netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ben Dooks <ben-linux@fluff.org>
To: netdev@vger.kernel.org
Cc: jeff@garzik.org, akpm@linux-foundation.org, daniel@caiaq.de,
	laurentp@cse-semaphore.com, Ben Dooks <ben-linux@fluff.org>
Subject: [PATCH 08/24 for-2.6.25] DM9000: Do not sleep with spinlock and IRQs held
Date: Tue, 05 Feb 2008 00:02:07 +0000	[thread overview]
Message-ID: <20080205000816.090539437@fluff.org.uk> (raw)
In-Reply-To: 20080205000159.432081941@fluff.org.uk

[-- Attachment #1: simtec/simtec-drivers-net-dm9000-phy-unlocksleep.patch --]
[-- Type: text/plain, Size: 1886 bytes --]

The phy read and write routines call udelay() with the board
lock held, and with the posibility of IRQs being disabled. Since
these delays can be up to 500usec, and are only required as we
have to save the chip's address register. 

To improve the behaviour, hold the lock whilst we are writing
and then restore the state before the delay and then repeat
the process once the delay has happened.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>

Index: linux-2.6.24-quilt3/drivers/net/dm9000.c
===================================================================
--- linux-2.6.24-quilt3.orig/drivers/net/dm9000.c
+++ linux-2.6.24-quilt3/drivers/net/dm9000.c
@@ -1130,7 +1130,15 @@ dm9000_phy_read(struct net_device *dev, 
 	iow(db, DM9000_EPAR, DM9000_PHY | reg);
 
 	iow(db, DM9000_EPCR, 0xc);	/* Issue phyxcer read command */
+
+	writeb(reg_save, db->io_addr);
+	spin_unlock_irqrestore(&db->lock,flags);
+
 	udelay(100);		/* Wait read complete */
+
+	spin_lock_irqsave(&db->lock,flags);
+	reg_save = readb(db->io_addr);
+
 	iow(db, DM9000_EPCR, 0x0);	/* Clear phyxcer read command */
 
 	/* The read data keeps on REG_0D & REG_0E */
@@ -1138,7 +1146,6 @@ dm9000_phy_read(struct net_device *dev, 
 
 	/* restore the previous address */
 	writeb(reg_save, db->io_addr);
-
 	spin_unlock_irqrestore(&db->lock,flags);
 
 	return ret;
@@ -1167,7 +1174,15 @@ dm9000_phy_write(struct net_device *dev,
 	iow(db, DM9000_EPDRH, ((value >> 8) & 0xff));
 
 	iow(db, DM9000_EPCR, 0xa);	/* Issue phyxcer write command */
+
+	writeb(reg_save, db->io_addr);
+	spin_unlock_irqrestore(&db->lock,flags);
+
 	udelay(500);		/* Wait write complete */
+
+	spin_lock_irqsave(&db->lock,flags);
+	reg_save = readb(db->io_addr);
+
 	iow(db, DM9000_EPCR, 0x0);	/* Clear phyxcer write command */
 
 	/* restore the previous address */

-- 
Ben (ben@fluff.org, http://www.fluff.org/)

  'a smiley only costs 4 bytes'

