[deliverable/linux.git] / arch / x86 / pci / xen.c

/*
 * Xen PCI - handle PCI (INTx) and MSI infrastructure calls for PV, HVM and
 * initial domain support. We also handle the DSDT _PRT callbacks for GSI's
 * used in HVM and initial domain mode (PV does not parse ACPI, so it has no
 * concept of GSIs). Under PV we hook under the pnbbios API for IRQs and
 * 0xcf8 PCI configuration read/write.
 *
 *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
 *           Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
 *           Stefano Stabellini <stefano.stabellini@eu.citrix.com>
 */
#include <linux/module.h>
#include <linux/init.h>
#include <linux/pci.h>
#include <linux/acpi.h>

#include <linux/io.h>
#include <asm/io_apic.h>
#include <asm/pci_x86.h>

#include <asm/xen/hypervisor.h>

#include <xen/features.h>
#include <xen/events.h>
#include <asm/xen/pci.h>

static int xen_pcifront_enable_irq(struct pci_dev *dev)
{
	int rc;
	int share = 1;
	int pirq;
	u8 gsi;

	rc = pci_read_config_byte(dev, PCI_INTERRUPT_LINE, &gsi);
	if (rc < 0) {
		dev_warn(&dev->dev, "Xen PCI: failed to read interrupt line: %d\n",
			 rc);
		return rc;
	}

	rc = xen_allocate_pirq_gsi(gsi);
	if (rc < 0) {
		dev_warn(&dev->dev, "Xen PCI: failed to allocate a PIRQ for GSI%d: %d\n",
			 gsi, rc);
		return rc;
	}
	pirq = rc;

	if (gsi < NR_IRQS_LEGACY)
		share = 0;

	rc = xen_bind_pirq_gsi_to_irq(gsi, pirq, share, "pcifront");
	if (rc < 0) {
		dev_warn(&dev->dev, "Xen PCI: failed to bind GSI%d (PIRQ%d) to IRQ: %d\n",
			 gsi, pirq, rc);
		return rc;
	}

	dev->irq = rc;
	dev_info(&dev->dev, "Xen PCI mapped GSI%d to IRQ%d\n", gsi, dev->irq);
	return 0;
}

#ifdef CONFIG_ACPI
static int xen_register_pirq(u32 gsi, int gsi_override, int triggering,
			     bool alloc_pirq)
{
	int rc, pirq = -1, irq = -1;
	struct physdev_map_pirq map_irq;
	int shareable = 0;
	char *name;

	if (alloc_pirq) {
		pirq = xen_allocate_pirq_gsi(gsi);
		if (pirq < 0)
			goto out;
	}
	map_irq.domid = DOMID_SELF;
	map_irq.type = MAP_PIRQ_TYPE_GSI;
	map_irq.index = gsi;
	map_irq.pirq = pirq;

	rc = HYPERVISOR_physdev_op(PHYSDEVOP_map_pirq, &map_irq);
	if (rc) {
		printk(KERN_WARNING "xen map irq failed %d\n", rc);
		return -1;
	}

	if (triggering == ACPI_EDGE_SENSITIVE) {
		shareable = 0;
		name = "ioapic-edge";
	} else {
		shareable = 1;
		name = "ioapic-level";
	}

	if (gsi_override >= 0)
		gsi = gsi_override;

	irq = xen_bind_pirq_gsi_to_irq(gsi, map_irq.pirq, shareable, name);
	if (irq < 0)
		goto out;

	printk(KERN_DEBUG "xen: --> pirq=%d -> irq=%d (gsi=%d)\n", map_irq.pirq, irq, gsi);
out:
	return irq;
}

static int acpi_register_gsi_xen_hvm(struct device *dev, u32 gsi,
				     int trigger, int polarity)
{
	if (!xen_hvm_domain())
		return -1;

	return xen_register_pirq(gsi, -1 /* no GSI override */,
				 trigger, false /* no PIRQ allocation */);
}

#ifdef CONFIG_XEN_DOM0
static int xen_register_gsi(u32 gsi, int gsi_override, int triggering, int polarity)
{
	int rc, irq;
	struct physdev_setup_gsi setup_gsi;

	if (!xen_pv_domain())
		return -1;

	printk(KERN_DEBUG "xen: registering gsi %u triggering %d polarity %d\n",
			gsi, triggering, polarity);

	irq = xen_register_pirq(gsi, gsi_override, triggering, true);

	setup_gsi.gsi = gsi;
	setup_gsi.triggering = (triggering == ACPI_EDGE_SENSITIVE ? 0 : 1);
	setup_gsi.polarity = (polarity == ACPI_ACTIVE_HIGH ? 0 : 1);

	rc = HYPERVISOR_physdev_op(PHYSDEVOP_setup_gsi, &setup_gsi);
	if (rc == -EEXIST)
		printk(KERN_INFO "Already setup the GSI :%d\n", gsi);
	else if (rc) {
		printk(KERN_ERR "Failed to setup GSI :%d, err_code:%d\n",
				gsi, rc);
	}

	return irq;
}

static int acpi_register_gsi_xen(struct device *dev, u32 gsi,
				 int trigger, int polarity)
{
	return xen_register_gsi(gsi, -1 /* no GSI override */, trigger, polarity);
}
#endif
#endif

