diff options
43 files changed, 2943 insertions, 317 deletions
diff --git a/Documentation/ABI/testing/sysfs-bus-pci b/Documentation/ABI/testing/sysfs-bus-pci index 349ecf26ce10..34f51100f029 100644 --- a/Documentation/ABI/testing/sysfs-bus-pci +++ b/Documentation/ABI/testing/sysfs-bus-pci @@ -66,6 +66,24 @@ Description: re-discover previously removed devices. Depends on CONFIG_HOTPLUG. +What: /sys/bus/pci/devices/.../msi_irqs/ +Date: September, 2011 +Contact: Neil Horman <nhorman@tuxdriver.com> +Description: + The /sys/devices/.../msi_irqs directory contains a variable set + of sub-directories, with each sub-directory being named after a + corresponding msi irq vector allocated to that device. Each + numbered sub-directory N contains attributes of that irq. + Note that this directory is not created for device drivers which + do not support msi irqs + +What: /sys/bus/pci/devices/.../msi_irqs/<N>/mode +Date: September 2011 +Contact: Neil Horman <nhorman@tuxdriver.com> +Description: + This attribute indicates the mode that the irq vector named by + the parent directory is in (msi vs. msix) + What: /sys/bus/pci/devices/.../remove Date: January 2009 Contact: Linux PCI developers <linux-pci@vger.kernel.org> diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index e69a461a06c2..9373d95319c1 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -329,6 +329,11 @@ bytes respectively. Such letter suffixes can also be entirely omitted. is a lot of faster off - do not initialize any AMD IOMMU found in the system + force_isolation - Force device isolation for all + devices. The IOMMU driver is not + allowed anymore to lift isolation + requirements as needed. This option + does not override iommu=pt amijoy.map= [HW,JOY] Amiga joystick support Map of devices attached to JOY0DAT and JOY1DAT @@ -1059,7 +1064,9 @@ bytes respectively. Such letter suffixes can also be entirely omitted. nomerge forcesac soft - pt [x86, IA-64] + pt [x86, IA-64] + group_mf [x86, IA-64] + io7= [HW] IO7 for Marvel based alpha systems See comment before marvel_specify_io7 in diff --git a/arch/arm/mach-omap2/devices.c b/arch/arm/mach-omap2/devices.c index 35d5dffab7e1..46dfd1ae8f71 100644 --- a/arch/arm/mach-omap2/devices.c +++ b/arch/arm/mach-omap2/devices.c @@ -28,6 +28,7 @@ #include <plat/board.h> #include <plat/mcbsp.h> #include <plat/mmc.h> +#include <plat/iommu.h> #include <plat/dma.h> #include <plat/omap_hwmod.h> #include <plat/omap_device.h> @@ -211,9 +212,15 @@ static struct platform_device omap3isp_device = { .resource = omap3isp_resources, }; +static struct omap_iommu_arch_data omap3_isp_iommu = { + .name = "isp", +}; + int omap3_init_camera(struct isp_platform_data *pdata) { omap3isp_device.dev.platform_data = pdata; + omap3isp_device.dev.archdata.iommu = &omap3_isp_iommu; + return platform_device_register(&omap3isp_device); } diff --git a/arch/arm/plat-omap/include/plat/iommu.h b/arch/arm/plat-omap/include/plat/iommu.h index a1d79ee19250..88be3e628b33 100644 --- a/arch/arm/plat-omap/include/plat/iommu.h +++ b/arch/arm/plat-omap/include/plat/iommu.h @@ -111,6 +111,32 @@ struct iommu_platform_data { u32 da_end; }; +/** + * struct