diff options
Diffstat (limited to 'drivers/iommu/amd/iommu.c')
| -rw-r--r-- | drivers/iommu/amd/iommu.c | 4265 |
1 files changed, 2159 insertions, 2106 deletions
diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c index 74cca1757172..9f1d56a5e145 100644 --- a/drivers/iommu/amd/iommu.c +++ b/drivers/iommu/amd/iommu.c @@ -11,97 +11,79 @@ #include <linux/ratelimit.h> #include <linux/pci.h> #include <linux/acpi.h> -#include <linux/amba/bus.h> -#include <linux/platform_device.h> #include <linux/pci-ats.h> #include <linux/bitmap.h> #include <linux/slab.h> +#include <linux/string_choices.h> #include <linux/debugfs.h> #include <linux/scatterlist.h> -#include <linux/dma-mapping.h> +#include <linux/dma-map-ops.h> #include <linux/dma-direct.h> -#include <linux/dma-iommu.h> +#include <linux/idr.h> #include <linux/iommu-helper.h> #include <linux/delay.h> #include <linux/amd-iommu.h> #include <linux/notifier.h> #include <linux/export.h> #include <linux/irq.h> +#include <linux/irqchip/irq-msi-lib.h> #include <linux/msi.h> -#include <linux/dma-contiguous.h> #include <linux/irqdomain.h> #include <linux/percpu.h> -#include <linux/iova.h> +#include <linux/cc_platform.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 <uapi/linux/iommufd.h> +#include <linux/generic_pt/iommu.h> #include "amd_iommu.h" #include "../irq_remapping.h" +#include "../iommu-pages.h" #define CMD_SET_TYPE(cmd, t) ((cmd)->data[1] |= ((t) << 28)) -#define LOOP_TIMEOUT 100000 - -/* IO virtual address start page frame number */ -#define IOVA_START_PFN (1) -#define IOVA_PFN(addr) ((addr) >> PAGE_SHIFT) - /* Reserved IOVA ranges */ #define MSI_RANGE_START (0xfee00000) #define MSI_RANGE_END (0xfeefffff) #define HT_RANGE_START (0xfd00000000ULL) #define HT_RANGE_END (0xffffffffffULL) -/* - * 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)) - -#define DEFAULT_PGTABLE_LEVEL PAGE_MODE_3_LEVEL - -static DEFINE_SPINLOCK(pd_bitmap_lock); - -/* List of all available dev_data structures */ -static LLIST_HEAD(dev_data_list); - LIST_HEAD(ioapic_map); LIST_HEAD(hpet_map); LIST_HEAD(acpihid_map); -/* - * Domain for untranslated devices - only allocated - * if iommu=pt passed on kernel cmd line. - */ const struct iommu_ops amd_iommu_ops; -static ATOMIC_NOTIFIER_HEAD(ppr_notifier); int amd_iommu_max_glx_val = -1; /* - * general struct to manage commands send to an IOMMU + * AMD IOMMU allows up to 2^16 different protection domains. This is a bitmap + * to know which ones are already in use. */ -struct iommu_cmd { - u32 data[4]; -}; +DEFINE_IDA(pdom_ids); + +static int amd_iommu_attach_device(struct iommu_domain *dom, struct device *dev, + struct iommu_domain *old); -struct kmem_cache *amd_iommu_irq_cache; +static void set_dte_entry(struct amd_iommu *iommu, + struct iommu_dev_data *dev_data, + phys_addr_t top_paddr, unsigned int top_level); -static void update_domain(struct protection_domain *domain); -static void detach_device(struct device *dev); -static void update_and_flush_device_table(struct protection_domain *domain, - struct domain_pgtable *pgtable); +static void amd_iommu_change_top(struct pt_iommu *iommu_table, + phys_addr_t top_paddr, unsigned int top_level); + +static void iommu_flush_dte_sync(struct amd_iommu *iommu, u16 devid); + +static struct iommu_dev_data *find_dev_data(struct amd_iommu *iommu, u16 devid); +static bool amd_iommu_enforce_cache_coherency(struct iommu_domain *domain); +static int amd_iommu_set_dirty_tracking(struct iommu_domain *domain, + bool enable); /**************************************************************************** * @@ -109,18 +91,159 @@ static void update_and_flush_device_table(struct protection_domain *domain, * ****************************************************************************/ -static inline u16 get_pci_device_id(struct device *dev) +static __always_inline void amd_iommu_atomic128_set(__int128 *ptr, __int128 val) { - struct pci_dev *pdev = to_pci_dev(dev); + /* + * Note: + * We use arch_cmpxchg128_local() because: + * - Need cmpxchg16b instruction mainly for 128-bit store to DTE + * (not necessary for cmpxchg since this function is already + * protected by a spin_lock for this DTE). + * - Neither need LOCK_PREFIX nor try loop because of the spin_lock. + */ + arch_cmpxchg128_local(ptr, *ptr, val); +} + +static void write_dte_upper128(struct dev_table_entry *ptr, struct dev_table_entry *new) +{ + struct dev_table_entry old; + + old.data128[1] = ptr->data128[1]; + /* + * Preserve DTE_DATA2_INTR_MASK. This needs to be + * done here since it requires to be inside + * spin_lock(&dev_data->dte_lock) context. + */ + new->data[2] &= ~DTE_DATA2_INTR_MASK; + new->data[2] |= old.data[2] & DTE_DATA2_INTR_MASK; + + amd_iommu_atomic128_set(&ptr->data128[1], new->data128[1]); +} + +static void write_dte_lower128(struct dev_table_entry *ptr, struct dev_table_entry *new) +{ + amd_iommu_atomic128_set(&ptr->data128[0], new->data128[0]); +} + +/* + * Note: + * IOMMU reads the entire Device Table entry in a single 256-bit transaction + * but the driver is programming DTE using 2 128-bit cmpxchg. So, the driver + * need to ensure the following: + * - DTE[V|GV] bit is being written last when setting. + * - DTE[V|GV] bit is being written first when clearing. + * + * This function is used only by code, which updates DMA translation part of the DTE. + * So, only consider control bits related to DMA when updating the entry. + */ +static void update_dte256(struct amd_iommu *iommu, struct iommu_dev_data *dev_data, + struct dev_table_entry *new) +{ + unsigned long flags; + struct dev_table_entry *dev_table = get_dev_table(iommu); + struct dev_table_entry *ptr = &dev_table[dev_data->devid]; + + spin_lock_irqsave(&dev_data->dte_lock, flags); + + if (!(ptr->data[0] & DTE_FLAG_V)) { + /* Existing DTE is not valid. */ + write_dte_upper128(ptr, new); + write_dte_lower128(ptr, new); + iommu_flush_dte_sync(iommu, dev_data->devid); + } else if (!(new->data[0] & DTE_FLAG_V)) { + /* Existing DTE is valid. New DTE is not valid. */ + write_dte_lower128(ptr, new); + write_dte_upper128(ptr, new); + iommu_flush_dte_sync(iommu, dev_data->devid); + } else if (!FIELD_GET(DTE_FLAG_GV, ptr->data[0])) { + /* + * Both DTEs are valid. + * Existing DTE has no guest page table. + */ + write_dte_upper128(ptr, new); + write_dte_lower128(ptr, new); + iommu_flush_dte_sync(iommu, dev_data->devid); + } else if (!FIELD_GET(DTE_FLAG_GV, new->data[0])) { + /* + * Both DTEs are valid. + * Existing DTE has guest page table, + * new DTE has no guest page table, + */ + write_dte_lower128(ptr, new); + write_dte_upper128(ptr, new); + iommu_flush_dte_sync(iommu, dev_data->devid); + } else if (FIELD_GET(DTE_GPT_LEVEL_MASK, ptr->data[2]) != + FIELD_GET(DTE_GPT_LEVEL_MASK, new->data[2])) { + /* + * Both DTEs are valid and have guest page table, + * but have different number of levels. So, we need + * to upadte both upper and lower 128-bit value, which + * require disabling and flushing. + */ + struct dev_table_entry clear = {}; + + /* First disable DTE */ + write_dte_lower128(ptr, &clear); + iommu_flush_dte_sync(iommu, dev_data->devid); + + /* Then update DTE */ + write_dte_upper128(ptr, new); + write_dte_lower128(ptr, new); + iommu_flush_dte_sync(iommu, dev_data->devid); + } else { + /* + * Both DTEs are valid and have guest page table, + * and same number of levels. We just need to only + * update the lower 128-bit. So no need to disable DTE. + */ + write_dte_lower128(ptr, new); + } + + spin_unlock_irqrestore(&dev_data->dte_lock, flags); +} + +static void get_dte256(struct amd_iommu *iommu, struct iommu_dev_data *dev_data, + struct dev_table_entry *dte) +{ + unsigned long flags; + struct dev_table_entry *ptr; + struct dev_table_entry *dev_table = get_dev_table(iommu); - return pci_dev_id(pdev); + ptr = &dev_table[dev_data->devid]; + + spin_lock_irqsave(&dev_data->dte_lock, flags); + dte->data128[0] = ptr->data128[0]; + dte->data128[1] = ptr->data128[1]; + spin_unlock_irqrestore(&dev_data->dte_lock, flags); +} + +static inline bool pdom_is_v2_pgtbl_mode(struct protection_domain *pdom) +{ + return (pdom && (pdom->pd_mode == PD_MODE_V2)); +} + +static inline bool pdom_is_in_pt_mode(struct protection_domain *pdom) +{ + return (pdom->domain.type == IOMMU_DOMAIN_IDENTITY); +} + +/* + * We cannot support PASID w/ existing v1 page table in the same domain + * since it will be nested. However, existing domain w/ v2 page table + * or passthrough mode can be used for PASID. + */ +static inline bool pdom_is_sva_capable(struct protection_domain *pdom) +{ + return pdom_is_v2_pgtbl_mode(pdom) || pdom_is_in_pt_mode(pdom); } static inline int get_acpihid_device_id(struct device *dev, struct acpihid_map_entry **entry) { struct acpi_device *adev = ACPI_COMPANION(dev); - struct acpihid_map_entry *p; + struct acpihid_map_entry *p, *p1 = NULL; + int hid_count = 0; + bool fw_bug; if (!adev) return -ENODEV; @@ -128,76 +251,133 @@ static inline int get_acpihid_device_id(struct device *dev, list_for_each_entry(p, &acpihid_map, list) { if (acpi_dev_hid_uid_match(adev, p->hid, p->uid[0] ? p->uid : NULL)) { - if (entry) - *entry = p; - return p->devid; + p1 = p; + fw_bug = false; + hid_count = 1; + break; + } + + /* + * Count HID matches w/o UID, raise FW_BUG but allow exactly one match + */ + if (acpi_dev_hid_match(adev, p->hid)) { + p1 = p; + hid_count++; + fw_bug = true; } } - return -EINVAL; + + if (!p1) + return -EINVAL; + if (fw_bug) + dev_err_once(dev, FW_BUG "No ACPI device matched UID, but %d device%s matched HID.\n", + hid_count, str_plural(hid_count)); + if (hid_count > 1) + return -EINVAL; + if (entry) + *entry = p1; + + return p1->devid; } -static inline int get_device_id(struct device *dev) +static inline int get_device_sbdf_id(struct device *dev) { - int devid; + int sbdf; if (dev_is_pci(dev)) - devid = get_pci_device_id(dev); + sbdf = get_pci_sbdf_id(to_pci_dev(dev)); else - devid = get_acpihid_device_id(dev, NULL); + sbdf = get_acpihid_device_id(dev, NULL); + + return sbdf; +} + +struct dev_table_entry *get_dev_table(struct amd_iommu *iommu) +{ + struct dev_table_entry *dev_table; + struct amd_iommu_pci_seg *pci_seg = iommu->pci_seg; + + BUG_ON(pci_seg == NULL); + dev_table = pci_seg->dev_table; + BUG_ON(dev_table == NULL); - return devid; + return dev_table; } -static struct protection_domain *to_pdomain(struct iommu_domain *dom) +static inline u16 get_device_segment(struct device *dev) { - return container_of(dom, struct protection_domain, domain); + u16 seg; + + if (dev_is_pci(dev)) { + struct pci_dev *pdev = to_pci_dev(dev); + + seg = pci_domain_nr(pdev->bus); + } else { + u32 devid = get_acpihid_device_id(dev, NULL); + + seg = PCI_SBDF_TO_SEGID(devid); + } + + return seg; } -static void amd_iommu_domain_get_pgtable(struct protection_domain *domain, - struct domain_pgtable *pgtable) +/* Writes the specific IOMMU for a device into the PCI segment rlookup table */ +void amd_iommu_set_rlookup_table(struct amd_iommu *iommu, u16 devid) { - u64 pt_root = atomic64_read(&domain->pt_root); + struct amd_iommu_pci_seg *pci_seg = iommu->pci_seg; - pgtable->root = (u64 *)(pt_root & PAGE_MASK); - pgtable->mode = pt_root & 7; /* lowest 3 bits encode pgtable mode */ + pci_seg->rlookup_table[devid] = iommu; } -static u64 amd_iommu_domain_encode_pgtable(u64 *root, int mode) +static struct amd_iommu *__rlookup_amd_iommu(u16 seg, u16 devid) { - u64 pt_root; + struct amd_iommu_pci_seg *pci_seg; - /* lowest 3 bits encode pgtable mode */ - pt_root = mode & 7; - pt_root |= (u64)root; + for_each_pci_segment(pci_seg) { + if (pci_seg->id == seg) + return pci_seg->rlookup_table[devid]; + } + return NULL; +} - return pt_root; +static struct amd_iommu *rlookup_amd_iommu(struct device *dev) +{ + u16 seg = get_device_segment(dev); + int devid = get_device_sbdf_id(dev); + + if (devid < 0) + return NULL; + return __rlookup_amd_iommu(seg, PCI_SBDF_TO_DEVID(devid)); } -static struct iommu_dev_data *alloc_dev_data(u16 devid) +static struct iommu_dev_data *alloc_dev_data(struct amd_iommu *iommu, u16 devid) { struct iommu_dev_data *dev_data; + struct amd_iommu_pci_seg *pci_seg = iommu->pci_seg; dev_data = kzalloc(sizeof(*dev_data), GFP_KERNEL); if (!dev_data) return NULL; - spin_lock_init(&dev_data->lock); + mutex_init(&dev_data->mutex); + spin_lock_init(&dev_data->dte_lock); dev_data->devid = devid; ratelimit_default_init(&dev_data->rs); - llist_add(&dev_data->dev_data_list, &dev_data_list); + llist_add(&dev_data->dev_data_list, &pci_seg->dev_data_list); return dev_data; } -static struct iommu_dev_data *search_dev_data(u16 devid) +struct iommu_dev_data *search_dev_data(struct amd_iommu *iommu, u16 devid) { struct iommu_dev_data *dev_data; struct llist_node *node; + struct amd_iommu_pci_seg *pci_seg = iommu->pci_seg; - if (llist_empty(&dev_data_list)) + if (llist_empty(&pci_seg->dev_data_list)) return NULL; - node = dev_data_list.first; + node = pci_seg->dev_data_list.first; llist_for_each_entry(dev_data, node, dev_data_list) { if (dev_data->devid == devid) return dev_data; @@ -208,67 +388,90 @@ static struct iommu_dev_data *search_dev_data(u16 devid) static int clone_alias(struct pci_dev *pdev, u16 alias, void *data) { + struct dev_table_entry new; + struct amd_iommu *iommu; + struct iommu_dev_data *dev_data, *alias_data; u16 devid = pci_dev_id(pdev); + int ret = 0; if (devid == alias) return 0; - amd_iommu_rlookup_table[alias] = - amd_iommu_rlookup_table[devid]; - memcpy(amd_iommu_dev_table[alias].data, - amd_iommu_dev_table[devid].data, - sizeof(amd_iommu_dev_table[alias].data)); + iommu = rlookup_amd_iommu(&pdev->dev); + if (!iommu) + return 0; - return 0; + /* Copy the data from pdev */ + dev_data = dev_iommu_priv_get(&pdev->dev); + if (!dev_data) { + pr_err("%s : Failed to get dev_data for 0x%x\n", __func__, devid); + ret = -EINVAL; + goto out; + } + get_dte256(iommu, dev_data, &new); + + /* Setup alias */ + alias_data = find_dev_data(iommu, alias); + if (!alias_data) { + pr_err("%s : Failed to get alias dev_data for 0x%x\n", __func__, alias); + ret = -EINVAL; + goto out; + } + update_dte256(iommu, alias_data, &new); + + amd_iommu_set_rlookup_table(iommu, alias); +out: + return ret; } -static void clone_aliases(struct pci_dev *pdev) +static void clone_aliases(struct amd_iommu *iommu, struct device *dev) { - if (!pdev) + struct pci_dev *pdev; + + if (!dev_is_pci(dev)) return; + pdev = to_pci_dev(dev); /* * The IVRS alias stored in the alias table may not be * part of the PCI DMA aliases if it's bus differs * from the original device. */ - clone_alias(pdev, amd_iommu_alias_table[pci_dev_id(pdev)], NULL); + clone_alias(pdev, iommu->pci_seg->alias_table[pci_dev_id(pdev)], NULL); pci_for_each_dma_alias(pdev, clone_alias, NULL); } -static struct pci_dev *setup_aliases(struct device *dev) +static void setup_aliases(struct amd_iommu *iommu, struct device *dev) { struct pci_dev *pdev = to_pci_dev(dev); + struct amd_iommu_pci_seg *pci_seg = iommu->pci_seg; u16 ivrs_alias; /* For ACPI HID devices, there are no aliases */ if (!dev_is_pci(dev)) - return NULL; + return; /* * Add the IVRS alias to the pci aliases if it is on the same * bus. The IVRS table may know about a quirk that we don't. */ - ivrs_alias = amd_iommu_alias_table[pci_dev_id(pdev)]; + ivrs_alias = pci_seg->alias_table[pci_dev_id(pdev)]; if (ivrs_alias != pci_dev_id(pdev) && PCI_BUS_NUM(ivrs_alias) == pdev->bus->number) pci_add_dma_alias(pdev, ivrs_alias & 0xff, 1); - clone_aliases(pdev); - - return pdev; + clone_aliases(iommu, dev); } -static struct iommu_dev_data *find_dev_data(u16 devid) +static struct iommu_dev_data *find_dev_data(struct amd_iommu *iommu, u16 devid) { struct iommu_dev_data *dev_data; - struct amd_iommu *iommu = amd_iommu_rlookup_table[devid]; - dev_data = search_dev_data(devid); + dev_data = search_dev_data(iommu, devid); if (dev_data == NULL) { - dev_data = alloc_dev_data(devid); + dev_data = alloc_dev_data(iommu, devid); if (!dev_data) return NULL; @@ -304,33 +507,143 @@ static struct iommu_group *acpihid_device_group(struct device *dev) return entry->group; } -static bool pci_iommuv2_capable(struct pci_dev *pdev) +static inline bool pdev_pasid_supported(struct iommu_dev_data *dev_data) { - static const int caps[] = { - PCI_EXT_CAP_ID_PRI, - PCI_EXT_CAP_ID_PASID, - }; - int i, pos; + return (dev_data->flags & AMD_IOMMU_DEVICE_FLAG_PASID_SUP); +} - if (!pci_ats_supported(pdev)) - return false; +static u32 pdev_get_caps(struct pci_dev *pdev) +{ + int features; + u32 flags = 0; + + if (pci_ats_supported(pdev)) + flags |= AMD_IOMMU_DEVICE_FLAG_ATS_SUP; + + if (pci_pri_supported(pdev)) + flags |= AMD_IOMMU_DEVICE_FLAG_PRI_SUP; + + features = pci_pasid_features(pdev); + if (features >= 0) { + flags |= AMD_IOMMU_DEVICE_FLAG_PASID_SUP; + + if (features & PCI_PASID_CAP_EXEC) + flags |= AMD_IOMMU_DEVICE_FLAG_EXEC_SUP; + + if (features & PCI_PASID_CAP_PRIV) + flags |= AMD_IOMMU_DEVICE_FLAG_PRIV_SUP; + } + + return flags; +} - for (i = 0; i < 2; ++i) { - pos = pci_find_ext_capability(pdev, caps[i]); - if (pos == 0) - return false; +static inline int pdev_enable_cap_ats(struct pci_dev *pdev) +{ + struct iommu_dev_data *dev_data = dev_iommu_priv_get(&pdev->dev); + int ret = -EINVAL; + + if (dev_data->ats_enabled) + return 0; + + if (amd_iommu_iotlb_sup && + (dev_data->flags & AMD_IOMMU_DEVICE_FLAG_ATS_SUP)) { + ret = pci_enable_ats(pdev, PAGE_SHIFT); + if (!ret) { + dev_data->ats_enabled = 1; + dev_data->ats_qdep = pci_ats_queue_depth(pdev); + } } - return true; + return ret; } -static bool pdev_pri_erratum(struct pci_dev *pdev, u32 erratum) +static inline void pdev_disable_cap_ats(struct pci_dev *pdev) { - struct iommu_dev_data *dev_data; + struct iommu_dev_data *dev_data = dev_iommu_priv_get(&pdev->dev); - dev_data = dev_iommu_priv_get(&pdev->dev); + if (dev_data->ats_enabled) { + pci_disable_ats(pdev); + dev_data->ats_enabled = 0; + } +} + +static inline int pdev_enable_cap_pri(struct pci_dev *pdev) +{ + struct iommu_dev_data *dev_data = dev_iommu_priv_get(&pdev->dev); + int ret = -EINVAL; + + if (dev_data->pri_enabled) + return 0; + + if (!dev_data->ats_enabled) + return 0; + + if (dev_data->flags & AMD_IOMMU_DEVICE_FLAG_PRI_SUP) { + /* + * First reset the PRI state of the device. + * FIXME: Hardcode number of outstanding requests for now + */ + if (!pci_reset_pri(pdev) && !pci_enable_pri(pdev, 32)) { + dev_data->pri_enabled = 1; + dev_data->pri_tlp = pci_prg_resp_pasid_required(pdev); - return dev_data->errata & (1 << erratum) ? true : false; + ret = 0; + } + } + + return ret; +} + +static inline void pdev_disable_cap_pri(struct pci_dev *pdev) +{ + struct iommu_dev_data *dev_data = dev_iommu_priv_get(&pdev->dev); + + if (dev_data->pri_enabled) { + pci_disable_pri(pdev); + dev_data->pri_enabled = 0; + } +} + +static inline int pdev_enable_cap_pasid(struct pci_dev *pdev) +{ + struct iommu_dev_data *dev_data = dev_iommu_priv_get(&pdev->dev); + int ret = -EINVAL; + + if (dev_data->pasid_enabled) + return 0; + + if (dev_data->flags & AMD_IOMMU_DEVICE_FLAG_PASID_SUP) { + /* Only allow access to user-accessible pages */ + ret = pci_enable_pasid(pdev, 0); + if (!ret) + dev_data->pasid_enabled = 1; + } + + return ret; +} + +static inline void pdev_disable_cap_pasid(struct pci_dev *pdev) +{ + struct iommu_dev_data *dev_data = dev_iommu_priv_get(&pdev->dev); + + if (dev_data->pasid_enabled) { + pci_disable_pasid(pdev); + dev_data->pasid_enabled = 0; + } +} + +static void