summaryrefslogtreecommitdiffstats
path: root/arch/sparc64/kernel/smp.c
blob: 88d7a8ecfd7fa2c44d3f8dd1af293d2a8e3b9bcf (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
/* smp.c: Sparc64 SMP support.
 *
 * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
 */

#define __KERNEL_SYSCALLS__
#include <linux/unistd.h>

extern int linux_num_cpus;
extern void calibrate_delay(void);

volatile int smp_processors_ready = 0;
unsigned long cpu_present_map = 0;
int smp_num_cpus = 1;
int smp_threads_ready = 0;

struct cpuinfo_sparc64 cpu_data[NR_CPUS];
static unsigned char boot_cpu_id = 0;
static int smp_activated = 0;

volatile int cpu_number_map[NR_CPUS];
volatile int cpu_logical_map[NR_CPUS];

struct klock_info klock_info = { KLOCK_CLEAR, 0 };

static volatile int smp_commenced = 0;

void smp_setup(char *str, int *ints)
{
	/* XXX implement me XXX */
}

static char smp_buf[512];

char *smp_info(void)
{
	/* XXX not SMP safe and need to support up to 64 penguins */
	sprintf(smp_buf,
"        CPU0\t\tCPU1\t\tCPU2\t\tCPU3\n"
"State:  %s\t\t%s\t\t%s\t\t%s\n",
(cpu_present_map & 1) ? ((klock_info.akp == 0) ? "akp" : "online") : "offline",
(cpu_present_map & 2) ? ((klock_info.akp == 1) ? "akp" : "online") : "offline",
(cpu_present_map & 4) ? ((klock_info.akp == 2) ? "akp" : "online") : "offline",
(cpu_present_map & 8) ? ((klock_info.akp == 3) ? "akp" : "online") : "offline");
	return smp_buf;
}

void smp_store_cpu_info(int id)
{
	cpu_data[id].udelay_val = loops_per_sec;
}

void smp_commence(void)
{
	local_flush_cache_all();
	local_flush_tlb_all();
	smp_commenced = 1;
	local_flush_cache_all();
	local_flush_tlb_all();
}

static void smp_setup_percpu_timer(void);

static volatile unsigned long callin_flag = 0;

void smp_callin(void)
{
	int cpuid = hard_smp_processor_id();

	local_flush_cache_all();
	local_flush_tlb_all();

	smp_setup_percpu_timer();

	calibrate_delay();
	smp_store_cpu_info(cpuid);
	callin_flag = 1;
	__asm__ __volatile__("membar #Sync\n\t"
			     "flush  %g6" : : : "memory");

	while(!task[cpuid])
		barrier();
	current = task[cpuid];

	while(!smp_commenced)
		barrier();

	__sti();
}

extern int cpu_idle(void *unused);
extern void init_IRQ(void);

void initialize_secondary(void)
{
}

int start_secondary(void *unused)
{
	trap_init();
	init_IRQ();
	smp_callin();
	return cpu_idle(NULL);
}

extern struct prom_cpuinfo linux_cpus[NR_CPUS];

void smp_boot_cpus(void)
{
	int cpucount = 0, i, first, prev;

	printk("Entering UltraSMPenguin Mode...\n");
	__sti();
	cpu_present_map = 0;
	for(i = 0; i < linux_num_cpus; i++)
		cpu_present_map |= (1 << i);
	for(i = 0; i < NR_CPUS; i++) {
		cpu_number_map[i] = -1;
		cpu_logical_map[i] = -1;
	}
	cpu_number_map[boot_cpu_id] = 0;
	cpu_logical_map[0] = boot_cpu_id;
	klock_info.akp = boot_cpu_id;
	current->processor = boot_cpu_id;
	smp_store_cpu_info(boot_cpu_id);
	smp_setup_percpu_timer();

	if(linux_num_cpus == 1)
		return;

	for(i = 0; i < NR_CPUS; i++) {
		if(i == boot_cpu_id)
			continue;

		if(cpu_present_map & (1 << i)) {
			extern unsigned long sparc64_cpu_startup;
			unsigned long entry = (unsigned long)&sparc_cpu_startup;
			struct task_struct *p;
			int timeout;

			kernel_thread(start_secondary, NULL, CLONE_PID);
			p = task[++cpucount];
			p->processor = i;
			prom_startcpu(linux_cpus[i].prom_node, entry, i);
			for(timeout = 0; timeout < 5000000; timeout++) {
				if(cpu_callin_map[i])
					break;
				udelay(100);
			}
			if(cpu_callin_map[i]) {
				/* XXX fix this */
				cpu_number_map[i] = i;
				cpu_logical_map[i] = i;
			} else {
				cpucount--;
				printk("Processor %d is stuck.\n", i);
			}
		}
		if(!(cpu_callin_map[i])) {
			cpu_present_map &= ~(1 << i);
			cpu_number_map[i] = -1;
		}
	}
	if(cpucount == 0) {
		printk("Error: only one processor found.\n");
		cpu_present_map = (1 << smp_processor_id());
	} else {
		unsigned long bogosum = 0;

		for(i = 0; i < NR_CPUS; i++) {
			if(cpu_present_map & (1 << i))
				bogosum += cpu_data[i].udelay_val;
		}
		printk("Total of %d processors activated (%lu.%02lu BogoMIPS).\n",
		       cpucount + 1,
		       (bogosum + 2500)/500000,
		       ((bogosum + 2500)/5000)%100);
		smp_activated = 1;
		smp_num_cpus = cpucount + 1;
	}
	smp_processors_ready = 1;
}

