NetBSD/sys/kern/subr_disk.c

431 lines
12 KiB
C
Raw Normal View History

/* $NetBSD: subr_disk.c,v 1.22 1997/10/05 18:39:51 thorpej Exp $ */
/*-
* Copyright (c) 1996, 1997 The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Jason R. Thorpe of the Numerical Aerospace Simulation Facility,
* NASA Ames Research Center.
*
* 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 NetBSD
* Foundation, Inc. and its contributors.
* 4. Neither the name of The NetBSD Foundation 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 NETBSD FOUNDATION, INC. 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 FOUNDATION 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.
*/
1994-05-19 07:43:13 +04:00
/*
* Copyright (c) 1982, 1986, 1988, 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.
*
* @(#)ufs_disksubr.c 8.5 (Berkeley) 1/21/94
1994-05-19 07:43:13 +04:00
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
1994-05-19 07:43:13 +04:00
#include <sys/buf.h>
#include <sys/syslog.h>
#include <sys/time.h>
#include <sys/disklabel.h>
#include <sys/disk.h>
/*
* A global list of all disks attached to the system. May grow or
* shrink over time.
*/
struct disklist_head disklist; /* TAILQ_HEAD */
int disk_count; /* number of drives in global disklist */
1994-05-19 07:43:13 +04:00
/*
* Seek sort for disks. We depend on the driver which calls us using b_resid
* as the current cylinder number.
*
* The argument ap structure holds a b_actf activity chain pointer on which we
* keep two queues, sorted in ascending cylinder order. The first queue holds
* those requests which are positioned after the current cylinder (in the first
* request); the second holds requests which came in after their cylinder number
* was passed. Thus we implement a one way scan, retracting after reaching the
* end of the drive to the first request on the second queue, at which time it
* becomes the first queue.
*
* A one-way scan is natural because of the way UNIX read-ahead blocks are
* allocated.
*/
void
disksort(ap, bp)
register struct buf *ap, *bp;
{
register struct buf *bq;
/* If the queue is empty, then it's easy. */
if (ap->b_actf == NULL) {
bp->b_actf = NULL;
ap->b_actf = bp;
return;
}
/*
* If we lie after the first (currently active) request, then we
* must locate the second request list and add ourselves to it.
*/
bq = ap->b_actf;
if (bp->b_cylinder < bq->b_cylinder) {
while (bq->b_actf) {
/*
* Check for an ``inversion'' in the normally ascending
* cylinder numbers, indicating the start of the second
* request list.
*/
if (bq->b_actf->b_cylinder < bq->b_cylinder) {
/*
* Search the second request list for the first
* request at a larger cylinder number. We go
* before that; if there is no such request, we
* go at end.
*/
do {
if (bp->b_cylinder <
bq->b_actf->b_cylinder)
goto insert;
if (bp->b_cylinder ==
bq->b_actf->b_cylinder &&
bp->b_blkno < bq->b_actf->b_blkno)
goto insert;
bq = bq->b_actf;
} while (bq->b_actf);
goto insert; /* after last */
}
bq = bq->b_actf;
}
/*
* No inversions... we will go after the last, and
* be the first request in the second request list.
*/
goto insert;
}
/*
* Request is at/after the current request...
* sort in the first request list.
*/
while (bq->b_actf) {
/*
* We want to go after the current request if there is an
* inversion after it (i.e. it is the end of the first
* request list), or if the next request is a larger cylinder
* than our request.
*/
if (bq->b_actf->b_cylinder < bq->b_cylinder ||
bp->b_cylinder < bq->b_actf->b_cylinder ||
(bp->b_cylinder == bq->b_actf->b_cylinder &&
bp->b_blkno < bq->b_actf->b_blkno))
goto insert;
bq = bq->b_actf;
}
/*
* Neither a second list nor a larger request... we go at the end of
* the first list, which is the same as the end of the whole schebang.
*/
insert: bp->b_actf = bq->b_actf;
bq->b_actf = bp;
}
/* encoding of disk minor numbers, should be elsewhere... */
#define dkunit(dev) (minor(dev) >> 3)
#define dkpart(dev) (minor(dev) & 07)
#define dkminor(unit, part) (((unit) << 3) | (part))
/*
* Compute checksum for disk label.
*/
u_int
dkcksum(lp)
register struct disklabel *lp;
{
register u_short *start, *end;
register u_short sum = 0;
start = (u_short *)lp;
end = (u_short *)&lp->d_partitions[lp->d_npartitions];
while (start < end)
sum ^= *start++;
return (sum);
}
/*
* Disk error is the preface to plaintive error messages
* about failing disk transfers. It prints messages of the form
hp0g: hard error reading fsbn 12345 of 12344-12347 (hp0 bn %d cn %d tn %d sn %d)
* if the offset of the error in the transfer and a disk label
* are both available. blkdone should be -1 if the position of the error
* is unknown; the disklabel pointer may be null from drivers that have not
1996-10-13 06:32:29 +04:00
* been converted to use them. The message is printed with printf
1994-05-19 07:43:13 +04:00
* if pri is LOG_PRINTF, otherwise it uses log at the specified priority.
1996-10-13 06:32:29 +04:00
* The message should be completed (with at least a newline) with printf
1994-05-19 07:43:13 +04:00
* or addlog, respectively. There is no trailing space.
