Merge http://mdomsch.bkbits.net/linux-2.5-edd

into home.osdl.org:/home/torvalds/v2.5/linux

Documentation/ia64/fsys.txt

@@ -4,7 +4,7 @@
 		-----------------------------------
 
 		        Started: 13-Jan-2003
-		    Last update: 11-Feb-2003
+		    Last update: 27-Sep-2003
 
 	              David Mosberger-Tang
 		      <davidm@hpl.hp.com>
@@ -146,6 +146,12 @@ speed comes a set of restrictions:
    task pointer is not considered sensitive: it's already exposed
    through ar.k6).
 
+ o Fsyscall-handlers MUST NOT access user-memory without first
+   validating access-permission (this can be done typically via
+   probe.r.fault and/or probe.w.fault) and without guarding against
+   memory access exceptions (this can be done with the EX() macros
+   defined by asmmacro.h).
+
 The above restrictions may seem draconian, but remember that it's
 possible to trade off some of the restrictions by paying a slightly
 higher overhead.  For example, if an fsyscall-handler could benefit
@@ -229,3 +235,52 @@ PSR.ed	Unchanged.  Note: This bit could only have an effect if an fsys-mode
 PSR.bn	Unchanged.  Note: fsys-mode handlers may clear the bit, if needed.
 	Doing so requires clearing PSR.i and PSR.ic as well.
 PSR.ia	Unchanged.  Note: the ia64 linux kernel never sets this bit.
+
+* Using fast system calls
+
+To use fast system calls, userspace applications need simply call
+__kernel_syscall_via_epc().  For example
+
+-- example fgettimeofday() call --
+-- fgettimeofday.S --
+
+#include <asm/asmmacro.h>
+
+GLOBAL_ENTRY(fgettimeofday)
+.prologue
+.save ar.pfs, r11
+mov r11 = ar.pfs
+.body 
+
+mov r2 = 0xa000000000020660;;  // gate address 
+			       // found by inspection of System.map for the 
+			       // __kernel_syscall_via_epc() function.  See
+			       // below for how to do this for real.
+
+mov b7 = r2
+mov r15 = 1087		       // gettimeofday syscall
+;;
+br.call.sptk.many b6 = b7
+;;
+
+.restore sp
+
+mov ar.pfs = r11
+br.ret.sptk.many rp;;	      // return to caller
+END(fgettimeofday)
+
+-- end fgettimeofday.S --
+
+In reality, getting the gate address is accomplished by two extra
+values passed via the ELF auxiliary vector (include/asm-ia64/elf.h)
+
+ o AT_SYSINFO : is the address of __kernel_syscall_via_epc()
+ o AT_SYSINFO_EHDR : is the address of the kernel gate ELF DSO
+
+The ELF DSO is a pre-linked library that is mapped in by the kernel at
+the gate page.  It is a proper ELF shared object so, with a dynamic
+loader that recognises the library, you should be able to make calls to
+the exported functions within it as with any other shared library.
+AT_SYSINFO points into the kernel DSO at the
+__kernel_syscall_via_epc() function for historical reasons (it was
+used before the kernel DSO) and as a convenience.

arch/arm/Makefile

@@ -149,7 +149,7 @@ maketools: include/asm-arm/.arch \
 bzImage: vmlinux
 	$(Q)$(MAKE) $(build)=$(boot) $(boot)/zImage
 
-zImage Image bootpImage: vmlinux
+zImage Image bootpImage uImage: vmlinux
 	$(Q)$(MAKE) $(build)=$(boot) $(boot)/$@
 
 zinstall install: vmlinux

arch/arm/boot/Makefile

@@ -8,6 +8,8 @@
 # Copyright (C) 1995-2002 Russell King
 #
 
+MKIMAGE         := $(srctree)/scripts/mkuboot.sh
+
 # Note: the following conditions must always be true:
 #   ZRELADDR == virt_to_phys(TEXTADDR)
 #   PARAMS_PHYS must be with 4MB of ZRELADDR
@@ -42,12 +44,14 @@ initrd_phys-$(CONFIG_ARCH_CDB89712)	:= 0x00700000
 ifeq ($(CONFIG_ARCH_SA1100),y)
    zreladdr-$(CONFIG_SA1111)		:= 0xc0208000
 endif
+params_phys-$(CONFIG_ARCH_SA1100)	:= 0xc0000100
+initrd_phys-$(CONFIG_ARCH_SA1100)	:= 0xc0800000
    zreladdr-$(CONFIG_ARCH_PXA)		:= 0xa0008000
    zreladdr-$(CONFIG_ARCH_ANAKIN)	:= 0x20008000
    zreladdr-$(CONFIG_ARCH_IOP3XX)	:= 0xa0008000
-params-phys-$(CONFIG_ARCH_IOP3XX)	:= 0xa0000100
+params_phys-$(CONFIG_ARCH_IOP3XX)	:= 0xa0000100
    zreladdr-$(CONFIG_ARCH_ADIFCC)	:= 0xc0008000
-params-phys-$(CONFIG_ARCH_ADIFCC)	:= 0xc0000100
+params_phys-$(CONFIG_ARCH_ADIFCC)	:= 0xc0000100
 
 ZRELADDR    := $(zreladdr-y)
 ZTEXTADDR   := $(ztextaddr-y)
@@ -78,6 +82,16 @@ $(obj)/zImage:	$(obj)/compressed/vmlinux FORCE
 	$(call if_changed,objcopy)
 	@echo '  Kernel: $@ is ready'
 
+quite_cmd_uimage = UIMAGE $@
+      cmd_uimage = $(CONFIG_SHELL) $(MKIMAGE) -A arm -O linux -T kernel \
+		   -C none -a $(ZRELADDR) -e $(ZRELADDR) \
+		   -n 'Linux-$(KERNELRELEASE)' -d $< $@
+
+targets += uImage
+$(obj)/uImage:	$(obj)/zImage
+	$(call if_changed,uimage)
+	@echo '  Image $@ is ready'
+
 $(obj)/bootpImage: $(obj)/bootp/bootp FORCE
 	$(call if_changed,objcopy)
 	@echo '  Kernel: $@ is ready'
@@ -86,7 +100,7 @@ $(obj)/compressed/vmlinux: vmlinux FORCE
 	$(Q)$(MAKE) $(build)=$(obj)/compressed $@
 
 $(obj)/bootp/bootp: $(obj)/zImage initrd FORCE
-	$(Q)$(MAKE) $(build)=$(obj)/compressed $@
+	$(Q)$(MAKE) $(build)=$(obj)/bootp $@
 
 .PHONY: initrd
 initrd:

arch/arm/boot/bootp/bootp.lds

@@ -16,10 +16,10 @@ SECTIONS
   .text : {
    _stext = .;
    *(.start)
-   kernel.o
+   arch/arm/boot/bootp/kernel.o
    . = ALIGN(32);
    initrd_start = .;
-   initrd.o
+   arch/arm/boot/bootp/initrd.o
    initrd_len = . - initrd_start;
    . = ALIGN(32);
    _etext = .;

arch/arm/kernel/armksyms.c

@@ -131,6 +131,7 @@ EXPORT_SYMBOL(fp_init);
 EXPORT_SYMBOL(__machine_arch_type);
 
 	/* networking */
+EXPORT_SYMBOL(csum_partial);
 EXPORT_SYMBOL(csum_partial_copy_nocheck);
 EXPORT_SYMBOL(__csum_ipv6_magic);
 

arch/arm/lib/div64.S

@@ -30,6 +30,8 @@ ENTRY(do_div64)
 	moveq	lr, #1			@ only divide low bits
 	moveq	nh, onl
 
+	tst	dh, #0x80000000
+	bne	2f
 1:	cmp	nh, dh
 	bls	2f
 	add	lr, lr, #1

arch/h8300/Kconfig

@@ -341,8 +341,8 @@ config UNIX98_PTY_COUNT
 
 endmenu
 
-#source drivers/misc/Config.in
 source "drivers/media/Kconfig"
+source "sound/Kconfig"
 
 source "fs/Kconfig"
 

arch/h8300/README

-uClinux-2.4 for H8/300 README
+linux-2.6 for H8/300 README
 Yoshinori Sato <ysato@users.sourceforge.jp>
 
 * Supported CPU
-H8/300H
-H8S is planning.
+H8/300H and H8S
 
 * Supported Target
 1.simulator of GDB
@@ -15,8 +14,11 @@ H8S is planning.
   Akizuki Denshi Tsusho Ltd. <http://www.akizuki.ne.jp> (Japanese Only)
 
 3.H8MAX 
-  Under development
-  see http://www.strawberry-linux.com (Japanese Only)
+  see http://ip-sol.jp/h8max/ (Japanese Only)
+
+4.EDOSK2674
+  see http://www.eu.renesas.com/products/mpumcu/tool/edk/support/edosk2674.html
+      http://www.azpower.com/H8-uClinux/
 
 * Toolchain Version
 gcc-3.1 or higher and patch
@@ -26,10 +28,10 @@ gdb-5.2 or higher
 The environment that can compile a h8300-elf binary is necessary.
 
 * Userland Develop environment
-Tempolary used h8300-hms(h8300-coff) Toolchain.
-I prepare toolchain corresponding to h8300-elf.
+used h8300-elf toolchains.
+see http://www.uclinux.org/pub/uClinux/ports/h8/
 
 * A few words of thanks
-Porting to H8/300H is support of Information-technology Promotion Agency, Japan.
+Porting to H8/300 serieses is support of Information-technology Promotion Agency, Japan.
 I thank support.
 and All developer/user.

arch/h8300/lib/romfs.S

 /* romfs move to __ebss */
 
 #include <asm/linkage.h>
+#include <linux/config.h>
 
 #if defined(__H8300H__) 
 	.h8300h
@@ -9,6 +10,8 @@
 	.h8300s
 #endif
 
+#define BLKOFFSET 512
+
 	.text
 .globl	__move_romfs
 _romfs_sig_len = 8
@@ -31,6 +34,9 @@ __move_romfs:
 	add.l	er0,er1			/* romfs image end */
 	mov.l	#__ebss,er2
 	add.l	er0,er2			/* distination address */
+#if defined(CONFIG_INTELFLASH)
+	add.l	#BLKOFFSET,er2
+#endif
 	adds	#2,er0
 	adds	#1,er0
 	shlr	er0

arch/h8300/platform/h8s/ints.c

 /*
- * linux/arch/h8300/platform/h8sh/ints.c
+ * linux/arch/h8300/platform/h8s/ints.c
  *
  * Yoshinori Sato <ysato@users.sourceforge.jp>
  *
@@ -20,7 +20,6 @@
 #include <linux/kernel_stat.h>
 #include <linux/seq_file.h>
 #include <linux/init.h>
-#include <linux/slab.h>
 #include <linux/bootmem.h>
 #include <linux/random.h>
 
@@ -75,7 +74,7 @@ const static struct irq_pins irq_assign_table1[16]={
 	{H8300_GPIO_P2,H8300_GPIO_B6},{H8300_GPIO_P2,H8300_GPIO_B7},
 };
 
-static int use_kmalloc;
+static short use_kmalloc = 0;
 
 extern unsigned long *interrupt_redirect_table;
 

arch/i386/kernel/apic.c

@@ -533,14 +533,19 @@ static int lapic_resume(struct sys_device *dev)
 	if (!apic_pm_state.active)
 		return 0;
 