#if defined(CONFIG_PCI_MSI)
#include <linux/msi.h>
#include <asm/msidef.h>

struct xen_pci_frontend_ops *xen_pci_frontend;
EXPORT_SYMBOL_GPL(xen_pci_frontend);

static int xen_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
{
	int irq, ret, i;
	struct msi_desc *msidesc;
	int *v;

	v = kzalloc(sizeof(int) * max(1, nvec), GFP_KERNEL);
	if (!v)
		return -ENOMEM;

	if (type == PCI_CAP_ID_MSIX)
		ret = xen_pci_frontend_enable_msix(dev, v, nvec);
	else
		ret = xen_pci_frontend_enable_msi(dev, v);
	if (ret)
		goto error;
	i = 0;
	list_for_each_entry(msidesc, &dev->msi_list, list) {
		irq = xen_bind_pirq_msi_to_irq(dev, msidesc, v[i], 0,
					       (type == PCI_CAP_ID_MSIX) ?
					       "pcifront-msi-x" :
					       "pcifront-msi",
						DOMID_SELF);
		if (irq < 0)
			goto free;
		i++;
	}
	kfree(v);
	return 0;

error:
	dev_err(&dev->dev, "Xen PCI frontend has not registered MSI/MSI-X support!\n");
free:
	kfree(v);
	return ret;
}

#define XEN_PIRQ_MSI_DATA  (MSI_DATA_TRIGGER_EDGE | \
		MSI_DATA_LEVEL_ASSERT | (3 << 8) | MSI_DATA_VECTOR(0))

static void xen_msi_compose_msg(struct pci_dev *pdev, unsigned int pirq,
		struct msi_msg *msg)
{
	/* We set vector == 0 to tell the hypervisor we don't care about it,
	 * but we want a pirq setup instead.
	 * We use the dest_id field to pass the pirq that we want. */
	msg->address_hi = MSI_ADDR_BASE_HI | MSI_ADDR_EXT_DEST_ID(pirq);
	msg->address_lo =
		MSI_ADDR_BASE_LO |
		MSI_ADDR_DEST_MODE_PHYSICAL |
		MSI_ADDR_REDIRECTION_CPU |
		MSI_ADDR_DEST_ID(pirq);

	msg->data = XEN_PIRQ_MSI_DATA;
}

static int xen_hvm_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
{
	int irq, pirq;
	struct msi_desc *msidesc;
	struct msi_msg msg;

	list_for_each_entry(msidesc, &dev->msi_list, list) {
		__read_msi_msg(msidesc, &msg);
		pirq = MSI_ADDR_EXT_DEST_ID(msg.address_hi) |
			((msg.address_lo >> MSI_ADDR_DEST_ID_SHIFT) & 0xff);
		if (msg.data != XEN_PIRQ_MSI_DATA ||
		    xen_irq_from_pirq(pirq) < 0) {
			pirq = xen_allocate_pirq_msi(dev, msidesc);
			if (pirq < 0)
				goto error;
			xen_msi_compose_msg(dev, pirq, &msg);
			__write_msi_msg(msidesc, &msg);
			dev_dbg(&dev->dev, "xen: msi bound to pirq=%d\n", pirq);
		} else {
			dev_dbg(&dev->dev,
				"xen: msi already bound to pirq=%d\n", pirq);
		}
		irq = xen_bind_pirq_msi_to_irq(dev, msidesc, pirq, 0,
					       (type == PCI_CAP_ID_MSIX) ?
					       "msi-x" : "msi",
					       DOMID_SELF);
		if (irq < 0)
			goto error;
		dev_dbg(&dev->dev,
			"xen: msi --> pirq=%d --> irq=%d\n", pirq, irq);
	}
	return 0;

error:
	dev_err(&dev->dev,
		"Xen PCI frontend has not registered MSI/MSI-X support!\n");
	return -ENODEV;
}

#ifdef CONFIG_XEN_DOM0
static int xen_initdom_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
{
	int ret = 0;
	struct msi_desc *msidesc;

	list_for_each_entry(msidesc, &dev->msi_list, list) {
		struct physdev_map_pirq map_irq;
		domid_t domid;

		domid = ret = xen_find_device_domain_owner(dev);
		/* N.B. Casting int's -ENODEV to uint16_t results in 0xFFED,
		 * hence check ret value for < 0. */
		if (ret < 0)
			domid = DOMID_SELF;

		memset(&map_irq, 0, sizeof(map_irq));
		map_irq.domid = domid;
		map_irq.type = MAP_PIRQ_TYPE_MSI;
		map_irq.index = -1;
		map_irq.pirq = -1;
		map_irq.bus = dev->bus->number;
		map_irq.devfn = dev->devfn;

		if (type == PCI_CAP_ID_MSIX) {
			int pos;
			u32 table_offset, bir;

			pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);

			pci_read_config_dword(dev, pos + PCI_MSIX_TABLE,
					      &table_offset);
			bir = (u8)(table_offset & PCI_MSIX_FLAGS_BIRMASK);

			map_irq.table_base = pci_resource_start(dev, bir);
			map_irq.entry_nr = msidesc->msi_attrib.entry_nr;
		}

		ret = HYPERVISOR_physdev_op(PHYSDEVOP_map_pirq, &map_irq);
		if (ret) {
			dev_warn(&dev->dev, "xen map irq failed %d for %d domain\n",
				 ret, domid);
			goto out;
		}

		ret = xen_bind_pirq_msi_to_irq(dev, msidesc,
					       map_irq.pirq, map_irq.index,
					       (type == PCI_CAP_ID_MSIX) ?
					       "msi-x" : "msi",
						domid);
		if (ret < 0)
			goto out;
	}
	ret = 0;
out:
	return ret;
}
#endif

static void xen_teardown_msi_irqs(struct pci_dev *dev)
{
	struct msi_desc *msidesc;

	msidesc = list_entry(dev->msi_list.next, struct msi_desc, list);
	if (msidesc->msi_attrib.is_msix)
		xen_pci_frontend_disable_msix(dev);
	else
		xen_pci_frontend_disable_msi(dev);

