linux-stable/drivers/md/dm-ioctl.c

2154 lines
48 KiB
C
Raw Normal View History

/*
* Copyright (C) 2001, 2002 Sistina Software (UK) Limited.
* Copyright (C) 2004 - 2006 Red Hat, Inc. All rights reserved.
*
* This file is released under the GPL.
*/
#include "dm-core.h"
#include <linux/module.h>
#include <linux/vmalloc.h>
#include <linux/miscdevice.h>
#include <linux/sched/mm.h>
#include <linux/init.h>
#include <linux/wait.h>
#include <linux/slab.h>
#include <linux/dm-ioctl.h>
#include <linux/hdreg.h>
#include <linux/compat.h>
#include <linux/uaccess.h>
#define DM_MSG_PREFIX "ioctl"
#define DM_DRIVER_EMAIL "dm-devel@redhat.com"
struct dm_file {
/*
* poll will wait until the global event number is greater than
* this value.
*/
volatile unsigned global_event_nr;
};
/*-----------------------------------------------------------------
* 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;
};
struct vers_iter {
size_t param_size;
struct dm_target_versions *vers, *old_vers;
char *end;
uint32_t flags;
};
#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];
static void dm_hash_remove_all(bool keep_open_devices, bool mark_deferred, bool only_deferred);
/*
* Guards access to both hash tables.
*/
static DECLARE_RWSEM(_hash_lock);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-10 23:51:52 +00:00
/*
* Protects use of mdptr to obtain hash cell name and uuid from mapped device.
*/
static DEFINE_MUTEX(dm_hash_cells_mutex);
static void init_buckets(struct list_head *buckets)
{
unsigned int i;
for (i = 0; i < NUM_BUCKETS; i++)
INIT_LIST_HEAD(buckets + i);
}
static int dm_hash_init(void)
{
init_buckets(_name_buckets);
init_buckets(_uuid_buckets);
return 0;
}
static void dm_hash_exit(void)
{
dm_hash_remove_all(false, false, false);
}
/*-----------------------------------------------------------------
* 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 hash_cell *hc;
unsigned int h = hash_str(str);
list_for_each_entry (hc, _name_buckets + h, name_list)
if (!strcmp(hc->name, str)) {
dm_get(hc->md);
return hc;
}
return NULL;
}
static struct hash_cell *__get_uuid_cell(const char *str)
{
struct hash_cell *hc;
unsigned int h = hash_str(str);
list_for_each_entry (hc, _uuid_buckets + h, uuid_list)
if (!strcmp(hc->uuid, str)) {
dm_get(hc->md);
return hc;
}
return NULL;
}
static struct hash_cell *__get_dev_cell(uint64_t dev)
{
struct mapped_device *md;
struct hash_cell *hc;
md = dm_get_md(huge_decode_dev(dev));
if (!md)
return NULL;
hc = dm_get_mdptr(md);
if (!hc) {
dm_put(md);
return NULL;
}
return hc;
}
/*-----------------------------------------------------------------
* Inserting, removing and renaming a device.
*---------------------------------------------------------------*/
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, GFP_KERNEL);
if (!hc->name) {
kfree(hc);
return NULL;
}
if (!uuid)
hc->uuid = NULL;
else {
hc->uuid = kstrdup(uuid, GFP_KERNEL);
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);
}
}
/*
* The kdev_t and uuid of a device can never change once it is
* initially inserted.
*/
static int dm_hash_insert(const char *name, const char *uuid, struct mapped_device *md)
{
struct hash_cell *cell, *hc;
/*
* Allocate the new cells.
*/
cell = alloc_cell(name, uuid, md);
if (!cell)
return -ENOMEM;
/*
* Insert the cell into both hash tables.
*/
down_write(&_hash_lock);
hc = __get_name_cell(name);
if (hc) {
dm_put(hc->md);
goto bad;
}
list_add(&cell->name_list, _name_buckets + hash_str(name));
if (uuid) {
hc = __get_uuid_cell(uuid);
if (hc) {
list_del(&cell->name_list);
dm_put(hc->md);
goto bad;
}
list_add(&cell->uuid_list, _uuid_buckets + hash_str(uuid));
}
dm_get(md);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-10 23:51:52 +00:00
mutex_lock(&dm_hash_cells_mutex);
dm_set_mdptr(md, cell);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-10 23:51:52 +00:00
mutex_unlock(&dm_hash_cells_mutex);
up_write(&_hash_lock);
return 0;
bad:
up_write(&_hash_lock);
free_cell(cell);
return -EBUSY;
}
static struct dm_table *__hash_remove(struct hash_cell *hc)
{
struct dm_table *table;
int srcu_idx;
/* remove from the dev hash */
list_del(&hc->uuid_list);
list_del(&hc->name_list);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-10 23:51:52 +00:00
mutex_lock(&dm_hash_cells_mutex);
dm_set_mdptr(hc->md, NULL);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-10 23:51:52 +00:00
mutex_unlock(&dm_hash_cells_mutex);
table = dm_get_live_table(hc->md, &srcu_idx);
if (table)
dm_table_event(table);
dm_put_live_table(hc->md, srcu_idx);
table = NULL;
if (hc->new_map)
table = hc->new_map;
dm_put(hc->md);
free_cell(hc);
return table;
}
static void dm_hash_remove_all(bool keep_open_devices, bool mark_deferred, bool only_deferred)
{
int i, dev_skipped;
struct hash_cell *hc;
struct mapped_device *md;
struct dm_table *t;
retry:
dev_skipped = 0;
down_write(&_hash_lock);
for (i = 0; i < NUM_BUCKETS; i++) {
list_for_each_entry(hc, _name_buckets + i, name_list) {
md = hc->md;
dm_get(md);
if (keep_open_devices &&
dm_lock_for_deletion(md, mark_deferred, only_deferred)) {
dm_put(md);
dev_skipped++;
continue;
}
t = __hash_remove(hc);
up_write(&_hash_lock);
if (t) {
dm_sync_table(md);
dm_table_destroy(t);
}
dm_put(md);
dm: separate device deletion from dm_put This patch separates the device deletion code from dm_put() to make sure the deletion happens in the process context. By this patch, device deletion always occurs in an ioctl (process) context and dm_put() can be called in interrupt context. As a result, the request-based dm's bad dm_put() usage pointed out by Mikulas below disappears. http://marc.info/?l=dm-devel&m=126699981019735&w=2 Without this patch, I confirmed there is a case to crash the system: dm_put() => dm_table_destroy() => vfree() => BUG_ON(in_interrupt()) Some more backgrounds and details: In request-based dm, a device opener can remove a mapped_device while the last request is still completing, because bios in the last request complete first and then the device opener can close and remove the mapped_device before the last request completes: CPU0 CPU1 ================================================================= <<INTERRUPT>> blk_end_request_all(clone_rq) blk_update_request(clone_rq) bio_endio(clone_bio) == end_clone_bio blk_update_request(orig_rq) bio_endio(orig_bio) <<I/O completed>> dm_blk_close() dev_remove() dm_put(md) <<Free md>> blk_finish_request(clone_rq) .... dm_end_request(clone_rq) free_rq_clone(clone_rq) blk_end_request_all(orig_rq) rq_completed(md) So request-based dm used dm_get()/dm_put() to hold md for each I/O until its request completion handling is fully done. However, the final dm_put() can call the device deletion code which must not be run in interrupt context and may cause kernel panic. To solve the problem, this patch moves the device deletion code, dm_destroy(), to predetermined places that is actually deleting the mapped_device in ioctl (process) context, and changes dm_put() just to decrement the reference count of the mapped_device. By this change, dm_put() can be used in any context and the symmetric model below is introduced: dm_create(): create a mapped_device dm_destroy(): destroy a mapped_device dm_get(): increment the reference count of a mapped_device dm_put(): decrement the reference count of a mapped_device dm_destroy() waits for all references of the mapped_device to disappear, then deletes the mapped_device. dm_destroy() uses active waiting with msleep(1), since deleting the mapped_device isn't performance-critical task. And since at this point, nobody opens the mapped_device and no new reference will be taken, the pending counts are just for racing completing activity and will eventually decrease to zero. For the unlikely case of the forced module unload, dm_destroy_immediate(), which doesn't wait and forcibly deletes the mapped_device, is also introduced and used in dm_hash_remove_all(). Otherwise, "rmmod -f" may be stuck and never return. And now, because the mapped_device is deleted at this point, subsequent accesses to the mapped_device may cause NULL pointer references. Cc: stable@kernel.org Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com> Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 03:13:56 +00:00
if (likely(keep_open_devices))
dm_destroy(md);
else
dm_destroy_immediate(md);
/*
* Some mapped devices may be using other mapped
* devices, so repeat until we make no further
* progress. If a new mapped device is created
* here it will also get removed.
*/
goto retry;
}
}
up_write(&_hash_lock);
if (dev_skipped)
DMWARN("remove_all left %d open device(s)", dev_skipped);
}
/*
* Set the uuid of a hash_cell that isn't already set.
*/
static void __set_cell_uuid(struct hash_cell *hc, char *new_uuid)
{
mutex_lock(&dm_hash_cells_mutex);
hc->uuid = new_uuid;
mutex_unlock(&dm_hash_cells_mutex);
list_add(&hc->uuid_list, _uuid_buckets + hash_str(new_uuid));
}
/*
* Changes the name of a hash_cell and returns the old name for
* the caller to free.
*/
static char *__change_cell_name(struct hash_cell *hc, char *new_name)
{
char *old_name;
/*
* Rename and move the name cell.
*/
list_del(&hc->name_list);
old_name = hc->name;
mutex_lock(&dm_hash_cells_mutex);
hc->name = new_name;
mutex_unlock(&dm_hash_cells_mutex);
list_add(&hc->name_list, _name_buckets + hash_str(new_name));
return old_name;
}
static struct mapped_device *dm_hash_rename(struct dm_ioctl *param,
const char *new)
{
char *new_data, *old_name = NULL;
struct hash_cell *hc;
struct dm_table *table;
struct mapped_device *md;
unsigned change_uuid = (param->flags & DM_UUID_FLAG) ? 1 : 0;
int srcu_idx;
/*
* duplicate new.
