v2.5.2.1 -> v2.5.2.1.1

- David Howells: abtract out "current->need_resched" as "need_resched()"
- Frank Davis: ide-tape update for bio
- various: header file fixups
- Jens Axboe: fix up bio/ide/highmem issues
- Kai Germaschewski: ISDN update
- Tim Waugh: parport update
- Patrik Mochel: initcall update
- Greg KH: USB and Compaq PCI hotplug updates

22 files changed, 200 insertions, 127 deletions

diff --git a/include/asm-arm/arch-arc/system.h b/include/asm-arm/arch-arc/system.h
index fce24985d3a1..5886cbeb8718 100644
--- a/include/asm-arm/arch-arc/system.h
+++ b/include/asm-arm/arch-arc/system.h
@@ -10,7 +10,7 @@
 
 static void arch_idle(void)
 {
-	while (!current->need_resched && !hlt_counter);
+	while (!need_resched() && !hlt_counter);
 }
 
 static inline void arch_reset(char mode)
diff --git a/include/asm-arm/arch-cl7500/system.h b/include/asm-arm/arch-cl7500/system.h
index 2bdd19913af2..6cb002194834 100644
--- a/include/asm-arm/arch-cl7500/system.h
+++ b/include/asm-arm/arch-cl7500/system.h
@@ -10,7 +10,7 @@
 
 static void arch_idle(void)
 {
-	while (!current->need_resched && !hlt_counter)
+	while (!need_resched() && !hlt_counter)
 		iomd_writeb(0, IOMD_SUSMODE);
 }
 
diff --git a/include/asm-arm/arch-ebsa110/system.h b/include/asm-arm/arch-ebsa110/system.h
index c4af29befbd9..e33951995e68 100644
--- a/include/asm-arm/arch-ebsa110/system.h
+++ b/include/asm-arm/arch-ebsa110/system.h
@@ -17,7 +17,7 @@
  * will stop our MCLK signal (which provides the clock for the glue
  * logic, and therefore the timer interrupt).
  *
- * Instead, we spin, waiting for either hlt_counter or need_resched
+ * Instead, we spin, waiting for either hlt_counter or need_resched()
  * to be set.  If we have been spinning for 2cs, then we drop the
  * core clock down to the memory clock.
  */
@@ -28,13 +28,13 @@ static void arch_idle(void)
 	start_idle = jiffies;
 
 	do {
-		if (current->need_resched || hlt_counter)
+		if (need_resched() || hlt_counter)
 			goto slow_out;
 	} while (time_before(jiffies, start_idle + HZ/50));
 
 	cpu_do_idle(IDLE_CLOCK_SLOW);
 
-	while (!current->need_resched && !hlt_counter) {
+	while (!need_resched() && !hlt_counter) {
 		/* do nothing slowly */
 	}
 
diff --git a/include/asm-arm/arch-ebsa285/system.h b/include/asm-arm/arch-ebsa285/system.h
index e9b4da1e93a7..9559db044860 100644
--- a/include/asm-arm/arch-ebsa285/system.h
+++ b/include/asm-arm/arch-ebsa285/system.h
@@ -20,14 +20,14 @@ static void arch_idle(void)
 	start_idle = jiffies;
 
 	do {
-		if (current->need_resched || hlt_counter)
+		if (need_resched() || hlt_counter)
 			goto slow_out;
 		cpu_do_idle(IDLE_WAIT_FAST);
 	} while (time_before(jiffies, start_idle + HZ/50));
 
 	cpu_do_idle(IDLE_CLOCK_SLOW);
 
-	while (!current->need_resched && !hlt_counter) {
+	while (!need_resched() && !hlt_counter) {
 		cpu_do_idle(IDLE_WAIT_SLOW);
 	}
 
diff --git a/include/asm-arm/arch-nexuspci/system.h b/include/asm-arm/arch-nexuspci/system.h
index 028788bdc743..c5257807042b 100644
--- a/include/asm-arm/arch-nexuspci/system.h
+++ b/include/asm-arm/arch-nexuspci/system.h
@@ -16,7 +16,7 @@
 
 static void arch_idle(void)
 {
-	while (!current->need_resched && !hlt_counter)
+	while (!need_resched() && !hlt_counter)
 		cpu_do_idle(IDLE_WAIT_SLOW);
 }
 
diff --git a/include/asm-arm/arch-rpc/system.h b/include/asm-arm/arch-rpc/system.h
index 6e070b9071a9..7c781a6d1848 100644
--- a/include/asm-arm/arch-rpc/system.h
+++ b/include/asm-arm/arch-rpc/system.h
@@ -18,14 +18,14 @@ static void arch_idle(void)
 	start_idle = jiffies;
 
 	do {
-		if (current->need_resched || hlt_counter)
+		if (need_resched() || hlt_counter)
 			goto slow_out;
 		cpu_do_idle(IDLE_WAIT_FAST);
 	} while (time_before(jiffies, start_idle + HZ/50));
 
 	cpu_do_idle(IDLE_CLOCK_SLOW);
 
