Merge from Linus' tree.

10 files changed, 132 insertions, 105 deletions

diff --git a/arch/ppc64/Makefile b/arch/ppc64/Makefile
index a8877881bb48..40675b3f924d 100644
--- a/arch/ppc64/Makefile
+++ b/arch/ppc64/Makefile
@@ -107,7 +107,7 @@ install: vmlinux
         $(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(BOOTIMAGE) $@
 
 defaultimage-$(CONFIG_PPC_PSERIES) := zImage
-defaultimage-$(CONFIG_PPC_PMAC) := vmlinux
+defaultimage-$(CONFIG_PPC_PMAC) := zImage.vmode
 defaultimage-$(CONFIG_PPC_MAPLE) := zImage
 defaultimage-$(CONFIG_PPC_ISERIES) := vmlinux
 KBUILD_IMAGE := $(defaultimage-y)
diff --git a/arch/ppc64/kernel/of_device.c b/arch/ppc64/kernel/of_device.c
index da580812ddfe..9f200f0f2ad5 100644
--- a/arch/ppc64/kernel/of_device.c
+++ b/arch/ppc64/kernel/of_device.c
@@ -233,7 +233,9 @@ void of_device_unregister(struct of_device *ofdev)
         device_unregister(&ofdev->dev);
 }
 
-struct of_device* of_platform_device_create(struct device_node *np, const char *bus_id)
+struct of_device* of_platform_device_create(struct device_node *np,
+                                            const char *bus_id,
+                                            struct device *parent)
 {
         struct of_device *dev;
 
@@ -245,7 +247,7 @@ struct of_device* of_platform_device_create(struct device_node *np, const char *
         dev->node = np;
         dev->dma_mask = 0xffffffffUL;
         dev->dev.dma_mask = &dev->dma_mask;
-        dev->dev.parent = NULL;
+        dev->dev.parent = parent;
         dev->dev.bus = &of_platform_bus_type;
         dev->dev.release = of_release_dev;
 
@@ -259,6 +261,7 @@ struct of_device* of_platform_device_create(struct device_node *np, const char *
         return dev;
 }
 
+
 EXPORT_SYMBOL(of_match_device);
 EXPORT_SYMBOL(of_platform_bus_type);
 EXPORT_SYMBOL(of_register_driver);
diff --git a/arch/ppc64/kernel/pSeries_iommu.c b/arch/ppc64/kernel/pSeries_iommu.c
index 2b5e622732f4..5914f61a152e 100644
--- a/arch/ppc64/kernel/pSeries_iommu.c
+++ b/arch/ppc64/kernel/pSeries_iommu.c
@@ -281,8 +281,10 @@ static void iommu_table_setparms(struct pci_controller *phb,
         tbl->it_offset = phb->dma_window_base_cur >> PAGE_SHIFT;
         
         /* Test if we are going over 2GB of DMA space */
-        if (phb->dma_window_base_cur + phb->dma_window_size > (1L << 31))
+        if (phb->dma_window_base_cur + phb->dma_window_size > 0x80000000ul) {
+                udbg_printf("PCI_DMA: Unexpected number of IOAs under this PHB.\n");
                 panic("PCI_DMA: Unexpected number of IOAs under this PHB.\n"); 
+        }
         
         phb->dma_window_base_cur += phb->dma_window_size;
 
@@ -326,92 +328,85 @@ static void iommu_table_setparms_lpar(struct pci_controller *phb,
 
 static void iommu_bus_setup_pSeries(struct pci_bus *bus)
 {
-        struct device_node *dn, *pdn;
-        struct pci_dn *pci;
+        struct device_node *dn;
         struct iommu_table *tbl;
+        struct device_node *isa_dn, *isa_dn_orig;
+        struct device_node *tmp;
+        struct pci_dn *pci;
+        int children;
 
         DBG("iommu_bus_setup_pSeries, bus %p, bus->self %p\n", bus, bus->self);
 
