[ARM] pxafb: cleanup of the timing checking code
[safe/jmp/linux-2.6] / drivers / pci / msi.c
index 202b750..74801f7 100644 (file)
  * Copyright (C) Tom Long Nguyen (tom.l.nguyen@intel.com)
  */
 
+#include <linux/err.h>
 #include <linux/mm.h>
 #include <linux/irq.h>
 #include <linux/interrupt.h>
 #include <linux/init.h>
-#include <linux/config.h>
 #include <linux/ioport.h>
-#include <linux/smp_lock.h>
 #include <linux/pci.h>
 #include <linux/proc_fs.h>
+#include <linux/msi.h>
+#include <linux/smp.h>
 
 #include <asm/errno.h>
 #include <asm/io.h>
-#include <asm/smp.h>
 
 #include "pci.h"
 #include "msi.h"
 
-#define MSI_TARGET_CPU         first_cpu(cpu_online_map)
-
-static DEFINE_SPINLOCK(msi_lock);
-static struct msi_desc* msi_desc[NR_IRQS] = { [0 ... NR_IRQS-1] = NULL };
-static kmem_cache_t* msi_cachep;
-
 static int pci_msi_enable = 1;
-static int last_alloc_vector;
-static int nr_released_vectors;
-static int nr_reserved_vectors = NR_HP_RESERVED_VECTORS;
-static int nr_msix_devices;
-
-#ifndef CONFIG_X86_IO_APIC
-int vector_irq[NR_VECTORS] = { [0 ... NR_VECTORS - 1] = -1};
-u8 irq_vector[NR_IRQ_VECTORS] = { FIRST_DEVICE_VECTOR , 0 };
-#endif
 
-static void msi_cache_ctor(void *p, kmem_cache_t *cache, unsigned long flags)
-{
-       memset(p, 0, NR_IRQS * sizeof(struct msi_desc));
-}
+/* Arch hooks */
 
-static int msi_cache_init(void)
+int __attribute__ ((weak))
+arch_msi_check_device(struct pci_dev *dev, int nvec, int type)
 {
-       msi_cachep = kmem_cache_create("msi_cache",
-                       NR_IRQS * sizeof(struct msi_desc),
-                       0, SLAB_HWCACHE_ALIGN, msi_cache_ctor, NULL);
-       if (!msi_cachep)
-               return -ENOMEM;
-
        return 0;
 }
 
-static void msi_set_mask_bit(unsigned int vector, int flag)
+int __attribute__ ((weak))
+arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *entry)
 {
-       struct msi_desc *entry;
-
-       entry = (struct msi_desc *)msi_desc[vector];
-       if (!entry || !entry->dev || !entry->mask_base)
-               return;
-       switch (entry->msi_attrib.type) {
-       case PCI_CAP_ID_MSI:
-       {
-               int             pos;
-               u32             mask_bits;
-
-               pos = (long)entry->mask_base;
-               pci_read_config_dword(entry->dev, pos, &mask_bits);
-               mask_bits &= ~(1);
-               mask_bits |= flag;
-               pci_write_config_dword(entry->dev, pos, mask_bits);
-               break;
-       }
-       case PCI_CAP_ID_MSIX:
-       {
-               int offset = entry->msi_attrib.entry_nr * PCI_MSIX_ENTRY_SIZE +
-                       PCI_MSIX_ENTRY_VECTOR_CTRL_OFFSET;
-               writel(flag, entry->mask_base + offset);
-               break;
-       }
-       default:
-               break;
-       }
+       return 0;
 }
 
-#ifdef CONFIG_SMP
-static void set_msi_affinity(unsigned int vector, cpumask_t cpu_mask)
+int __attribute__ ((weak))
+arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
 {
        struct msi_desc *entry;
-       struct msg_address address;
-       unsigned int irq = vector;
-       unsigned int dest_cpu = first_cpu(cpu_mask);
-
-       entry = (struct msi_desc *)msi_desc[vector];
-       if (!entry || !entry->dev)
-               return;
+       int ret;
 
-       switch (entry->msi_attrib.type) {
-       case PCI_CAP_ID_MSI:
-       {
-               int pos;
-
-               if (!(pos = pci_find_capability(entry->dev, PCI_CAP_ID_MSI)))
-                       return;
-
-               pci_read_config_dword(entry->dev, msi_lower_address_reg(pos),
-                       &address.lo_address.value);
-               address.lo_address.value &= MSI_ADDRESS_DEST_ID_MASK;
-               address.lo_address.value |= (cpu_physical_id(dest_cpu) <<
-                                                                       MSI_TARGET_CPU_SHIFT);
-               entry->msi_attrib.current_cpu = cpu_physical_id(dest_cpu);
-               pci_write_config_dword(entry->dev, msi_lower_address_reg(pos),
-                       address.lo_address.value);
-               set_native_irq_info(irq, cpu_mask);
-               break;
-       }
-       case PCI_CAP_ID_MSIX:
-       {
-               int offset = entry->msi_attrib.entry_nr * PCI_MSIX_ENTRY_SIZE +
-                       PCI_MSIX_ENTRY_LOWER_ADDR_OFFSET;
-
-               address.lo_address.value = readl(entry->mask_base + offset);
-               address.lo_address.value &= MSI_ADDRESS_DEST_ID_MASK;
-               address.lo_address.value |= (cpu_physical_id(dest_cpu) <<
-                                                                       MSI_TARGET_CPU_SHIFT);
-               entry->msi_attrib.current_cpu = cpu_physical_id(dest_cpu);
-               writel(address.lo_address.value, entry->mask_base + offset);
-               set_native_irq_info(irq, cpu_mask);
-               break;
+       list_for_each_entry(entry, &dev->msi_list, list) {
+               ret = arch_setup_msi_irq(dev, entry);
+               if (ret)
+                       return ret;
        }
-       default:
-               break;
-       }
-}
-#endif /* CONFIG_SMP */
 
-static void mask_MSI_irq(unsigned int vector)
-{
-       msi_set_mask_bit(vector, 1);
+       return 0;
 }
 
-static void unmask_MSI_irq(unsigned int vector)
+void __attribute__ ((weak)) arch_teardown_msi_irq(unsigned int irq)
 {
-       msi_set_mask_bit(vector, 0);
+       return;
 }
 
-static unsigned int startup_msi_irq_wo_maskbit(unsigned int vector)
+void __attribute__ ((weak))
+arch_teardown_msi_irqs(struct pci_dev *dev)
 {
        struct msi_desc *entry;
-       unsigned long flags;
 
-       spin_lock_irqsave(&msi_lock, flags);
-       entry = msi_desc[vector];
-       if (!entry || !entry->dev) {
-               spin_unlock_irqrestore(&msi_lock, flags);
-               return 0;
+       list_for_each_entry(entry, &dev->msi_list, list) {
+               if (entry->irq != 0)
+                       arch_teardown_msi_irq(entry->irq);
        }
-       entry->msi_attrib.state = 1;    /* Mark it active */
-       spin_unlock_irqrestore(&msi_lock, flags);
-
-       return 0;       /* never anything pending */
-}
-
-static unsigned int startup_msi_irq_w_maskbit(unsigned int vector)
-{
-       startup_msi_irq_wo_maskbit(vector);
-       unmask_MSI_irq(vector);
-       return 0;       /* never anything pending */
 }
 
-static void shutdown_msi_irq(unsigned int vector)
+static void __msi_set_enable(struct pci_dev *dev, int pos, int enable)
 {
-       struct msi_desc *entry;
-       unsigned long flags;
+       u16 control;
 
-       spin_lock_irqsave(&msi_lock, flags);
-       entry = msi_desc[vector];
-       if (entry && entry->dev)
-               entry->msi_attrib.state = 0;    /* Mark it not active */
-       spin_unlock_irqrestore(&msi_lock, flags);
+       if (pos) {
+               pci_read_config_word(dev, pos + PCI_MSI_FLAGS, &control);
+               control &= ~PCI_MSI_FLAGS_ENABLE;
+               if (enable)
+                       control |= PCI_MSI_FLAGS_ENABLE;
+               pci_write_config_word(dev, pos + PCI_MSI_FLAGS, control);
+       }
 }
 
