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

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

arch/ia64/config.in

@@ -136,6 +136,17 @@ else
 fi
 endmenu
 
+else # ! HP_SIM
+mainmenu_option next_comment
+comment 'Block devices'
+tristate 'Loopback device support' CONFIG_BLK_DEV_LOOP
+dep_tristate 'Network block device support' CONFIG_BLK_DEV_NBD $CONFIG_NET
+
+tristate 'RAM disk support' CONFIG_BLK_DEV_RAM
+if [ "$CONFIG_BLK_DEV_RAM" = "y" -o "$CONFIG_BLK_DEV_RAM" = "m" ]; then
+   int '  Default RAM disk size' CONFIG_BLK_DEV_RAM_SIZE 4096
+fi
+endmenu
 fi # !HP_SIM
 
 mainmenu_option next_comment

arch/ia64/ia32/ia32_entry.S

 #include <asm/asmmacro.h>
 #include <asm/offsets.h>
 #include <asm/signal.h>
+#include <asm/thread_info.h>
 
 #include "../kernel/minstate.h"
 
@@ -87,18 +88,21 @@ END(sys32_sigsuspend)
 
 GLOBAL_ENTRY(ia32_ret_from_clone)
 	PT_REGS_UNWIND_INFO(0)
+#if defined(CONFIG_SMP) || defined(CONFIG_PREEMPT)
 	/*
 	 * We need to call schedule_tail() to complete the scheduling process.
 	 * Called by ia64_switch_to after do_fork()->copy_thread().  r8 contains the
 	 * address of the previously executing task.
 	 */
 	br.call.sptk.many rp=ia64_invoke_schedule_tail
-.ret1:	adds r2=IA64_TASK_PTRACE_OFFSET,r13
+.ret1:
+#endif
+	adds r2=TI_FLAGS+IA64_TASK_SIZE,r13
 	;;
-	ld8 r2=[r2]
+	ld4 r2=[r2]
 	;;
 	mov r8=0
-	tbit.nz p6,p0=r2,PT_SYSCALLTRACE_BIT
+	tbit.nz p6,p0=r2,TIF_SYSCALL_TRACE
 (p6)	br.cond.spnt .ia32_strace_check_retval
 	;;					// prevent RAW on r8
 END(ia32_ret_from_clone)

arch/ia64/kernel/entry.S

@@ -486,6 +486,7 @@ END(ia64_trace_syscall)
 
 GLOBAL_ENTRY(ia64_ret_from_clone)
 	PT_REGS_UNWIND_INFO(0)
+#if defined(CONFIG_SMP) || defined(CONFIG_PREEMPT)
 	/*
 	 * We need to call schedule_tail() to complete the scheduling process.
 	 * Called by ia64_switch_to after do_fork()->copy_thread().  r8 contains the
@@ -493,6 +494,7 @@ GLOBAL_ENTRY(ia64_ret_from_clone)
 	 */
 	br.call.sptk.many rp=ia64_invoke_schedule_tail
 .ret8:
+#endif
 	adds r2=TI_FLAGS+IA64_TASK_SIZE,r13
 	;;
 	ld4 r2=[r2]
@@ -620,7 +622,7 @@ GLOBAL_ENTRY(ia64_leave_kernel)
 	shr.u r18=r19,16	// get byte size of existing "dirty" partition
 	;;
 	mov r16=ar.bsp		// get existing backing store pointer
-	movl r17=PERCPU_ADDR+IA64_CPU_PHYS_STACKED_SIZE_P8_OFFSET
+	movl r17=THIS_CPU(ia64_phys_stacked_size_p8)
 	;;
 	ld4 r17=[r17]		// r17 = cpu_data->phys_stacked_size_p8
 (pKern)	br.cond.dpnt skip_rbs_switch
@@ -756,6 +758,7 @@ ENTRY(handle_syscall_error)
 	br.cond.sptk ia64_leave_kernel
 END(handle_syscall_error)
 
