Merge wailua.hpl.hp.com:/bk/vanilla/linux-2.5

into wailua.hpl.hp.com:/bk/lia64/to-linus-2.5

arch/ia64/Config.help

@@ -557,3 +557,8 @@ CONFIG_DEBUG_SPINLOCK
   best used in conjunction with the NMI watchdog so that spinlock
   deadlocks are also debuggable.
 
+CONFIG_IA64_GRANULE_16MB
+  IA64 identity-mapped regions use a large page size called "granules".
+
+  Select "16MB" for a small granule size.
+  Select "64MB" for a large granule size.  This is the current default.

arch/ia64/config.in

@@ -83,7 +83,7 @@ fi
 define_bool CONFIG_KCORE_ELF y	# On IA-64, we always want an ELF /proc/kcore.
 
 bool 'SMP support' CONFIG_SMP
-tristate 'Support running of Linux/x86 binaries' CONFIG_IA32_SUPPORT
+bool 'Support running of Linux/x86 binaries' CONFIG_IA32_SUPPORT
 bool 'Performance monitor support' CONFIG_PERFMON
 tristate '/proc/pal support' CONFIG_IA64_PALINFO
 tristate '/proc/efi/vars support' CONFIG_EFI_VARS
@@ -123,6 +123,7 @@ source drivers/block/Config.in
 source drivers/ieee1394/Config.in
 source drivers/message/i2o/Config.in
 source drivers/md/Config.in
+source drivers/message/fusion/Config.in
 
 mainmenu_option next_comment
 comment 'ATA/IDE/MFM/RLL support'

arch/ia64/hp/zx1/hpzx1_misc.c

@@ -42,7 +42,7 @@ struct fake_pci_dev {
 
 static struct fake_pci_dev *fake_pci_head, **fake_pci_tail = &fake_pci_head;
 
-static struct pci_ops orig_pci_ops;
+static struct pci_ops *orig_pci_ops;
 
 static inline struct fake_pci_dev *
 fake_pci_find_slot(unsigned char bus, unsigned int devfn)
@@ -77,7 +77,7 @@ static int hp_cfg_read##sz (struct pci_dev *dev, int where, u##bits *value) \
 { \
 	struct fake_pci_dev *fake_dev; \
 	if (!(fake_dev = fake_pci_find_slot(dev->bus->number, dev->devfn))) \
-		return orig_pci_ops.name(dev, where, value); \
+		return orig_pci_ops->name(dev, where, value); \
 	\
 	switch (where) { \
 	case PCI_COMMAND: \
@@ -105,7 +105,7 @@ static int hp_cfg_write##sz (struct pci_dev *dev, int where, u##bits value) \
 { \
 	struct fake_pci_dev *fake_dev; \
 	if (!(fake_dev = fake_pci_find_slot(dev->bus->number, dev->devfn))) \
-		return orig_pci_ops.name(dev, where, value); \
+		return orig_pci_ops->name(dev, where, value); \
 	\
 	switch (where) { \
 	case PCI_BASE_ADDRESS_0: \
@@ -295,7 +295,7 @@ hpzx1_lba_probe(acpi_handle obj, u32 depth, void *context, void **ret)
 	if (status != AE_OK)
 		return status;
 
-	status = acpi_cf_evaluate_method(obj, METHOD_NAME__BBN, &busnum);
+	status = acpi_evaluate_integer(obj, METHOD_NAME__BBN, NULL, &busnum);
 	if (ACPI_FAILURE(status)) {
 		printk(KERN_ERR PFX "evaluate _BBN fail=0x%x\n", status);
 		busnum = 0;	// no _BBN; stick it on bus 0
@@ -313,7 +313,7 @@ hpzx1_lba_probe(acpi_handle obj, u32 depth, void *context, void **ret)
 static void
 hpzx1_acpi_dev_init(void)
 {
-	extern struct pci_ops pci_conf;
+	extern struct pci_ops *pci_root_ops;
 
 	/*
 	 * Make fake PCI devices for the following hardware in the
@@ -383,8 +383,8 @@ hpzx1_acpi_dev_init(void)
 	/*
 	 * Replace PCI ops, but only if we made fake devices.
 	 */
-	orig_pci_ops = pci_conf;
-	pci_conf = hp_pci_conf;
+	orig_pci_ops = pci_root_ops;
+	pci_root_ops = &hp_pci_conf;
 }
 
 extern void sba_init(void);

arch/ia64/kernel/acpi.c

@@ -56,18 +56,43 @@ asm (".weak iosapic_version");
 void (*pm_idle) (void);
 void (*pm_power_off) (void);
 
