Merge branch 'linus' into x86/i8259
[sfrench/cifs-2.6.git] / arch / x86 / kernel / traps_32.c
index b22c01e05a1841d4a3960a2e62a2881ee9c5babc..08d752de4eee307779ba20b31c3b1e7c006daa62 100644 (file)
@@ -9,26 +9,28 @@
  * 'Traps.c' handles hardware traps and faults after we have saved some
  * state in 'asm.s'.
  */
-#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <linux/kallsyms.h>
+#include <linux/spinlock.h>
+#include <linux/highmem.h>
+#include <linux/kprobes.h>
+#include <linux/uaccess.h>
+#include <linux/utsname.h>
+#include <linux/kdebug.h>
 #include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/ptrace.h>
 #include <linux/string.h>
+#include <linux/unwind.h>
+#include <linux/delay.h>
 #include <linux/errno.h>
+#include <linux/kexec.h>
+#include <linux/sched.h>
 #include <linux/timer.h>
-#include <linux/mm.h>
 #include <linux/init.h>
-#include <linux/delay.h>
-#include <linux/spinlock.h>
-#include <linux/interrupt.h>
-#include <linux/highmem.h>
-#include <linux/kallsyms.h>
-#include <linux/ptrace.h>
-#include <linux/utsname.h>
-#include <linux/kprobes.h>
-#include <linux/kexec.h>
-#include <linux/unwind.h>
-#include <linux/uaccess.h>
-#include <linux/nmi.h>
 #include <linux/bug.h>
+#include <linux/nmi.h>
+#include <linux/mm.h>
 
 #ifdef CONFIG_EISA
 #include <linux/ioport.h>
 #include <linux/edac.h>
 #endif
 
+#include <asm/arch_hooks.h>
+#include <asm/stacktrace.h>
 #include <asm/processor.h>
-#include <asm/system.h>
-#include <asm/io.h>
-#include <asm/atomic.h>
 #include <asm/debugreg.h>
+#include <asm/atomic.h>
+#include <asm/system.h>
+#include <asm/unwind.h>
 #include <asm/desc.h>
 #include <asm/i387.h>
 #include <asm/nmi.h>
-#include <asm/unwind.h>
 #include <asm/smp.h>
-#include <asm/arch_hooks.h>
-#include <linux/kdebug.h>
-#include <asm/stacktrace.h>
-
-#include <linux/module.h>
+#include <asm/io.h>
 
 #include "mach_traps.h"
 
@@ -69,7 +68,7 @@ EXPORT_SYMBOL_GPL(used_vectors);
 asmlinkage int system_call(void);
 
 /* Do we ignore FPU interrupts ? */
-char ignore_fpu_irq = 0;
+char ignore_fpu_irq;
 
 /*
  * The IDT has to be page-aligned to simplify the Pentium
@@ -105,12 +104,13 @@ static unsigned int code_bytes = 64;
 void printk_address(unsigned long address, int reliable)
 {
 #ifdef CONFIG_KALLSYMS
-       unsigned long offset = 0, symsize;
+       char namebuf[KSYM_NAME_LEN];
+       unsigned long offset = 0;
+       unsigned long symsize;
        const char *symname;
-       char *modname;
-       char *delim = ":";
-       char namebuf[128];
        char reliab[4] = "";
+       char *delim = ":";
+       char *modname;
 
        symname = kallsyms_lookup(address, &symsize, &offset,
                                        &modname, namebuf);
@@ -138,13 +138,14 @@ static inline int valid_stack_ptr(struct thread_info *tinfo, void *p, unsigned s
 
 /* The form of the top of the frame on the stack */
 struct stack_frame {
-       struct stack_frame *next_frame;
-       unsigned long return_address;
+       struct stack_frame      *next_frame;
+       unsigned long           return_address;
 };
 
-static inline unsigned long print_context_stack(struct thread_info *tinfo,
-                               unsigned long *stack, unsigned long bp,
-                               const struct stacktrace_ops *ops, void *data)
+static inline unsigned long
+print_context_stack(struct thread_info *tinfo,
+                   unsigned long *stack, unsigned long bp,
+                   const struct stacktrace_ops *ops, void *data)
 {
        struct stack_frame *frame = (struct stack_frame *)bp;
 
@@ -166,7 +167,7 @@ static inline unsigned long print_context_stack(struct thread_info *tinfo,
        return bp;
 }
 
-#define MSG(msg) ops->warning(data, msg)
+#define MSG(msg)               ops->warning(data, msg)
 
 void dump_trace(struct task_struct *task, struct pt_regs *regs,
                unsigned long *stack, unsigned long bp,
@@ -177,6 +178,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs,
 
        if (!stack) {
                unsigned long dummy;
+
                stack = &dummy;
                if (task != current)
                        stack = (unsigned long *)task->thread.sp;
@@ -186,7 +188,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs,
        if (!bp) {
                if (task == current) {
                        /* Grab bp right from our regs */
-                       asm ("movl %%ebp, %0" : "=r" (bp) : );
+                       asm("movl %%ebp, %0" : "=r" (bp) :);
                } else {
                        /* bp is the last reg pushed by switch_to */
                        bp = *(unsigned long *) task->thread.sp;
@@ -196,15 +198,18 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs,
 
