Merge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6: (27 commits)
  [IA64] kdump: Add crash_save_vmcoreinfo for INIT
  [IA64] Fix NUMA configuration issue
  [IA64] Itanium Spec updates
  [IA64] Untangle sync_icache_dcache() page size determination
  [IA64] arch/ia64/kernel/: use time_* macros
  [IA64] remove redundant display of free swap space in show_mem()
  [IA64] make IOMMU respect the segment boundary limits
  [IA64] kprobes: kprobe-booster for ia64
  [IA64] fix getpid and set_tid_address fast system calls for pid namespaces
  [IA64] Replace explicit jiffies tests with time_* macros.
  [IA64] use goto to jump out do/while_each_thread
  [IA64] Fix unlock ordering in smp_callin
  [IA64] pgd_offset() constfication.
  [IA64] kdump: crash.c coding style fix
  [IA64] kdump: add kdump_on_fatal_mca
  [IA64] Minimize per_cpu reservations.
  [IA64] Correct pernodesize calculation.
  [IA64] Kernel parameter for max number of concurrent global TLB purges
  [IA64] Multiple outstanding ptc.g instruction support
  [IA64] Implement smp_call_function_mask for ia64
  ...

34 files changed, 2645 insertions, 526 deletions

diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig
index 8fa3faf5ef1..ed21737a00c 100644
--- a/arch/ia64/Kconfig
+++ b/arch/ia64/Kconfig
@@ -283,6 +283,17 @@ config FORCE_MAX_ZONEORDER
 	default "17" if HUGETLB_PAGE
 	default "11"
 
+config VIRT_CPU_ACCOUNTING
+	bool "Deterministic task and CPU time accounting"
+	default n
+	help
+	  Select this option to enable more accurate task and CPU time
+	  accounting.  This is done by reading a CPU counter on each
+	  kernel entry and exit and on transitions within the kernel
+	  between system, softirq and hardirq state, so there is a
+	  small performance impact.
+	  If in doubt, say N here.
+
 config SMP
 	bool "Symmetric multi-processing support"
 	help
@@ -611,6 +622,9 @@ config IRQ_PER_CPU
 	bool
 	default y
 
+config IOMMU_HELPER
+	def_bool (IA64_HP_ZX1 || IA64_HP_ZX1_SWIOTLB || IA64_GENERIC)
+
 source "arch/ia64/hp/sim/Kconfig"
 
 source "arch/ia64/Kconfig.debug"
diff --git a/arch/ia64/hp/common/sba_iommu.c b/arch/ia64/hp/common/sba_iommu.c
index 523eae6d3e4..9409de5c944 100644
--- a/arch/ia64/hp/common/sba_iommu.c
+++ b/arch/ia64/hp/common/sba_iommu.c
@@ -35,6 +35,7 @@
 #include <linux/nodemask.h>
 #include <linux/bitops.h>         /* hweight64() */
 #include <linux/crash_dump.h>
+#include <linux/iommu-helper.h>
 
 #include <asm/delay.h>		/* ia64_get_itc() */
 #include <asm/io.h>
@@ -460,6 +461,13 @@ get_iovp_order (unsigned long size)
 	return order;
 }
 
+static unsigned long ptr_to_pide(struct ioc *ioc, unsigned long *res_ptr,
+				 unsigned int bitshiftcnt)
+{
+	return (((unsigned long)res_ptr - (unsigned long)ioc->res_map) << 3)
+		+ bitshiftcnt;
+}
+
 /**
  * sba_search_bitmap - find free space in IO PDIR resource bitmap
  * @ioc: IO MMU structure which owns the pdir we are interested in.
@@ -471,15 +479,25 @@ get_iovp_order (unsigned long size)
  * Cool perf optimization: search for log2(size) bits at a time.
  */
 static SBA_INLINE unsigned long
-sba_search_bitmap(struct ioc *ioc, unsigned long bits_wanted, int use_hint)
+sba_search_bitmap(struct ioc *ioc, struct device *dev,
+		  unsigned long bits_wanted, int use_hint)
 {
 	unsigned long *res_ptr;
 	unsigned long *res_end = (unsigned long *) &(ioc->res_map[ioc->res_size]);
-	unsigned long flags, pide = ~0UL;
+	unsigned long flags, pide = ~0UL, tpide;
+	unsigned long boundary_size;
+	unsigned long shift;
+	int ret;
 