-static void end_msi_irq_wo_maskbit(unsigned int vector)
+static void msi_set_enable(struct pci_dev *dev, int enable)
 {
-       move_native_irq(vector);
-       ack_APIC_irq();
+       __msi_set_enable(dev, pci_find_capability(dev, PCI_CAP_ID_MSI), enable);
 }
 
-static void end_msi_irq_w_maskbit(unsigned int vector)
+static void msix_set_enable(struct pci_dev *dev, int enable)
 {
-       move_native_irq(vector);
-       unmask_MSI_irq(vector);
-       ack_APIC_irq();
-}
+       int pos;
+       u16 control;
 
-static void do_nothing(unsigned int vector)
-{
+       pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
+       if (pos) {
+               pci_read_config_word(dev, pos + PCI_MSIX_FLAGS, &control);
+               control &= ~PCI_MSIX_FLAGS_ENABLE;
+               if (enable)
+                       control |= PCI_MSIX_FLAGS_ENABLE;
+               pci_write_config_word(dev, pos + PCI_MSIX_FLAGS, control);
+       }
 }
 
-/*
- * Interrupt Type for MSI-X PCI/PCI-X/PCI-Express Devices,
- * which implement the MSI-X Capability Structure.
- */
-static struct hw_interrupt_type msix_irq_type = {
-       .typename       = "PCI-MSI-X",
-       .startup        = startup_msi_irq_w_maskbit,
-       .shutdown       = shutdown_msi_irq,
-       .enable         = unmask_MSI_irq,
-       .disable        = mask_MSI_irq,
-       .ack            = mask_MSI_irq,
-       .end            = end_msi_irq_w_maskbit,
-       .set_affinity   = set_msi_irq_affinity
-};
-
-/*
- * Interrupt Type for MSI PCI/PCI-X/PCI-Express Devices,
- * which implement the MSI Capability Structure with
- * Mask-and-Pending Bits.
- */
-static struct hw_interrupt_type msi_irq_w_maskbit_type = {
-       .typename       = "PCI-MSI",
-       .startup        = startup_msi_irq_w_maskbit,
-       .shutdown       = shutdown_msi_irq,
-       .enable         = unmask_MSI_irq,
-       .disable        = mask_MSI_irq,
-       .ack            = mask_MSI_irq,
-       .end            = end_msi_irq_w_maskbit,
-       .set_affinity   = set_msi_irq_affinity
-};
-
-/*
- * Interrupt Type for MSI PCI/PCI-X/PCI-Express Devices,
- * which implement the MSI Capability Structure without
- * Mask-and-Pending Bits.
- */
-static struct hw_interrupt_type msi_irq_wo_maskbit_type = {
-       .typename       = "PCI-MSI",
-       .startup        = startup_msi_irq_wo_maskbit,
-       .shutdown       = shutdown_msi_irq,
-       .enable         = do_nothing,
-       .disable        = do_nothing,
-       .ack            = do_nothing,
-       .end            = end_msi_irq_wo_maskbit,
-       .set_affinity   = set_msi_irq_affinity
-};
-
-static void msi_data_init(struct msg_data *msi_data,
-                         unsigned int vector)
+static void msix_flush_writes(unsigned int irq)
 {
-       memset(msi_data, 0, sizeof(struct msg_data));
-       msi_data->vector = (u8)vector;
-       msi_data->delivery_mode = MSI_DELIVERY_MODE;
-       msi_data->level = MSI_LEVEL_MODE;
-       msi_data->trigger = MSI_TRIGGER_MODE;
-}
+       struct msi_desc *entry;
 
-static void msi_address_init(struct msg_address *msi_address)
-{
-       unsigned int    dest_id;
-       unsigned long   dest_phys_id = cpu_physical_id(MSI_TARGET_CPU);
-
-       memset(msi_address, 0, sizeof(struct msg_address));
-       msi_address->hi_address = (u32)0;
-       dest_id = (MSI_ADDRESS_HEADER << MSI_ADDRESS_HEADER_SHIFT);
-       msi_address->lo_address.u.dest_mode = MSI_PHYSICAL_MODE;
-       msi_address->lo_address.u.redirection_hint = MSI_REDIRECTION_HINT_MODE;
-       msi_address->lo_address.u.dest_id = dest_id;
-       msi_address->lo_address.value |= (dest_phys_id << MSI_TARGET_CPU_SHIFT);
+       entry = get_irq_msi(irq);
+       BUG_ON(!entry || !entry->dev);
+       switch (entry->msi_attrib.type) {
+       case PCI_CAP_ID_MSI:
+               /* nothing to do */
+               break;
+       case PCI_CAP_ID_MSIX:
+       {
+               int offset = entry->msi_attrib.entry_nr * PCI_MSIX_ENTRY_SIZE +
+                       PCI_MSIX_ENTRY_VECTOR_CTRL_OFFSET;
+               readl(entry->mask_base + offset);
+               break;
+       }
+       default:
+               BUG();
+               break;
+       }
 }
 
-static int msi_free_vector(struct pci_dev* dev, int vector, int reassign);
-static int assign_msi_vector(void)
+/*
+ * PCI 2.3 does not specify mask bits for each MSI interrupt.  Attempting to
+ * mask all MSI interrupts by clearing the MSI enable bit does not work
+ * reliably as devices without an INTx disable bit will then generate a
+ * level IRQ which will never be cleared.
+ *
+ * Returns 1 if it succeeded in masking the interrupt and 0 if the device
+ * doesn't support MSI masking.
+ */
+static int msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag)
 {
-       static int new_vector_avail = 1;
-       int vector;
-       unsigned long flags;
+       struct msi_desc *entry;
 
-       /*
-        * msi_lock is provided to ensure that successful allocation of MSI
-        * vector is assigned unique among drivers.
-        */
-       spin_lock_irqsave(&msi_lock, flags);
-
-       if (!new_vector_avail) {
-               int free_vector = 0;
-
-               /*
-                * vector_irq[] = -1 indicates that this specific vector is:
-                * - assigned for MSI (since MSI have no associated IRQ) or
-                * - assigned for legacy if less than 16, or
-                * - having no corresponding 1:1 vector-to-IOxAPIC IRQ mapping
-                * vector_irq[] = 0 indicates that this vector, previously
-                * assigned for MSI, is freed by hotplug removed operations.
-                * This vector will be reused for any subsequent hotplug added
-                * operations.
-                * vector_irq[] > 0 indicates that this vector is assigned for
-                * IOxAPIC IRQs. This vector and its value provides a 1-to-1
-                * vector-to-IOxAPIC IRQ mapping.
-                */
-               for (vector = FIRST_DEVICE_VECTOR; vector < NR_IRQS; vector++) {
-                       if (vector_irq[vector] != 0)
-                               continue;
-                       free_vector = vector;
-                       if (!msi_desc[vector])
-                               break;
-                       else
-                               continue;
-               }
-               if (!free_vector) {
-                       spin_unlock_irqrestore(&msi_lock, flags);
-                       return -EBUSY;
-               }
-               vector_irq[free_vector] = -1;
-               nr_released_vectors--;
-               spin_unlock_irqrestore(&msi_lock, flags);
-               if (msi_desc[free_vector] != NULL) {
-                       struct pci_dev *dev;
-                       int tail;
-
-                       /* free all linked vectors before re-assign */
-                       do {
-                               spin_lock_irqsave(&msi_lock, flags);
-                               dev = msi_desc[free_vector]->dev;
-                               tail = msi_desc[free_vector]->link.tail;
-                               spin_unlock_irqrestore(&msi_lock, flags);
-                               msi_free_vector(dev, tail, 1);
-                       } while (free_vector != tail);
+       entry = get_irq_msi(irq);
+       BUG_ON(!entry || !entry->dev);
+       switch (entry->msi_attrib.type) {
+       case PCI_CAP_ID_MSI:
+               if (entry->msi_attrib.maskbit) {
+                       int pos;
+                       u32 mask_bits;
+
+                       pos = (long)entry->mask_base;
+                       pci_read_config_dword(entry->dev, pos, &mask_bits);
+                       mask_bits &= ~(mask);
+                       mask_bits |= flag & mask;
+                       pci_write_config_dword(entry->dev, pos, mask_bits);
+               } else {
+                       return 0;
                }
-
-               return free_vector;
+               break;
+       case PCI_CAP_ID_MSIX:
+       {
+               int offset = entry->msi_attrib.entry_nr * PCI_MSIX_ENTRY_SIZE +
+                       PCI_MSIX_ENTRY_VECTOR_CTRL_OFFSET;
+               writel(flag, entry->mask_base + offset);
+               readl(entry->mask_base + offset);
+               break;
        }
-       vector = assign_irq_vector(AUTO_ASSIGN);
-       last_alloc_vector = vector;
-       if (vector  == LAST_DEVICE_VECTOR)
-               new_vector_avail = 0;
-
-       spin_unlock_irqrestore(&msi_lock, flags);
-       return vector;
+       default:
+               BUG();
+               break;
+       }
+       entry->msi_attrib.masked = !!flag;
+       return 1;
 }
 
