[SPARC64]: Virtualize IRQ numbers.

Inspired by PowerPC XICS interrupt support code.

All IRQs are virtualized in order to keep NR_IRQS from needing
to be too large.  Interrupts on sparc64 are arbitrary 11-bit
values, but we don't need to define NR_IRQS to 2048 if we
virtualize the IRQs.

As PCI and SBUS controller drivers build device IRQs, we divy
out virtual IRQ numbers incrementally starting at 1.  Zero is
a special virtual IRQ used for the timer interrupt.

So device drivers all see virtual IRQs, and all the normal
interfaces such as request_irq(), enable_irq(), etc. translate
that into a real IRQ number in order to configure the IRQ.

At this point knowledge of the struct ino_bucket is almost
entirely contained within arch/sparc64/kernel/irq.c  There are
a few small bits in the PCI controller drivers that need to
be swept away before we can remove ino_bucket's definition
out of asm-sparc64/irq.h and privately into kernel/irq.c

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 179 insertions, 84 deletions

diff --git a/arch/sparc64/kernel/irq.c b/arch/sparc64/kernel/irq.c
index f2668f2bed9c..49ad9cd0f8c8 100644
--- a/arch/sparc64/kernel/irq.c
+++ b/arch/sparc64/kernel/irq.c
@@ -70,7 +70,10 @@ struct ino_bucket ivector_table[NUM_IVECS] __attribute__ ((aligned (SMP_CACHE_BY
  */
 #define irq_work(__cpu) &(trap_block[(__cpu)].irq_worklist)
 
-static struct irqaction *irq_action[NR_IRQS];
+static struct irqaction timer_irq_action = {
+        .name = "timer",
+};
+static struct irqaction *irq_action[NR_IRQS] = { &timer_irq_action, };
 
 /* This only synchronizes entities which modify IRQ handler
  * state and some selected user-level spots that want to
@@ -79,6 +82,59 @@ static struct irqaction *irq_action[NR_IRQS];
  */
 static DEFINE_SPINLOCK(irq_action_lock);
 
+static unsigned int virt_to_real_irq_table[NR_IRQS];
+static unsigned char virt_irq_cur = 1;
+
+static unsigned char virt_irq_alloc(unsigned int real_irq)
+{
+        unsigned char ent;
+
+        BUILD_BUG_ON(NR_IRQS >= 256);
+
+        ent = virt_irq_cur;
+        if (ent >= NR_IRQS) {
+                printk(KERN_ERR "IRQ: Out of virtual IRQs.\n");
+                return 0;
+        }
+
+        virt_irq_cur = ent + 1;
+        virt_to_real_irq_table[ent] = real_irq;
+
+        return ent;
+}
+
+#if 0 /* Currently unused. */
+static unsigned char real_to_virt_irq(unsigned int real_irq)
+{
+        struct ino_bucket *bucket = __bucket(real_irq);
+
+        return bucket->virt_irq;
+}
+#endif
+
+static unsigned int virt_to_real_irq(unsigned char virt_irq)
+{
+        return virt_to_real_irq_table[virt_irq];
+}
+
+void irq_install_pre_handler(int virt_irq,
+                             void (*func)(struct ino_bucket *, void *, void *),
+                             void *arg1, void *arg2)
+{
+        unsigned int real_irq = virt_to_real_irq(virt_irq);
+        struct ino_bucket *bucket;
+        struct irq_desc *d;
+
+        if (unlikely(!real_irq))
+                return;
+
+        bucket = __bucket(real_irq);
+        d = bucket->irq_info;
+        d->pre_handler = func;
+        d->pre_handler_arg1 = arg1;
+        d->pre_handler_arg2 = arg2;
+}
+
 static void register_irq_proc (unsigned int irq);
 
 /*
@@ -164,14 +220,18 @@ static unsigned int sun4u_compute_tid(unsigned long imap, unsigned long cpuid)
         return tid;
 }
 
-/* Now these are always passed a true fully specified sun4u INO. */
-void enable_irq(unsigned int irq)
+void enable_irq(unsigned int virt_irq)
 {
-        struct ino_bucket *bucket = __bucket(irq);
+        unsigned int real_irq = virt_to_real_irq(virt_irq);
+        struct ino_bucket *bucket;
         unsigned long imap, cpuid;
 
