summaryrefslogtreecommitdiffstats
path: root/drivers/iommu
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/iommu')
-rw-r--r--drivers/iommu/Kconfig44
-rw-r--r--drivers/iommu/Makefile8
-rw-r--r--drivers/iommu/amd_iommu.c4
-rw-r--r--drivers/iommu/dmar.c246
-rw-r--r--drivers/iommu/intel-iommu.c280
-rw-r--r--drivers/iommu/intr_remapping.c93
-rw-r--r--drivers/iommu/iommu.c114
-rw-r--r--drivers/iommu/msm_iommu.c9
-rw-r--r--drivers/iommu/omap-iommu-debug.c418
-rw-r--r--drivers/iommu/omap-iommu.c1245
-rw-r--r--drivers/iommu/omap-iovmm.c742
11 files changed, 2878 insertions, 325 deletions
diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
index b57b3fa492f3..5414253b185a 100644
--- a/drivers/iommu/Kconfig
+++ b/drivers/iommu/Kconfig
@@ -59,10 +59,14 @@ config AMD_IOMMU_STATS
If unsure, say N.
# Intel IOMMU support
-config DMAR
- bool "Support for DMA Remapping Devices"
+config DMAR_TABLE
+ bool
+
+config INTEL_IOMMU
+ bool "Support for Intel IOMMU using DMA Remapping Devices"
depends on PCI_MSI && ACPI && (X86 || IA64_GENERIC)
select IOMMU_API
+ select DMAR_TABLE
help
DMA remapping (DMAR) devices support enables independent address
translations for Direct Memory Access (DMA) from devices.
@@ -70,18 +74,18 @@ config DMAR
and include PCI device scope covered by these DMA
remapping devices.
-config DMAR_DEFAULT_ON
+config INTEL_IOMMU_DEFAULT_ON
def_bool y
- prompt "Enable DMA Remapping Devices by default"
- depends on DMAR
+ prompt "Enable Intel DMA Remapping Devices by default"
+ depends on INTEL_IOMMU
help
Selecting this option will enable a DMAR device at boot time if
one is found. If this option is not selected, DMAR support can
be enabled by passing intel_iommu=on to the kernel.
-config DMAR_BROKEN_GFX_WA
+config INTEL_IOMMU_BROKEN_GFX_WA
bool "Workaround broken graphics drivers (going away soon)"
- depends on DMAR && BROKEN && X86
+ depends on INTEL_IOMMU && BROKEN && X86
---help---
Current Graphics drivers tend to use physical address
for DMA and avoid using DMA APIs. Setting this config
@@ -90,21 +94,41 @@ config DMAR_BROKEN_GFX_WA
to use physical addresses for DMA, at least until this
option is removed in the 2.6.32 kernel.
-config DMAR_FLOPPY_WA
+config INTEL_IOMMU_FLOPPY_WA
def_bool y
- depends on DMAR && X86
+ depends on INTEL_IOMMU && X86
---help---
Floppy disk drivers are known to bypass DMA API calls
thereby failing to work when IOMMU is enabled. This
workaround will setup a 1:1 mapping for the first
16MiB to make floppy (an ISA device) work.
-config INTR_REMAP
+config IRQ_REMAP
bool "Support for Interrupt Remapping (EXPERIMENTAL)"
depends on X86_64 && X86_IO_APIC && PCI_MSI && ACPI && EXPERIMENTAL
+ select DMAR_TABLE
---help---
Supports Interrupt remapping for IO-APIC and MSI devices.
To use x2apic mode in the CPU's which support x2APIC enhancements or
to support platforms with CPU's having > 8 bit APIC ID, say Y.
+# OMAP IOMMU support
+config OMAP_IOMMU
+ bool "OMAP IOMMU Support"
+ depends on ARCH_OMAP
+ select IOMMU_API
+
+config OMAP_IOVMM
+ tristate "OMAP IO Virtual Memory Manager Support"
+ depends on OMAP_IOMMU
+
+config OMAP_IOMMU_DEBUG
+ tristate "Export OMAP IOMMU/IOVMM internals in DebugFS"
+ depends on OMAP_IOVMM && DEBUG_FS
+ help
+ Select this to see extensive information about
+ the internal state of OMAP IOMMU/IOVMM in debugfs.
+
+ Say N unless you know you need this.
+
endif # IOMMU_SUPPORT
diff --git a/drivers/iommu/Makefile b/drivers/iommu/Makefile
index 4d4d77df7cac..2f4448794bc7 100644
--- a/drivers/iommu/Makefile
+++ b/drivers/iommu/Makefile
@@ -1,5 +1,9 @@
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_DMAR) += dmar.o iova.o intel-iommu.o
-obj-$(CONFIG_INTR_REMAP) += dmar.o intr_remapping.o
+obj-$(CONFIG_DMAR_TABLE) += dmar.o
+obj-$(CONFIG_INTEL_IOMMU) += iova.o intel-iommu.o
+obj-$(CONFIG_IRQ_REMAP) += intr_remapping.o
+obj-$(CONFIG_OMAP_IOMMU) += omap-iommu.o
+obj-$(CONFIG_OMAP_IOVMM) += omap-iovmm.o
+obj-$(CONFIG_OMAP_IOMMU_DEBUG) += omap-iommu-debug.o
diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c
index 0e4227f457af..4ee277a8521a 100644
--- a/drivers/iommu/amd_iommu.c
+++ b/drivers/iommu/amd_iommu.c
@@ -1283,7 +1283,7 @@ static int alloc_new_range(struct dma_ops_domain *dma_dom,
if (!pte || !IOMMU_PTE_PRESENT(*pte))
continue;
- dma_ops_reserve_addresses(dma_dom, i << PAGE_SHIFT, 1);
+ dma_ops_reserve_addresses(dma_dom, i >> PAGE_SHIFT, 1);
}
update_domain(&dma_dom->domain);
@@ -2495,7 +2495,7 @@ static unsigned device_dma_ops_init(void)
void __init amd_iommu_init_api(void)
{
- register_iommu(&amd_iommu_ops);
+ bus_set_iommu(&pci_bus_type, &amd_iommu_ops);
}
int __init amd_iommu_init_dma_ops(void)
diff --git a/drivers/iommu/dmar.c b/drivers/iommu/dmar.c
index 6dcc7e2d54de..35c1e17fce1d 100644
--- a/drivers/iommu/dmar.c
+++ b/drivers/iommu/dmar.c
@@ -46,7 +46,7 @@
*/
LIST_HEAD(dmar_drhd_units);
-static struct acpi_table_header * __initdata dmar_tbl;
+struct acpi_table_header * __initdata dmar_tbl;
static acpi_size dmar_tbl_size;
static void __init dmar_register_drhd_unit(struct dmar_drhd_unit *drhd)
@@ -118,8 +118,8 @@ static int __init dmar_parse_one_dev_scope(struct acpi_dmar_device_scope *scope,
return 0;
}
-static int __init dmar_parse_dev_scope(void *start, void *end, int *cnt,
- struct pci_dev ***devices, u16 segment)
+int __init dmar_parse_dev_scope(void *start, void *end, int *cnt,
+ struct pci_dev ***devices, u16 segment)
{
struct acpi_dmar_device_scope *scope;
void * tmp = start;
@@ -217,133 +217,6 @@ static int __init dmar_parse_dev(struct dmar_drhd_unit *dmaru)
return ret;
}
-#ifdef CONFIG_DMAR
-LIST_HEAD(dmar_rmrr_units);
-
-static void __init dmar_register_rmrr_unit(struct dmar_rmrr_unit *rmrr)
-{
- list_add(&rmrr->list, &dmar_rmrr_units);
-}
-
-
-static int __init
-dmar_parse_one_rmrr(struct acpi_dmar_header *header)
-{
- struct acpi_dmar_reserved_memory *rmrr;
- struct dmar_rmrr_unit *rmrru;
-
- rmrru = kzalloc(sizeof(*rmrru), GFP_KERNEL);
- if (!rmrru)
- return -ENOMEM;
-
- rmrru->hdr = header;
- rmrr = (struct acpi_dmar_reserved_memory *)header;
- rmrru->base_address = rmrr->base_address;
- rmrru->end_address = rmrr->end_address;
-
- dmar_register_rmrr_unit(rmrru);
- return 0;
-}
-
-static int __init
-rmrr_parse_dev(struct dmar_rmrr_unit *rmrru)
-{
- struct acpi_dmar_reserved_memory *rmrr;
- int ret;
-
- rmrr = (struct acpi_dmar_reserved_memory *) rmrru->hdr;
- ret = dmar_parse_dev_scope((void *)(rmrr + 1),
- ((void *)rmrr) + rmrr->header.length,
- &rmrru->devices_cnt, &rmrru->devices, rmrr->segment);
-
- if (ret || (rmrru->devices_cnt == 0)) {
- list_del(&rmrru->list);
- kfree(rmrru);
- }
- return ret;
-}
-
-static LIST_HEAD(dmar_atsr_units);
-
-static int __init dmar_parse_one_atsr(struct acpi_dmar_header *hdr)
-{
- struct acpi_dmar_atsr *atsr;
- struct dmar_atsr_unit *atsru;
-
- atsr = container_of(hdr, struct acpi_dmar_atsr, header);
- atsru = kzalloc(sizeof(*atsru), GFP_KERNEL);
- if (!atsru)
- return -ENOMEM;
-
- atsru->hdr = hdr;
- atsru->include_all = atsr->flags & 0x1;
-
- list_add(&atsru->list, &dmar_atsr_units);
-
- return 0;
-}
-
-static int __init atsr_parse_dev(struct dmar_atsr_unit *atsru)
-{
- int rc;
- struct acpi_dmar_atsr *atsr;
-
- if (atsru->include_all)
- return 0;
-
- atsr = container_of(atsru->hdr, struct acpi_dmar_atsr, header);
- rc = dmar_parse_dev_scope((void *)(atsr + 1),
- (void *)atsr + atsr->header.length,
- &atsru->devices_cnt, &atsru->devices,
- atsr->segment);
- if (rc || !atsru->devices_cnt) {
- list_del(&atsru->list);
- kfree(atsru);
- }
-
- return rc;
-}
-
-int dmar_find_matched_atsr_unit(struct pci_dev *dev)
-{
- int i;
- struct pci_bus *bus;
- struct acpi_dmar_atsr *atsr;
- struct dmar_atsr_unit *atsru;
-
- dev = pci_physfn(dev);
-
- list_for_each_entry(atsru, &dmar_atsr_units, list) {
- atsr = container_of(atsru->hdr, struct acpi_dmar_atsr, header);
- if (atsr->segment == pci_domain_nr(dev->bus))
- goto found;
- }
-
- return 0;
-
-found:
- for (bus = dev->bus; bus; bus = bus->parent) {
- struct pci_dev *bridge = bus->self;
-
- if (!bridge || !pci_is_pcie(bridge) ||
- bridge->pcie_type == PCI_EXP_TYPE_PCI_BRIDGE)
- return 0;
-
- if (bridge->pcie_type == PCI_EXP_TYPE_ROOT_PORT) {
- for (i = 0; i < atsru->devices_cnt; i++)
- if (atsru->devices[i] == bridge)
- return 1;
- break;
- }
- }
-
- if (atsru->include_all)
- return 1;
-
- return 0;
-}
-#endif
-
#ifdef CONFIG_ACPI_NUMA
static int __init
dmar_parse_one_rhsa(struct acpi_dmar_header *header)
@@ -484,14 +357,10 @@ parse_dmar_table(void)
ret = dmar_parse_one_drhd(entry_header);
break;
case ACPI_DMAR_TYPE_RESERVED_MEMORY:
-#ifdef CONFIG_DMAR
ret = dmar_parse_one_rmrr(entry_header);
-#endif
break;
case ACPI_DMAR_TYPE_ATSR:
-#ifdef CONFIG_DMAR
ret = dmar_parse_one_atsr(entry_header);
-#endif
break;
case ACPI_DMAR_HARDWARE_AFFINITY:
#ifdef CONFIG_ACPI_NUMA
@@ -557,34 +426,31 @@ dmar_find_matched_drhd_unit(struct pci_dev *dev)
int __init dmar_dev_scope_init(void)
{
+ static int dmar_dev_scope_initialized;
struct dmar_drhd_unit *drhd, *drhd_n;
int ret = -ENODEV;
+ if (dmar_dev_scope_initialized)
+ return dmar_dev_scope_initialized;
+
+ if (list_empty(&dmar_drhd_units))
+ goto fail;
+
list_for_each_entry_safe(drhd, drhd_n, &dmar_drhd_units, list) {
ret = dmar_parse_dev(drhd);
if (ret)
- return ret;
+ goto fail;
}
-#ifdef CONFIG_DMAR
- {
- struct dmar_rmrr_unit *rmrr, *rmrr_n;
- struct dmar_atsr_unit *atsr, *atsr_n;
-
- list_for_each_entry_safe(rmrr, rmrr_n, &dmar_rmrr_units, list) {
- ret = rmrr_parse_dev(rmrr);
- if (ret)
- return ret;
- }
+ ret = dmar_parse_rmrr_atsr_dev();
+ if (ret)
+ goto fail;
- list_for_each_entry_safe(atsr, atsr_n, &dmar_atsr_units, list) {
- ret = atsr_parse_dev(atsr);
- if (ret)
- return ret;
- }
- }
-#endif
+ dmar_dev_scope_initialized = 1;
+ return 0;
+fail:
+ dmar_dev_scope_initialized = ret;
return ret;
}
@@ -611,14 +477,6 @@ int __init dmar_table_init(void)
return -ENODEV;
}
-#ifdef CONFIG_DMAR
- if (list_empty(&dmar_rmrr_units))
- printk(KERN_INFO PREFIX "No RMRR found\n");
-
- if (list_empty(&dmar_atsr_units))
- printk(KERN_INFO PREFIX "No ATSR found\n");
-#endif
-
return 0;
}
@@ -682,9 +540,6 @@ int __init check_zero_address(void)
return 1;
failed:
-#ifdef CONFIG_DMAR
- dmar_disabled = 1;
-#endif
return 0;
}
@@ -696,22 +551,21 @@ int __init detect_intel_iommu(void)
if (ret)
ret = check_zero_address();
{
-#ifdef CONFIG_INTR_REMAP
struct acpi_table_dmar *dmar;
dmar = (struct acpi_table_dmar *) dmar_tbl;
- if (ret && cpu_has_x2apic && dmar->flags & 0x1)
+
+ if (ret && intr_remapping_enabled && cpu_has_x2apic &&
+ dmar->flags & 0x1)
printk(KERN_INFO
- "Queued invalidation will be enabled to support "
- "x2apic and Intr-remapping.\n");
-#endif
-#ifdef CONFIG_DMAR
+ "Queued invalidation will be enabled to support x2apic and Intr-remapping.\n");
+
if (ret && !no_iommu && !iommu_detected && !dmar_disabled) {
iommu_detected = 1;
/* Make sure ACS will be enabled */
pci_request_acs();
}
-#endif
+
#ifdef CONFIG_X86
if (ret)
x86_init.iommu.iommu_init = intel_iommu_init;
@@ -758,7 +612,6 @@ int alloc_iommu(struct dmar_drhd_unit *drhd)
goto err_unmap;
}
-#ifdef CONFIG_DMAR
agaw = iommu_calculate_agaw(iommu);
if (agaw < 0) {
printk(KERN_ERR
@@ -773,7 +626,6 @@ int alloc_iommu(struct dmar_drhd_unit *drhd)
iommu->seq_id);
goto err_unmap;
}
-#endif
iommu->agaw = agaw;
iommu->msagaw = msagaw;
@@ -800,7 +652,7 @@ int alloc_iommu(struct dmar_drhd_unit *drhd)
(unsigned long long)iommu->cap,
(unsigned long long)iommu->ecap);
- spin_lock_init(&iommu->register_lock);
+ raw_spin_lock_init(&iommu->register_lock);
drhd->iommu = iommu;
return 0;
@@ -817,9 +669,7 @@ void free_iommu(struct intel_iommu *iommu)
if (!iommu)
return;
-#ifdef CONFIG_DMAR
free_dmar_iommu(iommu);
-#endif
if (iommu->reg)
iounmap(iommu->reg);
@@ -921,11 +771,11 @@ int qi_submit_sync(struct qi_desc *desc, struct intel_iommu *iommu)
restart:
rc = 0;
- spin_lock_irqsave(&qi->q_lock, flags);
+ raw_spin_lock_irqsave(&qi->q_lock, flags);
while (qi->free_cnt < 3) {
- spin_unlock_irqrestore(&qi->q_lock, flags);
+ raw_spin_unlock_irqrestore(&qi->q_lock, flags);
cpu_relax();
- spin_lock_irqsave(&qi->q_lock, flags);
+ raw_spin_lock_irqsave(&qi->q_lock, flags);
}
index = qi->free_head;
@@ -965,15 +815,15 @@ restart:
if (rc)
break;
- spin_unlock(&qi->q_lock);
+ raw_spin_unlock(&qi->q_lock);
cpu_relax();
- spin_lock(&qi->q_lock);
+ raw_spin_lock(&qi->q_lock);
}
qi->desc_status[index] = QI_DONE;
reclaim_free_desc(qi);
- spin_unlock_irqrestore(&qi->q_lock, flags);
+ raw_spin_unlock_irqrestore(&qi->q_lock, flags);
if (rc == -EAGAIN)
goto restart;
@@ -1062,7 +912,7 @@ void dmar_disable_qi(struct intel_iommu *iommu)
if (!ecap_qis(iommu->ecap))
return;
- spin_lock_irqsave(&iommu->register_lock, flags);
+ raw_spin_lock_irqsave(&iommu->register_lock, flags);
sts = dmar_readq(iommu->reg + DMAR_GSTS_REG);
if (!(sts & DMA_GSTS_QIES))
@@ -1082,7 +932,7 @@ void dmar_disable_qi(struct intel_iommu *iommu)
IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG, readl,
!(sts & DMA_GSTS_QIES), sts);
end:
- spin_unlock_irqrestore(&iommu->register_lock, flags);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
}
/*
@@ -1097,7 +947,7 @@ static void __dmar_enable_qi(struct intel_iommu *iommu)
qi->free_head = qi->free_tail = 0;
qi->free_cnt = QI_LENGTH;
- spin_lock_irqsave(&iommu->register_lock, flags);
+ raw_spin_lock_irqsave(&iommu->register_lock, flags);
/* write zero to the tail reg */
writel(0, iommu->reg + DMAR_IQT_REG);
@@ -1110,7 +960,7 @@ static void __dmar_enable_qi(struct intel_iommu *iommu)
/* Make sure hardware complete it */
IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG, readl, (sts & DMA_GSTS_QIES), sts);
- spin_unlock_irqrestore(&iommu->register_lock, flags);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
}
/*
@@ -1159,7 +1009,7 @@ int dmar_enable_qi(struct intel_iommu *iommu)
qi->free_head = qi->free_tail = 0;
qi->free_cnt = QI_LENGTH;
- spin_lock_init(&qi->q_lock);
+ raw_spin_lock_init(&qi->q_lock);
__dmar_enable_qi(iommu);
@@ -1225,11 +1075,11 @@ void dmar_msi_unmask(struct irq_data *data)
unsigned long flag;
/* unmask it */
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
writel(0, iommu->reg + DMAR_FECTL_REG);
/* Read a reg to force flush the post write */
readl(iommu->reg + DMAR_FECTL_REG);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
void dmar_msi_mask(struct irq_data *data)
@@ -1238,11 +1088,11 @@ void dmar_msi_mask(struct irq_data *data)
struct intel_iommu *iommu = irq_data_get_irq_handler_data(data);
/* mask it */
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
writel(DMA_FECTL_IM, iommu->reg + DMAR_FECTL_REG);
/* Read a reg to force flush the post write */
readl(iommu->reg + DMAR_FECTL_REG);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
void dmar_msi_write(int irq, struct msi_msg *msg)
@@ -1250,11 +1100,11 @@ void dmar_msi_write(int irq, struct msi_msg *msg)
struct intel_iommu *iommu = irq_get_handler_data(irq);
unsigned long flag;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
writel(msg->data, iommu->reg + DMAR_FEDATA_REG);
writel(msg->address_lo, iommu->reg + DMAR_FEADDR_REG);