 	ASSERT(((unsigned long) ioc->res_hint & (sizeof(unsigned long) - 1UL)) == 0);
 	ASSERT(res_ptr < res_end);
 
+	boundary_size = (unsigned long long)dma_get_seg_boundary(dev) + 1;
+	boundary_size = ALIGN(boundary_size, 1ULL << iovp_shift) >> iovp_shift;
+
+	BUG_ON(ioc->ibase & ~iovp_mask);
+	shift = ioc->ibase >> iovp_shift;
+
 	spin_lock_irqsave(&ioc->res_lock, flags);
 
 	/* Allow caller to force a search through the entire resource space */
@@ -504,9 +522,7 @@ sba_search_bitmap(struct ioc *ioc, unsigned long bits_wanted, int use_hint)
 			if (likely(*res_ptr != ~0UL)) {
 				bitshiftcnt = ffz(*res_ptr);
 				*res_ptr |= (1UL << bitshiftcnt);
-				pide = ((unsigned long)res_ptr - (unsigned long)ioc->res_map);
-				pide <<= 3;	/* convert to bit address */
-				pide += bitshiftcnt;
+				pide = ptr_to_pide(ioc, res_ptr, bitshiftcnt);
 				ioc->res_bitshift = bitshiftcnt + bits_wanted;
 				goto found_it;
 			}
@@ -535,11 +551,13 @@ sba_search_bitmap(struct ioc *ioc, unsigned long bits_wanted, int use_hint)
 			DBG_RES("    %p %lx %lx\n", res_ptr, mask, *res_ptr);
 			ASSERT(0 != mask);
 			for (; mask ; mask <<= o, bitshiftcnt += o) {
-				if(0 == ((*res_ptr) & mask)) {
+				tpide = ptr_to_pide(ioc, res_ptr, bitshiftcnt);
+				ret = iommu_is_span_boundary(tpide, bits_wanted,
+							     shift,
+							     boundary_size);
+				if ((0 == ((*res_ptr) & mask)) && !ret) {
 					*res_ptr |= mask;     /* mark resources busy! */
-					pide = ((unsigned long)res_ptr - (unsigned long)ioc->res_map);
-					pide <<= 3;	/* convert to bit address */
-					pide += bitshiftcnt;
+					pide = tpide;
 					ioc->res_bitshift = bitshiftcnt + bits_wanted;
 					goto found_it;
 				}
@@ -560,6 +578,11 @@ sba_search_bitmap(struct ioc *ioc, unsigned long bits_wanted, int use_hint)
 		end = res_end - qwords;
 
 		for (; res_ptr < end; res_ptr++) {
+			tpide = ptr_to_pide(ioc, res_ptr, 0);
+			ret = iommu_is_span_boundary(tpide, bits_wanted,
+						     shift, boundary_size);
+			if (ret)
+				goto next_ptr;
 			for (i = 0 ; i < qwords ; i++) {
 				if (res_ptr[i] != 0)
 					goto next_ptr;
@@ -572,8 +595,7 @@ sba_search_bitmap(struct ioc *ioc, unsigned long bits_wanted, int use_hint)
 				res_ptr[i] = ~0UL;
 			res_ptr[i] |= RESMAP_MASK(bits);
 
-			pide = ((unsigned long)res_ptr - (unsigned long)ioc->res_map);
-			pide <<= 3;	/* convert to bit address */
+			pide = tpide;
 			res_ptr += qwords;
 			ioc->res_bitshift = bits;
 			goto found_it;
@@ -605,7 +627,7 @@ found_it:
  * resource bit map.
  */
 static int
-sba_alloc_range(struct ioc *ioc, size_t size)
+sba_alloc_range(struct ioc *ioc, struct device *dev, size_t size)
 {
 	unsigned int pages_needed = size >> iovp_shift;
 #ifdef PDIR_SEARCH_TIMING
@@ -622,9 +644,9 @@ sba_alloc_range(struct ioc *ioc, size_t size)
 	/*
 	** "seek and ye shall find"...praying never hurts either...
 	*/
-	pide = sba_search_bitmap(ioc, pages_needed, 1);
+	pide = sba_search_bitmap(ioc, dev, pages_needed, 1);
 	if (unlikely(pide >= (ioc->res_size << 3))) {
-		pide = sba_search_bitmap(ioc, pages_needed, 0);
+		pide = sba_search_bitmap(ioc, dev, pages_needed, 0);
 		if (unlikely(pide >= (ioc->res_size << 3))) {
 #if DELAYED_RESOURCE_CNT > 0
 			unsigned long flags;
@@ -653,7 +675,7 @@ sba_alloc_range(struct ioc *ioc, size_t size)
 			}
 			spin_unlock_irqrestore(&ioc->saved_lock, flags);
 
