aboutsummaryrefslogtreecommitdiff
path: root/arch/tile/kernel/process.c
blob: 80b2a18deb87a958c07b7c689dcb710b07de4ad3 (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
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
/*
 * Copyright 2010 Tilera Corporation. All Rights Reserved.
 *
 *   This program is free software; you can redistribute it and/or
 *   modify it under the terms of the GNU General Public License
 *   as published by the Free Software Foundation, version 2.
 *
 *   This program is distributed in the hope that it will be useful, but
 *   WITHOUT ANY WARRANTY; without even the implied warranty of
 *   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
 *   NON INFRINGEMENT.  See the GNU General Public License for
 *   more details.
 */

#include <linux/sched.h>
#include <linux/preempt.h>
#include <linux/module.h>
#include <linux/fs.h>
#include <linux/kprobes.h>
#include <linux/elfcore.h>
#include <linux/tick.h>
#include <linux/init.h>
#include <linux/mm.h>
#include <linux/compat.h>
#include <linux/hardirq.h>
#include <linux/syscalls.h>
#include <linux/kernel.h>
#include <linux/tracehook.h>
#include <linux/signal.h>
#include <asm/stack.h>
#include <asm/switch_to.h>
#include <asm/homecache.h>
#include <asm/syscalls.h>
#include <asm/traps.h>
#include <asm/setup.h>
#ifdef CONFIG_HARDWALL
#include <asm/hardwall.h>
#endif
#include <arch/chip.h>
#include <arch/abi.h>
#include <arch/sim_def.h>

/*
 * Use the (x86) "idle=poll" option to prefer low latency when leaving the
 * idle loop over low power while in the idle loop, e.g. if we have
 * one thread per core and we want to get threads out of futex waits fast.
 */
static int __init idle_setup(char *str)
{
	if (!str)
		return -EINVAL;

	if (!strcmp(str, "poll")) {
		pr_info("using polling idle threads.\n");
		cpu_idle_poll_ctrl(true);
		return 0;
	} else if (!strcmp(str, "halt")) {
		return 0;
	}
	return -1;
}
early_param("idle", idle_setup);

void arch_cpu_idle(void)
{
	__get_cpu_var(irq_stat).idle_timestamp = jiffies;
	_cpu_idle();
}

/*
 * Release a thread_info structure
 */
void arch_release_thread_info(struct thread_info *info)
{
	struct single_step_state *step_state = info->step_state;

#ifdef CONFIG_HARDWALL
	/*
	 * We free a thread_info from the context of the task that has
	 * been scheduled next, so the original task is already dead.
	 * Calling deactivate here just frees up the data structures.
	 * If the task we're freeing held the last reference to a
	 * hardwall fd, it would have been released prior to this point
	 * anyway via exit_files(), and the hardwall_task.info pointers
	 * would be NULL by now.
	 */
	hardwall_deactivate_all(info->task);
#endif

	if (step_state) {

		/*
		 * FIXME: we don't munmap step_state->buffer
		 * because the mm_struct for this process (info->task->mm)
		 * has already been zeroed in exit_mm().  Keeping a
		 * reference to it here seems like a bad move, so this
		 * means we can't munmap() the buffer, and therefore if we
		 * ptrace multiple threads in a process, we will slowly
		 * leak user memory.  (Note that as soon as the last
		 * thread in a process dies, we will reclaim all user
		 * memory including single-step buffers in the usual way.)
		 * We should either assign a kernel VA to this buffer
		 * somehow, or we should associate the buffer(s) with the
		 * mm itself so we can clean them up that way.
		 */
		kfree(step_state);
	}
}

static void save_arch_state(struct thread_struct *t);

int copy_thread(unsigned long clone_flags, unsigned long sp,
		unsigned long arg, struct task_struct *p)
{
	struct pt_regs *childregs = task_pt_regs(p);
	unsigned long ksp;
	unsigned long *callee_regs;

