[linux-2.6-block.git] / drivers / vfio / pci / vfio_pci.c

/*
 * Copyright (C) 2012 Red Hat, Inc.  All rights reserved.
 *     Author: Alex Williamson <alex.williamson@redhat.com>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 * Derived from original vfio:
 * Copyright 2010 Cisco Systems, Inc.  All rights reserved.
 * Author: Tom Lyon, pugs@cisco.com
 */

#include <linux/device.h>
#include <linux/eventfd.h>
#include <linux/interrupt.h>
#include <linux/iommu.h>
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/notifier.h>
#include <linux/pci.h>
#include <linux/pm_runtime.h>
#include <linux/slab.h>
#include <linux/types.h>
#include <linux/uaccess.h>
#include <linux/vfio.h>

#include "vfio_pci_private.h"

#define DRIVER_VERSION  "0.2"
#define DRIVER_AUTHOR   "Alex Williamson <alex.williamson@redhat.com>"
#define DRIVER_DESC     "VFIO PCI - User Level meta-driver"

static bool nointxmask;
module_param_named(nointxmask, nointxmask, bool, S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(nointxmask,
		  "Disable support for PCI 2.3 style INTx masking.  If this resolves problems for specific devices, report lspci -vvvxxx to linux-pci@vger.kernel.org so the device can be fixed automatically via the broken_intx_masking flag.");

static int vfio_pci_enable(struct vfio_pci_device *vdev)
{
	struct pci_dev *pdev = vdev->pdev;
	int ret;
	u16 cmd;
	u8 msix_pos;

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

	vdev->reset_works = (pci_reset_function(pdev) == 0);
	pci_save_state(pdev);
	vdev->pci_saved_state = pci_store_saved_state(pdev);
	if (!vdev->pci_saved_state)
		pr_debug("%s: Couldn't store %s saved state\n",
			 __func__, dev_name(&pdev->dev));

	ret = vfio_config_init(vdev);
	if (ret) {
		pci_load_and_free_saved_state(pdev, &vdev->pci_saved_state);
		pci_disable_device(pdev);
		return ret;
	}

	if (likely(!nointxmask))
		vdev->pci_2_3 = pci_intx_mask_supported(pdev);

	pci_read_config_word(pdev, PCI_COMMAND, &cmd);
	if (vdev->pci_2_3 && (cmd & PCI_COMMAND_INTX_DISABLE)) {
		cmd &= ~PCI_COMMAND_INTX_DISABLE;
		pci_write_config_word(pdev, PCI_COMMAND, cmd);
	}

	msix_pos = pci_find_capability(pdev, PCI_CAP_ID_MSIX);
	if (msix_pos) {
		u16 flags;
		u32 table;

		pci_read_config_word(pdev, msix_pos + PCI_MSIX_FLAGS, &flags);
		pci_read_config_dword(pdev, msix_pos + PCI_MSIX_TABLE, &table);

		vdev->msix_bar = table & PCI_MSIX_FLAGS_BIRMASK;
		vdev->msix_offset = table & ~PCI_MSIX_FLAGS_BIRMASK;
		vdev->msix_size = ((flags & PCI_MSIX_FLAGS_QSIZE) + 1) * 16;
	} else
		vdev->msix_bar = 0xFF;

	return 0;
}

static void vfio_pci_disable(struct vfio_pci_device *vdev)
{
	struct pci_dev *pdev = vdev->pdev;
	int bar;

	pci_disable_device(pdev);

	vfio_pci_set_irqs_ioctl(vdev, VFIO_IRQ_SET_DATA_NONE |
				VFIO_IRQ_SET_ACTION_TRIGGER,
				vdev->irq_type, 0, 0, NULL);

	vdev->virq_disabled = false;

	vfio_config_free(vdev);

	for (bar = PCI_STD_RESOURCES; bar <= PCI_STD_RESOURCE_END; bar++) {
		if (!vdev->barmap[bar])
			continue;
		pci_iounmap(pdev, vdev->barmap[bar]);
		pci_release_selected_regions(pdev, 1 << bar);
		vdev->barmap[bar] = NULL;
	}

	/*
	 * If we have saved state, restore it.  If we can reset the device,
	 * even better.  Resetting with current state seems better than
	 * nothing, but saving and restoring current state without reset
	 * is just busy work.
	 */
	if (pci_load_and_free_saved_state(pdev, &vdev->pci_saved_state)) {
		pr_info("%s: Couldn't reload %s saved state\n",
			__func__, dev_name(&pdev->dev));

		if (!vdev->reset_works)
			return;

		pci_save_state(pdev);
	}

	/*
	 * Disable INTx and MSI, presumably to avoid spurious interrupts
	 * during reset.  Stolen from pci_reset_function()
	 */
	pci_write_config_word(pdev, PCI_COMMAND, PCI_COMMAND_INTX_DISABLE);

	if (vdev->reset_works)
		__pci_reset_function(pdev);

	pci_restore_state(pdev);
}

static void vfio_pci_release(void *device_data)
{
	struct vfio_pci_device *vdev = device_data;

	if (atomic_dec_and_test(&vdev->refcnt))
		vfio_pci_disable(vdev);

	module_put(THIS_MODULE);
}

static int vfio_pci_open(void *device_data)
{
	struct vfio_pci_device *vdev = device_data;

	if (!try_module_get(THIS_MODULE))
		return -ENODEV;

	if (atomic_inc_return(&vdev->refcnt) == 1) {
		int ret = vfio_pci_enable(vdev);
		if (ret) {
			module_put(THIS_MODULE);
			return ret;
		}
	}

	return 0;
}

static int vfio_pci_get_irq_count(struct vfio_pci_device *vdev, int irq_type)
{
	if (irq_type == VFIO_PCI_INTX_IRQ_INDEX) {
		u8 pin;
		pci_read_config_byte(vdev->pdev, PCI_INTERRUPT_PIN, &pin);
		if (pin)
			return 1;

