2 * Copyright (C) 2009-2010 Nippon Telegraph and Telephone Corporation.
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public License version
6 * 2 as published by the Free Software Foundation.
8 * You should have received a copy of the GNU General Public License
9 * along with this program. If not, see <http://www.gnu.org/licenses/>.
11 * Contributions after 2012-01-13 are licensed under the terms of the
12 * GNU GPL, version 2 or (at your option) any later version.
15 #include "qemu-common.h"
17 #include "qemu/error-report.h"
18 #include "qemu/sockets.h"
19 #include "block/block_int.h"
20 #include "qemu/bitops.h"
22 #define SD_PROTO_VER 0x01
24 #define SD_DEFAULT_ADDR "localhost"
25 #define SD_DEFAULT_PORT 7000
27 #define SD_OP_CREATE_AND_WRITE_OBJ 0x01
28 #define SD_OP_READ_OBJ 0x02
29 #define SD_OP_WRITE_OBJ 0x03
30 /* 0x04 is used internally by Sheepdog */
31 #define SD_OP_DISCARD_OBJ 0x05
33 #define SD_OP_NEW_VDI 0x11
34 #define SD_OP_LOCK_VDI 0x12
35 #define SD_OP_RELEASE_VDI 0x13
36 #define SD_OP_GET_VDI_INFO 0x14
37 #define SD_OP_READ_VDIS 0x15
38 #define SD_OP_FLUSH_VDI 0x16
39 #define SD_OP_DEL_VDI 0x17
41 #define SD_FLAG_CMD_WRITE 0x01
42 #define SD_FLAG_CMD_COW 0x02
43 #define SD_FLAG_CMD_CACHE 0x04 /* Writeback mode for cache */
44 #define SD_FLAG_CMD_DIRECT 0x08 /* Don't use cache */
46 #define SD_RES_SUCCESS 0x00 /* Success */
47 #define SD_RES_UNKNOWN 0x01 /* Unknown error */
48 #define SD_RES_NO_OBJ 0x02 /* No object found */
49 #define SD_RES_EIO 0x03 /* I/O error */
50 #define SD_RES_VDI_EXIST 0x04 /* Vdi exists already */
51 #define SD_RES_INVALID_PARMS 0x05 /* Invalid parameters */
52 #define SD_RES_SYSTEM_ERROR 0x06 /* System error */
53 #define SD_RES_VDI_LOCKED 0x07 /* Vdi is locked */
54 #define SD_RES_NO_VDI 0x08 /* No vdi found */
55 #define SD_RES_NO_BASE_VDI 0x09 /* No base vdi found */
56 #define SD_RES_VDI_READ 0x0A /* Cannot read requested vdi */
57 #define SD_RES_VDI_WRITE 0x0B /* Cannot write requested vdi */
58 #define SD_RES_BASE_VDI_READ 0x0C /* Cannot read base vdi */
59 #define SD_RES_BASE_VDI_WRITE 0x0D /* Cannot write base vdi */
60 #define SD_RES_NO_TAG 0x0E /* Requested tag is not found */
61 #define SD_RES_STARTUP 0x0F /* Sheepdog is on starting up */
62 #define SD_RES_VDI_NOT_LOCKED 0x10 /* Vdi is not locked */
63 #define SD_RES_SHUTDOWN 0x11 /* Sheepdog is shutting down */
64 #define SD_RES_NO_MEM 0x12 /* Cannot allocate memory */
65 #define SD_RES_FULL_VDI 0x13 /* we already have the maximum vdis */
66 #define SD_RES_VER_MISMATCH 0x14 /* Protocol version mismatch */
67 #define SD_RES_NO_SPACE 0x15 /* Server has no room for new objects */
68 #define SD_RES_WAIT_FOR_FORMAT 0x16 /* Waiting for a format operation */
69 #define SD_RES_WAIT_FOR_JOIN 0x17 /* Waiting for other nodes joining */
70 #define SD_RES_JOIN_FAILED 0x18 /* Target node had failed to join sheepdog */
71 #define SD_RES_HALT 0x19 /* Sheepdog is stopped serving IO request */
72 #define SD_RES_READONLY 0x1A /* Object is read-only */
77 * 0 - 19 (20 bits): data object space
78 * 20 - 31 (12 bits): reserved data object space
79 * 32 - 55 (24 bits): vdi object space
80 * 56 - 59 ( 4 bits): reserved vdi object space
81 * 60 - 63 ( 4 bits): object type identifier space
84 #define VDI_SPACE_SHIFT 32
85 #define VDI_BIT (UINT64_C(1) << 63)
86 #define VMSTATE_BIT (UINT64_C(1) << 62)
87 #define MAX_DATA_OBJS (UINT64_C(1) << 20)
88 #define MAX_CHILDREN 1024
89 #define SD_MAX_VDI_LEN 256
90 #define SD_MAX_VDI_TAG_LEN 256
91 #define SD_NR_VDIS (1U << 24)
92 #define SD_DATA_OBJ_SIZE (UINT64_C(1) << 22)
93 #define SD_MAX_VDI_SIZE (SD_DATA_OBJ_SIZE * MAX_DATA_OBJS)
95 #define SD_INODE_SIZE (sizeof(SheepdogInode))
96 #define CURRENT_VDI_ID 0
98 typedef struct SheepdogReq
{
104 uint32_t data_length
;
105 uint32_t opcode_specific
[8];
108 typedef struct SheepdogRsp
{
114 uint32_t data_length
;
116 uint32_t opcode_specific
[7];
119 typedef struct SheepdogObjReq
{
125 uint32_t data_length
;
133 typedef struct SheepdogObjRsp
{
139 uint32_t data_length
;
145 typedef struct SheepdogVdiReq
{
151 uint32_t data_length
;
159 typedef struct SheepdogVdiRsp
{
165 uint32_t data_length
;
172 typedef struct SheepdogInode
{
173 char name
[SD_MAX_VDI_LEN
];
174 char tag
[SD_MAX_VDI_TAG_LEN
];
177 uint64_t vm_clock_nsec
;
179 uint64_t vm_state_size
;
180 uint16_t copy_policy
;
182 uint8_t block_size_shift
;
185 uint32_t parent_vdi_id
;
186 uint32_t child_vdi_id
[MAX_CHILDREN
];
187 uint32_t data_vdi_id
[MAX_DATA_OBJS
];
191 * 64 bit FNV-1a non-zero initial basis
193 #define FNV1A_64_INIT ((uint64_t)0xcbf29ce484222325ULL)
196 * 64 bit Fowler/Noll/Vo FNV-1a hash code
198 static inline uint64_t fnv_64a_buf(void *buf
, size_t len
, uint64_t hval
)
200 unsigned char *bp
= buf
;
201 unsigned char *be
= bp
+ len
;
203 hval
^= (uint64_t) *bp
++;
204 hval
+= (hval
<< 1) + (hval
<< 4) + (hval
<< 5) +
205 (hval
<< 7) + (hval
<< 8) + (hval
<< 40);
210 static inline bool is_data_obj_writable(SheepdogInode
*inode
, unsigned int idx
)
212 return inode
->vdi_id
== inode
->data_vdi_id
[idx
];
215 static inline bool is_data_obj(uint64_t oid
)
217 return !(VDI_BIT
& oid
);
220 static inline uint64_t data_oid_to_idx(uint64_t oid
)
222 return oid
& (MAX_DATA_OBJS
- 1);
225 static inline uint64_t vid_to_vdi_oid(uint32_t vid
)
227 return VDI_BIT
| ((uint64_t)vid
<< VDI_SPACE_SHIFT
);
230 static inline uint64_t vid_to_vmstate_oid(uint32_t vid
, uint32_t idx
)
232 return VMSTATE_BIT
| ((uint64_t)vid
<< VDI_SPACE_SHIFT
) | idx
;
235 static inline uint64_t vid_to_data_oid(uint32_t vid
, uint32_t idx
)
237 return ((uint64_t)vid
<< VDI_SPACE_SHIFT
) | idx
;
240 static inline bool is_snapshot(struct SheepdogInode
*inode
)
242 return !!inode
->snap_ctime
;
247 #define DPRINTF(fmt, args...) \
249 fprintf(stdout, "%s %d: " fmt, __func__, __LINE__, ##args); \
252 #define DPRINTF(fmt, args...)
