diff options
Diffstat (limited to 'arch/x86_64/kernel')
38 files changed, 2051 insertions, 615 deletions
diff --git a/arch/x86_64/kernel/Makefile b/arch/x86_64/kernel/Makefile index 0a3318e08ab..48f9e2c19cd 100644 --- a/arch/x86_64/kernel/Makefile +++ b/arch/x86_64/kernel/Makefile @@ -20,6 +20,7 @@ obj-$(CONFIG_SMP) += smp.o smpboot.o trampoline.o obj-$(CONFIG_X86_LOCAL_APIC) += apic.o nmi.o obj-$(CONFIG_X86_IO_APIC) += io_apic.o mpparse.o \ genapic.o genapic_cluster.o genapic_flat.o +obj-$(CONFIG_KEXEC) += machine_kexec.o relocate_kernel.o crash.o obj-$(CONFIG_PM) += suspend.o obj-$(CONFIG_SOFTWARE_SUSPEND) += suspend_asm.o obj-$(CONFIG_CPU_FREQ) += cpufreq/ @@ -28,6 +29,7 @@ obj-$(CONFIG_GART_IOMMU) += pci-gart.o aperture.o obj-$(CONFIG_DUMMY_IOMMU) += pci-nommu.o pci-dma.o obj-$(CONFIG_SWIOTLB) += swiotlb.o obj-$(CONFIG_KPROBES) += kprobes.o +obj-$(CONFIG_X86_PM_TIMER) += pmtimer.o obj-$(CONFIG_MODULES) += module.o diff --git a/arch/x86_64/kernel/acpi/wakeup.S b/arch/x86_64/kernel/acpi/wakeup.S index a4c630034cd..185faa911db 100644 --- a/arch/x86_64/kernel/acpi/wakeup.S +++ b/arch/x86_64/kernel/acpi/wakeup.S @@ -67,7 +67,7 @@ wakeup_code: shll $4, %eax addl $(gdta - wakeup_code), %eax movl %eax, gdt_48a +2 - wakeup_code - lgdt %ds:gdt_48a - wakeup_code # load gdt with whatever is + lgdtl %ds:gdt_48a - wakeup_code # load gdt with whatever is # appropriate movl $1, %eax # protected mode (PE) bit diff --git a/arch/x86_64/kernel/aperture.c b/arch/x86_64/kernel/aperture.c index a491f72cc96..c9a6b812e92 100644 --- a/arch/x86_64/kernel/aperture.c +++ b/arch/x86_64/kernel/aperture.c @@ -33,20 +33,14 @@ int fallback_aper_force __initdata = 0; int fix_aperture __initdata = 1; -#define NB_ID_3 (PCI_VENDOR_ID_AMD | (0x1103<<16)) +/* This code runs before the PCI subsystem is initialized, so just + access the northbridge directly. */ -static struct resource aper_res = { - .name = "Aperture", - .flags = IORESOURCE_MEM, -}; +#define NB_ID_3 (PCI_VENDOR_ID_AMD | (0x1103<<16)) static u32 __init allocate_aperture(void) { -#ifdef CONFIG_DISCONTIGMEM pg_data_t *nd0 = NODE_DATA(0); -#else - pg_data_t *nd0 = &contig_page_data; -#endif u32 aper_size; void *p; @@ -55,24 +49,11 @@ static u32 __init allocate_aperture(void) aper_size = (32 * 1024 * 1024) << fallback_aper_order; /* - * Aperture has to be naturally aligned. This means an 2GB - * aperture won't have much chances to find a place in the - * lower 4GB of memory. Unfortunately we cannot move it up - * because that would make the IOMMU useless. + * Aperture has to be naturally aligned. This means an 2GB aperture won't + * have much chances to find a place in the lower 4GB of memory. + * Unfortunately we cannot move it up because that would make the + * IOMMU useless. */ - - /* First try to find some free unused space */ - if (!allocate_resource(&iomem_resource, &aper_res, - aper_size, - 0, 0xffffffff, - aper_size, - NULL, NULL)) { - printk(KERN_INFO "Putting aperture at %lx-%lx\n", - aper_res.start, aper_res.end); - return aper_res.start; - } - - /* No free space found. Go on to waste some memory... */ p = __alloc_bootmem_node(nd0, aper_size, aper_size, 0); if (!p || __pa(p)+aper_size > 0xffffffff) { printk("Cannot allocate aperture memory hole (%p,%uK)\n", @@ -81,7 +62,7 @@ static u32 __init allocate_aperture(void) free_bootmem_node(nd0, (unsigned long)p, aper_size); return 0; } - printk("Mapping aperture over %d KB of precious RAM @ %lx\n", + printk("Mapping aperture over %d KB of RAM @ %lx\n", aper_size >> 10, __pa(p)); return (u32)__pa(p); } @@ -102,16 +83,10 @@ static int __init aperture_valid(char *name, u64 aper_base, u32 aper_size) printk("Aperture from %s pointing to e820 RAM. Ignoring.\n",name); return 0; } - /* Don't check the resource here because the aperture is usually - in an e820 reserved area, and we allocated these earlier. */ return 1; } -/* - * Find a PCI capability. - * This code runs before the PCI subsystem is initialized, so just - * access the northbridge directly. - */ +/* Find a PCI capability */ static __u32 __init find_cap(int num, int slot, int func, int cap) { u8 pos; @@ -276,6 +251,8 @@ void __init iommu_hole_init(void) fallback_aper_force) { printk("Your BIOS doesn't leave a aperture memory hole\n"); printk("Please enable the IOMMU option in the BIOS setup\n"); + printk("This costs you %d MB of RAM\n", + 32 << fallback_aper_order); aper_order = fallback_aper_order; aper_alloc = allocate_aperture(); diff --git a/arch/x86_64/kernel/apic.c b/arch/x86_64/kernel/apic.c index 7e13545748e..375d369570c 100644 --- a/arch/x86_64/kernel/apic.c +++ b/arch/x86_64/kernel/apic.c @@ -33,6 +33,7 @@ #include <asm/mpspec.h> #include <asm/pgalloc.h> #include <asm/mach_apic.h> +#include <asm/nmi.h> int apic_verbosity; @@ -132,7 +133,7 @@ void __init connect_bsp_APIC(void) } } -void disconnect_bsp_APIC(void) +void disconnect_bsp_APIC(int virt_wire_setup) { if (pic_mode) { /* @@ -145,6 +146,42 @@ void disconnect_bsp_APIC(void) outb(0x70, 0x22); outb(0x00, 0x23); } + else { + /* Go back to Virtual Wire compatibility mode */ + unsigned long value; + + /* For the spurious interrupt use vector F, and enable it */ + value = apic_read(APIC_SPIV); + value &= ~APIC_VECTOR_MASK; + value |= APIC_SPIV_APIC_ENABLED; + value |= 0xf; + apic_write_around(APIC_SPIV, value); + + if (!virt_wire_setup) { + /* For LVT0 make it edge triggered, active high, external and enabled */ + value = apic_read(APIC_LVT0); + value &= ~(APIC_MODE_MASK | APIC_SEND_PENDING | + APIC_INPUT_POLARITY | APIC_LVT_REMOTE_IRR | + APIC_LVT_LEVEL_TRIGGER | APIC_LVT_MASKED ); + value |= APIC_LVT_REMOTE_IRR | APIC_SEND_PENDING; + value = SET_APIC_DELIVERY_MODE(value, APIC_MODE_EXTINT); + apic_write_around(APIC_LVT0, value); + } + else { + /* Disable LVT0 */ + apic_write_around(APIC_LVT0, APIC_LVT_MASKED); + } + + /* For LVT1 make it edge triggered, active high, nmi and enabled */ + value = apic_read(APIC_LVT1); + value &= ~( + APIC_MODE_MASK | APIC_SEND_PENDING | + APIC_INPUT_POLARITY | APIC_LVT_REMOTE_IRR | + APIC_LVT_LEVEL_TRIGGER | APIC_LVT_MASKED); + value |= APIC_LVT_REMOTE_IRR | APIC_SEND_PENDING; + value = SET_APIC_DELIVERY_MODE(value, APIC_MODE_NMI); + apic_write_around(APIC_LVT1, value); + } } void disable_local_APIC(void) @@ -284,7 +321,7 @@ void __init init_bsp_APIC(void) apic_write_around(APIC_LVT1, value); } -void __init setup_local_APIC (void) +void __cpuinit setup_local_APIC (void) { unsigned int value, ver, maxlvt; @@ -533,7 +570,7 @@ static struct sys_device device_lapic = { .cls = &lapic_sysclass, }; -static void __init apic_pm_activate(void) +static void __cpuinit apic_pm_activate(void) { apic_pm_state.active = 1; } @@ -773,14 +810,14 @@ void __init setup_boot_APIC_clock (void) local_irq_enable(); } -void __init setup_secondary_APIC_clock(void) +void __cpuinit setup_secondary_APIC_clock(void) { local_irq_disable(); /* FIXME: Do we need this? --RR */ setup_APIC_timer(calibration_result); local_irq_enable(); } -void __init disable_APIC_timer(void) +void __cpuinit disable_APIC_timer(void) { if (using_apic_timer) { unsigned long v; @@ -925,7 +962,7 @@ __init int oem_force_hpet_timer(void) unsigned id; DECLARE_BITMAP(clustermap, NUM_APIC_CLUSTERS); - bitmap_empty(clustermap, NUM_APIC_CLUSTERS); + bitmap_zero(clustermap, NUM_APIC_CLUSTERS); for (i = 0; i < NR_CPUS; i++) { id = bios_cpu_apicid[i]; @@ -1056,7 +1093,7 @@ int __init APIC_init_uniprocessor (void) nr_ioapics = 0; #endif setup_boot_APIC_clock(); - + check_nmi_watchdog(); return 0; } diff --git a/arch/x86_64/kernel/crash.c b/arch/x86_64/kernel/crash.c new file mode 100644 index 00000000000..d7fa4248501 --- /dev/null +++ b/arch/x86_64/kernel/crash.c @@ -0,0 +1,35 @@ +/* + * Architecture specific (x86_64) functions for kexec based crash dumps. + * + * Created by: Hariprasad Nellitheertha (hari@in.ibm.com) + * + * Copyright (C) IBM Corporation, 2004. All rights reserved. + * + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/smp.h> +#include <linux/irq.h> +#include <linux/reboot.h> +#include <linux/kexec.h> + +#include <asm/processor.h> +#include <asm/hardirq.h> +#include <asm/nmi.h> +#include <asm/hw_irq.h> + +note_buf_t crash_notes[NR_CPUS]; + +void machine_crash_shutdown(struct pt_regs *regs) +{ + /* This function is only called after the system + * has paniced or is otherwise in a critical state. + * The minimum amount of code to allow a kexec'd kernel + * to run successfully needs to happen here. + * + * In practice this means shooting down the other cpus in + * an SMP system. + */ +} diff --git a/arch/x86_64/kernel/e820.c b/arch/x86_64/kernel/e820.c index 7c154dfff64..6ded3a50dfe 100644 --- a/arch/x86_64/kernel/e820.c +++ b/arch/x86_64/kernel/e820.c @@ -16,6 +16,7 @@ #include <linux/bootmem.h> #include <linux/ioport.h> #include <linux/string.h> +#include <linux/kexec.h> #include <asm/page.h> #include <asm/e820.h> #include <asm/proto.h> @@ -191,8 +192,6 @@ void __init e820_reserve_resources(void) int i; for (i = 0; i < e820.nr_map; i++) { struct resource *res; - if (e820.map[i].addr + e820.map[i].size > 0x100000000ULL) - continue; res = alloc_bootmem_low(sizeof(struct resource)); switch (e820.map[i].type) { case E820_RAM: res->name = "System RAM"; break; @@ -212,6 +211,9 @@ void __init e820_reserve_resources(void) */ request_resource(res, &code_resource); request_resource(res, &data_resource); +#ifdef CONFIG_KEXEC + request_resource(res, &crashk_res); +#endif } } } diff --git a/arch/x86_64/kernel/early_printk.c b/arch/x86_64/kernel/early_printk.c index e3a19e8ebbf..9631c747c5e 100644 --- a/arch/x86_64/kernel/early_printk.c +++ b/arch/x86_64/kernel/early_printk.c @@ -2,20 +2,24 @@ #include <linux/kernel.h> #include <linux/init.h> #include <linux/string.h> +#include <linux/tty.h> #include <asm/io.h> #include <asm/processor.h> /* Simple VGA output */ #ifdef __i386__ +#include <asm/setup.h> #define VGABASE (__ISA_IO_base + 0xb8000) #else +#include <asm/bootsetup.h> #define VGABASE ((void __iomem *)0xffffffff800b8000UL) #endif -#define MAX_YPOS 25 -#define MAX_XPOS 80 +#define MAX_YPOS max_ypos +#define MAX_XPOS max_xpos +static int max_ypos = 25, max_xpos = 80; static int current_ypos = 1, current_xpos = 0; static void early_vga_write(struct console *con, const char *str, unsigned n) @@ -196,7 +200,10 @@ int __init setup_early_printk(char *opt) } else if (!strncmp(buf, "ttyS", 4)) { early_serial_init(buf); early_console = &early_serial_console; - } else if (!strncmp(buf, "vga", 3)) { + } else if (!strncmp(buf, "vga", 3) + && SCREEN_INFO.orig_video_isVGA == 1) { + max_xpos = SCREEN_INFO.orig_video_cols; + max_ypos = SCREEN_INFO.orig_video_lines; early_console = &early_vga_console; } early_console_initialized = 1; diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S index 1086b5fcac2..28817490fdc 100644 --- a/arch/x86_64/kernel/entry.S +++ b/arch/x86_64/kernel/entry.S @@ -220,13 +220,18 @@ sysret_careful: jmp sysret_check /* Handle a signal */ - /* edx: work flags (arg3) */ sysret_signal: sti + testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx + jz 1f + + /* Really a signal */ + /* edx: work flags (arg3) */ leaq do_notify_resume(%rip),%rax leaq -ARGOFFSET(%rsp),%rdi # &pt_regs -> arg1 xorl %esi,%esi # oldset -> arg2 call ptregscall_common +1: movl $_TIF_NEED_RESCHED,%edi jmp sysret_check /* Do syscall tracing */ @@ -484,6 +489,8 @@ retint_careful: jmp retint_check retint_signal: + testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx + jz retint_swapgs sti SAVE_REST movq $-1,ORIG_RAX(%rsp) @@ -492,8 +499,8 @@ retint_signal: call do_notify_resume RESTORE_REST cli + movl $_TIF_NEED_RESCHED,%edi GET_THREAD_INFO(%rcx) - movl $_TIF_WORK_MASK,%edi jmp retint_check #ifdef CONFIG_PREEMPT diff --git a/arch/x86_64/kernel/genapic_flat.c b/arch/x86_64/kernel/genapic_flat.c index b4cbbad0422..28284696508 100644 --- a/arch/x86_64/kernel/genapic_flat.c +++ b/arch/x86_64/kernel/genapic_flat.c @@ -7,6 +7,8 @@ * Hacked for x86-64 by James Cleverdon from i386 architecture code by * Martin Bligh, Andi Kleen, James Bottomley, John Stultz, and * James Cleverdon. + * Ashok Raj <ashok.raj@intel.com> + * Removed IPI broadcast shortcut to support CPU hotplug */ #include <linux/config.h> #include <linux/threads.h> @@ -18,6 +20,46 @@ #include <asm/smp.h> #include <asm/ipi.h> +/* + * The following permit choosing broadcast IPI shortcut v.s sending IPI only + * to online cpus via the send_IPI_mask varient. + * The mask version is my preferred option, since it eliminates a lot of + * other extra code that would need to be written to cleanup intrs sent + * to a CPU while offline. + * + * Sending broadcast introduces lots of trouble in CPU hotplug situations. + * These IPI's are delivered to cpu's irrespective of their offline status + * and could pickup stale intr data when these CPUS are turned online. + * + * Not using broadcast is a cleaner approach IMO, but Andi Kleen disagrees with + * the idea of not using broadcast IPI's anymore. Hence the run time check + * is introduced, on his request so we can choose an alternate mechanism. + * + * Initial wacky performance tests that collect cycle counts show + * no increase in using mask v.s broadcast version. In fact they seem + * identical in terms of cycle counts. + * + * if we need to use broadcast, we need to do the following. + * + * cli; + * hold call_lock; + * clear any pending IPI, just ack and clear all pending intr + * set cpu_online_map; + * release call_lock; + * sti; + * + * The complicated dummy irq processing shown above is not required if + * we didnt sent IPI's to wrong CPU's in the first place. + * + * - Ashok Raj <ashok.raj@intel.com> + */ +#ifdef CONFIG_HOTPLUG_CPU +#define DEFAULT_SEND_IPI (1) +#else +#define DEFAULT_SEND_IPI (0) +#endif + +static int no_broadcast=DEFAULT_SEND_IPI; static cpumask_t flat_target_cpus(void) { @@ -45,22 +87,6 @@ static void flat_init_apic_ldr(void) apic_write_around(APIC_LDR, val); } -static void flat_send_IPI_allbutself(int vector) -{ - /* - * if there are no other CPUs in the system then - * we get an APIC send error if we try to broadcast. - * thus we have to avoid sending IPIs in this case. - */ - if (num_online_cpus() > 1) - __send_IPI_shortcut(APIC_DEST_ALLBUT, vector, APIC_DEST_LOGICAL); -} - -static void flat_send_IPI_all(int vector) -{ - __send_IPI_shortcut(APIC_DEST_ALLINC, vector, APIC_DEST_LOGICAL); -} - static void flat_send_IPI_mask(cpumask_t cpumask, int vector) { unsigned long mask = cpus_addr(cpumask)[0]; @@ -93,6 +119,39 @@ static void flat_send_IPI_mask(cpumask_t cpumask, int vector) local_irq_restore(flags); } +static inline void __local_flat_send_IPI_allbutself(int vector) +{ + if (no_broadcast) { + cpumask_t mask = cpu_online_map; + int this_cpu = get_cpu(); + + cpu_clear(this_cpu, mask); + flat_send_IPI_mask(mask, vector); + put_cpu(); + } + else + __send_IPI_shortcut(APIC_DEST_ALLBUT, vector, APIC_DEST_LOGICAL); +} + +static inline void __local_flat_send_IPI_all(int vector) +{ + if (no_broadcast) + flat_send_IPI_mask(cpu_online_map, vector); + else + __send_IPI_shortcut(APIC_DEST_ALLINC, vector, APIC_DEST_LOGICAL); +} + +static void flat_send_IPI_allbutself(int vector) +{ + if (((num_online_cpus()) - 1) >= 1) + __local_flat_send_IPI_allbutself(vector); +} + +static void flat_send_IPI_all(int vector) +{ + __local_flat_send_IPI_all(vector); +} + static int flat_apic_id_registered(void) { return physid_isset(GET_APIC_ID(apic_read(APIC_ID)), phys_cpu_present_map); @@ -111,6 +170,16 @@ static unsigned int phys_pkg_id(int index_msb) return ((ebx >> 24) & 0xFF) >> index_msb; } +static __init int no_ipi_broadcast(char *str) +{ + get_option(&str, &no_broadcast); + printk ("Using %s mode\n", no_broadcast ? "No IPI Broadcast" : + "IPI Broadcast"); + return 1; +} + +__setup("no_ipi_broadcast", no_ipi_broadcast); + struct genapic apic_flat = { .name = "flat", .int_delivery_mode = dest_LowestPrio, @@ -125,3 +194,12 @@ struct genapic apic_flat = { .cpu_mask_to_apicid = flat_cpu_mask_to_apicid, .phys_pkg_id = phys_pkg_id, }; + +static int __init print_ipi_mode(void) +{ + printk ("Using IPI %s mode\n", no_broadcast ? "No-Shortcut" : + "Shortcut"); + return 0; +} + +late_initcall(print_ipi_mode); diff --git a/arch/x86_64/kernel/head.S b/arch/x86_64/kernel/head.S index 9bd2e7a4b81..8d765aa77a2 100644 --- a/arch/x86_64/kernel/head.S +++ b/arch/x86_64/kernel/head.S @@ -248,23 +248,23 @@ ENTRY(_stext) */ .org 0x1000 ENTRY(init_level4_pgt) - .quad 0x0000000000102007 /* -> level3_ident_pgt */ + .quad 0x0000000000002007 + __PHYSICAL_START /* -> level3_ident_pgt */ .fill 255,8,0 - .quad 0x000000000010a007 + .quad 0x000000000000a007 + __PHYSICAL_START .fill 254,8,0 /* (2^48-(2*1024*1024*1024))/(2^39) = 511 */ - .quad 0x0000000000103007 /* -> level3_kernel_pgt */ + .quad 0x0000000000003007 + __PHYSICAL_START /* -> level3_kernel_pgt */ .org 0x2000 ENTRY(level3_ident_pgt) - .quad 0x0000000000104007 + .quad 0x0000000000004007 + __PHYSICAL_START .fill 511,8,0 .org 0x3000 ENTRY(level3_kernel_pgt) .fill 510,8,0 /* (2^48-(2*1024*1024*1024)-((2^39)*511))/(2^30) = 510 */ - .quad 0x0000000000105007 /* -> level2_kernel_pgt */ + .quad 0x0000000000005007 + __PHYSICAL_START /* -> level2_kernel_pgt */ .fill 1,8,0 .org 0x4000 @@ -337,17 +337,17 @@ ENTRY(empty_bad_pmd_table) .org 0xa000 ENTRY(level3_physmem_pgt) - .quad 0x0000000000105007 /* -> level2_kernel_pgt (so that __va works even before pagetable_init) */ + .quad 0x0000000000005007 + __PHYSICAL_START /* -> level2_kernel_pgt (so that __va works even before pagetable_init) */ .org 0xb000 #ifdef CONFIG_ACPI_SLEEP ENTRY(wakeup_level4_pgt) - .quad 0x0000000000102007 /* -> level3_ident_pgt */ + .quad 0x0000000000002007 + __PHYSICAL_START /* -> level3_ident_pgt */ .fill 255,8,0 - .quad 0x000000000010a007 + .quad 0x000000000000a007 + __PHYSICAL_START .fill 254,8,0 /* (2^48-(2*1024*1024*1024))/(2^39) = 511 */ - .quad 0x0000000000103007 /* -> level3_kernel_pgt */ + .quad 0x0000000000003007 + __PHYSICAL_START /* -> level3_kernel_pgt */ #endif .data diff --git a/arch/x86_64/kernel/head64.c b/arch/x86_64/kernel/head64.c index 0f8c78dcd38..cf6ab147a2a 100644 --- a/arch/x86_64/kernel/head64.c +++ b/arch/x86_64/kernel/head64.c @@ -94,7 +94,7 @@ void __init x86_64_start_kernel(char * real_mode_data) s = strstr(saved_command_line, "earlyprintk="); if (s != NULL) setup_early_printk(s); -#ifdef CONFIG_DISCONTIGMEM +#ifdef CONFIG_NUMA s = strstr(saved_command_line, "numa="); if (s != NULL) numa_setup(s+5); diff --git a/arch/x86_64/kernel/i387.c b/arch/x86_64/kernel/i387.c index ba139cac57c..d9b22b633e3 100644 --- a/arch/x86_64/kernel/i387.c +++ b/arch/x86_64/kernel/i387.c @@ -42,7 +42,7 @@ void mxcsr_feature_mask_init(void) * Called at bootup to set up the initial FPU state that is later cloned * into all processes. */ -void __init fpu_init(void) +void __cpuinit fpu_init(void) { unsigned long oldcr0 = read_cr0(); extern void __bad_fxsave_alignment(void); diff --git a/arch/x86_64/kernel/i8259.c b/arch/x86_64/kernel/i8259.c index 7873d9ba881..a8916909512 100644 --- a/arch/x86_64/kernel/i8259.c +++ b/arch/x86_64/kernel/i8259.c @@ -157,14 +157,13 @@ static unsigned int startup_8259A_irq(unsigned int irq) } static struct hw_interrupt_type i8259A_irq_type = { - "XT-PIC", - startup_8259A_irq, - shutdown_8259A_irq, - enable_8259A_irq, - disable_8259A_irq, - mask_and_ack_8259A, - end_8259A_irq, - NULL + .typename = "XT-PIC", + .startup = startup_8259A_irq, + .shutdown = shutdown_8259A_irq, + .enable = enable_8259A_irq, + .disable = disable_8259A_irq, + .ack = mask_and_ack_8259A, + .end = end_8259A_irq, }; /* @@ -415,10 +414,22 @@ static int i8259A_suspend(struct sys_device *dev, pm_message_t state) return 0; } +static int i8259A_shutdown(struct sys_device *dev) +{ + /* Put the i8259A into a quiescent state that + * the kernel initialization code can get it + * out of. + */ + outb(0xff, 0x21); /* mask all of 8259A-1 */ + outb(0xff, 0xA1); /* mask all of 8259A-1 */ + return 0; +} + static struct sysdev_class i8259_sysdev_class = { set_kset_name("i8259"), .suspend = i8259A_suspend, .resume = i8259A_resume, + .shutdown = i8259A_shutdown, }; static struct sys_device device_i8259A = { diff --git a/arch/x86_64/kernel/io_apic.c b/arch/x86_64/kernel/io_apic.c index 60be58617eb..157190d986b 100644 --- a/arch/x86_64/kernel/io_apic.c +++ b/arch/x86_64/kernel/io_apic.c @@ -37,11 +37,14 @@ #include <asm/desc.h> #include <asm/proto.h> #include <asm/mach_apic.h> +#include <asm/acpi.h> #define __apicdebuginit __init int sis_apic_bug; /* not actually supported, dummy for compile */ +static int no_timer_check; + static DEFINE_SPINLOCK(ioapic_lock); /* @@ -327,7 +330,7 @@ static int find_irq_entry(int apic, int pin, int type) /* * Find the pin to which IRQ[irq] (ISA) is connected */ -static int __init find_isa_irq_pin(int irq, int type) +static int find_isa_irq_pin(int irq, int type) { int i; @@ -1129,12 +1132,44 @@ static void __init enable_IO_APIC(void) */ void disable_IO_APIC(void) { + int pin; /* * Clear the IO-APIC before rebooting: */ clear_IO_APIC(); - disconnect_bsp_APIC(); + /* + * If the i82559 is routed through an IOAPIC + * Put that IOAPIC in virtual wire mode + * so legacy interrups can be delivered. + */ + pin = find_isa_irq_pin(0, mp_ExtINT); + if (pin != -1) { + struct IO_APIC_route_entry entry; + unsigned long flags; + + memset(&entry, 0, sizeof(entry)); + entry.mask = 0; /* Enabled */ + entry.trigger = 0; /* Edge */ + entry.irr = 0; + entry.polarity = 0; /* High */ + entry.delivery_status = 0; + entry.dest_mode = 0; /* Physical */ + entry.delivery_mode = 7; /* ExtInt */ + entry.vector = 0; + entry.dest.physical.physical_dest = 0; + + + /* + * Add it to the IO-APIC irq-routing table: + */ + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(0, 0x11+2*pin, *(((int *)&entry)+1)); + io_apic_write(0, 0x10+2*pin, *(((int *)&entry)+0)); + spin_unlock_irqrestore(&ioapic_lock, flags); + } + + disconnect_bsp_APIC(pin != -1); } /* @@ -1601,7 +1636,7 @@ static inline void check_timer(void) * Ok, does IRQ0 through the IOAPIC work? */ unmask_IO_APIC_irq(0); - if (timer_irq_works()) { + if (!no_timer_check && timer_irq_works()) { nmi_watchdog_default(); if (nmi_watchdog == NMI_IO_APIC) { disable_8259A_irq(0); @@ -1671,6 +1706,13 @@ static inline void check_timer(void) panic("IO-APIC + timer doesn't work! Try using the 'noapic' kernel parameter\n"); } +static int __init notimercheck(char *s) +{ + no_timer_check = 1; + return 1; +} +__setup("no_timer_check", notimercheck); + /* * * IRQ's that are handled by the PIC in the MPS IOAPIC case. @@ -1804,76 +1846,6 @@ device_initcall(ioapic_init_sysfs); #define IO_APIC_MAX_ID 0xFE -int __init io_apic_get_unique_id (int ioapic, int apic_id) -{ - union IO_APIC_reg_00 reg_00; - static physid_mask_t apic_id_map; - unsigned long flags; - int i = 0; - - /* - * The P4 platform supports up to 256 APIC IDs on two separate APIC - * buses (one for LAPICs, one for IOAPICs), where predecessors only - * supports up to 16 on one shared APIC bus. - * - * TBD: Expand LAPIC/IOAPIC support on P4-class systems to take full - * advantage of new APIC bus architecture. - */ - - if (physids_empty(apic_id_map)) - apic_id_map = phys_cpu_present_map; - - spin_lock_irqsave(&ioapic_lock, flags); - reg_00.raw = io_apic_read(ioapic, 0); - spin_unlock_irqrestore(&ioapic_lock, flags); - - if (apic_id >= IO_APIC_MAX_ID) { - apic_printk(APIC_QUIET, KERN_WARNING "IOAPIC[%d]: Invalid apic_id %d, trying " - "%d\n", ioapic, apic_id, reg_00.bits.ID); - apic_id = reg_00.bits.ID; - } - - /* - * Every APIC in a system must have a unique ID or we get lots of nice - * 'stuck on smp_invalidate_needed IPI wait' messages. - */ - if (physid_isset(apic_id, apic_id_map)) { - - for (i = 0; i < IO_APIC_MAX_ID; i++) { - if (!physid_isset(i, apic_id_map)) - break; - } - - if (i == IO_APIC_MAX_ID) - panic("Max apic_id exceeded!\n"); - - apic_printk(APIC_VERBOSE, KERN_WARNING "IOAPIC[%d]: apic_id %d already used, " - "trying %d\n", ioapic, apic_id, i); - - apic_id = i; - } - - physid_set(apic_id, apic_id_map); - - if (reg_00.bits.ID != apic_id) { - reg_00.bits.ID = apic_id; - - spin_lock_irqsave(&ioapic_lock, flags); - io_apic_write(ioapic, 0, reg_00.raw); - reg_00.raw = io_apic_read(ioapic, 0); - spin_unlock_irqrestore(&ioapic_lock, flags); - - /* Sanity check */ - if (reg_00.bits.ID != apic_id) - panic("IOAPIC[%d]: Unable change apic_id!\n", ioapic); - } - - apic_printk(APIC_VERBOSE,KERN_INFO "IOAPIC[%d]: Assigned apic_id %d\n", ioapic, apic_id); - - return apic_id; -} - - int __init io_apic_get_version (int ioapic) { union IO_APIC_reg_01 reg_01; diff --git a/arch/x86_64/kernel/irq.c b/arch/x86_64/kernel/irq.c index 62b112e4deb..cc3fb85f514 100644 --- a/arch/x86_64/kernel/irq.c +++ b/arch/x86_64/kernel/irq.c @@ -14,6 +14,7 @@ #include <linux/interrupt.h> #include <linux/seq_file.h> #include <linux/module.h> +#include <linux/delay.h> #include <asm/uaccess.h> #include <asm/io_apic.h> @@ -106,3 +107,31 @@ asmlinkage unsigned int do_IRQ(struct pt_regs *regs) return 1; } +#ifdef CONFIG_HOTPLUG_CPU +void fixup_irqs(cpumask_t map) +{ + unsigned int irq; + static int warned; + + for (irq = 0; irq < NR_IRQS; irq++) { + cpumask_t mask; + if (irq == 2) + continue; + + cpus_and(mask, irq_affinity[irq], map); + if (any_online_cpu(mask) == NR_CPUS) { + printk("Breaking affinity for irq %i\n", irq); + mask = map; + } + if (irq_desc[irq].handler->set_affinity) + irq_desc[irq].handler->set_affinity(irq, mask); + else if (irq_desc[irq].action && !(warned++)) + printk("Cannot set affinity for irq %i\n", irq); + } + + /* That doesn't seem sufficient. Give it 1ms. */ + local_irq_enable(); + mdelay(1); + local_irq_disable(); +} +#endif diff --git a/arch/x86_64/kernel/kprobes.c b/arch/x86_64/kernel/kprobes.c index f77f8a0ff18..4e680f87a75 100644 --- a/arch/x86_64/kernel/kprobes.c +++ b/arch/x86_64/kernel/kprobes.c @@ -27,6 +27,8 @@ * <prasanna@in.ibm.com> adapted for x86_64 * 2005-Mar Roland McGrath <roland@redhat.com> * Fixed to handle %rip-relative addressing mode correctly. + * 2005-May Rusty Lynch <rusty.lynch@intel.com> + * Added function return probes functionality */ #include <linux/config.h> @@ -37,18 +39,16 @@ #include <linux/slab.h> #include <linux/preempt.h> #include <linux/moduleloader.h> - +#include <asm/cacheflush.h> #include <asm/pgtable.h> #include <asm/kdebug.h> static DECLARE_MUTEX(kprobe_mutex); -/* kprobe_status settings */ -#define KPROBE_HIT_ACTIVE 0x00000001 -#define KPROBE_HIT_SS 0x00000002 - static struct kprobe *current_kprobe; static unsigned long kprobe_status, kprobe_old_rflags, kprobe_saved_rflags; +static struct kprobe *kprobe_prev; +static unsigned long kprobe_status_prev, kprobe_old_rflags_prev, kprobe_saved_rflags_prev; static struct pt_regs jprobe_saved_regs; static long *jprobe_saved_rsp; static kprobe_opcode_t *get_insn_slot(void); @@ -214,6 +214,21 @@ void arch_copy_kprobe(struct kprobe *p) BUG_ON((s64) (s32) disp != disp); /* Sanity check. */ *ripdisp = disp; } + p->opcode = *p->addr; +} + +void arch_arm_kprobe(struct kprobe *p) +{ + *p->addr = BREAKPOINT_INSTRUCTION; + flush_icache_range((unsigned long) p->addr, + (unsigned long) p->addr + sizeof(kprobe_opcode_t)); +} + +void arch_disarm_kprobe(struct kprobe *p) +{ + *p->addr = p->opcode; + flush_icache_range((unsigned long) p->addr, + (unsigned long) p->addr + sizeof(kprobe_opcode_t)); } void arch_remove_kprobe(struct kprobe *p) @@ -223,10 +238,29 @@ void arch_remove_kprobe(struct kprobe *p) down(&kprobe_mutex); } -static inline void disarm_kprobe(struct kprobe *p, struct pt_regs *regs) +static inline void save_previous_kprobe(void) { - *p->addr = p->opcode; - regs->rip = (unsigned long)p->addr; + kprobe_prev = current_kprobe; + kprobe_status_prev = kprobe_status; + kprobe_old_rflags_prev = kprobe_old_rflags; + kprobe_saved_rflags_prev = kprobe_saved_rflags; +} + +static inline void restore_previous_kprobe(void) +{ + current_kprobe = kprobe_prev; + kprobe_status = kprobe_status_prev; + kprobe_old_rflags = kprobe_old_rflags_prev; + kprobe_saved_rflags = kprobe_saved_rflags_prev; +} + +static inline void set_current_kprobe(struct kprobe *p, struct pt_regs *regs) +{ + current_kprobe = p; + kprobe_saved_rflags = kprobe_old_rflags + = (regs->eflags & (TF_MASK | IF_MASK)); + if (is_IF_modifier(p->ainsn.insn)) + kprobe_saved_rflags &= ~IF_MASK; } static void prepare_singlestep(struct kprobe *p, struct pt_regs *regs) @@ -240,6 +274,50 @@ static void prepare_singlestep(struct kprobe *p, struct pt_regs *regs) regs->rip = (unsigned long)p->ainsn.insn; } +struct task_struct *arch_get_kprobe_task(void *ptr) +{ + return ((struct thread_info *) (((unsigned long) ptr) & + (~(THREAD_SIZE -1))))->task; +} + +void arch_prepare_kretprobe(struct kretprobe *rp, struct pt_regs *regs) +{ + unsigned long *sara = (unsigned long *)regs->rsp; + struct kretprobe_instance *ri; + static void *orig_ret_addr; + + /* + * Save the return address when the return probe hits + * the first time, and use it to populate the (krprobe + * instance)->ret_addr for subsequent return probes at + * the same addrress since stack address would have + * the kretprobe_trampoline by then. + */ + if (((void*) *sara) != kretprobe_trampoline) + orig_ret_addr = (void*) *sara; + + if ((ri = get_free_rp_inst(rp)) != NULL) { + ri->rp = rp; + ri->stack_addr = sara; + ri->ret_addr = orig_ret_addr; + add_rp_inst(ri); + /* Replace the return addr with trampoline addr */ + *sara = (unsigned long) &kretprobe_trampoline; + } else { + rp->nmissed++; + } +} + +void arch_kprobe_flush_task(struct task_struct *tk) +{ + struct kretprobe_instance *ri; + while ((ri = get_rp_inst_tsk(tk)) != NULL) { + *((unsigned long *)(ri->stack_addr)) = + (unsigned long) ri->ret_addr; + recycle_rp_inst(ri); + } +} + /* * Interrupts are disabled on entry as trap3 is an interrupt gate and they * remain disabled thorough out this function. @@ -264,9 +342,30 @@ int kprobe_handler(struct pt_regs *regs) regs->eflags |= kprobe_saved_rflags; unlock_kprobes(); goto no_kprobe; + } else if (kprobe_status == KPROBE_HIT_SSDONE) { + /* TODO: Provide re-entrancy from + * post_kprobes_handler() and avoid exception + * stack corruption while single-stepping on + * the instruction of the new probe. + */ + arch_disarm_kprobe(p); + regs->rip = (unsigned long)p->addr; + ret = 1; + } else { + /* We have reentered the kprobe_handler(), since + * another probe was hit while within the + * handler. We here save the original kprobe + * variables and just single step on instruction + * of the new probe without calling any user + * handlers. + */ + save_previous_kprobe(); + set_current_kprobe(p, regs); + p->nmissed++; + prepare_singlestep(p, regs); + kprobe_status = KPROBE_REENTER; + return 1; } - disarm_kprobe(p, regs); - ret = 1; } else { p = current_kprobe; if (p->break_handler && p->break_handler(p, regs)) { @@ -296,11 +395,7 @@ int kprobe_handler(struct pt_regs *regs) } kprobe_status = KPROBE_HIT_ACTIVE; - current_kprobe = p; - kprobe_saved_rflags = kprobe_old_rflags - = (regs->eflags & (TF_MASK | IF_MASK)); - if (is_IF_modifier(p->ainsn.insn)) - kprobe_saved_rflags &= ~IF_MASK; + set_current_kprobe(p, regs); if (p->pre_handler && p->pre_handler(p, regs)) /* handler has already set things up, so skip ss setup */ @@ -317,6 +412,55 @@ no_kprobe: } /* + * For function-return probes, init_kprobes() establishes a probepoint + * here. When a retprobed function returns, this probe is hit and + * trampoline_probe_handler() runs, calling the kretprobe's handler. + */ + void kretprobe_trampoline_holder(void) + { + asm volatile ( ".global kretprobe_trampoline\n" + "kretprobe_trampoline: \n" + "nop\n"); + } + +/* + * Called when we hit the probe point at kretprobe_trampoline + */ +int trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) +{ + struct task_struct *tsk; + struct kretprobe_instance *ri; + struct hlist_head *head; + struct hlist_node *node; + unsigned long *sara = (unsigned long *)regs->rsp - 1; + + tsk = arch_get_kprobe_task(sara); + head = kretprobe_inst_table_head(tsk); + + hlist_for_each_entry(ri, node, head, hlist) { + if (ri->stack_addr == sara && ri->rp) { + if (ri->rp->handler) + ri->rp->handler(ri, regs); + } + } + return 0; +} + +void trampoline_post_handler(struct kprobe *p, struct pt_regs *regs, + unsigned long flags) +{ + struct kretprobe_instance *ri; + /* RA already popped */ + unsigned long *sara = ((unsigned long *)regs->rsp) - 1; + + while ((ri = get_rp_inst(sara))) { + regs->rip = (unsigned long)ri->ret_addr; + recycle_rp_inst(ri); + } + regs->eflags &= ~TF_MASK; +} + +/* * Called after single-stepping. p->addr is the address of the * instruction whose first byte has been replaced by the "int 3" * instruction. To avoid the SMP problems that can occur when we @@ -401,13 +545,23 @@ int post_kprobe_handler(struct pt_regs *regs) if (!kprobe_running()) return 0; - if (current_kprobe->post_handler) + if ((kprobe_status != KPROBE_REENTER) && current_kprobe->post_handler) { + kprobe_status = KPROBE_HIT_SSDONE; current_kprobe->post_handler(current_kprobe, regs, 0); + } - resume_execution(current_kprobe, regs); + if (current_kprobe->post_handler != trampoline_post_handler) + resume_execution(current_kprobe, regs); regs->eflags |= kprobe_saved_rflags; - unlock_kprobes(); + /* Restore the original saved kprobes variables and continue. */ + if (kprobe_status == KPROBE_REENTER) { + restore_previous_kprobe(); + goto out; + } else { + unlock_kprobes(); + } +out: preempt_enable_no_resched(); /* diff --git a/arch/x86_64/kernel/machine_kexec.c b/arch/x86_64/kernel/machine_kexec.c new file mode 100644 index 00000000000..60d1eff4156 --- /dev/null +++ b/arch/x86_64/kernel/machine_kexec.c @@ -0,0 +1,250 @@ +/* + * machine_kexec.c - handle transition of Linux booting another kernel + * Copyright (C) 2002-2005 Eric Biederman <ebiederm@xmission.com> + * + * This source code is licensed under the GNU General Public License, + * Version 2. See the file COPYING for more details. + */ + +#include <linux/mm.h> +#include <linux/kexec.h> +#include <linux/delay.h> +#include <linux/string.h> +#include <linux/reboot.h> +#include <asm/pda.h> +#include <asm/pgtable.h> +#include <asm/pgalloc.h> +#include <asm/tlbflush.h> +#include <asm/mmu_context.h> +#include <asm/io.h> +#include <asm/apic.h> +#include <asm/cpufeature.h> +#include <asm/hw_irq.h> + +#define LEVEL0_SIZE (1UL << 12UL) +#define LEVEL1_SIZE (1UL << 21UL) +#define LEVEL2_SIZE (1UL << 30UL) +#define LEVEL3_SIZE (1UL << 39UL) +#define LEVEL4_SIZE (1UL << 48UL) + +#define L0_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY) +#define L1_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_PSE) +#define L2_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY) +#define L3_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY) + +static void init_level2_page(u64 *level2p, unsigned long addr) +{ + unsigned long end_addr; + + addr &= PAGE_MASK; + end_addr = addr + LEVEL2_SIZE; + while (addr < end_addr) { + *(level2p++) = addr | L1_ATTR; + addr += LEVEL1_SIZE; + } +} + +static int init_level3_page(struct kimage *image, u64 *level3p, + unsigned long addr, unsigned long last_addr) +{ + unsigned long end_addr; + int result; + + result = 0; + addr &= PAGE_MASK; + end_addr = addr + LEVEL3_SIZE; + while ((addr < last_addr) && (addr < end_addr)) { + struct page *page; + u64 *level2p; + + page = kimage_alloc_control_pages(image, 0); + if (!page) { + result = -ENOMEM; + goto out; + } + level2p = (u64 *)page_address(page); + init_level2_page(level2p, addr); + *(level3p++) = __pa(level2p) | L2_ATTR; + addr += LEVEL2_SIZE; + } + /* clear the unused entries */ + while (addr < end_addr) { + *(level3p++) = 0; + addr += LEVEL2_SIZE; + } +out: + return result; +} + + +static int init_level4_page(struct kimage *image, u64 *level4p, + unsigned long addr, unsigned long last_addr) +{ + unsigned long end_addr; + int result; + + result = 0; + addr &= PAGE_MASK; + end_addr = addr + LEVEL4_SIZE; + while ((addr < last_addr) && (addr < end_addr)) { + struct page *page; + u64 *level3p; + + page = kimage_alloc_control_pages(image, 0); + if (!page) { + result = -ENOMEM; + goto out; + } + level3p = (u64 *)page_address(page); + result = init_level3_page(image, level3p, addr, last_addr); + if (result) { + goto out; + } + *(level4p++) = __pa(level3p) | L3_ATTR; + addr += LEVEL3_SIZE; + } + /* clear the unused entries */ + while (addr < end_addr) { + *(level4p++) = 0; + addr += LEVEL3_SIZE; + } +out: + return result; +} + + +static int init_pgtable(struct kimage *image, unsigned long start_pgtable) +{ + u64 *level4p; + level4p = (u64 *)__va(start_pgtable); + return init_level4_page(image, level4p, 0, end_pfn << PAGE_SHIFT); +} + +static void set_idt(void *newidt, u16 limit) +{ + unsigned char curidt[10]; + + /* x86-64 supports unaliged loads & stores */ + (*(u16 *)(curidt)) = limit; + (*(u64 *)(curidt +2)) = (unsigned long)(newidt); + + __asm__ __volatile__ ( + "lidt %0\n" + : "=m" (curidt) + ); +}; + + +static void set_gdt(void *newgdt, u16 limit) +{ + unsigned char curgdt[10]; + + /* x86-64 supports unaligned loads & stores */ + (*(u16 *)(curgdt)) = limit; + (*(u64 *)(curgdt +2)) = (unsigned long)(newgdt); + + __asm__ __volatile__ ( + "lgdt %0\n" + : "=m" (curgdt) + ); +}; + +static void load_segments(void) +{ + __asm__ __volatile__ ( + "\tmovl $"STR(__KERNEL_DS)",%eax\n" + "\tmovl %eax,%ds\n" + "\tmovl %eax,%es\n" + "\tmovl %eax,%ss\n" + "\tmovl %eax,%fs\n" + "\tmovl %eax,%gs\n" + ); +#undef STR +#undef __STR +} + +typedef NORET_TYPE void (*relocate_new_kernel_t)(unsigned long indirection_page, + unsigned long control_code_buffer, + unsigned long start_address, + unsigned long pgtable) ATTRIB_NORET; + +const extern unsigned char relocate_new_kernel[]; +const extern unsigned long relocate_new_kernel_size; + +int machine_kexec_prepare(struct kimage *image) +{ + unsigned long start_pgtable, control_code_buffer; + int result; + + /* Calculate the offsets */ + start_pgtable = page_to_pfn(image->control_code_page) << PAGE_SHIFT; + control_code_buffer = start_pgtable + 4096UL; + + /* Setup the identity mapped 64bit page table */ + result = init_pgtable(image, start_pgtable); + if (result) + return result; + + /* Place the code in the reboot code buffer */ + memcpy(__va(control_code_buffer), relocate_new_kernel, + relocate_new_kernel_size); + + return 0; +} + +void machine_kexec_cleanup(struct kimage *image) +{ + return; +} + +/* + * Do not allocate memory (or fail in any way) in machine_kexec(). + * We are past the point of no return, committed to rebooting now. + */ +NORET_TYPE void machine_kexec(struct kimage *image) +{ + unsigned long page_list; + unsigned long control_code_buffer; + unsigned long start_pgtable; + relocate_new_kernel_t rnk; + + /* Interrupts aren't acceptable while we reboot */ + local_irq_disable(); + + /* Calculate the offsets */ + page_list = image->head; + start_pgtable = page_to_pfn(image->control_code_page) << PAGE_SHIFT; + control_code_buffer = start_pgtable + 4096UL; + + /* Set the low half of the page table to my identity mapped + * page table for kexec. Leave the high half pointing at the + * kernel pages. Don't bother to flush the global pages + * as that will happen when I fully switch to my identity mapped + * page table anyway. + */ + memcpy(__va(read_cr3()), __va(start_pgtable), PAGE_SIZE/2); + __flush_tlb(); + + + /* The segment registers are funny things, they are + * automatically loaded from a table, in memory wherever you + * set them to a specific selector, but this table is never + * accessed again unless you set the segment to a different selector. + * + * The more common model are caches where the behide + * the scenes work is done, but is also dropped at arbitrary + * times. + * + * I take advantage of this here by force loading the + * segments, before I zap the gdt with an invalid value. + */ + load_segments(); + /* The gdt & idt are now invalid. + * If you want to load them you must set up your own idt & gdt. + */ + set_gdt(phys_to_virt(0),0); + set_idt(phys_to_virt(0),0); + /* now call it */ + rnk = (relocate_new_kernel_t) control_code_buffer; + (*rnk)(page_list, control_code_buffer, image->start, start_pgtable); +} diff --git a/arch/x86_64/kernel/mce.c b/arch/x86_64/kernel/mce.c index 3a89d735a4f..21e70625a49 100644 --- a/arch/x86_64/kernel/mce.c +++ b/arch/x86_64/kernel/mce.c @@ -327,7 +327,7 @@ static void mce_init(void *dummy) } /* Add per CPU specific workarounds here */ -static void __init mce_cpu_quirks(struct cpuinfo_x86 *c) +static void __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c) { /* This should be disabled by the BIOS, but isn't always */ if (c->x86_vendor == X86_VENDOR_AMD && c->x86 == 15) { @@ -337,7 +337,7 @@ static void __init mce_cpu_quirks(struct cpuinfo_x86 *c) } } -static void __init mce_cpu_features(struct cpuinfo_x86 *c) +static void __cpuinit mce_cpu_features(struct cpuinfo_x86 *c) { switch (c->x86_vendor) { case X86_VENDOR_INTEL: @@ -352,7 +352,7 @@ static void __init mce_cpu_features(struct cpuinfo_x86 *c) * Called for each booted CPU to set up machine checks. * Must be called with preempt off. */ -void __init mcheck_init(struct cpuinfo_x86 *c) +void __cpuinit mcheck_init(struct cpuinfo_x86 *c) { static cpumask_t mce_cpus __initdata = CPU_MASK_NONE; @@ -411,7 +411,7 @@ static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize, loff memset(mcelog.entry, 0, next * sizeof(struct mce)); mcelog.next = 0; - synchronize_kernel(); + synchronize_sched(); /* Collect entries that were still getting written before the synchronize. */ @@ -542,7 +542,7 @@ ACCESSOR(bank4ctl,bank[4],mce_restart()) ACCESSOR(tolerant,tolerant,) ACCESSOR(check_interval,check_interval,mce_restart()) -static __init int mce_init_device(void) +static __cpuinit int mce_init_device(void) { int err; if (!mce_available(&boot_cpu_data)) diff --git a/arch/x86_64/kernel/mce_intel.c b/arch/x86_64/kernel/mce_intel.c index 4db9a640069..0be0a795981 100644 --- a/arch/x86_64/kernel/mce_intel.c +++ b/arch/x86_64/kernel/mce_intel.c @@ -42,7 +42,7 @@ done: irq_exit(); } -static void __init intel_init_thermal(struct cpuinfo_x86 *c) +static void __cpuinit intel_init_thermal(struct cpuinfo_x86 *c) { u32 l, h; int tm2 = 0; @@ -93,7 +93,7 @@ static void __init intel_init_thermal(struct cpuinfo_x86 *c) return; } -void __init mce_intel_feature_init(struct cpuinfo_x86 *c) +void __cpuinit mce_intel_feature_init(struct cpuinfo_x86 *c) { intel_init_thermal(c); } diff --git a/arch/x86_64/kernel/mpparse.c b/arch/x86_64/kernel/mpparse.c index 7ec031c6ca1..9c5aa2a790c 100644 --- a/arch/x86_64/kernel/mpparse.c +++ b/arch/x86_64/kernel/mpparse.c @@ -23,6 +23,7 @@ #include <linux/kernel_stat.h> #include <linux/mc146818rtc.h> #include <linux/acpi.h> +#include <linux/module.h> #include <asm/smp.h> #include <asm/mtrr.h> @@ -30,6 +31,7 @@ #include <asm/pgalloc.h> #include <asm/io_apic.h> #include <asm/proto.h> +#include <asm/acpi.h> /* Have we found an MP table */ int smp_found_config; @@ -44,7 +46,8 @@ int acpi_found_madt; int apic_version [MAX_APICS]; unsigned char mp_bus_id_to_type [MAX_MP_BUSSES] = { [0 ... MAX_MP_BUSSES-1] = -1 }; int mp_bus_id_to_pci_bus [MAX_MP_BUSSES] = { [0 ... MAX_MP_BUSSES-1] = -1 }; -cpumask_t pci_bus_to_cpumask [256] = { [0 ... 255] = CPU_MASK_ALL }; +unsigned char pci_bus_to_node [256]; +EXPORT_SYMBOL(pci_bus_to_node); static int mp_current_pci_id = 0; /* I/O APIC entries */ @@ -107,6 +110,7 @@ static int __init mpf_checksum(unsigned char *mp, int len) static void __init MP_processor_info (struct mpc_config_processor *m) { int ver; + static int found_bsp=0; if (!(m->mpc_cpuflag & CPU_ENABLED)) return; @@ -126,11 +130,6 @@ static void __init MP_processor_info (struct mpc_config_processor *m) " Processor ignored.\n", NR_CPUS); return; } - if (num_processors >= maxcpus) { - printk(KERN_WARNING "WARNING: maxcpus limit of %i reached." - " Processor ignored.\n", maxcpus); - return; - } num_processors++; @@ -150,7 +149,19 @@ static void __init MP_processor_info (struct mpc_config_processor *m) ver = 0x10; } apic_version[m->mpc_apicid] = ver; - bios_cpu_apicid[num_processors - 1] = m->mpc_apicid; + if (m->mpc_cpuflag & CPU_BOOTPROCESSOR) { + /* + * bios_cpu_apicid is required to have processors listed + * in same order as logical cpu numbers. Hence the first + * entry is BSP, and so on. + */ + bios_cpu_apicid[0] = m->mpc_apicid; + x86_cpu_to_apicid[0] = m->mpc_apicid; + found_bsp = 1; + } else { + bios_cpu_apicid[num_processors - found_bsp] = m->mpc_apicid; + x86_cpu_to_apicid[num_processors - found_bsp] = m->mpc_apicid; + } } static void __init MP_bus_info (struct mpc_config_bus *m) @@ -759,7 +770,7 @@ void __init mp_register_ioapic ( mp_ioapics[idx].mpc_apicaddr = address; set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address); - mp_ioapics[idx].mpc_apicid = io_apic_get_unique_id(idx, id); + mp_ioapics[idx].mpc_apicid = id; mp_ioapics[idx].mpc_apicver = io_apic_get_version(idx); /* @@ -895,11 +906,20 @@ void __init mp_config_acpi_legacy_irqs (void) return; } +#define MAX_GSI_NUM 4096 + int mp_register_gsi(u32 gsi, int edge_level, int active_high_low) { int ioapic = -1; int ioapic_pin = 0; int idx, bit = 0; + static int pci_irq = 16; + /* + * Mapping between Global System Interrupts, which + * represent all possible interrupts, to the IRQs + * assigned to actual devices. + */ + static int gsi_to_irq[MAX_GSI_NUM]; if (acpi_irq_model != ACPI_IRQ_MODEL_IOAPIC) return gsi; @@ -934,11 +954,21 @@ int mp_register_gsi(u32 gsi, int edge_level, int active_high_low) if ((1<<bit) & mp_ioapic_routing[ioapic].pin_programmed[idx]) { Dprintk(KERN_DEBUG "Pin %d-%d already programmed\n", mp_ioapic_routing[ioapic].apic_id, ioapic_pin); - return gsi; + return gsi_to_irq[gsi]; } mp_ioapic_routing[ioapic].pin_programmed[idx] |= (1<<bit); + if (edge_level) { + /* + * For PCI devices assign IRQs in order, avoiding gaps + * due to unused I/O APIC pins. + */ + int irq = gsi; + gsi = pci_irq++; + gsi_to_irq[irq] = gsi; + } + io_apic_set_pci_routing(ioapic, ioapic_pin, gsi, edge_level == ACPI_EDGE_SENSITIVE ? 0 : 1, active_high_low == ACPI_ACTIVE_HIGH ? 