Merge branch 'master' into for-next

Fast-forward merge with Linus to be able to merge patches
based on more recent version of the tree.

35 files changed, 861 insertions, 315 deletions

diff --git a/arch/sparc/kernel/Makefile b/arch/sparc/kernel/Makefile
index b90b4a1d070..cb85458f89d 100644
--- a/arch/sparc/kernel/Makefile
+++ b/arch/sparc/kernel/Makefile
@@ -32,6 +32,7 @@ obj-$(CONFIG_SPARC32)   += sun4m_irq.o sun4c_irq.o sun4d_irq.o
 
 obj-y                   += process_$(BITS).o
 obj-y                   += signal_$(BITS).o
+obj-y                   += sigutil_$(BITS).o
 obj-$(CONFIG_SPARC32)   += ioport.o
 obj-y                   += setup_$(BITS).o
 obj-y                   += idprom.o
diff --git a/arch/sparc/kernel/cpu.c b/arch/sparc/kernel/cpu.c
index 138dbbc8dc8..9810fd88105 100644
--- a/arch/sparc/kernel/cpu.c
+++ b/arch/sparc/kernel/cpu.c
@@ -396,6 +396,7 @@ static int show_cpuinfo(struct seq_file *m, void *__unused)
 		   , cpu_data(0).clock_tick
 #endif
 		);
+	cpucap_info(m);
 #ifdef CONFIG_SMP
 	smp_bogo(m);
 #endif
@@ -474,11 +475,18 @@ static void __init sun4v_cpu_probe(void)
 		sparc_pmu_type = "niagara2";
 		break;
 
+	case SUN4V_CHIP_NIAGARA3:
+		sparc_cpu_type = "UltraSparc T3 (Niagara3)";
+		sparc_fpu_type = "UltraSparc T3 integrated FPU";
+		sparc_pmu_type = "niagara3";
+		break;
+
 	default:
 		printk(KERN_WARNING "CPU: Unknown sun4v cpu type [%s]\n",
 		       prom_cpu_compatible);
 		sparc_cpu_type = "Unknown SUN4V CPU";
 		sparc_fpu_type = "Unknown SUN4V FPU";
+		sparc_pmu_type = "Unknown SUN4V PMU";
 		break;
 	}
 }
diff --git a/arch/sparc/kernel/cpumap.c b/arch/sparc/kernel/cpumap.c
index d91fd782743..4197e8d62d4 100644
--- a/arch/sparc/kernel/cpumap.c
+++ b/arch/sparc/kernel/cpumap.c
@@ -324,6 +324,7 @@ static int iterate_cpu(struct cpuinfo_tree *t, unsigned int root_index)
 	switch (sun4v_chip_type) {
 	case SUN4V_CHIP_NIAGARA1:
 	case SUN4V_CHIP_NIAGARA2:
+	case SUN4V_CHIP_NIAGARA3:
 		rover_inc_table = niagara_iterate_method;
 		break;
 	default:
diff --git a/arch/sparc/kernel/ds.c b/arch/sparc/kernel/ds.c
index dd1342c0a3b..7429b47c3ac 100644
--- a/arch/sparc/kernel/ds.c
+++ b/arch/sparc/kernel/ds.c
@@ -15,12 +15,15 @@
 #include <linux/reboot.h>
 #include <linux/cpu.h>
 
+#include <asm/hypervisor.h>
 #include <asm/ldc.h>
 #include <asm/vio.h>
 #include <asm/mdesc.h>
 #include <asm/head.h>
 #include <asm/irq.h>
 
+#include "kernel.h"
+
 #define DRV_MODULE_NAME		"ds"
 #define PFX DRV_MODULE_NAME	": "
 #define DRV_MODULE_VERSION	"1.0"
@@ -828,18 +831,32 @@ void ldom_set_var(const char *var, const char *value)
 	}
 }
 
+static char full_boot_str[256] __attribute__((aligned(32)));
+static int reboot_data_supported;
+
 void ldom_reboot(const char *boot_command)
 {
 	/* Don't bother with any of this if the boot_command
 	 * is empty.
 	 */
 	if (boot_command && strlen(boot_command)) {
-		char full_boot_str[256];
+		unsigned long len;
 
 		strcpy(full_boot_str, "boot ");
 		strcpy(full_boot_str + strlen("boot "), boot_command);
+		len = strlen(full_boot_str);
 
-		ldom_set_var("reboot-command", full_boot_str);
+		if (reboot_data_supported) {
+			unsigned long ra = kimage_addr_to_ra(full_boot_str);
+			unsigned long hv_ret;
+
+			hv_ret = sun4v_reboot_data_set(ra, len);
+			if (hv_ret != HV_EOK)
+				pr_err("SUN4V: Unable to set reboot data "
+				       "hv_ret=%lu\n", hv_ret);
+		} else {
+			ldom_set_var("reboot-command", full_boot_str);
+		}
 	}
 	sun4v_mach_sir();
 }
@@ -1237,6 +1254,16 @@ static struct vio_driver ds_driver = {
 
 static int __init ds_init(void)
 {
+	unsigned long hv_ret, major, minor;
+
+	if (tlb_type == hypervisor) {
+		hv_ret = sun4v_get_version(HV_GRP_REBOOT_DATA, &major, &minor);
+		if (hv_ret == HV_EOK) {
+			pr_info("SUN4V: Reboot data supported (maj=%lu,min=%lu).\n",
+				major, minor);
+			reboot_data_supported = 1;
+		}
+	}
 	kthread_run(ds_thread, NULL, "kldomd");
 
 	return vio_register_driver(&ds_driver);
diff --git a/arch/sparc/kernel/entry.h b/arch/sparc/kernel/entry.h
index d1f1361c416..e27f8ea8656 100644
--- a/arch/sparc/kernel/entry.h
+++ b/arch/sparc/kernel/entry.h
@@ -42,6 +42,20 @@ extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
 extern void fpload(unsigned long *fpregs, unsigned long *fsr);
 
 #else /* CONFIG_SPARC32 */
+struct popc_3insn_patch_entry {
+	unsigned int	addr;
+	unsigned int	insns[3];
+};
+extern struct popc_3insn_patch_entry __popc_3insn_patch,
+	__popc_3insn_patch_end;
+
+struct popc_6insn_patch_entry {
+	unsigned int	addr;
+	unsigned int	insns[6];
+};
+extern struct popc_6insn_patch_entry __popc_6insn_patch,
+	__popc_6insn_patch_end;
+
 extern void __init per_cpu_patch(void);
 extern void __init sun4v_patch(void);
 extern void __init boot_cpu_id_too_large(int cpu);
diff --git a/arch/sparc/kernel/head_64.S b/arch/sparc/kernel/head_64.S
index aa594c792d1..0eac1b2fc53 100644
--- a/arch/sparc/kernel/head_64.S
+++ b/arch/sparc/kernel/head_64.S
@@ -132,6 +132,8 @@ prom_sun4v_name:
 	.asciz	"sun4v"
 prom_niagara_prefix:
 	.asciz	"SUNW,UltraSPARC-T"
+prom_sparc_prefix:
+	.asciz	"SPARC-T"
 	.align	4
 prom_root_compatible:
 	.skip	64
@@ -382,6 +384,22 @@ sun4v_chip_type:
 90:	ldub	[%g7], %g2
 	ldub	[%g1], %g4
 	cmp	%g2, %g4
+	bne,pn	%icc, 89f
+	 add	%g7, 1, %g7
+	subcc	%g3, 1, %g3
+	bne,pt	%xcc, 90b
+	 add	%g1, 1, %g1
+	ba,pt	%xcc, 91f
+	 nop
+
+89:	sethi	%hi(prom_cpu_compatible), %g1
+	or	%g1, %lo(prom_cpu_compatible), %g1
+	sethi	%hi(prom_sparc_prefix), %g7
+	or	%g7, %lo(prom_sparc_prefix), %g7
+	mov	7, %g3
+90:	ldub	[%g7], %g2
+	ldub	[%g1], %g4
+	cmp	%g2, %g4
 	bne,pn	%icc, 4f
 	 add	%g7, 1, %g7
 	subcc	%g3, 1, %g3
@@ -390,6 +408,15 @@ sun4v_chip_type:
 
 	sethi	%hi(prom_cpu_compatible), %g1
 	or	%g1, %lo(prom_cpu_compatible), %g1
+	ldub	[%g1 + 7], %g2
+	cmp	%g2, '3'
+	be,pt	%xcc, 5f
+	 mov	SUN4V_CHIP_NIAGARA3, %g4
+	ba,pt	%xcc, 4f
+	 nop
+
+91:	sethi	%hi(prom_cpu_compatible), %g1
+	or	%g1, %lo(prom_cpu_compatible), %g1
 	ldub	[%g1 + 17], %g2
 	cmp	%g2, '1'
 	be,pt	%xcc, 5f
@@ -397,6 +424,7 @@ sun4v_chip_type:
 	cmp	%g2, '2'
 	be,pt	%xcc, 5f
 	 mov	SUN4V_CHIP_NIAGARA2, %g4
+	
 4:
 	mov	SUN4V_CHIP_UNKNOWN, %g4
 5:	sethi	%hi(sun4v_chip_type), %g2
@@ -514,6 +542,9 @@ niagara_tlb_fixup:
 	 cmp	%g1, SUN4V_CHIP_NIAGARA2
 	be,pt	%xcc, niagara2_patch
 	 nop
+	cmp	%g1, SUN4V_CHIP_NIAGARA3
+	be,pt	%xcc, niagara2_patch
+	 nop
 
 	call	generic_patch_copyops
 	 nop
@@ -528,7 +559,7 @@ niagara2_patch:
 	 nop
 	call	niagara_patch_bzero
 	 nop
-	call	niagara2_patch_pageops
+	call	niagara_patch_pageops
 	 nop
 
 	ba,a,pt	%xcc, 80f
diff --git a/arch/sparc/kernel/hvapi.c b/arch/sparc/kernel/hvapi.c
index 7c60afb835b..c2d055d8ba9 100644
--- a/arch/sparc/kernel/hvapi.c
+++ b/arch/sparc/kernel/hvapi.c
@@ -28,16 +28,23 @@ static struct api_info api_table[] = {
 	{ .group = HV_GRP_CORE,		.flags = FLAG_PRE_API	},
 	{ .group = HV_GRP_INTR,					},
 	{ .group = HV_GRP_SOFT_STATE,				},
+	{ .group = HV_GRP_TM,					},
 	{ .group = HV_GRP_PCI,		.flags = FLAG_PRE_API	},
 	{ .group = HV_GRP_LDOM,					},
 	{ .group = HV_GRP_SVC_CHAN,	.flags = FLAG_PRE_API	},
 	{ .group = HV_GRP_NCS,		.flags = FLAG_PRE_API	},
 	{ .group = HV_GRP_RNG,					},
+	{ .group = HV_GRP_PBOOT,				},
+	{ .group = HV_GRP_TPM,					},
+	{ .group = HV_GRP_SDIO,					},
+	{ .group = HV_GRP_SDIO_ERR,				},
+	{ .group = HV_GRP_REBOOT_DATA,				},
 	{ .group = HV_GRP_NIAG_PERF,	.flags = FLAG_PRE_API	},
 	{ .group = HV_GRP_FIRE_PERF,				},
 	{ .group = HV_GRP_N2_CPU,				},
 	{ .group = HV_GRP_NIU,					},
 	{ .group = HV_GRP_VF_CPU,				},
+	{ .group = HV_GRP_KT_CPU,				},
 	{ .group = HV_GRP_DIAG,		.flags = FLAG_PRE_API	},
 };
 
diff --git a/arch/sparc/kernel/hvcalls.S b/arch/sparc/kernel/hvcalls.S
index 8a5f35ffb15..58d60de4d65 100644
--- a/arch/sparc/kernel/hvcalls.S
+++ b/arch/sparc/kernel/hvcalls.S
@@ -798,3 +798,10 @@ ENTRY(sun4v_niagara2_setperf)
 	retl
 	 nop
 ENDPROC(sun4v_niagara2_setperf)
+
+ENTRY(sun4v_reboot_data_set)
+	mov	HV_FAST_REBOOT_DATA_SET, %o5
+	ta	HV_FAST_TRAP
+	retl
+	 nop
+ENDPROC(sun4v_reboot_data_set)
diff --git a/arch/sparc/kernel/ioport.c b/arch/sparc/kernel/ioport.c
index 6ffccd6e015..d0479e2163f 100644
--- a/arch/sparc/kernel/ioport.c
+++ b/arch/sparc/kernel/ioport.c
@@ -65,9 +65,6 @@ static inline void dma_make_coherent(unsigned long pa, unsigned long len)
 }
 #endif
 