-        /* For each (root) bus, we carve up the available DMA space in 256MB
-         * pieces. Since each piece is used by one (sub) bus/device, that would
-         * give a maximum of 7 devices per PHB. In most cases, this is plenty.
-         *
-         * The exception is on Python PHBs (pre-POWER4). Here we don't have EADS
-         * bridges below the PHB to allocate the sectioned tables to, so instead
-         * we allocate a 1GB table at the PHB level.
+        dn = pci_bus_to_OF_node(bus);
+        pci = PCI_DN(dn);
+
+        if (bus->self) {
+                /* This is not a root bus, any setup will be done for the
+                 * device-side of the bridge in iommu_dev_setup_pSeries().
+                 */
+                return;
+        }
+
+        /* Check if the ISA bus on the system is under
+         * this PHB.
          */
+        isa_dn = isa_dn_orig = of_find_node_by_type(NULL, "isa");
 
-        dn = pci_bus_to_OF_node(bus);
-        pci = dn->data;
-
-        if (!bus->self) {
-                /* Root bus */
-                if (is_python(dn)) {
-                        unsigned int *iohole;
-
-                        DBG("Python root bus %s\n", bus->name);
-
-                        iohole = (unsigned int *)get_property(dn, "io-hole", 0);
-
-                        if (iohole) {
-                                /* On first bus we need to leave room for the
-                                 * ISA address space. Just skip the first 256MB
-                                 * alltogether. This leaves 768MB for the window.
-                                 */
-                                DBG("PHB has io-hole, reserving 256MB\n");
-                                pci->phb->dma_window_size = 3 << 28;
-                                pci->phb->dma_window_base_cur = 1 << 28;
-                        } else {
-                                /* 1GB window by default */
-                                pci->phb->dma_window_size = 1 << 30;
-                                pci->phb->dma_window_base_cur = 0;
-                        }
-
-                        tbl = kmalloc(sizeof(struct iommu_table), GFP_KERNEL);
-
-                        iommu_table_setparms(pci->phb, dn, tbl);
-                        pci->iommu_table = iommu_init_table(tbl);
-                } else {
-                        /* Do a 128MB table at root. This is used for the IDE
-                         * controller on some SMP-mode POWER4 machines. It
-                         * doesn't hurt to allocate it on other machines
-                         * -- it'll just be unused since new tables are
-                         * allocated on the EADS level.
-                         *
-                         * Allocate at offset 128MB to avoid having to deal
-                         * with ISA holes; 128MB table for IDE is plenty.
-                         */
-                        pci->phb->dma_window_size = 1 << 27;
-                        pci->phb->dma_window_base_cur = 1 << 27;
-
-                        tbl = kmalloc(sizeof(struct iommu_table), GFP_KERNEL);
-
-                        iommu_table_setparms(pci->phb, dn, tbl);
-                        pci->iommu_table = iommu_init_table(tbl);
-
-                        /* All child buses have 256MB tables */
-                        pci->phb->dma_window_size = 1 << 28;
-                }
-        } else {
-                pdn = pci_bus_to_OF_node(bus->parent);
+        while (isa_dn && isa_dn != dn)
+                isa_dn = isa_dn->parent;
+
+        if (isa_dn_orig)
+                of_node_put(isa_dn_orig);
 
-                if (!bus->parent->self && !is_python(pdn)) {
-                        struct iommu_table *tbl;
-                        /* First child and not python means this is the EADS
-                         * level. Allocate new table for this slot with 256MB
-                         * window.
-                         */
+        /* Count number of direct PCI children of the PHB.
+         * All PCI device nodes have class-code property, so it's
+         * an easy way to find them.
+         */
+        for (children = 0, tmp = dn->child; tmp; tmp = tmp->sibling)
+                if (get_property(tmp, "class-code", NULL))
+                        children++;
 
-                        tbl = kmalloc(sizeof(struct iommu_table), GFP_KERNEL);
+        DBG("Children: %d\n", children);
 
-                        iommu_table_setparms(pci->phb, dn, tbl);
+        /* Calculate amount of DMA window per slot. Each window must be
+         * a power of two (due to pci_alloc_consistent requirements).
+         *
+         * Keep 256MB aside for PHBs with ISA.
+         */
 
