patch-1.3.69 linux/drivers/block/md.c

• Lines: 815
• Date: Mon Feb 26 13:51:45 1996
• Orig file: v1.3.68/linux/drivers/block/md.c
• Orig date: Thu Jan 1 02:00:00 1970

diff -u --recursive --new-file v1.3.68/linux/drivers/block/md.c linux/drivers/block/md.c
@@ -0,0 +1,814 @@
+
+/*
+   md.c : Multiple Devices driver for Linux
+          Copyright (C) 1994-96 Marc ZYNGIER
+	  <zyngier@ufr-info-p7.ibp.fr> or
+	  <maz@gloups.fdn.fr>
+
+   A lot of inspiration came from hd.c ...
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2, or (at your option)
+   any later version.
+   
+   You should have received a copy of the GNU General Public License
+   (for example /usr/src/linux/COPYING); if not, write to the Free
+   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  
+*/
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/version.h>
+#include <linux/malloc.h>
+#include <linux/mm.h>
+#include <linux/md.h>
+#include <linux/hdreg.h>
+#include <linux/stat.h>
+#include <linux/fs.h>
+#include <linux/proc_fs.h>
+#include <linux/blkdev.h>
+#include <errno.h>
+
+#define MAJOR_NR MD_MAJOR
+#define MD_DRIVER
+
+#include <linux/blk.h>
+
+#ifdef CONFIG_MD_SUPPORT_RAID5
+int support_for_raid5; /* So raid-5 module won't be inserted if support
+			  was not set in the kernel */
+#endif
+
+#ifdef CONFIG_MD_SUPPORT_RAID1
+int support_for_raid1; /* So raid-1 module won't be inserted if support
+			  was not set in the kernel */
+#endif
+
+static struct hd_struct md_hd_struct[MAX_MD_DEV];
+static int md_blocksizes[MAX_MD_DEV];
+
+int md_size[MAX_MD_DEV]={0, };
+
+static void md_geninit (struct gendisk *);
+
+static struct gendisk md_gendisk=
+{
+  MD_MAJOR,
+  "md",
+  0,
+  1,
+  MAX_MD_DEV,
+  md_geninit,
+  md_hd_struct,
+  md_size,
+  MAX_MD_DEV,
+  NULL,
+  NULL
+};
+
+static struct md_personality *pers[MAX_PERSONALITY]={NULL, };
+
+struct real_dev devices[MAX_MD_DEV][MAX_REAL];
+struct md_dev md_dev[MAX_MD_DEV];
+
+static struct gendisk *find_gendisk (kdev_t dev)
+{
+  struct gendisk *tmp=gendisk_head;
+
+  while (tmp != NULL)
+  {
+    if (tmp->major==MAJOR(dev))
+      return (tmp);
+    
+    tmp=tmp->next;
+  }
+
+  return (NULL);
+}
+
+
+/* Picked up from genhd.c */
+char *partition_name (kdev_t dev)
+{
+  static char name[10];		/* This should be long
+				   enough for a device name ! */
+  struct gendisk *hd=find_gendisk (dev);
+  char base_name;
+  int minor=MINOR(dev);
+
+  if (!hd)
+  {
+    printk ("No gendisk entry for dev %04x\n", dev);
+    sprintf (name, "dev %04x", dev);
+    return (name);
+  }
+
+  base_name = (hd->major == IDE1_MAJOR) ? 'c' : 'a';
+  sprintf(name, "%s%c%d",
+	  hd->major_name,
+	  base_name + (minor >> hd->minor_shift),
+	  minor & ((1 << hd->minor_shift) - 1));
+  return (name);
+}
+
+
+static void set_ra (void)
+{
+  int i, j, minra=INT_MAX;
+
+  for (i=0; i<MAX_MD_DEV; i++)
+  {
+    if (!md_dev[i].pers)
+      continue;
+    
+    for (j=0; j<md_dev[i].nb_dev; j++)
+      if (read_ahead[MAJOR(devices[i][j].dev)]<minra)
+	minra=read_ahead[MAJOR(devices[i][j].dev)];
+  }
+  
+  read_ahead[MD_MAJOR]=minra;
+}
+
+
+static int md_ioctl (struct inode *inode, struct file *file,
+                     unsigned int cmd, unsigned long arg)
+{
+  int minor, index, err, current_ra;
+  struct gendisk *gen_real;
+  struct hd_geometry *loc = (struct hd_geometry *) arg;
+  kdev_t dev;
+
+  if (!suser())
+    return -EACCES;
+
+  if (((minor=MINOR(inode->i_rdev)) & 0x80) &&
+      (minor & 0x7f) < MAX_PERSONALITY &&
+      pers[minor & 0x7f] &&
+      pers[minor & 0x7f]->ioctl)
+    return (pers[minor & 0x7f]->ioctl (inode, file, cmd, arg));
+  
+  if (minor >= MAX_MD_DEV)
+    return -EINVAL;
+
+  switch (cmd)
+  {
+    case REGISTER_DEV:
+    dev=to_kdev_t ((dev_t) arg);
+    if (MAJOR(dev)==MD_MAJOR || md_dev[minor].nb_dev==MAX_REAL)
+      return -EINVAL;
+
+    if (!fs_may_mount (dev) || md_dev[minor].pers)
+      return -EBUSY;
+
+    if (!(gen_real=find_gendisk (dev)))
+      return -ENOENT;
+
+    index=md_dev[minor].nb_dev++;
+    devices[minor][index].dev=dev;
+
+    /* Lock the device by inserting a dummy inode. This doesn't
+       smeel very good, but I need to be consistent with the
+       mount stuff, specially with fs_may_mount. If someone have
+       a better idea, please help ! */
+    
+    devices[minor][index].inode=get_empty_inode ();
+    devices[minor][index].inode->i_dev=dev; /* don't care about
+					       other fields */
+    insert_inode_hash (devices[minor][index].inode);
+    
+    /* Devices sizes are rounded to a multiple of page (needed for
+       paging). This is NOT done by fdisk when partitionning,
+       but that's a DOS thing anyway... */
+    
+    devices[minor][index].size=gen_real->sizes[MINOR(dev)] & (PAGE_MASK>>10);
+    devices[minor][index].offset=index ?
+      (devices[minor][index-1].offset + devices[minor][index-1].size) : 0;
+
+    if (!index)
+      md_size[minor]=devices[minor][index].size;
+    else
+      md_size[minor]+=devices[minor][index].size;
+
+    printk("REGISTER_DEV %s to md%x done\n", partition_name(dev), minor);
+    break;
+
+    case START_MD:
+    if (!md_dev[minor].nb_dev)
+      return -EINVAL;
+
+    if (md_dev[minor].pers)
+      return -EBUSY;
+
+    md_dev[minor].repartition=(int) arg;
+    
+    if ((index=PERSONALITY(md_dev+minor) >> (PERSONALITY_SHIFT))
+	>= MAX_PERSONALITY ||
+	!pers[index])
+      return -EINVAL;
+
+    md_dev[minor].pers=pers[index];
+
+    if ((err=md_dev[minor].pers->run (minor, md_dev+minor)))
+    {
+      md_dev[minor].pers=NULL;
+      return (err);
+    }
+
+    /* FIXME : We assume here we have blocks
+       that are twice as large as sectors.
