Index: linux-2.6.9-rc4-ck3/arch/i386/kernel/cpu/mtrr/generic.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/arch/i386/kernel/cpu/mtrr/generic.c	2004-08-15 14:08:04.000000000 +1000
+++ linux-2.6.9-rc4-ck3/arch/i386/kernel/cpu/mtrr/generic.c	2004-10-16 23:38:18.908003558 +1000
@@ -240,11 +240,14 @@ static void prepare_set(void)
 	/*  Note that this is not ideal, since the cache is only flushed/disabled
 	   for this CPU while the MTRRs are changed, but changing this requires
 	   more invasive changes to the way the kernel boots  */
-	spin_lock(&set_atomicity_lock);
+	/*
+	 * Since we are disabling the cache dont allow any interrupts - they
+	 * would run extremely slow and would only increase the pain:
+	 */
+	spin_lock_irq(&set_atomicity_lock);
 
 	/*  Enter the no-fill (CD=1, NW=0) cache mode and flush caches. */
 	cr0 = read_cr0() | 0x40000000;	/* set CD flag */
-	wbinvd();
 	write_cr0(cr0);
 	wbinvd();
 
@@ -266,8 +269,7 @@ static void prepare_set(void)
 
 static void post_set(void)
 {
-	/*  Flush caches and TLBs  */
-	wbinvd();
+	/*  Flush TLBs (no need to flush caches - they are disabled)  */
 	__flush_tlb();
 
 	/* Intel (P6) standard MTRRs */
@@ -279,7 +281,7 @@ static void post_set(void)
 	/*  Restore value of CR4  */
 	if ( cpu_has_pge )
 		write_cr4(cr4);
-	spin_unlock(&set_atomicity_lock);
+	spin_unlock_irq(&set_atomicity_lock);
 }
 
 static void generic_set_all(void)
Index: linux-2.6.9-rc4-ck3/arch/i386/kernel/entry.S
===================================================================
--- linux-2.6.9-rc4-ck3.orig/arch/i386/kernel/entry.S	2004-10-11 16:15:58.000000000 +1000
+++ linux-2.6.9-rc4-ck3/arch/i386/kernel/entry.S	2004-10-16 23:38:18.938998719 +1000
@@ -218,10 +218,8 @@ need_resched:
 	jz restore_all
 	testl $IF_MASK,EFLAGS(%esp)     # interrupts off (exception path) ?
 	jz restore_all
-	movl $PREEMPT_ACTIVE,TI_preempt_count(%ebp)
 	sti
-	call schedule
-	movl $0,TI_preempt_count(%ebp)
+	call preempt_schedule
 	cli
 	jmp need_resched
 #endif
Index: linux-2.6.9-rc4-ck3/arch/i386/oprofile/nmi_int.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/arch/i386/oprofile/nmi_int.c	2004-05-23 12:54:46.000000000 +1000
+++ linux-2.6.9-rc4-ck3/arch/i386/oprofile/nmi_int.c	2004-10-16 23:38:18.954996221 +1000
@@ -311,7 +311,7 @@ struct oprofile_operations nmi_ops = {
 
 static int __init p4_init(void)
 {
-	__u8 cpu_model = current_cpu_data.x86_model;
+	__u8 cpu_model = boot_cpu_data.x86_model;
 
 	if (cpu_model > 3)
 		return 0;
@@ -342,7 +342,7 @@ static int __init p4_init(void)
 
 static int __init ppro_init(void)
 {
-	__u8 cpu_model = current_cpu_data.x86_model;
+	__u8 cpu_model = boot_cpu_data.x86_model;
 
 	if (cpu_model > 0xd)
 		return 0;
@@ -366,8 +366,8 @@ static int using_nmi;
 
 int __init nmi_init(struct oprofile_operations ** ops)
 {
-	__u8 vendor = current_cpu_data.x86_vendor;
-	__u8 family = current_cpu_data.x86;
+	__u8 vendor = boot_cpu_data.x86_vendor;
+	__u8 family = boot_cpu_data.x86;
  
 	if (!cpu_has_apic)
 		return -ENODEV;
Index: linux-2.6.9-rc4-ck3/drivers/char/pty.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/drivers/char/pty.c	2004-10-11 16:15:59.000000000 +1000
+++ linux-2.6.9-rc4-ck3/drivers/char/pty.c	2004-10-16 23:38:18.983991694 +1000
@@ -140,6 +140,7 @@ static int pty_write(struct tty_struct *
 			c     += n;
 			count -= n;
 			to->ldisc.receive_buf(to, temp_buffer, NULL, n);
+			cond_resched();
 		}
 		up(&tty->flip.pty_sem);
 	} else {
Index: linux-2.6.9-rc4-ck3/drivers/video/console/vgacon.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/drivers/video/console/vgacon.c	2004-08-15 14:08:09.000000000 +1000
+++ linux-2.6.9-rc4-ck3/drivers/video/console/vgacon.c	2004-10-16 23:38:18.994989977 +1000
@@ -49,6 +49,7 @@
 #include <linux/spinlock.h>
 #include <linux/ioport.h>
 #include <linux/init.h>
+#include <linux/smp_lock.h>
 #include <video/vga.h>
 #include <asm/io.h>
 
@@ -763,6 +764,7 @@ static int vgacon_do_font_op(struct vgas
 		charmap += 4 * cmapsz;
 #endif
 
+	unlock_kernel();
 	spin_lock_irq(&vga_lock);
 	/* First, the Sequencer */
 	vga_wseq(state->vgabase, VGA_SEQ_RESET, 0x1);
@@ -848,6 +850,7 @@ static int vgacon_do_font_op(struct vgas
 		vga_wattr(state->vgabase, VGA_AR_ENABLE_DISPLAY, 0);	
 	}
 	spin_unlock_irq(&vga_lock);
+	lock_kernel();
 	return 0;
 }
 
Index: linux-2.6.9-rc4-ck3/fs/dcache.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/fs/dcache.c	2004-10-11 16:16:00.000000000 +1000
+++ linux-2.6.9-rc4-ck3/fs/dcache.c	2004-10-16 23:38:19.008987791 +1000
@@ -155,7 +155,7 @@ repeat:
 		spin_unlock(&dcache_lock);
 		return;
 	}
-			
+
 	/*
 	 * AV: ->d_delete() is _NOT_ allowed to block now.
 	 */
@@ -387,6 +387,8 @@ static void prune_dcache(int count)
 		struct dentry *dentry;
 		struct list_head *tmp;
 
