[PATCH] x86-64 fixes for 2.6.4rc1

From: Andi Kleen <ak@suse.de>

Fix a few issues on x86-64 for 2.6.4rc1.

The 32bit emulation used 4GB/3 for the mmap break. This actually gave
programs less sbrk space than with a standard 32bit kernel. Move the 32bit
mmap break to 0xc5000000.

Uses the correct gcc option to optimize for Prescott now (requires a very
new mainline gcc)

Also merge the i386 sched_clock().  I hope this will fix some interactivity
problems with the scheduler.

- Change initializer to new style (Arnd Bergmann)
- Remove 2 sibling limit in HT support (from i386)
- Always log RIP in MCE records even when not exact
- Move 32bit program task break up to 0xc5000000 by default
- Use -march=prescott for Prescott optimized kernel if possible
- Don't divide by zero with report_lost_ticks on and HPET off
- Merge preempt/smp/debug_pagealloc oops printing from i386
- Add pfn_to_nid()
- Merge more accurate sched_clock from i386
- Remove traces of debugging code in mce.c
- Update defconfig

arch/x86_64/Makefile

@@ -38,7 +38,7 @@ OBJCOPYFLAGS	:= -O binary -R .note -R .comment -S
 LDFLAGS_vmlinux := -e stext
 
 cflags-$(CONFIG_MK8) += $(call check_gcc,-march=k8,)
-cflags-$(CONFIG_MPSC) += $(call check_gcc,-march=pentium4,)
+cflags-$(CONFIG_MPSC) += $(call check_gcc,-march=prescott,)
 CFLAGS += $(cflags-y)
 
 CFLAGS += -mno-red-zone

arch/x86_64/defconfig

@@ -27,6 +27,7 @@ CONFIG_SYSVIPC=y
 # CONFIG_BSD_PROCESS_ACCT is not set
 CONFIG_SYSCTL=y
 CONFIG_LOG_BUF_SHIFT=18
+# CONFIG_HOTPLUG is not set
 CONFIG_IKCONFIG=y
 CONFIG_IKCONFIG_PROC=y
 # CONFIG_EMBEDDED is not set
@@ -106,6 +107,7 @@ CONFIG_ACPI_POWER=y
 CONFIG_ACPI_PCI=y
 CONFIG_ACPI_SYSTEM=y
 # CONFIG_ACPI_RELAXED_AML is not set
+# CONFIG_X86_PM_TIMER is not set
 
 #
 # CPU Frequency scaling
@@ -119,7 +121,6 @@ CONFIG_PCI=y
 CONFIG_PCI_DIRECT=y
 # CONFIG_PCI_LEGACY_PROC is not set
 # CONFIG_PCI_NAMES is not set
-# CONFIG_HOTPLUG is not set
 
 #
 # Executable file formats / Emulations
@@ -168,6 +169,7 @@ CONFIG_BLK_DEV_RAM=y
 CONFIG_BLK_DEV_RAM_SIZE=4096
 CONFIG_BLK_DEV_INITRD=y
 CONFIG_LBD=y
+# CONFIG_DCSSBLK is not set
 
 #
 # ATA/ATAPI/MFM/RLL support
@@ -306,7 +308,7 @@ CONFIG_FUSION_MAX_SGE=40
 # CONFIG_FUSION_CTL is not set
 
 #
-# IEEE 1394 (FireWire) support (EXPERIMENTAL)
+# IEEE 1394 (FireWire) support
 #
 # CONFIG_IEEE1394 is not set
 
@@ -489,7 +491,7 @@ CONFIG_TIGON3=y
 #
 # ISDN subsystem
 #
-# CONFIG_ISDN_BOOL is not set
+# CONFIG_ISDN is not set
 
 #
 # Telephony Support
@@ -562,7 +564,8 @@ CONFIG_SERIAL_8250_NR_UARTS=4
 CONFIG_SERIAL_CORE=y
 CONFIG_SERIAL_CORE_CONSOLE=y
 CONFIG_UNIX98_PTYS=y
-CONFIG_UNIX98_PTY_COUNT=256
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
 