+#ifdef CONFIG_SMP
 	/*
 	 * Invoke schedule_tail(task) while preserving in0-in7, which may be needed
 	 * in case a system call gets restarted.
@@ -772,6 +775,8 @@ GLOBAL_ENTRY(ia64_invoke_schedule_tail)
 	br.ret.sptk.many rp
 END(ia64_invoke_schedule_tail)
 
+#endif /* CONFIG_SMP */
+
 #if __GNUC__ < 3
 
 	/*

arch/ia64/kernel/ia64_ksyms.c

@@ -59,7 +59,7 @@ EXPORT_SYMBOL(clear_page);
 
 #include <asm/processor.h>
 # ifndef CONFIG_NUMA
-EXPORT_SYMBOL(_cpu_data);
+EXPORT_SYMBOL(cpu_info);
 # endif
 EXPORT_SYMBOL(kernel_thread);
 

arch/ia64/kernel/ivt.S

@@ -645,7 +645,6 @@ ENTRY(break_fault)
 
 	mov r3=255
 	adds r15=-1024,r15			// r15 contains the syscall number---subtract 1024
-	adds r2=IA64_TASK_PTRACE_OFFSET,r13	// r2 = &current->ptrace
 	;;
 	cmp.geu p6,p7=r3,r15		// (syscall > 0 && syscall <= 1024+255) ?
 	movl r16=sys_call_table

arch/ia64/kernel/process.c

@@ -193,7 +193,10 @@ ia64_save_extra (struct task_struct *task)
 	if ((task->thread.flags & IA64_THREAD_PM_VALID) != 0)
 		pfm_save_regs(task);
 
-	if (local_cpu_data->pfm_syst_wide) pfm_syst_wide_update_task(task, 0);
+# ifdef CONFIG_SMP
+	if (local_cpu_data->pfm_syst_wide)
+		pfm_syst_wide_update_task(task, 0);
+# endif
 #endif
 
 	if (IS_IA32_PROCESS(ia64_task_regs(task)))
@@ -210,7 +213,9 @@ ia64_load_extra (struct task_struct *task)
 	if ((task->thread.flags & IA64_THREAD_PM_VALID) != 0)
 		pfm_load_regs(task);
 
+# ifdef CONFIG_SMP
 	if (local_cpu_data->pfm_syst_wide) pfm_syst_wide_update_task(task, 1);
+# endif
 #endif
 
 	if (IS_IA32_PROCESS(ia64_task_regs(task)))

arch/ia64/kernel/setup.c

@@ -54,12 +54,10 @@
 
 extern char _end;
 
-#ifdef CONFIG_NUMA
- struct cpuinfo_ia64 *boot_cpu_data;
-#else
- struct cpuinfo_ia64 _cpu_data[NR_CPUS] __attribute__ ((section ("__special_page_section")));
-#endif
+unsigned long __per_cpu_offset[NR_CPUS];
+struct cpuinfo_ia64 cpu_info __per_cpu_data;
 
+unsigned long ia64_phys_stacked_size_p8;
 unsigned long ia64_cycles_per_usec;
 struct ia64_boot_param *ia64_boot_param;
 struct screen_info screen_info;
@@ -511,6 +509,12 @@ identify_cpu (struct cpuinfo_ia64 *c)
 	c->unimpl_pa_mask = ~((1L<<63) | ((1L << phys_addr_size) - 1));
 }
 
+void
+setup_per_cpu_areas (void)
+{
+	/* start_kernel() requires this... */
+}
+
 /*
  * cpu_init() initializes state that is per-CPU.  This function acts
  * as a 'CPU state barrier', nothing should get across.
@@ -518,46 +522,21 @@ identify_cpu (struct cpuinfo_ia64 *c)
 void
 cpu_init (void)
 {
+	extern char __per_cpu_start[], __phys_per_cpu_start[], __per_cpu_end[];
 	extern void __init ia64_mmu_init (void *);
 	unsigned long num_phys_stacked;
 	pal_vm_info_2_u_t vmi;
 	unsigned int max_ctx;
-	struct cpuinfo_ia64 *my_cpu_data;
-#ifdef CONFIG_NUMA
-	int cpu, order;
+	struct cpuinfo_ia64 *my_cpu_info;
+	void *my_cpu_data;
+	int cpu = smp_processor_id();
 