writel(msg->address_hi, iommu->reg + DMAR_FEUADDR_REG);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
void dmar_msi_read(int irq, struct msi_msg *msg)
@@ -1262,11 +1112,11 @@ void dmar_msi_read(int irq, struct msi_msg *msg)
struct intel_iommu *iommu = irq_get_handler_data(irq);
unsigned long flag;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
msg->data = readl(iommu->reg + DMAR_FEDATA_REG);
msg->address_lo = readl(iommu->reg + DMAR_FEADDR_REG);
msg->address_hi = readl(iommu->reg + DMAR_FEUADDR_REG);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
static int dmar_fault_do_one(struct intel_iommu *iommu, int type,
@@ -1303,7 +1153,7 @@ irqreturn_t dmar_fault(int irq, void *dev_id)
u32 fault_status;
unsigned long flag;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
fault_status = readl(iommu->reg + DMAR_FSTS_REG);
if (fault_status)
printk(KERN_ERR "DRHD: handling fault status reg %x\n",
@@ -1342,7 +1192,7 @@ irqreturn_t dmar_fault(int irq, void *dev_id)
writel(DMA_FRCD_F, iommu->reg + reg +
fault_index * PRIMARY_FAULT_REG_LEN + 12);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
dmar_fault_do_one(iommu, type, fault_reason,
source_id, guest_addr);
@@ -1350,14 +1200,14 @@ irqreturn_t dmar_fault(int irq, void *dev_id)
fault_index++;
if (fault_index >= cap_num_fault_regs(iommu->cap))
fault_index = 0;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
}
clear_rest:
/* clear all the other faults */
fault_status = readl(iommu->reg + DMAR_FSTS_REG);
writel(fault_status, iommu->reg + DMAR_FSTS_REG);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
return IRQ_HANDLED;
}
diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index c621c98c99da..bb161d2fa03c 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -306,6 +306,11 @@ static inline bool dma_pte_present(struct dma_pte *pte)
return (pte->val & 3) != 0;
}
+static inline bool dma_pte_superpage(struct dma_pte *pte)
+{
+ return (pte->val & (1 << 7));
+}
+
static inline int first_pte_in_page(struct dma_pte *pte)
{
return !((unsigned long)pte & ~VTD_PAGE_MASK);
@@ -393,17 +398,20 @@ static long list_size;
static void domain_remove_dev_info(struct dmar_domain *domain);
-#ifdef CONFIG_DMAR_DEFAULT_ON
+#ifdef CONFIG_INTEL_IOMMU_DEFAULT_ON
int dmar_disabled = 0;
#else
int dmar_disabled = 1;
-#endif /*CONFIG_DMAR_DEFAULT_ON*/
+#endif /*CONFIG_INTEL_IOMMU_DEFAULT_ON*/
static int dmar_map_gfx = 1;
static int dmar_forcedac;
static int intel_iommu_strict;
static int intel_iommu_superpage = 1;
+int intel_iommu_gfx_mapped;
+EXPORT_SYMBOL_GPL(intel_iommu_gfx_mapped);
+
#define DUMMY_DEVICE_DOMAIN_INFO ((struct device_domain_info *)(-1))
static DEFINE_SPINLOCK(device_domain_lock);
static LIST_HEAD(device_domain_list);
@@ -577,17 +585,18 @@ static void domain_update_iommu_snooping(struct dmar_domain *domain)
static void domain_update_iommu_superpage(struct dmar_domain *domain)
{
- int i, mask = 0xf;
+ struct dmar_drhd_unit *drhd;
+ struct intel_iommu *iommu = NULL;
+ int mask = 0xf;
if (!intel_iommu_superpage) {
domain->iommu_superpage = 0;
return;
}
- domain->iommu_superpage = 4; /* 1TiB */
-
- for_each_set_bit(i, &domain->iommu_bmp, g_num_of_iommus) {
- mask |= cap_super_page_val(g_iommus[i]->cap);
+ /* set iommu_superpage to the smallest common denominator */
+ for_each_active_iommu(iommu, drhd) {
+ mask &= cap_super_page_val(iommu->cap);
if (!mask) {
break;
}
@@ -730,29 +739,23 @@ out:
}
static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain,
- unsigned long pfn, int large_level)
+ unsigned long pfn, int target_level)
{
int addr_width = agaw_to_width(domain->agaw) - VTD_PAGE_SHIFT;
struct dma_pte *parent, *pte = NULL;
int level = agaw_to_level(domain->agaw);
- int offset, target_level;
+ int offset;
BUG_ON(!domain->pgd);
BUG_ON(addr_width < BITS_PER_LONG && pfn >> addr_width);
parent = domain->pgd;
- /* Search pte */
- if (!large_level)
- target_level = 1;
- else
- target_level = large_level;
-
while (level > 0) {
void *tmp_page;
offset = pfn_level_offset(pfn, level);
pte = &parent[offset];
- if (!large_level && (pte->val & DMA_PTE_LARGE_PAGE))
+ if (!target_level && (dma_pte_superpage(pte) || !dma_pte_present(pte)))
break;
if (level == target_level)
break;
@@ -816,13 +819,14 @@ static struct dma_pte *dma_pfn_level_pte(struct dmar_domain *domain,
}
/* clear last level pte, a tlb flush should be followed */
-static void dma_pte_clear_range(struct dmar_domain *domain,
+static int dma_pte_clear_range(struct dmar_domain *domain,
unsigned long start_pfn,
unsigned long last_pfn)
{
int addr_width = agaw_to_width(domain->agaw) - VTD_PAGE_SHIFT;
unsigned int large_page = 1;
struct dma_pte *first_pte, *pte;
+ int order;
BUG_ON(addr_width < BITS_PER_LONG && start_pfn >> addr_width);
BUG_ON(addr_width < BITS_PER_LONG && last_pfn >> addr_width);
@@ -846,6 +850,9 @@ static void dma_pte_clear_range(struct dmar_domain *domain,
(void *)pte - (void *)first_pte);
} while (start_pfn && start_pfn <= last_pfn);
+
+ order = (large_page - 1) * 9;
+ return order;
}
/* free page table pages. last level pte should already be cleared */
@@ -932,7 +939,7 @@ static void iommu_set_root_entry(struct intel_iommu *iommu)
addr = iommu->root_entry;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
dmar_writeq(iommu->reg + DMAR_RTADDR_REG, virt_to_phys(addr));
writel(iommu->gcmd | DMA_GCMD_SRTP, iommu->reg + DMAR_GCMD_REG);
@@ -941,7 +948,7 @@ static void iommu_set_root_entry(struct intel_iommu *iommu)
IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG,
readl, (sts & DMA_GSTS_RTPS), sts);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
static void iommu_flush_write_buffer(struct intel_iommu *iommu)
@@ -952,14 +959,14 @@ static void iommu_flush_write_buffer(struct intel_iommu *iommu)
if (!rwbf_quirk && !cap_rwbf(iommu->cap))
return;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
writel(iommu->gcmd | DMA_GCMD_WBF, iommu->reg + DMAR_GCMD_REG);
/* Make sure hardware complete it */
IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG,
readl, (!(val & DMA_GSTS_WBFS)), val);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
/* return value determine if we need a write buffer flush */
@@ -986,14 +993,14 @@ static void __iommu_flush_context(struct intel_iommu *iommu,
}
val |= DMA_CCMD_ICC;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
dmar_writeq(iommu->reg + DMAR_CCMD_REG, val);
/* Make sure hardware complete it */
IOMMU_WAIT_OP(iommu, DMAR_CCMD_REG,
dmar_readq, (!(val & DMA_CCMD_ICC)), val);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
/* return value determine if we need a write buffer flush */
@@ -1032,7 +1039,7 @@ static void __iommu_flush_iotlb(struct intel_iommu *iommu, u16 did,
if (cap_write_drain(iommu->cap))
val |= DMA_TLB_WRITE_DRAIN;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
/* Note: Only uses first TLB reg currently */
if (val_iva)
dmar_writeq(iommu->reg + tlb_offset, val_iva);
@@ -1042,7 +1049,7 @@ static void __iommu_flush_iotlb(struct intel_iommu *iommu, u16 did,
IOMMU_WAIT_OP(iommu, tlb_offset + 8,
dmar_readq, (!(val & DMA_TLB_IVT)), val);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
/* check IOTLB invalidation granularity */
if (DMA_TLB_IAIG(val) == 0)
@@ -1158,7 +1165,7 @@ static void iommu_disable_protect_mem_regions(struct intel_iommu *iommu)
u32 pmen;
unsigned long flags;
- spin_lock_irqsave(&iommu->register_lock, flags);
+ raw_spin_lock_irqsave(&iommu->register_lock, flags);
pmen = readl(iommu->reg + DMAR_PMEN_REG);
pmen &= ~DMA_PMEN_EPM;
writel(pmen, iommu->reg + DMAR_PMEN_REG);
@@ -1167,7 +1174,7 @@ static void iommu_disable_protect_mem_regions(struct intel_iommu *iommu)
IOMMU_WAIT_OP(iommu, DMAR_PMEN_REG,
readl, !(pmen & DMA_PMEN_PRS), pmen);
- spin_unlock_irqrestore(&iommu->register_lock, flags);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
}
static int iommu_enable_translation(struct intel_iommu *iommu)
@@ -1175,7 +1182,7 @@ static int iommu_enable_translation(struct intel_iommu *iommu)
u32 sts;
unsigned long flags;
- spin_lock_irqsave(&iommu->register_lock, flags);
+ raw_spin_lock_irqsave(&iommu->register_lock, flags);
iommu->gcmd |= DMA_GCMD_TE;
writel(iommu->gcmd, iommu->reg + DMAR_GCMD_REG);
@@ -1183,7 +1190,7 @@ static int iommu_enable_translation(struct intel_iommu *iommu)
IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG,
readl, (sts & DMA_GSTS_TES), sts);
- spin_unlock_irqrestore(&iommu->register_lock, flags);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
return 0;
}
@@ -1192,7 +1199,7 @@ static int iommu_disable_translation(struct intel_iommu *iommu)
u32 sts;
unsigned long flag;
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
iommu->gcmd &= ~DMA_GCMD_TE;
writel(iommu->gcmd, iommu->reg + DMAR_GCMD_REG);
@@ -1200,7 +1207,7 @@ static int iommu_disable_translation(struct intel_iommu *iommu)
IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG,
readl, (!(sts & DMA_GSTS_TES)), sts);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
return 0;
}
@@ -2150,7 +2157,7 @@ static inline int iommu_prepare_rmrr_dev(struct dmar_rmrr_unit *rmrr,
rmrr->end_address);
}
-#ifdef CONFIG_DMAR_FLOPPY_WA
+#ifdef CONFIG_INTEL_IOMMU_FLOPPY_WA
static inline void iommu_prepare_isa(void)
{
struct pci_dev *pdev;
@@ -2173,7 +2180,7 @@ static inline void iommu_prepare_isa(void)
{
return;
}
-#endif /* !CONFIG_DMAR_FLPY_WA */
+#endif /* !CONFIG_INTEL_IOMMU_FLPY_WA */
static int md_domain_init(struct dmar_domain *domain, int guest_width);
@@ -2484,7 +2491,7 @@ static int __init init_dmars(void)
if (iommu_pass_through)
iommu_identity_mapping |= IDENTMAP_ALL;
-#ifdef CONFIG_DMAR_BROKEN_GFX_WA
+#ifdef CONFIG_INTEL_IOMMU_BROKEN_GFX_WA
iommu_identity_mapping |= IDENTMAP_GFX;
#endif
@@ -3226,9 +3233,6 @@ static void __init init_no_remapping_devices(void)
}
}
- if (dmar_map_gfx)
- return;
-
for_each_drhd_unit(drhd) {
int i;
if (drhd->ignored || drhd->include_all)
@@ -3236,18 +3240,23 @@ static void __init init_no_remapping_devices(void)
for (i = 0; i < drhd->devices_cnt; i++)
if (drhd->devices[i] &&
- !IS_GFX_DEVICE(drhd->devices[i]))
+ !IS_GFX_DEVICE(drhd->devices[i]))
break;
if (i < drhd->devices_cnt)
continue;
- /* bypass IOMMU if it is just for gfx devices */
- drhd->ignored = 1;
- for (i = 0; i < drhd->devices_cnt; i++) {
- if (!drhd->devices[i])
- continue;
- drhd->devices[i]->dev.archdata.iommu = DUMMY_DEVICE_DOMAIN_INFO;
+ /* This IOMMU has *only* gfx devices. Either bypass it or
+ set the gfx_mapped flag, as appropriate */
+ if (dmar_map_gfx) {
+ intel_iommu_gfx_mapped = 1;
+ } else {
+ drhd->ignored = 1;
+ for (i = 0; i < drhd->devices_cnt; i++) {
+ if (!drhd->devices[i])
+ continue;
+ drhd->devices[i]->dev.archdata.iommu = DUMMY_DEVICE_DOMAIN_INFO;
+ }
}
}
}
@@ -3320,7 +3329,7 @@ static int iommu_suspend(void)
for_each_active_iommu(iommu, drhd) {
iommu_disable_translation(iommu);
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
iommu->iommu_state[SR_DMAR_FECTL_REG] =
readl(iommu->reg + DMAR_FECTL_REG);
@@ -3331,7 +3340,7 @@ static int iommu_suspend(void)
iommu->iommu_state[SR_DMAR_FEUADDR_REG] =
readl(iommu->reg + DMAR_FEUADDR_REG);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
return 0;
@@ -3358,7 +3367,7 @@ static void iommu_resume(void)
for_each_active_iommu(iommu, drhd) {
- spin_lock_irqsave(&iommu->register_lock, flag);
+ raw_spin_lock_irqsave(&iommu->register_lock, flag);
writel(iommu->iommu_state[SR_DMAR_FECTL_REG],
iommu->reg + DMAR_FECTL_REG);
@@ -3369,7 +3378,7 @@ static void iommu_resume(void)
writel(iommu->iommu_state[SR_DMAR_FEUADDR_REG],
iommu->reg + DMAR_FEUADDR_REG);
- spin_unlock_irqrestore(&iommu->register_lock, flag);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
}
for_each_active_iommu(iommu, drhd)
@@ -3390,6 +3399,151 @@ static void __init init_iommu_pm_ops(void)
static inline void init_iommu_pm_ops(void) {}
#endif /* CONFIG_PM */
+LIST_HEAD(dmar_rmrr_units);
+
+static void __init dmar_register_rmrr_unit(struct dmar_rmrr_unit *rmrr)
+{
+ list_add(&rmrr->list, &dmar_rmrr_units);
+}
+
+
+int __init dmar_parse_one_rmrr(struct acpi_dmar_header *header)
+{
+ struct acpi_dmar_reserved_memory *rmrr;
+ struct dmar_rmrr_unit *rmrru;
+
+ rmrru = kzalloc(sizeof(*rmrru), GFP_KERNEL);
+ if (!rmrru)
+ return -ENOMEM;
+
+ rmrru->hdr = header;
+ rmrr = (struct acpi_dmar_reserved_memory *)header;
+ rmrru->base_address = rmrr->base_address;
+ rmrru->end_address = rmrr->end_address;
+
+ dmar_register_rmrr_unit(rmrru);
+ return 0;
+}
+
+static int __init
+rmrr_parse_dev(struct dmar_rmrr_unit *rmrru)
+{
+ struct acpi_dmar_reserved_memory *rmrr;
+ int ret;
+
+ rmrr = (struct acpi_dmar_reserved_memory *) rmrru->hdr;
+ ret = dmar_parse_dev_scope((void *)(rmrr + 1),
+ ((void *)rmrr) + rmrr->header.length,
+ &rmrru->devices_cnt, &rmrru->devices, rmrr->segment);
+
+ if (ret || (rmrru->devices_cnt == 0)) {
+ list_del(&rmrru->list);
+ kfree(rmrru);
+ }
+ return ret;
+}
+
+static LIST_HEAD(dmar_atsr_units);
+
+int __init dmar_parse_one_atsr(struct acpi_dmar_header *hdr)
+{
+ struct acpi_dmar_atsr *atsr;
+ struct dmar_atsr_unit *atsru;
+
+ atsr = container_of(hdr, struct acpi_dmar_atsr, header);
+ atsru = kzalloc(sizeof(*atsru), GFP_KERNEL);
+ if (!atsru)
+ return -ENOMEM;
+
+ atsru->hdr = hdr;
+ atsru->include_all = atsr->flags & 0x1;
+
+ list_add(&atsru->list, &dmar_atsr_units);
+
+ return 0;
+}
+
+static int __init atsr_parse_dev(struct dmar_atsr_unit *atsru)
+{
+ int rc;
+ struct acpi_dmar_atsr *atsr;
+
+ if (atsru->include_all)
+ return 0;
+
+ atsr = container_of(atsru->hdr, struct acpi_dmar_atsr, header);
+ rc = dmar_parse_dev_scope((void *)(atsr + 1),
+ (void *)atsr + atsr->header.length,
+ &atsru->devices_cnt, &atsru->devices,
+ atsr->segment);
+ if (rc || !atsru->devices_cnt) {
+ list_del(&atsru->list);
+ kfree(atsru);
+ }
+
+ return rc;
+}
+
+int dmar_find_matched_atsr_unit(struct pci_dev *dev)
+{
+ int i;
+ struct pci_bus *bus;
+ struct acpi_dmar_atsr *atsr;
+ struct dmar_atsr_unit *atsru;
+
+ dev = pci_physfn(dev);
+
+ list_for_each_entry(atsru, &dmar_atsr_units, list) {
+ atsr = container_of(atsru->hdr, struct acpi_dmar_atsr, header);
+ if (atsr->segment == pci_domain_nr(dev->bus))
+ goto found;
+ }
+
+ return 0;
+
+found:
+ for (bus = dev->bus; bus; bus = bus->parent) {
+ struct pci_dev *bridge = bus->self;
+
+ if (!bridge || !pci_is_pcie(bridge) ||
+ bridge->pcie_type == PCI_EXP_TYPE_PCI_BRIDGE)
+ return 0;
+
+ if (bridge->pcie_type == PCI_EXP_TYPE_ROOT_PORT) {
+ for (i = 0; i < atsru->devices_cnt; i++)
+ if (atsru->devices[i] == bridge)
+ return 1;
+ break;
+ }
+ }
+
+ if (atsru->include_all)
+ return 1;
+
+ return 0;
+}
+
+int dmar_parse_rmrr_atsr_dev(void)
+{
+ struct dmar_rmrr_unit *rmrr, *rmrr_n;
+ struct dmar_atsr_unit *atsr, *atsr_n;
+ int ret = 0;
+
+ list_for_each_entry_safe(rmrr, rmrr_n, &dmar_rmrr_units, list) {
+ ret = rmrr_parse_dev(rmrr);
+ if (ret)
+ return ret;
+ }
+
+ list_for_each_entry_safe(atsr, atsr_n, &dmar_atsr_units, list) {
+ ret = atsr_parse_dev(atsr);
+ if (ret)
+ return ret;
+ }
+
+ return ret;
+}
+
/*
* Here we only respond to action of unbound device from driver.
