summaryrefslogtreecommitdiffstats
path: root/drivers/iommu
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/iommu')
-rw-r--r--drivers/iommu/Kconfig2
-rw-r--r--drivers/iommu/amd_iommu.c98
-rw-r--r--drivers/iommu/amd_iommu_init.c10
-rw-r--r--drivers/iommu/amd_iommu_v2.c24
-rw-r--r--drivers/iommu/arm-smmu-v3.c60
-rw-r--r--drivers/iommu/intel-iommu.c41
-rw-r--r--drivers/iommu/iommu.c2
7 files changed, 125 insertions, 112 deletions
diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
index f1fb1d3ccc56..f491aec95160 100644
--- a/drivers/iommu/Kconfig
+++ b/drivers/iommu/Kconfig
@@ -222,7 +222,7 @@ config TEGRA_IOMMU_SMMU
select IOMMU_API
help
This driver supports the IOMMU hardware (SMMU) found on NVIDIA Tegra
- SoCs (Tegra30 up to Tegra132).
+ SoCs (Tegra30 up to Tegra210).
config EXYNOS_IOMMU
bool "Exynos IOMMU Support"
diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c
index a57e9b749895..658ee39e6569 100644
--- a/drivers/iommu/amd_iommu.c
+++ b/drivers/iommu/amd_iommu.c
@@ -76,8 +76,6 @@ LIST_HEAD(hpet_map);
* Domain for untranslated devices - only allocated
* if iommu=pt passed on kernel cmd line.
*/
-static struct protection_domain *pt_domain;
-
static const struct iommu_ops amd_iommu_ops;
static ATOMIC_NOTIFIER_HEAD(ppr_notifier);
@@ -96,7 +94,7 @@ struct iommu_dev_data {
struct protection_domain *domain; /* Domain the device is bound to */
u16 devid; /* PCI Device ID */
bool iommu_v2; /* Device can make use of IOMMUv2 */
- bool passthrough; /* Default for device is pt_domain */
+ bool passthrough; /* Device is identity mapped */
struct {
bool enabled;
int qdep;
@@ -116,7 +114,6 @@ struct iommu_cmd {
struct kmem_cache *amd_iommu_irq_cache;
static void update_domain(struct protection_domain *domain);
-static int alloc_passthrough_domain(void);
static int protection_domain_init(struct protection_domain *domain);
/****************************************************************************
@@ -2167,15 +2164,17 @@ static int attach_device(struct device *dev,
dev_data = get_dev_data(dev);
if (domain->flags & PD_IOMMUV2_MASK) {
- if (!dev_data->iommu_v2 || !dev_data->passthrough)
+ if (!dev_data->passthrough)
return -EINVAL;
- if (pdev_iommuv2_enable(pdev) != 0)
- return -EINVAL;
+ if (dev_data->iommu_v2) {
+ if (pdev_iommuv2_enable(pdev) != 0)
+ return -EINVAL;
- dev_data->ats.enabled = true;
- dev_data->ats.qdep = pci_ats_queue_depth(pdev);
- dev_data->pri_tlp = pci_pri_tlp_required(pdev);
+ dev_data->ats.enabled = true;
+ dev_data->ats.qdep = pci_ats_queue_depth(pdev);
+ dev_data->pri_tlp = pci_pri_tlp_required(pdev);
+ }
} else if (amd_iommu_iotlb_sup &&
pci_enable_ats(pdev, PAGE_SHIFT) == 0) {
dev_data->ats.enabled = true;
@@ -2221,15 +2220,6 @@ static void __detach_device(struct iommu_dev_data *dev_data)
do_detach(head);
spin_unlock_irqrestore(&domain->lock, flags);
-
- /*
- * If we run in passthrough mode the device must be assigned to the
- * passthrough domain if it is detached from any other domain.
- * Make sure we can deassign from the pt_domain itself.
