md.c

Linux内核源代码 为压缩文件 是<<Linux内核>>一书中的源代码

		printk("do_md_run() returned %d\n", err);
		/*
		 * prevent the writeback of an unrunnable array
		 */
		mddev->sb_dirty = 0;
		do_md_stop (mddev, 0);
	}
}

/*
 * lets try to run arrays based on all disks that have arrived
 * until now. (those are in the ->pending list)
 *
 * the method: pick the first pending disk, collect all disks with
 * the same UUID, remove all from the pending list and put them into
 * the 'same_array' list. Then order this list based on superblock
 * update time (freshest comes first), kick out 'old' disks and
 * compare superblocks. If everything's fine then run it.
 *
 * If "unit" is allocated, then bump its reference count
 */
static void autorun_devices (kdev_t countdev)
{
	struct md_list_head candidates;
	struct md_list_head *tmp;
	mdk_rdev_t *rdev0, *rdev;
	mddev_t *mddev;
	kdev_t md_kdev;


	printk("autorun ...\n");
	while (pending_raid_disks.next != &pending_raid_disks) {
		rdev0 = md_list_entry(pending_raid_disks.next,
					 mdk_rdev_t, pending);

		printk("considering %s ...\n", partition_name(rdev0->dev));
		MD_INIT_LIST_HEAD(&candidates);
		ITERATE_RDEV_PENDING(rdev,tmp) {
			if (uuid_equal(rdev0, rdev)) {
				if (!sb_equal(rdev0->sb, rdev->sb)) {
					printk("%s has same UUID as %s, but superblocks differ ...\n", partition_name(rdev->dev), partition_name(rdev0->dev));
					continue;
				}
				printk("  adding %s ...\n", partition_name(rdev->dev));
				md_list_del(&rdev->pending);
				md_list_add(&rdev->pending, &candidates);
			}
		}
		/*
		 * now we have a set of devices, with all of them having
		 * mostly sane superblocks. It's time to allocate the
		 * mddev.
		 */
		md_kdev = MKDEV(MD_MAJOR, rdev0->sb->md_minor);
		mddev = kdev_to_mddev(md_kdev);
		if (mddev) {
			printk("md%d already running, cannot run %s\n",
				 mdidx(mddev), partition_name(rdev0->dev));
			ITERATE_RDEV_GENERIC(candidates,pending,rdev,tmp)
				export_rdev(rdev);
			continue;
		}
		mddev = alloc_mddev(md_kdev);
 		if (mddev == NULL) {
 			printk("md: cannot allocate memory for md drive.\n");
 			break;
 		}
 		if (md_kdev == countdev)
 			atomic_inc(&mddev->active);
		printk("created md%d\n", mdidx(mddev));
		ITERATE_RDEV_GENERIC(candidates,pending,rdev,tmp) {
			bind_rdev_to_array(rdev, mddev);
			md_list_del(&rdev->pending);
			MD_INIT_LIST_HEAD(&rdev->pending);
		}
		autorun_array(mddev);
	}
	printk("... autorun DONE.\n");
}

/*
 * import RAID devices based on one partition
 * if possible, the array gets run as well.
 */

#define BAD_VERSION KERN_ERR \
"md: %s has RAID superblock version 0.%d, autodetect needs v0.90 or higher\n"

#define OUT_OF_MEM KERN_ALERT \
"md: out of memory.\n"

#define NO_DEVICE KERN_ERR \
"md: disabled device %s\n"

#define AUTOADD_FAILED KERN_ERR \
"md: auto-adding devices to md%d FAILED (error %d).\n"

#define AUTOADD_FAILED_USED KERN_ERR \
"md: cannot auto-add device %s to md%d, already used.\n"

#define AUTORUN_FAILED KERN_ERR \
"md: auto-running md%d FAILED (error %d).\n"

#define MDDEV_BUSY KERN_ERR \
"md: cannot auto-add to md%d, already running.\n"

#define AUTOADDING KERN_INFO \
"md: auto-adding devices to md%d, based on %s's superblock.\n"

