[deliverable/linux.git] / drivers / virtio / virtio_pci_common.c

/*
 * Virtio PCI driver - common functionality for all device versions
 *
 * This module allows virtio devices to be used over a virtual PCI device.
 * This can be used with QEMU based VMMs like KVM or Xen.
 *
 * Copyright IBM Corp. 2007
 * Copyright Red Hat, Inc. 2014
 *
 * Authors:
 *  Anthony Liguori  <aliguori@us.ibm.com>
 *  Rusty Russell <rusty@rustcorp.com.au>
 *  Michael S. Tsirkin <mst@redhat.com>
 *
 * This work is licensed under the terms of the GNU GPL, version 2 or later.
 * See the COPYING file in the top-level directory.
 *
 */

#include "virtio_pci_common.h"

/* wait for pending irq handlers */
void vp_synchronize_vectors(struct virtio_device *vdev)
{
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	int i;

	if (vp_dev->intx_enabled)
		synchronize_irq(vp_dev->pci_dev->irq);

	for (i = 0; i < vp_dev->msix_vectors; ++i)
		synchronize_irq(vp_dev->msix_entries[i].vector);
}

/* the notify function used when creating a virt queue */
bool vp_notify(struct virtqueue *vq)
{
	/* we write the queue's selector into the notification register to
	 * signal the other end */
	iowrite16(vq->index, (void __iomem *)vq->priv);
	return true;
}

/* Handle a configuration change: Tell driver if it wants to know. */
static irqreturn_t vp_config_changed(int irq, void *opaque)
{
	struct virtio_pci_device *vp_dev = opaque;

	virtio_config_changed(&vp_dev->vdev);
	return IRQ_HANDLED;
}

/* Notify all virtqueues on an interrupt. */
static irqreturn_t vp_vring_interrupt(int irq, void *opaque)
{
	struct virtio_pci_device *vp_dev = opaque;
	struct virtio_pci_vq_info *info;
	irqreturn_t ret = IRQ_NONE;
	unsigned long flags;

	spin_lock_irqsave(&vp_dev->lock, flags);
	list_for_each_entry(info, &vp_dev->virtqueues, node) {
		if (vring_interrupt(irq, info->vq) == IRQ_HANDLED)
			ret = IRQ_HANDLED;
	}
	spin_unlock_irqrestore(&vp_dev->lock, flags);

	return ret;
}

/* A small wrapper to also acknowledge the interrupt when it's handled.
 * I really need an EIO hook for the vring so I can ack the interrupt once we
 * know that we'll be handling the IRQ but before we invoke the callback since
 * the callback may notify the host which results in the host attempting to
 * raise an interrupt that we would then mask once we acknowledged the
 * interrupt. */
static irqreturn_t vp_interrupt(int irq, void *opaque)
{
	struct virtio_pci_device *vp_dev = opaque;
	u8 isr;

	/* reading the ISR has the effect of also clearing it so it's very
	 * important to save off the value. */
	isr = ioread8(vp_dev->isr);

	/* It's definitely not us if the ISR was not high */
	if (!isr)
		return IRQ_NONE;

	/* Configuration change?  Tell driver if it wants to know. */
	if (isr & VIRTIO_PCI_ISR_CONFIG)
		vp_config_changed(irq, opaque);

	return vp_vring_interrupt(irq, opaque);
}

static void vp_free_vectors(struct virtio_device *vdev)
{
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	int i;

	if (vp_dev->intx_enabled) {
		free_irq(vp_dev->pci_dev->irq, vp_dev);
		vp_dev->intx_enabled = 0;
	}

	for (i = 0; i < vp_dev->msix_used_vectors; ++i)
		free_irq(vp_dev->msix_entries[i].vector, vp_dev);

	for (i = 0; i < vp_dev->msix_vectors; i++)
		if (vp_dev->msix_affinity_masks[i])
			free_cpumask_var(vp_dev->msix_affinity_masks[i]);

	if (vp_dev->msix_enabled) {
		/* Disable the vector used for configuration */
		vp_dev->config_vector(vp_dev, VIRTIO_MSI_NO_VECTOR);

		pci_disable_msix(vp_dev->pci_dev);
		vp_dev->msix_enabled = 0;
	}

	vp_dev->msix_vectors = 0;
	vp_dev->msix_used_vectors = 0;
	kfree(vp_dev->msix_names);
	vp_dev->msix_names = NULL;
	kfree(vp_dev->msix_entries);
	vp_dev->msix_entries = NULL;
	kfree(vp_dev->msix_affinity_masks);
	vp_dev->msix_affinity_masks = NULL;
}

static int vp_request_msix_vectors(struct virtio_device *vdev, int nvectors,
				   bool per_vq_vectors)
{
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	const char *name = dev_name(&vp_dev->vdev.dev);
	unsigned i, v;
	int err = -ENOMEM;

	vp_dev->msix_vectors = nvectors;