255 typedef struct SheepdogAIOCB SheepdogAIOCB
;
257 typedef struct AIOReq
{
258 SheepdogAIOCB
*aiocb
;
259 unsigned int iov_offset
;
264 unsigned int data_len
;
268 QLIST_ENTRY(AIOReq
) aio_siblings
;
278 struct SheepdogAIOCB
{
279 BlockDriverAIOCB common
;
287 enum AIOCBState aiocb_type
;
289 Coroutine
*coroutine
;
290 void (*aio_done_func
)(SheepdogAIOCB
*);
296 typedef struct BDRVSheepdogState
{
299 uint32_t min_dirty_data_idx
;
300 uint32_t max_dirty_data_idx
;
302 char name
[SD_MAX_VDI_LEN
];
304 uint32_t cache_flags
;
305 bool discard_supported
;
315 uint32_t aioreq_seq_num
;
316 QLIST_HEAD(inflight_aio_head
, AIOReq
) inflight_aio_head
;
317 QLIST_HEAD(pending_aio_head
, AIOReq
) pending_aio_head
;
320 static const char * sd_strerror(int err
)
324 static const struct {
328 {SD_RES_SUCCESS
, "Success"},
329 {SD_RES_UNKNOWN
, "Unknown error"},
330 {SD_RES_NO_OBJ
, "No object found"},
331 {SD_RES_EIO
, "I/O error"},
332 {SD_RES_VDI_EXIST
, "VDI exists already"},
333 {SD_RES_INVALID_PARMS
, "Invalid parameters"},
334 {SD_RES_SYSTEM_ERROR
, "System error"},
335 {SD_RES_VDI_LOCKED
, "VDI is already locked"},
336 {SD_RES_NO_VDI
, "No vdi found"},
337 {SD_RES_NO_BASE_VDI
, "No base VDI found"},
338 {SD_RES_VDI_READ
, "Failed read the requested VDI"},
339 {SD_RES_VDI_WRITE
, "Failed to write the requested VDI"},
340 {SD_RES_BASE_VDI_READ
, "Failed to read the base VDI"},
341 {SD_RES_BASE_VDI_WRITE
, "Failed to write the base VDI"},
342 {SD_RES_NO_TAG
, "Failed to find the requested tag"},
343 {SD_RES_STARTUP
, "The system is still booting"},
344 {SD_RES_VDI_NOT_LOCKED
, "VDI isn't locked"},
345 {SD_RES_SHUTDOWN
, "The system is shutting down"},
346 {SD_RES_NO_MEM
, "Out of memory on the server"},
347 {SD_RES_FULL_VDI
, "We already have the maximum vdis"},
348 {SD_RES_VER_MISMATCH
, "Protocol version mismatch"},
349 {SD_RES_NO_SPACE
, "Server has no space for new objects"},
350 {SD_RES_WAIT_FOR_FORMAT
, "Sheepdog is waiting for a format operation"},
351 {SD_RES_WAIT_FOR_JOIN
, "Sheepdog is waiting for other nodes joining"},
352 {SD_RES_JOIN_FAILED
, "Target node had failed to join sheepdog"},
353 {SD_RES_HALT
, "Sheepdog is stopped serving IO request"},
354 {SD_RES_READONLY
, "Object is read-only"},
357 for (i
= 0; i
< ARRAY_SIZE(errors
); ++i
) {
358 if (errors
[i
].err
== err
) {
359 return errors
[i
].desc
;
363 return "Invalid error code";
367 * Sheepdog I/O handling:
369 * 1. In sd_co_rw_vector, we send the I/O requests to the server and
370 * link the requests to the inflight_list in the
371 * BDRVSheepdogState. The function exits without waiting for
372 * receiving the response.
374 * 2. We receive the response in aio_read_response, the fd handler to
375 * the sheepdog connection. If metadata update is needed, we send
376 * the write request to the vdi object in sd_write_done, the write
377 * completion function. We switch back to sd_co_readv/writev after
378 * all the requests belonging to the AIOCB are finished.
381 static inline AIOReq
*alloc_aio_req(BDRVSheepdogState
*s
, SheepdogAIOCB
*acb
,
382 uint64_t oid
, unsigned int data_len
,
383 uint64_t offset
, uint8_t flags
,
384 uint64_t base_oid
, unsigned int iov_offset
)
388 aio_req
= g_malloc(sizeof(*aio_req
));
389 aio_req
->aiocb
= acb
;
390 aio_req
->iov_offset
= iov_offset
;
392 aio_req
->base_oid
= base_oid
;
393 aio_req
->offset
= offset
;
394 aio_req
->data_len
= data_len
;
395 aio_req
->flags
= flags
;
396 aio_req
->id
= s
->aioreq_seq_num
++;
402 static inline void free_aio_req(BDRVSheepdogState
*s
, AIOReq
*aio_req
)
404 SheepdogAIOCB
*acb
= aio_req
->aiocb
;
406 QLIST_REMOVE(aio_req
, aio_siblings
);
412 static void coroutine_fn
sd_finish_aiocb(SheepdogAIOCB
*acb
)
414 if (!acb
->canceled
) {
415 qemu_coroutine_enter(acb
->coroutine
, NULL
);
417 qemu_aio_release(acb
);
420 static void sd_aio_cancel(BlockDriverAIOCB
*blockacb
)
422 SheepdogAIOCB
*acb
= (SheepdogAIOCB
*)blockacb
;
425 * Sheepdog cannot cancel the requests which are already sent to
426 * the servers, so we just complete the request with -EIO here.
429 qemu_coroutine_enter(acb
->coroutine
, NULL
);
430 acb
->canceled
= true;
433 static const AIOCBInfo sd_aiocb_info
= {
434 .aiocb_size
= sizeof(SheepdogAIOCB
),
435 .cancel
= sd_aio_cancel
,
438 static SheepdogAIOCB
*sd_aio_setup(BlockDriverState
*bs
, QEMUIOVector
*qiov
,
439 int64_t sector_num
, int nb_sectors
)
443 acb
= qemu_aio_get(&sd_aiocb_info
, bs
, NULL
, NULL
);
447 acb
->sector_num
= sector_num
;
448 acb
->nb_sectors
= nb_sectors
;
450 acb
->aio_done_func
= NULL
;
451 acb
->canceled
= false;
452 acb
->coroutine
= qemu_coroutine_self();
458 static int connect_to_sdog(BDRVSheepdogState
*s
)
464 fd
= unix_connect(s
->host_spec
, &err
);
466 fd
= inet_connect(s
->host_spec
, &err
);
469 int ret
= socket_set_nodelay(fd
);
471 error_report("%s", strerror(errno
));
477 qerror_report_err(err
);
480 qemu_set_nonblock(fd
);
486 static coroutine_fn
int send_co_req(int sockfd
, SheepdogReq
*hdr
, void *data
,
491 ret
= qemu_co_send(sockfd
, hdr
, sizeof(*hdr
));
492 if (ret
< sizeof(*hdr
)) {
493 error_report("failed to send a req, %s", strerror(errno
));
497 ret
= qemu_co_send(sockfd
, data
, *wlen
);
499 error_report("failed to send a req, %s", strerror(errno
));
505 static void restart_co_req(void *opaque
)
507 Coroutine
*co
= opaque
;
509 qemu_coroutine_enter(co
, NULL
);
512 typedef struct SheepdogReqCo
{
522 static coroutine_fn
void do_co_req(void *opaque
)
526 SheepdogReqCo
*srco
= opaque
;
527 int sockfd
= srco
->sockfd
;
528 SheepdogReq
*hdr
= srco
->hdr
;
529 void *data
= srco
->data
;
530 unsigned int *wlen
= srco
->wlen
;
531 unsigned int *rlen
= srco
->rlen
;
533 co
= qemu_coroutine_self();
534 qemu_aio_set_fd_handler(sockfd
, NULL
, restart_co_req
, co
);
536 ret
= send_co_req(sockfd
, hdr
, data
, wlen
);
541 qemu_aio_set_fd_handler(sockfd
, restart_co_req
, NULL
, co
);
543 ret
= qemu_co_recv(sockfd
, hdr
, sizeof(*hdr
));
544 if (ret
< sizeof(*hdr
)) {
545 error_report("failed to get a rsp, %s", strerror(errno
));
550 if (*rlen
> hdr
->data_length
) {
551 *rlen
= hdr
->data_length
;
555 ret
= qemu_co_recv(sockfd
, data
, *rlen
);
557 error_report("failed to get the data, %s", strerror(errno
));
564 /* there is at most one request for this sockfd, so it is safe to
565 * set each handler to NULL. */
566 qemu_aio_set_fd_handler(sockfd
, NULL
, NULL
, NULL
);
569 srco
->finished
= true;
572 static int do_req(int sockfd
, SheepdogReq
*hdr
, void *data
,
573 unsigned int *wlen
, unsigned int *rlen
)
576 SheepdogReqCo srco
= {
586 if (qemu_in_coroutine()) {
589 co
= qemu_coroutine_create(do_co_req
);
590 qemu_coroutine_enter(co
, &srco
);
591 while (!srco
.finished
) {
599 static int coroutine_fn
add_aio_request(BDRVSheepdogState
*s
, AIOReq
*aio_req
,
600 struct iovec
*iov
, int niov
, bool create
,
601 enum AIOCBState aiocb_type
);
602 static int coroutine_fn
resend_aioreq(BDRVSheepdogState
*s
, AIOReq
*aio_req
);
605 static AIOReq
*find_pending_req(BDRVSheepdogState
*s
, uint64_t oid
)
609 QLIST_FOREACH(aio_req
, &s
->pending_aio_head
, aio_siblings
) {
610 if (aio_req
->oid
== oid
) {
619 * This function searchs pending requests to the object `oid', and
622 static void coroutine_fn
send_pending_req(BDRVSheepdogState
*s
, uint64_t oid
)
628 while ((aio_req
= find_pending_req(s
, oid
)) != NULL
) {
629 acb
= aio_req
->aiocb
;
630 /* move aio_req from pending list to inflight one */
631 QLIST_REMOVE(aio_req
, aio_siblings
);
632 QLIST_INSERT_HEAD(&s
->inflight_aio_head
, aio_req
, aio_siblings
);
633 ret
= add_aio_request(s
, aio_req
, acb
->qiov
->iov
,
634 acb
->qiov
->niov
, false, acb
->aiocb_type
);
636 error_report("add_aio_request is failed");
637 free_aio_req(s
, aio_req
);
638 if (!acb
->nr_pending
) {
639 sd_finish_aiocb(acb
);
646 * Receive responses of the I/O requests.
