4.3BSD - /usr/src/sys/sys/ufs_bio.c

   1: /*
   2:  * Copyright (c) 1982, 1986 Regents of the University of California.
   3:  * All rights reserved.  The Berkeley software License Agreement
   4:  * specifies the terms and conditions for redistribution.
   5:  *
   6:  *	@(#)ufs_bio.c	7.1 (Berkeley) 6/5/86
   7:  */
   8: 
   9: #include "../machine/pte.h"
  10: 
  11: #include "param.h"
  12: #include "systm.h"
  13: #include "dir.h"
  14: #include "user.h"
  15: #include "buf.h"
  16: #include "conf.h"
  17: #include "proc.h"
  18: #include "seg.h"
  19: #include "vm.h"
  20: #include "trace.h"
  21: 
  22: /*
  23:  * Read in (if necessary) the block and return a buffer pointer.
  24:  */
  25: struct buf *
  26: bread(dev, blkno, size)
  27:     dev_t dev;
  28:     daddr_t blkno;
  29:     int size;
  30: {
  31:     register struct buf *bp;
  32: 
  33:     if (size == 0)
  34:         panic("bread: size 0");
  35:     bp = getblk(dev, blkno, size);
  36:     if (bp->b_flags&B_DONE) {
  37:         trace(TR_BREADHIT, pack(dev, size), blkno);
  38:         return (bp);
  39:     }
  40:     bp->b_flags |= B_READ;
  41:     if (bp->b_bcount > bp->b_bufsize)
  42:         panic("bread");
  43:     (*bdevsw[major(dev)].d_strategy)(bp);
  44:     trace(TR_BREADMISS, pack(dev, size), blkno);
  45:     u.u_ru.ru_inblock++;        /* pay for read */
  46:     biowait(bp);
  47:     return (bp);
  48: }
  49: 
  50: /*
  51:  * Read in the block, like bread, but also start I/O on the
  52:  * read-ahead block (which is not allocated to the caller)
  53:  */
  54: struct buf *
  55: breada(dev, blkno, size, rablkno, rabsize)
  56:     dev_t dev;
  57:     daddr_t blkno; int size;
  58:     daddr_t rablkno; int rabsize;
  59: {
  60:     register struct buf *bp, *rabp;
  61: 
  62:     bp = NULL;
  63:     /*
  64: 	 * If the block isn't in core, then allocate
  65: 	 * a buffer and initiate i/o (getblk checks
  66: 	 * for a cache hit).
  67: 	 */
  68:     if (!incore(dev, blkno)) {
  69:         bp = getblk(dev, blkno, size);
  70:         if ((bp->b_flags&B_DONE) == 0) {
  71:             bp->b_flags |= B_READ;
  72:             if (bp->b_bcount > bp->b_bufsize)
  73:                 panic("breada");
  74:             (*bdevsw[major(dev)].d_strategy)(bp);
  75:             trace(TR_BREADMISS, pack(dev, size), blkno);
  76:             u.u_ru.ru_inblock++;        /* pay for read */
  77:         } else
  78:             trace(TR_BREADHIT, pack(dev, size), blkno);
  79:     }
  80: 
  81:     /*
  82: 	 * If there's a read-ahead block, start i/o
  83: 	 * on it also (as above).
  84: 	 */
  85:     if (rablkno && !incore(dev, rablkno)) {
  86:         rabp = getblk(dev, rablkno, rabsize);
  87:         if (rabp->b_flags & B_DONE) {
  88:             brelse(rabp);
  89:             trace(TR_BREADHITRA, pack(dev, rabsize), blkno);
  90:         } else {
  91:             rabp->b_flags |= B_READ|B_ASYNC;
  92:             if (rabp->b_bcount > rabp->b_bufsize)
  93:                 panic("breadrabp");
  94:             (*bdevsw[major(dev)].d_strategy)(rabp);
  95:             trace(TR_BREADMISSRA, pack(dev, rabsize), rablock);
  96:             u.u_ru.ru_inblock++;        /* pay in advance */
  97:         }
  98:     }
  99: 
 100:     /*
 101: 	 * If block was in core, let bread get it.
