[SCSI] lpfc 8.1.1 : Add polled-mode support

- Add functionality to run in polled mode only. Includes run time
  attribute to enable mode.
- Enable runtime writable hba settings for coallescing and delay parameters

Customers have requested a mode in the driver to run strictly polled.
This is generally to support an environment where the server is extremely
loaded and is looking to reclaim some cpu cycles from adapter interrupt
handling.

This patch adds a new "poll" attribute, and the following behavior:

if value is 0 (default):
  The driver uses the normal method for i/o completion. It uses the
  firmware feature of interrupt coalesing. The firmware allows a
  minimum number of i/o completions before an interrupt, or a maximum
  time delay between interrupts.  By default, the driver sets these
  to no delay (disabled) or 1 i/o - meaning coalescing is disabled.

  Attributes were provided to change the coalescing values, but it was
  a module-load time only and global across all adapters.
  This patch allows them to be writable on a per-adapter basis.

if value is 1 :
  Interrupts are left enabled, expecting that the user has tuned the
  interrupt coalescing values. When this setting is enabled, the driver
  will attempt to service completed i/o whenever new i/o is submitted
  to the adapter. If the coalescing values are large, and the i/o
  generation rate steady, an interrupt will be avoided by servicing
  completed i/o prior to the coalescing thresholds kicking in. However,
  if the i/o completion load is high enough or i/o generation slow, the
  coalescion values will ensure that completed i/o is serviced in a timely
  fashion.

if value is 3 :
  Turns off FCP i/o interrupts altogether. The coalescing values now have
  no effect. A new attribute "poll_tmo" (default 10ms) exists to set
  the polling interval for i/o completion. When this setting is enabled,
  the driver will attempt to service completed i/o and restart the
  interval timer whenever new i/o is submitted. This behavior allows for
  servicing of completed i/o sooner than the interval timer, but ensures
  that if no i/o is being issued, then the interval timer will kick in
  to service the outstanding i/o.

Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

1 files changed, 90 insertions, 2 deletions

diff --git a/drivers/scsi/lpfc/lpfc_attr.c b/drivers/scsi/lpfc/lpfc_attr.c
index 89e8222bc7cc..5625a8c2a8fd 100644
--- a/drivers/scsi/lpfc/lpfc_attr.c
+++ b/drivers/scsi/lpfc/lpfc_attr.c
@@ -278,6 +278,71 @@ lpfc_board_online_store(struct class_device *cdev, const char *buf,
                 return -EIO;
 }
 
+static ssize_t
+lpfc_poll_show(struct class_device *cdev, char *buf)
+{
+        struct Scsi_Host *host = class_to_shost(cdev);
+        struct lpfc_hba *phba = (struct lpfc_hba*)host->hostdata[0];
+
+        return snprintf(buf, PAGE_SIZE, "%#x\n", phba->cfg_poll);
+}
+
+static ssize_t
+lpfc_poll_store(struct class_device *cdev, const char *buf,
+                size_t count)
+{
+        struct Scsi_Host *host = class_to_shost(cdev);
+        struct lpfc_hba *phba = (struct lpfc_hba*)host->hostdata[0];
+        uint32_t creg_val;
+        uint32_t old_val;
+        int val=0;
+
+        if (!isdigit(buf[0]))
+                return -EINVAL;
+
+        if (sscanf(buf, "%i", &val) != 1)
+                return -EINVAL;
+
+        if ((val & 0x3) != val)
+                return -EINVAL;
+
+        spin_lock_irq(phba->host->host_lock);
+
+        old_val = phba->cfg_poll;
+
+        if (val & ENABLE_FCP_RING_POLLING) {
+                if ((val & DISABLE_FCP_RING_INT) &&
+                    !(old_val & DISABLE_FCP_RING_INT)) {
+                        creg_val = readl(phba->HCregaddr);
+                        creg_val &= ~(HC_R0INT_ENA << LPFC_FCP_RING);
+                        writel(creg_val, phba->HCregaddr);
+                        readl(phba->HCregaddr); /* flush */
+
+                        lpfc_poll_start_timer(phba);
+                }
+        } else if (val != 0x0) {
+                spin_unlock_irq(phba->host->host_lock);
+                return -EINVAL;
+        }
+
+        if (!(val & DISABLE_FCP_RING_INT) &&
+            (old_val & DISABLE_FCP_RING_INT))
+        {
+                spin_unlock_irq(phba->host->host_lock);
+                del_timer(&phba->fcp_poll_timer);
+                spin_lock_irq(phba->host->host_lock);
+                creg_val = readl(phba->HCregaddr);
+                creg_val |= (HC_R0INT_ENA << LPFC_FCP_RING);
+                writel(creg_val, phba->HCregaddr);
+                readl(phba->HCregaddr); /* flush */
+        }
+
+        phba->cfg_poll = val;
+
+        spin_unlock_irq(phba->host->host_lock);
+
+        return strlen(buf);
+}
 
