static struct ioreq *ioreq_start(struct XenBlkDev *blkdev)
{
struct ioreq *ioreq = NULL;
if (QLIST_EMPTY(&blkdev->freelist)) {
if (blkdev->requests_total >= max_requests) {
goto out;
}
/* allocate new struct */
ioreq = g_malloc0(sizeof(*ioreq));
ioreq->blkdev = blkdev;
blkdev->requests_total++;
qemu_iovec_init(&ioreq->v, BLKIF_MAX_SEGMENTS_PER_REQUEST);
} else {
/* get one from freelist */
ioreq = QLIST_FIRST(&blkdev->freelist);
QLIST_REMOVE(ioreq, list);
qemu_iovec_reset(&ioreq->v);
}
QLIST_INSERT_HEAD(&blkdev->inflight, ioreq, list);
blkdev->requests_inflight++;
out:
return ioreq;
}
static XenBlockRequest *xen_block_start_request(XenBlockDataPlane *dataplane)
{
XenBlockRequest *request = NULL;
if (QLIST_EMPTY(&dataplane->freelist)) {
if (dataplane->requests_total >= dataplane->max_requests) {
goto out;
}
/* allocate new struct */
request = g_malloc0(sizeof(*request));
request->dataplane = dataplane;
/*
* We cannot need more pages per requests than this, and since we
* re-use requests, allocate the memory once here. It will be freed
* xen_block_dataplane_destroy() when the request list is freed.
*/
request->buf = qemu_memalign(XC_PAGE_SIZE,
BLKIF_MAX_SEGMENTS_PER_REQUEST *
XC_PAGE_SIZE);
dataplane->requests_total++;
qemu_iovec_init(&request->v, 1);
} else {
/* get one from freelist */
request = QLIST_FIRST(&dataplane->freelist);
QLIST_REMOVE(request, list);
}
QLIST_INSERT_HEAD(&dataplane->inflight, request, list);
dataplane->requests_inflight++;
out:
return request;
}
static int coroutine_fn
raw_co_writev_flags(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
QEMUIOVector *qiov, int flags)
{
void *buf = NULL;
BlockDriver *drv;
QEMUIOVector local_qiov;
int ret;
if (bs->probed && sector_num == 0) {
/* As long as these conditions are true, we can't get partial writes to
* the probe buffer and can just directly check the request. */
QEMU_BUILD_BUG_ON(BLOCK_PROBE_BUF_SIZE != 512);
QEMU_BUILD_BUG_ON(BDRV_SECTOR_SIZE != 512);
if (nb_sectors == 0) {
/* qemu_iovec_to_buf() would fail, but we want to return success
* instead of -EINVAL in this case. */
return 0;
}
buf = qemu_try_blockalign(bs->file->bs, 512);
if (!buf) {
ret = -ENOMEM;
goto fail;
}
ret = qemu_iovec_to_buf(qiov, 0, buf, 512);
if (ret != 512) {
ret = -EINVAL;
goto fail;
}
drv = bdrv_probe_all(buf, 512, NULL);
if (drv != bs->drv) {
ret = -EPERM;
goto fail;
}
/* Use the checked buffer, a malicious guest might be overwriting its
* original buffer in the background. */
qemu_iovec_init(&local_qiov, qiov->niov + 1);
qemu_iovec_add(&local_qiov, buf, 512);
qemu_iovec_concat(&local_qiov, qiov, 512, qiov->size - 512);
qiov = &local_qiov;
}
BLKDBG_EVENT(bs->file, BLKDBG_WRITE_AIO);
ret = bdrv_co_do_pwritev(bs->file->bs, sector_num * BDRV_SECTOR_SIZE,
nb_sectors * BDRV_SECTOR_SIZE, qiov, flags);
fail:
if (qiov == &local_qiov) {
qemu_iovec_destroy(&local_qiov);
}
qemu_vfree(buf);
return ret;
}
static inline void flash_sync_area(Flash *s, int64_t off, int64_t len)
{
QEMUIOVector *iov;
if (!s->blk || blk_is_read_only(s->blk)) {
return;
}
assert(!(len % BDRV_SECTOR_SIZE));
iov = g_new(QEMUIOVector, 1);
qemu_iovec_init(iov, 1);
qemu_iovec_add(iov, s->storage + off, len);
