Merge branch 'x86/uv' into x86/core

13 files changed, 452 insertions, 1 deletions

diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h
index 4ef949c1972..394d177d721 100644
--- a/arch/x86/include/asm/apic.h
+++ b/arch/x86/include/asm/apic.h
@@ -379,6 +379,7 @@ static inline u32 safe_apic_wait_icr_idle(void)
 
 static inline void ack_APIC_irq(void)
 {
+#ifdef CONFIG_X86_LOCAL_APIC
         /*
          * ack_APIC_irq() actually gets compiled as a single instruction
          * ... yummie.
@@ -386,6 +387,7 @@ static inline void ack_APIC_irq(void)
 
         /* Docs say use 0 for future compatibility */
         apic_write(APIC_EOI, 0);
+#endif
 }
 
 static inline unsigned default_get_apic_id(unsigned long x)
diff --git a/arch/x86/include/asm/entry_arch.h b/arch/x86/include/asm/entry_arch.h
index 854d538ae85..c2e6bedaf25 100644
--- a/arch/x86/include/asm/entry_arch.h
+++ b/arch/x86/include/asm/entry_arch.h
@@ -33,6 +33,8 @@ BUILD_INTERRUPT3(invalidate_interrupt7,INVALIDATE_TLB_VECTOR_START+7,
                  smp_invalidate_interrupt)
 #endif
 
+BUILD_INTERRUPT(generic_interrupt, GENERIC_INTERRUPT_VECTOR)
+
 /*
  * every pentium local APIC has two 'local interrupts', with a
  * soft-definable vector attached to both interrupts, one of
diff --git a/arch/x86/include/asm/hardirq.h b/arch/x86/include/asm/hardirq.h
index 176f058e715..039db6aa8e0 100644
--- a/arch/x86/include/asm/hardirq.h
+++ b/arch/x86/include/asm/hardirq.h
@@ -12,6 +12,7 @@ typedef struct {
         unsigned int apic_timer_irqs;   /* arch dependent */
         unsigned int irq_spurious_count;
 #endif
+        unsigned int generic_irqs;      /* arch dependent */
 #ifdef CONFIG_SMP
         unsigned int irq_resched_count;
         unsigned int irq_call_count;
diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h
index 370e1c83bb4..b762ea49bd7 100644
--- a/arch/x86/include/asm/hw_irq.h
+++ b/arch/x86/include/asm/hw_irq.h
@@ -27,6 +27,7 @@
 
 /* Interrupt handlers registered during init_IRQ */
 extern void apic_timer_interrupt(void);
+extern void generic_interrupt(void);
 extern void error_interrupt(void);
 extern void spurious_interrupt(void);
 extern void thermal_interrupt(void);
diff --git a/arch/x86/include/asm/irq.h b/arch/x86/include/asm/irq.h
index 107eb219669..f38481bcd45 100644
--- a/arch/x86/include/asm/irq.h
+++ b/arch/x86/include/asm/irq.h
@@ -36,6 +36,7 @@ static inline int irq_canonicalize(int irq)
 extern void fixup_irqs(void);
 #endif
 
+extern void (*generic_interrupt_extension)(void);
 extern void init_IRQ(void);
 extern void native_init_IRQ(void);
 extern bool handle_irq(unsigned irq, struct pt_regs *regs);
diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h
index 8a285f356f8..3cbd79bbb47 100644
--- a/arch/x86/include/asm/irq_vectors.h
+++ b/arch/x86/include/asm/irq_vectors.h
@@ -112,6 +112,11 @@
 #define LOCAL_PERF_VECTOR               0xee
 
 /*
+ * Generic system vector for platform specific use
+ */
+#define GENERIC_INTERRUPT_VECTOR        0xed
+
+/*
  * First APIC vector available to drivers: (vectors 0x30-0xee) we
  * start at 0x31(0x41) to spread out vectors evenly between priority
  * levels. (0x80 is the syscall vector)
diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h
index 777327ef05c..9f4dfba33b2 100644
--- a/arch/x86/include/asm/uv/uv_hub.h
+++ b/arch/x86/include/asm/uv/uv_hub.h
@@ -199,6 +199,10 @@ DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info);
 #define SCIR_CPU_ACTIVITY       0x02    /* not idle */
 #define SCIR_CPU_HB_INTERVAL    (HZ)    /* once per second */
 