iommu_arch_data - omap iommu private data + * @name: name of the iommu device + * @iommu_dev: handle of the iommu device + * + * This is an omap iommu private data object, which binds an iommu user + * to its iommu device. This object should be placed at the iommu user's + * dev_archdata so generic IOMMU API can be used without having to + * utilize omap-specific plumbing anymore. + */ +struct omap_iommu_arch_data { + const char *name; + struct omap_iommu *iommu_dev; +}; + +/** + * dev_to_omap_iommu() - retrieves an omap iommu object from a user device + * @dev: iommu client device + */ +static inline struct omap_iommu *dev_to_omap_iommu(struct device *dev) +{ + struct omap_iommu_arch_data *arch_data = dev->archdata.iommu; + + return arch_data->iommu_dev; +} + /* IOMMU errors */ #define OMAP_IOMMU_ERR_TLB_MISS (1 << 0) #define OMAP_IOMMU_ERR_TRANS_FAULT (1 << 1) @@ -163,8 +189,8 @@ extern int omap_iommu_set_isr(const char *name, void *priv), void *isr_priv); -extern void omap_iommu_save_ctx(struct omap_iommu *obj); -extern void omap_iommu_restore_ctx(struct omap_iommu *obj); +extern void omap_iommu_save_ctx(struct device *dev); +extern void omap_iommu_restore_ctx(struct device *dev); extern int omap_install_iommu_arch(const struct iommu_functions *ops); extern void omap_uninstall_iommu_arch(const struct iommu_functions *ops); @@ -176,6 +202,5 @@ extern ssize_t omap_iommu_dump_ctx(struct omap_iommu *obj, char *buf, ssize_t len); extern size_t omap_dump_tlb_entries(struct omap_iommu *obj, char *buf, ssize_t len); -struct device *omap_find_iommu_device(const char *name); #endif /* __MACH_IOMMU_H */ diff --git a/arch/arm/plat-omap/include/plat/iovmm.h b/arch/arm/plat-omap/include/plat/iovmm.h index 6af1a91c0f36..498e57cda6cd 100644 --- a/arch/arm/plat-omap/include/plat/iovmm.h +++ b/arch/arm/plat-omap/include/plat/iovmm.h @@ -72,18 +72,18 @@ struct iovm_struct { #define IOVMF_DA_FIXED (1 << (4 + IOVMF_SW_SHIFT)) -extern struct iovm_struct *omap_find_iovm_area(struct omap_iommu *obj, u32 da); +extern struct iovm_struct *omap_find_iovm_area(struct device *dev, u32 da); extern u32 -omap_iommu_vmap(struct iommu_domain *domain, struct omap_iommu *obj, u32 da, +omap_iommu_vmap(struct iommu_domain *domain, struct device *dev, u32 da, const struct sg_table *sgt, u32 flags); extern struct sg_table *omap_iommu_vunmap(struct iommu_domain *domain, - struct omap_iommu *obj, u32 da); + struct device *dev, u32 da); extern u32 -omap_iommu_vmalloc(struct iommu_domain *domain, struct omap_iommu *obj, +omap_iommu_vmalloc(struct iommu_domain *domain, struct device *dev, u32 da, size_t bytes, u32 flags); extern void -omap_iommu_vfree(struct iommu_domain *domain, struct omap_iommu *obj, +omap_iommu_vfree(struct iommu_domain *domain, struct device *dev, const u32 da); -extern void *omap_da_to_va(struct omap_iommu *obj, u32 da); +extern void *omap_da_to_va(struct device *dev, u32 da); #endif /* __IOMMU_MMAP_H */ diff --git a/arch/ia64/include/asm/iommu.h b/arch/ia64/include/asm/iommu.h index 105c93b00b1b..b6a809fa2995 100644 --- a/arch/ia64/include/asm/iommu.h +++ b/arch/ia64/include/asm/iommu.h @@ -11,10 +11,12 @@ extern