*/
new_data = kstrdup(new, GFP_KERNEL);
if (!new_data)
return ERR_PTR(-ENOMEM);
down_write(&_hash_lock);
/*
* Is new free ?
*/
if (change_uuid)
hc = __get_uuid_cell(new);
else
hc = __get_name_cell(new);
if (hc) {
DMWARN("Unable to change %s on mapped device %s to one that "
"already exists: %s",
change_uuid ? "uuid" : "name",
param->name, new);
dm_put(hc->md);
up_write(&_hash_lock);
kfree(new_data);
return ERR_PTR(-EBUSY);
}
/*
* Is there such a device as 'old' ?
*/
hc = __get_name_cell(param->name);
if (!hc) {
DMWARN("Unable to rename non-existent device, %s to %s%s",
param->name, change_uuid ? "uuid " : "", new);
up_write(&_hash_lock);
kfree(new_data);
return ERR_PTR(-ENXIO);
}
/*
* Does this device already have a uuid?
*/
if (change_uuid && hc->uuid) {
DMWARN("Unable to change uuid of mapped device %s to %s "
"because uuid is already set to %s",
param->name, new, hc->uuid);
dm_put(hc->md);
up_write(&_hash_lock);
kfree(new_data);
return ERR_PTR(-EINVAL);
}
if (change_uuid)
__set_cell_uuid(hc, new_data);
else
old_name = __change_cell_name(hc, new_data);
/*
* Wake up any dm event waiters.
*/
table = dm_get_live_table(hc->md, &srcu_idx);
if (table)
dm_table_event(table);
dm_put_live_table(hc->md, srcu_idx);
if (!dm_kobject_uevent(hc->md, KOBJ_CHANGE, param->event_nr))
param->flags |= DM_UEVENT_GENERATED_FLAG;
md = hc->md;
up_write(&_hash_lock);
kfree(old_name);
return md;
}
void dm_deferred_remove(void)
{
dm_hash_remove_all(true, false, true);
}
/*-----------------------------------------------------------------
* Implementation of the ioctl commands
*---------------------------------------------------------------*/
/*
* All the ioctl commands get dispatched to functions with this
* prototype.
*/
typedef int (*ioctl_fn)(struct file *filp, struct dm_ioctl *param, size_t param_size);
static int remove_all(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
dm_hash_remove_all(true, !!(param->flags & DM_DEFERRED_REMOVE), false);
param->data_size = 0;
return 0;
}
/*
* Round up the ptr to an 8-byte boundary.
*/
#define ALIGN_MASK 7
static inline size_t align_val(size_t val)
{
return (val + ALIGN_MASK) & ~ALIGN_MASK;
}
static inline void *align_ptr(void *ptr)
{
return (void *)align_val((size_t)ptr);
}
/*
* 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 file *filp, 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 *orig_nl, *nl, *old_nl = NULL;
uint32_t *event_nr;
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 += align_val(offsetof(struct dm_name_list, name) + strlen(hc->name) + 1);
needed += align_val(sizeof(uint32_t));
}
}
/*
* Grab our output buffer.
*/
nl = orig_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 = (uint32_t) ((void *) nl -
(void *) old_nl);
disk = dm_disk(hc->md);
nl->dev = huge_encode_dev(disk_devt(disk));
nl->next = 0;
strcpy(nl->name, hc->name);
old_nl = nl;
event_nr = align_ptr(nl->name + strlen(hc->name) + 1);
*event_nr = dm_get_event_nr(hc->md);
nl = align_ptr(event_nr + 1);
}
}
/*
* If mismatch happens, security may be compromised due to buffer
* overflow, so it's better to crash.
*/
BUG_ON((char *)nl - (char *)orig_nl != needed);
out:
up_write(&_hash_lock);
return 0;
}
static void list_version_get_needed(struct target_type *tt, void *needed_param)
{
size_t *needed = needed_param;
*needed += sizeof(struct dm_target_versions);
*needed += strlen(tt->name);
*needed += ALIGN_MASK;
}
static void list_version_get_info(struct target_type *tt, void *param)
{
struct vers_iter *info = param;
/* Check space - it might have changed since the first iteration */
if ((char *)info->vers + sizeof(tt->version) + strlen(tt->name) + 1 >
info->end) {
info->flags = DM_BUFFER_FULL_FLAG;
return;
}
if (info->old_vers)
info->old_vers->next = (uint32_t) ((void *)info->vers -
(void *)info->old_vers);
info->vers->version[0] = tt->version[0];
info->vers->version[1] = tt->version[1];
info->vers->version[2] = tt->version[2];
info->vers->next = 0;
strcpy(info->vers->name, tt->name);
info->old_vers = info->vers;
info->vers = align_ptr(((void *) ++info->vers) + strlen(tt->name) + 1);
}
static int __list_versions(struct dm_ioctl *param, size_t param_size, const char *name)
{
size_t len, needed = 0;
struct dm_target_versions *vers;
struct vers_iter iter_info;
struct target_type *tt = NULL;
if (name) {
tt = dm_get_target_type(name);
if (!tt)
return -EINVAL;
}
/*
* Loop through all the devices working out how much
* space we need.
*/
if (!tt)
dm_target_iterate(list_version_get_needed, &needed);
else
list_version_get_needed(tt, &needed);
/*
* Grab our output buffer.
*/
vers = 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;
iter_info.param_size = param_size;
iter_info.old_vers = NULL;
iter_info.vers = vers;
iter_info.flags = 0;
iter_info.end = (char *)vers+len;
/*
* Now loop through filling out the names & versions.
*/
if (!tt)
dm_target_iterate(list_version_get_info, &iter_info);
else
list_version_get_info(tt, &iter_info);
param->flags |= iter_info.flags;
out:
if (tt)
dm_put_target_type(tt);
return 0;
}
static int list_versions(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
return __list_versions(param, param_size, NULL);
}
static int get_target_version(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
return __list_versions(param, param_size, param->name);
}
static int check_name(const char *name)
{
if (strchr(name, '/')) {
DMWARN("invalid device name");
return -EINVAL;
}
return 0;
}
/*
* On successful return, the caller must not attempt to acquire
* _hash_lock without first calling dm_put_live_table, because dm_table_destroy
* waits for this dm_put_live_table and could be called under this lock.
*/
static struct dm_table *dm_get_inactive_table(struct mapped_device *md, int *srcu_idx)
{
struct hash_cell *hc;
struct dm_table *table = NULL;
/* increment rcu count, we don't care about the table pointer */
dm_get_live_table(md, srcu_idx);
down_read(&_hash_lock);
hc = dm_get_mdptr(md);
if (!hc || hc->md != md) {
DMWARN("device has been removed from the dev hash table.");
goto out;
}
table = hc->new_map;
out:
up_read(&_hash_lock);
return table;
}
static struct dm_table *dm_get_live_or_inactive_table(struct mapped_device *md,
struct dm_ioctl *param,
int *srcu_idx)
{
return (param->flags & DM_QUERY_INACTIVE_TABLE_FLAG) ?
dm_get_inactive_table(md, srcu_idx) : dm_get_live_table(md, srcu_idx);
}
/*
* Fills in a dm_ioctl structure, ready for sending back to
* userland.
*/
static void __dev_status(struct mapped_device *md, struct dm_ioctl *param)
{
struct gendisk *disk = dm_disk(md);
struct dm_table *table;
int srcu_idx;
param->flags &= ~(DM_SUSPEND_FLAG | DM_READONLY_FLAG |
dm: enhance internal suspend and resume interface Rename dm_internal_{suspend,resume} to dm_internal_{suspend,resume}_fast -- dm-stats will continue using these methods to avoid all the extra suspend/resume logic that is not needed in order to quickly flush IO. Introduce dm_internal_suspend_noflush() variant that actually calls the mapped_device's target callbacks -- otherwise target-specific hooks are avoided (e.g. dm-thin's thin_presuspend and thin_postsuspend). Common code between dm_internal_{suspend_noflush,resume} and dm_{suspend,resume} was factored out as __dm_{suspend,resume}. Update dm_internal_{suspend_noflush,resume} to always take and release the mapped_device's suspend_lock. Also update dm_{suspend,resume} to be aware of potential for DM_INTERNAL_SUSPEND_FLAG to be set and respond accordingly by interruptibly waiting for the DM_INTERNAL_SUSPEND_FLAG to be cleared. Add lockdep annotation to dm_suspend() and dm_resume(). The existing DM_SUSPEND_FLAG remains unchanged. DM_INTERNAL_SUSPEND_FLAG is set by dm_internal_suspend_noflush() and cleared by dm_internal_resume(). Both DM_SUSPEND_FLAG and DM_INTERNAL_SUSPEND_FLAG may be set if a device was already suspended when dm_internal_suspend_noflush() was called -- this can be thought of as a "nested suspend". A "nested suspend" can occur with legacy userspace dm-thin code that might suspend all active thin volumes before suspending the pool for resize. But otherwise, in the normal dm-thin-pool suspend case moving forward: the thin-pool will have DM_SUSPEND_FLAG set and all active thins from that thin-pool will have DM_INTERNAL_SUSPEND_FLAG set. Also add DM_INTERNAL_SUSPEND_FLAG to status report. This new DM_INTERNAL_SUSPEND_FLAG state is being reported to assist with debugging (e.g. 'dmsetup info' will report an internally suspended device accordingly). Signed-off-by: Mike Snitzer <snitzer@redhat.com> Acked-by: Joe Thornber <ejt@redhat.com>
2014-10-28 22:34:52 +00:00
DM_ACTIVE_PRESENT_FLAG | DM_INTERNAL_SUSPEND_FLAG);
if (dm_suspended_md(md))
param->flags |= DM_SUSPEND_FLAG;
dm: enhance internal suspend and resume interface Rename dm_internal_{suspend,resume} to dm_internal_{suspend,resume}_fast -- dm-stats will continue using these methods to avoid all the extra suspend/resume logic that is not needed in order to quickly flush IO. Introduce dm_internal_suspend_noflush() variant that actually calls the mapped_device's target callbacks -- otherwise target-specific hooks are avoided (e.g. dm-thin's thin_presuspend and thin_postsuspend). Common code between dm_internal_{suspend_noflush,resume} and dm_{suspend,resume} was factored out as __dm_{suspend,resume}. Update dm_internal_{suspend_noflush,resume} to always take and release the mapped_device's suspend_lock. Also update dm_{suspend,resume} to be aware of potential for DM_INTERNAL_SUSPEND_FLAG to be set and respond accordingly by interruptibly waiting for the DM_INTERNAL_SUSPEND_FLAG to be cleared. Add lockdep annotation to dm_suspend() and dm_resume(). The existing DM_SUSPEND_FLAG remains unchanged. DM_INTERNAL_SUSPEND_FLAG is set by dm_internal_suspend_noflush() and cleared by dm_internal_resume(). Both DM_SUSPEND_FLAG and DM_INTERNAL_SUSPEND_FLAG may be set if a device was already suspended when dm_internal_suspend_noflush() was called -- this can be thought of as a "nested suspend". A "nested suspend" can occur with legacy userspace dm-thin code that might suspend all active thin volumes before suspending the pool for resize. But otherwise, in the normal dm-thin-pool suspend case moving forward: the thin-pool will have DM_SUSPEND_FLAG set and all active thins from that thin-pool will have DM_INTERNAL_SUSPEND_FLAG set. Also add DM_INTERNAL_SUSPEND_FLAG to status report. This new DM_INTERNAL_SUSPEND_FLAG state is being reported to assist with debugging (e.g. 'dmsetup info' will report an internally suspended device accordingly). Signed-off-by: Mike Snitzer <snitzer@redhat.com> Acked-by: Joe Thornber <ejt@redhat.com>
2014-10-28 22:34:52 +00:00
if (dm_suspended_internally_md(md))
param->flags |= DM_INTERNAL_SUSPEND_FLAG;
if (dm_test_deferred_remove_flag(md))
param->flags |= DM_DEFERRED_REMOVE;
param->dev = huge_encode_dev(disk_devt(disk));
/*
* Yes, this will be out of date by the time it gets back
* to userland, but it is still very useful for
* debugging.
