diff options
Diffstat (limited to 'arch/x86/kernel/apic')
| -rw-r--r-- | arch/x86/kernel/apic/apic.c | 150 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/bigsmp_32.c | 2 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/es7000_32.c | 2 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/io_apic.c | 380 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/ipi.c | 2 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/nmi.c | 22 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/numaq_32.c | 57 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/probe_64.c | 21 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/summit_32.c | 2 | ||||
| -rw-r--r-- | arch/x86/kernel/apic/x2apic_uv_x.c | 11 |
10 files changed, 357 insertions, 292 deletions
diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 0a1c2830ec6..894aa97f071 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -14,7 +14,7 @@ * Mikael Pettersson : PM converted to driver model. */ -#include <linux/perf_counter.h> +#include <linux/perf_event.h> #include <linux/kernel_stat.h> #include <linux/mc146818rtc.h> #include <linux/acpi_pmtmr.h> @@ -35,7 +35,8 @@ #include <linux/smp.h> #include <linux/mm.h> -#include <asm/perf_counter.h> +#include <asm/perf_event.h> +#include <asm/x86_init.h> #include <asm/pgalloc.h> #include <asm/atomic.h> #include <asm/mpspec.h> @@ -49,6 +50,7 @@ #include <asm/mtrr.h> #include <asm/smp.h> #include <asm/mce.h> +#include <asm/kvm_para.h> unsigned int num_processors; @@ -60,7 +62,7 @@ unsigned int boot_cpu_physical_apicid = -1U; /* * The highest APIC ID seen during enumeration. * - * This determines the messaging protocol we can use: if all APIC IDs + * On AMD, this determines the messaging protocol we can use: if all APIC IDs * are in the 0 ... 7 range, then we can use logical addressing which * has some performance advantages (better broadcasting). * @@ -977,7 +979,7 @@ void lapic_shutdown(void) { unsigned long flags; - if (!cpu_has_apic) + if (!cpu_has_apic && !apic_from_smp_config()) return; local_irq_save(flags); @@ -1187,7 +1189,7 @@ void __cpuinit setup_local_APIC(void) apic_write(APIC_ESR, 0); } #endif - perf_counters_lapic_init(); + perf_events_lapic_init(); preempt_disable(); @@ -1195,8 +1197,7 @@ void __cpuinit setup_local_APIC(void) * Double-check whether this APIC is really registered. * This is meaningless in clustered apic mode, so we skip it. */ - if (!apic->apic_id_registered()) - BUG(); + BUG_ON(!apic->apic_id_registered()); /* * Intel recommends to set DFR, LDR and TPR before enabling @@ -1361,52 +1362,80 @@ void enable_x2apic(void) } #endif /* CONFIG_X86_X2APIC */ -void __init enable_IR_x2apic(void) +int __init enable_IR(void) { #ifdef CONFIG_INTR_REMAP - int ret; - unsigned long flags; - struct IO_APIC_route_entry **ioapic_entries = NULL; - - ret = dmar_table_init(); - if (ret) { - pr_debug("dmar_table_init() failed with %d:\n", ret); - goto ir_failed; - } - if (!intr_remapping_supported()) { pr_debug("intr-remapping not supported\n"); - goto ir_failed; + return 0; } - if (!x2apic_preenabled && skip_ioapic_setup) { pr_info("Skipped enabling intr-remap because of skipping " "io-apic setup\n"); - return; + return 0; } + if (enable_intr_remapping(x2apic_supported())) + return 0; + + pr_info("Enabled Interrupt-remapping\n"); + + return 1; + +#endif + return 0; +} + +void __init enable_IR_x2apic(void) +{ + unsigned long flags; + struct IO_APIC_route_entry **ioapic_entries = NULL; + int ret, x2apic_enabled = 0; + int dmar_table_init_ret = 0; + +#ifdef CONFIG_INTR_REMAP + dmar_table_init_ret = dmar_table_init(); + if (dmar_table_init_ret) + pr_debug("dmar_table_init() failed with %d:\n", + dmar_table_init_ret); +#endif + ioapic_entries = alloc_ioapic_entries(); if (!ioapic_entries) { - pr_info("Allocate ioapic_entries failed: %d\n", ret); - goto end; + pr_err("Allocate ioapic_entries failed\n"); + goto out; } ret = save_IO_APIC_setup(ioapic_entries); if (ret) { pr_info("Saving IO-APIC state failed: %d\n", ret); - goto end; + goto out; } local_irq_save(flags); - mask_IO_APIC_setup(ioapic_entries); mask_8259A(); + mask_IO_APIC_setup(ioapic_entries); - ret = enable_intr_remapping(x2apic_supported()); - if (ret) - goto end_restore; + if (dmar_table_init_ret) + ret = 0; + else + ret = enable_IR(); - pr_info("Enabled Interrupt-remapping\n"); + if (!ret) { + /* IR is required if there is APIC ID > 255 even when running + * under KVM + */ + if (max_physical_apicid > 255 || !kvm_para_available()) + goto nox2apic; + /* + * without IR all CPUs can be addressed by IOAPIC/MSI + * only in physical mode + */ + x2apic_force_phys(); + } + + x2apic_enabled = 1; if (x2apic_supported() && !x2apic_mode) { x2apic_mode = 1; @@ -1414,41 +1443,25 @@ void __init enable_IR_x2apic(void) pr_info("Enabled x2apic\n"); } -end_restore: - if (ret) - /* - * IR enabling failed - */ +nox2apic: + if (!ret) /* IR enabling failed */ restore_IO_APIC_setup(ioapic_entries); - unmask_8259A(); local_irq_restore(flags); -end: +out: if (ioapic_entries) free_ioapic_entries(ioapic_entries); - if (!ret) + if (x2apic_enabled) return; -ir_failed: if (x2apic_preenabled) - panic("x2apic enabled by bios. But IR enabling failed"); + panic("x2apic: enabled by BIOS but kernel init failed."); else if (cpu_has_x2apic) - pr_info("Not enabling x2apic,Intr-remapping\n"); -#else - if (!cpu_has_x2apic) - return; - - if (x2apic_preenabled) - panic("x2apic enabled prior OS handover," - " enable CONFIG_X86_X2APIC, CONFIG_INTR_REMAP"); -#endif - - return; + pr_info("Not enabling x2apic, Intr-remapping init failed.\n"); } - #ifdef CONFIG_X86_64 /* * Detect and enable local APICs on non-SMP boards. @@ -1549,8 +1562,6 @@ no_apic: #ifdef CONFIG_X86_64 void __init early_init_lapic_mapping(void) { - unsigned long phys_addr; - /* * If no local APIC can be found then go out * : it means there is no mpatable and MADT @@ -1558,11 +1569,9 @@ void __init early_init_lapic_mapping(void) if (!smp_found_config) return; - phys_addr = mp_lapic_addr; - - set_fixmap_nocache(FIX_APIC_BASE, phys_addr); + set_fixmap_nocache(FIX_APIC_BASE, mp_lapic_addr); apic_printk(APIC_VERBOSE, "mapped APIC to %16lx (%16lx)\n", - APIC_BASE, phys_addr); + APIC_BASE, mp_lapic_addr); /* * Fetch the APIC ID of the BSP in case we have a @@ -1651,7 +1660,6 @@ int __init APIC_init_uniprocessor(void) APIC_INTEGRATED(apic_version[boot_cpu_physical_apicid])) { pr_err("BIOS bug, local APIC 0x%x not detected!...\n", boot_cpu_physical_apicid); - clear_cpu_cap(&boot_cpu_data, X86_FEATURE_APIC); return -1; } #endif @@ -1701,7 +1709,7 @@ int __init APIC_init_uniprocessor(void) localise_nmi_watchdog(); #endif - setup_boot_clock(); + x86_init.timers.setup_percpu_clockev(); #ifdef CONFIG_X86_64 check_nmi_watchdog(); #endif @@ -1908,24 +1916,14 @@ void __cpuinit generic_processor_info(int apicid, int version) max_physical_apicid = apicid; #ifdef CONFIG_X86_32 - /* - * Would be preferable to switch to bigsmp when CONFIG_HOTPLUG_CPU=y - * but we need to work other dependencies like SMP_SUSPEND etc - * before this can be done without some confusion. - * if (CPU_HOTPLUG_ENABLED || num_processors > 8) - * - Ashok Raj <ashok.raj@intel.com> - */ - if (max_physical_apicid >= 8) { - switch (boot_cpu_data.x86_vendor) { - case X86_VENDOR_INTEL: - if (!APIC_XAPIC(version)) { - def_to_bigsmp = 0; - break; - } - /* If P4 and above fall through */ - case X86_VENDOR_AMD: + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_INTEL: + if (num_processors > 8) + def_to_bigsmp = 1; + break; + case X86_VENDOR_AMD: + if (max_physical_apicid >= 8) def_to_bigsmp = 1; - } } #endif diff --git a/arch/x86/kernel/apic/bigsmp_32.c b/arch/x86/kernel/apic/bigsmp_32.c index 676cdac385c..77a06413b6b 100644 --- a/arch/x86/kernel/apic/bigsmp_32.c +++ b/arch/x86/kernel/apic/bigsmp_32.c @@ -112,7 +112,7 @@ static physid_mask_t bigsmp_ioapic_phys_id_map(physid_mask_t phys_map) return physids_promote(0xFFL); } -static int bigsmp_check_phys_apicid_present(int boot_cpu_physical_apicid) +static int bigsmp_check_phys_apicid_present(int phys_apicid) { return 1; } diff --git a/arch/x86/kernel/apic/es7000_32.c b/arch/x86/kernel/apic/es7000_32.c index 8952a589028..89174f847b4 100644 --- a/arch/x86/kernel/apic/es7000_32.c +++ b/arch/x86/kernel/apic/es7000_32.c @@ -167,7 +167,7 @@ static int es7000_apic_is_cluster(void) { /* MPENTIUMIII */ if (boot_cpu_data.x86 == 6 && - (boot_cpu_data.x86_model >= 7 || boot_cpu_data.x86_model <= 11)) + (boot_cpu_data.x86_model >= 7 && boot_cpu_data.x86_model <= 11)) return 1; return 0; diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index d2ed6c5ddc8..64970b9885f 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -66,6 +66,8 @@ #include <asm/apic.h> #define __apicdebuginit(type) static type __init +#define for_each_irq_pin(entry, head) \ + for (entry = head; entry; entry = entry->next) /* * Is the SiS APIC rmw bug present ? @@ -85,12 +87,20 @@ int nr_ioapic_registers[MAX_IO_APICS]; struct mpc_ioapic mp_ioapics[MAX_IO_APICS]; int nr_ioapics; +/* IO APIC gsi routing info */ +struct mp_ioapic_gsi mp_gsi_routing[MAX_IO_APICS]; + /* MP IRQ source entries */ struct mpc_intsrc mp_irqs[MAX_IRQ_SOURCES]; /* # of MP IRQ source entries */ int mp_irq_entries; +/* Number