Merge branch 'master' of /pub/scm/linux/kernel/git/torvalds/linux-2.6

15 files changed, 466 insertions, 143 deletions

diff --git a/arch/powerpc/platforms/cell/Kconfig b/arch/powerpc/platforms/cell/Kconfig
index 3a963b4a9be0..2f169991896d 100644
--- a/arch/powerpc/platforms/cell/Kconfig
+++ b/arch/powerpc/platforms/cell/Kconfig
@@ -54,6 +54,13 @@ config SPU_FS_64K_LS
 	  uses 4K pages. This can improve performances of applications
 	  using multiple SPEs by lowering the TLB pressure on them.
 
+config SPU_TRACE
+	tristate "SPU event tracing support"
+	depends on SPU_FS && MARKERS
+	help
+	  This option allows reading a trace of spu-related events through
+	  the sputrace file in procfs.
+
 config SPU_BASE
 	bool
 	default n
diff --git a/arch/powerpc/platforms/cell/axon_msi.c b/arch/powerpc/platforms/cell/axon_msi.c
index 095988f13bf4..d95e71dee91f 100644
--- a/arch/powerpc/platforms/cell/axon_msi.c
+++ b/arch/powerpc/platforms/cell/axon_msi.c
@@ -13,7 +13,7 @@
 #include <linux/kernel.h>
 #include <linux/pci.h>
 #include <linux/msi.h>
-#include <linux/reboot.h>
+#include <linux/of_platform.h>
 
 #include <asm/dcr.h>
 #include <asm/machdep.h>
@@ -65,14 +65,12 @@
 
 struct axon_msic {
 	struct irq_host *irq_host;
-	__le32 *fifo;
+	__le32 *fifo_virt;
+	dma_addr_t fifo_phys;
 	dcr_host_t dcr_host;
-	struct list_head list;
 	u32 read_offset;
 };
 
-static LIST_HEAD(axon_msic_list);
-
 static void msic_dcr_write(struct axon_msic *msic, unsigned int dcr_n, u32 val)
 {
 	pr_debug("axon_msi: dcr_write(0x%x, 0x%x)\n", val, dcr_n);
@@ -94,7 +92,7 @@ static void axon_msi_cascade(unsigned int irq, struct irq_desc *desc)
 
 	while (msic->read_offset != write_offset) {
 		idx  = msic->read_offset / sizeof(__le32);
-		msi  = le32_to_cpu(msic->fifo[idx]);
+		msi  = le32_to_cpu(msic->fifo_virt[idx]);
 		msi &= 0xFFFF;
 
 		pr_debug("axon_msi: woff %x roff %x msi %x\n",
@@ -139,6 +137,7 @@ static struct axon_msic *find_msi_translator(struct pci_dev *dev)
 
 	tmp = dn;
 	dn = of_find_node_by_phandle(*ph);
+	of_node_put(tmp);
 	if (!dn) {
 		dev_dbg(&dev->dev,
 			"axon_msi: msi-translator doesn't point to a node\n");
@@ -156,7 +155,6 @@ static struct axon_msic *find_msi_translator(struct pci_dev *dev)
 
 out_error:
 	of_node_put(dn);
-	of_node_put(tmp);
 
 	return msic;
 }
@@ -292,30 +290,24 @@ static struct irq_host_ops msic_host_ops = {
 	.map	= msic_host_map,
 };
 
-static int axon_msi_notify_reboot(struct notifier_block *nb,
-				  unsigned long code, void *data)
+static int axon_msi_shutdown(struct of_device *device)
 {
-	struct axon_msic *msic;
+	struct axon_msic *msic = device->dev.platform_data;
 	u32 tmp;
 
-	list_for_each_entry(msic, &axon_msic_list, list) {
-		pr_debug("axon_msi: disabling %s\n",
-			  msic->irq_host->of_node->full_name);
-		tmp  = dcr_read(msic->dcr_host, MSIC_CTRL_REG);
-		tmp &= ~MSIC_CTRL_ENABLE & ~MSIC_CTRL_IRQ_ENABLE;
-		msic_dcr_write(msic, MSIC_CTRL_REG, tmp);
-	}
+	pr_debug("axon_msi: disabling %s\n",
+		  msic->irq_host->of_node->full_name);
+	tmp  = dcr_read(msic->dcr_host, MSIC_CTRL_REG);
+	tmp &= ~MSIC_CTRL_ENABLE & ~MSIC_CTRL_IRQ_ENABLE;
+	msic_dcr_write(msic, MSIC_CTRL_REG, tmp);
 
 	return 0;
 }
 
-static struct notifier_block axon_msi_reboot_notifier = {
-	.notifier_call = axon_msi_notify_reboot
-};
-
-static int axon_msi_setup_one(struct device_node *dn)
+static int axon_msi_probe(struct of_device *device,
+			  const struct of_device_id *device_id)
 {
-	struct page *page;
+	struct device_node *dn = device->node;
 	struct axon_msic *msic;
 	unsigned int virq;
 	int dcr_base, dcr_len;
@@ -346,16 +338,14 @@ static int axon_msi_setup_one(struct device_node *dn)
 		goto out_free_msic;
 	}
 
-	page = alloc_pages_node(of_node_to_nid(dn), GFP_KERNEL,
-				get_order(MSIC_FIFO_SIZE_BYTES));
-	if (!page) {
+	msic->fifo_virt = dma_alloc_coherent(&device->dev, MSIC_FIFO_SIZE_BYTES,
+					     &msic->fifo_phys, GFP_KERNEL);
+	if (!msic->fifo_virt) {
 		printk(KERN_ERR "axon_msi: couldn't allocate fifo for %s\n",
 		       dn->full_name);
 		goto out_free_msic;
 	}
 
-	msic->fifo = page_address(page);
-
 	msic->irq_host = irq_alloc_host(of_node_get(dn), IRQ_HOST_MAP_NOMAP,
 					NR_IRQS, &msic_host_ops, 0);
 	if (!msic->irq_host) {
@@ -378,14 +368,18 @@ static int axon_msi_setup_one(struct device_node *dn)
 	pr_debug("axon_msi: irq 0x%x setup for axon_msi\n", virq);
 