	} else if (irq_type == VFIO_PCI_MSI_IRQ_INDEX) {
		u8 pos;
		u16 flags;

		pos = pci_find_capability(vdev->pdev, PCI_CAP_ID_MSI);
		if (pos) {
			pci_read_config_word(vdev->pdev,
					     pos + PCI_MSI_FLAGS, &flags);

			return 1 << (flags & PCI_MSI_FLAGS_QMASK);
		}
	} else if (irq_type == VFIO_PCI_MSIX_IRQ_INDEX) {
		u8 pos;
		u16 flags;

		pos = pci_find_capability(vdev->pdev, PCI_CAP_ID_MSIX);
		if (pos) {
			pci_read_config_word(vdev->pdev,
					     pos + PCI_MSIX_FLAGS, &flags);

			return (flags & PCI_MSIX_FLAGS_QSIZE) + 1;
		}
	}

	return 0;
}

static long vfio_pci_ioctl(void *device_data,
			   unsigned int cmd, unsigned long arg)
{
	struct vfio_pci_device *vdev = device_data;
	unsigned long minsz;

	if (cmd == VFIO_DEVICE_GET_INFO) {
		struct vfio_device_info info;

		minsz = offsetofend(struct vfio_device_info, num_irqs);

		if (copy_from_user(&info, (void __user *)arg, minsz))
			return -EFAULT;

		if (info.argsz < minsz)
			return -EINVAL;

		info.flags = VFIO_DEVICE_FLAGS_PCI;

		if (vdev->reset_works)
			info.flags |= VFIO_DEVICE_FLAGS_RESET;

		info.num_regions = VFIO_PCI_NUM_REGIONS;
		info.num_irqs = VFIO_PCI_NUM_IRQS;

		return copy_to_user((void __user *)arg, &info, minsz);

	} else if (cmd == VFIO_DEVICE_GET_REGION_INFO) {
		struct pci_dev *pdev = vdev->pdev;
		struct vfio_region_info info;

		minsz = offsetofend(struct vfio_region_info, offset);

		if (copy_from_user(&info, (void __user *)arg, minsz))
			return -EFAULT;

		if (info.argsz < minsz)
			return -EINVAL;

		switch (info.index) {
		case VFIO_PCI_CONFIG_REGION_INDEX:
			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
			info.size = pdev->cfg_size;
			info.flags = VFIO_REGION_INFO_FLAG_READ |
				     VFIO_REGION_INFO_FLAG_WRITE;
			break;
		case VFIO_PCI_BAR0_REGION_INDEX ... VFIO_PCI_BAR5_REGION_INDEX:
			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
			info.size = pci_resource_len(pdev, info.index);
			if (!info.size) {
				info.flags = 0;
				break;
			}

			info.flags = VFIO_REGION_INFO_FLAG_READ |
				     VFIO_REGION_INFO_FLAG_WRITE;
			if (pci_resource_flags(pdev, info.index) &
			    IORESOURCE_MEM && info.size >= PAGE_SIZE)
				info.flags |= VFIO_REGION_INFO_FLAG_MMAP;
			break;
		case VFIO_PCI_ROM_REGION_INDEX:
		{
			void __iomem *io;
			size_t size;

			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
			info.flags = 0;

			/* Report the BAR size, not the ROM size */
			info.size = pci_resource_len(pdev, info.index);
			if (!info.size)
				break;

			/* Is it really there? */
			io = pci_map_rom(pdev, &size);
			if (!io || !size) {
				info.size = 0;
				break;
			}
			pci_unmap_rom(pdev, io);

			info.flags = VFIO_REGION_INFO_FLAG_READ;
			break;
		}
		default:
			return -EINVAL;
		}

		return copy_to_user((void __user *)arg, &info, minsz);

	} else if (cmd == VFIO_DEVICE_GET_IRQ_INFO) {
		struct vfio_irq_info info;

		minsz = offsetofend(struct vfio_irq_info, count);

		if (copy_from_user(&info, (void __user *)arg, minsz))
			return -EFAULT;

		if (info.argsz < minsz || info.index >= VFIO_PCI_NUM_IRQS)
			return -EINVAL;

		info.flags = VFIO_IRQ_INFO_EVENTFD;

		info.count = vfio_pci_get_irq_count(vdev, info.index);

		if (info.index == VFIO_PCI_INTX_IRQ_INDEX)
			info.flags |= (VFIO_IRQ_INFO_MASKABLE |
				       VFIO_IRQ_INFO_AUTOMASKED);
		else
			info.flags |= VFIO_IRQ_INFO_NORESIZE;

		return copy_to_user((void __user *)arg, &info, minsz);