-	while (!current->need_resched && !hlt_counter) {
+	while (!need_resched() && !hlt_counter) {
 		cpu_do_idle(IDLE_WAIT_SLOW);
 	}
 
diff --git a/include/asm-arm/arch-sa1100/system.h b/include/asm-arm/arch-sa1100/system.h
index d020a49268a2..6dbf2a45b061 100644
--- a/include/asm-arm/arch-sa1100/system.h
+++ b/include/asm-arm/arch-sa1100/system.h
@@ -10,7 +10,7 @@ static inline void arch_idle(void)
 	if (!hlt_counter) {
 		int flags;
 		local_irq_save(flags);
-		if (!current->need_resched)
+		if (!need_resched())
 			cpu_do_idle(0);
 		local_irq_restore(flags);
 	}
diff --git a/include/asm-arm/arch-tbox/system.h b/include/asm-arm/arch-tbox/system.h
index e52bb50fdf62..cf90aafef297 100644
--- a/include/asm-arm/arch-tbox/system.h
+++ b/include/asm-arm/arch-tbox/system.h
@@ -13,14 +13,14 @@ static void arch_idle(void)
 	start_idle = jiffies;
 
 	do {
-		if (current->need_resched || hlt_counter)
+		if (need_resched() || hlt_counter)
 			goto slow_out;
 		cpu_do_idle(IDLE_WAIT_FAST);
 	} while (time_before(jiffies, start_idle + HZ/50));
 
 	cpu_do_idle(IDLE_CLOCK_SLOW);
 
-	while (!current->need_resched && !hlt_counter) {
+	while (!need_resched() && !hlt_counter) {
 		cpu_do_idle(IDLE_WAIT_SLOW);
 	}
 
diff --git a/include/asm-i386/mmu_context.h b/include/asm-i386/mmu_context.h
index 2f76058591fb..6525a4b66495 100644
--- a/include/asm-i386/mmu_context.h
+++ b/include/asm-i386/mmu_context.h
@@ -1,5 +1,5 @@
-#ifndef __I386_MMU_CONTEXT_H
-#define __I386_MMU_CONTEXT_H
+#ifndef __I386_SCHED_H
+#define __I386_SCHED_H
 
 #include <linux/config.h>
 #include <asm/desc.h>
@@ -16,14 +16,13 @@
 # error update this function.
 #endif
 
-static inline int sched_find_first_zero_bit(char *bitmap)
+static inline int sched_find_first_zero_bit(unsigned long *b)
 {
-	unsigned int *b = (unsigned int *)bitmap;
 	unsigned int rt;
 
 	rt = b[0] & b[1] & b[2] & b[3];
 	if (unlikely(rt != 0xffffffff))
-		return find_first_zero_bit(bitmap, MAX_RT_PRIO);
+		return find_first_zero_bit(b, MAX_RT_PRIO);
 
 	if (b[4] != ~0)
 		return ffz(b[4]) + MAX_RT_PRIO;
diff --git a/include/asm-i386/smplock.h b/include/asm-i386/smplock.h
index 10cfc1fd0c43..c270defe9be4 100644
--- a/include/asm-i386/smplock.h
+++ b/include/asm-i386/smplock.h
@@ -19,8 +19,8 @@ extern spinlock_t kernel_flag;
 do {						\
 	if (unlikely(task->lock_depth >= 0)) {	\
 		spin_unlock(&kernel_flag);	\
-		release_irqlock(cpu);		\
-		__sti();			\
+		if (global_irq_holder == (cpu))	\
+			BUG();			\
 	}					\
 } while (0)
 
diff --git a/include/linux/blk.h b/include/linux/blk.h
index efede40e2408..b6507e62eb4b 100644
--- a/include/linux/blk.h
+++ b/include/linux/blk.h
@@ -90,9 +90,9 @@ extern inline struct request *elv_next_request(request_queue_t *q)
 