*/
param->open_count = dm_open_count(md);
param->event_nr = dm_get_event_nr(md);
param->target_count = 0;
table = dm_get_live_table(md, &srcu_idx);
if (table) {
if (!(param->flags & DM_QUERY_INACTIVE_TABLE_FLAG)) {
if (get_disk_ro(disk))
param->flags |= DM_READONLY_FLAG;
param->target_count = dm_table_get_num_targets(table);
}
param->flags |= DM_ACTIVE_PRESENT_FLAG;
}
dm_put_live_table(md, srcu_idx);
if (param->flags & DM_QUERY_INACTIVE_TABLE_FLAG) {
int srcu_idx;
table = dm_get_inactive_table(md, &srcu_idx);
if (table) {
if (!(dm_table_get_mode(table) & FMODE_WRITE))
param->flags |= DM_READONLY_FLAG;
param->target_count = dm_table_get_num_targets(table);
}
dm_put_live_table(md, srcu_idx);
}
}
static int dev_create(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
int r, m = DM_ANY_MINOR;
struct mapped_device *md;
r = check_name(param->name);
if (r)
return r;
if (param->flags & DM_PERSISTENT_DEV_FLAG)
m = MINOR(huge_decode_dev(param->dev));
r = dm_create(m, &md);
if (r)
return r;
r = dm_hash_insert(param->name, *param->uuid ? param->uuid : NULL, md);
dm: separate device deletion from dm_put This patch separates the device deletion code from dm_put() to make sure the deletion happens in the process context. By this patch, device deletion always occurs in an ioctl (process) context and dm_put() can be called in interrupt context. As a result, the request-based dm's bad dm_put() usage pointed out by Mikulas below disappears. http://marc.info/?l=dm-devel&m=126699981019735&w=2 Without this patch, I confirmed there is a case to crash the system: dm_put() => dm_table_destroy() => vfree() => BUG_ON(in_interrupt()) Some more backgrounds and details: In request-based dm, a device opener can remove a mapped_device while the last request is still completing, because bios in the last request complete first and then the device opener can close and remove the mapped_device before the last request completes: CPU0 CPU1 ================================================================= <<INTERRUPT>> blk_end_request_all(clone_rq) blk_update_request(clone_rq) bio_endio(clone_bio) == end_clone_bio blk_update_request(orig_rq) bio_endio(orig_bio) <<I/O completed>> dm_blk_close() dev_remove() dm_put(md) <<Free md>> blk_finish_request(clone_rq) .... dm_end_request(clone_rq) free_rq_clone(clone_rq) blk_end_request_all(orig_rq) rq_completed(md) So request-based dm used dm_get()/dm_put() to hold md for each I/O until its request completion handling is fully done. However, the final dm_put() can call the device deletion code which must not be run in interrupt context and may cause kernel panic. To solve the problem, this patch moves the device deletion code, dm_destroy(), to predetermined places that is actually deleting the mapped_device in ioctl (process) context, and changes dm_put() just to decrement the reference count of the mapped_device. By this change, dm_put() can be used in any context and the symmetric model below is introduced: dm_create(): create a mapped_device dm_destroy(): destroy a mapped_device dm_get(): increment the reference count of a mapped_device dm_put(): decrement the reference count of a mapped_device dm_destroy() waits for all references of the mapped_device to disappear, then deletes the mapped_device. dm_destroy() uses active waiting with msleep(1), since deleting the mapped_device isn't performance-critical task. And since at this point, nobody opens the mapped_device and no new reference will be taken, the pending counts are just for racing completing activity and will eventually decrease to zero. For the unlikely case of the forced module unload, dm_destroy_immediate(), which doesn't wait and forcibly deletes the mapped_device, is also introduced and used in dm_hash_remove_all(). Otherwise, "rmmod -f" may be stuck and never return. And now, because the mapped_device is deleted at this point, subsequent accesses to the mapped_device may cause NULL pointer references. Cc: stable@kernel.org Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com> Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 03:13:56 +00:00
if (r) {
dm_put(md);
dm_destroy(md);
return r;
}
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
__dev_status(md, param);
dm_put(md);
dm: separate device deletion from dm_put This patch separates the device deletion code from dm_put() to make sure the deletion happens in the process context. By this patch, device deletion always occurs in an ioctl (process) context and dm_put() can be called in interrupt context. As a result, the request-based dm's bad dm_put() usage pointed out by Mikulas below disappears. http://marc.info/?l=dm-devel&m=126699981019735&w=2 Without this patch, I confirmed there is a case to crash the system: dm_put() => dm_table_destroy() => vfree() => BUG_ON(in_interrupt()) Some more backgrounds and details: In request-based dm, a device opener can remove a mapped_device while the last request is still completing, because bios in the last request complete first and then the device opener can close and remove the mapped_device before the last request completes: CPU0 CPU1 ================================================================= <<INTERRUPT>> blk_end_request_all(clone_rq) blk_update_request(clone_rq) bio_endio(clone_bio) == end_clone_bio blk_update_request(orig_rq) bio_endio(orig_bio) <<I/O completed>> dm_blk_close() dev_remove() dm_put(md) <<Free md>> blk_finish_request(clone_rq) .... dm_end_request(clone_rq) free_rq_clone(clone_rq) blk_end_request_all(orig_rq) rq_completed(md) So request-based dm used dm_get()/dm_put() to hold md for each I/O until its request completion handling is fully done. However, the final dm_put() can call the device deletion code which must not be run in interrupt context and may cause kernel panic. To solve the problem, this patch moves the device deletion code, dm_destroy(), to predetermined places that is actually deleting the mapped_device in ioctl (process) context, and changes dm_put() just to decrement the reference count of the mapped_device. By this change, dm_put() can be used in any context and the symmetric model below is introduced: dm_create(): create a mapped_device dm_destroy(): destroy a mapped_device dm_get(): increment the reference count of a mapped_device dm_put(): decrement the reference count of a mapped_device dm_destroy() waits for all references of the mapped_device to disappear, then deletes the mapped_device. dm_destroy() uses active waiting with msleep(1), since deleting the mapped_device isn't performance-critical task. And since at this point, nobody opens the mapped_device and no new reference will be taken, the pending counts are just for racing completing activity and will eventually decrease to zero. For the unlikely case of the forced module unload, dm_destroy_immediate(), which doesn't wait and forcibly deletes the mapped_device, is also introduced and used in dm_hash_remove_all(). Otherwise, "rmmod -f" may be stuck and never return. And now, because the mapped_device is deleted at this point, subsequent accesses to the mapped_device may cause NULL pointer references. Cc: stable@kernel.org Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com> Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 03:13:56 +00:00
return 0;
}
/*
* Always use UUID for lookups if it's present, otherwise use name or dev.
*/
static struct hash_cell *__find_device_hash_cell(struct dm_ioctl *param)
{
struct hash_cell *hc = NULL;
if (*param->uuid) {
if (*param->name || param->dev)
return NULL;
hc = __get_uuid_cell(param->uuid);
if (!hc)
return NULL;
} else if (*param->name) {
if (param->dev)
return NULL;
hc = __get_name_cell(param->name);
if (!hc)
return NULL;
} else if (param->dev) {
hc = __get_dev_cell(param->dev);
if (!hc)
return NULL;
} else
return NULL;
/*
* 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';
if (hc->new_map)
param->flags |= DM_INACTIVE_PRESENT_FLAG;
else
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
return hc;
}
static 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;
up_read(&_hash_lock);
return md;
}
static int dev_remove(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
struct hash_cell *hc;
struct mapped_device *md;
int r;
struct dm_table *t;
down_write(&_hash_lock);
hc = __find_device_hash_cell(param);
if (!hc) {
DMDEBUG_LIMIT("device doesn't appear to be in the dev hash table.");
up_write(&_hash_lock);
return -ENXIO;
}
md = hc->md;
/*
* Ensure the device is not open and nothing further can open it.
