Commit | Line | Data |
---|---|---|
da7c5cc6 | 1 | /* |
1ad7f5c8 | 2 | * Copyright (c) 1982, 1986, 1988 Regents of the University of California. |
7188ac27 | 3 | * All rights reserved. |
da7c5cc6 | 4 | * |
af359dea C |
5 | * Redistribution and use in source and binary forms, with or without |
6 | * modification, are permitted provided that the following conditions | |
7 | * are met: | |
8 | * 1. Redistributions of source code must retain the above copyright | |
9 | * notice, this list of conditions and the following disclaimer. | |
10 | * 2. Redistributions in binary form must reproduce the above copyright | |
11 | * notice, this list of conditions and the following disclaimer in the | |
12 | * documentation and/or other materials provided with the distribution. | |
13 | * 3. All advertising materials mentioning features or use of this software | |
14 | * must display the following acknowledgement: | |
15 | * This product includes software developed by the University of | |
16 | * California, Berkeley and its contributors. | |
17 | * 4. Neither the name of the University nor the names of its contributors | |
18 | * may be used to endorse or promote products derived from this software | |
19 | * without specific prior written permission. | |
7188ac27 | 20 | * |
af359dea C |
21 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND |
22 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
23 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE | |
24 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE | |
25 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
26 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS | |
27 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) | |
28 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT | |
29 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY | |
30 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF | |
31 | * SUCH DAMAGE. | |
32 | * | |
33 | * @(#)ufs_disksubr.c 7.16 (Berkeley) 5/4/91 | |
da7c5cc6 | 34 | */ |
200a8f27 | 35 | |
97934dfa MK |
36 | #include "param.h" |
37 | #include "systm.h" | |
38 | #include "buf.h" | |
39 | #include "disklabel.h" | |
1ad7f5c8 | 40 | #include "syslog.h" |
9573c71c | 41 | |
200a8f27 | 42 | /* |
0d67fde4 BJ |
43 | * Seek sort for disks. We depend on the driver |
44 | * which calls us using b_resid as the current cylinder number. | |
45 | * | |
46 | * The argument dp structure holds a b_actf activity chain pointer | |
47 | * on which we keep two queues, sorted in ascending cylinder order. | |
48 | * The first queue holds those requests which are positioned after | |
49 | * the current cylinder (in the first request); the second holds | |
50 | * requests which came in after their cylinder number was passed. | |
51 | * Thus we implement a one way scan, retracting after reaching the | |
52 | * end of the drive to the first request on the second queue, | |
53 | * at which time it becomes the first queue. | |
54 | * | |
55 | * A one-way scan is natural because of the way UNIX read-ahead | |
56 | * blocks are allocated. | |
200a8f27 BJ |
57 | */ |
58 | ||
200a8f27 BJ |
59 | #define b_cylin b_resid |
60 | ||
61 | disksort(dp, bp) | |
0d67fde4 | 62 | register struct buf *dp, *bp; |
200a8f27 BJ |
63 | { |
64 | register struct buf *ap; | |
200a8f27 | 65 | |
0d67fde4 BJ |
66 | /* |
67 | * If nothing on the activity queue, then | |
68 | * we become the only thing. | |
69 | */ | |
200a8f27 BJ |
70 | ap = dp->b_actf; |
71 | if(ap == NULL) { | |
72 | dp->b_actf = bp; | |
