[mirror_ubuntu-artful-kernel.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 = pdev->msix_cap;
	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_TABLE_BIR;
		vdev->msix_offset = table & PCI_MSIX_TABLE_OFFSET;
		vdev->msix_size = ((flags & PCI_MSIX_FLAGS_QSIZE) + 1) * 16;
	} else
		vdev->msix_bar = 0xFF;

#ifdef CONFIG_VFIO_PCI_VGA
	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA)
		vdev->has_vga = true;
#endif

	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);

	/*
	 * Careful, device_lock may already be held.  This is the case if
	 * a driver unbind is blocked.  Try to get the locks ourselves to
	 * prevent a deadlock.
	 */
	if (vdev->reset_works) {
		bool reset_done = false;

		if (pci_cfg_access_trylock(pdev)) {
			if (device_trylock(&pdev->dev)) {
				__pci_reset_function_locked(pdev);
				reset_done = true;
				device_unlock(&pdev->dev);
			}
			pci_cfg_access_unlock(pdev);
		}

		if (!reset_done)
			pr_warn("%s: Unable to acquire locks for reset of %s\n",
				__func__, dev_name(&pdev->dev));
	}

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

			return (flags & PCI_MSIX_FLAGS_QSIZE) + 1;
		}
	} else if (irq_type == VFIO_PCI_ERR_IRQ_INDEX)
		if (pci_is_pcie(vdev->pdev))
			return 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;
		}
		case VFIO_PCI_VGA_REGION_INDEX:
			if (!vdev->has_vga)
				return -EINVAL;

			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
			info.size = 0xc0000;
			info.flags = VFIO_REGION_INFO_FLAG_READ |
				     VFIO_REGION_INFO_FLAG_WRITE;

			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;

		switch (info.index) {
		case VFIO_PCI_INTX_IRQ_INDEX ... VFIO_PCI_MSIX_IRQ_INDEX:
			break;
		case VFIO_PCI_ERR_IRQ_INDEX:
			if (pci_is_pcie(vdev->pdev))
				break;
		/* pass thru to return error */
		default:
			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;
			int max = vfio_pci_get_irq_count(vdev, hdr.index);

			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.start >= max || hdr.start + hdr.count > max)
				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);

	case VFIO_PCI_VGA_REGION_INDEX:
		return vfio_pci_vga_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_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 pci_ers_result_t vfio_pci_aer_err_detected(struct pci_dev *pdev,
						  pci_channel_state_t state)
{
	struct vfio_pci_device *vdev;
	struct vfio_device *device;

	device = vfio_device_get_from_dev(&pdev->dev);
	if (device == NULL)
		return PCI_ERS_RESULT_DISCONNECT;

	vdev = vfio_device_data(device);
	if (vdev == NULL) {
		vfio_device_put(device);
		return PCI_ERS_RESULT_DISCONNECT;
	}

	if (vdev->err_trigger)
		eventfd_signal(vdev->err_trigger, 1);

	vfio_device_put(device);

	return PCI_ERS_RESULT_CAN_RECOVER;
}

static struct pci_error_handlers vfio_err_handlers = {
	.error_detected = vfio_pci_aer_err_detected,
};