-
-/*
- * TBD: Should go away once we have an ACPI parser.
- */
 const char *
 acpi_get_sysname (void)
 {
 #ifdef CONFIG_IA64_GENERIC
-	return "hpsim";
+	unsigned long rsdp_phys = 0;
+	struct acpi20_table_rsdp *rsdp;
+	struct acpi_table_xsdt *xsdt;
+	struct acpi_table_header *hdr;
+
+	if ((0 != acpi_find_rsdp(&rsdp_phys)) || !rsdp_phys) {
+		printk("ACPI 2.0 RSDP not found, default to \"dig\"\n");
+		return "dig";
+	}
+
+	rsdp = (struct acpi20_table_rsdp *) __va(rsdp_phys);
+	if (strncmp(rsdp->signature, RSDP_SIG, sizeof(RSDP_SIG) - 1)) {
+		printk("ACPI 2.0 RSDP signature incorrect, default to \"dig\"\n");
+		return "dig";
+	}
+
+	xsdt = (struct acpi_table_xsdt *) __va(rsdp->xsdt_address);
+	hdr = &xsdt->header;
+	if (strncmp(hdr->signature, XSDT_SIG, sizeof(XSDT_SIG) - 1)) {
+		printk("ACPI 2.0 XSDT signature incorrect, default to \"dig\"\n");
+		return "dig";
+	}
+
+	if (!strcmp(hdr->oem_id, "HP")) {
+		return "hpzx1";
+	}
+
+	return "dig";
 #else
 # if defined (CONFIG_IA64_HP_SIM)
 	return "hpsim";
+# elif defined (CONFIG_IA64_HP_ZX1)
+	return "hpzx1";
 # elif defined (CONFIG_IA64_SGI_SN1)
 	return "sn1";
 # elif defined (CONFIG_IA64_SGI_SN2)
@@ -80,6 +105,69 @@ acpi_get_sysname (void)
 #endif
 }
 
+#ifdef CONFIG_ACPI
+
+/**
+ * acpi_get_crs - Return the current resource settings for a device
+ * obj: A handle for this device
+ * buf: A buffer to be populated by this call.
+ *
+ * Pass a valid handle, typically obtained by walking the namespace and a
+ * pointer to an allocated buffer, and this function will fill in the buffer
+ * with a list of acpi_resource structures.
+ */
+acpi_status
+acpi_get_crs (acpi_handle obj, acpi_buffer *buf)
+{
+	acpi_status result;
+	buf->length = 0;
+	buf->pointer = NULL;
+
+	result = acpi_get_current_resources(obj, buf);
+	if (result != AE_BUFFER_OVERFLOW)
+		return result;
+	buf->pointer = kmalloc(buf->length, GFP_KERNEL);
+	if (!buf->pointer)
+		return -ENOMEM;
+
+	result = acpi_get_current_resources(obj, buf);
+
+	return result;
+}
+
+acpi_resource *
+acpi_get_crs_next (acpi_buffer *buf, int *offset)
+{
+	acpi_resource *res;
+
+	if (*offset >= buf->length)
+		return NULL;
+
+	res = buf->pointer + *offset;
+	*offset += res->length;
+	return res;
+}
+
+acpi_resource_data *
+acpi_get_crs_type (acpi_buffer *buf, int *offset, int type)
+{
+	for (;;) {
+		acpi_resource *res = acpi_get_crs_next(buf, offset);
+		if (!res)
+			return NULL;
+		if (res->id == type)
+			return &res->data;
+	}
+}
+
+void
+acpi_dispose_crs (acpi_buffer *buf)
+{
+	kfree(buf->pointer);
+}
+
+#endif /* CONFIG_ACPI */
+
 #ifdef CONFIG_ACPI_BOOT
 
 #define ACPI_MAX_PLATFORM_IRQS	256

arch/ia64/kernel/gate.S

@@ -29,7 +29,7 @@
 #	define UNAT_OFF		IA64_SIGCONTEXT_AR_UNAT_OFFSET
 #	define FPSR_OFF		IA64_SIGCONTEXT_AR_FPSR_OFFSET
 #	define PR_OFF		IA64_SIGCONTEXT_PR_OFFSET
-#	define RP_OFF		IA64_SIGCONTEXT_B0_OFFSET
+#	define RP_OFF		IA64_SIGCONTEXT_IP_OFFSET
 #	define SP_OFF		IA64_SIGCONTEXT_R12_OFFSET
 #	define RBS_BASE_OFF	IA64_SIGCONTEXT_RBS_BASE_OFFSET
 #	define LOADRS_OFF	IA64_SIGCONTEXT_LOADRS_OFFSET

