From: Dan Williams <dan.j.williams@intel.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: "james.bottomley@suse.de" <james.bottomley@suse.de>,
"Jiang, Dave" <dave.jiang@intel.com>,
"linux-scsi@vger.kernel.org" <linux-scsi@vger.kernel.org>,
"Danecki, Jacek" <jacek.danecki@intel.com>,
"Ciechanowski, Ed" <ed.ciechanowski@intel.com>,
"Skirvin, Jeffrey D" <jeffrey.d.skirvin@intel.com>,
"Nadolski, Edmund" <edmund.nadolski@intel.com>
Subject: Re: [RFC PATCH 1/6] isci: initialization
Date: Fri, 18 Feb 2011 16:23:21 -0800 [thread overview]
Message-ID: <1298075001.19161.44.camel@dwillia2-linux> (raw)
In-Reply-To: <20110217082544.GA23662@infradead.org>
On Thu, 2011-02-17 at 00:25 -0800, Christoph Hellwig wrote:
> > +irqreturn_t isci_isr(int vec, void *data)
> > +{
> > + struct isci_host *isci_host
> > + = (struct isci_host *)data;
> > + struct scic_controller_handler_methods *handlers
> > + = &isci_host->scic_irq_handlers[SCI_MSIX_NORMAL_VECTOR];
> > + irqreturn_t ret = IRQ_NONE;
>
> > + if (isci_host_get_state(isci_host) != isci_starting
> > + && handlers->interrupt_handler) {
>
> Also there should be no need for a state check here. register_irq
> must not happen before you're ready to handle the interrupt.
>
Yes, the state checks are not needed. Here is the proposed clean up.
>From 2f89431766a7178e4a1beb87d3606d020f981e94 Mon Sep 17 00:00:00 2001
From: Dan Williams <dan.j.williams@intel.com>
Date: Fri, 18 Feb 2011 09:25:07 -0800
Subject: [PATCH] isci: cleanup "starting" state handling
The lldd actively disallows requests in the "starting" state. Retrying
or holding off commands in this state is sub-optimal:
1/ it adds another state check to the fast path
2/ retrying can cause libsas to give up
However, isci's ->lldd_dev_found() routine already waits for controller
start to complete before allowing further progress. Checking the
"starting" state in isci_task_execute_task and the isr is redundant and
misleading. Clean this up and introduce a controller-wide event queue
to start reeling in "completion" proliferation in the driver.
The "stopping" state cleanups are in a similar vein, rely on the the isr
and other paths being precluded from occurring rather than implementing
state checking logic.
Reported-by: Christoph Hellwig <hch@infradead.org>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Edmund Nadolski <edmund.nadolski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
drivers/scsi/isci/host.c | 145 +++++++++++++------------------------
drivers/scsi/isci/host.h | 17 ++++-
drivers/scsi/isci/remote_device.c | 4 +-
drivers/scsi/isci/task.c | 3 +-
4 files changed, 67 insertions(+), 102 deletions(-)
diff --git a/drivers/scsi/isci/host.c b/drivers/scsi/isci/host.c
index b66e620..dbdc3ba 100644
--- a/drivers/scsi/isci/host.c
+++ b/drivers/scsi/isci/host.c
@@ -67,15 +67,8 @@ irqreturn_t isci_msix_isr(int vec, void *data)
struct isci_host *ihost = data;
struct scic_sds_controller *scic = ihost->core_controller;
- if (isci_host_get_state(ihost) != isci_starting) {
- if (scic_sds_controller_isr(scic)) {
- if (isci_host_get_state(ihost) != isci_stopped)
- tasklet_schedule(&ihost->completion_tasklet);
- else
- dev_dbg(&ihost->pdev->dev,
- "%s: controller stopped\n", __func__);
- }
- }
+ if (scic_sds_controller_isr(scic))
+ tasklet_schedule(&ihost->completion_tasklet);
return IRQ_HANDLED;
}
@@ -89,22 +82,10 @@ irqreturn_t isci_intx_isr(int vec, void *data)
for_each_isci_host(ihost, pdev) {
struct scic_sds_controller *scic = ihost->core_controller;
- if (isci_host_get_state(ihost) != isci_starting) {
- if (scic_sds_controller_isr(scic)) {
- if (isci_host_get_state(ihost) != isci_stopped)
- tasklet_schedule(&ihost->completion_tasklet);
- else
- dev_dbg(&ihost->pdev->dev,
- "%s: controller stopped\n",
- __func__);
- ret = IRQ_HANDLED;
- }
- } else
- dev_warn(&ihost->pdev->dev,
- "%s: get_handler_methods failed, "
- "ihost->status = 0x%x\n",
- __func__,
- isci_host_get_state(ihost));
+ if (scic_sds_controller_isr(scic)) {
+ tasklet_schedule(&ihost->completion_tasklet);
+ ret = IRQ_HANDLED;
+ }
}
return ret;
}
@@ -118,26 +99,19 @@ irqreturn_t isci_intx_isr(int vec, void *data)
* core library.