-                        pci->iommu_table = iommu_init_table(tbl);
-                } else {
-                        /* Lower than first child or under python, use parent table */
-                        pci->iommu_table = PCI_DN(pdn)->iommu_table;
-                }
+        if (!isa_dn) {
+                /* No ISA/IDE - just set window size and return */
+                pci->phb->dma_window_size = 0x80000000ul; /* To be divided */
+
+                while (pci->phb->dma_window_size * children > 0x80000000ul)
+                        pci->phb->dma_window_size >>= 1;
+                DBG("No ISA/IDE, window size is 0x%lx\n",
+                        pci->phb->dma_window_size);
+                pci->phb->dma_window_base_cur = 0;
+
+                return;
         }
+
+        /* If we have ISA, then we probably have an IDE
+         * controller too. Allocate a 128MB table but
+         * skip the first 128MB to avoid stepping on ISA
+         * space.
+         */
+        pci->phb->dma_window_size = 0x8000000ul;
+        pci->phb->dma_window_base_cur = 0x8000000ul;
+
+        tbl = kmalloc(sizeof(struct iommu_table), GFP_KERNEL);
+
+        iommu_table_setparms(pci->phb, dn, tbl);
+        pci->iommu_table = iommu_init_table(tbl);
+
+        /* Divide the rest (1.75GB) among the children */
+        pci->phb->dma_window_size = 0x80000000ul;
+        while (pci->phb->dma_window_size * children > 0x70000000ul)
+                pci->phb->dma_window_size >>= 1;
+
+        DBG("ISA/IDE, window size is 0x%lx\n", pci->phb->dma_window_size);
+
 }
 
 
@@ -462,21 +457,36 @@ static void iommu_bus_setup_pSeriesLP(struct pci_bus *bus)
 static void iommu_dev_setup_pSeries(struct pci_dev *dev)
 {
         struct device_node *dn, *mydn;
+        struct iommu_table *tbl;
+
+        DBG("iommu_dev_setup_pSeries, dev %p (%s)\n", dev, pci_name(dev));
 
-        DBG("iommu_dev_setup_pSeries, dev %p (%s)\n", dev, dev->pretty_name);
-        /* Now copy the iommu_table ptr from the bus device down to the
-         * pci device_node.  This means get_iommu_table() won't need to search
-         * up the device tree to find it.
-         */
         mydn = dn = pci_device_to_OF_node(dev);
 
+        /* If we're the direct child of a root bus, then we need to allocate
+         * an iommu table ourselves. The bus setup code should have setup
+         * the window sizes already.
+         */
+        if (!dev->bus->self) {
+                DBG(" --> first child, no bridge. Allocating iommu table.\n");
+                tbl = kmalloc(sizeof(struct iommu_table), GFP_KERNEL);
+                iommu_table_setparms(PCI_DN(dn)->phb, dn, tbl);
+                PCI_DN(mydn)->iommu_table = iommu_init_table(tbl);
+
+                return;
+        }
+
+        /* If this device is further down the bus tree, search upwards until
+         * an already allocated iommu table is found and use that.
+         */
+
         while (dn && dn->data && PCI_DN(dn)->iommu_table == NULL)
                 dn = dn->parent;
 
         if (dn && dn->data) {
                 PCI_DN(mydn)->iommu_table = PCI_DN(dn)->iommu_table;
         } else {
-                DBG("iommu_dev_setup_pSeries, dev %p (%s) has no iommu table\n", dev, dev->pretty_name);
+                DBG("iommu_dev_setup_pSeries, dev %p (%s) has no iommu table\n", dev, pci_name(dev));
         }
 }
 
@@ -510,7 +520,7 @@ static void iommu_dev_setup_pSeriesLP(struct pci_dev *dev)
         int *dma_window = NULL;
         struct pci_dn *pci;
 
-        DBG("iommu_dev_setup_pSeriesLP, dev %p (%s)\n", dev, dev->pretty_name);
+        DBG("iommu_dev_setup_pSeriesLP, dev %p (%s)\n", dev, pci_name(dev));
 