pdev_enable_caps(struct pci_dev *pdev) +{ + pdev_enable_cap_pasid(pdev); + pdev_enable_cap_ats(pdev); + pdev_enable_cap_pri(pdev); +} + +static void pdev_disable_caps(struct pci_dev *pdev) +{ + pdev_disable_cap_ats(pdev); + pdev_disable_cap_pasid(pdev); + pdev_disable_cap_pri(pdev); } /* @@ -339,42 +652,55 @@ static bool pdev_pri_erratum(struct pci_dev *pdev, u32 erratum) */ static bool check_device(struct device *dev) { - int devid; + struct amd_iommu_pci_seg *pci_seg; + struct amd_iommu *iommu; + int devid, sbdf; if (!dev) return false; - devid = get_device_id(dev); - if (devid < 0) + sbdf = get_device_sbdf_id(dev); + if (sbdf < 0) return false; + devid = PCI_SBDF_TO_DEVID(sbdf); - /* Out of our scope? */ - if (devid > amd_iommu_last_bdf) + iommu = rlookup_amd_iommu(dev); + if (!iommu) return false; - if (amd_iommu_rlookup_table[devid] == NULL) + /* Out of our scope? */ + pci_seg = iommu->pci_seg; + if (devid > pci_seg->last_bdf) return false; return true; } -static int iommu_init_device(struct device *dev) +static int iommu_init_device(struct amd_iommu *iommu, struct device *dev) { struct iommu_dev_data *dev_data; - int devid; + int devid, sbdf; if (dev_iommu_priv_get(dev)) return 0; - devid = get_device_id(dev); - if (devid < 0) - return devid; + sbdf = get_device_sbdf_id(dev); + if (sbdf < 0) + return sbdf; - dev_data = find_dev_data(devid); + devid = PCI_SBDF_TO_DEVID(sbdf); + dev_data = find_dev_data(iommu, devid); if (!dev_data) return -ENOMEM; - dev_data->pdev = setup_aliases(dev); + dev_data->dev = dev; + + /* + * The dev_iommu_priv_set() needes to be called before setup_aliases. + * Otherwise, subsequent call to dev_iommu_priv_get() will fail. + */ + dev_iommu_priv_set(dev, dev_data); + setup_aliases(iommu, dev); /* * By default we use passthrough mode for IOMMUv2 capable device. @@ -383,73 +709,30 @@ static int iommu_init_device(struct device *dev) * it'll be forced to go into translation mode. */ if ((iommu_default_passthrough() || !amd_iommu_force_isolation) && - dev_is_pci(dev) && pci_iommuv2_capable(to_pci_dev(dev))) { - struct amd_iommu *iommu; - - iommu = amd_iommu_rlookup_table[dev_data->devid]; - dev_data->iommu_v2 = iommu->is_iommu_v2; + dev_is_pci(dev) && amd_iommu_gt_ppr_supported()) { + dev_data->flags = pdev_get_caps(to_pci_dev(dev)); } - dev_iommu_priv_set(dev, dev_data); - return 0; } -static void iommu_ignore_device(struct device *dev) +static void iommu_ignore_device(struct amd_iommu *iommu, struct device *dev) { - int devid; + struct amd_iommu_pci_seg *pci_seg = iommu->pci_seg; + struct dev_table_entry *dev_table = get_dev_table(iommu); + int devid, sbdf; - devid = get_device_id(dev); - if (devid < 0) + sbdf = get_device_sbdf_id(dev); + if (sbdf < 0) return; - amd_iommu_rlookup_table[devid] = NULL; - memset(&amd_iommu_dev_table[devid], 0, sizeof(struct dev_table_entry)); + devid = PCI_SBDF_TO_DEVID(sbdf); + pci_seg->rlookup_table[devid] = NULL; + memset(&dev_table[devid], 0, sizeof(struct dev_table_entry)); - setup_aliases(dev); + setup_aliases(iommu, dev); } -static void amd_iommu_uninit_device(struct device *dev) -{ - struct iommu_dev_data *dev_data; - - dev_data = dev_iommu_priv_get(dev); - if (!dev_data) - return; - - if (dev_data->domain) - detach_device(dev); - - dev_iommu_priv_set(dev, 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. - */ -} - -/* - * Helper function to get the first pte of a large mapping - */ -static u64 *first_pte_l7(u64 *pte, unsigned long *page_size, - unsigned long *count) -{ - unsigned long pte_mask, pg_size, cnt; - u64 *fpte; - - pg_size = PTE_PAGE_SIZE(*pte); - cnt = PAGE_SIZE_PTE_COUNT(pg_size); - pte_mask = ~((cnt << 3) - 1); - fpte = (u64 *)(((unsigned long)pte) & pte_mask); - - if (page_size) - *page_size = pg_size; - - if (count) - *count = cnt; - - return fpte; -} /**************************************************************************** * @@ -457,13 +740,16 @@ static u64 *first_pte_l7(u64 *pte, unsigned long *page_size, * ****************************************************************************/ -static void dump_dte_entry(u16 devid) +static void dump_dte_entry(struct amd_iommu *iommu, u16 devid) { int i; + struct dev_table_entry dte; + struct iommu_dev_data *dev_data = find_dev_data(iommu, devid); + + get_dte256(iommu, dev_data, &dte); for (i = 0; i < 4; ++i) - pr_err("DTE[%d]: %016llx\n", i, - amd_iommu_dev_table[devid].data[i]); + pr_err("DTE[%d]: %016llx\n", i, dte.data[i]); } static void dump_command(unsigned long phys_addr) @@ -475,26 +761,124 @@ static void dump_command(unsigned long phys_addr) pr_err("CMD[%d]: %08x\n", i, cmd->data[i]); } -static void amd_iommu_report_page_fault(u16 devid, u16 domain_id, +static void amd_iommu_report_rmp_hw_error(struct amd_iommu *iommu, volatile u32 *event) +{ + struct iommu_dev_data *dev_data = NULL; + int devid, vmg_tag, flags; + struct pci_dev *pdev; + u64 spa; + + devid = (event[0] >> EVENT_DEVID_SHIFT) & EVENT_DEVID_MASK; + vmg_tag = (event[1]) & 0xFFFF; + flags = (event[1] >> EVENT_FLAGS_SHIFT) & EVENT_FLAGS_MASK; + spa = ((u64)event[3] << 32) | (event[2] & 0xFFFFFFF8); + + pdev = pci_get_domain_bus_and_slot(iommu->pci_seg->id, PCI_BUS_NUM(devid), + devid & 0xff); + if (pdev) + dev_data = dev_iommu_priv_get(&pdev->dev); + + if (dev_data) { + if (__ratelimit(&dev_data->rs)) { + pci_err(pdev, "Event logged [RMP_HW_ERROR vmg_tag=0x%04x, spa=0x%llx, flags=0x%04x]\n", + vmg_tag, spa, flags); + } + } else { + pr_err_ratelimited("Event logged [RMP_HW_ERROR device=%04x:%02x:%02x.%x, vmg_tag=0x%04x, spa=0x%llx, flags=0x%04x]\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + vmg_tag, spa, flags); + } + + if (pdev) + pci_dev_put(pdev); +} + +static void amd_iommu_report_rmp_fault(struct amd_iommu *iommu, volatile u32 *event) +{ + struct iommu_dev_data *dev_data = NULL; + int devid, flags_rmp, vmg_tag, flags; + struct pci_dev *pdev; + u64 gpa; + + devid = (event[0] >> EVENT_DEVID_SHIFT) & EVENT_DEVID_MASK; + flags_rmp = (event[0] >> EVENT_FLAGS_SHIFT) & 0xFF; + vmg_tag = (event[1]) & 0xFFFF; + flags = (event[1] >> EVENT_FLAGS_SHIFT) & EVENT_FLAGS_MASK; + gpa = ((u64)event[3] << 32) | event[2]; + + pdev = pci_get_domain_bus_and_slot(iommu->pci_seg->id, PCI_BUS_NUM(devid), + devid & 0xff); + if (pdev) + dev_data = dev_iommu_priv_get(&pdev->dev); + + if (dev_data) { + if (__ratelimit(&dev_data->rs)) { + pci_err(pdev, "Event logged [RMP_PAGE_FAULT vmg_tag=0x%04x, gpa=0x%llx, flags_rmp=0x%04x, flags=0x%04x]\n", + vmg_tag, gpa, flags_rmp, flags); + } + } else { + pr_err_ratelimited("Event logged [RMP_PAGE_FAULT device=%04x:%02x:%02x.%x, vmg_tag=0x%04x, gpa=0x%llx, flags_rmp=0x%04x, flags=0x%04x]\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + vmg_tag, gpa, flags_rmp, flags); + } + + if (pdev) + pci_dev_put(pdev); +} + +#define IS_IOMMU_MEM_TRANSACTION(flags) \ + (((flags) & EVENT_FLAG_I) == 0) + +#define IS_WRITE_REQUEST(flags) \ + ((flags) & EVENT_FLAG_RW) + +static void amd_iommu_report_page_fault(struct amd_iommu *iommu, + u16 devid, u16 domain_id, u64 address, int flags) { struct iommu_dev_data *dev_data = NULL; struct pci_dev *pdev; - pdev = pci_get_domain_bus_and_slot(0, PCI_BUS_NUM(devid), + pdev = pci_get_domain_bus_and_slot(iommu->pci_seg->id, PCI_BUS_NUM(devid), devid & 0xff); if (pdev) dev_data = dev_iommu_priv_get(&pdev->dev); - if (dev_data && __ratelimit(&dev_data->rs)) { - pci_err(pdev, "Event logged [IO_PAGE_FAULT domain=0x%04x address=0x%llx flags=0x%04x]\n", - domain_id, address, flags); - } else if (printk_ratelimit()) { - pr_err("Event logged [IO_PAGE_FAULT device=%02x:%02x.%x domain=0x%04x address=0x%llx flags=0x%04x]\n", - PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + if (dev_data) { + /* + * If this is a DMA fault (for which the I(nterrupt) + * bit will be unset), allow report_iommu_fault() to + * prevent logging it. + */ + if (IS_IOMMU_MEM_TRANSACTION(flags)) { + /* Device not attached to domain properly */ + if (dev_data->domain == NULL) { + pr_err_ratelimited("Event logged [Device not attached to domain properly]\n"); + pr_err_ratelimited(" device=%04x:%02x:%02x.%x domain=0x%04x\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), + PCI_FUNC(devid), domain_id); + goto out; + } + + if (!report_iommu_fault(&dev_data->domain->domain, + &pdev->dev, address, + IS_WRITE_REQUEST(flags) ? + IOMMU_FAULT_WRITE : + IOMMU_FAULT_READ)) + goto out; + } + + if (__ratelimit(&dev_data->rs)) { + pci_err(pdev, "Event logged [IO_PAGE_FAULT domain=0x%04x address=0x%llx flags=0x%04x]\n", + domain_id, address, flags); + } + } else { + pr_err_ratelimited("Event logged [IO_PAGE_FAULT device=%04x:%02x:%02x.%x domain=0x%04x address=0x%llx flags=0x%04x]\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), domain_id, address, flags); } +out: if (pdev) pci_dev_put(pdev); } @@ -502,10 +886,11 @@ static void amd_iommu_report_page_fault(u16 devid, u16 domain_id, static void iommu_print_event(struct amd_iommu *iommu, void *__evt) { struct device *dev = iommu->iommu.dev; - int type, devid, pasid, flags, tag; + int type, devid, flags, tag; volatile u32 *event = __evt; int count = 0; - u64 address; + u64 address, ctrl; + u32 pasid; retry: type = (event[1] >> EVENT_TYPE_SHIFT) & EVENT_TYPE_MASK; @@ -514,6 +899,7 @@ retry: (event[1] & EVENT_DOMID_MASK_LO); flags = (event[1] >> EVENT_FLAGS_SHIFT) & EVENT_FLAGS_MASK; address = (u64)(((u64)event[3]) << 32) | event[2]; + ctrl = readq(iommu->mmio_base + MMIO_CONTROL_OFFSET); if (type == 0) { /* Did we hit the erratum? */ @@ -526,26 +912,27 @@ retry: } if (type == EVENT_TYPE_IO_FAULT) { - amd_iommu_report_page_fault(devid, pasid, address, flags); + amd_iommu_report_page_fault(iommu, devid, pasid, address, flags); return; } switch (type) { case EVENT_TYPE_ILL_DEV: - dev_err(dev, "Event logged [ILLEGAL_DEV_TABLE_ENTRY device=%02x:%02x.%x pasid=0x%05x address=0x%llx flags=0x%04x]\n", - PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + dev_err(dev, "Event logged [ILLEGAL_DEV_TABLE_ENTRY device=%04x:%02x:%02x.%x pasid=0x%05x address=0x%llx flags=0x%04x]\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), pasid, address, flags); - dump_dte_entry(devid); + dev_err(dev, "Control Reg : 0x%llx\n", ctrl); + dump_dte_entry(iommu, devid); break; case EVENT_TYPE_DEV_TAB_ERR: - dev_err(dev, "Event logged [DEV_TAB_HARDWARE_ERROR device=%02x:%02x.%x " + dev_err(dev, "Event logged [DEV_TAB_HARDWARE_ERROR device=%04x:%02x:%02x.%x " "address=0x%llx flags=0x%04x]\n", - PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), address, flags); break; case EVENT_TYPE_PAGE_TAB_ERR: - dev_err(dev, "Event logged [PAGE_TAB_HARDWARE_ERROR device=%02x:%02x.%x pasid=0x%04x address=0x%llx flags=0x%04x]\n", - PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + dev_err(dev, "Event logged [PAGE_TAB_HARDWARE_ERROR device=%04x:%02x:%02x.%x pasid=0x%04x address=0x%llx flags=0x%04x]\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), pasid, address, flags); break; case EVENT_TYPE_ILL_CMD: @@ -557,20 +944,26 @@ retry: address, flags); break; case EVENT_TYPE_IOTLB_INV_TO: - dev_err(dev, "Event logged [IOTLB_INV_TIMEOUT device=%02x:%02x.%x address=0x%llx]\n", - PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + dev_err(dev, "Event logged [IOTLB_INV_TIMEOUT device=%04x:%02x:%02x.%x address=0x%llx]\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), address); break; case EVENT_TYPE_INV_DEV_REQ: - dev_err(dev, "Event logged [INVALID_DEVICE_REQUEST device=%02x:%02x.%x pasid=0x%05x address=0x%llx flags=0x%04x]\n", - PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + dev_err(dev, "Event logged [INVALID_DEVICE_REQUEST device=%04x:%02x:%02x.%x pasid=0x%05x address=0x%llx flags=0x%04x]\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), pasid, address, flags); break; + case EVENT_TYPE_RMP_FAULT: + amd_iommu_report_rmp_fault(iommu, event); + break; + case EVENT_TYPE_RMP_HW_ERR: + amd_iommu_report_rmp_hw_error(iommu, event); + break; case EVENT_TYPE_INV_PPR_REQ: pasid = PPR_PASID(*((u64 *)__evt)); tag = event[1] & 0x03FF; - dev_err(dev, "Event logged [INVALID_PPR_REQUEST device=%02x:%02x.%x pasid=0x%05x address=0x%llx flags=0x%04x tag=0x%03x]\n", - PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), + dev_err(dev, "Event logged [INVALID_PPR_REQUEST device=%04x:%02x:%02x.%x pasid=0x%05x address=0x%llx flags=0x%04x tag=0x%03x]\n", + iommu->pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid), pasid, address, flags, tag); break; default: @@ -578,7 +971,14 @@ retry: event[0], event[1], event[2], event[3]); } - memset(__evt, 0, 4 * sizeof(u32)); + /* + * To detect the hardware errata 732 we need to clear the + * entry back to zero. This issue does not exist on SNP + * enabled system. Also this buffer is not writeable on + * SNP enabled system. + */ + if (!amd_iommu_snp_en) + memset(__evt, 0, 4 * sizeof(u32)); } static void iommu_poll_events(struct amd_iommu *iommu) @@ -590,79 +990,12 @@ static void iommu_poll_events(struct amd_iommu *iommu) while (head != tail) { iommu_print_event(iommu, iommu->evt_buf + head); - head = (head + EVENT_ENTRY_SIZE) % EVT_BUFFER_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; - - if (PPR_REQ_TYPE(raw[0]) != PPR_REQ_FAULT) { - pr_err_ratelimited("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); + head = (head + EVENT_ENTRY_SIZE) % EVT_BUFFER_SIZE; + writel(head, iommu->mmio_base + MMIO_EVT_HEAD_OFFSET); } + } #ifdef CONFIG_IRQ_REMAP @@ -672,13 +1005,21 @@ int amd_iommu_register_ga_log_notifier(int (*notifier)(u32)) { iommu_ga_log_notifier = notifier; + /* + * Ensure all in-flight IRQ handlers run to completion before returning + * to the caller, e.g. to ensure module code isn't unloaded while it's + * being executed in the IRQ handler. + */ + if (!notifier) + synchronize_rcu(); + return 0; } EXPORT_SYMBOL(amd_iommu_register_ga_log_notifier); static void iommu_poll_ga_log(struct amd_iommu *iommu) { - u32 head, tail, cnt = 0; + u32 head, tail; if (iommu->ga_log == NULL) return; @@ -691,7 +1032,6 @@ static void iommu_poll_ga_log(struct amd_iommu *iommu) u64 log_entry; raw = (u64 *)(iommu->ga_log + head); - cnt++; /* Avoid memcpy function-call overhead */ log_entry = *raw; @@ -718,39 +1058,43 @@ static void iommu_poll_ga_log(struct amd_iommu *iommu) } } } -#endif /* CONFIG_IRQ_REMAP */ -#define AMD_IOMMU_INT_MASK \ - (MMIO_STATUS_EVT_INT_MASK | \ - MMIO_STATUS_PPR_INT_MASK | \ - MMIO_STATUS_GALOG_INT_MASK) +static void +amd_iommu_set_pci_msi_domain(struct device *dev, struct amd_iommu *iommu) +{ + if (!irq_remapping_enabled || !dev_is_pci(dev) || + !pci_dev_has_default_msi_parent_domain(to_pci_dev(dev))) + return; + + dev_set_msi_domain(dev, iommu->ir_domain); +} + +#else /* CONFIG_IRQ_REMAP */ +static inline void +amd_iommu_set_pci_msi_domain(struct device *dev, struct amd_iommu *iommu) { } +#endif /* !CONFIG_IRQ_REMAP */ -irqreturn_t amd_iommu_int_thread(int irq, void *data) +static void amd_iommu_handle_irq(void *data, const char *evt_type, + u32 int_mask, u32 overflow_mask, + void (*int_handler)(struct amd_iommu *), + void (*overflow_handler)(struct amd_iommu *)) { struct amd_iommu *iommu = (struct amd_iommu *) data; u32 status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); + u32 mask = int_mask | overflow_mask; - while (status & AMD_IOMMU_INT_MASK) { - /* Enable EVT and PPR and GA interrupts again */ - writel(AMD_IOMMU_INT_MASK, - iommu->mmio_base + MMIO_STATUS_OFFSET); - - if (status & MMIO_STATUS_EVT_INT_MASK) { - pr_devel("Processing IOMMU Event Log\n"); - iommu_poll_events(iommu); - } + while (status & mask) { + /* Enable interrupt sources again */ + writel(mask, iommu->mmio_base + MMIO_STATUS_OFFSET); - if (status & MMIO_STATUS_PPR_INT_MASK) { - pr_devel("Processing IOMMU PPR Log\n"); - iommu_poll_ppr_log(iommu); + if (int_handler) { + pr_devel("Processing IOMMU (ivhd%d) %s Log\n", + iommu->index, evt_type); + int_handler(iommu); } -#ifdef CONFIG_IRQ_REMAP - if (status & MMIO_STATUS_GALOG_INT_MASK) { - pr_devel("Processing IOMMU GA Log\n"); - iommu_poll_ga_log(iommu); - } -#endif + if ((status & overflow_mask) && overflow_handler) + overflow_handler(iommu); /* * Hardware bug: ERBT1312 @@ -767,6 +1111,43 @@ irqreturn_t amd_iommu_int_thread(int irq, void *data) */ status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); } +} + +irqreturn_t amd_iommu_int_thread_evtlog(int irq, void *data) +{ + amd_iommu_handle_irq(data, "Evt", MMIO_STATUS_EVT_INT_MASK, + MMIO_STATUS_EVT_OVERFLOW_MASK, + iommu_poll_events, amd_iommu_restart_event_logging); + + return IRQ_HANDLED; +} + +irqreturn_t amd_iommu_int_thread_pprlog(int irq, void *data) +{ + amd_iommu_handle_irq(data, "PPR", MMIO_STATUS_PPR_INT_MASK, + MMIO_STATUS_PPR_OVERFLOW_MASK, + amd_iommu_poll_ppr_log, amd_iommu_restart_ppr_log); + + return IRQ_HANDLED; +} + +irqreturn_t amd_iommu_int_thread_galog(int irq, void *data) +{ +#ifdef CONFIG_IRQ_REMAP + amd_iommu_handle_irq(data, "GA", MMIO_STATUS_GALOG_INT_MASK, + MMIO_STATUS_GALOG_OVERFLOW_MASK, + iommu_poll_ga_log, amd_iommu_restart_ga_log); +#endif + + return IRQ_HANDLED; +} + +irqreturn_t amd_iommu_int_thread(int irq, void *data) +{ + amd_iommu_int_thread_evtlog(irq, data); + amd_iommu_int_thread_pprlog(irq, data); + amd_iommu_int_thread_galog(irq, data); + return IRQ_HANDLED; } @@ -781,17 +1162,43 @@ irqreturn_t amd_iommu_int_handler(int irq, void *data) * ****************************************************************************/ -static int wait_on_sem(volatile u64 *sem) +static void dump_command_buffer(struct amd_iommu *iommu) +{ + struct iommu_cmd *cmd; + u32 head, tail; + int i; + + head = readl(iommu->mmio_base + MMIO_CMD_HEAD_OFFSET); + tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET); + + pr_err("CMD Buffer head=%llu tail=%llu\n", MMIO_CMD_BUFFER_HEAD(head), + MMIO_CMD_BUFFER_TAIL(tail)); + + for (i = 0; i < CMD_BUFFER_ENTRIES; i++) { + cmd = (struct iommu_cmd *)(iommu->cmd_buf + i * sizeof(*cmd)); + pr_err("%3d: %08x %08x %08x %08x\n", i, cmd->data[0], cmd->data[1], cmd->data[2], + cmd->data[3]); + } +} + +static int wait_on_sem(struct amd_iommu *iommu, u64 data) { int i = 0; - while (*sem == 0 && i < LOOP_TIMEOUT) { + while (*iommu->cmd_sem != data && i < LOOP_TIMEOUT) { udelay(1); i += 1; } if (i == LOOP_TIMEOUT) { - pr_alert("Completion-Wait loop timed out\n"); + + pr_alert("IOMMU %04x:%02x:%02x.