Merge branches 'core/signal' and 'x86/spinlocks' into x86/xen

Conflicts:
	include/asm-x86/spinlock.h

arch/ia64/include/asm/siginfo.h

@@ -113,11 +113,6 @@ typedef struct siginfo {
 #undef NSIGSEGV
 #define NSIGSEGV	3
 
-/*
- * SIGTRAP si_codes
- */
-#define TRAP_BRANCH	(__SI_FAULT|3)	/* process taken branch trap */
-#define TRAP_HWBKPT	(__SI_FAULT|4)	/* hardware breakpoint or watchpoint */
 #undef NSIGTRAP
 #define NSIGTRAP	4
 

arch/powerpc/include/asm/siginfo.h

@@ -15,11 +15,6 @@
 
 #include <asm-generic/siginfo.h>
 
-/*
- * SIGTRAP si_codes
- */
-#define TRAP_BRANCH	(__SI_FAULT|3)	/* process taken branch trap */
-#define TRAP_HWBKPT	(__SI_FAULT|4)	/* hardware breakpoint or watchpoint */
 #undef NSIGTRAP
 #define NSIGTRAP	4
 

arch/x86/ia32/ia32_signal.c

@@ -351,31 +351,28 @@ static int ia32_setup_sigcontext(struct sigcontext_ia32 __user *sc,
 	savesegment(es, tmp);
 	err |= __put_user(tmp, (unsigned int __user *)&sc->es);
 
-	err |= __put_user((u32)regs->di, &sc->di);
-	err |= __put_user((u32)regs->si, &sc->si);
-	err |= __put_user((u32)regs->bp, &sc->bp);
-	err |= __put_user((u32)regs->sp, &sc->sp);
-	err |= __put_user((u32)regs->bx, &sc->bx);
-	err |= __put_user((u32)regs->dx, &sc->dx);
-	err |= __put_user((u32)regs->cx, &sc->cx);
-	err |= __put_user((u32)regs->ax, &sc->ax);
-	err |= __put_user((u32)regs->cs, &sc->cs);
-	err |= __put_user((u32)regs->ss, &sc->ss);
+	err |= __put_user(regs->di, &sc->di);
+	err |= __put_user(regs->si, &sc->si);
+	err |= __put_user(regs->bp, &sc->bp);
+	err |= __put_user(regs->sp, &sc->sp);
+	err |= __put_user(regs->bx, &sc->bx);
+	err |= __put_user(regs->dx, &sc->dx);
+	err |= __put_user(regs->cx, &sc->cx);
+	err |= __put_user(regs->ax, &sc->ax);
+	err |= __put_user(regs->cs, &sc->cs);
+	err |= __put_user(regs->ss, &sc->ss);
 	err |= __put_user(current->thread.trap_no, &sc->trapno);
 	err |= __put_user(current->thread.error_code, &sc->err);
-	err |= __put_user((u32)regs->ip, &sc->ip);
-	err |= __put_user((u32)regs->flags, &sc->flags);
-	err |= __put_user((u32)regs->sp, &sc->sp_at_signal);
+	err |= __put_user(regs->ip, &sc->ip);
+	err |= __put_user(regs->flags, &sc->flags);
+	err |= __put_user(regs->sp, &sc->sp_at_signal);
 
 	tmp = save_i387_xstate_ia32(fpstate);
 	if (tmp < 0)
 		err = -EFAULT;
-	else {
-		clear_used_math();
-		stts();
+	else
 		err |= __put_user(ptr_to_compat(tmp ? fpstate : NULL),
 					&sc->fpstate);
-	}
 
 	/* non-iBCS2 extensions.. */
 	err |= __put_user(mask, &sc->oldmask);
@@ -444,21 +441,18 @@ int ia32_setup_frame(int sig, struct k_sigaction *ka,
 	frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate);
 
 	if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame)))
-		goto give_sigsegv;
+		return -EFAULT;
 
-	err |= __put_user(sig, &frame->sig);
-	if (err)
-		goto give_sigsegv;
+	if (__put_user(sig, &frame->sig))
+		return -EFAULT;
 
-	err |= ia32_setup_sigcontext(&frame->sc, fpstate, regs, set->sig[0]);
-	if (err)
-		goto give_sigsegv;
+	if (ia32_setup_sigcontext(&frame->sc, fpstate, regs, set->sig[0]))
+		return -EFAULT;
 
 	if (_COMPAT_NSIG_WORDS > 1) {
-		err |= __copy_to_user(frame->extramask, &set->sig[1],
-				      sizeof(frame->extramask));
-		if (err)
-			goto give_sigsegv;
+		if (__copy_to_user(frame->extramask, &set->sig[1],
+				   sizeof(frame->extramask)))
+			return -EFAULT;
 	}
 
 	if (ka->sa.sa_flags & SA_RESTORER) {
@@ -479,7 +473,7 @@ int ia32_setup_frame(int sig, struct k_sigaction *ka,
 	 */
 	err |= __copy_to_user(frame->retcode, &code, 8);
 	if (err)
-		goto give_sigsegv;
+		return -EFAULT;
 