	vp_dev->msix_entries = kmalloc(nvectors * sizeof *vp_dev->msix_entries,
				       GFP_KERNEL);
	if (!vp_dev->msix_entries)
		goto error;
	vp_dev->msix_names = kmalloc(nvectors * sizeof *vp_dev->msix_names,
				     GFP_KERNEL);
	if (!vp_dev->msix_names)
		goto error;
	vp_dev->msix_affinity_masks
		= kzalloc(nvectors * sizeof *vp_dev->msix_affinity_masks,
			  GFP_KERNEL);
	if (!vp_dev->msix_affinity_masks)
		goto error;
	for (i = 0; i < nvectors; ++i)
		if (!alloc_cpumask_var(&vp_dev->msix_affinity_masks[i],
					GFP_KERNEL))
			goto error;

	for (i = 0; i < nvectors; ++i)
		vp_dev->msix_entries[i].entry = i;

	err = pci_enable_msix_exact(vp_dev->pci_dev,
				    vp_dev->msix_entries, nvectors);
	if (err)
		goto error;
	vp_dev->msix_enabled = 1;

	/* Set the vector used for configuration */
	v = vp_dev->msix_used_vectors;
	snprintf(vp_dev->msix_names[v], sizeof *vp_dev->msix_names,
		 "%s-config", name);
	err = request_irq(vp_dev->msix_entries[v].vector,
			  vp_config_changed, 0, vp_dev->msix_names[v],
			  vp_dev);
	if (err)
		goto error;
	++vp_dev->msix_used_vectors;

	v = vp_dev->config_vector(vp_dev, v);
	/* Verify we had enough resources to assign the vector */
	if (v == VIRTIO_MSI_NO_VECTOR) {
		err = -EBUSY;
		goto error;
	}

	if (!per_vq_vectors) {
		/* Shared vector for all VQs */
		v = vp_dev->msix_used_vectors;
		snprintf(vp_dev->msix_names[v], sizeof *vp_dev->msix_names,
			 "%s-virtqueues", name);
		err = request_irq(vp_dev->msix_entries[v].vector,
				  vp_vring_interrupt, 0, vp_dev->msix_names[v],
				  vp_dev);
		if (err)
			goto error;
		++vp_dev->msix_used_vectors;
	}
	return 0;
error:
	vp_free_vectors(vdev);
	return err;
}

static int vp_request_intx(struct virtio_device *vdev)
{
	int err;
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);

	err = request_irq(vp_dev->pci_dev->irq, vp_interrupt,
			  IRQF_SHARED, dev_name(&vdev->dev), vp_dev);
	if (!err)
		vp_dev->intx_enabled = 1;
	return err;
}

static struct virtqueue *vp_setup_vq(struct virtio_device *vdev, unsigned index,
				     void (*callback)(struct virtqueue *vq),
				     const char *name,
				     u16 msix_vec)
{
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	struct virtio_pci_vq_info *info = kmalloc(sizeof *info, GFP_KERNEL);
	struct virtqueue *vq;
	unsigned long flags;

	/* fill out our structure that represents an active queue */
	if (!info)
		return ERR_PTR(-ENOMEM);

	vq = vp_dev->setup_vq(vp_dev, info, index, callback, name, msix_vec);
	if (IS_ERR(vq))
		goto out_info;

	info->vq = vq;
	if (callback) {
		spin_lock_irqsave(&vp_dev->lock, flags);
		list_add(&info->node, &vp_dev->virtqueues);
		spin_unlock_irqrestore(&vp_dev->lock, flags);
	} else {
		INIT_LIST_HEAD(&info->node);
	}

	vp_dev->vqs[index] = info;
	return vq;

out_info:
	kfree(info);
	return vq;
}

static void vp_del_vq(struct virtqueue *vq)
{
	struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev);
	struct virtio_pci_vq_info *info = vp_dev->vqs[vq->index];
	unsigned long flags;

	spin_lock_irqsave(&vp_dev->lock, flags);
	list_del(&info->node);
	spin_unlock_irqrestore(&vp_dev->lock, flags);

	vp_dev->del_vq(info);
	kfree(info);
}

/* the config->del_vqs() implementation */
void vp_del_vqs(struct virtio_device *vdev)
{
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	struct virtqueue *vq, *n;
	struct virtio_pci_vq_info *info;

	list_for_each_entry_safe(vq, n, &vdev->vqs, list) {
		info = vp_dev->vqs[vq->index];
		if (vp_dev->per_vq_vectors &&
			info->msix_vector != VIRTIO_MSI_NO_VECTOR)
			free_irq(vp_dev->msix_entries[info->msix_vector].vector,
				 vq);
		vp_del_vq(vq);
	}
	vp_dev->per_vq_vectors = false;

