Merge linux-2.6 into linux-acpi-2.6 test

22 files changed, 1046 insertions, 177 deletions

diff --git a/include/asm-ia64/acpi.h b/include/asm-ia64/acpi.h
index 4c06d455139c..3a544ffc5008 100644
--- a/include/asm-ia64/acpi.h
+++ b/include/asm-ia64/acpi.h
@@ -116,6 +116,11 @@ extern int __initdata nid_to_pxm_map[MAX_NUMNODES];
 
 extern u16 ia64_acpiid_to_sapicid[];
 
+/*
+ * Refer Intel ACPI _PDC support document for bit definitions
+ */
+#define ACPI_PDC_EST_CAPABILITY_SMP     0x8
+
 #endif /*__KERNEL__*/
 
 #endif /*_ASM_ACPI_H*/
diff --git a/include/asm-ia64/fcntl.h b/include/asm-ia64/fcntl.h
index c9f8d835d0cc..cee16ea1780a 100644
--- a/include/asm-ia64/fcntl.h
+++ b/include/asm-ia64/fcntl.h
@@ -81,6 +81,7 @@ struct flock {
 
 #define F_LINUX_SPECIFIC_BASE	1024
 
-#define force_o_largefile() ( ! (current->personality & PER_LINUX32) )
+#define force_o_largefile()	\
+		(personality(current->personality) != PER_LINUX32)
 
 #endif /* _ASM_IA64_FCNTL_H */
diff --git a/include/asm-ia64/io.h b/include/asm-ia64/io.h
index 54e7637a326c..cf772a67f858 100644
--- a/include/asm-ia64/io.h
+++ b/include/asm-ia64/io.h
@@ -23,7 +23,7 @@
 #define __SLOW_DOWN_IO	do { } while (0)
 #define SLOW_DOWN_IO	do { } while (0)
 
-#define __IA64_UNCACHED_OFFSET	0xc000000000000000UL	/* region 6 */
+#define __IA64_UNCACHED_OFFSET	RGN_BASE(RGN_UNCACHED)
 
 /*
  * The legacy I/O space defined by the ia64 architecture supports only 65536 ports, but
@@ -41,7 +41,7 @@
 #define IO_SPACE_BASE(space)		((space) << IO_SPACE_BITS)
 #define IO_SPACE_PORT(port)		((port) & (IO_SPACE_SIZE - 1))
 
-#define IO_SPACE_SPARSE_ENCODING(p)	((((p) >> 2) << 12) | (p & 0xfff))
+#define IO_SPACE_SPARSE_ENCODING(p)	((((p) >> 2) << 12) | ((p) & 0xfff))
 
 struct io_space {
 	unsigned long mmio_base;	/* base in MMIO space */
diff --git a/include/asm-ia64/mmu.h b/include/asm-ia64/mmu.h
index ae1525352a25..611432ba579c 100644
--- a/include/asm-ia64/mmu.h
+++ b/include/asm-ia64/mmu.h
@@ -2,10 +2,12 @@
 #define __MMU_H
 
 /*
- * Type for a context number.  We declare it volatile to ensure proper ordering when it's
- * accessed outside of spinlock'd critical sections (e.g., as done in activate_mm() and
- * init_new_context()).
+ * Type for a context number.  We declare it volatile to ensure proper
+ * ordering when it's accessed outside of spinlock'd critical sections
+ * (e.g., as done in activate_mm() and init_new_context()).
  */
 typedef volatile unsigned long mm_context_t;
 
+typedef unsigned long nv_mm_context_t;
+
 #endif
diff --git a/include/asm-ia64/mmu_context.h b/include/asm-ia64/mmu_context.h
index e3e5fededb04..8d6e72f7b08e 100644
--- a/include/asm-ia64/mmu_context.h
+++ b/include/asm-ia64/mmu_context.h
@@ -19,6 +19,7 @@
 
 #define ia64_rid(ctx,addr)	(((ctx) << 3) | (addr >> 61))
 
+# include <asm/page.h>
 # ifndef __ASSEMBLY__
 
 #include <linux/compiler.h>
@@ -55,34 +56,46 @@ static inline void
 delayed_tlb_flush (void)
 {
 	extern void local_flush_tlb_all (void);
+	unsigned long flags;
 
 	if (unlikely(__ia64_per_cpu_var(ia64_need_tlb_flush))) {
-		local_flush_tlb_all();
-		__ia64_per_cpu_var(ia64_need_tlb_flush) = 0;
+		spin_lock_irqsave(&ia64_ctx.lock, flags);
+		{
+			if (__ia64_per_cpu_var(ia64_need_tlb_flush)) {
+				local_flush_tlb_all();
+				__ia64_per_cpu_var(ia64_need_tlb_flush) = 0;
+			}
+		}
+		spin_unlock_irqrestore(&ia64_ctx.lock, flags);
 	}
 }
 