of legacy interrupts */ +static int nr_legacy_irqs __read_mostly = NR_IRQS_LEGACY; +/* GSI interrupts */ +static int nr_irqs_gsi = NR_IRQS_LEGACY; + #if defined (CONFIG_MCA) || defined (CONFIG_EISA) int mp_bus_id_to_type[MAX_MP_BUSSES]; #endif @@ -116,15 +126,6 @@ static int __init parse_noapic(char *str) } early_param("noapic", parse_noapic); -struct irq_pin_list; - -/* - * This is performance-critical, we want to do it O(1) - * - * the indexing order of this array favors 1:1 mappings - * between pins and IRQs. - */ - struct irq_pin_list { int apic, pin; struct irq_pin_list *next; @@ -139,6 +140,11 @@ static struct irq_pin_list *get_one_free_irq_2_pin(int node) return pin; } +/* + * This is performance-critical, we want to do it O(1) + * + * Most irqs are mapped 1:1 with pins. + */ struct irq_cfg { struct irq_pin_list *irq_2_pin; cpumask_var_t domain; @@ -172,6 +178,12 @@ static struct irq_cfg irq_cfgx[NR_IRQS] = { [15] = { .vector = IRQ15_VECTOR, }, }; +void __init io_apic_disable_legacy(void) +{ + nr_legacy_irqs = 0; + nr_irqs_gsi = 0; +} + int __init arch_early_irq_init(void) { struct irq_cfg *cfg; @@ -189,7 +201,7 @@ int __init arch_early_irq_init(void) desc->chip_data = &cfg[i]; zalloc_cpumask_var_node(&cfg[i].domain, GFP_NOWAIT, node); zalloc_cpumask_var_node(&cfg[i].old_domain, GFP_NOWAIT, node); - if (i < NR_IRQS_LEGACY) + if (i < nr_legacy_irqs) cpumask_setall(cfg[i].domain); } @@ -414,13 +426,10 @@ static bool io_apic_level_ack_pending(struct irq_cfg *cfg) unsigned long flags; spin_lock_irqsave(&ioapic_lock, flags); - entry = cfg->irq_2_pin; - for (;;) { + for_each_irq_pin(entry, cfg->irq_2_pin) { unsigned int reg; int pin; - if (!entry) - break; pin = entry->pin; reg = io_apic_read(entry->apic, 0x10 + pin*2); /* Is the remote IRR bit set? */ @@ -428,9 +437,6 @@ static bool io_apic_level_ack_pending(struct irq_cfg *cfg) spin_unlock_irqrestore(&ioapic_lock, flags); return true; } - if (!entry->next) - break; - entry = entry->next; } spin_unlock_irqrestore(&ioapic_lock, flags); @@ -498,72 +504,68 @@ static void ioapic_mask_entry(int apic, int pin) * shared ISA-space IRQs, so we have to support them. We are super * fast in the common case, and fast for shared ISA-space IRQs. */ -static void add_pin_to_irq_node(struct irq_cfg *cfg, int node, int apic, int pin) +static int +add_pin_to_irq_node_nopanic(struct irq_cfg *cfg, int node, int apic, int pin) { - struct irq_pin_list *entry; + struct irq_pin_list **last, *entry; - entry = cfg->irq_2_pin; - if (!entry) { - entry = get_one_free_irq_2_pin(node); - if (!entry) { - printk(KERN_ERR "can not alloc irq_2_pin to add %d - %d\n", - apic, pin); - return; - } - cfg->irq_2_pin = entry; - entry->apic = apic; - entry->pin = pin; - return; - } - - while (entry->next) { - /* not again, please */ + /* don't allow duplicates */ + last = &cfg->irq_2_pin; + for_each_irq_pin(entry, cfg->irq_2_pin) { if (entry->apic == apic && entry->pin == pin) - return; - - entry = entry->next; + return 0; + last = &entry->next; } - entry->next = get_one_free_irq_2_pin(node); - entry = entry->next; + entry = get_one_free_irq_2_pin(node); + if (!entry) { + printk(KERN_ERR "can not alloc irq_pin_list (%d,%d,%d)\n", + node, apic, pin); + return -ENOMEM; + } entry->apic = apic; entry->pin = pin; + + *last = entry; + return 0; +} + +static void add_pin_to_irq_node(struct irq_cfg *cfg, int node, int apic, int pin) +{ + if (add_pin_to_irq_node_nopanic(cfg, node, apic, pin)) + panic("IO-APIC: failed to add irq-pin. Can not proceed\n"); } /* * Reroute an IRQ to a different pin. */ static void __init replace_pin_at_irq_node(struct irq_cfg *cfg, int node, - int oldapic, int oldpin, - int newapic, int newpin) + int oldapic, int oldpin, + int newapic, int newpin) { - struct irq_pin_list *entry = cfg->irq_2_pin; - int replaced = 0; + struct irq_pin_list *entry; - while (entry) { + for_each_irq_pin(entry, cfg->irq_2_pin) { if (entry->apic == oldapic && entry->pin == oldpin) { entry->apic = newapic; entry->pin = newpin; - replaced = 1; /* every one is different, right? */ - break; + return; } - entry = entry->next; } - /* why? call replace before add? */ - if (!replaced) - add_pin_to_irq_node(cfg, node, newapic, newpin); + /* old apic/pin didn't exist, so just add new ones */ + add_pin_to_irq_node(cfg, node, newapic, newpin); } -static inline void io_apic_modify_irq(struct irq_cfg *cfg, - int mask_and, int mask_or, - void (*final)(struct irq_pin_list *entry)) +static void io_apic_modify_irq(struct irq_cfg *cfg, + int mask_and, int mask_or, + void (*final)(struct irq_pin_list *entry)) { int pin; struct