 	/* Enable the MSIC hardware */
-	msic_dcr_write(msic, MSIC_BASE_ADDR_HI_REG, (u64)msic->fifo >> 32);
+	msic_dcr_write(msic, MSIC_BASE_ADDR_HI_REG, msic->fifo_phys >> 32);
 	msic_dcr_write(msic, MSIC_BASE_ADDR_LO_REG,
-				  (u64)msic->fifo & 0xFFFFFFFF);
+				  msic->fifo_phys & 0xFFFFFFFF);
 	msic_dcr_write(msic, MSIC_CTRL_REG,
 			MSIC_CTRL_IRQ_ENABLE | MSIC_CTRL_ENABLE |
 			MSIC_CTRL_FIFO_SIZE);
 
-	list_add(&msic->list, &axon_msic_list);
+	device->dev.platform_data = msic;
+
+	ppc_md.setup_msi_irqs = axon_msi_setup_msi_irqs;
+	ppc_md.teardown_msi_irqs = axon_msi_teardown_msi_irqs;
+	ppc_md.msi_check_device = axon_msi_check_device;
 
 	printk(KERN_DEBUG "axon_msi: setup MSIC on %s\n", dn->full_name);
 
@@ -394,7 +388,8 @@ static int axon_msi_setup_one(struct device_node *dn)
 out_free_host:
 	kfree(msic->irq_host);
 out_free_fifo:
-	__free_pages(virt_to_page(msic->fifo), get_order(MSIC_FIFO_SIZE_BYTES));
+	dma_free_coherent(&device->dev, MSIC_FIFO_SIZE_BYTES, msic->fifo_virt,
+			  msic->fifo_phys);
 out_free_msic:
 	kfree(msic);
 out:
@@ -402,28 +397,24 @@ out:
 	return -1;
 }
 
-static int axon_msi_init(void)
-{
-	struct device_node *dn;
-	int found = 0;
-
-	pr_debug("axon_msi: initialising ...\n");
-
-	for_each_compatible_node(dn, NULL, "ibm,axon-msic") {
-		if (axon_msi_setup_one(dn) == 0)
-			found++;
-	}
-
-	if (found) {
-		ppc_md.setup_msi_irqs = axon_msi_setup_msi_irqs;
-		ppc_md.teardown_msi_irqs = axon_msi_teardown_msi_irqs;
-		ppc_md.msi_check_device = axon_msi_check_device;
-
-		register_reboot_notifier(&axon_msi_reboot_notifier);
+static const struct of_device_id axon_msi_device_id[] = {
+	{
+		.compatible	= "ibm,axon-msic"
+	},
+	{}
+};
 
-		pr_debug("axon_msi: registered callbacks!\n");
-	}
+static struct of_platform_driver axon_msi_driver = {
+	.match_table	= axon_msi_device_id,
+	.probe		= axon_msi_probe,
+	.shutdown	= axon_msi_shutdown,
+	.driver		= {
+		.name	= "axon-msi"
+	},
+};
 
-	return 0;
+static int __init axon_msi_init(void)
+{
+	return of_register_platform_driver(&axon_msi_driver);
 }
-arch_initcall(axon_msi_init);
+subsys_initcall(axon_msi_init);
diff --git a/arch/powerpc/platforms/cell/iommu.c b/arch/powerpc/platforms/cell/iommu.c
index df330666ccc9..edab631a8dcb 100644
--- a/arch/powerpc/platforms/cell/iommu.c
+++ b/arch/powerpc/platforms/cell/iommu.c
@@ -26,6 +26,7 @@
 #include <linux/init.h>
 #include <linux/interrupt.h>
 #include <linux/notifier.h>
+#include <linux/of.h>
 #include <linux/of_platform.h>
 
 #include <asm/prom.h>
@@ -789,18 +790,16 @@ static int __init cell_iommu_init_disabled(void)
 static u64 cell_iommu_get_fixed_address(struct device *dev)
 {
 	u64 cpu_addr, size, best_size, pci_addr = OF_BAD_ADDR;
-	struct device_node *tmp, *np;
+	struct device_node *np;
 	const u32 *ranges = NULL;
 	int i, len, best;
 
-	np = dev->archdata.of_node;
-	of_node_get(np);
-	ranges = of_get_property(np, "dma-ranges", &len);
-	while (!ranges && np) {
-		tmp = of_get_parent(np);
-		of_node_put(np);
-		np = tmp;
+	np = of_node_get(dev->archdata.of_node);
+	while (np) {
 		ranges = of_get_property(np, "dma-ranges", &len);
+		if (ranges)
+			break;
+		np = of_get_next_parent(np);
 	}
 
 	if (!ranges) {
@@ -842,19 +841,18 @@ static int dma_set_mask_and_switch(struct device *dev, u64 dma_mask)
 	if (!dev->dma_mask || !dma_supported(dev, dma_mask))
 		return -EIO;
 
-	if (dma_mask == DMA_BIT_MASK(64)) {
-		if (cell_iommu_get_fixed_address(dev) == OF_BAD_ADDR)
-			dev_dbg(dev, "iommu: 64-bit OK, but bad addr\n");
-		else {
-			dev_dbg(dev, "iommu: 64-bit OK, using fixed ops\n");
-			set_dma_ops(dev, &dma_iommu_fixed_ops);
-			cell_dma_dev_setup(dev);
-		}
+	if (dma_mask == DMA_BIT_MASK(64) &&
+		cell_iommu_get_fixed_address(dev) != OF_BAD_ADDR)
+	{
+		dev_dbg(dev, "iommu: 64-bit OK, using fixed ops\n");
+		set_dma_ops(dev, &dma_iommu_fixed_ops);
 	} else {
 		dev_dbg(dev, "iommu: not 64-bit, using default ops\n");
 		set_dma_ops(dev, get_pci_dma_ops());
 	}
 
+	cell_dma_dev_setup(dev);
+
 	*dev->dma_mask = dma_mask;
 
 	return 0;
@@ -918,6 +916,18 @@ static int __init cell_iommu_fixed_mapping_init(void)
 		return -1;
 	}
 
+	/* We must have dma-ranges properties for fixed mapping to work */
+	for (np = NULL; (np = of_find_all_nodes(np));) {
+		if (of_find_property(np, "dma-ranges", NULL))
+			break;
+	}
+	of_node_put(np);
+
+	if (!np) {
+		pr_debug("iommu: no dma-ranges found, no fixed mapping\n");
+		return -1;
+	}
+
 	/* The default setup is to have the fixed mapping sit after the
 	 * dynamic region, so find the top of the largest IOMMU window
 	 * on any axon, then add the size of RAM and that's our max value.
@@ -981,8 +991,8 @@ static int __init cell_iommu_fixed_mapping_init(void)
 			dsize = htab_size_bytes;
 		}
 
