1 files changed, 98 insertions, 75 deletions

diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index 5459d148a0f..c091527efd8 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c
@@ -98,18 +98,14 @@ static void pmac_backlight_unblank(void)
 static inline void pmac_backlight_unblank(void) { }
 #endif
 
-int die(const char *str, struct pt_regs *regs, long err)
+static arch_spinlock_t die_lock = __ARCH_SPIN_LOCK_UNLOCKED;
+static int die_owner = -1;
+static unsigned int die_nest_count;
+static int die_counter;
+
+static unsigned __kprobes long oops_begin(struct pt_regs *regs)
 {
-	static struct {
-		raw_spinlock_t lock;
-		u32 lock_owner;
-		int lock_owner_depth;
-	} die = {
-		.lock =			__RAW_SPIN_LOCK_UNLOCKED(die.lock),
-		.lock_owner =		-1,
-		.lock_owner_depth =	0
-	};
-	static int die_counter;
+	int cpu;
 	unsigned long flags;
 
 	if (debugger(regs))
@@ -117,66 +113,109 @@ int die(const char *str, struct pt_regs *regs, long err)
 
 	oops_enter();
 
-	if (die.lock_owner != raw_smp_processor_id()) {
-		console_verbose();
-		raw_spin_lock_irqsave(&die.lock, flags);
-		die.lock_owner = smp_processor_id();
-		die.lock_owner_depth = 0;
-		bust_spinlocks(1);
-		if (machine_is(powermac))
-			pmac_backlight_unblank();
-	} else {
-		local_save_flags(flags);
+	/* racy, but better than risking deadlock. */
+	raw_local_irq_save(flags);
+	cpu = smp_processor_id();
+	if (!arch_spin_trylock(&die_lock)) {
+		if (cpu == die_owner)
+			/* nested oops. should stop eventually */;
+		else
+			arch_spin_lock(&die_lock);
 	}
+	die_nest_count++;
+	die_owner = cpu;
+	console_verbose();
+	bust_spinlocks(1);
+	if (machine_is(powermac))
+		pmac_backlight_unblank();
+	return flags;
+}
 
-	if (++die.lock_owner_depth < 3) {
-		printk("Oops: %s, sig: %ld [#%d]\n", str, err, ++die_counter);
-#ifdef CONFIG_PREEMPT
-		printk("PREEMPT ");
-#endif
-#ifdef CONFIG_SMP
-		printk("SMP NR_CPUS=%d ", NR_CPUS);
-#endif
-#ifdef CONFIG_DEBUG_PAGEALLOC
-		printk("DEBUG_PAGEALLOC ");
-#endif
-#ifdef CONFIG_NUMA
-		printk("NUMA ");
-#endif
-		printk("%s\n", ppc_md.name ? ppc_md.name : "");
+static void __kprobes oops_end(unsigned long flags, struct pt_regs *regs,
+			       int signr)
+{
+	bust_spinlocks(0);
+	die_owner = -1;
+	add_taint(TAINT_DIE);
+	die_nest_count--;
+	oops_exit();
+	printk("\n");
+	if (!die_nest_count)
+		/* Nest count reaches zero, release the lock. */
+		arch_spin_unlock(&die_lock);
+	raw_local_irq_restore(flags);
 
-		if (notify_die(DIE_OOPS, str, regs, err, 255,
-			       SIGSEGV) == NOTIFY_STOP)
-			return 1;
+	/*
+	 * A system reset (0x100) is a request to dump, so we always send
+	 * it through the crashdump code.
+	 */
+	if (kexec_should_crash(current) || (TRAP(regs) == 0x100)) {
+		crash_kexec(regs);
 
-		print_modules();
-		show_regs(regs);
-	} else {
-		printk("Recursive die() failure, output suppressed\n");
+		/*
+		 * We aren't the primary crash CPU. We need to send it
+		 * to a holding pattern to avoid it ending up in the panic
+		 * code.
+		 */
+		crash_kexec_secondary(regs);
 	}
 
-	bust_spinlocks(0);
-	die.lock_owner = -1;
-	add_taint(TAINT_DIE);
-	raw_spin_unlock_irqrestore(&die.lock, flags);
+	if (!signr)
+		return;
 