	/* Free the IRQ's and the msidesc using the generic code. */
	default_teardown_msi_irqs(dev);
}

static void xen_teardown_msi_irq(unsigned int irq)
{
	xen_destroy_irq(irq);
}

#endif

int __init pci_xen_init(void)
{
	if (!xen_pv_domain() || xen_initial_domain())
		return -ENODEV;

	printk(KERN_INFO "PCI: setting up Xen PCI frontend stub\n");

	pcibios_set_cache_line_size();

	pcibios_enable_irq = xen_pcifront_enable_irq;
	pcibios_disable_irq = NULL;

#ifdef CONFIG_ACPI
	/* Keep ACPI out of the picture */
	acpi_noirq = 1;
#endif

#ifdef CONFIG_PCI_MSI
	x86_msi.setup_msi_irqs = xen_setup_msi_irqs;
	x86_msi.teardown_msi_irq = xen_teardown_msi_irq;
	x86_msi.teardown_msi_irqs = xen_teardown_msi_irqs;
#endif
	return 0;
}

int __init pci_xen_hvm_init(void)
{
	if (!xen_feature(XENFEAT_hvm_pirqs))
		return 0;

#ifdef CONFIG_ACPI
	/*
	 * We don't want to change the actual ACPI delivery model,
	 * just how GSIs get registered.
	 */
	__acpi_register_gsi = acpi_register_gsi_xen_hvm;
#endif

#ifdef CONFIG_PCI_MSI
	x86_msi.setup_msi_irqs = xen_hvm_setup_msi_irqs;
	x86_msi.teardown_msi_irq = xen_teardown_msi_irq;
#endif
	return 0;
}

#ifdef CONFIG_XEN_DOM0
#ifdef CONFIG_ACPI
static __init void xen_setup_acpi_sci(void)
{
	int rc;
	int trigger, polarity;
	int gsi = acpi_sci_override_gsi;
	int irq = -1;
	int gsi_override = -1;

	if (!gsi)
		return;

	rc = acpi_get_override_irq(gsi, &trigger, &polarity);
	if (rc) {
		printk(KERN_WARNING "xen: acpi_get_override_irq failed for acpi"
				" sci, rc=%d\n", rc);
		return;
	}
	trigger = trigger ? ACPI_LEVEL_SENSITIVE : ACPI_EDGE_SENSITIVE;
	polarity = polarity ? ACPI_ACTIVE_LOW : ACPI_ACTIVE_HIGH;

	printk(KERN_INFO "xen: sci override: global_irq=%d trigger=%d "
			"polarity=%d\n", gsi, trigger, polarity);

	/* Before we bind the GSI to a Linux IRQ, check whether
	 * we need to override it with bus_irq (IRQ) value. Usually for
	 * IRQs below IRQ_LEGACY_IRQ this holds IRQ == GSI, as so:
	 *  ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 low level)
	 * but there are oddballs where the IRQ != GSI:
	 *  ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 20 low level)
	 * which ends up being: gsi_to_irq[9] == 20
	 * (which is what acpi_gsi_to_irq ends up calling when starting the
	 * the ACPI interpreter and keels over since IRQ 9 has not been
	 * setup as we had setup IRQ 20 for it).
	 */
	/* Check whether the GSI != IRQ */
	if (acpi_gsi_to_irq(gsi, &irq) == 0) {
		if (irq >= 0 && irq != gsi)
			/* Bugger, we MUST have that IRQ. */
			gsi_override = irq;
	}

	gsi = xen_register_gsi(gsi, gsi_override, trigger, polarity);
	printk(KERN_INFO "xen: acpi sci %d\n", gsi);

	return;
}
#endif
static int __init pci_xen_initial_domain(void)
{
#ifdef CONFIG_PCI_MSI
	x86_msi.setup_msi_irqs = xen_initdom_setup_msi_irqs;
	x86_msi.teardown_msi_irq = xen_teardown_msi_irq;
#endif
#ifdef CONFIG_ACPI
	xen_setup_acpi_sci();
	__acpi_register_gsi = acpi_register_gsi_xen;
#endif
	return 0;
}

void __init xen_setup_pirqs(void)
{
	int pirq, irq;

	pci_xen_initial_domain();

	if (0 == nr_ioapics) {
		for (irq = 0; irq < NR_IRQS_LEGACY; irq++) {
			pirq = xen_allocate_pirq_gsi(irq);
			if (WARN(pirq < 0,
				 "Could not allocate PIRQ for legacy interrupt\n"))
				break;
			irq = xen_bind_pirq_gsi_to_irq(irq, pirq, 0, "xt-pic");
		}
		return;
	}
#ifdef CONFIG_ACPI
	/* Pre-allocate legacy irqs */
	for (irq = 0; irq < NR_IRQS_LEGACY; irq++) {
		int trigger, polarity;

		if (acpi_get_override_irq(irq, &trigger, &polarity) == -1)
			continue;