648 * This function is registered as a fd handler, and called from the
649 * main loop when s->fd is ready for reading responses.
651 static void coroutine_fn
aio_read_response(void *opaque
)
654 BDRVSheepdogState
*s
= opaque
;
657 AIOReq
*aio_req
= NULL
;
661 if (QLIST_EMPTY(&s
->inflight_aio_head
)) {
666 ret
= qemu_co_recv(fd
, &rsp
, sizeof(rsp
));
668 error_report("failed to get the header, %s", strerror(errno
));
672 /* find the right aio_req from the inflight aio list */
673 QLIST_FOREACH(aio_req
, &s
->inflight_aio_head
, aio_siblings
) {
674 if (aio_req
->id
== rsp
.id
) {
679 error_report("cannot find aio_req %x", rsp
.id
);
683 acb
= aio_req
->aiocb
;
685 switch (acb
->aiocb_type
) {
686 case AIOCB_WRITE_UDATA
:
687 /* this coroutine context is no longer suitable for co_recv
688 * because we may send data to update vdi objects */
690 if (!is_data_obj(aio_req
->oid
)) {
693 idx
= data_oid_to_idx(aio_req
->oid
);
695 if (s
->inode
.data_vdi_id
[idx
] != s
->inode
.vdi_id
) {
697 * If the object is newly created one, we need to update
698 * the vdi object (metadata object). min_dirty_data_idx
699 * and max_dirty_data_idx are changed to include updated
700 * index between them.
702 if (rsp
.result
== SD_RES_SUCCESS
) {
703 s
->inode
.data_vdi_id
[idx
] = s
->inode
.vdi_id
;
704 s
->max_dirty_data_idx
= MAX(idx
, s
->max_dirty_data_idx
);
705 s
->min_dirty_data_idx
= MIN(idx
, s
->min_dirty_data_idx
);
708 * Some requests may be blocked because simultaneous
709 * create requests are not allowed, so we search the
710 * pending requests here.
712 send_pending_req(s
, aio_req
->oid
);
715 case AIOCB_READ_UDATA
:
716 ret
= qemu_co_recvv(fd
, acb
->qiov
->iov
, acb
->qiov
->niov
,
717 aio_req
->iov_offset
, rsp
.data_length
);
719 error_report("failed to get the data, %s", strerror(errno
));
723 case AIOCB_FLUSH_CACHE
:
724 if (rsp
.result
== SD_RES_INVALID_PARMS
) {
725 DPRINTF("disable cache since the server doesn't support it\n");
726 s
->cache_flags
= SD_FLAG_CMD_DIRECT
;
727 rsp
.result
= SD_RES_SUCCESS
;
730 case AIOCB_DISCARD_OBJ
:
731 switch (rsp
.result
) {
732 case SD_RES_INVALID_PARMS
:
733 error_report("sheep(%s) doesn't support discard command",
735 rsp
.result
= SD_RES_SUCCESS
;
736 s
->discard_supported
= false;
739 idx
= data_oid_to_idx(aio_req
->oid
);
740 s
->inode
.data_vdi_id
[idx
] = 0;
747 switch (rsp
.result
) {
750 case SD_RES_READONLY
:
751 ret
= resend_aioreq(s
, aio_req
);
752 if (ret
== SD_RES_SUCCESS
) {
758 error_report("%s", sd_strerror(rsp
.result
));
762 free_aio_req(s
, aio_req
);
763 if (!acb
->nr_pending
) {
765 * We've finished all requests which belong to the AIOCB, so
766 * we can switch back to sd_co_readv/writev now.
768 acb
->aio_done_func(acb
);
774 static void co_read_response(void *opaque
)
776 BDRVSheepdogState
*s
= opaque
;
779 s
->co_recv
= qemu_coroutine_create(aio_read_response
);
782 qemu_coroutine_enter(s
->co_recv
, opaque
);
785 static void co_write_request(void *opaque
)
787 BDRVSheepdogState
*s
= opaque
;
789 qemu_coroutine_enter(s
->co_send
, NULL
);
793 * Return a socket discriptor to read/write objects.
795 * We cannot use this discriptor for other operations because
796 * the block driver may be on waiting response from the server.
798 static int get_sheep_fd(BDRVSheepdogState
*s
)
802 fd
= connect_to_sdog(s
);
807 qemu_aio_set_fd_handler(fd
, co_read_response
, NULL
, s
);
811 static int sd_parse_uri(BDRVSheepdogState
*s
, const char *filename
,
812 char *vdi
, uint32_t *snapid
, char *tag
)
815 QueryParams
*qp
= NULL
;
818 uri
= uri_parse(filename
);
824 if (!strcmp(uri
->scheme
, "sheepdog")) {
826 } else if (!strcmp(uri
->scheme
, "sheepdog+tcp")) {
828 } else if (!strcmp(uri
->scheme
, "sheepdog+unix")) {
835 if (uri
->path
== NULL
|| !strcmp(uri
->path
, "/")) {
839 pstrcpy(vdi
, SD_MAX_VDI_LEN
, uri
->path
+ 1);
841 qp
= query_params_parse(uri
->query
);
842 if (qp
->n
> 1 || (s
->is_unix
&& !qp
->n
) || (!s
->is_unix
&& qp
->n
)) {
848 /* sheepdog+unix:///vdiname?socket=path */
849 if (uri
->server
|| uri
->port
|| strcmp(qp
->p
[0].name
, "socket")) {
853 s
->host_spec
= g_strdup(qp
->p
[0].value
);
855 /* sheepdog[+tcp]://[host:port]/vdiname */
856 s
->host_spec
= g_strdup_printf("%s:%d", uri
->server
?: SD_DEFAULT_ADDR
,
857 uri
->port
?: SD_DEFAULT_PORT
);
862 *snapid
= strtoul(uri
->fragment
, NULL
, 10);
864 pstrcpy(tag
, SD_MAX_VDI_TAG_LEN
, uri
->fragment
);
867 *snapid
= CURRENT_VDI_ID
; /* search current vdi */
872 query_params_free(qp
);
879 * Parse a filename (old syntax)
881 * filename must be one of the following formats:
883 * 2. [vdiname]:[snapid]
885 * 4. [hostname]:[port]:[vdiname]
886 * 5. [hostname]:[port]:[vdiname]:[snapid]
887 * 6. [hostname]:[port]:[vdiname]:[tag]
889 * You can boot from the snapshot images by specifying `snapid` or
892 * You can run VMs outside the Sheepdog cluster by specifying
893 * `hostname' and `port' (experimental).
895 static int parse_vdiname(BDRVSheepdogState
*s
, const char *filename
,
896 char *vdi
, uint32_t *snapid
, char *tag
)
899 const char *host_spec
, *vdi_spec
;
902 strstart(filename
, "sheepdog:", (const char **)&filename
);
903 p
= q
= g_strdup(filename
);
905 /* count the number of separators */
915 /* use the first two tokens as host_spec. */
928 p
= strchr(vdi_spec
, ':');
933 uri
= g_strdup_printf("sheepdog://%s/%s", host_spec
, vdi_spec
);
935 ret
= sd_parse_uri(s
, uri
, vdi
, snapid
, tag
);
943 static int find_vdi_name(BDRVSheepdogState
*s
, const char *filename
,
944 uint32_t snapid
, const char *tag
, uint32_t *vid
,
949 SheepdogVdiRsp
*rsp
= (SheepdogVdiRsp
*)&hdr
;
950 unsigned int wlen
, rlen
= 0;
951 char buf
[SD_MAX_VDI_LEN
+ SD_MAX_VDI_TAG_LEN
];
953 fd
= connect_to_sdog(s
);
958 /* This pair of strncpy calls ensures that the buffer is zero-filled,
959 * which is desirable since we'll soon be sending those bytes, and
960 * don't want the send_req to read uninitialized data.