 102: 	 * If block wasn't in core, then the read was started
 103: 	 * above, and just wait for it.
 104: 	 */
 105:     if (bp == NULL)
 106:         return (bread(dev, blkno, size));
 107:     biowait(bp);
 108:     return (bp);
 109: }
 110: 
 111: /*
 112:  * Write the buffer, waiting for completion.
 113:  * Then release the buffer.
 114:  */
 115: bwrite(bp)
 116:     register struct buf *bp;
 117: {
 118:     register flag;
 119: 
 120:     flag = bp->b_flags;
 121:     bp->b_flags &= ~(B_READ | B_DONE | B_ERROR | B_DELWRI);
 122:     if ((flag&B_DELWRI) == 0)
 123:         u.u_ru.ru_oublock++;        /* noone paid yet */
 124:     trace(TR_BWRITE, pack(bp->b_dev, bp->b_bcount), bp->b_blkno);
 125:     if (bp->b_bcount > bp->b_bufsize)
 126:         panic("bwrite");
 127:     (*bdevsw[major(bp->b_dev)].d_strategy)(bp);
 128: 
 129:     /*
 130: 	 * If the write was synchronous, then await i/o completion.
 131: 	 * If the write was "delayed", then we put the buffer on
 132: 	 * the q of blocks awaiting i/o completion status.
 133: 	 */
 134:     if ((flag&B_ASYNC) == 0) {
 135:         biowait(bp);
 136:         brelse(bp);
 137:     } else if (flag & B_DELWRI)
 138:         bp->b_flags |= B_AGE;
 139: }
 140: 
 141: /*
 142:  * Release the buffer, marking it so that if it is grabbed
 143:  * for another purpose it will be written out before being
 144:  * given up (e.g. when writing a partial block where it is
 145:  * assumed that another write for the same block will soon follow).
 146:  * This can't be done for magtape, since writes must be done
 147:  * in the same order as requested.
 148:  */
 149: bdwrite(bp)
 150:     register struct buf *bp;
 151: {
 152:     register int flags;
 153: 
 154:     if ((bp->b_flags&B_DELWRI) == 0)
 155:         u.u_ru.ru_oublock++;        /* noone paid yet */
 156:     flags = bdevsw[major(bp->b_dev)].d_flags;
 157:     if(flags & B_TAPE)
 158:         bawrite(bp);
 159:     else {
 160:         bp->b_flags |= B_DELWRI | B_DONE;
 161:         brelse(bp);
 162:     }
 163: }
 164: 
 165: /*
 166:  * Release the buffer, start I/O on it, but don't wait for completion.
 167:  */
 168: bawrite(bp)
 169:     register struct buf *bp;
 170: {
 171: 
 172:     bp->b_flags |= B_ASYNC;
 173:     bwrite(bp);
 174: }
 175: 
 176: /*
 177:  * Release the buffer, with no I/O implied.
 178:  */
 179: brelse(bp)
 180:     register struct buf *bp;
 181: {
 182:     register struct buf *flist;
 183:     register s;
 184: 
 185:     trace(TR_BRELSE, pack(bp->b_dev, bp->b_bufsize), bp->b_blkno);
 186:     /*
 187: 	 * If someone's waiting for the buffer, or
 188: 	 * is waiting for a buffer wake 'em up.
 189: 	 */
 190:     if (bp->b_flags&B_WANTED)
 191:         wakeup((caddr_t)bp);
 192:     if (bfreelist[0].b_flags&B_WANTED) {
 193:         bfreelist[0].b_flags &= ~B_WANTED;
 194:         wakeup((caddr_t)bfreelist);
 195:     }
 196:     if (bp->b_flags&B_ERROR)
 197:         if (bp->b_flags & B_LOCKED)
 198:             bp->b_flags &= ~B_ERROR;    /* try again later */
 199:         else
 200:             bp->b_dev = NODEV;          /* no assoc */
 201: 
 202:     /*
 203: 	 * Stick the buffer back on a free list.
