calibrate_delay() must be __cpuinit

calibrate_delay() must be __cpuinit, not __{dev,}init.

I've verified that this is correct for all users.

While doing the latter, I also did the following cleanups:
- remove pointless additional prototypes in C files
- ensure all users #include <linux/delay.h>

This fixes the following section mismatches with CONFIG_HOTPLUG=n,
CONFIG_HOTPLUG_CPU=y:

WARNING: vmlinux.o(.text+0x1128d): Section mismatch: reference to .init.text.1:calibrate_delay (between 'check_cx686_slop' and 'set_cx86_reorder')
WARNING: vmlinux.o(.text+0x25102): Section mismatch: reference to .init.text.1:calibrate_delay (between 'smp_callin' and 'cpu_coregroup_map')

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Richard Henderson <rth@twiddle.net>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Christian Zankel <chris@zankel.net>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
diff --git a/arch/alpha/kernel/smp.c b/arch/alpha/kernel/smp.c
index f4ab233..63c2073 100644
--- a/arch/alpha/kernel/smp.c
+++ b/arch/alpha/kernel/smp.c
@@ -77,10 +77,6 @@
 int smp_num_cpus = 1;		/* Number that came online.  */
 EXPORT_SYMBOL(smp_num_cpus);
 
-extern void calibrate_delay(void);
-
-
-
 /*
  * Called by both boot and secondaries to move global data into
  *  per-processor storage.
diff --git a/arch/frv/kernel/setup.c b/arch/frv/kernel/setup.c
index a74c087..b38ae1f 100644
--- a/arch/frv/kernel/setup.c
+++ b/arch/frv/kernel/setup.c
@@ -708,7 +708,7 @@
 /*
  * calibrate the delay loop
  */
-void __init calibrate_delay(void)
+void __cpuinit calibrate_delay(void)
 {
 	loops_per_jiffy = __delay_loops_MHz * (1000000 / HZ);
 
diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c
index 480b1a5..328fd2f 100644
--- a/arch/ia64/kernel/smpboot.c
+++ b/arch/ia64/kernel/smpboot.c
@@ -120,7 +120,6 @@
 
 #define DEBUG_ITC_SYNC	0
 
-extern void __devinit calibrate_delay (void);
 extern void start_ap (void);
 extern unsigned long ia64_iobase;
 
diff --git a/arch/mips/kernel/smp.c b/arch/mips/kernel/smp.c
index 1e5dfc2..9d41dab 100644
--- a/arch/mips/kernel/smp.c
+++ b/arch/mips/kernel/smp.c
@@ -52,7 +52,6 @@
 EXPORT_SYMBOL(phys_cpu_present_map);
 EXPORT_SYMBOL(cpu_online_map);
 
-extern void __init calibrate_delay(void);
 extern void cpu_idle(void);
 
 /* Number of TCs (or siblings in Intel speak) per CPU core */
diff --git a/arch/powerpc/platforms/powermac/cpufreq_32.c b/arch/powerpc/platforms/powermac/cpufreq_32.c
index c04abcc..792d3ce 100644
--- a/arch/powerpc/platforms/powermac/cpufreq_32.c
+++ b/arch/powerpc/platforms/powermac/cpufreq_32.c
@@ -113,8 +113,6 @@
 	 * result. We backup/restore the value to avoid affecting the
 	 * core cpufreq framework's own calculation.
 	 */
-	extern void calibrate_delay(void);
-
 	unsigned long save_lpj = loops_per_jiffy;
 	calibrate_delay();
 	loops_per_jiffy = save_lpj;
diff --git a/arch/sparc/kernel/sun4d_smp.c b/arch/sparc/kernel/sun4d_smp.c
index 89a6de95..0def481 100644
--- a/arch/sparc/kernel/sun4d_smp.c
+++ b/arch/sparc/kernel/sun4d_smp.c
@@ -19,12 +19,12 @@
 #include <linux/mm.h>
 #include <linux/swap.h>
 #include <linux/profile.h>
+#include <linux/delay.h>
 
 #include <asm/ptrace.h>
 #include <asm/atomic.h>
 #include <asm/irq_regs.h>
 
-#include <asm/delay.h>
 #include <asm/irq.h>
 #include <asm/page.h>
 #include <asm/pgalloc.h>
@@ -41,8 +41,6 @@
 
 extern ctxd_t *srmmu_ctx_table_phys;
 
-extern void calibrate_delay(void);
-
 static volatile int smp_processors_ready = 0;
 static int smp_highest_cpu;
 extern volatile unsigned long cpu_callin_map[NR_CPUS];
diff --git a/arch/sparc/kernel/sun4m_smp.c b/arch/sparc/kernel/sun4m_smp.c
index 730eb579..0b94072 100644
--- a/arch/sparc/kernel/sun4m_smp.c
+++ b/arch/sparc/kernel/sun4m_smp.c
@@ -16,6 +16,8 @@
 #include <linux/mm.h>
 #include <linux/swap.h>
 #include <linux/profile.h>
+#include <linux/delay.h>
+
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
 #include <asm/irq_regs.h>
@@ -23,7 +25,6 @@
 #include <asm/ptrace.h>
 #include <asm/atomic.h>
 