962 strncpy(buf
, filename
, SD_MAX_VDI_LEN
);
963 strncpy(buf
+ SD_MAX_VDI_LEN
, tag
, SD_MAX_VDI_TAG_LEN
);
965 memset(&hdr
, 0, sizeof(hdr
));
967 hdr
.opcode
= SD_OP_LOCK_VDI
;
969 hdr
.opcode
= SD_OP_GET_VDI_INFO
;
971 wlen
= SD_MAX_VDI_LEN
+ SD_MAX_VDI_TAG_LEN
;
972 hdr
.proto_ver
= SD_PROTO_VER
;
973 hdr
.data_length
= wlen
;
975 hdr
.flags
= SD_FLAG_CMD_WRITE
;
977 ret
= do_req(fd
, (SheepdogReq
*)&hdr
, buf
, &wlen
, &rlen
);
982 if (rsp
->result
!= SD_RES_SUCCESS
) {
983 error_report("cannot get vdi info, %s, %s %d %s",
984 sd_strerror(rsp
->result
), filename
, snapid
, tag
);
985 if (rsp
->result
== SD_RES_NO_VDI
) {
1000 static int coroutine_fn
add_aio_request(BDRVSheepdogState
*s
, AIOReq
*aio_req
,
1001 struct iovec
*iov
, int niov
, bool create
,
1002 enum AIOCBState aiocb_type
)
1004 int nr_copies
= s
->inode
.nr_copies
;
1006 unsigned int wlen
= 0;
1008 uint64_t oid
= aio_req
->oid
;
1009 unsigned int datalen
= aio_req
->data_len
;
1010 uint64_t offset
= aio_req
->offset
;
1011 uint8_t flags
= aio_req
->flags
;
1012 uint64_t old_oid
= aio_req
->base_oid
;
1015 error_report("bug");
1018 memset(&hdr
, 0, sizeof(hdr
));
1020 switch (aiocb_type
) {
1021 case AIOCB_FLUSH_CACHE
:
1022 hdr
.opcode
= SD_OP_FLUSH_VDI
;
1024 case AIOCB_READ_UDATA
:
1025 hdr
.opcode
= SD_OP_READ_OBJ
;
1028 case AIOCB_WRITE_UDATA
:
1030 hdr
.opcode
= SD_OP_CREATE_AND_WRITE_OBJ
;
1032 hdr
.opcode
= SD_OP_WRITE_OBJ
;
1035 hdr
.flags
= SD_FLAG_CMD_WRITE
| flags
;
1037 case AIOCB_DISCARD_OBJ
:
1038 hdr
.opcode
= SD_OP_DISCARD_OBJ
;
1042 if (s
->cache_flags
) {
1043 hdr
.flags
|= s
->cache_flags
;
1047 hdr
.cow_oid
= old_oid
;
1048 hdr
.copies
= s
->inode
.nr_copies
;
1050 hdr
.data_length
= datalen
;
1051 hdr
.offset
= offset
;
1053 hdr
.id
= aio_req
->id
;
1055 qemu_co_mutex_lock(&s
->lock
);
1056 s
->co_send
= qemu_coroutine_self();
1057 qemu_aio_set_fd_handler(s
->fd
, co_read_response
, co_write_request
, s
);
1058 socket_set_cork(s
->fd
, 1);
1061 ret
= qemu_co_send(s
->fd
, &hdr
, sizeof(hdr
));
1063 qemu_co_mutex_unlock(&s
->lock
);
1064 error_report("failed to send a req, %s", strerror(errno
));
1069 ret
= qemu_co_sendv(s
->fd
, iov
, niov
, aio_req
->iov_offset
, wlen
);
1071 qemu_co_mutex_unlock(&s
->lock
);
1072 error_report("failed to send a data, %s", strerror(errno
));
1077 socket_set_cork(s
->fd
, 0);
1078 qemu_aio_set_fd_handler(s
->fd
, co_read_response
, NULL
, s
);
1079 qemu_co_mutex_unlock(&s
->lock
);
1084 static int read_write_object(int fd
, char *buf
, uint64_t oid
, int copies
,
1085 unsigned int datalen
, uint64_t offset
,
1086 bool write
, bool create
, uint32_t cache_flags
)
1089 SheepdogObjRsp
*rsp
= (SheepdogObjRsp
*)&hdr
;
1090 unsigned int wlen
, rlen
;
1093 memset(&hdr
, 0, sizeof(hdr
));
1098 hdr
.flags
= SD_FLAG_CMD_WRITE
;
1100 hdr
.opcode
= SD_OP_CREATE_AND_WRITE_OBJ
;
1102 hdr
.opcode
= SD_OP_WRITE_OBJ
;
1107 hdr
.opcode
= SD_OP_READ_OBJ
;
1110 hdr
.flags
|= cache_flags
;
1113 hdr
.data_length
= datalen
;
1114 hdr
.offset
= offset
;
1115 hdr
.copies
= copies
;
1117 ret
= do_req(fd
, (SheepdogReq
*)&hdr
, buf
, &wlen
, &rlen
);
1119 error_report("failed to send a request to the sheep");
1123 switch (rsp
->result
) {
1124 case SD_RES_SUCCESS
:
1127 error_report("%s", sd_strerror(rsp
->result
));
1132 static int read_object(int fd
, char *buf
, uint64_t oid
, int copies
,
1133 unsigned int datalen
, uint64_t offset
,
1134 uint32_t cache_flags
)
1136 return read_write_object(fd
, buf
, oid
, copies
, datalen
, offset
, false,
1137 false, cache_flags
);
1140 static int write_object(int fd
, char *buf
, uint64_t oid
, int copies
,
1141 unsigned int datalen
, uint64_t offset
, bool create
,
1142 uint32_t cache_flags
)
1144 return read_write_object(fd
, buf
, oid
, copies
, datalen
, offset
, true,
1145 create
, cache_flags
);
1148 /* update inode with the latest state */
1149 static int reload_inode(BDRVSheepdogState
*s
, uint32_t snapid
, const char *tag
)
1151 SheepdogInode
*inode
;
1155 fd
= connect_to_sdog(s
);
1160 inode
= g_malloc(sizeof(s
->inode
));
1162 ret
= find_vdi_name(s
, s
->name
, snapid
, tag
, &vid
, false);
1167 ret
= read_object(fd
, (char *)inode
, vid_to_vdi_oid(vid
),
1168 s
->inode
.nr_copies
, sizeof(*inode
), 0, s
->cache_flags
);
1173 if (inode
->vdi_id
!= s
->inode
.vdi_id
) {
1174 memcpy(&s
->inode
, inode
, sizeof(s
->inode
));
1184 static int coroutine_fn
resend_aioreq(BDRVSheepdogState
*s
, AIOReq
*aio_req
)
1186 SheepdogAIOCB
*acb
= aio_req
->aiocb
;
1187 bool create
= false;
1190 ret
= reload_inode(s
, 0, "");
1195 aio_req
->oid
= vid_to_data_oid(s
->inode
.vdi_id
,
1196 data_oid_to_idx(aio_req
->oid
));
1198 /* check whether this request becomes a CoW one */
1199 if (acb
->aiocb_type
== AIOCB_WRITE_UDATA
) {
1200 int idx
= data_oid_to_idx(aio_req
->oid
);
1203 if (s
->inode
.data_vdi_id
[idx
] == 0) {
1207 if (is_data_obj_writable(&s
->inode
, idx
)) {
1211 /* link to the pending list if there is another CoW request to
1212 * the same object */
1213 QLIST_FOREACH(areq
, &s
->inflight_aio_head
, aio_siblings
) {
1214 if (areq
!= aio_req
&& areq
->oid
== aio_req
->oid
) {
1215 DPRINTF("simultaneous CoW to %" PRIx64
"\n", aio_req
->oid
);
1216 QLIST_REMOVE(aio_req
, aio_siblings
);
1217 QLIST_INSERT_HEAD(&s
->pending_aio_head
, aio_req
, aio_siblings
);
1218 return SD_RES_SUCCESS
;
1222 aio_req
->base_oid
= vid_to_data_oid(s
->inode
.data_vdi_id
[idx
], idx
);
1223 aio_req
->flags
|= SD_FLAG_CMD_COW
;
1227 return add_aio_request(s
, aio_req
, acb
->qiov
->iov
, acb
->qiov
->niov
,
1228 create
, acb
->aiocb_type
);
1231 /* TODO Convert to fine grained options */
1232 static QemuOptsList runtime_opts
= {
1234 .head
= QTAILQ_HEAD_INITIALIZER(runtime_opts
.head
),
1238 .type
= QEMU_OPT_STRING
,
1239 .help
= "URL to the sheepdog image",
1241 { /* end of list */ }
1245 static int sd_open(BlockDriverState
*bs
, QDict
*options
, int flags
,
1250 BDRVSheepdogState
*s
= bs
->opaque
;
1251 char vdi
[SD_MAX_VDI_LEN
], tag
[SD_MAX_VDI_TAG_LEN
];
1255 Error
*local_err
= NULL
;
1256 const char *filename
;
1258 opts
= qemu_opts_create_nofail(&runtime_opts
);
1259 qemu_opts_absorb_qdict(opts
, options
, &local_err
);
1260 if (error_is_set(&local_err
)) {
1261 qerror_report_err(local_err
);
1262 error_free(local_err
);
1267 filename
= qemu_opt_get(opts
, "filename");
1269 QLIST_INIT(&s
->inflight_aio_head
);
1270 QLIST_INIT(&s
->pending_aio_head
);
1273 memset(vdi
, 0, sizeof(vdi
));
1274 memset(tag
, 0, sizeof(tag
));
1276 if (strstr(filename
, "://")) {
1277 ret
= sd_parse_uri(s
, filename
, vdi
, &snapid
, tag
);
1279 ret
= parse_vdiname(s
, filename
, vdi
, &snapid
, tag
);
1284 s
->fd
= get_sheep_fd(s
);
1290 ret
= find_vdi_name(s
, vdi
, snapid
, tag
, &vid
, true);
1296 * QEMU block layer emulates writethrough cache as 'writeback + flush', so
1297 * we always set SD_FLAG_CMD_CACHE (writeback cache) as default.