+		cond_resched_lock(&dcache_lock);
+
 		tmp = dentry_unused.prev;
 		if (tmp == &dentry_unused)
 			break;
@@ -543,6 +545,13 @@ positive:
  * list for prune_dcache(). We descend to the next level
  * whenever the d_subdirs list is non-empty and continue
  * searching.
+ *
+ * It returns zero iff there are no unused children,
+ * otherwise  it returns the number of children moved to
+ * the end of the unused list. This may not be the total
+ * number of unused children, because select_parent can
+ * drop the lock and return early due to latency
+ * constraints.
  */
 static int select_parent(struct dentry * parent)
 {
@@ -572,6 +581,15 @@ resume:
 			dentry_stat.nr_unused++;
 			found++;
 		}
+
+		/*
+		 * We can return to the caller if we have found some (this
+		 * ensures forward progress). We'll be coming back to find
+		 * the rest.
+		 */
+		if (found && need_resched())
+			goto out;
+
 		/*
 		 * Descend a level if the d_subdirs list is non-empty.
 		 */
@@ -596,6 +614,7 @@ this_parent->d_parent->d_name.name, this
 #endif
 		goto resume;
 	}
+out:
 	spin_unlock(&dcache_lock);
 	return found;
 }
Index: linux-2.6.9-rc4-ck3/fs/exec.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/fs/exec.c	2004-10-11 16:16:00.000000000 +1000
+++ linux-2.6.9-rc4-ck3/fs/exec.c	2004-10-16 23:38:19.017986386 +1000
@@ -184,6 +184,7 @@ static int count(char __user * __user * 
 			argv++;
 			if(++i > max)
 				return -E2BIG;
+			cond_resched();
 		}
 	}
 	return i;
Index: linux-2.6.9-rc4-ck3/fs/fs-writeback.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/fs/fs-writeback.c	2004-10-11 16:16:00.000000000 +1000
+++ linux-2.6.9-rc4-ck3/fs/fs-writeback.c	2004-10-16 23:38:19.030984356 +1000
@@ -370,6 +370,7 @@ sync_sb_inodes(struct super_block *sb, s
 			list_move(&inode->i_list, &sb->s_dirty);
 		}
 		spin_unlock(&inode_lock);
+		cond_resched();
 		iput(inode);
 		spin_lock(&inode_lock);
 		if (wbc->nr_to_write <= 0)
Index: linux-2.6.9-rc4-ck3/fs/inode.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/fs/inode.c	2004-10-11 16:16:00.000000000 +1000
+++ linux-2.6.9-rc4-ck3/fs/inode.c	2004-10-16 23:38:19.043982327 +1000
@@ -306,6 +306,14 @@ static int invalidate_list(struct list_h
 		struct list_head * tmp = next;
 		struct inode * inode;
 
+		/*
+		 * We can reschedule here without worrying about the list's
+		 * consistency because the per-sb list of inodes must not
+		 * change during umount anymore, and because iprune_sem keeps
+		 * shrink_icache_memory() away.
+		 */
+		cond_resched_lock(&inode_lock);
+
 		next = next->next;
 		if (tmp == head)
 			break;
Index: linux-2.6.9-rc4-ck3/fs/select.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/fs/select.c	2004-10-11 16:16:00.000000000 +1000
+++ linux-2.6.9-rc4-ck3/fs/select.c	2004-10-16 23:38:19.044982171 +1000
@@ -239,6 +239,7 @@ int do_select(int n, fd_set_bits *fds, l
 						retval++;
 					}
 				}
+				cond_resched();
 			}
 			if (res_in)
 				*rinp = res_in;
Index: linux-2.6.9-rc4-ck3/include/asm-alpha/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-alpha/spinlock.h	2004-10-11 16:16:00.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-alpha/spinlock.h	2004-10-16 23:38:19.072977800 +1000
@@ -23,6 +23,9 @@ typedef struct {
 	struct task_struct * task;
 	const char *base_file;
 #endif
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #ifdef CONFIG_DEBUG_SPINLOCK
@@ -96,6 +99,9 @@ static inline int _raw_spin_trylock(spin
 
 typedef struct {
 	volatile unsigned int write_lock:1, read_counter:31;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } /*__attribute__((aligned(32)))*/ rwlock_t;
 
 #define RW_LOCK_UNLOCKED (rwlock_t) { 0, 0 }
Index: linux-2.6.9-rc4-ck3/include/asm-arm/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-arm/spinlock.h	2004-06-16 17:35:44.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-arm/spinlock.h	2004-10-16 23:38:19.087975458 +1000
@@ -17,6 +17,9 @@
  */
 typedef struct {
 	volatile unsigned int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #define SPIN_LOCK_UNLOCKED	(spinlock_t) { 0 }
@@ -70,6 +73,9 @@ static inline void _raw_spin_unlock(spin
  */
 typedef struct {
 	volatile unsigned int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #define RW_LOCK_UNLOCKED	(rwlock_t) { 0 }
@@ -143,6 +149,8 @@ static inline void _raw_read_unlock(rwlo
 	: "cc", "memory");
 }
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 static inline int _raw_write_trylock(rwlock_t *rw)
 {
 	unsigned long tmp;
Index: linux-2.6.9-rc4-ck3/include/asm-i386/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-i386/spinlock.h	2004-10-11 16:16:00.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-i386/spinlock.h	2004-10-16 23:38:19.103972960 +1000
@@ -19,6 +19,9 @@ typedef struct {
 #ifdef CONFIG_DEBUG_SPINLOCK
 	unsigned magic;
 #endif
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #define SPINLOCK_MAGIC	0xdead4ead
@@ -166,6 +169,9 @@ typedef struct {
 #ifdef CONFIG_DEBUG_SPINLOCK
 	unsigned magic;
 #endif
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #define RWLOCK_MAGIC	0xdeaf1eed
@@ -212,6 +218,16 @@ static inline void _raw_write_lock(rwloc
 #define _raw_read_unlock(rw)		asm volatile("lock ; incl %0" :"=m" ((rw)->lock) : : "memory")
 #define _raw_write_unlock(rw)	asm volatile("lock ; addl $" RW_LOCK_BIAS_STR ",%0":"=m" ((rw)->lock) : : "memory")
 
+static inline int _raw_read_trylock(rwlock_t *lock)
+{
+	atomic_t *count = (atomic_t *)lock;
+	atomic_dec(count);
+	if (atomic_read(count) >= 0)
+		return 1;
+	atomic_inc(count);
+	return 0;
+}
+
 static inline int _raw_write_trylock(rwlock_t *lock)
 {
 	atomic_t *count = (atomic_t *)lock;
Index: linux-2.6.9-rc4-ck3/include/asm-ia64/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-ia64/spinlock.h	2004-10-11 16:16:00.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-ia64/spinlock.h	2004-10-16 23:38:19.134968121 +1000
@@ -19,6 +19,9 @@
 
 typedef struct {
 	volatile unsigned int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #define SPIN_LOCK_UNLOCKED			(spinlock_t) { 0 }
@@ -116,6 +119,9 @@ do {											\
 typedef struct {
 	volatile unsigned int read_counter	: 31;
 	volatile unsigned int write_lock	:  1;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 #define RW_LOCK_UNLOCKED (rwlock_t) { 0, 0 }
 