-	if (kexec_should_crash(current) ||
-		kexec_sr_activated(smp_processor_id()))
-		crash_kexec(regs);
-	crash_kexec_secondary(regs);
+	/*
+	 * While our oops output is serialised by a spinlock, output
+	 * from panic() called below can race and corrupt it. If we
+	 * know we are going to panic, delay for 1 second so we have a
+	 * chance to get clean backtraces from all CPUs that are oopsing.
+	 */
+	if (in_interrupt() || panic_on_oops || !current->pid ||
+	    is_global_init(current)) {
+		mdelay(MSEC_PER_SEC);
+	}
 
 	if (in_interrupt())
 		panic("Fatal exception in interrupt");
-
 	if (panic_on_oops)
 		panic("Fatal exception");
+	do_exit(signr);
+}
 
-	oops_exit();
-	do_exit(err);
+static int __kprobes __die(const char *str, struct pt_regs *regs, long err)
+{
+	printk("Oops: %s, sig: %ld [#%d]\n", str, err, ++die_counter);
+#ifdef CONFIG_PREEMPT
+	printk("PREEMPT ");
+#endif
+#ifdef CONFIG_SMP
+	printk("SMP NR_CPUS=%d ", NR_CPUS);
+#endif
+#ifdef CONFIG_DEBUG_PAGEALLOC
+	printk("DEBUG_PAGEALLOC ");
+#endif
+#ifdef CONFIG_NUMA
+	printk("NUMA ");
+#endif
+	printk("%s\n", ppc_md.name ? ppc_md.name : "");
+
+	if (notify_die(DIE_OOPS, str, regs, err, 255, SIGSEGV) == NOTIFY_STOP)
+		return 1;
+
+	print_modules();
+	show_regs(regs);
 
 	return 0;
 }
 
+void die(const char *str, struct pt_regs *regs, long err)
+{
+	unsigned long flags = oops_begin(regs);
+
+	if (__die(str, regs, err))
+		err = 0;
+	oops_end(flags, regs, err);
+}
+
 void user_single_step_siginfo(struct task_struct *tsk,
 				struct pt_regs *regs, siginfo_t *info)
 {
@@ -195,10 +234,11 @@ void _exception(int signr, struct pt_regs *regs, int code, unsigned long addr)
 			"at %016lx nip %016lx lr %016lx code %x\n";
 
 	if (!user_mode(regs)) {
-		if (die("Exception in kernel mode", regs, signr))
-			return;
-	} else if (show_unhandled_signals &&
-		   unhandled_signal(current, signr)) {
+		die("Exception in kernel mode", regs, signr);
+		return;
+	}
+
+	if (show_unhandled_signals && unhandled_signal(current, signr)) {
 		printk_ratelimited(regs->msr & MSR_64BIT ? fmt64 : fmt32,
 				   current->comm, current->pid, signr,
 				   addr, regs->nip, regs->link, code);
@@ -220,25 +260,8 @@ void system_reset_exception(struct pt_regs *regs)
 			return;
 	}
 
-#ifdef CONFIG_KEXEC
-	cpumask_set_cpu(smp_processor_id(), &cpus_in_sr);
-#endif
-
 	die("System Reset", regs, SIGABRT);
 
-	/*
-	 * Some CPUs when released from the debugger will execute this path.
-	 * These CPUs entered the debugger via a soft-reset. If the CPU was
-	 * hung before entering the debugger it will return to the hung
-	 * state when exiting this function.  This causes a problem in
-	 * kdump since the hung CPU(s) will not respond to the IPI sent
-	 * from kdump. To prevent the problem we call crash_kexec_secondary()
-	 * here. If a kdump had not been initiated or we exit the debugger
-	 * with the "exit and recover" command (x) crash_kexec_secondary()
-	 * will return after 5ms and the CPU returns to its previous state.
-	 */
-	crash_kexec_secondary(regs);
-
 	/* Must die if the interrupt is not recoverable */
 	if (!(regs->msr & MSR_RI))
 		panic("Unrecoverable System Reset");