	/*
	 * Set up the stack and stack pointer appropriately for the
	 * new child to find itself woken up in __switch_to().
	 * The callee-saved registers must be on the stack to be read;
	 * the new task will then jump to assembly support to handle
	 * calling schedule_tail(), etc., and (for userspace tasks)
	 * returning to the context set up in the pt_regs.
	 */
	ksp = (unsigned long) childregs;
	ksp -= C_ABI_SAVE_AREA_SIZE;   /* interrupt-entry save area */
	((long *)ksp)[0] = ((long *)ksp)[1] = 0;
	ksp -= CALLEE_SAVED_REGS_COUNT * sizeof(unsigned long);
	callee_regs = (unsigned long *)ksp;
	ksp -= C_ABI_SAVE_AREA_SIZE;   /* __switch_to() save area */
	((long *)ksp)[0] = ((long *)ksp)[1] = 0;
	p->thread.ksp = ksp;

	/* Record the pid of the task that created this one. */
	p->thread.creator_pid = current->pid;

	if (unlikely(p->flags & PF_KTHREAD)) {
		/* kernel thread */
		memset(childregs, 0, sizeof(struct pt_regs));
		memset(&callee_regs[2], 0,
		       (CALLEE_SAVED_REGS_COUNT - 2) * sizeof(unsigned long));
		callee_regs[0] = sp;   /* r30 = function */
		callee_regs[1] = arg;  /* r31 = arg */
		childregs->ex1 = PL_ICS_EX1(KERNEL_PL, 0);
		p->thread.pc = (unsigned long) ret_from_kernel_thread;
		return 0;
	}

	/*
	 * Start new thread in ret_from_fork so it schedules properly
	 * and then return from interrupt like the parent.
	 */
	p->thread.pc = (unsigned long) ret_from_fork;

	/*
	 * Do not clone step state from the parent; each thread
	 * must make its own lazily.
	 */
	task_thread_info(p)->step_state = NULL;

	/*
	 * Copy the registers onto the kernel stack so the
	 * return-from-interrupt code will reload it into registers.
	 */
	*childregs = *current_pt_regs();
	childregs->regs[0] = 0;         /* return value is zero */
	if (sp)
		childregs->sp = sp;  /* override with new user stack pointer */
	memcpy(callee_regs, &childregs->regs[CALLEE_SAVED_FIRST_REG],
	       CALLEE_SAVED_REGS_COUNT * sizeof(unsigned long));

	/* Save user stack top pointer so we can ID the stack vm area later. */
	p->thread.usp0 = childregs->sp;

	/*
	 * If CLONE_SETTLS is set, set "tp" in the new task to "r4",
	 * which is passed in as arg #5 to sys_clone().
	 */
	if (clone_flags & CLONE_SETTLS)
		childregs->tp = childregs->regs[4];


#if CHIP_HAS_TILE_DMA()
	/*
	 * No DMA in the new thread.  We model this on the fact that
	 * fork() clears the pending signals, alarms, and aio for the child.
	 */
	memset(&p->thread.tile_dma_state, 0, sizeof(struct tile_dma_state));
	memset(&p->thread.dma_async_tlb, 0, sizeof(struct async_tlb));
#endif

#if CHIP_HAS_SN_PROC()
	/* Likewise, the new thread is not running static processor code. */
	p->thread.sn_proc_running = 0;
	memset(&p->thread.sn_async_tlb, 0, sizeof(struct async_tlb));
#endif

#if CHIP_HAS_PROC_STATUS_SPR()
	/* New thread has its miscellaneous processor state bits clear. */
	p->thread.proc_status = 0;
#endif

#ifdef CONFIG_HARDWALL
	/* New thread does not own any networks. */
	memset(&p->thread.hardwall[0], 0,
	       sizeof(struct hardwall_task) * HARDWALL_TYPES);
#endif


	/*
	 * Start the new thread with the current architecture state
	 * (user interrupt masks, etc.).
	 */
	save_arch_state(&p->thread);

	return 0;
}

/*
 * Return "current" if it looks plausible, or else a pointer to a dummy.
 * This can be helpful if we are just trying to emit a clean panic.
 */
struct task_struct *validate_current(void)
{
	static struct task_struct corrupt = { .comm = "<corrupt>" };
	struct task_struct *tsk = current;
	if (unlikely((unsigned long)tsk < PAGE_OFFSET ||
		     (high_memory && (void *)tsk > high_memory) ||
		     ((unsigned long)tsk & (__alignof__(*tsk) - 1)) != 0)) {
		pr_err("Corrupt 'current' %p (sp %#lx)\n", tsk, stack_pointer);
		tsk = &corrupt;
	}
	return tsk;
}