+       THIS MAY NOT BE TRUE !!! */
+    md_hd_struct[minor].start_sect=0;
+    md_hd_struct[minor].nr_sects=md_size[minor]<<1;
+
+    /* It would be better to have a per-md-dev read_ahead. Currently,
+       we only use the smallest read_ahead among md-attached devices */
+
+    current_ra=read_ahead[MD_MAJOR];
+    
+    for (index=0; index<md_dev[minor].nb_dev; index++)
+    {
+      if (current_ra>read_ahead[MAJOR(devices[minor][index].dev)])
+	current_ra=read_ahead[MAJOR(devices[minor][index].dev)];
+
+      devices[minor][index].fault_count=0;
+      devices[minor][index].invalid=VALID;
+    }
+
+    read_ahead[MD_MAJOR]=current_ra;
+
+    printk ("START_DEV md%x %s\n", minor, md_dev[minor].pers->name);
+    break;
+
+    case STOP_MD:
+    if (inode->i_count>1 || md_dev[minor].busy>1) /* ioctl : one open channel */
+    {
+      printk ("STOP_MD md%x failed : i_count=%d, busy=%d\n", minor, inode->i_count, md_dev[minor].busy);
+      return -EBUSY;
+    }
+
+    if (md_dev[minor].pers)
+    {
+      /*  The device won't exist anymore -> flush it now */
+      fsync_dev (inode->i_rdev);
+      invalidate_buffers (inode->i_rdev);
+      md_dev[minor].pers->stop (minor, md_dev+minor);
+    }
+
+    /* Remove locks. */
+    for (index=0; index<md_dev[minor].nb_dev; index++)
+      clear_inode (devices[minor][index].inode);
+
+    md_dev[minor].nb_dev=md_size[minor]=0;
+    md_dev[minor].pers=NULL;
+
+    set_ra ();			/* calculate new read_ahead */
+    
+    printk ("STOP_DEV md%x\n", minor);
+    break;
+
+#if defined(CONFIG_MD_SUPPORT_RAID1) || defined(CONFIG_MD_SUPPORT_RAID5)
+    case MD_INVALID:
+    dev=to_kdev_t ((dev_t) arg);
+    if (!(err=md_valid_device (minor, dev, INVALID_ALWAYS)))
+      printk ("md%d : %s disabled\n", minor, partition_name (dev));
+
+    return (err);
+
+    case MD_VALID:
+    dev=to_kdev_t ((dev_t) arg);
+    if (!(err=md_valid_device (minor, dev, VALID)))
+      printk ("md%d : %s enabled\n", minor, partition_name (dev));
+
+    return (err);
+#endif
+    
+    case BLKGETSIZE:   /* Return device size */
+    if  (!arg)  return -EINVAL;
+    err=verify_area (VERIFY_WRITE, (long *) arg, sizeof(long));
+    if (err)
+      return err;
+    put_user (md_hd_struct[MINOR(inode->i_rdev)].nr_sects, (long *) arg);
+    break;
+
+    case BLKFLSBUF:
+    fsync_dev (inode->i_rdev);
+    invalidate_buffers (inode->i_rdev);
+    break;
+
+    case BLKRASET:
+    if (arg > 0xff)
+      return -EINVAL;
+    read_ahead[MAJOR(inode->i_rdev)] = arg;
+    return 0;
+    
+    case BLKRAGET:
+    if  (!arg)  return -EINVAL;
+    err=verify_area (VERIFY_WRITE, (long *) arg, sizeof(long));
+    if (err)
+      return err;
+    put_user (read_ahead[MAJOR(inode->i_rdev)], (long *) arg);
+    break;
+
+    case HDIO_GETGEO:
+    if (!loc)  return -EINVAL;