-static inline mm_context_t
+static inline nv_mm_context_t
 get_mmu_context (struct mm_struct *mm)
 {
 	unsigned long flags;
-	mm_context_t context = mm->context;
-
-	if (context)
-		return context;
-
-	spin_lock_irqsave(&ia64_ctx.lock, flags);
-	{
-		/* re-check, now that we've got the lock: */
-		context = mm->context;
-		if (context == 0) {
-			cpus_clear(mm->cpu_vm_mask);
-			if (ia64_ctx.next >= ia64_ctx.limit)
-				wrap_mmu_context(mm);
-			mm->context = context = ia64_ctx.next++;
+	nv_mm_context_t context = mm->context;
+
+	if (unlikely(!context)) {
+		spin_lock_irqsave(&ia64_ctx.lock, flags);
+		{
+			/* re-check, now that we've got the lock: */
+			context = mm->context;
+			if (context == 0) {
+				cpus_clear(mm->cpu_vm_mask);
+				if (ia64_ctx.next >= ia64_ctx.limit)
+					wrap_mmu_context(mm);
+				mm->context = context = ia64_ctx.next++;
+			}
 		}
+		spin_unlock_irqrestore(&ia64_ctx.lock, flags);
 	}
-	spin_unlock_irqrestore(&ia64_ctx.lock, flags);
+	/*
+	 * Ensure we're not starting to use "context" before any old
+	 * uses of it are gone from our TLB.
+	 */
+	delayed_tlb_flush();
+
 	return context;
 }
 
@@ -104,13 +117,13 @@ destroy_context (struct mm_struct *mm)
 }
 
 static inline void
-reload_context (mm_context_t context)
+reload_context (nv_mm_context_t context)
 {
 	unsigned long rid;
 	unsigned long rid_incr = 0;
 	unsigned long rr0, rr1, rr2, rr3, rr4, old_rr4;
 
-	old_rr4 = ia64_get_rr(0x8000000000000000UL);
+	old_rr4 = ia64_get_rr(RGN_BASE(RGN_HPAGE));
 	rid = context << 3;	/* make space for encoding the region number */
 	rid_incr = 1 << 8;
 
@@ -122,6 +135,10 @@ reload_context (mm_context_t context)
 	rr4 = rr0 + 4*rid_incr;
 #ifdef  CONFIG_HUGETLB_PAGE
 	rr4 = (rr4 & (~(0xfcUL))) | (old_rr4 & 0xfc);
+
+#  if RGN_HPAGE != 4
+#    error "reload_context assumes RGN_HPAGE is 4"
+#  endif
 #endif
 
 	ia64_set_rr(0x0000000000000000UL, rr0);
@@ -138,7 +155,7 @@ reload_context (mm_context_t context)
 static inline void
 activate_context (struct mm_struct *mm)
 {
-	mm_context_t context;
+	nv_mm_context_t context;
 
 	do {
 		context = get_mmu_context(mm);
@@ -157,8 +174,6 @@ activate_context (struct mm_struct *mm)
 static inline void
 activate_mm (struct mm_struct *prev, struct mm_struct *next)
 {
-	delayed_tlb_flush();
-
 	/*
 	 * We may get interrupts here, but that's OK because interrupt handlers cannot
 	 * touch user-space.
diff --git a/include/asm-ia64/page.h b/include/asm-ia64/page.h
index 08894f73abf0..9edffad8c28b 100644
--- a/include/asm-ia64/page.h
+++ b/include/asm-ia64/page.h
@@ -13,6 +13,19 @@
 #include <asm/types.h>
 
 /*
+ * The top three bits of an IA64 address are its Region Number.
+ * Different regions are assigned to different purposes.
+ */
+#define RGN_SHIFT	(61)
+#define RGN_BASE(r)	(__IA64_UL_CONST(r)<<RGN_SHIFT)
+#define RGN_BITS	(RGN_BASE(-1))
+
+#define RGN_KERNEL	7	/* Identity mapped region */
+#define RGN_UNCACHED    6	/* Identity mapped I/O region */
+#define RGN_GATE	5	/* Gate page, Kernel text, etc */
+#define RGN_HPAGE	4	/* For Huge TLB pages */
+
+/*
  * PAGE_SHIFT determines the actual kernel page size.
  */
 #if defined(CONFIG_IA64_PAGE_SIZE_4KB)
@@ -36,10 +49,9 @@
 
 #define RGN_MAP_LIMIT	((1UL << (4*PAGE_SHIFT - 12)) - PAGE_SIZE)	/* per region addr limit */
 
+
 #ifdef CONFIG_HUGETLB_PAGE
-# define REGION_HPAGE		(4UL)	/* note: this is hardcoded in reload_context()!*/
-# define REGION_SHIFT		61
-# define HPAGE_REGION_BASE	(REGION_HPAGE << REGION_SHIFT)
+# define HPAGE_REGION_BASE	RGN_BASE(RGN_HPAGE)
 # define HPAGE_SHIFT		hpage_shift
 # define HPAGE_SHIFT_DEFAULT	28	/* check ia64 SDM for architecture supported size */
 # define HPAGE_SIZE		(__IA64_UL_CONST(1) << HPAGE_SHIFT)
@@ -130,16 +142,13 @@ typedef union ia64_va {
 #define REGION_NUMBER(x)	({ia64_va _v; _v.l = (long) (x); _v.f.reg;})
 #define REGION_OFFSET(x)	({ia64_va _v; _v.l = (long) (x); _v.f.off;})
 
-#define REGION_SIZE		REGION_NUMBER(1)
-#define REGION_KERNEL		7
-
 #ifdef CONFIG_HUGETLB_PAGE
 # define htlbpage_to_page(x)	(((unsigned long) REGION_NUMBER(x) << 61)			\
 				 | (REGION_OFFSET(x) >> (HPAGE_SHIFT-PAGE_SHIFT)))
 # define HUGETLB_PAGE_ORDER	(HPAGE_SHIFT - PAGE_SHIFT)
 # define is_hugepage_only_range(mm, addr, len)		\
-	 (REGION_NUMBER(addr) == REGION_HPAGE &&	\
-	  REGION_NUMBER((addr)+(len)-1) == REGION_HPAGE)
+	 (REGION_NUMBER(addr) == RGN_HPAGE &&	\
+	  REGION_NUMBER((addr)+(len)-1) == RGN_HPAGE)
 extern unsigned int hpage_shift;
 #endif
 
