Merge branch 'master' of /usr/src/ntfs-2.6/

16 files changed, 1239 insertions, 495 deletions

diff --git a/arch/powerpc/platforms/pseries/Makefile b/arch/powerpc/platforms/pseries/Makefile
index 06d5ef50121..61616d14407 100644
--- a/arch/powerpc/platforms/pseries/Makefile
+++ b/arch/powerpc/platforms/pseries/Makefile
@@ -1,10 +1,10 @@
 obj-y			:= pci.o lpar.o hvCall.o nvram.o reconfig.o \
-			   setup.o iommu.o ras.o rtasd.o
+			   setup.o iommu.o ras.o rtasd.o pci_dlpar.o
 obj-$(CONFIG_SMP)	+= smp.o
 obj-$(CONFIG_IBMVIO)	+= vio.o
 obj-$(CONFIG_XICS)	+= xics.o
 obj-$(CONFIG_SCANLOG)	+= scanlog.o
-obj-$(CONFIG_EEH)	+= eeh.o eeh_event.o
+obj-$(CONFIG_EEH)	+= eeh.o eeh_cache.o eeh_driver.o eeh_event.o
 
 obj-$(CONFIG_HVC_CONSOLE)	+= hvconsole.o
 obj-$(CONFIG_HVCS)		+= hvcserver.o
diff --git a/arch/powerpc/platforms/pseries/eeh.c b/arch/powerpc/platforms/pseries/eeh.c
index c8d2a40dc5b..83578313ee7 100644
--- a/arch/powerpc/platforms/pseries/eeh.c
+++ b/arch/powerpc/platforms/pseries/eeh.c
@@ -76,15 +76,14 @@
  */
 #define EEH_MAX_FAILS	100000
 
-/* Misc forward declaraions */
-static void eeh_save_bars(struct pci_dev * pdev, struct pci_dn *pdn);
-
 /* RTAS tokens */
 static int ibm_set_eeh_option;
 static int ibm_set_slot_reset;
 static int ibm_read_slot_reset_state;
 static int ibm_read_slot_reset_state2;
 static int ibm_slot_error_detail;
+static int ibm_get_config_addr_info;
+static int ibm_configure_bridge;
 
 int eeh_subsystem_enabled;
 EXPORT_SYMBOL(eeh_subsystem_enabled);
@@ -98,308 +97,23 @@ static DEFINE_SPINLOCK(slot_errbuf_lock);
 static int eeh_error_buf_size;
 
 /* System monitoring statistics */
-static DEFINE_PER_CPU(unsigned long, no_device);
-static DEFINE_PER_CPU(unsigned long, no_dn);
-static DEFINE_PER_CPU(unsigned long, no_cfg_addr);
-static DEFINE_PER_CPU(unsigned long, ignored_check);
-static DEFINE_PER_CPU(unsigned long, total_mmio_ffs);
-static DEFINE_PER_CPU(unsigned long, false_positives);
-static DEFINE_PER_CPU(unsigned long, ignored_failures);
-static DEFINE_PER_CPU(unsigned long, slot_resets);
-
-/**
- * The pci address cache subsystem.  This subsystem places
- * PCI device address resources into a red-black tree, sorted
- * according to the address range, so that given only an i/o
- * address, the corresponding PCI device can be **quickly**
- * found. It is safe to perform an address lookup in an interrupt
- * context; this ability is an important feature.
- *
- * Currently, the only customer of this code is the EEH subsystem;
- * thus, this code has been somewhat tailored to suit EEH better.
- * In particular, the cache does *not* hold the addresses of devices
- * for which EEH is not enabled.
- *
- * (Implementation Note: The RB tree seems to be better/faster
- * than any hash algo I could think of for this problem, even
- * with the penalty of slow pointer chases for d-cache misses).
- */
-struct pci_io_addr_range
-{
-	struct rb_node rb_node;
-	unsigned long addr_lo;
-	unsigned long addr_hi;
-	struct pci_dev *pcidev;
-	unsigned int flags;
-};
-
-static struct pci_io_addr_cache
-{
-	struct rb_root rb_root;
-	spinlock_t piar_lock;
-} pci_io_addr_cache_root;
-
-static inline struct pci_dev *__pci_get_device_by_addr(unsigned long addr)
-{
-	struct rb_node *n = pci_io_addr_cache_root.rb_root.rb_node;
-
-	while (n) {
-		struct pci_io_addr_range *piar;
-		piar = rb_entry(n, struct pci_io_addr_range, rb_node);
-
-		if (addr < piar->addr_lo) {
-			n = n->rb_left;
-		} else {
-			if (addr > piar->addr_hi) {
-				n = n->rb_right;
-			} else {
-				pci_dev_get(piar->pcidev);
-				return piar->pcidev;
-			}
-		}
-	}
-
-	return NULL;
-}
-
-/**
- * pci_get_device_by_addr - Get device, given only address
- * @addr: mmio (PIO) phys address or i/o port number
- *
- * Given an mmio phys address, or a port number, find a pci device
- * that implements this address.  Be sure to pci_dev_put the device
- * when finished.  I/O port numbers are assumed to be offset
- * from zero (that is, they do *not* have pci_io_addr added in).
- * It is safe to call this function within an interrupt.
- */
-static struct pci_dev *pci_get_device_by_addr(unsigned long addr)
-{
-	struct pci_dev *dev;
-	unsigned long flags;
-
-	spin_lock_irqsave(&pci_io_addr_cache_root.piar_lock, flags);
-	dev = __pci_get_device_by_addr(addr);
-	spin_unlock_irqrestore(&pci_io_addr_cache_root.piar_lock, flags);
-	return dev;
-}
-
-#ifdef DEBUG
-/*
- * Handy-dandy debug print routine, does nothing more
- * than print out the contents of our addr cache.
- */
-static void pci_addr_cache_print(struct pci_io_addr_cache *cache)
-{
-	struct rb_node *n;
-	int cnt = 0;
-
-	n = rb_first(&cache->rb_root);
-	while (n) {
-		struct pci_io_addr_range *piar;
-		piar = rb_entry(n, struct pci_io_addr_range, rb_node);
-		printk(KERN_DEBUG "PCI: %s addr range %d [%lx-%lx]: %s\n",
-		       (piar->flags & IORESOURCE_IO) ? "i/o" : "mem", cnt,
-		       piar->addr_lo, piar->addr_hi, pci_name(piar->pcidev));
-		cnt++;
-		n = rb_next(n);
-	}
-}
-#endif
-
-/* Insert address range into the rb tree. */
-static struct pci_io_addr_range *
-pci_addr_cache_insert(struct pci_dev *dev, unsigned long alo,
-		      unsigned long ahi, unsigned int flags)
-{
-	struct rb_node **p = &pci_io_addr_cache_root.rb_root.rb_node;
-	struct rb_node *parent = NULL;
-	struct pci_io_addr_range *piar;
-
-	/* Walk tree, find a place to insert into tree */
-	while (*p) {
-		parent = *p;
-		piar = rb_entry(parent, struct pci_io_addr_range, rb_node);
-		if (ahi < piar->addr_lo) {
-			p = &parent->rb_left;
-		} else if (alo > piar->addr_hi) {
-			p = &parent->rb_right;
-		} else {
-			if (dev != piar->pcidev ||
-			    alo != piar->addr_lo || ahi != piar->addr_hi) {
-				printk(KERN_WARNING "PIAR: overlapping address range\n");
-			}
-			return piar;
-		}
-	}
-	piar = (struct pci_io_addr_range *)kmalloc(sizeof(struct pci_io_addr_range), GFP_ATOMIC);
-	if (!piar)
-		return NULL;
-
-	piar->addr_lo = alo;
-	piar->addr_hi = ahi;
-	piar->pcidev = dev;
-	piar->flags = flags;
-
-#ifdef DEBUG
-	printk(KERN_DEBUG "PIAR: insert range=[%lx:%lx] dev=%s\n",
-	                  alo, ahi, pci_name (dev));
-#endif
+static unsigned long no_device;
+static unsigned long no_dn;
+static unsigned long no_cfg_addr;
+static unsigned long ignored_check;
+static unsigned long total_mmio_ffs;
+static unsigned long false_positives;
+static unsigned long ignored_failures;
+static unsigned long slot_resets;
 
-	rb_link_node(&piar->rb_node, parent, p);
-	rb_insert_color(&piar->rb_node, &pci_io_addr_cache_root.rb_root);
-
-	return piar;
-}
-
-static void __pci_addr_cache_insert_device(struct pci_dev *dev)
-{
-	struct device_node *dn;
-	struct pci_dn *pdn;
-	int i;
-	int inserted = 0;
-
-	dn = pci_device_to_OF_node(dev);
-	if (!dn) {
-		printk(KERN_WARNING "PCI: no pci dn found for dev=%s\n", pci_name(dev));
-		return;
-	}
-
-	/* Skip any devices for which EEH is not enabled. */
-	pdn = PCI_DN(dn);
-	if (!(pdn->eeh_mode & EEH_MODE_SUPPORTED) ||
-	    pdn->eeh_mode & EEH_MODE_NOCHECK) {
-#ifdef DEBUG
-		printk(KERN_INFO "PCI: skip building address cache for=%s - %s\n",
-		       pci_name(dev), pdn->node->full_name);
-#endif
-		return;
-	}
-
-	/* The cache holds a reference to the device... */
-	pci_dev_get(dev);
-
-	/* Walk resources on this device, poke them into the tree */
-	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
-		unsigned long start = pci_resource_start(dev,i);
-		unsigned long end = pci_resource_end(dev,i);
-		unsigned int flags = pci_resource_flags(dev,i);
-
-		/* We are interested only bus addresses, not dma or other stuff */
-		if (0 == (flags & (IORESOURCE_IO | IORESOURCE_MEM)))
-			continue;
-		if (start == 0 || ~start == 0 || end == 0 || ~end == 0)
-			 continue;
-		pci_addr_cache_insert(dev, start, end, flags);
-		inserted = 1;
-	}
-
-	/* If there was nothing to add, the cache has no reference... */
-	if (!inserted)
-		pci_dev_put(dev);
-}
-
-/**
- * pci_addr_cache_insert_device - Add a device to the address cache
- * @dev: PCI device whose I/O addresses we are interested in.
- *
- * In order to support the fast lookup of devices based on addresses,
- * we maintain a cache of devices that can be quickly searched.
- * This routine adds a device to that cache.
- */
-static void pci_addr_cache_insert_device(struct pci_dev *dev)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&pci_io_addr_cache_root.piar_lock, flags);
-	__pci_addr_cache_insert_device(dev);
-	spin_unlock_irqrestore(&pci_io_addr_cache_root.piar_lock, flags);
-}
-
-static inline void __pci_addr_cache_remove_device(struct pci_dev *dev)
-{
-	struct rb_node *n;
-	int removed = 0;
-
-restart:
-	n = rb_first(&pci_io_addr_cache_root.rb_root);
-	while (n) {
-		struct pci_io_addr_range *piar;
-		piar = rb_entry(n, struct pci_io_addr_range, rb_node);
-
-		if (piar->pcidev == dev) {
-			rb_erase(n, &pci_io_addr_cache_root.rb_root);
-			removed = 1;
-			kfree(piar);
-			goto restart;
-		}
-		n = rb_next(n);
-	}
-
-	/* The cache no longer holds its reference to this device... */
-	if (removed)
-		pci_dev_put(dev);
-}
-
-/**
- * pci_addr_cache_remove_device - remove pci device from addr cache
- * @dev: device to remove
- *
- * Remove a device from the addr-cache tree.
- * This is potentially expensive, since it will walk
- * the tree multiple times (once per resource).
- * But so what; device removal doesn't need to be that fast.
- */
-static void pci_addr_cache_remove_device(struct pci_dev *dev)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&pci_io_addr_cache_root.piar_lock, flags);
-	__pci_addr_cache_remove_device(dev);
-	spin_unlock_irqrestore(&pci_io_addr_cache_root.piar_lock, flags);
-}
-
-/**
- * pci_addr_cache_build - Build a cache of I/O addresses
- *
- * Build a cache of pci i/o addresses.  This cache will be used to
- * find the pci device that corresponds to a given address.
- * This routine scans all pci busses to build the cache.
- * Must be run late in boot process, after the pci controllers
- * have been scaned for devices (after all device resources are known).
- */
-void __init pci_addr_cache_build(void)
-{
-	struct device_node *dn;
-	struct pci_dev *dev = NULL;
-
-	if (!eeh_subsystem_enabled)
-		return;
-
-	spin_lock_init(&pci_io_addr_cache_root.piar_lock);
-
-	while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) {
-		/* Ignore PCI bridges ( XXX why ??) */
-		if ((dev->class >> 16) == PCI_BASE_CLASS_BRIDGE) {
-			continue;
-		}
-		pci_addr_cache_insert_device(dev);
-
-		/* Save the BAR's; firmware doesn't restore these after EEH reset */
-		dn = pci_device_to_OF_node(dev);
-		eeh_save_bars(dev, PCI_DN(dn));
-	}
-
-#ifdef DEBUG
-	/* Verify tree built up above, echo back the list of addrs. */
-	pci_addr_cache_print(&pci_io_addr_cache_root);
-#endif
-}
+#define IS_BRIDGE(class_code) (((class_code)<<16) == PCI_BASE_CLASS_BRIDGE)
 
