amd_iommu.c

/*
 * Copyright (C) 2007-2010 Advanced Micro Devices, Inc.
 * Author: Joerg Roedel <jroedel@suse.de>
 *         Leo Duran <leo.duran@amd.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.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
 */

#include <linux/ratelimit.h>
#include <linux/pci.h>
#include <linux/pci-ats.h>
#include <linux/bitmap.h>
#include <linux/slab.h>
#include <linux/debugfs.h>
#include <linux/scatterlist.h>
#include <linux/dma-mapping.h>
#include <linux/iommu-helper.h>
#include <linux/iommu.h>
#include <linux/delay.h>
#include <linux/amd-iommu.h>
#include <linux/notifier.h>
#include <linux/export.h>
#include <linux/irq.h>
#include <linux/msi.h>
#include <linux/dma-contiguous.h>
#include <asm/irq_remapping.h>
#include <asm/io_apic.h>
#include <asm/apic.h>
#include <asm/hw_irq.h>
#include <asm/msidef.h>
#include <asm/proto.h>
#include <asm/iommu.h>
#include <asm/gart.h>
#include <asm/dma.h>

#include "amd_iommu_proto.h"
#include "amd_iommu_types.h"
#include "irq_remapping.h"

#define CMD_SET_TYPE(cmd, t) ((cmd)->data[1] |= ((t) << 28))

#define LOOP_TIMEOUT	100000

/*
 * This bitmap is used to advertise the page sizes our hardware support
 * to the IOMMU core, which will then use this information to split
 * physically contiguous memory regions it is mapping into page sizes
 * that we support.
 *
 * 512GB Pages are not supported due to a hardware bug
 */
#define AMD_IOMMU_PGSIZES	((~0xFFFUL) & ~(2ULL << 38))

static DEFINE_RWLOCK(amd_iommu_devtable_lock);

/* A list of preallocated protection domains */
static LIST_HEAD(iommu_pd_list);
static DEFINE_SPINLOCK(iommu_pd_list_lock);

/* List of all available dev_data structures */
static LIST_HEAD(dev_data_list);
static DEFINE_SPINLOCK(dev_data_list_lock);

LIST_HEAD(ioapic_map);
LIST_HEAD(hpet_map);

/*
 * Domain for untranslated devices - only allocated
 * if iommu=pt passed on kernel cmd line.
 */
static struct protection_domain *pt_domain;

static const struct iommu_ops amd_iommu_ops;

static ATOMIC_NOTIFIER_HEAD(ppr_notifier);
int amd_iommu_max_glx_val = -1;

static struct dma_map_ops amd_iommu_dma_ops;

/*
 * This struct contains device specific data for the IOMMU
 */
struct iommu_dev_data {
	struct list_head list;		  /* For domain->dev_list */
	struct list_head dev_data_list;	  /* For global dev_data_list */
	struct list_head alias_list;      /* Link alias-groups together */
	struct iommu_dev_data *alias_data;/* The alias dev_data */
	struct protection_domain *domain; /* Domain the device is bound to */
	u16 devid;			  /* PCI Device ID */
	bool iommu_v2;			  /* Device can make use of IOMMUv2 */
	bool passthrough;		  /* Default for device is pt_domain */
	struct {
		bool enabled;
		int qdep;
	} ats;				  /* ATS state */
	bool pri_tlp;			  /* PASID TLB required for
					     PPR completions */
	u32 errata;			  /* Bitmap for errata to apply */
};

/*
 * general struct to manage commands send to an IOMMU
 */
struct iommu_cmd {
	u32 data[4];
};

struct kmem_cache *amd_iommu_irq_cache;

static void update_domain(struct protection_domain *domain);
static int __init alloc_passthrough_domain(void);

/****************************************************************************
 *
 * Helper functions
 *
 ****************************************************************************/

static struct protection_domain *to_pdomain(struct iommu_domain *dom)
{
	return container_of(dom, struct protection_domain, domain);
}

static struct iommu_dev_data *alloc_dev_data(u16 devid)
{
	struct iommu_dev_data *dev_data;
	unsigned long flags;

	dev_data = kzalloc(sizeof(*dev_data), GFP_KERNEL);
	if (!dev_data)
		return NULL;

