Following on from the header file for the v4 ioctl interface that I
posted a couple of weeks ago, here is the first cut at the
implementation (3 patches posted as a follow up to this mail). I hope
the v1 interface can be retired before 2.6. Tools are not yet
available to drive this, but should be later this week.
Thoughts ?
- Joe
Fix memory leak.
--- diff/drivers/md/dm-ioctl.c 2003-06-30 10:07:33.000000000 +0100
+++ source/drivers/md/dm-ioctl.c 2003-06-30 11:39:38.000000000 +0100
@@ -238,6 +238,7 @@
list_del(&hc->name_list);
unregister_with_devfs(hc);
dm_put(hc->md);
+ free_cell(hc);
}
void dm_hash_remove_all(void)
v4 of the ioctl interface. Note there never was a v2 or a v3 except
in an unofficial EVMS way.
There is now a config option to allow the user to select v1 or v4, it
defaults to v1. If you want to use v4 you will have to update your
tools (libdevmapper/lvm), these new tools have not been released at
the time of writing.
dm-ioctl.[hc] just #includes dm-ioctl-v[14].[hc] depending on the
config option.
--- diff/drivers/md/Kconfig 2003-05-21 11:50:15.000000000 +0100
+++ source/drivers/md/Kconfig 2003-07-01 15:36:42.000000000 +0100
@@ -145,5 +145,12 @@
If unsure, say N.
+config DM_IOCTL_V4
+ bool "ioctl interface version 4"
+ depends on BLK_DEV_DM
+ ---help---
+ Recent tools use a new version of the ioctl interface, only
+ select this option if you intend using such tools.
+
endmenu
--- diff/drivers/md/dm-ioctl.c 2003-06-30 17:12:49.000000000 +0100
+++ source/drivers/md/dm-ioctl.c 2003-07-01 15:36:42.000000000 +0100
@@ -1,1133 +1,13 @@
/*
- * Copyright (C) 2001, 2002 Sistina Software (UK) Limited.
+ * Copyright (C) 2003 Sistina Software (UK) Limited.
*
* This file is released under the GPL.
*/
-#include "dm.h"
-
-#include <linux/module.h>
-#include <linux/vmalloc.h>
-#include <linux/miscdevice.h>
#include <linux/dm-ioctl.h>
-#include <linux/init.h>
-#include <linux/wait.h>
-#include <linux/blk.h>
-#include <linux/slab.h>
-#include <linux/devfs_fs_kernel.h>
-
-#include <asm/uaccess.h>
-
-#define DM_DRIVER_EMAIL "[email protected]"
-
-/*-----------------------------------------------------------------
- * The ioctl interface needs to be able to look up devices by
- * name or uuid.
- *---------------------------------------------------------------*/
-struct hash_cell {
- struct list_head name_list;
- struct list_head uuid_list;
-
- char *name;
- char *uuid;
- struct mapped_device *md;
-};
-
-#define NUM_BUCKETS 64
-#define MASK_BUCKETS (NUM_BUCKETS - 1)
-static struct list_head _name_buckets[NUM_BUCKETS];
-static struct list_head _uuid_buckets[NUM_BUCKETS];
-
-void dm_hash_remove_all(void);
-
-/*
- * Guards access to all three tables.
- */
-static DECLARE_RWSEM(_hash_lock);
-
-static void init_buckets(struct list_head *buckets)
-{
- unsigned int i;
-
- for (i = 0; i < NUM_BUCKETS; i++)
- INIT_LIST_HEAD(buckets + i);
-}
-
-int dm_hash_init(void)
-{
- init_buckets(_name_buckets);
- init_buckets(_uuid_buckets);
- devfs_mk_dir(DM_DIR);
- return 0;
-}
-
-void dm_hash_exit(void)
-{
- dm_hash_remove_all();
- devfs_remove(DM_DIR);
-}
-
-/*-----------------------------------------------------------------
- * Hash function:
- * We're not really concerned with the str hash function being
- * fast since it's only used by the ioctl interface.
- *---------------------------------------------------------------*/
-static unsigned int hash_str(const char *str)
-{
- const unsigned int hash_mult = 2654435387U;
- unsigned int h = 0;
-
- while (*str)
- h = (h + (unsigned int) *str++) * hash_mult;
-
- return h & MASK_BUCKETS;
-}
-
-/*-----------------------------------------------------------------
- * Code for looking up a device by name
- *---------------------------------------------------------------*/
-static struct hash_cell *__get_name_cell(const char *str)
-{
- struct list_head *tmp;
- struct hash_cell *hc;
- unsigned int h = hash_str(str);
-
- list_for_each (tmp, _name_buckets + h) {
- hc = list_entry(tmp, struct hash_cell, name_list);
- if (!strcmp(hc->name, str))
- return hc;
- }
-
- return NULL;
-}
-
-static struct hash_cell *__get_uuid_cell(const char *str)
-{
- struct list_head *tmp;
- struct hash_cell *hc;
- unsigned int h = hash_str(str);
-
- list_for_each (tmp, _uuid_buckets + h) {
- hc = list_entry(tmp, struct hash_cell, uuid_list);
- if (!strcmp(hc->uuid, str))
- return hc;
- }
-
- return NULL;
-}
-
-/*-----------------------------------------------------------------
- * Inserting, removing and renaming a device.
- *---------------------------------------------------------------*/
-static inline char *kstrdup(const char *str)
-{
- char *r = kmalloc(strlen(str) + 1, GFP_KERNEL);
- if (r)
- strcpy(r, str);
- return r;
-}
-
-static struct hash_cell *alloc_cell(const char *name, const char *uuid,
- struct mapped_device *md)
-{
- struct hash_cell *hc;
-
- hc = kmalloc(sizeof(*hc), GFP_KERNEL);
- if (!hc)
- return NULL;
-
- hc->name = kstrdup(name);
- if (!hc->name) {
- kfree(hc);
- return NULL;
- }
-
- if (!uuid)
- hc->uuid = NULL;
-
- else {
- hc->uuid = kstrdup(uuid);
- if (!hc->uuid) {
- kfree(hc->name);
- kfree(hc);
- return NULL;
- }
- }
-
- INIT_LIST_HEAD(&hc->name_list);
- INIT_LIST_HEAD(&hc->uuid_list);
- hc->md = md;
- return hc;
-}
-
-static void free_cell(struct hash_cell *hc)
-{
- if (hc) {
- kfree(hc->name);
- kfree(hc->uuid);
- kfree(hc);
- }
-}
-
-/*
- * devfs stuff.
- */
-static int register_with_devfs(struct hash_cell *hc)
-{
- struct gendisk *disk = dm_disk(hc->md);
-
- devfs_mk_bdev(MKDEV(disk->major, disk->first_minor),
- S_IFBLK | S_IRUSR | S_IWUSR | S_IRGRP,
- DM_DIR "/%s", hc->name);
- return 0;
-}
-
-static int unregister_with_devfs(struct hash_cell *hc)
-{
- devfs_remove(DM_DIR"/%s", hc->name);
- return 0;
-}
-
-/*
- * The kdev_t and uuid of a device can never change once it is
- * initially inserted.
- */
-int dm_hash_insert(const char *name, const char *uuid, struct mapped_device *md)
-{
- struct hash_cell *cell;
-
- /*
- * Allocate the new cells.
- */
- cell = alloc_cell(name, uuid, md);
- if (!cell)
- return -ENOMEM;
-
- /*
- * Insert the cell into all three hash tables.
- */
- down_write(&_hash_lock);
- if (__get_name_cell(name))
- goto bad;
-
- list_add(&cell->name_list, _name_buckets + hash_str(name));
-
- if (uuid) {
- if (__get_uuid_cell(uuid)) {
- list_del(&cell->name_list);
- goto bad;
- }
- list_add(&cell->uuid_list, _uuid_buckets + hash_str(uuid));
- }
- register_with_devfs(cell);
- dm_get(md);
- up_write(&_hash_lock);
-
- return 0;
-
- bad:
- up_write(&_hash_lock);
- free_cell(cell);
- return -EBUSY;
-}
-
-void __hash_remove(struct hash_cell *hc)
-{
- /* remove from the dev hash */
- list_del(&hc->uuid_list);
- list_del(&hc->name_list);
- unregister_with_devfs(hc);
- dm_put(hc->md);
- free_cell(hc);
-}
-
-void dm_hash_remove_all(void)
-{
- int i;
- struct hash_cell *hc;
- struct list_head *tmp, *n;
-
- down_write(&_hash_lock);
- for (i = 0; i < NUM_BUCKETS; i++) {
- list_for_each_safe (tmp, n, _name_buckets + i) {
- hc = list_entry(tmp, struct hash_cell, name_list);
- __hash_remove(hc);
- }
- }
- up_write(&_hash_lock);
-}
-
-int dm_hash_rename(const char *old, const char *new)
-{
- char *new_name, *old_name;
- struct hash_cell *hc;
-
- /*
- * duplicate new.
- */
- new_name = kstrdup(new);
- if (!new_name)
- return -ENOMEM;
-
- down_write(&_hash_lock);
-
- /*
- * Is new free ?
- */
- hc = __get_name_cell(new);
- if (hc) {
- DMWARN("asked to rename to an already existing name %s -> %s",
- old, new);
- up_write(&_hash_lock);
- kfree(new_name);
- return -EBUSY;
- }
-
- /*
- * Is there such a device as 'old' ?
- */
- hc = __get_name_cell(old);
- if (!hc) {
- DMWARN("asked to rename a non existent device %s -> %s",
- old, new);
- up_write(&_hash_lock);
- kfree(new_name);
- return -ENXIO;
- }
-
- /*
- * rename and move the name cell.
- */
- unregister_with_devfs(hc);
-
- list_del(&hc->name_list);
- old_name = hc->name;
- hc->name = new_name;
- list_add(&hc->name_list, _name_buckets + hash_str(new_name));
-
- /* rename the device node in devfs */
- register_with_devfs(hc);
-
- up_write(&_hash_lock);
- kfree(old_name);
- return 0;
-}
-
-
-/*-----------------------------------------------------------------
- * Implementation of the ioctl commands
- *---------------------------------------------------------------*/
-
-/*
- * All the ioctl commands get dispatched to functions with this
- * prototype.
- */
-typedef int (*ioctl_fn)(struct dm_ioctl *param, struct dm_ioctl *user);
-
-/*
- * Check a string doesn't overrun the chunk of
- * memory we copied from userland.
- */
-static int valid_str(char *str, void *begin, void *end)
-{
- while (((void *) str >= begin) && ((void *) str < end))
- if (!*str++)
- return 0;
-
- return -EINVAL;
-}
-
-static int next_target(struct dm_target_spec *last, uint32_t next,
- void *begin, void *end,
- struct dm_target_spec **spec, char **params)
-{
- *spec = (struct dm_target_spec *)
- ((unsigned char *) last + next);
- *params = (char *) (*spec + 1);
-
- if (*spec < (last + 1) || ((void *) *spec > end))
- return -EINVAL;
-
- return valid_str(*params, begin, end);
-}
-
-static int populate_table(struct dm_table *table, struct dm_ioctl *args)
-{
- int r, first = 1;
- unsigned int i = 0;
- struct dm_target_spec *spec;
- char *params;
- void *begin, *end;
-
- if (!args->target_count) {
- DMWARN("populate_table: no targets specified");
- return -EINVAL;
- }
-
- begin = (void *) args;
- end = begin + args->data_size;
-
- for (i = 0; i < args->target_count; i++) {
-
- if (first)
- r = next_target((struct dm_target_spec *) args,
- args->data_start,
- begin, end, &spec, ¶ms);
- else
- r = next_target(spec, spec->next, begin, end,
- &spec, ¶ms);
-
- if (r) {
- DMWARN("unable to find target");
- return -EINVAL;
- }
-
- r = dm_table_add_target(table, spec->target_type,
- (sector_t) spec->sector_start,
- (sector_t) spec->length,
- params);
- if (r) {
- DMWARN("internal error adding target to table");
- return -EINVAL;
- }
-
- first = 0;
- }
-
- return dm_table_complete(table);
-}
-
-/*
- * Round up the ptr to the next 'align' boundary. Obviously
- * 'align' must be a power of 2.
- */
-static inline void *align_ptr(void *ptr, unsigned int align)
-{
- align--;
- return (void *) (((unsigned long) (ptr + align)) & ~align);
-}
-
-/*
- * Copies a dm_ioctl and an optional additional payload to
- * userland.
- */
-static int results_to_user(struct dm_ioctl *user, struct dm_ioctl *param,
- void *data, uint32_t len)
-{
- int r;
- void *ptr = NULL;
-
- if (data) {
- ptr = align_ptr(user + 1, sizeof(unsigned long));
- param->data_start = ptr - (void *) user;
- }
-
- /*
- * The version number has already been filled in, so we
- * just copy later fields.
- */
- r = copy_to_user(&user->data_size, ¶m->data_size,
- sizeof(*param) - sizeof(param->version));
- if (r)
- return -EFAULT;
-
- if (data) {
- if (param->data_start + len > param->data_size)
- return -ENOSPC;
-
- if (copy_to_user(ptr, data, len))
- r = -EFAULT;
- }
-
- return r;
-}
-
-/*
- * Fills in a dm_ioctl structure, ready for sending back to
- * userland.
- */
-static int __info(struct mapped_device *md, struct dm_ioctl *param)
-{
- struct dm_table *table;
- struct block_device *bdev;
- struct gendisk *disk = dm_disk(md);
-
- param->flags = DM_EXISTS_FLAG;
- if (dm_suspended(md))
- param->flags |= DM_SUSPEND_FLAG;
-
- bdev = bdget_disk(disk, 0);
- if (!bdev)
- return -ENXIO;
-
- param->dev = bdev->bd_dev;
- param->open_count = bdev->bd_openers;
- bdput(bdev);
-
- if (disk->policy)
- param->flags |= DM_READONLY_FLAG;
-
- table = dm_get_table(md);
- param->target_count = dm_table_get_num_targets(table);
- dm_table_put(table);
-
- return 0;
-}
-
-/*
- * Always use UUID for lookups if it's present, otherwise use name.
- */
-static inline struct mapped_device *find_device(struct dm_ioctl *param)
-{
- struct hash_cell *hc;
- struct mapped_device *md = NULL;
-
- down_read(&_hash_lock);
- hc = *param->uuid ? __get_uuid_cell(param->uuid) :
- __get_name_cell(param->name);
- if (hc) {
- md = hc->md;
-
- /*
- * Sneakily write in both the name and the uuid
- * while we have the cell.
- */
- strlcpy(param->name, hc->name, sizeof(param->name));
- if (hc->uuid)
- strlcpy(param->uuid, hc->uuid, sizeof(param->uuid));
- else
- param->uuid[0] = '\0';
-
- dm_get(md);
- }
- up_read(&_hash_lock);
-
- return md;
-}
-
-#define ALIGNMENT sizeof(int)
-static void *_align(void *ptr, unsigned int a)
-{
- register unsigned long align = --a;
-
- return (void *) (((unsigned long) ptr + align) & ~align);
-}
-
-/*
- * Copies device info back to user space, used by
- * the create and info ioctls.
- */
-static int info(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- struct mapped_device *md;
-
- param->flags = 0;
-
- md = find_device(param);
- if (!md)
- /*
- * Device not found - returns cleared exists flag.
- */
- goto out;
-
- __info(md, param);
- dm_put(md);
-
- out:
- return results_to_user(user, param, NULL, 0);
-}
-
-static inline int get_mode(struct dm_ioctl *param)
-{
- int mode = FMODE_READ | FMODE_WRITE;
-
- if (param->flags & DM_READONLY_FLAG)
- mode = FMODE_READ;
-
- return mode;
-}
-
-static int check_name(const char *name)
-{
- if (name[0] == '/') {
- DMWARN("invalid device name");
- return -EINVAL;
- }
-
- return 0;
-}
-
-static int create(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- int r;
- struct dm_table *t;
- struct mapped_device *md;
-
- r = check_name(param->name);
- if (r)
- return r;
-
- r = dm_table_create(&t, get_mode(param));
- if (r)
- return r;
-
- r = populate_table(t, param);
- if (r) {
- dm_table_put(t);
- return r;
- }
-
- if (param->flags & DM_PERSISTENT_DEV_FLAG)
- r = dm_create_with_minor(minor(to_kdev_t(param->dev)), t, &md);
- else
- r = dm_create(t, &md);
-
- if (r) {
- dm_table_put(t);
- return r;
- }
- dm_table_put(t); /* md will have grabbed its own reference */
-
- set_disk_ro(dm_disk(md), (param->flags & DM_READONLY_FLAG) ? 1 : 0);
- r = dm_hash_insert(param->name, *param->uuid ? param->uuid : NULL, md);
- dm_put(md);
-
- return r ? r : info(param, user);
-}
-
-/*
- * Build up the status struct for each target
- */
-static int __status(struct mapped_device *md, struct dm_ioctl *param,
- char *outbuf, size_t *len)
-{
- unsigned int i, num_targets;
- struct dm_target_spec *spec;
- char *outptr;
- status_type_t type;
- struct dm_table *table = dm_get_table(md);
-
- if (param->flags & DM_STATUS_TABLE_FLAG)
- type = STATUSTYPE_TABLE;
- else
- type = STATUSTYPE_INFO;
-
- outptr = outbuf;
-
- /* Get all the target info */
- num_targets = dm_table_get_num_targets(table);
- for (i = 0; i < num_targets; i++) {
- struct dm_target *ti = dm_table_get_target(table, i);
-
- if (outptr - outbuf +
- sizeof(struct dm_target_spec) > param->data_size) {
- dm_table_put(table);
- return -ENOMEM;
- }
-
- spec = (struct dm_target_spec *) outptr;
-
- spec->status = 0;
- spec->sector_start = ti->begin;
- spec->length = ti->len;
- strlcpy(spec->target_type, ti->type->name,
- sizeof(spec->target_type));
-
- outptr += sizeof(struct dm_target_spec);
-
- /* Get the status/table string from the target driver */
- if (ti->type->status)
- ti->type->status(ti, type, outptr,
- outbuf + param->data_size - outptr);
- else
- outptr[0] = '\0';
-
- outptr += strlen(outptr) + 1;
- _align(outptr, ALIGNMENT);
- spec->next = outptr - outbuf;
- }
-
- param->target_count = num_targets;
- *len = outptr - outbuf;
- dm_table_put(table);
-
- return 0;
-}
-
-/*
- * Return the status of a device as a text string for each
- * target.