	vp_free_vectors(vdev);
	kfree(vp_dev->vqs);
	vp_dev->vqs = NULL;
}

static int vp_try_to_find_vqs(struct virtio_device *vdev, unsigned nvqs,
			      struct virtqueue *vqs[],
			      vq_callback_t *callbacks[],
			      const char *names[],
			      bool use_msix,
			      bool per_vq_vectors)
{
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	u16 msix_vec;
	int i, err, nvectors, allocated_vectors;

	vp_dev->vqs = kmalloc(nvqs * sizeof *vp_dev->vqs, GFP_KERNEL);
	if (!vp_dev->vqs)
		return -ENOMEM;

	if (!use_msix) {
		/* Old style: one normal interrupt for change and all vqs. */
		err = vp_request_intx(vdev);
		if (err)
			goto error_find;
	} else {
		if (per_vq_vectors) {
			/* Best option: one for change interrupt, one per vq. */
			nvectors = 1;
			for (i = 0; i < nvqs; ++i)
				if (callbacks[i])
					++nvectors;
		} else {
			/* Second best: one for change, shared for all vqs. */
			nvectors = 2;
		}

		err = vp_request_msix_vectors(vdev, nvectors, per_vq_vectors);
		if (err)
			goto error_find;
	}

	vp_dev->per_vq_vectors = per_vq_vectors;
	allocated_vectors = vp_dev->msix_used_vectors;
	for (i = 0; i < nvqs; ++i) {
		if (!names[i]) {
			vqs[i] = NULL;
			continue;
		} else if (!callbacks[i] || !vp_dev->msix_enabled)
			msix_vec = VIRTIO_MSI_NO_VECTOR;
		else if (vp_dev->per_vq_vectors)
			msix_vec = allocated_vectors++;
		else
			msix_vec = VP_MSIX_VQ_VECTOR;
		vqs[i] = vp_setup_vq(vdev, i, callbacks[i], names[i], msix_vec);
		if (IS_ERR(vqs[i])) {
			err = PTR_ERR(vqs[i]);
			goto error_find;
		}

		if (!vp_dev->per_vq_vectors || msix_vec == VIRTIO_MSI_NO_VECTOR)
			continue;

		/* allocate per-vq irq if available and necessary */
		snprintf(vp_dev->msix_names[msix_vec],
			 sizeof *vp_dev->msix_names,
			 "%s-%s",
			 dev_name(&vp_dev->vdev.dev), names[i]);
		err = request_irq(vp_dev->msix_entries[msix_vec].vector,
				  vring_interrupt, 0,
				  vp_dev->msix_names[msix_vec],
				  vqs[i]);
		if (err) {
			vp_del_vq(vqs[i]);
			goto error_find;
		}
	}
	return 0;

error_find:
	vp_del_vqs(vdev);
	return err;
}

/* the config->find_vqs() implementation */
int vp_find_vqs(struct virtio_device *vdev, unsigned nvqs,
		struct virtqueue *vqs[],
		vq_callback_t *callbacks[],
		const char *names[])
{
	int err;

	/* Try MSI-X with one vector per queue. */
	err = vp_try_to_find_vqs(vdev, nvqs, vqs, callbacks, names, true, true);
	if (!err)
		return 0;
	/* Fallback: MSI-X with one vector for config, one shared for queues. */
	err = vp_try_to_find_vqs(vdev, nvqs, vqs, callbacks, names,
				 true, false);
	if (!err)
		return 0;
	/* Finally fall back to regular interrupts. */
	return vp_try_to_find_vqs(vdev, nvqs, vqs, callbacks, names,
				  false, false);
}

const char *vp_bus_name(struct virtio_device *vdev)
{
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);

	return pci_name(vp_dev->pci_dev);
}

/* Setup the affinity for a virtqueue:
 * - force the affinity for per vq vector
 * - OR over all affinities for shared MSI
 * - ignore the affinity request if we're using INTX
 */
int vp_set_vq_affinity(struct virtqueue *vq, int cpu)
{
	struct virtio_device *vdev = vq->vdev;
	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	struct virtio_pci_vq_info *info = vp_dev->vqs[vq->index];
	struct cpumask *mask;
	unsigned int irq;

	if (!vq->callback)
		return -EINVAL;

	if (vp_dev->msix_enabled) {
		mask = vp_dev->msix_affinity_masks[info->msix_vector];
		irq = vp_dev->msix_entries[info->msix_vector].vector;
		if (cpu == -1)
			irq_set_affinity_hint(irq, NULL);
		else {
			cpumask_set_cpu(cpu, mask);
			irq_set_affinity_hint(irq, mask);
		}
	}
	return 0;
}

