Linux CXL
 help / color / mirror / Atom feed
From: Alison Schofield <alison.schofield@intel.com>
To: Davidlohr Bueso <dave@stgolabs.net>
Cc: vishal.l.verma@intel.com, y-goto@fujitsu.com,
	dave.jiang@intel.com, dan.j.williams@intel.com,
	linux-cxl@vger.kernel.org, nvdimm@lists.linux.dev
Subject: Re: [PATCH -ndctl] cxl/memdev: Introduce sanitize-memdev functionality
Date: Mon, 30 Sep 2024 10:36:58 -0700	[thread overview]
Message-ID: <ZvrhusA7So_u51W_@aschofie-mobl2.lan> (raw)
In-Reply-To: <20240928211643.140264-1-dave@stgolabs.net>

+ nvdimm@lists.linux.dev

On Sat, Sep 28, 2024 at 02:16:42PM -0700, Davidlohr Bueso wrote:
> Add a new cxl_memdev_sanitize() to libcxl to support triggering memory
> device sanitation, in either Sanitize and/or Secure Erase, per the
> CXL 3.0 spec.
> 
> This is analogous to 'ndctl sanitize-dimm'.
> 
> Signed-off-by: Davidlohr Bueso <dave@stgolabs.net>

Hi David,

I'm wrangling patches for ndctl now and need your help with this one.

Looking at the lore history, you posted a patchset for wait-sanitize &
sanitize-memdev in Apr'03.[1] and update with rev2.[2] Later in Oct'23
Dan posted a patchset with wait-sanitize and a unit test that was merged
in ndctl v80.[3,4]  A quick look at the code tells me Dan did not just
grab your implementation. It differs.

Can you confirm that the two features as a set are what you want today?
Are the last comments from Vishal and I addressed? [2]
Can the existing unit test be expanded with a sanitize-memdev test case?

Thanks!

-- Alison


[1] https://lore.kernel.org/linux-cxl/20230423015920.11384-1-dave@stgolabs.net/
[2] https://lore.kernel.org/linux-cxl/20230713195455.19769-1-dave@stgolabs.net/
[3] https://lore.kernel.org/all/169657749402.1491881.12666757616880845510.stgit@dwillia2-xfh.jf.intel.com/
[4] https://lore.kernel.org/nvdimm/72bdf880b2cafd42163638d9e7e1d848c1d2d3a9.camel@intel.com/