		xen_register_pirq(irq, -1 /* no GSI override */,
			trigger ? ACPI_LEVEL_SENSITIVE : ACPI_EDGE_SENSITIVE,
			true /* allocate IRQ */);
	}
#endif
}

struct xen_device_domain_owner {
	domid_t domain;
	struct pci_dev *dev;
	struct list_head list;
};

static DEFINE_SPINLOCK(dev_domain_list_spinlock);
static struct list_head dev_domain_list = LIST_HEAD_INIT(dev_domain_list);

static struct xen_device_domain_owner *find_device(struct pci_dev *dev)
{
	struct xen_device_domain_owner *owner;

	list_for_each_entry(owner, &dev_domain_list, list) {
		if (owner->dev == dev)
			return owner;
	}
	return NULL;
}

int xen_find_device_domain_owner(struct pci_dev *dev)
{
	struct xen_device_domain_owner *owner;
	int domain = -ENODEV;

	spin_lock(&dev_domain_list_spinlock);
	owner = find_device(dev);
	if (owner)
		domain = owner->domain;
	spin_unlock(&dev_domain_list_spinlock);
	return domain;
}
EXPORT_SYMBOL_GPL(xen_find_device_domain_owner);

int xen_register_device_domain_owner(struct pci_dev *dev, uint16_t domain)
{
	struct xen_device_domain_owner *owner;

	owner = kzalloc(sizeof(struct xen_device_domain_owner), GFP_KERNEL);
	if (!owner)
		return -ENODEV;

	spin_lock(&dev_domain_list_spinlock);
	if (find_device(dev)) {
		spin_unlock(&dev_domain_list_spinlock);
		kfree(owner);
		return -EEXIST;
	}
	owner->domain = domain;
	owner->dev = dev;
	list_add_tail(&owner->list, &dev_domain_list);
	spin_unlock(&dev_domain_list_spinlock);
	return 0;
}
EXPORT_SYMBOL_GPL(xen_register_device_domain_owner);

int xen_unregister_device_domain_owner(struct pci_dev *dev)
{
	struct xen_device_domain_owner *owner;

	spin_lock(&dev_domain_list_spinlock);
	owner = find_device(dev);
	if (!owner) {
		spin_unlock(&dev_domain_list_spinlock);
		return -ENODEV;
	}
	list_del(&owner->list);
	spin_unlock(&dev_domain_list_spinlock);
	kfree(owner);
	return 0;
}
EXPORT_SYMBOL_GPL(xen_unregister_device_domain_owner);
#endif
Commit	Line	Data
b5401a96	1	/*
996c34ae KRW	2	* Xen PCI - handle PCI (INTx) and MSI infrastructure calls for PV, HVM and
	3	* initial domain support. We also handle the DSDT _PRT callbacks for GSI's
	4	* used in HVM and initial domain mode (PV does not parse ACPI, so it has no
	5	* concept of GSIs). Under PV we hook under the pnbbios API for IRQs and
	6	* 0xcf8 PCI configuration read/write.
b5401a96 AN	7	*
b5401a96 AN	8	* Author: Ryan Wilson <hap9@epoch.ncsc.mil>
996c34ae KRW	9	* Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
996c34ae KRW	10	* Stefano Stabellini <stefano.stabellini@eu.citrix.com>
b5401a96 AN	11	*/
	12	#include <linux/module.h>
	13	#include <linux/init.h>
	14	#include <linux/pci.h>
	15	#include <linux/acpi.h>
	16
	17	#include <linux/io.h>
0e058e52	18	#include <asm/io_apic.h>
b5401a96 AN	19	#include <asm/pci_x86.h>
	20
	21	#include <asm/xen/hypervisor.h>
	22
3942b740	23	#include <xen/features.h>
b5401a96 AN	24	#include <xen/events.h>
	25	#include <asm/xen/pci.h>
	26
fef6e262 KRW	27	static int xen_pcifront_enable_irq(struct pci_dev *dev)
	28	{
	29	int rc;
	30	int share = 1;
	31	int pirq;
	32	u8 gsi;
	33
	34	rc = pci_read_config_byte(dev, PCI_INTERRUPT_LINE, &gsi);
	35	if (rc < 0) {
	36	dev_warn(&dev->dev, "Xen PCI: failed to read interrupt line: %d\n",
	37	rc);
	38	return rc;
	39	}
	40
	41	rc = xen_allocate_pirq_gsi(gsi);
	42	if (rc < 0) {
	43	dev_warn(&dev->dev, "Xen PCI: failed to allocate a PIRQ for GSI%d: %d\n",
	44	gsi, rc);
	45	return rc;
	46	}
	47	pirq = rc;
	48
	49	if (gsi < NR_IRQS_LEGACY)
	50	share = 0;
	51
	52	rc = xen_bind_pirq_gsi_to_irq(gsi, pirq, share, "pcifront");
	53	if (rc < 0) {
	54	dev_warn(&dev->dev, "Xen PCI: failed to bind GSI%d (PIRQ%d) to IRQ: %d\n",
	55	gsi, pirq, rc);
	56	return rc;
	57	}
	58
	59	dev->irq = rc;
	60	dev_info(&dev->dev, "Xen PCI mapped GSI%d to IRQ%d\n", gsi, dev->irq);
	61	return 0;
	62	}
	63
42a1de56	64	#ifdef CONFIG_ACPI
ed89eb63 KRW	65	static int xen_register_pirq(u32 gsi, int gsi_override, int triggering,
ed89eb63 KRW	66	bool alloc_pirq)
42a1de56	67	{
ed89eb63	68	int rc, pirq = -1, irq = -1;
42a1de56 SS	69	struct physdev_map_pirq map_irq;
	70	int shareable = 0;
	71	char *name;
	72
ed89eb63 KRW	73	if (alloc_pirq) {
	74	pirq = xen_allocate_pirq_gsi(gsi);
	75	if (pirq < 0)
	76	goto out;
42a1de56	77	}
fef6e262 KRW	78	map_irq.domid = DOMID_SELF;
	79	map_irq.type = MAP_PIRQ_TYPE_GSI;
	80	map_irq.index = gsi;
	81	map_irq.pirq = pirq;
	82
	83	rc = HYPERVISOR_physdev_op(PHYSDEVOP_map_pirq, &map_irq);
	84	if (rc) {
	85	printk(KERN_WARNING "xen map irq failed %d\n", rc);
	86	return -1;
	87	}
	88
30bd35ed KRW	89	if (triggering == ACPI_EDGE_SENSITIVE) {
	90	shareable = 0;
	91	name = "ioapic-edge";
	92	} else {
	93	shareable = 1;
	94	name = "ioapic-level";
	95	}
	96
	97	if (gsi_override >= 0)
	98	gsi = gsi_override;
	99
ed89eb63	100	irq = xen_bind_pirq_gsi_to_irq(gsi, map_irq.pirq, shareable, name);
30bd35ed KRW	101	if (irq < 0)
	102	goto out;
	103
ed89eb63	104	printk(KERN_DEBUG "xen: --> pirq=%d -> irq=%d (gsi=%d)\n", map_irq.pirq, irq, gsi);
fef6e262 KRW	105	out:
	106	return irq;
	107	}
	108
ed89eb63 KRW	109	static int acpi_register_gsi_xen_hvm(struct device *dev, u32 gsi,
	110	int trigger, int polarity)
	111	{
	112	if (!xen_hvm_domain())
	113	return -1;
	114
	115	return xen_register_pirq(gsi, -1 /* no GSI override */,
	116	trigger, false /* no PIRQ allocation */);
	117	}
	118
	119	#ifdef CONFIG_XEN_DOM0
fef6e262 KRW	120	static int xen_register_gsi(u32 gsi, int gsi_override, int triggering, int polarity)
	121	{
	122	int rc, irq;
	123	struct physdev_setup_gsi setup_gsi;
	124
	125	if (!xen_pv_domain())
	126	return -1;
	127
	128	printk(KERN_DEBUG "xen: registering gsi %u triggering %d polarity %d\n",
	129	gsi, triggering, polarity);
	130
ed89eb63	131	irq = xen_register_pirq(gsi, gsi_override, triggering, true);
fef6e262 KRW	132
	133	setup_gsi.gsi = gsi;
	134	setup_gsi.triggering = (triggering == ACPI_EDGE_SENSITIVE ? 0 : 1);