irq_pin_list *entry; - for (entry = cfg->irq_2_pin; entry != NULL; entry = entry->next) { + for_each_irq_pin(entry, cfg->irq_2_pin) { unsigned int reg; pin = entry->pin; reg = io_apic_read(entry->apic, 0x10 + pin * 2); @@ -580,7 +582,6 @@ static void __unmask_IO_APIC_irq(struct irq_cfg *cfg) io_apic_modify_irq(cfg, ~IO_APIC_REDIR_MASKED, 0, NULL); } -#ifdef CONFIG_X86_64 static void io_apic_sync(struct irq_pin_list *entry) { /* @@ -596,11 +597,6 @@ static void __mask_IO_APIC_irq(struct irq_cfg *cfg) { io_apic_modify_irq(cfg, ~0, IO_APIC_REDIR_MASKED, &io_apic_sync); } -#else /* CONFIG_X86_32 */ -static void __mask_IO_APIC_irq(struct irq_cfg *cfg) -{ - io_apic_modify_irq(cfg, ~0, IO_APIC_REDIR_MASKED, NULL); -} static void __mask_and_edge_IO_APIC_irq(struct irq_cfg *cfg) { @@ -613,7 +609,6 @@ static void __unmask_and_level_IO_APIC_irq(struct irq_cfg *cfg) io_apic_modify_irq(cfg, ~IO_APIC_REDIR_MASKED, IO_APIC_REDIR_LEVEL_TRIGGER, NULL); } -#endif /* CONFIG_X86_32 */ static void mask_IO_APIC_irq_desc(struct irq_desc *desc) { @@ -883,7 +878,7 @@ static int __init find_isa_irq_apic(int irq, int type) */ static int EISA_ELCR(unsigned int irq) { - if (irq < NR_IRQS_LEGACY) { + if (irq < nr_legacy_irqs) { unsigned int port = 0x4d0 + (irq >> 3); return (inb(port) >> (irq & 7)) & 1; } @@ -1480,7 +1475,7 @@ static void setup_IO_APIC_irq(int apic_id, int pin, unsigned int irq, struct irq } ioapic_register_intr(irq, desc, trigger); - if (irq < NR_IRQS_LEGACY) + if (irq < nr_legacy_irqs) disable_8259A_irq(irq); ioapic_write_entry(apic_id, pin, entry); @@ -1702,12 +1697,8 @@ __apicdebuginit(void) print_IO_APIC(void) if (!entry) continue; printk(KERN_DEBUG "IRQ%d ", irq); - for (;;) { + for_each_irq_pin(entry, cfg->irq_2_pin) printk("-> %d:%d", entry->apic, entry->pin); - if (!entry->next) - break; - entry = entry->next; - } printk("\n"); } @@ -1851,7 +1842,7 @@ __apicdebuginit(void) print_PIC(void) unsigned int v; unsigned long flags; - if (apic_verbosity == APIC_QUIET) + if (apic_verbosity == APIC_QUIET || !nr_legacy_irqs) return; printk(KERN_DEBUG "\nprinting PIC contents\n"); @@ -1883,7 +1874,7 @@ __apicdebuginit(int) print_all_ICs(void) print_PIC(); /* don't print out if apic is not there */ - if (!cpu_has_apic || disable_apic) + if (!cpu_has_apic && !apic_from_smp_config()) return 0; print_all_local_APICs(); @@ -1914,6 +1905,10 @@ void __init enable_IO_APIC(void) spin_unlock_irqrestore(&ioapic_lock, flags); nr_ioapic_registers[apic] = reg_01.bits.entries+1; } + + if (!nr_legacy_irqs) + return; + for(apic = 0; apic < nr_ioapics; apic++) { int pin; /* See if any of the pins is in ExtINT mode */ @@ -1968,6 +1963,9 @@ void disable_IO_APIC(void) */ clear_IO_APIC(); + if (!nr_legacy_irqs) + return; + /* * If the i8259 is routed through an IOAPIC * Put that IOAPIC in virtual wire mode @@ -2001,7 +1999,7 @@ void disable_IO_APIC(void) /* * Use virtual wire A mode when interrupt remapping is enabled. */ - if (cpu_has_apic) + if (cpu_has_apic || apic_from_smp_config()) disconnect_bsp_APIC(!intr_remapping_enabled && ioapic_i8259.pin != -1); } @@ -2014,7 +2012,7 @@ void disable_IO_APIC(void) * by Matt Domsch <Matt_Domsch@dell.com> Tue Dec 21 12:25:05 CST 1999 */ -static void __init setup_ioapic_ids_from_mpc(void) +void __init setup_ioapic_ids_from_mpc(void) { union IO_APIC_reg_00 reg_00; physid_mask_t phys_id_present_map; @@ -2023,9 +2021,8 @@ static void __init setup_ioapic_ids_from_mpc(void) unsigned char old_id; unsigned long flags; - if (x86_quirks->setup_ioapic_ids && x86_quirks->setup_ioapic_ids()) + if (acpi_ioapic) return; - /* * Don't check I/O APIC IDs for xAPIC systems. They have * no meaning without the serial APIC bus. @@ -2199,7 +2196,7 @@ static unsigned int startup_ioapic_irq(unsigned int irq) struct irq_cfg *cfg; spin_lock_irqsave(&ioapic_lock, flags); - if (irq < NR_IRQS_LEGACY) { + if (irq < nr_legacy_irqs) { disable_8259A_irq(irq); if (i8259A_irq_pending(irq)) was_pending = 1; @@ -2211,7 +2208,6 @@ static unsigned int startup_ioapic_irq(unsigned int irq) return was_pending; } -#ifdef CONFIG_X86_64 static int ioapic_retrigger_irq(unsigned int irq) { @@ -2224,14 +2220,6 @@ static int ioapic_retrigger_irq(unsigned int irq) return 1; } -#else -static int ioapic_retrigger_irq(unsigned int irq) -{ - apic->send_IPI_self(irq_cfg(irq)->vector); - - return 1; -} -#endif /* * Level and edge triggered IO-APIC interrupts need different handling, @@ -2269,13 +2257,9 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq struct irq_pin_list *entry; u8 vector = cfg->vector; - entry = cfg->irq_2_pin; - for (;;) { + for_each_irq_pin(entry, cfg->irq_2_pin) { unsigned