subr_hash.c revision 137244 
1/*
2 * Copyright (c) 1982, 1986, 1991, 1993
3 *	The Regents of the University of California.  All rights reserved.
4 * (c) UNIX System Laboratories, Inc.
5 * All or some portions of this file are derived from material licensed
6 * to the University of California by American Telephone and Telegraph
7 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
8 * the permission of UNIX System Laboratories, Inc.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 *    notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 *    notice, this list of conditions and the following disclaimer in the
17 *    documentation and/or other materials provided with the distribution.
18 * 4. Neither the name of the University nor the names of its contributors
19 *    may be used to endorse or promote products derived from this software
20 *    without specific prior written permission.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * SUCH DAMAGE.
33 *
34 *	@(#)kern_subr.c	8.3 (Berkeley) 1/21/94
35 */
36
37#include <sys/cdefs.h>
38__FBSDID("$FreeBSD: head/sys/kern/kern_subr.c 137244 2004-11-05 06:52:29Z alc $");
39
40#include "opt_zero.h"
41
42#include <sys/param.h>
43#include <sys/systm.h>
44#include <sys/kernel.h>
45#include <sys/ktr.h>
46#include <sys/limits.h>
47#include <sys/lock.h>
48#include <sys/mutex.h>
49#include <sys/proc.h>
50#include <sys/malloc.h>
51#include <sys/resourcevar.h>
52#include <sys/sched.h>
53#include <sys/sysctl.h>
54#include <sys/vnode.h>
55
56#include <vm/vm.h>
57#include <vm/vm_page.h>
58#include <vm/vm_map.h>
59#ifdef ZERO_COPY_SOCKETS
60#include <vm/vm_param.h>
61#include <vm/vm_object.h>
62#endif
63
64SYSCTL_INT(_kern, KERN_IOV_MAX, iov_max, CTLFLAG_RD, NULL, UIO_MAXIOV,
65	"Maximum number of elements in an I/O vector; sysconf(_SC_IOV_MAX)");
66
67#ifdef ZERO_COPY_SOCKETS
68/* Declared in uipc_socket.c */
69extern int so_zero_copy_receive;
70
71static int
72vm_pgmoveco(vm_map_t mapa, vm_object_t srcobj, vm_offset_t kaddr,
73    vm_offset_t uaddr)
74{
75	vm_map_t map = mapa;
76	vm_page_t kern_pg, user_pg;
77	vm_object_t uobject;
78	vm_map_entry_t entry;
79	vm_pindex_t upindex;
80	vm_prot_t prot;
81	boolean_t wired;
82
83	/*
84	 * First lookup the kernel page.
85	 */
86	kern_pg = PHYS_TO_VM_PAGE(vtophys(kaddr));
87	/*
88	 * XXX The vm object containing kern_pg needs locking.
89	 */
90	if ((vm_map_lookup(&map, uaddr,
91			   VM_PROT_WRITE, &entry, &uobject,
92			   &upindex, &prot, &wired)) != KERN_SUCCESS) {
93		return(EFAULT);
94	}
95	VM_OBJECT_LOCK(uobject);
96	if ((user_pg = vm_page_lookup(uobject, upindex)) != NULL) {
97		do
98			vm_page_lock_queues();
99		while (vm_page_sleep_if_busy(user_pg, 1, "vm_pgmoveco"));
100		pmap_remove_all(user_pg);
101		vm_page_free(user_pg);
102	} else
103		vm_page_lock_queues();
104	if (kern_pg->busy || ((kern_pg->queue - kern_pg->pc) == PQ_FREE) ||
105	    (kern_pg->hold_count != 0)|| (kern_pg->flags & PG_BUSY)) {
106		printf("vm_pgmoveco: pindex(%lu), busy(%d), PG_BUSY(%d), "
107		       "hold(%d) paddr(0x%lx)\n", (u_long)kern_pg->pindex,
108			kern_pg->busy, (kern_pg->flags & PG_BUSY) ? 1 : 0,
109			kern_pg->hold_count, (u_long)kern_pg->phys_addr);
110		if ((kern_pg->queue - kern_pg->pc) == PQ_FREE)
111			panic("vm_pgmoveco: renaming free page");
112		else
113			panic("vm_pgmoveco: renaming busy page");
114	}
115	vm_page_rename(kern_pg, uobject, upindex);
116	kern_pg->valid = VM_PAGE_BITS_ALL;
117	vm_page_unlock_queues();
118	VM_OBJECT_UNLOCK(uobject);
119	vm_map_lookup_done(map, entry);
120	return(KERN_SUCCESS);