-		pr_debug("iommu: setting up %d, dynamic window %lx-%lx " \
-			 "fixed window %lx-%lx\n", iommu->nid, dbase,
+		printk(KERN_DEBUG "iommu: node %d, dynamic window 0x%lx-0x%lx "
+			"fixed window 0x%lx-0x%lx\n", iommu->nid, dbase,
 			 dbase + dsize, fbase, fbase + fsize);
 
 		cell_iommu_setup_page_tables(iommu, dbase, dsize, fbase, fsize);
@@ -998,8 +1008,6 @@ static int __init cell_iommu_fixed_mapping_init(void)
 	dma_iommu_ops.set_dma_mask = dma_set_mask_and_switch;
 	set_pci_dma_ops(&dma_iommu_ops);
 
-	printk(KERN_DEBUG "IOMMU fixed mapping established.\n");
-
 	return 0;
 }
 
diff --git a/arch/powerpc/platforms/cell/ras.c b/arch/powerpc/platforms/cell/ras.c
index b2494ebcdbe9..e43024c0392e 100644
--- a/arch/powerpc/platforms/cell/ras.c
+++ b/arch/powerpc/platforms/cell/ras.c
@@ -1,4 +1,13 @@
-#define DEBUG
+/*
+ * Copyright 2006-2008, 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.
+ */
+
+#undef DEBUG
 
 #include <linux/types.h>
 #include <linux/kernel.h>
diff --git a/arch/powerpc/platforms/cell/setup.c b/arch/powerpc/platforms/cell/setup.c
index e6534b519c9a..a7f609b3b876 100644
--- a/arch/powerpc/platforms/cell/setup.c
+++ b/arch/powerpc/platforms/cell/setup.c
@@ -98,7 +98,7 @@ static int __init cell_publish_devices(void)
 	}
 	return 0;
 }
-machine_device_initcall(cell, cell_publish_devices);
+machine_subsys_initcall(cell, cell_publish_devices);
 
 static void cell_mpic_cascade(unsigned int irq, struct irq_desc *desc)
 {
diff --git a/arch/powerpc/platforms/cell/spufs/Makefile b/arch/powerpc/platforms/cell/spufs/Makefile
index d3a349fb42e5..99610a6361f2 100644
--- a/arch/powerpc/platforms/cell/spufs/Makefile
+++ b/arch/powerpc/platforms/cell/spufs/Makefile
@@ -4,6 +4,8 @@ spufs-y += inode.o file.o context.o syscalls.o coredump.o
 spufs-y += sched.o backing_ops.o hw_ops.o run.o gang.o
 spufs-y += switch.o fault.o lscsa_alloc.o
 
+obj-$(CONFIG_SPU_TRACE)	+= sputrace.o
+
 # Rules to build switch.o with the help of SPU tool chain
 SPU_CROSS	:= spu-
 SPU_CC		:= $(SPU_CROSS)gcc
diff --git a/arch/powerpc/platforms/cell/spufs/backing_ops.c b/arch/powerpc/platforms/cell/spufs/backing_ops.c
index 50d98a154aaf..64eb15b22040 100644
--- a/arch/powerpc/platforms/cell/spufs/backing_ops.c
+++ b/arch/powerpc/platforms/cell/spufs/backing_ops.c
@@ -288,6 +288,12 @@ static void spu_backing_runcntl_write(struct spu_context *ctx, u32 val)
 	spin_lock(&ctx->csa.register_lock);
 	ctx->csa.prob.spu_runcntl_RW = val;
 	if (val & SPU_RUNCNTL_RUNNABLE) {
+		ctx->csa.prob.spu_status_R &=
+			~SPU_STATUS_STOPPED_BY_STOP &
+			~SPU_STATUS_STOPPED_BY_HALT &
+			~SPU_STATUS_SINGLE_STEP &
+			~SPU_STATUS_INVALID_INSTR &
+			~SPU_STATUS_INVALID_CH;
 		ctx->csa.prob.spu_status_R |= SPU_STATUS_RUNNING;
 	} else {
 		ctx->csa.prob.spu_status_R &= ~SPU_STATUS_RUNNING;
diff --git a/arch/powerpc/platforms/cell/spufs/fault.c b/arch/powerpc/platforms/cell/spufs/fault.c
index eff4d291ba85..e46d300e21a5 100644
--- a/arch/powerpc/platforms/cell/spufs/fault.c
+++ b/arch/powerpc/platforms/cell/spufs/fault.c
@@ -108,7 +108,7 @@ int spufs_handle_class1(struct spu_context *ctx)
 	u64 ea, dsisr, access;
 	unsigned long flags;
 	unsigned flt = 0;
-	int ret, ret2;
+	int ret;
 
 	/*
 	 * dar and dsisr get passed from the registers
@@ -148,13 +148,10 @@ int spufs_handle_class1(struct spu_context *ctx)
 		ret = spu_handle_mm_fault(current->mm, ea, dsisr, &flt);
 
 	/*
-	 * If spu_acquire fails due to a pending signal we just want to return
-	 * EINTR to userspace even if that means missing the dma restart or
-	 * updating the page fault statistics.
+	 * This is nasty: we need the state_mutex for all the bookkeeping even
+	 * if the syscall was interrupted by a signal. ewww.
 	 */
-	ret2 = spu_acquire(ctx);
-	if (ret2)
-		goto out;
+	mutex_lock(&ctx->state_mutex);
 
 	/*
 	 * Clear dsisr under ctxt lock after handling the fault, so that
@@ -185,7 +182,6 @@ int spufs_handle_class1(struct spu_context *ctx)
 	} else
 		spufs_handle_event(ctx, ea, SPE_EVENT_SPE_DATA_STORAGE);
 
- out:
 	spuctx_switch_state(ctx, SPU_UTIL_SYSTEM);
 	return ret;
 }
diff --git a/arch/powerpc/platforms/cell/spufs/file.c b/arch/powerpc/platforms/cell/spufs/file.c
index 3fcd06418b01..c66c3756970d 100644
--- a/arch/powerpc/platforms/cell/spufs/file.c
+++ b/arch/powerpc/platforms/cell/spufs/file.c
@@ -29,6 +29,7 @@
 #include <linux/poll.h>
 #include <linux/ptrace.h>
 #include <linux/seq_file.h>
+#include <linux/marker.h>
 
