4 * Copyright Red Hat, Inc. 2012
7 * Paolo Bonzini <pbonzini@redhat.com>
9 * This work is licensed under the terms of the GNU LGPL, version 2 or later.
10 * See the COPYING.LIB file in the top-level directory.
14 #include "qemu/osdep.h"
15 #include "qemu/cutils.h"
16 #include "qemu/coroutine.h"
17 #include "qemu/range.h"
19 #include "block/blockjob_int.h"
20 #include "block/block_int.h"
21 #include "block/dirty-bitmap.h"
22 #include "sysemu/block-backend.h"
23 #include "qapi/error.h"
24 #include "qemu/ratelimit.h"
25 #include "qemu/bitmap.h"
26 #include "qemu/memalign.h"
28 #define MAX_IN_FLIGHT 16
29 #define MAX_IO_BYTES (1 << 20) /* 1 Mb */
30 #define DEFAULT_MIRROR_BUF_SIZE (MAX_IN_FLIGHT * MAX_IO_BYTES)
32 /* The mirroring buffer is a list of granularity-sized chunks.
33 * Free chunks are organized in a list.
35 typedef struct MirrorBuffer
{
36 QSIMPLEQ_ENTRY(MirrorBuffer
) next
;
39 typedef struct MirrorOp MirrorOp
;
41 typedef struct MirrorBlockJob
{
44 BlockDriverState
*mirror_top_bs
;
45 BlockDriverState
*base
;
46 BlockDriverState
*base_overlay
;
48 /* The name of the graph node to replace */
50 /* The BDS to replace */
51 BlockDriverState
*to_replace
;
52 /* Used to block operations on the drive-mirror-replace target */
53 Error
*replace_blocker
;
55 BlockMirrorBackingMode backing_mode
;
56 /* Whether the target image requires explicit zero-initialization */
59 * To be accesssed with atomics. Written only under the BQL (required by the
60 * current implementation of mirror_change()).
62 MirrorCopyMode copy_mode
;
63 BlockdevOnError on_source_error
, on_target_error
;
65 * To be accessed with atomics.
67 * Set when the target is synced (dirty bitmap is clean, nothing in flight)
68 * and the job is running in active mode.
75 unsigned long *cow_bitmap
;
76 BdrvDirtyBitmap
*dirty_bitmap
;
77 BdrvDirtyBitmapIter
*dbi
;
79 QSIMPLEQ_HEAD(, MirrorBuffer
) buf_free
;
82 uint64_t last_pause_ns
;
83 unsigned long *in_flight_bitmap
;
85 int64_t bytes_in_flight
;
86 QTAILQ_HEAD(, MirrorOp
) ops_in_flight
;
89 int target_cluster_size
;
91 bool initial_zeroing_ongoing
;
92 int in_active_write_counter
;
93 int64_t active_write_bytes_in_flight
;
98 typedef struct MirrorBDSOpaque
{
110 /* The pointee is set by mirror_co_read(), mirror_co_zero(), and
111 * mirror_co_discard() before yielding for the first time */
112 int64_t *bytes_handled
;
115 bool is_active_write
;
117 CoQueue waiting_requests
;
119 MirrorOp
*waiting_for_op
;
121 QTAILQ_ENTRY(MirrorOp
) next
;
124 typedef enum MirrorMethod
{
127 MIRROR_METHOD_DISCARD
,
130 static BlockErrorAction
mirror_error_action(MirrorBlockJob
*s
, bool read
,
133 qatomic_set(&s
->actively_synced
, false);
135 return block_job_error_action(&s
->common
, s
->on_source_error
,
138 return block_job_error_action(&s
->common
, s
->on_target_error
,
143 static void coroutine_fn
mirror_wait_on_conflicts(MirrorOp
*self
,
148 uint64_t self_start_chunk
= offset
/ s
->granularity
;
149 uint64_t self_end_chunk
= DIV_ROUND_UP(offset
+ bytes
, s
->granularity
);
150 uint64_t self_nb_chunks
= self_end_chunk
- self_start_chunk
;
152 while (find_next_bit(s
->in_flight_bitmap
, self_end_chunk
,
153 self_start_chunk
) < self_end_chunk
&&
158 QTAILQ_FOREACH(op
, &s
->ops_in_flight
, next
) {
159 uint64_t op_start_chunk
= op
->offset
/ s
->granularity
;
160 uint64_t op_nb_chunks
= DIV_ROUND_UP(op
->offset
+ op
->bytes
,
168 if (ranges_overlap(self_start_chunk
, self_nb_chunks
,
169 op_start_chunk
, op_nb_chunks
))
173 * If the operation is already (indirectly) waiting for us,
174 * or will wait for us as soon as it wakes up, then just go
175 * on (instead of producing a deadlock in the former case).
177 if (op
->waiting_for_op
) {
181 self
->waiting_for_op
= op
;
184 qemu_co_queue_wait(&op
->waiting_requests
, NULL
);
187 self
->waiting_for_op
= NULL
;
196 static void coroutine_fn
mirror_iteration_done(MirrorOp
*op
, int ret
)
198 MirrorBlockJob
*s
= op
->s
;
203 trace_mirror_iteration_done(s
, op
->offset
, op
->bytes
, ret
);
206 s
->bytes_in_flight
-= op
->bytes
;
208 for (i
= 0; i
< op
->qiov
.niov
; i
++) {
209 MirrorBuffer
*buf
= (MirrorBuffer
*) iov
[i
].iov_base
;
210 QSIMPLEQ_INSERT_TAIL(&s
->buf_free
, buf
, next
);
214 chunk_num
= op
->offset
/ s
->granularity
;
215 nb_chunks
= DIV_ROUND_UP(op
->bytes
, s
->granularity
);
217 bitmap_clear(s
->in_flight_bitmap
, chunk_num
, nb_chunks
);
218 QTAILQ_REMOVE(&s
->ops_in_flight
, op
, next
);
221 bitmap_set(s
->cow_bitmap
, chunk_num
, nb_chunks
);
223 if (!s
->initial_zeroing_ongoing
) {
224 job_progress_update(&s
->common
.job
, op
->bytes
);
227 qemu_iovec_destroy(&op
->qiov
);
229 qemu_co_queue_restart_all(&op
->waiting_requests
);
233 static void coroutine_fn
mirror_write_complete(MirrorOp
*op
, int ret
)
235 MirrorBlockJob
*s
= op
->s
;
238 BlockErrorAction action
;
240 bdrv_set_dirty_bitmap(s
->dirty_bitmap
, op
->offset
, op
->bytes
);
241 action
= mirror_error_action(s
, false, -ret
);
242 if (action
== BLOCK_ERROR_ACTION_REPORT
&& s
->ret
>= 0) {
247 mirror_iteration_done(op
, ret
);
250 static void coroutine_fn
mirror_read_complete(MirrorOp
*op
, int ret
)
252 MirrorBlockJob
*s
= op
->s
;
255 BlockErrorAction action
;
257 bdrv_set_dirty_bitmap(s
->dirty_bitmap
, op
->offset
, op
->bytes
);
258 action
= mirror_error_action(s
, true, -ret
);
259 if (action
== BLOCK_ERROR_ACTION_REPORT
&& s
->ret
>= 0) {
263 mirror_iteration_done(op
, ret
);
267 ret
= blk_co_pwritev(s
->target
, op
->offset
, op
->qiov
.size
, &op
->qiov
, 0);
268 mirror_write_complete(op
, ret
);
271 /* Clip bytes relative to offset to not exceed end-of-file */
272 static inline int64_t mirror_clip_bytes(MirrorBlockJob
*s
,
276 return MIN(bytes
, s
->bdev_length
- offset
);
279 /* Round offset and/or bytes to target cluster if COW is needed, and
280 * return the offset of the adjusted tail against original. */
281 static int coroutine_fn
mirror_cow_align(MirrorBlockJob
*s
, int64_t *offset
,
286 int64_t align_offset
= *offset
;
287 int64_t align_bytes
= *bytes
;
288 int max_bytes
= s
->granularity
* s
->max_iov
;
290 need_cow
= !test_bit(*offset
/ s
->granularity
, s
->cow_bitmap
);
291 need_cow
|= !test_bit((*offset
+ *bytes
- 1) / s
->granularity
,
294 bdrv_round_to_subclusters(blk_bs(s
->target
), *offset
, *bytes
,
295 &align_offset
, &align_bytes
);
298 if (align_bytes
> max_bytes
) {
299 align_bytes
= max_bytes
;
301 align_bytes
= QEMU_ALIGN_DOWN(align_bytes
, s
->target_cluster_size
);
304 /* Clipping may result in align_bytes unaligned to chunk boundary, but
305 * that doesn't matter because it's already the end of source image. */
306 align_bytes
= mirror_clip_bytes(s
, align_offset
, align_bytes
);
308 ret
= align_offset
+ align_bytes
- (*offset
+ *bytes
);
309 *offset
= align_offset
;
310 *bytes
= align_bytes
;
315 static inline void coroutine_fn
316 mirror_wait_for_free_in_flight_slot(MirrorBlockJob
*s
)
320 QTAILQ_FOREACH(op
, &s
->ops_in_flight
, next
) {
322 * Do not wait on pseudo ops, because it may in turn wait on
323 * some other operation to start, which may in fact be the
324 * caller of this function. Since there is only one pseudo op
325 * at any given time, we will always find some real operation
327 * Also, do not wait on active operations, because they do not
328 * use up in-flight slots.