*
@@ -3439,16 +3593,12 @@ int __init intel_iommu_init(void)
return -ENODEV;
}
- if (dmar_dev_scope_init()) {
+ if (dmar_dev_scope_init() < 0) {
if (force_on)
panic("tboot: Failed to initialize DMAR device scope\n");
return -ENODEV;
}
- /*
- * Check the need for DMA-remapping initialization now.
- * Above initialization will also be used by Interrupt-remapping.
- */
if (no_iommu || dmar_disabled)
return -ENODEV;
@@ -3458,6 +3608,12 @@ int __init intel_iommu_init(void)
return -ENODEV;
}
+ if (list_empty(&dmar_rmrr_units))
+ printk(KERN_INFO "DMAR: No RMRR found\n");
+
+ if (list_empty(&dmar_atsr_units))
+ printk(KERN_INFO "DMAR: No ATSR found\n");
+
if (dmar_init_reserved_ranges()) {
if (force_on)
panic("tboot: Failed to reserve iommu ranges\n");
@@ -3486,7 +3642,7 @@ int __init intel_iommu_init(void)
init_iommu_pm_ops();
- register_iommu(&intel_iommu_ops);
+ bus_set_iommu(&pci_bus_type, &intel_iommu_ops);
bus_register_notifier(&pci_bus_type, &device_nb);
@@ -3568,6 +3724,8 @@ static void domain_remove_one_dev_info(struct dmar_domain *domain,
found = 1;
}
+ spin_unlock_irqrestore(&device_domain_lock, flags);
+
if (found == 0) {
unsigned long tmp_flags;
spin_lock_irqsave(&domain->iommu_lock, tmp_flags);
@@ -3584,8 +3742,6 @@ static void domain_remove_one_dev_info(struct dmar_domain *domain,
spin_unlock_irqrestore(&iommu->lock, tmp_flags);
}
}
-
- spin_unlock_irqrestore(&device_domain_lock, flags);
}
static void vm_domain_remove_all_dev_info(struct dmar_domain *domain)
@@ -3739,6 +3895,7 @@ static int intel_iommu_domain_init(struct iommu_domain *domain)
vm_domain_exit(dmar_domain);
return -ENOMEM;
}
+ domain_update_iommu_cap(dmar_domain);
domain->priv = dmar_domain;
return 0;
@@ -3864,14 +4021,15 @@ static int intel_iommu_unmap(struct iommu_domain *domain,
{
struct dmar_domain *dmar_domain = domain->priv;
size_t size = PAGE_SIZE << gfp_order;
+ int order;
- dma_pte_clear_range(dmar_domain, iova >> VTD_PAGE_SHIFT,
+ order = dma_pte_clear_range(dmar_domain, iova >> VTD_PAGE_SHIFT,
(iova + size - 1) >> VTD_PAGE_SHIFT);
if (dmar_domain->max_addr == iova + size)
dmar_domain->max_addr = iova;
- return gfp_order;
+ return order;
}
static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain,
@@ -3950,7 +4108,11 @@ static void __devinit quirk_calpella_no_shadow_gtt(struct pci_dev *dev)
if (!(ggc & GGC_MEMORY_VT_ENABLED)) {
printk(KERN_INFO "DMAR: BIOS has allocated no shadow GTT; disabling IOMMU for graphics\n");
dmar_map_gfx = 0;
- }
+ } else if (dmar_map_gfx) {
+ /* we have to ensure the gfx device is idle before we flush */
+ printk(KERN_INFO "DMAR: Disabling batched IOTLB flush on Ironlake\n");
+ intel_iommu_strict = 1;
+ }
}
DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x0040, quirk_calpella_no_shadow_gtt);
DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x0044, quirk_calpella_no_shadow_gtt);
diff --git a/drivers/iommu/intr_remapping.c b/drivers/iommu/intr_remapping.c
index 1a89d4a2cadf..07c9f189f314 100644
--- a/drivers/iommu/intr_remapping.c
+++ b/drivers/iommu/intr_remapping.c
@@ -21,6 +21,7 @@ int intr_remapping_enabled;
static int disable_intremap;
static int disable_sourceid_checking;
+static int no_x2apic_optout;
static __init int setup_nointremap(char *str)
{
@@ -34,18 +35,26 @@ static __init int setup_intremap(char *str)
if (!str)
return -EINVAL;
- if (!strncmp(str, "on", 2))
- disable_intremap = 0;
- else if (!strncmp(str, "off", 3))
- disable_intremap = 1;
- else if (!strncmp(str, "nosid", 5))
- disable_sourceid_checking = 1;
+ while (*str) {
+ if (!strncmp(str, "on", 2))
+ disable_intremap = 0;
+ else if (!strncmp(str, "off", 3))
+ disable_intremap = 1;
+ else if (!strncmp(str, "nosid", 5))
+ disable_sourceid_checking = 1;
+ else if (!strncmp(str, "no_x2apic_optout", 16))
+ no_x2apic_optout = 1;
+
+ str += strcspn(str, ",");
+ while (*str == ',')
+ str++;
+ }
return 0;
}
early_param("intremap", setup_intremap);
-static DEFINE_SPINLOCK(irq_2_ir_lock);
+static DEFINE_RAW_SPINLOCK(irq_2_ir_lock);
static struct irq_2_iommu *irq_2_iommu(unsigned int irq)
{
@@ -62,12 +71,12 @@ int get_irte(int irq, struct irte *entry)
if (!entry || !irq_iommu)
return -1;
- spin_lock_irqsave(&irq_2_ir_lock, flags);
+ raw_spin_lock_irqsave(&irq_2_ir_lock, flags);
index = irq_iommu->irte_index + irq_iommu->sub_handle;
*entry = *(irq_iommu->iommu->ir_table->base + index);
- spin_unlock_irqrestore(&irq_2_ir_lock, flags);
+ raw_spin_unlock_irqrestore(&irq_2_ir_lock, flags);
return 0;
}
@@ -101,7 +110,7 @@ int alloc_irte(struct intel_iommu *iommu, int irq, u16 count)
return -1;
}
- spin_lock_irqsave(&irq_2_ir_lock, flags);
+ raw_spin_lock_irqsave(&irq_2_ir_lock, flags);
do {
for (i = index; i < index + count; i++)
if (table->base[i].present)
@@ -113,7 +122,7 @@ int alloc_irte(struct intel_iommu *iommu, int irq, u16 count)
index = (index + count) % INTR_REMAP_TABLE_ENTRIES;
if (index == start_index) {
- spin_unlock_irqrestore(&irq_2_ir_lock, flags);
+ raw_spin_unlock_irqrestore(&irq_2_ir_lock, flags);
printk(KERN_ERR "can't allocate an IRTE\n");
return -1;
}
@@ -127,7 +136,7 @@ int alloc_irte(struct intel_iommu *iommu, int irq, u16 count)
irq_iommu->sub_handle = 0;
irq_iommu->irte_mask = mask;
- spin_unlock_irqrestore(&irq_2_ir_lock, flags);
+ raw_spin_unlock_irqrestore(&irq_2_ir_lock, flags);
return index;
}
@@ -152,10 +161,10 @@ int map_irq_to_irte_handle(int irq, u16 *sub_handle)
if (!irq_iommu)
return -1;
- spin_lock_irqsave(&irq_2_ir_lock, flags);
+ raw_spin_lock_irqsave(&irq_2_ir_lock, flags);
*sub_handle = irq_iommu->sub_handle;
index = irq_iommu->irte_index;
- spin_unlock_irqrestore(&irq_2_ir_lock, flags);
+ raw_spin_unlock_irqrestore(&irq_2_ir_lock, flags);
return index;
}
@@ -167,14 +176,14 @@ int set_irte_irq(int irq, struct intel_iommu *iommu, u16 index, u16 subhandle)
if (!irq_iommu)
return -1;
- spin_lock_irqsave(&irq_2_ir_lock, flags);
+ raw_spin_lock_irqsave(&irq_2_ir_lock, flags);
irq_iommu->iommu = iommu;
irq_iommu->irte_index = index;
irq_iommu->sub_handle = subhandle;
irq_iommu->irte_mask = 0;
- spin_unlock_irqrestore(&irq_2_ir_lock, flags);
+ raw_spin_unlock_irqrestore(&irq_2_ir_lock, flags);
return 0;
}
@@ -190,7 +199,7 @@ int modify_irte(int irq, struct irte *irte_modified)
if (!irq_iommu)
return -1;
- spin_lock_irqsave(&irq_2_ir_lock, flags);
+ raw_spin_lock_irqsave(&irq_2_ir_lock, flags);
iommu = irq_iommu->iommu;
@@ -202,7 +211,7 @@ int modify_irte(int irq, struct irte *irte_modified)
__iommu_flush_cache(iommu, irte, sizeof(*irte));
rc = qi_flush_iec(iommu, index, 0);
- spin_unlock_irqrestore(&irq_2_ir_lock, flags);
+ raw_spin_unlock_irqrestore(&irq_2_ir_lock, flags);
return rc;
}
@@ -270,7 +279,7 @@ int free_irte(int irq)
if (!irq_iommu)
return -1;
- spin_lock_irqsave(&irq_2_ir_lock, flags);
+ raw_spin_lock_irqsave(&irq_2_ir_lock, flags);
rc = clear_entries(irq_iommu);
@@ -279,7 +288,7 @@ int free_irte(int irq)
irq_iommu->sub_handle = 0;
irq_iommu->irte_mask = 0;
- spin_unlock_irqrestore(&irq_2_ir_lock, flags);
+ raw_spin_unlock_irqrestore(&irq_2_ir_lock, flags);
return rc;
}
@@ -409,7 +418,7 @@ static void iommu_set_intr_remapping(struct intel_iommu *iommu, int mode)
addr = virt_to_phys((void *)iommu->ir_table->base);
- spin_lock_irqsave(&iommu->register_lock, flags);
+ raw_spin_lock_irqsave(&iommu->register_lock, flags);
dmar_writeq(iommu->reg + DMAR_IRTA_REG,
(addr) | IR_X2APIC_MODE(mode) | INTR_REMAP_TABLE_REG_SIZE);
@@ -420,7 +429,7 @@ static void iommu_set_intr_remapping(struct intel_iommu *iommu, int mode)
IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG,
readl, (sts & DMA_GSTS_IRTPS), sts);
- spin_unlock_irqrestore(&iommu->register_lock, flags);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
/*
* global invalidation of interrupt entry cache before enabling
@@ -428,7 +437,7 @@ static void iommu_set_intr_remapping(struct intel_iommu *iommu, int mode)
*/
qi_global_iec(iommu);
- spin_lock_irqsave(&iommu->register_lock, flags);
+ raw_spin_lock_irqsave(&iommu->register_lock, flags);
/* Enable interrupt-remapping */
iommu->gcmd |= DMA_GCMD_IRE;
@@ -437,7 +446,7 @@ static void iommu_set_intr_remapping(struct intel_iommu *iommu, int mode)
IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG,
readl, (sts & DMA_GSTS_IRES), sts);
- spin_unlock_irqrestore(&iommu->register_lock, flags);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
}
@@ -485,7 +494,7 @@ static void iommu_disable_intr_remapping(struct intel_iommu *iommu)
*/
qi_global_iec(iommu);
- spin_lock_irqsave(&iommu->register_lock, flags);
+ raw_spin_lock_irqsave(&iommu->register_lock, flags);
sts = dmar_readq(iommu->reg + DMAR_GSTS_REG);
if (!(sts & DMA_GSTS_IRES))
@@ -498,7 +507,16 @@ static void iommu_disable_intr_remapping(struct intel_iommu *iommu)
readl, !(sts & DMA_GSTS_IRES), sts);
end:
- spin_unlock_irqrestore(&iommu->register_lock, flags);
+ raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
+}
+
+static int __init dmar_x2apic_optout(void)
+{
+ struct acpi_table_dmar *dmar;
+ dmar = (struct acpi_table_dmar *)dmar_tbl;
+ if (!dmar || no_x2apic_optout)
+ return 0;
+ return dmar->flags & DMAR_X2APIC_OPT_OUT;
}
int __init intr_remapping_supported(void)
@@ -521,16 +539,25 @@ int __init intr_remapping_supported(void)
return 1;
}
-int __init enable_intr_remapping(int eim)
+int __init enable_intr_remapping(void)
{
struct dmar_drhd_unit *drhd;
int setup = 0;
+ int eim = 0;
if (parse_ioapics_under_ir() != 1) {
printk(KERN_INFO "Not enable interrupt remapping\n");
return -1;
}
+ if (x2apic_supported()) {
+ eim = !dmar_x2apic_optout();
+ WARN(!eim, KERN_WARNING
+ "Your BIOS is broken and requested that x2apic be disabled\n"
+ "This will leave your machine vulnerable to irq-injection attacks\n"
+ "Use 'intremap=no_x2apic_optout' to override BIOS request\n");
+ }
+
for_each_drhd_unit(drhd) {
struct intel_iommu *iommu = drhd->iommu;
@@ -606,8 +633,9 @@ int __init enable_intr_remapping(int eim)
goto error;
intr_remapping_enabled = 1;
+ pr_info("Enabled IRQ remapping in %s mode\n", eim ? "x2apic" : "xapic");
- return 0;
+ return eim ? IRQ_REMAP_X2APIC_MODE : IRQ_REMAP_XAPIC_MODE;
error:
/*
@@ -745,6 +773,15 @@ int __init parse_ioapics_under_ir(void)
return ir_supported;
}
+int ir_dev_scope_init(void)
+{
+ if (!intr_remapping_enabled)
+ return 0;
+
+ return dmar_dev_scope_init();
+}
+rootfs_initcall(ir_dev_scope_init);
+
void disable_intr_remapping(void)
{
struct dmar_drhd_unit *drhd;
diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
index 6e6b6a11b3ce..2fb2963df553 100644
--- a/drivers/iommu/iommu.c
+++ b/drivers/iommu/iommu.c
@@ -16,6 +16,8 @@
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
+#include <linux/device.h>
+#include <linux/kernel.h>
#include <linux/bug.h>
#include <linux/types.h>
#include <linux/module.h>
@@ -23,32 +25,78 @@
#include <linux/errno.h>
#include <linux/iommu.h>
-static struct iommu_ops *iommu_ops;
+static void iommu_bus_init(struct bus_type *bus, struct iommu_ops *ops)
+{
+}
-void register_iommu(struct iommu_ops *ops)
+/**
+ * bus_set_iommu - set iommu-callbacks for the bus
+ * @bus: bus.