-	/* XXX: Pavel needs this for S3 resume, but can't explain why */
-	set_fixmap_nocache(FIX_APIC_BASE, APIC_DEFAULT_PHYS_BASE);
-
 	local_irq_save(flags);
+
+	/*
+	 * Make sure the APICBASE points to the right address
+	 *
+	 * FIXME! This will be wrong if we ever support suspend on
+	 * SMP! We'll need to do this as part of the CPU restore!
+	 */
 	rdmsr(MSR_IA32_APICBASE, l, h);
 	l &= ~MSR_IA32_APICBASE_BASE;
-	l |= MSR_IA32_APICBASE_ENABLE | APIC_DEFAULT_PHYS_BASE;
+	l |= MSR_IA32_APICBASE_ENABLE | mp_lapic_addr;
 	wrmsr(MSR_IA32_APICBASE, l, h);
+
 	apic_write(APIC_LVTERR, ERROR_APIC_VECTOR | APIC_LVT_MASKED);
 	apic_write(APIC_ID, apic_pm_state.apic_id);
 	apic_write(APIC_DFR, apic_pm_state.apic_dfr);
@@ -680,6 +685,12 @@ static int __init detect_init_APIC (void)
 	}
 	set_bit(X86_FEATURE_APIC, boot_cpu_data.x86_capability);
 	mp_lapic_addr = APIC_DEFAULT_PHYS_BASE;
+
+	/* The BIOS may have set up the APIC at some other address */
+	rdmsr(MSR_IA32_APICBASE, l, h);
+	if (l & MSR_IA32_APICBASE_ENABLE)
+		mp_lapic_addr = l & MSR_IA32_APICBASE_BASE;
+
 	if (nmi_watchdog != NMI_NONE)
 		nmi_watchdog = NMI_LOCAL_APIC;
 

arch/i386/kernel/irq.c

@@ -380,7 +380,7 @@ void enable_irq(unsigned int irq)
 	spin_lock_irqsave(&desc->lock, flags);
 	switch (desc->depth) {
 	case 1: {
-		unsigned int status = desc->status & ~IRQ_DISABLED;
+		unsigned int status = desc->status & ~(IRQ_DISABLED | IRQ_INPROGRESS);
 		desc->status = status;
 		if ((status & (IRQ_PENDING | IRQ_REPLAY)) == IRQ_PENDING) {
 			desc->status = status | IRQ_REPLAY;

arch/i386/kernel/mpparse.c

@@ -337,6 +337,16 @@ static void __init smp_read_mpc_oem(struct mp_config_oemtable *oemtable, \
 		}
        }
 }
+
+static inline void mps_oem_check(struct mp_config_table *mpc, char *oem,
+		char *productid)
+{
+	if (strncmp(oem, "IBM NUMA", 8))
+		printk("Warning!  May not be a NUMA-Q system!\n");
+	if (mpc->mpc_oemptr)
+		smp_read_mpc_oem((struct mp_config_oemtable *) mpc->mpc_oemptr,
+				mpc->mpc_oemsize);
+}
 #endif	/* CONFIG_X86_NUMAQ */
 
 /*

arch/i386/kernel/traps.c

@@ -104,7 +104,7 @@ void show_trace(struct task_struct *task, unsigned long * stack)
 #ifdef CONFIG_KALLSYMS
 	printk("\n");
 #endif
-	while (((long) stack & (THREAD_SIZE-1)) != 0) {
+	while (!kstack_end(stack)) {
 		addr = *stack++;
 		if (kernel_text_address(addr)) {
 			printk(" [<%08lx>] ", addr);
@@ -138,7 +138,7 @@ void show_stack(struct task_struct *task, unsigned long *esp)
 
 	stack = esp;
 	for(i = 0; i < kstack_depth_to_print; i++) {
-		if (((long) stack & (THREAD_SIZE-1)) == 0)
+		if (kstack_end(stack))
 			break;
 		if (i && ((i % 8) == 0))
 			printk("\n       ");
@@ -374,6 +374,9 @@ DO_ERROR_INFO(17, SIGBUS, "alignment check", alignment_check, BUS_ADRALN, get_cr
 
 asmlinkage void do_general_protection(struct pt_regs * regs, long error_code)
 {
+	if (regs->eflags & X86_EFLAGS_IF)
+		local_irq_enable();
+ 
 	if (regs->eflags & VM_MASK)
 		goto gp_in_vm86;
 
@@ -386,6 +389,7 @@ asmlinkage void do_general_protection(struct pt_regs * regs, long error_code)
 	return;
 
 gp_in_vm86:
+	local_irq_enable();
 	handle_vm86_fault((struct kernel_vm86_regs *) regs, error_code);
 	return;
 

arch/i386/mm/fault.c

@@ -223,7 +223,7 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long error_code)
 	__asm__("movl %%cr2,%0":"=r" (address));
 
 	/* It's safe to allow irq's after cr2 has been saved */
-	if (regs->eflags & X86_EFLAGS_IF)
+	if (regs->eflags & (X86_EFLAGS_IF|VM_MASK))
 		local_irq_enable();
 
 	tsk = current;

arch/ia64/Kconfig

@@ -57,6 +57,10 @@ choice
 
 config IA64_GENERIC
 	bool "generic"
+	select NUMA
+	select ACPI_NUMA
+	select VIRTUAL_MEM_MAP
+	select DISCONTIGMEM
 	---help---
 	  This selects the system type of your hardware.  A "generic" kernel
 	  will run on any supported IA-64 system.  However, if you configure
@@ -220,24 +224,8 @@ config NUMA
 	  Access).  This option is for configuring high-end multiprocessor
 	  server systems.  If in doubt, say N.
 
-choice
-	prompt "Maximum Memory per NUMA Node" if NUMA && IA64_DIG
-	depends on NUMA && IA64_DIG
-	default IA64_NODESIZE_16GB
-
-config IA64_NODESIZE_16GB
-	bool "16GB"
-
-config IA64_NODESIZE_64GB
-	bool "64GB"
-
-config IA64_NODESIZE_256GB
-	bool "256GB"
-
-endchoice
-
 config DISCONTIGMEM
-	bool "Discontiguous memory support" if (IA64_DIG || IA64_SGI_SN2 || IA64_GENERIC) && NUMA
+	bool "Discontiguous memory support" if (IA64_DIG || IA64_SGI_SN2 || IA64_GENERIC) && NUMA && VIRTUAL_MEM_MAP
 	default y if (IA64_SGI_SN2 || IA64_GENERIC) && NUMA
 	help
 	  Say Y to support efficient handling of discontiguous physical memory,
@@ -250,14 +238,10 @@ config VIRTUAL_MEM_MAP
 	default y if !IA64_HP_SIM
 	help
 	  Say Y to compile the kernel with support for a virtual mem map.
-	  This is an alternate method of supporting large holes in the
-	  physical address space on non NUMA machines. Since the DISCONTIGMEM
-	  option is not supported on machines with the ZX1 chipset, this is
-	  the only way of supporting more than 1 Gb of memory on those
-	  machines. This code also only takes effect if a memory hole of
-	  greater than 1 Gb is found during boot, so it is safe to enable
-	  unless you require the DISCONTIGMEM option for your machine. If you
-	  are unsure, say Y.
+	  This code also only takes effect if a memory hole of greater than
+	  1 Gb is found during boot.  You must turn this option on if you
+	  require the DISCONTIGMEM option for your machine. If you are
+	  unsure, say Y.
 
 config IA64_MCA
 	bool "Enable IA-64 Machine Check Abort"

arch/ia64/Makefile

@@ -64,7 +64,7 @@ core-$(CONFIG_IA64_SGI_SN2)	+= arch/ia64/sn/
 drivers-$(CONFIG_PCI)		+= arch/ia64/pci/
 drivers-$(CONFIG_IA64_HP_SIM)	+= arch/ia64/hp/sim/
 drivers-$(CONFIG_IA64_HP_ZX1)	+= arch/ia64/hp/common/ arch/ia64/hp/zx1/
-drivers-$(CONFIG_IA64_GENERIC)	+= arch/ia64/hp/common/ arch/ia64/hp/zx1/ arch/ia64/hp/sim/
+drivers-$(CONFIG_IA64_GENERIC)	+= arch/ia64/hp/common/ arch/ia64/hp/zx1/ arch/ia64/hp/sim/ arch/ia64/sn/
 drivers-$(CONFIG_OPROFILE)	+= arch/ia64/oprofile/
 
 boot := arch/ia64/hp/sim/boot

arch/ia64/ia32/sys_ia32.c

@@ -2486,11 +2486,14 @@ static int
 putstat64 (struct stat64 *ubuf, struct kstat *kbuf)
 {
 	int err;
+	u64 hdev;
 
 	if (clear_user(ubuf, sizeof(*ubuf)))
 		return -EFAULT;
 
-	err  = __put_user(huge_encode_dev(kbuf->dev), &ubuf->st_dev);
+	hdev = huge_encode_dev(kbuf->dev);
+	err  = __put_user(hdev, (u32*)&ubuf->st_dev);
+	err |= __put_user(hdev >> 32, ((u32*)&ubuf->st_dev) + 1);
 	err |= __put_user(kbuf->ino, &ubuf->__st_ino);
 	err |= __put_user(kbuf->ino, &ubuf->st_ino_lo);
 	err |= __put_user(kbuf->ino >> 32, &ubuf->st_ino_hi);
@@ -2498,7 +2501,9 @@ putstat64 (struct stat64 *ubuf, struct kstat *kbuf)
 	err |= __put_user(kbuf->nlink, &ubuf->st_nlink);
 	err |= __put_user(kbuf->uid, &ubuf->st_uid);
 	err |= __put_user(kbuf->gid, &ubuf->st_gid);
-	err |= __put_user(huge_encode_dev(kbuf->rdev), &ubuf->st_rdev);
+	hdev = huge_encode_dev(kbuf->rdev);
+	err  = __put_user(hdev, (u32*)&ubuf->st_rdev);
+	err |= __put_user(hdev >> 32, ((u32*)&ubuf->st_rdev) + 1);
 	err |= __put_user(kbuf->size, &ubuf->st_size_lo);
 	err |= __put_user((kbuf->size >> 32), &ubuf->st_size_hi);
 	err |= __put_user(kbuf->atime.tv_sec, &ubuf->st_atime);
@@ -2724,8 +2729,8 @@ sys32_open (const char * filename, int flags, int mode)
 struct epoll_event32
 {
 	u32 events;
-	u64 data;
-} __attribute__((packed));
+	u32 data[2];
+}; 
 
 asmlinkage long
 sys32_epoll_ctl(int epfd, int op, int fd, struct epoll_event32 *event)
@@ -2740,10 +2745,10 @@ sys32_epoll_ctl(int epfd, int op, int fd, struct epoll_event32 *event)
 		return error;
 
 	__get_user(event64.events, &event->events);
-	__get_user(data_halfword, (u32*)(&event->data));
+	__get_user(data_halfword, &event->data[0]);
 	event64.data = data_halfword;
-	__get_user(data_halfword, ((u32*)(&event->data) + 1));
- 	event64.data |= ((u64)data_halfword) << 32;
+	__get_user(data_halfword, &event->data[1]);
+ 	event64.data |= (u64)data_halfword << 32;
 
 	set_fs(KERNEL_DS);
 	error = sys_epoll_ctl(epfd, op, fd, &event64);
@@ -2758,8 +2763,9 @@ sys32_epoll_wait(int epfd, struct epoll_event32 *events, int maxevents,
 {
 	struct epoll_event *events64 = NULL;
 	mm_segment_t old_fs = get_fs();
-	int error;
+	int error, numevents, size;
 	int evt_idx;
+	int do_free_pages = 0;
 
 	if (maxevents <= 0) {
 		return -EINVAL;
@@ -2770,43 +2776,45 @@ sys32_epoll_wait(int epfd, struct epoll_event32 *events, int maxevents,
 				 maxevents * sizeof(struct epoll_event32))))
 		return error;
 