void no_iommu_init(void); extern int force_iommu, no_iommu; extern int iommu_pass_through; extern int iommu_detected; +extern int iommu_group_mf; #else #define iommu_pass_through (0) #define no_iommu (1) #define iommu_detected (0) +#define iommu_group_mf (0) #endif extern void iommu_dma_init(void); extern void machvec_init(const char *name); diff --git a/arch/ia64/kernel/pci-dma.c b/arch/ia64/kernel/pci-dma.c index c16162c70860..eb1175720050 100644 --- a/arch/ia64/kernel/pci-dma.c +++ b/arch/ia64/kernel/pci-dma.c @@ -33,6 +33,7 @@ int force_iommu __read_mostly; #endif int iommu_pass_through; +int iommu_group_mf; /* Dummy device used for NULL arguments (normally ISA). Better would be probably a smaller DMA mask, but this is bug-to-bug compatible diff --git a/arch/x86/include/asm/iommu.h b/arch/x86/include/asm/iommu.h index 345c99cef152..dffc38ee6255 100644 --- a/arch/x86/include/asm/iommu.h +++ b/arch/x86/include/asm/iommu.h @@ -5,6 +5,7 @@ extern struct dma_map_ops nommu_dma_ops; extern int force_iommu, no_iommu; extern int iommu_detected; extern int iommu_pass_through; +extern int iommu_group_mf; /* 10 seconds */ #define DMAR_OPERATION_TIMEOUT ((cycles_t) tsc_khz*10*1000) diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c index 80dc793b3f63..1c4d769e21ea 100644 --- a/arch/x86/kernel/pci-dma.c +++ b/arch/x86/kernel/pci-dma.c @@ -45,6 +45,15 @@ int iommu_detected __read_mostly = 0; */ int iommu_pass_through __read_mostly; +/* + * Group multi-function PCI devices into a single device-group for the + * iommu_device_group interface. This tells the iommu driver to pretend + * it cannot distinguish between functions of a device, exposing only one + * group for the device. Useful for disallowing use of individual PCI + * functions from userspace drivers. + */ +int iommu_group_mf __read_mostly; + extern struct iommu_table_entry __iommu_table[], __iommu_table_end[]; /* Dummy device used for NULL arguments (normally ISA). */ @@ -169,6 +178,8 @@ static __init int iommu_setup(char *p) #endif if (!strncmp(p, "pt", 2)) iommu_pass_through = 1; + if (!strncmp(p, "group_mf", 8)) + iommu_group_mf = 1; gart_parse_options(p); diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index 2672c798272f..7aff6312ce7c 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -596,6 +596,13 @@ static int __devinit acpi_pci_root_add(struct acpi_device *device) if (ACPI_SUCCESS(status)) { dev_info(root->bus->bridge, "ACPI _OSC control (0x%02x) granted\n", flags); + if (acpi_gbl_FADT.boot_flags & ACPI_FADT_NO_ASPM) { + /* + * We have ASPM control, but the FADT indicates + * that it's unsupported. Clear it. + */ + pcie_clear_aspm(root->bus); + } } else { dev_info(root->bus->bridge, "ACPI _OSC request failed (%s), " diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig index 5414253b185a..6bea6962f8ee 100644 --- a/drivers/iommu/Kconfig +++ b/drivers/iommu/Kconfig @@ -34,7 +34,9 @@ config AMD_IOMMU bool "AMD IOMMU support" select SWIOTLB select PCI_MSI - select PCI_IOV + select PCI_ATS + select PCI_PRI + select PCI_PASID select IOMMU_API depends on X86_64 && PCI && ACPI ---help--- @@ -58,6 +60,15 @@ config AMD_IOMMU_STATS