 #define _elv_add_request(q, rq, back, p) do {				      \
 	if ((back))							      \
-		_elv_add_request_core((q), (rq), (q)->queue_head.prev, (p)); \
+		_elv_add_request_core((q), (rq), (q)->queue_head.prev, (p));  \
 	else								      \
-		_elv_add_request_core((q), (rq), &(q)->queue_head, 0);	      \
+		_elv_add_request_core((q), (rq), &(q)->queue_head, (p));      \
 } while (0)
 
 #define elv_add_request(q, rq, back) _elv_add_request((q), (rq), (back), 1)
diff --git a/include/linux/file.h b/include/linux/file.h
index bac31250708c..5e006ad08df0 100644
--- a/include/linux/file.h
+++ b/include/linux/file.h
@@ -5,12 +5,9 @@
 #ifndef __LINUX_FILE_H
 #define __LINUX_FILE_H
 
-#ifndef _LINUX_POSIX_TYPES_H	/* __FD_CLR */
+#include <asm/atomic.h>
 #include <linux/posix_types.h>
-#endif
-#ifndef __LINUX_COMPILER_H	/* unlikely */
 #include <linux/compiler.h>
-#endif
 
 /*
  * The default fd array needs to be at least BITS_PER_LONG,
diff --git a/include/linux/fs.h b/include/linux/fs.h
index a01f0c3b4d34..4e5de1286d87 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -262,7 +262,7 @@ struct buffer_head {
 
 	wait_queue_head_t b_wait;
 
-	struct inode *	     b_inode;
+	int b_inode;				/* will go away */
 	struct list_head     b_inode_buffers;	/* doubly linked list of inode dirty buffers */
 };
 
@@ -859,6 +859,9 @@ struct seq_file;
  * without the big kernel lock held in all filesystems.
  */
 struct super_operations {
+   	struct inode *(*alloc_inode)(struct super_block *sb);
+	void (*destroy_inode)(struct inode *);
+
 	void (*read_inode) (struct inode *);
   
   	/* reiserfs kludge.  reiserfs needs 64 bits of information to
@@ -1151,7 +1154,16 @@ static inline void mark_buffer_clean(struct buffer_head * bh)
 extern void FASTCALL(__mark_dirty(struct buffer_head *bh));
 extern void FASTCALL(__mark_buffer_dirty(struct buffer_head *bh));
 extern void FASTCALL(mark_buffer_dirty(struct buffer_head *bh));
-extern void FASTCALL(buffer_insert_inode_data_queue(struct buffer_head *, struct inode *));
+extern void FASTCALL(buffer_insert_list(struct buffer_head *, struct list_head *));
+
+static inline void buffer_insert_inode_queue(struct buffer_head *bh, struct inode *inode)
+{
+	buffer_insert_list(bh, &inode->i_dirty_buffers);
+}
+static inline void buffer_insert_inode_data_queue(struct buffer_head *bh, struct inode *inode)
+{
+	buffer_insert_list(bh, &inode->i_dirty_data_buffers);
+}
 
 #define atomic_set_buffer_dirty(bh) test_and_set_bit(BH_Dirty, &(bh)->b_state)
 
@@ -1219,10 +1231,16 @@ extern int fsync_dev(kdev_t);
 extern int fsync_super(struct super_block *);
 extern int fsync_no_super(struct block_device *);
 extern void sync_inodes_sb(struct super_block *);
-extern int osync_inode_buffers(struct inode *);
-extern int osync_inode_data_buffers(struct inode *);
-extern int fsync_inode_buffers(struct inode *);
-extern int fsync_inode_data_buffers(struct inode *);
+extern int osync_buffers_list(struct list_head *);
+extern int fsync_buffers_list(struct list_head *);
+static inline int fsync_inode_buffers(struct inode *inode)
+{
+	return fsync_buffers_list(&inode->i_dirty_buffers);
+}
+static inline int fsync_inode_data_buffers(struct inode *inode)
+{
+	return fsync_buffers_list(&inode->i_dirty_data_buffers);
+}
 extern int inode_has_buffers(struct inode *);
 extern void filemap_fdatasync(struct address_space *);
 extern void filemap_fdatawait(struct address_space *);
@@ -1327,6 +1345,7 @@ extern struct dentry * lookup_hash(struct qstr *, struct dentry *);
 #define user_path_walk(name,nd)	 __user_walk(name, LOOKUP_FOLLOW|LOOKUP_POSITIVE, nd)
 #define user_path_walk_link(name,nd) __user_walk(name, LOOKUP_POSITIVE, nd)
 