-	/*
-	 * If NUMA is configured, the cpu_data array is not preallocated. The boot cpu
-	 * allocates entries for every possible cpu. As the remaining cpus come online,
-	 * they reallocate a new cpu_data structure on their local node. This extra work
-	 * is required because some boot code references all cpu_data structures
-	 * before the cpus are actually started.
-	 */
-	if (!boot_cpu_data) {
-		my_cpu_data = alloc_bootmem_pages_node(NODE_DATA(numa_node_id()),
-						       sizeof(struct cpuinfo_ia64));
-		boot_cpu_data = my_cpu_data;
-		my_cpu_data->cpu_data[0] = my_cpu_data;
-		for (cpu = 1; cpu < NR_CPUS; ++cpu)
-			my_cpu_data->cpu_data[cpu]
-				= alloc_bootmem_pages_node(NODE_DATA(numa_node_id()),
-							   sizeof(struct cpuinfo_ia64));
-		for (cpu = 1; cpu < NR_CPUS; ++cpu)
-			memcpy(my_cpu_data->cpu_data[cpu]->cpu_data,
-			       my_cpu_data->cpu_data, sizeof(my_cpu_data->cpu_data));
-	} else {
-		order = get_order(sizeof(struct cpuinfo_ia64));
-		my_cpu_data = page_address(alloc_pages_node(numa_node_id(), GFP_KERNEL, order));
-		memcpy(my_cpu_data, boot_cpu_data->cpu_data[smp_processor_id()],
-		       sizeof(struct cpuinfo_ia64));
-		__free_pages(virt_to_page(boot_cpu_data->cpu_data[smp_processor_id()]),
-			     order);
-		for (cpu = 0; cpu < NR_CPUS; ++cpu)
-			boot_cpu_data->cpu_data[cpu]->cpu_data[smp_processor_id()] = my_cpu_data;
-	}
-#else
-	my_cpu_data = cpu_data(smp_processor_id());
-#endif
+	my_cpu_data = alloc_bootmem_pages(__per_cpu_end - __per_cpu_start);
+	memcpy(my_cpu_data, __phys_per_cpu_start, __per_cpu_end - __per_cpu_start);
+
+	__per_cpu_offset[cpu] = (char *) my_cpu_data - __per_cpu_start;
+
+	my_cpu_info = my_cpu_data + ((char *) &cpu_info - __per_cpu_start);
 
 	/*
 	 * We can't pass "local_cpu_data" to identify_cpu() because we haven't called
@@ -565,7 +544,7 @@ cpu_init (void)
 	 * depends on the data returned by identify_cpu().  We break the dependency by
 	 * accessing cpu_data() the old way, through identity mapped space.
 	 */
-	identify_cpu(my_cpu_data);
+	identify_cpu(my_cpu_info);
 
 	/* Clear the stack memory reserved for pt_regs: */
 	memset(ia64_task_regs(current), 0, sizeof(struct pt_regs));
@@ -626,7 +605,7 @@ cpu_init (void)
 		printk ("cpu_init: PAL RSE info failed, assuming 96 physical stacked regs\n");
 		num_phys_stacked = 96;
 	}
-	local_cpu_data->phys_stacked_size_p8 = num_phys_stacked*8 + 8;
-
+	/* size of physical stacked register partition plus 8 bytes: */
+	ia64_phys_stacked_size_p8 = num_phys_stacked*8 + 8;
 	platform_cpu_init();
 }

arch/ia64/kernel/smp.c

@@ -75,12 +75,11 @@ struct call_data_struct {
 
 static volatile struct call_data_struct *call_data;
 