 /* --------------------------------------------------------------- */
-/* Above lies the PCI Address Cache. Below lies the EEH event infrastructure */
+/* Below lies the EEH event infrastructure */
 
 void eeh_slot_error_detail (struct pci_dn *pdn, int severity)
 {
+	int config_addr;
 	unsigned long flags;
 	int rc;
 
@@ -407,8 +121,13 @@ void eeh_slot_error_detail (struct pci_dn *pdn, int severity)
 	spin_lock_irqsave(&slot_errbuf_lock, flags);
 	memset(slot_errbuf, 0, eeh_error_buf_size);
 
+	/* Use PE configuration address, if present */
+	config_addr = pdn->eeh_config_addr;
+	if (pdn->eeh_pe_config_addr)
+		config_addr = pdn->eeh_pe_config_addr;
+
 	rc = rtas_call(ibm_slot_error_detail,
-	               8, 1, NULL, pdn->eeh_config_addr,
+	               8, 1, NULL, config_addr,
 	               BUID_HI(pdn->phb->buid),
 	               BUID_LO(pdn->phb->buid), NULL, 0,
 	               virt_to_phys(slot_errbuf),
@@ -428,6 +147,7 @@ void eeh_slot_error_detail (struct pci_dn *pdn, int severity)
 static int read_slot_reset_state(struct pci_dn *pdn, int rets[])
 {
 	int token, outputs;
+	int config_addr;
 
 	if (ibm_read_slot_reset_state2 != RTAS_UNKNOWN_SERVICE) {
 		token = ibm_read_slot_reset_state2;
@@ -438,7 +158,12 @@ static int read_slot_reset_state(struct pci_dn *pdn, int rets[])
 		outputs = 3;
 	}
 
-	return rtas_call(token, 3, outputs, rets, pdn->eeh_config_addr,
+	/* Use PE configuration address, if present */
+	config_addr = pdn->eeh_config_addr;
+	if (pdn->eeh_pe_config_addr)
+		config_addr = pdn->eeh_pe_config_addr;
+
+	return rtas_call(token, 3, outputs, rets, config_addr,
 			 BUID_HI(pdn->phb->buid), BUID_LO(pdn->phb->buid));
 }
 
@@ -462,7 +187,7 @@ static inline unsigned long eeh_token_to_phys(unsigned long token)
 /** 
  * Return the "partitionable endpoint" (pe) under which this device lies
  */
-static struct device_node * find_device_pe(struct device_node *dn)
+struct device_node * find_device_pe(struct device_node *dn)
 {
 	while ((dn->parent) && PCI_DN(dn->parent) &&
 	      (PCI_DN(dn->parent)->eeh_mode & EEH_MODE_SUPPORTED)) {
@@ -483,8 +208,14 @@ static void __eeh_mark_slot (struct device_node *dn, int mode_flag)
 {
 	while (dn) {
 		if (PCI_DN(dn)) {
+			/* Mark the pci device driver too */
+			struct pci_dev *dev = PCI_DN(dn)->pcidev;
+
 			PCI_DN(dn)->eeh_mode |= mode_flag;
 
+			if (dev && dev->driver)
+				dev->error_state = pci_channel_io_frozen;
+
 			if (dn->child)
 				__eeh_mark_slot (dn->child, mode_flag);
 		}
@@ -495,6 +226,11 @@ static void __eeh_mark_slot (struct device_node *dn, int mode_flag)
 void eeh_mark_slot (struct device_node *dn, int mode_flag)
 {
 	dn = find_device_pe (dn);
+
+	/* Back up one, since config addrs might be shared */
+	if (PCI_DN(dn) && PCI_DN(dn)->eeh_pe_config_addr)
+		dn = dn->parent;
+
 	PCI_DN(dn)->eeh_mode |= mode_flag;
 	__eeh_mark_slot (dn->child, mode_flag);
 }
@@ -516,7 +252,13 @@ void eeh_clear_slot (struct device_node *dn, int mode_flag)
 {
 	unsigned long flags;
 	spin_lock_irqsave(&confirm_error_lock, flags);
+	
 	dn = find_device_pe (dn);
+	
+	/* Back up one, since config addrs might be shared */
+	if (PCI_DN(dn) && PCI_DN(dn)->eeh_pe_config_addr)
+		dn = dn->parent;
+
 	PCI_DN(dn)->eeh_mode &= ~mode_flag;
 	PCI_DN(dn)->eeh_check_count = 0;
 	__eeh_clear_slot (dn->child, mode_flag);
@@ -544,15 +286,16 @@ int eeh_dn_check_failure(struct device_node *dn, struct pci_dev *dev)
 	int rets[3];
 	unsigned long flags;
 	struct pci_dn *pdn;
+	enum pci_channel_state state;
 	int rc = 0;
 
-	__get_cpu_var(total_mmio_ffs)++;
+	total_mmio_ffs++;
 
 	if (!eeh_subsystem_enabled)
 		return 0;
 
 	if (!dn) {
-		__get_cpu_var(no_dn)++;
+		no_dn++;
 		return 0;
 	}
 	pdn = PCI_DN(dn);
@@ -560,7 +303,7 @@ int eeh_dn_check_failure(struct device_node *dn, struct pci_dev *dev)
 	/* Access to IO BARs might get this far and still not want checking. */
 	if (!(pdn->eeh_mode & EEH_MODE_SUPPORTED) ||
 	    pdn->eeh_mode & EEH_MODE_NOCHECK) {
-		__get_cpu_var(ignored_check)++;
+		ignored_check++;
 #ifdef DEBUG
 		printk ("EEH:ignored check (%x) for %s %s\n", 
 		        pdn->eeh_mode, pci_name (dev), dn->full_name);
@@ -568,8 +311,8 @@ int eeh_dn_check_failure(struct device_node *dn, struct pci_dev *dev)
 		return 0;
 	}
 
-	if (!pdn->eeh_config_addr) {
-		__get_cpu_var(no_cfg_addr)++;
+	if (!pdn->eeh_config_addr && !pdn->eeh_pe_config_addr) {
+		no_cfg_addr++;
 		return 0;
 	}
 
@@ -611,7 +354,7 @@ int eeh_dn_check_failure(struct device_node *dn, struct pci_dev *dev)
 	if (ret != 0) {
 		printk(KERN_WARNING "EEH: read_slot_reset_state() failed; rc=%d dn=%s\n",
 		       ret, dn->full_name);
-		__get_cpu_var(false_positives)++;
+		false_positives++;
 		rc = 0;
 		goto dn_unlock;
 	}
@@ -620,14 +363,14 @@ int eeh_dn_check_failure(struct device_node *dn, struct pci_dev *dev)
 	if (rets[1] != 1) {
 		printk(KERN_WARNING "EEH: event on unsupported device, rc=%d dn=%s\n",
 		       ret, dn->full_name);
-		__get_cpu_var(false_positives)++;
+		false_positives++;
 		rc = 0;
 		goto dn_unlock;
 	}
 
 	/* If not the kind of error we know about, punt. */
 	if (rets[0] != 2 && rets[0] != 4 && rets[0] != 5) {
-		__get_cpu_var(false_positives)++;
+		false_positives++;
 		rc = 0;
 		goto dn_unlock;
 	}
@@ -635,12 +378,12 @@ int eeh_dn_check_failure(struct device_node *dn, struct pci_dev *dev)
 	/* Note that config-io to empty slots may fail;
 	 * we recognize empty because they don't have children. */
 	if ((rets[0] == 5) && (dn->child == NULL)) {
-		__get_cpu_var(false_positives)++;
+		false_positives++;
 		rc = 0;
 		goto dn_unlock;
 	}
 
-	__get_cpu_var(slot_resets)++;
+	slot_resets++;
  
 	/* Avoid repeated reports of this failure, including problems
 	 * with other functions on this device, and functions under
@@ -648,8 +391,13 @@ int eeh_dn_check_failure(struct device_node *dn, struct pci_dev *dev)
 	eeh_mark_slot (dn, EEH_MODE_ISOLATED);
 	spin_unlock_irqrestore(&confirm_error_lock, flags);
 
-	eeh_send_failure_event (dn, dev, rets[0], rets[2]);
-	
+	state = pci_channel_io_normal;
+	if ((rets[0] == 2) || (rets[0] == 4))
+		state = pci_channel_io_frozen;
+	if (rets[0] == 5)
+		state = pci_channel_io_perm_failure;
+	eeh_send_failure_event (dn, dev, state, rets[2]);
+
 	/* Most EEH events are due to device driver bugs.  Having
 	 * a stack trace will help the device-driver authors figure
 	 * out what happened.  So print that out. */
@@ -685,7 +433,7 @@ unsigned long eeh_check_failure(const volatile void __iomem *token, unsigned lon
 	addr = eeh_token_to_phys((unsigned long __force) token);
 	dev = pci_get_device_by_addr(addr);
 	if (!dev) {
-		__get_cpu_var(no_device)++;
+		no_device++;
 		return val;
 	}
 
@@ -716,11 +464,16 @@ eeh_slot_availability(struct pci_dn *pdn)
 	if (rc) return rc;
 
 	if (rets[1] == 0) return -1;  /* EEH is not supported */
-	if (rets[0] == 0)  return 0;  /* Oll Korrect */
+	if (rets[0] == 0) return 0;   /* Oll Korrect */
 	if (rets[0] == 5) {
 		if (rets[2] == 0) return -1; /* permanently unavailable */
 		return rets[2]; /* number of millisecs to wait */
 	}
+	if (rets[0] == 1)
+		return 250;
+
+	printk (KERN_ERR "EEH: Slot unavailable: rc=%d, rets=%d %d %d\n",
+		rc, rets[0], rets[1], rets[2]);
 	return -1;
 }
 
@@ -737,6 +490,7 @@ eeh_slot_availability(struct pci_dn *pdn)
 static void
 rtas_pci_slot_reset(struct pci_dn *pdn, int state)
 {
+	int config_addr;
 	int rc;
 
 	BUG_ON (pdn==NULL); 
@@ -747,8 +501,13 @@ rtas_pci_slot_reset(struct pci_dn *pdn, int state)
 		return;
 	}
 
+	/* Use PE configuration address, if present */
+	config_addr = pdn->eeh_config_addr;
+	if (pdn->eeh_pe_config_addr)
+		config_addr = pdn->eeh_pe_config_addr;
+
 	rc = rtas_call(ibm_set_slot_reset,4,1, NULL,
-	               pdn->eeh_config_addr,
+	               config_addr,
 	               BUID_HI(pdn->phb->buid),
 	               BUID_LO(pdn->phb->buid),
 	               state);
@@ -761,9 +520,11 @@ rtas_pci_slot_reset(struct pci_dn *pdn, int state)
 