 #include <asm/io.h>
 #include <asm/semaphore.h>
@@ -357,6 +358,9 @@ static unsigned long spufs_ps_nopfn(struct vm_area_struct *vma,
 {
 	struct spu_context *ctx = vma->vm_file->private_data;
 	unsigned long area, offset = address - vma->vm_start;
+	int ret = 0;
+
+	spu_context_nospu_trace(spufs_ps_nopfn__enter, ctx);
 
 	offset += vma->vm_pgoff << PAGE_SHIFT;
 	if (offset >= ps_size)
@@ -375,14 +379,18 @@ static unsigned long spufs_ps_nopfn(struct vm_area_struct *vma,
 
 	if (ctx->state == SPU_STATE_SAVED) {
 		up_read(&current->mm->mmap_sem);
-		spufs_wait(ctx->run_wq, ctx->state == SPU_STATE_RUNNABLE);
+		spu_context_nospu_trace(spufs_ps_nopfn__sleep, ctx);
+		ret = spufs_wait(ctx->run_wq, ctx->state == SPU_STATE_RUNNABLE);
+		spu_context_trace(spufs_ps_nopfn__wake, ctx, ctx->spu);
 		down_read(&current->mm->mmap_sem);
 	} else {
 		area = ctx->spu->problem_phys + ps_offs;
 		vm_insert_pfn(vma, address, (area + offset) >> PAGE_SHIFT);
+		spu_context_trace(spufs_ps_nopfn__insert, ctx, ctx->spu);
 	}
 
-	spu_release(ctx);
+	if (!ret)
+		spu_release(ctx);
 	return NOPFN_REFAULT;
 }
 
@@ -454,7 +462,7 @@ static int spufs_cntl_open(struct inode *inode, struct file *file)
 	if (!i->i_openers++)
 		ctx->cntl = inode->i_mapping;
 	mutex_unlock(&ctx->mapping_lock);
-	return spufs_attr_open(inode, file, spufs_cntl_get,
+	return simple_attr_open(inode, file, spufs_cntl_get,
 					spufs_cntl_set, "0x%08lx");
 }
 
@@ -464,7 +472,7 @@ spufs_cntl_release(struct inode *inode, struct file *file)
 	struct spufs_inode_info *i = SPUFS_I(inode);
 	struct spu_context *ctx = i->i_ctx;
 
-	spufs_attr_release(inode, file);
+	simple_attr_release(inode, file);
 
 	mutex_lock(&ctx->mapping_lock);
 	if (!--i->i_openers)
@@ -476,8 +484,8 @@ spufs_cntl_release(struct inode *inode, struct file *file)
 static const struct file_operations spufs_cntl_fops = {
 	.open = spufs_cntl_open,
 	.release = spufs_cntl_release,
-	.read = spufs_attr_read,
-	.write = spufs_attr_write,
+	.read = simple_attr_read,
+	.write = simple_attr_write,
 	.mmap = spufs_cntl_mmap,
 };
 
@@ -749,23 +757,25 @@ static ssize_t spufs_ibox_read(struct file *file, char __user *buf,
 
 	count = spu_acquire(ctx);
 	if (count)
-		return count;
+		goto out;
 
 	/* wait only for the first element */
 	count = 0;
 	if (file->f_flags & O_NONBLOCK) {
-		if (!spu_ibox_read(ctx, &ibox_data))
+		if (!spu_ibox_read(ctx, &ibox_data)) {
 			count = -EAGAIN;
+			goto out_unlock;
+		}
 	} else {
 		count = spufs_wait(ctx->ibox_wq, spu_ibox_read(ctx, &ibox_data));
+		if (count)
+			goto out;
 	}
-	if (count)
-		goto out;
 
 	/* if we can't write at all, return -EFAULT */
 	count = __put_user(ibox_data, udata);
 	if (count)
-		goto out;
+		goto out_unlock;
 
 	for (count = 4, udata++; (count + 4) <= len; count += 4, udata++) {
 		int ret;
@@ -782,9 +792,9 @@ static ssize_t spufs_ibox_read(struct file *file, char __user *buf,
 			break;
 	}
 
-out:
+out_unlock:
 	spu_release(ctx);
-
+out:
 	return count;
 }
 
@@ -899,7 +909,7 @@ static ssize_t spufs_wbox_write(struct file *file, const char __user *buf,
 
 	count = spu_acquire(ctx);
 	if (count)
-		return count;
+		goto out;
 
 	/*
 	 * make sure we can at least write one element, by waiting
@@ -907,14 +917,16 @@ static ssize_t spufs_wbox_write(struct file *file, const char __user *buf,
 	 */
 	count = 0;
 	if (file->f_flags & O_NONBLOCK) {
-		if (!spu_wbox_write(ctx, wbox_data))
+		if (!spu_wbox_write(ctx, wbox_data)) {
 			count = -EAGAIN;
+			goto out_unlock;
+		}
 	} else {
 		count = spufs_wait(ctx->wbox_wq, spu_wbox_write(ctx, wbox_data));
+		if (count)
+			goto out;
 	}
 
-	if (count)
-		goto out;
 
 	/* write as much as possible */
 	for (count = 4, udata++; (count + 4) <= len; count += 4, udata++) {
@@ -928,8 +940,9 @@ static ssize_t spufs_wbox_write(struct file *file, const char __user *buf,
 			break;
 	}
 
-out:
+out_unlock:
 	spu_release(ctx);
+out:
 	return count;
 }
 
@@ -1592,12 +1605,11 @@ static ssize_t spufs_mfc_read(struct file *file, char __user *buffer,
 	} else {
 		ret = spufs_wait(ctx->mfc_wq,
 			   spufs_read_mfc_tagstatus(ctx, &status));
+		if (ret)
+			goto out;
 	}
 	spu_release(ctx);
 