 	/* Set up registers for signal handler */
 	regs->sp = (unsigned long) frame;
@@ -502,10 +496,6 @@ int ia32_setup_frame(int sig, struct k_sigaction *ka,
 #endif
 
 	return 0;
-
-give_sigsegv:
-	force_sigsegv(sig, current);
-	return -EFAULT;
 }
 
 int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
@@ -533,14 +523,14 @@ int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
 	frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate);
 
 	if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame)))
-		goto give_sigsegv;
+		return -EFAULT;
 
 	err |= __put_user(sig, &frame->sig);
 	err |= __put_user(ptr_to_compat(&frame->info), &frame->pinfo);
 	err |= __put_user(ptr_to_compat(&frame->uc), &frame->puc);
 	err |= copy_siginfo_to_user32(&frame->info, info);
 	if (err)
-		goto give_sigsegv;
+		return -EFAULT;
 
 	/* Create the ucontext.  */
 	if (cpu_has_xsave)
@@ -556,7 +546,7 @@ int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
 				     regs, set->sig[0]);
 	err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set));
 	if (err)
-		goto give_sigsegv;
+		return -EFAULT;
 
 	if (ka->sa.sa_flags & SA_RESTORER)
 		restorer = ka->sa.sa_restorer;
@@ -571,7 +561,7 @@ int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
 	 */
 	err |= __copy_to_user(frame->retcode, &code, 8);
 	if (err)
-		goto give_sigsegv;
+		return -EFAULT;
 
 	/* Set up registers for signal handler */
 	regs->sp = (unsigned long) frame;
@@ -599,8 +589,4 @@ int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
 #endif
 
 	return 0;
-
-give_sigsegv:
-	force_sigsegv(sig, current);
-	return -EFAULT;
 }

arch/x86/kernel/process_64.c

@@ -736,12 +736,12 @@ unsigned long get_wchan(struct task_struct *p)
 	if (!p || p == current || p->state == TASK_RUNNING)
 		return 0;
 	stack = (unsigned long)task_stack_page(p);
-	if (p->thread.sp < stack || p->thread.sp > stack+THREAD_SIZE)
+	if (p->thread.sp < stack || p->thread.sp >= stack+THREAD_SIZE)
 		return 0;
 	fp = *(u64 *)(p->thread.sp);
 	do {
 		if (fp < (unsigned long)stack ||
-		    fp > (unsigned long)stack+THREAD_SIZE)
+		    fp >= (unsigned long)stack+THREAD_SIZE)
 			return 0;
 		ip = *(u64 *)(fp+8);
 		if (!in_sched_functions(ip))

arch/x86/kernel/ptrace.c

@@ -1452,7 +1452,8 @@ const struct user_regset_view *task_user_regset_view(struct task_struct *task)
 #endif
 }
 
-void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code)
+void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs,
+					 int error_code, int si_code)
 {
 	struct siginfo info;
 
@@ -1461,7 +1462,7 @@ void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code)
 
 	memset(&info, 0, sizeof(info));
 	info.si_signo = SIGTRAP;
-	info.si_code = TRAP_BRKPT;
+	info.si_code = si_code;
 
 	/* User-mode ip? */
 	info.si_addr = user_mode_vm(regs) ? (void __user *) regs->ip : NULL;
@@ -1548,5 +1549,5 @@ asmregparm void syscall_trace_leave(struct pt_regs *regs)
 	 */
 	if (test_thread_flag(TIF_SINGLESTEP) &&
 	    tracehook_consider_fatal_signal(current, SIGTRAP, SIG_DFL))
-		send_sigtrap(current, regs, 0);
+		send_sigtrap(current, regs, 0, TRAP_BRKPT);
 }

arch/x86/kernel/traps_32.c

@@ -891,6 +891,7 @@ void __kprobes do_debug(struct pt_regs *regs, long error_code)
 {
 	struct task_struct *tsk = current;
 	unsigned int condition;
+	int si_code;
 
 	trace_hardirqs_fixup();
 
@@ -935,8 +936,9 @@ void __kprobes do_debug(struct pt_regs *regs, long error_code)
 			goto clear_TF_reenable;
 	}
 
+	si_code = get_si_code((unsigned long)condition);
 	/* Ok, finally something we can handle */
-	send_sigtrap(tsk, regs, error_code);
+	send_sigtrap(tsk, regs, error_code, si_code);
 
