X-Git-Url: http://drtracing.org/?a=blobdiff_plain;f=arch%2Fx86%2Fkernel%2Fio_apic_64.c;h=940c4167b32541589d3a96a0cd4f73f95f65d52a;hb=047c8fdb8718890e3340073b178d0859d0c7f91f;hp=61a83b70c18fcc65ce60b965ee3a6e0456622dc4;hpb=8b53b57576292b92b27769f9e213df19b6e57786;p=deliverable%2Flinux.git diff --git a/arch/x86/kernel/io_apic_64.c b/arch/x86/kernel/io_apic_64.c index 61a83b70c18f..940c4167b325 100644 --- a/arch/x86/kernel/io_apic_64.c +++ b/arch/x86/kernel/io_apic_64.c @@ -27,16 +27,20 @@ #include #include #include +#include #include +#include #include #include #include -#include -#include +#include +#include +#include /* time_after() */ #ifdef CONFIG_ACPI #include #endif #include +#include #include #include @@ -45,15 +49,73 @@ #include #include #include +#include #include #include #include #include +#include +#include #include #include +#include + +#define __apicdebuginit(type) static type __init + +/* + * Is the SiS APIC rmw bug present ? + * -1 = don't know, 0 = no, 1 = yes + */ +int sis_apic_bug = -1; + +static DEFINE_SPINLOCK(ioapic_lock); +static DEFINE_SPINLOCK(vector_lock); + +int first_free_entry; +/* + * Rough estimation of how many shared IRQs there are, can + * be changed anytime. + */ +int pin_map_size; + +/* + * # of IRQ routing registers + */ +int nr_ioapic_registers[MAX_IO_APICS]; + +/* I/O APIC entries */ +struct mp_config_ioapic mp_ioapics[MAX_IO_APICS]; +int nr_ioapics; + +/* MP IRQ source entries */ +struct mp_config_intsrc mp_irqs[MAX_IRQ_SOURCES]; + +/* # of MP IRQ source entries */ +int mp_irq_entries; + +#if defined (CONFIG_MCA) || defined (CONFIG_EISA) +int mp_bus_id_to_type[MAX_MP_BUSSES]; +#endif + +DECLARE_BITMAP(mp_bus_not_pci, MAX_MP_BUSSES); + +int skip_ioapic_setup; +static int __init parse_noapic(char *str) +{ + /* disable IO-APIC */ + disable_ioapic_setup(); + return 0; +} +early_param("noapic", parse_noapic); + +struct irq_cfg; +struct irq_pin_list; struct irq_cfg { + unsigned int irq; + struct irq_cfg *next; + struct irq_pin_list *irq_2_pin; cpumask_t domain; cpumask_t old_domain; unsigned move_cleanup_count; @@ -62,70 +124,162 @@ struct irq_cfg { }; /* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */ -static struct irq_cfg irq_cfg[NR_IRQS] __read_mostly = { - [0] = { .domain = CPU_MASK_ALL, .vector = IRQ0_VECTOR, }, - [1] = { .domain = CPU_MASK_ALL, .vector = IRQ1_VECTOR, }, - [2] = { .domain = CPU_MASK_ALL, .vector = IRQ2_VECTOR, }, - [3] = { .domain = CPU_MASK_ALL, .vector = IRQ3_VECTOR, }, - [4] = { .domain = CPU_MASK_ALL, .vector = IRQ4_VECTOR, }, - [5] = { .domain = CPU_MASK_ALL, .vector = IRQ5_VECTOR, }, - [6] = { .domain = CPU_MASK_ALL, .vector = IRQ6_VECTOR, }, - [7] = { .domain = CPU_MASK_ALL, .vector = IRQ7_VECTOR, }, - [8] = { .domain = CPU_MASK_ALL, .vector = IRQ8_VECTOR, }, - [9] = { .domain = CPU_MASK_ALL, .vector = IRQ9_VECTOR, }, - [10] = { .domain = CPU_MASK_ALL, .vector = IRQ10_VECTOR, }, - [11] = { .domain = CPU_MASK_ALL, .vector = IRQ11_VECTOR, }, - [12] = { .domain = CPU_MASK_ALL, .vector = IRQ12_VECTOR, }, - [13] = { .domain = CPU_MASK_ALL, .vector = IRQ13_VECTOR, }, - [14] = { .domain = CPU_MASK_ALL, .vector = IRQ14_VECTOR, }, - [15] = { .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, }, +static struct irq_cfg irq_cfg_legacy[] __initdata = { + [0] = { .irq = 0, .domain = CPU_MASK_ALL, .vector = IRQ0_VECTOR, }, + [1] = { .irq = 1, .domain = CPU_MASK_ALL, .vector = IRQ1_VECTOR, }, + [2] = { .irq = 2, .domain = CPU_MASK_ALL, .vector = IRQ2_VECTOR, }, + [3] = { .irq = 3, .domain = CPU_MASK_ALL, .vector = IRQ3_VECTOR, }, + [4] = { .irq = 4, .domain = CPU_MASK_ALL, .vector = IRQ4_VECTOR, }, + [5] = { .irq = 5, .domain = CPU_MASK_ALL, .vector = IRQ5_VECTOR, }, + [6] = { .irq = 6, .domain = CPU_MASK_ALL, .vector = IRQ6_VECTOR, }, + [7] = { .irq = 7, .domain = CPU_MASK_ALL, .vector = IRQ7_VECTOR, }, + [8] = { .irq = 8, .domain = CPU_MASK_ALL, .vector = IRQ8_VECTOR, }, + [9] = { .irq = 9, .domain = CPU_MASK_ALL, .vector = IRQ9_VECTOR, }, + [10] = { .irq = 10, .domain = CPU_MASK_ALL, .vector = IRQ10_VECTOR, }, + [11] = { .irq = 11, .domain = CPU_MASK_ALL, .vector = IRQ11_VECTOR, }, + [12] = { .irq = 12, .domain = CPU_MASK_ALL, .vector = IRQ12_VECTOR, }, + [13] = { .irq = 13, .domain = CPU_MASK_ALL, .vector = IRQ13_VECTOR, }, + [14] = { .irq = 14, .domain = CPU_MASK_ALL, .vector = IRQ14_VECTOR, }, + [15] = { .irq = 15, .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, }, }; -static int assign_irq_vector(int irq, cpumask_t mask); +static struct irq_cfg irq_cfg_init = { .irq = -1U, }; +/* need to be biger than size of irq_cfg_legacy */ +static int nr_irq_cfg = 32; -int first_system_vector = 0xfe; +static int __init parse_nr_irq_cfg(char *arg) +{ + if (arg) { + nr_irq_cfg = simple_strtoul(arg, NULL, 0); + if (nr_irq_cfg < 32) + nr_irq_cfg = 32; + } + return 0; +} -char system_vectors[NR_VECTORS] = { [0 ... NR_VECTORS-1] = SYS_VECTOR_FREE}; +early_param("nr_irq_cfg", parse_nr_irq_cfg); -#define __apicdebuginit __init +static void init_one_irq_cfg(struct irq_cfg *cfg) +{ + memcpy(cfg, &irq_cfg_init, sizeof(struct irq_cfg)); +} -int sis_apic_bug; /* not actually supported, dummy for compile */ +static struct irq_cfg *irq_cfgx; +static struct irq_cfg *irq_cfgx_free; +static void __init init_work(void *data) +{ + struct dyn_array *da = data; + struct irq_cfg *cfg; + int legacy_count; + int i; -static int no_timer_check; + cfg = *da->name; -static int disable_timer_pin_1 __initdata; + memcpy(cfg, irq_cfg_legacy, sizeof(irq_cfg_legacy)); -int timer_through_8259 __initdata; + legacy_count = sizeof(irq_cfg_legacy)/sizeof(irq_cfg_legacy[0]); + for (i = legacy_count; i < *da->nr; i++) + init_one_irq_cfg(&cfg[i]); -/* Where if anywhere is the i8259 connect in external int mode */ -static struct { int pin, apic; } ioapic_i8259 = { -1, -1 }; + for (i = 1; i < *da->nr; i++) + cfg[i-1].next = &cfg[i]; -static DEFINE_SPINLOCK(ioapic_lock); -static DEFINE_SPINLOCK(vector_lock); + irq_cfgx_free = &irq_cfgx[legacy_count]; + irq_cfgx[legacy_count - 1].next = NULL; +} -/* - * # of IRQ routing registers - */ -int nr_ioapic_registers[MAX_IO_APICS]; +#define for_each_irq_cfg(cfg) \ + for (cfg = irq_cfgx; cfg; cfg = cfg->next) -/* I/O APIC entries */ -struct mp_config_ioapic mp_ioapics[MAX_IO_APICS]; -int nr_ioapics; +DEFINE_DYN_ARRAY(irq_cfgx, sizeof(struct irq_cfg), nr_irq_cfg, PAGE_SIZE, init_work); -/* MP IRQ source entries */ -struct mp_config_intsrc mp_irqs[MAX_IRQ_SOURCES]; +static struct irq_cfg *irq_cfg(unsigned int irq) +{ + struct irq_cfg *cfg; -/* # of MP IRQ source entries */ -int mp_irq_entries; + cfg = irq_cfgx; + while (cfg) { + if (cfg->irq == irq) + return cfg; -DECLARE_BITMAP(mp_bus_not_pci, MAX_MP_BUSSES); + cfg = cfg->next; + } -/* - * Rough estimation of how many shared IRQs there are, can - * be changed anytime. - */ -#define MAX_PLUS_SHARED_IRQS NR_IRQS -#define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS) + return NULL; +} + +static struct irq_cfg *irq_cfg_alloc(unsigned int irq) +{ + struct irq_cfg *cfg, *cfg_pri; + int i; + int count = 0; + + cfg_pri = cfg = irq_cfgx; + while (cfg) { + if (cfg->irq == irq) + return cfg; + + cfg_pri = cfg; + cfg = cfg->next; + count++; + } + + if (!irq_cfgx_free) { + unsigned long phys; + unsigned long total_bytes; + /* + * we run out of pre-allocate ones, allocate more + */ + printk(KERN_DEBUG "try to get more irq_cfg %d\n", nr_irq_cfg); + + total_bytes = sizeof(struct irq_cfg) * nr_irq_cfg; + if (after_bootmem) + cfg = kzalloc(total_bytes, GFP_ATOMIC); + else + cfg = __alloc_bootmem_nopanic(total_bytes, PAGE_SIZE, 0); + + if (!cfg) + panic("please boot with nr_irq_cfg= %d\n", count * 2); + + phys = __pa(cfg); + printk(KERN_DEBUG "irq_irq ==> [%#lx - %#lx]\n", phys, phys + total_bytes); + + for (i = 0; i < nr_irq_cfg; i++) + init_one_irq_cfg(&cfg[i]); + + for (i = 1; i < nr_irq_cfg; i++) + cfg[i-1].next = &cfg[i]; + + irq_cfgx_free = cfg; + } + + cfg = irq_cfgx_free; + irq_cfgx_free = irq_cfgx_free->next; + cfg->next = NULL; + if (cfg_pri) + cfg_pri->next = cfg; + else + irq_cfgx = cfg; + cfg->irq = irq; + printk(KERN_DEBUG "found new irq_cfg for irq %d\n", cfg->irq); +#ifdef CONFIG_HAVE_SPARSE_IRQ_DEBUG + { + /* dump the results */ + struct irq_cfg *cfg; + unsigned long phys; + unsigned long bytes = sizeof(struct irq_cfg); + + printk(KERN_DEBUG "=========================== %d\n", irq); + printk(KERN_DEBUG "irq_cfg dump after get that for %d\n", irq); + for_each_irq_cfg(cfg) { + phys = __pa(cfg); + printk(KERN_DEBUG "irq_cfg %d ==> [%#lx - %#lx]\n", cfg->irq, phys, phys + bytes); + } + printk(KERN_DEBUG "===========================\n"); + } +#endif + return cfg; +} /* * This is performance-critical, we want to do it O(1) @@ -134,9 +288,66 @@ DECLARE_BITMAP(mp_bus_not_pci, MAX_MP_BUSSES); * between pins and IRQs. */ -static struct irq_pin_list { - short apic, pin, next; -} irq_2_pin[PIN_MAP_SIZE]; +struct irq_pin_list { + int apic, pin; + struct irq_pin_list *next; +}; + +static struct irq_pin_list *irq_2_pin_head; +/* fill one page ? */ +static int