blk_aio_pwritev(s->blk, off, iov, 0, blk_sync_complete, iov);
}
static void flash_sync_page(Flash *s, int page)
{
QEMUIOVector *iov;
if (!s->blk || blk_is_read_only(s->blk)) {
return;
}
iov = g_new(QEMUIOVector, 1);
qemu_iovec_init(iov, 1);
qemu_iovec_add(iov, s->storage + page * s->pi->page_size,
s->pi->page_size);
blk_aio_pwritev(s->blk, page * s->pi->page_size, iov, 0,
blk_sync_complete, iov);
}
static void flash_sync_page(Flash *s, int page)
{
if (s->bdrv) {
int bdrv_sector, nb_sectors;
QEMUIOVector iov;
bdrv_sector = (page * s->pi->page_size) / BDRV_SECTOR_SIZE;
nb_sectors = DIV_ROUND_UP(s->pi->page_size, BDRV_SECTOR_SIZE);
qemu_iovec_init(&iov, 1);
qemu_iovec_add(&iov, s->storage + bdrv_sector * BDRV_SECTOR_SIZE,
nb_sectors * BDRV_SECTOR_SIZE);
bdrv_aio_writev(s->bdrv, bdrv_sector, &iov, nb_sectors,
bdrv_sync_complete, NULL);
}
}
/*
* Parse multiple length statements for vectored I/O, and construct an I/O
* vector matching it.
*/
static void *
create_iovec(QEMUIOVector *qiov, char **argv, int nr_iov, int pattern)
{
size_t *sizes = g_new0(size_t, nr_iov);
size_t count = 0;
void *buf = NULL;
void *p;
int i;
for (i = 0; i < nr_iov; i++) {
char *arg = argv[i];
int64_t len;
len = cvtnum(arg);
if (len < 0) {
printf("non-numeric length argument -- %s\n", arg);
goto fail;
}
/* should be SIZE_T_MAX, but that doesn't exist */
if (len > INT_MAX) {
printf("too large length argument -- %s\n", arg);
goto fail;
}
if (len & 0x1ff) {
printf("length argument %" PRId64
" is not sector aligned\n", len);
goto fail;
}
sizes[i] = len;
count += len;
}
qemu_iovec_init(qiov, nr_iov);
buf = p = qemu_io_alloc(count, pattern);
for (i = 0; i < nr_iov; i++) {
qemu_iovec_add(qiov, p, sizes[i]);
p += sizes[i];
}
fail:
g_free(sizes);
return buf;
}
static void flash_sync_page(Flash *s, int page)
{
int blk_sector, nb_sectors;
QEMUIOVector iov;
if (!s->blk || blk_is_read_only(s->blk)) {
return;
}
blk_sector = (page * s->pi->page_size) / BDRV_SECTOR_SIZE;
nb_sectors = DIV_ROUND_UP(s->pi->page_size, BDRV_SECTOR_SIZE);
qemu_iovec_init(&iov, 1);
qemu_iovec_add(&iov, s->storage + blk_sector * BDRV_SECTOR_SIZE,
nb_sectors * BDRV_SECTOR_SIZE);
blk_aio_writev(s->blk, blk_sector, &iov, nb_sectors, blk_sync_complete,
NULL);
}
static inline void submit_requests(BlockBackend *blk, MultiReqBuffer *mrb,
int start, int num_reqs, int niov)
{
QEMUIOVector *qiov = &mrb->reqs[start]->qiov;
int64_t sector_num = mrb->reqs[start]->sector_num;
int nb_sectors = mrb->reqs[start]->qiov.size / BDRV_SECTOR_SIZE;
bool is_write = mrb->is_write;
if (num_reqs > 1) {
int i;
struct iovec *tmp_iov = qiov->iov;
int tmp_niov = qiov->niov;
/* mrb->reqs[start]->qiov was initialized from external so we can't
* modifiy it here. We need to initialize it locally and then add the
* external iovecs. */
qemu_iovec_init(qiov, niov);
for (i = 0; i < tmp_niov; i++) {
qemu_iovec_add(qiov, tmp_iov[i].iov_base, tmp_iov[i].iov_len);
}
for (i = start + 1; i < start + num_reqs; i++) {
qemu_iovec_concat(qiov, &mrb->reqs[i]->qiov, 0,
mrb->reqs[i]->qiov.size);
mrb->reqs[i - 1]->mr_next = mrb->reqs[i];
nb_sectors += mrb->reqs[i]->qiov.size / BDRV_SECTOR_SIZE;
}
assert(nb_sectors == qiov->size / BDRV_SECTOR_SIZE);