information to userspace via debugfs. If unsure, say N. +config AMD_IOMMU_V2 + tristate "AMD IOMMU Version 2 driver (EXPERIMENTAL)" + depends on AMD_IOMMU && PROFILING && EXPERIMENTAL + select MMU_NOTIFIER + ---help--- + This option enables support for the AMD IOMMUv2 features of the IOMMU + hardware. Select this option if you want to use devices that support + the the PCI PRI and PASID interface. + # Intel IOMMU support config DMAR_TABLE bool diff --git a/drivers/iommu/Makefile b/drivers/iommu/Makefile index 2f4448794bc7..0e36b4934aff 100644 --- a/drivers/iommu/Makefile +++ b/drivers/iommu/Makefile @@ -1,6 +1,7 @@ obj-$(CONFIG_IOMMU_API) += iommu.o obj-$(CONFIG_MSM_IOMMU) += msm_iommu.o msm_iommu_dev.o obj-$(CONFIG_AMD_IOMMU) += amd_iommu.o amd_iommu_init.o +obj-$(CONFIG_AMD_IOMMU_V2) += amd_iommu_v2.o obj-$(CONFIG_DMAR_TABLE) += dmar.o obj-$(CONFIG_INTEL_IOMMU) += iova.o intel-iommu.o obj-$(CONFIG_IRQ_REMAP) += intr_remapping.o diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c index 4ee277a8521a..cce1f03b8895 100644 --- a/drivers/iommu/amd_iommu.c +++ b/drivers/iommu/amd_iommu.c @@ -17,6 +17,7 @@ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ +#include <linux/ratelimit.h> #include <linux/pci.h> #include <linux/pci-ats.h> #include <linux/bitmap.h> @@ -28,6 +29,8 @@ #include <linux/iommu.h> #include <linux/delay.h> #include <linux/amd-iommu.h> +#include <linux/notifier.h> +#include <linux/export.h> #include <asm/msidef.h> #include <asm/proto.h> #include <asm/iommu.h> @@ -41,6 +44,24 @@ #define LOOP_TIMEOUT 100000 +/* + * This bitmap is used to advertise the page sizes our hardware support + * to the IOMMU core, which will then use this information to split + * physically contiguous memory regions it is mapping into page sizes + * that we support. + * + * Traditionally the IOMMU core just handed us the mappings directly, + * after making sure the size is an order of a 4KiB page and that the + * mapping has natural alignment. + * + * To retain this behavior, we currently advertise that we support + * all page sizes that are an order of 4KiB. + * + * If at some point we'd like to utilize the IOMMU core's new behavior, + * we could change this to advertise the real page sizes we support. + */ +#define AMD_IOMMU_PGSIZES (~0xFFFUL) + static DEFINE_RWLOCK(amd_iommu_devtable_lock); /* A list of preallocated protection domains */ @@ -59,6 +80,9 @@ static struct protection_domain *pt_domain; static 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 */ @@ -67,6 +91,7 @@ struct iommu_cmd { }; static void update_domain(struct protection_domain *domain); +static int __init alloc_passthrough_domain(void); /**************************************************************************** * @@ -147,6 +172,33 @@ static struct iommu_dev_data *get_dev_data(struct device *dev) return dev->archdata.iommu; } +static bool pci_iommuv2_capable(struct pci_dev *pdev) +{ + static const int caps[] = { + PCI_EXT_CAP_ID_ATS, + PCI_EXT_CAP_ID_PRI, + PCI_EXT_CAP_ID_PASID, + }; + int i, pos; + + for (i = 0; i < 3; ++i) { + pos = pci_find_ext_capability(pdev, caps[i]); + if (pos == 0) + return false; + } + + return