	INIT_LIST_HEAD(&dev_data->alias_list);

	dev_data->devid = devid;

	spin_lock_irqsave(&dev_data_list_lock, flags);
	list_add_tail(&dev_data->dev_data_list, &dev_data_list);
	spin_unlock_irqrestore(&dev_data_list_lock, flags);

	return dev_data;
}

static void free_dev_data(struct iommu_dev_data *dev_data)
{
	unsigned long flags;

	spin_lock_irqsave(&dev_data_list_lock, flags);
	list_del(&dev_data->dev_data_list);
	spin_unlock_irqrestore(&dev_data_list_lock, flags);

	kfree(dev_data);
}

static struct iommu_dev_data *search_dev_data(u16 devid)
{
	struct iommu_dev_data *dev_data;
	unsigned long flags;

	spin_lock_irqsave(&dev_data_list_lock, flags);
	list_for_each_entry(dev_data, &dev_data_list, dev_data_list) {
		if (dev_data->devid == devid)
			goto out_unlock;
	}

	dev_data = NULL;

out_unlock:
	spin_unlock_irqrestore(&dev_data_list_lock, flags);

	return dev_data;
}

static struct iommu_dev_data *find_dev_data(u16 devid)
{
	struct iommu_dev_data *dev_data;

	dev_data = search_dev_data(devid);

	if (dev_data == NULL)
		dev_data = alloc_dev_data(devid);

	return dev_data;
}

static inline u16 get_device_id(struct device *dev)
{
	struct pci_dev *pdev = to_pci_dev(dev);

	return PCI_DEVID(pdev->bus->number, pdev->devfn);
}

static struct iommu_dev_data *get_dev_data(struct device *dev)
{
	return dev->archdata.iommu;
}

static bool pci_iommuv2_capable(struct pci_dev *pdev)
{
	static const int caps[] = {
		PCI_EXT_CAP_ID_ATS,
		PCI_EXT_CAP_ID_PRI,
		PCI_EXT_CAP_ID_PASID,
	};
	int i, pos;

	for (i = 0; i < 3; ++i) {
		pos = pci_find_ext_capability(pdev, caps[i]);
		if (pos == 0)
			return false;
	}

	return true;
}

static bool pdev_pri_erratum(struct pci_dev *pdev, u32 erratum)
{
	struct iommu_dev_data *dev_data;

	dev_data = get_dev_data(&pdev->dev);

	return dev_data->errata & (1 << erratum) ? true : false;
}

/*
 * In this function the list of preallocated protection domains is traversed to
 * find the domain for a specific device
 */
static struct dma_ops_domain *find_protection_domain(u16 devid)
{
	struct dma_ops_domain *entry, *ret = NULL;
	unsigned long flags;
	u16 alias = amd_iommu_alias_table[devid];

	if (list_empty(&iommu_pd_list))
		return NULL;

	spin_lock_irqsave(&iommu_pd_list_lock, flags);

	list_for_each_entry(entry, &iommu_pd_list, list) {
		if (entry->target_dev == devid ||
		    entry->target_dev == alias) {
			ret = entry;
			break;
		}
	}

	spin_unlock_irqrestore(&iommu_pd_list_lock, flags);

	return ret;
}

/*
 * This function checks if the driver got a valid device from the caller to
 * avoid dereferencing invalid pointers.
 */
static bool check_device(struct device *dev)
{
	u16 devid;

	if (!dev || !dev->dma_mask)
		return false;

	/* No PCI device */
	if (!dev_is_pci(dev))
		return false;

	devid = get_device_id(dev);

	/* Out of our scope? */
	if (devid > amd_iommu_last_bdf)
		return false;

	if (amd_iommu_rlookup_table[devid] == NULL)
		return false;

	return true;
}

static void init_iommu_group(struct device *dev)
{
	struct iommu_group *group;

	group = iommu_group_get_for_dev(dev);
	if (!IS_ERR(group))
		iommu_group_put(group);
}

static int __last_alias(struct pci_dev *pdev, u16 alias, void *data)
{
	*(u16 *)data = alias;
	return 0;
}

static u16 get_alias(struct device *dev)
{
	struct pci_dev *pdev = to_pci_dev(dev);
	u16 devid, ivrs_alias, pci_alias;

	devid = get_device_id(dev);
	ivrs_alias = amd_iommu_alias_table[devid];
	pci_for_each_dma_alias(pdev, __last_alias, &pci_alias);

	if (ivrs_alias == pci_alias)
		return ivrs_alias;

	/*
	 * DMA alias showdown
	 *
	 * The IVRS is fairly reliable in telling us about aliases, but it
	 * can't know about every screwy device.  If we don't have an IVRS
	 * reported alias, use the PCI reported alias.  In that case we may
	 * still need to initialize the rlookup and dev_table entries if the
	 * alias is to a non-existent device.
	 */
	if (ivrs_alias == devid) {
		if (!amd_iommu_rlookup_table[pci_alias]) {
			amd_iommu_rlookup_table[pci_alias] =
				amd_iommu_rlookup_table[devid];
			memcpy(amd_iommu_dev_table[pci_alias].data,
			       amd_iommu_dev_table[devid].data,
			       sizeof(amd_iommu_dev_table[pci_alias].data));
		}

		return pci_alias;
	}

	pr_info("AMD-Vi: Using IVRS reported alias %02x:%02x.%d "
		"for device %s[%04x:%04x], kernel reported alias "
		"%02x:%02x.%d\n", PCI_BUS_NUM(ivrs_alias), PCI_SLOT(ivrs_alias),
		PCI_FUNC(ivrs_alias), dev_name(dev), pdev->vendor, pdev->device,
		PCI_BUS_NUM(pci_alias), PCI_SLOT(pci_alias),
		PCI_FUNC(pci_alias));