+/* Loop through all installed blades */
+#define for_each_possible_blade(bid)            \
+        for ((bid) = 0; (bid) < uv_num_possible_blades(); (bid)++)
+
 /*
  * Macros for converting between kernel virtual addresses, socket local physical
  * addresses, and UV global physical addresses.
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index 95f216bbfaf..339ce35648e 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -111,7 +111,7 @@ obj-$(CONFIG_SWIOTLB)			+= pci-swiotlb_64.o # NB rename without _64
 ###
 # 64 bit specific files
 ifeq ($(CONFIG_X86_64),y)
-        obj-$(CONFIG_X86_UV)            += tlb_uv.o bios_uv.o uv_irq.o uv_sysfs.o
+        obj-$(CONFIG_X86_UV)            += tlb_uv.o bios_uv.o uv_irq.o uv_sysfs.o uv_time.o
         obj-$(CONFIG_X86_PM_TIMER)      += pmtimer_64.o
         obj-$(CONFIG_AUDIT)             += audit_64.o
 
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index 83d1836b946..7ba4621c0df 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -984,6 +984,8 @@ apicinterrupt UV_BAU_MESSAGE \
 #endif
 apicinterrupt LOCAL_TIMER_VECTOR \
         apic_timer_interrupt smp_apic_timer_interrupt
+apicinterrupt GENERIC_INTERRUPT_VECTOR \
+        generic_interrupt smp_generic_interrupt
 
 #ifdef CONFIG_SMP
 apicinterrupt INVALIDATE_TLB_VECTOR_START+0 \
diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
index f13ca1650aa..b864341dcc4 100644
--- a/arch/x86/kernel/irq.c
+++ b/arch/x86/kernel/irq.c
@@ -15,6 +15,9 @@
 
 atomic_t irq_err_count;
 
+/* Function pointer for generic interrupt vector handling */
+void (*generic_interrupt_extension)(void) = NULL;
+
 /*
  * 'what should we do if we get a hw irq event on an illegal vector'.
  * each architecture has to answer this themselves.
@@ -56,6 +59,12 @@ static int show_other_interrupts(struct seq_file *p)
                 seq_printf(p, "%10u ", irq_stats(j)->apic_timer_irqs);
         seq_printf(p, "  Local timer interrupts\n");
 #endif
+        if (generic_interrupt_extension) {
+                seq_printf(p, "PLT: ");
+                for_each_online_cpu(j)
+                        seq_printf(p, "%10u ", irq_stats(j)->generic_irqs);
+                seq_printf(p, "  Platform interrupts\n");
+        }
 #ifdef CONFIG_SMP
         seq_printf(p, "RES: ");
         for_each_online_cpu(j)
@@ -163,6 +172,8 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
 #ifdef CONFIG_X86_LOCAL_APIC
         sum += irq_stats(cpu)->apic_timer_irqs;
 #endif
+        if (generic_interrupt_extension)
+                sum += irq_stats(cpu)->generic_irqs;
 #ifdef CONFIG_SMP
         sum += irq_stats(cpu)->irq_resched_count;
         sum += irq_stats(cpu)->irq_call_count;
@@ -226,4 +237,27 @@ unsigned int __irq_entry do_IRQ(struct pt_regs *regs)
         return 1;
 }
 