#ifdef CONFIG_PM_SLEEP
static int virtio_pci_freeze(struct device *dev)
{
	struct pci_dev *pci_dev = to_pci_dev(dev);
	struct virtio_pci_device *vp_dev = pci_get_drvdata(pci_dev);
	int ret;

	ret = virtio_device_freeze(&vp_dev->vdev);

	if (!ret)
		pci_disable_device(pci_dev);
	return ret;
}

static int virtio_pci_restore(struct device *dev)
{
	struct pci_dev *pci_dev = to_pci_dev(dev);
	struct virtio_pci_device *vp_dev = pci_get_drvdata(pci_dev);
	int ret;

	ret = pci_enable_device(pci_dev);
	if (ret)
		return ret;

	pci_set_master(pci_dev);
	return virtio_device_restore(&vp_dev->vdev);
}

static const struct dev_pm_ops virtio_pci_pm_ops = {
	SET_SYSTEM_SLEEP_PM_OPS(virtio_pci_freeze, virtio_pci_restore)
};
#endif


/* Qumranet donated their vendor ID for devices 0x1000 thru 0x10FF. */
static const struct pci_device_id virtio_pci_id_table[] = {
	{ PCI_DEVICE(0x1af4, PCI_ANY_ID) },
	{ 0 }
};

MODULE_DEVICE_TABLE(pci, virtio_pci_id_table);

static int virtio_pci_probe(struct pci_dev *pci_dev,
			    const struct pci_device_id *id)
{
	return virtio_pci_legacy_probe(pci_dev, id);
}

static void virtio_pci_remove(struct pci_dev *pci_dev)
{
     virtio_pci_legacy_remove(pci_dev);
}

static struct pci_driver virtio_pci_driver = {
	.name		= "virtio-pci",
	.id_table	= virtio_pci_id_table,
	.probe		= virtio_pci_probe,
	.remove		= virtio_pci_remove,
#ifdef CONFIG_PM_SLEEP
	.driver.pm	= &virtio_pci_pm_ops,
#endif
};

module_pci_driver(virtio_pci_driver);