-	if (ret)
-		goto out;
-
 	ret = 4;
 	if (copy_to_user(buffer, &status, 4))
 		ret = -EFAULT;
@@ -1726,6 +1738,8 @@ static ssize_t spufs_mfc_write(struct file *file, const char __user *buffer,
 		int status;
 		ret = spufs_wait(ctx->mfc_wq,
 				 spu_send_mfc_command(ctx, cmd, &status));
+		if (ret)
+			goto out;
 		if (status)
 			ret = status;
 	}
@@ -1779,7 +1793,7 @@ static int spufs_mfc_flush(struct file *file, fl_owner_t id)
 
 	ret = spu_acquire(ctx);
 	if (ret)
-		return ret;
+		goto out;
 #if 0
 /* this currently hangs */
 	ret = spufs_wait(ctx->mfc_wq,
@@ -1788,12 +1802,13 @@ static int spufs_mfc_flush(struct file *file, fl_owner_t id)
 		goto out;
 	ret = spufs_wait(ctx->mfc_wq,
 			 ctx->ops->read_mfc_tagstatus(ctx) == ctx->tagwait);
-out:
+	if (ret)
+		goto out;
 #else
 	ret = 0;
 #endif
 	spu_release(ctx);
-
+out:
 	return ret;
 }
 
diff --git a/arch/powerpc/platforms/cell/spufs/inode.c b/arch/powerpc/platforms/cell/spufs/inode.c
index c0e968a4c211..6d1228c66c5e 100644
--- a/arch/powerpc/platforms/cell/spufs/inode.c
+++ b/arch/powerpc/platforms/cell/spufs/inode.c
@@ -1,3 +1,4 @@
+
 /*
  * SPU file system
  *
@@ -322,7 +323,7 @@ static struct spu_context *
 spufs_assert_affinity(unsigned int flags, struct spu_gang *gang,
 						struct file *filp)
 {
-	struct spu_context *tmp, *neighbor;
+	struct spu_context *tmp, *neighbor, *err;
 	int count, node;
 	int aff_supp;
 
@@ -354,11 +355,15 @@ spufs_assert_affinity(unsigned int flags, struct spu_gang *gang,
 		if (!list_empty(&neighbor->aff_list) && !(neighbor->aff_head) &&
 		    !list_is_last(&neighbor->aff_list, &gang->aff_list_head) &&
 		    !list_entry(neighbor->aff_list.next, struct spu_context,
-		    aff_list)->aff_head)
-			return ERR_PTR(-EEXIST);
+		    aff_list)->aff_head) {
+			err = ERR_PTR(-EEXIST);
+			goto out_put_neighbor;
+		}
 
-		if (gang != neighbor->gang)
-			return ERR_PTR(-EINVAL);
+		if (gang != neighbor->gang) {
+			err = ERR_PTR(-EINVAL);
+			goto out_put_neighbor;
+		}
 
 		count = 1;
 		list_for_each_entry(tmp, &gang->aff_list_head, aff_list)
@@ -372,11 +377,17 @@ spufs_assert_affinity(unsigned int flags, struct spu_gang *gang,
 				break;
 		}
 
-		if (node == MAX_NUMNODES)
-			return ERR_PTR(-EEXIST);
+		if (node == MAX_NUMNODES) {
+			err = ERR_PTR(-EEXIST);
+			goto out_put_neighbor;
+		}
 	}
 
 	return neighbor;
+
+out_put_neighbor:
+	put_spu_context(neighbor);
+	return err;
 }
 
 static void
@@ -454,9 +465,12 @@ spufs_create_context(struct inode *inode, struct dentry *dentry,
 	if (ret)
 		goto out_aff_unlock;
 
-	if (affinity)
+	if (affinity) {
 		spufs_set_affinity(flags, SPUFS_I(dentry->d_inode)->i_ctx,
 								neighbor);
+		if (neighbor)
+			put_spu_context(neighbor);
+	}
 
 	/*
 	 * get references for dget and mntget, will be released
@@ -579,7 +593,7 @@ long spufs_create(struct nameidata *nd, unsigned int flags, mode_t mode,
 
 	ret = -EINVAL;
 	/* check if we are on spufs */
-	if (nd->dentry->d_sb->s_type != &spufs_type)
+	if (nd->path.dentry->d_sb->s_type != &spufs_type)
 		goto out;
 
 	/* don't accept undefined flags */
@@ -587,9 +601,9 @@ long spufs_create(struct nameidata *nd, unsigned int flags, mode_t mode,
 		goto out;
 
 	/* only threads can be underneath a gang */
-	if (nd->dentry != nd->dentry->d_sb->s_root) {
+	if (nd->path.dentry != nd->path.dentry->d_sb->s_root) {
 		if ((flags & SPU_CREATE_GANG) ||
-		    !SPUFS_I(nd->dentry->d_inode)->i_gang)
+		    !SPUFS_I(nd->path.dentry->d_inode)->i_gang)
 			goto out;
 	}
 
@@ -605,16 +619,17 @@ long spufs_create(struct nameidata *nd, unsigned int flags, mode_t mode,
 	mode &= ~current->fs->umask;
 
 	if (flags & SPU_CREATE_GANG)
-		return spufs_create_gang(nd->dentry->d_inode,
-					dentry, nd->mnt, mode);
+		return spufs_create_gang(nd->path.dentry->d_inode,
+					 dentry, nd->path.mnt, mode);
 	else
-		return spufs_create_context(nd->dentry->d_inode,
-					dentry, nd->mnt, flags, mode, filp);
+		return spufs_create_context(nd->path.dentry->d_inode,
+					    dentry, nd->path.mnt, flags, mode,
+					    filp);
 
 out_dput:
 	dput(dentry);
 out_dir:
-	mutex_unlock(&nd->dentry->d_inode->i_mutex);
+	mutex_unlock(&nd->path.dentry->d_inode->i_mutex);
 out:
 	return ret;
 }
@@ -742,8 +757,11 @@ spufs_fill_super(struct super_block *sb, void *data, int silent)
 		.statfs = simple_statfs,
 		.delete_inode = spufs_delete_inode,
 		.drop_inode = generic_delete_inode,
+		.show_options = generic_show_options,
 	};
 
