[S390] nohz: Fix __udelay.

This fixes a regression that came with 934b2857
("[S390] nohz/sclp: disable timer on synchronous waits.").
If udelay() gets called from a disabled context it sets the clock comparator
to a value where it expects the next interrupt. When the interrupt happens
the clock comparator gets not reset and therefore the interrupt condition
doesn't get cleared. The result is an endless timer interrupt loop.

In addition this patch fixes also the following:

rcutorture reveals that our __udelay implementation is still buggy,
since it might schedule tasklets, but prevents their execution:

NOHZ: local_softirq_pending 42
NOHZ: local_softirq_pending 02
NOHZ: local_softirq_pending 142
NOHZ: local_softirq_pending 02

To fix this we make sure that only the clock comparator interrupt
is enabled when the enabled wait psw is loaded.
Also no code gets called anymore which might schedule tasklets.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

arch/s390/kernel/time.c

@@ -169,6 +169,8 @@ void init_cpu_timer(void)
 
 static void clock_comparator_interrupt(__u16 code)
 {
+	if (S390_lowcore.clock_comparator == -1ULL)
+		set_clock_comparator(S390_lowcore.clock_comparator);
 }
 
 static void etr_timing_alert(struct etr_irq_parm *);

arch/s390/lib/delay.c

 /*
- *  arch/s390/lib/delay.c
  *    Precise Delay Loops for S390
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
- *
- *  Derived from "arch/i386/lib/delay.c"
- *    Copyright (C) 1993 Linus Torvalds
- *    Copyright (C) 1997 Martin Mares <mj@atrey.karlin.mff.cuni.cz>
+ *    Copyright IBM Corp. 1999,2008
+ *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>,
+ *		 Heiko Carstens <heiko.carstens@de.ibm.com>,
  */
 
 #include <linux/sched.h>
@@ -29,30 +24,31 @@ void __delay(unsigned long loops)
 	asm volatile("0: brct %0,0b" : : "d" ((loops/2) + 1));
 }
 
-/*
- * Waits for 'usecs' microseconds using the TOD clock comparator.
- */
-void __udelay(unsigned long usecs)
+static void __udelay_disabled(unsigned long usecs)
 {
-	u64 end, time, old_cc = 0;
-	unsigned long flags, cr0, mask, dummy;
-	int irq_context;
+	unsigned long mask, cr0, cr0_saved;
+	u64 clock_saved;
 
-	irq_context = in_interrupt();
-	if (!irq_context)
-		local_bh_disable();
-	local_irq_save(flags);
-	if (raw_irqs_disabled_flags(flags)) {
-		old_cc = local_tick_disable();
-		S390_lowcore.clock_comparator = -1ULL;
-		__ctl_store(cr0, 0, 0);
-		dummy = (cr0 & 0xffff00e0) | 0x00000800;
-		__ctl_load(dummy , 0, 0);
+	clock_saved = local_tick_disable();
+	set_clock_comparator(get_clock() + ((u64) usecs << 12));
+	__ctl_store(cr0_saved, 0, 0);
+	cr0 = (cr0_saved & 0xffff00e0) | 0x00000800;
+	__ctl_load(cr0 , 0, 0);
 	mask = psw_kernel_bits | PSW_MASK_WAIT | PSW_MASK_EXT;
-	} else
-		mask = psw_kernel_bits | PSW_MASK_WAIT |
-			PSW_MASK_EXT | PSW_MASK_IO;
+	trace_hardirqs_on();
+	__load_psw_mask(mask);
+	local_irq_disable();
+	__ctl_load(cr0_saved, 0, 0);
+	local_tick_enable(clock_saved);
+	set_clock_comparator(S390_lowcore.clock_comparator);
+}
+
+static void __udelay_enabled(unsigned long usecs)
+{
+	unsigned long mask;
+	u64 end, time;
 
+	mask = psw_kernel_bits | PSW_MASK_WAIT | PSW_MASK_EXT | PSW_MASK_IO;
 	end = get_clock() + ((u64) usecs << 12);
 	do {
 		time = end < S390_lowcore.clock_comparator ?
@@ -62,13 +58,37 @@ void __udelay(unsigned long usecs)
 		__load_psw_mask(mask);
 		local_irq_disable();
 	} while (get_clock() < end);
+	set_clock_comparator(S390_lowcore.clock_comparator);
+}
 
-	if (raw_irqs_disabled_flags(flags)) {
-		__ctl_load(cr0, 0, 0);
-		local_tick_enable(old_cc);
+/*
+ * Waits for 'usecs' microseconds using the TOD clock comparator.
+ */
+void __udelay(unsigned long usecs)
+{
+	unsigned long flags;
+
+	preempt_disable();
+	local_irq_save(flags);
+	if (in_irq()) {
+		__udelay_disabled(usecs);
+		goto out;
 	}
-	if (!irq_context)
+	if (in_softirq()) {
+		if (raw_irqs_disabled_flags(flags))
+			__udelay_disabled(usecs);
+		else
+			__udelay_enabled(usecs);
+		goto out;
+	}
+	if (raw_irqs_disabled_flags(flags)) {
+		local_bh_disable();
+		__udelay_disabled(usecs);
 		_local_bh_enable();
-	set_clock_comparator(S390_lowcore.clock_comparator);
+		goto out;
+	}
+	__udelay_enabled(usecs);
+out:
 	local_irq_restore(flags);
+	preempt_enable();
 }