-static struct resource *_sparc_find_resource(struct resource *r,
-					     unsigned long);
-
 static void __iomem *_sparc_ioremap(struct resource *res, u32 bus, u32 pa, int sz);
 static void __iomem *_sparc_alloc_io(unsigned int busno, unsigned long phys,
     unsigned long size, char *name);
@@ -143,7 +140,11 @@ void iounmap(volatile void __iomem *virtual)
 	unsigned long vaddr = (unsigned long) virtual & PAGE_MASK;
 	struct resource *res;
 
-	if ((res = _sparc_find_resource(&sparc_iomap, vaddr)) == NULL) {
+	/*
+	 * XXX Too slow. Can have 8192 DVMA pages on sun4m in the worst case.
+	 * This probably warrants some sort of hashing.
+	*/
+	if ((res = lookup_resource(&sparc_iomap, vaddr)) == NULL) {
 		printk("free_io/iounmap: cannot free %lx\n", vaddr);
 		return;
 	}
@@ -319,7 +320,7 @@ static void sbus_free_coherent(struct device *dev, size_t n, void *p,
 	struct resource *res;
 	struct page *pgv;
 
-	if ((res = _sparc_find_resource(&_sparc_dvma,
+	if ((res = lookup_resource(&_sparc_dvma,
 	    (unsigned long)p)) == NULL) {
 		printk("sbus_free_consistent: cannot free %p\n", p);
 		return;
@@ -492,7 +493,7 @@ static void pci32_free_coherent(struct device *dev, size_t n, void *p,
 {
 	struct resource *res;
 
-	if ((res = _sparc_find_resource(&_sparc_dvma,
+	if ((res = lookup_resource(&_sparc_dvma,
 	    (unsigned long)p)) == NULL) {
 		printk("pci_free_consistent: cannot free %p\n", p);
 		return;
@@ -715,25 +716,6 @@ static const struct file_operations sparc_io_proc_fops = {
 };
 #endif /* CONFIG_PROC_FS */
 
-/*
- * This is a version of find_resource and it belongs to kernel/resource.c.
- * Until we have agreement with Linus and Martin, it lingers here.
- *
- * XXX Too slow. Can have 8192 DVMA pages on sun4m in the worst case.
- * This probably warrants some sort of hashing.
- */
-static struct resource *_sparc_find_resource(struct resource *root,
-					     unsigned long hit)
-{
-	struct resource *tmp;
-
-	for (tmp = root->child; tmp != 0; tmp = tmp->sibling) {
-		if (tmp->start <= hit && tmp->end >= hit)
-			return tmp;
-	}
-	return NULL;
-}
-
 static void register_proc_sparc_ioport(void)
 {
 #ifdef CONFIG_PROC_FS
diff --git a/arch/sparc/kernel/irq.h b/arch/sparc/kernel/irq.h
index 100b9c204e7..42851122bbd 100644
--- a/arch/sparc/kernel/irq.h
+++ b/arch/sparc/kernel/irq.h
@@ -88,7 +88,7 @@ BTFIXUPDEF_CALL(void, set_irq_udt, int)
 #define set_irq_udt(cpu) BTFIXUP_CALL(set_irq_udt)(cpu)
 
 /* All SUN4D IPIs are sent on this IRQ, may be shared with hard IRQs */
-#define SUN4D_IPI_IRQ 14
+#define SUN4D_IPI_IRQ 13
 
 extern void sun4d_ipi_interrupt(void);
 
diff --git a/arch/sparc/kernel/irq_64.c b/arch/sparc/kernel/irq_64.c
index 4e78862d12f..0dd8422a469 100644
--- a/arch/sparc/kernel/irq_64.c
+++ b/arch/sparc/kernel/irq_64.c
@@ -26,7 +26,7 @@
 
 #include <asm/ptrace.h>
 #include <asm/processor.h>
-#include <asm/atomic.h>
+#include <linux/atomic.h>
 #include <asm/system.h>
 #include <asm/irq.h>
 #include <asm/io.h>
diff --git a/arch/sparc/kernel/kernel.h b/arch/sparc/kernel/kernel.h
index 6f6544cfa0e..fd6c36b1df7 100644
--- a/arch/sparc/kernel/kernel.h
+++ b/arch/sparc/kernel/kernel.h
@@ -4,12 +4,27 @@
 #include <linux/interrupt.h>
 
 #include <asm/traps.h>
+#include <asm/head.h>
+#include <asm/io.h>
 
 /* cpu.c */
 extern const char *sparc_pmu_type;
 extern unsigned int fsr_storage;
 extern int ncpus_probed;
 
+#ifdef CONFIG_SPARC64
+/* setup_64.c */
+struct seq_file;
+extern void cpucap_info(struct seq_file *);
+
+static inline unsigned long kimage_addr_to_ra(const char *p)
+{
+	unsigned long val = (unsigned long) p;
+
+	return kern_base + (val - KERNBASE);
+}
+#endif
+
 #ifdef CONFIG_SPARC32
 /* cpu.c */
 extern void cpu_probe(void);
diff --git a/arch/sparc/kernel/ktlb.S b/arch/sparc/kernel/ktlb.S
index 1d361477d7d..79f31036484 100644
--- a/arch/sparc/kernel/ktlb.S
+++ b/arch/sparc/kernel/ktlb.S
@@ -47,16 +47,16 @@ kvmap_itlb_tsb_miss:
 kvmap_itlb_vmalloc_addr:
 	KERN_PGTABLE_WALK(%g4, %g5, %g2, kvmap_itlb_longpath)
 
-	KTSB_LOCK_TAG(%g1, %g2, %g7)
+	TSB_LOCK_TAG(%g1, %g2, %g7)
 
 	/* Load and check PTE.  */
 	ldxa		[%g5] ASI_PHYS_USE_EC, %g5
 	mov		1, %g7
 	sllx		%g7, TSB_TAG_INVALID_BIT, %g7
 	brgez,a,pn	%g5, kvmap_itlb_longpath
-	 KTSB_STORE(%g1, %g7)
+	 TSB_STORE(%g1, %g7)
 
-	KTSB_WRITE(%g1, %g5, %g6)
+	TSB_WRITE(%g1, %g5, %g6)
 
 	/* fallthrough to TLB load */
 
@@ -102,9 +102,9 @@ kvmap_itlb_longpath:
 kvmap_itlb_obp:
 	OBP_TRANS_LOOKUP(%g4, %g5, %g2, %g3, kvmap_itlb_longpath)
 
-	KTSB_LOCK_TAG(%g1, %g2, %g7)
+	TSB_LOCK_TAG(%g1, %g2, %g7)
 
-	KTSB_WRITE(%g1, %g5, %g6)
+	TSB_WRITE(%g1, %g5, %g6)
 
 	ba,pt		%xcc, kvmap_itlb_load
 	 nop
@@ -112,17 +112,17 @@ kvmap_itlb_obp:
 kvmap_dtlb_obp:
 	OBP_TRANS_LOOKUP(%g4, %g5, %g2, %g3, kvmap_dtlb_longpath)
 
-	KTSB_LOCK_TAG(%g1, %g2, %g7)
+	TSB_LOCK_TAG(%g1, %g2, %g7)
 
-	KTSB_WRITE(%g1, %g5, %g6)
+	TSB_WRITE(%g1, %g5, %g6)
 
 	ba,pt		%xcc, kvmap_dtlb_load
 	 nop
 
 	.align		32
 kvmap_dtlb_tsb4m_load:
-	KTSB_LOCK_TAG(%g1, %g2, %g7)
-	KTSB_WRITE(%g1, %g5, %g6)
+	TSB_LOCK_TAG(%g1, %g2, %g7)
+	TSB_WRITE(%g1, %g5, %g6)
 	ba,pt		%xcc, kvmap_dtlb_load
 	 nop
 
@@ -222,16 +222,16 @@ kvmap_linear_patch:
 kvmap_dtlb_vmalloc_addr:
 	KERN_PGTABLE_WALK(%g4, %g5, %g2, kvmap_dtlb_longpath)
 
-	KTSB_LOCK_TAG(%g1, %g2, %g7)
+	TSB_LOCK_TAG(%g1, %g2, %g7)
 
 	/* Load and check PTE.  */
 	ldxa		[%g5] ASI_PHYS_USE_EC, %g5
 	mov		1, %g7
 	sllx		%g7, TSB_TAG_INVALID_BIT, %g7
 	brgez,a,pn	%g5, kvmap_dtlb_longpath
-	 KTSB_STORE(%g1, %g7)
+	 TSB_STORE(%g1, %g7)
 
-	KTSB_WRITE(%g1, %g5, %g6)
+	TSB_WRITE(%g1, %g5, %g6)
 
 	/* fallthrough to TLB load */
 
diff --git a/arch/sparc/kernel/leon_pci_grpci2.c b/arch/sparc/kernel/leon_pci_grpci2.c
index 44dc093ee33..fad1bd07cb5 100644
--- a/arch/sparc/kernel/leon_pci_grpci2.c
+++ b/arch/sparc/kernel/leon_pci_grpci2.c
@@ -215,7 +215,7 @@ struct grpci2_priv {
 DEFINE_SPINLOCK(grpci2_dev_lock);
 struct grpci2_priv *grpci2priv;
 
-int grpci2_map_irq(struct pci_dev *dev, u8 slot, u8 pin)
+int grpci2_map_irq(const struct pci_dev *dev, u8 slot, u8 pin)
 {
 	struct grpci2_priv *priv = dev->bus->sysdata;
 	int irq_group;
diff --git a/arch/sparc/kernel/leon_smp.c b/arch/sparc/kernel/leon_smp.c
index fe8fb44c609..1210fde1874 100644
--- a/arch/sparc/kernel/leon_smp.c
+++ b/arch/sparc/kernel/leon_smp.c
@@ -28,7 +28,7 @@
 #include <asm/tlbflush.h>
 
 #include <asm/ptrace.h>
-#include <asm/atomic.h>
+#include <linux/atomic.h>
 #include <asm/irq_regs.h>
 #include <asm/traps.h>
 
diff --git a/arch/sparc/kernel/mdesc.c b/arch/sparc/kernel/mdesc.c
index 42f28c7420e..acaebb63c4f 100644
--- a/arch/sparc/kernel/mdesc.c
+++ b/arch/sparc/kernel/mdesc.c
@@ -508,6 +508,8 @@ const char *mdesc_node_name(struct mdesc_handle *hp, u64 node)
 }
 EXPORT_SYMBOL(mdesc_node_name);
 
+static u64 max_cpus = 64;
+
 static void __init report_platform_properties(void)
 {
 	struct mdesc_handle *hp = mdesc_grab();
@@ -543,8 +545,10 @@ static void __init report_platform_properties(void)
 	if (v)
 		printk("PLATFORM: watchdog-max-timeout [%llu ms]\n", *v);
 	v = mdesc_get_property(hp, pn, "max-cpus", NULL);
-	if (v)
-		printk("PLATFORM: max-cpus [%llu]\n", *v);
+	if (v) {
+		max_cpus = *v;
+		printk("PLATFORM: max-cpus [%llu]\n", max_cpus);
+	}
 