0 : 1); diff --git a/arch/x86_64/kernel/nmi.c b/arch/x86_64/kernel/nmi.c index 61de0b34a01..4e44d6e6b7e 100644 --- a/arch/x86_64/kernel/nmi.c +++ b/arch/x86_64/kernel/nmi.c @@ -33,6 +33,7 @@ #include <asm/msr.h> #include <asm/proto.h> #include <asm/kdebug.h> +#include <asm/local.h> /* * lapic_nmi_owner tracks the ownership of the lapic NMI hardware: @@ -59,7 +60,8 @@ int panic_on_timeout; unsigned int nmi_watchdog = NMI_DEFAULT; static unsigned int nmi_hz = HZ; -unsigned int nmi_perfctr_msr; /* the MSR to reset in NMI handler */ +static unsigned int nmi_perfctr_msr; /* the MSR to reset in NMI handler */ +static unsigned int nmi_p4_cccr_val; /* Note that these events don't tick when the CPU idles. This means the frequency varies with CPU load. */ @@ -71,61 +73,87 @@ unsigned int nmi_perfctr_msr; /* the MSR to reset in NMI handler */ #define K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING 0x76 #define K7_NMI_EVENT K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING -#define P6_EVNTSEL0_ENABLE (1 << 22) -#define P6_EVNTSEL_INT (1 << 20) -#define P6_EVNTSEL_OS (1 << 17) -#define P6_EVNTSEL_USR (1 << 16) -#define P6_EVENT_CPU_CLOCKS_NOT_HALTED 0x79 -#define P6_NMI_EVENT P6_EVENT_CPU_CLOCKS_NOT_HALTED +#define MSR_P4_MISC_ENABLE 0x1A0 +#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1<<7) +#define MSR_P4_MISC_ENABLE_PEBS_UNAVAIL (1<<12) +#define MSR_P4_PERFCTR0 0x300 +#define MSR_P4_CCCR0 0x360 +#define P4_ESCR_EVENT_SELECT(N) ((N)<<25) +#define P4_ESCR_OS (1<<3) +#define P4_ESCR_USR (1<<2) +#define P4_CCCR_OVF_PMI0 (1<<26) +#define P4_CCCR_OVF_PMI1 (1<<27) +#define P4_CCCR_THRESHOLD(N) ((N)<<20) +#define P4_CCCR_COMPLEMENT (1<<19) +#define P4_CCCR_COMPARE (1<<18) +#define P4_CCCR_REQUIRED (3<<16) +#define P4_CCCR_ESCR_SELECT(N) ((N)<<13) +#define P4_CCCR_ENABLE (1<<12) +/* Set up IQ_COUNTER0 to behave like a clock, by having IQ_CCCR0 filter + CRU_ESCR0 (with any non-null event selector) through a complemented + max threshold. [IA32-Vol3, Section 14.9.9] */ +#define MSR_P4_IQ_COUNTER0 0x30C +#define P4_NMI_CRU_ESCR0 (P4_ESCR_EVENT_SELECT(0x3F)|P4_ESCR_OS|P4_ESCR_USR) +#define P4_NMI_IQ_CCCR0 \ + (P4_CCCR_OVF_PMI0|P4_CCCR_THRESHOLD(15)|P4_CCCR_COMPLEMENT| \ + P4_CCCR_COMPARE|P4_CCCR_REQUIRED|P4_CCCR_ESCR_SELECT(4)|P4_CCCR_ENABLE) + +static __cpuinit inline int nmi_known_cpu(void) +{ + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_AMD: + return boot_cpu_data.x86 == 15; + case X86_VENDOR_INTEL: + return boot_cpu_data.x86 == 15; + } + return 0; +} /* Run after command line and cpu_init init, but before all other checks */ -void __init nmi_watchdog_default(void) +void __cpuinit nmi_watchdog_default(void) { if (nmi_watchdog != NMI_DEFAULT) return; - - /* For some reason the IO APIC watchdog doesn't work on the AMD - 8111 chipset. For now switch to local APIC mode using - perfctr0 there. On Intel CPUs we don't have code to handle - the perfctr and the IO-APIC seems to work, so use that. */ - - if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) { - nmi_watchdog = NMI_LOCAL_APIC; - printk(KERN_INFO - "Using local APIC NMI watchdog using perfctr0\n"); - } else { - printk(KERN_INFO "Using IO APIC NMI watchdog\n"); + if (nmi_known_cpu()) + nmi_watchdog = NMI_LOCAL_APIC; + else nmi_watchdog = NMI_IO_APIC; - } } -/* Why is there no CPUID flag for this? */ -static __init int cpu_has_lapic(void) +#ifdef CONFIG_SMP +/* The performance counters used by NMI_LOCAL_APIC don't trigger when + * the CPU is idle. To make sure the NMI watchdog really ticks on all + * CPUs during the test make them busy. + */ +static __init void nmi_cpu_busy(void *data) { - switch (boot_cpu_data.x86_vendor) { - case X86_VENDOR_INTEL: - case X86_VENDOR_AMD: - return boot_cpu_data.x86 >= 6; - /* .... add more cpus here or find a different way to figure this out. */ - default: - return 0; - } + volatile int *endflag = data; + local_irq_enable(); + /* Intentionally don't use cpu_relax here. This is + to make sure that the performance counter really ticks, + even if there is a simulator or similar that catches the + pause instruction. On a real HT machine this is fine because + all other CPUs are busy with "useless" delay loops and don't + care if they get somewhat less cycles. */ + while (*endflag == 0) + barrier(); } +#endif -static int __init check_nmi_watchdog (void) +int __init check_nmi_watchdog (void) { - int counts[NR_CPUS]; + volatile int endflag = 0; + int *counts; int cpu; - if (nmi_watchdog == NMI_NONE) - return 0; + counts = kmalloc(NR_CPUS * sizeof(int), GFP_KERNEL); + if (!counts) + return -1; - if (nmi_watchdog == NMI_LOCAL_APIC && !cpu_has_lapic()) { - nmi_watchdog = NMI_NONE; - return -1; - } + printk(KERN_INFO "testing NMI watchdog ... "); - printk(KERN_INFO "Testing NMI watchdog ... "); + if (nmi_watchdog == NMI_LOCAL_APIC) + smp_call_function(nmi_cpu_busy, (void *)&endflag, 0, 0); for (cpu = 0; cpu < NR_CPUS; cpu++) counts[cpu] = cpu_pda[cpu].__nmi_count; @@ -133,15 +161,22 @@ static int __init check_nmi_watchdog (void) mdelay((10*1000)/nmi_hz); // wait 10 ticks for (cpu = 0; cpu < NR_CPUS; cpu++) { + if (!cpu_online(cpu)) + continue; if (cpu_pda[cpu].__nmi_count - counts[cpu] <= 5) { - printk("CPU#%d: NMI appears to be stuck (%d)!\n", + endflag = 1; + printk("CPU#%d: NMI appears to be stuck (%d->%d)!\n", cpu, + counts[cpu], cpu_pda[cpu].__nmi_count); nmi_active = 0; lapic_nmi_owner &= ~LAPIC_NMI_WATCHDOG; + nmi_perfctr_msr = 0; + kfree(counts); return -1; } } + endflag = 1; printk("OK.\n"); /* now that we know it works we can reduce NMI frequency to @@ -149,10 +184,9 @@ static int __init check_nmi_watchdog (void) if (nmi_watchdog == NMI_LOCAL_APIC) nmi_hz = 1; + kfree(counts); return 0; } -/* Have this called later during boot so counters are updating */ -late_initcall(check_nmi_watchdog); int __init setup_nmi_watchdog(char *str) { @@ -170,7 +204,7 @@ int __init setup_nmi_watchdog(char *str) if (nmi >= NMI_INVALID) return 0; - nmi_watchdog = nmi; + nmi_watchdog = nmi; return 1; } @@ -185,7 +219,10 @@ static void disable_lapic_nmi_watchdog(void) wrmsr(MSR_K7_EVNTSEL0, 0, 0); break; case X86_VENDOR_INTEL: - wrmsr(MSR_IA32_EVNTSEL0, 0, 0); + if (boot_cpu_data.x86 == 15) { + wrmsr(MSR_P4_IQ_CCCR0, 0, 0); + wrmsr(MSR_P4_CRU_ESCR0, 0, 0); + } break; } nmi_active = -1; @@ -253,7 +290,7 @@ void enable_timer_nmi_watchdog(void) static int nmi_pm_active; /* nmi_active before suspend */ -static int lapic_nmi_suspend(struct sys_device *dev, pm_message_t state) +static int lapic_nmi_suspend(struct sys_device *dev, u32 state) { nmi_pm_active = nmi_active; disable_lapic_nmi_watchdog(); @@ -300,22 +337,27 @@ late_initcall(init_lapic_nmi_sysfs); * Original code written by Keith Owens. */ +static void clear_msr_range(unsigned int base, unsigned int n) +{ + unsigned int i; + + for(i = 0; i < n; ++i) + wrmsr(base+i, 0, 0); +} + static void setup_k7_watchdog(void) { int i; unsigned int evntsel; - /* No check, so can start with slow frequency */ - nmi_hz = 1; - - /* XXX should check these in EFER */ - nmi_perfctr_msr = MSR_K7_PERFCTR0; for(i = 0; i < 4; ++i) { /* Simulator may not support it */ - if (checking_wrmsrl(MSR_K7_EVNTSEL0+i, 0UL)) + if (checking_wrmsrl(MSR_K7_EVNTSEL0+i, 0UL)) { + nmi_perfctr_msr = 0; return; + } wrmsrl(MSR_K7_PERFCTR0+i, 0UL); } @@ -325,12 +367,54 @@ static void setup_k7_watchdog(void) | K7_NMI_EVENT; wrmsr(MSR_K7_EVNTSEL0, evntsel, 0); - wrmsrl(MSR_K7_PERFCTR0, -((u64)cpu_khz*1000) / nmi_hz); + wrmsr(MSR_K7_PERFCTR0, -(cpu_khz/nmi_hz*1000), -1); apic_write(APIC_LVTPC, APIC_DM_NMI); evntsel |= K7_EVNTSEL_ENABLE; wrmsr(MSR_K7_EVNTSEL0, evntsel, 0); } + +static int setup_p4_watchdog(void) +{ + unsigned int misc_enable, dummy; + + rdmsr(MSR_P4_MISC_ENABLE, misc_enable, dummy); + if (!(misc_enable & MSR_P4_MISC_ENABLE_PERF_AVAIL)) + return 0; + + nmi_perfctr_msr = MSR_P4_IQ_COUNTER0; + nmi_p4_cccr_val = P4_NMI_IQ_CCCR0; +#ifdef CONFIG_SMP + if (smp_num_siblings == 2) + nmi_p4_cccr_val |= P4_CCCR_OVF_PMI1; +#endif + + if (!(misc_enable & MSR_P4_MISC_ENABLE_PEBS_UNAVAIL)) + clear_msr_range(0x3F1, 2); + /* MSR 0x3F0 seems to have a default value of 0xFC00, but current + docs doesn't fully define it, so leave it alone for now. */ + if (boot_cpu_data.x86_model >= 0x3) { + /* MSR_P4_IQ_ESCR0/1 (0x3ba/0x3bb) removed */ + clear_msr_range(0x3A0, 26); + clear_msr_range(0x3BC, 3); + } else { + clear_msr_range(0x3A0, 31); + } + clear_msr_range(0x3C0, 6); + clear_msr_range(0x3C8, 6); + clear_msr_range(0x3E0, 2); + clear_msr_range(MSR_P4_CCCR0, 18); + clear_msr_range(MSR_P4_PERFCTR0, 18); + + wrmsr(MSR_P4_CRU_ESCR0, P4_NMI_CRU_ESCR0, 0); + wrmsr(MSR_P4_IQ_CCCR0, P4_NMI_IQ_CCCR0 & ~P4_CCCR_ENABLE, 0); + Dprintk("setting P4_IQ_COUNTER0 to 0x%08lx\n", -(cpu_khz/nmi_hz*1000)); + wrmsr(MSR_P4_IQ_COUNTER0, -(cpu_khz/nmi_hz*1000), -1); + apic_write(APIC_LVTPC, APIC_DM_NMI); + wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0); + return 1; +} + void setup_apic_nmi_watchdog(void) { switch (boot_cpu_data.x86_vendor) { @@ -341,6 +425,13 @@ void setup_apic_nmi_watchdog(void) return; setup_k7_watchdog(); break; + case X86_VENDOR_INTEL: + if (boot_cpu_data.x86 != 15) + return; + if (!setup_p4_watchdog()) + return; + break; + default: return; } @@ -355,56 +446,67 @@ void setup_apic_nmi_watchdog(void) * * as these watchdog NMI IRQs are generated on every CPU, we only * have to check the current processor. - * - * since NMIs don't listen to _any_ locks, we have to be extremely - * careful not to rely on unsafe variables. The printk might lock - * up though, so we have to break up any console locks first ... - * [when there will be more tty-related locks, break them up - * here too!] */ -static unsigned int - last_irq_sums [NR_CPUS], - alert_counter [NR_CPUS]; +static DEFINE_PER_CPU(unsigned, last_irq_sum); +static DEFINE_PER_CPU(local_t, alert_counter); +static DEFINE_PER_CPU(int, nmi_touch); void touch_nmi_watchdog (void) { int i; /* - * Just reset the alert counters, (other CPUs might be - * spinning on locks we hold): + * Tell other CPUs to reset their alert counters. We cannot + * do it ourselves because the alert count increase is not + * atomic. */ for (i = 0; i < NR_CPUS; i++) - alert_counter[i] = 0; + per_cpu(nmi_touch, i) = 1; } void nmi_watchdog_tick (struct pt_regs * regs, unsigned reason) { - int sum, cpu; + int sum; + int touched = 0; - cpu = safe_smp_processor_id(); sum = read_pda(apic_timer_irqs); - if (last_irq_sums[cpu] == sum) { + if (__get_cpu_var(nmi_touch)) { + __get_cpu_var(nmi_touch) = 0; + touched = 1; + } + if (!touched && __get_cpu_var(last_irq_sum) == sum) { /* * Ayiee, looks like this CPU is stuck ... * wait a few IRQs (5 seconds) before doing the oops ... */ - alert_counter[cpu]++; - if (alert_counter[cpu] == 5*nmi_hz) { + local_inc(&__get_cpu_var(alert_counter)); + if (local_read(&__get_cpu_var(alert_counter)) == 5*nmi_hz) { if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP) { - alert_counter[cpu] = 0; + local_set(&__get_cpu_var(alert_counter), 0); return; } die_nmi("NMI Watchdog detected LOCKUP on CPU%d", regs); } } else { - last_irq_sums[cpu] = sum; - alert_counter[cpu] = 0; + __get_cpu_var(last_irq_sum) = sum; + local_set(&__get_cpu_var(alert_counter), 0); } - if (nmi_perfctr_msr) + if (nmi_perfctr_msr) { + if (nmi_perfctr_msr == MSR_P4_IQ_COUNTER0) { + /* + * P4 quirks: + * - An overflown perfctr will assert its interrupt + * until the OVF flag in its CCCR is cleared. + * - LVTPC is masked on interrupt and must be + * unmasked by the LVTPC handler. + */ + wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0); + apic_write(APIC_LVTPC, APIC_DM_NMI); + } wrmsr(nmi_perfctr_msr, -(cpu_khz/nmi_hz*1000), -1); + } } static int dummy_nmi_callback(struct pt_regs * regs, int cpu) diff --git a/arch/x86_64/kernel/pmtimer.c b/arch/x86_64/kernel/pmtimer.c new file mode 100644 index 00000000000..feb5f108dd2 --- /dev/null +++ b/arch/x86_64/kernel/pmtimer.c @@ -0,0 +1,101 @@ +/* Ported over from i386 by AK, original copyright was: + * + * (C) Dominik Brodowski <linux@brodo.de> 2003 + * + * Driver to use the Power Management Timer (PMTMR) available in some + * southbridges as primary timing source for the Linux kernel. + * + * Based on parts of linux/drivers/acpi/hardware/hwtimer.c, timer_pit.c, + * timer_hpet.c, and on Arjan van de Ven's implementation for 2.4. + * + * This file is licensed under the GPL v2. + * + * Dropped all the hardware bug workarounds for now. Hopefully they + * are not needed on 64bit chipsets. + */ + +#include <linux/jiffies.h> +#include <linux/kernel.h> +#include <linux/time.h> +#include <linux/init.h> +#include <linux/cpumask.h> +#include <asm/io.h> +#include <asm/proto.h> +#include <asm/msr.h> +#include <asm/vsyscall.h> + +/* The I/O port the PMTMR resides at. + * The location is detected during setup_arch(), + * in arch/i386/kernel/acpi/boot.c */ +u32 pmtmr_ioport; + +/* value of the Power timer at last timer interrupt */ +static u32 offset_delay; +static u32 last_pmtmr_tick; + +#define ACPI_PM_MASK 0xFFFFFF /* limit it to 24 bits */ + +static inline u32 cyc2us(u32 cycles) +{ + /* The Power Management Timer ticks at 3.579545 ticks per microsecond. + * 1 / PM_TIMER_FREQUENCY == 0.27936511 =~ 286/1024 [error: 0.024%] + * + * Even with HZ = 100, delta is at maximum 35796 ticks, so it can + * easily be multiplied with 286 (=0x11E) without having to fear + * u32 overflows. + */ + cycles *= 286; + return (cycles >> 10); +} + +int pmtimer_mark_offset(void) +{ + static int first_run = 1; + unsigned long tsc; + u32 lost; + + u32 tick = inl(pmtmr_ioport); + u32 delta; + + delta = cyc2us((tick - last_pmtmr_tick) & ACPI_PM_MASK); + + last_pmtmr_tick = tick; + monotonic_base += delta * NSEC_PER_USEC; + + delta += offset_delay; + + lost = delta / (USEC_PER_SEC / HZ); + offset_delay = delta % (USEC_PER_SEC / HZ); + + rdtscll(tsc); + vxtime.last_tsc = tsc - offset_delay * cpu_khz; + + /* don't calculate delay for first run, + or if we've got less then a tick */ + if (first_run || (lost < 1)) { + first_run = 0; + offset_delay = 0; + } + + return lost - 1; +} + +unsigned int do_gettimeoffset_pm(void) +{ + u32 now, offset, delta = 0; + + offset = last_pmtmr_tick; + now = inl(pmtmr_ioport); + delta = (now - offset) & ACPI_PM_MASK; + + return offset_delay + cyc2us(delta); +} + + +static int __init nopmtimer_setup(char *s) +{ + pmtmr_ioport = 0; + return 0; +} + +__setup("nopmtimer", nopmtimer_setup); diff --git a/arch/x86_64/kernel/process.c b/arch/x86_64/kernel/process.c index 761b6d35e33..1d91271796e 100644 --- a/arch/x86_64/kernel/process.c +++ b/arch/x86_64/kernel/process.c @@ -8,7 +8,8 @@ * * X86-64 port * Andi Kleen. - * + * + * CPU hotplug support - ashok.raj@intel.com * $Id: process.c,v 1.38 2002/01/15 10:08:03 ak Exp $ */ @@ -18,6 +19,7 @@ #include <stdarg.h> +#include <linux/cpu.h> #include <linux/errno.h> #include <linux/sched.h> #include <linux/kernel.h> @@ -34,6 +36,7 @@ #include <linux/ptrace.h> #include <linux/utsname.h> #include <linux/random.h> +#include <linux/kprobes.h> #include <asm/uaccess.h> #include <asm/pgtable.h> @@ -153,6 +156,29 @@ void cpu_idle_wait(void) } EXPORT_SYMBOL_GPL(cpu_idle_wait); +#ifdef CONFIG_HOTPLUG_CPU +DECLARE_PER_CPU(int, cpu_state); + +#include <asm/nmi.h> +/* We don't actually take CPU down, just spin without interrupts. */ +static inline void play_dead(void) +{ + idle_task_exit(); + wbinvd(); + mb(); + /* Ack it */ + __get_cpu_var(cpu_state) = CPU_DEAD; + + while (1) + safe_halt(); +} +#else +static inline void play_dead(void) +{ + BUG(); +} +#endif /* CONFIG_HOTPLUG_CPU */ + /* * The idle thread. There's no useful work to be * done, so just try to conserve power and have a @@ -173,6 +199,8 @@ void cpu_idle (void) idle = pm_idle; if (!idle) idle = default_idle; + if (cpu_is_offline(smp_processor_id())) + play_dead(); idle(); } @@ -203,7 +231,7 @@ static void mwait_idle(void) } } -void __init select_idle_routine(const struct cpuinfo_x86 *c) +void __cpuinit select_idle_routine(const struct cpuinfo_x86 *c) { static int printed; if (cpu_has(c, X86_FEATURE_MWAIT)) { @@ -293,6 +321,14 @@ void exit_thread(void) { struct task_struct *me = current; struct thread_struct *t = &me->thread; + + /* + * Remove function-return probe instances associated with this task + * and put them back on the free list. Do not insert an exit probe for + * this function, it will be disabled by kprobe_flush_task if you do. + */ + kprobe_flush_task(me); + if (me->thread.io_bitmap_ptr) { struct tss_struct *tss = &per_cpu(init_tss, get_cpu()); @@ -312,6 +348,13 @@ void flush_thread(void) struct task_struct *tsk = current; struct thread_info *t = current_thread_info(); + /* + * Remove function-return probe instances associated with this task + * and put them back on the free list. Do not insert an exit probe for + * this function, it will be disabled by kprobe_flush_task if you do. + */ + kprobe_flush_task(tsk); + if (t->flags & _TIF_ABI_PENDING) t->flags ^= (_TIF_ABI_PENDING | _TIF_IA32); @@ -656,7 +699,7 @@ long do_arch_prctl(struct task_struct *task, int code, unsigned long addr) switch (code) { case ARCH_SET_GS: - if (addr >= TASK_SIZE) + if (addr >= TASK_SIZE_OF(task)) return -EPERM; cpu = get_cpu(); /* handle small bases via the GDT because that's faster to @@ -682,7 +725,7 @@ long do_arch_prctl(struct task_struct *task, int code, unsigned long addr) case ARCH_SET_FS: /* Not strictly needed for fs, but do it for symmetry with gs */ - if (addr >= TASK_SIZE) + if (addr >= TASK_SIZE_OF(task)) return -EPERM; cpu = get_cpu(); /* handle small bases via the GDT because that's faster to diff --git a/arch/x86_64/kernel/ptrace.c b/arch/x86_64/kernel/ptrace.c index e26e86bb56f..bbf64b59a21 100644 --- a/arch/x86_64/kernel/ptrace.c +++ b/arch/x86_64/kernel/ptrace.c @@ -257,13 +257,13 @@ static int putreg(struct task_struct *child, value &= 0xffff; return 0; case offsetof(struct user_regs_struct,fs_base): - if (!