Merge branch 'core' into x86/vt-d

Conflicts:
	drivers/iommu/intel-iommu.c

Documentation/ABI/testing/sysfs-class-iommu

@@ -0,0 +1,17 @@
+What:		/sys/class/iommu/<iommu>/devices/
+Date:		June 2014
+KernelVersion:	3.17
+Contact:	Alex Williamson <alex.williamson@redhat.com>
+Description:
+		IOMMU drivers are able to link devices managed by a
+		given IOMMU here to allow association of IOMMU to
+		device.
+
+What:		/sys/devices/.../iommu
+Date:		June 2014
+KernelVersion:	3.17
+Contact:	Alex Williamson <alex.williamson@redhat.com>
+Description:
+		IOMMU drivers are able to link the IOMMU for a
+		given device here to allow association of device to
+		IOMMU.

Documentation/ABI/testing/sysfs-class-iommu-amd-iommu

@@ -0,0 +1,14 @@
+What:		/sys/class/iommu/<iommu>/amd-iommu/cap
+Date:		June 2014
+KernelVersion:	3.17
+Contact:	Alex Williamson <alex.williamson@redhat.com>
+Description:
+		IOMMU capability header as documented in the AMD IOMMU
+		specification.  Format: %x
+
+What:		/sys/class/iommu/<iommu>/amd-iommu/features
+Date:		June 2014
+KernelVersion:	3.17
+Contact:	Alex Williamson <alex.williamson@redhat.com>
+Description:
+		Extended features of the IOMMU.  Format: %llx

Documentation/ABI/testing/sysfs-class-iommu-intel-iommu

@@ -0,0 +1,32 @@
+What:		/sys/class/iommu/<iommu>/intel-iommu/address
+Date:		June 2014
+KernelVersion:	3.17
+Contact:	Alex Williamson <alex.williamson@redhat.com>
+Description:
+		Physical address of the VT-d DRHD for this IOMMU.
+		Format: %llx.  This allows association of a sysfs
+		intel-iommu with a DMAR DRHD table entry.
+
+What:		/sys/class/iommu/<iommu>/intel-iommu/cap
+Date:		June 2014
+KernelVersion:	3.17
+Contact:	Alex Williamson <alex.williamson@redhat.com>
+Description:
+		The cached hardware capability register value
+		of this DRHD unit.  Format: %llx.
+
+What:		/sys/class/iommu/<iommu>/intel-iommu/ecap
+Date:		June 2014
+KernelVersion:	3.17
+Contact:	Alex Williamson <alex.williamson@redhat.com>
+Description:
+		The cached hardware extended capability register
+		value of this DRHD unit.  Format: %llx.
+
+What:		/sys/class/iommu/<iommu>/intel-iommu/version
+Date:		June 2014
+KernelVersion:	3.17
+Contact:	Alex Williamson <alex.williamson@redhat.com>
+Description:
+		The architecture version as reported from the
+		VT-d VER_REG.  Format: %d:%d, major:minor

drivers/iommu/Makefile

@@ -1,5 +1,6 @@
 obj-$(CONFIG_IOMMU_API) += iommu.o
 obj-$(CONFIG_IOMMU_API) += iommu-traces.o
+obj-$(CONFIG_IOMMU_API) += iommu-sysfs.o
 obj-$(CONFIG_OF_IOMMU)	+= of_iommu.o
 obj-$(CONFIG_MSM_IOMMU) += msm_iommu.o msm_iommu_dev.o
 obj-$(CONFIG_AMD_IOMMU) += amd_iommu.o amd_iommu_init.o

drivers/iommu/amd_iommu.c

@@ -46,7 +46,6 @@
 #include "amd_iommu_proto.h"
 #include "amd_iommu_types.h"
 #include "irq_remapping.h"
-#include "pci.h"
 
 #define CMD_SET_TYPE(cmd, t) ((cmd)->data[1] |= ((t) << 28))
 
@@ -81,7 +80,7 @@ LIST_HEAD(hpet_map);
  */
 static struct protection_domain *pt_domain;
 
-static struct iommu_ops amd_iommu_ops;
+static const struct iommu_ops amd_iommu_ops;
 
 static ATOMIC_NOTIFIER_HEAD(ppr_notifier);
 int amd_iommu_max_glx_val = -1;
@@ -133,9 +132,6 @@ static void free_dev_data(struct iommu_dev_data *dev_data)
 	list_del(&dev_data->dev_data_list);
 	spin_unlock_irqrestore(&dev_data_list_lock, flags);
 
-	if (dev_data->group)
-		iommu_group_put(dev_data->group);
-
 	kfree(dev_data);
 }
 
@@ -264,167 +260,79 @@ static bool check_device(struct device *dev)
 	return true;
 }
 
-static struct pci_bus *find_hosted_bus(struct pci_bus *bus)
-{
-	while (!bus->self) {
-		if (!pci_is_root_bus(bus))
-			bus = bus->parent;
-		else
-			return ERR_PTR(-ENODEV);
-	}
-
-	return bus;
-}
-
-#define REQ_ACS_FLAGS	(PCI_ACS_SV | PCI_ACS_RR | PCI_ACS_CR | PCI_ACS_UF)
-
-static struct pci_dev *get_isolation_root(struct pci_dev *pdev)
-{
-	struct pci_dev *dma_pdev = pdev;
-
-	/* Account for quirked devices */
-	swap_pci_ref(&dma_pdev, pci_get_dma_source(dma_pdev));
-
-	/*
-	 * If it's a multifunction device that does not support our
-	 * required ACS flags, add to the same group as lowest numbered
-	 * function that also does not suport the required ACS flags.
-	 */
-	if (dma_pdev->multifunction &&
-	    !pci_acs_enabled(dma_pdev, REQ_ACS_FLAGS)) {
-		u8 i, slot = PCI_SLOT(dma_pdev->devfn);
-
-		for (i = 0; i < 8; i++) {
-			struct pci_dev *tmp;
-
-			tmp = pci_get_slot(dma_pdev->bus, PCI_DEVFN(slot, i));
-			if (!tmp)
-				continue;
-
-			if (!pci_acs_enabled(tmp, REQ_ACS_FLAGS)) {
-				swap_pci_ref(&dma_pdev, tmp);
-				break;
-			}
-			pci_dev_put(tmp);
-		}
-	}
-
-	/*
-	 * Devices on the root bus go through the iommu.  If that's not us,
-	 * find the next upstream device and test ACS up to the root bus.
-	 * Finding the next device may require skipping virtual buses.
-	 */
-	while (!pci_is_root_bus(dma_pdev->bus)) {
-		struct pci_bus *bus = find_hosted_bus(dma_pdev->bus);
-		if (IS_ERR(bus))
-			break;
-
-		if (pci_acs_path_enabled(bus->self, NULL, REQ_ACS_FLAGS))
-			break;
-
-		swap_pci_ref(&dma_pdev, pci_dev_get(bus->self));
-	}
-
-	return dma_pdev;
-}
-
-static int use_pdev_iommu_group(struct pci_dev *pdev, struct device *dev)
+static int init_iommu_group(struct device *dev)
 {
-	struct iommu_group *group = iommu_group_get(&pdev->dev);
-	int ret;
+	struct iommu_group *group;
 
-	if (!group) {
-		group = iommu_group_alloc();
-		if (IS_ERR(group))
-			return PTR_ERR(group);
+	group = iommu_group_get_for_dev(dev);
 
-		WARN_ON(&pdev->dev != dev);
-	}
+	if (IS_ERR(group))
+		return PTR_ERR(group);
 
-	ret = iommu_group_add_device(group, dev);
 	iommu_group_put(group);
-	return ret;
+	return 0;
 }
 