121}
122#endif /* ZERO_COPY_SOCKETS */
123
124int
125uiomove(void *cp, int n, struct uio *uio)
126{
127	struct thread *td = curthread;
128	struct iovec *iov;
129	u_int cnt;
130	int error = 0;
131	int save = 0;
132
133	KASSERT(uio->uio_rw == UIO_READ || uio->uio_rw == UIO_WRITE,
134	    ("uiomove: mode"));
135	KASSERT(uio->uio_segflg != UIO_USERSPACE || uio->uio_td == curthread,
136	    ("uiomove proc"));
137	WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL,
138	    "Calling uiomove()");
139
140	save = td->td_pflags & TDP_DEADLKTREAT;
141	td->td_pflags |= TDP_DEADLKTREAT;
142
143	while (n > 0 && uio->uio_resid) {
144		iov = uio->uio_iov;
145		cnt = iov->iov_len;
146		if (cnt == 0) {
147			uio->uio_iov++;
148			uio->uio_iovcnt--;
149			continue;
150		}
151		if (cnt > n)
152			cnt = n;
153
154		switch (uio->uio_segflg) {
155
156		case UIO_USERSPACE:
157			if (ticks - PCPU_GET(switchticks) >= hogticks)
158				uio_yield();
159			if (uio->uio_rw == UIO_READ)
160				error = copyout(cp, iov->iov_base, cnt);
161			else
162				error = copyin(iov->iov_base, cp, cnt);
163			if (error)
164				goto out;
165			break;
166
167		case UIO_SYSSPACE:
168			if (uio->uio_rw == UIO_READ)
169				bcopy(cp, iov->iov_base, cnt);
170			else
171				bcopy(iov->iov_base, cp, cnt);
172			break;
173		case UIO_NOCOPY:
174			break;
175		}
176		iov->iov_base = (char *)iov->iov_base + cnt;
177		iov->iov_len -= cnt;
178		uio->uio_resid -= cnt;
179		uio->uio_offset += cnt;
180		cp = (char *)cp + cnt;
181		n -= cnt;
182	}
183out:
184	if (save == 0)
185		td->td_pflags &= ~TDP_DEADLKTREAT;
186	return (error);
187}
188
189/*
190 * Wrapper for uiomove() that validates the arguments against a known-good
191 * kernel buffer.  Currently, uiomove accepts a signed (n) argument, which
192 * is almost definitely a bad thing, so we catch that here as well.  We
193 * return a runtime failure, but it might be desirable to generate a runtime
194 * assertion failure instead.
195 */
196int
197uiomove_frombuf(void *buf, int buflen, struct uio *uio)
198{
199	unsigned int offset, n;
200
201	if (uio->uio_offset < 0 || uio->uio_resid < 0 ||
202	    (offset = uio->uio_offset) != uio->uio_offset)
203		return (EINVAL);
204	if (buflen <= 0 || offset >= buflen)
205		return (0);
206	if ((n = buflen - offset) > INT_MAX)
207		return (EINVAL);
208	return (uiomove((char *)buf + offset, n, uio));
209}
210
211#ifdef ZERO_COPY_SOCKETS
212/*
213 * Experimental support for zero-copy I/O
214 */
215static int
216userspaceco(void *cp, u_int cnt, struct uio *uio, struct vm_object *obj,
217    int disposable)
218{
219	struct iovec *iov;
220	int error;
221
222	iov = uio->uio_iov;
223	if (uio->uio_rw == UIO_READ) {
224		if ((so_zero_copy_receive != 0)
225		 && (obj != NULL)
226		 && ((cnt & PAGE_MASK) == 0)
227		 && ((((intptr_t) iov->iov_base) & PAGE_MASK) == 0)
228		 && ((uio->uio_offset & PAGE_MASK) == 0)
229		 && ((((intptr_t) cp) & PAGE_MASK) == 0)
230		 && (obj->type == OBJT_DEFAULT)
231		 && (disposable != 0)) {
232			/* SOCKET: use page-trading */
233			/*
234			 * We only want to call vm_pgmoveco() on
235			 * disposeable pages, since it gives the
236			 * kernel page to the userland process.
237			 */
238			error =	vm_pgmoveco(&curproc->p_vmspace->vm_map,
239					    obj, (vm_offset_t)cp,
240					    (vm_offset_t)iov->iov_base);
241
242			/*
243			 * If we get an error back, attempt
244			 * to use copyout() instead.  The
245			 * disposable page should be freed
246			 * automatically if we weren't able to move
247			 * it into userland.