330 if (!op
->is_pseudo_op
&& op
->is_in_flight
&& !op
->is_active_write
) {
331 qemu_co_queue_wait(&op
->waiting_requests
, NULL
);
338 /* Perform a mirror copy operation.
340 * *op->bytes_handled is set to the number of bytes copied after and
341 * including offset, excluding any bytes copied prior to offset due
342 * to alignment. This will be op->bytes if no alignment is necessary,
343 * or (new_end - op->offset) if the tail is rounded up or down due to
344 * alignment or buffer limit.
346 static void coroutine_fn
mirror_co_read(void *opaque
)
348 MirrorOp
*op
= opaque
;
349 MirrorBlockJob
*s
= op
->s
;
354 max_bytes
= s
->granularity
* s
->max_iov
;
356 /* We can only handle as much as buf_size at a time. */
357 op
->bytes
= MIN(s
->buf_size
, MIN(max_bytes
, op
->bytes
));
359 assert(op
->bytes
< BDRV_REQUEST_MAX_BYTES
);
360 *op
->bytes_handled
= op
->bytes
;
363 *op
->bytes_handled
+= mirror_cow_align(s
, &op
->offset
, &op
->bytes
);
365 /* Cannot exceed BDRV_REQUEST_MAX_BYTES + INT_MAX */
366 assert(*op
->bytes_handled
<= UINT_MAX
);
367 assert(op
->bytes
<= s
->buf_size
);
368 /* The offset is granularity-aligned because:
369 * 1) Caller passes in aligned values;
370 * 2) mirror_cow_align is used only when target cluster is larger. */
371 assert(QEMU_IS_ALIGNED(op
->offset
, s
->granularity
));
372 /* The range is sector-aligned, since bdrv_getlength() rounds up. */
373 assert(QEMU_IS_ALIGNED(op
->bytes
, BDRV_SECTOR_SIZE
));
374 nb_chunks
= DIV_ROUND_UP(op
->bytes
, s
->granularity
);
376 while (s
->buf_free_count
< nb_chunks
) {
377 trace_mirror_yield_in_flight(s
, op
->offset
, s
->in_flight
);
378 mirror_wait_for_free_in_flight_slot(s
);
381 /* Now make a QEMUIOVector taking enough granularity-sized chunks
384 qemu_iovec_init(&op
->qiov
, nb_chunks
);
385 while (nb_chunks
-- > 0) {
386 MirrorBuffer
*buf
= QSIMPLEQ_FIRST(&s
->buf_free
);
387 size_t remaining
= op
->bytes
- op
->qiov
.size
;
389 QSIMPLEQ_REMOVE_HEAD(&s
->buf_free
, next
);
391 qemu_iovec_add(&op
->qiov
, buf
, MIN(s
->granularity
, remaining
));
394 /* Copy the dirty cluster. */
396 s
->bytes_in_flight
+= op
->bytes
;
397 op
->is_in_flight
= true;
398 trace_mirror_one_iteration(s
, op
->offset
, op
->bytes
);
400 WITH_GRAPH_RDLOCK_GUARD() {
401 ret
= bdrv_co_preadv(s
->mirror_top_bs
->backing
, op
->offset
, op
->bytes
,
404 mirror_read_complete(op
, ret
);
407 static void coroutine_fn
mirror_co_zero(void *opaque
)
409 MirrorOp
*op
= opaque
;
413 op
->s
->bytes_in_flight
+= op
->bytes
;
414 *op
->bytes_handled
= op
->bytes
;
415 op
->is_in_flight
= true;
417 ret
= blk_co_pwrite_zeroes(op
->s
->target
, op
->offset
, op
->bytes
,
418 op
->s
->unmap
? BDRV_REQ_MAY_UNMAP
: 0);
419 mirror_write_complete(op
, ret
);
422 static void coroutine_fn
mirror_co_discard(void *opaque
)
424 MirrorOp
*op
= opaque
;
428 op
->s
->bytes_in_flight
+= op
->bytes
;
429 *op
->bytes_handled
= op
->bytes
;
430 op
->is_in_flight
= true;
432 ret
= blk_co_pdiscard(op
->s
->target
, op
->offset
, op
->bytes
);
433 mirror_write_complete(op
, ret
);
436 static unsigned mirror_perform(MirrorBlockJob
*s
, int64_t offset
,
437 unsigned bytes
, MirrorMethod mirror_method
)
441 int64_t bytes_handled
= -1;
443 op
= g_new(MirrorOp
, 1);
448 .bytes_handled
= &bytes_handled
,
450 qemu_co_queue_init(&op
->waiting_requests
);
452 switch (mirror_method
) {
453 case MIRROR_METHOD_COPY
:
454 co
= qemu_coroutine_create(mirror_co_read
, op
);
456 case MIRROR_METHOD_ZERO
:
457 co
= qemu_coroutine_create(mirror_co_zero
, op
);
459 case MIRROR_METHOD_DISCARD
:
460 co
= qemu_coroutine_create(mirror_co_discard
, op
);
467 QTAILQ_INSERT_TAIL(&s
->ops_in_flight
, op
, next
);
468 qemu_coroutine_enter(co
);
469 /* At this point, ownership of op has been moved to the coroutine
470 * and the object may already be freed */
472 /* Assert that this value has been set */
473 assert(bytes_handled
>= 0);
475 /* Same assertion as in mirror_co_read() (and for mirror_co_read()
476 * and mirror_co_discard(), bytes_handled == op->bytes, which
477 * is the @bytes parameter given to this function) */
478 assert(bytes_handled
<= UINT_MAX
);
479 return bytes_handled
;
482 static void coroutine_fn GRAPH_RDLOCK
mirror_iteration(MirrorBlockJob
*s
)
484 BlockDriverState
*source
= s
->mirror_top_bs
->backing
->bs
;
487 /* At least the first dirty chunk is mirrored in one iteration. */
489 bool write_zeroes_ok
= bdrv_can_write_zeroes_with_unmap(blk_bs(s
->target
));
490 int max_io_bytes
= MAX(s
->buf_size
/ MAX_IN_FLIGHT
, MAX_IO_BYTES
);
492 bdrv_dirty_bitmap_lock(s
->dirty_bitmap
);
493 offset
= bdrv_dirty_iter_next(s
->dbi
);
495 bdrv_set_dirty_iter(s
->dbi
, 0);
496 offset
= bdrv_dirty_iter_next(s
->dbi
);
497 trace_mirror_restart_iter(s
, bdrv_get_dirty_count(s
->dirty_bitmap
));
500 bdrv_dirty_bitmap_unlock(s
->dirty_bitmap
);
503 * Wait for concurrent requests to @offset. The next loop will limit the
504 * copied area based on in_flight_bitmap so we only copy an area that does
505 * not overlap with concurrent in-flight requests. Still, we would like to
506 * copy something, so wait until there are at least no more requests to the
507 * very beginning of the area.
509 mirror_wait_on_conflicts(NULL
, s
, offset
, 1);
511 job_pause_point(&s
->common
.job
);
513 /* Find the number of consecutive dirty chunks following the first dirty
514 * one, and wait for in flight requests in them. */
515 bdrv_dirty_bitmap_lock(s
->dirty_bitmap
);
516 while (nb_chunks
* s
->granularity
< s
->buf_size
) {
518 int64_t next_offset
= offset
+ nb_chunks
* s
->granularity
;
519 int64_t next_chunk
= next_offset
/ s
->granularity
;
520 if (next_offset
>= s
->bdev_length
||
521 !bdrv_dirty_bitmap_get_locked(s
->dirty_bitmap
, next_offset
)) {
524 if (test_bit(next_chunk
, s
->in_flight_bitmap
)) {
528 next_dirty
= bdrv_dirty_iter_next(s
->dbi
);
529 if (next_dirty
> next_offset
|| next_dirty
< 0) {
530 /* The bitmap iterator's cache is stale, refresh it */
531 bdrv_set_dirty_iter(s
->dbi
, next_offset
);
532 next_dirty
= bdrv_dirty_iter_next(s
->dbi
);
534 assert(next_dirty
== next_offset
);
538 /* Clear dirty bits before querying the block status, because
539 * calling bdrv_block_status_above could yield - if some blocks are
540 * marked dirty in this window, we need to know.