         /* dev setup for LPAR is a little tricky, since the device tree might
          * contain the dma-window properties per-device and not neccesarily
@@ -532,9 +542,8 @@ static void iommu_dev_setup_pSeriesLP(struct pci_dev *dev)
          * slots on POWER4 machines.
          */
         if (dma_window == NULL || pdn->parent == NULL) {
-                /* Fall back to regular (non-LPAR) dev setup */
-                DBG("No dma window for device, falling back to regular setup\n");
-                iommu_dev_setup_pSeries(dev);
+                DBG("No dma window for device, linking to parent\n");
+                PCI_DN(dn)->iommu_table = PCI_DN(pdn)->iommu_table;
                 return;
         } else {
                 DBG("Found DMA window, allocating table\n");
diff --git a/arch/ppc64/kernel/pci.c b/arch/ppc64/kernel/pci.c
index 861138ad092c..ff4be1da69d5 100644
--- a/arch/ppc64/kernel/pci.c
+++ b/arch/ppc64/kernel/pci.c
@@ -246,11 +246,14 @@ static unsigned int pci_parse_of_flags(u32 addr0)
         unsigned int flags = 0;
 
         if (addr0 & 0x02000000) {
-                flags |= IORESOURCE_MEM;
+                flags = IORESOURCE_MEM | PCI_BASE_ADDRESS_SPACE_MEMORY;
+                flags |= (addr0 >> 22) & PCI_BASE_ADDRESS_MEM_TYPE_64;
+                flags |= (addr0 >> 28) & PCI_BASE_ADDRESS_MEM_TYPE_1M;
                 if (addr0 & 0x40000000)
-                        flags |= IORESOURCE_PREFETCH;
+                        flags |= IORESOURCE_PREFETCH
+                                 | PCI_BASE_ADDRESS_MEM_PREFETCH;
         } else if (addr0 & 0x01000000)
-                flags |= IORESOURCE_IO;
+                flags = IORESOURCE_IO | PCI_BASE_ADDRESS_SPACE_IO;
         return flags;
 }
 
diff --git a/arch/ppc64/kernel/pmac_setup.c b/arch/ppc64/kernel/pmac_setup.c
index 1f5118078f7d..bb0c3bfbb7e2 100644
--- a/arch/ppc64/kernel/pmac_setup.c
+++ b/arch/ppc64/kernel/pmac_setup.c
@@ -434,15 +434,23 @@ static int pmac_check_legacy_ioport(unsigned int baseport)
 
 static int __init pmac_declare_of_platform_devices(void)
 {
-        struct device_node *np;
+        struct device_node *np, *npp;
 
-        np = find_devices("u3");
-        if (np) {
-                for (np = np->child; np != NULL; np = np->sibling)
+        npp = of_find_node_by_name(NULL, "u3");
+        if (npp) {
+                for (np = NULL; (np = of_get_next_child(npp, np)) != NULL;) {
                         if (strncmp(np->name, "i2c", 3) == 0) {
-                                of_platform_device_create(np, "u3-i2c");
+                                of_platform_device_create(np, "u3-i2c", NULL);
+                                of_node_put(np);
                                 break;
                         }
+                }
+                of_node_put(npp);
+        }
+        npp = of_find_node_by_type(NULL, "smu");
+        if (npp) {
+                of_platform_device_create(npp, "smu", NULL);
+                of_node_put(npp);
         }
 
         return 0;
diff --git a/arch/ppc64/kernel/pmac_time.c b/arch/ppc64/kernel/pmac_time.c
index 6c8c99295e7a..9d8c97decd32 100644
--- a/arch/ppc64/kernel/pmac_time.c
+++ b/arch/ppc64/kernel/pmac_time.c
@@ -84,7 +84,7 @@ void pmac_get_rtc_time(struct rtc_time *tm)
 
 #ifdef CONFIG_PMAC_SMU
         case SYS_CTRLER_SMU:
-                smu_get_rtc_time(tm);
+                smu_get_rtc_time(tm, 1);
                 break;
 #endif /* CONFIG_PMAC_SMU */
         default:
@@ -128,7 +128,7 @@ int pmac_set_rtc_time(struct rtc_time *tm)
 
