Driver core: convert block from raw kobjects to core devices

This moves the block devices to /sys/class/block. It will create a
flat list of all block devices, with the disks and partitions in one
directory. For compatibility /sys/block is created and contains symlinks
to the disks.

  /sys/class/block
  |-- sda -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda
  |-- sda1 -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda1
  |-- sda10 -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda10
  |-- sda5 -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda5
  |-- sda6 -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda6
  |-- sda7 -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda7
  |-- sda8 -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda8
  |-- sda9 -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda9
  `-- sr0 -> ../../devices/pci0000:00/0000:00:1f.2/host1/target1:0:0/1:0:0:0/block/sr0

  /sys/block/
  |-- sda -> ../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda
  `-- sr0 -> ../devices/pci0000:00/0000:00:1f.2/host1/target1:0:0/1:0:0:0/block/sr0
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

block/ll_rw_blk.c

@@ -4182,7 +4182,7 @@ int blk_register_queue(struct gendisk *disk)
 	if (!q || !q->request_fn)
 		return -ENXIO;
 
-	q->kobj.parent = kobject_get(&disk->kobj);
+	q->kobj.parent = kobject_get(&disk->dev.kobj);
 
 	ret = kobject_add(&q->kobj);
 	if (ret < 0)
@@ -4209,6 +4209,6 @@ void blk_unregister_queue(struct gendisk *disk)
 
 		kobject_uevent(&q->kobj, KOBJ_REMOVE);
 		kobject_del(&q->kobj);
-		kobject_put(&disk->kobj);
+		kobject_put(&disk->dev.kobj);
 	}
 }

drivers/base/class.c

@@ -17,6 +17,7 @@
 #include <linux/kdev_t.h>
 #include <linux/err.h>
 #include <linux/slab.h>
+#include <linux/genhd.h>
 #include "base.h"
 
 #define to_class_attr(_attr) container_of(_attr, struct class_attribute, attr)
@@ -149,7 +150,13 @@ int class_register(struct class * cls)
 	if (error)
 		return error;
 
+#ifdef CONFIG_SYSFS_DEPRECATED
+	/* let the block class directory show up in the root of sysfs */
+	if (cls != &block_class)
+		cls->subsys.kobj.kset = class_kset;
+#else
 	cls->subsys.kobj.kset = class_kset;
+#endif
 	cls->subsys.kobj.ktype = &class_ktype;
 
 	error = kset_register(&cls->subsys);

drivers/base/core.c

@@ -671,14 +671,15 @@ static int device_add_class_symlinks(struct device *dev)
 
 #ifdef CONFIG_SYSFS_DEPRECATED
 	/* stacked class devices need a symlink in the class directory */
-	if (dev->kobj.parent != &dev->class->subsys.kobj) {
+	if (dev->kobj.parent != &dev->class->subsys.kobj &&
+	    dev->type != &part_type) {
 		error = sysfs_create_link(&dev->class->subsys.kobj, &dev->kobj,
 					  dev->bus_id);
 		if (error)
 			goto out_subsys;
 	}
 
-	if (dev->parent) {
+	if (dev->parent && dev->type != &part_type) {
 		struct device *parent = dev->parent;
 		char *class_name;
 
@@ -707,10 +708,11 @@ static int device_add_class_symlinks(struct device *dev)
 	return 0;
 
 out_device:
-	if (dev->parent)
+	if (dev->parent && dev->type != &part_type)
 		sysfs_remove_link(&dev->kobj, "device");
 out_busid:
-	if (dev->kobj.parent != &dev->class->subsys.kobj)
+	if (dev->kobj.parent != &dev->class->subsys.kobj &&
+	    dev->type != &part_type)
 		sysfs_remove_link(&dev->class->subsys.kobj, dev->bus_id);
 #else
 	/* link in the class directory pointing to the device */
@@ -719,7 +721,7 @@ static int device_add_class_symlinks(struct device *dev)
 	if (error)
 		goto out_subsys;
 