> ---
>  Documentation/cxl/cxl-sanitize-memdev.txt | 59 +++++++++++++++++++++++
>  Documentation/cxl/meson.build             |  1 +
>  cxl/builtin.h                             |  1 +
>  cxl/cxl.c                                 |  1 +
>  cxl/lib/libcxl.c                          | 15 ++++++
>  cxl/lib/libcxl.sym                        |  1 +
>  cxl/libcxl.h                              |  1 +
>  cxl/memdev.c                              | 48 ++++++++++++++++++
>  8 files changed, 127 insertions(+)
>  create mode 100644 Documentation/cxl/cxl-sanitize-memdev.txt
> 
> diff --git a/Documentation/cxl/cxl-sanitize-memdev.txt b/Documentation/cxl/cxl-sanitize-memdev.txt
> new file mode 100644
> index 000000000000..22212898542a
> --- /dev/null
> +++ b/Documentation/cxl/cxl-sanitize-memdev.txt
> @@ -0,0 +1,59 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +cxl-sanitize-memdev(1)
> +======================
> +
> +NAME
> +----
> +cxl-sanitize-memdev - Perform a cryptographic destruction or sanitization
> +of the contents of the given memdev(s).
> +
> +SYNOPSIS
> +--------
> +[verse]
> +'cxl sanitize-memdev <mem0> [<mem1>..<memN>] [<options>]'
> +
> +DESCRIPTION
> +-----------
> +The 'sanitize-memdev' command performs two different methods of
> +sanitization, per the CXL 3.0+ specification. It is required that
> +the memdev be disabled before sanitizing, such that the device
> +cannot be actively decoding any HPA ranges at the time.
> +
> +The default is 'sanitize', but additionally, a 'secure-erase'
> +option is available. If both types of operations are supplied,
> +then the 'secure-erase' is performed before 'sanitize'.
> +
> +OPTIONS
> +-------
> +
> +include::bus-option.txt[]
> +
> +-e::
> +--secure-erase::
> +	Erase user data by changing the media encryption keys for all user
> +	data areas of the device.
> +
> +-s::
> +--sanitize::
> +	Sanitize the device to securely re-purpose or decommission it. This is
> +	done by ensuring that all user data and meta data, whether it resides
> +	in persistent capacity, volatile capacity, or the label storage area,
> +	is made permanently unavailable by whatever means is appropriate for
> +	the media type.
> +
> +	With this option, the sanitization request is merely submitted to the
> +	kernel, and the completion is asynchronous. Depending on the medium and
> +	capacity, sanitize may take tens of minutes to many hours. Subsequently,
> +	'cxl wait-sanitize’ can be used to wait for the memdevs that are under
> +	the sanitization.
> +
> +include::verbose-option.txt[]
> +
> +include::../copyright.txt[]
> +
> +SEE ALSO
> +--------
> +linkcxl:cxl-wait-sanitize[1],
> +linkcxl:cxl-disable-memdev[1],
> +linkcxl:cxl-list[1],
> diff --git a/Documentation/cxl/meson.build b/Documentation/cxl/meson.build
> index 8085c1c2c87e..99e6ee782a1c 100644
> --- a/Documentation/cxl/meson.build
> +++ b/Documentation/cxl/meson.build
> @@ -49,6 +49,7 @@ cxl_manpages = [
>    'cxl-monitor.txt',
>    'cxl-update-firmware.txt',
>    'cxl-set-alert-config.txt',
> +  'cxl-sanitize-memdev.txt',
>    'cxl-wait-sanitize.txt',
>  ]
>  
> diff --git a/cxl/builtin.h b/cxl/builtin.h
> index c483f301e5e0..29c8ad2a0ad9 100644
> --- a/cxl/builtin.h
> +++ b/cxl/builtin.h
> @@ -16,6 +16,7 @@ int cmd_reserve_dpa(int argc, const char **argv, struct cxl_ctx *ctx);
>  int cmd_free_dpa(int argc, const char **argv, struct cxl_ctx *ctx);
>  int cmd_update_fw(int argc, const char **argv, struct cxl_ctx *ctx);
>  int cmd_set_alert_config(int argc, const char **argv, struct cxl_ctx *ctx);
> +int cmd_sanitize_memdev(int argc, const char **argv, struct cxl_ctx *ctx);
>  int cmd_wait_sanitize(int argc, const char **argv, struct cxl_ctx *ctx);
>  int cmd_disable_port(int argc, const char **argv, struct cxl_ctx *ctx);
>  int cmd_enable_port(int argc, const char **argv, struct cxl_ctx *ctx);
> diff --git a/cxl/cxl.c b/cxl/cxl.c
> index 16436671dc53..9c9f217c5a93 100644
> --- a/cxl/cxl.c
> +++ b/cxl/cxl.c
> @@ -80,6 +80,7 @@ static struct cmd_struct commands[] = {
>  	{ "disable-region", .c_fn = cmd_disable_region },
>  	{ "destroy-region", .c_fn = cmd_destroy_region },
>  	{ "monitor", .c_fn = cmd_monitor },
> +	{ "sanitize-memdev", .c_fn = cmd_sanitize_memdev },
>  };
>  
>  int main(int argc, const char **argv)
> diff --git a/cxl/lib/libcxl.c b/cxl/lib/libcxl.c
> index 91eedd1c4688..4f44bf1b6185 100644
> --- a/cxl/lib/libcxl.c
> +++ b/cxl/lib/libcxl.c
> @@ -1414,6 +1414,21 @@ CXL_EXPORT int cxl_memdev_get_id(struct cxl_memdev *memdev)
>  	return memdev->id;
>  }
>  
> +CXL_EXPORT int cxl_memdev_sanitize(struct cxl_memdev *memdev, char *op)
> +{
> +	struct cxl_ctx *ctx = cxl_memdev_get_ctx(memdev);
> +	char *path = memdev->dev_buf;
> +	int len = memdev->buf_len;
> +
> +	if (snprintf(path, len,
> +		     "%s/security/%s", memdev->dev_path, op) >= len) {