-			pide = sba_search_bitmap(ioc, pages_needed, 0);
+			pide = sba_search_bitmap(ioc, dev, pages_needed, 0);
 			if (unlikely(pide >= (ioc->res_size << 3)))
 				panic(__FILE__ ": I/O MMU @ %p is out of mapping resources\n",
 				      ioc->ioc_hpa);
@@ -936,7 +958,7 @@ sba_map_single(struct device *dev, void *addr, size_t size, int dir)
 	spin_unlock_irqrestore(&ioc->res_lock, flags);
 #endif
 
-	pide = sba_alloc_range(ioc, size);
+	pide = sba_alloc_range(ioc, dev, size);
 
 	iovp = (dma_addr_t) pide << iovp_shift;
 
@@ -1373,7 +1395,7 @@ sba_coalesce_chunks(struct ioc *ioc, struct device *dev,
 		dma_len = (dma_len + dma_offset + ~iovp_mask) & iovp_mask;
 		ASSERT(dma_len <= DMA_CHUNK_SIZE);
 		dma_sg->dma_address = (dma_addr_t) (PIDE_FLAG
-			| (sba_alloc_range(ioc, dma_len) << iovp_shift)
+			| (sba_alloc_range(ioc, dev, dma_len) << iovp_shift)
 			| dma_offset);
 		n_mappings++;
 	}
diff --git a/arch/ia64/ia32/elfcore32.h b/arch/ia64/ia32/elfcore32.h
index 446c9aac924..9a3abf58cea 100644
--- a/arch/ia64/ia32/elfcore32.h
+++ b/arch/ia64/ia32/elfcore32.h
@@ -30,7 +30,19 @@ struct elf_siginfo
 	int	si_errno;			/* errno */
 };
 
-#define jiffies_to_timeval(a,b) do { (b)->tv_usec = 0; (b)->tv_sec = (a)/HZ; }while(0)
+#ifdef CONFIG_VIRT_CPU_ACCOUNTING
+/*
+ * Hacks are here since types between compat_timeval (= pair of s32) and
+ * ia64-native timeval (= pair of s64) are not compatible, at least a file
+ * arch/ia64/ia32/../../../fs/binfmt_elf.c will get warnings from compiler on
+ * use of cputime_to_timeval(), which usually an alias of jiffies_to_timeval().
+ */
+#define cputime_to_timeval(a,b) \
+	do { (b)->tv_usec = 0; (b)->tv_sec = (a)/NSEC_PER_SEC; } while(0)
+#else
+#define jiffies_to_timeval(a,b) \
+	do { (b)->tv_usec = 0; (b)->tv_sec = (a)/HZ; } while(0)
+#endif
 
 struct elf_prstatus
 {
diff --git a/arch/ia64/ia32/sys_ia32.c b/arch/ia64/ia32/sys_ia32.c
index b1bf51fe97b..7e028ceb93b 100644
--- a/arch/ia64/ia32/sys_ia32.c
+++ b/arch/ia64/ia32/sys_ia32.c
@@ -38,6 +38,7 @@
 #include <linux/eventpoll.h>
 #include <linux/personality.h>
 #include <linux/ptrace.h>
+#include <linux/regset.h>
 #include <linux/stat.h>
 #include <linux/ipc.h>
 #include <linux/capability.h>
@@ -2387,16 +2388,45 @@ get_free_idx (void)
 	return -ESRCH;
 }
 
