From: "Stephen M. Cameron" <scameron@beardog.cce.hp.com>
To: james.bottomley@parallels.com
Cc: webb.scales@hp.com, martin.petersen@oracle.com,
linux-scsi@vger.kernel.org, justin.lindley@hp.com,
stephenmcameron@gmail.com, joseph.t.handzik@hp.com,
thenzl@redhat.com, michael.miller@canonical.com,
scott.teel@hp.com, hch@lst.de, dan.carpenter@oracle.com
Subject: [PATCH 2 13/24] hpsa: Rearrange start_io to avoid one unlock/lock sequence in main io path
Date: Thu, 29 May 2014 10:53:28 -0500 [thread overview]
Message-ID: <20140529155328.8180.36330.stgit@beardog.cce.hp.com> (raw)
In-Reply-To: <20140529154739.8180.50710.stgit@beardog.cce.hp.com>
From: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Reviewed-by: Joe Handzik <joseph.t.handzik@hp.com>
---
drivers/scsi/hpsa.c | 24 ++++++++++++++++--------
1 files changed, 16 insertions(+), 8 deletions(-)
diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c
index ad73017..c6ed5ea 100644
--- a/drivers/scsi/hpsa.c
+++ b/drivers/scsi/hpsa.c
@@ -194,7 +194,8 @@ static int number_of_controllers;
static irqreturn_t do_hpsa_intr_intx(int irq, void *dev_id);
static irqreturn_t do_hpsa_intr_msi(int irq, void *dev_id);
static int hpsa_ioctl(struct scsi_device *dev, int cmd, void *arg);
-static void start_io(struct ctlr_info *h);
+static void lock_and_start_io(struct ctlr_info *h);
+static void start_io(struct ctlr_info *h, unsigned long *flags);
#ifdef CONFIG_COMPAT
static int hpsa_compat_ioctl(struct scsi_device *dev, int cmd, void *arg);
@@ -845,8 +846,8 @@ static void enqueue_cmd_and_start_io(struct ctlr_info *h,
spin_lock_irqsave(&h->lock, flags);
addQ(&h->reqQ, c);
h->Qdepth++;
+ start_io(h, &flags);
spin_unlock_irqrestore(&h->lock, flags);
- start_io(h);
}
static inline void removeQ(struct CommandList *c)
@@ -5452,13 +5453,12 @@ static void __iomem *remap_pci_mem(ulong base, ulong size)
/* Takes cmds off the submission queue and sends them to the hardware,
* then puts them on the queue of cmds waiting for completion.
+ * Assumes h->lock is held
*/
-static void start_io(struct ctlr_info *h)
+static void start_io(struct ctlr_info *h, unsigned long *flags)
{
struct CommandList *c;
- unsigned long flags;
- spin_lock_irqsave(&h->lock, flags);
while (!list_empty(&h->reqQ)) {
c = list_entry(h->reqQ.next, struct CommandList, list);
/* can't do anything if fifo is full */
@@ -5483,10 +5483,18 @@ static void start_io(struct ctlr_info *h)
h->commands_outstanding++;
/* Tell the controller execute command */
- spin_unlock_irqrestore(&h->lock, flags);
+ spin_unlock_irqrestore(&h->lock, *flags);
h->access.submit_command(h, c);
- spin_lock_irqsave(&h->lock, flags);
+ spin_lock_irqsave(&h->lock, *flags);
}
+}
+
+static void lock_and_start_io(struct ctlr_info *h)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&h->lock, flags);
+ start_io(h, &flags);
spin_unlock_irqrestore(&h->lock, flags);
}
@@ -5554,7 +5562,7 @@ static inline void finish_cmd(struct CommandList *c)
else if (c->cmd_type == CMD_IOCTL_PEND)
complete(c->waiting);
if (unlikely(io_may_be_stalled))
- start_io(h);
+ lock_and_start_io(h);
}
static inline u32 hpsa_tag_contains_index(u32 tag)
next prev parent reply other threads:[~2014-05-29 15:54 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-05-29 15:52 [PATCH 2 00/24] Resend of May 2014 patches for hpsa driver Stephen M. Cameron