- */
-static int get_status(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- struct mapped_device *md;
- size_t len = 0;
- int ret;
- char *outbuf = NULL;
-
- md = find_device(param);
- if (!md)
- /*
- * Device not found - returns cleared exists flag.
- */
- goto out;
-
- /* We haven't a clue how long the resultant data will be so
- just allocate as much as userland has allowed us and make sure
- we don't overun it */
- outbuf = kmalloc(param->data_size, GFP_KERNEL);
- if (!outbuf)
- goto out;
- /*
- * Get the status of all targets
- */
- __status(md, param, outbuf, &len);
-
- /*
- * Setup the basic dm_ioctl structure.
- */
- __info(md, param);
-
- out:
- if (md)
- dm_put(md);
-
- ret = results_to_user(user, param, outbuf, len);
-
- if (outbuf)
- kfree(outbuf);
-
- return ret;
-}
-
-/*
- * Wait for a device to report an event
- */
-static int wait_device_event(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- struct mapped_device *md;
- DECLARE_WAITQUEUE(wq, current);
-
- md = find_device(param);
- if (!md)
- /*
- * Device not found - returns cleared exists flag.
- */
- goto out;
-
- /*
- * Setup the basic dm_ioctl structure.
- */
- __info(md, param);
-
- /*
- * Wait for a notification event
- */
- set_current_state(TASK_INTERRUPTIBLE);
- if (!dm_add_wait_queue(md, &wq, dm_get_event_nr(md))) {
- yield();
- dm_remove_wait_queue(md, &wq);
- }
- set_current_state(TASK_RUNNING);
- dm_put(md);
-
- out:
- return results_to_user(user, param, NULL, 0);
-}
-
-/*
- * Retrieves a list of devices used by a particular dm device.
- */
-static int dep(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- int r;
- unsigned int count;
- struct mapped_device *md;
- struct list_head *tmp;
- size_t len = 0;
- struct dm_target_deps *deps = NULL;
- struct dm_table *table;
-
- md = find_device(param);
- if (!md)
- goto out;
- table = dm_get_table(md);
-
- /*
- * Setup the basic dm_ioctl structure.
- */
- __info(md, param);
-
- /*
- * Count the devices.
- */
- count = 0;
- list_for_each(tmp, dm_table_get_devices(table))
- count++;
-
- /*
- * Allocate a kernel space version of the dm_target_status
- * struct.
- */
- if (array_too_big(sizeof(*deps), sizeof(*deps->dev), count)) {
- dm_table_put(table);
- dm_put(md);
- return -ENOMEM;
- }
-
- len = sizeof(*deps) + (sizeof(*deps->dev) * count);
- deps = kmalloc(len, GFP_KERNEL);
- if (!deps) {
- dm_table_put(table);
- dm_put(md);
- return -ENOMEM;
- }
-
- /*
- * Fill in the devices.
- */
- deps->count = count;
- count = 0;
- list_for_each(tmp, dm_table_get_devices(table)) {
- struct dm_dev *dd = list_entry(tmp, struct dm_dev, list);
- deps->dev[count++] = dd->bdev->bd_dev;
- }
- dm_table_put(table);
- dm_put(md);
-
- out:
- r = results_to_user(user, param, deps, len);
-
- kfree(deps);
- return r;
-}
-
-static int remove(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- struct hash_cell *hc;
-
- down_write(&_hash_lock);
- hc = *param->uuid ? __get_uuid_cell(param->uuid) :
- __get_name_cell(param->name);
- if (!hc) {
- DMWARN("device doesn't appear to be in the dev hash table.");
- up_write(&_hash_lock);
- return -EINVAL;
- }
-
- /*
- * You may ask the interface to drop its reference to an
- * in use device. This is no different to unlinking a
- * file that someone still has open. The device will not
- * actually be destroyed until the last opener closes it.
- * The name and uuid of the device (both are interface
- * properties) will be available for reuse immediately.
- *
- * You don't want to drop a _suspended_ device from the
- * interface, since that will leave you with no way of
- * resuming it.
- */
- if (dm_suspended(hc->md)) {
- DMWARN("refusing to remove a suspended device.");
- up_write(&_hash_lock);
- return -EPERM;
- }
-
- __hash_remove(hc);
- up_write(&_hash_lock);
- return 0;
-}
-
-static int remove_all(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- dm_hash_remove_all();
- return 0;
-}
-
-static int suspend(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- int r;
- struct mapped_device *md;
-
- md = find_device(param);
- if (!md)
- return -ENXIO;
-
- if (param->flags & DM_SUSPEND_FLAG)
- r = dm_suspend(md);
- else
- r = dm_resume(md);
-
- dm_put(md);
- return r;
-}
-
-static int reload(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- int r;
- struct mapped_device *md;
- struct dm_table *t;
-
- r = dm_table_create(&t, get_mode(param));
- if (r)
- return r;
-
- r = populate_table(t, param);
- if (r) {
- dm_table_put(t);
- return r;
- }
-
- md = find_device(param);
- if (!md) {
- dm_table_put(t);
- return -ENXIO;
- }
-
- r = dm_swap_table(md, t);
- if (r) {
- dm_put(md);
- dm_table_put(t);
- return r;
- }
- dm_table_put(t); /* md will have taken its own reference */
-
- set_disk_ro(dm_disk(md), (param->flags & DM_READONLY_FLAG) ? 1 : 0);
- dm_put(md);
-
- r = info(param, user);
- return r;
-}
-
-static int rename(struct dm_ioctl *param, struct dm_ioctl *user)
-{
- int r;
- char *new_name = (char *) param + param->data_start;
-
- if (valid_str(new_name, (void *) param,
- (void *) param + param->data_size)) {
- DMWARN("Invalid new logical volume name supplied.");
- return -EINVAL;
- }
-
- r = check_name(new_name);
- if (r)
- return r;
-
- return dm_hash_rename(param->name, new_name);
-}
-
-
-/*-----------------------------------------------------------------
- * Implementation of open/close/ioctl on the special char
- * device.
- *---------------------------------------------------------------*/
-static ioctl_fn lookup_ioctl(unsigned int cmd)
-{
- static struct {
- int cmd;
- ioctl_fn fn;
- } _ioctls[] = {
- {DM_VERSION_CMD, NULL}, /* version is dealt with elsewhere */
- {DM_REMOVE_ALL_CMD, remove_all},
- {DM_DEV_CREATE_CMD, create},
- {DM_DEV_REMOVE_CMD, remove},
- {DM_DEV_RELOAD_CMD, reload},
- {DM_DEV_RENAME_CMD, rename},
- {DM_DEV_SUSPEND_CMD, suspend},
- {DM_DEV_DEPS_CMD, dep},
- {DM_DEV_STATUS_CMD, info},
- {DM_TARGET_STATUS_CMD, get_status},
- {DM_TARGET_WAIT_CMD, wait_device_event},
- };
-
- return (cmd >= ARRAY_SIZE(_ioctls)) ? NULL : _ioctls[cmd].fn;
-}
-
-/*
- * As well as checking the version compatibility this always
- * copies the kernel interface version out.
- */
-static int check_version(unsigned int cmd, struct dm_ioctl *user)
-{
- uint32_t version[3];
- int r = 0;
-
- if (copy_from_user(version, user->version, sizeof(version)))
- return -EFAULT;
-
- if ((DM_VERSION_MAJOR != version[0]) ||
- (DM_VERSION_MINOR < version[1])) {
- DMWARN("ioctl interface mismatch: "
- "kernel(%u.%u.%u), user(%u.%u.%u), cmd(%d)",
- DM_VERSION_MAJOR, DM_VERSION_MINOR,
- DM_VERSION_PATCHLEVEL,
- version[0], version[1], version[2], cmd);
- r = -EINVAL;
- }
-
- /*
- * Fill in the kernel version.
- */
- version[0] = DM_VERSION_MAJOR;
- version[1] = DM_VERSION_MINOR;
- version[2] = DM_VERSION_PATCHLEVEL;
- if (copy_to_user(user->version, version, sizeof(version)))
- return -EFAULT;
-
- return r;
-}
-
-static void free_params(struct dm_ioctl *param)
-{
- vfree(param);
-}
-
-static int copy_params(struct dm_ioctl *user, struct dm_ioctl **param)
-{
- struct dm_ioctl tmp, *dmi;
-
- if (copy_from_user(&tmp, user, sizeof(tmp)))
- return -EFAULT;
-
- if (tmp.data_size < sizeof(tmp))
- return -EINVAL;
-
- dmi = (struct dm_ioctl *) vmalloc(tmp.data_size);
- if (!dmi)
- return -ENOMEM;
-
- if (copy_from_user(dmi, user, tmp.data_size)) {
- vfree(dmi);
- return -EFAULT;
- }
-
- *param = dmi;
- return 0;
-}
-
-static int validate_params(uint cmd, struct dm_ioctl *param)
-{
- /* Ignores parameters */
- if (cmd == DM_REMOVE_ALL_CMD)
- return 0;
-
- /* Unless creating, either name of uuid but not both */
- if (cmd != DM_DEV_CREATE_CMD) {
- if ((!*param->uuid && !*param->name) ||
- (*param->uuid && *param->name)) {
- DMWARN("one of name or uuid must be supplied");
- return -EINVAL;
- }
- }
-
- /* Ensure strings are terminated */
- param->name[DM_NAME_LEN - 1] = '\0';
- param->uuid[DM_UUID_LEN - 1] = '\0';
-
- return 0;
-}
-
-static int ctl_ioctl(struct inode *inode, struct file *file,
- uint command, ulong u)
-{
- int r = 0;
- unsigned int cmd;
- struct dm_ioctl *param;
- struct dm_ioctl *user = (struct dm_ioctl *) u;
- ioctl_fn fn = NULL;
-
- /* only root can play with this */
- if (!capable(CAP_SYS_ADMIN))
- return -EACCES;
-
- if (_IOC_TYPE(command) != DM_IOCTL)
- return -ENOTTY;
-
- cmd = _IOC_NR(command);
-
- /*
- * Check the interface version passed in. This also
- * writes out the kernels interface version.
- */
- r = check_version(cmd, user);
- if (r)
- return r;
-
- /*
- * Nothing more to do for the version command.
- */
- if (cmd == DM_VERSION_CMD)
- return 0;
-
- fn = lookup_ioctl(cmd);
- if (!fn) {
- DMWARN("dm_ctl_ioctl: unknown command 0x%x", command);
- return -ENOTTY;
- }
-
- /*
- * Copy the parameters into kernel space.
- */
- r = copy_params(user, ¶m);
- if (r)
- return r;
-
- r = validate_params(cmd, param);
- if (r) {
- free_params(param);
- return r;
- }
-
- r = fn(param, user);
- free_params(param);
- return r;
-}
-
-static struct file_operations _ctl_fops = {
- .ioctl = ctl_ioctl,
- .owner = THIS_MODULE,
-};
-
-static struct miscdevice _dm_misc = {
- .minor = MISC_DYNAMIC_MINOR,
- .name = DM_NAME,
- .devfs_name = "mapper/control",
- .fops = &_ctl_fops
-};
-
-/*
- * Create misc character device and link to DM_DIR/control.
- */
-int __init dm_interface_init(void)
-{
- int r;
-
- r = dm_hash_init();
- if (r)
- return r;
-
- r = misc_register(&_dm_misc);
- if (r) {
- DMERR("misc_register failed for control device");
- dm_hash_exit();
- return r;
- }
-
- DMINFO("%d.%d.%d%s initialised: %s", DM_VERSION_MAJOR,
- DM_VERSION_MINOR, DM_VERSION_PATCHLEVEL, DM_VERSION_EXTRA,
- DM_DRIVER_EMAIL);
- return 0;
-
- if (misc_deregister(&_dm_misc) < 0)
- DMERR("misc_deregister failed for control device");
- dm_hash_exit();
- return r;
-}
-void dm_interface_exit(void)
-{
- if (misc_deregister(&_dm_misc) < 0)
- DMERR("misc_deregister failed for control device");
- dm_hash_exit();
-}
+#ifdef CONFIG_DM_IOCTL_V4
+#include "dm-ioctl-v4.c"
+#else
+#include "dm-ioctl-v1.c"
+#endif
--- diff/drivers/md/dm.c 2003-06-30 17:12:16.000000000 +0100
+++ source/drivers/md/dm.c 2003-07-01 15:36:42.000000000 +0100
@@ -516,6 +516,11 @@
down_read(&md->lock);
}
+ if (!md->map) {
+ bio_io_error(bio, bio->bi_size);
+ return 0;
+ }
+
__split_bio(md, bio);
up_read(&md->lock);
return 0;
@@ -672,6 +677,9 @@
static void __unbind(struct mapped_device *md)
{
+ if (!md->map)
+ return;
+
dm_table_event_callback(md->map, NULL, NULL);
dm_table_put(md->map);
md->map = NULL;
@@ -682,35 +690,26 @@
* Constructor for a new device.
*/
static int create_aux(unsigned int minor, int persistent,
- struct dm_table *table, struct mapped_device **result)
+ struct mapped_device **result)
{
- int r;
struct mapped_device *md;
md = alloc_dev(minor, persistent);
if (!md)
return -ENXIO;
- r = __bind(md, table);
- if (r) {
- free_dev(md);
- return r;
- }
- dm_table_resume_targets(md->map);
-
*result = md;
return 0;
}
-int dm_create(struct dm_table *table, struct mapped_device **result)
+int dm_create(struct mapped_device **result)
{
- return create_aux(0, 0, table, result);
+ return create_aux(0, 0, result);
}
-int dm_create_with_minor(unsigned int minor,
- struct dm_table *table, struct mapped_device **result)
+int dm_create_with_minor(unsigned int minor, struct mapped_device **result)
{
- return create_aux(minor, 1, table, result);
+ return create_aux(minor, 1, result);
}
void dm_get(struct mapped_device *md)
@@ -721,7 +720,7 @@
void dm_put(struct mapped_device *md)
{
if (atomic_dec_and_test(&md->holders)) {
- if (!test_bit(DMF_SUSPENDED, &md->flags))
+ if (!test_bit(DMF_SUSPENDED, &md->flags) && md->map)
dm_table_suspend_targets(md->map);
__unbind(md);
free_dev(md);
@@ -812,7 +811,8 @@
down_write(&md->lock);
remove_wait_queue(&md->wait, &wait);
set_bit(DMF_SUSPENDED, &md->flags);
- dm_table_suspend_targets(md->map);
+ if (md->map)
+ dm_table_suspend_targets(md->map);
up_write(&md->lock);
return 0;
@@ -823,7 +823,8 @@
struct deferred_io *def;
down_write(&md->lock);
- if (!test_bit(DMF_SUSPENDED, &md->flags) ||
+ if (!md->map ||
+ !test_bit(DMF_SUSPENDED, &md->flags) ||
!dm_table_get_size(md->map)) {
up_write(&md->lock);
return -EINVAL;
@@ -893,7 +894,8 @@
down_read(&md->lock);
t = md->map;
- dm_table_get(t);
+ if (t)
+ dm_table_get(t);
up_read(&md->lock);
return t;
--- diff/drivers/md/dm.h 2003-06-30 17:11:07.000000000 +0100
+++ source/drivers/md/dm.h 2003-07-01 15:36:42.000000000 +0100
@@ -51,9 +51,8 @@
* Functions for manipulating a struct mapped_device.
* Drop the reference with dm_put when you finish with the object.
*---------------------------------------------------------------*/
-int dm_create(struct dm_table *table, struct mapped_device **md);
-int dm_create_with_minor(unsigned int minor, struct dm_table *table,
- struct mapped_device **md);
+int dm_create(struct mapped_device **md);
+int dm_create_with_minor(unsigned int minor, struct mapped_device **md);
/*
* Reference counting for md.
--- diff/include/linux/dm-ioctl.h 2002-12-30 10:17:13.000000000 +0000
+++ source/include/linux/dm-ioctl.h 2003-07-01 15:36:42.000000000 +0100
@@ -1,5 +1,5 @@
/*
- * Copyright (C) 2001 Sistina Software (UK) Limited.
+ * Copyright (C) 2003 Sistina Software (UK) Limited.
*
* This file is released under the LGPL.
*/
@@ -7,143 +7,12 @@
#ifndef _LINUX_DM_IOCTL_H
#define _LINUX_DM_IOCTL_H
-#include <linux/types.h>
+#include <linux/config.h>
-#define DM_DIR "mapper" /* Slashes not supported */
-#define DM_MAX_TYPE_NAME 16
-#define DM_NAME_LEN 128
-#define DM_UUID_LEN 129
+#ifdef CONFIG_DM_IOCTL_V4
+#include "dm-ioctl-v4.h"
+#else
+#include "dm-ioctl-v1.h"
+#endif
-/*
- * Implements a traditional ioctl interface to the device mapper.
- */
-
-/*
- * All ioctl arguments consist of a single chunk of memory, with
- * this structure at the start. If a uuid is specified any
- * lookup (eg. for a DM_INFO) will be done on that, *not* the
- * name.