 204: 	 */
 205:     s = splbio();
 206:     if (bp->b_bufsize <= 0) {
 207:         /* block has no buffer ... put at front of unused buffer list */
 208:         flist = &bfreelist[BQ_EMPTY];
 209:         binsheadfree(bp, flist);
 210:     } else if (bp->b_flags & (B_ERROR|B_INVAL)) {
 211:         /* block has no info ... put at front of most free list */
 212:         flist = &bfreelist[BQ_AGE];
 213:         binsheadfree(bp, flist);
 214:     } else {
 215:         if (bp->b_flags & B_LOCKED)
 216:             flist = &bfreelist[BQ_LOCKED];
 217:         else if (bp->b_flags & B_AGE)
 218:             flist = &bfreelist[BQ_AGE];
 219:         else
 220:             flist = &bfreelist[BQ_LRU];
 221:         binstailfree(bp, flist);
 222:     }
 223:     bp->b_flags &= ~(B_WANTED|B_BUSY|B_ASYNC|B_AGE);
 224:     splx(s);
 225: }
 226: 
 227: /*
 228:  * See if the block is associated with some buffer
 229:  * (mainly to avoid getting hung up on a wait in breada)
 230:  */
 231: incore(dev, blkno)
 232:     dev_t dev;
 233:     daddr_t blkno;
 234: {
 235:     register struct buf *bp;
 236:     register struct buf *dp;
 237: 
 238:     dp = BUFHASH(dev, blkno);
 239:     for (bp = dp->b_forw; bp != dp; bp = bp->b_forw)
 240:         if (bp->b_blkno == blkno && bp->b_dev == dev &&
 241:             (bp->b_flags & B_INVAL) == 0)
 242:             return (1);
 243:     return (0);
 244: }
 245: 
 246: struct buf *
 247: baddr(dev, blkno, size)
 248:     dev_t dev;
 249:     daddr_t blkno;
 250:     int size;
 251: {
 252: 
 253:     if (incore(dev, blkno))
 254:         return (bread(dev, blkno, size));
 255:     return (0);
 256: }
 257: 
 258: /*
 259:  * Assign a buffer for the given block.  If the appropriate
 260:  * block is already associated, return it; otherwise search
 261:  * for the oldest non-busy buffer and reassign it.
 262:  *
 263:  * We use splx here because this routine may be called
 264:  * on the interrupt stack during a dump, and we don't
 265:  * want to lower the ipl back to 0.
 266:  */
 267: struct buf *
 268: getblk(dev, blkno, size)
 269:     dev_t dev;
 270:     daddr_t blkno;
 271:     int size;
 272: {
 273:     register struct buf *bp, *dp;
 274:     int s;
 275: 
 276:     if (size > MAXBSIZE)
 277:         panic("getblk: size too big");
 278:     /*
 279: 	 * To prevent overflow of 32-bit ints when converting block
 280: 	 * numbers to byte offsets, blknos > 2^32 / DEV_BSIZE are set
 281: 	 * to the maximum number that can be converted to a byte offset
 282: 	 * without overflow. This is historic code; what bug it fixed,
 283: 	 * or whether it is still a reasonable thing to do is open to
 284: 	 * dispute. mkm 9/85
 285: 	 */
 286:     if ((unsigned)blkno >= 1 << (sizeof(int)*NBBY-DEV_BSHIFT))
 287:         blkno = 1 << ((sizeof(int)*NBBY-DEV_BSHIFT) + 1);
 288:     /*
 289: 	 * Search the cache for the block.  If we hit, but
 290: 	 * the buffer is in use for i/o, then we wait until
 291: 	 * the i/o has completed.
 292: 	 */
 293:     dp = BUFHASH(dev, blkno);
 294: loop:
 295:     for (bp = dp->b_forw; bp != dp; bp = bp->b_forw) {
 296:         if (bp->b_blkno != blkno || bp->b_dev != dev ||
 297:             bp->b_flags&B_INVAL)
 298:             continue;
 299:         s = splbio();
 300:         if (bp->b_flags&B_BUSY) {
 301:             bp->b_flags |= B_WANTED;
 302:             sleep((caddr_t)bp, PRIBIO+1);
 303:             splx(s);
 304:             goto loop;
 305:         }
 306:         splx(s);
 307:         notavail(bp);
 308:         if (bp->b_bcount != size && brealloc(bp, size) == 0)
 309:             goto loop;
 310:         bp->b_flags |= B_CACHE;
 311:         return (bp);
 312:     }
 313:     if (major(dev) >= nblkdev)
 314:         panic("blkdev");
 315:     bp = getnewbuf();
 316:     bfree(bp);
 317:     bremhash(bp);
 318:     binshash(bp, dp);
 319:     bp->b_dev = dev;
 320:     bp->b_blkno = blkno;
 321:     bp->b_error = 0;
 322:     if (brealloc(bp, size) == 0)
 323:         goto loop;
 324:     return (bp);
 325: }
 326: 
 327: /*
 328:  * get an empty block,
 329:  * not assigned to any particular device
 330:  */
 331: struct buf *
 332: geteblk(size)
 333:     int size;
 334: {
 335:     register struct buf *bp, *flist;
 336: 
 337:     if (size > MAXBSIZE)
 338:         panic("geteblk: size too big");
 339: loop:
 340:     bp = getnewbuf();
 341:     bp->b_flags |= B_INVAL;
 342:     bfree(bp);
 343:     bremhash(bp);
 344:     flist = &bfreelist[BQ_AGE];
 345:     binshash(bp, flist);
 346:     bp->b_dev = (dev_t)NODEV;
 347:     bp->b_error = 0;
 348:     if (brealloc(bp, size) == 0)
 349:         goto loop;
 350:     return (bp);
 351: }
 352: 
 353: /*
 354:  * Allocate space associated with a buffer.
 355:  * If can't get space, buffer is released
 356:  */
 357: brealloc(bp, size)
 358:     register struct buf *bp;
 359:     int size;
 360: {
 361:     daddr_t start, last;
 362:     register struct buf *ep;
 363:     struct buf *dp;
 364:     int s;
 365: 
 366:     /*
 367: 	 * First need to make sure that all overlaping previous I/O
 368: 	 * is dispatched with.
 369: 	 */
 370:     if (size == bp->b_bcount)
 371:         return (1);
 372:     if (size < bp->b_bcount) {
 373:         if (bp->b_flags & B_DELWRI) {
 374:             bwrite(bp);
 375:             return (0);
 376:         }
 377:         if (bp->b_flags & B_LOCKED)
 378:             panic("brealloc");
 379:         return (allocbuf(bp, size));
 380:     }
 381:     bp->b_flags &= ~B_DONE;
 382:     if (bp->b_dev == NODEV)
 383:         return (allocbuf(bp, size));
 384: 
 385:     trace(TR_BREALLOC, pack(bp->b_dev, size), bp->b_blkno);
 386:     /*
 387: 	 * Search cache for any buffers that overlap the one that we
 388: 	 * are trying to allocate. Overlapping buffers must be marked
 389: 	 * invalid, after being written out if they are dirty. (indicated
 390: 	 * by B_DELWRI) A disk block must be mapped by at most one buffer
 391: 	 * at any point in time. Care must be taken to avoid deadlocking
 392: 	 * when two buffer are trying to get the same set of disk blocks.