arch/ia64/kernel/ia64_ksyms.c

@@ -6,11 +6,7 @@
 #include <linux/module.h>
 
 #include <linux/string.h>
-#undef memset
-extern void *memset (void *, int, size_t);
 EXPORT_SYMBOL_NOVERS(memset);			/* gcc generates direct calls to memset()... */
-EXPORT_SYMBOL_NOVERS(__memset_generic);
-EXPORT_SYMBOL_NOVERS(__bzero);
 EXPORT_SYMBOL(memchr);
 EXPORT_SYMBOL(memcmp);
 EXPORT_SYMBOL_NOVERS(memcpy);

arch/ia64/kernel/iosapic.c

@@ -23,6 +23,7 @@
  *                              iosapic_set_affinity(), initializations for
  *                              /proc/irq/#/smp_affinity
  * 02/04/02	P. Diefenbaugh	Cleaned up ACPI PCI IRQ routing.
+ * 02/04/18	J.I. Lee	bug fix in iosapic_init_pci_irq
  */
 /*
  * Here is what the interrupt logic between a PCI device and the CPU looks like:
@@ -676,6 +677,11 @@ iosapic_init_pci_irq (void)
 		       pci_irq.route[i].bus, pci_irq.route[i].pci_id>>16, pci_irq.route[i].pin,
 		       iosapic_irq[vector].base_irq + iosapic_irq[vector].pin, vector);
 #endif
+
+		/*
+		 * Forget not to program the IOSAPIC RTE per ACPI _PRT
+		 */
+		set_rte(vector, (ia64_get_lid() >> 16) & 0xffff);
 	}
 }
 

arch/ia64/kernel/ivt.S

@@ -330,12 +330,15 @@ ENTRY(alt_dtlb_miss)
 (p8)	br.cond.dptk dtlb_fault
 #endif
 	extr.u r23=r21,IA64_PSR_CPL0_BIT,2	// extract psr.cpl
+	and r22=IA64_ISR_CODE_MASK,r20		// get the isr.code field
 	tbit.nz p6,p7=r20,IA64_ISR_SP_BIT	// is speculation bit on?
-	and r19=r19,r16		// clear ed, reserved bits, and PTE control bits
 	shr.u r18=r16,57			// move address bit 61 to bit 4
+	and r19=r19,r16				// clear ed, reserved bits, and PTE control bits
+	tbit.nz p9,p0=r20,IA64_ISR_NA_BIT	// is non-access bit on?
 	;;
 	andcm r18=0x10,r18	// bit 4=~address-bit(61)
 	cmp.ne p8,p0=r0,r23
+(p9)	cmp.eq.or.andcm p6,p7=IA64_ISR_CODE_LFETCH,r22	// check isr.code field
 (p8)	br.cond.spnt page_fault
 
 	dep r21=-1,r21,IA64_PSR_ED_BIT,1

arch/ia64/kernel/setup.c

@@ -395,7 +395,7 @@ show_cpuinfo (struct seq_file *m, void *v)
 
 	switch (c->family) {
 	      case 0x07:	memcpy(family, "Itanium", 8); break;
-	      case 0x1f:	memcpy(family, "McKinley", 9); break;
+	      case 0x1f:	memcpy(family, "Itanium 2", 9); break;
 	      default:		sprintf(family, "%u", c->family); break;
 	}
 

arch/ia64/kernel/signal.c

@@ -559,7 +559,7 @@ ia64_do_signal (sigset_t *oldset, struct sigscratch *scr, long in_syscall)
 				continue;
 
 			switch (signr) {
-			      case SIGCONT: case SIGCHLD: case SIGWINCH:
+			      case SIGCONT: case SIGCHLD: case SIGWINCH: case SIGURG:
 				continue;
 
 			      case SIGTSTP: case SIGTTIN: case SIGTTOU:

arch/ia64/kernel/traps.c

@@ -186,6 +186,10 @@ ia64_bad_break (unsigned long break_num, struct pt_regs *regs)
 		sig = SIGSEGV; code = __SEGV_PSTKOVF;
 		break;
 