*/
void
diskerr(bp, dname, what, pri, blkdone, lp)
register struct buf *bp;
char *dname, *what;
int pri, blkdone;
register struct disklabel *lp;
{
int unit = dkunit(bp->b_dev), part = dkpart(bp->b_dev);
register void (*pr) __P((const char *, ...));
char partname = 'a' + part;
int sn;
if (pri != LOG_PRINTF) {
1996-03-17 02:17:04 +03:00
static const char fmt[] = "";
log(pri, fmt);
1994-05-19 07:43:13 +04:00
pr = addlog;
} else
1996-10-13 06:32:29 +04:00
pr = printf;
1994-05-19 07:43:13 +04:00
(*pr)("%s%d%c: %s %sing fsbn ", dname, unit, partname, what,
bp->b_flags & B_READ ? "read" : "writ");
sn = bp->b_blkno;
if (bp->b_bcount <= DEV_BSIZE)
(*pr)("%d", sn);
else {
if (blkdone >= 0) {
sn += blkdone;
(*pr)("%d of ", sn);
}
(*pr)("%d-%d", bp->b_blkno,
bp->b_blkno + (bp->b_bcount - 1) / DEV_BSIZE);
}
if (lp && (blkdone >= 0 || bp->b_bcount <= lp->d_secsize)) {
sn += lp->d_partitions[part].p_offset;
(*pr)(" (%s%d bn %d; cn %d", dname, unit, sn,
sn / lp->d_secpercyl);
sn %= lp->d_secpercyl;
(*pr)(" tn %d sn %d)", sn / lp->d_nsectors, sn % lp->d_nsectors);
}
}
/*
* Initialize the disklist. Called by main() before autoconfiguration.
*/
void
disk_init()
{
TAILQ_INIT(&disklist);
disk_count = 0;
}
/*
* Searches the disklist for the disk corresponding to the
* name provided.
*/
struct disk *
disk_find(name)
char *name;
{
struct disk *diskp;
if ((name == NULL) || (disk_count <= 0))
return (NULL);
for (diskp = disklist.tqh_first; diskp != NULL;
diskp = diskp->dk_link.tqe_next)
if (strcmp(diskp->dk_name, name) == 0)
return (diskp);
return (NULL);
}
/*
* Attach a disk.
*/
void
disk_attach(diskp)
struct disk *diskp;
{
int s;
/*
* Allocate and initialize the disklabel structures. Note that
* it's not safe to sleep here, since we're probably going to be
* called during autoconfiguration.
*/
diskp->dk_label = malloc(sizeof(struct disklabel), M_DEVBUF, M_NOWAIT);
diskp->dk_cpulabel = malloc(sizeof(struct cpu_disklabel), M_DEVBUF,
M_NOWAIT);
if ((diskp->dk_label == NULL) || (diskp->dk_cpulabel == NULL))
panic("disk_attach: can't allocate storage for disklabel");
bzero(diskp->dk_label, sizeof(struct disklabel));
bzero(diskp->dk_cpulabel, sizeof(struct cpu_disklabel));
/*
* Set the attached timestamp.
*/
s = splclock();
diskp->dk_attachtime = mono_time;
splx(s);
/*
* Link into the disklist.
*/
TAILQ_INSERT_TAIL(&disklist, diskp, dk_link);
++disk_count;
}
/*
1996-02-09 21:59:18 +03:00
* Detach a disk.
*/
void
1996-02-09 21:59:18 +03:00
disk_detach(diskp)
struct disk *diskp;
{
/*
* Free the space used by the disklabel structures.
*/
free(diskp->dk_label, M_DEVBUF);
free(diskp->dk_cpulabel, M_DEVBUF);
/*
* Remove from the disklist.
*/
TAILQ_REMOVE(&disklist, diskp, dk_link);
if (--disk_count < 0)
1996-02-09 21:59:18 +03:00
panic("disk_detach: disk_count < 0");
}
/*
* Increment a disk's busy counter. If the counter is going from
* 0 to 1, set the timestamp.
*/
void
disk_busy(diskp)
struct disk *diskp;
{
int s;
/*
* XXX We'd like to use something as accurate as microtime(),
* but that doesn't depend on the system TOD clock.
*/
if (diskp->dk_busy++ == 0) {
s = splclock();
diskp->dk_timestamp = mono_time;
splx(s);
}
}
/*
* Decrement a disk's busy counter, increment the byte count, total busy
* time, and reset the timestamp.
*/
void
disk_unbusy(diskp, bcount)
struct disk *diskp;
long bcount;
{
int s;
struct timeval dv_time, diff_time;
if (diskp->dk_busy-- == 0)
panic("disk_unbusy: %s: dk_busy < 0", diskp->dk_name);
s = splclock();
dv_time = mono_time;
splx(s);
timersub(&dv_time, &diskp->dk_timestamp, &diff_time);
timeradd(&diskp->dk_time, &diff_time, &diskp->dk_time);
diskp->dk_timestamp = dv_time;
if (bcount > 0) {
diskp->dk_bytes += bcount;
diskp->dk_xfer++;
}
}
/*
* Reset the metrics counters on the given disk. Note that we cannot
* reset the busy counter, as it may case a panic in disk_unbusy().
* We also must avoid playing with the timestamp information, as it
* may skew any pending transfer results.
*/
void
disk_resetstat(diskp)
struct disk *diskp;
{
int s = splbio(), t;
diskp->dk_xfer = 0;
diskp->dk_bytes = 0;
t = splclock();
diskp->dk_attachtime = mono_time;
splx(t);
timerclear(&diskp->dk_time);
splx(s);
}