%01x: Completion-Wait loop timed out\n", + iommu->pci_seg->id, PCI_BUS_NUM(iommu->devid), + PCI_SLOT(iommu->devid), PCI_FUNC(iommu->devid)); + + if (amd_iommu_dump) + DO_ONCE_LITE(dump_command_buffer, iommu); + return -EIO; } @@ -816,16 +1223,17 @@ static void copy_cmd_to_buffer(struct amd_iommu *iommu, writel(tail, iommu->mmio_base + MMIO_CMD_TAIL_OFFSET); } -static void build_completion_wait(struct iommu_cmd *cmd, u64 address) +static void build_completion_wait(struct iommu_cmd *cmd, + struct amd_iommu *iommu, + u64 data) { - u64 paddr = iommu_virt_to_phys((void *)address); - - WARN_ON(address & 0x7ULL); + u64 paddr = iommu->cmd_sem_paddr; memset(cmd, 0, sizeof(*cmd)); cmd->data[0] = lower_32_bits(paddr) | CMD_COMPL_WAIT_STORE_MASK; cmd->data[1] = upper_32_bits(paddr); - cmd->data[2] = 1; + cmd->data[2] = lower_32_bits(data); + cmd->data[3] = upper_32_bits(data); CMD_SET_TYPE(cmd, CMD_COMPL_WAIT); } @@ -836,108 +1244,92 @@ static void build_inv_dte(struct iommu_cmd *cmd, u16 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) +/* + * Builds an invalidation address which is suitable for one page or multiple + * pages. Sets the size bit (S) as needed is more than one page is flushed. + */ +static inline u64 build_inv_address(u64 address, size_t size) { - u64 pages; - bool s; + u64 pages, end, msb_diff; 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; - } + if (pages == 1) + return address & PAGE_MASK; - address &= PAGE_MASK; + end = address + size - 1; - 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; + /* + * msb_diff would hold the index of the most significant bit that + * flipped between the start and end. + */ + msb_diff = fls64(end ^ address) - 1; - if (pages > 1) { + /* + * Bits 63:52 are sign extended. If for some reason bit 51 is different + * between the start and the end, invalidate everything. + */ + if (unlikely(msb_diff > 51)) { + address = CMD_INV_IOMMU_ALL_PAGES_ADDRESS; + } else { /* - * If we have to flush more than one page, flush all - * TLB entries for this domain + * The msb-bit must be clear on the address. Just set all the + * lower bits. */ - address = CMD_INV_IOMMU_ALL_PAGES_ADDRESS; - s = true; + address |= (1ull << msb_diff) - 1; } + /* Clear bits 11:0 */ 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; + /* Set the size bit - we flush more than one 4kb page */ + return address | CMD_INV_IOMMU_PAGES_SIZE_MASK; } -static void build_inv_iommu_pasid(struct iommu_cmd *cmd, u16 domid, int pasid, - u64 address, bool size) +static void build_inv_iommu_pages(struct iommu_cmd *cmd, u64 address, + size_t size, u16 domid, + ioasid_t pasid, bool gn) { - memset(cmd, 0, sizeof(*cmd)); + u64 inv_address = build_inv_address(address, size); - address &= ~(0xfffULL); + memset(cmd, 0, sizeof(*cmd)); - cmd->data[0] = pasid; - cmd->data[1] = domid; - cmd->data[2] = lower_32_bits(address); - cmd->data[3] = upper_32_bits(address); + cmd->data[1] |= domid; + cmd->data[2] = lower_32_bits(inv_address); + cmd->data[3] = upper_32_bits(inv_address); + /* PDE bit - we want to flush everything, not only the PTEs */ 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; + if (gn) { + cmd->data[0] |= pasid; + cmd->data[2] |= CMD_INV_IOMMU_PAGES_GN_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) +static void build_inv_iotlb_pages(struct iommu_cmd *cmd, u16 devid, int qdep, + u64 address, size_t size, + ioasid_t pasid, bool gn) { - memset(cmd, 0, sizeof(*cmd)); + u64 inv_address = build_inv_address(address, size); - address &= ~(0xfffULL); + memset(cmd, 0, sizeof(*cmd)); cmd->data[0] = devid; - cmd->data[0] |= ((pasid >> 8) & 0xff) << 16; - cmd->data[0] |= (qdep & 0xff) << 24; + 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->data[2] = lower_32_bits(inv_address); + cmd->data[3] = upper_32_bits(inv_address); + if (gn) { + cmd->data[0] |= ((pasid >> 8) & 0xff) << 16; + cmd->data[1] |= (pasid & 0xff) << 16; + cmd->data[2] |= CMD_INV_IOMMU_PAGES_GN_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) +static void build_complete_ppr(struct iommu_cmd *cmd, u16 devid, u32 pasid, + int status, int tag, u8 gn) { memset(cmd, 0, sizeof(*cmd)); @@ -1034,22 +1426,21 @@ static int iommu_completion_wait(struct amd_iommu *iommu) struct iommu_cmd cmd; unsigned long flags; int ret; + u64 data; if (!iommu->need_sync) return 0; - - build_completion_wait(&cmd, (u64)&iommu->cmd_sem); + data = atomic64_inc_return(&iommu->cmd_sem_val); + build_completion_wait(&cmd, iommu, data); raw_spin_lock_irqsave(&iommu->lock, flags); - iommu->cmd_sem = 0; - ret = __iommu_queue_command_sync(iommu, &cmd, false); if (ret) goto out_unlock; - ret = wait_on_sem(&iommu->cmd_sem); + ret = wait_on_sem(iommu, data); out_unlock: raw_spin_unlock_irqrestore(&iommu->lock, flags); @@ -1057,6 +1448,21 @@ out_unlock: return ret; } +static void domain_flush_complete(struct protection_domain *domain) +{ + struct pdom_iommu_info *pdom_iommu_info; + unsigned long i; + + lockdep_assert_held(&domain->lock); + + /* + * Devices of this domain are behind this IOMMU + * We need to wait for completion of all commands. + */ + xa_for_each(&domain->iommu_array, i, pdom_iommu_info) + iommu_completion_wait(pdom_iommu_info->iommu); +} + static int iommu_flush_dte(struct amd_iommu *iommu, u16 devid) { struct iommu_cmd cmd; @@ -1066,11 +1472,21 @@ static int iommu_flush_dte(struct amd_iommu *iommu, u16 devid) return iommu_queue_command(iommu, &cmd); } +static void iommu_flush_dte_sync(struct amd_iommu *iommu, u16 devid) +{ + int ret; + + ret = iommu_flush_dte(iommu, devid); + if (!ret) + iommu_completion_wait(iommu); +} + static void amd_iommu_flush_dte_all(struct amd_iommu *iommu) { u32 devid; + u16 last_bdf = iommu->pci_seg->last_bdf; - for (devid = 0; devid <= 0xffff; ++devid) + for (devid = 0; devid <= last_bdf; ++devid) iommu_flush_dte(iommu, devid); iommu_completion_wait(iommu); @@ -1083,11 +1499,12 @@ static void amd_iommu_flush_dte_all(struct amd_iommu *iommu) static void amd_iommu_flush_tlb_all(struct amd_iommu *iommu) { u32 dom_id; + u16 last_bdf = iommu->pci_seg->last_bdf; - for (dom_id = 0; dom_id <= 0xffff; ++dom_id) { + for (dom_id = 0; dom_id <= last_bdf; ++dom_id) { struct iommu_cmd cmd; build_inv_iommu_pages(&cmd, 0, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, - dom_id, 1); + dom_id, IOMMU_NO_PASID, false); iommu_queue_command(iommu, &cmd); } @@ -1099,7 +1516,7 @@ static void amd_iommu_flush_tlb_domid(struct amd_iommu *iommu, u32 dom_id) struct iommu_cmd cmd; build_inv_iommu_pages(&cmd, 0, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, - dom_id, 1); + dom_id, IOMMU_NO_PASID, false); iommu_queue_command(iommu, &cmd); iommu_completion_wait(iommu); @@ -1127,16 +1544,20 @@ static void iommu_flush_irt(struct amd_iommu *iommu, u16 devid) static void amd_iommu_flush_irt_all(struct amd_iommu *iommu) { u32 devid; + u16 last_bdf = iommu->pci_seg->last_bdf; + + if (iommu->irtcachedis_enabled) + return; - for (devid = 0; devid <= MAX_DEV_TABLE_ENTRIES; devid++) + for (devid = 0; devid <= last_bdf; devid++) iommu_flush_irt(iommu, devid); iommu_completion_wait(iommu); } -void iommu_flush_all_caches(struct amd_iommu *iommu) +void amd_iommu_flush_all_caches(struct amd_iommu *iommu) { - if (iommu_feature(iommu, FEATURE_IA)) { + if (check_feature(FEATURE_IA)) { amd_iommu_flush_all(iommu); } else { amd_iommu_flush_dte_all(iommu); @@ -1148,17 +1569,15 @@ void iommu_flush_all_caches(struct amd_iommu *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) +static int device_flush_iotlb(struct iommu_dev_data *dev_data, u64 address, + size_t size, ioasid_t pasid, bool gn) { - struct amd_iommu *iommu; + struct amd_iommu *iommu = get_amd_iommu_from_dev_data(dev_data); struct iommu_cmd cmd; - int qdep; - - qdep = dev_data->ats.qdep; - iommu = amd_iommu_rlookup_table[dev_data->devid]; + int qdep = dev_data->ats_qdep; - build_inv_iotlb_pages(&cmd, dev_data->devid, qdep, address, size); + build_inv_iotlb_pages(&cmd, dev_data->devid, qdep, address, + size, pasid, gn); return iommu_queue_command(iommu, &cmd); } @@ -1175,696 +1594,534 @@ static int device_flush_dte_alias(struct pci_dev *pdev, u16 alias, void *data) */ static int device_flush_dte(struct iommu_dev_data *dev_data) { - struct amd_iommu *iommu; + struct amd_iommu *iommu = get_amd_iommu_from_dev_data(dev_data); + struct pci_dev *pdev = NULL; + struct amd_iommu_pci_seg *pci_seg; u16 alias; int ret; - iommu = amd_iommu_rlookup_table[dev_data->devid]; + if (dev_is_pci(dev_data->dev)) + pdev = to_pci_dev(dev_data->dev); - if (dev_data->pdev) - ret = pci_for_each_dma_alias(dev_data->pdev, + if (pdev) + ret = pci_for_each_dma_alias(pdev, device_flush_dte_alias, iommu); else ret = iommu_flush_dte(iommu, dev_data->devid); if (ret) return ret; - alias = amd_iommu_alias_table[dev_data->devid]; + pci_seg = iommu->pci_seg; + alias = pci_seg->alias_table[dev_data->devid]; if (alias != dev_data->devid) { ret = iommu_flush_dte(iommu, alias); if (ret) return ret; } - if (dev_data->ats.enabled) - ret = device_flush_iotlb(dev_data, 0, ~0UL); + if (dev_data->ats_enabled) { + /* Invalidate the entire contents of an IOTLB */ + ret = device_flush_iotlb(dev_data, 0, ~0UL, + IOMMU_NO_PASID, false); + } 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) +static int domain_flush_pages_v2(struct protection_domain *pdom, + u64 address, size_t size) { struct iommu_dev_data *dev_data; struct iommu_cmd cmd; - int ret = 0, i; + int ret = 0; - build_inv_iommu_pages(&cmd, address, size, domain->id, pde); + lockdep_assert_held(&pdom->lock); + list_for_each_entry(dev_data, &pdom->dev_list, list) { + struct amd_iommu *iommu = get_amd_iommu_from_dev(dev_data->dev); + u16 domid = dev_data->gcr3_info.domid; - for (i = 0; i < amd_iommu_get_num_iommus(); ++i) { - if (!domain->dev_iommu[i]) - continue; + build_inv_iommu_pages(&cmd, address, size, + domid, IOMMU_NO_PASID, true); - /* - * 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); + ret |= iommu_queue_command(iommu, &cmd); } - WARN_ON(ret); -} - -static void domain_flush_pages(struct protection_domain *domain, - u64 address, size_t size) -{ - __domain_flush_pages(domain, address, size, 0); + return ret; } -/* Flush the whole IO/TLB for a given protection domain - including PDE */ -static void domain_flush_tlb_pde(struct protection_domain *domain) +static int domain_flush_pages_v1(struct protection_domain *pdom, + u64 address, size_t size) { - __domain_flush_pages(domain, 0, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, 1); -} + struct pdom_iommu_info *pdom_iommu_info; + struct iommu_cmd cmd; + int ret = 0; + unsigned long i; -static void domain_flush_complete(struct protection_domain *domain) -{ - int i; + lockdep_assert_held(&pdom->lock); - for (i = 0; i < amd_iommu_get_num_iommus(); ++i) { - if (domain && !domain->dev_iommu[i]) - continue; + build_inv_iommu_pages(&cmd, address, size, + pdom->id, IOMMU_NO_PASID, false); + xa_for_each(&pdom->iommu_array, i, pdom_iommu_info) { /* * Devices of this domain are behind this IOMMU - * We need to wait for completion of all commands. + * We need a TLB flush */ - iommu_completion_wait(amd_iommus[i]); + ret |= iommu_queue_command(pdom_iommu_info->iommu, &cmd); } -} - -/* Flush the not present cache if it exists */ -static void domain_flush_np_cache(struct protection_domain *domain, - dma_addr_t iova, size_t size) -{ - if (unlikely(amd_iommu_np_cache)) { - unsigned long flags; - spin_lock_irqsave(&domain->lock, flags); - domain_flush_pages(domain, iova, size); - domain_flush_complete(domain); - spin_unlock_irqrestore(&domain->lock, flags); - } + return ret; } - /* - * This function flushes the DTEs for all devices in domain + * TLB invalidation function which is called from the mapping functions. + * It flushes range of PTEs of the domain. */ -static void domain_flush_devices(struct protection_domain *domain) +static void __domain_flush_pages(struct protection_domain *domain, + u64 address, size_t size) { struct iommu_dev_data *dev_data; + int ret = 0; + ioasid_t pasid = IOMMU_NO_PASID; + bool gn = false; - list_for_each_entry(dev_data, &domain->dev_list, list) - device_flush_dte(dev_data); -} - -/**************************************************************************** - * - * The functions below are used the create the page table mappings for - * unity mapped regions. - * - ****************************************************************************/ + lockdep_assert_held(&domain->lock); -static void free_page_list(struct page *freelist) -{ - while (freelist != NULL) { - unsigned long p = (unsigned long)page_address(freelist); - freelist = freelist->freelist; - free_page(p); + if (pdom_is_v2_pgtbl_mode(domain)) { + gn = true; + ret = domain_flush_pages_v2(domain, address, size); + } else { + ret = domain_flush_pages_v1(domain, address, size); } -} - -static struct page *free_pt_page(unsigned long pt, struct page *freelist) -{ - struct page *p = virt_to_page((void *)pt); - - p->freelist = freelist; - - return p; -} -#define DEFINE_FREE_PT_FN(LVL, FN) \ -static struct page *free_pt_##LVL (unsigned long __pt, struct page *freelist) \ -{ \ - unsigned long p; \ - u64 *pt; \ - int i; \ - \ - pt = (u64 *)__pt; \ - \ - for (i = 0; i < 512; ++i) { \ - /* PTE present? */ \ - if (!IOMMU_PTE_PRESENT(pt[i])) \ - continue; \ - \ - /* Large PTE? */ \ - if (PM_PTE_LEVEL(pt[i]) == 0 || \ - PM_PTE_LEVEL(pt[i]) == 7) \ - continue; \ - \ - p = (unsigned long)IOMMU_PTE_PAGE(pt[i]); \ - freelist = FN(p, freelist); \ - } \ - \ - return free_pt_page((unsigned long)pt, freelist); \ -} + list_for_each_entry(dev_data, &domain->dev_list, list) { -DEFINE_FREE_PT_FN(l2, free_pt_page) -DEFINE_FREE_PT_FN(l3, free_pt_l2) -DEFINE_FREE_PT_FN(l4, free_pt_l3) -DEFINE_FREE_PT_FN(l5, free_pt_l4) -DEFINE_FREE_PT_FN(l6, free_pt_l5) + if (!dev_data->ats_enabled) + continue; -static struct page *free_sub_pt(unsigned long root, int mode, - struct page *freelist) -{ - switch (mode) { - case PAGE_MODE_NONE: - case PAGE_MODE_7_LEVEL: - break; - case PAGE_MODE_1_LEVEL: - freelist = free_pt_page(root, freelist); - break; - case PAGE_MODE_2_LEVEL: - freelist = free_pt_l2(root, freelist); - break; - case PAGE_MODE_3_LEVEL: - freelist = free_pt_l3(root, freelist); - break; - case PAGE_MODE_4_LEVEL: - freelist = free_pt_l4(root, freelist); - break; - case PAGE_MODE_5_LEVEL: - freelist = free_pt_l5(root, freelist); - break; - case PAGE_MODE_6_LEVEL: - freelist = free_pt_l6(root, freelist); - break; - default: - BUG(); + ret |= device_flush_iotlb(dev_data, address, size, pasid, gn); } - return freelist; -} - -static void free_pagetable(struct domain_pgtable *pgtable) -{ - struct page *freelist = NULL; - unsigned long root; - - if (pgtable->mode == PAGE_MODE_NONE) - return; - - BUG_ON(pgtable->mode < PAGE_MODE_NONE || - pgtable->mode > PAGE_MODE_6_LEVEL); - - root = (unsigned long)pgtable->root; - freelist = free_sub_pt(root, pgtable->mode, freelist); - - free_page_list(freelist); + WARN_ON(ret); } -/* - * This function is used to add another level to an IO page table. Adding - * another level increases the size of the address space by 9 bits to a size up - * to 64 bits. - */ -static bool increase_address_space(struct protection_domain *domain, - unsigned long address, - gfp_t gfp) +void amd_iommu_domain_flush_pages(struct protection_domain *domain, + u64 address, size_t size) { - struct domain_pgtable pgtable; - unsigned long flags; - bool ret = true; - u64 *pte, root; + lockdep_assert_held(&domain->lock); - spin_lock_irqsave(&domain->lock, flags); - - amd_iommu_domain_get_pgtable(domain, &pgtable); - - if (address <= PM_LEVEL_SIZE(pgtable.mode)) - goto out; - - ret = false; - if (WARN_ON_ONCE(pgtable.mode == PAGE_MODE_6_LEVEL)) - goto out; + if (likely(!amd_iommu_np_cache)) { + __domain_flush_pages(domain, address, size); - pte = (void *)get_zeroed_page(gfp); - if (!pte) - goto out; - - *pte = PM_LEVEL_PDE(pgtable.mode, iommu_virt_to_phys(pgtable.root)); + /* Wait until IOMMU TLB and all device IOTLB flushes are complete */ + domain_flush_complete(domain); - pgtable.root = pte; - pgtable.mode += 1; - update_and_flush_device_table(domain, &pgtable); - domain_flush_complete(domain); + return; + } /* - * Device Table needs to be updated and flushed before the new root can - * be published. + * When NpCache is on, we infer that we run in a VM and use a vIOMMU. + * In such setups it is best to avoid flushes of ranges which are not + * naturally aligned, since it would lead to flushes of unmodified + * PTEs. Such flushes would require the hypervisor to do more work than + * necessary. Therefore, perform repeated flushes of aligned ranges + * until you cover the range. Each iteration flushes the smaller + * between the natural alignment of the address that we flush and the + * greatest naturally aligned region that fits in the range. */ - root = amd_iommu_domain_encode_pgtable(pte, pgtable.mode); - atomic64_set(&domain->pt_root, root); - - ret = true; - -out: - spin_unlock_irqrestore(&domain->lock, flags); - - return ret; -} - -static u64 *alloc_pte(struct protection_domain *domain, - unsigned long address, - unsigned long page_size, - u64 **pte_page, - gfp_t gfp, - bool *updated) -{ - struct domain_pgtable pgtable; - int level, end_lvl; - u64 *pte, *page; - - BUG_ON(!is_power_of_2(page_size)); + while (size != 0) { + int addr_alignment = __ffs(address); + int size_alignment = __fls(size); + int min_alignment; + size_t flush_size; - amd_iommu_domain_get_pgtable(domain, &pgtable); - - while (address > PM_LEVEL_SIZE(pgtable.mode)) { /* - * Return an error if there is no memory to update the - * page-table. + * size is always non-zero, but address might be zero, causing + * addr_alignment to be negative. As the casting of the + * argument in __ffs(address) to long might trim the high bits + * of the address on x86-32, cast to long when doing the check. */ - if (!increase_address_space(domain, address, gfp)) - return NULL; + if (likely((unsigned long)address != 0)) + min_alignment = min(addr_alignment, size_alignment); + else + min_alignment = size_alignment; - /* Read new values to check if update was successful */ - amd_iommu_domain_get_pgtable(domain, &pgtable); + flush_size = 1ul << min_alignment; + + __domain_flush_pages(domain, address, flush_size); + address += flush_size; + size -= flush_size; } + /* Wait until IOMMU TLB and all device IOTLB flushes are complete */ + domain_flush_complete(domain); +} - level = pgtable.mode - 1; - pte = &pgtable.root[PM_LEVEL_INDEX(level, address)]; - address = PAGE_SIZE_ALIGN(address, page_size); - end_lvl = PAGE_SIZE_LEVEL(page_size); +/* Flush the whole IO/TLB for a given protection domain - including PDE */ +static void amd_iommu_domain_flush_all(struct protection_domain *domain) +{ + amd_iommu_domain_flush_pages(domain, 0, + CMD_INV_IOMMU_ALL_PAGES_ADDRESS); +} - while (level > end_lvl) { - u64 __pte, __npte; - int pte_level; +void amd_iommu_dev_flush_pasid_pages(struct iommu_dev_data *dev_data, + ioasid_t pasid, u64 address, size_t size) +{ + struct iommu_cmd cmd; + struct amd_iommu *iommu = get_amd_iommu_from_dev(dev_data->dev); - __pte = *pte; - pte_level = PM_PTE_LEVEL(__pte); + build_inv_iommu_pages(&cmd, address, size, + dev_data->gcr3_info.domid, pasid, true); + iommu_queue_command(iommu, &cmd); - /* - * If we replace a series of large PTEs, we need - * to tear down all of them. - */ - if (IOMMU_PTE_PRESENT(__pte) && - pte_level == PAGE_MODE_7_LEVEL) { - unsigned long count, i; - u64 *lpte; + if (dev_data->ats_enabled) + device_flush_iotlb(dev_data, address, size, pasid, true); - lpte = first_pte_l7(pte, NULL, &count); + iommu_completion_wait(iommu); +} - /* - * Unmap the replicated PTEs that still match the - * original large mapping - */ - for (i = 0; i < count; ++i) - cmpxchg64(&lpte[i], __pte, 0ULL); +static void dev_flush_pasid_all(struct iommu_dev_data *dev_data, + ioasid_t pasid) +{ + amd_iommu_dev_flush_pasid_pages(dev_data, pasid, 0, + CMD_INV_IOMMU_ALL_PAGES_ADDRESS); +} - *updated = true; - continue; - } +int amd_iommu_complete_ppr(struct device *dev, u32 pasid, int status, int tag) +{ + struct iommu_dev_data *dev_data; + struct amd_iommu *iommu; + struct iommu_cmd cmd; - if (!IOMMU_PTE_PRESENT(__pte) || - pte_level == PAGE_MODE_NONE) { - page = (u64 *)get_zeroed_page(gfp); + dev_data = dev_iommu_priv_get(dev); + iommu = get_amd_iommu_from_dev(dev); - if (!page) - return NULL; + build_complete_ppr(&cmd, dev_data->devid, pasid, status, + tag, dev_data->pri_tlp); - __npte = PM_LEVEL_PDE(level, iommu_virt_to_phys(page)); + return iommu_queue_command(iommu, &cmd); +} - /* pte could have been changed somewhere. */ - if (cmpxchg64(pte, __pte, __npte) != __pte) - free_page((unsigned long)page); - else if (IOMMU_PTE_PRESENT(__pte)) - *updated = true; +/**************************************************************************** + * + * The next functions belong to the domain allocation. A domain is + * allocated for every IOMMU as the default domain. If device isolation + * is enabled, every device get its own domain. The most important thing + * about domains is the page table mapping the DMA address space they + * contain. + * + ****************************************************************************/ - continue; - } +static int pdom_id_alloc(void) +{ + return ida_alloc_range(&pdom_ids, 1, MAX_DOMAIN_ID - 1, GFP_ATOMIC); +} - /* No level skipping support yet */ - if (pte_level != level) - return NULL; +static void pdom_id_free(int id) +{ + ida_free(&pdom_ids, id); +} - level -= 1; +static void free_gcr3_tbl_level1(u64 *tbl) +{ + u64 *ptr; + int i; - pte = IOMMU_PTE_PAGE(__pte); + for (i = 0; i < 512; ++i) { + if (!