+	      case 0x3f000 ... 0x3ffff:	/* bundle-update in progress */
+		sig = SIGILL; code = __ILL_BNDMOD;
+		break;
+
 	      default:
 		if (break_num < 0x40000 || break_num > 0x100000)
 			die_if_kernel("Bad break", regs, break_num);
@@ -443,30 +447,14 @@ ia64_fault (unsigned long vector, unsigned long isr, unsigned long ifa,
 		"Unknown fault 13", "Unknown fault 14", "Unknown fault 15"
 	};
 
-#if 0
-	/* this is for minimal trust debugging; yeah this kind of stuff is useful at times... */
-
-	if (vector != 25) {
-		static unsigned long last_time;
-		static char count;
-		unsigned long n = vector;
-		char buf[32], *cp;
-
-		if (jiffies - last_time > 5*HZ)
-			count = 0;
-
-		if (count++ < 5) {
-			last_time = jiffies;
-			cp = buf + sizeof(buf);
-			*--cp = '\0';
-			while (n) {
-				*--cp = "0123456789abcdef"[n & 0xf];
-				n >>= 4;
-			}
-			printk("<0x%s>", cp);
-		}
+	if ((isr & IA64_ISR_NA) && ((isr & IA64_ISR_CODE_MASK) == IA64_ISR_CODE_LFETCH)) {
+		/*
+		 * This fault was due to lfetch.fault, set "ed" bit in the psr to cancel
+		 * the lfetch.
+		 */
+		ia64_psr(regs)->ed = 1;
+		return;
 	}
-#endif
 
 	switch (vector) {
 	      case 24: /* General Exception */

arch/ia64/lib/Makefile

@@ -13,7 +13,7 @@ obj-y := __divsi3.o __udivsi3.o __modsi3.o __umodsi3.o					\
 	__divdi3.o __udivdi3.o __moddi3.o __umoddi3.o					\
 	checksum.o clear_page.o csum_partial_copy.o copy_page.o				\
 	copy_user.o clear_user.o strncpy_from_user.o strlen_user.o strnlen_user.o	\
-	flush.o io.o do_csum.o								\
+	flush.o io.o ip_fast_csum.o do_csum.o						\
 	memcpy.o memset.o strlen.o swiotlb.o
 
 obj-$(CONFIG_ITANIUM) += copy_page.o

arch/ia64/lib/checksum.c

@@ -15,7 +15,7 @@
 #include <asm/byteorder.h>
 
 static inline unsigned short
-from64to16(unsigned long x)
+from64to16 (unsigned long x)
 {
 	/* add up 32-bit words for 33 bits */
 	x = (x & 0xffffffff) + (x >> 32);
@@ -32,22 +32,17 @@ from64to16(unsigned long x)
  * computes the checksum of the TCP/UDP pseudo-header
  * returns a 16-bit checksum, already complemented.
  */
-unsigned short int csum_tcpudp_magic(unsigned long saddr,
-				   unsigned long daddr,
-				   unsigned short len,
-				   unsigned short proto,
-				   unsigned int sum)
+unsigned short int
+csum_tcpudp_magic (unsigned long saddr, unsigned long daddr, unsigned short len,
+		   unsigned short proto, unsigned int sum)
 {
-	return ~from64to16(saddr + daddr + sum +
-		((unsigned long) ntohs(len) << 16) +
+	return ~from64to16(saddr + daddr + sum + ((unsigned long) ntohs(len) << 16) +
 			   ((unsigned long) proto << 8));
 }
 
-unsigned int csum_tcpudp_nofold(unsigned long saddr,
-				   unsigned long daddr,
-				   unsigned short len,
-				   unsigned short proto,
-				   unsigned int sum)
+unsigned int
+csum_tcpudp_nofold (unsigned long saddr, unsigned long daddr, unsigned short len,
+		    unsigned short proto, unsigned int sum)
 {
 	unsigned long result;
 
@@ -65,15 +60,6 @@ unsigned int csum_tcpudp_nofold(unsigned long saddr,
 
 extern unsigned long do_csum (const unsigned char *, long);
 
-/*
- *	This is a version of ip_compute_csum() optimized for IP headers,
- *	which always checksum on 4 octet boundaries.
- */
-unsigned short ip_fast_csum(unsigned char * iph, unsigned int ihl)
-{
-	return ~do_csum(iph, ihl*4);
-}
-
 /*
  * computes the checksum of a memory block at buff, length len,
  * and adds in "sum" (32-bit)
@@ -86,7 +72,8 @@ unsigned short ip_fast_csum(unsigned char * iph, unsigned int ihl)
  *
  * it's best to have buff aligned on a 32-bit boundary
  */
-unsigned int csum_partial(const unsigned char * buff, int len, unsigned int sum)
+unsigned int
+csum_partial (const unsigned char * buff, int len, unsigned int sum)
 {
 	unsigned long result = do_csum(buff, len);
 