((value >> 48) == 0 || (value >> 48) == 0xffff)) - return -EIO; + if (value >= TASK_SIZE_OF(child)) + return -EIO; child->thread.fs = value; return 0; case offsetof(struct user_regs_struct,gs_base): - if (!((value >> 48) == 0 || (value >> 48) == 0xffff)) - return -EIO; + if (value >= TASK_SIZE_OF(child)) + return -EIO; child->thread.gs = value; return 0; case offsetof(struct user_regs_struct, eflags): @@ -277,6 +277,11 @@ static int putreg(struct task_struct *child, return -EIO; value &= 0xffff; break; + case offsetof(struct user_regs_struct, rip): + /* Check if the new RIP address is canonical */ + if (value >= TASK_SIZE_OF(child)) + return -EIO; + break; } put_stack_long(child, regno - sizeof(struct pt_regs), value); return 0; @@ -375,7 +380,7 @@ asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, long data break; switch (addr) { - case 0 ... sizeof(struct user_regs_struct): + case 0 ... sizeof(struct user_regs_struct) - sizeof(long): tmp = getreg(child, addr); break; case offsetof(struct user, u_debugreg[0]): @@ -414,33 +419,35 @@ asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, long data break; case PTRACE_POKEUSR: /* write the word at location addr in the USER area */ + { + int dsize = test_tsk_thread_flag(child, TIF_IA32) ? 3 : 7; ret = -EIO; if ((addr & 7) || addr > sizeof(struct user) - 7) break; switch (addr) { - case 0 ... sizeof(struct user_regs_struct): + case 0 ... sizeof(struct user_regs_struct) - sizeof(long): ret = putreg(child, addr, data); break; /* Disallows to set a breakpoint into the vsyscall */ case offsetof(struct user, u_debugreg[0]): - if (data >= TASK_SIZE-7) break; + if (data >= TASK_SIZE_OF(child) - dsize) break; child->thread.debugreg0 = data; ret = 0; break; case offsetof(struct user, u_debugreg[1]): - if (data >= TASK_SIZE-7) break; + if (data >= TASK_SIZE_OF(child) - dsize) break; child->thread.debugreg1 = data; ret = 0; break; case offsetof(struct user, u_debugreg[2]): - if (data >= TASK_SIZE-7) break; + if (data >= TASK_SIZE_OF(child) - dsize) break; child->thread.debugreg2 = data; ret = 0; break; case offsetof(struct user, u_debugreg[3]): - if (data >= TASK_SIZE-7) break; + if (data >= TASK_SIZE_OF(child) - dsize) break; child->thread.debugreg3 = data; ret = 0; break; @@ -464,6 +471,7 @@ asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, long data break; } break; + } case PTRACE_SYSCALL: /* continue and stop at next (return from) syscall */ case PTRACE_CONT: /* restart after signal. */ diff --git a/arch/x86_64/kernel/reboot.c b/arch/x86_64/kernel/reboot.c index be4b36f762c..57e71dbdfd6 100644 --- a/arch/x86_64/kernel/reboot.c +++ b/arch/x86_64/kernel/reboot.c @@ -66,41 +66,47 @@ static int __init reboot_setup(char *str) __setup("reboot=", reboot_setup); -#ifdef CONFIG_SMP -static void smp_halt(void) +static inline void kb_wait(void) { - int cpuid = safe_smp_processor_id(); - static int first_entry = 1; + int i; - if (reboot_force) - return; + for (i=0; i<0x10000; i++) + if ((inb_p(0x64) & 0x02) == 0) + break; +} - if (first_entry) { - first_entry = 0; - smp_call_function((void *)machine_restart, NULL, 1, 0); - } - - smp_stop_cpu(); +void machine_shutdown(void) +{ + /* Stop the cpus and apics */ +#ifdef CONFIG_SMP + int reboot_cpu_id; - /* AP calling this. Just halt */ - if (cpuid != boot_cpu_id) { - for (;;) - asm("hlt"); + /* The boot cpu is always logical cpu 0 */ + reboot_cpu_id = 0; + + /* Make certain the cpu I'm about to reboot on is online */ + if (!cpu_isset(reboot_cpu_id, cpu_online_map)) { + reboot_cpu_id = smp_processor_id(); } - /* Wait for all other CPUs to have run smp_stop_cpu */ - while (!cpus_empty(cpu_online_map)) - rep_nop(); -} + /* Make certain I only run on the appropriate processor */ + set_cpus_allowed(current, cpumask_of_cpu(reboot_cpu_id)); + + /* O.K Now that I'm on the appropriate processor, + * stop all of the others. + */ + smp_send_stop(); #endif -static inline void kb_wait(void) -{ - int i; + local_irq_disable(); - for (i=0; i<0x10000; i++) - if ((inb_p(0x64) & 0x02) == 0) - break; +#ifndef CONFIG_SMP + disable_local_APIC(); +#endif + + disable_IO_APIC(); + + local_irq_enable(); } void machine_restart(char * __unused) @@ -109,9 +115,7 @@ void machine_restart(char * __unused) printk("machine restart\n"); -#ifdef CONFIG_SMP - smp_halt(); -#endif + machine_shutdown(); if (!reboot_force) { local_irq_disable(); diff --git a/arch/x86_64/kernel/relocate_kernel.S b/arch/x86_64/kernel/relocate_kernel.S new file mode 100644 index 00000000000..d24fa9b72a2 --- /dev/null +++ b/arch/x86_64/kernel/relocate_kernel.S @@ -0,0 +1,143 @@ +/* + * relocate_kernel.S - put the kernel image in place to boot + * Copyright (C) 2002-2005 Eric Biederman <ebiederm@xmission.com> + * + * This source code is licensed under the GNU General Public License, + * Version 2. See the file COPYING for more details. + */ + +#include <linux/linkage.h> + + /* + * Must be relocatable PIC code callable as a C function, that once + * it starts can not use the previous processes stack. + */ + .globl relocate_new_kernel + .code64 +relocate_new_kernel: + /* %rdi page_list + * %rsi reboot_code_buffer + * %rdx start address + * %rcx page_table + * %r8 arg5 + * %r9 arg6 + */ + + /* zero out flags, and disable interrupts */ + pushq $0 + popfq + + /* set a new stack at the bottom of our page... */ + lea 4096(%rsi), %rsp + + /* store the parameters back on the stack */ + pushq %rdx /* store the start address */ + + /* Set cr0 to a known state: + * 31 1 == Paging enabled + * 18 0 == Alignment check disabled + * 16 0 == Write protect disabled + * 3 0 == No task switch + * 2 0 == Don't do FP software emulation. + * 0 1 == Proctected mode enabled + */ + movq %cr0, %rax + andq $~((1<<18)|(1<<16)|(1<<3)|(1<<2)), %rax + orl $((1<<31)|(1<<0)), %eax + movq %rax, %cr0 + + /* Set cr4 to a known state: + * 10 0 == xmm exceptions disabled + * 9 0 == xmm registers instructions disabled + * 8 0 == performance monitoring counter disabled + * 7 0 == page global disabled + * 6 0 == machine check exceptions disabled + * 5 1 == physical address extension enabled + * 4 0 == page size extensions disabled + * 3 0 == Debug extensions disabled + * 2 0 == Time stamp disable (disabled) + * 1 0 == Protected mode virtual interrupts disabled + * 0 0 == VME disabled + */ + + movq $((1<<5)), %rax + movq %rax, %cr4 + + jmp 1f +1: + + /* Switch to the identity mapped page tables, + * and flush the TLB. + */ + movq %rcx, %cr3 + + /* Do the copies */ + movq %rdi, %rcx /* Put the page_list in %rcx */ + xorq %rdi, %rdi + xorq %rsi, %rsi + jmp 1f + +0: /* top, read another word for the indirection page */ + + movq (%rbx), %rcx + addq $8, %rbx +1: + testq $0x1, %rcx /* is it a destination page? */ + jz 2f + movq %rcx, %rdi + andq $0xfffffffffffff000, %rdi + jmp 0b +2: + testq $0x2, %rcx /* is it an indirection page? */ + jz 2f + movq %rcx, %rbx + andq $0xfffffffffffff000, %rbx + jmp 0b +2: + testq $0x4, %rcx /* is it the done indicator? */ + jz 2f + jmp 3f +2: + testq $0x8, %rcx /* is it the source indicator? */ + jz 0b /* Ignore it otherwise */ + movq %rcx, %rsi /* For ever source page do a copy */ + andq $0xfffffffffffff000, %rsi + + movq $512, %rcx + rep ; movsq + jmp 0b +3: + + /* To be certain of avoiding problems with self-modifying code + * I need to execute a serializing instruction here. + * So I flush the TLB by reloading %cr3 here, it's handy, + * and not processor dependent. + */ + movq %cr3, %rax + movq %rax, %cr3 + + /* set all of the registers to known values */ + /* leave %rsp alone */ + + xorq %rax, %rax + xorq %rbx, %rbx + xorq %rcx, %rcx + xorq %rdx, %rdx + xorq %rsi, %rsi + xorq %rdi, %rdi + xorq %rbp, %rbp + xorq %r8, %r8 + xorq %r9, %r9 + xorq %r10, %r9 + xorq %r11, %r11 + xorq %r12, %r12 + xorq %r13, %r13 + xorq %r14, %r14 + xorq %r15, %r15 + + ret +relocate_new_kernel_end: + + .globl relocate_new_kernel_size +relocate_new_kernel_size: + .quad relocate_new_kernel_end - relocate_new_kernel diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c index 2129cf9ba6b..b02d921da4f 100644 --- a/arch/x86_64/kernel/setup.c +++ b/arch/x86_64/kernel/setup.c @@ -40,6 +40,9 @@ #include <linux/acpi.h> #include <linux/kallsyms.h> #include <linux/edd.h> +#include <linux/mmzone.h> +#include <linux/kexec.h> + #include <asm/mtrr.h> #include <asm/uaccess.h> #include <asm/system.h> @@ -348,7 +351,7 @@ static __init void parse_cmdline_early (char ** cmdline_p) if (!memcmp(from, "mem=", 4)) parse_memopt(from+4, &from); -#ifdef CONFIG_DISCONTIGMEM +#ifdef CONFIG_NUMA if (!memcmp(from, "numa=", 5)) numa_setup(from+5); #endif @@ -365,6 +368,27 @@ static __init void parse_cmdline_early (char ** cmdline_p) if (!memcmp(from, "noexec=", 7)) nonx_setup(from + 7); +#ifdef CONFIG_KEXEC + /* crashkernel=size@addr specifies the location to reserve for + * a crash kernel. By reserving this memory we guarantee + * that linux never set's it up as a DMA target. + * Useful for holding code to do something appropriate + * after a kernel panic. + */ + else if (!memcmp(from, "crashkernel=", 12)) { + unsigned long size, base; + size = memparse(from+12, &from); + if (*from == '@') { + base = memparse(from+1, &from); + /* FIXME: Do I want a sanity check + * to validate the memory range? + */ + crashk_res.start = base; + crashk_res.end = base + size - 1; + } + } +#endif + next_char: c = *(from++); if (!c) @@ -377,17 +401,20 @@ static __init void parse_cmdline_early (char ** cmdline_p) *cmdline_p = command_line; } -#ifndef CONFIG_DISCONTIGMEM -static void __init contig_initmem_init(void) +#ifndef CONFIG_NUMA +static void __init +contig_initmem_init(unsigned long start_pfn, unsigned long end_pfn) { - unsigned long bootmap_size, bootmap; - bootmap_size = bootmem_bootmap_pages(end_pfn)<<PAGE_SHIFT; - bootmap = find_e820_area(0, end_pfn<<PAGE_SHIFT, bootmap_size); - if (bootmap == -1L) - panic("Cannot find bootmem map of size %ld\n",bootmap_size); - bootmap_size = init_bootmem(bootmap >> PAGE_SHIFT, end_pfn); - e820_bootmem_free(&contig_page_data, 0, end_pfn << PAGE_SHIFT); - reserve_bootmem(bootmap, bootmap_size); + unsigned long bootmap_size, bootmap; + + memory_present(0, start_pfn, end_pfn); + bootmap_size = bootmem_bootmap_pages(end_pfn)<<PAGE_SHIFT; + bootmap = find_e820_area(0, end_pfn<<PAGE_SHIFT, bootmap_size); + if (bootmap == -1L) + panic("Cannot find bootmem map of size %ld\n",bootmap_size); + bootmap_size = init_bootmem(bootmap >> PAGE_SHIFT, end_pfn); + e820_bootmem_free(NODE_DATA(0), 0, end_pfn << PAGE_SHIFT); + reserve_bootmem(bootmap, bootmap_size); } #endif @@ -554,10 +581,10 @@ void __init setup_arch(char **cmdline_p) acpi_numa_init(); #endif -#ifdef CONFIG_DISCONTIGMEM +#ifdef CONFIG_NUMA numa_initmem_init(0, end_pfn); #else - contig_initmem_init(); + contig_initmem_init(0, end_pfn); #endif /* Reserve direct mapping */ @@ -618,6 +645,15 @@ void __init setup_arch(char **cmdline_p) } } #endif + + sparse_init(); + +#ifdef CONFIG_KEXEC + if (crashk_res.start != crashk_res.end) { + reserve_bootmem(crashk_res.start, + crashk_res.end - crashk_res.start + 1); + } +#endif paging_init(); check_ioapic(); @@ -669,7 +705,7 @@ void __init setup_arch(char **cmdline_p) #endif } -static int __init get_model_name(struct cpuinfo_x86 *c) +static int __cpuinit get_model_name(struct cpuinfo_x86 *c) { unsigned int *v; @@ -685,7 +721,7 @@ static int __init get_model_name(struct cpuinfo_x86 *c) } -static void __init display_cacheinfo(struct cpuinfo_x86 *c) +static void __cpuinit display_cacheinfo(struct cpuinfo_x86 *c) { unsigned int n, dummy, eax, ebx, ecx, edx; @@ -719,7 +755,6 @@ static void __init display_cacheinfo(struct cpuinfo_x86 *c) } } -#ifdef CONFIG_SMP /* * On a AMD dual core setup the lower bits of the APIC id distingush the cores. * Assumes number of cores is a power of two. @@ -727,17 +762,26 @@ static void __init display_cacheinfo(struct cpuinfo_x86 *c) static void __init amd_detect_cmp(struct cpuinfo_x86 *c) { #ifdef CONFIG_SMP - int cpu = c->x86_apicid; + int cpu = smp_processor_id(); int node = 0; + unsigned bits; if (c->x86_num_cores == 1) return; - cpu_core_id[cpu] = cpu >> hweight32(c->x86_num_cores - 1); + + bits = 0; + while ((1 << bits) < c->x86_num_cores) + bits++; + + /* Low order bits define the core id (index of core in socket) */ + cpu_core_id[cpu] = phys_proc_id[cpu] & ((1 << bits)-1); + /* Convert the APIC ID into the socket ID */ + phys_proc_id[cpu] >>= bits; #ifdef CONFIG_NUMA /* When an ACPI SRAT table is available use the mappings from SRAT instead. */ if (acpi_numa <= 0) { - node = cpu_core_id[cpu]; + node = phys_proc_id[cpu]; if (!node_online(node)) node = first_node(node_online_map); cpu_to_node[cpu] = node; @@ -745,15 +789,11 @@ static void __init amd_detect_cmp(struct cpuinfo_x86 *c) node = cpu_to_node[cpu]; } #endif + printk(KERN_INFO "CPU %d(%d) -> Node %d -> Core %d\n", cpu, c->x86_num_cores, node, cpu_core_id[cpu]); #endif } -#else -static void __init amd_detect_cmp(struct cpuinfo_x86 *c) -{ -} -#endif static int __init init_amd(struct cpuinfo_x86 *c) { @@ -792,7 +832,7 @@ static int __init init_amd(struct cpuinfo_x86 *c) return r; } -static void __init detect_ht(struct cpuinfo_x86 *c) +static void __cpuinit detect_ht(struct cpuinfo_x86 *c) { #ifdef CONFIG_SMP u32 eax, ebx, ecx, edx; @@ -853,7 +893,7 @@ static void __init detect_ht(struct cpuinfo_x86 *c) /* * find out the number of processor cores on the die */ -static int __init intel_num_cpu_cores(struct cpuinfo_x86 *c) +static int __cpuinit intel_num_cpu_cores(struct cpuinfo_x86 *c) { unsigned int eax; @@ -871,7 +911,7 @@ static int __init intel_num_cpu_cores(struct cpuinfo_x86 *c) return 1; } -static void __init init_intel(struct cpuinfo_x86 *c) +static void __cpuinit init_intel(struct cpuinfo_x86 *c) { /* Cache sizes */ unsigned n; @@ -891,7 +931,7 @@ static void __init init_intel(struct cpuinfo_x86 *c) c->x86_num_cores = intel_num_cpu_cores(c); } -void __init get_cpu_vendor(struct cpuinfo_x86 *c) +void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c) { char *v = c->x86_vendor_id; @@ -912,7 +952,7 @@ struct cpu_model_info { /* Do some early cpuid on the boot CPU to get some parameter that are needed before check_bugs. Everything advanced is in identify_cpu below. */ -void __init early_identify_cpu(struct cpuinfo_x86 *c) +void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c) { u32 tfms; @@ -925,7 +965,6 @@ void __init early_identify_cpu(struct cpuinfo_x86 *c) c->x86_clflush_size = 64; c->x86_cache_alignment = c->x86_clflush_size; c->x86_num_cores = 1; - c->x86_apicid = c == &boot_cpu_data ? 