-static spinlock_t migration_lock = SPIN_LOCK_UNLOCKED;
-static task_t *migrating_task;
-
 #define IPI_CALL_FUNC		0
 #define IPI_CPU_STOP		1
-#define IPI_MIGRATE_TASK	2
+
+/* This needs to be cacheline aligned because it is written to by *other* CPUs.  */
+static __u64 ipi_operation __per_cpu_data ____cacheline_aligned;
 
 static void
 stop_this_cpu (void)
@@ -99,7 +98,7 @@ void
 handle_IPI (int irq, void *dev_id, struct pt_regs *regs)
 {
 	int this_cpu = smp_processor_id();
-	unsigned long *pending_ipis = &local_cpu_data->ipi_operation;
+	unsigned long *pending_ipis = &ipi_operation;
 	unsigned long ops;
 
 	/* Count this now; we may make a call that never returns. */
@@ -143,14 +142,6 @@ handle_IPI (int irq, void *dev_id, struct pt_regs *regs)
 			      }
 			      break;
 
-			      case IPI_MIGRATE_TASK:
-			      {
-				      task_t *p = migrating_task;
-				      spin_unlock(&migration_lock);
-				      sched_task_migrated(p);
-			      }
-			      break;
-
 			      case IPI_CPU_STOP:
 				stop_this_cpu();
 				break;
@@ -167,7 +158,7 @@ handle_IPI (int irq, void *dev_id, struct pt_regs *regs)
 static inline void
 send_IPI_single (int dest_cpu, int op)
 {
-	set_bit(op, &cpu_data(dest_cpu)->ipi_operation);
+	set_bit(op, &ipi_operation);
 	platform_send_ipi(dest_cpu, IA64_IPI_VECTOR, IA64_IPI_DM_INT, 0);
 }
 
@@ -350,15 +341,6 @@ smp_send_stop (void)
 	smp_num_cpus = 1;
 }
 
-void
-smp_migrate_task (int cpu, task_t *p)
-{
-	/* The target CPU will unlock the migration spinlock: */
-	spin_lock(&migration_lock);
-	migrating_task = p;
-	send_IPI_single(cpu, IPI_MIGRATE_TASK);
-}
-
 int __init
 setup_profiling_timer (unsigned int multiplier)
 {

arch/ia64/lib/clear_page.S

@@ -18,7 +18,7 @@
 # define PREFETCH_LINES	9	// magic number
 #else
 # define L3_LINE_SIZE	128	// McKinley L3 line size
-# define PREFETCH_LINES	7	// magic number
+# define PREFETCH_LINES	12	// magic number
 #endif
 
 #define saved_lc	r2

arch/ia64/mm/init.c

@@ -103,12 +103,12 @@ free_initmem (void)
 		free_page(addr);
 		++totalram_pages;
 	}
-	printk ("Freeing unused kernel memory: %ldkB freed\n",
+	printk(KERN_INFO "Freeing unused kernel memory: %ldkB freed\n",
 	       (&__init_end - &__init_begin) >> 10);
 }
 
 void
-free_initrd_mem(unsigned long start, unsigned long end)
+free_initrd_mem (unsigned long start, unsigned long end)
 {
 	/*
 	 * EFI uses 4KB pages while the kernel can use 4KB  or bigger.
@@ -145,7 +145,7 @@ free_initrd_mem(unsigned long start, unsigned long end)
 	end = end & PAGE_MASK;
 
 	if (start < end)
-		printk ("Freeing initrd memory: %ldkB freed\n", (end - start) >> 10);
+		printk(KERN_INFO "Freeing initrd memory: %ldkB freed\n", (end - start) >> 10);
 
 	for (; start < end; start += PAGE_SIZE) {
 		if (!VALID_PAGE(virt_to_page(start)))

arch/ia64/tools/print_offsets.awk

@@ -7,12 +7,6 @@ BEGIN {
 	print " * This file was generated by arch/ia64/tools/print_offsets.awk."
 	print " *"
 	print " */"
-	#
-	# This is a cheesy hack.  Make sure that
-	# PT_PTRACED == 1<<PT_PTRACED_BIT.
-	#
-	print "#define PT_PTRACED_BIT		0"
-	print "#define PT_SYSCALLTRACE_BIT	1"
 }
 