@@ -190,6 +196,8 @@ do {										\
 
 #endif /* !ASM_SUPPORTED */
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 #define _raw_write_unlock(x)								\
 ({											\
 	smp_mb__before_clear_bit();	/* need barrier before releasing lock... */	\
Index: linux-2.6.9-rc4-ck3/include/asm-mips/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-mips/spinlock.h	2004-06-16 17:35:45.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-mips/spinlock.h	2004-10-16 23:38:19.142966872 +1000
@@ -15,6 +15,9 @@
 
 typedef struct {
 	volatile unsigned int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #define SPIN_LOCK_UNLOCKED (spinlock_t) { 0 }
@@ -92,6 +95,9 @@ static inline unsigned int _raw_spin_try
 
 typedef struct {
 	volatile unsigned int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #define RW_LOCK_UNLOCKED (rwlock_t) { 0 }
@@ -168,6 +174,8 @@ static inline void _raw_write_unlock(rwl
 	: "memory");
 }
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 static inline int _raw_write_trylock(rwlock_t *rw)
 {
 	unsigned int tmp;
Index: linux-2.6.9-rc4-ck3/include/asm-parisc/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-parisc/spinlock.h	2004-08-15 14:08:18.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-parisc/spinlock.h	2004-10-16 23:38:19.167962969 +1000
@@ -128,6 +128,9 @@ do {									\
 typedef struct {
 	spinlock_t lock;
 	volatile int counter;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #define RW_LOCK_UNLOCKED (rwlock_t) { __SPIN_LOCK_UNLOCKED, 0 }
@@ -136,6 +139,8 @@ typedef struct {
 
 #define rwlock_is_locked(lp) ((lp)->counter != 0)
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 /* read_lock, read_unlock are pretty straightforward.  Of course it somehow
  * sucks we end up saving/restoring flags twice for read_lock_irqsave aso. */
 
Index: linux-2.6.9-rc4-ck3/include/asm-parisc/system.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-parisc/system.h	2004-08-15 14:08:18.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-parisc/system.h	2004-10-16 23:38:19.185960159 +1000
@@ -176,6 +176,9 @@ typedef struct {
 	void *previous;
 	struct task_struct * task;
 #endif
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #define __lock_aligned __attribute__((__section__(".data.lock_aligned")))
Index: linux-2.6.9-rc4-ck3/include/asm-ppc64/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-ppc64/spinlock.h	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-ppc64/spinlock.h	2004-10-16 23:38:19.203957349 +1000
@@ -23,10 +23,16 @@
 
 typedef struct {
 	volatile unsigned int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 typedef struct {
 	volatile signed int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #ifdef __KERNEL__
@@ -216,6 +222,8 @@ static void __inline__ _raw_read_unlock(
 	: "cr0", "memory");
 }
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 /*
  * This returns the old value in the lock,
  * so we got the write lock if the return value is 0.
Index: linux-2.6.9-rc4-ck3/include/asm-ppc/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-ppc/spinlock.h	2004-06-16 17:35:45.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-ppc/spinlock.h	2004-10-16 23:38:19.224954070 +1000
@@ -13,6 +13,9 @@ typedef struct {
 	volatile unsigned long owner_pc;
 	volatile unsigned long owner_cpu;
 #endif
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #ifdef __KERNEL__
@@ -83,6 +86,9 @@ typedef struct {
 #ifdef CONFIG_DEBUG_SPINLOCK
 	volatile unsigned long owner_pc;
 #endif
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #ifdef CONFIG_DEBUG_SPINLOCK
@@ -192,5 +198,7 @@ extern int _raw_write_trylock(rwlock_t *
 
 #endif
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 #endif /* __ASM_SPINLOCK_H */
 #endif /* __KERNEL__ */
Index: linux-2.6.9-rc4-ck3/include/asm-s390/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-s390/spinlock.h	2004-06-16 17:35:45.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-s390/spinlock.h	2004-10-16 23:38:19.241951416 +1000
@@ -36,6 +36,9 @@
 
 typedef struct {
 	volatile unsigned int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } __attribute__ ((aligned (4))) spinlock_t;
 
 #define SPIN_LOCK_UNLOCKED (spinlock_t) { 0 }
@@ -105,6 +108,9 @@ extern inline void _raw_spin_unlock(spin
 typedef struct {
 	volatile unsigned long lock;
 	volatile unsigned long owner_pc;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #define RW_LOCK_UNLOCKED (rwlock_t) { 0, 0 }
@@ -211,6 +217,8 @@ typedef struct {
 		       "m" ((rw)->lock) : "2", "3", "cc", "memory" )
 #endif /* __s390x__ */
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 extern inline int _raw_write_trylock(rwlock_t *rw)
 {
 	unsigned long result, reg;
Index: linux-2.6.9-rc4-ck3/include/asm-sh/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-sh/spinlock.h	2004-06-16 17:35:45.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-sh/spinlock.h	2004-10-16 23:38:19.259948606 +1000
@@ -17,6 +17,9 @@
  */
 typedef struct {
 	volatile unsigned long lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #define SPIN_LOCK_UNLOCKED	(spinlock_t) { 0 }
@@ -68,6 +71,9 @@ static inline void _raw_spin_unlock(spin
 typedef struct {
 	spinlock_t lock;
 	atomic_t counter;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #define RW_LOCK_BIAS		0x01000000
@@ -105,6 +111,8 @@ static inline void _raw_write_unlock(rwl
 	_raw_spin_unlock(&rw->lock);
 }
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 static inline int _raw_write_trylock(rwlock_t *rw)
 {
 	if (atomic_sub_and_test(RW_LOCK_BIAS, &rw->counter))
Index: linux-2.6.9-rc4-ck3/include/asm-sparc64/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-sparc64/spinlock.h	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-sparc64/spinlock.h	2004-10-16 23:38:19.294943142 +1000
@@ -304,6 +304,8 @@ do {	unsigned long flags; \
 
 #endif /* CONFIG_DEBUG_SPINLOCK */
 
+#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+
 #endif /* !(__ASSEMBLY__) */
 
 #endif /* !(__SPARC64_SPINLOCK_H) */
Index: linux-2.6.9-rc4-ck3/include/asm-sparc/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-sparc/spinlock.h	2004-06-16 17:35:45.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-sparc/spinlock.h	2004-10-16 23:38:19.335936742 +1000
@@ -16,6 +16,9 @@
 struct _spinlock_debug {
 	unsigned char lock;
 	unsigned long owner_pc;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 };
 typedef struct _spinlock_debug spinlock_t;
 
@@ -36,6 +39,9 @@ struct _rwlock_debug {
 	volatile unsigned int lock;
 	unsigned long owner_pc;
 	unsigned long reader_pc[NR_CPUS];
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 };
 typedef struct _rwlock_debug rwlock_t;
 