	135	setup_gsi.polarity = (polarity == ACPI_ACTIVE_HIGH ? 0 : 1);
	136
	137	rc = HYPERVISOR_physdev_op(PHYSDEVOP_setup_gsi, &setup_gsi);
	138	if (rc == -EEXIST)
	139	printk(KERN_INFO "Already setup the GSI :%d\n", gsi);
	140	else if (rc) {
	141	printk(KERN_ERR "Failed to setup GSI :%d, err_code:%d\n",
	142	gsi, rc);
	143	}
	144
	145	return irq;
	146	}
	147
	148	static int acpi_register_gsi_xen(struct device *dev, u32 gsi,
	149	int trigger, int polarity)
	150	{
	151	return xen_register_gsi(gsi, -1 /* no GSI override */, trigger, polarity);
	152	}
	153	#endif
d92edd81	154	#endif
fef6e262	155
b5401a96 AN	156	#if defined(CONFIG_PCI_MSI)
b5401a96 AN	157	#include <linux/msi.h>
809f9267	158	#include <asm/msidef.h>
b5401a96 AN	159
	160	struct xen_pci_frontend_ops *xen_pci_frontend;
	161	EXPORT_SYMBOL_GPL(xen_pci_frontend);
	162
fef6e262 KRW	163	static int xen_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
	164	{
	165	int irq, ret, i;
	166	struct msi_desc *msidesc;
	167	int *v;
	168
	169	v = kzalloc(sizeof(int) * max(1, nvec), GFP_KERNEL);
	170	if (!v)
	171	return -ENOMEM;
	172
	173	if (type == PCI_CAP_ID_MSIX)
	174	ret = xen_pci_frontend_enable_msix(dev, v, nvec);
	175	else
	176	ret = xen_pci_frontend_enable_msi(dev, v);
	177	if (ret)
	178	goto error;
	179	i = 0;
	180	list_for_each_entry(msidesc, &dev->msi_list, list) {
	181	irq = xen_bind_pirq_msi_to_irq(dev, msidesc, v[i], 0,
	182	(type == PCI_CAP_ID_MSIX) ?
	183	"pcifront-msi-x" :
	184	"pcifront-msi",
	185	DOMID_SELF);
	186	if (irq < 0)
	187	goto free;
	188	i++;
	189	}
	190	kfree(v);
	191	return 0;
	192
	193	error:
	194	dev_err(&dev->dev, "Xen PCI frontend has not registered MSI/MSI-X support!\n");
	195	free:
	196	kfree(v);
	197	return ret;
	198	}
	199
af42b8d1 SS	200	#define XEN_PIRQ_MSI_DATA (MSI_DATA_TRIGGER_EDGE \| \
	201	MSI_DATA_LEVEL_ASSERT \| (3 << 8) \| MSI_DATA_VECTOR(0))
	202
809f9267 SS	203	static void xen_msi_compose_msg(struct pci_dev *pdev, unsigned int pirq,
	204	struct msi_msg *msg)
	205	{
	206	/* We set vector == 0 to tell the hypervisor we don't care about it,
	207	* but we want a pirq setup instead.
	208	* We use the dest_id field to pass the pirq that we want. */
	209	msg->address_hi = MSI_ADDR_BASE_HI \| MSI_ADDR_EXT_DEST_ID(pirq);
	210	msg->address_lo =
	211	MSI_ADDR_BASE_LO \|
	212	MSI_ADDR_DEST_MODE_PHYSICAL \|
	213	MSI_ADDR_REDIRECTION_CPU \|
	214	MSI_ADDR_DEST_ID(pirq);
	215
af42b8d1	216	msg->data = XEN_PIRQ_MSI_DATA;
809f9267 SS	217	}
	218
	219	static int xen_hvm_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
	220	{
bf480d95	221	int irq, pirq;
809f9267 SS	222	struct msi_desc *msidesc;
	223	struct msi_msg msg;
	224
	225	list_for_each_entry(msidesc, &dev->msi_list, list) {
af42b8d1 SS	226	__read_msi_msg(msidesc, &msg);
	227	pirq = MSI_ADDR_EXT_DEST_ID(msg.address_hi) \|
	228	((msg.address_lo >> MSI_ADDR_DEST_ID_SHIFT) & 0xff);
bf480d95 IC	229	if (msg.data != XEN_PIRQ_MSI_DATA \|\|
	230	xen_irq_from_pirq(pirq) < 0) {
	231	pirq = xen_allocate_pirq_msi(dev, msidesc);
	232	if (pirq < 0)
af42b8d1	233	goto error;
bf480d95 IC	234	xen_msi_compose_msg(dev, pirq, &msg);
	235	__write_msi_msg(msidesc, &msg);
	236	dev_dbg(&dev->dev, "xen: msi bound to pirq=%d\n", pirq);
	237	} else {
	238	dev_dbg(&dev->dev,
	239	"xen: msi already bound to pirq=%d\n", pirq);
af42b8d1	240	}
ca1d8fe9	241	irq = xen_bind_pirq_msi_to_irq(dev, msidesc, pirq, 0,
bf480d95	242	(type == PCI_CAP_ID_MSIX) ?
beafbdc1 KRW	243	"msi-x" : "msi",
beafbdc1 KRW	244	DOMID_SELF);
bf480d95	245	if (irq < 0)
809f9267	246	goto error;
bf480d95 IC	247	dev_dbg(&dev->dev,
bf480d95 IC	248	"xen: msi --> pirq=%d --> irq=%d\n", pirq, irq);
809f9267 SS	249	}
	250	return 0;
	251
809f9267	252	error:
bf480d95 IC	253	dev_err(&dev->dev,
	254	"Xen PCI frontend has not registered MSI/MSI-X support!\n");
	255	return -ENODEV;
809f9267 SS	256	}
809f9267 SS	257
260a7d4c	258	#ifdef CONFIG_XEN_DOM0
f731e3ef QH	259	static int xen_initdom_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
f731e3ef QH	260	{
71eef7d1	261	int ret = 0;
f731e3ef QH	262	struct msi_desc *msidesc;
	263
	264	list_for_each_entry(msidesc, &dev->msi_list, list) {
71eef7d1	265	struct physdev_map_pirq map_irq;
beafbdc1 KRW	266	domid_t domid;
	267
	268	domid = ret = xen_find_device_domain_owner(dev);