@@ -197,7 +206,7 @@ get_order (unsigned long size)
 # define __pgprot(x)	(x)
 #endif /* !STRICT_MM_TYPECHECKS */
 
-#define PAGE_OFFSET			__IA64_UL_CONST(0xe000000000000000)
+#define PAGE_OFFSET			RGN_BASE(RGN_KERNEL)
 
 #define VM_DATA_DEFAULT_FLAGS		(VM_READ | VM_WRITE |					\
 					 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC |		\
diff --git a/include/asm-ia64/pal.h b/include/asm-ia64/pal.h
index 2303a10ee595..e828377ad295 100644
--- a/include/asm-ia64/pal.h
+++ b/include/asm-ia64/pal.h
@@ -75,6 +75,8 @@
 #define PAL_CACHE_READ		259	/* read tag & data of cacheline for diagnostic testing */
 #define PAL_CACHE_WRITE		260	/* write tag & data of cacheline for diagnostic testing */
 #define PAL_VM_TR_READ		261	/* read contents of translation register */
+#define PAL_GET_PSTATE		262	/* get the current P-state */
+#define PAL_SET_PSTATE		263	/* set the P-state */
 
 #ifndef __ASSEMBLY__
 
@@ -1111,6 +1113,25 @@ ia64_pal_halt_info (pal_power_mgmt_info_u_t *power_buf)
 	return iprv.status;
 }
 
+/* Get the current P-state information */
+static inline s64
+ia64_pal_get_pstate (u64 *pstate_index)
+{
+	struct ia64_pal_retval iprv;
+	PAL_CALL_STK(iprv, PAL_GET_PSTATE, 0, 0, 0);
+	*pstate_index = iprv.v0;
+	return iprv.status;
+}
+
+/* Set the P-state */
+static inline s64
+ia64_pal_set_pstate (u64 pstate_index)
+{
+	struct ia64_pal_retval iprv;
+	PAL_CALL_STK(iprv, PAL_SET_PSTATE, pstate_index, 0, 0);
+	return iprv.status;
+}
+
 /* Cause the processor to enter LIGHT HALT state, where prefetching and execution are
  * suspended, but cache and TLB coherency is maintained.
  */
diff --git a/include/asm-ia64/pgtable.h b/include/asm-ia64/pgtable.h
index 48586e08f432..2e34c06e6777 100644
--- a/include/asm-ia64/pgtable.h
+++ b/include/asm-ia64/pgtable.h
@@ -204,21 +204,18 @@ ia64_phys_addr_valid (unsigned long addr)
 #define set_pte(ptep, pteval)	(*(ptep) = (pteval))
 #define set_pte_at(mm,addr,ptep,pteval) set_pte(ptep,pteval)
 
-#define RGN_SIZE	(1UL << 61)
-#define RGN_KERNEL	7
-
-#define VMALLOC_START		0xa000000200000000UL
+#define VMALLOC_START		(RGN_BASE(RGN_GATE) + 0x200000000UL)
 #ifdef CONFIG_VIRTUAL_MEM_MAP
-# define VMALLOC_END_INIT	(0xa000000000000000UL + (1UL << (4*PAGE_SHIFT - 9)))
+# define VMALLOC_END_INIT	(RGN_BASE(RGN_GATE) + (1UL << (4*PAGE_SHIFT - 9)))
 # define VMALLOC_END		vmalloc_end
   extern unsigned long vmalloc_end;
 #else
-# define VMALLOC_END		(0xa000000000000000UL + (1UL << (4*PAGE_SHIFT - 9)))
+# define VMALLOC_END		(RGN_BASE(RGN_GATE) + (1UL << (4*PAGE_SHIFT - 9)))
 #endif
 
 /* fs/proc/kcore.c */
-#define	kc_vaddr_to_offset(v) ((v) - 0xa000000000000000UL)
-#define	kc_offset_to_vaddr(o) ((o) + 0xa000000000000000UL)
+#define	kc_vaddr_to_offset(v) ((v) - RGN_BASE(RGN_GATE))
+#define	kc_offset_to_vaddr(o) ((o) + RGN_BASE(RGN_GATE))
 
 /*
  * Conversion functions: convert page frame number (pfn) and a protection value to a page
diff --git a/include/asm-ia64/rwsem.h b/include/asm-ia64/rwsem.h
index 6ece5061dc19..e18b5ab0cb75 100644
--- a/include/asm-ia64/rwsem.h
+++ b/include/asm-ia64/rwsem.h
@@ -3,6 +3,7 @@
  *
  * Copyright (C) 2003 Ken Chen <kenneth.w.chen@intel.com>
  * Copyright (C) 2003 Asit Mallick <asit.k.mallick@intel.com>
+ * Copyright (C) 2005 Christoph Lameter <clameter@sgi.com>
  *
  * Based on asm-i386/rwsem.h and other architecture implementation.
  *
@@ -11,9 +12,9 @@
  *
  * The lock count is initialized to 0 (no active and no waiting lockers).
  *
- * When a writer subtracts WRITE_BIAS, it'll get 0xffff0001 for the case
- * of an uncontended lock. Readers increment by 1 and see a positive value
- * when uncontended, negative if there are writers (and maybe) readers
+ * When a writer subtracts WRITE_BIAS, it'll get 0xffffffff00000001 for
+ * the case of an uncontended lock. Readers increment by 1 and see a positive
+ * value when uncontended, negative if there are writers (and maybe) readers
  * waiting (in which case it goes to sleep).
  */
 
