summaryrefslogtreecommitdiffstats
path: root/sys/i386/i386/vm_machdep.c
blob: 10e262334d8c68188eae47c7ef497c07d0500dba (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
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
/*-
 * Copyright (c) 1982, 1986 The Regents of the University of California.
 * Copyright (c) 1989, 1990 William Jolitz
 * Copyright (c) 1994 John Dyson
 * All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * the Systems Programming Group of the University of Utah Computer
 * Science Department, and William Jolitz.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	from: @(#)vm_machdep.c	7.3 (Berkeley) 5/13/91
 *	Utah $Hdr: vm_machdep.c 1.16.1.1 89/06/23$
 */

#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");

#include "opt_isa.h"
#include "opt_npx.h"
#ifdef PC98
#include "opt_pc98.h"
#endif
#include "opt_reset.h"
#include "opt_cpu.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/bio.h>
#include <sys/buf.h>
#include <sys/kse.h>
#include <sys/kernel.h>
#include <sys/ktr.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/mutex.h>
#include <sys/pioctl.h>
#include <sys/proc.h>
#include <sys/sf_buf.h>
#include <sys/smp.h>
#include <sys/sysctl.h>
#include <sys/unistd.h>
#include <sys/vnode.h>
#include <sys/vmmeter.h>

#include <machine/cpu.h>
#include <machine/cputypes.h>
#include <machine/md_var.h>
#include <machine/pcb.h>
#include <machine/pcb_ext.h>
#include <machine/vm86.h>

#ifdef CPU_ELAN
#include <machine/elan_mmcr.h>
#endif

#include <vm/vm.h>
#include <vm/vm_extern.h>
#include <vm/vm_kern.h>
#include <vm/vm_page.h>
#include <vm/vm_map.h>
#include <vm/vm_param.h>

#ifdef PC98
#include <pc98/pc98/pc98.h>
#else
#include <i386/isa/isa.h>
#endif

#ifndef NSFBUFS
#define	NSFBUFS		(512 + maxusers * 16)
#endif

static void	cpu_reset_real(void);
static void	sf_buf_init(void *arg);
SYSINIT(sock_sf, SI_SUB_MBUF, SI_ORDER_ANY, sf_buf_init, NULL)

LIST_HEAD(sf_head, sf_buf);

/*
 * A hash table of active sendfile(2) buffers
 */
static struct sf_head *sf_buf_active;
static u_long sf_buf_hashmask;

#define	SF_BUF_HASH(m)	(((m) - vm_page_array) & sf_buf_hashmask)

static TAILQ_HEAD(, sf_buf) sf_buf_freelist;
static u_int	sf_buf_alloc_want;

/*
 * A lock used to synchronize access to the hash table and free list
 */
static struct mtx sf_buf_lock;

extern int	_ucodesel, _udatasel;

/*
 * Finish a fork operation, with process p2 nearly set up.
 * Copy and update the pcb, set up the stack so that the child
 * ready to run and return to user mode.
 */
void
cpu_fork(td1, p2, td2, flags)
	register struct thread *td1;
	register struct proc *p2;
	struct thread *td2;
	int flags;
{
	register struct proc *p1;
	struct pcb *pcb2;
	struct mdproc *mdp2;
#ifdef DEV_NPX
	register_t savecrit;
#endif

	p1 = td1->td_proc;
	if ((flags & RFPROC) == 0) {
		if ((flags & RFMEM) == 0) {
			/* unshare user LDT */
			struct mdproc *mdp1 = &p1->p_md;
			struct proc_ldt *pldt = mdp1->md_ldt;
			if (pldt && pldt->ldt_refcnt > 1) {
				pldt = user_ldt_alloc(mdp1, pldt->ldt_len);
				if (pldt == NULL)
					panic("could not copy LDT");
				mdp1->md_ldt = pldt;
				set_user_ldt(mdp1);
				user_ldt_free(td1);
			}
		}
		return;
	}

	/* Ensure that p1's pcb is up to date. */
#ifdef DEV_NPX
	if (td1 == curthread)
		td1->td_pcb->pcb_gs = rgs();
	savecrit = intr_disable();
	if (PCPU_GET(fpcurthread) == td1)
		npxsave(&td1->td_pcb->pcb_save);
	intr_restore(savecrit);
#endif

	/* Point the pcb to the top of the stack */
	pcb2 = (struct pcb *)(td2->td_kstack +
	    td2->td_kstack_pages * PAGE_SIZE) - 1;
	td2->td_pcb = pcb2;