#define AUTORUNNING KERN_INFO \
"md: auto-running md%d.\n"

static int autostart_array (kdev_t startdev, kdev_t countdev)
{
	int err = -EINVAL, i;
	mdp_super_t *sb = NULL;
	mdk_rdev_t *start_rdev = NULL, *rdev;

	if (md_import_device(startdev, 1)) {
		printk("could not import %s!\n", partition_name(startdev));
		goto abort;
	}

	start_rdev = find_rdev_all(startdev);
	if (!start_rdev) {
		MD_BUG();
		goto abort;
	}
	if (start_rdev->faulty) {
		printk("can not autostart based on faulty %s!\n",
						partition_name(startdev));
		goto abort;
	}
	md_list_add(&start_rdev->pending, &pending_raid_disks);

	sb = start_rdev->sb;

	err = detect_old_array(sb);
	if (err) {
		printk("array version is too old to be autostarted, use raidtools 0.90 mkraid --upgrade\nto upgrade the array without data loss!\n");
		goto abort;
	}

	for (i = 0; i < MD_SB_DISKS; i++) {
		mdp_disk_t *desc;
		kdev_t dev;

		desc = sb->disks + i;
		dev = MKDEV(desc->major, desc->minor);

		if (dev == MKDEV(0,0))
			continue;
		if (dev == startdev)
			continue;
		if (md_import_device(dev, 1)) {
			printk("could not import %s, trying to run array nevertheless.\n", partition_name(dev));
			continue;
		}
		rdev = find_rdev_all(dev);
		if (!rdev) {
			MD_BUG();
			goto abort;
		}
		md_list_add(&rdev->pending, &pending_raid_disks);
	}

	/*
	 * possibly return codes
	 */
	autorun_devices(countdev);
	return 0;

abort:
	if (start_rdev)
		export_rdev(start_rdev);
	return err;
}

#undef BAD_VERSION
#undef OUT_OF_MEM
#undef NO_DEVICE
#undef AUTOADD_FAILED_USED
#undef AUTOADD_FAILED
#undef AUTORUN_FAILED
#undef AUTOADDING
#undef AUTORUNNING

struct {
	int set;
	int noautodetect;

} raid_setup_args md__initdata = { 0, 0 };

void md_setup_drive(void) md__init;

/*
 * Searches all registered partitions for autorun RAID arrays
 * at boot time.
 */
#ifdef CONFIG_AUTODETECT_RAID
static int detected_devices[128] md__initdata = { 0, };
static int dev_cnt=0;
void md_autodetect_dev(kdev_t dev)
{
	if (dev_cnt >= 0 && dev_cnt < 127)
		detected_devices[dev_cnt++] = dev;
}
#endif

int md__init md_run_setup(void)
{
#ifdef CONFIG_AUTODETECT_RAID
	mdk_rdev_t *rdev;
	int i;

	if (raid_setup_args.noautodetect)
		printk(KERN_INFO "skipping autodetection of RAID arrays\n");
	else {

		printk(KERN_INFO "autodetecting RAID arrays\n");

		for (i=0; i<dev_cnt; i++) {
			kdev_t dev = detected_devices[i];

			if (md_import_device(dev,1)) {
				printk(KERN_ALERT "could not import %s!\n",
				       partition_name(dev));
				continue;
			}
			/*
			 * Sanity checks:
			 */
			rdev = find_rdev_all(dev);
			if (!rdev) {
				MD_BUG();
				continue;
			}
			if (rdev->faulty) {
				MD_BUG();
				continue;
			}
			md_list_add(&rdev->pending, &pending_raid_disks);
		}

		autorun_devices(-1);
	}

	dev_cnt = -1; /* make sure further calls to md_autodetect_dev are ignored */
#endif
#ifdef CONFIG_MD_BOOT
	md_setup_drive();
#endif
	return 0;
}

static int get_version (void * arg)
{
	mdu_version_t ver;

	ver.major = MD_MAJOR_VERSION;
	ver.minor = MD_MINOR_VERSION;
	ver.patchlevel = MD_PATCHLEVEL_VERSION;

	if (md_copy_to_user(arg, &ver, sizeof(ver)))
		return -EFAULT;

	return 0;
}

#define SET_FROM_SB(x) info.x = mddev->sb->x
static int get_array_info (mddev_t * mddev, void * arg)
{
	mdu_array_info_t info;

	if (!mddev->sb)
		return -EINVAL;

	SET_FROM_SB(major_version);
	SET_FROM_SB(minor_version);
	SET_FROM_SB(patch_version);
	SET_FROM_SB(ctime);
	SET_FROM_SB(level);
	SET_FROM_SB(size);
	SET_FROM_SB(nr_disks);
	SET_FROM_SB(raid_disks);
	SET_FROM_SB(md_minor);
	SET_FROM_SB(not_persistent);

	SET_FROM_SB(utime);
	SET_FROM_SB(state);
	SET_FROM_SB(active_disks);
	SET_FROM_SB(working_disks);
	SET_FROM_SB(failed_disks);
	SET_FROM_SB(spare_disks);