+ * @ops: the callbacks provided by the iommu-driver
+ *
+ * This function is called by an iommu driver to set the iommu methods
+ * used for a particular bus. Drivers for devices on that bus can use
+ * the iommu-api after these ops are registered.
+ * This special function is needed because IOMMUs are usually devices on
+ * the bus itself, so the iommu drivers are not initialized when the bus
+ * is set up. With this function the iommu-driver can set the iommu-ops
+ * afterwards.
+ */
+int bus_set_iommu(struct bus_type *bus, struct iommu_ops *ops)
{
- if (iommu_ops)
- BUG();
+ if (bus->iommu_ops != NULL)
+ return -EBUSY;
+
+ bus->iommu_ops = ops;
+
+ /* Do IOMMU specific setup for this bus-type */
+ iommu_bus_init(bus, ops);
- iommu_ops = ops;
+ return 0;
}
+EXPORT_SYMBOL_GPL(bus_set_iommu);
-bool iommu_found(void)
+bool iommu_present(struct bus_type *bus)
{
- return iommu_ops != NULL;
+ return bus->iommu_ops != NULL;
}
-EXPORT_SYMBOL_GPL(iommu_found);
+EXPORT_SYMBOL_GPL(iommu_present);
-struct iommu_domain *iommu_domain_alloc(void)
+/**
+ * iommu_set_fault_handler() - set a fault handler for an iommu domain
+ * @domain: iommu domain
+ * @handler: fault handler
+ *
+ * This function should be used by IOMMU users which want to be notified
+ * whenever an IOMMU fault happens.
+ *
+ * The fault handler itself should return 0 on success, and an appropriate
+ * error code otherwise.
+ */
+void iommu_set_fault_handler(struct iommu_domain *domain,
+ iommu_fault_handler_t handler)
+{
+ BUG_ON(!domain);
+
+ domain->handler = handler;
+}
+EXPORT_SYMBOL_GPL(iommu_set_fault_handler);
+
+struct iommu_domain *iommu_domain_alloc(struct bus_type *bus)
{
struct iommu_domain *domain;
int ret;
+ if (bus == NULL || bus->iommu_ops == NULL)
+ return NULL;
+
domain = kmalloc(sizeof(*domain), GFP_KERNEL);
if (!domain)
return NULL;
- ret = iommu_ops->domain_init(domain);
+ domain->ops = bus->iommu_ops;
+
+ ret = domain->ops->domain_init(domain);
if (ret)
goto out_free;
@@ -63,62 +111,78 @@ EXPORT_SYMBOL_GPL(iommu_domain_alloc);
void iommu_domain_free(struct iommu_domain *domain)
{
- iommu_ops->domain_destroy(domain);
+ if (likely(domain->ops->domain_destroy != NULL))
+ domain->ops->domain_destroy(domain);
+
kfree(domain);
}
EXPORT_SYMBOL_GPL(iommu_domain_free);
int iommu_attach_device(struct iommu_domain *domain, struct device *dev)
{
- return iommu_ops->attach_dev(domain, dev);
+ if (unlikely(domain->ops->attach_dev == NULL))
+ return -ENODEV;
+
+ return domain->ops->attach_dev(domain, dev);
}
EXPORT_SYMBOL_GPL(iommu_attach_device);
void iommu_detach_device(struct iommu_domain *domain, struct device *dev)
{
- iommu_ops->detach_dev(domain, dev);
+ if (unlikely(domain->ops->detach_dev == NULL))
+ return;
+
+ domain->ops->detach_dev(domain, dev);
}
EXPORT_SYMBOL_GPL(iommu_detach_device);
phys_addr_t iommu_iova_to_phys(struct iommu_domain *domain,
unsigned long iova)
{
- return iommu_ops->iova_to_phys(domain, iova);
+ if (unlikely(domain->ops->iova_to_phys == NULL))
+ return 0;
+
+ return domain->ops->iova_to_phys(domain, iova);
}
EXPORT_SYMBOL_GPL(iommu_iova_to_phys);
int iommu_domain_has_cap(struct iommu_domain *domain,
unsigned long cap)
{
- return iommu_ops->domain_has_cap(domain, cap);
+ if (unlikely(domain->ops->domain_has_cap == NULL))
+ return 0;
+
+ return domain->ops->domain_has_cap(domain, cap);
}
EXPORT_SYMBOL_GPL(iommu_domain_has_cap);
int iommu_map(struct iommu_domain *domain, unsigned long iova,
phys_addr_t paddr, int gfp_order, int prot)
{
- unsigned long invalid_mask;
size_t size;
- size = 0x1000UL << gfp_order;
- invalid_mask = size - 1;
+ if (unlikely(domain->ops->map == NULL))
+ return -ENODEV;
- BUG_ON((iova | paddr) & invalid_mask);
+ size = PAGE_SIZE << gfp_order;
- return iommu_ops->map(domain, iova, paddr, gfp_order, prot);
+ BUG_ON(!IS_ALIGNED(iova | paddr, size));
+
+ return domain->ops->map(domain, iova, paddr, gfp_order, prot);
}
EXPORT_SYMBOL_GPL(iommu_map);
int iommu_unmap(struct iommu_domain *domain, unsigned long iova, int gfp_order)
{
- unsigned long invalid_mask;
size_t size;
- size = 0x1000UL << gfp_order;
- invalid_mask = size - 1;
+ if (unlikely(domain->ops->unmap == NULL))
+ return -ENODEV;
+
+ size = PAGE_SIZE << gfp_order;
- BUG_ON(iova & invalid_mask);
+ BUG_ON(!IS_ALIGNED(iova, size));
- return iommu_ops->unmap(domain, iova, gfp_order);
+ return domain->ops->unmap(domain, iova, gfp_order);
}
EXPORT_SYMBOL_GPL(iommu_unmap);
diff --git a/drivers/iommu/msm_iommu.c b/drivers/iommu/msm_iommu.c
index 1a584e077c61..5865dd2e28f9 100644
--- a/drivers/iommu/msm_iommu.c
+++ b/drivers/iommu/msm_iommu.c
@@ -543,6 +543,13 @@ static int msm_iommu_unmap(struct iommu_domain *domain, unsigned long va,
}
ret = __flush_iotlb(domain);
+
+ /*
+ * the IOMMU API requires us to return the order of the unmapped
+ * page (on success).
+ */
+ if (!ret)
+ ret = order;
fail:
spin_unlock_irqrestore(&msm_iommu_lock, flags);
return ret;
@@ -721,7 +728,7 @@ static void __init setup_iommu_tex_classes(void)
static int __init msm_iommu_init(void)
{
setup_iommu_tex_classes();
- register_iommu(&msm_iommu_ops);
+ bus_set_iommu(&platform_bus_type, &msm_iommu_ops);
return 0;
}
diff --git a/drivers/iommu/omap-iommu-debug.c b/drivers/iommu/omap-iommu-debug.c
new file mode 100644
index 000000000000..9c192e79f806
--- /dev/null
+++ b/drivers/iommu/omap-iommu-debug.c
@@ -0,0 +1,418 @@
+/*
+ * omap iommu: debugfs interface
+ *
+ * Copyright (C) 2008-2009 Nokia Corporation
+ *
+ * Written by Hiroshi DOYU <Hiroshi.DOYU@nokia.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/err.h>
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+#include <linux/platform_device.h>
+#include <linux/debugfs.h>
+
+#include <plat/iommu.h>
+#include <plat/iovmm.h>
+
+#include <plat/iopgtable.h>
+
+#define MAXCOLUMN 100 /* for short messages */
+
+static DEFINE_MUTEX(iommu_debug_lock);
+
+static struct dentry *iommu_debug_root;
+
+static ssize_t debug_read_ver(struct file *file, char __user *userbuf,
+ size_t count, loff_t *ppos)
+{
+ u32 ver = omap_iommu_arch_version();
+ char buf[MAXCOLUMN], *p = buf;
+
+ p += sprintf(p, "H/W version: %d.%d\n", (ver >> 4) & 0xf , ver & 0xf);
+
+ return simple_read_from_buffer(userbuf, count, ppos, buf, p - buf);
+}
+
+static ssize_t debug_read_regs(struct file *file, char __user *userbuf,
+ size_t count, loff_t *ppos)
+{
+ struct omap_iommu *obj = file->private_data;
+ char *p, *buf;
+ ssize_t bytes;
+
+ buf = kmalloc(count, GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ p = buf;
+
+ mutex_lock(&iommu_debug_lock);
+
+ bytes = omap_iommu_dump_ctx(obj, p, count);
+ bytes = simple_read_from_buffer(userbuf, count, ppos, buf, bytes);
+
+ mutex_unlock(&iommu_debug_lock);
+ kfree(buf);
+
+ return bytes;
+}
+
+static ssize_t debug_read_tlb(struct file *file, char __user *userbuf,
+ size_t count, loff_t *ppos)
+{
+ struct omap_iommu *obj = file->private_data;
+ char *p, *buf;
+ ssize_t bytes, rest;
+
+ buf = kmalloc(count, GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ p = buf;
+
+ mutex_lock(&iommu_debug_lock);
+
+ p += sprintf(p, "%8s %8s\n", "cam:", "ram:");
+ p += sprintf(p, "-----------------------------------------\n");
+ rest = count - (p - buf);
+ p += omap_dump_tlb_entries(obj, p, rest);
+
+ bytes = simple_read_from_buffer(userbuf, count, ppos, buf, p - buf);
+
+ mutex_unlock(&iommu_debug_lock);
+ kfree(buf);
+
+ return bytes;
+}
+
+static ssize_t debug_write_pagetable(struct file *file,
+ const char __user *userbuf, size_t count, loff_t *ppos)
+{
+ struct iotlb_entry e;
+ struct cr_regs cr;
+ int err;
+ struct omap_iommu *obj = file->private_data;
+ char buf[MAXCOLUMN], *p = buf;
+
+ count = min(count, sizeof(buf));
+
+ mutex_lock(&iommu_debug_lock);
+ if (copy_from_user(p, userbuf, count)) {
+ mutex_unlock(&iommu_debug_lock);
+ return -EFAULT;
+ }
+
+ sscanf(p, "%x %x", &cr.cam, &cr.ram);
+ if (!cr.cam || !cr.ram) {
+ mutex_unlock(&iommu_debug_lock);
+ return -EINVAL;
+ }
+
+ omap_iotlb_cr_to_e(&cr, &e);
+ err = omap_iopgtable_store_entry(obj, &e);
+ if (err)
+ dev_err(obj->dev, "%s: fail to store cr\n", __func__);
+
+ mutex_unlock(&iommu_debug_lock);
+ return count;
+}
+
+#define dump_ioptable_entry_one(lv, da, val) \
+ ({ \
+ int __err = 0; \
+ ssize_t bytes; \
+ const int maxcol = 22; \
+ const char *str = "%d: %08x %08x\n"; \
+ bytes = snprintf(p, maxcol, str, lv, da, val); \
+ p += bytes; \
+ len -= bytes; \
+ if (len < maxcol) \
+ __err = -ENOMEM; \
+ __err; \
+ })
+
+static ssize_t dump_ioptable(struct omap_iommu *obj, char *buf, ssize_t len)
+{
+ int i;
+ u32 *iopgd;
+ char *p = buf;
+
+ spin_lock(&obj->page_table_lock);
+
+ iopgd = iopgd_offset(obj, 0);
+ for (i = 0; i < PTRS_PER_IOPGD; i++, iopgd++) {
+ int j, err;
+ u32 *iopte;
+ u32 da;
+
+ if (!*iopgd)
+ continue;
+
+ if (!(*iopgd & IOPGD_TABLE)) {
+ da = i << IOPGD_SHIFT;
+
+ err = dump_ioptable_entry_one(1, da, *iopgd);
+ if (err)
+ goto out;
+ continue;
+ }
+
+ iopte = iopte_offset(iopgd, 0);
+
+ for (j = 0; j < PTRS_PER_IOPTE; j++, iopte++) {
+ if (!*iopte)
+ continue;
+
+ da = (i << IOPGD_SHIFT) + (j << IOPTE_SHIFT);
+ err = dump_ioptable_entry_one(2, da, *iopgd);
+ if (err)
+ goto out;
+ }
+ }
+out:
+ spin_unlock(&obj->page_table_lock);
+
+ return p - buf;
+}
+
+static ssize_t debug_read_pagetable(struct file *file, char __user *userbuf,
+ size_t count, loff_t *ppos)
+{
+ struct omap_iommu *obj = file->private_data;
+ char *p, *buf;
+ size_t bytes;
+
+ buf = (char *)__get_free_page(GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ p = buf;
+
+ p += sprintf(p, "L: %8s %8s\n", "da:", "pa:");
+ p += sprintf(p, "-----------------------------------------\n");
+
+ mutex_lock(&iommu_debug_lock);
+
+ bytes = PAGE_SIZE - (p - buf);
+ p += dump_ioptable(obj, p, bytes);
+
+ bytes = simple_read_from_buffer(userbuf, count, ppos, buf, p - buf);
+
+ mutex_unlock(&iommu_debug_lock);
+ free_page((unsigned long)buf);
+
+ return bytes;
+}
+
+static ssize_t debug_read_mmap(struct file *file, char __user *userbuf,
+ size_t count, loff_t *ppos)
+{
+ struct omap_iommu *obj = file->private_data;
+ char *p, *buf;
+ struct iovm_struct *tmp;
+ int uninitialized_var(i);
+ ssize_t bytes;
+
+ buf = (char *)__get_free_page(GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ p = buf;
+
+ p += sprintf(p, "%-3s %-8s %-8s %6s %8s\n",
+ "No", "start", "end", "size", "flags");
+ p += sprintf(p, "-------------------------------------------------\n");
+
+ mutex_lock(&iommu_debug_lock);
+
+ list_for_each_entry(tmp, &obj->mmap, list) {
+ size_t len;
+ const char *str = "%3d %08x-%08x %6x %8x\n";
+ const int maxcol = 39;
+
+ len = tmp->da_end - tmp->da_start;
+ p += snprintf(p, maxcol, str,
+ i, tmp->da_start, tmp->da_end, len, tmp->flags);
+
+ if (PAGE_SIZE - (p - buf) < maxcol)
+ break;
+ i++;
+ }
+
+ bytes = simple_read_from_buffer(userbuf, count, ppos, buf, p - buf);
+
+ mutex_unlock(&iommu_debug_lock);
+ free_page((unsigned long)buf);
+
+ return bytes;
+}
+
+static ssize_t debug_read_mem(struct file *file, char __user *userbuf,