        while (1) {
                struct thread_info *context;
+
                context = (struct thread_info *)
                        ((unsigned long)stack & (~(THREAD_SIZE - 1)));
                bp = print_context_stack(context, stack, bp, ops, data);
-               /* Should be after the line below, but somewhere
-                  in early boot context comes out corrupted and we
-                  can't reference it -AK */
+               /*
+                * Should be after the line below, but somewhere
+                * in early boot context comes out corrupted and we
+                * can't reference it:
+                */
                if (ops->stack(data, "IRQ") < 0)
                        break;
-               stack = (unsigned long*)context->previous_esp;
+               stack = (unsigned long *)context->previous_esp;
                if (!stack)
                        break;
                touch_nmi_watchdog();
@@ -243,15 +248,15 @@ static void print_trace_address(void *data, unsigned long addr, int reliable)
 }
 
 static const struct stacktrace_ops print_trace_ops = {
-       .warning = print_trace_warning,
-       .warning_symbol = print_trace_warning_symbol,
-       .stack = print_trace_stack,
-       .address = print_trace_address,
+       .warning                = print_trace_warning,
+       .warning_symbol         = print_trace_warning_symbol,
+       .stack                  = print_trace_stack,
+       .address                = print_trace_address,
 };
 
 static void
 show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
-               unsigned long *stack, unsigned long bp, char *log_lvl)
+                  unsigned long *stack, unsigned long bp, char *log_lvl)
 {
        dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl);
        printk("%s =======================\n", log_lvl);
@@ -263,21 +268,22 @@ void show_trace(struct task_struct *task, struct pt_regs *regs,
        show_trace_log_lvl(task, regs, stack, bp, "");
 }
 
-static void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs,
-                      unsigned long *sp, unsigned long bp, char *log_lvl)
+static void
+show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs,
+                  unsigned long *sp, unsigned long bp, char *log_lvl)
 {
        unsigned long *stack;
        int i;
 
        if (sp == NULL) {
                if (task)
-                       sp = (unsigned long*)task->thread.sp;
+                       sp = (unsigned long *)task->thread.sp;
                else
                        sp = (unsigned long *)&sp;
        }
 
        stack = sp;
-       for(i = 0; i < kstack_depth_to_print; i++) {
+       for (i = 0; i < kstack_depth_to_print; i++) {
                if (kstack_end(stack))
                        break;
                if (i && ((i % 8) == 0))
@@ -285,6 +291,7 @@ static void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs,
                printk("%08lx ", *stack++);
        }
        printk("\n%sCall Trace:\n", log_lvl);
+
        show_trace_log_lvl(task, regs, sp, bp, log_lvl);
 }
 
@@ -299,8 +306,8 @@ void show_stack(struct task_struct *task, unsigned long *sp)
  */
 void dump_stack(void)
 {
-       unsigned long stack;
        unsigned long bp = 0;
+       unsigned long stack;
 
 #ifdef CONFIG_FRAME_POINTER
        if (!bp)
@@ -312,6 +319,7 @@ void dump_stack(void)
                init_utsname()->release,
                (int)strcspn(init_utsname()->version, " "),
                init_utsname()->version);
+
        show_trace(current, NULL, &stack, bp);
 }
 
@@ -323,6 +331,7 @@ void show_registers(struct pt_regs *regs)
 
        print_modules();
        __show_registers(regs, 0);
+
        printk(KERN_EMERG "Process %.*s (pid: %d, ti=%p task=%p task.ti=%p)",
                TASK_COMM_LEN, current->comm, task_pid_nr(current),
                current_thread_info(), current, task_thread_info(current));
@@ -331,10 +340,10 @@ void show_registers(struct pt_regs *regs)
         * time of the fault..
         */
        if (!user_mode_vm(regs)) {
-               u8 *ip;
                unsigned int code_prologue = code_bytes * 43 / 64;
                unsigned int code_len = code_bytes;
                unsigned char c;
+               u8 *ip;
 
                printk("\n" KERN_EMERG "Stack: ");
                show_stack_log_lvl(NULL, regs, &regs->sp, 0, KERN_EMERG);
@@ -361,7 +370,7 @@ void show_registers(struct pt_regs *regs)
                }
        }
        printk("\n");
-}      
+}
 
 int is_valid_bugaddr(unsigned long ip)
 {
@@ -377,10 +386,10 @@ int is_valid_bugaddr(unsigned long ip)
 
 static int die_counter;
 
-int __kprobes __die(const char * str, struct pt_regs * regs, long err)
+int __kprobes __die(const char *str, struct pt_regs *regs, long err)
 {
-       unsigned long sp;
        unsigned short ss;
+       unsigned long sp;
 
        printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter);
 #ifdef CONFIG_PREEMPT
@@ -395,8 +404,8 @@ int __kprobes __die(const char * str, struct pt_regs * regs, long err)
        printk("\n");
 
        if (notify_die(DIE_OOPS, str, regs, err,
-                               current->thread.trap_no, SIGSEGV) !=
-                       NOTIFY_STOP) {
+                       current->thread.trap_no, SIGSEGV) != NOTIFY_STOP) {
+
                show_registers(regs);
                /* Executive summary in case the oops scrolled away */
                sp = (unsigned long) (&regs->sp);
@@ -408,17 +417,18 @@ int __kprobes __die(const char * str, struct pt_regs * regs, long err)
                printk(KERN_EMERG "EIP: [<%08lx>] ", regs->ip);
                print_symbol("%s", regs->ip);
                printk(" SS:ESP %04x:%08lx\n", ss, sp);
+
                return 0;
-       } else {
-               return 1;
        }
+
+       return 1;
 }
 