+        if (unlikely(!real_irq))
+                return;
+
+        bucket = __bucket(real_irq);
         imap = bucket->imap;
-        if (imap == 0UL)
+        if (unlikely(imap == 0UL))
                 return;
 
         preempt_disable();
@@ -182,7 +242,7 @@ void enable_irq(unsigned int irq)
         cpuid = real_hard_smp_processor_id();
 
         if (tlb_type == hypervisor) {
-                unsigned int ino = __irq_ino(irq);
+                unsigned int ino = __irq_ino(real_irq);
                 int err;
 
                 err = sun4v_intr_settarget(ino, cpuid);
@@ -211,34 +271,39 @@ void enable_irq(unsigned int irq)
         preempt_enable();
 }
 
-/* This now gets passed true ino's as well. */
-void disable_irq(unsigned int irq)
+void disable_irq(unsigned int virt_irq)
 {
-        struct ino_bucket *bucket = __bucket(irq);
+        unsigned int real_irq = virt_to_real_irq(virt_irq);
+        struct ino_bucket *bucket;
         unsigned long imap;
 
+        if (unlikely(!real_irq))
+                return;
+
+        bucket = __bucket(real_irq);
         imap = bucket->imap;
-        if (imap != 0UL) {
-                if (tlb_type == hypervisor) {
-                        unsigned int ino = __irq_ino(irq);
-                        int err;
-
-                        err = sun4v_intr_setenabled(ino, HV_INTR_DISABLED);
-                        if (err != HV_EOK)
-                                printk("sun4v_intr_setenabled(%x): "
-                                       "err(%d)\n", ino, err);
-                } else {
-                        u32 tmp;
-
-                        /* NOTE: We do not want to futz with the IRQ clear registers
-                         *       and move the state to IDLE, the SCSI code does call
-                         *       disable_irq() to assure atomicity in the queue cmd
-                         *       SCSI adapter driver code.  Thus we'd lose interrupts.
-                         */
-                        tmp = upa_readl(imap);
-                        tmp &= ~IMAP_VALID;
-                        upa_writel(tmp, imap);
-                }
+        if (unlikely(imap == 0UL))
+                return;
+
+        if (tlb_type == hypervisor) {
+                unsigned int ino = __irq_ino(real_irq);
+                int err;
+
+                err = sun4v_intr_setenabled(ino, HV_INTR_DISABLED);
+                if (err != HV_EOK)
+                        printk("sun4v_intr_setenabled(%x): "
+                               "err(%d)\n", ino, err);
+        } else {
+                u32 tmp;
+
+                /* NOTE: We do not want to futz with the IRQ clear registers
+                 *       and move the state to IDLE, the SCSI code does call
+                 *       disable_irq() to assure atomicity in the queue cmd
+                 *       SCSI adapter driver code.  Thus we'd lose interrupts.
+                 */
+                tmp = upa_readl(imap);
+                tmp &= ~IMAP_VALID;
+                upa_writel(tmp, imap);
         }
 }
 
@@ -253,14 +318,14 @@ static void build_irq_error(const char *msg, unsigned int ino, int inofixup,
         prom_halt();
 }
 
-unsigned int build_irq(int inofixup, unsigned long iclr, unsigned long imap)
+unsigned int build_irq(int inofixup, unsigned long iclr, unsigned long imap, unsigned char flags)
 {
         struct ino_bucket *bucket;
         int ino;
 
         BUG_ON(tlb_type == hypervisor);
 
-        /* RULE: Both must be specified in all other cases. */
+        /* RULE: Both must be specified. */
         if (iclr == 0UL || imap == 0UL) {
                 prom_printf("Invalid build_irq %d %016lx %016lx\n",
                             inofixup, iclr, imap);
@@ -298,10 +363,12 @@ unsigned int build_irq(int inofixup, unsigned long iclr, unsigned long imap)
          */
         bucket->imap  = imap;
         bucket->iclr  = iclr;
-        bucket->flags = 0;
+        if (!bucket->virt_irq)
+                bucket->virt_irq = virt_irq_alloc(__irq(bucket));
+        bucket->flags = flags;
 
