[apple/xnu.git] / bsd / kern / kern_physio.c

/*
 * Copyright (c) 2000-2004 Apple Computer, Inc. All rights reserved.
 *
 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
 * 
 * This file contains Original Code and/or Modifications of Original Code
 * as defined in and that are subject to the Apple Public Source License
 * Version 2.0 (the 'License'). You may not use this file except in
 * compliance with the License. The rights granted to you under the License
 * may not be used to create, or enable the creation or redistribution of,
 * unlawful or unlicensed copies of an Apple operating system, or to
 * circumvent, violate, or enable the circumvention or violation of, any
 * terms of an Apple operating system software license agreement.
 * 
 * Please obtain a copy of the License at
 * http://www.opensource.apple.com/apsl/ and read it before using this file.
 * 
 * The Original Code and all software distributed under the License are
 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
 * Please see the License for the specific language governing rights and
 * limitations under the License.
 * 
 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
 */
/*-
 * Copyright (c) 1982, 1986, 1990, 1993
 *	The Regents of the University of California.  All rights reserved.
 * (c) UNIX System Laboratories, Inc.
 * All or some portions of this file are derived from material licensed
 * to the University of California by American Telephone and Telegraph
 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
 * the permission of UNIX System Laboratories, Inc.
 *
 * 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: @(#)kern_physio.c	8.1 (Berkeley) 6/10/93
 */
/*
 * HISTORY
 * 27-July-97  Umesh Vaishampayan  (umeshv@apple.com)
 *	Allow physio() to kernel space.
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/buf_internal.h>
#include <sys/conf.h>
#include <sys/proc_internal.h>
#include <sys/uio_internal.h>
#include <kern/assert.h>

int rawread(dev_t dev, struct uio *uio);
int rawwrite(dev_t dev, struct uio *uio);

int
physio( void (*f_strategy)(buf_t), 
	buf_t bp,
	dev_t dev,
	int flags,
	u_int (*f_minphys)(buf_t),
	struct uio *uio,
	int blocksize)
{
	struct proc *p = current_proc();
	int error, i, buf_allocated, todo, iosize;
	int orig_bflags = 0;
#if LP64KERN
	int64_t done;
#else
	int done;
#endif

	error = 0;
	flags &= B_READ | B_WRITE;
	buf_allocated = 0;

	/*
	 * [check user read/write access to the data buffer]
	 *
	 * Check each iov one by one.  Note that we know if we're reading or
	 * writing, so we ignore the uio's rw parameter.  Also note that if
	 * we're doing a read, that's a *write* to user-space.
	 */
	for (i = 0; i < uio->uio_iovcnt; i++) {
		if (UIO_SEG_IS_USER_SPACE(uio->uio_segflg)) {
			if (!useracc(uio_iov_base_at(uio, i),
					uio_iov_len_at(uio, i),
		    		(flags == B_READ) ? B_WRITE : B_READ))
			return (EFAULT);
		}
	}
	/*
	 * Make sure we have a buffer, creating one if necessary.
	 */
	if (bp == NULL) {
		bp = buf_alloc((vnode_t)0);
		buf_allocated = 1;
	} else
	        orig_bflags = buf_flags(bp);
	/*
	 * at this point we should have a buffer
	 * that is marked BL_BUSY... we either 
	 * acquired it via buf_alloc, or it was
	 * passed into us... if it was passed
	 * in, it needs to already be owned by
	 * the caller (i.e. BL_BUSY is set)
	 */
	assert(bp->b_lflags & BL_BUSY);

	/*
	 * [set up the fixed part of the buffer for a transfer]
	 */
	bp->b_dev = dev;
	bp->b_proc = p;

	/*
	 * [mark the buffer busy for physical I/O]
	 * (i.e. set B_PHYS (because it's an I/O to user
	 * memory, and B_RAW, because B_RAW is to be
	 * "Set by physio for raw transfers.", in addition
	 * to the read/write flag.)
	 */
        buf_setflags(bp, B_PHYS | B_RAW);

	/*
	 * [while there is data to transfer and no I/O error]
	 * Note that I/O errors are handled with a 'goto' at the bottom
	 * of the 'while' loop.
	 */
	for (i = 0; i < uio->uio_iovcnt; i++) {
		while (uio_iov_len_at(uio, i) > 0) {
			
			if ( (iosize = uio_iov_len_at(uio, i)) > MAXPHYSIO_WIRED)
			        iosize = MAXPHYSIO_WIRED;
			/*
			 * make sure we're set to issue a fresh I/O
			 * in the right direction
			 */
			buf_reset(bp, flags);

			/* [set up the buffer for a maximum-sized transfer] */
 			buf_setblkno(bp, uio->uio_offset / blocksize);
			buf_setcount(bp, iosize);
			// LP64todo - fix this!
			buf_setdataptr(bp, (uintptr_t)CAST_DOWN(caddr_t, uio_iov_base_at(uio, i)));
			