MODULE_AUTHOR("Anthony Liguori <aliguori@us.ibm.com>");
MODULE_DESCRIPTION("virtio-pci");
MODULE_LICENSE("GPL");
MODULE_VERSION("1");
Commit	Line	Data
3343660d	1	/*
a90fdce9	2	* Virtio PCI driver - common functionality for all device versions
3343660d AL	3	*
	4	* This module allows virtio devices to be used over a virtual PCI device.
	5	* This can be used with QEMU based VMMs like KVM or Xen.
	6	*
	7	* Copyright IBM Corp. 2007
a90fdce9	8	* Copyright Red Hat, Inc. 2014
3343660d AL	9	*
	10	* Authors:
	11	* Anthony Liguori <aliguori@us.ibm.com>
a90fdce9 MT	12	* Rusty Russell <rusty@rustcorp.com.au>
a90fdce9 MT	13	* Michael S. Tsirkin <mst@redhat.com>
3343660d AL	14	*
	15	* This work is licensed under the terms of the GNU GPL, version 2 or later.
	16	* See the COPYING file in the top-level directory.
	17	*
	18	*/
	19
5f4c9760	20	#include "virtio_pci_common.h"
3343660d	21
e6af578c	22	/* wait for pending irq handlers */
38eb4a29	23	void vp_synchronize_vectors(struct virtio_device *vdev)
e6af578c MT	24	{
	25	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	26	int i;
	27
	28	if (vp_dev->intx_enabled)
	29	synchronize_irq(vp_dev->pci_dev->irq);
	30
	31	for (i = 0; i < vp_dev->msix_vectors; ++i)
	32	synchronize_irq(vp_dev->msix_entries[i].vector);
	33	}
	34
3343660d	35	/* the notify function used when creating a virt queue */
38eb4a29	36	bool vp_notify(struct virtqueue *vq)
3343660d	37	{
3343660d AL	38	/* we write the queue's selector into the notification register to
3343660d AL	39	* signal the other end */
f30eaf4a	40	iowrite16(vq->index, (void __iomem *)vq->priv);
46f9c2b9	41	return true;
3343660d AL	42	}
3343660d AL	43
77cf5246 MT	44	/* Handle a configuration change: Tell driver if it wants to know. */
	45	static irqreturn_t vp_config_changed(int irq, void *opaque)
	46	{
	47	struct virtio_pci_device *vp_dev = opaque;
77cf5246	48
016c98c6	49	virtio_config_changed(&vp_dev->vdev);
77cf5246 MT	50	return IRQ_HANDLED;
	51	}
	52
	53	/* Notify all virtqueues on an interrupt. */
	54	static irqreturn_t vp_vring_interrupt(int irq, void *opaque)
	55	{
	56	struct virtio_pci_device *vp_dev = opaque;
	57	struct virtio_pci_vq_info *info;
	58	irqreturn_t ret = IRQ_NONE;
	59	unsigned long flags;
	60
	61	spin_lock_irqsave(&vp_dev->lock, flags);
	62	list_for_each_entry(info, &vp_dev->virtqueues, node) {
	63	if (vring_interrupt(irq, info->vq) == IRQ_HANDLED)
	64	ret = IRQ_HANDLED;
	65	}
	66	spin_unlock_irqrestore(&vp_dev->lock, flags);
	67
	68	return ret;
	69	}
	70
3343660d AL	71	/* A small wrapper to also acknowledge the interrupt when it's handled.
	72	* I really need an EIO hook for the vring so I can ack the interrupt once we
	73	* know that we'll be handling the IRQ but before we invoke the callback since
	74	* the callback may notify the host which results in the host attempting to
	75	* raise an interrupt that we would then mask once we acknowledged the
	76	* interrupt. */
	77	static irqreturn_t vp_interrupt(int irq, void *opaque)
	78	{
	79	struct virtio_pci_device *vp_dev = opaque;
3343660d AL	80	u8 isr;
	81
	82	/* reading the ISR has the effect of also clearing it so it's very
	83	* important to save off the value. */
af535722	84	isr = ioread8(vp_dev->isr);
3343660d AL	85
	86	/* It's definitely not us if the ISR was not high */
	87	if (!isr)
	88	return IRQ_NONE;
	89
	90	/* Configuration change? Tell driver if it wants to know. */
77cf5246 MT	91	if (isr & VIRTIO_PCI_ISR_CONFIG)
77cf5246 MT	92	vp_config_changed(irq, opaque);
3343660d	93
77cf5246	94	return vp_vring_interrupt(irq, opaque);
3343660d AL	95	}
3343660d AL	96
82af8ce8 MT	97	static void vp_free_vectors(struct virtio_device *vdev)
	98	{
	99	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	100	int i;
	101
	102	if (vp_dev->intx_enabled) {
	103	free_irq(vp_dev->pci_dev->irq, vp_dev);
	104	vp_dev->intx_enabled = 0;
	105	}
	106
	107	for (i = 0; i < vp_dev->msix_used_vectors; ++i)
	108	free_irq(vp_dev->msix_entries[i].vector, vp_dev);
82af8ce8	109
75a0a52b JW	110	for (i = 0; i < vp_dev->msix_vectors; i++)
	111	if (vp_dev->msix_affinity_masks[i])
	112	free_cpumask_var(vp_dev->msix_affinity_masks[i]);
	113