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

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
a9047f24	73	msix_pos = pdev->msix_cap;
89e1f7d4 AW	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
508d1aa6 BH	81	vdev->msix_bar = table & PCI_MSIX_TABLE_BIR;
508d1aa6 BH	82	vdev->msix_offset = table & PCI_MSIX_TABLE_OFFSET;
89e1f7d4 AW	83	vdev->msix_size = ((flags & PCI_MSIX_FLAGS_QSIZE) + 1) * 16;
	84	} else
	85	vdev->msix_bar = 0xFF;
	86
84237a82 AW	87	#ifdef CONFIG_VFIO_PCI_VGA
	88	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA)
	89	vdev->has_vga = true;
	90	#endif
	91
9a92c509	92	return 0;
89e1f7d4 AW	93	}
	94
	95	static void vfio_pci_disable(struct vfio_pci_device *vdev)
	96	{
2007722a	97	struct pci_dev *pdev = vdev->pdev;
89e1f7d4 AW	98	int bar;
89e1f7d4 AW	99
2007722a	100	pci_disable_device(pdev);
89e1f7d4 AW	101
	102	vfio_pci_set_irqs_ioctl(vdev, VFIO_IRQ_SET_DATA_NONE \|
	103	VFIO_IRQ_SET_ACTION_TRIGGER,
	104	vdev->irq_type, 0, 0, NULL);
	105
	106	vdev->virq_disabled = false;
	107
	108	vfio_config_free(vdev);
	109
89e1f7d4 AW	110	for (bar = PCI_STD_RESOURCES; bar <= PCI_STD_RESOURCE_END; bar++) {
	111	if (!vdev->barmap[bar])
	112	continue;
2007722a AW	113	pci_iounmap(pdev, vdev->barmap[bar]);
2007722a AW	114	pci_release_selected_regions(pdev, 1 << bar);
89e1f7d4 AW	115	vdev->barmap[bar] = NULL;
89e1f7d4 AW	116	}
2007722a AW	117
	118	/*
	119	* If we have saved state, restore it. If we can reset the device,
	120	* even better. Resetting with current state seems better than
	121	* nothing, but saving and restoring current state without reset
	122	* is just busy work.
	123	*/
	124	if (pci_load_and_free_saved_state(pdev, &vdev->pci_saved_state)) {
	125	pr_info("%s: Couldn't reload %s saved state\n",
	126	__func__, dev_name(&pdev->dev));
	127
	128	if (!vdev->reset_works)
	129	return;
	130
	131	pci_save_state(pdev);
	132	}
	133
	134	/*
	135	* Disable INTx and MSI, presumably to avoid spurious interrupts
	136	* during reset. Stolen from pci_reset_function()
	137	*/
	138	pci_write_config_word(pdev, PCI_COMMAND, PCI_COMMAND_INTX_DISABLE);
	139
d24cdbfd AW	140	/*
	141	* Careful, device_lock may already be held. This is the case if
	142	* a driver unbind is blocked. Try to get the locks ourselves to
	143	* prevent a deadlock.
	144	*/
	145	if (vdev->reset_works) {
	146	bool reset_done = false;
	147
	148	if (pci_cfg_access_trylock(pdev)) {
	149	if (device_trylock(&pdev->dev)) {
	150	__pci_reset_function_locked(pdev);
	151	reset_done = true;
	152	device_unlock(&pdev->dev);
	153	}
	154	pci_cfg_access_unlock(pdev);
	155	}
	156
	157	if (!reset_done)
	158	pr_warn("%s: Unable to acquire locks for reset of %s\n",
	159	__func__, dev_name(&pdev->dev));
	160	}
2007722a AW	161
2007722a AW	162	pci_restore_state(pdev);
89e1f7d4 AW	163	}
	164
	165	static void vfio_pci_release(void *device_data)
	166	{
	167	struct vfio_pci_device *vdev = device_data;
	168
	169	if (atomic_dec_and_test(&vdev->refcnt))
	170	vfio_pci_disable(vdev);
	171
	172	module_put(THIS_MODULE);
	173	}
	174
	175	static int vfio_pci_open(void *device_data)
	176	{
	177	struct vfio_pci_device *vdev = device_data;
	178
	179	if (!try_module_get(THIS_MODULE))
	180	return -ENODEV;
	181
	182	if (atomic_inc_return(&vdev->refcnt) == 1) {
	183	int ret = vfio_pci_enable(vdev);
	184	if (ret) {
	185	module_put(THIS_MODULE);
	186	return ret;
	187	}
	188	}
	189
	190	return 0;
