public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Stephen Hemminger <shemminger@vyatta.com>
To: Jesse Barnes <jbarnes@virtuousgeek.org>,
	Ben Hutchings <bhutchings@solarflare.com>
Cc: linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH 1/3] PCI: vpd handle longer delays in access
Date: Thu, 04 Sep 2008 13:56:37 -0700	[thread overview]
Message-ID: <20080904205718.543005986@vyatta.com> (raw)
In-Reply-To: 20080904205636.130211023@vyatta.com

[-- Attachment #1: vpd-delay.patch --]
[-- Type: text/plain, Size: 3108 bytes --]

Accessing the VPD area can take a long time.  The existing 
VPD access code fails consistently on my hardware. There are comments
in the SysKonnect vendor driver that it can take up to 13ms per word. 

Change the access routines to:
  * use a mutex rather than spinning with IRQ's disabled and lock held
  * have a longer timeout
  * call schedule while spinning to provide some responsivness

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>


--- a/drivers/pci/access.c	2008-09-04 09:06:51.000000000 -0700
+++ b/drivers/pci/access.c	2008-09-04 10:16:52.000000000 -0700
@@ -133,7 +133,7 @@ PCI_USER_WRITE_CONFIG(dword, u32)
 
 struct pci_vpd_pci22 {
 	struct pci_vpd base;
-	spinlock_t lock; /* controls access to hardware and the flags */
+	struct mutex lock;
 	u8	cap;
 	bool	busy;
 	bool	flag; /* value of F bit to wait for */
@@ -144,29 +144,33 @@ static int pci_vpd_pci22_wait(struct pci
 {
 	struct pci_vpd_pci22 *vpd =
 		container_of(dev->vpd, struct pci_vpd_pci22, base);
-	u16 flag, status;
-	int wait;
+	u16 flag = vpd->flag ? PCI_VPD_ADDR_F : 0;
+	unsigned long timeout = jiffies + (vpd->flag ? HZ/50 : HZ/10);
+	u16 status;
 	int ret;
 
 	if (!vpd->busy)
 		return 0;
 
-	flag = vpd->flag ? PCI_VPD_ADDR_F : 0;
-	wait = vpd->flag ? 10 : 1000; /* read: 100 us; write: 10 ms */
 	for (;;) {
-		ret = pci_user_read_config_word(dev,
-						vpd->cap + PCI_VPD_ADDR,
+		ret = pci_user_read_config_word(dev, vpd->cap + PCI_VPD_ADDR,
 						&status);
 		if (ret < 0)
-			return ret;
+			break;
+
 		if ((status & PCI_VPD_ADDR_F) == flag) {
 			vpd->busy = false;
-			return 0;
+			break;
 		}
-		if (wait-- == 0)
+
+		if (time_after(jiffies, timeout))
 			return -ETIMEDOUT;
-		udelay(10);
+		if (signal_pending(current))
+			return -EINTR;
+		yield();
 	}
+
+	return ret;
 }
 
 static int pci_vpd_pci22_read(struct pci_dev *dev, int pos, int size,
@@ -183,7 +187,9 @@ static int pci_vpd_pci22_read(struct pci
 	if (size == 0)
 		return 0;
 
-	spin_lock_irq(&vpd->lock);
+	ret = mutex_lock_killable(&vpd->lock);
+	if (ret)
+		return ret;
 	ret = pci_vpd_pci22_wait(dev);
 	if (ret < 0)
 		goto out;
@@ -199,7 +205,7 @@ static int pci_vpd_pci22_read(struct pci
 	ret = pci_user_read_config_dword(dev, vpd->cap + PCI_VPD_DATA,
 					 &val);
 out:
-	spin_unlock_irq(&vpd->lock);
+	mutex_unlock(&vpd->lock);
 	if (ret < 0)
 		return ret;
 
@@ -231,7 +237,9 @@ static int pci_vpd_pci22_write(struct pc
 	val |= ((u8) *buf++) << 16;
 	val |= ((u32)(u8) *buf++) << 24;
 
-	spin_lock_irq(&vpd->lock);
+	ret = mutex_lock_killable(&vpd->lock);
+	if (ret)
+		return ret;
 	ret = pci_vpd_pci22_wait(dev);
 	if (ret < 0)
 		goto out;
@@ -247,7 +255,7 @@ static int pci_vpd_pci22_write(struct pc
 	vpd->flag = 0;
 	ret = pci_vpd_pci22_wait(dev);
 out:
-	spin_unlock_irq(&vpd->lock);
+	mutex_unlock(&vpd->lock);
 	if (ret < 0)
 		return ret;
 
@@ -279,7 +287,7 @@ int pci_vpd_pci22_init(struct pci_dev *d
 
 	vpd->base.len = PCI_VPD_PCI22_SIZE;
 	vpd->base.ops = &pci_vpd_pci22_ops;
-	spin_lock_init(&vpd->lock);
+	mutex_init(&vpd->lock);
 	vpd->cap = cap;
 	vpd->busy = false;
 	dev->vpd = &vpd->base;

-- 


  reply	other threads:[~2008-09-04 21:49 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-09-04 20:56 [PATCH 0/3] PCI VPD changes Stephen Hemminger
2008-09-04 20:56 ` Stephen Hemminger [this message]
2008-09-05  9:11   ` [PATCH 1/3] PCI: vpd handle longer delays in access Peter Zijlstra
2008-09-05 12:40   ` Matthew Wilcox
2008-09-08 20:40   ` Andrew Morton
2008-09-08 21:08     ` Stephen Hemminger
2008-09-08 21:26       ` Arjan van de Ven
2008-09-09 16:55         ` Stephen Hemminger
2008-09-09 17:01           ` Arjan van de Ven
2008-09-04 20:56 ` [PATCH 2/3] PCI: revise VPD access interface Stephen Hemminger
2008-09-05 11:02   ` Ben Hutchings
2008-09-08 20:46   ` Andrew Morton
2008-09-04 20:56 ` [PATCH 3/3] PCI: add interface to set visible size of VPD Stephen Hemminger
2008-09-05 11:07   ` Ben Hutchings

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=20080904205718.543005986@vyatta.com \
    --to=shemminger@vyatta.com \
    --cc=bhutchings@solarflare.com \
    --cc=jbarnes@virtuousgeek.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@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