-	if (dev->parent) {
+	if (dev->parent && dev->type != &part_type) {
 		error = sysfs_create_link(&dev->kobj, &dev->parent->kobj,
 					  "device");
 		if (error)
@@ -743,7 +745,7 @@ static void device_remove_class_symlinks(struct device *dev)
 		return;
 
 #ifdef CONFIG_SYSFS_DEPRECATED
-	if (dev->parent) {
+	if (dev->parent && dev->type != &part_type) {
 		char *class_name;
 
 		class_name = make_class_name(dev->class->name, &dev->kobj);
@@ -754,10 +756,11 @@ static void device_remove_class_symlinks(struct device *dev)
 		sysfs_remove_link(&dev->kobj, "device");
 	}
 
-	if (dev->kobj.parent != &dev->class->subsys.kobj)
+	if (dev->kobj.parent != &dev->class->subsys.kobj &&
+	    dev->type != &part_type)
 		sysfs_remove_link(&dev->class->subsys.kobj, dev->bus_id);
 #else
-	if (dev->parent)
+	if (dev->parent && dev->type != &part_type)
 		sysfs_remove_link(&dev->kobj, "device");
 
 	sysfs_remove_link(&dev->class->subsys.kobj, dev->bus_id);
@@ -925,6 +928,7 @@ struct device * get_device(struct device * dev)
  */
 void put_device(struct device * dev)
 {
+	/* might_sleep(); */
 	if (dev)
 		kobject_put(&dev->kobj);
 }

drivers/block/aoe/aoeblk.c

@@ -15,8 +15,10 @@
 
 static struct kmem_cache *buf_pool_cache;
 
-static ssize_t aoedisk_show_state(struct gendisk * disk, char *page)
+static ssize_t aoedisk_show_state(struct device *dev,
+				  struct device_attribute *attr, char *page)
 {
+	struct gendisk *disk = dev_to_disk(dev);
 	struct aoedev *d = disk->private_data;
 
 	return snprintf(page, PAGE_SIZE,
@@ -26,50 +28,47 @@ static ssize_t aoedisk_show_state(struct gendisk * disk, char *page)
 			(d->nopen && !(d->flags & DEVFL_UP)) ? ",closewait" : "");
 	/* I'd rather see nopen exported so we can ditch closewait */
 }
-static ssize_t aoedisk_show_mac(struct gendisk * disk, char *page)
+static ssize_t aoedisk_show_mac(struct device *dev,
+				struct device_attribute *attr, char *page)
 {
+	struct gendisk *disk = dev_to_disk(dev);
 	struct aoedev *d = disk->private_data;
 
 	return snprintf(page, PAGE_SIZE, "%012llx\n",
 			(unsigned long long)mac_addr(d->addr));
 }
-static ssize_t aoedisk_show_netif(struct gendisk * disk, char *page)
+static ssize_t aoedisk_show_netif(struct device *dev,
+				  struct device_attribute *attr, char *page)
 {
+	struct gendisk *disk = dev_to_disk(dev);
 	struct aoedev *d = disk->private_data;
 
 	return snprintf(page, PAGE_SIZE, "%s\n", d->ifp->name);
 }
 /* firmware version */
-static ssize_t aoedisk_show_fwver(struct gendisk * disk, char *page)
+static ssize_t aoedisk_show_fwver(struct device *dev,
+				  struct device_attribute *attr, char *page)
 {
+	struct gendisk *disk = dev_to_disk(dev);
 	struct aoedev *d = disk->private_data;
 