+extern void inode_init_once(struct inode *);
 extern void iput(struct inode *);
 extern void force_delete(struct inode *);
 extern struct inode * igrab(struct inode *);
@@ -1340,20 +1359,8 @@ static inline struct inode *iget(struct super_block *sb, unsigned long ino)
 }
 
 extern void clear_inode(struct inode *);
-extern struct inode * get_empty_inode(void);
-
-static inline struct inode * new_inode(struct super_block *sb)
-{
-	struct inode *inode = get_empty_inode();
-	if (inode) {
-		inode->i_sb = sb;
-		inode->i_dev = sb->s_dev;
-		inode->i_blkbits = sb->s_blocksize_bits;
-	}
-	return inode;
-}
+extern struct inode *new_inode(struct super_block *);
 extern void remove_suid(struct inode *inode);
-
 extern void insert_inode_hash(struct inode *);
 extern void remove_inode_hash(struct inode *);
 extern struct file * get_empty_filp(void);
diff --git a/include/linux/ide.h b/include/linux/ide.h
index b5e752a49e7d..8b81bcc98cb3 100644
--- a/include/linux/ide.h
+++ b/include/linux/ide.h
@@ -884,6 +884,9 @@ typedef enum {
  */
 #define ide_rq_offset(rq) (((rq)->hard_cur_sectors - (rq)->current_nr_sectors) << 9)
 
+#define task_rq_offset(rq) \
+	(((rq)->nr_sectors - (rq)->current_nr_sectors) * SECTOR_SIZE)
+
 extern inline void *ide_map_buffer(struct request *rq, unsigned long *flags)
 {
 	return bio_kmap_irq(rq->bio, flags) + ide_rq_offset(rq);
@@ -895,6 +898,24 @@ extern inline void ide_unmap_buffer(char *buffer, unsigned long *flags)
 }
 
 /*
+ * for now, taskfile requests are special :/
+ */
+extern inline char *ide_map_rq(struct request *rq, unsigned long *flags)
+{
+	if (rq->bio)
+		return ide_map_buffer(rq, flags);
+	else
+		return rq->buffer + task_rq_offset(rq);
+}
+
+extern inline void ide_unmap_rq(struct request *rq, char *buf,
+				unsigned long *flags)
+{
+	if (rq->bio)
+		ide_unmap_buffer(buf, flags);
+}
+
+/*
  * This function issues a special IDE device request
  * onto the request queue.
  *
diff --git a/include/linux/init.h b/include/linux/init.h
index f0644ca302e8..0a9c2eba880e 100644
--- a/include/linux/init.h
+++ b/include/linux/init.h
@@ -50,8 +50,26 @@ typedef void (*exitcall_t)(void);
 
 extern initcall_t __initcall_start, __initcall_end;
 
-#define __initcall(fn)								\
-	static initcall_t __initcall_##fn __init_call = fn
+/* initcalls are now grouped by functionality into separate 
+ * subsections. Ordering inside the subsections is determined
+ * by link order. 
+ * For backwards compatability, initcall() puts the call in 
+ * the device init subsection.
+ */
+
+#define __define_initcall(level,fn) \
+	static initcall_t __initcall_##fn __attribute__ ((unused,__section__ (".initcall" level ".init"))) = fn
+
+#define early_arch_initcall(fn)		__define_initcall("1",fn)
+#define mem_initcall(fn)		__define_initcall("2",fn)
+#define subsys_initcall(fn)		__define_initcall("3",fn)
+#define arch_initcall(fn)		__define_initcall("4",fn)
+#define fs_initcall(fn)			__define_initcall("5",fn)
+#define device_initcall(fn)		__define_initcall("6",fn)
+#define late_initcall(fn)		__define_initcall("7",fn)
+
+#define __initcall(fn) device_initcall(fn)
+
 #define __exitcall(fn)								\
 	static exitcall_t __exitcall_##fn __exit_call = fn
 
@@ -80,7 +98,7 @@ extern struct kernel_param __setup_start, __setup_end;
 #define __initdata	__attribute__ ((__section__ (".data.init")))
 #define __exitdata	__attribute__ ((unused, __section__ (".data.exit")))
 #define __initsetup	__attribute__ ((unused,__section__ (".setup.init")))
-#define __init_call	__attribute__ ((unused,__section__ (".initcall.init")))
+#define __init_call(level)  __attribute__ ((unused,__section__ (".initcall" level ".init")))
 #define __exit_call	__attribute__ ((unused,__section__ (".exitcall.exit")))
 