- */
- if (dev_data->passthrough &&
- (dev_data->domain == NULL && domain != pt_domain))
- __attach_device(dev_data, pt_domain);
}
/*
@@ -2249,7 +2239,7 @@ static void detach_device(struct device *dev)
__detach_device(dev_data);
write_unlock_irqrestore(&amd_iommu_devtable_lock, flags);
- if (domain->flags & PD_IOMMUV2_MASK)
+ 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));
@@ -2287,17 +2277,15 @@ static int amd_iommu_add_device(struct device *dev)
BUG_ON(!dev_data);
- if (dev_data->iommu_v2)
+ if (iommu_pass_through || dev_data->iommu_v2)
iommu_request_dm_for_dev(dev);
/* 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_IDENTITY) {
+ if (domain->type == IOMMU_DOMAIN_IDENTITY)
dev_data->passthrough = true;
- dev->archdata.dma_ops = &nommu_dma_ops;
- } else {
+ else
dev->archdata.dma_ops = &amd_iommu_dma_ops;
- }
out:
iommu_completion_wait(iommu);
@@ -2862,8 +2850,17 @@ int __init amd_iommu_init_api(void)
int __init amd_iommu_init_dma_ops(void)
{
+ swiotlb = iommu_pass_through ? 1 : 0;
iommu_detected = 1;
- swiotlb = 0;
+
+ /*
+ * In case we don't initialize SWIOTLB (actually the common case
+ * when AMD IOMMU is enabled), make sure there are global
+ * dma_ops set as a fall-back for devices not handled by this
+ * driver (for example non-PCI devices).
+ */
+ if (!swiotlb)
+ dma_ops = &nommu_dma_ops;
amd_iommu_stats_init();
@@ -2947,21 +2944,6 @@ out_err:
return NULL;
}
-static int alloc_passthrough_domain(void)
-{
- if (pt_domain != NULL)
- return 0;
-
- /* allocate passthrough domain */
- pt_domain = protection_domain_alloc();
- if (!pt_domain)
- return -ENOMEM;
-
- pt_domain->mode = PAGE_MODE_NONE;
-
- return 0;
-}
-
static struct iommu_domain *amd_iommu_domain_alloc(unsigned type)
{
struct protection_domain *pdomain;
@@ -3222,33 +3204,6 @@ static const struct iommu_ops amd_iommu_ops = {
*
*****************************************************************************/
-int __init amd_iommu_init_passthrough(void)
-{
- struct iommu_dev_data *dev_data;
- struct pci_dev *dev = NULL;
- int ret;
-
- ret = alloc_passthrough_domain();
- if (ret)
- return ret;
-
- for_each_pci_dev(dev) {
- if (!check_device(&dev->dev))
- continue;
-
- dev_data = get_dev_data(&dev->dev);
- dev_data->passthrough = true;
-
- attach_device(&dev->dev, pt_domain);
- }
-
- amd_iommu_stats_init();
-
- pr_info("AMD-Vi: Initialized for Passthrough Mode\n");
-
- return 0;
-}
-
/* IOMMUv2 specific functions */
int amd_iommu_register_ppr_notifier(struct notifier_block *nb)
{
@@ -3363,7 +3318,12 @@ static int __flush_pasid(struct protection_domain *domain, int pasid,
struct amd_iommu *iommu;
int qdep;
- BUG_ON(!dev_data->ats.enabled);
+ /*
+ 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];
diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
index dbda9ae68c5d..a24495eb4e26 100644
--- a/drivers/iommu/amd_iommu_init.c
+++ b/drivers/iommu/amd_iommu_init.c
@@ -2026,14 +2026,6 @@ static bool detect_ivrs(void)
return true;
}
-static int amd_iommu_init_dma(void)
-{
- if (iommu_pass_through)
- return amd_iommu_init_passthrough();
- else
- return amd_iommu_init_dma_ops();
-}
-
/****************************************************************************
*
* AMD IOMMU Initialization State Machine
@@ -2073,7 +2065,7 @@ static int __init state_next(void)
init_state = ret ? IOMMU_INIT_ERROR : IOMMU_INTERRUPTS_EN;
break;
case IOMMU_INTERRUPTS_EN:
- ret = amd_iommu_init_dma();
+ ret = amd_iommu_init_dma_ops();
init_state = ret ? IOMMU_INIT_ERROR : IOMMU_DMA_OPS;
break;
case IOMMU_DMA_OPS:
diff --git a/drivers/iommu/amd_iommu_v2.c b/drivers/iommu/amd_iommu_v2.c
index 3465faf1809e..f7b875bb70d4 100644
--- a/drivers/iommu/amd_iommu_v2.c
+++ b/drivers/iommu/amd_iommu_v2.c
@@ -132,11 +132,19 @@ static struct device_state *get_device_state(u16 devid)
static void free_device_state(struct device_state *dev_state)
{
+ struct iommu_group *group;
+
/*
* First detach device from domain - No more PRI requests will arrive
* from that device after it is unbound from the IOMMUv2 domain.