(tbl[i] & GCR3_VALID)) + continue; - if (pte_page && level == end_lvl) - *pte_page = pte; + ptr = iommu_phys_to_virt(tbl[i] & PAGE_MASK); - pte = &pte[PM_LEVEL_INDEX(level, address)]; + iommu_free_pages(ptr); } - - return pte; } -/* - * This function checks if there is a PTE for a given dma address. If - * there is one, it returns the pointer to it. - */ -static u64 *fetch_pte(struct protection_domain *domain, - unsigned long address, - unsigned long *page_size) +static void free_gcr3_tbl_level2(u64 *tbl) { - struct domain_pgtable pgtable; - int level; - u64 *pte; - - *page_size = 0; - - amd_iommu_domain_get_pgtable(domain, &pgtable); - - if (address > PM_LEVEL_SIZE(pgtable.mode)) - return NULL; - - level = pgtable.mode - 1; - pte = &pgtable.root[PM_LEVEL_INDEX(level, address)]; - *page_size = PTE_LEVEL_PAGE_SIZE(level); - - while (level > 0) { - - /* Not Present */ - if (!IOMMU_PTE_PRESENT(*pte)) - return NULL; - - /* Large PTE */ - if (PM_PTE_LEVEL(*pte) == 7 || - PM_PTE_LEVEL(*pte) == 0) - break; + u64 *ptr; + int i; - /* No level skipping support yet */ - if (PM_PTE_LEVEL(*pte) != level) - return NULL; + for (i = 0; i < 512; ++i) { + if (!(tbl[i] & GCR3_VALID)) + continue; - level -= 1; + ptr = iommu_phys_to_virt(tbl[i] & PAGE_MASK); - /* Walk to the next level */ - pte = IOMMU_PTE_PAGE(*pte); - pte = &pte[PM_LEVEL_INDEX(level, address)]; - *page_size = PTE_LEVEL_PAGE_SIZE(level); + free_gcr3_tbl_level1(ptr); } - - /* - * If we have a series of large PTEs, make - * sure to return a pointer to the first one. - */ - if (PM_PTE_LEVEL(*pte) == PAGE_MODE_7_LEVEL) - pte = first_pte_l7(pte, page_size, NULL); - - return pte; } -static struct page *free_clear_pte(u64 *pte, u64 pteval, struct page *freelist) +static void free_gcr3_table(struct gcr3_tbl_info *gcr3_info) { - unsigned long pt; - int mode; - - while (cmpxchg64(pte, pteval, 0) != pteval) { - pr_warn("AMD-Vi: IOMMU pte changed since we read it\n"); - pteval = *pte; - } + if (gcr3_info->glx == 2) + free_gcr3_tbl_level2(gcr3_info->gcr3_tbl); + else if (gcr3_info->glx == 1) + free_gcr3_tbl_level1(gcr3_info->gcr3_tbl); + else + WARN_ON_ONCE(gcr3_info->glx != 0); - if (!IOMMU_PTE_PRESENT(pteval)) - return freelist; + gcr3_info->glx = 0; - pt = (unsigned long)IOMMU_PTE_PAGE(pteval); - mode = IOMMU_PTE_MODE(pteval); + /* Free per device domain ID */ + pdom_id_free(gcr3_info->domid); - return free_sub_pt(pt, mode, freelist); + iommu_free_pages(gcr3_info->gcr3_tbl); + gcr3_info->gcr3_tbl = NULL; } /* - * Generic mapping functions. It maps a physical address into a DMA - * address space. It allocates the page table pages if necessary. - * In the future it can be extended to a generic mapping function - * supporting all features of AMD IOMMU page tables like level skipping - * and full 64 bit address spaces. + * Number of GCR3 table levels required. Level must be 4-Kbyte + * page and can contain up to 512 entries. */ -static int iommu_map_page(struct protection_domain *dom, - unsigned long bus_addr, - unsigned long phys_addr, - unsigned long page_size, - int prot, - gfp_t gfp) -{ - struct page *freelist = NULL; - bool updated = false; - u64 __pte, *pte; - int ret, i, count; - - BUG_ON(!IS_ALIGNED(bus_addr, page_size)); - BUG_ON(!IS_ALIGNED(phys_addr, page_size)); - - ret = -EINVAL; - if (!(prot & IOMMU_PROT_MASK)) - goto out; - - count = PAGE_SIZE_PTE_COUNT(page_size); - pte = alloc_pte(dom, bus_addr, page_size, NULL, gfp, &updated); - - ret = -ENOMEM; - if (!pte) - goto out; +static int get_gcr3_levels(int pasids) +{ + int levels; - for (i = 0; i < count; ++i) - freelist = free_clear_pte(&pte[i], pte[i], freelist); + if (pasids == -1) + return amd_iommu_max_glx_val; - if (freelist != NULL) - updated = true; + levels = get_count_order(pasids); - if (count > 1) { - __pte = PAGE_SIZE_PTE(__sme_set(phys_addr), page_size); - __pte |= PM_LEVEL_ENC(7) | IOMMU_PTE_PR | IOMMU_PTE_FC; - } else - __pte = __sme_set(phys_addr) | IOMMU_PTE_PR | IOMMU_PTE_FC; + return levels ? (DIV_ROUND_UP(levels, 9) - 1) : levels; +} - if (prot & IOMMU_PROT_IR) - __pte |= IOMMU_PTE_IR; - if (prot & IOMMU_PROT_IW) - __pte |= IOMMU_PTE_IW; +static int setup_gcr3_table(struct gcr3_tbl_info *gcr3_info, + struct amd_iommu *iommu, int pasids) +{ + int levels = get_gcr3_levels(pasids); + int nid = iommu ? dev_to_node(&iommu->dev->dev) : NUMA_NO_NODE; + int domid; - for (i = 0; i < count; ++i) - pte[i] = __pte; + if (levels > amd_iommu_max_glx_val) + return -EINVAL; - ret = 0; + if (gcr3_info->gcr3_tbl) + return -EBUSY; -out: - if (updated) { - unsigned long flags; + /* Allocate per device domain ID */ + domid = pdom_id_alloc(); + if (domid <= 0) + return -ENOSPC; + gcr3_info->domid = domid; - spin_lock_irqsave(&dom->lock, flags); - /* - * Flush domain TLB(s) and wait for completion. Any Device-Table - * Updates and flushing already happened in - * increase_address_space(). - */ - domain_flush_tlb_pde(dom); - domain_flush_complete(dom); - spin_unlock_irqrestore(&dom->lock, flags); + gcr3_info->gcr3_tbl = iommu_alloc_pages_node_sz(nid, GFP_ATOMIC, SZ_4K); + if (gcr3_info->gcr3_tbl == NULL) { + pdom_id_free(domid); + return -ENOMEM; } - /* Everything flushed out, free pages now */ - free_page_list(freelist); + gcr3_info->glx = levels; - return ret; + return 0; } -static unsigned long iommu_unmap_page(struct protection_domain *dom, - unsigned long bus_addr, - unsigned long page_size) +static u64 *__get_gcr3_pte(struct gcr3_tbl_info *gcr3_info, + ioasid_t pasid, bool alloc) { - unsigned long long unmapped; - unsigned long unmap_size; + int index; u64 *pte; + u64 *root = gcr3_info->gcr3_tbl; + int level = gcr3_info->glx; - BUG_ON(!is_power_of_2(page_size)); + while (true) { - unmapped = 0; + index = (pasid >> (9 * level)) & 0x1ff; + pte = &root[index]; - while (unmapped < page_size) { + if (level == 0) + break; - pte = fetch_pte(dom, bus_addr, &unmap_size); + if (!(*pte & GCR3_VALID)) { + if (!alloc) + return NULL; - if (pte) { - int i, count; + root = (void *)get_zeroed_page(GFP_ATOMIC); + if (root == NULL) + return NULL; - count = PAGE_SIZE_PTE_COUNT(unmap_size); - for (i = 0; i < count; i++) - pte[i] = 0ULL; + *pte = iommu_virt_to_phys(root) | GCR3_VALID; } - bus_addr = (bus_addr & ~(unmap_size - 1)) + unmap_size; - unmapped += unmap_size; - } + root = iommu_phys_to_virt(*pte & PAGE_MASK); - BUG_ON(unmapped && !is_power_of_2(unmapped)); + level -= 1; + } - return unmapped; + return pte; } -/**************************************************************************** - * - * The next functions belong to the domain allocation. A domain is - * allocated for every IOMMU as the default domain. If device isolation - * is enabled, every device get its own domain. The most important thing - * about domains is the page table mapping the DMA address space they - * contain. - * - ****************************************************************************/ - -static u16 domain_id_alloc(void) +static int update_gcr3(struct iommu_dev_data *dev_data, + ioasid_t pasid, unsigned long gcr3, bool set) { - int id; + struct gcr3_tbl_info *gcr3_info = &dev_data->gcr3_info; + u64 *pte; + + pte = __get_gcr3_pte(gcr3_info, pasid, true); + if (pte == NULL) + return -ENOMEM; - spin_lock(&pd_bitmap_lock); - id = find_first_zero_bit(amd_iommu_pd_alloc_bitmap, MAX_DOMAIN_ID); - BUG_ON(id == 0); - if (id > 0 && id < MAX_DOMAIN_ID) - __set_bit(id, amd_iommu_pd_alloc_bitmap); + if (set) + *pte = (gcr3 & PAGE_MASK) | GCR3_VALID; else - id = 0; - spin_unlock(&pd_bitmap_lock); + *pte = 0; - return id; + dev_flush_pasid_all(dev_data, pasid); + return 0; } -static void domain_id_free(int id) +int amd_iommu_set_gcr3(struct iommu_dev_data *dev_data, ioasid_t pasid, + unsigned long gcr3) { - spin_lock(&pd_bitmap_lock); - if (id > 0 && id < MAX_DOMAIN_ID) - __clear_bit(id, amd_iommu_pd_alloc_bitmap); - spin_unlock(&pd_bitmap_lock); + struct gcr3_tbl_info *gcr3_info = &dev_data->gcr3_info; + int ret; + + iommu_group_mutex_assert(dev_data->dev); + + ret = update_gcr3(dev_data, pasid, gcr3, true); + if (ret) + return ret; + + gcr3_info->pasid_cnt++; + return ret; } -static void free_gcr3_tbl_level1(u64 *tbl) +int amd_iommu_clear_gcr3(struct iommu_dev_data *dev_data, ioasid_t pasid) { - u64 *ptr; - int i; + struct gcr3_tbl_info *gcr3_info = &dev_data->gcr3_info; + int ret; - for (i = 0; i < 512; ++i) { - if (!(tbl[i] & GCR3_VALID)) - continue; + iommu_group_mutex_assert(dev_data->dev); - ptr = iommu_phys_to_virt(tbl[i] & PAGE_MASK); + ret = update_gcr3(dev_data, pasid, 0, false); + if (ret) + return ret; - free_page((unsigned long)ptr); - } + gcr3_info->pasid_cnt--; + return ret; } -static void free_gcr3_tbl_level2(u64 *tbl) +static void make_clear_dte(struct iommu_dev_data *dev_data, struct dev_table_entry *ptr, + struct dev_table_entry *new) { - u64 *ptr; - int i; + /* All existing DTE must have V bit set */ + new->data128[0] = DTE_FLAG_V; + new->data128[1] = 0; +} - for (i = 0; i < 512; ++i) { - if (!(tbl[i] & GCR3_VALID)) - continue; +/* + * Note: + * The old value for GCR3 table and GPT have been cleared from caller. + */ +static void set_dte_gcr3_table(struct amd_iommu *iommu, + struct iommu_dev_data *dev_data, + struct dev_table_entry *target) +{ + struct gcr3_tbl_info *gcr3_info = &dev_data->gcr3_info; + u64 gcr3; - ptr = iommu_phys_to_virt(tbl[i] & PAGE_MASK); + if (!gcr3_info->gcr3_tbl) + return; - free_gcr3_tbl_level1(ptr); - } -} + pr_debug("%s: devid=%#x, glx=%#x, gcr3_tbl=%#llx\n", + __func__, dev_data->devid, gcr3_info->glx, + (unsigned long long)gcr3_info->gcr3_tbl); -static void free_gcr3_table(struct protection_domain *domain) -{ - if (domain->glx == 2) - free_gcr3_tbl_level2(domain->gcr3_tbl); - else if (domain->glx == 1) - free_gcr3_tbl_level1(domain->gcr3_tbl); - else - BUG_ON(domain->glx != 0); + gcr3 = iommu_virt_to_phys(gcr3_info->gcr3_tbl); - free_page((unsigned long)domain->gcr3_tbl); + target->data[0] |= DTE_FLAG_GV | + FIELD_PREP(DTE_GLX, gcr3_info->glx) | + FIELD_PREP(DTE_GCR3_14_12, gcr3 >> 12); + if (pdom_is_v2_pgtbl_mode(dev_data->domain)) + target->data[0] |= DTE_FLAG_GIOV; + + target->data[1] |= FIELD_PREP(DTE_GCR3_30_15, gcr3 >> 15) | + FIELD_PREP(DTE_GCR3_51_31, gcr3 >> 31); + + /* Guest page table can only support 4 and 5 levels */ + if (amd_iommu_gpt_level == PAGE_MODE_5_LEVEL) + target->data[2] |= FIELD_PREP(DTE_GPT_LEVEL_MASK, GUEST_PGTABLE_5_LEVEL); + else + target->data[2] |= FIELD_PREP(DTE_GPT_LEVEL_MASK, GUEST_PGTABLE_4_LEVEL); } -static void set_dte_entry(u16 devid, struct protection_domain *domain, - struct domain_pgtable *pgtable, - bool ats, bool ppr) +static void set_dte_entry(struct amd_iommu *iommu, + struct iommu_dev_data *dev_data, + phys_addr_t top_paddr, unsigned int top_level) { - u64 pte_root = 0; - u64 flags = 0; + u16 domid; u32 old_domid; + struct dev_table_entry *initial_dte; + struct dev_table_entry new = {}; + struct protection_domain *domain = dev_data->domain; + struct gcr3_tbl_info *gcr3_info = &dev_data->gcr3_info; + struct dev_table_entry *dte = &get_dev_table(iommu)[dev_data->devid]; + struct pt_iommu_amdv1_hw_info pt_info; - if (pgtable->mode != PAGE_MODE_NONE) - pte_root = iommu_virt_to_phys(pgtable->root); - - pte_root |= (pgtable->mode & DEV_ENTRY_MODE_MASK) - << DEV_ENTRY_MODE_SHIFT; - pte_root |= DTE_FLAG_IR | DTE_FLAG_IW | DTE_FLAG_V | DTE_FLAG_TV; - - flags = amd_iommu_dev_table[devid].data[1]; + make_clear_dte(dev_data, dte, &new); - if (ats) - flags |= DTE_FLAG_IOTLB; + if (gcr3_info && gcr3_info->gcr3_tbl) + domid = dev_data->gcr3_info.domid; + else { + domid = domain->id; - if (ppr) { - struct amd_iommu *iommu = amd_iommu_rlookup_table[devid]; + if (domain->domain.type & __IOMMU_DOMAIN_PAGING) { + /* + * When updating the IO pagetable, the new top and level + * are provided as parameters. For other operations i.e. + * device attach, retrieve the current pagetable info + * via the IOMMU PT API. + */ + if (top_paddr) { + pt_info.host_pt_root = top_paddr; + pt_info.mode = top_level + 1; + } else { + WARN_ON(top_paddr || top_level); + pt_iommu_amdv1_hw_info(&domain->amdv1, + &pt_info); + } - if (iommu_feature(iommu, FEATURE_EPHSUP)) - pte_root |= 1ULL << DEV_ENTRY_PPR; + new.data[0] |= __sme_set(pt_info.host_pt_root) | + (pt_info.mode & DEV_ENTRY_MODE_MASK) + << DEV_ENTRY_MODE_SHIFT; + } } - if (domain->flags & PD_IOMMUV2_MASK) { - u64 gcr3 = iommu_virt_to_phys(domain->gcr3_tbl); - u64 glx = domain->glx; - u64 tmp; + new.data[0] |= DTE_FLAG_IR | DTE_FLAG_IW; - pte_root |= DTE_FLAG_GV; - pte_root |= (glx & DTE_GLX_MASK) << DTE_GLX_SHIFT; + /* + * When SNP is enabled, we can only support TV=1 with non-zero domain ID. + * This is prevented by the SNP-enable and IOMMU_DOMAIN_IDENTITY check in + * do_iommu_domain_alloc(). + */ + WARN_ON(amd_iommu_snp_en && (domid == 0)); + new.data[0] |= DTE_FLAG_TV; - /* First mask out possible old values for GCR3 table */ - tmp = DTE_GCR3_VAL_B(~0ULL) << DTE_GCR3_SHIFT_B; - flags &= ~tmp; + if (dev_data->ppr) + new.data[0] |= 1ULL << DEV_ENTRY_PPR; - tmp = DTE_GCR3_VAL_C(~0ULL) << DTE_GCR3_SHIFT_C; - flags &= ~tmp; + if (domain->dirty_tracking) + new.data[0] |= DTE_FLAG_HAD; - /* Encode GCR3 table into DTE */ - tmp = DTE_GCR3_VAL_A(gcr3) << DTE_GCR3_SHIFT_A; - pte_root |= tmp; + if (dev_data->ats_enabled) + new.data[1] |= DTE_FLAG_IOTLB; - tmp = DTE_GCR3_VAL_B(gcr3) << DTE_GCR3_SHIFT_B; - flags |= tmp; + old_domid = READ_ONCE(dte->data[1]) & DEV_DOMID_MASK; + new.data[1] |= domid; - tmp = DTE_GCR3_VAL_C(gcr3) << DTE_GCR3_SHIFT_C; - flags |= tmp; + /* + * Restore cached persistent DTE bits, which can be set by information + * in IVRS table. See set_dev_entry_from_acpi(). + */ + initial_dte = amd_iommu_get_ivhd_dte_flags(iommu->pci_seg->id, dev_data->devid); + if (initial_dte) { + new.data128[0] |= initial_dte->data128[0]; + new.data128[1] |= initial_dte->data128[1]; } - flags &= ~DEV_DOMID_MASK; - flags |= domain->id; + set_dte_gcr3_table(iommu, dev_data, &new); - old_domid = amd_iommu_dev_table[devid].data[1] & DEV_DOMID_MASK; - amd_iommu_dev_table[devid].data[1] = flags; - amd_iommu_dev_table[devid].data[0] = pte_root; + update_dte256(iommu, dev_data, &new); /* * A kdump kernel might be replacing a domain ID that was copied from @@ -1872,142 +2129,149 @@ static void set_dte_entry(u16 devid, struct protection_domain *domain, * entries for the old domain ID that is being overwritten */ if (old_domid) { - struct amd_iommu *iommu = amd_iommu_rlookup_table[devid]; - amd_iommu_flush_tlb_domid(iommu, old_domid); } } -static void clear_dte_entry(u16 devid) +/* + * Clear DMA-remap related flags to block all DMA (blockeded domain) + */ +static void clear_dte_entry(struct amd_iommu *iommu, struct iommu_dev_data *dev_data) { - /* remove entry from the device table seen by the hardware */ - amd_iommu_dev_table[devid].data[0] = DTE_FLAG_V | DTE_FLAG_TV; - amd_iommu_dev_table[devid].data[1] &= DTE_FLAG_MASK; + struct dev_table_entry new = {}; + struct dev_table_entry *dte = &get_dev_table(iommu)[dev_data->devid]; - amd_iommu_apply_erratum_63(devid); + make_clear_dte(dev_data, dte, &new); + update_dte256(iommu, dev_data, &new); } -static void do_attach(struct iommu_dev_data *dev_data, - struct protection_domain *domain) +/* Update and flush DTE for the given device */ +static void dev_update_dte(struct iommu_dev_data *dev_data, bool set) { - struct domain_pgtable pgtable; - struct amd_iommu *iommu; - bool ats; - - iommu = amd_iommu_rlookup_table[dev_data->devid]; - ats = dev_data->ats.enabled; - - /* Update data structures */ - dev_data->domain = domain; - list_add(&dev_data->list, &domain->dev_list); - - /* Do reference counting */ - domain->dev_iommu[iommu->index] += 1; - domain->dev_cnt += 1; + struct amd_iommu *iommu = get_amd_iommu_from_dev(dev_data->dev); - /* Update device table */ - amd_iommu_domain_get_pgtable(domain, &pgtable); - set_dte_entry(dev_data->devid, domain, &pgtable, - ats, dev_data->iommu_v2); - clone_aliases(dev_data->pdev); + if (set) + set_dte_entry(iommu, dev_data, 0, 0); + else + clear_dte_entry(iommu, dev_data); + clone_aliases(iommu, dev_data->dev); device_flush_dte(dev_data); + iommu_completion_wait(iommu); } -static void do_detach(struct iommu_dev_data *dev_data) -{ - struct protection_domain *domain = dev_data->domain; - struct amd_iommu *iommu; - - iommu = amd_iommu_rlookup_table[dev_data->devid]; - - /* Update data structures */ - dev_data->domain = NULL; - list_del(&dev_data->list); - clear_dte_entry(dev_data->devid); - clone_aliases(dev_data->pdev); - - /* Flush the DTE entry */ - device_flush_dte(dev_data); +/* + * If domain is SVA capable then initialize GCR3 table. Also if domain is + * in v2 page table mode then update GCR3[0]. + */ +static int init_gcr3_table(struct iommu_dev_data *dev_data, + struct protection_domain *pdom) +{ + struct amd_iommu *iommu = get_amd_iommu_from_dev_data(dev_data); + int max_pasids = dev_data->max_pasids; + struct pt_iommu_x86_64_hw_info pt_info; + int ret = 0; + + /* + * If domain is in pt mode then setup GCR3 table only if device + * is PASID capable + */ + if (pdom_is_in_pt_mode(pdom) && !pdev_pasid_supported(dev_data)) + return ret; - /* Flush IOTLB */ - domain_flush_tlb_pde(domain); + /* + * By default, setup GCR3 table to support MAX PASIDs + * supported by the device/IOMMU. + */ + ret = setup_gcr3_table(&dev_data->gcr3_info, iommu, + max_pasids > 0 ? max_pasids : 1); + if (ret) + return ret; - /* Wait for the flushes to finish */ - domain_flush_complete(domain); + /* Setup GCR3[0] only if domain is setup with v2 page table mode */ + if (!pdom_is_v2_pgtbl_mode(pdom)) + return ret; - /* decrease reference counters - needs to happen after the flushes */ - domain->dev_iommu[iommu->index] -= 1; - domain->dev_cnt -= 1; -} + pt_iommu_x86_64_hw_info(&pdom->amdv2, &pt_info); + ret = update_gcr3(dev_data, 0, __sme_set(pt_info.gcr3_pt), true); + if (ret) + free_gcr3_table(&dev_data->gcr3_info); -static void pdev_iommuv2_disable(struct pci_dev *pdev) -{ - pci_disable_ats(pdev); - pci_disable_pri(pdev); - pci_disable_pasid(pdev); + return ret; } -/* FIXME: Change generic reset-function to do the same */ -static int pri_reset_while_enabled(struct pci_dev *pdev) +static void destroy_gcr3_table(struct iommu_dev_data *dev_data, + struct protection_domain *pdom) { - u16 control; - int pos; + struct