+ size_t count, loff_t *ppos)
+{
+ struct omap_iommu *obj = file->private_data;
+ char *p, *buf;
+ struct iovm_struct *area;
+ ssize_t bytes;
+
+ count = min_t(ssize_t, count, PAGE_SIZE);
+
+ buf = (char *)__get_free_page(GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ p = buf;
+
+ mutex_lock(&iommu_debug_lock);
+
+ area = omap_find_iovm_area(obj, (u32)ppos);
+ if (IS_ERR(area)) {
+ bytes = -EINVAL;
+ goto err_out;
+ }
+ memcpy(p, area->va, count);
+ p += count;
+
+ bytes = simple_read_from_buffer(userbuf, count, ppos, buf, p - buf);
+err_out:
+ mutex_unlock(&iommu_debug_lock);
+ free_page((unsigned long)buf);
+
+ return bytes;
+}
+
+static ssize_t debug_write_mem(struct file *file, const char __user *userbuf,
+ size_t count, loff_t *ppos)
+{
+ struct omap_iommu *obj = file->private_data;
+ struct iovm_struct *area;
+ char *p, *buf;
+
+ count = min_t(size_t, count, PAGE_SIZE);
+
+ buf = (char *)__get_free_page(GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ p = buf;
+
+ mutex_lock(&iommu_debug_lock);
+
+ if (copy_from_user(p, userbuf, count)) {
+ count = -EFAULT;
+ goto err_out;
+ }
+
+ area = omap_find_iovm_area(obj, (u32)ppos);
+ if (IS_ERR(area)) {
+ count = -EINVAL;
+ goto err_out;
+ }
+ memcpy(area->va, p, count);
+err_out:
+ mutex_unlock(&iommu_debug_lock);
+ free_page((unsigned long)buf);
+
+ return count;
+}
+
+static int debug_open_generic(struct inode *inode, struct file *file)
+{
+ file->private_data = inode->i_private;
+ return 0;
+}
+
+#define DEBUG_FOPS(name) \
+ static const struct file_operations debug_##name##_fops = { \
+ .open = debug_open_generic, \
+ .read = debug_read_##name, \
+ .write = debug_write_##name, \
+ .llseek = generic_file_llseek, \
+ };
+
+#define DEBUG_FOPS_RO(name) \
+ static const struct file_operations debug_##name##_fops = { \
+ .open = debug_open_generic, \
+ .read = debug_read_##name, \
+ .llseek = generic_file_llseek, \
+ };
+
+DEBUG_FOPS_RO(ver);
+DEBUG_FOPS_RO(regs);
+DEBUG_FOPS_RO(tlb);
+DEBUG_FOPS(pagetable);
+DEBUG_FOPS_RO(mmap);
+DEBUG_FOPS(mem);
+
+#define __DEBUG_ADD_FILE(attr, mode) \
+ { \
+ struct dentry *dent; \
+ dent = debugfs_create_file(#attr, mode, parent, \
+ obj, &debug_##attr##_fops); \
+ if (!dent) \
+ return -ENOMEM; \
+ }
+
+#define DEBUG_ADD_FILE(name) __DEBUG_ADD_FILE(name, 600)
+#define DEBUG_ADD_FILE_RO(name) __DEBUG_ADD_FILE(name, 400)
+
+static int iommu_debug_register(struct device *dev, void *data)
+{
+ struct platform_device *pdev = to_platform_device(dev);
+ struct omap_iommu *obj = platform_get_drvdata(pdev);
+ struct dentry *d, *parent;
+
+ if (!obj || !obj->dev)
+ return -EINVAL;
+
+ d = debugfs_create_dir(obj->name, iommu_debug_root);
+ if (!d)
+ return -ENOMEM;
+ parent = d;
+
+ d = debugfs_create_u8("nr_tlb_entries", 400, parent,
+ (u8 *)&obj->nr_tlb_entries);
+ if (!d)
+ return -ENOMEM;
+
+ DEBUG_ADD_FILE_RO(ver);
+ DEBUG_ADD_FILE_RO(regs);
+ DEBUG_ADD_FILE_RO(tlb);
+ DEBUG_ADD_FILE(pagetable);
+ DEBUG_ADD_FILE_RO(mmap);
+ DEBUG_ADD_FILE(mem);
+
+ return 0;
+}
+
+static int __init iommu_debug_init(void)
+{
+ struct dentry *d;
+ int err;
+
+ d = debugfs_create_dir("iommu", NULL);
+ if (!d)
+ return -ENOMEM;
+ iommu_debug_root = d;
+
+ err = omap_foreach_iommu_device(d, iommu_debug_register);
+ if (err)
+ goto err_out;
+ return 0;
+
+err_out:
+ debugfs_remove_recursive(iommu_debug_root);
+ return err;
+}
+module_init(iommu_debug_init)
+
+static void __exit iommu_debugfs_exit(void)
+{
+ debugfs_remove_recursive(iommu_debug_root);
+}
+module_exit(iommu_debugfs_exit)
+
+MODULE_DESCRIPTION("omap iommu: debugfs interface");
+MODULE_AUTHOR("Hiroshi DOYU <Hiroshi.DOYU@nokia.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/iommu/omap-iommu.c b/drivers/iommu/omap-iommu.c
new file mode 100644
index 000000000000..8f32b2bf7587
--- /dev/null
+++ b/drivers/iommu/omap-iommu.c
@@ -0,0 +1,1245 @@
+/*
+ * omap iommu: tlb and pagetable primitives
+ *
+ * Copyright (C) 2008-2010 Nokia Corporation
+ *
+ * Written by Hiroshi DOYU <Hiroshi.DOYU@nokia.com>,
+ * Paul Mundt and Toshihiro Kobayashi
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/err.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/interrupt.h>
+#include <linux/ioport.h>
+#include <linux/clk.h>
+#include <linux/platform_device.h>
+#include <linux/iommu.h>
+#include <linux/mutex.h>
+#include <linux/spinlock.h>
+
+#include <asm/cacheflush.h>
+
+#include <plat/iommu.h>
+
+#include <plat/iopgtable.h>
+
+#define for_each_iotlb_cr(obj, n, __i, cr) \
+ for (__i = 0; \
+ (__i < (n)) && (cr = __iotlb_read_cr((obj), __i), true); \
+ __i++)
+
+/**
+ * struct omap_iommu_domain - omap iommu domain
+ * @pgtable: the page table
+ * @iommu_dev: an omap iommu device attached to this domain. only a single
+ * iommu device can be attached for now.
+ * @lock: domain lock, should be taken when attaching/detaching
+ */
+struct omap_iommu_domain {
+ u32 *pgtable;
+ struct omap_iommu *iommu_dev;
+ spinlock_t lock;
+};
+
+/* accommodate the difference between omap1 and omap2/3 */
+static const struct iommu_functions *arch_iommu;
+
+static struct platform_driver omap_iommu_driver;
+static struct kmem_cache *iopte_cachep;
+
+/**
+ * omap_install_iommu_arch - Install archtecure specific iommu functions
+ * @ops: a pointer to architecture specific iommu functions
+ *
+ * There are several kind of iommu algorithm(tlb, pagetable) among
+ * omap series. This interface installs such an iommu algorighm.
+ **/
+int omap_install_iommu_arch(const struct iommu_functions *ops)
+{
+ if (arch_iommu)
+ return -EBUSY;
+
+ arch_iommu = ops;
+ return 0;
+}
+EXPORT_SYMBOL_GPL(omap_install_iommu_arch);
+
+/**
+ * omap_uninstall_iommu_arch - Uninstall archtecure specific iommu functions
+ * @ops: a pointer to architecture specific iommu functions
+ *
+ * This interface uninstalls the iommu algorighm installed previously.
+ **/
+void omap_uninstall_iommu_arch(const struct iommu_functions *ops)
+{
+ if (arch_iommu != ops)
+ pr_err("%s: not your arch\n", __func__);
+
+ arch_iommu = NULL;
+}
+EXPORT_SYMBOL_GPL(omap_uninstall_iommu_arch);
+
+/**
+ * omap_iommu_save_ctx - Save registers for pm off-mode support
+ * @obj: target iommu
+ **/
+void omap_iommu_save_ctx(struct omap_iommu *obj)
+{
+ arch_iommu->save_ctx(obj);
+}
+EXPORT_SYMBOL_GPL(omap_iommu_save_ctx);
+
+/**
+ * omap_iommu_restore_ctx - Restore registers for pm off-mode support
+ * @obj: target iommu
+ **/
+void omap_iommu_restore_ctx(struct omap_iommu *obj)
+{
+ arch_iommu->restore_ctx(obj);
+}
+EXPORT_SYMBOL_GPL(omap_iommu_restore_ctx);
+
+/**
+ * omap_iommu_arch_version - Return running iommu arch version
+ **/
+u32 omap_iommu_arch_version(void)
+{
+ return arch_iommu->version;
+}
+EXPORT_SYMBOL_GPL(omap_iommu_arch_version);
+
+static int iommu_enable(struct omap_iommu *obj)
+{
+ int err;
+
+ if (!obj)
+ return -EINVAL;
+
+ if (!arch_iommu)
+ return -ENODEV;
+
+ clk_enable(obj->clk);
+
+ err = arch_iommu->enable(obj);
+
+ clk_disable(obj->clk);
+ return err;
+}
+
+static void iommu_disable(struct omap_iommu *obj)
+{
+ if (!obj)
+ return;
+
+ clk_enable(obj->clk);
+
+ arch_iommu->disable(obj);
+
+ clk_disable(obj->clk);
+}
+
+/*
+ * TLB operations
+ */
+void omap_iotlb_cr_to_e(struct cr_regs *cr, struct iotlb_entry *e)
+{
+ BUG_ON(!cr || !e);
+
+ arch_iommu->cr_to_e(cr, e);
+}
+EXPORT_SYMBOL_GPL(omap_iotlb_cr_to_e);
+
+static inline int iotlb_cr_valid(struct cr_regs *cr)
+{
+ if (!cr)
+ return -EINVAL;
+
+ return arch_iommu->cr_valid(cr);
+}
+
+static inline struct cr_regs *iotlb_alloc_cr(struct omap_iommu *obj,
+ struct iotlb_entry *e)
+{
+ if (!e)
+ return NULL;
+
+ return arch_iommu->alloc_cr(obj, e);
+}
+
+static u32 iotlb_cr_to_virt(struct cr_regs *cr)
+{
+ return arch_iommu->cr_to_virt(cr);
+}
+
+static u32 get_iopte_attr(struct iotlb_entry *e)
+{
+ return arch_iommu->get_pte_attr(e);
+}
+
+static u32 iommu_report_fault(struct omap_iommu *obj, u32 *da)
+{
+ return arch_iommu->fault_isr(obj, da);
+}
+
+static void iotlb_lock_get(struct omap_iommu *obj, struct iotlb_lock *l)
+{
+ u32 val;
+
+ val = iommu_read_reg(obj, MMU_LOCK);
+
+ l->base = MMU_LOCK_BASE(val);
+ l->vict = MMU_LOCK_VICT(val);
+
+}
+
+static void iotlb_lock_set(struct omap_iommu *obj, struct iotlb_lock *l)
+{
+ u32 val;
+
+ val = (l->base << MMU_LOCK_BASE_SHIFT);
+ val |= (l->vict << MMU_LOCK_VICT_SHIFT);
+
+ iommu_write_reg(obj, val, MMU_LOCK);
+}
+
+static void iotlb_read_cr(struct omap_iommu *obj, struct cr_regs *cr)
+{
+ arch_iommu->tlb_read_cr(obj, cr);
+}
+
+static void iotlb_load_cr(struct omap_iommu *obj, struct cr_regs *cr)
+{
+ arch_iommu->tlb_load_cr(obj, cr);
+
+ iommu_write_reg(obj, 1, MMU_FLUSH_ENTRY);
+ iommu_write_reg(obj, 1, MMU_LD_TLB);
+}
+
+/**
+ * iotlb_dump_cr - Dump an iommu tlb entry into buf
+ * @obj: target iommu
+ * @cr: contents of cam and ram register
+ * @buf: output buffer
+ **/
+static inline ssize_t iotlb_dump_cr(struct omap_iommu *obj, struct cr_regs *cr,
+ char *buf)
+{
+ BUG_ON(!cr || !buf);
+
+ return arch_iommu->dump_cr(obj, cr, buf);
+}
+
+/* only used in iotlb iteration for-loop */
+static struct cr_regs __iotlb_read_cr(struct omap_iommu *obj, int n)
+{
+ struct cr_regs cr;
+ struct iotlb_lock l;
+
+ iotlb_lock_get(obj, &l);
+ l.vict = n;
+ iotlb_lock_set(obj, &l);
+ iotlb_read_cr(obj, &cr);
+
+ return cr;
+}
+
+/**
+ * load_iotlb_entry - Set an iommu tlb entry
+ * @obj: target iommu
+ * @e: an iommu tlb entry info
+ **/
+#ifdef PREFETCH_IOTLB
+static int load_iotlb_entry(struct omap_iommu *obj, struct iotlb_entry *e)
+{
+ int err = 0;
+ struct iotlb_lock l;
+ struct cr_regs *cr;
+
+ if (!obj || !obj->nr_tlb_entries || !e)
+ return -EINVAL;
+
+ clk_enable(obj->clk);
+
+ iotlb_lock_get(obj, &l);
+ if (l.base == obj->nr_tlb_entries) {
+ dev_warn(obj->dev, "%s: preserve entries full\n", __func__);
+ err = -EBUSY;
+ goto out;
+ }
+ if (!e->prsvd) {
+ int i;
+ struct cr_regs tmp;
+
+ for_each_iotlb_cr(obj, obj->nr_tlb_entries, i, tmp)
+ if (!iotlb_cr_valid(&tmp))
+ break;
+
+ if (i == obj->nr_tlb_entries) {
+ dev_dbg(obj->dev, "%s: full: no entry\n", __func__);
+ err = -EBUSY;
+ goto out;
+ }
+
+ iotlb_lock_get(obj, &l);
+ } else {
+ l.vict = l.base;
+ iotlb_lock_set(obj, &l);
+ }
+
+ cr = iotlb_alloc_cr(obj, e);
+ if (IS_ERR(cr)) {
+ clk_disable(obj->clk);
+ return PTR_ERR(cr);
+ }
+
+ iotlb_load_cr(obj, cr);
+ kfree(cr);
+
+ if (e->prsvd)
+ l.base++;
+ /* increment victim for next tlb load */
+ if (++l.vict == obj->nr_tlb_entries)
+ l.vict = l.base;
+ iotlb_lock_set(obj, &l);
+out:
+ clk_disable(obj->clk);
+ return err;
+}
+
+#else /* !PREFETCH_IOTLB */
+
+static int load_iotlb_entry(struct omap_iommu *obj, struct iotlb_entry *e)
+{
+ return 0;
+}
+
+#endif /* !PREFETCH_IOTLB */
+
+static int prefetch_iotlb_entry(struct omap_iommu *obj, struct iotlb_entry *e)
+{
+ return load_iotlb_entry(obj, e);
+}
+
+/**
+ * flush_iotlb_page - Clear an iommu tlb entry
+ * @obj: target iommu
+ * @da: iommu device virtual address
+ *
+ * Clear an iommu tlb entry which includes 'da' address.
+ **/
+static void flush_iotlb_page(struct omap_iommu *obj, u32 da)
+{
+ int i;
+ struct cr_regs cr;
+
+ clk_enable(obj->clk);
+
+ for_each_iotlb_cr(obj, obj->nr_tlb_entries, i, cr) {
+ u32 start;
+ size_t bytes;
+
+ if (!iotlb_cr_valid(&cr))
+ continue;
+
+ start = iotlb_cr_to_virt(&cr);
+ bytes = iopgsz_to_bytes(cr.cam & 3);
+
+ if ((start <= da) && (da < start + bytes)) {
+ dev_dbg(obj->dev, "%s: %08x<=%08x(%x)\n",
+ __func__, start, da, bytes);
+ iotlb_load_cr(obj, &cr);
+ iommu_write_reg(obj, 1, MMU_FLUSH_ENTRY);
+ }
+ }
+ clk_disable(obj->clk);
+
+ if (i == obj->nr_tlb_entries)
+ dev_dbg(obj->dev, "%s: no page for %08x\n", __func__, da);
+}
+
+/**
+ * flush_iotlb_all - Clear all iommu tlb entries
+ * @obj: target iommu
+ **/
+static void flush_iotlb_all(struct omap_iommu *obj)
+{
+ struct iotlb_lock l;
+
+ clk_enable(obj->clk);
+
+ l.base = 0;
+ l.vict = 0;
+ iotlb_lock_set(obj, &l);
+
+ iommu_write_reg(obj, 1, MMU_GFLUSH);
+
+ clk_disable(obj->clk);
+}
+
+#if defined(CONFIG_OMAP_IOMMU_DEBUG) || defined(CONFIG_OMAP_IOMMU_DEBUG_MODULE)
+
+ssize_t omap_iommu_dump_ctx(struct omap_iommu *obj, char *buf, ssize_t bytes)
+{
+ if (!obj || !buf)
+ return -EINVAL;
+
+ clk_enable(obj->clk);
+
+ bytes = arch_iommu->dump_ctx(obj, buf, bytes);
+
+ clk_disable(obj->clk);
+
+ return bytes;
+}
+EXPORT_SYMBOL_GPL(omap_iommu_dump_ctx);
+
+static int
+__dump_tlb_entries(struct omap_iommu *obj, struct cr_regs *crs, int num)
+{
+ int i;
+ struct iotlb_lock saved;
+ struct cr_regs tmp;
+ struct cr_regs *p = crs;
+
+ clk_enable(obj->clk);
+ iotlb_lock_get(obj, &saved);
+
+ for_each_iotlb_cr(obj, num, i, tmp) {
+ if (!iotlb_cr_valid(&tmp))
+ continue;
+ *p++ = tmp;
+ }
+
+ iotlb_lock_set(obj, &saved);
+ clk_disable(obj->clk);
+
+ return p - crs;
+}
+
+/**
+ * omap_dump_tlb_entries - dump cr arrays to given buffer
+ * @obj: target iommu
+ * @buf: output buffer
+ **/
+size_t omap_dump_tlb_entries(struct omap_iommu *obj, char *buf, ssize_t bytes)
+{
+ int i, num;
+ struct cr_regs *cr;
+ char *p = buf;
+
+ num = bytes / sizeof(*cr);
+ num = min(obj->nr_tlb_entries, num);
+
+ cr = kcalloc(num, sizeof(*cr), GFP_KERNEL);
+ if (!cr)
+ return 0;
+
+ num = __dump_tlb_entries(obj, cr, num);
+ for (i = 0; i < num; i++)
+ p += iotlb_dump_cr(obj, cr + i, p);
+ kfree(cr);
+
+ return p - buf;
+}
+EXPORT_SYMBOL_GPL(omap_dump_tlb_entries);
+
+int omap_foreach_iommu_device(void *data, int (*fn)(struct device *, void *))
+{
+ return driver_for_each_device(&omap_iommu_driver.driver,
+ NULL, data, fn);
+}
+EXPORT_SYMBOL_GPL(omap_foreach_iommu_device);
+
+#endif /* CONFIG_OMAP_IOMMU_DEBUG_MODULE */
+
+/*
+ * H/W pagetable operations
+ */
+static void flush_iopgd_range(u32 *first, u32 *last)
+{
+ /* FIXME: L2 cache should be taken care of if it exists */
+ do {
+ asm("mcr p15, 0, %0, c7, c10, 1 @ flush_pgd"
+ : : "r" (first));
+ first += L1_CACHE_BYTES / sizeof(*first);
+ } while (first <= last);
+}
+
+static void flush_iopte_range(u32 *first, u32 *last)
+{
+ /* FIXME: L2 cache should be taken care of if it exists */
+ do {
+ asm("mcr p15, 0, %0, c7, c10, 1 @ flush_pte"
+ : : "r" (first));
+ first += L1_CACHE_BYTES / sizeof(*first);
+ } while (first <= last);
+}
+
+static void iopte_free(u32 *iopte)
+{
+ /* Note: freed iopte's must be clean ready for re-use */
+ kmem_cache_free(iopte_cachep, iopte);
+}
+
+static u32 *iopte_alloc(struct omap_iommu *obj, u32 *iopgd, u32 da)
+{
+ u32 *iopte;
+
+ /* a table has already existed */
+ if (*iopgd)
+ goto pte_ready;
+
+ /*
+ * do the allocation outside the page table lock
+ */
+ spin_unlock(&obj->page_table_lock);
+ iopte = kmem_cache_zalloc(iopte_cachep, GFP_KERNEL);
+ spin_lock(&obj->page_table_lock);
+
+ if (!*iopgd) {
+ if (!iopte)
+ return ERR_PTR(-ENOMEM);
+
+ *iopgd = virt_to_phys(iopte) | IOPGD_TABLE;
+ flush_iopgd_range(iopgd, iopgd);
+
+ dev_vdbg(obj->dev, "%s: a new pte:%p\n", __func__, iopte);
+ } else {
+ /* We raced, free the reduniovant table */
+ iopte_free(iopte);
+ }
+
+pte_ready:
+ iopte = iopte_offset(iopgd, da);
+
+ dev_vdbg(obj->dev,
+ "%s: da:%08x pgd:%p *pgd:%08x pte:%p *pte:%08x\n",
+ __func__, da, iopgd, *iopgd, iopte, *iopte);
+
+ return iopte;
+}
+
+static int iopgd_alloc_section(struct omap_iommu *obj, u32 da, u32 pa, u32 prot)
+{
+ u32 *iopgd = iopgd_offset(obj, da);
+
+ if ((da | pa) & ~IOSECTION_MASK) {
+ dev_err(obj->dev, "%s: %08x:%08x should aligned on %08lx\n",
+ __func__, da, pa, IOSECTION_SIZE);
+ return -EINVAL;
+ }
+
+ *iopgd = (pa & IOSECTION_MASK) | prot | IOPGD_SECTION;
+ flush_iopgd_range(iopgd, iopgd);
+ return 0;
+}
+
+static int iopgd_alloc_super(struct omap_iommu *obj, u32 da, u32 pa, u32 prot)
+{
+ u32 *iopgd = iopgd_offset(obj, da);
+ int i;
+
+ if ((da | pa) & ~IOSUPER_MASK) {
+ dev_err(obj->dev, "%s: %08x:%08x should aligned on %08lx\n",
+ __func__, da, pa, IOSUPER_SIZE);
+ return -EINVAL;
+ }
+
+ for (i = 0; i < 16; i++)
+ *(iopgd + i) = (pa & IOSUPER_MASK) | prot | IOPGD_SUPER;
+ flush_iopgd_range(iopgd, iopgd + 15);
+ return 0;
+}
+
+static int iopte_alloc_page(struct omap_iommu *obj, u32 da, u32 pa, u32 prot)
+{
+ u32 *iopgd = iopgd_offset(obj, da);
+ u32 *iopte = iopte_alloc(obj, iopgd, da);
+
+ if (IS_ERR(iopte))
+ return PTR_ERR(iopte);
+
+ *iopte = (pa & IOPAGE_MASK) | prot | IOPTE_SMALL;
+ flush_iopte_range(iopte, iopte);
+
+ dev_vdbg(obj->dev, "%s: da:%08x pa:%08x pte:%p *pte:%08x\n",
+ __func__, da, pa, iopte, *iopte);
+
+ return 0;
+}
+
+static int iopte_alloc_large(struct omap_iommu *obj, u32 da, u32 pa, u32 prot)
+{
+ u32 *iopgd = iopgd_offset(obj, da);
+ u32 *iopte = iopte_alloc(obj, iopgd, da);
+ int i;
+
+ if ((da | pa) & ~IOLARGE_MASK) {
+ dev_err(obj->dev, "%s: %08x:%08x should aligned on %08lx\n",
+ __func__, da, pa, IOLARGE_SIZE);
+ return -EINVAL;
+ }
+
+ if (IS_ERR(iopte))
+ return PTR_ERR(iopte);
+
+ for (i = 0; i < 16; i++)
+ *(iopte + i) = (pa & IOLARGE_MASK) | prot | IOPTE_LARGE;
+ flush_iopte_range(iopte, iopte + 15);
+ return 0;
+}
+
+static int
+iopgtable_store_entry_core(struct omap_iommu *obj, struct iotlb_entry *e)
+{
+ int (*fn)(struct omap_iommu *, u32, u32, u32);
+ u32 prot;
+ int err;
+
+ if (!obj || !e)
+ return -EINVAL;
+
+ switch (e->pgsz) {
+ case MMU_CAM_PGSZ_16M:
+ fn = iopgd_alloc_super;
+ break;
+ case MMU_CAM_PGSZ_1M:
+ fn = iopgd_alloc_section;
+ break;
+ case MMU_CAM_PGSZ_64K:
+ fn = iopte_alloc_large;
+ break;
+ case MMU_CAM_PGSZ_4K:
+ fn = iopte_alloc_page;
+ break;
+ default:
+ fn = NULL;
+ BUG();
+ break;
+ }
+
+ prot = get_iopte_attr(e);
+
+ spin_lock(&obj->page_table_lock);
+ err = fn(obj, e->da, e->pa, prot);
+ spin_unlock(&obj->page_table_lock);
+
+ return err;
+}
+
+/**
+ * omap_iopgtable_store_entry - Make an iommu pte entry
+ * @obj: target iommu
+ * @e: an iommu tlb entry info
+ **/
+int omap_iopgtable_store_entry(struct omap_iommu *obj, struct iotlb_entry *e)
+{
+ int err;
+
+ flush_iotlb_page(obj, e->da);
+ err = iopgtable_store_entry_core(obj, e);
+ if (!err)
+ prefetch_iotlb_entry(obj, e);
+ return err;
+}
+EXPORT_SYMBOL_GPL(omap_iopgtable_store_entry);
+
+/**
+ * iopgtable_lookup_entry - Lookup an iommu pte entry
+ * @obj: target iommu
+ * @da: iommu device virtual address
+ * @ppgd: iommu pgd entry pointer to be returned
+ * @ppte: iommu pte entry pointer to be returned
+ **/
+static void
+iopgtable_lookup_entry(struct omap_iommu *obj, u32 da, u32 **ppgd, u32 **ppte)
+{
+ u32 *iopgd, *iopte = NULL;
+
+ iopgd = iopgd_offset(obj, da);
+ if (!*iopgd)
+ goto out;
+
+ if (iopgd_is_table(*iopgd))
+ iopte = iopte_offset(iopgd, da);
+out:
+ *ppgd = iopgd;
+ *ppte = iopte;
+}
+
+static size_t iopgtable_clear_entry_core(struct omap_iommu *obj, u32 da)
+{
+ size_t bytes;
+ u32 *iopgd = iopgd_offset(obj, da);
+ int nent = 1;
+
+ if (!*iopgd)
+ return 0;
+
+ if (iopgd_is_table(*iopgd)) {
+ int i;
+ u32 *iopte = iopte_offset(iopgd, da);
+
+ bytes = IOPTE_SIZE;
+ if (*iopte & IOPTE_LARGE) {
+ nent *= 16;
+ /* rewind to the 1st entry */
+ iopte = iopte_offset(iopgd, (da & IOLARGE_MASK));
+ }
+ bytes *= nent;
+ memset(iopte, 0, nent * sizeof(*iopte));
+ flush_iopte_range(iopte, iopte + (nent - 1) * sizeof(*iopte));
+
+ /*
+ * do table walk to check if this table is necessary or not
+ */
+ iopte = iopte_offset(iopgd, 0);
+ for (i = 0; i < PTRS_PER_IOPTE; i++)
+ if (iopte[i])
+ goto out;
+
+ iopte_free(iopte);
+ nent = 1; /* for the next L1 entry */
+ } else {
+ bytes = IOPGD_SIZE;
+ if ((*iopgd & IOPGD_SUPER) == IOPGD_SUPER) {
+ nent *= 16;
+ /* rewind to the 1st entry */
+ iopgd = iopgd_offset(obj, (da & IOSUPER_MASK));
+ }
+ bytes *= nent;
+ }
+ memset(iopgd, 0, nent * sizeof(*iopgd));
+ flush_iopgd_range(iopgd, iopgd + (nent - 1) * sizeof(*iopgd));
+out:
+ return bytes;
+}
+
+/**
+ * iopgtable_clear_entry - Remove an iommu pte entry
+ * @obj: target iommu
+ * @da: iommu device virtual address
+ **/
+static size_t iopgtable_clear_entry(struct omap_iommu *obj, u32 da)
+{
+ size_t bytes;
+
+ spin_lock(&obj->page_table_lock);
+
+ bytes = iopgtable_clear_entry_core(obj, da);
+ flush_iotlb_page(obj, da);
+
+ spin_unlock(&obj->page_table_lock);
+
+ return bytes;
+}
+
+static void iopgtable_clear_entry_all(struct omap_iommu *obj)
+{
+ int i;
+
+ spin_lock(&obj->page_table_lock);
+
+ for (i = 0; i < PTRS_PER_IOPGD; i++) {
+ u32 da;
+ u32 *iopgd;
+
+ da = i << IOPGD_SHIFT;
+ iopgd = iopgd_offset(obj, da);
+
+ if (!*iopgd)
+ continue;
+
+ if (iopgd_is_table(*iopgd))
+ iopte_free(iopte_offset(iopgd, 0));
+
+ *iopgd = 0;
+ flush_iopgd_range(iopgd, iopgd);
+ }
+
+ flush_iotlb_all(obj);
+
+ spin_unlock(&obj->page_table_lock);
+}
+
+/*
+ * Device IOMMU generic operations
+ */
+static irqreturn_t iommu_fault_handler(int irq, void *data)
+{
+ u32 da, errs;
+ u32 *iopgd, *iopte;
+ struct omap_iommu *obj = data;
+ struct iommu_domain *domain = obj->domain;
+
+ if (!obj->refcount)
+ return IRQ_NONE;
+
+ clk_enable(obj->clk);
+ errs = iommu_report_fault(obj, &da);
+ clk_disable(obj->clk);
+ if (errs == 0)
+ return IRQ_HANDLED;
+
+ /* Fault callback or TLB/PTE Dynamic loading */
+ if (!report_iommu_fault(domain, obj->dev, da, 0))
+ return IRQ_HANDLED;
+
+ iommu_disable(obj);
+
+ iopgd = iopgd_offset(obj, da);
+
+ if (!iopgd_is_table(*iopgd)) {
+ dev_err(obj->dev, "%s: errs:0x%08x da:0x%08x pgd:0x%p "
+ "*pgd:px%08x\n", obj->name, errs, da, iopgd, *iopgd);
+ return IRQ_NONE;
+ }
+
+ iopte = iopte_offset(iopgd, da);
+
+ dev_err(obj->dev, "%s: errs:0x%08x da:0x%08x pgd:0x%p *pgd:0x%08x "
+ "pte:0x%p *pte:0x%08x\n", obj->name, errs, da, iopgd, *iopgd,
+ iopte, *iopte);
+
+ return IRQ_NONE;
+}
+
+static int device_match_by_alias(struct device *dev, void *data)
+{
+ struct omap_iommu *obj = to_iommu(dev);
+ const char *name = data;
+
+ pr_debug("%s: %s %s\n", __func__, obj->name, name);
+
+ return strcmp(obj->name, name) == 0;
+}
+
+/**
+ * omap_find_iommu_device() - find an omap iommu device by name
+ * @name: name of the iommu device
+ *
+ * The generic iommu API requires the caller to provide the device
+ * he wishes to attach to a certain iommu domain.
+ *
+ * Drivers generally should not bother with this as it should just
+ * be taken care of by the DMA-API using dev_archdata.
+ *
+ * This function is provided as an interim solution until the latter
+ * materializes, and omap3isp is fully migrated to the DMA-API.