0 : c - cpu_data; c->extended_cpuid_level = 0; memset(&c->x86_capability, 0, sizeof c->x86_capability); @@ -954,17 +993,20 @@ void __init early_identify_cpu(struct cpuinfo_x86 *c) } if (c->x86_capability[0] & (1<<19)) c->x86_clflush_size = ((misc >> 8) & 0xff) * 8; - c->x86_apicid = misc >> 24; } else { /* Have CPUID level 0 only - unheard of */ c->x86 = 4; } + +#ifdef CONFIG_SMP + phys_proc_id[smp_processor_id()] = (cpuid_ebx(1) >> 24) & 0xff; +#endif } /* * This does the hard work of actually picking apart the CPU stuff... */ -void __init identify_cpu(struct cpuinfo_x86 *c) +void __cpuinit identify_cpu(struct cpuinfo_x86 *c) { int i; u32 xlvl; @@ -1041,7 +1083,7 @@ void __init identify_cpu(struct cpuinfo_x86 *c) } -void __init print_cpu_info(struct cpuinfo_x86 *c) +void __cpuinit print_cpu_info(struct cpuinfo_x86 *c) { if (c->x86_model_id[0]) printk("%s", c->x86_model_id); @@ -1088,7 +1130,7 @@ static int show_cpuinfo(struct seq_file *m, void *v) NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, /* Other (Linux-defined) */ - "cxmmx", NULL, "cyrix_arr", "centaur_mcr", "k8c+", + "cxmmx", NULL, "cyrix_arr", "centaur_mcr", NULL, "constant_tsc", NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, diff --git a/arch/x86_64/kernel/setup64.c b/arch/x86_64/kernel/setup64.c index 678b7ac33b8..f2b9c6bc999 100644 --- a/arch/x86_64/kernel/setup64.c +++ b/arch/x86_64/kernel/setup64.c @@ -29,7 +29,7 @@ char x86_boot_params[BOOT_PARAM_SIZE] __initdata = {0,}; -cpumask_t cpu_initialized __initdata = CPU_MASK_NONE; +cpumask_t cpu_initialized __cpuinitdata = CPU_MASK_NONE; struct x8664_pda cpu_pda[NR_CPUS] __cacheline_aligned; @@ -171,7 +171,7 @@ void syscall_init(void) wrmsrl(MSR_SYSCALL_MASK, EF_TF|EF_DF|EF_IE|0x3000); } -void __init check_efer(void) +void __cpuinit check_efer(void) { unsigned long efer; @@ -188,7 +188,7 @@ void __init check_efer(void) * 'CPU state barrier', nothing should get across. * A lot of state is already set up in PDA init. */ -void __init cpu_init (void) +void __cpuinit cpu_init (void) { #ifdef CONFIG_SMP int cpu = stack_smp_processor_id(); diff --git a/arch/x86_64/kernel/signal.c b/arch/x86_64/kernel/signal.c index d439ced150c..98590a989f3 100644 --- a/arch/x86_64/kernel/signal.c +++ b/arch/x86_64/kernel/signal.c @@ -28,14 +28,15 @@ #include <asm/uaccess.h> #include <asm/i387.h> #include <asm/proto.h> +#include <asm/ia32_unistd.h> /* #define DEBUG_SIG 1 */ #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) -void ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, +int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, sigset_t *set, struct pt_regs * regs); -void ia32_setup_frame(int sig, struct k_sigaction *ka, +int ia32_setup_frame(int sig, struct k_sigaction *ka, sigset_t *set, struct pt_regs * regs); asmlinkage long @@ -237,7 +238,7 @@ get_stack(struct k_sigaction *ka, struct pt_regs *regs, unsigned long size) return (void __user *)round_down(rsp - size, 16); } -static void setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, +static int setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, sigset_t *set, struct pt_regs * regs) { struct rt_sigframe __user *frame; @@ -326,20 +327,23 @@ static void setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, current->comm, current->pid, frame, regs->rip, frame->pretcode); #endif - return; + return 1; give_sigsegv: force_sigsegv(sig, current); + return 0; } /* * OK, we're invoking a handler */ -static void +static int handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, sigset_t *oldset, struct pt_regs *regs) { + int ret; + #ifdef DEBUG_SIG printk("handle_signal pid:%d sig:%lu rip:%lx rsp:%lx regs=%p\n", current->pid, sig, @@ -383,20 +387,22 @@ handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, #ifdef CONFIG_IA32_EMULATION if (test_thread_flag(TIF_IA32)) { if (ka->sa.sa_flags & SA_SIGINFO) - ia32_setup_rt_frame(sig, ka, info, oldset, regs); + ret = ia32_setup_rt_frame(sig, ka, info, oldset, regs); else - ia32_setup_frame(sig, ka, oldset, regs); + ret = ia32_setup_frame(sig, ka, oldset, regs); } else #endif - setup_rt_frame(sig, ka, info, oldset, regs); + ret = setup_rt_frame(sig, ka, info, oldset, regs); - if (!(ka->sa.sa_flags & SA_NODEFER)) { + if (ret && !(ka->sa.sa_flags & SA_NODEFER)) { spin_lock_irq(¤t->sighand->siglock); sigorsets(¤t->blocked,¤t->blocked,&ka->sa.sa_mask); sigaddset(¤t->blocked,sig); recalc_sigpending(); spin_unlock_irq(¤t->sighand->siglock); } + + return ret; } /* @@ -416,10 +422,10 @@ int do_signal(struct pt_regs *regs, sigset_t *oldset) * kernel mode. Just return without doing anything * if so. */ - if ((regs->cs & 3) != 3) + if (!user_mode(regs)) return 1; - if (try_to_freeze(0)) + if (try_to_freeze()) goto no_signal; if (!oldset) @@ -433,11 +439,10 @@ int do_signal(struct pt_regs *regs, sigset_t *oldset) * inside the kernel. */ if (current->thread.debugreg7) - asm volatile("movq %0,%%db7" : : "r" (current->thread.debugreg7)); + set_debugreg(current->thread.debugreg7, 7); /* Whee! Actually deliver the signal. */ - handle_signal(signr, &info, &ka, oldset, regs); - return 1; + return handle_signal(signr, &info, &ka, oldset, regs); } no_signal: @@ -452,7 +457,9 @@ int do_signal(struct pt_regs *regs, sigset_t *oldset) regs->rip -= 2; } if (regs->rax == (unsigned long)-ERESTART_RESTARTBLOCK) { - regs->rax = __NR_restart_syscall; + regs->rax = test_thread_flag(TIF_IA32) ? + __NR_ia32_restart_syscall : + __NR_restart_syscall; regs->rip -= 2; } } diff --git a/arch/x86_64/kernel/smp.c b/arch/x86_64/kernel/smp.c index 1e379ed17b1..ccae392886a 100644 --- a/arch/x86_64/kernel/smp.c +++ b/arch/x86_64/kernel/smp.c @@ -283,6 +283,16 @@ struct call_data_struct { static struct call_data_struct * call_data; +void lock_ipi_call_lock(void) +{ + spin_lock_irq(&call_lock); +} + +void unlock_ipi_call_lock(void) +{ + spin_unlock_irq(&call_lock); +} + /* * this function sends a 'generic call function' IPI to all other CPUs * in the system. diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c index 73f7e8b9543..b969ee12872 100644 --- a/arch/x86_64/kernel/smpboot.c +++ b/arch/x86_64/kernel/smpboot.c @@ -34,6 +34,7 @@ * Andi Kleen : Converted to new state machine. * Various cleanups. * Probably mostly hotplug CPU ready now. + * Ashok Raj : CPU hotplug support */ @@ -56,11 +57,7 @@ #include <asm/kdebug.h> #include <asm/tlbflush.h> #include <asm/proto.h> - -/* Change for real CPU hotplug. Note other files need to be fixed - first too. */ -#define __cpuinit __init -#define __cpuinitdata __initdata +#include <asm/nmi.h> /* Number of siblings per CPU package */ int smp_num_siblings = 1; @@ -93,6 +90,7 @@ int smp_threads_ready; cpumask_t cpu_sibling_map[NR_CPUS] __cacheline_aligned; cpumask_t cpu_core_map[NR_CPUS] __cacheline_aligned; +EXPORT_SYMBOL(cpu_core_map); /* * Trampoline 80x86 program as an array. @@ -101,6 +99,37 @@ cpumask_t cpu_core_map[NR_CPUS] __cacheline_aligned; extern unsigned char trampoline_data[]; extern unsigned char trampoline_end[]; +/* State of each CPU */ +DEFINE_PER_CPU(int, cpu_state) = { 0 }; + +/* + * Store all idle threads, this can be reused instead of creating + * a new thread. Also avoids complicated thread destroy functionality + * for idle threads. + */ +struct task_struct *idle_thread_array[NR_CPUS] __cpuinitdata ; + +#define get_idle_for_cpu(x) (idle_thread_array[(x)]) +#define set_idle_for_cpu(x,p) (idle_thread_array[(x)] = (p)) + +/* + * cpu_possible_map should be static, it cannot change as cpu's + * are onlined, or offlined. The reason is per-cpu data-structures + * are allocated by some modules at init time, and dont expect to + * do this dynamically on cpu arrival/departure. + * cpu_present_map on the other hand can change dynamically. + * In case when cpu_hotplug is not compiled, then we resort to current + * behaviour, which is cpu_possible == cpu_present. + * If cpu-hotplug is supported, then we need to preallocate for all + * those NR_CPUS, hence cpu_possible_map represents entire NR_CPUS range. + * - Ashok Raj + */ +#ifdef CONFIG_HOTPLUG_CPU +#define fixup_cpu_possible_map(x) cpu_set((x), cpu_possible_map) +#else +#define fixup_cpu_possible_map(x) +#endif + /* * Currently trivial. Write the real->protected mode * bootstrap into the page concerned. The caller @@ -125,96 +154,210 @@ static void __cpuinit smp_store_cpu_info(int id) *c = boot_cpu_data; identify_cpu(c); + print_cpu_info(c); } /* - * Synchronize TSCs of CPUs + * New Funky TSC sync algorithm borrowed from IA64. + * Main advantage is that it doesn't reset the TSCs fully and + * in general looks more robust and it works better than my earlier + * attempts. I believe it was written by David Mosberger. Some minor + * adjustments for x86-64 by me -AK + * + * Original comment reproduced below. * - * This new algorithm is less accurate than the old "zero TSCs" - * one, but we cannot zero TSCs anymore in the new hotplug CPU - * model. + * Synchronize TSC of the current (slave) CPU with the TSC of the + * MASTER CPU (normally the time-keeper CPU). We use a closed loop to + * eliminate the possibility of unaccounted-for errors (such as + * getting a machine check in the middle of a calibration step). The + * basic idea is for the slave to ask the master what itc value it has + * and to read its own itc before and after the master responds. Each + * iteration gives us three timestamps: + * + * slave master + * + * t0 ---\ + * ---\ + * ---> + * tm + * /--- + * /--- + * t1 <--- + * + * + * The goal is to adjust the slave's TSC such that tm falls exactly + * half-way between t0 and t1. If we achieve this, the clocks are + * synchronized provided the interconnect between the slave and the + * master is symmetric. Even if the interconnect were asymmetric, we + * would still know that the synchronization error is smaller than the + * roundtrip latency (t0 - t1). + * + * When the interconnect is quiet and symmetric, this lets us + * synchronize the TSC to within one or two cycles. However, we can + * only *guarantee* that the synchronization is accurate to within a + * round-trip time, which is typically in the range of several hundred + * cycles (e.g., ~500 cycles). In practice, this means that the TSCs + * are usually almost perfectly synchronized, but we shouldn't assume + * that the accuracy is much better than half a micro second or so. + * + * [there are other errors like the latency of RDTSC and of the + * WRMSR. These can also account to hundreds of cycles. So it's + * probably worse. It claims 153 cycles error on a dual Opteron, + * but I suspect the numbers are actually somewhat worse -AK] */ -static atomic_t __cpuinitdata tsc_flag; +#define MASTER 0 +#define SLAVE (SMP_CACHE_BYTES/8) + +/* Intentionally don't use cpu_relax() while TSC synchronization + because we don't want to go into funky power save modi or cause + hypervisors to schedule us away. Going to sleep would likely affect + latency and low latency is the primary objective here. -AK */ +#define no_cpu_relax() barrier() + static __cpuinitdata DEFINE_SPINLOCK(tsc_sync_lock); -static unsigned long long __cpuinitdata bp_tsc, ap_tsc; +static volatile __cpuinitdata unsigned long go[SLAVE + 1]; +static int notscsync __cpuinitdata; -#define NR_LOOPS 5 +#undef DEBUG_TSC_SYNC -static void __cpuinit sync_tsc_bp_init(int init) +#define NUM_ROUNDS 64 /* magic value */ +#define NUM_ITERS 5 /* likewise */ + +/* Callback on boot CPU */ +static __cpuinit void sync_master(void *arg) { - if (init) - _raw_spin_lock(&tsc_sync_lock); - else - _raw_spin_unlock(&tsc_sync_lock); - atomic_set(&tsc_flag, 0); + unsigned long flags, i; + + if (smp_processor_id() != boot_cpu_id) + return; + + go[MASTER] = 0; + + local_irq_save(flags); + { + for (i = 0; i < NUM_ROUNDS*NUM_ITERS; ++i) { + while (!go[MASTER]) + no_cpu_relax(); + go[MASTER] = 0; + rdtscll(go[SLAVE]); + } + } + local_irq_restore(flags); } /* - * Synchronize TSC on AP with BP. + * Return the number of cycles by which our tsc differs from the tsc + * on the master (time-keeper) CPU. A positive number indicates our + * tsc is ahead of the master, negative that it is behind. */ -static void __cpuinit __sync_tsc_ap(void) +static inline long +get_delta(long *rt, long *master) { - if (!cpu_has_tsc) - return; - Dprintk("AP %d syncing TSC\n", smp_processor_id()); + unsigned long best_t0 = 0, best_t1 = ~0UL, best_tm = 0; + unsigned long tcenter, t0, t1, tm; + int i; - while (atomic_read(&tsc_flag) != 0) - cpu_relax(); - atomic_inc(&tsc_flag); - mb(); - _raw_spin_lock(&tsc_sync_lock); - wrmsrl(MSR_IA32_TSC, bp_tsc); - _raw_spin_unlock(&tsc_sync_lock); - rdtscll(ap_tsc); - mb(); - atomic_inc(&tsc_flag); - mb(); + for (i = 0; i < NUM_ITERS; ++i) { + rdtscll(t0); + go[MASTER] = 1; + while (!(tm = go[SLAVE])) + no_cpu_relax(); + go[SLAVE] = 0; + rdtscll(t1); + + if (t1 - t0 < best_t1 - best_t0) + best_t0 = t0, best_t1 = t1, best_tm = tm; + } + + *rt = best_t1 - best_t0; + *master = best_tm - best_t0; + + /* average best_t0 and best_t1 without overflow: */ + tcenter = (best_t0/2 + best_t1/2); + if (best_t0 % 2 + best_t1 % 2 == 2) + ++tcenter; + return tcenter - best_tm; } -static void __cpuinit sync_tsc_ap(void) +static __cpuinit void sync_tsc(void) { - int i; - for (i = 0; i < NR_LOOPS; i++) - __sync_tsc_ap(); + int i, done = 0; + long delta, adj, adjust_latency = 0; + unsigned long flags, rt, master_time_stamp, bound; +#if DEBUG_TSC_SYNC + static struct syncdebug { + long rt; /* roundtrip time */ + long master; /* master's timestamp */ + long diff; /* difference between midpoint and master's timestamp */ + long lat; /* estimate of tsc adjustment latency */ + } t[NUM_ROUNDS] __cpuinitdata; +#endif + + go[MASTER] = 1; + + smp_call_function(sync_master, NULL, 1, 0); + + while (go[MASTER]) /* wait for master to be ready */ + no_cpu_relax(); + + spin_lock_irqsave(&tsc_sync_lock, flags); + { + for (i = 0; i < NUM_ROUNDS; ++i) { + delta = get_delta(&rt, &master_time_stamp); + if (delta == 0) { + done = 1; /* let's lock on to this... */ + bound = rt; + } + + if (!done) { + unsigned long t; + if (i > 0) { + adjust_latency += -delta; + adj = -delta + adjust_latency/4; + } else + adj = -delta; + + rdtscll(t); + wrmsrl(MSR_IA32_TSC, t + adj); + } +#if DEBUG_TSC_SYNC + t[i].rt = rt; + t[i].master = master_time_stamp; + t[i].diff = delta; + t[i].lat = adjust_latency/4; +#endif + } + } + spin_unlock_irqrestore(&tsc_sync_lock, flags); + +#if DEBUG_TSC_SYNC + for (i = 0; i < NUM_ROUNDS; ++i) + printk("rt=%5ld master=%5ld diff=%5ld adjlat=%5ld\n", + t[i].rt, t[i].master, t[i].diff, t[i].lat); +#endif + + printk(KERN_INFO + "CPU %d: synchronized TSC with CPU %u (last diff %ld cycles, " + "maxerr %lu cycles)\n", + smp_processor_id(), boot_cpu_id, delta, rt); } -/* - * Synchronize TSC from BP to AP. - */ -static void __cpuinit __sync_tsc_bp(int cpu) +static void __cpuinit tsc_sync_wait(void) { - if (!cpu_has_tsc) + if (notscsync || !cpu_has_tsc) return; - - /* Wait for AP */ - while (atomic_read(&tsc_flag) == 0) - cpu_relax(); - /* Save BPs TSC */ - sync_core(); - rdtscll(bp_tsc); - /* Don't do the sync core here to avoid too much latency. */ - mb(); - /* Start the AP */ - _raw_spin_unlock(&tsc_sync_lock); - /* Wait for AP again */ - while (atomic_read(&tsc_flag) < 2) - cpu_relax(); - rdtscl(bp_tsc); - barrier(); + printk(KERN_INFO "CPU %d: Syncing TSC to CPU %u.