-	/* Allocate the space needed for the intermediate copy */
-	events64 = kmalloc(maxevents * sizeof(struct epoll_event), GFP_KERNEL);
+	/* 
+ 	 * Allocate space for the intermediate copy.  If the space needed 
+	 * is large enough to cause kmalloc to fail, then try again with
+	 * __get_free_pages.
+	 */
+	size = maxevents * sizeof(struct epoll_event);
+	events64 = kmalloc(size, GFP_KERNEL);
 	if (events64 == NULL) {
-		return -ENOMEM;
-	}
-
-	/* Expand the 32-bit structures into the 64-bit structures */
-	for (evt_idx = 0; evt_idx < maxevents; evt_idx++) {
-		u32 data_halfword;
-		__get_user(events64[evt_idx].events, &events[evt_idx].events);
-		__get_user(data_halfword, (u32*)(&events[evt_idx].data));
-		events64[evt_idx].data = data_halfword;
-		__get_user(data_halfword, ((u32*)(&events[evt_idx].data) + 1));
-		events64[evt_idx].data |= ((u64)data_halfword) << 32;
+		events64 = (struct epoll_event *)
+				__get_free_pages(GFP_KERNEL, get_order(size));
+		if (events64 == NULL) 
+			return -ENOMEM;
+		do_free_pages = 1;
 	}
 
 	/* Do the system call */
 	set_fs(KERNEL_DS); /* copy_to/from_user should work on kernel mem*/
-	error = sys_epoll_wait(epfd, events64, maxevents, timeout);
+	numevents = sys_epoll_wait(epfd, events64, maxevents, timeout);
 	set_fs(old_fs);
 
 	/* Don't modify userspace memory if we're returning an error */
-	if (!error) {
+	if (numevents > 0) {
 		/* Translate the 64-bit structures back into the 32-bit
 		   structures */
-		for (evt_idx = 0; evt_idx < maxevents; evt_idx++) {
+		for (evt_idx = 0; evt_idx < numevents; evt_idx++) {
 			__put_user(events64[evt_idx].events,
 				   &events[evt_idx].events);
-			__put_user((u32)(events64[evt_idx].data),
-				   (u32*)(&events[evt_idx].data));
+			__put_user((u32)events64[evt_idx].data,
+				   &events[evt_idx].data[0]);
 			__put_user((u32)(events64[evt_idx].data >> 32),
-				   ((u32*)(&events[evt_idx].data) + 1));
+				   &events[evt_idx].data[1]);
 		}
 	}
 
-	kfree(events64);
-	return error;
+	if (do_free_pages)
+		free_pages((unsigned long) events64, get_order(size));
+	else
+		kfree(events64);
+	return numevents;
 }
 
 #ifdef	NOTYET  /* UNTESTED FOR IA64 FROM HERE DOWN */

arch/ia64/kernel/acpi.c

@@ -56,6 +56,7 @@ void (*pm_idle) (void);
 void (*pm_power_off) (void);
 
 unsigned char acpi_kbd_controller_present = 1;
+unsigned char acpi_legacy_devices;
 
 int acpi_disabled;	/* XXX this shouldn't be needed---we can't boot without ACPI! */
 
@@ -380,7 +381,7 @@ acpi_numa_processor_affinity_init (struct acpi_table_processor_affinity *pa)
 void __init
 acpi_numa_memory_affinity_init (struct acpi_table_memory_affinity *ma)
 {
-	unsigned long paddr, size, hole_size, min_hole_size;
+	unsigned long paddr, size;
 	u8 pxm;
 	struct node_memblk_s *p, *q, *pend;
 
@@ -402,34 +403,6 @@ acpi_numa_memory_affinity_init (struct acpi_table_memory_affinity *ma)
 	if (!ma->flags.enabled)
 		return;
 
-	/*
-	 * When the chunk is not the first one in the node, check distance
-	 * from the other chunks. When the hole is too huge ignore the chunk.
-	 * This restriction should be removed when multiple chunks per node
-	 * is supported.
-	 */
-	pend = &node_memblk[num_memblks];
-	min_hole_size = 0;
-	for (p = &node_memblk[0]; p < pend; p++) {
-		if (p->nid != pxm)
-			continue;
-		if (p->start_paddr < paddr)
-			hole_size = paddr - (p->start_paddr + p->size);
-		else
-			hole_size = p->start_paddr - (paddr + size);
-
-		if (!min_hole_size || hole_size < min_hole_size)
-			min_hole_size = hole_size;
-	}
-
-	if (min_hole_size) {
-		if (min_hole_size > size) {
-			printk(KERN_ERR "Too huge memory hole. Ignoring %ld MBytes at %lx\n",
-			       size/(1024*1024), paddr);
-			return;
-		}
-	}
-
 	/* record this node in proximity bitmap */
 	pxm_bit_set(pxm);
 
@@ -454,6 +427,12 @@ acpi_numa_arch_fixup (void)
 {
 	int i, j, node_from, node_to;
 
+	/* If there's no SRAT, fix the phys_id */
+	if (srat_num_cpus == 0) {
+		node_cpuid[0].phys_id = hard_smp_processor_id();
+		return;
+	}
+
 	/* calculate total number of nodes in system from PXM bitmap */
 	numnodes = 0;		/* init total nodes in system */
 
@@ -531,6 +510,9 @@ acpi_parse_fadt (unsigned long phys_addr, unsigned long size)
 	if (!(fadt->iapc_boot_arch & BAF_8042_KEYBOARD_CONTROLLER))
 		acpi_kbd_controller_present = 0;
 
+	if (fadt->iapc_boot_arch & BAF_LEGACY_DEVICES)
+		acpi_legacy_devices = 1;
+
 	acpi_register_irq(fadt->sci_int, ACPI_ACTIVE_LOW, ACPI_LEVEL_SENSITIVE);
 	return 0;
 }
@@ -614,6 +596,12 @@ acpi_boot_init (void)
 
 	smp_build_cpu_map();
 # ifdef CONFIG_NUMA
+	if (srat_num_cpus == 0) {
+		int cpu, i = 1;
+		for (cpu = 0; cpu < smp_boot_data.cpu_count; cpu++)
+			if (smp_boot_data.cpu_phys_id[cpu] != hard_smp_processor_id())
+				node_cpuid[i++].phys_id = smp_boot_data.cpu_phys_id[cpu];
+	}
 	build_cpu_to_node_map();
 # endif
 #endif

arch/ia64/kernel/asm-offsets.c

@@ -33,16 +33,30 @@ void foo(void)
 
 	BLANK();
 
+	DEFINE(IA64_TASK_BLOCKED_OFFSET,offsetof (struct task_struct, blocked));
 	DEFINE(IA64_TASK_CLEAR_CHILD_TID_OFFSET,offsetof (struct task_struct, clear_child_tid));
 	DEFINE(IA64_TASK_GROUP_LEADER_OFFSET, offsetof (struct task_struct, group_leader));
+	DEFINE(IA64_TASK_PENDING_OFFSET,offsetof (struct task_struct, pending));
 	DEFINE(IA64_TASK_PID_OFFSET, offsetof (struct task_struct, pid));
 	DEFINE(IA64_TASK_REAL_PARENT_OFFSET, offsetof (struct task_struct, real_parent));
+	DEFINE(IA64_TASK_SIGHAND_OFFSET,offsetof (struct task_struct, sighand));
+	DEFINE(IA64_TASK_SIGNAL_OFFSET,offsetof (struct task_struct, signal));
 	DEFINE(IA64_TASK_TGID_OFFSET, offsetof (struct task_struct, tgid));
 	DEFINE(IA64_TASK_THREAD_KSP_OFFSET, offsetof (struct task_struct, thread.ksp));
 	DEFINE(IA64_TASK_THREAD_ON_USTACK_OFFSET, offsetof (struct task_struct, thread.on_ustack));
 
 	BLANK();
 
+	DEFINE(IA64_SIGHAND_SIGLOCK_OFFSET,offsetof (struct sighand_struct, siglock));
+
+	BLANK();
+
+	DEFINE(IA64_SIGNAL_GROUP_STOP_COUNT_OFFSET,offsetof (struct signal_struct,
+							     group_stop_count));
+	DEFINE(IA64_SIGNAL_SHARED_PENDING_OFFSET,offsetof (struct signal_struct, shared_pending));
+
+	BLANK();
+
 	DEFINE(IA64_PT_REGS_B6_OFFSET, offsetof (struct pt_regs, b6));
 	DEFINE(IA64_PT_REGS_B7_OFFSET, offsetof (struct pt_regs, b7));
 	DEFINE(IA64_PT_REGS_AR_CSD_OFFSET, offsetof (struct pt_regs, ar_csd));
@@ -158,6 +172,10 @@ void foo(void)
 
 	BLANK();
 
+	DEFINE(IA64_SIGPENDING_SIGNAL_OFFSET, offsetof (struct sigpending, signal));
+
+	BLANK();
+
 	DEFINE(IA64_SIGFRAME_ARG0_OFFSET, offsetof (struct sigframe, arg0));
 	DEFINE(IA64_SIGFRAME_ARG1_OFFSET, offsetof (struct sigframe, arg1));
 	DEFINE(IA64_SIGFRAME_ARG2_OFFSET, offsetof (struct sigframe, arg2));

arch/ia64/kernel/fsys.S

@@ -4,6 +4,7 @@
  * Copyright (C) 2003 Hewlett-Packard Co
  * 	David Mosberger-Tang <davidm@hpl.hp.com>
  *
+ * 25-Sep-03 davidm	Implement fsys_rt_sigprocmask().
  * 18-Feb-03 louisk	Implement fsys_gettimeofday().
  * 28-Feb-03 davidm	Fixed several bugs in fsys_gettimeofday().  Tuned it some more,
  *			probably broke it along the way... ;-)
@@ -15,6 +16,7 @@
 #include <asm/percpu.h>
 #include <asm/thread_info.h>
 #include <asm/sal.h>
+#include <asm/signal.h>
 #include <asm/system.h>
 #include <asm/unistd.h>
 
@@ -48,8 +50,7 @@ ENTRY(fsys_ni_syscall)
 	.body
 	mov r8=ENOSYS
 	mov r10=-1
-	MCKINLEY_E9_WORKAROUND
-	br.ret.sptk.many b6
+	FSYS_RETURN
 END(fsys_ni_syscall)
 
 ENTRY(fsys_getpid)
@@ -66,8 +67,7 @@ ENTRY(fsys_getpid)
 	;;
 	cmp.ne p8,p0=0,r9
 (p8)	br.spnt.many fsys_fallback_syscall
-	MCKINLEY_E9_WORKAROUND
-	br.ret.sptk.many b6
+	FSYS_RETURN
 END(fsys_getpid)
 
 ENTRY(fsys_getppid)
@@ -114,8 +114,7 @@ ENTRY(fsys_getppid)
 	mov r18=0			// i must not leak kernel bits...
 	mov r19=0			// i must not leak kernel bits...
 #endif
-	MCKINLEY_E9_WORKAROUND
-	br.ret.sptk.many b6
+	FSYS_RETURN
 END(fsys_getppid)
 
 ENTRY(fsys_set_tid_address)
@@ -141,8 +140,7 @@ ENTRY(fsys_set_tid_address)
 	;;
 	mov r17=0			// i must not leak kernel bits...
 	mov r18=0			// i must not leak kernel bits...
-	MCKINLEY_E9_WORKAROUND
-	br.ret.sptk.many b6
+	FSYS_RETURN
 END(fsys_set_tid_address)
 
 /*
@@ -199,7 +197,7 @@ ENTRY(fsys_gettimeofday)
 
 	adds r10=IA64_CPUINFO_ITM_DELTA_OFFSET, r10
 (p7)	tnat.nz p6,p0=r33
-(p6)	br.cond.spnt.few .fail
+(p6)	br.cond.spnt.few .fail_einval
 
 	adds r8=IA64_CPUINFO_NSEC_PER_CYC_OFFSET, r3
 	movl r24=2361183241434822607	// for division hack (only for / 1000)
@@ -225,8 +223,8 @@ ENTRY(fsys_gettimeofday)
 	 * to store the result.  That's OK as long as the stores are also
 	 * protect by EX().
 	 */
-EX(.fail, probe.w.fault r32, 3)		// this must come _after_ NaT-check
-EX(.fail, probe.w.fault r10, 3)		// this must come _after_ NaT-check
+EX(.fail_efault, probe.w.fault r32, 3)		// this must come _after_ NaT-check
+EX(.fail_efault, probe.w.fault r10, 3)		// this must come _after_ NaT-check
 	nop 0
 
 	ldf8 f10=[r8]			// f10 <- local_cpu_data->nsec_per_cyc value
@@ -311,14 +309,13 @@ EX(.fail, probe.w.fault r10, 3)		// this must come _after_ NaT-check
 (p7)	br.spnt.many 1b
 
 	// finally: r2 = sec, r3 = usec
-EX(.fail, st8 [r32]=r2)
+EX(.fail_efault, st8 [r32]=r2)
 	adds r9=8, r32
 	mov r8=r0			// success
 	;;
-EX(.fail, st8 [r9]=r3)			// store them in the timeval struct
+EX(.fail_efault, st8 [r9]=r3)		// store them in the timeval struct
 	mov r10=0
-	MCKINLEY_E9_WORKAROUND
-	br.ret.sptk.many b6		// return to caller
+	FSYS_RETURN
 	/*
 	 * Note: We are NOT clearing the scratch registers here.  Since the only things
 	 *	 in those registers are time-related variables and some addresses (which
@@ -326,12 +323,183 @@ EX(.fail, st8 [r9]=r3)			// store them in the timeval struct
 	 *	 and we should be fine.
 	 */
 
-.fail:	adds r8=EINVAL, r0		// r8 = EINVAL
-	adds r10=-1, r0			// r10 = -1
-	MCKINLEY_E9_WORKAROUND
-	br.ret.spnt.many b6		// return with r8 set to EINVAL
+.fail_einval:
+	mov r8=EINVAL			// r8 = EINVAL
+	mov r10=-1			// r10 = -1
+	FSYS_RETURN
+
+.fail_efault:
+	mov r8=EFAULT			// r8 = EFAULT
+	mov r10=-1			// r10 = -1
+	FSYS_RETURN
 END(fsys_gettimeofday)
 
+/*
+ * long fsys_rt_sigprocmask (int how, sigset_t *set, sigset_t *oset, size_t sigsetsize).
+ */
+#if _NSIG_WORDS != 1
+# error Sorry, fsys_rt_sigprocmask() needs to be updated for _NSIG_WORDS != 1.
+#endif
+ENTRY(fsys_rt_sigprocmask)
+	.prologue
+	.altrp b6
+	.body
+
+	mf					// ensure reading of current->blocked is ordered
+	add r2=IA64_TASK_BLOCKED_OFFSET,r16
+	add r9=TI_FLAGS+IA64_TASK_SIZE,r16
+	;;
+	/*
+	 * Since we're only reading a single word, we can do it
+	 * atomically without acquiring current->sighand->siglock.  To
+	 * be on the safe side, we need a fully-ordered load, though:
+	 */
+	ld8.acq r3=[r2]				// read/prefetch current->blocked
+	ld4 r9=[r9]
+	add r31=IA64_TASK_SIGHAND_OFFSET,r16
+	;;
+#ifdef CONFIG_SMP
+	ld8 r31=[r31]				// r31 <- current->sighand
+#endif
+	and r9=TIF_ALLWORK_MASK,r9
+	tnat.nz p6,p0=r32
+	;;
+	cmp.ne p7,p0=0,r9
+	tnat.nz.or p6,p0=r35
+	tnat.nz p8,p0=r34
+	;;
+	cmp.ne p15,p0=r0,r34			// oset != NULL?
+	cmp.ne.or p6,p0=_NSIG_WORDS*8,r35
+	tnat.nz.or p8,p0=r33
+
+(p6)	br.spnt.few .fail_einval		// fail with EINVAL
+(p7)	br.spnt.many fsys_fallback_syscall	// got pending kernel work...
+(p8)	br.spnt.few .fail_efault		// fail with EFAULT
+	;;
+
+	cmp.eq p6,p7=r0,r33			// set == NULL?
+	add r31=IA64_SIGHAND_SIGLOCK_OFFSET,r31	// r31 <- current->sighand->siglock
+(p6)	br.dpnt.many .store_mask		// -> short-circuit to just reading the signal mask
+
+	/* Argh, we actually have to do some work and _update_ the signal mask: */
+
+EX(.fail_efault, probe.r.fault r33, 3)		// verify user has read-access to *set
+EX(.fail_efault, ld8 r14=[r33])			// r14 <- *set
+	mov r17=(1 << (SIGKILL - 1)) | (1 << (SIGSTOP - 1))
+	;;
+
+	rsm psr.i				// mask interrupt delivery
+	mov ar.ccv=0
+	andcm r14=r14,r17			// filter out SIGKILL & SIGSTOP
+
+#ifdef CONFIG_SMP
+	mov r17=1
+	;;
+	cmpxchg4.acq r18=[r31],r17,ar.ccv	// try to acquire the lock
+	mov r8=EINVAL			// default to EINVAL
+	;;
+	ld8 r3=[r2]			// re-read current->blocked now that we hold the lock
+	cmp4.ne p6,p0=r18,r0
+(p6)	br.cond.spnt.many .lock_contention
+	;;
+#else
+	ld8 r3=[r2]			// re-read current->blocked now that we hold the lock
+	mov r8=EINVAL			// default to EINVAL
+#endif
+	add r18=IA64_TASK_PENDING_OFFSET+IA64_SIGPENDING_SIGNAL_OFFSET,r16
+	add r19=IA64_TASK_SIGNAL_OFFSET,r16
+	cmp4.eq p6,p0=SIG_BLOCK,r32
+	;;
+	ld8 r19=[r19]			// r19 <- current->signal
+	cmp4.eq p7,p0=SIG_UNBLOCK,r32
+	cmp4.eq p8,p0=SIG_SETMASK,r32
+	;;
+	ld8 r18=[r18]			// r18 <- current->pending.signal
+	.pred.rel.mutex p6,p7,p8
+(p6)	or r3=r3,r14			// SIG_BLOCK
+(p7)	andcm r3=r3,r14			// SIG_UNBLOCK
+
+(p8)	mov r3=r14			// SIG_SETMASK
+(p6)	mov r8=0			// clear error code
+	// recalc_sigpending()
+	add r17=IA64_SIGNAL_GROUP_STOP_COUNT_OFFSET,r19
+
+	add r19=IA64_SIGNAL_SHARED_PENDING_OFFSET+IA64_SIGPENDING_SIGNAL_OFFSET,r19
+	;;
+	ld4 r17=[r17]		// r17 <- current->signal->group_stop_count
+(p7)	mov r8=0		// clear error code
+
+	ld8 r19=[r19]		// r19 <- current->signal->shared_pending
+	;;
+	cmp4.gt p6,p7=r17,r0	// p6/p7 <- (current->signal->group_stop_count > 0)?
+(p8)	mov r8=0		// clear error code
+
+	or r18=r18,r19		// r18 <- current->pending | current->signal->shared_pending
+	;;
+	// r18 <- (current->pending | current->signal->shared_pending) & ~current->blocked:
+	andcm r18=r18,r3
+	add r9=TI_FLAGS+IA64_TASK_SIZE,r16
+	;;
+
+(p7)	cmp.ne.or.andcm p6,p7=r18,r0		// p6/p7 <- signal pending
+	mov r19=0					// i must not leak kernel bits...
+(p6)	br.cond.dpnt.many .sig_pending
+	;;
+
+1:	ld4 r17=[r9]				// r17 <- current->thread_info->flags
+	;;
+	mov ar.ccv=r17
+	and r18=~_TIF_SIGPENDING,r17		// r18 <- r17 & ~(1 << TIF_SIGPENDING)
+	;;
+
+	st8 [r2]=r3				// update current->blocked with new mask
+	cmpxchg4.acq r14=[r9],r18,ar.ccv	// current->thread_info->flags <- r18
+	;;
+	cmp.ne p6,p0=r17,r14			// update failed?
+(p6)	br.cond.spnt.few 1b			// yes -> retry
+
+#ifdef CONFIG_SMP
+	st4.rel [r31]=r0			// release the lock
+#endif
+	ssm psr.i
+	cmp.ne p9,p0=r8,r0			// check for bad HOW value
+	;;
+
+	srlz.d					// ensure psr.i is set again
+	mov r18=0					// i must not leak kernel bits...
+(p9)	br.spnt.few .fail_einval		// bail out for bad HOW value
+
+.store_mask:
+EX(.fail_efault, (p15) probe.w.fault r34, 3)	// verify user has write-access to *oset
+EX(.fail_efault, (p15) st8 [r34]=r3)
+	mov r2=0					// i must not leak kernel bits...
+	mov r3=0					// i must not leak kernel bits...
+	mov r8=0				// return 0
+	mov r9=0					// i must not leak kernel bits...
+	mov r14=0					// i must not leak kernel bits...
+	mov r17=0					// i must not leak kernel bits...
+	mov r31=0					// i must not leak kernel bits...
+	FSYS_RETURN
+
+.sig_pending:
+#ifdef CONFIG_SMP
+	st4.rel [r31]=r0			// release the lock
+#endif
+	ssm psr.i
+	;;
+	srlz.d
+	br.sptk.many fsys_fallback_syscall	// with signal pending, do the heavy-weight syscall
+
+#ifdef CONFIG_SMP
+.lock_contention:
+	/* Rather than spinning here, fall back on doing a heavy-weight syscall.  */
+	ssm psr.i
+	;;
+	srlz.d
+	br.sptk.many fsys_fallback_syscall
+#endif
+END(fsys_rt_sigprocmask)
+
 ENTRY(fsys_fallback_syscall)
 	.prologue
 	.altrp b6
@@ -600,7 +768,7 @@ fsyscall_table:
 	data8 0				// sigaltstack
 	data8 0				// rt_sigaction
 	data8 0				// rt_sigpending
-	data8 0				// rt_sigprocmask
+	data8 fsys_rt_sigprocmask	// rt_sigprocmask
 	data8 0				// rt_sigqueueinfo	// 1180
 	data8 0				// rt_sigreturn
 	data8 0				// rt_sigsuspend

arch/ia64/kernel/gate.S

@@ -118,8 +118,7 @@ GLOBAL_ENTRY(__kernel_syscall_via_epc)
 
 	mov r10=-1
 	mov r8=ENOSYS
-	MCKINLEY_E9_WORKAROUND
-	br.ret.sptk.many b6
+	FSYS_RETURN
 END(__kernel_syscall_via_epc)
 
 #	define ARG0_OFF		(16 + IA64_SIGFRAME_ARG0_OFFSET)

arch/ia64/kernel/head.S

@@ -797,6 +797,25 @@ GLOBAL_ENTRY(ia64_switch_mode_virt)
 	br.ret.sptk.many rp
 END(ia64_switch_mode_virt)
 
+GLOBAL_ENTRY(ia64_delay_loop)
+	.prologue
+{	nop 0			// work around GAS unwind info generation bug...
+	.save ar.lc,r2
+	mov r2=ar.lc
+	.body
+	;;
+	mov ar.lc=r32
+}
+	;;
+	// force loop to be 32-byte aligned (GAS bug means we cannot use .align
+	// inside function body without corrupting unwind info).
+{	nop 0 }
+1:	br.cloop.sptk.few 1b
+	;;
+	mov ar.lc=r2
+	br.ret.sptk.many rp
+END(ia64_delay_loop)
+
 #ifdef CONFIG_IA64_BRL_EMU
 
 /*

arch/ia64/kernel/mca.c

@@ -81,8 +81,6 @@ u64				ia64_init_stack[KERNEL_STACK_SIZE/8] __attribute__((aligned(16)));
 u64				ia64_mca_sal_data_area[1356];
 u64				ia64_tlb_functional;
 u64				ia64_os_mca_recovery_successful;
-/* TODO: need to assign min-state structure to UC memory */
-u64				ia64_mca_min_state_save_info[MIN_STATE_AREA_SIZE] __attribute__((aligned(512)));
 static void			ia64_mca_wakeup_ipi_wait(void);
 static void			ia64_mca_wakeup(int cpu);
 static void			ia64_mca_wakeup_all(void);
@@ -465,26 +463,6 @@ ia64_mca_register_cpev (int cpev)
 