1299 s
->cache_flags
= SD_FLAG_CMD_CACHE
;
1300 if (flags
& BDRV_O_NOCACHE
) {
1301 s
->cache_flags
= SD_FLAG_CMD_DIRECT
;
1303 s
->discard_supported
= true;
1305 if (snapid
|| tag
[0] != '\0') {
1306 DPRINTF("%" PRIx32
" snapshot inode was open.\n", vid
);
1307 s
->is_snapshot
= true;
1310 fd
= connect_to_sdog(s
);
1316 buf
= g_malloc(SD_INODE_SIZE
);
1317 ret
= read_object(fd
, buf
, vid_to_vdi_oid(vid
), 0, SD_INODE_SIZE
, 0,
1326 memcpy(&s
->inode
, buf
, sizeof(s
->inode
));
1327 s
->min_dirty_data_idx
= UINT32_MAX
;
1328 s
->max_dirty_data_idx
= 0;
1330 bs
->total_sectors
= s
->inode
.vdi_size
/ BDRV_SECTOR_SIZE
;
1331 pstrcpy(s
->name
, sizeof(s
->name
), vdi
);
1332 qemu_co_mutex_init(&s
->lock
);
1333 qemu_opts_del(opts
);
1337 qemu_aio_set_fd_handler(s
->fd
, NULL
, NULL
, NULL
);
1341 qemu_opts_del(opts
);
1346 static int do_sd_create(BDRVSheepdogState
*s
, char *filename
, int64_t vdi_size
,
1347 uint32_t base_vid
, uint32_t *vdi_id
, int snapshot
)
1350 SheepdogVdiRsp
*rsp
= (SheepdogVdiRsp
*)&hdr
;
1352 unsigned int wlen
, rlen
= 0;
1353 char buf
[SD_MAX_VDI_LEN
];
1355 fd
= connect_to_sdog(s
);
1360 /* FIXME: would it be better to fail (e.g., return -EIO) when filename
1361 * does not fit in buf? For now, just truncate and avoid buffer overrun.
1363 memset(buf
, 0, sizeof(buf
));
1364 pstrcpy(buf
, sizeof(buf
), filename
);
1366 memset(&hdr
, 0, sizeof(hdr
));
1367 hdr
.opcode
= SD_OP_NEW_VDI
;
1368 hdr
.vdi_id
= base_vid
;
1370 wlen
= SD_MAX_VDI_LEN
;
1372 hdr
.flags
= SD_FLAG_CMD_WRITE
;
1373 hdr
.snapid
= snapshot
;
1375 hdr
.data_length
= wlen
;
1376 hdr
.vdi_size
= vdi_size
;
1378 ret
= do_req(fd
, (SheepdogReq
*)&hdr
, buf
, &wlen
, &rlen
);
1386 if (rsp
->result
!= SD_RES_SUCCESS
) {
1387 error_report("%s, %s", sd_strerror(rsp
->result
), filename
);
1392 *vdi_id
= rsp
->vdi_id
;
1398 static int sd_prealloc(const char *filename
)
1400 BlockDriverState
*bs
= NULL
;
1401 uint32_t idx
, max_idx
;
1403 void *buf
= g_malloc0(SD_DATA_OBJ_SIZE
);
1406 ret
= bdrv_file_open(&bs
, filename
, NULL
, BDRV_O_RDWR
);
1411 vdi_size
= bdrv_getlength(bs
);
1416 max_idx
= DIV_ROUND_UP(vdi_size
, SD_DATA_OBJ_SIZE
);
1418 for (idx
= 0; idx
< max_idx
; idx
++) {
1420 * The created image can be a cloned image, so we need to read
1421 * a data from the source image.
1423 ret
= bdrv_pread(bs
, idx
* SD_DATA_OBJ_SIZE
, buf
, SD_DATA_OBJ_SIZE
);
1427 ret
= bdrv_pwrite(bs
, idx
* SD_DATA_OBJ_SIZE
, buf
, SD_DATA_OBJ_SIZE
);
1441 static int sd_create(const char *filename
, QEMUOptionParameter
*options
,
1445 uint32_t vid
= 0, base_vid
= 0;
1446 int64_t vdi_size
= 0;
1447 char *backing_file
= NULL
;
1448 BDRVSheepdogState
*s
;
1449 char vdi
[SD_MAX_VDI_LEN
], tag
[SD_MAX_VDI_TAG_LEN
];
1451 bool prealloc
= false;
1453 s
= g_malloc0(sizeof(BDRVSheepdogState
));
1455 memset(vdi
, 0, sizeof(vdi
));
1456 memset(tag
, 0, sizeof(tag
));
1457 if (strstr(filename
, "://")) {
1458 ret
= sd_parse_uri(s
, filename
, vdi
, &snapid
, tag
);
1460 ret
= parse_vdiname(s
, filename
, vdi
, &snapid
, tag
);
1466 while (options
&& options
->name
) {
1467 if (!strcmp(options
->name
, BLOCK_OPT_SIZE
)) {
1468 vdi_size
= options
->value
.n
;
1469 } else if (!strcmp(options
->name
, BLOCK_OPT_BACKING_FILE
)) {
1470 backing_file
= options
->value
.s
;
1471 } else if (!strcmp(options
->name
, BLOCK_OPT_PREALLOC
)) {
1472 if (!options
->value
.s
|| !strcmp(options
->value
.s
, "off")) {
1474 } else if (!strcmp(options
->value
.s
, "full")) {
1477 error_report("Invalid preallocation mode: '%s'",
1486 if (vdi_size
> SD_MAX_VDI_SIZE
) {
1487 error_report("too big image size");
1493 BlockDriverState
*bs
;
1494 BDRVSheepdogState
*s
;
1497 /* Currently, only Sheepdog backing image is supported. */
1498 drv
= bdrv_find_protocol(backing_file
, true);
1499 if (!drv
|| strcmp(drv
->protocol_name
, "sheepdog") != 0) {
1500 error_report("backing_file must be a sheepdog image");
1505 ret
= bdrv_file_open(&bs
, backing_file
, NULL
, 0);
1512 if (!is_snapshot(&s
->inode
)) {
1513 error_report("cannot clone from a non snapshot vdi");
1519 base_vid
= s
->inode
.vdi_id
;
1523 ret
= do_sd_create(s
, vdi
, vdi_size
, base_vid
, &vid
, 0);
1524 if (!prealloc
|| ret
) {
1528 ret
= sd_prealloc(filename
);
1534 static void sd_close(BlockDriverState
*bs
)
1536 BDRVSheepdogState
*s
= bs
->opaque
;
1538 SheepdogVdiRsp
*rsp
= (SheepdogVdiRsp
*)&hdr
;
1539 unsigned int wlen
, rlen
= 0;
1542 DPRINTF("%s\n", s
->name
);
1544 fd
= connect_to_sdog(s
);
1549 memset(&hdr
, 0, sizeof(hdr
));
1551 hdr
.opcode
= SD_OP_RELEASE_VDI
;
1552 hdr
.vdi_id
= s
->inode
.vdi_id
;
1553 wlen
= strlen(s
->name
) + 1;
1554 hdr
.data_length
= wlen
;
1555 hdr
.flags
= SD_FLAG_CMD_WRITE
;
1557 ret
= do_req(fd
, (SheepdogReq
*)&hdr
, s
->name
, &wlen
, &rlen
);
1561 if (!ret
&& rsp
->result
!= SD_RES_SUCCESS
&&
1562 rsp
->result
!= SD_RES_VDI_NOT_LOCKED
) {
1563 error_report("%s, %s", sd_strerror(rsp
->result
), s
->name
);
1566 qemu_aio_set_fd_handler(s
->fd
, NULL
, NULL
, NULL
);
1568 g_free(s
->host_spec
);
1571 static int64_t sd_getlength(BlockDriverState
*bs
)
1573 BDRVSheepdogState
*s
= bs
->opaque
;
1575 return s
->inode
.vdi_size
;
1578 static int sd_truncate(BlockDriverState
*bs
, int64_t offset
)
1580 BDRVSheepdogState
*s
= bs
->opaque
;
1582 unsigned int datalen
;
1584 if (offset
< s
->inode
.vdi_size
) {
1585 error_report("shrinking is not supported");
1587 } else if (offset
> SD_MAX_VDI_SIZE
) {
1588 error_report("too big image size");
1592 fd
= connect_to_sdog(s
);
1597 /* we don't need to update entire object */
1598 datalen
= SD_INODE_SIZE
- sizeof(s
->inode
.data_vdi_id
);
1599 s
->inode
.vdi_size
= offset
;
1600 ret
= write_object(fd
, (char *)&s
->inode
, vid_to_vdi_oid(s
->inode
.vdi_id
),
1601 s
->inode
.nr_copies
, datalen
, 0, false, s
->cache_flags
);
1605 error_report("failed to update an inode.");
1612 * This function is called after writing data objects. If we need to
1613 * update metadata, this sends a write request to the vdi object.