+	save_mount_options(sb, data);
+
 	sb->s_maxbytes = MAX_LFS_FILESIZE;
 	sb->s_blocksize = PAGE_CACHE_SIZE;
 	sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
diff --git a/arch/powerpc/platforms/cell/spufs/run.c b/arch/powerpc/platforms/cell/spufs/run.c
index c01a09da1e56..fca22e18069a 100644
--- a/arch/powerpc/platforms/cell/spufs/run.c
+++ b/arch/powerpc/platforms/cell/spufs/run.c
@@ -53,7 +53,7 @@ int spu_stopped(struct spu_context *ctx, u32 *stat)
 
 	stopped = SPU_STATUS_INVALID_INSTR | SPU_STATUS_SINGLE_STEP |
 		SPU_STATUS_STOPPED_BY_HALT | SPU_STATUS_STOPPED_BY_STOP;
-	if (*stat & stopped)
+	if (!(*stat & SPU_STATUS_RUNNING) && (*stat & stopped))
 		return 1;
 
 	dsisr = ctx->csa.dsisr;
@@ -354,8 +354,15 @@ long spufs_run_spu(struct spu_context *ctx, u32 *npc, u32 *event)
 
 	do {
 		ret = spufs_wait(ctx->stop_wq, spu_stopped(ctx, &status));
-		if (unlikely(ret))
+		if (unlikely(ret)) {
+			/*
+			 * This is nasty: we need the state_mutex for all the
+			 * bookkeeping even if the syscall was interrupted by
+			 * a signal. ewww.
+			 */
+			mutex_lock(&ctx->state_mutex);
 			break;
+		}
 		spu = ctx->spu;
 		if (unlikely(test_and_clear_bit(SPU_SCHED_NOTIFY_ACTIVE,
 						&ctx->sched_flags))) {
@@ -388,16 +395,14 @@ long spufs_run_spu(struct spu_context *ctx, u32 *npc, u32 *event)
 				      SPU_STATUS_STOPPED_BY_HALT |
 				       SPU_STATUS_SINGLE_STEP)));
 
-	if ((status & SPU_STATUS_STOPPED_BY_STOP) &&
-	    (((status >> SPU_STOP_STATUS_SHIFT) & 0x3f00) == 0x2100) &&
-	    (ctx->state == SPU_STATE_RUNNABLE))
-		ctx->stats.libassist++;
-
-
 	spu_disable_spu(ctx);
 	ret = spu_run_fini(ctx, npc, &status);
 	spu_yield(ctx);
 
+	if ((status & SPU_STATUS_STOPPED_BY_STOP) &&
+	    (((status >> SPU_STOP_STATUS_SHIFT) & 0x3f00) == 0x2100))
+		ctx->stats.libassist++;
+
 	if ((ret == 0) ||
 	    ((ret == -ERESTARTSYS) &&
 	     ((status & SPU_STATUS_STOPPED_BY_HALT) ||
@@ -410,8 +415,11 @@ long spufs_run_spu(struct spu_context *ctx, u32 *npc, u32 *event)
 	 * since we have TIF_SINGLESTEP set, thus the kernel will do
 	 * it upon return from the syscall anyawy
 	 */
-	if ((status & SPU_STATUS_STOPPED_BY_STOP)
-	    && (status >> SPU_STOP_STATUS_SHIFT) == 0x3fff) {
+	if (unlikely(status & SPU_STATUS_SINGLE_STEP))
+		ret = -ERESTARTSYS;
+
+	else if (unlikely((status & SPU_STATUS_STOPPED_BY_STOP)
+	    && (status >> SPU_STOP_STATUS_SHIFT) == 0x3fff)) {
 		force_sig(SIGTRAP, current);
 		ret = -ERESTARTSYS;
 	}
diff --git a/arch/powerpc/platforms/cell/spufs/sched.c b/arch/powerpc/platforms/cell/spufs/sched.c
index 00d914232af1..5915343e2599 100644
--- a/arch/powerpc/platforms/cell/spufs/sched.c
+++ b/arch/powerpc/platforms/cell/spufs/sched.c
@@ -39,6 +39,7 @@
 #include <linux/pid_namespace.h>
 #include <linux/proc_fs.h>
 #include <linux/seq_file.h>
+#include <linux/marker.h>
 
 #include <asm/io.h>
 #include <asm/mmu_context.h>
@@ -216,8 +217,8 @@ void do_notify_spus_active(void)
  */
 static void spu_bind_context(struct spu *spu, struct spu_context *ctx)
 {
-	pr_debug("%s: pid=%d SPU=%d NODE=%d\n", __FUNCTION__, current->pid,
-		 spu->number, spu->node);
+	spu_context_trace(spu_bind_context__enter, ctx, spu);
+
 	spuctx_switch_state(ctx, SPU_UTIL_SYSTEM);
 
 	if (ctx->flags & SPU_CREATE_NOSCHED)
@@ -399,8 +400,8 @@ static int has_affinity(struct spu_context *ctx)
  */
 static void spu_unbind_context(struct spu *spu, struct spu_context *ctx)
 {
-	pr_debug("%s: unbind pid=%d SPU=%d NODE=%d\n", __FUNCTION__,
-		 spu->pid, spu->number, spu->node);
+	spu_context_trace(spu_unbind_context__enter, ctx, spu);
+
 	spuctx_switch_state(ctx, SPU_UTIL_SYSTEM);
 
  	if (spu->ctx->flags & SPU_CREATE_NOSCHED)
@@ -528,6 +529,8 @@ static struct spu *spu_get_idle(struct spu_context *ctx)
 	struct spu *spu, *aff_ref_spu;
 	int node, n;
 
+	spu_context_nospu_trace(spu_get_idle__enter, ctx);
+
 	if (ctx->gang) {
 		mutex_lock(&ctx->gang->aff_mutex);
 		if (has_affinity(ctx)) {
@@ -546,8 +549,7 @@ static struct spu *spu_get_idle(struct spu_context *ctx)
 			if (atomic_dec_and_test(&ctx->gang->aff_sched_count))
 				ctx->gang->aff_ref_spu = NULL;
 			mutex_unlock(&ctx->gang->aff_mutex);
-
-			return NULL;
+			goto not_found;
 		}
 		mutex_unlock(&ctx->gang->aff_mutex);
 	}
@@ -565,12 +567,14 @@ static struct spu *spu_get_idle(struct spu_context *ctx)
 		mutex_unlock(&cbe_spu_info[node].list_mutex);
 	}
 
+ not_found:
+	spu_context_nospu_trace(spu_get_idle__not_found, ctx);
 	return NULL;
 
  found:
 	spu->alloc_state = SPU_USED;
 	mutex_unlock(&cbe_spu_info[node].list_mutex);
-	pr_debug("Got SPU %d %d\n", spu->number, spu->node);
+	spu_context_trace(spu_get_idle__found, ctx, spu);
 	spu_init_channels(spu);
 	return spu;
 }
@@ -587,6 +591,8 @@ static struct spu *find_victim(struct spu_context *ctx)
 	struct spu *spu;
 	int node, n;
 