+/*
+ * Handler for GENERIC_INTERRUPT_VECTOR.
+ */
+void smp_generic_interrupt(struct pt_regs *regs)
+{
+        struct pt_regs *old_regs = set_irq_regs(regs);
+
+        ack_APIC_irq();
+
+        exit_idle();
+
+        irq_enter();
+
+        inc_irq_stat(generic_irqs);
+
+        if (generic_interrupt_extension)
+                generic_interrupt_extension();
+
+        irq_exit();
+
+        set_irq_regs(old_regs);
+}
+
 EXPORT_SYMBOL_GPL(vector_used_by_percpu_irq);
diff --git a/arch/x86/kernel/irqinit_32.c b/arch/x86/kernel/irqinit_32.c
index 50b8c3a3006..bc132610544 100644
--- a/arch/x86/kernel/irqinit_32.c
+++ b/arch/x86/kernel/irqinit_32.c
@@ -175,6 +175,9 @@ void __init native_init_IRQ(void)
         /* self generated IPI for local APIC timer */
         alloc_intr_gate(LOCAL_TIMER_VECTOR, apic_timer_interrupt);
 
+        /* generic IPI for platform specific use */
+        alloc_intr_gate(GENERIC_INTERRUPT_VECTOR, generic_interrupt);
+
         /* IPI vectors for APIC spurious and error interrupts */
         alloc_intr_gate(SPURIOUS_APIC_VECTOR, spurious_interrupt);
         alloc_intr_gate(ERROR_APIC_VECTOR, error_interrupt);
diff --git a/arch/x86/kernel/irqinit_64.c b/arch/x86/kernel/irqinit_64.c
index da481a1e3f3..c7a49e0ffbf 100644
--- a/arch/x86/kernel/irqinit_64.c
+++ b/arch/x86/kernel/irqinit_64.c
@@ -147,6 +147,9 @@ static void __init apic_intr_init(void)
         /* self generated IPI for local APIC timer */
         alloc_intr_gate(LOCAL_TIMER_VECTOR, apic_timer_interrupt);
 