@@ -29,7 +30,7 @@
  * the semaphore definition
  */
 struct rw_semaphore {
-	signed int		count;
+	signed long		count;
 	spinlock_t		wait_lock;
 	struct list_head	wait_list;
 #if RWSEM_DEBUG
@@ -37,10 +38,10 @@ struct rw_semaphore {
 #endif
 };
 
-#define RWSEM_UNLOCKED_VALUE		0x00000000
-#define RWSEM_ACTIVE_BIAS		0x00000001
-#define RWSEM_ACTIVE_MASK		0x0000ffff
-#define RWSEM_WAITING_BIAS		(-0x00010000)
+#define RWSEM_UNLOCKED_VALUE		__IA64_UL_CONST(0x0000000000000000)
+#define RWSEM_ACTIVE_BIAS		__IA64_UL_CONST(0x0000000000000001)
+#define RWSEM_ACTIVE_MASK		__IA64_UL_CONST(0x00000000ffffffff)
+#define RWSEM_WAITING_BIAS		-__IA64_UL_CONST(0x0000000100000000)
 #define RWSEM_ACTIVE_READ_BIAS		RWSEM_ACTIVE_BIAS
 #define RWSEM_ACTIVE_WRITE_BIAS		(RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS)
 
@@ -83,7 +84,7 @@ init_rwsem (struct rw_semaphore *sem)
 static inline void
 __down_read (struct rw_semaphore *sem)
 {
-	int result = ia64_fetchadd4_acq((unsigned int *)&sem->count, 1);
+	long result = ia64_fetchadd8_acq((unsigned long *)&sem->count, 1);
 
 	if (result < 0)
 		rwsem_down_read_failed(sem);
@@ -95,7 +96,7 @@ __down_read (struct rw_semaphore *sem)
 static inline void
 __down_write (struct rw_semaphore *sem)
 {
-	int old, new;
+	long old, new;
 
 	do {
 		old = sem->count;
@@ -112,7 +113,7 @@ __down_write (struct rw_semaphore *sem)
 static inline void
 __up_read (struct rw_semaphore *sem)
 {
-	int result = ia64_fetchadd4_rel((unsigned int *)&sem->count, -1);
+	long result = ia64_fetchadd8_rel((unsigned long *)&sem->count, -1);
 
 	if (result < 0 && (--result & RWSEM_ACTIVE_MASK) == 0)
 		rwsem_wake(sem);
@@ -124,7 +125,7 @@ __up_read (struct rw_semaphore *sem)
 static inline void
 __up_write (struct rw_semaphore *sem)
 {
-	int old, new;
+	long old, new;
 
 	do {
 		old = sem->count;
@@ -141,7 +142,7 @@ __up_write (struct rw_semaphore *sem)
 static inline int
 __down_read_trylock (struct rw_semaphore *sem)
 {
-	int tmp;
+	long tmp;
 	while ((tmp = sem->count) >= 0) {
 		if (tmp == cmpxchg_acq(&sem->count, tmp, tmp+1)) {
 			return 1;
@@ -156,7 +157,7 @@ __down_read_trylock (struct rw_semaphore *sem)
 static inline int
 __down_write_trylock (struct rw_semaphore *sem)
 {
-	int tmp = cmpxchg_acq(&sem->count, RWSEM_UNLOCKED_VALUE,
+	long tmp = cmpxchg_acq(&sem->count, RWSEM_UNLOCKED_VALUE,
 			      RWSEM_ACTIVE_WRITE_BIAS);
 	return tmp == RWSEM_UNLOCKED_VALUE;
 }
@@ -167,7 +168,7 @@ __down_write_trylock (struct rw_semaphore *sem)
 static inline void
 __downgrade_write (struct rw_semaphore *sem)
 {
-	int old, new;
+	long old, new;
 
 	do {
 		old = sem->count;
@@ -182,7 +183,7 @@ __downgrade_write (struct rw_semaphore *sem)
  * Implement atomic add functionality.  These used to be "inline" functions, but GCC v3.1
  * doesn't quite optimize this stuff right and ends up with bad calls to fetchandadd.
  */
-#define rwsem_atomic_add(delta, sem)	atomic_add(delta, (atomic_t *)(&(sem)->count))
-#define rwsem_atomic_update(delta, sem)	atomic_add_return(delta, (atomic_t *)(&(sem)->count))
+#define rwsem_atomic_add(delta, sem)	atomic64_add(delta, (atomic64_t *)(&(sem)->count))
+#define rwsem_atomic_update(delta, sem)	atomic64_add_return(delta, (atomic64_t *)(&(sem)->count))
 
 #endif /* _ASM_IA64_RWSEM_H */
diff --git a/include/asm-ia64/sn/addrs.h b/include/asm-ia64/sn/addrs.h
index 103d745dc5f2..2c32e4b77b54 100644
--- a/include/asm-ia64/sn/addrs.h
+++ b/include/asm-ia64/sn/addrs.h
@@ -3,7 +3,7 @@
  * License.  See the file "COPYING" in the main directory of this archive
  * for more details.
  *
- * Copyright (c) 1992-1999,2001-2004 Silicon Graphics, Inc. All rights reserved.
+ * Copyright (c) 1992-1999,2001-2005 Silicon Graphics, Inc. All rights reserved.
  */
 