	} else if (cmd == VFIO_DEVICE_SET_IRQS) {
		struct vfio_irq_set hdr;
		u8 *data = NULL;
		int ret = 0;

		minsz = offsetofend(struct vfio_irq_set, count);

		if (copy_from_user(&hdr, (void __user *)arg, minsz))
			return -EFAULT;

		if (hdr.argsz < minsz || hdr.index >= VFIO_PCI_NUM_IRQS ||
		    hdr.flags & ~(VFIO_IRQ_SET_DATA_TYPE_MASK |
				  VFIO_IRQ_SET_ACTION_TYPE_MASK))
			return -EINVAL;

		if (!(hdr.flags & VFIO_IRQ_SET_DATA_NONE)) {
			size_t size;

			if (hdr.flags & VFIO_IRQ_SET_DATA_BOOL)
				size = sizeof(uint8_t);
			else if (hdr.flags & VFIO_IRQ_SET_DATA_EVENTFD)
				size = sizeof(int32_t);
			else
				return -EINVAL;

			if (hdr.argsz - minsz < hdr.count * size ||
			    hdr.count > vfio_pci_get_irq_count(vdev, hdr.index))
				return -EINVAL;

			data = memdup_user((void __user *)(arg + minsz),
					   hdr.count * size);
			if (IS_ERR(data))
				return PTR_ERR(data);
		}

		mutex_lock(&vdev->igate);

		ret = vfio_pci_set_irqs_ioctl(vdev, hdr.flags, hdr.index,
					      hdr.start, hdr.count, data);

		mutex_unlock(&vdev->igate);
		kfree(data);

		return ret;

	} else if (cmd == VFIO_DEVICE_RESET)
		return vdev->reset_works ?
			pci_reset_function(vdev->pdev) : -EINVAL;

	return -ENOTTY;
}

static ssize_t vfio_pci_rw(void *device_data, char __user *buf,
			   size_t count, loff_t *ppos, bool iswrite)
{
	unsigned int index = VFIO_PCI_OFFSET_TO_INDEX(*ppos);
	struct vfio_pci_device *vdev = device_data;

	if (index >= VFIO_PCI_NUM_REGIONS)
		return -EINVAL;

	switch (index) {
	case VFIO_PCI_CONFIG_REGION_INDEX:
		return vfio_pci_config_rw(vdev, buf, count, ppos, iswrite);

	case VFIO_PCI_ROM_REGION_INDEX:
		if (iswrite)
			return -EINVAL;
		return vfio_pci_bar_rw(vdev, buf, count, ppos, false);

	case VFIO_PCI_BAR0_REGION_INDEX ... VFIO_PCI_BAR5_REGION_INDEX:
		return vfio_pci_bar_rw(vdev, buf, count, ppos, iswrite);
	}

	return -EINVAL;
}

static ssize_t vfio_pci_read(void *device_data, char __user *buf,
			     size_t count, loff_t *ppos)
{
	if (!count)
		return 0;

	return vfio_pci_rw(device_data, buf, count, ppos, false);
}

static ssize_t vfio_pci_write(void *device_data, const char __user *buf,
			      size_t count, loff_t *ppos)
{
	if (!count)
		return 0;

	return vfio_pci_rw(device_data, (char __user *)buf, count, ppos, true);
}

static int vfio_pci_mmap(void *device_data, struct vm_area_struct *vma)
{
	struct vfio_pci_device *vdev = device_data;
	struct pci_dev *pdev = vdev->pdev;
	unsigned int index;
	u64 phys_len, req_len, pgoff, req_start;
	int ret;

	index = vma->vm_pgoff >> (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT);

	if (vma->vm_end < vma->vm_start)
		return -EINVAL;
	if ((vma->vm_flags & VM_SHARED) == 0)
		return -EINVAL;
	if (index >= VFIO_PCI_ROM_REGION_INDEX)
		return -EINVAL;
	if (!(pci_resource_flags(pdev, index) & IORESOURCE_MEM))
		return -EINVAL;

	phys_len = pci_resource_len(pdev, index);
	req_len = vma->vm_end - vma->vm_start;
	pgoff = vma->vm_pgoff &
		((1U << (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT)) - 1);
	req_start = pgoff << PAGE_SHIFT;

	if (phys_len < PAGE_SIZE || req_start + req_len > phys_len)
		return -EINVAL;

	if (index == vdev->msix_bar) {
		/*
		 * Disallow mmaps overlapping the MSI-X table; users don't
		 * get to touch this directly.  We could find somewhere
		 * else to map the overlap, but page granularity is only
		 * a recommendation, not a requirement, so the user needs
		 * to know which bits are real.  Requiring them to mmap
		 * around the table makes that clear.
		 */

		/* If neither entirely above nor below, then it overlaps */
		if (!(req_start >= vdev->msix_offset + vdev->msix_size ||
		      req_start + req_len <= vdev->msix_offset))
			return -EINVAL;
	}