			/*
			 * [call f_minphys to bound the tranfer size]
			 * and remember the amount of data to transfer,
			 * for later comparison.
			 */
			(*f_minphys)(bp);
			todo = buf_count(bp);

			/*
			 * [lock the part of the user address space involved
			 *    in the transfer]
			 */

			if(UIO_SEG_IS_USER_SPACE(uio->uio_segflg)) {
				error = vslock(CAST_USER_ADDR_T(buf_dataptr(bp)),
					       (user_size_t)todo);
				if (error)
					goto done;
			}
			
			/* [call f_strategy to start the transfer] */
			(*f_strategy)(bp);


			/* [wait for the transfer to complete] */
			error = (int)buf_biowait(bp);

			/*
			 * [unlock the part of the address space previously
			 *    locked]
			 */
			if(UIO_SEG_IS_USER_SPACE(uio->uio_segflg))
				vsunlock(CAST_USER_ADDR_T(buf_dataptr(bp)),
					 (user_size_t)todo,
					 (flags & B_READ));

			/*
			 * [deduct the transfer size from the total number
			 *    of data to transfer]
			 */
			done = buf_count(bp) - buf_resid(bp);
			uio_iov_len_add_at(uio, -done, i);
			uio_iov_base_add_at(uio, done, i);
			uio->uio_offset += done;
			uio_setresid(uio, (uio_resid(uio) - done));

			/*
			 * Now, check for an error.
			 * Also, handle weird end-of-disk semantics.
			 */
			if (error || done < todo)
				goto done;
		}
	}

done:
	if (buf_allocated)
	        buf_free(bp);
	else
		buf_setflags(bp, orig_bflags);

	return (error);
}

/*
 * Leffler, et al., says on p. 231:
 * "The minphys() routine is called by physio() to adjust the
 * size of each I/O transfer before the latter is passed to
 * the strategy routine..." 
 *
 * so, just adjust the buffer's count accounting to MAXPHYS here,
 * and return the new count;
 */
u_int
minphys(struct buf *bp)
{

	buf_setcount(bp, min(MAXPHYS, buf_count(bp)));
        return buf_count(bp);
}

/*
 * Do a read on a device for a user process.
 */
int
rawread(dev_t dev, struct uio *uio)
{
	return (physio(cdevsw[major(dev)].d_strategy, (struct buf *)NULL,
	    dev, B_READ, minphys, uio, DEV_BSIZE));
}