nr_irq_2_pin = 0x100; +static struct irq_pin_list *irq_2_pin_ptr; +static void __init irq_2_pin_init_work(void *data) +{ + struct dyn_array *da = data; + struct irq_pin_list *pin; + int i; + + pin = *da->name; + + for (i = 1; i < *da->nr; i++) + pin[i-1].next = &pin[i]; + + irq_2_pin_ptr = &pin[0]; +} +DEFINE_DYN_ARRAY(irq_2_pin_head, sizeof(struct irq_pin_list), nr_irq_2_pin, PAGE_SIZE, irq_2_pin_init_work); + +static struct irq_pin_list *get_one_free_irq_2_pin(void) +{ + struct irq_pin_list *pin; + int i; + + pin = irq_2_pin_ptr; + + if (pin) { + irq_2_pin_ptr = pin->next; + pin->next = NULL; + return pin; + } + + /* + * we run out of pre-allocate ones, allocate more + */ + printk(KERN_DEBUG "try to get more irq_2_pin %d\n", nr_irq_2_pin); + + if (after_bootmem) + pin = kzalloc(sizeof(struct irq_pin_list)*nr_irq_2_pin, + GFP_ATOMIC); + else + pin = __alloc_bootmem_nopanic(sizeof(struct irq_pin_list) * + nr_irq_2_pin, PAGE_SIZE, 0); + + if (!pin) + panic("can not get more irq_2_pin\n"); + + for (i = 1; i < nr_irq_2_pin; i++) + pin[i-1].next = &pin[i]; + + irq_2_pin_ptr = pin->next; + pin->next = NULL; + + return pin; +} struct io_apic { unsigned int index; @@ -167,27 +378,33 @@ static inline void io_apic_write(unsigned int apic, unsigned int reg, unsigned i /* * Re-write a value: to be used for read-modify-write * cycles where the read already set up the index register. + * + * Older SiS APIC requires we rewrite the index register */ -static inline void io_apic_modify(unsigned int apic, unsigned int value) +static inline void io_apic_modify(unsigned int apic, unsigned int reg, unsigned int value) { struct io_apic __iomem *io_apic = io_apic_base(apic); + if (sis_apic_bug) + writel(reg, &io_apic->index); writel(value, &io_apic->data); } +#ifdef CONFIG_X86_64 static bool io_apic_level_ack_pending(unsigned int irq) { struct irq_pin_list *entry; unsigned long flags; + struct irq_cfg *cfg = irq_cfg(irq); spin_lock_irqsave(&ioapic_lock, flags); - entry = irq_2_pin + irq; + entry = cfg->irq_2_pin; for (;;) { unsigned int reg; int pin; - pin = entry->pin; - if (pin == -1) + if (!entry) break; + pin = entry->pin; reg = io_apic_read(entry->apic, 0x10 + pin*2); /* Is the remote IRR bit set? */ if (reg & IO_APIC_REDIR_REMOTE_IRR) { @@ -196,44 +413,13 @@ static bool io_apic_level_ack_pending(unsigned int irq) } if (!entry->next) break; - entry = irq_2_pin + entry->next; + entry = entry->next; } spin_unlock_irqrestore(&ioapic_lock, flags); return false; } - -/* - * Synchronize the IO-APIC and the CPU by doing - * a dummy read from the IO-APIC - */ -static inline void io_apic_sync(unsigned int apic) -{ - struct io_apic __iomem *io_apic = io_apic_base(apic); - readl(&io_apic->data); -} - -#define __DO_ACTION(R, ACTION, FINAL) \ - \ -{ \ - int pin; \ - struct irq_pin_list *entry = irq_2_pin + irq; \ - \ - BUG_ON(irq >= NR_IRQS); \ - for (;;) { \ - unsigned int reg; \ - pin = entry->pin; \ - if (pin == -1) \ - break; \ - reg = io_apic_read(entry->apic, 0x10 + R + pin*2); \ - reg ACTION; \ - io_apic_modify(entry->apic, reg); \ - FINAL; \ - if (!entry->next) \ - break; \ - entry = irq_2_pin + entry->next; \ - } \ -} +#endif union entry_union { struct { u32 w1, w2; }; @@ -294,54 +480,71 @@ static void ioapic_mask_entry(int apic, int pin) static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, u8 vector) { int apic, pin; - struct irq_pin_list *entry = irq_2_pin + irq; + struct irq_cfg *cfg; + struct irq_pin_list *entry; - BUG_ON(irq >= NR_IRQS); + cfg = irq_cfg(irq); + entry = cfg->irq_2_pin; for (;;) { unsigned int reg; + + if (!entry) + break; + apic = entry->apic; pin = entry->pin; - if (pin == -1) - break; +#ifdef CONFIG_INTR_REMAP + /* + * With interrupt-remapping, destination information comes + * from interrupt-remapping table entry. + */ + if (!irq_remapped(irq)) + io_apic_write(apic, 0x11 + pin*2, dest); +#else io_apic_write(apic, 0x11 + pin*2, dest); +#endif reg = io_apic_read(apic, 0x10 + pin*2); reg &= ~IO_APIC_REDIR_VECTOR_MASK; reg |= vector; - io_apic_modify(apic, reg); + io_apic_modify(apic, 0x10 + pin*2, reg); if (!entry->next) break; - entry = irq_2_pin + entry->next; + entry = entry->next; } } +static int assign_irq_vector(int irq, cpumask_t mask); + static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) { - struct irq_cfg *cfg = irq_cfg + irq; + struct irq_cfg *cfg; unsigned long flags; unsigned int dest; cpumask_t tmp; + struct irq_desc *desc; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) return; + cfg = irq_cfg(irq); if (assign_irq_vector(irq, mask)) return; cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); - /* * Only the high 8 bits are valid. */ dest = SET_APIC_LOGICAL_ID(dest); + desc = irq_to_desc(irq); spin_lock_irqsave(&ioapic_lock, flags); __target_IO_APIC_irq(irq, dest, cfg->vector); - irq_desc[irq].affinity = mask; + desc->affinity = mask; spin_unlock_irqrestore(&ioapic_lock, flags); } -#endif +#endif /* CONFIG_SMP */ /* * The common case is 1:1 IRQ<->pin mappings. Sometimes there are @@ -350,21 +553,34 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) */ static void add_pin_to_irq(unsigned int irq, int apic, int pin) { - static int first_free_entry = NR_IRQS; - struct irq_pin_list *entry = irq_2_pin + irq; + struct irq_cfg *cfg; + struct irq_pin_list *entry; + + /* first time to refer irq_cfg, so with new */ + cfg = irq_cfg_alloc(irq); + entry = cfg->irq_2_pin; + if (!entry) { + entry = get_one_free_irq_2_pin(); + cfg->irq_2_pin = entry; + entry->apic = apic; + entry->pin = pin; + printk(KERN_DEBUG " 0 add_pin_to_irq: irq %d --> apic %d pin %d\n", irq, apic, pin); + return; + } - BUG_ON(irq >= NR_IRQS); - while (entry->next) - entry = irq_2_pin + entry->next; + while (entry->next) { + /* not again, please */ + if (entry->apic == apic && entry->pin == pin) + return; - if (entry->pin != -1) { - entry->next = first_free_entry; - entry = irq_2_pin + entry->next; - if (++first_free_entry >= PIN_MAP_SIZE) - panic("io_apic.c: ran out of irq_2_pin entries!"); + entry = entry->next; } + + entry->next = get_one_free_irq_2_pin(); + entry = entry->next; entry->apic = apic; entry->pin = pin; + printk(KERN_DEBUG " x add_pin_to_irq: irq %d --> apic %d pin %d\n", irq, apic, pin); } /* @@ -374,19 +590,60 @@ static void __init replace_pin_at_irq(unsigned int irq, int oldapic, int oldpin, int newapic, int newpin) { - struct irq_pin_list *entry = irq_2_pin + irq; + struct irq_cfg *cfg = irq_cfg(irq); + struct irq_pin_list *entry = cfg->irq_2_pin; + int replaced = 0; - while (1) { + while (entry) { if (entry->apic == oldapic && entry->pin == oldpin) { entry->apic = newapic; entry->pin = newpin; - } - if (!entry->next) + replaced = 1; + /* every one is different, right? */ break; - entry = irq_2_pin + entry->next; + } + entry = entry->next; } + + /* why? call replace before add? */ + if (!replaced) + add_pin_to_irq(irq, newapic, newpin); } +#ifdef CONFIG_X86_64 +/* + * Synchronize the IO-APIC and the CPU by doing + * a dummy read from the IO-APIC + */ +static inline void io_apic_sync(unsigned int apic) +{ + struct io_apic __iomem *io_apic = io_apic_base(apic); + readl(&io_apic->data); +} + +#define __DO_ACTION(R, ACTION, FINAL) \ + \ +{ \ + int pin; \ + struct irq_cfg *cfg; \ + struct irq_pin_list *entry; \ + \ + cfg = irq_cfg(irq); \ + entry = cfg->irq_2_pin; \ + for (;;) { \ + unsigned int reg; \ + if (!entry) \ + break; \ + pin = entry->pin; \ + reg = io_apic_read(entry->apic, 0x10 + R + pin*2); \ + reg ACTION; \ + io_apic_modify(entry->apic, 0x10 + R + pin*2, reg); \ + FINAL; \ + if (!entry->next) \ + break; \ + entry = entry->next; \ + } \ +} #define DO_ACTION(name,R,ACTION, FINAL) \ \ @@ -399,6 +656,58 @@ DO_ACTION(__mask, 0, |= IO_APIC_REDIR_MASKED, io_apic_sync(entry->apic)) /* mask = 0 */ DO_ACTION(__unmask, 0, &= ~IO_APIC_REDIR_MASKED, ) +#else + +static void __modify_IO_APIC_irq(unsigned int irq, unsigned long enable, unsigned long disable) +{ + struct irq_cfg *cfg; + struct irq_pin_list *entry; + unsigned int pin, reg; + + cfg = irq_cfg(irq); + entry = cfg->irq_2_pin; + for (;;) { + if (!entry) + break; + pin = entry->pin; + reg = io_apic_read(entry->apic, 0x10 + pin*2); + reg &= ~disable; + reg |= enable; + io_apic_modify(entry->apic, 0x10 + pin*2, reg); + if (!entry->next) + break; + entry = entry->next; + } +} + +/* mask = 1 */ +static void __mask_IO_APIC_irq(unsigned int irq) +{ + __modify_IO_APIC_irq(irq, IO_APIC_REDIR_MASKED, 0); +} + +/* mask = 0 */ +static void __unmask_IO_APIC_irq(unsigned int irq) +{ + __modify_IO_APIC_irq(irq, 0, IO_APIC_REDIR_MASKED); +} + +/* mask = 1, trigger = 0 */ +static void __mask_and_edge_IO_APIC_irq(unsigned int irq) +{ + __modify_IO_APIC_irq(irq, IO_APIC_REDIR_MASKED, + IO_APIC_REDIR_LEVEL_TRIGGER); +} + +/* mask = 0, trigger = 1 */ +static void __unmask_and_level_IO_APIC_irq(unsigned int irq) +{ + __modify_IO_APIC_irq(irq, IO_APIC_REDIR_LEVEL_TRIGGER, + IO_APIC_REDIR_MASKED); +} + +#endif + static void mask_IO_APIC_irq (unsigned int irq) { unsigned long flags; @@ -440,24 +749,131 @@ static void clear_IO_APIC (void) clear_IO_APIC_pin(apic, pin); } -int skip_ioapic_setup; -int ioapic_force; +#if !defined(CONFIG_SMP) && defined(CONFIG_X86_32) +void send_IPI_self(int vector) +{ + unsigned int cfg; -static int __init parse_noapic(char *str) + /* + * Wait for idle. + */ + apic_wait_icr_idle(); + cfg = APIC_DM_FIXED | APIC_DEST_SELF | vector | APIC_DEST_LOGICAL; + /* + * Send the IPI. The write to APIC_ICR fires this off. + */ + apic_write(APIC_ICR, cfg); +} +#endif /* !CONFIG_SMP && CONFIG_X86_32*/ + +#ifdef CONFIG_X86_32 +/* + * support for broken MP BIOSs, enables hand-redirection of PIRQ0-7 to + * specific CPU-side IRQs. + */ + +#define MAX_PIRQS 8 +static int pirq_entries [MAX_PIRQS]; +static int pirqs_enabled; + +static int __init ioapic_pirq_setup(char *str) { - disable_ioapic_setup(); + int i, max; + int ints[MAX_PIRQS+1]; + + get_options(str, ARRAY_SIZE(ints), ints); + + for (i = 0; i < MAX_PIRQS; i++) + pirq_entries[i] = -1; + + pirqs_enabled = 1; + apic_printk(APIC_VERBOSE, KERN_INFO + "PIRQ redirection, working around broken MP-BIOS.