 #
 # Mice
@@ -718,7 +721,6 @@ CONFIG_ISO9660_FS=y
 CONFIG_PROC_FS=y
 CONFIG_PROC_KCORE=y
 # CONFIG_DEVFS_FS is not set
-CONFIG_DEVPTS_FS=y
 # CONFIG_DEVPTS_FS_XATTR is not set
 CONFIG_TMPFS=y
 CONFIG_HUGETLBFS=y
@@ -731,6 +733,7 @@ CONFIG_RAMFS=y
 # CONFIG_ADFS_FS is not set
 # CONFIG_AFFS_FS is not set
 # CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
 # CONFIG_BEFS_FS is not set
 # CONFIG_BFS_FS is not set
 # CONFIG_EFS_FS is not set
@@ -761,7 +764,6 @@ CONFIG_SUNRPC=y
 # CONFIG_CIFS is not set
 # CONFIG_NCP_FS is not set
 # CONFIG_CODA_FS is not set
-# CONFIG_INTERMEZZO_FS is not set
 # CONFIG_AFS_FS is not set
 
 #

arch/x86_64/kernel/i8259.c

@@ -423,7 +423,7 @@ static struct sysdev_class timer_sysclass = {
 
 static struct sys_device device_timer = {
 	.id		= 0,
-	.cls		&timer_sysclass,
+	.cls		= &timer_sysclass,
 };
 
 static int __init init_timer_sysfs(void)

arch/x86_64/kernel/mce.c

@@ -73,7 +73,9 @@ static void print_mce(struct mce *m)
 	printk("CPU %d: Machine Check Exception: %16Lx Bank %d: %016Lx\n",
 	       m->cpu, m->mcgstatus, m->bank, m->status);
 	if (m->rip) {
-		printk("RIP %02x:<%016Lx> ", m->cs, m->rip);
+		printk("RIP%s %02x:<%016Lx> ",
+		       !(m->mcgstatus & MCG_STATUS_EIPV) ? " !INEXACT!" : "",
+		       m->cs, m->rip);
 		if (m->cs == __KERNEL_CS)
 			print_symbol("{%s}", m->rip);
 		printk("\n");
@@ -133,7 +135,7 @@ void do_machine_check(struct pt_regs * regs, long error_code)
 		return;
 	if (!(m.mcgstatus & MCG_STATUS_RIPV))
 		kill_it = 1;
-	if (regs && (m.mcgstatus & MCG_STATUS_EIPV)) {
+	if (regs) {
 		m.rip = regs->rip;
 		m.cs = regs->cs;
 	}

arch/x86_64/kernel/setup.c

@@ -610,8 +610,7 @@ static void __init detect_ht(void)
 		 * At this point we only support two siblings per
 		 * processor package.
 		 */
-#define NR_SIBLINGS	2
-		if (smp_num_siblings != NR_SIBLINGS) {
+		if (smp_num_siblings > NR_CPUS) {
 			printk(KERN_WARNING "CPU: Unsupported number of the siblings %d", smp_num_siblings);
 			smp_num_siblings = 1;
 			return;

arch/x86_64/kernel/time.c

@@ -26,6 +26,7 @@
 #include <linux/device.h>
 #include <linux/sysdev.h>
 #include <linux/bcd.h>
+#include <linux/kallsyms.h>
 #include <asm/pgtable.h>
 #include <asm/vsyscall.h>
 #include <asm/timex.h>
@@ -353,11 +354,11 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs)
 	}
 
 	if (lost) {
-		if (report_lost_ticks)
+		if (report_lost_ticks) {
 			printk(KERN_WARNING "time.c: Lost %ld timer "
-			       "tick(s)! (rip %016lx)\n",
-			       (offset - vxtime.last) / hpet_tick - 1,
-			       regs->rip);
+			       "tick(s)! ", lost);
+			print_symbol("rip %s)\n", regs->rip);
+		}
 		jiffies += lost;
 	}
 
@@ -399,8 +400,19 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs)
 	return IRQ_HANDLED;
 }
 