1614 * Otherwise, this switches back to sd_co_readv/writev.
1616 static void coroutine_fn
sd_write_done(SheepdogAIOCB
*acb
)
1619 BDRVSheepdogState
*s
= acb
->common
.bs
->opaque
;
1622 uint32_t offset
, data_len
, mn
, mx
;
1624 mn
= s
->min_dirty_data_idx
;
1625 mx
= s
->max_dirty_data_idx
;
1627 /* we need to update the vdi object. */
1628 offset
= sizeof(s
->inode
) - sizeof(s
->inode
.data_vdi_id
) +
1629 mn
* sizeof(s
->inode
.data_vdi_id
[0]);
1630 data_len
= (mx
- mn
+ 1) * sizeof(s
->inode
.data_vdi_id
[0]);
1632 s
->min_dirty_data_idx
= UINT32_MAX
;
1633 s
->max_dirty_data_idx
= 0;
1635 iov
.iov_base
= &s
->inode
;
1636 iov
.iov_len
= sizeof(s
->inode
);
1637 aio_req
= alloc_aio_req(s
, acb
, vid_to_vdi_oid(s
->inode
.vdi_id
),
1638 data_len
, offset
, 0, 0, offset
);
1639 QLIST_INSERT_HEAD(&s
->inflight_aio_head
, aio_req
, aio_siblings
);
1640 ret
= add_aio_request(s
, aio_req
, &iov
, 1, false, AIOCB_WRITE_UDATA
);
1642 free_aio_req(s
, aio_req
);
1647 acb
->aio_done_func
= sd_finish_aiocb
;
1648 acb
->aiocb_type
= AIOCB_WRITE_UDATA
;
1652 sd_finish_aiocb(acb
);
1655 /* Delete current working VDI on the snapshot chain */
1656 static bool sd_delete(BDRVSheepdogState
*s
)
1658 unsigned int wlen
= SD_MAX_VDI_LEN
, rlen
= 0;
1659 SheepdogVdiReq hdr
= {
1660 .opcode
= SD_OP_DEL_VDI
,
1661 .vdi_id
= s
->inode
.vdi_id
,
1662 .data_length
= wlen
,
1663 .flags
= SD_FLAG_CMD_WRITE
,
1665 SheepdogVdiRsp
*rsp
= (SheepdogVdiRsp
*)&hdr
;
1668 fd
= connect_to_sdog(s
);
1673 ret
= do_req(fd
, (SheepdogReq
*)&hdr
, s
->name
, &wlen
, &rlen
);
1678 switch (rsp
->result
) {
1680 error_report("%s was already deleted", s
->name
);
1682 case SD_RES_SUCCESS
:
1685 error_report("%s, %s", sd_strerror(rsp
->result
), s
->name
);
1693 * Create a writable VDI from a snapshot
1695 static int sd_create_branch(BDRVSheepdogState
*s
)
1702 DPRINTF("%" PRIx32
" is snapshot.\n", s
->inode
.vdi_id
);
1704 buf
= g_malloc(SD_INODE_SIZE
);
1707 * Even If deletion fails, we will just create extra snapshot based on
1708 * the workding VDI which was supposed to be deleted. So no need to
1711 deleted
= sd_delete(s
);
1712 ret
= do_sd_create(s
, s
->name
, s
->inode
.vdi_size
, s
->inode
.vdi_id
, &vid
,
1718 DPRINTF("%" PRIx32
" is created.\n", vid
);
1720 fd
= connect_to_sdog(s
);
1726 ret
= read_object(fd
, buf
, vid_to_vdi_oid(vid
), s
->inode
.nr_copies
,
1727 SD_INODE_SIZE
, 0, s
->cache_flags
);
1735 memcpy(&s
->inode
, buf
, sizeof(s
->inode
));
1737 s
->is_snapshot
= false;
1739 DPRINTF("%" PRIx32
" was newly created.\n", s
->inode
.vdi_id
);
1748 * Send I/O requests to the server.
1750 * This function sends requests to the server, links the requests to
1751 * the inflight_list in BDRVSheepdogState, and exits without
1752 * waiting the response. The responses are received in the
1753 * `aio_read_response' function which is called from the main loop as
1756 * Returns 1 when we need to wait a response, 0 when there is no sent
1757 * request and -errno in error cases.
1759 static int coroutine_fn
sd_co_rw_vector(void *p
)
1761 SheepdogAIOCB
*acb
= p
;
1763 unsigned long len
, done
= 0, total
= acb
->nb_sectors
* BDRV_SECTOR_SIZE
;
1764 unsigned long idx
= acb
->sector_num
* BDRV_SECTOR_SIZE
/ SD_DATA_OBJ_SIZE
;
1766 uint64_t offset
= (acb
->sector_num
* BDRV_SECTOR_SIZE
) % SD_DATA_OBJ_SIZE
;
1767 BDRVSheepdogState
*s
= acb
->common
.bs
->opaque
;
1768 SheepdogInode
*inode
= &s
->inode
;
1771 if (acb
->aiocb_type
== AIOCB_WRITE_UDATA
&& s
->is_snapshot
) {
1773 * In the case we open the snapshot VDI, Sheepdog creates the
1774 * writable VDI when we do a write operation first.
1776 ret
= sd_create_branch(s
);
1784 * Make sure we don't free the aiocb before we are done with all requests.
1785 * This additional reference is dropped at the end of this function.
1789 while (done
!= total
) {
1791 uint64_t old_oid
= 0;
1792 bool create
= false;
1794 oid
= vid_to_data_oid(inode
->data_vdi_id
[idx
], idx
);
1796 len
= MIN(total
- done
, SD_DATA_OBJ_SIZE
- offset
);
1798 switch (acb
->aiocb_type
) {
1799 case AIOCB_READ_UDATA
:
1800 if (!inode
->data_vdi_id
[idx
]) {
1801 qemu_iovec_memset(acb
->qiov
, done
, 0, len
);
1805 case AIOCB_WRITE_UDATA
:
1806 if (!inode
->data_vdi_id
[idx
]) {
1808 } else if (!is_data_obj_writable(inode
, idx
)) {
1812 flags
= SD_FLAG_CMD_COW
;
1815 case AIOCB_DISCARD_OBJ
:
1817 * We discard the object only when the whole object is
1818 * 1) allocated 2) trimmed. Otherwise, simply skip it.