 #define lpfc_param_show(attr)   \
 static ssize_t \
@@ -416,6 +481,15 @@ static CLASS_DEVICE_ATTR(management_version, S_IRUGO, management_version_show,
 static CLASS_DEVICE_ATTR(board_online, S_IRUGO | S_IWUSR,
                          lpfc_board_online_show, lpfc_board_online_store);
 
+static int lpfc_poll = 0;
+module_param(lpfc_poll, int, 0);
+MODULE_PARM_DESC(lpfc_poll, "FCP ring polling mode control:"
+                 " 0 - none,"
+                 " 1 - poll with interrupts enabled"
+                 " 3 - poll and disable FCP ring interrupts");
+
+static CLASS_DEVICE_ATTR(lpfc_poll, S_IRUGO | S_IWUSR,
+                         lpfc_poll_show, lpfc_poll_store);
 
 /*
 # lpfc_log_verbose: Only turn this flag on if you are willing to risk being
@@ -523,10 +597,10 @@ LPFC_ATTR_R(ack0, 0, 0, 1, "Enable ACK0 support");
 # is 0. Default value of cr_count is 1. The cr_count feature is disabled if
 # cr_delay is set to 0.
 */
-LPFC_ATTR(cr_delay, 0, 0, 63, "A count of milliseconds after which an"
+LPFC_ATTR_RW(cr_delay, 0, 0, 63, "A count of milliseconds after which an"
                 "interrupt response is generated");
 
-LPFC_ATTR(cr_count, 1, 1, 255, "A count of I/O completions after which an"
+LPFC_ATTR_RW(cr_count, 1, 1, 255, "A count of I/O completions after which an"
                 "interrupt response is generated");
 
 /*
@@ -553,6 +627,13 @@ LPFC_ATTR(discovery_threads, 32, 1, 64, "Maximum number of ELS commands"
 LPFC_ATTR_R(max_luns, 256, 1, 32768,
              "Maximum number of LUNs per target driver will support");
 
+/*
+# lpfc_poll_tmo: .Milliseconds driver will wait between polling FCP ring.
+# Value range is [1,255], default value is 10.
+*/
+LPFC_ATTR_RW(poll_tmo, 10, 1, 255,
+             "Milliseconds driver will wait between polling FCP ring");
+
 struct class_device_attribute *lpfc_host_attrs[] = {
         &class_device_attr_info,
         &class_device_attr_serialnum,
@@ -575,11 +656,15 @@ struct class_device_attribute *lpfc_host_attrs[] = {
         &class_device_attr_lpfc_topology,
         &class_device_attr_lpfc_scan_down,
         &class_device_attr_lpfc_link_speed,
+        &class_device_attr_lpfc_cr_delay,
+        &class_device_attr_lpfc_cr_count,
         &class_device_attr_lpfc_fdmi_on,
         &class_device_attr_lpfc_max_luns,
         &class_device_attr_nport_evt_cnt,
         &class_device_attr_management_version,
         &class_device_attr_board_online,
+        &class_device_attr_lpfc_poll,
+        &class_device_attr_lpfc_poll_tmo,
         NULL,
 };
 
@@ -1292,6 +1377,9 @@ lpfc_get_cfgparam(struct lpfc_hba *phba)
         lpfc_fdmi_on_init(phba, lpfc_fdmi_on);
         lpfc_discovery_threads_init(phba, lpfc_discovery_threads);
         lpfc_max_luns_init(phba, lpfc_max_luns);
+        lpfc_poll_tmo_init(phba, lpfc_poll_tmo);
+
+        phba->cfg_poll = lpfc_poll;
 
         /*
          * The total number of segments is the configuration value plus 2