-static int use_dev_data_iommu_group(struct iommu_dev_data *dev_data,
-				    struct device *dev)
+static int __last_alias(struct pci_dev *pdev, u16 alias, void *data)
 {
-	if (!dev_data->group) {
-		struct iommu_group *group = iommu_group_alloc();
-		if (IS_ERR(group))
-			return PTR_ERR(group);
-
-		dev_data->group = group;
-	}
-
-	return iommu_group_add_device(dev_data->group, dev);
+	*(u16 *)data = alias;
+	return 0;
 }
 
-static int init_iommu_group(struct device *dev)
+static u16 get_alias(struct device *dev)
 {
-	struct iommu_dev_data *dev_data;
-	struct iommu_group *group;
-	struct pci_dev *dma_pdev;
-	int ret;
-
-	group = iommu_group_get(dev);
-	if (group) {
-		iommu_group_put(group);
-		return 0;
-	}
-
-	dev_data = find_dev_data(get_device_id(dev));
-	if (!dev_data)
-		return -ENOMEM;
+	struct pci_dev *pdev = to_pci_dev(dev);
+	u16 devid, ivrs_alias, pci_alias;
 
-	if (dev_data->alias_data) {
-		u16 alias;
-		struct pci_bus *bus;
+	devid = get_device_id(dev);
+	ivrs_alias = amd_iommu_alias_table[devid];
+	pci_for_each_dma_alias(pdev, __last_alias, &pci_alias);
 
-		if (dev_data->alias_data->group)
-			goto use_group;
+	if (ivrs_alias == pci_alias)
+		return ivrs_alias;
 
-		/*
-		 * If the alias device exists, it's effectively just a first
-		 * level quirk for finding the DMA source.
-		 */
-		alias = amd_iommu_alias_table[dev_data->devid];
-		dma_pdev = pci_get_bus_and_slot(alias >> 8, alias & 0xff);
-		if (dma_pdev) {
-			dma_pdev = get_isolation_root(dma_pdev);
-			goto use_pdev;
+	/*
+	 * 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));
 		}
 
-		/*
-		 * If the alias is virtual, try to find a parent device
-		 * and test whether the IOMMU group is actualy rooted above
-		 * the alias.  Be careful to also test the parent device if
-		 * we think the alias is the root of the group.
-		 */
-		bus = pci_find_bus(0, alias >> 8);
-		if (!bus)
-			goto use_group;
-
-		bus = find_hosted_bus(bus);
-		if (IS_ERR(bus) || !bus->self)
-			goto use_group;
+		return pci_alias;
+	}
 
-		dma_pdev = get_isolation_root(pci_dev_get(bus->self));
-		if (dma_pdev != bus->self || (dma_pdev->multifunction &&
-		    !pci_acs_enabled(dma_pdev, REQ_ACS_FLAGS)))
-			goto use_pdev;
+	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));
 
-		pci_dev_put(dma_pdev);
-		goto use_group;
+	/*
+	 * 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));
 	}
 
-	dma_pdev = get_isolation_root(pci_dev_get(to_pci_dev(dev)));
-use_pdev:
-	ret = use_pdev_iommu_group(dma_pdev, dev);
-	pci_dev_put(dma_pdev);
-	return ret;
-use_group:
-	return use_dev_data_iommu_group(dev_data->alias_data, dev);
+	return ivrs_alias;
 }
 
 static int iommu_init_device(struct device *dev)
@@ -441,7 +349,8 @@ static int iommu_init_device(struct device *dev)
 	if (!dev_data)
 		return -ENOMEM;
 
-	alias = amd_iommu_alias_table[dev_data->devid];
+	alias = get_alias(dev);
+
 	if (alias != dev_data->devid) {
 		struct iommu_dev_data *alias_data;
 
@@ -470,6 +379,9 @@ static int iommu_init_device(struct device *dev)
 
 	dev->archdata.iommu = dev_data;
 
+	iommu_device_link(amd_iommu_rlookup_table[dev_data->devid]->iommu_dev,
+			  dev);
+
 	return 0;
 }
 
@@ -489,12 +401,22 @@ static void iommu_ignore_device(struct device *dev)
 
 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;
+
 	/*
-	 * Nothing to do here - 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.
+	 * 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.
 	 */
 }
 
@@ -3473,7 +3395,7 @@ static int amd_iommu_domain_has_cap(struct iommu_domain *domain,
 	return 0;
 }
 
-static struct iommu_ops amd_iommu_ops = {
+static const struct iommu_ops amd_iommu_ops = {
 	.domain_init = amd_iommu_domain_init,
 	.domain_destroy = amd_iommu_domain_destroy,
 	.attach_dev = amd_iommu_attach_device,

drivers/iommu/amd_iommu_init.c

@@ -26,6 +26,7 @@
 #include <linux/msi.h>
 #include <linux/amd-iommu.h>
 #include <linux/export.h>
+#include <linux/iommu.h>
 #include <asm/pci-direct.h>
 #include <asm/iommu.h>
 #include <asm/gart.h>
@@ -1197,6 +1198,39 @@ static void init_iommu_perf_ctr(struct amd_iommu *iommu)
 	iommu->max_counters = (u8) ((val >> 7) & 0xf);
 }
 
+static ssize_t amd_iommu_show_cap(struct device *dev,
+				  struct device_attribute *attr,
+				  char *buf)
+{
+	struct amd_iommu *iommu = dev_get_drvdata(dev);
+	return sprintf(buf, "%x\n", iommu->cap);
+}
+static DEVICE_ATTR(cap, S_IRUGO, amd_iommu_show_cap, NULL);
+
+static ssize_t amd_iommu_show_features(struct device *dev,
+				       struct device_attribute *attr,
+				       char *buf)
+{
+	struct amd_iommu *iommu = dev_get_drvdata(dev);
+	return sprintf(buf, "%llx\n", iommu->features);
+}
+static DEVICE_ATTR(features, S_IRUGO, amd_iommu_show_features, NULL);
+
+static struct attribute *amd_iommu_attrs[] = {
+	&dev_attr_cap.attr,
+	&dev_attr_features.attr,
+	NULL,
+};
+
+static struct attribute_group amd_iommu_group = {
+	.name = "amd-iommu",
+	.attrs = amd_iommu_attrs,
+};
+
+static const struct attribute_group *amd_iommu_groups[] = {
+	&amd_iommu_group,
+	NULL,
+};
 
 static int iommu_init_pci(struct amd_iommu *iommu)
 {
@@ -1297,6 +1331,10 @@ static int iommu_init_pci(struct amd_iommu *iommu)
 
 	amd_iommu_erratum_746_workaround(iommu);
 
+	iommu->iommu_dev = iommu_device_create(&iommu->dev->dev, iommu,
+					       amd_iommu_groups, "ivhd%d",
+					       iommu->index);
+
 	return pci_enable_device(iommu->dev);
 }
 

drivers/iommu/amd_iommu_types.h

@@ -432,7 +432,6 @@ struct iommu_dev_data {
 	struct iommu_dev_data *alias_data;/* The alias dev_data */
 	struct protection_domain *domain; /* Domain the device is bound to */
 	atomic_t bind;			  /* Domain attach reference count */
-	struct iommu_group *group;	  /* IOMMU group for virtual aliases */
 	u16 devid;			  /* PCI Device ID */
 	bool iommu_v2;			  /* Device can make use of IOMMUv2 */
 	bool passthrough;		  /* Default for device is pt_domain */
@@ -578,6 +577,9 @@ struct amd_iommu {
 	/* default dma_ops domain for that IOMMU */
 	struct dma_ops_domain *default_dom;
 
+	/* IOMMU sysfs device */
+	struct device *iommu_dev;
+
 	/*
 	 * We can't rely on the BIOS to restore all values on reinit, so we
 	 * need to stash them

drivers/iommu/arm-smmu.c

@@ -1609,7 +1609,7 @@ static void arm_smmu_remove_device(struct device *dev)
 	iommu_group_remove_device(dev);
 }
 
