diff options
Diffstat (limited to 'arch/x86/kernel/cpu')
-rw-r--r-- | arch/x86/kernel/cpu/Makefile | 4 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/addon_cpuid_features.c | 17 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/amd.c | 42 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/amd_64.c | 222 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/bugs.c | 27 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/bugs_64.c | 33 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/centaur_64.c | 35 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/common.c | 2 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/common_64.c | 681 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/cpu.h | 5 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/cpufreq/cpufreq-nforce2.c | 44 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/intel.c | 4 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/intel_64.c | 95 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/intel_cacheinfo.c | 1 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/mcheck/k7.c | 36 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/mcheck/mce_64.c | 28 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/mcheck/p4.c | 90 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/mtrr/generic.c | 38 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/mtrr/main.c | 901 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/mtrr/mtrr.h | 3 | ||||
-rw-r--r-- | arch/x86/kernel/cpu/perfctr-watchdog.c | 206 |
21 files changed, 2239 insertions, 275 deletions
diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile index a0c6f81..ee76eaa 100644 --- a/arch/x86/kernel/cpu/Makefile +++ b/arch/x86/kernel/cpu/Makefile @@ -6,11 +6,15 @@ obj-y := intel_cacheinfo.o addon_cpuid_features.o obj-y += proc.o feature_names.o obj-$(CONFIG_X86_32) += common.o bugs.o +obj-$(CONFIG_X86_64) += common_64.o bugs_64.o obj-$(CONFIG_X86_32) += amd.o +obj-$(CONFIG_X86_64) += amd_64.o obj-$(CONFIG_X86_32) += cyrix.o obj-$(CONFIG_X86_32) += centaur.o +obj-$(CONFIG_X86_64) += centaur_64.o obj-$(CONFIG_X86_32) += transmeta.o obj-$(CONFIG_X86_32) += intel.o +obj-$(CONFIG_X86_64) += intel_64.o obj-$(CONFIG_X86_32) += umc.o obj-$(CONFIG_X86_MCE) += mcheck/ diff --git a/arch/x86/kernel/cpu/addon_cpuid_features.c b/arch/x86/kernel/cpu/addon_cpuid_features.c index c2e1ce3..84a8220 100644 --- a/arch/x86/kernel/cpu/addon_cpuid_features.c +++ b/arch/x86/kernel/cpu/addon_cpuid_features.c @@ -1,9 +1,7 @@ - /* * Routines to indentify additional cpu features that are scattered in * cpuid space. */ - #include <linux/cpu.h> #include <asm/pat.h> @@ -53,19 +51,20 @@ void __cpuinit init_scattered_cpuid_features(struct cpuinfo_x86 *c) #ifdef CONFIG_X86_PAT void __cpuinit validate_pat_support(struct cpuinfo_x86 *c) { + if (!cpu_has_pat) + pat_disable("PAT not supported by CPU."); + switch (c->x86_vendor) { - case X86_VENDOR_AMD: - if (c->x86 >= 0xf && c->x86 <= 0x11) - return; - break; case X86_VENDOR_INTEL: if (c->x86 == 0xF || (c->x86 == 6 && c->x86_model >= 15)) return; break; + case X86_VENDOR_AMD: + case X86_VENDOR_CENTAUR: + case X86_VENDOR_TRANSMETA: + return; } - pat_disable(cpu_has_pat ? - "PAT disabled. Not yet verified on this CPU type." : - "PAT not supported by CPU."); + pat_disable("PAT disabled. Not yet verified on this CPU type."); } #endif diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 2458668..81a07ca 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -24,43 +24,6 @@ extern void vide(void); __asm__(".align 4\nvide: ret"); -#ifdef CONFIG_X86_LOCAL_APIC -#define ENABLE_C1E_MASK 0x18000000 -#define CPUID_PROCESSOR_SIGNATURE 1 -#define CPUID_XFAM 0x0ff00000 -#define CPUID_XFAM_K8 0x00000000 -#define CPUID_XFAM_10H 0x00100000 -#define CPUID_XFAM_11H 0x00200000 -#define CPUID_XMOD 0x000f0000 -#define CPUID_XMOD_REV_F 0x00040000 - -/* AMD systems with C1E don't have a working lAPIC timer. Check for that. */ -static __cpuinit int amd_apic_timer_broken(void) -{ - u32 lo, hi; - u32 eax = cpuid_eax(CPUID_PROCESSOR_SIGNATURE); - switch (eax & CPUID_XFAM) { - case CPUID_XFAM_K8: - if ((eax & CPUID_XMOD) < CPUID_XMOD_REV_F) - break; - case CPUID_XFAM_10H: - case CPUID_XFAM_11H: - rdmsr(MSR_K8_ENABLE_C1E, lo, hi); - if (lo & ENABLE_C1E_MASK) { - if (smp_processor_id() != boot_cpu_physical_apicid) - printk(KERN_INFO "AMD C1E detected late. " - " Force timer broadcast.\n"); - return 1; - } - break; - default: - /* err on the side of caution */ - return 1; - } - return 0; -} -#endif - int force_mwait __cpuinitdata; static void __cpuinit early_init_amd(struct cpuinfo_x86 *c) @@ -297,11 +260,6 @@ static void __cpuinit init_amd(struct cpuinfo_x86 *c) num_cache_leaves = 3; } -#ifdef CONFIG_X86_LOCAL_APIC - if (amd_apic_timer_broken()) - local_apic_timer_disabled = 1; -#endif - /* K6s reports MCEs but don't actually have all the MSRs */ if (c->x86 < 6) clear_cpu_cap(c, X86_FEATURE_MCE); diff --git a/arch/x86/kernel/cpu/amd_64.c b/arch/x86/kernel/cpu/amd_64.c new file mode 100644 index 0000000..7c36fb8 --- /dev/null +++ b/arch/x86/kernel/cpu/amd_64.c @@ -0,0 +1,222 @@ +#include <linux/init.h> +#include <linux/mm.h> + +#include <asm/numa_64.h> +#include <asm/mmconfig.h> +#include <asm/cacheflush.h> + +#include <mach_apic.h> + +#include "cpu.h" + +int force_mwait __cpuinitdata; + +#ifdef CONFIG_NUMA +static int __cpuinit nearby_node(int apicid) +{ + int i, node; + + for (i = apicid - 1; i >= 0; i--) { + node = apicid_to_node[i]; + if (node != NUMA_NO_NODE && node_online(node)) + return node; + } + for (i = apicid + 1; i < MAX_LOCAL_APIC; i++) { + node = apicid_to_node[i]; + if (node != NUMA_NO_NODE && node_online(node)) + return node; + } + return first_node(node_online_map); /* Shouldn't happen */ +} +#endif + +/* + * 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. + */ +static void __cpuinit amd_detect_cmp(struct cpuinfo_x86 *c) +{ +#ifdef CONFIG_SMP + unsigned bits; +#ifdef CONFIG_NUMA + int cpu = smp_processor_id(); + int node = 0; + unsigned apicid = hard_smp_processor_id(); +#endif + bits = c->x86_coreid_bits; + + /* Low order bits define the core id (index of core in socket) */ + c->cpu_core_id = c->initial_apicid & ((1 << bits)-1); + /* Convert the initial APIC ID into the socket ID */ + c->phys_proc_id = c->initial_apicid >> bits; + +#ifdef CONFIG_NUMA + node = c->phys_proc_id; + if (apicid_to_node[apicid] != NUMA_NO_NODE) + node = apicid_to_node[apicid]; + if (!node_online(node)) { + /* Two possibilities here: + - The CPU is missing memory and no node was created. + In that case try picking one from a nearby CPU + - The APIC IDs differ from the HyperTransport node IDs + which the K8 northbridge parsing fills in. + Assume they are all increased by a constant offset, + but in the same order as the HT nodeids. + If that doesn't result in a usable node fall back to the + path for the previous case. */ + + int ht_nodeid = c->initial_apicid; + + if (ht_nodeid >= 0 && + apicid_to_node[ht_nodeid] != NUMA_NO_NODE) + node = apicid_to_node[ht_nodeid]; + /* Pick a nearby node */ + if (!node_online(node)) + node = nearby_node(apicid); + } + numa_set_node(cpu, node); + + printk(KERN_INFO "CPU %d/%x -> Node %d\n", cpu, apicid, node); +#endif +#endif +} + +static void __cpuinit early_init_amd_mc(struct cpuinfo_x86 *c) +{ +#ifdef CONFIG_SMP + unsigned bits, ecx; + + /* Multi core CPU? */ + if (c->extended_cpuid_level < 0x80000008) + return; + + ecx = cpuid_ecx(0x80000008); + + c->x86_max_cores = (ecx & 0xff) + 1; + + /* CPU telling us the core id bits shift? */ + bits = (ecx >> 12) & 0xF; + + /* Otherwise recompute */ + if (bits == 0) { + while ((1 << bits) < c->x86_max_cores) + bits++; + } + + c->x86_coreid_bits = bits; + +#endif +} + +static void __cpuinit early_init_amd(struct cpuinfo_x86 *c) +{ + early_init_amd_mc(c); + + /* c->x86_power is 8000_0007 edx. Bit 8 is constant TSC */ + if (c->x86_power & (1<<8)) + set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); +} + +static void __cpuinit init_amd(struct cpuinfo_x86 *c) +{ + unsigned level; + +#ifdef CONFIG_SMP + unsigned long value; + + /* + * Disable TLB flush filter by setting HWCR.FFDIS on K8 + * bit 6 of msr C001_0015 + * + * Errata 63 for SH-B3 steppings + * Errata 122 for all steppings (F+ have it disabled by default) + */ + if (c->x86 == 0xf) { + rdmsrl(MSR_K8_HWCR, value); + value |= 1 << 6; + wrmsrl(MSR_K8_HWCR, value); + } +#endif + + /* Bit 31 in normal CPUID used for nonstandard 3DNow ID; + 3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */ + clear_cpu_cap(c, 0*32+31); + + /* On C+ stepping K8 rep microcode works well for copy/memset */ + if (c->x86 == 0xf) { + level = cpuid_eax(1); + if((level >= 0x0f48 && level < 0x0f50) || level >= 0x0f58) + set_cpu_cap(c, X86_FEATURE_REP_GOOD); + } + if (c->x86 == 0x10 || c->x86 == 0x11) + set_cpu_cap(c, X86_FEATURE_REP_GOOD); + + /* Enable workaround for FXSAVE leak */ + if (c->x86 >= 6) + set_cpu_cap(c, X86_FEATURE_FXSAVE_LEAK); + + level = get_model_name(c); + if (!level) { + switch (c->x86) { + case 0xf: + /* Should distinguish Models here, but this is only + a fallback anyways. */ + strcpy(c->x86_model_id, "Hammer"); + break; + } + } + display_cacheinfo(c); + + /* Multi core CPU? */ + if (c->extended_cpuid_level >= 0x80000008) + amd_detect_cmp(c); + + if (c->extended_cpuid_level >= 0x80000006 && + (cpuid_edx(0x80000006) & 0xf000)) + num_cache_leaves = 4; + else + num_cache_leaves = 3; + + if (c->x86 >= 0xf && c->x86 <= 0x11) + set_cpu_cap(c, X86_FEATURE_K8); + + /* MFENCE stops RDTSC speculation */ + set_cpu_cap(c, X86_FEATURE_MFENCE_RDTSC); + + if (c->x86 == 0x10) { + /* do this for boot cpu */ + if (c == &boot_cpu_data) + check_enable_amd_mmconf_dmi(); + + fam10h_check_enable_mmcfg(); + } + + if (c == &boot_cpu_data && c->x86 >= 0xf && c->x86 <= 0x11) { + unsigned long long tseg; + + /* + * Split up direct mapping around the TSEG SMM area. + * Don't do it for gbpages because there seems very little + * benefit in doing so. + */ + if (!rdmsrl_safe(MSR_K8_TSEG_ADDR, &tseg)) { + printk(KERN_DEBUG "tseg: %010llx\n", tseg); + if ((tseg>>PMD_SHIFT) < + (max_low_pfn_mapped>>(PMD_SHIFT-PAGE_SHIFT)) || + ((tseg>>PMD_SHIFT) < + (max_pfn_mapped>>(PMD_SHIFT-PAGE_SHIFT)) && + (tseg>>PMD_SHIFT) >= (1ULL<<(32 - PMD_SHIFT)))) + set_memory_4k((unsigned long)__va(tseg), 1); + } + } +} + +static struct cpu_dev amd_cpu_dev __cpuinitdata = { + .c_vendor = "AMD", + .c_ident = { "AuthenticAMD" }, + .c_early_init = early_init_amd, + .c_init = init_amd, +}; + +cpu_vendor_dev_register(X86_VENDOR_AMD, &amd_cpu_dev); + diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c index 170d2f5..1b1c56b 100644 --- a/arch/x86/kernel/cpu/bugs.c +++ b/arch/x86/kernel/cpu/bugs.c @@ -59,8 +59,12 @@ static void __init check_fpu(void) return; } -/* trap_init() enabled FXSR and company _before_ testing for FP problems here. */ - /* Test for the divl bug.. */ + /* + * trap_init() enabled FXSR and company _before_ testing for FP + * problems here. + * + * Test for the divl bug.. + */ __asm__("fninit\n\t" "fldl %1\n\t" "fdivl %2\n\t" @@ -108,10 +112,15 @@ static void __init check_popad(void) "movl $12345678,%%eax; movl $0,%%edi; pusha; popa; movl (%%edx,%%edi),%%ecx " : "=&a" (res) : "d" (inp) - : "ecx", "edi" ); - /* If this fails, it means that any user program may lock the CPU hard. Too bad. */ - if (res != 12345678) printk( "Buggy.\n" ); - else printk( "OK.