/*
 * Do a write on a device for a user process.
 */
int
rawwrite(dev_t dev, struct uio *uio)
{
	return (physio(cdevsw[major(dev)].d_strategy, (struct buf *)NULL,
	    dev, B_WRITE, minphys, uio, DEV_BSIZE));
}
Commit	Line	Data
1c79356b	1	/*
5d5c5d0d A	2	* Copyright (c) 2000-2004 Apple Computer, Inc. All rights reserved.
5d5c5d0d A	3	*
2d21ac55	4	* @APPLE_OSREFERENCE_LICENSE_HEADER_START@
1c79356b	5	*
2d21ac55 A	6	* This file contains Original Code and/or Modifications of Original Code
	7	* as defined in and that are subject to the Apple Public Source License
	8	* Version 2.0 (the 'License'). You may not use this file except in
	9	* compliance with the License. The rights granted to you under the License
	10	* may not be used to create, or enable the creation or redistribution of,
	11	* unlawful or unlicensed copies of an Apple operating system, or to
	12	* circumvent, violate, or enable the circumvention or violation of, any
	13	* terms of an Apple operating system software license agreement.
8f6c56a5	14	*
2d21ac55 A	15	* Please obtain a copy of the License at
	16	* http://www.opensource.apple.com/apsl/ and read it before using this file.
	17	*
	18	* The Original Code and all software distributed under the License are
	19	* distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
8f6c56a5 A	20	* EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
8f6c56a5 A	21	* INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
2d21ac55 A	22	* FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
	23	* Please see the License for the specific language governing rights and
	24	* limitations under the License.
8f6c56a5	25	*
2d21ac55	26	* @APPLE_OSREFERENCE_LICENSE_HEADER_END@
1c79356b A	27	*/
	28	/*-
	29	* Copyright (c) 1982, 1986, 1990, 1993
	30	* The Regents of the University of California. All rights reserved.
	31	* (c) UNIX System Laboratories, Inc.
	32	* All or some portions of this file are derived from material licensed
	33	* to the University of California by American Telephone and Telegraph
	34	* Co. or Unix System Laboratories, Inc. and are reproduced herein with
	35	* the permission of UNIX System Laboratories, Inc.
	36	*
	37	* Redistribution and use in source and binary forms, with or without
	38	* modification, are permitted provided that the following conditions
	39	* are met:
	40	* 1. Redistributions of source code must retain the above copyright
	41	* notice, this list of conditions and the following disclaimer.
	42	* 2. Redistributions in binary form must reproduce the above copyright
	43	* notice, this list of conditions and the following disclaimer in the
	44	* documentation and/or other materials provided with the distribution.
	45	* 3. All advertising materials mentioning features or use of this software
	46	* must display the following acknowledgement:
	47	* This product includes software developed by the University of
	48	* California, Berkeley and its contributors.
	49	* 4. Neither the name of the University nor the names of its contributors
	50	* may be used to endorse or promote products derived from this software
	51	* without specific prior written permission.
	52	*
	53	* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
	54	* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
	55	* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
	56	* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
	57	* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
	58	* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
	59	* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
	60	* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
	61	* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
	62	* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
	63	* SUCH DAMAGE.
	64	*
	65	* from: @(#)kern_physio.c 8.1 (Berkeley) 6/10/93
	66	*/
	67	/*
	68	* HISTORY
	69	* 27-July-97 Umesh Vaishampayan (umeshv@apple.com)
	70	* Allow physio() to kernel space.
	71	*/
	72
	73	#include <sys/param.h>
	74	#include <sys/systm.h>
91447636	75	#include <sys/buf_internal.h>
1c79356b	76	#include <sys/conf.h>
91447636 A	77	#include <sys/proc_internal.h>
91447636 A	78	#include <sys/uio_internal.h>
2d21ac55 A	79	#include <kern/assert.h>
	80
	81	int rawread(dev_t dev, struct uio *uio);
	82	int rawwrite(dev_t dev, struct uio *uio);
1c79356b A	83
1c79356b A	84	int
2d21ac55 A	85	physio( void (*f_strategy)(buf_t),
	86	buf_t bp,
	87	dev_t dev,
	88	int flags,
	89	u_int (*f_minphys)(buf_t),
	90	struct uio *uio,
	91	int blocksize)
1c79356b	92	{
1c79356b	93	struct proc *p = current_proc();
2d21ac55 A	94	int error, i, buf_allocated, todo, iosize;
2d21ac55 A	95	int orig_bflags = 0;
91447636 A	96	#if LP64KERN
	97	int64_t done;
	98	#else
	99	int done;
	100	#endif
1c79356b A	101
	102	error = 0;
	103	flags &= B_READ \| B_WRITE;
2d21ac55	104	buf_allocated = 0;
1c79356b A	105
	106	/*
	107	* [check user read/write access to the data buffer]
	108	*
	109	* Check each iov one by one. Note that we know if we're reading or
	110	* writing, so we ignore the uio's rw parameter. Also note that if
