Merge tp1.ruhr-uni-bochum.de:/home/kai/src/kernel/v2.5/linux-2.5

into tp1.ruhr-uni-bochum.de:/home/kai/src/kernel/v2.5/linux-2.5.make

Documentation/filesystems/ext3.txt

+
+Ext3 Filesystem
+===============
+
+ext3 was originally released in September 1999. Written by Stephen Tweedie
+for 2.2 branch, and ported to 2.4 kernels by Peter Braam, Andreas Dilger, 
+Andrew Morton, Alexander Viro, Ted Ts'o and Stephen Tweedie.
+
+ext3 is ext2 filesystem enhanced with journalling capabilities. 
+
+Options
+=======
+
+When mounting an ext3 filesystem, the following option are accepted:
+(*) == default
+
+jounal=update		Update the ext3 file system's journal to the 
+			current format.
+
+journal=inum		When a journal already exists, this option is 
+			ignored. Otherwise, it specifies the number of
+			the inode which will represent the ext3 file
+			system's journal file.
+
+bsddf 		(*)	Make 'df' act like BSD.
+minixdf			Make 'df' act like Minix.
+
+check=none		Don't do extra checking of bitmaps on mount.
+nocheck		
+
+debug			Extra debugging information is sent to syslog.
+
+noload			Don't load the journal on mounting.
+
+errors=remount-ro(*)	Remount the filesystem read-only on an error.
+errors=continue		Keep going on a filesystem error.
+errors=panic		Panic and halt the machine if an error occurs.
+
+grpid			Give objects the same group ID as their creator.
+bsdgroups		
+
+nogrpid		(*)	New objects have the group ID of their creator.
+sysvgroups
+
+resgid=n		The group ID which may use the reserved blocks.
+
+resuid=n		The user ID which may use the reserved blocks.
+
+sb=n			Use alternate superblock at this location.
+
+data=journal		All data are commited into the journal prior 
+			to being written into the main file system.
+		
+data=ordered	(*)	All data are forced directly out to the main file 
+			system prior to its metadata being commited to 
+			the journal.
+		
+data=writeback  	Data ordering is not preserved, data may be 
+			written into the main file system after its
+			metadata has been committed to the journal.
+
+quota			Quota options are currently silently ignored.
+noquota			(see fs/ext3/super.c, line 594)
+grpquota
+usrquota
+
+
+Specification
+=============
+ext3 shares all disk implementation with ext2 filesystem, and add
+transactions capabilities to ext2.  Journaling is done by the
+Journaling block device layer.
+
+Journaling Block Device layer
+-----------------------------
+The Journaling Block Device layer (JBD) isn't ext3 specific.  It was
+design to add journaling capabilities on a block device.  The ext3
+filesystem code will inform the JBD of modifications it is performing
+(Call a transaction).  the journal support the transactions start and
+stop, and in case of crash, the journal can replayed the transactions
+to put the partition on a consistant state fastly.
+
+handles represent a single atomic update to a filesystem.  JBD can
+handle external journal on a block device.
+
+Data Mode
+---------
+There's 3 different data modes:
+
+* writeback mode
+In data=writeback mode, ext3 does not journal data at all.  This mode
+provides a similar level of journaling as XFS, JFS, and ReiserFS in its
+default mode - metadata journaling.  A crash+recovery can cause
+incorrect data to appear in files which were written shortly before the
+crash.  This mode will typically provide the best ext3 performance.
+
+* ordered mode
+In data=ordered mode, ext3 only officially journals metadata, but it
+logically groups metadata and data blocks into a single unit called a
+transaction.  When it's time to write the new metadata out to disk, the
+associated data blocks are written first.  In general, this mode
+perform slightly slower than writeback but significantly faster than
+journal mode.
+
+* journal mode
+data=journal mode provides full data and metadata journaling.  All new
+data is written to the journal first, and then to its final location. 
+In the event of a crash, the journal can be replayed, bringing both
+data and metadata into a consistent state.  This mode is the slowest
+except when data needs to be read from and written to disk at the same
+time where it outperform all others mode.
+
+Compatibility
+-------------
+
+Ext2 partitions can be easily convert to ext3, with `tune2fs -j <dev>`.
+ Ext3 is fully compatible with Ext2.  Ext3 partitions can easily be
+mounted as Ext2.
+
+External Tools
+==============
+see manual pages to know more.
+
+tune2fs: 	create a ext3 journal on a ext2 partition with the -j flags
+mke2fs: 	create a ext3 partition with the -j flags
+debugfs: 	ext2 and ext3 file system debugger
+
+References
+==========
+
+kernel source:	file:/usr/src/linux/fs/ext3
+		file:/usr/src/linux/fs/jbd
+
+programs: 	http://e2fsprogs.sourceforge.net
+
+useful link:
+		http://www.zip.com.au/~akpm/linux/ext3/ext3-usage.html
+		http://www-106.ibm.com/developerworks/linux/library/l-fs7/
+		http://www-106.ibm.com/developerworks/linux/library/l-fs8/

MAINTAINERS

@@ -1433,6 +1433,12 @@ M:	Kai.Makisara@metla.fi
 L:	linux-scsi@vger.kernel.org
 S:	Maintained
 
+SCx200 CPU SUPPORT
+P:	Christer Weinigel
+M:	christer@weinigel.se
+W:	http://www.weinigel.se
+S:	Supported
+
 SGI VISUAL WORKSTATION 320 AND 540
 P:	Bent Hagemark
 M:	bh@sgi.com

arch/i386/Config.help

@@ -1058,3 +1058,12 @@ CONFIG_SOFTWARE_SUSPEND
   absence of features.
 
   For more information take a look at Documentation/swsusp.txt.
+
+CONFIG_SCx200
+  This provides basic support for the National Semiconductor SCx200 
+  processor.  Right now this is just a driver for the GPIO pins.
+
+  If you don't know what to do here, say N.
+
+  This support is also available as a module.  If compiled as a
+  module, it will be called scx200.o.

arch/i386/config.in

@@ -293,6 +293,8 @@ else
    fi
 fi
 
+tristate 'NatSemi SCx200 support' CONFIG_SCx200
+
 source drivers/pci/Config.in
 
 bool 'EISA support' CONFIG_EISA

arch/i386/kernel/Makefile

@@ -29,4 +29,7 @@ obj-$(CONFIG_X86_NUMAQ)		+= numaq.o
 
 EXTRA_AFLAGS   := -traditional
 
+export-objs += scx200.o
+obj-$(CONFIG_SCx200)		+= scx200.o
+
 include $(TOPDIR)/Rules.make

arch/i386/kernel/scx200.c

+/* linux/arch/i386/kernel/scx200.c 
+
+   Copyright (c) 2001,2002 Christer Weinigel <wingel@nano-system.com>
+
+   National Semiconductor SCx200 support. */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/pci.h>
+
+#include <linux/scx200.h>
+#include <linux/scx200.h>
+
+#define NAME "scx200"
+
+MODULE_AUTHOR("Christer Weinigel <wingel@nano-system.com>");
+MODULE_DESCRIPTION("NatSemi SCx200 Driver");
+MODULE_LICENSE("GPL");
+
+unsigned scx200_gpio_base = 0;
+long scx200_gpio_shadow[2];
+
+spinlock_t scx200_gpio_lock = SPIN_LOCK_UNLOCKED;
+static spinlock_t scx200_gpio_config_lock = SPIN_LOCK_UNLOCKED;
+
+u32 scx200_gpio_configure(int index, u32 mask, u32 bits)
+{
+	u32 config, new_config;
+	unsigned long flags;
+
+	spin_lock_irqsave(&scx200_gpio_config_lock, flags);
+
+	outl(index, scx200_gpio_base + 0x20);
+	config = inl(scx200_gpio_base + 0x24);
+
+	new_config = (config & mask) | bits;
+	outl(new_config, scx200_gpio_base + 0x24);
+
+	spin_unlock_irqrestore(&scx200_gpio_config_lock, flags);
+
+	return config;
+}
+
+void scx200_gpio_dump(unsigned index)
+{
+	u32 config = scx200_gpio_configure(index, ~0, 0);
+	printk(KERN_DEBUG "GPIO%02u: 0x%08lx", index, (unsigned long)config);
+	
+	if (config & 1) 
+		printk(" OE"); /* output enabled */
+	else
+		printk(" TS"); /* tristate */
+	if (config & 2) 
+		printk(" PP"); /* push pull */
+	else
+		printk(" OD"); /* open drain */
+	if (config & 4) 
+		printk(" PUE"); /* pull up enabled */
+	else
+		printk(" PUD"); /* pull up disabled */
+	if (config & 8) 
+		printk(" LOCKED"); /* locked */
+	if (config & 16) 
+		printk(" LEVEL"); /* level input */
+	else
+		printk(" EDGE"); /* edge input */
+	if (config & 32) 
+		printk(" HI"); /* trigger on rising edge */
+	else
+		printk(" LO"); /* trigger on falling edge */
+	if (config & 64) 
+		printk(" DEBOUNCE"); /* debounce */
+	printk("\n");
+}
+
+int __init scx200_init(void)
+{
+	struct pci_dev *bridge;
+	int bank;
+	unsigned base;
+
+	printk(KERN_INFO NAME ": NatSemi SCx200 Driver\n");
+
+	if ((bridge = pci_find_device(PCI_VENDOR_ID_NS, 
+				      PCI_DEVICE_ID_NS_SCx200_BRIDGE,
+				      NULL)) == NULL)
+		return -ENODEV;
+
+	base = pci_resource_start(bridge, 0);
+	printk(KERN_INFO NAME ": GPIO base 0x%x\n", base);
+
+	if (request_region(base, SCx200_GPIO_SIZE, "NatSemi SCx200 GPIO") == 0) {
+		printk(KERN_ERR NAME ": can't allocate I/O for GPIOs\n");
+		return -EBUSY;
+	}
+
+	scx200_gpio_base = base;
+
+	/* read the current values driven on the GPIO signals */
+	for (bank = 0; bank < 2; ++bank)
+		scx200_gpio_shadow[bank] = inl(scx200_gpio_base + 0x10 * bank);
+
+	return 0;
+}
+
+void __exit scx200_cleanup(void)
+{
+	release_region(scx200_gpio_base, SCx200_GPIO_SIZE);
+}
+
+module_init(scx200_init);
+module_exit(scx200_cleanup);
+
+EXPORT_SYMBOL(scx200_gpio_base);
+EXPORT_SYMBOL(scx200_gpio_shadow);
+EXPORT_SYMBOL(scx200_gpio_lock);
+EXPORT_SYMBOL(scx200_gpio_configure);
+EXPORT_SYMBOL(scx200_gpio_dump);
+
+/*
+    Local variables:
+        compile-command: "make -k -C ../../.. SUBDIRS=arch/i386/kernel modules"
+        c-basic-offset: 8
+    End:
+*/

arch/i386/mm/hugetlbpage.c

@@ -44,24 +44,22 @@ find_key_inode(int key)
 static struct page *
 alloc_hugetlb_page(void)
 {
-	struct list_head *curr, *head;
+	int i;
 	struct page *page;
 
 	spin_lock(&htlbpage_lock);
-
-	head = &htlbpage_freelist;
-	curr = head->next;
-
-	if (curr == head) {
+	if (list_empty(&htlbpage_freelist)) {
 		spin_unlock(&htlbpage_lock);
 		return NULL;
 	}
-	page = list_entry(curr, struct page, list);
-	list_del(curr);
+
+	page = list_entry(htlbpage_freelist.next, struct page, list);
+	list_del(&page->list);
 	htlbpagemem--;
 	spin_unlock(&htlbpage_lock);
 	set_page_count(page, 1);
-	memset(page_address(page), 0, HPAGE_SIZE);
+	for (i = 0; i < (HPAGE_SIZE/PAGE_SIZE); ++i)
+		clear_highpage(&page[i]);
 	return page;
 }
 

drivers/block/ll_rw_blk.c

@@ -1876,7 +1876,7 @@ void generic_make_request(struct bio *bio)
  */
 int submit_bio(int rw, struct bio *bio)
 {
-	int count = bio_sectors(bio);
+	int count = bio_sectors(bio) >> 1;
 
 	BUG_ON(!bio->bi_end_io);
 	BIO_BUG_ON(!bio->bi_size);

drivers/char/Config.help

@@ -818,6 +818,12 @@ CONFIG_MIXCOMWD
   module, say M here and read <file:Documentation/modules.txt>.  Most
   people will say N.
 
+CONFIG_SCx200_WDT
+  Enable the built-in watchdog timer support on the National 
+  Semiconductor SCx200 processors.
+
+  If compiled as a module, it will be called scx200_watchdog.o.
+
 CONFIG_MACHZ_WDT
   If you are using a ZF Micro MachZ processor, say Y here, otherwise
   N.  This is the driver for the watchdog timer builtin on that
@@ -1021,3 +1027,9 @@ CONFIG_RAW_DRIVER
   Once bound, I/O against /dev/raw/rawN uses efficient zero-copy I/O. 
   See the raw(8) manpage for more details.
 
+CONFIG_SCx200_GPIO
+  Give userspace access to the GPIO pins on the National
+  Semiconductor SCx200 processors.
+
+  If compiled as a module, it will be called scx200_gpio.o.
+

drivers/char/Config.in

@@ -131,6 +131,7 @@ if [ "$CONFIG_WATCHDOG" != "n" ]; then
    tristate '  IB700 SBC Watchdog Timer' CONFIG_IB700_WDT
    tristate '  Intel i810 TCO timer / Watchdog' CONFIG_I810_TCO
    tristate '  Mixcom Watchdog' CONFIG_MIXCOMWD 
+   tristate '  NatSemi SCx200 Watchdog' CONFIG_SCx200_WDT
    tristate '  SBC-60XX Watchdog Timer' CONFIG_60XX_WDT
    tristate '  W83877F (EMACS) Watchdog Timer' CONFIG_W83877F_WDT
    tristate '  ZF MachZ Watchdog' CONFIG_MACHZ_WDT
@@ -193,6 +194,8 @@ if [ "$CONFIG_X86" = "y" ]; then
    tristate 'ACP Modem (Mwave) support' CONFIG_MWAVE
 fi
 
+dep_tristate 'NatSemi SCx200 GPIO Support' CONFIG_SCx200_GPIO $CONFIG_SCx200
+
 tristate '  RAW driver (/dev/raw/rawN)' CONFIG_RAW_DRIVER
 
 endmenu

drivers/char/Makefile

@@ -75,6 +75,7 @@ obj-$(CONFIG_PPDEV) += ppdev.o
 obj-$(CONFIG_DZ) += dz.o
 obj-$(CONFIG_NWBUTTON) += nwbutton.o
 obj-$(CONFIG_NWFLASH) += nwflash.o
+obj-$(CONFIG_SCx200_GPIO) += scx200_gpio.o
 
 # Only one watchdog can succeed. We probe the hardware watchdog
 # drivers first, then the softdog driver.  This means if your hardware
@@ -86,6 +87,7 @@ obj-$(CONFIG_ACQUIRE_WDT) += acquirewdt.o
 obj-$(CONFIG_ADVANTECH_WDT) += advantechwdt.o
 obj-$(CONFIG_IB700_WDT) += ib700wdt.o
 obj-$(CONFIG_MIXCOMWD) += mixcomwd.o
+obj-$(CONFIG_SCx200_WDT) += scx200_wdt.o
 obj-$(CONFIG_60XX_WDT) += sbc60xxwdt.o
 obj-$(CONFIG_WDT) += wdt.o
 obj-$(CONFIG_WDTPCI) += wdt_pci.o

drivers/char/scx200_gpio.c

+/* linux/drivers/char/scx200_gpio.c 
+
+   National Semiconductor SCx200 GPIO driver.  Allows a user space
+   process to play with the GPIO pins.
+
+   Copyright (c) 2001,2002 Christer Weinigel <wingel@nano-system.com> */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+
+#include <linux/scx200_gpio.h>
+
+#define NAME "scx200_gpio"
+
+MODULE_AUTHOR("Christer Weinigel <wingel@nano-system.com>");
+MODULE_DESCRIPTION("NatSemi SCx200 GPIO Pin Driver");
+MODULE_LICENSE("GPL");
+
+static int major = 0;		/* default to dynamic major */
+MODULE_PARM(major, "i");
+MODULE_PARM_DESC(major, "Major device number");
+
+static ssize_t scx200_gpio_write(struct file *file, const char *data, 
+				 size_t len, loff_t *ppos)
+{
+	unsigned m = minor(file->f_dentry->d_inode->i_rdev);
+	size_t i;
+
+	if (ppos != &file->f_pos)
+		return -ESPIPE;
+
+	for (i = 0; i < len; ++i) {
+		char c;
+		if (get_user(c, data+i))
+			return -EFAULT;
+		switch (c)
+		{
+		case '0': 
+			scx200_gpio_set(m, 0); 
+			break;
+		case '1': 
+			scx200_gpio_set(m, 1); 
+			break;
+		case 'O':
+			printk(KERN_INFO NAME ": GPIO%d output enabled\n", m);
+			scx200_gpio_configure(m, ~1, 1);
+			break;
+		case 'o':
+			printk(KERN_INFO NAME ": GPIO%d output disabled\n", m);
+			scx200_gpio_configure(m, ~1, 0);
+			break;
+		case 'T':
+			printk(KERN_INFO NAME ": GPIO%d output is push pull\n", m);
+			scx200_gpio_configure(m, ~2, 2);
+			break;
+		case 't':
+			printk(KERN_INFO NAME ": GPIO%d output is open drain\n", m);
+			scx200_gpio_configure(m, ~2, 0);
+			break;
+		case 'P':
+			printk(KERN_INFO NAME ": GPIO%d pull up enabled\n", m);
+			scx200_gpio_configure(m, ~4, 4);
+			break;
+		case 'p':
+			printk(KERN_INFO NAME ": GPIO%d pull up disabled\n", m);
+			scx200_gpio_configure(m, ~4, 0);
+			break;
+		}
+	}
+
+	return len;
+}
+
+static ssize_t scx200_gpio_read(struct file *file, char *buf,
+				size_t len, loff_t *ppos)
+{
+	unsigned m = minor(file->f_dentry->d_inode->i_rdev);
+	int value;
+
+	if (ppos != &file->f_pos)
+		return -ESPIPE;
+
+	value = scx200_gpio_get(m);
+	if (put_user(value ? '1' : '0', buf))
+		return -EFAULT;
+	
+	return 1;
+}
+
+static int scx200_gpio_open(struct inode *inode, struct file *file)
+{
+	unsigned m = minor(inode->i_rdev);
+	if (m > 63)
+		return -EINVAL;
+	return 0;
+}
+
+static int scx200_gpio_release(struct inode *inode, struct file *file)
+{
+	return 0;
+}
+
+
+static struct file_operations scx200_gpio_fops = {
+	.owner   = THIS_MODULE,
+	.write   = scx200_gpio_write,
+	.read    = scx200_gpio_read,
+	.open    = scx200_gpio_open,
+	.release = scx200_gpio_release,
+};
+
+static int __init scx200_gpio_init(void)
+{
+	int r;
+
+	printk(KERN_DEBUG NAME ": NatSemi SCx200 GPIO Driver\n");
+
+	if (!scx200_gpio_present()) {
+		printk(KERN_ERR NAME ": no SCx200 gpio pins available\n");
+		return -ENODEV;
+	}
+
+	r = register_chrdev(major, NAME, &scx200_gpio_fops);
+	if (r < 0) {
+		printk(KERN_ERR NAME ": unable to register character device\n");
+		return r;
+	}
+	if (!major) {
+		major = r;
+		printk(KERN_DEBUG NAME ": got dynamic major %d\n", major);
+	}
+
+	return 0;
+}
+
+static void __exit scx200_gpio_cleanup(void)
+{
+	unregister_chrdev(major, NAME);
+}
+
+module_init(scx200_gpio_init);
+module_exit(scx200_gpio_cleanup);
+
+/*
+    Local variables:
+        compile-command: "make -k -C ../.. SUBDIRS=drivers/char modules"
+        c-basic-offset: 8
+    End:
+*/

drivers/char/scx200_wdt.c

+/* linux/drivers/char/scx200_wdt.c 
+
+   National Semiconductor SCx200 Watchdog support
+
+   Copyright (c) 2001,2002 Christer Weinigel <wingel@nano-system.com>
+
+   Som code taken from:
+   National Semiconductor PC87307/PC97307 (ala SC1200) WDT driver
+   (c) Copyright 2002 Zwane Mwaikambo <zwane@commfireservices.com>
+
+   This program is free software; you can redistribute it and/or
+   modify it under the terms of the GNU General Public License as
+   published by the Free Software Foundation; either version 2 of the
+   License, or (at your option) any later version.
+
+   The author(s) of this software shall not be held liable for damages
+   of any nature resulting due to the use of this software. This
+   software is provided AS-IS with no warranties. */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/miscdevice.h>
+#include <linux/watchdog.h>
+#include <linux/notifier.h>
+#include <linux/reboot.h>
+#include <linux/pci.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+
+#include <linux/scx200.h>
+
+#define NAME "scx200_wdt"
+
+MODULE_AUTHOR("Christer Weinigel <wingel@nano-system.com>");
+MODULE_DESCRIPTION("NatSemi SCx200 Watchdog Driver");
+MODULE_LICENSE("GPL");
+
+#ifndef CONFIG_WATCHDOG_NOWAYOUT
+#define CONFIG_WATCHDOG_NOWAYOUT 0
+#endif
+
+static int margin = 60;		/* in seconds */
+MODULE_PARM(margin, "i");
+MODULE_PARM_DESC(margin, "Watchdog margin in seconds");
+
+static int nowayout = CONFIG_WATCHDOG_NOWAYOUT;
+MODULE_PARM(nowayout, "i");
+MODULE_PARM_DESC(nowayout, "Disable watchdog shutdown on close");
+
+static u16 wdto_restart;
+static struct semaphore open_semaphore;
+static unsigned expect_close;
+
+/* Bits of the WDCNFG register */
+#define W_ENABLE 0x00fa		/* Enable watchdog */
+#define W_DISABLE 0x0000	/* Disable watchdog */
+
+/* The scaling factor for the timer, this depends on the value of W_ENABLE */
+#define W_SCALE (32768/1024)
+
+static void scx200_wdt_ping(void)
+{
+	outw(wdto_restart, SCx200_CB_BASE + SCx200_WDT_WDTO);
+}
+
+static void scx200_wdt_update_margin(void)
+{
+	printk(KERN_INFO NAME ": timer margin %d seconds\n", margin);
+	wdto_restart = margin * W_SCALE;
+}
+
+static void scx200_wdt_enable(void)
+{
+	printk(KERN_DEBUG NAME ": enabling watchdog timer, wdto_restart = %d\n", 
+	       wdto_restart);
+
+	outw(0, SCx200_CB_BASE + SCx200_WDT_WDTO);
+	outb(SCx200_WDT_WDSTS_WDOVF, SCx200_CB_BASE + SCx200_WDT_WDSTS);
+	outw(W_ENABLE, SCx200_CB_BASE + SCx200_WDT_WDCNFG);
+
+	scx200_wdt_ping();
+}
+
+static void scx200_wdt_disable(void)
+{
+	printk(KERN_DEBUG NAME ": disabling watchdog timer\n");
+		
+	outw(0, SCx200_CB_BASE + SCx200_WDT_WDTO);
+	outb(SCx200_WDT_WDSTS_WDOVF, SCx200_CB_BASE + SCx200_WDT_WDSTS);
+	outw(W_DISABLE, SCx200_CB_BASE + SCx200_WDT_WDCNFG);
+}
+
+static int scx200_wdt_open(struct inode *inode, struct file *file)
+{
+        /* only allow one at a time */
+        if (down_trylock(&open_semaphore))
+                return -EBUSY;
+	scx200_wdt_enable();
+	expect_close = 0;
+
+	return 0;
+}
+
+static int scx200_wdt_release(struct inode *inode, struct file *file)
+{
+	if (!expect_close) {
+		printk(KERN_WARNING NAME ": watchdog device closed unexpectedly, will not disable the watchdog timer\n");
+	} else if (!nowayout) {
+		scx200_wdt_disable();
+	}
+        up(&open_semaphore);
+
+	return 0;
+}
+
+static int scx200_wdt_notify_sys(struct notifier_block *this, 
+				      unsigned long code, void *unused)
+{
+        if (code == SYS_HALT || code == SYS_POWER_OFF)
+		if (!nowayout)
+			scx200_wdt_disable();
+
+        return NOTIFY_DONE;
+}
+
+static struct notifier_block scx200_wdt_notifier =
+{
+        notifier_call: scx200_wdt_notify_sys
+};
+
+static ssize_t scx200_wdt_write(struct file *file, const char *data, 
+				     size_t len, loff_t *ppos)
+{
+	if (ppos != &file->f_pos)
+		return -ESPIPE;
+
+	/* check for a magic close character */
+	if (len) 
+	{
+		size_t i;
+
+		scx200_wdt_ping();
+
+		expect_close = 0;
+		for (i = 0; i < len; ++i) {
+			char c;
+			if (get_user(c, data+i))
+				return -EFAULT;
+			if (c == 'V')
+				expect_close = 1;
+		}
+
+		return len;
+	}
+
+	return 0;
+}
+
+static int scx200_wdt_ioctl(struct inode *inode, struct file *file,
+	unsigned int cmd, unsigned long arg)
+{
+	static struct watchdog_info ident = {
+		.identity = "NatSemi SCx200 Watchdog",
+		.firmware_version = 1, 
+		.options = (WDIOF_SETTIMEOUT | WDIOF_KEEPALIVEPING),
+	};
+	int new_margin;
+	
+	switch (cmd) {
+	default:
+		return -ENOTTY;
+	case WDIOC_GETSUPPORT:
+		if(copy_to_user((struct watchdog_info *)arg, &ident, 
+				sizeof(ident)))
+			return -EFAULT;
+		return 0;
+	case WDIOC_GETSTATUS:
+	case WDIOC_GETBOOTSTATUS:
+		if (put_user(0, (int *)arg))
+			return -EFAULT;
+		return 0;
+	case WDIOC_KEEPALIVE:
+		scx200_wdt_ping();
+		return 0;
+	case WDIOC_SETTIMEOUT:
+		if (get_user(new_margin, (int *)arg))
+			return -EFAULT;
+		if (new_margin < 1)
+			return -EINVAL;
+		margin = new_margin;
+		scx200_wdt_update_margin();
+		scx200_wdt_ping();
+	case WDIOC_GETTIMEOUT:
+		if (put_user(margin, (int *)arg))
+			return -EFAULT;
+		return 0;
+	}
+}
+
+static struct file_operations scx200_wdt_fops = {
+	.owner	 = THIS_MODULE,
+	.write   = scx200_wdt_write,
+	.ioctl   = scx200_wdt_ioctl,
+	.open    = scx200_wdt_open,
+	.release = scx200_wdt_release,
+};
+
+static struct miscdevice scx200_wdt_miscdev = {
+	.minor = WATCHDOG_MINOR,
+	.name  = NAME,
+	.fops  = &scx200_wdt_fops,
+};
+
+static int __init scx200_wdt_init(void)
+{
+	int r;
+
+	printk(KERN_DEBUG NAME ": NatSemi SCx200 Watchdog Driver\n");
+
+	/* First check that this really is a NatSemi SCx200 CPU */
+	if ((pci_find_device(PCI_VENDOR_ID_NS, 
+			     PCI_DEVICE_ID_NS_SCx200_BRIDGE,
+			     NULL)) == NULL)
+		return -ENODEV;
+
+	/* More sanity checks, verify that the configuration block is there */
+	if (!scx200_cb_probe(SCx200_CB_BASE)) {
+		printk(KERN_WARNING NAME ": no configuration block found\n");
+		return -ENODEV;
+	}
+
+	if (!request_region(SCx200_CB_BASE + SCx200_WDT_OFFSET, 
+			    SCx200_WDT_SIZE, 
+			    "NatSemi SCx200 Watchdog")) {
+		printk(KERN_WARNING NAME ": watchdog I/O region busy\n");
+		return -EBUSY;
+	}
+
+	scx200_wdt_update_margin();
+	scx200_wdt_disable();
+
+	sema_init(&open_semaphore, 1);
+
+	r = misc_register(&scx200_wdt_miscdev);
+	if (r)
+		return r;
+
+	r = register_reboot_notifier(&scx200_wdt_notifier);
+        if (r) {
+                printk(KERN_ERR NAME ": unable to register reboot notifier");
+		misc_deregister(&scx200_wdt_miscdev);
+                return r;
+        }
+
+	return 0;
+}
+
+static void __exit scx200_wdt_cleanup(void)
+{
+        unregister_reboot_notifier(&scx200_wdt_notifier);
+	misc_deregister(&scx200_wdt_miscdev);
+	release_region(SCx200_CB_BASE + SCx200_WDT_OFFSET,
+		       SCx200_WDT_SIZE);
+}
+
+module_init(scx200_wdt_init);
+module_exit(scx200_wdt_cleanup);
+
+/*
+    Local variables:
+        compile-command: "make -k -C ../.. SUBDIRS=drivers/char modules"
+        c-basic-offset: 8
+    End:
+*/

