linux-scsi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Douglas Gilbert <dgilbert@interlog.com>
To: linux-scsi@vger.kernel.org
Cc: martin.petersen@oracle.com, jejb@linux.vnet.ibm.com, hare@suse.de
Subject: [PATCH v5 9/9] scsi_host: switch ida to idr to hold shost ptr
Date: Sun, 19 Jul 2020 22:57:42 -0400	[thread overview]
Message-ID: <20200720025742.349296-10-dgilbert@interlog.com> (raw)
In-Reply-To: <20200720025742.349296-1-dgilbert@interlog.com>

Previously the scsi_host code was using a global static "ida"
array to make sure host numbers where unique and issued in a
predictable ascending integer order. Extend that ida to an
idr array so that it can additionally hold the pointer to
the scsi_host object. This enables scsi_host_lookup() to do
a O(ln(n)) search replacing class_find_device() which is O(n).
This makes all the SCSI exported " lookup" functions O(ln(n)).

iSCSI seems to be the largest user of scsi_host_lookup().

Signed-off-by: Douglas Gilbert <dgilbert@interlog.com>
---
 drivers/scsi/hosts.c | 37 +++++++++++--------------------------
 1 file changed, 11 insertions(+), 26 deletions(-)

diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
index aebef37684e8..7734dc63606d 100644
--- a/drivers/scsi/hosts.c
+++ b/drivers/scsi/hosts.c
@@ -50,7 +50,7 @@ module_param_named(eh_deadline, shost_eh_deadline, int, S_IRUGO|S_IWUSR);
 MODULE_PARM_DESC(eh_deadline,
 		 "SCSI EH timeout in seconds (should be between 0 and 2^31-1)");
 
-static DEFINE_IDA(host_index_ida);
+static DEFINE_IDR(host_index_idr);	/* store index and shost pointer */
 
 
 static void scsi_host_cls_release(struct device *dev)
@@ -343,7 +343,7 @@ static void scsi_host_dev_release(struct device *dev)
 
 	kfree(shost->shost_data);
 
-	ida_simple_remove(&host_index_ida, shost->host_no);
+	idr_remove(&host_index_idr, shost->host_no);
 
 	if (parent)
 		put_device(parent);
@@ -390,8 +390,8 @@ struct Scsi_Host *scsi_host_alloc(struct scsi_host_template *sht, int privsize)
 	init_waitqueue_head(&shost->host_wait);
 	mutex_init(&shost->scan_mutex);
 
-	index = ida_simple_get(&host_index_ida, 0, 0, GFP_KERNEL);
-	if (index < 0)
+	index = idr_alloc(&host_index_idr, shost, 0, 0, GFP_KERNEL);
+	if (unlikely(index < 0))
 		goto fail_kfree;
 	shost->host_no = index;
 
@@ -501,22 +501,13 @@ struct Scsi_Host *scsi_host_alloc(struct scsi_host_template *sht, int privsize)
  fail_kthread:
 	kthread_stop(shost->ehandler);
  fail_index_remove:
-	ida_simple_remove(&host_index_ida, shost->host_no);
+	idr_remove(&host_index_idr, shost->host_no);
  fail_kfree:
 	kfree(shost);
 	return NULL;
 }
 EXPORT_SYMBOL(scsi_host_alloc);
 
-static int __scsi_host_match(struct device *dev, const void *data)
-{
-	struct Scsi_Host *p;
-	const unsigned short *hostnum = data;
-
-	p = class_to_shost(dev);
-	return p->host_no == *hostnum;
-}
-
 /**
  * scsi_host_lookup - get a reference to a Scsi_Host by host no
  * @hostnum:	host number to locate
@@ -525,20 +516,14 @@ static int __scsi_host_match(struct device *dev, const void *data)
  *	A pointer to located Scsi_Host or NULL.
  *
  *	The caller must do a scsi_host_put() to drop the reference
- *	that scsi_host_get() took. The put_device() below dropped
- *	the reference from class_find_device().
+ *	that scsi_host_get() took.
  **/
 struct Scsi_Host *scsi_host_lookup(unsigned short hostnum)
 {
-	struct device *cdev;
-	struct Scsi_Host *shost = NULL;
-
-	cdev = class_find_device(&shost_class, NULL, &hostnum,
-				 __scsi_host_match);
-	if (cdev) {
-		shost = scsi_host_get(class_to_shost(cdev));
-		put_device(cdev);
-	}
+	struct Scsi_Host *shost = idr_find(&host_index_idr, hostnum);
+
+	if (shost)
+		shost = scsi_host_get(shost);
 	return shost;
 }
 EXPORT_SYMBOL(scsi_host_lookup);
@@ -600,7 +585,7 @@ int scsi_init_hosts(void)
 void scsi_exit_hosts(void)
 {
 	class_unregister(&shost_class);
-	ida_destroy(&host_index_ida);
+	idr_destroy(&host_index_idr);
 }
 
 int scsi_is_host_device(const struct device *dev)
-- 
2.25.1


      parent reply	other threads:[~2020-07-20  2:58 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-20  2:57 [PATCH v5 0/9] scsi: use xarray for devices and targets Douglas Gilbert
2020-07-20  2:57 ` [PATCH v5 1/9] scsi: convert target lookup to xarray Douglas Gilbert
2020-07-20  2:57 ` [PATCH v5 2/9] target_core_pscsi: use __scsi_device_lookup() Douglas Gilbert
2020-07-20  2:57 ` [PATCH v5 3/9] scsi: move target device list to xarray Douglas Gilbert
2020-07-20  2:57 ` [PATCH v5 4/9] scsi: remove direct device lookup per host Douglas Gilbert
2020-07-20  2:57 ` [PATCH v5 5/9] scsi_error: use xarray lookup instead of wrappers Douglas Gilbert
2020-07-20  2:57 ` [PATCH v5 6/9] scsi: avoid pointless memory allocation in scsi_alloc_target() Douglas Gilbert
2020-07-20  2:57 ` [PATCH v5 7/9] scsi: add starget_to_shost() specialization Douglas Gilbert
2020-07-20  2:57 ` [PATCH v5 8/9] scsi: simplify scsi_target() inline Douglas Gilbert
2020-07-20  2:57 ` Douglas Gilbert [this message]

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=20200720025742.349296-10-dgilbert@interlog.com \
    --to=dgilbert@interlog.com \
    --cc=hare@suse.de \
    --cc=jejb@linux.vnet.ibm.com \
    --cc=linux-scsi@vger.kernel.org \
    --cc=martin.petersen@oracle.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;
as well as URLs for NNTP newsgroup(s).