-static struct iommu_ops arm_smmu_ops = {
+static const struct iommu_ops arm_smmu_ops = {
 	.domain_init	= arm_smmu_domain_init,
 	.domain_destroy	= arm_smmu_domain_destroy,
 	.attach_dev	= arm_smmu_attach_dev,

drivers/iommu/dmar.c

@@ -38,6 +38,7 @@
 #include <linux/tboot.h>
 #include <linux/dmi.h>
 #include <linux/slab.h>
+#include <linux/iommu.h>
 #include <asm/irq_remapping.h>
 #include <asm/iommu_table.h>
 
@@ -980,6 +981,12 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd)
 	raw_spin_lock_init(&iommu->register_lock);
 
 	drhd->iommu = iommu;
+
+	if (intel_iommu_enabled)
+		iommu->iommu_dev = iommu_device_create(NULL, iommu,
+						       intel_iommu_groups,
+						       iommu->name);
+
 	return 0;
 
  err_unmap:
@@ -991,6 +998,8 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd)
 
 static void free_iommu(struct intel_iommu *iommu)
 {
+	iommu_device_destroy(iommu->iommu_dev);
+
 	if (iommu->irq) {
 		free_irq(iommu->irq, iommu);
 		irq_set_handler_data(iommu->irq, NULL);

drivers/iommu/exynos-iommu.c

@@ -1170,7 +1170,7 @@ static void exynos_iommu_remove_device(struct device *dev)
 	iommu_group_remove_device(dev);
 }
 
-static struct iommu_ops exynos_iommu_ops = {
+static const struct iommu_ops exynos_iommu_ops = {
 	.domain_init = exynos_iommu_domain_init,
 	.domain_destroy = exynos_iommu_domain_destroy,
 	.attach_dev = exynos_iommu_attach_device,

drivers/iommu/fsl_pamu_domain.c

@@ -38,7 +38,6 @@
 #include <sysdev/fsl_pci.h>
 
 #include "fsl_pamu_domain.h"
-#include "pci.h"
 
 /*
  * Global spinlock that needs to be held while
@@ -892,8 +891,6 @@ static int fsl_pamu_get_domain_attr(struct iommu_domain *domain,
 	return ret;
 }
 
-#define REQ_ACS_FLAGS	(PCI_ACS_SV | PCI_ACS_RR | PCI_ACS_CR | PCI_ACS_UF)
-
 static struct iommu_group *get_device_iommu_group(struct device *dev)
 {
 	struct iommu_group *group;
@@ -950,74 +947,13 @@ static struct iommu_group *get_pci_device_group(struct pci_dev *pdev)
 	struct pci_controller *pci_ctl;
 	bool pci_endpt_partioning;
 	struct iommu_group *group = NULL;
-	struct pci_dev *bridge, *dma_pdev = NULL;
 
 	pci_ctl = pci_bus_to_host(pdev->bus);
 	pci_endpt_partioning = check_pci_ctl_endpt_part(pci_ctl);
 	/* We can partition PCIe devices so assign device group to the device */
 	if (pci_endpt_partioning) {
-		bridge = pci_find_upstream_pcie_bridge(pdev);
-		if (bridge) {
-			if (pci_is_pcie(bridge))
-				dma_pdev = pci_get_domain_bus_and_slot(
-						pci_domain_nr(pdev->bus),
-						bridge->subordinate->number, 0);
-			if (!dma_pdev)
-				dma_pdev = pci_dev_get(bridge);
-		} else
-			dma_pdev = pci_dev_get(pdev);
-
-		/* Account for quirked devices */
-		swap_pci_ref(&dma_pdev, pci_get_dma_source(dma_pdev));
-
-		/*
-		 * If it's a multifunction device that does not support our
-		 * required ACS flags, add to the same group as lowest numbered
-		 * function that also does not suport the required ACS flags.
-		 */
-		if (dma_pdev->multifunction &&
-		    !pci_acs_enabled(dma_pdev, REQ_ACS_FLAGS)) {
-			u8 i, slot = PCI_SLOT(dma_pdev->devfn);
-
-			for (i = 0; i < 8; i++) {
-				struct pci_dev *tmp;
-
-				tmp = pci_get_slot(dma_pdev->bus, PCI_DEVFN(slot, i));
-				if (!tmp)
-					continue;
-
-				if (!pci_acs_enabled(tmp, REQ_ACS_FLAGS)) {
-					swap_pci_ref(&dma_pdev, tmp);
-					break;
-				}
-				pci_dev_put(tmp);
-			}
-		}
-
-		/*
-		 * Devices on the root bus go through the iommu.  If that's not us,
-		 * find the next upstream device and test ACS up to the root bus.
-		 * Finding the next device may require skipping virtual buses.
-		 */
-		while (!pci_is_root_bus(dma_pdev->bus)) {
-			struct pci_bus *bus = dma_pdev->bus;
-
-			while (!bus->self) {
-				if (!pci_is_root_bus(bus))
-					bus = bus->parent;
-				else
-					goto root_bus;
-			}
-
-			if (pci_acs_path_enabled(bus->self, NULL, REQ_ACS_FLAGS))
-				break;
-
-			swap_pci_ref(&dma_pdev, pci_dev_get(bus->self));
-		}
+		group = iommu_group_get_for_dev(&pdev->dev);
 
-root_bus:
-		group = get_device_iommu_group(&dma_pdev->dev);
-		pci_dev_put(dma_pdev);
 		/*
 		 * PCIe controller is not a paritionable entity
 		 * free the controller device iommu_group.
@@ -1140,7 +1076,7 @@ static u32 fsl_pamu_get_windows(struct iommu_domain *domain)
 	return dma_domain->win_cnt;
 }
 
-static struct iommu_ops fsl_pamu_ops = {
+static const struct iommu_ops fsl_pamu_ops = {
 	.domain_init	= fsl_pamu_domain_init,
 	.domain_destroy = fsl_pamu_domain_destroy,
 	.attach_dev	= fsl_pamu_attach_device,

drivers/iommu/intel-iommu.c

@@ -45,7 +45,6 @@
 #include <asm/iommu.h>
 
 #include "irq_remapping.h"
-#include "pci.h"
 
 #define ROOT_SIZE		VTD_PAGE_SIZE
 #define CONTEXT_SIZE		VTD_PAGE_SIZE
@@ -451,7 +450,7 @@ EXPORT_SYMBOL_GPL(intel_iommu_gfx_mapped);
 static DEFINE_SPINLOCK(device_domain_lock);
 static LIST_HEAD(device_domain_list);
 
-static struct iommu_ops intel_iommu_ops;
+static const struct iommu_ops intel_iommu_ops;
 
 static int __init intel_iommu_setup(char *str)
 {
@@ -1840,54 +1839,56 @@ static int domain_context_mapping_one(struct dmar_domain *domain,
 	return 0;
 }
 
+struct domain_context_mapping_data {
+	struct dmar_domain *domain;
+	struct intel_iommu *iommu;
+	int translation;
+};
+
+static int domain_context_mapping_cb(struct pci_dev *pdev,
+				     u16 alias, void *opaque)
+{
+	struct domain_context_mapping_data *data = opaque;
+
+	return domain_context_mapping_one(data->domain, data->iommu,
+					  PCI_BUS_NUM(alias), alias & 0xff,
+					  data->translation);
+}
+
 static int
 domain_context_mapping(struct dmar_domain *domain, struct device *dev,
 		       int translation)
 {
-	int ret;
-	struct pci_dev *pdev, *tmp, *parent;
 	struct intel_iommu *iommu;
 	u8 bus, devfn;
+	struct domain_context_mapping_data data;
 
 	iommu = device_to_iommu(dev, &bus, &devfn);
 	if (!iommu)
 		return -ENODEV;
 