gcr3_tbl_info *gcr3_info = &dev_data->gcr3_info; - pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_PRI); - if (!pos) - return -EINVAL; + if (pdom_is_v2_pgtbl_mode(pdom)) + update_gcr3(dev_data, 0, 0, false); - pci_read_config_word(pdev, pos + PCI_PRI_CTRL, &control); - control |= PCI_PRI_CTRL_RESET; - pci_write_config_word(pdev, pos + PCI_PRI_CTRL, control); + if (gcr3_info->gcr3_tbl == NULL) + return; - return 0; + free_gcr3_table(gcr3_info); } -static int pdev_iommuv2_enable(struct pci_dev *pdev) +static int pdom_attach_iommu(struct amd_iommu *iommu, + struct protection_domain *pdom) { - bool reset_enable; - int reqs, ret; + struct pdom_iommu_info *pdom_iommu_info, *curr; + unsigned long flags; + int ret = 0; - /* FIXME: Hardcode number of outstanding requests for now */ - reqs = 32; - if (pdev_pri_erratum(pdev, AMD_PRI_DEV_ERRATUM_LIMIT_REQ_ONE)) - reqs = 1; - reset_enable = pdev_pri_erratum(pdev, AMD_PRI_DEV_ERRATUM_ENABLE_RESET); + spin_lock_irqsave(&pdom->lock, flags); - /* Only allow access to user-accessible pages */ - ret = pci_enable_pasid(pdev, 0); - if (ret) - goto out_err; + pdom_iommu_info = xa_load(&pdom->iommu_array, iommu->index); + if (pdom_iommu_info) { + pdom_iommu_info->refcnt++; + goto out_unlock; + } - /* First reset the PRI state of the device */ - ret = pci_reset_pri(pdev); - if (ret) - goto out_err; + pdom_iommu_info = kzalloc(sizeof(*pdom_iommu_info), GFP_ATOMIC); + if (!pdom_iommu_info) { + ret = -ENOMEM; + goto out_unlock; + } - /* Enable PRI */ - ret = pci_enable_pri(pdev, reqs); - if (ret) - goto out_err; + pdom_iommu_info->iommu = iommu; + pdom_iommu_info->refcnt = 1; - if (reset_enable) { - ret = pri_reset_while_enabled(pdev); - if (ret) - goto out_err; + curr = xa_cmpxchg(&pdom->iommu_array, iommu->index, + NULL, pdom_iommu_info, GFP_ATOMIC); + if (curr) { + kfree(pdom_iommu_info); + ret = -ENOSPC; + goto out_unlock; } - ret = pci_enable_ats(pdev, PAGE_SHIFT); - if (ret) - goto out_err; +out_unlock: + spin_unlock_irqrestore(&pdom->lock, flags); + return ret; +} - return 0; +static void pdom_detach_iommu(struct amd_iommu *iommu, + struct protection_domain *pdom) +{ + struct pdom_iommu_info *pdom_iommu_info; + unsigned long flags; -out_err: - pci_disable_pri(pdev); - pci_disable_pasid(pdev); + spin_lock_irqsave(&pdom->lock, flags); - return ret; + pdom_iommu_info = xa_load(&pdom->iommu_array, iommu->index); + if (!pdom_iommu_info) { + spin_unlock_irqrestore(&pdom->lock, flags); + return; + } + + pdom_iommu_info->refcnt--; + if (pdom_iommu_info->refcnt == 0) { + xa_erase(&pdom->iommu_array, iommu->index); + kfree(pdom_iommu_info); + } + + spin_unlock_irqrestore(&pdom->lock, flags); } /* @@ -2017,64 +2281,59 @@ out_err: static int attach_device(struct device *dev, struct protection_domain *domain) { - struct iommu_dev_data *dev_data; + struct iommu_dev_data *dev_data = dev_iommu_priv_get(dev); + struct amd_iommu *iommu = get_amd_iommu_from_dev_data(dev_data); struct pci_dev *pdev; unsigned long flags; - int ret; - - spin_lock_irqsave(&domain->lock, flags); + int ret = 0; - dev_data = dev_iommu_priv_get(dev); - - spin_lock(&dev_data->lock); + mutex_lock(&dev_data->mutex); - ret = -EBUSY; - if (dev_data->domain != NULL) + if (dev_data->domain != NULL) { + ret = -EBUSY; goto out; + } - if (!dev_is_pci(dev)) - goto skip_ats_check; - - pdev = to_pci_dev(dev); - if (domain->flags & PD_IOMMUV2_MASK) { - struct iommu_domain *def_domain = iommu_get_dma_domain(dev); + /* Do reference counting */ + ret = pdom_attach_iommu(iommu, domain); + if (ret) + goto out; - ret = -EINVAL; - if (def_domain->type != IOMMU_DOMAIN_IDENTITY) + /* Setup GCR3 table */ + if (pdom_is_sva_capable(domain)) { + ret = init_gcr3_table(dev_data, domain); + if (ret) { + pdom_detach_iommu(iommu, domain); goto out; - - if (dev_data->iommu_v2) { - if (pdev_iommuv2_enable(pdev) != 0) - goto out; - - dev_data->ats.enabled = true; - dev_data->ats.qdep = pci_ats_queue_depth(pdev); - dev_data->pri_tlp = pci_prg_resp_pasid_required(pdev); } - } else if (amd_iommu_iotlb_sup && - pci_enable_ats(pdev, PAGE_SHIFT) == 0) { - dev_data->ats.enabled = true; - dev_data->ats.qdep = pci_ats_queue_depth(pdev); } -skip_ats_check: - ret = 0; + pdev = dev_is_pci(dev_data->dev) ? to_pci_dev(dev_data->dev) : NULL; + if (pdev && pdom_is_sva_capable(domain)) { + pdev_enable_caps(pdev); - do_attach(dev_data, domain); + /* + * Device can continue to function even if IOPF + * enablement failed. Hence in error path just + * disable device PRI support. + */ + if (amd_iommu_iopf_add_device(iommu, dev_data)) + pdev_disable_cap_pri(pdev); + } else if (pdev) { + pdev_enable_cap_ats(pdev); + } - /* - * We might boot into a crash-kernel here. The crashed kernel - * left the caches in the IOMMU dirty. So we have to flush - * here to evict all dirty stuff. - */ - domain_flush_tlb_pde(domain); + /* Update data structures */ + dev_data->domain = domain; + spin_lock_irqsave(&domain->lock, flags); + list_add(&dev_data->list, &domain->dev_list); + spin_unlock_irqrestore(&domain->lock, flags); - domain_flush_complete(domain); + /* Update device table */ + dev_update_dte(dev_data, true); out: - spin_unlock(&dev_data->lock); - - spin_unlock_irqrestore(&domain->lock, flags); + mutex_unlock(&dev_data->mutex); return ret; } @@ -2084,16 +2343,12 @@ out: */ static void detach_device(struct device *dev) { - struct protection_domain *domain; - struct iommu_dev_data *dev_data; + struct iommu_dev_data *dev_data = dev_iommu_priv_get(dev); + struct amd_iommu *iommu = get_amd_iommu_from_dev_data(dev_data); + struct protection_domain *domain = dev_data->domain; unsigned long flags; - dev_data = dev_iommu_priv_get(dev); - domain = dev_data->domain; - - spin_lock_irqsave(&domain->lock, flags); - - spin_lock(&dev_data->lock); + mutex_lock(&dev_data->mutex); /* * First check if the device is still attached. It might already @@ -2104,79 +2359,113 @@ static void detach_device(struct device *dev) if (WARN_ON(!dev_data->domain)) goto out; - do_detach(dev_data); + /* Remove IOPF handler */ + if (dev_data->ppr) { + iopf_queue_flush_dev(dev); + amd_iommu_iopf_remove_device(iommu, dev_data); + } - if (!dev_is_pci(dev)) - goto out; + if (dev_is_pci(dev)) + pdev_disable_caps(to_pci_dev(dev)); - if (domain->flags & PD_IOMMUV2_MASK && dev_data->iommu_v2) - pdev_iommuv2_disable(to_pci_dev(dev)); - else if (dev_data->ats.enabled) - pci_disable_ats(to_pci_dev(dev)); + /* Clear DTE and flush the entry */ + dev_update_dte(dev_data, false); - dev_data->ats.enabled = false; + /* Flush IOTLB and wait for the flushes to finish */ + spin_lock_irqsave(&domain->lock, flags); + amd_iommu_domain_flush_all(domain); + list_del(&dev_data->list); + spin_unlock_irqrestore(&domain->lock, flags); -out: - spin_unlock(&dev_data->lock); + /* Clear GCR3 table */ + if (pdom_is_sva_capable(domain)) + destroy_gcr3_table(dev_data, domain); - spin_unlock_irqrestore(&domain->lock, flags); + /* Update data structures */ + dev_data->domain = NULL; + + /* decrease reference counters - needs to happen after the flushes */ + pdom_detach_iommu(iommu, domain); + +out: + mutex_unlock(&dev_data->mutex); } static struct iommu_device *amd_iommu_probe_device(struct device *dev) { struct iommu_device *iommu_dev; struct amd_iommu *iommu; - int ret, devid; + struct iommu_dev_data *dev_data; + int ret; if (!check_device(dev)) return ERR_PTR(-ENODEV); - devid = get_device_id(dev); - if (devid < 0) - return ERR_PTR(devid); + iommu = rlookup_amd_iommu(dev); + if (!iommu) + return ERR_PTR(-ENODEV); - iommu = amd_iommu_rlookup_table[devid]; + /* Not registered yet? */ + if (!iommu->iommu.ops) + return ERR_PTR(-ENODEV); if (dev_iommu_priv_get(dev)) return &iommu->iommu; - ret = iommu_init_device(dev); + ret = iommu_init_device(iommu, dev); if (ret) { - if (ret != -ENOTSUPP) - dev_err(dev, "Failed to initialize - trying to proceed anyway\n"); + dev_err(dev, "Failed to initialize - trying to proceed anyway\n"); iommu_dev = ERR_PTR(ret); - iommu_ignore_device(dev); - } else { - iommu_dev = &iommu->iommu; + iommu_ignore_device(iommu, dev); + goto out_err; + } + + amd_iommu_set_pci_msi_domain(dev, iommu); + iommu_dev = &iommu->iommu; + + /* + * If IOMMU and device supports PASID then it will contain max + * supported PASIDs, else it will be zero. + */ + dev_data = dev_iommu_priv_get(dev); + if (amd_iommu_pasid_supported() && dev_is_pci(dev) && + pdev_pasid_supported(dev_data)) { + dev_data->max_pasids = min_t(u32, iommu->iommu.max_pasids, + pci_max_pasids(to_pci_dev(dev))); + } + + if (amd_iommu_pgtable == PD_MODE_NONE) { + pr_warn_once("%s: DMA translation not supported by iommu.\n", + __func__); + iommu_dev = ERR_PTR(-ENODEV); + goto out_err; } +out_err: + iommu_completion_wait(iommu); - return iommu_dev; -} + if (FEATURE_NUM_INT_REMAP_SUP_2K(amd_iommu_efr2)) + dev_data->max_irqs = MAX_IRQS_PER_TABLE_2K; + else + dev_data->max_irqs = MAX_IRQS_PER_TABLE_512; -static void amd_iommu_probe_finalize(struct device *dev) -{ - struct iommu_domain *domain; + if (dev_is_pci(dev)) + pci_prepare_ats(to_pci_dev(dev), PAGE_SHIFT); - /* Domains are initialized for this device - have a look what we ended up with */ - domain = iommu_get_domain_for_dev(dev); - if (domain->type == IOMMU_DOMAIN_DMA) - iommu_setup_dma_ops(dev, IOVA_START_PFN << PAGE_SHIFT, 0); + return iommu_dev; } static void amd_iommu_release_device(struct device *dev) { - int devid = get_device_id(dev); - struct amd_iommu *iommu; - - if (!check_device(dev)) - return; + struct iommu_dev_data *dev_data = dev_iommu_priv_get(dev); - iommu = amd_iommu_rlookup_table[devid]; + WARN_ON(dev_data->domain); - amd_iommu_uninit_device(dev); - iommu_completion_wait(iommu); + /* + * 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. + */ } static struct iommu_group *amd_iommu_device_group(struct device *dev) @@ -2187,284 +2476,410 @@ static struct iommu_group *amd_iommu_device_group(struct device *dev) return acpihid_device_group(dev); } -static int amd_iommu_domain_get_attr(struct iommu_domain *domain, - enum iommu_attr attr, void *data) -{ - switch (domain->type) { - case IOMMU_DOMAIN_UNMANAGED: - return -ENODEV; - case IOMMU_DOMAIN_DMA: - switch (attr) { - case DOMAIN_ATTR_DMA_USE_FLUSH_QUEUE: - *(int *)data = !amd_iommu_unmap_flush; - return 0; - default: - return -ENODEV; - } - break; - default: - return -EINVAL; - } -} - /***************************************************************************** * - * The next functions belong to the dma_ops mapping/unmapping code. + * The following functions belong to the exported interface of AMD IOMMU + * + * This interface allows access to lower level functions of the IOMMU + * like protection domain handling and assignement of devices to domains + * which is not possible with the dma_ops interface. * *****************************************************************************/ -static void update_device_table(struct protection_domain *domain, - struct domain_pgtable *pgtable) +static void protection_domain_init(struct protection_domain *domain) { - struct iommu_dev_data *dev_data; - - list_for_each_entry(dev_data, &domain->dev_list, list) { - set_dte_entry(dev_data->devid, domain, pgtable, - dev_data->ats.enabled, dev_data->iommu_v2); - clone_aliases(dev_data->pdev); - } + spin_lock_init(&domain->lock); + INIT_LIST_HEAD(&domain->dev_list); + INIT_LIST_HEAD(&domain->dev_data_list); + xa_init(&domain->iommu_array); } -static void update_and_flush_device_table(struct protection_domain *domain, - struct domain_pgtable *pgtable) +struct protection_domain *protection_domain_alloc(void) { - update_device_table(domain, pgtable); - domain_flush_devices(domain); -} + struct protection_domain *domain; + int domid; -static void update_domain(struct protection_domain *domain) -{ - struct domain_pgtable pgtable; + domain = kzalloc(sizeof(*domain), GFP_KERNEL); + if (!domain) + return NULL; - /* Update device table */ - amd_iommu_domain_get_pgtable(domain, &pgtable); - update_and_flush_device_table(domain, &pgtable); + domid = pdom_id_alloc(); + if (domid <= 0) { + kfree(domain); + return NULL; + } + domain->id = domid; - /* Flush domain TLB(s) and wait for completion */ - domain_flush_tlb_pde(domain); - domain_flush_complete(domain); + protection_domain_init(domain); + + return domain; } -int __init amd_iommu_init_api(void) +static bool amd_iommu_hd_support(struct amd_iommu *iommu) { - int ret, err = 0; + if (amd_iommu_hatdis) + return false; - ret = iova_cache_get(); - if (ret) - return ret; + return iommu && (iommu->features & FEATURE_HDSUP); +} - err = bus_set_iommu(&pci_bus_type, &amd_iommu_ops); - if (err) - return err; -#ifdef CONFIG_ARM_AMBA - err = bus_set_iommu(&amba_bustype, &amd_iommu_ops); - if (err) - return err; -#endif - err = bus_set_iommu(&platform_bus_type, &amd_iommu_ops); - if (err) - return err; +static spinlock_t *amd_iommu_get_top_lock(struct pt_iommu *iommupt) +{ + struct protection_domain *pdom = + container_of(iommupt, struct protection_domain, iommu); - return 0; + return &pdom->lock; } -int __init amd_iommu_init_dma_ops(void) +/* + * Update all HW references to the domain with a new pgtable configuration. + */ +static void amd_iommu_change_top(struct pt_iommu *iommu_table, + phys_addr_t top_paddr, unsigned int top_level) { - swiotlb = (iommu_default_passthrough() || sme_me_mask) ? 1 : 0; + struct protection_domain *pdom = + container_of(iommu_table, struct protection_domain, iommu); + struct iommu_dev_data *dev_data; - if (amd_iommu_unmap_flush) - pr_info("IO/TLB flush on unmap enabled\n"); - else - pr_info("Lazy IO/TLB flushing enabled\n"); + lockdep_assert_held(&pdom->lock); - return 0; + /* Update the DTE for all devices attached to this domain */ + list_for_each_entry(dev_data, &pdom->dev_list, list) { + struct amd_iommu *iommu = rlookup_amd_iommu(dev_data->dev); -} + /* Update the HW references with the new level and top ptr */ + set_dte_entry(iommu, dev_data, top_paddr, top_level); + clone_aliases(iommu, dev_data->dev); + } -/***************************************************************************** - * - * The following functions belong to the exported interface of AMD IOMMU - * - * This interface allows access to lower level functions of the IOMMU - * like protection domain handling and assignement of devices to domains - * which is not possible with the dma_ops interface. - * - *****************************************************************************/ + list_for_each_entry(dev_data, &pdom->dev_list, list) + device_flush_dte(dev_data); -static void cleanup_domain(struct protection_domain *domain) + domain_flush_complete(pdom); +} + +/* + * amd_iommu_iotlb_sync_map() is used to generate flushes for non-present to + * present (ie mapping) operations. It is a NOP if the IOMMU doesn't have non + * present caching (like hypervisor shadowing). + */ +static int amd_iommu_iotlb_sync_map(struct iommu_domain *dom, + unsigned long iova, size_t size) { - struct iommu_dev_data *entry; + struct protection_domain *domain = to_pdomain(dom); unsigned long flags; - spin_lock_irqsave(&domain->lock, flags); - - while (!list_empty(&domain->dev_list)) { - entry = list_first_entry(&domain->dev_list, - struct iommu_dev_data, list); - BUG_ON(!entry->domain); - do_detach(entry); - } + if (likely(!amd_iommu_np_cache)) + return 0; + spin_lock_irqsave(&domain->lock, flags); + amd_iommu_domain_flush_pages(domain, iova, size); spin_unlock_irqrestore(&domain->lock, flags); + return 0; } -static void protection_domain_free(struct protection_domain *domain) +static void amd_iommu_flush_iotlb_all(struct iommu_domain *domain) { - struct domain_pgtable pgtable; - - if (!domain) - return; - - if (domain->id) - domain_id_free(domain->id); - - amd_iommu_domain_get_pgtable(domain, &pgtable); - atomic64_set(&domain->pt_root, 0); - free_pagetable(&pgtable); + struct protection_domain *dom = to_pdomain(domain); + unsigned long flags; - kfree(domain); + spin_lock_irqsave(&dom->lock, flags); + amd_iommu_domain_flush_all(dom); + spin_unlock_irqrestore(&dom->lock, flags); } -static int protection_domain_init(struct protection_domain *domain, int mode) +static void amd_iommu_iotlb_sync(struct iommu_domain *domain, + struct iommu_iotlb_gather *gather) { - u64 *pt_root = NULL, root; - - BUG_ON(mode < PAGE_MODE_NONE || mode > PAGE_MODE_6_LEVEL); + struct protection_domain *dom = to_pdomain(domain); + unsigned long flags; - spin_lock_init(&domain->lock); - domain->id = domain_id_alloc(); - if (!domain->id) - return -ENOMEM; - INIT_LIST_HEAD(&domain->dev_list); + spin_lock_irqsave(&dom->lock, flags); + amd_iommu_domain_flush_pages(dom, gather->start, + gather->end - gather->start + 1); + spin_unlock_irqrestore(&dom->lock, flags); + iommu_put_pages_list(&gather->freelist); +} - if (mode != PAGE_MODE_NONE) { - pt_root = (void *)get_zeroed_page(GFP_KERNEL); - if (!pt_root) - return -ENOMEM; - } +static const struct pt_iommu_driver_ops amd_hw_driver_ops_v1 = { + .get_top_lock = amd_iommu_get_top_lock, + .change_top = amd_iommu_change_top, +}; - root = amd_iommu_domain_encode_pgtable(pt_root, mode); - atomic64_set(&domain->pt_root, root); +static const struct iommu_domain_ops amdv1_ops = { + IOMMU_PT_DOMAIN_OPS(amdv1), + .iotlb_sync_map = amd_iommu_iotlb_sync_map, + .flush_iotlb_all = amd_iommu_flush_iotlb_all, + .iotlb_sync = amd_iommu_iotlb_sync, + .attach_dev = amd_iommu_attach_device, + .free = amd_iommu_domain_free, + .enforce_cache_coherency = amd_iommu_enforce_cache_coherency, +}; - return 0; -} +static const struct iommu_dirty_ops amdv1_dirty_ops = { + IOMMU_PT_DIRTY_OPS(amdv1), + .set_dirty_tracking = amd_iommu_set_dirty_tracking, +}; -static struct protection_domain *protection_domain_alloc(int mode) +static struct iommu_domain *amd_iommu_domain_alloc_paging_v1(struct device *dev, + u32 flags) { + struct pt_iommu_amdv1_cfg cfg = {}; struct protection_domain *domain; + int ret; - domain = kzalloc(sizeof(*domain), GFP_KERNEL); + if (amd_iommu_hatdis) + return ERR_PTR(-EOPNOTSUPP); + + domain = protection_domain_alloc(); if (!domain) - return NULL; + return ERR_PTR(-ENOMEM); - if (protection_domain_init(domain, mode)) - goto out_err; + domain->pd_mode = PD_MODE_V1; + domain->iommu.driver_ops = &amd_hw_driver_ops_v1; + domain->iommu.nid = dev_to_node(dev); + if (flags & IOMMU_HWPT_ALLOC_DIRTY_TRACKING) + domain->domain.dirty_ops = &amdv1_dirty_ops; - return domain; + /* + * Someday FORCE_COHERENCE should be set by + * amd_iommu_enforce_cache_coherency() like VT-d does. + */ + cfg.common.features = BIT(PT_FEAT_DYNAMIC_TOP) | + BIT(PT_FEAT_AMDV1_ENCRYPT_TABLES) | + BIT(PT_FEAT_AMDV1_FORCE_COHERENCE); -out_err: - kfree(domain); + /* + * AMD's IOMMU can flush as many pages as necessary in a single flush. + * Unless we run in a virtual machine, which can be inferred according + * to whether "non-present cache" is on, it is probably best to prefer + * (potentially) too extensive TLB flushing (i.e., more misses) over + * multiple TLB flushes (i.e., more flushes). For virtual machines the + * hypervisor needs to synchronize the host IOMMU PTEs with those of + * the guest, and the trade-off is different: unnecessary TLB flushes + * should be avoided. + */ + if (amd_iommu_np_cache) + cfg.common.features |= BIT(PT_FEAT_FLUSH_RANGE_NO_GAPS); + else + cfg.common.features |= BIT(PT_FEAT_FLUSH_RANGE); - return NULL; + cfg.common.hw_max_vasz_lg2 = + min(64, (amd_iommu_hpt_level - 1) * 9 + 