542 bdrv_reset_dirty_bitmap_locked(s
->dirty_bitmap
, offset
,
543 nb_chunks
* s
->granularity
);
544 bdrv_dirty_bitmap_unlock(s
->dirty_bitmap
);
546 /* Before claiming an area in the in-flight bitmap, we have to
547 * create a MirrorOp for it so that conflicting requests can wait
548 * for it. mirror_perform() will create the real MirrorOps later,
549 * for now we just create a pseudo operation that will wake up all
550 * conflicting requests once all real operations have been
552 pseudo_op
= g_new(MirrorOp
, 1);
553 *pseudo_op
= (MirrorOp
){
555 .bytes
= nb_chunks
* s
->granularity
,
556 .is_pseudo_op
= true,
558 qemu_co_queue_init(&pseudo_op
->waiting_requests
);
559 QTAILQ_INSERT_TAIL(&s
->ops_in_flight
, pseudo_op
, next
);
561 bitmap_set(s
->in_flight_bitmap
, offset
/ s
->granularity
, nb_chunks
);
562 while (nb_chunks
> 0 && offset
< s
->bdev_length
) {
565 int64_t io_bytes_acct
;
566 MirrorMethod mirror_method
= MIRROR_METHOD_COPY
;
568 assert(!(offset
% s
->granularity
));
569 WITH_GRAPH_RDLOCK_GUARD() {
570 ret
= bdrv_co_block_status_above(source
, NULL
, offset
,
571 nb_chunks
* s
->granularity
,
572 &io_bytes
, NULL
, NULL
);
575 io_bytes
= MIN(nb_chunks
* s
->granularity
, max_io_bytes
);
576 } else if (ret
& BDRV_BLOCK_DATA
) {
577 io_bytes
= MIN(io_bytes
, max_io_bytes
);
580 io_bytes
-= io_bytes
% s
->granularity
;
581 if (io_bytes
< s
->granularity
) {
582 io_bytes
= s
->granularity
;
583 } else if (ret
>= 0 && !(ret
& BDRV_BLOCK_DATA
)) {
584 int64_t target_offset
;
585 int64_t target_bytes
;
586 WITH_GRAPH_RDLOCK_GUARD() {
587 bdrv_round_to_subclusters(blk_bs(s
->target
), offset
, io_bytes
,
588 &target_offset
, &target_bytes
);
590 if (target_offset
== offset
&&
591 target_bytes
== io_bytes
) {
592 mirror_method
= ret
& BDRV_BLOCK_ZERO
?
594 MIRROR_METHOD_DISCARD
;
598 while (s
->in_flight
>= MAX_IN_FLIGHT
) {
599 trace_mirror_yield_in_flight(s
, offset
, s
->in_flight
);
600 mirror_wait_for_free_in_flight_slot(s
);
608 io_bytes
= mirror_clip_bytes(s
, offset
, io_bytes
);
609 io_bytes
= mirror_perform(s
, offset
, io_bytes
, mirror_method
);
610 if (mirror_method
!= MIRROR_METHOD_COPY
&& write_zeroes_ok
) {
613 io_bytes_acct
= io_bytes
;
617 nb_chunks
-= DIV_ROUND_UP(io_bytes
, s
->granularity
);
618 block_job_ratelimit_processed_bytes(&s
->common
, io_bytes_acct
);
622 QTAILQ_REMOVE(&s
->ops_in_flight
, pseudo_op
, next
);
623 qemu_co_queue_restart_all(&pseudo_op
->waiting_requests
);
627 static void mirror_free_init(MirrorBlockJob
*s
)
629 int granularity
= s
->granularity
;
630 size_t buf_size
= s
->buf_size
;
631 uint8_t *buf
= s
->buf
;
633 assert(s
->buf_free_count
== 0);
634 QSIMPLEQ_INIT(&s
->buf_free
);
635 while (buf_size
!= 0) {
636 MirrorBuffer
*cur
= (MirrorBuffer
*)buf
;
637 QSIMPLEQ_INSERT_TAIL(&s
->buf_free
, cur
, next
);
639 buf_size
-= granularity
;
644 /* This is also used for the .pause callback. There is no matching
645 * mirror_resume() because mirror_run() will begin iterating again
646 * when the job is resumed.
648 static void coroutine_fn
mirror_wait_for_all_io(MirrorBlockJob
*s
)
650 while (s
->in_flight
> 0) {
651 mirror_wait_for_free_in_flight_slot(s
);
656 * mirror_exit_common: handle both abort() and prepare() cases.
657 * for .prepare, returns 0 on success and -errno on failure.
658 * for .abort cases, denoted by abort = true, MUST return 0.
660 static int mirror_exit_common(Job
*job
)
662 MirrorBlockJob
*s
= container_of(job
, MirrorBlockJob
, common
.job
);
663 BlockJob
*bjob
= &s
->common
;
664 MirrorBDSOpaque
*bs_opaque
;
665 AioContext
*replace_aio_context
= NULL
;
666 BlockDriverState
*src
;
667 BlockDriverState
*target_bs
;
668 BlockDriverState
*mirror_top_bs
;
669 Error
*local_err
= NULL
;
670 bool abort
= job
->ret
< 0;
680 aio_context_acquire(qemu_get_aio_context());
681 bdrv_graph_rdlock_main_loop();
683 mirror_top_bs
= s
->mirror_top_bs
;
684 bs_opaque
= mirror_top_bs
->opaque
;
685 src
= mirror_top_bs
->backing
->bs
;
686 target_bs
= blk_bs(s
->target
);
688 if (bdrv_chain_contains(src
, target_bs
)) {
689 bdrv_unfreeze_backing_chain(mirror_top_bs
, target_bs
);
692 bdrv_release_dirty_bitmap(s
->dirty_bitmap
);
694 /* Make sure that the source BDS doesn't go away during bdrv_replace_node,
695 * before we can call bdrv_drained_end */
697 bdrv_ref(mirror_top_bs
);
700 bdrv_graph_rdunlock_main_loop();
703 * Remove target parent that still uses BLK_PERM_WRITE/RESIZE before
704 * inserting target_bs at s->to_replace, where we might not be able to get
707 blk_unref(s
->target
);
710 /* We don't access the source any more. Dropping any WRITE/RESIZE is
711 * required before it could become a backing file of target_bs. Not having
712 * these permissions any more means that we can't allow any new requests on
713 * mirror_top_bs from now on, so keep it drained. */
714 bdrv_drained_begin(mirror_top_bs
);
715 bdrv_drained_begin(target_bs
);
716 bs_opaque
->stop
= true;
718 bdrv_graph_rdlock_main_loop();
719 bdrv_child_refresh_perms(mirror_top_bs
, mirror_top_bs
->backing
,
722 if (!abort
&& s
->backing_mode
== MIRROR_SOURCE_BACKING_CHAIN
) {
723 BlockDriverState
*backing
= s
->is_none_mode
? src
: s
->base
;
724 BlockDriverState
*unfiltered_target
= bdrv_skip_filters(target_bs
);
726 if (bdrv_cow_bs(unfiltered_target
) != backing
) {
727 bdrv_set_backing_hd(unfiltered_target
, backing
, &local_err
);
729 error_report_err(local_err
);
734 } else if (!abort
&& s
->backing_mode
== MIRROR_OPEN_BACKING_CHAIN
) {
735 assert(!bdrv_backing_chain_next(target_bs
));
736 ret
= bdrv_open_backing_file(bdrv_skip_filters(target_bs
), NULL
,
737 "backing", &local_err
);
739 error_report_err(local_err
);
743 bdrv_graph_rdunlock_main_loop();
746 replace_aio_context
= bdrv_get_aio_context(s
->to_replace
);
747 aio_context_acquire(replace_aio_context
);
750 if (s
->should_complete
&& !abort
) {
751 BlockDriverState
*to_replace
= s
->to_replace
?: src
;
752 bool ro
= bdrv_is_read_only(to_replace
);
754 if (ro
!= bdrv_is_read_only(target_bs
)) {
755 bdrv_reopen_set_read_only(target_bs
, ro
, NULL
);
758 /* The mirror job has no requests in flight any more, but we need to
759 * drain potential other users of the BDS before changing the graph. */
761 bdrv_drained_begin(to_replace
);
763 * Cannot use check_to_replace_node() here, because that would
764 * check for an op blocker on @to_replace, and we have our own
767 bdrv_graph_wrlock(target_bs
);
768 if (bdrv_recurse_can_replace(src
, to_replace
)) {
769 bdrv_replace_node(to_replace
, target_bs
, &local_err
);
771 error_setg(&local_err
, "Can no longer replace '%s' by '%s', "
772 "because it can no longer be guaranteed that doing so "
773 "would not lead to an abrupt change of visible data",
774 to_replace
->node_name
, target_bs
->node_name
);
776 bdrv_graph_wrunlock();
777 bdrv_drained_end(to_replace
);
779 error_report_err(local_err
);
784 bdrv_op_unblock_all(s
->to_replace
, s
->replace_blocker
);
785 error_free(s
->replace_blocker
);
786 bdrv_unref(s
->to_replace
);
788 if (replace_aio_context
) {
789 aio_context_release(replace_aio_context
);
794 * Remove the mirror filter driver from the graph. Before this, get rid of
795 * the blockers on the intermediate nodes so that the resulting state is
798 block_job_remove_all_bdrv(bjob
);
799 bdrv_graph_wrlock(mirror_top_bs
);
800 bdrv_replace_node(mirror_top_bs
, mirror_top_bs
->backing
->bs
, &error_abort
);