82af8ce8 MT	114	if (vp_dev->msix_enabled) {
82af8ce8 MT	115	/* Disable the vector used for configuration */
6f8f23d6	116	vp_dev->config_vector(vp_dev, VIRTIO_MSI_NO_VECTOR);
82af8ce8	117
82af8ce8	118	pci_disable_msix(vp_dev->pci_dev);
ff52c3fc	119	vp_dev->msix_enabled = 0;
82af8ce8	120	}
ff52c3fc	121
f11335db	122	vp_dev->msix_vectors = 0;
ff52c3fc MT	123	vp_dev->msix_used_vectors = 0;
	124	kfree(vp_dev->msix_names);
	125	vp_dev->msix_names = NULL;
	126	kfree(vp_dev->msix_entries);
	127	vp_dev->msix_entries = NULL;
75a0a52b JW	128	kfree(vp_dev->msix_affinity_masks);
75a0a52b JW	129	vp_dev->msix_affinity_masks = NULL;
82af8ce8 MT	130	}
82af8ce8 MT	131
f68d2408 RR	132	static int vp_request_msix_vectors(struct virtio_device *vdev, int nvectors,
f68d2408 RR	133	bool per_vq_vectors)
82af8ce8 MT	134	{
	135	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	136	const char *name = dev_name(&vp_dev->vdev.dev);
	137	unsigned i, v;
	138	int err = -ENOMEM;
e969fed5	139
f11335db AV	140	vp_dev->msix_vectors = nvectors;
f11335db AV	141
82af8ce8 MT	142	vp_dev->msix_entries = kmalloc(nvectors * sizeof *vp_dev->msix_entries,
	143	GFP_KERNEL);
	144	if (!vp_dev->msix_entries)
ff52c3fc	145	goto error;
82af8ce8 MT	146	vp_dev->msix_names = kmalloc(nvectors * sizeof *vp_dev->msix_names,
	147	GFP_KERNEL);
	148	if (!vp_dev->msix_names)
ff52c3fc	149	goto error;
75a0a52b JW	150	vp_dev->msix_affinity_masks
	151	= kzalloc(nvectors * sizeof *vp_dev->msix_affinity_masks,
	152	GFP_KERNEL);
	153	if (!vp_dev->msix_affinity_masks)
	154	goto error;
	155	for (i = 0; i < nvectors; ++i)
	156	if (!alloc_cpumask_var(&vp_dev->msix_affinity_masks[i],
	157	GFP_KERNEL))
	158	goto error;
82af8ce8 MT	159
	160	for (i = 0; i < nvectors; ++i)
	161	vp_dev->msix_entries[i].entry = i;
	162
5e37f670 AG	163	err = pci_enable_msix_exact(vp_dev->pci_dev,
5e37f670 AG	164	vp_dev->msix_entries, nvectors);
e969fed5 MT	165	if (err)
e969fed5 MT	166	goto error;
e969fed5 MT	167	vp_dev->msix_enabled = 1;
	168
	169	/* Set the vector used for configuration */
	170	v = vp_dev->msix_used_vectors;
	171	snprintf(vp_dev->msix_names[v], sizeof *vp_dev->msix_names,
	172	"%s-config", name);
	173	err = request_irq(vp_dev->msix_entries[v].vector,
	174	vp_config_changed, 0, vp_dev->msix_names[v],
	175	vp_dev);
	176	if (err)
	177	goto error;
	178	++vp_dev->msix_used_vectors;
82af8ce8	179
6f8f23d6	180	v = vp_dev->config_vector(vp_dev, v);
e969fed5	181	/* Verify we had enough resources to assign the vector */
e969fed5 MT	182	if (v == VIRTIO_MSI_NO_VECTOR) {
	183	err = -EBUSY;
	184	goto error;
82af8ce8 MT	185	}
82af8ce8 MT	186
e969fed5	187	if (!per_vq_vectors) {
82af8ce8 MT	188	/* Shared vector for all VQs */
	189	v = vp_dev->msix_used_vectors;
	190	snprintf(vp_dev->msix_names[v], sizeof *vp_dev->msix_names,
	191	"%s-virtqueues", name);
	192	err = request_irq(vp_dev->msix_entries[v].vector,
	193	vp_vring_interrupt, 0, vp_dev->msix_names[v],
	194	vp_dev);
	195	if (err)
ff52c3fc	196	goto error;
82af8ce8 MT	197	++vp_dev->msix_used_vectors;
	198	}
	199	return 0;
ff52c3fc	200	error:
82af8ce8	201	vp_free_vectors(vdev);
82af8ce8 MT	202	return err;
	203	}
	204
f68d2408 RR	205	static int vp_request_intx(struct virtio_device *vdev)
	206	{
	207	int err;
	208	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	209
	210	err = request_irq(vp_dev->pci_dev->irq, vp_interrupt,
	211	IRQF_SHARED, dev_name(&vdev->dev), vp_dev);
	212	if (!err)
	213	vp_dev->intx_enabled = 1;
	214	return err;
	215	}
	216
b09f00bb MT	217	static struct virtqueue vp_setup_vq(struct virtio_device vdev, unsigned index,
	218	void (callback)(struct virtqueue vq),
	219	const char *name,
	220	u16 msix_vec)
	221	{
	222	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	223	struct virtio_pci_vq_info info = kmalloc(sizeof info, GFP_KERNEL);
	224	struct virtqueue *vq;
	225	unsigned long flags;
	226
	227	/* fill out our structure that represents an active queue */
	228	if (!info)
	229	return ERR_PTR(-ENOMEM);
	230
	231	vq = vp_dev->setup_vq(vp_dev, info, index, callback, name, msix_vec);
	232	if (IS_ERR(vq))
	233	goto out_info;
	234
	235	info->vq = vq;
005b20a8 KK	236	if (callback) {