int reg; - if (!entry) - break; - apic = entry->apic; pin = entry->pin; /* @@ -2288,9 +2272,6 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq reg &= ~IO_APIC_REDIR_VECTOR_MASK; reg |= vector; io_apic_modify(apic, 0x10 + pin*2, reg); - if (!entry->next) - break; - entry = entry->next; } } @@ -2515,11 +2496,8 @@ atomic_t irq_mis_count; static void ack_apic_level(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - -#ifdef CONFIG_X86_32 unsigned long v; int i; -#endif struct irq_cfg *cfg; int do_unmask_irq = 0; @@ -2532,31 +2510,28 @@ static void ack_apic_level(unsigned int irq) } #endif -#ifdef CONFIG_X86_32 /* - * It appears there is an erratum which affects at least version 0x11 - * of I/O APIC (that's the 82093AA and cores integrated into various - * chipsets). Under certain conditions a level-triggered interrupt is - * erroneously delivered as edge-triggered one but the respective IRR - * bit gets set nevertheless. As a result the I/O unit expects an EOI - * message but it will never arrive and further interrupts are blocked - * from the source. The exact reason is so far unknown, but the - * phenomenon was observed when two consecutive interrupt requests - * from a given source get delivered to the same CPU and the source is - * temporarily disabled in between. - * - * A workaround is to simulate an EOI message manually. We achieve it - * by setting the trigger mode to edge and then to level when the edge - * trigger mode gets detected in the TMR of a local APIC for a - * level-triggered interrupt. We mask the source for the time of the - * operation to prevent an edge-triggered interrupt escaping meanwhile. - * The idea is from Manfred Spraul. --macro - */ + * It appears there is an erratum which affects at least version 0x11 + * of I/O APIC (that's the 82093AA and cores integrated into various + * chipsets). Under certain conditions a level-triggered interrupt is + * erroneously delivered as edge-triggered one but the respective IRR + * bit gets set nevertheless. As a result the I/O unit expects an EOI + * message but it will never arrive and further interrupts are blocked + * from the source. The exact reason is so far unknown, but the + * phenomenon was observed when two consecutive interrupt requests + * from a given source get delivered to the same CPU and the source is + * temporarily disabled in between. + * + * A workaround is to simulate an EOI message manually. We achieve it + * by setting the trigger mode to edge and then to level when the edge + * trigger mode gets detected in the TMR of a local APIC for a + * level-triggered interrupt. We mask the source for the time of the + * operation to prevent an edge-triggered interrupt escaping meanwhile. + * The idea is from Manfred Spraul. --macro + */ cfg = desc->chip_data; i = cfg->vector; - v = apic_read(APIC_TMR + ((i & ~0x1f) >> 1)); -#endif /* * We must acknowledge the irq before we move it or the acknowledge will @@ -2598,7 +2573,7 @@ static void ack_apic_level(unsigned int irq) unmask_IO_APIC_irq_desc(desc); } -#ifdef CONFIG_X86_32 + /* Tail end of version 0x11 I/O APIC bug workaround */ if (!(v & (1 << (i & 0x1f)))) { atomic_inc(&irq_mis_count); spin_lock(&ioapic_lock); @@ -2606,26 +2581,15 @@ static void ack_apic_level(unsigned int irq) __unmask_and_level_IO_APIC_irq(cfg); spin_unlock(&ioapic_lock); } -#endif } #ifdef CONFIG_INTR_REMAP static void __eoi_ioapic_irq(unsigned int irq, struct irq_cfg *cfg) { - int apic, pin; struct irq_pin_list *entry; - entry = cfg->irq_2_pin; - for (;;) { - - if (!entry) - break; - - apic = entry->apic; - pin = entry->pin; - io_apic_eoi(apic, pin); - entry = entry->next; - } + for_each_irq_pin(entry, cfg->irq_2_pin) + io_apic_eoi(entry->apic, entry->pin); } static void @@ -2710,7 +2674,7 @@ static inline void init_IO_APIC_traps(void) * so default to an old-fashioned 8259 * interrupt if we can.. */ - if (irq < NR_IRQS_LEGACY) + if (irq < nr_legacy_irqs) make_8259A_irq(irq); else /* Strange. Oh, well.. */ @@ -3046,7 +3010,7 @@ out: * the I/O APIC in all cases now. No actual device should request * it anyway. --macro */ -#define PIC_IRQS (1 << PIC_CASCADE_IR) +#define PIC_IRQS (1UL << PIC_CASCADE_IR) void __init setup_IO_APIC(void) { @@ -3054,21 +3018,19 @@ void __init setup_IO_APIC(void) /* * calling enable_IO_APIC() is moved to setup_local_APIC for BP */ - - io_apic_irqs = ~PIC_IRQS; + io_apic_irqs = nr_legacy_irqs ? ~PIC_IRQS : ~0UL; apic_printk(APIC_VERBOSE, "ENABLING IO-APIC IRQs\n"); /* * Set up IO-APIC IRQ routing. */ -#ifdef CONFIG_X86_32 - if (!acpi_ioapic) - setup_ioapic_ids_from_mpc(); -#endif + x86_init.mpparse.setup_ioapic_ids(); + sync_Arb_IDs(); setup_IO_APIC_irqs(); init_IO_APIC_traps(); - check_timer(); + if (nr_legacy_irqs) + check_timer(); } /* @@ -3169,7 +3131,6 @@ static int __init ioapic_init_sysfs(void) device_initcall(ioapic_init_sysfs); -static int nr_irqs_gsi = NR_IRQS_LEGACY; /* * Dynamic irq allocate and deallocation */ @@ -3241,8 +3202,7 @@ void destroy_irq(unsigned int irq) cfg = desc->chip_data; dynamic_irq_cleanup(irq); /* connect back irq_cfg */ - if (desc) - desc->chip_data = cfg; + desc->chip_data = cfg; free_irte(irq); spin_lock_irqsave(&vector_lock, flags); @@ -3910,9 +3870,13 @@ static int __io_apic_set_pci_routing(struct device *dev, int irq, /* * IRQs < 16 are already in the irq_2_pin[] map */ - if (irq >= NR_IRQS_LEGACY) { + if (irq >= nr_legacy_irqs) { cfg = desc->chip_data; - add_pin_to_irq_node(cfg, node, ioapic, pin); + if (add_pin_to_irq_node_nopanic(cfg, node, ioapic, pin)) { + printk(KERN_INFO "can not add pin %d for irq %d\n", + pin, irq); + return 0; + } } setup_IO_APIC_irq(ioapic, pin, irq, desc, trigger, polarity); @@ -3941,11 +3905,28 @@ int io_apic_set_pci_routing(struct device *dev, int irq, return __io_apic_set_pci_routing(dev, irq, irq_attr); } -/* -------------------------------------------------------------------------- - ACPI-based IOAPIC Configuration - -------------------------------------------------------------------------- */ +u8 __init io_apic_unique_id(u8 id) +{ +#ifdef CONFIG_X86_32 + if ((boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) && + !APIC_XAPIC(apic_version[boot_cpu_physical_apicid])) + return io_apic_get_unique_id(nr_ioapics, id); + else + return id; +#else + int i; + DECLARE_BITMAP(used, 256); -#ifdef CONFIG_ACPI + bitmap_zero(used, 256); + for (i = 0; i < nr_ioapics; i++) { + struct mpc_ioapic *ia = &mp_ioapics[i]; + __set_bit(ia->apicid, used); + } + if (!test_bit(id, used)) + return id; + return find_first_zero_bit(used, 256); +#endif +} #ifdef CONFIG_X86_32 int __init io_apic_get_unique_id(int ioapic, int apic_id) @@ -4054,8 +4035,6 @@ int acpi_get_override_irq(int bus_irq, int *trigger, int *polarity) return 0; } -#endif /* CONFIG_ACPI */ - /* * This function currently is only a helper for the i386 smp boot process where * we need to reprogram the ioredtbls to cater for the cpus which have come online @@ -4109,7 +4088,7 @@ void __init setup_ioapic_dest(void) static struct resource *ioapic_resources; -static struct resource * __init ioapic_setup_resources(void) +static struct resource * __init ioapic_setup_resources(int nr_ioapics) { unsigned long n; struct resource *res; @@ -4125,15 +4104,13 @@ static struct resource * __init ioapic_setup_resources(void) mem = alloc_bootmem(n); res = (void *)mem; - if (mem != NULL) { - mem += sizeof(struct resource) * nr_ioapics; + mem += sizeof(struct resource) * nr_ioapics; - for (i = 0; i < nr_ioapics; i++) { - res[i].name = mem; - res[i].flags = IORESOURCE_MEM | IORESOURCE_BUSY; - sprintf(mem, "IOAPIC %u", i); - mem += IOAPIC_RESOURCE_NAME_SIZE; - } + for (i = 0; i < nr_ioapics; i++) { + res[i].name = mem; + res[i].flags = IORESOURCE_MEM | IORESOURCE_BUSY; + sprintf(mem, "IOAPIC %u", i); + mem += IOAPIC_RESOURCE_NAME_SIZE; } ioapic_resources = res; @@ -4147,7 +4124,7 @@ void __init ioapic_init_mappings(void) struct resource *ioapic_res; int i; - ioapic_res = ioapic_setup_resources(); + ioapic_res = ioapic_setup_resources(nr_ioapics); for (i = 0; i < nr_ioapics; i++) { if (smp_found_config) { ioapic_phys = mp_ioapics[i].apicaddr; @@ -4176,11 +4153,9 @@ fake_ioapic_page: __fix_to_virt(idx), ioapic_phys); idx++; - if (ioapic_res != NULL) { - ioapic_res->start = ioapic_phys; - ioapic_res->end = ioapic_phys + (4 * 1024) - 1; - ioapic_res++; - } + ioapic_res->start = ioapic_phys; + ioapic_res->end = ioapic_phys + (4 * 1024) - 1; + ioapic_res++; } } @@ -4201,3 +4176,76 @@ void __init ioapic_insert_resources(void) r++; } } + +int mp_find_ioapic(int gsi) +{ + int i = 0; + + /* Find the IOAPIC that manages this GSI. */ + for (i = 0; i < nr_ioapics; i++) { + if ((gsi >= mp_gsi_routing[i].gsi_base) + && (gsi <= mp_gsi_routing[i].gsi_end)) + return i; + } + + printk(KERN_ERR "ERROR: Unable to locate IOAPIC for GSI %d\n", gsi); + return -1; +} + +int mp_find_ioapic_pin(int ioapic, int gsi) +{ + if (WARN_ON(ioapic == -1)) + return -1; + if (WARN_ON(gsi > mp_gsi_routing[ioapic].gsi_end)) + return -1; + + return gsi - mp_gsi_routing[ioapic].gsi_base; +} + +static int bad_ioapic(unsigned long address) +{ + if (nr_ioapics >= MAX_IO_APICS) { + printk(KERN_WARNING "WARING: Max # of I/O APICs (%d) exceeded " + "(found %d), skipping\n", MAX_IO_APICS, nr_ioapics); + return 1; + } + if (!address) { + printk(KERN_WARNING "WARNING: Bogus (zero) I/O APIC address" + " found in table, skipping!