,6 +184,17 @@ config SAMPLE_TIMER
>  	bool "Timer sample"
>  	depends on CC_CAN_LINK && HEADERS_INSTALL
>  
> +config SAMPLE_TSM_MR
> +	tristate "TSM measurement sample"
> +	select TSM_MEASUREMENTS
> +	select VIRT_DRIVERS
> +	help
> +	  Build a sample module that emulates MRs (Measurement Registers) and
> +	  exposes them to user mode applications through the TSM sysfs
> +	  interface (/sys/class/misc/tsm_mr_sample/emulated_mr/).
> +
> +	  The module name will be tsm-mr-sample when built as a module.
> +
>  config SAMPLE_UHID
>  	bool "UHID sample"
>  	depends on CC_CAN_LINK && HEADERS_INSTALL
> diff --git a/samples/Makefile b/samples/Makefile
> index bf6e6fca5410..c95bac31851c 100644
> --- a/samples/Makefile
> +++ b/samples/Makefile
> @@ -43,3 +43,4 @@ obj-$(CONFIG_SAMPLES_RUST)		+= rust/
>  obj-$(CONFIG_SAMPLE_DAMON_WSSE)		+= damon/
>  obj-$(CONFIG_SAMPLE_DAMON_PRCL)		+= damon/
>  obj-$(CONFIG_SAMPLE_HUNG_TASK)		+= hung_task/
> +obj-$(CONFIG_SAMPLE_TSM_MR)		+= tsm-mr/
> diff --git a/samples/tsm-mr/Makefile b/samples/tsm-mr/Makefile
> new file mode 100644
> index 000000000000..587c3947b3a7
> --- /dev/null
> +++ b/samples/tsm-mr/Makefile
> @@ -0,0 +1,2 @@
> +# SPDX-License-Identifier: GPL-2.0-only
> +obj-$(CONFIG_SAMPLE_TSM_MR) += tsm_mr_sample.o
> diff --git a/samples/tsm-mr/tsm_mr_sample.c b/samples/tsm-mr/tsm_mr_sample.c
> new file mode 100644
> index 000000000000..3ee35d87d275
> --- /dev/null
> +++ b/samples/tsm-mr/tsm_mr_sample.c
> @@ -0,0 +1,137 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/* Copyright(c) 2024 Intel Corporation. All rights reserved. */

2024-2025

> +
> +#define DEBUG

No need for this. Either the person trying out this sample driver will
define DEBUG at compile time, manually add this line, or otherwise use
dynamic-debug to enable the pr_debug() statements when loading the
module.

If you put #define DEBUG directly in the file then pr_debug() just
becomes printk() which defeats the purpose.

> +#define pr_fmt(x) KBUILD_MODNAME ": " x
> +
> +#include <linux/module.h>
> +#include <linux/tsm-mr.h>
> +#include <linux/miscdevice.h>
> +#include <crypto/hash.h>
> +
> +struct {
> +	u8 static_mr[SHA384_DIGEST_SIZE];
> +	u8 config_mr[SHA512_DIGEST_SIZE];
> +	u8 rtmr0[SHA256_DIGEST_SIZE];
> +	u8 rtmr1[SHA384_DIGEST_SIZE];
> +	u8 report_digest[SHA512_DIGEST_SIZE];
> +} sample_report = {
> +	.static_mr = "static_mr",
> +	.config_mr = "config_mr",
> +	.rtmr0 = "rtmr0",
> +	.rtmr1 = "rtmr1",
> +};
> +
> +static int sample_report_refresh(const struct tsm_measurements *tm,
> +				 const struct tsm_measurement_register *mr)
> +{
> +	struct crypto_shash *tfm;
> +	int rc;
> +
> +	pr_debug("%s(%s) is called\n", __func__, mr ? mr->mr_name : "<nil>");

With dynamic-debug the function name can be enabled, so no need to
include that in the debug statement.

See "Decorator flags add to the message-prefix" in
Documentation/admin-guide/dynamic-debug-howto.rst.

Other than that this looks good to me:

Reviewed-by: Dan Williams <dan.j.williams@intel.com>