+static void set_tls_desc(struct task_struct *p, int idx,
+		const struct ia32_user_desc *info, int n)
+{
+	struct thread_struct *t = &p->thread;
+	struct desc_struct *desc = &t->tls_array[idx - GDT_ENTRY_TLS_MIN];
+	int cpu;
+
+	/*
+	 * We must not get preempted while modifying the TLS.
+	 */
+	cpu = get_cpu();
+
+	while (n-- > 0) {
+		if (LDT_empty(info)) {
+			desc->a = 0;
+			desc->b = 0;
+		} else {
+			desc->a = LDT_entry_a(info);
+			desc->b = LDT_entry_b(info);
+		}
+
+		++info;
+		++desc;
+	}
+
+	if (t == &current->thread)
+		load_TLS(t, cpu);
+
+	put_cpu();
+}
+
 /*
  * Set a given TLS descriptor:
  */
 asmlinkage int
 sys32_set_thread_area (struct ia32_user_desc __user *u_info)
 {
-	struct thread_struct *t = &current->thread;
 	struct ia32_user_desc info;
-	struct desc_struct *desc;
-	int cpu, idx;
+	int idx;
 
 	if (copy_from_user(&info, u_info, sizeof(info)))
 		return -EFAULT;
@@ -2416,18 +2446,7 @@ sys32_set_thread_area (struct ia32_user_desc __user *u_info)
 	if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX)
 		return -EINVAL;
 
-	desc = t->tls_array + idx - GDT_ENTRY_TLS_MIN;
-
-	cpu = smp_processor_id();
-
-	if (LDT_empty(&info)) {
-		desc->a = 0;
-		desc->b = 0;
-	} else {
-		desc->a = LDT_entry_a(&info);
-		desc->b = LDT_entry_b(&info);
-	}
-	load_TLS(t, cpu);
+	set_tls_desc(current, idx, &info, 1);
 	return 0;
 }
 
@@ -2451,6 +2470,20 @@ sys32_set_thread_area (struct ia32_user_desc __user *u_info)
 #define GET_PRESENT(desc)	(((desc)->b >> 15) & 1)
 #define GET_USEABLE(desc)	(((desc)->b >> 20) & 1)
 
+static void fill_user_desc(struct ia32_user_desc *info, int idx,
+		const struct desc_struct *desc)
+{
+	info->entry_number = idx;
+	info->base_addr = GET_BASE(desc);
+	info->limit = GET_LIMIT(desc);
+	info->seg_32bit = GET_32BIT(desc);
+	info->contents = GET_CONTENTS(desc);
+	info->read_exec_only = !GET_WRITABLE(desc);
+	info->limit_in_pages = GET_LIMIT_PAGES(desc);
+	info->seg_not_present = !GET_PRESENT(desc);
+	info->useable = GET_USEABLE(desc);
+}
+
 asmlinkage int
 sys32_get_thread_area (struct ia32_user_desc __user *u_info)
 {
@@ -2464,22 +2497,588 @@ sys32_get_thread_area (struct ia32_user_desc __user *u_info)
 		return -EINVAL;
 
 	desc = current->thread.tls_array + idx - GDT_ENTRY_TLS_MIN;
-
-	info.entry_number = idx;
-	info.base_addr = GET_BASE(desc);
-	info.limit = GET_LIMIT(desc);
-	info.seg_32bit = GET_32BIT(desc);
-	info.contents = GET_CONTENTS(desc);
-	info.read_exec_only = !GET_WRITABLE(desc);
-	info.limit_in_pages = GET_LIMIT_PAGES(desc);
-	info.seg_not_present = !GET_PRESENT(desc);
-	info.useable = GET_USEABLE(desc);
+	fill_user_desc(&info, idx, desc);
 