@@ -102,7 +89,8 @@ unsigned int csum_partial(const unsigned char * buff, int len, unsigned int sum)
  * this routine is used for miscellaneous IP-like checksums, mainly
  * in icmp.c
  */
-unsigned short ip_compute_csum(unsigned char * buff, int len)
+unsigned short
+ip_compute_csum (unsigned char * buff, int len)
 {
 	return ~do_csum(buff,len);
 }

arch/ia64/lib/copy_page.S

@@ -30,6 +30,7 @@
 #define tgt2		r23
 #define srcf		r24
 #define tgtf		r25
+#define tgt_last	r26
 
 #define Nrot		((8*PIPE_DEPTH+7)&~7)
 
@@ -55,18 +56,21 @@ GLOBAL_ENTRY(copy_page)
 
 	mov src1=in1
 	adds src2=8,in1
+	mov tgt_last = PAGE_SIZE
 	;;
 	adds tgt2=8,in0
 	add srcf=512,in1
 	mov ar.lc=lcount
 	mov tgt1=in0
 	add tgtf=512,in0
+	add tgt_last = tgt_last, in0
 	;;
 1:
 (p[0])	ld8 t1[0]=[src1],16
 (EPI)	st8 [tgt1]=t1[PIPE_DEPTH-1],16
 (p[0])	ld8 t2[0]=[src2],16
 (EPI)	st8 [tgt2]=t2[PIPE_DEPTH-1],16
+	cmp.ltu p6,p0 = tgtf, tgt_last
 	;;
 (p[0])	ld8 t3[0]=[src1],16
 (EPI)	st8 [tgt1]=t3[PIPE_DEPTH-1],16
@@ -83,8 +87,8 @@ GLOBAL_ENTRY(copy_page)
 (p[0])	ld8 t8[0]=[src2],16
 (EPI)	st8 [tgt2]=t8[PIPE_DEPTH-1],16
 
-	lfetch [srcf], 64
-	lfetch [tgtf], 64
+(p6)	lfetch [srcf], 64
+(p6)	lfetch [tgtf], 64
 	br.ctop.sptk.few 1b
 	;;
 	mov pr=saved_pr,0xffffffffffff0000	// restore predicates

arch/ia64/lib/do_csum.S

@@ -11,6 +11,9 @@
  * Copyright (C) 1999, 2001-2002 Hewlett-Packard Co
  *	Stephane Eranian <eranian@hpl.hp.com>
  *
+ * 02/04/22	Ken Chen <kenneth.w.chen@intel.com>
+ *		Data locality study on the checksum buffer.
+ *		More optimization cleanup - remove excessive stop bits.
  * 02/04/08	David Mosberger <davidm@hpl.hp.com>
  *		More cleanup and tuning.
  * 01/04/18	Jun Nakajima <jun.nakajima@intel.com>
@@ -80,6 +83,12 @@
 //	  type of packet or alignment we get. Like the ip_fast_csum() routine
 //	  where we know we have at least 20bytes worth of data to checksum.
 //	- Do a better job of handling small packets.
+//	- Note on prefetching: it was found that under various load, i.e. ftp read/write,
+//	  nfs read/write, the L1 cache hit rate is at 60% and L2 cache hit rate is at 99.8%
+//	  on the data that buffer points to (partly because the checksum is often preceded by
+//	  a copy_from_user()).  This finding indiate that lfetch will not be beneficial since
+//	  the data is already in the cache.
+//
 
 #define saved_pfs	r11
 #define hmask		r16
@@ -117,7 +126,7 @@
 GLOBAL_ENTRY(do_csum)
 	.prologue
 	.save ar.pfs, saved_pfs
-	alloc saved_pfs=ar.pfs,2,16,1,16
+	alloc saved_pfs=ar.pfs,2,16,0,16
 	.rotr word1[4], word2[4],result1[LOAD_LATENCY+2],result2[LOAD_LATENCY+2]
 	.rotp p[PIPE_DEPTH], pC1[2], pC2[2]
 	mov ret0=r0		// in case we have zero length
@@ -197,22 +206,21 @@ GLOBAL_ENTRY(do_csum)
 	// Calculate the checksum loading two 8-byte words per loop.
 	//
 .do_csum16:
-	mov saved_lc=ar.lc
 	shr.u count=count,1	// we do 16 bytes per loop
+	brp.loop.imp 1f,2f
 	;;
 	cmp.eq p9,p10=r0,count	// if (count == 0)
 	adds count=-1,count
-	brp.loop.imp 1f,2f
-	;;
 	mov ar.ec=PIPE_DEPTH
-	mov ar.lc=count	// set lc
 