 393: 	 */
 394:     start = bp->b_blkno;
 395:     last = start + btodb(size) - 1;
 396:     dp = BUFHASH(bp->b_dev, bp->b_blkno);
 397: loop:
 398:     for (ep = dp->b_forw; ep != dp; ep = ep->b_forw) {
 399:         if (ep == bp || ep->b_dev != bp->b_dev || (ep->b_flags&B_INVAL))
 400:             continue;
 401:         /* look for overlap */
 402:         if (ep->b_bcount == 0 || ep->b_blkno > last ||
 403:             ep->b_blkno + btodb(ep->b_bcount) <= start)
 404:             continue;
 405:         s = splbio();
 406:         if (ep->b_flags&B_BUSY) {
 407:             ep->b_flags |= B_WANTED;
 408:             sleep((caddr_t)ep, PRIBIO+1);
 409:             splx(s);
 410:             goto loop;
 411:         }
 412:         splx(s);
 413:         notavail(ep);
 414:         if (ep->b_flags & B_DELWRI) {
 415:             bwrite(ep);
 416:             goto loop;
 417:         }
 418:         ep->b_flags |= B_INVAL;
 419:         brelse(ep);
 420:     }
 421:     return (allocbuf(bp, size));
 422: }
 423: 
 424: /*
 425:  * Find a buffer which is available for use.
 426:  * Select something from a free list.
 427:  * Preference is to AGE list, then LRU list.
 428:  */
 429: struct buf *
 430: getnewbuf()
 431: {
 432:     register struct buf *bp, *dp;
 433:     int s;
 434: 
 435: loop:
 436:     s = splbio();
 437:     for (dp = &bfreelist[BQ_AGE]; dp > bfreelist; dp--)
 438:         if (dp->av_forw != dp)
 439:             break;
 440:     if (dp == bfreelist) {      /* no free blocks */
 441:         dp->b_flags |= B_WANTED;
 442:         sleep((caddr_t)dp, PRIBIO+1);
 443:         splx(s);
 444:         goto loop;
 445:     }
 446:     splx(s);
 447:     bp = dp->av_forw;
 448:     notavail(bp);
 449:     if (bp->b_flags & B_DELWRI) {
 450:         bp->b_flags |= B_ASYNC;
 451:         bwrite(bp);
 452:         goto loop;
 453:     }
 454:     trace(TR_BRELSE, pack(bp->b_dev, bp->b_bufsize), bp->b_blkno);
 455:     bp->b_flags = B_BUSY;
 456:     return (bp);
 457: }
 458: 
 459: /*
 460:  * Wait for I/O completion on the buffer; return errors
 461:  * to the user.
 462:  */
 463: biowait(bp)
 464:     register struct buf *bp;
 465: {
 466:     int s;
 467: 
 468:     s = splbio();
 469:     while ((bp->b_flags&B_DONE)==0)
 470:         sleep((caddr_t)bp, PRIBIO);
 471:     splx(s);
 472:     if (u.u_error == 0)         /* XXX */
 473:         u.u_error = geterror(bp);
 474: }
 475: 
 476: /*
 477:  * Mark I/O complete on a buffer.
 478:  * If someone should be called, e.g. the pageout
 479:  * daemon, do so.  Otherwise, wake up anyone
 480:  * waiting for it.
 481:  */
 482: biodone(bp)
 483:     register struct buf *bp;
 484: {
 485: 
 486:     if (bp->b_flags & B_DONE)
 487:         panic("dup biodone");
 488:     bp->b_flags |= B_DONE;
 489:     if (bp->b_flags & B_CALL) {
 490:         bp->b_flags &= ~B_CALL;
 491:         (*bp->b_iodone)(bp);
 492:         return;
 493:     }
 494:     if (bp->b_flags&B_ASYNC)
 495:         brelse(bp);
 496:     else {
 497:         bp->b_flags &= ~B_WANTED;
 498:         wakeup((caddr_t)bp);
 499:     }
 500: }
 501: 
 502: /*
 503:  * Insure that no part of a specified block is in an incore buffer.