\n"); + return 1; + } + return 0; +} + +void __init mp_register_ioapic(int id, u32 address, u32 gsi_base) +{ + int idx = 0; + + if (bad_ioapic(address)) + return; + + idx = nr_ioapics; + + mp_ioapics[idx].type = MP_IOAPIC; + mp_ioapics[idx].flags = MPC_APIC_USABLE; + mp_ioapics[idx].apicaddr = address; + + set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address); + mp_ioapics[idx].apicid = io_apic_unique_id(id); + mp_ioapics[idx].apicver = io_apic_get_version(idx); + + /* + * Build basic GSI lookup table to facilitate gsi->io_apic lookups + * and to prevent reprogramming of IOAPIC pins (PCI GSIs). + */ + mp_gsi_routing[idx].gsi_base = gsi_base; + mp_gsi_routing[idx].gsi_end = gsi_base + + io_apic_get_redir_entries(idx); + + printk(KERN_INFO "IOAPIC[%d]: apic_id %d, version %d, address 0x%x, " + "GSI %d-%d\n", idx, mp_ioapics[idx].apicid, + mp_ioapics[idx].apicver, mp_ioapics[idx].apicaddr, + mp_gsi_routing[idx].gsi_base, mp_gsi_routing[idx].gsi_end); + + nr_ioapics++; +} diff --git a/arch/x86/kernel/apic/ipi.c b/arch/x86/kernel/apic/ipi.c index 6ef00ba4c88..08385e090a6 100644 --- a/arch/x86/kernel/apic/ipi.c +++ b/arch/x86/kernel/apic/ipi.c @@ -153,7 +153,7 @@ int safe_smp_processor_id(void) { int apicid, cpuid; - if (!boot_cpu_has(X86_FEATURE_APIC)) + if (!cpu_has_apic) return 0; apicid = hard_smp_processor_id(); diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c index b3025b43b63..cb66a22d98a 100644 --- a/arch/x86/kernel/apic/nmi.c +++ b/arch/x86/kernel/apic/nmi.c @@ -39,7 +39,7 @@ int unknown_nmi_panic; int nmi_watchdog_enabled; -static cpumask_var_t backtrace_mask; +static cpumask_t backtrace_mask __read_mostly; /* nmi_active: * >0: the lapic NMI watchdog is active, but can be disabled @@ -66,7 +66,7 @@ static inline unsigned int get_nmi_count(int cpu) static inline int mce_in_progress(void) { -#if defined(CONFIG_X86_NEW_MCE) +#if defined(CONFIG_X86_MCE) return atomic_read(&mce_entry) > 0; #endif return 0; @@ -138,7 +138,6 @@ int __init check_nmi_watchdog(void) if (!prev_nmi_count) goto error; - alloc_cpumask_var(&backtrace_mask, GFP_KERNEL|__GFP_ZERO); printk(KERN_INFO "Testing NMI watchdog ... "); #ifdef CONFIG_SMP @@ -415,14 +414,17 @@ nmi_watchdog_tick(struct pt_regs *regs, unsigned reason) } /* We can be called before check_nmi_watchdog, hence NULL check. */ - if (backtrace_mask != NULL && cpumask_test_cpu(cpu, backtrace_mask)) { + if (cpumask_test_cpu(cpu, &backtrace_mask)) { static DEFINE_SPINLOCK(lock); /* Serialise the printks */ spin_lock(&lock); printk(KERN_WARNING "NMI backtrace for cpu %d\n", cpu); + show_regs(regs); dump_stack(); spin_unlock(&lock); - cpumask_clear_cpu(cpu, backtrace_mask); + cpumask_clear_cpu(cpu, &backtrace_mask); + + rc = 1; } /* Could check oops_in_progress here too, but it's safer not to */ @@ -552,14 +554,18 @@ int do_nmi_callback(struct pt_regs *regs, int cpu) return 0; } -void __trigger_all_cpu_backtrace(void) +void arch_trigger_all_cpu_backtrace(void) { int i; - cpumask_copy(backtrace_mask, cpu_online_mask); + cpumask_copy(&backtrace_mask, cpu_online_mask); + + printk(KERN_INFO "sending NMI to all CPUs:\n"); + apic->send_IPI_all(NMI_VECTOR); + /* Wait for up to 10 seconds for all CPUs to do the backtrace */ for (i = 0; i < 10 * 1000; i++) { - if (cpumask_empty(backtrace_mask)) + if (cpumask_empty(&backtrace_mask)) break; mdelay(1); } diff --git a/arch/x86/kernel/apic/numaq_32.c b/arch/x86/kernel/apic/numaq_32.c index ca96e68f0d2..efa00e2b850 100644 --- a/arch/x86/kernel/apic/numaq_32.c +++ b/arch/x86/kernel/apic/numaq_32.c @@ -66,7 +66,6 @@ struct mpc_trans { unsigned short trans_reserved; }; -/* x86_quirks member */ static int mpc_record; static struct mpc_trans *translation_table[MAX_MPC_ENTRY]; @@ -130,10 +129,9 @@ void __cpuinit numaq_tsc_disable(void) } } -static int __init numaq_pre_time_init(void) +static void __init numaq_tsc_init(void) { numaq_tsc_disable(); - return 0; } static inline int generate_logical_apicid(int quad, int phys_apicid) @@ -177,6 +175,19 @@ static void mpc_oem_pci_bus(struct mpc_bus *m) quad_local_to_mp_bus_id[quad][local] = m->busid; } +/* + * Called from mpparse code. + * mode = 0: prescan + * mode = 1: one mpc entry scanned + */ +static void numaq_mpc_record(unsigned int mode) +{ + if (!mode) + mpc_record = 0; + else + mpc_record++; +} + static void __init MP_translation_info(struct mpc_trans *m) { printk(KERN_INFO @@ -206,9 +217,9 @@ static