 /* For assembly routines */
@@ -141,6 +159,14 @@ typedef void (*__cleanup_module_func_t)(void);
 
 #define __setup(str,func) /* nothing */
 
+#define early_arch_initcall(fn)		module_init(fn)
+#define mem_initcall(fn)		module_init(fn)
+#define subsys_initcall(fn)		module_init(fn)
+#define arch_initcall(fn)		module_init(fn)
+#define fs_initcall(fn)			module_init(fn)
+#define device_initcall(fn)		module_init(fn)
+#define late_initcall(fn)		module_init(fn)
+
 #endif
 
 #ifdef CONFIG_HOTPLUG
diff --git a/include/linux/init_task.h b/include/linux/init_task.h
index 3c4e71fc0a60..b496169b6023 100644
--- a/include/linux/init_task.h
+++ b/include/linux/init_task.h
@@ -1,9 +1,7 @@
 #ifndef _LINUX__INIT_TASK_H
 #define _LINUX__INIT_TASK_H
 
-#ifndef __LINUX_FILE_H
 #include <linux/file.h>
-#endif
 
 #define INIT_FILES \
 { 							\
diff --git a/include/linux/isdn.h b/include/linux/isdn.h
index 6f6ae516028f..b59350f01f32 100644
--- a/include/linux/isdn.h
+++ b/include/linux/isdn.h
@@ -178,8 +178,6 @@ typedef struct {
  * the correspondent code in isdn.c
  */
 
-#define ISDN_MINOR_B        0
-#define ISDN_MINOR_BMAX     (ISDN_MAX_CHANNELS-1)
 #define ISDN_MINOR_CTRL     64
 #define ISDN_MINOR_CTRLMAX  (64 + (ISDN_MAX_CHANNELS-1))
 #define ISDN_MINOR_PPP      128
diff --git a/include/linux/mtd/cfi.h b/include/linux/mtd/cfi.h
index 7d6b31638321..0ad3d186597d 100644
--- a/include/linux/mtd/cfi.h
+++ b/include/linux/mtd/cfi.h
@@ -368,7 +368,7 @@ static inline __u8 cfi_read_query(struct map_info *map, __u32 addr)
 static inline void cfi_udelay(int us)
 {
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,2,0)
-	if (current->need_resched) {
+	if (need_resched()) {
 		unsigned long t = us * HZ / 1000000;
 		if (t < 1)
 			t = 1;
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 4971fa3c3586..7660a0bf8e9b 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -525,7 +525,6 @@ int pcibios_find_device (unsigned short vendor, unsigned short dev_id,
 
 /* Generic PCI functions used internally */
 
-void pci_init(void);
 int pci_bus_exists(const struct list_head *list, int nr);
 struct pci_bus *pci_scan_bus(int bus, struct pci_ops *ops, void *sysdata);
 struct pci_bus *pci_alloc_primary_bus(int bus);
diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h
index a78f78986723..73d6d823fbbb 100644
--- a/include/linux/reiserfs_fs_sb.h
+++ b/include/linux/reiserfs_fs_sb.h
@@ -300,7 +300,7 @@ struct reiserfs_journal {
   int j_free_bitmap_nodes ;
   int j_used_bitmap_nodes ;
   struct list_head j_bitmap_nodes ;
-  struct inode j_dummy_inode ;
+  struct list_head j_dirty_buffers ;
   struct reiserfs_list_bitmap j_list_bitmap[JOURNAL_NUM_BITMAPS] ;	/* array of bitmaps to record the deleted blocks */
   struct reiserfs_journal_list j_journal_list[JOURNAL_LIST_COUNT] ;	    /* array of all the journal lists */
   struct reiserfs_journal_cnode *j_hash_table[JOURNAL_HASH_SIZE] ; 	    /* hash table for real buffer heads in current trans */ 
diff --git a/include/linux/sched.h b/include/linux/sched.h
index b8d564fd0c08..3797423a4fbe 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -25,6 +25,7 @@ extern unsigned long event;
 #include <linux/signal.h>
 #include <linux/securebits.h>
 #include <linux/fs_struct.h>
+#include <linux/compiler.h>
 
 struct exec_domain;
 