 /*
- * This is gone through when something in the kernel has done something bad and
- * is about to be terminated.
+ * This is gone through when something in the kernel has done something bad
+ * and is about to be terminated:
  */
-void die(const char * str, struct pt_regs * regs, long err)
+void die(const char *str, struct pt_regs *regs, long err)
 {
        static struct {
                raw_spinlock_t lock;
@@ -440,8 +450,9 @@ void die(const char * str, struct pt_regs * regs, long err)
                die.lock_owner = smp_processor_id();
                die.lock_owner_depth = 0;
                bust_spinlocks(1);
-       } else
+       } else {
                raw_local_irq_save(flags);
+       }
 
        if (++die.lock_owner_depth < 3) {
                report_bug(regs->ip, regs);
@@ -474,19 +485,20 @@ void die(const char * str, struct pt_regs * regs, long err)
        do_exit(SIGSEGV);
 }
 
-static inline void die_if_kernel(const char * str, struct pt_regs * regs, long err)
+static inline void
+die_if_kernel(const char *str, struct pt_regs *regs, long err)
 {
        if (!user_mode_vm(regs))
                die(str, regs, err);
 }
 
-static void __kprobes do_trap(int trapnr, int signr, char *str, int vm86,
-                             struct pt_regs * regs, long error_code,
-                             siginfo_t *info)
+static void __kprobes
+do_trap(int trapnr, int signr, char *str, int vm86, struct pt_regs *regs,
+       long error_code, siginfo_t *info)
 {
        struct task_struct *tsk = current;
 
-       if (regs->flags & VM_MASK) {
+       if (regs->flags & X86_VM_MASK) {
                if (vm86)
                        goto vm86_trap;
                goto trap_signal;
@@ -495,111 +507,113 @@ static void __kprobes do_trap(int trapnr, int signr, char *str, int vm86,
        if (!user_mode(regs))
                goto kernel_trap;
 
-       trap_signal: {
-               /*
-                * We want error_code and trap_no set for userspace faults and
-                * kernelspace faults which result in die(), but not
-                * kernelspace faults which are fixed up.  die() gives the
-                * process no chance to handle the signal and notice the
-                * kernel fault information, so that won't result in polluting
-                * the information about previously queued, but not yet
-                * delivered, faults.  See also do_general_protection below.
-                */
-               tsk->thread.error_code = error_code;
-               tsk->thread.trap_no = trapnr;
+trap_signal:
+       /*
+        * We want error_code and trap_no set for userspace faults and
+        * kernelspace faults which result in die(), but not
+        * kernelspace faults which are fixed up.  die() gives the
+        * process no chance to handle the signal and notice the
+        * kernel fault information, so that won't result in polluting
+        * the information about previously queued, but not yet
+        * delivered, faults.  See also do_general_protection below.
+        */
+       tsk->thread.error_code = error_code;
+       tsk->thread.trap_no = trapnr;
 
-               if (info)
-                       force_sig_info(signr, info, tsk);
-               else
-                       force_sig(signr, tsk);
-               return;
-       }
+       if (info)
+               force_sig_info(signr, info, tsk);
+       else
+               force_sig(signr, tsk);
+       return;
 
-       kernel_trap: {
-               if (!fixup_exception(regs)) {
-                       tsk->thread.error_code = error_code;
-                       tsk->thread.trap_no = trapnr;
-                       die(str, regs, error_code);
-               }
-               return;
+kernel_trap:
+       if (!fixup_exception(regs)) {
+               tsk->thread.error_code = error_code;
+               tsk->thread.trap_no = trapnr;
+               die(str, regs, error_code);
        }
+       return;
 
-       vm86_trap: {
-               int ret = handle_vm86_trap((struct kernel_vm86_regs *) regs, error_code, trapnr);
-               if (ret) goto trap_signal;
-               return;
-       }
+vm86_trap:
+       if (handle_vm86_trap((struct kernel_vm86_regs *) regs,
+                                               error_code, trapnr))
+               goto trap_signal;
+       return;
 }
 
-#define DO_ERROR(trapnr, signr, str, name) \
-void do_##name(struct pt_regs * regs, long error_code) \
-{ \
-       if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
-                                               == NOTIFY_STOP) \
-               return; \
-       do_trap(trapnr, signr, str, 0, regs, error_code, NULL); \
+#define DO_ERROR(trapnr, signr, str, name)                             \
+void do_##name(struct pt_regs *regs, long error_code)                  \
+{                                                                      \
+       trace_hardirqs_fixup();                                         \
+       if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr)  \
+                                               == NOTIFY_STOP)         \
+               return;                                                 \
+       do_trap(trapnr, signr, str, 0, regs, error_code, NULL);         \
 }
 
-#define DO_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr, irq) \
-void do_##name(struct pt_regs * regs, long error_code) \
-{ \
-       siginfo_t info; \
-       if (irq) \
-               local_irq_enable(); \
-       info.si_signo = signr; \
-       info.si_errno = 0; \
-       info.si_code = sicode; \
-       info.si_addr = (void __user *)siaddr; \
-       if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
-                                               == NOTIFY_STOP) \
-               return; \
-       do_trap(trapnr, signr, str, 0, regs, error_code, &info); \
+#define DO_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr, irq)   \
+void do_##name(struct pt_regs *regs, long error_code)                  \
+{                                                                      \
+       siginfo_t info;                                                 \
+       if (irq)                                                        \
+               local_irq_enable();                                     \
+       info.si_signo = signr;                                          \
+       info.si_errno = 0;                                              \
+       info.si_code = sicode;                                          \
+       info.si_addr = (void __user *)siaddr;                           \
+       if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr)  \
+                                               == NOTIFY_STOP)         \
+               return;                                                 \
+       do_trap(trapnr, signr, str, 0, regs, error_code, &info);        \
 }
 