- */
-struct dm_ioctl {
- /*
- * The version number is made up of three parts:
- * major - no backward or forward compatibility,
- * minor - only backwards compatible,
- * patch - both backwards and forwards compatible.
- *
- * All clients of the ioctl interface should fill in the
- * version number of the interface that they were
- * compiled with.
- *
- * All recognised ioctl commands (ie. those that don't
- * return -ENOTTY) fill out this field, even if the
- * command failed.
- */
- uint32_t version[3]; /* in/out */
- uint32_t data_size; /* total size of data passed in
- * including this struct */
-
- uint32_t data_start; /* offset to start of data
- * relative to start of this struct */
-
- uint32_t target_count; /* in/out */
- uint32_t open_count; /* out */
- uint32_t flags; /* in/out */
-
- __kernel_dev_t dev; /* in/out */
-
- char name[DM_NAME_LEN]; /* device name */
- char uuid[DM_UUID_LEN]; /* unique identifier for
- * the block device */
-};
-
-/*
- * Used to specify tables. These structures appear after the
- * dm_ioctl.
- */
-struct dm_target_spec {
- int32_t status; /* used when reading from kernel only */
- uint64_t sector_start;
- uint32_t length;
-
- /*
- * Offset in bytes (from the start of this struct) to
- * next target_spec.
- */
- uint32_t next;
-
- char target_type[DM_MAX_TYPE_NAME];
-
- /*
- * Parameter string starts immediately after this object.
- * Be careful to add padding after string to ensure correct
- * alignment of subsequent dm_target_spec.
- */
-};
-
-/*
- * Used to retrieve the target dependencies.
- */
-struct dm_target_deps {
- uint32_t count;
-
- __kernel_dev_t dev[0]; /* out */
-};
-
-/*
- * If you change this make sure you make the corresponding change
- * to dm-ioctl.c:lookup_ioctl()
- */
-enum {
- /* Top level cmds */
- DM_VERSION_CMD = 0,
- DM_REMOVE_ALL_CMD,
-
- /* device level cmds */
- DM_DEV_CREATE_CMD,
- DM_DEV_REMOVE_CMD,
- DM_DEV_RELOAD_CMD,
- DM_DEV_RENAME_CMD,
- DM_DEV_SUSPEND_CMD,
- DM_DEV_DEPS_CMD,
- DM_DEV_STATUS_CMD,
-
- /* target level cmds */
- DM_TARGET_STATUS_CMD,
- DM_TARGET_WAIT_CMD
-};
-
-#define DM_IOCTL 0xfd
-
-#define DM_VERSION _IOWR(DM_IOCTL, DM_VERSION_CMD, struct dm_ioctl)
-#define DM_REMOVE_ALL _IOWR(DM_IOCTL, DM_REMOVE_ALL_CMD, struct dm_ioctl)
-
-#define DM_DEV_CREATE _IOWR(DM_IOCTL, DM_DEV_CREATE_CMD, struct dm_ioctl)
-#define DM_DEV_REMOVE _IOWR(DM_IOCTL, DM_DEV_REMOVE_CMD, struct dm_ioctl)
-#define DM_DEV_RELOAD _IOWR(DM_IOCTL, DM_DEV_RELOAD_CMD, struct dm_ioctl)
-#define DM_DEV_SUSPEND _IOWR(DM_IOCTL, DM_DEV_SUSPEND_CMD, struct dm_ioctl)
-#define DM_DEV_RENAME _IOWR(DM_IOCTL, DM_DEV_RENAME_CMD, struct dm_ioctl)
-#define DM_DEV_DEPS _IOWR(DM_IOCTL, DM_DEV_DEPS_CMD, struct dm_ioctl)
-#define DM_DEV_STATUS _IOWR(DM_IOCTL, DM_DEV_STATUS_CMD, struct dm_ioctl)
-
-#define DM_TARGET_STATUS _IOWR(DM_IOCTL, DM_TARGET_STATUS_CMD, struct dm_ioctl)
-#define DM_TARGET_WAIT _IOWR(DM_IOCTL, DM_TARGET_WAIT_CMD, struct dm_ioctl)
-
-#define DM_VERSION_MAJOR 1
-#define DM_VERSION_MINOR 0
-#define DM_VERSION_PATCHLEVEL 6
-#define DM_VERSION_EXTRA "-ioctl (2002-10-15)"
-
-/* Status bits */
-#define DM_READONLY_FLAG 0x00000001
-#define DM_SUSPEND_FLAG 0x00000002
-#define DM_EXISTS_FLAG 0x00000004
-#define DM_PERSISTENT_DEV_FLAG 0x00000008
-
-/*
- * Flag passed into ioctl STATUS command to get table information
- * rather than current status.
- */
-#define DM_STATUS_TABLE_FLAG 0x00000010
-
-#endif /* _LINUX_DM_IOCTL_H */
+#endif
--- diff/drivers/md/dm-ioctl-v1.c 1970-01-01 01:00:00.000000000 +0100
+++ source/drivers/md/dm-ioctl-v1.c 2003-07-01 15:36:42.000000000 +0100
@@ -0,0 +1,1160 @@
+/*
+ * Copyright (C) 2001, 2002 Sistina Software (UK) Limited.
+ *
+ * This file is released under the GPL.
+ */
+
+#include "dm.h"
+
+#include <linux/module.h>
+#include <linux/vmalloc.h>
+#include <linux/miscdevice.h>
+#include <linux/dm-ioctl.h>
+#include <linux/init.h>
+#include <linux/wait.h>
+#include <linux/blk.h>
+#include <linux/slab.h>
+#include <linux/devfs_fs_kernel.h>
+
+#include <asm/uaccess.h>
+
+#define DM_DRIVER_EMAIL "[email protected]"
+
+/*-----------------------------------------------------------------
+ * The ioctl interface needs to be able to look up devices by
+ * name or uuid.
+ *---------------------------------------------------------------*/
+struct hash_cell {
+ struct list_head name_list;
+ struct list_head uuid_list;
+
+ char *name;
+ char *uuid;
+ struct mapped_device *md;
+};
+
+#define NUM_BUCKETS 64
+#define MASK_BUCKETS (NUM_BUCKETS - 1)
+static struct list_head _name_buckets[NUM_BUCKETS];
+static struct list_head _uuid_buckets[NUM_BUCKETS];
+
+void dm_hash_remove_all(void);
+
+/*
+ * Guards access to all three tables.
+ */
+static DECLARE_RWSEM(_hash_lock);
+
+static void init_buckets(struct list_head *buckets)
+{
+ unsigned int i;
+
+ for (i = 0; i < NUM_BUCKETS; i++)
+ INIT_LIST_HEAD(buckets + i);
+}
+
+int dm_hash_init(void)
+{
+ init_buckets(_name_buckets);
+ init_buckets(_uuid_buckets);
+ devfs_mk_dir(DM_DIR);
+ return 0;
+}
+
+void dm_hash_exit(void)
+{
+ dm_hash_remove_all();
+ devfs_remove(DM_DIR);
+}
+
+/*-----------------------------------------------------------------
+ * Hash function:
+ * We're not really concerned with the str hash function being
+ * fast since it's only used by the ioctl interface.
+ *---------------------------------------------------------------*/
+static unsigned int hash_str(const char *str)
+{
+ const unsigned int hash_mult = 2654435387U;
+ unsigned int h = 0;
+
+ while (*str)
+ h = (h + (unsigned int) *str++) * hash_mult;
+
+ return h & MASK_BUCKETS;
+}
+
+/*-----------------------------------------------------------------
+ * Code for looking up a device by name
+ *---------------------------------------------------------------*/
+static struct hash_cell *__get_name_cell(const char *str)
+{
+ struct list_head *tmp;
+ struct hash_cell *hc;
+ unsigned int h = hash_str(str);
+
+ list_for_each (tmp, _name_buckets + h) {
+ hc = list_entry(tmp, struct hash_cell, name_list);
+ if (!strcmp(hc->name, str))
+ return hc;
+ }
+
+ return NULL;
+}
+
+static struct hash_cell *__get_uuid_cell(const char *str)
+{
+ struct list_head *tmp;
+ struct hash_cell *hc;
+ unsigned int h = hash_str(str);
+
+ list_for_each (tmp, _uuid_buckets + h) {
+ hc = list_entry(tmp, struct hash_cell, uuid_list);
+ if (!strcmp(hc->uuid, str))
+ return hc;
+ }
+
+ return NULL;
+}
+
+/*-----------------------------------------------------------------
+ * Inserting, removing and renaming a device.
+ *---------------------------------------------------------------*/
+static inline char *kstrdup(const char *str)
+{
+ char *r = kmalloc(strlen(str) + 1, GFP_KERNEL);
+ if (r)
+ strcpy(r, str);
+ return r;
+}
+
+static struct hash_cell *alloc_cell(const char *name, const char *uuid,
+ struct mapped_device *md)
+{
+ struct hash_cell *hc;
+
+ hc = kmalloc(sizeof(*hc), GFP_KERNEL);
+ if (!hc)
+ return NULL;
+
+ hc->name = kstrdup(name);
+ if (!hc->name) {
+ kfree(hc);
+ return NULL;
+ }
+
+ if (!uuid)
+ hc->uuid = NULL;
+
+ else {
+ hc->uuid = kstrdup(uuid);
+ if (!hc->uuid) {
+ kfree(hc->name);
+ kfree(hc);
+ return NULL;
+ }
+ }
+
+ INIT_LIST_HEAD(&hc->name_list);
+ INIT_LIST_HEAD(&hc->uuid_list);
+ hc->md = md;
+ return hc;
+}
+
+static void free_cell(struct hash_cell *hc)
+{
+ if (hc) {
+ kfree(hc->name);
+ kfree(hc->uuid);
+ kfree(hc);
+ }
+}
+
+/*
+ * devfs stuff.
+ */
+static int register_with_devfs(struct hash_cell *hc)
+{
+ struct gendisk *disk = dm_disk(hc->md);
+
+ devfs_mk_bdev(MKDEV(disk->major, disk->first_minor),
+ S_IFBLK | S_IRUSR | S_IWUSR | S_IRGRP,
+ DM_DIR "/%s", hc->name);
+ return 0;
+}
+
+static int unregister_with_devfs(struct hash_cell *hc)
+{
+ devfs_remove(DM_DIR"/%s", hc->name);
+ return 0;
+}
+
+/*
+ * The kdev_t and uuid of a device can never change once it is
+ * initially inserted.
+ */
+int dm_hash_insert(const char *name, const char *uuid, struct mapped_device *md)
+{
+ struct hash_cell *cell;
+
+ /*
+ * Allocate the new cells.
+ */
+ cell = alloc_cell(name, uuid, md);
+ if (!cell)
+ return -ENOMEM;
+
+ /*
+ * Insert the cell into all three hash tables.
+ */
+ down_write(&_hash_lock);
+ if (__get_name_cell(name))
+ goto bad;
+
+ list_add(&cell->name_list, _name_buckets + hash_str(name));
+
+ if (uuid) {
+ if (__get_uuid_cell(uuid)) {
+ list_del(&cell->name_list);
+ goto bad;
+ }
+ list_add(&cell->uuid_list, _uuid_buckets + hash_str(uuid));
+ }
+ register_with_devfs(cell);
+ dm_get(md);
+ up_write(&_hash_lock);
+
+ return 0;
+
+ bad:
+ up_write(&_hash_lock);
+ free_cell(cell);
+ return -EBUSY;
+}
+
+void __hash_remove(struct hash_cell *hc)
+{
+ /* remove from the dev hash */
+ list_del(&hc->uuid_list);
+ list_del(&hc->name_list);
+ unregister_with_devfs(hc);
+ dm_put(hc->md);
+ free_cell(hc);
+}
+
+void dm_hash_remove_all(void)
+{
+ int i;
+ struct hash_cell *hc;
+ struct list_head *tmp, *n;
+
+ down_write(&_hash_lock);
+ for (i = 0; i < NUM_BUCKETS; i++) {
+ list_for_each_safe (tmp, n, _name_buckets + i) {
+ hc = list_entry(tmp, struct hash_cell, name_list);
+ __hash_remove(hc);
+ }
+ }
+ up_write(&_hash_lock);
+}
+
+int dm_hash_rename(const char *old, const char *new)
+{
+ char *new_name, *old_name;
+ struct hash_cell *hc;
+
+ /*
+ * duplicate new.
+ */
+ new_name = kstrdup(new);
+ if (!new_name)
+ return -ENOMEM;
+
+ down_write(&_hash_lock);
+
+ /*
+ * Is new free ?
+ */
+ hc = __get_name_cell(new);
+ if (hc) {
+ DMWARN("asked to rename to an already existing name %s -> %s",
+ old, new);
+ up_write(&_hash_lock);
+ kfree(new_name);
+ return -EBUSY;
+ }
+
+ /*
+ * Is there such a device as 'old' ?
+ */
+ hc = __get_name_cell(old);
+ if (!hc) {
+ DMWARN("asked to rename a non existent device %s -> %s",
+ old, new);
+ up_write(&_hash_lock);
+ kfree(new_name);
+ return -ENXIO;
+ }
+
+ /*
+ * rename and move the name cell.
+ */
+ unregister_with_devfs(hc);
+
+ list_del(&hc->name_list);
+ old_name = hc->name;
+ hc->name = new_name;
+ list_add(&hc->name_list, _name_buckets + hash_str(new_name));
+
+ /* rename the device node in devfs */
+ register_with_devfs(hc);
+
+ up_write(&_hash_lock);
+ kfree(old_name);
+ return 0;
+}
+
+
+/*-----------------------------------------------------------------
+ * Implementation of the ioctl commands
+ *---------------------------------------------------------------*/
+
+/*
+ * All the ioctl commands get dispatched to functions with this
+ * prototype.
+ */
+typedef int (*ioctl_fn)(struct dm_ioctl *param, struct dm_ioctl *user);
+
+/*
+ * Check a string doesn't overrun the chunk of
+ * memory we copied from userland.
+ */
+static int valid_str(char *str, void *begin, void *end)
+{
+ while (((void *) str >= begin) && ((void *) str < end))
+ if (!*str++)
+ return 0;
+
+ return -EINVAL;
+}
+
+static int next_target(struct dm_target_spec *last, uint32_t next,
+ void *begin, void *end,
+ struct dm_target_spec **spec, char **params)
+{
+ *spec = (struct dm_target_spec *)
+ ((unsigned char *) last + next);
+ *params = (char *) (*spec + 1);
+
+ if (*spec < (last + 1) || ((void *) *spec > end))
+ return -EINVAL;
+
+ return valid_str(*params, begin, end);
+}
+
+static int populate_table(struct dm_table *table, struct dm_ioctl *args)
+{
+ int r, first = 1;
+ unsigned int i = 0;
+ struct dm_target_spec *spec;
+ char *params;
+ void *begin, *end;
+
+ if (!args->target_count) {
+ DMWARN("populate_table: no targets specified");
+ return -EINVAL;
+ }
+
+ begin = (void *) args;
+ end = begin + args->data_size;
+
+ for (i = 0; i < args->target_count; i++) {
+
+ if (first)
+ r = next_target((struct dm_target_spec *) args,
+ args->data_start,
+ begin, end, &spec, ¶ms);
+ else
+ r = next_target(spec, spec->next, begin, end,
+ &spec, ¶ms);
+
+ if (r) {
+ DMWARN("unable to find target");
+ return -EINVAL;
+ }
+
+ r = dm_table_add_target(table, spec->target_type,
+ (sector_t) spec->sector_start,
+ (sector_t) spec->length,
+ params);
+ if (r) {
+ DMWARN("internal error adding target to table");
+ return -EINVAL;
+ }
+
+ first = 0;
+ }
+
+ return dm_table_complete(table);
+}
+
+/*
+ * Round up the ptr to the next 'align' boundary. Obviously
+ * 'align' must be a power of 2.
+ */
+static inline void *align_ptr(void *ptr, unsigned int align)
+{
+ align--;
+ return (void *) (((unsigned long) (ptr + align)) & ~align);
+}
+
+/*
+ * Copies a dm_ioctl and an optional additional payload to
+ * userland.
+ */
+static int results_to_user(struct dm_ioctl *user, struct dm_ioctl *param,
+ void *data, uint32_t len)
+{
+ int r;
+ void *ptr = NULL;
+
+ if (data) {
+ ptr = align_ptr(user + 1, sizeof(unsigned long));
+ param->data_start = ptr - (void *) user;
+ }
+
+ /*
+ * The version number has already been filled in, so we
+ * just copy later fields.
+ */
+ r = copy_to_user(&user->data_size, ¶m->data_size,
+ sizeof(*param) - sizeof(param->version));
+ if (r)
+ return -EFAULT;
+
+ if (data) {
+ if (param->data_start + len > param->data_size)
+ return -ENOSPC;
+
+ if (copy_to_user(ptr, data, len))
+ r = -EFAULT;
+ }
+
+ return r;
+}
+
+/*
+ * Fills in a dm_ioctl structure, ready for sending back to
+ * userland.
+ */
+static int __info(struct mapped_device *md, struct dm_ioctl *param)
+{
+ struct dm_table *table;
+ struct block_device *bdev;
+ struct gendisk *disk = dm_disk(md);
+
+ param->flags = DM_EXISTS_FLAG;
+ if (dm_suspended(md))
+ param->flags |= DM_SUSPEND_FLAG;
+
+ bdev = bdget_disk(disk, 0);
+ if (!bdev)
+ return -ENXIO;
+
+ param->dev = bdev->bd_dev;
+ param->open_count = bdev->bd_openers;
+ bdput(bdev);
+
+ if (disk->policy)
+ param->flags |= DM_READONLY_FLAG;
+
+ table = dm_get_table(md);
+ param->target_count = dm_table_get_num_targets(table);
+ dm_table_put(table);
+
+ return 0;
+}
+
+/*
+ * Always use UUID for lookups if it's present, otherwise use name.