*
*/
-void isci_host_start_complete(
- struct isci_host *isci_host,
- enum sci_status completion_status)
+void isci_host_start_complete(struct isci_host *ihost, enum sci_status completion_status)
{
- if (completion_status == SCI_SUCCESS) {
- dev_dbg(&isci_host->pdev->dev,
- "%s: completion_status: SCI_SUCCESS\n", __func__);
- isci_host_change_state(isci_host, isci_ready);
- complete_all(&isci_host->start_complete);
- } else
- dev_err(&isci_host->pdev->dev,
- "controller start failed with "
- "completion_status = 0x%x;",
- completion_status);
-
+ if (completion_status != SCI_SUCCESS)
+ dev_info(&ihost->pdev->dev,
+ "controller start timed out, continuing...\n");
+ isci_host_change_state(ihost, isci_ready);
+ clear_bit(IHOST_START_PENDING, &ihost->flags);
+ wake_up(&ihost->eventq);
}
int isci_host_scan_finished(struct Scsi_Host *shost, unsigned long time)
{
- struct isci_host *isci_host = isci_host_from_sas_ha(SHOST_TO_SAS_HA(shost));
+ struct isci_host *ihost = isci_host_from_sas_ha(SHOST_TO_SAS_HA(shost));
/**
* check interrupt_handler's status and call completion_handler if true,
@@ -148,61 +122,44 @@ int isci_host_scan_finished(struct Scsi_Host *shost, unsigned long time)
* continue to return zero from thee scan_finished routine until
* the scic_cb_controller_start_complete() call comes from the core.
**/
- if (scic_sds_controller_isr(isci_host->core_controller))
- scic_sds_controller_completion_handler(isci_host->core_controller);
+ if (scic_sds_controller_isr(ihost->core_controller))
+ scic_sds_controller_completion_handler(ihost->core_controller);
- if (isci_starting == isci_host_get_state(isci_host)
- && time < (HZ * 10)) {
- dev_dbg(&isci_host->pdev->dev,
- "%s: isci_host->status = %d, time = %ld\n",
- __func__, isci_host_get_state(isci_host), time);
+ if (test_bit(IHOST_START_PENDING, &ihost->flags) && time < HZ*10) {
+ dev_dbg(&ihost->pdev->dev,
+ "%s: ihost->status = %d, time = %ld\n",
+ __func__, isci_host_get_state(ihost), time);
return 0;
}
- dev_dbg(&isci_host->pdev->dev,
- "%s: isci_host->status = %d, time = %ld\n",
- __func__, isci_host_get_state(isci_host), time);
+ dev_dbg(&ihost->pdev->dev,
+ "%s: ihost->status = %d, time = %ld\n",
+ __func__, isci_host_get_state(ihost), time);
- scic_controller_enable_interrupts(isci_host->core_controller);
+ scic_controller_enable_interrupts(ihost->core_controller);
return 1;
}
-
-/**
- * isci_host_scan_start() - This function is one of the SCSI Host Template
- * function, called by the SCSI mid layer berfore a target scan begins. The
- * core library controller start routine is called from here.