drivers/i2c/Config.help

@@ -91,6 +91,28 @@ CONFIG_I2C_CHARDEV
   <file:Documentation/modules.txt>.
   The module will be called i2c-dev.o.
 
+CONFIG_SCx200_I2C
+  Enable the use of two GPIO pins of a SCx200 processor as an I2C bus.
+
+  If you don't know what to do here, say N.
+
+  If compiled as a module, it will be called scx200_i2c.o.
+
+CONFIG_SCx200_I2C_SCL
+  Enter the GPIO pin number used for the SCL signal.  This value can
+  also be specified with a module parameter.
+
+CONFIG_SCx200_I2C_SDA
+  Enter the GPIO pin number used for the SSA signal.  This value can
+  also be specified with a module parameter.
+
+CONFIG_SCx200_ACB
+  Enable the use of the ACCESS.bus controllers of a SCx200 processor.
+
+  If you don't know what to do here, say N.
+
+  If compiled as a module, it will be called scx200_acb.o.
+
 CONFIG_I2C_PROC
   This provides support for i2c device entries in the /proc filesystem.
   The entries will be found in /proc/sys/dev/sensors.

drivers/i2c/Config.in

@@ -13,6 +13,12 @@ if [ "$CONFIG_I2C" != "n" ]; then
       dep_tristate '  Philips style parallel port adapter' CONFIG_I2C_PHILIPSPAR $CONFIG_I2C_ALGOBIT $CONFIG_PARPORT
       dep_tristate '  ELV adapter' CONFIG_I2C_ELV $CONFIG_I2C_ALGOBIT
       dep_tristate '  Velleman K9000 adapter' CONFIG_I2C_VELLEMAN $CONFIG_I2C_ALGOBIT
+      dep_tristate '  NatSemi SCx200 I2C using GPIO pins' CONFIG_SCx200_I2C $CONFIG_SCx200 $CONFIG_I2C_ALGOBIT
+      if [ "$CONFIG_SCx200_I2C" != "n" ]; then
+         int  '    GPIO pin used for SCL' CONFIG_SCx200_I2C_SCL 12
+         int  '    GPIO pin used for SDA' CONFIG_SCx200_I2C_SDA 13
+      fi
+      dep_tristate '  NatSemi SCx200 ACCESS.bus' CONFIG_SCx200_ACB $CONFIG_I2C
    fi
 
    dep_tristate 'I2C PCF 8584 interfaces' CONFIG_I2C_ALGOPCF $CONFIG_I2C

drivers/i2c/Makefile

@@ -15,6 +15,8 @@ obj-$(CONFIG_I2C_ALGOPCF)	+= i2c-algo-pcf.o
 obj-$(CONFIG_I2C_ELEKTOR)	+= i2c-elektor.o
 obj-$(CONFIG_ITE_I2C_ALGO)	+= i2c-algo-ite.o
 obj-$(CONFIG_ITE_I2C_ADAP)	+= i2c-adap-ite.o
+obj-$(CONFIG_SCx200_I2C)	+= scx200_i2c.o
+obj-$(CONFIG_SCx200_ACB)	+= scx200_acb.o
 obj-$(CONFIG_I2C_PROC)		+= i2c-proc.o
 
 # This is needed for automatic patch generation: sensors code starts here

drivers/i2c/scx200_acb.c

+/*  linux/drivers/i2c/scx200_acb.c 
+
+    Copyright (c) 2001,2002 Christer Weinigel <wingel@nano-system.com>
+
+    National Semiconductor SCx200 ACCESS.bus support
+    
+    Based on i2c-keywest.c which is:
+        Copyright (c) 2001 Benjamin Herrenschmidt <benh@kernel.crashing.org>
+        Copyright (c) 2000 Philip Edelbrock <phil@stimpy.netroedge.com>
+    
+    This program is free software; you can redistribute it and/or
+    modify it under the terms of the GNU General Public License as
+    published by the Free Software Foundation; either version 2 of the
+    License, or (at your option) any later version.
+   
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+    General Public License for more details.
+   
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+*/
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/i2c.h>
+#include <linux/smp_lock.h>
+#include <linux/pci.h>
+#include <asm/io.h>
+
+#include <linux/scx200.h>
+
+#define NAME "scx200_acb"
+
+MODULE_AUTHOR("Christer Weinigel <wingel@nano-system.com>");
+MODULE_DESCRIPTION("NatSemi SCx200 ACCESS.bus Driver");
+MODULE_LICENSE("GPL");
+
+#define MAX_DEVICES 4
+static int base[MAX_DEVICES] = { 0x840 };
+MODULE_PARM(base, "1-4i");
+MODULE_PARM_DESC(base, "Base addresses for the ACCESS.bus controllers");
+
+#define DEBUG 0
+
+#if DEBUG
+#define DBG(x...) printk(KERN_DEBUG NAME ": " x)
+#else
+#define DBG(x...)
+#endif
+
+/* The hardware supports interrupt driven mode too, but I haven't
+   implemented that. */
+#define POLLED_MODE 1
+#define POLL_TIMEOUT (HZ)
+
+enum scx200_acb_state {
+	state_idle,
+	state_address,
+	state_command,
+	state_repeat_start,
+	state_quick,
+	state_read,
+	state_write,
+};
+
+static const char *scx200_acb_state_name[] = {
+	"idle",
+	"address",
+	"command",
+	"repeat_start",
+	"quick",
+	"read",
+	"write",
+};
+
+/* Physical interface */
+struct scx200_acb_iface
+{
+	struct scx200_acb_iface *next;
+	struct i2c_adapter adapter;
+	unsigned base;
+	struct semaphore sem;
+
+	/* State machine data */
+	enum scx200_acb_state state;
+	int result;
+	u8 address_byte;
+	u8 command;
+	u8 *ptr;
+	char needs_reset;
+	unsigned len;
+};
+
+/* Register Definitions */
+#define ACBSDA		(iface->base + 0)
+#define ACBST		(iface->base + 1)
+#define    ACBST_SDAST		0x40 /* SDA Status */
+#define    ACBST_BER		0x20 
+#define    ACBST_NEGACK		0x10 /* Negative Acknowledge */
+#define    ACBST_STASTR		0x08 /* Stall After Start */
+#define    ACBST_MASTER		0x02
+#define ACBCST		(iface->base + 2)
+#define    ACBCST_BB		0x02
+#define ACBCTL1		(iface->base + 3)
+#define    ACBCTL1_STASTRE	0x80
+#define    ACBCTL1_NMINTE	0x40
+#define	   ACBCTL1_ACK		0x10
+#define	   ACBCTL1_STOP		0x02
+#define	   ACBCTL1_START	0x01
+#define ACBADDR		(iface->base + 4)
+#define ACBCTL2		(iface->base + 5)
+#define    ACBCTL2_ENABLE	0x01
+
+/************************************************************************/
+
+static void scx200_acb_machine(struct scx200_acb_iface *iface, u8 status)
+{
+	const char *errmsg;
+
+	DBG("state %s, status = 0x%02x\n", 
+	    scx200_acb_state_name[iface->state], status);
+
+	if (status & ACBST_BER) {
+		errmsg = "bus error";
+		goto error;
+	}
+	if (!(status & ACBST_MASTER)) {
+		errmsg = "not master";
+		goto error;
+	}
+	if (status & ACBST_NEGACK)
+		goto negack;
+
+	switch (iface->state) {
+	case state_idle:
+		printk(KERN_WARNING NAME ": %s, interrupt in idle state\n", 
+		       iface->adapter.name);
+		break;
+
+	case state_address:
+		/* Do a pointer write first */
+		outb(iface->address_byte & ~1, ACBSDA);
+
+		iface->state = state_command;
+		break;
+
+	case state_command:
+		outb(iface->command, ACBSDA);
+
+		if (iface->address_byte & 1)
+			iface->state = state_repeat_start;
+		else
+			iface->state = state_write;
+		break;
+
+	case state_repeat_start:
+		outb(inb(ACBCTL1) | ACBCTL1_START, ACBCTL1);
+		/* fallthrough */
+		
+	case state_quick:
+		if (iface->address_byte & 1) {
+			if (iface->len == 1) 
+				outb(inb(ACBCTL1) | ACBCTL1_ACK, ACBCTL1);
+			else
+				outb(inb(ACBCTL1) & ~ACBCTL1_ACK, ACBCTL1);
+			outb(iface->address_byte, ACBSDA);
+
+			iface->state = state_read;
+		} else {
+			outb(iface->address_byte, ACBSDA);
+
+			iface->state = state_write;
+		}
+		break;
+
+	case state_read:
+		/* Set ACK if receiving the last byte */
+		if (iface->len == 1)
+			outb(inb(ACBCTL1) | ACBCTL1_ACK, ACBCTL1);
+		else
+			outb(inb(ACBCTL1) & ~ACBCTL1_ACK, ACBCTL1);
+
+		*iface->ptr++ = inb(ACBSDA);
+		--iface->len;
+
+		if (iface->len == 0) {
+			iface->result = 0;
+			iface->state = state_idle;
+			outb(inb(ACBCTL1) | ACBCTL1_STOP, ACBCTL1);
+		}
+
+		break;
+
+	case state_write:
+		if (iface->len == 0) {
+			iface->result = 0;
+			iface->state = state_idle;
+			outb(inb(ACBCTL1) | ACBCTL1_STOP, ACBCTL1);
+			break;
+		}
+		
+		outb(*iface->ptr++, ACBSDA);
+		--iface->len;
+		
+		break;
+	}
+
+	return;
+
+ negack:
+	DBG("negative acknowledge in state %s\n", 
+	    scx200_acb_state_name[iface->state]);
+
+	iface->state = state_idle;
+	iface->result = -ENXIO;
+
+	outb(inb(ACBCTL1) | ACBCTL1_STOP, ACBCTL1);
+	outb(ACBST_STASTR | ACBST_NEGACK, ACBST);
+	return;
+
+ error:
+	printk(KERN_ERR NAME ": %s, %s in state %s\n", iface->adapter.name, 
+	       errmsg, scx200_acb_state_name[iface->state]);
+
+	iface->state = state_idle;
+	iface->result = -EIO;
+	iface->needs_reset = 1;
+}
+
+static void scx200_acb_timeout(struct scx200_acb_iface *iface) 
+{
+	printk(KERN_ERR NAME ": %s, timeout in state %s\n", 
+	       iface->adapter.name, scx200_acb_state_name[iface->state]);
+
+	iface->state = state_idle;
+	iface->result = -EIO;
+	iface->needs_reset = 1;
+}
+
+#ifdef POLLED_MODE
+static void scx200_acb_poll(struct scx200_acb_iface *iface)
+{
+	u8 status = 0;
+	unsigned long timeout;
+
+	timeout = jiffies + POLL_TIMEOUT;
+	while (time_before(jiffies, timeout)) {
+		status = inb(ACBST);
+		if ((status & (ACBST_SDAST|ACBST_BER|ACBST_NEGACK)) != 0) {
+			scx200_acb_machine(iface, status);
+			return;
+		}
+		schedule_timeout(HZ/100+1);
+	}
+
+	scx200_acb_timeout(iface);
+}
+#endif /* POLLED_MODE */
+
+static void scx200_acb_reset(struct scx200_acb_iface *iface)
+{
+	/* Disable the ACCESS.bus device and Configure the SCL
+           frequency: 16 clock cycles */
+	outb(0x70, ACBCTL2);
+	/* Polling mode */
+	outb(0, ACBCTL1);
+	/* Disable slave address */
+	outb(0, ACBADDR);
+	/* Enable the ACCESS.bus device */
+	outb(inb(ACBCTL2) | ACBCTL2_ENABLE, ACBCTL2);
+	/* Free STALL after START */
+	outb(inb(ACBCTL1) & ~(ACBCTL1_STASTRE | ACBCTL1_NMINTE), ACBCTL1);
+	/* Send a STOP */
+	outb(inb(ACBCTL1) | ACBCTL1_STOP, ACBCTL1);
+	/* Clear BER, NEGACK and STASTR bits */
+	outb(ACBST_BER | ACBST_NEGACK | ACBST_STASTR, ACBST);
+	/* Clear BB bit */
+	outb(inb(ACBCST) | ACBCST_BB, ACBCST);
+}
+
+static s32 scx200_acb_smbus_xfer(struct i2c_adapter *adapter,
+				u16 address, unsigned short flags,	
+				char rw, u8 command, int size, 
+				union i2c_smbus_data *data)
+{
+	struct scx200_acb_iface *iface = adapter->data;
+	int len;
+	u8 *buffer;
+	u16 cur_word;
+	int rc;
+
+	switch (size) {
+	case I2C_SMBUS_QUICK:
+	    	len = 0;
+	    	buffer = NULL;
+	    	break;
+	case I2C_SMBUS_BYTE:
+		if (rw == I2C_SMBUS_READ) {
+			len = 1;
+			buffer = &data->byte;
+		} else {
+			len = 1;
+			buffer = &command;
+		}
+	    	break;
+	case I2C_SMBUS_BYTE_DATA:
+	    	len = 1;
+	    	buffer = &data->byte;
+	    	break;
+	case I2C_SMBUS_WORD_DATA:
+		len = 2;
+	    	cur_word = cpu_to_le16(data->word);
+	    	buffer = (u8 *)&cur_word;
+		break;
+	case I2C_SMBUS_BLOCK_DATA:
+	    	len = data->block[0];
+	    	buffer = &data->block[1];
+		break;
+	default:
+	    	return -EINVAL;
+	}
+
+	DBG("size=%d, address=0x%x, command=0x%x, len=%d, read=%d\n",
+	    size, address, command, len, rw == I2C_SMBUS_READ);
+
+	if (!len && rw == I2C_SMBUS_READ) {
+		printk(KERN_WARNING NAME ": %s, zero length read\n", 
+		       adapter->name);
+		return -EINVAL;
+	}
+
+	if (len && !buffer) {
+		printk(KERN_WARNING NAME ": %s, nonzero length but no buffer\n", adapter->name);
+		return -EFAULT;
+	}
+
+	down(&iface->sem);
+
+	iface->address_byte = address<<1;
+	if (rw == I2C_SMBUS_READ)
+		iface->address_byte |= 1;
+	iface->command = command;
+	iface->ptr = buffer;
+	iface->len = len;
+	iface->result = -EINVAL;
+	iface->needs_reset = 0;
+
+	outb(inb(ACBCTL1) | ACBCTL1_START, ACBCTL1);
+
+	if (size == I2C_SMBUS_QUICK || size == I2C_SMBUS_BYTE)
+		iface->state = state_quick;
+	else
+		iface->state = state_address;
+
+#ifdef POLLED_MODE
+	while (iface->state != state_idle)
+		scx200_acb_poll(iface);
+#else /* POLLED_MODE */
+#error Interrupt driven mode not implemented
+#endif /* POLLED_MODE */	
+
+	if (iface->needs_reset)
+		scx200_acb_reset(iface);
+
+	rc = iface->result;
+
+	up(&iface->sem);
+
+	if (rc == 0 && size == I2C_SMBUS_WORD_DATA && rw == I2C_SMBUS_READ)
+	    	data->word = le16_to_cpu(cur_word);
+
+#if DEBUG
+	printk(KERN_DEBUG NAME ": transfer done, result: %d", rc);
+	if (buffer) {
+		int i;
+		printk(" data:");
+		for (i = 0; i < len; ++i)
+			printk(" %02x", buffer[i]);
+	}
+	printk("\n");
+#endif
+
+	return rc;
+}
+
+static u32 scx200_acb_func(struct i2c_adapter *adapter)
+{
+	return I2C_FUNC_SMBUS_QUICK | I2C_FUNC_SMBUS_BYTE |
+	       I2C_FUNC_SMBUS_BYTE_DATA | I2C_FUNC_SMBUS_WORD_DATA |
+	       I2C_FUNC_SMBUS_BLOCK_DATA;
+}
+
+static int scx200_acb_reg(struct i2c_client *client)
+{
+	return 0;
+}
+
+static int scx200_acb_unreg(struct i2c_client *client)
+{
+	return 0;
+}
+
+static void scx200_acb_inc_use(struct i2c_adapter *adapter)
+{
+	MOD_INC_USE_COUNT;
+}
+
+static void scx200_acb_dec_use(struct i2c_adapter *adapter)
+{
+	MOD_DEC_USE_COUNT;
+}
+
+/* For now, we only handle combined mode (smbus) */
+static struct i2c_algorithm scx200_acb_algorithm = {
+	name:		"NatSemi SCx200 ACCESS.bus",
+	id:		I2C_ALGO_SMBUS,
+	smbus_xfer:	scx200_acb_smbus_xfer,
+	functionality:	scx200_acb_func,
+};
+
+struct scx200_acb_iface *scx200_acb_list;
+
+int scx200_acb_probe(struct scx200_acb_iface *iface)
+{
+	u8 val;
+
+	/* Disable the ACCESS.bus device and Configure the SCL
+           frequency: 16 clock cycles */
+	outb(0x70, ACBCTL2);
+
+	if (inb(ACBCTL2) != 0x70) {
+		DBG("ACBCTL2 readback failed\n");
+		return -ENXIO;
+	}
+
+	outb(inb(ACBCTL1) | ACBCTL1_NMINTE, ACBCTL1);
+
+	val = inb(ACBCTL1);
+	if (val) {
+		DBG("disabled, but ACBCTL1=0x%02x\n", val);
+		return -ENXIO;
+	}
+
+	outb(inb(ACBCTL2) | ACBCTL2_ENABLE, ACBCTL2);
+
+	outb(inb(ACBCTL1) | ACBCTL1_NMINTE, ACBCTL1);
+
+	val = inb(ACBCTL1);
+	if ((val & ACBCTL1_NMINTE) != ACBCTL1_NMINTE) {
+		DBG("enabled, but NMINTE won't be set, ACBCTL1=0x%02x\n", val);
+		return -ENXIO;
+	}
+
+	return 0;
+}
+
+static int  __init scx200_acb_create(int base, int index)
+{
+	struct scx200_acb_iface *iface;
+	struct i2c_adapter *adapter;
+	int rc = 0;
+	char description[64];
+
+	iface = kmalloc(sizeof(*iface), GFP_KERNEL);
+	if (!iface) {
+		printk(KERN_ERR NAME ": can't allocate memory\n");
+		rc = -ENOMEM;
+		goto errout;
+	}
+
+	memset(iface, 0, sizeof(*iface));
+	adapter = &iface->adapter;
+	adapter->data = iface;
+	sprintf(adapter->name, "SCx200 ACB%d", index);
+	adapter->id = I2C_ALGO_SMBUS;
+	adapter->algo = &scx200_acb_algorithm;
+	adapter->inc_use = scx200_acb_inc_use;
+	adapter->dec_use = scx200_acb_dec_use;
+	adapter->client_register = scx200_acb_reg;
+	adapter->client_unregister = scx200_acb_unreg;
+
+	init_MUTEX(&iface->sem);
+
+	sprintf(description, "NatSemi SCx200 ACCESS.bus [%s]", adapter->name);
+	if (request_region(base, 8, description) == 0) {
+		printk(KERN_ERR NAME ": %s, can't allocate io 0x%x-0x%x\n", 
+		       adapter->name, base, base + 8-1);
+		rc = -EBUSY;
+		goto errout;
+	}
+	iface->base = base;
+
+	rc = scx200_acb_probe(iface);
+	if (rc) {
+		printk(KERN_WARNING NAME ": %s, probe failed\n", adapter->name);
+		goto errout;
+	}
+
+	scx200_acb_reset(iface);
+
+	if (i2c_add_adapter(adapter) < 0) {
+		printk(KERN_ERR NAME ": %s, failed to register\n", adapter->name);
+		rc = -ENODEV;
+		goto errout;
+	}
+
+	lock_kernel();
+	iface->next = scx200_acb_list;
+	scx200_acb_list = iface;
+	unlock_kernel();
+
+	return 0;
+
+ errout:
+	if (iface) {
+		if (iface->base)
+			release_region(iface->base, 8);
+		kfree(iface);
+	}
+	return rc;
+}
+
+static int __init scx200_acb_init(void)
+{
+	int i;
+	int rc;
+
+	printk(KERN_DEBUG NAME ": NatSemi SCx200 ACCESS.bus Driver\n");
+
+	/* Verify that this really is a SCx200 processor */
+	if (pci_find_device(PCI_VENDOR_ID_NS,
+			    PCI_DEVICE_ID_NS_SCx200_BRIDGE,
+			    NULL) == NULL)
+		return -ENODEV;
+
+	rc = -ENXIO;
+	for (i = 0; i < MAX_DEVICES; ++i) {
+		if (base[i] > 0)
+			rc = scx200_acb_create(base[i], i);
+	}
+	if (scx200_acb_list)
+		return 0;
+	return rc;
+}
+
+static void __exit scx200_acb_cleanup(void)
+{
+	struct scx200_acb_iface *iface;
+	lock_kernel();
+	while ((iface = scx200_acb_list) != NULL) {
+		scx200_acb_list = iface->next;
+		unlock_kernel();
+
+		i2c_del_adapter(&iface->adapter);
+		release_region(iface->base, 8);
+		kfree(iface);
+		lock_kernel();
+	}
+	unlock_kernel();
+}
+
+module_init(scx200_acb_init);
+module_exit(scx200_acb_cleanup);
+
+/*
+    Local variables:
+        compile-command: "make -k -C ../.. SUBDIRS=drivers/i2c modules"
+        c-basic-offset: 8
+    End:
+*/
+