*/
- iommu_detach_device(dev_state->domain, &dev_state->pdev->dev);
+ group = iommu_group_get(&dev_state->pdev->dev);
+ if (WARN_ON(!group))
+ return;
+
+ iommu_detach_group(dev_state->domain, group);
+
+ iommu_group_put(group);
/* Everything is down now, free the IOMMUv2 domain */
iommu_domain_free(dev_state->domain);
@@ -731,6 +739,7 @@ EXPORT_SYMBOL(amd_iommu_unbind_pasid);
int amd_iommu_init_device(struct pci_dev *pdev, int pasids)
{
struct device_state *dev_state;
+ struct iommu_group *group;
unsigned long flags;
int ret, tmp;
u16 devid;
@@ -776,10 +785,16 @@ int amd_iommu_init_device(struct pci_dev *pdev, int pasids)
if (ret)
goto out_free_domain;
- ret = iommu_attach_device(dev_state->domain, &pdev->dev);
- if (ret != 0)
+ group = iommu_group_get(&pdev->dev);
+ if (!group)
goto out_free_domain;
+ ret = iommu_attach_group(dev_state->domain, group);
+ if (ret != 0)
+ goto out_drop_group;
+
+ iommu_group_put(group);
+
spin_lock_irqsave(&state_lock, flags);
if (__get_device_state(devid) != NULL) {
@@ -794,6 +809,9 @@ int amd_iommu_init_device(struct pci_dev *pdev, int pasids)
return 0;
+out_drop_group:
+ iommu_group_put(group);
+
out_free_domain:
iommu_domain_free(dev_state->domain);
diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c
index 8e9ec81ce4bb..da902baaa794 100644
--- a/drivers/iommu/arm-smmu-v3.c
+++ b/drivers/iommu/arm-smmu-v3.c
@@ -199,9 +199,10 @@
* Stream table.
*
* Linear: Enough to cover 1 << IDR1.SIDSIZE entries
- * 2lvl: 8k L1 entries, 256 lazy entries per table (each table covers a PCI bus)
+ * 2lvl: 128k L1 entries,
+ * 256 lazy entries per table (each table covers a PCI bus)
*/
-#define STRTAB_L1_SZ_SHIFT 16
+#define STRTAB_L1_SZ_SHIFT 20
#define STRTAB_SPLIT 8
#define STRTAB_L1_DESC_DWORDS 1
@@ -269,10 +270,10 @@
#define ARM64_TCR_TG0_SHIFT 14
#define ARM64_TCR_TG0_MASK 0x3UL
#define CTXDESC_CD_0_TCR_IRGN0_SHIFT 8
-#define ARM64_TCR_IRGN0_SHIFT 24
+#define ARM64_TCR_IRGN0_SHIFT 8
#define ARM64_TCR_IRGN0_MASK 0x3UL
#define CTXDESC_CD_0_TCR_ORGN0_SHIFT 10
-#define ARM64_TCR_ORGN0_SHIFT 26
+#define ARM64_TCR_ORGN0_SHIFT 10
#define ARM64_TCR_ORGN0_MASK 0x3UL
#define CTXDESC_CD_0_TCR_SH0_SHIFT 12
#define ARM64_TCR_SH0_SHIFT 12
@@ -542,6 +543,9 @@ struct arm_smmu_device {
#define ARM_SMMU_FEAT_HYP (1 << 12)
u32 features;
+#define ARM_SMMU_OPT_SKIP_PREFETCH (1 << 0)
+ u32 options;
+
struct arm_smmu_cmdq cmdq;
struct arm_smmu_evtq evtq;
struct arm_smmu_priq priq;
@@ -602,11 +606,35 @@ struct arm_smmu_domain {
static DEFINE_SPINLOCK(arm_smmu_devices_lock);
static LIST_HEAD(arm_smmu_devices);
+struct arm_smmu_option_prop {
+ u32 opt;
+ const char *prop;
+};
+
+static struct arm_smmu_option_prop arm_smmu_options[] = {
+ { ARM_SMMU_OPT_SKIP_PREFETCH, "hisilicon,broken-prefetch-cmd" },
+ { 0, NULL},
+};
+
static struct arm_smmu_domain *to_smmu_domain(struct iommu_domain *dom)
{
return container_of(dom, struct arm_smmu_domain, domain);
}
+static void parse_driver_options(struct arm_smmu_device *smmu)
+{
+ int i = 0;
+
+ do {
+ if (of_property_read_bool(smmu->dev->of_node,
+ arm_smmu_options[i].prop)) {
+ smmu->options |= arm_smmu_options[i].opt;
+ dev_notice(smmu->dev, "option %s\n",
+ arm_smmu_options[i].prop);
+ }
+ } while (arm_smmu_options[++i].opt);