	/*
	 * Even though we don't make use of the barmap for the mmap,
	 * we need to request the region and the barmap tracks that.
	 */
	if (!vdev->barmap[index]) {
		ret = pci_request_selected_regions(pdev,
						   1 << index, "vfio-pci");
		if (ret)
			return ret;

		vdev->barmap[index] = pci_iomap(pdev, index, 0);
	}

	vma->vm_private_data = vdev;
	vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP;
	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
	vma->vm_pgoff = (pci_resource_start(pdev, index) >> PAGE_SHIFT) + pgoff;

	return remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
			       req_len, vma->vm_page_prot);
}

static const struct vfio_device_ops vfio_pci_ops = {
	.name		= "vfio-pci",
	.open		= vfio_pci_open,
	.release	= vfio_pci_release,
	.ioctl		= vfio_pci_ioctl,
	.read		= vfio_pci_read,
	.write		= vfio_pci_write,
	.mmap		= vfio_pci_mmap,
};

static int vfio_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
{
	u8 type;
	struct vfio_pci_device *vdev;
	struct iommu_group *group;
	int ret;

	pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
	if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
		return -EINVAL;

	group = iommu_group_get(&pdev->dev);
	if (!group)
		return -EINVAL;

	vdev = kzalloc(sizeof(*vdev), GFP_KERNEL);
	if (!vdev) {
		iommu_group_put(group);
		return -ENOMEM;
	}

	vdev->pdev = pdev;
	vdev->irq_type = VFIO_PCI_NUM_IRQS;
	mutex_init(&vdev->igate);
	spin_lock_init(&vdev->irqlock);
	atomic_set(&vdev->refcnt, 0);

	ret = vfio_add_group_dev(&pdev->dev, &vfio_pci_ops, vdev);
	if (ret) {
		iommu_group_put(group);
		kfree(vdev);
	}

	return ret;
}

static void vfio_pci_remove(struct pci_dev *pdev)
{
	struct vfio_pci_device *vdev;

	vdev = vfio_del_group_dev(&pdev->dev);
	if (!vdev)
		return;

	iommu_group_put(pdev->dev.iommu_group);
	kfree(vdev);
}

static struct pci_driver vfio_pci_driver = {
	.name		= "vfio-pci",
	.id_table	= NULL, /* only dynamic ids */
	.probe		= vfio_pci_probe,
	.remove		= vfio_pci_remove,
};

static void __exit vfio_pci_cleanup(void)
{
	pci_unregister_driver(&vfio_pci_driver);
	vfio_pci_virqfd_exit();
	vfio_pci_uninit_perm_bits();
}

static int __init vfio_pci_init(void)
{
	int ret;

	/* Allocate shared config space permision data used by all devices */
	ret = vfio_pci_init_perm_bits();
	if (ret)
		return ret;

	/* Start the virqfd cleanup handler */
	ret = vfio_pci_virqfd_init();
	if (ret)
		goto out_virqfd;

	/* Register and scan for devices */
	ret = pci_register_driver(&vfio_pci_driver);
	if (ret)
		goto out_driver;

	return 0;

out_driver:
	vfio_pci_virqfd_exit();
out_virqfd:
	vfio_pci_uninit_perm_bits();
	return ret;
}

module_init(vfio_pci_init);
module_exit(vfio_pci_cleanup);