 /** rtas_set_slot_reset -- assert the pci #RST line for 1/4 second
  *  dn -- device node to be reset.
+ *
+ *  Return 0 if success, else a non-zero value.
  */
 
-void
+int
 rtas_set_slot_reset(struct pci_dn *pdn)
 {
 	int i, rc;
@@ -793,10 +554,21 @@ rtas_set_slot_reset(struct pci_dn *pdn)
 	 * ready to be used; if not, wait for recovery. */
 	for (i=0; i<10; i++) {
 		rc = eeh_slot_availability (pdn);
-		if (rc <= 0) break;
+		if (rc < 0)
+			printk (KERN_ERR "EEH: failed (%d) to reset slot %s\n", rc, pdn->node->full_name);
+		if (rc == 0)
+			return 0;
+		if (rc < 0)
+			return -1;
 
 		msleep (rc+100);
 	}
+
+	rc = eeh_slot_availability (pdn);
+	if (rc)
+		printk (KERN_ERR "EEH: timeout resetting slot %s\n", pdn->node->full_name);
+
+	return rc;
 }
 
 /* ------------------------------------------------------- */
@@ -851,7 +623,7 @@ void eeh_restore_bars(struct pci_dn *pdn)
 	if (!pdn) 
 		return;
 	
-	if (! pdn->eeh_is_bridge)
+	if ((pdn->eeh_mode & EEH_MODE_SUPPORTED) && !IS_BRIDGE(pdn->class_code))
 		__restore_bars (pdn);
 
 	dn = pdn->node->child;
@@ -869,30 +641,30 @@ void eeh_restore_bars(struct pci_dn *pdn)
  * PCI devices are added individuallly; but, for the restore,
  * an entire slot is reset at a time.
  */
-static void eeh_save_bars(struct pci_dev * pdev, struct pci_dn *pdn)
+static void eeh_save_bars(struct pci_dn *pdn)
 {
 	int i;
 
-	if (!pdev || !pdn )
+	if (!pdn )
 		return;
 	
 	for (i = 0; i < 16; i++)
-		pci_read_config_dword(pdev, i * 4, &pdn->config_space[i]);
-
-	if (pdev->hdr_type == PCI_HEADER_TYPE_BRIDGE)
-		pdn->eeh_is_bridge = 1;
+		rtas_read_config(pdn, i * 4, 4, &pdn->config_space[i]);
 }
 
 void
 rtas_configure_bridge(struct pci_dn *pdn)
 {
-	int token = rtas_token ("ibm,configure-bridge");
+	int config_addr;
 	int rc;
 
-	if (token == RTAS_UNKNOWN_SERVICE)
-		return;
-	rc = rtas_call(token,3,1, NULL,
-	               pdn->eeh_config_addr,
+	/* Use PE configuration address, if present */
+	config_addr = pdn->eeh_config_addr;
+	if (pdn->eeh_pe_config_addr)
+		config_addr = pdn->eeh_pe_config_addr;
+
+	rc = rtas_call(ibm_configure_bridge,3,1, NULL,
+	               config_addr,
 	               BUID_HI(pdn->phb->buid),
 	               BUID_LO(pdn->phb->buid));
 	if (rc) {
@@ -927,6 +699,7 @@ static void *early_enable_eeh(struct device_node *dn, void *data)
 	int enable;
 	struct pci_dn *pdn = PCI_DN(dn);
 
+	pdn->class_code = 0;
 	pdn->eeh_mode = 0;
 	pdn->eeh_check_count = 0;
 	pdn->eeh_freeze_count = 0;
@@ -943,6 +716,7 @@ static void *early_enable_eeh(struct device_node *dn, void *data)
 		pdn->eeh_mode |= EEH_MODE_NOCHECK;
 		return NULL;
 	}
+	pdn->class_code = *class_code;
 
 	/*
 	 * Now decide if we are going to "Disable" EEH checking
@@ -953,8 +727,10 @@ static void *early_enable_eeh(struct device_node *dn, void *data)
 	 * But there are a few cases like display devices that make sense.
 	 */
 	enable = 1;	/* i.e. we will do checking */
+#if 0
 	if ((*class_code >> 16) == PCI_BASE_CLASS_DISPLAY)
 		enable = 0;
+#endif
 
 	if (!enable)
 		pdn->eeh_mode |= EEH_MODE_NOCHECK;
@@ -973,8 +749,22 @@ static void *early_enable_eeh(struct device_node *dn, void *data)
 			eeh_subsystem_enabled = 1;
 			pdn->eeh_mode |= EEH_MODE_SUPPORTED;
 			pdn->eeh_config_addr = regs[0];
+
+			/* If the newer, better, ibm,get-config-addr-info is supported, 
+			 * then use that instead. */
+			pdn->eeh_pe_config_addr = 0;
+			if (ibm_get_config_addr_info != RTAS_UNKNOWN_SERVICE) {
+				unsigned int rets[2];
+				ret = rtas_call (ibm_get_config_addr_info, 4, 2, rets, 
+					pdn->eeh_config_addr, 
+					info->buid_hi, info->buid_lo,
+					0);
+				if (ret == 0)
+					pdn->eeh_pe_config_addr = rets[0];
+			}
 #ifdef DEBUG
-			printk(KERN_DEBUG "EEH: %s: eeh enabled\n", dn->full_name);
+			printk(KERN_DEBUG "EEH: %s: eeh enabled, config=%x pe_config=%x\n",
+			       dn->full_name, pdn->eeh_config_addr, pdn->eeh_pe_config_addr);
 #endif
 		} else {
 
@@ -993,6 +783,7 @@ static void *early_enable_eeh(struct device_node *dn, void *data)
 		       dn->full_name);
 	}
 
+	eeh_save_bars(pdn);
 	return NULL;
 }
 
@@ -1026,6 +817,8 @@ void __init eeh_init(void)
 	ibm_read_slot_reset_state2 = rtas_token("ibm,read-slot-reset-state2");
 	ibm_read_slot_reset_state = rtas_token("ibm,read-slot-reset-state");
 	ibm_slot_error_detail = rtas_token("ibm,slot-error-detail");
+	ibm_get_config_addr_info = rtas_token("ibm,get-config-addr-info");
+	ibm_configure_bridge = rtas_token ("ibm,configure-bridge");
 
 	if (ibm_set_eeh_option == RTAS_UNKNOWN_SERVICE)
 		return;
@@ -1080,12 +873,10 @@ void eeh_add_device_early(struct device_node *dn)
 	if (!dn || !PCI_DN(dn))
 		return;
 	phb = PCI_DN(dn)->phb;
-	if (NULL == phb || 0 == phb->buid) {
-		printk(KERN_WARNING "EEH: Expected buid but found none for %s\n",
-		       dn->full_name);
-		dump_stack();
+
+	/* USB Bus children of PCI devices will not have BUID's */
+	if (NULL == phb || 0 == phb->buid)
 		return;
-	}
 
 	info.buid_hi = BUID_HI(phb->buid);
 	info.buid_lo = BUID_LO(phb->buid);
@@ -1093,6 +884,15 @@ void eeh_add_device_early(struct device_node *dn)
 }
 EXPORT_SYMBOL_GPL(eeh_add_device_early);
 
+void eeh_add_device_tree_early(struct device_node *dn)
+{
+	struct device_node *sib;
+	for (sib = dn->child; sib; sib = sib->sibling)
+		eeh_add_device_tree_early(sib);
+	eeh_add_device_early(dn);
+}
+EXPORT_SYMBOL_GPL(eeh_add_device_tree_early);
+
 /**
  * eeh_add_device_late - perform EEH initialization for the indicated pci device
  * @dev: pci device for which to set up EEH
@@ -1118,7 +918,6 @@ void eeh_add_device_late(struct pci_dev *dev)
 	pdn->pcidev = dev;
 
 	pci_addr_cache_insert_device (dev);
-	eeh_save_bars(dev, pdn);
 }
 EXPORT_SYMBOL_GPL(eeh_add_device_late);
 
@@ -1147,27 +946,28 @@ void eeh_remove_device(struct pci_dev *dev)
 }
 EXPORT_SYMBOL_GPL(eeh_remove_device);
 
-static int proc_eeh_show(struct seq_file *m, void *v)
-{
-	unsigned int cpu;
-	unsigned long ffs = 0, positives = 0, failures = 0;
-	unsigned long resets = 0;
-	unsigned long no_dev = 0, no_dn = 0, no_cfg = 0, no_check = 0;
-
-	for_each_cpu(cpu) {
-		ffs += per_cpu(total_mmio_ffs, cpu);
-		positives += per_cpu(false_positives, cpu);
-		failures += per_cpu(ignored_failures, cpu);
-		resets += per_cpu(slot_resets, cpu);
-		no_dev += per_cpu(no_device, cpu);
-		no_dn += per_cpu(no_dn, cpu);
-		no_cfg += per_cpu(no_cfg_addr, cpu);
-		no_check += per_cpu(ignored_check, cpu);
+void eeh_remove_bus_device(struct pci_dev *dev)
+{
+	eeh_remove_device(dev);
+	if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE) {
+		struct pci_bus *bus = dev->subordinate;
+		struct list_head *ln;
+		if (!bus)
+			return; 
+		for (ln = bus->devices.next; ln != &bus->devices; ln = ln->next) {
+			struct pci_dev *pdev = pci_dev_b(ln);
+			if (pdev)
+				eeh_remove_bus_device(pdev);
+		}
 	}
+}
+EXPORT_SYMBOL_GPL(eeh_remove_bus_device);
 
+static int proc_eeh_show(struct seq_file *m, void *v)
+{
 	if (0 == eeh_subsystem_enabled) {
 		seq_printf(m, "EEH Subsystem is globally disabled\n");
-		seq_printf(m, "eeh_total_mmio_ffs=%ld\n", ffs);
+		seq_printf(m, "eeh_total_mmio_ffs=%ld\n", total_mmio_ffs);
 	} else {
 		seq_printf(m, "EEH Subsystem is enabled\n");
 		seq_printf(m,
@@ -1179,8 +979,10 @@ static int proc_eeh_show(struct seq_file *m, void *v)
 				"eeh_false_positives=%ld\n"
 				"eeh_ignored_failures=%ld\n"
 				"eeh_slot_resets=%ld\n",
-				no_dev, no_dn, no_cfg, no_check,
-				ffs, positives, failures, resets);
+				no_device, no_dn, no_cfg_addr, 
+				ignored_check, total_mmio_ffs, 
+				false_positives, ignored_failures, 
+				slot_resets);
 	}
 