	/*
	 * If we don't have a PCI DMA alias and the IVRS alias is on the same
	 * bus, then the IVRS table may know about a quirk that we don't.
	 */
	if (pci_alias == devid &&
	    PCI_BUS_NUM(ivrs_alias) == pdev->bus->number) {
		pdev->dev_flags |= PCI_DEV_FLAGS_DMA_ALIAS_DEVFN;
		pdev->dma_alias_devfn = ivrs_alias & 0xff;
		pr_info("AMD-Vi: Added PCI DMA alias %02x.%d for %s\n",
			PCI_SLOT(ivrs_alias), PCI_FUNC(ivrs_alias),
			dev_name(dev));
	}

	return ivrs_alias;
}

static int iommu_init_device(struct device *dev)
{
	struct pci_dev *pdev = to_pci_dev(dev);
	struct iommu_dev_data *dev_data;
	u16 alias;

	if (dev->archdata.iommu)
		return 0;

	dev_data = find_dev_data(get_device_id(dev));
	if (!dev_data)
		return -ENOMEM;

	alias = get_alias(dev);

	if (alias != dev_data->devid) {
		struct iommu_dev_data *alias_data;

		alias_data = find_dev_data(alias);
		if (alias_data == NULL) {
			pr_err("AMD-Vi: Warning: Unhandled device %s\n",
					dev_name(dev));
			free_dev_data(dev_data);
			return -ENOTSUPP;
		}
		dev_data->alias_data = alias_data;

		/* Add device to the alias_list */
		list_add(&dev_data->alias_list, &alias_data->alias_list);
	}

	if (pci_iommuv2_capable(pdev)) {
		struct amd_iommu *iommu;

		iommu              = amd_iommu_rlookup_table[dev_data->devid];
		dev_data->iommu_v2 = iommu->is_iommu_v2;
	}

	dev->archdata.iommu = dev_data;

	iommu_device_link(amd_iommu_rlookup_table[dev_data->devid]->iommu_dev,
			  dev);

	return 0;
}

static void iommu_ignore_device(struct device *dev)
{
	u16 devid, alias;

	devid = get_device_id(dev);
	alias = amd_iommu_alias_table[devid];

	memset(&amd_iommu_dev_table[devid], 0, sizeof(struct dev_table_entry));
	memset(&amd_iommu_dev_table[alias], 0, sizeof(struct dev_table_entry));

	amd_iommu_rlookup_table[devid] = NULL;
	amd_iommu_rlookup_table[alias] = NULL;
}

static void iommu_uninit_device(struct device *dev)
{
	struct iommu_dev_data *dev_data = search_dev_data(get_device_id(dev));

	if (!dev_data)
		return;

	iommu_device_unlink(amd_iommu_rlookup_table[dev_data->devid]->iommu_dev,
			    dev);

	iommu_group_remove_device(dev);

	/* Unlink from alias, it may change if another device is re-plugged */
	dev_data->alias_data = NULL;

	/*
	 * We keep dev_data around for unplugged devices and reuse it when the
	 * device is re-plugged - not doing so would introduce a ton of races.
	 */
}

void __init amd_iommu_uninit_devices(void)
{
	struct iommu_dev_data *dev_data, *n;
	struct pci_dev *pdev = NULL;

	for_each_pci_dev(pdev) {

		if (!check_device(&pdev->dev))
			continue;

		iommu_uninit_device(&pdev->dev);
	}

	/* Free all of our dev_data structures */
	list_for_each_entry_safe(dev_data, n, &dev_data_list, dev_data_list)
		free_dev_data(dev_data);
}

int __init amd_iommu_init_devices(void)
{
	struct pci_dev *pdev = NULL;
	int ret = 0;

	for_each_pci_dev(pdev) {

		if (!check_device(&pdev->dev))
			continue;

		ret = iommu_init_device(&pdev->dev);
		if (ret == -ENOTSUPP)
			iommu_ignore_device(&pdev->dev);
		else if (ret)
			goto out_free;
	}