-static int get_new_vector(void)
+void read_msi_msg(unsigned int irq, struct msi_msg *msg)
 {
-       int vector;
-
-       if ((vector = assign_msi_vector()) > 0)
-               set_intr_gate(vector, interrupt[vector]);
+       struct msi_desc *entry = get_irq_msi(irq);
+       switch(entry->msi_attrib.type) {
+       case PCI_CAP_ID_MSI:
+       {
+               struct pci_dev *dev = entry->dev;
+               int pos = entry->msi_attrib.pos;
+               u16 data;
+
+               pci_read_config_dword(dev, msi_lower_address_reg(pos),
+                                       &msg->address_lo);
+               if (entry->msi_attrib.is_64) {
+                       pci_read_config_dword(dev, msi_upper_address_reg(pos),
+                                               &msg->address_hi);
+                       pci_read_config_word(dev, msi_data_reg(pos, 1), &data);
+               } else {
+                       msg->address_hi = 0;
+                       pci_read_config_word(dev, msi_data_reg(pos, 0), &data);
+               }
+               msg->data = data;
+               break;
+       }
+       case PCI_CAP_ID_MSIX:
+       {
+               void __iomem *base;
+               base = entry->mask_base +
+                       entry->msi_attrib.entry_nr * PCI_MSIX_ENTRY_SIZE;
 
-       return vector;
+               msg->address_lo = readl(base + PCI_MSIX_ENTRY_LOWER_ADDR_OFFSET);
+               msg->address_hi = readl(base + PCI_MSIX_ENTRY_UPPER_ADDR_OFFSET);
+               msg->data = readl(base + PCI_MSIX_ENTRY_DATA_OFFSET);
+               break;
+       }
+       default:
+               BUG();
+       }
 }
 
-static int msi_init(void)
+void write_msi_msg(unsigned int irq, struct msi_msg *msg)
 {
-       static int status = -ENOMEM;
-
-       if (!status)
-               return status;
-
-       if (pci_msi_quirk) {
-               pci_msi_enable = 0;
-               printk(KERN_WARNING "PCI: MSI quirk detected. MSI disabled.\n");
-               status = -EINVAL;
-               return status;
+       struct msi_desc *entry = get_irq_msi(irq);
+       switch (entry->msi_attrib.type) {
+       case PCI_CAP_ID_MSI:
+       {
+               struct pci_dev *dev = entry->dev;
+               int pos = entry->msi_attrib.pos;
+
+               pci_write_config_dword(dev, msi_lower_address_reg(pos),
+                                       msg->address_lo);
+               if (entry->msi_attrib.is_64) {
+                       pci_write_config_dword(dev, msi_upper_address_reg(pos),
+                                               msg->address_hi);
+                       pci_write_config_word(dev, msi_data_reg(pos, 1),
+                                               msg->data);
+               } else {
+                       pci_write_config_word(dev, msi_data_reg(pos, 0),
+                                               msg->data);
+               }
+               break;
        }
-
-       if ((status = msi_cache_init()) < 0) {
-               pci_msi_enable = 0;
-               printk(KERN_WARNING "PCI: MSI cache init failed\n");
-               return status;
+       case PCI_CAP_ID_MSIX:
+       {
+               void __iomem *base;
+               base = entry->mask_base +
+                       entry->msi_attrib.entry_nr * PCI_MSIX_ENTRY_SIZE;
+
+               writel(msg->address_lo,
+                       base + PCI_MSIX_ENTRY_LOWER_ADDR_OFFSET);
+               writel(msg->address_hi,
+                       base + PCI_MSIX_ENTRY_UPPER_ADDR_OFFSET);
+               writel(msg->data, base + PCI_MSIX_ENTRY_DATA_OFFSET);
+               break;
        }
-       last_alloc_vector = assign_irq_vector(AUTO_ASSIGN);
-       if (last_alloc_vector < 0) {
-               pci_msi_enable = 0;
-               printk(KERN_WARNING "PCI: No interrupt vectors available for MSI\n");
-               status = -EBUSY;
-               return status;
+       default:
+               BUG();
        }
-       vector_irq[last_alloc_vector] = 0;
-       nr_released_vectors++;
+       entry->msg = *msg;
+}
 
-       return status;
+void mask_msi_irq(unsigned int irq)
+{
+       msi_set_mask_bits(irq, 1, 1);
+       msix_flush_writes(irq);
 }
 
-static int get_msi_vector(struct pci_dev *dev)
+void unmask_msi_irq(unsigned int irq)
 {
-       return get_new_vector();
+       msi_set_mask_bits(irq, 1, 0);
+       msix_flush_writes(irq);
 }
 
+static int msi_free_irqs(struct pci_dev* dev);
+
+
 static struct msi_desc* alloc_msi_entry(void)
 {
        struct msi_desc *entry;
 
-       entry = kmem_cache_alloc(msi_cachep, SLAB_KERNEL);
+       entry = kzalloc(sizeof(struct msi_desc), GFP_KERNEL);
        if (!entry)
                return NULL;
 
-       memset(entry, 0, sizeof(struct msi_desc));
-       entry->link.tail = entry->link.head = 0;        /* single message */
+       INIT_LIST_HEAD(&entry->list);
+       entry->irq = 0;
        entry->dev = NULL;
 
        return entry;
 }
 
-static void attach_msi_entry(struct msi_desc *entry, int vector)
+static void pci_intx_for_msi(struct pci_dev *dev, int enable)
 {
-       unsigned long flags;
-
-       spin_lock_irqsave(&msi_lock, flags);
-       msi_desc[vector] = entry;
-       spin_unlock_irqrestore(&msi_lock, flags);
-}
-
-static void irq_handler_init(int cap_id, int pos, int mask)
-{
-       spin_lock(&irq_desc[pos].lock);
-       if (cap_id == PCI_CAP_ID_MSIX)
-               irq_desc[pos].handler = &msix_irq_type;
-       else {
-               if (!mask)
-                       irq_desc[pos].handler = &msi_irq_wo_maskbit_type;
-               else
-                       irq_desc[pos].handler = &msi_irq_w_maskbit_type;
-       }
-       spin_unlock(&irq_desc[pos].lock);
+       if (!(dev->dev_flags & PCI_DEV_FLAGS_MSI_INTX_DISABLE_BUG))
+               pci_intx(dev, enable);
 }
 
-static void enable_msi_mode(struct pci_dev *dev, int pos, int type)
+static void __pci_restore_msi_state(struct pci_dev *dev)
 {
+       int pos;
        u16 control;
+       struct msi_desc *entry;
 