+    err = verify_area(VERIFY_WRITE, loc, sizeof(*loc));
+    if (err)
+      return err;
+    put_user (2, (char *) &loc->heads);
+    put_user (4, (char *) &loc->sectors);
+    put_user (md_hd_struct[minor].nr_sects/8, (short *) &loc->cylinders);
+    put_user (md_hd_struct[MINOR(inode->i_rdev)].start_sect,
+		(long *) &loc->start);
+    break;
+    
+    RO_IOCTLS(inode->i_rdev,arg);
+    
+    default:
+    printk ("Unknown md_ioctl %d\n", cmd);
+    return -EINVAL;
+  }
+
+  return (0);
+}
+
+
+static int md_open (struct inode *inode, struct file *file)
+{
+  int minor=MINOR(inode->i_rdev);
+
+  md_dev[minor].busy++;
+  return (0);			/* Always succed */
+}
+
+
+static void md_release (struct inode *inode, struct file *file)
+{
+  int minor=MINOR(inode->i_rdev);
+
+  sync_dev (inode->i_rdev);
+  md_dev[minor].busy--;
+}
+
+
+static struct file_operations md_fops=
+{
+  NULL,
+  block_read,
+  block_write,
+  NULL,
+  NULL,
+  md_ioctl,
+  NULL,
+  md_open,
+  md_release,
+  block_fsync
+};
+
+
+static inline int remap_request (int minor, struct request *req)
+{
+  if (!md_dev[minor].pers)
+  {
+    printk ("Oops ! md%d not running, giving up !\n", minor);
+    return -1;
+  }
+
+  return (md_dev[minor].pers->map(minor, md_dev+minor, req));
+}
+
+static void do_md_request (void)
+{
+  int minor;
+  struct request *req;
+    
+  while (1)
+  {
+#ifdef MD_COUNT_SIZE
+    int reqsize, chunksize;
+#endif
+    
+    cli ();
+    req = blk_dev[MD_MAJOR].current_request;
+    if (!req || (req->rq_status == RQ_INACTIVE))
+    {
+      sti ();
+      return;
+    }
+    
+#ifdef MD_COUNT_SIZE
+    reqsize=req->nr_sectors>>1;
+    chunksize=1 << FACTOR_SHIFT(FACTOR(md_dev+MINOR(req->rq_dev)));
+    if (reqsize==chunksize) (md_dev+MINOR(req->rq_dev))->equal_count++;
+    if (reqsize<chunksize) (md_dev+MINOR(req->rq_dev))->smallest_count++;
+    if (reqsize>chunksize) (md_dev+MINOR(req->rq_dev))->biggest_count++;
+#endif
+    
+    blk_dev[MD_MAJOR].current_request = req->next;
+    sti ();
+
+    minor = MINOR(req->rq_dev);
+    if ((MAJOR(req->rq_dev) != MD_MAJOR) || (minor >= MAX_REAL))
+    {
+      printk("md: bad device number: 0x%04x\n", req->rq_dev);
+      end_request(0, req);
+      continue;
+    }
+
+    switch (remap_request (minor, req))
+    {
+      case REDIRECTED_BHREQ:	/* Allright, redirection was succesful */
+      req->rq_status=RQ_INACTIVE;
+      wake_up (&wait_for_request);
+      break;
+
+      case REDIRECTED_REQ:
+      break;			/* Redirected whole request (for swaping) */
+      
+      case REDIRECT_FAILED:	/* Swap redirection failed in RAID-[15] */
+      end_request (0, req);
+      break;
+      
+      default:
+      printk ("remap_request returned strange value !\n");
+    }
+  }
+}
+
+extern struct semaphore request_lock;
+