 	if (copy_to_user(u_info, &info, sizeof(info)))
 		return -EFAULT;
 	return 0;
 }
 
+struct regset_get {
+	void *kbuf;
+	void __user *ubuf;
+};
+
+struct regset_set {
+	const void *kbuf;
+	const void __user *ubuf;
+};
+
+struct regset_getset {
+	struct task_struct *target;
+	const struct user_regset *regset;
+	union {
+		struct regset_get get;
+		struct regset_set set;
+	} u;
+	unsigned int pos;
+	unsigned int count;
+	int ret;
+};
+
+static void getfpreg(struct task_struct *task, int regno, int *val)
+{
+	switch (regno / sizeof(int)) {
+	case 0:
+		*val = task->thread.fcr & 0xffff;
+		break;
+	case 1:
+		*val = task->thread.fsr & 0xffff;
+		break;
+	case 2:
+		*val = (task->thread.fsr>>16) & 0xffff;
+		break;
+	case 3:
+		*val = task->thread.fir;
+		break;
+	case 4:
+		*val = (task->thread.fir>>32) & 0xffff;
+		break;
+	case 5:
+		*val = task->thread.fdr;
+		break;
+	case 6:
+		*val = (task->thread.fdr >> 32) & 0xffff;
+		break;
+	}
+}
+
+static void setfpreg(struct task_struct *task, int regno, int val)
+{
+	switch (regno / sizeof(int)) {
+	case 0:
+		task->thread.fcr = (task->thread.fcr & (~0x1f3f))
+			| (val & 0x1f3f);
+		break;
+	case 1:
+		task->thread.fsr = (task->thread.fsr & (~0xffff)) | val;
+		break;
+	case 2:
+		task->thread.fsr = (task->thread.fsr & (~0xffff0000))
+			| (val << 16);
+		break;
+	case 3:
+		task->thread.fir = (task->thread.fir & (~0xffffffff)) | val;
+		break;
+	case 5:
+		task->thread.fdr = (task->thread.fdr & (~0xffffffff)) | val;
+		break;
+	}
+}
+
+static void access_fpreg_ia32(int regno, void *reg,
+		struct pt_regs *pt, struct switch_stack *sw,
+		int tos, int write)
+{
+	void *f;
+
+	if ((regno += tos) >= 8)
+		regno -= 8;
+	if (regno < 4)
+		f = &pt->f8 + regno;
+	else if (regno <= 7)
+		f = &sw->f12 + (regno - 4);
+	else {
+		printk(KERN_ERR "regno must be less than 7 \n");
+		 return;
+	}
+
+	if (write)
+		memcpy(f, reg, sizeof(struct _fpreg_ia32));
+	else
+		memcpy(reg, f, sizeof(struct _fpreg_ia32));
+}
+
+static void do_fpregs_get(struct unw_frame_info *info, void *arg)
+{
+	struct regset_getset *dst = arg;
+	struct task_struct *task = dst->target;
+	struct pt_regs *pt;
+	int start, end, tos;
+	char buf[80];
+
+	if (dst->count == 0 || unw_unwind_to_user(info) < 0)
+		return;
+	if (dst->pos < 7 * sizeof(int)) {
+		end = min((dst->pos + dst->count),
+			(unsigned int)(7 * sizeof(int)));
+		for (start = dst->pos; start < end; start += sizeof(int))
+			getfpreg(task, start, (int *)(buf + start));
+		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
+				&dst->u.get.kbuf, &dst->u.get.ubuf, buf,
+				0, 7 * sizeof(int));
+		if (dst->ret || dst->count == 0)
+			return;
+	}
+	if (dst->pos < sizeof(struct ia32_user_i387_struct)) {
+		pt = task_pt_regs(task);
+		tos = (task->thread.fsr >> 11) & 7;
+		end = min(dst->pos + dst->count,
+			(unsigned int)(sizeof(struct ia32_user_i387_struct)));
+		start = (dst->pos - 7 * sizeof(int)) /
+			sizeof(struct _fpreg_ia32);
+		end = (end - 7 * sizeof(int)) / sizeof(struct _fpreg_ia32);
+		for (; start < end; start++)
+			access_fpreg_ia32(start,
+				(struct _fpreg_ia32 *)buf + start,
+				pt, info->sw, tos, 0);
+		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
+				&dst->u.get.kbuf, &dst->u.get.ubuf,
+				buf, 7 * sizeof(int),