+ */
+struct device *omap_find_iommu_device(const char *name)
+{
+ return driver_find_device(&omap_iommu_driver.driver, NULL,
+ (void *)name,
+ device_match_by_alias);
+}
+EXPORT_SYMBOL_GPL(omap_find_iommu_device);
+
+/**
+ * omap_iommu_attach() - attach iommu device to an iommu domain
+ * @dev: target omap iommu device
+ * @iopgd: page table
+ **/
+static struct omap_iommu *omap_iommu_attach(struct device *dev, u32 *iopgd)
+{
+ int err = -ENOMEM;
+ struct omap_iommu *obj = to_iommu(dev);
+
+ spin_lock(&obj->iommu_lock);
+
+ /* an iommu device can only be attached once */
+ if (++obj->refcount > 1) {
+ dev_err(dev, "%s: already attached!\n", obj->name);
+ err = -EBUSY;
+ goto err_enable;
+ }
+
+ obj->iopgd = iopgd;
+ err = iommu_enable(obj);
+ if (err)
+ goto err_enable;
+ flush_iotlb_all(obj);
+
+ if (!try_module_get(obj->owner))
+ goto err_module;
+
+ spin_unlock(&obj->iommu_lock);
+
+ dev_dbg(obj->dev, "%s: %s\n", __func__, obj->name);
+ return obj;
+
+err_module:
+ if (obj->refcount == 1)
+ iommu_disable(obj);
+err_enable:
+ obj->refcount--;
+ spin_unlock(&obj->iommu_lock);
+ return ERR_PTR(err);
+}
+
+/**
+ * omap_iommu_detach - release iommu device
+ * @obj: target iommu
+ **/
+static void omap_iommu_detach(struct omap_iommu *obj)
+{
+ if (!obj || IS_ERR(obj))
+ return;
+
+ spin_lock(&obj->iommu_lock);
+
+ if (--obj->refcount == 0)
+ iommu_disable(obj);
+
+ module_put(obj->owner);
+
+ obj->iopgd = NULL;
+
+ spin_unlock(&obj->iommu_lock);
+
+ dev_dbg(obj->dev, "%s: %s\n", __func__, obj->name);
+}
+
+/*
+ * OMAP Device MMU(IOMMU) detection
+ */
+static int __devinit omap_iommu_probe(struct platform_device *pdev)
+{
+ int err = -ENODEV;
+ int irq;
+ struct omap_iommu *obj;
+ struct resource *res;
+ struct iommu_platform_data *pdata = pdev->dev.platform_data;
+
+ if (pdev->num_resources != 2)
+ return -EINVAL;
+
+ obj = kzalloc(sizeof(*obj) + MMU_REG_SIZE, GFP_KERNEL);
+ if (!obj)
+ return -ENOMEM;
+
+ obj->clk = clk_get(&pdev->dev, pdata->clk_name);
+ if (IS_ERR(obj->clk))
+ goto err_clk;
+
+ obj->nr_tlb_entries = pdata->nr_tlb_entries;
+ obj->name = pdata->name;
+ obj->dev = &pdev->dev;
+ obj->ctx = (void *)obj + sizeof(*obj);
+ obj->da_start = pdata->da_start;
+ obj->da_end = pdata->da_end;
+
+ spin_lock_init(&obj->iommu_lock);
+ mutex_init(&obj->mmap_lock);
+ spin_lock_init(&obj->page_table_lock);
+ INIT_LIST_HEAD(&obj->mmap);
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ if (!res) {
+ err = -ENODEV;
+ goto err_mem;
+ }
+
+ res = request_mem_region(res->start, resource_size(res),
+ dev_name(&pdev->dev));
+ if (!res) {
+ err = -EIO;
+ goto err_mem;
+ }
+
+ obj->regbase = ioremap(res->start, resource_size(res));
+ if (!obj->regbase) {
+ err = -ENOMEM;
+ goto err_ioremap;
+ }
+
+ irq = platform_get_irq(pdev, 0);
+ if (irq < 0) {
+ err = -ENODEV;
+ goto err_irq;
+ }
+ err = request_irq(irq, iommu_fault_handler, IRQF_SHARED,
+ dev_name(&pdev->dev), obj);
+ if (err < 0)
+ goto err_irq;
+ platform_set_drvdata(pdev, obj);
+
+ dev_info(&pdev->dev, "%s registered\n", obj->name);
+ return 0;
+
+err_irq:
+ iounmap(obj->regbase);
+err_ioremap:
+ release_mem_region(res->start, resource_size(res));
+err_mem:
+ clk_put(obj->clk);
+err_clk:
+ kfree(obj);
+ return err;
+}
+
+static int __devexit omap_iommu_remove(struct platform_device *pdev)
+{
+ int irq;
+ struct resource *res;
+ struct omap_iommu *obj = platform_get_drvdata(pdev);
+
+ platform_set_drvdata(pdev, NULL);
+
+ iopgtable_clear_entry_all(obj);
+
+ irq = platform_get_irq(pdev, 0);
+ free_irq(irq, obj);
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ release_mem_region(res->start, resource_size(res));
+ iounmap(obj->regbase);
+
+ clk_put(obj->clk);
+ dev_info(&pdev->dev, "%s removed\n", obj->name);
+ kfree(obj);
+ return 0;
+}
+
+static struct platform_driver omap_iommu_driver = {
+ .probe = omap_iommu_probe,
+ .remove = __devexit_p(omap_iommu_remove),
+ .driver = {
+ .name = "omap-iommu",
+ },
+};
+
+static void iopte_cachep_ctor(void *iopte)
+{
+ clean_dcache_area(iopte, IOPTE_TABLE_SIZE);
+}
+
+static int omap_iommu_map(struct iommu_domain *domain, unsigned long da,
+ phys_addr_t pa, int order, int prot)
+{
+ struct omap_iommu_domain *omap_domain = domain->priv;
+ struct omap_iommu *oiommu = omap_domain->iommu_dev;
+ struct device *dev = oiommu->dev;
+ size_t bytes = PAGE_SIZE << order;
+ struct iotlb_entry e;
+ int omap_pgsz;
+ u32 ret, flags;
+
+ /* we only support mapping a single iommu page for now */
+ omap_pgsz = bytes_to_iopgsz(bytes);
+ if (omap_pgsz < 0) {
+ dev_err(dev, "invalid size to map: %d\n", bytes);
+ return -EINVAL;
+ }
+
+ dev_dbg(dev, "mapping da 0x%lx to pa 0x%x size 0x%x\n", da, pa, bytes);
+
+ flags = omap_pgsz | prot;
+
+ iotlb_init_entry(&e, da, pa, flags);
+
+ ret = omap_iopgtable_store_entry(oiommu, &e);
+ if (ret)
+ dev_err(dev, "omap_iopgtable_store_entry failed: %d\n", ret);
+
+ return ret;
+}
+
+static int omap_iommu_unmap(struct iommu_domain *domain, unsigned long da,
+ int order)
+{
+ struct omap_iommu_domain *omap_domain = domain->priv;
+ struct omap_iommu *oiommu = omap_domain->iommu_dev;
+ struct device *dev = oiommu->dev;
+ size_t unmap_size;
+
+ dev_dbg(dev, "unmapping da 0x%lx order %d\n", da, order);
+
+ unmap_size = iopgtable_clear_entry(oiommu, da);
+
+ return unmap_size ? get_order(unmap_size) : -EINVAL;
+}
+
+static int
+omap_iommu_attach_dev(struct iommu_domain *domain, struct device *dev)
+{
+ struct omap_iommu_domain *omap_domain = domain->priv;
+ struct omap_iommu *oiommu;
+ int ret = 0;
+
+ spin_lock(&omap_domain->lock);
+
+ /* only a single device is supported per domain for now */
+ if (omap_domain->iommu_dev) {
+ dev_err(dev, "iommu domain is already attached\n");
+ ret = -EBUSY;
+ goto out;
+ }
+
+ /* get a handle to and enable the omap iommu */
+ oiommu = omap_iommu_attach(dev, omap_domain->pgtable);
+ if (IS_ERR(oiommu)) {
+ ret = PTR_ERR(oiommu);
+ dev_err(dev, "can't get omap iommu: %d\n", ret);
+ goto out;
+ }
+
+ omap_domain->iommu_dev = oiommu;
+ oiommu->domain = domain;
+
+out:
+ spin_unlock(&omap_domain->lock);
+ return ret;
+}
+
+static void omap_iommu_detach_dev(struct iommu_domain *domain,
+ struct device *dev)
+{
+ struct omap_iommu_domain *omap_domain = domain->priv;
+ struct omap_iommu *oiommu = to_iommu(dev);
+
+ spin_lock(&omap_domain->lock);
+
+ /* only a single device is supported per domain for now */
+ if (omap_domain->iommu_dev != oiommu) {
+ dev_err(dev, "invalid iommu device\n");
+ goto out;
+ }
+
+ iopgtable_clear_entry_all(oiommu);
+
+ omap_iommu_detach(oiommu);
+
+ omap_domain->iommu_dev = NULL;
+
+out:
+ spin_unlock(&omap_domain->lock);
+}
+
+static int omap_iommu_domain_init(struct iommu_domain *domain)
+{
+ struct omap_iommu_domain *omap_domain;
+
+ omap_domain = kzalloc(sizeof(*omap_domain), GFP_KERNEL);
+ if (!omap_domain) {
+ pr_err("kzalloc failed\n");
+ goto out;
+ }
+
+ omap_domain->pgtable = kzalloc(IOPGD_TABLE_SIZE, GFP_KERNEL);
+ if (!omap_domain->pgtable) {
+ pr_err("kzalloc failed\n");
+ goto fail_nomem;
+ }
+
+ /*
+ * should never fail, but please keep this around to ensure
+ * we keep the hardware happy
+ */
+ BUG_ON(!IS_ALIGNED((long)omap_domain->pgtable, IOPGD_TABLE_SIZE));
+
+ clean_dcache_area(omap_domain->pgtable, IOPGD_TABLE_SIZE);
+ spin_lock_init(&omap_domain->lock);
+
+ domain->priv = omap_domain;
+
+ return 0;
+
+fail_nomem:
+ kfree(omap_domain);
+out:
+ return -ENOMEM;
+}
+
+/* assume device was already detached */
+static void omap_iommu_domain_destroy(struct iommu_domain *domain)
+{
+ struct omap_iommu_domain *omap_domain = domain->priv;
+
+ domain->priv = NULL;
+
+ kfree(omap_domain->pgtable);
+ kfree(omap_domain);
+}
+
+static phys_addr_t omap_iommu_iova_to_phys(struct iommu_domain *domain,
+ unsigned long da)
+{
+ struct omap_iommu_domain *omap_domain = domain->priv;
+ struct omap_iommu *oiommu = omap_domain->iommu_dev;
+ struct device *dev = oiommu->dev;
+ u32 *pgd, *pte;
+ phys_addr_t ret = 0;
+
+ iopgtable_lookup_entry(oiommu, da, &pgd, &pte);
+
+ if (pte) {
+ if (iopte_is_small(*pte))
+ ret = omap_iommu_translate(*pte, da, IOPTE_MASK);
+ else if (iopte_is_large(*pte))
+ ret = omap_iommu_translate(*pte, da, IOLARGE_MASK);
+ else
+ dev_err(dev, "bogus pte 0x%x", *pte);
+ } else {
+ if (iopgd_is_section(*pgd))
+ ret = omap_iommu_translate(*pgd, da, IOSECTION_MASK);
+ else if (iopgd_is_super(*pgd))
+ ret = omap_iommu_translate(*pgd, da, IOSUPER_MASK);
+ else
+ dev_err(dev, "bogus pgd 0x%x", *pgd);
+ }
+
+ return ret;
+}
+
+static int omap_iommu_domain_has_cap(struct iommu_domain *domain,
+ unsigned long cap)
+{
+ return 0;
+}
+
+static struct iommu_ops omap_iommu_ops = {
+ .domain_init = omap_iommu_domain_init,
+ .domain_destroy = omap_iommu_domain_destroy,
+ .attach_dev = omap_iommu_attach_dev,
+ .detach_dev = omap_iommu_detach_dev,
+ .map = omap_iommu_map,
+ .unmap = omap_iommu_unmap,
+ .iova_to_phys = omap_iommu_iova_to_phys,
+ .domain_has_cap = omap_iommu_domain_has_cap,
+};
+
+static int __init omap_iommu_init(void)
+{
+ struct kmem_cache *p;
+ const unsigned long flags = SLAB_HWCACHE_ALIGN;
+ size_t align = 1 << 10; /* L2 pagetable alignement */
+
+ p = kmem_cache_create("iopte_cache", IOPTE_TABLE_SIZE, align, flags,
+ iopte_cachep_ctor);
+ if (!p)
+ return -ENOMEM;
+ iopte_cachep = p;
+
+ bus_set_iommu(&platform_bus_type, &omap_iommu_ops);
+
+ return platform_driver_register(&omap_iommu_driver);
+}
+module_init(omap_iommu_init);
+
+static void __exit omap_iommu_exit(void)
+{
+ kmem_cache_destroy(iopte_cachep);
+
+ platform_driver_unregister(&omap_iommu_driver);
+}
+module_exit(omap_iommu_exit);
+
+MODULE_DESCRIPTION("omap iommu: tlb and pagetable primitives");
+MODULE_ALIAS("platform:omap-iommu");
+MODULE_AUTHOR("Hiroshi DOYU, Paul Mundt and Toshihiro Kobayashi");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/iommu/omap-iovmm.c b/drivers/iommu/omap-iovmm.c
new file mode 100644
index 000000000000..e8fdb8830f69
--- /dev/null
+++ b/drivers/iommu/omap-iovmm.c
@@ -0,0 +1,742 @@
+/*
+ * omap iommu: simple virtual address space management
+ *
+ * Copyright (C) 2008-2009 Nokia Corporation
+ *
+ * Written by Hiroshi DOYU <Hiroshi.DOYU@nokia.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/err.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/device.h>
+#include <linux/scatterlist.h>
+#include <linux/iommu.h>
+
+#include <asm/cacheflush.h>
+#include <asm/mach/map.h>
+
+#include <plat/iommu.h>
+#include <plat/iovmm.h>
+
+#include <plat/iopgtable.h>
+
+static struct kmem_cache *iovm_area_cachep;
+
+/* return the offset of the first scatterlist entry in a sg table */
+static unsigned int sgtable_offset(const struct sg_table *sgt)
+{
+ if (!sgt || !sgt->nents)
+ return 0;
+
+ return sgt->sgl->offset;
+}
+
+/* return total bytes of sg buffers */
+static size_t sgtable_len(const struct sg_table *sgt)
+{
+ unsigned int i, total = 0;
+ struct scatterlist *sg;
+
+ if (!sgt)
+ return 0;
+
+ for_each_sg(sgt->sgl, sg, sgt->nents, i) {
+ size_t bytes;
+
+ bytes = sg->length + sg->offset;
+
+ if (!iopgsz_ok(bytes)) {
+ pr_err("%s: sg[%d] not iommu pagesize(%u %u)\n",
+ __func__, i, bytes, sg->offset);
+ return 0;
+ }
+
+ if (i && sg->offset) {
+ pr_err("%s: sg[%d] offset not allowed in internal "
+ "entries\n", __func__, i);
+ return 0;
+ }
+
+ total += bytes;
+ }
+
+ return total;
+}
+#define sgtable_ok(x) (!!sgtable_len(x))
+
+static unsigned max_alignment(u32 addr)
+{
+ int i;
+ unsigned pagesize[] = { SZ_16M, SZ_1M, SZ_64K, SZ_4K, };
+ for (i = 0; i < ARRAY_SIZE(pagesize) && addr & (pagesize[i] - 1); i++)
+ ;
+ return (i < ARRAY_SIZE(pagesize)) ? pagesize[i] : 0;
+}
+
+/*
+ * calculate the optimal number sg elements from total bytes based on
+ * iommu superpages
+ */
+static unsigned sgtable_nents(size_t bytes, u32 da, u32 pa)
+{
+ unsigned nr_entries = 0, ent_sz;
+
+ if (!IS_ALIGNED(bytes, PAGE_SIZE)) {
+ pr_err("%s: wrong size %08x\n", __func__, bytes);
+ return 0;
+ }
+
+ while (bytes) {
+ ent_sz = max_alignment(da | pa);
+ ent_sz = min_t(unsigned, ent_sz, iopgsz_max(bytes));
+ nr_entries++;
+ da += ent_sz;
+ pa += ent_sz;
+ bytes -= ent_sz;
+ }
+
+ return nr_entries;
+}
+
+/* allocate and initialize sg_table header(a kind of 'superblock') */
+static struct sg_table *sgtable_alloc(const size_t bytes, u32 flags,
+ u32 da, u32 pa)
+{
+ unsigned int nr_entries;
+ int err;
+ struct sg_table *sgt;
+
+ if (!bytes)
+ return ERR_PTR(-EINVAL);
+
+ if (!IS_ALIGNED(bytes, PAGE_SIZE))
+ return ERR_PTR(-EINVAL);
+
+ if (flags & IOVMF_LINEAR) {
+ nr_entries = sgtable_nents(bytes, da, pa);
+ if (!nr_entries)
+ return ERR_PTR(-EINVAL);
+ } else
+ nr_entries = bytes / PAGE_SIZE;
+
+ sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
+ if (!sgt)
+ return ERR_PTR(-ENOMEM);
+
+ err = sg_alloc_table(sgt, nr_entries, GFP_KERNEL);
+ if (err) {
+ kfree(sgt);
+ return ERR_PTR(err);
+ }
+
+ pr_debug("%s: sgt:%p(%d entries)\n", __func__, sgt, nr_entries);
+
+ return sgt;
+}
+
+/* free sg_table header(a kind of superblock) */
+static void sgtable_free(struct sg_table *sgt)
+{
+ if (!sgt)
+ return;
+
+ sg_free_table(sgt);
+ kfree(sgt);
+
+ pr_debug("%s: sgt:%p\n", __func__, sgt);
+}
+
+/* map 'sglist' to a contiguous mpu virtual area and return 'va' */
+static void *vmap_sg(const struct sg_table *sgt)
+{
+ u32 va;
+ size_t total;
+ unsigned int i;
+ struct scatterlist *sg;
+ struct vm_struct *new;
+ const struct mem_type *mtype;
+
+ mtype = get_mem_type(MT_DEVICE);
+ if (!mtype)
+ return ERR_PTR(-EINVAL);
+
+ total = sgtable_len(sgt);
+ if (!total)
+ return ERR_PTR(-EINVAL);
+
+ new = __get_vm_area(total, VM_IOREMAP, VMALLOC_START, VMALLOC_END);
+ if (!new)
+ return ERR_PTR(-ENOMEM);
+ va = (u32)new->addr;
+
+ for_each_sg(sgt->sgl, sg, sgt->nents, i) {
+ size_t bytes;
+ u32 pa;
+ int err;
+
+ pa = sg_phys(sg) - sg->offset;
+ bytes = sg->length + sg->offset;
+
+ BUG_ON(bytes != PAGE_SIZE);
+
+ err = ioremap_page(va, pa, mtype);
+ if (err)
+ goto err_out;
+
+ va += bytes;
+ }
+
+ flush_cache_vmap((unsigned long)new->addr,
+ (unsigned long)(new->addr + total));
+ return new->addr;
+
+err_out:
+ WARN_ON(1); /* FIXME: cleanup some mpu mappings */
+ vunmap(new->addr);
+ return ERR_PTR(-EAGAIN);
+}
+
+static inline void vunmap_sg(const void *va)
+{
+ vunmap(va);
+}
+
+static struct iovm_struct *__find_iovm_area(struct omap_iommu *obj,
+ const u32 da)
+{
+ struct iovm_struct *tmp;
+
+ list_for_each_entry(tmp, &obj->mmap, list) {
+ if ((da >= tmp->da_start) && (da < tmp->da_end)) {
+ size_t len;
+
+ len = tmp->da_end - tmp->da_start;
+
+ dev_dbg(obj->dev, "%s: %08x-%08x-%08x(%x) %08x\n",
+ __func__, tmp->da_start, da, tmp->da_end, len,
+ tmp->flags);
+
+ return tmp;
+ }
+ }
+
+ return NULL;
+}
+
+/**
+ * omap_find_iovm_area - find iovma which includes @da
+ * @da: iommu device virtual address
+ *
+ * Find the existing iovma starting at @da
+ */
+struct iovm_struct *omap_find_iovm_area(struct omap_iommu *obj, u32 da)
+{
+ struct iovm_struct *area;
+
+ mutex_lock(&obj->mmap_lock);
+ area = __find_iovm_area(obj, da);
+ mutex_unlock(&obj->mmap_lock);
+
+ return area;
+}
+EXPORT_SYMBOL_GPL(omap_find_iovm_area);
+
+/*
+ * This finds the hole(area) which fits the requested address and len
+ * in iovmas mmap, and returns the new allocated iovma.