\n" ); + : "ecx", "edi"); + /* + * If this fails, it means that any user program may lock the + * CPU hard. Too bad. + */ + if (res != 12345678) + printk("Buggy.\n"); + else + printk("OK.\n"); #endif } @@ -137,7 +146,8 @@ static void __init check_config(void) * i486+ only features! (WP works in supervisor mode and the * new "invlpg" and "bswap" instructions) */ -#if defined(CONFIG_X86_WP_WORKS_OK) || defined(CONFIG_X86_INVLPG) || defined(CONFIG_X86_BSWAP) +#if defined(CONFIG_X86_WP_WORKS_OK) || defined(CONFIG_X86_INVLPG) || \ + defined(CONFIG_X86_BSWAP) if (boot_cpu_data.x86 == 3) panic("Kernel requires i486+ for 'invlpg' and other features"); #endif @@ -170,6 +180,7 @@ void __init check_bugs(void) check_fpu(); check_hlt(); check_popad(); - init_utsname()->machine[1] = '0' + (boot_cpu_data.x86 > 6 ? 6 : boot_cpu_data.x86); + init_utsname()->machine[1] = + '0' + (boot_cpu_data.x86 > 6 ? 6 : boot_cpu_data.x86); alternative_instructions(); } diff --git a/arch/x86/kernel/cpu/bugs_64.c b/arch/x86/kernel/cpu/bugs_64.c new file mode 100644 index 0000000..9a3ed06 --- /dev/null +++ b/arch/x86/kernel/cpu/bugs_64.c @@ -0,0 +1,33 @@ +/* + * Copyright (C) 1994 Linus Torvalds + * Copyright (C) 2000 SuSE + */ + +#include <linux/kernel.h> +#include <linux/init.h> +#include <asm/alternative.h> +#include <asm/bugs.h> +#include <asm/processor.h> +#include <asm/mtrr.h> +#include <asm/cacheflush.h> + +void __init check_bugs(void) +{ + identify_boot_cpu(); +#if !defined(CONFIG_SMP) + printk("CPU: "); + print_cpu_info(&boot_cpu_data); +#endif + alternative_instructions(); + + /* + * Make sure the first 2MB area is not mapped by huge pages + * There are typically fixed size MTRRs in there and overlapping + * MTRRs into large pages causes slow downs. + * + * Right now we don't do that with gbpages because there seems + * very little benefit for that case. + */ + if (!direct_gbpages) + set_memory_4k((unsigned long)__va(0), 1); +} diff --git a/arch/x86/kernel/cpu/centaur_64.c b/arch/x86/kernel/cpu/centaur_64.c new file mode 100644 index 0000000..1d181c4 --- /dev/null +++ b/arch/x86/kernel/cpu/centaur_64.c @@ -0,0 +1,35 @@ +#include <linux/init.h> +#include <linux/smp.h> + +#include <asm/cpufeature.h> +#include <asm/processor.h> + +#include "cpu.h" + +static void __cpuinit early_init_centaur(struct cpuinfo_x86 *c) +{ + if (c->x86 == 0x6 && c->x86_model >= 0xf) + set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + + set_cpu_cap(c, X86_FEATURE_SYSENTER32); +} + +static void __cpuinit init_centaur(struct cpuinfo_x86 *c) +{ + if (c->x86 == 0x6 && c->x86_model >= 0xf) { + c->x86_cache_alignment = c->x86_clflush_size * 2; + set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + set_cpu_cap(c, X86_FEATURE_REP_GOOD); + } + set_cpu_cap(c, X86_FEATURE_LFENCE_RDTSC); +} + +static struct cpu_dev centaur_cpu_dev __cpuinitdata = { + .c_vendor = "Centaur", + .c_ident = { "CentaurHauls" }, + .c_early_init = early_init_centaur, + .c_init = init_centaur, +}; + +cpu_vendor_dev_register(X86_VENDOR_CENTAUR, ¢aur_cpu_dev); + diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index d0463a9..80ab20d 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -427,7 +427,7 @@ __setup("serialnumber", x86_serial_nr_setup); /* * This does the hard work of actually picking apart the CPU stuff... */ -void __cpuinit identify_cpu(struct cpuinfo_x86 *c) +static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) { int i; diff --git a/arch/x86/kernel/cpu/common_64.c b/arch/x86/kernel/cpu/common_64.c new file mode 100644 index 0000000..7b8cc72 --- /dev/null +++ b/arch/x86/kernel/cpu/common_64.c @@ -0,0 +1,681 @@ +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/sched.h> +#include <linux/string.h> +#include <linux/bootmem.h> +#include <linux/bitops.h> +#include <linux/module.h> +#include <linux/kgdb.h> +#include <linux/topology.h> +#include <linux/string.h> +#include <linux/delay.h> +#include <linux/smp.h> +#include <linux/module.h> +#include <linux/percpu.h> +#include <asm/processor.h> +#include <asm/i387.h> +#include <asm/msr.h> +#include <asm/io.h> +#include <asm/mmu_context.h> +#include <asm/mtrr.h> +#include <asm/mce.h> +#include <asm/pat.h> +#include <asm/numa.h> +#ifdef CONFIG_X86_LOCAL_APIC +#include <asm/mpspec.h> +#include <asm/apic.h> +#include <mach_apic.h> +#endif +#include <asm/pda.h> +#include <asm/pgtable.h> +#include <asm/processor.h> +#include <asm/desc.h> +#include <asm/atomic.h> +#include <asm/proto.h> +#include <asm/sections.h> +#include <asm/setup.h> +#include <asm/genapic.h> + +#include "cpu.h" + +/* We need valid kernel segments for data and code in long mode too + * IRET will check the segment types kkeil 2000/10/28 + * Also sysret mandates a special GDT layout + */ +/* The TLS descriptors are currently at a different place compared to i386. + Hopefully nobody expects them at a fixed place (Wine?) */ +DEFINE_PER_CPU(struct gdt_page, gdt_page) = { .gdt = { + [GDT_ENTRY_KERNEL32_CS] = { { { 0x0000ffff, 0x00cf9b00 } } }, + [GDT_ENTRY_KERNEL_CS] = { { { 0x0000ffff, 0x00af9b00 } } }, + [GDT_ENTRY_KERNEL_DS] = { { { 0x0000ffff, 0x00cf9300 } } }, + [GDT_ENTRY_DEFAULT_USER32_CS] = { { { 0x0000ffff, 0x00cffb00 } } }, + [GDT_ENTRY_DEFAULT_USER_DS] = { { { 0x0000ffff, 0x00cff300 } } }, + [GDT_ENTRY_DEFAULT_USER_CS] = { { { 0x0000ffff, 0x00affb00 } } }, +} }; +EXPORT_PER_CPU_SYMBOL_GPL(gdt_page); + +__u32 cleared_cpu_caps[NCAPINTS] __cpuinitdata; + +/* Current gdt points %fs at the "master" per-cpu area: after this, + * it's on the real one. */ +void switch_to_new_gdt(void) +{ + struct desc_ptr gdt_descr; + + gdt_descr.address = (long)get_cpu_gdt_table(smp_processor_id()); + gdt_descr.size = GDT_SIZE - 1; + load_gdt(&gdt_descr); +} + +struct cpu_dev *cpu_devs[X86_VENDOR_NUM] = {}; + +static void __cpuinit default_init(struct cpuinfo_x86 *c) +{ + display_cacheinfo(c); +} + +static struct cpu_dev __cpuinitdata default_cpu = { + .c_init = default_init, + .c_vendor = "Unknown", +}; +static struct cpu_dev *this_cpu __cpuinitdata = &default_cpu; + +int __cpuinit get_model_name(struct cpuinfo_x86 *c) +{ + unsigned int *v; + + if (c->extended_cpuid_level < 0x80000004) + return 0; + + v = (unsigned int *) c->x86_model_id; + cpuid(0x80000002, &v[0], &v[1], &v[2], &v[3]); + cpuid(0x80000003, &v[4], &v[5], &v[6], &v[7]); + cpuid(0x80000004, &v[8], &v[9], &v[10], &v[11]); + c->x86_model_id[48] = 0; + return 1; +} + + +void __cpuinit display_cacheinfo(struct cpuinfo_x86 *c) +{ + unsigned int n, dummy, ebx, ecx, edx; + + n = c->extended_cpuid_level; + + if (n >= 0x80000005) { + cpuid(0x80000005, &dummy, &ebx, &ecx, &edx); + printk(KERN_INFO "CPU: L1 I Cache: %dK (%d bytes/line), " + "D cache %dK (%d bytes/line)\n", + edx>>24, edx&0xFF, ecx>>24, ecx&0xFF); + c->x86_cache_size = (ecx>>24) + (edx>>24); + /* On K8 L1 TLB is inclusive, so don't count it */ + c->x86_tlbsize = 0; + } + + if (n >= 0x80000006) { + cpuid(0x80000006, &dummy, &ebx, &ecx, &edx); + ecx = cpuid_ecx(0x80000006); + c->x86_cache_size = ecx >> 16; + c->x86_tlbsize += ((ebx >> 16) & 0xfff) + (ebx & 0xfff); + + printk(KERN_INFO "CPU: L2 Cache: %dK (%d bytes/line)\n", + c->x86_cache_size, ecx & 0xFF); + } +} + +void __cpuinit detect_ht(struct cpuinfo_x86 *c) +{ +#ifdef CONFIG_SMP + u32 eax, ebx, ecx, edx; + int index_msb, core_bits; + + cpuid(1, &eax, &ebx, &ecx, &edx); + + + if (!cpu_has(c, X86_FEATURE_HT)) + return; + if (cpu_has(c, X86_FEATURE_CMP_LEGACY)) + goto out; + + smp_num_siblings = (ebx & 0xff0000) >> 16; + + if (smp_num_siblings == 1) { + printk(KERN_INFO "CPU: Hyper-Threading is disabled\n"); + } else if (smp_num_siblings > 1) { + + if (smp_num_siblings > NR_CPUS) { + printk(KERN_WARNING "CPU: Unsupported number of " + "siblings %d", smp_num_siblings); + smp_num_siblings = 1; + return; + } + + index_msb = get_count_order(smp_num_siblings); + c->phys_proc_id = phys_pkg_id(index_msb); + + smp_num_siblings = smp_num_siblings / c->x86_max_cores; + + index_msb = get_count_order(smp_num_siblings); + + core_bits = get_count_order(c->x86_max_cores); + + c->cpu_core_id = phys_pkg_id(index_msb) & + ((1 << core_bits) - 1); + } +out: + if ((c->x86_max_cores * smp_num_siblings) > 1) { + printk(KERN_INFO "CPU: Physical Processor ID: %d\n", + c->phys_proc_id); + printk(KERN_INFO "CPU: Processor Core ID: %d\n", + c->cpu_core_id); + } + +#endif +} + +static void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c) +{ + char *v = c->x86_vendor_id; + int i; + static int printed; + + for (i = 0; i < X86_VENDOR_NUM; i++) { + if (cpu_devs[i]) { + if (!strcmp(v, cpu_devs[i]->c_ident[0]) || + (cpu_devs[i]->c_ident[1] && + !strcmp(v, cpu_devs[i]->c_ident[1]))) { + c->x86_vendor = i; + this_cpu = cpu_devs[i]; + return; + } + } + } + if (!printed) { + printed++; + printk(KERN_ERR "CPU: Vendor unknown, using generic init.\n"); + printk(KERN_ERR "CPU: Your system may be unstable.\n"); + } + c->x86_vendor = X86_VENDOR_UNKNOWN; +} + +static void __init early_cpu_support_print(void) +{ + int i,j; + struct cpu_dev *cpu_devx; + + printk("KERNEL supported cpus:\n"); + for (i = 0; i < X86_VENDOR_NUM; i++) { + cpu_devx = cpu_devs[i]; + if (!cpu_devx) + continue; + for (j = 0; j < 2; j++) { + if (!cpu_devx->c_ident[j]) + continue; + printk(" %s %s\n", cpu_devx->c_vendor, + cpu_devx->c_ident[j]); + } + } +} + +static void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c); + +void __init early_cpu_init(void) +{ + struct cpu_vendor_dev *cvdev; + + for (cvdev = __x86cpuvendor_start ; + cvdev < __x86cpuvendor_end ; + cvdev++) + cpu_devs[cvdev->vendor] = cvdev->cpu_dev; + early_cpu_support_print(); + early_identify_cpu(&boot_cpu_data); +} + +/* 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. */ +static void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c) +{ + u32 tfms, xlvl; + + c->loops_per_jiffy = loops_per_jiffy; + c->x86_cache_size = -1; + c->x86_vendor = X86_VENDOR_UNKNOWN; + c->x86_model = c->x86_mask = 0; /* So far unknown... */ + c->x86_vendor_id[0] = '\0'; /* Unset */ + c->x86_model_id[0] = '\0'; /* Unset */ + c->x86_clflush_size = 64; + c->x86_cache_alignment = c->x86_clflush_size; + c->x86_max_cores = 1; + c->x86_coreid_bits = 0; + c->extended_cpuid_level = 0; + memset(&c->x86_capability, 0, sizeof c->x86_capability); + + /* Get vendor name */ + cpuid(0x00000000, (unsigned int *)&c->cpuid_level, + (unsigned int *)&c->x86_vendor_id[0], + (unsigned int *)&c->x86_vendor_id[8], + (unsigned int *)&c->x86_vendor_id[4]); + + get_cpu_vendor(c); + + /* Initialize the standard set of capabilities */ + /* Note that the vendor-specific code below might override */ + + /* Intel-defined flags: level 0x00000001 */ + if (c->cpuid_level >= 0x00000001) { + __u32 misc; + cpuid(0x00000001, &tfms, &misc, &c->x86_capability[4], + &c->x86_capability[0]); + c->x86 = (tfms >> 8) & 0xf; + c->x86_model = (tfms >> 4) & 0xf; + c->x86_mask = tfms & 0xf; + if (c->x86 == 0xf) + c->x86 += (tfms >> 20) & 0xff; + if (c->x86 >= 0x6) + c->x86_model += ((tfms >> 16) & 0xF) << 4; + if (test_cpu_cap(c, X86_FEATURE_CLFLSH)) + c->x86_clflush_size = ((misc >> 8) & 0xff) * 8; + } else { + /* Have CPUID level 0 only - unheard of */ + c->x86 = 4; + } + + c->initial_apicid = (cpuid_ebx(1) >> 24) & 0xff; +#ifdef CONFIG_SMP + c->phys_proc_id = c->initial_apicid; +#endif + /* AMD-defined flags: level 0x80000001 */ + xlvl = cpuid_eax(0x80000000); + c->extended_cpuid_level = xlvl; + if ((xlvl & 0xffff0000) == 0x80000000) { + if (xlvl >= 0x80000001) { + c->x86_capability[1] = cpuid_edx(0x80000001); + c->x86_capability[6] = cpuid_ecx(0x80000001); + } + if (xlvl >= 