 #ifndef _ASM_IA64_SN_ADDRS_H
@@ -65,7 +65,6 @@
 
 #define NASID_MASK              ((u64)NASID_BITMASK << NASID_SHIFT)
 #define AS_MASK			((u64)AS_BITMASK << AS_SHIFT)
-#define REGION_BITS		0xe000000000000000UL
 
 
 /*
@@ -79,38 +78,30 @@
 #define AS_CAC_SPACE		(AS_CAC_VAL << AS_SHIFT)
 
 
-/*
- * Base addresses for various address ranges.
- */
-#define CACHED			0xe000000000000000UL
-#define UNCACHED                0xc000000000000000UL
-#define UNCACHED_PHYS           0x8000000000000000UL
-
-
 /* 
  * Virtual Mode Local & Global MMR space.  
  */
 #define SH1_LOCAL_MMR_OFFSET	0x8000000000UL
 #define SH2_LOCAL_MMR_OFFSET	0x0200000000UL
 #define LOCAL_MMR_OFFSET	(is_shub2() ? SH2_LOCAL_MMR_OFFSET : SH1_LOCAL_MMR_OFFSET)
-#define LOCAL_MMR_SPACE		(UNCACHED | LOCAL_MMR_OFFSET)
-#define LOCAL_PHYS_MMR_SPACE	(UNCACHED_PHYS | LOCAL_MMR_OFFSET)
+#define LOCAL_MMR_SPACE		(__IA64_UNCACHED_OFFSET | LOCAL_MMR_OFFSET)
+#define LOCAL_PHYS_MMR_SPACE	(RGN_BASE(RGN_HPAGE) | LOCAL_MMR_OFFSET)
 
 #define SH1_GLOBAL_MMR_OFFSET	0x0800000000UL
 #define SH2_GLOBAL_MMR_OFFSET	0x0300000000UL
 #define GLOBAL_MMR_OFFSET	(is_shub2() ? SH2_GLOBAL_MMR_OFFSET : SH1_GLOBAL_MMR_OFFSET)
-#define GLOBAL_MMR_SPACE	(UNCACHED | GLOBAL_MMR_OFFSET)
+#define GLOBAL_MMR_SPACE	(__IA64_UNCACHED_OFFSET | GLOBAL_MMR_OFFSET)
 
 /*
  * Physical mode addresses
  */
-#define GLOBAL_PHYS_MMR_SPACE	(UNCACHED_PHYS | GLOBAL_MMR_OFFSET)
+#define GLOBAL_PHYS_MMR_SPACE	(RGN_BASE(RGN_HPAGE) | GLOBAL_MMR_OFFSET)
 
 
 /*
  * Clear region & AS bits.
  */
-#define TO_PHYS_MASK		(~(REGION_BITS | AS_MASK))
+#define TO_PHYS_MASK		(~(RGN_BITS | AS_MASK))
 
 
 /*
@@ -126,6 +117,7 @@
 #define GLOBAL_MMR_PHYS_ADDR(n,a) (GLOBAL_PHYS_MMR_SPACE | REMOTE_ADDR(n,a))
 #define GLOBAL_CAC_ADDR(n,a)	(CAC_BASE | REMOTE_ADDR(n,a))
 #define CHANGE_NASID(n,x)	((void *)(((u64)(x) & ~NASID_MASK) | NASID_SPACE(n)))
+#define IS_TIO_NASID(n)		((n) & 1)
 
 
 /* non-II mmr's start at top of big window space (4G) */
@@ -134,10 +126,10 @@
 /*
  * general address defines
  */
-#define CAC_BASE		(CACHED   | AS_CAC_SPACE)
-#define AMO_BASE		(UNCACHED | AS_AMO_SPACE)
-#define AMO_PHYS_BASE		(UNCACHED_PHYS | AS_AMO_SPACE)
-#define GET_BASE		(CACHED   | AS_GET_SPACE)
+#define CAC_BASE		(PAGE_OFFSET | AS_CAC_SPACE)
+#define AMO_BASE		(__IA64_UNCACHED_OFFSET | AS_AMO_SPACE)
+#define AMO_PHYS_BASE		(RGN_BASE(RGN_HPAGE) | AS_AMO_SPACE)
+#define GET_BASE		(PAGE_OFFSET | AS_GET_SPACE)
 