 #endif /* PLATFORM_MCA_HANDLERS */
 
-/*
- * routine to process and prepare to dump min_state_save
- * information for debugging purposes.
- */
-void
-ia64_process_min_state_save (pal_min_state_area_t *pmss)
-{
-	int i, max = MIN_STATE_AREA_SIZE;
-	u64 *tpmss_ptr = (u64 *)pmss;
-	u64 *return_min_state_ptr = ia64_mca_min_state_save_info;
-
-	for (i=0;i<max;i++) {
-
-		/* copy min-state register info for eventual return to PAL */
-		*return_min_state_ptr++ = *tpmss_ptr;
-
-		tpmss_ptr++;  /* skip to next entry */
-	}
-}
-
 /*
  * ia64_mca_cmc_vector_setup
  *
@@ -828,7 +806,7 @@ ia64_mca_wakeup_ipi_wait(void)
 			irr = ia64_getreg(_IA64_REG_CR_IRR3);
 			break;
 		}
-	} while (!(irr & (1 << irr_bit))) ;
+	} while (!(irr & (1UL << irr_bit))) ;
 }
 
 /*
@@ -961,9 +939,8 @@ ia64_return_to_sal_check(void)
 	/* Default = tell SAL to return to same context */
 	ia64_os_to_sal_handoff_state.imots_context = IA64_MCA_SAME_CONTEXT;
 
-	/* Register pointer to new min state values */
 	ia64_os_to_sal_handoff_state.imots_new_min_state =
-		ia64_mca_min_state_save_info;
+		(u64 *)ia64_sal_to_os_handoff_state.pal_min_state;
 }
 
 /*
@@ -2154,9 +2131,6 @@ ia64_log_proc_dev_err_info_print (sal_log_processor_info_t  *slpi,
 	if (slpi->valid.psi_static_struct) {
 		spsi = (sal_processor_static_info_t *)p_data;
 
-		/* copy interrupted context PAL min-state info */
-		ia64_process_min_state_save(&spsi->min_state_area);
-
 		/* Print branch register contents if valid */
 		if (spsi->valid.br)
 			ia64_log_processor_regs_print(spsi->br, 8, "Branch", "br",

arch/ia64/kernel/patch.c

@@ -130,9 +130,11 @@ ia64_patch_mckinley_e9 (unsigned long start, unsigned long end)
 
 	while (offp < (s32 *) end) {
 		wp = (u64 *) ia64_imva((char *) offp + *offp);
-		wp[0] = 0x0000000100000000;
+		wp[0] = 0x0000000100000000; /* nop.m 0; nop.i 0; nop.i 0 */
 		wp[1] = 0x0004000000000200;
-		ia64_fc(wp);
+		wp[2] = 0x0000000100000011; /* nop.m 0; nop.i 0; br.ret.sptk.many b6 */
+		wp[3] = 0x0084006880000200;
+		ia64_fc(wp); ia64_fc(wp + 2);
 		++offp;
 	}
 	ia64_sync_i();

arch/ia64/kernel/perfmon_itanium.h

@@ -81,6 +81,8 @@ pfm_ita_pmc_check(struct task_struct *task, pfm_context_t *ctx, unsigned int cnu
 	 */
 	if (cnum == 13 && ((*val & 0x1) == 0UL) && ctx->ctx_fl_using_dbreg == 0) {
 
+		DPRINT(("pmc[%d]=0x%lx has active pmc13.ta cleared, clearing ibr\n", cnum, *val));
+
 		/* don't mix debug with perfmon */
 		if (task && (task->thread.flags & IA64_THREAD_DBG_VALID) != 0) return -EINVAL;
 
@@ -98,6 +100,8 @@ pfm_ita_pmc_check(struct task_struct *task, pfm_context_t *ctx, unsigned int cnu
 	 */
 	if (cnum == 11 && ((*val >> 28)& 0x1) == 0 && ctx->ctx_fl_using_dbreg == 0) {
 
+		DPRINT(("pmc[%d]=0x%lx has active pmc11.pt cleared, clearing dbr\n", cnum, *val));
+
 		/* don't mix debug with perfmon */
 		if (task && (task->thread.flags & IA64_THREAD_DBG_VALID) != 0) return -EINVAL;
 

arch/ia64/kernel/perfmon_mckinley.h

@@ -109,10 +109,20 @@ pfm_mck_pmc_check(struct task_struct *task, pfm_context_t *ctx, unsigned int cnu
 	if (ctx == NULL) return -EINVAL;
 
 	/*
-	 * we must clear the debug registers if any pmc13.ena_dbrpX bit is enabled
-	 * before they are written (fl_using_dbreg==0) to avoid picking up stale information.
+	 * we must clear the debug registers if pmc13 has a value which enable
+	 * memory pipeline event constraints. In this case we need to clear the
+	 * the debug registers if they have not yet been accessed. This is required
+	 * to avoid picking stale state.
+	 * PMC13 is "active" if:
+	 * 	one of the pmc13.cfg_dbrpXX field is different from 0x3
+	 * AND
+	 * 	at the corresponding pmc13.ena_dbrpXX is set.
+	 *
+	 * For now, we just check on cfg_dbrXX != 0x3.
 	 */
-	if (cnum == 13 && (*val & (0xfUL << 45)) && ctx->ctx_fl_using_dbreg == 0) {
+	if (cnum == 13 && ((*val & 0x18181818UL) != 0x18181818UL) && ctx->ctx_fl_using_dbreg == 0) {
+
+		DPRINT(("pmc[%d]=0x%lx has active pmc13 settings, clearing dbr\n", cnum, *val));
 
 		/* don't mix debug with perfmon */
 		if (task && (task->thread.flags & IA64_THREAD_DBG_VALID) != 0) return -EINVAL;
@@ -128,7 +138,9 @@ pfm_mck_pmc_check(struct task_struct *task, pfm_context_t *ctx, unsigned int cnu
 	 * we must clear the (instruction) debug registers if any pmc14.ibrpX bit is enabled
 	 * before they are (fl_using_dbreg==0) to avoid picking up stale information.
 	 */
-	if (cnum == 14 && ((*val & 0x2222) != 0x2222) && ctx->ctx_fl_using_dbreg == 0) {
+	if (cnum == 14 && ((*val & 0x2222UL) != 0x2222UL) && ctx->ctx_fl_using_dbreg == 0) {
+
+		DPRINT(("pmc[%d]=0x%lx has active pmc14 settings, clearing ibr\n", cnum, *val));
 
 		/* don't mix debug with perfmon */
 		if (task && (task->thread.flags & IA64_THREAD_DBG_VALID) != 0) return -EINVAL;
@@ -170,7 +182,7 @@ pfm_mck_pmc_check(struct task_struct *task, pfm_context_t *ctx, unsigned int cnu
 		   && ((((val14>>1) & 0x3) == 0x2 || ((val14>>1) & 0x3) == 0x0)
 		       ||(((val14>>4) & 0x3) == 0x2 || ((val14>>4) & 0x3) == 0x0));
 
-		if (ret) printk("perfmon: failure check_case1\n");
+		if (ret) DPRINT((KERN_DEBUG "perfmon: failure check_case1\n"));
 	}
 
 	return ret ? -EINVAL : 0;

arch/ia64/kernel/setup.c

@@ -30,6 +30,8 @@
 #include <linux/string.h>
 #include <linux/threads.h>
 #include <linux/tty.h>
+#include <linux/serial.h>
+#include <linux/serial_core.h>
 #include <linux/efi.h>
 #include <linux/initrd.h>
 
@@ -43,6 +45,7 @@
 #include <asm/processor.h>
 #include <asm/sal.h>
 #include <asm/sections.h>
+#include <asm/serial.h>
 #include <asm/smp.h>
 #include <asm/system.h>
 #include <asm/unistd.h>
@@ -101,7 +104,7 @@ int
 filter_rsvd_memory (unsigned long start, unsigned long end, void *arg)
 {
 	unsigned long range_start, range_end, prev_start;
-	void (*func)(unsigned long, unsigned long);
+	void (*func)(unsigned long, unsigned long, int);
 	int i;
 
 #if IGNORE_PFN0
@@ -122,11 +125,7 @@ filter_rsvd_memory (unsigned long start, unsigned long end, void *arg)
 		range_end   = min(end, rsvd_region[i].start);
 
 		if (range_start < range_end)
-#ifdef CONFIG_DISCONTIGMEM
-			call_pernode_memory(__pa(range_start), __pa(range_end), func);
-#else
-			(*func)(__pa(range_start), range_end - range_start);
-#endif
+			call_pernode_memory(__pa(range_start), range_end - range_start, func);
 
 		/* nothing more available in this segment */
 		if (range_end == end) return 0;
@@ -225,6 +224,25 @@ find_initrd (void)
 #endif
 }
 
+#ifdef CONFIG_SERIAL_8250_CONSOLE
+static void __init
+setup_serial_legacy (void)
+{
+	struct uart_port port;
+	unsigned int i, iobase[] = {0x3f8, 0x2f8};
+
+	printk(KERN_INFO "Registering legacy COM ports for serial console\n");
+	memset(&port, 0, sizeof(port));
+	port.iotype = SERIAL_IO_PORT;
+	port.uartclk = BASE_BAUD * 16;
+	for (i = 0; i < ARRAY_SIZE(iobase); i++) {
+		port.line = i;
+		port.iobase = iobase[i];
+		early_serial_setup(&port);
+	}
+}
+#endif
+
 void __init
 setup_arch (char **cmdline_p)
 {
@@ -239,7 +257,6 @@ setup_arch (char **cmdline_p)
 	strlcpy(saved_command_line, *cmdline_p, sizeof(saved_command_line));
 
 	efi_init();
-	find_memory();
 
 #ifdef CONFIG_ACPI_BOOT
 	/* Initialize the ACPI boot-time table parser */
@@ -253,6 +270,8 @@ setup_arch (char **cmdline_p)
 # endif
 #endif /* CONFIG_APCI_BOOT */
 
+	find_memory();
+
 	/* process SAL system table: */
 	ia64_sal_init(efi.sal_systab);
 
@@ -297,11 +316,22 @@ setup_arch (char **cmdline_p)
 #ifdef CONFIG_SERIAL_8250_HCDP
 	if (efi.hcdp) {
 		void setup_serial_hcdp(void *);
-
-		/* Setup the serial ports described by HCDP */
 		setup_serial_hcdp(efi.hcdp);
 	}
 #endif
+#ifdef CONFIG_SERIAL_8250_CONSOLE
+	/*
+	 * Without HCDP, we won't discover any serial ports until the serial driver looks
+	 * in the ACPI namespace.  If ACPI claims there are some legacy devices, register
+	 * the legacy COM ports so serial console works earlier.  This is slightly dangerous
+	 * because we don't *really* know whether there's anything there, but we hope that
+	 * all new boxes will implement HCDP.
+	 */
+	extern unsigned char acpi_legacy_devices;
+	if (!efi.hcdp && acpi_legacy_devices)
+		setup_serial_legacy();
+#endif
+
 #ifdef CONFIG_VT
 # if defined(CONFIG_DUMMY_CONSOLE)
 	conswitchp = &dummy_con;
@@ -544,28 +574,7 @@ cpu_init (void)
 	struct cpuinfo_ia64 *cpu_info;
 	void *cpu_data;
 