	191	}
	192
	193	static int vfio_pci_get_irq_count(struct vfio_pci_device *vdev, int irq_type)
	194	{
	195	if (irq_type == VFIO_PCI_INTX_IRQ_INDEX) {
	196	u8 pin;
	197	pci_read_config_byte(vdev->pdev, PCI_INTERRUPT_PIN, &pin);
	198	if (pin)
	199	return 1;
	200
	201	} else if (irq_type == VFIO_PCI_MSI_IRQ_INDEX) {
	202	u8 pos;
	203	u16 flags;
	204
a9047f24	205	pos = vdev->pdev->msi_cap;
89e1f7d4 AW	206	if (pos) {
	207	pci_read_config_word(vdev->pdev,
	208	pos + PCI_MSI_FLAGS, &flags);
	209
	210	return 1 << (flags & PCI_MSI_FLAGS_QMASK);
	211	}
	212	} else if (irq_type == VFIO_PCI_MSIX_IRQ_INDEX) {
	213	u8 pos;
	214	u16 flags;
	215
a9047f24	216	pos = vdev->pdev->msix_cap;
89e1f7d4 AW	217	if (pos) {
	218	pci_read_config_word(vdev->pdev,
	219	pos + PCI_MSIX_FLAGS, &flags);
	220
	221	return (flags & PCI_MSIX_FLAGS_QSIZE) + 1;
	222	}
dad9f897 VMP	223	} else if (irq_type == VFIO_PCI_ERR_IRQ_INDEX)
	224	if (pci_is_pcie(vdev->pdev))
	225	return 1;
89e1f7d4 AW	226
	227	return 0;
	228	}
	229
	230	static long vfio_pci_ioctl(void *device_data,
	231	unsigned int cmd, unsigned long arg)
	232	{
	233	struct vfio_pci_device *vdev = device_data;
	234	unsigned long minsz;
	235
	236	if (cmd == VFIO_DEVICE_GET_INFO) {
	237	struct vfio_device_info info;
	238
	239	minsz = offsetofend(struct vfio_device_info, num_irqs);
	240
	241	if (copy_from_user(&info, (void __user *)arg, minsz))
	242	return -EFAULT;
	243
	244	if (info.argsz < minsz)
	245	return -EINVAL;
	246
	247	info.flags = VFIO_DEVICE_FLAGS_PCI;
	248
	249	if (vdev->reset_works)
	250	info.flags \|= VFIO_DEVICE_FLAGS_RESET;
	251
	252	info.num_regions = VFIO_PCI_NUM_REGIONS;
	253	info.num_irqs = VFIO_PCI_NUM_IRQS;
	254
	255	return copy_to_user((void __user *)arg, &info, minsz);
	256
	257	} else if (cmd == VFIO_DEVICE_GET_REGION_INFO) {
	258	struct pci_dev *pdev = vdev->pdev;
	259	struct vfio_region_info info;
	260
	261	minsz = offsetofend(struct vfio_region_info, offset);
	262
	263	if (copy_from_user(&info, (void __user *)arg, minsz))
	264	return -EFAULT;
	265
	266	if (info.argsz < minsz)
	267	return -EINVAL;
	268
	269	switch (info.index) {
	270	case VFIO_PCI_CONFIG_REGION_INDEX:
	271	info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
	272	info.size = pdev->cfg_size;
	273	info.flags = VFIO_REGION_INFO_FLAG_READ \|
	274	VFIO_REGION_INFO_FLAG_WRITE;
	275	break;
	276	case VFIO_PCI_BAR0_REGION_INDEX ... VFIO_PCI_BAR5_REGION_INDEX:
	277	info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
	278	info.size = pci_resource_len(pdev, info.index);
	279	if (!info.size) {
	280	info.flags = 0;
	281	break;
	282	}
	283
	284	info.flags = VFIO_REGION_INFO_FLAG_READ \|
	285	VFIO_REGION_INFO_FLAG_WRITE;
	286	if (pci_resource_flags(pdev, info.index) &
	287	IORESOURCE_MEM && info.size >= PAGE_SIZE)
	288	info.flags \|= VFIO_REGION_INFO_FLAG_MMAP;
	289	break;
290	case VFIO_PCI_ROM_REGION_INDEX:
291	{
292	void __iomem *io;
293	size_t size;
294
295	info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
296	info.flags = 0;
297
298	/* Report the BAR size, not the ROM size */
299	info.size = pci_resource_len(pdev, info.index);
300	if (!info.size)
301	break;
302
303	/* Is it really there? */
304	io = pci_map_rom(pdev, &size);
305	if (!io \|\| !size) {
306	info.size = 0;
307	break;
308	}
309	pci_unmap_rom(pdev, io);
310
311	info.flags = VFIO_REGION_INFO_FLAG_READ;
312	break;
313	}
84237a82 AW	314	case VFIO_PCI_VGA_REGION_INDEX:
	315	if (!vdev->has_vga)
	316	return -EINVAL;
	317
	318	info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
	319	info.size = 0xc0000;
	320	info.flags = VFIO_REGION_INFO_FLAG_READ \|
	321	VFIO_REGION_INFO_FLAG_WRITE;
	322
	323	break;
89e1f7d4 AW	324	default:
	325	return -EINVAL;
	326	}
	327
	328	return copy_to_user((void __user *)arg, &info, minsz);
	329
	330	} else if (cmd == VFIO_DEVICE_GET_IRQ_INFO) {
	331	struct vfio_irq_info info;
	332
	333	minsz = offsetofend(struct vfio_irq_info, count);
	334
	335	if (copy_from_user(&info, (void __user *)arg, minsz))
	336	return -EFAULT;
	337
	338	if (info.argsz < minsz \|\| info.index >= VFIO_PCI_NUM_IRQS)
	339	return -EINVAL;
	340
dad9f897 VMP	341	switch (info.index) {
	342	case VFIO_PCI_INTX_IRQ_INDEX ... VFIO_PCI_MSIX_IRQ_INDEX:
	343	break;
	344	case VFIO_PCI_ERR_IRQ_INDEX:
	345	if (pci_is_pcie(vdev->pdev))
	346	break;
	347	/* pass thru to return error */
	348	default:
	349	return -EINVAL;
	350	}
	351
89e1f7d4 AW	352	info.flags = VFIO_IRQ_INFO_EVENTFD;
	353
	354	info.count = vfio_pci_get_irq_count(vdev, info.index);
	355
	356	if (info.index == VFIO_PCI_INTX_IRQ_INDEX)
	357	info.flags \|= (VFIO_IRQ_INFO_MASKABLE \|
	358	VFIO_IRQ_INFO_AUTOMASKED);
	359	else
	360	info.flags \|= VFIO_IRQ_INFO_NORESIZE;
	361
	362	return copy_to_user((void __user *)arg, &info, minsz);
	363
	364	} else if (cmd == VFIO_DEVICE_SET_IRQS) {
	365	struct vfio_irq_set hdr;
	366	u8 *data = NULL;
	367	int ret = 0;
	368
	369	minsz = offsetofend(struct vfio_irq_set, count);
	370
	371	if (copy_from_user(&hdr, (void __user *)arg, minsz))
	372	return -EFAULT;
	373
	374	if (hdr.argsz < minsz \|\| hdr.index >= VFIO_PCI_NUM_IRQS \|\|
	375	hdr.flags & ~(VFIO_IRQ_SET_DATA_TYPE_MASK \|
	376	VFIO_IRQ_SET_ACTION_TYPE_MASK))
	377	return -EINVAL;
	378
	379	if (!(hdr.flags & VFIO_IRQ_SET_DATA_NONE)) {
	380	size_t size;
904c680c	381	int max = vfio_pci_get_irq_count(vdev, hdr.index);
89e1f7d4 AW	382
	383	if (hdr.flags & VFIO_IRQ_SET_DATA_BOOL)
	384	size = sizeof(uint8_t);
	385	else if (hdr.flags & VFIO_IRQ_SET_DATA_EVENTFD)
	386	size = sizeof(int32_t);
	387	else
	388	return -EINVAL;
	389
	390	if (hdr.argsz - minsz < hdr.count * size \|\|
904c680c	391	hdr.start >= max \|\| hdr.start + hdr.count > max)
89e1f7d4 AW	392	return -EINVAL;
89e1f7d4 AW	393
3a1f7041 FW	394	data = memdup_user((void __user *)(arg + minsz),
	395	hdr.count * size);
	396	if (IS_ERR(data))
	397	return PTR_ERR(data);
89e1f7d4 AW	398	}
	399
	400	mutex_lock(&vdev->igate);
	401
	402	ret = vfio_pci_set_irqs_ioctl(vdev, hdr.flags, hdr.index,
	403	hdr.start, hdr.count, data);
	404
	405	mutex_unlock(&vdev->igate);
	406	kfree(data);
	407
	408	return ret;
	409
	410	} else if (cmd == VFIO_DEVICE_RESET)
	411	return vdev->reset_works ?
	412	pci_reset_function(vdev->pdev) : -EINVAL;
	413
	414	return -ENOTTY;
	415	}
	416
5b279a11 AW	417	static ssize_t vfio_pci_rw(void device_data, char __user buf,
5b279a11 AW	418	size_t count, loff_t *ppos, bool iswrite)
89e1f7d4 AW	419	{
	420	unsigned int index = VFIO_PCI_OFFSET_TO_INDEX(*ppos);
	421	struct vfio_pci_device *vdev = device_data;
89e1f7d4 AW	422
	423	if (index >= VFIO_PCI_NUM_REGIONS)
	424	return -EINVAL;
	425
5b279a11 AW	426	switch (index) {
5b279a11 AW	427	case VFIO_PCI_CONFIG_REGION_INDEX:
906ee99d AW	428	return vfio_pci_config_rw(vdev, buf, count, ppos, iswrite);
906ee99d AW	429
5b279a11 AW	430	case VFIO_PCI_ROM_REGION_INDEX:
	431	if (iswrite)
	432	return -EINVAL;
906ee99d	433	return vfio_pci_bar_rw(vdev, buf, count, ppos, false);
89e1f7d4	434
5b279a11	435	case VFIO_PCI_BAR0_REGION_INDEX ... VFIO_PCI_BAR5_REGION_INDEX:
906ee99d	436	return vfio_pci_bar_rw(vdev, buf, count, ppos, iswrite);
84237a82 AW	437
	438	case VFIO_PCI_VGA_REGION_INDEX:
	439	return vfio_pci_vga_rw(vdev, buf, count, ppos, iswrite);
5b279a11 AW	440	}
5b279a11 AW	441
89e1f7d4 AW	442	return -EINVAL;
	443	}
	444
5b279a11 AW	445	static ssize_t vfio_pci_read(void device_data, char __user buf,
	446	size_t count, loff_t *ppos)
	447	{
906ee99d AW	448	if (!count)
	449	return 0;
	450
5b279a11 AW	451	return vfio_pci_rw(device_data, buf, count, ppos, false);
	452	}
	453
89e1f7d4 AW	454	static ssize_t vfio_pci_write(void device_data, const char __user buf,
	455	size_t count, loff_t *ppos)
	456	{
906ee99d AW	457	if (!count)
	458	return 0;
	459
	460	return vfio_pci_rw(device_data, (char __user *)buf, count, ppos, true);
89e1f7d4 AW	461	}
	462
	463	static int vfio_pci_mmap(void device_data, struct vm_area_struct vma)
	464	{
	465	struct vfio_pci_device *vdev = device_data;
	466	struct pci_dev *pdev = vdev->pdev;
	467	unsigned int index;
34002f54	468	u64 phys_len, req_len, pgoff, req_start;
89e1f7d4 AW	469	int ret;
	470
	471	index = vma->vm_pgoff >> (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT);
	472
	473	if (vma->vm_end < vma->vm_start)
	474	return -EINVAL;
	475	if ((vma->vm_flags & VM_SHARED) == 0)
	476	return -EINVAL;
	477	if (index >= VFIO_PCI_ROM_REGION_INDEX)
	478	return -EINVAL;
	479	if (!(pci_resource_flags(pdev, index) & IORESOURCE_MEM))
	480	return -EINVAL;
	481
	482	phys_len = pci_resource_len(pdev, index);
	483	req_len = vma->vm_end - vma->vm_start;
	484	pgoff = vma->vm_pgoff &
	485	((1U << (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT)) - 1);
	486	req_start = pgoff << PAGE_SHIFT;
	487
	488	if (phys_len < PAGE_SIZE \|\| req_start + req_len > phys_len)
	489	return -EINVAL;
	490
	491	if (index == vdev->msix_bar) {
	492	/*
	493	* Disallow mmaps overlapping the MSI-X table; users don't
	494	* get to touch this directly. We could find somewhere
	495	* else to map the overlap, but page granularity is only
	496	* a recommendation, not a requirement, so the user needs
	497	* to know which bits are real. Requiring them to mmap
	498	* around the table makes that clear.
	499	*/
	500
	501	/* If neither entirely above nor below, then it overlaps */
	502	if (!(req_start >= vdev->msix_offset + vdev->msix_size \|\|
	503	req_start + req_len <= vdev->msix_offset))
	504	return -EINVAL;
	505	}
	506
	507	/*
	508	* Even though we don't make use of the barmap for the mmap,
	509	* we need to request the region and the barmap tracks that.
	510	*/
	511	if (!vdev->barmap[index]) {
	512	ret = pci_request_selected_regions(pdev,
	513	1 << index, "vfio-pci");
	514	if (ret)
	515	return ret;
	516
	517	vdev->barmap[index] = pci_iomap(pdev, index, 0);
	518	}
	519
	520	vma->vm_private_data = vdev;
89e1f7d4	521	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
34002f54	522	vma->vm_pgoff = (pci_resource_start(pdev, index) >> PAGE_SHIFT) + pgoff;
89e1f7d4	523
34002f54	524	return remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
89e1f7d4 AW	525	req_len, vma->vm_page_prot);
	526	}
	527
	528	static const struct vfio_device_ops vfio_pci_ops = {
	529	.name = "vfio-pci",