true; +} + +static bool pdev_pri_erratum(struct pci_dev *pdev, u32 erratum) +{ + struct iommu_dev_data *dev_data; + + dev_data = get_dev_data(&pdev->dev); + + return dev_data->errata & (1 << erratum) ? true : false; +} + /* * In this function the list of preallocated protection domains is traversed to * find the domain for a specific device @@ -204,6 +256,7 @@ static bool check_device(struct device *dev) static int iommu_init_device(struct device *dev) { + struct pci_dev *pdev = to_pci_dev(dev); struct iommu_dev_data *dev_data; u16 alias; @@ -228,6 +281,13 @@ static int iommu_init_device(struct device *dev) dev_data->alias_data = alias_data; } + if (pci_iommuv2_capable(pdev)) { + struct amd_iommu *iommu; + + iommu = amd_iommu_rlookup_table[dev_data->devid]; + dev_data->iommu_v2 = iommu->is_iommu_v2; + } + dev->archdata.iommu = dev_data; return 0; @@ -317,6 +377,11 @@ DECLARE_STATS_COUNTER(domain_flush_single); DECLARE_STATS_COUNTER(domain_flush_all); DECLARE_STATS_COUNTER(alloced_io_mem); DECLARE_STATS_COUNTER(total_map_requests); +DECLARE_STATS_COUNTER(complete_ppr); +DECLARE_STATS_COUNTER(invalidate_iotlb); +DECLARE_STATS_COUNTER(invalidate_iotlb_all); +DECLARE_STATS_COUNTER(pri_requests); + static struct dentry *stats_dir; static struct dentry *de_fflush; @@ -351,6 +416,10 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&domain_flush_all); amd_iommu_stats_add(&alloced_io_mem); amd_iommu_stats_add(&total_map_requests); + amd_iommu_stats_add(&complete_ppr); + amd_iommu_stats_add(&invalidate_iotlb); + amd_iommu_stats_add(&invalidate_iotlb_all); + amd_iommu_stats_add(&pri_requests); } #endif @@ -365,8 +434,8 @@ static void dump_dte_entry(u16 devid) { int i; - for (i = 0; i < 8; ++i) - pr_err("AMD-Vi: DTE[%d]: %08x\n", i, + for (i = 0; i < 4; ++i) + pr_err("AMD-Vi: DTE[%d]: %016llx\n", i, amd_iommu_dev_table[devid].data[i]); } @@ -461,12 +530,84 @@ static void iommu_poll_events(struct amd_iommu *iommu) spin_unlock_irqrestore(&iommu->lock, flags); } +static void iommu_handle_ppr_entry(struct amd_iommu *iommu, u32 head) +{ + struct amd_iommu_fault fault; + volatile u64 *raw; + int i; + + INC_STATS_COUNTER(pri_requests); + + 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); + } + + if (PPR_REQ_TYPE(raw[0]) != PPR_REQ_FAULT) { + pr_err_ratelimited("AMD-Vi: Unknown PPR request received\n"); + return; + } + + fault.address = raw[1]; + fault.pasid = PPR_PASID(raw[0]); + fault.device_id = PPR_DEVID(raw[0]); + fault.tag = PPR_TAG(raw[0]); + fault.flags = PPR_FLAGS(raw[0]); + + /* + * To detect the hardware bug we need to clear the entry + * to back to zero. + */ + raw[0] = raw[1] = 0; + + atomic_notifier_call_chain(&ppr_notifier, 0, &fault); +} + +static void iommu_poll_ppr_log(struct amd_iommu *iommu) +{ + unsigned long flags; + u32 head, tail; + + if (iommu->ppr_log == NULL) + return; + + spin_lock_irqsave(&iommu->lock, flags); + + head = readl(iommu->mmio_base + MMIO_PPR_HEAD_OFFSET); + tail = readl(iommu->mmio_base + MMIO_PPR_TAIL_OFFSET); + + while (head != tail) { + + /* Handle PPR entry */ + iommu_handle_ppr_entry(iommu, head); + + /* Update and refresh ring-buffer state*/ + head = (head + PPR_ENTRY_SIZE) % PPR_LOG_SIZE; + writel(head, iommu->mmio_base + MMIO_PPR_HEAD_OFFSET); + tail = readl(iommu->mmio_base + MMIO_PPR_TAIL_OFFSET); + } + + /* enable ppr interrupts again */ + writel(MMIO_STATUS_PPR_INT_MASK, iommu->mmio_base + MMIO_STATUS_OFFSET); + + spin_unlock_irqrestore(&iommu->lock, flags); +} + irqreturn_t amd_iommu_int_thread(int irq, void *data) { struct amd_iommu *iommu; - for_each_iommu(iommu) + for_each_iommu(iommu) { iommu_poll_events(iommu); + iommu_poll_ppr_log(iommu); + } return IRQ_HANDLED; } @@ -595,6 +736,60 @@ static void build_inv_iotlb_pages(struct iommu_cmd *cmd, u16 devid, int qdep, cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK; } +static void build_inv_iommu_pasid(struct iommu_cmd *cmd, u16 domid, int pasid, + u64 address, bool size) +{ + memset(cmd, 0, sizeof(*cmd)); + + address &= ~(0xfffULL); + + cmd->data[0] = pasid & PASID_MASK; + cmd->data[1] = domid; + cmd->data[2] = lower_32_bits(address); + cmd->data[3] = upper_32_bits(address); + cmd->data[2] |= CMD_INV_IOMMU_PAGES_PDE_MASK; + cmd->data[2] |= CMD_INV_IOMMU_PAGES_GN_MASK; + if (size) + cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK; + CMD_SET_TYPE(cmd, CMD_INV_IOMMU_PAGES); +} + +static void build_inv_iotlb_pasid(struct iommu_cmd *cmd, u16 devid, int pasid, + int qdep, u64 address, bool size) +{ + memset(cmd, 0, sizeof(*cmd)); + + address &= ~(0xfffULL); + + cmd->data[0] = devid; + cmd->data[0] |= (pasid & 0xff) << 16; + cmd->data[0] |= (qdep & 0xff) << 24; + cmd->data[1] = devid; + cmd->data[1] |= ((pasid >> 8) & 0xfff) << 16; + cmd->data[2] = lower_32_bits(address); + cmd->data[2] |= CMD_INV_IOMMU_PAGES_GN_MASK; + cmd->data[3] = upper_32_bits(address); + if (size) + cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK; + CMD_SET_TYPE(cmd, CMD_INV_IOTLB_PAGES); +} + +static void build_complete_ppr(struct iommu_cmd *cmd, u16 devid, int pasid, + int status, int tag, bool gn) +{ + memset(cmd, 0, sizeof(*cmd)); + + cmd->data[0] = devid; + if (gn) { + cmd->data[1] = pasid & PASID_MASK; + cmd->data[2] = CMD_INV_IOMMU_PAGES_GN_MASK; + } + cmd->data[3] = tag & 0x1ff; + cmd->data[3] |= (status & PPR_STATUS_MASK) << PPR_STATUS_SHIFT; + + CMD_SET_TYPE(cmd, CMD_COMPLETE_PPR); +} + static void build_inv_all(struct iommu_cmd *cmd) { memset(cmd, 0, sizeof(*cmd)); @@ -1496,6 +1691,48 @@ static void free_pagetable(struct protection_domain *domain) domain->pt_root = NULL; } +static void free_gcr3_tbl_level1(u64 *tbl) +{ + u64 *ptr; + int i; + + for (i = 0; i < 512; ++i) { + if (!(tbl[i] & GCR3_VALID)) + continue; + + ptr = __va(tbl[i] & PAGE_MASK); + + free_page((unsigned long)ptr); + } +} + +static void free_gcr3_tbl_level2(u64 *tbl) +{ + u64 *ptr; + int i; + + for (i = 0; i < 512; ++i) { + if (!