-#define DO_VM86_ERROR(trapnr, signr, str, name) \
-void do_##name(struct pt_regs * regs, long error_code) \
-{ \
-       if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
-                                               == NOTIFY_STOP) \
-               return; \
-       do_trap(trapnr, signr, str, 1, regs, error_code, NULL); \
+#define DO_VM86_ERROR(trapnr, signr, str, name)                                \
+void do_##name(struct pt_regs *regs, long error_code)                  \
+{                                                                      \
+       if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr)  \
+                                               == NOTIFY_STOP)         \
+               return;                                                 \
+       do_trap(trapnr, signr, str, 1, regs, error_code, NULL);         \
 }
 
-#define DO_VM86_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr) \
-void do_##name(struct pt_regs * regs, long error_code) \
-{ \
-       siginfo_t info; \
-       info.si_signo = signr; \
-       info.si_errno = 0; \
-       info.si_code = sicode; \
-       info.si_addr = (void __user *)siaddr; \
-       trace_hardirqs_fixup(); \
-       if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
-                                               == NOTIFY_STOP) \
-               return; \
-       do_trap(trapnr, signr, str, 1, regs, error_code, &info); \
+#define DO_VM86_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr)   \
+void do_##name(struct pt_regs *regs, long error_code)                  \
+{                                                                      \
+       siginfo_t info;                                                 \
+       info.si_signo = signr;                                          \
+       info.si_errno = 0;                                              \
+       info.si_code = sicode;                                          \
+       info.si_addr = (void __user *)siaddr;                           \
+       trace_hardirqs_fixup();                                         \
+       if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr)  \
+                                               == NOTIFY_STOP)         \
+               return;                                                 \
+       do_trap(trapnr, signr, str, 1, regs, error_code, &info);        \
 }
 
-DO_VM86_ERROR_INFO( 0, SIGFPE,  "divide error", divide_error, FPE_INTDIV, regs->ip)
+DO_VM86_ERROR_INFO(0, SIGFPE,  "divide error", divide_error, FPE_INTDIV, regs->ip)
 #ifndef CONFIG_KPROBES
-DO_VM86_ERROR( 3, SIGTRAP, "int3", int3)
+DO_VM86_ERROR(3, SIGTRAP, "int3", int3)
 #endif
-DO_VM86_ERROR( 4, SIGSEGV, "overflow", overflow)
-DO_VM86_ERROR( 5, SIGSEGV, "bounds", bounds)
-DO_ERROR_INFO( 6, SIGILL,  "invalid opcode", invalid_op, ILL_ILLOPN, regs->ip, 0)
-DO_ERROR( 9, SIGFPE,  "coprocessor segment overrun", coprocessor_segment_overrun)
+DO_VM86_ERROR(4, SIGSEGV, "overflow", overflow)
+DO_VM86_ERROR(5, SIGSEGV, "bounds", bounds)
+DO_ERROR_INFO(6, SIGILL,  "invalid opcode", invalid_op, ILL_ILLOPN, regs->ip, 0)
+DO_ERROR(9, SIGFPE,  "coprocessor segment overrun", coprocessor_segment_overrun)
 DO_ERROR(10, SIGSEGV, "invalid TSS", invalid_TSS)
 DO_ERROR(11, SIGBUS,  "segment not present", segment_not_present)
 DO_ERROR(12, SIGBUS,  "stack segment", stack_segment)
 DO_ERROR_INFO(17, SIGBUS, "alignment check", alignment_check, BUS_ADRALN, 0, 0)
-DO_ERROR_INFO(32, SIGSEGV, "iret exception", iret_error, ILL_BADSTK, 0, 1)
+DO_ERROR_INFO(32, SIGILL, "iret exception", iret_error, ILL_BADSTK, 0, 1)
 
-void __kprobes do_general_protection(struct pt_regs * regs,
-                                             long error_code)
+void __kprobes do_general_protection(struct pt_regs *regs, long error_code)
 {
-       int cpu = get_cpu();
-       struct tss_struct *tss = &per_cpu(init_tss, cpu);
-       struct thread_struct *thread = &current->thread;
+       struct thread_struct *thread;
+       struct tss_struct *tss;
+       int cpu;
+
+       cpu = get_cpu();
+       tss = &per_cpu(init_tss, cpu);
+       thread = &current->thread;
 
        /*
         * Perform the lazy TSS's I/O bitmap copy. If the TSS has an
@@ -616,19 +630,21 @@ void __kprobes do_general_protection(struct pt_regs * regs,
                 * If the previously set map was extending to higher ports
                 * than the current one, pad extra space with 0xff (no access).
                 */
-               if (thread->io_bitmap_max < tss->io_bitmap_max)
+               if (thread->io_bitmap_max < tss->io_bitmap_max) {
                        memset((char *) tss->io_bitmap +
                                thread->io_bitmap_max, 0xff,
                                tss->io_bitmap_max - thread->io_bitmap_max);
+               }
                tss->io_bitmap_max = thread->io_bitmap_max;
                tss->x86_tss.io_bitmap_base = IO_BITMAP_OFFSET;
                tss->io_bitmap_owner = thread;
                put_cpu();
+
                return;
        }
        put_cpu();
 
-       if (regs->flags & VM_MASK)
+       if (regs->flags & X86_VM_MASK)
                goto gp_in_vm86;
 
        if (!user_mode(regs))
@@ -636,6 +652,7 @@ void __kprobes do_general_protection(struct pt_regs * regs,
 
        current->thread.error_code = error_code;
        current->thread.trap_no = 13;
+
        if (show_unhandled_signals && unhandled_signal(current, SIGSEGV) &&
            printk_ratelimit()) {
                printk(KERN_INFO
@@ -665,22 +682,25 @@ gp_in_kernel:
        }
 }
 