 	return 0;
diff --git a/arch/powerpc/platforms/pseries/eeh_cache.c b/arch/powerpc/platforms/pseries/eeh_cache.c
new file mode 100644
index 00000000000..d4a402c5866
--- /dev/null
+++ b/arch/powerpc/platforms/pseries/eeh_cache.c
@@ -0,0 +1,316 @@
+/*
+ * eeh_cache.c
+ * PCI address cache; allows the lookup of PCI devices based on I/O address
+ *
+ * Copyright (C) 2004 Linas Vepstas <linas@austin.ibm.com> IBM Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+#include <linux/list.h>
+#include <linux/pci.h>
+#include <linux/rbtree.h>
+#include <linux/spinlock.h>
+#include <asm/atomic.h>
+#include <asm/pci-bridge.h>
+#include <asm/ppc-pci.h>
+
+#undef DEBUG
+
+/**
+ * The pci address cache subsystem.  This subsystem places
+ * PCI device address resources into a red-black tree, sorted
+ * according to the address range, so that given only an i/o
+ * address, the corresponding PCI device can be **quickly**
+ * found. It is safe to perform an address lookup in an interrupt
+ * context; this ability is an important feature.
+ *
+ * Currently, the only customer of this code is the EEH subsystem;
+ * thus, this code has been somewhat tailored to suit EEH better.
+ * In particular, the cache does *not* hold the addresses of devices
+ * for which EEH is not enabled.
+ *
+ * (Implementation Note: The RB tree seems to be better/faster
+ * than any hash algo I could think of for this problem, even
+ * with the penalty of slow pointer chases for d-cache misses).
+ */
+struct pci_io_addr_range
+{
+	struct rb_node rb_node;
+	unsigned long addr_lo;
+	unsigned long addr_hi;
+	struct pci_dev *pcidev;
+	unsigned int flags;
+};
+
+static struct pci_io_addr_cache
+{
+	struct rb_root rb_root;
+	spinlock_t piar_lock;
+} pci_io_addr_cache_root;
+
+static inline struct pci_dev *__pci_get_device_by_addr(unsigned long addr)
+{
+	struct rb_node *n = pci_io_addr_cache_root.rb_root.rb_node;
+
+	while (n) {
+		struct pci_io_addr_range *piar;
+		piar = rb_entry(n, struct pci_io_addr_range, rb_node);
+
+		if (addr < piar->addr_lo) {
+			n = n->rb_left;
+		} else {
+			if (addr > piar->addr_hi) {
+				n = n->rb_right;
+			} else {
+				pci_dev_get(piar->pcidev);
+				return piar->pcidev;
+			}
+		}
+	}
+
+	return NULL;
+}
+
+/**
+ * pci_get_device_by_addr - Get device, given only address
+ * @addr: mmio (PIO) phys address or i/o port number
+ *
+ * Given an mmio phys address, or a port number, find a pci device
+ * that implements this address.  Be sure to pci_dev_put the device
+ * when finished.  I/O port numbers are assumed to be offset
+ * from zero (that is, they do *not* have pci_io_addr added in).
+ * It is safe to call this function within an interrupt.
+ */
+struct pci_dev *pci_get_device_by_addr(unsigned long addr)
+{
+	struct pci_dev *dev;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pci_io_addr_cache_root.piar_lock, flags);
+	dev = __pci_get_device_by_addr(addr);
+	spin_unlock_irqrestore(&pci_io_addr_cache_root.piar_lock, flags);
+	return dev;
+}
+
+#ifdef DEBUG
+/*
+ * Handy-dandy debug print routine, does nothing more
+ * than print out the contents of our addr cache.
+ */
+static void pci_addr_cache_print(struct pci_io_addr_cache *cache)
+{
+	struct rb_node *n;
+	int cnt = 0;
+
+	n = rb_first(&cache->rb_root);
+	while (n) {
+		struct pci_io_addr_range *piar;
+		piar = rb_entry(n, struct pci_io_addr_range, rb_node);
+		printk(KERN_DEBUG "PCI: %s addr range %d [%lx-%lx]: %s\n",
+		       (piar->flags & IORESOURCE_IO) ? "i/o" : "mem", cnt,
+		       piar->addr_lo, piar->addr_hi, pci_name(piar->pcidev));
+		cnt++;
+		n = rb_next(n);
+	}
+}
+#endif
+
+/* Insert address range into the rb tree. */
+static struct pci_io_addr_range *
+pci_addr_cache_insert(struct pci_dev *dev, unsigned long alo,
+		      unsigned long ahi, unsigned int flags)
+{
+	struct rb_node **p = &pci_io_addr_cache_root.rb_root.rb_node;
+	struct rb_node *parent = NULL;
+	struct pci_io_addr_range *piar;
+
+	/* Walk tree, find a place to insert into tree */
+	while (*p) {
+		parent = *p;
+		piar = rb_entry(parent, struct pci_io_addr_range, rb_node);
+		if (ahi < piar->addr_lo) {
+			p = &parent->rb_left;
+		} else if (alo > piar->addr_hi) {
+			p = &parent->rb_right;
+		} else {
+			if (dev != piar->pcidev ||
+			    alo != piar->addr_lo || ahi != piar->addr_hi) {
+				printk(KERN_WARNING "PIAR: overlapping address range\n");
+			}
+			return piar;
+		}
+	}
+	piar = (struct pci_io_addr_range *)kmalloc(sizeof(struct pci_io_addr_range), GFP_ATOMIC);
+	if (!piar)
+		return NULL;
+
+	piar->addr_lo = alo;
+	piar->addr_hi = ahi;
+	piar->pcidev = dev;
+	piar->flags = flags;
+
+#ifdef DEBUG
+	printk(KERN_DEBUG "PIAR: insert range=[%lx:%lx] dev=%s\n",
+	                  alo, ahi, pci_name (dev));
+#endif
+
+	rb_link_node(&piar->rb_node, parent, p);
+	rb_insert_color(&piar->rb_node, &pci_io_addr_cache_root.rb_root);
+
+	return piar;
+}
+
+static void __pci_addr_cache_insert_device(struct pci_dev *dev)
+{
+	struct device_node *dn;
+	struct pci_dn *pdn;
+	int i;
+	int inserted = 0;
+
+	dn = pci_device_to_OF_node(dev);
+	if (!dn) {
+		printk(KERN_WARNING "PCI: no pci dn found for dev=%s\n", pci_name(dev));
+		return;
+	}
+
+	/* Skip any devices for which EEH is not enabled. */
+	pdn = PCI_DN(dn);
+	if (!(pdn->eeh_mode & EEH_MODE_SUPPORTED) ||
+	    pdn->eeh_mode & EEH_MODE_NOCHECK) {
+#ifdef DEBUG
+		printk(KERN_INFO "PCI: skip building address cache for=%s - %s\n",
+		       pci_name(dev), pdn->node->full_name);
+#endif
+		return;
+	}
+
+	/* The cache holds a reference to the device... */
+	pci_dev_get(dev);
+
+	/* Walk resources on this device, poke them into the tree */
+	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
+		unsigned long start = pci_resource_start(dev,i);
+		unsigned long end = pci_resource_end(dev,i);
+		unsigned int flags = pci_resource_flags(dev,i);
+
+		/* We are interested only bus addresses, not dma or other stuff */
+		if (0 == (flags & (IORESOURCE_IO | IORESOURCE_MEM)))
+			continue;
+		if (start == 0 || ~start == 0 || end == 0 || ~end == 0)
+			 continue;
+		pci_addr_cache_insert(dev, start, end, flags);
+		inserted = 1;
+	}
+
+	/* If there was nothing to add, the cache has no reference... */
+	if (!inserted)
+		pci_dev_put(dev);
+}
+
+/**
+ * pci_addr_cache_insert_device - Add a device to the address cache
+ * @dev: PCI device whose I/O addresses we are interested in.
+ *
+ * In order to support the fast lookup of devices based on addresses,
+ * we maintain a cache of devices that can be quickly searched.
+ * This routine adds a device to that cache.
+ */
+void pci_addr_cache_insert_device(struct pci_dev *dev)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&pci_io_addr_cache_root.piar_lock, flags);
+	__pci_addr_cache_insert_device(dev);
+	spin_unlock_irqrestore(&pci_io_addr_cache_root.piar_lock, flags);
+}
+
+static inline void __pci_addr_cache_remove_device(struct pci_dev *dev)
+{
+	struct rb_node *n;
+	int removed = 0;
+
+restart:
+	n = rb_first(&pci_io_addr_cache_root.rb_root);
+	while (n) {
+		struct pci_io_addr_range *piar;
+		piar = rb_entry(n, struct pci_io_addr_range, rb_node);
+
+		if (piar->pcidev == dev) {
+			rb_erase(n, &pci_io_addr_cache_root.rb_root);
+			removed = 1;
+			kfree(piar);
+			goto restart;
+		}
+		n = rb_next(n);
+	}
+
+	/* The cache no longer holds its reference to this device... */
+	if (removed)
+		pci_dev_put(dev);
+}
+
+/**
+ * pci_addr_cache_remove_device - remove pci device from addr cache
+ * @dev: device to remove
+ *
+ * Remove a device from the addr-cache tree.
+ * This is potentially expensive, since it will walk
+ * the tree multiple times (once per resource).
+ * But so what; device removal doesn't need to be that fast.
+ */
+void pci_addr_cache_remove_device(struct pci_dev *dev)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&pci_io_addr_cache_root.piar_lock, flags);
+	__pci_addr_cache_remove_device(dev);
+	spin_unlock_irqrestore(&pci_io_addr_cache_root.piar_lock, flags);
+}
+
+/**
+ * pci_addr_cache_build - Build a cache of I/O addresses
+ *
+ * Build a cache of pci i/o addresses.  This cache will be used to
+ * find the pci device that corresponds to a given address.
+ * This routine scans all pci busses to build the cache.
+ * Must be run late in boot process, after the pci controllers
+ * have been scaned for devices (after all device resources are known).
+ */
+void __init pci_addr_cache_build(void)
+{
+	struct device_node *dn;
+	struct pci_dev *dev = NULL;
+
+	spin_lock_init(&pci_io_addr_cache_root.piar_lock);
+
+	while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) {
+		/* Ignore PCI bridges */
+		if ((dev->class >> 16) == PCI_BASE_CLASS_BRIDGE)
+			continue;
+
+		pci_addr_cache_insert_device(dev);
+
+		dn = pci_device_to_OF_node(dev);
+		pci_dev_get (dev);  /* matching put is in eeh_remove_device() */
+		PCI_DN(dn)->pcidev = dev;
+	}
+
+#ifdef DEBUG
+	/* Verify tree built up above, echo back the list of addrs. */
+	pci_addr_cache_print(&pci_io_addr_cache_root);
+#endif
+}
+
diff --git a/arch/powerpc/platforms/pseries/eeh_driver.c b/arch/powerpc/platforms/pseries/eeh_driver.c
new file mode 100644
index 00000000000..6373372932b
--- /dev/null
+++ b/arch/powerpc/platforms/pseries/eeh_driver.c
@@ -0,0 +1,376 @@
+/*
+ * PCI Error Recovery Driver for RPA-compliant PPC64 platform.
+ * Copyright (C) 2004, 2005 Linas Vepstas <linas@linas.org>
+ *
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or (at
+ * your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT.  See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ * Send feedback to <linas@us.ibm.com>
+ *
+ */
+#include <linux/delay.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/notifier.h>
+#include <linux/pci.h>
+#include <asm/eeh.h>
+#include <asm/eeh_event.h>
+#include <asm/ppc-pci.h>
+#include <asm/pci-bridge.h>
+#include <asm/prom.h>
+#include <asm/rtas.h>
+
+
+static inline const char * pcid_name (struct pci_dev *pdev)
+{
+	if (pdev->dev.driver)
+		return pdev->dev.driver->name;
+	return "";
+}
+
+#ifdef DEBUG
+static void print_device_node_tree (struct pci_dn *pdn, int dent)
+{
+	int i;
+	if (!pdn) return;
+	for (i=0;i<dent; i++)
+		printk(" ");