-/* RED-PEN: calculation is done in 32bits with multiply for performance
-   and could overflow, it may be better (but slower)to use an 64bit division. */
+static unsigned int cyc2ns_scale;
+#define CYC2NS_SCALE_FACTOR 10 /* 2^10, carefully chosen */
+
+static inline void set_cyc2ns_scale(unsigned long cpu_mhz)
+{
+	cyc2ns_scale = (1000 << CYC2NS_SCALE_FACTOR)/cpu_mhz;
+}
+
+static inline unsigned long long cycles_2_ns(unsigned long long cyc)
+{
+	return (cyc * cyc2ns_scale) >> CYC2NS_SCALE_FACTOR;
+}
+
 unsigned long long sched_clock(void)
 {
 	unsigned long a = 0;
@@ -420,7 +432,7 @@ unsigned long long sched_clock(void)
 	   purposes. */
 
 	rdtscll(a);
-	return (a * vxtime.tsc_quot) >> 32;
+	return cycles_2_ns(a);
 }
 
 unsigned long get_cmos_time(void)
@@ -527,6 +539,8 @@ static int time_cpufreq_notifier(struct notifier_block *nb, unsigned long val,
 		vxtime.tsc_quot = (1000L << 32) / cpu_khz;
 	}
 	
+	set_cyc2ns_scale(cpu_khz_ref / 1000);
+
 	return 0;
 }
  
@@ -725,6 +739,8 @@ void __init time_init(void)
 	rdtscll_sync(&vxtime.last_tsc);
 	setup_irq(0, &irq0);
 
+	set_cyc2ns_scale(cpu_khz / 1000);
+
 #ifdef CONFIG_CPU_FREQ
 	cpufreq_register_notifier(&time_cpufreq_notifier_block, 
 				  CPUFREQ_TRANSITION_NOTIFIER);

arch/x86_64/kernel/traps.c

@@ -351,9 +351,24 @@ void oops_end(void)
 
 void __die(const char * str, struct pt_regs * regs, long err)
 {
+	int nl = 0;
 	static int die_counter;
 	printk(KERN_EMERG "%s: %04lx [%u]\n", str, err & 0xffff,++die_counter);
 	notify_die(DIE_OOPS, (char *)str, regs, err, 255, SIGSEGV);
+#ifdef CONFIG_PREEMPT
+	printk("PREEMPT ");
+	nl = 1;
+#endif
+#ifdef CONFIG_SMP
+	printk("SMP ");
+	nl = 1;
+#endif
+#ifdef CONFIG_DEBUG_PAGEALLOC
+	printk("DEBUG_PAGEALLOC");
+	nl = 1;
+#endif
+	if (nl)
+		printk("\n");
 	show_registers(regs);
 	/* Executive summary in case the oops scrolled away */
 	printk("RIP "); 

include/asm-x86_64/mmzone.h

@@ -31,6 +31,8 @@ static inline __attribute__((pure)) int phys_to_nid(unsigned long addr)
 	return nid; 
 } 
 
+#define pfn_to_nid(pfn) phys_to_nid((unsigned long)(pfn) << PAGE_SHIFT)
+
 #define kvaddr_to_nid(kaddr)	phys_to_nid(__pa(kaddr))
 #define NODE_DATA(nid)		(node_data[nid])
 

include/asm-x86_64/processor.h

@@ -173,7 +173,7 @@ static inline void clear_in_cr4 (unsigned long mask)
  * space during mmap's.
  */
 #define IA32_PAGE_OFFSET ((current->personality & ADDR_LIMIT_3GB) ? 0xc0000000 : 0xFFFFe000)
-#define TASK_UNMAPPED_32 (PAGE_ALIGN(IA32_PAGE_OFFSET / 3))
+#define TASK_UNMAPPED_32 (PAGE_ALIGN(0xc5000000))
 #define TASK_UNMAPPED_64 PAGE_ALIGN(TASK_SIZE/3) 
 #define TASK_UNMAPPED_BASE	\
 	(test_thread_flag(TIF_IA32) ? TASK_UNMAPPED_32 : TASK_UNMAPPED_64)