73 | dp->b_actl = bp; | |
74 | bp->av_forw = NULL; | |
75 | return; | |
76 | } | |
0d67fde4 BJ |
77 | /* |
78 | * If we lie after the first (currently active) | |
79 | * request, then we must locate the second request list | |
80 | * and add ourselves to it. | |
81 | */ | |
82 | if (bp->b_cylin < ap->b_cylin) { | |
83 | while (ap->av_forw) { | |
84 | /* | |
85 | * Check for an ``inversion'' in the | |
86 | * normally ascending cylinder numbers, | |
87 | * indicating the start of the second request list. | |
88 | */ | |
89 | if (ap->av_forw->b_cylin < ap->b_cylin) { | |
90 | /* | |
91 | * Search the second request list | |
92 | * for the first request at a larger | |
93 | * cylinder number. We go before that; | |
94 | * if there is no such request, we go at end. | |
95 | */ | |
96 | do { | |
97 | if (bp->b_cylin < ap->av_forw->b_cylin) | |
98 | goto insert; | |
667e5665 MK |
99 | if (bp->b_cylin == ap->av_forw->b_cylin && |
100 | bp->b_blkno < ap->av_forw->b_blkno) | |
101 | goto insert; | |
0d67fde4 BJ |
102 | ap = ap->av_forw; |
103 | } while (ap->av_forw); | |
104 | goto insert; /* after last */ | |
105 | } | |
106 | ap = ap->av_forw; | |
200a8f27 | 107 | } |
0d67fde4 BJ |
108 | /* |
109 | * No inversions... we will go after the last, and | |
110 | * be the first request in the second request list. | |
111 | */ | |
112 | goto insert; | |
200a8f27 | 113 | } |
0d67fde4 BJ |
114 | /* |
115 | * Request is at/after the current request... | |
116 | * sort in the first request list. | |
117 | */ | |
118 | while (ap->av_forw) { | |
119 | /* | |
120 | * We want to go after the current request | |
121 | * if there is an inversion after it (i.e. it is | |
122 | * the end of the first request list), or if | |
123 | * the next request is a larger cylinder than our request. | |
124 | */ | |
125 | if (ap->av_forw->b_cylin < ap->b_cylin || | |
667e5665 MK |
126 | bp->b_cylin < ap->av_forw->b_cylin || |
127 | (bp->b_cylin == ap->av_forw->b_cylin && | |
128 | bp->b_blkno < ap->av_forw->b_blkno)) | |
0d67fde4 BJ |
129 | goto insert; |
130 | ap = ap->av_forw; | |
131 | } | |
132 | /* | |
133 | * Neither a second list nor a larger | |
134 | * request... we go at the end of the first list, | |
135 | * which is the same as the end of the whole schebang. | |
136 | */ | |
137 | insert: | |
138 | bp->av_forw = ap->av_forw; | |
139 | ap->av_forw = bp; | |
140 | if (ap == dp->b_actl) | |
200a8f27 BJ |
141 | dp->b_actl = bp; |
142 | } | |
97934dfa | 143 | |
9573c71c MK |
144 | /* |
145 | * Attempt to read a disk label from a device | |
146 | * using the indicated stategy routine. | |
147 | * The label must be partly set up before this: | |
148 | * secpercyl and anything required in the strategy routine | |
149 | * (e.g., sector size) must be filled in before calling us. | |
150 | * Returns null on success and an error string on failure. | |
151 | */ | |
152 | char * | |
153 | readdisklabel(dev, strat, lp) | |
154 | dev_t dev; | |
155 | int (*strat)(); | |
156 | register struct disklabel *lp; | |
157 | { | |
158 | register struct buf *bp; | |
159 | struct disklabel *dlp; | |
160 | char *msg = NULL; | |
161 | ||
162 | if (lp->d_secperunit == 0) | |
163 | lp->d_secperunit = 0x1fffffff; | |
164 | lp->d_npartitions = 1; | |
165 | if (lp->d_partitions[0].p_size == 0) | |
166 | lp->d_partitions[0].p_size = 0x1fffffff; | |
167 | lp->d_partitions[0].p_offset = 0; | |
168 | ||
d3633d66 | 169 | bp = geteblk((int)lp->d_secsize); |
9573c71c MK |
170 | bp->b_dev = dev; |