MODULE_VERSION(DRIVER_VERSION);
MODULE_LICENSE("GPL v2");
MODULE_AUTHOR(DRIVER_AUTHOR);
MODULE_DESCRIPTION(DRIVER_DESC);
Commit	Line	Data
89e1f7d4 AW	1	/*
	2	* Copyright (C) 2012 Red Hat, Inc. All rights reserved.
	3	* Author: Alex Williamson <alex.williamson@redhat.com>
	4	*
	5	* This program is free software; you can redistribute it and/or modify
	6	* it under the terms of the GNU General Public License version 2 as
	7	* published by the Free Software Foundation.
	8	*
	9	* Derived from original vfio:
	10	* Copyright 2010 Cisco Systems, Inc. All rights reserved.
	11	* Author: Tom Lyon, pugs@cisco.com
	12	*/
	13
	14	#include <linux/device.h>
	15	#include <linux/eventfd.h>
	16	#include <linux/interrupt.h>
	17	#include <linux/iommu.h>
	18	#include <linux/module.h>
	19	#include <linux/mutex.h>
	20	#include <linux/notifier.h>
	21	#include <linux/pci.h>
	22	#include <linux/pm_runtime.h>
	23	#include <linux/slab.h>
	24	#include <linux/types.h>
	25	#include <linux/uaccess.h>
	26	#include <linux/vfio.h>
	27
	28	#include "vfio_pci_private.h"
	29
	30	#define DRIVER_VERSION "0.2"
	31	#define DRIVER_AUTHOR "Alex Williamson <alex.williamson@redhat.com>"
	32	#define DRIVER_DESC "VFIO PCI - User Level meta-driver"
	33
	34	static bool nointxmask;
	35	module_param_named(nointxmask, nointxmask, bool, S_IRUGO \| S_IWUSR);
	36	MODULE_PARM_DESC(nointxmask,
	37	"Disable support for PCI 2.3 style INTx masking. If this resolves problems for specific devices, report lspci -vvvxxx to linux-pci@vger.kernel.org so the device can be fixed automatically via the broken_intx_masking flag.");
	38
	39	static int vfio_pci_enable(struct vfio_pci_device *vdev)
	40	{
	41	struct pci_dev *pdev = vdev->pdev;
	42	int ret;
	43	u16 cmd;
	44	u8 msix_pos;
	45
9a92c509 AW	46	ret = pci_enable_device(pdev);
	47	if (ret)
	48	return ret;
	49
89e1f7d4 AW	50	vdev->reset_works = (pci_reset_function(pdev) == 0);
	51	pci_save_state(pdev);
	52	vdev->pci_saved_state = pci_store_saved_state(pdev);
	53	if (!vdev->pci_saved_state)
	54	pr_debug("%s: Couldn't store %s saved state\n",
	55	__func__, dev_name(&pdev->dev));
	56
	57	ret = vfio_config_init(vdev);
9a92c509 AW	58	if (ret) {
	59	pci_load_and_free_saved_state(pdev, &vdev->pci_saved_state);
	60	pci_disable_device(pdev);
	61	return ret;
	62	}
89e1f7d4 AW	63
	64	if (likely(!nointxmask))
	65	vdev->pci_2_3 = pci_intx_mask_supported(pdev);
	66
	67	pci_read_config_word(pdev, PCI_COMMAND, &cmd);
	68	if (vdev->pci_2_3 && (cmd & PCI_COMMAND_INTX_DISABLE)) {
	69	cmd &= ~PCI_COMMAND_INTX_DISABLE;
	70	pci_write_config_word(pdev, PCI_COMMAND, cmd);
	71	}
	72
	73	msix_pos = pci_find_capability(pdev, PCI_CAP_ID_MSIX);
	74	if (msix_pos) {
	75	u16 flags;
	76	u32 table;
	77
	78	pci_read_config_word(pdev, msix_pos + PCI_MSIX_FLAGS, &flags);
	79	pci_read_config_dword(pdev, msix_pos + PCI_MSIX_TABLE, &table);
	80
	81	vdev->msix_bar = table & PCI_MSIX_FLAGS_BIRMASK;
	82	vdev->msix_offset = table & ~PCI_MSIX_FLAGS_BIRMASK;
	83	vdev->msix_size = ((flags & PCI_MSIX_FLAGS_QSIZE) + 1) * 16;
	84	} else
	85	vdev->msix_bar = 0xFF;
	86
9a92c509	87	return 0;
89e1f7d4 AW	88	}
	89
	90	static void vfio_pci_disable(struct vfio_pci_device *vdev)
	91	{
2007722a	92	struct pci_dev *pdev = vdev->pdev;
89e1f7d4 AW	93	int bar;
89e1f7d4 AW	94
2007722a	95	pci_disable_device(pdev);
89e1f7d4 AW	96
	97	vfio_pci_set_irqs_ioctl(vdev, VFIO_IRQ_SET_DATA_NONE \|
	98	VFIO_IRQ_SET_ACTION_TRIGGER,
	99	vdev->irq_type, 0, 0, NULL);
	100
	101	vdev->virq_disabled = false;
	102
	103	vfio_config_free(vdev);
	104
89e1f7d4 AW	105	for (bar = PCI_STD_RESOURCES; bar <= PCI_STD_RESOURCE_END; bar++) {
	106	if (!vdev->barmap[bar])
	107	continue;
2007722a AW	108	pci_iounmap(pdev, vdev->barmap[bar]);
2007722a AW	109	pci_release_selected_regions(pdev, 1 << bar);
89e1f7d4 AW	110	vdev->barmap[bar] = NULL;
89e1f7d4 AW	111	}
2007722a AW	112
	113	/*
	114	* If we have saved state, restore it. If we can reset the device,
	115	* even better. Resetting with current state seems better than
	116	* nothing, but saving and restoring current state without reset
	117	* is just busy work.
	118	*/
	119	if (pci_load_and_free_saved_state(pdev, &vdev->pci_saved_state)) {
	120	pr_info("%s: Couldn't reload %s saved state\n",
	121	__func__, dev_name(&pdev->dev));
	122