0x80000004) + get_model_name(c); /* Default name */ + } + + /* Transmeta-defined flags: level 0x80860001 */ + xlvl = cpuid_eax(0x80860000); + if ((xlvl & 0xffff0000) == 0x80860000) { + /* Don't set x86_cpuid_level here for now to not confuse. */ + if (xlvl >= 0x80860001) + c->x86_capability[2] = cpuid_edx(0x80860001); + } + + c->extended_cpuid_level = cpuid_eax(0x80000000); + if (c->extended_cpuid_level >= 0x80000007) + c->x86_power = cpuid_edx(0x80000007); + + if (c->extended_cpuid_level >= 0x80000008) { + u32 eax = cpuid_eax(0x80000008); + + c->x86_virt_bits = (eax >> 8) & 0xff; + c->x86_phys_bits = eax & 0xff; + } + + /* Assume all 64-bit CPUs support 32-bit syscall */ + set_cpu_cap(c, X86_FEATURE_SYSCALL32); + + if (c->x86_vendor != X86_VENDOR_UNKNOWN && + cpu_devs[c->x86_vendor]->c_early_init) + cpu_devs[c->x86_vendor]->c_early_init(c); + + validate_pat_support(c); + + /* early_param could clear that, but recall get it set again */ + if (disable_apic) + clear_cpu_cap(c, X86_FEATURE_APIC); +} + +/* + * This does the hard work of actually picking apart the CPU stuff... + */ +static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) +{ + int i; + + early_identify_cpu(c); + + init_scattered_cpuid_features(c); + + c->apicid = phys_pkg_id(0); + + /* + * Vendor-specific initialization. In this section we + * canonicalize the feature flags, meaning if there are + * features a certain CPU supports which CPUID doesn't + * tell us, CPUID claiming incorrect flags, or other bugs, + * we handle them here. + * + * At the end of this section, c->x86_capability better + * indicate the features this CPU genuinely supports! + */ + if (this_cpu->c_init) + this_cpu->c_init(c); + + detect_ht(c); + + /* + * On SMP, boot_cpu_data holds the common feature set between + * all CPUs; so make sure that we indicate which features are + * common between the CPUs. The first time this routine gets + * executed, c == &boot_cpu_data. + */ + if (c != &boot_cpu_data) { + /* AND the already accumulated flags with these */ + for (i = 0; i < NCAPINTS; i++) + boot_cpu_data.x86_capability[i] &= c->x86_capability[i]; + } + + /* Clear all flags overriden by options */ + for (i = 0; i < NCAPINTS; i++) + c->x86_capability[i] &= ~cleared_cpu_caps[i]; + +#ifdef CONFIG_X86_MCE + mcheck_init(c); +#endif + select_idle_routine(c); + +#ifdef CONFIG_NUMA + numa_add_cpu(smp_processor_id()); +#endif + +} + +void __cpuinit identify_boot_cpu(void) +{ + identify_cpu(&boot_cpu_data); +} + +void __cpuinit identify_secondary_cpu(struct cpuinfo_x86 *c) +{ + BUG_ON(c == &boot_cpu_data); + identify_cpu(c); + mtrr_ap_init(); +} + +static __init int setup_noclflush(char *arg) +{ + setup_clear_cpu_cap(X86_FEATURE_CLFLSH); + return 1; +} +__setup("noclflush", setup_noclflush); + +void __cpuinit print_cpu_info(struct cpuinfo_x86 *c) +{ + if (c->x86_model_id[0]) + printk(KERN_CONT "%s", c->x86_model_id); + + if (c->x86_mask || c->cpuid_level >= 0) + printk(KERN_CONT " stepping %02x\n", c->x86_mask); + else + printk(KERN_CONT "\n"); +} + +static __init int setup_disablecpuid(char *arg) +{ + int bit; + if (get_option(&arg, &bit) && bit < NCAPINTS*32) + setup_clear_cpu_cap(bit); + else + return 0; + return 1; +} +__setup("clearcpuid=", setup_disablecpuid); + +cpumask_t cpu_initialized __cpuinitdata = CPU_MASK_NONE; + +struct x8664_pda **_cpu_pda __read_mostly; +EXPORT_SYMBOL(_cpu_pda); + +struct desc_ptr idt_descr = { 256 * 16 - 1, (unsigned long) idt_table }; + +char boot_cpu_stack[IRQSTACKSIZE] __page_aligned_bss; + +unsigned long __supported_pte_mask __read_mostly = ~0UL; +EXPORT_SYMBOL_GPL(__supported_pte_mask); + +static int do_not_nx __cpuinitdata; + +/* noexec=on|off +Control non executable mappings for 64bit processes. + +on Enable(default) +off Disable +*/ +static int __init nonx_setup(char *str) +{ + if (!str) + return -EINVAL; + if (!strncmp(str, "on", 2)) { + __supported_pte_mask |= _PAGE_NX; + do_not_nx = 0; + } else if (!strncmp(str, "off", 3)) { + do_not_nx = 1; + __supported_pte_mask &= ~_PAGE_NX; + } + return 0; +} +early_param("noexec", nonx_setup); + +int force_personality32; + +/* noexec32=on|off +Control non executable heap for 32bit processes. +To control the stack too use noexec=off + +on PROT_READ does not imply PROT_EXEC for 32bit processes (default) +off PROT_READ implies PROT_EXEC +*/ +static int __init nonx32_setup(char *str) +{ + if (!strcmp(str, "on")) + force_personality32 &= ~READ_IMPLIES_EXEC; + else if (!strcmp(str, "off")) + force_personality32 |= READ_IMPLIES_EXEC; + return 1; +} +__setup("noexec32=", nonx32_setup); + +void pda_init(int cpu) +{ + struct x8664_pda *pda = cpu_pda(cpu); + + /* Setup up data that may be needed in __get_free_pages early */ + loadsegment(fs, 0); + loadsegment(gs, 0); + /* Memory clobbers used to order PDA accessed */ + mb(); + wrmsrl(MSR_GS_BASE, pda); + mb(); + + pda->cpunumber = cpu; + pda->irqcount = -1; + pda->kernelstack = (unsigned long)stack_thread_info() - + PDA_STACKOFFSET + THREAD_SIZE; + pda->active_mm = &init_mm; + pda->mmu_state = 0; + + if (cpu == 0) { + /* others are initialized in smpboot.c */ + pda->pcurrent = &init_task; + pda->irqstackptr = boot_cpu_stack; + } else { + pda->irqstackptr = (char *) + __get_free_pages(GFP_ATOMIC, IRQSTACK_ORDER); + if (!pda->irqstackptr) + panic("cannot allocate irqstack for cpu %d", cpu); + + if (pda->nodenumber == 0 && cpu_to_node(cpu) != NUMA_NO_NODE) + pda->nodenumber = cpu_to_node(cpu); + } + + pda->irqstackptr += IRQSTACKSIZE-64; +} + +char boot_exception_stacks[(N_EXCEPTION_STACKS - 1) * EXCEPTION_STKSZ + + DEBUG_STKSZ] +__attribute__((section(".bss.page_aligned"))); + +extern asmlinkage void ignore_sysret(void); + +/* May not be marked __init: used by software suspend */ +void syscall_init(void) +{ + /* + * LSTAR and STAR live in a bit strange symbiosis. + * They both write to the same internal register. STAR allows to + * set CS/DS but only a 32bit target. LSTAR sets the 64bit rip. + */ + wrmsrl(MSR_STAR, ((u64)__USER32_CS)<<48 | ((u64)__KERNEL_CS)<<32); + wrmsrl(MSR_LSTAR, system_call); + wrmsrl(MSR_CSTAR, ignore_sysret); + +#ifdef CONFIG_IA32_EMULATION + syscall32_cpu_init(); +#endif + + /* Flags to clear on syscall */ + wrmsrl(MSR_SYSCALL_MASK, + X86_EFLAGS_TF|X86_EFLAGS_DF|X86_EFLAGS_IF|X86_EFLAGS_IOPL); +} + +void __cpuinit check_efer(void) +{ + unsigned long efer; + + rdmsrl(MSR_EFER, efer); + if (!(efer & EFER_NX) || do_not_nx) + __supported_pte_mask &= ~_PAGE_NX; +} + +unsigned long kernel_eflags; + +/* + * Copies of the original ist values from the tss are only accessed during + * debugging, no special alignment required. + */ +DEFINE_PER_CPU(struct orig_ist, orig_ist); + +/* + * cpu_init() initializes state that is per-CPU. Some data is already + * initialized (naturally) in the bootstrap process, such as the GDT + * and IDT. We reload them nevertheless, this function acts as a + * 'CPU state barrier', nothing should get across. + * A lot of state is already set up in PDA init. + */ +void __cpuinit cpu_init(void) +{ + int cpu = stack_smp_processor_id(); + struct tss_struct *t = &per_cpu(init_tss, cpu); + struct orig_ist *orig_ist = &per_cpu(orig_ist, cpu); + unsigned long v; + char *estacks = NULL; + struct task_struct *me; + int i; + + /* CPU 0 is initialised in head64.c */ + if (cpu != 0) + pda_init(cpu); + else + estacks = boot_exception_stacks; + + me = current; + + if (cpu_test_and_set(cpu, cpu_initialized)) + panic("CPU#%d already initialized!\n", cpu); + + printk(KERN_INFO "Initializing CPU#%d\n", cpu); + + clear_in_cr4(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE); + + /* + * Initialize the per-CPU GDT with the boot GDT, + * and set up the GDT descriptor: + */ + + switch_to_new_gdt(); + load_idt((const struct desc_ptr *)&idt_descr); + + memset(me->thread.tls_array, 0, GDT_ENTRY_TLS_ENTRIES * 8); + syscall_init(); + + wrmsrl(MSR_FS_BASE, 0); + wrmsrl(MSR_KERNEL_GS_BASE, 0); + barrier(); + + check_efer(); + + /* + * set up and load the per-CPU TSS + */ + for (v = 0; v < N_EXCEPTION_STACKS; v++) { + static const unsigned int order[N_EXCEPTION_STACKS] = { + [0 ... N_EXCEPTION_STACKS - 1] = EXCEPTION_STACK_ORDER, + [DEBUG_STACK - 1] = DEBUG_STACK_ORDER + }; + if (cpu) { + estacks = (char *)__get_free_pages(GFP_ATOMIC, order[v]); + if (!estacks) + panic("Cannot allocate exception stack %ld %d\n", + v, cpu); + } + estacks += PAGE_SIZE << order[v]; + orig_ist->ist[v] = t->x86_tss.ist[v] = (unsigned long)estacks; + } + + t->x86_tss.io_bitmap_base = offsetof(struct tss_struct, io_bitmap); + /* + * <= is required because the CPU will access up to + * 8 bits beyond the end of the IO permission bitmap. + */ + for (i = 0; i <= IO_BITMAP_LONGS; i++) + t->io_bitmap[i] = ~0UL; + + atomic_inc(&init_mm.mm_count); + me->active_mm = &init_mm; + if (me->mm) + BUG(); + enter_lazy_tlb(&init_mm, me); + + load_sp0(t, ¤t->thread); + set_tss_desc(cpu, t); + load_TR_desc(); + load_LDT(&init_mm.context); + +#ifdef CONFIG_KGDB + /* + * If the kgdb is connected no debug regs should be altered. This + * is only applicable when KGDB and a KGDB I/O module are built + * into the kernel and you are using early debugging with + * kgdbwait. KGDB will control the kernel HW breakpoint registers. + */ + if (kgdb_connected && arch_kgdb_ops.correct_hw_break) + arch_kgdb_ops.correct_hw_break(); + else { +#endif + /* + * Clear all 6 debug registers: + */ + + set_debugreg(0UL, 0); + set_debugreg(0UL, 1); + set_debugreg(0UL, 2); + set_debugreg(0UL, 3); + set_debugreg(0UL, 6); + set_debugreg(0UL, 7); +#ifdef CONFIG_KGDB + /* If the kgdb is connected no debug regs should be altered. */ + } +#endif + + fpu_init(); + + raw_local_save_flags(kernel_eflags); + + if (is_uv_system()) + uv_cpu_init(); +} diff --git a/arch/x86/kernel/cpu/cpu.h b/arch/x86/kernel/cpu/cpu.h index 783691b..4d894e8 100644 --- a/arch/x86/kernel/cpu/cpu.h +++ b/arch/x86/kernel/cpu/cpu.h @@ -1,3 +1,6 @@ +#ifndef ARCH_X86_CPU_H + +#define ARCH_X86_CPU_H struct cpu_model_info { int vendor; @@ -36,3 +39,5 @@ extern struct cpu_vendor_dev __x86cpuvendor_start[], __x86cpuvendor_end[]; extern int get_model_name(struct cpuinfo_x86 *c); extern void display_cacheinfo(struct cpuinfo_x86 *c); + +#endif diff --git a/arch/x86/kernel/cpu/cpufreq/cpufreq-nforce2.c b/arch/x86/kernel/cpu/cpufreq/cpufreq-nforce2.c index f03e9153..965ea52 100644 --- a/arch/x86/kernel/cpu/cpufreq/cpufreq-nforce2.c +++ b/arch/x86/kernel/cpu/cpufreq/cpufreq-nforce2.c @@ -26,9 +26,10 @@ #define NFORCE2_SAFE_DISTANCE 50 /* Delay in ms between FSB changes */ -//#define NFORCE2_DELAY 10 +/* #define NFORCE2_DELAY 10 */ -/* nforce2_chipset: +/* + * nforce2_chipset: * FSB is changed using the chipset */ static struct pci_dev *nforce2_chipset_dev; @@ -36,13 +37,13 @@ static struct pci_dev *nforce2_chipset_dev; /* fid: * multiplier * 10 */ -static int fid = 0; +static int fid; /* min_fsb, max_fsb: * minimum and maximum FSB (= FSB at boot time) */ -static int min_fsb = 0; -static int max_fsb = 0; +static int min_fsb; +static int max_fsb; MODULE_AUTHOR("Sebastian Witt <se.witt@gmx.net>"); MODULE_DESCRIPTION("nForce2 FSB changing cpufreq driver"); @@ -53,7 +54,7 @@ module_param(min_fsb, int, 0444); MODULE_PARM_DESC(fid, "CPU multiplier to use (11.5 = 115)"); MODULE_PARM_DESC(min_fsb, - "Minimum FSB to use, if not defined: current FSB - 50"); + "Minimum FSB to use, if not defined: current FSB - 50"); #define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "cpufreq-nforce2", msg) @@ -139,7 +140,7 @@ static unsigned