171 | bp->b_blkno = LABELSECTOR; | |
cd6cd535 | 172 | bp->b_bcount = lp->d_secsize; |
9573c71c MK |
173 | bp->b_flags = B_BUSY | B_READ; |
174 | bp->b_cylin = LABELSECTOR / lp->d_secpercyl; | |
175 | (*strat)(bp); | |
7188ac27 | 176 | if (biowait(bp)) { |
9573c71c | 177 | msg = "I/O error"; |
cd6cd535 MK |
178 | } else for (dlp = (struct disklabel *)bp->b_un.b_addr; |
179 | dlp <= (struct disklabel *)(bp->b_un.b_addr+DEV_BSIZE-sizeof(*dlp)); | |
180 | dlp = (struct disklabel *)((char *)dlp + sizeof(long))) { | |
181 | if (dlp->d_magic != DISKMAGIC || dlp->d_magic2 != DISKMAGIC) { | |
182 | if (msg == NULL) | |
183 | msg = "no disk label"; | |
0328954d KM |
184 | } else if (dlp->d_npartitions > MAXPARTITIONS || |
185 | dkcksum(dlp) != 0) | |
9573c71c | 186 | msg = "disk label corrupted"; |
cd6cd535 | 187 | else { |
9573c71c | 188 | *lp = *dlp; |
cd6cd535 MK |
189 | msg = NULL; |
190 | break; | |
191 | } | |
9573c71c MK |
192 | } |
193 | bp->b_flags = B_INVAL | B_AGE; | |
194 | brelse(bp); | |
195 | return (msg); | |
196 | } | |
197 | ||
667e5665 MK |
198 | /* |
199 | * Check new disk label for sensibility | |
200 | * before setting it. | |
201 | */ | |
202 | setdisklabel(olp, nlp, openmask) | |
203 | register struct disklabel *olp, *nlp; | |
204 | u_long openmask; | |
205 | { | |
206 | register i; | |
207 | register struct partition *opp, *npp; | |
208 | ||
209 | if (nlp->d_magic != DISKMAGIC || nlp->d_magic2 != DISKMAGIC || | |
210 | dkcksum(nlp) != 0) | |
211 | return (EINVAL); | |
5adcb337 | 212 | while ((i = ffs((long)openmask)) != 0) { |
667e5665 MK |
213 | i--; |
214 | openmask &= ~(1 << i); | |
215 | if (nlp->d_npartitions <= i) | |
216 | return (EBUSY); | |
217 | opp = &olp->d_partitions[i]; | |
218 | npp = &nlp->d_partitions[i]; | |
219 | if (npp->p_offset != opp->p_offset || npp->p_size < opp->p_size) | |
220 | return (EBUSY); | |
221 | /* | |
222 | * Copy internally-set partition information | |
223 | * if new label doesn't include it. XXX | |
224 | */ | |
225 | if (npp->p_fstype == FS_UNUSED && opp->p_fstype != FS_UNUSED) { | |
226 | npp->p_fstype = opp->p_fstype; | |
227 | npp->p_fsize = opp->p_fsize; | |
228 | npp->p_frag = opp->p_frag; | |
229 | npp->p_cpg = opp->p_cpg; | |
230 | } | |
231 | } | |
d3633d66 MK |
232 | nlp->d_checksum = 0; |
233 | nlp->d_checksum = dkcksum(nlp); | |
667e5665 MK |
234 | *olp = *nlp; |
235 | return (0); | |
236 | } | |
237 | ||
238 | /* encoding of disk minor numbers, should be elsewhere... */ | |
239 | #define dkunit(dev) (minor(dev) >> 3) | |
240 | #define dkpart(dev) (minor(dev) & 07) | |
241 | #define dkminor(unit, part) (((unit) << 3) | (part)) | |
242 | ||
243 | /* | |
244 | * Write disk label back to device after modification. | |
245 | */ | |
246 | writedisklabel(dev, strat, lp) | |
247 | dev_t dev; | |
248 | int (*strat)(); | |
249 | register struct disklabel *lp; | |
250 | { | |
251 | struct buf *bp; | |
252 | struct disklabel *dlp; | |
253 | int labelpart; | |
254 | int error = 0; | |
255 | ||
256 | labelpart = dkpart(dev); | |
257 | if (lp->d_partitions[labelpart].p_offset != 0) { | |
258 | if (lp->d_partitions[0].p_offset != 0) | |
259 | return (EXDEV); /* not quite right */ | |
260 | labelpart = 0; | |
261 | } | |
d3633d66 | 262 | bp = geteblk((int)lp->d_secsize); |
667e5665 MK |
263 | bp->b_dev = makedev(major(dev), dkminor(dkunit(dev), labelpart)); |
264 | bp->b_blkno = LABELSECTOR; | |
265 | bp->b_bcount = lp->d_secsize; | |
266 | bp->b_flags = B_READ; | |
267 | (*strat)(bp); | |