-	// result1[0] must be initialized in advance.
-	mov result2[0]=r0
-	mov pr.rot=1<<16
 	mov carry1=r0
 	mov carry2=r0
 	add first2=8,first1
+	;;
+	mov ar.lc=count	// set lc
+	mov pr.rot=1<<16
+	// result1[0] must be initialized in advance.
+	mov result2[0]=r0
 (p9)	br.cond.sptk .do_csum_exit
 	;;
 	.align 32
@@ -223,7 +231,7 @@ GLOBAL_ENTRY(do_csum)
 (pC2[1])adds carry2=1,carry2
 (ELD)	add result1[LOAD_LATENCY-1]=result1[LOAD_LATENCY],word1[LOAD_LATENCY]
 (ELD)	add result2[LOAD_LATENCY-1]=result2[LOAD_LATENCY],word2[LOAD_LATENCY]
-[2:]
+2:
 (p[0])	ld8 word1[0]=[first1],16
 (p[0])	ld8 word2[0]=[first2],16
 	br.ctop.sptk 1b

arch/ia64/lib/ip_fast_csum.S

+/*
+ * Optmized version of the ip_fast_csum() function
+ * Used for calculating IP header checksum
+ *
+ * Return: 16bit checksum, complemented
+ *
+ * Inputs:
+ *      in0: address of buffer to checksum (char *)
+ *      in1: length of the buffer (int)
+ *
+ * Copyright (C) 2002 Intel Corp.
+ * Copyright (C) 2002 Ken Chen <kenneth.w.chen@intel.com>
+ */
+
+#include <asm/asmmacro.h>
+
+/*
+ * Since we know that most likely this function is called with buf aligned
+ * on 4-byte boundary and 20 bytes in length, we can execution rather quickly
+ * versus calling generic version of do_csum, which has lots of overhead in
+ * handling various alignments and sizes.  However, due to lack of constrains
+ * put on the function input argument, cases with alignment not on 4-byte or
+ * size not equal to 20 bytes will be handled by the generic do_csum function.
+ */
+
+#define in0	r32
+#define in1	r33
+#define ret0	r8
+
+GLOBAL_ENTRY(ip_fast_csum)
+	.prologue
+	.body
+	cmp.ne	p6,p7=5,in1	// size other than 20 byte?
+	and	r14=3,in0	// is it aligned on 4-byte?
+	add	r15=4,in0	// second source pointer
+	;;
+	cmp.ne.or.andcm p6,p7=r14,r0
+	;;
+(p7)	ld4	r20=[in0],8
+(p7)	ld4	r21=[r15],8
+(p6)	br.spnt	.generic
+	;;
+	ld4	r22=[in0],8
+	ld4	r23=[r15],8
+	;;
+	ld4	r24=[in0]
+	add	r20=r20,r21
+	add	r22=r22,r23
+	;;
+	add	r20=r20,r22
+	;;
+	add	r20=r20,r24
+	;;
+	shr.u	ret0=r20,16	// now need to add the carry
+	zxt2	r20=r20
+	;;
+	add	r20=ret0,r20
+	;;
+	shr.u	ret0=r20,16	// add carry again
+	zxt2	r20=r20
+	;;
+	add	r20=ret0,r20
+	;;
+	shr.u	ret0=r20,16
+	zxt2	r20=r20
+	;;
+	add	r20=ret0,r20
+	;;
+	andcm	ret0=-1,r20
+	.restore sp		// reset frame state
+	br.ret.sptk.many b0
+	;;
+
+.generic:
+	.prologue
+	.save ar.pfs, r35
+	alloc	r35=ar.pfs,2,2,2,0
+	.save rp, r34
+	mov	r34=b0
+	.body
+	dep.z	out1=in1,2,30
+	mov	out0=in0
+	;;
+	br.call.sptk.many b0=do_csum
+	;;
+	andcm	ret0=-1,ret0
+	mov	ar.pfs=r35
+	mov	b0=r34
+	br.ret.sptk.many b0
+END(ip_fast_csum)

arch/ia64/mm/fault.c

@@ -137,10 +137,13 @@ ia64_do_page_fault (unsigned long address, unsigned long isr, struct pt_regs *re
 