int nforce2_fsb_read(int bootfsb) /* Get chipset boot FSB from subdevice 5 (FSB at boot-time) */ nforce2_sub5 = pci_get_subsys(PCI_VENDOR_ID_NVIDIA, - 0x01EF,PCI_ANY_ID,PCI_ANY_ID,NULL); + 0x01EF, PCI_ANY_ID, PCI_ANY_ID, NULL); if (!nforce2_sub5) return 0; @@ -147,13 +148,13 @@ static unsigned int nforce2_fsb_read(int bootfsb) fsb /= 1000000; /* Check if PLL register is already set */ - pci_read_config_byte(nforce2_chipset_dev,NFORCE2_PLLENABLE, (u8 *)&temp); + pci_read_config_byte(nforce2_chipset_dev, NFORCE2_PLLENABLE, (u8 *)&temp); - if(bootfsb || !temp) + if (bootfsb || !temp) return fsb; - + /* Use PLL register FSB value */ - pci_read_config_dword(nforce2_chipset_dev,NFORCE2_PLLREG, &temp); + pci_read_config_dword(nforce2_chipset_dev, NFORCE2_PLLREG, &temp); fsb = nforce2_calc_fsb(temp); return fsb; @@ -184,7 +185,7 @@ static int nforce2_set_fsb(unsigned int fsb) } /* First write? Then set actual value */ - pci_read_config_byte(nforce2_chipset_dev,NFORCE2_PLLENABLE, (u8 *)&temp); + pci_read_config_byte(nforce2_chipset_dev, NFORCE2_PLLENABLE, (u8 *)&temp); if (!temp) { pll = nforce2_calc_pll(tfsb); @@ -210,7 +211,8 @@ static int nforce2_set_fsb(unsigned int fsb) tfsb--; /* Calculate the PLL reg. value */ - if ((pll = nforce2_calc_pll(tfsb)) == -1) + pll = nforce2_calc_pll(tfsb); + if (pll == -1) return -EINVAL; nforce2_write_pll(pll); @@ -249,7 +251,7 @@ static unsigned int nforce2_get(unsigned int cpu) static int nforce2_target(struct cpufreq_policy *policy, unsigned int target_freq, unsigned int relation) { -// unsigned long flags; +/* unsigned long flags; */ struct cpufreq_freqs freqs; unsigned int target_fsb; @@ -271,17 +273,17 @@ static int nforce2_target(struct cpufreq_policy *policy, cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); /* Disable IRQs */ - //local_irq_save(flags); + /* local_irq_save(flags); */ if (nforce2_set_fsb(target_fsb) < 0) printk(KERN_ERR "cpufreq: Changing FSB to %d failed\n", - target_fsb); + target_fsb); else dprintk("Changed FSB successfully to %d\n", - target_fsb); + target_fsb); /* Enable IRQs */ - //local_irq_restore(flags); + /* local_irq_restore(flags); */ cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); @@ -302,8 +304,8 @@ static int nforce2_verify(struct cpufreq_policy *policy) policy->max = (fsb_pol_max + 1) * fid * 100; cpufreq_verify_within_limits(policy, - policy->cpuinfo.min_freq, - policy->cpuinfo.max_freq); + policy->cpuinfo.min_freq, + policy->cpuinfo.max_freq); return 0; } @@ -347,7 +349,7 @@ static int nforce2_cpu_init(struct cpufreq_policy *policy) /* Set maximum FSB to FSB at boot time */ max_fsb = nforce2_fsb_read(1); - if(!max_fsb) + if (!max_fsb) return -EIO; if (!min_fsb) diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index fe9224c..70609ef 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -226,6 +226,10 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) if (cpu_has_bts) ds_init_intel(c); + +#ifdef CONFIG_X86_NUMAQ + numaq_tsc_disable(); +#endif } static unsigned int __cpuinit intel_size_cache(struct cpuinfo_x86 *c, unsigned int size) diff --git a/arch/x86/kernel/cpu/intel_64.c b/arch/x86/kernel/cpu/intel_64.c new file mode 100644 index 0000000..1019c58 --- /dev/null +++ b/arch/x86/kernel/cpu/intel_64.c @@ -0,0 +1,95 @@ +#include <linux/init.h> +#include <linux/smp.h> +#include <asm/processor.h> +#include <asm/ptrace.h> +#include <asm/topology.h> +#include <asm/numa_64.h> + +#include "cpu.h" + +static void __cpuinit early_init_intel(struct cpuinfo_x86 *c) +{ + if ((c->x86 == 0xf && c->x86_model >= 0x03) || + (c->x86 == 0x6 && c->x86_model >= 0x0e)) + set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + + set_cpu_cap(c, X86_FEATURE_SYSENTER32); +} + +/* + * find out the number of processor cores on the die + */ +static int __cpuinit intel_num_cpu_cores(struct cpuinfo_x86 *c) +{ + unsigned int eax, t; + + if (c->cpuid_level < 4) + return 1; + + cpuid_count(4, 0, &eax, &t, &t, &t); + + if (eax & 0x1f) + return ((eax >> 26) + 1); + else + return 1; +} + +static void __cpuinit srat_detect_node(void) +{ +#ifdef CONFIG_NUMA + unsigned node; + int cpu = smp_processor_id(); + int apicid = hard_smp_processor_id(); + + /* Don't do the funky fallback heuristics the AMD version employs + for now. */ + node = apicid_to_node[apicid]; + if (node == NUMA_NO_NODE || !node_online(node)) + node = first_node(node_online_map); + numa_set_node(cpu, node); + + printk(KERN_INFO "CPU %d/%x -> Node %d\n", cpu, apicid, node); +#endif +} + +static void __cpuinit init_intel(struct cpuinfo_x86 *c) +{ + init_intel_cacheinfo(c); + if (c->cpuid_level > 9) { + unsigned eax = cpuid_eax(10); + /* Check for version and the number of counters */ + if ((eax & 0xff) && (((eax>>8) & 0xff) > 1)) + set_cpu_cap(c, X86_FEATURE_ARCH_PERFMON); + } + + if (cpu_has_ds) { + unsigned int l1, l2; + rdmsr(MSR_IA32_MISC_ENABLE, l1, l2); + if (!(l1 & (1<<11))) + set_cpu_cap(c, X86_FEATURE_BTS); + if (!(l1 & (1<<12))) + set_cpu_cap(c, X86_FEATURE_PEBS); + } + + + if (cpu_has_bts) + ds_init_intel(c); + + if (c->x86 == 15) + c->x86_cache_alignment = c->x86_clflush_size * 2; + if (c->x86 == 6) + set_cpu_cap(c, X86_FEATURE_REP_GOOD); + set_cpu_cap(c, X86_FEATURE_LFENCE_RDTSC); + c->x86_max_cores = intel_num_cpu_cores(c); + + srat_detect_node(); +} + +static struct cpu_dev intel_cpu_dev __cpuinitdata = { + .c_vendor = "Intel", + .c_ident = { "GenuineIntel" }, + .c_early_init = early_init_intel, + .c_init = init_intel, +}; +cpu_vendor_dev_register(X86_VENDOR_INTEL, &intel_cpu_dev); + diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 26d615d..2c8afaf 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -62,6 +62,7 @@ static struct _cache_table cache_table[] __cpuinitdata = { 0x4b, LVL_3, 8192 }, /* 16-way set assoc, 64 byte line size */ { 0x4c, LVL_3, 12288 }, /* 12-way set assoc, 64 byte line size */ { 0x4d, LVL_3, 16384 }, /* 16-way set assoc, 64 byte line size */ + { 0x4e, LVL_2, 6144 }, /* 24-way set assoc, 64 byte line size */ { 0x60, LVL_1_DATA, 16 }, /* 8-way set assoc, sectored cache, 64 byte line size */ { 0x66, LVL_1_DATA, 8 }, /* 4-way set assoc, sectored cache, 64 byte line size */ { 0x67, LVL_1_DATA, 16 }, /* 4-way set assoc, sectored cache, 64 byte line size */ diff --git a/arch/x86/kernel/cpu/mcheck/k7.c b/arch/x86/kernel/cpu/mcheck/k7.c index e633c9c..f390c9f 100644 --- a/arch/x86/kernel/cpu/mcheck/k7.c +++ b/arch/x86/kernel/cpu/mcheck/k7.c @@ -9,23 +9,23 @@ #include <linux/interrupt.h> #include <linux/smp.h> -#include <asm/processor.h> +#include <asm/processor.h> #include <asm/system.h> #include <asm/msr.h> #include "mce.h" /* Machine Check Handler For AMD Athlon/Duron */ -static void k7_machine_check(struct pt_regs * regs, long error_code) +static void k7_machine_check(struct pt_regs *regs, long error_code) { - int recover=1; + int recover = 1; u32 alow, ahigh, high, low; u32 mcgstl, mcgsth; int i; - rdmsr (MSR_IA32_MCG_STATUS, mcgstl, mcgsth); + rdmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); if (mcgstl & (1<<0)) /* Recoverable ? */ - recover=0; + recover = 0; printk(KERN_EMERG "CPU %d: Machine Check Exception: %08x%08x\n", smp_processor_id(), mcgsth, mcgstl); @@ -60,12 +60,12 @@ static void k7_machine_check(struct pt_regs * regs, long error_code) } if (recover&2) - panic ("CPU context corrupt"); + panic("CPU context corrupt"); if (recover&1) - panic ("Unable to continue"); - printk (KERN_EMERG "Attempting to continue.\n"); + panic("Unable to continue"); + printk(KERN_EMERG "Attempting to continue.\n"); mcgstl &= ~(1<<2); - wrmsr (MSR_IA32_MCG_STATUS,mcgstl, mcgsth); + wrmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); } @@ -81,25 +81,25 @@ void amd_mcheck_init(struct cpuinfo_x86 *c) machine_check_vector = k7_machine_check; wmb(); - printk (KERN_INFO "Intel machine check architecture supported.\n"); - rdmsr (MSR_IA32_MCG_CAP, l, h); + printk(KERN_INFO "Intel machine check architecture supported.\n"); + rdmsr(MSR_IA32_MCG_CAP, l, h); if (l & (1<<8)) /* Control register present ? */ - wrmsr (MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); + wrmsr(MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); nr_mce_banks = l & 0xff; /* Clear status for MC index 0 separately, we don't touch CTL, * as some K7 Athlons cause spurious MCEs when its enabled. */ if (boot_cpu_data.x86 == 6) { - wrmsr (MSR_IA32_MC0_STATUS, 0x0, 0x0); + wrmsr(MSR_IA32_MC0_STATUS, 0x0, 0x0); i = 1; } else i = 0; - for (; i<nr_mce_banks; i++) { - wrmsr (MSR_IA32_MC0_CTL+4*i, 0xffffffff, 0xffffffff); - wrmsr (MSR_IA32_MC0_STATUS+4*i, 0x0, 0x0); + for (; i < nr_mce_banks; i++) { + wrmsr(MSR_IA32_MC0_CTL+4*i, 0xffffffff, 0xffffffff); + wrmsr(MSR_IA32_MC0_STATUS+4*i, 0x0, 0x0); } - set_in_cr4 (X86_CR4_MCE); - printk (KERN_INFO "Intel machine check reporting enabled on CPU#%d.\n", + set_in_cr4(X86_CR4_MCE); + printk(KERN_INFO "Intel machine check reporting enabled on CPU#%d.\n", smp_processor_id()); } diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c index 43b7cb5..c4a7ec3 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_64.c @@ -9,6 +9,7 @@ #include <linux/types.h> #include <linux/kernel.h> #include <linux/sched.h> +#include <linux/smp_lock.h> #include <linux/string.h> #include <linux/rcupdate.h> #include <linux/kallsyms.h> @@ -31,7 +32,7 @@ #include <asm/idle.h> #define MISC_MCELOG_MINOR 227 -#define NR_BANKS 6 +#define NR_SYSFS_BANKS 6 atomic_t mce_entry; @@ -46,7 +47,7 @@ static int mce_dont_init; */ static int tolerant = 1; static int banks; -static unsigned long bank[NR_BANKS] = { [0 ... NR_BANKS-1] = ~0UL }; +static unsigned long bank[NR_SYSFS_BANKS] = { [0 ... NR_SYSFS_BANKS-1] = ~0UL }; static unsigned long notify_user; static int rip_msr; static int mce_bootlog = -1; @@ -209,7 +210,7 @@ void do_machine_check(struct pt_regs * regs, long error_code) barrier(); for (i = 0; i < banks; i++) { - if (!bank[i]) + if (i < NR_SYSFS_BANKS && !bank[i]) continue; m.misc = 0; @@ -444,9 +445,10 @@ static void mce_init(void *dummy) rdmsrl(MSR_IA32_MCG_CAP, cap); banks = cap & 0xff; - if (banks > NR_BANKS) { - printk(KERN_INFO "MCE: warning: using only %d banks\n", banks); - banks = NR_BANKS; + if (banks > MCE_EXTENDED_BANK) { + banks = MCE_EXTENDED_BANK; + printk(KERN_INFO "MCE: warning: using only %d banks\n", + MCE_EXTENDED_BANK); } /* Use accurate RIP reporting if available. */ if ((cap & (1<<9)) && ((cap >> 16) & 0xff) >= 9) @@ -462,7 +464,11 @@ static void mce_init(void *dummy) wrmsr(MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); for (i = 0; i < banks; i++) { - wrmsrl(MSR_IA32_MC0_CTL+4*i, bank[i]); + if (i < NR_SYSFS_BANKS) + wrmsrl(MSR_IA32_MC0_CTL+4*i, bank[i]); + else + wrmsrl(MSR_IA32_MC0_CTL+4*i, ~0UL); + wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0); } } @@ -527,10 +533,12 @@ static int open_exclu; /* already open exclusive? */ static int mce_open(struct inode *inode, struct file *file) { + lock_kernel(); spin_lock(&mce_state_lock); if (open_exclu || (open_count && (file->f_flags & O_EXCL))) { spin_unlock(&mce_state_lock); + unlock_kernel(); return -EBUSY; } @@ -539,6 +547,7 @@ static int mce_open(struct inode *inode, struct file *file) open_count++; spin_unlock(&mce_state_lock); + unlock_kernel(); return nonseekable_open(inode, file); } @@ -766,7 +775,10 @@ DEFINE_PER_CPU(struct sys_device, device_mce); } \ static SYSDEV_ATTR(name, 0644, show_ ## name, set_ ## name); -/* TBD should generate these dynamically based on number of available banks */ +/* + * TBD should generate these dynamically based on number of available banks. + * Have only 6 contol banks in /sysfs until then. + */ ACCESSOR(bank0ctl,bank[0],mce_restart()) ACCESSOR(bank1ctl,bank[1],mce_restart()) ACCESSOR(bank2ctl,bank[2],mce_restart()) diff --git a/arch/x86/kernel/cpu/mcheck/p4.c b/arch/x86/kernel/cpu/mcheck/p4.c index cb03345..eef001a 100644 --- a/arch/x86/kernel/cpu/mcheck/p4.c +++ b/arch/x86/kernel/cpu/mcheck/p4.c @@ -8,7 +8,7 @@ #include <linux/interrupt.h> #include <linux/smp.h> -#include <asm/processor.h> +#include <asm/processor.h> #include <asm/system.h> #include <asm/msr.h> #include <asm/apic.h> @@ -32,12 +32,12 @@ struct intel_mce_extended_msrs { /* u32 *reserved[]; */ }; -static int mce_num_extended_msrs = 0; +static int mce_num_extended_msrs; #ifdef CONFIG_X86_MCE_P4THERMAL static void unexpected_thermal_interrupt(struct pt_regs *regs) -{ +{ printk(KERN_ERR "CPU%d: Unexpected LVT TMR interrupt!