/*
* simp_blkdev_make_request
*/
static void simp_blkdev_make_request(struct request_queue *q, struct bio *bio)
{
struct bio_vec *bvec;
int i;
void *dsk_mem;
/* 判断 IO 请求是否超出实际 RanDisk 大小 */
if((bio->bi_sector << 9) + bio->bi_size > SIMP_BLKDEV_BYTES)
{
printk(KERN_ERR SIMP_BLKDEV_DISKNAME
": bad request: block=%llu, count=%u\n",
(unsigned long long)bio->bi_sector, bio->bi_size);
/* 条件编译，对于 2.4 内核之后的变量多了一个，伤不起阿！ */
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 24)
/* 返回这个对 bio 请求的处理结果，－EIO 表示出错 */
bio_endio(bio, 0, -EIO);
#else
bio_endio(bio, -EIO);
#endif
return ;
}
dsk_mem = simp_blkdev_data + (bio->bi_sector << 9);
/* 遍历 bio 中所有的 bvec */
bio_for_each_segment(bvec, bio, i)
{
void *iovec_mem;
switch(bio_rw(bio))
{
case READ:
case READA: /* 读和预读都进行同样的处理 */
/* 用 kmalloc 将请求页映射到非线性映射区域进行
* 访问，这种方法主要是为了兼容高端内存，
* (bvec->bv_page 可能源于高端内存） */
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
memcpy(iovec_mem, dsk_mem, bvec->bv_len);
kunmap(bvec->bv_page);
break;
case WRITE:
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
memcpy(dsk_mem, iovec_mem, bvec->bv_len);
kunmap(bvec->bv_page);
break;
default:
printk(KERN_ERR SIMP_BLKDEV_DISKNAME
":Unknown value of bio_rw: %lu\n",
bio_rw(bio));
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 24)
bio_endio(bio, 0, -EIO);
#else
bio_endio(bio, -EIO);
#endif
return ;
}
dsk_mem += bvec->bv_len;
}
static int null_lnvm_submit_io(struct nvm_dev *dev, struct nvm_rq *rqd)
{
struct request_queue *q = dev->q;
struct request *rq;
struct bio *bio = rqd->bio;
rq = blk_mq_alloc_request(q, bio_rw(bio), 0);
if (IS_ERR(rq))
return -ENOMEM;
rq->cmd_type = REQ_TYPE_DRV_PRIV;
rq->__sector = bio->bi_iter.bi_sector;
rq->ioprio = bio_prio(bio);
if (bio_has_data(bio))
rq->nr_phys_segments = bio_phys_segments(q, bio);
rq->__data_len = bio->bi_iter.bi_size;
rq->bio = rq->biotail = bio;
rq->end_io_data = rqd;
blk_execute_rq_nowait(q, NULL, rq, 0, null_lnvm_end_io);
return 0;
}
static int Virtual_blkdev_make_request(struct request_queue *q, struct bio *bio)
{
struct bio_vec *bvec;
int i;
void *dsk_mem;
if ((bio->bi_sector << 9) + bio->bi_size > VIRTUAL_BLKDEV_BYTES) {
printk(KERN_ERR VIRTUAL_BLKDEV_DISKNAME
": bad request: block=%llu, count=%u\n",
(unsigned long long)bio->bi_sector, bio->bi_size);
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 24)
bio_endio(bio, 0, -EIO);
#else
bio_endio(bio, -EIO);
#endif
return 0;
}
dsk_mem = Virtual_blkdev_data + (bio->bi_sector << 9);
bio_for_each_segment(bvec, bio, i) {
void *iovec_mem;
switch (bio_rw(bio)) {
case READ:
case READA:
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
memcpy(iovec_mem, dsk_mem, bvec->bv_len);
kunmap(bvec->bv_page);
break;
case WRITE:
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
memcpy(dsk_mem, iovec_mem, bvec->bv_len);
kunmap(bvec->bv_page);
break;
default:
printk(KERN_ERR VIRTUAL_BLKDEV_DISKNAME
": unknown value of bio_rw: %lu\n",
bio_rw(bio));
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 24)
bio_endio(bio, 0, -EIO);
#else
bio_endio(bio, -EIO);
#endif
return 0;
}
dsk_mem += bvec->bv_len;
}
/*
* The request function that just remaps the bio built up by
* dm_merge_bvec.
*/
static int dm_request(request_queue_t *q, struct bio *bio)
{
int r;
int rw = bio_data_dir(bio);
struct mapped_device *md = q->queuedata;
/*
* There is no use in forwarding any barrier request since we can't
* guarantee it is (or can be) handled by the targets correctly.
*/
if (unlikely(bio_barrier(bio))) {
bio_endio(bio, bio->bi_size, -EOPNOTSUPP);
return 0;
}
down_read(&md->io_lock);
disk_stat_inc(dm_disk(md), ios[rw]);
disk_stat_add(dm_disk(md), sectors[rw], bio_sectors(bio));
/*
* If we're suspended we have to queue
* this io for later.
*/
while (test_bit(DMF_BLOCK_IO, &md->flags)) {
up_read(&md->io_lock);
if (bio_rw(bio) == READA) {
bio_io_error(bio, bio->bi_size);
return 0;
}
r = queue_io(md, bio);
if (r < 0) {
bio_io_error(bio, bio->bi_size);
return 0;
} else if (r == 0)
return 0; /* deferred successfully */
/*
* We're in a while loop, because someone could suspend
* before we get to the following read lock.
*/
down_read(&md->io_lock);
}
__split_bio(md, bio);
up_read(&md->io_lock);
return 0;
}
static int rrpc_setup_rq(struct rrpc *rrpc, struct bio *bio,
struct nvm_rq *rqd, unsigned long flags, uint8_t npages)
{
if (npages > 1) {
rqd->ppa_list = nvm_dev_dma_alloc(rrpc->dev, GFP_KERNEL,
&rqd->dma_ppa_list);
if (!rqd->ppa_list) {
pr_err("rrpc: not able to allocate ppa list\n");
return NVM_IO_ERR;
}
if (bio_rw(bio) == WRITE)
return rrpc_write_ppalist_rq(rrpc, bio, rqd, flags,
npages);
return rrpc_read_ppalist_rq(rrpc, bio, rqd, flags, npages);
}
if (bio_rw(bio) == WRITE)
return rrpc_write_rq(rrpc, bio, rqd, flags);
return rrpc_read_rq(rrpc, bio, rqd, flags);
}
/*
* Return zeros only on reads
*/
static int zero_map(struct dm_target *ti, struct bio *bio)
{
switch(bio_rw(bio)) {
case READ:
zero_fill_bio(bio);
break;
case READA:
/* readahead of null bytes only wastes buffer cache */
return -EIO;
case WRITE:
/* writes get silently dropped */
break;
}
bio_endio(bio, 0);
/* accepted bio, don't make new request */
return DM_MAPIO_SUBMITTED;
}
/*
* The request function that just remaps the bio built up by
* dm_merge_bvec.
*/
static int dm_request(request_queue_t *q, struct bio *bio)
{
int r;
int rw = bio_data_dir(bio);
struct mapped_device *md = q->queuedata;
down_read(&md->io_lock);
disk_stat_inc(dm_disk(md), ios[rw]);
disk_stat_add(dm_disk(md), sectors[rw], bio_sectors(bio));
/*
* If we're suspended we have to queue
* this io for later.
*/
while (test_bit(DMF_BLOCK_IO, &md->flags)) {
up_read(&md->io_lock);
if (bio_rw(bio) == READA) {
bio_io_error(bio, bio->bi_size);
return 0;
}
r = queue_io(md, bio);
if (r < 0) {
bio_io_error(bio, bio->bi_size);
return 0;
} else if (r == 0)
return 0; /* deferred successfully */
/*
* We're in a while loop, because someone could suspend
* before we get to the following read lock.
*/
down_read(&md->io_lock);
}
__split_bio(md, bio);
up_read(&md->io_lock);
return 0;
}
/*
* Return zeros only on reads
*/
static int zero_map(struct dm_target *ti, struct bio *bio,
union map_info *map_context)
{
switch(bio_rw(bio)) {
case READ:
zero_fill_bio(bio);
break;
case READA:
/* readahead of null bytes only wastes buffer cache */
return -EIO;
case WRITE:
/* writes get silently dropped */
break;
}
bio_endio(bio, bio->bi_size, 0);
/* accepted bio, don't make new request */
return 0;
}
/* read, readA or write requests on R_PRIMARY coming from drbd_make_request
*/
void drbd_endio_pri(struct bio *bio, int error)
{
unsigned long flags;
struct drbd_request *req = bio->bi_private;
struct drbd_conf *mdev = req->mdev;
struct bio_and_error m;
enum drbd_req_event what;
int uptodate = bio_flagged(bio, BIO_UPTODATE);
if (error)
dev_warn(DEV, "p %s: error=%d\n",
bio_data_dir(bio) == WRITE ? "write" : "read", error);
if (!error && !uptodate) {
dev_warn(DEV, "p %s: setting error to -EIO\n",
bio_data_dir(bio) == WRITE ? "write" : "read");
/* strange behavior of some lower level drivers...
* fail the request by clearing the uptodate flag,
* but do not return any error?! */
error = -EIO;
}
/* to avoid recursion in __req_mod */
if (unlikely(error)) {
what = (bio_data_dir(bio) == WRITE)
? write_completed_with_error
: (bio_rw(bio) == READ)
? read_completed_with_error
: read_ahead_completed_with_error;
} else
what = completed_ok;
bio_put(req->private_bio);
req->private_bio = ERR_PTR(error);
spin_lock_irqsave(&mdev->req_lock, flags);
__req_mod(req, what, &m);
spin_unlock_irqrestore(&mdev->req_lock, flags);
if (m.bio)
complete_master_bio(mdev, &m);
}
static int nvme_nvm_submit_io(struct nvm_dev *dev, struct nvm_rq *rqd)
{
struct request_queue *q = dev->q;
struct nvme_ns *ns = q->queuedata;
struct request *rq;
struct bio *bio = rqd->bio;
struct nvme_nvm_command *cmd;
rq = blk_mq_alloc_request(q, bio_rw(bio), 0);
if (IS_ERR(rq))
return -ENOMEM;
cmd = kzalloc(sizeof(struct nvme_nvm_command) +
sizeof(struct nvme_nvm_completion), GFP_KERNEL);
if (!cmd) {
blk_mq_free_request(rq);
return -ENOMEM;
}
rq->cmd_type = REQ_TYPE_DRV_PRIV;
rq->ioprio = bio_prio(bio);
if (bio_has_data(bio))
rq->nr_phys_segments = bio_phys_segments(q, bio);
rq->__data_len = bio->bi_iter.bi_size;
rq->bio = rq->biotail = bio;
nvme_nvm_rqtocmd(rq, rqd, ns, cmd);
rq->cmd = (unsigned char *)cmd;
rq->cmd_len = sizeof(struct nvme_nvm_command);
rq->special = cmd + 1;
rq->end_io_data = rqd;
blk_execute_rq_nowait(q, NULL, rq, 0, nvme_nvm_end_io);
return 0;
}
static int sbd_make_request(struct request_queue *q, struct bio *bio)
{
struct bio_vec *bvec;
int i;
unsigned long long dsk_offset;
dsk_offset = bio->bi_sector * 512;
bio_for_each_segment(bvec, bio, i){
unsigned int count_done, count_current;
void *dsk_mem;
void *iovec_mem;
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
count_done = 0;
while(count_done < bvec->bv_len){
count_current = min(bvec->bv_len - count_done,
PAGE_SIZE -
(dsk_offset + count_done) % PAGE_SIZE);
dsk_mem = radix_tree_lookup(&sbd_data,
(dsk_offset + count_done) / PAGE_SIZE);
dsk_mem += (dsk_offset + count_done) % PAGE_SIZE;
switch(bio_rw(bio)){
case READ:
case READA:
memcpy(iovec_mem, dsk_mem, count_current);
break;
case WRITE:
memcpy(dsk_mem, iovec_mem, count_current);
break;
}
count_done += count_current;
}
kunmap(bvec->bv_page);
dsk_offset += bvec->bv_len;
}
int simp_blkdev_make_request(struct request_queue *q, struct bio *bio)
{
struct bio_vec * bvec;
int i;
void *disk_mem;
if( (bio->bi_sector<<9)+bio->bi_size > DEV_CAPACITY )
{
printk("simp_blkdev: bad request: S=%lu, nS=%u\n",
(unsigned long)bio->bi_sector, bio->bi_size);
bio_endio(bio, -EIO);
return 0;
}
disk_mem = simp_blkdev_data + (bio->bi_sector<<9);
bio_for_each_segment(bvec, bio, i)
{
void *iovec_mem;
switch(bio_rw(bio))
{
case READ:
case READA:/*预读*/
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
memcpy(iovec_mem, disk_mem, bvec->bv_len);
kunmap(bvec->bv_page);
break;
case WRITE:
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
memcpy(disk_mem, iovec_mem, bvec->bv_len);
kunmap(bvec->bv_page);
break;
default:
printk("blkdev: unspported bio type\n");
bio_endio(bio, -EIO);
return 0;
}
disk_mem += bvec->bv_len;
}
static int simpleblk_make_request(struct request_queue *q, struct bio *bio)
{
struct bio_vec *bvec;
int i;
void *dsk_mem;
if ((bio->bi_sector << 9) + bio->bi_size > MEMBLK_BYTES)
{
#if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 24)) //XXX not very accurate
bio_endio(bio, 0, -EIO);
#else
bio_endio(bio, -EIO);
#endif
return -EFAULT;
}
dsk_mem = blkdev_data + (bio->bi_sector << 9);//in fact the bio->bi_sector is sector offset not real addr.
switch (bio_rw(bio))
{
case READ:
case READA:
bio_for_each_segment(bvec, bio, i)
{
void *iovec_mem;
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
memcpy(iovec_mem, dsk_mem, bvec->bv_len);
kunmap(bvec->bv_page);
dsk_mem += bvec->bv_len;
}
break;
case WRITE:
bio_for_each_segment(bvec, bio, i)
{
void *iovec_mem;
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
memcpy(dsk_mem, iovec_mem, bvec->bv_len);
kunmap(bvec->bv_page);
dsk_mem += bvec->bv_len;
}
/* Helper for __req_mod().
* Set m->bio to the master bio, if it is fit to be completed,
* or leave it alone (it is initialized to NULL in __req_mod),
* if it has already been completed, or cannot be completed yet.
* If m->bio is set, the error status to be returned is placed in m->error.
*/
static
void drbd_req_complete(struct drbd_request *req, struct bio_and_error *m)
{
const unsigned s = req->rq_state;
struct drbd_conf *mdev = req->w.mdev;
int rw;
int error, ok;
/* we must not complete the master bio, while it is
* still being processed by _drbd_send_zc_bio (drbd_send_dblock)
* not yet acknowledged by the peer
* not yet completed by the local io subsystem
* these flags may get cleared in any order by
* the worker,
* the receiver,
* the bio_endio completion callbacks.
*/
if ((s & RQ_LOCAL_PENDING && !(s & RQ_LOCAL_ABORTED)) ||
(s & RQ_NET_QUEUED) || (s & RQ_NET_PENDING) ||
(s & RQ_COMPLETION_SUSP)) {
dev_err(DEV, "drbd_req_complete: Logic BUG rq_state = 0x%x\n", s);
return;
}
if (!req->master_bio) {
dev_err(DEV, "drbd_req_complete: Logic BUG, master_bio == NULL!\n");
return;
}
rw = bio_rw(req->master_bio);
/*
* figure out whether to report success or failure.
*
* report success when at least one of the operations succeeded.
* or, to put the other way,
* only report failure, when both operations failed.
*
* what to do about the failures is handled elsewhere.
* what we need to do here is just: complete the master_bio.
*
* local completion error, if any, has been stored as ERR_PTR
* in private_bio within drbd_request_endio.
*/
ok = (s & RQ_LOCAL_OK) || (s & RQ_NET_OK);
error = PTR_ERR(req->private_bio);
/* remove the request from the conflict detection
* respective block_id verification hash */
if (!drbd_interval_empty(&req->i)) {
struct rb_root *root;
if (rw == WRITE)
root = &mdev->write_requests;
else
root = &mdev->read_requests;
drbd_remove_request_interval(root, req);
}
/* Before we can signal completion to the upper layers,
* we may need to close the current transfer log epoch.
* We are within the request lock, so we can simply compare
* the request epoch number with the current transfer log
* epoch number. If they match, increase the current_tle_nr,
* and reset the transfer log epoch write_cnt.
*/
if (rw == WRITE &&
req->epoch == atomic_read(&mdev->tconn->current_tle_nr))
start_new_tl_epoch(mdev->tconn);
/* Update disk stats */
_drbd_end_io_acct(mdev, req);
/* If READ failed,
* have it be pushed back to the retry work queue,
* so it will re-enter __drbd_make_request(),
* and be re-assigned to a suitable local or remote path,
* or failed if we do not have access to good data anymore.
*
* Unless it was failed early by __drbd_make_request(),
* because no path was available, in which case
* it was not even added to the transfer_log.
*
* READA may fail, and will not be retried.
*
* WRITE should have used all available paths already.
*/
if (!ok && rw == READ && !list_empty(&req->tl_requests))
req->rq_state |= RQ_POSTPONED;
if (!(req->rq_state & RQ_POSTPONED)) {
m->error = ok ? 0 : (error ?: -EIO);
m->bio = req->master_bio;
req->master_bio = NULL;
}
/*
static int blkdev_make_request(struct request_queue *q,struct bio *bio)
{
struct bio_vec *bvec;
int i;
unsigned long long dsk_offset;
if((bio->bi_sector<<PAGE_SHIFT)+bio->bi_size >BLKDEV_BYTES)
{
printk(KERN_ERR "bad request!bi_sector:%llu,bi_size:%u\n",(unsigned long)bio->bi_sector,bio->bi_size);
bio_endio(bio,-EIO);
}
dsk_offset=bio->bi_sector << PAGE_SHIFT;
bio_for_each_segment(bvec,bio,i)
{
unsigned int count_done,count_current;
void *iovec_mem;
void *dsk_mem;
iovec_mem=kmap(bvec->bv_page) + bvec->bv_offset;
count_done = 0;
while(count_done < bvec->bv_len)
{
count_current = min(bvec->bv_len - count_done,(unsigned int)(PAGE_SIZE-(dsk_offset + count_done)%PAGE_SIZE));
dsk_mem = radix_tree_lookup(&blkdev_data,(dsk_offset + count_done)%PAGE_SIZE);
dsk_mem+=(dsk_offset + count_done)%PAGE_SIZE;
switch(bio_rw(bio))
{
case READ:
case READA:
memcpy(iovec_mem +count_done,dsk_mem,count_current);
break;
case WRITE:
memcpy(dsk_mem,iovec_mem+count_done,count_current);
}
count_done +=count_current;
}
kunmap(bvec->bv_page);
dsk_offset += bvec->bv_len;
}
bio_endio(bio,0);
return 0;
}
*/
static int blkdev_make_request(struct request_queue *q, struct bio *bio)
{
struct bio_vec *bvec;
int i;
unsigned long long dsk_offset;
if ((bio->bi_sector << 9) + bio->bi_size > BLKDEV_BYTES) {
printk(KERN_ERR BLKDEV_DISKNAME
": bad request: block=%llu, count=%u\n",
(unsigned long long)bio->bi_sector, bio->bi_size);
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 24)
bio_endio(bio, 0, -EIO);
#else
bio_endio(bio, -EIO);
#endif
return 0;
}
dsk_offset = bio->bi_sector << 9;
bio_for_each_segment(bvec, bio, i) {
unsigned int count_done, count_current;
void *iovec_mem;
void *dsk_mem;
iovec_mem = kmap(bvec->bv_page) + bvec->bv_offset;
count_done = 0;
while (count_done < bvec->bv_len) {
count_current = min(bvec->bv_len - count_done,
(unsigned int)(PAGE_SIZE
- ((dsk_offset + count_done) & ~PAGE_MASK)));
dsk_mem = radix_tree_lookup(&blkdev_data,
(dsk_offset + count_done) >> PAGE_SHIFT);
if (!dsk_mem) {
printk(KERN_ERR BLKDEV_DISKNAME
": search memory failed: %llu\n",
(dsk_offset + count_done)
>> PAGE_SHIFT);
kunmap(bvec->bv_page);
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 24)
bio_endio(bio, 0, -EIO);
#else
bio_endio(bio, -EIO);
#endif
return 0;
}
dsk_mem += (dsk_offset + count_done) & ~PAGE_MASK;
switch (bio_rw(bio)) {
case READ:
case READA:
memcpy(iovec_mem + count_done, dsk_mem,
count_current);
break;
case WRITE:
memcpy(dsk_mem, iovec_mem + count_done,
count_current);
break;
default:
printk(KERN_ERR BLKDEV_DISKNAME
": unknown value of bio_rw: %lu\n",
bio_rw(bio));
kunmap(bvec->bv_page);
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 24)
bio_endio(bio, 0, -EIO);
#else
bio_endio(bio, -EIO);
#endif
return 0;
}
count_done += count_current;
}