-	ret = domain_context_mapping_one(domain, iommu, bus, devfn,
-					 translation);
-	if (ret || !dev_is_pci(dev))
-		return ret;
-
-	/* dependent device mapping */
-	pdev = to_pci_dev(dev);
-	tmp = pci_find_upstream_pcie_bridge(pdev);
-	if (!tmp)
-		return 0;
-	/* Secondary interface's bus number and devfn 0 */
-	parent = pdev->bus->self;
-	while (parent != tmp) {
-		ret = domain_context_mapping_one(domain, iommu,
-						 parent->bus->number,
-						 parent->devfn, translation);
-		if (ret)
-			return ret;
-		parent = parent->bus->self;
-	}
-	if (pci_is_pcie(tmp)) /* this is a PCIe-to-PCI bridge */
-		return domain_context_mapping_one(domain, iommu,
-					tmp->subordinate->number, 0,
-					translation);
-	else /* this is a legacy PCI bridge */
-		return domain_context_mapping_one(domain, iommu,
-						  tmp->bus->number,
-						  tmp->devfn,
+	if (!dev_is_pci(dev))
+		return domain_context_mapping_one(domain, iommu, bus, devfn,
 						  translation);
+
+	data.domain = domain;
+	data.iommu = iommu;
+	data.translation = translation;
+
+	return pci_for_each_dma_alias(to_pci_dev(dev),
+				      &domain_context_mapping_cb, &data);
+}
+
+static int domain_context_mapped_cb(struct pci_dev *pdev,
+				    u16 alias, void *opaque)
+{
+	struct intel_iommu *iommu = opaque;
+
+	return !device_context_mapped(iommu, PCI_BUS_NUM(alias), alias & 0xff);
 }
 
 static int domain_context_mapped(struct device *dev)
 {
-	int ret;
-	struct pci_dev *pdev, *tmp, *parent;
 	struct intel_iommu *iommu;
 	u8 bus, devfn;
 
@@ -1895,30 +1896,11 @@ static int domain_context_mapped(struct device *dev)
 	if (!iommu)
 		return -ENODEV;
 
-	ret = device_context_mapped(iommu, bus, devfn);
-	if (!ret || !dev_is_pci(dev))
-		return ret;
+	if (!dev_is_pci(dev))
+		return device_context_mapped(iommu, bus, devfn);
 
-	/* dependent device mapping */
-	pdev = to_pci_dev(dev);
-	tmp = pci_find_upstream_pcie_bridge(pdev);
-	if (!tmp)
-		return ret;
-	/* Secondary interface's bus number and devfn 0 */
-	parent = pdev->bus->self;
-	while (parent != tmp) {
-		ret = device_context_mapped(iommu, parent->bus->number,
-					    parent->devfn);
-		if (!ret)
-			return ret;
-		parent = parent->bus->self;
-	}
-	if (pci_is_pcie(tmp))
-		return device_context_mapped(iommu, tmp->subordinate->number,
-					     0);
-	else
-		return device_context_mapped(iommu, tmp->bus->number,
-					     tmp->devfn);
+	return !pci_for_each_dma_alias(to_pci_dev(dev),
+				       domain_context_mapped_cb, iommu);
 }
 
 /* Returns a number of VTD pages, but aligned to MM page size */
@@ -2205,80 +2187,86 @@ static struct dmar_domain *dmar_insert_dev_info(struct intel_iommu *iommu,
 	return domain;
 }
 
+static int get_last_alias(struct pci_dev *pdev, u16 alias, void *opaque)
+{
+	*(u16 *)opaque = alias;
+	return 0;
+}
+
 /* domain is initialized */
 static struct dmar_domain *get_domain_for_dev(struct device *dev, int gaw)
 {
-	struct dmar_domain *domain, *free = NULL;
-	struct intel_iommu *iommu = NULL;
+	struct dmar_domain *domain, *tmp;
+	struct intel_iommu *iommu;
 	struct device_domain_info *info;
-	struct pci_dev *dev_tmp = NULL;
+	u16 dma_alias;
 	unsigned long flags;
 	u8 bus, devfn;
-	u8 bridge_bus = 0, bridge_devfn = 0;
 
 	domain = find_domain(dev);
 	if (domain)
 		return domain;
 
+	iommu = device_to_iommu(dev, &bus, &devfn);
+	if (!iommu)
+		return NULL;
+
 	if (dev_is_pci(dev)) {
 		struct pci_dev *pdev = to_pci_dev(dev);
-		u16 segment;
 
-		segment = pci_domain_nr(pdev->bus);
-		dev_tmp = pci_find_upstream_pcie_bridge(pdev);
-		if (dev_tmp) {
-			if (pci_is_pcie(dev_tmp)) {
-				bridge_bus = dev_tmp->subordinate->number;
-				bridge_devfn = 0;
-			} else {
-				bridge_bus = dev_tmp->bus->number;
-				bridge_devfn = dev_tmp->devfn;
-			}
-			spin_lock_irqsave(&device_domain_lock, flags);
-			info = dmar_search_domain_by_dev_info(segment,
-							      bridge_bus,
-							      bridge_devfn);
-			if (info) {
-				iommu = info->iommu;
-				domain = info->domain;
-			}
-			spin_unlock_irqrestore(&device_domain_lock, flags);
-			/* pcie-pci bridge already has a domain, uses it */
-			if (info)
-				goto found_domain;
+		pci_for_each_dma_alias(pdev, get_last_alias, &dma_alias);
+
+		spin_lock_irqsave(&device_domain_lock, flags);
+		info = dmar_search_domain_by_dev_info(pci_domain_nr(pdev->bus),
+						      PCI_BUS_NUM(dma_alias),
+						      dma_alias & 0xff);
+		if (info) {
+			iommu = info->iommu;
+			domain = info->domain;
 		}
-	}
+		spin_unlock_irqrestore(&device_domain_lock, flags);
 
-	iommu = device_to_iommu(dev, &bus, &devfn);
-	if (!iommu)
-		goto error;
+		/* DMA alias already has a domain, uses it */
+		if (info)
+			goto found_domain;
+	}
 
 	/* Allocate and initialize new domain for the device */
 	domain = alloc_domain(false);
 	if (!domain)
-		goto error;
+		return NULL;
+
 	if (iommu_attach_domain(domain, iommu)) {
 		free_domain_mem(domain);
-		domain = NULL;
-		goto error;
+		return NULL;
 	}
-	free = domain;
-	if (domain_init(domain, gaw))
-		goto error;
 
-	/* register pcie-to-pci device */
-	if (dev_tmp) {
-		domain = dmar_insert_dev_info(iommu, bridge_bus, bridge_devfn,
-					      NULL, domain);
+	if (domain_init(domain, gaw)) {
+		domain_exit(domain);
+		return NULL;
+	}
+
+	/* register PCI DMA alias device */
+	if (dev_is_pci(dev)) {
+		tmp = dmar_insert_dev_info(iommu, PCI_BUS_NUM(dma_alias),
+					   dma_alias & 0xff, NULL, domain);
+
+		if (!tmp || tmp != domain) {
+			domain_exit(domain);
+			domain = tmp;
+		}
+
 		if (!domain)
-			goto error;
+			return NULL;
 	}
 
 found_domain:
-	domain = dmar_insert_dev_info(iommu, bus, devfn, dev, domain);
-error:
-	if (free != domain)
-		domain_exit(free);
+	tmp = dmar_insert_dev_info(iommu, bus, devfn, dev, domain);
+
+	if (!tmp || tmp != domain) {
+		domain_exit(domain);
+		domain = tmp;
+	}
 
 	return domain;
 }
@@ -3953,6 +3941,63 @@ static struct notifier_block intel_iommu_memory_nb = {
 	.priority = 0
 };
 