\n", smp_processor_id()); add_taint(TAINT_MACHINE_CHECK); @@ -83,7 +83,7 @@ static void intel_init_thermal(struct cpuinfo_x86 *c) * be some SMM goo which handles it, so we can't even put a handler * since it might be delivered via SMI already -zwanem. */ - rdmsr (MSR_IA32_MISC_ENABLE, l, h); + rdmsr(MSR_IA32_MISC_ENABLE, l, h); h = apic_read(APIC_LVTTHMR); if ((l & (1<<3)) && (h & APIC_DM_SMI)) { printk(KERN_DEBUG "CPU%d: Thermal monitoring handled by SMI\n", @@ -91,7 +91,7 @@ static void intel_init_thermal(struct cpuinfo_x86 *c) return; /* -EBUSY */ } - /* check whether a vector already exists, temporarily masked? */ + /* check whether a vector already exists, temporarily masked? */ if (h & APIC_VECTOR_MASK) { printk(KERN_DEBUG "CPU%d: Thermal LVT vector (%#x) already " "installed\n", @@ -104,18 +104,18 @@ static void intel_init_thermal(struct cpuinfo_x86 *c) h |= (APIC_DM_FIXED | APIC_LVT_MASKED); /* we'll mask till we're ready */ apic_write_around(APIC_LVTTHMR, h); - rdmsr (MSR_IA32_THERM_INTERRUPT, l, h); - wrmsr (MSR_IA32_THERM_INTERRUPT, l | 0x03 , h); + rdmsr(MSR_IA32_THERM_INTERRUPT, l, h); + wrmsr(MSR_IA32_THERM_INTERRUPT, l | 0x03 , h); /* ok we're good to go... */ vendor_thermal_interrupt = intel_thermal_interrupt; - - rdmsr (MSR_IA32_MISC_ENABLE, l, h); - wrmsr (MSR_IA32_MISC_ENABLE, l | (1<<3), h); - l = apic_read (APIC_LVTTHMR); - apic_write_around (APIC_LVTTHMR, l & ~APIC_LVT_MASKED); - printk (KERN_INFO "CPU%d: Thermal monitoring enabled\n", cpu); + rdmsr(MSR_IA32_MISC_ENABLE, l, h); + wrmsr(MSR_IA32_MISC_ENABLE, l | (1<<3), h); + + l = apic_read(APIC_LVTTHMR); + apic_write_around(APIC_LVTTHMR, l & ~APIC_LVT_MASKED); + printk(KERN_INFO "CPU%d: Thermal monitoring enabled\n", cpu); /* enable thermal throttle processing */ atomic_set(&therm_throt_en, 1); @@ -129,28 +129,28 @@ static inline void intel_get_extended_msrs(struct intel_mce_extended_msrs *r) { u32 h; - rdmsr (MSR_IA32_MCG_EAX, r->eax, h); - rdmsr (MSR_IA32_MCG_EBX, r->ebx, h); - rdmsr (MSR_IA32_MCG_ECX, r->ecx, h); - rdmsr (MSR_IA32_MCG_EDX, r->edx, h); - rdmsr (MSR_IA32_MCG_ESI, r->esi, h); - rdmsr (MSR_IA32_MCG_EDI, r->edi, h); - rdmsr (MSR_IA32_MCG_EBP, r->ebp, h); - rdmsr (MSR_IA32_MCG_ESP, r->esp, h); - rdmsr (MSR_IA32_MCG_EFLAGS, r->eflags, h); - rdmsr (MSR_IA32_MCG_EIP, r->eip, h); + rdmsr(MSR_IA32_MCG_EAX, r->eax, h); + rdmsr(MSR_IA32_MCG_EBX, r->ebx, h); + rdmsr(MSR_IA32_MCG_ECX, r->ecx, h); + rdmsr(MSR_IA32_MCG_EDX, r->edx, h); + rdmsr(MSR_IA32_MCG_ESI, r->esi, h); + rdmsr(MSR_IA32_MCG_EDI, r->edi, h); + rdmsr(MSR_IA32_MCG_EBP, r->ebp, h); + rdmsr(MSR_IA32_MCG_ESP, r->esp, h); + rdmsr(MSR_IA32_MCG_EFLAGS, r->eflags, h); + rdmsr(MSR_IA32_MCG_EIP, r->eip, h); } -static void intel_machine_check(struct pt_regs * regs, long error_code) +static void intel_machine_check(struct pt_regs *regs, long error_code) { - int recover=1; + int recover = 1; u32 alow, ahigh, high, low; u32 mcgstl, mcgsth; int i; - rdmsr (MSR_IA32_MCG_STATUS, mcgstl, mcgsth); + rdmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); if (mcgstl & (1<<0)) /* Recoverable ? */ - recover=0; + recover = 0; printk(KERN_EMERG "CPU %d: Machine Check Exception: %08x%08x\n", smp_processor_id(), mcgsth, mcgstl); @@ -191,20 +191,20 @@ static void intel_machine_check(struct pt_regs * regs, long error_code) } if (recover & 2) - panic ("CPU context corrupt"); + panic("CPU context corrupt"); if (recover & 1) - panic ("Unable to continue"); + panic("Unable to continue"); printk(KERN_EMERG "Attempting to continue.\n"); - /* - * Do not clear the MSR_IA32_MCi_STATUS if the error is not + /* + * Do not clear the MSR_IA32_MCi_STATUS if the error is not * recoverable/continuable.This will allow BIOS to look at the MSRs * for errors if the OS could not log the error. */ - for (i=0; i<nr_mce_banks; i++) { + for (i = 0; i < nr_mce_banks; i++) { u32 msr; msr = MSR_IA32_MC0_STATUS+i*4; - rdmsr (msr, low, high); + rdmsr(msr, low, high); if (high&(1<<31)) { /* Clear it */ wrmsr(msr, 0UL, 0UL); @@ -214,7 +214,7 @@ static void intel_machine_check(struct pt_regs * regs, long error_code) } } mcgstl &= ~(1<<2); - wrmsr (MSR_IA32_MCG_STATUS,mcgstl, mcgsth); + wrmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); } @@ -222,30 +222,30 @@ void intel_p4_mcheck_init(struct cpuinfo_x86 *c) { u32 l, h; int i; - + machine_check_vector = intel_machine_check; wmb(); - printk (KERN_INFO "Intel machine check architecture supported.\n"); - rdmsr (MSR_IA32_MCG_CAP, l, h); + printk(KERN_INFO "Intel machine check architecture supported.\n"); + rdmsr(MSR_IA32_MCG_CAP, l, h); if (l & (1<<8)) /* Control register present ? */ - wrmsr (MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); + wrmsr(MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); nr_mce_banks = l & 0xff; - for (i=0; i<nr_mce_banks; i++) { - wrmsr (MSR_IA32_MC0_CTL+4*i, 0xffffffff, 0xffffffff); - wrmsr (MSR_IA32_MC0_STATUS+4*i, 0x0, 0x0); + for (i = 0; i < nr_mce_banks; i++) { + wrmsr(MSR_IA32_MC0_CTL+4*i, 0xffffffff, 0xffffffff); + wrmsr(MSR_IA32_MC0_STATUS+4*i, 0x0, 0x0); } - set_in_cr4 (X86_CR4_MCE); - printk (KERN_INFO "Intel machine check reporting enabled on CPU#%d.\n", + set_in_cr4(X86_CR4_MCE); + printk(KERN_INFO "Intel machine check reporting enabled on CPU#%d.\n", smp_processor_id()); /* Check for P4/Xeon extended MCE MSRs */ - rdmsr (MSR_IA32_MCG_CAP, l, h); + rdmsr(MSR_IA32_MCG_CAP, l, h); if (l & (1<<9)) {/* MCG_EXT_P */ mce_num_extended_msrs = (l >> 16) & 0xff; - printk (KERN_INFO "CPU%d: Intel P4/Xeon Extended MCE MSRs (%d)" + printk(KERN_INFO "CPU%d: Intel P4/Xeon Extended MCE MSRs (%d)" " available\n", smp_processor_id(), mce_num_extended_msrs); diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 5d241ce..509bd3d 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -37,7 +37,7 @@ static struct fixed_range_block fixed_range_blocks[] = { static unsigned long smp_changes_mask; static struct mtrr_state mtrr_state = {}; static int mtrr_state_set; -static u64 tom2; +u64 mtrr_tom2; #undef MODULE_PARAM_PREFIX #define MODULE_PARAM_PREFIX "mtrr." @@ -139,8 +139,8 @@ u8 mtrr_type_lookup(u64 start, u64 end) } } - if (tom2) { - if (start >= (1ULL<<32) && (end < tom2)) + if (mtrr_tom2) { + if (start >= (1ULL<<32) && (end < mtrr_tom2)) return MTRR_TYPE_WRBACK; } @@ -158,6 +158,20 @@ get_mtrr_var_range(unsigned int index, struct mtrr_var_range *vr) rdmsr(MTRRphysMask_MSR(index), vr->mask_lo, vr->mask_hi); } +/* fill the MSR pair relating to a var range */ +void fill_mtrr_var_range(unsigned int index, + u32 base_lo, u32 base_hi, u32 mask_lo, u32 mask_hi) +{ + struct mtrr_var_range *vr; + + vr = mtrr_state.var_ranges; + + vr[index].base_lo = base_lo; + vr[index].base_hi = base_hi; + vr[index].mask_lo = mask_lo; + vr[index].mask_hi = mask_hi; +} + static void get_fixed_ranges(mtrr_type * frs) { @@ -213,13 +227,13 @@ void __init get_mtrr_state(void) mtrr_state.enabled = (lo & 0xc00) >> 10; if (amd_special_default_mtrr()) { - unsigned lo, hi; + unsigned low, high; /* TOP_MEM2 */ - rdmsr(MSR_K8_TOP_MEM2, lo, hi); - tom2 = hi; - tom2 <<= 32; - tom2 |= lo; - tom2 &= 0xffffff8000000ULL; + rdmsr(MSR_K8_TOP_MEM2, low, high); + mtrr_tom2 = high; + mtrr_tom2 <<= 32; + mtrr_tom2 |= low; + mtrr_tom2 &= 0xffffff800000ULL; } if (mtrr_show) { int high_width; @@ -251,9 +265,9 @@ void __init get_mtrr_state(void) else printk(KERN_INFO "MTRR %u disabled\n", i); } - if (tom2) { + if (mtrr_tom2) { printk(KERN_INFO "TOM2: %016llx aka %lldM\n", - tom2, tom2>>20); + mtrr_tom2, mtrr_tom2>>20); } } mtrr_state_set = 1; @@ -328,7 +342,7 @@ static void set_fixed_range(int msr, bool *changed, unsigned int *msrwords) if (lo != msrwords[0] || hi != msrwords[1]) { if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD && - boot_cpu_data.x86 == 15 && + (boot_cpu_data.x86 >= 0x0f && boot_cpu_data.x86 <= 0x11) && ((msrwords[0] | msrwords[1]) & K8_MTRR_RDMEM_WRMEM_MASK)) k8_enable_fixed_iorrs(); mtrr_wrmsr(msr, msrwords[0], msrwords[1]); diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c index 290652c..6f23969 100644 --- a/arch/x86/kernel/cpu/mtrr/main.c +++ b/arch/x86/kernel/cpu/mtrr/main.c @@ -37,6 +37,7 @@ #include <linux/smp.h> #include <linux/cpu.h> #include <linux/mutex.h> +#include <linux/sort.h> #include <asm/e820.h> #include <asm/mtrr.h> @@ -609,6 +610,787 @@ static struct sysdev_driver mtrr_sysdev_driver = { .resume = mtrr_restore, }; +/* should be related to MTRR_VAR_RANGES nums */ +#define RANGE_NUM 256 + +struct res_range { + unsigned long start; + unsigned long end; +}; + +static int __init +add_range(struct res_range *range, int nr_range, unsigned long start, + unsigned long end) +{ + /* out of slots */ + if (nr_range >= RANGE_NUM) + return nr_range; + + range[nr_range].start = start; + range[nr_range].end = end; + + nr_range++; + + return nr_range; +} + +static int __init +add_range_with_merge(struct res_range *range, int nr_range, unsigned long start, + unsigned long end) +{ + int i; + + /* try to merge it with old one */ + for (i = 0; i < nr_range; i++) { + unsigned long final_start, final_end; + unsigned long common_start, common_end; + + if (!range[i].end) + continue; + + common_start = max(range[i].start, start); + common_end = min(range[i].end, end); + if (common_start > common_end + 1) + continue; + + final_start = min(range[i].start, start); + final_end = max(range[i].end, end); + + range[i].start = final_start; + range[i].end = final_end; + return nr_range; + } + + /* need to add that */ + return add_range(range, nr_range, start, end); +} + +static void __init +subtract_range(struct res_range *range, unsigned long start, unsigned long end) +{ + int i, j; + + for (j = 0; j < RANGE_NUM; j++) { + if (!range[j].end) + continue; + + if (start <= range[j].start && end >= range[j].end) { + range[j].start = 0; + range[j].end = 0; + continue; + } + + if (start <= range[j].start && end < range[j].end && + range[j].start < end + 1) { + range[j].start = end + 1; + continue; + } + + + if (start > range[j].start && end >= range[j].end && + range[j].end > start - 1) { + range[j].end = start - 1; + continue; + } + + if (start > range[j].start && end < range[j].end) { + /* find the new spare */ + for (i = 0; i < RANGE_NUM; i++) { + if (range[i].end == 0) + break; + } + if (i < RANGE_NUM) { + range[i].end = range[j].end; + range[i].start = end + 1; + } else { + printk(KERN_ERR "run of slot in ranges\n"); + } + range[j].end = start - 1; + continue; + } + } +} + +static int __init cmp_range(const void *x1, const void *x2) +{ + const struct res_range *r1 = x1; + const struct res_range *r2 = x2; + long start1, start2; + + start1 = r1->start; + start2 = r2->start; + + return start1 - start2; +} + +struct var_mtrr_range_state { + unsigned long base_pfn; + unsigned long size_pfn; + mtrr_type type; +}; + +struct var_mtrr_range_state __initdata range_state[RANGE_NUM]; +static int __initdata debug_print; + +static int __init +x86_get_mtrr_mem_range(struct res_range *range, int nr_range, + unsigned long extra_remove_base, + unsigned long extra_remove_size) +{ + unsigned long i, base, size; + mtrr_type type; + + for (i = 0; i < num_var_ranges; i++) { + type = range_state[i].type; + if (type != MTRR_TYPE_WRBACK) + continue; + base = range_state[i].base_pfn; + size = range_state[i].size_pfn; + nr_range = add_range_with_merge(range, nr_range, base, + base + size - 1); + } + if (debug_print) { + printk(KERN_DEBUG "After WB checking\n"); + for (i = 0; i < nr_range; i++) + printk(KERN_DEBUG "MTRR MAP PFN: %016lx - %016lx\n", + range[i].start, range[i].end + 1); + } + + /* take out UC ranges */ + for (i = 0; i < num_var_ranges; i++) { + type = range_state[i].type; + if (type != MTRR_TYPE_UNCACHABLE) + continue; + size = range_state[i].size_pfn; + if (!size) + continue; + base = range_state[i].base_pfn; + subtract_range(range, base, base + size - 1); + } + if (extra_remove_size) + subtract_range(range, extra_remove_base, + extra_remove_base + extra_remove_size - 1); + + /* get new range num */ + nr_range = 0; + for (i = 0; i < RANGE_NUM; i++) { + if (!range[i].end) + continue; + nr_range++; + } + if (debug_print) { + printk(KERN_DEBUG "After UC checking\n"); + for (i = 0; i < nr_range; i++) + printk(KERN_DEBUG "MTRR MAP PFN: %016lx - %016lx\n", + range[i].start, range[i].end + 1); + } + + /* sort the ranges */ + sort(range, nr_range, sizeof(struct res_range), cmp_range, NULL); + if (debug_print) { + printk(KERN_DEBUG "After sorting\n"); + for (i = 0; i < nr_range; i++) + printk(KERN_DEBUG "MTRR MAP PFN: %016lx - %016lx\n", + range[i].start, range[i].end + 1); + } + + /* clear those is not used */ + for (i = nr_range; i < RANGE_NUM; i++) + memset(&range[i], 0, sizeof(range[i])); + + return nr_range; +} + +static struct res_range __initdata range[RANGE_NUM]; + +#ifdef CONFIG_MTRR_SANITIZER + +static unsigned long __init sum_ranges(struct res_range *range, int nr_range) +{ + unsigned long sum; + int i; + + sum = 0; + for (i = 0; i < nr_range; i++) + sum += range[i].end + 1 - range[i].start; + + return sum; +} + +static int enable_mtrr_cleanup __initdata = + CONFIG_MTRR_SANITIZER_ENABLE_DEFAULT; + +static int __init disable_mtrr_cleanup_setup(char *str) +{ + if (enable_mtrr_cleanup != -1) + enable_mtrr_cleanup = 0; + return 0; +} +early_param("disable_mtrr_cleanup", disable_mtrr_cleanup_setup); + +static int __init enable_mtrr_cleanup_setup(char *str) +{ + if (enable_mtrr_cleanup != -1) + enable_mtrr_cleanup = 1; + return 0; +} +early_param("enble_mtrr_cleanup", enable_mtrr_cleanup_setup); + +struct var_mtrr_state { + unsigned long range_startk; + unsigned long range_sizek; + unsigned long chunk_sizek; + unsigned long gran_sizek; + unsigned int reg; +}; + +static void __init +set_var_mtrr(unsigned int reg, unsigned long basek, unsigned long sizek, + unsigned char type, unsigned int address_bits) +{ + u32 base_lo, base_hi, mask_lo, mask_hi; + u64 base, mask; + + if (!sizek) { + fill_mtrr_var_range(reg, 0, 0, 0, 0); + return; + } + + mask = (1ULL << address_bits) - 1; + mask &= ~((((u64)sizek) << 10) - 1); + + base = ((u64)basek) << 10; + + base |= type; + mask |= 0x800; + + base_lo = base & ((1ULL<<32) - 1); + base_hi = base >> 32; + + mask_lo = mask & ((1ULL<<32) - 1); + mask_hi = mask >> 32; + + fill_mtrr_var_range(reg, base_lo, base_hi, mask_lo, mask_hi); +} + +static void __init +save_var_mtrr(unsigned int reg, unsigned long basek, unsigned long sizek, + unsigned char type) +{ + range_state[reg].base_pfn = basek >> (PAGE_SHIFT - 10); + range_state[reg].size_pfn = sizek >> (PAGE_SHIFT - 10); + range_state[reg].type = type; +} + +static void __init +set_var_mtrr_all(unsigned int address_bits) +{ + unsigned long basek, sizek; + unsigned char type; + unsigned int reg; + + for (reg = 0; reg < num_var_ranges; reg++) { + basek = range_state[reg].base_pfn << (PAGE_SHIFT - 10); + sizek = range_state[reg].size_pfn << (PAGE_SHIFT - 10); + type = range_state[reg].type; + + set_var_mtrr(reg, basek, sizek, type, address_bits); + } +} + +static unsigned int __init +range_to_mtrr(unsigned int reg, unsigned long range_startk, + unsigned long range_sizek, unsigned char type) +{ + if (!range_sizek || (reg >= num_var_ranges)) + return reg; + + while (range_sizek) { + unsigned long max_align, align; + unsigned long sizek; + + /* Compute the maximum size I can make a range */ + if (range_startk) + max_align = ffs(range_startk) - 1; + else + max_align = 32; + align = fls(range_sizek) - 1; + if (align > max_align) + align = max_align; + + sizek = 1 << align; + if (debug_print) + printk(KERN_DEBUG "Setting variable MTRR %d, " + "base: %ldMB, range: %ldMB, type %s\n", + reg, range_startk >> 10, sizek >> 10, + (type == MTRR_TYPE_UNCACHABLE)?"UC": + ((type == MTRR_TYPE_WRBACK)?"WB":"Other") + ); + save_var_mtrr(reg++, range_startk, sizek, type); + range_startk += sizek; + range_sizek -= sizek; + if (reg >= num_var_ranges) + break; + } + return reg; +} + +static unsigned __init +range_to_mtrr_with_hole(struct var_mtrr_state *state, unsigned long basek, + unsigned long sizek) +{ + unsigned long hole_basek, hole_sizek; + unsigned long second_basek, second_sizek; + unsigned long range0_basek, range0_sizek; + unsigned long range_basek, range_sizek; + unsigned long chunk_sizek; + unsigned long gran_sizek; + + hole_basek = 0; + hole_sizek = 0; + second_basek = 0; + second_sizek = 0; + chunk_sizek = state->chunk_sizek; + gran_sizek = state->gran_sizek; + + /* align with gran size, prevent small block used up MTRRs */ + range_basek = ALIGN(state->range_startk, gran_sizek); + if ((range_basek > basek) && basek) + return second_sizek; + state->range_sizek -= (range_basek - state->range_startk); + range_sizek = ALIGN(state->range_sizek, gran_sizek); + + while (range_sizek > state->range_sizek) { + range_sizek -= gran_sizek; + if (!range_sizek) + return 0; + } + state->range_sizek = range_sizek; + + /* try to append some small hole */ + range0_basek = state->range_startk; + range0_sizek = ALIGN(state->range_sizek, chunk_sizek); + if (range0_sizek == state->range_sizek) { + if (debug_print) + printk(KERN_DEBUG "rangeX: %016lx - %016lx\n", + range0_basek<<10, + (range0_basek + state->range_sizek)<<10); + state->reg = range_to_mtrr(state->reg, range0_basek, + state->range_sizek, MTRR_TYPE_WRBACK); + return 0; + } + + range0_sizek -= chunk_sizek; + if (range0_sizek && sizek) { + while (range0_basek + range0_sizek > (basek + sizek)) { + range0_sizek -= chunk_sizek; + if (!range0_sizek) + break; + } + } + + if (range0_sizek) { + if (debug_print) + printk(KERN_DEBUG "range0: %016lx - %016lx\n", + range0_basek<<10, + (range0_basek + range0_sizek)<<10); + state->reg = range_to_mtrr(state->reg, range0_basek, + range0_sizek, MTRR_TYPE_WRBACK); + + } + + range_basek = range0_basek + range0_sizek; + range_sizek = chunk_sizek; + + if (range_basek + range_sizek > basek && + range_basek + range_sizek <= (basek + sizek)) { + /* one hole */ + second_basek = basek; + second_sizek = range_basek + range_sizek - basek; + } + + /* if last piece, only could one hole near end */ + if ((second_basek || !basek) && + range_sizek - (state->range_sizek - range0_sizek) - second_sizek < + (chunk_sizek >> 1)) { + /* + * one hole in middle (second_sizek is 0) or at end + * (second_sizek is 0 ) + */ + hole_sizek = range_sizek - (state->range_sizek - range0_sizek) + - second_sizek; + hole_basek = range_basek + range_sizek - hole_sizek + - second_sizek; + } else { + /* fallback for big hole, or several holes */ + range_sizek = state->range_sizek - range0_sizek; + second_basek = 0; + second_sizek = 0; + } + + if (debug_print) + printk(KERN_DEBUG "range: %016lx - %016lx\n", range_basek<<10, + (range_basek + range_sizek)<<10); + state->reg = range_to_mtrr(state->reg, range_basek, range_sizek, + MTRR_TYPE_WRBACK); + if (hole_sizek) { + if (debug_print) + printk(KERN_DEBUG "hole: %016lx - %016lx\n", + hole_basek<<10, (hole_basek + hole_sizek)<<10); + state->reg = range_to_mtrr(state->reg, hole_basek, hole_sizek, + MTRR_TYPE_UNCACHABLE); + + } + + return second_sizek; +} + +static void __init +set_var_mtrr_range(struct var_mtrr_state *state, unsigned long base_pfn, + unsigned long size_pfn) +{ + unsigned long basek, sizek; + unsigned long second_sizek = 0; + + if (state->reg >= num_var_ranges) + return; + + basek = base_pfn << (PAGE_SHIFT - 10); + sizek = size_pfn << (PAGE_SHIFT - 10); + + /* See if I can merge with the last range */ + if ((basek <= 1024) || + (state->range_startk + state->range_sizek == basek)) { + unsigned long endk = basek + sizek; + state->range_sizek = endk - state->range_startk; + return; + } + /* Write the range mtrrs */ + if (state->range_sizek != 0) + second_sizek = range_to_mtrr_with_hole(state, basek, sizek); + + /* Allocate an msr */ + state->range_startk = basek + second_sizek; + state->range_sizek = sizek - second_sizek; +} + +/* mininum size of mtrr block that can take hole */ +static u64 mtrr_chunk_size __initdata = (256ULL<<20); + +static int __init parse_mtrr_chunk_size_opt(char *p) +{ + if (!p) + return -EINVAL; + mtrr_chunk_size = memparse(p, &p); + return 0; +} +early_param("mtrr_chunk_size", parse_mtrr_chunk_size_opt); + +/* granity of mtrr of block */ +static u64 mtrr_gran_size __initdata; + +static int __init parse_mtrr_gran_size_opt(char *p) +{ + if (!p) + return -EINVAL; + mtrr_gran_size = memparse(p, &p); + return 0; +} +early_param("mtrr_gran_size", parse_mtrr_gran_size_opt); + +static int nr_mtrr_spare_reg __initdata = + CONFIG_MTRR_SANITIZER_SPARE_REG_NR_DEFAULT; + +static int __init parse_mtrr_spare_reg(char *arg) +{ + if (arg) + nr_mtrr_spare_reg = simple_strtoul(arg, NULL, 0); + return 0; +} + +early_param("mtrr_spare_reg_nr", parse_mtrr_spare_reg); + +static int __init +x86_setup_var_mtrrs(struct res_range *range, int nr_range, + u64 chunk_size, u64 gran_size) +{ + struct var_mtrr_state var_state; + int i; + int num_reg; + + var_state.range_startk = 0; + var_state.range_sizek = 0; + var_state.reg = 0; + var_state.chunk_sizek = chunk_size >> 10; + var_state.gran_sizek = gran_size >> 10; + + memset(range_state, 0, sizeof(range_state)); + + /* Write the range etc */ + for (i = 0; i < nr_range; i++) + set_var_mtrr_range(&var_state, range[i].start, + range[i].end - range[i].start + 1); + + /* Write the last range */ + if (var_state.range_sizek != 0) + range_to_mtrr_with_hole(&var_state, 0, 0); + + num_reg = var_state.reg; + /* Clear out the extra MTRR's */ + while (var_state.reg < num_var_ranges) { + save_var_mtrr(var_state.reg, 0, 0, 0); + var_state.reg++; + } + + return num_reg; +} + +struct mtrr_cleanup_result { + unsigned long gran_sizek; + unsigned long chunk_sizek; + unsigned long lose_cover_sizek; + unsigned int num_reg; + int bad; +}; + +/* + * gran_size: 1M, 2M, ..., 2G + * chunk size: gran_size, ..., 4G + * so we need (2+13)*6 + */ +#define NUM_RESULT 90 +#define PSHIFT (PAGE_SHIFT - 10) + +static struct mtrr_cleanup_result __initdata result[NUM_RESULT]; +static struct res_range __initdata range_new[RANGE_NUM]; +static unsigned long __initdata min_loss_pfn[RANGE_NUM]; + +static int __init mtrr_cleanup(unsigned