-       pci_read_config_word(dev, msi_control_reg(pos), &control);
-       if (type == PCI_CAP_ID_MSI) {
-               /* Set enabled bits to single MSI & enable MSI_enable bit */
-               msi_enable(control, 1);
-               pci_write_config_word(dev, msi_control_reg(pos), control);
-       } else {
-               msix_enable(control);
-               pci_write_config_word(dev, msi_control_reg(pos), control);
-       }
-       if (pci_find_capability(dev, PCI_CAP_ID_EXP)) {
-               /* PCI Express Endpoint device detected */
-               pci_intx(dev, 0);  /* disable intx */
-       }
+       if (!dev->msi_enabled)
+               return;
+
+       entry = get_irq_msi(dev->irq);
+       pos = entry->msi_attrib.pos;
+
+       pci_intx_for_msi(dev, 0);
+       msi_set_enable(dev, 0);
+       write_msi_msg(dev->irq, &entry->msg);
+       if (entry->msi_attrib.maskbit)
+               msi_set_mask_bits(dev->irq, entry->msi_attrib.maskbits_mask,
+                                 entry->msi_attrib.masked);
+
+       pci_read_config_word(dev, pos + PCI_MSI_FLAGS, &control);
+       control &= ~PCI_MSI_FLAGS_QSIZE;
+       control |= PCI_MSI_FLAGS_ENABLE;
+       pci_write_config_word(dev, pos + PCI_MSI_FLAGS, control);
 }
 
-void disable_msi_mode(struct pci_dev *dev, int pos, int type)
+static void __pci_restore_msix_state(struct pci_dev *dev)
 {
+       int pos;
+       struct msi_desc *entry;
        u16 control;
 
-       pci_read_config_word(dev, msi_control_reg(pos), &control);
-       if (type == PCI_CAP_ID_MSI) {
-               /* Set enabled bits to single MSI & enable MSI_enable bit */
-               msi_disable(control);
-               pci_write_config_word(dev, msi_control_reg(pos), control);
-       } else {
-               msix_disable(control);
-               pci_write_config_word(dev, msi_control_reg(pos), control);
-       }
-       if (pci_find_capability(dev, PCI_CAP_ID_EXP)) {
-               /* PCI Express Endpoint device detected */
-               pci_intx(dev, 1);  /* enable intx */
-       }
-}
+       if (!dev->msix_enabled)
+               return;
 
-static int msi_lookup_vector(struct pci_dev *dev, int type)
-{
-       int vector;
-       unsigned long flags;
-
-       spin_lock_irqsave(&msi_lock, flags);
-       for (vector = FIRST_DEVICE_VECTOR; vector < NR_IRQS; vector++) {
-               if (!msi_desc[vector] || msi_desc[vector]->dev != dev ||
-                       msi_desc[vector]->msi_attrib.type != type ||
-                       msi_desc[vector]->msi_attrib.default_vector != dev->irq)
-                       continue;
-               spin_unlock_irqrestore(&msi_lock, flags);
-               /* This pre-assigned MSI vector for this device
-                  already exits. Override dev->irq with this vector */
-               dev->irq = vector;
-               return 0;
+       /* route the table */
+       pci_intx_for_msi(dev, 0);
+       msix_set_enable(dev, 0);
+
+       list_for_each_entry(entry, &dev->msi_list, list) {
+               write_msi_msg(entry->irq, &entry->msg);
+               msi_set_mask_bits(entry->irq, 1, entry->msi_attrib.masked);
        }
-       spin_unlock_irqrestore(&msi_lock, flags);
 
-       return -EACCES;
+       BUG_ON(list_empty(&dev->msi_list));
+       entry = list_entry(dev->msi_list.next, struct msi_desc, list);
+       pos = entry->msi_attrib.pos;
+       pci_read_config_word(dev, pos + PCI_MSIX_FLAGS, &control);
+       control &= ~PCI_MSIX_FLAGS_MASKALL;
+       control |= PCI_MSIX_FLAGS_ENABLE;
+       pci_write_config_word(dev, pos + PCI_MSIX_FLAGS, control);
 }
 
-void pci_scan_msi_device(struct pci_dev *dev)
+void pci_restore_msi_state(struct pci_dev *dev)
 {
-       if (!dev)
-               return;
-
-       if (pci_find_capability(dev, PCI_CAP_ID_MSIX) > 0)
-               nr_msix_devices++;
-       else if (pci_find_capability(dev, PCI_CAP_ID_MSI) > 0)
-               nr_reserved_vectors++;
+       __pci_restore_msi_state(dev);
+       __pci_restore_msix_state(dev);
 }
+EXPORT_SYMBOL_GPL(pci_restore_msi_state);
 
 /**
  * msi_capability_init - configure device's MSI capability structure
  * @dev: pointer to the pci_dev data structure of MSI device function
  *
  * Setup the MSI capability structure of device function with a single
- * MSI vector, regardless of device function is capable of handling
+ * MSI irq, regardless of device function is capable of handling
  * multiple messages. A return of zero indicates the successful setup
- * of an entry zero with the new MSI vector or non-zero for otherwise.
+ * of an entry zero with the new MSI irq or non-zero for otherwise.
  **/
 static int msi_capability_init(struct pci_dev *dev)
 {
        struct msi_desc *entry;
-       struct msg_address address;
-       struct msg_data data;
-       int pos, vector;
+       int pos, ret;
        u16 control;
 
+       msi_set_enable(dev, 0); /* Ensure msi is disabled as I set it up */
+
        pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
        pci_read_config_word(dev, msi_control_reg(pos), &control);
        /* MSI Entry Initialization */
-       if (!(entry = alloc_msi_entry()))
+       entry = alloc_msi_entry();
+       if (!entry)
                return -ENOMEM;
 
-       if ((vector = get_msi_vector(dev)) < 0) {
-               kmem_cache_free(msi_cachep, entry);
-               return -EBUSY;
-       }
-       entry->link.head = vector;
-       entry->link.tail = vector;
        entry->msi_attrib.type = PCI_CAP_ID_MSI;
-       entry->msi_attrib.state = 0;                    /* Mark it not active */
+       entry->msi_attrib.is_64 = is_64bit_address(control);
        entry->msi_attrib.entry_nr = 0;
        entry->msi_attrib.maskbit = is_mask_bit_support(control);
-       entry->msi_attrib.default_vector = dev->irq;    /* Save IOAPIC IRQ */
-       dev->irq = vector;
-       entry->dev = dev;
-       if (is_mask_bit_support(control)) {
+       entry->msi_attrib.masked = 1;
+       entry->msi_attrib.default_irq = dev->irq;       /* Save IOAPIC IRQ */
+       entry->msi_attrib.pos = pos;
+       if (entry->msi_attrib.maskbit) {
                entry->mask_base = (void __iomem *)(long)msi_mask_bits_reg(pos,
-                               is_64bit_address(control));
+                               entry->msi_attrib.is_64);
        }
-       /* Replace with MSI handler */
-       irq_handler_init(PCI_CAP_ID_MSI, vector, entry->msi_attrib.maskbit);
-       /* Configure MSI capability structure */
-       msi_address_init(&address);
-       msi_data_init(&data, vector);
-       entry->msi_attrib.current_cpu = ((address.lo_address.u.dest_id >>
-                               MSI_TARGET_CPU_SHIFT) & MSI_TARGET_CPU_MASK);
-       pci_write_config_dword(dev, msi_lower_address_reg(pos),
-                       address.lo_address.value);
-       if (is_64bit_address(control)) {
-               pci_write_config_dword(dev,
-                       msi_upper_address_reg(pos), address.hi_address);
-               pci_write_config_word(dev,
-                       msi_data_reg(pos, 1), *((u32*)&data));
-       } else
-               pci_write_config_word(dev,
-                       msi_data_reg(pos, 0), *((u32*)&data));
+       entry->dev = dev;
        if (entry->msi_attrib.maskbit) {
                unsigned int maskbits, temp;
                /* All MSIs are unmasked by default, Mask them all */
                pci_read_config_dword(dev,
-                       msi_mask_bits_reg(pos, is_64bit_address(control)),
+                       msi_mask_bits_reg(pos, entry->msi_attrib.is_64),
                        &maskbits);
                temp = (1 << multi_msi_capable(control));
                temp = ((temp - 1) & ~temp);
                maskbits |= temp;
-               pci_write_config_dword(dev,
-                       msi_mask_bits_reg(pos, is_64bit_address(control)),
-                       maskbits);
+               pci_write_config_dword(dev, entry->msi_attrib.is_64, maskbits);
+               entry->msi_attrib.maskbits_mask = temp;
+       }
+       list_add_tail(&entry->list, &dev->msi_list);
+
+       /* Configure MSI capability structure */
+       ret = arch_setup_msi_irqs(dev, 1, PCI_CAP_ID_MSI);
+       if (ret) {
+               msi_free_irqs(dev);
+               return ret;
        }
-       attach_msi_entry(entry, vector);
+
        /* Set MSI enabled bits  */
-       enable_msi_mode(dev, pos, PCI_CAP_ID_MSI);
+       pci_intx_for_msi(dev, 0);
+       msi_set_enable(dev, 1);
+       dev->msi_enabled = 1;
 