 	return snprintf(page, PAGE_SIZE, "0x%04x\n", (unsigned int) d->fw_ver);
 }
 
-static struct disk_attribute disk_attr_state = {
-	.attr = {.name = "state", .mode = S_IRUGO },
-	.show = aoedisk_show_state
-};
-static struct disk_attribute disk_attr_mac = {
-	.attr = {.name = "mac", .mode = S_IRUGO },
-	.show = aoedisk_show_mac
-};
-static struct disk_attribute disk_attr_netif = {
-	.attr = {.name = "netif", .mode = S_IRUGO },
-	.show = aoedisk_show_netif
-};
-static struct disk_attribute disk_attr_fwver = {
-	.attr = {.name = "firmware-version", .mode = S_IRUGO },
-	.show = aoedisk_show_fwver
+static DEVICE_ATTR(state, S_IRUGO, aoedisk_show_state, NULL);
+static DEVICE_ATTR(mac, S_IRUGO, aoedisk_show_mac, NULL);
+static DEVICE_ATTR(netif, S_IRUGO, aoedisk_show_netif, NULL);
+static struct device_attribute dev_attr_firmware_version = {
+	.attr = { .name = "firmware-version", .mode = S_IRUGO, .owner = THIS_MODULE },
+	.show = aoedisk_show_fwver,
 };
 
 static struct attribute *aoe_attrs[] = {
-	&disk_attr_state.attr,
-	&disk_attr_mac.attr,
-	&disk_attr_netif.attr,
-	&disk_attr_fwver.attr,
-	NULL
+	&dev_attr_state.attr,
+	&dev_attr_mac.attr,
+	&dev_attr_netif.attr,
+	&dev_attr_firmware_version.attr,
+	NULL,
 };
 
 static const struct attribute_group attr_group = {
@@ -79,12 +78,12 @@ static const struct attribute_group attr_group = {
 static int
 aoedisk_add_sysfs(struct aoedev *d)
 {
-	return sysfs_create_group(&d->gd->kobj, &attr_group);
+	return sysfs_create_group(&d->gd->dev.kobj, &attr_group);
 }
 void
 aoedisk_rm_sysfs(struct aoedev *d)
 {
-	sysfs_remove_group(&d->gd->kobj, &attr_group);
+	sysfs_remove_group(&d->gd->dev.kobj, &attr_group);
 }
 
 static int

drivers/block/nbd.c

@@ -375,14 +375,17 @@ static struct request *nbd_read_stat(struct nbd_device *lo)
 	return NULL;
 }
 
-static ssize_t pid_show(struct gendisk *disk, char *page)
+static ssize_t pid_show(struct device *dev,
+			struct device_attribute *attr, char *buf)
 {
-	return sprintf(page, "%ld\n",
+	struct gendisk *disk = dev_to_disk(dev);
+
+	return sprintf(buf, "%ld\n",
 		(long) ((struct nbd_device *)disk->private_data)->pid);
 }
 
-static struct disk_attribute pid_attr = {
-	.attr = { .name = "pid", .mode = S_IRUGO },
+static struct device_attribute pid_attr = {
+	.attr = { .name = "pid", .mode = S_IRUGO, .owner = THIS_MODULE },
 	.show = pid_show,
 };
 
@@ -394,7 +397,7 @@ static int nbd_do_it(struct nbd_device *lo)
 	BUG_ON(lo->magic != LO_MAGIC);
 
 	lo->pid = current->pid;
-	ret = sysfs_create_file(&lo->disk->kobj, &pid_attr.attr);
+	ret = sysfs_create_file(&lo->disk->dev.kobj, &pid_attr.attr);
 	if (ret) {
 		printk(KERN_ERR "nbd: sysfs_create_file failed!");
 		return ret;
@@ -403,7 +406,7 @@ static int nbd_do_it(struct nbd_device *lo)
 	while ((req = nbd_read_stat(lo)) != NULL)
 		nbd_end_request(req);
 
-	sysfs_remove_file(&lo->disk->kobj, &pid_attr.attr);
+	sysfs_remove_file(&lo->disk->dev.kobj, &pid_attr.attr);
 	return 0;
 }
 

drivers/ide/ide-probe.c

@@ -1173,7 +1173,7 @@ static struct kobject *exact_match(dev_t dev, int *part, void *data)
 {
 	struct gendisk *p = data;
 	*part &= (1 << PARTN_BITS) - 1;
-	return &p->kobj;
+	return &p->dev.kobj;
 }
 
 static int exact_lock(dev_t dev, void *data)

drivers/md/dm.c

@@ -1109,7 +1109,7 @@ static void event_callback(void *context)
 	list_splice_init(&md->uevent_list, &uevents);
 	spin_unlock_irqrestore(&md->uevent_lock, flags);
 
