Linux CXL
 help / color / mirror / Atom feed
From: Jonathan Cameron <Jonathan.Cameron@huawei.com>
To: Dave Jiang <dave.jiang@intel.com>
Cc: <linux-cxl@vger.kernel.org>, <dan.j.williams@intel.com>,
	<ira.weiny@intel.com>, <vishal.l.verma@intel.com>,
	<alison.schofield@intel.com>, <dave@stgolabs.net>,
	<jgg@nvidia.com>, <shiju.jose@huawei.com>
Subject: Re: [PATCH v2 08/16] cxl: Add FWCTL support to the CXL memdev driver
Date: Mon, 3 Feb 2025 12:42:32 +0000	[thread overview]
Message-ID: <20250203124232.00003dd2@huawei.com> (raw)
In-Reply-To: <20250201004459.466499-9-dave.jiang@intel.com>

On Fri, 31 Jan 2025 17:42:01 -0700
Dave Jiang <dave.jiang@intel.com> wrote:

> Add fwctl support code to allow sending of CXL feature commands from
> userspace through as ioctls via FWCTL. Provide initial setup bits.
> 
> Signed-off-by: Dave Jiang <dave.jiang@intel.com>
> ---
> v2:
> - Associate FWCTL with the 'cxl_memdev' through the memdev driver. (Dan)
> - I'm leaving Jonathan's tag off to get this looked at again.
Wise move ;)  Comments inline.  Mind you I'm rereading the lot
as been off a few days and this is all gone from my head!



> diff --git a/drivers/cxl/features.c b/drivers/cxl/features.c
> index 82598a409acc..2dcfefdfa6b3 100644
> --- a/drivers/cxl/features.c
> +++ b/drivers/cxl/features.c

>  
> +/* The __weak symbol gets replaced by FWCTL allocation function in fwctl.c */
> +struct cxl_features_state * __weak devm_cxlfs_allocate(struct cxl_memdev *cxlmd)
> +{
> +	struct device *dev = &cxlmd->dev;
> +	struct cxl_features_state *cxlfs =
> +		devm_kzalloc(dev, sizeof(*cxlfs), GFP_KERNEL);
> +
> +	if (!cxlfs)
> +		return ERR_PTR(-ENOMEM);
This is returning an error pointer, but the DEFINE FREE assumes
NULL I think for the failed to allocate case. Or just return NULL
and don't modify that.

> +
> +	cxlfs->cxlmd = cxlmd;
> +
> +	return cxlfs;
> +}
> +
> +/* The __weak symbol gets replaced by FWCTL free function in fwctl.c */
> +void __weak devm_cxlfs_free(struct cxl_memdev *cxlmd)
> +{
> +	devm_kfree(&cxlmd->dev, cxlmd->cxlfs);
> +	cxlmd->cxlfs = NULL;

This missing clear of hanging pointer I mentioned back in earlier patch.

> +}
> +
>  static void cxl_cxlfs_free(struct cxl_features_state *cxlfs)
>  {
> -	struct device *dev = &cxlfs->cxlmd->dev;
> -
> -	devm_kfree(dev, cxlfs);
> +	devm_cxlfs_free(cxlfs->cxlmd);
>  }
>  
>  DEFINE_FREE(free_cxlfs, struct cxl_features_state *, if (_T) cxl_cxlfs_free(_T))
> @@ -185,15 +208,13 @@ DEFINE_FREE(free_cxlfs, struct cxl_features_state *, if (_T) cxl_cxlfs_free(_T))

As above, needs to be if (!IS_ERR_OR_NULL(_T)) I think.

>   */
>  int devm_cxl_add_features(struct cxl_memdev *cxlmd)
>  {
> -	struct device *dev = &cxlmd->dev;
>  	int rc;
>  
>  	struct cxl_features_state *cxlfs __free(free_cxlfs) =
> -		devm_kzalloc(dev, sizeof(*cxlfs), GFP_KERNEL);
> +		devm_cxlfs_allocate(cxlmd);

As per early comments. I'd drag this much earlier in the series
though make only add the __weak markings and comments in this
patch.  That will reduce churn.

>  	if (!cxlfs)
>  		return -ENOMEM;
>  
> -	cxlfs->cxlmd = cxlmd;
>  	enumerate_feature_cmds(cxlmd, cxlfs);
>  	rc = get_supported_features(cxlmd, cxlfs);
>  	if (rc)
> @@ -204,3 +225,5 @@ int devm_cxl_add_features(struct cxl_memdev *cxlmd)
>  	return 0;
>  }
>  EXPORT_SYMBOL_NS_GPL(devm_cxl_add_features, "CXL");
> +
> +MODULE_IMPORT_NS("CXL");

> diff --git a/drivers/cxl/fwctl.c b/drivers/cxl/fwctl.c
> new file mode 100644
> index 000000000000..d8fd68ac8982
> --- /dev/null
> +++ b/drivers/cxl/fwctl.c
> @@ -0,0 +1,88 @@

> +static void remove_cxlfs(void *_cxlfs)
> +{
> +	struct cxl_features_state *cxlfs = _cxlfs;
> +	struct cxl_memdev *cxlmd = cxlfs->cxlmd;
> +	struct fwctl_device *fwctl = &cxlfs->fwctl;
> +
> +	cxlmd->cxlfs = NULL;

This is a little nasty.  Originally set in enumerate commands, so not
obvious it should be unwound down in here. Perhaps a comment is
enough.

> +	fwctl_unregister(fwctl);
> +	fwctl_put(fwctl);
> +}
> +
> +DEFINE_FREE(free_cxlfs, struct cxl_features_state *, if (_T) fwctl_put(&_T->fwctl))
> +
> +static struct cxl_features_state *
> +__devm_cxl_cfs_allocate(struct cxl_memdev *cxlmd, const struct fwctl_ops *ops)
> +{
> +	struct device *dev = &cxlmd->dev;
> +	int rc;
> +
> +	struct cxl_features_state *cxlfs __free(free_cxlfs) =
> +		fwctl_alloc_device(dev, ops, struct cxl_features_state, fwctl);
> +	if (!cxlfs)
> +		return ERR_PTR(-ENOMEM);

As above, the caller of this is used with __free that is assuming NULL
for failed.

> +
> +	cxlfs->cxlmd = cxlmd;
> +	rc = fwctl_register(&cxlfs->fwctl);
> +	if (rc)
> +		return ERR_PTR(rc);
> +
> +	rc = devm_add_action_or_reset(dev, remove_cxlfs, cxlfs);
> +	if (rc)
> +		return ERR_PTR(rc);
> +
> +	return no_free_ptr(cxlfs);
> +}

This use of weak functions is a bit too clever for me to be entirely happy but it
seems right so fair enough.

> +
> +struct cxl_features_state *devm_cxlfs_allocate(struct cxl_memdev *cxlmd)
> +{
> +	return __devm_cxl_cfs_allocate(cxlmd, &cxlctl_ops);
> +}
> +EXPORT_SYMBOL_NS_GPL(devm_cxlfs_allocate, "CXL");
> +
> +void devm_cxlfs_free(struct cxl_memdev *cxlmd)
> +{
> +	devm_release_action(&cxlmd->dev, remove_cxlfs, cxlmd);
> +}
> +EXPORT_SYMBOL_NS_GPL(devm_cxlfs_free, "CXL");
> +
> +MODULE_IMPORT_NS("FWCTL");


  parent reply	other threads:[~2025-02-03 12:42 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-01  0:41 [PATCH v2 0/16] cxl: Add CXL feature commands support via fwctl Dave Jiang
2025-02-01  0:41 ` [PATCH v2 01/16] cxl: Refactor user ioctl command path from mds to mailbox Dave Jiang
2025-02-01  0:41 ` [PATCH v2 02/16] cxl: Enumerate feature commands Dave Jiang
2025-02-03 12:06   ` Jonathan Cameron
2025-02-03 23:23     ` Dave Jiang
2025-02-01  0:41 ` [PATCH v2 03/16] cxl: Add Get Supported Features command for kernel usage Dave Jiang
2025-02-03 12:19   ` Jonathan Cameron
2025-02-01  0:41 ` [PATCH v2 04/16] cxl/test: Add Get Supported Features mailbox command support Dave Jiang
2025-02-03 12:22   ` Jonathan Cameron
2025-02-01  0:41 ` [PATCH v2 05/16] cxl/mbox: Add GET_FEATURE mailbox command Dave Jiang
2025-02-01  0:41 ` [PATCH v2 06/16] cxl/mbox: Add SET_FEATURE " Dave Jiang
2025-02-03 12:27   ` Jonathan Cameron
2025-02-01  0:42 ` [PATCH v2 07/16] cxl: Setup exclusive CXL features that are reserved for the kernel Dave Jiang
2025-02-01  0:42 ` [PATCH v2 08/16] cxl: Add FWCTL support to the CXL memdev driver Dave Jiang
2025-02-01  1:04   ` Dave Jiang
2025-02-03 12:42   ` Jonathan Cameron [this message]
2025-02-03 14:25     ` Jason Gunthorpe
2025-02-03 16:22       ` Dave Jiang
2025-02-03 16:30         ` Jason Gunthorpe
2025-02-01  0:42 ` [PATCH v2 09/16] cxl: Add support for FWCTL get driver information callback Dave Jiang
2025-02-03 12:43   ` Jonathan Cameron
2025-02-01  0:42 ` [PATCH v2 10/16] cxl: Move cxl feature command structs to user header Dave Jiang
2025-02-01  0:42 ` [PATCH v2 11/16] cxl: Add support for fwctl RPC command to enable CXL feature commands Dave Jiang
2025-02-03 13:03   ` Jonathan Cameron
2025-02-01  0:42 ` [PATCH v2 12/16] cxl: Add support to handle user feature commands for get feature Dave Jiang
2025-02-01  0:42 ` [PATCH v2 13/16] cxl: Add support to handle user feature commands for set feature Dave Jiang
2025-02-01  0:42 ` [PATCH v2 14/16] cxl/test: Add Get Feature support to cxl_test Dave Jiang
2025-02-01  0:42 ` [PATCH v2 15/16] cxl/test: Add Set " Dave Jiang
2025-02-01  0:42 ` [PATCH v2 16/16] fwctl/cxl: Add documentation to FWCTL CXL Dave Jiang
2025-02-03 13:08   ` Jonathan Cameron

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=20250203124232.00003dd2@huawei.com \
    --to=jonathan.cameron@huawei.com \
    --cc=alison.schofield@intel.com \
    --cc=dan.j.williams@intel.com \
    --cc=dave.jiang@intel.com \
    --cc=dave@stgolabs.net \
    --cc=ira.weiny@intel.com \
    --cc=jgg@nvidia.com \
    --cc=linux-cxl@vger.kernel.org \
    --cc=shiju.jose@huawei.com \
    --cc=vishal.l.verma@intel.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