/* Take and return the pointer to the previous task, for schedule_tail(). */
struct task_struct *sim_notify_fork(struct task_struct *prev)
{
	struct task_struct *tsk = current;
	__insn_mtspr(SPR_SIM_CONTROL, SIM_CONTROL_OS_FORK_PARENT |
		     (tsk->thread.creator_pid << _SIM_CONTROL_OPERATOR_BITS));
	__insn_mtspr(SPR_SIM_CONTROL, SIM_CONTROL_OS_FORK |
		     (tsk->pid << _SIM_CONTROL_OPERATOR_BITS));
	return prev;
}

int dump_task_regs(struct task_struct *tsk, elf_gregset_t *regs)
{
	struct pt_regs *ptregs = task_pt_regs(tsk);
	elf_core_copy_regs(regs, ptregs);
	return 1;
}

#if CHIP_HAS_TILE_DMA()

/* Allow user processes to access the DMA SPRs */
void grant_dma_mpls(void)
{
#if CONFIG_KERNEL_PL == 2
	__insn_mtspr(SPR_MPL_DMA_CPL_SET_1, 1);
	__insn_mtspr(SPR_MPL_DMA_NOTIFY_SET_1, 1);
#else
	__insn_mtspr(SPR_MPL_DMA_CPL_SET_0, 1);
	__insn_mtspr(SPR_MPL_DMA_NOTIFY_SET_0, 1);
#endif
}

/* Forbid user processes from accessing the DMA SPRs */
void restrict_dma_mpls(void)
{
#if CONFIG_KERNEL_PL == 2
	__insn_mtspr(SPR_MPL_DMA_CPL_SET_2, 1);
	__insn_mtspr(SPR_MPL_DMA_NOTIFY_SET_2, 1);
#else
	__insn_mtspr(SPR_MPL_DMA_CPL_SET_1, 1);
	__insn_mtspr(SPR_MPL_DMA_NOTIFY_SET_1, 1);
#endif
}

/* Pause the DMA engine, then save off its state registers. */
static void save_tile_dma_state(struct tile_dma_state *dma)
{
	unsigned long state = __insn_mfspr(SPR_DMA_USER_STATUS);
	unsigned long post_suspend_state;

	/* If we're running, suspend the engine. */
	if ((state & DMA_STATUS_MASK) == SPR_DMA_STATUS__RUNNING_MASK)
		__insn_mtspr(SPR_DMA_CTR, SPR_DMA_CTR__SUSPEND_MASK);

	/*
	 * Wait for the engine to idle, then save regs.  Note that we
	 * want to record the "running" bit from before suspension,
	 * and the "done" bit from after, so that we can properly
	 * distinguish a case where the user suspended the engine from
	 * the case where the kernel suspended as part of the context
	 * swap.
	 */
	do {
		post_suspend_state = __insn_mfspr(SPR_DMA_USER_STATUS);
	} while (post_suspend_state & SPR_DMA_STATUS__BUSY_MASK);

	dma->src = __insn_mfspr(SPR_DMA_SRC_ADDR);
	dma->src_chunk = __insn_mfspr(SPR_DMA_SRC_CHUNK_ADDR);
	dma->dest = __insn_mfspr(SPR_DMA_DST_ADDR);
	dma->dest_chunk = __insn_mfspr(SPR_DMA_DST_CHUNK_ADDR);
	dma->strides = __insn_mfspr(SPR_DMA_STRIDE);
	dma->chunk_size = __insn_mfspr(SPR_DMA_CHUNK_SIZE);
	dma->byte = __insn_mfspr(SPR_DMA_BYTE);
	dma->status = (state & SPR_DMA_STATUS__RUNNING_MASK) |
		(post_suspend_state & SPR_DMA_STATUS__DONE_MASK);
}