   bad_area:
 	up_read(&mm->mmap_sem);
-	if (isr & IA64_ISR_SP) {
+	if ((isr & IA64_ISR_SP)
+	    || ((isr & IA64_ISR_NA) && (isr & IA64_ISR_CODE_MASK) == IA64_ISR_CODE_LFETCH))
+	{
 		/*
-		 * This fault was due to a speculative load set the "ed" bit in the psr to
-		 * ensure forward progress (target register will get a NaT).
+		 * This fault was due to a speculative load or lfetch.fault, set the "ed"
+		 * bit in the psr to ensure forward progress.  (Target register will get a
+		 * NaT for ld.s, lfetch will be canceled.)
 		 */
 		ia64_psr(regs)->ed = 1;
 		return;

arch/ia64/tools/print_offsets.c

@@ -143,6 +143,7 @@ tab[] =
     { "IA64_SWITCH_STACK_AR_RNAT_OFFSET",	offsetof (struct switch_stack, ar_rnat) },
     { "IA64_SWITCH_STACK_AR_BSPSTORE_OFFSET",	offsetof (struct switch_stack, ar_bspstore) },
     { "IA64_SWITCH_STACK_PR_OFFSET",	offsetof (struct switch_stack, pr) },
+    { "IA64_SIGCONTEXT_IP_OFFSET",	offsetof (struct sigcontext, sc_ip) },
     { "IA64_SIGCONTEXT_AR_BSP_OFFSET",	offsetof (struct sigcontext, sc_ar_bsp) },
     { "IA64_SIGCONTEXT_AR_FPSR_OFFSET", offsetof (struct sigcontext, sc_ar_fpsr) },
     { "IA64_SIGCONTEXT_AR_RNAT_OFFSET",	offsetof (struct sigcontext, sc_ar_rnat) },

include/asm-ia64/acpi.h

@@ -32,6 +32,7 @@
 
 #define __acpi_map_table(phys_addr, size) __va(phys_addr)
 
+const char *acpi_get_sysname (void);
 int acpi_boot_init (char *cdline);
 int acpi_find_rsdp (unsigned long *phys_addr);
 int acpi_request_vector (u32 int_type);

include/asm-ia64/errno.h

@@ -4,8 +4,8 @@
 /*
  * This is derived from the Linux/x86 version.
  *
- * Copyright (C) 1998, 1999 Hewlett-Packard Co
- * Copyright (C) 1998, 1999 David Mosberger-Tang <davidm@hpl.hp.com>
+ * Copyright (C) 1998, 1999, 2002 Hewlett-Packard Co
+ *	David Mosberger-Tang <davidm@hpl.hp.com>
  */
 
 #define	EPERM		 1	/* Operation not permitted */
@@ -14,7 +14,7 @@
 #define	EINTR		 4	/* Interrupted system call */
 #define	EIO		 5	/* I/O error */
 #define	ENXIO		 6	/* No such device or address */
-#define	E2BIG		 7	/* Arg list too long */
+#define	E2BIG		 7	/* Argument list too long */
 #define	ENOEXEC		 8	/* Exec format error */
 #define	EBADF		 9	/* Bad file number */
 #define	ECHILD		10	/* No child processes */

include/asm-ia64/ioctls.h

@@ -2,8 +2,8 @@
 #define _ASM_IA64_IOCTLS_H
 
 /*
- * Copyright (C) 1998, 1999 Hewlett-Packard Co
- * Copyright (C) 1998, 1999 David Mosberger-Tang <davidm@hpl.hp.com>
+ * Copyright (C) 1998, 1999, 2002 Hewlett-Packard Co
+ *	David Mosberger-Tang <davidm@hpl.hp.com>
  */
 
 #include <asm/ioctl.h>
@@ -11,7 +11,7 @@
 /* 0x54 is just a magic number to make these relatively unique ('T') */
 
 #define TCGETS		0x5401
-#define TCSETS		0x5402
+#define TCSETS		0x5402	/* Clashes with SNDCTL_TMR_START sound ioctl */
 #define TCSETSW		0x5403
 #define TCSETSF		0x5404
 #define TCGETA		0x5405

include/asm-ia64/page.h

@@ -71,7 +71,7 @@ do {						\
 #ifdef CONFIG_IA64_GENERIC
 # include <asm/machvec.h>
 # define virt_to_page(kaddr)	(mem_map + platform_map_nr(kaddr))
-# define page_to_phys(page)	XXX fix me
+# define page_to_phys(page)	((page - mem_map) << PAGE_SHIFT)
 #elif defined (CONFIG_IA64_SGI_SN1)
 # ifndef CONFIG_DISCONTIGMEM
 #  define virt_to_page(kaddr)	(mem_map + MAP_NR_DENSE(kaddr))

include/asm-ia64/processor.h

@@ -154,6 +154,7 @@
 #define IA64_ISR_SP_BIT		36	/* speculative load exception */
 #define IA64_ISR_RS_BIT		37	/* mandatory register-stack exception */
 #define IA64_ISR_IR_BIT		38	/* invalid register frame exception */
+#define IA64_ISR_CODE_MASK	0xf
 