 	/*
 	 * Disable additional traps. They'll be re-enabled when

arch/x86/kernel/traps_64.c

@@ -940,7 +940,7 @@ asmlinkage void __kprobes do_debug(struct pt_regs *regs,
 	tsk->thread.error_code = error_code;
 	info.si_signo = SIGTRAP;
 	info.si_errno = 0;
-	info.si_code = TRAP_BRKPT;
+	info.si_code = get_si_code(condition);
 	info.si_addr = user_mode(regs) ? (void __user *)regs->ip : NULL;
 	force_sig_info(SIGTRAP, &info, tsk);
 

arch/x86/kernel/vmlinux_64.lds.S

@@ -172,8 +172,8 @@ SECTIONS
   .x86_cpu_dev.init : AT(ADDR(.x86_cpu_dev.init) - LOAD_OFFSET) {
 	*(.x86_cpu_dev.init)
   }
-  SECURITY_INIT
   __x86_cpu_dev_end = .;
+  SECURITY_INIT
 
   . = ALIGN(8);
   .parainstructions : AT(ADDR(.parainstructions) - LOAD_OFFSET) {

arch/x86/mm/ioremap.c

@@ -614,7 +614,7 @@ void __init *early_ioremap(unsigned long phys_addr, unsigned long size)
 	 */
 	offset = phys_addr & ~PAGE_MASK;
 	phys_addr &= PAGE_MASK;
-	size = PAGE_ALIGN(last_addr) - phys_addr;
+	size = PAGE_ALIGN(last_addr + 1) - phys_addr;
 
 	/*
 	 * Mappings have to fit in the FIX_BTMAP area.

include/asm-generic/siginfo.h

@@ -199,6 +199,8 @@ typedef struct siginfo {
  */
 #define TRAP_BRKPT	(__SI_FAULT|1)	/* process breakpoint */
 #define TRAP_TRACE	(__SI_FAULT|2)	/* process trace trap */
+#define TRAP_BRANCH     (__SI_FAULT|3)  /* process taken branch trap */
+#define TRAP_HWBKPT     (__SI_FAULT|4)  /* hardware breakpoint/watchpoint */
 #define NSIGTRAP	2
 
 /*

include/asm-parisc/siginfo.h

@@ -3,11 +3,6 @@
 
 #include <asm-generic/siginfo.h>
 
-/*
- * SIGTRAP si_codes
- */
-#define TRAP_BRANCH	(__SI_FAULT|3)	/* process taken branch trap */
-#define TRAP_HWBKPT	(__SI_FAULT|4)	/* hardware breakpoint or watchpoint */
 #undef NSIGTRAP
 #define NSIGTRAP	4
 

include/asm-x86/ptrace.h

@@ -177,11 +177,11 @@ convert_ip_to_linear(struct task_struct *child, struct pt_regs *regs);
 
 #ifdef CONFIG_X86_32
 extern void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs,
-			 int error_code);
-#else
-void signal_fault(struct pt_regs *regs, void __user *frame, char *where);
+			 int error_code, int si_code);
 #endif
 
+void signal_fault(struct pt_regs *regs, void __user *frame, char *where);
+
 extern long syscall_trace_enter(struct pt_regs *);
 extern void syscall_trace_leave(struct pt_regs *);
 

include/asm-x86/spinlock.h

@@ -21,8 +21,10 @@
 
 #ifdef CONFIG_X86_32
 # define LOCK_PTR_REG "a"
+# define REG_PTR_MODE "k"
 #else
 # define LOCK_PTR_REG "D"
+# define REG_PTR_MODE "q"
 #endif
 
 #if defined(CONFIG_X86_32) && \
@@ -54,19 +56,7 @@
  * much between them in performance though, especially as locks are out of line.
  */
 #if (NR_CPUS < 256)
-static inline int __ticket_spin_is_locked(raw_spinlock_t *lock)
-{
-	int tmp = ACCESS_ONCE(lock->slock);
-
-	return (((tmp >> 8) & 0xff) != (tmp & 0xff));
-}
-
-static inline int __ticket_spin_is_contended(raw_spinlock_t *lock)
-{
-	int tmp = ACCESS_ONCE(lock->slock);
-
-	return (((tmp >> 8) - tmp) & 0xff) > 1;
-}
+#define TICKET_SHIFT 8
 
 static __always_inline void __ticket_spin_lock(raw_spinlock_t *lock)
 {
@@ -89,19 +79,17 @@ static __always_inline void __ticket_spin_lock(raw_spinlock_t *lock)
 
 static __always_inline int __ticket_spin_trylock(raw_spinlock_t *lock)
 {
-	int tmp;
-	short new;
+	int tmp, new;
 