	/*
	 * Initialize IOMMU groups only after iommu_init_device() has
	 * had a chance to populate any IVRS defined aliases.
	 */
	for_each_pci_dev(pdev) {
		if (check_device(&pdev->dev))
			init_iommu_group(&pdev->dev);
	}

	return 0;

out_free:

	amd_iommu_uninit_devices();

	return ret;
}
#ifdef CONFIG_AMD_IOMMU_STATS

/*
 * Initialization code for statistics collection
 */

DECLARE_STATS_COUNTER(compl_wait);
DECLARE_STATS_COUNTER(cnt_map_single);
DECLARE_STATS_COUNTER(cnt_unmap_single);
DECLARE_STATS_COUNTER(cnt_map_sg);
DECLARE_STATS_COUNTER(cnt_unmap_sg);
DECLARE_STATS_COUNTER(cnt_alloc_coherent);
DECLARE_STATS_COUNTER(cnt_free_coherent);
DECLARE_STATS_COUNTER(cross_page);
DECLARE_STATS_COUNTER(domain_flush_single);
DECLARE_STATS_COUNTER(domain_flush_all);
DECLARE_STATS_COUNTER(alloced_io_mem);
DECLARE_STATS_COUNTER(total_map_requests);
DECLARE_STATS_COUNTER(complete_ppr);
DECLARE_STATS_COUNTER(invalidate_iotlb);
DECLARE_STATS_COUNTER(invalidate_iotlb_all);
DECLARE_STATS_COUNTER(pri_requests);

static struct dentry *stats_dir;
static struct dentry *de_fflush;

static void amd_iommu_stats_add(struct __iommu_counter *cnt)
{
	if (stats_dir == NULL)
		return;

	cnt->dent = debugfs_create_u64(cnt->name, 0444, stats_dir,
				       &cnt->value);
}

static void amd_iommu_stats_init(void)
{
	stats_dir = debugfs_create_dir("amd-iommu", NULL);
	if (stats_dir == NULL)
		return;

	de_fflush  = debugfs_create_bool("fullflush", 0444, stats_dir,
					 &amd_iommu_unmap_flush);

	amd_iommu_stats_add(&compl_wait);
	amd_iommu_stats_add(&cnt_map_single);
	amd_iommu_stats_add(&cnt_unmap_single);
	amd_iommu_stats_add(&cnt_map_sg);
	amd_iommu_stats_add(&cnt_unmap_sg);
	amd_iommu_stats_add(&cnt_alloc_coherent);
	amd_iommu_stats_add(&cnt_free_coherent);
	amd_iommu_stats_add(&cross_page);
	amd_iommu_stats_add(&domain_flush_single);
	amd_iommu_stats_add(&domain_flush_all);
	amd_iommu_stats_add(&alloced_io_mem);
	amd_iommu_stats_add(&total_map_requests);
	amd_iommu_stats_add(&complete_ppr);
	amd_iommu_stats_add(&invalidate_iotlb);
	amd_iommu_stats_add(&invalidate_iotlb_all);
	amd_iommu_stats_add(&pri_requests);
}

#endif

/****************************************************************************
 *
 * Interrupt handling functions
 *
 ****************************************************************************/

static void dump_dte_entry(u16 devid)
{
	int i;

	for (i = 0; i < 4; ++i)
		pr_err("AMD-Vi: DTE[%d]: %016llx\n", i,
			amd_iommu_dev_table[devid].data[i]);
}

static void dump_command(unsigned long phys_addr)
{
	struct iommu_cmd *cmd = phys_to_virt(phys_addr);
	int i;

	for (i = 0; i < 4; ++i)
		pr_err("AMD-Vi: CMD[%d]: %08x\n", i, cmd->data[i]);
}

static void iommu_print_event(struct amd_iommu *iommu, void *__evt)
{
	int type, devid, domid, flags;
	volatile u32 *event = __evt;
	int count = 0;
	u64 address;

retry:
	type    = (event[1] >> EVENT_TYPE_SHIFT)  & EVENT_TYPE_MASK;
	devid   = (event[0] >> EVENT_DEVID_SHIFT) & EVENT_DEVID_MASK;
	domid   = (event[1] >> EVENT_DOMID_SHIFT) & EVENT_DOMID_MASK;
	flags   = (event[1] >> EVENT_FLAGS_SHIFT) & EVENT_FLAGS_MASK;
	address = (u64)(((u64)event[3]) << 32) | event[2];

	if (type == 0) {
		/* Did we hit the erratum? */
		if (++count == LOOP_TIMEOUT) {
			pr_err("AMD-Vi: No event written to event log\n");
			return;
		}
		udelay(1);
		goto retry;
	}

	printk(KERN_ERR "AMD-Vi: Event logged [");