+        /* generic IPI for platform specific use */
+        alloc_intr_gate(GENERIC_INTERRUPT_VECTOR, generic_interrupt);
+
         /* IPI vectors for APIC spurious and error interrupts */
         alloc_intr_gate(SPURIOUS_APIC_VECTOR, spurious_interrupt);
         alloc_intr_gate(ERROR_APIC_VECTOR, error_interrupt);
diff --git a/arch/x86/kernel/uv_time.c b/arch/x86/kernel/uv_time.c
new file mode 100644
index 00000000000..2ffb6c53326
--- /dev/null
+++ b/arch/x86/kernel/uv_time.c
@@ -0,0 +1,393 @@
+/*
+ * SGI RTC clock/timer routines.
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ *  Copyright (c) 2009 Silicon Graphics, Inc.  All Rights Reserved.
+ *  Copyright (c) Dimitri Sivanich
+ */
+#include <linux/clockchips.h>
+
+#include <asm/uv/uv_mmrs.h>
+#include <asm/uv/uv_hub.h>
+#include <asm/uv/bios.h>
+#include <asm/uv/uv.h>
+#include <asm/apic.h>
+#include <asm/cpu.h>
+
+#define RTC_NAME                "sgi_rtc"
+
+static cycle_t uv_read_rtc(void);
+static int uv_rtc_next_event(unsigned long, struct clock_event_device *);
+static void uv_rtc_timer_setup(enum clock_event_mode,
+                                struct clock_event_device *);
+
+static struct clocksource clocksource_uv = {
+        .name           = RTC_NAME,
+        .rating         = 400,
+        .read           = uv_read_rtc,
+        .mask           = (cycle_t)UVH_RTC_REAL_TIME_CLOCK_MASK,
+        .shift          = 10,
+        .flags          = CLOCK_SOURCE_IS_CONTINUOUS,
+};
+
+static struct clock_event_device clock_event_device_uv = {
+        .name           = RTC_NAME,
+        .features       = CLOCK_EVT_FEAT_ONESHOT,
+        .shift          = 20,
+        .rating         = 400,
+        .irq            = -1,
+        .set_next_event = uv_rtc_next_event,
+        .set_mode       = uv_rtc_timer_setup,
+        .event_handler  = NULL,
+};
+
+static DEFINE_PER_CPU(struct clock_event_device, cpu_ced);
+
+/* There is one of these allocated per node */
+struct uv_rtc_timer_head {
+        spinlock_t      lock;
+        /* next cpu waiting for timer, local node relative: */
+        int             next_cpu;
+        /* number of cpus on this node: */
+        int             ncpus;
+        struct {
+                int     lcpu;           /* systemwide logical cpu number */
+                u64     expires;        /* next timer expiration for this cpu */
+        } cpu[1];
+};
+
+/*
+ * Access to uv_rtc_timer_head via blade id.
+ */
+static struct uv_rtc_timer_head         **blade_info __read_mostly;
+
+static int                              uv_rtc_enable;
+
+/*
+ * Hardware interface routines
+ */
+
+/* Send IPIs to another node */
+static void uv_rtc_send_IPI(int cpu)
+{
+        unsigned long apicid, val;
+        int pnode;
+
+        apicid = cpu_physical_id(cpu);
+        pnode = uv_apicid_to_pnode(apicid);
+        val = (1UL << UVH_IPI_INT_SEND_SHFT) |
+              (apicid << UVH_IPI_INT_APIC_ID_SHFT) |
+              (GENERIC_INTERRUPT_VECTOR << UVH_IPI_INT_VECTOR_SHFT);
+
+        uv_write_global_mmr64(pnode, UVH_IPI_INT, val);
+}
+
+/* Check for an RTC interrupt pending */
+static int uv_intr_pending(int pnode)
+{
+        return uv_read_global_mmr64(pnode, UVH_EVENT_OCCURRED0) &
+                UVH_EVENT_OCCURRED0_RTC1_MASK;
+}
+
+/* Setup interrupt and return non-zero if early expiration occurred. */
+static int uv_setup_intr(int cpu, u64 expires)
+{
+        u64 val;
+        int pnode = uv_cpu_to_pnode(cpu);
+
+        uv_write_global_mmr64(pnode, UVH_RTC1_INT_CONFIG,
+                UVH_RTC1_INT_CONFIG_M_MASK);
+        uv_write_global_mmr64(pnode, UVH_INT_CMPB, -1L);
+
+        uv_write_global_mmr64(pnode, UVH_EVENT_OCCURRED0_ALIAS,
+                UVH_EVENT_OCCURRED0_RTC1_MASK);
+
+        val = (GENERIC_INTERRUPT_VECTOR << UVH_RTC1_INT_CONFIG_VECTOR_SHFT) |
+                ((u64)cpu_physical_id(cpu) << UVH_RTC1_INT_CONFIG_APIC_ID_SHFT);
+
+        /* Set configuration */
+        uv_write_global_mmr64(pnode, UVH_RTC1_INT_CONFIG, val);
+        /* Initialize comparator value */