1820 if (len
!= SD_DATA_OBJ_SIZE
|| inode
->data_vdi_id
[idx
] == 0) {
1829 DPRINTF("update ino (%" PRIu32
") %" PRIu64
" %" PRIu64
" %ld\n",
1831 vid_to_data_oid(inode
->data_vdi_id
[idx
], idx
), idx
);
1832 oid
= vid_to_data_oid(inode
->vdi_id
, idx
);
1833 DPRINTF("new oid %" PRIx64
"\n", oid
);
1836 aio_req
= alloc_aio_req(s
, acb
, oid
, len
, offset
, flags
, old_oid
, done
);
1840 QLIST_FOREACH(areq
, &s
->inflight_aio_head
, aio_siblings
) {
1841 if (areq
->oid
== oid
) {
1843 * Sheepdog cannot handle simultaneous create
1844 * requests to the same object. So we cannot send
1845 * the request until the previous request
1849 aio_req
->base_oid
= 0;
1850 QLIST_INSERT_HEAD(&s
->pending_aio_head
, aio_req
,
1857 QLIST_INSERT_HEAD(&s
->inflight_aio_head
, aio_req
, aio_siblings
);
1858 ret
= add_aio_request(s
, aio_req
, acb
->qiov
->iov
, acb
->qiov
->niov
,
1859 create
, acb
->aiocb_type
);
1861 error_report("add_aio_request is failed");
1862 free_aio_req(s
, aio_req
);
1872 if (!--acb
->nr_pending
) {
1878 static coroutine_fn
int sd_co_writev(BlockDriverState
*bs
, int64_t sector_num
,
1879 int nb_sectors
, QEMUIOVector
*qiov
)
1884 if (bs
->growable
&& sector_num
+ nb_sectors
> bs
->total_sectors
) {
1885 ret
= sd_truncate(bs
, (sector_num
+ nb_sectors
) * BDRV_SECTOR_SIZE
);
1889 bs
->total_sectors
= sector_num
+ nb_sectors
;
1892 acb
= sd_aio_setup(bs
, qiov
, sector_num
, nb_sectors
);
1893 acb
->aio_done_func
= sd_write_done
;
1894 acb
->aiocb_type
= AIOCB_WRITE_UDATA
;
1896 ret
= sd_co_rw_vector(acb
);
1898 qemu_aio_release(acb
);
1902 qemu_coroutine_yield();
1907 static coroutine_fn
int sd_co_readv(BlockDriverState
*bs
, int64_t sector_num
,
1908 int nb_sectors
, QEMUIOVector
*qiov
)
1913 acb
= sd_aio_setup(bs
, qiov
, sector_num
, nb_sectors
);
1914 acb
->aiocb_type
= AIOCB_READ_UDATA
;
1915 acb
->aio_done_func
= sd_finish_aiocb
;
1917 ret
= sd_co_rw_vector(acb
);
1919 qemu_aio_release(acb
);
1923 qemu_coroutine_yield();
1928 static int coroutine_fn
sd_co_flush_to_disk(BlockDriverState
*bs
)
1930 BDRVSheepdogState
*s
= bs
->opaque
;
1935 if (s
->cache_flags
!= SD_FLAG_CMD_CACHE
) {
1939 acb
= sd_aio_setup(bs
, NULL
, 0, 0);
1940 acb
->aiocb_type
= AIOCB_FLUSH_CACHE
;
1941 acb
->aio_done_func
= sd_finish_aiocb
;
1943 aio_req
= alloc_aio_req(s
, acb
, vid_to_vdi_oid(s
->inode
.vdi_id
),
1945 QLIST_INSERT_HEAD(&s
->inflight_aio_head
, aio_req
, aio_siblings
);
1946 ret
= add_aio_request(s
, aio_req
, NULL
, 0, false, acb
->aiocb_type
);
1948 error_report("add_aio_request is failed");
1949 free_aio_req(s
, aio_req
);
1950 qemu_aio_release(acb
);
1954 qemu_coroutine_yield();
1958 static int sd_snapshot_create(BlockDriverState
*bs
, QEMUSnapshotInfo
*sn_info
)
1960 BDRVSheepdogState
*s
= bs
->opaque
;
1963 SheepdogInode
*inode
;
1964 unsigned int datalen
;
1966 DPRINTF("sn_info: name %s id_str %s s: name %s vm_state_size %" PRId64
" "
1967 "is_snapshot %d\n", sn_info
->name
, sn_info
->id_str
,
1968 s
->name
, sn_info
->vm_state_size
, s
->is_snapshot
);
1970 if (s
->is_snapshot
) {
1971 error_report("You can't create a snapshot of a snapshot VDI, "
1972 "%s (%" PRIu32
").", s
->name
, s
->inode
.vdi_id
);
1977 DPRINTF("%s %s\n", sn_info
->name
, sn_info
->id_str
);
1979 s
->inode
.vm_state_size
= sn_info
->vm_state_size
;
1980 s
->inode
.vm_clock_nsec
= sn_info
->vm_clock_nsec
;
1981 /* It appears that inode.tag does not require a NUL terminator,
1982 * which means this use of strncpy is ok.
1984 strncpy(s
->inode
.tag
, sn_info
->name
, sizeof(s
->inode
.tag
));
1985 /* we don't need to update entire object */
1986 datalen
= SD_INODE_SIZE
- sizeof(s
->inode
.data_vdi_id
);
1988 /* refresh inode. */
1989 fd
= connect_to_sdog(s
);
1995 ret
= write_object(fd
, (char *)&s
->inode
, vid_to_vdi_oid(s
->inode
.vdi_id
),
1996 s
->inode
.nr_copies
, datalen
, 0, false, s
->cache_flags
);
1998 error_report("failed to write snapshot's inode.");
2002 ret
= do_sd_create(s
, s
->name
, s
->inode
.vdi_size
, s
->inode
.vdi_id
, &new_vid
,
2005 error_report("failed to create inode for snapshot. %s",
2010 inode
= (SheepdogInode
*)g_malloc(datalen
);
2012 ret
= read_object(fd
, (char *)inode
, vid_to_vdi_oid(new_vid
),
2013 s
->inode
.nr_copies
, datalen
, 0, s
->cache_flags
);
2016 error_report("failed to read new inode info. %s", strerror(errno
));
2020 memcpy(&s
->inode
, inode
, datalen
);
2021 DPRINTF("s->inode: name %s snap_id %x oid %x\n",
2022 s
->inode
.name
, s
->inode
.snap_id
, s
->inode
.vdi_id
);
2030 * We implement rollback(loadvm) operation to the specified snapshot by
2031 * 1) switch to the snapshot
2032 * 2) rely on sd_create_branch to delete working VDI and
2033 * 3) create a new working VDI based on the speicified snapshot
2035 static int sd_snapshot_goto(BlockDriverState
*bs
, const char *snapshot_id
)
2037 BDRVSheepdogState
*s
= bs
->opaque
;
2038 BDRVSheepdogState
*old_s
;
2039 char tag
[SD_MAX_VDI_TAG_LEN
];
2040 uint32_t snapid
= 0;
2043 old_s
= g_malloc(sizeof(BDRVSheepdogState
));
2045 memcpy(old_s
, s
, sizeof(BDRVSheepdogState
));
2047 snapid
= strtoul(snapshot_id
, NULL
, 10);
2051 pstrcpy(tag
, sizeof(tag
), snapshot_id
);
2054 ret
= reload_inode(s
, snapid
, tag
);
2059 ret
= sd_create_branch(s
);
2068 /* recover bdrv_sd_state */
2069 memcpy(s
, old_s
, sizeof(BDRVSheepdogState
));
2072 error_report("failed to open. recover old bdrv_sd_state.");
2077 static int sd_snapshot_delete(BlockDriverState
*bs
,
2078 const char *snapshot_id
,
2082 /* FIXME: Delete specified snapshot id. */
2086 static int sd_snapshot_list(BlockDriverState
*bs
, QEMUSnapshotInfo
**psn_tab
)
2088 BDRVSheepdogState
*s
= bs
->opaque
;
2090 int fd
, nr
= 1024, ret
, max
= BITS_TO_LONGS(SD_NR_VDIS
) * sizeof(long);
2091 QEMUSnapshotInfo
*sn_tab
= NULL
;
2092 unsigned wlen
, rlen
;
2094 static SheepdogInode inode
;
2095 unsigned long *vdi_inuse
;
2096 unsigned int start_nr
;
2100 vdi_inuse
= g_malloc(max
);
2102 fd
= connect_to_sdog(s
);
2111 memset(&req
, 0, sizeof(req
));
2113 req
.opcode
= SD_OP_READ_VDIS
;
2114 req
.data_length
= max
;
2116 ret
= do_req(fd
, (SheepdogReq
*)&req
, vdi_inuse
, &wlen
, &rlen
);
2123 sn_tab
= g_malloc0(nr
* sizeof(*sn_tab
));
2125 /* calculate a vdi id with hash function */
2126 hval
= fnv_64a_buf(s
->name
, strlen(s
->name
), FNV1A_64_INIT
);
2127 start_nr
= hval
& (SD_NR_VDIS
- 1);
2129 fd
= connect_to_sdog(s
);
2135 for (vid
= start_nr
; found
< nr
; vid
= (vid
+ 1) % SD_NR_VDIS
) {
2136 if (!test_bit(vid
, vdi_inuse
)) {