-static __kprobes void
-mem_parity_error(unsigned char reason, struct pt_regs * regs)
+static notrace __kprobes void
+mem_parity_error(unsigned char reason, struct pt_regs *regs)
 {
-       printk(KERN_EMERG "Uhhuh. NMI received for unknown reason %02x on "
-               "CPU %d.\n", reason, smp_processor_id());
-       printk(KERN_EMERG "You have some hardware problem, likely on the PCI bus.\n");
+       printk(KERN_EMERG
+               "Uhhuh. NMI received for unknown reason %02x on CPU %d.\n",
+                       reason, smp_processor_id());
+
+       printk(KERN_EMERG
+               "You have some hardware problem, likely on the PCI bus.\n");
 
 #if defined(CONFIG_EDAC)
-       if(edac_handler_set()) {
+       if (edac_handler_set()) {
                edac_atomic_assert_error();
                return;
        }
 #endif
 
        if (panic_on_unrecovered_nmi)
-                panic("NMI: Not continuing");
+               panic("NMI: Not continuing");
 
        printk(KERN_EMERG "Dazed and confused, but trying to continue\n");
 
@@ -688,8 +708,8 @@ mem_parity_error(unsigned char reason, struct pt_regs * regs)
        clear_mem_error(reason);
 }
 
-static __kprobes void
-io_check_error(unsigned char reason, struct pt_regs * regs)
+static notrace __kprobes void
+io_check_error(unsigned char reason, struct pt_regs *regs)
 {
        unsigned long i;
 
@@ -699,44 +719,52 @@ io_check_error(unsigned char reason, struct pt_regs * regs)
        /* Re-enable the IOCK line, wait for a few seconds */
        reason = (reason & 0xf) | 8;
        outb(reason, 0x61);
+
        i = 2000;
-       while (--i) udelay(1000);
+       while (--i)
+               udelay(1000);
+
        reason &= ~8;
        outb(reason, 0x61);
 }
 
-static __kprobes void
-unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
+static notrace __kprobes void
+unknown_nmi_error(unsigned char reason, struct pt_regs *regs)
 {
+       if (notify_die(DIE_NMIUNKNOWN, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP)
+               return;
 #ifdef CONFIG_MCA
-       /* Might actually be able to figure out what the guilty party
-       * is. */
-       if( MCA_bus ) {
+       /*
+        * Might actually be able to figure out what the guilty party
+        * is:
+        */
+       if (MCA_bus) {
                mca_handle_nmi();
                return;
        }
 #endif
-       printk(KERN_EMERG "Uhhuh. NMI received for unknown reason %02x on "
-               "CPU %d.\n", reason, smp_processor_id());
+       printk(KERN_EMERG
+               "Uhhuh. NMI received for unknown reason %02x on CPU %d.\n",
+                       reason, smp_processor_id());
+
        printk(KERN_EMERG "Do you have a strange power saving mode enabled?\n");
        if (panic_on_unrecovered_nmi)
-                panic("NMI: Not continuing");
+               panic("NMI: Not continuing");
 
        printk(KERN_EMERG "Dazed and confused, but trying to continue\n");
 }
 
 static DEFINE_SPINLOCK(nmi_print_lock);
 
-void __kprobes die_nmi(struct pt_regs *regs, const char *msg)
+void notrace __kprobes die_nmi(struct pt_regs *regs, const char *msg)
 {
-       if (notify_die(DIE_NMIWATCHDOG, msg, regs, 0, 2, SIGINT) ==
-           NOTIFY_STOP)
+       if (notify_die(DIE_NMIWATCHDOG, msg, regs, 0, 2, SIGINT) == NOTIFY_STOP)
                return;
 
        spin_lock(&nmi_print_lock);
        /*
        * We are in trouble anyway, lets at least try
-       * to get a message out.
+       * to get a message out:
        */
        bust_spinlocks(1);
        printk(KERN_EMERG "%s", msg);
@@ -747,9 +775,10 @@ void __kprobes die_nmi(struct pt_regs *regs, const char *msg)
        spin_unlock(&nmi_print_lock);
        bust_spinlocks(0);
 
-       /* If we are in kernel we are probably nested up pretty bad
-        * and might aswell get out now while we still can.
-       */
+       /*
+        * If we are in kernel we are probably nested up pretty bad
+        * and might aswell get out now while we still can:
+        */
        if (!user_mode_vm(regs)) {
                current->thread.trap_no = 2;
                crash_kexec(regs);
@@ -758,14 +787,14 @@ void __kprobes die_nmi(struct pt_regs *regs, const char *msg)
        do_exit(SIGSEGV);
 }
 
-static __kprobes void default_do_nmi(struct pt_regs * regs)
+static notrace __kprobes void default_do_nmi(struct pt_regs *regs)
 {
        unsigned char reason = 0;
 
-       /* Only the BSP gets external NMIs from the system */
+       /* Only the BSP gets external NMIs from the system: */
        if (!smp_processor_id())
                reason = get_nmi_reason();
+
        if (!(reason & 0xc0)) {
                if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 2, SIGINT)
                                                        == NOTIFY_STOP)
@@ -778,8 +807,10 @@ static __kprobes void default_do_nmi(struct pt_regs * regs)
                if (nmi_watchdog_tick(regs, reason))
                        return;
                if (!do_nmi_callback(regs, smp_processor_id()))
-#endif
                        unknown_nmi_error(reason, regs);
+#else
+               unknown_nmi_error(reason, regs);
+#endif
 
                return;
        }
@@ -791,14 +822,14 @@ static __kprobes void default_do_nmi(struct pt_regs * regs)
                io_check_error(reason, regs);
        /*
         * Reassert NMI in case it became active meanwhile
-        * as it's edge-triggered.
+        * as it's edge-triggered:
         */
        reassert_nmi();
 }
 
 static int ignore_nmis;
 