\n"); + max = MAX_PIRQS; + if (ints[0] < MAX_PIRQS) + max = ints[0]; + + for (i = 0; i < max; i++) { + apic_printk(APIC_VERBOSE, KERN_DEBUG + "... PIRQ%d -> IRQ %d\n", i, ints[i+1]); + /* + * PIRQs are mapped upside down, usually. + */ + pirq_entries[MAX_PIRQS-i-1] = ints[i+1]; + } + return 1; +} + +__setup("pirq=", ioapic_pirq_setup); +#endif /* CONFIG_X86_32 */ + +#ifdef CONFIG_INTR_REMAP +/* I/O APIC RTE contents at the OS boot up */ +static struct IO_APIC_route_entry *early_ioapic_entries[MAX_IO_APICS]; + +/* + * Saves and masks all the unmasked IO-APIC RTE's + */ +int save_mask_IO_APIC_setup(void) +{ + union IO_APIC_reg_01 reg_01; + unsigned long flags; + int apic, pin; + + /* + * The number of IO-APIC IRQ registers (== #pins): + */ + for (apic = 0; apic < nr_ioapics; apic++) { + spin_lock_irqsave(&ioapic_lock, flags); + reg_01.raw = io_apic_read(apic, 1); + spin_unlock_irqrestore(&ioapic_lock, flags); + nr_ioapic_registers[apic] = reg_01.bits.entries+1; + } + + for (apic = 0; apic < nr_ioapics; apic++) { + early_ioapic_entries[apic] = + kzalloc(sizeof(struct IO_APIC_route_entry) * + nr_ioapic_registers[apic], GFP_KERNEL); + if (!early_ioapic_entries[apic]) + return -ENOMEM; + } + + for (apic = 0; apic < nr_ioapics; apic++) + for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) { + struct IO_APIC_route_entry entry; + + entry = early_ioapic_entries[apic][pin] = + ioapic_read_entry(apic, pin); + if (!entry.mask) { + entry.mask = 1; + ioapic_write_entry(apic, pin, entry); + } + } return 0; } -early_param("noapic", parse_noapic); -/* Actually the next is obsolete, but keep it for paranoid reasons -AK */ -static int __init disable_timer_pin_setup(char *arg) +void restore_IO_APIC_setup(void) { - disable_timer_pin_1 = 1; - return 1; + int apic, pin; + + for (apic = 0; apic < nr_ioapics; apic++) + for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) + ioapic_write_entry(apic, pin, + early_ioapic_entries[apic][pin]); } -__setup("disable_timer_pin_1", disable_timer_pin_setup); +void reinit_intr_remapped_IO_APIC(int intr_remapping) +{ + /* + * for now plain restore of previous settings. + * TBD: In the case of OS enabling interrupt-remapping, + * IO-APIC RTE's need to be setup to point to interrupt-remapping + * table entries. for now, do a plain restore, and wait for + * the setup_IO_APIC_irqs() to do proper initialization. + */ + restore_IO_APIC_setup(); +} +#endif /* * Find the IRQ entry number of a certain pin. @@ -561,22 +977,54 @@ int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin) best_guess = irq; } } - BUG_ON(best_guess >= NR_IRQS); return best_guess; } +EXPORT_SYMBOL(IO_APIC_get_PCI_irq_vector); + +#if defined(CONFIG_EISA) || defined(CONFIG_MCA) +/* + * EISA Edge/Level control register, ELCR + */ +static int EISA_ELCR(unsigned int irq) +{ + if (irq < 16) { + unsigned int port = 0x4d0 + (irq >> 3); + return (inb(port) >> (irq & 7)) & 1; + } + apic_printk(APIC_VERBOSE, KERN_INFO + "Broken MPtable reports ISA irq %d\n", irq); + return 0; +} + +#endif + /* ISA interrupts are always polarity zero edge triggered, * when listed as conforming in the MP table. */ #define default_ISA_trigger(idx) (0) #define default_ISA_polarity(idx) (0) +/* EISA interrupts are always polarity zero and can be edge or level + * trigger depending on the ELCR value. If an interrupt is listed as + * EISA conforming in the MP table, that means its trigger type must + * be read in from the ELCR */ + +#define default_EISA_trigger(idx) (EISA_ELCR(mp_irqs[idx].mp_srcbusirq)) +#define default_EISA_polarity(idx) default_ISA_polarity(idx) + /* PCI interrupts are always polarity one level triggered, * when listed as conforming in the MP table. */ #define default_PCI_trigger(idx) (1) #define default_PCI_polarity(idx) (1) +/* MCA interrupts are always polarity zero level triggered, + * when listed as conforming in the MP table. */ + +#define default_MCA_trigger(idx) (1) +#define default_MCA_polarity(idx) default_ISA_polarity(idx) + static int MPBIOS_polarity(int idx) { int bus = mp_irqs[idx].mp_srcbus; @@ -634,6 +1082,36 @@ static int MPBIOS_trigger(int idx) trigger = default_ISA_trigger(idx); else trigger = default_PCI_trigger(idx); +#if defined(CONFIG_EISA) || defined(CONFIG_MCA) + switch (mp_bus_id_to_type[bus]) { + case MP_BUS_ISA: /* ISA pin */ + { + /* set before the switch */ + break; + } + case MP_BUS_EISA: /* EISA pin */ + { + trigger = default_EISA_trigger(idx); + break; + } + case MP_BUS_PCI: /* PCI pin */ + { + /* set before the switch */ + break; + } + case MP_BUS_MCA: /* MCA pin */ + { + trigger = default_MCA_trigger(idx); + break; + } + default: + { + printk(KERN_WARNING "broken BIOS!!\n"); + trigger = 1; + break; + } + } +#endif break; case 1: /* edge */ { @@ -671,6 +1149,7 @@ static inline int irq_trigger(int idx) return MPBIOS_trigger(idx); } +int (*ioapic_renumber_irq)(int ioapic, int irq); static int pin_2_irq(int idx, int apic, int pin) { int irq, i; @@ -692,8 +1171,32 @@ static int pin_2_irq(int idx, int apic, int pin) while (i < apic) irq += nr_ioapic_registers[i++]; irq += pin; + /* + * For MPS mode, so far only needed by ES7000 platform + */ + if (ioapic_renumber_irq) + irq = ioapic_renumber_irq(apic, irq); } - BUG_ON(irq >= NR_IRQS); + +#ifdef CONFIG_X86_32 + /* + * PCI IRQ command line redirection. Yes, limits are hardcoded. + */ + if ((pin >= 16) && (pin <= 23)) { + if (pirq_entries[pin-16] != -1) { + if (!pirq_entries[pin-16]) { + apic_printk(APIC_VERBOSE, KERN_DEBUG + "disabling PIRQ%d\n", pin-16); + } else { + irq = pirq_entries[pin-16]; + apic_printk(APIC_VERBOSE, KERN_DEBUG + "using PIRQ%d -> IRQ %d\n", + pin-16, irq); + } + } + } +#endif + return irq; } @@ -728,8 +1231,7 @@ static int __assign_irq_vector(int irq, cpumask_t mask) int cpu; struct irq_cfg *cfg; - BUG_ON((unsigned)irq >= NR_IRQS); - cfg = &irq_cfg[irq]; + cfg = irq_cfg(irq); /* Only try and allocate irqs on cpus that are present */ cpus_and(mask, mask, cpu_online_map); @@ -764,8 +1266,13 @@ next: } if (unlikely(current_vector == vector)) continue; +#ifdef CONFIG_X86_64 if (vector == IA32_SYSCALL_VECTOR) goto next; +#else + if (vector == SYSCALL_VECTOR) + goto next; +#endif for_each_cpu_mask_nr(new_cpu, new_mask) if (per_cpu(vector_irq, new_cpu)[vector] != -1) goto next; @@ -802,8 +1309,7 @@ static void __clear_irq_vector(int irq) cpumask_t mask; int cpu, vector; - BUG_ON((unsigned)irq >= NR_IRQS); - cfg = &irq_cfg[irq]; + cfg = irq_cfg(irq); BUG_ON(!cfg->vector); vector = cfg->vector; @@ -820,12 +1326,14 @@ void __setup_vector_irq(int cpu) /* Initialize vector_irq on a new cpu */ /* This function must be called with vector_lock held */ int irq, vector; + struct irq_cfg *cfg; /* Mark the inuse vectors */ - for (irq = 0; irq < NR_IRQS; ++irq) { - if (!cpu_isset(cpu, irq_cfg[irq].domain)) + for_each_irq_cfg(cfg) { + if (!cpu_isset(cpu, cfg->domain)) continue; - vector = irq_cfg[irq].vector; + vector = cfg->vector; + irq = cfg->irq; per_cpu(vector_irq, cpu)[vector] = irq; } /* Mark the free vectors */ @@ -833,36 +1341,158 @@ void __setup_vector_irq(int cpu) irq = per_cpu(vector_irq, cpu)[vector]; if (irq < 0) continue; - if (!cpu_isset(cpu, irq_cfg[irq].domain)) + + cfg = irq_cfg(irq); + if (!cpu_isset(cpu, cfg->domain)) per_cpu(vector_irq, cpu)[vector] = -1; } } static struct irq_chip ioapic_chip; +#ifdef CONFIG_INTR_REMAP +static struct irq_chip ir_ioapic_chip; +#endif + +#define IOAPIC_AUTO -1 +#define IOAPIC_EDGE 0 +#define IOAPIC_LEVEL 1 + +#ifdef CONFIG_X86_32 +static inline int IO_APIC_irq_trigger(int irq) +{ + int apic, idx, pin; + + for (apic = 0; apic < nr_ioapics; apic++) { + for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) { + idx = find_irq_entry(apic, pin, mp_INT); + if ((idx != -1) && (irq == pin_2_irq(idx, apic, pin))) + return irq_trigger(idx); + } + } + /* + * nonexistent IRQs are edge default + */ + return 0; +} +#else +static inline int IO_APIC_irq_trigger(int irq) +{ + return 1; +} +#endif static void