+ */
+static struct iovm_struct *alloc_iovm_area(struct omap_iommu *obj, u32 da,
+ size_t bytes, u32 flags)
+{
+ struct iovm_struct *new, *tmp;
+ u32 start, prev_end, alignment;
+
+ if (!obj || !bytes)
+ return ERR_PTR(-EINVAL);
+
+ start = da;
+ alignment = PAGE_SIZE;
+
+ if (~flags & IOVMF_DA_FIXED) {
+ /* Don't map address 0 */
+ start = obj->da_start ? obj->da_start : alignment;
+
+ if (flags & IOVMF_LINEAR)
+ alignment = iopgsz_max(bytes);
+ start = roundup(start, alignment);
+ } else if (start < obj->da_start || start > obj->da_end ||
+ obj->da_end - start < bytes) {
+ return ERR_PTR(-EINVAL);
+ }
+
+ tmp = NULL;
+ if (list_empty(&obj->mmap))
+ goto found;
+
+ prev_end = 0;
+ list_for_each_entry(tmp, &obj->mmap, list) {
+
+ if (prev_end > start)
+ break;
+
+ if (tmp->da_start > start && (tmp->da_start - start) >= bytes)
+ goto found;
+
+ if (tmp->da_end >= start && ~flags & IOVMF_DA_FIXED)
+ start = roundup(tmp->da_end + 1, alignment);
+
+ prev_end = tmp->da_end;
+ }
+
+ if ((start >= prev_end) && (obj->da_end - start >= bytes))
+ goto found;
+
+ dev_dbg(obj->dev, "%s: no space to fit %08x(%x) flags: %08x\n",
+ __func__, da, bytes, flags);
+
+ return ERR_PTR(-EINVAL);
+
+found:
+ new = kmem_cache_zalloc(iovm_area_cachep, GFP_KERNEL);
+ if (!new)
+ return ERR_PTR(-ENOMEM);
+
+ new->iommu = obj;
+ new->da_start = start;
+ new->da_end = start + bytes;
+ new->flags = flags;
+
+ /*
+ * keep ascending order of iovmas
+ */
+ if (tmp)
+ list_add_tail(&new->list, &tmp->list);
+ else
+ list_add(&new->list, &obj->mmap);
+
+ dev_dbg(obj->dev, "%s: found %08x-%08x-%08x(%x) %08x\n",
+ __func__, new->da_start, start, new->da_end, bytes, flags);
+
+ return new;
+}
+
+static void free_iovm_area(struct omap_iommu *obj, struct iovm_struct *area)
+{
+ size_t bytes;
+
+ BUG_ON(!obj || !area);
+
+ bytes = area->da_end - area->da_start;
+
+ dev_dbg(obj->dev, "%s: %08x-%08x(%x) %08x\n",
+ __func__, area->da_start, area->da_end, bytes, area->flags);
+
+ list_del(&area->list);
+ kmem_cache_free(iovm_area_cachep, area);
+}
+
+/**
+ * omap_da_to_va - convert (d) to (v)
+ * @obj: objective iommu
+ * @da: iommu device virtual address
+ * @va: mpu virtual address
+ *
+ * Returns mpu virtual addr which corresponds to a given device virtual addr
+ */
+void *omap_da_to_va(struct omap_iommu *obj, u32 da)
+{
+ void *va = NULL;
+ struct iovm_struct *area;
+
+ mutex_lock(&obj->mmap_lock);
+
+ area = __find_iovm_area(obj, da);
+ if (!area) {
+ dev_dbg(obj->dev, "%s: no da area(%08x)\n", __func__, da);
+ goto out;
+ }
+ va = area->va;
+out:
+ mutex_unlock(&obj->mmap_lock);
+
+ return va;
+}
+EXPORT_SYMBOL_GPL(omap_da_to_va);
+
+static void sgtable_fill_vmalloc(struct sg_table *sgt, void *_va)
+{
+ unsigned int i;
+ struct scatterlist *sg;
+ void *va = _va;
+ void *va_end;
+
+ for_each_sg(sgt->sgl, sg, sgt->nents, i) {
+ struct page *pg;
+ const size_t bytes = PAGE_SIZE;
+
+ /*
+ * iommu 'superpage' isn't supported with 'omap_iommu_vmalloc()'
+ */
+ pg = vmalloc_to_page(va);
+ BUG_ON(!pg);
+ sg_set_page(sg, pg, bytes, 0);
+
+ va += bytes;
+ }
+
+ va_end = _va + PAGE_SIZE * i;
+}
+
+static inline void sgtable_drain_vmalloc(struct sg_table *sgt)
+{
+ /*
+ * Actually this is not necessary at all, just exists for
+ * consistency of the code readability.
+ */
+ BUG_ON(!sgt);
+}
+
+/* create 'da' <-> 'pa' mapping from 'sgt' */
+static int map_iovm_area(struct iommu_domain *domain, struct iovm_struct *new,
+ const struct sg_table *sgt, u32 flags)
+{
+ int err;
+ unsigned int i, j;
+ struct scatterlist *sg;
+ u32 da = new->da_start;
+ int order;
+
+ if (!domain || !sgt)
+ return -EINVAL;
+
+ BUG_ON(!sgtable_ok(sgt));
+
+ for_each_sg(sgt->sgl, sg, sgt->nents, i) {
+ u32 pa;
+ size_t bytes;
+
+ pa = sg_phys(sg) - sg->offset;
+ bytes = sg->length + sg->offset;
+
+ flags &= ~IOVMF_PGSZ_MASK;
+
+ if (bytes_to_iopgsz(bytes) < 0)
+ goto err_out;
+
+ order = get_order(bytes);
+
+ pr_debug("%s: [%d] %08x %08x(%x)\n", __func__,
+ i, da, pa, bytes);
+
+ err = iommu_map(domain, da, pa, order, flags);
+ if (err)
+ goto err_out;
+
+ da += bytes;
+ }
+ return 0;
+
+err_out:
+ da = new->da_start;
+
+ for_each_sg(sgt->sgl, sg, i, j) {
+ size_t bytes;
+
+ bytes = sg->length + sg->offset;
+ order = get_order(bytes);
+
+ /* ignore failures.. we're already handling one */
+ iommu_unmap(domain, da, order);
+
+ da += bytes;
+ }
+ return err;
+}
+
+/* release 'da' <-> 'pa' mapping */
+static void unmap_iovm_area(struct iommu_domain *domain, struct omap_iommu *obj,
+ struct iovm_struct *area)
+{
+ u32 start;
+ size_t total = area->da_end - area->da_start;
+ const struct sg_table *sgt = area->sgt;
+ struct scatterlist *sg;
+ int i, err;
+
+ BUG_ON(!sgtable_ok(sgt));
+ BUG_ON((!total) || !IS_ALIGNED(total, PAGE_SIZE));
+
+ start = area->da_start;
+ for_each_sg(sgt->sgl, sg, sgt->nents, i) {
+ size_t bytes;
+ int order;
+
+ bytes = sg->length + sg->offset;
+ order = get_order(bytes);
+
+ err = iommu_unmap(domain, start, order);
+ if (err < 0)
+ break;
+
+ dev_dbg(obj->dev, "%s: unmap %08x(%x) %08x\n",
+ __func__, start, bytes, area->flags);
+
+ BUG_ON(!IS_ALIGNED(bytes, PAGE_SIZE));
+
+ total -= bytes;
+ start += bytes;
+ }
+ BUG_ON(total);
+}
+
+/* template function for all unmapping */
+static struct sg_table *unmap_vm_area(struct iommu_domain *domain,
+ struct omap_iommu *obj, const u32 da,
+ void (*fn)(const void *), u32 flags)
+{
+ struct sg_table *sgt = NULL;
+ struct iovm_struct *area;
+
+ if (!IS_ALIGNED(da, PAGE_SIZE)) {
+ dev_err(obj->dev, "%s: alignment err(%08x)\n", __func__, da);
+ return NULL;
+ }
+
+ mutex_lock(&obj->mmap_lock);
+
+ area = __find_iovm_area(obj, da);
+ if (!area) {
+ dev_dbg(obj->dev, "%s: no da area(%08x)\n", __func__, da);
+ goto out;
+ }
+
+ if ((area->flags & flags) != flags) {
+ dev_err(obj->dev, "%s: wrong flags(%08x)\n", __func__,
+ area->flags);
+ goto out;
+ }
+ sgt = (struct sg_table *)area->sgt;
+
+ unmap_iovm_area(domain, obj, area);
+
+ fn(area->va);
+
+ dev_dbg(obj->dev, "%s: %08x-%08x-%08x(%x) %08x\n", __func__,
+ area->da_start, da, area->da_end,
+ area->da_end - area->da_start, area->flags);
+
+ free_iovm_area(obj, area);
+out:
+ mutex_unlock(&obj->mmap_lock);
+
+ return sgt;
+}
+
+static u32 map_iommu_region(struct iommu_domain *domain, struct omap_iommu *obj,
+ u32 da, const struct sg_table *sgt, void *va,
+ size_t bytes, u32 flags)
+{
+ int err = -ENOMEM;
+ struct iovm_struct *new;
+
+ mutex_lock(&obj->mmap_lock);
+
+ new = alloc_iovm_area(obj, da, bytes, flags);
+ if (IS_ERR(new)) {
+ err = PTR_ERR(new);
+ goto err_alloc_iovma;
+ }
+ new->va = va;
+ new->sgt = sgt;
+
+ if (map_iovm_area(domain, new, sgt, new->flags))
+ goto err_map;
+
+ mutex_unlock(&obj->mmap_lock);
+
+ dev_dbg(obj->dev, "%s: da:%08x(%x) flags:%08x va:%p\n",
+ __func__, new->da_start, bytes, new->flags, va);
+
+ return new->da_start;
+
+err_map:
+ free_iovm_area(obj, new);
+err_alloc_iovma:
+ mutex_unlock(&obj->mmap_lock);
+ return err;
+}
+
+static inline u32
+__iommu_vmap(struct iommu_domain *domain, struct omap_iommu *obj,
+ u32 da, const struct sg_table *sgt,
+ void *va, size_t bytes, u32 flags)
+{
+ return map_iommu_region(domain, obj, da, sgt, va, bytes, flags);
+}
+
+/**
+ * omap_iommu_vmap - (d)-(p)-(v) address mapper
+ * @obj: objective iommu
+ * @sgt: address of scatter gather table
+ * @flags: iovma and page property
+ *
+ * Creates 1-n-1 mapping with given @sgt and returns @da.
+ * All @sgt element must be io page size aligned.
+ */
+u32 omap_iommu_vmap(struct iommu_domain *domain, struct omap_iommu *obj, u32 da,
+ const struct sg_table *sgt, u32 flags)
+{
+ size_t bytes;
+ void *va = NULL;
+
+ if (!obj || !obj->dev || !sgt)
+ return -EINVAL;
+
+ bytes = sgtable_len(sgt);
+ if (!bytes)
+ return -EINVAL;
+ bytes = PAGE_ALIGN(bytes);
+
+ if (flags & IOVMF_MMIO) {
+ va = vmap_sg(sgt);
+ if (IS_ERR(va))
+ return PTR_ERR(va);
+ }
+
+ flags |= IOVMF_DISCONT;
+ flags |= IOVMF_MMIO;
+
+ da = __iommu_vmap(domain, obj, da, sgt, va, bytes, flags);
+ if (IS_ERR_VALUE(da))
+ vunmap_sg(va);
+
+ return da + sgtable_offset(sgt);
+}
+EXPORT_SYMBOL_GPL(omap_iommu_vmap);
+
+/**
+ * omap_iommu_vunmap - release virtual mapping obtained by 'omap_iommu_vmap()'
+ * @obj: objective iommu
+ * @da: iommu device virtual address
+ *
+ * Free the iommu virtually contiguous memory area starting at
+ * @da, which was returned by 'omap_iommu_vmap()'.
+ */
+struct sg_table *
+omap_iommu_vunmap(struct iommu_domain *domain, struct omap_iommu *obj, u32 da)
+{
+ struct sg_table *sgt;
+ /*
+ * 'sgt' is allocated before 'omap_iommu_vmalloc()' is called.
+ * Just returns 'sgt' to the caller to free
+ */
+ da &= PAGE_MASK;
+ sgt = unmap_vm_area(domain, obj, da, vunmap_sg,
+ IOVMF_DISCONT | IOVMF_MMIO);
+ if (!sgt)
+ dev_dbg(obj->dev, "%s: No sgt\n", __func__);
+ return sgt;
+}
+EXPORT_SYMBOL_GPL(omap_iommu_vunmap);
+
+/**
+ * omap_iommu_vmalloc - (d)-(p)-(v) address allocator and mapper
+ * @obj: objective iommu
+ * @da: contiguous iommu virtual memory
+ * @bytes: allocation size
+ * @flags: iovma and page property
+ *
+ * Allocate @bytes linearly and creates 1-n-1 mapping and returns
+ * @da again, which might be adjusted if 'IOVMF_DA_FIXED' is not set.
+ */
+u32
+omap_iommu_vmalloc(struct iommu_domain *domain, struct omap_iommu *obj, u32 da,
+ size_t bytes, u32 flags)
+{
+ void *va;
+ struct sg_table *sgt;
+
+ if (!obj || !obj->dev || !bytes)
+ return -EINVAL;
+
+ bytes = PAGE_ALIGN(bytes);
+
+ va = vmalloc(bytes);
+ if (!va)
+ return -ENOMEM;
+
+ flags |= IOVMF_DISCONT;
+ flags |= IOVMF_ALLOC;
+
+ sgt = sgtable_alloc(bytes, flags, da, 0);
+ if (IS_ERR(sgt)) {
+ da = PTR_ERR(sgt);
+ goto err_sgt_alloc;
+ }
+ sgtable_fill_vmalloc(sgt, va);
+
+ da = __iommu_vmap(domain, obj, da, sgt, va, bytes, flags);
+ if (IS_ERR_VALUE(da))
+ goto err_iommu_vmap;
+
+ return da;
+
+err_iommu_vmap:
+ sgtable_drain_vmalloc(sgt);
+ sgtable_free(sgt);
+err_sgt_alloc:
+ vfree(va);
+ return da;
+}
+EXPORT_SYMBOL_GPL(omap_iommu_vmalloc);
+
+/**
+ * omap_iommu_vfree - release memory allocated by 'omap_iommu_vmalloc()'
+ * @obj: objective iommu
+ * @da: iommu device virtual address
+ *
+ * Frees the iommu virtually continuous memory area starting at
+ * @da, as obtained from 'omap_iommu_vmalloc()'.
+ */
+void omap_iommu_vfree(struct iommu_domain *domain, struct omap_iommu *obj,
+ const u32 da)
+{
+ struct sg_table *sgt;
+
+ sgt = unmap_vm_area(domain, obj, da, vfree,
+ IOVMF_DISCONT | IOVMF_ALLOC);
+ if (!sgt)
+ dev_dbg(obj->dev, "%s: No sgt\n", __func__);
+ sgtable_free(sgt);
+}
+EXPORT_SYMBOL_GPL(omap_iommu_vfree);
+
+static int __init iovmm_init(void)
+{
+ const unsigned long flags = SLAB_HWCACHE_ALIGN;
+ struct kmem_cache *p;
+
+ p = kmem_cache_create("iovm_area_cache", sizeof(struct iovm_struct), 0,
+ flags, NULL);
+ if (!p)
+ return -ENOMEM;
+ iovm_area_cachep = p;
+
+ return 0;
+}
+module_init(iovmm_init);
+
+static void __exit iovmm_exit(void)
+{
+ kmem_cache_destroy(iovm_area_cachep);
+}
+module_exit(iovmm_exit);
+
+MODULE_DESCRIPTION("omap iommu: simple virtual address space management");
+MODULE_AUTHOR("Hiroshi DOYU <Hiroshi.DOYU@nokia.com>");
+MODULE_LICENSE("GPL v2");