address_bits) +{ + unsigned long extra_remove_base, extra_remove_size; + unsigned long i, base, size, def, dummy; + mtrr_type type; + int nr_range, nr_range_new; + u64 chunk_size, gran_size; + unsigned long range_sums, range_sums_new; + int index_good; + int num_reg_good; + + /* extra one for all 0 */ + int num[MTRR_NUM_TYPES + 1]; + + if (!is_cpu(INTEL) || enable_mtrr_cleanup < 1) + return 0; + rdmsr(MTRRdefType_MSR, def, dummy); + def &= 0xff; + if (def != MTRR_TYPE_UNCACHABLE) + return 0; + + /* get it and store it aside */ + memset(range_state, 0, sizeof(range_state)); + for (i = 0; i < num_var_ranges; i++) { + mtrr_if->get(i, &base, &size, &type); + range_state[i].base_pfn = base; + range_state[i].size_pfn = size; + range_state[i].type = type; + } + + /* check entries number */ + memset(num, 0, sizeof(num)); + for (i = 0; i < num_var_ranges; i++) { + type = range_state[i].type; + size = range_state[i].size_pfn; + if (type >= MTRR_NUM_TYPES) + continue; + if (!size) + type = MTRR_NUM_TYPES; + num[type]++; + } + + /* check if we got UC entries */ + if (!num[MTRR_TYPE_UNCACHABLE]) + return 0; + + /* check if we only had WB and UC */ + if (num[MTRR_TYPE_WRBACK] + num[MTRR_TYPE_UNCACHABLE] != + num_var_ranges - num[MTRR_NUM_TYPES]) + return 0; + + memset(range, 0, sizeof(range)); + extra_remove_size = 0; + if (mtrr_tom2) { + extra_remove_base = 1 << (32 - PAGE_SHIFT); + extra_remove_size = + (mtrr_tom2 >> PAGE_SHIFT) - extra_remove_base; + } + nr_range = x86_get_mtrr_mem_range(range, 0, extra_remove_base, + extra_remove_size); + range_sums = sum_ranges(range, nr_range); + printk(KERN_INFO "total RAM coverred: %ldM\n", + range_sums >> (20 - PAGE_SHIFT)); + + if (mtrr_chunk_size && mtrr_gran_size) { + int num_reg; + + debug_print = 1; + /* convert ranges to var ranges state */ + num_reg = x86_setup_var_mtrrs(range, nr_range, mtrr_chunk_size, + mtrr_gran_size); + + /* we got new setting in range_state, check it */ + memset(range_new, 0, sizeof(range_new)); + nr_range_new = x86_get_mtrr_mem_range(range_new, 0, + extra_remove_base, + extra_remove_size); + range_sums_new = sum_ranges(range_new, nr_range_new); + + i = 0; + result[i].chunk_sizek = mtrr_chunk_size >> 10; + result[i].gran_sizek = mtrr_gran_size >> 10; + result[i].num_reg = num_reg; + if (range_sums < range_sums_new) { + result[i].lose_cover_sizek = + (range_sums_new - range_sums) << PSHIFT; + result[i].bad = 1; + } else + result[i].lose_cover_sizek = + (range_sums - range_sums_new) << PSHIFT; + + printk(KERN_INFO "%sgran_size: %ldM \tchunk_size: %ldM \t", + result[i].bad?"*BAD*":" ", result[i].gran_sizek >> 10, + result[i].chunk_sizek >> 10); + printk(KERN_CONT "num_reg: %d \tlose cover RAM: %s%ldM \n", + result[i].num_reg, result[i].bad?"-":"", + result[i].lose_cover_sizek >> 10); + if (!result[i].bad) { + set_var_mtrr_all(address_bits); + return 1; + } + printk(KERN_INFO "invalid mtrr_gran_size or mtrr_chunk_size, " + "will find optimal one\n"); + debug_print = 0; + memset(result, 0, sizeof(result[0])); + } + + i = 0; + memset(min_loss_pfn, 0xff, sizeof(min_loss_pfn)); + memset(result, 0, sizeof(result)); + for (gran_size = (1ULL<<20); gran_size < (1ULL<<32); gran_size <<= 1) { + for (chunk_size = gran_size; chunk_size < (1ULL<<33); + chunk_size <<= 1) { + int num_reg; + + if (debug_print) + printk(KERN_INFO + "\ngran_size: %lldM chunk_size_size: %lldM\n", + gran_size >> 20, chunk_size >> 20); + if (i >= NUM_RESULT) + continue; + + /* convert ranges to var ranges state */ + num_reg = x86_setup_var_mtrrs(range, nr_range, + chunk_size, gran_size); + + /* we got new setting in range_state, check it */ + memset(range_new, 0, sizeof(range_new)); + nr_range_new = x86_get_mtrr_mem_range(range_new, 0, + extra_remove_base, extra_remove_size); + range_sums_new = sum_ranges(range_new, nr_range_new); + + result[i].chunk_sizek = chunk_size >> 10; + result[i].gran_sizek = gran_size >> 10; + result[i].num_reg = num_reg; + if (range_sums < range_sums_new) { + result[i].lose_cover_sizek = + (range_sums_new - range_sums) << PSHIFT; + result[i].bad = 1; + } else + result[i].lose_cover_sizek = + (range_sums - range_sums_new) << PSHIFT; + + /* double check it */ + if (!result[i].bad && !result[i].lose_cover_sizek) { + if (nr_range_new != nr_range || + memcmp(range, range_new, sizeof(range))) + result[i].bad = 1; + } + + if (!result[i].bad && (range_sums - range_sums_new < + min_loss_pfn[num_reg])) { + min_loss_pfn[num_reg] = + range_sums - range_sums_new; + } + i++; + } + } + + /* print out all */ + for (i = 0; i < NUM_RESULT; i++) { + printk(KERN_INFO "%sgran_size: %ldM \tchunk_size: %ldM \t", + result[i].bad?"*BAD* ":" ", result[i].gran_sizek >> 10, + result[i].chunk_sizek >> 10); + printk(KERN_CONT "num_reg: %d \tlose RAM: %s%ldM\n", + result[i].num_reg, result[i].bad?"-":"", + result[i].lose_cover_sizek >> 10); + } + + /* try to find the optimal index */ + if (nr_mtrr_spare_reg >= num_var_ranges) + nr_mtrr_spare_reg = num_var_ranges - 1; + num_reg_good = -1; + for (i = num_var_ranges - nr_mtrr_spare_reg; i > 0; i--) { + if (!min_loss_pfn[i]) { + num_reg_good = i; + break; + } + } + + index_good = -1; + if (num_reg_good != -1) { + for (i = 0; i < NUM_RESULT; i++) { + if (!result[i].bad && + result[i].num_reg == num_reg_good && + !result[i].lose_cover_sizek) { + index_good = i; + break; + } + } + } + + if (index_good != -1) { + printk(KERN_INFO "Found optimal setting for mtrr clean up\n"); + i = index_good; + printk(KERN_INFO "gran_size: %ldM \tchunk_size: %ldM \t", + result[i].gran_sizek >> 10, + result[i].chunk_sizek >> 10); + printk(KERN_CONT "num_reg: %d \tlose RAM: %ldM\n", + result[i].num_reg, + result[i].lose_cover_sizek >> 10); + /* convert ranges to var ranges state */ + chunk_size = result[i].chunk_sizek; + chunk_size <<= 10; + gran_size = result[i].gran_sizek; + gran_size <<= 10; + debug_print = 1; + x86_setup_var_mtrrs(range, nr_range, chunk_size, gran_size); + set_var_mtrr_all(address_bits); + return 1; + } + + printk(KERN_INFO "mtrr_cleanup: can not find optimal value\n"); + printk(KERN_INFO "please specify mtrr_gran_size/mtrr_chunk_size\n"); + + return 0; +} +#else +static int __init mtrr_cleanup(unsigned address_bits) +{ + return 0; +} +#endif + +static int __initdata changed_by_mtrr_cleanup; + static int disable_mtrr_trim; static int __init disable_mtrr_trim_setup(char *str) @@ -648,6 +1430,19 @@ int __init amd_special_default_mtrr(void) return 0; } +static u64 __init real_trim_memory(unsigned long start_pfn, + unsigned long limit_pfn) +{ + u64 trim_start, trim_size; + trim_start = start_pfn; + trim_start <<= PAGE_SHIFT; + trim_size = limit_pfn; + trim_size <<= PAGE_SHIFT; + trim_size -= trim_start; + + return e820_update_range(trim_start, trim_size, E820_RAM, + E820_RESERVED); +} /** * mtrr_trim_uncached_memory - trim RAM not covered by MTRRs * @end_pfn: ending page frame number @@ -663,8 +1458,11 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn) { unsigned long i, base, size, highest_pfn = 0, def, dummy; mtrr_type type; - u64 trim_start, trim_size; + int nr_range; + u64 total_trim_size; + /* extra one for all 0 */ + int num[MTRR_NUM_TYPES + 1]; /* * Make sure we only trim uncachable memory on machines that * support the Intel MTRR architecture: @@ -676,14 +1474,22 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn) if (def != MTRR_TYPE_UNCACHABLE) return 0; - if (amd_special_default_mtrr()) - return 0; + /* get it and store it aside */ + memset(range_state, 0, sizeof(range_state)); + for (i = 0; i < num_var_ranges; i++) { + mtrr_if->get(i, &base, &size, &type); + range_state[i].base_pfn = base; + range_state[i].size_pfn = size; + range_state[i].type = type; + } /* Find highest cached pfn */ for (i = 0; i < num_var_ranges; i++) { - mtrr_if->get(i, &base, &size, &type); + type = range_state[i].type; if (type != MTRR_TYPE_WRBACK) continue; + base = range_state[i].base_pfn; + size = range_state[i].size_pfn; if (highest_pfn < base + size) highest_pfn = base + size; } @@ -698,22 +1504,65 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn) return 0; } - if (highest_pfn < end_pfn) { + /* check entries number */ + memset(num, 0, sizeof(num)); + for (i = 0; i < num_var_ranges; i++) { + type = range_state[i].type; + if (type >= MTRR_NUM_TYPES) + continue; + size = range_state[i].size_pfn; + if (!size) + type = MTRR_NUM_TYPES; + num[type]++; + } + + /* no entry for WB? */ + if (!num[MTRR_TYPE_WRBACK]) + return 0; + + /* check if we only had WB and UC */ + if (num[MTRR_TYPE_WRBACK] + num[MTRR_TYPE_UNCACHABLE] != + num_var_ranges - num[MTRR_NUM_TYPES]) + return 0; + + memset(range, 0, sizeof(range)); + nr_range = 0; + if (mtrr_tom2) { + range[nr_range].start = (1ULL<<(32 - PAGE_SHIFT)); + range[nr_range].end = (mtrr_tom2 >> PAGE_SHIFT) - 1; + if (highest_pfn < range[nr_range].end + 1) + highest_pfn = range[nr_range].end + 1; + nr_range++; + } + nr_range = x86_get_mtrr_mem_range(range, nr_range, 0, 0); + + total_trim_size = 0; + /* check the head */ + if (range[0].start) + total_trim_size += real_trim_memory(0, range[0].start); + /* check the holes */ + for (i = 0; i < nr_range - 1; i++) { + if (range[i].end + 1 < range[i+1].start) + total_trim_size += real_trim_memory(range[i].end + 1, + range[i+1].start); + } + /* check the top */ + i = nr_range - 1; + if (range[i].end + 1 < end_pfn) + total_trim_size += real_trim_memory(range[i].end + 1, + end_pfn); + + if (total_trim_size) { printk(KERN_WARNING "WARNING: BIOS bug: CPU MTRRs don't cover" - " all of memory, losing %luMB of RAM.\n", - (end_pfn - highest_pfn) >> (20 - PAGE_SHIFT)); + " all of memory, losing %lluMB of RAM.\n", + total_trim_size >> 20); - WARN_ON(1); + if (!changed_by_mtrr_cleanup) + WARN_ON(1); printk(KERN_INFO "update e820 for mtrr\n"); - trim_start = highest_pfn; - trim_start <<= PAGE_SHIFT; - trim_size = end_pfn; - trim_size <<= PAGE_SHIFT; - trim_size -= trim_start; - update_memory_range(trim_start, trim_size, E820_RAM, - E820_RESERVED); update_e820(); + return 1; } @@ -729,18 +1578,21 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn) */ void __init mtrr_bp_init(void) { + u32 phys_addr; init_ifs(); + phys_addr = 32; + if (cpu_has_mtrr) { mtrr_if = &generic_mtrr_ops; size_or_mask = 0xff000000; /* 36 bits */ size_and_mask = 0x00f00000; + phys_addr = 36; /* This is an AMD specific MSR, but we assume(hope?) that Intel will implement it to when they extend the address bus of the Xeon. */ if (cpuid_eax(0x80000000) >= 0x80000008) { - u32 phys_addr; phys_addr = cpuid_eax(0x80000008) & 0xff; /* CPUID workaround for Intel 0F33/0F34 CPU */ if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL && @@ -758,6 +1610,7 @@ void __init mtrr_bp_init(void) don't support PAE */ size_or_mask = 0xfff00000; /* 32 bits */ size_and_mask = 0; + phys_addr = 32; } } else { switch (boot_cpu_data.x86_vendor) { @@ -791,8 +1644,15 @@ void __init mtrr_bp_init(void) if (mtrr_if) { set_num_var_ranges(); init_table(); - if (use_intel()) + if (use_intel()) { get_mtrr_state(); + + if (mtrr_cleanup(phys_addr)) { + changed_by_mtrr_cleanup = 1; + mtrr_if->set_all(); + } + + } } } @@ -829,9 +1689,10 @@ static int __init mtrr_init_finialize(void) { if (!mtrr_if) return 0; - if (use_intel()) - mtrr_state_warn(); - else { + if (use_intel()) { + if (!changed_by_mtrr_cleanup) + mtrr_state_warn(); + } else { /* The CPUs haven't MTRR and seem to not support SMP. They have * specific drivers, we use a tricky method to support * suspend/resume for them. diff --git a/arch/x86/kernel/cpu/mtrr/mtrr.h