drivers/i2c/scx200_i2c.c

+/* linux/drivers/i2c/scx200_i2c.c 
+
+   Copyright (c) 2001,2002 Christer Weinigel <wingel@nano-system.com>
+
+   National Semiconductor SCx200 I2C bus on GPIO pins
+
+   Based on i2c-velleman.c Copyright (C) 1995-96, 2000 Simon G. Vogl
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2 of the License, or
+   (at your option) any later version.
+   
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+   
+   You should have received a copy of the GNU General Public License
+   along with this program; if not, write to the Free Software
+   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.		     
+*/
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/i2c.h>
+#include <linux/i2c-algo-bit.h>
+#include <asm/io.h>
+
+#include <linux/scx200_gpio.h>
+
+#define NAME "scx200_i2c"
+
+MODULE_AUTHOR("Christer Weinigel <wingel@nano-system.com>");
+MODULE_DESCRIPTION("NatSemi SCx200 I2C Driver");
+MODULE_LICENSE("GPL");
+
+MODULE_PARM(scl, "i");
+MODULE_PARM_DESC(scl, "GPIO line for SCL");
+MODULE_PARM(sda, "i");
+MODULE_PARM_DESC(sda, "GPIO line for SDA");
+
+static int scl = CONFIG_SCx200_I2C_SCL;
+static int sda = CONFIG_SCx200_I2C_SDA;
+
+static void scx200_i2c_setscl(void *data, int state)
+{
+	scx200_gpio_set(scl, state);
+}
+
+static void scx200_i2c_setsda(void *data, int state)
+{
+	scx200_gpio_set(sda, state);
+} 
+
+static int scx200_i2c_getscl(void *data)
+{
+	return scx200_gpio_get(scl);
+}
+
+static int scx200_i2c_getsda(void *data)
+{
+	return scx200_gpio_get(sda);
+}
+
+static int scx200_i2c_reg(struct i2c_client *client)
+{
+	return 0;
+}
+
+static int scx200_i2c_unreg(struct i2c_client *client)
+{
+	return 0;
+}
+
+static void scx200_i2c_inc_use(struct i2c_adapter *adap)
+{
+	MOD_INC_USE_COUNT;
+}
+
+static void scx200_i2c_dec_use(struct i2c_adapter *adap)
+{
+	MOD_DEC_USE_COUNT;
+}
+
+/* ------------------------------------------------------------------------
+ * Encapsulate the above functions in the correct operations structure.
+ * This is only done when more than one hardware adapter is supported.
+ */
+
+static struct i2c_algo_bit_data scx200_i2c_data = {
+	NULL,
+	scx200_i2c_setsda,
+	scx200_i2c_setscl,
+	scx200_i2c_getsda,
+	scx200_i2c_getscl,
+	10, 10, 100,		/* waits, timeout */
+};
+
+static struct i2c_adapter scx200_i2c_ops = {
+	.name              = "NatSemi SCx200 I2C",
+	.id		   = I2C_HW_B_VELLE,
+	.algo_data	   = &scx200_i2c_data,
+	.inc_use	   = scx200_i2c_inc_use,
+	.dec_use	   = scx200_i2c_dec_use,
+	.client_register   = scx200_i2c_reg,
+	.client_unregister = scx200_i2c_unreg,
+};
+
+int scx200_i2c_init(void)
+{
+	printk(KERN_DEBUG NAME ": NatSemi SCx200 I2C Driver\n");
+
+	if (!scx200_gpio_present()) {
+		printk(KERN_ERR NAME ": no SCx200 gpio pins available\n");
+		return -ENODEV;
+	}
+
+	printk(KERN_DEBUG NAME ": SCL=GPIO%02u, SDA=GPIO%02u\n", 
+	       scl, sda);
+
+	if (scl == -1 || sda == -1 || scl == sda) {
+		printk(KERN_ERR NAME ": scl and sda must be specified\n");
+		return -EINVAL;
+	}
+
+	/* Configure GPIOs as open collector outputs */
+	scx200_gpio_configure(scl, ~2, 5);
+	scx200_gpio_configure(sda, ~2, 5);
+
+	if (i2c_bit_add_bus(&scx200_i2c_ops) < 0) {
+		printk(KERN_ERR NAME ": adapter %s registration failed\n", 
+		       scx200_i2c_ops.name);
+		return -ENODEV;
+	}
+	
+	return 0;
+}
+
+void scx200_i2c_cleanup(void)
+{
+	i2c_bit_del_bus(&scx200_i2c_ops);
+}
+
+module_init(scx200_i2c_init);
+module_exit(scx200_i2c_cleanup);
+
+/*
+    Local variables:
+        compile-command: "make -k -C ../.. SUBDIRS=drivers/i2c modules"
+        c-basic-offset: 8
+    End:
+*/

drivers/mtd/maps/Config.help

@@ -127,6 +127,14 @@ CONFIG_MTD_MIXMEM
   you probably want to enable this mapping driver. More info is at
   <http://www.itc.hu/>.
 
+CONFIG_MTD_SCx200_DOCFLASH
+  Enable support for a flash chip mapped using the DOCCS signal on a
+  National Semiconductor SCx200 processor.
+
+  If you don't know what to do here, say N.
+
+  If compiled as a module, it will be called scx200_docflash.o.
+
 CONFIG_MTD_OCTAGON
   This provides a 'mapping' driver which supports the way in which
   the flash chips are connected in the Octagon-5066 Single Board

drivers/mtd/maps/Config.in

@@ -26,6 +26,7 @@ if [ "$CONFIG_X86" = "y" ]; then
    dep_tristate '  JEDEC Flash device mapped on Mixcom piggyback card' CONFIG_MTD_MIXMEM $CONFIG_MTD_JEDEC
    dep_tristate '  JEDEC Flash device mapped on Octagon 5066 SBC' CONFIG_MTD_OCTAGON $CONFIG_MTD_JEDEC
    dep_tristate '  JEDEC Flash device mapped on Tempustech VMAX SBC301' CONFIG_MTD_VMAX $CONFIG_MTD_JEDEC
+   dep_tristate '  Flash device mapped with DOCCS on NatSemi SCx200' CONFIG_MTD_SCx200_DOCFLASH $CONFIG_MTD_CFI
    dep_tristate '  BIOS flash chip on Intel L440GX boards' CONFIG_MTD_L440GX $CONFIG_MTD_JEDEC
 fi
 

drivers/mtd/maps/Makefile

@@ -24,6 +24,7 @@ obj-$(CONFIG_MTD_SC520CDP)	+= sc520cdp.o
 obj-$(CONFIG_MTD_NETSC520)	+= netsc520.o
 obj-$(CONFIG_MTD_SUN_UFLASH)    += sun_uflash.o
 obj-$(CONFIG_MTD_VMAX)		+= vmax301.o
+obj-$(CONFIG_MTD_SCx200_DOCFLASH)+= scx200_docflash.o
 obj-$(CONFIG_MTD_DBOX2)		+= dbox2-flash.o
 obj-$(CONFIG_MTD_OCELOT)	+= ocelot.o
 obj-$(CONFIG_MTD_SOLUTIONENGINE)+= solutionengine.o

drivers/mtd/maps/scx200_docflash.c

+/* linux/drivers/mtd/maps/scx200_docflash.c 
+
+   Copyright (c) 2001,2002 Christer Weinigel <wingel@nano-system.com>
+
+   National Semiconductor SCx200 flash mapped with DOCCS
+*/
+
+#include <linux/module.h>
+#include <linux/config.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <asm/io.h>
+#include <linux/mtd/mtd.h>
+#include <linux/mtd/map.h>
+#include <linux/mtd/partitions.h>
+
+#include <linux/pci.h>
+#include <linux/scx200.h>
+
+#define NAME "scx200_docflash"
+
+MODULE_AUTHOR("Christer Weinigel <wingel@hack.org>");
+MODULE_DESCRIPTION("NatSemi SCx200 DOCCS Flash Driver");
+MODULE_LICENSE("GPL");
+
+/* Set this to one if you want to partition the flash */
+#define PARTITION 1
+
+MODULE_PARM(probe, "i");
+MODULE_PARM_DESC(probe, "Probe for a BIOS mapping");
+MODULE_PARM(size, "i");
+MODULE_PARM_DESC(size, "Size of the flash mapping");
+MODULE_PARM(width, "i");
+MODULE_PARM_DESC(width, "Data width of the flash mapping (8/16)");
+MODULE_PARM(flashtype, "s");
+MODULE_PARM_DESC(flashtype, "Type of MTD probe to do");
+
+static int probe = 0;		/* Don't autoprobe */
+static unsigned size = 0x1000000; /* 16 MB the whole ISA address space */
+static unsigned width = 8;	/* Default to 8 bits wide */
+static char *flashtype = "cfi_probe";
+
+static struct resource docmem = {
+	.flags = IORESOURCE_MEM,
+	.name  = "NatSemi SCx200 DOCCS Flash",
+};
+
+static struct mtd_info *mymtd;
+
+#if PARTITION
+static struct mtd_partition partition_info[] = {
+	{ 
+		.name   = "DOCCS Boot kernel", 
+		.offset = 0, 
+		.size   = 0xc0000
+	},
+	{ 
+		.name   = "DOCCS Low BIOS", 
+		.offset = 0xc0000, 
+		.size   = 0x40000
+	},
+	{ 
+		.name   = "DOCCS File system", 
+		.offset = 0x100000, 
+		.size   = ~0	/* calculate from flash size */
+	},
+	{ 
+		.name   = "DOCCS High BIOS", 
+		.offset = ~0, 	/* calculate from flash size */
+		.size   = 0x80000
+	},
+};
+#define NUM_PARTITIONS (sizeof(partition_info)/sizeof(partition_info[0]))
+#endif
+
+static __u8 scx200_docflash_read8(struct map_info *map, unsigned long ofs)
+{
+	return __raw_readb(map->map_priv_1 + ofs);
+}
+
+static __u16 scx200_docflash_read16(struct map_info *map, unsigned long ofs)
+{
+	return __raw_readw(map->map_priv_1 + ofs);
+}
+
+static void scx200_docflash_copy_from(struct map_info *map, void *to, unsigned long from, ssize_t len)
+{
+	memcpy_fromio(to, map->map_priv_1 + from, len);
+}
+
+static void scx200_docflash_write8(struct map_info *map, __u8 d, unsigned long adr)
+{
+	__raw_writeb(d, map->map_priv_1 + adr);
+	mb();
+}
+
+static void scx200_docflash_write16(struct map_info *map, __u16 d, unsigned long adr)
+{
+	__raw_writew(d, map->map_priv_1 + adr);
+	mb();
+}
+
+static void scx200_docflash_copy_to(struct map_info *map, unsigned long to, const void *from, ssize_t len)
+{
+	memcpy_toio(map->map_priv_1 + to, from, len);
+}
+
+static struct map_info scx200_docflash_map = {
+	.name      = "NatSemi SCx200 DOCCS Flash",
+	.read8     = scx200_docflash_read8,
+	.read16    = scx200_docflash_read16,
+	.copy_from = scx200_docflash_copy_from,
+	.write8    = scx200_docflash_write8,
+	.write16   = scx200_docflash_write16,
+	.copy_to   = scx200_docflash_copy_to
+};
+
+int __init init_scx200_docflash(void)
+{
+	unsigned u;
+	unsigned base;
+	unsigned ctrl;
+	unsigned pmr;
+	struct pci_dev *bridge;
+
+	printk(KERN_DEBUG NAME ": NatSemi SCx200 DOCCS Flash Driver\n");
+
+	if ((bridge = pci_find_device(PCI_VENDOR_ID_NS, 
+				      PCI_DEVICE_ID_NS_SCx200_BRIDGE,
+				      NULL)) == NULL)
+		return -ENODEV;
+	
+	if (!scx200_cb_probe(SCx200_CB_BASE)) {
+		printk(KERN_WARNING NAME ": no configuration block found\n");
+		return -ENODEV;
+	}
+
+	if (probe) {
+		/* Try to use the present flash mapping if any */
+		pci_read_config_dword(bridge, SCx200_DOCCS_BASE, &base);
+		pci_read_config_dword(bridge, SCx200_DOCCS_CTRL, &ctrl);
+		pmr = inl(SCx200_CB_BASE + SCx200_PMR);
+
+		if (base == 0
+		    || (ctrl & 0x07000000) != 0x07000000
+		    || (ctrl & 0x0007ffff) == 0)
+			return -ENODEV;
+
+		size = ((ctrl&0x1fff)<<13) + (1<<13);
+
+		for (u = size; u > 1; u >>= 1)
+			;
+		if (u != 1)
+			return -ENODEV;
+
+		if (pmr & (1<<6))
+			width = 16;
+		else
+			width = 8;
+
+		docmem.start = base;
+		docmem.end = base + size;
+
+		if (request_resource(&iomem_resource, &docmem)) {
+			printk(KERN_ERR NAME ": unable to allocate memory for flash mapping\n");
+			return -ENOMEM;
+		}
+	} else {
+		for (u = size; u > 1; u >>= 1)
+			;
+		if (u != 1) {
+			printk(KERN_ERR NAME ": invalid size for flash mapping\n");
+			return -EINVAL;
+		}
+		
+		if (width != 8 && width != 16) {
+			printk(KERN_ERR NAME ": invalid bus width for flash mapping\n");
+			return -EINVAL;
+		}
+		
+		if (allocate_resource(&iomem_resource, &docmem, 
+				      size,
+				      0xc0000000, 0xffffffff, 
+				      size, NULL, NULL)) {
+			printk(KERN_ERR NAME ": unable to allocate memory for flash mapping\n");
+			return -ENOMEM;
+		}
+		
+		ctrl = 0x07000000 | ((size-1) >> 13);
+
+		printk(KERN_INFO "DOCCS BASE=0x%08lx, CTRL=0x%08lx\n", (long)docmem.start, (long)ctrl);
+		
+		pci_write_config_dword(bridge, SCx200_DOCCS_BASE, docmem.start);
+		pci_write_config_dword(bridge, SCx200_DOCCS_CTRL, ctrl);
+		pmr = inl(SCx200_CB_BASE + SCx200_PMR);
+		
+		if (width == 8) {
+			pmr &= ~(1<<6);
+		} else {
+			pmr |= (1<<6);
+		}
+		outl(pmr, SCx200_CB_BASE + SCx200_PMR);
+	}
+	
+       	printk(KERN_INFO NAME ": DOCCS mapped at 0x%lx-0x%lx, width %d\n", 
+	       docmem.start, docmem.end, width);
+
+	scx200_docflash_map.size = size;
+	if (width == 8)
+		scx200_docflash_map.buswidth = 1;
+	else
+		scx200_docflash_map.buswidth = 2;
+
+	scx200_docflash_map.map_priv_1 = (unsigned long)ioremap(docmem.start, scx200_docflash_map.size);
+	if (!scx200_docflash_map.map_priv_1) {
+		printk(KERN_ERR NAME ": failed to ioremap the flash\n");
+		release_resource(&docmem);
+		return -EIO;
+	}
+
+	mymtd = do_map_probe(flashtype, &scx200_docflash_map);
+	if (!mymtd) {
+		printk(KERN_ERR NAME ": unable to detect flash\n");
+		iounmap((void *)scx200_docflash_map.map_priv_1);
+		release_resource(&docmem);
+		return -ENXIO;
+	}
+
+	if (size < mymtd->size)
+		printk(KERN_WARNING NAME ": warning, flash mapping is smaller than flash size\n");
+
+	mymtd->module = THIS_MODULE;
+
+#if PARTITION
+	partition_info[3].offset = mymtd->size-partition_info[3].size;
+	partition_info[2].size = partition_info[3].offset-partition_info[2].offset;
+	add_mtd_partitions(mymtd, partition_info, NUM_PARTITIONS);
+#else
+	add_mtd_device(mymtd);
+#endif
+	return 0;
+}
+
+static void __exit cleanup_scx200_docflash(void)
+{
+	if (mymtd) {
+#if PARTITION
+		del_mtd_partitions(mymtd);
+#else
+		del_mtd_device(mymtd);
+#endif
+		map_destroy(mymtd);
+	}
+	if (scx200_docflash_map.map_priv_1) {
+		iounmap((void *)scx200_docflash_map.map_priv_1);
+		release_resource(&docmem);
+	}
+}
+
+module_init(init_scx200_docflash);
+module_exit(cleanup_scx200_docflash);
+
+/*
+    Local variables:
+        compile-command: "make -k -C ../../.. SUBDIRS=drivers/mtd/maps modules"
+        c-basic-offset: 8
+    End:
+*/

drivers/scsi/st.c

@@ -3832,14 +3832,14 @@ static int st_attach(Scsi_Device * SDp)
 	    tpnt->driverfs_dev_r[mode].parent = &SDp->sdev_driverfs_dev;
 	    tpnt->driverfs_dev_r[mode].bus = &scsi_driverfs_bus_type;
 	    tpnt->driverfs_dev_r[mode].driver_data =
-			(void *)(long)__mkdev(MAJOR_NR, i + (mode << 5));
+			(void *)(long)__mkdev(MAJOR_NR, dev_num + (mode << 5));
 	    device_register(&tpnt->driverfs_dev_r[mode]);
 	    device_create_file(&tpnt->driverfs_dev_r[mode], 
 			       &dev_attr_type);
 	    device_create_file(&tpnt->driverfs_dev_r[mode], &dev_attr_kdev);
 	    tpnt->de_r[mode] =
 		devfs_register (SDp->de, name, DEVFS_FL_DEFAULT,
-				MAJOR_NR, i + (mode << 5),
+				MAJOR_NR, dev_num + (mode << 5),
 				S_IFCHR | S_IRUGO | S_IWUGO,
 				&st_fops, NULL);
 	    /*  No-rewind entry  */
@@ -3851,7 +3851,7 @@ static int st_attach(Scsi_Device * SDp)
 	    tpnt->driverfs_dev_n[mode].parent= &SDp->sdev_driverfs_dev;
 	    tpnt->driverfs_dev_n[mode].bus = &scsi_driverfs_bus_type;
 	    tpnt->driverfs_dev_n[mode].driver_data =
-			(void *)(long)__mkdev(MAJOR_NR, i + (mode << 5) + 128);
+			(void *)(long)__mkdev(MAJOR_NR, dev_num + (mode << 5) + 128);
 	    device_register(&tpnt->driverfs_dev_n[mode]);
 	    device_create_file(&tpnt->driverfs_dev_n[mode], 
 			       &dev_attr_type);
@@ -3859,7 +3859,7 @@ static int st_attach(Scsi_Device * SDp)
 			       &dev_attr_kdev);
 	    tpnt->de_n[mode] =
 		devfs_register (SDp->de, name, DEVFS_FL_DEFAULT,
-				MAJOR_NR, i + (mode << 5) + 128,
+				MAJOR_NR, dev_num + (mode << 5) + 128,
 				S_IFCHR | S_IRUGO | S_IWUGO,
 				&st_fops, NULL);
 	}

fs/buffer.c

@@ -735,80 +735,22 @@ int sync_mapping_buffers(struct address_space *mapping)
 }
 EXPORT_SYMBOL(sync_mapping_buffers);
 
-/**
- * write_mapping_buffers - Start writeout of a mapping's "associated" buffers.
- * @mapping - the mapping which wants those buffers written.
- *
- * Starts I/O against dirty buffers which are on @mapping->private_list.
- * Those buffers must be backed by @mapping->assoc_mapping.
- *
- * The private_list buffers generally contain filesystem indirect blocks.
- * The idea is that the filesystem can start I/O against the indirects at
- * the same time as running generic_writepages(), so the indirect's
- * I/O will be merged with the data.
- *
- * We sneakliy write the buffers in probable tail-to-head order.  This is
- * because generic_writepages() writes in probable head-to-tail
- * order.  If the file is so huge that the data or the indirects overflow
- * the request queue we will at least get some merging this way.
- *
- * Any clean+unlocked buffers are de-listed.  clean/locked buffers must be
- * left on the list for an fsync() to wait on.
- *
- * Couldn't think of a smart way of avoiding livelock, so chose the dumb
- * way instead.
- *
- * FIXME: duplicates fsync_inode_buffers() functionality a bit.
+/*
+ * Called when we've recently written block `bblock', and it is known that
+ * `bblock' was for a buffer_boundary() buffer.  This means that the block at
+ * `bblock + 1' is probably a dirty indirect block.  Hunt it down and, if it's
+ * dirty, schedule it for IO.  So that indirects merge nicely with their data.
  */
-int write_mapping_buffers(struct address_space *mapping)
+void write_boundary_block(struct block_device *bdev,
+			sector_t bblock, unsigned blocksize)
 {
-	spinlock_t *lock;
-	struct address_space *buffer_mapping;
-	unsigned nr_to_write;	/* livelock avoidance */
-	struct list_head *lh;
-	int ret = 0;
-
-	if (list_empty(&mapping->private_list))
-		goto out;
-
-	buffer_mapping = mapping->assoc_mapping;
-	lock = &buffer_mapping->private_lock;
-	spin_lock(lock);
-	nr_to_write = 0;
-	lh = mapping->private_list.next;
-	while (lh != &mapping->private_list) {
-		lh = lh->next;
-		nr_to_write++;
-	}
-	nr_to_write *= 2;	/* Allow for some late additions */
-
-	while (nr_to_write-- && !list_empty(&mapping->private_list)) {
-		struct buffer_head *bh;
-
-		bh = BH_ENTRY(mapping->private_list.prev);
-		list_del_init(&bh->b_assoc_buffers);
-		if (!buffer_dirty(bh) && !buffer_locked(bh))
-			continue;
-		/* Stick it on the far end of the list. Order is preserved. */
-		list_add(&bh->b_assoc_buffers, &mapping->private_list);
-		if (test_set_buffer_locked(bh))
-			continue;
-		get_bh(bh);
-		spin_unlock(lock);
-		if (test_clear_buffer_dirty(bh)) {
-			bh->b_end_io = end_buffer_io_sync;
-			submit_bh(WRITE, bh);
-		} else {
-			unlock_buffer(bh);
-			put_bh(bh);
-		}
-		spin_lock(lock);
+	struct buffer_head *bh = __find_get_block(bdev, bblock + 1, blocksize);
+	if (bh) {
+		if (buffer_dirty(bh))
+			ll_rw_block(WRITE, 1, &bh);
+		put_bh(bh);
 	}
-	spin_unlock(lock);
-out:
-	return ret;
 }
-EXPORT_SYMBOL(write_mapping_buffers);
 
 void mark_buffer_dirty_inode(struct buffer_head *bh, struct inode *inode)
 {
@@ -2487,77 +2429,32 @@ int submit_bh(int rw, struct buffer_head * bh)
  * All of the buffers must be for the same device, and must also be a
  * multiple of the current approved size for the device.
  */
-void ll_rw_block(int rw, int nr, struct buffer_head * bhs[])
+void ll_rw_block(int rw, int nr, struct buffer_head *bhs[])
 {
-	unsigned int major;
-	int correct_size;
 	int i;
 
-	if (!nr)
-		return;
-
-	major = major(to_kdev_t(bhs[0]->b_bdev->bd_dev));
-
-	/* Determine correct block size for this device. */
-	correct_size = bdev_hardsect_size(bhs[0]->b_bdev);
-
-	/* Verify requested block sizes. */
-	for (i = 0; i < nr; i++) {
-		struct buffer_head *bh = bhs[i];
-		if (bh->b_size & (correct_size - 1)) {
-			printk(KERN_NOTICE "ll_rw_block: device %s: "
-			       "only %d-char blocks implemented (%u)\n",
-			       bdevname(bhs[0]->b_bdev),
-			       correct_size, bh->b_size);
-			goto sorry;
-		}
-	}
-
-	if ((rw & WRITE) && bdev_read_only(bhs[0]->b_bdev)) {
-		printk(KERN_NOTICE "Can't write to read-only device %s\n",
-		       bdevname(bhs[0]->b_bdev));
-		goto sorry;
-	}
-
 	for (i = 0; i < nr; i++) {
 		struct buffer_head *bh = bhs[i];
 
-		/* Only one thread can actually submit the I/O. */
 		if (test_set_buffer_locked(bh))
 			continue;
 