-	dm_send_uevents(&uevents, &md->disk->kobj);
+	dm_send_uevents(&uevents, &md->disk->dev.kobj);
 
 	atomic_inc(&md->event_nr);
 	wake_up(&md->eventq);
@@ -1530,7 +1530,7 @@ int dm_resume(struct mapped_device *md)
  *---------------------------------------------------------------*/
 void dm_kobject_uevent(struct mapped_device *md)
 {
-	kobject_uevent(&md->disk->kobj, KOBJ_CHANGE);
+	kobject_uevent(&md->disk->dev.kobj, KOBJ_CHANGE);
 }
 
 uint32_t dm_next_uevent_seq(struct mapped_device *md)

drivers/md/md.c

@@ -1396,9 +1396,9 @@ static int bind_rdev_to_array(mdk_rdev_t * rdev, mddev_t * mddev)
 		goto fail;
 
 	if (rdev->bdev->bd_part)
-		ko = &rdev->bdev->bd_part->kobj;
+		ko = &rdev->bdev->bd_part->dev.kobj;
 	else
-		ko = &rdev->bdev->bd_disk->kobj;
+		ko = &rdev->bdev->bd_disk->dev.kobj;
 	if ((err = sysfs_create_link(&rdev->kobj, ko, "block"))) {
 		kobject_del(&rdev->kobj);
 		goto fail;
@@ -3083,7 +3083,7 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data)
 	add_disk(disk);
 	mddev->gendisk = disk;
 	mutex_unlock(&disks_mutex);
-	error = kobject_init_and_add(&mddev->kobj, &md_ktype, &disk->kobj,
+	error = kobject_init_and_add(&mddev->kobj, &md_ktype, &disk->dev.kobj,
 				     "%s", "md");
 	if (error)
 		printk(KERN_WARNING "md: cannot register %s/md - name in use\n",
@@ -3361,7 +3361,7 @@ static int do_md_run(mddev_t * mddev)
 
 	mddev->changed = 1;
 	md_new_event(mddev);
-	kobject_uevent(&mddev->gendisk->kobj, KOBJ_CHANGE);
+	kobject_uevent(&mddev->gendisk->dev.kobj, KOBJ_CHANGE);
 	return 0;
 }
 

fs/block_dev.c

@@ -738,9 +738,9 @@ EXPORT_SYMBOL(bd_release);
 static struct kobject *bdev_get_kobj(struct block_device *bdev)
 {
 	if (bdev->bd_contains != bdev)
-		return kobject_get(&bdev->bd_part->kobj);
+		return kobject_get(&bdev->bd_part->dev.kobj);
 	else
-		return kobject_get(&bdev->bd_disk->kobj);
+		return kobject_get(&bdev->bd_disk->dev.kobj);
 }
 
 static struct kobject *bdev_get_holder(struct block_device *bdev)
@@ -1176,7 +1176,7 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part)
 				ret = -ENXIO;
 				goto out_first;
 			}
-			kobject_get(&p->kobj);
+			kobject_get(&p->dev.kobj);
 			bdev->bd_part = p;
 			bd_set_size(bdev, (loff_t) p->nr_sects << 9);
 		}
@@ -1299,7 +1299,7 @@ static int __blkdev_put(struct block_device *bdev, int for_part)
 		module_put(owner);
 
 		if (bdev->bd_contains != bdev) {
-			kobject_put(&bdev->bd_part->kobj);
+			kobject_put(&bdev->bd_part->dev.kobj);
 			bdev->bd_part = NULL;
 		}
 		bdev->bd_disk = NULL;