 #ifdef CONFIG_SMP
 	{
@@ -715,7 +719,7 @@ static void __cpuinit set_proc_ids(struct mdesc_handle *hp)
 }
 
 static void __cpuinit get_one_mondo_bits(const u64 *p, unsigned int *mask,
-					 unsigned char def)
+					 unsigned long def, unsigned long max)
 {
 	u64 val;
 
@@ -726,6 +730,9 @@ static void __cpuinit get_one_mondo_bits(const u64 *p, unsigned int *mask,
 	if (!val || val >= 64)
 		goto use_default;
 
+	if (val > max)
+		val = max;
+
 	*mask = ((1U << val) * 64U) - 1U;
 	return;
 
@@ -736,19 +743,28 @@ use_default:
 static void __cpuinit get_mondo_data(struct mdesc_handle *hp, u64 mp,
 				     struct trap_per_cpu *tb)
 {
+	static int printed;
 	const u64 *val;
 
 	val = mdesc_get_property(hp, mp, "q-cpu-mondo-#bits", NULL);
-	get_one_mondo_bits(val, &tb->cpu_mondo_qmask, 7);
+	get_one_mondo_bits(val, &tb->cpu_mondo_qmask, 7, ilog2(max_cpus * 2));
 
 	val = mdesc_get_property(hp, mp, "q-dev-mondo-#bits", NULL);
-	get_one_mondo_bits(val, &tb->dev_mondo_qmask, 7);
+	get_one_mondo_bits(val, &tb->dev_mondo_qmask, 7, 8);
 
 	val = mdesc_get_property(hp, mp, "q-resumable-#bits", NULL);
-	get_one_mondo_bits(val, &tb->resum_qmask, 6);
+	get_one_mondo_bits(val, &tb->resum_qmask, 6, 7);
 
 	val = mdesc_get_property(hp, mp, "q-nonresumable-#bits", NULL);
-	get_one_mondo_bits(val, &tb->nonresum_qmask, 2);
+	get_one_mondo_bits(val, &tb->nonresum_qmask, 2, 2);
+	if (!printed++) {
+		pr_info("SUN4V: Mondo queue sizes "
+			"[cpu(%u) dev(%u) r(%u) nr(%u)]\n",
+			tb->cpu_mondo_qmask + 1,
+			tb->dev_mondo_qmask + 1,
+			tb->resum_qmask + 1,
+			tb->nonresum_qmask + 1);
+	}
 }
 
 static void * __cpuinit mdesc_iterate_over_cpus(void *(*func)(struct mdesc_handle *, u64, int, void *), void *arg, cpumask_t *mask)
diff --git a/arch/sparc/kernel/pcic.c b/arch/sparc/kernel/pcic.c
index a19f0419547..1aaf8c180be 100644
--- a/arch/sparc/kernel/pcic.c
+++ b/arch/sparc/kernel/pcic.c
@@ -352,8 +352,8 @@ int __init pcic_probe(void)
 	strcpy(pbm->prom_name, namebuf);
 
 	{
-		extern volatile int t_nmi[1];
-		extern int pcic_nmi_trap_patch[1];
+		extern volatile int t_nmi[4];
+		extern int pcic_nmi_trap_patch[4];
 
 		t_nmi[0] = pcic_nmi_trap_patch[0];
 		t_nmi[1] = pcic_nmi_trap_patch[1];
diff --git a/arch/sparc/kernel/pcr.c b/arch/sparc/kernel/pcr.c
index 8ac23e66008..343b0f9e2e7 100644
--- a/arch/sparc/kernel/pcr.c
+++ b/arch/sparc/kernel/pcr.c
@@ -80,8 +80,11 @@ static void n2_pcr_write(u64 val)
 {
 	unsigned long ret;
 
-	ret = sun4v_niagara2_setperf(HV_N2_PERF_SPARC_CTL, val);
-	if (ret != HV_EOK)
+	if (val & PCR_N2_HTRACE) {
+		ret = sun4v_niagara2_setperf(HV_N2_PERF_SPARC_CTL, val);
+		if (ret != HV_EOK)
+			write_pcr(val);
+	} else
 		write_pcr(val);
 }
 
@@ -106,6 +109,10 @@ static int __init register_perf_hsvc(void)
 			perf_hsvc_group = HV_GRP_N2_CPU;
 			break;
 
+		case SUN4V_CHIP_NIAGARA3:
+			perf_hsvc_group = HV_GRP_KT_CPU;
+			break;
+
 		default:
 			return -ENODEV;
 		}
diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c
index 62a034318b1..614da624330 100644
--- a/arch/sparc/kernel/perf_event.c
+++ b/arch/sparc/kernel/perf_event.c
@@ -22,7 +22,7 @@
 #include <asm/stacktrace.h>
 #include <asm/cpudata.h>
 #include <asm/uaccess.h>
-#include <asm/atomic.h>
+#include <linux/atomic.h>
 #include <asm/nmi.h>
 #include <asm/pcr.h>
 
@@ -1343,7 +1343,8 @@ static bool __init supported_pmu(void)
 		sparc_pmu = &niagara1_pmu;
 		return true;
 	}
-	if (!strcmp(sparc_pmu_type, "niagara2")) {
+	if (!strcmp(sparc_pmu_type, "niagara2") ||
+	    !strcmp(sparc_pmu_type, "niagara3")) {
 		sparc_pmu = &niagara2_pmu;
 		return true;
 	}
diff --git a/arch/sparc/kernel/setup_64.c b/arch/sparc/kernel/setup_64.c
index c4dd0999da8..3c5bb784214 100644
--- a/arch/sparc/kernel/setup_64.c
+++ b/arch/sparc/kernel/setup_64.c
@@ -29,6 +29,7 @@
 #include <linux/interrupt.h>
 #include <linux/cpu.h>
 #include <linux/initrd.h>
+#include <linux/module.h>
 
 #include <asm/system.h>
 #include <asm/io.h>
@@ -46,6 +47,8 @@
 #include <asm/mmu.h>
 #include <asm/ns87303.h>
 #include <asm/btext.h>
+#include <asm/elf.h>
+#include <asm/mdesc.h>
 
 #ifdef CONFIG_IP_PNP
 #include <net/ipconfig.h>
@@ -269,6 +272,40 @@ void __init sun4v_patch(void)
 	sun4v_hvapi_init();
 }
 
+static void __init popc_patch(void)
+{
+	struct popc_3insn_patch_entry *p3;
+	struct popc_6insn_patch_entry *p6;
+
+	p3 = &__popc_3insn_patch;
+	while (p3 < &__popc_3insn_patch_end) {
+		unsigned long i, addr = p3->addr;
+
+		for (i = 0; i < 3; i++) {
+			*(unsigned int *) (addr +  (i * 4)) = p3->insns[i];
+			wmb();
+			__asm__ __volatile__("flush	%0"
+					     : : "r" (addr +  (i * 4)));
+		}
+
+		p3++;
+	}
+
+	p6 = &__popc_6insn_patch;
+	while (p6 < &__popc_6insn_patch_end) {
+		unsigned long i, addr = p6->addr;
+
+		for (i = 0; i < 6; i++) {
+			*(unsigned int *) (addr +  (i * 4)) = p6->insns[i];
+			wmb();
+			__asm__ __volatile__("flush	%0"
+					     : : "r" (addr +  (i * 4)));
+		}
+
+		p6++;
+	}
+}
+
 #ifdef CONFIG_SMP
 void __init boot_cpu_id_too_large(int cpu)
 {
@@ -278,6 +315,160 @@ void __init boot_cpu_id_too_large(int cpu)
 }
 #endif
 
+/* On Ultra, we support all of the v8 capabilities. */
+unsigned long sparc64_elf_hwcap = (HWCAP_SPARC_FLUSH | HWCAP_SPARC_STBAR |
+				   HWCAP_SPARC_SWAP | HWCAP_SPARC_MULDIV |
+				   HWCAP_SPARC_V9);
+EXPORT_SYMBOL(sparc64_elf_hwcap);
+
+static const char *hwcaps[] = {
+	"flush", "stbar", "swap", "muldiv", "v9",
+	"ultra3", "blkinit", "n2",
+
+	/* These strings are as they appear in the machine description
+	 * 'hwcap-list' property for cpu nodes.
+	 */
+	"mul32", "div32", "fsmuld", "v8plus", "popc", "vis", "vis2",
+	"ASIBlkInit", "fmaf", "vis3", "hpc", "random", "trans", "fjfmau",
+	"ima", "cspare",
+};
+
+void cpucap_info(struct seq_file *m)
+{
+	unsigned long caps = sparc64_elf_hwcap;
+	int i, printed = 0;
+
+	seq_puts(m, "cpucaps\t\t: ");
+	for (i = 0; i < ARRAY_SIZE(hwcaps); i++) {
+		unsigned long bit = 1UL << i;
+		if (caps & bit) {
+			seq_printf(m, "%s%s",
+				   printed ? "," : "", hwcaps[i]);
+			printed++;
+		}
+	}
+	seq_putc(m, '\n');
+}
+
+static void __init report_hwcaps(unsigned long caps)
+{
+	int i, printed = 0;
+
+	printk(KERN_INFO "CPU CAPS: [");
+	for (i = 0; i < ARRAY_SIZE(hwcaps); i++) {
+		unsigned long bit = 1UL << i;
+		if (caps & bit) {
+			printk(KERN_CONT "%s%s",
+			       printed ? "," : "", hwcaps[i]);
+			if (++printed == 8) {
+				printk(KERN_CONT "]\n");
+				printk(KERN_INFO "CPU CAPS: [");
+				printed = 0;
+			}
+		}
+	}
+	printk(KERN_CONT "]\n");
+}
+
+static unsigned long __init mdesc_cpu_hwcap_list(void)
+{
+	struct mdesc_handle *hp;
+	unsigned long caps = 0;
+	const char *prop;
+	int len;
+	u64 pn;
+
+	hp = mdesc_grab();
+	if (!hp)
+		return 0;
+
+	pn = mdesc_node_by_name(hp, MDESC_NODE_NULL, "cpu");
+	if (pn == MDESC_NODE_NULL)
+		goto out;
+
+	prop = mdesc_get_property(hp, pn, "hwcap-list", &len);
+	if (!prop)
+		goto out;
+
+	while (len) {
+		int i, plen;
+
+		for (i = 0; i < ARRAY_SIZE(hwcaps); i++) {
+			unsigned long bit = 1UL << i;
+
+			if (!strcmp(prop, hwcaps[i])) {
+				caps |= bit;
+				break;
+			}
+		}
+
+		plen = strlen(prop) + 1;
+		prop += plen;
+		len -= plen;
+	}
+
+out:
+	mdesc_release(hp);
+	return caps;
+}
+
+/* This yields a mask that user programs can use to figure out what
+ * instruction set this cpu supports.
+ */
+static void __init init_sparc64_elf_hwcap(void)
+{
+	unsigned long cap = sparc64_elf_hwcap;
+	unsigned long mdesc_caps;
+
+	if (tlb_type == cheetah || tlb_type == cheetah_plus)
+		cap |= HWCAP_SPARC_ULTRA3;
+	else if (tlb_type == hypervisor) {
+		if (sun4v_chip_type == SUN4V_CHIP_NIAGARA1 ||
+		    sun4v_chip_type == SUN4V_CHIP_NIAGARA2 ||
+		    sun4v_chip_type == SUN4V_CHIP_NIAGARA3)
+			cap |= HWCAP_SPARC_BLKINIT;
+		if (sun4v_chip_type == SUN4V_CHIP_NIAGARA2 ||
+		    sun4v_chip_type == SUN4V_CHIP_NIAGARA3)
+			cap |= HWCAP_SPARC_N2;
+	}
+
+	cap |= (AV_SPARC_MUL32 | AV_SPARC_DIV32 | AV_SPARC_V8PLUS);
+
+	mdesc_caps = mdesc_cpu_hwcap_list();
+	if (!mdesc_caps) {
+		if (tlb_type == spitfire)
+			cap |= AV_SPARC_VIS;
+		if (tlb_type == cheetah || tlb_type == cheetah_plus)
+			cap |= AV_SPARC_VIS | AV_SPARC_VIS2;
+		if (tlb_type == cheetah_plus) {
+			unsigned long impl, ver;
+
+			__asm__ __volatile__("rdpr %%ver, %0" : "=r" (ver));
+			impl = ((ver >> 32) & 0xffff);
+			if (impl == PANTHER_IMPL)
+				cap |= AV_SPARC_POPC;
+		}
+		if (tlb_type == hypervisor) {
+			if (sun4v_chip_type == SUN4V_CHIP_NIAGARA1)
+				cap |= AV_SPARC_ASI_BLK_INIT;
+			if (sun4v_chip_type == SUN4V_CHIP_NIAGARA2 ||
+			    sun4v_chip_type == SUN4V_CHIP_NIAGARA3)
+				cap |= (AV_SPARC_VIS | AV_SPARC_VIS2 |
+					AV_SPARC_ASI_BLK_INIT |
+					AV_SPARC_POPC);
+			if (sun4v_chip_type == SUN4V_CHIP_NIAGARA3)
+				cap |= (AV_SPARC_VIS3 | AV_SPARC_HPC |
+					AV_SPARC_FMAF);
+		}
+	}
+	sparc64_elf_hwcap = cap | mdesc_caps;
+
+	report_hwcaps(sparc64_elf_hwcap);
+
+	if (sparc64_elf_hwcap & AV_SPARC_POPC)
+		popc_patch();
+}
+
 void __init setup_arch(char **cmdline_p)
 {
 	/* Initialize PROM console and command line. */
@@ -337,6 +528,7 @@ void __init setup_arch(char **cmdline_p)
 	init_cur_cpu_trap(current_thread_info());
 