trace_virtio_blk_submit_multireq(mrb, start, num_reqs, sector_num,
nb_sectors, is_write);
block_acct_merge_done(blk_get_stats(blk),
is_write ? BLOCK_ACCT_WRITE : BLOCK_ACCT_READ,
num_reqs - 1);
}
if (is_write) {
blk_aio_writev(blk, sector_num, qiov, nb_sectors,
virtio_blk_rw_complete, mrb->reqs[start]);
} else {
blk_aio_readv(blk, sector_num, qiov, nb_sectors,
virtio_blk_rw_complete, mrb->reqs[start]);
}
}
static inline void flash_sync_area(Flash *s, int64_t off, int64_t len)
{
int64_t start, end, nb_sectors;
QEMUIOVector iov;
if (!s->blk || blk_is_read_only(s->blk)) {
return;
}
assert(!(len % BDRV_SECTOR_SIZE));
start = off / BDRV_SECTOR_SIZE;
end = (off + len) / BDRV_SECTOR_SIZE;
nb_sectors = end - start;
qemu_iovec_init(&iov, 1);
qemu_iovec_add(&iov, s->storage + (start * BDRV_SECTOR_SIZE),
nb_sectors * BDRV_SECTOR_SIZE);
blk_aio_writev(s->blk, start, &iov, nb_sectors, blk_sync_complete, NULL);
}
static BlockDriverAIOCB *blkverify_aio_readv(BlockDriverState *bs,
int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
BlockDriverCompletionFunc *cb, void *opaque)
{
BDRVBlkverifyState *s = bs->opaque;
BlkverifyAIOCB *acb = blkverify_aio_get(bs, false, sector_num, qiov,
nb_sectors, cb, opaque);
acb->verify = blkverify_verify_readv;
acb->buf = qemu_blockalign(bs->file, qiov->size);
qemu_iovec_init(&acb->raw_qiov, acb->qiov->niov);
blkverify_iovec_clone(&acb->raw_qiov, qiov, acb->buf);
bdrv_aio_readv(s->test_file, sector_num, qiov, nb_sectors,
blkverify_aio_cb, acb);
bdrv_aio_readv(bs->file, sector_num, &acb->raw_qiov, nb_sectors,
blkverify_aio_cb, acb);
return &acb->common;
}
BlockDriverAIOCB *dma_bdrv_io(
BlockDriverState *bs, QEMUSGList *sg, uint64_t sector_num,
DMAIOFunc *io_func, BlockDriverCompletionFunc *cb,
void *opaque, bool to_dev)
{
DMAAIOCB *dbs = qemu_aio_get(&dma_aio_pool, bs, cb, opaque);
dbs->acb = NULL;
dbs->bs = bs;
dbs->sg = sg;
dbs->sector_num = sector_num;
dbs->sg_cur_index = 0;
dbs->sg_cur_byte = 0;
dbs->to_dev = to_dev;
dbs->io_func = io_func;
dbs->bh = NULL;
qemu_iovec_init(&dbs->iov, sg->nsg);
dma_bdrv_cb(dbs, 0);
return &dbs->common;
}
BlockAIOCB *dma_blk_io(
BlockBackend *blk, QEMUSGList *sg, uint64_t sector_num,
DMAIOFunc *io_func, BlockCompletionFunc *cb,
void *opaque, DMADirection dir)
{
DMAAIOCB *dbs = blk_aio_get(&dma_aiocb_info, blk, cb, opaque);
trace_dma_blk_io(dbs, blk, sector_num, (dir == DMA_DIRECTION_TO_DEVICE));
dbs->acb = NULL;
dbs->blk = blk;
dbs->sg = sg;
dbs->sector_num = sector_num;
dbs->sg_cur_index = 0;
dbs->sg_cur_byte = 0;
dbs->dir = dir;
dbs->io_func = io_func;
dbs->bh = NULL;
qemu_iovec_init(&dbs->iov, sg->nsg);
dma_blk_cb(dbs, 0);
return &dbs->common;
}
static BlockDriverAIOCB *dma_bdrv_io(
BlockDriverState *bs, QEMUSGList *sg, uint64_t sector_num,
BlockDriverCompletionFunc *cb, void *opaque,
int is_write)
{
DMAAIOCB *dbs = qemu_aio_get(&dma_aio_pool, bs, cb, opaque);
dbs->acb = NULL;
dbs->bs = bs;
dbs->sg = sg;
dbs->sector_num = sector_num;
dbs->sg_cur_index = 0;
dbs->sg_cur_byte = 0;
dbs->is_write = is_write;
dbs->bh = NULL;
qemu_iovec_init(&dbs->iov, sg->nsg);
dma_bdrv_cb(dbs, 0);
if (!dbs->acb) {
qemu_aio_release(dbs);
return NULL;
}
return &dbs->common;
}
static coroutine_fn int vhdx_co_readv(BlockDriverState *bs, int64_t sector_num,
int nb_sectors, QEMUIOVector *qiov)
{
BDRVVHDXState *s = bs->opaque;
int ret = 0;
VHDXSectorInfo sinfo;
uint64_t bytes_done = 0;
QEMUIOVector hd_qiov;
qemu_iovec_init(&hd_qiov, qiov->niov);
qemu_co_mutex_lock(&s->lock);
while (nb_sectors > 0) {
/* We are a differencing file, so we need to inspect the sector bitmap
* to see if we have the data or not */
if (s->params.data_bits & VHDX_PARAMS_HAS_PARENT) {
/* not supported yet */
ret = -ENOTSUP;
goto exit;
} else {
vhdx_block_translate(s, sector_num, nb_sectors, &sinfo);
qemu_iovec_reset(&hd_qiov);
qemu_iovec_concat(&hd_qiov, qiov, bytes_done, sinfo.bytes_avail);
/* check the payload block state */
switch (s->bat[sinfo.bat_idx] & VHDX_BAT_STATE_BIT_MASK) {
case PAYLOAD_BLOCK_NOT_PRESENT: /* fall through */
case PAYLOAD_BLOCK_UNDEFINED: /* fall through */
case PAYLOAD_BLOCK_UNMAPPED: /* fall through */
case PAYLOAD_BLOCK_ZERO:
/* return zero */
qemu_iovec_memset(&hd_qiov, 0, 0, sinfo.bytes_avail);
break;
case PAYLOAD_BLOCK_FULL_PRESENT:
qemu_co_mutex_unlock(&s->lock);
ret = bdrv_co_readv(bs->file,
sinfo.file_offset >> BDRV_SECTOR_BITS,
sinfo.sectors_avail, &hd_qiov);
qemu_co_mutex_lock(&s->lock);
if (ret < 0) {
goto exit;
}
break;
case PAYLOAD_BLOCK_PARTIALLY_PRESENT:
/* we don't yet support difference files, fall through
* to error */
default:
ret = -EIO;
goto exit;
break;
}
nb_sectors -= sinfo.sectors_avail;
sector_num += sinfo.sectors_avail;
bytes_done += sinfo.bytes_avail;
}
}
ret = 0;
exit:
qemu_co_mutex_unlock(&s->lock);
qemu_iovec_destroy(&hd_qiov);
return ret;
}
void usb_packet_init(USBPacket *p)
{
qemu_iovec_init(&p->iov, 1);
}
/* Submit async read while handling COW.
* Returns: The number of bytes copied after and including offset,
* excluding any bytes copied prior to offset due to alignment.
* This will be @bytes if no alignment is necessary, or
* (new_end - offset) if tail is rounded up or down due to
* alignment or buffer limit.
*/
static uint64_t mirror_do_read(MirrorBlockJob *s, int64_t offset,
uint64_t bytes)
{
BlockBackend *source = s->common.blk;
int nb_chunks;
uint64_t ret;
MirrorOp *op;
uint64_t max_bytes;
max_bytes = s->granularity * s->max_iov;
/* We can only handle as much as buf_size at a time. */
bytes = MIN(s->buf_size, MIN(max_bytes, bytes));
assert(bytes);
assert(bytes < BDRV_REQUEST_MAX_BYTES);
ret = bytes;
if (s->cow_bitmap) {
ret += mirror_cow_align(s, &offset, &bytes);
}
assert(bytes <= s->buf_size);
/* The offset is granularity-aligned because:
* 1) Caller passes in aligned values;
* 2) mirror_cow_align is used only when target cluster is larger. */
assert(QEMU_IS_ALIGNED(offset, s->granularity));
/* The range is sector-aligned, since bdrv_getlength() rounds up. */
assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
nb_chunks = DIV_ROUND_UP(bytes, s->granularity);
while (s->buf_free_count < nb_chunks) {
trace_mirror_yield_in_flight(s, offset, s->in_flight);
mirror_wait_for_io(s);
}
/* Allocate a MirrorOp that is used as an AIO callback. */
op = g_new(MirrorOp, 1);
op->s = s;
op->offset = offset;
op->bytes = bytes;
/* Now make a QEMUIOVector taking enough granularity-sized chunks
* from s->buf_free.
*/
qemu_iovec_init(&op->qiov, nb_chunks);
while (nb_chunks-- > 0) {
MirrorBuffer *buf = QSIMPLEQ_FIRST(&s->buf_free);
size_t remaining = bytes - op->qiov.size;
QSIMPLEQ_REMOVE_HEAD(&s->buf_free, next);
s->buf_free_count--;
qemu_iovec_add(&op->qiov, buf, MIN(s->granularity, remaining));
}
/* Copy the dirty cluster. */
s->in_flight++;
s->bytes_in_flight += bytes;
trace_mirror_one_iteration(s, offset, bytes);
blk_aio_preadv(source, offset, &op->qiov, 0, mirror_read_complete, op);
return ret;
}
static coroutine_fn int
block_crypto_co_writev(BlockDriverState *bs, int64_t sector_num,
int remaining_sectors, QEMUIOVector *qiov)
{
BlockCrypto *crypto = bs->opaque;
int cur_nr_sectors; /* number of sectors in current iteration */
uint64_t bytes_done = 0;
uint8_t *cipher_data = NULL;
QEMUIOVector hd_qiov;
int ret = 0;
size_t payload_offset =
qcrypto_block_get_payload_offset(crypto->block) / 512;
qemu_iovec_init(&hd_qiov, qiov->niov);
/* Bounce buffer so we have a linear mem region for
* entire sector. XXX optimize so we avoid bounce
* buffer in case that qiov->niov == 1
*/
cipher_data =
qemu_try_blockalign(bs->file->bs, MIN(BLOCK_CRYPTO_MAX_SECTORS * 512,
qiov->size));
if (cipher_data == NULL) {
ret = -ENOMEM;
goto cleanup;
}
while (remaining_sectors) {
cur_nr_sectors = remaining_sectors;
if (cur_nr_sectors > BLOCK_CRYPTO_MAX_SECTORS) {
cur_nr_sectors = BLOCK_CRYPTO_MAX_SECTORS;
}
qemu_iovec_to_buf(qiov, bytes_done,
cipher_data, cur_nr_sectors * 512);
if (qcrypto_block_encrypt(crypto->block,
sector_num,
cipher_data, cur_nr_sectors * 512,
NULL) < 0) {
ret = -EIO;
goto cleanup;
}
qemu_iovec_reset(&hd_qiov);
qemu_iovec_add(&hd_qiov, cipher_data, cur_nr_sectors * 512);
ret = bdrv_co_writev(bs->file,
payload_offset + sector_num,
cur_nr_sectors, &hd_qiov);
if (ret < 0) {
goto cleanup;
}
remaining_sectors -= cur_nr_sectors;
sector_num += cur_nr_sectors;
bytes_done += cur_nr_sectors * 512;
}
cleanup:
qemu_iovec_destroy(&hd_qiov);
qemu_vfree(cipher_data);
return ret;
}
static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
{
BlockDriverState *source = s->common.bs;
int nb_sectors, sectors_per_chunk, nb_chunks;
int64_t end, sector_num, next_chunk, next_sector, hbitmap_next_sector;
uint64_t delay_ns = 0;
MirrorOp *op;
int pnum;
int64_t ret;
s->sector_num = hbitmap_iter_next(&s->hbi);
if (s->sector_num < 0) {
bdrv_dirty_iter_init(s->dirty_bitmap, &s->hbi);
s->sector_num = hbitmap_iter_next(&s->hbi);
trace_mirror_restart_iter(s, bdrv_get_dirty_count(s->dirty_bitmap));
assert(s->sector_num >= 0);
}
hbitmap_next_sector = s->sector_num;
sector_num = s->sector_num;
sectors_per_chunk = s->granularity >> BDRV_SECTOR_BITS;
end = s->bdev_length / BDRV_SECTOR_SIZE;
/* Extend the QEMUIOVector to include all adjacent blocks that will
* be copied in this operation.
*
* We have to do this if we have no backing file yet in the destination,
* and the cluster size is very large. Then we need to do COW ourselves.
* The first time a cluster is copied, copy it entirely. Note that,
* because both the granularity and the cluster size are powers of two,
* the number of sectors to copy cannot exceed one cluster.
*
* We also want to extend the QEMUIOVector to include more adjacent
* dirty blocks if possible, to limit the number of I/O operations and
* run efficiently even with a small granularity.
*/
nb_chunks = 0;
nb_sectors = 0;
next_sector = sector_num;
next_chunk = sector_num / sectors_per_chunk;
/* Wait for I/O to this cluster (from a previous iteration) to be done. */
while (test_bit(next_chunk, s->in_flight_bitmap)) {
trace_mirror_yield_in_flight(s, sector_num, s->in_flight);
s->waiting_for_io = true;
qemu_coroutine_yield();
s->waiting_for_io = false;
}
do {
int added_sectors, added_chunks;
if (!bdrv_get_dirty(source, s->dirty_bitmap, next_sector) ||
test_bit(next_chunk, s->in_flight_bitmap)) {
assert(nb_sectors > 0);
break;
}
added_sectors = sectors_per_chunk;
if (s->cow_bitmap && !test_bit(next_chunk, s->cow_bitmap)) {
bdrv_round_to_clusters(s->target,
next_sector, added_sectors,
&next_sector, &added_sectors);
/* On the first iteration, the rounding may make us copy
* sectors before the first dirty one.
*/
if (next_sector < sector_num) {
assert(nb_sectors == 0);
sector_num = next_sector;
next_chunk = next_sector / sectors_per_chunk;
}
}
added_sectors = MIN(added_sectors, end - (sector_num + nb_sectors));
added_chunks = (added_sectors + sectors_per_chunk - 1) / sectors_per_chunk;
/* When doing COW, it may happen that there is not enough space for
* a full cluster. Wait if that is the case.
*/
while (nb_chunks == 0 && s->buf_free_count < added_chunks) {
trace_mirror_yield_buf_busy(s, nb_chunks, s->in_flight);
s->waiting_for_io = true;
qemu_coroutine_yield();
s->waiting_for_io = false;
}
if (s->buf_free_count < nb_chunks + added_chunks) {
trace_mirror_break_buf_busy(s, nb_chunks, s->in_flight);
break;
}
if (IOV_MAX < nb_chunks + added_chunks) {
trace_mirror_break_iov_max(s, nb_chunks, added_chunks);
break;
}
/* We have enough free space to copy these sectors. */
bitmap_set(s->in_flight_bitmap, next_chunk, added_chunks);
nb_sectors += added_sectors;
nb_chunks += added_chunks;
next_sector += added_sectors;
next_chunk += added_chunks;
if (!s->synced && s->common.speed) {
delay_ns = ratelimit_calculate_delay(&s->limit, added_sectors);
}
} while (delay_ns == 0 && next_sector < end);
/* Allocate a MirrorOp that is used as an AIO callback. */
op = g_new(MirrorOp, 1);
op->s = s;
op->sector_num = sector_num;
op->nb_sectors = nb_sectors;
/* Now make a QEMUIOVector taking enough granularity-sized chunks
* from s->buf_free.
*/
qemu_iovec_init(&op->qiov, nb_chunks);
next_sector = sector_num;
while (nb_chunks-- > 0) {
MirrorBuffer *buf = QSIMPLEQ_FIRST(&s->buf_free);
size_t remaining = (nb_sectors * BDRV_SECTOR_SIZE) - op->qiov.size;
QSIMPLEQ_REMOVE_HEAD(&s->buf_free, next);
s->buf_free_count--;
qemu_iovec_add(&op->qiov, buf, MIN(s->granularity, remaining));
/* Advance the HBitmapIter in parallel, so that we do not examine
* the same sector twice.