 /*
  * Convert Memory addresses between various addressing modes.
@@ -155,17 +147,35 @@
  *           the chiplet id is zero.  If we implement TIO-TIO dma, we might need
  *           to insert a chiplet id into this macro.  However, it is our belief
  *           right now that this chiplet id will be ICE, which is also zero.
- *           Nasid starts on bit 40.
  */
-#define PHYS_TO_TIODMA(x)	( (((u64)(NASID_GET(x))) << 40) | NODE_OFFSET(x))
-#define PHYS_TO_DMA(x)          ( (((u64)(x) & NASID_MASK) >> 2) | NODE_OFFSET(x))
+#define SH1_TIO_PHYS_TO_DMA(x) 						\
+	((((u64)(NASID_GET(x))) << 40) | NODE_OFFSET(x))
+
+#define SH2_NETWORK_BANK_OFFSET(x) 					\
+        ((u64)(x) & ((1UL << (sn_hub_info->nasid_shift - 4)) -1))
+
+#define SH2_NETWORK_BANK_SELECT(x) 					\
+        ((((u64)(x) & (0x3UL << (sn_hub_info->nasid_shift - 4)))	\
+        	>> (sn_hub_info->nasid_shift - 4)) << 36)
+
+#define SH2_NETWORK_ADDRESS(x) 						\
+	(SH2_NETWORK_BANK_OFFSET(x) | SH2_NETWORK_BANK_SELECT(x))
+
+#define SH2_TIO_PHYS_TO_DMA(x) 						\
+        (((u64)(NASID_GET(x)) << 40) | 	SH2_NETWORK_ADDRESS(x))
+
+#define PHYS_TO_TIODMA(x)						\
+	(is_shub1() ? SH1_TIO_PHYS_TO_DMA(x) : SH2_TIO_PHYS_TO_DMA(x))
+
+#define PHYS_TO_DMA(x)							\
+	((((u64)(x) & NASID_MASK) >> 2) | NODE_OFFSET(x))
 
 
 /*
  * Macros to test for address type.
  */
-#define IS_AMO_ADDRESS(x)	(((u64)(x) & (REGION_BITS | AS_MASK)) == AMO_BASE)
-#define IS_AMO_PHYS_ADDRESS(x)	(((u64)(x) & (REGION_BITS | AS_MASK)) == AMO_PHYS_BASE)
+#define IS_AMO_ADDRESS(x)	(((u64)(x) & (RGN_BITS | AS_MASK)) == AMO_BASE)
+#define IS_AMO_PHYS_ADDRESS(x)	(((u64)(x) & (RGN_BITS | AS_MASK)) == AMO_PHYS_BASE)
 
 
 /*
@@ -180,18 +190,20 @@
 #define TIO_SWIN_BASE(n, w) 		(TIO_IO_BASE(n) + \
 					    ((u64) (w) << TIO_SWIN_SIZE_BITS))
 #define NODE_IO_BASE(n)			(GLOBAL_MMR_SPACE | NASID_SPACE(n))
-#define TIO_IO_BASE(n)                  (UNCACHED | NASID_SPACE(n))
+#define TIO_IO_BASE(n)                  (__IA64_UNCACHED_OFFSET | NASID_SPACE(n))
 #define BWIN_SIZE			(1UL << BWIN_SIZE_BITS)
 #define NODE_BWIN_BASE0(n)		(NODE_IO_BASE(n) + BWIN_SIZE)
 #define NODE_BWIN_BASE(n, w)		(NODE_BWIN_BASE0(n) + ((u64) (w) << BWIN_SIZE_BITS))
 #define RAW_NODE_SWIN_BASE(n, w)	(NODE_IO_BASE(n) + ((u64) (w) << SWIN_SIZE_BITS))
 #define BWIN_WIDGET_MASK		0x7
 #define BWIN_WINDOWNUM(x)		(((x) >> BWIN_SIZE_BITS) & BWIN_WIDGET_MASK)
+#define SH1_IS_BIG_WINDOW_ADDR(x)	((x) & BWIN_TOP)
 
 #define TIO_BWIN_WINDOW_SELECT_MASK	0x7
 #define TIO_BWIN_WINDOWNUM(x)		(((x) >> TIO_BWIN_SIZE_BITS) & TIO_BWIN_WINDOW_SELECT_MASK)
 
-
+#define TIO_HWIN_SHIFT_BITS		33
+#define TIO_HWIN(x)			(NODE_OFFSET(x) >> TIO_HWIN_SHIFT_BITS)
 
 /*
  * The following definitions pertain to the IO special address
@@ -216,10 +228,6 @@
 #define TIO_SWIN_WIDGETNUM(x)		(((x)  >> TIO_SWIN_SIZE_BITS) & TIO_SWIN_WIDGET_MASK)
 
 
-#define TIO_IOSPACE_ADDR(n,x)					\
-	/* Move in the Chiplet ID for TIO Local Block MMR */	\
-	(REMOTE_ADDR(n,x) | 1UL << (NASID_SHIFT - 2))
-
 /*
  * The following macros produce the correct base virtual address for
  * the hub registers. The REMOTE_HUB_* macro produce
@@ -234,18 +242,40 @@
  *	Otherwise, the recommended approach is to use *_HUB_L() and *_HUB_S().
  *	They're always safe.
  */
+/* Shub1 TIO & MMR addressing macros */
+#define SH1_TIO_IOSPACE_ADDR(n,x)					\
+	GLOBAL_MMR_ADDR(n,x)
+
+#define SH1_REMOTE_BWIN_MMR(n,x)					\
+	GLOBAL_MMR_ADDR(n,x)
+
+#define SH1_REMOTE_SWIN_MMR(n,x)					\
+	(NODE_SWIN_BASE(n,1) + 0x800000UL + (x))
+
+#define SH1_REMOTE_MMR(n,x)						\
+	(SH1_IS_BIG_WINDOW_ADDR(x) ? SH1_REMOTE_BWIN_MMR(n,x) :		\
+	 	SH1_REMOTE_SWIN_MMR(n,x))
+
+/* Shub1 TIO & MMR addressing macros */
+#define SH2_TIO_IOSPACE_ADDR(n,x)					\
+	((__IA64_UNCACHED_OFFSET | REMOTE_ADDR(n,x) | 1UL << (NASID_SHIFT - 2)))
+
+#define SH2_REMOTE_MMR(n,x)						\
+	GLOBAL_MMR_ADDR(n,x)
+
+
+/* TIO & MMR addressing macros that work on both shub1 & shub2 */
+#define TIO_IOSPACE_ADDR(n,x)						\
+	((u64 *)(is_shub1() ? SH1_TIO_IOSPACE_ADDR(n,x) :		\
+		 SH2_TIO_IOSPACE_ADDR(n,x)))
+
+#define SH_REMOTE_MMR(n,x)						\
+	(is_shub1() ? SH1_REMOTE_MMR(n,x) : SH2_REMOTE_MMR(n,x))
+
 #define REMOTE_HUB_ADDR(n,x)						\
-	((n & 1) ?							\
-	/* TIO: */							\
-	(is_shub2() ?							\
-	/* TIO on Shub2 */						\
-	(volatile u64 *)(TIO_IOSPACE_ADDR(n,x))				\
-	: /* TIO on shub1 */						\
-	(volatile u64 *)(GLOBAL_MMR_ADDR(n,x)))				\
-									\
-	: /* SHUB1 and SHUB2 MMRs: */					\
-	(((x) & BWIN_TOP) ? ((volatile u64 *)(GLOBAL_MMR_ADDR(n,x)))	\
-	: ((volatile u64 *)(NODE_SWIN_BASE(n,1) + 0x800000 + (x)))))
+	(IS_TIO_NASID(n) ?  ((volatile u64*)TIO_IOSPACE_ADDR(n,x)) :	\
+	 ((volatile u64*)SH_REMOTE_MMR(n,x)))
+
 