@@ -136,6 +137,8 @@ struct completion;
 extern rwlock_t tasklist_lock;
 extern spinlock_t mmlist_lock;
 
+typedef struct task_struct task_t;
+
 extern void sched_init(void);
 extern void init_idle(void);
 extern void show_state(void);
@@ -144,8 +147,7 @@ extern void trap_init(void);
 extern void update_process_times(int user);
 extern void update_one_process(struct task_struct *p, unsigned long user,
 			       unsigned long system, int cpu);
-extern void expire_task(struct task_struct *p);
-extern void idle_tick(void);
+extern void scheduler_tick(struct task_struct *p);
 
 #define	MAX_SCHEDULE_TIMEOUT	LONG_MAX
 extern signed long FASTCALL(schedule_timeout(signed long timeout));
@@ -221,7 +223,6 @@ struct user_struct {
 extern struct user_struct root_user;
 #define INIT_USER (&root_user)
 
-typedef struct task_struct task_t;
 typedef struct prio_array prio_array_t;
 
 struct task_struct {
@@ -251,7 +252,10 @@ struct task_struct {
 	prio_array_t *array;
 
 	unsigned int time_slice;
-	unsigned long sleep_jtime;
+
+	#define MAX_SLEEP_AVG (2*HZ)
+	unsigned long sleep_avg;
+	unsigned long sleep_timestamp;
 
 	unsigned long policy;
 	unsigned long cpus_allowed;
@@ -390,30 +394,47 @@ struct task_struct {
  * them out at 128 to make it easier to search the
  * scheduler bitmap.
  */
-#define MAX_RT_PRIO	128
+#define MAX_RT_PRIO		128
 /*
  * The lower the priority of a process, the more likely it is
  * to run. Priority of a process goes from 0 to 167. The 0-99
  * priority range is allocated to RT tasks, the 128-167 range
  * is for SCHED_OTHER tasks.
  */
-#define MAX_PRIO	(MAX_RT_PRIO+40)
-#define DEF_USER_NICE	0
+#define MAX_PRIO		(MAX_RT_PRIO + 40)
 
 /*
- * Default timeslice is 80 msecs, maximum is 160 msecs.
+ * Scales user-nice values [ -20 ... 0 ... 19 ]
+ * to static priority [ 128 ... 167 (MAX_PRIO-1) ]
+ *
+ * User-nice value of -20 == static priority 128, and
+ * user-nice value 19 == static priority 167. The lower
+ * the priority value, the higher the task's priority.
+ */
+#define NICE_TO_PRIO(n)		(MAX_RT_PRIO + (n) + 20)
+#define DEF_USER_NICE		0
+
+/*
+ * Default timeslice is 90 msecs, maximum is 180 msecs.
  * Minimum timeslice is 10 msecs.
  */
-#define MIN_TIMESLICE	(10 * HZ / 1000)
-#define MAX_TIMESLICE	(160 * HZ / 1000)
+#define MIN_TIMESLICE		( 10 * HZ / 1000)
+#define MAX_TIMESLICE		(180 * HZ / 1000)
 
-#define USER_PRIO(p) ((p)-MAX_RT_PRIO)
-#define MAX_USER_PRIO (USER_PRIO(MAX_PRIO))
-#define DEF_PRIO	(MAX_RT_PRIO + MAX_USER_PRIO / 3)
-#define NICE_TO_PRIO(n) (MAX_PRIO-1 + (n) - 19)
+#define USER_PRIO(p)		((p)-MAX_RT_PRIO)
+#define MAX_USER_PRIO		(USER_PRIO(MAX_PRIO))
 
-#define NICE_TO_TIMESLICE(n)   (MIN_TIMESLICE + \
-	((MAX_TIMESLICE - MIN_TIMESLICE) * (19 - (n))) / 39)
+/*
+ * NICE_TO_TIMESLICE scales nice values [ -20 ... 19 ]
+ * to time slice values.
+ *
+ * The higher a process's priority, the bigger timeslices
+ * it gets during one round of execution. But even the lowest
+ * priority process gets MIN_TIMESLICE worth of execution time.
+ */
+
+#define NICE_TO_TIMESLICE(n) (MIN_TIMESLICE + \
+	((MAX_TIMESLICE - MIN_TIMESLICE) * (19-(n))) / 39)
 
 extern void set_cpus_allowed(task_t *p, unsigned long new_mask);
 extern void set_user_nice(task_t *p, long nice);
@@ -542,6 +563,17 @@ static inline int signal_pending(struct task_struct *p)
 {
 	return (p->sigpending != 0);
 }
+  
+static inline int need_resched(void)
+{
+	return unlikely(current->need_resched != 0);
+}
+
+static inline void cond_resched(void)
+{
+	if (need_resched())
+		schedule();
+}
 