	switch (type) {
	case EVENT_TYPE_ILL_DEV:
		printk("ILLEGAL_DEV_TABLE_ENTRY device=%02x:%02x.%x "
		       "address=0x%016llx flags=0x%04x]\n",
		       PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid),
		       address, flags);
		dump_dte_entry(devid);
		break;
	case EVENT_TYPE_IO_FAULT:
		printk("IO_PAGE_FAULT device=%02x:%02x.%x "
		       "domain=0x%04x address=0x%016llx flags=0x%04x]\n",
		       PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid),
		       domid, address, flags);
		break;
	case EVENT_TYPE_DEV_TAB_ERR:
		printk("DEV_TAB_HARDWARE_ERROR device=%02x:%02x.%x "
		       "address=0x%016llx flags=0x%04x]\n",
		       PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid),
		       address, flags);
		break;
	case EVENT_TYPE_PAGE_TAB_ERR:
		printk("PAGE_TAB_HARDWARE_ERROR device=%02x:%02x.%x "
		       "domain=0x%04x address=0x%016llx flags=0x%04x]\n",
		       PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid),
		       domid, address, flags);
		break;
	case EVENT_TYPE_ILL_CMD:
		printk("ILLEGAL_COMMAND_ERROR address=0x%016llx]\n", address);
		dump_command(address);
		break;
	case EVENT_TYPE_CMD_HARD_ERR:
		printk("COMMAND_HARDWARE_ERROR address=0x%016llx "
		       "flags=0x%04x]\n", address, flags);
		break;
	case EVENT_TYPE_IOTLB_INV_TO:
		printk("IOTLB_INV_TIMEOUT device=%02x:%02x.%x "
		       "address=0x%016llx]\n",
		       PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid),
		       address);
		break;
	case EVENT_TYPE_INV_DEV_REQ:
		printk("INVALID_DEVICE_REQUEST device=%02x:%02x.%x "
		       "address=0x%016llx flags=0x%04x]\n",
		       PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid),
		       address, flags);
		break;
	default:
		printk(KERN_ERR "UNKNOWN type=0x%02x]\n", type);
	}

	memset(__evt, 0, 4 * sizeof(u32));
}

static void iommu_poll_events(struct amd_iommu *iommu)
{
	u32 head, tail;

	head = readl(iommu->mmio_base + MMIO_EVT_HEAD_OFFSET);
	tail = readl(iommu->mmio_base + MMIO_EVT_TAIL_OFFSET);

	while (head != tail) {
		iommu_print_event(iommu, iommu->evt_buf + head);
		head = (head + EVENT_ENTRY_SIZE) % iommu->evt_buf_size;
	}

	writel(head, iommu->mmio_base + MMIO_EVT_HEAD_OFFSET);
}

static void iommu_handle_ppr_entry(struct amd_iommu *iommu, u64 *raw)
{
	struct amd_iommu_fault fault;

	INC_STATS_COUNTER(pri_requests);

	if (PPR_REQ_TYPE(raw[0]) != PPR_REQ_FAULT) {
		pr_err_ratelimited("AMD-Vi: Unknown PPR request received\n");
		return;
	}

	fault.address   = raw[1];
	fault.pasid     = PPR_PASID(raw[0]);
	fault.device_id = PPR_DEVID(raw[0]);
	fault.tag       = PPR_TAG(raw[0]);
	fault.flags     = PPR_FLAGS(raw[0]);

	atomic_notifier_call_chain(&ppr_notifier, 0, &fault);
}

static void iommu_poll_ppr_log(struct amd_iommu *iommu)
{
	u32 head, tail;

	if (iommu->ppr_log == NULL)
		return;

	head = readl(iommu->mmio_base + MMIO_PPR_HEAD_OFFSET);
	tail = readl(iommu->mmio_base + MMIO_PPR_TAIL_OFFSET);

	while (head != tail) {
		volatile u64 *raw;
		u64 entry[2];
		int i;

		raw = (u64 *)(iommu->ppr_log + head);

		/*
		 * Hardware bug: Interrupt may arrive before the entry is
		 * written to memory. If this happens we need to wait for the
		 * entry to arrive.
		 */
		for (i = 0; i < LOOP_TIMEOUT; ++i) {
			if (PPR_REQ_TYPE(raw[0]) != 0)
				break;
			udelay(1);
		}

		/* Avoid memcpy function-call overhead */
		entry[0] = raw[0];
		entry[1] = raw[1];

		/*
		 * To detect the hardware bug we need to clear the entry
		 * back to zero.
		 */
		raw[0] = raw[1] = 0UL;

		/* Update head pointer of hardware ring-buffer */
		head = (head + PPR_ENTRY_SIZE) % PPR_LOG_SIZE;
		writel(head, iommu->mmio_base + MMIO_PPR_HEAD_OFFSET);

		/* Handle PPR entry */
		iommu_handle_ppr_entry(iommu, entry);

		/* Refresh ring-buffer information */
		head = readl(iommu->mmio_base + MMIO_PPR_HEAD_OFFSET);
		tail = readl(iommu->mmio_base + MMIO_PPR_TAIL_OFFSET);
	}
}

irqreturn_t amd_iommu_int_thread(int irq, void *data)
{
	struct amd_iommu *iommu = (struct amd_iommu *) data;
	u32 status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);

	while (status & (MMIO_STATUS_EVT_INT_MASK | MMIO_STATUS_PPR_INT_MASK)) {
		/* Enable EVT and PPR interrupts again */
		writel((MMIO_STATUS_EVT_INT_MASK | MMIO_STATUS_PPR_INT_MASK),
			iommu->mmio_base + MMIO_STATUS_OFFSET);

		if (status & MMIO_STATUS_EVT_INT_MASK) {
			pr_devel("AMD-Vi: Processing IOMMU Event Log\n");
			iommu_poll_events(iommu);
		}

		if (status & MMIO_STATUS_PPR_INT_MASK) {
			pr_devel("AMD-Vi: Processing IOMMU PPR Log\n");
			iommu_poll_ppr_log(iommu);
		}