- * @shost: This parameter specifies the SCSI host to be scanned
- *
- */
void isci_host_scan_start(struct Scsi_Host *shost)
{
- struct isci_host *isci_host;
-
- isci_host = isci_host_from_sas_ha(SHOST_TO_SAS_HA(shost));
- isci_host_change_state(isci_host, isci_starting);
+ struct isci_host *ihost = isci_host_from_sas_ha(SHOST_TO_SAS_HA(shost));
+ struct scic_sds_controller *scic = ihost->core_controller;
+ unsigned long tmo = scic_controller_get_suggested_start_timeout(scic);
- scic_controller_disable_interrupts(isci_host->core_controller);
- init_completion(&isci_host->start_complete);
- scic_controller_start(
- isci_host->core_controller,
- scic_controller_get_suggested_start_timeout(
- isci_host->core_controller)
- );
+ set_bit(IHOST_START_PENDING, &ihost->flags);
+ scic_controller_disable_interrupts(ihost->core_controller);
+ scic_controller_start(scic, tmo);
}
-void isci_host_stop_complete(
- struct isci_host *isci_host,
- enum sci_status completion_status)
+void isci_host_stop_complete(struct isci_host *ihost, enum sci_status completion_status)
{
- isci_host_change_state(isci_host, isci_stopped);
- scic_controller_disable_interrupts(
- isci_host->core_controller
- );
- complete(&isci_host->stop_complete);
+ isci_host_change_state(ihost, isci_stopped);
+ scic_controller_disable_interrupts(ihost->core_controller);
+ clear_bit(IHOST_STOP_PENDING, &ihost->flags);
+ wake_up(&ihost->eventq);
}
static struct coherent_memory_info *isci_host_alloc_mdl_struct(
@@ -370,31 +327,26 @@ static void isci_host_completion_routine(unsigned long data)
}
-void isci_host_deinit(
- struct isci_host *isci_host)
+void isci_host_deinit(struct isci_host *ihost)
{
+ struct scic_sds_controller *scic = ihost->core_controller;
int i;
- isci_host_change_state(isci_host, isci_stopping);
+ isci_host_change_state(ihost, isci_stopping);
for (i = 0; i < SCI_MAX_PORTS; i++) {
- struct isci_port *port = &isci_host->isci_ports[i];
- struct isci_remote_device *device, *tmpdev;
- list_for_each_entry_safe(device, tmpdev,
- &port->remote_dev_list, node) {
- isci_remote_device_change_state(device, isci_stopping);
- isci_remote_device_stop(device);
+ struct isci_port *port = &ihost->isci_ports[i];
+ struct isci_remote_device *idev, *d;
+
+ list_for_each_entry_safe(idev, d, &port->remote_dev_list, node) {
+ isci_remote_device_change_state(idev, isci_stopping);
+ isci_remote_device_stop(idev);
}
}
- /* stop the comtroller and wait for completion. */
- init_completion(&isci_host->stop_complete);
- scic_controller_stop(
- isci_host->core_controller,
- SCIC_CONTROLLER_STOP_TIMEOUT
- );
- wait_for_completion(&isci_host->stop_complete);
- /* next, reset the controller. */
- scic_controller_reset(isci_host->core_controller);
+ set_bit(IHOST_STOP_PENDING, &ihost->flags);
+ scic_controller_stop(scic, SCIC_CONTROLLER_STOP_TIMEOUT);
+ wait_for_stop(ihost);
+ scic_controller_reset(scic);
}
static int isci_verify_firmware(const struct firmware *fw,
@@ -506,6 +458,7 @@ int isci_host_init(struct isci_host *isci_host)
spin_lock_init(&isci_host->state_lock);
spin_lock_init(&isci_host->scic_lock);
spin_lock_init(&isci_host->queue_lock);
+ init_waitqueue_head(&isci_host->eventq);
isci_host_change_state(isci_host, isci_starting);
isci_host->can_queue = ISCI_CAN_QUEUE_VAL;
diff --git a/drivers/scsi/isci/host.h b/drivers/scsi/isci/host.h
index 421d3de..26768c5 100644
--- a/drivers/scsi/isci/host.h
+++ b/drivers/scsi/isci/host.h
@@ -109,13 +109,15 @@ struct isci_host {
u8 sas_addr[SAS_ADDR_SIZE];
enum isci_status status;
+ #define IHOST_START_PENDING 0
+ #define IHOST_STOP_PENDING 1
+ unsigned long flags;
+ wait_queue_head_t eventq;
struct Scsi_Host *shost;
struct tasklet_struct completion_tasklet;
struct list_head mdl_struct_list;
struct list_head requests_to_complete;
struct list_head requests_to_abort;
- struct completion stop_complete;
- struct completion start_complete;
spinlock_t scic_lock;
struct isci_host *next;
};
@@ -202,6 +204,17 @@ static inline void isci_host_can_dequeue(
spin_unlock_irqrestore(&isci_host->queue_lock, flags);
}
+static inline void wait_for_start(struct isci_host *ihost)
+{
+ wait_event(ihost->eventq, !test_bit(IHOST_START_PENDING, &ihost->flags));
+}
+
+static inline void wait_for_stop(struct isci_host *ihost)
+{
+ wait_event(ihost->eventq, !test_bit(IHOST_STOP_PENDING, &ihost->flags));
+}
+
+
/**
* isci_host_from_sas_ha() - This accessor retrieves the isci_host object
* reference from the Linux sas_ha_struct reference.