801 bdrv_graph_wrunlock();
803 bdrv_drained_end(target_bs
);
804 bdrv_unref(target_bs
);
806 bs_opaque
->job
= NULL
;
808 bdrv_drained_end(src
);
809 bdrv_drained_end(mirror_top_bs
);
811 bdrv_unref(mirror_top_bs
);
814 aio_context_release(qemu_get_aio_context());
819 static int mirror_prepare(Job
*job
)
821 return mirror_exit_common(job
);
824 static void mirror_abort(Job
*job
)
826 int ret
= mirror_exit_common(job
);
830 static void coroutine_fn
mirror_throttle(MirrorBlockJob
*s
)
832 int64_t now
= qemu_clock_get_ns(QEMU_CLOCK_REALTIME
);
834 if (now
- s
->last_pause_ns
> BLOCK_JOB_SLICE_TIME
) {
835 s
->last_pause_ns
= now
;
836 job_sleep_ns(&s
->common
.job
, 0);
838 job_pause_point(&s
->common
.job
);
842 static int coroutine_fn GRAPH_UNLOCKED
mirror_dirty_init(MirrorBlockJob
*s
)
845 BlockDriverState
*bs
;
846 BlockDriverState
*target_bs
= blk_bs(s
->target
);
850 bdrv_graph_co_rdlock();
851 bs
= s
->mirror_top_bs
->backing
->bs
;
852 bdrv_graph_co_rdunlock();
854 if (s
->zero_target
) {
855 if (!bdrv_can_write_zeroes_with_unmap(target_bs
)) {
856 bdrv_set_dirty_bitmap(s
->dirty_bitmap
, 0, s
->bdev_length
);
860 s
->initial_zeroing_ongoing
= true;
861 for (offset
= 0; offset
< s
->bdev_length
; ) {
862 int bytes
= MIN(s
->bdev_length
- offset
,
863 QEMU_ALIGN_DOWN(INT_MAX
, s
->granularity
));
867 if (job_is_cancelled(&s
->common
.job
)) {
868 s
->initial_zeroing_ongoing
= false;
872 if (s
->in_flight
>= MAX_IN_FLIGHT
) {
873 trace_mirror_yield(s
, UINT64_MAX
, s
->buf_free_count
,
875 mirror_wait_for_free_in_flight_slot(s
);
879 mirror_perform(s
, offset
, bytes
, MIRROR_METHOD_ZERO
);
883 mirror_wait_for_all_io(s
);
884 s
->initial_zeroing_ongoing
= false;
887 /* First part, loop on the sectors and initialize the dirty bitmap. */
888 for (offset
= 0; offset
< s
->bdev_length
; ) {
889 /* Just to make sure we are not exceeding int limit. */
890 int bytes
= MIN(s
->bdev_length
- offset
,
891 QEMU_ALIGN_DOWN(INT_MAX
, s
->granularity
));
895 if (job_is_cancelled(&s
->common
.job
)) {
899 WITH_GRAPH_RDLOCK_GUARD() {
900 ret
= bdrv_co_is_allocated_above(bs
, s
->base_overlay
, true, offset
,
909 bdrv_set_dirty_bitmap(s
->dirty_bitmap
, offset
, count
);
916 /* Called when going out of the streaming phase to flush the bulk of the
917 * data to the medium, or just before completing.
919 static int coroutine_fn
mirror_flush(MirrorBlockJob
*s
)
921 int ret
= blk_co_flush(s
->target
);
923 if (mirror_error_action(s
, false, -ret
) == BLOCK_ERROR_ACTION_REPORT
) {
930 static int coroutine_fn
mirror_run(Job
*job
, Error
**errp
)
932 MirrorBlockJob
*s
= container_of(job
, MirrorBlockJob
, common
.job
);
933 BlockDriverState
*bs
;
934 MirrorBDSOpaque
*mirror_top_opaque
= s
->mirror_top_bs
->opaque
;
935 BlockDriverState
*target_bs
= blk_bs(s
->target
);
936 bool need_drain
= true;
937 BlockDeviceIoStatus iostatus
;
939 int64_t target_length
;
941 char backing_filename
[2]; /* we only need 2 characters because we are only
942 checking for a NULL string */
945 bdrv_graph_co_rdlock();
946 bs
= bdrv_filter_bs(s
->mirror_top_bs
);
947 bdrv_graph_co_rdunlock();
949 if (job_is_cancelled(&s
->common
.job
)) {
953 bdrv_graph_co_rdlock();
954 s
->bdev_length
= bdrv_co_getlength(bs
);
955 bdrv_graph_co_rdunlock();
957 if (s
->bdev_length
< 0) {
958 ret
= s
->bdev_length
;
962 target_length
= blk_co_getlength(s
->target
);
963 if (target_length
< 0) {
968 /* Active commit must resize the base image if its size differs from the
970 if (s
->base
== blk_bs(s
->target
)) {
971 if (s
->bdev_length
> target_length
) {
972 ret
= blk_co_truncate(s
->target
, s
->bdev_length
, false,
973 PREALLOC_MODE_OFF
, 0, NULL
);
978 } else if (s
->bdev_length
!= target_length
) {
979 error_setg(errp
, "Source and target image have different sizes");
984 if (s
->bdev_length
== 0) {
985 /* Transition to the READY state and wait for complete. */
986 job_transition_to_ready(&s
->common
.job
);
987 qatomic_set(&s
->actively_synced
, true);
988 while (!job_cancel_requested(&s
->common
.job
) && !s
->should_complete
) {
989 job_yield(&s
->common
.job
);
994 length
= DIV_ROUND_UP(s
->bdev_length
, s
->granularity
);
995 s
->in_flight_bitmap
= bitmap_new(length
);
997 /* If we have no backing file yet in the destination, we cannot let
998 * the destination do COW. Instead, we copy sectors around the
999 * dirty data if needed. We need a bitmap to do that.
1001 bdrv_get_backing_filename(target_bs
, backing_filename
,
1002 sizeof(backing_filename
));
1003 bdrv_graph_co_rdlock();
1004 if (!bdrv_co_get_info(target_bs
, &bdi
) && bdi
.cluster_size
) {
1005 s
->target_cluster_size
= bdi
.cluster_size
;
1007 s
->target_cluster_size
= BDRV_SECTOR_SIZE
;
1009 if (backing_filename
[0] && !bdrv_backing_chain_next(target_bs
) &&
1010 s
->granularity
< s
->target_cluster_size
) {
1011 s
->buf_size
= MAX(s
->buf_size
, s
->target_cluster_size
);
1012 s
->cow_bitmap
= bitmap_new(length
);
1014 s
->max_iov
= MIN(bs
->bl
.max_iov
, target_bs
->bl
.max_iov
);
1015 bdrv_graph_co_rdunlock();
1017 s
->buf
= qemu_try_blockalign(bs
, s
->buf_size
);
1018 if (s
->buf
== NULL
) {
1020 goto immediate_exit
;
1023 mirror_free_init(s
);
1025 s
->last_pause_ns
= qemu_clock_get_ns(QEMU_CLOCK_REALTIME
);
1026 if (!s
->is_none_mode
) {
1027 ret
= mirror_dirty_init(s
);
1028 if (ret
< 0 || job_is_cancelled(&s
->common
.job
)) {
1029 goto immediate_exit
;
1034 * Only now the job is fully initialised and mirror_top_bs should start
1037 mirror_top_opaque
->job
= s
;
1040 s
->dbi
= bdrv_dirty_iter_new(s
->dirty_bitmap
);
1043 bool should_complete
;
1047 goto immediate_exit
;
1050 job_pause_point(&s
->common
.job
);
1052 if (job_is_cancelled(&s
->common
.job
)) {
1054 goto immediate_exit
;
1057 cnt
= bdrv_get_dirty_count(s
->dirty_bitmap
);
1058 /* cnt is the number of dirty bytes remaining and s->bytes_in_flight is
1059 * the number of bytes currently being processed; together those are
1060 * the current remaining operation length */
1061 job_progress_set_remaining(&s
->common
.job
,
1062 s
->bytes_in_flight
+ cnt
+
1063 s
->active_write_bytes_in_flight
);
1065 /* Note that even when no rate limit is applied we need to yield
1066 * periodically with no pending I/O so that bdrv_drain_all() returns.
1067 * We do so every BLKOCK_JOB_SLICE_TIME nanoseconds, or when there is
1068 * an error, or when the source is clean, whichever comes first. */
1069 delta
= qemu_clock_get_ns(QEMU_CLOCK_REALTIME
) - s
->last_pause_ns
;
1070 WITH_JOB_LOCK_GUARD() {
1071 iostatus
= s
->common
.iostatus
;
1073 if (delta
< BLOCK_JOB_SLICE_TIME
&&
1074 iostatus
== BLOCK_DEVICE_IO_STATUS_OK
) {
1075 if (s
->in_flight
>= MAX_IN_FLIGHT
|| s
->buf_free_count
== 0 ||
1076 (cnt
== 0 && s
->in_flight
> 0)) {
1077 trace_mirror_yield(s
, cnt
, s
->buf_free_count
, s
->in_flight
);
1078 mirror_wait_for_free_in_flight_slot(s
);
1080 } else if (cnt
!= 0) {
1081 bdrv_graph_co_rdlock();
1082 mirror_iteration(s
);
1083 bdrv_graph_co_rdunlock();
1087 should_complete
= false;
1088 if (s
->in_flight
== 0 && cnt
== 0) {
1089 trace_mirror_before_flush(s
);
1090 if (!job_is_ready(&s
->common
.job
)) {
1091 if (mirror_flush(s
) < 0) {
1092 /* Go check s->ret. */
1095 /* We're out of the streaming phase. From now on, if the job
1096 * is cancelled we will actually complete all pending I/O and
1097 * report completion. This way, block-job-cancel will leave
1098 * the target in a consistent state.