 /*
  * Re-calculate pending state from the set of locally pending
diff --git a/include/scsi/sg.h b/include/scsi/sg.h
index 26ee7fea25d2..20624abce6c7 100644
--- a/include/scsi/sg.h
+++ b/include/scsi/sg.h
@@ -11,78 +11,71 @@ Original driver (sg.h):
 Version 2 and 3 extensions to driver:
 *       Copyright (C) 1998 - 2001 Douglas Gilbert
 
-    Version: 3.1.20 (20010814)
-    This version is for 2.4 series kernels.
-
-    Changes since 3.1.19 (20010623)
-	- add SG_GET_ACCESS_COUNT ioctl 
-	- make open() increment and close() decrement access_count
-	- only register first 256 devices, reject subsequent devices
-    Changes since 3.1.18 (20010505)
-	- fix bug that caused long wait when large buffer requested
-	- fix leak in error case of sg_new_read() [report: Eric Barton]
-	- add 'online' column to /proc/scsi/sg/devices
-    Changes since 3.1.17 (20000921)
-    	- add CAP_SYS_RAWIO capability for sensitive stuff
-    	- compile in dio stuff, procfs 'allow_dio' defaulted off (0)
-	- make premature close and detach more robust
-	- lun masked into commands <= SCSI_2
-	- poll() and async notification now yield POLL_HUP on detach
-	- various 3rd party tweaks tracking lk 2.4 internal changes
+    Version: 3.5.23 (20011231)
+    This version is for 2.5 series kernels.
+
+    Changes since 3.1.22 (20011208)
+    	- branch sg driver for lk 2.5 series
+    	- remove lock_kernel() from sg_close()
+    	- remove code based on scsi mid level dma pool
+    	- change scatterlist 'address' to use page + offset
+    	- add SG_INTERFACE_ID_ORIG
 
 Map of SG verions to the Linux kernels in which they appear:
        ----------        ----------------------------------
        original          all kernels < 2.2.6
-       2.1.38            2.2.16
-       2.1.39            2.2.17 - 2.2.19
+       2.1.40            2.2.20
        3.0.x             optional version 3 sg driver for 2.2 series
-       3.1.17            2.4.0 ++
+       3.1.17++          2.4.0++
+       3.5.23++          2.5.0++
 
 Major new features in SG 3.x driver (cf SG 2.x drivers)
 	- SG_IO ioctl() combines function if write() and read()
 	- new interface (sg_io_hdr_t) but still supports old interface
-	- scatter/gather in user space and direct IO supported
-
- The term "indirect IO" refers a method by which data is DMAed into kernel
- buffers from the hardware and afterwards is transferred into the user
- space (or vice versa if you are writing). Transfer speeds of up to 20 to
- 30MBytes/sec have been measured using indirect IO. For faster throughputs
- "direct IO" which cuts out the double handling of data is required.
- Direct IO is supported by the SG 3.x drivers on 2.4 series Linux kernels
- and requires the use of the new interface.
-
- Requests for direct IO with the new interface will automatically fall back
- to indirect IO mode if they cannot be fulfilled. An example of such a case
- is an ISA SCSI adapter which is only capable of DMAing to the lower 16MB of
- memory due to the architecture of ISA. The 'info' field in the new
- interface indicates whether a direct or indirect data transfer took place.
-
- Obtaining memory for the kernel buffers used in indirect IO is done by
- first checking if the "reserved buffer" for the current file descriptor
- is available and large enough. If these conditions are _not_ met then
- kernel memory is obtained on a per SCSI command basis. This corresponds
- to a write(), read() sequence or a SG_IO ioctl() call. Further, the
- kernel memory that is suitable for DMA may be constrained by the
- architecture of the SCSI adapter (e.g. ISA adapters).
+	- scatter/gather in user space, direct IO, and mmap supported
+
+ The normal action of this driver is to use the adapter (HBA) driver to DMA
+ data into kernel buffers and then use the CPU to copy the data into the 
+ user space (vice versa for writes). That is called "indirect" IO due to 
+ the double handling of data. There are two methods offered to remove the
+ redundant copy: 1) direct IO which uses the kernel kiobuf mechanism and 
+ 2) using the mmap() system call to map the reserve buffer (this driver has 
+ one reserve buffer per fd) into the user space. Both have their advantages.
+ In terms of absolute speed mmap() is faster. If speed is not a concern, 
+ indirect IO should be fine. Read the documentation for more information.
 