+ */
+static inline struct mapped_device *find_device(struct dm_ioctl *param)
+{
+ struct hash_cell *hc;
+ struct mapped_device *md = NULL;
+
+ down_read(&_hash_lock);
+ hc = *param->uuid ? __get_uuid_cell(param->uuid) :
+ __get_name_cell(param->name);
+ if (hc) {
+ md = hc->md;
+
+ /*
+ * Sneakily write in both the name and the uuid
+ * while we have the cell.
+ */
+ strlcpy(param->name, hc->name, sizeof(param->name));
+ if (hc->uuid)
+ strlcpy(param->uuid, hc->uuid, sizeof(param->uuid));
+ else
+ param->uuid[0] = '\0';
+
+ dm_get(md);
+ }
+ up_read(&_hash_lock);
+
+ return md;
+}
+
+#define ALIGNMENT sizeof(int)
+static void *_align(void *ptr, unsigned int a)
+{
+ register unsigned long align = --a;
+
+ return (void *) (((unsigned long) ptr + align) & ~align);
+}
+
+/*
+ * Copies device info back to user space, used by
+ * the create and info ioctls.
+ */
+static int info(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ struct mapped_device *md;
+
+ param->flags = 0;
+
+ md = find_device(param);
+ if (!md)
+ /*
+ * Device not found - returns cleared exists flag.
+ */
+ goto out;
+
+ __info(md, param);
+ dm_put(md);
+
+ out:
+ return results_to_user(user, param, NULL, 0);
+}
+
+static inline int get_mode(struct dm_ioctl *param)
+{
+ int mode = FMODE_READ | FMODE_WRITE;
+
+ if (param->flags & DM_READONLY_FLAG)
+ mode = FMODE_READ;
+
+ return mode;
+}
+
+static int check_name(const char *name)
+{
+ if (name[0] == '/') {
+ DMWARN("invalid device name");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static int create(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ int r;
+ struct dm_table *t;
+ struct mapped_device *md;
+
+ r = check_name(param->name);
+ if (r)
+ return r;
+
+ r = dm_table_create(&t, get_mode(param));
+ if (r)
+ return r;
+
+ r = populate_table(t, param);
+ if (r) {
+ dm_table_put(t);
+ return r;
+ }
+
+ if (param->flags & DM_PERSISTENT_DEV_FLAG)
+ r = dm_create_with_minor(minor(to_kdev_t(param->dev)), &md);
+ else
+ r = dm_create(&md);
+
+ if (r) {
+ dm_table_put(t);
+ return r;
+ }
+
+ /* suspend the device */
+ r = dm_suspend(md);
+ if (r) {
+ DMWARN("suspend failed");
+ dm_table_put(t);
+ dm_put(md);
+ return r;
+ }
+ /* swap in the table */
+ r = dm_swap_table(md, t);
+ if (r) {
+ DMWARN("table swap failed");
+ dm_table_put(t);
+ dm_put(md);
+ return r;
+ }
+
+ /* resume the device */
+ r = dm_resume(md);
+ if (r) {
+ DMWARN("resume failed");
+ dm_table_put(t);
+ dm_put(md);
+ return r;
+ }
+
+ dm_table_put(t); /* md will have grabbed its own reference */
+
+ set_disk_ro(dm_disk(md), (param->flags & DM_READONLY_FLAG) ? 1 : 0);
+ r = dm_hash_insert(param->name, *param->uuid ? param->uuid : NULL, md);
+ dm_put(md);
+
+ return r ? r : info(param, user);
+}
+
+/*
+ * Build up the status struct for each target
+ */
+static int __status(struct mapped_device *md, struct dm_ioctl *param,
+ char *outbuf, size_t *len)
+{
+ unsigned int i, num_targets;
+ struct dm_target_spec *spec;
+ char *outptr;
+ status_type_t type;
+ struct dm_table *table = dm_get_table(md);
+
+ if (param->flags & DM_STATUS_TABLE_FLAG)
+ type = STATUSTYPE_TABLE;
+ else
+ type = STATUSTYPE_INFO;
+
+ outptr = outbuf;
+
+ /* Get all the target info */
+ num_targets = dm_table_get_num_targets(table);
+ for (i = 0; i < num_targets; i++) {
+ struct dm_target *ti = dm_table_get_target(table, i);
+
+ if (outptr - outbuf +
+ sizeof(struct dm_target_spec) > param->data_size) {
+ dm_table_put(table);
+ return -ENOMEM;
+ }
+
+ spec = (struct dm_target_spec *) outptr;
+
+ spec->status = 0;
+ spec->sector_start = ti->begin;
+ spec->length = ti->len;
+ strlcpy(spec->target_type, ti->type->name,
+ sizeof(spec->target_type));
+
+ outptr += sizeof(struct dm_target_spec);
+
+ /* Get the status/table string from the target driver */
+ if (ti->type->status)
+ ti->type->status(ti, type, outptr,
+ outbuf + param->data_size - outptr);
+ else
+ outptr[0] = '\0';
+
+ outptr += strlen(outptr) + 1;
+ _align(outptr, ALIGNMENT);
+ spec->next = outptr - outbuf;
+ }
+
+ param->target_count = num_targets;
+ *len = outptr - outbuf;
+ dm_table_put(table);
+
+ return 0;
+}
+
+/*
+ * Return the status of a device as a text string for each
+ * target.
+ */
+static int get_status(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ struct mapped_device *md;
+ size_t len = 0;
+ int ret;
+ char *outbuf = NULL;
+
+ md = find_device(param);
+ if (!md)
+ /*
+ * Device not found - returns cleared exists flag.
+ */
+ goto out;
+
+ /* We haven't a clue how long the resultant data will be so
+ just allocate as much as userland has allowed us and make sure
+ we don't overun it */
+ outbuf = kmalloc(param->data_size, GFP_KERNEL);
+ if (!outbuf)
+ goto out;
+ /*
+ * Get the status of all targets
+ */
+ __status(md, param, outbuf, &len);
+
+ /*
+ * Setup the basic dm_ioctl structure.
+ */
+ __info(md, param);
+
+ out:
+ if (md)
+ dm_put(md);
+
+ ret = results_to_user(user, param, outbuf, len);
+
+ if (outbuf)
+ kfree(outbuf);
+
+ return ret;
+}
+
+/*
+ * Wait for a device to report an event
+ */
+static int wait_device_event(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ struct mapped_device *md;
+ DECLARE_WAITQUEUE(wq, current);
+
+ md = find_device(param);
+ if (!md)
+ /*
+ * Device not found - returns cleared exists flag.
+ */
+ goto out;
+
+ /*
+ * Setup the basic dm_ioctl structure.
+ */
+ __info(md, param);
+
+ /*
+ * Wait for a notification event
+ */
+ set_current_state(TASK_INTERRUPTIBLE);
+ if (!dm_add_wait_queue(md, &wq, dm_get_event_nr(md))) {
+ yield();
+ dm_remove_wait_queue(md, &wq);
+ }
+ set_current_state(TASK_RUNNING);
+ dm_put(md);
+
+ out:
+ return results_to_user(user, param, NULL, 0);
+}
+
+/*
+ * Retrieves a list of devices used by a particular dm device.
+ */
+static int dep(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ int r;
+ unsigned int count;
+ struct mapped_device *md;
+ struct list_head *tmp;
+ size_t len = 0;
+ struct dm_target_deps *deps = NULL;
+ struct dm_table *table;
+
+ md = find_device(param);
+ if (!md)
+ goto out;
+ table = dm_get_table(md);
+
+ /*
+ * Setup the basic dm_ioctl structure.
+ */
+ __info(md, param);
+
+ /*
+ * Count the devices.
+ */
+ count = 0;
+ list_for_each(tmp, dm_table_get_devices(table))
+ count++;
+
+ /*
+ * Allocate a kernel space version of the dm_target_status
+ * struct.
+ */
+ if (array_too_big(sizeof(*deps), sizeof(*deps->dev), count)) {
+ dm_table_put(table);
+ dm_put(md);
+ return -ENOMEM;
+ }
+
+ len = sizeof(*deps) + (sizeof(*deps->dev) * count);
+ deps = kmalloc(len, GFP_KERNEL);
+ if (!deps) {
+ dm_table_put(table);
+ dm_put(md);
+ return -ENOMEM;
+ }
+
+ /*
+ * Fill in the devices.
+ */
+ deps->count = count;
+ count = 0;
+ list_for_each(tmp, dm_table_get_devices(table)) {
+ struct dm_dev *dd = list_entry(tmp, struct dm_dev, list);
+ deps->dev[count++] = dd->bdev->bd_dev;
+ }
+ dm_table_put(table);
+ dm_put(md);
+
+ out:
+ r = results_to_user(user, param, deps, len);
+
+ kfree(deps);
+ return r;
+}
+
+static int remove(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ struct hash_cell *hc;
+
+ down_write(&_hash_lock);
+ hc = *param->uuid ? __get_uuid_cell(param->uuid) :
+ __get_name_cell(param->name);
+ if (!hc) {
+ DMWARN("device doesn't appear to be in the dev hash table.");
+ up_write(&_hash_lock);
+ return -EINVAL;
+ }
+
+ /*
+ * You may ask the interface to drop its reference to an
+ * in use device. This is no different to unlinking a
+ * file that someone still has open. The device will not
+ * actually be destroyed until the last opener closes it.
+ * The name and uuid of the device (both are interface
+ * properties) will be available for reuse immediately.
+ *
+ * You don't want to drop a _suspended_ device from the
+ * interface, since that will leave you with no way of
+ * resuming it.
+ */
+ if (dm_suspended(hc->md)) {
+ DMWARN("refusing to remove a suspended device.");
+ up_write(&_hash_lock);
+ return -EPERM;
+ }
+
+ __hash_remove(hc);
+ up_write(&_hash_lock);
+ return 0;
+}
+
+static int remove_all(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ dm_hash_remove_all();
+ return 0;
+}
+
+static int suspend(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ int r;
+ struct mapped_device *md;
+
+ md = find_device(param);
+ if (!md)
+ return -ENXIO;
+
+ if (param->flags & DM_SUSPEND_FLAG)
+ r = dm_suspend(md);
+ else
+ r = dm_resume(md);
+
+ dm_put(md);
+ return r;
+}
+
+static int reload(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ int r;
+ struct mapped_device *md;
+ struct dm_table *t;
+
+ r = dm_table_create(&t, get_mode(param));
+ if (r)
+ return r;
+
+ r = populate_table(t, param);
+ if (r) {
+ dm_table_put(t);
+ return r;
+ }
+
+ md = find_device(param);
+ if (!md) {
+ dm_table_put(t);
+ return -ENXIO;
+ }
+
+ r = dm_swap_table(md, t);
+ if (r) {
+ dm_put(md);
+ dm_table_put(t);
+ return r;
+ }
+ dm_table_put(t); /* md will have taken its own reference */
+
+ set_disk_ro(dm_disk(md), (param->flags & DM_READONLY_FLAG) ? 1 : 0);
+ dm_put(md);
+
+ r = info(param, user);
+ return r;
+}
+
+static int rename(struct dm_ioctl *param, struct dm_ioctl *user)
+{
+ int r;
+ char *new_name = (char *) param + param->data_start;
+
+ if (valid_str(new_name, (void *) param,
+ (void *) param + param->data_size)) {
+ DMWARN("Invalid new logical volume name supplied.");
+ return -EINVAL;
+ }
+
+ r = check_name(new_name);
+ if (r)
+ return r;
+
+ return dm_hash_rename(param->name, new_name);
+}
+
+
+/*-----------------------------------------------------------------
+ * Implementation of open/close/ioctl on the special char
+ * device.
+ *---------------------------------------------------------------*/
+static ioctl_fn lookup_ioctl(unsigned int cmd)
+{
+ static struct {
+ int cmd;
+ ioctl_fn fn;
+ } _ioctls[] = {
+ {DM_VERSION_CMD, NULL}, /* version is dealt with elsewhere */
+ {DM_REMOVE_ALL_CMD, remove_all},
+ {DM_DEV_CREATE_CMD, create},
+ {DM_DEV_REMOVE_CMD, remove},
+ {DM_DEV_RELOAD_CMD, reload},
+ {DM_DEV_RENAME_CMD, rename},
+ {DM_DEV_SUSPEND_CMD, suspend},
+ {DM_DEV_DEPS_CMD, dep},
+ {DM_DEV_STATUS_CMD, info},
+ {DM_TARGET_STATUS_CMD, get_status},
+ {DM_TARGET_WAIT_CMD, wait_device_event},
+ };
+
+ return (cmd >= ARRAY_SIZE(_ioctls)) ? NULL : _ioctls[cmd].fn;
+}
+
+/*
+ * As well as checking the version compatibility this always
+ * copies the kernel interface version out.
+ */
+static int check_version(unsigned int cmd, struct dm_ioctl *user)
+{
+ uint32_t version[3];
+ int r = 0;
+
+ if (copy_from_user(version, user->version, sizeof(version)))
+ return -EFAULT;
+
+ if ((DM_VERSION_MAJOR != version[0]) ||
+ (DM_VERSION_MINOR < version[1])) {
+ DMWARN("ioctl interface mismatch: "
+ "kernel(%u.%u.%u), user(%u.%u.%u), cmd(%d)",
+ DM_VERSION_MAJOR, DM_VERSION_MINOR,
+ DM_VERSION_PATCHLEVEL,
+ version[0], version[1], version[2], cmd);
+ r = -EINVAL;
+ }
+
+ /*
+ * Fill in the kernel version.
+ */
+ version[0] = DM_VERSION_MAJOR;
+ version[1] = DM_VERSION_MINOR;
+ version[2] = DM_VERSION_PATCHLEVEL;
+ if (copy_to_user(user->version, version, sizeof(version)))
+ return -EFAULT;
+
+ return r;
+}
+
+static void free_params(struct dm_ioctl *param)
+{
+ vfree(param);
+}
+
+static int copy_params(struct dm_ioctl *user, struct dm_ioctl **param)
+{
+ struct dm_ioctl tmp, *dmi;
+
+ if (copy_from_user(&tmp, user, sizeof(tmp)))
+ return -EFAULT;
+
+ if (tmp.data_size < sizeof(tmp))
+ return -EINVAL;
+
+ dmi = (struct dm_ioctl *) vmalloc(tmp.data_size);
+ if (!dmi)
+ return -ENOMEM;
+
+ if (copy_from_user(dmi, user, tmp.data_size)) {
+ vfree(dmi);
+ return -EFAULT;
+ }
+
+ *param = dmi;
+ return 0;
+}
+
+static int validate_params(uint cmd, struct dm_ioctl *param)
+{
+ /* Ignores parameters */
+ if (cmd == DM_REMOVE_ALL_CMD)
+ return 0;
+
+ /* Unless creating, either name of uuid but not both */
+ if (cmd != DM_DEV_CREATE_CMD) {
+ if ((!*param->uuid && !*param->name) ||
+ (*param->uuid && *param->name)) {
+ DMWARN("one of name or uuid must be supplied");
+ return -EINVAL;
+ }
+ }
+
+ /* Ensure strings are terminated */
+ param->name[DM_NAME_LEN - 1] = '\0';
+ param->uuid[DM_UUID_LEN - 1] = '\0';
+
+ return 0;
+}
+
+static int ctl_ioctl(struct inode *inode, struct file *file,
+ uint command, ulong u)
+{
+ int r = 0;
+ unsigned int cmd;
+ struct dm_ioctl *param;
+ struct dm_ioctl *user = (struct dm_ioctl *) u;
+ ioctl_fn fn = NULL;
+
+ /* only root can play with this */
+ if (!capable(CAP_SYS_ADMIN))
+ return -EACCES;
+
+ if (_IOC_TYPE(command) != DM_IOCTL)
+ return -ENOTTY;
+
+ cmd = _IOC_NR(command);
+
+ /*
+ * Check the interface version passed in. This also
+ * writes out the kernels interface version.
+ */
+ r = check_version(cmd, user);
+ if (r)
+ return r;
+
+ /*
+ * Nothing more to do for the version command.
+ */
+ if (cmd == DM_VERSION_CMD)
+ return 0;
+
+ fn = lookup_ioctl(cmd);
+ if (!fn) {
+ DMWARN("dm_ctl_ioctl: unknown command 0x%x", command);
+ return -ENOTTY;
+ }
+
+ /*
+ * Copy the parameters into kernel space.
+ */
+ r = copy_params(user, ¶m);
+ if (r)
+ return r;
+
+ r = validate_params(cmd, param);
+ if (r) {
+ free_params(param);
+ return r;
+ }
+
+ r = fn(param, user);
+ free_params(param);
+ return r;
+}
+
+static struct file_operations _ctl_fops = {
+ .ioctl = ctl_ioctl,
+ .owner = THIS_MODULE,
+};
+
+static struct miscdevice _dm_misc = {
+ .minor = MISC_DYNAMIC_MINOR,
+ .name = DM_NAME,
+ .devfs_name = "mapper/control",
+ .fops = &_ctl_fops
+};
+
+/*
+ * Create misc character device and link to DM_DIR/control.