*/
r = dm_lock_for_deletion(md, !!(param->flags & DM_DEFERRED_REMOVE), false);
if (r) {
if (r == -EBUSY && param->flags & DM_DEFERRED_REMOVE) {
up_write(&_hash_lock);
dm_put(md);
return 0;
}
DMDEBUG_LIMIT("unable to remove open device %s", hc->name);
up_write(&_hash_lock);
dm_put(md);
return r;
}
t = __hash_remove(hc);
up_write(&_hash_lock);
if (t) {
dm_sync_table(md);
dm_table_destroy(t);
}
param->flags &= ~DM_DEFERRED_REMOVE;
if (!dm_kobject_uevent(md, KOBJ_REMOVE, param->event_nr))
param->flags |= DM_UEVENT_GENERATED_FLAG;
dm_put(md);
dm: separate device deletion from dm_put This patch separates the device deletion code from dm_put() to make sure the deletion happens in the process context. By this patch, device deletion always occurs in an ioctl (process) context and dm_put() can be called in interrupt context. As a result, the request-based dm's bad dm_put() usage pointed out by Mikulas below disappears. http://marc.info/?l=dm-devel&m=126699981019735&w=2 Without this patch, I confirmed there is a case to crash the system: dm_put() => dm_table_destroy() => vfree() => BUG_ON(in_interrupt()) Some more backgrounds and details: In request-based dm, a device opener can remove a mapped_device while the last request is still completing, because bios in the last request complete first and then the device opener can close and remove the mapped_device before the last request completes: CPU0 CPU1 ================================================================= <<INTERRUPT>> blk_end_request_all(clone_rq) blk_update_request(clone_rq) bio_endio(clone_bio) == end_clone_bio blk_update_request(orig_rq) bio_endio(orig_bio) <<I/O completed>> dm_blk_close() dev_remove() dm_put(md) <<Free md>> blk_finish_request(clone_rq) .... dm_end_request(clone_rq) free_rq_clone(clone_rq) blk_end_request_all(orig_rq) rq_completed(md) So request-based dm used dm_get()/dm_put() to hold md for each I/O until its request completion handling is fully done. However, the final dm_put() can call the device deletion code which must not be run in interrupt context and may cause kernel panic. To solve the problem, this patch moves the device deletion code, dm_destroy(), to predetermined places that is actually deleting the mapped_device in ioctl (process) context, and changes dm_put() just to decrement the reference count of the mapped_device. By this change, dm_put() can be used in any context and the symmetric model below is introduced: dm_create(): create a mapped_device dm_destroy(): destroy a mapped_device dm_get(): increment the reference count of a mapped_device dm_put(): decrement the reference count of a mapped_device dm_destroy() waits for all references of the mapped_device to disappear, then deletes the mapped_device. dm_destroy() uses active waiting with msleep(1), since deleting the mapped_device isn't performance-critical task. And since at this point, nobody opens the mapped_device and no new reference will be taken, the pending counts are just for racing completing activity and will eventually decrease to zero. For the unlikely case of the forced module unload, dm_destroy_immediate(), which doesn't wait and forcibly deletes the mapped_device, is also introduced and used in dm_hash_remove_all(). Otherwise, "rmmod -f" may be stuck and never return. And now, because the mapped_device is deleted at this point, subsequent accesses to the mapped_device may cause NULL pointer references. Cc: stable@kernel.org Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com> Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 03:13:56 +00:00
dm_destroy(md);
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 file *filp, struct dm_ioctl *param, size_t param_size)
{
int r;
char *new_data = (char *) param + param->data_start;
struct mapped_device *md;
unsigned change_uuid = (param->flags & DM_UUID_FLAG) ? 1 : 0;
if (new_data < param->data ||
invalid_str(new_data, (void *) param + param_size) || !*new_data ||
strlen(new_data) > (change_uuid ? DM_UUID_LEN - 1 : DM_NAME_LEN - 1)) {
DMWARN("Invalid new mapped device name or uuid string supplied.");
return -EINVAL;
}
if (!change_uuid) {
r = check_name(new_data);
if (r)
return r;
}
md = dm_hash_rename(param, new_data);
if (IS_ERR(md))
return PTR_ERR(md);
__dev_status(md, param);
dm_put(md);
return 0;
}
static int dev_set_geometry(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
int r = -EINVAL, x;
struct mapped_device *md;
struct hd_geometry geometry;
unsigned long indata[4];
char *geostr = (char *) param + param->data_start;
char dummy;
md = find_device(param);
if (!md)
return -ENXIO;
if (geostr < param->data ||
invalid_str(geostr, (void *) param + param_size)) {
DMWARN("Invalid geometry supplied.");
goto out;
}
x = sscanf(geostr, "%lu %lu %lu %lu%c", indata,
indata + 1, indata + 2, indata + 3, &dummy);
if (x != 4) {
DMWARN("Unable to interpret geometry settings.");
goto out;
}
if (indata[0] > 65535 || indata[1] > 255 ||
indata[2] > 255 || indata[3] > ULONG_MAX) {
DMWARN("Geometry exceeds range limits.");
goto out;
}
geometry.cylinders = indata[0];
geometry.heads = indata[1];
geometry.sectors = indata[2];
geometry.start = indata[3];
r = dm_set_geometry(md, &geometry);
param->data_size = 0;
out:
dm_put(md);
return r;
}
static int do_suspend(struct dm_ioctl *param)
{
int r = 0;
unsigned suspend_flags = DM_SUSPEND_LOCKFS_FLAG;
struct mapped_device *md;
md = find_device(param);
if (!md)
return -ENXIO;
if (param->flags & DM_SKIP_LOCKFS_FLAG)
suspend_flags &= ~DM_SUSPEND_LOCKFS_FLAG;
if (param->flags & DM_NOFLUSH_FLAG)
suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
if (!dm_suspended_md(md)) {
r = dm_suspend(md, suspend_flags);
if (r)
goto out;
}
__dev_status(md, param);
out:
dm_put(md);
return r;
}
static int do_resume(struct dm_ioctl *param)
{
int r = 0;
unsigned suspend_flags = DM_SUSPEND_LOCKFS_FLAG;
struct hash_cell *hc;
struct mapped_device *md;
struct dm_table *new_map, *old_map = NULL;
down_write(&_hash_lock);
hc = __find_device_hash_cell(param);
if (!hc) {
DMDEBUG_LIMIT("device doesn't appear to be in the dev hash table.");
up_write(&_hash_lock);
return -ENXIO;
}
md = hc->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 (param->flags & DM_SKIP_LOCKFS_FLAG)
suspend_flags &= ~DM_SUSPEND_LOCKFS_FLAG;
if (param->flags & DM_NOFLUSH_FLAG)
suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
if (!dm_suspended_md(md))
dm_suspend(md, suspend_flags);
old_map = dm_swap_table(md, new_map);
if (IS_ERR(old_map)) {
dm_sync_table(md);
dm table: rework reference counting Rework table reference counting. The existing code uses a reference counter. When the last reference is dropped and the counter reaches zero, the table destructor is called. Table reference counters are acquired/released from upcalls from other kernel code (dm_any_congested, dm_merge_bvec, dm_unplug_all). If the reference counter reaches zero in one of the upcalls, the table destructor is called from almost random kernel code. This leads to various problems: * dm_any_congested being called under a spinlock, which calls the destructor, which calls some sleeping function. * the destructor attempting to take a lock that is already taken by the same process. * stale reference from some other kernel code keeps the table constructed, which keeps some devices open, even after successful return from "dmsetup remove". This can confuse lvm and prevent closing of underlying devices or reusing device minor numbers. The patch changes reference counting so that the table destructor can be called only at predetermined places. The table has always exactly one reference from either mapped_device->map or hash_cell->new_map. After this patch, this reference is not counted in table->holders. A pair of dm_create_table/dm_destroy_table functions is used for table creation/destruction. Temporary references from the other code increase table->holders. A pair of dm_table_get/dm_table_put functions is used to manipulate it. When the table is about to be destroyed, we wait for table->holders to reach 0. Then, we call the table destructor. We use active waiting with msleep(1), because the situation happens rarely (to one user in 5 years) and removing the device isn't performance-critical task: the user doesn't care if it takes one tick more or not. This way, the destructor is called only at specific points (dm_table_destroy function) and the above problems associated with lazy destruction can't happen. Finally remove the temporary protection added to dm_any_congested(). Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-01-06 03:05:10 +00:00
dm_table_destroy(new_map);
dm_put(md);
return PTR_ERR(old_map);
}
if (dm_table_get_mode(new_map) & FMODE_WRITE)
set_disk_ro(dm_disk(md), 0);
else
set_disk_ro(dm_disk(md), 1);
}
if (dm_suspended_md(md)) {
r = dm_resume(md);
if (!r && !dm_kobject_uevent(md, KOBJ_CHANGE, param->event_nr))
param->flags |= DM_UEVENT_GENERATED_FLAG;
}
/*
* Since dm_swap_table synchronizes RCU, nobody should be in
* read-side critical section already.
*/
if (old_map)
dm_table_destroy(old_map);
if (!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 file *filp, struct dm_ioctl *param, size_t param_size)
{
if (param->flags & DM_SUSPEND_FLAG)
return do_suspend(param);
return do_resume(param);
}
/*
* Copies device info back to user space, used by
* the create and info ioctls.