-		/* We have the buffer lock */
-		atomic_inc(&bh->b_count);
+		get_bh(bh);
 		bh->b_end_io = end_buffer_io_sync;
-
-		switch(rw) {
-		case WRITE:
-			if (!test_clear_buffer_dirty(bh))
-				/* Hmmph! Nothing to write */
-				goto end_io;
-			break;
-
-		case READA:
-		case READ:
-			if (buffer_uptodate(bh))
-				/* Hmmph! Already have it */
-				goto end_io;
-			break;
-		default:
-			BUG();
-	end_io:
-			bh->b_end_io(bh, buffer_uptodate(bh));
-			continue;
+		if (rw == WRITE) {
+			if (test_clear_buffer_dirty(bh)) {
+				submit_bh(WRITE, bh);
+				continue;
+			}
+		} else {
+			if (!buffer_uptodate(bh)) {
+				submit_bh(READ, bh);
+				continue;
+			}
 		}
-
-		submit_bh(rw, bh);
+		unlock_buffer(bh);
+		put_bh(bh);
 	}
-	return;
-
-sorry:
-	/* Make sure we don't get infinite dirty retries.. */
-	for (i = 0; i < nr; i++)
-		clear_buffer_dirty(bhs[i]);
 }
 
 /*

fs/direct-io.c

@@ -49,6 +49,7 @@ struct dio {
 	/* Page fetching state */
 	int curr_page;			/* changes */
 	int total_pages;		/* doesn't change */
+	int pages_left;			/* approximate total IO pages */
 	unsigned long curr_user_address;/* changes */
 
 	/* Page queue */
@@ -396,7 +397,8 @@ static int dio_new_bio(struct dio *dio)
 	if (ret)
 		goto out;
 	sector = dio->next_block_in_bio << (dio->blkbits - 9);
-	nr_pages = min(dio->total_pages, BIO_MAX_PAGES);
+	nr_pages = min(dio->pages_left, bio_get_nr_vecs(dio->map_bh.b_bdev));
+	BUG_ON(nr_pages <= 0);
 	ret = dio_bio_alloc(dio, dio->map_bh.b_bdev, sector, nr_pages);
 	dio->boundary = 0;
 out:
@@ -423,6 +425,7 @@ dio_bio_add_page(struct dio *dio, struct page *page,
 		}
 	}
 	page_cache_release(page);
+	dio->pages_left--;
 out:
 	return ret;
 }
@@ -566,6 +569,10 @@ direct_io_worker(int rw, struct inode *inode, const struct iovec *iov,
 	spin_lock_init(&dio.bio_list_lock);
 	dio.bio_list = NULL;
 	dio.waiter = NULL;
+	dio.pages_left = 0;
+
+	for (seg = 0; seg < nr_segs; seg++) 
+		dio.pages_left += (iov[seg].iov_len / PAGE_SIZE) + 2; 
 
 	for (seg = 0; seg < nr_segs; seg++) {
 		user_addr = (unsigned long)iov[seg].iov_base;
@@ -620,13 +627,11 @@ generic_direct_IO(int rw, struct inode *inode, const struct iovec *iov,
 	int seg;
 	size_t size;
 	unsigned long addr;
-	struct address_space *mapping = inode->i_mapping;
 	unsigned blocksize_mask = (1 << inode->i_blkbits) - 1;
 	ssize_t retval = -EINVAL;
 
-	if (offset & blocksize_mask) {
+	if (offset & blocksize_mask)
 		goto out;
-	}
 
 	/* Check the memory alignment.  Blocks cannot straddle pages */
 	for (seg = 0; seg < nr_segs; seg++) {
@@ -636,14 +641,6 @@ generic_direct_IO(int rw, struct inode *inode, const struct iovec *iov,
 			goto out;	
 	}
 
-	if (mapping->nrpages) {
-		retval = filemap_fdatawrite(mapping);
-		if (retval == 0)
-			retval = filemap_fdatawait(mapping);
-		if (retval)
-			goto out;
-	}
-
 	retval = direct_io_worker(rw, inode, iov, offset, nr_segs, get_blocks);
 out:
 	return retval;
@@ -656,8 +653,17 @@ generic_file_direct_IO(int rw, struct inode *inode, const struct iovec *iov,
 	struct address_space *mapping = inode->i_mapping;
 	ssize_t retval;
 
+	if (mapping->nrpages) {
+		retval = filemap_fdatawrite(mapping);
+		if (retval == 0)
+			retval = filemap_fdatawait(mapping);
+		if (retval)
+			goto out;
+	}
+
 	retval = mapping->a_ops->direct_IO(rw, inode, iov, offset, nr_segs);
 	if (inode->i_mapping->nrpages)
 		invalidate_inode_pages2(inode->i_mapping);
+out:
 	return retval;
 }

fs/ext2/inode.c

@@ -629,14 +629,7 @@ ext2_direct_IO(int rw, struct inode *inode, const struct iovec *iov,
 static int
 ext2_writepages(struct address_space *mapping, struct writeback_control *wbc)
 {
-	int ret;
-	int err;
-
-	ret = write_mapping_buffers(mapping);
-	err = mpage_writepages(mapping, wbc, ext2_get_block);
-	if (!ret)
-		ret = err;
-	return ret;
+	return mpage_writepages(mapping, wbc, ext2_get_block);
 }
 
 struct address_space_operations ext2_aops = {

fs/ext3/inode.c

@@ -1477,14 +1477,7 @@ struct address_space_operations ext3_aops = {
 static int
 ext3_writepages(struct address_space *mapping, struct writeback_control *wbc)
 {
-	int ret;
-	int err;
-
-	ret = write_mapping_buffers(mapping);
-	err = mpage_writepages(mapping, wbc, ext3_get_block);
-	if (!ret)
-		ret = err;
-	return ret;
+	return mpage_writepages(mapping, wbc, ext3_get_block);
 }
 #endif
 

fs/fat/inode.c

@@ -228,8 +228,6 @@ static int parse_options(char *options, int *debug,
 	save = 0;
 	savep = NULL;
 	while ((this_char = strsep(&options,",")) != NULL) {
-		if (!*this_char)
-			continue;
 		if ((value = strchr(this_char,'=')) != NULL) {
 			save = *value;
 			savep = value;
@@ -351,7 +349,7 @@ static int parse_options(char *options, int *debug,
 			strncpy(cvf_options,value,100);
 		}
 
-		if (this_char != options) *(this_char-1) = ',';
+		if (options) *(options-1) = ',';
 		if (value) *savep = save;
 		if (ret == 0)
 			break;

fs/hpfs/alloc.c

@@ -24,8 +24,8 @@ static int chk_if_allocated(struct super_block *s, secno sec, char *msg)
 		goto fail1;
 	}
 	hpfs_brelse4(&qbh);
-	if (sec >= s->s_hpfs_dirband_start && sec < s->s_hpfs_dirband_start + s->s_hpfs_dirband_size) {
-		unsigned ssec = (sec - s->s_hpfs_dirband_start) / 4;
+	if (sec >= hpfs_sb(s)->sb_dirband_start && sec < hpfs_sb(s)->sb_dirband_start + hpfs_sb(s)->sb_dirband_size) {
+		unsigned ssec = (sec - hpfs_sb(s)->sb_dirband_start) / 4;
 		if (!(bmp = hpfs_map_dnode_bitmap(s, &qbh))) goto fail;
 		if ((bmp[ssec >> 5] >> (ssec & 0x1f)) & 1) {
 			hpfs_error(s, "sector '%s' - %08x not allocated in directory bitmap", msg, sec);
@@ -48,11 +48,11 @@ static int chk_if_allocated(struct super_block *s, secno sec, char *msg)
 int hpfs_chk_sectors(struct super_block *s, secno start, int len, char *msg)
 {
 	if (start + len < start || start < 0x12 ||
-	    start + len > s->s_hpfs_fs_size) {
+	    start + len > hpfs_sb(s)->sb_fs_size) {
 	    	hpfs_error(s, "sector(s) '%s' badly placed at %08x", msg, start);
 		return 1;
 	}
-	if (s->s_hpfs_chk>=2) {
+	if (hpfs_sb(s)->sb_chk>=2) {
 		int i;
 		for (i = 0; i < len; i++)
 			if (chk_if_allocated(s, start + i, msg)) return 1;
@@ -127,7 +127,7 @@ static secno alloc_in_bmp(struct super_block *s, secno near, unsigned n, unsigne
 	}
 	rt:
 	if (ret) {
-		if (s->s_hpfs_chk && ((ret >> 14) != (bs >> 14) || (bmp[(ret & 0x3fff) >> 5] | ~(((1 << n) - 1) << (ret & 0x1f))) != 0xffffffff)) {
+		if (hpfs_sb(s)->sb_chk && ((ret >> 14) != (bs >> 14) || (bmp[(ret & 0x3fff) >> 5] | ~(((1 << n) - 1) << (ret & 0x1f))) != 0xffffffff)) {
 			hpfs_error(s, "Allocation doesn't work! Wanted %d, allocated at %08x", n, ret);
 			ret = 0;
 			goto b;
@@ -155,14 +155,15 @@ secno hpfs_alloc_sector(struct super_block *s, secno near, unsigned n, int forwa
 	secno sec;
 	unsigned i;
 	unsigned n_bmps;
-	int b = s->s_hpfs_c_bitmap;
+	struct hpfs_sb_info *sbi = hpfs_sb(s);
+	int b = sbi->sb_c_bitmap;
 	int f_p = 0;
 	if (forward < 0) {
 		forward = -forward;
 		f_p = 1;
 	}
 	if (lock) hpfs_lock_creation(s);
-	if (near && near < s->s_hpfs_fs_size)
+	if (near && near < sbi->sb_fs_size)
 		if ((sec = alloc_in_bmp(s, near, n, f_p ? forward : forward/4))) goto ret;
 	if (b != -1) {
 		if ((sec = alloc_in_bmp(s, b<<14, n, f_p ? forward : forward/2))) {
@@ -171,25 +172,25 @@ secno hpfs_alloc_sector(struct super_block *s, secno near, unsigned n, int forwa
 		}
 		if (b > 0x10000000) if ((sec = alloc_in_bmp(s, (b&0xfffffff)<<14, n, f_p ? forward : 0))) goto ret;
 	}	
-	n_bmps = (s->s_hpfs_fs_size + 0x4000 - 1) >> 14;
+	n_bmps = (sbi->sb_fs_size + 0x4000 - 1) >> 14;
 	for (i = 0; i < n_bmps / 2; i++) {
 		if ((sec = alloc_in_bmp(s, (n_bmps/2+i) << 14, n, forward))) {
-			s->s_hpfs_c_bitmap = n_bmps/2+i;
+			sbi->sb_c_bitmap = n_bmps/2+i;
 			goto ret;
 		}	
 		if ((sec = alloc_in_bmp(s, (n_bmps/2-i-1) << 14, n, forward))) {
-			s->s_hpfs_c_bitmap = n_bmps/2-i-1;
+			sbi->sb_c_bitmap = n_bmps/2-i-1;
 			goto ret;
 		}
 	}
 	if ((sec = alloc_in_bmp(s, (n_bmps-1) << 14, n, forward))) {
-		s->s_hpfs_c_bitmap = n_bmps-1;
+		sbi->sb_c_bitmap = n_bmps-1;
 		goto ret;
 	}
 	if (!f_p) {
 		for (i = 0; i < n_bmps; i++)
 			if ((sec = alloc_in_bmp(s, i << 14, n, 0))) {
-				s->s_hpfs_c_bitmap = 0x10000000 + i;
+				sbi->sb_c_bitmap = 0x10000000 + i;
 				goto ret;
 			}
 	}
@@ -212,17 +213,18 @@ static secno alloc_in_dirband(struct super_block *s, secno near, int lock)
 {
 	unsigned nr = near;
 	secno sec;
-	if (nr < s->s_hpfs_dirband_start)
-		nr = s->s_hpfs_dirband_start;
-	if (nr >= s->s_hpfs_dirband_start + s->s_hpfs_dirband_size)
-		nr = s->s_hpfs_dirband_start + s->s_hpfs_dirband_size - 4;
-	nr -= s->s_hpfs_dirband_start;
+	struct hpfs_sb_info *sbi = hpfs_sb(s);
+	if (nr < sbi->sb_dirband_start)
+		nr = sbi->sb_dirband_start;
+	if (nr >= sbi->sb_dirband_start + sbi->sb_dirband_size)
+		nr = sbi->sb_dirband_start + sbi->sb_dirband_size - 4;
+	nr -= sbi->sb_dirband_start;
 	nr >>= 2;
 	if (lock) hpfs_lock_creation(s);
 	sec = alloc_in_bmp(s, (~0x3fff) | nr, 1, 0);
 	if (lock) hpfs_unlock_creation(s);
 	if (!sec) return 0;
-	return ((sec & 0x3fff) << 2) + s->s_hpfs_dirband_start;
+	return ((sec & 0x3fff) << 2) + sbi->sb_dirband_start;
 }
 
 /* Alloc sector if it's free */
@@ -303,8 +305,8 @@ void hpfs_free_sectors(struct super_block *s, secno sec, unsigned n)
 
 int hpfs_check_free_dnodes(struct super_block *s, int n)
 {
-	int n_bmps = (s->s_hpfs_fs_size + 0x4000 - 1) >> 14;
-	int b = s->s_hpfs_c_bitmap & 0x0fffffff;
+	int n_bmps = (hpfs_sb(s)->sb_fs_size + 0x4000 - 1) >> 14;
+	int b = hpfs_sb(s)->sb_c_bitmap & 0x0fffffff;
 	int i, j;
 	unsigned *bmp;
 	struct quad_buffer_head qbh;
@@ -320,7 +322,7 @@ int hpfs_check_free_dnodes(struct super_block *s, int n)
 	}
 	hpfs_brelse4(&qbh);
 	i = 0;
-	if (s->s_hpfs_c_bitmap != -1 ) {
+	if (hpfs_sb(s)->sb_c_bitmap != -1 ) {
 		bmp = hpfs_map_bitmap(s, b, &qbh, "chkdn1");
 		goto chk_bmp;
 	}
@@ -349,17 +351,17 @@ int hpfs_check_free_dnodes(struct super_block *s, int n)
 
 void hpfs_free_dnode(struct super_block *s, dnode_secno dno)
 {
-	if (s->s_hpfs_chk) if (dno & 3) {
+	if (hpfs_sb(s)->sb_chk) if (dno & 3) {
 		hpfs_error(s, "hpfs_free_dnode: dnode %08x not aligned", dno);
 		return;
 	}
-	if (dno < s->s_hpfs_dirband_start ||
-	    dno >= s->s_hpfs_dirband_start + s->s_hpfs_dirband_size) {
+	if (dno < hpfs_sb(s)->sb_dirband_start ||
+	    dno >= hpfs_sb(s)->sb_dirband_start + hpfs_sb(s)->sb_dirband_size) {
 		hpfs_free_sectors(s, dno, 4);
 	} else {
 		struct quad_buffer_head qbh;
 		unsigned *bmp;
-		unsigned ssec = (dno - s->s_hpfs_dirband_start) / 4;
+		unsigned ssec = (dno - hpfs_sb(s)->sb_dirband_start) / 4;
 		lock_super(s);
 		if (!(bmp = hpfs_map_dnode_bitmap(s, &qbh))) {
 			unlock_super(s);
@@ -377,7 +379,7 @@ struct dnode *hpfs_alloc_dnode(struct super_block *s, secno near,
 			 int lock)
 {
 	struct dnode *d;
-	if (hpfs_count_one_bitmap(s, s->s_hpfs_dmap) > FREE_DNODES_ADD) {
+	if (hpfs_count_one_bitmap(s, hpfs_sb(s)->sb_dmap) > FREE_DNODES_ADD) {
 		if (!(*dno = alloc_in_dirband(s, near, lock)))
 			if (!(*dno = hpfs_alloc_sector(s, near, 4, 0, lock))) return NULL;
 	} else {

fs/hpfs/anode.c

@@ -20,7 +20,7 @@ secno hpfs_bplus_lookup(struct super_block *s, struct inode *inode,
 	int i;
 	int c1, c2 = 0;
 	go_down:
-	if (s->s_hpfs_chk) if (hpfs_stop_cycles(s, a, &c1, &c2, "hpfs_bplus_lookup")) return -1;
+	if (hpfs_sb(s)->sb_chk) if (hpfs_stop_cycles(s, a, &c1, &c2, "hpfs_bplus_lookup")) return -1;
 	if (btree->internal) {
 		for (i = 0; i < btree->n_used_nodes; i++)
 			if (btree->u.internal[i].file_secno > sec) {
@@ -38,7 +38,7 @@ secno hpfs_bplus_lookup(struct super_block *s, struct inode *inode,
 		if (btree->u.external[i].file_secno <= sec &&
 		    btree->u.external[i].file_secno + btree->u.external[i].length > sec) {
 			a = btree->u.external[i].disk_secno + sec - btree->u.external[i].file_secno;
-			if (s->s_hpfs_chk) if (hpfs_chk_sectors(s, a, 1, "data")) {
+			if (hpfs_sb(s)->sb_chk) if (hpfs_chk_sectors(s, a, 1, "data")) {
 				brelse(bh);
 				return -1;
 			}
@@ -88,7 +88,7 @@ secno hpfs_add_sector_to_btree(struct super_block *s, secno node, int fnod, unsi
 		btree->u.internal[n].file_secno = -1;
 		mark_buffer_dirty(bh);
 		brelse(bh);
-		if (s->s_hpfs_chk)
+		if (hpfs_sb(s)->sb_chk)
 			if (hpfs_stop_cycles(s, a, &c1, &c2, "hpfs_add_sector_to_btree #1")) return -1;
 		if (!(anode = hpfs_map_anode(s, a, &bh))) return -1;
 		btree = &anode->btree;
@@ -164,7 +164,7 @@ secno hpfs_add_sector_to_btree(struct super_block *s, secno node, int fnod, unsi
 	c2 = 0;
 	while (up != -1) {
 		struct anode *new_anode;
-		if (s->s_hpfs_chk)
+		if (hpfs_sb(s)->sb_chk)
 			if (hpfs_stop_cycles(s, up, &c1, &c2, "hpfs_add_sector_to_btree #2")) return -1;
 		if (up != node || !fnod) {
 			if (!(anode = hpfs_map_anode(s, up, &bh))) return -1;
@@ -283,7 +283,7 @@ void hpfs_remove_btree(struct super_block *s, struct bplus_header *btree)
 	while (btree1->internal) {
 		ano = btree1->u.internal[pos].down;
 		if (level) brelse(bh);
-		if (s->s_hpfs_chk)
+		if (hpfs_sb(s)->sb_chk)
 			if (hpfs_stop_cycles(s, ano, &d1, &d2, "hpfs_remove_btree #1"))
 				return;
 		if (!(anode = hpfs_map_anode(s, ano, &bh))) return;
@@ -296,7 +296,7 @@ void hpfs_remove_btree(struct super_block *s, struct bplus_header *btree)
 	go_up:
 	if (!level) return;
 	brelse(bh);
-	if (s->s_hpfs_chk)
+	if (hpfs_sb(s)->sb_chk)
 		if (hpfs_stop_cycles(s, ano, &c1, &c2, "hpfs_remove_btree #2")) return;
 	hpfs_free_sectors(s, ano, 1);
 	oano = ano;
@@ -343,7 +343,7 @@ int hpfs_ea_read(struct super_block *s, secno a, int ano, unsigned pos,
 			if ((sec = anode_lookup(s, a, pos >> 9)) == -1)
 				return -1;
 		} else sec = a + (pos >> 9);
-		if (s->s_hpfs_chk) if (hpfs_chk_sectors(s, sec, 1, "ea #1")) return -1;
+		if (hpfs_sb(s)->sb_chk) if (hpfs_chk_sectors(s, sec, 1, "ea #1")) return -1;
 		if (!(data = hpfs_map_sector(s, sec, &bh, (len - 1) >> 9)))
 			return -1;
 		l = 0x200 - (pos & 0x1ff); if (l > len) l = len;
@@ -366,7 +366,7 @@ int hpfs_ea_write(struct super_block *s, secno a, int ano, unsigned pos,
 			if ((sec = anode_lookup(s, a, pos >> 9)) == -1)
 				return -1;
 		} else sec = a + (pos >> 9);
-		if (s->s_hpfs_chk) if (hpfs_chk_sectors(s, sec, 1, "ea #2")) return -1;
+		if (hpfs_sb(s)->sb_chk) if (hpfs_chk_sectors(s, sec, 1, "ea #2")) return -1;
 		if (!(data = hpfs_map_sector(s, sec, &bh, (len - 1) >> 9)))
 			return -1;
 		l = 0x200 - (pos & 0x1ff); if (l > len) l = len;
@@ -440,7 +440,7 @@ void hpfs_truncate_btree(struct super_block *s, secno f, int fno, unsigned secs)
 		}
 		node = btree->u.internal[i].down;
 		brelse(bh);
-		if (s->s_hpfs_chk)
+		if (hpfs_sb(s)->sb_chk)
 			if (hpfs_stop_cycles(s, node, &c1, &c2, "hpfs_truncate_btree"))
 				return;
 		if (!(anode = hpfs_map_anode(s, node, &bh))) return;

fs/hpfs/buffer.c

@@ -15,7 +15,7 @@ void hpfs_lock_creation(struct super_block *s)
 #ifdef DEBUG_LOCKS
 	printk("lock creation\n");
 #endif
-	down(&s->u.hpfs_sb.hpfs_creation_de);
+	down(&hpfs_sb(s)->hpfs_creation_de);
 }
 
 void hpfs_unlock_creation(struct super_block *s)
@@ -23,7 +23,7 @@ void hpfs_unlock_creation(struct super_block *s)
 #ifdef DEBUG_LOCKS
 	printk("unlock creation\n");
 #endif
-	up(&s->u.hpfs_sb.hpfs_creation_de);
+	up(&hpfs_sb(s)->hpfs_creation_de);
 }
 
 void hpfs_lock_iget(struct super_block *s, int mode)
@@ -31,8 +31,8 @@ void hpfs_lock_iget(struct super_block *s, int mode)
 #ifdef DEBUG_LOCKS
 	printk("lock iget\n");
 #endif
-	while (s->s_hpfs_rd_inode) sleep_on(&s->s_hpfs_iget_q);
-	s->s_hpfs_rd_inode = mode;
+	while (hpfs_sb(s)->sb_rd_inode) sleep_on(&hpfs_sb(s)->sb_iget_q);
+	hpfs_sb(s)->sb_rd_inode = mode;
 }
 
 void hpfs_unlock_iget(struct super_block *s)
@@ -40,8 +40,8 @@ void hpfs_unlock_iget(struct super_block *s)
 #ifdef DEBUG_LOCKS
 	printk("unlock iget\n");
 #endif
-	s->s_hpfs_rd_inode = 0;
-	wake_up(&s->s_hpfs_iget_q);
+	hpfs_sb(s)->sb_rd_inode = 0;
+	wake_up(&hpfs_sb(s)->sb_iget_q);
 }
 
 void hpfs_lock_inode(struct inode *i)

fs/hpfs/dentry.c

@@ -28,7 +28,7 @@ int hpfs_hash_dentry(struct dentry *dentry, struct qstr *qstr)
 
 	hash = init_name_hash();
 	for (i = 0; i < l; i++)
-		hash = partial_name_hash(hpfs_upcase(dentry->d_sb->s_hpfs_cp_table,qstr->name[i]), hash);
+		hash = partial_name_hash(hpfs_upcase(hpfs_sb(dentry->d_sb)->sb_cp_table,qstr->name[i]), hash);
 	qstr->hash = end_name_hash(hash);
 
 	return 0;

fs/hpfs/dir.c

@@ -65,7 +65,7 @@ int hpfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	int c1, c2 = 0;
 	int ret = 0;
 