+ */
+int __init dm_interface_init(void)
+{
+ int r;
+
+ r = dm_hash_init();
+ if (r)
+ return r;
+
+ r = misc_register(&_dm_misc);
+ if (r) {
+ DMERR("misc_register failed for control device");
+ dm_hash_exit();
+ return r;
+ }
+
+ DMINFO("%d.%d.%d%s initialised: %s", DM_VERSION_MAJOR,
+ DM_VERSION_MINOR, DM_VERSION_PATCHLEVEL, DM_VERSION_EXTRA,
+ DM_DRIVER_EMAIL);
+ return 0;
+
+ if (misc_deregister(&_dm_misc) < 0)
+ DMERR("misc_deregister failed for control device");
+ dm_hash_exit();
+ return r;
+}
+
+void dm_interface_exit(void)
+{
+ if (misc_deregister(&_dm_misc) < 0)
+ DMERR("misc_deregister failed for control device");
+ dm_hash_exit();
+}
--- diff/drivers/md/dm-ioctl-v4.c 1970-01-01 01:00:00.000000000 +0100
+++ source/drivers/md/dm-ioctl-v4.c 2003-07-01 15:36:42.000000000 +0100
@@ -0,0 +1,1248 @@
+/*
+ * Copyright (C) 2001, 2002 Sistina Software (UK) Limited.
+ *
+ * This file is released under the GPL.
+ */
+
+#include "dm.h"
+
+#include <linux/module.h>
+#include <linux/vmalloc.h>
+#include <linux/miscdevice.h>
+#include <linux/init.h>
+#include <linux/wait.h>
+#include <linux/blk.h>
+#include <linux/slab.h>
+#include <linux/devfs_fs_kernel.h>
+
+#include <asm/uaccess.h>
+
+#define DM_DRIVER_EMAIL "[email protected]"
+
+/*-----------------------------------------------------------------
+ * The ioctl interface needs to be able to look up devices by
+ * name or uuid.
+ *---------------------------------------------------------------*/
+struct hash_cell {
+ struct list_head name_list;
+ struct list_head uuid_list;
+
+ char *name;
+ char *uuid;
+ struct mapped_device *md;
+ struct dm_table *new_map;
+};
+
+#define NUM_BUCKETS 64
+#define MASK_BUCKETS (NUM_BUCKETS - 1)
+static struct list_head _name_buckets[NUM_BUCKETS];
+static struct list_head _uuid_buckets[NUM_BUCKETS];
+
+void dm_hash_remove_all(void);
+
+/*
+ * Guards access to all three tables.
+ */
+static DECLARE_RWSEM(_hash_lock);
+
+static void init_buckets(struct list_head *buckets)
+{
+ unsigned int i;
+
+ for (i = 0; i < NUM_BUCKETS; i++)
+ INIT_LIST_HEAD(buckets + i);
+}
+
+int dm_hash_init(void)
+{
+ init_buckets(_name_buckets);
+ init_buckets(_uuid_buckets);
+ devfs_mk_dir(DM_DIR);
+ return 0;
+}
+
+void dm_hash_exit(void)
+{
+ dm_hash_remove_all();
+ devfs_remove(DM_DIR);
+}
+
+/*-----------------------------------------------------------------
+ * Hash function:
+ * We're not really concerned with the str hash function being
+ * fast since it's only used by the ioctl interface.
+ *---------------------------------------------------------------*/
+static unsigned int hash_str(const char *str)
+{
+ const unsigned int hash_mult = 2654435387U;
+ unsigned int h = 0;
+
+ while (*str)
+ h = (h + (unsigned int) *str++) * hash_mult;
+
+ return h & MASK_BUCKETS;
+}
+
+/*-----------------------------------------------------------------
+ * Code for looking up a device by name
+ *---------------------------------------------------------------*/
+static struct hash_cell *__get_name_cell(const char *str)
+{
+ struct list_head *tmp;
+ struct hash_cell *hc;
+ unsigned int h = hash_str(str);
+
+ list_for_each (tmp, _name_buckets + h) {
+ hc = list_entry(tmp, struct hash_cell, name_list);
+ if (!strcmp(hc->name, str))
+ return hc;
+ }
+
+ return NULL;
+}
+
+static struct hash_cell *__get_uuid_cell(const char *str)
+{
+ struct list_head *tmp;
+ struct hash_cell *hc;
+ unsigned int h = hash_str(str);
+
+ list_for_each (tmp, _uuid_buckets + h) {
+ hc = list_entry(tmp, struct hash_cell, uuid_list);
+ if (!strcmp(hc->uuid, str))
+ return hc;
+ }
+
+ return NULL;
+}
+
+/*-----------------------------------------------------------------
+ * Inserting, removing and renaming a device.
+ *---------------------------------------------------------------*/
+static inline char *kstrdup(const char *str)
+{
+ char *r = kmalloc(strlen(str) + 1, GFP_KERNEL);
+ if (r)
+ strcpy(r, str);
+ return r;
+}
+
+static struct hash_cell *alloc_cell(const char *name, const char *uuid,
+ struct mapped_device *md)
+{
+ struct hash_cell *hc;
+
+ hc = kmalloc(sizeof(*hc), GFP_KERNEL);
+ if (!hc)
+ return NULL;
+
+ hc->name = kstrdup(name);
+ if (!hc->name) {
+ kfree(hc);
+ return NULL;
+ }
+
+ if (!uuid)
+ hc->uuid = NULL;
+
+ else {
+ hc->uuid = kstrdup(uuid);
+ if (!hc->uuid) {
+ kfree(hc->name);
+ kfree(hc);
+ return NULL;
+ }
+ }
+
+ INIT_LIST_HEAD(&hc->name_list);
+ INIT_LIST_HEAD(&hc->uuid_list);
+ hc->md = md;
+ hc->new_map = NULL;
+ return hc;
+}
+
+static void free_cell(struct hash_cell *hc)
+{
+ if (hc) {
+ kfree(hc->name);
+ kfree(hc->uuid);
+ kfree(hc);
+ }
+}
+
+/*
+ * devfs stuff.
+ */
+static int register_with_devfs(struct hash_cell *hc)
+{
+ struct gendisk *disk = dm_disk(hc->md);
+
+ devfs_mk_bdev(MKDEV(disk->major, disk->first_minor),
+ S_IFBLK | S_IRUSR | S_IWUSR | S_IRGRP,
+ DM_DIR "/%s", hc->name);
+ return 0;
+}
+
+static int unregister_with_devfs(struct hash_cell *hc)
+{
+ devfs_remove(DM_DIR"/%s", hc->name);
+ return 0;
+}
+
+/*
+ * The kdev_t and uuid of a device can never change once it is
+ * initially inserted.
+ */
+int dm_hash_insert(const char *name, const char *uuid, struct mapped_device *md)
+{
+ struct hash_cell *cell;
+
+ /*
+ * Allocate the new cells.
+ */
+ cell = alloc_cell(name, uuid, md);
+ if (!cell)
+ return -ENOMEM;
+
+ /*
+ * Insert the cell into all three hash tables.
+ */
+ down_write(&_hash_lock);
+ if (__get_name_cell(name))
+ goto bad;
+
+ list_add(&cell->name_list, _name_buckets + hash_str(name));
+
+ if (uuid) {
+ if (__get_uuid_cell(uuid)) {
+ list_del(&cell->name_list);
+ goto bad;
+ }
+ list_add(&cell->uuid_list, _uuid_buckets + hash_str(uuid));
+ }
+ register_with_devfs(cell);
+ dm_get(md);
+ up_write(&_hash_lock);
+
+ return 0;
+
+ bad:
+ up_write(&_hash_lock);
+ free_cell(cell);
+ return -EBUSY;
+}
+
+void __hash_remove(struct hash_cell *hc)
+{
+ /* remove from the dev hash */
+ list_del(&hc->uuid_list);
+ list_del(&hc->name_list);
+ unregister_with_devfs(hc);
+ dm_put(hc->md);
+ if (hc->new_map)
+ dm_table_put(hc->new_map);
+ free_cell(hc);
+}
+
+void dm_hash_remove_all(void)
+{
+ int i;
+ struct hash_cell *hc;
+ struct list_head *tmp, *n;
+
+ down_write(&_hash_lock);
+ for (i = 0; i < NUM_BUCKETS; i++) {
+ list_for_each_safe (tmp, n, _name_buckets + i) {
+ hc = list_entry(tmp, struct hash_cell, name_list);
+ __hash_remove(hc);
+ }
+ }
+ up_write(&_hash_lock);
+}
+
+int dm_hash_rename(const char *old, const char *new)
+{
+ char *new_name, *old_name;
+ struct hash_cell *hc;
+
+ /*
+ * duplicate new.
+ */
+ new_name = kstrdup(new);
+ if (!new_name)
+ return -ENOMEM;
+
+ down_write(&_hash_lock);
+
+ /*
+ * Is new free ?
+ */
+ hc = __get_name_cell(new);
+ if (hc) {
+ DMWARN("asked to rename to an already existing name %s -> %s",
+ old, new);
+ up_write(&_hash_lock);
+ kfree(new_name);
+ return -EBUSY;
+ }
+
+ /*
+ * Is there such a device as 'old' ?
+ */
+ hc = __get_name_cell(old);
+ if (!hc) {
+ DMWARN("asked to rename a non existent device %s -> %s",
+ old, new);
+ up_write(&_hash_lock);
+ kfree(new_name);
+ return -ENXIO;
+ }
+
+ /*
+ * rename and move the name cell.
+ */
+ list_del(&hc->name_list);
+ old_name = hc->name;
+ hc->name = new_name;
+ list_add(&hc->name_list, _name_buckets + hash_str(new_name));
+
+ /* rename the device node in devfs */
+ unregister_with_devfs(hc);
+ register_with_devfs(hc);
+
+ up_write(&_hash_lock);
+ kfree(old_name);
+ return 0;
+}
+
+/*-----------------------------------------------------------------
+ * Implementation of the ioctl commands
+ *---------------------------------------------------------------*/
+/*
+ * All the ioctl commands get dispatched to functions with this
+ * prototype.
+ */
+typedef int (*ioctl_fn)(struct dm_ioctl *param, size_t param_size);
+
+static int remove_all(struct dm_ioctl *param, size_t param_size)
+{
+ dm_hash_remove_all();
+ param->data_size = 0;
+ return 0;
+}
+
+/*
+ * Round up the ptr to an 8-byte boundary.
+ */
+#define ALIGN_MASK 7
+static inline void *align_ptr(void *ptr)
+{
+ return (void *) (((size_t) (ptr + ALIGN_MASK)) & ~ALIGN_MASK);
+}
+
+/*
+ * Retrieves the data payload buffer from an already allocated
+ * struct dm_ioctl.
+ */
+static void *get_result_buffer(struct dm_ioctl *param, size_t param_size,
+ size_t *len)
+{
+ param->data_start = align_ptr(param + 1) - (void *) param;
+
+ if (param->data_start < param_size)
+ *len = param_size - param->data_start;
+ else
+ *len = 0;
+
+ return ((void *) param) + param->data_start;
+}
+
+static int list_devices(struct dm_ioctl *param, size_t param_size)
+{
+ unsigned int i;
+ struct hash_cell *hc;
+ size_t len, needed = 0;
+ struct gendisk *disk;
+ struct dm_name_list *nl, *old_nl = NULL;
+
+ down_write(&_hash_lock);
+
+ /*
+ * Loop through all the devices working out how much
+ * space we need.
+ */
+ for (i = 0; i < NUM_BUCKETS; i++) {
+ list_for_each_entry (hc, _name_buckets + i, name_list) {
+ needed += sizeof(struct dm_name_list);
+ needed += strlen(hc->name);
+ needed += ALIGN_MASK;
+ }
+ }
+
+ /*
+ * Grab our output buffer.
+ */
+ nl = get_result_buffer(param, param_size, &len);
+ if (len < needed) {
+ param->flags |= DM_BUFFER_FULL_FLAG;
+ goto out;
+ }
+ param->data_size = param->data_start + needed;
+
+ nl->dev = 0; /* Flags no data */
+
+ /*
+ * Now loop through filling out the names.
+ */
+ for (i = 0; i < NUM_BUCKETS; i++) {
+ list_for_each_entry (hc, _name_buckets + i, name_list) {
+ if (old_nl)
+ old_nl->next = (__u32) ((void *) nl -
+ (void *) old_nl);
+ disk = dm_disk(hc->md);
+ nl->dev = MKDEV(disk->major, disk->first_minor);
+ nl->next = 0;
+ strcpy(nl->name, hc->name);
+
+ old_nl = nl;
+ nl = align_ptr(((void *) ++nl) + strlen(hc->name) + 1);
+ }
+ }
+
+ out:
+ up_write(&_hash_lock);
+ return 0;
+}
+
+static int check_name(const char *name)
+{
+ if (strchr(name, '/')) {
+ DMWARN("invalid device name");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/*
+ * Fills in a dm_ioctl structure, ready for sending back to
+ * userland.
+ */
+static int __dev_status(struct mapped_device *md, struct dm_ioctl *param)
+{
+ struct gendisk *disk = dm_disk(md);
+ struct dm_table *table;
+ struct block_device *bdev;
+
+ param->flags &= ~(DM_SUSPEND_FLAG | DM_READONLY_FLAG |
+ DM_ACTIVE_PRESENT_FLAG);
+
+ if (dm_suspended(md))
+ param->flags |= DM_SUSPEND_FLAG;
+
+ bdev = bdget_disk(disk, 0);
+ if (!bdev)
+ return -ENXIO;
+
+ param->dev = MKDEV(disk->major, disk->first_minor);
+
+ /*
+ * Yes, this will be out of date by the time it gets back
+ * to userland, but it is still very useful ofr
+ * debugging.
+ */
+ param->open_count = bdev->bd_openers;
+ bdput(bdev);
+
+ if (disk->policy)
+ param->flags |= DM_READONLY_FLAG;
+
+ param->event_nr = dm_get_event_nr(md);
+
+ table = dm_get_table(md);
+ if (table) {
+ param->flags |= DM_ACTIVE_PRESENT_FLAG;
+ param->target_count = dm_table_get_num_targets(table);
+ dm_table_put(table);
+ } else
+ param->target_count = 0;
+
+ return 0;
+}
+
+static int dev_create(struct dm_ioctl *param, size_t param_size)
+{
+ int r;
+ struct mapped_device *md;
+
+ r = check_name(param->name);
+ if (r)
+ return r;
+
+ if (param->flags & DM_PERSISTENT_DEV_FLAG)
+ r = dm_create_with_minor(minor(to_kdev_t(param->dev)), &md);
+ else
+ r = dm_create(&md);
+
+ if (r)
+ return r;
+
+ r = dm_hash_insert(param->name, *param->uuid ? param->uuid : NULL, md);
+ if (r) {
+ dm_put(md);
+ return r;
+ }
+
+ param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
+
+ r = __dev_status(md, param);
+ dm_put(md);
+
+ return r;
+}
+
+/*
+ * Always use UUID for lookups if it's present, otherwise use name.
+ */
+static inline struct hash_cell *__find_device_hash_cell(struct dm_ioctl *param)
+{
+ return *param->uuid ?
+ __get_uuid_cell(param->uuid) : __get_name_cell(param->name);
+}
+
+static inline struct mapped_device *find_device(struct dm_ioctl *param)
+{
+ struct hash_cell *hc;
+ struct mapped_device *md = NULL;
+
+ down_read(&_hash_lock);
+ hc = __find_device_hash_cell(param);
+ if (hc) {
+ md = hc->md;
+
+ /*
+ * Sneakily write in both the name and the uuid
+ * while we have the cell.
+ */
+ strncpy(param->name, hc->name, sizeof(param->name));
+ if (hc->uuid)
+ strncpy(param->uuid, hc->uuid, sizeof(param->uuid)-1);
+ else
+ param->uuid[0] = '\0';
+
+ if (hc->new_map)
+ param->flags |= DM_INACTIVE_PRESENT_FLAG;
+ else
+ param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
+
+ dm_get(md);
+ }
+ up_read(&_hash_lock);
+
+ return md;
+}
+
+static int dev_remove(struct dm_ioctl *param, size_t param_size)
+{
+ struct hash_cell *hc;
+
+ down_write(&_hash_lock);
+ hc = __find_device_hash_cell(param);
+
+ if (!hc) {
+ DMWARN("device doesn't appear to be in the dev hash table.");
+ up_write(&_hash_lock);
+ return -ENXIO;
+ }
+
+ __hash_remove(hc);
+ up_write(&_hash_lock);
+ param->data_size = 0;
+ return 0;
+}
+
+/*
+ * Check a string doesn't overrun the chunk of
+ * memory we copied from userland.
+ */
+static int invalid_str(char *str, void *end)
+{
+ while ((void *) str < end)
+ if (!*str++)
+ return 0;
+
+ return -EINVAL;
+}
+
+static int dev_rename(struct dm_ioctl *param, size_t param_size)
+{
+ int r;
+ char *new_name = (char *) param + param->data_start;
+
+ if (new_name < (char *) (param + 1) ||
+ invalid_str(new_name, (void *) param + param_size)) {
+ DMWARN("Invalid new logical volume name supplied.");
+ return -EINVAL;
+ }
+
+ r = check_name(new_name);
+ if (r)
+ return r;
+
+ param->data_size = 0;
+ return dm_hash_rename(param->name, new_name);
+}
+
+static int suspend(struct dm_ioctl *param)
+{
+ int r = 0;
+ struct mapped_device *md;
+
+ md = find_device(param);
+ if (!md)
+ return -ENXIO;
+
+ if (!dm_suspended(md))
+ r = dm_suspend(md);
+
+ if (!r)
+ r = __dev_status(md, param);
+
+ dm_put(md);
+ return r;
+}
+
+static int resume(struct dm_ioctl *param)
+{
+ int r = 0;
+ struct hash_cell *hc;
+ struct mapped_device *md;
+ struct dm_table *new_map;
+
+ down_write(&_hash_lock);
+
+ hc = __find_device_hash_cell(param);
+ if (!hc) {
+ DMWARN("device doesn't appear to be in the dev hash table.");
+ up_write(&_hash_lock);
+ return -ENXIO;
+ }
+
+ md = hc->md;
+ dm_get(md);
+
+ new_map = hc->new_map;
+ hc->new_map = NULL;
+ param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
+
+ up_write(&_hash_lock);
+
+ /* Do we need to load a new map ? */
+ if (new_map) {
+ /* Suspend if it isn't already suspended */
+ if (!dm_suspended(md))
+ dm_suspend(md);
+
+ r = dm_swap_table(md, new_map);
+ if (r) {
+ dm_put(md);
+ dm_table_put(new_map);
+ return r;
+ }
+
+ if (!(dm_table_get_mode(new_map) & FMODE_WRITE))
+ set_disk_ro(dm_disk(md), 1);
+
+ dm_table_put(new_map);
+ }
+
+ if (dm_suspended(md))
+ r = dm_resume(md);
+
+ if (!r)
+ r = __dev_status(md, param);
+
+ dm_put(md);
+ return r;
+}
+
+/*
+ * Set or unset the suspension state of a device.