1100 job_transition_to_ready(&s
->common
.job
);
1102 if (qatomic_read(&s
->copy_mode
) != MIRROR_COPY_MODE_BACKGROUND
) {
1103 qatomic_set(&s
->actively_synced
, true);
1106 should_complete
= s
->should_complete
||
1107 job_cancel_requested(&s
->common
.job
);
1108 cnt
= bdrv_get_dirty_count(s
->dirty_bitmap
);
1111 if (cnt
== 0 && should_complete
) {
1112 /* The dirty bitmap is not updated while operations are pending.
1113 * If we're about to exit, wait for pending operations before
1114 * calling bdrv_get_dirty_count(bs), or we may exit while the
1115 * source has dirty data to copy!
1117 * Note that I/O can be submitted by the guest while
1118 * mirror_populate runs, so pause it now. Before deciding
1119 * whether to switch to target check one last time if I/O has
1120 * come in the meanwhile, and if not flush the data to disk.
1122 trace_mirror_before_drain(s
, cnt
);
1125 bdrv_drained_begin(bs
);
1127 /* Must be zero because we are drained */
1128 assert(s
->in_active_write_counter
== 0);
1130 cnt
= bdrv_get_dirty_count(s
->dirty_bitmap
);
1131 if (cnt
> 0 || mirror_flush(s
) < 0) {
1132 bdrv_drained_end(bs
);
1133 s
->in_drain
= false;
1137 /* The two disks are in sync. Exit and report successful
1140 assert(QLIST_EMPTY(&bs
->tracked_requests
));
1145 if (job_is_ready(&s
->common
.job
) && !should_complete
) {
1146 if (s
->in_flight
== 0 && cnt
== 0) {
1147 trace_mirror_before_sleep(s
, cnt
, job_is_ready(&s
->common
.job
),
1148 BLOCK_JOB_SLICE_TIME
);
1149 job_sleep_ns(&s
->common
.job
, BLOCK_JOB_SLICE_TIME
);
1152 block_job_ratelimit_sleep(&s
->common
);
1154 s
->last_pause_ns
= qemu_clock_get_ns(QEMU_CLOCK_REALTIME
);
1158 if (s
->in_flight
> 0) {
1159 /* We get here only if something went wrong. Either the job failed,
1160 * or it was cancelled prematurely so that we do not guarantee that
1161 * the target is a copy of the source.
1163 assert(ret
< 0 || job_is_cancelled(&s
->common
.job
));
1165 mirror_wait_for_all_io(s
);
1168 assert(s
->in_flight
== 0);
1170 g_free(s
->cow_bitmap
);
1171 g_free(s
->in_flight_bitmap
);
1172 bdrv_dirty_iter_free(s
->dbi
);
1176 bdrv_drained_begin(bs
);
1182 static void mirror_complete(Job
*job
, Error
**errp
)
1184 MirrorBlockJob
*s
= container_of(job
, MirrorBlockJob
, common
.job
);
1186 if (!job_is_ready(job
)) {
1187 error_setg(errp
, "The active block job '%s' cannot be completed",
1192 /* block all operations on to_replace bs */
1194 AioContext
*replace_aio_context
;
1196 s
->to_replace
= bdrv_find_node(s
->replaces
);
1197 if (!s
->to_replace
) {
1198 error_setg(errp
, "Node name '%s' not found", s
->replaces
);
1202 replace_aio_context
= bdrv_get_aio_context(s
->to_replace
);
1203 aio_context_acquire(replace_aio_context
);
1205 /* TODO Translate this into child freeze system. */
1206 error_setg(&s
->replace_blocker
,
1207 "block device is in use by block-job-complete");
1208 bdrv_op_block_all(s
->to_replace
, s
->replace_blocker
);
1209 bdrv_ref(s
->to_replace
);
1211 aio_context_release(replace_aio_context
);
1214 s
->should_complete
= true;
1216 /* If the job is paused, it will be re-entered when it is resumed */
1217 WITH_JOB_LOCK_GUARD() {
1219 job_enter_cond_locked(job
, NULL
);
1224 static void coroutine_fn
mirror_pause(Job
*job
)
1226 MirrorBlockJob
*s
= container_of(job
, MirrorBlockJob
, common
.job
);
1228 mirror_wait_for_all_io(s
);
1231 static bool mirror_drained_poll(BlockJob
*job
)
1233 MirrorBlockJob
*s
= container_of(job
, MirrorBlockJob
, common
);
1235 /* If the job isn't paused nor cancelled, we can't be sure that it won't
1236 * issue more requests. We make an exception if we've reached this point
1237 * from one of our own drain sections, to avoid a deadlock waiting for
1240 WITH_JOB_LOCK_GUARD() {
1241 if (!s
->common
.job
.paused
&& !job_is_cancelled_locked(&job
->job
)
1247 return !!s
->in_flight
;
1250 static bool mirror_cancel(Job
*job
, bool force
)
1252 MirrorBlockJob
*s
= container_of(job
, MirrorBlockJob
, common
.job
);
1253 BlockDriverState
*target
= blk_bs(s
->target
);
1256 * Before the job is READY, we treat any cancellation like a
1257 * force-cancellation.
1259 force
= force
|| !job_is_ready(job
);
1262 bdrv_cancel_in_flight(target
);
1267 static bool commit_active_cancel(Job
*job
, bool force
)
1269 /* Same as above in mirror_cancel() */
1270 return force
|| !job_is_ready(job
);
1273 static void mirror_change(BlockJob
*job
, BlockJobChangeOptions
*opts
,
1276 MirrorBlockJob
*s
= container_of(job
, MirrorBlockJob
, common
);
1277 BlockJobChangeOptionsMirror
*change_opts
= &opts
->u
.mirror
;
1278 MirrorCopyMode current
;
1281 * The implementation relies on the fact that copy_mode is only written
1282 * under the BQL. Otherwise, further synchronization would be required.
1285 GLOBAL_STATE_CODE();
1287 if (qatomic_read(&s
->copy_mode
) == change_opts
->copy_mode
) {
1291 if (change_opts
->copy_mode
!= MIRROR_COPY_MODE_WRITE_BLOCKING
) {
1292 error_setg(errp
, "Change to copy mode '%s' is not implemented",
1293 MirrorCopyMode_str(change_opts
->copy_mode
));
1297 current
= qatomic_cmpxchg(&s
->copy_mode
, MIRROR_COPY_MODE_BACKGROUND
,
1298 change_opts
->copy_mode
);
1299 if (current
!= MIRROR_COPY_MODE_BACKGROUND
) {
1300 error_setg(errp
, "Expected current copy mode '%s', got '%s'",
1301 MirrorCopyMode_str(MIRROR_COPY_MODE_BACKGROUND
),
1302 MirrorCopyMode_str(current
));
1306 static void mirror_query(BlockJob
*job
, BlockJobInfo
*info
)
1308 MirrorBlockJob
*s
= container_of(job
, MirrorBlockJob
, common
);
1310 info
->u
.mirror
= (BlockJobInfoMirror
) {
1311 .actively_synced
= qatomic_read(&s
->actively_synced
),
1315 static const BlockJobDriver mirror_job_driver
= {
1317 .instance_size
= sizeof(MirrorBlockJob
),
1318 .job_type
= JOB_TYPE_MIRROR
,
1319 .free
= block_job_free
,
1320 .user_resume
= block_job_user_resume
,
1322 .prepare
= mirror_prepare
,
1323 .abort
= mirror_abort
,
1324 .pause
= mirror_pause
,
1325 .complete
= mirror_complete
,
1326 .cancel
= mirror_cancel
,
1328 .drained_poll
= mirror_drained_poll
,
1329 .change
= mirror_change
,
1330 .query
= mirror_query
,
1333 static const BlockJobDriver commit_active_job_driver
= {
1335 .instance_size
= sizeof(MirrorBlockJob
),
1336 .job_type
= JOB_TYPE_COMMIT
,
1337 .free
= block_job_free
,
1338 .user_resume
= block_job_user_resume
,
1340 .prepare
= mirror_prepare
,
1341 .abort
= mirror_abort
,
1342 .pause
= mirror_pause
,
1343 .complete
= mirror_complete
,
1344 .cancel
= commit_active_cancel
,
1346 .drained_poll
= mirror_drained_poll
,
1349 static void coroutine_fn
1350 do_sync_target_write(MirrorBlockJob
*job
, MirrorMethod method
,
1351 uint64_t offset
, uint64_t bytes
,
1352 QEMUIOVector
*qiov
, int flags
)
1355 size_t qiov_offset
= 0;
1356 int64_t bitmap_offset
, bitmap_end
;
1358 if (!QEMU_IS_ALIGNED(offset
, job
->granularity
) &&
1359 bdrv_dirty_bitmap_get(job
->dirty_bitmap
, offset
))
1362 * Dirty unaligned padding: ignore it.