21); + cfg.common.hw_max_oasz_lg2 = 52; + cfg.starting_level = 2; + domain->domain.ops = &amdv1_ops; + + ret = pt_iommu_amdv1_init(&domain->amdv1, &cfg, GFP_KERNEL); + if (ret) { + amd_iommu_domain_free(&domain->domain); + return ERR_PTR(ret); + } + + /* + * Narrow the supported page sizes to those selected by the kernel + * command line. + */ + domain->domain.pgsize_bitmap &= amd_iommu_pgsize_bitmap; + return &domain->domain; } -static struct iommu_domain *amd_iommu_domain_alloc(unsigned type) +static const struct iommu_domain_ops amdv2_ops = { + IOMMU_PT_DOMAIN_OPS(x86_64), + .iotlb_sync_map = amd_iommu_iotlb_sync_map, + .flush_iotlb_all = amd_iommu_flush_iotlb_all, + .iotlb_sync = amd_iommu_iotlb_sync, + .attach_dev = amd_iommu_attach_device, + .free = amd_iommu_domain_free, + /* + * Note the AMDv2 page table format does not support a Force Coherency + * bit, so enforce_cache_coherency should not be set. However VFIO is + * not prepared to handle a case where some domains will support + * enforcement and others do not. VFIO and iommufd will have to be fixed + * before it can fully use the V2 page table. See the comment in + * iommufd_hwpt_paging_alloc(). For now leave things as they have + * historically been and lie about enforce_cache_coherencey. + */ + .enforce_cache_coherency = amd_iommu_enforce_cache_coherency, +}; + +static struct iommu_domain *amd_iommu_domain_alloc_paging_v2(struct device *dev, + u32 flags) { + struct pt_iommu_x86_64_cfg cfg = {}; struct protection_domain *domain; - int mode = DEFAULT_PGTABLE_LEVEL; + int ret; - if (type == IOMMU_DOMAIN_IDENTITY) - mode = PAGE_MODE_NONE; + if (!amd_iommu_v2_pgtbl_supported()) + return ERR_PTR(-EOPNOTSUPP); - domain = protection_domain_alloc(mode); + domain = protection_domain_alloc(); if (!domain) - return NULL; - - domain->domain.geometry.aperture_start = 0; - domain->domain.geometry.aperture_end = ~0ULL; - domain->domain.geometry.force_aperture = true; + return ERR_PTR(-ENOMEM); - if (type == IOMMU_DOMAIN_DMA && - iommu_get_dma_cookie(&domain->domain) == -ENOMEM) - goto free_domain; + domain->pd_mode = PD_MODE_V2; + domain->iommu.nid = dev_to_node(dev); - return &domain->domain; + cfg.common.features = BIT(PT_FEAT_X86_64_AMD_ENCRYPT_TABLES); + if (amd_iommu_np_cache) + cfg.common.features |= BIT(PT_FEAT_FLUSH_RANGE_NO_GAPS); + else + cfg.common.features |= BIT(PT_FEAT_FLUSH_RANGE); -free_domain: - protection_domain_free(domain); + /* + * The v2 table behaves differently if it is attached to PASID 0 vs a + * non-zero PASID. On PASID 0 it has no sign extension and the full + * 57/48 bits decode the lower addresses. Otherwise it behaves like a + * normal sign extended x86 page table. Since we want the domain to work + * in both modes the top bit is removed and PT_FEAT_SIGN_EXTEND is not + * set which creates a table that is compatible in both modes. + */ + if (amd_iommu_gpt_level == PAGE_MODE_5_LEVEL) { + cfg.common.hw_max_vasz_lg2 = 56; + cfg.top_level = 4; + } else { + cfg.common.hw_max_vasz_lg2 = 47; + cfg.top_level = 3; + } + cfg.common.hw_max_oasz_lg2 = 52; + domain->domain.ops = &amdv2_ops; - return NULL; + ret = pt_iommu_x86_64_init(&domain->amdv2, &cfg, GFP_KERNEL); + if (ret) { + amd_iommu_domain_free(&domain->domain); + return ERR_PTR(ret); + } + return &domain->domain; } -static void amd_iommu_domain_free(struct iommu_domain *dom) +static struct iommu_domain * +amd_iommu_domain_alloc_paging_flags(struct device *dev, u32 flags, + const struct iommu_user_data *user_data) + { - struct protection_domain *domain; + struct amd_iommu *iommu = get_amd_iommu_from_dev(dev); + const u32 supported_flags = IOMMU_HWPT_ALLOC_DIRTY_TRACKING | + IOMMU_HWPT_ALLOC_PASID; + + if ((flags & ~supported_flags) || user_data) + return ERR_PTR(-EOPNOTSUPP); - domain = to_pdomain(dom); + switch (flags & supported_flags) { + case IOMMU_HWPT_ALLOC_DIRTY_TRACKING: + /* Allocate domain with v1 page table for dirty tracking */ + if (!amd_iommu_hd_support(iommu)) + break; + return amd_iommu_domain_alloc_paging_v1(dev, flags); + case IOMMU_HWPT_ALLOC_PASID: + /* Allocate domain with v2 page table if IOMMU supports PASID. */ + if (!amd_iommu_pasid_supported()) + break; + return amd_iommu_domain_alloc_paging_v2(dev, flags); + case 0: { + struct iommu_domain *ret; + + /* If nothing specific is required use the kernel commandline default */ + if (amd_iommu_pgtable == PD_MODE_V1) { + ret = amd_iommu_domain_alloc_paging_v1(dev, flags); + if (ret != ERR_PTR(-EOPNOTSUPP)) + return ret; + return amd_iommu_domain_alloc_paging_v2(dev, flags); + } + ret = amd_iommu_domain_alloc_paging_v2(dev, flags); + if (ret != ERR_PTR(-EOPNOTSUPP)) + return ret; + return amd_iommu_domain_alloc_paging_v1(dev, flags); + } + default: + break; + } + return ERR_PTR(-EOPNOTSUPP); +} - if (domain->dev_cnt > 0) - cleanup_domain(domain); +void amd_iommu_domain_free(struct iommu_domain *dom) +{ + struct protection_domain *domain = to_pdomain(dom); - BUG_ON(domain->dev_cnt != 0); + WARN_ON(!list_empty(&domain->dev_list)); + pt_iommu_deinit(&domain->iommu); + pdom_id_free(domain->id); + kfree(domain); +} - if (!dom) - return; +static int blocked_domain_attach_device(struct iommu_domain *domain, + struct device *dev, + struct iommu_domain *old) +{ + struct iommu_dev_data *dev_data = dev_iommu_priv_get(dev); - if (dom->type == IOMMU_DOMAIN_DMA) - iommu_put_dma_cookie(&domain->domain); + if (dev_data->domain) + detach_device(dev); - if (domain->flags & PD_IOMMUV2_MASK) - free_gcr3_table(domain); + /* Clear DTE and flush the entry */ + mutex_lock(&dev_data->mutex); + dev_update_dte(dev_data, false); + mutex_unlock(&dev_data->mutex); - protection_domain_free(domain); + return 0; } -static void amd_iommu_detach_device(struct iommu_domain *dom, - struct device *dev) +static int blocked_domain_set_dev_pasid(struct iommu_domain *domain, + struct device *dev, ioasid_t pasid, + struct iommu_domain *old) { - struct iommu_dev_data *dev_data = dev_iommu_priv_get(dev); - struct amd_iommu *iommu; - int devid; + amd_iommu_remove_dev_pasid(dev, pasid, old); + return 0; +} - if (!check_device(dev)) - return; +static struct iommu_domain blocked_domain = { + .type = IOMMU_DOMAIN_BLOCKED, + .ops = &(const struct iommu_domain_ops) { + .attach_dev = blocked_domain_attach_device, + .set_dev_pasid = blocked_domain_set_dev_pasid, + } +}; - devid = get_device_id(dev); - if (devid < 0) - return; +static struct protection_domain identity_domain; - if (dev_data->domain != NULL) - detach_device(dev); +static const struct iommu_domain_ops identity_domain_ops = { + .attach_dev = amd_iommu_attach_device, +}; - iommu = amd_iommu_rlookup_table[devid]; - if (!iommu) - return; +void amd_iommu_init_identity_domain(void) +{ + struct iommu_domain *domain = &identity_domain.domain; -#ifdef CONFIG_IRQ_REMAP - if (AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir) && - (dom->type == IOMMU_DOMAIN_UNMANAGED)) - dev_data->use_vapic = 0; -#endif + domain->type = IOMMU_DOMAIN_IDENTITY; + domain->ops = &identity_domain_ops; + domain->owner = &amd_iommu_ops; - iommu_completion_wait(iommu); + identity_domain.id = pdom_id_alloc(); + + protection_domain_init(&identity_domain); } -static int amd_iommu_attach_device(struct iommu_domain *dom, - struct device *dev) +static int amd_iommu_attach_device(struct iommu_domain *dom, struct device *dev, + struct iommu_domain *old) { + struct iommu_dev_data *dev_data = dev_iommu_priv_get(dev); struct protection_domain *domain = to_pdomain(dom); - struct iommu_dev_data *dev_data; - struct amd_iommu *iommu; + struct amd_iommu *iommu = get_amd_iommu_from_dev(dev); int ret; - if (!check_device(dev)) - return -EINVAL; + /* + * Skip attach device to domain if new domain is same as + * devices current domain + */ + if (dev_data->domain == domain) + return 0; - dev_data = dev_iommu_priv_get(dev); dev_data->defer_attach = false; - iommu = amd_iommu_rlookup_table[dev_data->devid]; - if (!iommu) + /* + * Restrict to devices with compatible IOMMU hardware support + * when enforcement of dirty tracking is enabled. + */ + if (dom->dirty_ops && !amd_iommu_hd_support(iommu)) return -EINVAL; if (dev_data->domain) @@ -2481,87 +2896,73 @@ static int amd_iommu_attach_device(struct iommu_domain *dom, } #endif - iommu_completion_wait(iommu); - return ret; } -static int amd_iommu_map(struct iommu_domain *dom, unsigned long iova, - phys_addr_t paddr, size_t page_size, int iommu_prot, - gfp_t gfp) +static bool amd_iommu_capable(struct device *dev, enum iommu_cap cap) { - struct protection_domain *domain = to_pdomain(dom); - struct domain_pgtable pgtable; - int prot = 0; - int ret; - - amd_iommu_domain_get_pgtable(domain, &pgtable); - if (pgtable.mode == PAGE_MODE_NONE) - return -EINVAL; - - if (iommu_prot & IOMMU_READ) - prot |= IOMMU_PROT_IR; - if (iommu_prot & IOMMU_WRITE) - prot |= IOMMU_PROT_IW; - - ret = iommu_map_page(domain, iova, paddr, page_size, prot, gfp); + switch (cap) { + case IOMMU_CAP_CACHE_COHERENCY: + return true; + case IOMMU_CAP_NOEXEC: + return false; + case IOMMU_CAP_PRE_BOOT_PROTECTION: + return amdr_ivrs_remap_support; + case IOMMU_CAP_ENFORCE_CACHE_COHERENCY: + return true; + case IOMMU_CAP_DEFERRED_FLUSH: + return true; + case IOMMU_CAP_DIRTY_TRACKING: { + struct amd_iommu *iommu = get_amd_iommu_from_dev(dev); - domain_flush_np_cache(domain, iova, page_size); + return amd_iommu_hd_support(iommu); + } + default: + break; + } - return ret; + return false; } -static size_t amd_iommu_unmap(struct iommu_domain *dom, unsigned long iova, - size_t page_size, - struct iommu_iotlb_gather *gather) +static int amd_iommu_set_dirty_tracking(struct iommu_domain *domain, + bool enable) { - struct protection_domain *domain = to_pdomain(dom); - struct domain_pgtable pgtable; + struct protection_domain *pdomain = to_pdomain(domain); + struct dev_table_entry *dte; + struct iommu_dev_data *dev_data; + bool domain_flush = false; + struct amd_iommu *iommu; + unsigned long flags; + u64 new; - amd_iommu_domain_get_pgtable(domain, &pgtable); - if (pgtable.mode == PAGE_MODE_NONE) + spin_lock_irqsave(&pdomain->lock, flags); + if (!(pdomain->dirty_tracking ^ enable)) { + spin_unlock_irqrestore(&pdomain->lock, flags); return 0; + } - return iommu_unmap_page(domain, iova, page_size); -} - -static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom, - dma_addr_t iova) -{ - struct protection_domain *domain = to_pdomain(dom); - unsigned long offset_mask, pte_pgsize; - struct domain_pgtable pgtable; - u64 *pte, __pte; - - amd_iommu_domain_get_pgtable(domain, &pgtable); - if (pgtable.mode == PAGE_MODE_NONE) - return iova; - - pte = fetch_pte(domain, iova, &pte_pgsize); + list_for_each_entry(dev_data, &pdomain->dev_list, list) { + spin_lock(&dev_data->dte_lock); + iommu = get_amd_iommu_from_dev_data(dev_data); + dte = &get_dev_table(iommu)[dev_data->devid]; + new = dte->data[0]; + new = (enable ? new | DTE_FLAG_HAD : new & ~DTE_FLAG_HAD); + dte->data[0] = new; + spin_unlock(&dev_data->dte_lock); - if (!pte || !IOMMU_PTE_PRESENT(*pte)) - return 0; + /* Flush device DTE */ + device_flush_dte(dev_data); + domain_flush = true; + } - offset_mask = pte_pgsize - 1; - __pte = __sme_clr(*pte & PM_ADDR_MASK); + /* Flush IOTLB to mark IOPTE dirty on the next translation(s) */ + if (domain_flush) + amd_iommu_domain_flush_all(pdomain); - return (__pte & ~offset_mask) | (iova & offset_mask); -} + pdomain->dirty_tracking = enable; + spin_unlock_irqrestore(&pdomain->lock, flags); -static bool amd_iommu_capable(enum iommu_cap cap) -{ - switch (cap) { - case IOMMU_CAP_CACHE_COHERENCY: - return true; - case IOMMU_CAP_INTR_REMAP: - return (irq_remapping_enabled == 1); - case IOMMU_CAP_NOEXEC: - return false; - default: - break; - } - - return false; + return 0; } static void amd_iommu_get_resv_regions(struct device *dev, @@ -2569,13 +2970,19 @@ static void amd_iommu_get_resv_regions(struct device *dev, { struct iommu_resv_region *region; struct unity_map_entry *entry; - int devid; + struct amd_iommu *iommu; + struct amd_iommu_pci_seg *pci_seg; + int devid, sbdf; - devid = get_device_id(dev); - if (devid < 0) + sbdf = get_device_sbdf_id(dev); + if (sbdf < 0) return; - list_for_each_entry(entry, &amd_iommu_unity_map, list) { + devid = PCI_SBDF_TO_DEVID(sbdf); + iommu = get_amd_iommu_from_dev(dev); + pci_seg = iommu->pci_seg; + + list_for_each_entry(entry, &pci_seg->unity_map, list) { int type, prot = 0; size_t length; @@ -2593,7 +3000,8 @@ static void amd_iommu_get_resv_regions(struct device *dev, type = IOMMU_RESV_RESERVED; region = iommu_alloc_resv_region(entry->address_start, - length, prot, type); + length, prot, type, + GFP_KERNEL); if (!region) { dev_err(dev, "Out of memory allocating dm-regions\n"); return; @@ -2603,44 +3011,28 @@ static void amd_iommu_get_resv_regions(struct device *dev, region = iommu_alloc_resv_region(MSI_RANGE_START, MSI_RANGE_END - MSI_RANGE_START + 1, - 0, IOMMU_RESV_MSI); + 0, IOMMU_RESV_MSI, GFP_KERNEL); if (!region) return; list_add_tail(®ion->list, head); + if (amd_iommu_ht_range_ignore()) + return; + region = iommu_alloc_resv_region(HT_RANGE_START, HT_RANGE_END - HT_RANGE_START + 1, - 0, IOMMU_RESV_RESERVED); + 0, IOMMU_RESV_RESERVED, GFP_KERNEL); if (!region) return; list_add_tail(®ion->list, head); } -bool amd_iommu_is_attach_deferred(struct iommu_domain *domain, - struct device *dev) +static bool amd_iommu_is_attach_deferred(struct device *dev) { struct iommu_dev_data *dev_data = dev_iommu_priv_get(dev); return dev_data->defer_attach; } -EXPORT_SYMBOL_GPL(amd_iommu_is_attach_deferred); - -static void amd_iommu_flush_iotlb_all(struct iommu_domain *domain) -{ - struct protection_domain *dom = to_pdomain(domain); - unsigned long flags; - - spin_lock_irqsave(&dom->lock, flags); - domain_flush_tlb_pde(dom); - domain_flush_complete(dom); - spin_unlock_irqrestore(&dom->lock, flags); -} - -static void amd_iommu_iotlb_sync(struct iommu_domain *domain, - struct iommu_iotlb_gather *gather) -{ - amd_iommu_flush_iotlb_all(domain); -} static int amd_iommu_def_domain_type(struct device *dev) { @@ -2650,474 +3042,133 @@ static int amd_iommu_def_domain_type(struct device *dev) if (!dev_data) return 0; - if (dev_data->iommu_v2) + /* Always use DMA domain for untrusted device */ + if (dev_is_pci(dev) && to_pci_dev(dev)->untrusted) + return IOMMU_DOMAIN_DMA; + + /* + * Do not identity map IOMMUv2 capable devices when: + * - memory encryption is active, because some of those devices + * (AMD GPUs) don't have the encryption bit in their DMA-mask + * and require remapping. + * - SNP is enabled, because it prohibits DTE[Mode]=0. + */ + if (pdev_pasid_supported(dev_data) && + !cc_platform_has(CC_ATTR_MEM_ENCRYPT) && + !amd_iommu_snp_en) { return IOMMU_DOMAIN_IDENTITY; + } return 0; } +static bool amd_iommu_enforce_cache_coherency(struct iommu_domain *domain) +{ + /* IOMMU_PTE_FC is always set */ + return true; +} + const struct iommu_ops amd_iommu_ops = { .capable = amd_iommu_capable, - .domain_alloc = amd_iommu_domain_alloc, - .domain_free = amd_iommu_domain_free, - .attach_dev = amd_iommu_attach_device, - .detach_dev = amd_iommu_detach_device, - .map = amd_iommu_map, - .unmap = amd_iommu_unmap, - .iova_to_phys = amd_iommu_iova_to_phys, + .blocked_domain = &blocked_domain, + .release_domain = &blocked_domain, + .identity_domain = &identity_domain.domain, + .domain_alloc_paging_flags = amd_iommu_domain_alloc_paging_flags, + .domain_alloc_sva = amd_iommu_domain_alloc_sva, .probe_device = amd_iommu_probe_device, .release_device = amd_iommu_release_device, - .probe_finalize = amd_iommu_probe_finalize, .device_group = amd_iommu_device_group, - .domain_get_attr = amd_iommu_domain_get_attr, .get_resv_regions = amd_iommu_get_resv_regions, - .put_resv_regions = generic_iommu_put_resv_regions, .is_attach_deferred = amd_iommu_is_attach_deferred, - .pgsize_bitmap = AMD_IOMMU_PGSIZES, - .flush_iotlb_all = amd_iommu_flush_iotlb_all, - .iotlb_sync = amd_iommu_iotlb_sync, .def_domain_type = amd_iommu_def_domain_type, + .page_response = amd_iommu_page_response, }; +#ifdef CONFIG_IRQ_REMAP + /***************************************************************************** * - * The next functions do a basic initialization of IOMMU for pass through - * mode - * - * In passthrough mode the IOMMU is initialized and enabled but not used for - * DMA-API translation. + * Interrupt Remapping Implementation * *****************************************************************************/ -/* IOMMUv2 specific functions */ -int amd_iommu_register_ppr_notifier(struct notifier_block *nb) -{ - return atomic_notifier_chain_register(&ppr_notifier, nb); -} -EXPORT_SYMBOL(amd_iommu_register_ppr_notifier); - -int amd_iommu_unregister_ppr_notifier(struct notifier_block *nb) -{ - return atomic_notifier_chain_unregister(&ppr_notifier, nb); -} -EXPORT_SYMBOL(amd_iommu_unregister_ppr_notifier); - -void amd_iommu_domain_direct_map(struct iommu_domain *dom) -{ - struct protection_domain *domain = to_pdomain(dom); - struct domain_pgtable pgtable; - unsigned long flags; - - spin_lock_irqsave(&domain->lock, flags); - - /* First save pgtable configuration*/ - amd_iommu_domain_get_pgtable(domain, &pgtable); - - /* Update data structure */ - atomic64_set(&domain->pt_root, 0); - - /* Make changes visible to IOMMUs */ - update_domain(domain); - - /* Page-table is not visible to IOMMU anymore, so free it */ - free_pagetable(&pgtable); - - spin_unlock_irqrestore(&domain->lock, flags); -} -EXPORT_SYMBOL(amd_iommu_domain_direct_map); +static struct irq_chip amd_ir_chip; +static DEFINE_SPINLOCK(iommu_table_lock); -int amd_iommu_domain_enable_v2(struct iommu_domain *dom, int pasids) +static void iommu_flush_irt_and_complete(struct amd_iommu *iommu, u16 devid) { - struct protection_domain *domain = to_pdomain(dom); + int ret; + u64 data; unsigned long flags; - int levels, ret; - - if (pasids <= 0 || pasids > (PASID_MASK + 1)) - return -EINVAL; + struct iommu_cmd cmd, cmd2; - /* Number of GCR3 table levels required */ - for (levels = 0; (pasids - 1) & ~0x1ff; pasids >>= 9) - levels += 1; - - if (levels > amd_iommu_max_glx_val) - return -EINVAL; + if (iommu->irtcachedis_enabled) + return; - spin_lock_irqsave(&domain->lock, flags); + build_inv_irt(&cmd, devid); + data = atomic64_inc_return(&iommu->cmd_sem_val); + build_completion_wait(&cmd2, iommu, data); - /* - * Save us all sanity checks whether devices already in the - * domain support IOMMUv2. Just force that the domain has no - * devices attached when it is switched into IOMMUv2 mode. - */ - ret = -EBUSY; - if (domain->dev_cnt > 0 || domain->flags & PD_IOMMUV2_MASK) + raw_spin_lock_irqsave(&iommu->lock, flags); + ret = __iommu_queue_command_sync(iommu, &cmd, true); + if (ret) goto out; - - ret = -ENOMEM; - domain->gcr3_tbl = (void *)get_zeroed_page(GFP_ATOMIC); - if (domain->gcr3_tbl == NULL) + ret = __iommu_queue_command_sync(iommu, &cmd2, false); + if (ret) goto out; - - domain->glx = levels; - domain->flags |= PD_IOMMUV2_MASK; - - update_domain(domain); - - ret = 0; - -out: - spin_unlock_irqrestore(&domain->lock, flags); - - return ret; -} -EXPORT_SYMBOL(amd_iommu_domain_enable_v2); - -static int __flush_pasid(struct protection_domain *domain, int pasid, - u64 address, bool size) -{ - struct iommu_dev_data *dev_data; - struct iommu_cmd cmd; - int i, ret; - - if (!