+ * If the device already is in the requested state we just return its status.
+ */
+static int dev_suspend(struct dm_ioctl *param, size_t param_size)
+{
+ if (param->flags & DM_SUSPEND_FLAG)
+ return suspend(param);
+
+ return resume(param);
+}
+
+/*
+ * Copies device info back to user space, used by
+ * the create and info ioctls.
+ */
+static int dev_status(struct dm_ioctl *param, size_t param_size)
+{
+ int r;
+ struct mapped_device *md;
+
+ md = find_device(param);
+ if (!md)
+ return -ENXIO;
+
+ r = __dev_status(md, param);
+ dm_put(md);
+ return r;
+}
+
+/*
+ * Wait for a device to report an event
+ */
+static int dev_wait(struct dm_ioctl *param, size_t param_size)
+{
+ int r;
+ struct mapped_device *md;
+ DECLARE_WAITQUEUE(wq, current);
+
+ md = find_device(param);
+ if (!md)
+ return -ENXIO;
+
+ /*
+ * Wait for a notification event
+ */
+ set_current_state(TASK_INTERRUPTIBLE);
+ if (!dm_add_wait_queue(md, &wq, param->event_nr)) {
+ schedule();
+ dm_remove_wait_queue(md, &wq);
+ }
+
+ /*
+ * The userland program is going to want to know what
+ * changed to trigger the event, so we may as well tell
+ * him and save an ioctl.
+ */
+ r = __dev_status(md, param);
+
+ dm_put(md);
+ return r;
+}
+
+static inline int get_mode(struct dm_ioctl *param)
+{
+ int mode = FMODE_READ | FMODE_WRITE;
+
+ if (param->flags & DM_READONLY_FLAG)
+ mode = FMODE_READ;
+
+ return mode;
+}
+
+static int next_target(struct dm_target_spec *last, __u32 next, void *end,
+ struct dm_target_spec **spec, char **target_params)
+{
+ *spec = (struct dm_target_spec *) ((unsigned char *) last + next);
+ *target_params = (char *) (*spec + 1);
+
+ if (*spec < (last + 1))
+ return -EINVAL;
+
+ return invalid_str(*target_params, end);
+}
+
+static int populate_table(struct dm_table *table,
+ struct dm_ioctl *param, size_t param_size)
+{
+ int r;
+ unsigned int i = 0;
+ struct dm_target_spec *spec = (struct dm_target_spec *) param;
+ __u32 next = param->data_start;
+ void *end = (void *) param + param_size;
+ char *target_params;
+
+ if (!param->target_count) {
+ DMWARN("populate_table: no targets specified");
+ return -EINVAL;
+ }
+
+ for (i = 0; i < param->target_count; i++) {
+
+ r = next_target(spec, next, end, &spec, &target_params);
+ if (r) {
+ DMWARN("unable to find target");
+ return r;
+ }
+
+ r = dm_table_add_target(table, spec->target_type,
+ (sector_t) spec->sector_start,
+ (sector_t) spec->length,
+ target_params);
+ if (r) {
+ DMWARN("error adding target to table");
+ return r;
+ }
+
+ next = spec->next;
+ }
+
+ return dm_table_complete(table);
+}
+
+static int table_load(struct dm_ioctl *param, size_t param_size)
+{
+ int r;
+ struct hash_cell *hc;
+ struct dm_table *t;
+
+ r = dm_table_create(&t, get_mode(param));
+ if (r)
+ return r;
+
+ r = populate_table(t, param, param_size);
+ if (r) {
+ dm_table_put(t);
+ return r;
+ }
+
+ down_write(&_hash_lock);
+ hc = __find_device_hash_cell(param);
+ if (!hc) {
+ DMWARN("device doesn't appear to be in the dev hash table.");
+ up_write(&_hash_lock);
+ return -ENXIO;
+ }
+
+ hc->new_map = t;
+ param->flags |= DM_INACTIVE_PRESENT_FLAG;
+
+ r = __dev_status(hc->md, param);
+ up_write(&_hash_lock);
+ return r;
+}
+
+static int table_clear(struct dm_ioctl *param, size_t param_size)
+{
+ int r;
+ struct hash_cell *hc;
+
+ down_write(&_hash_lock);
+
+ hc = __find_device_hash_cell(param);
+ if (!hc) {
+ DMWARN("device doesn't appear to be in the dev hash table.");
+ up_write(&_hash_lock);
+ return -ENXIO;
+ }
+
+ if (hc->new_map) {
+ dm_table_put(hc->new_map);
+ hc->new_map = NULL;
+ }
+
+ param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
+
+ r = __dev_status(hc->md, param);
+ up_write(&_hash_lock);
+ return r;
+}
+
+/*
+ * Retrieves a list of devices used by a particular dm device.
+ */
+static void retrieve_deps(struct dm_table *table,
+ struct dm_ioctl *param, size_t param_size)
+{
+ unsigned int count = 0;
+ struct list_head *tmp;
+ size_t len, needed;
+ struct dm_target_deps *deps;
+
+ deps = get_result_buffer(param, param_size, &len);
+
+ /*
+ * Count the devices.
+ */
+ list_for_each(tmp, dm_table_get_devices(table))
+ count++;
+
+ /*
+ * Check we have enough space.
+ */
+ needed = sizeof(*deps) + (sizeof(*deps->dev) * count);
+ if (len < needed) {
+ param->flags |= DM_BUFFER_FULL_FLAG;
+ return;
+ }
+
+ /*
+ * Fill in the devices.
+ */
+ deps->count = count;
+ count = 0;
+ list_for_each(tmp, dm_table_get_devices(table)) {
+ struct dm_dev *dd = list_entry(tmp, struct dm_dev, list);
+ deps->dev[count++] = dd->bdev->bd_dev;
+ }
+
+ param->data_size = param->data_start + needed;
+}
+
+static int table_deps(struct dm_ioctl *param, size_t param_size)
+{
+ int r = 0;
+ struct mapped_device *md;
+ struct dm_table *table;
+
+ md = find_device(param);
+ if (!md)
+ return -ENXIO;
+
+ r = __dev_status(md, param);
+ if (r)
+ goto out;
+
+ table = dm_get_table(md);
+ if (table) {
+ retrieve_deps(table, param, param_size);
+ dm_table_put(table);
+ }
+
+ out:
+ dm_put(md);
+ return r;
+}
+
+/*
+ * Build up the status struct for each target
+ */
+static void retrieve_status(struct dm_table *table,
+ struct dm_ioctl *param, size_t param_size)
+{
+ unsigned int i, num_targets;
+ struct dm_target_spec *spec;
+ char *outbuf, *outptr;
+ status_type_t type;
+ size_t remaining, len, used = 0;
+
+ outptr = outbuf = get_result_buffer(param, param_size, &len);
+
+ if (param->flags & DM_STATUS_TABLE_FLAG)
+ type = STATUSTYPE_TABLE;
+ else
+ type = STATUSTYPE_INFO;
+
+ /* Get all the target info */
+ num_targets = dm_table_get_num_targets(table);
+ for (i = 0; i < num_targets; i++) {
+ struct dm_target *ti = dm_table_get_target(table, i);
+
+ remaining = len - (outptr - outbuf);
+ if (remaining < sizeof(struct dm_target_spec)) {
+ param->flags |= DM_BUFFER_FULL_FLAG;
+ break;
+ }
+
+ spec = (struct dm_target_spec *) outptr;
+
+ spec->status = 0;
+ spec->sector_start = ti->begin;
+ spec->length = ti->len;
+ strncpy(spec->target_type, ti->type->name,
+ sizeof(spec->target_type));
+
+ outptr += sizeof(struct dm_target_spec);
+ remaining = len - (outptr - outbuf);
+
+ /* Get the status/table string from the target driver */
+ if (ti->type->status) {
+ if (ti->type->status(ti, type, outptr, remaining)) {
+ param->flags |= DM_BUFFER_FULL_FLAG;
+ break;
+ }
+ } else
+ outptr[0] = '\0';
+
+ outptr += strlen(outptr) + 1;
+ used = param->data_start + (outptr - outbuf);
+
+ align_ptr(outptr);
+ spec->next = outptr - outbuf;
+ }
+
+ if (used)
+ param->data_size = used;
+
+ param->target_count = num_targets;
+}
+
+/*
+ * Return the status of a device as a text string for each
+ * target.
+ */
+static int table_status(struct dm_ioctl *param, size_t param_size)
+{
+ int r;
+ struct mapped_device *md;
+ struct dm_table *table;
+
+ md = find_device(param);
+ if (!md)
+ return -ENXIO;
+
+ r = __dev_status(md, param);
+ if (r)
+ goto out;
+
+ table = dm_get_table(md);
+ if (table) {
+ retrieve_status(table, param, param_size);
+ dm_table_put(table);
+ }
+
+ out:
+ dm_put(md);
+ return r;
+}
+
+/*-----------------------------------------------------------------
+ * Implementation of open/close/ioctl on the special char
+ * device.
+ *---------------------------------------------------------------*/
+static ioctl_fn lookup_ioctl(unsigned int cmd)
+{
+ static struct {
+ int cmd;
+ ioctl_fn fn;
+ } _ioctls[] = {
+ {DM_VERSION_CMD, NULL}, /* version is dealt with elsewhere */
+ {DM_REMOVE_ALL_CMD, remove_all},
+ {DM_LIST_DEVICES_CMD, list_devices},
+
+ {DM_DEV_CREATE_CMD, dev_create},
+ {DM_DEV_REMOVE_CMD, dev_remove},
+ {DM_DEV_RENAME_CMD, dev_rename},
+ {DM_DEV_SUSPEND_CMD, dev_suspend},
+ {DM_DEV_STATUS_CMD, dev_status},
+ {DM_DEV_WAIT_CMD, dev_wait},
+
+ {DM_TABLE_LOAD_CMD, table_load},
+ {DM_TABLE_CLEAR_CMD, table_clear},
+ {DM_TABLE_DEPS_CMD, table_deps},
+ {DM_TABLE_STATUS_CMD, table_status}
+ };
+
+ return (cmd >= ARRAY_SIZE(_ioctls)) ? NULL : _ioctls[cmd].fn;
+}
+
+/*
+ * As well as checking the version compatibility this always
+ * copies the kernel interface version out.
+ */
+static int check_version(unsigned int cmd, struct dm_ioctl *user)
+{
+ __u32 version[3];
+ int r = 0;
+
+ if (copy_from_user(version, user->version, sizeof(version)))
+ return -EFAULT;
+
+ if ((DM_VERSION_MAJOR != version[0]) ||
+ (DM_VERSION_MINOR < version[1])) {
+ DMWARN("ioctl interface mismatch: "
+ "kernel(%u.%u.%u), user(%u.%u.%u), cmd(%d)",
+ DM_VERSION_MAJOR, DM_VERSION_MINOR,
+ DM_VERSION_PATCHLEVEL,
+ version[0], version[1], version[2], cmd);
+ r = -EINVAL;
+ }
+
+ /*
+ * Fill in the kernel version.
+ */
+ version[0] = DM_VERSION_MAJOR;
+ version[1] = DM_VERSION_MINOR;
+ version[2] = DM_VERSION_PATCHLEVEL;
+ if (copy_to_user(user->version, version, sizeof(version)))
+ return -EFAULT;
+
+ return r;
+}
+
+static void free_params(struct dm_ioctl *param)
+{
+ vfree(param);
+}
+
+static int copy_params(struct dm_ioctl *user, struct dm_ioctl **param)
+{
+ struct dm_ioctl tmp, *dmi;
+
+ if (copy_from_user(&tmp, user, sizeof(tmp)))
+ return -EFAULT;
+
+ if (tmp.data_size < sizeof(tmp))
+ return -EINVAL;
+
+ dmi = (struct dm_ioctl *) vmalloc(tmp.data_size);
+ if (!dmi)
+ return -ENOMEM;
+
+ if (copy_from_user(dmi, user, tmp.data_size)) {
+ vfree(dmi);
+ return -EFAULT;
+ }
+
+ *param = dmi;
+ return 0;
+}
+
+static int validate_params(uint cmd, struct dm_ioctl *param)
+{
+ /* Always clear this flag */
+ param->flags &= ~DM_BUFFER_FULL_FLAG;
+
+ /* Ignores parameters */
+ if (cmd == DM_REMOVE_ALL_CMD || cmd == DM_LIST_DEVICES_CMD)
+ return 0;
+
+ /* Unless creating, either name or uuid but not both */
+ if (cmd != DM_DEV_CREATE_CMD) {
+ if ((!*param->uuid && !*param->name) ||
+ (*param->uuid && *param->name)) {
+ DMWARN("one of name or uuid must be supplied, cmd(%u)",
+ cmd);
+ return -EINVAL;
+ }
+ }
+
+ /* Ensure strings are terminated */
+ param->name[DM_NAME_LEN - 1] = '\0';
+ param->uuid[DM_UUID_LEN - 1] = '\0';
+
+ return 0;
+}
+
+static int ctl_ioctl(struct inode *inode, struct file *file,
+ uint command, ulong u)
+{
+ int r = 0;
+ unsigned int cmd;
+ struct dm_ioctl *param;
+ struct dm_ioctl *user = (struct dm_ioctl *) u;
+ ioctl_fn fn = NULL;
+ size_t param_size;
+
+ /* only root can play with this */
+ if (!capable(CAP_SYS_ADMIN))
+ return -EACCES;
+
+ if (_IOC_TYPE(command) != DM_IOCTL)
+ return -ENOTTY;
+
+ cmd = _IOC_NR(command);
+
+ /*
+ * Check the interface version passed in. This also
+ * writes out the kernel's interface version.
+ */
+ r = check_version(cmd, user);
+ if (r)
+ return r;
+
+ /*
+ * Nothing more to do for the version command.
+ */
+ if (cmd == DM_VERSION_CMD)
+ return 0;
+
+ fn = lookup_ioctl(cmd);
+ if (!fn) {
+ DMWARN("dm_ctl_ioctl: unknown command 0x%x", command);
+ return -ENOTTY;
+ }
+
+ /*
+ * Trying to avoid low memory issues when a device is
+ * suspended.
+ */
+ current->flags |= PF_MEMALLOC;
+
+ /*
+ * Copy the parameters into kernel space.
+ */
+ r = copy_params(user, ¶m);
+ if (r) {
+ current->flags &= ~PF_MEMALLOC;
+ return r;
+ }
+
+ /*
+ * FIXME: eventually we will remove the PF_MEMALLOC flag
+ * here. However the tools still do nasty things like
+ * 'load' while a device is suspended.
+ */
+
+ r = validate_params(cmd, param);
+ if (r)
+ goto out;
+
+ param_size = param->data_size;
+ param->data_size = sizeof(*param);
+ r = fn(param, param_size);
+
+ /*
+ * Copy the results back to userland.
+ */
+ if (!r && copy_to_user(user, param, param->data_size))
+ r = -EFAULT;
+
+ out:
+ free_params(param);
+ current->flags &= ~PF_MEMALLOC;
+ return r;
+}
+
+static struct file_operations _ctl_fops = {
+ .ioctl = ctl_ioctl,
+ .owner = THIS_MODULE,
+};
+
+static struct miscdevice _dm_misc = {
+ .minor = MISC_DYNAMIC_MINOR,
+ .name = DM_NAME,
+ .devfs_name = "mapper/control",
+ .fops = &_ctl_fops
+};
+
+/*
+ * Create misc character device and link to DM_DIR/control.
+ */
+int __init dm_interface_init(void)
+{
+ int r;
+
+ r = dm_hash_init();
+ if (r)
+ return r;
+
+ r = misc_register(&_dm_misc);
+ if (r) {
+ DMERR("misc_register failed for control device");
+ dm_hash_exit();
+ return r;
+ }
+
+ DMINFO("%d.%d.%d%s initialised: %s", DM_VERSION_MAJOR,
+ DM_VERSION_MINOR, DM_VERSION_PATCHLEVEL, DM_VERSION_EXTRA,
+ DM_DRIVER_EMAIL);
+ return 0;
+}
+
+void dm_interface_exit(void)
+{
+ if (misc_deregister(&_dm_misc) < 0)
+ DMERR("misc_deregister failed for control device");
+
+ dm_hash_exit();
+}
--- diff/include/linux/dm-ioctl-v1.h 1970-01-01 01:00:00.000000000 +0100
+++ source/include/linux/dm-ioctl-v1.h 2003-07-01 15:36:42.000000000 +0100
@@ -0,0 +1,149 @@
+/*
+ * Copyright (C) 2001 Sistina Software (UK) Limited.
+ *
+ * This file is released under the LGPL.
+ */
+
+#ifndef _LINUX_DM_IOCTL_V1_H
+#define _LINUX_DM_IOCTL_V1_H
+
+#include <linux/types.h>
+
+#define DM_DIR "mapper" /* Slashes not supported */
+#define DM_MAX_TYPE_NAME 16
+#define DM_NAME_LEN 128
+#define DM_UUID_LEN 129
+
+/*
+ * Implements a traditional ioctl interface to the device mapper.