+				sizeof(struct ia32_user_i387_struct));
+		if (dst->ret || dst->count == 0)
+			return;
+	}
+}
+
+static void do_fpregs_set(struct unw_frame_info *info, void *arg)
+{
+	struct regset_getset *dst = arg;
+	struct task_struct *task = dst->target;
+	struct pt_regs *pt;
+	char buf[80];
+	int end, start, tos;
+
+	if (dst->count == 0 || unw_unwind_to_user(info) < 0)
+		return;
+
+	if (dst->pos < 7 * sizeof(int)) {
+		start = dst->pos;
+		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
+				&dst->u.set.kbuf, &dst->u.set.ubuf, buf,
+				0, 7 * sizeof(int));
+		if (dst->ret)
+			return;
+		for (; start < dst->pos; start += sizeof(int))
+			setfpreg(task, start, *((int *)(buf + start)));
+		if (dst->count == 0)
+			return;
+	}
+	if (dst->pos < sizeof(struct ia32_user_i387_struct)) {
+		start = (dst->pos - 7 * sizeof(int)) /
+			sizeof(struct _fpreg_ia32);
+		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
+				&dst->u.set.kbuf, &dst->u.set.ubuf,
+				buf, 7 * sizeof(int),
+				sizeof(struct ia32_user_i387_struct));
+		if (dst->ret)
+			return;
+		pt = task_pt_regs(task);
+		tos = (task->thread.fsr >> 11) & 7;
+		end = (dst->pos - 7 * sizeof(int)) / sizeof(struct _fpreg_ia32);
+		for (; start < end; start++)
+			access_fpreg_ia32(start,
+				(struct _fpreg_ia32 *)buf + start,
+				pt, info->sw, tos, 1);
+		if (dst->count == 0)
+			return;
+	}
+}
+
+#define OFFSET(member) ((int)(offsetof(struct ia32_user_fxsr_struct, member)))
+static void getfpxreg(struct task_struct *task, int start, int end, char *buf)
+{
+	int min_val;
+
+	min_val = min(end, OFFSET(fop));
+	while (start < min_val) {
+		if (start == OFFSET(cwd))
+			*((short *)buf) = task->thread.fcr & 0xffff;
+		else if (start == OFFSET(swd))
+			*((short *)buf) = task->thread.fsr & 0xffff;
+		else if (start == OFFSET(twd))
+			*((short *)buf) = (task->thread.fsr>>16) & 0xffff;
+		buf += 2;
+		start += 2;
+	}
+	/* skip fop element */
+	if (start == OFFSET(fop)) {
+		start += 2;
+		buf += 2;
+	}
+	while (start < end) {
+		if (start == OFFSET(fip))
+			*((int *)buf) = task->thread.fir;
+		else if (start == OFFSET(fcs))
+			*((int *)buf) = (task->thread.fir>>32) & 0xffff;
+		else if (start == OFFSET(foo))
+			*((int *)buf) = task->thread.fdr;
+		else if (start == OFFSET(fos))
+			*((int *)buf) = (task->thread.fdr>>32) & 0xffff;
+		else if (start == OFFSET(mxcsr))
+			*((int *)buf) = ((task->thread.fcr>>32) & 0xff80)
+					 | ((task->thread.fsr>>32) & 0x3f);
+		buf += 4;
+		start += 4;
+	}
+}
+
+static void setfpxreg(struct task_struct *task, int start, int end, char *buf)
+{
+	int min_val, num32;
+	short num;
+	unsigned long num64;
+
+	min_val = min(end, OFFSET(fop));
+	while (start < min_val) {
+		num = *((short *)buf);
+		if (start == OFFSET(cwd)) {
+			task->thread.fcr = (task->thread.fcr & (~0x1f3f))
+						| (num & 0x1f3f);
+		} else if (start == OFFSET(swd)) {
+			task->thread.fsr = (task->thread.fsr & (~0xffff)) | num;
+		} else if (start == OFFSET(twd)) {
+			task->thread.fsr = (task->thread.fsr & (~0xffff0000))
+				| (((int)num) << 16);
+		}
+		buf += 2;
+		start += 2;
+	}
+	/* skip fop element */
+	if (start == OFFSET(fop)) {
+		start += 2;
+		buf += 2;
+	}
+	while (start < end) {
+		num32 = *((int *)buf);
+		if (start == OFFSET(fip))
+			task->thread.fir = (task->thread.fir & (~0xffffffff))