 # look for .tab:

arch/ia64/tools/print_offsets.c

@@ -52,14 +52,7 @@ tab[] =
     { "SIGFRAME_SIZE",			sizeof (struct sigframe) },
     { "UNW_FRAME_INFO_SIZE",		sizeof (struct unw_frame_info) },
     { "", 0 },			/* spacer */
-    { "IA64_TASK_PTRACE_OFFSET",	offsetof (struct task_struct, ptrace) },
-    { "IA64_TASK_THREAD_OFFSET",	offsetof (struct task_struct, thread) },
     { "IA64_TASK_THREAD_KSP_OFFSET",	offsetof (struct task_struct, thread.ksp) },
-#ifdef CONFIG_PERFMON
-    { "IA64_TASK_PFM_OVFL_BLOCK_RESET_OFFSET",offsetof(struct task_struct, thread.pfm_ovfl_block_reset) },
-#endif
-    { "IA64_TASK_PID_OFFSET",		offsetof (struct task_struct, pid) },
-    { "IA64_TASK_MM_OFFSET",		offsetof (struct task_struct, mm) },
     { "IA64_PT_REGS_CR_IPSR_OFFSET",	offsetof (struct pt_regs, cr_ipsr) },
     { "IA64_PT_REGS_CR_IIP_OFFSET",	offsetof (struct pt_regs, cr_iip) },
     { "IA64_PT_REGS_CR_IFS_OFFSET",	offsetof (struct pt_regs, cr_ifs) },
@@ -169,9 +162,6 @@ tab[] =
     { "IA64_SIGFRAME_SIGCONTEXT_OFFSET",	offsetof (struct sigframe, sc) },
     { "IA64_CLONE_VFORK",		CLONE_VFORK },
     { "IA64_CLONE_VM",			CLONE_VM },
-    { "IA64_CPU_IRQ_COUNT_OFFSET",	offsetof (struct cpuinfo_ia64, irq_stat.f.irq_count) },
-    { "IA64_CPU_BH_COUNT_OFFSET",	offsetof (struct cpuinfo_ia64, irq_stat.f.bh_count) },
-    { "IA64_CPU_PHYS_STACKED_SIZE_P8_OFFSET",offsetof (struct cpuinfo_ia64, phys_stacked_size_p8)},
 };
 
 static const char *tabs = "\t\t\t\t\t\t\t\t\t\t";
@@ -189,16 +179,6 @@ main (int argc, char **argv)
   printf ("/*\n * DO NOT MODIFY\n *\n * This file was generated by "
 	  "arch/ia64/tools/print_offsets.\n *\n */\n\n");
 
-  /* This is stretching things a bit, but entry.S needs the bit number
-     for PT_PTRACED and it can't include <linux/sched.h> so this seems
-     like a reasonably solution.  At least the code won't break in
-     subtle ways should PT_PTRACED ever change.  Ditto for
-     PT_TRACESYS_BIT. */
-  printf ("#define PT_PTRACED_BIT\t\t\t%u\n", ffs (PT_PTRACED) - 1);
-#if 0
-  printf ("#define PT_SYSCALLTRACE_BIT\t\t\t%u\n\n", ffs (PT_SYSCALLTRACE) - 1);
-#endif
-
   for (i = 0; i < sizeof (tab) / sizeof (tab[0]); ++i)
     {
       if (tab[i].name[0] == '\0')

arch/ia64/vmlinux.lds.S

 #include <linux/config.h>
 
+#include <asm/cache.h>
 #include <asm/ptrace.h>
 #include <asm/system.h>
 
@@ -65,16 +66,6 @@ SECTIONS
   machvec_end = .;
 #endif
 
-  __start___ksymtab = .;	/* Kernel symbol table */
-  __ksymtab : AT(ADDR(__ksymtab) - PAGE_OFFSET)
-	{ *(__ksymtab) }
-  __stop___ksymtab = .;
-
-  __start___kallsyms = .;	/* All kernel symbols for debugging */
-  __kallsyms : AT(ADDR(__kallsyms) - PAGE_OFFSET)
-	{ *(__kallsyms) }
-  __stop___kallsyms = .;
-
   /* Unwind info & table: */
   . = ALIGN(8);
   .IA_64.unwind_info : AT(ADDR(.IA_64.unwind_info) - PAGE_OFFSET)
@@ -124,10 +115,7 @@ SECTIONS
   .data.init_task : AT(ADDR(.data.init_task) - PAGE_OFFSET)
 	{ *(.data.init_task) }
 
-  .data.page_aligned : AT(ADDR(.data.page_aligned) - PAGE_OFFSET)
-        { *(.data.idt) }
-
-  . = ALIGN(64);
+  . = ALIGN(SMP_CACHE_BYTES);
   .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - PAGE_OFFSET)
         { *(.data.cacheline_aligned) }
 
@@ -135,6 +123,17 @@ SECTIONS
   .kstrtab : AT(ADDR(.kstrtab) - PAGE_OFFSET)
 	{ *(.kstrtab) }
 
+  /* Per-cpu data: */
+  . = ALIGN(PAGE_SIZE);
+  __phys_per_cpu_start = .;
+  .data.percpu PERCPU_ADDR : AT(__phys_per_cpu_start - PAGE_OFFSET)
+	{
+		__per_cpu_start = .;
+		*(.data.percpu)
+		__per_cpu_end = .;
+	}
+  . = __phys_per_cpu_start + 4096;	/* ensure percpu fits into smallest page size (4KB) */
+
   .data : AT(ADDR(.data) - PAGE_OFFSET)
 	{ *(.data) *(.gnu.linkonce.d*) CONSTRUCTORS }
 
@@ -151,7 +150,7 @@ SECTIONS
 	{ *(.sbss) *(.scommon) }
   .bss : AT(ADDR(.bss) - PAGE_OFFSET)
 	{ *(.bss) *(COMMON) }
-  . = ALIGN(64 / 8);
+
   _end = .;
 
   /* Stabs debugging sections.  */

include/asm-ia64/cache.h

@@ -5,7 +5,7 @@
 
 /*
  * Copyright (C) 1998-2000 Hewlett-Packard Co
- * Copyright (C) 1998-2000 David Mosberger-Tang <davidm@hpl.hp.com>
+ *	David Mosberger-Tang <davidm@hpl.hp.com>
  */
 
 /* Bytes per L1 (data) cache line.  */

include/asm-ia64/hardirq.h

@@ -2,8 +2,8 @@
 #define _ASM_IA64_HARDIRQ_H
 
 /*
- * Copyright (C) 1998-2001 Hewlett-Packard Co
- * Copyright (C) 1998-2001 David Mosberger-Tang <davidm@hpl.hp.com>
+ * Copyright (C) 1998-2002 Hewlett-Packard Co
+ *	David Mosberger-Tang <davidm@hpl.hp.com>
  */
 
 #include <linux/config.h>

include/asm-ia64/offsets.h

@@ -6,23 +6,16 @@
  * This file was generated by arch/ia64/tools/print_offsets.awk.
  *
  */
-#define PT_PTRACED_BIT		0
-#define PT_SYSCALLTRACE_BIT	1
 #define IA64_TASK_SIZE			3936	/* 0xf60 */
 #define IA64_THREAD_INFO_SIZE		24	/* 0x18 */
 #define IA64_PT_REGS_SIZE		400	/* 0x190 */
 #define IA64_SWITCH_STACK_SIZE		560	/* 0x230 */
 #define IA64_SIGINFO_SIZE		128	/* 0x80 */
-#define IA64_CPU_SIZE			16384	/* 0x4000 */
+#define IA64_CPU_SIZE			224	/* 0xe0 */
 #define SIGFRAME_SIZE			2816	/* 0xb00 */
 #define UNW_FRAME_INFO_SIZE		448	/* 0x1c0 */
 