+void make_md_request (struct request *pending, int n)
+{
+  int i, j, max_req, major=0, rw, found;
+  kdev_t dev;
+  struct buffer_head *bh;
+  struct request *req;
+  
+  down (&request_lock);
+
+  for (i=0; i<n; i++)
+  {
+    if (!pending[i].bh)
+      continue;
+
+    cli();
+
+    found=0;
+    rw=pending[i].cmd;
+    bh=pending[i].bh;
+    major=MAJOR(dev=pending[i].rq_dev);
+    
+    max_req = (rw == READ) ? NR_REQUEST : ((NR_REQUEST*2)/3);
+ 
+    if ((   major == IDE0_MAJOR	/* same as HD_MAJOR */
+	 || major == IDE1_MAJOR
+	 || major == SCSI_DISK_MAJOR
+	 || major == IDE2_MAJOR
+	 || major == IDE3_MAJOR)
+	&& (req = blk_dev[major].current_request))
+    {
+#ifdef CONFIG_BLK_DEV_HD
+      if (major == HD_MAJOR)
+	req = req->next;
+#endif CONFIG_BLK_DEV_HD
+
+      while (req && !found)
+      {
+	if (req->rq_status!=RQ_INACTIVE && req->rq_status!=RQ_ACTIVE)
+	  printk ("Saw bad status request !\n");
+
+	if (req->rq_dev == dev &&
+	    !req->sem &&
+	    req->cmd == rw &&
+	    req->sector + req->nr_sectors == pending[i].sector &&
+	    (req->nr_sectors + pending[i].nr_sectors) < 245)
+	{
+	  req->bhtail->b_reqnext = bh;
+	  req->bhtail = pending[i].bhtail;
+	  req->nr_sectors += pending[i].nr_sectors;
+	  found=1;
+	  continue;
+	}
+	
+	if (!found &&
+	    req->rq_dev == dev &&
+	    !req->sem &&
+	    req->cmd == rw &&
+	    req->sector - pending[i].nr_sectors == pending[i].sector &&
+	    (req->nr_sectors + pending[i].nr_sectors) < 245)
+	{
+	  req->nr_sectors += pending[i].nr_sectors;
+	  bh->b_reqnext = req->bh;
+	  req->buffer = bh->b_data;
+	  req->current_nr_sectors = bh->b_size >> 9;
+	  req->sector = pending[i].sector;
+	  req->bh = bh;
+	  found=1;
+	  continue;
+	}    
+
+	req = req->next;
+      }
+    }
+
+    if (found)
+      continue;
+
+    up (&request_lock);
+    req=get_md_request (max_req, dev);
+    
+    /* Build it up... */
+    req->cmd = rw;
+    req->errors = 0;
+#if defined (CONFIG_MD_SUPPORT_RAID1)
+    req->shared_count = 0;
+#endif
+    req->sector = pending[i].sector;
+    req->nr_sectors = pending[i].nr_sectors;
+    req->current_nr_sectors = bh->b_size >> 9;
+    req->buffer = bh->b_data;
+    req->sem = NULL;
+    req->bh = bh;
+    req->bhtail = pending[i].bhtail;
+    req->next = NULL;
+
+    add_request (blk_dev + MAJOR(dev), req);
+    down (&request_lock);
+  }
+
+  up (&request_lock);
+  for (j=0; j<n; j++)
+  {
+    if (!pending[j].bh)
+      continue;
+    
+    pending[j].bh=NULL;
+  }
+
+  sti ();
+}
+
+
+static struct symbol_table md_symbol_table=
+{
+#include <linux/symtab_begin.h>
+
+  X(devices),
+  X(md_size),
+  X(add_request),
+  X(make_md_request),
+
+#ifdef CONFIG_MD_SUPPORT_RAID1
+  X(support_for_raid1),
+#endif
+
+#ifdef CONFIG_MD_SUPPORT_RAID5
+  X(support_for_raid5),
+#endif
+
+  X(register_md_personality),