+        uv_write_global_mmr64(pnode, UVH_INT_CMPB, expires);
+
+        return (expires < uv_read_rtc() && !uv_intr_pending(pnode));
+}
+
+/*
+ * Per-cpu timer tracking routines
+ */
+
+static __init void uv_rtc_deallocate_timers(void)
+{
+        int bid;
+
+        for_each_possible_blade(bid) {
+                kfree(blade_info[bid]);
+        }
+        kfree(blade_info);
+}
+
+/* Allocate per-node list of cpu timer expiration times. */
+static __init int uv_rtc_allocate_timers(void)
+{
+        int cpu;
+
+        blade_info = kmalloc(uv_possible_blades * sizeof(void *), GFP_KERNEL);
+        if (!blade_info)
+                return -ENOMEM;
+        memset(blade_info, 0, uv_possible_blades * sizeof(void *));
+
+        for_each_present_cpu(cpu) {
+                int nid = cpu_to_node(cpu);
+                int bid = uv_cpu_to_blade_id(cpu);
+                int bcpu = uv_cpu_hub_info(cpu)->blade_processor_id;
+                struct uv_rtc_timer_head *head = blade_info[bid];
+
+                if (!head) {
+                        head = kmalloc_node(sizeof(struct uv_rtc_timer_head) +
+                                (uv_blade_nr_possible_cpus(bid) *
+                                        2 * sizeof(u64)),
+                                GFP_KERNEL, nid);
+                        if (!head) {
+                                uv_rtc_deallocate_timers();
+                                return -ENOMEM;
+                        }
+                        spin_lock_init(&head->lock);
+                        head->ncpus = uv_blade_nr_possible_cpus(bid);
+                        head->next_cpu = -1;
+                        blade_info[bid] = head;
+                }
+
+                head->cpu[bcpu].lcpu = cpu;
+                head->cpu[bcpu].expires = ULLONG_MAX;
+        }
+
+        return 0;
+}
+
+/* Find and set the next expiring timer.  */
+static void uv_rtc_find_next_timer(struct uv_rtc_timer_head *head, int pnode)
+{
+        u64 lowest = ULLONG_MAX;
+        int c, bcpu = -1;
+
+        head->next_cpu = -1;
+        for (c = 0; c < head->ncpus; c++) {
+                u64 exp = head->cpu[c].expires;
+                if (exp < lowest) {
+                        bcpu = c;
+                        lowest = exp;
+                }
+        }
+        if (bcpu >= 0) {
+                head->next_cpu = bcpu;
+                c = head->cpu[bcpu].lcpu;
+                if (uv_setup_intr(c, lowest))
+                        /* If we didn't set it up in time, trigger */
+                        uv_rtc_send_IPI(c);
+        } else {
+                uv_write_global_mmr64(pnode, UVH_RTC1_INT_CONFIG,
+                        UVH_RTC1_INT_CONFIG_M_MASK);
+        }
+}
+
+/*
+ * Set expiration time for current cpu.
+ *
+ * Returns 1 if we missed the expiration time.
+ */
+static int uv_rtc_set_timer(int cpu, u64 expires)
+{
+        int pnode = uv_cpu_to_pnode(cpu);
+        int bid = uv_cpu_to_blade_id(cpu);
+        struct uv_rtc_timer_head *head = blade_info[bid];
+        int bcpu = uv_cpu_hub_info(cpu)->blade_processor_id;
+        u64 *t = &head->cpu[bcpu].expires;
+        unsigned long flags;
+        int next_cpu;
+
+        spin_lock_irqsave(&head->lock, flags);
+
+        next_cpu = head->next_cpu;
+        *t = expires;
+        /* Will this one be next to go off? */
+        if (next_cpu < 0 || bcpu == next_cpu ||
+                        expires < head->cpu[next_cpu].expires) {
+                head->next_cpu = bcpu;
+                if (uv_setup_intr(cpu, expires)) {
+                        *t = ULLONG_MAX;
+                        uv_rtc_find_next_timer(head, pnode);
+                        spin_unlock_irqrestore(&head->lock, flags);
+                        return 1;
+                }
+        }
+
+        spin_unlock_irqrestore(&head->lock, flags);
+        return 0;
+}
+
+/*
+ * Unset expiration time for current cpu.
+ *
+ * Returns 1 if this timer was pending.
+ */
+static int uv_rtc_unset_timer(int cpu)
+{