/* Restart a DMA that was running before we were context-switched out. */
static void restore_tile_dma_state(struct thread_struct *t)
{
	const struct tile_dma_state *dma = &t->tile_dma_state;

	/*
	 * The only way to restore the done bit is to run a zero
	 * length transaction.
	 */
	if ((dma->status & SPR_DMA_STATUS__DONE_MASK) &&
	    !(__insn_mfspr(SPR_DMA_USER_STATUS) & SPR_DMA_STATUS__DONE_MASK)) {
		__insn_mtspr(SPR_DMA_BYTE, 0);
		__insn_mtspr(SPR_DMA_CTR, SPR_DMA_CTR__REQUEST_MASK);
		while (__insn_mfspr(SPR_DMA_USER_STATUS) &
		       SPR_DMA_STATUS__BUSY_MASK)
			;
	}

	__insn_mtspr(SPR_DMA_SRC_ADDR, dma->src);
	__insn_mtspr(SPR_DMA_SRC_CHUNK_ADDR, dma->src_chunk);
	__insn_mtspr(SPR_DMA_DST_ADDR, dma->dest);
	__insn_mtspr(SPR_DMA_DST_CHUNK_ADDR, dma->dest_chunk);
	__insn_mtspr(SPR_DMA_STRIDE, dma->strides);
	__insn_mtspr(SPR_DMA_CHUNK_SIZE, dma->chunk_size);
	__insn_mtspr(SPR_DMA_BYTE, dma->byte);

	/*
	 * Restart the engine if we were running and not done.
	 * Clear a pending async DMA fault that we were waiting on return
	 * to user space to execute, since we expect the DMA engine
	 * to regenerate those faults for us now.  Note that we don't
	 * try to clear the TIF_ASYNC_TLB flag, since it's relatively
	 * harmless if set, and it covers both DMA and the SN processor.
	 */
	if ((dma->status & DMA_STATUS_MASK) == SPR_DMA_STATUS__RUNNING_MASK) {
		t->dma_async_tlb.fault_num = 0;
		__insn_mtspr(SPR_DMA_CTR, SPR_DMA_CTR__REQUEST_MASK);
	}
}

#endif

static void save_arch_state(struct thread_struct *t)
{
#if CHIP_HAS_SPLIT_INTR_MASK()
	t->interrupt_mask = __insn_mfspr(SPR_INTERRUPT_MASK_0_0) |
		((u64)__insn_mfspr(SPR_INTERRUPT_MASK_0_1) << 32);
#else
	t->interrupt_mask = __insn_mfspr(SPR_INTERRUPT_MASK_0);
#endif
	t->ex_context[0] = __insn_mfspr(SPR_EX_CONTEXT_0_0);
	t->ex_context[1] = __insn_mfspr(SPR_EX_CONTEXT_0_1);
	t->system_save[0] = __insn_mfspr(SPR_SYSTEM_SAVE_0_0);
	t->system_save[1] = __insn_mfspr(SPR_SYSTEM_SAVE_0_1);
	t->system_save[2] = __insn_mfspr(SPR_SYSTEM_SAVE_0_2);
	t->system_save[3] = __insn_mfspr(SPR_SYSTEM_SAVE_0_3);
	t->intctrl_0 = __insn_mfspr(SPR_INTCTRL_0_STATUS);
#if CHIP_HAS_PROC_STATUS_SPR()
	t->proc_status = __insn_mfspr(SPR_PROC_STATUS);
#endif
#if !CHIP_HAS_FIXED_INTVEC_BASE()
	t->interrupt_vector_base = __insn_mfspr(SPR_INTERRUPT_VECTOR_BASE_0);
#endif
#if CHIP_HAS_TILE_RTF_HWM()
	t->tile_rtf_hwm = __insn_mfspr(SPR_TILE_RTF_HWM);
#endif
#if CHIP_HAS_DSTREAM_PF()
	t->dstream_pf = __insn_mfspr(SPR_DSTREAM_PF);
#endif
}