	/* Copy p1's pcb */
	bcopy(td1->td_pcb, pcb2, sizeof(*pcb2));

	/* Point mdproc and then copy over td1's contents */
	mdp2 = &p2->p_md;
	bcopy(&p1->p_md, mdp2, sizeof(*mdp2));

	/*
	 * Create a new fresh stack for the new process.
	 * Copy the trap frame for the return to user mode as if from a
	 * syscall.  This copies most of the user mode register values.
	 * The -16 is so we can expand the trapframe if we go to vm86.
	 */
	td2->td_frame = (struct trapframe *)((caddr_t)td2->td_pcb - 16) - 1;
	bcopy(td1->td_frame, td2->td_frame, sizeof(struct trapframe));

	td2->td_frame->tf_eax = 0;		/* Child returns zero */
	td2->td_frame->tf_eflags &= ~PSL_C;	/* success */
	td2->td_frame->tf_edx = 1;

	/*
	 * If the parent process has the trap bit set (i.e. a debugger had
	 * single stepped the process to the system call), we need to clear
	 * the trap flag from the new frame unless the debugger had set PF_FORK
	 * on the parent.  Otherwise, the child will receive a (likely
	 * unexpected) SIGTRAP when it executes the first instruction after
	 * returning  to userland.
	 */
	if ((p1->p_pfsflags & PF_FORK) == 0)
		td2->td_frame->tf_eflags &= ~PSL_T;

	/*
	 * Set registers for trampoline to user mode.  Leave space for the
	 * return address on stack.  These are the kernel mode register values.
	 */
#ifdef PAE
	pcb2->pcb_cr3 = vtophys(vmspace_pmap(p2->p_vmspace)->pm_pdpt);
#else
	pcb2->pcb_cr3 = vtophys(vmspace_pmap(p2->p_vmspace)->pm_pdir);
#endif
	pcb2->pcb_edi = 0;
	pcb2->pcb_esi = (int)fork_return;	/* fork_trampoline argument */
	pcb2->pcb_ebp = 0;
	pcb2->pcb_esp = (int)td2->td_frame - sizeof(void *);
	pcb2->pcb_ebx = (int)td2;		/* fork_trampoline argument */
	pcb2->pcb_eip = (int)fork_trampoline;
	pcb2->pcb_psl = PSL_KERNEL;		/* ints disabled */
	pcb2->pcb_gs = rgs();
	/*-
	 * pcb2->pcb_dr*:	cloned above.
	 * pcb2->pcb_savefpu:	cloned above.
	 * pcb2->pcb_flags:	cloned above.
	 * pcb2->pcb_onfault:	cloned above (always NULL here?).
	 * pcb2->pcb_gs:	cloned above.
	 * pcb2->pcb_ext:	cleared below.
	 */

	/*
	 * XXX don't copy the i/o pages.  this should probably be fixed.
	 */
	pcb2->pcb_ext = 0;

        /* Copy the LDT, if necessary. */
	mtx_lock_spin(&sched_lock);
        if (mdp2->md_ldt != 0) {
		if (flags & RFMEM) {
			mdp2->md_ldt->ldt_refcnt++;
		} else {
			mdp2->md_ldt = user_ldt_alloc(mdp2,
			    mdp2->md_ldt->ldt_len);
			if (mdp2->md_ldt == NULL)
				panic("could not copy LDT");
		}
        }
	mtx_unlock_spin(&sched_lock);

	/*
	 * Now, cpu_switch() can schedule the new process.
	 * pcb_esp is loaded pointing to the cpu_switch() stack frame
	 * containing the return address when exiting cpu_switch.
	 * This will normally be to fork_trampoline(), which will have
	 * %ebx loaded with the new proc's pointer.  fork_trampoline()
	 * will set up a stack to call fork_return(p, frame); to complete
	 * the return to user-mode.
	 */
}

/*
 * Intercept the return address from a freshly forked process that has NOT
 * been scheduled yet.
 *
 * This is needed to make kernel threads stay in kernel mode.
 */
void
cpu_set_fork_handler(td, func, arg)
	struct thread *td;
	void (*func)(void *);
	void *arg;
{
	/*
	 * Note that the trap frame follows the args, so the function
	 * is really called like this:  func(arg, frame);
	 */
	td->td_pcb->pcb_esi = (int) func;	/* function */
	td->td_pcb->pcb_ebx = (int) arg;	/* first arg */
}

void
cpu_exit(struct thread *td)
{
	struct mdproc *mdp;
	struct pcb *pcb = td->td_pcb; 