@@ -79,8 +85,14 @@ do {	unsigned long flags; \
 
 #else /* !CONFIG_DEBUG_SPINLOCK */
 
-typedef unsigned char spinlock_t;
-#define SPIN_LOCK_UNLOCKED	0
+typedef struct {
+	unsigned char lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
+} spinlock_t;
+
+#define SPIN_LOCK_UNLOCKED	{ 0, }
 
 #define spin_lock_init(lock)   (*((unsigned char *)(lock)) = 0)
 #define spin_is_locked(lock)    (*((volatile unsigned char *)(lock)) != 0)
@@ -137,7 +149,12 @@ extern __inline__ void _raw_spin_unlock(
  * XXX This might create some problems with my dual spinlock
  * XXX scheme, deadlocks etc. -DaveM
  */
-typedef struct { volatile unsigned int lock; } rwlock_t;
+typedef struct {
+	volatile unsigned int lock;
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
+} rwlock_t;
 
 #define RW_LOCK_UNLOCKED (rwlock_t) { 0 }
 
Index: linux-2.6.9-rc4-ck3/include/asm-x86_64/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/asm-x86_64/spinlock.h	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/asm-x86_64/spinlock.h	2004-10-16 23:38:19.357933307 +1000
@@ -18,6 +18,9 @@ typedef struct {
 #ifdef CONFIG_DEBUG_SPINLOCK
 	unsigned magic;
 #endif
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } spinlock_t;
 
 #define SPINLOCK_MAGIC	0xdead4ead
@@ -139,6 +142,9 @@ typedef struct {
 #ifdef CONFIG_DEBUG_SPINLOCK
 	unsigned magic;
 #endif
+#ifdef CONFIG_PREEMPT
+	unsigned int break_lock;
+#endif
 } rwlock_t;
 
 #define RWLOCK_MAGIC	0xdeaf1eed
@@ -185,6 +191,16 @@ static inline void _raw_write_lock(rwloc
 #define _raw_read_unlock(rw)		asm volatile("lock ; incl %0" :"=m" ((rw)->lock) : : "memory")
 #define _raw_write_unlock(rw)	asm volatile("lock ; addl $" RW_LOCK_BIAS_STR ",%0":"=m" ((rw)->lock) : : "memory")
 
+static inline int _raw_read_trylock(rwlock_t *lock)
+{
+	atomic_t *count = (atomic_t *)lock;
+	atomic_dec(count);
+	if (atomic_read(count) >= 0)
+		return 1;
+	atomic_inc(count);
+	return 0;
+}
+
 static inline int _raw_write_trylock(rwlock_t *lock)
 {
 	atomic_t *count = (atomic_t *)lock;
Index: linux-2.6.9-rc4-ck3/include/linux/hardirq.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/linux/hardirq.h	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/linux/hardirq.h	2004-10-16 23:38:19.377930185 +1000
@@ -32,7 +32,7 @@
 # define preemptible()	(preempt_count() == 0 && !irqs_disabled())
 # define IRQ_EXIT_OFFSET (HARDIRQ_OFFSET-1)
 #else
-# define in_atomic()	(preempt_count() != 0)
+# define in_atomic()	((preempt_count() & ~PREEMPT_ACTIVE) != 0)
 # define preemptible()	0
 # define IRQ_EXIT_OFFSET HARDIRQ_OFFSET
 #endif
Index: linux-2.6.9-rc4-ck3/include/linux/mmzone.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/linux/mmzone.h	2004-10-16 15:19:11.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/linux/mmzone.h	2004-10-16 23:38:19.393927687 +1000
@@ -370,7 +370,7 @@ int lower_zone_protection_sysctl_handler
 
 #include <linux/topology.h>
 /* Returns the number of the current Node. */
-#define numa_node_id()		(cpu_to_node(smp_processor_id()))
+#define numa_node_id()		(cpu_to_node(_smp_processor_id()))
 
 #ifndef CONFIG_DISCONTIGMEM
 
Index: linux-2.6.9-rc4-ck3/include/linux/sched.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/linux/sched.h	2004-10-16 16:10:11.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/linux/sched.h	2004-10-16 23:38:19.405925814 +1000
@@ -963,29 +963,36 @@ static inline int need_resched(void)
 	return unlikely(test_thread_flag(TIF_NEED_RESCHED));
 }
 
-extern void __cond_resched(void);
-static inline void cond_resched(void)
-{
-	if (need_resched())
-		__cond_resched();
-}
+/*
+ * cond_resched() and cond_resched_lock(): latency reduction via
+ * explicit rescheduling in places that are safe. The return
+ * value indicates whether a reschedule was done in fact.
+ * cond_resched_lock() will drop the spinlock before scheduling,
+ * cond_resched_softirq() will enable bhs before scheduling.
+ */
+extern int cond_resched(void);
+extern int cond_resched_lock(spinlock_t * lock);
+extern int cond_resched_softirq(void);
+
+/*
+ * Does a critical section need to be broken due to another
+ * task waiting?:
+ */
+#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
+# define need_lockbreak(lock) ((lock)->break_lock)
+#else
+# define need_lockbreak(lock) 0
+#endif
 
 /*
- * cond_resched_lock() - if a reschedule is pending, drop the given lock,
- * call schedule, and on return reacquire the lock.
- *
- * This works OK both with and without CONFIG_PREEMPT.  We do strange low-level
- * operations here to prevent schedule() from being called twice (once via
- * spin_unlock(), once by hand).
+ * Does a critical section need to be broken due to another
+ * task waiting or preemption being signalled:
  */
-static inline void cond_resched_lock(spinlock_t * lock)
+static inline int lock_need_resched(spinlock_t *lock)
 {
-	if (need_resched()) {
-		_raw_spin_unlock(lock);
-		preempt_enable_no_resched();
-		__cond_resched();
-		spin_lock(lock);
-	}
+	if (need_lockbreak(lock) || need_resched())
+		return 1;
+	return 0;
 }
 
 /* Reevaluate whether the task has signals pending delivery.
Index: linux-2.6.9-rc4-ck3/include/linux/spinlock.h
===================================================================
--- linux-2.6.9-rc4-ck3.orig/include/linux/spinlock.h	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/include/linux/spinlock.h	2004-10-16 23:38:19.422923160 +1000
@@ -41,6 +41,7 @@
 #define __lockfunc fastcall __attribute__((section(".spinlock.text")))
 
 int __lockfunc _spin_trylock(spinlock_t *lock);
+int __lockfunc _read_trylock(rwlock_t *lock);
 int __lockfunc _write_trylock(rwlock_t *lock);
 void __lockfunc _spin_lock(spinlock_t *lock);
 void __lockfunc _write_lock(rwlock_t *lock);
@@ -68,7 +69,9 @@ void __lockfunc _write_unlock_irqrestore
 void __lockfunc _write_unlock_irq(rwlock_t *lock);
 void __lockfunc _write_unlock_bh(rwlock_t *lock);
 int __lockfunc _spin_trylock_bh(spinlock_t *lock);
+int __lockfunc generic_raw_read_trylock(rwlock_t *lock);
 int in_lock_functions(unsigned long addr);
+
 #else
 