+
+static ssize_t intel_iommu_show_version(struct device *dev,
+					struct device_attribute *attr,
+					char *buf)
+{
+	struct intel_iommu *iommu = dev_get_drvdata(dev);
+	u32 ver = readl(iommu->reg + DMAR_VER_REG);
+	return sprintf(buf, "%d:%d\n",
+		       DMAR_VER_MAJOR(ver), DMAR_VER_MINOR(ver));
+}
+static DEVICE_ATTR(version, S_IRUGO, intel_iommu_show_version, NULL);
+
+static ssize_t intel_iommu_show_address(struct device *dev,
+					struct device_attribute *attr,
+					char *buf)
+{
+	struct intel_iommu *iommu = dev_get_drvdata(dev);
+	return sprintf(buf, "%llx\n", iommu->reg_phys);
+}
+static DEVICE_ATTR(address, S_IRUGO, intel_iommu_show_address, NULL);
+
+static ssize_t intel_iommu_show_cap(struct device *dev,
+				    struct device_attribute *attr,
+				    char *buf)
+{
+	struct intel_iommu *iommu = dev_get_drvdata(dev);
+	return sprintf(buf, "%llx\n", iommu->cap);
+}
+static DEVICE_ATTR(cap, S_IRUGO, intel_iommu_show_cap, NULL);
+
+static ssize_t intel_iommu_show_ecap(struct device *dev,
+				    struct device_attribute *attr,
+				    char *buf)
+{
+	struct intel_iommu *iommu = dev_get_drvdata(dev);
+	return sprintf(buf, "%llx\n", iommu->ecap);
+}
+static DEVICE_ATTR(ecap, S_IRUGO, intel_iommu_show_ecap, NULL);
+
+static struct attribute *intel_iommu_attrs[] = {
+	&dev_attr_version.attr,
+	&dev_attr_address.attr,
+	&dev_attr_cap.attr,
+	&dev_attr_ecap.attr,
+	NULL,
+};
+
+static struct attribute_group intel_iommu_group = {
+	.name = "intel-iommu",
+	.attrs = intel_iommu_attrs,
+};
+
+const struct attribute_group *intel_iommu_groups[] = {
+	&intel_iommu_group,
+	NULL,
+};
+
 int __init intel_iommu_init(void)
 {
 	int ret = -ENODEV;
@@ -4024,6 +4069,11 @@ int __init intel_iommu_init(void)
 
 	init_iommu_pm_ops();
 
+	for_each_active_iommu(iommu, drhd)
+		iommu->iommu_dev = iommu_device_create(NULL, iommu,
+						       intel_iommu_groups,
+						       iommu->name);
+
 	bus_set_iommu(&pci_bus_type, &intel_iommu_ops);
 	bus_register_notifier(&pci_bus_type, &device_nb);
 	if (si_domain && !hw_pass_through)
@@ -4042,33 +4092,27 @@ out_free_dmar:
 	return ret;
 }
 
+static int iommu_detach_dev_cb(struct pci_dev *pdev, u16 alias, void *opaque)
+{
+	struct intel_iommu *iommu = opaque;
+
+	iommu_detach_dev(iommu, PCI_BUS_NUM(alias), alias & 0xff);
+	return 0;
+}
+
+/*
+ * NB - intel-iommu lacks any sort of reference counting for the users of
+ * dependent devices.  If multiple endpoints have intersecting dependent
+ * devices, unbinding the driver from any one of them will possibly leave
+ * the others unable to operate.
+ */
 static void iommu_detach_dependent_devices(struct intel_iommu *iommu,
 					   struct device *dev)
 {
-	struct pci_dev *tmp, *parent, *pdev;
-
 	if (!iommu || !dev || !dev_is_pci(dev))
 		return;
 
-	pdev = to_pci_dev(dev);
-
-	/* dependent device detach */
-	tmp = pci_find_upstream_pcie_bridge(pdev);
-	/* Secondary interface's bus number and devfn 0 */
-	if (tmp) {
-		parent = pdev->bus->self;
-		while (parent != tmp) {
-			iommu_detach_dev(iommu, parent->bus->number,
-					 parent->devfn);
-			parent = parent->bus->self;
-		}
-		if (pci_is_pcie(tmp)) /* this is a PCIe-to-PCI bridge */
-			iommu_detach_dev(iommu,
-				tmp->subordinate->number, 0);
-		else /* this is a legacy PCI bridge */
-			iommu_detach_dev(iommu, tmp->bus->number,
-					 tmp->devfn);
-	}
+	pci_for_each_dma_alias(to_pci_dev(dev), &iommu_detach_dev_cb, iommu);
 }
 
 static void domain_remove_one_dev_info(struct dmar_domain *domain,
@@ -4370,99 +4414,42 @@ static int intel_iommu_domain_has_cap(struct iommu_domain *domain,
 	return 0;
 }
 
-#define REQ_ACS_FLAGS	(PCI_ACS_SV | PCI_ACS_RR | PCI_ACS_CR | PCI_ACS_UF)
-
 static int intel_iommu_add_device(struct device *dev)
 {
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct pci_dev *bridge, *dma_pdev = NULL;
+	struct intel_iommu *iommu;
 	struct iommu_group *group;
-	int ret;
 	u8 bus, devfn;
 
-	if (!device_to_iommu(dev, &bus, &devfn))
+	iommu = device_to_iommu(dev, &bus, &devfn);
+	if (!iommu)
 		return -ENODEV;
 
-	bridge = pci_find_upstream_pcie_bridge(pdev);
-	if (bridge) {
-		if (pci_is_pcie(bridge))
-			dma_pdev = pci_get_domain_bus_and_slot(
-						pci_domain_nr(pdev->bus),
-						bridge->subordinate->number, 0);
-		if (!dma_pdev)
-			dma_pdev = pci_dev_get(bridge);
-	} else
-		dma_pdev = pci_dev_get(pdev);
-
-	/* Account for quirked devices */
-	swap_pci_ref(&dma_pdev, pci_get_dma_source(dma_pdev));
-
-	/*
-	 * If it's a multifunction device that does not support our
-	 * required ACS flags, add to the same group as lowest numbered
-	 * function that also does not suport the required ACS flags.
-	 */
-	if (dma_pdev->multifunction &&
-	    !pci_acs_enabled(dma_pdev, REQ_ACS_FLAGS)) {
-		u8 i, slot = PCI_SLOT(dma_pdev->devfn);
-
-		for (i = 0; i < 8; i++) {
-			struct pci_dev *tmp;
-
-			tmp = pci_get_slot(dma_pdev->bus, PCI_DEVFN(slot, i));
-			if (!tmp)
-				continue;
-
-			if (!pci_acs_enabled(tmp, REQ_ACS_FLAGS)) {
-				swap_pci_ref(&dma_pdev, tmp);
-				break;
-			}
-			pci_dev_put(tmp);
-		}
-	}
-
-	/*
-	 * Devices on the root bus go through the iommu.  If that's not us,
-	 * find the next upstream device and test ACS up to the root bus.
-	 * Finding the next device may require skipping virtual buses.
-	 */
-	while (!pci_is_root_bus(dma_pdev->bus)) {
-		struct pci_bus *bus = dma_pdev->bus;
-
-		while (!bus->self) {
-			if (!pci_is_root_bus(bus))
-				bus = bus->parent;
-			else
-				goto root_bus;
-		}
+	iommu_device_link(iommu->iommu_dev, dev);
 
-		if (pci_acs_path_enabled(bus->self, NULL, REQ_ACS_FLAGS))
-			break;
+	group = iommu_group_get_for_dev(dev);
 