+       dev->irq = entry->irq;
        return 0;
 }
 
@@ -583,30 +420,31 @@ static int msi_capability_init(struct pci_dev *dev)
  * @nvec: number of @entries
  *
  * Setup the MSI-X capability structure of device function with a
- * single MSI-X vector. A return of zero indicates the successful setup of
- * requested MSI-X entries with allocated vectors or non-zero for otherwise.
+ * single MSI-X irq. A return of zero indicates the successful setup of
+ * requested MSI-X entries with allocated irqs or non-zero for otherwise.
  **/
 static int msix_capability_init(struct pci_dev *dev,
                                struct msix_entry *entries, int nvec)
 {
-       struct msi_desc *head = NULL, *tail = NULL, *entry = NULL;
-       struct msg_address address;
-       struct msg_data data;
-       int vector, pos, i, j, nr_entries, temp = 0;
-       u32 phys_addr, table_offset;
+       struct msi_desc *entry;
+       int pos, i, j, nr_entries, ret;
+       unsigned long phys_addr;
+       u32 table_offset;
        u16 control;
        u8 bir;
        void __iomem *base;
 
+       msix_set_enable(dev, 0);/* Ensure msix is disabled as I set it up */
+
        pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
        /* Request & Map MSI-X table region */
        pci_read_config_word(dev, msi_control_reg(pos), &control);
        nr_entries = multi_msix_capable(control);
-       pci_read_config_dword(dev, msix_table_offset_reg(pos),
-               &table_offset);
+
+       pci_read_config_dword(dev, msix_table_offset_reg(pos), &table_offset);
        bir = (u8)(table_offset & PCI_MSIX_FLAGS_BIRMASK);
-       phys_addr = pci_resource_start (dev, bir);
-       phys_addr += (u32)(table_offset & ~PCI_MSIX_FLAGS_BIRMASK);
+       table_offset &= ~PCI_MSIX_FLAGS_BIRMASK;
+       phys_addr = pci_resource_start (dev, bir) + table_offset;
        base = ioremap_nocache(phys_addr, nr_entries * PCI_MSIX_ENTRY_SIZE);
        if (base == NULL)
                return -ENOMEM;
@@ -616,60 +454,97 @@ static int msix_capability_init(struct pci_dev *dev,
                entry = alloc_msi_entry();
                if (!entry)
                        break;
-               if ((vector = get_msi_vector(dev)) < 0)
-                       break;
 
                j = entries[i].entry;
-               entries[i].vector = vector;
                entry->msi_attrib.type = PCI_CAP_ID_MSIX;
-               entry->msi_attrib.state = 0;            /* Mark it not active */
+               entry->msi_attrib.is_64 = 1;
                entry->msi_attrib.entry_nr = j;
                entry->msi_attrib.maskbit = 1;
-               entry->msi_attrib.default_vector = dev->irq;
+               entry->msi_attrib.masked = 1;
+               entry->msi_attrib.default_irq = dev->irq;
+               entry->msi_attrib.pos = pos;
                entry->dev = dev;
                entry->mask_base = base;
-               if (!head) {
-                       entry->link.head = vector;
-                       entry->link.tail = vector;
-                       head = entry;
-               } else {
-                       entry->link.head = temp;
-                       entry->link.tail = tail->link.tail;
-                       tail->link.tail = vector;
-                       head->link.head = vector;
-               }
-               temp = vector;
-               tail = entry;
-               /* Replace with MSI-X handler */
-               irq_handler_init(PCI_CAP_ID_MSIX, vector, 1);
-               /* Configure MSI-X capability structure */
-               msi_address_init(&address);
-               msi_data_init(&data, vector);
-               entry->msi_attrib.current_cpu =
-                       ((address.lo_address.u.dest_id >>
-                       MSI_TARGET_CPU_SHIFT) & MSI_TARGET_CPU_MASK);
-               writel(address.lo_address.value,
-                       base + j * PCI_MSIX_ENTRY_SIZE +
-                       PCI_MSIX_ENTRY_LOWER_ADDR_OFFSET);
-               writel(address.hi_address,
-                       base + j * PCI_MSIX_ENTRY_SIZE +
-                       PCI_MSIX_ENTRY_UPPER_ADDR_OFFSET);
-               writel(*(u32*)&data,
-                       base + j * PCI_MSIX_ENTRY_SIZE +
-                       PCI_MSIX_ENTRY_DATA_OFFSET);
-               attach_msi_entry(entry, vector);
+
+               list_add_tail(&entry->list, &dev->msi_list);
        }
-       if (i != nvec) {
-               i--;
-               for (; i >= 0; i--) {
-                       vector = (entries + i)->vector;
-                       msi_free_vector(dev, vector, 0);
-                       (entries + i)->vector = 0;
+
+       ret = arch_setup_msi_irqs(dev, nvec, PCI_CAP_ID_MSIX);
+       if (ret) {
+               int avail = 0;
+               list_for_each_entry(entry, &dev->msi_list, list) {
+                       if (entry->irq != 0) {
+                               avail++;
+                       }
                }
-               return -EBUSY;
+
+               msi_free_irqs(dev);
+
+               /* If we had some success report the number of irqs
+                * we succeeded in setting up.
+                */
+               if (avail == 0)
+                       avail = ret;
+               return avail;
+       }
+
+       i = 0;
+       list_for_each_entry(entry, &dev->msi_list, list) {
+               entries[i].vector = entry->irq;
+               set_irq_msi(entry->irq, entry);
+               i++;
        }
        /* Set MSI-X enabled bits */
-       enable_msi_mode(dev, pos, PCI_CAP_ID_MSIX);
+       pci_intx_for_msi(dev, 0);
+       msix_set_enable(dev, 1);
+       dev->msix_enabled = 1;
+
+       return 0;
+}
+
+/**
+ * pci_msi_check_device - check whether MSI may be enabled on a device
+ * @dev: pointer to the pci_dev data structure of MSI device function
+ * @nvec: how many MSIs have been requested ?
+ * @type: are we checking for MSI or MSI-X ?
+ *
+ * Look at global flags, the device itself, and its parent busses
+ * to determine if MSI/-X are supported for the device. If MSI/-X is
+ * supported return 0, else return an error code.
+ **/
+static int pci_msi_check_device(struct pci_dev* dev, int nvec, int type)
+{
+       struct pci_bus *bus;
+       int ret;
+
+       /* MSI must be globally enabled and supported by the device */
+       if (!pci_msi_enable || !dev || dev->no_msi)
+               return -EINVAL;
+
+       /*
+        * You can't ask to have 0 or less MSIs configured.
+        *  a) it's stupid ..
+        *  b) the list manipulation code assumes nvec >= 1.
+        */
+       if (nvec < 1)
+               return -ERANGE;
+
+       /* Any bridge which does NOT route MSI transactions from it's
+        * secondary bus to it's primary bus must set NO_MSI flag on
+        * the secondary pci_bus.
+        * We expect only arch-specific PCI host bus controller driver
+        * or quirks for specific PCI bridges to be setting NO_MSI.
+        */
+       for (bus = dev->bus; bus; bus = bus->parent)
+               if (bus->bus_flags & PCI_BUS_FLAGS_NO_MSI)
+                       return -EINVAL;
+
+       ret = arch_msi_check_device(dev, nvec, type);
+       if (ret)
+               return ret;
+
+       if (!pci_find_capability(dev, type))
+               return -EINVAL;
 
        return 0;
 }
@@ -679,218 +554,96 @@ static int msix_capability_init(struct pci_dev *dev,
  * @dev: pointer to the pci_dev data structure of MSI device function
  *
  * Setup the MSI capability structure of device function with
- * a single MSI vector upon its software driver call to request for
+ * a single MSI irq upon its software driver call to request for
  * MSI mode enabled on its hardware device function. A return of zero
  * indicates the successful setup of an entry zero with the new MSI
- * vector or non-zero for otherwise.
+ * irq or non-zero for otherwise.
  **/
 int pci_enable_msi(struct pci_dev* dev)
 {
-       int pos, temp, status = -EINVAL;
-       u16 control;
-
-       if (!pci_msi_enable || !dev)
-               return status;
+       int status;
 