-	if (inode->i_sb->s_hpfs_chk) {
+	if (hpfs_sb(inode->i_sb)->sb_chk) {
 		if (hpfs_chk_sectors(inode->i_sb, inode->i_ino, 1, "dir_fnode")) {
 			ret = -EFSERROR;
 			goto out;
@@ -75,7 +75,7 @@ int hpfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 			goto out;
 		}
 	}
-	if (inode->i_sb->s_hpfs_chk >= 2) {
+	if (hpfs_sb(inode->i_sb)->sb_chk >= 2) {
 		struct buffer_head *bh;
 		struct fnode *fno;
 		int e = 0;
@@ -97,7 +97,7 @@ int hpfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 			goto out;
 		}
 	}
-	lc = inode->i_sb->s_hpfs_lowercase;
+	lc = hpfs_sb(inode->i_sb)->sb_lowercase;
 	if (filp->f_pos == 12) { /* diff -r requires this (note, that diff -r */
 		filp->f_pos = 13; /* also fails on msdos filesystem in 2.0) */
 		goto out;
@@ -114,7 +114,7 @@ int hpfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		/* This won't work when cycle is longer than number of dirents
 		   accepted by filldir, but what can I do?
 		   maybe killall -9 ls helps */
-		if (inode->i_sb->s_hpfs_chk)
+		if (hpfs_sb(inode->i_sb)->sb_chk)
 			if (hpfs_stop_cycles(inode->i_sb, filp->f_pos, &c1, &c2, "hpfs_readdir")) {
 				hpfs_unlock_inode(inode);
 				ret = -EFSERROR;
@@ -160,7 +160,7 @@ int hpfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 				goto out;
 			}
 			if (de->first || de->last) {
-				if (inode->i_sb->s_hpfs_chk) {
+				if (hpfs_sb(inode->i_sb)->sb_chk) {
 					if (de->first && !de->last && (de->namelen != 2 || de ->name[0] != 1 || de->name[1] != 1)) hpfs_error(inode->i_sb, "hpfs_readdir: bad ^A^A entry; pos = %08x", old_pos);
 					if (de->last && (de->namelen != 1 || de ->name[0] != 255)) hpfs_error(inode->i_sb, "hpfs_readdir: bad \\377 entry; pos = %08x", old_pos);
 				}
@@ -241,7 +241,7 @@ struct dentry *hpfs_lookup(struct inode *dir, struct dentry *dentry)
 	 * Go find or make an inode.
 	 */
 
-	hpfs_lock_iget(dir->i_sb, de->directory || (de->ea_size && dir->i_sb->s_hpfs_eas) ? 1 : 2);
+	hpfs_lock_iget(dir->i_sb, de->directory || (de->ea_size && hpfs_sb(dir->i_sb)->sb_eas) ? 1 : 2);
 	if (!(result = iget(dir->i_sb, ino))) {
 		hpfs_unlock_iget(dir->i_sb);
 		hpfs_error(dir->i_sb, "hpfs_lookup: can't get inode");

fs/hpfs/dnode.c

@@ -134,7 +134,7 @@ static void set_last_pointer(struct super_block *s, struct dnode *d, dnode_secno
 		hpfs_error(s, "set_last_pointer: empty dnode %08x", d->self);
 		return;
 	}
-	if (s->s_hpfs_chk) {
+	if (hpfs_sb(s)->sb_chk) {
 		if (de->down) {
 			hpfs_error(s, "set_last_pointer: dnode %08x has already last pointer %08x",
 				d->self, de_down_pointer(de));
@@ -253,7 +253,7 @@ int hpfs_add_to_dnode(struct inode *i, dnode_secno dno, unsigned char *name, uns
 		return 1;
 	}
 	go_up_a:
-	if (i->i_sb->s_hpfs_chk)
+	if (hpfs_sb(i->i_sb)->sb_chk)
 		if (hpfs_stop_cycles(i->i_sb, dno, &c1, &c2, "hpfs_add_to_dnode")) {
 			hpfs_brelse4(&qbh);
 			if (nd) kfree(nd);
@@ -379,7 +379,7 @@ int hpfs_add_dirent(struct inode *i, unsigned char *name, unsigned namelen,
 	int c1, c2 = 0;
 	dno = hpfs_inode->i_dno;
 	down:
-	if (i->i_sb->s_hpfs_chk)
+	if (hpfs_sb(i->i_sb)->sb_chk)
 		if (hpfs_stop_cycles(i->i_sb, dno, &c1, &c2, "hpfs_add_dirent")) return 1;
 	if (!(d = hpfs_map_dnode(i->i_sb, dno, &qbh))) return 1;
 	de_end = dnode_end_de(d);
@@ -427,11 +427,11 @@ static secno move_to_top(struct inode *i, dnode_secno from, dnode_secno to)
 	int c1, c2 = 0;
 	dno = from;
 	while (1) {
-		if (i->i_sb->s_hpfs_chk)
+		if (hpfs_sb(i->i_sb)->sb_chk)
 			if (hpfs_stop_cycles(i->i_sb, dno, &c1, &c2, "move_to_top"))
 				return 0;
 		if (!(dnode = hpfs_map_dnode(i->i_sb, dno, &qbh))) return 0;
-		if (i->i_sb->s_hpfs_chk) {
+		if (hpfs_sb(i->i_sb)->sb_chk) {
 			if (dnode->up != chk_up) {
 				hpfs_error(i->i_sb, "move_to_top: up pointer from %08x should be %08x, is %08x",
 					dno, chk_up, dnode->up);
@@ -519,7 +519,7 @@ static void delete_empty_dnode(struct inode *i, dnode_secno dno)
 		up = dnode->up;
 		de = dnode_first_de(dnode);
 		down = de->down ? de_down_pointer(de) : 0;
-		if (i->i_sb->s_hpfs_chk) if (root && !down) {
+		if (hpfs_sb(i->i_sb)->sb_chk) if (root && !down) {
 			hpfs_error(i->i_sb, "delete_empty_dnode: root dnode %08x is empty", dno);
 			goto end;
 		}
@@ -532,7 +532,7 @@ static void delete_empty_dnode(struct inode *i, dnode_secno dno)
 			struct buffer_head *bh;
 			struct dnode *d1;
 			struct quad_buffer_head qbh1;
-			if (i->i_sb->s_hpfs_chk) if (up != i->i_ino) {
+			if (hpfs_sb(i->i_sb)->sb_chk) if (up != i->i_ino) {
 				hpfs_error(i->i_sb, "bad pointer to fnode, dnode %08x, pointing to %08x, should be %08x", dno, up, i->i_ino);
 				return;
 			}
@@ -628,14 +628,14 @@ static void delete_empty_dnode(struct inode *i, dnode_secno dno)
 			dlp = del->down ? de_down_pointer(del) : 0;
 			if (!dlp && down) {
 				if (d1->first_free > 2044) {
-					if (i->i_sb->s_hpfs_chk >= 2) {
+					if (hpfs_sb(i->i_sb)->sb_chk >= 2) {
 						printk("HPFS: warning: unbalanced dnode tree, see hpfs.txt 4 more info\n");
 						printk("HPFS: warning: terminating balancing operation\n");
 					}
 					hpfs_brelse4(&qbh1);
 					goto endm;
 				}
-				if (i->i_sb->s_hpfs_chk >= 2) {
+				if (hpfs_sb(i->i_sb)->sb_chk >= 2) {
 					printk("HPFS: warning: unbalanced dnode tree, see hpfs.txt 4 more info\n");
 					printk("HPFS: warning: goin'on\n");
 				}
@@ -738,12 +738,12 @@ void hpfs_count_dnodes(struct super_block *s, dnode_secno dno, int *n_dnodes,
 	int d1, d2 = 0;
 	go_down:
 	if (n_dnodes) (*n_dnodes)++;
-	if (s->s_hpfs_chk)
+	if (hpfs_sb(s)->sb_chk)
 		if (hpfs_stop_cycles(s, dno, &c1, &c2, "hpfs_count_dnodes #1")) return;
 	ptr = 0;
 	go_up:
 	if (!(dnode = hpfs_map_dnode(s, dno, &qbh))) return;
-	if (s->s_hpfs_chk) if (odno && odno != -1 && dnode->up != odno)
+	if (hpfs_sb(s)->sb_chk) if (odno && odno != -1 && dnode->up != odno)
 		hpfs_error(s, "hpfs_count_dnodes: bad up pointer; dnode %08x, down %08x points to %08x", odno, dno, dnode->up);
 	de = dnode_first_de(dnode);
 	if (ptr) while(1) {
@@ -774,7 +774,7 @@ void hpfs_count_dnodes(struct super_block *s, dnode_secno dno, int *n_dnodes,
 		return;
 	}
 	hpfs_brelse4(&qbh);
-	if (s->s_hpfs_chk)
+	if (hpfs_sb(s)->sb_chk)
 		if (hpfs_stop_cycles(s, ptr, &d1, &d2, "hpfs_count_dnodes #2")) return;
 	odno = -1;
 	goto go_up;
@@ -811,11 +811,11 @@ dnode_secno hpfs_de_as_down_as_possible(struct super_block *s, dnode_secno dno)
 	int c1, c2 = 0;
 
 	again:
-	if (s->s_hpfs_chk)
+	if (hpfs_sb(s)->sb_chk)
 		if (hpfs_stop_cycles(s, d, &c1, &c2, "hpfs_de_as_down_as_possible"))
 			return d;
 	if (!(de = map_nth_dirent(s, d, 1, &qbh, NULL))) return dno;
-	if (s->s_hpfs_chk)
+	if (hpfs_sb(s)->sb_chk)
 		if (up && ((struct dnode *)qbh.data)->up != up)
 			hpfs_error(s, "hpfs_de_as_down_as_possible: bad up pointer; dnode %08x, down %08x points to %08x", up, d, ((struct dnode *)qbh.data)->up);
 	if (!de->down) {
@@ -901,7 +901,7 @@ struct hpfs_dirent *map_dirent(struct inode *inode, dnode_secno dno, char *name,
 
 	if (!S_ISDIR(inode->i_mode)) hpfs_error(inode->i_sb, "map_dirent: not a directory\n");
 	again:
-	if (inode->i_sb->s_hpfs_chk)
+	if (hpfs_sb(inode->i_sb)->sb_chk)
 		if (hpfs_stop_cycles(inode->i_sb, dno, &c1, &c2, "map_dirent")) return NULL;
 	if (!(dnode = hpfs_map_dnode(inode->i_sb, dno, qbh))) return NULL;
 	
@@ -1046,7 +1046,7 @@ struct hpfs_dirent *map_fnode_dirent(struct super_block *s, fnode_secno fno,
 	if (c < 0 && de->down) {
 		dno = de_down_pointer(de);
 		hpfs_brelse4(qbh);
-		if (s->s_hpfs_chk)
+		if (hpfs_sb(s)->sb_chk)
 			if (hpfs_stop_cycles(s, dno, &c1, &c2, "map_fnode_dirent #1")) {
 			kfree(name2);
 			return NULL;
@@ -1065,7 +1065,7 @@ struct hpfs_dirent *map_fnode_dirent(struct super_block *s, fnode_secno fno,
 	downd = dno;
 	dno = d->up;
 	hpfs_brelse4(qbh);
-	if (s->s_hpfs_chk)
+	if (hpfs_sb(s)->sb_chk)
 		if (hpfs_stop_cycles(s, downd, &d1, &d2, "map_fnode_dirent #2")) {
 			kfree(name2);
 			return NULL;

fs/hpfs/hpfs_fn.h

@@ -13,6 +13,7 @@
 #include <linux/fs.h>
 #include <linux/hpfs_fs.h>
 #include <linux/hpfs_fs_i.h>
+#include <linux/hpfs_fs_sb.h>
 #include <linux/errno.h>
 #include <linux/slab.h>
 #include <linux/kernel.h>
@@ -57,22 +58,6 @@
 
 typedef void nonconst; /* What this is for ? */
 
-/*
- * local time (HPFS) to GMT (Unix)
- */
-
-extern inline time_t local_to_gmt(struct super_block *s, time_t t)
-{
-	extern struct timezone sys_tz;
-	return t + sys_tz.tz_minuteswest * 60 + s->s_hpfs_timeshift;
-}
-
-extern inline time_t gmt_to_local(struct super_block *s, time_t t)
-{
-	extern struct timezone sys_tz;
-	return t - sys_tz.tz_minuteswest * 60 - s->s_hpfs_timeshift;
-}
-
 /*
  * conv= options
  */
@@ -309,6 +294,11 @@ static inline struct hpfs_inode_info *hpfs_i(struct inode *inode)
 	return list_entry(inode, struct hpfs_inode_info, vfs_inode);
 }
 
+static inline struct hpfs_sb_info *hpfs_sb(struct super_block *sb)
+{
+	return sb->u.generic_sbp;
+}
+
 /* super.c */
 
 void hpfs_error(struct super_block *, char *, ...);
@@ -319,3 +309,20 @@ unsigned hpfs_count_one_bitmap(struct super_block *, secno);
 int hpfs_statfs(struct super_block *, struct statfs *);
 
 extern struct address_space_operations hpfs_aops;
+
+/*
+ * local time (HPFS) to GMT (Unix)
+ */
+
+extern inline time_t local_to_gmt(struct super_block *s, time_t t)
+{
+	extern struct timezone sys_tz;
+	return t + sys_tz.tz_minuteswest * 60 + hpfs_sb(s)->sb_timeshift;
+}
+
+extern inline time_t gmt_to_local(struct super_block *s, time_t t)
+{
+	extern struct timezone sys_tz;
+	return t - sys_tz.tz_minuteswest * 60 - hpfs_sb(s)->sb_timeshift;
+}
+

fs/hpfs/inode.c

@@ -66,10 +66,10 @@ void hpfs_read_inode(struct inode *i)
 	unsigned char *ea;
 	int ea_size;
 
-	i->i_uid = sb->s_hpfs_uid;
-	i->i_gid = sb->s_hpfs_gid;
-	i->i_mode = sb->s_hpfs_mode;
-	hpfs_inode->i_conv = sb->s_hpfs_conv;
+	i->i_uid = hpfs_sb(sb)->sb_uid;
+	i->i_gid = hpfs_sb(sb)->sb_gid;
+	i->i_mode = hpfs_sb(sb)->sb_mode;
+	hpfs_inode->i_conv = hpfs_sb(sb)->sb_conv;
 	i->i_blksize = 512;
 	i->i_size = -1;
 	i->i_blocks = -1;
@@ -93,9 +93,9 @@ void hpfs_read_inode(struct inode *i)
 	i->i_mtime = 0;
 	i->i_ctime = 0;
 
-	if (!i->i_sb->s_hpfs_rd_inode)
-		hpfs_error(i->i_sb, "read_inode: s_hpfs_rd_inode == 0");
-	if (i->i_sb->s_hpfs_rd_inode == 2) {
+	if (!hpfs_sb(i->i_sb)->sb_rd_inode)
+		hpfs_error(i->i_sb, "read_inode: sb_rd_inode == 0");
+	if (hpfs_sb(i->i_sb)->sb_rd_inode == 2) {
 		i->i_mode |= S_IFREG;
 		i->i_mode &= ~0111;
 		i->i_op = &hpfs_file_iops;
@@ -112,7 +112,7 @@ void hpfs_read_inode(struct inode *i)
 		make_bad_inode(i);
 		return;
 	}
-	if (i->i_sb->s_hpfs_eas) {
+	if (hpfs_sb(i->i_sb)->sb_eas) {
 		if ((ea = hpfs_get_ea(i->i_sb, fnode, "UID", &ea_size))) {
 			if (ea_size == 2) {
 				i->i_uid = ea[0] + (ea[1] << 8);
@@ -140,7 +140,7 @@ void hpfs_read_inode(struct inode *i)
 		}
 		if ((ea = hpfs_get_ea(i->i_sb, fnode, "MODE", &ea_size))) {
 			int rdev = 0;
-			umode_t mode = sb->s_hpfs_mode;
+			umode_t mode = hpfs_sb(sb)->sb_mode;
 			if (ea_size == 2) {
 				mode = ea[0] + (ea[1] << 8);
 				hpfs_inode->i_ea_mode = 1;
@@ -171,7 +171,7 @@ void hpfs_read_inode(struct inode *i)
 		i->i_fop = &hpfs_dir_ops;
 		hpfs_inode->i_parent_dir = fnode->up;
 		hpfs_inode->i_dno = fnode->u.external[0].disk_secno;
-		if (sb->s_hpfs_chk >= 2) {
+		if (hpfs_sb(sb)->sb_chk >= 2) {
 			struct buffer_head *bh0;
 			if (hpfs_map_fnode(sb, hpfs_inode->i_parent_dir, &bh0)) brelse(bh0);
 		}
@@ -201,24 +201,24 @@ void hpfs_write_inode_ea(struct inode *i, struct fnode *fnode)
 		/* Some unknown structures like ACL may be in fnode,
 		   we'd better not overwrite them */
 		hpfs_error(i->i_sb, "fnode %08x has some unknown HPFS386 stuctures", i->i_ino);
-	} else if (i->i_sb->s_hpfs_eas >= 2) {
+	} else if (hpfs_sb(i->i_sb)->sb_eas >= 2) {
 		unsigned char ea[4];
-		if ((i->i_uid != i->i_sb->s_hpfs_uid) || hpfs_inode->i_ea_uid) {
+		if ((i->i_uid != hpfs_sb(i->i_sb)->sb_uid) || hpfs_inode->i_ea_uid) {
 			ea[0] = i->i_uid & 0xff;
 			ea[1] = i->i_uid >> 8;
 			hpfs_set_ea(i, fnode, "UID", ea, 2);
 			hpfs_inode->i_ea_uid = 1;
 		}
-		if ((i->i_gid != i->i_sb->s_hpfs_gid) || hpfs_inode->i_ea_gid) {
+		if ((i->i_gid != hpfs_sb(i->i_sb)->sb_gid) || hpfs_inode->i_ea_gid) {
 			ea[0] = i->i_gid & 0xff;
 			ea[1] = i->i_gid >> 8;
 			hpfs_set_ea(i, fnode, "GID", ea, 2);
 			hpfs_inode->i_ea_gid = 1;
 		}
 		if (!S_ISLNK(i->i_mode))
-			if ((i->i_mode != ((i->i_sb->s_hpfs_mode & ~(S_ISDIR(i->i_mode) ? 0 : 0111))
+			if ((i->i_mode != ((hpfs_sb(i->i_sb)->sb_mode & ~(S_ISDIR(i->i_mode) ? 0 : 0111))
 			  | (S_ISDIR(i->i_mode) ? S_IFDIR : S_IFREG))
-			  && i->i_mode != ((i->i_sb->s_hpfs_mode & ~(S_ISDIR(i->i_mode) ? 0222 : 0333))
+			  && i->i_mode != ((hpfs_sb(i->i_sb)->sb_mode & ~(S_ISDIR(i->i_mode) ? 0222 : 0333))
 			  | (S_ISDIR(i->i_mode) ? S_IFDIR : S_IFREG))) || hpfs_inode->i_ea_mode) {
 				ea[0] = i->i_mode & 0xff;
 				ea[1] = i->i_mode >> 8;
@@ -241,7 +241,7 @@ void hpfs_write_inode(struct inode *i)
 	struct hpfs_inode_info *hpfs_inode = hpfs_i(i);
 	struct inode *parent;
 	if (!i->i_nlink) return;
-	if (i->i_ino == i->i_sb->s_hpfs_root) return;
+	if (i->i_ino == hpfs_sb(i->i_sb)->sb_root) return;
 	if (hpfs_inode->i_rddir_off && !atomic_read(&i->i_count)) {
 		if (*hpfs_inode->i_rddir_off) printk("HPFS: write_inode: some position still there\n");
 		kfree(hpfs_inode->i_rddir_off);
@@ -264,9 +264,9 @@ void hpfs_write_inode_nolock(struct inode *i)
 	struct fnode *fnode;
 	struct quad_buffer_head qbh;
 	struct hpfs_dirent *de;
-	if (i->i_ino == i->i_sb->s_hpfs_root) return;
+	if (i->i_ino == hpfs_sb(i->i_sb)->sb_root) return;
 	if (!(fnode = hpfs_map_fnode(i->i_sb, i->i_ino, &bh))) return;
-	if (i->i_ino != i->i_sb->s_hpfs_root) {
+	if (i->i_ino != hpfs_sb(i->i_sb)->sb_root) {
 		if (!(de = map_fnode_dirent(i->i_sb, i->i_ino, fnode, &qbh))) {
 			brelse(bh);
 			return;
@@ -309,7 +309,7 @@ int hpfs_notify_change(struct dentry *dentry, struct iattr *attr)
 	int error=0;
 	lock_kernel();
 	if ( ((attr->ia_valid & ATTR_SIZE) && attr->ia_size > inode->i_size) ||
-	     (inode->i_sb->s_hpfs_root == inode->i_ino) ) {
+	     (hpfs_sb(inode->i_sb)->sb_root == inode->i_ino) ) {
 		error = -EINVAL;
 	} else if ((error = inode_change_ok(inode, attr))) {
 	} else if ((error = inode_setattr(inode, attr))) {

fs/hpfs/map.c

@@ -11,19 +11,19 @@
 
 unsigned *hpfs_map_dnode_bitmap(struct super_block *s, struct quad_buffer_head *qbh)
 {
-	return hpfs_map_4sectors(s, s->s_hpfs_dmap, qbh, 0);
+	return hpfs_map_4sectors(s, hpfs_sb(s)->sb_dmap, qbh, 0);
 }
 
 unsigned int *hpfs_map_bitmap(struct super_block *s, unsigned bmp_block,
 			 struct quad_buffer_head *qbh, char *id)
 {
 	secno sec;
-	if (s->s_hpfs_chk) if (bmp_block * 16384 > s->s_hpfs_fs_size) {
+	if (hpfs_sb(s)->sb_chk) if (bmp_block * 16384 > hpfs_sb(s)->sb_fs_size) {
 		hpfs_error(s, "hpfs_map_bitmap called with bad parameter: %08x at %s", bmp_block, id);
 		return NULL;
 	}
-	sec = s->s_hpfs_bmp_dir[bmp_block];
-	if (!sec || sec > s->s_hpfs_fs_size-4) {
+	sec = hpfs_sb(s)->sb_bmp_dir[bmp_block];
+	if (!sec || sec > hpfs_sb(s)->sb_fs_size-4) {
 		hpfs_error(s, "invalid bitmap block pointer %08x -> %08x at %s", bmp_block, sec, id);
 		return NULL;
 	}
@@ -93,7 +93,7 @@ char *hpfs_load_code_page(struct super_block *s, secno cps)
 secno *hpfs_load_bitmap_directory(struct super_block *s, secno bmp)
 {
 	struct buffer_head *bh;
-	int n = (s->s_hpfs_fs_size + 0x200000 - 1) >> 21;
+	int n = (hpfs_sb(s)->sb_fs_size + 0x200000 - 1) >> 21;
 	int i;
 	secno *b;
 	if (!(b = kmalloc(n * 512, GFP_KERNEL))) {
@@ -119,11 +119,11 @@ secno *hpfs_load_bitmap_directory(struct super_block *s, secno bmp)
 struct fnode *hpfs_map_fnode(struct super_block *s, ino_t ino, struct buffer_head **bhp)
 {
 	struct fnode *fnode;
-	if (s->s_hpfs_chk) if (hpfs_chk_sectors(s, ino, 1, "fnode")) {
+	if (hpfs_sb(s)->sb_chk) if (hpfs_chk_sectors(s, ino, 1, "fnode")) {
 		return NULL;
 	}
 	if ((fnode = hpfs_map_sector(s, ino, bhp, FNODE_RD_AHEAD))) {
-		if (s->s_hpfs_chk) {
+		if (hpfs_sb(s)->sb_chk) {
 			struct extended_attribute *ea;
 			struct extended_attribute *ea_end;
 			if (fnode->magic != FNODE_MAGIC) {
@@ -168,9 +168,9 @@ struct fnode *hpfs_map_fnode(struct super_block *s, ino_t ino, struct buffer_hea
 struct anode *hpfs_map_anode(struct super_block *s, anode_secno ano, struct buffer_head **bhp)
 {
 	struct anode *anode;
-	if (s->s_hpfs_chk) if (hpfs_chk_sectors(s, ano, 1, "anode")) return NULL;
+	if (hpfs_sb(s)->sb_chk) if (hpfs_chk_sectors(s, ano, 1, "anode")) return NULL;
 	if ((anode = hpfs_map_sector(s, ano, bhp, ANODE_RD_AHEAD)))
-		if (s->s_hpfs_chk) {
+		if (hpfs_sb(s)->sb_chk) {
 			if (anode->magic != ANODE_MAGIC || anode->self != ano) {
 				hpfs_error(s, "bad magic on anode %08x", ano);
 				goto bail;
@@ -200,7 +200,7 @@ struct dnode *hpfs_map_dnode(struct super_block *s, unsigned secno,
 			     struct quad_buffer_head *qbh)
 {
 	struct dnode *dnode;
-	if (s->s_hpfs_chk) {
+	if (hpfs_sb(s)->sb_chk) {
 		if (hpfs_chk_sectors(s, secno, 4, "dnode")) return NULL;
 		if (secno & 3) {
 			hpfs_error(s, "dnode %08x not byte-aligned", secno);
@@ -208,7 +208,7 @@ struct dnode *hpfs_map_dnode(struct super_block *s, unsigned secno,
 		}	
 	}
 	if ((dnode = hpfs_map_4sectors(s, secno, qbh, DNODE_RD_AHEAD)))
-		if (s->s_hpfs_chk) {
+		if (hpfs_sb(s)->sb_chk) {
 			unsigned p, pp = 0;
 			unsigned char *d = (char *)dnode;
 			int b = 0;
@@ -234,7 +234,7 @@ struct dnode *hpfs_map_dnode(struct super_block *s, unsigned secno,
 					hpfs_error(s, "namelen does not match dirent size in dnode %08x, dirent %03x, last %03x", secno, p, pp);
 					goto bail;
 				}
-				if (s->s_hpfs_chk >= 2) b |= 1 << de->down;
+				if (hpfs_sb(s)->sb_chk >= 2) b |= 1 << de->down;
 				if (de->down) if (de_down_pointer(de) < 0x10) {
 					hpfs_error(s, "bad down pointer in dnode %08x, dirent %03x, last %03x", secno, p, pp);
 					goto bail;

fs/hpfs/name.c

@@ -89,7 +89,7 @@ char *hpfs_translate_name(struct super_block *s, unsigned char *from,
 {
 	char *to;
 	int i;
-	if (s->s_hpfs_chk >= 2) if (hpfs_is_name_long(from, len) != lng) {
+	if (hpfs_sb(s)->sb_chk >= 2) if (hpfs_is_name_long(from, len) != lng) {
 		printk("HPFS: Long name flag mismatch - name ");
 		for (i=0; i<len; i++) printk("%c", from[i]);
 		printk(" misidentified as %s.\n", lng ? "short" : "long");
@@ -100,7 +100,7 @@ char *hpfs_translate_name(struct super_block *s, unsigned char *from,
 		printk("HPFS: can't allocate memory for name conversion buffer\n");
 		return from;
 	}
-	for (i = 0; i < len; i++) to[i] = locase(s->s_hpfs_cp_table,from[i]);
+	for (i = 0; i < len; i++) to[i] = locase(hpfs_sb(s)->sb_cp_table,from[i]);
 	return to;
 }
 