		/*
		 * Hardware bug: ERBT1312
		 * When re-enabling interrupt (by writing 1
		 * to clear the bit), the hardware might also try to set
		 * the interrupt bit in the event status register.
		 * In this scenario, the bit will be set, and disable
		 * subsequent interrupts.
		 *
		 * Workaround: The IOMMU driver should read back the
		 * status register and check if the interrupt bits are cleared.
		 * If not, driver will need to go through the interrupt handler
		 * again and re-clear the bits
		 */
		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
	}
	return IRQ_HANDLED;
}

irqreturn_t amd_iommu_int_handler(int irq, void *data)
{
	return IRQ_WAKE_THREAD;
}

/****************************************************************************
 *
 * IOMMU command queuing functions
 *
 ****************************************************************************/

static int wait_on_sem(volatile u64 *sem)
{
	int i = 0;

	while (*sem == 0 && i < LOOP_TIMEOUT) {
		udelay(1);
		i += 1;
	}

	if (i == LOOP_TIMEOUT) {
		pr_alert("AMD-Vi: Completion-Wait loop timed out\n");
		return -EIO;
	}

	return 0;
}

static void copy_cmd_to_buffer(struct amd_iommu *iommu,
			       struct iommu_cmd *cmd,
			       u32 tail)
{
	u8 *target;

	target = iommu->cmd_buf + tail;
	tail   = (tail + sizeof(*cmd)) % iommu->cmd_buf_size;

	/* Copy command to buffer */
	memcpy(target, cmd, sizeof(*cmd));

	/* Tell the IOMMU about it */
	writel(tail, iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
}

static void build_completion_wait(struct iommu_cmd *cmd, u64 address)
{
	WARN_ON(address & 0x7ULL);

	memset(cmd, 0, sizeof(*cmd));
	cmd->data[0] = lower_32_bits(__pa(address)) | CMD_COMPL_WAIT_STORE_MASK;
	cmd->data[1] = upper_32_bits(__pa(address));
	cmd->data[2] = 1;
	CMD_SET_TYPE(cmd, CMD_COMPL_WAIT);
}

static void build_inv_dte(struct iommu_cmd *cmd, u16 devid)
{
	memset(cmd, 0, sizeof(*cmd));
	cmd->data[0] = devid;
	CMD_SET_TYPE(cmd, CMD_INV_DEV_ENTRY);
}

static void build_inv_iommu_pages(struct iommu_cmd *cmd, u64 address,
				  size_t size, u16 domid, int pde)
{
	u64 pages;
	bool s;

	pages = iommu_num_pages(address, size, PAGE_SIZE);
	s     = false;

	if (pages > 1) {
		/*
		 * If we have to flush more than one page, flush all
		 * TLB entries for this domain
		 */
		address = CMD_INV_IOMMU_ALL_PAGES_ADDRESS;
		s = true;
	}

	address &= PAGE_MASK;

	memset(cmd, 0, sizeof(*cmd));
	cmd->data[1] |= domid;
	cmd->data[2]  = lower_32_bits(address);
	cmd->data[3]  = upper_32_bits(address);
	CMD_SET_TYPE(cmd, CMD_INV_IOMMU_PAGES);
	if (s) /* size bit - we flush more than one 4kb page */
		cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK;
	if (pde) /* PDE bit - we want to flush everything, not only the PTEs */
		cmd->data[2] |= CMD_INV_IOMMU_PAGES_PDE_MASK;
}

static void build_inv_iotlb_pages(struct iommu_cmd *cmd, u16 devid, int qdep,
				  u64 address, size_t size)
{
	u64 pages;
	bool s;

	pages = iommu_num_pages(address, size, PAGE_SIZE);
	s     = false;

	if (pages > 1) {
		/*
		 * If we have to flush more than one page, flush all
		 * TLB entries for this domain
		 */
		address = CMD_INV_IOMMU_ALL_PAGES_ADDRESS;
		s = true;
	}

	address &= PAGE_MASK;

	memset(cmd, 0, sizeof(*cmd));
	cmd->data[0]  = devid;
	cmd->data[0] |= (qdep & 0xff) << 24;
	cmd->data[1]  = devid;
	cmd->data[2]  = lower_32_bits(address);
	cmd->data[3]  = upper_32_bits(address);
	CMD_SET_TYPE(cmd, CMD_INV_IOTLB_PAGES);
	if (s)
		cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK;
}