-		swap_pci_ref(&dma_pdev, pci_dev_get(bus->self));
-	}
-
-root_bus:
-	group = iommu_group_get(&dma_pdev->dev);
-	pci_dev_put(dma_pdev);
-	if (!group) {
-		group = iommu_group_alloc();
-		if (IS_ERR(group))
-			return PTR_ERR(group);
-	}
-
-	ret = iommu_group_add_device(group, dev);
+	if (IS_ERR(group))
+		return PTR_ERR(group);
 
 	iommu_group_put(group);
-	return ret;
+	return 0;
 }
 
 static void intel_iommu_remove_device(struct device *dev)
 {
+	struct intel_iommu *iommu;
+	u8 bus, devfn;
+
+	iommu = device_to_iommu(dev, &bus, &devfn);
+	if (!iommu)
+		return;
+
 	iommu_group_remove_device(dev);
+
+	iommu_device_unlink(iommu->iommu_dev, dev);
 }
 
-static struct iommu_ops intel_iommu_ops = {
+static const struct iommu_ops intel_iommu_ops = {
 	.domain_init	= intel_iommu_domain_init,
 	.domain_destroy = intel_iommu_domain_destroy,
 	.attach_dev	= intel_iommu_attach_device,

drivers/iommu/intel_irq_remapping.c

@@ -369,29 +369,52 @@ static int set_hpet_sid(struct irte *irte, u8 id)
 	return 0;
 }
 
+struct set_msi_sid_data {
+	struct pci_dev *pdev;
+	u16 alias;
+};
+
+static int set_msi_sid_cb(struct pci_dev *pdev, u16 alias, void *opaque)
+{
+	struct set_msi_sid_data *data = opaque;
+
+	data->pdev = pdev;
+	data->alias = alias;
+
+	return 0;
+}
+
 static int set_msi_sid(struct irte *irte, struct pci_dev *dev)
 {
-	struct pci_dev *bridge;
+	struct set_msi_sid_data data;
 
 	if (!irte || !dev)
 		return -1;
 
-	/* PCIe device or Root Complex integrated PCI device */
-	if (pci_is_pcie(dev) || !dev->bus->parent) {
-		set_irte_sid(irte, SVT_VERIFY_SID_SQ, SQ_ALL_16,
-			     (dev->bus->number << 8) | dev->devfn);
-		return 0;
-	}
+	pci_for_each_dma_alias(dev, set_msi_sid_cb, &data);
 
-	bridge = pci_find_upstream_pcie_bridge(dev);
-	if (bridge) {
-		if (pci_is_pcie(bridge))/* this is a PCIe-to-PCI/PCIX bridge */
-			set_irte_sid(irte, SVT_VERIFY_BUS, SQ_ALL_16,
-				(bridge->bus->number << 8) | dev->bus->number);
-		else /* this is a legacy PCI bridge */
-			set_irte_sid(irte, SVT_VERIFY_SID_SQ, SQ_ALL_16,
-				(bridge->bus->number << 8) | bridge->devfn);
-	}
+	/*
+	 * DMA alias provides us with a PCI device and alias.  The only case
+	 * where the it will return an alias on a different bus than the
+	 * device is the case of a PCIe-to-PCI bridge, where the alias is for
+	 * the subordinate bus.  In this case we can only verify the bus.
+	 *
+	 * If the alias device is on a different bus than our source device
+	 * then we have a topology based alias, use it.
+	 *
+	 * Otherwise, the alias is for a device DMA quirk and we cannot
+	 * assume that MSI uses the same requester ID.  Therefore use the
+	 * original device.
+	 */
+	if (PCI_BUS_NUM(data.alias) != data.pdev->bus->number)
+		set_irte_sid(irte, SVT_VERIFY_BUS, SQ_ALL_16,
+			     PCI_DEVID(PCI_BUS_NUM(data.alias),
+				       dev->bus->number));
+	else if (data.pdev->bus->number != dev->bus->number)
+		set_irte_sid(irte, SVT_VERIFY_SID_SQ, SQ_ALL_16, data.alias);
+	else
+		set_irte_sid(irte, SVT_VERIFY_SID_SQ, SQ_ALL_16,
+			     PCI_DEVID(dev->bus->number, dev->devfn));
 
 	return 0;
 }

drivers/iommu/iommu-sysfs.c

@@ -0,0 +1,134 @@
+/*
+ * IOMMU sysfs class support
+ *
+ * Copyright (C) 2014 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.
+ */
+
+#include <linux/device.h>
+#include <linux/iommu.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+
+/*
+ * We provide a common class "devices" group which initially has no attributes.
+ * As devices are added to the IOMMU, we'll add links to the group.
+ */
+static struct attribute *devices_attr[] = {
+	NULL,
+};
+
+static const struct attribute_group iommu_devices_attr_group = {
+	.name = "devices",
+	.attrs = devices_attr,
+};
+
+static const struct attribute_group *iommu_dev_groups[] = {
+	&iommu_devices_attr_group,
+	NULL,
+};
+
+static void iommu_release_device(struct device *dev)
+{
+	kfree(dev);
+}
+
+static struct class iommu_class = {
+	.name = "iommu",
+	.dev_release = iommu_release_device,
+	.dev_groups = iommu_dev_groups,
+};
+
+static int __init iommu_dev_init(void)
+{
+	return class_register(&iommu_class);
+}
+postcore_initcall(iommu_dev_init);
+
+/*
+ * Create an IOMMU device and return a pointer to it.  IOMMU specific
+ * attributes can be provided as an attribute group, allowing a unique
+ * namespace per IOMMU type.
+ */
+struct device *iommu_device_create(struct device *parent, void *drvdata,
+				   const struct attribute_group **groups,
+				   const char *fmt, ...)
+{
+	struct device *dev;
+	va_list vargs;
+	int ret;
+
+	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
+	if (!dev)
+		return ERR_PTR(-ENOMEM);
+
+	device_initialize(dev);
+
+	dev->class = &iommu_class;
+	dev->parent = parent;
+	dev->groups = groups;
+	dev_set_drvdata(dev, drvdata);
+
+	va_start(vargs, fmt);
+	ret = kobject_set_name_vargs(&dev->kobj, fmt, vargs);
+	va_end(vargs);
+	if (ret)
+		goto error;
+
+	ret = device_add(dev);
+	if (ret)
+		goto error;
+
+	return dev;
+
+error:
+	put_device(dev);
+	return ERR_PTR(ret);
+}
+
+void iommu_device_destroy(struct device *dev)
+{
+	if (!dev || IS_ERR(dev))
+		return;
+
+	device_unregister(dev);
+}
+
+/*
+ * IOMMU drivers can indicate a device is managed by a given IOMMU using
+ * this interface.  A link to the device will be created in the "devices"
+ * directory of the IOMMU device in sysfs and an "iommu" link will be
+ * created under the linked device, pointing back at the IOMMU device.
+ */
+int iommu_device_link(struct device *dev, struct device *link)
+{
+	int ret;
+
+	if (!dev || IS_ERR(dev))
+		return -ENODEV;
+
+	ret = sysfs_add_link_to_group(&dev->kobj, "devices",
+				      &link->kobj, dev_name(link));
+	if (ret)
+		return ret;
+
+	ret = sysfs_create_link_nowarn(&link->kobj, &dev->kobj, "iommu");
+	if (ret)
+		sysfs_remove_link_from_group(&dev->kobj, "devices",
+					     dev_name(link));
+
+	return ret;
+}
+
+void iommu_device_unlink(struct device *dev, struct device *link)
+{
+	if (!dev || IS_ERR(dev))
+		return;
+
+	sysfs_remove_link(&link->kobj, "iommu");
+	sysfs_remove_link_from_group(&dev->kobj, "devices", dev_name(link));
+}

drivers/iommu/iommu.c

@@ -29,12 +29,17 @@
 #include <linux/idr.h>
 #include <linux/notifier.h>
 #include <linux/err.h>
+#include <linux/pci.h>
 #include <trace/events/iommu.h>
 
 static struct kset *iommu_group_kset;
 static struct ida iommu_group_ida;
 static struct mutex iommu_group_mutex;
 