\n", smp_processor_id(), + boot_cpu_id); + sync_tsc(); } -static void __cpuinit sync_tsc_bp(int cpu) +static __init int notscsync_setup(char *s) { - int i; - for (i = 0; i < NR_LOOPS - 1; i++) { - __sync_tsc_bp(cpu); - sync_tsc_bp_init(1); - } - __sync_tsc_bp(cpu); - printk(KERN_INFO "Synced TSC of CPU %d difference %Ld\n", - cpu, ap_tsc - bp_tsc); + notscsync = 1; + return 0; } +__setup("notscsync", notscsync_setup); static atomic_t init_deasserted __cpuinitdata; @@ -302,6 +445,33 @@ void __cpuinit smp_callin(void) cpu_set(cpuid, cpu_callin_map); } +static inline void set_cpu_sibling_map(int cpu) +{ + int i; + + if (smp_num_siblings > 1) { + for_each_cpu(i) { + if (cpu_core_id[cpu] == cpu_core_id[i]) { + cpu_set(i, cpu_sibling_map[cpu]); + cpu_set(cpu, cpu_sibling_map[i]); + } + } + } else { + cpu_set(cpu, cpu_sibling_map[cpu]); + } + + if (current_cpu_data.x86_num_cores > 1) { + for_each_cpu(i) { + if (phys_proc_id[cpu] == phys_proc_id[i]) { + cpu_set(i, cpu_core_map[cpu]); + cpu_set(cpu, cpu_core_map[i]); + } + } + } else { + cpu_core_map[cpu] = cpu_sibling_map[cpu]; + } +} + /* * Setup code on secondary processor (after comming out of the trampoline) */ @@ -315,11 +485,6 @@ void __cpuinit start_secondary(void) cpu_init(); smp_callin(); - /* - * Synchronize the TSC with the BP - */ - sync_tsc_ap(); - /* otherwise gcc will move up the smp_processor_id before the cpu_init */ barrier(); @@ -334,15 +499,38 @@ void __cpuinit start_secondary(void) enable_8259A_irq(0); } - enable_APIC_timer(); /* + * The sibling maps must be set before turing the online map on for + * this cpu + */ + set_cpu_sibling_map(smp_processor_id()); + + /* + * We need to hold call_lock, so there is no inconsistency + * between the time smp_call_function() determines number of + * IPI receipients, and the time when the determination is made + * for which cpus receive the IPI in genapic_flat.c. Holding this + * lock helps us to not include this cpu in a currently in progress + * smp_call_function(). + */ + lock_ipi_call_lock(); + + /* * Allow the master to continue. */ cpu_set(smp_processor_id(), cpu_online_map); + per_cpu(cpu_state, smp_processor_id()) = CPU_ONLINE; + unlock_ipi_call_lock(); + mb(); + /* Wait for TSC sync to not schedule things before. + We still process interrupts, which could see an inconsistent + time in that window unfortunately. */ + tsc_sync_wait(); + cpu_idle(); } @@ -513,34 +701,77 @@ static int __cpuinit wakeup_secondary_via_INIT(int phys_apicid, unsigned int sta return (send_status | accept_status); } +struct create_idle { + struct task_struct *idle; + struct completion done; + int cpu; +}; + +void do_fork_idle(void *_c_idle) +{ + struct create_idle *c_idle = _c_idle; + + c_idle->idle = fork_idle(c_idle->cpu); + complete(&c_idle->done); +} + /* * Boot one CPU. */ static int __cpuinit do_boot_cpu(int cpu, int apicid) { - struct task_struct *idle; unsigned long boot_error; int timeout; unsigned long start_rip; + struct create_idle c_idle = { + .cpu = cpu, + .done = COMPLETION_INITIALIZER(c_idle.done), + }; + DECLARE_WORK(work, do_fork_idle, &c_idle); + + c_idle.idle = get_idle_for_cpu(cpu); + + if (c_idle.idle) { + c_idle.idle->thread.rsp = (unsigned long) (((struct pt_regs *) + (THREAD_SIZE + (unsigned long) c_idle.idle->thread_info)) - 1); + init_idle(c_idle.idle, cpu); + goto do_rest; + } + /* - * We can't use kernel_thread since we must avoid to - * reschedule the child. + * During cold boot process, keventd thread is not spun up yet. + * When we do cpu hot-add, we create idle threads on the fly, we should + * not acquire any attributes from the calling context. Hence the clean + * way to create kernel_threads() is to do that from keventd(). + * We do the current_is_keventd() due to the fact that ACPI notifier + * was also queuing to keventd() and when the caller is already running + * in context of keventd(), we would end up with locking up the keventd + * thread. */ - idle = fork_idle(cpu); - if (IS_ERR(idle)) { + if (!keventd_up() || current_is_keventd()) + work.func(work.data); + else { + schedule_work(&work); + wait_for_completion(&c_idle.done); + } + + if (IS_ERR(c_idle.idle)) { printk("failed fork for CPU %d\n", cpu); - return PTR_ERR(idle); + return PTR_ERR(c_idle.idle); } - x86_cpu_to_apicid[cpu] = apicid; - cpu_pda[cpu].pcurrent = idle; + set_idle_for_cpu(cpu, c_idle.idle); + +do_rest: + + cpu_pda[cpu].pcurrent = c_idle.idle; start_rip = setup_trampoline(); - init_rsp = idle->thread.rsp; + init_rsp = c_idle.idle->thread.rsp; per_cpu(init_tss,cpu).rsp0 = init_rsp; initial_code = start_secondary; - clear_ti_thread_flag(idle->thread_info, TIF_FORK); + clear_ti_thread_flag(c_idle.idle->thread_info, TIF_FORK); printk(KERN_INFO "Booting processor %d/%d rip %lx rsp %lx\n", cpu, apicid, start_rip, init_rsp); @@ -600,8 +831,6 @@ static int __cpuinit do_boot_cpu(int cpu, int apicid) if (cpu_isset(cpu, cpu_callin_map)) { /* number CPUs logically, starting from 1 (BSP is 0) */ - Dprintk("OK.\n"); - print_cpu_info(&cpu_data[cpu]); Dprintk("CPU has booted.\n"); } else { boot_error = 1; @@ -634,51 +863,6 @@ cycles_t cacheflush_time; unsigned long cache_decay_ticks; /* - * Construct cpu_sibling_map[], so that we can tell the sibling CPU - * on SMT systems efficiently. - */ -static __cpuinit void detect_siblings(void) -{ - int cpu; - - for (cpu = 0; cpu < NR_CPUS; cpu++) { - cpus_clear(cpu_sibling_map[cpu]); - cpus_clear(cpu_core_map[cpu]); - } - - for_each_online_cpu (cpu) { - struct cpuinfo_x86 *c = cpu_data + cpu; - int siblings = 0; - int i; - if (smp_num_siblings > 1) { - for_each_online_cpu (i) { - if (cpu_core_id[cpu] == cpu_core_id[i]) { - siblings++; - cpu_set(i, cpu_sibling_map[cpu]); - } - } - } else { - siblings++; - cpu_set(cpu, cpu_sibling_map[cpu]); - } - - if (siblings != smp_num_siblings) { - printk(KERN_WARNING - "WARNING: %d siblings found for CPU%d, should be %d\n", - siblings, cpu, smp_num_siblings); - smp_num_siblings = siblings; - } - if (c->x86_num_cores > 1) { - for_each_online_cpu(i) { - if (phys_proc_id[cpu] == phys_proc_id[i]) - cpu_set(i, cpu_core_map[cpu]); - } - } else - cpu_core_map[cpu] = cpu_sibling_map[cpu]; - } -} - -/* * Cleanup possible dangling ends... */ static __cpuinit void smp_cleanup_boot(void) @@ -711,7 +895,7 @@ static __cpuinit void smp_cleanup_boot(void) * * RED-PEN audit/test this more. I bet there is more state messed up here. */ -static __cpuinit void disable_smp(void) +static __init void disable_smp(void) { cpu_present_map = cpumask_of_cpu(0); cpu_possible_map = cpumask_of_cpu(0); @@ -726,7 +910,7 @@ static __cpuinit void disable_smp(void) /* * Handle user cpus=... parameter. */ -static __cpuinit void enforce_max_cpus(unsigned max_cpus) +static __init void enforce_max_cpus(unsigned max_cpus) { int i, k; k = 0; @@ -743,7 +927,7 @@ static __cpuinit void enforce_max_cpus(unsigned max_cpus) /* * Various sanity checks. */ -static int __cpuinit smp_sanity_check(unsigned max_cpus) +static int __init smp_sanity_check(unsigned max_cpus) { if (!physid_isset(hard_smp_processor_id(), phys_cpu_present_map)) { printk("weird, boot CPU (#%d) not listed by the BIOS.\n", @@ -801,7 +985,7 @@ static int __cpuinit smp_sanity_check(unsigned max_cpus) * Prepare for SMP bootup. The MP table or ACPI has been read * earlier. Just do some sanity checking here and enable APIC mode. */ -void __cpuinit smp_prepare_cpus(unsigned int max_cpus) +void __init smp_prepare_cpus(unsigned int max_cpus) { int i; @@ -818,10 +1002,9 @@ void __cpuinit smp_prepare_cpus(unsigned int max_cpus) int apicid = cpu_present_to_apicid(i); if (physid_isset(apicid, phys_cpu_present_map)) { cpu_set(i, cpu_present_map); - /* possible map would be different if we supported real - CPU hotplug. */ cpu_set(i, cpu_possible_map); } + fixup_cpu_possible_map(i); } if (smp_sanity_check(max_cpus) < 0) { @@ -842,7 +1025,6 @@ void __cpuinit smp_prepare_cpus(unsigned int max_cpus) GET_APIC_ID(apic_read(APIC_ID)), boot_cpu_id); /* Or can we switch back to PIC here? */ } - x86_cpu_to_apicid[0] = boot_cpu_id; /* * Now start the IO-APICs @@ -867,13 +1049,13 @@ void __init smp_prepare_boot_cpu(void) int me = smp_processor_id(); cpu_set(me, cpu_online_map); cpu_set(me, cpu_callout_map); + cpu_set(0, cpu_sibling_map[0]); + cpu_set(0, cpu_core_map[0]); + per_cpu(cpu_state, me) = CPU_ONLINE; } /* * Entry point to boot a CPU. - * - * This is all __cpuinit, not __devinit for now because we don't support - * CPU hotplug (yet). */ int __cpuinit __cpu_up(unsigned int cpu) { @@ -889,38 +1071,140 @@ int __cpuinit __cpu_up(unsigned int cpu) printk("__cpu_up: bad cpu %d\n", cpu); return -EINVAL; } - sync_tsc_bp_init(1); + /* + * Already booted CPU? + */ + if (cpu_isset(cpu, cpu_callin_map)) { + Dprintk("do_boot_cpu %d Already started\n", cpu); + return -ENOSYS; + } + + per_cpu(cpu_state, cpu) = CPU_UP_PREPARE; /* Boot it! */ err = do_boot_cpu(cpu, apicid); if (err < 0) { - sync_tsc_bp_init(0); Dprintk("do_boot_cpu failed %d\n", err); return err; } - sync_tsc_bp(cpu); - /* Unleash the CPU! */ Dprintk("waiting for cpu %d\n", cpu); while (!cpu_isset(cpu, cpu_online_map)) cpu_relax(); - return 0; + err = 0; + + return err; } /* * Finish the SMP boot. */ -void __cpuinit smp_cpus_done(unsigned int max_cpus) +void __init smp_cpus_done(unsigned int max_cpus) { +#ifndef CONFIG_HOTPLUG_CPU zap_low_mappings(); +#endif smp_cleanup_boot(); #ifdef CONFIG_X86_IO_APIC setup_ioapic_dest(); #endif - detect_siblings(); time_init_gtod(); + + check_nmi_watchdog(); +} + +#ifdef CONFIG_HOTPLUG_CPU + +static void remove_siblinginfo(int cpu) +{ + int sibling; + + for_each_cpu_mask(sibling, cpu_sibling_map[cpu]) + cpu_clear(cpu, cpu_sibling_map[sibling]); + for_each_cpu_mask(sibling, cpu_core_map[cpu]) + cpu_clear(cpu, cpu_core_map[sibling]); + cpus_clear(cpu_sibling_map[cpu]); + cpus_clear(cpu_core_map[cpu]); + phys_proc_id[cpu] = BAD_APICID; + cpu_core_id[cpu] = BAD_APICID; +} + +void remove_cpu_from_maps(void) +{ + int cpu = smp_processor_id(); + + cpu_clear(cpu, cpu_callout_map); + cpu_clear(cpu, cpu_callin_map); + clear_bit(cpu, &cpu_initialized); /* was set by cpu_init() */ +} + +int __cpu_disable(void) +{ + int cpu = smp_processor_id(); + + /* + * Perhaps use cpufreq to drop frequency, but that could go + * into generic code. + * + * We won't take down the boot processor on i386 due to some + * interrupts only being able to be serviced by the BSP. + * Especially so if we're not using an IOAPIC -zwane + */ + if (cpu == 0) + return -EBUSY; + + disable_APIC_timer(); + + /* + * HACK: + * Allow any queued timer interrupts to get serviced + * This is only a temporary solution until we cleanup + * fixup_irqs as we do for IA64. + */ + local_irq_enable(); + mdelay(1); + + local_irq_disable(); + remove_siblinginfo(cpu); + + /* It's now safe to remove this processor from the online map */ + cpu_clear(cpu, cpu_online_map); + remove_cpu_from_maps(); + fixup_irqs(cpu_online_map); + return 0; +} + +void __cpu_die(unsigned int cpu) +{ + /* We don't do anything here: idle task is faking death itself. */ + unsigned int i; + + for (i = 0; i < 10; i++) { + /* They ack this in play_dead by setting CPU_DEAD */ + if (per_cpu(cpu_state, cpu) == CPU_DEAD) { + printk ("CPU %d is now offline\n", cpu); + return; + } + current->state = TASK_UNINTERRUPTIBLE; + schedule_timeout(HZ/10); + } + printk(KERN_ERR "CPU %u didn't die...\n", cpu); +} + +#else /* ... !CONFIG_HOTPLUG_CPU */ + +int __cpu_disable(void) +{ + return -ENOSYS; +} + +void __cpu_die(unsigned int cpu) +{ + /* We said "no" in __cpu_disable */ + BUG(); } +#endif /* CONFIG_HOTPLUG_CPU */ diff --git a/arch/x86_64/kernel/suspend.c b/arch/x86_64/kernel/suspend.c index ebaa1e37d65..6c0f402e3a8 100644 --- a/arch/x86_64/kernel/suspend.c +++ b/arch/x86_64/kernel/suspend.c @@ -44,7 +44,6 @@ void __save_processor_state(struct saved_context *ctxt) */ asm volatile ("sgdt %0" : "=m" (ctxt->gdt_limit)); asm volatile ("sidt %0" : "=m" (ctxt->idt_limit)); - asm volatile ("sldt %0" : "=m" (ctxt->ldt)); asm volatile ("str %0" : "=m" (ctxt->tr)); /* XMM0..XMM15 should be handled by kernel_fpu_begin(). */ @@ -69,6 +68,7 @@ void __save_processor_state(struct saved_context *ctxt) asm volatile ("movq %%cr2, %0" : "=r" (ctxt->cr2)); asm volatile ("movq %%cr3, %0" : "=r" (ctxt->cr3)); asm volatile ("movq %%cr4, %0" : "=r" (ctxt->cr4)); + asm volatile ("movq %%cr8, %0" : "=r" (ctxt->cr8)); } void save_processor_state(void) @@ -90,12 +90,20 @@ void __restore_processor_state(struct saved_context *ctxt) /* * control registers */ + asm volatile ("movq %0, %%cr8" :: "r" (ctxt->cr8)); asm volatile ("movq %0, %%cr4" :: "r" (ctxt->cr4)); asm volatile ("movq %0, %%cr3" :: "r" (ctxt->cr3)); asm volatile ("movq %0, %%cr2" :: "r" (ctxt->cr2)); asm volatile ("movq %0, %%cr0" :: "r" (ctxt->cr0)); /* + * now restore the descriptor tables to their proper values + * ltr is done i fix_processor_context(). + */ + asm volatile ("lgdt %0" :: "m" (ctxt->gdt_limit)); + asm volatile ("lidt %0" :: "m" (ctxt->idt_limit)); + + /* * segment registers */ asm volatile ("movw %0, %%ds" :: "r" (ctxt->ds)); @@ -108,14 +116,6 @@ void __restore_processor_state(struct saved_context *ctxt) wrmsrl(MSR_GS_BASE, ctxt->gs_base); wrmsrl(MSR_KERNEL_GS_BASE, ctxt->gs_kernel_base); - /* - * now restore the descriptor tables to their proper values - * ltr is done i fix_processor_context(). - */ - asm volatile ("lgdt %0" :: "m" (ctxt->gdt_limit)); - asm volatile ("lidt %0" :: "m" (ctxt->idt_limit)); - asm volatile ("lldt %0" :: "m" (ctxt->ldt)); - fix_processor_context(); do_fpu_end(); diff --git a/arch/x86_64/kernel/sys_x86_64.c b/arch/x86_64/kernel/sys_x86_64.c index dbebd5ccba6..cc7821c6885 100644 --- a/arch/x86_64/kernel/sys_x86_64.c +++ b/arch/x86_64/kernel/sys_x86_64.c @@ -68,13 +68,7 @@ out: static void find_start_end(unsigned long flags, unsigned long *begin, unsigned long *end) { -#ifdef CONFIG_IA32_EMULATION - if (test_thread_flag(TIF_IA32)) { - *begin = TASK_UNMAPPED_32; - *end = IA32_PAGE_OFFSET; - } else -#endif - if (flags & MAP_32BIT) { + if (!test_thread_flag(TIF_IA32) && (flags & MAP_32BIT)) { /* This is usually used needed to map code in small model, so it needs to be in the first 31bit. Limit it to that. This means we need to move the @@ -84,10 +78,10 @@ static void find_start_end(unsigned long flags, unsigned long *begin, of playground for now. -AK */ *begin = 0x40000000; *end = 0x80000000; - } else { - *begin = TASK_UNMAPPED_64; + } else { + *begin = TASK_UNMAPPED_BASE; *end = TASK_SIZE; - } + } } unsigned long @@ -111,6 +105,11 @@ arch_get_unmapped_area(struct file *filp, unsigned long addr, (!vma || addr + len <= vma->vm_start)) return addr; } + if (((flags & MAP_32BIT) || test_thread_flag(TIF_IA32)) + && len <= mm->cached_hole_size) { + mm->cached_hole_size = 0; + mm->free_area_cache = begin; + } addr = mm->free_area_cache; if (addr < begin) addr = begin; @@ -126,6 +125,7 @@ full_search: */ if (start_addr != begin) { start_addr = addr = begin; + mm->cached_hole_size = 0; goto full_search; } return -ENOMEM; @@ -137,6 +137,9 @@ full_search: mm->free_area_cache = addr + len; return addr; } + if (addr + mm->cached_hole_size < vma->vm_start) + mm->cached_hole_size = vma->vm_start - addr; + addr = vma->vm_end; } } diff --git a/arch/x86_64/kernel/time.c b/arch/x86_64/kernel/time.c index 0e3b60e1ac5..66bf6ddeb0c 100644 --- a/arch/x86_64/kernel/time.c +++ b/arch/x86_64/kernel/time.c @@ -26,6 +26,10 @@ #include <linux/sysdev.h> #include <linux/bcd.h> #include <linux/kallsyms.h> +#include <linux/acpi.h> +#ifdef CONFIG_ACPI +#include <acpi/achware.h> /* for PM timer frequency */ +#endif #include <asm/8253pit.h> #include <asm/pgtable.h> #include <asm/vsyscall.h> @@ -60,6 +64,7 @@ static int notsc __initdata = 0; unsigned int cpu_khz; /* TSC clocks / usec, not used here */ static unsigned long hpet_period; /* fsecs / HPET clock */ unsigned long hpet_tick; /* HPET clocks / interrupt */ +static int hpet_use_timer; unsigned long vxtime_hz = PIT_TICK_RATE; int report_lost_ticks; /* command line option */ unsigned long long monotonic_base; @@ -101,7 +106,9 @@ static inline unsigned int do_gettimeoffset_tsc(void) static inline unsigned int do_gettimeoffset_hpet(void) { - return ((hpet_readl(HPET_COUNTER) - vxtime.last) * vxtime.quot) >> 32; + /* cap counter read to one tick to avoid inconsistencies */ + unsigned long counter = hpet_readl(HPET_COUNTER) - vxtime.last; + return (min(counter,hpet_tick) * vxtime.quot) >> 32; } unsigned int (*do_gettimeoffset)(void) = do_gettimeoffset_tsc; @@ -297,7 +304,7 @@ unsigned long long monotonic_clock(void) last_offset = vxtime.last; base = monotonic_base; - this_offset = hpet_readl(HPET_T0_CMP) - hpet_tick; + this_offset = hpet_readl(HPET_COUNTER); } while (read_seqretry(&xtime_lock, seq)); offset = (this_offset - last_offset); @@ -373,7 +380,14 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs) write_seqlock(&xtime_lock); - if (vxtime.hpet_address) { + if (vxtime.hpet_address) + offset = hpet_readl(HPET_COUNTER); + + if (hpet_use_timer) { + /* if we're using the hpet timer functionality, + * we can more accurately know the counter value + * when the timer interrupt occured. + */ offset = hpet_readl(HPET_T0_CMP) - hpet_tick; delay = hpet_readl(HPET_COUNTER) - offset; } else { @@ -396,6 +410,10 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs) (offset - vxtime.last)*(NSEC_PER_SEC/HZ) / hpet_tick; vxtime.last = offset; +#ifdef CONFIG_X86_PM_TIMER + } else if (vxtime.mode == VXTIME_PMTMR) { + lost = pmtimer_mark_offset(); +#endif } else { offset = (((tsc - vxtime.last_tsc) * vxtime.tsc_quot) >> 32) - (USEC_PER_SEC / HZ); @@ -795,17 +813,18 @@ static int hpet_timer_stop_set_go(unsigned long tick) * Set up timer 0, as periodic with first interrupt to happen at hpet_tick, * and period also hpet_tick. */ - - hpet_writel(HPET_TN_ENABLE | HPET_TN_PERIODIC | HPET_TN_SETVAL | + if (hpet_use_timer) { + hpet_writel(HPET_TN_ENABLE | HPET_TN_PERIODIC | HPET_TN_SETVAL | HPET_TN_32BIT, HPET_T0_CFG); - hpet_writel(hpet_tick, HPET_T0_CMP); - hpet_writel(hpet_tick, HPET_T0_CMP); /* AK: why twice? */ - + hpet_writel(hpet_tick, HPET_T0_CMP); + hpet_writel(hpet_tick, HPET_T0_CMP); /* AK: why twice? */ + cfg |= HPET_CFG_LEGACY; + } /* * Go! */ - cfg |= HPET_CFG_ENABLE | HPET_CFG_LEGACY; + cfg |= HPET_CFG_ENABLE; hpet_writel(cfg, HPET_CFG); return 0; @@ -826,8 +845,7 @@ static int hpet_init(void) id = hpet_readl(HPET_ID); - if (!