+ */
+
+/*
+ * All ioctl arguments consist of a single chunk of memory, with
+ * this structure at the start. If a uuid is specified any
+ * lookup (eg. for a DM_INFO) will be done on that, *not* the
+ * name.
+ */
+struct dm_ioctl {
+ /*
+ * The version number is made up of three parts:
+ * major - no backward or forward compatibility,
+ * minor - only backwards compatible,
+ * patch - both backwards and forwards compatible.
+ *
+ * All clients of the ioctl interface should fill in the
+ * version number of the interface that they were
+ * compiled with.
+ *
+ * All recognised ioctl commands (ie. those that don't
+ * return -ENOTTY) fill out this field, even if the
+ * command failed.
+ */
+ uint32_t version[3]; /* in/out */
+ uint32_t data_size; /* total size of data passed in
+ * including this struct */
+
+ uint32_t data_start; /* offset to start of data
+ * relative to start of this struct */
+
+ uint32_t target_count; /* in/out */
+ uint32_t open_count; /* out */
+ uint32_t flags; /* in/out */
+
+ __kernel_dev_t dev; /* in/out */
+
+ char name[DM_NAME_LEN]; /* device name */
+ char uuid[DM_UUID_LEN]; /* unique identifier for
+ * the block device */
+};
+
+/*
+ * Used to specify tables. These structures appear after the
+ * dm_ioctl.
+ */
+struct dm_target_spec {
+ int32_t status; /* used when reading from kernel only */
+ uint64_t sector_start;
+ uint32_t length;
+
+ /*
+ * Offset in bytes (from the start of this struct) to
+ * next target_spec.
+ */
+ uint32_t next;
+
+ char target_type[DM_MAX_TYPE_NAME];
+
+ /*
+ * Parameter string starts immediately after this object.
+ * Be careful to add padding after string to ensure correct
+ * alignment of subsequent dm_target_spec.
+ */
+};
+
+/*
+ * Used to retrieve the target dependencies.
+ */
+struct dm_target_deps {
+ uint32_t count;
+
+ __kernel_dev_t dev[0]; /* out */
+};
+
+/*
+ * If you change this make sure you make the corresponding change
+ * to dm-ioctl.c:lookup_ioctl()
+ */
+enum {
+ /* Top level cmds */
+ DM_VERSION_CMD = 0,
+ DM_REMOVE_ALL_CMD,
+
+ /* device level cmds */
+ DM_DEV_CREATE_CMD,
+ DM_DEV_REMOVE_CMD,
+ DM_DEV_RELOAD_CMD,
+ DM_DEV_RENAME_CMD,
+ DM_DEV_SUSPEND_CMD,
+ DM_DEV_DEPS_CMD,
+ DM_DEV_STATUS_CMD,
+
+ /* target level cmds */
+ DM_TARGET_STATUS_CMD,
+ DM_TARGET_WAIT_CMD
+};
+
+#define DM_IOCTL 0xfd
+
+#define DM_VERSION _IOWR(DM_IOCTL, DM_VERSION_CMD, struct dm_ioctl)
+#define DM_REMOVE_ALL _IOWR(DM_IOCTL, DM_REMOVE_ALL_CMD, struct dm_ioctl)
+
+#define DM_DEV_CREATE _IOWR(DM_IOCTL, DM_DEV_CREATE_CMD, struct dm_ioctl)
+#define DM_DEV_REMOVE _IOWR(DM_IOCTL, DM_DEV_REMOVE_CMD, struct dm_ioctl)
+#define DM_DEV_RELOAD _IOWR(DM_IOCTL, DM_DEV_RELOAD_CMD, struct dm_ioctl)
+#define DM_DEV_SUSPEND _IOWR(DM_IOCTL, DM_DEV_SUSPEND_CMD, struct dm_ioctl)
+#define DM_DEV_RENAME _IOWR(DM_IOCTL, DM_DEV_RENAME_CMD, struct dm_ioctl)
+#define DM_DEV_DEPS _IOWR(DM_IOCTL, DM_DEV_DEPS_CMD, struct dm_ioctl)
+#define DM_DEV_STATUS _IOWR(DM_IOCTL, DM_DEV_STATUS_CMD, struct dm_ioctl)
+
+#define DM_TARGET_STATUS _IOWR(DM_IOCTL, DM_TARGET_STATUS_CMD, struct dm_ioctl)
+#define DM_TARGET_WAIT _IOWR(DM_IOCTL, DM_TARGET_WAIT_CMD, struct dm_ioctl)
+
+#define DM_VERSION_MAJOR 1
+#define DM_VERSION_MINOR 0
+#define DM_VERSION_PATCHLEVEL 6
+#define DM_VERSION_EXTRA "-ioctl (2002-10-15)"
+
+/* Status bits */
+#define DM_READONLY_FLAG 0x00000001
+#define DM_SUSPEND_FLAG 0x00000002
+#define DM_EXISTS_FLAG 0x00000004
+#define DM_PERSISTENT_DEV_FLAG 0x00000008
+
+/*
+ * Flag passed into ioctl STATUS command to get table information
+ * rather than current status.
+ */
+#define DM_STATUS_TABLE_FLAG 0x00000010
+
+#endif /* _LINUX_DM_IOCTL_H */
--- diff/include/linux/dm-ioctl-v4.h 1970-01-01 01:00:00.000000000 +0100
+++ source/include/linux/dm-ioctl-v4.h 2003-07-01 15:36:42.000000000 +0100
@@ -0,0 +1,237 @@
+/*
+ * Copyright (C) 2001 - 2003 Sistina Software (UK) Limited.
+ *
+ * This file is released under the LGPL.
+ */
+
+#ifndef _LINUX_DM_IOCTL_V4_H
+#define _LINUX_DM_IOCTL_V4_H
+
+#include <linux/types.h>
+
+#define DM_DIR "mapper" /* Slashes not supported */
+#define DM_MAX_TYPE_NAME 16
+#define DM_NAME_LEN 128
+#define DM_UUID_LEN 129
+
+/*
+ * A traditional ioctl interface for the device mapper.
+ *
+ * Each device can have two tables associated with it, an
+ * 'active' table which is the one currently used by io passing
+ * through the device, and an 'inactive' one which is a table
+ * that is being prepared as a replacement for the 'active' one.
+ *
+ * DM_VERSION:
+ * Just get the version information for the ioctl interface.
+ *
+ * DM_REMOVE_ALL:
+ * Remove all dm devices, destroy all tables. Only really used
+ * for debug.
+ *
+ * DM_LIST_DEVICES:
+ * Get a list of all the dm device names.
+ *
+ * DM_DEV_CREATE:
+ * Create a new device, neither the 'active' or 'inactive' table
+ * slots will be filled. The device will be in suspended state
+ * after creation, however any io to the device will get errored
+ * since it will be out-of-bounds.
+ *
+ * DM_DEV_REMOVE:
+ * Remove a device, destroy any tables.
+ *
+ * DM_DEV_RENAME:
+ * Rename a device.
+ *
+ * DM_SUSPEND:
+ * This performs both suspend and resume, depending which flag is
+ * passed in.
+ * Suspend: This command will not return until all pending io to
+ * the device has completed. Further io will be deferred until
+ * the device is resumed.
+ * Resume: It is no longer an error to issue this command on an
+ * unsuspended device. If a table is present in the 'inactive'
+ * slot, it will be moved to the active slot, then the old table
+ * from the active slot will be _destroyed_. Finally the device
+ * is resumed.
+ *
+ * DM_DEV_STATUS:
+ * Retrieves the status for the table in the 'active' slot.
+ *
+ * DM_DEV_WAIT:
+ * Wait for a significant event to occur to the device. This
+ * could either be caused by an event triggered by one of the
+ * targets of the table in the 'active' slot, or a table change.
+ *
+ * DM_TABLE_LOAD:
+ * Load a table into the 'inactive' slot for the device. The
+ * device does _not_ need to be suspended prior to this command.
+ *
+ * DM_TABLE_CLEAR:
+ * Destroy any table in the 'inactive' slot (ie. abort).
+ *
+ * DM_TABLE_DEPS:
+ * Return a set of device dependencies for the 'active' table.
+ *
+ * DM_TABLE_STATUS:
+ * Return the targets status for the 'active' table.
+ */
+
+/*
+ * All ioctl arguments consist of a single chunk of memory, with
+ * this structure at the start. If a uuid is specified any
+ * lookup (eg. for a DM_INFO) will be done on that, *not* the
+ * name.
+ */
+struct dm_ioctl {
+ /*
+ * The version number is made up of three parts:
+ * major - no backward or forward compatibility,
+ * minor - only backwards compatible,
+ * patch - both backwards and forwards compatible.
+ *
+ * All clients of the ioctl interface should fill in the
+ * version number of the interface that they were
+ * compiled with.
+ *
+ * All recognised ioctl commands (ie. those that don't
+ * return -ENOTTY) fill out this field, even if the
+ * command failed.
+ */
+ __u32 version[3]; /* in/out */
+ __u32 data_size; /* total size of data passed in
+ * including this struct */
+
+ __u32 data_start; /* offset to start of data
+ * relative to start of this struct */
+
+ __u32 target_count; /* in/out */
+ __s32 open_count; /* out */
+ __u32 flags; /* in/out */
+ __u32 event_nr; /* in/out */
+ __u32 padding;
+
+ __u64 dev; /* in/out */
+
+ char name[DM_NAME_LEN]; /* device name */
+ char uuid[DM_UUID_LEN]; /* unique identifier for
+ * the block device */
+};
+
+/*
+ * Used to specify tables. These structures appear after the
+ * dm_ioctl.
+ */
+struct dm_target_spec {
+ __u64 sector_start;
+ __u64 length;
+ __s32 status; /* used when reading from kernel only */
+
+ /*
+ * Offset in bytes (from the start of this struct) to
+ * next target_spec.
+ */
+ __u32 next;
+
+ char target_type[DM_MAX_TYPE_NAME];
+
+ /*
+ * Parameter string starts immediately after this object.
+ * Be careful to add padding after string to ensure correct
+ * alignment of subsequent dm_target_spec.
+ */
+};
+
+/*
+ * Used to retrieve the target dependencies.
+ */
+struct dm_target_deps {
+ __u32 count; /* Array size */
+ __u32 padding; /* unused */
+ __u64 dev[0]; /* out */
+};
+
+/*
+ * Used to get a list of all dm devices.
+ */
+struct dm_name_list {
+ __u64 dev;
+ __u32 next; /* offset to the next record from
+ the _start_ of this */
+ char name[0];
+};
+
+/*
+ * If you change this make sure you make the corresponding change
+ * to dm-ioctl.c:lookup_ioctl()
+ */
+enum {
+ /* Top level cmds */
+ DM_VERSION_CMD = 0,
+ DM_REMOVE_ALL_CMD,
+ DM_LIST_DEVICES_CMD,
+
+ /* device level cmds */
+ DM_DEV_CREATE_CMD,
+ DM_DEV_REMOVE_CMD,
+ DM_DEV_RENAME_CMD,
+ DM_DEV_SUSPEND_CMD,
+ DM_DEV_STATUS_CMD,
+ DM_DEV_WAIT_CMD,
+
+ /* Table level cmds */
+ DM_TABLE_LOAD_CMD,
+ DM_TABLE_CLEAR_CMD,
+ DM_TABLE_DEPS_CMD,
+ DM_TABLE_STATUS_CMD,
+};
+
+#define DM_IOCTL 0xfd
+
+#define DM_VERSION _IOWR(DM_IOCTL, DM_VERSION_CMD, struct dm_ioctl)
+#define DM_REMOVE_ALL _IOWR(DM_IOCTL, DM_REMOVE_ALL_CMD, struct dm_ioctl)
+#define DM_LIST_DEVICES _IOWR(DM_IOCTL, DM_LIST_DEVICES_CMD, struct dm_ioctl)
+
+#define DM_DEV_CREATE _IOWR(DM_IOCTL, DM_DEV_CREATE_CMD, struct dm_ioctl)
+#define DM_DEV_REMOVE _IOWR(DM_IOCTL, DM_DEV_REMOVE_CMD, struct dm_ioctl)
+#define DM_DEV_RENAME _IOWR(DM_IOCTL, DM_DEV_RENAME_CMD, struct dm_ioctl)
+#define DM_DEV_SUSPEND _IOWR(DM_IOCTL, DM_DEV_SUSPEND_CMD, struct dm_ioctl)
+#define DM_DEV_STATUS _IOWR(DM_IOCTL, DM_DEV_STATUS_CMD, struct dm_ioctl)
+#define DM_DEV_WAIT _IOWR(DM_IOCTL, DM_DEV_WAIT_CMD, struct dm_ioctl)
+
+#define DM_TABLE_LOAD _IOWR(DM_IOCTL, DM_TABLE_LOAD_CMD, struct dm_ioctl)
+#define DM_TABLE_CLEAR _IOWR(DM_IOCTL, DM_TABLE_CLEAR_CMD, struct dm_ioctl)
+#define DM_TABLE_DEPS _IOWR(DM_IOCTL, DM_TABLE_DEPS_CMD, struct dm_ioctl)
+#define DM_TABLE_STATUS _IOWR(DM_IOCTL, DM_TABLE_STATUS_CMD, struct dm_ioctl)
+
+#define DM_VERSION_MAJOR 4
+#define DM_VERSION_MINOR 0
+#define DM_VERSION_PATCHLEVEL 0
+#define DM_VERSION_EXTRA "-ioctl (2003-06-04)"
+
+/* Status bits */
+#define DM_READONLY_FLAG (1 << 0) /* In/Out */
+#define DM_SUSPEND_FLAG (1 << 1) /* In/Out */
+#define DM_PERSISTENT_DEV_FLAG (1 << 3) /* In */
+
+/*
+ * Flag passed into ioctl STATUS command to get table information
+ * rather than current status.
+ */
+#define DM_STATUS_TABLE_FLAG (1 << 4) /* In */
+
+/*
+ * Flags that indicate whether a table is present in either of
+ * the two table slots that a device has.
+ */
+#define DM_ACTIVE_PRESENT_FLAG (1 << 5) /* Out */
+#define DM_INACTIVE_PRESENT_FLAG (1 << 6) /* Out */
+
+/*
+ * Indicates that the buffer passed in wasn't big enough for the
+ * results.
+ */
+#define DM_BUFFER_FULL_FLAG (1 << 8) /* Out */
+
+#endif /* _LINUX_DM_IOCTL_H */
There was a race associated with the 'wait for a significant event'
functionality. Basically userland could read the status table, then
wait for another event, but the event it was waiting for could have
occurred in the gap between reading and waiting. To solve this we
assign identifiers to events, in order to successfully wait for an
event both userland and the kernel driver must be in agreement about
what the last event identifier was. If they don't agree the wait call
will return immediately, allowing userland to re-read the status and
see what it missed. The new ioctl interface will use this properly.
--- diff/drivers/md/dm-ioctl.c 2003-06-30 11:39:38.000000000 +0100
+++ source/drivers/md/dm-ioctl.c 2003-06-30 17:12:49.000000000 +0100
@@ -706,7 +706,6 @@
static int wait_device_event(struct dm_ioctl *param, struct dm_ioctl *user)
{
struct mapped_device *md;
- struct dm_table *table;
DECLARE_WAITQUEUE(wq, current);
md = find_device(param);
@@ -725,13 +724,12 @@
* Wait for a notification event
*/
set_current_state(TASK_INTERRUPTIBLE);
- table = dm_get_table(md);
- dm_table_add_wait_queue(table, &wq);
- dm_table_put(table);
- dm_put(md);
-
- yield();
+ if (!dm_add_wait_queue(md, &wq, dm_get_event_nr(md))) {
+ yield();
+ dm_remove_wait_queue(md, &wq);
+ }
set_current_state(TASK_RUNNING);
+ dm_put(md);
out:
return results_to_user(user, param, NULL, 0);
--- diff/drivers/md/dm-table.c 2003-06-30 17:03:17.000000000 +0100
+++ source/drivers/md/dm-table.c 2003-06-30 17:03:09.000000000 +0100
@@ -48,11 +48,9 @@
*/
struct io_restrictions limits;
- /*
- * A waitqueue for processes waiting for something
- * interesting to happen to this table.
- */
- wait_queue_head_t eventq;
+ /* events get handed up using this callback */
+ void (*event_fn)(void *);
+ void *event_context;
};
/*
@@ -222,7 +220,6 @@
return -ENOMEM;
}
- init_waitqueue_head(&t->eventq);
t->mode = mode;
*result = t;
return 0;
@@ -243,9 +240,6 @@
{
unsigned int i;
- /* destroying the table counts as an event */
- dm_table_event(t);
-
/* free the indexes (see dm_table_complete) */
if (t->depth >= 2)
vfree(t->index[t->depth - 2]);
@@ -694,9 +688,22 @@
return r;
}
+static spinlock_t _event_lock = SPIN_LOCK_UNLOCKED;
+void dm_table_event_callback(struct dm_table *t,
+ void (*fn)(void *), void *context)
+{
+ spin_lock_irq(&_event_lock);
+ t->event_fn = fn;
+ t->event_context = context;
+ spin_unlock_irq(&_event_lock);
+}
+
void dm_table_event(struct dm_table *t)
{
- wake_up_interruptible(&t->eventq);
+ spin_lock(&_event_lock);
+ if (t->event_fn)
+ t->event_fn(t->event_context);
+ spin_unlock(&_event_lock);
}
sector_t dm_table_get_size(struct dm_table *t)
@@ -761,11 +768,6 @@
return t->mode;
}
-void dm_table_add_wait_queue(struct dm_table *t, wait_queue_t *wq)
-{
- add_wait_queue(&t->eventq, wq);
-}
-
void dm_table_suspend_targets(struct dm_table *t)
{
int i;
--- diff/drivers/md/dm.c 2003-06-30 17:03:17.000000000 +0100
+++ source/drivers/md/dm.c 2003-06-30 17:12:16.000000000 +0100
@@ -63,6 +63,12 @@
* io objects are allocated from here.