+  X(unregister_md_personality),
+  X(partition_name),
+
+#if defined(CONFIG_MD_SUPPORT_RAID1) || defined(CONFIG_MD_SUPPORT_RAID5)
+  X(md_valid_device),
+  X(md_can_reemit),
+#endif
+
+#include <linux/symtab_end.h>
+};
+
+
+static void md_geninit (struct gendisk *gdisk)
+{
+  int i;
+  
+  for(i=0;i<MAX_MD_DEV;i++)
+  {
+    md_blocksizes[i] = 1024;
+    md_gendisk.part[i].start_sect=-1;
+    md_dev[i].pers=NULL;
+#ifdef MD_COUNT_SIZES
+    md_dev[i].smallest_count=md_dev[i].biggest_count=md_dev[i].equal_count=0;
+#endif
+  }
+
+  blksize_size[MAJOR_NR] = md_blocksizes;
+  register_symtab (&md_symbol_table);
+
+  proc_register(&proc_root,
+		&(struct proc_dir_entry)
+	      {
+		PROC_MD, 6, "mdstat",
+		S_IFREG | S_IRUGO, 1, 0, 0,
+	      });
+}
+
+
+int get_md_status (char *page)
+{
+  int sz=0, i, j;
+
+  sz+=sprintf( page+sz, "Personalities : ");
+  for (i=0; i<MAX_PERSONALITY; i++)
+    if (pers[i])
+      sz+=sprintf (page+sz, "[%d %s] ", i, pers[i]->name);
+
+  page[sz-1]='\n';
+
+  sz+=sprintf (page+sz, "read_ahead ");
+  if (read_ahead[MD_MAJOR]==INT_MAX)
+    sz+=sprintf (page+sz, "not set\n");
+  else
+    sz+=sprintf (page+sz, "%d sectors\n", read_ahead[MD_MAJOR]);
+  
+  for (i=0; i<MAX_MD_DEV; i++)
+  {
+    sz+=sprintf (page+sz, "md%d : %sactive", i, md_dev[i].pers ? "" : "in");
+
+    if (md_dev[i].pers)
+      sz+=sprintf (page+sz, " %s", md_dev[i].pers->name);
+
+    for (j=0; j<md_dev[i].nb_dev; j++)
+      sz+=sprintf (page+sz, " %s%s%s",
+		   (devices[i][j].invalid==VALID) ? "" : "(",
+		   partition_name(devices[i][j].dev),
+		   (devices[i][j].invalid==VALID) ? "" : ")");
+    
+    if (md_dev[i].nb_dev)
+      sz+=sprintf (page+sz, " %d blocks", md_size[i]);
+
+    if (!md_dev[i].pers)
+    {
+      sz+=sprintf (page+sz, "\n");
+      continue;
+    }
+
+    if (md_dev[i].pers->max_invalid_dev)
+      sz+=sprintf (page+sz, " maxfault=%ld", MAX_FAULT(md_dev+i));
+
+    if (md_dev[i].pers != pers[(LINEAR>>PERSONALITY_SHIFT)])
+    {
+      sz+=sprintf (page+sz, " %dk chunks", 1<<FACTOR_SHIFT(FACTOR(md_dev+i)));
+#ifdef MD_COUNT_SIZES
+      sz+=sprintf (page+sz, " (%d/%d/%d)",
+		   md_dev[i].smallest_count,
+		   md_dev[i].equal_count,
+		   md_dev[i].biggest_count);
+#endif
+    }
+    sz+=sprintf (page+sz, "\n");
+    sz+=md_dev[i].pers->status (page+sz, i, md_dev+i);
+  }
+  
+  return (sz);
+}
+
+#if defined(CONFIG_MD_SUPPORT_RAID1) || defined(CONFIG_MD_SUPPORT_RAID5)
+
+int md_valid_device (int minor, kdev_t dev, int mode)
+{
+  int i;
+
+  for (i=0; i<md_dev[minor].nb_dev; i++)
+    if (devices[minor][i].dev==dev)
+      break;
+
+  if (i>md_dev[minor].nb_dev)
+  {
+    printk ("Oops, dev %04x not found in md_valid_device\n", dev);
+    return -EINVAL;
+  }
+
+  switch (mode)
+  {
+    case VALID:
+    /* Don't consider INVALID_NEXT as a real invalidation.