-	asm volatile("movw %2,%w0\n\t"
+	asm volatile("movzwl %2, %0\n\t"
 		     "cmpb %h0,%b0\n\t"
+		     "leal 0x100(%" REG_PTR_MODE "0), %1\n\t"
 		     "jne 1f\n\t"
-		     "movw %w0,%w1\n\t"
-		     "incb %h1\n\t"
 		     LOCK_PREFIX "cmpxchgw %w1,%2\n\t"
 		     "1:"
 		     "sete %b1\n\t"
 		     "movzbl %b1,%0\n\t"
-		     : "=&a" (tmp), "=Q" (new), "+m" (lock->slock)
+		     : "=&a" (tmp), "=&q" (new), "+m" (lock->slock)
 		     :
 		     : "memory", "cc");
 
@@ -116,19 +104,7 @@ static __always_inline void __ticket_spin_unlock(raw_spinlock_t *lock)
 		     : "memory", "cc");
 }
 #else
-static inline int __ticket_spin_is_locked(raw_spinlock_t *lock)
-{
-	int tmp = ACCESS_ONCE(lock->slock);
-
-	return (((tmp >> 16) & 0xffff) != (tmp & 0xffff));
-}
-
-static inline int __ticket_spin_is_contended(raw_spinlock_t *lock)
-{
-	int tmp = ACCESS_ONCE(lock->slock);
-
-	return (((tmp >> 16) - tmp) & 0xffff) > 1;
-}
+#define TICKET_SHIFT 16
 
 static __always_inline void __ticket_spin_lock(raw_spinlock_t *lock)
 {
@@ -146,7 +122,7 @@ static __always_inline void __ticket_spin_lock(raw_spinlock_t *lock)
 		     /* don't need lfence here, because loads are in-order */
 		     "jmp 1b\n"
 		     "2:"
-		     : "+Q" (inc), "+m" (lock->slock), "=r" (tmp)
+		     : "+r" (inc), "+m" (lock->slock), "=&r" (tmp)
 		     :
 		     : "memory", "cc");
 }
@@ -160,13 +136,13 @@ static __always_inline int __ticket_spin_trylock(raw_spinlock_t *lock)
 		     "movl %0,%1\n\t"
 		     "roll $16, %0\n\t"
 		     "cmpl %0,%1\n\t"
+		     "leal 0x00010000(%" REG_PTR_MODE "0), %1\n\t"
 		     "jne 1f\n\t"
-		     "addl $0x00010000, %1\n\t"
 		     LOCK_PREFIX "cmpxchgl %1,%2\n\t"
 		     "1:"
 		     "sete %b1\n\t"
 		     "movzbl %b1,%0\n\t"
-		     : "=&a" (tmp), "=r" (new), "+m" (lock->slock)
+		     : "=&a" (tmp), "=&q" (new), "+m" (lock->slock)
 		     :
 		     : "memory", "cc");
 
@@ -182,6 +158,20 @@ static __always_inline void __ticket_spin_unlock(raw_spinlock_t *lock)
 }
 #endif
 
+static inline int __ticket_spin_is_locked(raw_spinlock_t *lock)
+{
+	int tmp = ACCESS_ONCE(lock->slock);
+
+	return !!(((tmp >> TICKET_SHIFT) ^ tmp) & ((1 << TICKET_SHIFT) - 1));
+}
+
+static inline int __ticket_spin_is_contended(raw_spinlock_t *lock)
+{
+	int tmp = ACCESS_ONCE(lock->slock);
+
+	return (((tmp >> TICKET_SHIFT) - tmp) & ((1 << TICKET_SHIFT) - 1)) > 1;
+}
+
 #ifdef CONFIG_PARAVIRT
 /*
  * Define virtualization-friendly old-style lock byte lock, for use in

include/asm-x86/traps.h

 #ifndef ASM_X86__TRAPS_H
 #define ASM_X86__TRAPS_H
 
+#include <asm/debugreg.h>
+
 /* Common in X86_32 and X86_64 */
 asmlinkage void divide_error(void);
 asmlinkage void debug(void);
@@ -36,6 +38,16 @@ void do_invalid_op(struct pt_regs *, long);
 void do_general_protection(struct pt_regs *, long);
 void do_nmi(struct pt_regs *, long);
 
+static inline int get_si_code(unsigned long condition)
+{
+	if (condition & DR_STEP)
+		return TRAP_TRACE;
+	else if (condition & (DR_TRAP0|DR_TRAP1|DR_TRAP2|DR_TRAP3))
+		return TRAP_HWBKPT;
+	else
+		return TRAP_BRKPT;
+}
+
 extern int panic_on_unrecovered_nmi;
 extern int kstack_depth_to_print;