	123	if (!vdev->reset_works)
	124	return;
	125
	126	pci_save_state(pdev);
	127	}
	128
	129	/*
	130	* Disable INTx and MSI, presumably to avoid spurious interrupts
	131	* during reset. Stolen from pci_reset_function()
	132	*/
	133	pci_write_config_word(pdev, PCI_COMMAND, PCI_COMMAND_INTX_DISABLE);
	134
	135	if (vdev->reset_works)
	136	__pci_reset_function(pdev);
	137
	138	pci_restore_state(pdev);
89e1f7d4 AW	139	}
	140
	141	static void vfio_pci_release(void *device_data)
	142	{
	143	struct vfio_pci_device *vdev = device_data;
	144
	145	if (atomic_dec_and_test(&vdev->refcnt))
	146	vfio_pci_disable(vdev);
	147
	148	module_put(THIS_MODULE);
	149	}
	150
	151	static int vfio_pci_open(void *device_data)
	152	{
	153	struct vfio_pci_device *vdev = device_data;
	154
	155	if (!try_module_get(THIS_MODULE))
	156	return -ENODEV;
	157
	158	if (atomic_inc_return(&vdev->refcnt) == 1) {
	159	int ret = vfio_pci_enable(vdev);
	160	if (ret) {
	161	module_put(THIS_MODULE);
	162	return ret;
	163	}
	164	}
	165
	166	return 0;
	167	}
	168
	169	static int vfio_pci_get_irq_count(struct vfio_pci_device *vdev, int irq_type)
	170	{
	171	if (irq_type == VFIO_PCI_INTX_IRQ_INDEX) {
	172	u8 pin;
	173	pci_read_config_byte(vdev->pdev, PCI_INTERRUPT_PIN, &pin);
	174	if (pin)
	175	return 1;
	176
	177	} else if (irq_type == VFIO_PCI_MSI_IRQ_INDEX) {
	178	u8 pos;
	179	u16 flags;
	180
	181	pos = pci_find_capability(vdev->pdev, PCI_CAP_ID_MSI);
	182	if (pos) {
	183	pci_read_config_word(vdev->pdev,
	184	pos + PCI_MSI_FLAGS, &flags);
	185
	186	return 1 << (flags & PCI_MSI_FLAGS_QMASK);
	187	}
	188	} else if (irq_type == VFIO_PCI_MSIX_IRQ_INDEX) {
	189	u8 pos;
	190	u16 flags;
	191
	192	pos = pci_find_capability(vdev->pdev, PCI_CAP_ID_MSIX);
	193	if (pos) {
	194	pci_read_config_word(vdev->pdev,
	195	pos + PCI_MSIX_FLAGS, &flags);
	196
	197	return (flags & PCI_MSIX_FLAGS_QSIZE) + 1;
	198	}
	199	}
	200
	201	return 0;
	202	}
203
204	static long vfio_pci_ioctl(void *device_data,
205	unsigned int cmd, unsigned long arg)
206	{
207	struct vfio_pci_device *vdev = device_data;
208	unsigned long minsz;
209
210	if (cmd == VFIO_DEVICE_GET_INFO) {
211	struct vfio_device_info info;
212
213	minsz = offsetofend(struct vfio_device_info, num_irqs);
214
215	if (copy_from_user(&info, (void __user *)arg, minsz))
216	return -EFAULT;
217
218	if (info.argsz < minsz)
219	return -EINVAL;
220
221	info.flags = VFIO_DEVICE_FLAGS_PCI;
222
223	if (vdev->reset_works)
224	info.flags \|= VFIO_DEVICE_FLAGS_RESET;
225
226	info.num_regions = VFIO_PCI_NUM_REGIONS;
227	info.num_irqs = VFIO_PCI_NUM_IRQS;
228
229	return copy_to_user((void __user *)arg, &info, minsz);
230
231	} else if (cmd == VFIO_DEVICE_GET_REGION_INFO) {
232	struct pci_dev *pdev = vdev->pdev;
233	struct vfio_region_info info;
234
235	minsz = offsetofend(struct vfio_region_info, offset);
236
237	if (copy_from_user(&info, (void __user *)arg, minsz))
238	return -EFAULT;
239
240	if (info.argsz < minsz)
241	return -EINVAL;
242
243	switch (info.index) {
244	case VFIO_PCI_CONFIG_REGION_INDEX:
245	info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
246	info.size = pdev->cfg_size;
247	info.flags = VFIO_REGION_INFO_FLAG_READ \|
248	VFIO_REGION_INFO_FLAG_WRITE;
249	break;
250	case VFIO_PCI_BAR0_REGION_INDEX ... VFIO_PCI_BAR5_REGION_INDEX:
251	info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
252	info.size = pci_resource_len(pdev, info.index);
253	if (!info.size) {
254	info.flags = 0;
255	break;
256	}
257
258	info.flags = VFIO_REGION_INFO_FLAG_READ \|
259	VFIO_REGION_INFO_FLAG_WRITE;
260	if (pci_resource_flags(pdev, info.index) &
261	IORESOURCE_MEM && info.size >= PAGE_SIZE)
262	info.flags \|= VFIO_REGION_INFO_FLAG_MMAP;
263	break;
264	case VFIO_PCI_ROM_REGION_INDEX:
265	{
266	void __iomem *io;
267	size_t size;
268
269	info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
270	info.flags = 0;
271
272	/* Report the BAR size, not the ROM size */
273	info.size = pci_resource_len(pdev, info.index);
274	if (!info.size)
275	break;
276
277	/* Is it really there? */
278	io = pci_map_rom(pdev, &size);
279	if (!io \|\| !size) {
280	info.size = 0;
281	break;
282	}
283	pci_unmap_rom(pdev, io);
284
285	info.flags = VFIO_REGION_INFO_FLAG_READ;
286	break;
287	}
288	default:
289	return -EINVAL;
290	}
291