 504:  */
 505: blkflush(dev, blkno, size)
 506:     dev_t dev;
 507:     daddr_t blkno;
 508:     long size;
 509: {
 510:     register struct buf *ep;
 511:     struct buf *dp;
 512:     daddr_t start, last;
 513:     int s;
 514: 
 515:     start = blkno;
 516:     last = start + btodb(size) - 1;
 517:     dp = BUFHASH(dev, blkno);
 518: loop:
 519:     for (ep = dp->b_forw; ep != dp; ep = ep->b_forw) {
 520:         if (ep->b_dev != dev || (ep->b_flags&B_INVAL))
 521:             continue;
 522:         /* look for overlap */
 523:         if (ep->b_bcount == 0 || ep->b_blkno > last ||
 524:             ep->b_blkno + btodb(ep->b_bcount) <= start)
 525:             continue;
 526:         s = splbio();
 527:         if (ep->b_flags&B_BUSY) {
 528:             ep->b_flags |= B_WANTED;
 529:             sleep((caddr_t)ep, PRIBIO+1);
 530:             splx(s);
 531:             goto loop;
 532:         }
 533:         if (ep->b_flags & B_DELWRI) {
 534:             splx(s);
 535:             notavail(ep);
 536:             bwrite(ep);
 537:             goto loop;
 538:         }
 539:         splx(s);
 540:     }
 541: }
 542: 
 543: /*
 544:  * Make sure all write-behind blocks
 545:  * on dev (or NODEV for all)
 546:  * are flushed out.
 547:  * (from umount and update)
 548:  */
 549: bflush(dev)
 550:     dev_t dev;
 551: {
 552:     register struct buf *bp;
 553:     register struct buf *flist;
 554:     int s;
 555: 
 556: loop:
 557:     s = splbio();
 558:     for (flist = bfreelist; flist < &bfreelist[BQ_EMPTY]; flist++)
 559:     for (bp = flist->av_forw; bp != flist; bp = bp->av_forw) {
 560:         if ((bp->b_flags & B_DELWRI) == 0)
 561:             continue;
 562:         if (dev == NODEV || dev == bp->b_dev) {
 563:             bp->b_flags |= B_ASYNC;
 564:             notavail(bp);
 565:             bwrite(bp);
 566:             splx(s);
 567:             goto loop;
 568:         }
 569:     }
 570:     splx(s);
 571: }
 572: 
 573: /*
 574:  * Pick up the device's error number and pass it to the user;
 575:  * if there is an error but the number is 0 set a generalized code.
 576:  */
 577: geterror(bp)
 578:     register struct buf *bp;
 579: {
 580:     int error = 0;
 581: 
 582:     if (bp->b_flags&B_ERROR)
 583:         if ((error = bp->b_error)==0)
 584:             return (EIO);
 585:     return (error);
 586: }
 587: 
 588: /*
 589:  * Invalidate in core blocks belonging to closed or umounted filesystem
 590:  *
 591:  * This is not nicely done at all - the buffer ought to be removed from the
 592:  * hash chains & have its dev/blkno fields clobbered, but unfortunately we
 593:  * can't do that here, as it is quite possible that the block is still
 594:  * being used for i/o. Eventually, all disc drivers should be forced to
 595:  * have a close routine, which ought ensure that the queue is empty, then
 596:  * properly flush the queues. Until that happy day, this suffices for
 597:  * correctness.						... kre
 598:  */
 599: binval(dev)
 600:     dev_t dev;
 601: {
 602:     register struct buf *bp;
 603:     register struct bufhd *hp;
 604: #define dp ((struct buf *)hp)
 605: 
 606:     for (hp = bufhash; hp < &bufhash[BUFHSZ]; hp++)
 607:         for (bp = dp->b_forw; bp != dp; bp = bp->b_forw)
 608:             if (bp->b_dev == dev)
 609:                 bp->b_flags |= B_INVAL;
 610: }

Defined functions

Defined macros