 	paging_init();
+	init_sparc64_elf_hwcap();
 }
 
 extern int stop_a_enabled;
diff --git a/arch/sparc/kernel/signal32.c b/arch/sparc/kernel/signal32.c
index 75fad425e24..1ba95aff5d5 100644
--- a/arch/sparc/kernel/signal32.c
+++ b/arch/sparc/kernel/signal32.c
@@ -29,6 +29,8 @@
 #include <asm/visasm.h>
 #include <asm/compat_signal.h>
 
+#include "sigutil.h"
+
 #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
 
 /* This magic should be in g_upper[0] for all upper parts
@@ -44,14 +46,14 @@ typedef struct {
 struct signal_frame32 {
 	struct sparc_stackf32	ss;
 	__siginfo32_t		info;
-	/* __siginfo_fpu32_t * */ u32 fpu_save;
+	/* __siginfo_fpu_t * */ u32 fpu_save;
 	unsigned int		insns[2];
 	unsigned int		extramask[_COMPAT_NSIG_WORDS - 1];
 	unsigned int		extra_size; /* Should be sizeof(siginfo_extra_v8plus_t) */
 	/* Only valid if (info.si_regs.psr & (PSR_VERS|PSR_IMPL)) == PSR_V8PLUS */
 	siginfo_extra_v8plus_t	v8plus;
-	__siginfo_fpu_t		fpu_state;
-};
+	/* __siginfo_rwin_t * */u32 rwin_save;
+} __attribute__((aligned(8)));
 
 typedef struct compat_siginfo{
 	int si_signo;
@@ -110,18 +112,14 @@ struct rt_signal_frame32 {
 	compat_siginfo_t	info;
 	struct pt_regs32	regs;
 	compat_sigset_t		mask;
-	/* __siginfo_fpu32_t * */ u32 fpu_save;
+	/* __siginfo_fpu_t * */ u32 fpu_save;
 	unsigned int		insns[2];
 	stack_t32		stack;
 	unsigned int		extra_size; /* Should be sizeof(siginfo_extra_v8plus_t) */
 	/* Only valid if (regs.psr & (PSR_VERS|PSR_IMPL)) == PSR_V8PLUS */
 	siginfo_extra_v8plus_t	v8plus;
-	__siginfo_fpu_t		fpu_state;
-};
-
-/* Align macros */
-#define SF_ALIGNEDSZ  (((sizeof(struct signal_frame32) + 15) & (~15)))
-#define RT_ALIGNEDSZ  (((sizeof(struct rt_signal_frame32) + 15) & (~15)))
+	/* __siginfo_rwin_t * */u32 rwin_save;
+} __attribute__((aligned(8)));
 
 int copy_siginfo_to_user32(compat_siginfo_t __user *to, siginfo_t *from)
 {
@@ -192,30 +190,13 @@ int copy_siginfo_from_user32(siginfo_t *to, compat_siginfo_t __user *from)
 	return 0;
 }
 
-static int restore_fpu_state32(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
-{
-	unsigned long *fpregs = current_thread_info()->fpregs;
-	unsigned long fprs;
-	int err;
-	
-	err = __get_user(fprs, &fpu->si_fprs);
-	fprs_write(0);
-	regs->tstate &= ~TSTATE_PEF;
-	if (fprs & FPRS_DL)
-		err |= copy_from_user(fpregs, &fpu->si_float_regs[0], (sizeof(unsigned int) * 32));
-	if (fprs & FPRS_DU)
-		err |= copy_from_user(fpregs+16, &fpu->si_float_regs[32], (sizeof(unsigned int) * 32));
-	err |= __get_user(current_thread_info()->xfsr[0], &fpu->si_fsr);
-	err |= __get_user(current_thread_info()->gsr[0], &fpu->si_gsr);
-	current_thread_info()->fpsaved[0] |= fprs;
-	return err;
-}
-
 void do_sigreturn32(struct pt_regs *regs)
 {
 	struct signal_frame32 __user *sf;
+	compat_uptr_t fpu_save;
+	compat_uptr_t rwin_save;
 	unsigned int psr;
-	unsigned pc, npc, fpu_save;
+	unsigned pc, npc;
 	sigset_t set;
 	unsigned seta[_COMPAT_NSIG_WORDS];
 	int err, i;
@@ -273,8 +254,13 @@ void do_sigreturn32(struct pt_regs *regs)
 	pt_regs_clear_syscall(regs);
 
 	err |= __get_user(fpu_save, &sf->fpu_save);
-	if (fpu_save)
-		err |= restore_fpu_state32(regs, &sf->fpu_state);
+	if (!err && fpu_save)
+		err |= restore_fpu_state(regs, compat_ptr(fpu_save));
+	err |= __get_user(rwin_save, &sf->rwin_save);
+	if (!err && rwin_save) {
+		if (restore_rwin_state(compat_ptr(rwin_save)))
+			goto segv;
+	}
 	err |= __get_user(seta[0], &sf->info.si_mask);
 	err |= copy_from_user(seta+1, &sf->extramask,
 			      (_COMPAT_NSIG_WORDS - 1) * sizeof(unsigned int));
@@ -300,7 +286,9 @@ segv:
 asmlinkage void do_rt_sigreturn32(struct pt_regs *regs)
 {
 	struct rt_signal_frame32 __user *sf;
-	unsigned int psr, pc, npc, fpu_save, u_ss_sp;
+	unsigned int psr, pc, npc, u_ss_sp;
+	compat_uptr_t fpu_save;
+	compat_uptr_t rwin_save;
 	mm_segment_t old_fs;
 	sigset_t set;
 	compat_sigset_t seta;
@@ -359,8 +347,8 @@ asmlinkage void do_rt_sigreturn32(struct pt_regs *regs)
 	pt_regs_clear_syscall(regs);
 
 	err |= __get_user(fpu_save, &sf->fpu_save);
-	if (fpu_save)
-		err |= restore_fpu_state32(regs, &sf->fpu_state);
+	if (!err && fpu_save)
+		err |= restore_fpu_state(regs, compat_ptr(fpu_save));
 	err |= copy_from_user(&seta, &sf->mask, sizeof(compat_sigset_t));
 	err |= __get_user(u_ss_sp, &sf->stack.ss_sp);
 	st.ss_sp = compat_ptr(u_ss_sp);
@@ -376,6 +364,12 @@ asmlinkage void do_rt_sigreturn32(struct pt_regs *regs)
 	do_sigaltstack((stack_t __user *) &st, NULL, (unsigned long)sf);
 	set_fs(old_fs);
 	
+	err |= __get_user(rwin_save, &sf->rwin_save);
+	if (!err && rwin_save) {
+		if (restore_rwin_state(compat_ptr(rwin_save)))
+			goto segv;
+	}
+
 	switch (_NSIG_WORDS) {
 		case 4: set.sig[3] = seta.sig[6] + (((long)seta.sig[7]) << 32);
 		case 3: set.sig[2] = seta.sig[4] + (((long)seta.sig[5]) << 32);
@@ -433,26 +427,6 @@ static void __user *get_sigframe(struct sigaction *sa, struct pt_regs *regs, uns
 	return (void __user *) sp;
 }
 
-static int save_fpu_state32(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
-{
-	unsigned long *fpregs = current_thread_info()->fpregs;
-	unsigned long fprs;
-	int err = 0;
-	
-	fprs = current_thread_info()->fpsaved[0];
-	if (fprs & FPRS_DL)
-		err |= copy_to_user(&fpu->si_float_regs[0], fpregs,
-				    (sizeof(unsigned int) * 32));
-	if (fprs & FPRS_DU)
-		err |= copy_to_user(&fpu->si_float_regs[32], fpregs+16,
-				    (sizeof(unsigned int) * 32));
-	err |= __put_user(current_thread_info()->xfsr[0], &fpu->si_fsr);
-	err |= __put_user(current_thread_info()->gsr[0], &fpu->si_gsr);
-	err |= __put_user(fprs, &fpu->si_fprs);
-
-	return err;
-}
-
 /* The I-cache flush instruction only works in the primary ASI, which
  * right now is the nucleus, aka. kernel space.
  *
@@ -515,18 +489,23 @@ static int setup_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 			 int signo, sigset_t *oldset)
 {
 	struct signal_frame32 __user *sf;
+	int i, err, wsaved;
+	void __user *tail;
 	int sigframe_size;
 	u32 psr;
-	int i, err;
 	unsigned int seta[_COMPAT_NSIG_WORDS];
 
 	/* 1. Make sure everything is clean */
 	synchronize_user_stack();
 	save_and_clear_fpu();
 	
-	sigframe_size = SF_ALIGNEDSZ;
-	if (!(current_thread_info()->fpsaved[0] & FPRS_FEF))
-		sigframe_size -= sizeof(__siginfo_fpu_t);
+	wsaved = get_thread_wsaved();
+
+	sigframe_size = sizeof(*sf);
+	if (current_thread_info()->fpsaved[0] & FPRS_FEF)
+		sigframe_size += sizeof(__siginfo_fpu_t);
+	if (wsaved)
+		sigframe_size += sizeof(__siginfo_rwin_t);
 
 	sf = (struct signal_frame32 __user *)
 		get_sigframe(&ka->sa, regs, sigframe_size);
@@ -534,8 +513,7 @@ static int setup_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 	if (invalid_frame_pointer(sf, sigframe_size))
 		goto sigill;
 
-	if (get_thread_wsaved() != 0)
-		goto sigill;
+	tail = (sf + 1);
 