 out:
-        return __irq(bucket);
+        return bucket->virt_irq;
 }
 
 unsigned int sun4v_build_irq(u32 devhandle, unsigned int devino, unsigned char flags)
@@ -322,7 +389,8 @@ unsigned int sun4v_build_irq(u32 devhandle, unsigned int devino, unsigned char f
          */
         bucket->imap = ~0UL - sysino;
         bucket->iclr = ~0UL - sysino;
-
+        if (!bucket->virt_irq)
+                bucket->virt_irq = virt_irq_alloc(__irq(bucket));
         bucket->flags = flags;
 
         bucket->irq_info = kzalloc(sizeof(struct irq_desc), GFP_ATOMIC);
@@ -331,7 +399,7 @@ unsigned int sun4v_build_irq(u32 devhandle, unsigned int devino, unsigned char f
                 prom_halt();
         }
 
-        return __irq(bucket);
+        return bucket->virt_irq;
 }
 
 static void atomic_bucket_insert(struct ino_bucket *bucket)
@@ -390,37 +458,42 @@ static struct irqaction *get_action_slot(struct ino_bucket *bucket)
         return NULL;
 }
 
-int request_irq(unsigned int irq, irqreturn_t (*handler)(int, void *, struct pt_regs *),
+int request_irq(unsigned int virt_irq,
+                irqreturn_t (*handler)(int, void *, struct pt_regs *),
                 unsigned long irqflags, const char *name, void *dev_id)
 {
         struct irqaction *action;
-        struct ino_bucket *bucket = __bucket(irq);
+        struct ino_bucket *bucket;
         unsigned long flags;
+        unsigned int real_irq;
         int pending = 0;
 
+        real_irq = virt_to_real_irq(virt_irq);
+        if (unlikely(!real_irq))
+                return -EINVAL;
+
         if (unlikely(!handler))
                 return -EINVAL;
 
+        bucket = __bucket(real_irq);
         if (unlikely(!bucket->irq_info))
                 return -ENODEV;
 
         if (irqflags & SA_SAMPLE_RANDOM) {
                 /*
-                 * This function might sleep, we want to call it first,
-                 * outside of the atomic block. In SA_STATIC_ALLOC case,
-                 * random driver's kmalloc will fail, but it is safe.
-                 * If already initialized, random driver will not reinit.
-                 * Yes, this might clear the entropy pool if the wrong
-                 * driver is attempted to be loaded, without actually
-                 * installing a new handler, but is this really a problem,
-                 * only the sysadmin is able to do this.
-                 */
-                rand_initialize_irq(PIL_DEVICE_IRQ);
+                 * This function might sleep, we want to call it first,
+                 * outside of the atomic block.
+                 * Yes, this might clear the entropy pool if the wrong
+                 * driver is attempted to be loaded, without actually
+                 * installing a new handler, but is this really a problem,
+                 * only the sysadmin is able to do this.
+                 */
+                rand_initialize_irq(virt_irq);
         }
 
         spin_lock_irqsave(&irq_action_lock, flags);
 
-        if (check_irq_sharing(PIL_DEVICE_IRQ, irqflags)) {
+        if (check_irq_sharing(virt_irq, irqflags)) {
                 spin_unlock_irqrestore(&irq_action_lock, flags);
                 return -EBUSY;
         }
@@ -441,12 +514,12 @@ int request_irq(unsigned int irq, irqreturn_t (*handler)(int, void *, struct pt_
         action->name = name;
         action->next = NULL;
         action->dev_id = dev_id;
-        put_ino_in_irqaction(action, irq);
+        put_ino_in_irqaction(action, __irq_ino(real_irq));
         put_smpaff_in_irqaction(action, CPU_MASK_NONE);
 
-        append_irq_action(PIL_DEVICE_IRQ, action);
+        append_irq_action(virt_irq, action);
 
-        enable_irq(irq);
+        enable_irq(virt_irq);
 
         /* We ate the IVEC already, this makes sure it does not get lost. */
         if (pending) {
@@ -456,7 +529,7 @@ int request_irq(unsigned int irq, irqreturn_t (*handler)(int, void *, struct pt_
 
         spin_unlock_irqrestore(&irq_action_lock, flags);
 
-        register_irq_proc(__irq_ino(irq));
+        register_irq_proc(virt_irq);
 
 #ifdef CONFIG_SMP
         distribute_irqs();
@@ -466,17 +539,17 @@ int request_irq(unsigned int irq, irqreturn_t (*handler)(int, void *, struct pt_
 