292	return copy_to_user((void __user *)arg, &info, minsz);
293
294	} else if (cmd == VFIO_DEVICE_GET_IRQ_INFO) {
295	struct vfio_irq_info info;
296
297	minsz = offsetofend(struct vfio_irq_info, count);
298
299	if (copy_from_user(&info, (void __user *)arg, minsz))
300	return -EFAULT;
301
302	if (info.argsz < minsz \|\| info.index >= VFIO_PCI_NUM_IRQS)
303	return -EINVAL;
304
305	info.flags = VFIO_IRQ_INFO_EVENTFD;
306
307	info.count = vfio_pci_get_irq_count(vdev, info.index);
308
309	if (info.index == VFIO_PCI_INTX_IRQ_INDEX)
310	info.flags \|= (VFIO_IRQ_INFO_MASKABLE \|
311	VFIO_IRQ_INFO_AUTOMASKED);
312	else
313	info.flags \|= VFIO_IRQ_INFO_NORESIZE;
314
315	return copy_to_user((void __user *)arg, &info, minsz);
316
317	} else if (cmd == VFIO_DEVICE_SET_IRQS) {
318	struct vfio_irq_set hdr;
319	u8 *data = NULL;
320	int ret = 0;
321
322	minsz = offsetofend(struct vfio_irq_set, count);
323
324	if (copy_from_user(&hdr, (void __user *)arg, minsz))
325	return -EFAULT;
326
327	if (hdr.argsz < minsz \|\| hdr.index >= VFIO_PCI_NUM_IRQS \|\|
328	hdr.flags & ~(VFIO_IRQ_SET_DATA_TYPE_MASK \|
329	VFIO_IRQ_SET_ACTION_TYPE_MASK))
330	return -EINVAL;
331
332	if (!(hdr.flags & VFIO_IRQ_SET_DATA_NONE)) {
333	size_t size;
334
335	if (hdr.flags & VFIO_IRQ_SET_DATA_BOOL)
336	size = sizeof(uint8_t);
337	else if (hdr.flags & VFIO_IRQ_SET_DATA_EVENTFD)
338	size = sizeof(int32_t);
339	else
340	return -EINVAL;
341
342	if (hdr.argsz - minsz < hdr.count * size \|\|
343	hdr.count > vfio_pci_get_irq_count(vdev, hdr.index))
344	return -EINVAL;
345
3a1f7041 FW	346	data = memdup_user((void __user *)(arg + minsz),
	347	hdr.count * size);
	348	if (IS_ERR(data))
	349	return PTR_ERR(data);
89e1f7d4 AW	350	}
	351
	352	mutex_lock(&vdev->igate);
	353
	354	ret = vfio_pci_set_irqs_ioctl(vdev, hdr.flags, hdr.index,
	355	hdr.start, hdr.count, data);
	356
	357	mutex_unlock(&vdev->igate);
	358	kfree(data);
	359
	360	return ret;
	361
	362	} else if (cmd == VFIO_DEVICE_RESET)
	363	return vdev->reset_works ?
	364	pci_reset_function(vdev->pdev) : -EINVAL;
	365
	366	return -ENOTTY;
	367	}
	368
5b279a11 AW	369	static ssize_t vfio_pci_rw(void device_data, char __user buf,
5b279a11 AW	370	size_t count, loff_t *ppos, bool iswrite)
89e1f7d4 AW	371	{
	372	unsigned int index = VFIO_PCI_OFFSET_TO_INDEX(*ppos);
	373	struct vfio_pci_device *vdev = device_data;
89e1f7d4 AW	374
	375	if (index >= VFIO_PCI_NUM_REGIONS)
	376	return -EINVAL;
	377
5b279a11 AW	378	switch (index) {
5b279a11 AW	379	case VFIO_PCI_CONFIG_REGION_INDEX:
906ee99d AW	380	return vfio_pci_config_rw(vdev, buf, count, ppos, iswrite);
906ee99d AW	381
5b279a11 AW	382	case VFIO_PCI_ROM_REGION_INDEX:
	383	if (iswrite)
	384	return -EINVAL;
906ee99d	385	return vfio_pci_bar_rw(vdev, buf, count, ppos, false);
89e1f7d4	386
5b279a11	387	case VFIO_PCI_BAR0_REGION_INDEX ... VFIO_PCI_BAR5_REGION_INDEX:
906ee99d	388	return vfio_pci_bar_rw(vdev, buf, count, ppos, iswrite);
5b279a11 AW	389	}
5b279a11 AW	390
89e1f7d4 AW	391	return -EINVAL;
	392	}
	393
5b279a11 AW	394	static ssize_t vfio_pci_read(void device_data, char __user buf,
	395	size_t count, loff_t *ppos)
	396	{
906ee99d AW	397	if (!count)
	398	return 0;
	399
5b279a11 AW	400	return vfio_pci_rw(device_data, buf, count, ppos, false);
	401	}
	402
89e1f7d4 AW	403	static ssize_t vfio_pci_write(void device_data, const char __user buf,
	404	size_t count, loff_t *ppos)
	405	{
906ee99d AW	406	if (!count)
	407	return 0;
	408
	409	return vfio_pci_rw(device_data, (char __user *)buf, count, ppos, true);
89e1f7d4 AW	410	}
	411
	412	static int vfio_pci_mmap(void device_data, struct vm_area_struct vma)
	413	{
	414	struct vfio_pci_device *vdev = device_data;
	415	struct pci_dev *pdev = vdev->pdev;
	416	unsigned int index;
34002f54	417	u64 phys_len, req_len, pgoff, req_start;
89e1f7d4 AW	418	int ret;
	419
	420	index = vma->vm_pgoff >> (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT);
	421
	422	if (vma->vm_end < vma->vm_start)
	423	return -EINVAL;
	424	if ((vma->vm_flags & VM_SHARED) == 0)
	425	return -EINVAL;
	426	if (index >= VFIO_PCI_ROM_REGION_INDEX)
	427	return -EINVAL;
	428	if (!(pci_resource_flags(pdev, index) & IORESOURCE_MEM))
	429	return -EINVAL;
	430
	431	phys_len = pci_resource_len(pdev, index);
	432	req_len = vma->vm_end - vma->vm_start;