	111	* we're doing a read, that's a write to user-space.
	112	*/
	113	for (i = 0; i < uio->uio_iovcnt; i++) {
2d21ac55	114	if (UIO_SEG_IS_USER_SPACE(uio->uio_segflg)) {
91447636 A	115	if (!useracc(uio_iov_base_at(uio, i),
91447636 A	116	uio_iov_len_at(uio, i),
1c79356b A	117	(flags == B_READ) ? B_WRITE : B_READ))
	118	return (EFAULT);
	119	}
	120	}
2d21ac55 A	121	/*
	122	* Make sure we have a buffer, creating one if necessary.
	123	*/
	124	if (bp == NULL) {
	125	bp = buf_alloc((vnode_t)0);
	126	buf_allocated = 1;
	127	} else
	128	orig_bflags = buf_flags(bp);
	129	/*
	130	* at this point we should have a buffer
	131	* that is marked BL_BUSY... we either
	132	* acquired it via buf_alloc, or it was
	133	* passed into us... if it was passed
	134	* in, it needs to already be owned by
	135	* the caller (i.e. BL_BUSY is set)
	136	*/
	137	assert(bp->b_lflags & BL_BUSY);
1c79356b	138
2d21ac55 A	139	/*
	140	* [set up the fixed part of the buffer for a transfer]
	141	*/
1c79356b	142	bp->b_dev = dev;
1c79356b A	143	bp->b_proc = p;
1c79356b A	144
2d21ac55 A	145	/*
	146	* [mark the buffer busy for physical I/O]
	147	* (i.e. set B_PHYS (because it's an I/O to user
	148	* memory, and B_RAW, because B_RAW is to be
	149	* "Set by physio for raw transfers.", in addition
	150	* to the read/write flag.)
	151	*/
	152	buf_setflags(bp, B_PHYS \| B_RAW);
	153
1c79356b	154	/*
91447636	155	* [while there is data to transfer and no I/O error]
1c79356b A	156	* Note that I/O errors are handled with a 'goto' at the bottom
	157	* of the 'while' loop.
	158	*/
	159	for (i = 0; i < uio->uio_iovcnt; i++) {
91447636	160	while (uio_iov_len_at(uio, i) > 0) {
91447636 A	161
	162	if ( (iosize = uio_iov_len_at(uio, i)) > MAXPHYSIO_WIRED)
	163	iosize = MAXPHYSIO_WIRED;
2d21ac55 A	164	/*
	165	* make sure we're set to issue a fresh I/O
	166	* in the right direction
	167	*/
	168	buf_reset(bp, flags);
1c79356b A	169
1c79356b A	170	/* [set up the buffer for a maximum-sized transfer] */
91447636 A	171	buf_setblkno(bp, uio->uio_offset / blocksize);
	172	buf_setcount(bp, iosize);
	173	// LP64todo - fix this!
2d21ac55	174	buf_setdataptr(bp, (uintptr_t)CAST_DOWN(caddr_t, uio_iov_base_at(uio, i)));
1c79356b A	175
1c79356b A	176	/*
2d21ac55	177	* [call f_minphys to bound the tranfer size]
1c79356b A	178	* and remember the amount of data to transfer,
	179	* for later comparison.
	180	*/
2d21ac55	181	(*f_minphys)(bp);
91447636	182	todo = buf_count(bp);
1c79356b A	183
	184	/*
	185	* [lock the part of the user address space involved
	186	* in the transfer]
1c79356b A	187	*/
1c79356b A	188
2d21ac55 A	189	if(UIO_SEG_IS_USER_SPACE(uio->uio_segflg)) {
	190	error = vslock(CAST_USER_ADDR_T(buf_dataptr(bp)),
	191	(user_size_t)todo);
	192	if (error)
	193	goto done;
	194	}
1c79356b	195
2d21ac55 A	196	/* [call f_strategy to start the transfer] */
2d21ac55 A	197	(*f_strategy)(bp);
1c79356b	198
1c79356b A	199
1c79356b A	200	/* [wait for the transfer to complete] */
91447636	201	error = (int)buf_biowait(bp);
1c79356b A	202
	203	/*
	204	* [unlock the part of the address space previously
	205	* locked]
	206	*/
91447636 A	207	if(UIO_SEG_IS_USER_SPACE(uio->uio_segflg))
	208	vsunlock(CAST_USER_ADDR_T(buf_dataptr(bp)),
	209	(user_size_t)todo,
	210	(flags & B_READ));
1c79356b A	211
	212	/*
	213	* [deduct the transfer size from the total number
	214	* of data to transfer]
	215	*/
91447636 A	216	done = buf_count(bp) - buf_resid(bp);
	217	uio_iov_len_add_at(uio, -done, i);
	218	uio_iov_base_add_at(uio, done, i);
	219	uio->uio_offset += done;
	220	uio_setresid(uio, (uio_resid(uio) - done));
1c79356b A	221
	222	/*
	223	* Now, check for an error.
	224	* Also, handle weird end-of-disk semantics.
	225	*/
	226	if (error \|\| done < todo)
	227	goto done;
	228	}
	229	}
	230
	231	done:
2d21ac55 A	232	if (buf_allocated)
	233	buf_free(bp);
	234	else
	235	buf_setflags(bp, orig_bflags);
1c79356b A	236
	237	return (error);
	238	}
	239
	240	/*
	241	* Leffler, et al., says on p. 231:
	242	* "The minphys() routine is called by physio() to adjust the
	243	* size of each I/O transfer before the latter is passed to
	244	* the strategy routine..."
	245	*
	246	* so, just adjust the buffer's count accounting to MAXPHYS here,
	247	* and return the new count;
	248	*/
	249	u_int
2d21ac55	250	minphys(struct buf *bp)
1c79356b A	251	{
1c79356b A	252
91447636 A	253	buf_setcount(bp, min(MAXPHYS, buf_count(bp)));
91447636 A	254	return buf_count(bp);
1c79356b A	255	}
	256
	257	/*
	258	* Do a read on a device for a user process.
	259	*/
2d21ac55 A	260	int
2d21ac55 A	261	rawread(dev_t dev, struct uio *uio)
1c79356b A	262	{
	263	return (physio(cdevsw[major(dev)].d_strategy, (struct buf *)NULL,
	264	dev, B_READ, minphys, uio, DEV_BSIZE));
	265	}
	266
	267	/*
	268	* Do a write on a device for a user process.
	269	*/
2d21ac55 A	270	int
2d21ac55 A	271	rawwrite(dev_t dev, struct uio *uio)
1c79356b A	272	{
	273	return (physio(cdevsw[major(dev)].d_strategy, (struct buf *)NULL,
	274	dev, B_WRITE, minphys, uio, DEV_BSIZE));
	275	}