7188ac27 | 268 | if (error = biowait(bp)) |
d3633d66 | 269 | goto done; |
d3633d66 MK |
270 | for (dlp = (struct disklabel *)bp->b_un.b_addr; |
271 | dlp <= (struct disklabel *) | |
272 | (bp->b_un.b_addr + lp->d_secsize - sizeof(*dlp)); | |
273 | dlp = (struct disklabel *)((char *)dlp + sizeof(long))) { | |
274 | if (dlp->d_magic == DISKMAGIC && dlp->d_magic2 == DISKMAGIC && | |
275 | dkcksum(dlp) == 0) { | |
276 | *dlp = *lp; | |
277 | bp->b_flags = B_WRITE; | |
278 | (*strat)(bp); | |
7188ac27 | 279 | error = biowait(bp); |
d3633d66 MK |
280 | goto done; |
281 | } | |
667e5665 | 282 | } |
d3633d66 MK |
283 | error = ESRCH; |
284 | done: | |
667e5665 MK |
285 | brelse(bp); |
286 | return (error); | |
287 | } | |
288 | ||
97934dfa MK |
289 | /* |
290 | * Compute checksum for disk label. | |
291 | */ | |
292 | dkcksum(lp) | |
293 | register struct disklabel *lp; | |
294 | { | |
295 | register u_short *start, *end; | |
296 | register u_short sum = 0; | |
297 | ||
298 | start = (u_short *)lp; | |
299 | end = (u_short *)&lp->d_partitions[lp->d_npartitions]; | |
300 | while (start < end) | |
301 | sum ^= *start++; | |
302 | return (sum); | |
303 | } | |
1ad7f5c8 MK |
304 | |
305 | /* | |
306 | * Disk error is the preface to plaintive error messages | |
307 | * about failing disk transfers. It prints messages of the form | |
05de9101 MK |
308 | |
309 | hp0g: hard error reading fsbn 12345 of 12344-12347 (hp0 bn %d cn %d tn %d sn %d) | |
310 | ||
1ad7f5c8 MK |
311 | * if the offset of the error in the transfer and a disk label |
312 | * are both available. blkdone should be -1 if the position of the error | |
313 | * is unknown; the disklabel pointer may be null from drivers that have not | |
314 | * been converted to use them. The message is printed with printf | |
315 | * if pri is LOG_PRINTF, otherwise it uses log at the specified priority. | |
316 | * The message should be completed (with at least a newline) with printf | |
317 | * or addlog, respectively. There is no trailing space. | |
318 | */ | |
319 | diskerr(bp, dname, what, pri, blkdone, lp) | |
320 | register struct buf *bp; | |
321 | char *dname, *what; | |
322 | int pri, blkdone; | |
323 | register struct disklabel *lp; | |
324 | { | |
325 | int unit = dkunit(bp->b_dev), part = dkpart(bp->b_dev); | |
af359dea | 326 | register void (*pr) __P((const char *, ...)); |
1ad7f5c8 | 327 | char partname = 'a' + part; |
af359dea | 328 | int sn; |
1ad7f5c8 MK |
329 | |
330 | if (pri != LOG_PRINTF) { | |
331 | log(pri, ""); | |
332 | pr = addlog; | |
333 | } else | |
334 | pr = printf; | |
335 | (*pr)("%s%d%c: %s %sing fsbn ", dname, unit, partname, what, | |
336 | bp->b_flags & B_READ ? "read" : "writ"); | |
337 | sn = bp->b_blkno; | |
338 | if (bp->b_bcount <= DEV_BSIZE) | |
339 | (*pr)("%d", sn); | |
340 | else { | |
341 | if (blkdone >= 0) { | |
342 | sn += blkdone; | |
343 | (*pr)("%d of ", sn); | |
344 | } | |
345 | (*pr)("%d-%d", bp->b_blkno, | |
346 | bp->b_blkno + (bp->b_bcount - 1) / DEV_BSIZE); | |
347 | } | |
f2a9617b MK |
348 | if (lp && (blkdone >= 0 || bp->b_bcount <= lp->d_secsize)) { |
349 | #ifdef tahoe | |
350 | sn *= DEV_BSIZE / lp->d_secsize; /* XXX */ | |
351 | #endif | |
1ad7f5c8 | 352 | sn += lp->d_partitions[part].p_offset; |
05de9101 MK |
353 | (*pr)(" (%s%d bn %d; cn %d", dname, unit, sn, |
354 | sn / lp->d_secpercyl); | |
355 | sn %= lp->d_secpercyl; | |
9f79ca31 | 356 | (*pr)(" tn %d sn %d)", sn / lp->d_nsectors, sn % lp->d_nsectors); |
1ad7f5c8 MK |
357 | } |
358 | } |