 	/* 2. Save the current process state */
 	if (test_thread_flag(TIF_32BIT)) {
@@ -560,11 +538,22 @@ static int setup_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 			  &sf->v8plus.asi);
 
 	if (psr & PSR_EF) {
-		err |= save_fpu_state32(regs, &sf->fpu_state);
-		err |= __put_user((u64)&sf->fpu_state, &sf->fpu_save);
+		__siginfo_fpu_t __user *fp = tail;
+		tail += sizeof(*fp);
+		err |= save_fpu_state(regs, fp);
+		err |= __put_user((u64)fp, &sf->fpu_save);
 	} else {
 		err |= __put_user(0, &sf->fpu_save);
 	}
+	if (wsaved) {
+		__siginfo_rwin_t __user *rwp = tail;
+		tail += sizeof(*rwp);
+		err |= save_rwin_state(wsaved, rwp);
+		err |= __put_user((u64)rwp, &sf->rwin_save);
+		set_thread_wsaved(0);
+	} else {
+		err |= __put_user(0, &sf->rwin_save);
+	}
 
 	switch (_NSIG_WORDS) {
 	case 4: seta[7] = (oldset->sig[3] >> 32);
@@ -580,10 +569,21 @@ static int setup_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 	err |= __copy_to_user(sf->extramask, seta + 1,
 			      (_COMPAT_NSIG_WORDS - 1) * sizeof(unsigned int));
 
-	err |= copy_in_user((u32 __user *)sf,
-			    (u32 __user *)(regs->u_regs[UREG_FP]),
-			    sizeof(struct reg_window32));
-	
+	if (!wsaved) {
+		err |= copy_in_user((u32 __user *)sf,
+				    (u32 __user *)(regs->u_regs[UREG_FP]),
+				    sizeof(struct reg_window32));
+	} else {
+		struct reg_window *rp;
+
+		rp = &current_thread_info()->reg_window[wsaved - 1];
+		for (i = 0; i < 8; i++)
+			err |= __put_user(rp->locals[i], &sf->ss.locals[i]);
+		for (i = 0; i < 6; i++)
+			err |= __put_user(rp->ins[i], &sf->ss.ins[i]);
+		err |= __put_user(rp->ins[6], &sf->ss.fp);
+		err |= __put_user(rp->ins[7], &sf->ss.callers_pc);
+	}	
 	if (err)
 		goto sigsegv;
 
@@ -613,7 +613,6 @@ static int setup_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 		err |= __put_user(0x91d02010, &sf->insns[1]); /*t 0x10*/
 		if (err)
 			goto sigsegv;
-
 		flush_signal_insns(address);
 	}
 	return 0;
@@ -632,18 +631,23 @@ static int setup_rt_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 			    siginfo_t *info)
 {
 	struct rt_signal_frame32 __user *sf;
+	int i, err, wsaved;
+	void __user *tail;
 	int sigframe_size;
 	u32 psr;
-	int i, err;
 	compat_sigset_t seta;
 
 	/* 1. Make sure everything is clean */
 	synchronize_user_stack();
 	save_and_clear_fpu();
 	
-	sigframe_size = RT_ALIGNEDSZ;
-	if (!(current_thread_info()->fpsaved[0] & FPRS_FEF))
-		sigframe_size -= sizeof(__siginfo_fpu_t);
+	wsaved = get_thread_wsaved();
+
+	sigframe_size = sizeof(*sf);
+	if (current_thread_info()->fpsaved[0] & FPRS_FEF)
+		sigframe_size += sizeof(__siginfo_fpu_t);
+	if (wsaved)
+		sigframe_size += sizeof(__siginfo_rwin_t);
 
 	sf = (struct rt_signal_frame32 __user *)
 		get_sigframe(&ka->sa, regs, sigframe_size);
@@ -651,8 +655,7 @@ static int setup_rt_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 	if (invalid_frame_pointer(sf, sigframe_size))
 		goto sigill;
 
-	if (get_thread_wsaved() != 0)
-		goto sigill;
+	tail = (sf + 1);
 
 	/* 2. Save the current process state */
 	if (test_thread_flag(TIF_32BIT)) {
@@ -677,11 +680,22 @@ static int setup_rt_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 			  &sf->v8plus.asi);
 
 	if (psr & PSR_EF) {
-		err |= save_fpu_state32(regs, &sf->fpu_state);
-		err |= __put_user((u64)&sf->fpu_state, &sf->fpu_save);
+		__siginfo_fpu_t __user *fp = tail;
+		tail += sizeof(*fp);
+		err |= save_fpu_state(regs, fp);
+		err |= __put_user((u64)fp, &sf->fpu_save);
 	} else {
 		err |= __put_user(0, &sf->fpu_save);
 	}
+	if (wsaved) {
+		__siginfo_rwin_t __user *rwp = tail;
+		tail += sizeof(*rwp);
+		err |= save_rwin_state(wsaved, rwp);
+		err |= __put_user((u64)rwp, &sf->rwin_save);
+		set_thread_wsaved(0);
+	} else {
+		err |= __put_user(0, &sf->rwin_save);
+	}
 
 	/* Update the siginfo structure.  */
 	err |= copy_siginfo_to_user32(&sf->info, info);
@@ -703,9 +717,21 @@ static int setup_rt_frame32(struct k_sigaction *ka, struct pt_regs *regs,
 	}
 	err |= __copy_to_user(&sf->mask, &seta, sizeof(compat_sigset_t));
 
-	err |= copy_in_user((u32 __user *)sf,
-			    (u32 __user *)(regs->u_regs[UREG_FP]),
-			    sizeof(struct reg_window32));
+	if (!wsaved) {
+		err |= copy_in_user((u32 __user *)sf,
+				    (u32 __user *)(regs->u_regs[UREG_FP]),
+				    sizeof(struct reg_window32));
+	} else {
+		struct reg_window *rp;
+
+		rp = &current_thread_info()->reg_window[wsaved - 1];
+		for (i = 0; i < 8; i++)
+			err |= __put_user(rp->locals[i], &sf->ss.locals[i]);
+		for (i = 0; i < 6; i++)
+			err |= __put_user(rp->ins[i], &sf->ss.ins[i]);
+		err |= __put_user(rp->ins[6], &sf->ss.fp);
+		err |= __put_user(rp->ins[7], &sf->ss.callers_pc);
+	}
 	if (err)
 		goto sigsegv;
 	
diff --git a/arch/sparc/kernel/signal_32.c b/arch/sparc/kernel/signal_32.c
index 5e5c5fd0378..04ede8f04ad 100644
--- a/arch/sparc/kernel/signal_32.c
+++ b/arch/sparc/kernel/signal_32.c
@@ -26,6 +26,8 @@
 #include <asm/pgtable.h>
 #include <asm/cacheflush.h>	/* flush_sig_insns */
 
+#include "sigutil.h"
+
 #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
 
 extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
@@ -39,8 +41,8 @@ struct signal_frame {
 	unsigned long		insns[2] __attribute__ ((aligned (8)));
 	unsigned int		extramask[_NSIG_WORDS - 1];
 	unsigned int		extra_size; /* Should be 0 */
-	__siginfo_fpu_t		fpu_state;
-};
+	__siginfo_rwin_t __user	*rwin_save;
+} __attribute__((aligned(8)));
 
 struct rt_signal_frame {
 	struct sparc_stackf	ss;
@@ -51,8 +53,8 @@ struct rt_signal_frame {
 	unsigned int		insns[2];
 	stack_t			stack;
 	unsigned int		extra_size; /* Should be 0 */
-	__siginfo_fpu_t		fpu_state;
-};
+	__siginfo_rwin_t __user	*rwin_save;
+} __attribute__((aligned(8)));
 
 /* Align macros */
 #define SF_ALIGNEDSZ  (((sizeof(struct signal_frame) + 7) & (~7)))
@@ -79,43 +81,13 @@ asmlinkage int sys_sigsuspend(old_sigset_t set)
 	return _sigpause_common(set);
 }
 
-static inline int
-restore_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
-{
-	int err;
-#ifdef CONFIG_SMP
-	if (test_tsk_thread_flag(current, TIF_USEDFPU))
-		regs->psr &= ~PSR_EF;
-#else
-	if (current == last_task_used_math) {
-		last_task_used_math = NULL;
-		regs->psr &= ~PSR_EF;
-	}
-#endif
-	set_used_math();
-	clear_tsk_thread_flag(current, TIF_USEDFPU);
-
-	if (!access_ok(VERIFY_READ, fpu, sizeof(*fpu)))
-		return -EFAULT;
-
-	err = __copy_from_user(&current->thread.float_regs[0], &fpu->si_float_regs[0],
-			       (sizeof(unsigned long) * 32));
-	err |= __get_user(current->thread.fsr, &fpu->si_fsr);
-	err |= __get_user(current->thread.fpqdepth, &fpu->si_fpqdepth);
-	if (current->thread.fpqdepth != 0)
-		err |= __copy_from_user(&current->thread.fpqueue[0],
-					&fpu->si_fpqueue[0],
-					((sizeof(unsigned long) +
-					(sizeof(unsigned long *)))*16));
-	return err;
-}
-
 asmlinkage void do_sigreturn(struct pt_regs *regs)
 {
 	struct signal_frame __user *sf;
 	unsigned long up_psr, pc, npc;
 	sigset_t set;
 	__siginfo_fpu_t __user *fpu_save;
+	__siginfo_rwin_t __user *rwin_save;
 	int err;
 
 	/* Always make any pending restarted system calls return -EINTR */
@@ -150,9 +122,11 @@ asmlinkage void do_sigreturn(struct pt_regs *regs)
 	pt_regs_clear_syscall(regs);
 
 	err |= __get_user(fpu_save, &sf->fpu_save);
-
 	if (fpu_save)
 		err |= restore_fpu_state(regs, fpu_save);
+	err |= __get_user(rwin_save, &sf->rwin_save);
+	if (rwin_save)
+		err |= restore_rwin_state(rwin_save);
 
 	/* This is pretty much atomic, no amount locking would prevent
 	 * the races which exist anyways.
@@ -180,6 +154,7 @@ asmlinkage void do_rt_sigreturn(struct pt_regs *regs)
 	struct rt_signal_frame __user *sf;
 	unsigned int psr, pc, npc;
 	__siginfo_fpu_t __user *fpu_save;
+	__siginfo_rwin_t __user *rwin_save;
 	mm_segment_t old_fs;
 	sigset_t set;
 	stack_t st;
@@ -207,8 +182,7 @@ asmlinkage void do_rt_sigreturn(struct pt_regs *regs)
 	pt_regs_clear_syscall(regs);
 
 	err |= __get_user(fpu_save, &sf->fpu_save);
-
-	if (fpu_save)
+	if (!err && fpu_save)
 		err |= restore_fpu_state(regs, fpu_save);
 	err |= __copy_from_user(&set, &sf->mask, sizeof(sigset_t));
 	
@@ -228,6 +202,12 @@ asmlinkage void do_rt_sigreturn(struct pt_regs *regs)
 	do_sigaltstack((const stack_t __user *) &st, NULL, (unsigned long)sf);
 	set_fs(old_fs);
 
+	err |= __get_user(rwin_save, &sf->rwin_save);
+	if (!err && rwin_save) {
+		if (restore_rwin_state(rwin_save))
+			goto segv;
+	}
+
 	sigdelsetmask(&set, ~_BLOCKABLE);
 	spin_lock_irq(&current->sighand->siglock);
 	current->blocked = set;
@@ -280,53 +260,23 @@ static inline void __user *get_sigframe(struct sigaction *sa, struct pt_regs *re
 	return (void __user *) sp;
 }
 