+	printk("dn=%s mode=%x \tcfg_addr=%x pe_addr=%x \tfull=%s\n",
+		pdn->node->name, pdn->eeh_mode, pdn->eeh_config_addr,
+		pdn->eeh_pe_config_addr, pdn->node->full_name);
+	dent += 3;
+	struct device_node *pc = pdn->node->child;
+	while (pc) {
+		print_device_node_tree(PCI_DN(pc), dent);
+		pc = pc->sibling;
+	}
+}
+#endif
+
+/** 
+ * irq_in_use - return true if this irq is being used 
+ */
+static int irq_in_use(unsigned int irq)
+{
+	int rc = 0;
+	unsigned long flags;
+   struct irq_desc *desc = irq_desc + irq;
+
+	spin_lock_irqsave(&desc->lock, flags);
+	if (desc->action)
+		rc = 1;
+	spin_unlock_irqrestore(&desc->lock, flags);
+	return rc;
+}
+
+/* ------------------------------------------------------- */
+/** eeh_report_error - report an EEH error to each device,
+ *  collect up and merge the device responses.
+ */
+
+static void eeh_report_error(struct pci_dev *dev, void *userdata)
+{
+	enum pci_ers_result rc, *res = userdata;
+	struct pci_driver *driver = dev->driver;
+
+	dev->error_state = pci_channel_io_frozen;
+
+	if (!driver)
+		return;
+
+	if (irq_in_use (dev->irq)) {
+		struct device_node *dn = pci_device_to_OF_node(dev);
+		PCI_DN(dn)->eeh_mode |= EEH_MODE_IRQ_DISABLED;
+		disable_irq_nosync(dev->irq);
+	}
+	if (!driver->err_handler)
+		return;
+	if (!driver->err_handler->error_detected)
+		return;
+
+	rc = driver->err_handler->error_detected (dev, pci_channel_io_frozen);
+	if (*res == PCI_ERS_RESULT_NONE) *res = rc;
+	if (*res == PCI_ERS_RESULT_NEED_RESET) return;
+	if (*res == PCI_ERS_RESULT_DISCONNECT &&
+	     rc == PCI_ERS_RESULT_NEED_RESET) *res = rc;
+}
+
+/** eeh_report_reset -- tell this device that the pci slot
+ *  has been reset.
+ */
+
+static void eeh_report_reset(struct pci_dev *dev, void *userdata)
+{
+	struct pci_driver *driver = dev->driver;
+	struct device_node *dn = pci_device_to_OF_node(dev);
+
+	if (!driver)
+		return;
+
+	if ((PCI_DN(dn)->eeh_mode) & EEH_MODE_IRQ_DISABLED) {
+		PCI_DN(dn)->eeh_mode &= ~EEH_MODE_IRQ_DISABLED;
+		enable_irq(dev->irq);
+	}
+	if (!driver->err_handler)
+		return;
+	if (!driver->err_handler->slot_reset)
+		return;
+
+	driver->err_handler->slot_reset(dev);
+}
+
+static void eeh_report_resume(struct pci_dev *dev, void *userdata)
+{
+	struct pci_driver *driver = dev->driver;
+
+	dev->error_state = pci_channel_io_normal;
+
+	if (!driver)
+		return;
+	if (!driver->err_handler)
+		return;
+	if (!driver->err_handler->resume)
+		return;
+
+	driver->err_handler->resume(dev);
+}
+
+static void eeh_report_failure(struct pci_dev *dev, void *userdata)
+{
+	struct pci_driver *driver = dev->driver;
+
+	dev->error_state = pci_channel_io_perm_failure;
+
+	if (!driver)
+		return;
+
+	if (irq_in_use (dev->irq)) {
+		struct device_node *dn = pci_device_to_OF_node(dev);
+		PCI_DN(dn)->eeh_mode |= EEH_MODE_IRQ_DISABLED;
+		disable_irq_nosync(dev->irq);
+	}
+	if (!driver->err_handler)
+		return;
+	if (!driver->err_handler->error_detected)
+		return;
+	driver->err_handler->error_detected(dev, pci_channel_io_perm_failure);
+}
+
+/* ------------------------------------------------------- */
+/**
+ * handle_eeh_events -- reset a PCI device after hard lockup.
+ *
+ * pSeries systems will isolate a PCI slot if the PCI-Host
+ * bridge detects address or data parity errors, DMA's
+ * occuring to wild addresses (which usually happen due to
+ * bugs in device drivers or in PCI adapter firmware).
+ * Slot isolations also occur if #SERR, #PERR or other misc
+ * PCI-related errors are detected.
+ *
+ * Recovery process consists of unplugging the device driver
+ * (which generated hotplug events to userspace), then issuing
+ * a PCI #RST to the device, then reconfiguring the PCI config
+ * space for all bridges & devices under this slot, and then
+ * finally restarting the device drivers (which cause a second
+ * set of hotplug events to go out to userspace).
+ */
+
+/**
+ * eeh_reset_device() -- perform actual reset of a pci slot
+ * Args: bus: pointer to the pci bus structure corresponding
+ *            to the isolated slot. A non-null value will
+ *            cause all devices under the bus to be removed
+ *            and then re-added.
+ *     pe_dn: pointer to a "Partionable Endpoint" device node.
+ *            This is the top-level structure on which pci
+ *            bus resets can be performed.
+ */
+
+static int eeh_reset_device (struct pci_dn *pe_dn, struct pci_bus *bus)
+{
+	int rc;
+	if (bus)
+		pcibios_remove_pci_devices(bus);
+
+	/* Reset the pci controller. (Asserts RST#; resets config space).
+	 * Reconfigure bridges and devices. Don't try to bring the system
+	 * up if the reset failed for some reason. */
+	rc = rtas_set_slot_reset(pe_dn);
+	if (rc)
+		return rc;
+
+ 	/* New-style config addrs might be shared across multiple devices,
+ 	 * Walk over all functions on this device */
+ 	if (pe_dn->eeh_pe_config_addr) {
+ 		struct device_node *pe = pe_dn->node;
+ 		pe = pe->parent->child;
+ 		while (pe) {
+ 			struct pci_dn *ppe = PCI_DN(pe);
+ 			if (pe_dn->eeh_pe_config_addr == ppe->eeh_pe_config_addr) {
+ 				rtas_configure_bridge(ppe);
+ 				eeh_restore_bars(ppe);
+ 			}
+ 			pe = pe->sibling;
+ 		}
+ 	} else {
+ 		rtas_configure_bridge(pe_dn);
+ 		eeh_restore_bars(pe_dn);
+ 	}
+
+	/* Give the system 5 seconds to finish running the user-space
+	 * hotplug shutdown scripts, e.g. ifdown for ethernet.  Yes, 
+	 * this is a hack, but if we don't do this, and try to bring 
+	 * the device up before the scripts have taken it down, 
+	 * potentially weird things happen.
+	 */
+	if (bus) {
+		ssleep (5);
+		pcibios_add_pci_devices(bus);
+	}
+
+	return 0;
+}
+
+/* The longest amount of time to wait for a pci device
+ * to come back on line, in seconds.
+ */
+#define MAX_WAIT_FOR_RECOVERY 15
+
+void handle_eeh_events (struct eeh_event *event)
+{
+	struct device_node *frozen_dn;
+	struct pci_dn *frozen_pdn;
+	struct pci_bus *frozen_bus;
+	int rc = 0;
+	enum pci_ers_result result = PCI_ERS_RESULT_NONE;
+
+	frozen_dn = find_device_pe(event->dn);
+	frozen_bus = pcibios_find_pci_bus(frozen_dn);
+
+	if (!frozen_dn) {
+		printk(KERN_ERR "EEH: Error: Cannot find partition endpoint for %s\n",
+		        pci_name(event->dev));
+		return;
+	}
+
+	/* There are two different styles for coming up with the PE.
+	 * In the old style, it was the highest EEH-capable device
+	 * which was always an EADS pci bridge.  In the new style,
+	 * there might not be any EADS bridges, and even when there are,
+	 * the firmware marks them as "EEH incapable". So another
+	 * two-step is needed to find the pci bus.. */
+	if (!frozen_bus)
+		frozen_bus = pcibios_find_pci_bus (frozen_dn->parent);
+
+	if (!frozen_bus) {
+		printk(KERN_ERR "EEH: Cannot find PCI bus for %s\n",
+		        frozen_dn->full_name);
+		return;
+	}
+
+#if 0
+	/* We may get "permanent failure" messages on empty slots.
+	 * These are false alarms. Empty slots have no child dn. */
+	if ((event->state == pci_channel_io_perm_failure) && (frozen_device == NULL))
+		return;
+#endif
+
+	frozen_pdn = PCI_DN(frozen_dn);
+	frozen_pdn->eeh_freeze_count++;
+	
+	if (frozen_pdn->eeh_freeze_count > EEH_MAX_ALLOWED_FREEZES)
+		goto hard_fail;
+
+	/* If the reset state is a '5' and the time to reset is 0 (infinity)
+	 * or is more then 15 seconds, then mark this as a permanent failure.
+	 */
+	if ((event->state == pci_channel_io_perm_failure) &&
+	    ((event->time_unavail <= 0) ||
+	     (event->time_unavail > MAX_WAIT_FOR_RECOVERY*1000)))
+		goto hard_fail;
+
+	eeh_slot_error_detail(frozen_pdn, 1 /* Temporary Error */);
+	printk(KERN_WARNING
+	   "EEH: This PCI device has failed %d times since last reboot: %s - %s\n",
+		frozen_pdn->eeh_freeze_count,
+		pci_name (frozen_pdn->pcidev), 
+		pcid_name(frozen_pdn->pcidev));
+
+	/* Walk the various device drivers attached to this slot through
+	 * a reset sequence, giving each an opportunity to do what it needs
+	 * to accomplish the reset.  Each child gets a report of the
+	 * status ... if any child can't handle the reset, then the entire
+	 * slot is dlpar removed and added.
+	 */
+	pci_walk_bus(frozen_bus, eeh_report_error, &result);
+
+	/* If all device drivers were EEH-unaware, then shut
+	 * down all of the device drivers, and hope they
+	 * go down willingly, without panicing the system.
+	 */
+	if (result == PCI_ERS_RESULT_NONE) {
+		rc = eeh_reset_device(frozen_pdn, frozen_bus);
+		if (rc)
+			goto hard_fail;
+	}
+
+	/* If any device called out for a reset, then reset the slot */
+	if (result == PCI_ERS_RESULT_NEED_RESET) {
+		rc = eeh_reset_device(frozen_pdn, NULL);
+		if (rc)
+			goto hard_fail;
+		pci_walk_bus(frozen_bus, eeh_report_reset, 0);
+	}
+
+	/* If all devices reported they can proceed, the re-enable PIO */
+	if (result == PCI_ERS_RESULT_CAN_RECOVER) {
+		/* XXX Not supported; we brute-force reset the device */
+		rc = eeh_reset_device(frozen_pdn, NULL);
+		if (rc)
+			goto hard_fail;
+		pci_walk_bus(frozen_bus, eeh_report_reset, 0);
+	}
+
+	/* Tell all device drivers that they can resume operations */
+	pci_walk_bus(frozen_bus, eeh_report_resume, 0);
+
+	return;
+	
+hard_fail:
+	/*
+	 * About 90% of all real-life EEH failures in the field
+	 * are due to poorly seated PCI cards. Only 10% or so are
+	 * due to actual, failed cards.
+	 */
+	printk(KERN_ERR
+	   "EEH: PCI device %s - %s has failed %d times \n"
+	   "and has been permanently disabled.  Please try reseating\n"
+	   "this device or replacing it.\n",
+		pci_name (frozen_pdn->pcidev), 
+		pcid_name(frozen_pdn->pcidev), 
+		frozen_pdn->eeh_freeze_count);
+
+	eeh_slot_error_detail(frozen_pdn, 2 /* Permanent Error */);
+
+	/* Notify all devices that they're about to go down. */
+	pci_walk_bus(frozen_bus, eeh_report_failure, 0);
+
+	/* Shut down the device drivers for good. */
+	pcibios_remove_pci_devices(frozen_bus);
+}
+
+/* ---------- end of file ---------- */
diff --git a/arch/powerpc/platforms/pseries/eeh_event.c b/arch/powerpc/platforms/pseries/eeh_event.c
index 92497333c2b..9a9961f2748 100644
--- a/arch/powerpc/platforms/pseries/eeh_event.c
+++ b/arch/powerpc/platforms/pseries/eeh_event.c
@@ -21,6 +21,7 @@
 #include <linux/list.h>
 #include <linux/pci.h>
 #include <asm/eeh_event.h>
+#include <asm/ppc-pci.h>
 