static void restore_arch_state(const struct thread_struct *t)
{
#if CHIP_HAS_SPLIT_INTR_MASK()
	__insn_mtspr(SPR_INTERRUPT_MASK_0_0, (u32) t->interrupt_mask);
	__insn_mtspr(SPR_INTERRUPT_MASK_0_1, t->interrupt_mask >> 32);
#else
	__insn_mtspr(SPR_INTERRUPT_MASK_0, t->interrupt_mask);
#endif
	__insn_mtspr(SPR_EX_CONTEXT_0_0, t->ex_context[0]);
	__insn_mtspr(SPR_EX_CONTEXT_0_1, t->ex_context[1]);
	__insn_mtspr(SPR_SYSTEM_SAVE_0_0, t->system_save[0]);
	__insn_mtspr(SPR_SYSTEM_SAVE_0_1, t->system_save[1]);
	__insn_mtspr(SPR_SYSTEM_SAVE_0_2, t->system_save[2]);
	__insn_mtspr(SPR_SYSTEM_SAVE_0_3, t->system_save[3]);
	__insn_mtspr(SPR_INTCTRL_0_STATUS, t->intctrl_0);
#if CHIP_HAS_PROC_STATUS_SPR()
	__insn_mtspr(SPR_PROC_STATUS, t->proc_status);
#endif
#if !CHIP_HAS_FIXED_INTVEC_BASE()
	__insn_mtspr(SPR_INTERRUPT_VECTOR_BASE_0, t->interrupt_vector_base);
#endif
#if CHIP_HAS_TILE_RTF_HWM()
	__insn_mtspr(SPR_TILE_RTF_HWM, t->tile_rtf_hwm);
#endif
#if CHIP_HAS_DSTREAM_PF()
	__insn_mtspr(SPR_DSTREAM_PF, t->dstream_pf);
#endif
}


void _prepare_arch_switch(struct task_struct *next)
{
#if CHIP_HAS_SN_PROC()
	int snctl;
#endif
#if CHIP_HAS_TILE_DMA()
	struct tile_dma_state *dma = &current->thread.tile_dma_state;
	if (dma->enabled)
		save_tile_dma_state(dma);
#endif
#if CHIP_HAS_SN_PROC()
	/*
	 * Suspend the static network processor if it was running.
	 * We do not suspend the fabric itself, just like we don't
	 * try to suspend the UDN.
	 */
	snctl = __insn_mfspr(SPR_SNCTL);
	current->thread.sn_proc_running =
		(snctl & SPR_SNCTL__FRZPROC_MASK) == 0;
	if (current->thread.sn_proc_running)
		__insn_mtspr(SPR_SNCTL, snctl | SPR_SNCTL__FRZPROC_MASK);
#endif
}


struct task_struct *__sched _switch_to(struct task_struct *prev,
				       struct task_struct *next)
{
	/* DMA state is already saved; save off other arch state. */
	save_arch_state(&prev->thread);

#if CHIP_HAS_TILE_DMA()
	/*
	 * Restore DMA in new task if desired.
	 * Note that it is only safe to restart here since interrupts
	 * are disabled, so we can't take any DMATLB miss or access
	 * interrupts before we have finished switching stacks.
	 */
	if (next->thread.tile_dma_state.enabled) {
		restore_tile_dma_state(&next->thread);
		grant_dma_mpls();
	} else {
		restrict_dma_mpls();
	}
#endif

	/* Restore other arch state. */
	restore_arch_state(&next->thread);

#if CHIP_HAS_SN_PROC()
	/*
	 * Restart static network processor in the new process
	 * if it was running before.
	 */
	if (next->thread.sn_proc_running) {
		int snctl = __insn_mfspr(SPR_SNCTL);
		__insn_mtspr(SPR_SNCTL, snctl & ~SPR_SNCTL__FRZPROC_MASK);
	}
#endif

#ifdef CONFIG_HARDWALL
	/* Enable or disable access to the network registers appropriately. */
	hardwall_switch_tasks(prev, next);
#endif

	/*
	 * Switch kernel SP, PC, and callee-saved registers.
	 * In the context of the new task, return the old task pointer
	 * (i.e. the task that actually called __switch_to).
	 * Pass the value to use for SYSTEM_SAVE_K_0 when we reset our sp.
	 */
	return __switch_to(prev, next, next_current_ksp0(next));
}