 #define HUB_L(x)			(*((volatile typeof(*x) *)x))
 #define	HUB_S(x,d)			(*((volatile typeof(*x) *)x) = (d))
diff --git a/include/asm-ia64/sn/geo.h b/include/asm-ia64/sn/geo.h
index 84b254603b8d..f083c9434066 100644
--- a/include/asm-ia64/sn/geo.h
+++ b/include/asm-ia64/sn/geo.h
@@ -3,7 +3,7 @@
  * License.  See the file "COPYING" in the main directory of this archive
  * for more details.
  *
- * Copyright (C) 1992 - 1997, 2000-2004 Silicon Graphics, Inc. All rights reserved.
+ * Copyright (C) 1992 - 1997, 2000-2005 Silicon Graphics, Inc. All rights reserved.
  */
 
 #ifndef _ASM_IA64_SN_GEO_H
@@ -108,7 +108,6 @@ typedef union geoid_u {
 #define INVALID_SLAB            (slabid_t)-1
 #define INVALID_SLOT            (slotid_t)-1
 #define INVALID_MODULE          ((moduleid_t)-1)
-#define INVALID_PARTID          ((partid_t)-1)
 
 static inline slabid_t geo_slab(geoid_t g)
 {
diff --git a/include/asm-ia64/sn/intr.h b/include/asm-ia64/sn/intr.h
index e190dd4213d5..e35074f526d9 100644
--- a/include/asm-ia64/sn/intr.h
+++ b/include/asm-ia64/sn/intr.h
@@ -12,13 +12,12 @@
 #include <linux/rcupdate.h>
 
 #define SGI_UART_VECTOR		(0xe9)
-#define SGI_PCIBR_ERROR		(0x33)
 
 /* Reserved IRQs : Note, not to exceed IA64_SN2_FIRST_DEVICE_VECTOR */
 #define SGI_XPC_ACTIVATE                (0x30)
 #define SGI_II_ERROR                    (0x31)
 #define SGI_XBOW_ERROR                  (0x32)
-#define SGI_PCIBR_ERROR                 (0x33)
+#define SGI_PCIASIC_ERROR               (0x33)
 #define SGI_ACPI_SCI_INT                (0x34)
 #define SGI_TIOCA_ERROR                 (0x35)
 #define SGI_TIO_ERROR                   (0x36)
diff --git a/include/asm-ia64/sn/nodepda.h b/include/asm-ia64/sn/nodepda.h
index 7138b1eafd6b..47bb8100fd00 100644
--- a/include/asm-ia64/sn/nodepda.h
+++ b/include/asm-ia64/sn/nodepda.h
@@ -37,7 +37,6 @@ struct phys_cpuid {
 
 struct nodepda_s {
 	void 		*pdinfo;	/* Platform-dependent per-node info */
-	spinlock_t		bist_lock;
 
 	/*
 	 * The BTEs on this node are shared by the local cpus
@@ -55,6 +54,8 @@ struct nodepda_s {
 	 * Array of physical cpu identifiers. Indexed by cpuid.
 	 */
 	struct phys_cpuid	phys_cpuid[NR_CPUS];
+	spinlock_t		ptc_lock ____cacheline_aligned_in_smp;
+	spinlock_t		bist_lock;
 };
 
 typedef struct nodepda_s nodepda_t;
diff --git a/include/asm-ia64/sn/pcibus_provider_defs.h b/include/asm-ia64/sn/pcibus_provider_defs.h
index 976f5eff0539..ad0e8e8ae53f 100644
--- a/include/asm-ia64/sn/pcibus_provider_defs.h
+++ b/include/asm-ia64/sn/pcibus_provider_defs.h
@@ -18,8 +18,9 @@
 #define PCIIO_ASIC_TYPE_PIC	2
 #define PCIIO_ASIC_TYPE_TIOCP	3
 #define PCIIO_ASIC_TYPE_TIOCA	4
+#define PCIIO_ASIC_TYPE_TIOCE	5
 