 /** Overview:
  *  EEH error states may be detected within exception handlers;
@@ -37,31 +38,6 @@ static void eeh_thread_launcher(void *);
 DECLARE_WORK(eeh_event_wq, eeh_thread_launcher, NULL);
 
 /**
- * eeh_panic - call panic() for an eeh event that cannot be handled.
- * The philosophy of this routine is that it is better to panic and
- * halt the OS than it is to risk possible data corruption by
- * oblivious device drivers that don't know better.
- *
- * @dev pci device that had an eeh event
- * @reset_state current reset state of the device slot
- */
-static void eeh_panic(struct pci_dev *dev, int reset_state)
-{
-	/*
-	 * Since the panic_on_oops sysctl is used to halt the system
-	 * in light of potential corruption, we can use it here.
-	 */
-	if (panic_on_oops) {
-		panic("EEH: MMIO failure (%d) on device:%s\n", reset_state,
-		      pci_name(dev));
-	}
-	else {
-		printk(KERN_INFO "EEH: Ignored MMIO failure (%d) on device:%s\n",
-		       reset_state, pci_name(dev));
-	}
-}
-
-/**
  * eeh_event_handler - dispatch EEH events.  The detection of a frozen
  * slot can occur inside an interrupt, where it can be hard to do
  * anything about it.  The goal of this routine is to pull these
@@ -82,10 +58,16 @@ static int eeh_event_handler(void * dummy)
 
 		spin_lock_irqsave(&eeh_eventlist_lock, flags);
 		event = NULL;
+
+		/* Unqueue the event, get ready to process. */
 		if (!list_empty(&eeh_eventlist)) {
 			event = list_entry(eeh_eventlist.next, struct eeh_event, list);
 			list_del(&event->list);
 		}
+		
+		if (event)
+			eeh_mark_slot(event->dn, EEH_MODE_RECOVERING);
+
 		spin_unlock_irqrestore(&eeh_eventlist_lock, flags);
 		if (event == NULL)
 			break;
@@ -93,8 +75,11 @@ static int eeh_event_handler(void * dummy)
 		printk(KERN_INFO "EEH: Detected PCI bus error on device %s\n",
 		       pci_name(event->dev));
 
-		eeh_panic (event->dev, event->state);
+		handle_eeh_events(event);
+
+		eeh_clear_slot(event->dn, EEH_MODE_RECOVERING);
 
+		pci_dev_put(event->dev);
 		kfree(event);
 	}
 
@@ -122,7 +107,7 @@ static void eeh_thread_launcher(void *dummy)
  */
 int eeh_send_failure_event (struct device_node *dn,
                             struct pci_dev *dev,
-                            int state,
+                            enum pci_channel_state state,
                             int time_unavail)
 {
 	unsigned long flags;
diff --git a/arch/powerpc/platforms/pseries/hvcserver.c b/arch/powerpc/platforms/pseries/hvcserver.c
index 4d584172055..22bfb5c89db 100644
--- a/arch/powerpc/platforms/pseries/hvcserver.c
+++ b/arch/powerpc/platforms/pseries/hvcserver.c
@@ -40,7 +40,7 @@ MODULE_VERSION(HVCS_ARCH_VERSION);
  * functions aren't performance sensitive, so this conversion isn't an
  * issue.
  */
-int hvcs_convert(long to_convert)
+static int hvcs_convert(long to_convert)
 {
 	switch (to_convert) {
 		case H_Success:
@@ -91,7 +91,7 @@ int hvcs_free_partner_info(struct list_head *head)
 EXPORT_SYMBOL(hvcs_free_partner_info);
 
 /* Helper function for hvcs_get_partner_info */
-int hvcs_next_partner(uint32_t unit_address,
+static int hvcs_next_partner(uint32_t unit_address,
 		unsigned long last_p_partition_ID,
 		unsigned long last_p_unit_address, unsigned long *pi_buff)
 
diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c
index 2043659ea7b..48cfbfc43f9 100644
--- a/arch/powerpc/platforms/pseries/iommu.c
+++ b/arch/powerpc/platforms/pseries/iommu.c
@@ -51,8 +51,6 @@
 
 #define DBG(fmt...)
 
-extern int is_python(struct device_node *);
-
 static void tce_build_pSeries(struct iommu_table *tbl, long index, 
 			      long npages, unsigned long uaddr, 
 			      enum dma_data_direction direction)
@@ -436,7 +434,7 @@ static void iommu_bus_setup_pSeriesLP(struct pci_bus *bus)
 		return;
 	}
 
-	ppci = pdn->data;
+	ppci = PCI_DN(pdn);
 	if (!ppci->iommu_table) {
 		/* Bussubno hasn't been copied yet.
 		 * Do it now because iommu_table_setparms_lpar needs it.
@@ -483,10 +481,10 @@ static void iommu_dev_setup_pSeries(struct pci_dev *dev)
 	 * an already allocated iommu table is found and use that.
 	 */
 
-	while (dn && dn->data && PCI_DN(dn)->iommu_table == NULL)
+	while (dn && PCI_DN(dn) && PCI_DN(dn)->iommu_table == NULL)
 		dn = dn->parent;
 
-	if (dn && dn->data) {
+	if (dn && PCI_DN(dn)) {
 		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, pci_name(dev));
@@ -497,7 +495,7 @@ static int iommu_reconfig_notifier(struct notifier_block *nb, unsigned long acti
 {
 	int err = NOTIFY_OK;
 	struct device_node *np = node;
-	struct pci_dn *pci = np->data;
+	struct pci_dn *pci = PCI_DN(np);
 
 	switch (action) {
 	case PSERIES_RECONFIG_REMOVE:
@@ -533,7 +531,7 @@ static void iommu_dev_setup_pSeriesLP(struct pci_dev *dev)
 	 */
 	dn = pci_device_to_OF_node(dev);
 
-	for (pdn = dn; pdn && pdn->data && !PCI_DN(pdn)->iommu_table;
+	for (pdn = dn; pdn && PCI_DN(pdn) && !PCI_DN(pdn)->iommu_table;
 	     pdn = pdn->parent) {
 		dma_window = (unsigned int *)
 			get_property(pdn, "ibm,dma-window", NULL);
@@ -552,7 +550,7 @@ static void iommu_dev_setup_pSeriesLP(struct pci_dev *dev)
 		DBG("Found DMA window, allocating table\n");
 	}
 
-	pci = pdn->data;
+	pci = PCI_DN(pdn);
 	if (!pci->iommu_table) {
 		/* iommu_table_setparms_lpar needs bussubno. */
 		pci->bussubno = pci->phb->bus->number;
diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c
index cf1bc11b334..8952528d31a 100644
--- a/arch/powerpc/platforms/pseries/lpar.c
+++ b/arch/powerpc/platforms/pseries/lpar.c
@@ -24,6 +24,7 @@
 #include <linux/config.h>
 #include <linux/kernel.h>
 #include <linux/dma-mapping.h>
+#include <linux/console.h>
 #include <asm/processor.h>
 #include <asm/mmu.h>
 #include <asm/page.h>
@@ -60,7 +61,7 @@ extern void pSeries_find_serial_port(void);
 int vtermno;	/* virtual terminal# for udbg  */
 
 #define __ALIGNED__ __attribute__((__aligned__(sizeof(long))))
-static void udbg_hvsi_putc(unsigned char c)
+static void udbg_hvsi_putc(char c)
 {
 	/* packet's seqno isn't used anyways */
 	uint8_t packet[] __ALIGNED__ = { 0xff, 5, 0, 0, c };
@@ -111,7 +112,7 @@ static int udbg_hvsi_getc_poll(void)
 	return ch;
 }
 
-static unsigned char udbg_hvsi_getc(void)
+static int udbg_hvsi_getc(void)
 {
 	int ch;
 	for (;;) {
@@ -127,7 +128,7 @@ static unsigned char udbg_hvsi_getc(void)
 	}
 }
 
-static void udbg_putcLP(unsigned char c)
+static void udbg_putcLP(char c)
 {
 	char buf[16];
 	unsigned long rc;
@@ -172,7 +173,7 @@ static int udbg_getc_pollLP(void)
 	return ch;
 }
 
-static unsigned char udbg_getcLP(void)
+static int udbg_getcLP(void)
 {
 	int ch;
 	for (;;) {
@@ -191,7 +192,7 @@ static unsigned char udbg_getcLP(void)
 /* call this from early_init() for a working debug console on
  * vterm capable LPAR machines
  */
-void udbg_init_debug_lpar(void)
+void __init udbg_init_debug_lpar(void)
 {
 	vtermno = 0;
 	udbg_putc = udbg_putcLP;
@@ -200,73 +201,64 @@ void udbg_init_debug_lpar(void)
 }
 
 /* returns 0 if couldn't find or use /chosen/stdout as console */
-int find_udbg_vterm(void)
+void __init find_udbg_vterm(void)
 {
 	struct device_node *stdout_node;
 	u32 *termno;
 	char *name;
-	int found = 0;
+	int add_console;
 
 	/* find the boot console from /chosen/stdout */
 	if (!of_chosen)
-		return 0;
+		return;
 	name = (char *)get_property(of_chosen, "linux,stdout-path", NULL);
 	if (name == NULL)
-		return 0;
+		return;
 	stdout_node = of_find_node_by_path(name);
 	if (!stdout_node)
-		return 0;
-
-	/* now we have the stdout node; figure out what type of device it is. */
+		return;
 	name = (char *)get_property(stdout_node, "name", NULL);
 	if (!name) {
 		printk(KERN_WARNING "stdout node missing 'name' property!\n");
 		goto out;
 	}
+	/* The user has requested a console so this is already set up. */
+	add_console = !strstr(cmd_line, "console=");
 
-	if (strncmp(name, "vty", 3) == 0) {
-		if (device_is_compatible(stdout_node, "hvterm1")) {
-			termno = (u32 *)get_property(stdout_node, "reg", NULL);
-			if (termno) {
-				vtermno = termno[0];
-				udbg_putc = udbg_putcLP;
-				udbg_getc = udbg_getcLP;
-				udbg_getc_poll = udbg_getc_pollLP;
-				found = 1;
-			}
-		} else if (device_is_compatible(stdout_node, "hvterm-protocol")) {
-			termno = (u32 *)get_property(stdout_node, "reg", NULL);
-			if (termno) {
-				vtermno = termno[0];
-				udbg_putc = udbg_hvsi_putc;
-				udbg_getc = udbg_hvsi_getc;
-				udbg_getc_poll = udbg_hvsi_getc_poll;
-				found = 1;
-			}
-		}
-	} else if (strncmp(name, "serial", 6)) {
-		/* XXX fix ISA serial console */
-		printk(KERN_WARNING "serial stdout on LPAR ('%s')! "
-				"can't print udbg messages\n",
-		       stdout_node->full_name);
-	} else {
-		printk(KERN_WARNING "don't know how to print to stdout '%s'\n",
-		       stdout_node->full_name);
+	/* Check if it's a virtual terminal */
+	if (strncmp(name, "vty", 3) != 0)
+		goto out;
+	termno = (u32 *)get_property(stdout_node, "reg", NULL);
+	if (termno == NULL)
+		goto out;
+	vtermno = termno[0];
+
+	if (device_is_compatible(stdout_node, "hvterm1")) {
+		udbg_putc = udbg_putcLP;
+		udbg_getc = udbg_getcLP;
+		udbg_getc_poll = udbg_getc_pollLP;
+		if (add_console)
+			add_preferred_console("hvc", termno[0] & 0xff, NULL);
+	} else if (device_is_compatible(stdout_node, "hvterm-protocol")) {
+		vtermno = termno[0];
+		udbg_putc = udbg_hvsi_putc;
+		udbg_getc = udbg_hvsi_getc;
+		udbg_getc_poll = udbg_hvsi_getc_poll;
+		if (add_console)
+			add_preferred_console("hvsi", termno[0] & 0xff, NULL);
 	}
-
 out:
 	of_node_put(stdout_node);
-	return found;
 }
 
 void vpa_init(int cpu)
 {
 	int hwcpu = get_hard_smp_processor_id(cpu);
-	unsigned long vpa = __pa(&paca[cpu].lppaca);
+	unsigned long vpa = __pa(&lppaca[cpu]);
 	long ret;
 