(domain->flags & PD_IOMMUV2_MASK)) - return -EINVAL; - - build_inv_iommu_pasid(&cmd, domain->id, pasid, address, size); - - /* - * IOMMU TLB needs to be flushed before Device TLB to - * prevent device TLB refill from IOMMU TLB - */ - for (i = 0; i < amd_iommu_get_num_iommus(); ++i) { - if (domain->dev_iommu[i] == 0) - continue; - - ret = iommu_queue_command(amd_iommus[i], &cmd); - if (ret != 0) - goto out; - } - - /* Wait until IOMMU TLB flushes are complete */ - domain_flush_complete(domain); - - /* Now flush device TLBs */ - list_for_each_entry(dev_data, &domain->dev_list, list) { - struct amd_iommu *iommu; - int qdep; - - /* - There might be non-IOMMUv2 capable devices in an IOMMUv2 - * domain. - */ - if (!dev_data->ats.enabled) - continue; - - qdep = dev_data->ats.qdep; - iommu = amd_iommu_rlookup_table[dev_data->devid]; - - build_inv_iotlb_pasid(&cmd, dev_data->devid, pasid, - qdep, address, size); - - ret = iommu_queue_command(iommu, &cmd); - if (ret != 0) - goto out; - } - - /* Wait until all device TLBs are flushed */ - domain_flush_complete(domain); - - ret = 0; - + wait_on_sem(iommu, data); out: - - return ret; -} - -static int __amd_iommu_flush_page(struct protection_domain *domain, int pasid, - u64 address) -{ - return __flush_pasid(domain, pasid, address, false); -} - -int amd_iommu_flush_page(struct iommu_domain *dom, int pasid, - u64 address) -{ - struct protection_domain *domain = to_pdomain(dom); - unsigned long flags; - int ret; - - spin_lock_irqsave(&domain->lock, flags); - ret = __amd_iommu_flush_page(domain, pasid, address); - spin_unlock_irqrestore(&domain->lock, flags); - - return ret; -} -EXPORT_SYMBOL(amd_iommu_flush_page); - -static int __amd_iommu_flush_tlb(struct protection_domain *domain, int pasid) -{ - return __flush_pasid(domain, pasid, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, - true); -} - -int amd_iommu_flush_tlb(struct iommu_domain *dom, int pasid) -{ - struct protection_domain *domain = to_pdomain(dom); - unsigned long flags; - int ret; - - spin_lock_irqsave(&domain->lock, flags); - ret = __amd_iommu_flush_tlb(domain, pasid); - spin_unlock_irqrestore(&domain->lock, flags); - - return ret; -} -EXPORT_SYMBOL(amd_iommu_flush_tlb); - -static u64 *__get_gcr3_pte(u64 *root, int level, int pasid, bool alloc) -{ - int index; - u64 *pte; - - while (true) { - - index = (pasid >> (9 * level)) & 0x1ff; - pte = &root[index]; - - if (level == 0) - break; - - if (!(*pte & GCR3_VALID)) { - if (!alloc) - return NULL; - - root = (void *)get_zeroed_page(GFP_ATOMIC); - if (root == NULL) - return NULL; - - *pte = iommu_virt_to_phys(root) | GCR3_VALID; - } - - root = iommu_phys_to_virt(*pte & PAGE_MASK); - - level -= 1; - } - - return pte; -} - -static int __set_gcr3(struct protection_domain *domain, int pasid, - unsigned long cr3) -{ - struct domain_pgtable pgtable; - u64 *pte; - - amd_iommu_domain_get_pgtable(domain, &pgtable); - if (pgtable.mode != PAGE_MODE_NONE) - return -EINVAL; - - pte = __get_gcr3_pte(domain->gcr3_tbl, domain->glx, pasid, true); - if (pte == NULL) - return -ENOMEM; - - *pte = (cr3 & PAGE_MASK) | GCR3_VALID; - - return __amd_iommu_flush_tlb(domain, pasid); -} - -static int __clear_gcr3(struct protection_domain *domain, int pasid) -{ - struct domain_pgtable pgtable; - u64 *pte; - - amd_iommu_domain_get_pgtable(domain, &pgtable); - if (pgtable.mode != PAGE_MODE_NONE) - return -EINVAL; - - pte = __get_gcr3_pte(domain->gcr3_tbl, domain->glx, pasid, false); - if (pte == NULL) - return 0; - - *pte = 0; - - return __amd_iommu_flush_tlb(domain, pasid); -} - -int amd_iommu_domain_set_gcr3(struct iommu_domain *dom, int pasid, - unsigned long cr3) -{ - struct protection_domain *domain = to_pdomain(dom); - unsigned long flags; - int ret; - - spin_lock_irqsave(&domain->lock, flags); - ret = __set_gcr3(domain, pasid, cr3); - spin_unlock_irqrestore(&domain->lock, flags); - - return ret; -} -EXPORT_SYMBOL(amd_iommu_domain_set_gcr3); - -int amd_iommu_domain_clear_gcr3(struct iommu_domain *dom, int pasid) -{ - struct protection_domain *domain = to_pdomain(dom); - unsigned long flags; - int ret; - - spin_lock_irqsave(&domain->lock, flags); - ret = __clear_gcr3(domain, pasid); - spin_unlock_irqrestore(&domain->lock, flags); - - return ret; -} -EXPORT_SYMBOL(amd_iommu_domain_clear_gcr3); - -int amd_iommu_complete_ppr(struct pci_dev *pdev, int pasid, - int status, int tag) -{ - struct iommu_dev_data *dev_data; - struct amd_iommu *iommu; - struct iommu_cmd cmd; - - dev_data = dev_iommu_priv_get(&pdev->dev); - iommu = amd_iommu_rlookup_table[dev_data->devid]; - - build_complete_ppr(&cmd, dev_data->devid, pasid, status, - tag, dev_data->pri_tlp); - - return iommu_queue_command(iommu, &cmd); + raw_spin_unlock_irqrestore(&iommu->lock, flags); } -EXPORT_SYMBOL(amd_iommu_complete_ppr); -struct iommu_domain *amd_iommu_get_v2_domain(struct pci_dev *pdev) +static inline u8 iommu_get_int_tablen(struct iommu_dev_data *dev_data) { - struct protection_domain *pdomain; - struct iommu_dev_data *dev_data; - struct device *dev = &pdev->dev; - struct iommu_domain *io_domain; - - if (!check_device(dev)) - return NULL; - - dev_data = dev_iommu_priv_get(&pdev->dev); - pdomain = dev_data->domain; - io_domain = iommu_get_domain_for_dev(dev); - - if (pdomain == NULL && dev_data->defer_attach) { - dev_data->defer_attach = false; - pdomain = to_pdomain(io_domain); - attach_device(dev, pdomain); - } - - if (pdomain == NULL) - return NULL; - - if (io_domain->type != IOMMU_DOMAIN_DMA) - return NULL; - - /* Only return IOMMUv2 domains */ - if (!(pdomain->flags & PD_IOMMUV2_MASK)) - return NULL; - - return &pdomain->domain; + if (dev_data && dev_data->max_irqs == MAX_IRQS_PER_TABLE_2K) + return DTE_INTTABLEN_2K; + return DTE_INTTABLEN_512; } -EXPORT_SYMBOL(amd_iommu_get_v2_domain); -void amd_iommu_enable_device_erratum(struct pci_dev *pdev, u32 erratum) +static void set_dte_irq_entry(struct amd_iommu *iommu, u16 devid, + struct irq_remap_table *table) { - struct iommu_dev_data *dev_data; - - if (!amd_iommu_v2_supported()) - return; - - dev_data = dev_iommu_priv_get(&pdev->dev); - dev_data->errata |= (1 << erratum); -} -EXPORT_SYMBOL(amd_iommu_enable_device_erratum); - -int amd_iommu_device_info(struct pci_dev *pdev, - struct amd_iommu_device_info *info) -{ - int max_pasids; - int pos; - - if (pdev == NULL || info == NULL) - return -EINVAL; - - if (!amd_iommu_v2_supported()) - return -EINVAL; - - memset(info, 0, sizeof(*info)); - - if (pci_ats_supported(pdev)) - info->flags |= AMD_IOMMU_DEVICE_FLAG_ATS_SUP; - - pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_PRI); - if (pos) - info->flags |= AMD_IOMMU_DEVICE_FLAG_PRI_SUP; - - pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_PASID); - if (pos) { - int features; - - max_pasids = 1 << (9 * (amd_iommu_max_glx_val + 1)); - max_pasids = min(max_pasids, (1 << 20)); - - info->flags |= AMD_IOMMU_DEVICE_FLAG_PASID_SUP; - info->max_pasids = min(pci_max_pasids(pdev), max_pasids); - - features = pci_pasid_features(pdev); - if (features & PCI_PASID_CAP_EXEC) - info->flags |= AMD_IOMMU_DEVICE_FLAG_EXEC_SUP; - if (features & PCI_PASID_CAP_PRIV) - info->flags |= AMD_IOMMU_DEVICE_FLAG_PRIV_SUP; - } - - return 0; -} -EXPORT_SYMBOL(amd_iommu_device_info); - -#ifdef CONFIG_IRQ_REMAP - -/***************************************************************************** - * - * Interrupt Remapping Implementation - * - *****************************************************************************/ + u64 new; + struct dev_table_entry *dte = &get_dev_table(iommu)[devid]; + struct iommu_dev_data *dev_data = search_dev_data(iommu, devid); -static struct irq_chip amd_ir_chip; -static DEFINE_SPINLOCK(iommu_table_lock); + if (dev_data) + spin_lock(&dev_data->dte_lock); -static void set_dte_irq_entry(u16 devid, struct irq_remap_table *table) -{ - u64 dte; - - dte = amd_iommu_dev_table[devid].data[2]; - dte &= ~DTE_IRQ_PHYS_ADDR_MASK; - dte |= iommu_virt_to_phys(table->table); - dte |= DTE_IRQ_REMAP_INTCTL; - dte |= DTE_IRQ_TABLE_LEN; - dte |= DTE_IRQ_REMAP_ENABLE; + new = READ_ONCE(dte->data[2]); + new &= ~DTE_IRQ_PHYS_ADDR_MASK; + new |= iommu_virt_to_phys(table->table); + new |= DTE_IRQ_REMAP_INTCTL; + new |= iommu_get_int_tablen(dev_data); + new |= DTE_IRQ_REMAP_ENABLE; + WRITE_ONCE(dte->data[2], new); - amd_iommu_dev_table[devid].data[2] = dte; + if (dev_data) + spin_unlock(&dev_data->dte_lock); } -static struct irq_remap_table *get_irq_table(u16 devid) +static struct irq_remap_table *get_irq_table(struct amd_iommu *iommu, u16 devid) { struct irq_remap_table *table; + struct amd_iommu_pci_seg *pci_seg = iommu->pci_seg; - if (WARN_ONCE(!amd_iommu_rlookup_table[devid], - "%s: no iommu for devid %x\n", __func__, devid)) + if (WARN_ONCE(!pci_seg->rlookup_table[devid], + "%s: no iommu for devid %x:%x\n", + __func__, pci_seg->id, devid)) return NULL; - table = irq_lookup_table[devid]; - if (WARN_ONCE(!table, "%s: no table for devid %x\n", __func__, devid)) + table = pci_seg->irq_lookup_table[devid]; + if (WARN_ONCE(!table, "%s: no table for devid %x:%x\n", + __func__, pci_seg->id, devid)) return NULL; return table; } -static struct irq_remap_table *__alloc_irq_table(void) +static struct irq_remap_table *__alloc_irq_table(int nid, size_t size) { struct irq_remap_table *table; @@ -3125,27 +3176,24 @@ static struct irq_remap_table *__alloc_irq_table(void) if (!table) return NULL; - table->table = kmem_cache_alloc(amd_iommu_irq_cache, GFP_KERNEL); + table->table = iommu_alloc_pages_node_sz( + nid, GFP_KERNEL, max(DTE_INTTAB_ALIGNMENT, size)); if (!table->table) { kfree(table); return NULL; } raw_spin_lock_init(&table->lock); - if (!AMD_IOMMU_GUEST_IR_GA(amd_iommu_guest_ir)) - memset(table->table, 0, - MAX_IRQS_PER_TABLE * sizeof(u32)); - else - memset(table->table, 0, - (MAX_IRQS_PER_TABLE * (sizeof(u64) * 2))); return table; } static void set_remap_table_entry(struct amd_iommu *iommu, u16 devid, struct irq_remap_table *table) { - irq_lookup_table[devid] = table; - set_dte_irq_entry(devid, table); + struct amd_iommu_pci_seg *pci_seg = iommu->pci_seg; + + pci_seg->irq_lookup_table[devid] = table; + set_dte_irq_entry(iommu, devid, table); iommu_flush_dte(iommu, devid); } @@ -3153,35 +3201,48 @@ static int set_remap_table_entry_alias(struct pci_dev *pdev, u16 alias, void *data) { struct irq_remap_table *table = data; + struct amd_iommu_pci_seg *pci_seg; + struct amd_iommu *iommu = rlookup_amd_iommu(&pdev->dev); - irq_lookup_table[alias] = table; - set_dte_irq_entry(alias, table); + if (!iommu) + return -EINVAL; - iommu_flush_dte(amd_iommu_rlookup_table[alias], alias); + pci_seg = iommu->pci_seg; + pci_seg->irq_lookup_table[alias] = table; + set_dte_irq_entry(iommu, alias, table); + iommu_flush_dte(pci_seg->rlookup_table[alias], alias); return 0; } -static struct irq_remap_table *alloc_irq_table(u16 devid, struct pci_dev *pdev) +static inline size_t get_irq_table_size(unsigned int max_irqs) +{ + if (!AMD_IOMMU_GUEST_IR_GA(amd_iommu_guest_ir)) + return max_irqs * sizeof(u32); + + return max_irqs * (sizeof(u64) * 2); +} + +static struct irq_remap_table *alloc_irq_table(struct amd_iommu *iommu, + u16 devid, struct pci_dev *pdev, + unsigned int max_irqs) { struct irq_remap_table *table = NULL; struct irq_remap_table *new_table = NULL; - struct amd_iommu *iommu; + struct amd_iommu_pci_seg *pci_seg; unsigned long flags; + int nid = iommu && iommu->dev ? dev_to_node(&iommu->dev->dev) : NUMA_NO_NODE; u16 alias; spin_lock_irqsave(&iommu_table_lock, flags); - iommu = amd_iommu_rlookup_table[devid]; - if (!iommu) - goto out_unlock; - - table = irq_lookup_table[devid]; + pci_seg = iommu->pci_seg; + table = pci_seg->irq_lookup_table[devid]; if (table) goto out_unlock; - alias = amd_iommu_alias_table[devid]; - table = irq_lookup_table[alias]; + alias = pci_seg->alias_table[devid]; + table = pci_seg->irq_lookup_table[alias]; if (table) { set_remap_table_entry(iommu, devid, table); goto out_wait; @@ -3189,17 +3250,17 @@ static struct irq_remap_table *alloc_irq_table(u16 devid, struct pci_dev *pdev) spin_unlock_irqrestore(&iommu_table_lock, flags); /* Nothing there yet, allocate new irq remapping table */ - new_table = __alloc_irq_table(); + new_table = __alloc_irq_table(nid, get_irq_table_size(max_irqs)); if (!new_table) return NULL; spin_lock_irqsave(&iommu_table_lock, flags); - table = irq_lookup_table[devid]; + table = pci_seg->irq_lookup_table[devid]; if (table) goto out_unlock; - table = irq_lookup_table[alias]; + table = pci_seg->irq_lookup_table[alias]; if (table) { set_remap_table_entry(iommu, devid, table); goto out_wait; @@ -3224,24 +3285,21 @@ out_unlock: spin_unlock_irqrestore(&iommu_table_lock, flags); if (new_table) { - kmem_cache_free(amd_iommu_irq_cache, new_table->table); + iommu_free_pages(new_table->table); kfree(new_table); } return table; } -static int alloc_irq_index(u16 devid, int count, bool align, - struct pci_dev *pdev) +static int alloc_irq_index(struct amd_iommu *iommu, u16 devid, int count, + bool align, struct pci_dev *pdev, + unsigned long max_irqs) { struct irq_remap_table *table; int index, c, alignment = 1; unsigned long flags; - struct amd_iommu *iommu = amd_iommu_rlookup_table[devid]; - - if (!iommu) - return -ENODEV; - table = alloc_irq_table(devid, pdev); + table = alloc_irq_table(iommu, devid, pdev, max_irqs); if (!table) return -ENODEV; @@ -3252,7 +3310,7 @@ static int alloc_irq_index(u16 devid, int count, bool align, /* Scan table for free entries */ for (index = ALIGN(table->min_index, alignment), c = 0; - index < MAX_IRQS_PER_TABLE;) { + index < max_irqs;) { if (!iommu->irte_ops->is_allocated(table, index)) { c += 1; } else { @@ -3280,19 +3338,15 @@ out: return index; } -static int modify_irte_ga(u16 devid, int index, struct irte_ga *irte, - struct amd_ir_data *data) +static int __modify_irte_ga(struct amd_iommu *iommu, u16 devid, int index, + struct irte_ga *irte) { struct irq_remap_table *table; - struct amd_iommu *iommu; - unsigned long flags; struct irte_ga *entry; + unsigned long flags; + u128 old; - iommu = amd_iommu_rlookup_table[devid]; - if (iommu == NULL) - return -EINVAL; - - table = get_irq_table(devid); + table = get_irq_table(iommu, devid); if (!table) return -ENOMEM; @@ -3300,32 +3354,42 @@ static int modify_irte_ga(u16 devid, int index, struct irte_ga *irte, entry = (struct irte_ga *)table->table; entry = &entry[index]; - entry->lo.fields_remap.valid = 0; - entry->hi.val = irte->hi.val; - entry->lo.val = irte->lo.val; - entry->lo.fields_remap.valid = 1; - if (data) - data->ref = entry; + + /* + * We use cmpxchg16 to atomically update the 128-bit IRTE, + * and it cannot be updated by the hardware or other processors + * behind us, so the return value of cmpxchg16 should be the + * same as the old value. + */ + old = entry->irte; + WARN_ON(!try_cmpxchg128(&entry->irte, &old, irte->irte)); raw_spin_unlock_irqrestore(&table->lock, flags); - iommu_flush_irt(iommu, devid); - iommu_completion_wait(iommu); + return 0; +} + +static int modify_irte_ga(struct amd_iommu *iommu, u16 devid, int index, + struct irte_ga *irte) +{ + int ret; + + ret = __modify_irte_ga(iommu, devid, index, irte); + if (ret) + return ret; + + iommu_flush_irt_and_complete(iommu, devid); return 0; } -static int modify_irte(u16 devid, int index, union irte *irte) +static int modify_irte(struct amd_iommu *iommu, + u16 devid, int index, union irte *irte) { struct irq_remap_table *table; - struct amd_iommu *iommu; unsigned long flags; - iommu = amd_iommu_rlookup_table[devid]; - if (iommu == NULL) - return -EINVAL; - - table = get_irq_table(devid); + table = get_irq_table(iommu, devid); if (!table) return -ENOMEM; @@ -3333,23 +3397,17 @@ static int modify_irte(u16 devid, int index, union irte *irte) table->table[index] = irte->val; raw_spin_unlock_irqrestore(&table->lock, flags); - iommu_flush_irt(iommu, devid); - iommu_completion_wait(iommu); + iommu_flush_irt_and_complete(iommu, devid); return 0; } -static void free_irte(u16 devid, int index) +static void free_irte(struct amd_iommu *iommu, u16 devid, int index) { struct irq_remap_table *table; - struct amd_iommu *iommu; unsigned long flags; - iommu = amd_iommu_rlookup_table[devid]; - if (iommu == NULL) - return; - - table = get_irq_table(devid); + table = get_irq_table(iommu, devid); if (!table) return; @@ -3357,12 +3415,11 @@ static void free_irte(u16 devid, int index) iommu->irte_ops->clear_allocated(table, index); raw_spin_unlock_irqrestore(&table->lock, flags); - iommu_flush_irt(iommu, devid); - iommu_completion_wait(iommu); + iommu_flush_irt_and_complete(iommu, devid); } static void irte_prepare(void *entry, - u32 delivery_mode, u32 dest_mode, + u32 delivery_mode, bool dest_mode, u8 vector, u32 dest_apicid, int devid) { union irte *irte = (union irte *) entry; @@ -3376,7 +3433,7 @@ static void irte_prepare(void *entry, } static void irte_ga_prepare(void *entry, - u32 delivery_mode, u32 dest_mode, + u32 delivery_mode, bool dest_mode, u8 vector, u32 dest_apicid, int devid) { struct irte_ga *irte = (struct irte_ga *) entry; @@ -3391,49 +3448,49 @@ static void irte_ga_prepare(void *entry, irte->lo.fields_remap.valid = 1; } -static void irte_activate(void *entry, u16 devid, u16 index) +static void irte_activate(struct amd_iommu *iommu, void *entry, u16 devid, u16 index) { union irte *irte = (union irte *) entry; irte->fields.valid = 1; - modify_irte(devid, index, irte); + modify_irte(iommu, devid, index, irte); } -static void irte_ga_activate(void *entry, u16 devid, u16 index) +static void irte_ga_activate(struct amd_iommu *iommu, void *entry, u16 devid, u16 index) { struct irte_ga *irte = (struct irte_ga *) entry; irte->lo.fields_remap.valid = 1; - modify_irte_ga(devid, index, irte, NULL); + modify_irte_ga(iommu, devid, index, irte); } -static void irte_deactivate(void *entry, u16 devid, u16 index) +static void irte_deactivate(struct amd_iommu *iommu, void *entry, u16 devid, u16 index) { union irte *irte = (union irte *) entry; irte->fields.valid = 0; - modify_irte(devid, index, irte); + modify_irte(iommu, devid, index, irte); } -static void irte_ga_deactivate(void *entry, u16 devid, u16 index) +static void irte_ga_deactivate(struct amd_iommu *iommu, void *entry, u16 devid, u16 index) { struct irte_ga *irte = (struct irte_ga *) entry; irte->lo.fields_remap.valid = 0; - modify_irte_ga(devid, index, irte, NULL); + modify_irte_ga(iommu, devid, index, irte); } -static void irte_set_affinity(void *entry, u16 devid, u16 index, +static void irte_set_affinity(struct amd_iommu *iommu, void *entry, u16 devid, u16 index, u8 vector, u32 dest_apicid) { union irte *irte = (union irte *) entry; irte->fields.vector = vector; irte->fields.destination = dest_apicid; - modify_irte(devid, index, irte); + modify_irte(iommu, devid, index, irte); } -static void irte_ga_set_affinity(void *entry, u16 devid, u16 index, +static void irte_ga_set_affinity(struct amd_iommu *iommu, void *entry, u16 devid, u16 index, u8 vector, u32 dest_apicid) { struct irte_ga *irte = (struct irte_ga *) entry; @@ -3444,7 +3501,7 @@ static void irte_ga_set_affinity(void *entry, u16 devid, u16 index, APICID_TO_IRTE_DEST_LO(dest_apicid); irte->hi.fields.destination = APICID_TO_IRTE_DEST_HI(dest_apicid); - modify_irte_ga(devid, index, irte, NULL); + modify_irte_ga(iommu, devid, index, irte); } } @@ -3496,69 +3553,18 @@ static void irte_ga_clear_allocated(struct irq_remap_table *table, int index) static int get_devid(struct irq_alloc_info *info) { - int devid = -1; - switch (info->type) { case X86_IRQ_ALLOC_TYPE_IOAPIC: - devid = get_ioapic_devid(info->ioapic_id); - break; + return get_ioapic_devid(info->devid); case X86_IRQ_ALLOC_TYPE_HPET: - devid = get_hpet_devid(info->hpet_id); - break; - case X86_IRQ_ALLOC_TYPE_MSI: - case X86_IRQ_ALLOC_TYPE_MSIX: - devid = get_device_id(&info->msi_dev->dev); - break; - default: - BUG_ON(1); - break; - } - - return devid; -} - -static struct irq_domain *get_ir_irq_domain(struct irq_alloc_info *info) -{ - struct amd_iommu *iommu; - int devid; - - if (!info) - return NULL; - - devid = get_devid(info); - if (devid >= 0) { - iommu = amd_iommu_rlookup_table[devid]; - if (iommu) - return iommu->ir_domain; - } - - return NULL; -} - -static struct irq_domain *get_irq_domain(struct irq_alloc_info *info) -{ - struct amd_iommu *iommu; - int devid; - - if (!info) - return NULL; - - switch (info->type) { - case X86_IRQ_ALLOC_TYPE_MSI: - case X86_IRQ_ALLOC_TYPE_MSIX: - devid = get_device_id(&info->msi_dev->dev); - if (devid < 0) - return NULL; - - iommu = amd_iommu_rlookup_table[devid]; - if (iommu) - return iommu->msi_domain; - break; + return get_hpet_devid(info->devid); + case X86_IRQ_ALLOC_TYPE_PCI_MSI: + case X86_IRQ_ALLOC_TYPE_PCI_MSIX: + return get_device_sbdf_id(msi_desc_to_dev(info->desc)); default: - break; + WARN_ON_ONCE(1); + return -1; } - - return NULL; } struct irq_remap_ops amd_iommu_irq_ops = { @@ -3567,50 +3573,47 @@ struct irq_remap_ops amd_iommu_irq_ops = { .disable = amd_iommu_disable, .reenable = amd_iommu_reenable, .enable_faulting = amd_iommu_enable_faulting, - .get_ir_irq_domain = get_ir_irq_domain, - .get_irq_domain = get_irq_domain, }; +static void fill_msi_msg(struct msi_msg *msg, u32 index) +{ + msg->data = index; + msg->address_lo = 0; + msg->arch_addr_lo.base_address = X86_MSI_BASE_ADDRESS_LOW; + /* + * The struct msi_msg.dest_mode_logical is used to set the DM bit + * in MSI Message Address Register. For device w/ 2K int-remap support, + * this is bit must be set to 1 regardless of the actual destination + * mode, which is signified by the IRTE[DM]. + */ + if (FEATURE_NUM_INT_REMAP_SUP_2K(amd_iommu_efr2)) + msg->arch_addr_lo.dest_mode_logical = true; + msg->address_hi = X86_MSI_BASE_ADDRESS_HIGH; +} + static void irq_remapping_prepare_irte(struct amd_ir_data *data, struct irq_cfg *irq_cfg, struct irq_alloc_info *info, int devid, int index, int sub_handle) { struct irq_2_irte *irte_info = &data->irq_2_irte; - struct msi_msg *msg = &data->msi_entry; - struct IO_APIC_route_entry *entry; - struct amd_iommu *iommu = amd_iommu_rlookup_table[devid]; + struct amd_iommu *iommu = data->iommu; if (!iommu) return; data->irq_2_irte.devid = devid; data->irq_2_irte.index = index + sub_handle; - iommu->irte_ops->prepare(data->entry, apic->irq_delivery_mode, - apic->irq_dest_mode, irq_cfg->vector, + iommu->irte_ops->prepare(data->entry, APIC_DELIVERY_MODE_FIXED, + apic->dest_mode_logical, irq_cfg->vector, irq_cfg->dest_apicid, devid); switch (info->type) { case X86_IRQ_ALLOC_TYPE_IOAPIC: - /* Setup IOAPIC entry */ - entry = info->ioapic_entry; - info->ioapic_entry = NULL; - memset(entry, 0, sizeof(*entry)); - entry->vector = index; - entry->mask = 0; - entry->trigger = info->ioapic_trigger; - entry->polarity = info->ioapic_polarity; - /* Mask level triggered irqs. */ - if (info->ioapic_trigger) - entry->mask = 1; - break; - case X86_IRQ_ALLOC_TYPE_HPET: - case X86_IRQ_ALLOC_TYPE_MSI: - case X86_IRQ_ALLOC_TYPE_MSIX: - msg->address_hi = MSI_ADDR_BASE_HI; - msg->address_lo = MSI_ADDR_BASE_LO; - msg->data = irte_info->index; + case X86_IRQ_ALLOC_TYPE_PCI_MSI: + case X86_IRQ_ALLOC_TYPE_PCI_MSIX: + fill_msi_msg(&data->msi_entry, irte_info->index); break; default: @@ -3645,36 +3648,39 @@ static int irq_remapping_alloc(struct irq_domain *domain, unsigned int virq, struct irq_alloc_info *info = arg; struct irq_data *irq_data; struct amd_ir_data *data = NULL; + struct amd_iommu *iommu; struct irq_cfg *cfg; - int i, ret, devid; + struct iommu_dev_data *dev_data; + unsigned long max_irqs; + int i, ret, devid, seg, sbdf; int index; if (!info) return -EINVAL; - if (nr_irqs > 1 && info->type != X86_IRQ_ALLOC_TYPE_MSI && - info->type != X86_IRQ_ALLOC_TYPE_MSIX) + if (nr_irqs > 1 && info->type != X86_IRQ_ALLOC_TYPE_PCI_MSI) return -EINVAL; - /* - * With IRQ remapping enabled, don't need contiguous CPU vectors - * to support multiple MSI interrupts. - */ - if (info->type == X86_IRQ_ALLOC_TYPE_MSI) - info->flags &= ~X86_IRQ_ALLOC_CONTIGUOUS_VECTORS; + sbdf = get_devid(info); + if (sbdf < 0) + return -EINVAL; - devid = get_devid(info); - if (devid < 0) + seg = PCI_SBDF_TO_SEGID(sbdf); + devid = PCI_SBDF_TO_DEVID(sbdf); + iommu = __rlookup_amd_iommu(seg, devid); + if (!iommu) return -EINVAL; + dev_data = search_dev_data(iommu, devid); + max_irqs = dev_data ? dev_data->max_irqs : MAX_IRQS_PER_TABLE_512; + ret = irq_domain_alloc_irqs_parent(domain, virq, nr_irqs, arg); if (ret < 0) return ret; if (info->type == X86_IRQ_ALLOC_TYPE_IOAPIC) { struct irq_remap_table *table; - struct amd_iommu *iommu; - table = alloc_irq_table(devid, NULL); + table = alloc_irq_table(iommu, devid, NULL, max_irqs); if (table) { if (!table->min_index) { /* @@ -3682,22 +3688,24 @@ static int irq_remapping_alloc(struct irq_domain *domain, unsigned int virq, * interrupts. */ table->min_index = 32; - iommu = amd_iommu_rlookup_table[devid]; for (i = 0; i < 32; ++i) iommu->irte_ops->set_allocated(table, i); } WARN_ON(table->min_index != 32); - index = info->ioapic_pin; + index = info->ioapic.pin; } else { index = -ENOMEM; } - } else if (info->type == X86_IRQ_ALLOC_TYPE_MSI || - info->type == X86_IRQ_ALLOC_TYPE_MSIX) { - bool align = (info->type == X86_IRQ_ALLOC_TYPE_MSI); + } else if (info->type == X86_IRQ_ALLOC_TYPE_PCI_MSI || + info->type == X86_IRQ_ALLOC_TYPE_PCI_MSIX) { + bool align = (info->type == X86_IRQ_ALLOC_TYPE_PCI_MSI); - index = alloc_irq_index(devid, nr_irqs, align, info->msi_dev); + index = alloc_irq_index(iommu, devid, nr_irqs, align, + msi_desc_to_pci_dev(info->desc), + max_irqs); } else { - index = alloc_irq_index(devid, nr_irqs, false, NULL); + index = alloc_irq_index(iommu, devid, nr_irqs, false, NULL, + max_irqs); } if (index < 0) { @@ -3708,8 +3716,8 @@ static int irq_remapping_alloc(struct irq_domain *domain, unsigned int virq, for (i = 0; i < nr_irqs; i++) { irq_data = irq_domain_get_irq_data(domain, virq + i); - cfg = irqd_cfg(irq_data); - if (!irq_data || !cfg) { + cfg = irq_data ? irqd_cfg(irq_data) : NULL; + if (!cfg) { ret = -EINVAL; goto out_free_data; } @@ -3729,11 +3737,11 @@ static int irq_remapping_alloc(struct irq_domain *domain, unsigned int virq, goto out_free_data; } + data->iommu = iommu; irq_data->hwirq = (devid << 16) + i; irq_data->chip_data = data; irq_data->chip = &amd_ir_chip; irq_remapping_prepare_irte(data, cfg, info, devid, index, i); - irq_set_status_flags(virq + i, IRQ_MOVE_PCNTXT); } return 0; @@ -3745,7 +3753,7 @@ out_free_data: kfree(irq_data->chip_data); } for (i = 0; i < nr_irqs; i++) - free_irte(devid, index + i); + free_irte(iommu, devid, index + i); out_free_parent: irq_domain_free_irqs_common(domain, virq, nr_irqs); return ret; @@ -3764,7 +3772,7 @@ static void irq_remapping_free(struct irq_domain *domain, unsigned int virq, if (irq_data && irq_data->chip_data) { data = irq_data->chip_data; irte_info = &data->irq_2_irte; - free_irte(irte_info->devid, irte_info->index); + free_irte(data->iommu, irte_info->devid, irte_info->index); kfree(data->entry); kfree(data); } @@ -3782,13 +3790,13 @@ static int irq_remapping_activate(struct irq_domain *domain, { struct amd_ir_data *data = irq_data->chip_data; struct irq_2_irte *irte_info = &data->irq_2_irte; - struct amd_iommu *iommu = amd_iommu_rlookup_table[irte_info->devid]; + struct amd_iommu *iommu = data->iommu; struct irq_cfg *cfg = irqd_cfg(irq_data); if (!iommu) return 0; - iommu->irte_ops->activate(data->entry, irte_info->devid, + iommu->irte_ops->activate(iommu, data->entry, irte_info->devid, irte_info->index); amd_ir_update_irte(irq_data, iommu, data, irte_info, cfg); return 0; @@ -3799,40 +3807,123 @@ static void irq_remapping_deactivate(struct irq_domain *domain, { struct amd_ir_data *data = irq_data->chip_data; struct irq_2_irte *irte_info = &data->irq_2_irte; - struct amd_iommu *iommu = amd_iommu_rlookup_table[irte_info->devid]; + struct amd_iommu *iommu = data->iommu; if (iommu) - iommu->irte_ops->deactivate(data->entry, irte_info->devid, + iommu->irte_ops->deactivate(iommu, data->entry, irte_info->devid, irte_info->index); } +static int irq_remapping_select(struct irq_domain *d, struct irq_fwspec *fwspec, + enum irq_domain_bus_token bus_token) +{ + struct amd_iommu *iommu; + int devid = -1; + + if (!amd_iommu_irq_remap) + return 0; + + if (x86_fwspec_is_ioapic(fwspec)) + devid = get_ioapic_devid(fwspec->param[0]); + else if (x86_fwspec_is_hpet(fwspec)) + devid = get_hpet_devid(fwspec->param[0]); + + if (devid < 0) + return 0; + iommu = __rlookup_amd_iommu((devid >> 16), (devid & 0xffff)); + + return iommu && iommu->ir_domain == d; +} + static const struct irq_domain_ops amd_ir_domain_ops = { + .select = irq_remapping_select, .alloc = irq_remapping_alloc, .free = irq_remapping_free, .activate = irq_remapping_activate, .deactivate = irq_remapping_deactivate, }; -int amd_iommu_activate_guest_mode(void *data) +static void __amd_iommu_update_ga(struct irte_ga *entry, int cpu, + bool ga_log_intr) +{ + if (cpu >= 0) { + entry->lo.fields_vapic.destination = + APICID_TO_IRTE_DEST_LO(cpu); + entry->hi.fields.destination = + APICID_TO_IRTE_DEST_HI(cpu); + entry->lo.fields_vapic.is_run = true; + entry->lo.fields_vapic.ga_log_intr = false; + } else { + entry->lo.fields_vapic.is_run = false; + entry->lo.fields_vapic.ga_log_intr = ga_log_intr; + } +} + +/* + * Update the pCPU information for an IRTE that is configured to post IRQs to + * a vCPU, without issuing an IOMMU invalidation for the IRTE. + * + * If the vCPU is associated with a pCPU (@cpu >= 0), configure the Destination + * with the pCPU's APIC ID, set IsRun, and clear GALogIntr. If the vCPU isn't + * associated with a pCPU (@cpu < 0), clear IsRun and set/clear GALogIntr based + * on input from the caller (e.g. KVM only requests GALogIntr when the vCPU is + * blocking and requires a notification wake event). I.e. treat vCPUs that are + * associated with a pCPU as running. This API is intended to be used when a + * vCPU is scheduled in/out (or stops running for any reason), to do a fast + * update of IsRun, GALogIntr, and (conditionally) Destination. + * + * Per the IOMMU spec, the Destination, IsRun, and GATag fields are not cached + * and thus don't require an invalidation to ensure the IOMMU consumes fresh + * information. + */ +int amd_iommu_update_ga(void *data, int cpu, bool ga_log_intr) +{ + struct amd_ir_data *ir_data = (struct amd_ir_data *)data; + struct irte_ga *entry = (struct irte_ga *) ir_data->entry; + + if (WARN_ON_ONCE(!AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir))) + return -EINVAL; + + if (!entry || !entry->lo.fields_vapic.guest_mode) + return 0; + + if (!ir_data->iommu) + return -ENODEV; + + __amd_iommu_update_ga(entry, cpu, ga_log_intr); + + return __modify_irte_ga(ir_data->iommu, ir_data->irq_2_irte.devid, + ir_data->irq_2_irte.index, entry); +} +EXPORT_SYMBOL(amd_iommu_update_ga); + +int amd_iommu_activate_guest_mode(void *data, int cpu, bool ga_log_intr) { struct amd_ir_data *ir_data = (struct amd_ir_data *)data; struct irte_ga *entry = (struct irte_ga *) ir_data->entry; + u64 valid; + + if (WARN_ON_ONCE(!AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir))) + return -EINVAL; - if (!AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir) || - !entry || entry->lo.fields_vapic.guest_mode) + if (!entry) return 0; + valid = entry->lo.fields_vapic.valid; + entry->lo.val = 0; entry->hi.val = 0; + entry->lo.fields_vapic.valid = valid; entry->lo.fields_vapic.guest_mode = 1; - entry->lo.fields_vapic.ga_log_intr = 1; entry->hi.fields.ga_root_ptr = ir_data->ga_root_ptr; entry->hi.fields.vector = ir_data->ga_vector; entry->lo.fields_vapic.ga_tag = ir_data->ga_tag; - return modify_irte_ga(ir_data->irq_2_irte.devid, - ir_data->irq_2_irte.index, entry, ir_data); + __amd_iommu_update_ga(entry, cpu, ga_log_intr); + + return modify_irte_ga(ir_data->iommu, ir_data->irq_2_irte.devid, + ir_data->irq_2_irte.index, entry); } EXPORT_SYMBOL(amd_iommu_activate_guest_mode); @@ -3841,78 +3932,71 @@ int amd_iommu_deactivate_guest_mode(void *data) struct amd_ir_data *ir_data = (struct amd_ir_data *)data; struct irte_ga *entry = (struct irte_ga *) ir_data->entry; struct irq_cfg *cfg = ir_data->cfg; + u64 valid; - if (!AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir) || - !entry || !entry->lo.fields_vapic.guest_mode) + if (WARN_ON_ONCE(!AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir))) + return -EINVAL; + + if (!entry || !entry->lo.fields_vapic.guest_mode) return 0; + valid = entry->lo.fields_remap.valid; + entry->lo.val = 0; entry->hi.val = 0; - entry->lo.fields_remap.dm = apic->irq_dest_mode; - entry->lo.fields_remap.int_type = apic->irq_delivery_mode; + entry->lo.fields_remap.valid = valid; + entry->lo.fields_remap.dm = apic->dest_mode_logical; + entry->lo.fields_remap.int_type = APIC_DELIVERY_MODE_FIXED; entry->hi.fields.vector = cfg->vector; entry->lo.fields_remap.destination = APICID_TO_IRTE_DEST_LO(cfg->dest_apicid); entry->hi.fields.destination = APICID_TO_IRTE_DEST_HI(cfg->dest_apicid); - return modify_irte_ga(ir_data->irq_2_irte.devid, - ir_data->irq_2_irte.index, entry, ir_data); + return modify_irte_ga(ir_data->iommu, ir_data->irq_2_irte.devid, + ir_data->irq_2_irte.index, entry); } EXPORT_SYMBOL(amd_iommu_deactivate_guest_mode); -static int amd_ir_set_vcpu_affinity(struct irq_data *data, void *vcpu_info) +static int amd_ir_set_vcpu_affinity(struct irq_data *data, void *info) { int ret; - struct amd_iommu *iommu; - struct amd_iommu_pi_data *pi_data = vcpu_info; - struct vcpu_data *vcpu_pi_info = pi_data->vcpu_data; + struct amd_iommu_pi_data *pi_data = info; struct amd_ir_data *ir_data = data->chip_data; struct irq_2_irte *irte_info = &ir_data->irq_2_irte; - struct iommu_dev_data *dev_data = search_dev_data(irte_info->devid); + struct iommu_dev_data *dev_data; + + if (WARN_ON_ONCE(!AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir))) + return -EINVAL; + + if (ir_data->iommu == NULL) + return -EINVAL; + + dev_data = search_dev_data(ir_data->iommu, irte_info->devid); /* Note: * This device has never been set up for guest mode. * we should not modify the IRTE */ if (!dev_data || !dev_data->use_vapic) - return 0; + return -EINVAL; ir_data->cfg = irqd_cfg(data); - pi_data->ir_data = ir_data; - /* Note: - * SVM tries to set up for VAPIC mode, but we are in - * legacy mode. So, we force legacy mode instead. - */ - if (!AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir)) { - pr_debug("%s: Fall back to using intr legacy remap\n", - __func__); - pi_data->is_guest_mode = false; - } - - iommu = amd_iommu_rlookup_table[irte_info->devid]; - if (iommu == NULL) - return -EINVAL; + if (pi_data) { + pi_data->ir_data = ir_data; - pi_data->prev_ga_tag = ir_data->cached_ga_tag; - if (pi_data->is_guest_mode) { - ir_data->ga_root_ptr = (pi_data->base >> 12); - ir_data->ga_vector = vcpu_pi_info->vector; + ir_data->ga_root_ptr = (pi_data->vapic_addr >> 12); + ir_data->ga_vector = pi_data->vector; ir_data->ga_tag = pi_data->ga_tag; - ret = amd_iommu_activate_guest_mode(ir_data); - if (!ret) - ir_data->cached_ga_tag = pi_data->ga_tag; + if (pi_data->is_guest_mode) + ret = amd_iommu_activate_guest_mode(ir_data, pi_data->cpu, + pi_data->ga_log_intr); + else + ret = amd_iommu_deactivate_guest_mode(ir_data); } else { ret = amd_iommu_deactivate_guest_mode(ir_data); - - /* - * This communicates the ga_tag back to the caller - * so that it can do all the necessary clean up. - */ - if (!ret) - ir_data->cached_ga_tag = 0; } return ret; @@ -3929,7 +4013,7 @@ static void amd_ir_update_irte(struct irq_data *irqd, struct amd_iommu *iommu, * Atomically updates the IRTE with the new destination, vector * and flushes the interrupt entry cache. */ - iommu->irte_ops->set_affinity(ir_data->entry, irte_info->devid, + iommu->irte_ops->set_affinity(iommu, ir_data->entry, irte_info->devid, irte_info->index, cfg->vector, cfg->dest_apicid); } @@ -3941,7 +4025,7 @@ static int amd_ir_set_affinity(struct irq_data *data, struct irq_2_irte *irte_info = &ir_data->irq_2_irte; struct irq_cfg *cfg = irqd_cfg(data); struct irq_data *parent = data->parent_data; - struct amd_iommu *iommu = amd_iommu_rlookup_table[irte_info->devid]; + struct amd_iommu *iommu = ir_data->iommu; int ret; if (!iommu) @@ -3957,7 +4041,7 @@ static int amd_ir_set_affinity(struct irq_data *data, * at the new destination. So, time to cleanup the previous * vector allocation. */ - send_cleanup_vector(cfg); + vector_schedule_cleanup(cfg); return IRQ_SET_MASK_OK_DONE; } @@ -3977,65 +4061,34 @@ static struct irq_chip amd_ir_chip = { .irq_compose_msi_msg = ir_compose_msi_msg, }; +static const struct msi_parent_ops amdvi_msi_parent_ops = { + .supported_flags = X86_VECTOR_MSI_FLAGS_SUPPORTED | MSI_FLAG_MULTI_PCI_MSI, + .bus_select_token = DOMAIN_BUS_AMDVI, + .bus_select_mask = MATCH_PCI_MSI, + .prefix = "IR-", + .init_dev_msi_info = msi_parent_init_dev_msi_info, +}; + int amd_iommu_create_irq_domain(struct amd_iommu *iommu) { - struct fwnode_handle *fn; + struct irq_domain_info info = { + .fwnode = irq_domain_alloc_named_id_fwnode("AMD-IR", iommu->index), + .ops = &amd_ir_domain_ops, + .domain_flags = IRQ_DOMAIN_FLAG_ISOLATED_MSI, + .host_data = iommu, + .parent = arch_get_ir_parent_domain(), + }; - fn = irq_domain_alloc_named_id_fwnode("AMD-IR", iommu->index); - if (!fn) - return -ENOMEM; - iommu->ir_domain = irq_domain_create_tree(fn, &amd_ir_domain_ops, iommu); - irq_domain_free_fwnode(fn); - if (!iommu->ir_domain) + if (!info.fwnode) return -ENOMEM; - iommu->ir_domain->parent = arch_get_ir_parent_domain(); - iommu->msi_domain = arch_create_remap_msi_irq_domain(iommu->ir_domain, - "AMD-IR-MSI", - iommu->index); - return 0; -} - -int amd_iommu_update_ga(int cpu, bool is_run, void *data) -{ - unsigned long flags; - struct amd_iommu *iommu; - struct irq_remap_table *table; - struct amd_ir_data *ir_data = (struct amd_ir_data *)data; - int devid = ir_data->irq_2_irte.devid; - struct irte_ga *entry = (struct irte_ga *) ir_data->entry; - struct irte_ga *ref = (struct irte_ga *) ir_data->ref; - - if (!AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir) || - !ref || !entry || !entry->lo.fields_vapic.guest_mode) - return 0; - - iommu = amd_iommu_rlookup_table[devid]; - if (!iommu) - return -ENODEV; - - table = get_irq_table(devid); - if (!table) - return -ENODEV; - - raw_spin_lock_irqsave(&table->lock, flags); - - if (ref->lo.fields_vapic.guest_mode) { - if (cpu >= 0) { - ref->lo.fields_vapic.destination = - APICID_TO_IRTE_DEST_LO(cpu); - ref->hi.fields.destination = - APICID_TO_IRTE_DEST_HI(cpu); - } - ref->lo.fields_vapic.is_run = is_run; - barrier(); + iommu->ir_domain = msi_create_parent_irq_domain(&info, &amdvi_msi_parent_ops); + if (!iommu->ir_domain) { + irq_domain_free_fwnode(info.fwnode); + return -ENOMEM; } - - raw_spin_unlock_irqrestore(&table->lock, flags); - - iommu_flush_irt(iommu, devid); - iommu_completion_wait(iommu); return 0; } -EXPORT_SYMBOL(amd_iommu_update_ga); #endif + +MODULE_IMPORT_NS("GENERIC_PT_IOMMU"); |