	/* Reset pc->pcb_gs and %gs before possibly invalidating it. */
	mdp = &td->td_proc->p_md;
	if (mdp->md_ldt) {
		td->td_pcb->pcb_gs = _udatasel;
		load_gs(_udatasel);
		user_ldt_free(td);
	}
	if (pcb->pcb_flags & PCB_DBREGS) {
		/* disable all hardware breakpoints */
		reset_dbregs();
		pcb->pcb_flags &= ~PCB_DBREGS;
	}
}

void
cpu_thread_exit(struct thread *td)
{
	struct pcb *pcb = td->td_pcb; 
#ifdef DEV_NPX
	if (td == PCPU_GET(fpcurthread))
		npxdrop();
#endif
        if (pcb->pcb_flags & PCB_DBREGS) {
		/* disable all hardware breakpoints */
                reset_dbregs();
                pcb->pcb_flags &= ~PCB_DBREGS;
        }
}

void
cpu_thread_clean(struct thread *td)
{
	struct pcb *pcb;

	pcb = td->td_pcb; 
	if (pcb->pcb_ext != 0) {
		/* XXXKSE  XXXSMP  not SMP SAFE.. what locks do we have? */
		/* if (pcb->pcb_ext->ext_refcount-- == 1) ?? */
		/*
		 * XXX do we need to move the TSS off the allocated pages
		 * before freeing them?  (not done here)
		 */
		kmem_free(kernel_map, (vm_offset_t)pcb->pcb_ext,
		    ctob(IOPAGES + 1));
		pcb->pcb_ext = 0;
	}
}

void
cpu_thread_swapin(struct thread *td)
{
}

void
cpu_thread_swapout(struct thread *td)
{
}

void
cpu_thread_setup(struct thread *td)
{

	td->td_pcb = (struct pcb *)(td->td_kstack +
	    td->td_kstack_pages * PAGE_SIZE) - 1;
	td->td_frame = (struct trapframe *)((caddr_t)td->td_pcb - 16) - 1;
	td->td_pcb->pcb_ext = NULL; 
}

/*
 * Initialize machine state (pcb and trap frame) for a new thread about to
 * upcall. Pu t enough state in the new thread's PCB to get it to go back 
 * userret(), where we can intercept it again to set the return (upcall)
 * Address and stack, along with those from upcals that are from other sources
 * such as those generated in thread_userret() itself.
 */
void
cpu_set_upcall(struct thread *td, struct thread *td0)
{
	struct pcb *pcb2;

	/* Point the pcb to the top of the stack. */
	pcb2 = td->td_pcb;

	/*
	 * Copy the upcall pcb.  This loads kernel regs.
	 * Those not loaded individually below get their default
	 * values here.
	 *
	 * XXXKSE It might be a good idea to simply skip this as
	 * the values of the other registers may be unimportant.
	 * This would remove any requirement for knowing the KSE
	 * at this time (see the matching comment below for
	 * more analysis) (need a good safe default).
	 */
	bcopy(td0->td_pcb, pcb2, sizeof(*pcb2));
	pcb2->pcb_flags &= ~(PCB_NPXTRAP|PCB_NPXINITDONE);

	/*
	 * Create a new fresh stack for the new thread.
	 * The -16 is so we can expand the trapframe if we go to vm86.
	 * Don't forget to set this stack value into whatever supplies
	 * the address for the fault handlers.
	 * The contexts are filled in at the time we actually DO the
	 * upcall as only then do we know which KSE we got.
	 */
	bcopy(td0->td_frame, td->td_frame, sizeof(struct trapframe));

	/*
	 * Set registers for trampoline to user mode.  Leave space for the
	 * return address on stack.  These are the kernel mode register values.
	 */
#ifdef PAE
	pcb2->pcb_cr3 = vtophys(vmspace_pmap(td->td_proc->p_vmspace)->pm_pdpt);
#else
	pcb2->pcb_cr3 = vtophys(vmspace_pmap(td->td_proc->p_vmspace)->pm_pdir);
#endif
	pcb2->pcb_edi = 0;
	pcb2->pcb_esi = (int)fork_return;		    /* trampoline arg */
	pcb2->pcb_ebp = 0;
	pcb2->pcb_esp = (int)td->td_frame - sizeof(void *); /* trampoline arg */
	pcb2->pcb_ebx = (int)td;			    /* trampoline arg */
	pcb2->pcb_eip = (int)fork_trampoline;
	pcb2->pcb_psl &= ~(PSL_I);	/* interrupts must be disabled */
	pcb2->pcb_gs = rgs();
	/*
	 * If we didn't copy the pcb, we'd need to do the following registers:
	 * pcb2->pcb_dr*:	cloned above.
	 * pcb2->pcb_savefpu:	cloned above.
	 * pcb2->pcb_flags:	cloned above.
	 * pcb2->pcb_onfault:	cloned above (always NULL here?).
	 * pcb2->pcb_gs:	cloned above.  XXXKSE ???
	 * pcb2->pcb_ext:	cleared below.
	 */
	 pcb2->pcb_ext = NULL;
}