+       Maybe that's not the good way to treat such a thing,
+       we'll see. */
+    if (devices[minor][i].invalid==INVALID_ALWAYS)
+    {
+      devices[minor][i].fault_count=0; /* reset fault count */
+      if (md_dev[minor].invalid_dev_count)
+	md_dev[minor].invalid_dev_count--;
+    }
+    break;
+
+    case INVALID:
+    if (devices[minor][i].invalid != VALID )
+      return 0;			/* Don't invalidate twice */
+    
+    if (++devices[minor][i].fault_count > MAX_FAULT(md_dev+minor) &&
+	MAX_FAULT(md_dev+minor)!=0xFF)
+    {
+      /* We cannot tolerate this fault.
+	 So sing a song, and say GoodBye to this device... */
+      
+      mode=INVALID_ALWAYS;
+      md_dev[minor].invalid_dev_count++;
+    }
+    else
+      /* FIXME :
+	 If we reached the max_invalid_dev count, doing one
+	 more invalidation will kill the md_dev. So we choose
+	 not to invalid the physical dev in such a case. But
+	 next access will probably fail... */
+      if (md_dev[minor].invalid_dev_count<=md_dev[minor].pers->max_invalid_dev)
+	mode=INVALID_NEXT;
+      else
+	mode=VALID;
+    break;
+
+    case INVALID_ALWAYS:	/* Only used via MD_INVALID ioctl */
+    md_dev[minor].invalid_dev_count++;
+  }
+  
+  devices[minor][i].invalid=mode;
+  return 0;
+}
+
+
+int md_can_reemit (int minor)
+{
+  /* FIXME :
+     If the device is raid-1 (md_dev[minor].pers->max_invalid_dev=-1),
+     always pretend that we can reemit the request.
+     Problem : if the 2 devices in the pair are dead, will loop
+     forever. Maybe having a per-personality can_reemit function would
+     help. */
+
+  if (!md_dev[minor].pers)
+    return (0);
+  
+  return(md_dev[minor].pers->max_invalid_dev &&
+	 ((md_dev[minor].pers->max_invalid_dev==-1) ?
+	 1 :
+	 md_dev[minor].invalid_dev_count<=md_dev[minor].pers->max_invalid_dev));
+}
+
+#endif
+
+int register_md_personality (int p_num, struct md_personality *p)
+{
+  int i=(p_num >> PERSONALITY_SHIFT);
+
+  if (i >= MAX_PERSONALITY)
+    return -EINVAL;
+
+  if (pers[i])
+    return -EBUSY;
+  
+  pers[i]=p;
+  printk ("%s personality registered\n", p->name);
+  return 0;
+}
+
+int unregister_md_personality (int p_num)
+{
+  int i=(p_num >> PERSONALITY_SHIFT);
+
+  if (i >= MAX_PERSONALITY)
+    return -EINVAL;
+
+  printk ("%s personality unregistered\n", pers[i]->name);
+  pers[i]=NULL;
+  return 0;
+} 
+
+void linear_init (void);
+void raid0_init (void);
+void raid1_init (void);
+void raid5_init (void);
+
+int md_init (void)
+{
+  printk ("md driver %s MAX_MD_DEV=%d, MAX_REAL=%d\n", MD_VERSION, MAX_MD_DEV, MAX_REAL);
+
+  if (register_blkdev (MD_MAJOR, "md", &md_fops))
+  {
+    printk ("Unable to get major %d for md\n", MD_MAJOR);
+    return (-1);
+  }
+
+  blk_dev[MD_MAJOR].request_fn=DEVICE_REQUEST;
+  blk_dev[MD_MAJOR].current_request=NULL;
+  read_ahead[MD_MAJOR]=INT_MAX;
+  md_gendisk.next=gendisk_head;
+
+  gendisk_head=&md_gendisk;
+
+#ifdef CONFIG_MD_LINEAR
+  linear_init ();
+#endif
+#ifdef CONFIG_MD_STRIPED
+  raid0_init ();
+#endif
+#ifdef CONFIG_MD_RAID1
+  raid1_init ();
+#endif
+#ifdef CONFIG_MD_RAID5
+  raid5_init ();
+#endif
+  
+  return (0);
+}