248			 */
249			if (error != 0)
250				error = copyout(cp, iov->iov_base, cnt);
251		} else {
252			error = copyout(cp, iov->iov_base, cnt);
253		}
254	} else {
255		error = copyin(iov->iov_base, cp, cnt);
256	}
257	return (error);
258}
259
260int
261uiomoveco(void *cp, int n, struct uio *uio, struct vm_object *obj,
262    int disposable)
263{
264	struct iovec *iov;
265	u_int cnt;
266	int error;
267
268	KASSERT(uio->uio_rw == UIO_READ || uio->uio_rw == UIO_WRITE,
269	    ("uiomoveco: mode"));
270	KASSERT(uio->uio_segflg != UIO_USERSPACE || uio->uio_td == curthread,
271	    ("uiomoveco proc"));
272
273	while (n > 0 && uio->uio_resid) {
274		iov = uio->uio_iov;
275		cnt = iov->iov_len;
276		if (cnt == 0) {
277			uio->uio_iov++;
278			uio->uio_iovcnt--;
279			continue;
280		}
281		if (cnt > n)
282			cnt = n;
283
284		switch (uio->uio_segflg) {
285
286		case UIO_USERSPACE:
287			if (ticks - PCPU_GET(switchticks) >= hogticks)
288				uio_yield();
289
290			error = userspaceco(cp, cnt, uio, obj, disposable);
291
292			if (error)
293				return (error);
294			break;
295
296		case UIO_SYSSPACE:
297			if (uio->uio_rw == UIO_READ)
298				bcopy(cp, iov->iov_base, cnt);
299			else
300				bcopy(iov->iov_base, cp, cnt);
301			break;
302		case UIO_NOCOPY:
303			break;
304		}
305		iov->iov_base = (char *)iov->iov_base + cnt;
306		iov->iov_len -= cnt;
307		uio->uio_resid -= cnt;
308		uio->uio_offset += cnt;
309		cp = (char *)cp + cnt;
310		n -= cnt;
311	}
312	return (0);
313}
314#endif /* ZERO_COPY_SOCKETS */
315
316/*
317 * Give next character to user as result of read.
318 */
319int
320ureadc(int c, struct uio *uio)
321{
322	struct iovec *iov;
323	char *iov_base;
324
325again:
326	if (uio->uio_iovcnt == 0 || uio->uio_resid == 0)
327		panic("ureadc");
328	iov = uio->uio_iov;
329	if (iov->iov_len == 0) {
330		uio->uio_iovcnt--;
331		uio->uio_iov++;
332		goto again;
333	}
334	switch (uio->uio_segflg) {
335
336	case UIO_USERSPACE:
337		if (subyte(iov->iov_base, c) < 0)
338			return (EFAULT);
339		break;
340
341	case UIO_SYSSPACE:
342		iov_base = iov->iov_base;
343		*iov_base = c;
344		iov->iov_base = iov_base;
345		break;
346
347	case UIO_NOCOPY:
348		break;
349	}
350	iov->iov_base = (char *)iov->iov_base + 1;
351	iov->iov_len--;
352	uio->uio_resid--;
353	uio->uio_offset++;
354	return (0);
355}
356
357/*
358 * General routine to allocate a hash table.
359 */
360void *
361hashinit(int elements, struct malloc_type *type, u_long *hashmask)
362{
363	long hashsize;
364	LIST_HEAD(generic, generic) *hashtbl;
365	int i;
366
367	if (elements <= 0)
368		panic("hashinit: bad elements");
369	for (hashsize = 1; hashsize <= elements; hashsize <<= 1)
370		continue;
371	hashsize >>= 1;
372	hashtbl = malloc((u_long)hashsize * sizeof(*hashtbl), type, M_WAITOK);
373	for (i = 0; i < hashsize; i++)
374		LIST_INIT(&hashtbl[i]);
375	*hashmask = hashsize - 1;
376	return (hashtbl);
377}
378
379void
380hashdestroy(void *vhashtbl, struct malloc_type *type, u_long hashmask)
381{
382	LIST_HEAD(generic, generic) *hashtbl, *hp;
383
384	hashtbl = vhashtbl;
385	for (hp = hashtbl; hp <= &hashtbl[hashmask]; hp++)
386		if (!LIST_EMPTY(hp))
387			panic("hashdestroy: hash not empty");
388	free(hashtbl, type);
389}
390
391static int primes[] = { 1, 13, 31, 61, 127, 251, 509, 761, 1021, 1531, 2039,
392			2557, 3067, 3583, 4093, 4603, 5119, 5623, 6143, 6653,
393			7159, 7673, 8191, 12281, 16381, 24571, 32749 };
394#define NPRIMES (sizeof(primes) / sizeof(primes[0]))
395
396/*
397 * General routine to allocate a prime number sized hash table.