  parent reply	other threads:[~2008-02-05  0:08 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-02-05  0:01 [PATCH 00/24 for-2.6.25] DM9000 updates for 2.6.25 Ben Dooks
2008-02-05  0:02 ` [PATCH 01/24 for-2.6.25] DM9000: Fix endian-ness of data accesses. Patch from: Laurent Pinchart <laurentp@cse-semaphore.com> Ben Dooks
2008-02-05  0:13   ` [PATCH 01/24 for-2.6.25] DM9000: Fix endian-ness of data accesses Ben Dooks
2008-02-05 22:57   ` [PATCH 01/24 for-2.6.25] DM9000: Fix endian-ness of data accesses. Patch from: Laurent Pinchart <laurentp@cse-semaphore.com> Francois Romieu
2008-02-06 11:46   ` Jeff Garzik
2008-02-07 13:28     ` Laurent Pinchart
2008-02-07 14:54       ` Jeff Garzik
2008-02-07 13:30   ` Christoph Hellwig
2008-02-05  0:02 ` [PATCH 02/24 for-2.6.25] DM9000: Add platform data to specify external phy " Ben Dooks
2008-02-05  0:14   ` [PATCH 02/24 for-2.6.25] DM9000: Add platform data to specify external phy Ben Dooks
2008-02-05  0:02 ` [PATCH 03/24 for-2.6.25] DM9000 use dev_xxx() instead of printk for output Ben Dooks
2008-02-05  0:02 ` [PATCH 04/24 for-2.6.25] DM9000 update debugging macros to use debug level Ben Dooks
2008-02-05  0:02 ` [PATCH 05/24 for-2.6.25] DM9000: Pass IRQ flags via platform resources Ben Dooks
2008-02-05  0:02 ` [PATCH 06/24 for-2.6.25] DM9000: Remove old timer based poll routines Ben Dooks
2008-02-05  0:02 ` [PATCH 07/24 for-2.6.25] DM9000: Add initial ethtool support Ben Dooks
2008-02-05 22:50   ` Francois Romieu
2008-02-06 22:03     ` Ben Dooks
2008-02-06 21:52       ` Francois Romieu
2008-02-06 22:16         ` Ben Dooks
2008-02-05  0:02 ` Ben Dooks [this message]
2008-02-05  0:02 ` [PATCH 09/24 for-2.6.25] DM9000: Use msleep() instead of udelay() Ben Dooks
2008-02-05  0:02 ` [PATCH 10/24 for-2.6.25] DM9000: Remove barely used SROM array read Ben Dooks
2008-02-05  0:02 ` [PATCH 11/24 for-2.6.25] DM9000: Add mutex to protect access Ben Dooks
2008-02-05  0:02 ` [PATCH 12/24 for-2.6.25] DM9000: Add ethtool support for reading and writing EEPROM Ben Dooks
2008-02-05  0:02 ` [PATCH 13/24 for-2.6.25] DM9000: Add ethtool control of msg_enable value Ben Dooks
2008-02-05  0:02 ` [PATCH 14/24 for-2.6.25] DM9000: Remove EEPROM initialisation code Ben Dooks
2008-02-05  0:02 ` [PATCH 15/24 for-2.6.25] DM9000: Ensure spinlock held whilst accessing EEPROM registers Ben Dooks
2008-02-05  0:02 ` [PATCH 16/24 for-2.6.25] DM9000: Remove unnecessary changelog in header comment Ben Dooks
2008-02-05  0:02 ` [PATCH 17/24 for-2.6.25] DM9000: Use netif_msg to enable debugging options Ben Dooks
2008-02-05  0:02 ` [PATCH 18/24 for-2.6.25] DM9000: Fix delays used by EEPROM read and write Ben Dooks
2008-02-05  0:02 ` [PATCH 19/24 for-2.6.25] DM9000: Remove cal_CRC() and use ether_crc_le instead Ben Dooks
2008-02-05  0:02 ` [PATCH 20/24 for-2.6.25] DM9000: Remove redudant use of "& 0xff" Ben Dooks
2008-02-05  0:02 ` [PATCH 21/24 for-2.6.25] DM9000: Add platform flag for no attached EEPROM Ben Dooks
2008-02-05  0:02 ` [PATCH 22/24 for-2.6.25] DM9000: Add support for MII ioctl() calls Ben Dooks
2008-02-05  0:02 ` [PATCH 23/24 for-2.6.25] DM9000: Update retry count whilst identifying chip Ben Dooks
2008-02-05  0:02 ` [PATCH 24/24 for-2.6.25] DM9000: Show the MAC address source after printing MAC Ben Dooks
2008-02-05  0:19 ` [PATCH 00/24 for-2.6.25] DM9000 updates for 2.6.25 Andrew Morton
2008-02-05  1:44   ` Ben Dooks
2008-02-05  9:52 ` Laurent Pinchart

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=20080205000816.090539437@fluff.org.uk \
    --to=ben-linux@fluff.org \
    --cc=akpm@linux-foundation.org \
    --cc=daniel@caiaq.de \
    --cc=jeff@garzik.org \
    --cc=laurentp@cse-semaphore.com \
    --cc=netdev@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).