-static inline int
-save_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
-{
-	int err = 0;
-#ifdef CONFIG_SMP
-	if (test_tsk_thread_flag(current, TIF_USEDFPU)) {
-		put_psr(get_psr() | PSR_EF);
-		fpsave(&current->thread.float_regs[0], &current->thread.fsr,
-		       &current->thread.fpqueue[0], &current->thread.fpqdepth);
-		regs->psr &= ~(PSR_EF);
-		clear_tsk_thread_flag(current, TIF_USEDFPU);
-	}
-#else
-	if (current == last_task_used_math) {
-		put_psr(get_psr() | PSR_EF);
-		fpsave(&current->thread.float_regs[0], &current->thread.fsr,
-		       &current->thread.fpqueue[0], &current->thread.fpqdepth);
-		last_task_used_math = NULL;
-		regs->psr &= ~(PSR_EF);
-	}
-#endif
-	err |= __copy_to_user(&fpu->si_float_regs[0],
-			      &current->thread.float_regs[0],
-			      (sizeof(unsigned long) * 32));
-	err |= __put_user(current->thread.fsr, &fpu->si_fsr);
-	err |= __put_user(current->thread.fpqdepth, &fpu->si_fpqdepth);
-	if (current->thread.fpqdepth != 0)
-		err |= __copy_to_user(&fpu->si_fpqueue[0],
-				      &current->thread.fpqueue[0],
-				      ((sizeof(unsigned long) +
-				      (sizeof(unsigned long *)))*16));
-	clear_used_math();
-	return err;
-}
-
 static int setup_frame(struct k_sigaction *ka, struct pt_regs *regs,
 		       int signo, sigset_t *oldset)
 {
 	struct signal_frame __user *sf;
-	int sigframe_size, err;
+	int sigframe_size, err, wsaved;
+	void __user *tail;
 
 	/* 1. Make sure everything is clean */
 	synchronize_user_stack();
 
-	sigframe_size = SF_ALIGNEDSZ;
-	if (!used_math())
-		sigframe_size -= sizeof(__siginfo_fpu_t);
+	wsaved = current_thread_info()->w_saved;
+
+	sigframe_size = sizeof(*sf);
+	if (used_math())
+		sigframe_size += sizeof(__siginfo_fpu_t);
+	if (wsaved)
+		sigframe_size += sizeof(__siginfo_rwin_t);
 
 	sf = (struct signal_frame __user *)
 		get_sigframe(&ka->sa, regs, sigframe_size);
@@ -334,8 +284,7 @@ static int setup_frame(struct k_sigaction *ka, struct pt_regs *regs,
 	if (invalid_frame_pointer(sf, sigframe_size))
 		goto sigill_and_return;
 
-	if (current_thread_info()->w_saved != 0)
-		goto sigill_and_return;
+	tail = sf + 1;
 
 	/* 2. Save the current process state */
 	err = __copy_to_user(&sf->info.si_regs, regs, sizeof(struct pt_regs));
@@ -343,17 +292,34 @@ static int setup_frame(struct k_sigaction *ka, struct pt_regs *regs,
 	err |= __put_user(0, &sf->extra_size);
 
 	if (used_math()) {
-		err |= save_fpu_state(regs, &sf->fpu_state);
-		err |= __put_user(&sf->fpu_state, &sf->fpu_save);
+		__siginfo_fpu_t __user *fp = tail;
+		tail += sizeof(*fp);
+		err |= save_fpu_state(regs, fp);
+		err |= __put_user(fp, &sf->fpu_save);
 	} else {
 		err |= __put_user(0, &sf->fpu_save);
 	}
+	if (wsaved) {
+		__siginfo_rwin_t __user *rwp = tail;
+		tail += sizeof(*rwp);
+		err |= save_rwin_state(wsaved, rwp);
+		err |= __put_user(rwp, &sf->rwin_save);
+	} else {
+		err |= __put_user(0, &sf->rwin_save);
+	}
 
 	err |= __put_user(oldset->sig[0], &sf->info.si_mask);
 	err |= __copy_to_user(sf->extramask, &oldset->sig[1],
 			      (_NSIG_WORDS - 1) * sizeof(unsigned int));
-	err |= __copy_to_user(sf, (char *) regs->u_regs[UREG_FP],
-			      sizeof(struct reg_window32));
+	if (!wsaved) {
+		err |= __copy_to_user(sf, (char *) regs->u_regs[UREG_FP],
+				      sizeof(struct reg_window32));
+	} else {
+		struct reg_window32 *rp;
+
+		rp = &current_thread_info()->reg_window[wsaved - 1];
+		err |= __copy_to_user(sf, rp, sizeof(struct reg_window32));
+	}
 	if (err)
 		goto sigsegv;
 	
@@ -399,21 +365,24 @@ static int setup_rt_frame(struct k_sigaction *ka, struct pt_regs *regs,
 			  int signo, sigset_t *oldset, siginfo_t *info)
 {
 	struct rt_signal_frame __user *sf;
-	int sigframe_size;
+	int sigframe_size, wsaved;
+	void __user *tail;
 	unsigned int psr;
 	int err;
 
 	synchronize_user_stack();
-	sigframe_size = RT_ALIGNEDSZ;
-	if (!used_math())
-		sigframe_size -= sizeof(__siginfo_fpu_t);
+	wsaved = current_thread_info()->w_saved;
+	sigframe_size = sizeof(*sf);
+	if (used_math())
+		sigframe_size += sizeof(__siginfo_fpu_t);
+	if (wsaved)
+		sigframe_size += sizeof(__siginfo_rwin_t);
 	sf = (struct rt_signal_frame __user *)
 		get_sigframe(&ka->sa, regs, sigframe_size);
 	if (invalid_frame_pointer(sf, sigframe_size))
 		goto sigill;
-	if (current_thread_info()->w_saved != 0)
-		goto sigill;
 
+	tail = sf + 1;
 	err  = __put_user(regs->pc, &sf->regs.pc);
 	err |= __put_user(regs->npc, &sf->regs.npc);
 	err |= __put_user(regs->y, &sf->regs.y);
@@ -425,11 +394,21 @@ static int setup_rt_frame(struct k_sigaction *ka, struct pt_regs *regs,
 	err |= __put_user(0, &sf->extra_size);
 
 	if (psr & PSR_EF) {
-		err |= save_fpu_state(regs, &sf->fpu_state);
-		err |= __put_user(&sf->fpu_state, &sf->fpu_save);
+		__siginfo_fpu_t *fp = tail;
+		tail += sizeof(*fp);
+		err |= save_fpu_state(regs, fp);
+		err |= __put_user(fp, &sf->fpu_save);
 	} else {
 		err |= __put_user(0, &sf->fpu_save);
 	}
+	if (wsaved) {
+		__siginfo_rwin_t *rwp = tail;
+		tail += sizeof(*rwp);
+		err |= save_rwin_state(wsaved, rwp);
+		err |= __put_user(rwp, &sf->rwin_save);
+	} else {
+		err |= __put_user(0, &sf->rwin_save);
+	}
 	err |= __copy_to_user(&sf->mask, &oldset->sig[0], sizeof(sigset_t));
 	
 	/* Setup sigaltstack */
@@ -437,8 +416,15 @@ static int setup_rt_frame(struct k_sigaction *ka, struct pt_regs *regs,
 	err |= __put_user(sas_ss_flags(regs->u_regs[UREG_FP]), &sf->stack.ss_flags);
 	err |= __put_user(current->sas_ss_size, &sf->stack.ss_size);
 	
-	err |= __copy_to_user(sf, (char *) regs->u_regs[UREG_FP],
-			      sizeof(struct reg_window32));
+	if (!wsaved) {
+		err |= __copy_to_user(sf, (char *) regs->u_regs[UREG_FP],
+				      sizeof(struct reg_window32));
+	} else {
+		struct reg_window32 *rp;
+
+		rp = &current_thread_info()->reg_window[wsaved - 1];
+		err |= __copy_to_user(sf, rp, sizeof(struct reg_window32));
+	}
 
 	err |= copy_siginfo_to_user(&sf->info, info);
 
diff --git a/arch/sparc/kernel/signal_64.c b/arch/sparc/kernel/signal_64.c
index 006fe451588..47509df3b89 100644
--- a/arch/sparc/kernel/signal_64.c
+++ b/arch/sparc/kernel/signal_64.c
@@ -34,6 +34,7 @@
 
 #include "entry.h"
 #include "systbls.h"
+#include "sigutil.h"
 
 #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
 
@@ -236,7 +237,7 @@ struct rt_signal_frame {
 	__siginfo_fpu_t __user	*fpu_save;
 	stack_t			stack;
 	sigset_t		mask;
-	__siginfo_fpu_t		fpu_state;
+	__siginfo_rwin_t	*rwin_save;
 };
 
 static long _sigpause_common(old_sigset_t set)
@@ -266,33 +267,12 @@ asmlinkage long sys_sigsuspend(old_sigset_t set)
 	return _sigpause_common(set);
 }
 
-static inline int
-restore_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
-{
-	unsigned long *fpregs = current_thread_info()->fpregs;
-	unsigned long fprs;
-	int err;
-
-	err = __get_user(fprs, &fpu->si_fprs);
-	fprs_write(0);
-	regs->tstate &= ~TSTATE_PEF;
-	if (fprs & FPRS_DL)
-		err |= copy_from_user(fpregs, &fpu->si_float_regs[0],
-		       	       (sizeof(unsigned int) * 32));
-	if (fprs & FPRS_DU)
-		err |= copy_from_user(fpregs+16, &fpu->si_float_regs[32],
-		       	       (sizeof(unsigned int) * 32));
-	err |= __get_user(current_thread_info()->xfsr[0], &fpu->si_fsr);
-	err |= __get_user(current_thread_info()->gsr[0], &fpu->si_gsr);
-	current_thread_info()->fpsaved[0] |= fprs;
-	return err;
-}
-
 void do_rt_sigreturn(struct pt_regs *regs)
 {
 	struct rt_signal_frame __user *sf;
 	unsigned long tpc, tnpc, tstate;
 	__siginfo_fpu_t __user *fpu_save;
+	__siginfo_rwin_t __user *rwin_save;
 	sigset_t set;
 	int err;
 
@@ -325,8 +305,8 @@ void do_rt_sigreturn(struct pt_regs *regs)
 	regs->tstate |= (tstate & (TSTATE_ASI | TSTATE_ICC | TSTATE_XCC));
 
 	err |= __get_user(fpu_save, &sf->fpu_save);
-	if (fpu_save)
-		err |= restore_fpu_state(regs, &sf->fpu_state);
+	if (!err && fpu_save)
+		err |= restore_fpu_state(regs, fpu_save);
 
 	err |= __copy_from_user(&set, &sf->mask, sizeof(sigset_t));
 	err |= do_sigaltstack(&sf->stack, NULL, (unsigned long)sf);
@@ -334,6 +314,12 @@ void do_rt_sigreturn(struct pt_regs *regs)
 	if (err)
 		goto segv;
 
+	err |= __get_user(rwin_save, &sf->rwin_save);
+	if (!err && rwin_save) {
+		if (restore_rwin_state(rwin_save))
+			goto segv;
+	}
+
 	regs->tpc = tpc;
 	regs->tnpc = tnpc;
 
@@ -351,34 +337,13 @@ segv:
 }
 
 /* Checks if the fp is valid */
-static int invalid_frame_pointer(void __user *fp, int fplen)
+static int invalid_frame_pointer(void __user *fp)
 {
 	if (((unsigned long) fp) & 15)
 		return 1;
 	return 0;
 }
 