  ** N.B. To use direct IO 'echo 1 > /proc/scsi/sg/allow_dio' may be
          needed. That pseudo file's content is defaulted to 0. **
+ 
+ Historical note: this SCSI pass-through driver has been known as "sg" for 
+ a decade. In broader kernel discussions "sg" is used to refer to scatter
+ gather techniques. The context should clarify which "sg" is referred to.
 
  Documentation
  =============
- A web site for SG device drivers can be found at:
+ A web site for the SG device driver can be found at:
 	http://www.torque.net/sg  [alternatively check the MAINTAINERS file]
- The main documents are still based on 2.x versions:
+ The documentation for the sg version 3 driver can be found at:
+ 	http://www.torque.net/sg/p/sg_v3_ho.html
+ This is a rendering from DocBook source [change the extension to "sgml"
+ or "xml"]. There are renderings in "ps", "pdf", "rtf" and "txt" (soon).
+
+ The older, version 2 documents discuss the original sg interface in detail:
 	http://www.torque.net/sg/p/scsi-generic.txt
 	http://www.torque.net/sg/p/scsi-generic_long.txt
- Documentation on the changes and additions in 3.x version of the sg driver
- can be found at: http://www.torque.net/sg/p/scsi-generic_v3.txt
  A version of this document (potentially out of date) may also be found in
  the kernel source tree, probably at:
         /usr/src/linux/Documentation/scsi-generic.txt .
- Utility and test programs are available at the sg web site.
+
+ Utility and test programs are available at the sg web site. They are 
+ bundled as sg_utils (for the lk 2.2 series) and sg3_utils (for the
+ lk 2.4 series).
+
+ There is a HOWTO on the Linux SCSI subsystem in the lk 2.4 series at:
+ 	http://www.linuxdoc.org/HOWTO/SCSI-2.4-HOWTO
 */
 
+
 /* New interface introduced in the 3.x SG drivers follows */
 
 typedef struct sg_iovec /* same structure as used by readv() Linux system */
@@ -119,20 +112,23 @@ typedef struct sg_io_hdr
     unsigned int info;          /* [o] auxiliary information */
 } sg_io_hdr_t;  /* 64 bytes long (on i386) */
 
+#define SG_INTERFACE_ID_ORIG 'S'
+
 /* Use negative values to flag difference from original sg_header structure */
-#define SG_DXFER_NONE -1        /* e.g. a SCSI Test Unit Ready command */
-#define SG_DXFER_TO_DEV -2      /* e.g. a SCSI WRITE command */
-#define SG_DXFER_FROM_DEV -3    /* e.g. a SCSI READ command */
-#define SG_DXFER_TO_FROM_DEV -4 /* treated like SG_DXFER_FROM_DEV with the
+#define SG_DXFER_NONE (-1)      /* e.g. a SCSI Test Unit Ready command */
+#define SG_DXFER_TO_DEV (-2)    /* e.g. a SCSI WRITE command */
+#define SG_DXFER_FROM_DEV (-3)  /* e.g. a SCSI READ command */
+#define SG_DXFER_TO_FROM_DEV (-4) /* treated like SG_DXFER_FROM_DEV with the
 				   additional property than during indirect
 				   IO the user buffer is copied into the
 				   kernel buffers before the transfer */
-#define SG_DXFER_UNKNOWN -5     /* Unknown data direction */
+#define SG_DXFER_UNKNOWN (-5)   /* Unknown data direction */
 
 /* following flag values can be "or"-ed together */
 #define SG_FLAG_DIRECT_IO 1     /* default is indirect IO */
-#define SG_FLAG_LUN_INHIBIT 2   /* default is to put device's lun into */
-				/* the 2nd byte of SCSI command */
+#define SG_FLAG_LUN_INHIBIT 2   /* default is overwrite lun in SCSI */
+				/* command block (when <= SCSI_2) */
+#define SG_FLAG_MMAP_IO 4       /* request memory mapped IO */
 #define SG_FLAG_NO_DXFER 0x10000 /* no transfer of kernel buffers to/from */
 				/* user space (debug indirect IO) */