ioapic_register_intr(int irq, unsigned long trigger) { - if (trigger) { - irq_desc[irq].status |= IRQ_LEVEL; + struct irq_desc *desc; + + /* first time to use this irq_desc */ + if (irq < 16) + desc = irq_to_desc(irq); + else + desc = irq_to_desc_alloc(irq); + + if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) || + trigger == IOAPIC_LEVEL) + desc->status |= IRQ_LEVEL; + else + desc->status &= ~IRQ_LEVEL; + +#ifdef CONFIG_INTR_REMAP + if (irq_remapped(irq)) { + desc->status |= IRQ_MOVE_PCNTXT; + if (trigger) + set_irq_chip_and_handler_name(irq, &ir_ioapic_chip, + handle_fasteoi_irq, + "fasteoi"); + else + set_irq_chip_and_handler_name(irq, &ir_ioapic_chip, + handle_edge_irq, "edge"); + return; + } +#endif + if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) || + trigger == IOAPIC_LEVEL) set_irq_chip_and_handler_name(irq, &ioapic_chip, - handle_fasteoi_irq, "fasteoi"); - } else { - irq_desc[irq].status &= ~IRQ_LEVEL; + handle_fasteoi_irq, + "fasteoi"); + else set_irq_chip_and_handler_name(irq, &ioapic_chip, handle_edge_irq, "edge"); +} + +static int setup_ioapic_entry(int apic, int irq, + struct IO_APIC_route_entry *entry, + unsigned int destination, int trigger, + int polarity, int vector) +{ + /* + * add it to the IO-APIC irq-routing table: + */ + memset(entry,0,sizeof(*entry)); + +#ifdef CONFIG_INTR_REMAP + if (intr_remapping_enabled) { + struct intel_iommu *iommu = map_ioapic_to_ir(apic); + struct irte irte; + struct IR_IO_APIC_route_entry *ir_entry = + (struct IR_IO_APIC_route_entry *) entry; + int index; + + if (!iommu) + panic("No mapping iommu for ioapic %d\n", apic); + + index = alloc_irte(iommu, irq, 1); + if (index < 0) + panic("Failed to allocate IRTE for ioapic %d\n", apic); + + memset(&irte, 0, sizeof(irte)); + + irte.present = 1; + irte.dst_mode = INT_DEST_MODE; + irte.trigger_mode = trigger; + irte.dlvry_mode = INT_DELIVERY_MODE; + irte.vector = vector; + irte.dest_id = IRTE_DEST(destination); + + modify_irte(irq, &irte); + + ir_entry->index2 = (index >> 15) & 0x1; + ir_entry->zero = 0; + ir_entry->format = 1; + ir_entry->index = (index & 0x7fff); + } else +#endif + { + entry->delivery_mode = INT_DELIVERY_MODE; + entry->dest_mode = INT_DEST_MODE; + entry->dest = destination; } + + entry->mask = 0; /* enable IRQ */ + entry->trigger = trigger; + entry->polarity = polarity; + entry->vector = vector; + + /* Mask level triggered irqs. + * Use IRQ_DELAYED_DISABLE for edge triggered irqs. + */ + if (trigger) + entry->mask = 1; + return 0; } static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, int trigger, int polarity) { - struct irq_cfg *cfg = irq_cfg + irq; + struct irq_cfg *cfg; struct IO_APIC_route_entry entry; cpumask_t mask; if (!IO_APIC_IRQ(irq)) return; + cfg = irq_cfg(irq); + mask = TARGET_CPUS; if (assign_irq_vector(irq, mask)) return; @@ -875,24 +1505,15 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, apic, mp_ioapics[apic].mp_apicid, pin, cfg->vector, irq, trigger, polarity); - /* - * add it to the IO-APIC irq-routing table: - */ - memset(&entry,0,sizeof(entry)); - - entry.delivery_mode = INT_DELIVERY_MODE; - entry.dest_mode = INT_DEST_MODE; - entry.dest = cpu_mask_to_apicid(mask); - entry.mask = 0; /* enable IRQ */ - entry.trigger = trigger; - entry.polarity = polarity; - entry.vector = cfg->vector; - /* Mask level triggered irqs. - * Use IRQ_DELAYED_DISABLE for edge triggered irqs. - */ - if (trigger) - entry.mask = 1; + if (setup_ioapic_entry(mp_ioapics[apic].mp_apicid, irq, &entry, + cpu_mask_to_apicid(mask), trigger, polarity, + cfg->vector)) { + printk("Failed to setup ioapic entry for ioapic %d, pin %d\n", + mp_ioapics[apic].mp_apicid, pin); + __clear_irq_vector(irq); + return; + } ioapic_register_intr(irq, trigger); if (irq < 16) @@ -925,6 +1546,10 @@ static void __init setup_IO_APIC_irqs(void) } irq = pin_2_irq(idx, apic, pin); +#ifdef CONFIG_X86_32 + if (multi_timer_check(apic, irq)) + continue; +#endif add_pin_to_irq(irq, apic, pin); setup_IO_APIC_irq(apic, pin, irq, @@ -944,6 +1569,11 @@ static void __init setup_timer_IRQ0_pin(unsigned int apic, unsigned int pin, { struct IO_APIC_route_entry entry; +#ifdef CONFIG_INTR_REMAP + if (intr_remapping_enabled) + return; +#endif + memset(&entry, 0, sizeof(entry)); /* @@ -970,13 +1600,16 @@ static void __init setup_timer_IRQ0_pin(unsigned int apic, unsigned int pin, ioapic_write_entry(apic, pin, entry); } -void __apicdebuginit print_IO_APIC(void) + +__apicdebuginit(void) print_IO_APIC(void) { int apic, i; union IO_APIC_reg_00 reg_00; union IO_APIC_reg_01 reg_01; union IO_APIC_reg_02 reg_02; + union IO_APIC_reg_03 reg_03; unsigned long flags; + struct irq_cfg *cfg; if (apic_verbosity == APIC_QUIET) return; @@ -999,12 +1632,16 @@ void __apicdebuginit print_IO_APIC(void) reg_01.raw = io_apic_read(apic, 1); if (reg_01.bits.version >= 0x10) reg_02.raw = io_apic_read(apic, 2); + if (reg_01.bits.version >= 0x20) + reg_03.raw = io_apic_read(apic, 3); spin_unlock_irqrestore(&ioapic_lock, flags); printk("\n"); printk(KERN_DEBUG "IO APIC #%d......\n", mp_ioapics[apic].mp_apicid); printk(KERN_DEBUG ".... register #00: %08X\n", reg_00.raw); printk(KERN_DEBUG "....... : physical APIC id: %02X\n", reg_00.bits.ID); + printk(KERN_DEBUG "....... : Delivery Type: %X\n", reg_00.bits.delivery_type); + printk(KERN_DEBUG "....... : LTS : %X\n", reg_00.bits.LTS); printk(KERN_DEBUG ".... register #01: %08X\n", *(int *)®_01); printk(KERN_DEBUG "....... : max redirection entries: %04X\n", reg_01.bits.entries); @@ -1012,11 +1649,27 @@ void __apicdebuginit print_IO_APIC(void) printk(KERN_DEBUG "....... : PRQ implemented: %X\n", reg_01.bits.PRQ); printk(KERN_DEBUG "....... : IO APIC version: %04X\n", reg_01.bits.version); - if (reg_01.bits.version >= 0x10) { + /* + * Some Intel chipsets with IO APIC VERSION of 0x1? don't have reg_02, + * but the value of reg_02 is read as the previous read register + * value, so ignore it if reg_02 == reg_01. + */ + if (reg_01.bits.version >= 0x10 && reg_02.raw != reg_01.raw) { printk(KERN_DEBUG ".... register #02: %08X\n", reg_02.raw); printk(KERN_DEBUG "....... : arbitration: %02X\n", reg_02.bits.arbitration); } + /* + * Some Intel chipsets with IO APIC VERSION of 0x2? don't have reg_02 + * or reg_03, but the value of reg_0[23] is read as the previous read + * register value, so ignore it if reg_03 == reg_0[12]. + */ + if (reg_01.bits.version >= 0x20 && reg_03.raw != reg_02.raw && + reg_03.raw != reg_01.raw) { + printk(KERN_DEBUG ".... register #03: %08X\n", reg_03.raw); + printk(KERN_DEBUG "....... : Boot DT : %X\n", reg_03.bits.boot_DT); + } + printk(KERN_DEBUG ".... IRQ redirection table:\n"); printk(KERN_DEBUG " NR Dst Mask Trig IRR Pol" @@ -1045,16 +1698,16 @@ void __apicdebuginit print_IO_APIC(void) } } printk(KERN_DEBUG "IRQ to pin mappings:\n"); - for (i = 0; i < NR_IRQS; i++) { - struct irq_pin_list *entry = irq_2_pin + i; - if (entry->pin < 0) + for_each_irq_cfg(cfg) { + struct irq_pin_list *entry = cfg->irq_2_pin; + if (!entry) continue; - printk(KERN_DEBUG "IRQ%d ", i); + printk(KERN_DEBUG "IRQ%d ", cfg->irq); for (;;) { printk("-> %d:%d", entry->apic, entry->pin); if (!entry->next) break; - entry = irq_2_pin + entry->next; + entry = entry->next; } printk("\n"); } @@ -1064,9 +1717,7 @@ void __apicdebuginit print_IO_APIC(void) return; } -#if 0 - -static __apicdebuginit void print_APIC_bitfield (int base) +__apicdebuginit(void) print_APIC_bitfield(int base) { unsigned int v; int i, j; @@ -1087,9 +1738,10 @@ static __apicdebuginit void print_APIC_bitfield (int base) } } -void __apicdebuginit print_local_APIC(void * dummy) +__apicdebuginit(void) print_local_APIC(void *dummy) { unsigned int v, ver, maxlvt; + u64 icr; if (apic_verbosity == APIC_QUIET) return; @@ -1097,7 +1749,7 @@ void __apicdebuginit print_local_APIC(void * dummy) printk("\n" KERN_DEBUG "printing local APIC contents on CPU#%d/%d:\n", smp_processor_id(), hard_smp_processor_id()); v = apic_read(APIC_ID); - printk(KERN_INFO "... APIC ID: %08x (%01x)\n", v, GET_APIC_ID(read_apic_id())); + printk(KERN_INFO "... APIC ID: %08x (%01x)\n", v, read_apic_id()); v = apic_read(APIC_LVR); printk(KERN_INFO "... APIC VERSION: %08x\n", v); ver = GET_APIC_VERSION(v); @@ -1106,11 +1758,13 @@ void __apicdebuginit print_local_APIC(void * dummy) v = apic_read(APIC_TASKPRI); printk(KERN_DEBUG "... APIC TASKPRI: %08x (%02x)\n", v, v & APIC_TPRI_MASK); - v = apic_read(APIC_ARBPRI); - printk(KERN_DEBUG "... APIC ARBPRI: %08x (%02x)\n", v, - v & APIC_ARBPRI_MASK); - v = apic_read(APIC_PROCPRI); - printk(KERN_DEBUG "... APIC PROCPRI: %08x\n", v); + if (APIC_INTEGRATED(ver)) { /* !82489DX */ + v = apic_read(APIC_ARBPRI); + printk(KERN_DEBUG "... APIC ARBPRI: %08x (%02x)\n", v, + v & APIC_ARBPRI_MASK); + v = apic_read(APIC_PROCPRI); + printk(KERN_DEBUG "... APIC PROCPRI: %08x\n", v); + } v = apic_read(APIC_EOI); printk(KERN_DEBUG "... APIC EOI: %08x\n", v); @@ -1130,13 +1784,17 @@ void __apicdebuginit print_local_APIC(void * dummy) printk(KERN_DEBUG "... APIC IRR field:\n"); print_APIC_bitfield(APIC_IRR); - v = apic_read(APIC_ESR); - printk(KERN_DEBUG "... APIC ESR: %08x\n", v); + if (APIC_INTEGRATED(ver)) { /* !82489DX */ + if (maxlvt > 3) /* Due to the Pentium erratum 3AP. */ + apic_write(APIC_ESR, 0); + + v = apic_read(APIC_ESR); + printk(KERN_DEBUG "... APIC ESR: %08x\n", v); + } - v = apic_read(APIC_ICR); - printk(KERN_DEBUG "... APIC ICR: %08x\n", v); - v = apic_read(APIC_ICR2); - printk(KERN_DEBUG "... APIC ICR2: %08x\n", v); + icr = apic_icr_read(); + printk(KERN_DEBUG "... APIC ICR: %08x\n", (u32)icr); + printk(KERN_DEBUG "... APIC ICR2: %08x\n", (u32)(icr >> 32)); v = apic_read(APIC_LVTT); printk(KERN_DEBUG "... APIC LVTT: %08x\n", v); @@ -1164,12 +1822,12 @@ void __apicdebuginit print_local_APIC(void * dummy) printk("\n"); } -void print_all_local_APICs (void) +__apicdebuginit(void) print_all_local_APICs(void) { on_each_cpu(print_local_APIC, NULL, 1); } -void __apicdebuginit print_PIC(void) +__apicdebuginit(void) print_PIC(void) { unsigned int v; unsigned long flags; @@ -1201,19 +1859,34 @@ void __apicdebuginit print_PIC(void) printk(KERN_DEBUG "... PIC ELCR: %04x\n", v); } -#endif /* 0 */ +__apicdebuginit(int) print_all_ICs(void) +{ + print_PIC(); + print_all_local_APICs(); + print_IO_APIC(); + + return 0; +} + +fs_initcall(print_all_ICs); + + +/* Where if anywhere is the i8259 connect in external int mode */ +static struct { int pin, apic; } ioapic_i8259 = { -1, -1 }; void __init enable_IO_APIC(void) { union IO_APIC_reg_01 reg_01; int i8259_apic, i8259_pin; - int i, apic; + int apic; unsigned long flags; - for (i = 0; i < PIN_MAP_SIZE; i++) { - irq_2_pin[i].pin = -1; - irq_2_pin[i].next = 0; - } +#ifdef CONFIG_X86_32 + int i; + if (!pirqs_enabled) + for (i = 0; i < MAX_PIRQS; i++) + pirq_entries[i] = -1; +#endif /* * The number of IO-APIC IRQ registers (== #pins): @@ -1243,6 +1916,10 @@ void __init enable_IO_APIC(void) } found_i8259: /* Look to see what if the MP table has reported the ExtINT */ + /* If we could not find the appropriate pin by looking at the ioapic + * the i8259 probably is not connected the ioapic but give the + * mptable a chance anyway. + */ i8259_pin = find_isa_irq_pin(0, mp_ExtINT); i8259_apic = find_isa_irq_apic(0, mp_ExtINT); /* Trust the MP table if nothing is setup in the hardware */ @@ -1291,7 +1968,7 @@ void disable_IO_APIC(void) entry.dest_mode = 0; /* Physical */ entry.delivery_mode = dest_ExtINT; /* ExtInt */ entry.vector = 0; - entry.dest = GET_APIC_ID(read_apic_id()); + entry.dest = read_apic_id(); /* * Add it to the IO-APIC irq-routing table: @@ -1302,6 +1979,131 @@ void disable_IO_APIC(void) disconnect_bsp_APIC(ioapic_i8259.pin != -1); } +#ifdef CONFIG_X86_32 +/* + * function to set the IO-APIC physical IDs based on the + * values stored in the MPC table. + * + * by Matt Domsch Tue Dec 21 12:25:05 CST 1999 + */ + +static void __init setup_ioapic_ids_from_mpc(void) +{ + union IO_APIC_reg_00 reg_00; + physid_mask_t phys_id_present_map; + int apic; + int i; + unsigned char old_id; + unsigned long flags; + + if (x86_quirks->setup_ioapic_ids && x86_quirks->setup_ioapic_ids()) + return; + + /* + * Don't check I/O APIC IDs for xAPIC systems. They have + * no meaning without the serial APIC bus. + */ + if (!(boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) + || APIC_XAPIC(apic_version[boot_cpu_physical_apicid])) + return; + /* + * This is broken; anything with a real cpu count has to + * circumvent this idiocy regardless. + */ + phys_id_present_map = ioapic_phys_id_map(phys_cpu_present_map); + + /* + * Set the IOAPIC ID to the value stored in the MPC table. + */ + for (apic = 0; apic < nr_ioapics; apic++) { + + /* Read the register 0 value */ + spin_lock_irqsave(&ioapic_lock, flags); + reg_00.raw = io_apic_read(apic, 0); + spin_unlock_irqrestore(&ioapic_lock, flags); + + old_id = mp_ioapics[apic].mp_apicid; + + if (mp_ioapics[apic].mp_apicid >= get_physical_broadcast()) { + printk(KERN_ERR "BIOS bug, IO-APIC#%d ID is %d in the MPC table!...\n", + apic, mp_ioapics[apic].mp_apicid); + printk(KERN_ERR "... fixing up to %d. (tell your hw vendor)\n", + reg_00.bits.ID); + mp_ioapics[apic].mp_apicid = reg_00.bits.ID; + } + + /* + * Sanity check, is the ID really free? Every APIC in a + * system must have a unique ID or we get lots of nice + * 'stuck on smp_invalidate_needed IPI wait' messages. + */ + if (check_apicid_used(phys_id_present_map, + mp_ioapics[apic].mp_apicid)) { + printk(KERN_ERR "BIOS bug, IO-APIC#%d ID %d is already used!...\n", + apic, mp_ioapics[apic].mp_apicid); + for (i = 0; i < get_physical_broadcast(); i++) + if (!physid_isset(i, phys_id_present_map)) + break; + if (i >= get_physical_broadcast()) + panic("Max APIC ID exceeded!\n"); + printk(KERN_ERR "... fixing up to %d. (tell your hw vendor)\n", + i); + physid_set(i, phys_id_present_map); + mp_ioapics[apic].mp_apicid = i; + } else { + physid_mask_t tmp; + tmp = apicid_to_cpu_present(mp_ioapics[apic].mp_apicid); + apic_printk(APIC_VERBOSE, "Setting %d in the " + "phys_id_present_map\n", + mp_ioapics[apic].mp_apicid); + physids_or(phys_id_present_map, phys_id_present_map, tmp); + } + + + /* + * We need to adjust the IRQ routing table + * if the ID changed. + */ + if (old_id != mp_ioapics[apic].mp_apicid) + for (i = 0; i < mp_irq_entries; i++) + if (mp_irqs[i].mp_dstapic == old_id) + mp_irqs[i].mp_dstapic + = mp_ioapics[apic].mp_apicid; + + /* + * Read the right value from the MPC table and + * write it into the ID register. + */ + apic_printk(APIC_VERBOSE, KERN_INFO + "...changing IO-APIC physical APIC ID to %d ...", + mp_ioapics[apic].mp_apicid); + + reg_00.bits.ID = mp_ioapics[apic].mp_apicid; + spin_lock_irqsave(&ioapic_lock, flags); + + /* + * Sanity check + */ + spin_lock_irqsave(&ioapic_lock, flags); + reg_00.raw = io_apic_read(apic, 0); + spin_unlock_irqrestore(&ioapic_lock, flags); + if (reg_00.bits.ID != mp_ioapics[apic].mp_apicid) + printk("could not set ID!\n"); + else + apic_printk(APIC_VERBOSE, " ok.\n"); + } +} +#endif + +int no_timer_check __initdata; + +static int __init notimercheck(char *s) +{ + no_timer_check = 1; + return 1; +} +__setup("no_timer_check", notimercheck); + /* * There is a nasty bug in some older SMP boards, their mptable lies * about the timer IRQ. We do the following to work around the situation: @@ -1315,6 +2117,9 @@ static int __init timer_irq_works(void) unsigned long t1 = jiffies; unsigned long flags; + if (no_timer_check) + return 1; + local_save_flags(flags); local_irq_enable(); /* Let ten ticks pass... */ @@ -1375,9 +2180,11 @@ static unsigned int startup_ioapic_irq(unsigned int irq) return was_pending; } +#ifdef CONFIG_X86_64 static int ioapic_retrigger_irq(unsigned int irq) { - struct irq_cfg *cfg = &irq_cfg[irq]; + + struct irq_cfg *cfg = irq_cfg(irq); unsigned long flags; spin_lock_irqsave(&vector_lock, flags); @@ -1386,6 +2193,14 @@ static int ioapic_retrigger_irq(unsigned int irq) return 1; } +#else +static int ioapic_retrigger_irq(unsigned int irq) +{ + send_IPI_self(irq_cfg(irq)->vector); + + return 1; +} +#endif /* * Level and edge triggered IO-APIC interrupts need different handling, @@ -1397,11 +2212,159 @@ static int ioapic_retrigger_irq(unsigned int irq) */ #ifdef CONFIG_SMP + +#ifdef CONFIG_INTR_REMAP +static void ir_irq_migration(struct work_struct *work); + +static DECLARE_DELAYED_WORK(ir_migration_work, ir_irq_migration); + +/* + * Migrate the IO-APIC irq in the presence of intr-remapping. + * + * For edge triggered, irq migration is a simple atomic update(of vector + * and cpu destination) of IRTE and flush the hardware cache. + * + * For level triggered, we need to modify the io-apic RTE aswell with the update + * vector information, along with modifying IRTE with vector and destination. + * So irq migration for level triggered is little bit more complex compared to + * edge triggered migration. But the good news is, we use the same algorithm + * for level triggered migration as we have today, only difference being, + * we now initiate the irq migration from process context instead of the + * interrupt context. + * + * In future, when we do a directed EOI (combined with cpu EOI broadcast + * suppression) to the IO-APIC, level triggered irq migration will also be + * as simple as edge triggered migration and we can do the irq migration + * with a simple atomic update to IO-APIC RTE. + */ +static void migrate_ioapic_irq(int irq, cpumask_t mask) +{ + struct irq_cfg *cfg; + struct irq_desc *desc; + cpumask_t tmp, cleanup_mask; + struct irte irte; + int modify_ioapic_rte; + unsigned int dest; + unsigned long flags; + + cpus_and(tmp, mask, cpu_online_map); + if (cpus_empty(tmp)) + return; + + if (get_irte(irq, &irte)) + return; + + if (assign_irq_vector(irq, mask)) + return; + + cfg = irq_cfg(irq); + cpus_and(tmp, cfg->domain, mask); + dest = cpu_mask_to_apicid(tmp); + + desc = irq_to_desc(irq); + modify_ioapic_rte = desc->status & IRQ_LEVEL; + if (modify_ioapic_rte) { + spin_lock_irqsave(&ioapic_lock, flags); + __target_IO_APIC_irq(irq, dest, cfg->vector); + spin_unlock_irqrestore(&ioapic_lock, flags); + } + + irte.vector = cfg->vector; + irte.dest_id = IRTE_DEST(dest); + + /* + * Modified the IRTE and flushes the Interrupt entry cache. + */ + modify_irte(irq, &irte); + + if (cfg->move_in_progress) { + cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); + cfg->move_cleanup_count = cpus_weight(cleanup_mask); + send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); + cfg->move_in_progress = 0; + } + + desc->affinity = mask; +} + +static int migrate_irq_remapped_level(int irq) +{ + int ret = -1; + struct irq_desc *desc = irq_to_desc(irq); + + mask_IO_APIC_irq(irq); + + if (io_apic_level_ack_pending(irq)) { + /* + * Interrupt in progress. Migrating irq now will change the + * vector information in the IO-APIC RTE and that will confuse + * the EOI broadcast performed by cpu. + * So, delay the irq migration to the next instance. + */ + schedule_delayed_work(&ir_migration_work, 1); + goto unmask; + } + + /* everthing is clear. we have right of way */ + migrate_ioapic_irq(irq, desc->pending_mask); + + ret = 0; + desc->status &= ~IRQ_MOVE_PENDING; + cpus_clear(desc->pending_mask); + +unmask: + unmask_IO_APIC_irq(irq); + return ret; +} + +static void ir_irq_migration(struct work_struct *work) +{ + unsigned int irq; + struct irq_desc *desc; + + for_each_irq_desc(irq, desc) { + if (desc->status & IRQ_MOVE_PENDING) { + unsigned long flags; + + spin_lock_irqsave(&desc->lock, flags); + if (!desc->chip->set_affinity || + !(desc->status & IRQ_MOVE_PENDING)) { + desc->status &= ~IRQ_MOVE_PENDING; + spin_unlock_irqrestore(&desc->lock, flags); + continue; + } + + desc->chip->set_affinity(irq, desc->pending_mask); + spin_unlock_irqrestore(&desc->lock, flags); + } + } +} + +/* + * Migrates the IRQ destination in the process context. + */ +static void set_ir_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) +{ + struct irq_desc *desc = irq_to_desc(irq); + + if (desc->status & IRQ_LEVEL) { + desc->status |= IRQ_MOVE_PENDING; + desc->pending_mask = mask; + migrate_irq_remapped_level(irq); + return; + } + + migrate_ioapic_irq(irq, mask); +} +#endif + asmlinkage void smp_irq_move_cleanup_interrupt(void) { unsigned vector, me; ack_APIC_irq(); +#ifdef CONFIG_X86_64 exit_idle(); +#endif irq_enter(); me = smp_processor_id(); @@ -1410,11 +2373,12 @@ asmlinkage void smp_irq_move_cleanup_interrupt(void) struct irq_desc *desc; struct irq_cfg *cfg; irq = __get_cpu_var(vector_irq)[vector]; - if (irq >= NR_IRQS) + + desc = irq_to_desc(irq); + if (!desc) continue; - desc = irq_desc + irq; - cfg = irq_cfg + irq; + cfg = irq_cfg(irq); spin_lock(&desc->lock); if (!cfg->move_cleanup_count) goto unlock; @@ -1433,7 +2397,7 @@ unlock: static void irq_complete_move(unsigned int irq) { - struct irq_cfg *cfg = irq_cfg + irq; + struct irq_cfg *cfg = irq_cfg(irq); unsigned vector, me; if (likely(!cfg->move_in_progress)) @@ -1453,6 +2417,17 @@ static void irq_complete_move(unsigned int irq) #else static inline void irq_complete_move(unsigned int irq) {} #endif +#ifdef CONFIG_INTR_REMAP +static void ack_x2apic_level(unsigned int irq) +{ + ack_x2APIC_irq(); +} + +static void ack_x2apic_edge(unsigned int irq) +{ + ack_x2APIC_irq(); +} +#endif static void ack_apic_edge(unsigned int irq) { @@ -1461,6 +2436,7 @@ static void ack_apic_edge(unsigned int irq) ack_APIC_irq(); } +#ifdef CONFIG_X86_64 static void ack_apic_level(unsigned int irq) { int do_unmask_irq = 0; @@ -1468,7 +2444,7 @@ static void ack_apic_level(unsigned int irq) irq_complete_move(irq); #ifdef CONFIG_GENERIC_PENDING_IRQ /* If we are moving the irq we need to mask it */ - if (unlikely(irq_desc[irq].status & IRQ_MOVE_PENDING)) { + if (unlikely(irq_to_desc(irq)->status & IRQ_MOVE_PENDING)) { do_unmask_irq = 1; mask_IO_APIC_irq(irq); } @@ -1513,6 +2489,49 @@ static void ack_apic_level(unsigned int irq) unmask_IO_APIC_irq(irq); } } +#else +atomic_t irq_mis_count; +static void ack_apic_level(unsigned int irq) +{ + unsigned long v; + int i; + + irq_complete_move(irq); + move_native_irq(irq); + /* + * 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 + */ + i = irq_cfg(irq)->vector; + + v = apic_read(APIC_TMR + ((i & ~0x1f) >> 1)); + + ack_APIC_irq(); + + if (!(v & (1 << (i & 0x1f)))) { + atomic_inc(&irq_mis_count); + spin_lock(&ioapic_lock); + __mask_and_edge_IO_APIC_irq(irq); + __unmask_and_level_IO_APIC_irq(irq); + spin_unlock(&ioapic_lock); + } +} +#endif static struct irq_chip ioapic_chip __read_mostly = { .name = "IO-APIC", @@ -1527,9 +2546,26 @@ static struct irq_chip ioapic_chip __read_mostly = { .retrigger = ioapic_retrigger_irq, }; +#ifdef CONFIG_INTR_REMAP +static struct irq_chip ir_ioapic_chip __read_mostly = { + .name = "IR-IO-APIC", + .startup = startup_ioapic_irq, + .mask = mask_IO_APIC_irq, + .unmask = unmask_IO_APIC_irq, + .ack = ack_x2apic_edge, + .eoi = ack_x2apic_level, +#ifdef CONFIG_SMP + .set_affinity = set_ir_ioapic_affinity_irq, +#endif + .retrigger = ioapic_retrigger_irq, +}; +#endif + static inline void init_IO_APIC_traps(void) { int irq; + struct irq_desc *desc; + struct irq_cfg *cfg; /* * NOTE! The local APIC isn't very good at handling @@ -1542,8 +2578,9 @@ static inline void init_IO_APIC_traps(void) * Also, we've got to be careful not to trash gate * 0x80, because int 0x80 is hm, kind of importantish. ;) */ - for (irq = 0; irq < NR_IRQS ; irq++) { - if (IO_APIC_IRQ(irq) && !irq_cfg[irq].vector) { + for_each_irq_cfg(cfg) { + irq = cfg->irq; + if (IO_APIC_IRQ(irq) && !cfg->vector) { /* * Hmm.. We don't have an entry for this, * so default to an old-fashioned 8259 @@ -1551,27 +2588,33 @@ static inline void init_IO_APIC_traps(void) */ if (irq < 16) make_8259A_irq(irq); - else + else { + desc = irq_to_desc(irq); /* Strange. Oh, well.. */ - irq_desc[irq].chip = &no_irq_chip; + desc->chip = &no_irq_chip; + } } } } -static void unmask_lapic_irq(unsigned int irq) +/* + * The local APIC irq-chip implementation: + */ + +static void mask_lapic_irq(unsigned int irq) { unsigned long v; v = apic_read(APIC_LVT0); - apic_write(APIC_LVT0, v & ~APIC_LVT_MASKED); + apic_write(APIC_LVT0, v | APIC_LVT_MASKED); } -static void mask_lapic_irq(unsigned int irq) +static void unmask_lapic_irq(unsigned int irq) { unsigned long v; v = apic_read(APIC_LVT0); - apic_write(APIC_LVT0, v | APIC_LVT_MASKED); + apic_write(APIC_LVT0, v & ~APIC_LVT_MASKED); } static void ack_lapic_irq (unsigned int irq) @@ -1588,7 +2631,10 @@ static struct irq_chip lapic_chip __read_mostly = { static void lapic_register_intr(int irq) { - irq_desc[irq].status &= ~IRQ_LEVEL; + struct irq_desc *desc; + + desc = irq_to_desc(irq); + desc->status &= ~IRQ_LEVEL; set_irq_chip_and_handler_name(irq, &lapic_chip, handle_edge_irq, "edge"); } @@ -1596,19 +2642,19 @@ static void lapic_register_intr(int irq) static void __init setup_nmi(void) { /* - * Dirty trick to enable the NMI watchdog ... + * Dirty trick to enable the NMI watchdog ... * We put the 8259A master into AEOI mode and * unmask on all local APICs LVT0 as NMI. * * The idea to use the 8259A in AEOI mode ('8259A Virtual Wire') * is from Maciej W. Rozycki - so we do not have to EOI from * the NMI handler or the timer interrupt. - */ - printk(KERN_INFO "activating NMI Watchdog ..."); + */ + apic_printk(APIC_VERBOSE, KERN_INFO "activating NMI Watchdog ..."); enable_NMI_through_LVT0(); - printk(" done.\n"); + apic_printk(APIC_VERBOSE, " done.\n"); } /* @@ -1625,12 +2671,17 @@ static inline void __init unlock_ExtINT_logic(void) unsigned char save_control, save_freq_select; pin = find_isa_irq_pin(8, mp_INT); + if (pin == -1) { + WARN_ON_ONCE(1); + return; + } apic = find_isa_irq_apic(8, mp_INT); - if (pin == -1) + if (apic == -1) { + WARN_ON_ONCE(1); return; + } entry0 = ioapic_read_entry(apic, pin); - clear_IO_APIC_pin(apic, pin); memset(&entry1, 0, sizeof(entry1)); @@ -1665,23 +2716,38 @@ static inline void __init unlock_ExtINT_logic(void) ioapic_write_entry(apic, pin, entry0); } +static int disable_timer_pin_1 __initdata; +/* Actually the next is obsolete, but keep it for paranoid reasons -AK */ +static int __init disable_timer_pin_setup(char *arg) +{ + disable_timer_pin_1 = 1; + return 0; +} +early_param("disable_timer_pin_1", disable_timer_pin_setup); + +int timer_through_8259 __initdata; + /* * This code may look a bit paranoid, but it's supposed to cooperate with * a wide range of boards and BIOS bugs. Fortunately only the timer IRQ * is so screwy. Thanks to Brian Perkins for testing/hacking this beast * fanatically on his truly buggy board. * - * FIXME: really need to revamp this for modern platforms only. + * FIXME: really need to revamp this for all platforms. */ static inline void __init check_timer(void) { - struct irq_cfg *cfg = irq_cfg + 0; + struct irq_cfg *cfg = irq_cfg(0); int apic1, pin1, apic2, pin2; unsigned long flags; + unsigned int ver; int no_pin1 = 0; local_irq_save(flags); + ver = apic_read(APIC_LVR); + ver = GET_APIC_VERSION(ver); + /* * get/set the timer IRQ vector: */ @@ -1690,10 +2756,18 @@ static inline void __init check_timer(void) /* * As IRQ0 is to be enabled in the 8259A, the