+struct iommu_callback_data {
+	const struct iommu_ops *ops;
+};
+
 struct iommu_group {
 	struct kobject kobj;
 	struct kobject *devices_kobj;
@@ -514,9 +519,191 @@ int iommu_group_id(struct iommu_group *group)
 }
 EXPORT_SYMBOL_GPL(iommu_group_id);
 
+/*
+ * To consider a PCI device isolated, we require ACS to support Source
+ * Validation, Request Redirection, Completer Redirection, and Upstream
+ * Forwarding.  This effectively means that devices cannot spoof their
+ * requester ID, requests and completions cannot be redirected, and all
+ * transactions are forwarded upstream, even as it passes through a
+ * bridge where the target device is downstream.
+ */
+#define REQ_ACS_FLAGS   (PCI_ACS_SV | PCI_ACS_RR | PCI_ACS_CR | PCI_ACS_UF)
+
+struct group_for_pci_data {
+	struct pci_dev *pdev;
+	struct iommu_group *group;
+};
+
+/*
+ * DMA alias iterator callback, return the last seen device.  Stop and return
+ * the IOMMU group if we find one along the way.
+ */
+static int get_pci_alias_or_group(struct pci_dev *pdev, u16 alias, void *opaque)
+{
+	struct group_for_pci_data *data = opaque;
+
+	data->pdev = pdev;
+	data->group = iommu_group_get(&pdev->dev);
+
+	return data->group != NULL;
+}
+
+/*
+ * Use standard PCI bus topology, isolation features, and DMA alias quirks
+ * to find or create an IOMMU group for a device.
+ */
+static struct iommu_group *iommu_group_get_for_pci_dev(struct pci_dev *pdev)
+{
+	struct group_for_pci_data data;
+	struct pci_bus *bus;
+	struct iommu_group *group = NULL;
+	struct pci_dev *tmp;
+
+	/*
+	 * Find the upstream DMA alias for the device.  A device must not
+	 * be aliased due to topology in order to have its own IOMMU group.
+	 * If we find an alias along the way that already belongs to a
+	 * group, use it.
+	 */
+	if (pci_for_each_dma_alias(pdev, get_pci_alias_or_group, &data))
+		return data.group;
+
+	pdev = data.pdev;
+
+	/*
+	 * Continue upstream from the point of minimum IOMMU granularity
+	 * due to aliases to the point where devices are protected from
+	 * peer-to-peer DMA by PCI ACS.  Again, if we find an existing
+	 * group, use it.
+	 */
+	for (bus = pdev->bus; !pci_is_root_bus(bus); bus = bus->parent) {
+		if (!bus->self)
+			continue;
+
+		if (pci_acs_path_enabled(bus->self, NULL, REQ_ACS_FLAGS))
+			break;
+
+		pdev = bus->self;
+
+		group = iommu_group_get(&pdev->dev);
+		if (group)
+			return group;
+	}
+
+	/*
+	 * Next we need to consider DMA alias quirks.  If one device aliases
+	 * to another, they should be grouped together.  It's theoretically
+	 * possible that aliases could create chains of devices where each
+	 * device aliases another device.  If we then factor in multifunction
+	 * ACS grouping requirements, each alias could incorporate a new slot
+	 * with multiple functions, each with aliases.  This is all extremely
+	 * unlikely as DMA alias quirks are typically only used for PCIe
+	 * devices where we usually have a single slot per bus.  Furthermore,
+	 * the alias quirk is usually to another function within the slot
+	 * (and ACS multifunction is not supported) or to a different slot
+	 * that doesn't physically exist.  The likely scenario is therefore
+	 * that everything on the bus gets grouped together.  To reduce the
+	 * problem space, share the IOMMU group for all devices on the bus
+	 * if a DMA alias quirk is present on the bus.
+	 */
+	tmp = NULL;
+	for_each_pci_dev(tmp) {
+		if (tmp->bus != pdev->bus ||
+		    !(tmp->dev_flags & PCI_DEV_FLAGS_DMA_ALIAS_DEVFN))
+			continue;
+
+		pci_dev_put(tmp);
+		tmp = NULL;
+
+		/* We have an alias quirk, search for an existing group */
+		for_each_pci_dev(tmp) {
+			struct iommu_group *group_tmp;
+
+			if (tmp->bus != pdev->bus)
+				continue;
+
+			group_tmp = iommu_group_get(&tmp->dev);
+			if (!group) {
+				group = group_tmp;
+				continue;
+			}
+
+			if (group_tmp) {
+				WARN_ON(group != group_tmp);
+				iommu_group_put(group_tmp);
+			}
+		}
+
+		return group ? group : iommu_group_alloc();
+	}
+
+	/*
+	 * Non-multifunction devices or multifunction devices supporting
+	 * ACS get their own group.
+	 */
+	if (!pdev->multifunction || pci_acs_enabled(pdev, REQ_ACS_FLAGS))
+		return iommu_group_alloc();
+
+	/*
+	 * Multifunction devices not supporting ACS share a group with other
+	 * similar devices in the same slot.
+	 */
+	tmp = NULL;
+	for_each_pci_dev(tmp) {
+		if (tmp == pdev || tmp->bus != pdev->bus ||
+		    PCI_SLOT(tmp->devfn) !=  PCI_SLOT(pdev->devfn) ||
+		    pci_acs_enabled(tmp, REQ_ACS_FLAGS))
+			continue;
+
+		group = iommu_group_get(&tmp->dev);
+		if (group) {
+			pci_dev_put(tmp);
+			return group;
+		}
+	}
+
+	/* No shared group found, allocate new */
+	return iommu_group_alloc();
+}
+
+/**
+ * iommu_group_get_for_dev - Find or create the IOMMU group for a device
+ * @dev: target device
+ *
+ * This function is intended to be called by IOMMU drivers and extended to
+ * support common, bus-defined algorithms when determining or creating the
+ * IOMMU group for a device.  On success, the caller will hold a reference
+ * to the returned IOMMU group, which will already include the provided
+ * device.  The reference should be released with iommu_group_put().
+ */
+struct iommu_group *iommu_group_get_for_dev(struct device *dev)
+{
+	struct iommu_group *group = ERR_PTR(-EIO);
+	int ret;
+
+	group = iommu_group_get(dev);
+	if (group)
+		return group;
+
+	if (dev_is_pci(dev))
+		group = iommu_group_get_for_pci_dev(to_pci_dev(dev));
+
+	if (IS_ERR(group))
+		return group;
+
+	ret = iommu_group_add_device(group, dev);
+	if (ret) {
+		iommu_group_put(group);
+		return ERR_PTR(ret);
+	}
+
+	return group;
+}
+
 static int add_iommu_group(struct device *dev, void *data)
 {
-	struct iommu_ops *ops = data;
+	struct iommu_callback_data *cb = data;
+	const struct iommu_ops *ops = cb->ops;
 
 	if (!ops->add_device)
 		return -ENODEV;
@@ -532,7 +719,7 @@ static int iommu_bus_notifier(struct notifier_block *nb,
 			      unsigned long action, void *data)
 {
 	struct device *dev = data;
-	struct iommu_ops *ops = dev->bus->iommu_ops;
+	const struct iommu_ops *ops = dev->bus->iommu_ops;
 	struct iommu_group *group;
 	unsigned long group_action = 0;
 
@@ -585,10 +772,14 @@ static struct notifier_block iommu_bus_nb = {
 	.notifier_call = iommu_bus_notifier,
 };
 
-static void iommu_bus_init(struct bus_type *bus, struct iommu_ops *ops)
+static void iommu_bus_init(struct bus_type *bus, const struct iommu_ops *ops)
 {
+	struct iommu_callback_data cb = {
+		.ops = ops,
+	};
+
 	bus_register_notifier(bus, &iommu_bus_nb);
-	bus_for_each_dev(bus, NULL, ops, add_iommu_group);
+	bus_for_each_dev(bus, NULL, &cb, add_iommu_group);
 }
 