-#define IA64_TASK_PTRACE_OFFSET		32	/* 0x20 */
-#define IA64_TASK_THREAD_OFFSET		1472	/* 0x5c0 */
 #define IA64_TASK_THREAD_KSP_OFFSET	1480	/* 0x5c8 */
-#define IA64_TASK_PFM_OVFL_BLOCK_RESET_OFFSET 2096	/* 0x830 */
-#define IA64_TASK_PID_OFFSET		212	/* 0xd4 */
-#define IA64_TASK_MM_OFFSET		136	/* 0x88 */
 #define IA64_PT_REGS_CR_IPSR_OFFSET	0	/* 0x0 */
 #define IA64_PT_REGS_CR_IIP_OFFSET	8	/* 0x8 */
 #define IA64_PT_REGS_CR_IFS_OFFSET	16	/* 0x10 */
@@ -132,8 +125,5 @@
 #define IA64_SIGFRAME_SIGCONTEXT_OFFSET	160	/* 0xa0 */
 #define IA64_CLONE_VFORK		16384	/* 0x4000 */
 #define IA64_CLONE_VM			256	/* 0x100 */
-#define IA64_CPU_IRQ_COUNT_OFFSET	0	/* 0x0 */
-#define IA64_CPU_BH_COUNT_OFFSET	4	/* 0x4 */
-#define IA64_CPU_PHYS_STACKED_SIZE_P8_OFFSET 12	/* 0xc */
 
 #endif /* _ASM_IA64_OFFSETS_H */

include/asm-ia64/processor.h

@@ -15,6 +15,8 @@
 
 #include <linux/config.h>
 
+#include <linux/compiler.h>
+
 #include <asm/ptrace.h>
 #include <asm/kregs.h>
 #include <asm/system.h>
@@ -184,6 +186,10 @@
  */
 #define IA64_USEC_PER_CYC_SHIFT	41
 
+#define __HAVE_ARCH_PER_CPU
+
+#define THIS_CPU(var)	(var)
+
 #ifndef __ASSEMBLY__
 
 #include <linux/threads.h>
@@ -196,6 +202,11 @@
 #include <asm/unwind.h>
 #include <asm/atomic.h>
 
+extern unsigned long __per_cpu_offset[NR_CPUS];
+
+#define per_cpu(var, cpu)	(*(__typeof__(&(var))) ((void *) &(var) + __per_cpu_offset[cpu]))
+#define this_cpu(var)		(var)
+
 /* like above but expressed as bitfields for more efficient access: */
 struct ia64_psr {
 	__u64 reserved0 : 1;
@@ -239,7 +250,7 @@ struct ia64_psr {
  * CPU type, hardware bug flags, and per-CPU state.  Frequently used
  * state comes earlier:
  */
-struct cpuinfo_ia64 {
+extern struct cpuinfo_ia64 {
 	/* irq_stat must be 64-bit aligned */
 	union {
 		struct {
@@ -249,7 +260,6 @@ struct cpuinfo_ia64 {
 		__u64 irq_and_bh_counts;
 	} irq_stat;
 	__u32 softirq_pending;
-	__u32 phys_stacked_size_p8;	/* size of physical stacked registers + 8 */
 	__u64 itm_delta;	/* # of clock cycles between clock ticks */
 	__u64 itm_next;		/* interval timer mask value to use for next clock tick */
 	__u64 *pgd_quick;
@@ -282,41 +292,15 @@ struct cpuinfo_ia64 {
 	__u64 prof_multiplier;
 	__u32 pfm_syst_wide;
 	__u32 pfm_dcr_pp;
-	/* this is written to by *other* CPUs: */
-	__u64 ipi_operation ____cacheline_aligned;
-#endif
-#ifdef CONFIG_NUMA
-	void *node_directory;
-	int numa_node_id;
-	struct cpuinfo_ia64 *cpu_data[NR_CPUS];
 #endif
-	/* Platform specific word.  MUST BE LAST IN STRUCT */
-	__u64 platform_specific;
-} __attribute__ ((aligned (PAGE_SIZE))) ;
+} cpu_info __per_cpu_data;
 