b/arch/x86/kernel/cpu/mtrr/mtrr.h index 2cc77eb..2dc4ec6 100644 --- a/arch/x86/kernel/cpu/mtrr/mtrr.h +++ b/arch/x86/kernel/cpu/mtrr/mtrr.h @@ -81,6 +81,8 @@ void set_mtrr_done(struct set_mtrr_context *ctxt); void set_mtrr_cache_disable(struct set_mtrr_context *ctxt); void set_mtrr_prepare_save(struct set_mtrr_context *ctxt); +void fill_mtrr_var_range(unsigned int index, + u32 base_lo, u32 base_hi, u32 mask_lo, u32 mask_hi); void get_mtrr_state(void); extern void set_mtrr_ops(struct mtrr_ops * ops); @@ -92,6 +94,7 @@ extern struct mtrr_ops * mtrr_if; #define use_intel() (mtrr_if && mtrr_if->use_intel_if == 1) extern unsigned int num_var_ranges; +extern u64 mtrr_tom2; void mtrr_state_warn(void); const char *mtrr_attrib_to_str(int x); diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c index 58043f0..6d4bdc0 100644 --- a/arch/x86/kernel/cpu/perfctr-watchdog.c +++ b/arch/x86/kernel/cpu/perfctr-watchdog.c @@ -1,11 +1,15 @@ -/* local apic based NMI watchdog for various CPUs. - This file also handles reservation of performance counters for coordination - with other users (like oprofile). - - Note that these events normally don't tick when the CPU idles. This means - the frequency varies with CPU load. - - Original code for K7/P6 written by Keith Owens */ +/* + * local apic based NMI watchdog for various CPUs. + * + * This file also handles reservation of performance counters for coordination + * with other users (like oprofile). + * + * Note that these events normally don't tick when the CPU idles. This means + * the frequency varies with CPU load. + * + * Original code for K7/P6 written by Keith Owens + * + */ #include <linux/percpu.h> #include <linux/module.h> @@ -36,12 +40,16 @@ struct wd_ops { static const struct wd_ops *wd_ops; -/* this number is calculated from Intel's MSR_P4_CRU_ESCR5 register and it's - * offset from MSR_P4_BSU_ESCR0. It will be the max for all platforms (for now) +/* + * this number is calculated from Intel's MSR_P4_CRU_ESCR5 register and it's + * offset from MSR_P4_BSU_ESCR0. + * + * It will be the max for all platforms (for now) */ #define NMI_MAX_COUNTER_BITS 66 -/* perfctr_nmi_owner tracks the ownership of the perfctr registers: +/* + * perfctr_nmi_owner tracks the ownership of the perfctr registers: * evtsel_nmi_owner tracks the ownership of the event selection * - different performance counters/ event selection may be reserved for * different subsystems this reservation system just tries to coordinate @@ -73,8 +81,10 @@ static inline unsigned int nmi_perfctr_msr_to_bit(unsigned int msr) return 0; } -/* converts an msr to an appropriate reservation bit */ -/* returns the bit offset of the event selection register */ +/* + * converts an msr to an appropriate reservation bit + * returns the bit offset of the event selection register + */ static inline unsigned int nmi_evntsel_msr_to_bit(unsigned int msr) { /* returns the bit offset of the event selection register */ @@ -114,6 +124,7 @@ int avail_to_resrv_perfctr_nmi(unsigned int msr) return (!test_bit(counter, perfctr_nmi_owner)); } +EXPORT_SYMBOL(avail_to_resrv_perfctr_nmi_bit); int reserve_perfctr_nmi(unsigned int msr) { @@ -128,6 +139,7 @@ int reserve_perfctr_nmi(unsigned int msr) return 1; return 0; } +EXPORT_SYMBOL(reserve_perfctr_nmi); void release_perfctr_nmi(unsigned int msr) { @@ -140,6 +152,7 @@ void release_perfctr_nmi(unsigned int msr) clear_bit(counter, perfctr_nmi_owner); } +EXPORT_SYMBOL(release_perfctr_nmi); int reserve_evntsel_nmi(unsigned int msr) { @@ -154,6 +167,7 @@ int reserve_evntsel_nmi(unsigned int msr) return 1; return 0; } +EXPORT_SYMBOL(reserve_evntsel_nmi); void release_evntsel_nmi(unsigned int msr) { @@ -166,11 +180,6 @@ void release_evntsel_nmi(unsigned int msr) clear_bit(counter, evntsel_nmi_owner); } - -EXPORT_SYMBOL(avail_to_resrv_perfctr_nmi_bit); -EXPORT_SYMBOL(reserve_perfctr_nmi); -EXPORT_SYMBOL(release_perfctr_nmi); -EXPORT_SYMBOL(reserve_evntsel_nmi); EXPORT_SYMBOL(release_evntsel_nmi); void disable_lapic_nmi_watchdog(void) @@ -181,7 +190,9 @@ void disable_lapic_nmi_watchdog(void) return; on_each_cpu(stop_apic_nmi_watchdog, NULL, 1); - wd_ops->unreserve(); + + if (wd_ops) + wd_ops->unreserve(); BUG_ON(atomic_read(&nmi_active) != 0); } @@ -232,8 +243,8 @@ static unsigned int adjust_for_32bit_ctr(unsigned int hz) return retval; } -static void -write_watchdog_counter(unsigned int perfctr_msr, const char *descr, unsigned nmi_hz) +static void write_watchdog_counter(unsigned int perfctr_msr, + const char *descr, unsigned nmi_hz) { u64 count = (u64)cpu_khz * 1000; @@ -244,7 +255,7 @@ write_watchdog_counter(unsigned int perfctr_msr, const char *descr, unsigned nmi } static void write_watchdog_counter32(unsigned int perfctr_msr, - const char *descr, unsigned nmi_hz) + const char *descr, unsigned nmi_hz) { u64 count = (u64)cpu_khz * 1000; @@ -254,9 +265,10 @@ static void write_watchdog_counter32(unsigned int perfctr_msr, wrmsr(perfctr_msr, (u32)(-count), 0); } -/* AMD K7/K8/Family10h/Family11h support. AMD keeps this interface - nicely stable so there is not much variety */ - +/* + * AMD K7/K8/Family10h/Family11h support. + * AMD keeps this interface nicely stable so there is not much variety + */ #define K7_EVNTSEL_ENABLE (1 << 22) #define K7_EVNTSEL_INT (1 << 20) #define K7_EVNTSEL_OS (1 << 17) @@ -289,7 +301,7 @@ static int setup_k7_watchdog(unsigned nmi_hz) wd->perfctr_msr = perfctr_msr; wd->evntsel_msr = evntsel_msr; - wd->cccr_msr = 0; //unused + wd->cccr_msr = 0; /* unused */ return 1; } @@ -325,18 +337,19 @@ static void single_msr_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz) } static const struct wd_ops k7_wd_ops = { - .reserve = single_msr_reserve, - .unreserve = single_msr_unreserve, - .setup = setup_k7_watchdog, - .rearm = single_msr_rearm, - .stop = single_msr_stop_watchdog, - .perfctr = MSR_K7_PERFCTR0, - .evntsel = MSR_K7_EVNTSEL0, - .checkbit = 1ULL<<47, + .reserve = single_msr_reserve, + .unreserve = single_msr_unreserve, + .setup = setup_k7_watchdog, + .rearm = single_msr_rearm, + .stop = single_msr_stop_watchdog, + .perfctr = MSR_K7_PERFCTR0, + .evntsel = MSR_K7_EVNTSEL0, + .checkbit = 1ULL << 47, }; -/* Intel Model 6 (PPro+,P2,P3,P-M,Core1) */ - +/* + * Intel Model 6 (PPro+,P2,P3,P-M,Core1) + */ #define P6_EVNTSEL0_ENABLE (1 << 22) #define P6_EVNTSEL_INT (1 << 20) #define P6_EVNTSEL_OS (1 << 17) @@ -372,52 +385,58 @@ static int setup_p6_watchdog(unsigned nmi_hz) wd->perfctr_msr = perfctr_msr; wd->evntsel_msr = evntsel_msr; - wd->cccr_msr = 0; //unused + wd->cccr_msr = 0; /* unused */ return 1; } static void p6_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz) { - /* P6 based Pentium M need to re-unmask + /* + * P6 based Pentium M need to re-unmask * the apic vector but it doesn't hurt * other P6 variant. - * ArchPerfom/Core Duo also needs this */ + * ArchPerfom/Core Duo also needs this + */ apic_write(APIC_LVTPC, APIC_DM_NMI); + /* P6/ARCH_PERFMON has 32 bit counter write */ write_watchdog_counter32(wd->perfctr_msr, NULL,nmi_hz); } static const struct wd_ops p6_wd_ops = { - .reserve = single_msr_reserve, - .unreserve = single_msr_unreserve, - .setup = setup_p6_watchdog, - .rearm = p6_rearm, - .stop = single_msr_stop_watchdog, - .perfctr = MSR_P6_PERFCTR0, - .evntsel = MSR_P6_EVNTSEL0, - .checkbit = 1ULL<<39, + .reserve = single_msr_reserve, + .unreserve = single_msr_unreserve, + .setup = setup_p6_watchdog, + .rearm = p6_rearm, + .stop = single_msr_stop_watchdog, + .perfctr = MSR_P6_PERFCTR0, + .evntsel = MSR_P6_EVNTSEL0, + .checkbit = 1ULL << 39, }; -/* Intel P4 performance counters. By far the most complicated of all. */ - -#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1<<7) -#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) -#define P4_CCCR_OVF (1<<31) - -/* 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] */ +/* + * Intel P4 performance counters. + * By far the most complicated of all. + */ +#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1 << 7) +#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) +#define P4_CCCR_OVF (1 << 31) +/* + * 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] + */ static int setup_p4_watchdog(unsigned nmi_hz) { unsigned int perfctr_msr, evntsel_msr, cccr_msr; @@ -442,7 +461,8 @@ static int setup_p4_watchdog(unsigned nmi_hz) #endif ht_num = 0; - /* performance counters are shared resources + /* + * performance counters are shared resources * assign each hyperthread its own set * (re-use the ESCR0 register, seems safe * and keeps the cccr_val the same) @@ -540,20 +560,21 @@ static void p4_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz) } static const struct wd_ops p4_wd_ops = { - .reserve = p4_reserve, - .unreserve = p4_unreserve, - .setup = setup_p4_watchdog, - .rearm = p4_rearm, - .stop = stop_p4_watchdog, + .reserve = p4_reserve, + .unreserve = p4_unreserve, + .setup = setup_p4_watchdog, + .rearm = p4_rearm, + .stop = stop_p4_watchdog, /* RED-PEN this is wrong for the other sibling */ - .perfctr = MSR_P4_BPU_PERFCTR0, - .evntsel = MSR_P4_BSU_ESCR0, - .checkbit = 1ULL<<39, + .perfctr = MSR_P4_BPU_PERFCTR0, + .evntsel = MSR_P4_BSU_ESCR0, + .checkbit = 1ULL << 39, }; -/* Watchdog using the Intel architected PerfMon. Used for Core2 and hopefully - all future Intel CPUs. */ - +/* + * Watchdog using the Intel architected PerfMon. + * Used for Core2 and hopefully all future Intel CPUs. + */ #define ARCH_PERFMON_NMI_EVENT_SEL ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL #define ARCH_PERFMON_NMI_EVENT_UMASK ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK @@ -599,19 +620,19 @@ static int setup_intel_arch_watchdog(unsigned nmi_hz) wd->perfctr_msr = perfctr_msr; wd->evntsel_msr = evntsel_msr; - wd->cccr_msr = 0; //unused + wd->cccr_msr = 0; /* unused */ intel_arch_wd_ops.checkbit = 1ULL << (eax.split.bit_width - 1); return 1; } static struct wd_ops intel_arch_wd_ops __read_mostly = { - .reserve = single_msr_reserve, - .unreserve = single_msr_unreserve, - .setup = setup_intel_arch_watchdog, - .rearm = p6_rearm, - .stop = single_msr_stop_watchdog, - .perfctr = MSR_ARCH_PERFMON_PERFCTR1, - .evntsel = MSR_ARCH_PERFMON_EVENTSEL1, + .reserve = single_msr_reserve, + .unreserve = single_msr_unreserve, + .setup = setup_intel_arch_watchdog, + .rearm = p6_rearm, + .stop = single_msr_stop_watchdog, + .perfctr = MSR_ARCH_PERFMON_PERFCTR1, + .evntsel = MSR_ARCH_PERFMON_EVENTSEL1, }; static void probe_nmi_watchdog(void) @@ -624,8 +645,10 @@ static void probe_nmi_watchdog(void) wd_ops = &k7_wd_ops; break; case X86_VENDOR_INTEL: - /* Work around Core Duo (Yonah) errata AE49 where perfctr1 - doesn't have a working enable bit. */ + /* + * Work around Core Duo (Yonah) errata AE49 where perfctr1 + * doesn't have a working enable bit. + */ if (boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model == 14) { intel_arch_wd_ops.perfctr = MSR_ARCH_PERFMON_PERFCTR0; intel_arch_wd_ops.evntsel = MSR_ARCH_PERFMON_EVENTSEL0; @@ -636,7 +659,7 @@ static void probe_nmi_watchdog(void) } switch (boot_cpu_data.x86) { case 6: - if (boot_cpu_data.x86_model > 0xd) + if (boot_cpu_data.x86_model > 13) return; wd_ops = &p6_wd_ops; @@ -697,10 +720,11 @@ int lapic_wd_event(unsigned nmi_hz) { struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk); u64 ctr; + rdmsrl(wd->perfctr_msr, ctr); - if (ctr & wd_ops->checkbit) { /* perfctr still running? */ + if (ctr & wd_ops->checkbit) /* perfctr still running? */ return 0; - } + wd_ops->rearm(wd, nmi_hz); return 1; } |