+	spu_context_nospu_trace(spu_find_vitim__enter, ctx);
+
 	/*
 	 * Look for a possible preemption candidate on the local node first.
 	 * If there is no candidate look at the other nodes.  This isn't
@@ -640,6 +646,8 @@ static struct spu *find_victim(struct spu_context *ctx)
 				goto restart;
 			}
 
+			spu_context_trace(__spu_deactivate__unload, ctx, spu);
+
 			mutex_lock(&cbe_spu_info[node].list_mutex);
 			cbe_spu_info[node].nr_active--;
 			spu_unbind_context(spu, victim);
@@ -822,6 +830,7 @@ static int __spu_deactivate(struct spu_context *ctx, int force, int max_prio)
  */
 void spu_deactivate(struct spu_context *ctx)
 {
+	spu_context_nospu_trace(spu_deactivate__enter, ctx);
 	__spu_deactivate(ctx, 1, MAX_PRIO);
 }
 
@@ -835,6 +844,7 @@ void spu_deactivate(struct spu_context *ctx)
  */
 void spu_yield(struct spu_context *ctx)
 {
+	spu_context_nospu_trace(spu_yield__enter, ctx);
 	if (!(ctx->flags & SPU_CREATE_NOSCHED)) {
 		mutex_lock(&ctx->state_mutex);
 		__spu_deactivate(ctx, 0, MAX_PRIO);
@@ -864,11 +874,15 @@ static noinline void spusched_tick(struct spu_context *ctx)
 		goto out;
 
 	spu = ctx->spu;
+
+	spu_context_trace(spusched_tick__preempt, ctx, spu);
+
 	new = grab_runnable_context(ctx->prio + 1, spu->node);
 	if (new) {
 		spu_unschedule(spu, ctx);
 		spu_add_to_rq(ctx);
 	} else {
+		spu_context_nospu_trace(spusched_tick__newslice, ctx);
 		ctx->time_slice++;
 	}
 out:
diff --git a/arch/powerpc/platforms/cell/spufs/spufs.h b/arch/powerpc/platforms/cell/spufs/spufs.h
index 0e114038ea6f..2c2fe3c07d72 100644
--- a/arch/powerpc/platforms/cell/spufs/spufs.h
+++ b/arch/powerpc/platforms/cell/spufs/spufs.h
@@ -268,6 +268,9 @@ extern char *isolated_loader;
  *	Same as wait_event_interruptible(), except that here
  *	we need to call spu_release(ctx) before sleeping, and
  *	then spu_acquire(ctx) when awoken.
+ *
+ * 	Returns with state_mutex re-acquired when successfull or
+ * 	with -ERESTARTSYS and the state_mutex dropped when interrupted.
  */
 
 #define spufs_wait(wq, condition)					\
@@ -278,11 +281,11 @@ extern char *isolated_loader;
 		prepare_to_wait(&(wq), &__wait, TASK_INTERRUPTIBLE);	\
 		if (condition)						\
 			break;						\
+		spu_release(ctx);					\
 		if (signal_pending(current)) {				\
 			__ret = -ERESTARTSYS;				\
 			break;						\
 		}							\
-		spu_release(ctx);					\
 		schedule();						\
 		__ret = spu_acquire(ctx);				\
 		if (__ret)						\
@@ -325,4 +328,9 @@ extern void spu_free_lscsa(struct spu_state *csa);
 extern void spuctx_switch_state(struct spu_context *ctx,
 		enum spu_utilization_state new_state);
 