	530	.open = vfio_pci_open,
	531	.release = vfio_pci_release,
	532	.ioctl = vfio_pci_ioctl,
	533	.read = vfio_pci_read,
	534	.write = vfio_pci_write,
	535	.mmap = vfio_pci_mmap,
	536	};
	537
	538	static int vfio_pci_probe(struct pci_dev pdev, const struct pci_device_id id)
	539	{
	540	u8 type;
	541	struct vfio_pci_device *vdev;
	542	struct iommu_group *group;
	543	int ret;
	544
	545	pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
	546	if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
	547	return -EINVAL;
	548
	549	group = iommu_group_get(&pdev->dev);
	550	if (!group)
	551	return -EINVAL;
	552
	553	vdev = kzalloc(sizeof(*vdev), GFP_KERNEL);
	554	if (!vdev) {
	555	iommu_group_put(group);
	556	return -ENOMEM;
	557	}
	558
	559	vdev->pdev = pdev;
	560	vdev->irq_type = VFIO_PCI_NUM_IRQS;
	561	mutex_init(&vdev->igate);
	562	spin_lock_init(&vdev->irqlock);
	563	atomic_set(&vdev->refcnt, 0);
	564
	565	ret = vfio_add_group_dev(&pdev->dev, &vfio_pci_ops, vdev);
	566	if (ret) {
	567	iommu_group_put(group);
	568	kfree(vdev);
	569	}
	570
	571	return ret;
	572	}
	573
	574	static void vfio_pci_remove(struct pci_dev *pdev)
	575	{
	576	struct vfio_pci_device *vdev;
	577
	578	vdev = vfio_del_group_dev(&pdev->dev);
	579	if (!vdev)
	580	return;
	581
	582	iommu_group_put(pdev->dev.iommu_group);
	583	kfree(vdev);
	584	}
	585
dad9f897 VMP	586	static pci_ers_result_t vfio_pci_aer_err_detected(struct pci_dev *pdev,
	587	pci_channel_state_t state)
	588	{
	589	struct vfio_pci_device *vdev;
	590	struct vfio_device *device;
	591
	592	device = vfio_device_get_from_dev(&pdev->dev);
	593	if (device == NULL)
	594	return PCI_ERS_RESULT_DISCONNECT;
	595
	596	vdev = vfio_device_data(device);
	597	if (vdev == NULL) {
	598	vfio_device_put(device);
	599	return PCI_ERS_RESULT_DISCONNECT;
	600	}
	601
	602	if (vdev->err_trigger)
	603	eventfd_signal(vdev->err_trigger, 1);
	604
	605	vfio_device_put(device);
	606
	607	return PCI_ERS_RESULT_CAN_RECOVER;
	608	}
	609
	610	static struct pci_error_handlers vfio_err_handlers = {
	611	.error_detected = vfio_pci_aer_err_detected,
	612	};
	613
89e1f7d4 AW	614	static struct pci_driver vfio_pci_driver = {
	615	.name = "vfio-pci",
	616	.id_table = NULL, /* only dynamic ids */
	617	.probe = vfio_pci_probe,
	618	.remove = vfio_pci_remove,
dad9f897	619	.err_handler = &vfio_err_handlers,
89e1f7d4 AW	620	};
	621
	622	static void __exit vfio_pci_cleanup(void)
	623	{
	624	pci_unregister_driver(&vfio_pci_driver);
	625	vfio_pci_virqfd_exit();
	626	vfio_pci_uninit_perm_bits();
	627	}
	628
	629	static int __init vfio_pci_init(void)
	630	{
	631	int ret;
	632
	633	/* Allocate shared config space permision data used by all devices */
	634	ret = vfio_pci_init_perm_bits();
	635	if (ret)
	636	return ret;
	637
	638	/* Start the virqfd cleanup handler */
	639	ret = vfio_pci_virqfd_init();
	640	if (ret)
	641	goto out_virqfd;
	642
	643	/* Register and scan for devices */
	644	ret = pci_register_driver(&vfio_pci_driver);
	645	if (ret)
	646	goto out_driver;
	647
	648	return 0;
	649
89e1f7d4	650	out_driver:
05bf3aac JL	651	vfio_pci_virqfd_exit();
05bf3aac JL	652	out_virqfd:
89e1f7d4 AW	653	vfio_pci_uninit_perm_bits();
	654	return ret;
	655	}
	656
	657	module_init(vfio_pci_init);
	658	module_exit(vfio_pci_cleanup);
	659
	660	MODULE_VERSION(DRIVER_VERSION);
	661	MODULE_LICENSE("GPL v2");
	662	MODULE_AUTHOR(DRIVER_AUTHOR);
	663	MODULE_DESCRIPTION(DRIVER_DESC);