/* XXX deprecated interface... */
void smp_message_pass(int target, int msg, unsigned long data, int wait)
{
	printk("smp_message_pass() called, this is bad, spinning.\n");
	__sti();
	while(1)
		barrier();
}

/* XXX Make it fast later. */
void smp_cross_call(unsigned long *func, u32 ctx, u64 data1, u64 data2)
{
	if(smp_processors_ready) {
		unsigned long mask;
		u64 data0 = (((unsigned long)ctx)<<32 |
			     (((unsigned long)func) & 0xffffffff));
		u64 pstate;
		int i, ncpus = smp_num_cpus;

		__asm__ __volatile__("rdpr %%pstate, %0" : "=r" (pstate));
		mask = (cpu_present_map & ~(1 << smp_processor_id()));
		for(i = 0; i < ncpus; i++) {
			if(mask & (1 << i)) {
				u64 target = mid<<14 | 0x70;
				u64 result;

				__asm__ __volatile__("
				wrpr	%0, %1, %%pstate
				wrpr	%%g0, %2, %%asi
				stxa	%3, [0x40] %%asi
				stxa	%4, [0x50] %%asi
				stxa	%5, [0x60] %%asi
				stxa	%%g0, [%6] %7
				membar	#Sync"
				: /* No outputs */
				: "r" (pstate), "i" (PSTATE_IE), "i" (ASI_UDB_INTR_W),
				  "r" (data0), "r" (data1), "r" (data2),
				  "r" (target), "i" (ASI_UDB_INTR_W));

				/* NOTE: PSTATE_IE is still clear. */
				do {
					__asm__ __volatile__("ldxa [%%g0] %1, %0",
						: "=r" (result)
						: "i" (ASI_INTR_DISPATCH_STAT));
				} while(result & 0x1);
				__asm__ __volatile__("wrpr %0, 0x0, %%pstate"
						     : : "r" (pstate));
				if(result & 0x2)
					panic("Penguin NACK's master!");
			}
		}

		/* NOTE: Caller runs local copy on master. */
	}
}

extern unsigned long xcall_flush_tlb_page;
extern unsigned long xcall_flush_tlb_mm;
extern unsigned long xcall_flush_tlb_range;
extern unsigned long xcall_flush_tlb_all;
extern unsigned long xcall_flush_cache_all;

void smp_flush_cache_all(void)
{
	smp_cross_call(&xcall_flush_cache_all, 0, 0, 0);
}

void smp_flush_tlb_all(void)
{
	smp_cross_call(&xcall_flush_tlb_all, 0, 0, 0);
}

void smp_flush_tlb_mm(struct mm_struct *mm)
{
	u32 ctx = mm->context & 0x1fff;
	if(mm->cpu_vm_mask != (1 << smp_processor_id()))
		smp_cross_call(&xcall_flush_tlb_mm, ctx, 0, 0);
	__flush_tlb_mm(ctx);
}

void smp_flush_tlb_range(struct mm_struct *mm, unsigned long start,
			 unsigned long end)
{
	u32 ctx = mm->context & 0x1fff;
	if(mm->cpu_vm_mask != (1 << smp_processor_id()))
		smp_cross_call(&xcall_flush_tlb_range, ctx, start, end);
	__flush_tlb_range(ctx, start, end);
}

void smp_flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
{
	struct mm_struct *mm = vma->vm_mm;
	u32 ctx = mm->context & 0x1fff;

	if(mm->cpu_vm_mask != (1 << smp_processor_id()))
		smp_cross_call(&xcall_flush_tlb_page, ctx, page, 0);
	__flush_tlb_page(ctx, page);
}

static spinlock_t ticker_lock = SPIN_LOCK_UNLOCKED;

static inline void sparc64_do_profile(unsigned long pc)
{
	if(prof_buffer && current->pid) {
		extern int _stext;

		pc -= (unsigned long) &_stext;
		pc >>= prof_shift;

		spin_lock(&ticker_lock);
		if(pc < prof_len)
			prof_buffer[pc]++;
		else
			prof_buffer[prof_len - 1]++;
		spin_unlock(&ticker_lock);
	}
}

unsigned int prof_multiplier[NR_CPUS];
unsigned int prof_counter[NR_CPUS];

extern void update_one_process(struct task_struct *p, unsigned long ticks,
			       unsigned long user, unsigned long system);

void smp_percpu_timer_interrupt(struct pt_regs *regs)
{
	int cpu = smp_processor_id();

	clear_profile_irq(cpu);
	if(!user_mode(regs))
		sparc_do_profile(regs->pc);
	if(!--prof_counter[cpu]) {
		int user = user_mode(regs);
		if(current->pid) {
			update_one_process(current, 1, user, !user);
			if(--current->counter < 0) {
				current->counter = 0;
				need_resched = 1;
			}

			spin_lock(&ticker_lock);
			if(user) {
				if(current->priority < DEF_PRIORITY)
					kstat.cpu_nice++;
				else
					kstat.cpu_user++;
			} else {
				kstat.cpu_system++;
			}
			spin_unlock(&ticker_lock);
		}
		prof_counter[cpu] = prof_multiplier[cpu];
	}
}

static void smp_setup_percpu_timer(void)
{
	/* XXX implement me */
}

int setup_profiling_timer(unsigned int multiplier)
{
	/* XXX implement me */
}