 #define in_lock_functions(ADDR) 0
@@ -213,11 +216,15 @@ typedef struct {
 #define _raw_read_unlock(lock)	do { (void)(lock); } while(0)
 #define _raw_write_lock(lock)	do { (void)(lock); } while(0)
 #define _raw_write_unlock(lock)	do { (void)(lock); } while(0)
+#define _raw_read_trylock(lock) ({ (void)(lock); (1); })
 #define _raw_write_trylock(lock) ({ (void)(lock); (1); })
 
 #define _spin_trylock(lock)	({preempt_disable(); _raw_spin_trylock(lock) ? \
 				1 : ({preempt_enable(); 0;});})
 
+#define _read_trylock(lock)	({preempt_disable();_raw_read_trylock(lock) ? \
+				1 : ({preempt_enable(); 0;});})
+
 #define _write_trylock(lock)	({preempt_disable(); _raw_write_trylock(lock) ? \
 				1 : ({preempt_enable(); 0;});})
 
@@ -401,16 +408,12 @@ do { \
  * methods are defined as nops in the case they are not required.
  */
 #define spin_trylock(lock)	_spin_trylock(lock)
+#define read_trylock(lock)	_read_trylock(lock)
 #define write_trylock(lock)	_write_trylock(lock)
 
-/* Where's read_trylock? */
-
 #define spin_lock(lock)		_spin_lock(lock)
 #define write_lock(lock)	_write_lock(lock)
 #define read_lock(lock)		_read_lock(lock)
-#define spin_unlock(lock)	_spin_unlock(lock)
-#define write_unlock(lock)	_write_unlock(lock)
-#define read_unlock(lock)	_read_unlock(lock)
 
 #ifdef CONFIG_SMP
 #define spin_lock_irqsave(lock, flags)	flags = _spin_lock_irqsave(lock)
@@ -430,6 +433,11 @@ do { \
 
 #define write_lock_irq(lock)		_write_lock_irq(lock)
 #define write_lock_bh(lock)		_write_lock_bh(lock)
+
+#define spin_unlock(lock)	_spin_unlock(lock)
+#define write_unlock(lock)	_write_unlock(lock)
+#define read_unlock(lock)	_read_unlock(lock)
+
 #define spin_unlock_irqrestore(lock, flags)	_spin_unlock_irqrestore(lock, flags)
 #define spin_unlock_irq(lock)		_spin_unlock_irq(lock)
 #define spin_unlock_bh(lock)		_spin_unlock_bh(lock)
@@ -452,6 +460,7 @@ extern void _metered_read_lock    (rwloc
 extern void _metered_read_unlock  (rwlock_t *lock);
 extern void _metered_write_lock   (rwlock_t *lock);
 extern void _metered_write_unlock (rwlock_t *lock);
+extern int  _metered_read_trylock (rwlock_t *lock);
 extern int  _metered_write_trylock(rwlock_t *lock);
 #endif
 
@@ -479,8 +488,11 @@ static inline void bit_spin_lock(int bit
 	preempt_disable();
 #if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_SPINLOCK)
 	while (test_and_set_bit(bitnum, addr)) {
-		while (test_bit(bitnum, addr))
+		while (test_bit(bitnum, addr)) {
+			preempt_enable();
 			cpu_relax();
+			preempt_disable();
+		}
 	}
 #endif
 }
Index: linux-2.6.9-rc4-ck3/init/main.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/init/main.c	2004-10-16 16:10:11.000000000 +1000
+++ linux-2.6.9-rc4-ck3/init/main.c	2004-10-16 23:38:19.427922379 +1000
@@ -436,6 +436,12 @@ static void noinline rest_init(void)
 {
 	kernel_thread(init, NULL, CLONE_FS | CLONE_SIGHAND);
 	numa_default_policy();
+	/*
+	 * Re-enable preemption but disable interrupts to make sure
+	 * we dont get preempted until we schedule() in cpu_idle().
+	 */
+	local_irq_disable();
+	preempt_enable_no_resched();
 	unlock_kernel();
  	cpu_idle();
 } 
@@ -502,6 +508,7 @@ asmlinkage void __init start_kernel(void
 	 * time - but meanwhile we still have a functioning scheduler.
 	 */
 	sched_init();
+	preempt_disable();
 	build_all_zonelists();
 	page_alloc_init();
 	printk("Kernel command line: %s\n", saved_command_line);
Index: linux-2.6.9-rc4-ck3/kernel/kthread.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/kernel/kthread.c	2004-08-15 14:08:19.000000000 +1000
+++ linux-2.6.9-rc4-ck3/kernel/kthread.c	2004-10-16 23:38:19.446919413 +1000
@@ -14,6 +14,12 @@
 #include <linux/module.h>
 #include <asm/semaphore.h>
 
+/*
+ * We dont want to execute off keventd since it might
+ * hold a semaphore our callers hold too:
+ */
+static struct workqueue_struct *helper_wq;
+
 struct kthread_create_info
 {
 	/* Information passed to kthread() from keventd. */
@@ -126,12 +132,13 @@ struct task_struct *kthread_create(int (
 	init_completion(&create.started);
 	init_completion(&create.done);
 
-	/* If we're being called to start the first workqueue, we
-	 * can't use keventd. */
-	if (!keventd_up())
+	/*
+	 * The workqueue needs to start up first:
+	 */
+	if (!helper_wq)
 		work.func(work.data);
 	else {
-		schedule_work(&work);
+		queue_work(helper_wq, &work);
 		wait_for_completion(&create.done);
 	}
 	if (!IS_ERR(create.result)) {
@@ -183,3 +190,13 @@ int kthread_stop(struct task_struct *k)
 	return ret;
 }
 EXPORT_SYMBOL(kthread_stop);
+
+static __init int helper_init(void)
+{
+	helper_wq = create_singlethread_workqueue("kthread");
+	BUG_ON(!helper_wq);
+
+	return 0;
+}
+core_initcall(helper_init);
+
Index: linux-2.6.9-rc4-ck3/kernel/printk.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/kernel/printk.c	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/kernel/printk.c	2004-10-16 23:38:19.464916603 +1000
@@ -282,6 +282,7 @@ int do_syslog(int type, char __user * bu
 			error = __put_user(c,buf);
 			buf++;
 			i++;
+			cond_resched();
 			spin_lock_irq(&logbuf_lock);
 		}
 		spin_unlock_irq(&logbuf_lock);
@@ -323,6 +324,7 @@ int do_syslog(int type, char __user * bu
 			c = LOG_BUF(j);
 			spin_unlock_irq(&logbuf_lock);
 			error = __put_user(c,&buf[count-1-i]);
+			cond_resched();
 			spin_lock_irq(&logbuf_lock);
 		}
 		spin_unlock_irq(&logbuf_lock);
@@ -338,6 +340,7 @@ int do_syslog(int type, char __user * bu
 					error = -EFAULT;
 					break;
 				}
+				cond_resched();
 			}
 		}
 		break;
Index: linux-2.6.9-rc4-ck3/kernel/sched.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/kernel/sched.c	2004-10-16 16:10:11.000000000 +1000
+++ linux-2.6.9-rc4-ck3/kernel/sched.c	2004-10-16 23:38:19.487913012 +1000
@@ -3297,13 +3297,71 @@ asmlinkage long sys_sched_yield(void)
 	return 0;
 }
 
-void __sched __cond_resched(void)
+static inline void __cond_resched(void)
 {
-	set_current_state(TASK_RUNNING);
-	schedule();
+	do {
+		preempt_count() += PREEMPT_ACTIVE;
+		schedule();
+		preempt_count() -= PREEMPT_ACTIVE;
+	} while (need_resched());
 }
 
-EXPORT_SYMBOL(__cond_resched);
+int __sched cond_resched(void)
+{
+	if (need_resched()) {
+		__cond_resched();
+		return 1;
+	}
+	return 0;
+}
+
+EXPORT_SYMBOL(cond_resched);
+
+/*
+ * cond_resched_lock() - if a reschedule is pending, drop the given lock,
+ * call schedule, and on return reacquire the lock.
+ *
+ * This works OK both with and without CONFIG_PREEMPT.  We do strange low-level
+ * operations here to prevent schedule() from being called twice (once via
+ * spin_unlock(), once by hand).
+ */
+int cond_resched_lock(spinlock_t * lock)
+{
+#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT)
+	if (lock->break_lock) {
+		lock->break_lock = 0;
+		spin_unlock(lock);
+		cpu_relax();
+		spin_lock(lock);
+	}
+#endif
+	if (need_resched()) {
+		_raw_spin_unlock(lock);
+		preempt_enable_no_resched();
+		__cond_resched();
+		spin_lock(lock);
+		return 1;
+	}
+	return 0;
+}
+
+EXPORT_SYMBOL(cond_resched_lock);
+
+int __sched cond_resched_softirq(void)
+{
+	BUG_ON(!in_softirq());
+
+	if (need_resched()) {
+		__local_bh_enable();
+		__cond_resched();
+		local_bh_disable();
+		return 1;
+	}
+	return 0;
+}
+
+EXPORT_SYMBOL(cond_resched_softirq);
+
 
 /**
  * yield - yield the current processor to other threads.
Index: linux-2.6.9-rc4-ck3/kernel/spinlock.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/kernel/spinlock.c	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/kernel/spinlock.c	2004-10-16 23:38:19.489912700 +1000
@@ -2,6 +2,8 @@
  * Copyright (2004) Linus Torvalds
  *
  * Author: Zwane Mwaikambo <zwane@fsmlabs.com>
+ *
+ * Copyright (2004) Ingo Molnar
  */
 
 #include <linux/config.h>
@@ -11,6 +13,17 @@
 #include <linux/interrupt.h>
 #include <linux/module.h>
 
+/*
+ * Generic declaration of the raw read_trylock() function,
+ * architectures are supposed to optimize this:
+ */
+int __lockfunc generic_raw_read_trylock(rwlock_t *lock)
+{
+	_raw_read_lock(lock);
+	return 1;
+}
+EXPORT_SYMBOL(generic_raw_read_trylock);
+
 int __lockfunc _spin_trylock(spinlock_t *lock)
 {
 	preempt_disable();
@@ -22,86 +35,29 @@ int __lockfunc _spin_trylock(spinlock_t 
 }
 EXPORT_SYMBOL(_spin_trylock);
 
-int __lockfunc _write_trylock(rwlock_t *lock)
+int __lockfunc _read_trylock(rwlock_t *lock)
 {
 	preempt_disable();
-	if (_raw_write_trylock(lock))
+	if (_raw_read_trylock(lock))
 		return 1;
 
 	preempt_enable();
 	return 0;
 }
-EXPORT_SYMBOL(_write_trylock);
-
-#ifdef CONFIG_PREEMPT
-/*
- * This could be a long-held lock.  If another CPU holds it for a long time,
- * and that CPU is not asked to reschedule then *this* CPU will spin on the
- * lock for a long time, even if *this* CPU is asked to reschedule.
- *
- * So what we do here, in the slow (contended) path is to spin on the lock by
- * hand while permitting preemption.
- *
- * Called inside preempt_disable().
- */
-static inline void __preempt_spin_lock(spinlock_t *lock)
-{
-	if (preempt_count() > 1) {
-		_raw_spin_lock(lock);
-		return;
-	}
-
-	do {
-		preempt_enable();
-		while (spin_is_locked(lock))
-			cpu_relax();
-		preempt_disable();
-	} while (!_raw_spin_trylock(lock));
-}
+EXPORT_SYMBOL(_read_trylock);
 
-void __lockfunc _spin_lock(spinlock_t *lock)
+int __lockfunc _write_trylock(rwlock_t *lock)
 {
 	preempt_disable();
-	if (unlikely(!_raw_spin_trylock(lock)))
-		__preempt_spin_lock(lock);
-}
-
-static inline void __preempt_write_lock(rwlock_t *lock)
-{
-	if (preempt_count() > 1) {
-		_raw_write_lock(lock);
-		return;
-	}
-
-	do {
-		preempt_enable();
-		while (rwlock_is_locked(lock))
-			cpu_relax();
-		preempt_disable();
-	} while (!_raw_write_trylock(lock));
-}
+	if (_raw_write_trylock(lock))
+		return 1;
 
-void __lockfunc _write_lock(rwlock_t *lock)
-{
-	preempt_disable();
-	if (unlikely(!_raw_write_trylock(lock)))
-		__preempt_write_lock(lock);
-}
-#else
-void __lockfunc _spin_lock(spinlock_t *lock)
-{
-	preempt_disable();
-	_raw_spin_lock(lock);
+	preempt_enable();
+	return 0;
 }
+EXPORT_SYMBOL(_write_trylock);
 
-void __lockfunc _write_lock(rwlock_t *lock)
-{
-	preempt_disable();
-	_raw_write_lock(lock);
-}
-#endif
-EXPORT_SYMBOL(_spin_lock);
-EXPORT_SYMBOL(_write_lock);
+#ifndef CONFIG_PREEMPT
 
 void __lockfunc _read_lock(rwlock_t *lock)
 {
@@ -110,27 +66,6 @@ void __lockfunc _read_lock(rwlock_t *loc
 }
 EXPORT_SYMBOL(_read_lock);
 