virtual - * wire has to be disabled in the local APIC. + * wire has to be disabled in the local APIC. Also + * timer interrupts need to be acknowledged manually in + * the 8259A for the i82489DX when using the NMI + * watchdog as that APIC treats NMIs as level-triggered. + * The AEOI mode will finish them in the 8259A + * automatically. */ apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT); init_8259A(1); +#ifdef CONFIG_X86_32 + timer_ack = (nmi_watchdog == NMI_IO_APIC && !APIC_INTEGRATED(ver)); +#endif pin1 = find_isa_irq_pin(0, mp_INT); apic1 = find_isa_irq_apic(0, mp_INT); @@ -1712,6 +2786,10 @@ static inline void __init check_timer(void) * 8259A. */ if (pin1 == -1) { +#ifdef CONFIG_INTR_REMAP + if (intr_remapping_enabled) + panic("BIOS bug: timer not connected to IO-APIC"); +#endif pin1 = pin2; apic1 = apic2; no_pin1 = 1; @@ -1729,7 +2807,7 @@ static inline void __init check_timer(void) setup_timer_IRQ0_pin(apic1, pin1, cfg->vector); } unmask_IO_APIC_irq(0); - if (!no_timer_check && timer_irq_works()) { + if (timer_irq_works()) { if (nmi_watchdog == NMI_IO_APIC) { setup_nmi(); enable_8259A_irq(0); @@ -1738,6 +2816,10 @@ static inline void __init check_timer(void) clear_IO_APIC_pin(0, pin1); goto out; } +#ifdef CONFIG_INTR_REMAP + if (intr_remapping_enabled) + panic("timer doesn't work through Interrupt-remapped IO-APIC"); +#endif clear_IO_APIC_pin(apic1, pin1); if (!no_pin1) apic_printk(APIC_QUIET, KERN_ERR "..MP-BIOS bug: " @@ -1777,6 +2859,9 @@ static inline void __init check_timer(void) "through the IO-APIC - disabling NMI Watchdog!\n"); nmi_watchdog = NMI_NONE; } +#ifdef CONFIG_X86_32 + timer_ack = 0; +#endif apic_printk(APIC_QUIET, KERN_INFO "...trying to set up timer as Virtual Wire IRQ...\n"); @@ -1813,13 +2898,6 @@ out: local_irq_restore(flags); } -static int __init notimercheck(char *s) -{ - no_timer_check = 1; - return 1; -} -__setup("no_timer_check", notimercheck); - /* * Traditionally ISA IRQ2 is the cascade IRQ, and is not available * to devices. However there may be an I/O APIC pin available for @@ -1837,27 +2915,49 @@ __setup("no_timer_check", notimercheck); * the I/O APIC in all cases now. No actual device should request * it anyway. --macro */ -#define PIC_IRQS (1<<2) +#define PIC_IRQS (1 << PIC_CASCADE_IR) void __init setup_IO_APIC(void) { +#ifdef CONFIG_X86_32 + enable_IO_APIC(); +#else /* * calling enable_IO_APIC() is moved to setup_local_APIC for BP */ +#endif io_apic_irqs = ~PIC_IRQS; 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 sync_Arb_IDs(); setup_IO_APIC_irqs(); init_IO_APIC_traps(); check_timer(); - if (!acpi_ioapic) - print_IO_APIC(); } +/* + * Called after all the initialization is done. If we didnt find any + * APIC bugs then we can allow the modify fast path + */ + +static int __init io_apic_bug_finalize(void) +{ + if (sis_apic_bug == -1) + sis_apic_bug = 0; + return 0; +} + +late_initcall(io_apic_bug_finalize); + struct sysfs_ioapic_data { struct sys_device dev; struct IO_APIC_route_entry entry[0]; @@ -1945,38 +3045,62 @@ device_initcall(ioapic_init_sysfs); /* * Dynamic irq allocate and deallocation */ -int create_irq(void) +unsigned int create_irq_nr(unsigned int irq_want) { /* Allocate an unused irq */ - int irq; - int new; + unsigned int irq; + unsigned int new; unsigned long flags; + struct irq_cfg *cfg_new; + +#ifndef CONFIG_HAVE_SPARSE_IRQ + irq_want = nr_irqs - 1; +#endif - irq = -ENOSPC; + irq = 0; spin_lock_irqsave(&vector_lock, flags); - for (new = (NR_IRQS - 1); new >= 0; new--) { + for (new = irq_want; new > 0; new--) { if (platform_legacy_irq(new)) continue; - if (irq_cfg[new].vector != 0) + cfg_new = irq_cfg(new); + if (cfg_new && cfg_new->vector != 0) continue; + /* check if need to create one */ + if (!cfg_new) + cfg_new = irq_cfg_alloc(new); if (__assign_irq_vector(new, TARGET_CPUS) == 0) irq = new; break; } spin_unlock_irqrestore(&vector_lock, flags); - if (irq >= 0) { + if (irq > 0) { dynamic_irq_init(irq); } return irq; } +int create_irq(void) +{ + int irq; + + irq = create_irq_nr(nr_irqs - 1); + + if (irq == 0) + irq = -1; + + return irq; +} + void destroy_irq(unsigned int irq) { unsigned long flags; dynamic_irq_cleanup(irq); +#ifdef CONFIG_INTR_REMAP + free_irte(irq); +#endif spin_lock_irqsave(&vector_lock, flags); __clear_irq_vector(irq); spin_unlock_irqrestore(&vector_lock, flags); @@ -1988,17 +3112,49 @@ void destroy_irq(unsigned int irq) #ifdef CONFIG_PCI_MSI static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_msg *msg) { - struct irq_cfg *cfg = irq_cfg + irq; + struct irq_cfg *cfg; int err; unsigned dest; cpumask_t tmp; tmp = TARGET_CPUS; err = assign_irq_vector(irq, tmp); - if (!err) { - cpus_and(tmp, cfg->domain, tmp); - dest = cpu_mask_to_apicid(tmp); + if (err) + return err; + + cfg = irq_cfg(irq); + cpus_and(tmp, cfg->domain, tmp); + dest = cpu_mask_to_apicid(tmp); + +#ifdef CONFIG_INTR_REMAP + if (irq_remapped(irq)) { + struct irte irte; + int ir_index; + u16 sub_handle; + + ir_index = map_irq_to_irte_handle(irq, &sub_handle); + BUG_ON(ir_index == -1); + + memset (&irte, 0, sizeof(irte)); + + irte.present = 1; + irte.dst_mode = INT_DEST_MODE; + irte.trigger_mode = 0; /* edge */ + irte.dlvry_mode = INT_DELIVERY_MODE; + irte.vector = cfg->vector; + irte.dest_id = IRTE_DEST(dest); + modify_irte(irq, &irte); + + msg->address_hi = MSI_ADDR_BASE_HI; + msg->data = sub_handle; + msg->address_lo = MSI_ADDR_BASE_LO | MSI_ADDR_IR_EXT_INT | + MSI_ADDR_IR_SHV | + MSI_ADDR_IR_INDEX1(ir_index) | + MSI_ADDR_IR_INDEX2(ir_index); + } else +#endif + { msg->address_hi = MSI_ADDR_BASE_HI; msg->address_lo = MSI_ADDR_BASE_LO | @@ -2024,10 +3180,11 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms #ifdef CONFIG_SMP static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) { - struct irq_cfg *cfg = irq_cfg + irq; + struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; cpumask_t tmp; + struct irq_desc *desc; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) @@ -2036,6 +3193,7 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) if (assign_irq_vector(irq, mask)) return; + cfg = irq_cfg(irq); cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); @@ -2047,8 +3205,61 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); write_msi_msg(irq, &msg); - irq_desc[irq].affinity = mask; + desc = irq_to_desc(irq); + desc->affinity = mask; +} + +#ifdef CONFIG_INTR_REMAP +/* + * Migrate the MSI irq to another cpumask. This migration is + * done in the process context using interrupt-remapping hardware. + */ +static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) +{ + struct irq_cfg *cfg; + unsigned int dest; + cpumask_t tmp, cleanup_mask; + struct irte irte; + struct irq_desc *desc; + + cpus_and(tmp, mask, cpu_online_map); + if (cpus_empty(tmp)) + return; + + if (get_irte(irq, &irte)) + return; + + if (assign_irq_vector(irq, mask)) + return; + + cfg = irq_cfg(irq); + cpus_and(tmp, cfg->domain, mask); + dest = cpu_mask_to_apicid(tmp); + + irte.vector = cfg->vector; + irte.dest_id = IRTE_DEST(dest); + + /* + * atomically update the IRTE with the new destination and vector. + */ + modify_irte(irq, &irte); + + /* + * After this point, all the interrupts will start arriving + * at the new destination. So, time to cleanup the previous + * vector allocation. + */ + if (cfg->move_in_progress) { + cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); + cfg->move_cleanup_count = cpus_weight(cleanup_mask); + send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); + cfg->move_in_progress = 0; + } + + desc = irq_to_desc(irq); + desc->affinity = mask; } +#endif #endif /* CONFIG_SMP */ /* @@ -2066,26 +3277,177 @@ static struct irq_chip msi_chip = { .retrigger = ioapic_retrigger_irq, }; -int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc) +#ifdef CONFIG_INTR_REMAP +static struct irq_chip msi_ir_chip = { + .name = "IR-PCI-MSI", + .unmask = unmask_msi_irq, + .mask = mask_msi_irq, + .ack = ack_x2apic_edge, +#ifdef CONFIG_SMP + .set_affinity = ir_set_msi_irq_affinity, +#endif + .retrigger = ioapic_retrigger_irq, +}; + +/* + * Map the PCI dev to the corresponding remapping hardware unit + * and allocate 'nvec' consecutive interrupt-remapping table entries + * in it. + */ +static int msi_alloc_irte(struct pci_dev *dev, int irq, int nvec) +{ + struct intel_iommu *iommu; + int index; + + iommu = map_dev_to_ir(dev); + if (!iommu) { + printk(KERN_ERR + "Unable to map PCI %s to iommu\n", pci_name(dev)); + return -ENOENT; + } + + index = alloc_irte(iommu, irq, nvec); + if (index < 0) { + printk(KERN_ERR + "Unable to allocate %d IRTE for PCI %s\n", nvec, + pci_name(dev)); + return -ENOSPC; + } + return index; +} +#endif + +static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc, int irq) { + int ret; struct msi_msg msg; - int irq, ret; - irq = create_irq(); - if (irq < 0) - return irq; ret = msi_compose_msg(dev, irq, &msg); + if (ret < 0) + return ret; + + set_irq_msi(irq, desc); + write_msi_msg(irq, &msg); + +#ifdef CONFIG_INTR_REMAP + if (irq_remapped(irq)) { + struct irq_desc *desc = irq_to_desc(irq); + /* + * irq migration in process context + */ + desc->status |= IRQ_MOVE_PCNTXT; + set_irq_chip_and_handler_name(irq, &msi_ir_chip, handle_edge_irq, "edge"); + } else +#endif + set_irq_chip_and_handler_name(irq, &msi_chip, handle_edge_irq, "edge"); + + return 0; +} + +static unsigned int build_irq_for_pci_dev(struct pci_dev *dev) +{ + unsigned int irq; + + irq = dev->bus->number; + irq <<= 8; + irq |= dev->devfn; + irq <<= 12; + + return irq; +} + +int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc) +{ + unsigned int irq; + int ret; + unsigned int irq_want; + + irq_want = build_irq_for_pci_dev(dev) + 0x100; + + irq = create_irq_nr(irq_want); + if (irq == 0) + return -1; + +#ifdef CONFIG_INTR_REMAP + if (!intr_remapping_enabled) + goto no_ir; + + ret = msi_alloc_irte(dev, irq, 1); + if (ret < 0) + goto error; +no_ir: +#endif + ret = setup_msi_irq(dev, desc, irq); if (ret < 0) { destroy_irq(irq); return ret; } + return 0; - set_irq_msi(irq, desc); - write_msi_msg(irq, &msg); +#ifdef CONFIG_INTR_REMAP +error: + destroy_irq(irq); + return ret; +#endif +} + +int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) +{ + unsigned int irq; + int ret, sub_handle; + struct msi_desc *desc; + unsigned int irq_want; - set_irq_chip_and_handler_name(irq, &msi_chip, handle_edge_irq, "edge"); +#ifdef CONFIG_INTR_REMAP + struct intel_iommu *iommu = 0; + int index = 0; +#endif + irq_want = build_irq_for_pci_dev(dev) + 0x100; + sub_handle = 0; + list_for_each_entry(desc, &dev->msi_list, list) { + irq = create_irq_nr(irq_want--); + if (irq == 0) + return -1; +#ifdef CONFIG_INTR_REMAP + if (!intr_remapping_enabled) + goto no_ir; + + if (!sub_handle) { + /* + * allocate the consecutive block of IRTE's + * for 'nvec' + */ + index = msi_alloc_irte(dev, irq, nvec); + if (index < 0) { + ret = index; + goto error; + } + } else { + iommu = map_dev_to_ir(dev); + if (!iommu) { + ret = -ENOENT; + goto error; + } + /* + * setup the mapping between the irq and the IRTE + * base index, the sub_handle pointing to the + * appropriate interrupt remap table entry. + */ + set_irte_irq(irq, iommu, index, sub_handle); + } +no_ir: +#endif + ret = setup_msi_irq(dev, desc, irq); + if (ret < 0) + goto error; + sub_handle++; + } return 0; + +error: + destroy_irq(irq); + return ret; } void arch_teardown_msi_irq(unsigned int irq) @@ -2097,10 +3459,11 @@ void arch_teardown_msi_irq(unsigned int irq) #ifdef CONFIG_SMP static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) { - struct irq_cfg *cfg = irq_cfg + irq; + struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; cpumask_t tmp; + struct irq_desc *desc; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) @@ -2109,6 +3472,7 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) if (assign_irq_vector(irq, mask)) return; + cfg = irq_cfg(irq); cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); @@ -2120,7 +3484,8 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); dmar_msi_write(irq, &msg); - irq_desc[irq].affinity = mask; + desc = irq_to_desc(irq); + desc->affinity = mask; } #endif /* CONFIG_SMP */ @@ -2174,9 +3539,10 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector) static void set_ht_irq_affinity(unsigned int irq, cpumask_t mask) { - struct irq_cfg *cfg = irq_cfg + irq; + struct irq_cfg *cfg; unsigned int dest; cpumask_t tmp; + struct irq_desc *desc; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) @@ -2185,11 +3551,13 @@ static void set_ht_irq_affinity(unsigned int irq, cpumask_t mask) if (assign_irq_vector(irq, mask)) return; + cfg = irq_cfg(irq); cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); target_ht_irq(irq, dest, cfg->vector); - irq_desc[irq].affinity = mask; + desc = irq_to_desc(irq); + desc->affinity = mask; } #endif @@ -2206,7 +3574,7 @@ static struct irq_chip ht_irq_chip = { int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) { - struct irq_cfg *cfg = irq_cfg + irq; + struct irq_cfg *cfg; int err; cpumask_t tmp; @@ -2216,6 +3584,7 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) struct ht_irq_msg msg; unsigned dest; + cfg = irq_cfg(irq); cpus_and(tmp, cfg->domain, tmp); dest = cpu_mask_to_apicid(tmp); @@ -2249,7 +3618,93 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) #ifdef CONFIG_ACPI -#define IO_APIC_MAX_ID 0xFE +#ifdef CONFIG_X86_32 +int __init io_apic_get_unique_id(int ioapic, int apic_id) +{ + union IO_APIC_reg_00 reg_00; + static physid_mask_t apic_id_map = PHYSID_MASK_NONE; + physid_mask_t tmp; + unsigned long flags; + int i = 0; + + /* + * The P4 platform supports up to 256 APIC IDs on two separate APIC + * buses (one for LAPICs, one for IOAPICs), where predecessors only + * supports up to 16 on one shared APIC bus. + * + * TBD: Expand LAPIC/IOAPIC support on P4-class systems to take full + * advantage of new APIC bus architecture. + */ + + if (physids_empty(apic_id_map)) + apic_id_map = ioapic_phys_id_map(phys_cpu_present_map); + + spin_lock_irqsave(&ioapic_lock, flags); + reg_00.raw = io_apic_read(ioapic, 0); + spin_unlock_irqrestore(&ioapic_lock, flags); + + if (apic_id >= get_physical_broadcast()) { + printk(KERN_WARNING "IOAPIC[%d]: Invalid apic_id %d, trying " + "%d\n", ioapic, apic_id, reg_00.bits.ID); + apic_id = reg_00.bits.ID; + } + + /* + * Every APIC in a system must have a unique ID or we get lots of nice + * 'stuck on smp_invalidate_needed IPI wait' messages. + */ + if (check_apicid_used(apic_id_map, apic_id)) { + + for (i = 0; i < get_physical_broadcast(); i++) { + if (!check_apicid_used(apic_id_map, i)) + break; + } + + if (i == get_physical_broadcast()) + panic("Max apic_id exceeded!\n"); + + printk(KERN_WARNING "IOAPIC[%d]: apic_id %d already used, " + "trying %d\n", ioapic, apic_id, i); + + apic_id = i; + } + + tmp = apicid_to_cpu_present(apic_id); + physids_or(apic_id_map, apic_id_map, tmp); + + if (reg_00.bits.ID != apic_id) { + reg_00.bits.ID = apic_id; + + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(ioapic, 0, reg_00.raw); + reg_00.raw = io_apic_read(ioapic, 0); + spin_unlock_irqrestore(&ioapic_lock, flags); + + /* Sanity check */ + if (reg_00.bits.ID != apic_id) { + printk("IOAPIC[%d]: Unable to change apic_id!\n", ioapic); + return -1; + } + } + + apic_printk(APIC_VERBOSE, KERN_INFO + "IOAPIC[%d]: Assigned apic_id %d\n", ioapic, apic_id); + + return apic_id; +} + +int __init io_apic_get_version(int ioapic) +{ + union IO_APIC_reg_01 reg_01; + unsigned long flags; + + spin_lock_irqsave(&ioapic_lock, flags); + reg_01.raw = io_apic_read(ioapic, 1); + spin_unlock_irqrestore(&ioapic_lock, flags); + + return reg_01.bits.version; +} +#endif int __init io_apic_get_redir_entries (int ioapic) { @@ -2314,6 +3769,7 @@ int acpi_get_override_irq(int bus_irq, int *trigger, int *polarity) void __init setup_ioapic_dest(void) { int pin, ioapic, irq, irq_entry; + struct irq_cfg *cfg; if (skip_ioapic_setup == 1) return; @@ -2329,10 +3785,15 @@ void __init setup_ioapic_dest(void) * when you have too many devices, because at that time only boot * cpu is online. */ - if (!irq_cfg[irq].vector) + cfg = irq_cfg(irq); + if (!cfg->vector) setup_IO_APIC_irq(ioapic, pin, irq, irq_trigger(irq_entry), irq_polarity(irq_entry)); +#ifdef CONFIG_INTR_REMAP + else if (intr_remapping_enabled) + set_ir_ioapic_affinity_irq(irq, TARGET_CPUS); +#endif else set_ioapic_affinity_irq(irq, TARGET_CPUS); } @@ -2341,6 +3802,7 @@ void __init setup_ioapic_dest(void) } #endif +#ifdef CONFIG_X86_64 #define IOAPIC_RESOURCE_NAME_SIZE 11 static struct resource *ioapic_resources; @@ -2376,36 +3838,56 @@ static struct resource * __init ioapic_setup_resources(void) return res; } +#endif void __init ioapic_init_mappings(void) { unsigned long ioapic_phys, idx = FIX_IO_APIC_BASE_0; - struct resource *ioapic_res; int i; +#ifdef CONFIG_X86_64 + struct resource *ioapic_res; ioapic_res = ioapic_setup_resources(); +#endif for (i = 0; i < nr_ioapics; i++) { if (smp_found_config) { ioapic_phys = mp_ioapics[i].mp_apicaddr; +#ifdef CONFIG_X86_32 + if (!ioapic_phys) { + printk(KERN_ERR + "WARNING: bogus zero IO-APIC " + "address found in MPTABLE, " + "disabling IO/APIC support!\n"); + smp_found_config = 0; + skip_ioapic_setup = 1; + goto fake_ioapic_page; + } +#endif } else { +#ifdef CONFIG_X86_32 +fake_ioapic_page: +#endif ioapic_phys = (unsigned long) alloc_bootmem_pages(PAGE_SIZE); ioapic_phys = __pa(ioapic_phys); } set_fixmap_nocache(idx, ioapic_phys); apic_printk(APIC_VERBOSE, - "mapped IOAPIC to %016lx (%016lx)\n", + "mapped IOAPIC to %08lx (%08lx)\n", __fix_to_virt(idx), ioapic_phys); idx++; +#ifdef CONFIG_X86_64 if (ioapic_res != NULL) { ioapic_res->start = ioapic_phys; ioapic_res->end = ioapic_phys + (4 * 1024) - 1; ioapic_res++; } +#endif } } +#ifdef CONFIG_X86_64 static int __init ioapic_insert_resources(void) { int i; @@ -2428,4 +3910,4 @@ static int __init ioapic_insert_resources(void) /* Insert the IO APIC resources after PCI initialization has occured to handle * IO APICS that are mapped in on a BAR in PCI space. */ late_initcall(ioapic_insert_resources); - +#endif