*/
if (next_sector > hbitmap_next_sector
&& bdrv_get_dirty(source, s->dirty_bitmap, next_sector)) {
hbitmap_next_sector = hbitmap_iter_next(&s->hbi);
}
next_sector += sectors_per_chunk;
}
bdrv_reset_dirty_bitmap(s->dirty_bitmap, sector_num, nb_sectors);
/* Copy the dirty cluster. */
s->in_flight++;
s->sectors_in_flight += nb_sectors;
trace_mirror_one_iteration(s, sector_num, nb_sectors);
ret = bdrv_get_block_status_above(source, NULL, sector_num,
nb_sectors, &pnum);
if (ret < 0 || pnum < nb_sectors ||
(ret & BDRV_BLOCK_DATA && !(ret & BDRV_BLOCK_ZERO))) {
bdrv_aio_readv(source, sector_num, &op->qiov, nb_sectors,
mirror_read_complete, op);
} else if (ret & BDRV_BLOCK_ZERO) {
bdrv_aio_write_zeroes(s->target, sector_num, op->nb_sectors,
s->unmap ? BDRV_REQ_MAY_UNMAP : 0,
mirror_write_complete, op);
} else {
assert(!(ret & BDRV_BLOCK_DATA));
bdrv_aio_discard(s->target, sector_num, op->nb_sectors,
mirror_write_complete, op);
}
return delay_ns;
}
static uint16_t nvme_map_prp(QEMUSGList *qsg, QEMUIOVector *iov, uint64_t prp1,
uint64_t prp2, uint32_t len, NvmeCtrl *n)
{
hwaddr trans_len = n->page_size - (prp1 % n->page_size);
trans_len = MIN(len, trans_len);
int num_prps = (len >> n->page_bits) + 1;
if (unlikely(!prp1)) {
trace_nvme_err_invalid_prp();
return NVME_INVALID_FIELD | NVME_DNR;
} else if (n->cmbsz && prp1 >= n->ctrl_mem.addr &&
prp1 < n->ctrl_mem.addr + int128_get64(n->ctrl_mem.size)) {
qsg->nsg = 0;
qemu_iovec_init(iov, num_prps);
qemu_iovec_add(iov, (void *)&n->cmbuf[prp1 - n->ctrl_mem.addr], trans_len);
} else {
pci_dma_sglist_init(qsg, &n->parent_obj, num_prps);
qemu_sglist_add(qsg, prp1, trans_len);
}
len -= trans_len;
if (len) {
if (unlikely(!prp2)) {
trace_nvme_err_invalid_prp2_missing();
goto unmap;
}
if (len > n->page_size) {
uint64_t prp_list[n->max_prp_ents];
uint32_t nents, prp_trans;
int i = 0;
nents = (len + n->page_size - 1) >> n->page_bits;
prp_trans = MIN(n->max_prp_ents, nents) * sizeof(uint64_t);
nvme_addr_read(n, prp2, (void *)prp_list, prp_trans);
while (len != 0) {
uint64_t prp_ent = le64_to_cpu(prp_list[i]);
if (i == n->max_prp_ents - 1 && len > n->page_size) {
if (unlikely(!prp_ent || prp_ent & (n->page_size - 1))) {
trace_nvme_err_invalid_prplist_ent(prp_ent);
goto unmap;
}
i = 0;
nents = (len + n->page_size - 1) >> n->page_bits;
prp_trans = MIN(n->max_prp_ents, nents) * sizeof(uint64_t);
nvme_addr_read(n, prp_ent, (void *)prp_list,
prp_trans);
prp_ent = le64_to_cpu(prp_list[i]);
}
if (unlikely(!prp_ent || prp_ent & (n->page_size - 1))) {
trace_nvme_err_invalid_prplist_ent(prp_ent);
goto unmap;
}
trans_len = MIN(len, n->page_size);
if (qsg->nsg){
qemu_sglist_add(qsg, prp_ent, trans_len);
} else {
qemu_iovec_add(iov, (void *)&n->cmbuf[prp_ent - n->ctrl_mem.addr], trans_len);
}
len -= trans_len;
i++;
}
} else {
if (unlikely(prp2 & (n->page_size - 1))) {
static void coroutine_fn mirror_iteration(MirrorBlockJob *s)
{
BlockDriverState *source = s->common.bs;
int nb_sectors, sectors_per_chunk, nb_chunks;
int64_t end, sector_num, next_chunk, next_sector, hbitmap_next_sector;
MirrorOp *op;
s->sector_num = hbitmap_iter_next(&s->hbi);
if (s->sector_num < 0) {
bdrv_dirty_iter_init(source, &s->hbi);
s->sector_num = hbitmap_iter_next(&s->hbi);
trace_mirror_restart_iter(s, bdrv_get_dirty_count(source));
assert(s->sector_num >= 0);
}
hbitmap_next_sector = s->sector_num;
sector_num = s->sector_num;
sectors_per_chunk = s->granularity >> BDRV_SECTOR_BITS;
end = s->common.len >> BDRV_SECTOR_BITS;
/* Extend the QEMUIOVector to include all adjacent blocks that will
* be copied in this operation.
*
* We have to do this if we have no backing file yet in the destination,
* and the cluster size is very large. Then we need to do COW ourselves.
* The first time a cluster is copied, copy it entirely. Note that,
* because both the granularity and the cluster size are powers of two,
* the number of sectors to copy cannot exceed one cluster.
*
* We also want to extend the QEMUIOVector to include more adjacent
* dirty blocks if possible, to limit the number of I/O operations and
* run efficiently even with a small granularity.
*/
nb_chunks = 0;
nb_sectors = 0;
next_sector = sector_num;
next_chunk = sector_num / sectors_per_chunk;
/* Wait for I/O to this cluster (from a previous iteration) to be done. */
while (test_bit(next_chunk, s->in_flight_bitmap)) {
trace_mirror_yield_in_flight(s, sector_num, s->in_flight);
qemu_coroutine_yield();
}
do {
int added_sectors, added_chunks;
if (!bdrv_get_dirty(source, next_sector) ||
test_bit(next_chunk, s->in_flight_bitmap)) {
assert(nb_sectors > 0);
break;
}
added_sectors = sectors_per_chunk;
if (s->cow_bitmap && !test_bit(next_chunk, s->cow_bitmap)) {
bdrv_round_to_clusters(s->target,
next_sector, added_sectors,
&next_sector, &added_sectors);
/* On the first iteration, the rounding may make us copy
* sectors before the first dirty one.
*/
if (next_sector < sector_num) {
assert(nb_sectors == 0);
sector_num = next_sector;
next_chunk = next_sector / sectors_per_chunk;
}
}
added_sectors = MIN(added_sectors, end - (sector_num + nb_sectors));
added_chunks = (added_sectors + sectors_per_chunk - 1) / sectors_per_chunk;
/* When doing COW, it may happen that there is not enough space for
* a full cluster. Wait if that is the case.
*/
while (nb_chunks == 0 && s->buf_free_count < added_chunks) {
trace_mirror_yield_buf_busy(s, nb_chunks, s->in_flight);
qemu_coroutine_yield();
}
if (s->buf_free_count < nb_chunks + added_chunks) {
trace_mirror_break_buf_busy(s, nb_chunks, s->in_flight);
break;
}
/* We have enough free space to copy these sectors. */
bitmap_set(s->in_flight_bitmap, next_chunk, added_chunks);
nb_sectors += added_sectors;
nb_chunks += added_chunks;
next_sector += added_sectors;
next_chunk += added_chunks;
} while (next_sector < end);
/* Allocate a MirrorOp that is used as an AIO callback. */
op = g_slice_new(MirrorOp);
op->s = s;
op->sector_num = sector_num;
op->nb_sectors = nb_sectors;
/* Now make a QEMUIOVector taking enough granularity-sized chunks
* from s->buf_free.
*/
qemu_iovec_init(&op->qiov, nb_chunks);
next_sector = sector_num;
while (nb_chunks-- > 0) {
MirrorBuffer *buf = QSIMPLEQ_FIRST(&s->buf_free);
QSIMPLEQ_REMOVE_HEAD(&s->buf_free, next);
s->buf_free_count--;
qemu_iovec_add(&op->qiov, buf, s->granularity);
/* Advance the HBitmapIter in parallel, so that we do not examine
* the same sector twice.
*/
if (next_sector > hbitmap_next_sector && bdrv_get_dirty(source, next_sector)) {
hbitmap_next_sector = hbitmap_iter_next(&s->hbi);
}
next_sector += sectors_per_chunk;
}
bdrv_reset_dirty(source, sector_num, nb_sectors);
/* Copy the dirty cluster. */
s->in_flight++;
trace_mirror_one_iteration(s, sector_num, nb_sectors);
bdrv_aio_readv(source, sector_num, &op->qiov, nb_sectors,
mirror_read_complete, op);
}