@@ -111,8 +111,8 @@ int hpfs_compare_names(struct super_block *s, unsigned char *n1, unsigned l1,
 	unsigned i;
 	if (last) return -1;
 	for (i = 0; i < l; i++) {
-		unsigned char c1 = upcase(s->s_hpfs_cp_table,n1[i]);
-		unsigned char c2 = upcase(s->s_hpfs_cp_table,n2[i]);
+		unsigned char c1 = upcase(hpfs_sb(s)->sb_cp_table,n1[i]);
+		unsigned char c2 = upcase(hpfs_sb(s)->sb_cp_table,n2[i]);
 		if (c1 < c2) return -1;
 		if (c1 > c2) return 1;
 	}

fs/hpfs/namei.c

@@ -187,7 +187,7 @@ int hpfs_mknod(struct inode *dir, struct dentry *dentry, int mode, int rdev)
 	struct inode *result = NULL;
 	int err;
 	if ((err = hpfs_chk_name((char *)name, &len))) return err==-ENOENT ? -EINVAL : err;
-	if (dir->i_sb->s_hpfs_eas < 2) return -EPERM;
+	if (hpfs_sb(dir->i_sb)->sb_eas < 2) return -EPERM;
 	lock_kernel();
 	if (!(fnode = hpfs_alloc_fnode(dir->i_sb, hpfs_i(dir)->i_dno, &fno, &bh))) goto bail;
 	memset(&dee, 0, sizeof dee);
@@ -255,7 +255,7 @@ int hpfs_symlink(struct inode *dir, struct dentry *dentry, const char *symlink)
 	int err;
 	if ((err = hpfs_chk_name((char *)name, &len))) return err==-ENOENT ? -EINVAL : err;
 	lock_kernel();
-	if (dir->i_sb->s_hpfs_eas < 2) {
+	if (hpfs_sb(dir->i_sb)->sb_eas < 2) {
 		unlock_kernel();
 		return -EPERM;
 	}
@@ -559,7 +559,7 @@ int hpfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		mark_buffer_dirty(bh);
 		brelse(bh);
 	}
-	hpfs_i(i)->i_conv = i->i_sb->s_hpfs_conv;
+	hpfs_i(i)->i_conv = hpfs_sb(i->i_sb)->sb_conv;
 	hpfs_decide_conv(i, (char *)new_name, new_len);
 	end1:
 	hpfs_unlock_3inodes(old_dir, new_dir, i);

fs/hpfs/super.c

@@ -16,7 +16,7 @@
 
 static void mark_dirty(struct super_block *s)
 {
-	if (s->s_hpfs_chkdsk && !(s->s_flags & MS_RDONLY)) {
+	if (hpfs_sb(s)->sb_chkdsk && !(s->s_flags & MS_RDONLY)) {
 		struct buffer_head *bh;
 		struct hpfs_spare_block *sb;
 		if ((sb = hpfs_map_sector(s, 17, &bh, 0))) {
@@ -37,8 +37,8 @@ static void unmark_dirty(struct super_block *s)
 	struct hpfs_spare_block *sb;
 	if (s->s_flags & MS_RDONLY) return;
 	if ((sb = hpfs_map_sector(s, 17, &bh, 0))) {
-		sb->dirty = s->s_hpfs_chkdsk > 1 - s->s_hpfs_was_error;
-		sb->old_wrote = s->s_hpfs_chkdsk >= 2 && !s->s_hpfs_was_error;
+		sb->dirty = hpfs_sb(s)->sb_chkdsk > 1 - hpfs_sb(s)->sb_was_error;
+		sb->old_wrote = hpfs_sb(s)->sb_chkdsk >= 2 && !hpfs_sb(s)->sb_was_error;
 		mark_buffer_dirty(bh);
 		brelse(bh);
 	}
@@ -60,12 +60,12 @@ void hpfs_error(struct super_block *s, char *m,...)
 	printk("HPFS: filesystem error: ");
 	if (buf) printk("%s", buf);
 	else printk("%s\n",m);
-	if (!s->s_hpfs_was_error) {
-		if (s->s_hpfs_err == 2) {
+	if (!hpfs_sb(s)->sb_was_error) {
+		if (hpfs_sb(s)->sb_err == 2) {
 			printk("; crashing the system because you wanted it\n");
 			mark_dirty(s);
 			panic("HPFS panic");
-		} else if (s->s_hpfs_err == 1) {
+		} else if (hpfs_sb(s)->sb_err == 1) {
 			if (s->s_flags & MS_RDONLY) printk("; already mounted read-only\n");
 			else {
 				printk("; remounting read-only\n");
@@ -76,7 +76,7 @@ void hpfs_error(struct super_block *s, char *m,...)
 		else printk("; corrupted filesystem mounted read/write - your computer will explode within 20 seconds ... but you wanted it so!\n");
 	} else printk("\n");
 	if (buf) kfree(buf);
-	s->s_hpfs_was_error = 1;
+	hpfs_sb(s)->sb_was_error = 1;
 }
 
 /* 
@@ -101,9 +101,12 @@ int hpfs_stop_cycles(struct super_block *s, int key, int *c1, int *c2,
 
 void hpfs_put_super(struct super_block *s)
 {
-	if (s->s_hpfs_cp_table) kfree(s->s_hpfs_cp_table);
-	if (s->s_hpfs_bmp_dir) kfree(s->s_hpfs_bmp_dir);
+	struct hpfs_sb_info *sbi = hpfs_sb(s);
+	if (sbi->sb_cp_table) kfree(sbi->sb_cp_table);
+	if (sbi->sb_bmp_dir) kfree(sbi->sb_bmp_dir);
 	unmark_dirty(s);
+	s->u.generic_sbp = NULL;
+	kfree(sbi);
 }
 
 unsigned hpfs_count_one_bitmap(struct super_block *s, secno secno)
@@ -125,28 +128,29 @@ unsigned hpfs_count_one_bitmap(struct super_block *s, secno secno)
 static unsigned count_bitmaps(struct super_block *s)
 {
 	unsigned n, count, n_bands;
-	n_bands = (s->s_hpfs_fs_size + 0x3fff) >> 14;
+	n_bands = (hpfs_sb(s)->sb_fs_size + 0x3fff) >> 14;
 	count = 0;
 	for (n = 0; n < n_bands; n++)
-		count += hpfs_count_one_bitmap(s, s->s_hpfs_bmp_dir[n]);
+		count += hpfs_count_one_bitmap(s, hpfs_sb(s)->sb_bmp_dir[n]);
 	return count;
 }
 
 int hpfs_statfs(struct super_block *s, struct statfs *buf)
 {
+	struct hpfs_sb_info *sbi = hpfs_sb(s);
 	lock_kernel();
 
-	/*if (s->s_hpfs_n_free == -1) {*/
-		s->s_hpfs_n_free = count_bitmaps(s);
-		s->s_hpfs_n_free_dnodes = hpfs_count_one_bitmap(s, s->s_hpfs_dmap);
+	/*if (sbi->sb_n_free == -1) {*/
+		sbi->sb_n_free = count_bitmaps(s);
+		sbi->sb_n_free_dnodes = hpfs_count_one_bitmap(s, sbi->sb_dmap);
 	/*}*/
 	buf->f_type = s->s_magic;
 	buf->f_bsize = 512;
-	buf->f_blocks = s->s_hpfs_fs_size;
-	buf->f_bfree = s->s_hpfs_n_free;
-	buf->f_bavail = s->s_hpfs_n_free;
-	buf->f_files = s->s_hpfs_dirband_size / 4;
-	buf->f_ffree = s->s_hpfs_n_free_dnodes;
+	buf->f_blocks = sbi->sb_fs_size;
+	buf->f_bfree = sbi->sb_n_free;
+	buf->f_bavail = sbi->sb_n_free;
+	buf->f_files = sbi->sb_dirband_size / 4;
+	buf->f_ffree = sbi->sb_n_free_dnodes;
 	buf->f_namelen = 254;
 
 	unlock_kernel();
@@ -377,14 +381,15 @@ int hpfs_remount_fs(struct super_block *s, int *flags, char *data)
 	umode_t umask;
 	int lowercase, conv, eas, chk, errs, chkdsk, timeshift;
 	int o;
+	struct hpfs_sb_info *sbi = hpfs_sb(s);
 	
 	*flags |= MS_NOATIME;
 	
-	uid = s->s_hpfs_uid; gid = s->s_hpfs_gid;
-	umask = 0777 & ~s->s_hpfs_mode;
-	lowercase = s->s_hpfs_lowercase; conv = s->s_hpfs_conv;
-	eas = s->s_hpfs_eas; chk = s->s_hpfs_chk; chkdsk = s->s_hpfs_chkdsk;
-	errs = s->s_hpfs_err; timeshift = s->s_hpfs_timeshift;
+	uid = sbi->sb_uid; gid = sbi->sb_gid;
+	umask = 0777 & ~sbi->sb_mode;
+	lowercase = sbi->sb_lowercase; conv = sbi->sb_conv;
+	eas = sbi->sb_eas; chk = sbi->sb_chk; chkdsk = sbi->sb_chkdsk;
+	errs = sbi->sb_err; timeshift = sbi->sb_timeshift;
 
 	if (!(o = parse_opts(data, &uid, &gid, &umask, &lowercase, &conv,
 	    &eas, &chk, &errs, &chkdsk, &timeshift))) {
@@ -395,18 +400,18 @@ int hpfs_remount_fs(struct super_block *s, int *flags, char *data)
 		hpfs_help();
 		return 1;
 	}
-	if (timeshift != s->s_hpfs_timeshift) {
+	if (timeshift != sbi->sb_timeshift) {
 		printk("HPFS: timeshift can't be changed using remount.\n");
 		return 1;
 	}
 
 	unmark_dirty(s);
 
-	s->s_hpfs_uid = uid; s->s_hpfs_gid = gid;
-	s->s_hpfs_mode = 0777 & ~umask;
-	s->s_hpfs_lowercase = lowercase; s->s_hpfs_conv = conv;
-	s->s_hpfs_eas = eas; s->s_hpfs_chk = chk; s->s_hpfs_chkdsk = chkdsk;
-	s->s_hpfs_err = errs; s->s_hpfs_timeshift = timeshift;
+	sbi->sb_uid = uid; sbi->sb_gid = gid;
+	sbi->sb_mode = 0777 & ~umask;
+	sbi->sb_lowercase = lowercase; sbi->sb_conv = conv;
+	sbi->sb_eas = eas; sbi->sb_chk = chk; sbi->sb_chkdsk = chkdsk;
+	sbi->sb_err = errs; sbi->sb_timeshift = timeshift;
 
 	if (!(*flags & MS_RDONLY)) mark_dirty(s);
 
@@ -419,6 +424,7 @@ static int hpfs_fill_super(struct super_block *s, void *options, int silent)
 	struct hpfs_boot_block *bootblock;
 	struct hpfs_super_block *superblock;
 	struct hpfs_spare_block *spareblock;
+	struct hpfs_sb_info *sbi;
 
 	uid_t uid;
 	gid_t gid;
@@ -431,12 +437,18 @@ static int hpfs_fill_super(struct super_block *s, void *options, int silent)
 
 	int o;
 
-	s->s_hpfs_bmp_dir = NULL;
-	s->s_hpfs_cp_table = NULL;
+	sbi = kmalloc(sizeof(*sbi), GFP_KERNEL);
+	if (!sbi)
+		return -ENOMEM;
+	s->u.generic_sbp = sbi;
+	memset(sbi, 0, sizeof(*sbi));
+
+	sbi->sb_bmp_dir = NULL;
+	sbi->sb_cp_table = NULL;
 
-	s->s_hpfs_rd_inode = 0;
-	init_MUTEX(&s->u.hpfs_sb.hpfs_creation_de);
-	init_waitqueue_head(&s->s_hpfs_iget_q);
+	sbi->sb_rd_inode = 0;
+	init_MUTEX(&sbi->hpfs_creation_de);
+	init_waitqueue_head(&sbi->sb_iget_q);
 
 	uid = current->uid;
 	gid = current->gid;
@@ -459,9 +471,9 @@ static int hpfs_fill_super(struct super_block *s, void *options, int silent)
 		goto bail0;
 	}
 
-	/*s->s_hpfs_mounting = 1;*/
+	/*sbi->sb_mounting = 1;*/
 	sb_set_blocksize(s, 512);
-	s->s_hpfs_fs_size = -1;
+	sbi->sb_fs_size = -1;
 	if (!(bootblock = hpfs_map_sector(s, 0, &bh0, 0))) goto bail1;
 	if (!(superblock = hpfs_map_sector(s, 16, &bh1, 1))) goto bail2;
 	if (!(spareblock = hpfs_map_sector(s, 17, &bh2, 0))) goto bail3;
@@ -489,30 +501,30 @@ static int hpfs_fill_super(struct super_block *s, void *options, int silent)
 	s->s_magic = HPFS_SUPER_MAGIC;
 	s->s_op = &hpfs_sops;
 
-	s->s_hpfs_root = superblock->root;
-	s->s_hpfs_fs_size = superblock->n_sectors;
-	s->s_hpfs_bitmaps = superblock->bitmaps;
-	s->s_hpfs_dirband_start = superblock->dir_band_start;
-	s->s_hpfs_dirband_size = superblock->n_dir_band;
-	s->s_hpfs_dmap = superblock->dir_band_bitmap;
-	s->s_hpfs_uid = uid;
-	s->s_hpfs_gid = gid;
-	s->s_hpfs_mode = 0777 & ~umask;
-	s->s_hpfs_n_free = -1;
-	s->s_hpfs_n_free_dnodes = -1;
-	s->s_hpfs_lowercase = lowercase;
-	s->s_hpfs_conv = conv;
-	s->s_hpfs_eas = eas;
-	s->s_hpfs_chk = chk;
-	s->s_hpfs_chkdsk = chkdsk;
-	s->s_hpfs_err = errs;
-	s->s_hpfs_timeshift = timeshift;
-	s->s_hpfs_was_error = 0;
-	s->s_hpfs_cp_table = NULL;
-	s->s_hpfs_c_bitmap = -1;
+	sbi->sb_root = superblock->root;
+	sbi->sb_fs_size = superblock->n_sectors;
+	sbi->sb_bitmaps = superblock->bitmaps;
+	sbi->sb_dirband_start = superblock->dir_band_start;
+	sbi->sb_dirband_size = superblock->n_dir_band;
+	sbi->sb_dmap = superblock->dir_band_bitmap;
+	sbi->sb_uid = uid;
+	sbi->sb_gid = gid;
+	sbi->sb_mode = 0777 & ~umask;
+	sbi->sb_n_free = -1;
+	sbi->sb_n_free_dnodes = -1;
+	sbi->sb_lowercase = lowercase;
+	sbi->sb_conv = conv;
+	sbi->sb_eas = eas;
+	sbi->sb_chk = chk;
+	sbi->sb_chkdsk = chkdsk;
+	sbi->sb_err = errs;
+	sbi->sb_timeshift = timeshift;
+	sbi->sb_was_error = 0;
+	sbi->sb_cp_table = NULL;
+	sbi->sb_c_bitmap = -1;
 	
 	/* Load bitmap directory */
-	if (!(s->s_hpfs_bmp_dir = hpfs_load_bitmap_directory(s, superblock->bitmaps)))
+	if (!(sbi->sb_bmp_dir = hpfs_load_bitmap_directory(s, superblock->bitmaps)))
 		goto bail4;
 	
 	/* Check for general fs errors*/
@@ -557,20 +569,20 @@ static int hpfs_fill_super(struct super_block *s, void *options, int silent)
 				superblock->dir_band_start, superblock->dir_band_end, superblock->n_dir_band);
 			goto bail4;
 		}
-		a = s->s_hpfs_dirband_size;
-		s->s_hpfs_dirband_size = 0;
+		a = sbi->sb_dirband_size;
+		sbi->sb_dirband_size = 0;
 		if (hpfs_chk_sectors(s, superblock->dir_band_start, superblock->n_dir_band, "dir_band") ||
 		    hpfs_chk_sectors(s, superblock->dir_band_bitmap, 4, "dir_band_bitmap") ||
 		    hpfs_chk_sectors(s, superblock->bitmaps, 4, "bitmaps")) {
 			mark_dirty(s);
 			goto bail4;
 		}
-		s->s_hpfs_dirband_size = a;
+		sbi->sb_dirband_size = a;
 	} else printk("HPFS: You really don't want any checks? You are crazy...\n");
 
 	/* Load code page table */
 	if (spareblock->n_code_pages)
-		if (!(s->s_hpfs_cp_table = hpfs_load_code_page(s, spareblock->code_page_dir)))
+		if (!(sbi->sb_cp_table = hpfs_load_code_page(s, spareblock->code_page_dir)))
 			printk("HPFS: Warning: code page support is disabled\n");
 
 	brelse(bh2);
@@ -578,7 +590,7 @@ static int hpfs_fill_super(struct super_block *s, void *options, int silent)
 	brelse(bh0);
 
 	hpfs_lock_iget(s, 1);
-	s->s_root = d_alloc_root(iget(s, s->s_hpfs_root));
+	s->s_root = d_alloc_root(iget(s, sbi->sb_root));
 	hpfs_unlock_iget(s);
 	if (!s->s_root || !s->s_root->d_inode) {
 		printk("HPFS: iget failed. Why???\n");
@@ -590,7 +602,7 @@ static int hpfs_fill_super(struct super_block *s, void *options, int silent)
 	 * find the root directory's . pointer & finish filling in the inode
 	 */
 
-	root_dno = hpfs_fnode_dno(s, s->s_hpfs_root);
+	root_dno = hpfs_fnode_dno(s, sbi->sb_root);
 	if (root_dno)
 		de = map_dirent(s->s_root->d_inode, root_dno, "\001\001", 2, NULL, &qbh);
 	if (!root_dno || !de) hpfs_error(s, "unable to find root dir");
@@ -612,8 +624,10 @@ bail3:	brelse(bh1);
 bail2:	brelse(bh0);
 bail1:
 bail0:
-	if (s->s_hpfs_bmp_dir) kfree(s->s_hpfs_bmp_dir);
-	if (s->s_hpfs_cp_table) kfree(s->s_hpfs_cp_table);
+	if (sbi->sb_bmp_dir) kfree(sbi->sb_bmp_dir);
+	if (sbi->sb_cp_table) kfree(sbi->sb_cp_table);
+	s->u.generic_sbp = NULL;
+	kfree(sbi);
 	return -EINVAL;
 }
 

fs/mpage.c

@@ -331,6 +331,8 @@ mpage_writepage(struct bio *bio, struct page *page, get_block_t get_block,
 	unsigned first_unmapped = blocks_per_page;
 	struct block_device *bdev = NULL;
 	int boundary = 0;
+	sector_t boundary_block = 0;
+	struct block_device *boundary_bdev = NULL;
 
 	if (page_has_buffers(page)) {
 		struct buffer_head *head = page_buffers(page);
@@ -363,6 +365,10 @@ mpage_writepage(struct bio *bio, struct page *page, get_block_t get_block,
 			}
 			blocks[page_block++] = bh->b_blocknr;
 			boundary = buffer_boundary(bh);
+			if (boundary) {
+				boundary_block = bh->b_blocknr;
+				boundary_bdev = bh->b_bdev;
+			}
 			bdev = bh->b_bdev;
 		} while ((bh = bh->b_this_page) != head);
 
@@ -393,6 +399,10 @@ mpage_writepage(struct bio *bio, struct page *page, get_block_t get_block,
 		if (buffer_new(&map_bh))
 			unmap_underlying_metadata(map_bh.b_bdev,
 						map_bh.b_blocknr);
+		if (buffer_boundary(&map_bh)) {
+			boundary_block = map_bh.b_blocknr;
+			boundary_bdev = map_bh.b_bdev;
+		}
 		if (page_block) {
 			if (map_bh.b_blocknr != blocks[page_block-1] + 1)
 				goto confused;
@@ -430,10 +440,8 @@ mpage_writepage(struct bio *bio, struct page *page, get_block_t get_block,
 
 alloc_new:
 	if (bio == NULL) {
-		const unsigned __nr_pages = 64;	/* FIXME */
-
 		bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
-					__nr_pages, GFP_NOFS|__GFP_HIGH);
+				bio_get_nr_vecs(bdev), GFP_NOFS|__GFP_HIGH);
 		if (bio == NULL)
 			goto confused;
 	}
@@ -466,10 +474,15 @@ mpage_writepage(struct bio *bio, struct page *page, get_block_t get_block,
 	BUG_ON(PageWriteback(page));
 	SetPageWriteback(page);
 	unlock_page(page);
-	if (boundary || (first_unmapped != blocks_per_page))
+	if (boundary || (first_unmapped != blocks_per_page)) {
 		bio = mpage_bio_submit(WRITE, bio);
-	else
+		if (boundary_block) {
+			write_boundary_block(boundary_bdev,
+					boundary_block, 1 << blkbits);
+		}
+	} else {
 		*last_block_in_bio = blocks[blocks_per_page - 1];
+	}
 	goto out;
 
 confused:

fs/vfat/namei.c

@@ -117,8 +117,6 @@ static int parse_options(char *options,	struct fat_mount_options *opts)
 	savep = NULL;
 	ret = 1;
 	while ((this_char = strsep(&options,",")) != NULL) {
-		if (!*this_char)
-			continue;
 		if ((value = strchr(this_char,'=')) != NULL) {
 			save = *value;
 			savep = value;
@@ -154,8 +152,8 @@ static int parse_options(char *options,	struct fat_mount_options *opts)
 			else
 				ret = 0;
 		}
-		if (this_char != options)
-			*(this_char-1) = ',';
+		if (options)
+			*(options-1) = ',';
 		if (value) {
 			*savep = save;
 		}

include/linux/buffer_head.h

@@ -140,7 +140,6 @@ void end_buffer_io_sync(struct buffer_head *bh, int uptodate);
 void buffer_insert_list(spinlock_t *lock,
 			struct buffer_head *, struct list_head *);
 void mark_buffer_dirty_inode(struct buffer_head *bh, struct inode *inode);
-int write_mapping_buffers(struct address_space *mapping);
 int inode_has_buffers(struct inode *);
 void invalidate_inode_buffers(struct inode *);
 int fsync_buffers_list(spinlock_t *lock, struct list_head *);
@@ -168,6 +167,9 @@ void free_buffer_head(struct buffer_head * bh);
 void FASTCALL(unlock_buffer(struct buffer_head *bh));
 void ll_rw_block(int, int, struct buffer_head * bh[]);
 int submit_bh(int, struct buffer_head *);
+void write_boundary_block(struct block_device *bdev,
+			sector_t bblock, unsigned blocksize);
+
 extern int buffer_heads_over_limit;
 
 /*

include/linux/ext3_fs.h

@@ -17,6 +17,7 @@
 #define _LINUX_EXT3_FS_H
 
 #include <linux/types.h>
+#include <linux/ext3_fs_sb.h>
 
 /*
  * The second extended filesystem constants/structures

include/linux/fs.h

@@ -627,9 +627,6 @@ extern int send_sigurg(struct fown_struct *fown);
 #define MNT_FORCE	0x00000001	/* Attempt to forcibily umount */
 #define MNT_DETACH	0x00000002	/* Just detach from the tree */
 
-#include <linux/ext3_fs_sb.h>
-#include <linux/hpfs_fs_sb.h>
-
 extern struct list_head super_blocks;
 extern spinlock_t sb_lock;
 