+}
+
/* Low-level queue manipulation functions */
static bool queue_full(struct arm_smmu_queue *q)
{
@@ -1036,7 +1064,8 @@ static void arm_smmu_write_strtab_ent(struct arm_smmu_device *smmu, u32 sid,
arm_smmu_sync_ste_for_sid(smmu, sid);
/* It's likely that we'll want to use the new STE soon */
- arm_smmu_cmdq_issue_cmd(smmu, &prefetch_cmd);
+ if (!(smmu->options & ARM_SMMU_OPT_SKIP_PREFETCH))
+ arm_smmu_cmdq_issue_cmd(smmu, &prefetch_cmd);
}
static void arm_smmu_init_bypass_stes(u64 *strtab, unsigned int nent)
@@ -1064,7 +1093,7 @@ static int arm_smmu_init_l2_strtab(struct arm_smmu_device *smmu, u32 sid)
return 0;
size = 1 << (STRTAB_SPLIT + ilog2(STRTAB_STE_DWORDS) + 3);
- strtab = &cfg->strtab[sid >> STRTAB_SPLIT << STRTAB_L1_DESC_DWORDS];
+ strtab = &cfg->strtab[(sid >> STRTAB_SPLIT) * STRTAB_L1_DESC_DWORDS];
desc->span = STRTAB_SPLIT + 1;
desc->l2ptr = dma_zalloc_coherent(smmu->dev, size, &desc->l2ptr_dma,
@@ -2020,21 +2049,23 @@ static int arm_smmu_init_strtab_2lvl(struct arm_smmu_device *smmu)
{
void *strtab;
u64 reg;
- u32 size;
+ u32 size, l1size;
int ret;
struct arm_smmu_strtab_cfg *cfg = &smmu->strtab_cfg;
/* Calculate the L1 size, capped to the SIDSIZE */
size = STRTAB_L1_SZ_SHIFT - (ilog2(STRTAB_L1_DESC_DWORDS) + 3);
size = min(size, smmu->sid_bits - STRTAB_SPLIT);
- if (size + STRTAB_SPLIT < smmu->sid_bits)
+ cfg->num_l1_ents = 1 << size;
+
+ size += STRTAB_SPLIT;
+ if (size < smmu->sid_bits)
dev_warn(smmu->dev,
"2-level strtab only covers %u/%u bits of SID\n",
- size + STRTAB_SPLIT, smmu->sid_bits);
+ size, smmu->sid_bits);
- cfg->num_l1_ents = 1 << size;
- size = cfg->num_l1_ents * (STRTAB_L1_DESC_DWORDS << 3);
- strtab = dma_zalloc_coherent(smmu->dev, size, &cfg->strtab_dma,
+ l1size = cfg->num_l1_ents * (STRTAB_L1_DESC_DWORDS << 3);
+ strtab = dma_zalloc_coherent(smmu->dev, l1size, &cfg->strtab_dma,
GFP_KERNEL);
if (!strtab) {
dev_err(smmu->dev,
@@ -2055,8 +2086,7 @@ static int arm_smmu_init_strtab_2lvl(struct arm_smmu_device *smmu)
ret = arm_smmu_init_l1_strtab(smmu);
if (ret)
dma_free_coherent(smmu->dev,
- cfg->num_l1_ents *
- (STRTAB_L1_DESC_DWORDS << 3),
+ l1size,
strtab,
cfg->strtab_dma);
return ret;
@@ -2573,6 +2603,8 @@ static int arm_smmu_device_dt_probe(struct platform_device *pdev)
if (irq > 0)
smmu->gerr_irq = irq;
+ parse_driver_options(smmu);
+
/* Probe the h/w */
ret = arm_smmu_device_probe(smmu);
if (ret)
diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index a98a7b27aca1..c82ebee6c7e5 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -408,6 +408,10 @@ struct device_domain_info {
struct list_head global; /* link to global list */
u8 bus; /* PCI bus number */
u8 devfn; /* PCI devfn number */
+ struct {
+ u8 enabled:1;
+ u8 qdep;
+ } ats; /* ATS state */
struct device *dev; /* it's NULL for PCIe-to-PCI bridge */
struct intel_iommu *iommu; /* IOMMU used by this device */
struct dmar_domain *domain; /* pointer to domain */
@@ -1391,19 +1395,26 @@ iommu_support_dev_iotlb (struct dmar_domain *domain, struct intel_iommu *iommu,
static void iommu_enable_dev_iotlb(struct device_domain_info *info)
{
+ struct pci_dev *pdev;
+
if (!info || !dev_is_pci(info->dev))
return;
- pci_enable_ats(to_pci_dev(info->dev), VTD_PAGE_SHIFT);
+ pdev = to_pci_dev(info->dev);
+ if (pci_enable_ats(pdev, VTD_PAGE_SHIFT))
+ return;
+
+ info->ats.enabled = 1;
+ info->ats.qdep = pci_ats_queue_depth(pdev);
}
static void iommu_disable_dev_iotlb(struct device_domain_info *info)
{
- if (!info->dev || !dev_is_pci(info->dev) ||
- !pci_ats_enabled(to_pci_dev(info->dev)))
+ if (!info->ats.enabled)
return;
pci_disable_ats(to_pci_dev(info->dev));
+ info->ats.enabled = 0;
}
static void iommu_flush_dev_iotlb(struct dmar_domain *domain,
@@ -1415,16 +1426,11 @@ static void iommu_flush_dev_iotlb(struct dmar_domain *domain,
spin_lock_irqsave(&device_domain_lock, flags);
list_for_each_entry(info, &domain->devices, link) {
- struct pci_dev *pdev;
- if (!info->dev || !dev_is_pci(info->dev))
- continue;
-
- pdev = to_pci_dev(info->dev);
- if (!pci_ats_enabled(pdev))
+ if (!info->ats.enabled)
continue;
sid = info->bus << 8 | info->devfn;
- qdep = pci_ats_queue_depth(pdev);
+ qdep = info->ats.qdep;
qi_flush_dev_iotlb(info->iommu, sid, qdep, addr, mask);
}
spin_unlock_irqrestore(&device_domain_lock, flags);
@@ -1830,8 +1836,9 @@ static int domain_init(struct dmar_domain *domain, int guest_width)
static void domain_exit(struct dmar_domain *domain)
{
+ struct dmar_drhd_unit *drhd;
+ struct intel_iommu *iommu;
struct page *freelist = NULL;
- int i;
/* Domain 0 is reserved, so dont process it */
if (!domain)
@@ -1851,8 +1858,10 @@ static void domain_exit(struct dmar_domain *domain)
/* clear attached or cached domains */
rcu_read_lock();
- for_each_set_bit(i, domain->iommu_bmp, g_num_of_iommus)
- iommu_detach_domain(domain, g_iommus[i]);
+ for_each_active_iommu(iommu, drhd)
+ if (domain_type_is_vm(domain) ||
+ test_bit(iommu->seq_id, domain->iommu_bmp))
+ iommu_detach_domain(domain, iommu);
rcu_read_unlock();
dma_free_pagelist(freelist);
@@ -2094,7 +2103,7 @@ static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
sg_res = aligned_nrpages(sg->offset, sg->length);
sg->dma_address = ((dma_addr_t)iov_pfn << VTD_PAGE_SHIFT) + sg->offset;
sg->dma_length = sg->length;
- pteval = page_to_phys(sg_page(sg)) | prot;
+ pteval = (sg_phys(sg) & PAGE_MASK) | prot;
phys_pfn = pteval >> VTD_PAGE_SHIFT;
}
@@ -2272,6 +2281,8 @@ static struct dmar_domain *dmar_insert_dev_info(struct intel_iommu *iommu,
info->bus = bus;
info->devfn = devfn;
+ info->ats.enabled = 0;
+ info->ats.qdep = 0;
info->dev = dev;
info->domain = domain;
info->iommu = iommu;
@@ -3620,7 +3631,7 @@ static int intel_nontranslate_map_sg(struct device *hddev,
for_each_sg(sglist, sg, nelems, i) {
BUG_ON(!sg_page(sg));
- sg->dma_address = page_to_phys(sg_page(sg)) + sg->offset;
+ sg->dma_address = sg_phys(sg);
sg->dma_length = sg->length;
}
return nelems;
diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
index f286090931cc..049df495c274 100644
--- a/drivers/iommu/iommu.c
+++ b/drivers/iommu/iommu.c
@@ -1408,7 +1408,7 @@ size_t default_iommu_map_sg(struct iommu_domain *domain, unsigned long iova,
min_pagesz = 1 << __ffs(domain->ops->pgsize_bitmap);
for_each_sg(sg, s, nents, i) {
- phys_addr_t phys = page_to_phys(sg_page(s)) + s->offset;
+ phys_addr_t phys = sg_phys(s);
/*
* We are mapping on IOMMU page boundaries, so offset within
generated by cgit v1.2.3 (git 2.39.1) at 2025-01-05 22:40:59 +0100