Linux-HyperV List
 help / color / mirror / Atom feed
From: Melanie Plageman <melanieplageman@gmail.com>
To: linux-scsi@vger.kernel.org
Cc: andres@anarazel.de, kys@microsoft.com, haiyangz@microsoft.com,
	sthemmin@microsoft.com, wei.liu@kernel.org, jejb@linux.ibm.com,
	martin.petersen@oracle.com, linux-hyperv@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	"Melanie Plageman (Microsoft)" <melanieplageman@gmail.com>
Subject: [PATCH v3] scsi: storvsc: Parameterize number hardware queues
Date: Thu, 11 Feb 2021 23:18:03 +0000	[thread overview]
Message-ID: <20210211231803.25463-1-melanieplageman@gmail.com> (raw)

From: "Melanie Plageman (Microsoft)" <melanieplageman@gmail.com>

Add ability to set the number of hardware queues with new module parameter,
storvsc_max_hw_queues. The default value remains the number of CPUs.  This
functionality is useful in some environments (e.g. Microsoft Azure) where
decreasing the number of hardware queues has been shown to improve
performance.

Signed-off-by: Melanie Plageman (Microsoft) <melanieplageman@gmail.com>
---
 drivers/scsi/storvsc_drv.c | 28 ++++++++++++++++++++++++++--
 1 file changed, 26 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/storvsc_drv.c b/drivers/scsi/storvsc_drv.c
index 2e4fa77445fd..a64e6664c915 100644
--- a/drivers/scsi/storvsc_drv.c
+++ b/drivers/scsi/storvsc_drv.c
@@ -378,10 +378,14 @@ static u32 max_outstanding_req_per_channel;
 static int storvsc_change_queue_depth(struct scsi_device *sdev, int queue_depth);
 
 static int storvsc_vcpus_per_sub_channel = 4;
+static int storvsc_max_hw_queues = -1;
 
 module_param(storvsc_ringbuffer_size, int, S_IRUGO);
 MODULE_PARM_DESC(storvsc_ringbuffer_size, "Ring buffer size (bytes)");
 
+module_param(storvsc_max_hw_queues, int, S_IRUGO|S_IWUSR);
+MODULE_PARM_DESC(storvsc_max_hw_queues, "Maximum number of hardware queues");
+
 module_param(storvsc_vcpus_per_sub_channel, int, S_IRUGO);
 MODULE_PARM_DESC(storvsc_vcpus_per_sub_channel, "Ratio of VCPUs to subchannels");
 
@@ -1897,6 +1901,7 @@ static int storvsc_probe(struct hv_device *device,
 {
 	int ret;
 	int num_cpus = num_online_cpus();
+	int num_present_cpus = num_present_cpus();
 	struct Scsi_Host *host;
 	struct hv_host_device *host_dev;
 	bool dev_is_ide = ((dev_id->driver_data == IDE_GUID) ? true : false);
@@ -2004,8 +2009,19 @@ static int storvsc_probe(struct hv_device *device,
 	 * For non-IDE disks, the host supports multiple channels.
 	 * Set the number of HW queues we are supporting.
 	 */
-	if (!dev_is_ide)
-		host->nr_hw_queues = num_present_cpus();
+	if (!dev_is_ide) {
+		if (storvsc_max_hw_queues == -1)
+			host->nr_hw_queues = num_present_cpus;
+		else if (storvsc_max_hw_queues > num_present_cpus ||
+			 storvsc_max_hw_queues == 0 ||
+			storvsc_max_hw_queues < -1) {
+			storvsc_log(device, STORVSC_LOGGING_WARN,
+				"Resetting invalid storvsc_max_hw_queues value to default.\n");
+			host->nr_hw_queues = num_present_cpus;
+			storvsc_max_hw_queues = -1;
+		} else
+			host->nr_hw_queues = storvsc_max_hw_queues;
+	}
 
 	/*
 	 * Set the error handler work queue.
@@ -2169,6 +2185,14 @@ static int __init storvsc_drv_init(void)
 		vmscsi_size_delta,
 		sizeof(u64)));
 
+	if (storvsc_max_hw_queues > num_present_cpus() ||
+		storvsc_max_hw_queues == 0 ||
+		storvsc_max_hw_queues < -1) {
+		pr_warn("Setting storvsc_max_hw_queues to -1. %d is invalid.\n",
+			storvsc_max_hw_queues);
+		storvsc_max_hw_queues = -1;
+	}
+
 #if IS_ENABLED(CONFIG_SCSI_FC_ATTRS)
 	fc_transport_template = fc_attach_transport(&fc_transport_functions);
 	if (!fc_transport_template)
-- 
2.20.1


             reply	other threads:[~2021-02-11 23:19 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-11 23:18 Melanie Plageman [this message]
2021-02-12 16:35 ` [PATCH v3] scsi: storvsc: Parameterize number hardware queues Michael Kelley
2021-02-18  0:04   ` Melanie Plageman
2021-02-18  5:24     ` Michael Kelley

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=20210211231803.25463-1-melanieplageman@gmail.com \
    --to=melanieplageman@gmail.com \
    --cc=andres@anarazel.de \
    --cc=haiyangz@microsoft.com \
    --cc=jejb@linux.ibm.com \
    --cc=kys@microsoft.com \
    --cc=linux-hyperv@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=martin.petersen@oracle.com \
    --cc=sthemmin@microsoft.com \
    --cc=wei.liu@kernel.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