@@ -670,8 +667,6 @@ struct super_block {
 	char s_id[32];				/* Informational name */
 
 	union {
-		struct ext3_sb_info	ext3_sb;
-		struct hpfs_sb_info	hpfs_sb;
 		void			*generic_sbp;
 	} u;
 	/*

include/linux/hpfs_fs_sb.h

@@ -36,31 +36,4 @@ struct hpfs_sb_info {
 	int sb_timeshift;
 };
 
-#define s_hpfs_root u.hpfs_sb.sb_root
-#define s_hpfs_fs_size u.hpfs_sb.sb_fs_size
-#define s_hpfs_bitmaps u.hpfs_sb.sb_bitmaps
-#define s_hpfs_dirband_start u.hpfs_sb.sb_dirband_start
-#define s_hpfs_dirband_size u.hpfs_sb.sb_dirband_size
-#define s_hpfs_dmap u.hpfs_sb.sb_dmap
-#define s_hpfs_uid u.hpfs_sb.sb_uid
-#define s_hpfs_gid u.hpfs_sb.sb_gid
-#define s_hpfs_mode u.hpfs_sb.sb_mode
-#define s_hpfs_n_free u.hpfs_sb.sb_n_free
-#define s_hpfs_n_free_dnodes u.hpfs_sb.sb_n_free_dnodes
-#define s_hpfs_lowercase u.hpfs_sb.sb_lowercase
-#define s_hpfs_conv u.hpfs_sb.sb_conv
-#define s_hpfs_eas u.hpfs_sb.sb_eas
-#define s_hpfs_err u.hpfs_sb.sb_err
-#define s_hpfs_chk u.hpfs_sb.sb_chk
-#define s_hpfs_was_error u.hpfs_sb.sb_was_error
-#define s_hpfs_chkdsk u.hpfs_sb.sb_chkdsk
-/*#define s_hpfs_rd_fnode u.hpfs_sb.sb_rd_fnode*/
-#define s_hpfs_rd_inode u.hpfs_sb.sb_rd_inode
-#define s_hpfs_cp_table u.hpfs_sb.sb_cp_table
-#define s_hpfs_bmp_dir u.hpfs_sb.sb_bmp_dir
-#define s_hpfs_c_bitmap u.hpfs_sb.sb_c_bitmap
-#define s_hpfs_iget_q u.hpfs_sb.sb_iget_q
-/*#define s_hpfs_mounting u.hpfs_sb.sb_mounting*/
-#define s_hpfs_timeshift u.hpfs_sb.sb_timeshift
-
 #endif

include/linux/list.h

@@ -86,13 +86,12 @@ static inline void __list_del(struct list_head * prev, struct list_head * next)
 /**
  * list_del - deletes entry from list.
  * @entry: the element to delete from the list.
- * Note: list_empty on entry does not return true after this, the entry is in an undefined state.
+ * Note: list_empty on entry does not return true after this, the entry is
+ * in an undefined state.
  */
 static inline void list_del(struct list_head *entry)
 {
 	__list_del(entry->prev, entry->next);
-	entry->next = (void *) 0;
-	entry->prev = (void *) 0;
 }
 
 /**

include/linux/mm.h

@@ -176,7 +176,7 @@ struct page {
 	 * Architectures with slow multiplication can define
 	 * WANT_PAGE_VIRTUAL in asm/page.h
 	 */
-#if defined(CONFIG_HIGHMEM) || defined(WANT_PAGE_VIRTUAL)
+#if defined(WANT_PAGE_VIRTUAL)
 	void *virtual;			/* Kernel virtual address (NULL if
 					   not kmapped, ie. highmem) */
 #endif /* CONFIG_HIGMEM || WANT_PAGE_VIRTUAL */
@@ -289,38 +289,34 @@ static inline void set_page_zone(struct page *page, unsigned long zone_num)
 	page->flags |= zone_num << ZONE_SHIFT;
 }
 
-/*
- * In order to avoid #ifdefs within C code itself, we define
- * set_page_address to a noop for non-highmem machines, where
- * the field isn't useful.
- * The same is true for page_address() in arch-dependent code.
- */
-#if defined(CONFIG_HIGHMEM) || defined(WANT_PAGE_VIRTUAL)
+#define lowmem_page_address(page)					\
+	__va( ( ((page) - page_zone(page)->zone_mem_map)		\
+			+ page_zone(page)->zone_start_pfn) << PAGE_SHIFT)
+
+#if defined(CONFIG_HIGHMEM) && !defined(WANT_PAGE_VIRTUAL)
+#define HASHED_PAGE_VIRTUAL
+#endif
 
+#if defined(WANT_PAGE_VIRTUAL)
+#define page_address(page) ((page)->virtual)
 #define set_page_address(page, address)			\
 	do {						\
 		(page)->virtual = (address);		\
 	} while(0)
+#define page_address_init()  do { } while(0)
+#endif
 
-#else /* CONFIG_HIGHMEM || WANT_PAGE_VIRTUAL */
-#define set_page_address(page, address)  do { } while(0)
-#endif /* CONFIG_HIGHMEM || WANT_PAGE_VIRTUAL */
-
-/*
- * Permanent address of a page. Obviously must never be
- * called on a highmem page.
- */
-#if defined(CONFIG_HIGHMEM) || defined(WANT_PAGE_VIRTUAL)
-
-#define page_address(page) ((page)->virtual)
-
-#else /* CONFIG_HIGHMEM || WANT_PAGE_VIRTUAL */
-
-#define page_address(page)						\
-	__va( ( ((page) - page_zone(page)->zone_mem_map)		\
-			+ page_zone(page)->zone_start_pfn) << PAGE_SHIFT)
+#if defined(HASHED_PAGE_VIRTUAL)
+void *page_address(struct page *page);
+void set_page_address(struct page *page, void *virtual);
+void page_address_init(void);
+#endif
 
-#endif /* CONFIG_HIGHMEM || WANT_PAGE_VIRTUAL */
+#if !defined(HASHED_PAGE_VIRTUAL) && !defined(WANT_PAGE_VIRTUAL)
+#define page_address(page) lowmem_page_address(page)
+#define set_page_address(page, address)  do { } while(0)
+#define page_address_init()  do { } while(0)
+#endif
 
 /*
  * Return true if this page is mapped into pagetables.  Subtle: test pte.direct

include/linux/mmzone.h

@@ -62,7 +62,6 @@ struct zone {
 	spinlock_t		lock;
 	unsigned long		free_pages;
 	unsigned long		pages_min, pages_low, pages_high;
-	int			need_balance;
 
 	ZONE_PADDING(_pad1_)
 
@@ -120,7 +119,8 @@ struct zone {
 	 * rarely used fields:
 	 */
 	char			*name;
-	unsigned long		size;
+	unsigned long		spanned_pages;	/* total size, including holes */
+	unsigned long		present_pages;	/* amount of memory (excluding holes) */
 } ____cacheline_maxaligned_in_smp;
 
 #define ZONE_DMA		0

include/linux/pci_ids.h

@@ -288,6 +288,12 @@
 #define PCI_DEVICE_ID_NS_87560_USB	0x0012
 #define PCI_DEVICE_ID_NS_83815		0x0020
 #define PCI_DEVICE_ID_NS_83820		0x0022
+#define PCI_DEVICE_ID_NS_SCx200_BRIDGE	0x0500
+#define PCI_DEVICE_ID_NS_SCx200_SMI	0x0501
+#define PCI_DEVICE_ID_NS_SCx200_IDE	0x0502
+#define PCI_DEVICE_ID_NS_SCx200_AUDIO	0x0503
+#define PCI_DEVICE_ID_NS_SCx200_VIDEO	0x0504
+#define PCI_DEVICE_ID_NS_SCx200_XBUS	0x0505
 #define PCI_DEVICE_ID_NS_87410		0xd001
 
 #define PCI_VENDOR_ID_TSENG		0x100c

include/linux/scx200.h

+/* linux/include/linux/scx200.h
+
+   Copyright (c) 2001,2002 Christer Weinigel <wingel@nano-system.com>
+
+   Defines for the National Semiconductor SCx200 Processors
+*/
+
+/* Interesting stuff for the National Semiconductor SCx200 CPU */
+
+/* F0 PCI Header/Bridge Configuration Registers */
+#define SCx200_DOCCS_BASE 0x78	/* DOCCS Base Address Register */
+#define SCx200_DOCCS_CTRL 0x7c	/* DOCCS Control Register */
+
+/* GPIO Register Block */
+#define SCx200_GPIO_SIZE 0x2c	/* Size of GPIO register block */
+
+/* General Configuration Block */
+#define SCx200_CB_BASE 0x9000	/* Base fixed at 0x9000 according to errata */
+
+/* Watchdog Timer */
+#define SCx200_WDT_OFFSET 0x00	/* offset within configuration block */
+#define SCx200_WDT_SIZE 0x05	/* size */
+
+#define SCx200_WDT_WDTO 0x00	/* Time-Out Register */
+#define SCx200_WDT_WDCNFG 0x02	/* Configuration Register */
+#define SCx200_WDT_WDSTS 0x04	/* Status Register */
+#define SCx200_WDT_WDSTS_WDOVF (1<<0) /* Overflow bit */
+
+/* High Resolution Timer */
+#define SCx200_TIMER_OFFSET 0x08
+#define SCx200_TIMER_SIZE 0x05
+
+/* Clock Generators */
+#define SCx200_CLOCKGEN_OFFSET 0x10
+#define SCx200_CLOCKGEN_SIZE 0x10
+
+/* Pin Multiplexing and Miscellaneous Configuration Registers */
+#define SCx200_MISC_OFFSET 0x30
+#define SCx200_MISC_SIZE 0x10
+
+#define SCx200_PMR 0x30		/* Pin Multiplexing Register */
+#define SCx200_MCR 0x34		/* Miscellaneous Configuration Register */
+#define SCx200_INTSEL 0x38	/* Interrupt Selection Register */
+#define SCx200_IID 0x3c		/* IA On a Chip Identification Number Reg */
+#define SCx200_REV 0x3d		/* Revision Register */
+#define SCx200_CBA 0x3e		/* Configuration Base Address Register */
+
+/* Verify that the configuration block really is there */
+#define scx200_cb_probe(base) (inw((base) + SCx200_CBA) == (base))
+
+/*
+    Local variables:
+        compile-command: "make -C ../.. bzImage modules"
+        c-basic-offset: 8
+    End:
+*/

include/linux/scx200_gpio.h

+#include <linux/spinlock.h>
+
+u32 scx200_gpio_configure(int index, u32 set, u32 clear);
+void scx200_gpio_dump(unsigned index);
+
+extern unsigned scx200_gpio_base;
+extern spinlock_t scx200_gpio_lock;
+extern long scx200_gpio_shadow[2];
+
+#define scx200_gpio_present() (scx200_gpio_base!=0)
+
+/* Definitions to make sure I do the same thing in all functions */
+#define __SCx200_GPIO_BANK unsigned bank = index>>5
+#define __SCx200_GPIO_IOADDR unsigned short ioaddr = scx200_gpio_base+0x10*bank
+#define __SCx200_GPIO_SHADOW long *shadow = scx200_gpio_shadow+bank
+#define __SCx200_GPIO_INDEX index &= 31
+
+#define __SCx200_GPIO_OUT __asm__ __volatile__("outsl":"=mS" (shadow):"d" (ioaddr), "0" (shadow))
+
+/* returns the value of the GPIO pin */
+
+static inline int scx200_gpio_get(int index) {
+	__SCx200_GPIO_BANK;
+	__SCx200_GPIO_IOADDR + 0x04;
+	__SCx200_GPIO_INDEX;
+		
+	return (inl(ioaddr) & (1<<index)) ? 1 : 0;
+}
+
+/* return the value driven on the GPIO signal (the value that will be
+   driven if the GPIO is configured as an output, it might not be the
+   state of the GPIO right now if the GPIO is configured as an input) */
+
+static inline int scx200_gpio_current(int index) {
+        __SCx200_GPIO_BANK;
+	__SCx200_GPIO_INDEX;
+		
+	return (scx200_gpio_shadow[bank] & (1<<index)) ? 1 : 0;
+}
+
+/* drive the GPIO signal high */
+
+static inline void scx200_gpio_set_high(int index) {
+	__SCx200_GPIO_BANK;
+	__SCx200_GPIO_IOADDR;
+	__SCx200_GPIO_SHADOW;
+	__SCx200_GPIO_INDEX;
+	set_bit(index, shadow);
+	__SCx200_GPIO_OUT;
+}
+
+/* drive the GPIO signal low */
+
+static inline void scx200_gpio_set_low(int index) {
+	__SCx200_GPIO_BANK;
+	__SCx200_GPIO_IOADDR;
+	__SCx200_GPIO_SHADOW;
+	__SCx200_GPIO_INDEX;
+	clear_bit(index, shadow);
+	__SCx200_GPIO_OUT;
+}
+
+/* drive the GPIO signal to state */
+
+static inline void scx200_gpio_set(int index, int state) {
+	__SCx200_GPIO_BANK;
+	__SCx200_GPIO_IOADDR;
+	__SCx200_GPIO_SHADOW;
+	__SCx200_GPIO_INDEX;
+	if (state)
+		set_bit(index, shadow);
+	else
+		clear_bit(index, shadow);
+	__SCx200_GPIO_OUT;
+}
+
+/* toggle the GPIO signal */
+static inline void scx200_gpio_change(int index) {
+	__SCx200_GPIO_BANK;
+	__SCx200_GPIO_IOADDR;
+	__SCx200_GPIO_SHADOW;
+	__SCx200_GPIO_INDEX;
+	change_bit(index, shadow);
+	__SCx200_GPIO_OUT;
+}
+
+#undef __SCx200_GPIO_BANK
+#undef __SCx200_GPIO_IOADDR
+#undef __SCx200_GPIO_SHADOW
+#undef __SCx200_GPIO_INDEX
+#undef __SCx200_GPIO_OUT
+
+/*
+    Local variables:
+        compile-command: "make -C ../.. bzImage modules"
+        c-basic-offset: 8
+    End:
+*/

init/main.c

@@ -433,6 +433,7 @@ asmlinkage void __init start_kernel(void)
 		initrd_start = 0;
 	}
 #endif
+	page_address_init();
 	mem_init();
 	kmem_cache_sizes_init();
 	pidhash_init();

kernel/ksyms.c

@@ -132,6 +132,9 @@ EXPORT_SYMBOL(highmem_start_page);
 EXPORT_SYMBOL(kmap_prot);
 EXPORT_SYMBOL(kmap_pte);
 #endif
+#ifdef HASHED_PAGE_VIRTUAL
+EXPORT_SYMBOL(page_address);
+#endif
 EXPORT_SYMBOL(get_user_pages);
 
 /* filesystem internal functions */

mm/highmem.c

@@ -22,6 +22,7 @@
 #include <linux/mempool.h>
 #include <linux/blkdev.h>
 #include <linux/init.h>
+#include <linux/hash.h>
 #include <asm/pgalloc.h>
 
 static mempool_t *page_pool, *isa_page_pool;
@@ -88,7 +89,7 @@ static void flush_all_zero_pkmaps(void)
 		page = pte_page(pkmap_page_table[i]);
 		pte_clear(&pkmap_page_table[i]);
 
-		page->virtual = NULL;
+		set_page_address(page, NULL);
 	}
 	flush_tlb_kernel_range(PKMAP_ADDR(0), PKMAP_ADDR(LAST_PKMAP));
 }
@@ -126,8 +127,8 @@ static inline unsigned long map_new_virtual(struct page *page)
 			spin_lock(&kmap_lock);
 
 			/* Somebody else might have mapped it while we slept */
-			if (page->virtual)
-				return (unsigned long) page->virtual;
+			if (page_address(page))
+				return (unsigned long)page_address(page);
 
 			/* Re-start */
 			goto start;
@@ -137,7 +138,7 @@ static inline unsigned long map_new_virtual(struct page *page)
 	set_pte(&(pkmap_page_table[last_pkmap_nr]), mk_pte(page, kmap_prot));
 
 	pkmap_count[last_pkmap_nr] = 1;
-	page->virtual = (void *) vaddr;
+	set_page_address(page, (void *)vaddr);
 
 	return vaddr;
 }
@@ -153,7 +154,7 @@ void *kmap_high(struct page *page)
 	 * We cannot call this from interrupts, as it may block
 	 */
 	spin_lock(&kmap_lock);
-	vaddr = (unsigned long) page->virtual;
+	vaddr = (unsigned long)page_address(page);
 	if (!vaddr)
 		vaddr = map_new_virtual(page);
 	pkmap_count[PKMAP_NR(vaddr)]++;
@@ -170,7 +171,7 @@ void kunmap_high(struct page *page)
 	int need_wakeup;
 
 	spin_lock(&kmap_lock);
-	vaddr = (unsigned long) page->virtual;
+	vaddr = (unsigned long)page_address(page);
 	if (!vaddr)
 		BUG();
 	nr = PKMAP_NR(vaddr);
@@ -467,7 +468,7 @@ void blk_queue_bounce(request_queue_t *q, struct bio **bio_orig)
 	*bio_orig = bio;
 }
 
-#if CONFIG_DEBUG_HIGHMEM
+#if defined(CONFIG_DEBUG_HIGHMEM) && defined(CONFIG_HIGHMEM)
 void check_highmem_ptes(void)
 {
 	int idx, type;
@@ -484,3 +485,121 @@ void check_highmem_ptes(void)
 }
 #endif
 
+#if defined(HASHED_PAGE_VIRTUAL)
+
+#define PA_HASH_ORDER	7
+
+/*
+ * Describes one page->virtual association
+ */
+struct page_address_map {
+	struct page *page;
+	void *virtual;
+	struct list_head list;
+};
+
+/*
+ * page_address_map freelist, allocated from page_address_maps.
+ */
+static struct list_head page_address_pool;	/* freelist */
+static spinlock_t pool_lock;			/* protects page_address_pool */
+
+/*
+ * Hash table bucket
+ */
+static struct page_address_slot {
+	struct list_head lh;			/* List of page_address_maps */
+	spinlock_t lock;			/* Protect this bucket's list */
+} ____cacheline_aligned_in_smp page_address_htable[1<<PA_HASH_ORDER];
+
+static struct page_address_slot *page_slot(struct page *page)
+{
+	return &page_address_htable[hash_ptr(page, PA_HASH_ORDER)];
+}
+
+void *page_address(struct page *page)
+{
+	unsigned long flags;
+	void *ret;
+	struct page_address_slot *pas;
+
+	if (!PageHighMem(page))
+		return lowmem_page_address(page);
+
+	pas = page_slot(page);
+	ret = NULL;
+	spin_lock_irqsave(&pas->lock, flags);
+	if (!list_empty(&pas->lh)) {
+		struct page_address_map *pam;
+
+		list_for_each_entry(pam, &pas->lh, list) {
+			if (pam->page == page) {
+				ret = pam->virtual;
+				goto done;
+			}
+		}
+	}
+done:
+	spin_unlock_irqrestore(&pas->lock, flags);
+	return ret;
+}
+
+void set_page_address(struct page *page, void *virtual)
+{
+	unsigned long flags;
+	struct page_address_slot *pas;
+	struct page_address_map *pam;
+
+	BUG_ON(!PageHighMem(page));
+
+	pas = page_slot(page);
+	if (virtual) {		/* Add */
+		BUG_ON(list_empty(&page_address_pool));
+
+		spin_lock_irqsave(&pool_lock, flags);
+		pam = list_entry(page_address_pool.next,
+				struct page_address_map, list);
+		list_del(&pam->list);
+		spin_unlock_irqrestore(&pool_lock, flags);
+
+		pam->page = page;
+		pam->virtual = virtual;
+
+		spin_lock_irqsave(&pas->lock, flags);
+		list_add_tail(&pam->list, &pas->lh);
+		spin_unlock_irqrestore(&pas->lock, flags);
+	} else {		/* Remove */
+		spin_lock_irqsave(&pas->lock, flags);
+		list_for_each_entry(pam, &pas->lh, list) {
+			if (pam->page == page) {
+				list_del(&pam->list);
+				spin_unlock_irqrestore(&pas->lock, flags);
+				spin_lock_irqsave(&pool_lock, flags);
+				list_add_tail(&pam->list, &page_address_pool);
+				spin_unlock_irqrestore(&pool_lock, flags);
+				goto done;
+			}
+		}
+		spin_unlock_irqrestore(&pas->lock, flags);
+	}
+done:
+	return;
+}
+
+static struct page_address_map page_address_maps[LAST_PKMAP];
+
+void __init page_address_init(void)
+{
+	int i;
+
+	INIT_LIST_HEAD(&page_address_pool);
+	for (i = 0; i < ARRAY_SIZE(page_address_maps); i++)
+		list_add(&page_address_maps[i].list, &page_address_pool);
+	for (i = 0; i < ARRAY_SIZE(page_address_htable); i++) {
+		INIT_LIST_HEAD(&page_address_htable[i].lh);
+		spin_lock_init(&page_address_htable[i].lock);
+	}
+	spin_lock_init(&pool_lock);
+}
+
+#endif	/* defined(CONFIG_HIGHMEM) && !defined(WANT_PAGE_VIRTUAL) */

mm/mempool.c

@@ -12,6 +12,7 @@
 #include <linux/slab.h>
 #include <linux/module.h>
 #include <linux/mempool.h>
+#include <linux/blkdev.h>
 #include <linux/writeback.h>
 
 static void add_element(mempool_t *pool, void *element)
@@ -184,8 +185,7 @@ void * mempool_alloc(mempool_t *pool, int gfp_mask)
 {
 	void *element;
 	unsigned long flags;
-	int curr_nr;
-	DECLARE_WAITQUEUE(wait, current);
+	DEFINE_WAIT(wait);
 	int gfp_nowait = gfp_mask & ~(__GFP_WAIT | __GFP_IO);
 	int pf_flags = current->flags;
 
@@ -226,18 +226,10 @@ void * mempool_alloc(mempool_t *pool, int gfp_mask)
 
 	blk_run_queues();
 
-	add_wait_queue_exclusive(&pool->wait, &wait);
-	set_task_state(current, TASK_UNINTERRUPTIBLE);
-
-	spin_lock_irqsave(&pool->lock, flags);
-	curr_nr = pool->curr_nr;
-	spin_unlock_irqrestore(&pool->lock, flags);
-
-	if (!curr_nr)
-		schedule();
-
-	current->state = TASK_RUNNING;
-	remove_wait_queue(&pool->wait, &wait);
+	prepare_to_wait(&pool->wait, &wait, TASK_UNINTERRUPTIBLE);
+	if (!pool->curr_nr)
+		io_schedule();
+	finish_wait(&pool->wait, &wait);
 
 	goto repeat_alloc;
 }

mm/page-writeback.c

@@ -41,10 +41,8 @@
  */
 static long ratelimit_pages = 32;
 
-/*
- * The total number of pages in the machine.
- */
-static long total_pages;
+static long total_pages;	/* The total number of pages in the machine. */
+static int dirty_exceeded;	/* Dirty mem may be over limit */
 
 /*
  * When balance_dirty_pages decides that the caller needs to perform some
@@ -60,16 +58,12 @@ static inline long sync_writeback_pages(void)
 /* The following parameters are exported via /proc/sys/vm */
 
 /*
- * Dirty memory thresholds, in percentages
- */
-
-/*
- * Start background writeback (via pdflush) at this level
+ * Start background writeback (via pdflush) at this percentage
  */
 int dirty_background_ratio = 10;
 
 /*
- * The generator of dirty data starts async writeback at this level
+ * The generator of dirty data starts async writeback at this percentage
  */
 int dirty_async_ratio = 40;
 
@@ -80,7 +74,7 @@ int dirty_async_ratio = 40;
 int dirty_writeback_centisecs = 5 * 100;
 
 /*
- * The longest amount of time for which data is allowed to remain dirty
+ * The longest number of centiseconds for which data is allowed to remain dirty
  */
 int dirty_expire_centisecs = 30 * 100;
 
@@ -90,22 +84,17 @@ int dirty_expire_centisecs = 30 * 100;
 static void background_writeout(unsigned long _min_pages);
 
 /*
- * balance_dirty_pages() must be called by processes which are
- * generating dirty data.  It looks at the number of dirty pages
- * in the machine and either:
- *
- * - Starts background writeback or
- * - Causes the caller to perform async writeback or
- * - Causes the caller to perform synchronous writeback, then
- *   tells a pdflush thread to perform more writeback or
- * - Does nothing at all.
- *
- * balance_dirty_pages() can sleep.
+ * balance_dirty_pages() must be called by processes which are generating dirty
+ * data.  It looks at the number of dirty pages in the machine and will force
+ * the caller to perform writeback if the system is over `async_thresh'.
+ * If we're over `background_thresh' then pdflush is woken to perform some
+ * writeout.
  */
 void balance_dirty_pages(struct address_space *mapping)
 {
 	struct page_state ps;
-	long background_thresh, async_thresh;
+	long background_thresh;
+	long async_thresh;
 	unsigned long dirty_and_writeback;
 	struct backing_dev_info *bdi;
 