include/linux/genhd.h

@@ -10,9 +10,19 @@
  */
 
 #include <linux/types.h>
+#include <linux/kdev_t.h>
 
 #ifdef CONFIG_BLOCK
 
+#define kobj_to_dev(k) container_of(k, struct device, kobj)
+#define dev_to_disk(device) container_of(device, struct gendisk, dev)
+#define dev_to_part(device) container_of(device, struct hd_struct, dev)
+
+extern struct device_type disk_type;
+extern struct device_type part_type;
+extern struct kobject *block_depr;
+extern struct class block_class;
+
 enum {
 /* These three have identical behaviour; use the second one if DOS FDISK gets
    confused about extended/logical partitions starting past cylinder 1023. */
@@ -84,7 +94,7 @@ struct partition {
 struct hd_struct {
 	sector_t start_sect;
 	sector_t nr_sects;
-	struct kobject kobj;
+	struct device dev;
 	struct kobject *holder_dir;
 	unsigned ios[2], sectors[2];	/* READs and WRITEs */
 	int policy, partno;
@@ -117,15 +127,14 @@ struct gendisk {
                                          * disks that can't be partitioned. */
 	char disk_name[32];		/* name of major driver */
 	struct hd_struct **part;	/* [indexed by minor] */
-	int part_uevent_suppress;
 	struct block_device_operations *fops;
 	struct request_queue *queue;
 	void *private_data;
 	sector_t capacity;
 
 	int flags;
-	struct device *driverfs_dev;
-	struct kobject kobj;
+	struct device *driverfs_dev;  // FIXME: remove
+	struct device dev;
 	struct kobject *holder_dir;
 	struct kobject *slave_dir;
 
@@ -143,13 +152,6 @@ struct gendisk {
 	struct work_struct async_notify;
 };
 
-/* Structure for sysfs attributes on block devices */
-struct disk_attribute {
-	struct attribute attr;
-	ssize_t (*show)(struct gendisk *, char *);
-	ssize_t (*store)(struct gendisk *, const char *, size_t);
-};
-
 /* 
  * Macros to operate on percpu disk statistics:
  *
@@ -411,7 +413,8 @@ struct unixware_disklabel {
 #define ADDPART_FLAG_RAID	1
 #define ADDPART_FLAG_WHOLEDISK	2
 
-char *disk_name (struct gendisk *hd, int part, char *buf);
+extern dev_t blk_lookup_devt(const char *name);
+extern char *disk_name (struct gendisk *hd, int part, char *buf);
 
 extern int rescan_partitions(struct gendisk *disk, struct block_device *bdev);
 extern void add_partition(struct gendisk *, int, sector_t, sector_t, int);
@@ -423,12 +426,12 @@ extern struct gendisk *alloc_disk(int minors);
 extern struct kobject *get_disk(struct gendisk *disk);
 extern void put_disk(struct gendisk *disk);
 extern void genhd_media_change_notify(struct gendisk *disk);
-extern void blk_register_region(dev_t dev, unsigned long range,
+extern void blk_register_region(dev_t devt, unsigned long range,
 			struct module *module,
 			struct kobject *(*probe)(dev_t, int *, void *),
 			int (*lock)(dev_t, void *),
 			void *data);
-extern void blk_unregister_region(dev_t dev, unsigned long range);
+extern void blk_unregister_region(dev_t devt, unsigned long range);
 
 static inline struct block_device *bdget_disk(struct gendisk *disk, int index)
 {
@@ -441,6 +444,12 @@ static inline struct block_device *bdget_disk(struct gendisk *disk, int index)
 
 static inline void printk_all_partitions(void) { }
 
+static inline dev_t blk_lookup_devt(const char *name)
+{
+	dev_t devt = MKDEV(0, 0);
+	return devt;
+}
+
 #endif /* CONFIG_BLOCK */
 
 #endif

init/do_mounts.c

@@ -55,69 +55,6 @@ static int __init readwrite(char *str)
 __setup("ro", readonly);
 __setup("rw", readwrite);
 
