#include <linux/config.h>
#include <linux/init.h>
#include <linux/spinlock.h>
#include <linux/threads.h>
#include <linux/time.h>
#include <linux/timex.h>
#include <linux/sched.h>
#include <linux/cache.h>

#include <asm/atomic.h>
#include <asm/processor.h>
#include <asm/system.h>
#include <asm/hardirq.h>
#include <asm/softirq.h>
#include <asm/mmu_context.h>
#include <asm/irq.h>

#ifdef CONFIG_SGI_IP27

#include <asm/sn/arch.h>
#include <asm/sn/intr.h>
#include <asm/sn/addrs.h>
#include <asm/sn/agent.h>
#include <asm/sn/sn0/ip27.h>

#define DORESCHED	0xab
#define DOCALL		0xbc

static void sendintr(int destid, unsigned char status)
{
	int irq;

#if (CPUS_PER_NODE == 2)
	switch (status) {
		case DORESCHED:	irq = CPU_RESCHED_A_IRQ; break;
		case DOCALL:	irq = CPU_CALL_A_IRQ; break;
		default:	panic("sendintr");
	}
	irq += cputoslice(destid);

	/*
	 * Convert the compact hub number to the NASID to get the correct
	 * part of the address space.  Then set the interrupt bit associated
	 * with the CPU we want to send the interrupt to.
	 */
	REMOTE_HUB_SEND_INTR(COMPACT_TO_NASID_NODEID(cputocnode(destid)),
			FAST_IRQ_TO_LEVEL(irq));
#else
	<< Bomb!  Must redefine this for more than 2 CPUS. >>
#endif
}

#endif /* CONFIG_SGI_IP27 */

/* The 'big kernel lock' */
spinlock_t kernel_flag __cacheline_aligned_in_smp = SPIN_LOCK_UNLOCKED;
int smp_threads_ready;	/* Not used */
atomic_t smp_commenced = ATOMIC_INIT(0);
struct cpuinfo_mips cpu_data[NR_CPUS];
int smp_num_cpus = 1;		/* Number that came online.  */
int __cpu_number_map[NR_CPUS];
int __cpu_logical_map[NR_CPUS];
cycles_t cacheflush_time;

static void smp_tune_scheduling (void)
{
}

void __init smp_boot_cpus(void)
{
	extern void allowboot(void);

	init_new_context(current, &init_mm);
	current->processor = 0;
	init_idle();
	smp_tune_scheduling();
	allowboot();
}

void __init smp_commence(void)
{
	wmb();
	atomic_set(&smp_commenced,1);
}

static void stop_this_cpu(void *dummy)
{
	/*
	 * Remove this CPU
	 */
	for (;;);
}

void smp_send_stop(void)
{
	smp_call_function(stop_this_cpu, NULL, 1, 0);
	smp_num_cpus = 1;
}

/*
 * this function sends a 'reschedule' IPI to another CPU.
 * it goes straight through and wastes no time serializing
 * anything. Worst case is that we lose a reschedule ...
 */
void smp_send_reschedule(int cpu)
{
	sendintr(cpu, DORESCHED);
}

/* Not really SMP stuff ... */
int setup_profiling_timer(unsigned int multiplier)
{
	return 0;
}

/*
 * Run a function on all other CPUs.
 *  <func>      The function to run. This must be fast and non-blocking.
 *  <info>      An arbitrary pointer to pass to the function.
 *  <retry>     If true, keep retrying until ready.
 *  <wait>      If true, wait until function has completed on other CPUs.
 *  [RETURNS]   0 on success, else a negative status code.
 *
 * Does not return until remote CPUs are nearly ready to execute <func>
 * or are or have executed.
 */
static volatile struct call_data_struct {
	void (*func) (void *info);
	void *info;
	atomic_t started;
	atomic_t finished;
	int wait;
} *call_data;

int smp_call_function (void (*func) (void *info), void *info, int retry, 
								int wait)
{
	struct call_data_struct data;
	int i, cpus = smp_num_cpus-1;
	static spinlock_t lock = SPIN_LOCK_UNLOCKED;

	if (cpus == 0)
		return 0;

	data.func = func;
	data.info = info;
	atomic_set(&data.started, 0);
	data.wait = wait;
	if (wait)
		atomic_set(&data.finished, 0);

	spin_lock_bh(&lock);
	call_data = &data;
	/* Send a message to all other CPUs and wait for them to respond */
	for (i = 0; i < smp_num_cpus; i++)
		if (smp_processor_id() != i)
			sendintr(i, DOCALL);