+#define spu_context_trace(name, ctx, spu) \
+	trace_mark(name, "%p %p", ctx, spu);
+#define spu_context_nospu_trace(name, ctx) \
+	trace_mark(name, "%p", ctx);
+
 #endif
diff --git a/arch/powerpc/platforms/cell/spufs/sputrace.c b/arch/powerpc/platforms/cell/spufs/sputrace.c
new file mode 100644
index 000000000000..01974f7776e1
--- /dev/null
+++ b/arch/powerpc/platforms/cell/spufs/sputrace.c
@@ -0,0 +1,241 @@
+/*
+ * Copyright (C) 2007 IBM Deutschland Entwicklung GmbH
+ *	Released under GPL v2.
+ *
+ * Partially based on net/ipv4/tcp_probe.c.
+ *
+ * Simple tracing facility for spu contexts.
+ */
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/marker.h>
+#include <linux/proc_fs.h>
+#include <linux/wait.h>
+#include <asm/atomic.h>
+#include <asm/uaccess.h>
+#include "spufs.h"
+
+struct spu_probe {
+	const char *name;
+	const char *format;
+	marker_probe_func *probe_func;
+};
+
+struct sputrace {
+	ktime_t tstamp;
+	int owner_tid; /* owner */
+	int curr_tid;
+	const char *name;
+	int number;
+};
+
+static int bufsize __read_mostly = 16384;
+MODULE_PARM_DESC(bufsize, "Log buffer size (number of records)");
+module_param(bufsize, int, 0);
+
+
+static DEFINE_SPINLOCK(sputrace_lock);
+static DECLARE_WAIT_QUEUE_HEAD(sputrace_wait);
+static ktime_t sputrace_start;
+static unsigned long sputrace_head, sputrace_tail;
+static struct sputrace *sputrace_log;
+
+static int sputrace_used(void)
+{
+	return (sputrace_head - sputrace_tail) % bufsize;
+}
+
+static inline int sputrace_avail(void)
+{
+	return bufsize - sputrace_used();
+}
+
+static int sputrace_sprint(char *tbuf, int n)
+{
+	const struct sputrace *t = sputrace_log + sputrace_tail % bufsize;
+	struct timespec tv =
+		ktime_to_timespec(ktime_sub(t->tstamp, sputrace_start));
+
+	return snprintf(tbuf, n,
+		"[%lu.%09lu] %d: %s (thread = %d, spu = %d)\n",
+		(unsigned long) tv.tv_sec,
+		(unsigned long) tv.tv_nsec,
+		t->owner_tid,
+		t->name,
+		t->curr_tid,
+		t->number);
+}
+
+static ssize_t sputrace_read(struct file *file, char __user *buf,
+		size_t len, loff_t *ppos)
+{
+	int error = 0, cnt = 0;
+
+	if (!buf || len < 0)
+		return -EINVAL;
+
+	while (cnt < len) {
+		char tbuf[128];
+		int width;
+
+		error = wait_event_interruptible(sputrace_wait,
+						 sputrace_used() > 0);
+		if (error)
+			break;
+
+		spin_lock(&sputrace_lock);
+		if (sputrace_head == sputrace_tail) {
+			spin_unlock(&sputrace_lock);
+			continue;
+		}
+
+		width = sputrace_sprint(tbuf, sizeof(tbuf));
+		if (width < len)
+			sputrace_tail = (sputrace_tail + 1) % bufsize;
+		spin_unlock(&sputrace_lock);
+
+		if (width >= len)
+			break;
+
+		error = copy_to_user(buf + cnt, tbuf, width);
+		if (error)
+			break;
+		cnt += width;
+	}
+
+	return cnt == 0 ? error : cnt;
+}
+
+static int sputrace_open(struct inode *inode, struct file *file)
+{
+	spin_lock(&sputrace_lock);
+	sputrace_head = sputrace_tail = 0;
+	sputrace_start = ktime_get();
+	spin_unlock(&sputrace_lock);
+
+	return 0;
+}
+
+static const struct file_operations sputrace_fops = {
+	.owner	= THIS_MODULE,
+	.open	= sputrace_open,
+	.read	= sputrace_read,
+};
+
+static void sputrace_log_item(const char *name, struct spu_context *ctx,
+		struct spu *spu)
+{
+	spin_lock(&sputrace_lock);
+	if (sputrace_avail() > 1) {
+		struct sputrace *t = sputrace_log + sputrace_head;
+
+		t->tstamp = ktime_get();
+		t->owner_tid = ctx->tid;
+		t->name = name;
+		t->curr_tid = current->pid;
+		t->number = spu ? spu->number : -1;
+
+		sputrace_head = (sputrace_head + 1) % bufsize;
+	} else {
+		printk(KERN_WARNING
+		       "sputrace: lost samples due to full buffer.\n");
+	}
+	spin_unlock(&sputrace_lock);
+
+	wake_up(&sputrace_wait);
+}
+
+static void spu_context_event(void *probe_private, void *call_data,
+		const char *format, va_list *args)
+{
+	struct spu_probe *p = probe_private;
+	struct spu_context *ctx;
+	struct spu *spu;
+
+	ctx = va_arg(*args, struct spu_context *);
+	spu = va_arg(*args, struct spu *);
+
+	sputrace_log_item(p->name, ctx, spu);
+}
+
+static void spu_context_nospu_event(void *probe_private, void *call_data,
+		const char *format, va_list *args)
+{
+	struct spu_probe *p = probe_private;
+	struct spu_context *ctx;
+
+	ctx = va_arg(*args, struct spu_context *);
+
+	sputrace_log_item(p->name, ctx, NULL);
+}
+
+struct spu_probe spu_probes[] = {
+	{ "spu_bind_context__enter", "%p %p", spu_context_event },
+	{ "spu_unbind_context__enter", "%p %p", spu_context_event },
+	{ "spu_get_idle__enter", "%p", spu_context_nospu_event },
+	{ "spu_get_idle__found", "%p %p", spu_context_event },
+	{ "spu_get_idle__not_found", "%p", spu_context_nospu_event },
+	{ "spu_find_victim__enter", "%p", spu_context_nospu_event },
+	{ "spusched_tick__preempt", "%p %p", spu_context_event },
+	{ "spusched_tick__newslice", "%p", spu_context_nospu_event },
+	{ "spu_yield__enter", "%p", spu_context_nospu_event },
+	{ "spu_deactivate__enter", "%p", spu_context_nospu_event },
+	{ "__spu_deactivate__unload", "%p %p", spu_context_event },
+	{ "spufs_ps_nopfn__enter", "%p", spu_context_nospu_event },
+	{ "spufs_ps_nopfn__sleep", "%p", spu_context_nospu_event },
+	{ "spufs_ps_nopfn__wake", "%p %p", spu_context_event },
+	{ "spufs_ps_nopfn__insert", "%p %p", spu_context_event },
+	{ "spu_acquire_saved__enter", "%p", spu_context_nospu_event },
+	{ "destroy_spu_context__enter", "%p", spu_context_nospu_event },
+};
+
+static int __init sputrace_init(void)
+{
+	struct proc_dir_entry *entry;
+	int i, error = -ENOMEM;
+
+	sputrace_log = kcalloc(sizeof(struct sputrace),
+				bufsize, GFP_KERNEL);
+	if (!sputrace_log)
+		goto out;
+
+	entry = create_proc_entry("sputrace", S_IRUSR, NULL);
+	if (!entry)
+		goto out_free_log;
+	entry->proc_fops = &sputrace_fops;
+
+	for (i = 0; i < ARRAY_SIZE(spu_probes); i++) {
+		struct spu_probe *p = &spu_probes[i];
+
+		error = marker_probe_register(p->name, p->format,
+					      p->probe_func, p);
+		if (error)
+			printk(KERN_INFO "Unable to register probe %s\n",
+					p->name);
+	}
+
+	return 0;
+
+out_free_log:
+	kfree(sputrace_log);
+out:
+	return -ENOMEM;
+}
+
+static void __exit sputrace_exit(void)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(spu_probes); i++)
+		marker_probe_unregister(spu_probes[i].name,
+			spu_probes[i].probe_func, &spu_probes[i]);
+
+	remove_proc_entry("sputrace", NULL);
+	kfree(sputrace_log);
+}
+
+module_init(sputrace_init);
+module_exit(sputrace_exit);
+
+MODULE_LICENSE("GPL");
diff --git a/arch/powerpc/platforms/cell/spufs/syscalls.c b/arch/powerpc/platforms/cell/spufs/syscalls.c
index 430404413178..49c87769b1f8 100644
--- a/arch/powerpc/platforms/cell/spufs/syscalls.c
+++ b/arch/powerpc/platforms/cell/spufs/syscalls.c
@@ -73,7 +73,7 @@ static long do_spu_create(const char __user *pathname, unsigned int flags,
 				LOOKUP_OPEN|LOOKUP_CREATE, &nd);
 		if (!ret) {
 			ret = spufs_create(&nd, flags, mode, neighbor);
-			path_release(&nd);
+			path_put(&nd.path);
 		}
 		putname(tmp);
 	}