2140 /* we don't need to read entire object */
2141 ret
= read_object(fd
, (char *)&inode
, vid_to_vdi_oid(vid
),
2142 0, SD_INODE_SIZE
- sizeof(inode
.data_vdi_id
), 0,
2149 if (!strcmp(inode
.name
, s
->name
) && is_snapshot(&inode
)) {
2150 sn_tab
[found
].date_sec
= inode
.snap_ctime
>> 32;
2151 sn_tab
[found
].date_nsec
= inode
.snap_ctime
& 0xffffffff;
2152 sn_tab
[found
].vm_state_size
= inode
.vm_state_size
;
2153 sn_tab
[found
].vm_clock_nsec
= inode
.vm_clock_nsec
;
2155 snprintf(sn_tab
[found
].id_str
, sizeof(sn_tab
[found
].id_str
), "%u",
2157 pstrcpy(sn_tab
[found
].name
,
2158 MIN(sizeof(sn_tab
[found
].name
), sizeof(inode
.tag
)),
2177 static int do_load_save_vmstate(BDRVSheepdogState
*s
, uint8_t *data
,
2178 int64_t pos
, int size
, int load
)
2181 int fd
, ret
= 0, remaining
= size
;
2182 unsigned int data_len
;
2183 uint64_t vmstate_oid
;
2186 uint32_t vdi_id
= load
? s
->inode
.parent_vdi_id
: s
->inode
.vdi_id
;
2188 fd
= connect_to_sdog(s
);
2194 vdi_index
= pos
/ SD_DATA_OBJ_SIZE
;
2195 offset
= pos
% SD_DATA_OBJ_SIZE
;
2197 data_len
= MIN(remaining
, SD_DATA_OBJ_SIZE
- offset
);
2199 vmstate_oid
= vid_to_vmstate_oid(vdi_id
, vdi_index
);
2201 create
= (offset
== 0);
2203 ret
= read_object(fd
, (char *)data
, vmstate_oid
,
2204 s
->inode
.nr_copies
, data_len
, offset
,
2207 ret
= write_object(fd
, (char *)data
, vmstate_oid
,
2208 s
->inode
.nr_copies
, data_len
, offset
, create
,
2213 error_report("failed to save vmstate %s", strerror(errno
));
2219 remaining
-= data_len
;
2227 static int sd_save_vmstate(BlockDriverState
*bs
, QEMUIOVector
*qiov
,
2230 BDRVSheepdogState
*s
= bs
->opaque
;
2234 buf
= qemu_blockalign(bs
, qiov
->size
);
2235 qemu_iovec_to_buf(qiov
, 0, buf
, qiov
->size
);
2236 ret
= do_load_save_vmstate(s
, (uint8_t *) buf
, pos
, qiov
->size
, 0);
2242 static int sd_load_vmstate(BlockDriverState
*bs
, uint8_t *data
,
2243 int64_t pos
, int size
)
2245 BDRVSheepdogState
*s
= bs
->opaque
;
2247 return do_load_save_vmstate(s
, data
, pos
, size
, 1);
2251 static coroutine_fn
int sd_co_discard(BlockDriverState
*bs
, int64_t sector_num
,
2256 BDRVSheepdogState
*s
= bs
->opaque
;
2259 if (!s
->discard_supported
) {
2263 acb
= sd_aio_setup(bs
, &dummy
, sector_num
, nb_sectors
);
2264 acb
->aiocb_type
= AIOCB_DISCARD_OBJ
;
2265 acb
->aio_done_func
= sd_finish_aiocb
;
2267 ret
= sd_co_rw_vector(acb
);
2269 qemu_aio_release(acb
);
2273 qemu_coroutine_yield();
2278 static coroutine_fn
int64_t
2279 sd_co_get_block_status(BlockDriverState
*bs
, int64_t sector_num
, int nb_sectors
,
2282 BDRVSheepdogState
*s
= bs
->opaque
;
2283 SheepdogInode
*inode
= &s
->inode
;
2284 unsigned long start
= sector_num
* BDRV_SECTOR_SIZE
/ SD_DATA_OBJ_SIZE
,
2285 end
= DIV_ROUND_UP((sector_num
+ nb_sectors
) *
2286 BDRV_SECTOR_SIZE
, SD_DATA_OBJ_SIZE
);
2288 int64_t ret
= BDRV_BLOCK_DATA
;
2290 for (idx
= start
; idx
< end
; idx
++) {
2291 if (inode
->data_vdi_id
[idx
] == 0) {
2296 /* Get the longest length of unallocated sectors */
2298 for (idx
= start
+ 1; idx
< end
; idx
++) {
2299 if (inode
->data_vdi_id
[idx
] != 0) {
2305 *pnum
= (idx
- start
) * SD_DATA_OBJ_SIZE
/ BDRV_SECTOR_SIZE
;
2306 if (*pnum
> nb_sectors
) {
2312 static QEMUOptionParameter sd_create_options
[] = {
2314 .name
= BLOCK_OPT_SIZE
,
2316 .help
= "Virtual disk size"
2319 .name
= BLOCK_OPT_BACKING_FILE
,
2321 .help
= "File name of a base image"
2324 .name
= BLOCK_OPT_PREALLOC
,
2326 .help
= "Preallocation mode (allowed values: off, full)"
2331 static BlockDriver bdrv_sheepdog
= {
2332 .format_name
= "sheepdog",
2333 .protocol_name
= "sheepdog",
2334 .instance_size
= sizeof(BDRVSheepdogState
),
2335 .bdrv_file_open
= sd_open
,
2336 .bdrv_close
= sd_close
,
2337 .bdrv_create
= sd_create
,
2338 .bdrv_has_zero_init
= bdrv_has_zero_init_1
,
2339 .bdrv_getlength
= sd_getlength
,
2340 .bdrv_truncate
= sd_truncate
,
2342 .bdrv_co_readv
= sd_co_readv
,
2343 .bdrv_co_writev
= sd_co_writev
,
2344 .bdrv_co_flush_to_disk
= sd_co_flush_to_disk
,
2345 .bdrv_co_discard
= sd_co_discard
,
2346 .bdrv_co_get_block_status
= sd_co_get_block_status
,
2348 .bdrv_snapshot_create
= sd_snapshot_create
,
2349 .bdrv_snapshot_goto
= sd_snapshot_goto
,
2350 .bdrv_snapshot_delete
= sd_snapshot_delete
,
2351 .bdrv_snapshot_list
= sd_snapshot_list
,
2353 .bdrv_save_vmstate
= sd_save_vmstate
,
2354 .bdrv_load_vmstate
= sd_load_vmstate
,
2356 .create_options
= sd_create_options
,
2359 static BlockDriver bdrv_sheepdog_tcp
= {
2360 .format_name
= "sheepdog",
2361 .protocol_name
= "sheepdog+tcp",
2362 .instance_size
= sizeof(BDRVSheepdogState
),
2363 .bdrv_file_open
= sd_open
,
2364 .bdrv_close
= sd_close
,
2365 .bdrv_create
= sd_create
,
2366 .bdrv_has_zero_init
= bdrv_has_zero_init_1
,
2367 .bdrv_getlength
= sd_getlength
,
2368 .bdrv_truncate
= sd_truncate
,
2370 .bdrv_co_readv
= sd_co_readv
,
2371 .bdrv_co_writev
= sd_co_writev
,
2372 .bdrv_co_flush_to_disk
= sd_co_flush_to_disk
,
2373 .bdrv_co_discard
= sd_co_discard
,
2374 .bdrv_co_get_block_status
= sd_co_get_block_status
,
2376 .bdrv_snapshot_create
= sd_snapshot_create
,
2377 .bdrv_snapshot_goto
= sd_snapshot_goto
,
2378 .bdrv_snapshot_delete
= sd_snapshot_delete
,
2379 .bdrv_snapshot_list
= sd_snapshot_list
,
2381 .bdrv_save_vmstate
= sd_save_vmstate
,
2382 .bdrv_load_vmstate
= sd_load_vmstate
,
2384 .create_options
= sd_create_options
,
2387 static BlockDriver bdrv_sheepdog_unix
= {
2388 .format_name
= "sheepdog",
2389 .protocol_name
= "sheepdog+unix",
2390 .instance_size
= sizeof(BDRVSheepdogState
),
2391 .bdrv_file_open
= sd_open
,
2392 .bdrv_close
= sd_close
,
2393 .bdrv_create
= sd_create
,
2394 .bdrv_has_zero_init
= bdrv_has_zero_init_1
,
2395 .bdrv_getlength
= sd_getlength
,
2396 .bdrv_truncate
= sd_truncate
,
2398 .bdrv_co_readv
= sd_co_readv
,
2399 .bdrv_co_writev
= sd_co_writev
,
2400 .bdrv_co_flush_to_disk
= sd_co_flush_to_disk
,
2401 .bdrv_co_discard
= sd_co_discard
,
2402 .bdrv_co_get_block_status
= sd_co_get_block_status
,
2404 .bdrv_snapshot_create
= sd_snapshot_create
,
2405 .bdrv_snapshot_goto
= sd_snapshot_goto
,
2406 .bdrv_snapshot_delete
= sd_snapshot_delete
,
2407 .bdrv_snapshot_list
= sd_snapshot_list
,
2409 .bdrv_save_vmstate
= sd_save_vmstate
,
2410 .bdrv_load_vmstate
= sd_load_vmstate
,
2412 .create_options
= sd_create_options
,
2415 static void bdrv_sheepdog_init(void)
2417 bdrv_register(&bdrv_sheepdog
);
2418 bdrv_register(&bdrv_sheepdog_tcp
);
2419 bdrv_register(&bdrv_sheepdog_unix
);
2421 block_init(bdrv_sheepdog_init
);