-void __lockfunc _spin_unlock(spinlock_t *lock)
-{
-	_raw_spin_unlock(lock);
-	preempt_enable();
-}
-EXPORT_SYMBOL(_spin_unlock);
-
-void __lockfunc _write_unlock(rwlock_t *lock)
-{
-	_raw_write_unlock(lock);
-	preempt_enable();
-}
-EXPORT_SYMBOL(_write_unlock);
-
-void __lockfunc _read_unlock(rwlock_t *lock)
-{
-	_raw_read_unlock(lock);
-	preempt_enable();
-}
-EXPORT_SYMBOL(_read_unlock);
-
 unsigned long __lockfunc _spin_lock_irqsave(spinlock_t *lock)
 {
 	unsigned long flags;
@@ -212,6 +147,130 @@ void __lockfunc _write_lock_bh(rwlock_t 
 }
 EXPORT_SYMBOL(_write_lock_bh);
 
+void __lockfunc _spin_lock(spinlock_t *lock)
+{
+	preempt_disable();
+	_raw_spin_lock(lock);
+}
+
+EXPORT_SYMBOL(_spin_lock);
+
+void __lockfunc _write_lock(rwlock_t *lock)
+{
+	preempt_disable();
+	_raw_write_lock(lock);
+}
+
+EXPORT_SYMBOL(_write_lock);
+
+#else /* CONFIG_PREEMPT: */
+
+/*
+ * This could be a long-held lock. We both prepare to spin for a long
+ * time (making _this_ CPU preemptable if possible), and we also signal
+ * towards that other CPU that it should break the lock ASAP.
+ *
+ * (We do this in a function because inlining it would be excessive.)
+ */
+
+#define BUILD_LOCK_OPS(op, locktype)					\
+void __lockfunc _##op##_lock(locktype *lock)				\
+{									\
+	preempt_disable();						\
+	for (;;) {							\
+		if (likely(_raw_##op##_trylock(lock)))			\
+			break;						\
+		preempt_enable();					\
+		if (!(lock)->break_lock)				\
+			(lock)->break_lock = 1;				\
+		cpu_relax();						\
+		preempt_disable();					\
+	}								\
+}									\
+									\
+EXPORT_SYMBOL(_##op##_lock);						\
+									\
+unsigned long __lockfunc _##op##_lock_irqsave(locktype *lock)		\
+{									\
+	unsigned long flags;						\
+									\
+	preempt_disable();						\
+	for (;;) {							\
+		local_irq_save(flags);					\
+		if (likely(_raw_##op##_trylock(lock)))			\
+			break;						\
+		local_irq_restore(flags);				\
+									\
+		preempt_enable();					\
+		if (!(lock)->break_lock)				\
+			(lock)->break_lock = 1;				\
+		cpu_relax();						\
+		preempt_disable();					\
+	}								\
+	return flags;							\
+}									\
+									\
+EXPORT_SYMBOL(_##op##_lock_irqsave);					\
+									\
+void __lockfunc _##op##_lock_irq(locktype *lock)			\
+{									\
+	_##op##_lock_irqsave(lock);					\
+}									\
+									\
+EXPORT_SYMBOL(_##op##_lock_irq);					\
+									\
+void __lockfunc _##op##_lock_bh(locktype *lock)				\
+{									\
+	unsigned long flags;						\
+									\
+	/*							*/	\
+	/* Careful: we must exclude softirqs too, hence the	*/	\
+	/* irq-disabling. We use the generic preemption-aware	*/	\
+	/* function:						*/	\
+	/**/								\
+	flags = _##op##_lock_irqsave(lock);				\
+	local_bh_disable();						\
+	local_irq_restore(flags);					\
+}									\
+									\
+EXPORT_SYMBOL(_##op##_lock_bh)
+
+/*
+ * Build preemption-friendly versions of the following
+ * lock-spinning functions:
+ *
+ *         _[spin|read|write]_lock()
+ *         _[spin|read|write]_lock_irq()
+ *         _[spin|read|write]_lock_irqsave()
+ *         _[spin|read|write]_lock_bh()
+ */
+BUILD_LOCK_OPS(spin, spinlock_t);
+BUILD_LOCK_OPS(read, rwlock_t);
+BUILD_LOCK_OPS(write, rwlock_t);
+
+#endif /* CONFIG_PREEMPT */
+
+void __lockfunc _spin_unlock(spinlock_t *lock)
+{
+	_raw_spin_unlock(lock);
+	preempt_enable();
+}
+EXPORT_SYMBOL(_spin_unlock);
+
+void __lockfunc _write_unlock(rwlock_t *lock)
+{
+	_raw_write_unlock(lock);
+	preempt_enable();
+}
+EXPORT_SYMBOL(_write_unlock);
+
+void __lockfunc _read_unlock(rwlock_t *lock)
+{
+	_raw_read_unlock(lock);
+	preempt_enable();
+}
+EXPORT_SYMBOL(_read_unlock);
+
 void __lockfunc _spin_unlock_irqrestore(spinlock_t *lock, unsigned long flags)
 {
 	_raw_spin_unlock(lock);
Index: linux-2.6.9-rc4-ck3/mm/memory.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/mm/memory.c	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/mm/memory.c	2004-10-16 23:38:19.520907861 +1000
@@ -329,6 +329,25 @@ skip_copy_pte_range:
 				set_pte(dst_pte, pte);
 				page_dup_rmap(page);
 cont_copy_pte_range_noset:
+				/*
+				 * We are holding two locks at this point -
+				 * any of them could generate latencies in
+				 * another task on another CPU:
+				 */
+				if (need_resched() ||
+					need_lockbreak(&src->page_table_lock) ||
+					need_lockbreak(&dst->page_table_lock))
+				{
+					pte_unmap_nested(src_pte);
+					pte_unmap(dst_pte);
+					spin_unlock(&src->page_table_lock);
+					spin_unlock(&dst->page_table_lock);
+					cond_resched();
+					spin_lock(&dst->page_table_lock);
+					spin_lock(&src->page_table_lock);
+					dst_pte = pte_offset_map(dst_pmd, address);
+					src_pte = pte_offset_map_nested(src_pmd, address);
+				}
 				address += PAGE_SIZE;
 				if (address >= end) {
 					pte_unmap_nested(src_pte);
@@ -478,19 +497,15 @@ static void unmap_page_range(struct mmu_
 	tlb_end_vma(tlb, vma);
 }
 