-#define PCIIO_ASIC_MAX_TYPES	5
+#define PCIIO_ASIC_MAX_TYPES	6
 
 /*
  * Common pciio bus provider data.  There should be one of these as the
@@ -30,7 +31,8 @@
 struct pcibus_bussoft {
 	uint32_t		bs_asic_type;	/* chipset type */
 	uint32_t		bs_xid;		/* xwidget id */
-	uint64_t		bs_persist_busnum; /* Persistent Bus Number */
+	uint32_t		bs_persist_busnum; /* Persistent Bus Number */
+	uint32_t		bs_persist_segment; /* Segment Number */
 	uint64_t		bs_legacy_io;	/* legacy io pio addr */
 	uint64_t		bs_legacy_mem;	/* legacy mem pio addr */
 	uint64_t		bs_base;	/* widget base */
@@ -47,6 +49,8 @@ struct sn_pcibus_provider {
 	dma_addr_t	(*dma_map_consistent)(struct pci_dev *, unsigned long, size_t);
 	void		(*dma_unmap)(struct pci_dev *, dma_addr_t, int);
 	void *		(*bus_fixup)(struct pcibus_bussoft *, struct pci_controller *);
+ 	void		(*force_interrupt)(struct sn_irq_info *);
+ 	void		(*target_interrupt)(struct sn_irq_info *);
 };
 
 extern struct sn_pcibus_provider *sn_pci_provider[];
diff --git a/include/asm-ia64/sn/pda.h b/include/asm-ia64/sn/pda.h
index ea5590c76ca4..1c5108d44d8b 100644
--- a/include/asm-ia64/sn/pda.h
+++ b/include/asm-ia64/sn/pda.h
@@ -39,7 +39,6 @@ typedef struct pda_s {
 	unsigned long pio_write_status_val;
 	volatile unsigned long *pio_shub_war_cam_addr;
 
-	unsigned long	sn_soft_irr[4];
 	unsigned long	sn_in_service_ivecs[4];
 	int		sn_lb_int_war_ticks;
 	int		sn_last_irq;
diff --git a/include/asm-ia64/sn/sn2/sn_hwperf.h b/include/asm-ia64/sn/sn2/sn_hwperf.h
index df75f4c4aec3..291ef3d69da2 100644
--- a/include/asm-ia64/sn/sn2/sn_hwperf.h
+++ b/include/asm-ia64/sn/sn2/sn_hwperf.h
@@ -43,6 +43,7 @@ struct sn_hwperf_object_info {
 
 /* macros for object classification */
 #define SN_HWPERF_IS_NODE(x)		((x) && strstr((x)->name, "SHub"))
+#define SN_HWPERF_IS_NODE_SHUB2(x)	((x) && strstr((x)->name, "SHub 2."))
 #define SN_HWPERF_IS_IONODE(x)		((x) && strstr((x)->name, "TIO"))
 #define SN_HWPERF_IS_ROUTER(x)		((x) && strstr((x)->name, "Router"))
 #define SN_HWPERF_IS_NL3ROUTER(x)	((x) && strstr((x)->name, "NL3Router"))
@@ -214,6 +215,15 @@ struct sn_hwperf_ioctl_args {
  */
 #define SN_HWPERF_GET_NODE_NASID	(102|SN_HWPERF_OP_MEM_COPYOUT)
 
+/*
+ * Given a node id, determine the id of the nearest node with CPUs
+ * and the id of the nearest node that has memory. The argument
+ * node would normally be a "headless" node, e.g. an "IO node".
+ * Return 0 on success.
+ */
+extern int sn_hwperf_get_nearest_node(cnodeid_t node,
+	cnodeid_t *near_mem, cnodeid_t *near_cpu);
+
 /* return codes */
 #define SN_HWPERF_OP_OK			0
 #define SN_HWPERF_OP_NOMEM		1
diff --git a/include/asm-ia64/sn/sn_sal.h b/include/asm-ia64/sn/sn_sal.h
index 27976d223186..e67825ad1930 100644
--- a/include/asm-ia64/sn/sn_sal.h
+++ b/include/asm-ia64/sn/sn_sal.h
@@ -55,7 +55,6 @@
 #define  SN_SAL_BUS_CONFIG		   	   0x02000037
 #define  SN_SAL_SYS_SERIAL_GET			   0x02000038
 #define  SN_SAL_PARTITION_SERIAL_GET		   0x02000039
-#define  SN_SAL_SYSCTL_PARTITION_GET		   0x0200003a
 #define  SN_SAL_SYSTEM_POWER_DOWN		   0x0200003b
 #define  SN_SAL_GET_MASTER_BASEIO_NASID		   0x0200003c
 #define  SN_SAL_COHERENCE                          0x0200003d
@@ -78,7 +77,8 @@
 
 #define SN_SAL_HUB_ERROR_INTERRUPT		   0x02000060
 #define SN_SAL_BTE_RECOVER			   0x02000061
-#define SN_SAL_IOIF_GET_PCI_TOPOLOGY	           0x02000062
+#define SN_SAL_RESERVED_DO_NOT_USE		   0x02000062
+#define SN_SAL_IOIF_GET_PCI_TOPOLOGY		   0x02000064
 
 /*
  * Service-specific constants
@@ -586,35 +586,6 @@ sn_partition_serial_number_val(void) {
 }
 
 /*
- * Returns the partition id of the nasid passed in as an argument,
- * or IN