2014-05-29 15:52 ` [PATCH 2 01/24] hpsa: add new Smart Array PCI IDs (May 2014) Stephen M. Cameron
2014-05-29 15:52 ` [PATCH 2 02/24] hpsa: fix missing check of kzalloc return value Stephen M. Cameron
2014-05-29 15:52 ` [PATCH 2 03/24] hpsa: remove unused fields from struct ctlr_info Stephen M. Cameron
2014-05-29 15:52 ` [PATCH 2 04/24] hpsa: allow passthru ioctls to work with bidirectional commands Stephen M. Cameron
2014-05-29 15:52 ` [PATCH 2 05/24] hpsa: change doorbell reset delay to ten seconds Stephen M. Cameron
2014-05-29 15:52 ` [PATCH 2 06/24] hpsa: use gcc aligned attribute instead of manually padding structs Stephen M. Cameron
2014-05-29 15:52 ` [PATCH 2 07/24] hpsa: remove dev_dbg() calls from hot paths Stephen M. Cameron
2014-05-29 15:53 ` [PATCH 2 08/24] hpsa: choose number of reply queues more intelligently Stephen M. Cameron
2014-06-02 9:27 ` Christoph Hellwig
2014-06-02 14:52 ` scameron
2014-06-02 15:00 ` scameron
2014-05-29 15:53 ` [PATCH 2 09/24] hpsa: allocate reply queues individually Stephen M. Cameron
2014-05-29 15:53 ` [PATCH 2 10/24] hpsa: set irq affinity hints to route MSI-X vectors across CPUs Stephen M. Cameron
2014-05-29 15:53 ` [PATCH 2 11/24] hpsa: use per-cpu variable for lockup_detected Stephen M. Cameron
2014-05-29 15:53 ` [PATCH 2 12/24] hpsa: avoid unnecessary readl on every command submission Stephen M. Cameron
2014-05-29 15:53 ` Stephen M. Cameron [this message]
2014-05-29 15:53 ` [PATCH 2 14/24] hpsa: define extended_report_lun_entry data structure Stephen M. Cameron
2014-05-29 15:53 ` [PATCH 2 15/24] hpsa: kill annoying messages about SSD Smart Path retries Stephen M. Cameron
2014-05-29 15:53 ` [PATCH 2 16/24] hpsa: fix event filtering to prevent excessive rescans with old firmware Stephen M. Cameron
2014-05-29 15:53 ` [PATCH 2 17/24] hpsa: remove bad unlikely annotation from device list updating code Stephen M. Cameron
2014-05-29 15:53 ` [PATCH 2 18/24] hpsa: report check condition even if no sense data present for ioaccel2 mode Stephen M. Cameron
2014-05-29 15:54 ` [PATCH 2 19/24] hpsa: fix memory leak in hpsa_hba_mode_enabled Stephen M. Cameron
2014-05-29 15:54 ` [PATCH 2 20/24] hpsa: do not ignore failure of sense controller parameters command Stephen M. Cameron
2014-05-29 15:54 ` [PATCH 2 21/24] hpsa: remove messages about volume status VPD inquiry page not supported Stephen M. Cameron
2014-05-29 15:54 ` [PATCH 2 22/24] hpsa: fix bad comparison of signed with unsigned in hpsa_update_scsi_devices Stephen M. Cameron
2014-05-29 15:54 ` [PATCH 2 23/24] hpsa: return -ENOMEM not -1 on kzalloc failure in hpsa_get_device_id Stephen M. Cameron
2014-05-29 15:54 ` [PATCH 2 24/24] hpsa: fix handling of hpsa_volume_offline return value Stephen M. Cameron
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=20140529155328.8180.36330.stgit@beardog.cce.hp.com \
--to=scameron@beardog.cce.hp.com \
--cc=dan.carpenter@oracle.com \
--cc=hch@lst.de \
--cc=james.bottomley@parallels.com \
--cc=joseph.t.handzik@hp.com \
--cc=justin.lindley@hp.com \
--cc=linux-scsi@vger.kernel.org \
--cc=martin.petersen@oracle.com \
--cc=michael.miller@canonical.com \
--cc=scott.teel@hp.com \
--cc=stephenmcameron@gmail.com \
--cc=thenzl@redhat.com \
--cc=webb.scales@hp.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).