(id & HPET_ID_VENDOR) || !(id & HPET_ID_NUMBER) || - !(id & HPET_ID_LEGSUP)) + if (!(id & HPET_ID_VENDOR) || !(id & HPET_ID_NUMBER)) return -1; hpet_period = hpet_readl(HPET_PERIOD); @@ -837,6 +855,8 @@ static int hpet_init(void) hpet_tick = (1000000000L * (USEC_PER_SEC / HZ) + hpet_period / 2) / hpet_period; + hpet_use_timer = (id & HPET_ID_LEGSUP); + return hpet_timer_stop_set_go(hpet_tick); } @@ -893,11 +913,20 @@ void __init time_init(void) set_normalized_timespec(&wall_to_monotonic, -xtime.tv_sec, -xtime.tv_nsec); - if (!hpet_init()) { + if (!hpet_init()) vxtime_hz = (1000000000000000L + hpet_period / 2) / hpet_period; + + if (hpet_use_timer) { cpu_khz = hpet_calibrate_tsc(); timename = "HPET"; +#ifdef CONFIG_X86_PM_TIMER + } else if (pmtmr_ioport) { + vxtime_hz = PM_TIMER_FREQUENCY; + timename = "PM"; + pit_init(); + cpu_khz = pit_calibrate_tsc(); +#endif } else { pit_init(); cpu_khz = pit_calibrate_tsc(); @@ -923,37 +952,52 @@ void __init time_init(void) } /* + * Make an educated guess if the TSC is trustworthy and synchronized + * over all CPUs. + */ +static __init int unsynchronized_tsc(void) +{ +#ifdef CONFIG_SMP + if (oem_force_hpet_timer()) + return 1; + /* Intel systems are normally all synchronized. Exceptions + are handled in the OEM check above. */ + if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) + return 0; + /* All in a single socket - should be synchronized */ + if (cpus_weight(cpu_core_map[0]) == num_online_cpus()) + return 0; +#endif + /* Assume multi socket systems are not synchronized */ + return num_online_cpus() > 1; +} + +/* * Decide after all CPUs are booted what mode gettimeofday should use. */ void __init time_init_gtod(void) { char *timetype; - /* - * AMD systems with more than one CPU don't have fully synchronized - * TSCs. Always use HPET gettimeofday for these, although it is slower. - * Intel SMP systems usually have synchronized TSCs, so use always - * the TSC. - * - * Exceptions: - * IBM Summit2 checked by oem_force_hpet_timer(). - * AMD dual core may also not need HPET. Check me. - * - * Can be turned off with "notsc". - */ - if (num_online_cpus() > 1 && - boot_cpu_data.x86_vendor == X86_VENDOR_AMD) - notsc = 1; - /* Some systems will want to disable TSC and use HPET. */ - if (oem_force_hpet_timer()) + if (unsynchronized_tsc()) notsc = 1; if (vxtime.hpet_address && notsc) { - timetype = "HPET"; + timetype = hpet_use_timer ? "HPET" : "PIT/HPET"; vxtime.last = hpet_readl(HPET_T0_CMP) - hpet_tick; vxtime.mode = VXTIME_HPET; do_gettimeoffset = do_gettimeoffset_hpet; +#ifdef CONFIG_X86_PM_TIMER + /* Using PM for gettimeofday is quite slow, but we have no other + choice because the TSC is too unreliable on some systems. */ + } else if (pmtmr_ioport && !vxtime.hpet_address && notsc) { + timetype = "PM"; + do_gettimeoffset = do_gettimeoffset_pm; + vxtime.mode = VXTIME_PMTMR; + sysctl_vsyscall = 0; + printk(KERN_INFO "Disabling vsyscall due to use of PM timer\n"); +#endif } else { - timetype = vxtime.hpet_address ? "HPET/TSC" : "PIT/TSC"; + timetype = hpet_use_timer ? "HPET/TSC" : "PIT/TSC"; vxtime.mode = VXTIME_TSC; } diff --git a/arch/x86_64/kernel/traps.c b/arch/x86_64/kernel/traps.c index 65a37f52c56..10273663000 100644 --- a/arch/x86_64/kernel/traps.c +++ b/arch/x86_64/kernel/traps.c @@ -274,7 +274,7 @@ EXPORT_SYMBOL(dump_stack); void show_registers(struct pt_regs *regs) { int i; - int in_kernel = (regs->cs & 3) == 0; + int in_kernel = !user_mode(regs); unsigned long rsp; const int cpu = safe_smp_processor_id(); struct task_struct *cur = cpu_pda[cpu].pcurrent; @@ -318,7 +318,7 @@ void handle_BUG(struct pt_regs *regs) struct bug_frame f; char tmp; - if (regs->cs & 3) + if (user_mode(regs)) return; if (__copy_from_user(&f, (struct bug_frame *) regs->rip, sizeof(struct bug_frame))) @@ -332,10 +332,12 @@ void handle_BUG(struct pt_regs *regs) printk(KERN_ALERT "Kernel BUG at %.50s:%d\n", f.filename, f.line); } +#ifdef CONFIG_BUG void out_of_line_bug(void) { BUG(); } +#endif static DEFINE_SPINLOCK(die_lock); static int die_owner = -1; @@ -435,7 +437,7 @@ static void do_trap(int trapnr, int signr, char *str, } #endif - if ((regs->cs & 3) != 0) { + if (user_mode(regs)) { struct task_struct *tsk = current; if (exception_trace && unhandled_signal(tsk, signr)) @@ -520,7 +522,7 @@ asmlinkage void do_general_protection(struct pt_regs * regs, long error_code) } #endif - if ((regs->cs & 3)!=0) { + if (user_mode(regs)) { struct task_struct *tsk = current; if (exception_trace && unhandled_signal(tsk, SIGSEGV)) @@ -584,11 +586,17 @@ static void unknown_nmi_error(unsigned char reason, struct pt_regs * regs) asmlinkage void default_do_nmi(struct pt_regs *regs) { unsigned char reason = 0; + int cpu; + + cpu = smp_processor_id(); /* Only the BSP gets external NMIs from the system. */ - if (!smp_processor_id()) + if (!cpu) reason = get_nmi_reason(); + if (!cpu_online(cpu)) + return; + if (!(reason & 0xc0)) { if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 0, SIGINT) == NOTIFY_STOP) @@ -636,7 +644,7 @@ asmlinkage struct pt_regs *sync_regs(struct pt_regs *eregs) if (eregs == (struct pt_regs *)eregs->rsp) ; /* Exception from user space */ - else if (eregs->cs & 3) + else if (user_mode(eregs)) regs = ((struct pt_regs *)current->thread.rsp0) - 1; /* Exception from kernel and interrupts are enabled. Move to kernel process stack. */ @@ -667,7 +675,7 @@ asmlinkage void do_debug(struct pt_regs * regs, unsigned long error_code) } #endif - asm("movq %%db6,%0" : "=r" (condition)); + get_debugreg(condition, 6); if (notify_die(DIE_DEBUG, "debug", regs, condition, error_code, SIGTRAP) == NOTIFY_STOP) @@ -695,7 +703,7 @@ asmlinkage void do_debug(struct pt_regs * regs, unsigned long error_code) * allowing programs to debug themselves without the ptrace() * interface. */ - if ((regs->cs & 3) == 0) + if (!user_mode(regs)) goto clear_TF_reenable; /* * Was the TF flag set by a debugger? If so, clear it now, @@ -713,13 +721,13 @@ asmlinkage void do_debug(struct pt_regs * regs, unsigned long error_code) info.si_signo = SIGTRAP; info.si_errno = 0; info.si_code = TRAP_BRKPT; - if ((regs->cs & 3) == 0) + if (!user_mode(regs)) goto clear_dr7; info.si_addr = (void __user *)regs->rip; force_sig_info(SIGTRAP, &info, tsk); clear_dr7: - asm volatile("movq %0,%%db7"::"r"(0UL)); + set_debugreg(0UL, 7); return; clear_TF_reenable: @@ -754,7 +762,7 @@ asmlinkage void do_coprocessor_error(struct pt_regs *regs) unsigned short cwd, swd; conditional_sti(regs); - if ((regs->cs & 3) == 0 && + if (!user_mode(regs) && kernel_math_error(regs, "kernel x87 math error")) return; @@ -820,7 +828,7 @@ asmlinkage void do_simd_coprocessor_error(struct pt_regs *regs) unsigned short mxcsr; conditional_sti(regs); - if ((regs->cs & 3) == 0 && + if (!user_mode(regs) && kernel_math_error(regs, "kernel simd math error")) return; diff --git a/arch/x86_64/kernel/vmlinux.lds.S b/arch/x86_64/kernel/vmlinux.lds.S index 59ebd5beda8..73389f51c4e 100644 --- a/arch/x86_64/kernel/vmlinux.lds.S +++ b/arch/x86_64/kernel/vmlinux.lds.S @@ -2,7 +2,10 @@ * Written by Martin Mares <mj@atrey.karlin.mff.cuni.cz>; */ +#define LOAD_OFFSET __START_KERNEL_map + #include <asm-generic/vmlinux.lds.h> +#include <asm/page.h> #include <linux/config.h> OUTPUT_FORMAT("elf64-x86-64", "elf64-x86-64", "elf64-x86-64") @@ -11,28 +14,30 @@ ENTRY(phys_startup_64) jiffies_64 = jiffies; SECTIONS { - . = 0xffffffff80100000; + . = __START_KERNEL; phys_startup_64 = startup_64 - LOAD_OFFSET; _text = .; /* Text and read-only data */ - .text : { + .text : AT(ADDR(.text) - LOAD_OFFSET) { *(.text) SCHED_TEXT LOCK_TEXT *(.fixup) *(.gnu.warning) } = 0x9090 - .text.lock : { *(.text.lock) } /* out-of-line lock text */ + /* out-of-line lock text */ + .text.lock : AT(ADDR(.text.lock) - LOAD_OFFSET) { *(.text.lock) } _etext = .; /* End of text section */ . = ALIGN(16); /* Exception table */ __start___ex_table = .; - __ex_table : { *(__ex_table) } + __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) { *(__ex_table) } __stop___ex_table = .; RODATA - .data : { /* Data */ + /* Data */ + .data : AT(ADDR(.data) - LOAD_OFFSET) { *(.data) CONSTRUCTORS } @@ -40,62 +45,95 @@ SECTIONS _edata = .; /* End of data section */ __bss_start = .; /* BSS */ - .bss : { + .bss : AT(ADDR(.bss) - LOAD_OFFSET) { *(.bss.page_aligned) *(.bss) } __bss_end = .; + . = ALIGN(PAGE_SIZE); . = ALIGN(CONFIG_X86_L1_CACHE_BYTES); - .data.cacheline_aligned : { *(.data.cacheline_aligned) } + .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) { + *(.data.cacheline_aligned) + } + +#define VSYSCALL_ADDR (-10*1024*1024) +#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data.cacheline_aligned) + SIZEOF(.data.cacheline_aligned) + 4095) & ~(4095)) +#define VSYSCALL_VIRT_ADDR ((ADDR(.data.cacheline_aligned) + SIZEOF(.data.cacheline_aligned) + 4095) & ~(4095)) + +#define VLOAD_OFFSET (VSYSCALL_ADDR - VSYSCALL_PHYS_ADDR) +#define VLOAD(x) (ADDR(x) - VLOAD_OFFSET) + +#define VVIRT_OFFSET (VSYSCALL_ADDR - VSYSCALL_VIRT_ADDR) +#define VVIRT(x) (ADDR(x) - VVIRT_OFFSET) -#define AFTER(x) BINALIGN(LOADADDR(x) + SIZEOF(x), 16) -#define BINALIGN(x,y) (((x) + (y) - 1) & ~((y) - 1)) -#define CACHE_ALIGN(x) BINALIGN(x, CONFIG_X86_L1_CACHE_BYTES) + . = VSYSCALL_ADDR; + .vsyscall_0 : AT(VSYSCALL_PHYS_ADDR) { *(.vsyscall_0) } + __vsyscall_0 = VSYSCALL_VIRT_ADDR; - .vsyscall_0 -10*1024*1024: AT ((LOADADDR(.data.cacheline_aligned) + SIZEOF(.data.cacheline_aligned) + 4095) & ~(4095)) { *(.vsyscall_0) } - __vsyscall_0 = LOADADDR(.vsyscall_0); . = ALIGN(CONFIG_X86_L1_CACHE_BYTES); - .xtime_lock : AT CACHE_ALIGN(AFTER(.vsyscall_0)) { *(.xtime_lock) } - xtime_lock = LOADADDR(.xtime_lock); - .vxtime : AT AFTER(.xtime_lock) { *(.vxtime) } - vxtime = LOADADDR(.vxtime); - .wall_jiffies : AT AFTER(.vxtime) { *(.wall_jiffies) } - wall_jiffies = LOADADDR(.wall_jiffies); - .sys_tz : AT AFTER(.wall_jiffies) { *(.sys_tz) } - sys_tz = LOADADDR(.sys_tz); - .sysctl_vsyscall : AT AFTER(.sys_tz) { *(.sysctl_vsyscall) } - sysctl_vsyscall = LOADADDR(.sysctl_vsyscall); - .xtime : AT AFTER(.sysctl_vsyscall) { *(.xtime) } - xtime = LOADADDR(.xtime); + .xtime_lock : AT(VLOAD(.xtime_lock)) { *(.xtime_lock) } + xtime_lock = VVIRT(.xtime_lock); + + .vxtime : AT(VLOAD(.vxtime)) { *(.vxtime) } + vxtime = VVIRT(.vxtime); + + .wall_jiffies : AT(VLOAD(.wall_jiffies)) { *(.wall_jiffies) } + wall_jiffies = VVIRT(.wall_jiffies); + + .sys_tz : AT(VLOAD(.sys_tz)) { *(.sys_tz) } + sys_tz = VVIRT(.sys_tz); + + .sysctl_vsyscall : AT(VLOAD(.sysctl_vsyscall)) { *(.sysctl_vsyscall) } + sysctl_vsyscall = VVIRT(.sysctl_vsyscall); + + .xtime : AT(VLOAD(.xtime)) { *(.xtime) } + xtime = VVIRT(.xtime); + . = ALIGN(CONFIG_X86_L1_CACHE_BYTES); - .jiffies : AT CACHE_ALIGN(AFTER(.xtime)) { *(.jiffies) } - jiffies = LOADADDR(.jiffies); - .vsyscall_1 ADDR(.vsyscall_0) + 1024: AT (LOADADDR(.vsyscall_0) + 1024) { *(.vsyscall_1) } - . = LOADADDR(.vsyscall_0) + 4096; + .jiffies : AT(VLOAD(.jiffies)) { *(.jiffies) } + jiffies = VVIRT(.jiffies); + + .vsyscall_1 ADDR(.vsyscall_0) + 1024: AT(VLOAD(.vsyscall_1)) { *(.vsyscall_1) } + .vsyscall_2 ADDR(.vsyscall_0) + 2048: AT(VLOAD(.vsyscall_2)) { *(.vsyscall_2) } + .vsyscall_3 ADDR(.vsyscall_0) + 3072: AT(VLOAD(.vsyscall_3)) { *(.vsyscall_3) } + + . = VSYSCALL_VIRT_ADDR + 4096; + +#undef VSYSCALL_ADDR +#undef VSYSCALL_PHYS_ADDR +#undef VSYSCALL_VIRT_ADDR +#undef VLOAD_OFFSET +#undef VLOAD +#undef VVIRT_OFFSET +#undef VVIRT . = ALIGN(8192); /* init_task */ - .data.init_task : { *(.data.init_task) } + .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) { + *(.data.init_task) + } . = ALIGN(4096); - .data.page_aligned : { *(.data.page_aligned) } + .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) { + *(.data.page_aligned) + } . = ALIGN(4096); /* Init code and data */ __init_begin = .; - .init.text : { + .init.text : AT(ADDR(.init.text) - LOAD_OFFSET) { _sinittext = .; *(.init.text) _einittext = .; } __initdata_begin = .; - .init.data : { *(.init.data) } + .init.data : AT(ADDR(.init.data) - LOAD_OFFSET) { *(.init.data) } __initdata_end = .; . = ALIGN(16); __setup_start = .; - .init.setup : { *(.init.setup) } + .init.setup : AT(ADDR(.init.setup) - LOAD_OFFSET) { *(.init.setup) } __setup_end = .; __initcall_start = .; - .initcall.init : { + .initcall.init : AT(ADDR(.initcall.init) - LOAD_OFFSET) { *(.initcall1.init) *(.initcall2.init) *(.initcall3.init) @@ -106,32 +144,38 @@ SECTIONS } __initcall_end = .; __con_initcall_start = .; - .con_initcall.init : { *(.con_initcall.init) } + .con_initcall.init : AT(ADDR(.con_initcall.init) - LOAD_OFFSET) { + *(.con_initcall.init) + } __con_initcall_end = .; SECURITY_INIT . = ALIGN(8); __alt_instructions = .; - .altinstructions : { *(.altinstructions) } + .altinstructions : AT(ADDR(.altinstructions) - LOAD_OFFSET) { + *(.altinstructions) + } __alt_instructions_end = .; - .altinstr_replacement : { *(.altinstr_replacement) } + .altinstr_replacement : AT(ADDR(.altinstr_replacement) - LOAD_OFFSET) { + *(.altinstr_replacement) + } /* .exit.text is discard at runtime, not link time, to deal with references from .altinstructions and .eh_frame */ - .exit.text : { *(.exit.text) } - .exit.data : { *(.exit.data) } + .exit.text : AT(ADDR(.exit.text) - LOAD_OFFSET) { *(.exit.text) } + .exit.data : AT(ADDR(.exit.data) - LOAD_OFFSET) { *(.exit.data) } . = ALIGN(4096); __initramfs_start = .; - .init.ramfs : { *(.init.ramfs) } + .init.ramfs : AT(ADDR(.init.ramfs) - LOAD_OFFSET) { *(.init.ramfs) } __initramfs_end = .; . = ALIGN(32); __per_cpu_start = .; - .data.percpu : { *(.data.percpu) } + .data.percpu : AT(ADDR(.data.percpu) - LOAD_OFFSET) { *(.data.percpu) } __per_cpu_end = .; . = ALIGN(4096); __init_end = .; . = ALIGN(4096); __nosave_begin = .; - .data_nosave : { *(.data.nosave) } + .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) { *(.data.nosave) } . = ALIGN(4096); __nosave_end = .; diff --git a/arch/x86_64/kernel/vsyscall.c b/arch/x86_64/kernel/vsyscall.c index b4b8dc59663..2e573442594 100644 --- a/arch/x86_64/kernel/vsyscall.c +++ b/arch/x86_64/kernel/vsyscall.c @@ -65,7 +65,7 @@ static force_inline void do_vgettimeofday(struct timeval * tv) usec = (__xtime.tv_nsec / 1000) + (__jiffies - __wall_jiffies) * (1000000 / HZ); - if (__vxtime.mode == VXTIME_TSC) { + if (__vxtime.mode != VXTIME_HPET) { sync_core(); rdtscll(t); if (t < __vxtime.last_tsc) @@ -217,8 +217,9 @@ static int __init vsyscall_init(void) BUG_ON((unsigned long) &vtime != VSYSCALL_ADDR(__NR_vtime)); BUG_ON((VSYSCALL_ADDR(0) != __fix_to_virt(VSYSCALL_FIRST_PAGE))); map_vsyscall(); - sysctl_vsyscall = 1; +#ifdef CONFIG_SYSCTL register_sysctl_table(kernel_root_table2, 0); +#endif return 0; } diff --git a/arch/x86_64/kernel/x8664_ksyms.c b/arch/x86_64/kernel/x8664_ksyms.c index a43dedb58fa..68ec03070e5 100644 --- a/arch/x86_64/kernel/x8664_ksyms.c +++ b/arch/x86_64/kernel/x8664_ksyms.c @@ -193,8 +193,9 @@ EXPORT_SYMBOL(smp_num_siblings); extern void do_softirq_thunk(void); EXPORT_SYMBOL(do_softirq_thunk); -void out_of_line_bug(void); +#ifdef CONFIG_BUG EXPORT_SYMBOL(out_of_line_bug); +#endif EXPORT_SYMBOL(init_level4_pgt); |