*/
static int dev_status(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
struct mapped_device *md;
md = find_device(param);
if (!md)
return -ENXIO;
__dev_status(md, param);
dm_put(md);
return 0;
}
/*
* 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;
unsigned status_flags = 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);
dm: fix truncated status strings Avoid returning a truncated table or status string instead of setting the DM_BUFFER_FULL_FLAG when the last target of a table fills the buffer. When processing a table or status request, the function retrieve_status calls ti->type->status. If ti->type->status returns non-zero, retrieve_status assumes that the buffer overflowed and sets DM_BUFFER_FULL_FLAG. However, targets don't return non-zero values from their status method on overflow. Most targets returns always zero. If a buffer overflow happens in a target that is not the last in the table, it gets noticed during the next iteration of the loop in retrieve_status; but if a buffer overflow happens in the last target, it goes unnoticed and erroneously truncated data is returned. In the current code, the targets behave in the following way: * dm-crypt returns -ENOMEM if there is not enough space to store the key, but it returns 0 on all other overflows. * dm-thin returns errors from the status method if a disk error happened. This is incorrect because retrieve_status doesn't check the error code, it assumes that all non-zero values mean buffer overflow. * all the other targets always return 0. This patch changes the ti->type->status function to return void (because most targets don't use the return code). Overflow is detected in retrieve_status: if the status method fills up the remaining space completely, it is assumed that buffer overflow happened. Cc: stable@vger.kernel.org Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2013-03-01 22:45:44 +00:00
size_t l;
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) - 1);
outptr += sizeof(struct dm_target_spec);
remaining = len - (outptr - outbuf);
if (remaining <= 0) {
param->flags |= DM_BUFFER_FULL_FLAG;
break;
}
/* Get the status/table string from the target driver */
if (ti->type->status) {
if (param->flags & DM_NOFLUSH_FLAG)
status_flags |= DM_STATUS_NOFLUSH_FLAG;
dm: fix truncated status strings Avoid returning a truncated table or status string instead of setting the DM_BUFFER_FULL_FLAG when the last target of a table fills the buffer. When processing a table or status request, the function retrieve_status calls ti->type->status. If ti->type->status returns non-zero, retrieve_status assumes that the buffer overflowed and sets DM_BUFFER_FULL_FLAG. However, targets don't return non-zero values from their status method on overflow. Most targets returns always zero. If a buffer overflow happens in a target that is not the last in the table, it gets noticed during the next iteration of the loop in retrieve_status; but if a buffer overflow happens in the last target, it goes unnoticed and erroneously truncated data is returned. In the current code, the targets behave in the following way: * dm-crypt returns -ENOMEM if there is not enough space to store the key, but it returns 0 on all other overflows. * dm-thin returns errors from the status method if a disk error happened. This is incorrect because retrieve_status doesn't check the error code, it assumes that all non-zero values mean buffer overflow. * all the other targets always return 0. This patch changes the ti->type->status function to return void (because most targets don't use the return code). Overflow is detected in retrieve_status: if the status method fills up the remaining space completely, it is assumed that buffer overflow happened. Cc: stable@vger.kernel.org Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2013-03-01 22:45:44 +00:00
ti->type->status(ti, type, status_flags, outptr, remaining);
} else
outptr[0] = '\0';
dm: fix truncated status strings Avoid returning a truncated table or status string instead of setting the DM_BUFFER_FULL_FLAG when the last target of a table fills the buffer. When processing a table or status request, the function retrieve_status calls ti->type->status. If ti->type->status returns non-zero, retrieve_status assumes that the buffer overflowed and sets DM_BUFFER_FULL_FLAG. However, targets don't return non-zero values from their status method on overflow. Most targets returns always zero. If a buffer overflow happens in a target that is not the last in the table, it gets noticed during the next iteration of the loop in retrieve_status; but if a buffer overflow happens in the last target, it goes unnoticed and erroneously truncated data is returned. In the current code, the targets behave in the following way: * dm-crypt returns -ENOMEM if there is not enough space to store the key, but it returns 0 on all other overflows. * dm-thin returns errors from the status method if a disk error happened. This is incorrect because retrieve_status doesn't check the error code, it assumes that all non-zero values mean buffer overflow. * all the other targets always return 0. This patch changes the ti->type->status function to return void (because most targets don't use the return code). Overflow is detected in retrieve_status: if the status method fills up the remaining space completely, it is assumed that buffer overflow happened. Cc: stable@vger.kernel.org Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2013-03-01 22:45:44 +00:00
l = strlen(outptr) + 1;
if (l == remaining) {
param->flags |= DM_BUFFER_FULL_FLAG;
break;
}
outptr += l;
used = param->data_start + (outptr - outbuf);
outptr = align_ptr(outptr);
spec->next = outptr - outbuf;
}
if (used)
param->data_size = used;
param->target_count = num_targets;
}
/*
* Wait for a device to report an event
*/
static int dev_wait(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
int r = 0;
struct mapped_device *md;
struct dm_table *table;
int srcu_idx;
md = find_device(param);
if (!md)
return -ENXIO;
/*
* Wait for a notification event
*/
if (dm_wait_event(md, param->event_nr)) {
r = -ERESTARTSYS;
goto out;
}
/*
* 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.
*/
__dev_status(md, param);
table = dm_get_live_or_inactive_table(md, param, &srcu_idx);
if (table)
retrieve_status(table, param, param_size);
dm_put_live_table(md, srcu_idx);
out:
dm_put(md);
return r;
}
/*
* Remember the global event number and make it possible to poll
* for further events.
*/
static int dev_arm_poll(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
struct dm_file *priv = filp->private_data;
priv->global_event_nr = atomic_read(&dm_global_event_nr);
return 0;
}
static inline fmode_t get_mode(struct dm_ioctl *param)
{
fmode_t 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, uint32_t 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;
uint32_t 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 bool is_valid_type(enum dm_queue_mode cur, enum dm_queue_mode new)
{
if (cur == new ||
(cur == DM_TYPE_BIO_BASED && new == DM_TYPE_DAX_BIO_BASED))
return true;
return false;
}
static int table_load(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
int r;
struct hash_cell *hc;
struct dm_table *t, *old_map = NULL;
struct mapped_device *md;
struct target_type *immutable_target_type;
md = find_device(param);
if (!md)
return -ENXIO;
r = dm_table_create(&t, get_mode(param), param->target_count, md);
if (r)
goto err;
/* Protect md->type and md->queue against concurrent table loads. */
dm_lock_md_type(md);
r = populate_table(t, param, param_size);
if (r)
goto err_unlock_md_type;
immutable_target_type = dm_get_immutable_target_type(md);
if (immutable_target_type &&
(immutable_target_type != dm_table_get_immutable_target_type(t)) &&
!dm_table_get_wildcard_target(t)) {
DMWARN("can't replace immutable target type %s",
immutable_target_type->name);
r = -EINVAL;
goto err_unlock_md_type;
}
if (dm_get_md_type(md) == DM_TYPE_NONE) {
/* Initial table load: acquire type of table. */
dm_set_md_type(md, dm_table_get_type(t));
/* setup md->queue to reflect md's type (may block) */
r = dm_setup_md_queue(md, t);
if (r) {
DMWARN("unable to set up device queue for new table.");
goto err_unlock_md_type;
}
} else if (!is_valid_type(dm_get_md_type(md), dm_table_get_type(t))) {
DMWARN("can't change device type (old=%u vs new=%u) after initial table load.",
dm_get_md_type(md), dm_table_get_type(t));
r = -EINVAL;
goto err_unlock_md_type;
}
dm: do not initialise full request queue when bio based Change bio-based mapped devices no longer to have a fully initialized request_queue (request_fn, elevator, etc). This means bio-based DM devices no longer register elevator sysfs attributes ('iosched/' tree or 'scheduler' other than "none"). In contrast, a request-based DM device will continue to have a full request_queue and will register elevator sysfs attributes. Therefore a user can determine a DM device's type by checking if elevator sysfs attributes exist. First allocate a minimalist request_queue structure for a DM device (needed for both bio and request-based DM). Initialization of a full request_queue is deferred until it is known that the DM device is request-based, at the end of the table load sequence. Factor DM device's request_queue initialization: - common to both request-based and bio-based into dm_init_md_queue(). - specific to request-based into dm_init_request_based_queue(). The md->type_lock mutex is used to protect md->queue, in addition to md->type, during table_load(). A DM device's first table_load will establish the immutable md->type. But md->queue initialization, based on md->type, may fail at that time (because blk_init_allocated_queue cannot allocate memory). Therefore any subsequent table_load must (re)try dm_setup_md_queue independently of establishing md->type. Signed-off-by: Mike Snitzer <snitzer@redhat.com> Acked-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 03:14:02 +00:00
dm_unlock_md_type(md);
/* stage inactive table */
down_write(&_hash_lock);
hc = dm_get_mdptr(md);
if (!hc || hc->md != md) {
DMWARN("device has been removed from the dev hash table.");
up_write(&_hash_lock);
r = -ENXIO;
goto err_destroy_table;
}
if (hc->new_map)
old_map = hc->new_map;
hc->new_map = t;
up_write(&_hash_lock);
param->flags |= DM_INACTIVE_PRESENT_FLAG;
__dev_status(md, param);
if (old_map) {
dm_sync_table(md);
dm_table_destroy(old_map);
}
dm_put(md);
return 0;
err_unlock_md_type:
dm_unlock_md_type(md);
err_destroy_table:
dm_table_destroy(t);
err:
dm_put(md);
return r;
}
static int table_clear(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
struct hash_cell *hc;
struct mapped_device *md;
struct dm_table *old_map = NULL;
down_write(&_hash_lock);
hc = __find_device_hash_cell(param);
if (!hc) {
DMDEBUG_LIMIT("device doesn't appear to be in the dev hash table.");
up_write(&_hash_lock);
return -ENXIO;
}
if (hc->new_map) {
old_map = hc->new_map;
hc->new_map = NULL;
}
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
__dev_status(hc->md, param);
md = hc->md;
up_write(&_hash_lock);
if (old_map) {
dm_sync_table(md);
dm_table_destroy(old_map);
}
dm_put(md);
return 0;
}
/*
* 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_dev_internal *dd;
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 = struct_size(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_entry (dd, dm_table_get_devices(table), list)
dm: allow active and inactive tables to share dm_devs Until this change, when loading a new DM table, DM core would re-open all of the devices in the DM table. Now, DM core will avoid redundant device opens (and closes when destroying the old table) if the old table already has a device open using the same mode. This is achieved by managing reference counts on the table_devices that DM core now stores in the mapped_device structure (rather than in the dm_table structure). So a mapped_device's active and inactive dm_tables' dm_dev lists now just point to the dm_devs stored in the mapped_device's table_devices list. This improvement in DM core's device reference counting has the side-effect of fixing a long-standing limitation of the multipath target: a DM multipath table couldn't include any paths that were unusable (failed). For example: if all paths have failed and you add a new, working, path to the table; you can't use it since the table load would fail due to it still containing failed paths. Now a re-load of a multipath table can include failed devices and when those devices become active again they can be used instantly. The device list code in dm.c isn't a straight copy/paste from the code in dm-table.c, but it's very close (aside from some variable renames). One subtle difference is that find_table_device for the tables_devices list will only match devices with the same name and mode. This is because we don't want to upgrade a device's mode in the active table when an inactive table is loaded. Access to the mapped_device structure's tables_devices list requires a mutex (tables_devices_lock), so that tables cannot be created and destroyed concurrently. Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com> Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2014-08-13 18:53:43 +00:00
deps->dev[count++] = huge_encode_dev(dd->dm_dev->bdev->bd_dev);
param->data_size = param->data_start + needed;
}
static int table_deps(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
struct mapped_device *md;
struct dm_table *table;
int srcu_idx;
md = find_device(param);
if (!md)
return -ENXIO;
__dev_status(md, param);
table = dm_get_live_or_inactive_table(md, param, &srcu_idx);
if (table)
retrieve_deps(table, param, param_size);
dm_put_live_table(md, srcu_idx);
dm_put(md);
return 0;
}
/*
* Return the status of a device as a text string for each
* target.