@@ -123,9 +112,13 @@ void balance_dirty_pages(struct address_space *mapping)
 			.older_than_this = NULL,
 			.nr_to_write	= sync_writeback_pages(),
 		};
-
+		if (!dirty_exceeded)
+			dirty_exceeded = 1;
 		writeback_inodes(&wbc);
 		get_page_state(&ps);
+	} else {
+		if (dirty_exceeded)
+			dirty_exceeded = 0;
 	}
 
 	if (!writeback_in_progress(bdi) && ps.nr_dirty > background_thresh)
@@ -141,17 +134,25 @@ EXPORT_SYMBOL_GPL(balance_dirty_pages);
  * which was newly dirtied.  The function will periodically check the system's
  * dirty state and will initiate writeback if needed.
  *
- * balance_dirty_pages_ratelimited() may sleep.
+ * On really big machines, get_page_state is expensive, so try to avoid calling
+ * it too often (ratelimiting).  But once we're over the dirty memory limit we
+ * decrease the ratelimiting by a lot, to prevent individual processes from
+ * overshooting the limit by (ratelimit_pages) each.
  */
 void balance_dirty_pages_ratelimited(struct address_space *mapping)
 {
 	static struct rate_limit_struct {
 		int count;
-	} ____cacheline_aligned ratelimits[NR_CPUS];
+	} ____cacheline_aligned_in_smp ratelimits[NR_CPUS];
 	int cpu;
+	long ratelimit;
+
+	ratelimit = ratelimit_pages;
+	if (dirty_exceeded)
+		ratelimit = 8;
 
 	cpu = get_cpu();
-	if (ratelimits[cpu].count++ >= ratelimit_pages) {
+	if (ratelimits[cpu].count++ >= ratelimit) {
 		ratelimits[cpu].count = 0;
 		put_cpu();
 		balance_dirty_pages(mapping);

mm/page_alloc.c

@@ -48,7 +48,7 @@ static int zone_balance_max[MAX_NR_ZONES] __initdata = { 255 , 255, 255, };
  */
 static inline int bad_range(struct zone *zone, struct page *page)
 {
-	if (page_to_pfn(page) >= zone->zone_start_pfn + zone->size)
+	if (page_to_pfn(page) >= zone->zone_start_pfn + zone->spanned_pages)
 		return 1;
 	if (page_to_pfn(page) < zone->zone_start_pfn)
 		return 1;
@@ -346,8 +346,6 @@ __alloc_pages(unsigned int gfp_mask, unsigned int order,
 		}
 	}
 
-	classzone->need_balance = 1;
-	mb();
 	/* we're somewhat low on memory, failed to find what we needed */
 	for (i = 0; zones[i] != NULL; i++) {
 		struct zone *z = zones[i];
@@ -509,7 +507,7 @@ static unsigned int nr_free_zone_pages(int offset)
 		struct zone *zone;
 
 		for (zone = *zonep++; zone; zone = *zonep++) {
-			unsigned long size = zone->size;
+			unsigned long size = zone->present_pages;
 			unsigned long high = zone->pages_high;
 			if (size > high)
 				sum += size - high;
@@ -681,7 +679,7 @@ void show_free_areas(void)
 			struct zone *zone = &pgdat->node_zones[type];
  			unsigned long nr, flags, order, total = 0;
 
-			if (!zone->size)
+			if (!zone->present_pages)
 				continue;
 
 			spin_lock_irqsave(&zone->lock, flags);
@@ -710,7 +708,7 @@ static int __init build_zonelists_node(pg_data_t *pgdat, struct zonelist *zoneli
 		BUG();
 	case ZONE_HIGHMEM:
 		zone = pgdat->node_zones + ZONE_HIGHMEM;
-		if (zone->size) {
+		if (zone->present_pages) {
 #ifndef CONFIG_HIGHMEM
 			BUG();
 #endif
@@ -718,11 +716,11 @@ static int __init build_zonelists_node(pg_data_t *pgdat, struct zonelist *zoneli
 		}
 	case ZONE_NORMAL:
 		zone = pgdat->node_zones + ZONE_NORMAL;
-		if (zone->size)
+		if (zone->present_pages)
 			zonelist->zones[j++] = zone;
 	case ZONE_DMA:
 		zone = pgdat->node_zones + ZONE_DMA;
-		if (zone->size)
+		if (zone->present_pages)
 			zonelist->zones[j++] = zone;
 	}
 
@@ -866,13 +864,13 @@ void __init free_area_init_core(pg_data_t *pgdat,
 			realsize -= zholes_size[j];
 
 		printk("  %s zone: %lu pages\n", zone_names[j], realsize);
-		zone->size = size;
+		zone->spanned_pages = size;
+		zone->present_pages = realsize;
 		zone->name = zone_names[j];
 		spin_lock_init(&zone->lock);
 		spin_lock_init(&zone->lru_lock);
 		zone->zone_pgdat = pgdat;
 		zone->free_pages = 0;
-		zone->need_balance = 0;
 		INIT_LIST_HEAD(&zone->active_list);
 		INIT_LIST_HEAD(&zone->inactive_list);
 		atomic_set(&zone->refill_counter, 0);
@@ -923,12 +921,15 @@ void __init free_area_init_core(pg_data_t *pgdat,
 			set_page_count(page, 0);
 			SetPageReserved(page);
 			INIT_LIST_HEAD(&page->list);
+#ifdef WANT_PAGE_VIRTUAL
 			if (j != ZONE_HIGHMEM)
 				/*
 				 * The shift left won't overflow because the
 				 * ZONE_NORMAL is below 4G.
 				 */
-				set_page_address(page, __va(zone_start_pfn << PAGE_SHIFT));
+				set_page_address(page,
+					__va(zone_start_pfn << PAGE_SHIFT));
+#endif
 			zone_start_pfn++;
 		}
 
@@ -1034,7 +1035,7 @@ static int frag_show(struct seq_file *m, void *arg)
 	int order;
 
 	for (zone = node_zones; zone - node_zones < MAX_NR_ZONES; ++zone) {
-		if (!zone->size)
+		if (!zone->present_pages)
 			continue;
 
 		spin_lock_irqsave(&zone->lock, flags);

mm/truncate.c

@@ -35,9 +35,18 @@ static inline void truncate_partial_page(struct page *page, unsigned partial)
  * If truncate cannot remove the fs-private metadata from the page, the page
  * becomes anonymous.  It will be left on the LRU and may even be mapped into
  * user pagetables if we're racing with filemap_nopage().
+ *
+ * We need to bale out if page->mapping is no longer equal to the original
+ * mapping.  This happens a) when the VM reclaimed the page while we waited on
+ * its lock, b) when a concurrent invalidate_inode_pages got there first and
+ * c) when tmpfs swizzles a page between a tmpfs inode and swapper_space.
  */
-static void truncate_complete_page(struct page *page)
+static void
+truncate_complete_page(struct address_space *mapping, struct page *page)
 {
+	if (page->mapping != mapping)
+		return;
+
 	if (PagePrivate(page))
 		do_invalidatepage(page, 0);
 
@@ -61,6 +70,9 @@ static void truncate_complete_page(struct page *page)
  * The first pass will remove most pages, so the search cost of the second pass
  * is low.
  *
+ * When looking at page->index outside the page lock we need to be careful to
+ * copy it into a local to avoid races (it could change at any time).
+ *
  * Called under (and serialised by) inode->i_sem.
  */
 void truncate_inode_pages(struct address_space *mapping, loff_t lstart)
@@ -76,15 +88,18 @@ void truncate_inode_pages(struct address_space *mapping, loff_t lstart)
 	while (pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
 		for (i = 0; i < pagevec_count(&pvec); i++) {
 			struct page *page = pvec.pages[i];
+			pgoff_t page_index = page->index;
 
-			next = page->index + 1;
+			if (page_index > next)
+				next = page_index;
+			next++;
 			if (TestSetPageLocked(page))
 				continue;
 			if (PageWriteback(page)) {
 				unlock_page(page);
 				continue;
 			}
-			truncate_complete_page(page);
+			truncate_complete_page(mapping, page);
 			unlock_page(page);
 		}
 		pagevec_release(&pvec);
@@ -114,8 +129,10 @@ void truncate_inode_pages(struct address_space *mapping, loff_t lstart)
 
 			lock_page(page);
 			wait_on_page_writeback(page);
-			next = page->index + 1;
-			truncate_complete_page(page);
+			if (page->index > next)
+				next = page->index;
+			next++;
+			truncate_complete_page(mapping, page);
 			unlock_page(page);
 		}
 		pagevec_release(&pvec);
@@ -150,14 +167,16 @@ void invalidate_inode_pages(struct address_space *mapping)
 				next++;
 				continue;
 			}
-			next = page->index + 1;
+			if (page->index > next)
+				next = page->index;
+			next++;
 			if (PageDirty(page) || PageWriteback(page))
 				goto unlock;
 			if (PagePrivate(page) && !try_to_release_page(page, 0))
 				goto unlock;
 			if (page_mapped(page))
 				goto unlock;
-			truncate_complete_page(page);
+			truncate_complete_page(mapping, page);
 unlock:
 			unlock_page(page);
 		}
@@ -183,18 +202,18 @@ void invalidate_inode_pages2(struct address_space *mapping)
 	int i;
 
 	pagevec_init(&pvec);
-	while (!pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
+	while (pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
 		for (i = 0; i < pagevec_count(&pvec); i++) {
 			struct page *page = pvec.pages[i];
 
 			lock_page(page);
-			if (page->mapping) {	/* truncate race? */
+			if (page->mapping == mapping) {	/* truncate race? */
 				wait_on_page_writeback(page);
 				next = page->index + 1;
 				if (page_mapped(page))
 					clear_page_dirty(page);
 				else
-					truncate_complete_page(page);
+					truncate_complete_page(mapping, page);
 			}
 			unlock_page(page);
 		}

mm/vmscan.c

@@ -101,15 +101,19 @@ static inline int is_page_cache_freeable(struct page *page)
 	return page_count(page) - !!PagePrivate(page) == 2;
 }
 
+
+/*
+ * shrink_list returns the number of reclaimed pages
+ */
 static /* inline */ int
-shrink_list(struct list_head *page_list, int nr_pages,
-		unsigned int gfp_mask, int *max_scan, int *nr_mapped)
+shrink_list(struct list_head *page_list, unsigned int gfp_mask,
+		int *max_scan, int *nr_mapped)
 {
 	struct address_space *mapping;
 	LIST_HEAD(ret_pages);
 	struct pagevec freed_pvec;
-	const int nr_pages_in = nr_pages;
 	int pgactivate = 0;
+	int ret = 0;
 
 	pagevec_init(&freed_pvec);
 	while (!list_empty(page_list)) {
@@ -295,7 +299,7 @@ shrink_list(struct list_head *page_list, int nr_pages,
 		__put_page(page);	/* The pagecache ref */
 free_it:
 		unlock_page(page);
-		nr_pages--;
+		ret++;
 		if (!pagevec_add(&freed_pvec, page))
 			__pagevec_release_nonlru(&freed_pvec);
 		continue;
@@ -312,11 +316,11 @@ shrink_list(struct list_head *page_list, int nr_pages,
 	list_splice(&ret_pages, page_list);
 	if (pagevec_count(&freed_pvec))
 		__pagevec_release_nonlru(&freed_pvec);
-	mod_page_state(pgsteal, nr_pages_in - nr_pages);
+	mod_page_state(pgsteal, ret);
 	if (current->flags & PF_KSWAPD)
-		mod_page_state(kswapd_steal, nr_pages_in - nr_pages);
+		mod_page_state(kswapd_steal, ret);
 	mod_page_state(pgactivate, pgactivate);
-	return nr_pages;
+	return ret;
 }
 
 /*
@@ -325,18 +329,19 @@ shrink_list(struct list_head *page_list, int nr_pages,
  * not freed will be added back to the LRU.
  *
  * shrink_cache() is passed the number of pages to try to free, and returns
- * the number which are yet-to-free.
+ * the number of pages which were reclaimed.
  *
  * For pagecache intensive workloads, the first loop here is the hottest spot
  * in the kernel (apart from the copy_*_user functions).
  */
 static /* inline */ int
-shrink_cache(int nr_pages, struct zone *zone,
+shrink_cache(const int nr_pages, struct zone *zone,
 		unsigned int gfp_mask, int max_scan, int *nr_mapped)
 {
 	LIST_HEAD(page_list);
 	struct pagevec pvec;
 	int nr_to_process;
+	int ret = 0;
 
 	/*
 	 * Try to ensure that we free `nr_pages' pages in one pass of the loop.
@@ -349,10 +354,11 @@ shrink_cache(int nr_pages, struct zone *zone,
 
 	lru_add_drain();
 	spin_lock_irq(&zone->lru_lock);
-	while (max_scan > 0 && nr_pages > 0) {
+	while (max_scan > 0 && ret < nr_pages) {
 		struct page *page;
 		int nr_taken = 0;
 		int nr_scan = 0;
+		int nr_freed;
 
 		while (nr_scan++ < nr_to_process &&
 				!list_empty(&zone->inactive_list)) {
@@ -383,10 +389,10 @@ shrink_cache(int nr_pages, struct zone *zone,
 
 		max_scan -= nr_scan;
 		mod_page_state(pgscan, nr_scan);
-		nr_pages = shrink_list(&page_list, nr_pages,
-				gfp_mask, &max_scan, nr_mapped);
-
-		if (nr_pages <= 0 && list_empty(&page_list))
+		nr_freed = shrink_list(&page_list, gfp_mask,
+					&max_scan, nr_mapped);
+		ret += nr_freed;
+		if (nr_freed <= 0 && list_empty(&page_list))
 			goto done;
 
 		spin_lock_irq(&zone->lru_lock);
@@ -412,7 +418,7 @@ shrink_cache(int nr_pages, struct zone *zone,
 	spin_unlock_irq(&zone->lru_lock);
 done:
 	pagevec_release(&pvec);
-	return nr_pages;	
+	return ret;
 }
 
 /*
@@ -533,9 +539,14 @@ refill_inactive_zone(struct zone *zone, const int nr_pages_in)
 	mod_page_state(pgdeactivate, pgdeactivate);
 }
 
+/*
+ * Try to reclaim `nr_pages' from this zone.  Returns the number of reclaimed
+ * pages.  This is a basic per-zone page freer.  Used by both kswapd and
+ * direct reclaim.
+ */
 static /* inline */ int
-shrink_zone(struct zone *zone, int max_scan,
-		unsigned int gfp_mask, int nr_pages, int *nr_mapped)
+shrink_zone(struct zone *zone, int max_scan, unsigned int gfp_mask,
+		const int nr_pages, int *nr_mapped)
 {
 	unsigned long ratio;
 
@@ -556,36 +567,60 @@ shrink_zone(struct zone *zone, int max_scan,
 		atomic_sub(SWAP_CLUSTER_MAX, &zone->refill_counter);
 		refill_inactive_zone(zone, SWAP_CLUSTER_MAX);
 	}
-	nr_pages = shrink_cache(nr_pages, zone, gfp_mask,
-				max_scan, nr_mapped);
-	return nr_pages;
+	return shrink_cache(nr_pages, zone, gfp_mask, max_scan, nr_mapped);
+}
+
+/*
+ * FIXME: don't do this for ZONE_HIGHMEM
+ */
+/*
+ * Here we assume it costs one seek to replace a lru page and that it also
+ * takes a seek to recreate a cache object.  With this in mind we age equal
+ * percentages of the lru and ageable caches.  This should balance the seeks
+ * generated by these structures.
+ *
+ * NOTE: for now I do this for all zones.  If we find this is too aggressive
+ * on large boxes we may want to exclude ZONE_HIGHMEM.
+ *
+ * If we're encountering mapped pages on the LRU then increase the pressure on
+ * slab to avoid swapping.
+ */
+static void shrink_slab(int total_scanned, int gfp_mask)
+{
+	int shrink_ratio;
+	int pages = nr_used_zone_pages();
+
+	shrink_ratio = (pages / (total_scanned + 1)) + 1;
+	shrink_dcache_memory(shrink_ratio, gfp_mask);
+	shrink_icache_memory(shrink_ratio, gfp_mask);
+	shrink_dqcache_memory(shrink_ratio, gfp_mask);
 }
 
+/*
+ * This is the direct reclaim path, for page-allocating processes.  We only
+ * try to reclaim pages from zones which will satisfy the caller's allocation
+ * request.
+ */
 static int
-shrink_caches(struct zone *classzone, int priority,
-		int *total_scanned, int gfp_mask, int nr_pages)
+shrink_caches(struct zone *classzone, int priority, int *total_scanned,
+		int gfp_mask, const int nr_pages, int order)
 {
 	struct zone *first_classzone;
 	struct zone *zone;
-	int ratio;
 	int nr_mapped = 0;
-	int pages = nr_used_zone_pages();
+	int ret = 0;
 
 	first_classzone = classzone->zone_pgdat->node_zones;
 	for (zone = classzone; zone >= first_classzone; zone--) {
 		int max_scan;
 		int to_reclaim;
-		int unreclaimed;
 
 		to_reclaim = zone->pages_high - zone->free_pages;
-		if (to_reclaim < 0)
+		if (order == 0 && to_reclaim < 0)
 			continue;	/* zone has enough memory */
 
-		if (to_reclaim > SWAP_CLUSTER_MAX)
-			to_reclaim = SWAP_CLUSTER_MAX;
-
-		if (to_reclaim < nr_pages)
-			to_reclaim = nr_pages;
+		to_reclaim = min(to_reclaim, SWAP_CLUSTER_MAX);
+		to_reclaim = max(to_reclaim, nr_pages);
 
 		/*
 		 * If we cannot reclaim `nr_pages' pages by scanning twice
@@ -594,33 +629,18 @@ shrink_caches(struct zone *classzone, int priority,
 		max_scan = zone->nr_inactive >> priority;
 		if (max_scan < to_reclaim * 2)
 			max_scan = to_reclaim * 2;
-		unreclaimed = shrink_zone(zone, max_scan,
-				gfp_mask, to_reclaim, &nr_mapped);
-		nr_pages -= to_reclaim - unreclaimed;
+		ret += shrink_zone(zone, max_scan, gfp_mask,
+				to_reclaim, &nr_mapped);
 		*total_scanned += max_scan;
+		*total_scanned += nr_mapped;
+		if (ret >= nr_pages)
+			break;
 	}
-
-	/*
-	 * Here we assume it costs one seek to replace a lru page and that
-	 * it also takes a seek to recreate a cache object.  With this in
-	 * mind we age equal percentages of the lru and ageable caches.
-	 * This should balance the seeks generated by these structures.
-	 *
-	 * NOTE: for now I do this for all zones.  If we find this is too
-	 * aggressive on large boxes we may want to exclude ZONE_HIGHMEM
-	 *
-	 * If we're encountering mapped pages on the LRU then increase the
-	 * pressure on slab to avoid swapping.
-	 */
-	ratio = (pages / (*total_scanned + nr_mapped + 1)) + 1;
-	shrink_dcache_memory(ratio, gfp_mask);
-	shrink_icache_memory(ratio, gfp_mask);
-	shrink_dqcache_memory(ratio, gfp_mask);
-	return nr_pages;
+	return ret;
 }
 
 /*
- * This is the main entry point to page reclaim.
+ * This is the main entry point to direct page reclaim.
  *
  * If a full scan of the inactive list fails to free enough memory then we
  * are "out of memory" and something needs to be killed.
@@ -640,17 +660,19 @@ int
 try_to_free_pages(struct zone *classzone,
 		unsigned int gfp_mask, unsigned int order)
 {
-	int priority = DEF_PRIORITY;
-	int nr_pages = SWAP_CLUSTER_MAX;
+	int priority;
+	const int nr_pages = SWAP_CLUSTER_MAX;
+	int nr_reclaimed = 0;
 
 	inc_page_state(pageoutrun);
 
 	for (priority = DEF_PRIORITY; priority; priority--) {
 		int total_scanned = 0;
 
-		nr_pages = shrink_caches(classzone, priority, &total_scanned,
-					gfp_mask, nr_pages);
-		if (nr_pages <= 0)
+		nr_reclaimed += shrink_caches(classzone, priority,
+					&total_scanned, gfp_mask,
+					nr_pages, order);
+		if (nr_reclaimed >= nr_pages)
 			return 1;
 		if (total_scanned == 0)
 			return 1;	/* All zones had enough free memory */
@@ -665,62 +687,46 @@ try_to_free_pages(struct zone *classzone,
 
 		/* Take a nap, wait for some writeback to complete */
 		blk_congestion_wait(WRITE, HZ/4);
+		shrink_slab(total_scanned, gfp_mask);
 	}
 	if (gfp_mask & __GFP_FS)
 		out_of_memory();
 	return 0;
 }
 
-static int check_classzone_need_balance(struct zone *classzone)
+/*
+ * kswapd will work across all this node's zones until they are all at
+ * pages_high.
+ */
+static void kswapd_balance_pgdat(pg_data_t *pgdat)
 {
-	struct zone *first_classzone;
+	int priority = DEF_PRIORITY;
+	int i;
 
-	first_classzone = classzone->zone_pgdat->node_zones;
-	while (classzone >= first_classzone) {
-		if (classzone->free_pages > classzone->pages_high)
-			return 0;
-		classzone--;
-	}
-	return 1;
-}
+	for (priority = DEF_PRIORITY; priority; priority--) {
+		int success = 1;
 
-static int kswapd_balance_pgdat(pg_data_t * pgdat)
-{
-	int need_more_balance = 0, i;
-	struct zone *zone;
+		for (i = 0; i < pgdat->nr_zones; i++) {
+			struct zone *zone = pgdat->node_zones + i;
+			int nr_mapped = 0;
+			int max_scan;
+			int to_reclaim;
 
-	for (i = pgdat->nr_zones-1; i >= 0; i--) {
-		zone = pgdat->node_zones + i;
-		cond_resched();
-		if (!zone->need_balance)
-			continue;
-		if (!try_to_free_pages(zone, GFP_KSWAPD, 0)) {
-			zone->need_balance = 0;
-			__set_current_state(TASK_INTERRUPTIBLE);
-			schedule_timeout(HZ);
-			continue;
+			to_reclaim = zone->pages_high - zone->free_pages;
+			if (to_reclaim <= 0)
+				continue;
+			success = 0;
+			max_scan = zone->nr_inactive >> priority;
+			if (max_scan < to_reclaim * 2)
+				max_scan = to_reclaim * 2;
+			shrink_zone(zone, max_scan, GFP_KSWAPD,
+					to_reclaim, &nr_mapped);
+			shrink_slab(max_scan + nr_mapped, GFP_KSWAPD);
 		}
-		if (check_classzone_need_balance(zone))
-			need_more_balance = 1;
-		else
-			zone->need_balance = 0;
-	}
-
-	return need_more_balance;
-}
-
-static int kswapd_can_sleep_pgdat(pg_data_t * pgdat)
-{
-	struct zone *zone;
-	int i;
-
-	for (i = pgdat->nr_zones-1; i >= 0; i--) {
-		zone = pgdat->node_zones + i;
-		if (zone->need_balance)
-			return 0;
+		if (success)
+			break;	/* All zones are at pages_high */
+		blk_congestion_wait(WRITE, HZ/4);
 	}
-
-	return 1;
 }
 
 /*
@@ -740,7 +746,7 @@ int kswapd(void *p)
 {
 	pg_data_t *pgdat = (pg_data_t*)p;
 	struct task_struct *tsk = current;
-	DECLARE_WAITQUEUE(wait, tsk);
+	DEFINE_WAIT(wait);
 
 	daemonize();
 	set_cpus_allowed(tsk, __node_to_cpu_mask(pgdat->node_id));
@@ -761,27 +767,12 @@ int kswapd(void *p)
 	 */
 	tsk->flags |= PF_MEMALLOC|PF_KSWAPD;
 
-	/*
-	 * Kswapd main loop.
-	 */
-	for (;;) {
+	for ( ; ; ) {
 		if (current->flags & PF_FREEZE)
 			refrigerator(PF_IOTHREAD);
-		__set_current_state(TASK_INTERRUPTIBLE);
-		add_wait_queue(&pgdat->kswapd_wait, &wait);
-
-		mb();
-		if (kswapd_can_sleep_pgdat(pgdat))
-			schedule();
-
-		__set_current_state(TASK_RUNNING);
-		remove_wait_queue(&pgdat->kswapd_wait, &wait);
-
-		/*
-		 * If we actually get into a low-memory situation,
-		 * the processes needing more memory will wake us
-		 * up on a more timely basis.
-		 */
+		prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
+		schedule();
+		finish_wait(&pgdat->kswapd_wait, &wait);
 		kswapd_balance_pgdat(pgdat);
 		blk_run_queues();
 	}