(tbl[i] & GCR3_VALID)) + continue; + + ptr = __va(tbl[i] & PAGE_MASK); + + free_gcr3_tbl_level1(ptr); + } +} + +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 if (domain->glx != 0) + BUG(); + + free_page((unsigned long)domain->gcr3_tbl); +} + /* * Free a domain, only used if something went wrong in the * allocation path and we need to free an already allocated page table @@ -1582,20 +1819,52 @@ static bool dma_ops_domain(struct protection_domain *domain) static void set_dte_entry(u16 devid, struct protection_domain *domain, bool ats) { - u64 pte_root = virt_to_phys(domain->pt_root); - u32 flags = 0; + u64 pte_root = 0; + u64 flags = 0; + + if (domain->mode != PAGE_MODE_NONE) + pte_root = virt_to_phys(domain->pt_root); pte_root |= (domain->mode & DEV_ENTRY_MODE_MASK) << DEV_ENTRY_MODE_SHIFT; pte_root |= IOMMU_PTE_IR | IOMMU_PTE_IW | IOMMU_PTE_P | IOMMU_PTE_TV; + flags = amd_iommu_dev_table[devid].data[1]; + if (ats) flags |= DTE_FLAG_IOTLB; - amd_iommu_dev_table[devid].data[3] |= flags; - amd_iommu_dev_table[devid].data[2] = domain->id; - amd_iommu_dev_table[devid].data[1] = upper_32_bits(pte_root); - amd_iommu_dev_table[devid].data[0] = lower_32_bits(pte_root); + if (domain->flags & PD_IOMMUV2_MASK) { + u64 gcr3 = __pa(domain->gcr3_tbl); + u64 glx = domain->glx; + u64 tmp; + + pte_root |= DTE_FLAG_GV; + pte_root |= (glx & DTE_GLX_MASK) << DTE_GLX_SHIFT; + + /* First mask out possible old values for GCR3 table */ + tmp = DTE_GCR3_VAL_B(~0ULL) << DTE_GCR3_SHIFT_B; + flags &= ~tmp; + + tmp = DTE_GCR3_VAL_C(~0ULL) << DTE_GCR3_SHIFT_C; + flags &= ~tmp; + + /* Encode GCR3 table into DTE */ + tmp = DTE_GCR3_VAL_A(gcr3) << DTE_GCR3_SHIFT_A; + pte_root |= tmp; + + tmp = DTE_GCR3_VAL_B(gcr3) << DTE_GCR3_SHIFT_B; + flags |= tmp; + + tmp = DTE_GCR3_VAL_C(gcr3) << DTE_GCR3_SHIFT_C; + flags |= tmp; + } + + flags &= ~(0xffffUL); + flags |= domain->id; + + amd_iommu_dev_table[devid].data[1] = flags; + amd_iommu_dev_table[devid].data[0] = pte_root; } static void clear_dte_entry(u16 devid) @@ -1603,7 +1872,6 @@ static void clear_dte_entry(u16 devid) /* remove entry from the device table seen by the hardware */ amd_iommu_dev_table[devid].data[0] = IOMMU_PTE_P | IOMMU_PTE_TV; amd_iommu_dev_table[devid].data[1] = 0; - amd_iommu_dev_table[devid].data[2] = 0; amd_iommu_apply_erratum_63(devid); } @@ -1696,6 +1964,93 @@ out_unlock: return ret; } + +static void pdev_iommuv2_disable(struct pci_dev *pdev) +{ + pci_disable_ats(pdev); + pci_disable_pri(pdev); + pci_disable_pasid(pdev); +} + +/* FIXME: Change generic reset-function to do the same */ +static int pri_reset_while_enabled(struct pci_dev *pdev) +{ + u16 control; + int pos; + + pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_PRI); + if (!pos) + return -EINVAL; + + 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); + + return 0; +} + +static int pdev_iommuv2_enable(struct pci_dev *pdev) +{ + bool reset_enable; + int reqs, ret; + + /* 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); + + /* Only allow access to user-accessible pages */ + ret = pci_enable_pasid(pdev, 0); + if (ret) + goto out_err; + + /* First reset the PRI state of the device */ + ret = pci_reset_pri(pdev); + if (ret) + goto out_err; + + /* Enable PRI */ + ret = pci_enable_pri(pdev, reqs); + if (ret) + goto out_err; + + if (reset_enable) { + ret = pri_reset_while_enabled(pdev); + if (ret) + goto out_err; + } + + ret = pci_enable_ats(pdev, PAGE_SHIFT); + if (ret) + goto out_err; + + return 0; + +out_err: + pci_disable_pri(pdev); + pci_disable_pasid(pdev); + + return ret; +} + +/* FIX |