-       if (dev->no_msi)
+       status = pci_msi_check_device(dev, 1, PCI_CAP_ID_MSI);
+       if (status)
                return status;
 
-       temp = dev->irq;
+       WARN_ON(!!dev->msi_enabled);
 
-       if ((status = msi_init()) < 0)
-               return status;
-
-       if (!(pos = pci_find_capability(dev, PCI_CAP_ID_MSI)))
+       /* Check whether driver already requested for MSI-X irqs */
+       if (dev->msix_enabled) {
+               dev_info(&dev->dev, "can't enable MSI "
+                        "(MSI-X already enabled)\n");
                return -EINVAL;
-
-       pci_read_config_word(dev, msi_control_reg(pos), &control);
-       if (control & PCI_MSI_FLAGS_ENABLE)
-               return 0;                       /* Already in MSI mode */
-
-       if (!msi_lookup_vector(dev, PCI_CAP_ID_MSI)) {
-               /* Lookup Sucess */
-               unsigned long flags;
-
-               spin_lock_irqsave(&msi_lock, flags);
-               if (!vector_irq[dev->irq]) {
-                       msi_desc[dev->irq]->msi_attrib.state = 0;
-                       vector_irq[dev->irq] = -1;
-                       nr_released_vectors--;
-                       spin_unlock_irqrestore(&msi_lock, flags);
-                       enable_msi_mode(dev, pos, PCI_CAP_ID_MSI);
-                       return 0;
-               }
-               spin_unlock_irqrestore(&msi_lock, flags);
-               dev->irq = temp;
-       }
-       /* Check whether driver already requested for MSI-X vectors */
-       if ((pos = pci_find_capability(dev, PCI_CAP_ID_MSIX)) > 0 &&
-               !msi_lookup_vector(dev, PCI_CAP_ID_MSIX)) {
-                       printk(KERN_INFO "PCI: %s: Can't enable MSI.  "
-                              "Device already has MSI-X vectors assigned\n",
-                              pci_name(dev));
-                       dev->irq = temp;
-                       return -EINVAL;
        }
        status = msi_capability_init(dev);
-       if (!status) {
-               if (!pos)
-                       nr_reserved_vectors--;  /* Only MSI capable */
-               else if (nr_msix_devices > 0)
-                       nr_msix_devices--;      /* Both MSI and MSI-X capable,
-                                                  but choose enabling MSI */
-       }
-
        return status;
 }
+EXPORT_SYMBOL(pci_enable_msi);
 
-void pci_disable_msi(struct pci_dev* dev)
+void pci_msi_shutdown(struct pci_dev* dev)
 {
        struct msi_desc *entry;
-       int pos, default_vector;
-       u16 control;
-       unsigned long flags;
 
-       if (!dev || !(pos = pci_find_capability(dev, PCI_CAP_ID_MSI)))
+       if (!pci_msi_enable || !dev || !dev->msi_enabled)
                return;
 
-       pci_read_config_word(dev, msi_control_reg(pos), &control);
-       if (!(control & PCI_MSI_FLAGS_ENABLE))
-               return;
+       msi_set_enable(dev, 0);
+       pci_intx_for_msi(dev, 1);
+       dev->msi_enabled = 0;
 
-       spin_lock_irqsave(&msi_lock, flags);
-       entry = msi_desc[dev->irq];
-       if (!entry || !entry->dev || entry->msi_attrib.type != PCI_CAP_ID_MSI) {
-               spin_unlock_irqrestore(&msi_lock, flags);
-               return;
-       }
-       if (entry->msi_attrib.state) {
-               spin_unlock_irqrestore(&msi_lock, flags);
-               printk(KERN_WARNING "PCI: %s: pci_disable_msi() called without "
-                      "free_irq() on MSI vector %d\n",
-                      pci_name(dev), dev->irq);
-               BUG_ON(entry->msi_attrib.state > 0);
-       } else {
-               vector_irq[dev->irq] = 0; /* free it */
-               nr_released_vectors++;
-               default_vector = entry->msi_attrib.default_vector;
-               spin_unlock_irqrestore(&msi_lock, flags);
-               /* Restore dev->irq to its default pin-assertion vector */
-               dev->irq = default_vector;
-               disable_msi_mode(dev, pci_find_capability(dev, PCI_CAP_ID_MSI),
-                                       PCI_CAP_ID_MSI);
+       BUG_ON(list_empty(&dev->msi_list));
+       entry = list_entry(dev->msi_list.next, struct msi_desc, list);
+       /* Return the the pci reset with msi irqs unmasked */
+       if (entry->msi_attrib.maskbit) {
+               u32 mask = entry->msi_attrib.maskbits_mask;
+               msi_set_mask_bits(dev->irq, mask, ~mask);
        }
-}
+       if (!entry->dev || entry->msi_attrib.type != PCI_CAP_ID_MSI)
+               return;
 
-static int msi_free_vector(struct pci_dev* dev, int vector, int reassign)
+       /* Restore dev->irq to its default pin-assertion irq */
+       dev->irq = entry->msi_attrib.default_irq;
+}
+void pci_disable_msi(struct pci_dev* dev)
 {
        struct msi_desc *entry;
-       int head, entry_nr, type;
-       void __iomem *base;
-       unsigned long flags;
 
-       spin_lock_irqsave(&msi_lock, flags);
-       entry = msi_desc[vector];
-       if (!entry || entry->dev != dev) {
-               spin_unlock_irqrestore(&msi_lock, flags);
-               return -EINVAL;
-       }
-       type = entry->msi_attrib.type;
-       entry_nr = entry->msi_attrib.entry_nr;
-       head = entry->link.head;
-       base = entry->mask_base;
-       msi_desc[entry->link.head]->link.tail = entry->link.tail;
-       msi_desc[entry->link.tail]->link.head = entry->link.head;
-       entry->dev = NULL;
-       if (!reassign) {
-               vector_irq[vector] = 0;
-               nr_released_vectors++;
-       }
-       msi_desc[vector] = NULL;
-       spin_unlock_irqrestore(&msi_lock, flags);
-
-       kmem_cache_free(msi_cachep, entry);
-
-       if (type == PCI_CAP_ID_MSIX) {
-               if (!reassign)
-                       writel(1, base +
-                               entry_nr * PCI_MSIX_ENTRY_SIZE +
-                               PCI_MSIX_ENTRY_VECTOR_CTRL_OFFSET);
-
-               if (head == vector) {
-                       /*
-                        * Detect last MSI-X vector to be released.
-                        * Release the MSI-X memory-mapped table.
-                        */
-                       int pos, nr_entries;
-                       u32 phys_addr, table_offset;
-                       u16 control;
-                       u8 bir;
-
-                       pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
-                       pci_read_config_word(dev, msi_control_reg(pos),
-                               &control);
-                       nr_entries = multi_msix_capable(control);
-                       pci_read_config_dword(dev, msix_table_offset_reg(pos),
-                               &table_offset);
-                       bir = (u8)(table_offset & PCI_MSIX_FLAGS_BIRMASK);
-                       phys_addr = pci_resource_start (dev, bir);
-                       phys_addr += (u32)(table_offset &
-                               ~PCI_MSIX_FLAGS_BIRMASK);
-                       iounmap(base);
-               }
-       }
+       if (!pci_msi_enable || !dev || !dev->msi_enabled)
+               return;
 