1365 * 1. If we copy it, we can't reset corresponding bit in
1366 * dirty_bitmap as there may be some "dirty" bytes still not
1368 * 2. It's already dirty, so skipping it we don't diverge mirror
1371 * Note, that because of this, guest write may have no contribution
1372 * into mirror converge, but that's not bad, as we have background
1373 * process of mirroring. If under some bad circumstances (high guest
1374 * IO load) background process starve, we will not converge anyway,
1375 * even if each write will contribute, as guest is not guaranteed to
1376 * rewrite the whole disk.
1378 qiov_offset
= QEMU_ALIGN_UP(offset
, job
->granularity
) - offset
;
1379 if (bytes
<= qiov_offset
) {
1380 /* nothing to do after shrink */
1383 offset
+= qiov_offset
;
1384 bytes
-= qiov_offset
;
1387 if (!QEMU_IS_ALIGNED(offset
+ bytes
, job
->granularity
) &&
1388 bdrv_dirty_bitmap_get(job
->dirty_bitmap
, offset
+ bytes
- 1))
1390 uint64_t tail
= (offset
+ bytes
) % job
->granularity
;
1392 if (bytes
<= tail
) {
1393 /* nothing to do after shrink */
1400 * Tails are either clean or shrunk, so for bitmap resetting
1401 * we safely align the range down.
1403 bitmap_offset
= QEMU_ALIGN_UP(offset
, job
->granularity
);
1404 bitmap_end
= QEMU_ALIGN_DOWN(offset
+ bytes
, job
->granularity
);
1405 if (bitmap_offset
< bitmap_end
) {
1406 bdrv_reset_dirty_bitmap(job
->dirty_bitmap
, bitmap_offset
,
1407 bitmap_end
- bitmap_offset
);
1410 job_progress_increase_remaining(&job
->common
.job
, bytes
);
1411 job
->active_write_bytes_in_flight
+= bytes
;
1414 case MIRROR_METHOD_COPY
:
1415 ret
= blk_co_pwritev_part(job
->target
, offset
, bytes
,
1416 qiov
, qiov_offset
, flags
);
1419 case MIRROR_METHOD_ZERO
:
1421 ret
= blk_co_pwrite_zeroes(job
->target
, offset
, bytes
, flags
);
1424 case MIRROR_METHOD_DISCARD
:
1426 ret
= blk_co_pdiscard(job
->target
, offset
, bytes
);
1433 job
->active_write_bytes_in_flight
-= bytes
;
1435 job_progress_update(&job
->common
.job
, bytes
);
1437 BlockErrorAction action
;
1440 * We failed, so we should mark dirty the whole area, aligned up.
1441 * Note that we don't care about shrunk tails if any: they were dirty
1442 * at function start, and they must be still dirty, as we've locked
1443 * the region for in-flight op.
1445 bitmap_offset
= QEMU_ALIGN_DOWN(offset
, job
->granularity
);
1446 bitmap_end
= QEMU_ALIGN_UP(offset
+ bytes
, job
->granularity
);
1447 bdrv_set_dirty_bitmap(job
->dirty_bitmap
, bitmap_offset
,
1448 bitmap_end
- bitmap_offset
);
1449 qatomic_set(&job
->actively_synced
, false);
1451 action
= mirror_error_action(job
, false, -ret
);
1452 if (action
== BLOCK_ERROR_ACTION_REPORT
) {
1460 static MirrorOp
*coroutine_fn
active_write_prepare(MirrorBlockJob
*s
,
1465 uint64_t start_chunk
= offset
/ s
->granularity
;
1466 uint64_t end_chunk
= DIV_ROUND_UP(offset
+ bytes
, s
->granularity
);
1468 op
= g_new(MirrorOp
, 1);
1473 .is_active_write
= true,
1474 .is_in_flight
= true,
1475 .co
= qemu_coroutine_self(),
1477 qemu_co_queue_init(&op
->waiting_requests
);
1478 QTAILQ_INSERT_TAIL(&s
->ops_in_flight
, op
, next
);
1480 s
->in_active_write_counter
++;
1483 * Wait for concurrent requests affecting the area. If there are already
1484 * running requests that are copying off now-to-be stale data in the area,
1485 * we must wait for them to finish before we begin writing fresh data to the
1486 * target so that the write operations appear in the correct order.
1487 * Note that background requests (see mirror_iteration()) in contrast only
1488 * wait for conflicting requests at the start of the dirty area, and then
1489 * (based on the in_flight_bitmap) truncate the area to copy so it will not
1490 * conflict with any requests beyond that. For active writes, however, we
1491 * cannot truncate that area. The request from our parent must be blocked
1492 * until the area is copied in full. Therefore, we must wait for the whole
1493 * area to become free of concurrent requests.
1495 mirror_wait_on_conflicts(op
, s
, offset
, bytes
);
1497 bitmap_set(s
->in_flight_bitmap
, start_chunk
, end_chunk
- start_chunk
);
1502 static void coroutine_fn GRAPH_RDLOCK
active_write_settle(MirrorOp
*op
)
1504 uint64_t start_chunk
= op
->offset
/ op
->s
->granularity
;
1505 uint64_t end_chunk
= DIV_ROUND_UP(op
->offset
+ op
->bytes
,
1506 op
->s
->granularity
);
1508 if (!--op
->s
->in_active_write_counter
&&
1509 qatomic_read(&op
->s
->actively_synced
)) {
1510 BdrvChild
*source
= op
->s
->mirror_top_bs
->backing
;
1512 if (QLIST_FIRST(&source
->bs
->parents
) == source
&&
1513 QLIST_NEXT(source
, next_parent
) == NULL
)
1515 /* Assert that we are back in sync once all active write
1516 * operations are settled.
1517 * Note that we can only assert this if the mirror node
1518 * is the source node's only parent. */
1519 assert(!bdrv_get_dirty_count(op
->s
->dirty_bitmap
));
1522 bitmap_clear(op
->s
->in_flight_bitmap
, start_chunk
, end_chunk
- start_chunk
);
1523 QTAILQ_REMOVE(&op
->s
->ops_in_flight
, op
, next
);
1524 qemu_co_queue_restart_all(&op
->waiting_requests
);
1528 static int coroutine_fn GRAPH_RDLOCK
1529 bdrv_mirror_top_preadv(BlockDriverState
*bs
, int64_t offset
, int64_t bytes
,
1530 QEMUIOVector
*qiov
, BdrvRequestFlags flags
)
1532 return bdrv_co_preadv(bs
->backing
, offset
, bytes
, qiov
, flags
);
1535 static bool should_copy_to_target(MirrorBDSOpaque
*s
)
1537 return s
->job
&& s
->job
->ret
>= 0 &&
1538 !job_is_cancelled(&s
->job
->common
.job
) &&
1539 qatomic_read(&s
->job
->copy_mode
) == MIRROR_COPY_MODE_WRITE_BLOCKING
;
1542 static int coroutine_fn GRAPH_RDLOCK
1543 bdrv_mirror_top_do_write(BlockDriverState
*bs
, MirrorMethod method
,
1544 bool copy_to_target
, uint64_t offset
, uint64_t bytes
,
1545 QEMUIOVector
*qiov
, int flags
)
1547 MirrorOp
*op
= NULL
;
1548 MirrorBDSOpaque
*s
= bs
->opaque
;
1551 if (copy_to_target
) {
1552 op
= active_write_prepare(s
->job
, offset
, bytes
);
1556 case MIRROR_METHOD_COPY
:
1557 ret
= bdrv_co_pwritev(bs
->backing
, offset
, bytes
, qiov
, flags
);
1560 case MIRROR_METHOD_ZERO
:
1561 ret
= bdrv_co_pwrite_zeroes(bs
->backing
, offset
, bytes
, flags
);
1564 case MIRROR_METHOD_DISCARD
:
1565 ret
= bdrv_co_pdiscard(bs
->backing
, offset
, bytes
);
1572 if (!copy_to_target
&& s
->job
&& s
->job
->dirty_bitmap
) {
1573 qatomic_set(&s
->job
->actively_synced
, false);
1574 bdrv_set_dirty_bitmap(s
->job
->dirty_bitmap
, offset
, bytes
);
1581 if (copy_to_target
) {
1582 do_sync_target_write(s
->job
, method
, offset
, bytes
, qiov
, flags
);
1586 if (copy_to_target
) {
1587 active_write_settle(op
);
1592 static int coroutine_fn GRAPH_RDLOCK
1593 bdrv_mirror_top_pwritev(BlockDriverState
*bs
, int64_t offset
, int64_t bytes
,
1594 QEMUIOVector
*qiov
, BdrvRequestFlags flags
)
1596 QEMUIOVector bounce_qiov
;
1599 bool copy_to_target
= should_copy_to_target(bs
->opaque
);
1601 if (copy_to_target
) {