*/
static int table_status(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
struct mapped_device *md;
struct dm_table *table;
int srcu_idx;
md = find_device(param);
if (!md)
return -ENXIO;
__dev_status(md, param);
table = dm_get_live_or_inactive_table(md, param, &srcu_idx);
if (table)
retrieve_status(table, param, param_size);
dm_put_live_table(md, srcu_idx);
dm_put(md);
return 0;
}
/*
* Process device-mapper dependent messages. Messages prefixed with '@'
* are processed by the DM core. All others are delivered to the target.
* Returns a number <= 1 if message was processed by device mapper.
* Returns 2 if message should be delivered to the target.
*/
static int message_for_md(struct mapped_device *md, unsigned argc, char **argv,
char *result, unsigned maxlen)
{
int r;
if (**argv != '@')
return 2; /* no '@' prefix, deliver to target */
if (!strcasecmp(argv[0], "@cancel_deferred_remove")) {
if (argc != 1) {
DMERR("Invalid arguments for @cancel_deferred_remove");
return -EINVAL;
}
return dm_cancel_deferred_remove(md);
}
r = dm_stats_message(md, argc, argv, result, maxlen);
if (r < 2)
return r;
DMERR("Unsupported message sent to DM core: %s", argv[0]);
return -EINVAL;
}
/*
* Pass a message to the target that's at the supplied device offset.
*/
static int target_message(struct file *filp, struct dm_ioctl *param, size_t param_size)
{
int r, argc;
char **argv;
struct mapped_device *md;
struct dm_table *table;
struct dm_target *ti;
struct dm_target_msg *tmsg = (void *) param + param->data_start;
size_t maxlen;
char *result = get_result_buffer(param, param_size, &maxlen);
int srcu_idx;
md = find_device(param);
if (!md)
return -ENXIO;
if (tmsg < (struct dm_target_msg *) param->data ||
invalid_str(tmsg->message, (void *) param + param_size)) {
DMWARN("Invalid target message parameters.");
r = -EINVAL;
goto out;
}
r = dm_split_args(&argc, &argv, tmsg->message);
if (r) {
DMWARN("Failed to split target message parameters");
goto out;
}
if (!argc) {
DMWARN("Empty message received.");
goto out_argv;
}
r = message_for_md(md, argc, argv, result, maxlen);
if (r <= 1)
goto out_argv;
table = dm_get_live_table(md, &srcu_idx);
if (!table)
goto out_table;
if (dm_deleting_md(md)) {
r = -ENXIO;
goto out_table;
}
ti = dm_table_find_target(table, tmsg->sector);
if (!ti) {
DMWARN("Target message sector outside device.");
r = -EINVAL;
} else if (ti->type->message)
r = ti->type->message(ti, argc, argv, result, maxlen);
else {
DMWARN("Target type does not support messages");
r = -EINVAL;
}
out_table:
dm_put_live_table(md, srcu_idx);
out_argv:
kfree(argv);
out:
if (r >= 0)
__dev_status(md, param);
if (r == 1) {
param->flags |= DM_DATA_OUT_FLAG;
if (dm_message_test_buffer_overflow(result, maxlen))
param->flags |= DM_BUFFER_FULL_FLAG;
else
param->data_size = param->data_start + strlen(result) + 1;
r = 0;
}
dm_put(md);
return r;
}
/*
* The ioctl parameter block consists of two parts, a dm_ioctl struct
* followed by a data buffer. This flag is set if the second part,
* which has a variable size, is not used by the function processing
* the ioctl.
*/
#define IOCTL_FLAGS_NO_PARAMS 1
#define IOCTL_FLAGS_ISSUE_GLOBAL_EVENT 2
/*-----------------------------------------------------------------
* Implementation of open/close/ioctl on the special char
* device.
*---------------------------------------------------------------*/
static ioctl_fn lookup_ioctl(unsigned int cmd, int *ioctl_flags)
{
static const struct {
int cmd;
int flags;
ioctl_fn fn;
} _ioctls[] = {
{DM_VERSION_CMD, 0, NULL}, /* version is dealt with elsewhere */
{DM_REMOVE_ALL_CMD, IOCTL_FLAGS_NO_PARAMS | IOCTL_FLAGS_ISSUE_GLOBAL_EVENT, remove_all},
{DM_LIST_DEVICES_CMD, 0, list_devices},
{DM_DEV_CREATE_CMD, IOCTL_FLAGS_NO_PARAMS | IOCTL_FLAGS_ISSUE_GLOBAL_EVENT, dev_create},
{DM_DEV_REMOVE_CMD, IOCTL_FLAGS_NO_PARAMS | IOCTL_FLAGS_ISSUE_GLOBAL_EVENT, dev_remove},
{DM_DEV_RENAME_CMD, IOCTL_FLAGS_ISSUE_GLOBAL_EVENT, dev_rename},
{DM_DEV_SUSPEND_CMD, IOCTL_FLAGS_NO_PARAMS, dev_suspend},
{DM_DEV_STATUS_CMD, IOCTL_FLAGS_NO_PARAMS, dev_status},
{DM_DEV_WAIT_CMD, 0, dev_wait},
{DM_TABLE_LOAD_CMD, 0, table_load},
{DM_TABLE_CLEAR_CMD, IOCTL_FLAGS_NO_PARAMS, table_clear},
{DM_TABLE_DEPS_CMD, 0, table_deps},
{DM_TABLE_STATUS_CMD, 0, table_status},
{DM_LIST_VERSIONS_CMD, 0, list_versions},
{DM_TARGET_MSG_CMD, 0, target_message},
{DM_DEV_SET_GEOMETRY_CMD, 0, dev_set_geometry},
{DM_DEV_ARM_POLL, IOCTL_FLAGS_NO_PARAMS, dev_arm_poll},
{DM_GET_TARGET_VERSION, 0, get_target_version},
};
if (unlikely(cmd >= ARRAY_SIZE(_ioctls)))
return NULL;
*ioctl_flags = _ioctls[cmd].flags;
return _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 *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;
}
#define DM_PARAMS_MALLOC 0x0001 /* Params allocated with kvmalloc() */
#define DM_WIPE_BUFFER 0x0010 /* Wipe input buffer before returning from ioctl */
static void free_params(struct dm_ioctl *param, size_t param_size, int param_flags)
{
if (param_flags & DM_WIPE_BUFFER)
memset(param, 0, param_size);
if (param_flags & DM_PARAMS_MALLOC)
kvfree(param);
}
static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl *param_kernel,
int ioctl_flags, struct dm_ioctl **param, int *param_flags)
{
struct dm_ioctl *dmi;
int secure_data;
const size_t minimum_data_size = offsetof(struct dm_ioctl, data);
unsigned noio_flag;
if (copy_from_user(param_kernel, user, minimum_data_size))
return -EFAULT;
if (param_kernel->data_size < minimum_data_size)
return -EINVAL;
secure_data = param_kernel->flags & DM_SECURE_DATA_FLAG;
*param_flags = secure_data ? DM_WIPE_BUFFER : 0;
if (ioctl_flags & IOCTL_FLAGS_NO_PARAMS) {
dmi = param_kernel;
dmi->data_size = minimum_data_size;
goto data_copied;
}
/*
* Use __GFP_HIGH to avoid low memory issues when a device is
* suspended and the ioctl is needed to resume it.
* Use kmalloc() rather than vmalloc() when we can.