-       return 0;
+       pci_msi_shutdown(dev);
+
+       entry = list_entry(dev->msi_list.next, struct msi_desc, list);
+       if (!entry->dev || entry->msi_attrib.type != PCI_CAP_ID_MSI)
+               return;
+
+       msi_free_irqs(dev);
 }
+EXPORT_SYMBOL(pci_disable_msi);
 
-static int reroute_msix_table(int head, struct msix_entry *entries, int *nvec)
+static int msi_free_irqs(struct pci_dev* dev)
 {
-       int vector = head, tail = 0;
-       int i, j = 0, nr_entries = 0;
-       void __iomem *base;
-       unsigned long flags;
-
-       spin_lock_irqsave(&msi_lock, flags);
-       while (head != tail) {
-               nr_entries++;
-               tail = msi_desc[vector]->link.tail;
-               if (entries[0].entry == msi_desc[vector]->msi_attrib.entry_nr)
-                       j = vector;
-               vector = tail;
-       }
-       if (*nvec > nr_entries) {
-               spin_unlock_irqrestore(&msi_lock, flags);
-               *nvec = nr_entries;
-               return -EINVAL;
+       struct msi_desc *entry, *tmp;
+
+       list_for_each_entry(entry, &dev->msi_list, list) {
+               if (entry->irq)
+                       BUG_ON(irq_has_action(entry->irq));
        }
-       vector = ((j > 0) ? j : head);
-       for (i = 0; i < *nvec; i++) {
-               j = msi_desc[vector]->msi_attrib.entry_nr;
-               msi_desc[vector]->msi_attrib.state = 0; /* Mark it not active */
-               vector_irq[vector] = -1;                /* Mark it busy */
-               nr_released_vectors--;
-               entries[i].vector = vector;
-               if (j != (entries + i)->entry) {
-                       base = msi_desc[vector]->mask_base;
-                       msi_desc[vector]->msi_attrib.entry_nr =
-                               (entries + i)->entry;
-                       writel( readl(base + j * PCI_MSIX_ENTRY_SIZE +
-                               PCI_MSIX_ENTRY_LOWER_ADDR_OFFSET), base +
-                               (entries + i)->entry * PCI_MSIX_ENTRY_SIZE +
-                               PCI_MSIX_ENTRY_LOWER_ADDR_OFFSET);
-                       writel( readl(base + j * PCI_MSIX_ENTRY_SIZE +
-                               PCI_MSIX_ENTRY_UPPER_ADDR_OFFSET), base +
-                               (entries + i)->entry * PCI_MSIX_ENTRY_SIZE +
-                               PCI_MSIX_ENTRY_UPPER_ADDR_OFFSET);
-                       writel( (readl(base + j * PCI_MSIX_ENTRY_SIZE +
-                               PCI_MSIX_ENTRY_DATA_OFFSET) & 0xff00) | vector,
-                               base + (entries+i)->entry*PCI_MSIX_ENTRY_SIZE +
-                               PCI_MSIX_ENTRY_DATA_OFFSET);
+
+       arch_teardown_msi_irqs(dev);
+
+       list_for_each_entry_safe(entry, tmp, &dev->msi_list, list) {
+               if (entry->msi_attrib.type == PCI_CAP_ID_MSIX) {
+                       writel(1, entry->mask_base + entry->msi_attrib.entry_nr
+                                 * PCI_MSIX_ENTRY_SIZE
+                                 + PCI_MSIX_ENTRY_VECTOR_CTRL_OFFSET);
+
+                       if (list_is_last(&entry->list, &dev->msi_list))
+                               iounmap(entry->mask_base);
                }
-               vector = msi_desc[vector]->link.tail;
+               list_del(&entry->list);
+               kfree(entry);
        }
-       spin_unlock_irqrestore(&msi_lock, flags);
 
        return 0;
 }
@@ -899,37 +652,32 @@ static int reroute_msix_table(int head, struct msix_entry *entries, int *nvec)
  * pci_enable_msix - configure device's MSI-X capability structure
  * @dev: pointer to the pci_dev data structure of MSI-X device function
  * @entries: pointer to an array of MSI-X entries
- * @nvec: number of MSI-X vectors requested for allocation by device driver
+ * @nvec: number of MSI-X irqs requested for allocation by device driver
  *
  * Setup the MSI-X capability structure of device function with the number
- * of requested vectors upon its software driver call to request for
+ * of requested irqs upon its software driver call to request for
  * MSI-X mode enabled on its hardware device function. A return of zero
  * indicates the successful configuration of MSI-X capability structure
- * with new allocated MSI-X vectors. A return of < 0 indicates a failure.
+ * with new allocated MSI-X irqs. A return of < 0 indicates a failure.
  * Or a return of > 0 indicates that driver request is exceeding the number
- * of vectors available. Driver should use the returned value to re-send
+ * of irqs available. Driver should use the returned value to re-send
  * its request.
  **/
 int pci_enable_msix(struct pci_dev* dev, struct msix_entry *entries, int nvec)
 {
-       int status, pos, nr_entries, free_vectors;
-       int i, j, temp;
+       int status, pos, nr_entries;
+       int i, j;
        u16 control;
-       unsigned long flags;
 
-       if (!pci_msi_enable || !dev || !entries)
+       if (!entries)
                return -EINVAL;
 
-       if ((status = msi_init()) < 0)
+       status = pci_msi_check_device(dev, nvec, PCI_CAP_ID_MSIX);
+       if (status)
                return status;
 
-       if (!(pos = pci_find_capability(dev, PCI_CAP_ID_MSIX)))
-               return -EINVAL;
-
+       pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
        pci_read_config_word(dev, msi_control_reg(pos), &control);
-       if (control & PCI_MSIX_FLAGS_ENABLE)
-               return -EINVAL;                 /* Already in MSI-X mode */
-
        nr_entries = multi_msix_capable(control);
        if (nvec > nr_entries)
                return -EINVAL;
@@ -943,183 +691,92 @@ int pci_enable_msix(struct pci_dev* dev, struct msix_entry *entries, int nvec)
                                return -EINVAL; /* duplicate entry */
                }
        }
-       temp = dev->irq;
-       if (!msi_lookup_vector(dev, PCI_CAP_ID_MSIX)) {
-               /* Lookup Sucess */
-               nr_entries = nvec;
-               /* Reroute MSI-X table */
-               if (reroute_msix_table(dev->irq, entries, &nr_entries)) {
-                       /* #requested > #previous-assigned */
-                       dev->irq = temp;
-                       return nr_entries;
-               }
-               dev->irq = temp;
-               enable_msi_mode(dev, pos, PCI_CAP_ID_MSIX);
-               return 0;
-       }
-       /* Check whether driver already requested for MSI vector */
-       if (pci_find_capability(dev, PCI_CAP_ID_MSI) > 0 &&
-               !msi_lookup_vector(dev, PCI_CAP_ID_MSI)) {
-               printk(KERN_INFO "PCI: %s: Can't enable MSI-X.  "
-                      "Device already has an MSI vector assigned\n",
-                      pci_name(dev));
-               dev->irq = temp;
-               return -EINVAL;
-       }
+       WARN_ON(!!dev->msix_enabled);
 