/*
 * This routine is called on return from interrupt if any of the
 * TIF_WORK_MASK flags are set in thread_info->flags.  It is
 * entered with interrupts disabled so we don't miss an event
 * that modified the thread_info flags.  If any flag is set, we
 * handle it and return, and the calling assembly code will
 * re-disable interrupts, reload the thread flags, and call back
 * if more flags need to be handled.
 *
 * We return whether we need to check the thread_info flags again
 * or not.  Note that we don't clear TIF_SINGLESTEP here, so it's
 * important that it be tested last, and then claim that we don't
 * need to recheck the flags.
 */
int do_work_pending(struct pt_regs *regs, u32 thread_info_flags)
{
	/* If we enter in kernel mode, do nothing and exit the caller loop. */
	if (!user_mode(regs))
		return 0;

	/* Enable interrupts; they are disabled again on return to caller. */
	local_irq_enable();

	if (thread_info_flags & _TIF_NEED_RESCHED) {
		schedule();
		return 1;
	}
#if CHIP_HAS_TILE_DMA() || CHIP_HAS_SN_PROC()
	if (thread_info_flags & _TIF_ASYNC_TLB) {
		do_async_page_fault(regs);
		return 1;
	}
#endif
	if (thread_info_flags & _TIF_SIGPENDING) {
		do_signal(regs);
		return 1;
	}
	if (thread_info_flags & _TIF_NOTIFY_RESUME) {
		clear_thread_flag(TIF_NOTIFY_RESUME);
		tracehook_notify_resume(regs);
		return 1;
	}
	if (thread_info_flags & _TIF_SINGLESTEP) {
		single_step_once(regs);
		return 0;
	}
	panic("work_pending: bad flags %#x\n", thread_info_flags);
}

unsigned long get_wchan(struct task_struct *p)
{
	struct KBacktraceIterator kbt;

	if (!p || p == current || p->state == TASK_RUNNING)
		return 0;

	for (KBacktraceIterator_init(&kbt, p, NULL);
	     !KBacktraceIterator_end(&kbt);
	     KBacktraceIterator_next(&kbt)) {
		if (!in_sched_functions(kbt.it.pc))
			return kbt.it.pc;
	}

	return 0;
}

/* Flush thread state. */
void flush_thread(void)
{
	/* Nothing */
}

/*
 * Free current thread data structures etc..
 */
void exit_thread(void)
{
	/* Nothing */
}

void show_regs(struct pt_regs *regs)
{
	struct task_struct *tsk = validate_current();
	int i;

	pr_err("\n");
	pr_err(" Pid: %d, comm: %20s, CPU: %d\n",
	       tsk->pid, tsk->comm, smp_processor_id());
#ifdef __tilegx__
	for (i = 0; i < 51; i += 3)
		pr_err(" r%-2d: "REGFMT" r%-2d: "REGFMT" r%-2d: "REGFMT"\n",
		       i, regs->regs[i], i+1, regs->regs[i+1],
		       i+2, regs->regs[i+2]);
	pr_err(" r51: "REGFMT" r52: "REGFMT" tp : "REGFMT"\n",
	       regs->regs[51], regs->regs[52], regs->tp);
	pr_err(" sp : "REGFMT" lr : "REGFMT"\n", regs->sp, regs->lr);
#else
	for (i = 0; i < 52; i += 4)
		pr_err(" r%-2d: "REGFMT" r%-2d: "REGFMT
		       " r%-2d: "REGFMT" r%-2d: "REGFMT"\n",
		       i, regs->regs[i], i+1, regs->regs[i+1],
		       i+2, regs->regs[i+2], i+3, regs->regs[i+3]);
	pr_err(" r52: "REGFMT" tp : "REGFMT" sp : "REGFMT" lr : "REGFMT"\n",
	       regs->regs[52], regs->tp, regs->sp, regs->lr);
#endif
	pr_err(" pc : "REGFMT" ex1: %ld     faultnum: %ld\n",
	       regs->pc, regs->ex1, regs->faultnum);

	dump_stack_regs(regs);
}