 #define IA64_ISR_X	(__IA64_UL(1) << IA64_ISR_X_BIT)
 #define IA64_ISR_W	(__IA64_UL(1) << IA64_ISR_W_BIT)
@@ -163,6 +164,14 @@
 #define IA64_ISR_RS	(__IA64_UL(1) << IA64_ISR_RS_BIT)
 #define IA64_ISR_IR	(__IA64_UL(1) << IA64_ISR_IR_BIT)
 
+/* ISR code field for non-access instructions */
+#define IA64_ISR_CODE_TPA	0
+#define IA64_ISR_CODE_FC	1
+#define IA64_ISR_CODE_PROBE	2
+#define IA64_ISR_CODE_TAK	3
+#define IA64_ISR_CODE_LFETCH	4
+#define IA64_ISR_CODE_PROBEF	5
+
 #define IA64_THREAD_FPH_VALID	(__IA64_UL(1) << 0)	/* floating-point high state valid? */
 #define IA64_THREAD_DBG_VALID	(__IA64_UL(1) << 1)	/* debug registers valid? */
 #define IA64_THREAD_PM_VALID	(__IA64_UL(1) << 2)	/* performance registers valid? */

include/asm-ia64/siginfo.h

@@ -154,7 +154,8 @@ typedef struct siginfo {
 #define ILL_BADSTK	(__SI_FAULT|8)	/* internal stack error */
 #define ILL_BADIADDR	(__SI_FAULT|9)	/* unimplemented instruction address */
 #define __ILL_BREAK	(__SI_FAULT|10)	/* illegal break */
-#define NSIGILL		10
+#define __ILL_BNDMOD	(__SI_FAULT|11)	/* bundle-update (modification) in progress */
+#define NSIGILL		11
 
 /*
  * SIGFPE si_codes

include/asm-ia64/sn/sn2/shub_md.h

-/**************************************************************************
- *                                                                        *
- *  Copyright (C) 2001 Silicon Graphics, Inc. All rights reserved.        *
- *                                                                        *
- *  These coded instructions, statements, and computer programs  contain  *
- *  unpublished  proprietary  information of Silicon Graphics, Inc., and  *
- *  are protected by Federal copyright law.  They  may  not be disclosed  *
- *  to  third  parties  or copied or duplicated in any form, in whole or  *
- *  in part, without the prior written consent of Silicon Graphics, Inc.  *
- *                                                                        *
- **************************************************************************/
+/*
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (c) 2001, 2002 Silicon Graphics, Inc.  All rights reserved.
+ */
+
 
 #ifndef _SHUB_MD_H
 #define _SHUB_MD_H

include/asm-ia64/string.h

@@ -18,20 +18,6 @@
 
 extern __kernel_size_t strlen (const char *);
 extern void *memcpy (void *, const void *, __kernel_size_t);
-
-extern void *__memset_generic (void *, int, __kernel_size_t);
-extern void __bzero (void *, __kernel_size_t);
-
-#define memset(s, c, count)				\
-({							\
-	void *_s = (s);					\
-	int _c = (c);					\
-	__kernel_size_t _count = (count);		\
-							\
-	if (__builtin_constant_p(_c) && _c == 0)	\
-		__bzero(_s, _count);			\
-	else						\
-		__memset_generic(_s, _c, _count);	\
-})
+extern void *memset (void *, int, __kernel_size_t);
 
 #endif /* _ASM_IA64_STRING_H */

include/asm-ia64/system.h

@@ -353,6 +353,9 @@ extern long __cmpxchg_called_with_bad_pointer(void);
 # define IS_IA32_PROCESS(regs)	(ia64_psr(regs)->is != 0)
 #else
 # define IS_IA32_PROCESS(regs)		0
+struct task_struct;
+static inline void ia32_save_state(struct task_struct *t __attribute__((unused))){}
+static inline void ia32_load_state(struct task_struct *t __attribute__((unused))){}
 #endif
 
 /*