-#include <asm/delay.h>
 #include <asm/irq.h>
 #include <asm/page.h>
 #include <asm/pgalloc.h>
@@ -39,8 +40,6 @@
 
 extern ctxd_t *srmmu_ctx_table_phys;
 
-extern void calibrate_delay(void);
-
 extern volatile unsigned long cpu_callin_map[NR_CPUS];
 extern unsigned char boot_cpu_id;
 
diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c
index c399449..a8052b7 100644
--- a/arch/sparc64/kernel/smp.c
+++ b/arch/sparc64/kernel/smp.c
@@ -46,8 +46,6 @@
 #include <asm/ldc.h>
 #include <asm/hypervisor.h>
 
-extern void calibrate_delay(void);
-
 int sparc64_multi_core __read_mostly;
 
 cpumask_t cpu_possible_map __read_mostly = CPU_MASK_NONE;
diff --git a/arch/x86/kernel/cpu/cyrix.c b/arch/x86/kernel/cpu/cyrix.c
index 404a6a2..7139b02 100644
--- a/arch/x86/kernel/cpu/cyrix.c
+++ b/arch/x86/kernel/cpu/cyrix.c
@@ -83,8 +83,6 @@
  * FIXME: our newer udelay uses the tsc. We don't need to frob with SLOP
  */
 
-extern void calibrate_delay(void) __init;
-
 static void __cpuinit check_cx686_slop(struct cpuinfo_x86 *c)
 {
 	unsigned long flags;
diff --git a/arch/x86/kernel/smpboot_32.c b/arch/x86/kernel/smpboot_32.c
index 5787a0c..579b9b7 100644
--- a/arch/x86/kernel/smpboot_32.c
+++ b/arch/x86/kernel/smpboot_32.c
@@ -202,8 +202,6 @@
 	;
 }
 
-extern void calibrate_delay(void);
-
 static atomic_t init_deasserted;
 
 static void __cpuinit smp_callin(void)
diff --git a/arch/x86/mach-voyager/voyager_smp.c b/arch/x86/mach-voyager/voyager_smp.c
index dffa786..3cc8eb2 100644
--- a/arch/x86/mach-voyager/voyager_smp.c
+++ b/arch/x86/mach-voyager/voyager_smp.c
@@ -444,8 +444,6 @@
 static void __init start_secondary(void *unused)
 {
 	__u8 cpuid = hard_smp_processor_id();
-	/* external functions not defined in the headers */
-	extern void calibrate_delay(void);
 
 	cpu_init();
 
diff --git a/arch/xtensa/kernel/time.c b/arch/xtensa/kernel/time.c
index 60d29fe..8df1e842 100644
--- a/arch/xtensa/kernel/time.c
+++ b/arch/xtensa/kernel/time.c
@@ -204,7 +204,7 @@
 }
 
 #ifndef CONFIG_GENERIC_CALIBRATE_DELAY
-void __devinit calibrate_delay(void)
+void __cpuinit calibrate_delay(void)
 {
 	loops_per_jiffy = CCOUNT_PER_JIFFY;
 	printk("Calibrating delay loop (skipped)... "
diff --git a/drivers/s390/sysinfo.c b/drivers/s390/sysinfo.c
index 19343f9..291ff62 100644
--- a/drivers/s390/sysinfo.c
+++ b/drivers/s390/sysinfo.c
@@ -422,7 +422,7 @@
 /*
  * calibrate the delay loop
  */
-void __init calibrate_delay(void)
+void __cpuinit calibrate_delay(void)
 {
 	s390_adjust_jiffies();
 	/* Print the good old Bogomips line .. */
diff --git a/init/calibrate.c b/init/calibrate.c
index 1d87891..ecb3822 100644
--- a/init/calibrate.c
+++ b/init/calibrate.c
@@ -28,7 +28,7 @@
 #define DELAY_CALIBRATION_TICKS			((HZ < 100) ? 1 : (HZ/100))
 #define MAX_DIRECT_CALIBRATION_RETRIES		5
 
-static unsigned long __devinit calibrate_delay_direct(void)
+static unsigned long __cpuinit calibrate_delay_direct(void)
 {
 	unsigned long pre_start, start, post_start;
 	unsigned long pre_end, end, post_end;
@@ -101,7 +101,7 @@
 	return 0;
 }
 #else
-static unsigned long __devinit calibrate_delay_direct(void) {return 0;}
+static unsigned long __cpuinit calibrate_delay_direct(void) {return 0;}
 #endif
 
 /*
@@ -111,7 +111,7 @@
  */
 #define LPS_PREC 8
 
-void __devinit calibrate_delay(void)
+void __cpuinit calibrate_delay(void)
 {
 	unsigned long ticks, loopbit;
 	int lps_precision = LPS_PREC;