*/
mempool_t *io_pool;
+
+ /*
+ * Event handling.
+ */
+ __u32 event_nr;
+ wait_queue_head_t eventq;
};
#define MIN_IOS 256
@@ -619,6 +625,8 @@
atomic_set(&md->pending, 0);
init_waitqueue_head(&md->wait);
+ init_waitqueue_head(&md->eventq);
+
return md;
}
@@ -634,6 +642,16 @@
/*
* Bind a table to the device.
*/
+static void event_callback(void *context)
+{
+ struct mapped_device *md = (struct mapped_device *) context;
+
+ down_write(&md->lock);
+ md->event_nr++;
+ wake_up_interruptible(&md->eventq);
+ up_write(&md->lock);
+}
+
static int __bind(struct mapped_device *md, struct dm_table *t)
{
request_queue_t *q = &md->queue;
@@ -645,6 +663,8 @@
if (size == 0)
return 0;
+ dm_table_event_callback(md->map, event_callback, md);
+
dm_table_get(t);
dm_table_set_restrictions(t, q);
return 0;
@@ -652,6 +672,7 @@
static void __unbind(struct mapped_device *md)
{
+ dm_table_event_callback(md->map, NULL, NULL);
dm_table_put(md->map);
md->map = NULL;
set_capacity(md->disk, 0);
@@ -821,6 +842,42 @@
return 0;
}
+/*-----------------------------------------------------------------
+ * Event notification.
+ *---------------------------------------------------------------*/
+__u32 dm_get_event_nr(struct mapped_device *md)
+{
+ __u32 r;
+
+ down_read(&md->lock);
+ r = md->event_nr;
+ up_read(&md->lock);
+
+ return r;
+}
+
+int dm_add_wait_queue(struct mapped_device *md, wait_queue_t *wq,
+ __u32 event_nr)
+{
+ down_write(&md->lock);
+ if (event_nr != md->event_nr) {
+ up_write(&md->lock);
+ return 1;
+ }
+
+ add_wait_queue(&md->eventq, wq);
+ up_write(&md->lock);
+
+ return 0;
+}
+
+void dm_remove_wait_queue(struct mapped_device *md, wait_queue_t *wq)
+{
+ down_write(&md->lock);
+ remove_wait_queue(&md->eventq, wq);
+ up_write(&md->lock);
+}
+
/*
* The gendisk is only valid as long as you have a reference
* count on 'md'.
--- diff/drivers/md/dm.h 2003-06-30 17:03:17.000000000 +0100
+++ source/drivers/md/dm.h 2003-06-30 17:11:07.000000000 +0100
@@ -79,6 +79,14 @@
struct dm_table *dm_get_table(struct mapped_device *md);
/*
+ * Event functions.
+ */
+__u32 dm_get_event_nr(struct mapped_device *md);
+int dm_add_wait_queue(struct mapped_device *md, wait_queue_t *wq,
+ __u32 event_nr);
+void dm_remove_wait_queue(struct mapped_device *md, wait_queue_t *wq);
+
+/*
* Info functions.
*/
struct gendisk *dm_disk(struct mapped_device *md);
@@ -96,6 +104,8 @@
int dm_table_add_target(struct dm_table *t, const char *type,
sector_t start, sector_t len, char *params);
int dm_table_complete(struct dm_table *t);
+void dm_table_event_callback(struct dm_table *t,
+ void (*fn)(void *), void *context);
void dm_table_event(struct dm_table *t);
sector_t dm_table_get_size(struct dm_table *t);
struct dm_target *dm_table_get_target(struct dm_table *t, unsigned int index);
@@ -104,7 +114,6 @@
unsigned int dm_table_get_num_targets(struct dm_table *t);
struct list_head *dm_table_get_devices(struct dm_table *t);
int dm_table_get_mode(struct dm_table *t);
-void dm_table_add_wait_queue(struct dm_table *t, wait_queue_t *wq);
void dm_table_suspend_targets(struct dm_table *t);
void dm_table_resume_targets(struct dm_table *t);
On Tuesday 01 July 2003 10:02, Joe Thornber wrote:
> --- diff/drivers/md/dm-ioctl-v4.c 1970-01-01 01:00:00.000000000 +0100
> +++ source/drivers/md/dm-ioctl-v4.c 2003-07-01 15:36:42.000000000 +0100
> +#define NUM_BUCKETS 64
> +#define MASK_BUCKETS (NUM_BUCKETS - 1)
> +static struct list_head _name_buckets[NUM_BUCKETS];
> +static struct list_head _uuid_buckets[NUM_BUCKETS];
> +
> +void dm_hash_remove_all(void);
> +
> +/*
> + * Guards access to all three tables.
Guards access to both tables. (The third table used to be for looking up
devices based on kdev_t.)
> + */
> +static DECLARE_RWSEM(_hash_lock);
> +
> +/*
> + * The kdev_t and uuid of a device can never change once it is
> + * initially inserted.
> + */
> +int dm_hash_insert(const char *name, const char *uuid, struct
> mapped_device *md) +{
> + struct hash_cell *cell;
> +
> + /*
> + * Allocate the new cells.
> + */
> + cell = alloc_cell(name, uuid, md);
> + if (!cell)
> + return -ENOMEM;
> +
> + /*
> + * Insert the cell into all three hash tables.
Again, just two tables. :)
> + */
> + down_write(&_hash_lock);
> + if (__get_name_cell(name))
> + goto bad;
> +
> +int dm_hash_rename(const char *old, const char *new)
> +{
> + char *new_name, *old_name;
> + struct hash_cell *hc;
> +
> + /*
> + * duplicate new.
> + */
> + new_name = kstrdup(new);
> + if (!new_name)
> + return -ENOMEM;
> +
> + down_write(&_hash_lock);
> +
> + /*
> + * Is new free ?
> + */
> + hc = __get_name_cell(new);
> + if (hc) {
> + DMWARN("asked to rename to an already existing name %s -> %s",
> + old, new);
> + up_write(&_hash_lock);
> + kfree(new_name);
> + return -EBUSY;
> + }
> +
> + /*
> + * Is there such a device as 'old' ?
> + */
> + hc = __get_name_cell(old);
> + if (!hc) {
> + DMWARN("asked to rename a non existent device %s -> %s",
> + old, new);
> + up_write(&_hash_lock);
> + kfree(new_name);
> + return -ENXIO;
> + }
> +
> + /*
> + * rename and move the name cell.
> + */
> + list_del(&hc->name_list);
> + old_name = hc->name;
> + hc->name = new_name;
> + list_add(&hc->name_list, _name_buckets + hash_str(new_name));
> +
> + /* rename the device node in devfs */
> + unregister_with_devfs(hc);
The "unregister" call needs to be before the actual rename. Same patch as a
couple weeks ago.
> + register_with_devfs(hc);
> +
> + up_write(&_hash_lock);
> + kfree(old_name);
> + return 0;
> +}
> +static int check_name(const char *name)
> +{
> + if (strchr(name, '/')) {
> + DMWARN("invalid device name");
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
Can't we allow slashes in device names? I thought we discussed this before
(http://marc.theaimsgroup.com/?t=104628092700011&r=1&w=2). Any reason for the
change?
> +static int resume(struct dm_ioctl *param)
> +{
> + int r = 0;
> + struct hash_cell *hc;
> + struct mapped_device *md;
> + struct dm_table *new_map;
> +
> + down_write(&_hash_lock);
> +
> + hc = __find_device_hash_cell(param);
> + if (!hc) {
> + DMWARN("device doesn't appear to be in the dev hash table.");
> + up_write(&_hash_lock);
> + return -ENXIO;
> + }
> +
> + md = hc->md;
> + dm_get(md);
> +
> + new_map = hc->new_map;
> + hc->new_map = NULL;
> + param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
> +
> + up_write(&_hash_lock);
> +
> + /* Do we need to load a new map ? */
> + if (new_map) {
> + /* Suspend if it isn't already suspended */
> + if (!dm_suspended(md))
> + dm_suspend(md);
> +
> + r = dm_swap_table(md, new_map);
> + if (r) {
> + dm_put(md);
> + dm_table_put(new_map);
> + return r;
> + }
Does this imply that if the dm_swap_table() call fails, then the "inactive"
mapping is automatically deleted? Although....looking at the dm_swap_table()
code, it looks like it can't actually fail (the only reason it could fail is
if the device wasn't suspended, which is verified before calling
dm_swap_table - and the _hash_lock prevent another process from trying to
unsuspend the device in-between the calls).
As a side note, the __bind() function in dm.c currently will never return an
error, so dm_swap_table() doesn't necessarily need to check for one.
> +
> + if (!(dm_table_get_mode(new_map) & FMODE_WRITE))
> + set_disk_ro(dm_disk(md), 1);
> +
> + dm_table_put(new_map);
> + }
> +
> + if (dm_suspended(md))
> + r = dm_resume(md);
> +
> + if (!r)
> + r = __dev_status(md, param);
> +
> + dm_put(md);
> + return r;
> +}
> +
On first pass through this file, everything looks pretty good. As long as the
actual definition of the interface isn't going to change, I'd say go ahead
and include this. After it's merged, we'll put together a version of EVMS
that will support the new interface.
--
Kevin Corry
[email protected]
http://evms.sourceforge.net/
On Tue, Jul 01, 2003 at 03:05:07PM -0500, Kevin Corry wrote:
> The "unregister" call needs to be before the actual rename. Same patch as a
> couple weeks ago.
Agreed.
> > +static int check_name(const char *name)
> > +{
> > + if (strchr(name, '/')) {
> > + DMWARN("invalid device name");
> > + return -EINVAL;
> > + }
> > +
> > + return 0;
> > +}
>
> Can't we allow slashes in device names? I thought we discussed this before
> (http://marc.theaimsgroup.com/?t=104628092700011&r=1&w=2). Any reason for the
> change?
I think I made the wrong decision before. Still thinking about it though.
> Does this imply that if the dm_swap_table() call fails, then the "inactive"
> mapping is automatically deleted?
Yes, that is the behaviour ATM. Would you rather it didn't ?
> As a side note, the __bind() function in dm.c currently will never return an
> error, so dm_swap_table() doesn't necessarily need to check for one.
y, __bind can become void.
- Joe
Clear the read-only flag if the table is not read only.
--- diff/drivers/md/dm-ioctl-v4.c 2003-07-01 15:36:42.000000000 +0100
+++ source/drivers/md/dm-ioctl-v4.c 2003-07-02 11:27:50.000000000 +0100
@@ -650,7 +650,9 @@
return r;
}
- if (!(dm_table_get_mode(new_map) & FMODE_WRITE))
+ if (dm_table_get_mode(new_map) & FMODE_WRITE)
+ set_disk_ro(dm_disk(md), 0);
+ else
set_disk_ro(dm_disk(md), 1);
dm_table_put(new_map);
Correct comments.
--- diff/drivers/md/dm-ioctl-v4.c 2003-07-02 11:31:05.000000000 +0100
+++ source/drivers/md/dm-ioctl-v4.c 2003-07-02 11:30:56.000000000 +0100
@@ -41,7 +41,7 @@
void dm_hash_remove_all(void);
/*
- * Guards access to all three tables.
+ * Guards access to both hash tables.
*/
static DECLARE_RWSEM(_hash_lock);
@@ -205,7 +205,7 @@
return -ENOMEM;
/*
- * Insert the cell into all three hash tables.
+ * Insert the cell into both hash tables.
*/
down_write(&_hash_lock);
if (__get_name_cell(name))
dm_swap_table() will now fail for a table with no targets.
--- diff/drivers/md/dm.c 2003-07-01 15:36:42.000000000 +0100
+++ source/drivers/md/dm.c 2003-07-02 11:53:22.000000000 +0100
@@ -664,10 +664,10 @@
md->map = t;
size = dm_table_get_size(t);
- set_capacity(md->disk, size);
- if (size == 0)
- return 0;
+ if (!size)
+ return -EINVAL;
+ set_capacity(md->disk, size);
dm_table_event_callback(md->map, event_callback, md);
dm_table_get(t);
@@ -759,8 +759,10 @@
__unbind(md);
r = __bind(md, table);
- if (r)
+ if (r) {
+ up_write(&md->lock);
return r;
+ }
up_write(&md->lock);
return 0;
If dm_add_wait_queue() fails we need to set TASK_RUNNING.
--- diff/drivers/md/dm-ioctl-v4.c 2003-07-02 11:32:42.000000000 +0100
+++ source/drivers/md/dm-ioctl-v4.c 2003-07-02 11:32:23.000000000 +0100
@@ -719,6 +719,7 @@
schedule();
dm_remove_wait_queue(md, &wq);
}
+ set_current_state(TASK_RUNNING);
/*
* The userland program is going to want to know what
Move unregister_with_devfs() to before the rename.
--- diff/drivers/md/dm-ioctl-v4.c 2003-07-02 11:35:31.000000000 +0100
+++ source/drivers/md/dm-ioctl-v4.c 2003-07-02 11:34:24.000000000 +0100
@@ -301,13 +301,14 @@
/*
* rename and move the name cell.
*/
+ unregister_with_devfs(hc);
+
list_del(&hc->name_list);
old_name = hc->name;
hc->name = new_name;
list_add(&hc->name_list, _name_buckets + hash_str(new_name));
/* rename the device node in devfs */
- unregister_with_devfs(hc);
register_with_devfs(hc);
up_write(&_hash_lock);
This definitely seems to make more sense.
On Wednesday 02 July 2003 06:00, Joe Thornber wrote:
> dm_swap_table() will now fail for a table with no targets.
> --- diff/drivers/md/dm.c 2003-07-01 15:36:42.000000000 +0100
> +++ source/drivers/md/dm.c 2003-07-02 11:53:22.000000000 +0100
> @@ -664,10 +664,10 @@
> md->map = t;
>
> size = dm_table_get_size(t);
> - set_capacity(md->disk, size);
> - if (size == 0)
> - return 0;
> + if (!size)
> + return -EINVAL;
>
> + set_capacity(md->disk, size);
> dm_table_event_callback(md->map, event_callback, md);
>
> dm_table_get(t);
> @@ -759,8 +759,10 @@
>
> __unbind(md);
> r = __bind(md, table);
> - if (r)
> + if (r) {
> + up_write(&md->lock);
> return r;
> + }
>
> up_write(&md->lock);
> return 0;
>
--
Kevin Corry
[email protected]
http://evms.sourceforge.net/
On Wednesday 02 July 2003 03:59, Joe Thornber wrote:
> On Tue, Jul 01, 2003 at 03:05:07PM -0500, Kevin Corry wrote:
> > > +static int check_name(const char *name)
> > > +{
> > > + if (strchr(name, '/')) {
> > > + DMWARN("invalid device name");
> > > + return -EINVAL;
> > > + }
> > > +
> > > + return 0;
> > > +}
> >
> > Can't we allow slashes in device names? I thought we discussed this
> > before (http://marc.theaimsgroup.com/?t=104628092700011&r=1&w=2). Any
> > reason for the change?
>
> I think I made the wrong decision before. Still thinking about it though.
Well, what are your current thoughts?
I guess the way I see it is that this name is simply a way for user-space to
identify a device to the driver, and it seems kind of unnecessary for the
driver to put restrictions on that name.
The only reason I can really think of to restrict the contents of this name is
due to using the name for devfs. But devfs doesn't seem to mind dealing with
names with embedded slashes (I always run with devfs enabled, and I've never
seen it complain).
On a side note, looking through the devfs code reveals a potential error. In
DM, names can be 127 characters. But devfs_mk_bdev() (in fs/devfs/base.c)
will only accept names up to 63 characters. If it gets a name longer than
that, it just returns an error, which DM will ignore. This isn't a fatal
error by any means, but if any user-space tools rely on those devfs devices
being created, there may be cases where the tools won't work correctly.
> > Does this imply that if the dm_swap_table() call fails, then the
> > "inactive" mapping is automatically deleted?
>
> Yes, that is the behaviour ATM. Would you rather it didn't ?
Well...I can't say that it makes much difference to me either way at the
moment. :) I just wanted to clarify the behavior, since it wasn't explicitly
stated anywhere.
--
Kevin Corry
[email protected]
http://evms.sourceforge.net/
On Tue, Jul 01, 2003 at 03:58:12PM +0100, Joe Thornber wrote:
> Following on from the header file for the v4 ioctl interface that I
> posted a couple of weeks ago, here is the first cut at the
> implementation (3 patches posted as a follow up to this mail). I hope
> the v1 interface can be retired before 2.6. Tools are not yet
> available to drive this, but should be later this week.
Updated device-mapper tools (dmsetup + libdevmapper) are now
available for testing at:
ftp://ftp.sistina.com/pub/LVM2/device-mapper/device-mapper-testing-new-version4-interface.tgz
This tarball also includes the complete device-mapper patches
for 2.4.20 and 2.4.21, also available outside the tarball at:
ftp://ftp.sistina.com/pub/LVM2/device-mapper/patches/combined*
Updated LVM2 tools to work alongside the above are at:
ftp://ftp.sistina.com/pub/LVM2/tools/LVM2.0-testing.tgz
If the new tools detect the old version 1 driver, they try to
fall back and use the version 1 API, so you shouldn't
need to keep two sets of tools around if you're regularly
swapping kernels. [But this hasn't been thoroughly tested.]
Alasdair
--
[email protected]