1602 /* The guest might concurrently modify the data to write; but
1603 * the data on source and destination must match, so we have
1604 * to use a bounce buffer if we are going to write to the
1606 bounce_buf
= qemu_blockalign(bs
, bytes
);
1607 iov_to_buf_full(qiov
->iov
, qiov
->niov
, 0, bounce_buf
, bytes
);
1609 qemu_iovec_init(&bounce_qiov
, 1);
1610 qemu_iovec_add(&bounce_qiov
, bounce_buf
, bytes
);
1611 qiov
= &bounce_qiov
;
1613 flags
&= ~BDRV_REQ_REGISTERED_BUF
;
1616 ret
= bdrv_mirror_top_do_write(bs
, MIRROR_METHOD_COPY
, copy_to_target
,
1617 offset
, bytes
, qiov
, flags
);
1619 if (copy_to_target
) {
1620 qemu_iovec_destroy(&bounce_qiov
);
1621 qemu_vfree(bounce_buf
);
1627 static int coroutine_fn GRAPH_RDLOCK
bdrv_mirror_top_flush(BlockDriverState
*bs
)
1629 if (bs
->backing
== NULL
) {
1630 /* we can be here after failed bdrv_append in mirror_start_job */
1633 return bdrv_co_flush(bs
->backing
->bs
);
1636 static int coroutine_fn GRAPH_RDLOCK
1637 bdrv_mirror_top_pwrite_zeroes(BlockDriverState
*bs
, int64_t offset
,
1638 int64_t bytes
, BdrvRequestFlags flags
)
1640 bool copy_to_target
= should_copy_to_target(bs
->opaque
);
1641 return bdrv_mirror_top_do_write(bs
, MIRROR_METHOD_ZERO
, copy_to_target
,
1642 offset
, bytes
, NULL
, flags
);
1645 static int coroutine_fn GRAPH_RDLOCK
1646 bdrv_mirror_top_pdiscard(BlockDriverState
*bs
, int64_t offset
, int64_t bytes
)
1648 bool copy_to_target
= should_copy_to_target(bs
->opaque
);
1649 return bdrv_mirror_top_do_write(bs
, MIRROR_METHOD_DISCARD
, copy_to_target
,
1650 offset
, bytes
, NULL
, 0);
1653 static void GRAPH_RDLOCK
bdrv_mirror_top_refresh_filename(BlockDriverState
*bs
)
1655 if (bs
->backing
== NULL
) {
1656 /* we can be here after failed bdrv_attach_child in
1657 * bdrv_set_backing_hd */
1660 pstrcpy(bs
->exact_filename
, sizeof(bs
->exact_filename
),
1661 bs
->backing
->bs
->filename
);
1664 static void bdrv_mirror_top_child_perm(BlockDriverState
*bs
, BdrvChild
*c
,
1666 BlockReopenQueue
*reopen_queue
,
1667 uint64_t perm
, uint64_t shared
,
1668 uint64_t *nperm
, uint64_t *nshared
)
1670 MirrorBDSOpaque
*s
= bs
->opaque
;
1674 * If the job is to be stopped, we do not need to forward
1675 * anything to the real image.
1678 *nshared
= BLK_PERM_ALL
;
1682 bdrv_default_perms(bs
, c
, role
, reopen_queue
,
1683 perm
, shared
, nperm
, nshared
);
1687 * For commit jobs, we cannot take CONSISTENT_READ, because
1688 * that permission is unshared for everything above the base
1689 * node (except for filters on the base node).
1690 * We also have to force-share the WRITE permission, or
1691 * otherwise we would block ourselves at the base node (if
1692 * writes are blocked for a node, they are also blocked for
1693 * its backing file).
1694 * (We could also share RESIZE, because it may be needed for
1695 * the target if its size is less than the top node's; but
1696 * bdrv_default_perms_for_cow() automatically shares RESIZE
1697 * for backing nodes if WRITE is shared, so there is no need
1700 *nperm
&= ~BLK_PERM_CONSISTENT_READ
;
1701 *nshared
|= BLK_PERM_WRITE
;
1705 /* Dummy node that provides consistent read to its users without requiring it
1706 * from its backing file and that allows writes on the backing file chain. */
1707 static BlockDriver bdrv_mirror_top
= {
1708 .format_name
= "mirror_top",
1709 .bdrv_co_preadv
= bdrv_mirror_top_preadv
,
1710 .bdrv_co_pwritev
= bdrv_mirror_top_pwritev
,
1711 .bdrv_co_pwrite_zeroes
= bdrv_mirror_top_pwrite_zeroes
,
1712 .bdrv_co_pdiscard
= bdrv_mirror_top_pdiscard
,
1713 .bdrv_co_flush
= bdrv_mirror_top_flush
,
1714 .bdrv_refresh_filename
= bdrv_mirror_top_refresh_filename
,
1715 .bdrv_child_perm
= bdrv_mirror_top_child_perm
,
1718 .filtered_child_is_backing
= true,
1721 static BlockJob
*mirror_start_job(
1722 const char *job_id
, BlockDriverState
*bs
,
1723 int creation_flags
, BlockDriverState
*target
,
1724 const char *replaces
, int64_t speed
,
1725 uint32_t granularity
, int64_t buf_size
,
1726 BlockMirrorBackingMode backing_mode
,
1728 BlockdevOnError on_source_error
,
1729 BlockdevOnError on_target_error
,
1731 BlockCompletionFunc
*cb
,
1733 const BlockJobDriver
*driver
,
1734 bool is_none_mode
, BlockDriverState
*base
,
1735 bool auto_complete
, const char *filter_node_name
,
1736 bool is_mirror
, MirrorCopyMode copy_mode
,
1740 MirrorBDSOpaque
*bs_opaque
;
1741 BlockDriverState
*mirror_top_bs
;
1742 bool target_is_backing
;
1743 uint64_t target_perms
, target_shared_perms
;
1746 GLOBAL_STATE_CODE();
1748 if (granularity
== 0) {
1749 granularity
= bdrv_get_default_bitmap_granularity(target
);
1752 assert(is_power_of_2(granularity
));
1755 error_setg(errp
, "Invalid parameter 'buf-size'");
1759 if (buf_size
== 0) {
1760 buf_size
= DEFAULT_MIRROR_BUF_SIZE
;
1763 bdrv_graph_rdlock_main_loop();
1764 if (bdrv_skip_filters(bs
) == bdrv_skip_filters(target
)) {
1765 error_setg(errp
, "Can't mirror node into itself");
1766 bdrv_graph_rdunlock_main_loop();
1770 target_is_backing
= bdrv_chain_contains(bs
, target
);
1771 bdrv_graph_rdunlock_main_loop();
1773 /* In the case of active commit, add dummy driver to provide consistent
1774 * reads on the top, while disabling it in the intermediate nodes, and make
1775 * the backing chain writable. */
1776 mirror_top_bs
= bdrv_new_open_driver(&bdrv_mirror_top
, filter_node_name
,
1778 if (mirror_top_bs
== NULL
) {
1781 if (!filter_node_name
) {
1782 mirror_top_bs
->implicit
= true;
1785 /* So that we can always drop this node */
1786 mirror_top_bs
->never_freeze
= true;
1788 mirror_top_bs
->total_sectors
= bs
->total_sectors
;
1789 mirror_top_bs
->supported_write_flags
= BDRV_REQ_WRITE_UNCHANGED
;
1790 mirror_top_bs
->supported_zero_flags
= BDRV_REQ_WRITE_UNCHANGED
|
1791 BDRV_REQ_NO_FALLBACK
;
1792 bs_opaque
= g_new0(MirrorBDSOpaque
, 1);
1793 mirror_top_bs
->opaque
= bs_opaque
;
1795 bs_opaque
->is_commit
= target_is_backing
;
1797 bdrv_drained_begin(bs
);
1798 ret
= bdrv_append(mirror_top_bs
, bs
, errp
);
1799 bdrv_drained_end(bs
);
1802 bdrv_unref(mirror_top_bs
);
1806 /* Make sure that the source is not resized while the job is running */
1807 s
= block_job_create(job_id
, driver
, NULL
, mirror_top_bs
,
1808 BLK_PERM_CONSISTENT_READ
,
1809 BLK_PERM_CONSISTENT_READ
| BLK_PERM_WRITE_UNCHANGED
|
1810 BLK_PERM_WRITE
, speed
,
1811 creation_flags
, cb
, opaque
, errp
);
1816 /* The block job now has a reference to this node */
1817 bdrv_unref(mirror_top_bs
);
1819 s
->mirror_top_bs
= mirror_top_bs
;
1821 /* No resize for the target either; while the mirror is still running, a
1822 * consistent read isn't necessarily possible. We could possibly allow
1823 * writes and graph modifications, though it would likely defeat the
1824 * purpose of a mirror, so leave them blocked for now.
1826 * In the case of active commit, things look a bit different, though,
1827 * because the target is an already populated backing file in active use.
