NPTL, round one.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
diff --git a/arch/mips/kernel/asm-offsets.c b/arch/mips/kernel/asm-offsets.c
index 2c11abb..af69cdb 100644
--- a/arch/mips/kernel/asm-offsets.c
+++ b/arch/mips/kernel/asm-offsets.c
@@ -95,6 +95,7 @@
 	offset("#define TI_PRE_COUNT       ", struct thread_info, preempt_count);
 	offset("#define TI_ADDR_LIMIT      ", struct thread_info, addr_limit);
 	offset("#define TI_RESTART_BLOCK   ", struct thread_info, restart_block);
+	offset("#define TI_TP_VALUE        ", struct thread_info, tp_value);
 	constant("#define _THREAD_SIZE_ORDER ", THREAD_SIZE_ORDER);
 	constant("#define _THREAD_SIZE       ", THREAD_SIZE);
 	constant("#define _THREAD_MASK       ", THREAD_MASK);
diff --git a/arch/mips/kernel/linux32.c b/arch/mips/kernel/linux32.c
index e8e886d..330cf84 100644
--- a/arch/mips/kernel/linux32.c
+++ b/arch/mips/kernel/linux32.c
@@ -1468,3 +1468,30 @@
 	}
 	return sys_rt_sigtimedwait(uthese, uinfo, uts, sigsetsize);
 }
+
+save_static_function(sys32_clone);
+__attribute_used__ noinline static int
+_sys32_clone(nabi_no_regargs struct pt_regs regs)
+{
+	unsigned long clone_flags;
+	unsigned long newsp;
+	int __user *parent_tidptr, *child_tidptr;
+
+	clone_flags = regs.regs[4];
+	newsp = regs.regs[5];
+	if (!newsp)
+		newsp = regs.regs[29];
+	parent_tidptr = (int *) regs.regs[6];
+
+	/* Use __dummy4 instead of getting it off the stack, so that
+	   syscall() works.  */
+	child_tidptr = (int __user *) __dummy4;
+	return do_fork(clone_flags, newsp, &regs, 0,
+	               parent_tidptr, child_tidptr);
+}
+
+extern asmlinkage void sys_set_thread_area(u32 addr);
+asmlinkage void sys32_set_thread_area(u32 addr)
+{
+	sys_set_thread_area(AA(addr));
+}
diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c
index 2b7a44d..368526a 100644
--- a/arch/mips/kernel/process.c
+++ b/arch/mips/kernel/process.c
@@ -89,6 +89,7 @@
 	struct thread_info *ti = p->thread_info;
 	struct pt_regs *childregs;
 	long childksp;
+	p->set_child_tid = p->clear_child_tid = NULL;
 
 	childksp = (unsigned long)ti + THREAD_SIZE - 32;
 
@@ -134,6 +135,9 @@
 	childregs->cp0_status &= ~(ST0_CU2|ST0_CU1);
 	clear_tsk_thread_flag(p, TIF_USEDFPU);
 
+	if (clone_flags & CLONE_SETTLS)
+		ti->tp_value = regs->regs[7];
+
 	return 0;
 }
 
diff --git a/arch/mips/kernel/ptrace.c b/arch/mips/kernel/ptrace.c
index 2c7fc74..649c90d 100644
--- a/arch/mips/kernel/ptrace.c
+++ b/arch/mips/kernel/ptrace.c
@@ -289,6 +289,11 @@
 		ret = ptrace_detach(child, data);
 		break;
 
+	case PTRACE_GET_THREAD_AREA:
+		ret = put_user(child->thread_info->tp_value,
+				(unsigned long __user *) data);
+		break;
+
 	default:
 		ret = ptrace_request(child, request, addr, data);
 		break;
diff --git a/arch/mips/kernel/ptrace32.c b/arch/mips/kernel/ptrace32.c
index a8a72c9..eb446e5 100644
--- a/arch/mips/kernel/ptrace32.c
+++ b/arch/mips/kernel/ptrace32.c
@@ -268,6 +268,11 @@
 		wake_up_process(child);
 		break;
 
+	case PTRACE_GET_THREAD_AREA:
+		ret = put_user(child->thread_info->tp_value,
+				(unsigned int __user *) (unsigned long) data);
+		break;
+
 	case PTRACE_DETACH: /* detach a process that was attached. */
 		ret = ptrace_detach(child, data);
 		break;
diff --git a/arch/mips/kernel/scall32-o32.S b/arch/mips/kernel/scall32-o32.S
index 9c4bb91..6fa1112 100644
--- a/arch/mips/kernel/scall32-o32.S
+++ b/arch/mips/kernel/scall32-o32.S
@@ -623,6 +623,7 @@
 	sys	sys_add_key		5
 	sys	sys_request_key		4
 	sys	sys_keyctl		5
+	sys	sys_set_thread_area	1
 