-static inline int
-save_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
-{
-	unsigned long *fpregs = current_thread_info()->fpregs;
-	unsigned long fprs;
-	int err = 0;
-	
-	fprs = current_thread_info()->fpsaved[0];
-	if (fprs & FPRS_DL)
-		err |= copy_to_user(&fpu->si_float_regs[0], fpregs,
-				    (sizeof(unsigned int) * 32));
-	if (fprs & FPRS_DU)
-		err |= copy_to_user(&fpu->si_float_regs[32], fpregs+16,
-				    (sizeof(unsigned int) * 32));
-	err |= __put_user(current_thread_info()->xfsr[0], &fpu->si_fsr);
-	err |= __put_user(current_thread_info()->gsr[0], &fpu->si_gsr);
-	err |= __put_user(fprs, &fpu->si_fprs);
-
-	return err;
-}
-
 static inline void __user *get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, unsigned long framesize)
 {
 	unsigned long sp = regs->u_regs[UREG_FP] + STACK_BIAS;
@@ -414,34 +379,48 @@ setup_rt_frame(struct k_sigaction *ka, struct pt_regs *regs,
 	       int signo, sigset_t *oldset, siginfo_t *info)
 {
 	struct rt_signal_frame __user *sf;
-	int sigframe_size, err;
+	int wsaved, err, sf_size;
+	void __user *tail;
 
 	/* 1. Make sure everything is clean */
 	synchronize_user_stack();
 	save_and_clear_fpu();
 	
-	sigframe_size = sizeof(struct rt_signal_frame);
-	if (!(current_thread_info()->fpsaved[0] & FPRS_FEF))
-		sigframe_size -= sizeof(__siginfo_fpu_t);
+	wsaved = get_thread_wsaved();
 
+	sf_size = sizeof(struct rt_signal_frame);
+	if (current_thread_info()->fpsaved[0] & FPRS_FEF)
+		sf_size += sizeof(__siginfo_fpu_t);
+	if (wsaved)
+		sf_size += sizeof(__siginfo_rwin_t);
 	sf = (struct rt_signal_frame __user *)
-		get_sigframe(ka, regs, sigframe_size);
-	
-	if (invalid_frame_pointer (sf, sigframe_size))
-		goto sigill;
+		get_sigframe(ka, regs, sf_size);
 
-	if (get_thread_wsaved() != 0)
+	if (invalid_frame_pointer (sf))
 		goto sigill;
 
+	tail = (sf + 1);
+
 	/* 2. Save the current process state */
 	err = copy_to_user(&sf->regs, regs, sizeof (*regs));
 
 	if (current_thread_info()->fpsaved[0] & FPRS_FEF) {
-		err |= save_fpu_state(regs, &sf->fpu_state);
-		err |= __put_user((u64)&sf->fpu_state, &sf->fpu_save);
+		__siginfo_fpu_t __user *fpu_save = tail;
+		tail += sizeof(__siginfo_fpu_t);
+		err |= save_fpu_state(regs, fpu_save);
+		err |= __put_user((u64)fpu_save, &sf->fpu_save);
 	} else {
 		err |= __put_user(0, &sf->fpu_save);
 	}
+	if (wsaved) {
+		__siginfo_rwin_t __user *rwin_save = tail;
+		tail += sizeof(__siginfo_rwin_t);
+		err |= save_rwin_state(wsaved, rwin_save);
+		err |= __put_user((u64)rwin_save, &sf->rwin_save);
+		set_thread_wsaved(0);
+	} else {
+		err |= __put_user(0, &sf->rwin_save);
+	}
 	
 	/* Setup sigaltstack */
 	err |= __put_user(current->sas_ss_sp, &sf->stack.ss_sp);
@@ -450,10 +429,17 @@ setup_rt_frame(struct k_sigaction *ka, struct pt_regs *regs,
 
 	err |= copy_to_user(&sf->mask, oldset, sizeof(sigset_t));
 
-	err |= copy_in_user((u64 __user *)sf,
-			    (u64 __user *)(regs->u_regs[UREG_FP]+STACK_BIAS),
-			    sizeof(struct reg_window));
+	if (!wsaved) {
+		err |= copy_in_user((u64 __user *)sf,
+				    (u64 __user *)(regs->u_regs[UREG_FP] +
+						   STACK_BIAS),
+				    sizeof(struct reg_window));
+	} else {
+		struct reg_window *rp;
 
+		rp = &current_thread_info()->reg_window[wsaved - 1];
+		err |= copy_to_user(sf, rp, sizeof(struct reg_window));
+	}
 	if (info)
 		err |= copy_siginfo_to_user(&sf->info, info);
 	else {
diff --git a/arch/sparc/kernel/sigutil.h b/arch/sparc/kernel/sigutil.h
new file mode 100644
index 00000000000..d223aa432bb
--- /dev/null
+++ b/arch/sparc/kernel/sigutil.h
@@ -0,0 +1,9 @@
+#ifndef _SIGUTIL_H
+#define _SIGUTIL_H
+
+int save_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu);
+int restore_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu);
+int save_rwin_state(int wsaved, __siginfo_rwin_t __user *rwin);
+int restore_rwin_state(__siginfo_rwin_t __user *rp);
+
+#endif /* _SIGUTIL_H */
diff --git a/arch/sparc/kernel/sigutil_32.c b/arch/sparc/kernel/sigutil_32.c
new file mode 100644
index 00000000000..35c7897b009
--- /dev/null
+++ b/arch/sparc/kernel/sigutil_32.c
@@ -0,0 +1,120 @@
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/thread_info.h>
+#include <linux/uaccess.h>
+#include <linux/sched.h>
+
+#include <asm/sigcontext.h>
+#include <asm/fpumacro.h>
+#include <asm/ptrace.h>
+
+#include "sigutil.h"
+
+int save_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
+{
+	int err = 0;
+#ifdef CONFIG_SMP
+	if (test_tsk_thread_flag(current, TIF_USEDFPU)) {
+		put_psr(get_psr() | PSR_EF);
+		fpsave(&current->thread.float_regs[0], &current->thread.fsr,
+		       &current->thread.fpqueue[0], &current->thread.fpqdepth);
+		regs->psr &= ~(PSR_EF);
+		clear_tsk_thread_flag(current, TIF_USEDFPU);
+	}
+#else
+	if (current == last_task_used_math) {
+		put_psr(get_psr() | PSR_EF);
+		fpsave(&current->thread.float_regs[0], &current->thread.fsr,
+		       &current->thread.fpqueue[0], &current->thread.fpqdepth);
+		last_task_used_math = NULL;
+		regs->psr &= ~(PSR_EF);
+	}
+#endif
+	err |= __copy_to_user(&fpu->si_float_regs[0],
+			      &current->thread.float_regs[0],
+			      (sizeof(unsigned long) * 32));
+	err |= __put_user(current->thread.fsr, &fpu->si_fsr);
+	err |= __put_user(current->thread.fpqdepth, &fpu->si_fpqdepth);
+	if (current->thread.fpqdepth != 0)
+		err |= __copy_to_user(&fpu->si_fpqueue[0],
+				      &current->thread.fpqueue[0],
+				      ((sizeof(unsigned long) +
+				      (sizeof(unsigned long *)))*16));
+	clear_used_math();
+	return err;
+}
+
+int restore_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
+{
+	int err;
+#ifdef CONFIG_SMP
+	if (test_tsk_thread_flag(current, TIF_USEDFPU))
+		regs->psr &= ~PSR_EF;
+#else
+	if (current == last_task_used_math) {
+		last_task_used_math = NULL;
+		regs->psr &= ~PSR_EF;
+	}
+#endif
+	set_used_math();
+	clear_tsk_thread_flag(current, TIF_USEDFPU);
+
+	if (!access_ok(VERIFY_READ, fpu, sizeof(*fpu)))
+		return -EFAULT;
+
+	err = __copy_from_user(&current->thread.float_regs[0], &fpu->si_float_regs[0],
+			       (sizeof(unsigned long) * 32));
+	err |= __get_user(current->thread.fsr, &fpu->si_fsr);
+	err |= __get_user(current->thread.fpqdepth, &fpu->si_fpqdepth);
+	if (current->thread.fpqdepth != 0)
+		err |= __copy_from_user(&current->thread.fpqueue[0],
+					&fpu->si_fpqueue[0],
+					((sizeof(unsigned long) +
+					(sizeof(unsigned long *)))*16));
+	return err;
+}
+
+int save_rwin_state(int wsaved, __siginfo_rwin_t __user *rwin)
+{
+	int i, err = __put_user(wsaved, &rwin->wsaved);
+
+	for (i = 0; i < wsaved; i++) {
+		struct reg_window32 *rp;
+		unsigned long fp;
+
+		rp = &current_thread_info()->reg_window[i];
+		fp = current_thread_info()->rwbuf_stkptrs[i];
+		err |= copy_to_user(&rwin->reg_window[i], rp,
+				    sizeof(struct reg_window32));
+		err |= __put_user(fp, &rwin->rwbuf_stkptrs[i]);
+	}
+	return err;
+}
+
+int restore_rwin_state(__siginfo_rwin_t __user *rp)
+{
+	struct thread_info *t = current_thread_info();
+	int i, wsaved, err;
+
+	__get_user(wsaved, &rp->wsaved);
+	if (wsaved > NSWINS)
+		return -EFAULT;
+
+	err = 0;
+	for (i = 0; i < wsaved; i++) {
+		err |= copy_from_user(&t->reg_window[i],
+				      &rp->reg_window[i],
+				      sizeof(struct reg_window32));
+		err |= __get_user(t->rwbuf_stkptrs[i],
+				  &rp->rwbuf_stkptrs[i]);
+	}
+	if (err)
+		return err;
+
+	t->w_saved = wsaved;
+	synchronize_user_stack();
+	if (t->w_saved)
+		return -EFAULT;
+	return 0;
+
+}
diff --git a/arch/sparc/kernel/sigutil_64.c b/arch/sparc/kernel/sigutil_64.c
new file mode 100644
index 00000000000..e7dc508c38e
--- /dev/null
+++ b/arch/sparc/kernel/sigutil_64.c
@@ -0,0 +1,93 @@
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/thread_info.h>
+#include <linux/uaccess.h>
+
+#include <asm/sigcontext.h>
+#include <asm/fpumacro.h>
+#include <asm/ptrace.h>
+
+#include "sigutil.h"
+
+int save_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
+{
+	unsigned long *fpregs = current_thread_info()->fpregs;
+	unsigned long fprs;
+	int err = 0;
+	
+	fprs = current_thread_info()->fpsaved[0];
+	if (fprs & FPRS_DL)
+		err |= copy_to_user(&fpu->si_float_regs[0], fpregs,
+				    (sizeof(unsigned int) * 32));
+	if (fprs & FPRS_DU)
+		err |= copy_to_user(&fpu->si_float_regs[32], fpregs+16,
+				    (sizeof(unsigned int) * 32));
+	err |= __put_user(current_thread_info()->xfsr[0], &fpu->si_fsr);
+	err |= __put_user(current_thread_info()->gsr[0], &fpu->si_gsr);
+	err |= __put_user(fprs, &fpu->si_fprs);
+
+	return err;
+}
+
+int restore_fpu_state(struct pt_regs *regs, __siginfo_fpu_t __user *fpu)
+{
+	unsigned long *fpregs = current_thread_info()->fpregs;
+	unsigned long fprs;
+	int err;
+
+	err = __get_user(fprs, &fpu->si_fprs);
+	fprs_write(0);
+	regs->tstate &= ~TSTATE_PEF;
+	if (fprs & FPRS_DL)
+		err |= copy_from_user(fpregs, &fpu->si_float_regs[0],
+		       	       (sizeof(unsigned int) * 32));
+	if (fprs & FPRS_DU)
+		err |= copy_from_user(fpregs+16, &fpu->si_float_regs[32],
+		       	       (sizeof(unsigned int) * 32));
+	err |= __get_user(current_thread_info()->xfsr[0], &fpu->si_fsr);
+	err |= __get_user(current_thread_info()->gsr[0], &fpu->si_gsr);
+	current_thread_info()->fpsaved[0] |= fprs;
+	return err;
+}
+
+int save_rwin_state(int wsaved, __siginfo_rwin_t __user *rwin)
+{
+	int i, err = __put_user(wsaved, &rwin->wsaved);
+
+	for (i = 0; i < wsaved; i++) {
+		struct reg_window *rp = &current_thread_info()->reg_window[i];
+		unsigned long fp = current_thread_info()->rwbuf_stkptrs[i];
+
+		err |= copy_to_user(&rwin->reg_window[i], rp,
+				    sizeof(struct reg_window));
+		err |= __put_user(fp, &rwin->rwbuf_stkptrs[i]);
+	}
+	return err;
+}
+
+int restore_rwin_state(__siginfo_rwin_t __user *rp)
+{
+	struct thread_info *t = current_thread_info();
+	int i, wsaved, err;
+
+	__get_user(wsaved, &rp->wsaved);
+	if (wsaved > NSWINS)
+		return -EFAULT;
+
+	err = 0;
+	for (i = 0; i < wsaved; i++) {
+		err |= copy_from_user(&t->reg_window[i],
+				      &rp->reg_window[i],
+				      sizeof(struct reg_window));
+		err |= __get_user(t->rwbuf_stkptrs[i],
+				  &rp->rwbuf_stkptrs[i]);
+	}
+	if (err)
+		return err;
+
+	set_thread_wsaved(wsaved);
+	synchronize_user_stack();
+	if (get_thread_wsaved())
+		return -EFAULT;
+	return 0;
+}
diff --git a/arch/sparc/kernel/smp_32.c b/arch/sparc/kernel/smp_32.c
index 21b125341bf..f671e7fd6dd 100644
--- a/arch/sparc/kernel/smp_32.c
+++ b/arch/sparc/kernel/smp_32.c
@@ -22,7 +22,7 @@
 #include <linux/delay.h>
 