-/* Dispose of an entire struct mmu_gather per rescheduling point */
-#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT)
-#define ZAP_BLOCK_SIZE	(FREE_PTE_NR * PAGE_SIZE)
-#endif
-
-/* For UP, 256 pages at a time gives nice low latency */
-#if !defined(CONFIG_SMP) && defined(CONFIG_PREEMPT)
-#define ZAP_BLOCK_SIZE	(256 * PAGE_SIZE)
-#endif
-
+#ifdef CONFIG_PREEMPT
+/*
+ * It's not an issue to have a small zap block size - TLB flushes
+ * only happen once normally, due to the tlb->need_flush optimization.
+ */
+# define ZAP_BLOCK_SIZE	(8 * PAGE_SIZE)
+#else
 /* No preempt: go for improved straight-line efficiency */
-#if !defined(CONFIG_PREEMPT)
-#define ZAP_BLOCK_SIZE	(1024 * PAGE_SIZE)
+# define ZAP_BLOCK_SIZE	(1024 * PAGE_SIZE)
 #endif
 
 /**
@@ -565,15 +580,15 @@ int unmap_vmas(struct mmu_gather **tlbp,
 
 			start += block;
 			zap_bytes -= block;
-			if ((long)zap_bytes > 0)
-				continue;
-			if (!atomic && need_resched()) {
+			if (!atomic) {
 				int fullmm = tlb_is_full_mm(*tlbp);
 				tlb_finish_mmu(*tlbp, tlb_start, start);
 				cond_resched_lock(&mm->page_table_lock);
 				*tlbp = tlb_gather_mmu(mm, fullmm);
 				tlb_start_valid = 0;
 			}
+			if ((long)zap_bytes > 0)
+				continue;
 			zap_bytes = ZAP_BLOCK_SIZE;
 		}
 	}
@@ -1499,6 +1514,7 @@ do_no_page(struct mm_struct *mm, struct 
 	}
 	smp_rmb();  /* Prevent CPU from reordering lock-free ->nopage() */
 retry:
+	cond_resched();
 	new_page = vma->vm_ops->nopage(vma, address & PAGE_MASK, &ret);
 
 	/* no page was available -- either SIGBUS or OOM */
Index: linux-2.6.9-rc4-ck3/mm/msync.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/mm/msync.c	2004-08-15 14:08:19.000000000 +1000
+++ linux-2.6.9-rc4-ck3/mm/msync.c	2004-10-16 23:38:19.527906768 +1000
@@ -92,8 +92,8 @@ static inline int filemap_sync_pmd_range
 	return error;
 }
 
-static int filemap_sync(struct vm_area_struct * vma, unsigned long address,
-	size_t size, unsigned int flags)
+static int __filemap_sync(struct vm_area_struct *vma, unsigned long address,
+			size_t size, unsigned int flags)
 {
 	pgd_t * dir;
 	unsigned long end = address + size;
@@ -131,6 +131,31 @@ static int filemap_sync(struct vm_area_s
 	return error;
 }
 
+#ifdef CONFIG_PREEMPT
+static int filemap_sync(struct vm_area_struct *vma, unsigned long address,
+			size_t size, unsigned int flags)
+{
+	const size_t chunk = 64 * 1024;	/* bytes */
+	int error = 0;
+
+	while (size) {
+		size_t sz = min(size, chunk);
+
+		error |= __filemap_sync(vma, address, sz, flags);
+		cond_resched();
+		address += sz;
+		size -= sz;
+	}
+	return error;
+}
+#else
+static int filemap_sync(struct vm_area_struct *vma, unsigned long address,
+			size_t size, unsigned int flags)
+{
+	return __filemap_sync(vma, address, size, flags);
+}
+#endif
+
 /*
  * MS_SYNC syncs the entire file - including mappings.
  *
Index: linux-2.6.9-rc4-ck3/mm/slab.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/mm/slab.c	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/mm/slab.c	2004-10-16 23:38:19.535905519 +1000
@@ -2790,7 +2790,7 @@ static void cache_reap(void *unused)
 next_unlock:
 		spin_unlock_irq(&searchp->spinlock);
 next:
-		;
+		cond_resched();
 	}
 	check_irq_on();
 	up(&cache_chain_sem);
Index: linux-2.6.9-rc4-ck3/mm/vmscan.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/mm/vmscan.c	2004-10-16 16:10:02.000000000 +1000
+++ linux-2.6.9-rc4-ck3/mm/vmscan.c	2004-10-16 23:38:19.547903646 +1000
@@ -359,6 +359,8 @@ static int shrink_list(struct list_head 
 		int may_enter_fs;
 		int referenced;
 
+		cond_resched();
+
 		page = lru_to_page(page_list);
 		list_del(&page->lru);
 
@@ -711,6 +713,7 @@ refill_inactive_zone(struct zone *zone, 
 		reclaim_mapped = 1;
 
 	while (!list_empty(&l_hold)) {
+		cond_resched();
 		page = lru_to_page(&l_hold);
 		list_del(&page->lru);
 		if (page_mapped(page)) {
Index: linux-2.6.9-rc4-ck3/net/core/sock.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/net/core/sock.c	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/net/core/sock.c	2004-10-16 23:38:19.548903489 +1000
@@ -933,6 +933,15 @@ void __release_sock(struct sock *sk)
 
 			skb->next = NULL;
 			sk->sk_backlog_rcv(sk, skb);
+
+			/*
+			 * We are in process context here with softirqs
+			 * disabled, use cond_resched_softirq() to preempt.
+			 * This is safe to do because we've taken the backlog
+			 * queue private:
+			 */
+			cond_resched_softirq();
+
 			skb = next;
 		} while (skb != NULL);
 
Index: linux-2.6.9-rc4-ck3/net/ipv4/tcp_ipv4.c
===================================================================
--- linux-2.6.9-rc4-ck3.orig/net/ipv4/tcp_ipv4.c	2004-10-11 16:16:01.000000000 +1000
+++ linux-2.6.9-rc4-ck3/net/ipv4/tcp_ipv4.c	2004-10-16 23:38:19.549903333 +1000
@@ -2230,7 +2230,10 @@ static void *established_get_first(struc
 		struct sock *sk;
 		struct hlist_node *node;
 		struct tcp_tw_bucket *tw;
-	       
+
+		/* We can reschedule _before_ having picked the target: */
+		cond_resched_softirq();
+
 		read_lock(&tcp_ehash[st->bucket].lock);
 		sk_for_each(sk, node, &tcp_ehash[st->bucket].chain) {
 			if (sk->sk_family != st->family) {
@@ -2277,6 +2280,10 @@ get_tw:
 		}
 		read_unlock(&tcp_ehash[st->bucket].lock);
 		st->state = TCP_SEQ_STATE_ESTABLISHED;
+
+		/* We can reschedule between buckets: */
+		cond_resched_softirq();
+
 		if (++st->bucket < tcp_ehash_size) {
 			read_lock(&tcp_ehash[st->bucket].lock);
 			sk = sk_head(&tcp_ehash[st->bucket].chain);