 /*
  * The "local" data pointer.  It points to the per-CPU data of the currently executing
  * CPU, much like "current" points to the per-task data of the currently executing task.
  */
-#define local_cpu_data		((struct cpuinfo_ia64 *) PERCPU_ADDR)
-
-/*
- * On NUMA systems, cpu_data for each cpu is allocated during cpu_init() & is allocated on
- * the node that contains the cpu. This minimizes off-node memory references.  cpu_data
- * for each cpu contains an array of pointers to the cpu_data structures of each of the
- * other cpus.
- *
- * On non-NUMA systems, cpu_data is a static array allocated at compile time.  References
- * to the cpu_data of another cpu is done by direct references to the appropriate entry of
- * the array.
- */
-#ifdef CONFIG_NUMA
-# define cpu_data(cpu)		local_cpu_data->cpu_data[cpu]
-# define numa_node_id()		(local_cpu_data->numa_node_id)
-#else
-  extern struct cpuinfo_ia64 _cpu_data[NR_CPUS];
-# define cpu_data(cpu)		(&_cpu_data[cpu])
-#endif
+#define local_cpu_data		(&this_cpu(cpu_info))
+#define cpu_data(cpu)		(&per_cpu(cpu_info, cpu))
 
 extern void identify_cpu (struct cpuinfo_ia64 *);
 extern void print_cpu_info (struct cpuinfo_ia64 *);

include/asm-ia64/siginfo.h

@@ -136,6 +136,7 @@ typedef struct siginfo {
 #define SI_ASYNCIO	(-4)		/* sent by AIO completion */
 #define SI_SIGIO	(-5)		/* sent by queued SIGIO */
 #define SI_TKILL	(-6)		/* sent by tkill system call */
+#define SI_DETHREAD	(-7)		/* sent by execve() killing subsidiary threads */
 
 #define SI_FROMUSER(siptr)	((siptr)->si_code <= 0)
 #define SI_FROMKERNEL(siptr)	((siptr)->si_code > 0)

include/asm-ia64/smplock.h

@@ -12,30 +12,36 @@
 
 extern spinlock_t kernel_flag;
 
-#define kernel_locked()		spin_is_locked(&kernel_flag)
+#ifdef CONFIG_SMP
+# define kernel_locked()	spin_is_locked(&kernel_flag)
+# define check_irq_holder(cpu)			\
+do {						\
+	if (global_irq_holder == (cpu))		\
+		BUG();				\
+} while (0)
+#else
+# define kernel_locked()	(1)
+#endif
 
 /*
  * Release global kernel lock and global interrupt lock
  */
-static __inline__ void
-release_kernel_lock(struct task_struct *task, int cpu)
-{
-	if (unlikely(task->lock_depth >= 0)) {
-		spin_unlock(&kernel_flag);
-		if (global_irq_holder == (cpu))	\
-			BUG();			\
-	}
-}
+#define release_kernel_lock(task, cpu)		\
+do {						\
+	if (unlikely(task->lock_depth >= 0)) {	\
+		spin_unlock(&kernel_flag);	\
+		check_irq_holder(cpu);		\
+	}					\
+} while (0)
 
 /*
  * Re-acquire the kernel lock
  */
-static __inline__ void
-reacquire_kernel_lock(struct task_struct *task)
-{
-	if (unlikely(task->lock_depth >= 0))
-		spin_lock(&kernel_flag);
-}
+#define reacquire_kernel_lock(task)		\
+do {						\
+	if (unlikely(task->lock_depth >= 0))	\
+		spin_lock(&kernel_flag);	\
+} while (0)
 
 /*
  * Getting the big kernel lock.

kernel/sched.c

@@ -1451,8 +1451,10 @@ void __init init_idle(task_t *idle, int cpu)
 	set_tsk_need_resched(idle);
 	__restore_flags(flags);
 
+#ifdef CONFIG_PREEMPT
 	/* Set the preempt count _outside_ the spinlocks! */
 	idle->thread_info->preempt_count = (idle->lock_depth >= 0);
+#endif
 }
 
 extern void init_timervecs(void);