 #include <asm/ptrace.h>
-#include <asm/atomic.h>
+#include <linux/atomic.h>
 
 #include <asm/irq.h>
 #include <asm/page.h>
diff --git a/arch/sparc/kernel/smp_64.c b/arch/sparc/kernel/smp_64.c
index 99cb17251bb..4a442c32e11 100644
--- a/arch/sparc/kernel/smp_64.c
+++ b/arch/sparc/kernel/smp_64.c
@@ -28,7 +28,7 @@
 
 #include <asm/head.h>
 #include <asm/ptrace.h>
-#include <asm/atomic.h>
+#include <linux/atomic.h>
 #include <asm/tlbflush.h>
 #include <asm/mmu_context.h>
 #include <asm/cpudata.h>
diff --git a/arch/sparc/kernel/sparc_ksyms_64.c b/arch/sparc/kernel/sparc_ksyms_64.c
index 372ad59c4cb..83b47ab02d9 100644
--- a/arch/sparc/kernel/sparc_ksyms_64.c
+++ b/arch/sparc/kernel/sparc_ksyms_64.c
@@ -8,6 +8,7 @@
 #include <linux/module.h>
 #include <linux/pci.h>
 #include <linux/init.h>
+#include <linux/bitops.h>
 
 #include <asm/system.h>
 #include <asm/cpudata.h>
@@ -38,5 +39,15 @@ EXPORT_SYMBOL(sun4v_niagara_setperf);
 EXPORT_SYMBOL(sun4v_niagara2_getperf);
 EXPORT_SYMBOL(sun4v_niagara2_setperf);
 
+/* from hweight.S */
+EXPORT_SYMBOL(__arch_hweight8);
+EXPORT_SYMBOL(__arch_hweight16);
+EXPORT_SYMBOL(__arch_hweight32);
+EXPORT_SYMBOL(__arch_hweight64);
+
+/* from ffs_ffz.S */
+EXPORT_SYMBOL(ffs);
+EXPORT_SYMBOL(__ffs);
+
 /* Exporting a symbol from /init/main.c */
 EXPORT_SYMBOL(saved_command_line);
diff --git a/arch/sparc/kernel/sstate.c b/arch/sparc/kernel/sstate.c
index 8cdbe5946b4..c59af546f52 100644
--- a/arch/sparc/kernel/sstate.c
+++ b/arch/sparc/kernel/sstate.c
@@ -14,14 +14,9 @@
 #include <asm/head.h>
 #include <asm/io.h>
 
-static int hv_supports_soft_state;
-
-static unsigned long kimage_addr_to_ra(const char *p)
-{
-	unsigned long val = (unsigned long) p;
+#include "kernel.h"
 
-	return kern_base + (val - KERNBASE);
-}
+static int hv_supports_soft_state;
 
 static void do_set_sstate(unsigned long state, const char *msg)
 {
diff --git a/arch/sparc/kernel/sys32.S b/arch/sparc/kernel/sys32.S
index 44e5faf1ad5..d97f3eb72e0 100644
--- a/arch/sparc/kernel/sys32.S
+++ b/arch/sparc/kernel/sys32.S
@@ -81,7 +81,6 @@ SIGN2(sys32_fadvise64, compat_sys_fadvise64, %o0, %o4)
 SIGN2(sys32_fadvise64_64, compat_sys_fadvise64_64, %o0, %o5)
 SIGN2(sys32_bdflush, sys_bdflush, %o0, %o1)
 SIGN1(sys32_mlockall, sys_mlockall, %o0)
-SIGN1(sys32_nfsservctl, compat_sys_nfsservctl, %o0)
 SIGN1(sys32_clock_nanosleep, compat_sys_clock_nanosleep, %o1)
 SIGN1(sys32_timer_settime, compat_sys_timer_settime, %o1)
 SIGN1(sys32_io_submit, compat_sys_io_submit, %o1)
diff --git a/arch/sparc/kernel/systbls_32.S b/arch/sparc/kernel/systbls_32.S
index 6e492d59f6b..09d8ec45445 100644
--- a/arch/sparc/kernel/systbls_32.S
+++ b/arch/sparc/kernel/systbls_32.S
@@ -67,7 +67,7 @@ sys_call_table:
 /*235*/	.long sys_fstatfs64, sys_llseek, sys_mlock, sys_munlock, sys_mlockall
 /*240*/	.long sys_munlockall, sys_sched_setparam, sys_sched_getparam, sys_sched_setscheduler, sys_sched_getscheduler
 /*245*/	.long sys_sched_yield, sys_sched_get_priority_max, sys_sched_get_priority_min, sys_sched_rr_get_interval, sys_nanosleep
-/*250*/	.long sys_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
+/*250*/	.long sys_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_ni_syscall
 /*255*/	.long sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
 /*260*/	.long sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
 /*265*/	.long sys_timer_delete, sys_timer_create, sys_nis_syscall, sys_io_setup, sys_io_destroy
diff --git a/arch/sparc/kernel/systbls_64.S b/arch/sparc/kernel/systbls_64.S
index f566518483b..edbec45d468 100644
--- a/arch/sparc/kernel/systbls_64.S
+++ b/arch/sparc/kernel/systbls_64.S
@@ -68,7 +68,7 @@ sys_call_table32:
 	.word compat_sys_fstatfs64, sys_llseek, sys_mlock, sys_munlock, sys32_mlockall
 /*240*/	.word sys_munlockall, sys32_sched_setparam, sys32_sched_getparam, sys32_sched_setscheduler, sys32_sched_getscheduler
 	.word sys_sched_yield, sys32_sched_get_priority_max, sys32_sched_get_priority_min, sys32_sched_rr_get_interval, compat_sys_nanosleep
-/*250*/	.word sys_mremap, compat_sys_sysctl, sys32_getsid, sys_fdatasync, sys32_nfsservctl
+/*250*/	.word sys_mremap, compat_sys_sysctl, sys32_getsid, sys_fdatasync, sys_nis_syscall
 	.word sys32_sync_file_range, compat_sys_clock_settime, compat_sys_clock_gettime, compat_sys_clock_getres, sys32_clock_nanosleep
 /*260*/	.word compat_sys_sched_getaffinity, compat_sys_sched_setaffinity, sys32_timer_settime, compat_sys_timer_gettime, sys_timer_getoverrun
 	.word sys_timer_delete, compat_sys_timer_create, sys_ni_syscall, compat_sys_io_setup, sys_io_destroy
@@ -145,7 +145,7 @@ sys_call_table:
 	.word sys_fstatfs64, sys_llseek, sys_mlock, sys_munlock, sys_mlockall
 /*240*/	.word sys_munlockall, sys_sched_setparam, sys_sched_getparam, sys_sched_setscheduler, sys_sched_getscheduler
 	.word sys_sched_yield, sys_sched_get_priority_max, sys_sched_get_priority_min, sys_sched_rr_get_interval, sys_nanosleep
-/*250*/	.word sys_64_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
+/*250*/	.word sys_64_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nis_syscall
 	.word sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
 /*260*/	.word sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
 	.word sys_timer_delete, sys_timer_create, sys_ni_syscall, sys_io_setup, sys_io_destroy
diff --git a/arch/sparc/kernel/unaligned_64.c b/arch/sparc/kernel/unaligned_64.c
index 35cff1673aa..76e4ac1a13e 100644
--- a/arch/sparc/kernel/unaligned_64.c
+++ b/arch/sparc/kernel/unaligned_64.c
@@ -22,6 +22,7 @@
 #include <linux/bitops.h>
 #include <linux/perf_event.h>
 #include <linux/ratelimit.h>
+#include <linux/bitops.h>
 #include <asm/fpumacro.h>
 
 enum direction {
@@ -373,16 +374,11 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
 	}
 }
 
-static char popc_helper[] = {
-0, 1, 1, 2, 1, 2, 2, 3,
-1, 2, 2, 3, 2, 3, 3, 4, 
-};
-
 int handle_popc(u32 insn, struct pt_regs *regs)
 {
-	u64 value;
-	int ret, i, rd = ((insn >> 25) & 0x1f);
 	int from_kernel = (regs->tstate & TSTATE_PRIV) != 0;
+	int ret, rd = ((insn >> 25) & 0x1f);
+	u64 value;
 	                        
 	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 	if (insn & 0x2000) {
@@ -392,10 +388,7 @@ int handle_popc(u32 insn, struct pt_regs *regs)
 		maybe_flush_windows(0, insn & 0x1f, rd, from_kernel);
 		value = fetch_reg(insn & 0x1f, regs);
 	}
-	for (ret = 0, i = 0; i < 16; i++) {
-		ret += popc_helper[value & 0xf];
-		value >>= 4;
-	}
+	ret = hweight64(value);
 	if (rd < 16) {
 		if (rd)
 			regs->u_regs[rd] = ret;
diff --git a/arch/sparc/kernel/vmlinux.lds.S b/arch/sparc/kernel/vmlinux.lds.S
index c0220759003..0e1605697b4 100644
--- a/arch/sparc/kernel/vmlinux.lds.S
+++ b/arch/sparc/kernel/vmlinux.lds.S
@@ -107,7 +107,26 @@ SECTIONS
 		*(.sun4v_2insn_patch)
 		__sun4v_2insn_patch_end = .;
 	}
-
+	.swapper_tsb_phys_patch : {
+		__swapper_tsb_phys_patch = .;
+		*(.swapper_tsb_phys_patch)
+		__swapper_tsb_phys_patch_end = .;
+	}
+	.swapper_4m_tsb_phys_patch : {
+		__swapper_4m_tsb_phys_patch = .;
+		*(.swapper_4m_tsb_phys_patch)
+		__swapper_4m_tsb_phys_patch_end = .;
+	}
+	.popc_3insn_patch : {
+		__popc_3insn_patch = .;
+		*(.popc_3insn_patch)
+		__popc_3insn_patch_end = .;
+	}
+	.popc_6insn_patch : {
+		__popc_6insn_patch = .;
+		*(.popc_6insn_patch)
+		__popc_6insn_patch_end = .;
+	}
 	PERCPU_SECTION(SMP_CACHE_BYTES)
 
 	. = ALIGN(PAGE_SIZE);