> +		err(ctx, "%s: buffer too small!\n",
> +		    cxl_memdev_get_devname(memdev));
> +		return -ERANGE;
> +	}
> +	return sysfs_write_attr(ctx, path, "1\n");
> +}
> +
>  CXL_EXPORT int cxl_memdev_wait_sanitize(struct cxl_memdev *memdev,
>  					int timeout_ms)
>  {
> diff --git a/cxl/lib/libcxl.sym b/cxl/lib/libcxl.sym
> index 304d7fa735d4..89a4c63cb874 100644
> --- a/cxl/lib/libcxl.sym
> +++ b/cxl/lib/libcxl.sym
> @@ -281,4 +281,5 @@ global:
>  	cxl_memdev_get_ram_qos_class;
>  	cxl_region_qos_class_mismatch;
>  	cxl_port_decoders_committed;
> +	cxl_memdev_sanitize;
>  } LIBCXL_6;
> diff --git a/cxl/libcxl.h b/cxl/libcxl.h
> index fc6dd0085440..a722bab8a65b 100644
> --- a/cxl/libcxl.h
> +++ b/cxl/libcxl.h
> @@ -79,6 +79,7 @@ bool cxl_memdev_fw_update_in_progress(struct cxl_memdev *memdev);
>  size_t cxl_memdev_fw_update_get_remaining(struct cxl_memdev *memdev);
>  int cxl_memdev_update_fw(struct cxl_memdev *memdev, const char *fw_path);
>  int cxl_memdev_cancel_fw_update(struct cxl_memdev *memdev);
> +int cxl_memdev_sanitize(struct cxl_memdev *memdev, char *op);
>  int cxl_memdev_wait_sanitize(struct cxl_memdev *memdev, int timeout_ms);
>  
>  /* ABI spelling mistakes are forever */
> diff --git a/cxl/memdev.c b/cxl/memdev.c
> index 6e44d1578d03..60d1515b19f3 100644
> --- a/cxl/memdev.c
> +++ b/cxl/memdev.c
> @@ -35,6 +35,8 @@ static struct parameters {
>  	bool align;
>  	bool cancel;
>  	bool wait;
> +	bool sanitize;
> +	bool secure_erase;
>  	const char *type;
>  	const char *size;
>  	const char *decoder_filter;
> @@ -160,6 +162,12 @@ OPT_STRING('\0', "pmem-err-alert",                                            \
>  	   &param.corrected_pmem_err_alert, "'on' or 'off'",                  \
>  	   "enable or disable corrected pmem error warning alert")
>  
> +#define SANITIZE_OPTIONS()			      \
> +OPT_BOOLEAN('e', "secure-erase", &param.secure_erase, \
> +	    "secure erase a memdev"),		      \
> +OPT_BOOLEAN('s', "sanitize", &param.sanitize,	      \
> +	    "sanitize a memdev")
> +
>  #define WAIT_SANITIZE_OPTIONS()                \
>  OPT_INTEGER('t', "timeout", &param.timeout,    \
>  	    "time in milliseconds to wait for overwrite completion (default: infinite)")
> @@ -226,6 +234,12 @@ static const struct option set_alert_options[] = {
>  	OPT_END(),
>  };
>  
> +static const struct option sanitize_options[] = {
> +	BASE_OPTIONS(),
> +	SANITIZE_OPTIONS(),
> +	OPT_END(),
> +};
> +
>  static const struct option wait_sanitize_options[] = {
>  	BASE_OPTIONS(),
>  	WAIT_SANITIZE_OPTIONS(),
> @@ -772,6 +786,27 @@ out_err:
>  	return rc;
>  }
>  
> +static int action_sanitize_memdev(struct cxl_memdev *memdev,
> +				  struct action_context *actx)
> +{
> +	int rc = 0;
> +
> +	if (cxl_memdev_is_enabled(memdev))
> +		return -EBUSY;
> +
> +	/* let Sanitize be the default */
> +	if (!param.secure_erase && !param.sanitize)
> +		param.sanitize = true;
> +	if (param.secure_erase)
> +		rc = cxl_memdev_sanitize(memdev, "erase");
> +	if (param.sanitize)
> +		rc = cxl_memdev_sanitize(memdev, "sanitize");
> +	else
> +		rc = -EINVAL;
> +
> +	return rc;
> +}
> +
>  static int action_wait_sanitize(struct cxl_memdev *memdev,
>  				struct action_context *actx)
>  {
> @@ -1228,6 +1263,19 @@ int cmd_set_alert_config(int argc, const char **argv, struct cxl_ctx *ctx)
>  	return count >= 0 ? 0 : EXIT_FAILURE;
>  }
>  
> +int cmd_sanitize_memdev(int argc, const char **argv, struct cxl_ctx *ctx)
> +{
> +	int count = memdev_action(
> +		argc, argv, ctx, action_sanitize_memdev, sanitize_options,
> +		"cxl sanitize-memdev <mem0> [<mem1>..<memn>] [<options>]");
> +
> +	log_info(&ml, "sanitize %s on %d mem device%s\n",
> +		 count >= 0 ? "completed/started" : "failed",
> +		 count >= 0 ? count : 0,  count > 1 ? "s" : "");
> +
> +	return count >= 0 ? 0 : EXIT_FAILURE;
> +}
> +
>  int cmd_wait_sanitize(int argc, const char **argv, struct cxl_ctx *ctx)
>  {
>  	int count = memdev_action(
> -- 
> 2.46.1
> 
> 

  reply	other threads:[~2024-09-30 17:37 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-09-28 21:16 [PATCH -ndctl] cxl/memdev: Introduce sanitize-memdev functionality Davidlohr Bueso
2024-09-30 17:36 ` Alison Schofield [this message]
2024-09-30 19:29   ` Davidlohr Bueso
2024-09-30 21:39     ` Davidlohr Bueso
2025-02-18 16:34       ` Alison Schofield
2025-02-18 16:55         ` Davidlohr Bueso
2024-12-10  7:38 ` Yasunori Gotou (Fujitsu)

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ZvrhusA7So_u51W_@aschofie-mobl2.lan \
    --to=alison.schofield@intel.com \
    --cc=dan.j.williams@intel.com \
    --cc=dave.jiang@intel.com \
    --cc=dave@stgolabs.net \
    --cc=linux-cxl@vger.kernel.org \
    --cc=nvdimm@lists.linux.dev \
    --cc=vishal.l.verma@intel.com \
    --cc=y-goto@fujitsu.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox