2015-03-20 14:50:37 +00:00
|
|
|
/*
|
|
|
|
* Copyright (C) 2014 Facebook. All rights reserved.
|
|
|
|
*
|
|
|
|
* This file is released under the GPL.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/device-mapper.h>
|
|
|
|
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/blkdev.h>
|
|
|
|
#include <linux/bio.h>
|
2017-10-20 05:24:04 +00:00
|
|
|
#include <linux/dax.h>
|
2015-03-20 14:50:37 +00:00
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/kthread.h>
|
|
|
|
#include <linux/freezer.h>
|
2017-10-20 05:24:04 +00:00
|
|
|
#include <linux/uio.h>
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
#define DM_MSG_PREFIX "log-writes"
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This target will sequentially log all writes to the target device onto the
|
|
|
|
* log device. This is helpful for replaying writes to check for fs consistency
|
|
|
|
* at all times. This target provides a mechanism to mark specific events to
|
|
|
|
* check data at a later time. So for example you would:
|
|
|
|
*
|
|
|
|
* write data
|
|
|
|
* fsync
|
|
|
|
* dmsetup message /dev/whatever mark mymark
|
|
|
|
* unmount /mnt/test
|
|
|
|
*
|
|
|
|
* Then replay the log up to mymark and check the contents of the replay to
|
|
|
|
* verify it matches what was written.
|
|
|
|
*
|
|
|
|
* We log writes only after they have been flushed, this makes the log describe
|
|
|
|
* close to the order in which the data hits the actual disk, not its cache. So
|
|
|
|
* for example the following sequence (W means write, C means complete)
|
|
|
|
*
|
|
|
|
* Wa,Wb,Wc,Cc,Ca,FLUSH,FUAd,Cb,CFLUSH,CFUAd
|
|
|
|
*
|
|
|
|
* Would result in the log looking like this:
|
|
|
|
*
|
dm log writes: fix incorrect comment about the logged sequence example
dm-log-writes records the sequence of completion, not submission, thus
for the following sequence (W=write, C=complete):
Wa,Wb,Wc,Cc,Ca,FLUSH,FUAd,Cb,CFLUSH,CFUAd
The logged results in log device should be:
c,a,b,flush,fua
Fix the comment to give a better example.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-06-18 05:39:38 +00:00
|
|
|
* c,a,b,flush,fuad,<other writes>,<next flush>
|
2015-03-20 14:50:37 +00:00
|
|
|
*
|
|
|
|
* This is meant to help expose problems where file systems do not properly wait
|
|
|
|
* on data being written before invoking a FLUSH. FUA bypasses cache so once it
|
|
|
|
* completes it is added to the log as it should be on disk.
|
|
|
|
*
|
|
|
|
* We treat DISCARDs as if they don't bypass cache so that they are logged in
|
|
|
|
* order of completion along with the normal writes. If we didn't do it this
|
|
|
|
* way we would process all the discards first and then write all the data, when
|
|
|
|
* in fact we want to do the data and the discard in the order that they
|
|
|
|
* completed.
|
|
|
|
*/
|
2018-02-28 07:32:47 +00:00
|
|
|
#define LOG_FLUSH_FLAG (1 << 0)
|
|
|
|
#define LOG_FUA_FLAG (1 << 1)
|
|
|
|
#define LOG_DISCARD_FLAG (1 << 2)
|
|
|
|
#define LOG_MARK_FLAG (1 << 3)
|
|
|
|
#define LOG_METADATA_FLAG (1 << 4)
|
2015-03-20 14:50:37 +00:00
|
|
|
|
2015-04-18 22:07:30 +00:00
|
|
|
#define WRITE_LOG_VERSION 1ULL
|
|
|
|
#define WRITE_LOG_MAGIC 0x6a736677736872ULL
|
2019-06-05 13:27:08 +00:00
|
|
|
#define WRITE_LOG_SUPER_SECTOR 0
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* The disk format for this is braindead simple.
|
|
|
|
*
|
|
|
|
* At byte 0 we have our super, followed by the following sequence for
|
|
|
|
* nr_entries:
|
|
|
|
*
|
|
|
|
* [ 1 sector ][ entry->nr_sectors ]
|
|
|
|
* [log_write_entry][ data written ]
|
|
|
|
*
|
|
|
|
* The log_write_entry takes up a full sector so we can have arbitrary length
|
|
|
|
* marks and it leaves us room for extra content in the future.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Basic info about the log for userspace.
|
|
|
|
*/
|
|
|
|
struct log_write_super {
|
|
|
|
__le64 magic;
|
|
|
|
__le64 version;
|
|
|
|
__le64 nr_entries;
|
|
|
|
__le32 sectorsize;
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* sector - the sector we wrote.
|
|
|
|
* nr_sectors - the number of sectors we wrote.
|
|
|
|
* flags - flags for this log entry.
|
|
|
|
* data_len - the size of the data in this log entry, this is for private log
|
|
|
|
* entry stuff, the MARK data provided by userspace for example.
|
|
|
|
*/
|
|
|
|
struct log_write_entry {
|
|
|
|
__le64 sector;
|
|
|
|
__le64 nr_sectors;
|
|
|
|
__le64 flags;
|
|
|
|
__le64 data_len;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct log_writes_c {
|
|
|
|
struct dm_dev *dev;
|
|
|
|
struct dm_dev *logdev;
|
|
|
|
u64 logged_entries;
|
|
|
|
u32 sectorsize;
|
2017-07-28 14:42:25 +00:00
|
|
|
u32 sectorshift;
|
2015-03-20 14:50:37 +00:00
|
|
|
atomic_t io_blocks;
|
|
|
|
atomic_t pending_blocks;
|
|
|
|
sector_t next_sector;
|
|
|
|
sector_t end_sector;
|
|
|
|
bool logging_enabled;
|
|
|
|
bool device_supports_discard;
|
|
|
|
spinlock_t blocks_lock;
|
|
|
|
struct list_head unflushed_blocks;
|
|
|
|
struct list_head logging_blocks;
|
|
|
|
wait_queue_head_t wait;
|
|
|
|
struct task_struct *log_kthread;
|
2019-06-05 13:27:08 +00:00
|
|
|
struct completion super_done;
|
2015-03-20 14:50:37 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
struct pending_block {
|
|
|
|
int vec_cnt;
|
|
|
|
u64 flags;
|
|
|
|
sector_t sector;
|
|
|
|
sector_t nr_sectors;
|
|
|
|
char *data;
|
|
|
|
u32 datalen;
|
|
|
|
struct list_head list;
|
2020-05-07 18:51:58 +00:00
|
|
|
struct bio_vec vecs[];
|
2015-03-20 14:50:37 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
struct per_bio_data {
|
|
|
|
struct pending_block *block;
|
|
|
|
};
|
|
|
|
|
2017-07-28 14:42:25 +00:00
|
|
|
static inline sector_t bio_to_dev_sectors(struct log_writes_c *lc,
|
|
|
|
sector_t sectors)
|
|
|
|
{
|
|
|
|
return sectors >> (lc->sectorshift - SECTOR_SHIFT);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline sector_t dev_to_bio_sectors(struct log_writes_c *lc,
|
|
|
|
sector_t sectors)
|
|
|
|
{
|
|
|
|
return sectors << (lc->sectorshift - SECTOR_SHIFT);
|
|
|
|
}
|
|
|
|
|
2015-03-20 14:50:37 +00:00
|
|
|
static void put_pending_block(struct log_writes_c *lc)
|
|
|
|
{
|
|
|
|
if (atomic_dec_and_test(&lc->pending_blocks)) {
|
|
|
|
smp_mb__after_atomic();
|
|
|
|
if (waitqueue_active(&lc->wait))
|
|
|
|
wake_up(&lc->wait);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void put_io_block(struct log_writes_c *lc)
|
|
|
|
{
|
|
|
|
if (atomic_dec_and_test(&lc->io_blocks)) {
|
|
|
|
smp_mb__after_atomic();
|
|
|
|
if (waitqueue_active(&lc->wait))
|
|
|
|
wake_up(&lc->wait);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-07-20 13:29:37 +00:00
|
|
|
static void log_end_io(struct bio *bio)
|
2015-03-20 14:50:37 +00:00
|
|
|
{
|
|
|
|
struct log_writes_c *lc = bio->bi_private;
|
|
|
|
|
2017-06-03 07:38:06 +00:00
|
|
|
if (bio->bi_status) {
|
2015-03-20 14:50:37 +00:00
|
|
|
unsigned long flags;
|
|
|
|
|
2017-06-03 07:38:06 +00:00
|
|
|
DMERR("Error writing log block, error=%d", bio->bi_status);
|
2015-03-20 14:50:37 +00:00
|
|
|
spin_lock_irqsave(&lc->blocks_lock, flags);
|
|
|
|
lc->logging_enabled = false;
|
|
|
|
spin_unlock_irqrestore(&lc->blocks_lock, flags);
|
|
|
|
}
|
|
|
|
|
2016-09-22 07:10:01 +00:00
|
|
|
bio_free_pages(bio);
|
2015-03-20 14:50:37 +00:00
|
|
|
put_io_block(lc);
|
|
|
|
bio_put(bio);
|
|
|
|
}
|
|
|
|
|
2019-06-05 13:27:08 +00:00
|
|
|
static void log_end_super(struct bio *bio)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = bio->bi_private;
|
|
|
|
|
|
|
|
complete(&lc->super_done);
|
|
|
|
log_end_io(bio);
|
|
|
|
}
|
|
|
|
|
2015-03-20 14:50:37 +00:00
|
|
|
/*
|
|
|
|
* Meant to be called if there is an error, it will free all the pages
|
|
|
|
* associated with the block.
|
|
|
|
*/
|
|
|
|
static void free_pending_block(struct log_writes_c *lc,
|
|
|
|
struct pending_block *block)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < block->vec_cnt; i++) {
|
|
|
|
if (block->vecs[i].bv_page)
|
|
|
|
__free_page(block->vecs[i].bv_page);
|
|
|
|
}
|
|
|
|
kfree(block->data);
|
|
|
|
kfree(block);
|
|
|
|
put_pending_block(lc);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_metadata(struct log_writes_c *lc, void *entry,
|
|
|
|
size_t entrylen, void *data, size_t datalen,
|
|
|
|
sector_t sector)
|
|
|
|
{
|
|
|
|
struct bio *bio;
|
|
|
|
struct page *page;
|
|
|
|
void *ptr;
|
|
|
|
size_t ret;
|
|
|
|
|
|
|
|
bio = bio_alloc(GFP_KERNEL, 1);
|
|
|
|
if (!bio) {
|
|
|
|
DMERR("Couldn't alloc log bio");
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
bio->bi_iter.bi_size = 0;
|
|
|
|
bio->bi_iter.bi_sector = sector;
|
2017-08-23 17:10:32 +00:00
|
|
|
bio_set_dev(bio, lc->logdev->bdev);
|
2019-06-05 13:27:08 +00:00
|
|
|
bio->bi_end_io = (sector == WRITE_LOG_SUPER_SECTOR) ?
|
|
|
|
log_end_super : log_end_io;
|
2015-03-20 14:50:37 +00:00
|
|
|
bio->bi_private = lc;
|
2016-06-05 19:32:04 +00:00
|
|
|
bio_set_op_attrs(bio, REQ_OP_WRITE, 0);
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
page = alloc_page(GFP_KERNEL);
|
|
|
|
if (!page) {
|
|
|
|
DMERR("Couldn't alloc log page");
|
|
|
|
bio_put(bio);
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
|
|
|
|
ptr = kmap_atomic(page);
|
|
|
|
memcpy(ptr, entry, entrylen);
|
|
|
|
if (datalen)
|
|
|
|
memcpy(ptr + entrylen, data, datalen);
|
|
|
|
memset(ptr + entrylen + datalen, 0,
|
|
|
|
lc->sectorsize - entrylen - datalen);
|
|
|
|
kunmap_atomic(ptr);
|
|
|
|
|
|
|
|
ret = bio_add_page(bio, page, lc->sectorsize, 0);
|
|
|
|
if (ret != lc->sectorsize) {
|
|
|
|
DMERR("Couldn't add page to the log block");
|
|
|
|
goto error_bio;
|
|
|
|
}
|
2016-06-05 19:31:41 +00:00
|
|
|
submit_bio(bio);
|
2015-03-20 14:50:37 +00:00
|
|
|
return 0;
|
|
|
|
error_bio:
|
|
|
|
bio_put(bio);
|
|
|
|
__free_page(page);
|
|
|
|
error:
|
|
|
|
put_io_block(lc);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2017-10-20 05:24:03 +00:00
|
|
|
static int write_inline_data(struct log_writes_c *lc, void *entry,
|
|
|
|
size_t entrylen, void *data, size_t datalen,
|
|
|
|
sector_t sector)
|
|
|
|
{
|
|
|
|
int num_pages, bio_pages, pg_datalen, pg_sectorlen, i;
|
|
|
|
struct page *page;
|
|
|
|
struct bio *bio;
|
|
|
|
size_t ret;
|
|
|
|
void *ptr;
|
|
|
|
|
|
|
|
while (datalen) {
|
|
|
|
num_pages = ALIGN(datalen, PAGE_SIZE) >> PAGE_SHIFT;
|
|
|
|
bio_pages = min(num_pages, BIO_MAX_PAGES);
|
|
|
|
|
|
|
|
atomic_inc(&lc->io_blocks);
|
|
|
|
|
|
|
|
bio = bio_alloc(GFP_KERNEL, bio_pages);
|
|
|
|
if (!bio) {
|
|
|
|
DMERR("Couldn't alloc inline data bio");
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
|
|
|
|
bio->bi_iter.bi_size = 0;
|
|
|
|
bio->bi_iter.bi_sector = sector;
|
|
|
|
bio_set_dev(bio, lc->logdev->bdev);
|
|
|
|
bio->bi_end_io = log_end_io;
|
|
|
|
bio->bi_private = lc;
|
|
|
|
bio_set_op_attrs(bio, REQ_OP_WRITE, 0);
|
|
|
|
|
|
|
|
for (i = 0; i < bio_pages; i++) {
|
|
|
|
pg_datalen = min_t(int, datalen, PAGE_SIZE);
|
|
|
|
pg_sectorlen = ALIGN(pg_datalen, lc->sectorsize);
|
|
|
|
|
|
|
|
page = alloc_page(GFP_KERNEL);
|
|
|
|
if (!page) {
|
|
|
|
DMERR("Couldn't alloc inline data page");
|
|
|
|
goto error_bio;
|
|
|
|
}
|
|
|
|
|
|
|
|
ptr = kmap_atomic(page);
|
|
|
|
memcpy(ptr, data, pg_datalen);
|
|
|
|
if (pg_sectorlen > pg_datalen)
|
|
|
|
memset(ptr + pg_datalen, 0, pg_sectorlen - pg_datalen);
|
|
|
|
kunmap_atomic(ptr);
|
|
|
|
|
|
|
|
ret = bio_add_page(bio, page, pg_sectorlen, 0);
|
|
|
|
if (ret != pg_sectorlen) {
|
|
|
|
DMERR("Couldn't add page of inline data");
|
|
|
|
__free_page(page);
|
|
|
|
goto error_bio;
|
|
|
|
}
|
|
|
|
|
|
|
|
datalen -= pg_datalen;
|
|
|
|
data += pg_datalen;
|
|
|
|
}
|
|
|
|
submit_bio(bio);
|
|
|
|
|
|
|
|
sector += bio_pages * PAGE_SECTORS;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
error_bio:
|
|
|
|
bio_free_pages(bio);
|
|
|
|
bio_put(bio);
|
|
|
|
error:
|
|
|
|
put_io_block(lc);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2015-03-20 14:50:37 +00:00
|
|
|
static int log_one_block(struct log_writes_c *lc,
|
|
|
|
struct pending_block *block, sector_t sector)
|
|
|
|
{
|
|
|
|
struct bio *bio;
|
|
|
|
struct log_write_entry entry;
|
2017-10-20 05:24:03 +00:00
|
|
|
size_t metadatalen, ret;
|
2015-03-20 14:50:37 +00:00
|
|
|
int i;
|
|
|
|
|
|
|
|
entry.sector = cpu_to_le64(block->sector);
|
|
|
|
entry.nr_sectors = cpu_to_le64(block->nr_sectors);
|
|
|
|
entry.flags = cpu_to_le64(block->flags);
|
|
|
|
entry.data_len = cpu_to_le64(block->datalen);
|
2017-10-20 05:24:03 +00:00
|
|
|
|
|
|
|
metadatalen = (block->flags & LOG_MARK_FLAG) ? block->datalen : 0;
|
2015-03-20 14:50:37 +00:00
|
|
|
if (write_metadata(lc, &entry, sizeof(entry), block->data,
|
2017-10-20 05:24:03 +00:00
|
|
|
metadatalen, sector)) {
|
2015-03-20 14:50:37 +00:00
|
|
|
free_pending_block(lc, block);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2017-10-20 05:24:03 +00:00
|
|
|
sector += dev_to_bio_sectors(lc, 1);
|
|
|
|
|
|
|
|
if (block->datalen && metadatalen == 0) {
|
|
|
|
if (write_inline_data(lc, &entry, sizeof(entry), block->data,
|
|
|
|
block->datalen, sector)) {
|
|
|
|
free_pending_block(lc, block);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
/* we don't support both inline data & bio data */
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2015-03-20 14:50:37 +00:00
|
|
|
if (!block->vec_cnt)
|
|
|
|
goto out;
|
|
|
|
|
2016-08-30 20:11:53 +00:00
|
|
|
atomic_inc(&lc->io_blocks);
|
2016-08-30 20:20:55 +00:00
|
|
|
bio = bio_alloc(GFP_KERNEL, min(block->vec_cnt, BIO_MAX_PAGES));
|
2015-03-20 14:50:37 +00:00
|
|
|
if (!bio) {
|
|
|
|
DMERR("Couldn't alloc log bio");
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
bio->bi_iter.bi_size = 0;
|
|
|
|
bio->bi_iter.bi_sector = sector;
|
2017-08-23 17:10:32 +00:00
|
|
|
bio_set_dev(bio, lc->logdev->bdev);
|
2015-03-20 14:50:37 +00:00
|
|
|
bio->bi_end_io = log_end_io;
|
|
|
|
bio->bi_private = lc;
|
2016-06-05 19:32:04 +00:00
|
|
|
bio_set_op_attrs(bio, REQ_OP_WRITE, 0);
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
for (i = 0; i < block->vec_cnt; i++) {
|
|
|
|
/*
|
|
|
|
* The page offset is always 0 because we allocate a new page
|
|
|
|
* for every bvec in the original bio for simplicity sake.
|
|
|
|
*/
|
|
|
|
ret = bio_add_page(bio, block->vecs[i].bv_page,
|
|
|
|
block->vecs[i].bv_len, 0);
|
|
|
|
if (ret != block->vecs[i].bv_len) {
|
|
|
|
atomic_inc(&lc->io_blocks);
|
2016-06-05 19:31:41 +00:00
|
|
|
submit_bio(bio);
|
2016-08-30 20:20:55 +00:00
|
|
|
bio = bio_alloc(GFP_KERNEL, min(block->vec_cnt - i, BIO_MAX_PAGES));
|
2015-03-20 14:50:37 +00:00
|
|
|
if (!bio) {
|
|
|
|
DMERR("Couldn't alloc log bio");
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
bio->bi_iter.bi_size = 0;
|
|
|
|
bio->bi_iter.bi_sector = sector;
|
2017-08-23 17:10:32 +00:00
|
|
|
bio_set_dev(bio, lc->logdev->bdev);
|
2015-03-20 14:50:37 +00:00
|
|
|
bio->bi_end_io = log_end_io;
|
|
|
|
bio->bi_private = lc;
|
2016-06-05 19:32:04 +00:00
|
|
|
bio_set_op_attrs(bio, REQ_OP_WRITE, 0);
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
ret = bio_add_page(bio, block->vecs[i].bv_page,
|
|
|
|
block->vecs[i].bv_len, 0);
|
|
|
|
if (ret != block->vecs[i].bv_len) {
|
|
|
|
DMERR("Couldn't add page on new bio?");
|
|
|
|
bio_put(bio);
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
sector += block->vecs[i].bv_len >> SECTOR_SHIFT;
|
|
|
|
}
|
2016-06-05 19:31:41 +00:00
|
|
|
submit_bio(bio);
|
2015-03-20 14:50:37 +00:00
|
|
|
out:
|
|
|
|
kfree(block->data);
|
|
|
|
kfree(block);
|
|
|
|
put_pending_block(lc);
|
|
|
|
return 0;
|
|
|
|
error:
|
|
|
|
free_pending_block(lc, block);
|
|
|
|
put_io_block(lc);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int log_super(struct log_writes_c *lc)
|
|
|
|
{
|
|
|
|
struct log_write_super super;
|
|
|
|
|
|
|
|
super.magic = cpu_to_le64(WRITE_LOG_MAGIC);
|
|
|
|
super.version = cpu_to_le64(WRITE_LOG_VERSION);
|
|
|
|
super.nr_entries = cpu_to_le64(lc->logged_entries);
|
|
|
|
super.sectorsize = cpu_to_le32(lc->sectorsize);
|
|
|
|
|
2019-06-05 13:27:08 +00:00
|
|
|
if (write_metadata(lc, &super, sizeof(super), NULL, 0,
|
|
|
|
WRITE_LOG_SUPER_SECTOR)) {
|
2015-03-20 14:50:37 +00:00
|
|
|
DMERR("Couldn't write super");
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2019-06-05 13:27:08 +00:00
|
|
|
/*
|
|
|
|
* Super sector should be writen in-order, otherwise the
|
|
|
|
* nr_entries could be rewritten incorrectly by an old bio.
|
|
|
|
*/
|
|
|
|
wait_for_completion_io(&lc->super_done);
|
|
|
|
|
2015-03-20 14:50:37 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline sector_t logdev_last_sector(struct log_writes_c *lc)
|
|
|
|
{
|
|
|
|
return i_size_read(lc->logdev->bdev->bd_inode) >> SECTOR_SHIFT;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int log_writes_kthread(void *arg)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = (struct log_writes_c *)arg;
|
|
|
|
sector_t sector = 0;
|
|
|
|
|
|
|
|
while (!kthread_should_stop()) {
|
|
|
|
bool super = false;
|
|
|
|
bool logging_enabled;
|
|
|
|
struct pending_block *block = NULL;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
if (!list_empty(&lc->logging_blocks)) {
|
|
|
|
block = list_first_entry(&lc->logging_blocks,
|
|
|
|
struct pending_block, list);
|
|
|
|
list_del_init(&block->list);
|
|
|
|
if (!lc->logging_enabled)
|
|
|
|
goto next;
|
|
|
|
|
|
|
|
sector = lc->next_sector;
|
2017-07-28 14:42:25 +00:00
|
|
|
if (!(block->flags & LOG_DISCARD_FLAG))
|
|
|
|
lc->next_sector += dev_to_bio_sectors(lc, block->nr_sectors);
|
|
|
|
lc->next_sector += dev_to_bio_sectors(lc, 1);
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Apparently the size of the device may not be known
|
|
|
|
* right away, so handle this properly.
|
|
|
|
*/
|
|
|
|
if (!lc->end_sector)
|
|
|
|
lc->end_sector = logdev_last_sector(lc);
|
|
|
|
if (lc->end_sector &&
|
|
|
|
lc->next_sector >= lc->end_sector) {
|
|
|
|
DMERR("Ran out of space on the logdev");
|
|
|
|
lc->logging_enabled = false;
|
|
|
|
goto next;
|
|
|
|
}
|
|
|
|
lc->logged_entries++;
|
|
|
|
atomic_inc(&lc->io_blocks);
|
|
|
|
|
|
|
|
super = (block->flags & (LOG_FUA_FLAG | LOG_MARK_FLAG));
|
|
|
|
if (super)
|
|
|
|
atomic_inc(&lc->io_blocks);
|
|
|
|
}
|
|
|
|
next:
|
|
|
|
logging_enabled = lc->logging_enabled;
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
|
|
|
if (block) {
|
|
|
|
if (logging_enabled) {
|
|
|
|
ret = log_one_block(lc, block, sector);
|
|
|
|
if (!ret && super)
|
|
|
|
ret = log_super(lc);
|
|
|
|
if (ret) {
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
lc->logging_enabled = false;
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
|
|
|
}
|
|
|
|
} else
|
|
|
|
free_pending_block(lc, block);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!try_to_freeze()) {
|
|
|
|
set_current_state(TASK_INTERRUPTIBLE);
|
|
|
|
if (!kthread_should_stop() &&
|
2017-07-28 14:42:24 +00:00
|
|
|
list_empty(&lc->logging_blocks))
|
2015-03-20 14:50:37 +00:00
|
|
|
schedule();
|
|
|
|
__set_current_state(TASK_RUNNING);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Construct a log-writes mapping:
|
|
|
|
* log-writes <dev_path> <log_dev_path>
|
|
|
|
*/
|
|
|
|
static int log_writes_ctr(struct dm_target *ti, unsigned int argc, char **argv)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc;
|
|
|
|
struct dm_arg_set as;
|
|
|
|
const char *devname, *logdevname;
|
2015-07-31 13:20:36 +00:00
|
|
|
int ret;
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
as.argc = argc;
|
|
|
|
as.argv = argv;
|
|
|
|
|
|
|
|
if (argc < 2) {
|
|
|
|
ti->error = "Invalid argument count";
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
lc = kzalloc(sizeof(struct log_writes_c), GFP_KERNEL);
|
|
|
|
if (!lc) {
|
|
|
|
ti->error = "Cannot allocate context";
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
spin_lock_init(&lc->blocks_lock);
|
|
|
|
INIT_LIST_HEAD(&lc->unflushed_blocks);
|
|
|
|
INIT_LIST_HEAD(&lc->logging_blocks);
|
|
|
|
init_waitqueue_head(&lc->wait);
|
2019-06-05 13:27:08 +00:00
|
|
|
init_completion(&lc->super_done);
|
2015-03-20 14:50:37 +00:00
|
|
|
atomic_set(&lc->io_blocks, 0);
|
|
|
|
atomic_set(&lc->pending_blocks, 0);
|
|
|
|
|
|
|
|
devname = dm_shift_arg(&as);
|
2015-07-31 13:20:36 +00:00
|
|
|
ret = dm_get_device(ti, devname, dm_table_get_mode(ti->table), &lc->dev);
|
|
|
|
if (ret) {
|
2015-03-20 14:50:37 +00:00
|
|
|
ti->error = "Device lookup failed";
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
logdevname = dm_shift_arg(&as);
|
2015-07-31 13:20:36 +00:00
|
|
|
ret = dm_get_device(ti, logdevname, dm_table_get_mode(ti->table),
|
|
|
|
&lc->logdev);
|
|
|
|
if (ret) {
|
2015-03-20 14:50:37 +00:00
|
|
|
ti->error = "Log device lookup failed";
|
|
|
|
dm_put_device(ti, lc->dev);
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
2017-07-28 14:42:25 +00:00
|
|
|
lc->sectorsize = bdev_logical_block_size(lc->dev->bdev);
|
|
|
|
lc->sectorshift = ilog2(lc->sectorsize);
|
2015-03-20 14:50:37 +00:00
|
|
|
lc->log_kthread = kthread_run(log_writes_kthread, lc, "log-write");
|
2016-03-09 23:22:19 +00:00
|
|
|
if (IS_ERR(lc->log_kthread)) {
|
|
|
|
ret = PTR_ERR(lc->log_kthread);
|
2015-03-20 14:50:37 +00:00
|
|
|
ti->error = "Couldn't alloc kthread";
|
|
|
|
dm_put_device(ti, lc->dev);
|
|
|
|
dm_put_device(ti, lc->logdev);
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
2017-07-28 14:42:25 +00:00
|
|
|
/*
|
|
|
|
* next_sector is in 512b sectors to correspond to what bi_sector expects.
|
|
|
|
* The super starts at sector 0, and the next_sector is the next logical
|
|
|
|
* one based on the sectorsize of the device.
|
|
|
|
*/
|
|
|
|
lc->next_sector = lc->sectorsize >> SECTOR_SHIFT;
|
2015-03-20 14:50:37 +00:00
|
|
|
lc->logging_enabled = true;
|
|
|
|
lc->end_sector = logdev_last_sector(lc);
|
|
|
|
lc->device_supports_discard = true;
|
|
|
|
|
|
|
|
ti->num_flush_bios = 1;
|
|
|
|
ti->flush_supported = true;
|
|
|
|
ti->num_discard_bios = 1;
|
|
|
|
ti->discards_supported = true;
|
2016-01-31 18:28:26 +00:00
|
|
|
ti->per_io_data_size = sizeof(struct per_bio_data);
|
2015-03-20 14:50:37 +00:00
|
|
|
ti->private = lc;
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
bad:
|
|
|
|
kfree(lc);
|
2015-07-31 13:20:36 +00:00
|
|
|
return ret;
|
2015-03-20 14:50:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int log_mark(struct log_writes_c *lc, char *data)
|
|
|
|
{
|
|
|
|
struct pending_block *block;
|
|
|
|
size_t maxsize = lc->sectorsize - sizeof(struct log_write_entry);
|
|
|
|
|
|
|
|
block = kzalloc(sizeof(struct pending_block), GFP_KERNEL);
|
|
|
|
if (!block) {
|
|
|
|
DMERR("Error allocating pending block");
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2017-12-12 09:39:10 +00:00
|
|
|
block->data = kstrndup(data, maxsize - 1, GFP_KERNEL);
|
2015-03-20 14:50:37 +00:00
|
|
|
if (!block->data) {
|
|
|
|
DMERR("Error copying mark data");
|
|
|
|
kfree(block);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
atomic_inc(&lc->pending_blocks);
|
|
|
|
block->datalen = strlen(block->data);
|
|
|
|
block->flags |= LOG_MARK_FLAG;
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
list_add_tail(&block->list, &lc->logging_blocks);
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
|
|
|
wake_up_process(lc->log_kthread);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void log_writes_dtr(struct dm_target *ti)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
list_splice_init(&lc->unflushed_blocks, &lc->logging_blocks);
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is just nice to have since it'll update the super to include the
|
|
|
|
* unflushed blocks, if it fails we don't really care.
|
|
|
|
*/
|
|
|
|
log_mark(lc, "dm-log-writes-end");
|
|
|
|
wake_up_process(lc->log_kthread);
|
|
|
|
wait_event(lc->wait, !atomic_read(&lc->io_blocks) &&
|
|
|
|
!atomic_read(&lc->pending_blocks));
|
|
|
|
kthread_stop(lc->log_kthread);
|
|
|
|
|
|
|
|
WARN_ON(!list_empty(&lc->logging_blocks));
|
|
|
|
WARN_ON(!list_empty(&lc->unflushed_blocks));
|
|
|
|
dm_put_device(ti, lc->dev);
|
|
|
|
dm_put_device(ti, lc->logdev);
|
|
|
|
kfree(lc);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void normal_map_bio(struct dm_target *ti, struct bio *bio)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
|
2017-08-23 17:10:32 +00:00
|
|
|
bio_set_dev(bio, lc->dev->bdev);
|
2015-03-20 14:50:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int log_writes_map(struct dm_target *ti, struct bio *bio)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
struct per_bio_data *pb = dm_per_bio_data(bio, sizeof(struct per_bio_data));
|
|
|
|
struct pending_block *block;
|
|
|
|
struct bvec_iter iter;
|
|
|
|
struct bio_vec bv;
|
|
|
|
size_t alloc_size;
|
|
|
|
int i = 0;
|
2016-08-05 21:35:16 +00:00
|
|
|
bool flush_bio = (bio->bi_opf & REQ_PREFLUSH);
|
|
|
|
bool fua_bio = (bio->bi_opf & REQ_FUA);
|
2016-06-05 19:32:04 +00:00
|
|
|
bool discard_bio = (bio_op(bio) == REQ_OP_DISCARD);
|
2018-02-28 07:32:47 +00:00
|
|
|
bool meta_bio = (bio->bi_opf & REQ_META);
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
pb->block = NULL;
|
|
|
|
|
|
|
|
/* Don't bother doing anything if logging has been disabled */
|
|
|
|
if (!lc->logging_enabled)
|
|
|
|
goto map_bio;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Map reads as normal.
|
|
|
|
*/
|
|
|
|
if (bio_data_dir(bio) == READ)
|
|
|
|
goto map_bio;
|
|
|
|
|
|
|
|
/* No sectors and not a flush? Don't care */
|
|
|
|
if (!bio_sectors(bio) && !flush_bio)
|
|
|
|
goto map_bio;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Discards will have bi_size set but there's no actual data, so just
|
|
|
|
* allocate the size of the pending block.
|
|
|
|
*/
|
|
|
|
if (discard_bio)
|
|
|
|
alloc_size = sizeof(struct pending_block);
|
|
|
|
else
|
2019-06-12 06:14:46 +00:00
|
|
|
alloc_size = struct_size(block, vecs, bio_segments(bio));
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
block = kzalloc(alloc_size, GFP_NOIO);
|
|
|
|
if (!block) {
|
|
|
|
DMERR("Error allocating pending block");
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
lc->logging_enabled = false;
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
2017-06-03 07:38:02 +00:00
|
|
|
return DM_MAPIO_KILL;
|
2015-03-20 14:50:37 +00:00
|
|
|
}
|
|
|
|
INIT_LIST_HEAD(&block->list);
|
|
|
|
pb->block = block;
|
|
|
|
atomic_inc(&lc->pending_blocks);
|
|
|
|
|
|
|
|
if (flush_bio)
|
|
|
|
block->flags |= LOG_FLUSH_FLAG;
|
|
|
|
if (fua_bio)
|
|
|
|
block->flags |= LOG_FUA_FLAG;
|
|
|
|
if (discard_bio)
|
|
|
|
block->flags |= LOG_DISCARD_FLAG;
|
2018-02-28 07:32:47 +00:00
|
|
|
if (meta_bio)
|
|
|
|
block->flags |= LOG_METADATA_FLAG;
|
2015-03-20 14:50:37 +00:00
|
|
|
|
2017-07-28 14:42:25 +00:00
|
|
|
block->sector = bio_to_dev_sectors(lc, bio->bi_iter.bi_sector);
|
|
|
|
block->nr_sectors = bio_to_dev_sectors(lc, bio_sectors(bio));
|
2015-03-20 14:50:37 +00:00
|
|
|
|
|
|
|
/* We don't need the data, just submit */
|
|
|
|
if (discard_bio) {
|
|
|
|
WARN_ON(flush_bio || fua_bio);
|
|
|
|
if (lc->device_supports_discard)
|
|
|
|
goto map_bio;
|
2015-07-20 13:29:37 +00:00
|
|
|
bio_endio(bio);
|
2015-03-20 14:50:37 +00:00
|
|
|
return DM_MAPIO_SUBMITTED;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Flush bio, splice the unflushed blocks onto this list and submit */
|
|
|
|
if (flush_bio && !bio_sectors(bio)) {
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
list_splice_init(&lc->unflushed_blocks, &block->list);
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
|
|
|
goto map_bio;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We will write this bio somewhere else way later so we need to copy
|
|
|
|
* the actual contents into new pages so we know the data will always be
|
|
|
|
* there.
|
|
|
|
*
|
|
|
|
* We do this because this could be a bio from O_DIRECT in which case we
|
|
|
|
* can't just hold onto the page until some later point, we have to
|
|
|
|
* manually copy the contents.
|
|
|
|
*/
|
|
|
|
bio_for_each_segment(bv, bio, iter) {
|
|
|
|
struct page *page;
|
|
|
|
void *src, *dst;
|
|
|
|
|
|
|
|
page = alloc_page(GFP_NOIO);
|
|
|
|
if (!page) {
|
|
|
|
DMERR("Error allocing page");
|
|
|
|
free_pending_block(lc, block);
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
lc->logging_enabled = false;
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
2017-06-03 07:38:02 +00:00
|
|
|
return DM_MAPIO_KILL;
|
2015-03-20 14:50:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
src = kmap_atomic(bv.bv_page);
|
|
|
|
dst = kmap_atomic(page);
|
|
|
|
memcpy(dst, src + bv.bv_offset, bv.bv_len);
|
|
|
|
kunmap_atomic(dst);
|
|
|
|
kunmap_atomic(src);
|
|
|
|
block->vecs[i].bv_page = page;
|
|
|
|
block->vecs[i].bv_len = bv.bv_len;
|
|
|
|
block->vec_cnt++;
|
|
|
|
i++;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Had a flush with data in it, weird */
|
|
|
|
if (flush_bio) {
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
list_splice_init(&lc->unflushed_blocks, &block->list);
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
|
|
|
}
|
|
|
|
map_bio:
|
|
|
|
normal_map_bio(ti, bio);
|
|
|
|
return DM_MAPIO_REMAPPED;
|
|
|
|
}
|
|
|
|
|
2017-06-03 07:38:06 +00:00
|
|
|
static int normal_end_io(struct dm_target *ti, struct bio *bio,
|
|
|
|
blk_status_t *error)
|
2015-03-20 14:50:37 +00:00
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
struct per_bio_data *pb = dm_per_bio_data(bio, sizeof(struct per_bio_data));
|
|
|
|
|
|
|
|
if (bio_data_dir(bio) == WRITE && pb->block) {
|
|
|
|
struct pending_block *block = pb->block;
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&lc->blocks_lock, flags);
|
|
|
|
if (block->flags & LOG_FLUSH_FLAG) {
|
|
|
|
list_splice_tail_init(&block->list, &lc->logging_blocks);
|
|
|
|
list_add_tail(&block->list, &lc->logging_blocks);
|
|
|
|
wake_up_process(lc->log_kthread);
|
|
|
|
} else if (block->flags & LOG_FUA_FLAG) {
|
|
|
|
list_add_tail(&block->list, &lc->logging_blocks);
|
|
|
|
wake_up_process(lc->log_kthread);
|
|
|
|
} else
|
|
|
|
list_add_tail(&block->list, &lc->unflushed_blocks);
|
|
|
|
spin_unlock_irqrestore(&lc->blocks_lock, flags);
|
|
|
|
}
|
|
|
|
|
2017-06-03 07:38:03 +00:00
|
|
|
return DM_ENDIO_DONE;
|
2015-03-20 14:50:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* INFO format: <logged entries> <highest allocated sector>
|
|
|
|
*/
|
|
|
|
static void log_writes_status(struct dm_target *ti, status_type_t type,
|
|
|
|
unsigned status_flags, char *result,
|
|
|
|
unsigned maxlen)
|
|
|
|
{
|
|
|
|
unsigned sz = 0;
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
|
|
|
|
switch (type) {
|
|
|
|
case STATUSTYPE_INFO:
|
|
|
|
DMEMIT("%llu %llu", lc->logged_entries,
|
|
|
|
(unsigned long long)lc->next_sector - 1);
|
|
|
|
if (!lc->logging_enabled)
|
|
|
|
DMEMIT(" logging_disabled");
|
|
|
|
break;
|
|
|
|
|
|
|
|
case STATUSTYPE_TABLE:
|
|
|
|
DMEMIT("%s %s", lc->dev->name, lc->logdev->name);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-10-15 12:10:50 +00:00
|
|
|
static int log_writes_prepare_ioctl(struct dm_target *ti,
|
2018-04-03 20:54:10 +00:00
|
|
|
struct block_device **bdev)
|
2015-03-20 14:50:37 +00:00
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
struct dm_dev *dev = lc->dev;
|
|
|
|
|
2015-10-15 12:10:50 +00:00
|
|
|
*bdev = dev->bdev;
|
2015-03-20 14:50:37 +00:00
|
|
|
/*
|
|
|
|
* Only pass ioctls through if the device sizes match exactly.
|
|
|
|
*/
|
|
|
|
if (ti->len != i_size_read(dev->bdev->bd_inode) >> SECTOR_SHIFT)
|
2015-10-15 12:10:50 +00:00
|
|
|
return 1;
|
|
|
|
return 0;
|
2015-03-20 14:50:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int log_writes_iterate_devices(struct dm_target *ti,
|
|
|
|
iterate_devices_callout_fn fn,
|
|
|
|
void *data)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
|
|
|
|
return fn(ti, lc->dev, 0, ti->len, data);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Messages supported:
|
|
|
|
* mark <mark data> - specify the marked data.
|
|
|
|
*/
|
2018-02-28 20:59:59 +00:00
|
|
|
static int log_writes_message(struct dm_target *ti, unsigned argc, char **argv,
|
|
|
|
char *result, unsigned maxlen)
|
2015-03-20 14:50:37 +00:00
|
|
|
{
|
|
|
|
int r = -EINVAL;
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
|
|
|
|
if (argc != 2) {
|
|
|
|
DMWARN("Invalid log-writes message arguments, expect 2 arguments, got %d", argc);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!strcasecmp(argv[0], "mark"))
|
|
|
|
r = log_mark(lc, argv[1]);
|
|
|
|
else
|
|
|
|
DMWARN("Unrecognised log writes target message received: %s", argv[0]);
|
|
|
|
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void log_writes_io_hints(struct dm_target *ti, struct queue_limits *limits)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
struct request_queue *q = bdev_get_queue(lc->dev->bdev);
|
|
|
|
|
|
|
|
if (!q || !blk_queue_discard(q)) {
|
|
|
|
lc->device_supports_discard = false;
|
2017-07-28 14:42:25 +00:00
|
|
|
limits->discard_granularity = lc->sectorsize;
|
2015-03-20 14:50:37 +00:00
|
|
|
limits->max_discard_sectors = (UINT_MAX >> SECTOR_SHIFT);
|
|
|
|
}
|
2017-07-28 14:42:25 +00:00
|
|
|
limits->logical_block_size = bdev_logical_block_size(lc->dev->bdev);
|
|
|
|
limits->physical_block_size = bdev_physical_block_size(lc->dev->bdev);
|
|
|
|
limits->io_min = limits->physical_block_size;
|
2015-03-20 14:50:37 +00:00
|
|
|
}
|
|
|
|
|
2018-03-30 00:22:13 +00:00
|
|
|
#if IS_ENABLED(CONFIG_DAX_DRIVER)
|
|
|
|
static int log_dax(struct log_writes_c *lc, sector_t sector, size_t bytes,
|
|
|
|
struct iov_iter *i)
|
|
|
|
{
|
|
|
|
struct pending_block *block;
|
|
|
|
|
|
|
|
if (!bytes)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
block = kzalloc(sizeof(struct pending_block), GFP_KERNEL);
|
|
|
|
if (!block) {
|
|
|
|
DMERR("Error allocating dax pending block");
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
|
|
|
block->data = kzalloc(bytes, GFP_KERNEL);
|
|
|
|
if (!block->data) {
|
|
|
|
DMERR("Error allocating dax data space");
|
|
|
|
kfree(block);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* write data provided via the iterator */
|
|
|
|
if (!copy_from_iter(block->data, bytes, i)) {
|
|
|
|
DMERR("Error copying dax data");
|
|
|
|
kfree(block->data);
|
|
|
|
kfree(block);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* rewind the iterator so that the block driver can use it */
|
|
|
|
iov_iter_revert(i, bytes);
|
|
|
|
|
|
|
|
block->datalen = bytes;
|
|
|
|
block->sector = bio_to_dev_sectors(lc, sector);
|
|
|
|
block->nr_sectors = ALIGN(bytes, lc->sectorsize) >> lc->sectorshift;
|
|
|
|
|
|
|
|
atomic_inc(&lc->pending_blocks);
|
|
|
|
spin_lock_irq(&lc->blocks_lock);
|
|
|
|
list_add_tail(&block->list, &lc->unflushed_blocks);
|
|
|
|
spin_unlock_irq(&lc->blocks_lock);
|
|
|
|
wake_up_process(lc->log_kthread);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2017-10-20 05:24:04 +00:00
|
|
|
static long log_writes_dax_direct_access(struct dm_target *ti, pgoff_t pgoff,
|
|
|
|
long nr_pages, void **kaddr, pfn_t *pfn)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
sector_t sector = pgoff * PAGE_SECTORS;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = bdev_dax_pgoff(lc->dev->bdev, sector, nr_pages * PAGE_SIZE, &pgoff);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
return dax_direct_access(lc->dev->dax_dev, pgoff, nr_pages, kaddr, pfn);
|
|
|
|
}
|
|
|
|
|
|
|
|
static size_t log_writes_dax_copy_from_iter(struct dm_target *ti,
|
|
|
|
pgoff_t pgoff, void *addr, size_t bytes,
|
|
|
|
struct iov_iter *i)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
sector_t sector = pgoff * PAGE_SECTORS;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (bdev_dax_pgoff(lc->dev->bdev, sector, ALIGN(bytes, PAGE_SIZE), &pgoff))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* Don't bother doing anything if logging has been disabled */
|
|
|
|
if (!lc->logging_enabled)
|
|
|
|
goto dax_copy;
|
|
|
|
|
|
|
|
err = log_dax(lc, sector, bytes, i);
|
|
|
|
if (err) {
|
|
|
|
DMWARN("Error %d logging DAX write", err);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
dax_copy:
|
|
|
|
return dax_copy_from_iter(lc->dev->dax_dev, pgoff, addr, bytes, i);
|
|
|
|
}
|
2018-05-02 13:46:33 +00:00
|
|
|
|
|
|
|
static size_t log_writes_dax_copy_to_iter(struct dm_target *ti,
|
|
|
|
pgoff_t pgoff, void *addr, size_t bytes,
|
|
|
|
struct iov_iter *i)
|
|
|
|
{
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
sector_t sector = pgoff * PAGE_SECTORS;
|
|
|
|
|
|
|
|
if (bdev_dax_pgoff(lc->dev->bdev, sector, ALIGN(bytes, PAGE_SIZE), &pgoff))
|
|
|
|
return 0;
|
|
|
|
return dax_copy_to_iter(lc->dev->dax_dev, pgoff, addr, bytes, i);
|
|
|
|
}
|
|
|
|
|
2020-02-28 16:34:54 +00:00
|
|
|
static int log_writes_dax_zero_page_range(struct dm_target *ti, pgoff_t pgoff,
|
|
|
|
size_t nr_pages)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
struct log_writes_c *lc = ti->private;
|
|
|
|
sector_t sector = pgoff * PAGE_SECTORS;
|
|
|
|
|
|
|
|
ret = bdev_dax_pgoff(lc->dev->bdev, sector, nr_pages << PAGE_SHIFT,
|
|
|
|
&pgoff);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
return dax_zero_page_range(lc->dev->dax_dev, pgoff,
|
|
|
|
nr_pages << PAGE_SHIFT);
|
|
|
|
}
|
|
|
|
|
2018-03-30 00:22:13 +00:00
|
|
|
#else
|
|
|
|
#define log_writes_dax_direct_access NULL
|
|
|
|
#define log_writes_dax_copy_from_iter NULL
|
2018-05-02 13:46:33 +00:00
|
|
|
#define log_writes_dax_copy_to_iter NULL
|
2020-02-28 16:34:54 +00:00
|
|
|
#define log_writes_dax_zero_page_range NULL
|
2018-03-30 00:22:13 +00:00
|
|
|
#endif
|
2017-10-20 05:24:04 +00:00
|
|
|
|
2015-03-20 14:50:37 +00:00
|
|
|
static struct target_type log_writes_target = {
|
|
|
|
.name = "log-writes",
|
2017-10-20 05:24:04 +00:00
|
|
|
.version = {1, 1, 0},
|
2015-03-20 14:50:37 +00:00
|
|
|
.module = THIS_MODULE,
|
|
|
|
.ctr = log_writes_ctr,
|
|
|
|
.dtr = log_writes_dtr,
|
|
|
|
.map = log_writes_map,
|
|
|
|
.end_io = normal_end_io,
|
|
|
|
.status = log_writes_status,
|
2015-10-15 12:10:50 +00:00
|
|
|
.prepare_ioctl = log_writes_prepare_ioctl,
|
2015-03-20 14:50:37 +00:00
|
|
|
.message = log_writes_message,
|
|
|
|
.iterate_devices = log_writes_iterate_devices,
|
|
|
|
.io_hints = log_writes_io_hints,
|
2017-10-20 05:24:04 +00:00
|
|
|
.direct_access = log_writes_dax_direct_access,
|
|
|
|
.dax_copy_from_iter = log_writes_dax_copy_from_iter,
|
2018-05-02 13:46:33 +00:00
|
|
|
.dax_copy_to_iter = log_writes_dax_copy_to_iter,
|
2020-02-28 16:34:54 +00:00
|
|
|
.dax_zero_page_range = log_writes_dax_zero_page_range,
|
2015-03-20 14:50:37 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
static int __init dm_log_writes_init(void)
|
|
|
|
{
|
|
|
|
int r = dm_register_target(&log_writes_target);
|
|
|
|
|
|
|
|
if (r < 0)
|
|
|
|
DMERR("register failed %d", r);
|
|
|
|
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __exit dm_log_writes_exit(void)
|
|
|
|
{
|
|
|
|
dm_unregister_target(&log_writes_target);
|
|
|
|
}
|
|
|
|
|
|
|
|
module_init(dm_log_writes_init);
|
|
|
|
module_exit(dm_log_writes_exit);
|
|
|
|
|
|
|
|
MODULE_DESCRIPTION(DM_NAME " log writes target");
|
|
|
|
MODULE_AUTHOR("Josef Bacik <jbacik@fb.com>");
|
|
|
|
MODULE_LICENSE("GPL");
|