static void build_inv_iommu_pasid(struct iommu_cmd *cmd, u16 domid, int pasid,
				  u64 address, bool size)
{
	memset(cmd, 0, sizeof(*cmd));

	address &= ~(0xfffULL);

	cmd->data[0]  = pasid;
	cmd->data[1]  = domid;
	cmd->data[2]  = lower_32_bits(address);
	cmd->data[3]  = upper_32_bits(address);
	cmd->data[2] |= CMD_INV_IOMMU_PAGES_PDE_MASK;
	cmd->data[2] |= CMD_INV_IOMMU_PAGES_GN_MASK;
	if (size)
		cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK;
	CMD_SET_TYPE(cmd, CMD_INV_IOMMU_PAGES);
}

static void build_inv_iotlb_pasid(struct iommu_cmd *cmd, u16 devid, int pasid,
				  int qdep, u64 address, bool size)
{
	memset(cmd, 0, sizeof(*cmd));

	address &= ~(0xfffULL);

	cmd->data[0]  = devid;
	cmd->data[0] |= ((pasid >> 8) & 0xff) << 16;
	cmd->data[0] |= (qdep  & 0xff) << 24;
	cmd->data[1]  = devid;
	cmd->data[1] |= (pasid & 0xff) << 16;
	cmd->data[2]  = lower_32_bits(address);
	cmd->data[2] |= CMD_INV_IOMMU_PAGES_GN_MASK;
	cmd->data[3]  = upper_32_bits(address);
	if (size)
		cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK;
	CMD_SET_TYPE(cmd, CMD_INV_IOTLB_PAGES);
}

static void build_complete_ppr(struct iommu_cmd *cmd, u16 devid, int pasid,
			       int status, int tag, bool gn)
{
	memset(cmd, 0, sizeof(*cmd));

	cmd->data[0]  = devid;
	if (gn) {
		cmd->data[1]  = pasid;
		cmd->data[2]  = CMD_INV_IOMMU_PAGES_GN_MASK;
	}
	cmd->data[3]  = tag & 0x1ff;
	cmd->data[3] |= (status & PPR_STATUS_MASK) << PPR_STATUS_SHIFT;

	CMD_SET_TYPE(cmd, CMD_COMPLETE_PPR);
}

static void build_inv_all(struct iommu_cmd *cmd)
{
	memset(cmd, 0, sizeof(*cmd));
	CMD_SET_TYPE(cmd, CMD_INV_ALL);
}

static void build_inv_irt(struct iommu_cmd *cmd, u16 devid)
{
	memset(cmd, 0, sizeof(*cmd));
	cmd->data[0] = devid;
	CMD_SET_TYPE(cmd, CMD_INV_IRT);
}

/*
 * Writes the command to the IOMMUs command buffer and informs the
 * hardware about the new command.
 */
static int iommu_queue_command_sync(struct amd_iommu *iommu,
				    struct iommu_cmd *cmd,
				    bool sync)
{
	u32 left, tail, head, next_tail;
	unsigned long flags;

	WARN_ON(iommu->cmd_buf_size & CMD_BUFFER_UNINITIALIZED);

again:
	spin_lock_irqsave(&iommu->lock, flags);

	head      = readl(iommu->mmio_base + MMIO_CMD_HEAD_OFFSET);
	tail      = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
	next_tail = (tail + sizeof(*cmd)) % iommu->cmd_buf_size;
	left      = (head - next_tail) % iommu->cmd_buf_size;

	if (left <= 2) {
		struct iommu_cmd sync_cmd;
		volatile u64 sem = 0;
		int ret;

		build_completion_wait(&sync_cmd, (u64)&sem);
		copy_cmd_to_buffer(iommu, &sync_cmd, tail);

		spin_unlock_irqrestore(&iommu->lock, flags);

		if ((ret = wait_on_sem(&sem)) != 0)
			return ret;

		goto again;
	}

	copy_cmd_to_buffer(iommu, cmd, tail);

	/* We need to sync now to make sure all commands are processed */
	iommu->need_sync = sync;

	spin_unlock_irqrestore(&iommu->lock, flags);

	return 0;
}

static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd)
{
	return iommu_queue_command_sync(iommu, cmd, true);
}