diff --git a/drivers/scsi/isci/remote_device.c b/drivers/scsi/isci/remote_device.c
index dbf3c82..936f229 100644
--- a/drivers/scsi/isci/remote_device.c
+++ b/drivers/scsi/isci/remote_device.c
@@ -507,6 +507,8 @@ int isci_remote_device_found(struct domain_device *domain_dev)
dev_dbg(&isci_host->pdev->dev,
"%s: domain_device = %p\n", __func__, domain_dev);
+ wait_for_start(isci_host);
+
sas_port = domain_dev->port;
sas_phy = list_first_entry(&sas_port->phy_list, struct asd_sas_phy,
port_phy_el);
@@ -560,8 +562,6 @@ int isci_remote_device_found(struct domain_device *domain_dev)
return -ENODEV;
}
- wait_for_completion(&isci_host->start_complete);
-
return 0;
}
/**
diff --git a/drivers/scsi/isci/task.c b/drivers/scsi/isci/task.c
index 5e6f558..6f98f6c 100644
--- a/drivers/scsi/isci/task.c
+++ b/drivers/scsi/isci/task.c
@@ -164,8 +164,7 @@ int isci_task_execute_task(struct sas_task *task, int num, gfp_t gfp_flags)
* for the quiesce spinlock.
*/
- if (isci_host_get_state(isci_host) == isci_starting ||
- (device && ((isci_remote_device_get_state(device) == isci_ready) ||
+ if ((device && ((isci_remote_device_get_state(device) == isci_ready) ||
(isci_remote_device_get_state(device) == isci_host_quiesce)))) {
/* Forces a retry from scsi mid layer. */
--
1.7.2.3
next prev parent reply other threads:[~2011-02-19 0:02 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-02-07 0:34 [RFC PATCH 0/6] isci: initial driver release (part1: intro and lldd) Dan Williams
2011-02-07 0:34 ` [RFC PATCH 1/6] isci: initialization Dan Williams
2011-02-17 8:22 ` Jeff Garzik
2011-02-19 0:12 ` Dan Williams
2011-02-17 8:25 ` Christoph Hellwig
2011-02-19 0:23 ` Dan Williams [this message]
2011-03-04 23:35 ` James Bottomley
2011-03-08 1:51 ` Dan Williams
2011-03-18 16:51 ` Christoph Hellwig
2011-02-07 0:34 ` [RFC PATCH 2/6] isci: task (libsas interface support) Dan Williams
2011-02-09 15:01 ` David Milburn
2011-02-14 7:14 ` Dan Williams
2011-02-16 18:48 ` David Milburn
2011-02-16 19:35 ` David Milburn
2011-02-07 0:34 ` [RFC PATCH 3/6] isci: request (core request infrastructure) Dan Williams
2011-03-18 16:41 ` Christoph Hellwig
2011-02-07 0:34 ` [RFC PATCH 4/6] isci: hardware / topology event handling Dan Williams
2011-03-18 16:18 ` Christoph Hellwig
2011-03-23 8:15 ` Dan Williams
2011-03-23 8:40 ` Christoph Hellwig
2011-03-23 9:04 ` Dan Williams
2011-03-23 9:08 ` Christoph Hellwig
2011-03-24 0:07 ` Dan Williams
2011-03-24 6:26 ` Christoph Hellwig
2011-03-25 0:57 ` Dan Williams
2011-03-25 19:45 ` Christoph Hellwig
2011-03-25 21:39 ` Dan Williams
2011-03-25 22:07 ` Christoph Hellwig
2011-03-25 22:34 ` Dan Williams
2011-03-27 22:28 ` Christoph Hellwig
2011-03-29 1:11 ` Dan Williams
2011-03-30 0:37 ` Dan Williams
2011-02-07 0:35 ` [RFC PATCH 5/6] isci: phy, port, and remote device Dan Williams
2011-02-07 0:35 ` [RFC PATCH 6/6] isci: sata support and phy settings via request_firmware() Dan Williams
2011-02-07 7:58 ` [RFC PATCH 1/6] isci: initialization jack_wang
2011-02-14 7:49 ` 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=1298075001.19161.44.camel@dwillia2-linux \
--to=dan.j.williams@intel.com \
--cc=dave.jiang@intel.com \
--cc=ed.ciechanowski@intel.com \
--cc=edmund.nadolski@intel.com \
--cc=hch@infradead.org \
--cc=jacek.danecki@intel.com \
--cc=james.bottomley@suse.de \
--cc=jeffrey.d.skirvin@intel.com \
--cc=linux-scsi@vger.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;
as well as URLs for NNTP newsgroup(s).