 EXPORT_SYMBOL(request_irq);
 
-static struct irqaction *unlink_irq_action(unsigned int irq, void *dev_id)
+static struct irqaction *unlink_irq_action(unsigned int virt_irq, void *dev_id)
 {
         struct irqaction *action, **pp;
 
-        pp = irq_action + PIL_DEVICE_IRQ;
+        pp = irq_action + virt_irq;
         action = *pp;
         if (unlikely(!action))
                 return NULL;
 
         if (unlikely(!action->handler)) {
-                printk("Freeing free IRQ %d\n", PIL_DEVICE_IRQ);
+                printk("Freeing free IRQ %d\n", virt_irq);
                 return NULL;
         }
 
@@ -491,28 +564,33 @@ static struct irqaction *unlink_irq_action(unsigned int irq, void *dev_id)
         return action;
 }
 
-void free_irq(unsigned int irq, void *dev_id)
+void free_irq(unsigned int virt_irq, void *dev_id)
 {
         struct irqaction *action;
         struct ino_bucket *bucket;
         struct irq_desc *desc;
         unsigned long flags;
+        unsigned int real_irq;
         int ent, i;
 
+        real_irq = virt_to_real_irq(virt_irq);
+        if (unlikely(!real_irq))
+                return;
+
         spin_lock_irqsave(&irq_action_lock, flags);
 
-        action = unlink_irq_action(irq, dev_id);
+        action = unlink_irq_action(virt_irq, dev_id);
 
         spin_unlock_irqrestore(&irq_action_lock, flags);
 
         if (unlikely(!action))
                 return;
 
-        synchronize_irq(irq);
+        synchronize_irq(virt_irq);
 
         spin_lock_irqsave(&irq_action_lock, flags);
 
-        bucket = __bucket(irq);
+        bucket = __bucket(real_irq);
         desc = bucket->irq_info;
 
         for (i = 0; i < MAX_IRQ_DESC_ACTION; i++) {
@@ -545,7 +623,7 @@ void free_irq(unsigned int irq, void *dev_id)
                  * the same IMAP are active.
                  */
                 if (ent == NUM_IVECS)
-                        disable_irq(irq);
+                        disable_irq(virt_irq);
         }
 
         spin_unlock_irqrestore(&irq_action_lock, flags);
@@ -554,10 +632,15 @@ void free_irq(unsigned int irq, void *dev_id)
 EXPORT_SYMBOL(free_irq);
 
 #ifdef CONFIG_SMP
-void synchronize_irq(unsigned int irq)
+void synchronize_irq(unsigned int virt_irq)
 {
-        struct ino_bucket *bucket = __bucket(irq);
+        unsigned int real_irq = virt_to_real_irq(virt_irq);
+        struct ino_bucket *bucket;
+
+        if (unlikely(!real_irq))
+                return;
 
+        bucket = __bucket(real_irq);
 #if 0
         /* The following is how I wish I could implement this.
          * Unfortunately the ICLR registers are read-only, you can
@@ -616,7 +699,7 @@ static void process_bucket(struct ino_bucket *bp, struct pt_regs *regs)
 
                 action_mask &= ~mask;
 
-                if (p->handler(__irq(bp), p->dev_id, regs) == IRQ_HANDLED)
+                if (p->handler(bp->virt_irq, p->dev_id, regs) == IRQ_HANDLED)
                         random |= p->flags;
 
                 if (!action_mask)
@@ -637,7 +720,7 @@ static void process_bucket(struct ino_bucket *bp, struct pt_regs *regs)
 
         /* Test and add entropy */
         if (random & SA_SAMPLE_RANDOM)
-                add_interrupt_randomness(PIL_DEVICE_IRQ);
+                add_interrupt_randomness(bp->virt_irq);
 out:
         bp->flags &= ~IBF_INPROGRESS;
 }
@@ -657,7 +740,7 @@ void timer_irq(int irq, struct pt_regs *regs)
         clear_softint(clr_mask);
 
         irq_enter();
-        kstat_this_cpu.irqs[irq]++;
+        kstat_this_cpu.irqs[0]++;
         timer_interrupt(irq, NULL, regs);
         irq_exit();
 }
@@ -1022,13 +1105,13 @@ void __init init_IRQ(void)
                              : "g1");
 }
 