-#ifdef CONFIG_SMP
-	int cpu;
-
-	/*
-	 * get_free_pages() cannot be used before cpu_init() done.  BSP allocates
-	 * "NR_CPUS" pages for all CPUs to avoid that AP calls get_zeroed_page().
-	 */
-	if (smp_processor_id() == 0) {
-		cpu_data = __alloc_bootmem(PERCPU_PAGE_SIZE * NR_CPUS, PERCPU_PAGE_SIZE,
-					   __pa(MAX_DMA_ADDRESS));
-		for (cpu = 0; cpu < NR_CPUS; cpu++) {
-			memcpy(cpu_data, __phys_per_cpu_start, __per_cpu_end - __per_cpu_start);
-			__per_cpu_offset[cpu] = (char *) cpu_data - __per_cpu_start;
-			cpu_data += PERCPU_PAGE_SIZE;
-
-			per_cpu(local_per_cpu_offset, cpu) = __per_cpu_offset[cpu];
-		}
-	}
-	cpu_data = __per_cpu_start + __per_cpu_offset[smp_processor_id()];
-#else /* !CONFIG_SMP */
-	cpu_data = __phys_per_cpu_start;
-#endif /* !CONFIG_SMP */
+	cpu_data = per_cpu_init();
 
 	get_max_cacheline_size();
 
@@ -576,9 +585,6 @@ cpu_init (void)
 	 * accessing cpu_data() through the canonical per-CPU address.
 	 */
 	cpu_info = cpu_data + ((char *) &__ia64_per_cpu_var(cpu_info) - __per_cpu_start);
-#ifdef CONFIG_NUMA
-	cpu_info->node_data = get_node_data_ptr();
-#endif
 	identify_cpu(cpu_info);
 
 #ifdef CONFIG_MCKINLEY

arch/ia64/kernel/time.c

@@ -65,8 +65,12 @@ itc_update (long delta_nsec)
 }
 
 /*
- * Return the number of nano-seconds that elapsed since the last update to jiffy.  The
- * xtime_lock must be at least read-locked when calling this routine.
+ * Return the number of nano-seconds that elapsed since the last
+ * update to jiffy.  It is quite possible that the timer interrupt
+ * will interrupt this and result in a race for any of jiffies,
+ * wall_jiffies or itm_next.  Thus, the xtime_lock must be at least
+ * read synchronised when calling this routine (see do_gettimeofday()
+ * below for an example).
  */
 unsigned long
 itc_get_offset (void)
@@ -77,11 +81,6 @@ itc_get_offset (void)
 	last_tick = (cpu_data(TIME_KEEPER_ID)->itm_next
 		     - (lost + 1)*cpu_data(TIME_KEEPER_ID)->itm_delta);
 
-	if (unlikely((long) (now - last_tick) < 0)) {
-		printk(KERN_ERR "CPU %d: now < last_tick (now=0x%lx,last_tick=0x%lx)!\n",
-		       smp_processor_id(), now, last_tick);
-		return last_nsec_offset;
-	}
 	elapsed_cycles = now - last_tick;
 	return (elapsed_cycles*local_cpu_data->nsec_per_cyc) >> IA64_NSEC_PER_CYC_SHIFT;
 }

arch/ia64/mm/contig.c

@@ -25,6 +25,10 @@
 #include <asm/pgtable.h>
 #include <asm/sections.h>
 
+#ifdef CONFIG_VIRTUAL_MEM_MAP
+static unsigned long num_dma_physpages;
+#endif
+
 /**
  * show_mem - display a memory statistics summary
  *
@@ -161,3 +165,133 @@ find_memory (void)
 
 	find_initrd();
 }
+
+#ifdef CONFIG_SMP
+/**
+ * per_cpu_init - setup per-cpu variables
+ *
+ * Allocate and setup per-cpu data areas.
+ */
+void *
+per_cpu_init (void)
+{
+	void *cpu_data;
+	int cpu;
+
+	/*
+	 * get_free_pages() cannot be used before cpu_init() done.  BSP
+	 * allocates "NR_CPUS" pages for all CPUs to avoid that AP calls
+	 * get_zeroed_page().
+	 */
+	if (smp_processor_id() == 0) {
+		cpu_data = __alloc_bootmem(PERCPU_PAGE_SIZE * NR_CPUS,
+					   PERCPU_PAGE_SIZE, __pa(MAX_DMA_ADDRESS));
+		for (cpu = 0; cpu < NR_CPUS; cpu++) {
+			memcpy(cpu_data, __phys_per_cpu_start, __per_cpu_end - __per_cpu_start);
+			__per_cpu_offset[cpu] = (char *) cpu_data - __per_cpu_start;
+			cpu_data += PERCPU_PAGE_SIZE;
+			per_cpu(local_per_cpu_offset, cpu) = __per_cpu_offset[cpu];
+		}
+	}
+	return __per_cpu_start + __per_cpu_offset[smp_processor_id()];
+}
+#endif /* CONFIG_SMP */
+
+static int
+count_pages (u64 start, u64 end, void *arg)
+{
+	unsigned long *count = arg;
+
+	*count += (end - start) >> PAGE_SHIFT;
+	return 0;
+}
+
+#ifdef CONFIG_VIRTUAL_MEM_MAP
+static int
+count_dma_pages (u64 start, u64 end, void *arg)
+{
+	unsigned long *count = arg;
+
+	if (end <= MAX_DMA_ADDRESS)
+		*count += (end - start) >> PAGE_SHIFT;
+	return 0;
+}
+#endif
+
+/*
+ * Set up the page tables.
+ */
+
+void
+paging_init (void)
+{
+	unsigned long max_dma;
+	unsigned long zones_size[MAX_NR_ZONES];
+#ifdef CONFIG_VIRTUAL_MEM_MAP
+	unsigned long zholes_size[MAX_NR_ZONES];
+	unsigned long max_gap;
+#endif
+
+	/* initialize mem_map[] */
+
+	memset(zones_size, 0, sizeof(zones_size));
+
+	num_physpages = 0;
+	efi_memmap_walk(count_pages, &num_physpages);
+
+	max_dma = virt_to_phys((void *) MAX_DMA_ADDRESS) >> PAGE_SHIFT;
+
+#ifdef CONFIG_VIRTUAL_MEM_MAP
+	memset(zholes_size, 0, sizeof(zholes_size));
+
+	num_dma_physpages = 0;
+	efi_memmap_walk(count_dma_pages, &num_dma_physpages);
+
+	if (max_low_pfn < max_dma) {
+		zones_size[ZONE_DMA] = max_low_pfn;
+		zholes_size[ZONE_DMA] = max_low_pfn - num_dma_physpages;
+	} else {
+		zones_size[ZONE_DMA] = max_dma;
+		zholes_size[ZONE_DMA] = max_dma - num_dma_physpages;
+		if (num_physpages > num_dma_physpages) {
+			zones_size[ZONE_NORMAL] = max_low_pfn - max_dma;
+			zholes_size[ZONE_NORMAL] =
+				((max_low_pfn - max_dma) -
+				 (num_physpages - num_dma_physpages));
+		}
+	}
+
+	max_gap = 0;
+	efi_memmap_walk(find_largest_hole, (u64 *)&max_gap);
+	if (max_gap < LARGE_GAP) {
+		vmem_map = (struct page *) 0;
+		free_area_init_node(0, &contig_page_data, NULL, zones_size, 0,
+				    zholes_size);
+		mem_map = contig_page_data.node_mem_map;
+	} else {
+		unsigned long map_size;
+
+		/* allocate virtual_mem_map */
+
+		map_size = PAGE_ALIGN(max_low_pfn * sizeof(struct page));
+		vmalloc_end -= map_size;
+		vmem_map = (struct page *) vmalloc_end;
+		efi_memmap_walk(create_mem_map_page_table, 0);
+
+		free_area_init_node(0, &contig_page_data, vmem_map, zones_size,
+				    0, zholes_size);
+
+		mem_map = contig_page_data.node_mem_map;
+		printk("Virtual mem_map starts at 0x%p\n", mem_map);
+	}
+#else /* !CONFIG_VIRTUAL_MEM_MAP */
+	if (max_low_pfn < max_dma)
+		zones_size[ZONE_DMA] = max_low_pfn;
+	else {
+		zones_size[ZONE_DMA] = max_dma;
+		zones_size[ZONE_NORMAL] = max_low_pfn - max_dma;
+	}
+	free_area_init(zones_size);
+#endif /* !CONFIG_VIRTUAL_MEM_MAP */
+	zero_page_memmap_ptr = virt_to_page(ia64_imva(empty_zero_page));
+}

arch/ia64/mm/hugetlbpage.c

@@ -20,13 +20,46 @@
 
 #define TASK_HPAGE_BASE (REGION_HPAGE << REGION_SHIFT)
 
-static long    htlbpagemem;
-int     htlbpage_max;
-static long    htlbzone_pages;
+static long	htlbpagemem;
+int		htlbpage_max;
+static long	htlbzone_pages;
 
-static LIST_HEAD(htlbpage_freelist);
+static struct list_head hugepage_freelists[MAX_NUMNODES];
 static spinlock_t htlbpage_lock = SPIN_LOCK_UNLOCKED;
 
+static void enqueue_huge_page(struct page *page)
+{
+	list_add(&page->list,
+		&hugepage_freelists[page_zone(page)->zone_pgdat->node_id]);
+}
+
+static struct page *dequeue_huge_page(void)
+{
+	int nid = numa_node_id();
+	struct page *page = NULL;
+
+	if (list_empty(&hugepage_freelists[nid])) {
+		for (nid = 0; nid < MAX_NUMNODES; ++nid)
+			if (!list_empty(&hugepage_freelists[nid]))
+				break;
+	}
+	if (nid >= 0 && nid < MAX_NUMNODES &&
+	    !list_empty(&hugepage_freelists[nid])) {
+		page = list_entry(hugepage_freelists[nid].next, struct page, list);
+		list_del(&page->list);
+	}
+	return page;
+}
+
+static struct page *alloc_fresh_huge_page(void)
+{
+	static int nid = 0;
+	struct page *page;
+	page = alloc_pages_node(nid, GFP_HIGHUSER, HUGETLB_PAGE_ORDER);
+	nid = (nid + 1) % numnodes;
+	return page;
+}
+
 void free_huge_page(struct page *page);
 
 static struct page *alloc_hugetlb_page(void)
@@ -35,13 +68,11 @@ static struct page *alloc_hugetlb_page(void)
 	struct page *page;
 
 	spin_lock(&htlbpage_lock);
-	if (list_empty(&htlbpage_freelist)) {
+	page = dequeue_huge_page();
+	if (!page) {
 		spin_unlock(&htlbpage_lock);
 		return NULL;
 	}
-
-	page = list_entry(htlbpage_freelist.next, struct page, list);
-	list_del(&page->list);
 	htlbpagemem--;
 	spin_unlock(&htlbpage_lock);
 	set_page_count(page, 1);
@@ -228,7 +259,7 @@ void free_huge_page(struct page *page)
 	INIT_LIST_HEAD(&page->list);
 
 	spin_lock(&htlbpage_lock);
-	list_add(&page->list, &htlbpage_freelist);
+	enqueue_huge_page(page);
 	htlbpagemem++;
 	spin_unlock(&htlbpage_lock);
 }
@@ -371,7 +402,7 @@ int try_to_free_low(int count)
 
 	map = NULL;
 	spin_lock(&htlbpage_lock);
-	list_for_each(p, &htlbpage_freelist) {
+	list_for_each(p, &hugepage_freelists[0]) {
 		if (map) {
 			list_del(&map->list);
 			update_and_free_page(map);
@@ -408,11 +439,11 @@ int set_hugetlb_mem_size(int count)
 		return (int)htlbzone_pages;
 	if (lcount > 0) {	/* Increase the mem size. */
 		while (lcount--) {
-			page = alloc_pages(__GFP_HIGHMEM, HUGETLB_PAGE_ORDER);
+			page = alloc_fresh_huge_page();
 			if (page == NULL)
 				break;
 			spin_lock(&htlbpage_lock);
-			list_add(&page->list, &htlbpage_freelist);
+			enqueue_huge_page(page);
 			htlbpagemem++;
 			htlbzone_pages++;
 			spin_unlock(&htlbpage_lock);
@@ -449,17 +480,18 @@ __setup("hugepages=", hugetlb_setup);
 
 static int __init hugetlb_init(void)
 {
-	int i, j;
+	int i;
 	struct page *page;
 