 /**
@@ -604,7 +795,7 @@ static void iommu_bus_init(struct bus_type *bus, struct iommu_ops *ops)
  * is set up. With this function the iommu-driver can set the iommu-ops
  * afterwards.
  */
-int bus_set_iommu(struct bus_type *bus, struct iommu_ops *ops)
+int bus_set_iommu(struct bus_type *bus, const struct iommu_ops *ops)
 {
 	if (bus->iommu_ops != NULL)
 		return -EBUSY;

drivers/iommu/ipmmu-vmsa.c

@@ -1120,7 +1120,7 @@ static void ipmmu_remove_device(struct device *dev)
 	dev->archdata.iommu = NULL;
 }
 
-static struct iommu_ops ipmmu_ops = {
+static const struct iommu_ops ipmmu_ops = {
 	.domain_init = ipmmu_domain_init,
 	.domain_destroy = ipmmu_domain_destroy,
 	.attach_dev = ipmmu_attach_device,

drivers/iommu/msm_iommu.c

@@ -674,7 +674,7 @@ fail:
 	return 0;
 }
 
-static struct iommu_ops msm_iommu_ops = {
+static const struct iommu_ops msm_iommu_ops = {
 	.domain_init = msm_iommu_domain_init,
 	.domain_destroy = msm_iommu_domain_destroy,
 	.attach_dev = msm_iommu_attach_dev,

drivers/iommu/omap-iommu.c

@@ -1291,7 +1291,7 @@ static void omap_iommu_remove_device(struct device *dev)
 	kfree(arch_data);
 }
 
-static struct iommu_ops omap_iommu_ops = {
+static const struct iommu_ops omap_iommu_ops = {
 	.domain_init	= omap_iommu_domain_init,
 	.domain_destroy	= omap_iommu_domain_destroy,
 	.attach_dev	= omap_iommu_attach_dev,

drivers/iommu/pci.h

@@ -1,29 +0,0 @@
-/*
- * 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, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
- *
- * Copyright (C) 2013 Red Hat, Inc.
- * Copyright (C) 2013 Freescale Semiconductor, Inc.
- *
- */
-#ifndef __IOMMU_PCI_H
-#define __IOMMU_PCI_H
-
-/* Helper function for swapping pci device reference */
-static inline void swap_pci_ref(struct pci_dev **from, struct pci_dev *to)
-{
-	pci_dev_put(*from);
-	*from = to;
-}
-
-#endif  /* __IOMMU_PCI_H */

drivers/iommu/shmobile-iommu.c

@@ -354,7 +354,7 @@ static int shmobile_iommu_add_device(struct device *dev)
 	return 0;
 }
 
-static struct iommu_ops shmobile_iommu_ops = {
+static const struct iommu_ops shmobile_iommu_ops = {
 	.domain_init = shmobile_iommu_domain_init,
 	.domain_destroy = shmobile_iommu_domain_destroy,
 	.attach_dev = shmobile_iommu_attach_device,

drivers/iommu/tegra-gart.c

@@ -309,7 +309,7 @@ static int gart_iommu_domain_has_cap(struct iommu_domain *domain,
 	return 0;
 }
 
-static struct iommu_ops gart_iommu_ops = {
+static const struct iommu_ops gart_iommu_ops = {
 	.domain_init	= gart_iommu_domain_init,
 	.domain_destroy	= gart_iommu_domain_destroy,
 	.attach_dev	= gart_iommu_attach_dev,

drivers/iommu/tegra-smmu.c

@@ -947,7 +947,7 @@ static void smmu_iommu_domain_destroy(struct iommu_domain *domain)
 	dev_dbg(smmu->dev, "smmu_as@%p\n", as);
 }
 
-static struct iommu_ops smmu_iommu_ops = {
+static const struct iommu_ops smmu_iommu_ops = {
 	.domain_init	= smmu_iommu_domain_init,
 	.domain_destroy	= smmu_iommu_domain_destroy,
 	.attach_dev	= smmu_iommu_attach_dev,

include/linux/device.h

@@ -124,7 +124,7 @@ struct bus_type {
 
 	const struct dev_pm_ops *pm;
 
-	struct iommu_ops *iommu_ops;
+	const struct iommu_ops *iommu_ops;
 
 	struct subsys_private *p;
 	struct lock_class_key lock_key;

include/linux/intel-iommu.h

@@ -336,6 +336,7 @@ struct intel_iommu {
 #ifdef CONFIG_IRQ_REMAP
 	struct ir_table *ir_table;	/* Interrupt remapping info */
 #endif
+	struct device	*iommu_dev; /* IOMMU-sysfs device */
 	int		node;
 };
 
@@ -365,4 +366,6 @@ extern int qi_submit_sync(struct qi_desc *desc, struct intel_iommu *iommu);
 
 extern int dmar_ir_support(void);
 
+extern const struct attribute_group *intel_iommu_groups[];
+
 #endif

include/linux/iommu.h

@@ -50,7 +50,7 @@ struct iommu_domain_geometry {
 };
 
 struct iommu_domain {
-	struct iommu_ops *ops;
+	const struct iommu_ops *ops;
 	void *priv;
 	iommu_fault_handler_t handler;
 	void *handler_token;
@@ -140,7 +140,7 @@ struct iommu_ops {
 #define IOMMU_GROUP_NOTIFY_UNBIND_DRIVER	5 /* Pre Driver unbind */
 #define IOMMU_GROUP_NOTIFY_UNBOUND_DRIVER	6 /* Post Driver unbind */
 
-extern int bus_set_iommu(struct bus_type *bus, struct iommu_ops *ops);
+extern int bus_set_iommu(struct bus_type *bus, const struct iommu_ops *ops);
 extern bool iommu_present(struct bus_type *bus);
 extern struct iommu_domain *iommu_domain_alloc(struct bus_type *bus);
 extern struct iommu_group *iommu_group_get_by_id(int id);
@@ -181,11 +181,18 @@ extern int iommu_group_register_notifier(struct iommu_group *group,
 extern int iommu_group_unregister_notifier(struct iommu_group *group,
 					   struct notifier_block *nb);
 extern int iommu_group_id(struct iommu_group *group);
+extern struct iommu_group *iommu_group_get_for_dev(struct device *dev);
 
 extern int iommu_domain_get_attr(struct iommu_domain *domain, enum iommu_attr,
 				 void *data);
 extern int iommu_domain_set_attr(struct iommu_domain *domain, enum iommu_attr,
 				 void *data);
+struct device *iommu_device_create(struct device *parent, void *drvdata,
+				   const struct attribute_group **groups,
+				   const char *fmt, ...);
+void iommu_device_destroy(struct device *dev);
+int iommu_device_link(struct device *dev, struct device *link);
+void iommu_device_unlink(struct device *dev, struct device *link);
 
 /* Window handling function prototypes */
 extern int iommu_domain_window_enable(struct iommu_domain *domain, u32 wnd_nr,
@@ -396,6 +403,27 @@ static inline int iommu_domain_set_attr(struct iommu_domain *domain,
 	return -EINVAL;
 }
 
+static inline struct device *iommu_device_create(struct device *parent,
+					void *drvdata,
+					const struct attribute_group **groups,
+					const char *fmt, ...)
+{
+	return ERR_PTR(-ENODEV);
+}
+
+static inline void iommu_device_destroy(struct device *dev)
+{
+}
+
+static inline int iommu_device_link(struct device *dev, struct device *link)
+{
+	return -EINVAL;
+}
+
+static inline void iommu_device_unlink(struct device *dev, struct device *link)
+{
+}
+
 #endif /* CONFIG_IOMMU_API */
 
 #endif /* __LINUX_IOMMU_H */