-static struct proc_dir_entry * root_irq_dir;
-static struct proc_dir_entry * irq_dir [NUM_IVECS];
+static struct proc_dir_entry *root_irq_dir;
+static struct proc_dir_entry *irq_dir[NR_IRQS];
 
 #ifdef CONFIG_SMP
 
-static int irq_affinity_read_proc (char *page, char **start, off_t off,
-                        int count, int *eof, void *data)
+static int irq_affinity_read_proc(char *page, char **start, off_t off,
+                                  int count, int *eof, void *data)
 {
         struct ino_bucket *bp = ivector_table + (long)data;
         struct irq_desc *desc = bp->irq_info;
@@ -1047,11 +1130,20 @@ static int irq_affinity_read_proc (char *page, char **start, off_t off,
         return len;
 }
 
-static inline void set_intr_affinity(int irq, cpumask_t hw_aff)
+static inline void set_intr_affinity(int virt_irq, cpumask_t hw_aff)
 {
-        struct ino_bucket *bp = ivector_table + irq;
-        struct irq_desc *desc = bp->irq_info;
-        struct irqaction *ap = desc->action;
+        struct ino_bucket *bp;
+        struct irq_desc *desc;
+        struct irqaction *ap;
+        unsigned int real_irq;
+
+        real_irq = virt_to_real_irq(virt_irq);
+        if (unlikely(!real_irq))
+                return;
+
+        bp = __bucket(real_irq);
+        desc = bp->irq_info;
+        ap = desc->action;
 
         /* Users specify affinity in terms of hw cpu ids.
          * As soon as we do this, handler_irq() might see and take action.
@@ -1060,13 +1152,16 @@ static inline void set_intr_affinity(int irq, cpumask_t hw_aff)
 
         /* Migration is simply done by the next cpu to service this
          * interrupt.
+         *
+         * XXX Broken, this doesn't happen anymore...
          */
 }
 
-static int irq_affinity_write_proc (struct file *file, const char __user *buffer,
-                                        unsigned long count, void *data)
+static int irq_affinity_write_proc(struct file *file,
+                                   const char __user *buffer,
+                                   unsigned long count, void *data)
 {
-        int irq = (long) data, full_count = count, err;
+        int virt_irq = (long) data, full_count = count, err;
         cpumask_t new_value;
 
         err = cpumask_parse(buffer, count, new_value);
@@ -1080,7 +1175,7 @@ static int irq_affinity_write_proc (struct file *file, const char __user *buffer
         if (cpus_empty(new_value))
                 return -EINVAL;
 
-        set_intr_affinity(irq, new_value);
+        set_intr_affinity(virt_irq, new_value);
 
         return full_count;
 }
@@ -1089,18 +1184,18 @@ static int irq_affinity_write_proc (struct file *file, const char __user *buffer
 
 #define MAX_NAMELEN 10
 
-static void register_irq_proc (unsigned int irq)
+static void register_irq_proc(unsigned int virt_irq)
 {
         char name [MAX_NAMELEN];
 
-        if (!root_irq_dir || irq_dir[irq])
+        if (!root_irq_dir || irq_dir[virt_irq])
                 return;
 
         memset(name, 0, MAX_NAMELEN);
-        sprintf(name, "%x", irq);
+        sprintf(name, "%d", virt_irq);
 
         /* create /proc/irq/1234 */
-        irq_dir[irq] = proc_mkdir(name, root_irq_dir);
+        irq_dir[virt_irq] = proc_mkdir(name, root_irq_dir);
 
 #ifdef CONFIG_SMP
         /* XXX SMP affinity not supported on starfire yet. */
@@ -1112,7 +1207,7 @@ static void register_irq_proc (unsigned int irq)
 
                 if (entry) {
                         entry->nlink = 1;
-                        entry->data = (void *)(long)irq;
+                        entry->data = (void *)(long)virt_irq;
                         entry->read_proc = irq_affinity_read_proc;
                         entry->write_proc = irq_affinity_write_proc;
                 }
@@ -1120,7 +1215,7 @@ static void register_irq_proc (unsigned int irq)
 #endif
 }
 
-void init_irq_proc (void)
+void init_irq_proc(void)
 {
         /* create /proc/irq */
         root_irq_dir = proc_mkdir("irq", NULL);