	269	/* N.B. Casting int's -ENODEV to uint16_t results in 0xFFED,
	270	* hence check ret value for < 0. */
	271	if (ret < 0)
	272	domid = DOMID_SELF;
71eef7d1 IC	273
71eef7d1 IC	274	memset(&map_irq, 0, sizeof(map_irq));
beafbdc1	275	map_irq.domid = domid;
71eef7d1 IC	276	map_irq.type = MAP_PIRQ_TYPE_MSI;
	277	map_irq.index = -1;
	278	map_irq.pirq = -1;
	279	map_irq.bus = dev->bus->number;
	280	map_irq.devfn = dev->devfn;
	281
	282	if (type == PCI_CAP_ID_MSIX) {
	283	int pos;
	284	u32 table_offset, bir;
	285
	286	pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
	287
	288	pci_read_config_dword(dev, pos + PCI_MSIX_TABLE,
	289	&table_offset);
	290	bir = (u8)(table_offset & PCI_MSIX_FLAGS_BIRMASK);
	291
	292	map_irq.table_base = pci_resource_start(dev, bir);
	293	map_irq.entry_nr = msidesc->msi_attrib.entry_nr;
	294	}
	295
	296	ret = HYPERVISOR_physdev_op(PHYSDEVOP_map_pirq, &map_irq);
	297	if (ret) {
beafbdc1 KRW	298	dev_warn(&dev->dev, "xen map irq failed %d for %d domain\n",
beafbdc1 KRW	299	ret, domid);
71eef7d1 IC	300	goto out;
	301	}
	302
	303	ret = xen_bind_pirq_msi_to_irq(dev, msidesc,
	304	map_irq.pirq, map_irq.index,
	305	(type == PCI_CAP_ID_MSIX) ?
beafbdc1 KRW	306	"msi-x" : "msi",
beafbdc1 KRW	307	domid);
71eef7d1 IC	308	if (ret < 0)
71eef7d1 IC	309	goto out;
f731e3ef	310	}
71eef7d1 IC	311	ret = 0;
	312	out:
	313	return ret;
f731e3ef	314	}
b5401a96 AN	315	#endif
b5401a96 AN	316
fef6e262	317	static void xen_teardown_msi_irqs(struct pci_dev *dev)
b5401a96	318	{
fef6e262	319	struct msi_desc *msidesc;
b5401a96	320
fef6e262 KRW	321	msidesc = list_entry(dev->msi_list.next, struct msi_desc, list);
	322	if (msidesc->msi_attrib.is_msix)
	323	xen_pci_frontend_disable_msix(dev);
	324	else
	325	xen_pci_frontend_disable_msi(dev);
b5401a96	326
fef6e262 KRW	327	/* Free the IRQ's and the msidesc using the generic code. */
	328	default_teardown_msi_irqs(dev);
	329	}
f4d0635b	330
fef6e262 KRW	331	static void xen_teardown_msi_irq(unsigned int irq)
	332	{
	333	xen_destroy_irq(irq);
	334	}
b5401a96	335
fef6e262	336	#endif
3f2a230c	337
b5401a96 AN	338	int __init pci_xen_init(void)
	339	{
	340	if (!xen_pv_domain() \|\| xen_initial_domain())
	341	return -ENODEV;
	342
	343	printk(KERN_INFO "PCI: setting up Xen PCI frontend stub\n");
	344
	345	pcibios_set_cache_line_size();
	346
	347	pcibios_enable_irq = xen_pcifront_enable_irq;
	348	pcibios_disable_irq = NULL;
	349
	350	#ifdef CONFIG_ACPI
	351	/* Keep ACPI out of the picture */
	352	acpi_noirq = 1;
	353	#endif
	354
b5401a96 AN	355	#ifdef CONFIG_PCI_MSI
	356	x86_msi.setup_msi_irqs = xen_setup_msi_irqs;
	357	x86_msi.teardown_msi_irq = xen_teardown_msi_irq;
	358	x86_msi.teardown_msi_irqs = xen_teardown_msi_irqs;
	359	#endif
	360	return 0;
	361	}
3942b740 SS	362
	363	int __init pci_xen_hvm_init(void)
	364	{
	365	if (!xen_feature(XENFEAT_hvm_pirqs))
	366	return 0;
	367
	368	#ifdef CONFIG_ACPI
	369	/*
	370	* We don't want to change the actual ACPI delivery model,
	371	* just how GSIs get registered.
	372	*/
	373	__acpi_register_gsi = acpi_register_gsi_xen_hvm;
	374	#endif
809f9267 SS	375
	376	#ifdef CONFIG_PCI_MSI
	377	x86_msi.setup_msi_irqs = xen_hvm_setup_msi_irqs;
	378	x86_msi.teardown_msi_irq = xen_teardown_msi_irq;
	379	#endif
3942b740 SS	380	return 0;
3942b740 SS	381	}
38aa66fc JF	382
38aa66fc JF	383	#ifdef CONFIG_XEN_DOM0
d92edd81	384	#ifdef CONFIG_ACPI
38aa66fc JF	385	static __init void xen_setup_acpi_sci(void)
	386	{
	387	int rc;
	388	int trigger, polarity;
	389	int gsi = acpi_sci_override_gsi;
ee339fe6 KRW	390	int irq = -1;
ee339fe6 KRW	391	int gsi_override = -1;
38aa66fc JF	392
	393	if (!gsi)
	394	return;
	395
	396	rc = acpi_get_override_irq(gsi, &trigger, &polarity);
	397	if (rc) {
	398	printk(KERN_WARNING "xen: acpi_get_override_irq failed for acpi"
	399	" sci, rc=%d\n", rc);
	400	return;
	401	}
	402	trigger = trigger ? ACPI_LEVEL_SENSITIVE : ACPI_EDGE_SENSITIVE;
	403	polarity = polarity ? ACPI_ACTIVE_LOW : ACPI_ACTIVE_HIGH;
996c34ae	404
38aa66fc JF	405	printk(KERN_INFO "xen: sci override: global_irq=%d trigger=%d "
	406	"polarity=%d\n", gsi, trigger, polarity);
	407
ee339fe6 KRW	408	/* Before we bind the GSI to a Linux IRQ, check whether
	409	* we need to override it with bus_irq (IRQ) value. Usually for
	410	* IRQs below IRQ_LEGACY_IRQ this holds IRQ == GSI, as so:
	411	* ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 low level)
	412	* but there are oddballs where the IRQ != GSI:
	413	* ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 20 low level)
	414	* which ends up being: gsi_to_irq[9] == 20
	415	* (which is what acpi_gsi_to_irq ends up calling when starting the
	416	* the ACPI interpreter and keels over since IRQ 9 has not been
	417	* setup as we had setup IRQ 20 for it).
	418	*/
	419	/* Check whether the GSI != IRQ */
	420	if (acpi_gsi_to_irq(gsi, &irq) == 0) {
	421	if (irq >= 0 && irq != gsi)
	422	/* Bugger, we MUST have that IRQ. */
	423	gsi_override = irq;
	424	}
	425
	426	gsi = xen_register_gsi(gsi, gsi_override, trigger, polarity);
38aa66fc JF	427	printk(KERN_INFO "xen: acpi sci %d\n", gsi);
	428
	429	return;
	430	}
d92edd81	431	#endif
38aa66fc JF	432	static int __init pci_xen_initial_domain(void)
38aa66fc JF	433	{
f731e3ef QH	434	#ifdef CONFIG_PCI_MSI
	435	x86_msi.setup_msi_irqs = xen_initdom_setup_msi_irqs;
	436	x86_msi.teardown_msi_irq = xen_teardown_msi_irq;
	437	#endif
d92edd81	438	#ifdef CONFIG_ACPI
38aa66fc JF	439	xen_setup_acpi_sci();
38aa66fc JF	440	__acpi_register_gsi = acpi_register_gsi_xen;
d92edd81	441	#endif
38aa66fc JF	442	return 0;
	443	}
	444
	445	void __init xen_setup_pirqs(void)
	446	{
f4d0635b	447	int pirq, irq;
38aa66fc JF	448
	449	pci_xen_initial_domain();
	450
	451	if (0 == nr_ioapics) {
f4d0635b IC	452	for (irq = 0; irq < NR_IRQS_LEGACY; irq++) {
	453	pirq = xen_allocate_pirq_gsi(irq);
	454	if (WARN(pirq < 0,
	455	"Could not allocate PIRQ for legacy interrupt\n"))
	456	break;
	457	irq = xen_bind_pirq_gsi_to_irq(irq, pirq, 0, "xt-pic");
	458	}
38aa66fc JF	459	return;
38aa66fc JF	460	}
d92edd81	461	#ifdef CONFIG_ACPI
38aa66fc JF	462	/* Pre-allocate legacy irqs */
	463	for (irq = 0; irq < NR_IRQS_LEGACY; irq++) {
	464	int trigger, polarity;
	465
	466	if (acpi_get_override_irq(irq, &trigger, &polarity) == -1)
	467	continue;
	468
ee339fe6	469	xen_register_pirq(irq, -1 /* no GSI override */,
ed89eb63 KRW	470	trigger ? ACPI_LEVEL_SENSITIVE : ACPI_EDGE_SENSITIVE,
ed89eb63 KRW	471	true /* allocate IRQ */);
38aa66fc	472	}
d92edd81	473	#endif
38aa66fc	474	}
c55fa78b KRW	475
	476	struct xen_device_domain_owner {
	477	domid_t domain;
	478	struct pci_dev *dev;
	479	struct list_head list;
	480	};
	481
	482	static DEFINE_SPINLOCK(dev_domain_list_spinlock);
	483	static struct list_head dev_domain_list = LIST_HEAD_INIT(dev_domain_list);
	484
	485	static struct xen_device_domain_owner find_device(struct pci_dev dev)
	486	{
	487	struct xen_device_domain_owner *owner;
	488
	489	list_for_each_entry(owner, &dev_domain_list, list) {
	490	if (owner->dev == dev)
	491	return owner;
	492	}
	493	return NULL;
	494	}
	495
	496	int xen_find_device_domain_owner(struct pci_dev *dev)
	497	{
	498	struct xen_device_domain_owner *owner;
	499	int domain = -ENODEV;
	500
	501	spin_lock(&dev_domain_list_spinlock);
	502	owner = find_device(dev);
	503	if (owner)
	504	domain = owner->domain;
	505	spin_unlock(&dev_domain_list_spinlock);
	506	return domain;
	507	}
	508	EXPORT_SYMBOL_GPL(xen_find_device_domain_owner);
	509
	510	int xen_register_device_domain_owner(struct pci_dev *dev, uint16_t domain)
	511	{
	512	struct xen_device_domain_owner *owner;
	513
	514	owner = kzalloc(sizeof(struct xen_device_domain_owner), GFP_KERNEL);
	515	if (!owner)
	516	return -ENODEV;
	517
	518	spin_lock(&dev_domain_list_spinlock);
	519	if (find_device(dev)) {
	520	spin_unlock(&dev_domain_list_spinlock);
	521	kfree(owner);
	522	return -EEXIST;
	523	}
	524	owner->domain = domain;
	525	owner->dev = dev;
	526	list_add_tail(&owner->list, &dev_domain_list);
	527	spin_unlock(&dev_domain_list_spinlock);
	528	return 0;
	529	}
	530	EXPORT_SYMBOL_GPL(xen_register_device_domain_owner);
	531
	532	int xen_unregister_device_domain_owner(struct pci_dev *dev)
	533	{
	534	struct xen_device_domain_owner *owner;
	535
	536	spin_lock(&dev_domain_list_spinlock);
	537	owner = find_device(dev);
	538	if (!owner) {
539	spin_unlock(&dev_domain_list_spinlock);
540	return -ENODEV;
541	}
542	list_del(&owner->list);
543	spin_unlock(&dev_domain_list_spinlock);
544	kfree(owner);
545	return 0;
546	}
547	EXPORT_SYMBOL_GPL(xen_unregister_device_domain_owner);
7c1bfd68	548	#endif