	SET_FROM_SB(layout);
	SET_FROM_SB(chunk_size);

	if (md_copy_to_user(arg, &info, sizeof(info)))
		return -EFAULT;

	return 0;
}
#undef SET_FROM_SB

#define SET_FROM_SB(x) info.x = mddev->sb->disks[nr].x
static int get_disk_info (mddev_t * mddev, void * arg)
{
	mdu_disk_info_t info;
	unsigned int nr;

	if (!mddev->sb)
		return -EINVAL;

	if (md_copy_from_user(&info, arg, sizeof(info)))
		return -EFAULT;

	nr = info.number;
	if (nr >= mddev->sb->nr_disks)
		return -EINVAL;

	SET_FROM_SB(major);
	SET_FROM_SB(minor);
	SET_FROM_SB(raid_disk);
	SET_FROM_SB(state);

	if (md_copy_to_user(arg, &info, sizeof(info)))
		return -EFAULT;

	return 0;
}
#undef SET_FROM_SB

#define SET_SB(x) mddev->sb->disks[nr].x = info->x

static int add_new_disk (mddev_t * mddev, mdu_disk_info_t *info)
{
	int err, size, persistent;
	mdk_rdev_t *rdev;
	unsigned int nr;
	kdev_t dev;
	dev = MKDEV(info->major,info->minor);

	if (find_rdev_all(dev)) {
		printk("device %s already used in a RAID array!\n",
				partition_name(dev));
		return -EBUSY;
	}
	if (!mddev->sb) {
		/* expecting a device which has a superblock */
		err = md_import_device(dev, 1);
		if (err) {
			printk("md error, md_import_device returned %d\n", err);
			return -EINVAL;
		}
		rdev = find_rdev_all(dev);
		if (!rdev) {
			MD_BUG();
			return -EINVAL;
		}
		if (mddev->nb_dev) {
			mdk_rdev_t *rdev0 = md_list_entry(mddev->disks.next,
							  mdk_rdev_t, same_set);
			if (!uuid_equal(rdev0, rdev)) {
				printk("md: %s has different UUID to %s\n", partition_name(rdev->dev), partition_name(rdev0->dev));
				export_rdev(rdev);
				return -EINVAL;
			}
			if (!sb_equal(rdev0->sb, rdev->sb)) {
				printk("md: %s has same UUID but different superblock to %s\n", partition_name(rdev->dev), partition_name(rdev0->dev));
				export_rdev(rdev);
				return -EINVAL;
			}
		}
		bind_rdev_to_array(rdev, mddev);
		return 0;
	}

	nr = info->number;
	if (nr >= mddev->sb->nr_disks)
		return -EINVAL;

	SET_SB(number);
	SET_SB(major);
	SET_SB(minor);
	SET_SB(raid_disk);
	SET_SB(state);

	if ((info->state & (1<<MD_DISK_FAULTY))==0) {
		err = md_import_device (dev, 0);
		if (err) {
			printk("md: error, md_import_device() returned %d\n", err);
			return -EINVAL;
		}
		rdev = find_rdev_all(dev);
		if (!rdev) {
			MD_BUG();
			return -EINVAL;
		}

		rdev->old_dev = dev;
		rdev->desc_nr = info->number;

		bind_rdev_to_array(rdev, mddev);

		persistent = !mddev->sb->not_persistent;
		if (!persistent)
			printk("nonpersistent superblock ...\n");
		if (!mddev->sb->chunk_size)
			printk("no chunksize?\n");

		size = calc_dev_size(dev, mddev, persistent);
		rdev->sb_offset = calc_dev_sboffset(dev, mddev, persistent);

		if (!mddev->sb->size || (mddev->sb->size > size))
			mddev->sb->size = size;
	}

	/*
	 * sync all other superblocks with the main superblock
	 */
	sync_sbs(mddev);

	return 0;
}
#undef SET_SB

static int hot_remove_disk (mddev_t * mddev, kdev_t dev)
{
	int err;
	mdk_rdev_t *rdev;
	mdp_disk_t *disk;

	if (!mddev->pers)
		return -ENODEV;

	printk("trying to remove %s from md%d ... \n",
		partition_name(dev), mdidx(mddev));

	if (!mddev->pers->diskop) {
		printk("md%d: personality does not support diskops!\n",
								 mdidx(mddev));
		return -EINVAL;
	}

	rdev = find_rdev(mddev, dev);
	if (!rdev)
		return -ENXIO;