-__kprobes void do_nmi(struct pt_regs * regs, long error_code)
+notrace __kprobes void do_nmi(struct pt_regs *regs, long error_code)
 {
        int cpu;
 
@@ -834,9 +865,12 @@ void __kprobes do_int3(struct pt_regs *regs, long error_code)
        if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP)
                        == NOTIFY_STOP)
                return;
-       /* This is an interrupt gate, because kprobes wants interrupts
-       disabled.  Normal trap handlers don't. */
+       /*
+        * This is an interrupt gate, because kprobes wants interrupts
+        * disabled. Normal trap handlers don't.
+        */
        restore_interrupts(regs);
+
        do_trap(3, SIGTRAP, "int3", 1, regs, error_code, NULL);
 }
 #endif
@@ -851,7 +885,7 @@ void __kprobes do_int3(struct pt_regs *regs, long error_code)
  * from user space. Such code must not hold kernel locks (since it
  * can equally take a page fault), therefore it is safe to call
  * force_sig_info even though that claims and releases locks.
- * 
+ *
  * Code in ./signal.c ensures that the debug control register
  * is restored before we deliver any signal, and therefore that
  * user code runs with the correct debug control register even though
@@ -863,10 +897,10 @@ void __kprobes do_int3(struct pt_regs *regs, long error_code)
  * find every occurrence of the TF bit that could be saved away even
  * by user code)
  */
-void __kprobes do_debug(struct pt_regs * regs, long error_code)
+void __kprobes do_debug(struct pt_regs *regs, long error_code)
 {
-       unsigned int condition;
        struct task_struct *tsk = current;
+       unsigned int condition;
 
        trace_hardirqs_fixup();
 
@@ -891,7 +925,7 @@ void __kprobes do_debug(struct pt_regs * regs, long error_code)
                        goto clear_dr7;
        }
 
-       if (regs->flags & VM_MASK)
+       if (regs->flags & X86_VM_MASK)
                goto debug_vm86;
 
        /* Save debug status register where ptrace can see it */
@@ -914,7 +948,8 @@ void __kprobes do_debug(struct pt_regs * regs, long error_code)
        /* Ok, finally something we can handle */
        send_sigtrap(tsk, regs, error_code);
 
-       /* Disable additional traps. They'll be re-enabled when
+       /*
+        * Disable additional traps. They'll be re-enabled when
         * the signal is delivered.
         */
 clear_dr7:
@@ -927,7 +962,7 @@ debug_vm86:
 
 clear_TF_reenable:
        set_tsk_thread_flag(tsk, TIF_SINGLESTEP);
-       regs->flags &= ~TF_MASK;
+       regs->flags &= ~X86_EFLAGS_TF;
        return;
 }
 
@@ -938,9 +973,10 @@ clear_TF_reenable:
  */
 void math_error(void __user *ip)
 {
-       struct task_struct * task;
+       struct task_struct *task;
+       unsigned short cwd;
+       unsigned short swd;
        siginfo_t info;
-       unsigned short cwd, swd;
 
        /*
         * Save the info for the exception handler and clear the error.
@@ -966,36 +1002,36 @@ void math_error(void __user *ip)
        cwd = get_fpu_cwd(task);
        swd = get_fpu_swd(task);
        switch (swd & ~cwd & 0x3f) {
-               case 0x000: /* No unmasked exception */
-                       return;
-               default:    /* Multiple exceptions */
-                       break;
-               case 0x001: /* Invalid Op */
-                       /*
-                        * swd & 0x240 == 0x040: Stack Underflow
-                        * swd & 0x240 == 0x240: Stack Overflow
-                        * User must clear the SF bit (0x40) if set
-                        */
-                       info.si_code = FPE_FLTINV;
-                       break;
-               case 0x002: /* Denormalize */
-               case 0x010: /* Underflow */
-                       info.si_code = FPE_FLTUND;
-                       break;
-               case 0x004: /* Zero Divide */
-                       info.si_code = FPE_FLTDIV;
-                       break;
-               case 0x008: /* Overflow */
-                       info.si_code = FPE_FLTOVF;
-                       break;
-               case 0x020: /* Precision */
-                       info.si_code = FPE_FLTRES;
-                       break;
+       case 0x000: /* No unmasked exception */
+               return;
+       default:    /* Multiple exceptions */
+               break;
+       case 0x001: /* Invalid Op */
+               /*
+                * swd & 0x240 == 0x040: Stack Underflow
+                * swd & 0x240 == 0x240: Stack Overflow
+                * User must clear the SF bit (0x40) if set
+                */
+               info.si_code = FPE_FLTINV;
+               break;
+       case 0x002: /* Denormalize */
+       case 0x010: /* Underflow */
+               info.si_code = FPE_FLTUND;
+               break;
+       case 0x004: /* Zero Divide */
+               info.si_code = FPE_FLTDIV;
+               break;
+       case 0x008: /* Overflow */
+               info.si_code = FPE_FLTOVF;
+               break;
+       case 0x020: /* Precision */
+               info.si_code = FPE_FLTRES;
+               break;
        }
        force_sig_info(SIGFPE, &info, task);
 }
 
-void do_coprocessor_error(struct pt_regs * regs, long error_code)
+void do_coprocessor_error(struct pt_regs *regs, long error_code)
 {
        ignore_fpu_irq = 1;
        math_error((void __user *)regs->ip);
@@ -1003,9 +1039,9 @@ void do_coprocessor_error(struct pt_regs * regs, long error_code)
 
 static void simd_math_error(void __user *ip)
 {
-       struct task_struct * task;
-       siginfo_t info;
+       struct task_struct *task;
        unsigned short mxcsr;
+       siginfo_t info;
 
        /*
         * Save the info for the exception handler and clear the error.
@@ -1026,82 +1062,80 @@ static void simd_math_error(void __user *ip)
         */
        mxcsr = get_fpu_mxcsr(task);
        switch (~((mxcsr & 0x1f80) >> 7) & (mxcsr & 0x3f)) {
-               case 0x000:
-               default:
-                       break;
-               case 0x001: /* Invalid Op */
-                       info.si_code = FPE_FLTINV;
-                       break;
-               case 0x002: /* Denormalize */
-               case 0x010: /* Underflow */
-                       info.si_code = FPE_FLTUND;
-                       break;
-               case 0x004: /* Zero Divide */
-                       info.si_code = FPE_FLTDIV;
-                       break;
-               case 0x008: /* Overflow */
-                       info.si_code = FPE_FLTOVF;
-                       break;
-               case 0x020: /* Precision */
-                       info.si_code = FPE_FLTRES;
-                       break;
+       case 0x000:
+       default:
+               break;
+       case 0x001: /* Invalid Op */
+               info.si_code = FPE_FLTINV;
+               break;
+       case 0x002: /* Denormalize */
+       case 0x010: /* Underflow */
+               info.si_code = FPE_FLTUND;
+               break;
+       case 0x004: /* Zero Divide */
+               info.si_code = FPE_FLTDIV;
+               break;
+       case 0x008: /* Overflow */
+               info.si_code = FPE_FLTOVF;
+               break;
+       case 0x020: /* Precision */
+               info.si_code = FPE_FLTRES;
+               break;
        }
        force_sig_info(SIGFPE, &info, task);
 }
 
-void do_simd_coprocessor_error(struct pt_regs * regs,
-                                         long error_code)
+void do_simd_coprocessor_error(struct pt_regs *regs, long error_code)
 {
        if (cpu_has_xmm) {
                /* Handle SIMD FPU exceptions on PIII+ processors. */
                ignore_fpu_irq = 1;
                simd_math_error((void __user *)regs->ip);
-       } else {
-               /*
-                * Handle strange cache flush from user space exception
-                * in all other cases.  This is undocumented behaviour.
-                */
-               if (regs->flags & VM_MASK) {
-                       handle_vm86_fault((struct kernel_vm86_regs *)regs,
-                                         error_code);
-                       return;
-               }
-               current->thread.trap_no = 19;
-               current->thread.error_code = error_code;
-               die_if_kernel("cache flush denied", regs, error_code);
-               force_sig(SIGSEGV, current);
+               return;
        }
+       /*
+        * Handle strange cache flush from user space exception
+        * in all other cases.  This is undocumented behaviour.
+        */
+       if (regs->flags & X86_VM_MASK) {
+               handle_vm86_fault((struct kernel_vm86_regs *)regs, error_code);
+               return;
+       }
+       current->thread.trap_no = 19;
+       current->thread.error_code = error_code;
+       die_if_kernel("cache flush denied", regs, error_code);
+       force_sig(SIGSEGV, current);
 }
 
-void do_spurious_interrupt_bug(struct pt_regs * regs,
-                                         long error_code)
+void do_spurious_interrupt_bug(struct pt_regs *regs, long error_code)
 {
 #if 0
        /* No need to warn about this any longer. */
-       printk("Ignoring P6 Local APIC Spurious Interrupt Bug...\n");
+       printk(KERN_INFO "Ignoring P6 Local APIC Spurious Interrupt Bug...\n");
 #endif
 }
 
-unsigned long patch_espfix_desc(unsigned long uesp,
-                                         unsigned long kesp)
+unsigned long patch_espfix_desc(unsigned long uesp, unsigned long kesp)
 {
        struct desc_struct *gdt = __get_cpu_var(gdt_page).gdt;
        unsigned long base = (kesp - uesp) & -THREAD_SIZE;
        unsigned long new_kesp = kesp - base;
        unsigned long lim_pages = (new_kesp | (THREAD_SIZE - 1)) >> PAGE_SHIFT;
        __u64 desc = *(__u64 *)&gdt[GDT_ENTRY_ESPFIX_SS];
+
        /* Set up base for espfix segment */
-       desc &= 0x00f0ff0000000000ULL;
-       desc |= ((((__u64)base) << 16) & 0x000000ffffff0000ULL) |
+       desc &= 0x00f0ff0000000000ULL;
+       desc |= ((((__u64)base) << 16) & 0x000000ffffff0000ULL) |
                ((((__u64)base) << 32) & 0xff00000000000000ULL) |
                ((((__u64)lim_pages) << 32) & 0x000f000000000000ULL) |
                (lim_pages & 0xffff);
        *(__u64 *)&gdt[GDT_ENTRY_ESPFIX_SS] = desc;
+
        return new_kesp;
 }
 
 /*
- *  'math_state_restore()' saves the current math information in the
+ * 'math_state_restore()' saves the current math information in the
  * old math state array, and gets the new ones from the current task
  *
  * Careful.. There are problems with IBM-designed IRQ13 behaviour.
@@ -1115,9 +1149,22 @@ asmlinkage void math_state_restore(void)
        struct thread_info *thread = current_thread_info();
        struct task_struct *tsk = thread->task;
 