 #ifdef CONFIG_PMAC_SMU
         case SYS_CTRLER_SMU:
-                return smu_set_rtc_time(tm);
+                return smu_set_rtc_time(tm, 1);
 #endif /* CONFIG_PMAC_SMU */
         default:
                 return -ENODEV;
diff --git a/arch/ppc64/kernel/prom_init.c b/arch/ppc64/kernel/prom_init.c
index 9979919cdf92..f252670874a4 100644
--- a/arch/ppc64/kernel/prom_init.c
+++ b/arch/ppc64/kernel/prom_init.c
@@ -1711,6 +1711,7 @@ static void __init flatten_device_tree(void)
         unsigned long offset = reloc_offset();
         unsigned long mem_start, mem_end, room;
         struct boot_param_header *hdr;
+        struct prom_t *_prom = PTRRELOC(&prom);
         char *namep;
         u64 *rsvmap;
 
@@ -1765,6 +1766,7 @@ static void __init flatten_device_tree(void)
         RELOC(dt_struct_end) = PAGE_ALIGN(mem_start);
 
         /* Finish header */
+        hdr->boot_cpuid_phys = _prom->cpu;
         hdr->magic = OF_DT_HEADER;
         hdr->totalsize = RELOC(dt_struct_end) - RELOC(dt_header_start);
         hdr->off_dt_struct = RELOC(dt_struct_start) - RELOC(dt_header_start);
@@ -1854,7 +1856,6 @@ static void __init prom_find_boot_cpu(void)
 
         cpu_pkg = call_prom("instance-to-package", 1, 1, prom_cpu);
 
-        prom_setprop(cpu_pkg, "linux,boot-cpu", NULL, 0);
         prom_getprop(cpu_pkg, "reg", &getprop_rval, sizeof(getprop_rval));
         _prom->cpu = getprop_rval;
 
diff --git a/arch/ppc64/kernel/ptrace.c b/arch/ppc64/kernel/ptrace.c
index 85ed3188a91d..b1c044ca5756 100644
--- a/arch/ppc64/kernel/ptrace.c
+++ b/arch/ppc64/kernel/ptrace.c
@@ -219,6 +219,7 @@ int sys_ptrace(long request, long pid, long addr, long data)
 
         case PTRACE_SET_DEBUGREG:
                 ret = ptrace_set_debugreg(child, addr, data);
+                break;
 
         case PTRACE_DETACH:
                 ret = ptrace_detach(child, data);
diff --git a/arch/ppc64/mm/hash_native.c b/arch/ppc64/mm/hash_native.c
index 29b074505d3e..874cf96938ff 100644
--- a/arch/ppc64/mm/hash_native.c
+++ b/arch/ppc64/mm/hash_native.c
@@ -342,15 +342,14 @@ static void native_flush_hash_range(unsigned long number, int local)
         hpte_t *hptep;
         unsigned long hpte_v;
         struct ppc64_tlb_batch *batch = &__get_cpu_var(ppc64_tlb_batch);
-
-        /* XXX fix for large ptes */
-        unsigned long large = 0;
+        unsigned long large;
 
         local_irq_save(flags);
 
         j = 0;
         for (i = 0; i < number; i++) {
                 va = batch->vaddr[j];
+                large = pte_huge(batch->pte[i]);
                 if (large)
                         vpn = va >> HPAGE_SHIFT;
                 else
diff --git a/arch/ppc64/mm/hugetlbpage.c b/arch/ppc64/mm/hugetlbpage.c
index 338771ec70d7..0ea0994ed974 100644
--- a/arch/ppc64/mm/hugetlbpage.c
+++ b/arch/ppc64/mm/hugetlbpage.c
@@ -710,10 +710,13 @@ repeat:
                         hpte_group = ((~hash & htab_hash_mask) *
                                       HPTES_PER_GROUP) & ~0x7UL; 
                         slot = ppc_md.hpte_insert(hpte_group, va, prpn,
-                                                  HPTE_V_LARGE, rflags);
+                                                  HPTE_V_LARGE |
+                                                  HPTE_V_SECONDARY,
+                                                  rflags);
                         if (slot == -1) {
                                 if (mftb() & 0x1)
-                                        hpte_group = ((hash & htab_hash_mask) * HPTES_PER_GROUP) & ~0x7UL;
+                                        hpte_group = ((hash & htab_hash_mask) *
+                                                      HPTES_PER_GROUP)&~0x7UL;
 
                                 ppc_md.hpte_remove(hpte_group);
                                 goto repeat;