+	for (i = 0; i < MAX_NUMNODES; ++i)
+		INIT_LIST_HEAD(&hugepage_freelists[i]);
+
 	for (i = 0; i < htlbpage_max; ++i) {
-		page = alloc_pages(__GFP_HIGHMEM, HUGETLB_PAGE_ORDER);
+		page = alloc_fresh_huge_page();
 		if (!page)
 			break;
-		for (j = 0; j < HPAGE_SIZE/PAGE_SIZE; ++j)
-			SetPageReserved(&page[j]);
 		spin_lock(&htlbpage_lock);
-		list_add(&page->list, &htlbpage_freelist);
+		enqueue_huge_page(page);
 		spin_unlock(&htlbpage_lock);
 	}
 	htlbpage_max = htlbpagemem = htlbzone_pages = i;

arch/ia64/mm/init.c

@@ -24,6 +24,7 @@
 #include <asm/ia32.h>
 #include <asm/io.h>
 #include <asm/machvec.h>
+#include <asm/numa.h>
 #include <asm/patch.h>
 #include <asm/pgalloc.h>
 #include <asm/sal.h>
@@ -40,10 +41,8 @@ extern void ia64_tlb_init (void);
 unsigned long MAX_DMA_ADDRESS = PAGE_OFFSET + 0x100000000UL;
 
 #ifdef CONFIG_VIRTUAL_MEM_MAP
-# define LARGE_GAP	0x40000000	/* Use virtual mem map if hole is > than this */
   unsigned long vmalloc_end = VMALLOC_END_INIT;
-  static struct page *vmem_map;
-  static unsigned long num_dma_physpages;
+  struct page *vmem_map;
 #endif
 
 static int pgt_cache_water[2] = { 25, 50 };
@@ -337,11 +336,12 @@ ia64_mmu_init (void *my_cpu_data)
 
 #ifdef CONFIG_VIRTUAL_MEM_MAP
 
-static int
+int
 create_mem_map_page_table (u64 start, u64 end, void *arg)
 {
 	unsigned long address, start_page, end_page;
 	struct page *map_start, *map_end;
+	int node;
 	pgd_t *pgd;
 	pmd_t *pmd;
 	pte_t *pte;
@@ -351,19 +351,20 @@ create_mem_map_page_table (u64 start, u64 end, void *arg)
 
 	start_page = (unsigned long) map_start & PAGE_MASK;
 	end_page = PAGE_ALIGN((unsigned long) map_end);
+	node = paddr_to_nid(__pa(start));
 
 	for (address = start_page; address < end_page; address += PAGE_SIZE) {
 		pgd = pgd_offset_k(address);
 		if (pgd_none(*pgd))
-			pgd_populate(&init_mm, pgd, alloc_bootmem_pages(PAGE_SIZE));
+			pgd_populate(&init_mm, pgd, alloc_bootmem_pages_node(NODE_DATA(node), PAGE_SIZE));
 		pmd = pmd_offset(pgd, address);
 
 		if (pmd_none(*pmd))
-			pmd_populate_kernel(&init_mm, pmd, alloc_bootmem_pages(PAGE_SIZE));
+			pmd_populate_kernel(&init_mm, pmd, alloc_bootmem_pages_node(NODE_DATA(node), PAGE_SIZE));
 		pte = pte_offset_kernel(pmd, address);
 
 		if (pte_none(*pte))
-			set_pte(pte, pfn_pte(__pa(alloc_bootmem_pages(PAGE_SIZE)) >> PAGE_SHIFT,
+			set_pte(pte, pfn_pte(__pa(alloc_bootmem_pages_node(NODE_DATA(node), PAGE_SIZE)) >> PAGE_SHIFT,
 					     PAGE_KERNEL));
 	}
 	return 0;
@@ -433,17 +434,7 @@ ia64_pfn_valid (unsigned long pfn)
 	return __get_user(byte, (char *) pfn_to_page(pfn)) == 0;
 }
 
-static int
-count_dma_pages (u64 start, u64 end, void *arg)
-{
-	unsigned long *count = arg;
-
-	if (end <= MAX_DMA_ADDRESS)
-		*count += (end - start) >> PAGE_SHIFT;
-	return 0;
-}
-
-static int
+int
 find_largest_hole (u64 start, u64 end, void *arg)
 {
 	u64 *max_gap = arg;
@@ -459,103 +450,6 @@ find_largest_hole (u64 start, u64 end, void *arg)
 }
 #endif /* CONFIG_VIRTUAL_MEM_MAP */
 
-static int
-count_pages (u64 start, u64 end, void *arg)
-{
-	unsigned long *count = arg;
-
-	*count += (end - start) >> PAGE_SHIFT;
-	return 0;
-}
-
-/*
- * Set up the page tables.
- */
-
-#ifdef CONFIG_DISCONTIGMEM
-void
-paging_init (void)
-{
-	extern void discontig_paging_init(void);
-
-	discontig_paging_init();
-	efi_memmap_walk(count_pages, &num_physpages);
-	zero_page_memmap_ptr = virt_to_page(ia64_imva(empty_zero_page));
-}
-#else /* !CONFIG_DISCONTIGMEM */
-void
-paging_init (void)
-{
-	unsigned long max_dma;
-	unsigned long zones_size[MAX_NR_ZONES];
-#  ifdef CONFIG_VIRTUAL_MEM_MAP
-	unsigned long zholes_size[MAX_NR_ZONES];
-	unsigned long max_gap;
-#  endif
-
-	/* initialize mem_map[] */
-
-	memset(zones_size, 0, sizeof(zones_size));
-
-	num_physpages = 0;
-	efi_memmap_walk(count_pages, &num_physpages);
-
-	max_dma = virt_to_phys((void *) MAX_DMA_ADDRESS) >> PAGE_SHIFT;
-
-#  ifdef CONFIG_VIRTUAL_MEM_MAP
-	memset(zholes_size, 0, sizeof(zholes_size));
-
-	num_dma_physpages = 0;
-	efi_memmap_walk(count_dma_pages, &num_dma_physpages);
-
-	if (max_low_pfn < max_dma) {
-		zones_size[ZONE_DMA] = max_low_pfn;
-		zholes_size[ZONE_DMA] = max_low_pfn - num_dma_physpages;
-	} else {
-		zones_size[ZONE_DMA] = max_dma;
-		zholes_size[ZONE_DMA] = max_dma - num_dma_physpages;
-		if (num_physpages > num_dma_physpages) {
-			zones_size[ZONE_NORMAL] = max_low_pfn - max_dma;
-			zholes_size[ZONE_NORMAL] = ((max_low_pfn - max_dma)
-						    - (num_physpages - num_dma_physpages));
-		}
-	}
-
-	max_gap = 0;
-	efi_memmap_walk(find_largest_hole, (u64 *)&max_gap);
-	if (max_gap < LARGE_GAP) {
-		vmem_map = (struct page *) 0;
-		free_area_init_node(0, &contig_page_data, NULL, zones_size, 0, zholes_size);
-		mem_map = contig_page_data.node_mem_map;
-	}
-	else {
-		unsigned long map_size;
-
-		/* allocate virtual_mem_map */
-
-		map_size = PAGE_ALIGN(max_low_pfn * sizeof(struct page));
-		vmalloc_end -= map_size;
-		vmem_map = (struct page *) vmalloc_end;
-		efi_memmap_walk(create_mem_map_page_table, 0);
-
-		free_area_init_node(0, &contig_page_data, vmem_map, zones_size, 0, zholes_size);
-
-		mem_map = contig_page_data.node_mem_map;
-		printk("Virtual mem_map starts at 0x%p\n", mem_map);
-	}
-#  else /* !CONFIG_VIRTUAL_MEM_MAP */
-	if (max_low_pfn < max_dma)
-		zones_size[ZONE_DMA] = max_low_pfn;
-	else {
-		zones_size[ZONE_DMA] = max_dma;
-		zones_size[ZONE_NORMAL] = max_low_pfn - max_dma;
-	}
-	free_area_init(zones_size);
-#  endif /* !CONFIG_VIRTUAL_MEM_MAP */
-	zero_page_memmap_ptr = virt_to_page(ia64_imva(empty_zero_page));
-}
-#endif /* !CONFIG_DISCONTIGMEM */
-
 static int
 count_reserved_pages (u64 start, u64 end, void *arg)
 {

arch/ia64/sn/io/machvec/pci_bus_cvlink.c

@@ -867,6 +867,9 @@ sn_pci_init (void)
 	int i = 0;
 	struct pci_controller *controller;
 
+	if (!ia64_platform_is("sn2"))
+	    return 0;
+
 	/*
 	 * set pci_raw_ops, etc.
 	 */

arch/ia64/sn/io/sn2/ml_SN_intr.c

@@ -285,7 +285,6 @@ static cpuid_t intr_cpu_choose_node(void)
 cpuid_t intr_heuristic(vertex_hdl_t dev, int req_bit, int *resp_bit)
 {
 	cpuid_t		cpuid;
-	cpuid_t		candidate = CPU_NONE;
 	vertex_hdl_t	pconn_vhdl;
 	pcibr_soft_t	pcibr_soft;
 	int 		bit;
@@ -293,30 +292,32 @@ cpuid_t intr_heuristic(vertex_hdl_t dev, int req_bit, int *resp_bit)
 	/* XXX: gross layering violation.. */
 	if (hwgraph_edge_get(dev, EDGE_LBL_PCI, &pconn_vhdl) == GRAPH_SUCCESS) {
 		pcibr_soft = pcibr_soft_get(pconn_vhdl);
-		if (pcibr_soft && pcibr_soft->bsi_err_intr)
-			candidate = ((hub_intr_t)pcibr_soft->bsi_err_intr)->i_cpuid;
-	}
-
-	if (candidate != CPU_NONE) {
-		/*
-		 * The cpu was chosen already when we assigned
-		 * the error interrupt.
-		 */
-		bit = intr_reserve_level(candidate, req_bit);
-		if (bit >= 0) {
-			*resp_bit = bit;
-			return candidate;
+		if (pcibr_soft && pcibr_soft->bsi_err_intr) {
+			/*
+			 * The cpu was chosen already when we assigned
+			 * the error interrupt.
+			 */
+			cpuid = ((hub_intr_t)pcibr_soft->bsi_err_intr)->i_cpuid;
+			goto done;
 		}
-
-		printk("Cannot target interrupt to target node (%ld).\n",candidate);
-		return CPU_NONE;
 	}
 
 	/*
 	 * Need to choose one.  Try the controlling c-brick first.
 	 */
 	cpuid = intr_cpu_choose_from_node(master_node_get(dev));
-	if (cpuid != CPU_NONE)
-		return cpuid;
-	return intr_cpu_choose_node();
+	if (cpuid == CPU_NONE)
+		cpuid = intr_cpu_choose_node();
+
+ done:
+	if (cpuid != CPU_NONE) {
+		bit = intr_reserve_level(cpuid, req_bit);
+		if (bit >= 0) {
+			*resp_bit = bit;
+			return cpuid;
+		}
+	}
+
+	printk("Cannot target interrupt to target cpu (%ld).\n", cpuid);
+	return CPU_NONE;
 }

arch/ia64/sn/kernel/setup.c

@@ -147,7 +147,6 @@ char drive_info[4*16];
  * Sets up an initial console to aid debugging.  Intended primarily
  * for bringup.  See start_kernel() in init/main.c.
  */
-#if defined(CONFIG_IA64_EARLY_PRINTK_SGI_SN) || defined(CONFIG_IA64_SGI_SN_SIM)
 
 void __init
 early_sn_setup(void)
@@ -189,7 +188,6 @@ early_sn_setup(void)
 		printk(KERN_DEBUG "early_sn_setup: setting master_node_bedrock_address to 0x%lx\n", master_node_bedrock_address);
 	}
 }
-#endif /* CONFIG_IA64_EARLY_PRINTK_SGI_SN */
 