	433	pgoff = vma->vm_pgoff &
	434	((1U << (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT)) - 1);
	435	req_start = pgoff << PAGE_SHIFT;
	436
	437	if (phys_len < PAGE_SIZE \|\| req_start + req_len > phys_len)
	438	return -EINVAL;
	439
	440	if (index == vdev->msix_bar) {
	441	/*
	442	* Disallow mmaps overlapping the MSI-X table; users don't
	443	* get to touch this directly. We could find somewhere
	444	* else to map the overlap, but page granularity is only
	445	* a recommendation, not a requirement, so the user needs
	446	* to know which bits are real. Requiring them to mmap
	447	* around the table makes that clear.
	448	*/
	449
	450	/* If neither entirely above nor below, then it overlaps */
	451	if (!(req_start >= vdev->msix_offset + vdev->msix_size \|\|
	452	req_start + req_len <= vdev->msix_offset))
	453	return -EINVAL;
	454	}
	455
	456	/*
	457	* Even though we don't make use of the barmap for the mmap,
	458	* we need to request the region and the barmap tracks that.
	459	*/
	460	if (!vdev->barmap[index]) {
	461	ret = pci_request_selected_regions(pdev,
	462	1 << index, "vfio-pci");
	463	if (ret)
	464	return ret;
	465
	466	vdev->barmap[index] = pci_iomap(pdev, index, 0);
	467	}
	468
	469	vma->vm_private_data = vdev;
547b1e81	470	vma->vm_flags \|= VM_IO \| VM_DONTEXPAND \| VM_DONTDUMP;
89e1f7d4	471	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
34002f54	472	vma->vm_pgoff = (pci_resource_start(pdev, index) >> PAGE_SHIFT) + pgoff;
89e1f7d4	473
34002f54	474	return remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
89e1f7d4 AW	475	req_len, vma->vm_page_prot);
	476	}
	477
	478	static const struct vfio_device_ops vfio_pci_ops = {
	479	.name = "vfio-pci",
	480	.open = vfio_pci_open,
	481	.release = vfio_pci_release,
	482	.ioctl = vfio_pci_ioctl,
	483	.read = vfio_pci_read,
	484	.write = vfio_pci_write,
	485	.mmap = vfio_pci_mmap,
	486	};
	487
	488	static int vfio_pci_probe(struct pci_dev pdev, const struct pci_device_id id)
	489	{
	490	u8 type;
	491	struct vfio_pci_device *vdev;
	492	struct iommu_group *group;
	493	int ret;
	494
	495	pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
	496	if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
	497	return -EINVAL;
	498
	499	group = iommu_group_get(&pdev->dev);
	500	if (!group)
	501	return -EINVAL;
	502
	503	vdev = kzalloc(sizeof(*vdev), GFP_KERNEL);
	504	if (!vdev) {
	505	iommu_group_put(group);
	506	return -ENOMEM;
	507	}
	508
	509	vdev->pdev = pdev;
	510	vdev->irq_type = VFIO_PCI_NUM_IRQS;
	511	mutex_init(&vdev->igate);
	512	spin_lock_init(&vdev->irqlock);
	513	atomic_set(&vdev->refcnt, 0);
	514
	515	ret = vfio_add_group_dev(&pdev->dev, &vfio_pci_ops, vdev);
	516	if (ret) {
	517	iommu_group_put(group);
	518	kfree(vdev);
	519	}
	520
	521	return ret;
	522	}
	523
	524	static void vfio_pci_remove(struct pci_dev *pdev)
	525	{
	526	struct vfio_pci_device *vdev;
	527
	528	vdev = vfio_del_group_dev(&pdev->dev);
	529	if (!vdev)
	530	return;
	531
	532	iommu_group_put(pdev->dev.iommu_group);
	533	kfree(vdev);
	534	}
	535
	536	static struct pci_driver vfio_pci_driver = {
	537	.name = "vfio-pci",
	538	.id_table = NULL, /* only dynamic ids */
539	.probe = vfio_pci_probe,
540	.remove = vfio_pci_remove,
541	};
542
543	static void __exit vfio_pci_cleanup(void)
544	{
545	pci_unregister_driver(&vfio_pci_driver);
546	vfio_pci_virqfd_exit();
547	vfio_pci_uninit_perm_bits();
548	}
549
550	static int __init vfio_pci_init(void)
551	{
552	int ret;
553
554	/* Allocate shared config space permision data used by all devices */
555	ret = vfio_pci_init_perm_bits();
556	if (ret)
557	return ret;
558
559	/* Start the virqfd cleanup handler */
560	ret = vfio_pci_virqfd_init();
561	if (ret)
562	goto out_virqfd;
563
564	/* Register and scan for devices */
565	ret = pci_register_driver(&vfio_pci_driver);
566	if (ret)
567	goto out_driver;
568
569	return 0;
570
89e1f7d4	571	out_driver:
05bf3aac JL	572	vfio_pci_virqfd_exit();
05bf3aac JL	573	out_virqfd:
89e1f7d4 AW	574	vfio_pci_uninit_perm_bits();
	575	return ret;
	576	}
	577
	578	module_init(vfio_pci_init);
	579	module_exit(vfio_pci_cleanup);
	580
	581	MODULE_VERSION(DRIVER_VERSION);
	582	MODULE_LICENSE("GPL v2");
	583	MODULE_AUTHOR(DRIVER_AUTHOR);
	584	MODULE_DESCRIPTION(DRIVER_DESC);