-       spin_lock_irqsave(&msi_lock, flags);
-       /*
-        * msi_lock is provided to ensure that enough vectors resources are
-        * available before granting.
-        */
-       free_vectors = pci_vector_resources(last_alloc_vector,
-                               nr_released_vectors);
-       /* Ensure that each MSI/MSI-X device has one vector reserved by
-          default to avoid any MSI-X driver to take all available
-          resources */
-       free_vectors -= nr_reserved_vectors;
-       /* Find the average of free vectors among MSI-X devices */
-       if (nr_msix_devices > 0)
-               free_vectors /= nr_msix_devices;
-       spin_unlock_irqrestore(&msi_lock, flags);
-
-       if (nvec > free_vectors) {
-               if (free_vectors > 0)
-                       return free_vectors;
-               else
-                       return -EBUSY;
+       /* Check whether driver already requested for MSI irq */
+       if (dev->msi_enabled) {
+               dev_info(&dev->dev, "can't enable MSI-X "
+                      "(MSI IRQ already assigned)\n");
+               return -EINVAL;
        }
-
        status = msix_capability_init(dev, entries, nvec);
-       if (!status && nr_msix_devices > 0)
-               nr_msix_devices--;
-
        return status;
 }
+EXPORT_SYMBOL(pci_enable_msix);
 
-void pci_disable_msix(struct pci_dev* dev)
+static void msix_free_all_irqs(struct pci_dev *dev)
 {
-       int pos, temp;
-       u16 control;
+       msi_free_irqs(dev);
+}
 
-       if (!dev || !(pos = pci_find_capability(dev, PCI_CAP_ID_MSIX)))
+void pci_msix_shutdown(struct pci_dev* dev)
+{
+       if (!pci_msi_enable || !dev || !dev->msix_enabled)
                return;
 
-       pci_read_config_word(dev, msi_control_reg(pos), &control);
-       if (!(control & PCI_MSIX_FLAGS_ENABLE))
+       msix_set_enable(dev, 0);
+       pci_intx_for_msi(dev, 1);
+       dev->msix_enabled = 0;
+}
+void pci_disable_msix(struct pci_dev* dev)
+{
+       if (!pci_msi_enable || !dev || !dev->msix_enabled)
                return;
 
-       temp = dev->irq;
-       if (!msi_lookup_vector(dev, PCI_CAP_ID_MSIX)) {
-               int state, vector, head, tail = 0, warning = 0;
-               unsigned long flags;
-
-               vector = head = dev->irq;
-               spin_lock_irqsave(&msi_lock, flags);
-               while (head != tail) {
-                       state = msi_desc[vector]->msi_attrib.state;
-                       if (state)
-                               warning = 1;
-                       else {
-                               vector_irq[vector] = 0; /* free it */
-                               nr_released_vectors++;
-                       }
-                       tail = msi_desc[vector]->link.tail;
-                       vector = tail;
-               }
-               spin_unlock_irqrestore(&msi_lock, flags);
-               if (warning) {
-                       dev->irq = temp;
-                       printk(KERN_WARNING "PCI: %s: pci_disable_msix() called without "
-                              "free_irq() on all MSI-X vectors\n",
-                              pci_name(dev));
-                       BUG_ON(warning > 0);
-               } else {
-                       dev->irq = temp;
-                       disable_msi_mode(dev,
-                               pci_find_capability(dev, PCI_CAP_ID_MSIX),
-                               PCI_CAP_ID_MSIX);
+       pci_msix_shutdown(dev);
 
-               }
-       }
+       msix_free_all_irqs(dev);
 }
+EXPORT_SYMBOL(pci_disable_msix);
 
 /**
- * msi_remove_pci_irq_vectors - reclaim MSI(X) vectors to unused state
+ * msi_remove_pci_irq_vectors - reclaim MSI(X) irqs to unused state
  * @dev: pointer to the pci_dev data structure of MSI(X) device function
  *
  * Being called during hotplug remove, from which the device function
- * is hot-removed. All previous assigned MSI/MSI-X vectors, if
+ * is hot-removed. All previous assigned MSI/MSI-X irqs, if
  * allocated for this device function, are reclaimed to unused state,
  * which may be used later on.
  **/
 void msi_remove_pci_irq_vectors(struct pci_dev* dev)
 {
-       int state, pos, temp;
-       unsigned long flags;
-
        if (!pci_msi_enable || !dev)
                return;
 
-       temp = dev->irq;                /* Save IOAPIC IRQ */
-       if ((pos = pci_find_capability(dev, PCI_CAP_ID_MSI)) > 0 &&
-               !msi_lookup_vector(dev, PCI_CAP_ID_MSI)) {
-               spin_lock_irqsave(&msi_lock, flags);
-               state = msi_desc[dev->irq]->msi_attrib.state;
-               spin_unlock_irqrestore(&msi_lock, flags);
-               if (state) {
-                       printk(KERN_WARNING "PCI: %s: msi_remove_pci_irq_vectors() "
-                              "called without free_irq() on MSI vector %d\n",
-                              pci_name(dev), dev->irq);
-                       BUG_ON(state > 0);
-               } else /* Release MSI vector assigned to this device */
-                       msi_free_vector(dev, dev->irq, 0);
-               dev->irq = temp;                /* Restore IOAPIC IRQ */
-       }
-       if ((pos = pci_find_capability(dev, PCI_CAP_ID_MSIX)) > 0 &&
-               !msi_lookup_vector(dev, PCI_CAP_ID_MSIX)) {
-               int vector, head, tail = 0, warning = 0;
-               void __iomem *base = NULL;
-
-               vector = head = dev->irq;
-               while (head != tail) {
-                       spin_lock_irqsave(&msi_lock, flags);
-                       state = msi_desc[vector]->msi_attrib.state;
-                       tail = msi_desc[vector]->link.tail;
-                       base = msi_desc[vector]->mask_base;
-                       spin_unlock_irqrestore(&msi_lock, flags);
-                       if (state)
-                               warning = 1;
-                       else if (vector != head) /* Release MSI-X vector */
-                               msi_free_vector(dev, vector, 0);
-                       vector = tail;
-               }
-               msi_free_vector(dev, vector, 0);
-               if (warning) {
-                       /* Force to release the MSI-X memory-mapped table */
-                       u32 phys_addr, table_offset;
-                       u16 control;
-                       u8 bir;
-
-                       pci_read_config_word(dev, msi_control_reg(pos),
-                               &control);
-                       pci_read_config_dword(dev, msix_table_offset_reg(pos),
-                               &table_offset);
-                       bir = (u8)(table_offset & PCI_MSIX_FLAGS_BIRMASK);
-                       phys_addr = pci_resource_start (dev, bir);
-                       phys_addr += (u32)(table_offset &
-                               ~PCI_MSIX_FLAGS_BIRMASK);
-                       iounmap(base);
-                       printk(KERN_WARNING "PCI: %s: msi_remove_pci_irq_vectors() "
-                              "called without free_irq() on all MSI-X vectors\n",
-                              pci_name(dev));
-                       BUG_ON(warning > 0);
-               }
-               dev->irq = temp;                /* Restore IOAPIC IRQ */
-       }
+       if (dev->msi_enabled)
+               msi_free_irqs(dev);
+
+       if (dev->msix_enabled)
+               msix_free_all_irqs(dev);
 }
 
-EXPORT_SYMBOL(pci_enable_msi);
-EXPORT_SYMBOL(pci_disable_msi);
-EXPORT_SYMBOL(pci_enable_msix);
-EXPORT_SYMBOL(pci_disable_msix);
+void pci_no_msi(void)
+{
+       pci_msi_enable = 0;
+}
+
+void pci_msi_init_pci_dev(struct pci_dev *dev)
+{
+       INIT_LIST_HEAD(&dev->msi_list);
+}
+
+#ifdef CONFIG_ACPI
+#include <linux/acpi.h>
+#include <linux/pci-acpi.h>
+static void __devinit msi_acpi_init(void)
+{
+       if (acpi_pci_disabled)
+               return;
+       pci_osc_support_set(OSC_MSI_SUPPORT);
+       pcie_osc_support_set(OSC_MSI_SUPPORT);
+}
+#else
+static inline void msi_acpi_init(void) { }
+#endif /* CONFIG_ACPI */
+
+void __devinit msi_init(void)
+{
+       if (!pci_msi_enable)
+               return;
+       msi_acpi_init();
+}