/*
 * This function queues a completion wait command into the command
 * buffer of an IOMMU
 */
static int iommu_completion_wait(struct amd_iommu *iommu)
{
	struct iommu_cmd cmd;
	volatile u64 sem = 0;
	int ret;

	if (!iommu->need_sync)
		return 0;

	build_completion_wait(&cmd, (u64)&sem);

	ret = iommu_queue_command_sync(iommu, &cmd, false);
	if (ret)
		return ret;

	return wait_on_sem(&sem);
}

static int iommu_flush_dte(struct amd_iommu *iommu, u16 devid)
{
	struct iommu_cmd cmd;

	build_inv_dte(&cmd, devid);

	return iommu_queue_command(iommu, &cmd);
}

static void iommu_flush_dte_all(struct amd_iommu *iommu)
{
	u32 devid;

	for (devid = 0; devid <= 0xffff; ++devid)
		iommu_flush_dte(iommu, devid);

	iommu_completion_wait(iommu);
}

/*
 * This function uses heavy locking and may disable irqs for some time. But
 * this is no issue because it is only called during resume.
 */
static void iommu_flush_tlb_all(struct amd_iommu *iommu)
{
	u32 dom_id;

	for (dom_id = 0; dom_id <= 0xffff; ++dom_id) {
		struct iommu_cmd cmd;
		build_inv_iommu_pages(&cmd, 0, CMD_INV_IOMMU_ALL_PAGES_ADDRESS,
				      dom_id, 1);
		iommu_queue_command(iommu, &cmd);
	}

	iommu_completion_wait(iommu);
}

static void iommu_flush_all(struct amd_iommu *iommu)
{
	struct iommu_cmd cmd;

	build_inv_all(&cmd);

	iommu_queue_command(iommu, &cmd);
	iommu_completion_wait(iommu);
}

static void iommu_flush_irt(struct amd_iommu *iommu, u16 devid)
{
	struct iommu_cmd cmd;

	build_inv_irt(&cmd, devid);

	iommu_queue_command(iommu, &cmd);
}

static void iommu_flush_irt_all(struct amd_iommu *iommu)
{
	u32 devid;

	for (devid = 0; devid <= MAX_DEV_TABLE_ENTRIES; devid++)
		iommu_flush_irt(iommu, devid);

	iommu_completion_wait(iommu);
}

void iommu_flush_all_caches(struct amd_iommu *iommu)
{
	if (iommu_feature(iommu, FEATURE_IA)) {
		iommu_flush_all(iommu);
	} else {
		iommu_flush_dte_all(iommu);
		iommu_flush_irt_all(iommu);
		iommu_flush_tlb_all(iommu);
	}
}

/*
 * Command send function for flushing on-device TLB
 */
static int device_flush_iotlb(struct iommu_dev_data *dev_data,
			      u64 address, size_t size)
{
	struct amd_iommu *iommu;
	struct iommu_cmd cmd;
	int qdep;

	qdep     = dev_data->ats.qdep;
	iommu    = amd_iommu_rlookup_table[dev_data->devid];

	build_inv_iotlb_pages(&cmd, dev_data->devid, qdep, address, size);

	return iommu_queue_command(iommu, &cmd);
}

/*
 * Command send function for invalidating a device table entry
 */
static int device_flush_dte(struct iommu_dev_data *dev_data)
{
	struct amd_iommu *iommu;
	int ret;

	iommu = amd_iommu_rlookup_table[dev_data->devid];

	ret = iommu_flush_dte(iommu, dev_data->devid);
	if (ret)
		return ret;

	if (dev_data->ats.enabled)
		ret = device_flush_iotlb(dev_data, 0, ~0UL);

	return ret;
}

/*
 * TLB invalidation function which is called from the mapping functions.
 * It invalidates a single PTE if the range to flush is within a single
 * page. Otherwise it flushes the whole TLB of the IOMMU.
 */
static void __domain_flush_pages(struct protection_domain *domain,
				 u64 address, size_t size, int pde)
{
	struct iommu_dev_data *dev_data;
	struct iommu_cmd cmd;
	int ret = 0, i;

	build_inv_iommu_pages(&cmd, address, size, domain->id, pde);

	for (i = 0; i < amd_iommus_present; ++i) {
		if (!domain->dev_iommu[i])
			continue;

		/*
		 * Devices of this domain are behind this IOMMU
		 * We need a TLB flush
		 */
		ret |= iommu_queue_command(amd_iommus[i], &cmd);
	}

	list_for_each_entry(dev_data, &domain->dev_list, list) {

		if (!dev_data->ats.enabled)
			continue;

		ret |= device_flush_iotlb(dev_data, address, size);
	}

	WARN_ON(ret);
}

static void domain_flush_pages(struct protection_domain *domain,
			       u64 address, size_t size)
{
	__domain_flush_pages(domain, address, size, 0);
}

/* Flush the whole IO/TLB for a given protection domain */
static void domain_flush_tlb(struct protection_domain *domain)
{
	__domain_flush_pages(domain, 0, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, 0);
}

/* Flush the whole IO/TLB for a given protection domain - including PDE */
static void domain_flush_tlb_pde(struct protection_domain *domain)
{
	__domain_flush_pages(domain, 0, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, 1);
}

static void domain_flush_complete(struct protection_domain *domain)
{
	int i;

	for (i = 0; i < amd_iommus_present; ++i) {
		if (!domain->dev_iommu[i