/*
 * Set that machine state for performing an upcall that has to
 * be done in thread_userret() so that those upcalls generated
 * in thread_userret() itself can be done as well.
 */
void
cpu_set_upcall_kse(struct thread *td, struct kse_upcall *ku)
{

	/* 
	 * Do any extra cleaning that needs to be done.
	 * The thread may have optional components
	 * that are not present in a fresh thread.
	 * This may be a recycled thread so make it look
	 * as though it's newly allocated.
	 */
	cpu_thread_clean(td);

	/*
	 * Set the trap frame to point at the beginning of the uts
	 * function.
	 */
	td->td_frame->tf_ebp = 0; 
	td->td_frame->tf_esp =
	    (int)ku->ku_stack.ss_sp + ku->ku_stack.ss_size - 16;
	td->td_frame->tf_eip = (int)ku->ku_func;

	/*
	 * Pass the address of the mailbox for this kse to the uts
	 * function as a parameter on the stack.
	 */
	suword((void *)(td->td_frame->tf_esp + sizeof(void *)),
	    (int)ku->ku_mailbox);
}

/*
 * Convert kernel VA to physical address
 */
vm_paddr_t
kvtop(void *addr)
{
	vm_paddr_t pa;

	pa = pmap_kextract((vm_offset_t)addr);
	if (pa == 0)
		panic("kvtop: zero page frame");
	return (pa);
}

void
cpu_reset()
{
#ifdef SMP
	u_int map;

	if (smp_active) {
		map = PCPU_GET(other_cpus) & ~stopped_cpus;
		if (map != 0) {
			printf("cpu_reset: Stopping other CPUs\n");
			stop_cpus(map);
		}
		DELAY(1000000);
	}
#endif
	cpu_reset_real();
	/* NOTREACHED */
}

static void
cpu_reset_real()
{

#ifdef CPU_ELAN
	if (elan_mmcr != NULL)
		elan_mmcr->RESCFG = 1;
#endif

	if (cpu == CPU_GEODE1100) {
		/* Attempt Geode's own reset */
		outl(0xcf8, 0x80009044ul);
		outl(0xcfc, 0xf);
	}

#ifdef PC98
	/*
	 * Attempt to do a CPU reset via CPU reset port.
	 */
	disable_intr();
	if ((inb(0x35) & 0xa0) != 0xa0) {
		outb(0x37, 0x0f);		/* SHUT0 = 0. */
		outb(0x37, 0x0b);		/* SHUT1 = 0. */
	}
	outb(0xf0, 0x00);		/* Reset. */
#else
#if !defined(BROKEN_KEYBOARD_RESET)
	/*
	 * Attempt to do a CPU reset via the keyboard controller,
	 * do not turn off GateA20, as any machine that fails
	 * to do the reset here would then end up in no man's land.
	 */
	outb(IO_KBD + 4, 0xFE);
	DELAY(500000);	/* wait 0.5 sec to see if that did it */
	printf("Keyboard reset did not work, attempting CPU shutdown\n");
	DELAY(1000000);	/* wait 1 sec for printf to complete */
#endif
#endif /* PC98 */

	/* Force a shutdown by unmapping entire address space. */
	bzero((caddr_t)PTD, NBPTD);

	/* "good night, sweet prince .... <THUNK!>" */
	invltlb();
	/* NOTREACHED */
	while(1);
}

/*
 * Allocate a pool of sf_bufs (sendfile(2) or "super-fast" if you prefer. :-))
 */
static void
sf_buf_init(void *arg)
{
	struct sf_buf *sf_bufs;
	vm_offset_t sf_base;
	int i;

	nsfbufs = NSFBUFS;
	TUNABLE_INT_FETCH("kern.ipc.nsfbufs", &nsfbufs);

	sf_buf_active = hashinit(nsfbufs, M_TEMP, &sf_buf_hashmask);
	TAILQ_INIT(&sf_buf_freelist);
	sf_base = kmem_alloc_nofault(kernel_map, nsfbufs * PAGE_SIZE);
	sf_bufs = malloc(nsfbufs * sizeof(struct sf_buf), M_TEMP,
	    M_NOWAIT | M_ZERO);
	for (i = 0; i < nsfbufs; i++) {
		sf_bufs[i].kva = sf_base + i * PAGE_SIZE;
		TAILQ_INSERT_TAIL(&sf_buf_freelist, &sf_bufs[i], free_entry);
	}
	sf_buf_alloc_want = 0;
	mtx_init(&sf_buf_lock, "sf_buf", NULL, MTX_DEF);
}