	/* Wait for response */
	/* FIXME: lock-up detection, backtrace on lock-up */
	while (atomic_read(&data.started) != cpus)
		barrier();

	if (wait)
		while (atomic_read(&data.finished) != cpus)
			barrier();
	spin_unlock_bh(&lock);
	return 0;
}

extern void smp_call_function_interrupt(int irq, void *d, struct pt_regs *r)
{
	void (*func) (void *info) = call_data->func;
	void *info = call_data->info;
	int wait = call_data->wait;

	/*
	 * Notify initiating CPU that I've grabbed the data and am
	 * about to execute the function.
	 */
	atomic_inc(&call_data->started);

	/*
	 * At this point the info structure may be out of scope unless wait==1.
	 */
	(*func)(info);
	if (wait)
		atomic_inc(&call_data->finished);
}
	

static void flush_tlb_all_ipi(void *info)
{
	_flush_tlb_all();
}

void flush_tlb_all(void)
{
	smp_call_function(flush_tlb_all_ipi, 0, 1, 1);
	_flush_tlb_all();
}

static void flush_tlb_mm_ipi(void *mm)
{
	_flush_tlb_mm((struct mm_struct *)mm);
}

/*
 * The following tlb flush calls are invoked when old translations are 
 * being torn down, or pte attributes are changing. For single threaded
 * address spaces, a new context is obtained on the current cpu, and tlb
 * context on other cpus are invalidated to force a new context allocation
 * at switch_mm time, should the mm ever be used on other cpus. For 
 * multithreaded address spaces, intercpu interrupts have to be sent.
 * Another case where intercpu interrupts are required is when the target
 * mm might be active on another cpu (eg debuggers doing the flushes on
 * behalf of debugees, kswapd stealing pages from another process etc).
 * Kanoj 07/00.
 */

void flush_tlb_mm(struct mm_struct *mm)
{
	if ((atomic_read(&mm->mm_users) != 1) || (current->mm != mm)) {
		smp_call_function(flush_tlb_mm_ipi, (void *)mm, 1, 1);
	} else {
		int i;
		for (i = 0; i < smp_num_cpus; i++)
			if (smp_processor_id() != i)
				CPU_CONTEXT(i, mm) = 0;
	}
	_flush_tlb_mm(mm);
}

struct flush_tlb_data {
	struct mm_struct *mm;
	struct vm_area_struct *vma;
	unsigned long addr1;
	unsigned long addr2;
};

static void flush_tlb_range_ipi(void *info)
{
	struct flush_tlb_data *fd = (struct flush_tlb_data *)info;

	_flush_tlb_range(fd->mm, fd->addr1, fd->addr2);
}

void flush_tlb_range(struct mm_struct *mm, unsigned long start, unsigned long end)
{
	if ((atomic_read(&mm->mm_users) != 1) || (current->mm != mm)) {
		struct flush_tlb_data fd;

		fd.mm = mm;
		fd.addr1 = start;
		fd.addr2 = end;
		smp_call_function(flush_tlb_range_ipi, (void *)&fd, 1, 1);
	} else {
		int i;
		for (i = 0; i < smp_num_cpus; i++)
			if (smp_processor_id() != i)
				CPU_CONTEXT(i, mm) = 0;
	}
	_flush_tlb_range(mm, start, end);
}

static void flush_tlb_page_ipi(void *info)
{
	struct flush_tlb_data *fd = (struct flush_tlb_data *)info;

	_flush_tlb_page(fd->vma, fd->addr1);
}

void flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
{
	if ((atomic_read(&vma->vm_mm->mm_users) != 1) || (current->mm != vma->vm_mm)) {
		struct flush_tlb_data fd;

		fd.vma = vma;
		fd.addr1 = page;
		smp_call_function(flush_tlb_page_ipi, (void *)&fd, 1, 1);
	} else {
		int i;
		for (i = 0; i < smp_num_cpus; i++)
			if (smp_processor_id() != i)
				CPU_CONTEXT(i, vma->vm_mm) = 0;
	}
	_flush_tlb_page(vma, page);
}