*/
dmi = NULL;
noio_flag = memalloc_noio_save();
dmi = kvmalloc(param_kernel->data_size, GFP_KERNEL | __GFP_HIGH);
memalloc_noio_restore(noio_flag);
if (!dmi) {
if (secure_data && clear_user(user, param_kernel->data_size))
return -EFAULT;
return -ENOMEM;
}
*param_flags |= DM_PARAMS_MALLOC;
/* Copy from param_kernel (which was already copied from user) */
memcpy(dmi, param_kernel, minimum_data_size);
if (copy_from_user(&dmi->data, (char __user *)user + minimum_data_size,
param_kernel->data_size - minimum_data_size))
2012-12-21 20:23:30 +00:00
goto bad;
data_copied:
/* Wipe the user buffer so we do not return it to userspace */
if (secure_data && clear_user(user, param_kernel->data_size))
goto bad;
*param = dmi;
return 0;
bad:
free_params(dmi, param_kernel->data_size, *param_flags);
return -EFAULT;
}
static int validate_params(uint cmd, struct dm_ioctl *param)
{
/* Always clear this flag */
param->flags &= ~DM_BUFFER_FULL_FLAG;
param->flags &= ~DM_UEVENT_GENERATED_FLAG;
param->flags &= ~DM_SECURE_DATA_FLAG;
param->flags &= ~DM_DATA_OUT_FLAG;
/* Ignores parameters */
if (cmd == DM_REMOVE_ALL_CMD ||
cmd == DM_LIST_DEVICES_CMD ||
cmd == DM_LIST_VERSIONS_CMD)
return 0;
if (cmd == DM_DEV_CREATE_CMD) {
if (!*param->name) {
DMWARN("name not supplied when creating device");
return -EINVAL;
}
} else if (*param->uuid && *param->name) {
DMWARN("only supply one of name or uuid, 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 file *file, uint command, struct dm_ioctl __user *user)
{
int r = 0;
int ioctl_flags;
int param_flags;
unsigned int cmd;
treewide: Remove uninitialized_var() usage Using uninitialized_var() is dangerous as it papers over real bugs[1] (or can in the future), and suppresses unrelated compiler warnings (e.g. "unused variable"). If the compiler thinks it is uninitialized, either simply initialize the variable or make compiler changes. In preparation for removing[2] the[3] macro[4], remove all remaining needless uses with the following script: git grep '\buninitialized_var\b' | cut -d: -f1 | sort -u | \ xargs perl -pi -e \ 's/\buninitialized_var\(([^\)]+)\)/\1/g; s:\s*/\* (GCC be quiet|to make compiler happy) \*/$::g;' drivers/video/fbdev/riva/riva_hw.c was manually tweaked to avoid pathological white-space. No outstanding warnings were found building allmodconfig with GCC 9.3.0 for x86_64, i386, arm64, arm, powerpc, powerpc64le, s390x, mips, sparc64, alpha, and m68k. [1] https://lore.kernel.org/lkml/20200603174714.192027-1-glider@google.com/ [2] https://lore.kernel.org/lkml/CA+55aFw+Vbj0i=1TGqCR5vQkCzWJ0QxK6CernOU6eedsudAixw@mail.gmail.com/ [3] https://lore.kernel.org/lkml/CA+55aFwgbgqhbp1fkxvRKEpzyR5J8n1vKT1VZdz9knmPuXhOeg@mail.gmail.com/ [4] https://lore.kernel.org/lkml/CA+55aFz2500WfbKXAx8s67wrm9=yVJu65TpLgN_ybYNv0VEOKA@mail.gmail.com/ Reviewed-by: Leon Romanovsky <leonro@mellanox.com> # drivers/infiniband and mlx4/mlx5 Acked-by: Jason Gunthorpe <jgg@mellanox.com> # IB Acked-by: Kalle Valo <kvalo@codeaurora.org> # wireless drivers Reviewed-by: Chao Yu <yuchao0@huawei.com> # erofs Signed-off-by: Kees Cook <keescook@chromium.org>
2020-06-03 20:09:38 +00:00
struct dm_ioctl *param;
ioctl_fn fn = NULL;
size_t input_param_size;
struct dm_ioctl param_kernel;
/* 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, &ioctl_flags);
if (!fn) {
DMWARN("dm_ctl_ioctl: unknown command 0x%x", command);
return -ENOTTY;
}
/*
* Copy the parameters into kernel space.
*/
r = copy_params(user, &param_kernel, ioctl_flags, &param, &param_flags);
if (r)
return r;
input_param_size = param->data_size;
r = validate_params(cmd, param);
if (r)
goto out;
param->data_size = offsetof(struct dm_ioctl, data);
r = fn(file, param, input_param_size);
if (unlikely(param->flags & DM_BUFFER_FULL_FLAG) &&
unlikely(ioctl_flags & IOCTL_FLAGS_NO_PARAMS))
DMERR("ioctl %d tried to output some data but has IOCTL_FLAGS_NO_PARAMS set", cmd);
if (!r && ioctl_flags & IOCTL_FLAGS_ISSUE_GLOBAL_EVENT)
dm_issue_global_event();
/*
* Copy the results back to userland.
*/
if (!r && copy_to_user(user, param, param->data_size))
r = -EFAULT;
out:
free_params(param, input_param_size, param_flags);
return r;
}
static long dm_ctl_ioctl(struct file *file, uint command, ulong u)
{
return (long)ctl_ioctl(file, command, (struct dm_ioctl __user *)u);
}
#ifdef CONFIG_COMPAT
static long dm_compat_ctl_ioctl(struct file *file, uint command, ulong u)
{
return (long)dm_ctl_ioctl(file, command, (ulong) compat_ptr(u));
}
#else
#define dm_compat_ctl_ioctl NULL
#endif
static int dm_open(struct inode *inode, struct file *filp)
{
int r;
struct dm_file *priv;
r = nonseekable_open(inode, filp);
if (unlikely(r))
return r;
priv = filp->private_data = kmalloc(sizeof(struct dm_file), GFP_KERNEL);
if (!priv)
return -ENOMEM;
priv->global_event_nr = atomic_read(&dm_global_event_nr);
return 0;
}
static int dm_release(struct inode *inode, struct file *filp)
{
kfree(filp->private_data);
return 0;
}
static __poll_t dm_poll(struct file *filp, poll_table *wait)
{
struct dm_file *priv = filp->private_data;
__poll_t mask = 0;
poll_wait(filp, &dm_global_eventq, wait);
if ((int)(atomic_read(&dm_global_event_nr) - priv->global_event_nr) > 0)
mask |= EPOLLIN;
return mask;
}
static const struct file_operations _ctl_fops = {
.open = dm_open,
.release = dm_release,
.poll = dm_poll,
.unlocked_ioctl = dm_ctl_ioctl,
.compat_ioctl = dm_compat_ctl_ioctl,
.owner = THIS_MODULE,
llseek: automatically add .llseek fop All file_operations should get a .llseek operation so we can make nonseekable_open the default for future file operations without a .llseek pointer. The three cases that we can automatically detect are no_llseek, seq_lseek and default_llseek. For cases where we can we can automatically prove that the file offset is always ignored, we use noop_llseek, which maintains the current behavior of not returning an error from a seek. New drivers should normally not use noop_llseek but instead use no_llseek and call nonseekable_open at open time. Existing drivers can be converted to do the same when the maintainer knows for certain that no user code relies on calling seek on the device file. The generated code is often incorrectly indented and right now contains comments that clarify for each added line why a specific variant was chosen. In the version that gets submitted upstream, the comments will be gone and I will manually fix the indentation, because there does not seem to be a way to do that using coccinelle. Some amount of new code is currently sitting in linux-next that should get the same modifications, which I will do at the end of the merge window. Many thanks to Julia Lawall for helping me learn to write a semantic patch that does all this. ===== begin semantic patch ===== // This adds an llseek= method to all file operations, // as a preparation for making no_llseek the default. // // The rules are // - use no_llseek explicitly if we do nonseekable_open // - use seq_lseek for sequential files // - use default_llseek if we know we access f_pos // - use noop_llseek if we know we don't access f_pos, // but we still want to allow users to call lseek // @ open1 exists @ identifier nested_open; @@ nested_open(...) { <+... nonseekable_open(...) ...+> } @ open exists@ identifier open_f; identifier i, f; identifier open1.nested_open; @@ int open_f(struct inode *i, struct file *f) { <+... ( nonseekable_open(...) | nested_open(...) ) ...+> } @ read disable optional_qualifier exists @ identifier read_f; identifier f, p, s, off; type ssize_t, size_t, loff_t; expression E; identifier func; @@ ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off) { <+... ( *off = E | *off += E | func(..., off, ...) | E = *off ) ...+> } @ read_no_fpos disable optional_qualifier exists @ identifier read_f; identifier f, p, s, off; type ssize_t, size_t, loff_t; @@ ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off) { ... when != off } @ write @ identifier write_f; identifier f, p, s, off; type ssize_t, size_t, loff_t; expression E; identifier func; @@ ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off) { <+... ( *off = E | *off += E | func(..., off, ...) | E = *off ) ...+> } @ write_no_fpos @ identifier write_f; identifier f, p, s, off; type ssize_t, size_t, loff_t; @@ ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off) { ... when != off } @ fops0 @ identifier fops; @@ struct file_operations fops = { ... }; @ has_llseek depends on fops0 @ identifier fops0.fops; identifier llseek_f; @@ struct file_operations fops = { ... .llseek = llseek_f, ... }; @ has_read depends on fops0 @ identifier fops0.fops; identifier read_f; @@ struct file_operations fops = { ... .read = read_f, ... }; @ has_write depends on fops0 @ identifier fops0.fops; identifier write_f; @@ struct file_operations fops = { ... .write = write_f, ... }; @ has_open depends on fops0 @ identifier fops0.fops; identifier open_f; @@ struct file_operations fops = { ... .open = open_f, ... }; // use no_llseek if we call nonseekable_open //////////////////////////////////////////// @ nonseekable1 depends on !has_llseek && has_open @ identifier fops0.fops; identifier nso ~= "nonseekable_open"; @@ struct file_operations fops = { ... .open = nso, ... +.llseek = no_llseek, /* nonseekable */ }; @ nonseekable2 depends on !has_llseek @ identifier fops0.fops; identifier open.open_f; @@ struct file_operations fops = { ... .open = open_f, ... +.llseek = no_llseek, /* open uses nonseekable */ }; // use seq_lseek for sequential files ///////////////////////////////////// @ seq depends on !has_llseek @ identifier fops0.fops; identifier sr ~= "seq_read"; @@ struct file_operations fops = { ... .read = sr, ... +.llseek = seq_lseek, /* we have seq_read */ }; // use default_llseek if there is a readdir /////////////////////////////////////////// @ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier readdir_e; @@ // any other fop is used that changes pos struct file_operations fops = { ... .readdir = readdir_e, ... +.llseek = default_llseek, /* readdir is present */ }; // use default_llseek if at least one of read/write touches f_pos ///////////////////////////////////////////////////////////////// @ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier read.read_f; @@ // read fops use offset struct file_operations fops = { ... .read = read_f, ... +.llseek = default_llseek, /* read accesses f_pos */ }; @ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier write.write_f; @@ // write fops use offset struct file_operations fops = { ... .write = write_f, ... + .llseek = default_llseek, /* write accesses f_pos */ }; // Use noop_llseek if neither read nor write accesses f_pos /////////////////////////////////////////////////////////// @ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier read_no_fpos.read_f; identifier write_no_fpos.write_f; @@ // write fops use offset struct file_operations fops = { ... .write = write_f, .read = read_f, ... +.llseek = noop_llseek, /* read and write both use no f_pos */ }; @ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier write_no_fpos.write_f; @@ struct file_operations fops = { ... .write = write_f, ... +.llseek = noop_llseek, /* write uses no f_pos */ }; @ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier read_no_fpos.read_f; @@ struct file_operations fops = { ... .read = read_f, ... +.llseek = noop_llseek, /* read uses no f_pos */ }; @ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; @@ struct file_operations fops = { ... +.llseek = noop_llseek, /* no read or write fn */ }; ===== End semantic patch ===== Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Julia Lawall <julia@diku.dk> Cc: Christoph Hellwig <hch@infradead.org>