 	if (cpu_has_feature(CPU_FTR_ALTIVEC))
-		paca[cpu].lppaca.vmxregs_in_use = 1;
+		lppaca[cpu].vmxregs_in_use = 1;
 
 	ret = register_vpa(hwcpu, vpa);
 
diff --git a/arch/powerpc/platforms/pseries/pci_dlpar.c b/arch/powerpc/platforms/pseries/pci_dlpar.c
new file mode 100644
index 00000000000..21934784f93
--- /dev/null
+++ b/arch/powerpc/platforms/pseries/pci_dlpar.c
@@ -0,0 +1,174 @@
+/*
+ * PCI Dynamic LPAR, PCI Hot Plug and PCI EEH recovery code
+ * for RPA-compliant PPC64 platform.
+ * Copyright (C) 2003 Linda Xie <lxie@us.ibm.com>
+ * Copyright (C) 2005 International Business Machines
+ *
+ * Updates, 2005, John Rose <johnrose@austin.ibm.com>
+ * Updates, 2005, Linas Vepstas <linas@austin.ibm.com>
+ *
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or (at
+ * your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT.  See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <linux/pci.h>
+#include <asm/pci-bridge.h>
+
+static struct pci_bus *
+find_bus_among_children(struct pci_bus *bus,
+                        struct device_node *dn)
+{
+	struct pci_bus *child = NULL;
+	struct list_head *tmp;
+	struct device_node *busdn;
+
+	busdn = pci_bus_to_OF_node(bus);
+	if (busdn == dn)
+		return bus;
+
+	list_for_each(tmp, &bus->children) {
+		child = find_bus_among_children(pci_bus_b(tmp), dn);
+		if (child)
+			break;
+	};
+	return child;
+}
+
+struct pci_bus *
+pcibios_find_pci_bus(struct device_node *dn)
+{
+	struct pci_dn *pdn = dn->data;
+
+	if (!pdn  || !pdn->phb || !pdn->phb->bus)
+		return NULL;
+
+	return find_bus_among_children(pdn->phb->bus, dn);
+}
+
+/**
+ * pcibios_remove_pci_devices - remove all devices under this bus
+ *
+ * Remove all of the PCI devices under this bus both from the
+ * linux pci device tree, and from the powerpc EEH address cache.
+ */
+void
+pcibios_remove_pci_devices(struct pci_bus *bus)
+{
+	struct pci_dev *dev, *tmp;
+
+	list_for_each_entry_safe(dev, tmp, &bus->devices, bus_list) {
+		eeh_remove_bus_device(dev);
+		pci_remove_bus_device(dev);
+	}
+}
+
+/* Must be called before pci_bus_add_devices */
+void
+pcibios_fixup_new_pci_devices(struct pci_bus *bus, int fix_bus)
+{
+	struct pci_dev *dev;
+
+	list_for_each_entry(dev, &bus->devices, bus_list) {
+		/*
+		 * Skip already-present devices (which are on the
+		 * global device list.)
+		 */
+		if (list_empty(&dev->global_list)) {
+			int i;
+
+			/* Need to setup IOMMU tables */
+			ppc_md.iommu_dev_setup(dev);
+
+			if(fix_bus)
+				pcibios_fixup_device_resources(dev, bus);
+			pci_read_irq_line(dev);
+			for (i = 0; i < PCI_NUM_RESOURCES; i++) {
+				struct resource *r = &dev->resource[i];
+
+				if (r->parent || !r->start || !r->flags)
+					continue;
+				pci_claim_resource(dev, i);
+			}
+		}
+	}
+}
+
+static int
+pcibios_pci_config_bridge(struct pci_dev *dev)
+{
+	u8 sec_busno;
+	struct pci_bus *child_bus;
+	struct pci_dev *child_dev;
+
+	/* Get busno of downstream bus */
+	pci_read_config_byte(dev, PCI_SECONDARY_BUS, &sec_busno);
+
+	/* Add to children of PCI bridge dev->bus */
+	child_bus = pci_add_new_bus(dev->bus, dev, sec_busno);
+	if (!child_bus) {
+		printk (KERN_ERR "%s: could not add second bus\n", __FUNCTION__);
+		return -EIO;
+	}
+	sprintf(child_bus->name, "PCI Bus #%02x", child_bus->number);
+
+	pci_scan_child_bus(child_bus);
+
+	list_for_each_entry(child_dev, &child_bus->devices, bus_list) {
+		eeh_add_device_late(child_dev);
+	}
+
+	/* Fixup new pci devices without touching bus struct */
+	pcibios_fixup_new_pci_devices(child_bus, 0);
+
+	/* Make the discovered devices available */
+	pci_bus_add_devices(child_bus);
+	return 0;
+}
+
+/**
+ * pcibios_add_pci_devices - adds new pci devices to bus
+ *
+ * This routine will find and fixup new pci devices under
+ * the indicated bus. This routine presumes that there
+ * might already be some devices under this bridge, so
+ * it carefully tries to add only new devices.  (And that
+ * is how this routine differs from other, similar pcibios
+ * routines.)
+ */
+void
+pcibios_add_pci_devices(struct pci_bus * bus)
+{
+	int slotno, num;
+	struct pci_dev *dev;
+	struct device_node *dn = pci_bus_to_OF_node(bus);
+
+	eeh_add_device_tree_early(dn);
+
+	/* pci_scan_slot should find all children */
+	slotno = PCI_SLOT(PCI_DN(dn->child)->devfn);
+	num = pci_scan_slot(bus, PCI_DEVFN(slotno, 0));
+	if (num) {
+		pcibios_fixup_new_pci_devices(bus, 1);
+		pci_bus_add_devices(bus);
+	}
+
+	list_for_each_entry(dev, &bus->devices, bus_list) {
+		eeh_add_device_late (dev);
+		if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE)
+			pcibios_pci_config_bridge(dev);
+	}
+}
diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c
index fbd214d68b0..b046bcf7443 100644
--- a/arch/powerpc/platforms/pseries/ras.c
+++ b/arch/powerpc/platforms/pseries/ras.c
@@ -49,14 +49,14 @@
 #include <asm/machdep.h>
 #include <asm/rtas.h>
 #include <asm/udbg.h>
+#include <asm/firmware.h>
+
+#include "ras.h"
 
 static unsigned char ras_log_buf[RTAS_ERROR_LOG_MAX];
 static DEFINE_SPINLOCK(ras_log_buf_lock);
 
-char mce_data_buf[RTAS_ERROR_LOG_MAX]
-;
-/* This is true if we are using the firmware NMI handler (typically LPAR) */
-extern int fwnmi_active;
+char mce_data_buf[RTAS_ERROR_LOG_MAX];
 
 static int ras_get_sensor_state_token;
 static int ras_check_exception_token;
@@ -280,7 +280,7 @@ static void fwnmi_release_errinfo(void)
 		printk("FWNMI: nmi-interlock failed: %d\n", ret);
 }
 
-void pSeries_system_reset_exception(struct pt_regs *regs)
+int pSeries_system_reset_exception(struct pt_regs *regs)
 {
 	if (fwnmi_active) {
 		struct rtas_error_log *errhdr = fwnmi_get_errinfo(regs);
@@ -289,6 +289,7 @@ void pSeries_system_reset_exception(struct pt_regs *regs)
 		}
 		fwnmi_release_errinfo();
 	}
+	return 0; /* need to perform reset */
 }
 
 /*
diff --git a/arch/powerpc/platforms/pseries/ras.h b/arch/powerpc/platforms/pseries/ras.h
new file mode 100644
index 00000000000..0e66b0da55e
--- /dev/null
+++ b/arch/powerpc/platforms/pseries/ras.h
@@ -0,0 +1,9 @@
+#ifndef _PSERIES_RAS_H
+#define _PSERIES_RAS_H
+
+struct pt_regs;
+
+extern int pSeries_system_reset_exception(struct pt_regs *regs);
+extern int pSeries_machine_check_exception(struct pt_regs *regs);
+
+#endif /* _PSERIES_RAS_H */
diff --git a/arch/powerpc/platforms/pseries/reconfig.c b/arch/powerpc/platforms/pseries/reconfig.c
index d8864164dbe..86cfa6ecdcf 100644
--- a/arch/powerpc/platforms/pseries/reconfig.c
+++ b/arch/powerpc/platforms/pseries/reconfig.c
@@ -350,6 +350,100 @@ static int do_remove_node(char *buf)
 	return rv;
 }
 
+static char *parse_node(char *buf, size_t bufsize, struct device_node **npp)
+{
+	char *handle_str;
+	phandle handle;
+	*npp = NULL;
+
+	handle_str = buf;
+
+	buf = strchr(buf, ' ');
+	if (!buf)
+		return NULL;
+	*buf = '\0';
+	buf++;
+
+	handle = simple_strtoul(handle_str, NULL, 10);
+
+	*npp = of_find_node_by_phandle(handle);
+	return buf;
+}
+
+static int do_add_property(char *buf, size_t bufsize)
+{
+	struct property *prop = NULL;
+	struct device_node *np;
+	unsigned char *value;
+	char *name, *end;
+	int length;
+	end = buf + bufsize;
+	buf = parse_node(buf, bufsize, &np);
+
+	if (!np)
+		return -ENODEV;
+
+	if (parse_next_property(buf, end, &name, &length, &value) == NULL)
+		return -EINVAL;
+
+	prop = new_property(name, length, value, NULL);
+	if (!prop)
+		return -ENOMEM;
+
+	prom_add_property(np, prop);
+
+	return 0;
+}
+
+static int do_remove_property(char *buf, size_t bufsize)
+{
+	struct device_node *np;
+	char *tmp;
+	struct property *prop;
+	buf = parse_node(buf, bufsize, &np);
+
+	if (!np)
+		return -ENODEV;
+
+	tmp = strchr(buf,' ');
+	if (tmp)
+		*tmp = '\0';
+
+	if (strlen(buf) == 0)
+		return -EINVAL;
+
+	prop = of_find_property(np, buf, NULL);
+
+	return prom_remove_property(np, prop);
+}
+
+static int do_update_property(char *buf, size_t bufsize)
+{
+	struct device_node *np;
+	unsigned char *value;
+	char *name, *end;
+	int length;
+	struct property *newprop, *oldprop;
+	buf = parse_node(buf, bufsize, &np);
+	end = buf + bufsize;
+
+	if (!np)
+		return -ENODEV;
+
+	if (parse_next_property(buf, end, &name, &length, &value) == NULL)
+		return -EINVAL;
+
+	newprop = new_property(name, length, value, NULL);
+	if (!newprop)
+		return -ENOMEM;
+
+	oldprop = of_find_property(np, name,NULL);
+	if (!oldprop)
+		return -ENODEV;
+
+	return prom_update_property(np, newprop, oldprop);
+}
+
 /**
  * ofdt_write - perform operations on the Open Firmware device tree
  *
@@ -392,6 +486,12 @@ static ssize_t ofdt_write(struct file *file, const char __user *buf, size_t coun
 		rv = do_add_node(tmp, count - (tmp - kbuf));
 	else if (!strcmp(kbuf, "remove_node"))
 		rv = do_remove_node(tmp);
+	else if (!strcmp(kbuf, "add_property"))
+		rv = do_add_property(tmp, count - (tmp - kbuf));
+	else if (!strcmp(kbuf, "remove_property"))
+		rv = do_remove_property(tmp, count - (tmp - kbuf));
+	else if (!strcmp(kbuf, "update_property"))
+		rv = do_update_property(tmp, count - (tmp - kbuf));
 	else
 		rv = -EINVAL;
 out:
diff --git a/arch/powerpc/platforms/pseries/scanlog.c b/arch/powerpc/platforms/pseries/scanlog.c
index 2edc947f7c4..50643496eb6 100644
--- a/arch/powerpc/platforms/pseries/scanlog.c
+++ b/arch/powerpc/platforms/pseries/scanlog.c
@@ -192,7 +192,7 @@ struct file_operations scanlog_fops = {
 	.release	= scanlog_release,
 };
 