/*
 * Get an sf_buf from the freelist. Will block if none are available.
 */
struct sf_buf *
sf_buf_alloc(struct vm_page *m, int flags)
{
	pt_entry_t opte, *ptep;
	struct sf_head *hash_list;
	struct sf_buf *sf;
	int error;

	hash_list = &sf_buf_active[SF_BUF_HASH(m)];
	mtx_lock(&sf_buf_lock);
	LIST_FOREACH(sf, hash_list, list_entry) {
		if (sf->m == m) {
			sf->ref_count++;
			if (sf->ref_count == 1) {
				TAILQ_REMOVE(&sf_buf_freelist, sf, free_entry);
				nsfbufsused++;
				nsfbufspeak = imax(nsfbufspeak, nsfbufsused);
			}
			goto done;
		}
	}
	while ((sf = TAILQ_FIRST(&sf_buf_freelist)) == NULL) {
		if (flags & SFB_NOWAIT)
			goto done;
		sf_buf_alloc_want++;
		mbstat.sf_allocwait++;
		error = msleep(&sf_buf_freelist, &sf_buf_lock,
		    (flags & SFB_CATCH) ? PCATCH | PVM : PVM, "sfbufa", 0);
		sf_buf_alloc_want--;

		/*
		 * If we got a signal, don't risk going back to sleep. 
		 */
		if (error)
			goto done;
	}
	TAILQ_REMOVE(&sf_buf_freelist, sf, free_entry);
	if (sf->m != NULL)
		LIST_REMOVE(sf, list_entry);
	LIST_INSERT_HEAD(hash_list, sf, list_entry);
	sf->ref_count = 1;
	sf->m = m;
	nsfbufsused++;
	nsfbufspeak = imax(nsfbufspeak, nsfbufsused);

	/*
	 * Update the sf_buf's virtual-to-physical mapping, flushing the
	 * virtual address from the TLB only if the PTE implies that the old
	 * mapping has been used.  Since the reference count for the sf_buf's
	 * old mapping was zero, that mapping is not currently in use.
	 * Consequently, there is no need to exchange the old and new PTEs
	 * atomically, even under PAE.
	 */
	ptep = vtopte(sf->kva);
	opte = *ptep;
	*ptep = VM_PAGE_TO_PHYS(m) | pgeflag | PG_RW | PG_V;
	if ((opte & (PG_A | PG_V)) == (PG_A | PG_V))
		pmap_invalidate_page(kernel_pmap, sf->kva);
done:
	mtx_unlock(&sf_buf_lock);
	return (sf);
}

/*
 * Remove a reference from the given sf_buf, adding it to the free
 * list when its reference count reaches zero.  A freed sf_buf still,
 * however, retains its virtual-to-physical mapping until it is
 * recycled or reactivated by sf_buf_alloc(9).
 */
void
sf_buf_free(struct sf_buf *sf)
{

	mtx_lock(&sf_buf_lock);
	sf->ref_count--;
	if (sf->ref_count == 0) {
		TAILQ_INSERT_TAIL(&sf_buf_freelist, sf, free_entry);
		nsfbufsused--;
		if (sf_buf_alloc_want > 0)
			wakeup_one(&sf_buf_freelist);
	}
	mtx_unlock(&sf_buf_lock);
}

/*
 * Software interrupt handler for queued VM system processing.
 */   
void  
swi_vm(void *dummy) 
{     
	if (busdma_swi_pending != 0)
		busdma_swi();
}

/*
 * Tell whether this address is in some physical memory region.
 * Currently used by the kernel coredump code in order to avoid
 * dumping the ``ISA memory hole'' which could cause indefinite hangs,
 * or other unpredictable behaviour.
 */

int
is_physical_memory(vm_paddr_t addr)
{

#ifdef DEV_ISA
	/* The ISA ``memory hole''. */
	if (addr >= 0xa0000 && addr < 0x100000)
		return 0;
#endif

	/*
	 * stuff other tests for known memory-mapped devices (PCI?)
	 * here
	 */

	return 1;
}
OpenPOWER on IntegriCloud