398 */
399void *
400phashinit(int elements, struct malloc_type *type, u_long *nentries)
401{
402	long hashsize;
403	LIST_HEAD(generic, generic) *hashtbl;
404	int i;
405
406	if (elements <= 0)
407		panic("phashinit: bad elements");
408	for (i = 1, hashsize = primes[1]; hashsize <= elements;) {
409		i++;
410		if (i == NPRIMES)
411			break;
412		hashsize = primes[i];
413	}
414	hashsize = primes[i - 1];
415	hashtbl = malloc((u_long)hashsize * sizeof(*hashtbl), type, M_WAITOK);
416	for (i = 0; i < hashsize; i++)
417		LIST_INIT(&hashtbl[i]);
418	*nentries = hashsize;
419	return (hashtbl);
420}
421
422void
423uio_yield(void)
424{
425	struct thread *td;
426
427	td = curthread;
428	mtx_lock_spin(&sched_lock);
429	DROP_GIANT();
430	sched_prio(td, td->td_ksegrp->kg_user_pri); /* XXXKSE */
431	mi_switch(SW_INVOL, NULL);
432	mtx_unlock_spin(&sched_lock);
433	PICKUP_GIANT();
434}
435
436int
437copyinfrom(const void * __restrict src, void * __restrict dst, size_t len,
438    int seg)
439{
440	int error = 0;
441
442	switch (seg) {
443	case UIO_USERSPACE:
444		error = copyin(src, dst, len);
445		break;
446	case UIO_SYSSPACE:
447		bcopy(src, dst, len);
448		break;
449	default:
450		panic("copyinfrom: bad seg %d\n", seg);
451	}
452	return (error);
453}
454
455int
456copyinstrfrom(const void * __restrict src, void * __restrict dst, size_t len,
457    size_t * __restrict copied, int seg)
458{
459	int error = 0;
460
461	switch (seg) {
462	case UIO_USERSPACE:
463		error = copyinstr(src, dst, len, copied);
464		break;
465	case UIO_SYSSPACE:
466		error = copystr(src, dst, len, copied);
467		break;
468	default:
469		panic("copyinstrfrom: bad seg %d\n", seg);
470	}
471	return (error);
472}
473
474int
475copyiniov(struct iovec *iovp, u_int iovcnt, struct iovec **iov, int error)
476{
477	u_int iovlen;
478
479	*iov = NULL;
480	if (iovcnt > UIO_MAXIOV)
481		return (error);
482	iovlen = iovcnt * sizeof (struct iovec);
483	*iov = malloc(iovlen, M_IOV, M_WAITOK);
484	error = copyin(iovp, *iov, iovlen);
485	if (error) {
486		free(*iov, M_IOV);
487		*iov = NULL;
488	}
489	return (error);
490}
491
492int
493copyinuio(struct iovec *iovp, u_int iovcnt, struct uio **uiop)
494{
495	struct iovec *iov;
496	struct uio *uio;
497	u_int iovlen;
498	int error, i;
499
500	*uiop = NULL;
501	if (iovcnt > UIO_MAXIOV)
502		return (EINVAL);
503	iovlen = iovcnt * sizeof (struct iovec);
504	uio = malloc(iovlen + sizeof *uio, M_IOV, M_WAITOK);
505	iov = (struct iovec *)(uio + 1);
506	error = copyin(iovp, iov, iovlen);
507	if (error) {
508		free(uio, M_IOV);
509		return (error);
510	}
511	uio->uio_iov = iov;
512	uio->uio_iovcnt = iovcnt;
513	uio->uio_segflg = UIO_USERSPACE;
514	uio->uio_offset = -1;
515	uio->uio_resid = 0;
516	for (i = 0; i < iovcnt; i++) {
517		if (iov->iov_len > INT_MAX - uio->uio_resid) {
518			free(uio, M_IOV);
519			return (EINVAL);
520		}
521		uio->uio_resid += iov->iov_len;
522		iov++;
523	}
524	*uiop = uio;
525	return (0);
526}
527
528struct uio *
529cloneuio(struct uio *uiop)
530{
531	struct uio *uio;
532	int iovlen;
533
534	iovlen = uiop->uio_iovcnt * sizeof (struct iovec);
535	uio = malloc(iovlen + sizeof *uio, M_IOV, M_WAITOK);
536	*uio = *uiop;
537	uio->uio_iov = (struct iovec *)(uio + 1);
538	bcopy(uiop->uio_iov, uio->uio_iov, iovlen);
539	return (uio);
540}
541