-int __init scanlog_init(void)
+static int __init scanlog_init(void)
 {
 	struct proc_dir_entry *ent;
 
@@ -222,7 +222,7 @@ int __init scanlog_init(void)
 	return 0;
 }
 
-void __exit scanlog_cleanup(void)
+static void __exit scanlog_cleanup(void)
 {
 	if (proc_ppc64_scan_log_dump) {
 		kfree(proc_ppc64_scan_log_dump->data);
diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index 4a465f067ed..da6cebaf72c 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -56,6 +56,7 @@
 #include <asm/dma.h>
 #include <asm/machdep.h>
 #include <asm/irq.h>
+#include <asm/kexec.h>
 #include <asm/time.h>
 #include <asm/nvram.h>
 #include "xics.h"
@@ -68,6 +69,7 @@
 #include <asm/smp.h>
 
 #include "plpar_wrappers.h"
+#include "ras.h"
 
 #ifdef DEBUG
 #define DBG(fmt...) udbg_printf(fmt)
@@ -76,22 +78,15 @@
 #endif
 
 extern void find_udbg_vterm(void);
-extern void system_reset_fwnmi(void);	/* from head.S */
-extern void machine_check_fwnmi(void);	/* from head.S */
-extern void generic_find_legacy_serial_ports(u64 *physport,
-		unsigned int *default_speed);
 
 int fwnmi_active;  /* TRUE if an FWNMI handler is present */
 
-extern void pSeries_system_reset_exception(struct pt_regs *regs);
-extern int pSeries_machine_check_exception(struct pt_regs *regs);
-
 static void pseries_shared_idle(void);
 static void pseries_dedicated_idle(void);
 
 struct mpic *pSeries_mpic;
 
-void pSeries_show_cpuinfo(struct seq_file *m)
+static void pSeries_show_cpuinfo(struct seq_file *m)
 {
 	struct device_node *root;
 	const char *model = "";
@@ -105,18 +100,22 @@ void pSeries_show_cpuinfo(struct seq_file *m)
 
 /* Initialize firmware assisted non-maskable interrupts if
  * the firmware supports this feature.
- *
  */
 static void __init fwnmi_init(void)
 {
-	int ret;
+	unsigned long system_reset_addr, machine_check_addr;
+
 	int ibm_nmi_register = rtas_token("ibm,nmi-register");
 	if (ibm_nmi_register == RTAS_UNKNOWN_SERVICE)
 		return;
-	ret = rtas_call(ibm_nmi_register, 2, 1, NULL,
-			__pa((unsigned long)system_reset_fwnmi),
-			__pa((unsigned long)machine_check_fwnmi));
-	if (ret == 0)
+
+	/* If the kernel's not linked at zero we point the firmware at low
+	 * addresses anyway, and use a trampoline to get to the real code. */
+	system_reset_addr  = __pa(system_reset_fwnmi) - PHYSICAL_START;
+	machine_check_addr = __pa(machine_check_fwnmi) - PHYSICAL_START;
+
+	if (0 == rtas_call(ibm_nmi_register, 2, 1, NULL, system_reset_addr,
+				machine_check_addr))
 		fwnmi_active = 1;
 }
 
@@ -191,7 +190,7 @@ static void pseries_lpar_enable_pmcs(void)
 
 	/* instruct hypervisor to maintain PMCs */
 	if (firmware_has_feature(FW_FEATURE_SPLPAR))
-		get_paca()->lppaca.pmcregs_in_use = 1;
+		get_lppaca()->pmcregs_in_use = 1;
 }
 
 static void __init pSeries_setup_arch(void)
@@ -235,7 +234,7 @@ static void __init pSeries_setup_arch(void)
 	/* Choose an idle loop */
 	if (firmware_has_feature(FW_FEATURE_SPLPAR)) {
 		vpa_init(boot_cpuid);
-		if (get_paca()->lppaca.shared_proc) {
+		if (get_lppaca()->shared_proc) {
 			printk(KERN_INFO "Using shared processor idle loop\n");
 			ppc_md.idle_loop = pseries_shared_idle;
 		} else {
@@ -323,15 +322,18 @@ static  void __init pSeries_discover_pic(void)
 	ppc64_interrupt_controller = IC_INVALID;
 	for (np = NULL; (np = of_find_node_by_name(np, "interrupt-controller"));) {
 		typep = (char *)get_property(np, "compatible", NULL);
-		if (strstr(typep, "open-pic"))
+		if (strstr(typep, "open-pic")) {
 			ppc64_interrupt_controller = IC_OPEN_PIC;
-		else if (strstr(typep, "ppc-xicp"))
+			break;
+		} else if (strstr(typep, "ppc-xicp")) {
 			ppc64_interrupt_controller = IC_PPC_XIC;
-		else
-			printk("pSeries_discover_pic: failed to recognize"
-			       " interrupt-controller\n");
-		break;
+			break;
+		}
 	}
+	if (ppc64_interrupt_controller == IC_INVALID)
+		printk("pSeries_discover_pic: failed to recognize"
+			" interrupt-controller\n");
+
 }
 
 static void pSeries_mach_cpu_die(void)
@@ -365,10 +367,7 @@ static int pseries_set_xdabr(unsigned long dabr)
  */
 static void __init pSeries_init_early(void)
 {
-	void *comport;
 	int iommu_off = 0;
-	unsigned int default_speed;
-	u64 physport;
 
 	DBG(" -> pSeries_init_early()\n");
 
@@ -382,17 +381,8 @@ static void __init pSeries_init_early(void)
 			     get_property(of_chosen, "linux,iommu-off", NULL));
 	}
 
-	generic_find_legacy_serial_ports(&physport, &default_speed);
-
 	if (platform_is_lpar())
 		find_udbg_vterm();
-	else if (physport) {
-		/* Map the uart for udbg. */
-		comport = (void *)ioremap(physport, 16);
-		udbg_init_uart(comport, default_speed);
-
-		DBG("Hello World !\n");
-	}
 
 	if (firmware_has_feature(FW_FEATURE_DABR))
 		ppc_md.set_dabr = pseries_set_dabr;
@@ -454,10 +444,10 @@ DECLARE_PER_CPU(unsigned long, smt_snooze_delay);
 
 static inline void dedicated_idle_sleep(unsigned int cpu)
 {
-	struct paca_struct *ppaca = &paca[cpu ^ 1];
+	struct lppaca *plppaca = &lppaca[cpu ^ 1];
 
 	/* Only sleep if the other thread is not idle */
-	if (!(ppaca->lppaca.idle)) {
+	if (!(plppaca->idle)) {
 		local_irq_disable();
 
 		/*
@@ -490,7 +480,6 @@ static inline void dedicated_idle_sleep(unsigned int cpu)
 
 static void pseries_dedicated_idle(void)
 { 
-	struct paca_struct *lpaca = get_paca();
 	unsigned int cpu = smp_processor_id();
 	unsigned long start_snooze;
 	unsigned long *smt_snooze_delay = &__get_cpu_var(smt_snooze_delay);
@@ -501,7 +490,7 @@ static void pseries_dedicated_idle(void)
 		 * Indicate to the HV that we are idle. Now would be
 		 * a good time to find other work to dispatch.
 		 */
-		lpaca->lppaca.idle = 1;
+		get_lppaca()->idle = 1;
 
 		if (!need_resched()) {
 			start_snooze = get_tb() +
@@ -528,7 +517,7 @@ static void pseries_dedicated_idle(void)
 			HMT_medium();
 		}
 
-		lpaca->lppaca.idle = 0;
+		get_lppaca()->idle = 0;
 		ppc64_runlatch_on();
 
 		preempt_enable_no_resched();
@@ -542,7 +531,6 @@ static void pseries_dedicated_idle(void)
 
 static void pseries_shared_idle(void)
 {
-	struct paca_struct *lpaca = get_paca();
 	unsigned int cpu = smp_processor_id();
 
 	while (1) {
@@ -550,7 +538,7 @@ static void pseries_shared_idle(void)
 		 * Indicate to the HV that we are idle. Now would be
 		 * a good time to find other work to dispatch.
 		 */
-		lpaca->lppaca.idle = 1;
+		get_lppaca()->idle = 1;
 
 		while (!need_resched() && !cpu_is_offline(cpu)) {
 			local_irq_disable();
@@ -574,7 +562,7 @@ static void pseries_shared_idle(void)
 			HMT_medium();
 		}
 
-		lpaca->lppaca.idle = 0;
+		get_lppaca()->idle = 0;
 		ppc64_runlatch_on();
 
 		preempt_enable_no_resched();
@@ -598,7 +586,7 @@ static void pseries_kexec_cpu_down(int crash_shutdown, int secondary)
 {
 	/* Don't risk a hypervisor call if we're crashing */
 	if (!crash_shutdown) {
-		unsigned long vpa = __pa(&get_paca()->lppaca);
+		unsigned long vpa = __pa(get_lppaca());
 
 		if (unregister_vpa(hard_smp_processor_id(), vpa)) {
 			printk("VPA deregistration of cpu %u (hw_cpu_id %d) "
@@ -638,5 +626,8 @@ struct machdep_calls __initdata pSeries_md = {
 	.machine_check_exception = pSeries_machine_check_exception,
 #ifdef CONFIG_KEXEC
 	.kexec_cpu_down		= pseries_kexec_cpu_down,
+	.machine_kexec		= default_machine_kexec,
+	.machine_kexec_prepare	= default_machine_kexec_prepare,
+	.machine_crash_shutdown	= default_machine_crash_shutdown,
 #endif
 };
diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c
index 25181c594d7..8e6b1ed1396 100644
--- a/arch/powerpc/platforms/pseries/smp.c
+++ b/arch/powerpc/platforms/pseries/smp.c
@@ -93,7 +93,7 @@ static int query_cpu_stopped(unsigned int pcpu)
 	return cpu_status;
 }
 
-int pSeries_cpu_disable(void)
+static int pSeries_cpu_disable(void)
 {
 	int cpu = smp_processor_id();
 
@@ -109,7 +109,7 @@ int pSeries_cpu_disable(void)
 	return 0;
 }
 
-void pSeries_cpu_die(unsigned int cpu)
+static void pSeries_cpu_die(unsigned int cpu)
 {
 	int tries;
 	int cpu_status;
@@ -282,7 +282,7 @@ static inline int __devinit smp_startup_cpu(unsigned int lcpu)
 	pcpu = get_hard_smp_processor_id(lcpu);
 
 	/* Fixup atomic count: it exited inside IRQ handler. */
-	paca[lcpu].__current->thread_info->preempt_count	= 0;
+	task_thread_info(paca[lcpu].__current)->preempt_count	= 0;
 
 	/* 
 	 * If the RTAS start-cpu token does not exist then presume the
diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c
index 0377decc071..fd823c7c9ac 100644
--- a/arch/powerpc/platforms/pseries/xics.c
+++ b/arch/powerpc/platforms/pseries/xics.c
@@ -381,7 +381,7 @@ int xics_get_irq(struct pt_regs *regs)
 
 #ifdef CONFIG_SMP
 
-irqreturn_t xics_ipi_action(int irq, void *dev_id, struct pt_regs *regs)
+static irqreturn_t xics_ipi_action(int irq, void *dev_id, struct pt_regs *regs)
 {
 	int cpu = smp_processor_id();
 
@@ -407,7 +407,7 @@ irqreturn_t xics_ipi_action(int irq, void *dev_id, struct pt_regs *regs)
 			smp_message_recv(PPC_MSG_MIGRATE_TASK, regs);
 		}
 #endif
-#ifdef CONFIG_DEBUGGER
+#if defined(CONFIG_DEBUGGER) || defined(CONFIG_KEXEC)
 		if (test_and_clear_bit(PPC_MSG_DEBUGGER_BREAK,
 				       &xics_ipi_message[cpu].value)) {
 			mb();