-       clts();         /* Allow maths ops (or we recurse) */
-       if (!tsk_used_math(tsk))
-               init_fpu(tsk);
+       if (!tsk_used_math(tsk)) {
+               local_irq_enable();
+               /*
+                * does a slab alloc which can sleep
+                */
+               if (init_fpu(tsk)) {
+                       /*
+                        * ran out of memory!
+                        */
+                       do_group_exit(SIGKILL);
+                       return;
+               }
+               local_irq_disable();
+       }
+
+       clts();                         /* Allow maths ops (or we recurse) */
        restore_fpu(tsk);
        thread->status |= TS_USEDFPU;   /* So we fnsave on switch_to() */
        tsk->fpu_counter++;
@@ -1128,80 +1175,76 @@ EXPORT_SYMBOL_GPL(math_state_restore);
 
 asmlinkage void math_emulate(long arg)
 {
-       printk(KERN_EMERG "math-emulation not enabled and no coprocessor found.\n");
-       printk(KERN_EMERG "killing %s.\n",current->comm);
-       force_sig(SIGFPE,current);
+       printk(KERN_EMERG
+               "math-emulation not enabled and no coprocessor found.\n");
+       printk(KERN_EMERG "killing %s.\n", current->comm);
+       force_sig(SIGFPE, current);
        schedule();
 }
 
 #endif /* CONFIG_MATH_EMULATION */
 
-
 void __init trap_init(void)
 {
        int i;
 
 #ifdef CONFIG_EISA
        void __iomem *p = early_ioremap(0x0FFFD9, 4);
-       if (readl(p) == 'E'+('I'<<8)+('S'<<16)+('A'<<24)) {
+
+       if (readl(p) == 'E' + ('I'<<8) + ('S'<<16) + ('A'<<24))
                EISA_bus = 1;
-       }
        early_iounmap(p, 4);
 #endif
 
 #ifdef CONFIG_X86_LOCAL_APIC
        init_apic_mappings();
 #endif
-
-       set_trap_gate(0,&divide_error);
-       set_intr_gate(1,&debug);
-       set_intr_gate(2,&nmi);
+       set_trap_gate(0,  &divide_error);
+       set_intr_gate(1,  &debug);
+       set_intr_gate(2,  &nmi);
        set_system_intr_gate(3, &int3); /* int3/4 can be called from all */
-       set_system_gate(4,&overflow);
-       set_trap_gate(5,&bounds);
-       set_trap_gate(6,&invalid_op);
-       set_trap_gate(7,&device_not_available);
-       set_task_gate(8,GDT_ENTRY_DOUBLEFAULT_TSS);
-       set_trap_gate(9,&coprocessor_segment_overrun);
-       set_trap_gate(10,&invalid_TSS);
-       set_trap_gate(11,&segment_not_present);
-       set_trap_gate(12,&stack_segment);
-       set_trap_gate(13,&general_protection);
-       set_intr_gate(14,&page_fault);
-       set_trap_gate(15,&spurious_interrupt_bug);
-       set_trap_gate(16,&coprocessor_error);
-       set_trap_gate(17,&alignment_check);
+       set_system_gate(4, &overflow);
+       set_trap_gate(5,  &bounds);
+       set_trap_gate(6,  &invalid_op);
+       set_trap_gate(7,  &device_not_available);
+       set_task_gate(8,  GDT_ENTRY_DOUBLEFAULT_TSS);
+       set_trap_gate(9,  &coprocessor_segment_overrun);
+       set_trap_gate(10, &invalid_TSS);
+       set_trap_gate(11, &segment_not_present);
+       set_trap_gate(12, &stack_segment);
+       set_trap_gate(13, &general_protection);
+       set_intr_gate(14, &page_fault);
+       set_trap_gate(15, &spurious_interrupt_bug);
+       set_trap_gate(16, &coprocessor_error);
+       set_trap_gate(17, &alignment_check);
 #ifdef CONFIG_X86_MCE
-       set_trap_gate(18,&machine_check);
+       set_trap_gate(18, &machine_check);
 #endif
-       set_trap_gate(19,&simd_coprocessor_error);
+       set_trap_gate(19, &simd_coprocessor_error);
 
-       /*
-        * Verify that the FXSAVE/FXRSTOR data will be 16-byte aligned.
-        * Generate a build-time error if the alignment is wrong.
-        */
-       BUILD_BUG_ON(offsetof(struct task_struct, thread.i387.fxsave) & 15);
        if (cpu_has_fxsr) {
                printk(KERN_INFO "Enabling fast FPU save and restore... ");
                set_in_cr4(X86_CR4_OSFXSR);
                printk("done.\n");
        }
        if (cpu_has_xmm) {
-               printk(KERN_INFO "Enabling unmasked SIMD FPU exception "
-                               "support... ");
+               printk(KERN_INFO
+                       "Enabling unmasked SIMD FPU exception support... ");
                set_in_cr4(X86_CR4_OSXMMEXCPT);
                printk("done.\n");
        }
 
-       set_system_gate(SYSCALL_VECTOR,&system_call);
+       set_system_gate(SYSCALL_VECTOR, &system_call);
 
-       /* Reserve all the builtin and the syscall vector. */
+       /* Reserve all the builtin and the syscall vector: */
        for (i = 0; i < FIRST_EXTERNAL_VECTOR; i++)
                set_bit(i, used_vectors);
+
        set_bit(SYSCALL_VECTOR, used_vectors);
 
+       init_thread_xstate();
        /*
-        * Should be a barrier for any external CPU state.
+        * Should be a barrier for any external CPU state:
         */
        cpu_init();
 
@@ -1211,6 +1254,7 @@ void __init trap_init(void)
 static int __init kstack_setup(char *s)
 {
        kstack_depth_to_print = simple_strtoul(s, NULL, 0);
+
        return 1;
 }
 __setup("kstack=", kstack_setup);