	if (rdev->desc_nr == -1) {
		MD_BUG();
		return -EINVAL;
	}
	disk = &mddev->sb->disks[rdev->desc_nr];
	if (disk_active(disk))
		goto busy;
	if (disk_removed(disk)) {
		MD_BUG();
		return -EINVAL;
	}
	
	err = mddev->pers->diskop(mddev, &disk, DISKOP_HOT_REMOVE_DISK);
	if (err == -EBUSY)
		goto busy;
	if (err) {
		MD_BUG();
		return -EINVAL;
	}

	remove_descriptor(disk, mddev->sb);
	kick_rdev_from_array(rdev);
	mddev->sb_dirty = 1;
	md_update_sb(mddev);

	return 0;
busy:
	printk("cannot remove active disk %s from md%d ... \n",
		partition_name(dev), mdidx(mddev));
	return -EBUSY;
}

static int hot_add_disk (mddev_t * mddev, kdev_t dev)
{
	int i, err, persistent;
	unsigned int size;
	mdk_rdev_t *rdev;
	mdp_disk_t *disk;

	if (!mddev->pers)
		return -ENODEV;

	printk("trying to hot-add %s to md%d ... \n",
		partition_name(dev), mdidx(mddev));

	if (!mddev->pers->diskop) {
		printk("md%d: personality does not support diskops!\n",
								 mdidx(mddev));
		return -EINVAL;
	}

	persistent = !mddev->sb->not_persistent;
	size = calc_dev_size(dev, mddev, persistent);

	if (size < mddev->sb->size) {
		printk("md%d: disk size %d blocks < array size %d\n",
				mdidx(mddev), size, mddev->sb->size);
		return -ENOSPC;
	}

	rdev = find_rdev(mddev, dev);
	if (rdev)
		return -EBUSY;

	err = md_import_device (dev, 0);
	if (err) {
		printk("md: error, md_import_device() returned %d\n", err);
		return -EINVAL;
	}
	rdev = find_rdev_all(dev);
	if (!rdev) {
		MD_BUG();
		return -EINVAL;
	}
	if (rdev->faulty) {
		printk("md: can not hot-add faulty %s disk to md%d!\n",
				partition_name(dev), mdidx(mddev));
		err = -EINVAL;
		goto abort_export;
	}
	bind_rdev_to_array(rdev, mddev);

	/*
	 * The rest should better be atomic, we can have disk failures
	 * noticed in interrupt contexts ...
	 */
	rdev->old_dev = dev;
	rdev->size = size;
	rdev->sb_offset = calc_dev_sboffset(dev, mddev, persistent);

	disk = mddev->sb->disks + mddev->sb->raid_disks;
	for (i = mddev->sb->raid_disks; i < MD_SB_DISKS; i++) {
		disk = mddev->sb->disks + i;

		if (!disk->major && !disk->minor)
			break;
		if (disk_removed(disk))
			break;
	}
	if (i == MD_SB_DISKS) {
		printk("md%d: can not hot-add to full array!\n", mdidx(mddev));
		err = -EBUSY;
		goto abort_unbind_export;
	}

	if (disk_removed(disk)) {
		/*
		 * reuse slot
		 */
		if (disk->number != i) {
			MD_BUG();
			err = -EINVAL;
			goto abort_unbind_export;
		}
	} else {
		disk->number = i;
	}

	disk->raid_disk = disk->number;
	disk->major = MAJOR(dev);
	disk->minor = MINOR(dev);

	if (mddev->pers->diskop(mddev, &disk, DISKOP_HOT_ADD_DISK)) {
		MD_BUG();
		err = -EINVAL;
		goto abort_unbind_export;
	}

	mark_disk_spare(disk);
	mddev->sb->nr_disks++;
	mddev->sb->spare_disks++;
	mddev->sb->working_disks++;

	mddev->sb_dirty = 1;

	md_update_sb(mddev);

	/*
	 * Kick recovery, maybe this spare has to be added to the
	 * array immediately.
	 */
	md_recover_arrays();

	return 0;

abort_unbind_export:
	unbind_rdev_from_array(rdev);

abort_export:
	export_rdev(rdev);
	return err;
}

#define SET_SB(x) mddev->sb->x = info->x
static int set_array_info (mddev_t * mddev, mdu_array_info_t *info)
{

	if (alloc_array_sb(mddev))
		return -ENOMEM;

	mddev->sb->major_version = MD_MAJOR_VERSION;
	mddev->sb->minor_version = MD_MINOR_VERSION;
	mddev->sb->patch_version = MD_PATCHLEVEL_VERSION;
	mddev->sb->ctime = CURRENT_TIME;

	SET_SB(level);