2010-08-15 16:52:59 +00:00
.llseek = noop_llseek,
};
static struct miscdevice _dm_misc = {
.minor = MAPPER_CTRL_MINOR,
.name = DM_NAME,
.nodename = DM_DIR "/" DM_CONTROL_NODE,
.fops = &_ctl_fops
};
MODULE_ALIAS_MISCDEV(MAPPER_CTRL_MINOR);
MODULE_ALIAS("devname:" DM_DIR "/" DM_CONTROL_NODE);
/*
* 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)
{
misc_deregister(&_dm_misc);
dm_hash_exit();
}
/**
* dm_copy_name_and_uuid - Copy mapped device name & uuid into supplied buffers
* @md: Pointer to mapped_device
* @name: Buffer (size DM_NAME_LEN) for name
* @uuid: Buffer (size DM_UUID_LEN) for uuid or empty string if uuid not defined
*/
int dm_copy_name_and_uuid(struct mapped_device *md, char *name, char *uuid)
{
int r = 0;
struct hash_cell *hc;
if (!md)
return -ENXIO;
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-10 23:51:52 +00:00
mutex_lock(&dm_hash_cells_mutex);
hc = dm_get_mdptr(md);
if (!hc || hc->md != md) {
r = -ENXIO;
goto out;
}
if (name)
strcpy(name, hc->name);
if (uuid)
strcpy(uuid, hc->uuid ? : "");
out:
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-10 23:51:52 +00:00
mutex_unlock(&dm_hash_cells_mutex);
return r;
}
EXPORT_SYMBOL_GPL(dm_copy_name_and_uuid);
dm: add support to directly boot to a mapped device Add a "create" module parameter, which allows device-mapper targets to be configured at boot time. This enables early use of DM targets in the boot process (as the root device or otherwise) without the need of an initramfs. The syntax used in the boot param is based on the concise format from the dmsetup tool to follow the rule of least surprise: dmsetup table --concise /dev/mapper/lroot Which is: dm-mod.create=<name>,<uuid>,<minor>,<flags>,<table>[,<table>+][;<name>,<uuid>,<minor>,<flags>,<table>[,<table>+]+] Where, <name> ::= The device name. <uuid> ::= xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | "" <minor> ::= The device minor number | "" <flags> ::= "ro" | "rw" <table> ::= <start_sector> <num_sectors> <target_type> <target_args> <target_type> ::= "verity" | "linear" | ... For example, the following could be added in the boot parameters: dm-mod.create="lroot,,,rw, 0 4096 linear 98:16 0, 4096 4096 linear 98:32 0" root=/dev/dm-0 Only the targets that were tested are allowed and the ones that don't change any block device when the device is create as read-only. For example, mirror and cache targets are not allowed. The rationale behind this is that if the user makes a mistake, choosing the wrong device to be the mirror or the cache can corrupt data. The only targets initially allowed are: * crypt * delay * linear * snapshot-origin * striped * verity Co-developed-by: Will Drewry <wad@chromium.org> Co-developed-by: Kees Cook <keescook@chromium.org> Co-developed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com> Signed-off-by: Helen Koike <helen.koike@collabora.com> Reviewed-by: Kees Cook <keescook@chromium.org> Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-02-21 20:33:34 +00:00
/**
* dm_early_create - create a mapped device in early boot.
*
* @dmi: Contains main information of the device mapping to be created.
* @spec_array: array of pointers to struct dm_target_spec. Describes the
* mapping table of the device.
* @target_params_array: array of strings with the parameters to a specific
* target.
*
* Instead of having the struct dm_target_spec and the parameters for every
* target embedded at the end of struct dm_ioctl (as performed in a normal
* ioctl), pass them as arguments, so the caller doesn't need to serialize them.
* The size of the spec_array and target_params_array is given by
* @dmi->target_count.
* This function is supposed to be called in early boot, so locking mechanisms
* to protect against concurrent loads are not required.
*/
int __init dm_early_create(struct dm_ioctl *dmi,
struct dm_target_spec **spec_array,
char **target_params_array)
{
int r, m = DM_ANY_MINOR;
struct dm_table *t, *old_map;
struct mapped_device *md;
unsigned int i;
if (!dmi->target_count)
return -EINVAL;
r = check_name(dmi->name);
if (r)
return r;
if (dmi->flags & DM_PERSISTENT_DEV_FLAG)
m = MINOR(huge_decode_dev(dmi->dev));
/* alloc dm device */
r = dm_create(m, &md);
if (r)
return r;
/* hash insert */
r = dm_hash_insert(dmi->name, *dmi->uuid ? dmi->uuid : NULL, md);
if (r)
goto err_destroy_dm;
/* alloc table */
r = dm_table_create(&t, get_mode(dmi), dmi->target_count, md);
if (r)
goto err_hash_remove;
dm: add support to directly boot to a mapped device Add a "create" module parameter, which allows device-mapper targets to be configured at boot time. This enables early use of DM targets in the boot process (as the root device or otherwise) without the need of an initramfs. The syntax used in the boot param is based on the concise format from the dmsetup tool to follow the rule of least surprise: dmsetup table --concise /dev/mapper/lroot Which is: dm-mod.create=<name>,<uuid>,<minor>,<flags>,<table>[,<table>+][;<name>,<uuid>,<minor>,<flags>,<table>[,<table>+]+] Where, <name> ::= The device name. <uuid> ::= xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | "" <minor> ::= The device minor number | "" <flags> ::= "ro" | "rw" <table> ::= <start_sector> <num_sectors> <target_type> <target_args> <target_type> ::= "verity" | "linear" | ... For example, the following could be added in the boot parameters: dm-mod.create="lroot,,,rw, 0 4096 linear 98:16 0, 4096 4096 linear 98:32 0" root=/dev/dm-0 Only the targets that were tested are allowed and the ones that don't change any block device when the device is create as read-only. For example, mirror and cache targets are not allowed. The rationale behind this is that if the user makes a mistake, choosing the wrong device to be the mirror or the cache can corrupt data. The only targets initially allowed are: * crypt * delay * linear * snapshot-origin * striped * verity Co-developed-by: Will Drewry <wad@chromium.org> Co-developed-by: Kees Cook <keescook@chromium.org> Co-developed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com> Signed-off-by: Helen Koike <helen.koike@collabora.com> Reviewed-by: Kees Cook <keescook@chromium.org> Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-02-21 20:33:34 +00:00
/* add targets */
for (i = 0; i < dmi->target_count; i++) {
r = dm_table_add_target(t, spec_array[i]->target_type,
(sector_t) spec_array[i]->sector_start,
(sector_t) spec_array[i]->length,
target_params_array[i]);
if (r) {
DMWARN("error adding target to table");
goto err_destroy_table;
}
}
/* finish table */
r = dm_table_complete(t);
if (r)
goto err_destroy_table;
md->type = dm_table_get_type(t);
/* setup md->queue to reflect md's type (may block) */
r = dm_setup_md_queue(md, t);
if (r) {
DMWARN("unable to set up device queue for new table.");
goto err_destroy_table;
}
/* Set new map */
dm_suspend(md, 0);
old_map = dm_swap_table(md, t);
if (IS_ERR(old_map)) {
r = PTR_ERR(old_map);
goto err_destroy_table;
}
set_disk_ro(dm_disk(md), !!(dmi->flags & DM_READONLY_FLAG));
/* resume device */
r = dm_resume(md);
if (r)
goto err_destroy_table;
DMINFO("%s (%s) is ready", md->disk->disk_name, dmi->name);
dm_put(md);
return 0;
err_destroy_table:
dm_table_destroy(t);
err_hash_remove:
(void) __hash_remove(__get_name_cell(dmi->name));
/* release reference from __get_name_cell */
dm_put(md);
dm: add support to directly boot to a mapped device Add a "create" module parameter, which allows device-mapper targets to be configured at boot time. This enables early use of DM targets in the boot process (as the root device or otherwise) without the need of an initramfs. The syntax used in the boot param is based on the concise format from the dmsetup tool to follow the rule of least surprise: dmsetup table --concise /dev/mapper/lroot Which is: dm-mod.create=<name>,<uuid>,<minor>,<flags>,<table>[,<table>+][;<name>,<uuid>,<minor>,<flags>,<table>[,<table>+]+] Where, <name> ::= The device name. <uuid> ::= xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | "" <minor> ::= The device minor number | "" <flags> ::= "ro" | "rw" <table> ::= <start_sector> <num_sectors> <target_type> <target_args> <target_type> ::= "verity" | "linear" | ... For example, the following could be added in the boot parameters: dm-mod.create="lroot,,,rw, 0 4096 linear 98:16 0, 4096 4096 linear 98:32 0" root=/dev/dm-0 Only the targets that were tested are allowed and the ones that don't change any block device when the device is create as read-only. For example, mirror and cache targets are not allowed. The rationale behind this is that if the user makes a mistake, choosing the wrong device to be the mirror or the cache can corrupt data. The only targets initially allowed are: * crypt * delay * linear * snapshot-origin * striped * verity Co-developed-by: Will Drewry <wad@chromium.org> Co-developed-by: Kees Cook <keescook@chromium.org> Co-developed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com> Signed-off-by: Helen Koike <helen.koike@collabora.com> Reviewed-by: Kees Cook <keescook@chromium.org> Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-02-21 20:33:34 +00:00
err_destroy_dm:
dm_put(md);
dm_destroy(md);
return r;
}