	237	spin_lock_irqsave(&vp_dev->lock, flags);
	238	list_add(&info->node, &vp_dev->virtqueues);
	239	spin_unlock_irqrestore(&vp_dev->lock, flags);
	240	} else {
	241	INIT_LIST_HEAD(&info->node);
	242	}
3343660d	243
3ec7a77b	244	vp_dev->vqs[index] = info;
3343660d AL	245	return vq;
3343660d AL	246
3343660d AL	247	out_info:
3343660d AL	248	kfree(info);
b09f00bb	249	return vq;
3343660d AL	250	}
3343660d AL	251
5386cef2 MT	252	static void vp_del_vq(struct virtqueue *vq)
	253	{
	254	struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev);
	255	struct virtio_pci_vq_info *info = vp_dev->vqs[vq->index];
	256	unsigned long flags;
	257
	258	spin_lock_irqsave(&vp_dev->lock, flags);
	259	list_del(&info->node);
	260	spin_unlock_irqrestore(&vp_dev->lock, flags);
	261
	262	vp_dev->del_vq(info);
3343660d AL	263	kfree(info);
	264	}
	265
82af8ce8	266	/* the config->del_vqs() implementation */
38eb4a29	267	void vp_del_vqs(struct virtio_device *vdev)
d2a7ddda	268	{
e969fed5	269	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
d2a7ddda	270	struct virtqueue vq, n;
e969fed5	271	struct virtio_pci_vq_info *info;
d2a7ddda	272
e969fed5	273	list_for_each_entry_safe(vq, n, &vdev->vqs, list) {
3ec7a77b	274	info = vp_dev->vqs[vq->index];
31198159 MT	275	if (vp_dev->per_vq_vectors &&
31198159 MT	276	info->msix_vector != VIRTIO_MSI_NO_VECTOR)
f68d2408 RR	277	free_irq(vp_dev->msix_entries[info->msix_vector].vector,
f68d2408 RR	278	vq);
d2a7ddda	279	vp_del_vq(vq);
e969fed5 MT	280	}
e969fed5 MT	281	vp_dev->per_vq_vectors = false;
82af8ce8 MT	282
82af8ce8 MT	283	vp_free_vectors(vdev);
3ec7a77b	284	kfree(vp_dev->vqs);
80e9541f	285	vp_dev->vqs = NULL;
d2a7ddda MT	286	}
d2a7ddda MT	287
e969fed5 MT	288	static int vp_try_to_find_vqs(struct virtio_device *vdev, unsigned nvqs,
	289	struct virtqueue *vqs[],
	290	vq_callback_t *callbacks[],
	291	const char *names[],
f68d2408	292	bool use_msix,
e969fed5	293	bool per_vq_vectors)
d2a7ddda	294	{
e969fed5	295	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
f68d2408 RR	296	u16 msix_vec;
f68d2408 RR	297	int i, err, nvectors, allocated_vectors;
82af8ce8	298
3ec7a77b MT	299	vp_dev->vqs = kmalloc(nvqs * sizeof *vp_dev->vqs, GFP_KERNEL);
	300	if (!vp_dev->vqs)
	301	return -ENOMEM;
	302
f68d2408 RR	303	if (!use_msix) {
	304	/* Old style: one normal interrupt for change and all vqs. */
	305	err = vp_request_intx(vdev);
	306	if (err)
3ec7a77b	307	goto error_find;
f68d2408 RR	308	} else {
	309	if (per_vq_vectors) {
	310	/* Best option: one for change interrupt, one per vq. */
	311	nvectors = 1;
	312	for (i = 0; i < nvqs; ++i)
	313	if (callbacks[i])
	314	++nvectors;
	315	} else {
	316	/* Second best: one for change, shared for all vqs. */
	317	nvectors = 2;
	318	}
	319
	320	err = vp_request_msix_vectors(vdev, nvectors, per_vq_vectors);
	321	if (err)
3ec7a77b	322	goto error_find;
f68d2408	323	}
d2a7ddda	324
e969fed5 MT	325	vp_dev->per_vq_vectors = per_vq_vectors;
e969fed5 MT	326	allocated_vectors = vp_dev->msix_used_vectors;
d2a7ddda	327	for (i = 0; i < nvqs; ++i) {
6457f126 MT	328	if (!names[i]) {
	329	vqs[i] = NULL;
	330	continue;
	331	} else if (!callbacks[i] \|\| !vp_dev->msix_enabled)
f68d2408	332	msix_vec = VIRTIO_MSI_NO_VECTOR;
e969fed5	333	else if (vp_dev->per_vq_vectors)
f68d2408	334	msix_vec = allocated_vectors++;
e969fed5	335	else
f68d2408	336	msix_vec = VP_MSIX_VQ_VECTOR;
b09f00bb	337	vqs[i] = vp_setup_vq(vdev, i, callbacks[i], names[i], msix_vec);
e969fed5 MT	338	if (IS_ERR(vqs[i])) {
e969fed5 MT	339	err = PTR_ERR(vqs[i]);
82af8ce8	340	goto error_find;
e969fed5	341	}
0b22bd0b MT	342
	343	if (!vp_dev->per_vq_vectors \|\| msix_vec == VIRTIO_MSI_NO_VECTOR)
	344	continue;
	345
e969fed5	346	/* allocate per-vq irq if available and necessary */
0b22bd0b MT	347	snprintf(vp_dev->msix_names[msix_vec],
	348	sizeof *vp_dev->msix_names,
	349	"%s-%s",
	350	dev_name(&vp_dev->vdev.dev), names[i]);
	351	err = request_irq(vp_dev->msix_entries[msix_vec].vector,
	352	vring_interrupt, 0,
	353	vp_dev->msix_names[msix_vec],
	354	vqs[i]);
	355	if (err) {
	356	vp_del_vq(vqs[i]);
	357	goto error_find;
e969fed5	358	}
d2a7ddda MT	359	}
	360	return 0;
	361
82af8ce8	362	error_find:
d2a7ddda	363	vp_del_vqs(vdev);
e969fed5 MT	364	return err;
	365	}
	366
	367	/* the config->find_vqs() implementation */
38eb4a29 MT	368	int vp_find_vqs(struct virtio_device *vdev, unsigned nvqs,
	369	struct virtqueue *vqs[],
	370	vq_callback_t *callbacks[],
	371	const char *names[])
e969fed5	372	{
f68d2408	373	int err;
e969fed5	374
f68d2408 RR	375	/* Try MSI-X with one vector per queue. */
f68d2408 RR	376	err = vp_try_to_find_vqs(vdev, nvqs, vqs, callbacks, names, true, true);
e969fed5 MT	377	if (!err)
e969fed5 MT	378	return 0;
f68d2408	379	/* Fallback: MSI-X with one vector for config, one shared for queues. */
e969fed5	380	err = vp_try_to_find_vqs(vdev, nvqs, vqs, callbacks, names,
f68d2408	381	true, false);
e969fed5 MT	382	if (!err)
	383	return 0;
	384	/* Finally fall back to regular interrupts. */
f68d2408 RR	385	return vp_try_to_find_vqs(vdev, nvqs, vqs, callbacks, names,
f68d2408 RR	386	false, false);
d2a7ddda MT	387	}
d2a7ddda MT	388
38eb4a29	389	const char vp_bus_name(struct virtio_device vdev)
66846048 RJ	390	{
	391	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
	392
	393	return pci_name(vp_dev->pci_dev);
	394	}
	395
75a0a52b JW	396	/* Setup the affinity for a virtqueue:
	397	* - force the affinity for per vq vector
	398	* - OR over all affinities for shared MSI
	399	* - ignore the affinity request if we're using INTX
	400	*/
38eb4a29	401	int vp_set_vq_affinity(struct virtqueue *vq, int cpu)
75a0a52b JW	402	{
	403	struct virtio_device *vdev = vq->vdev;
	404	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
3ec7a77b	405	struct virtio_pci_vq_info *info = vp_dev->vqs[vq->index];
75a0a52b JW	406	struct cpumask *mask;
	407	unsigned int irq;
	408
	409	if (!vq->callback)
	410	return -EINVAL;
	411
	412	if (vp_dev->msix_enabled) {
	413	mask = vp_dev->msix_affinity_masks[info->msix_vector];
	414	irq = vp_dev->msix_entries[info->msix_vector].vector;
	415	if (cpu == -1)
	416	irq_set_affinity_hint(irq, NULL);
	417	else {
	418	cpumask_set_cpu(cpu, mask);
	419	irq_set_affinity_hint(irq, mask);
	420	}
	421	}
	422	return 0;
	423	}
	424
9e266ece	425	#ifdef CONFIG_PM_SLEEP
f0fe6f11 AS	426	static int virtio_pci_freeze(struct device *dev)
	427	{
	428	struct pci_dev *pci_dev = to_pci_dev(dev);
	429	struct virtio_pci_device *vp_dev = pci_get_drvdata(pci_dev);
f0fe6f11 AS	430	int ret;
f0fe6f11 AS	431
c6716bae	432	ret = virtio_device_freeze(&vp_dev->vdev);
f0fe6f11 AS	433
	434	if (!ret)
	435	pci_disable_device(pci_dev);
	436	return ret;
	437	}
	438
0517fdd1	439	static int virtio_pci_restore(struct device *dev)
f0fe6f11 AS	440	{
	441	struct pci_dev *pci_dev = to_pci_dev(dev);
	442	struct virtio_pci_device *vp_dev = pci_get_drvdata(pci_dev);
	443	int ret;
	444
	445	ret = pci_enable_device(pci_dev);
	446	if (ret)
	447	return ret;
0517fdd1	448
f0fe6f11	449	pci_set_master(pci_dev);
c6716bae	450	return virtio_device_restore(&vp_dev->vdev);
f0fe6f11 AS	451	}
f0fe6f11 AS	452
9a4253db	453	static const struct dev_pm_ops virtio_pci_pm_ops = {
f878d0be	454	SET_SYSTEM_SLEEP_PM_OPS(virtio_pci_freeze, virtio_pci_restore)
d0775363	455	};
3343660d	456	#endif
9a4253db MT	457
	458
	459	/* Qumranet donated their vendor ID for devices 0x1000 thru 0x10FF. */
	460	static const struct pci_device_id virtio_pci_id_table[] = {
	461	{ PCI_DEVICE(0x1af4, PCI_ANY_ID) },
	462	{ 0 }
	463	};
	464
	465	MODULE_DEVICE_TABLE(pci, virtio_pci_id_table);
	466
	467	static int virtio_pci_probe(struct pci_dev *pci_dev,
	468	const struct pci_device_id *id)
	469	{
	470	return virtio_pci_legacy_probe(pci_dev, id);
	471	}
	472
	473	static void virtio_pci_remove(struct pci_dev *pci_dev)
	474	{
	475	virtio_pci_legacy_remove(pci_dev);
	476	}
	477
	478	static struct pci_driver virtio_pci_driver = {
	479	.name = "virtio-pci",
	480	.id_table = virtio_pci_id_table,
	481	.probe = virtio_pci_probe,
	482	.remove = virtio_pci_remove,
	483	#ifdef CONFIG_PM_SLEEP
	484	.driver.pm = &virtio_pci_pm_ops,
	485	#endif
	486	};
	487
	488	module_pci_driver(virtio_pci_driver);
5ff16110 HX	489
	490	MODULE_AUTHOR("Anthony Liguori <aliguori@us.ibm.com>");
	491	MODULE_DESCRIPTION("virtio-pci");
	492	MODULE_LICENSE("GPL");
	493	MODULE_VERSION("1");