+        int pnode = uv_cpu_to_pnode(cpu);
+        int bid = uv_cpu_to_blade_id(cpu);
+        struct uv_rtc_timer_head *head = blade_info[bid];
+        int bcpu = uv_cpu_hub_info(cpu)->blade_processor_id;
+        u64 *t = &head->cpu[bcpu].expires;
+        unsigned long flags;
+        int rc = 0;
+
+        spin_lock_irqsave(&head->lock, flags);
+
+        if (head->next_cpu == bcpu && uv_read_rtc() >= *t)
+                rc = 1;
+
+        *t = ULLONG_MAX;
+
+        /* Was the hardware setup for this timer? */
+        if (head->next_cpu == bcpu)
+                uv_rtc_find_next_timer(head, pnode);
+
+        spin_unlock_irqrestore(&head->lock, flags);
+
+        return rc;
+}
+
+
+/*
+ * Kernel interface routines.
+ */
+
+/*
+ * Read the RTC.
+ */
+static cycle_t uv_read_rtc(void)
+{
+        return (cycle_t)uv_read_local_mmr(UVH_RTC);
+}
+
+/*
+ * Program the next event, relative to now
+ */
+static int uv_rtc_next_event(unsigned long delta,
+                             struct clock_event_device *ced)
+{
+        int ced_cpu = cpumask_first(ced->cpumask);
+
+        return uv_rtc_set_timer(ced_cpu, delta + uv_read_rtc());
+}
+
+/*
+ * Setup the RTC timer in oneshot mode
+ */
+static void uv_rtc_timer_setup(enum clock_event_mode mode,
+                               struct clock_event_device *evt)
+{
+        int ced_cpu = cpumask_first(evt->cpumask);
+
+        switch (mode) {
+        case CLOCK_EVT_MODE_PERIODIC:
+        case CLOCK_EVT_MODE_ONESHOT:
+        case CLOCK_EVT_MODE_RESUME:
+                /* Nothing to do here yet */
+                break;
+        case CLOCK_EVT_MODE_UNUSED:
+        case CLOCK_EVT_MODE_SHUTDOWN:
+                uv_rtc_unset_timer(ced_cpu);
+                break;
+        }
+}
+
+static void uv_rtc_interrupt(void)
+{
+        struct clock_event_device *ced = &__get_cpu_var(cpu_ced);
+        int cpu = smp_processor_id();
+
+        if (!ced || !ced->event_handler)
+                return;
+
+        if (uv_rtc_unset_timer(cpu) != 1)
+                return;
+
+        ced->event_handler(ced);
+}
+
+static int __init uv_enable_rtc(char *str)
+{
+        uv_rtc_enable = 1;
+
+        return 1;
+}
+__setup("uvrtc", uv_enable_rtc);
+
+static __init void uv_rtc_register_clockevents(struct work_struct *dummy)
+{
+        struct clock_event_device *ced = &__get_cpu_var(cpu_ced);
+
+        *ced = clock_event_device_uv;
+        ced->cpumask = cpumask_of(smp_processor_id());
+        clockevents_register_device(ced);
+}
+
+static __init int uv_rtc_setup_clock(void)
+{
+        int rc;
+
+        if (!uv_rtc_enable || !is_uv_system() || generic_interrupt_extension)
+                return -ENODEV;
+
+        generic_interrupt_extension = uv_rtc_interrupt;
+
+        clocksource_uv.mult = clocksource_hz2mult(sn_rtc_cycles_per_second,
+                                clocksource_uv.shift);
+
+        rc = clocksource_register(&clocksource_uv);
+        if (rc) {
+                generic_interrupt_extension = NULL;
+                return rc;
+        }
+
+        /* Setup and register clockevents */
+        rc = uv_rtc_allocate_timers();
+        if (rc) {
+                clocksource_unregister(&clocksource_uv);
+                generic_interrupt_extension = NULL;
+                return rc;
+        }
+
+        clock_event_device_uv.mult = div_sc(sn_rtc_cycles_per_second,
+                                NSEC_PER_SEC, clock_event_device_uv.shift);
+
+        clock_event_device_uv.min_delta_ns = NSEC_PER_SEC /
+                                                sn_rtc_cycles_per_second;
+
+        clock_event_device_uv.max_delta_ns = clocksource_uv.mask *
+                                (NSEC_PER_SEC / sn_rtc_cycles_per_second);
+
+        rc = schedule_on_each_cpu(uv_rtc_register_clockevents);
+        if (rc) {
+                clocksource_unregister(&clocksource_uv);
+                generic_interrupt_extension = NULL;
+                uv_rtc_deallocate_timers();
+        }
+
+        return rc;
+}
+arch_initcall(uv_rtc_setup_clock);