1828 * We can allow anything except resize there.*/
1830 target_perms
= BLK_PERM_WRITE
;
1831 target_shared_perms
= BLK_PERM_WRITE_UNCHANGED
;
1833 if (target_is_backing
) {
1834 int64_t bs_size
, target_size
;
1835 bs_size
= bdrv_getlength(bs
);
1837 error_setg_errno(errp
, -bs_size
,
1838 "Could not inquire top image size");
1842 target_size
= bdrv_getlength(target
);
1843 if (target_size
< 0) {
1844 error_setg_errno(errp
, -target_size
,
1845 "Could not inquire base image size");
1849 if (target_size
< bs_size
) {
1850 target_perms
|= BLK_PERM_RESIZE
;
1853 target_shared_perms
|= BLK_PERM_CONSISTENT_READ
| BLK_PERM_WRITE
;
1855 bdrv_graph_rdlock_main_loop();
1856 if (bdrv_chain_contains(bs
, bdrv_skip_filters(target
))) {
1858 * We may want to allow this in the future, but it would
1859 * require taking some extra care.
1861 error_setg(errp
, "Cannot mirror to a filter on top of a node in "
1862 "the source's backing chain");
1863 bdrv_graph_rdunlock_main_loop();
1866 bdrv_graph_rdunlock_main_loop();
1869 s
->target
= blk_new(s
->common
.job
.aio_context
,
1870 target_perms
, target_shared_perms
);
1871 ret
= blk_insert_bs(s
->target
, target
, errp
);
1876 /* XXX: Mirror target could be a NBD server of target QEMU in the case
1877 * of non-shared block migration. To allow migration completion, we
1878 * have to allow "inactivate" of the target BB. When that happens, we
1879 * know the job is drained, and the vcpus are stopped, so no write
1880 * operation will be performed. Block layer already has assertions to
1882 blk_set_force_allow_inactivate(s
->target
);
1884 blk_set_allow_aio_context_change(s
->target
, true);
1885 blk_set_disable_request_queuing(s
->target
, true);
1887 bdrv_graph_rdlock_main_loop();
1888 s
->replaces
= g_strdup(replaces
);
1889 s
->on_source_error
= on_source_error
;
1890 s
->on_target_error
= on_target_error
;
1891 s
->is_none_mode
= is_none_mode
;
1892 s
->backing_mode
= backing_mode
;
1893 s
->zero_target
= zero_target
;
1894 qatomic_set(&s
->copy_mode
, copy_mode
);
1896 s
->base_overlay
= bdrv_find_overlay(bs
, base
);
1897 s
->granularity
= granularity
;
1898 s
->buf_size
= ROUND_UP(buf_size
, granularity
);
1900 if (auto_complete
) {
1901 s
->should_complete
= true;
1903 bdrv_graph_rdunlock_main_loop();
1905 s
->dirty_bitmap
= bdrv_create_dirty_bitmap(s
->mirror_top_bs
, granularity
,
1907 if (!s
->dirty_bitmap
) {
1912 * The dirty bitmap is set by bdrv_mirror_top_do_write() when not in active
1915 bdrv_disable_dirty_bitmap(s
->dirty_bitmap
);
1917 bdrv_graph_wrlock(bs
);
1918 ret
= block_job_add_bdrv(&s
->common
, "source", bs
, 0,
1919 BLK_PERM_WRITE_UNCHANGED
| BLK_PERM_WRITE
|
1920 BLK_PERM_CONSISTENT_READ
,
1923 bdrv_graph_wrunlock();
1927 /* Required permissions are already taken with blk_new() */
1928 block_job_add_bdrv(&s
->common
, "target", target
, 0, BLK_PERM_ALL
,
1931 /* In commit_active_start() all intermediate nodes disappear, so
1932 * any jobs in them must be blocked */
1933 if (target_is_backing
) {
1934 BlockDriverState
*iter
, *filtered_target
;
1935 uint64_t iter_shared_perms
;
1938 * The topmost node with
1939 * bdrv_skip_filters(filtered_target) == bdrv_skip_filters(target)
1941 filtered_target
= bdrv_cow_bs(bdrv_find_overlay(bs
, target
));
1943 assert(bdrv_skip_filters(filtered_target
) ==
1944 bdrv_skip_filters(target
));
1947 * XXX BLK_PERM_WRITE needs to be allowed so we don't block
1948 * ourselves at s->base (if writes are blocked for a node, they are
1949 * also blocked for its backing file). The other options would be a
1950 * second filter driver above s->base (== target).
1952 iter_shared_perms
= BLK_PERM_WRITE_UNCHANGED
| BLK_PERM_WRITE
;
1954 for (iter
= bdrv_filter_or_cow_bs(bs
); iter
!= target
;
1955 iter
= bdrv_filter_or_cow_bs(iter
))
1957 if (iter
== filtered_target
) {
1959 * From here on, all nodes are filters on the base.
1960 * This allows us to share BLK_PERM_CONSISTENT_READ.
1962 iter_shared_perms
|= BLK_PERM_CONSISTENT_READ
;
1965 ret
= block_job_add_bdrv(&s
->common
, "intermediate node", iter
, 0,
1966 iter_shared_perms
, errp
);
1968 bdrv_graph_wrunlock();
1973 if (bdrv_freeze_backing_chain(mirror_top_bs
, target
, errp
) < 0) {
1974 bdrv_graph_wrunlock();
1978 bdrv_graph_wrunlock();
1980 QTAILQ_INIT(&s
->ops_in_flight
);
1982 trace_mirror_start(bs
, s
, opaque
);
1983 job_start(&s
->common
.job
);
1989 /* Make sure this BDS does not go away until we have completed the graph
1991 bdrv_ref(mirror_top_bs
);
1993 g_free(s
->replaces
);
1994 blk_unref(s
->target
);
1995 bs_opaque
->job
= NULL
;
1996 if (s
->dirty_bitmap
) {
1997 bdrv_release_dirty_bitmap(s
->dirty_bitmap
);
1999 job_early_fail(&s
->common
.job
);
2002 bs_opaque
->stop
= true;
2003 bdrv_drained_begin(bs
);
2004 bdrv_graph_wrlock(bs
);
2005 assert(mirror_top_bs
->backing
->bs
== bs
);
2006 bdrv_child_refresh_perms(mirror_top_bs
, mirror_top_bs
->backing
,
2008 bdrv_replace_node(mirror_top_bs
, bs
, &error_abort
);
2009 bdrv_graph_wrunlock();
2010 bdrv_drained_end(bs
);
2012 bdrv_unref(mirror_top_bs
);
2017 void mirror_start(const char *job_id
, BlockDriverState
*bs
,
2018 BlockDriverState
*target
, const char *replaces
,
2019 int creation_flags
, int64_t speed
,
2020 uint32_t granularity
, int64_t buf_size
,
2021 MirrorSyncMode mode
, BlockMirrorBackingMode backing_mode
,
2023 BlockdevOnError on_source_error
,
2024 BlockdevOnError on_target_error
,
2025 bool unmap
, const char *filter_node_name
,
2026 MirrorCopyMode copy_mode
, Error
**errp
)
2029 BlockDriverState
*base
;
2031 GLOBAL_STATE_CODE();
2033 if ((mode
== MIRROR_SYNC_MODE_INCREMENTAL
) ||
2034 (mode
== MIRROR_SYNC_MODE_BITMAP
)) {
2035 error_setg(errp
, "Sync mode '%s' not supported",
2036 MirrorSyncMode_str(mode
));
2040 bdrv_graph_rdlock_main_loop();
2041 is_none_mode
= mode
== MIRROR_SYNC_MODE_NONE
;
2042 base
= mode
== MIRROR_SYNC_MODE_TOP
? bdrv_backing_chain_next(bs
) : NULL
;
2043 bdrv_graph_rdunlock_main_loop();
2045 mirror_start_job(job_id
, bs
, creation_flags
, target
, replaces
,
2046 speed
, granularity
, buf_size
, backing_mode
, zero_target
,
2047 on_source_error
, on_target_error
, unmap
, NULL
, NULL
,
2048 &mirror_job_driver
, is_none_mode
, base
, false,
2049 filter_node_name
, true, copy_mode
, errp
);
2052 BlockJob
*commit_active_start(const char *job_id
, BlockDriverState
*bs
,
2053 BlockDriverState
*base
, int creation_flags
,
2054 int64_t speed
, BlockdevOnError on_error
,
2055 const char *filter_node_name
,
2056 BlockCompletionFunc
*cb
, void *opaque
,
2057 bool auto_complete
, Error
**errp
)
2059 bool base_read_only
;
2062 GLOBAL_STATE_CODE();
2064 base_read_only
= bdrv_is_read_only(base
);
2066 if (base_read_only
) {
2067 if (bdrv_reopen_set_read_only(base
, false, errp
) < 0) {
2072 job
= mirror_start_job(
2073 job_id
, bs
, creation_flags
, base
, NULL
, speed
, 0, 0,
2074 MIRROR_LEAVE_BACKING_CHAIN
, false,
2075 on_error
, on_error
, true, cb
, opaque
,
2076 &commit_active_job_driver
, false, base
, auto_complete
,
2077 filter_node_name
, false, MIRROR_COPY_MODE_BACKGROUND
,
2080 goto error_restore_flags
;
2085 error_restore_flags
:
2086 /* ignore error and errp for bdrv_reopen, because we want to propagate
2087 * the original error */
2088 if (base_read_only
) {
2089 bdrv_reopen_set_read_only(base
, true, NULL
);