int __init mpf_checksum(unsigned char *mp, int len) /* * Read/parse the MPC oem tables */ -static void __init - smp_read_mpc_oem(struct mpc_oemtable *oemtable, unsigned short oemsize) +static void __init smp_read_mpc_oem(struct mpc_table *mpc) { + struct mpc_oemtable *oemtable = (void *)(long)mpc->oemptr; int count = sizeof(*oemtable); /* the header size */ unsigned char *oemptr = ((unsigned char *)oemtable) + count; @@ -250,29 +261,6 @@ static void __init } } -static int __init numaq_setup_ioapic_ids(void) -{ - /* so can skip it */ - return 1; -} - -static struct x86_quirks numaq_x86_quirks __initdata = { - .arch_pre_time_init = numaq_pre_time_init, - .arch_time_init = NULL, - .arch_pre_intr_init = NULL, - .arch_memory_setup = NULL, - .arch_intr_init = NULL, - .arch_trap_init = NULL, - .mach_get_smp_config = NULL, - .mach_find_smp_config = NULL, - .mpc_record = &mpc_record, - .mpc_apic_id = mpc_apic_id, - .mpc_oem_bus_info = mpc_oem_bus_info, - .mpc_oem_pci_bus = mpc_oem_pci_bus, - .smp_read_mpc_oem = smp_read_mpc_oem, - .setup_ioapic_ids = numaq_setup_ioapic_ids, -}; - static __init void early_check_numaq(void) { /* @@ -286,8 +274,15 @@ static __init void early_check_numaq(void) if (smp_found_config) early_get_smp_config(); - if (found_numaq) - x86_quirks = &numaq_x86_quirks; + if (found_numaq) { + x86_init.mpparse.mpc_record = numaq_mpc_record; + x86_init.mpparse.setup_ioapic_ids = x86_init_noop; + x86_init.mpparse.mpc_apic_id = mpc_apic_id; + x86_init.mpparse.smp_read_mpc_oem = smp_read_mpc_oem; + x86_init.mpparse.mpc_oem_pci_bus = mpc_oem_pci_bus; + x86_init.mpparse.mpc_oem_bus_info = mpc_oem_bus_info; + x86_init.timers.tsc_pre_init = numaq_tsc_init; + } } int __init get_memcfg_numaq(void) @@ -418,7 +413,7 @@ static inline physid_mask_t numaq_apicid_to_cpu_present(int logical_apicid) /* Where the IO area was mapped on multiquad, always 0 otherwise */ void *xquad_portio; -static inline int numaq_check_phys_apicid_present(int boot_cpu_physical_apicid) +static inline int numaq_check_phys_apicid_present(int phys_apicid) { return 1; } diff --git a/arch/x86/kernel/apic/probe_64.c b/arch/x86/kernel/apic/probe_64.c index fcec2f1d34a..c4cbd3080c1 100644 --- a/arch/x86/kernel/apic/probe_64.c +++ b/arch/x86/kernel/apic/probe_64.c @@ -55,25 +55,32 @@ static int apicid_phys_pkg_id(int initial_apic_id, int index_msb) void __init default_setup_apic_routing(void) { #ifdef CONFIG_X86_X2APIC - if (x2apic_mode && (apic != &apic_x2apic_phys && + if (x2apic_mode #ifdef CONFIG_X86_UV - apic != &apic_x2apic_uv_x && + && apic != &apic_x2apic_uv_x #endif - apic != &apic_x2apic_cluster)) { + ) { if (x2apic_phys) apic = &apic_x2apic_phys; else apic = &apic_x2apic_cluster; - printk(KERN_INFO "Setting APIC routing to %s\n", apic->name); } #endif if (apic == &apic_flat) { - if (max_physical_apicid >= 8) - apic = &apic_physflat; - printk(KERN_INFO "Setting APIC routing to %s\n", apic->name); + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_INTEL: + if (num_processors > 8) + apic = &apic_physflat; + break; + case X86_VENDOR_AMD: + if (max_physical_apicid >= 8) + apic = &apic_physflat; + } } + printk(KERN_INFO "Setting APIC routing to %s\n", apic->name); + if (is_vsmp_box()) { /* need to update phys_pkg_id */ apic->phys_pkg_id = apicid_phys_pkg_id; diff --git a/arch/x86/kernel/apic/summit_32.c b/arch/x86/kernel/apic/summit_32.c index eafdfbd1ea9..645ecc4ff0b 100644 --- a/arch/x86/kernel/apic/summit_32.c +++ b/arch/x86/kernel/apic/summit_32.c @@ -272,7 +272,7 @@ static physid_mask_t summit_apicid_to_cpu_present(int apicid) return physid_mask_of_physid(0); } -static int summit_check_phys_apicid_present(int boot_cpu_physical_apicid) +static int summit_check_phys_apicid_present(int physical_apicid) { return 1; } diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index 601159374e8..f5f5886a6b5 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -389,6 +389,16 @@ static __init void map_gru_high(int max_pnode) map_high("GRU", gru.s.base, shift, max_pnode, map_wb); } +static __init void map_mmr_high(int max_pnode) +{ + union uvh_rh_gam_mmr_overlay_config_mmr_u mmr; + int shift = UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR_BASE_SHFT; + + mmr.v = uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR); + if (mmr.s.enable) + map_high("MMR", mmr.s.base, shift, max_pnode, map_uc); +} + static __init void map_mmioh_high(int max_pnode) { union uvh_rh_gam_mmioh_overlay_config_mmr_u mmioh; @@ -643,6 +653,7 @@ void __init uv_system_init(void) } map_gru_high(max_pnode); + map_mmr_high(max_pnode); map_mmioh_high(max_pnode); uv_cpu_init(); |