 	.endm
 
diff --git a/arch/mips/kernel/scall64-64.S b/arch/mips/kernel/scall64-64.S
index ffb22a2..d11f99b 100644
--- a/arch/mips/kernel/scall64-64.S
+++ b/arch/mips/kernel/scall64-64.S
@@ -449,3 +449,4 @@
 	PTR	sys_add_key
 	PTR	sys_request_key			/* 5240 */
 	PTR	sys_keyctl
+	PTR	sys_set_thread_area
diff --git a/arch/mips/kernel/scall64-n32.S b/arch/mips/kernel/scall64-n32.S
index d912218..ce03041 100644
--- a/arch/mips/kernel/scall64-n32.S
+++ b/arch/mips/kernel/scall64-n32.S
@@ -363,3 +363,4 @@
 	PTR	sys_add_key
 	PTR	sys_request_key
 	PTR	sys_keyctl			/* 6245 */
+	PTR	sys_set_thread_area
diff --git a/arch/mips/kernel/scall64-o32.S b/arch/mips/kernel/scall64-o32.S
index 1017176..f49182ea 100644
--- a/arch/mips/kernel/scall64-o32.S
+++ b/arch/mips/kernel/scall64-o32.S
@@ -322,7 +322,7 @@
 	PTR	sys32_ipc
 	PTR	sys_fsync
 	PTR	sys32_sigreturn
-	PTR	sys_clone			/* 4120 */
+	PTR	sys32_clone			/* 4120 */
 	PTR	sys_setdomainname
 	PTR	sys32_newuname
 	PTR	sys_ni_syscall			/* sys_modify_ldt */
@@ -485,4 +485,5 @@
 	PTR	sys_add_key			/* 4280 */
 	PTR	sys_request_key
 	PTR	sys_keyctl
+	PTR	sys_set_thread_area
 	.size	sys_call_table,.-sys_call_table
diff --git a/arch/mips/kernel/syscall.c b/arch/mips/kernel/syscall.c
index 8fde242..ee98eeb 100644
--- a/arch/mips/kernel/syscall.c
+++ b/arch/mips/kernel/syscall.c
@@ -7,6 +7,7 @@
  * Copyright (C) 1999, 2000 Silicon Graphics, Inc.
  * Copyright (C) 2001 MIPS Technologies, Inc.
  */
+#include <linux/config.h>
 #include <linux/a.out.h>
 #include <linux/errno.h>
 #include <linux/linkage.h>
@@ -176,14 +177,28 @@
 {
 	unsigned long clone_flags;
 	unsigned long newsp;
-	int *parent_tidptr, *child_tidptr;
+	int __user *parent_tidptr, *child_tidptr;
 
 	clone_flags = regs.regs[4];
 	newsp = regs.regs[5];
 	if (!newsp)
 		newsp = regs.regs[29];
-	parent_tidptr = (int *) regs.regs[6];
-	child_tidptr = (int *) regs.regs[7];
+	parent_tidptr = (int __user *) regs.regs[6];
+#ifdef CONFIG_32BIT
+	/* We need to fetch the fifth argument off the stack.  */
+	child_tidptr = NULL;
+	if (clone_flags & (CLONE_CHILD_SETTID | CLONE_CHILD_CLEARTID)) {
+		int __user *__user *usp = (int __user *__user *) regs.regs[29];
+		if (regs.regs[2] == __NR_syscall) {
+			if (get_user (child_tidptr, &usp[5]))
+				return -EFAULT;
+		}
+		else if (get_user (child_tidptr, &usp[4]))
+			return -EFAULT;
+	}
+#else
+	child_tidptr = (int __user *) regs.regs[8];
+#endif
 	return do_fork(clone_flags, newsp, &regs, 0,
 	               parent_tidptr, child_tidptr);
 }
@@ -245,6 +260,16 @@
 	return error;
 }
 
+void sys_set_thread_area(unsigned long addr)
+{
+	struct thread_info *ti = current->thread_info;
+
+	ti->tp_value = addr;
+
+	/* If some future MIPS implementation has this register in hardware,
+	 * we will need to update it here (and in context switches).  */
+}
+
 asmlinkage int _sys_sysmips(int cmd, long arg1, int arg2, int arg3)
 {
 	int	tmp, len;
diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
index 94d9141..15fed02 100644
--- a/arch/mips/kernel/traps.c
+++ b/arch/mips/kernel/traps.c
@@ -360,6 +360,10 @@
 #define OFFSET 0x0000ffff
 #define LL     0xc0000000
 #define SC     0xe0000000
+#define SPEC3  0x7c000000
+#define RD     0x0000f800
+#define FUNC   0x0000003f
+#define RDHWR  0x0000003b
 
 /*
  * The ll_bit is cleared by r*_switch.S
@@ -495,6 +499,37 @@
 	return -EFAULT;			/* Strange things going on ... */
 }
 
+/*
+ * Simulate trapping 'rdhwr' instructions to provide user accessible
+ * registers not implemented in hardware.  The only current use of this
+ * is the thread area pointer.
+ */
+static inline int simulate_rdhwr(struct pt_regs *regs)
+{
+	struct thread_info *ti = current->thread_info;
+	unsigned int opcode;
+
+	if (unlikely(get_insn_opcode(regs, &opcode)))
+		return -EFAULT;
+
+	if (unlikely(compute_return_epc(regs)))
+		return -EFAULT;
+
+	if ((opcode & OPCODE) == SPEC3 && (opcode & FUNC) == RDHWR) {
+		int rd = (opcode & RD) >> 11;
+		int rt = (opcode & RT) >> 16;
+		switch (rd) {
+			case 29:
+				regs->regs[rt] = ti->tp_value;
+				break;
+			default:
+				return -EFAULT;
+		}
+	}
+
+	return 0;
+}
+
 asmlinkage void do_ov(struct pt_regs *regs)
 {
 	siginfo_t info;
@@ -641,6 +676,9 @@
 		if (!simulate_llsc(regs))
 			return;
 
+	if (!simulate_rdhwr(regs))
+		return;
+
 	force_sig(SIGILL, current);
 }
 
@@ -654,11 +692,13 @@
 
 	switch (cpid) {
 	case 0:
-		if (cpu_has_llsc)
-			break;
+		if (!cpu_has_llsc)
+			if (!simulate_llsc(regs))
+				return;
 
-		if (!simulate_llsc(regs))
+		if (!simulate_rdhwr(regs))
 			return;
+
 		break;
 
 	case 1: