linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dan Williams <dan.j.williams@intel.com>
To: linux-nvdimm@ml01.01.org
Cc: linux-kernel@vger.kernel.org
Subject: [PATCH 3/8] dax: register seed device
Date: Sat, 10 Dec 2016 22:28:41 -0800	[thread overview]
Message-ID: <148143772092.10950.13962446376991710280.stgit@dwillia2-desk3.amr.corp.intel.com> (raw)
In-Reply-To: <148143770485.10950.13227732273892953675.stgit@dwillia2-desk3.amr.corp.intel.com>

Towards adding support for sub-dividing device-dax regions, add a seed
device.  Similar to libnvdimm a 'seed' device is un-configured device
that is enabled after setting configuration parameters.  After a given
seed device is enabled another is created and this process repeats until
no more seed devices can be activated due to dax_available_size being
exhausted.

For now, this simply registers a seed instance after the first dax
device instance is established.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
 drivers/dax/dax.c |   41 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 41 insertions(+)

diff --git a/drivers/dax/dax.c b/drivers/dax/dax.c
index da2962b6f8de..c9ba011e333b 100644
--- a/drivers/dax/dax.c
+++ b/drivers/dax/dax.c
@@ -40,8 +40,10 @@ MODULE_PARM_DESC(nr_dax, "max number of device-dax instances");
  * @kref: to pin while other agents have a need to do lookups
  * @lock: synchronize changes / consistent-access to the resource tree (@res)
  * @dev: parent device backing this region
+ * @seed: next device for dynamic allocation / configuration
  * @align: allocation and mapping alignment for child dax devices
  * @res: physical address range of the region
+ * @child_count: number of registered dax device instances
  * @pfn_flags: identify whether the pfns are paged back or not
  */
 struct dax_region {
@@ -51,8 +53,10 @@ struct dax_region {
 	struct kref kref;
 	struct mutex lock;
 	struct device *dev;
+	struct device *seed;
 	unsigned int align;
 	struct resource res;
+	atomic_t child_count;
 	unsigned long pfn_flags;
 };
 
@@ -161,10 +165,31 @@ static ssize_t align_show(struct device *dev,
 }
 static DEVICE_ATTR_RO(align);
 
+static ssize_t seed_show(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct dax_region *dax_region;
+	ssize_t rc = -ENXIO;
+
+	device_lock(dev);
+	dax_region = dev_get_drvdata(dev);
+	if (dax_region) {
+		mutex_lock(&dax_region->lock);
+		if (dax_region->seed)
+			rc = sprintf(buf, "%s\n", dev_name(dax_region->seed));
+		mutex_unlock(&dax_region->lock);
+	}
+	device_unlock(dev);
+
+	return rc;
+}
+static DEVICE_ATTR_RO(seed);
+
 static struct attribute *dax_region_attributes[] = {
 	&dev_attr_available_size.attr,
 	&dev_attr_region_size.attr,
 	&dev_attr_align.attr,
+	&dev_attr_seed.attr,
 	&dev_attr_id.attr,
 	NULL,
 };
@@ -295,6 +320,9 @@ static void dax_region_free(struct kref *kref)
 	struct dax_region *dax_region;
 
 	dax_region = container_of(kref, struct dax_region, kref);
+	WARN(atomic_read(&dax_region->child_count),
+			"%s: child count not zero\n",
+			dev_name(dax_region->dev));
 	kfree(dax_region);
 }
 
@@ -691,7 +719,10 @@ static void unregister_dax_dev(void *dev)
 	for (i = 0; i < dax_dev->num_resources; i++)
 		__release_region(&dax_region->res, dax_dev->res[i]->start,
 				resource_size(dax_dev->res[i]));
+	if (dax_region->seed == dev)
+		dax_region->seed = NULL;
 	mutex_unlock(&dax_region->lock);
+	atomic_dec(&dax_region->child_count);
 
 	cdev_del(cdev);
 	device_unregister(dev);
@@ -796,6 +827,16 @@ struct dax_dev *devm_create_dax_dev(struct dax_region *dax_region,
 	if (rc)
 		return ERR_PTR(rc);
 
+	if (atomic_inc_return(&dax_region->child_count) == 1) {
+		struct dax_dev *seed;
+
+		seed = devm_create_dax_dev(dax_region, NULL, 0);
+		if (IS_ERR(seed))
+			dev_warn(parent, "failed to create region seed\n");
+		else
+			dax_region->seed = &seed->dev;
+	}
+
 	return dax_dev;
 
  err_cdev:

  parent reply	other threads:[~2016-12-11  6:32 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-12-11  6:28 [PATCH 0/8] device-dax: sub-division support Dan Williams
2016-12-11  6:28 ` [PATCH 1/8] dax: add region-available-size attribute Dan Williams
2016-12-14 14:38   ` Johannes Thumshirn
2016-12-14 15:53     ` Dan Williams
2016-12-15  6:47       ` Dan Williams
2016-12-11  6:28 ` [PATCH 2/8] dax: add region 'id', 'size', and 'align' attributes Dan Williams
2016-12-11  6:28 ` Dan Williams [this message]
2016-12-11  6:28 ` [PATCH 4/8] dax: use multi-order radix for resource lookup Dan Williams
2016-12-11  6:28 ` [PATCH 5/8] dax: refactor locking out of size calculation routines Dan Williams
2016-12-14 15:01   ` Johannes Thumshirn
2016-12-14 15:55     ` Dan Williams
2016-12-11  6:28 ` [PATCH 6/8] dax: sub-division support Dan Williams
2016-12-11  6:29 ` [PATCH 7/8] dax: add / remove dax devices after provisioning Dan Williams
2016-12-11  6:29 ` [PATCH 8/8] dax: add debug for region available_size Dan Williams
2016-12-12 17:15 ` [PATCH 0/8] device-dax: sub-division support Jeff Moyer
2016-12-12 18:46   ` Dan Williams
2016-12-13 23:46     ` Jeff Moyer
2016-12-14  1:17       ` Dan Williams
2016-12-15 16:50         ` Jeff Moyer
2016-12-15 23:48           ` Dan Williams
2016-12-16  2:33             ` Dan Williams

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=148143772092.10950.13962446376991710280.stgit@dwillia2-desk3.amr.corp.intel.com \
    --to=dan.j.williams@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-nvdimm@ml01.01.org \
    /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;
as well as URLs for NNTP newsgroup(s).