+						 | num32;
+		else if (start == OFFSET(foo))
+			task->thread.fdr = (task->thread.fdr & (~0xffffffff))
+						 | num32;
+		else if (start == OFFSET(mxcsr)) {
+			num64 = num32 & 0xff10;
+			task->thread.fcr = (task->thread.fcr &
+				(~0xff1000000000UL)) | (num64<<32);
+			num64 = num32 & 0x3f;
+			task->thread.fsr = (task->thread.fsr &
+				(~0x3f00000000UL)) | (num64<<32);
+		}
+		buf += 4;
+		start += 4;
+	}
+}
+
+static void do_fpxregs_get(struct unw_frame_info *info, void *arg)
+{
+	struct regset_getset *dst = arg;
+	struct task_struct *task = dst->target;
+	struct pt_regs *pt;
+	char buf[128];
+	int start, end, tos;
+
+	if (dst->count == 0 || unw_unwind_to_user(info) < 0)
+		return;
+	if (dst->pos < OFFSET(st_space[0])) {
+		end = min(dst->pos + dst->count, (unsigned int)32);
+		getfpxreg(task, dst->pos, end, buf);
+		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
+				&dst->u.get.kbuf, &dst->u.get.ubuf, buf,
+				0, OFFSET(st_space[0]));
+		if (dst->ret || dst->count == 0)
+			return;
+	}
+	if (dst->pos < OFFSET(xmm_space[0])) {
+		pt = task_pt_regs(task);
+		tos = (task->thread.fsr >> 11) & 7;
+		end = min(dst->pos + dst->count,
+				(unsigned int)OFFSET(xmm_space[0]));
+		start = (dst->pos - OFFSET(st_space[0])) / 16;
+		end = (end - OFFSET(st_space[0])) / 16;
+		for (; start < end; start++)
+			access_fpreg_ia32(start, buf + 16 * start, pt,
+						info->sw, tos, 0);
+		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
+				&dst->u.get.kbuf, &dst->u.get.ubuf,
+				buf, OFFSET(st_space[0]), OFFSET(xmm_space[0]));
+		if (dst->ret || dst->count == 0)
+			return;
+	}
+	if (dst->pos < OFFSET(padding[0]))
+		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
+				&dst->u.get.kbuf, &dst->u.get.ubuf,
+				&info->sw->f16, OFFSET(xmm_space[0]),
+				OFFSET(padding[0]));
+}
+
+static void do_fpxregs_set(struct unw_frame_info *info, void *arg)
+{
+	struct regset_getset *dst = arg;
+	struct task_struct *task = dst->target;
+	char buf[128];
+	int start, end;
+
+	if (dst->count == 0 || unw_unwind_to_user(info) < 0)
+		return;
+
+	if (dst->pos < OFFSET(st_space[0])) {
+		start = dst->pos;
+		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
+				&dst->u.set.kbuf, &dst->u.set.ubuf,
+				buf, 0, OFFSET(st_space[0]));
+		if (dst->ret)
+			return;
+		setfpxreg(task, start, dst->pos, buf);
+		if (dst->count == 0)
+			return;
+	}
+	if (dst->pos < OFFSET(xmm_space[0])) {
+		struct pt_regs *pt;
+		int tos;
+		pt = task_pt_regs(task);
+		tos = (task->thread.fsr >> 11) & 7;
+		start = (dst->pos - OFFSET(st_space[0])) / 16;
+		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
+				&dst->u.set.kbuf, &dst->u.set.ubuf,
+				buf, OFFSET(st_space[0]), OFFSET(xmm_space[0]));
+		if (dst->ret)
+			return;
+		end = (dst->pos - OFFSET(st_space[0])) / 16;
+		for (; start < end; start++)
+			access_fpreg_ia32(start, buf + 16 * start, pt, info->sw,
+						 tos, 1);
+		if (dst->count == 0)
+			return;
+	}
+	if (dst->pos < OFFSET(padding[0]))
+		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
+				&dst->u.set.kbuf, &dst->u.set.ubuf,
+				&info->sw->f16, OFFSET(xmm_space[0]),
+				 OFFSET(padding[0]));
+}
+#undef OFFSET
+
+static int do_regset_call(void (*call)(struct unw_frame_info *, void *),