 #ifdef CONFIG_IA64_MCA
 extern int platform_intr_list[];

arch/ia64/sn/kernel/sn2/io.c

@@ -11,6 +11,8 @@
 
 #include <asm/sn/sn2/io.h>
 
+#ifdef CONFIG_IA64_GENERIC
+
 #undef __sn_inb
 #undef __sn_inw
 #undef __sn_inl
@@ -81,3 +83,5 @@ __sn_readq (void *addr)
 {
 	return ___sn_readq (addr);
 }
+
+#endif

arch/m68k/lib/checksum.c

@@ -125,6 +125,7 @@ csum_partial (const unsigned char *buff, int len, unsigned int sum)
 	return(sum);
 }
 
+EXPORT_SYMBOL(csum_partial);
 
 
 /*

arch/m68k/sun3/sbus.c

@@ -10,6 +10,7 @@
  */
 
 #include <linux/types.h>
+#include <linux/compiler.h>
 #include <linux/init.h>
 
 int __init sbus_init(void)

arch/ppc/boot/of1275/map.c

@@ -23,11 +23,10 @@ map(unsigned int phys, unsigned int virt, unsigned int size)
 	char *method;
 	ihandle mmu_ihandle;
 	int misc;
-	unsigned int phys;
-	unsigned int virt;
 	unsigned int size;
+	unsigned int virt;
+	unsigned int phys;
 	int ret0;
-	int ret1;
     } args;
 
     if (of_prom_mmu == 0) {
@@ -36,10 +35,10 @@ map(unsigned int phys, unsigned int virt, unsigned int size)
     }
     args.service = "call-method";
     args.nargs = 6;
-    args.nret = 2;
+    args.nret = 1;
     args.method = "map";
     args.mmu_ihandle = of_prom_mmu;
-    args.misc = -1;
+    args.misc = 0;
     args.phys = phys;
     args.virt = virt;
     args.size = size;

arch/ppc/boot/openfirmware/coffmain.c

@@ -38,9 +38,9 @@ static char heap[SCRATCH_SIZE];
 
 static unsigned long ram_start = 0;
 static unsigned long ram_end = 0x1000000;
-static unsigned long prog_start = 0x800000;
-static unsigned long prog_size = 0x800000;
 
+static unsigned long prog_start = 0x900000;
+static unsigned long prog_size = 0x700000;
 
 typedef void (*kernel_start_t)(int, int, void *);
 

arch/ppc/kernel/head.S

@@ -181,26 +181,8 @@ __after_mmu_off:
 	bl	setup_disp_bat
 #endif
 #else /* CONFIG_POWER4 */
-/*
- * Load up the SDR1 and segment register values now
- * since we don't have the BATs.
- * Also make sure we are running in 32-bit mode.
- */
 	bl	reloc_offset
-	addis	r14,r3,_SDR1@ha		/* get the value from _SDR1 */
-	lwz	r14,_SDR1@l(r14)	/* assume hash table below 4GB */
-	mtspr	SDR1,r14
-	slbia
-	lis	r4,0x2000		/* set pseudo-segment reg 12 */
-	ori	r5,r4,0x0ccc
-	mtsr	12,r5
-	ori	r4,r4,0x0888		/* set pseudo-segment reg 8 */
-	mtsr	8,r4			/* (for access to serial port) */
-	mfmsr	r0
-	clrldi	r0,r0,1
-	sync
-	mtmsr	r0
-	isync
+	bl	initial_mm_power4
 #endif /* CONFIG_POWER4 */
 
 /*
@@ -1637,6 +1619,34 @@ setup_disp_bat:
 #endif /* !defined(CONFIG_APUS) && defined(CONFIG_BOOTX_TEXT) */
 
 #else /* CONFIG_POWER4 */
+/*
+ * Load up the SDR1 and segment register values now
+ * since we don't have the BATs.
+ * Also make sure we are running in 32-bit mode.
+ */
+
+initial_mm_power4:
+	addis	r14,r3,_SDR1@ha		/* get the value from _SDR1 */
+	lwz	r14,_SDR1@l(r14)	/* assume hash table below 4GB */
+	mtspr	SDR1,r14
+	slbia
+	lis	r4,0x2000		/* set pseudo-segment reg 12 */
+	ori	r5,r4,0x0ccc
+	mtsr	12,r5
+	ori	r5,r4,0x0888		/* set pseudo-segment reg 8 */
+	mtsr	8,r5			/* (for access to serial port) */
+	ori	r5,r4,0x0999		/* set pseudo-segment reg 8 */
+	mtsr	9,r5			/* (for access to screen) */
+	mfmsr	r0
+	clrldi	r0,r0,1
+	sync
+	mtmsr	r0
+	isync
+	blr
+
+/*
+ * On 970 (G5), we pre-set a few bits in HID0 & HID1
+ */
 ppc970_setup_hid:
 	li	r0,0
 	sync

arch/ppc/kernel/ppc_ksyms.c

@@ -367,11 +367,17 @@ EXPORT_SYMBOL(next_mmu_context);
 EXPORT_SYMBOL(set_context);
 EXPORT_SYMBOL(handle_mm_fault); /* For MOL */
 EXPORT_SYMBOL_NOVERS(disarm_decr);
+extern long mol_trampoline;
+EXPORT_SYMBOL(mol_trampoline); /* For MOL */
 #ifdef CONFIG_PPC_STD_MMU
 EXPORT_SYMBOL(flush_hash_pages); /* For MOL */
+#ifdef CONFIG_SMP
+extern int mmu_hash_lock;
+EXPORT_SYMBOL(mmu_hash_lock); /* For MOL */
+#endif /* CONFIG_SMP */
 extern long *intercept_table;
 EXPORT_SYMBOL(intercept_table);
-#endif
+#endif /* CONFIG_PPC_STD_MMU */
 EXPORT_SYMBOL(cur_cpu_spec);
 #ifdef CONFIG_PPC_PMAC
 extern unsigned long agp_special_page;

arch/ppc/kernel/process.c

@@ -56,6 +56,7 @@ static struct files_struct init_files = INIT_FILES;
 static struct signal_struct init_signals = INIT_SIGNALS(init_signals);
 static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand);
 struct mm_struct init_mm = INIT_MM(init_mm);
+EXPORT_SYMBOL(init_mm);
 
 /* this is 8kB-aligned so we can get to the thread_info struct
    at the base of it from the stack pointer with 1 integer instruction. */

arch/ppc/mm/hashtable.S

@@ -417,6 +417,21 @@ _GLOBAL(hash_page_patch_C)
 	lwz	r6,next_slot@l(r4)
 	addi	r6,r6,PTE_SIZE
 	andi.	r6,r6,7*PTE_SIZE
+#ifdef CONFIG_POWER4
+	/*
+	 * Since we don't have BATs on POWER4, we rely on always having
+	 * PTEs in the hash table to map the hash table and the code
+	 * that manipulates it in virtual mode, namely flush_hash_page and
+	 * flush_hash_segments.  Otherwise we can get a DSI inside those
+	 * routines which leads to a deadlock on the hash_table_lock on
+	 * SMP machines.  We avoid this by never overwriting the first
+	 * PTE of each PTEG if it is already valid.
+	 *	-- paulus.
+	 */
+	bne	102f
+	li	r6,PTE_SIZE
+102:
+#endif /* CONFIG_POWER4 */
 	stw	r6,next_slot@l(r4)
 	add	r4,r3,r6
 

arch/ppc/mm/ppc_mmu.c

@@ -211,6 +211,17 @@ void __init MMU_init_hw(void)
 #define MIN_N_HPTEG	1024		/* min 64kB hash table */
 #endif
 
+#ifdef CONFIG_POWER4
+	/* The hash table has already been allocated and initialized
+	   in prom.c */
+	n_hpteg = Hash_size >> LG_HPTEG_SIZE;
+	lg_n_hpteg = __ilog2(n_hpteg);
+
+	/* Remove the hash table from the available memory */
+	if (Hash)
+		reserve_phys_mem(__pa(Hash), Hash_size);
+
+#else /* CONFIG_POWER4 */
 	/*
 	 * Allow 1 HPTE (1/8 HPTEG) for each page of memory.
 	 * This is less than the recommended amount, but then
@@ -224,13 +235,7 @@ void __init MMU_init_hw(void)
 		++lg_n_hpteg;		/* round up if not power of 2 */
 		n_hpteg = 1 << lg_n_hpteg;
 	}
-
 	Hash_size = n_hpteg << LG_HPTEG_SIZE;
-	Hash_mask = n_hpteg - 1;
-	hmask = Hash_mask >> (16 - LG_HPTEG_SIZE);
-	mb2 = mb = 32 - LG_HPTEG_SIZE - lg_n_hpteg;
-	if (lg_n_hpteg > 16)
-		mb2 = 16 - LG_HPTEG_SIZE;
 
 	/*
 	 * Find some memory for the hash table.
@@ -240,6 +245,7 @@ void __init MMU_init_hw(void)
 	cacheable_memzero(Hash, Hash_size);
 	_SDR1 = __pa(Hash) | SDR1_LOW_BITS;
 	Hash_end = (PTE *) ((unsigned long)Hash + Hash_size);
+#endif /* CONFIG_POWER4 */
 
 	printk("Total memory = %ldMB; using %ldkB for hash table (at %p)\n",
 	       total_memory >> 20, Hash_size >> 10, Hash);
@@ -249,6 +255,12 @@ void __init MMU_init_hw(void)
 	 * Patch up the instructions in hashtable.S:create_hpte
 	 */
 	if ( ppc_md.progress ) ppc_md.progress("hash:patch", 0x345);
+	Hash_mask = n_hpteg - 1;
+	hmask = Hash_mask >> (16 - LG_HPTEG_SIZE);
+	mb2 = mb = 32 - LG_HPTEG_SIZE - lg_n_hpteg;
+	if (lg_n_hpteg > 16)
+		mb2 = 16 - LG_HPTEG_SIZE;
+
 	hash_page_patch_A[0] = (hash_page_patch_A[0] & ~0xffff)
 		| ((unsigned int)(Hash) >> 16);
 	hash_page_patch_A[1] = (hash_page_patch_A[1] & ~0x7c0) | (mb << 6);

arch/ppc/platforms/pmac_time.c

@@ -166,7 +166,7 @@ via_calibrate_decr(void)
 {
 	struct device_node *vias;
 	volatile unsigned char *via;
-	int count = VIA_TIMER_FREQ_6 / HZ;
+	int count = VIA_TIMER_FREQ_6 / 100;
 	unsigned int dstart, dend;
 
 	vias = find_devices("via-cuda");
@@ -196,7 +196,7 @@ via_calibrate_decr(void)
 		;
 	dend = get_dec();
 
-	tb_ticks_per_jiffy = (dstart - dend) / 6;
+	tb_ticks_per_jiffy = (dstart - dend) / (6 * (HZ/100));
 	tb_to_us = mulhwu_scale_factor(dstart - dend, 60000);
 
 	printk(KERN_INFO "via_calibrate_decr: ticks per jiffy = %u (%u ticks)\n",
@@ -260,7 +260,9 @@ pmac_calibrate_decr(void)
 	 * calibration. That's better since the VIA itself seems
 	 * to be slightly off. --BenH
 	 */
-	if (!machine_is_compatible("MacRISC2"))
+	if (!machine_is_compatible("MacRISC2") &&
+	    !machine_is_compatible("MacRISC3") &&
+	    !machine_is_compatible("MacRISC4"))
 		if (via_calibrate_decr())
 			return;
 

arch/ppc64/kernel/head.S

@@ -937,7 +937,6 @@ _GLOBAL(do_stab_bolted)
 	mfspr	r20,SPRG2
 	mfspr	r21,SPRG1
 	rfid
-_TRACEBACK(do_stab_bolted)
 
 /*
  * r20 points to the PACA, r21 to the exception frame,
@@ -1052,7 +1051,6 @@ SLB_NUM_ENTRIES = 64
 	mfspr	r20,SPRG2
 	mfspr	r21,SPRG1
 	rfid
-_TRACEBACK(do_slb_bolted)
 
 _GLOBAL(do_stab_SI)
 	mflr	r21			/* Save LR in r21 */