-static dev_t try_name(char *name, int part)
-{
-	char path[64];
-	char buf[32];
-	int range;
-	dev_t res;
-	char *s;
-	int len;
-	int fd;
-	unsigned int maj, min;
-
-	/* read device number from .../dev */
-
-	sprintf(path, "/sys/block/%s/dev", name);
-	fd = sys_open(path, 0, 0);
-	if (fd < 0)
-		goto fail;
-	len = sys_read(fd, buf, 32);
-	sys_close(fd);
-	if (len <= 0 || len == 32 || buf[len - 1] != '\n')
-		goto fail;
-	buf[len - 1] = '\0';
-	if (sscanf(buf, "%u:%u", &maj, &min) == 2) {
-		/*
-		 * Try the %u:%u format -- see print_dev_t()
-		 */
-		res = MKDEV(maj, min);
-		if (maj != MAJOR(res) || min != MINOR(res))
-			goto fail;
-	} else {
-		/*
-		 * Nope.  Try old-style "0321"
-		 */
-		res = new_decode_dev(simple_strtoul(buf, &s, 16));
-		if (*s)
-			goto fail;
-	}
-
-	/* if it's there and we are not looking for a partition - that's it */
-	if (!part)
-		return res;
-
-	/* otherwise read range from .../range */
-	sprintf(path, "/sys/block/%s/range", name);
-	fd = sys_open(path, 0, 0);
-	if (fd < 0)
-		goto fail;
-	len = sys_read(fd, buf, 32);
-	sys_close(fd);
-	if (len <= 0 || len == 32 || buf[len - 1] != '\n')
-		goto fail;
-	buf[len - 1] = '\0';
-	range = simple_strtoul(buf, &s, 10);
-	if (*s)
-		goto fail;
-
-	/* if partition is within range - we got it */
-	if (part < range)
-		return res + part;
-fail:
-	return 0;
-}
-
 /*
  *	Convert a name into device number.  We accept the following variants:
  *
@@ -129,12 +66,10 @@ static dev_t try_name(char *name, int part)
  *	5) /dev/<disk_name>p<decimal> - same as the above, that form is
  *	   used when disk name of partitioned disk ends on a digit.
  *
- *	If name doesn't have fall into the categories above, we return 0.
- *	Sysfs is used to check if something is a disk name - it has
- *	all known disks under bus/block/devices.  If the disk name
- *	contains slashes, name of sysfs node has them replaced with
- *	bangs.  try_name() does the actual checks, assuming that sysfs
- *	is mounted on rootfs /sys.
+ *	If name doesn't have fall into the categories above, we return (0,0).
+ *	block_class is used to check if something is a disk name. If the disk
+ *	name contains slashes, the device name has them replaced with
+ *	bangs.
  */
 
 dev_t name_to_dev_t(char *name)
@@ -142,13 +77,6 @@ dev_t name_to_dev_t(char *name)
 	char s[32];
 	char *p;
 	dev_t res = 0;
-	int part;
-
-#ifdef CONFIG_SYSFS
-	int mkdir_err = sys_mkdir("/sys", 0700);
-	if (sys_mount("sysfs", "/sys", "sysfs", 0, NULL) < 0)
-		goto out;
-#endif
 
 	if (strncmp(name, "/dev/", 5) != 0) {
 		unsigned maj, min;
@@ -164,6 +92,7 @@ dev_t name_to_dev_t(char *name)
 		}
 		goto done;
 	}
+
 	name += 5;
 	res = Root_NFS;
 	if (strcmp(name, "nfs") == 0)
@@ -178,35 +107,14 @@ dev_t name_to_dev_t(char *name)
 	for (p = s; *p; p++)
 		if (*p == '/')
 			*p = '!';
-	res = try_name(s, 0);
+	res = blk_lookup_devt(s);
 	if (res)
 		goto done;
 
-	while (p > s && isdigit(p[-1]))
-		p--;
-	if (p == s || !*p || *p == '0')
-		goto fail;
-	part = simple_strtoul(p, NULL, 10);
-	*p = '\0';
-	res = try_name(s, part);
-	if (res)
-		goto done;
-
-	if (p < s + 2 || !isdigit(p[-2]) || p[-1] != 'p')
-		goto fail;
-	p[-1] = '\0';
-	res = try_name(s, part);
+fail:
+	return 0;
 done:
-#ifdef CONFIG_SYSFS
-	sys_umount("/sys", 0);
-out:
-	if (!mkdir_err)
-		sys_rmdir("/sys");
-#endif
 	return res;
-fail:
-	res = 0;
-	goto done;
 }
 
 static int __init root_dev_setup(char *line)