+		struct task_struct *target,
+		const struct user_regset *regset,
+		unsigned int pos, unsigned int count,
+		const void *kbuf, const void __user *ubuf)
+{
+	struct regset_getset info = { .target = target, .regset = regset,
+		.pos = pos, .count = count,
+		.u.set = { .kbuf = kbuf, .ubuf = ubuf },
+		.ret = 0 };
+
+	if (target == current)
+		unw_init_running(call, &info);
+	else {
+		struct unw_frame_info ufi;
+		memset(&ufi, 0, sizeof(ufi));
+		unw_init_from_blocked_task(&ufi, target);
+		(*call)(&ufi, &info);
+	}
+
+	return info.ret;
+}
+
+static int ia32_fpregs_get(struct task_struct *target,
+		const struct user_regset *regset,
+		unsigned int pos, unsigned int count,
+		void *kbuf, void __user *ubuf)
+{
+	return do_regset_call(do_fpregs_get, target, regset, pos, count,
+		kbuf, ubuf);
+}
+
+static int ia32_fpregs_set(struct task_struct *target,
+		const struct user_regset *regset,
+		unsigned int pos, unsigned int count,
+		const void *kbuf, const void __user *ubuf)
+{
+	return do_regset_call(do_fpregs_set, target, regset, pos, count,
+		kbuf, ubuf);
+}
+
+static int ia32_fpxregs_get(struct task_struct *target,
+		const struct user_regset *regset,
+		unsigned int pos, unsigned int count,
+		void *kbuf, void __user *ubuf)
+{
+	return do_regset_call(do_fpxregs_get, target, regset, pos, count,
+		kbuf, ubuf);
+}
+
+static int ia32_fpxregs_set(struct task_struct *target,
+		const struct user_regset *regset,
+		unsigned int pos, unsigned int count,
+		const void *kbuf, const void __user *ubuf)
+{
+	return do_regset_call(do_fpxregs_set, target, regset, pos, count,
+		kbuf, ubuf);
+}
+
+static int ia32_genregs_get(struct task_struct *target,
+		const struct user_regset *regset,
+		unsigned int pos, unsigned int count,
+		void *kbuf, void __user *ubuf)
+{
+	if (kbuf) {
+		u32 *kp = kbuf;
+		while (count > 0) {
+			*kp++ = getreg(target, pos);
+			pos += 4;
+			count -= 4;
+		}
+	} else {
+		u32 __user *up = ubuf;
+		while (count > 0) {
+			if (__put_user(getreg(target, pos), up++))
+				return -EFAULT;
+			pos += 4;
+			count -= 4;
+		}
+	}
+	return 0;
+}
+
+static int ia32_genregs_set(struct task_struct *target,
+		const struct user_regset *regset,
+		unsigned int pos, unsigned int count,
+		const void *kbuf, const void __user *ubuf)
+{
+	int ret = 0;
+
+	if (kbuf) {
+		const u32 *kp = kbuf;
+		while (!ret && count > 0) {
+			putreg(target, pos, *kp++);
+			pos += 4;
+			count -= 4;
+		}
+	} else {
+		const u32 __user *up = ubuf;
+		u32 val;
+		while (!ret && count > 0) {
+			ret = __get_user(val, up++);
+			if (!ret)
+				putreg(target, pos, val);
+			pos += 4;
+			count -= 4;
+		}
+	}
+	return ret;
+}
+
+static int ia32_tls_active(struct task_struct *target,
+		const struct user_regset *regset)
+{
+	struct thread_struct *t = &target->thread;
+	int n = GDT_ENTRY_TLS_ENTRIES;
+	while (n > 0 && desc_empty(&t->tls_array[n -1]))
+		--n;
+	return n;
+}
+
+static int ia32_tls_get(struct task_struct *target,
+		const struct user_regset *regset, unsigned int pos,
+		unsigned int count, void *kbuf, void __user *ubuf)
+{
+	const struct desc_struct *tls;
+
+	if (pos > GDT_ENTRY_TLS_ENTRIES * sizeof(struct ia32_user_desc) ||
+			(pos % sizeof(struct ia32_user_desc)) != 0 ||
+			(count % sizeof(struct ia32_user_desc)) != 0)
+		return -EINVAL;
+